From 891456227881da9c565c455010380a40d385a478 Mon Sep 17 00:00:00 2001 From: Marcin Slusarz Date: Wed, 13 Feb 2008 15:34:20 -0600 Subject: jfs: le*_add_cpu conversion replace all: little_endian_variable = cpu_to_leX(leX_to_cpu(little_endian_variable) + expression_in_cpu_byteorder); with: leX_add_cpu(&little_endian_variable, expression_in_cpu_byteorder); generated with semantic patch Signed-off-by: Marcin Slusarz Signed-off-by: Dave Kleikamp Cc: jfs-discussion@lists.sourceforge.net --- fs/jfs/jfs_dmap.c | 11 +++++------ fs/jfs/jfs_imap.c | 15 ++++++--------- fs/jfs/jfs_xtree.c | 26 ++++++++------------------ 3 files changed, 19 insertions(+), 33 deletions(-) (limited to 'fs') diff --git a/fs/jfs/jfs_dmap.c b/fs/jfs/jfs_dmap.c index e1985066b1c..2bc7d8aa574 100644 --- a/fs/jfs/jfs_dmap.c +++ b/fs/jfs/jfs_dmap.c @@ -2172,7 +2172,7 @@ static void dbAllocBits(struct bmap * bmp, struct dmap * dp, s64 blkno, } /* update the free count for this dmap */ - dp->nfree = cpu_to_le32(le32_to_cpu(dp->nfree) - nblocks); + le32_add_cpu(&dp->nfree, -nblocks); BMAP_LOCK(bmp); @@ -2316,7 +2316,7 @@ static int dbFreeBits(struct bmap * bmp, struct dmap * dp, s64 blkno, /* update the free count for this dmap. */ - dp->nfree = cpu_to_le32(le32_to_cpu(dp->nfree) + nblocks); + le32_add_cpu(&dp->nfree, nblocks); BMAP_LOCK(bmp); @@ -3226,7 +3226,7 @@ static int dbAllocDmapBU(struct bmap * bmp, struct dmap * dp, s64 blkno, } /* update the free count for this dmap */ - dp->nfree = cpu_to_le32(le32_to_cpu(dp->nfree) - nblocks); + le32_add_cpu(&dp->nfree, -nblocks); /* reconstruct summary tree */ dbInitDmapTree(dp); @@ -3660,9 +3660,8 @@ static int dbInitDmap(struct dmap * dp, s64 Blkno, int nblocks) goto initTree; } } else { - dp->nblocks = - cpu_to_le32(le32_to_cpu(dp->nblocks) + nblocks); - dp->nfree = cpu_to_le32(le32_to_cpu(dp->nfree) + nblocks); + le32_add_cpu(&dp->nblocks, nblocks); + le32_add_cpu(&dp->nfree, nblocks); } /* word number containing start block number */ diff --git a/fs/jfs/jfs_imap.c b/fs/jfs/jfs_imap.c index 9bf29f77173..734ec916bea 100644 --- a/fs/jfs/jfs_imap.c +++ b/fs/jfs/jfs_imap.c @@ -1019,8 +1019,7 @@ int diFree(struct inode *ip) /* update the free inode counts at the iag, ag and * map level. */ - iagp->nfreeinos = - cpu_to_le32(le32_to_cpu(iagp->nfreeinos) + 1); + le32_add_cpu(&iagp->nfreeinos, 1); imap->im_agctl[agno].numfree += 1; atomic_inc(&imap->im_numfree); @@ -1219,9 +1218,8 @@ int diFree(struct inode *ip) /* update the number of free inodes and number of free extents * for the iag. */ - iagp->nfreeinos = cpu_to_le32(le32_to_cpu(iagp->nfreeinos) - - (INOSPEREXT - 1)); - iagp->nfreeexts = cpu_to_le32(le32_to_cpu(iagp->nfreeexts) + 1); + le32_add_cpu(&iagp->nfreeinos, -(INOSPEREXT - 1)); + le32_add_cpu(&iagp->nfreeexts, 1); /* update the number of free inodes and backed inodes * at the ag and inode map level. @@ -2124,7 +2122,7 @@ static int diAllocBit(struct inomap * imap, struct iag * iagp, int ino) /* update the free inode count at the iag, ag, inode * map levels. */ - iagp->nfreeinos = cpu_to_le32(le32_to_cpu(iagp->nfreeinos) - 1); + le32_add_cpu(&iagp->nfreeinos, -1); imap->im_agctl[agno].numfree -= 1; atomic_dec(&imap->im_numfree); @@ -2378,9 +2376,8 @@ static int diNewExt(struct inomap * imap, struct iag * iagp, int extno) /* update the free inode and free extent counts for the * iag. */ - iagp->nfreeinos = cpu_to_le32(le32_to_cpu(iagp->nfreeinos) + - (INOSPEREXT - 1)); - iagp->nfreeexts = cpu_to_le32(le32_to_cpu(iagp->nfreeexts) - 1); + le32_add_cpu(&iagp->nfreeinos, (INOSPEREXT - 1)); + le32_add_cpu(&iagp->nfreeexts, -1); /* update the free and backed inode counts for the ag. */ diff --git a/fs/jfs/jfs_xtree.c b/fs/jfs/jfs_xtree.c index a000aaa7513..5a61ebf2cbc 100644 --- a/fs/jfs/jfs_xtree.c +++ b/fs/jfs/jfs_xtree.c @@ -905,8 +905,7 @@ int xtInsert(tid_t tid, /* transaction id */ XT_PUTENTRY(xad, xflag, xoff, xlen, xaddr); /* advance next available entry index */ - p->header.nextindex = - cpu_to_le16(le16_to_cpu(p->header.nextindex) + 1); + le16_add_cpu(&p->header.nextindex, 1); /* Don't log it if there are no links to the file */ if (!test_cflag(COMMIT_Nolink, ip)) { @@ -997,8 +996,7 @@ xtSplitUp(tid_t tid, split->addr); /* advance next available entry index */ - sp->header.nextindex = - cpu_to_le16(le16_to_cpu(sp->header.nextindex) + 1); + le16_add_cpu(&sp->header.nextindex, 1); /* Don't log it if there are no links to the file */ if (!test_cflag(COMMIT_Nolink, ip)) { @@ -1167,9 +1165,7 @@ xtSplitUp(tid_t tid, JFS_SBI(ip->i_sb)->nbperpage, rcbn); /* advance next available entry index. */ - sp->header.nextindex = - cpu_to_le16(le16_to_cpu(sp->header.nextindex) + - 1); + le16_add_cpu(&sp->header.nextindex, 1); /* Don't log it if there are no links to the file */ if (!test_cflag(COMMIT_Nolink, ip)) { @@ -1738,8 +1734,7 @@ int xtExtend(tid_t tid, /* transaction id */ XT_PUTENTRY(xad, XAD_NEW, xoff, len, xaddr); /* advance next available entry index */ - p->header.nextindex = - cpu_to_le16(le16_to_cpu(p->header.nextindex) + 1); + le16_add_cpu(&p->header.nextindex, 1); } /* get back old entry */ @@ -1905,8 +1900,7 @@ printf("xtTailgate: xoff:0x%lx xlen:0x%x xaddr:0x%lx\n", XT_PUTENTRY(xad, XAD_NEW, xoff, xlen, xaddr); /* advance next available entry index */ - p->header.nextindex = - cpu_to_le16(le16_to_cpu(p->header.nextindex) + 1); + le16_add_cpu(&p->header.nextindex, 1); } /* get back old XAD */ @@ -2567,8 +2561,7 @@ int xtAppend(tid_t tid, /* transaction id */ XT_PUTENTRY(xad, xflag, xoff, xlen, xaddr); /* advance next available entry index */ - p->header.nextindex = - cpu_to_le16(le16_to_cpu(p->header.nextindex) + 1); + le16_add_cpu(&p->header.nextindex, 1); xtlck->lwm.offset = (xtlck->lwm.offset) ? min(index,(int) xtlck->lwm.offset) : index; @@ -2631,8 +2624,7 @@ int xtDelete(tid_t tid, struct inode *ip, s64 xoff, s32 xlen, int flag) * delete the entry from the leaf page */ nextindex = le16_to_cpu(p->header.nextindex); - p->header.nextindex = - cpu_to_le16(le16_to_cpu(p->header.nextindex) - 1); + le16_add_cpu(&p->header.nextindex, -1); /* * if the leaf page bocome empty, free the page @@ -2795,9 +2787,7 @@ xtDeleteUp(tid_t tid, struct inode *ip, (nextindex - index - 1) << L2XTSLOTSIZE); - p->header.nextindex = - cpu_to_le16(le16_to_cpu(p->header.nextindex) - - 1); + le16_add_cpu(&p->header.nextindex, -1); jfs_info("xtDeleteUp(entry): 0x%lx[%d]", (ulong) parent->bn, index); } -- cgit v1.2.3 From 2785259631697ebb0749a3782cca206e2e542939 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Mon, 4 Feb 2008 23:48:37 -0800 Subject: nfs: use GFP_NOFS preloads for radix-tree insertion NFS should use GFP_NOFS mode radix tree preloads rather than GFP_ATOMIC allocations at radix-tree insertion-time. This is important to reduce the atomic memory requirement. Signed-off-by: Nick Piggin Cc: Trond Myklebust Cc: "J. Bruce Fields" Signed-off-by: Andrew Morton Signed-off-by: Trond Myklebust --- fs/nfs/write.c | 34 +++++++++++++++++++--------------- 1 file changed, 19 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/nfs/write.c b/fs/nfs/write.c index f55c437124a..7be42e6eb63 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -360,15 +360,13 @@ int nfs_writepages(struct address_space *mapping, struct writeback_control *wbc) /* * Insert a write request into an inode */ -static int nfs_inode_add_request(struct inode *inode, struct nfs_page *req) +static void nfs_inode_add_request(struct inode *inode, struct nfs_page *req) { struct nfs_inode *nfsi = NFS_I(inode); int error; error = radix_tree_insert(&nfsi->nfs_page_tree, req->wb_index, req); - BUG_ON(error == -EEXIST); - if (error) - return error; + BUG_ON(error); if (!nfsi->npages) { igrab(inode); if (nfs_have_delegation(inode, FMODE_WRITE)) @@ -378,8 +376,8 @@ static int nfs_inode_add_request(struct inode *inode, struct nfs_page *req) set_page_private(req->wb_page, (unsigned long)req); nfsi->npages++; kref_get(&req->wb_kref); - radix_tree_tag_set(&nfsi->nfs_page_tree, req->wb_index, NFS_PAGE_TAG_LOCKED); - return 0; + radix_tree_tag_set(&nfsi->nfs_page_tree, req->wb_index, + NFS_PAGE_TAG_LOCKED); } /* @@ -591,6 +589,13 @@ static struct nfs_page * nfs_update_request(struct nfs_open_context* ctx, /* Loop over all inode entries and see if we find * A request for the page we wish to update */ + if (new) { + if (radix_tree_preload(GFP_NOFS)) { + nfs_release_request(new); + return ERR_PTR(-ENOMEM); + } + } + spin_lock(&inode->i_lock); req = nfs_page_find_request_locked(page); if (req) { @@ -601,28 +606,27 @@ static struct nfs_page * nfs_update_request(struct nfs_open_context* ctx, error = nfs_wait_on_request(req); nfs_release_request(req); if (error < 0) { - if (new) + if (new) { + radix_tree_preload_end(); nfs_release_request(new); + } return ERR_PTR(error); } continue; } spin_unlock(&inode->i_lock); - if (new) + if (new) { + radix_tree_preload_end(); nfs_release_request(new); + } break; } if (new) { - int error; nfs_lock_request_dontget(new); - error = nfs_inode_add_request(inode, new); - if (error) { - spin_unlock(&inode->i_lock); - nfs_unlock_request(new); - return ERR_PTR(error); - } + nfs_inode_add_request(inode, new); spin_unlock(&inode->i_lock); + radix_tree_preload_end(); req = new; goto zero_page; } -- cgit v1.2.3 From b28ba9fa0154f78f3d36f5ae9a42f7bb01663cca Mon Sep 17 00:00:00 2001 From: David Woodhouse Date: Mon, 25 Feb 2008 15:20:50 +0000 Subject: [JFFS2] Set i_blocks when truncating files Addresses OLPC trac #6480 Signed-off-by: David Woodhouse --- fs/jffs2/fs.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/jffs2/fs.c b/fs/jffs2/fs.c index e26ea78c789..3f49562dc50 100644 --- a/fs/jffs2/fs.c +++ b/fs/jffs2/fs.c @@ -149,6 +149,7 @@ int jffs2_do_setattr (struct inode *inode, struct iattr *iattr) if (ivalid & ATTR_SIZE && inode->i_size < iattr->ia_size) { jffs2_add_full_dnode_to_inode(c, f, new_metadata); inode->i_size = iattr->ia_size; + inode->i_blocks = (inode->i_size + 511) >> 9; f->metadata = NULL; } else { f->metadata = new_metadata; @@ -167,8 +168,10 @@ int jffs2_do_setattr (struct inode *inode, struct iattr *iattr) We are protected from a simultaneous write() extending i_size back past iattr->ia_size, because do_truncate() holds the generic inode semaphore. */ - if (ivalid & ATTR_SIZE && inode->i_size > iattr->ia_size) - vmtruncate(inode, iattr->ia_size); + if (ivalid & ATTR_SIZE && inode->i_size > iattr->ia_size) { + vmtruncate(inode, iattr->ia_size); + inode->i_blocks = (inode->i_size + 511) >> 9; + } return 0; } -- cgit v1.2.3 From dd919660aacdf4adfcd279556aa03e595f7f0fc2 Mon Sep 17 00:00:00 2001 From: David Woodhouse Date: Mon, 25 Feb 2008 15:25:25 +0000 Subject: [JFFS2] Use ALLOC_DELETION priority for truncation to zero length This is going to obsolete all previous nodes, so treat it as deletion. Signed-off-by: David Woodhouse --- fs/jffs2/fs.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/jffs2/fs.c b/fs/jffs2/fs.c index 3f49562dc50..9dafb53fb1d 100644 --- a/fs/jffs2/fs.c +++ b/fs/jffs2/fs.c @@ -36,6 +36,7 @@ int jffs2_do_setattr (struct inode *inode, struct iattr *iattr) unsigned int ivalid; uint32_t alloclen; int ret; + int alloc_type = ALLOC_NORMAL; D1(printk(KERN_DEBUG "jffs2_setattr(): ino #%lu\n", inode->i_ino)); @@ -115,6 +116,10 @@ int jffs2_do_setattr (struct inode *inode, struct iattr *iattr) ri->compr = JFFS2_COMPR_ZERO; ri->dsize = cpu_to_je32(iattr->ia_size - inode->i_size); ri->offset = cpu_to_je32(inode->i_size); + } else if (ivalid & ATTR_SIZE && !iattr->ia_size) { + /* For truncate-to-zero, treat it as deletion because + it'll always be obsoleting all previous nodes */ + alloc_type = ALLOC_DELETION; } ri->node_crc = cpu_to_je32(crc32(0, ri, sizeof(*ri)-8)); if (mdatalen) @@ -122,7 +127,7 @@ int jffs2_do_setattr (struct inode *inode, struct iattr *iattr) else ri->data_crc = cpu_to_je32(0); - new_metadata = jffs2_write_dnode(c, f, ri, mdata, mdatalen, ALLOC_NORMAL); + new_metadata = jffs2_write_dnode(c, f, ri, mdata, mdatalen, alloc_type); if (S_ISLNK(inode->i_mode)) kfree(mdata); -- cgit v1.2.3 From 4b5621f6b127bce9218998c187bd25bf7f9fc371 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 25 Feb 2008 15:56:29 -0800 Subject: NFS: Fix an f_mode/f_flags confusion in fs/nfs/write.c O_SYNC is stored in filp->f_flags. Thanks to Al Viro for pointing out the bug. Signed-off-by: Trond Myklebust --- fs/nfs/write.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/write.c b/fs/nfs/write.c index f55c437124a..80c61fdb272 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -734,7 +734,7 @@ int nfs_updatepage(struct file *file, struct page *page, */ if (nfs_write_pageuptodate(page, inode) && inode->i_flock == NULL && - !(file->f_mode & O_SYNC)) { + !(file->f_flags & O_SYNC)) { count = max(count + offset, nfs_page_length(page)); offset = 0; } -- cgit v1.2.3 From 383ba71938519959be8e0b598ec658f0c211ff45 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 19 Feb 2008 20:04:20 -0500 Subject: NFS: Fix a deadlock with lazy umount We can't allow rpc callback functions like task->tk_ops->rpc_call_prepare() and task->tk_ops->rpc_call_done() to call mntput() in any way, since that will cause a deadlock when the call to rpc_shutdown_client() attempts to wait on 'task' to complete. We can avoid the above deadlock by moving calls to mntput to task->tk_ops->rpc_release() callback, since at that time the task will be marked as completed, and so rpc_shutdown_client won't attempt to wait on it. Signed-off-by: Trond Myklebust --- fs/nfs/direct.c | 5 +++-- fs/nfs/inode.c | 6 +++++- fs/nfs/read.c | 7 +++++-- fs/nfs/write.c | 13 ++++++++++--- 4 files changed, 23 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c index 16844f98f50..e0170407a88 100644 --- a/fs/nfs/direct.c +++ b/fs/nfs/direct.c @@ -323,7 +323,7 @@ static ssize_t nfs_direct_read_schedule_segment(struct nfs_direct_req *dreq, data->inode = inode; data->cred = msg.rpc_cred; data->args.fh = NFS_FH(inode); - data->args.context = ctx; + data->args.context = get_nfs_open_context(ctx); data->args.offset = pos; data->args.pgbase = pgbase; data->args.pages = data->pagevec; @@ -546,6 +546,7 @@ static void nfs_direct_commit_schedule(struct nfs_direct_req *dreq) data->args.fh = NFS_FH(data->inode); data->args.offset = 0; data->args.count = 0; + data->args.context = get_nfs_open_context(dreq->ctx); data->res.count = 0; data->res.fattr = &data->fattr; data->res.verf = &data->verf; @@ -728,7 +729,7 @@ static ssize_t nfs_direct_write_schedule_segment(struct nfs_direct_req *dreq, data->inode = inode; data->cred = msg.rpc_cred; data->args.fh = NFS_FH(inode); - data->args.context = ctx; + data->args.context = get_nfs_open_context(ctx); data->args.offset = pos; data->args.pgbase = pgbase; data->args.pages = data->pagevec; diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 966a8850aa3..a499fb58d85 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -521,8 +521,12 @@ struct nfs_open_context *get_nfs_open_context(struct nfs_open_context *ctx) static void __put_nfs_open_context(struct nfs_open_context *ctx, int wait) { - struct inode *inode = ctx->path.dentry->d_inode; + struct inode *inode; + if (ctx == NULL) + return; + + inode = ctx->path.dentry->d_inode; if (!atomic_dec_and_lock(&ctx->count, &inode->i_lock)) return; list_del(&ctx->list); diff --git a/fs/nfs/read.c b/fs/nfs/read.c index 3d7d9631e12..fab0d3720a0 100644 --- a/fs/nfs/read.c +++ b/fs/nfs/read.c @@ -73,7 +73,10 @@ static void nfs_readdata_free(struct nfs_read_data *rdata) void nfs_readdata_release(void *data) { - nfs_readdata_free(data); + struct nfs_read_data *rdata = data; + + put_nfs_open_context(rdata->args.context); + nfs_readdata_free(rdata); } static @@ -186,7 +189,7 @@ static void nfs_read_rpcsetup(struct nfs_page *req, struct nfs_read_data *data, data->args.pgbase = req->wb_pgbase + offset; data->args.pages = data->pagevec; data->args.count = count; - data->args.context = req->wb_context; + data->args.context = get_nfs_open_context(req->wb_context); data->res.fattr = &data->fattr; data->res.count = count; diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 80c61fdb272..69b4158d9a1 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -105,8 +105,11 @@ static void nfs_writedata_free(struct nfs_write_data *wdata) call_rcu_bh(&wdata->task.u.tk_rcu, nfs_writedata_rcu_free); } -void nfs_writedata_release(void *wdata) +void nfs_writedata_release(void *data) { + struct nfs_write_data *wdata = data; + + put_nfs_open_context(wdata->args.context); nfs_writedata_free(wdata); } @@ -816,7 +819,7 @@ static void nfs_write_rpcsetup(struct nfs_page *req, data->args.pgbase = req->wb_pgbase + offset; data->args.pages = data->pagevec; data->args.count = count; - data->args.context = req->wb_context; + data->args.context = get_nfs_open_context(req->wb_context); data->args.stable = NFS_UNSTABLE; if (how & FLUSH_STABLE) { data->args.stable = NFS_DATA_SYNC; @@ -1153,8 +1156,11 @@ int nfs_writeback_done(struct rpc_task *task, struct nfs_write_data *data) #if defined(CONFIG_NFS_V3) || defined(CONFIG_NFS_V4) -void nfs_commit_release(void *wdata) +void nfs_commit_release(void *data) { + struct nfs_write_data *wdata = data; + + put_nfs_open_context(wdata->args.context); nfs_commit_free(wdata); } @@ -1197,6 +1203,7 @@ static void nfs_commit_rpcsetup(struct list_head *head, /* Note: we always request a commit of the entire inode */ data->args.offset = 0; data->args.count = 0; + data->args.context = get_nfs_open_context(first->wb_context); data->res.count = 0; data->res.fattr = &data->fattr; data->res.verf = &data->verf; -- cgit v1.2.3 From 5746006f1d17d9d5a3015051ea54de4341cb31f9 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 19 Feb 2008 20:04:22 -0500 Subject: NFS: Add an nfsiod workqueue NFS post-rpciod cleanups often involve tasks that cannot be safely performed within the rpciod context (due to deadlock concerns). We therefore add a dedicated NFS workqueue that can perform tasks like cleaning up state after an interrupted NFSv4 open() call, or calling put_nfs_open_context() after an asynchronous read or write call. Signed-off-by: Trond Myklebust --- fs/nfs/inode.c | 37 +++++++++++++++++++++++++++++++++++++ fs/nfs/internal.h | 1 + 2 files changed, 38 insertions(+) (limited to 'fs') diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index a499fb58d85..36cb99985d2 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -1219,6 +1219,36 @@ static void nfs_destroy_inodecache(void) kmem_cache_destroy(nfs_inode_cachep); } +struct workqueue_struct *nfsiod_workqueue; + +/* + * start up the nfsiod workqueue + */ +static int nfsiod_start(void) +{ + struct workqueue_struct *wq; + dprintk("RPC: creating workqueue nfsiod\n"); + wq = create_singlethread_workqueue("nfsiod"); + if (wq == NULL) + return -ENOMEM; + nfsiod_workqueue = wq; + return 0; +} + +/* + * Destroy the nfsiod workqueue + */ +static void nfsiod_stop(void) +{ + struct workqueue_struct *wq; + + wq = nfsiod_workqueue; + if (wq == NULL) + return; + nfsiod_workqueue = NULL; + destroy_workqueue(wq); +} + /* * Initialize NFS */ @@ -1226,6 +1256,10 @@ static int __init init_nfs_fs(void) { int err; + err = nfsiod_start(); + if (err) + goto out6; + err = nfs_fs_proc_init(); if (err) goto out5; @@ -1272,6 +1306,8 @@ out3: out4: nfs_fs_proc_exit(); out5: + nfsiod_stop(); +out6: return err; } @@ -1287,6 +1323,7 @@ static void __exit exit_nfs_fs(void) #endif unregister_nfs_fs(); nfs_fs_proc_exit(); + nfsiod_stop(); } /* Not quite true; I just maintain it */ diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index 0f5619611b8..985dc293103 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -143,6 +143,7 @@ extern struct rpc_procinfo nfs4_procedures[]; extern int nfs_access_cache_shrinker(int nr_to_scan, gfp_t gfp_mask); /* inode.c */ +extern struct workqueue_struct *nfsiod_workqueue; extern struct inode *nfs_alloc_inode(struct super_block *sb); extern void nfs_destroy_inode(struct inode *); extern int nfs_write_inode(struct inode *,int); -- cgit v1.2.3 From 101070ca2fe67186f5f5517b66cb4757b17f4e29 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 19 Feb 2008 20:04:23 -0500 Subject: NFS: Ensure that the asynchronous RPC calls complete on nfsiod. We want to ensure that rpc_call_ops that involve mntput() are run on nfsiod rather than on rpciod, so that they don't deadlock when the resulting umount calls rpc_shutdown_client(). Hence we specify that read, write and commit calls must complete on nfsiod. Ditto for NFSv4 open, lock, locku and close asynchronous calls. Signed-off-by: Trond Myklebust --- fs/nfs/direct.c | 4 ++++ fs/nfs/nfs4proc.c | 6 ++++++ fs/nfs/read.c | 1 + fs/nfs/write.c | 2 ++ 4 files changed, 13 insertions(+) (limited to 'fs') diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c index e0170407a88..e44200579c8 100644 --- a/fs/nfs/direct.c +++ b/fs/nfs/direct.c @@ -280,6 +280,7 @@ static ssize_t nfs_direct_read_schedule_segment(struct nfs_direct_req *dreq, .rpc_client = NFS_CLIENT(inode), .rpc_message = &msg, .callback_ops = &nfs_read_direct_ops, + .workqueue = nfsiod_workqueue, .flags = RPC_TASK_ASYNC, }; unsigned int pgbase; @@ -446,6 +447,7 @@ static void nfs_direct_write_reschedule(struct nfs_direct_req *dreq) struct rpc_task_setup task_setup_data = { .rpc_client = NFS_CLIENT(inode), .callback_ops = &nfs_write_direct_ops, + .workqueue = nfsiod_workqueue, .flags = RPC_TASK_ASYNC, }; @@ -537,6 +539,7 @@ static void nfs_direct_commit_schedule(struct nfs_direct_req *dreq) .rpc_message = &msg, .callback_ops = &nfs_commit_direct_ops, .callback_data = data, + .workqueue = nfsiod_workqueue, .flags = RPC_TASK_ASYNC, }; @@ -683,6 +686,7 @@ static ssize_t nfs_direct_write_schedule_segment(struct nfs_direct_req *dreq, .rpc_client = NFS_CLIENT(inode), .rpc_message = &msg, .callback_ops = &nfs_write_direct_ops, + .workqueue = nfsiod_workqueue, .flags = RPC_TASK_ASYNC, }; size_t wsize = NFS_SERVER(inode)->wsize; diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 7ce07862c2f..b6db833f33e 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -51,6 +51,7 @@ #include "nfs4_fs.h" #include "delegation.h" +#include "internal.h" #include "iostat.h" #define NFSDBG_FACILITY NFSDBG_PROC @@ -773,6 +774,7 @@ static int _nfs4_proc_open_confirm(struct nfs4_opendata *data) .rpc_message = &msg, .callback_ops = &nfs4_open_confirm_ops, .callback_data = data, + .workqueue = nfsiod_workqueue, .flags = RPC_TASK_ASYNC, }; int status; @@ -910,6 +912,7 @@ static int _nfs4_proc_open(struct nfs4_opendata *data) .rpc_message = &msg, .callback_ops = &nfs4_open_ops, .callback_data = data, + .workqueue = nfsiod_workqueue, .flags = RPC_TASK_ASYNC, }; int status; @@ -1315,6 +1318,7 @@ int nfs4_do_close(struct path *path, struct nfs4_state *state, int wait) .rpc_client = server->client, .rpc_message = &msg, .callback_ops = &nfs4_close_ops, + .workqueue = nfsiod_workqueue, .flags = RPC_TASK_ASYNC, }; int status = -ENOMEM; @@ -3235,6 +3239,7 @@ static struct rpc_task *nfs4_do_unlck(struct file_lock *fl, .rpc_client = NFS_CLIENT(lsp->ls_state->inode), .rpc_message = &msg, .callback_ops = &nfs4_locku_ops, + .workqueue = nfsiod_workqueue, .flags = RPC_TASK_ASYNC, }; @@ -3419,6 +3424,7 @@ static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *f .rpc_client = NFS_CLIENT(state->inode), .rpc_message = &msg, .callback_ops = &nfs4_lock_ops, + .workqueue = nfsiod_workqueue, .flags = RPC_TASK_ASYNC, }; int ret; diff --git a/fs/nfs/read.c b/fs/nfs/read.c index fab0d3720a0..87546cd277d 100644 --- a/fs/nfs/read.c +++ b/fs/nfs/read.c @@ -177,6 +177,7 @@ static void nfs_read_rpcsetup(struct nfs_page *req, struct nfs_read_data *data, .rpc_message = &msg, .callback_ops = call_ops, .callback_data = data, + .workqueue = nfsiod_workqueue, .flags = RPC_TASK_ASYNC | swap_flags, }; diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 69b4158d9a1..84495ef9072 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -803,6 +803,7 @@ static void nfs_write_rpcsetup(struct nfs_page *req, .rpc_message = &msg, .callback_ops = call_ops, .callback_data = data, + .workqueue = nfsiod_workqueue, .flags = flags, .priority = priority, }; @@ -1187,6 +1188,7 @@ static void nfs_commit_rpcsetup(struct list_head *head, .rpc_message = &msg, .callback_ops = &nfs_commit_ops, .callback_data = data, + .workqueue = nfsiod_workqueue, .flags = flags, .priority = priority, }; -- cgit v1.2.3 From fda1393938035559b417dd5b26b9cc293a7aee00 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 22 Feb 2008 16:34:12 -0500 Subject: SUNRPC: Convert users of rpc_wake_up_task to use rpc_wake_up_queued_task Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index b6db833f33e..54743396b66 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -2768,7 +2768,7 @@ nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server) rpc_sleep_on(&clp->cl_rpcwaitq, task, NULL, NULL); nfs4_schedule_state_recovery(clp); if (test_bit(NFS4CLNT_STATE_RECOVER, &clp->cl_state) == 0) - rpc_wake_up_task(task); + rpc_wake_up_queued_task(&clp->cl_rpcwaitq, task); task->tk_status = 0; return -EAGAIN; case -NFS4ERR_DELAY: -- cgit v1.2.3 From 5d00837b90340af9106dcd93af75fd664c8eb87f Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 22 Feb 2008 16:34:17 -0500 Subject: SUNRPC: Run rpc timeout functions as callbacks instead of in softirqs An audit of the current RPC timeout functions shows that they don't really ever need to run in the softirq context. As long as the softirq is able to signal that the wakeup is due to a timeout (which it can do by setting task->tk_status to -ETIMEDOUT) then the callback functions can just run as standard task->tk_callback functions (in the rpciod/process context). The only possible border-line case would be xprt_timer() for the case of UDP, when the callback is used to reduce the size of the transport congestion window. In testing, however, the effect of moving that update to a callback would appear to be minor. Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 2 +- fs/nfs/nfs4state.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 54743396b66..bbb0d58ee6a 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -2765,7 +2765,7 @@ nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server) case -NFS4ERR_STALE_CLIENTID: case -NFS4ERR_STALE_STATEID: case -NFS4ERR_EXPIRED: - rpc_sleep_on(&clp->cl_rpcwaitq, task, NULL, NULL); + rpc_sleep_on(&clp->cl_rpcwaitq, task, NULL); nfs4_schedule_state_recovery(clp); if (test_bit(NFS4CLNT_STATE_RECOVER, &clp->cl_state) == 0) rpc_wake_up_queued_task(&clp->cl_rpcwaitq, task); diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index b962397004c..a2ef02824aa 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -731,7 +731,7 @@ int nfs_wait_on_sequence(struct nfs_seqid *seqid, struct rpc_task *task) list_add_tail(&seqid->list, &sequence->list); if (list_first_entry(&sequence->list, struct nfs_seqid, list) == seqid) goto unlock; - rpc_sleep_on(&sequence->wait, task, NULL, NULL); + rpc_sleep_on(&sequence->wait, task, NULL); status = -EAGAIN; unlock: spin_unlock(&sequence->lock); -- cgit v1.2.3 From f6a1cc89309f0ae847a9b6fe418d1c4215e5bc55 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 22 Feb 2008 17:06:55 -0500 Subject: SUNRPC: Add a (empty for the moment) destructor for rpc_wait_queues Signed-off-by: Trond Myklebust --- fs/nfs/client.c | 2 ++ fs/nfs/nfs4state.c | 6 +++++- 2 files changed, 7 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/client.c b/fs/nfs/client.c index c5c0175898f..06f064d8fbb 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -170,6 +170,8 @@ static void nfs4_shutdown_client(struct nfs_client *clp) BUG_ON(!RB_EMPTY_ROOT(&clp->cl_state_owners)); if (__test_and_clear_bit(NFS_CS_IDMAP, &clp->cl_res_state)) nfs_idmap_delete(clp); + + rpc_destroy_wait_queue(&clp->cl_rpcwaitq); #endif } diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index a2ef02824aa..7775435ea7a 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -292,8 +292,10 @@ struct nfs4_state_owner *nfs4_get_state_owner(struct nfs_server *server, struct spin_unlock(&clp->cl_lock); if (sp == new) get_rpccred(cred); - else + else { + rpc_destroy_wait_queue(&new->so_sequence.wait); kfree(new); + } return sp; } @@ -310,6 +312,7 @@ void nfs4_put_state_owner(struct nfs4_state_owner *sp) return; nfs4_remove_state_owner(clp, sp); spin_unlock(&clp->cl_lock); + rpc_destroy_wait_queue(&sp->so_sequence.wait); put_rpccred(cred); kfree(sp); } @@ -529,6 +532,7 @@ static void nfs4_free_lock_state(struct nfs4_lock_state *lsp) spin_lock(&clp->cl_lock); nfs_free_unique_id(&clp->cl_lockowner_id, &lsp->ls_id); spin_unlock(&clp->cl_lock); + rpc_destroy_wait_queue(&lsp->ls_sequence.wait); kfree(lsp); } -- cgit v1.2.3 From 5e4424af9a1f062c6451681dff24a26e27741cc6 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 25 Feb 2008 21:53:49 -0800 Subject: SUNRPC: Remove now-redundant RCU-safe rpc_task free path Now that we've tightened up the locking rules for RPC queue wakeups, we can remove the RCU-safe kfree calls... Signed-off-by: Trond Myklebust --- fs/nfs/read.c | 8 +------- fs/nfs/write.c | 16 ++-------------- 2 files changed, 3 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/nfs/read.c b/fs/nfs/read.c index 87546cd277d..be9e8270f4d 100644 --- a/fs/nfs/read.c +++ b/fs/nfs/read.c @@ -58,19 +58,13 @@ struct nfs_read_data *nfs_readdata_alloc(unsigned int pagecount) return p; } -static void nfs_readdata_rcu_free(struct rcu_head *head) +static void nfs_readdata_free(struct nfs_read_data *p) { - struct nfs_read_data *p = container_of(head, struct nfs_read_data, task.u.tk_rcu); if (p && (p->pagevec != &p->page_array[0])) kfree(p->pagevec); mempool_free(p, nfs_rdata_mempool); } -static void nfs_readdata_free(struct nfs_read_data *rdata) -{ - call_rcu_bh(&rdata->task.u.tk_rcu, nfs_readdata_rcu_free); -} - void nfs_readdata_release(void *data) { struct nfs_read_data *rdata = data; diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 84495ef9072..30513029c00 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -58,19 +58,13 @@ struct nfs_write_data *nfs_commit_alloc(void) return p; } -static void nfs_commit_rcu_free(struct rcu_head *head) +void nfs_commit_free(struct nfs_write_data *p) { - struct nfs_write_data *p = container_of(head, struct nfs_write_data, task.u.tk_rcu); if (p && (p->pagevec != &p->page_array[0])) kfree(p->pagevec); mempool_free(p, nfs_commit_mempool); } -void nfs_commit_free(struct nfs_write_data *wdata) -{ - call_rcu_bh(&wdata->task.u.tk_rcu, nfs_commit_rcu_free); -} - struct nfs_write_data *nfs_writedata_alloc(unsigned int pagecount) { struct nfs_write_data *p = mempool_alloc(nfs_wdata_mempool, GFP_NOFS); @@ -92,19 +86,13 @@ struct nfs_write_data *nfs_writedata_alloc(unsigned int pagecount) return p; } -static void nfs_writedata_rcu_free(struct rcu_head *head) +static void nfs_writedata_free(struct nfs_write_data *p) { - struct nfs_write_data *p = container_of(head, struct nfs_write_data, task.u.tk_rcu); if (p && (p->pagevec != &p->page_array[0])) kfree(p->pagevec); mempool_free(p, nfs_wdata_mempool); } -static void nfs_writedata_free(struct nfs_write_data *wdata) -{ - call_rcu_bh(&wdata->task.u.tk_rcu, nfs_writedata_rcu_free); -} - void nfs_writedata_release(void *data) { struct nfs_write_data *wdata = data; -- cgit v1.2.3 From 15732a1cb5f9078d460a254449eb59391e531ffc Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 5 Mar 2008 14:38:22 -0600 Subject: jfs: replace __inline with inline Signed-off-by: Harvey Harrison Signed-off-by: Dave Kleikamp --- fs/jfs/jfs_dmap.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/jfs/jfs_dmap.h b/fs/jfs/jfs_dmap.h index 11e6d471b36..1a6eb41569b 100644 --- a/fs/jfs/jfs_dmap.h +++ b/fs/jfs/jfs_dmap.h @@ -61,7 +61,7 @@ * determine the maximum free string for four (lower level) nodes * of the tree. */ -static __inline signed char TREEMAX(signed char *cp) +static inline signed char TREEMAX(signed char *cp) { signed char tmp1, tmp2; -- cgit v1.2.3 From 98a8e3239427051f5d44f2025b398bdcc3918f37 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 12 Mar 2008 12:25:28 -0400 Subject: SUNRPC: Add a helper rpcauth_lookup_generic_cred() The NFSv4 protocol allows clients to negotiate security protocols on the fly in the case where an administrator on the server changes the export settings and/or in the case where we may have a filesystem migration event. Instead of having the NFS client code cache credentials that are tied to a particular AUTH method it is therefore preferable to have a generic credential that can be converted into whatever AUTH is in use by the RPC client when the read/write/sillyrename/... is put on the wire. We do this by means of the new "generic" credential, which basically just caches the minimal information that is needed to look up an RPCSEC_GSS, AUTH_SYS, or AUTH_NULL credential. Signed-off-by: Trond Myklebust --- fs/nfs/dir.c | 2 +- fs/nfs/inode.c | 2 +- fs/nfs/nfs4proc.c | 8 ++++---- fs/nfs/unlink.c | 2 +- 4 files changed, 7 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index 6cea7479c5b..d583654a0b3 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -1966,7 +1966,7 @@ force_lookup: if (!NFS_PROTO(inode)->access) goto out_notsup; - cred = rpcauth_lookupcred(NFS_CLIENT(inode)->cl_auth, 0); + cred = rpc_lookup_cred(); if (!IS_ERR(cred)) { res = nfs_do_access(inode, cred, mask); put_rpccred(cred); diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index c49f6d8b42d..15f787355d2 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -613,7 +613,7 @@ int nfs_open(struct inode *inode, struct file *filp) struct nfs_open_context *ctx; struct rpc_cred *cred; - cred = rpcauth_lookupcred(NFS_CLIENT(inode)->cl_auth, 0); + cred = rpc_lookup_cred(); if (IS_ERR(cred)) return PTR_ERR(cred); ctx = alloc_nfs_open_context(filp->f_path.mnt, filp->f_path.dentry, cred); diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index bbb0d58ee6a..f38d0573be1 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -1408,7 +1408,7 @@ nfs4_atomic_open(struct inode *dir, struct dentry *dentry, struct nameidata *nd) BUG_ON(nd->intent.open.flags & O_CREAT); } - cred = rpcauth_lookupcred(NFS_CLIENT(dir)->cl_auth, 0); + cred = rpc_lookup_cred(); if (IS_ERR(cred)) return (struct dentry *)cred; parent = dentry->d_parent; @@ -1443,7 +1443,7 @@ nfs4_open_revalidate(struct inode *dir, struct dentry *dentry, int openflags, st struct rpc_cred *cred; struct nfs4_state *state; - cred = rpcauth_lookupcred(NFS_CLIENT(dir)->cl_auth, 0); + cred = rpc_lookup_cred(); if (IS_ERR(cred)) return PTR_ERR(cred); state = nfs4_do_open(dir, &path, openflags, NULL, cred); @@ -1660,7 +1660,7 @@ nfs4_proc_setattr(struct dentry *dentry, struct nfs_fattr *fattr, nfs_fattr_init(fattr); - cred = rpcauth_lookupcred(NFS_CLIENT(inode)->cl_auth, 0); + cred = rpc_lookup_cred(); if (IS_ERR(cred)) return PTR_ERR(cred); @@ -1896,7 +1896,7 @@ nfs4_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr, struct rpc_cred *cred; int status = 0; - cred = rpcauth_lookupcred(NFS_CLIENT(dir)->cl_auth, 0); + cred = rpc_lookup_cred(); if (IS_ERR(cred)) { status = PTR_ERR(cred); goto out; diff --git a/fs/nfs/unlink.c b/fs/nfs/unlink.c index 75741536342..3adf8b26646 100644 --- a/fs/nfs/unlink.c +++ b/fs/nfs/unlink.c @@ -234,7 +234,7 @@ nfs_async_unlink(struct inode *dir, struct dentry *dentry) if (data == NULL) goto out; - data->cred = rpcauth_lookupcred(NFS_CLIENT(dir)->cl_auth, 0); + data->cred = rpc_lookup_cred(); if (IS_ERR(data->cred)) { status = PTR_ERR(data->cred); goto out_free; -- cgit v1.2.3 From 2f42b5d043ee271d1e5d30ecd77186b6c4d4e534 Mon Sep 17 00:00:00 2001 From: Fred Isaman Date: Thu, 13 Mar 2008 15:26:30 +0200 Subject: NFS: fix encode_fsinfo_maxsz The previous value was not taking into account space for bitmap array size. Signed-off-by: Fred Isaman Signed-off-by: Benny Halevy Signed-off-by: Trond Myklebust --- fs/nfs/nfs4xdr.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index db1ed9c46ed..37421dd4805 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -110,7 +110,7 @@ static int nfs4_stat_to_errno(int); #define decode_savefh_maxsz (op_decode_hdr_maxsz) #define encode_restorefh_maxsz (op_encode_hdr_maxsz) #define decode_restorefh_maxsz (op_decode_hdr_maxsz) -#define encode_fsinfo_maxsz (op_encode_hdr_maxsz + 2) +#define encode_fsinfo_maxsz (encode_getattr_maxsz) #define decode_fsinfo_maxsz (op_decode_hdr_maxsz + 11) #define encode_renew_maxsz (op_encode_hdr_maxsz + 3) #define decode_renew_maxsz (op_decode_hdr_maxsz) -- cgit v1.2.3 From 6d884e8fc8114dc8877218f06a9a9a1d801901e4 Mon Sep 17 00:00:00 2001 From: Fred Date: Wed, 19 Mar 2008 11:24:38 -0400 Subject: nfs: nfs_redirty_request Both flush functions have the same error handling routine. Pull it out as a function. Signed-off-by: Fred Isaman Signed-off-by: Trond Myklebust --- fs/nfs/write.c | 23 ++++++++++++++--------- 1 file changed, 14 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 4cb88df12f8..ce40cadb15d 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -282,8 +282,6 @@ static int nfs_page_async_flush(struct nfs_pageio_descriptor *pgio, spin_unlock(&inode->i_lock); if (!nfs_pageio_add_request(pgio, req)) { nfs_redirty_request(req); - nfs_end_page_writeback(page); - nfs_clear_page_tag_locked(req); return pgio->pg_error; } return 0; @@ -402,7 +400,7 @@ static void nfs_inode_remove_request(struct nfs_page *req) } static void -nfs_redirty_request(struct nfs_page *req) +nfs_mark_request_dirty(struct nfs_page *req) { __set_page_dirty_nobuffers(req->wb_page); } @@ -456,7 +454,7 @@ int nfs_reschedule_unstable_write(struct nfs_page *req) return 1; } if (test_and_clear_bit(PG_NEED_RESCHED, &req->wb_flags)) { - nfs_redirty_request(req); + nfs_mark_request_dirty(req); return 1; } return 0; @@ -847,6 +845,17 @@ static void nfs_write_rpcsetup(struct nfs_page *req, rpc_put_task(task); } +/* If a nfs_flush_* function fails, it should remove reqs from @head and + * call this on each, which will prepare them to be retried on next + * writeback using standard nfs. + */ +static void nfs_redirty_request(struct nfs_page *req) +{ + nfs_mark_request_dirty(req); + nfs_end_page_writeback(req->wb_page); + nfs_clear_page_tag_locked(req); +} + /* * Generate multiple small requests to write out a single * contiguous dirty area on one page. @@ -902,8 +911,6 @@ out_bad: nfs_writedata_release(data); } nfs_redirty_request(req); - nfs_end_page_writeback(req->wb_page); - nfs_clear_page_tag_locked(req); return -ENOMEM; } @@ -944,8 +951,6 @@ static int nfs_flush_one(struct inode *inode, struct list_head *head, unsigned i req = nfs_list_entry(head->next); nfs_list_remove_request(req); nfs_redirty_request(req); - nfs_end_page_writeback(req->wb_page); - nfs_clear_page_tag_locked(req); } return -ENOMEM; } @@ -1298,7 +1303,7 @@ static void nfs_commit_done(struct rpc_task *task, void *calldata) } /* We have a mismatch. Write the page again */ dprintk(" mismatch\n"); - nfs_redirty_request(req); + nfs_mark_request_dirty(req); next: nfs_clear_page_tag_locked(req); } -- cgit v1.2.3 From 4af68bffac444a23f027e18ff244101e63b79227 Mon Sep 17 00:00:00 2001 From: Fred Isaman Date: Wed, 19 Mar 2008 11:54:04 -0400 Subject: nfs: remove duplicate initializations of nfs_read_data field Signed-off-by: Fred Isaman Signed-off-by: Trond Myklebust --- fs/nfs/read.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'fs') diff --git a/fs/nfs/read.c b/fs/nfs/read.c index ab2f7d233e0..d333f5fedca 100644 --- a/fs/nfs/read.c +++ b/fs/nfs/read.c @@ -251,7 +251,6 @@ static int nfs_pagein_multi(struct inode *inode, struct list_head *head, unsigne data = nfs_readdata_alloc(1); if (!data) goto out_bad; - INIT_LIST_HEAD(&data->pages); list_add(&data->pages, &list); requests++; nbytes -= len; @@ -298,7 +297,6 @@ static int nfs_pagein_one(struct inode *inode, struct list_head *head, unsigned if (!data) goto out_bad; - INIT_LIST_HEAD(&data->pages); pages = data->pagevec; while (!list_empty(head)) { req = nfs_list_entry(head->next); -- cgit v1.2.3 From caa02bd540618e4b447a1f776363ba27c4c79090 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Fri, 22 Feb 2008 14:49:59 -0500 Subject: NFS: clean up short packet handling for NFSv2 readdir Currently, the NFS readdir decoders have a workaround for buggy servers that send an empty readdir response with the EOF bit unset. If the server sends a malformed response in some cases, this workaround kicks in and just returns an empty response rather than returning a proper error to the caller. This patch does 3 things: 1) have malformed responses with no entries return error (-EIO) 2) preserve existing workaround for servers that send empty responses with the EOF marker unset. 3) Add some comments to clarify the logic in nfs_xdr_readdirres(). Signed-off-by: Jeff Layton Signed-off-by: Trond Myklebust --- fs/nfs/nfs2xdr.c | 37 ++++++++++++++++++++++++++++--------- 1 file changed, 28 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs2xdr.c b/fs/nfs/nfs2xdr.c index 1f7ea675e0c..86a80b33ec8 100644 --- a/fs/nfs/nfs2xdr.c +++ b/fs/nfs/nfs2xdr.c @@ -428,7 +428,7 @@ nfs_xdr_readdirres(struct rpc_rqst *req, __be32 *p, void *dummy) size_t hdrlen; unsigned int pglen, recvd; u32 len; - int status, nr; + int status, nr = 0; __be32 *end, *entry, *kaddr; if ((status = ntohl(*p++))) @@ -452,7 +452,12 @@ nfs_xdr_readdirres(struct rpc_rqst *req, __be32 *p, void *dummy) kaddr = p = kmap_atomic(*page, KM_USER0); end = (__be32 *)((char *)p + pglen); entry = p; - for (nr = 0; *p++; nr++) { + + /* Make sure the packet actually has a value_follows and EOF entry */ + if ((entry + 1) > end) + goto short_pkt; + + for (; *p++; nr++) { if (p + 2 > end) goto short_pkt; p++; /* fileid */ @@ -467,18 +472,32 @@ nfs_xdr_readdirres(struct rpc_rqst *req, __be32 *p, void *dummy) goto short_pkt; entry = p; } - if (!nr && (entry[0] != 0 || entry[1] == 0)) - goto short_pkt; + + /* + * Apparently some server sends responses that are a valid size, but + * contain no entries, and have value_follows==0 and EOF==0. For + * those, just set the EOF marker. + */ + if (!nr && entry[1] == 0) { + dprintk("NFS: readdir reply truncated!\n"); + entry[1] = 1; + } out: kunmap_atomic(kaddr, KM_USER0); return nr; short_pkt: + /* + * When we get a short packet there are 2 possibilities. We can + * return an error, or fix up the response to look like a valid + * response and return what we have so far. If there are no + * entries and the packet was short, then return -EIO. If there + * are valid entries in the response, return them and pretend that + * the call was successful, but incomplete. The caller can retry the + * readdir starting at the last cookie. + */ entry[0] = entry[1] = 0; - /* truncate listing ? */ - if (!nr) { - dprintk("NFS: readdir reply truncated!\n"); - entry[1] = 1; - } + if (!nr) + nr = -errno_NFSERR_IO; goto out; err_unmap: nr = -errno_NFSERR_IO; -- cgit v1.2.3 From 643f81115baca3630e544f6874567648b605efae Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Fri, 22 Feb 2008 14:50:00 -0500 Subject: NFS: clean up short packet handling for NFSv3 readdir Currently, the NFS readdir decoders have a workaround for buggy servers that send an empty readdir response with the EOF bit unset. If the server sends a malformed response in some cases, this workaround kicks in and just returns an empty response rather than returning a proper error to the caller. This patch does 3 things: 1) have malformed responses with no entries return error (-EIO) 2) preserve existing workaround for servers that send empty responses with the EOF marker unset. 3) Add some comments to clarify the logic in nfs3_xdr_readdirres(). Signed-off-by: Jeff Layton Signed-off-by: Trond Myklebust --- fs/nfs/nfs3xdr.c | 37 ++++++++++++++++++++++++++++--------- 1 file changed, 28 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs3xdr.c b/fs/nfs/nfs3xdr.c index 3917e2fa4e4..fb03048ac65 100644 --- a/fs/nfs/nfs3xdr.c +++ b/fs/nfs/nfs3xdr.c @@ -508,7 +508,7 @@ nfs3_xdr_readdirres(struct rpc_rqst *req, __be32 *p, struct nfs3_readdirres *res struct page **page; size_t hdrlen; u32 len, recvd, pglen; - int status, nr; + int status, nr = 0; __be32 *entry, *end, *kaddr; status = ntohl(*p++); @@ -542,7 +542,12 @@ nfs3_xdr_readdirres(struct rpc_rqst *req, __be32 *p, struct nfs3_readdirres *res kaddr = p = kmap_atomic(*page, KM_USER0); end = (__be32 *)((char *)p + pglen); entry = p; - for (nr = 0; *p++; nr++) { + + /* Make sure the packet actually has a value_follows and EOF entry */ + if ((entry + 1) > end) + goto short_pkt; + + for (; *p++; nr++) { if (p + 3 > end) goto short_pkt; p += 2; /* inode # */ @@ -581,18 +586,32 @@ nfs3_xdr_readdirres(struct rpc_rqst *req, __be32 *p, struct nfs3_readdirres *res goto short_pkt; entry = p; } - if (!nr && (entry[0] != 0 || entry[1] == 0)) - goto short_pkt; + + /* + * Apparently some server sends responses that are a valid size, but + * contain no entries, and have value_follows==0 and EOF==0. For + * those, just set the EOF marker. + */ + if (!nr && entry[1] == 0) { + dprintk("NFS: readdir reply truncated!\n"); + entry[1] = 1; + } out: kunmap_atomic(kaddr, KM_USER0); return nr; short_pkt: + /* + * When we get a short packet there are 2 possibilities. We can + * return an error, or fix up the response to look like a valid + * response and return what we have so far. If there are no + * entries and the packet was short, then return -EIO. If there + * are valid entries in the response, return them and pretend that + * the call was successful, but incomplete. The caller can retry the + * readdir starting at the last cookie. + */ entry[0] = entry[1] = 0; - /* truncate listing ? */ - if (!nr) { - dprintk("NFS: readdir reply truncated!\n"); - entry[1] = 1; - } + if (!nr) + nr = -errno_NFSERR_IO; goto out; err_unmap: nr = -errno_NFSERR_IO; -- cgit v1.2.3 From 7bda2cdf484a00e52b0ed925e99d4bf4696b2c7a Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Fri, 22 Feb 2008 14:50:01 -0500 Subject: NFS: clean up short packet handling for NFSv4 readdir Currently, the NFS readdir decoders have a workaround for buggy servers that send an empty readdir response with the EOF bit unset. If the server sends a malformed response in some cases, this workaround kicks in and just returns an empty response rather than returning a proper error to the caller. This patch does 3 things: 1) have malformed responses with no entries return error (-EIO) 2) preserve existing workaround for servers that send empty responses with the EOF marker unset. 3) Add some comments to clarify the logic in decode_readdir(). Signed-off-by: Jeff Layton Signed-off-by: Trond Myklebust --- fs/nfs/nfs4xdr.c | 37 +++++++++++++++++++++++++++---------- 1 file changed, 27 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index 37421dd4805..2b519f6325f 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -3481,7 +3481,7 @@ static int decode_readdir(struct xdr_stream *xdr, struct rpc_rqst *req, struct n size_t hdrlen; u32 recvd, pglen = rcvbuf->page_len; __be32 *end, *entry, *p, *kaddr; - unsigned int nr; + unsigned int nr = 0; int status; status = decode_op_hdr(xdr, OP_READDIR); @@ -3505,7 +3505,12 @@ static int decode_readdir(struct xdr_stream *xdr, struct rpc_rqst *req, struct n kaddr = p = kmap_atomic(page, KM_USER0); end = p + ((pglen + readdir->pgbase) >> 2); entry = p; - for (nr = 0; *p++; nr++) { + + /* Make sure the packet actually has a value_follows and EOF entry */ + if ((entry + 1) > end) + goto short_pkt; + + for (; *p++; nr++) { u32 len, attrlen, xlen; if (end - p < 3) goto short_pkt; @@ -3532,20 +3537,32 @@ static int decode_readdir(struct xdr_stream *xdr, struct rpc_rqst *req, struct n p += attrlen; /* attributes */ entry = p; } - if (!nr && (entry[0] != 0 || entry[1] == 0)) - goto short_pkt; + /* + * Apparently some server sends responses that are a valid size, but + * contain no entries, and have value_follows==0 and EOF==0. For + * those, just set the EOF marker. + */ + if (!nr && entry[1] == 0) { + dprintk("NFS: readdir reply truncated!\n"); + entry[1] = 1; + } out: kunmap_atomic(kaddr, KM_USER0); return 0; short_pkt: + /* + * When we get a short packet there are 2 possibilities. We can + * return an error, or fix up the response to look like a valid + * response and return what we have so far. If there are no + * entries and the packet was short, then return -EIO. If there + * are valid entries in the response, return them and pretend that + * the call was successful, but incomplete. The caller can retry the + * readdir starting at the last cookie. + */ dprintk("%s: short packet at entry %d\n", __FUNCTION__, nr); entry[0] = entry[1] = 0; - /* truncate listing ? */ - if (!nr) { - dprintk("NFS: readdir reply truncated!\n"); - entry[1] = 1; - } - goto out; + if (nr) + goto out; err_unmap: kunmap_atomic(kaddr, KM_USER0); return -errno_NFSERR_IO; -- cgit v1.2.3 From 2d76743227028a26f884284aade03d1e43f4f249 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 14 Mar 2008 14:10:08 -0400 Subject: NFS: numeric mount parameters are unsigned Clean up: use %u instead of %d when displaying NFS mount options. Nit: Fix reporting of "namlen=" option in nfs_show_mount_stats. The mount option is called "namlen" without the "e". Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfs/super.c | 26 +++++++++++++------------- 1 file changed, 13 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/nfs/super.c b/fs/nfs/super.c index fcf4b982c88..53a67c6d4d2 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -463,17 +463,17 @@ static void nfs_show_mount_options(struct seq_file *m, struct nfs_server *nfss, const struct proc_nfs_info *nfs_infop; struct nfs_client *clp = nfss->nfs_client; - seq_printf(m, ",vers=%d", clp->rpc_ops->version); - seq_printf(m, ",rsize=%d", nfss->rsize); - seq_printf(m, ",wsize=%d", nfss->wsize); + seq_printf(m, ",vers=%u", clp->rpc_ops->version); + seq_printf(m, ",rsize=%u", nfss->rsize); + seq_printf(m, ",wsize=%u", nfss->wsize); if (nfss->acregmin != 3*HZ || showdefaults) - seq_printf(m, ",acregmin=%d", nfss->acregmin/HZ); + seq_printf(m, ",acregmin=%u", nfss->acregmin/HZ); if (nfss->acregmax != 60*HZ || showdefaults) - seq_printf(m, ",acregmax=%d", nfss->acregmax/HZ); + seq_printf(m, ",acregmax=%u", nfss->acregmax/HZ); if (nfss->acdirmin != 30*HZ || showdefaults) - seq_printf(m, ",acdirmin=%d", nfss->acdirmin/HZ); + seq_printf(m, ",acdirmin=%u", nfss->acdirmin/HZ); if (nfss->acdirmax != 60*HZ || showdefaults) - seq_printf(m, ",acdirmax=%d", nfss->acdirmax/HZ); + seq_printf(m, ",acdirmax=%u", nfss->acdirmax/HZ); for (nfs_infop = nfs_info; nfs_infop->flag; nfs_infop++) { if (nfss->flags & nfs_infop->flag) seq_puts(m, nfs_infop->str); @@ -529,10 +529,10 @@ static int nfs_show_stats(struct seq_file *m, struct vfsmount *mnt) seq_printf(m, "\n\tcaps:\t"); seq_printf(m, "caps=0x%x", nfss->caps); - seq_printf(m, ",wtmult=%d", nfss->wtmult); - seq_printf(m, ",dtsize=%d", nfss->dtsize); - seq_printf(m, ",bsize=%d", nfss->bsize); - seq_printf(m, ",namelen=%d", nfss->namelen); + seq_printf(m, ",wtmult=%u", nfss->wtmult); + seq_printf(m, ",dtsize=%u", nfss->dtsize); + seq_printf(m, ",bsize=%u", nfss->bsize); + seq_printf(m, ",namlen=%u", nfss->namelen); #ifdef CONFIG_NFS_V4 if (nfss->nfs_client->rpc_ops->version == 4) { @@ -546,9 +546,9 @@ static int nfs_show_stats(struct seq_file *m, struct vfsmount *mnt) /* * Display security flavor in effect for this mount */ - seq_printf(m, "\n\tsec:\tflavor=%d", auth->au_ops->au_flavor); + seq_printf(m, "\n\tsec:\tflavor=%u", auth->au_ops->au_flavor); if (auth->au_flavor) - seq_printf(m, ",pseudoflavor=%d", auth->au_flavor); + seq_printf(m, ",pseudoflavor=%u", auth->au_flavor); /* * Display superblock I/O counters -- cgit v1.2.3 From 78fa701f341564e60461de91cd08ff5f7fb09b31 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 14 Mar 2008 14:10:15 -0400 Subject: NFS: Fix up data types of fields in nfs_parsed_mount_options Clean up: make data types of fields in nfs_parsed_mount_options more consistent with other uses. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/nfs/internal.h | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index 4c1122a1384..e89688a955b 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -46,9 +46,9 @@ struct nfs_parsed_mount_data { struct sockaddr_storage address; size_t addrlen; char *hostname; - unsigned int version; + u32 version; unsigned short port; - int protocol; + unsigned short protocol; } mount_server; struct { @@ -56,7 +56,7 @@ struct nfs_parsed_mount_data { size_t addrlen; char *hostname; char *export_path; - int protocol; + unsigned short protocol; } nfs_server; struct security_mnt_opts lsm_opts; -- cgit v1.2.3 From f22d6d79fe227245363a8849ea8c85fe6c6598c3 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 14 Mar 2008 14:10:22 -0400 Subject: NFS: Save the value of the "port=" mount option During a remount based on the mount options displayed in /proc/mounts, we want to preserve the original behavior of the mount request. Let's save the original setting of the "port=" mount option in the mount's nfs_server structure. This allows us to simplify the default behavior of port setting for NFSv4 mounts: by default, NFSv2/3 mounts first try an RPC bind to determine the NFS server's port, unless the user specified the "port=" mount option; Users can force the client to skip the RPC bind by explicitly specifying "port=". NFSv4, by contrast, assumes the NFS server port is 2049 and skips the RPC bind, unless the user specifies "port=". Users can force an RPC bind for NFSv4 by explicitly specifying "port=0". I added a couple of extra comments to clarify this behavior. Signed-off-by: Chuck Lever Cc: Miklos Szeredi Signed-off-by: Trond Myklebust --- fs/nfs/client.c | 4 ++++ fs/nfs/internal.h | 1 + fs/nfs/super.c | 37 ++++++------------------------------- 3 files changed, 11 insertions(+), 31 deletions(-) (limited to 'fs') diff --git a/fs/nfs/client.c b/fs/nfs/client.c index 06f064d8fbb..874018113d0 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -682,6 +682,8 @@ static int nfs_init_server(struct nfs_server *server, if (error < 0) goto error; + server->port = data->nfs_server.port; + error = nfs_init_server_rpcclient(server, &timeparms, data->auth_flavors[0]); if (error < 0) goto error; @@ -1064,6 +1066,8 @@ static int nfs4_init_server(struct nfs_server *server, server->acdirmin = data->acdirmin * HZ; server->acdirmax = data->acdirmax * HZ; + server->port = data->nfs_server.port; + error = nfs_init_server_rpcclient(server, &timeparms, data->auth_flavors[0]); error: diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index e89688a955b..999ad8ee064 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -56,6 +56,7 @@ struct nfs_parsed_mount_data { size_t addrlen; char *hostname; char *export_path; + unsigned short port; unsigned short protocol; } nfs_server; diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 53a67c6d4d2..3c6f53aa731 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -685,7 +685,6 @@ static int nfs_parse_mount_options(char *raw, struct nfs_parsed_mount_data *mnt) { char *p, *string, *secdata; - unsigned short port = 0; int rc; if (!raw) { @@ -800,7 +799,7 @@ static int nfs_parse_mount_options(char *raw, return 0; if (option < 0 || option > 65535) return 0; - port = option; + mnt->nfs_server.port = option; break; case Opt_rsize: if (match_int(args, &mnt->rsize)) @@ -1050,7 +1049,8 @@ static int nfs_parse_mount_options(char *raw, } } - nfs_set_port((struct sockaddr *)&mnt->nfs_server.address, port); + nfs_set_port((struct sockaddr *)&mnt->nfs_server.address, + mnt->nfs_server.port); return 1; @@ -1171,7 +1171,9 @@ static int nfs_validate_mount_data(void *options, args->acregmax = 60; args->acdirmin = 30; args->acdirmax = 60; + args->mount_server.port = 0; /* autobind unless user sets port */ args->mount_server.protocol = XPRT_TRANSPORT_UDP; + args->nfs_server.port = 0; /* autobind unless user sets port */ args->nfs_server.protocol = XPRT_TRANSPORT_TCP; switch (data->version) { @@ -1707,28 +1709,6 @@ static void nfs4_fill_super(struct super_block *sb) nfs_initialise_sb(sb); } -/* - * If the user didn't specify a port, set the port number to - * the NFS version 4 default port. - */ -static void nfs4_default_port(struct sockaddr *sap) -{ - switch (sap->sa_family) { - case AF_INET: { - struct sockaddr_in *ap = (struct sockaddr_in *)sap; - if (ap->sin_port == 0) - ap->sin_port = htons(NFS_PORT); - break; - } - case AF_INET6: { - struct sockaddr_in6 *ap = (struct sockaddr_in6 *)sap; - if (ap->sin6_port == 0) - ap->sin6_port = htons(NFS_PORT); - break; - } - } -} - /* * Validate NFSv4 mount options */ @@ -1753,6 +1733,7 @@ static int nfs4_validate_mount_data(void *options, args->acregmax = 60; args->acdirmin = 30; args->acdirmax = 60; + args->nfs_server.port = NFS_PORT; /* 2049 unless user set port= */ args->nfs_server.protocol = XPRT_TRANSPORT_TCP; switch (data->version) { @@ -1769,9 +1750,6 @@ static int nfs4_validate_mount_data(void *options, &args->nfs_server.address)) goto out_no_address; - nfs4_default_port((struct sockaddr *) - &args->nfs_server.address); - switch (data->auth_flavourlen) { case 0: args->auth_flavors[0] = RPC_AUTH_UNIX; @@ -1829,9 +1807,6 @@ static int nfs4_validate_mount_data(void *options, &args->nfs_server.address)) return -EINVAL; - nfs4_default_port((struct sockaddr *) - &args->nfs_server.address); - switch (args->auth_flavor_len) { case 0: args->auth_flavors[0] = RPC_AUTH_UNIX; -- cgit v1.2.3 From 3f8400d1f1f9d5fb175bdbf6236e564dde454f28 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 14 Mar 2008 14:10:30 -0400 Subject: NFS: Save the values of the "mount*=" mount options Save the value of the mountproto= mountport= mountvers= and mountaddr= options so that these values can be displayed later via nfs_show_options(). This preserves the intent of the original mount options, should the file system need to be remounted based on what's displayed in /proc/mounts. Signed-off-by: Chuck Lever Cc: Miklos Szeredi Signed-off-by: Trond Myklebust --- fs/nfs/client.c | 10 ++++++++++ 1 file changed, 10 insertions(+) (limited to 'fs') diff --git a/fs/nfs/client.c b/fs/nfs/client.c index 874018113d0..93dfd75aba7 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -688,6 +688,16 @@ static int nfs_init_server(struct nfs_server *server, if (error < 0) goto error; + /* Preserve the values of mount_server-related mount options */ + if (data->mount_server.addrlen) { + memcpy(&server->mountd_address, &data->mount_server.address, + data->mount_server.addrlen); + server->mountd_addrlen = data->mount_server.addrlen; + } + server->mountd_version = data->mount_server.version; + server->mountd_port = data->mount_server.port; + server->mountd_protocol = data->mount_server.protocol; + server->namelen = data->namlen; /* Create a client RPC handle for the NFSv3 ACL management interface */ nfs_init_server_aclclient(server); -- cgit v1.2.3 From 82d101d58a2312297ee79f96d44c1d8c7fe1032d Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 14 Mar 2008 14:10:37 -0400 Subject: NFS: Show most mount options via nfs_show_options() Display all mount options in /proc/mount which may be needed to reconstruct a previous mount. Signed-off-by: Chuck Lever Cc: Miklos Szeredi Signed-off-by: Trond Myklebust --- fs/nfs/super.c | 67 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 65 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 3c6f53aa731..c99ca1f992c 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -441,10 +441,52 @@ static const char *nfs_pseudoflavour_to_name(rpc_authflavor_t flavour) return sec_flavours[i].str; } +static void nfs_show_mountd_options(struct seq_file *m, struct nfs_server *nfss, + int showdefaults) +{ + struct sockaddr *sap = (struct sockaddr *)&nfss->mountd_address; + + switch (sap->sa_family) { + case AF_INET: { + struct sockaddr_in *sin = (struct sockaddr_in *)sap; + seq_printf(m, ",mountaddr=" NIPQUAD_FMT, + NIPQUAD(sin->sin_addr.s_addr)); + break; + } + case AF_INET6: { + struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)sap; + seq_printf(m, ",mountaddr=" NIP6_FMT, + NIP6(sin6->sin6_addr)); + break; + } + default: + if (showdefaults) + seq_printf(m, ",mountaddr=unspecified"); + } + + if (nfss->mountd_version || showdefaults) + seq_printf(m, ",mountvers=%u", nfss->mountd_version); + if (nfss->mountd_port || showdefaults) + seq_printf(m, ",mountport=%u", nfss->mountd_port); + + switch (nfss->mountd_protocol) { + case IPPROTO_UDP: + seq_printf(m, ",mountproto=udp"); + break; + case IPPROTO_TCP: + seq_printf(m, ",mountproto=tcp"); + break; + default: + if (showdefaults) + seq_printf(m, ",mountproto=auto"); + } +} + /* * Describe the mount options in force on this server representation */ -static void nfs_show_mount_options(struct seq_file *m, struct nfs_server *nfss, int showdefaults) +static void nfs_show_mount_options(struct seq_file *m, struct nfs_server *nfss, + int showdefaults) { static const struct proc_nfs_info { int flag; @@ -452,6 +494,8 @@ static void nfs_show_mount_options(struct seq_file *m, struct nfs_server *nfss, const char *nostr; } nfs_info[] = { { NFS_MOUNT_SOFT, ",soft", ",hard" }, + { NFS_MOUNT_INTR, ",intr", ",nointr" }, + { NFS_MOUNT_POSIX, ",posix", "" }, { NFS_MOUNT_NOCTO, ",nocto", "" }, { NFS_MOUNT_NOAC, ",noac", "" }, { NFS_MOUNT_NONLM, ",nolock", "" }, @@ -462,10 +506,14 @@ static void nfs_show_mount_options(struct seq_file *m, struct nfs_server *nfss, }; const struct proc_nfs_info *nfs_infop; struct nfs_client *clp = nfss->nfs_client; + u32 version = clp->rpc_ops->version; - seq_printf(m, ",vers=%u", clp->rpc_ops->version); + seq_printf(m, ",vers=%u", version); seq_printf(m, ",rsize=%u", nfss->rsize); seq_printf(m, ",wsize=%u", nfss->wsize); + if (nfss->bsize != 0) + seq_printf(m, ",bsize=%u", nfss->bsize); + seq_printf(m, ",namlen=%u", nfss->namelen); if (nfss->acregmin != 3*HZ || showdefaults) seq_printf(m, ",acregmin=%u", nfss->acregmin/HZ); if (nfss->acregmax != 60*HZ || showdefaults) @@ -482,9 +530,24 @@ static void nfs_show_mount_options(struct seq_file *m, struct nfs_server *nfss, } seq_printf(m, ",proto=%s", rpc_peeraddr2str(nfss->client, RPC_DISPLAY_PROTO)); + if (version == 4) { + if (nfss->port != NFS_PORT) + seq_printf(m, ",port=%u", nfss->port); + } else + if (nfss->port) + seq_printf(m, ",port=%u", nfss->port); + seq_printf(m, ",timeo=%lu", 10U * nfss->client->cl_timeout->to_initval / HZ); seq_printf(m, ",retrans=%u", nfss->client->cl_timeout->to_retries); seq_printf(m, ",sec=%s", nfs_pseudoflavour_to_name(nfss->client->cl_auth->au_flavor)); + + if (version != 4) + nfs_show_mountd_options(m, nfss, showdefaults); + +#ifdef CONFIG_NFS_V4 + if (clp->rpc_ops->version == 4) + seq_printf(m, ",clientaddr=%s", clp->cl_ipaddr); +#endif } /* -- cgit v1.2.3 From ecfc555a8327ff09b07066d73a98c04115007eec Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 14 Mar 2008 14:14:56 -0400 Subject: NFS: Always enable NFS direct I/O Since O_DIRECT is a standard feature that is enabled in most distros, eliminate the CONFIG_NFS_DIRECTIO build option, and change the fs/nfs/Makefile to always build in the NFS direct I/O engine. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/Kconfig | 24 ------------------------ fs/nfs/Makefile | 3 +-- fs/nfs/file.c | 6 ------ fs/nfs/internal.h | 5 ----- 4 files changed, 1 insertion(+), 37 deletions(-) (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index d7312825592..55c1a0374e1 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -1637,30 +1637,6 @@ config NFS_V4 If unsure, say N. -config NFS_DIRECTIO - bool "Allow direct I/O on NFS files" - depends on NFS_FS - help - This option enables applications to perform uncached I/O on files - in NFS file systems using the O_DIRECT open() flag. When O_DIRECT - is set for a file, its data is not cached in the system's page - cache. Data is moved to and from user-level application buffers - directly. Unlike local disk-based file systems, NFS O_DIRECT has - no alignment restrictions. - - Unless your program is designed to use O_DIRECT properly, you are - much better off allowing the NFS client to manage data caching for - you. Misusing O_DIRECT can cause poor server performance or network - storms. This kernel build option defaults OFF to avoid exposing - system administrators unwittingly to a potentially hazardous - feature. - - For more details on NFS O_DIRECT, see fs/nfs/direct.c. - - If unsure, say N. This reduces the size of the NFS client, and - causes open() to return EINVAL if a file residing in NFS is - opened with the O_DIRECT flag. - config NFSD tristate "NFS server support" depends on INET diff --git a/fs/nfs/Makefile b/fs/nfs/Makefile index df0f41e0988..ac6170c594a 100644 --- a/fs/nfs/Makefile +++ b/fs/nfs/Makefile @@ -5,7 +5,7 @@ obj-$(CONFIG_NFS_FS) += nfs.o nfs-y := client.o dir.o file.o getroot.o inode.o super.o nfs2xdr.o \ - pagelist.o proc.o read.o symlink.o unlink.o \ + direct.o pagelist.o proc.o read.o symlink.o unlink.o \ write.o namespace.o mount_clnt.o nfs-$(CONFIG_ROOT_NFS) += nfsroot.o nfs-$(CONFIG_NFS_V3) += nfs3proc.o nfs3xdr.o @@ -14,5 +14,4 @@ nfs-$(CONFIG_NFS_V4) += nfs4proc.o nfs4xdr.o nfs4state.o nfs4renewd.o \ delegation.o idmap.o \ callback.o callback_xdr.o callback_proc.o \ nfs4namespace.o -nfs-$(CONFIG_NFS_DIRECTIO) += direct.o nfs-$(CONFIG_SYSCTL) += sysctl.o diff --git a/fs/nfs/file.c b/fs/nfs/file.c index ef57a5ae590..10e8b807e7f 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -234,10 +234,8 @@ nfs_file_read(struct kiocb *iocb, const struct iovec *iov, ssize_t result; size_t count = iov_length(iov, nr_segs); -#ifdef CONFIG_NFS_DIRECTIO if (iocb->ki_filp->f_flags & O_DIRECT) return nfs_file_direct_read(iocb, iov, nr_segs, pos); -#endif dfprintk(VFS, "nfs: read(%s/%s, %lu@%lu)\n", dentry->d_parent->d_name.name, dentry->d_name.name, @@ -383,9 +381,7 @@ const struct address_space_operations nfs_file_aops = { .write_end = nfs_write_end, .invalidatepage = nfs_invalidate_page, .releasepage = nfs_release_page, -#ifdef CONFIG_NFS_DIRECTIO .direct_IO = nfs_direct_IO, -#endif .launder_page = nfs_launder_page, }; @@ -443,10 +439,8 @@ static ssize_t nfs_file_write(struct kiocb *iocb, const struct iovec *iov, ssize_t result; size_t count = iov_length(iov, nr_segs); -#ifdef CONFIG_NFS_DIRECTIO if (iocb->ki_filp->f_flags & O_DIRECT) return nfs_file_direct_write(iocb, iov, nr_segs, pos); -#endif dfprintk(VFS, "nfs: write(%s/%s(%ld), %lu@%Ld)\n", dentry->d_parent->d_name.name, dentry->d_name.name, diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index 999ad8ee064..04ae867dddb 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -116,13 +116,8 @@ extern void nfs_destroy_readpagecache(void); extern int __init nfs_init_writepagecache(void); extern void nfs_destroy_writepagecache(void); -#ifdef CONFIG_NFS_DIRECTIO extern int __init nfs_init_directcache(void); extern void nfs_destroy_directcache(void); -#else -#define nfs_init_directcache() (0) -#define nfs_destroy_directcache() do {} while(0) -#endif /* nfs2xdr.c */ extern int nfs_stat_to_errno(int); -- cgit v1.2.3 From 327a299d8abd0f4580098cbd6c58b0f017417005 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 14 Mar 2008 14:15:11 -0400 Subject: SUNRPC: Update help Kconfig text Clean up: refresh the help text for Kconfig items related to the sunrpc module. Remove obsolete URLs, and make the language consistent among the options. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/Kconfig | 45 ++++++++++++++++++++++++++++++++------------- 1 file changed, 32 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index 55c1a0374e1..83201856b36 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -1757,15 +1757,32 @@ config SUNRPC_XPRT_RDMA tristate depends on SUNRPC && INFINIBAND && EXPERIMENTAL default SUNRPC && INFINIBAND + help + This option enables an RPC client transport capability that + allows the NFS client to mount servers via an RDMA-enabled + transport. + + To compile RPC client RDMA transport support as a module, + choose M here: the module will be called xprtrdma. + + If unsure, say N. config SUNRPC_BIND34 bool "Support for rpcbind versions 3 & 4 (EXPERIMENTAL)" depends on SUNRPC && EXPERIMENTAL help - Provides kernel support for querying rpcbind servers via versions 3 - and 4 of the rpcbind protocol. The kernel automatically falls back - to version 2 if a remote rpcbind service does not support versions - 3 or 4. + RPC requests over IPv6 networks require support for larger + addresses when performing an RPC bind. Sun added support for + IPv6 addressing by creating two new versions of the rpcbind + protocol (RFC 1833). + + This option enables support in the kernel RPC client for + querying rpcbind servers via versions 3 and 4 of the rpcbind + protocol. The kernel automatically falls back to version 2 + if a remote rpcbind service does not support versions 3 or 4. + By themselves, these new versions do not provide support for + RPC over IPv6, but the new protocol versions are necessary to + support it. If unsure, say N to get traditional behavior (version 2 rpcbind requests only). @@ -1779,12 +1796,13 @@ config RPCSEC_GSS_KRB5 select CRYPTO_DES select CRYPTO_CBC help - Provides for secure RPC calls by means of a gss-api - mechanism based on Kerberos V5. This is required for - NFSv4. + Choose Y here to enable Secure RPC using the Kerberos version 5 + GSS-API mechanism (RFC 1964). - Note: Requires an auxiliary userspace daemon which may be found on - http://www.citi.umich.edu/projects/nfsv4/ + Secure RPC calls with Kerberos require an auxiliary user-space + daemon which may be found in the Linux nfs-utils package + available from http://linux-nfs.org/. In addition, user-space + Kerberos support should be installed. If unsure, say N. @@ -1798,11 +1816,12 @@ config RPCSEC_GSS_SPKM3 select CRYPTO_CAST5 select CRYPTO_CBC help - Provides for secure RPC calls by means of a gss-api - mechanism based on the SPKM3 public-key mechanism. + Choose Y here to enable Secure RPC using the SPKM3 public key + GSS-API mechansim (RFC 2025). - Note: Requires an auxiliary userspace daemon which may be found on - http://www.citi.umich.edu/projects/nfsv4/ + Secure RPC calls with SPKM3 require an auxiliary userspace + daemon which may be found in the Linux nfs-utils package + available from http://linux-nfs.org/. If unsure, say N. -- cgit v1.2.3 From 1e40316bc44851c5920490df01482ed862e4dbc4 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 14 Mar 2008 14:15:18 -0400 Subject: SUNRPC: Add a default setting for CONFIG_SUNRPC_BIND34 Most distros will want support for rpcbind protocols 3 and 4 to default off until they have integrated user-space support for the new rpcbind daemon which supports IPv6 RPC services. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/Kconfig | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index 83201856b36..4115deb567e 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -1770,6 +1770,7 @@ config SUNRPC_XPRT_RDMA config SUNRPC_BIND34 bool "Support for rpcbind versions 3 & 4 (EXPERIMENTAL)" depends on SUNRPC && EXPERIMENTAL + default n help RPC requests over IPv6 networks require support for larger addresses when performing an RPC bind. Sun added support for -- cgit v1.2.3 From 90d5b18061656993410dfd57ddb88aa5a3f34563 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 14 Mar 2008 14:18:30 -0400 Subject: NLM: LOCKD fails to load if CONFIG_SYSCTL is not set Bruce Fields says: "By the way, we've got another config-related nit here: http://bugzilla.linux-nfs.org/show_bug.cgi?id=156 You can build lockd without CONFIG_SYSCTL set, but then the module will fail to load." For now, disable the sysctl registration calls in lockd if CONFIG_SYSCTL is not enabled. This allows the kernel to build properly if PROC_FS or SYSCTL is not enabled, but an NFS client is desired. In the long run, we would like to be able to build the kernel with an NFS client but without lockd. This makes sense, for example, if you want an NFSv4-only NFS client, as NFSv4 doesn't use NLM at all. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/lockd/svc.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'fs') diff --git a/fs/lockd/svc.c b/fs/lockd/svc.c index 1ed8bd4de94..38c2f0b1dd7 100644 --- a/fs/lockd/svc.c +++ b/fs/lockd/svc.c @@ -74,7 +74,9 @@ static const unsigned long nlm_timeout_min = 3; static const unsigned long nlm_timeout_max = 20; static const int nlm_port_min = 0, nlm_port_max = 65535; +#ifdef CONFIG_SYSCTL static struct ctl_table_header * nlm_sysctl_table; +#endif static unsigned long get_lockd_grace_period(void) { @@ -359,6 +361,8 @@ out: } EXPORT_SYMBOL(lockd_down); +#ifdef CONFIG_SYSCTL + /* * Sysctl parameters (same as module parameters, different interface). */ @@ -443,6 +447,8 @@ static ctl_table nlm_sysctl_root[] = { { .ctl_name = 0 } }; +#endif /* CONFIG_SYSCTL */ + /* * Module (and sysfs) parameters. */ @@ -516,15 +522,21 @@ module_param(nsm_use_hostnames, bool, 0644); static int __init init_nlm(void) { +#ifdef CONFIG_SYSCTL nlm_sysctl_table = register_sysctl_table(nlm_sysctl_root); return nlm_sysctl_table ? 0 : -ENOMEM; +#else + return 0; +#endif } static void __exit exit_nlm(void) { /* FIXME: delete all NLM clients */ nlm_shutdown_hosts(); +#ifdef CONFIG_SYSCTL unregister_sysctl_table(nlm_sysctl_table); +#endif } module_init(init_nlm); -- cgit v1.2.3 From eb18860e1385bfc7f08fcb7ba362e4a5156c8324 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 14 Mar 2008 14:18:37 -0400 Subject: NLM: NLM protocol version numbers are u32 Clean up: RPC protocol version numbers are u32. Make sure we use an appropriate type for NLM version numbers when calling nlm_lookup_host(). Eliminates a harmless mixed sign comparison in nlm_host_lookup(). Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/lockd/host.c | 20 +++++++++++--------- 1 file changed, 11 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/lockd/host.c b/fs/lockd/host.c index f1ef49fff11..f23750db165 100644 --- a/fs/lockd/host.c +++ b/fs/lockd/host.c @@ -42,11 +42,12 @@ static struct nsm_handle * nsm_find(const struct sockaddr_in *sin, /* * Common host lookup routine for server & client */ -static struct nlm_host * -nlm_lookup_host(int server, const struct sockaddr_in *sin, - int proto, int version, const char *hostname, - unsigned int hostname_len, - const struct sockaddr_in *ssin) +static struct nlm_host *nlm_lookup_host(int server, + const struct sockaddr_in *sin, + int proto, u32 version, + const char *hostname, + unsigned int hostname_len, + const struct sockaddr_in *ssin) { struct hlist_head *chain; struct hlist_node *pos; @@ -55,7 +56,7 @@ nlm_lookup_host(int server, const struct sockaddr_in *sin, int hash; dprintk("lockd: nlm_lookup_host("NIPQUAD_FMT"->"NIPQUAD_FMT - ", p=%d, v=%d, my role=%s, name=%.*s)\n", + ", p=%d, v=%u, my role=%s, name=%.*s)\n", NIPQUAD(ssin->sin_addr.s_addr), NIPQUAD(sin->sin_addr.s_addr), proto, version, server? "server" : "client", @@ -175,9 +176,10 @@ nlm_destroy_host(struct nlm_host *host) /* * Find an NLM server handle in the cache. If there is none, create it. */ -struct nlm_host * -nlmclnt_lookup_host(const struct sockaddr_in *sin, int proto, int version, - const char *hostname, unsigned int hostname_len) +struct nlm_host *nlmclnt_lookup_host(const struct sockaddr_in *sin, + int proto, u32 version, + const char *hostname, + unsigned int hostname_len) { struct sockaddr_in ssin = {0}; -- cgit v1.2.3 From 099bd05f27ff24a3041d54e7ed42d2eb681484b9 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 14 Mar 2008 14:25:32 -0400 Subject: lockd: Ensure NSM strings aren't longer than protocol allows Introduce a special helper function to check the length of NSM strings before they are placed on the wire. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/lockd/mon.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'fs') diff --git a/fs/lockd/mon.c b/fs/lockd/mon.c index 908b23fadd0..84fd84cb67b 100644 --- a/fs/lockd/mon.c +++ b/fs/lockd/mon.c @@ -149,6 +149,15 @@ nsm_create(void) * XDR functions for NSM. */ +static __be32 *xdr_encode_nsm_string(__be32 *p, char *string) +{ + size_t len = strlen(string); + + if (len > SM_MAXSTRLEN) + len = SM_MAXSTRLEN; + return xdr_encode_opaque(p, string, len); +} + static __be32 * xdr_encode_common(struct rpc_rqst *rqstp, __be32 *p, struct nsm_args *argp) { -- cgit v1.2.3 From 496951743100b2d2ccd8b268257e79425e489851 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 14 Mar 2008 14:25:39 -0400 Subject: lockd: refactor SM_MON mon_name argument encoder Clean up: introduce a new XDR encoder specifically for the mon_name argument of SM_MON requests. This will be updated later to support IPv6 addresses in addition to IPv4 addresses. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/lockd/mon.c | 25 +++++++++++++++++++++++++ 1 file changed, 25 insertions(+) (limited to 'fs') diff --git a/fs/lockd/mon.c b/fs/lockd/mon.c index 84fd84cb67b..87bd4e4d65d 100644 --- a/fs/lockd/mon.c +++ b/fs/lockd/mon.c @@ -18,6 +18,8 @@ #define NLMDBG_FACILITY NLMDBG_MONITOR +#define XDR_ADDRBUF_LEN (20) + static struct rpc_clnt * nsm_create(void); static struct rpc_program nsm_program; @@ -158,6 +160,29 @@ static __be32 *xdr_encode_nsm_string(__be32 *p, char *string) return xdr_encode_opaque(p, string, len); } +/* + * "mon_name" specifies the host to be monitored. + * + * Linux uses a text version of the IP address of the remote + * host as the host identifier (the "mon_name" argument). + * + * Linux statd always looks up the canonical hostname first for + * whatever remote hostname it receives, so this works alright. + */ +static __be32 *xdr_encode_mon_name(__be32 *p, struct nsm_args *argp) +{ + char buffer[XDR_ADDRBUF_LEN + 1]; + char *name = argp->mon_name; + + if (!nsm_use_hostnames) { + snprintf(buffer, XDR_ADDRBUF_LEN, + NIPQUAD_FMT, NIPQUAD(argp->addr)); + name = buffer; + } + + return xdr_encode_nsm_string(p, name); +} + static __be32 * xdr_encode_common(struct rpc_rqst *rqstp, __be32 *p, struct nsm_args *argp) { -- cgit v1.2.3 From 850c95fd07b63704d06909c9a081c51272d32db1 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 14 Mar 2008 14:25:46 -0400 Subject: lockd: refactor SM_MON my_id argument encoder Clean up: introduce a new XDR encoder specifically for the my_id argument of SM_MON requests. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/lockd/mon.c | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) (limited to 'fs') diff --git a/fs/lockd/mon.c b/fs/lockd/mon.c index 87bd4e4d65d..06216d6715f 100644 --- a/fs/lockd/mon.c +++ b/fs/lockd/mon.c @@ -183,6 +183,25 @@ static __be32 *xdr_encode_mon_name(__be32 *p, struct nsm_args *argp) return xdr_encode_nsm_string(p, name); } +/* + * The "my_id" argument specifies the hostname and RPC procedure + * to be called when the status manager receives notification + * (via the SM_NOTIFY call) that the state of host "mon_name" + * has changed. + */ +static __be32 *xdr_encode_my_id(__be32 *p, struct nsm_args *argp) +{ + p = xdr_encode_nsm_string(p, utsname()->nodename); + if (!p) + return ERR_PTR(-EIO); + + *p++ = htonl(argp->prog); + *p++ = htonl(argp->vers); + *p++ = htonl(argp->proc); + + return p; +} + static __be32 * xdr_encode_common(struct rpc_rqst *rqstp, __be32 *p, struct nsm_args *argp) { -- cgit v1.2.3 From ea72a7f170e686baf00ceee57b6197bef686889c Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 14 Mar 2008 14:25:53 -0400 Subject: lockd: document use of mon_id argument in SM_MON requests Clean up: document the argument type that xdr_encode_common() is marshalling by introducing a new function. The new function will replace xdr_encode_common() in just a sec. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/lockd/mon.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'fs') diff --git a/fs/lockd/mon.c b/fs/lockd/mon.c index 06216d6715f..3935d7b604f 100644 --- a/fs/lockd/mon.c +++ b/fs/lockd/mon.c @@ -202,6 +202,19 @@ static __be32 *xdr_encode_my_id(__be32 *p, struct nsm_args *argp) return p; } +/* + * The "mon_id" argument specifies the non-private arguments + * of an SM_MON or SM_UNMON call. + */ +static __be32 *xdr_encode_mon_id(__be32 *p, struct nsm_args *argp) +{ + p = xdr_encode_mon_name(p, argp); + if (!p) + return ERR_PTR(-EIO); + + return xdr_encode_my_id(p, argp); +} + static __be32 * xdr_encode_common(struct rpc_rqst *rqstp, __be32 *p, struct nsm_args *argp) { -- cgit v1.2.3 From 2ca7754d4c96d68e1475690422a202ba5f0443d8 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 14 Mar 2008 14:26:01 -0400 Subject: lockd: Fix up incorrect RPC buffer size calculations. Switch to using the new mon_id encoder function. Now that we've refactored the encoding of SM_MON requests, we've discovered that the pre-computed buffer length maximums are incorrect! Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/lockd/mon.c | 39 ++++++++------------------------------- 1 file changed, 8 insertions(+), 31 deletions(-) (limited to 'fs') diff --git a/fs/lockd/mon.c b/fs/lockd/mon.c index 3935d7b604f..f2507fec7c6 100644 --- a/fs/lockd/mon.c +++ b/fs/lockd/mon.c @@ -149,6 +149,9 @@ nsm_create(void) /* * XDR functions for NSM. + * + * See http://www.opengroup.org/ for details on the Network + * Status Monitor wire protocol. */ static __be32 *xdr_encode_nsm_string(__be32 *p, char *string) @@ -215,37 +218,10 @@ static __be32 *xdr_encode_mon_id(__be32 *p, struct nsm_args *argp) return xdr_encode_my_id(p, argp); } -static __be32 * -xdr_encode_common(struct rpc_rqst *rqstp, __be32 *p, struct nsm_args *argp) -{ - char buffer[20], *name; - - /* - * Use the dotted-quad IP address of the remote host as - * identifier. Linux statd always looks up the canonical - * hostname first for whatever remote hostname it receives, - * so this works alright. - */ - if (nsm_use_hostnames) { - name = argp->mon_name; - } else { - sprintf(buffer, "%u.%u.%u.%u", NIPQUAD(argp->addr)); - name = buffer; - } - if (!(p = xdr_encode_string(p, name)) - || !(p = xdr_encode_string(p, utsname()->nodename))) - return ERR_PTR(-EIO); - *p++ = htonl(argp->prog); - *p++ = htonl(argp->vers); - *p++ = htonl(argp->proc); - - return p; -} - static int xdr_encode_mon(struct rpc_rqst *rqstp, __be32 *p, struct nsm_args *argp) { - p = xdr_encode_common(rqstp, p, argp); + p = xdr_encode_mon_id(p, argp); if (IS_ERR(p)) return PTR_ERR(p); @@ -261,7 +237,7 @@ xdr_encode_mon(struct rpc_rqst *rqstp, __be32 *p, struct nsm_args *argp) static int xdr_encode_unmon(struct rpc_rqst *rqstp, __be32 *p, struct nsm_args *argp) { - p = xdr_encode_common(rqstp, p, argp); + p = xdr_encode_mon_id(p, argp); if (IS_ERR(p)) return PTR_ERR(p); rqstp->rq_slen = xdr_adjust_iovec(rqstp->rq_svec, p); @@ -286,8 +262,9 @@ xdr_decode_stat(struct rpc_rqst *rqstp, __be32 *p, struct nsm_res *resp) } #define SM_my_name_sz (1+XDR_QUADLEN(SM_MAXSTRLEN)) -#define SM_my_id_sz (3+1+SM_my_name_sz) -#define SM_mon_id_sz (1+XDR_QUADLEN(20)+SM_my_id_sz) +#define SM_my_id_sz (SM_my_name_sz+3) +#define SM_mon_name_sz (1+XDR_QUADLEN(SM_MAXSTRLEN)) +#define SM_mon_id_sz (SM_mon_name_sz+SM_my_id_sz) #define SM_mon_sz (SM_mon_id_sz+4) #define SM_monres_sz 2 #define SM_unmonres_sz 1 -- cgit v1.2.3 From 0490a54a00c14212f22c5948c8c13a4553d745bd Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Fri, 14 Mar 2008 14:26:08 -0400 Subject: lockd: introduce new function to encode private argument in SM_MON requests Clean up: refactor the encoding of the opaque 16-byte private argument in xdr_encode_mon(). This will be updated later to support IPv6 addresses. Signed-off-by: Chuck Lever Signed-off-by: Trond Myklebust --- fs/lockd/mon.c | 30 ++++++++++++++++++++++++------ 1 file changed, 24 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/lockd/mon.c b/fs/lockd/mon.c index f2507fec7c6..e4d563543b1 100644 --- a/fs/lockd/mon.c +++ b/fs/lockd/mon.c @@ -218,6 +218,24 @@ static __be32 *xdr_encode_mon_id(__be32 *p, struct nsm_args *argp) return xdr_encode_my_id(p, argp); } +/* + * The "priv" argument may contain private information required + * by the SM_MON call. This information will be supplied in the + * SM_NOTIFY call. + * + * Linux provides the raw IP address of the monitored host, + * left in network byte order. + */ +static __be32 *xdr_encode_priv(__be32 *p, struct nsm_args *argp) +{ + *p++ = argp->addr; + *p++ = 0; + *p++ = 0; + *p++ = 0; + + return p; +} + static int xdr_encode_mon(struct rpc_rqst *rqstp, __be32 *p, struct nsm_args *argp) { @@ -225,11 +243,10 @@ xdr_encode_mon(struct rpc_rqst *rqstp, __be32 *p, struct nsm_args *argp) if (IS_ERR(p)) return PTR_ERR(p); - /* Surprise - there may even be room for an IPv6 address now */ - *p++ = argp->addr; - *p++ = 0; - *p++ = 0; - *p++ = 0; + p = xdr_encode_priv(p, argp); + if (IS_ERR(p)) + return PTR_ERR(p); + rqstp->rq_slen = xdr_adjust_iovec(rqstp->rq_svec, p); return 0; } @@ -265,7 +282,8 @@ xdr_decode_stat(struct rpc_rqst *rqstp, __be32 *p, struct nsm_res *resp) #define SM_my_id_sz (SM_my_name_sz+3) #define SM_mon_name_sz (1+XDR_QUADLEN(SM_MAXSTRLEN)) #define SM_mon_id_sz (SM_mon_name_sz+SM_my_id_sz) -#define SM_mon_sz (SM_mon_id_sz+4) +#define SM_priv_sz (XDR_QUADLEN(SM_PRIV_SIZE)) +#define SM_mon_sz (SM_mon_id_sz+SM_priv_sz) #define SM_monres_sz 2 #define SM_unmonres_sz 1 -- cgit v1.2.3 From 1218854afa6f659be90b748cf1bc7badee954a35 Mon Sep 17 00:00:00 2001 From: YOSHIFUJI Hideaki Date: Wed, 26 Mar 2008 02:36:06 +0900 Subject: [NET] NETNS: Omit seq_net_private->net without CONFIG_NET_NS. Without CONFIG_NET_NS, no namespace other than &init_net exists, no need to store net in seq_net_private. Signed-off-by: YOSHIFUJI Hideaki --- fs/proc/proc_net.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/proc/proc_net.c b/fs/proc/proc_net.c index 4caa5f774fb..13cd7835d0d 100644 --- a/fs/proc/proc_net.c +++ b/fs/proc/proc_net.c @@ -44,7 +44,9 @@ int seq_open_net(struct inode *ino, struct file *f, put_net(net); return -ENOMEM; } +#ifdef CONFIG_NET_NS p->net = net; +#endif return 0; } EXPORT_SYMBOL_GPL(seq_open_net); @@ -52,12 +54,10 @@ EXPORT_SYMBOL_GPL(seq_open_net); int seq_release_net(struct inode *ino, struct file *f) { struct seq_file *seq; - struct seq_net_private *p; seq = f->private_data; - p = seq->private; - put_net(p->net); + put_net(seq_file_net(seq)); seq_release_private(ino, f); return 0; } -- cgit v1.2.3 From 0e5f8be1388093edc324a78ebf241170b258eba3 Mon Sep 17 00:00:00 2001 From: "Denis V. Lunev" Date: Thu, 27 Mar 2008 14:25:53 -0700 Subject: [NETNS]: Compile NET /proc support only if CONFIG_NET is set. This fix broken compilation for 'allnoconfig'. This was introduced by Introduced by commit 1218854afa6f659be90b748cf1bc7badee954a35 ("[NET] NETNS: Omit seq_net_private->net without CONFIG_NET_NS.") Signed-off-by: Denis V. Lunev Acked-by: YOSHIFUJI Hideaki Signed-off-by: David S. Miller --- fs/proc/proc_net.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/proc/proc_net.c b/fs/proc/proc_net.c index 13cd7835d0d..7034facf8b8 100644 --- a/fs/proc/proc_net.c +++ b/fs/proc/proc_net.c @@ -51,6 +51,7 @@ int seq_open_net(struct inode *ino, struct file *f, } EXPORT_SYMBOL_GPL(seq_open_net); +#ifdef CONFIG_NET int seq_release_net(struct inode *ino, struct file *f) { struct seq_file *seq; @@ -218,3 +219,4 @@ int __init proc_net_init(void) return register_pernet_subsys(&proc_net_ns_ops); } +#endif /* CONFIG_NET */ -- cgit v1.2.3 From 941e6d7d09aaf455c0d7ad383f7f5ae67e4ccf16 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Mon, 28 Jan 2008 08:47:38 +0000 Subject: [GFS2] Speed up gfs2_write_alloc_required, deprecate gfs2_extent_map This patch removes the call to gfs2_extent_map from gfs2_write_alloc_required, instead we call gfs2_block_map directly. This results in fewer overall calls to gfs2_block_map in the multi-block case. Also, gfs2_extent_map is marked as deprecated so that people know that its going away as soon as all the callers have been converted. Signed-off-by: Steven Whitehouse --- fs/gfs2/bmap.c | 36 +++++++++++++++++++----------------- 1 file changed, 19 insertions(+), 17 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index e9456ebd3bb..a25444ac648 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -548,6 +548,9 @@ out_fail: return error; } +/* + * Deprecated: do not use in new code + */ int gfs2_extent_map(struct inode *inode, u64 lblock, int *new, u64 *dblock, unsigned *extlen) { struct buffer_head bh = { .b_state = 0, .b_blocknr = 0 }; @@ -1197,10 +1200,9 @@ int gfs2_write_alloc_required(struct gfs2_inode *ip, u64 offset, unsigned int len, int *alloc_required) { struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); - u64 lblock, lblock_stop, dblock; - u32 extlen; - int new = 0; - int error = 0; + struct buffer_head bh; + unsigned int shift; + u64 lblock, lblock_stop, size; *alloc_required = 0; @@ -1214,6 +1216,8 @@ int gfs2_write_alloc_required(struct gfs2_inode *ip, u64 offset, return 0; } + *alloc_required = 1; + shift = sdp->sd_sb.sb_bsize_shift; if (gfs2_is_dir(ip)) { unsigned int bsize = sdp->sd_jbsize; lblock = offset; @@ -1221,27 +1225,25 @@ int gfs2_write_alloc_required(struct gfs2_inode *ip, u64 offset, lblock_stop = offset + len + bsize - 1; do_div(lblock_stop, bsize); } else { - unsigned int shift = sdp->sd_sb.sb_bsize_shift; u64 end_of_file = (ip->i_di.di_size + sdp->sd_sb.sb_bsize - 1) >> shift; lblock = offset >> shift; lblock_stop = (offset + len + sdp->sd_sb.sb_bsize - 1) >> shift; - if (lblock_stop > end_of_file) { - *alloc_required = 1; + if (lblock_stop > end_of_file) return 0; - } } - for (; lblock < lblock_stop; lblock += extlen) { - error = gfs2_extent_map(&ip->i_inode, lblock, &new, &dblock, &extlen); - if (error) - return error; - - if (!dblock) { - *alloc_required = 1; + size = (lblock_stop - lblock) << shift; + do { + bh.b_state = 0; + bh.b_size = size; + gfs2_block_map(&ip->i_inode, lblock, &bh, 0); + if (!buffer_mapped(&bh)) return 0; - } - } + size -= bh.b_size; + lblock += (bh.b_size >> ip->i_inode.i_blkbits); + } while(size > 0); + *alloc_required = 0; return 0; } -- cgit v1.2.3 From ecc30c79157103e8bd7492043ee992b763443832 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Mon, 28 Jan 2008 10:37:35 +0000 Subject: [GFS2] Streamline indirect pointer tree height calculation This patch improves the calculation of the tree height in order to reduce the number of operations which are carried out on each call to gfs2_block_map. In the common case, we now make a single comparison, rather than calculating the required tree height from scratch each time. Also in the case that the tree does need some extra height, we start from the current height rather from zero when we work out what the new height ought to be. In addition the di_height field is moved into the inode proper and reduced in size to a u8 since the value must be between 0 and GFS2_MAX_META_HEIGHT (10). Signed-off-by: Steven Whitehouse --- fs/gfs2/bmap.c | 108 +++++++++++++++++++------------------------------------ fs/gfs2/incore.h | 6 ++-- fs/gfs2/inode.c | 20 ++++++----- fs/gfs2/inode.h | 2 +- fs/gfs2/super.c | 2 ++ 5 files changed, 55 insertions(+), 83 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index a25444ac648..5a3187049dd 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -166,7 +166,7 @@ int gfs2_unstuff_dinode(struct gfs2_inode *ip, struct page *page) di->di_blocks = cpu_to_be64(ip->i_di.di_blocks); } - ip->i_di.di_height = 1; + ip->i_height = 1; di->di_height = cpu_to_be16(1); out_brelse: @@ -176,43 +176,6 @@ out: return error; } -/** - * calc_tree_height - Calculate the height of a metadata tree - * @ip: The GFS2 inode - * @size: The proposed size of the file - * - * Work out how tall a metadata tree needs to be in order to accommodate a - * file of a particular size. If size is less than the current size of - * the inode, then the current size of the inode is used instead of the - * supplied one. - * - * Returns: the height the tree should be - */ - -static unsigned int calc_tree_height(struct gfs2_inode *ip, u64 size) -{ - struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); - u64 *arr; - unsigned int max, height; - - if (ip->i_di.di_size > size) - size = ip->i_di.di_size; - - if (gfs2_is_dir(ip)) { - arr = sdp->sd_jheightsize; - max = sdp->sd_max_jheight; - } else { - arr = sdp->sd_heightsize; - max = sdp->sd_max_height; - } - - for (height = 0; height < max; height++) - if (arr[height] >= size) - break; - - return height; -} - /** * build_height - Build a metadata tree of the requested height * @ip: The GFS2 inode @@ -225,7 +188,7 @@ static unsigned int calc_tree_height(struct gfs2_inode *ip, u64 size) static int build_height(struct inode *inode, unsigned height) { struct gfs2_inode *ip = GFS2_I(inode); - unsigned new_height = height - ip->i_di.di_height; + unsigned new_height = height - ip->i_height; struct buffer_head *dibh; struct buffer_head *blocks[GFS2_MAX_META_HEIGHT]; struct gfs2_dinode *di; @@ -234,7 +197,7 @@ static int build_height(struct inode *inode, unsigned height) u64 bn; unsigned n; - if (height <= ip->i_di.di_height) + if (height <= ip->i_height) return 0; error = gfs2_meta_inode_buffer(ip, &dibh); @@ -270,10 +233,10 @@ static int build_height(struct inode *inode, unsigned height) di = (struct gfs2_dinode *)dibh->b_data; gfs2_buffer_clear_tail(dibh, sizeof(struct gfs2_dinode)); *(__be64 *)(di + 1) = cpu_to_be64(bn); - ip->i_di.di_height += new_height; + ip->i_height += new_height; ip->i_di.di_blocks += new_height; gfs2_set_inode_blocks(&ip->i_inode); - di->di_height = cpu_to_be16(ip->i_di.di_height); + di->di_height = cpu_to_be16(ip->i_height); di->di_blocks = cpu_to_be64(ip->i_di.di_blocks); brelse(dibh); return error; @@ -345,7 +308,7 @@ static void find_metapath(struct gfs2_inode *ip, u64 block, u64 b = block; unsigned int i; - for (i = ip->i_di.di_height; i--;) + for (i = ip->i_height; i--;) mp->mp_list[i] = do_div(b, sdp->sd_inptrs); } @@ -407,7 +370,7 @@ static int lookup_block(struct gfs2_inode *ip, struct buffer_head *bh, if (!create) return 0; - if (height == ip->i_di.di_height - 1 && !gfs2_is_dir(ip)) + if (height == ip->i_height - 1 && !gfs2_is_dir(ip)) *block = gfs2_alloc_data(ip); else *block = gfs2_alloc_meta(ip); @@ -458,8 +421,7 @@ int gfs2_block_map(struct inode *inode, sector_t lblock, struct gfs2_inode *ip = GFS2_I(inode); struct gfs2_sbd *sdp = GFS2_SB(inode); struct buffer_head *bh; - unsigned int bsize; - unsigned int height; + unsigned int bsize = sdp->sd_sb.sb_bsize; unsigned int end_of_metadata; unsigned int x; int error = 0; @@ -470,7 +432,7 @@ int gfs2_block_map(struct inode *inode, sector_t lblock, struct metapath mp; u64 size; struct buffer_head *dibh = NULL; - + const u64 *arr = sdp->sd_heightsize; BUG_ON(maxlen == 0); if (gfs2_assert_warn(sdp, !gfs2_is_stuffed(ip))) @@ -480,23 +442,25 @@ int gfs2_block_map(struct inode *inode, sector_t lblock, clear_buffer_mapped(bh_map); clear_buffer_new(bh_map); clear_buffer_boundary(bh_map); - bsize = gfs2_is_dir(ip) ? sdp->sd_jbsize : sdp->sd_sb.sb_bsize; + if (gfs2_is_dir(ip)) { + bsize = sdp->sd_jbsize; + arr = sdp->sd_jheightsize; + } size = (lblock + 1) * bsize; - if (size > ip->i_di.di_size) { - height = calc_tree_height(ip, size); - if (ip->i_di.di_height < height) { - if (!create) - goto out_ok; - - error = build_height(inode, height); - if (error) - goto out_fail; - } + if (size > arr[ip->i_height]) { + u8 height = ip->i_height; + if (!create) + goto out_ok; + while (size > arr[height]) + height++; + error = build_height(inode, height); + if (error) + goto out_fail; } find_metapath(ip, lblock, &mp); - end_of_metadata = ip->i_di.di_height - 1; + end_of_metadata = ip->i_height - 1; error = gfs2_meta_inode_buffer(ip, &bh); if (error) goto out_fail; @@ -624,7 +588,7 @@ static int recursive_scan(struct gfs2_inode *ip, struct buffer_head *dibh, if (error) goto out; - if (height < ip->i_di.di_height - 1) + if (height < ip->i_height - 1) for (; top < bottom; top++, first = 0) { if (!*top) continue; @@ -682,7 +646,7 @@ static int do_strip(struct gfs2_inode *ip, struct buffer_head *dibh, sm->sm_first = 0; } - metadata = (height != ip->i_di.di_height - 1); + metadata = (height != ip->i_height - 1); if (metadata) revokes = (height) ? sdp->sd_inptrs : sdp->sd_diptrs; @@ -807,7 +771,6 @@ static int do_grow(struct gfs2_inode *ip, u64 size) struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); struct gfs2_alloc *al; struct buffer_head *dibh; - unsigned int h; int error; al = gfs2_alloc_get(ip); @@ -833,20 +796,23 @@ static int do_grow(struct gfs2_inode *ip, u64 size) goto out_ipres; if (size > sdp->sd_sb.sb_bsize - sizeof(struct gfs2_dinode)) { + const u64 *arr = sdp->sd_heightsize; if (gfs2_is_stuffed(ip)) { error = gfs2_unstuff_dinode(ip, NULL); if (error) goto out_end_trans; } - h = calc_tree_height(ip, size); - if (ip->i_di.di_height < h) { - down_write(&ip->i_rw_mutex); - error = build_height(&ip->i_inode, h); - up_write(&ip->i_rw_mutex); - if (error) - goto out_end_trans; + down_write(&ip->i_rw_mutex); + if (size > arr[ip->i_height]) { + u8 height = ip->i_height; + while(size > arr[height]) + height++; + error = build_height(&ip->i_inode, height); } + up_write(&ip->i_rw_mutex); + if (error) + goto out_end_trans; } ip->i_di.di_size = size; @@ -989,7 +955,7 @@ out: static int trunc_dealloc(struct gfs2_inode *ip, u64 size) { - unsigned int height = ip->i_di.di_height; + unsigned int height = ip->i_height; u64 lblock; struct metapath mp; int error; @@ -1040,7 +1006,7 @@ static int trunc_end(struct gfs2_inode *ip) goto out; if (!ip->i_di.di_size) { - ip->i_di.di_height = 0; + ip->i_height = 0; ip->i_di.di_goal_meta = ip->i_di.di_goal_data = ip->i_no_addr; diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index 525dcae352d..43472baa92e 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -246,7 +246,6 @@ struct gfs2_dinode_host { u64 di_goal_data; /* data block goal */ u64 di_generation; /* generation number for NFS */ u32 di_flags; /* GFS2_DIF_... */ - u16 di_height; /* height of metadata */ /* These only apply to directories */ u16 di_depth; /* Number of bits in the table */ u32 di_entries; /* The number of entries in the directory */ @@ -268,6 +267,7 @@ struct gfs2_inode { u64 i_last_rg_alloc; struct rw_semaphore i_rw_mutex; + u8 i_height; }; /* @@ -490,9 +490,9 @@ struct gfs2_sbd { u32 sd_qc_per_block; u32 sd_max_dirres; /* Max blocks needed to add a directory entry */ u32 sd_max_height; /* Max height of a file's metadata tree */ - u64 sd_heightsize[GFS2_MAX_META_HEIGHT]; + u64 sd_heightsize[GFS2_MAX_META_HEIGHT + 1]; u32 sd_max_jheight; /* Max height of journaled file's meta tree */ - u64 sd_jheightsize[GFS2_MAX_META_HEIGHT]; + u64 sd_jheightsize[GFS2_MAX_META_HEIGHT + 1]; struct gfs2_args sd_args; /* Mount arguments */ struct gfs2_tune sd_tune; /* Filesystem tuning structure */ diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c index 37725ade3c5..ff66ab7a17c 100644 --- a/fs/gfs2/inode.c +++ b/fs/gfs2/inode.c @@ -248,12 +248,10 @@ static int gfs2_dinode_in(struct gfs2_inode *ip, const void *buf) { struct gfs2_dinode_host *di = &ip->i_di; const struct gfs2_dinode *str = buf; + u16 height; - if (ip->i_no_addr != be64_to_cpu(str->di_num.no_addr)) { - if (gfs2_consist_inode(ip)) - gfs2_dinode_print(ip); - return -EIO; - } + if (unlikely(ip->i_no_addr != be64_to_cpu(str->di_num.no_addr))) + goto corrupt; ip->i_no_formal_ino = be64_to_cpu(str->di_num.no_formal_ino); ip->i_inode.i_mode = be32_to_cpu(str->di_mode); ip->i_inode.i_rdev = 0; @@ -290,7 +288,10 @@ static int gfs2_dinode_in(struct gfs2_inode *ip, const void *buf) di->di_flags = be32_to_cpu(str->di_flags); gfs2_set_inode_flags(&ip->i_inode); - di->di_height = be16_to_cpu(str->di_height); + height = be16_to_cpu(str->di_height); + if (unlikely(height > GFS2_MAX_META_HEIGHT)) + goto corrupt; + ip->i_height = (u8)height; di->di_depth = be16_to_cpu(str->di_depth); di->di_entries = be32_to_cpu(str->di_entries); @@ -300,6 +301,10 @@ static int gfs2_dinode_in(struct gfs2_inode *ip, const void *buf) gfs2_set_aops(&ip->i_inode); return 0; +corrupt: + if (gfs2_consist_inode(ip)) + gfs2_dinode_print(ip); + return -EIO; } /** @@ -1401,7 +1406,7 @@ void gfs2_dinode_out(const struct gfs2_inode *ip, void *buf) str->di_generation = cpu_to_be64(di->di_generation); str->di_flags = cpu_to_be32(di->di_flags); - str->di_height = cpu_to_be16(di->di_height); + str->di_height = cpu_to_be16(ip->i_height); str->di_payload_format = cpu_to_be32(S_ISDIR(ip->i_inode.i_mode) && !(ip->i_di.di_flags & GFS2_DIF_EXHASH) ? GFS2_FORMAT_DE : 0); @@ -1430,7 +1435,6 @@ void gfs2_dinode_print(const struct gfs2_inode *ip) printk(KERN_INFO " di_goal_data = %llu\n", (unsigned long long)di->di_goal_data); printk(KERN_INFO " di_flags = 0x%.8X\n", di->di_flags); - printk(KERN_INFO " di_height = %u\n", di->di_height); printk(KERN_INFO " di_depth = %u\n", di->di_depth); printk(KERN_INFO " di_entries = %u\n", di->di_entries); printk(KERN_INFO " di_eattr = %llu\n", diff --git a/fs/gfs2/inode.h b/fs/gfs2/inode.h index d4465066261..db738686ca1 100644 --- a/fs/gfs2/inode.h +++ b/fs/gfs2/inode.h @@ -12,7 +12,7 @@ static inline int gfs2_is_stuffed(const struct gfs2_inode *ip) { - return !ip->i_di.di_height; + return !ip->i_height; } static inline int gfs2_is_jdata(const struct gfs2_inode *ip) diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c index ef0562c3bc7..88497b0b733 100644 --- a/fs/gfs2/super.c +++ b/fs/gfs2/super.c @@ -316,6 +316,7 @@ int gfs2_read_sb(struct gfs2_sbd *sdp, struct gfs2_glock *gl, int silent) sdp->sd_heightsize[x] = space; } sdp->sd_max_height = x; + sdp->sd_heightsize[x] = ~0; gfs2_assert(sdp, sdp->sd_max_height <= GFS2_MAX_META_HEIGHT); sdp->sd_jheightsize[0] = sdp->sd_sb.sb_bsize - @@ -334,6 +335,7 @@ int gfs2_read_sb(struct gfs2_sbd *sdp, struct gfs2_glock *gl, int silent) sdp->sd_jheightsize[x] = space; } sdp->sd_max_jheight = x; + sdp->sd_jheightsize[x] = ~0; gfs2_assert(sdp, sdp->sd_max_jheight <= GFS2_MAX_META_HEIGHT); return 0; -- cgit v1.2.3 From fe6c991c52a0dd07d4a19d392fd65048226cb1bc Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Mon, 28 Jan 2008 11:13:02 -0600 Subject: [GFS2] Get rid of unneeded parameter in gfs2_rlist_alloc This patch removed the unnecessary parameter from function gfs2_rlist_alloc. The parameter was always passed in as 0. Signed-off-by: Bob Peterson Signed-off-by: Steven Whitehouse --- fs/gfs2/bmap.c | 2 +- fs/gfs2/dir.c | 2 +- fs/gfs2/eattr.c | 2 +- fs/gfs2/rgrp.c | 7 +++---- fs/gfs2/rgrp.h | 5 ++--- 5 files changed, 8 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index 5a3187049dd..e8e48b690ce 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -680,7 +680,7 @@ static int do_strip(struct gfs2_inode *ip, struct buffer_head *dibh, else goto out; /* Nothing to do */ - gfs2_rlist_alloc(&rlist, LM_ST_EXCLUSIVE, 0); + gfs2_rlist_alloc(&rlist, LM_ST_EXCLUSIVE); for (x = 0; x < rlist.rl_rgrps; x++) { struct gfs2_rgrpd *rgd; diff --git a/fs/gfs2/dir.c b/fs/gfs2/dir.c index c34709512b1..78c236fb34e 100644 --- a/fs/gfs2/dir.c +++ b/fs/gfs2/dir.c @@ -1894,7 +1894,7 @@ static int leaf_dealloc(struct gfs2_inode *dip, u32 index, u32 len, l_blocks++; } - gfs2_rlist_alloc(&rlist, LM_ST_EXCLUSIVE, 0); + gfs2_rlist_alloc(&rlist, LM_ST_EXCLUSIVE); for (x = 0; x < rlist.rl_rgrps; x++) { struct gfs2_rgrpd *rgd; diff --git a/fs/gfs2/eattr.c b/fs/gfs2/eattr.c index bee99704ea1..04febbc17a1 100644 --- a/fs/gfs2/eattr.c +++ b/fs/gfs2/eattr.c @@ -1347,7 +1347,7 @@ static int ea_dealloc_indirect(struct gfs2_inode *ip) else goto out; - gfs2_rlist_alloc(&rlist, LM_ST_EXCLUSIVE, 0); + gfs2_rlist_alloc(&rlist, LM_ST_EXCLUSIVE); for (x = 0; x < rlist.rl_rgrps; x++) { struct gfs2_rgrpd *rgd; diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index 3552110b2e5..7b9d6f1d152 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -1,6 +1,6 @@ /* * Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved. - * Copyright (C) 2004-2007 Red Hat, Inc. All rights reserved. + * Copyright (C) 2004-2008 Red Hat, Inc. All rights reserved. * * This copyrighted material is made available to anyone wishing to use, * modify, copy, or redistribute it subject to the terms and conditions @@ -1699,8 +1699,7 @@ void gfs2_rlist_add(struct gfs2_sbd *sdp, struct gfs2_rgrp_list *rlist, * */ -void gfs2_rlist_alloc(struct gfs2_rgrp_list *rlist, unsigned int state, - int flags) +void gfs2_rlist_alloc(struct gfs2_rgrp_list *rlist, unsigned int state) { unsigned int x; @@ -1708,7 +1707,7 @@ void gfs2_rlist_alloc(struct gfs2_rgrp_list *rlist, unsigned int state, GFP_NOFS | __GFP_NOFAIL); for (x = 0; x < rlist->rl_rgrps; x++) gfs2_holder_init(rlist->rl_rgd[x]->rd_gl, - state, flags, + state, 0, &rlist->rl_ghs[x]); } diff --git a/fs/gfs2/rgrp.h b/fs/gfs2/rgrp.h index 149bb161f4b..5683605695f 100644 --- a/fs/gfs2/rgrp.h +++ b/fs/gfs2/rgrp.h @@ -1,6 +1,6 @@ /* * Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved. - * Copyright (C) 2004-2006 Red Hat, Inc. All rights reserved. + * Copyright (C) 2004-2008 Red Hat, Inc. All rights reserved. * * This copyrighted material is made available to anyone wishing to use, * modify, copy, or redistribute it subject to the terms and conditions @@ -64,8 +64,7 @@ struct gfs2_rgrp_list { void gfs2_rlist_add(struct gfs2_sbd *sdp, struct gfs2_rgrp_list *rlist, u64 block); -void gfs2_rlist_alloc(struct gfs2_rgrp_list *rlist, unsigned int state, - int flags); +void gfs2_rlist_alloc(struct gfs2_rgrp_list *rlist, unsigned int state); void gfs2_rlist_free(struct gfs2_rgrp_list *rlist); u64 gfs2_ri_total(struct gfs2_sbd *sdp); -- cgit v1.2.3 From ca390601a8bbb4ab8301a9469d23cdb1cf77e7cb Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Mon, 28 Jan 2008 11:15:57 -0600 Subject: [GFS2] Fix debug inode printing I noticed that the latest change to i_height got rid of the value from the inode dump. This patch adds it back. Signed-off-by: Bob Peterson Signed-off-by: Steven Whitehouse --- fs/gfs2/inode.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c index ff66ab7a17c..db5961a9aa5 100644 --- a/fs/gfs2/inode.c +++ b/fs/gfs2/inode.c @@ -1,6 +1,6 @@ /* * Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved. - * Copyright (C) 2004-2006 Red Hat, Inc. All rights reserved. + * Copyright (C) 2004-2008 Red Hat, Inc. All rights reserved. * * This copyrighted material is made available to anyone wishing to use, * modify, copy, or redistribute it subject to the terms and conditions @@ -1435,6 +1435,7 @@ void gfs2_dinode_print(const struct gfs2_inode *ip) printk(KERN_INFO " di_goal_data = %llu\n", (unsigned long long)di->di_goal_data); printk(KERN_INFO " di_flags = 0x%.8X\n", di->di_flags); + printk(KERN_INFO " i_height = %u\n", ip->i_height); printk(KERN_INFO " di_depth = %u\n", di->di_depth); printk(KERN_INFO " di_entries = %u\n", di->di_entries); printk(KERN_INFO " di_eattr = %llu\n", -- cgit v1.2.3 From d0109bfa84d6603becac8c2e87b3716f557f2039 Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Mon, 28 Jan 2008 11:20:10 -0600 Subject: [GFS2] Only do lo_incore_commit once This patch is performance related. When we're doing a log flush, I noticed we were calling buf_lo_incore_commit twice: once for data bufs and once for metadata bufs. Since this is the same function and does the same thing in both cases, there should be no reason to call it twice. Since we only need to call it once, we can also make it faster by removing it from the generic "lops" code and making it a stand-along static function. Signed-off-by: Bob Peterson Signed-off-by: Steven Whitehouse --- fs/gfs2/incore.h | 1 - fs/gfs2/log.c | 17 ++++++++++++++++- fs/gfs2/lops.c | 17 ----------------- fs/gfs2/lops.h | 11 +---------- 4 files changed, 17 insertions(+), 29 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index 43472baa92e..7ae12d2b626 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -44,7 +44,6 @@ struct gfs2_log_header_host { struct gfs2_log_operations { void (*lo_add) (struct gfs2_sbd *sdp, struct gfs2_log_element *le); - void (*lo_incore_commit) (struct gfs2_sbd *sdp, struct gfs2_trans *tr); void (*lo_before_commit) (struct gfs2_sbd *sdp); void (*lo_after_commit) (struct gfs2_sbd *sdp, struct gfs2_ail *ai); void (*lo_before_scan) (struct gfs2_jdesc *jd, diff --git a/fs/gfs2/log.c b/fs/gfs2/log.c index 161ab6f2058..b335304fc5d 100644 --- a/fs/gfs2/log.c +++ b/fs/gfs2/log.c @@ -779,6 +779,21 @@ static void log_refund(struct gfs2_sbd *sdp, struct gfs2_trans *tr) gfs2_log_unlock(sdp); } +static void buf_lo_incore_commit(struct gfs2_sbd *sdp, struct gfs2_trans *tr) +{ + struct list_head *head = &tr->tr_list_buf; + struct gfs2_bufdata *bd; + + gfs2_log_lock(sdp); + while (!list_empty(head)) { + bd = list_entry(head->next, struct gfs2_bufdata, bd_list_tr); + list_del_init(&bd->bd_list_tr); + tr->tr_num_buf--; + } + gfs2_log_unlock(sdp); + gfs2_assert_warn(sdp, !tr->tr_num_buf); +} + /** * gfs2_log_commit - Commit a transaction to the log * @sdp: the filesystem @@ -790,7 +805,7 @@ static void log_refund(struct gfs2_sbd *sdp, struct gfs2_trans *tr) void gfs2_log_commit(struct gfs2_sbd *sdp, struct gfs2_trans *tr) { log_refund(sdp, tr); - lops_incore_commit(sdp, tr); + buf_lo_incore_commit(sdp, tr); sdp->sd_vfs->s_dirt = 1; up_read(&sdp->sd_log_flush_lock); diff --git a/fs/gfs2/lops.c b/fs/gfs2/lops.c index fae59d69d01..71387372c88 100644 --- a/fs/gfs2/lops.c +++ b/fs/gfs2/lops.c @@ -152,21 +152,6 @@ out: unlock_buffer(bd->bd_bh); } -static void buf_lo_incore_commit(struct gfs2_sbd *sdp, struct gfs2_trans *tr) -{ - struct list_head *head = &tr->tr_list_buf; - struct gfs2_bufdata *bd; - - gfs2_log_lock(sdp); - while (!list_empty(head)) { - bd = list_entry(head->next, struct gfs2_bufdata, bd_list_tr); - list_del_init(&bd->bd_list_tr); - tr->tr_num_buf--; - } - gfs2_log_unlock(sdp); - gfs2_assert_warn(sdp, !tr->tr_num_buf); -} - static void buf_lo_before_commit(struct gfs2_sbd *sdp) { struct buffer_head *bh; @@ -737,7 +722,6 @@ static void databuf_lo_after_commit(struct gfs2_sbd *sdp, struct gfs2_ail *ai) const struct gfs2_log_operations gfs2_buf_lops = { .lo_add = buf_lo_add, - .lo_incore_commit = buf_lo_incore_commit, .lo_before_commit = buf_lo_before_commit, .lo_after_commit = buf_lo_after_commit, .lo_before_scan = buf_lo_before_scan, @@ -763,7 +747,6 @@ const struct gfs2_log_operations gfs2_rg_lops = { const struct gfs2_log_operations gfs2_databuf_lops = { .lo_add = databuf_lo_add, - .lo_incore_commit = buf_lo_incore_commit, .lo_before_commit = databuf_lo_before_commit, .lo_after_commit = databuf_lo_after_commit, .lo_scan_elements = databuf_lo_scan_elements, diff --git a/fs/gfs2/lops.h b/fs/gfs2/lops.h index 41a00df7558..3c0b2737658 100644 --- a/fs/gfs2/lops.h +++ b/fs/gfs2/lops.h @@ -1,6 +1,6 @@ /* * Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved. - * Copyright (C) 2004-2006 Red Hat, Inc. All rights reserved. + * Copyright (C) 2004-2008 Red Hat, Inc. All rights reserved. * * This copyrighted material is made available to anyone wishing to use, * modify, copy, or redistribute it subject to the terms and conditions @@ -57,15 +57,6 @@ static inline void lops_add(struct gfs2_sbd *sdp, struct gfs2_log_element *le) le->le_ops->lo_add(sdp, le); } -static inline void lops_incore_commit(struct gfs2_sbd *sdp, - struct gfs2_trans *tr) -{ - int x; - for (x = 0; gfs2_log_ops[x]; x++) - if (gfs2_log_ops[x]->lo_incore_commit) - gfs2_log_ops[x]->lo_incore_commit(sdp, tr); -} - static inline void lops_before_commit(struct gfs2_sbd *sdp) { int x; -- cgit v1.2.3 From 7eabb77e65c559d9c284da232b9ba5354898028a Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Mon, 28 Jan 2008 11:24:35 -0600 Subject: [GFS2] Misc fixups This patch contains two small fixups that didn't fit elsewhere. They are: (1) get rid of temp variable in find_metapath. (2) Remove vestigial "ret" variable from gfs2_writepage_common. Signed-off-by: Bob Peterson Signed-off-by: Steven Whitehouse --- fs/gfs2/bmap.c | 3 +-- fs/gfs2/ops_address.c | 4 +--- 2 files changed, 2 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index e8e48b690ce..e84e3845a39 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -305,11 +305,10 @@ static void find_metapath(struct gfs2_inode *ip, u64 block, struct metapath *mp) { struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); - u64 b = block; unsigned int i; for (i = ip->i_height; i--;) - mp->mp_list[i] = do_div(b, sdp->sd_inptrs); + mp->mp_list[i] = do_div(block, sdp->sd_inptrs); } diff --git a/fs/gfs2/ops_address.c b/fs/gfs2/ops_address.c index ac772b6d9db..7523999afc5 100644 --- a/fs/gfs2/ops_address.c +++ b/fs/gfs2/ops_address.c @@ -1,6 +1,6 @@ /* * Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved. - * Copyright (C) 2004-2007 Red Hat, Inc. All rights reserved. + * Copyright (C) 2004-2008 Red Hat, Inc. All rights reserved. * * This copyrighted material is made available to anyone wishing to use, * modify, copy, or redistribute it subject to the terms and conditions @@ -104,11 +104,9 @@ static int gfs2_writepage_common(struct page *page, loff_t i_size = i_size_read(inode); pgoff_t end_index = i_size >> PAGE_CACHE_SHIFT; unsigned offset; - int ret = -EIO; if (gfs2_assert_withdraw(sdp, gfs2_glock_is_held_excl(ip->i_gl))) goto out; - ret = 0; if (current->journal_info) goto redirty; /* Is the page fully outside i_size? (truncate in progress) */ -- cgit v1.2.3 From ef8c441cb7fece75dbbdb1f59d3f82b6a4be7474 Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Mon, 28 Jan 2008 14:54:16 -0600 Subject: [GFS2] Only wake the reclaim daemon if we need to This patch only wakes up the glock reclaim daemon if there is actually something to be reclaimed. Signed-off-by: Bob Peterson Signed-off-by: Steven Whitehouse --- fs/gfs2/glock.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index 7175a4d0643..5752dec017c 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -1594,10 +1594,10 @@ void gfs2_glock_schedule_for_reclaim(struct gfs2_glock *gl) gfs2_glock_hold(gl); list_add(&gl->gl_reclaim, &sdp->sd_reclaim_list); atomic_inc(&sdp->sd_reclaim_count); - } - spin_unlock(&sdp->sd_reclaim_lock); - - wake_up(&sdp->sd_reclaim_wq); + spin_unlock(&sdp->sd_reclaim_lock); + wake_up(&sdp->sd_reclaim_wq); + } else + spin_unlock(&sdp->sd_reclaim_lock); } /** -- cgit v1.2.3 From 048786f1e6042022a8fb2035157a8c8c3a82a4f2 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Tue, 29 Jan 2008 00:11:34 +0200 Subject: [GFS2] make gfs2_glock_hold() static gfs2_glock_hold() can now become static. Signed-off-by: Adrian Bunk Signed-off-by: Steven Whitehouse --- fs/gfs2/glock.c | 2 +- fs/gfs2/glock.h | 1 - 2 files changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index 5752dec017c..befcda0e53a 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -197,7 +197,7 @@ static void glock_free(struct gfs2_glock *gl) * */ -void gfs2_glock_hold(struct gfs2_glock *gl) +static void gfs2_glock_hold(struct gfs2_glock *gl) { atomic_inc(&gl->gl_ref); } diff --git a/fs/gfs2/glock.h b/fs/gfs2/glock.h index 2f9c6d136b3..ace5770760c 100644 --- a/fs/gfs2/glock.h +++ b/fs/gfs2/glock.h @@ -79,7 +79,6 @@ static inline int gfs2_glock_is_blocking(struct gfs2_glock *gl) int gfs2_glock_get(struct gfs2_sbd *sdp, u64 number, const struct gfs2_glock_operations *glops, int create, struct gfs2_glock **glp); -void gfs2_glock_hold(struct gfs2_glock *gl); int gfs2_glock_put(struct gfs2_glock *gl); void gfs2_holder_init(struct gfs2_glock *gl, unsigned int state, unsigned flags, struct gfs2_holder *gh); -- cgit v1.2.3 From 3ad62e87cd38817361e165cf4ad496ab76e19e81 Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Mon, 28 Jan 2008 16:35:13 -0600 Subject: [GFS2] Plug an unlikely leak Signed-off-by: Bob Peterson Signed-off-by: Steven Whitehouse --- fs/gfs2/lops.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/gfs2/lops.c b/fs/gfs2/lops.c index 71387372c88..4390f6f4047 100644 --- a/fs/gfs2/lops.c +++ b/fs/gfs2/lops.c @@ -404,8 +404,10 @@ static int revoke_lo_scan_elements(struct gfs2_jdesc *jd, unsigned int start, blkno = be64_to_cpu(*(__be64 *)(bh->b_data + offset)); error = gfs2_revoke_add(sdp, blkno, start); - if (error < 0) + if (error < 0) { + brelse(bh); return error; + } else if (error) sdp->sd_found_revokes++; -- cgit v1.2.3 From 6bdd9be628fa5f4dd14eb89ebddc12840d684277 Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Mon, 28 Jan 2008 17:20:26 -0600 Subject: [GFS2] Allocate gfs2_rgrpd from slab memory This patch moves the gfs2_rgrpd structure to its own slab memory. This makes it easier to control and monitor, and yields less memory fragmentation. Signed-off-by: Bob Peterson Signed-off-by: Steven Whitehouse --- fs/gfs2/main.c | 10 ++++++++++ fs/gfs2/rgrp.c | 4 ++-- fs/gfs2/util.c | 1 + fs/gfs2/util.h | 1 + 4 files changed, 14 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/main.c b/fs/gfs2/main.c index 9c7765c12d6..053e2ebbbd5 100644 --- a/fs/gfs2/main.c +++ b/fs/gfs2/main.c @@ -89,6 +89,12 @@ static int __init init_gfs2_fs(void) if (!gfs2_bufdata_cachep) goto fail; + gfs2_rgrpd_cachep = kmem_cache_create("gfs2_rgrpd", + sizeof(struct gfs2_rgrpd), + 0, 0, NULL); + if (!gfs2_rgrpd_cachep) + goto fail; + error = register_filesystem(&gfs2_fs_type); if (error) goto fail; @@ -108,6 +114,9 @@ fail_unregister: fail: gfs2_glock_exit(); + if (gfs2_rgrpd_cachep) + kmem_cache_destroy(gfs2_rgrpd_cachep); + if (gfs2_bufdata_cachep) kmem_cache_destroy(gfs2_bufdata_cachep); @@ -133,6 +142,7 @@ static void __exit exit_gfs2_fs(void) unregister_filesystem(&gfs2_fs_type); unregister_filesystem(&gfs2meta_fs_type); + kmem_cache_destroy(gfs2_rgrpd_cachep); kmem_cache_destroy(gfs2_bufdata_cachep); kmem_cache_destroy(gfs2_inode_cachep); kmem_cache_destroy(gfs2_glock_cachep); diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index 7b9d6f1d152..dc7e83eed32 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -353,7 +353,7 @@ static void clear_rgrpdi(struct gfs2_sbd *sdp) } kfree(rgd->rd_bits); - kfree(rgd); + kmem_cache_free(gfs2_rgrpd_cachep, rgd); } } @@ -516,7 +516,7 @@ static int read_rindex_entry(struct gfs2_inode *ip, return error; } - rgd = kzalloc(sizeof(struct gfs2_rgrpd), GFP_NOFS); + rgd = kmem_cache_zalloc(gfs2_rgrpd_cachep, GFP_NOFS); error = -ENOMEM; if (!rgd) return error; diff --git a/fs/gfs2/util.c b/fs/gfs2/util.c index 424a0774eda..fe9c28ef77b 100644 --- a/fs/gfs2/util.c +++ b/fs/gfs2/util.c @@ -25,6 +25,7 @@ struct kmem_cache *gfs2_glock_cachep __read_mostly; struct kmem_cache *gfs2_inode_cachep __read_mostly; struct kmem_cache *gfs2_bufdata_cachep __read_mostly; +struct kmem_cache *gfs2_rgrpd_cachep __read_mostly; void gfs2_assert_i(struct gfs2_sbd *sdp) { diff --git a/fs/gfs2/util.h b/fs/gfs2/util.h index 28938a46cf4..ac0c567ebc3 100644 --- a/fs/gfs2/util.h +++ b/fs/gfs2/util.h @@ -147,6 +147,7 @@ gfs2_io_error_bh_i((sdp), (bh), __FUNCTION__, __FILE__, __LINE__); extern struct kmem_cache *gfs2_glock_cachep; extern struct kmem_cache *gfs2_inode_cachep; extern struct kmem_cache *gfs2_bufdata_cachep; +extern struct kmem_cache *gfs2_rgrpd_cachep; static inline unsigned int gfs2_tune_get_i(struct gfs2_tune *gt, unsigned int *p) -- cgit v1.2.3 From 42d52e3818751633656fb90df1bd5cb5362fa8cc Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Mon, 28 Jan 2008 18:38:07 -0600 Subject: [GFS2] Combine rg_flags and rd_flags This patch reduces the memory required by GFS2 by combining the rd_flags and rg_flags (in core only). Signed-off-by: Bob Peterson Signed-off-by: Steven Whitehouse --- fs/gfs2/incore.h | 6 +++--- fs/gfs2/rgrp.c | 36 +++++++++++++++++++++++------------- 2 files changed, 26 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index 7ae12d2b626..39bab7b213e 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -69,7 +69,6 @@ struct gfs2_bitmap { }; struct gfs2_rgrp_host { - u32 rg_flags; u32 rg_free; u32 rg_dinodes; u64 rg_igeneration; @@ -95,8 +94,9 @@ struct gfs2_rgrpd { u32 rd_last_alloc_data; u32 rd_last_alloc_meta; struct gfs2_sbd *rd_sbd; - unsigned long rd_flags; -#define GFS2_RDF_CHECK 0x0001 /* Need to check for unlinked inodes */ + unsigned char rd_flags; +#define GFS2_RDF_CHECK 0x01 /* Need to check for unlinked inodes */ +#define GFS2_RDF_NOALLOC 0x02 /* rg prohibits allocation */ }; enum gfs2_state_bits { diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index dc7e83eed32..da60ce8c5d7 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -655,21 +655,31 @@ int gfs2_rindex_hold(struct gfs2_sbd *sdp, struct gfs2_holder *ri_gh) return error; } -static void gfs2_rgrp_in(struct gfs2_rgrp_host *rg, const void *buf) +static void gfs2_rgrp_in(struct gfs2_rgrpd *rgd, const void *buf) { const struct gfs2_rgrp *str = buf; + struct gfs2_rgrp_host *rg = &rgd->rd_rg; + u32 rg_flags; - rg->rg_flags = be32_to_cpu(str->rg_flags); + rg_flags = be32_to_cpu(str->rg_flags); + if (rg_flags & GFS2_RGF_NOALLOC) + rgd->rd_flags |= GFS2_RDF_NOALLOC; + else + rgd->rd_flags &= ~GFS2_RDF_NOALLOC; rg->rg_free = be32_to_cpu(str->rg_free); rg->rg_dinodes = be32_to_cpu(str->rg_dinodes); rg->rg_igeneration = be64_to_cpu(str->rg_igeneration); } -static void gfs2_rgrp_out(const struct gfs2_rgrp_host *rg, void *buf) +static void gfs2_rgrp_out(struct gfs2_rgrpd *rgd, void *buf) { struct gfs2_rgrp *str = buf; + struct gfs2_rgrp_host *rg = &rgd->rd_rg; + u32 rg_flags = 0; - str->rg_flags = cpu_to_be32(rg->rg_flags); + if (rgd->rd_flags & GFS2_RDF_NOALLOC) + rg_flags |= GFS2_RGF_NOALLOC; + str->rg_flags = cpu_to_be32(rg_flags); str->rg_free = cpu_to_be32(rg->rg_free); str->rg_dinodes = cpu_to_be32(rg->rg_dinodes); str->__pad = cpu_to_be32(0); @@ -727,7 +737,7 @@ int gfs2_rgrp_bh_get(struct gfs2_rgrpd *rgd) } if (rgd->rd_rg_vn != gl->gl_vn) { - gfs2_rgrp_in(&rgd->rd_rg, (rgd->rd_bits[0].bi_bh)->b_data); + gfs2_rgrp_in(rgd, (rgd->rd_bits[0].bi_bh)->b_data); rgd->rd_rg_vn = gl->gl_vn; } @@ -840,7 +850,7 @@ static int try_rgrp_fit(struct gfs2_rgrpd *rgd, struct gfs2_alloc *al) struct gfs2_sbd *sdp = rgd->rd_sbd; int ret = 0; - if (rgd->rd_rg.rg_flags & GFS2_RGF_NOALLOC) + if (rgd->rd_flags & GFS2_RDF_NOALLOC) return 0; spin_lock(&sdp->sd_rindex_spin); @@ -1431,7 +1441,7 @@ u64 gfs2_alloc_data(struct gfs2_inode *ip) rgd->rd_rg.rg_free--; gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1); - gfs2_rgrp_out(&rgd->rd_rg, rgd->rd_bits[0].bi_bh->b_data); + gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data); al->al_alloced++; @@ -1476,7 +1486,7 @@ u64 gfs2_alloc_meta(struct gfs2_inode *ip) rgd->rd_rg.rg_free--; gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1); - gfs2_rgrp_out(&rgd->rd_rg, rgd->rd_bits[0].bi_bh->b_data); + gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data); al->al_alloced++; @@ -1519,7 +1529,7 @@ u64 gfs2_alloc_di(struct gfs2_inode *dip, u64 *generation) rgd->rd_rg.rg_dinodes++; *generation = rgd->rd_rg.rg_igeneration++; gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1); - gfs2_rgrp_out(&rgd->rd_rg, rgd->rd_bits[0].bi_bh->b_data); + gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data); al->al_alloced++; @@ -1553,7 +1563,7 @@ void gfs2_free_data(struct gfs2_inode *ip, u64 bstart, u32 blen) rgd->rd_rg.rg_free += blen; gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1); - gfs2_rgrp_out(&rgd->rd_rg, rgd->rd_bits[0].bi_bh->b_data); + gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data); gfs2_trans_add_rg(rgd); @@ -1581,7 +1591,7 @@ void gfs2_free_meta(struct gfs2_inode *ip, u64 bstart, u32 blen) rgd->rd_rg.rg_free += blen; gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1); - gfs2_rgrp_out(&rgd->rd_rg, rgd->rd_bits[0].bi_bh->b_data); + gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data); gfs2_trans_add_rg(rgd); @@ -1601,7 +1611,7 @@ void gfs2_unlink_di(struct inode *inode) if (!rgd) return; gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1); - gfs2_rgrp_out(&rgd->rd_rg, rgd->rd_bits[0].bi_bh->b_data); + gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data); gfs2_trans_add_rg(rgd); } @@ -1621,7 +1631,7 @@ static void gfs2_free_uninit_di(struct gfs2_rgrpd *rgd, u64 blkno) rgd->rd_rg.rg_free++; gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1); - gfs2_rgrp_out(&rgd->rd_rg, rgd->rd_bits[0].bi_bh->b_data); + gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data); gfs2_statfs_change(sdp, 0, +1, -1); gfs2_trans_add_rg(rgd); -- cgit v1.2.3 From 29d38cd16358dcaef4a6c50866ecee28025b481a Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Mon, 28 Jan 2008 22:31:39 -0600 Subject: [GFS2] Get rid of gl_waiters2 This patch reduces memory by replacing the int variable gl_waiters2 by a single bit in the gl_flags. Signed-off-by: Bob Peterson Signed-off-by: Steven Whitehouse --- fs/gfs2/glock.c | 7 ++++--- fs/gfs2/incore.h | 4 ++-- 2 files changed, 6 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index befcda0e53a..951cb91e7dd 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -595,11 +595,12 @@ static void run_queue(struct gfs2_glock *gl) blocked = rq_mutex(gh); } else if (test_bit(GLF_DEMOTE, &gl->gl_flags)) { blocked = rq_demote(gl); - if (gl->gl_waiters2 && !blocked) { + if (test_bit(GLF_WAITERS2, &gl->gl_flags) && + !blocked) { set_bit(GLF_DEMOTE, &gl->gl_flags); gl->gl_demote_state = LM_ST_UNLOCKED; } - gl->gl_waiters2 = 0; + clear_bit(GLF_WAITERS2, &gl->gl_flags); } else if (!list_empty(&gl->gl_waiters3)) { gh = list_entry(gl->gl_waiters3.next, struct gfs2_holder, gh_list); @@ -710,7 +711,7 @@ static void handle_callback(struct gfs2_glock *gl, unsigned int state, } else if (gl->gl_demote_state != LM_ST_UNLOCKED && gl->gl_demote_state != state) { if (test_bit(GLF_DEMOTE_IN_PROGRESS, &gl->gl_flags)) - gl->gl_waiters2 = 1; + set_bit(GLF_WAITERS2, &gl->gl_flags); else gl->gl_demote_state = LM_ST_UNLOCKED; } diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index 39bab7b213e..fe14f6a417b 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -1,6 +1,6 @@ /* * Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved. - * Copyright (C) 2004-2007 Red Hat, Inc. All rights reserved. + * Copyright (C) 2004-2008 Red Hat, Inc. All rights reserved. * * This copyrighted material is made available to anyone wishing to use, * modify, copy, or redistribute it subject to the terms and conditions @@ -167,6 +167,7 @@ enum { GLF_DIRTY = 5, GLF_DEMOTE_IN_PROGRESS = 6, GLF_LFLUSH = 7, + GLF_WAITERS2 = 8, }; struct gfs2_glock { @@ -186,7 +187,6 @@ struct gfs2_glock { struct list_head gl_holders; struct list_head gl_waiters1; /* HIF_MUTEX */ struct list_head gl_waiters3; /* HIF_PROMOTE */ - int gl_waiters2; /* GIF_DEMOTE */ const struct gfs2_glock_operations *gl_ops; -- cgit v1.2.3 From 11707ea05e85290d10c482b87e195c198f5eb3cf Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Mon, 28 Jan 2008 15:10:29 +0000 Subject: [GFS2] Move part of gfs2_block_map into a separate function This is required to enable future changes to the block mapping code. Signed-off-by: Steven Whitehouse --- fs/gfs2/bmap.c | 57 ++++++++++++++++++++++++++++++++++++--------------------- 1 file changed, 36 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index e84e3845a39..08d1be492ef 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -384,6 +384,35 @@ static int lookup_block(struct gfs2_inode *ip, struct buffer_head *bh, return 0; } +static int lookup_metapath(struct inode *inode, struct metapath *mp, + int create, int *new, u64 *dblock, + struct buffer_head **dibh, struct buffer_head **bh) +{ + struct gfs2_inode *ip = GFS2_I(inode); + unsigned int end_of_metadata = ip->i_height - 1; + unsigned int x; + int ret = gfs2_meta_inode_buffer(ip, bh); + if (ret) + return ret; + + *dibh = *bh; + get_bh(*dibh); + + for (x = 0; x < end_of_metadata; x++) { + lookup_block(ip, *bh, x, mp, create, new, dblock); + brelse(*bh); + *bh = NULL; + if (!dblock) + return 0; + + ret = gfs2_meta_indirect_buffer(ip, x+1, *dblock, *new, bh); + if (ret) + return ret; + } + + return lookup_block(ip, *bh, end_of_metadata, mp, create, new, dblock); +} + static inline void bmap_lock(struct inode *inode, int create) { struct gfs2_inode *ip = GFS2_I(inode); @@ -419,10 +448,8 @@ int gfs2_block_map(struct inode *inode, sector_t lblock, { struct gfs2_inode *ip = GFS2_I(inode); struct gfs2_sbd *sdp = GFS2_SB(inode); - struct buffer_head *bh; unsigned int bsize = sdp->sd_sb.sb_bsize; - unsigned int end_of_metadata; - unsigned int x; + struct buffer_head *bh = NULL; int error = 0; int new = 0; u64 dblock = 0; @@ -459,25 +486,11 @@ int gfs2_block_map(struct inode *inode, sector_t lblock, } find_metapath(ip, lblock, &mp); - end_of_metadata = ip->i_height - 1; - error = gfs2_meta_inode_buffer(ip, &bh); - if (error) + error = lookup_metapath(inode, &mp, create, &new, &dblock, &dibh, &bh); + if (error < 0) goto out_fail; - dibh = bh; - get_bh(dibh); - - for (x = 0; x < end_of_metadata; x++) { - lookup_block(ip, bh, x, &mp, create, &new, &dblock); - brelse(bh); - if (!dblock) - goto out_ok; + boundary = error; - error = gfs2_meta_indirect_buffer(ip, x+1, dblock, new, &bh); - if (error) - goto out_fail; - } - - boundary = lookup_block(ip, bh, end_of_metadata, &mp, create, &new, &dblock); if (dblock) { map_bh(bh_map, inode->i_sb, dblock); if (boundary) @@ -489,6 +502,7 @@ int gfs2_block_map(struct inode *inode, sector_t lblock, goto out_brelse; } while(--maxlen && !buffer_boundary(bh_map)) { + unsigned int end_of_metadata = ip->i_height - 1; u64 eblock; mp.mp_list[end_of_metadata]++; @@ -501,7 +515,8 @@ int gfs2_block_map(struct inode *inode, sector_t lblock, } } out_brelse: - brelse(bh); + if (bh) + brelse(bh); out_ok: error = 0; out_fail: -- cgit v1.2.3 From dbac6710a6dfcec7fbe7d9571c183d86a4237623 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Tue, 29 Jan 2008 09:12:55 +0000 Subject: [GFS2] Introduce array of buffers to struct metapath The reason for doing this is to allow all the block mapping code to share the same array. As a result we can remove two arguments from lookup_metapath since they are now returned via the array. We also add a function to drop all refs to buffer heads when we are done with the metapath. The build_height function shares the struct metapath, but currently still frees its own buffers, and this will change in a future patch. Signed-off-by: Steven Whitehouse --- fs/gfs2/bmap.c | 97 +++++++++++++++++++++++++++++----------------------------- 1 file changed, 48 insertions(+), 49 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index 08d1be492ef..2011dd27f8d 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -33,6 +33,7 @@ * keep it small. */ struct metapath { + struct buffer_head *mp_bh[GFS2_MAX_META_HEIGHT]; __u16 mp_list[GFS2_MAX_META_HEIGHT]; }; @@ -185,12 +186,11 @@ out: * Returns: errno */ -static int build_height(struct inode *inode, unsigned height) +static int build_height(struct inode *inode, struct metapath *mp, unsigned height) { struct gfs2_inode *ip = GFS2_I(inode); unsigned new_height = height - ip->i_height; struct buffer_head *dibh; - struct buffer_head *blocks[GFS2_MAX_META_HEIGHT]; struct gfs2_dinode *di; int error; __be64 *bp; @@ -206,29 +206,29 @@ static int build_height(struct inode *inode, unsigned height) for(n = 0; n < new_height; n++) { bn = gfs2_alloc_meta(ip); - blocks[n] = gfs2_meta_new(ip->i_gl, bn); - gfs2_trans_add_bh(ip->i_gl, blocks[n], 1); + mp->mp_bh[n] = gfs2_meta_new(ip->i_gl, bn); + gfs2_trans_add_bh(ip->i_gl, mp->mp_bh[n], 1); } n = 0; - bn = blocks[0]->b_blocknr; + bn = mp->mp_bh[0]->b_blocknr; if (new_height > 1) { for(; n < new_height-1; n++) { - gfs2_metatype_set(blocks[n], GFS2_METATYPE_IN, + gfs2_metatype_set(mp->mp_bh[n], GFS2_METATYPE_IN, GFS2_FORMAT_IN); - gfs2_buffer_clear_tail(blocks[n], + gfs2_buffer_clear_tail(mp->mp_bh[n], sizeof(struct gfs2_meta_header)); - bp = (__be64 *)(blocks[n]->b_data + + bp = (__be64 *)(mp->mp_bh[n]->b_data + sizeof(struct gfs2_meta_header)); - *bp = cpu_to_be64(blocks[n+1]->b_blocknr); - brelse(blocks[n]); - blocks[n] = NULL; + *bp = cpu_to_be64(mp->mp_bh[n+1]->b_blocknr); + brelse(mp->mp_bh[n]); + mp->mp_bh[n] = NULL; } } - gfs2_metatype_set(blocks[n], GFS2_METATYPE_IN, GFS2_FORMAT_IN); - gfs2_buffer_copy_tail(blocks[n], sizeof(struct gfs2_meta_header), + gfs2_metatype_set(mp->mp_bh[n], GFS2_METATYPE_IN, GFS2_FORMAT_IN); + gfs2_buffer_copy_tail(mp->mp_bh[n], sizeof(struct gfs2_meta_header), dibh, sizeof(struct gfs2_dinode)); - brelse(blocks[n]); + brelse(mp->mp_bh[n]); gfs2_trans_add_bh(ip->i_gl, dibh, 1); di = (struct gfs2_dinode *)dibh->b_data; gfs2_buffer_clear_tail(dibh, sizeof(struct gfs2_dinode)); @@ -314,7 +314,6 @@ static void find_metapath(struct gfs2_inode *ip, u64 block, /** * metapointer - Return pointer to start of metadata in a buffer - * @bh: The buffer * @height: The metadata height (0 = dinode) * @mp: The metapath * @@ -323,9 +322,10 @@ static void find_metapath(struct gfs2_inode *ip, u64 block, * metadata tree. */ -static inline __be64 *metapointer(struct buffer_head *bh, int *boundary, - unsigned int height, const struct metapath *mp) +static inline __be64 *metapointer(int *boundary, unsigned int height, + const struct metapath *mp) { + struct buffer_head *bh = mp->mp_bh[height]; unsigned int head_size = (height > 0) ? sizeof(struct gfs2_meta_header) : sizeof(struct gfs2_dinode); __be64 *ptr; @@ -339,7 +339,6 @@ static inline __be64 *metapointer(struct buffer_head *bh, int *boundary, /** * lookup_block - Get the next metadata block in metadata tree * @ip: The GFS2 inode - * @bh: Buffer containing the pointers to metadata blocks * @height: The height of the tree (0 = dinode) * @mp: The metapath * @create: Non-zero if we may create a new meatdata block @@ -352,12 +351,12 @@ static inline __be64 *metapointer(struct buffer_head *bh, int *boundary, * */ -static int lookup_block(struct gfs2_inode *ip, struct buffer_head *bh, - unsigned int height, struct metapath *mp, int create, +static int lookup_block(struct gfs2_inode *ip, unsigned int height, + struct metapath *mp, int create, int *new, u64 *block) { int boundary; - __be64 *ptr = metapointer(bh, &boundary, height, mp); + __be64 *ptr = metapointer(&boundary, height, mp); if (*ptr) { *block = be64_to_cpu(*ptr); @@ -374,7 +373,7 @@ static int lookup_block(struct gfs2_inode *ip, struct buffer_head *bh, else *block = gfs2_alloc_meta(ip); - gfs2_trans_add_bh(ip->i_gl, bh, 1); + gfs2_trans_add_bh(ip->i_gl, mp->mp_bh[height], 1); *ptr = cpu_to_be64(*block); ip->i_di.di_blocks++; @@ -385,32 +384,38 @@ static int lookup_block(struct gfs2_inode *ip, struct buffer_head *bh, } static int lookup_metapath(struct inode *inode, struct metapath *mp, - int create, int *new, u64 *dblock, - struct buffer_head **dibh, struct buffer_head **bh) + int create, int *new, u64 *dblock) { + struct buffer_head *bh; struct gfs2_inode *ip = GFS2_I(inode); unsigned int end_of_metadata = ip->i_height - 1; unsigned int x; - int ret = gfs2_meta_inode_buffer(ip, bh); + int ret = gfs2_meta_inode_buffer(ip, &bh); if (ret) return ret; - *dibh = *bh; - get_bh(*dibh); + mp->mp_bh[0] = bh; for (x = 0; x < end_of_metadata; x++) { - lookup_block(ip, *bh, x, mp, create, new, dblock); - brelse(*bh); - *bh = NULL; + lookup_block(ip, x, mp, create, new, dblock); if (!dblock) return 0; - ret = gfs2_meta_indirect_buffer(ip, x+1, *dblock, *new, bh); + ret = gfs2_meta_indirect_buffer(ip, x+1, *dblock, *new, &mp->mp_bh[x+1]); if (ret) return ret; } - return lookup_block(ip, *bh, end_of_metadata, mp, create, new, dblock); + return lookup_block(ip, end_of_metadata, mp, create, new, dblock); +} + +static void release_metapath(struct metapath *mp) +{ + int i; + + for (i = 0; i < GFS2_MAX_META_HEIGHT; i++) + if (mp->mp_bh[i]) + brelse(mp->mp_bh[i]); } static inline void bmap_lock(struct inode *inode, int create) @@ -449,7 +454,6 @@ int gfs2_block_map(struct inode *inode, sector_t lblock, struct gfs2_inode *ip = GFS2_I(inode); struct gfs2_sbd *sdp = GFS2_SB(inode); unsigned int bsize = sdp->sd_sb.sb_bsize; - struct buffer_head *bh = NULL; int error = 0; int new = 0; u64 dblock = 0; @@ -457,13 +461,13 @@ int gfs2_block_map(struct inode *inode, sector_t lblock, unsigned int maxlen = bh_map->b_size >> inode->i_blkbits; struct metapath mp; u64 size; - struct buffer_head *dibh = NULL; const u64 *arr = sdp->sd_heightsize; BUG_ON(maxlen == 0); if (gfs2_assert_warn(sdp, !gfs2_is_stuffed(ip))) return 0; + memset(mp.mp_bh, 0, sizeof(mp.mp_bh)); bmap_lock(inode, create); clear_buffer_mapped(bh_map); clear_buffer_new(bh_map); @@ -480,13 +484,13 @@ int gfs2_block_map(struct inode *inode, sector_t lblock, goto out_ok; while (size > arr[height]) height++; - error = build_height(inode, height); + error = build_height(inode, &mp, height); if (error) goto out_fail; } find_metapath(ip, lblock, &mp); - error = lookup_metapath(inode, &mp, create, &new, &dblock, &dibh, &bh); + error = lookup_metapath(inode, &mp, create, &new, &dblock); if (error < 0) goto out_fail; boundary = error; @@ -496,17 +500,15 @@ int gfs2_block_map(struct inode *inode, sector_t lblock, if (boundary) set_buffer_boundary(bh_map); if (new) { - gfs2_trans_add_bh(ip->i_gl, dibh, 1); - gfs2_dinode_out(ip, dibh->b_data); + gfs2_trans_add_bh(ip->i_gl, mp.mp_bh[0], 1); + gfs2_dinode_out(ip, mp.mp_bh[0]->b_data); set_buffer_new(bh_map); - goto out_brelse; + goto out_ok; } while(--maxlen && !buffer_boundary(bh_map)) { - unsigned int end_of_metadata = ip->i_height - 1; u64 eblock; - - mp.mp_list[end_of_metadata]++; - boundary = lookup_block(ip, bh, end_of_metadata, &mp, 0, &new, &eblock); + mp.mp_list[ip->i_height - 1]++; + boundary = lookup_block(ip, ip->i_height - 1, &mp, 0, &new, &eblock); if (eblock != ++dblock) break; bh_map->b_size += (1 << inode->i_blkbits); @@ -514,14 +516,10 @@ int gfs2_block_map(struct inode *inode, sector_t lblock, set_buffer_boundary(bh_map); } } -out_brelse: - if (bh) - brelse(bh); out_ok: error = 0; out_fail: - if (dibh) - brelse(dibh); + release_metapath(&mp); bmap_unlock(inode, create); return error; } @@ -819,10 +817,11 @@ static int do_grow(struct gfs2_inode *ip, u64 size) down_write(&ip->i_rw_mutex); if (size > arr[ip->i_height]) { + struct metapath mp; u8 height = ip->i_height; while(size > arr[height]) height++; - error = build_height(&ip->i_inode, height); + error = build_height(&ip->i_inode, &mp, height); } up_write(&ip->i_rw_mutex); if (error) -- cgit v1.2.3 From 110acf38377b5b341b11644bfe98389eccec627d Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Tue, 29 Jan 2008 13:30:20 +0000 Subject: [GFS2] Add consts to various bits of rgrp.c There are a couple of routines which scan bitmaps where we can mark the bitmaps const, plus a couple of call sites that can be updated too. Signed-off-by: Steven Whitehouse --- fs/gfs2/rgrp.c | 44 ++++++++++++++++++++++---------------------- 1 file changed, 22 insertions(+), 22 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index da60ce8c5d7..5fd87104e59 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -126,23 +126,24 @@ static unsigned char gfs2_testbit(struct gfs2_rgrpd *rgd, unsigned char *buffer, * Return: the block number (bitmap buffer scope) that was found */ -static u32 gfs2_bitfit(unsigned char *buffer, unsigned int buflen, u32 goal, - unsigned char old_state) +static u32 gfs2_bitfit(const u8 *buffer, unsigned int buflen, u32 goal, + u8 old_state) { - unsigned char *byte; + const u8 *byte; u32 blk = goal; unsigned int bit, bitlong; - unsigned long *plong, plong55; + const unsigned long *plong; +#if BITS_PER_LONG == 32 + const unsigned long plong55 = 0x55555555; +#else + const unsigned long plong55 = 0x5555555555555555; +#endif byte = buffer + (goal / GFS2_NBBY); - plong = (unsigned long *)(buffer + (goal / GFS2_NBBY)); + plong = (const unsigned long *)(buffer + (goal / GFS2_NBBY)); bit = (goal % GFS2_NBBY) * GFS2_BIT_SIZE; bitlong = bit; -#if BITS_PER_LONG == 32 - plong55 = 0x55555555; -#else - plong55 = 0x5555555555555555; -#endif + while (byte < buffer + buflen) { if (bitlong == 0 && old_state == 0 && *plong == plong55) { @@ -179,14 +180,14 @@ static u32 gfs2_bitfit(unsigned char *buffer, unsigned int buflen, u32 goal, * Returns: The number of bits */ -static u32 gfs2_bitcount(struct gfs2_rgrpd *rgd, unsigned char *buffer, - unsigned int buflen, unsigned char state) +static u32 gfs2_bitcount(struct gfs2_rgrpd *rgd, const u8 *buffer, + unsigned int buflen, u8 state) { - unsigned char *byte = buffer; - unsigned char *end = buffer + buflen; - unsigned char state1 = state << 2; - unsigned char state2 = state << 4; - unsigned char state3 = state << 6; + const u8 *byte = buffer; + const u8 *end = buffer + buflen; + const u8 state1 = state << 2; + const u8 state2 = state << 4; + const u8 state3 = state << 6; u32 count = 0; for (; byte < end; byte++) { @@ -1327,12 +1328,11 @@ static u32 rgblk_search(struct gfs2_rgrpd *rgd, u32 goal, for (x = 0; x <= length; x++) { /* The GFS2_BLKST_UNLINKED state doesn't apply to the clone bitmaps, so we must search the originals for that. */ + const u8 *buffer = bi->bi_bh->b_data + bi->bi_offset; if (old_state != GFS2_BLKST_UNLINKED && bi->bi_clone) - blk = gfs2_bitfit(bi->bi_clone + bi->bi_offset, - bi->bi_len, goal, old_state); - else - blk = gfs2_bitfit(bi->bi_bh->b_data + bi->bi_offset, - bi->bi_len, goal, old_state); + buffer = bi->bi_clone + bi->bi_offset; + + blk = gfs2_bitfit(buffer, bi->bi_len, goal, old_state); if (blk != BFITNOENT) break; -- cgit v1.2.3 From ab0d756681c9502a2ab9e2e4ab3685bc0567f4ee Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Tue, 29 Jan 2008 13:56:15 -0600 Subject: [GFS2] Eliminate gl_req_bh This patch further reduces the memory needs of GFS2 by eliminating the gl_req_bh variable from struct gfs2_glock. Signed-off-by: Bob Peterson Signed-off-by: Steven Whitehouse --- fs/gfs2/glock.c | 107 ++++++++++++++++++++++++++----------------------------- fs/gfs2/incore.h | 1 - 2 files changed, 51 insertions(+), 57 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index 951cb91e7dd..a8387e0b506 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -338,7 +338,6 @@ int gfs2_glock_get(struct gfs2_sbd *sdp, u64 number, gl->gl_ip = 0; gl->gl_ops = glops; gl->gl_req_gh = NULL; - gl->gl_req_bh = NULL; gl->gl_vn = 0; gl->gl_stamp = jiffies; gl->gl_tchange = jiffies; @@ -743,6 +742,50 @@ static void state_change(struct gfs2_glock *gl, unsigned int new_state) gl->gl_tchange = jiffies; } +/** + * drop_bh - Called after a lock module unlock completes + * @gl: the glock + * @ret: the return status + * + * Doesn't wake up the process waiting on the struct gfs2_holder (if any) + * Doesn't drop the reference on the glock the top half took out + * + */ + +static void drop_bh(struct gfs2_glock *gl, unsigned int ret) +{ + struct gfs2_sbd *sdp = gl->gl_sbd; + const struct gfs2_glock_operations *glops = gl->gl_ops; + struct gfs2_holder *gh = gl->gl_req_gh; + + gfs2_assert_warn(sdp, test_bit(GLF_LOCK, &gl->gl_flags)); + gfs2_assert_warn(sdp, list_empty(&gl->gl_holders)); + gfs2_assert_warn(sdp, !ret); + + state_change(gl, LM_ST_UNLOCKED); + + if (glops->go_inval) + glops->go_inval(gl, DIO_METADATA); + + if (gh) { + spin_lock(&gl->gl_spin); + list_del_init(&gh->gh_list); + gh->gh_error = 0; + spin_unlock(&gl->gl_spin); + } + + spin_lock(&gl->gl_spin); + gfs2_demote_wake(gl); + gl->gl_req_gh = NULL; + clear_bit(GLF_LOCK, &gl->gl_flags); + spin_unlock(&gl->gl_spin); + + gfs2_glock_put(gl); + + if (gh) + gfs2_holder_wake(gh); +} + /** * xmote_bh - Called after the lock module is done acquiring a lock * @gl: The glock in question @@ -758,6 +801,11 @@ static void xmote_bh(struct gfs2_glock *gl, unsigned int ret) int prev_state = gl->gl_state; int op_done = 1; + if ((ret & LM_OUT_ST_MASK) == LM_ST_UNLOCKED) { + drop_bh(gl, ret); + return; + } + gfs2_assert_warn(sdp, test_bit(GLF_LOCK, &gl->gl_flags)); gfs2_assert_warn(sdp, list_empty(&gl->gl_holders)); gfs2_assert_warn(sdp, !(ret & LM_OUT_ASYNC)); @@ -783,7 +831,6 @@ static void xmote_bh(struct gfs2_glock *gl, unsigned int ret) } else { spin_lock(&gl->gl_spin); if (gl->gl_state != gl->gl_demote_state) { - gl->gl_req_bh = NULL; spin_unlock(&gl->gl_spin); gfs2_glock_drop_th(gl); gfs2_glock_put(gl); @@ -825,7 +872,6 @@ out: if (op_done) { spin_lock(&gl->gl_spin); gl->gl_req_gh = NULL; - gl->gl_req_bh = NULL; clear_bit(GLF_LOCK, &gl->gl_flags); spin_unlock(&gl->gl_spin); } @@ -864,7 +910,6 @@ static void gfs2_glock_xmote_th(struct gfs2_glock *gl, struct gfs2_holder *gh) gfs2_assert_warn(sdp, state != gl->gl_state); gfs2_glock_hold(gl); - gl->gl_req_bh = xmote_bh; lck_ret = gfs2_lm_lock(sdp, gl->gl_lock, gl->gl_state, state, lck_flags); @@ -877,51 +922,6 @@ static void gfs2_glock_xmote_th(struct gfs2_glock *gl, struct gfs2_holder *gh) xmote_bh(gl, lck_ret); } -/** - * drop_bh - Called after a lock module unlock completes - * @gl: the glock - * @ret: the return status - * - * Doesn't wake up the process waiting on the struct gfs2_holder (if any) - * Doesn't drop the reference on the glock the top half took out - * - */ - -static void drop_bh(struct gfs2_glock *gl, unsigned int ret) -{ - struct gfs2_sbd *sdp = gl->gl_sbd; - const struct gfs2_glock_operations *glops = gl->gl_ops; - struct gfs2_holder *gh = gl->gl_req_gh; - - gfs2_assert_warn(sdp, test_bit(GLF_LOCK, &gl->gl_flags)); - gfs2_assert_warn(sdp, list_empty(&gl->gl_holders)); - gfs2_assert_warn(sdp, !ret); - - state_change(gl, LM_ST_UNLOCKED); - - if (glops->go_inval) - glops->go_inval(gl, DIO_METADATA); - - if (gh) { - spin_lock(&gl->gl_spin); - list_del_init(&gh->gh_list); - gh->gh_error = 0; - spin_unlock(&gl->gl_spin); - } - - spin_lock(&gl->gl_spin); - gfs2_demote_wake(gl); - gl->gl_req_gh = NULL; - gl->gl_req_bh = NULL; - clear_bit(GLF_LOCK, &gl->gl_flags); - spin_unlock(&gl->gl_spin); - - gfs2_glock_put(gl); - - if (gh) - gfs2_holder_wake(gh); -} - /** * gfs2_glock_drop_th - call into the lock module to unlock a lock * @gl: the glock @@ -942,7 +942,6 @@ static void gfs2_glock_drop_th(struct gfs2_glock *gl) gfs2_assert_warn(sdp, gl->gl_state != LM_ST_UNLOCKED); gfs2_glock_hold(gl); - gl->gl_req_bh = drop_bh; ret = gfs2_lm_unlock(sdp, gl->gl_lock, gl->gl_state); @@ -971,8 +970,7 @@ static void do_cancels(struct gfs2_holder *gh) while (gl->gl_req_gh != gh && !test_bit(HIF_HOLDER, &gh->gh_iflags) && !list_empty(&gh->gh_list)) { - if (gl->gl_req_bh && !(gl->gl_req_gh && - (gl->gl_req_gh->gh_flags & GL_NOCANCEL))) { + if (!(gl->gl_req_gh && (gl->gl_req_gh->gh_flags & GL_NOCANCEL))) { spin_unlock(&gl->gl_spin); gfs2_lm_cancel(gl->gl_sbd, gl->gl_lock); msleep(100); @@ -1042,7 +1040,6 @@ static int glock_wait_internal(struct gfs2_holder *gh) spin_lock(&gl->gl_spin); gl->gl_req_gh = NULL; - gl->gl_req_bh = NULL; clear_bit(GLF_LOCK, &gl->gl_flags); run_queue(gl); spin_unlock(&gl->gl_spin); @@ -1535,8 +1532,7 @@ void gfs2_glock_cb(void *cb_data, unsigned int type, void *data) gl = gfs2_glock_find(sdp, &async->lc_name); if (gfs2_assert_warn(sdp, gl)) return; - if (!gfs2_assert_warn(sdp, gl->gl_req_bh)) - gl->gl_req_bh(gl, async->lc_ret); + xmote_bh(gl, async->lc_ret); if (queue_delayed_work(glock_workqueue, &gl->gl_work, 0) == 0) gfs2_glock_put(gl); up_read(&gfs2_umount_flush_sem); @@ -1898,7 +1894,6 @@ static int dump_glock(struct glock_iter *gi, struct gfs2_glock *gl) print_dbg(gi, " gl_owner = -1\n"); print_dbg(gi, " gl_ip = %lu\n", gl->gl_ip); print_dbg(gi, " req_gh = %s\n", (gl->gl_req_gh) ? "yes" : "no"); - print_dbg(gi, " req_bh = %s\n", (gl->gl_req_bh) ? "yes" : "no"); print_dbg(gi, " lvb_count = %d\n", atomic_read(&gl->gl_lvb_count)); print_dbg(gi, " object = %s\n", (gl->gl_object) ? "yes" : "no"); print_dbg(gi, " reclaim = %s\n", diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index fe14f6a417b..65aa46acb08 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -191,7 +191,6 @@ struct gfs2_glock { const struct gfs2_glock_operations *gl_ops; struct gfs2_holder *gl_req_gh; - gfs2_glop_bh_t gl_req_bh; void *gl_lock; char *gl_lvb; -- cgit v1.2.3 From da755fdb414470d6dce3df12ad188de9131cf96c Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Wed, 30 Jan 2008 15:34:04 +0000 Subject: [GFS2] Remove lm.[ch] and distribute content The functions in lm.c were just wrappers which were mostly only used in one other file. By moving the functions to the files where they are being used, they can be marked static and also this will usually result in them being inlined since they are often only used from one point in the code. A couple of really trivial functions have been inlined by hand into the function which called them as it makes the code clearer to do that. We also gain from one fewer function call in the glock lock and unlock paths. Signed-off-by: Steven Whitehouse --- fs/gfs2/Makefile | 2 +- fs/gfs2/glock.c | 51 ++++++++++++- fs/gfs2/incore.h | 2 + fs/gfs2/lm.c | 210 --------------------------------------------------- fs/gfs2/lm.h | 42 ----------- fs/gfs2/ops_file.c | 31 +++++++- fs/gfs2/ops_fstype.c | 71 ++++++++++++++++- fs/gfs2/ops_super.c | 1 - fs/gfs2/recovery.c | 11 ++- fs/gfs2/super.h | 1 + fs/gfs2/sys.c | 1 - fs/gfs2/util.c | 23 +++++- fs/gfs2/util.h | 1 + 13 files changed, 184 insertions(+), 263 deletions(-) delete mode 100644 fs/gfs2/lm.c delete mode 100644 fs/gfs2/lm.h (limited to 'fs') diff --git a/fs/gfs2/Makefile b/fs/gfs2/Makefile index 8fff11058ce..e2350df02a0 100644 --- a/fs/gfs2/Makefile +++ b/fs/gfs2/Makefile @@ -1,6 +1,6 @@ obj-$(CONFIG_GFS2_FS) += gfs2.o gfs2-y := acl.o bmap.o daemon.o dir.o eaops.o eattr.o glock.o \ - glops.o inode.o lm.o log.o lops.o locking.o main.o meta_io.o \ + glops.o inode.o log.o lops.o locking.o main.o meta_io.o \ mount.o ops_address.o ops_dentry.o ops_export.o ops_file.o \ ops_fstype.o ops_inode.o ops_super.o quota.o \ recovery.o rgrp.o super.o sys.o trans.o util.o diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index a8387e0b506..611f84d2257 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -35,7 +35,6 @@ #include "glock.h" #include "glops.h" #include "inode.h" -#include "lm.h" #include "lops.h" #include "meta_io.h" #include "quota.h" @@ -183,7 +182,8 @@ static void glock_free(struct gfs2_glock *gl) struct gfs2_sbd *sdp = gl->gl_sbd; struct inode *aspace = gl->gl_aspace; - gfs2_lm_put_lock(sdp, gl->gl_lock); + if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) + sdp->sd_lockstruct.ls_ops->lm_put_lock(gl->gl_lock); if (aspace) gfs2_aspace_put(aspace); @@ -293,6 +293,16 @@ static void glock_work_func(struct work_struct *work) gfs2_glock_put(gl); } +static int gfs2_lm_get_lock(struct gfs2_sbd *sdp, struct lm_lockname *name, + void **lockp) +{ + int error = -EIO; + if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) + error = sdp->sd_lockstruct.ls_ops->lm_get_lock( + sdp->sd_lockstruct.ls_lockspace, name, lockp); + return error; +} + /** * gfs2_glock_get() - Get a glock, or create one if one doesn't exist * @sdp: The GFS2 superblock @@ -882,6 +892,17 @@ out: gfs2_holder_wake(gh); } +static unsigned int gfs2_lm_lock(struct gfs2_sbd *sdp, void *lock, + unsigned int cur_state, unsigned int req_state, + unsigned int flags) +{ + int ret = 0; + if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) + ret = sdp->sd_lockstruct.ls_ops->lm_lock(lock, cur_state, + req_state, flags); + return ret; +} + /** * gfs2_glock_xmote_th - Call into the lock module to acquire or change a glock * @gl: The glock in question @@ -922,6 +943,15 @@ static void gfs2_glock_xmote_th(struct gfs2_glock *gl, struct gfs2_holder *gh) xmote_bh(gl, lck_ret); } +static unsigned int gfs2_lm_unlock(struct gfs2_sbd *sdp, void *lock, + unsigned int cur_state) +{ + int ret = 0; + if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) + ret = sdp->sd_lockstruct.ls_ops->lm_unlock(lock, cur_state); + return ret; +} + /** * gfs2_glock_drop_th - call into the lock module to unlock a lock * @gl: the glock @@ -964,6 +994,7 @@ static void gfs2_glock_drop_th(struct gfs2_glock *gl) static void do_cancels(struct gfs2_holder *gh) { struct gfs2_glock *gl = gh->gh_gl; + struct gfs2_sbd *sdp = gl->gl_sbd; spin_lock(&gl->gl_spin); @@ -972,7 +1003,8 @@ static void do_cancels(struct gfs2_holder *gh) !list_empty(&gh->gh_list)) { if (!(gl->gl_req_gh && (gl->gl_req_gh->gh_flags & GL_NOCANCEL))) { spin_unlock(&gl->gl_spin); - gfs2_lm_cancel(gl->gl_sbd, gl->gl_lock); + if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) + sdp->sd_lockstruct.ls_ops->lm_cancel(gl->gl_lock); msleep(100); spin_lock(&gl->gl_spin); } else { @@ -1426,6 +1458,14 @@ void gfs2_glock_dq_uninit_m(unsigned int num_gh, struct gfs2_holder *ghs) gfs2_glock_dq_uninit(&ghs[x]); } +static int gfs2_lm_hold_lvb(struct gfs2_sbd *sdp, void *lock, char **lvbp) +{ + int error = -EIO; + if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) + error = sdp->sd_lockstruct.ls_ops->lm_hold_lvb(lock, lvbp); + return error; +} + /** * gfs2_lvb_hold - attach a LVB from a glock * @gl: The glock in question @@ -1461,12 +1501,15 @@ int gfs2_lvb_hold(struct gfs2_glock *gl) void gfs2_lvb_unhold(struct gfs2_glock *gl) { + struct gfs2_sbd *sdp = gl->gl_sbd; + gfs2_glock_hold(gl); gfs2_glmutex_lock(gl); gfs2_assert(gl->gl_sbd, atomic_read(&gl->gl_lvb_count) > 0); if (atomic_dec_and_test(&gl->gl_lvb_count)) { - gfs2_lm_unhold_lvb(gl->gl_sbd, gl->gl_lock, gl->gl_lvb); + if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) + sdp->sd_lockstruct.ls_ops->lm_unhold_lvb(gl->gl_lock, gl->gl_lvb); gl->gl_lvb = NULL; gfs2_glock_put(gl); } diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index 65aa46acb08..8dee4672c3d 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -211,6 +211,8 @@ struct gfs2_glock { struct delayed_work gl_work; }; +#define GFS2_MIN_LVB_SIZE 32 /* Min size of LVB that gfs2 supports */ + struct gfs2_alloc { /* Quota stuff */ diff --git a/fs/gfs2/lm.c b/fs/gfs2/lm.c deleted file mode 100644 index cfcc39b86a5..00000000000 --- a/fs/gfs2/lm.c +++ /dev/null @@ -1,210 +0,0 @@ -/* - * Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved. - * Copyright (C) 2004-2006 Red Hat, Inc. All rights reserved. - * - * This copyrighted material is made available to anyone wishing to use, - * modify, copy, or redistribute it subject to the terms and conditions - * of the GNU General Public License version 2. - */ - -#include -#include -#include -#include -#include -#include -#include - -#include "gfs2.h" -#include "incore.h" -#include "glock.h" -#include "lm.h" -#include "super.h" -#include "util.h" - -/** - * gfs2_lm_mount - mount a locking protocol - * @sdp: the filesystem - * @args: mount arguements - * @silent: if 1, don't complain if the FS isn't a GFS2 fs - * - * Returns: errno - */ - -int gfs2_lm_mount(struct gfs2_sbd *sdp, int silent) -{ - char *proto = sdp->sd_proto_name; - char *table = sdp->sd_table_name; - int flags = 0; - int error; - - if (sdp->sd_args.ar_spectator) - flags |= LM_MFLAG_SPECTATOR; - - fs_info(sdp, "Trying to join cluster \"%s\", \"%s\"\n", proto, table); - - error = gfs2_mount_lockproto(proto, table, sdp->sd_args.ar_hostdata, - gfs2_glock_cb, sdp, - GFS2_MIN_LVB_SIZE, flags, - &sdp->sd_lockstruct, &sdp->sd_kobj); - if (error) { - fs_info(sdp, "can't mount proto=%s, table=%s, hostdata=%s\n", - proto, table, sdp->sd_args.ar_hostdata); - goto out; - } - - if (gfs2_assert_warn(sdp, sdp->sd_lockstruct.ls_lockspace) || - gfs2_assert_warn(sdp, sdp->sd_lockstruct.ls_ops) || - gfs2_assert_warn(sdp, sdp->sd_lockstruct.ls_lvb_size >= - GFS2_MIN_LVB_SIZE)) { - gfs2_unmount_lockproto(&sdp->sd_lockstruct); - goto out; - } - - if (sdp->sd_args.ar_spectator) - snprintf(sdp->sd_fsname, GFS2_FSNAME_LEN, "%s.s", table); - else - snprintf(sdp->sd_fsname, GFS2_FSNAME_LEN, "%s.%u", table, - sdp->sd_lockstruct.ls_jid); - - fs_info(sdp, "Joined cluster. Now mounting FS...\n"); - - if ((sdp->sd_lockstruct.ls_flags & LM_LSFLAG_LOCAL) && - !sdp->sd_args.ar_ignore_local_fs) { - sdp->sd_args.ar_localflocks = 1; - sdp->sd_args.ar_localcaching = 1; - } - -out: - return error; -} - -void gfs2_lm_others_may_mount(struct gfs2_sbd *sdp) -{ - if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) - sdp->sd_lockstruct.ls_ops->lm_others_may_mount( - sdp->sd_lockstruct.ls_lockspace); -} - -void gfs2_lm_unmount(struct gfs2_sbd *sdp) -{ - if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) - gfs2_unmount_lockproto(&sdp->sd_lockstruct); -} - -int gfs2_lm_withdraw(struct gfs2_sbd *sdp, char *fmt, ...) -{ - va_list args; - - if (test_and_set_bit(SDF_SHUTDOWN, &sdp->sd_flags)) - return 0; - - va_start(args, fmt); - vprintk(fmt, args); - va_end(args); - - fs_err(sdp, "about to withdraw this file system\n"); - BUG_ON(sdp->sd_args.ar_debug); - - fs_err(sdp, "telling LM to withdraw\n"); - gfs2_withdraw_lockproto(&sdp->sd_lockstruct); - fs_err(sdp, "withdrawn\n"); - dump_stack(); - - return -1; -} - -int gfs2_lm_get_lock(struct gfs2_sbd *sdp, struct lm_lockname *name, - void **lockp) -{ - int error = -EIO; - if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) - error = sdp->sd_lockstruct.ls_ops->lm_get_lock( - sdp->sd_lockstruct.ls_lockspace, name, lockp); - return error; -} - -void gfs2_lm_put_lock(struct gfs2_sbd *sdp, void *lock) -{ - if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) - sdp->sd_lockstruct.ls_ops->lm_put_lock(lock); -} - -unsigned int gfs2_lm_lock(struct gfs2_sbd *sdp, void *lock, - unsigned int cur_state, unsigned int req_state, - unsigned int flags) -{ - int ret = 0; - if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) - ret = sdp->sd_lockstruct.ls_ops->lm_lock(lock, cur_state, - req_state, flags); - return ret; -} - -unsigned int gfs2_lm_unlock(struct gfs2_sbd *sdp, void *lock, - unsigned int cur_state) -{ - int ret = 0; - if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) - ret = sdp->sd_lockstruct.ls_ops->lm_unlock(lock, cur_state); - return ret; -} - -void gfs2_lm_cancel(struct gfs2_sbd *sdp, void *lock) -{ - if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) - sdp->sd_lockstruct.ls_ops->lm_cancel(lock); -} - -int gfs2_lm_hold_lvb(struct gfs2_sbd *sdp, void *lock, char **lvbp) -{ - int error = -EIO; - if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) - error = sdp->sd_lockstruct.ls_ops->lm_hold_lvb(lock, lvbp); - return error; -} - -void gfs2_lm_unhold_lvb(struct gfs2_sbd *sdp, void *lock, char *lvb) -{ - if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) - sdp->sd_lockstruct.ls_ops->lm_unhold_lvb(lock, lvb); -} - -int gfs2_lm_plock_get(struct gfs2_sbd *sdp, struct lm_lockname *name, - struct file *file, struct file_lock *fl) -{ - int error = -EIO; - if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) - error = sdp->sd_lockstruct.ls_ops->lm_plock_get( - sdp->sd_lockstruct.ls_lockspace, name, file, fl); - return error; -} - -int gfs2_lm_plock(struct gfs2_sbd *sdp, struct lm_lockname *name, - struct file *file, int cmd, struct file_lock *fl) -{ - int error = -EIO; - if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) - error = sdp->sd_lockstruct.ls_ops->lm_plock( - sdp->sd_lockstruct.ls_lockspace, name, file, cmd, fl); - return error; -} - -int gfs2_lm_punlock(struct gfs2_sbd *sdp, struct lm_lockname *name, - struct file *file, struct file_lock *fl) -{ - int error = -EIO; - if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) - error = sdp->sd_lockstruct.ls_ops->lm_punlock( - sdp->sd_lockstruct.ls_lockspace, name, file, fl); - return error; -} - -void gfs2_lm_recovery_done(struct gfs2_sbd *sdp, unsigned int jid, - unsigned int message) -{ - if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) - sdp->sd_lockstruct.ls_ops->lm_recovery_done( - sdp->sd_lockstruct.ls_lockspace, jid, message); -} - diff --git a/fs/gfs2/lm.h b/fs/gfs2/lm.h deleted file mode 100644 index 21cdc30ee08..00000000000 --- a/fs/gfs2/lm.h +++ /dev/null @@ -1,42 +0,0 @@ -/* - * Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved. - * Copyright (C) 2004-2006 Red Hat, Inc. All rights reserved. - * - * This copyrighted material is made available to anyone wishing to use, - * modify, copy, or redistribute it subject to the terms and conditions - * of the GNU General Public License version 2. - */ - -#ifndef __LM_DOT_H__ -#define __LM_DOT_H__ - -struct gfs2_sbd; - -#define GFS2_MIN_LVB_SIZE 32 - -int gfs2_lm_mount(struct gfs2_sbd *sdp, int silent); -void gfs2_lm_others_may_mount(struct gfs2_sbd *sdp); -void gfs2_lm_unmount(struct gfs2_sbd *sdp); -int gfs2_lm_withdraw(struct gfs2_sbd *sdp, char *fmt, ...) - __attribute__ ((format(printf, 2, 3))); -int gfs2_lm_get_lock(struct gfs2_sbd *sdp, struct lm_lockname *name, - void **lockp); -void gfs2_lm_put_lock(struct gfs2_sbd *sdp, void *lock); -unsigned int gfs2_lm_lock(struct gfs2_sbd *sdp, void *lock, - unsigned int cur_state, unsigned int req_state, - unsigned int flags); -unsigned int gfs2_lm_unlock(struct gfs2_sbd *sdp, void *lock, - unsigned int cur_state); -void gfs2_lm_cancel(struct gfs2_sbd *sdp, void *lock); -int gfs2_lm_hold_lvb(struct gfs2_sbd *sdp, void *lock, char **lvbp); -void gfs2_lm_unhold_lvb(struct gfs2_sbd *sdp, void *lock, char *lvb); -int gfs2_lm_plock_get(struct gfs2_sbd *sdp, struct lm_lockname *name, - struct file *file, struct file_lock *fl); -int gfs2_lm_plock(struct gfs2_sbd *sdp, struct lm_lockname *name, - struct file *file, int cmd, struct file_lock *fl); -int gfs2_lm_punlock(struct gfs2_sbd *sdp, struct lm_lockname *name, - struct file *file, struct file_lock *fl); -void gfs2_lm_recovery_done(struct gfs2_sbd *sdp, unsigned int jid, - unsigned int message); - -#endif /* __LM_DOT_H__ */ diff --git a/fs/gfs2/ops_file.c b/fs/gfs2/ops_file.c index f4842f2548c..f97a8b86c48 100644 --- a/fs/gfs2/ops_file.c +++ b/fs/gfs2/ops_file.c @@ -30,7 +30,6 @@ #include "glock.h" #include "glops.h" #include "inode.h" -#include "lm.h" #include "log.h" #include "meta_io.h" #include "quota.h" @@ -596,6 +595,36 @@ static int gfs2_setlease(struct file *file, long arg, struct file_lock **fl) return generic_setlease(file, arg, fl); } +static int gfs2_lm_plock_get(struct gfs2_sbd *sdp, struct lm_lockname *name, + struct file *file, struct file_lock *fl) +{ + int error = -EIO; + if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) + error = sdp->sd_lockstruct.ls_ops->lm_plock_get( + sdp->sd_lockstruct.ls_lockspace, name, file, fl); + return error; +} + +static int gfs2_lm_plock(struct gfs2_sbd *sdp, struct lm_lockname *name, + struct file *file, int cmd, struct file_lock *fl) +{ + int error = -EIO; + if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) + error = sdp->sd_lockstruct.ls_ops->lm_plock( + sdp->sd_lockstruct.ls_lockspace, name, file, cmd, fl); + return error; +} + +static int gfs2_lm_punlock(struct gfs2_sbd *sdp, struct lm_lockname *name, + struct file *file, struct file_lock *fl) +{ + int error = -EIO; + if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) + error = sdp->sd_lockstruct.ls_ops->lm_punlock( + sdp->sd_lockstruct.ls_lockspace, name, file, fl); + return error; +} + /** * gfs2_lock - acquire/release a posix lock on a file * @file: the file pointer diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c index 4bee6aa845e..5b6a3451716 100644 --- a/fs/gfs2/ops_fstype.c +++ b/fs/gfs2/ops_fstype.c @@ -26,7 +26,6 @@ #include "glock.h" #include "glops.h" #include "inode.h" -#include "lm.h" #include "mount.h" #include "ops_fstype.h" #include "ops_dentry.h" @@ -363,6 +362,13 @@ static int map_journal_extents(struct gfs2_sbd *sdp) return rc; } +static void gfs2_lm_others_may_mount(struct gfs2_sbd *sdp) +{ + if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) + sdp->sd_lockstruct.ls_ops->lm_others_may_mount( + sdp->sd_lockstruct.ls_lockspace); +} + static int init_journal(struct gfs2_sbd *sdp, int undo) { struct gfs2_holder ji_gh; @@ -704,6 +710,69 @@ fail: return error; } +/** + * gfs2_lm_mount - mount a locking protocol + * @sdp: the filesystem + * @args: mount arguements + * @silent: if 1, don't complain if the FS isn't a GFS2 fs + * + * Returns: errno + */ + +static int gfs2_lm_mount(struct gfs2_sbd *sdp, int silent) +{ + char *proto = sdp->sd_proto_name; + char *table = sdp->sd_table_name; + int flags = 0; + int error; + + if (sdp->sd_args.ar_spectator) + flags |= LM_MFLAG_SPECTATOR; + + fs_info(sdp, "Trying to join cluster \"%s\", \"%s\"\n", proto, table); + + error = gfs2_mount_lockproto(proto, table, sdp->sd_args.ar_hostdata, + gfs2_glock_cb, sdp, + GFS2_MIN_LVB_SIZE, flags, + &sdp->sd_lockstruct, &sdp->sd_kobj); + if (error) { + fs_info(sdp, "can't mount proto=%s, table=%s, hostdata=%s\n", + proto, table, sdp->sd_args.ar_hostdata); + goto out; + } + + if (gfs2_assert_warn(sdp, sdp->sd_lockstruct.ls_lockspace) || + gfs2_assert_warn(sdp, sdp->sd_lockstruct.ls_ops) || + gfs2_assert_warn(sdp, sdp->sd_lockstruct.ls_lvb_size >= + GFS2_MIN_LVB_SIZE)) { + gfs2_unmount_lockproto(&sdp->sd_lockstruct); + goto out; + } + + if (sdp->sd_args.ar_spectator) + snprintf(sdp->sd_fsname, GFS2_FSNAME_LEN, "%s.s", table); + else + snprintf(sdp->sd_fsname, GFS2_FSNAME_LEN, "%s.%u", table, + sdp->sd_lockstruct.ls_jid); + + fs_info(sdp, "Joined cluster. Now mounting FS...\n"); + + if ((sdp->sd_lockstruct.ls_flags & LM_LSFLAG_LOCAL) && + !sdp->sd_args.ar_ignore_local_fs) { + sdp->sd_args.ar_localflocks = 1; + sdp->sd_args.ar_localcaching = 1; + } + +out: + return error; +} + +void gfs2_lm_unmount(struct gfs2_sbd *sdp) +{ + if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) + gfs2_unmount_lockproto(&sdp->sd_lockstruct); +} + /** * fill_super - Read in superblock * @sb: The VFS superblock diff --git a/fs/gfs2/ops_super.c b/fs/gfs2/ops_super.c index 5e524217944..2278c68b7e3 100644 --- a/fs/gfs2/ops_super.c +++ b/fs/gfs2/ops_super.c @@ -25,7 +25,6 @@ #include "incore.h" #include "glock.h" #include "inode.h" -#include "lm.h" #include "log.h" #include "mount.h" #include "ops_super.h" diff --git a/fs/gfs2/recovery.c b/fs/gfs2/recovery.c index 6fb07d67ca8..b17d3b8b232 100644 --- a/fs/gfs2/recovery.c +++ b/fs/gfs2/recovery.c @@ -20,7 +20,6 @@ #include "bmap.h" #include "glock.h" #include "glops.h" -#include "lm.h" #include "lops.h" #include "meta_io.h" #include "recovery.h" @@ -425,6 +424,16 @@ static int clean_journal(struct gfs2_jdesc *jd, struct gfs2_log_header_host *hea return error; } + +static void gfs2_lm_recovery_done(struct gfs2_sbd *sdp, unsigned int jid, + unsigned int message) +{ + if (likely(!test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) + sdp->sd_lockstruct.ls_ops->lm_recovery_done( + sdp->sd_lockstruct.ls_lockspace, jid, message); +} + + /** * gfs2_recover_journal - recovery a given journal * @jd: the struct gfs2_jdesc describing the journal diff --git a/fs/gfs2/super.h b/fs/gfs2/super.h index 60a870e430b..44361ecc44f 100644 --- a/fs/gfs2/super.h +++ b/fs/gfs2/super.h @@ -17,6 +17,7 @@ void gfs2_tune_init(struct gfs2_tune *gt); int gfs2_check_sb(struct gfs2_sbd *sdp, struct gfs2_sb_host *sb, int silent); int gfs2_read_sb(struct gfs2_sbd *sdp, struct gfs2_glock *gl, int silent); int gfs2_read_super(struct gfs2_sbd *sdp, sector_t sector); +void gfs2_lm_unmount(struct gfs2_sbd *sdp); static inline unsigned int gfs2_jindex_size(struct gfs2_sbd *sdp) { diff --git a/fs/gfs2/sys.c b/fs/gfs2/sys.c index eaa3b7b2f99..cc35ec862ee 100644 --- a/fs/gfs2/sys.c +++ b/fs/gfs2/sys.c @@ -20,7 +20,6 @@ #include "gfs2.h" #include "incore.h" -#include "lm.h" #include "sys.h" #include "super.h" #include "glock.h" diff --git a/fs/gfs2/util.c b/fs/gfs2/util.c index fe9c28ef77b..d31e355c61f 100644 --- a/fs/gfs2/util.c +++ b/fs/gfs2/util.c @@ -19,7 +19,6 @@ #include "gfs2.h" #include "incore.h" #include "glock.h" -#include "lm.h" #include "util.h" struct kmem_cache *gfs2_glock_cachep __read_mostly; @@ -33,6 +32,28 @@ void gfs2_assert_i(struct gfs2_sbd *sdp) sdp->sd_fsname); } +int gfs2_lm_withdraw(struct gfs2_sbd *sdp, char *fmt, ...) +{ + va_list args; + + if (test_and_set_bit(SDF_SHUTDOWN, &sdp->sd_flags)) + return 0; + + va_start(args, fmt); + vprintk(fmt, args); + va_end(args); + + fs_err(sdp, "about to withdraw this file system\n"); + BUG_ON(sdp->sd_args.ar_debug); + + fs_err(sdp, "telling LM to withdraw\n"); + gfs2_withdraw_lockproto(&sdp->sd_lockstruct); + fs_err(sdp, "withdrawn\n"); + dump_stack(); + + return -1; +} + /** * gfs2_assert_withdraw_i - Cause the machine to withdraw if @assertion is false * Returns: -1 if this call withdrew the machine, diff --git a/fs/gfs2/util.h b/fs/gfs2/util.h index ac0c567ebc3..509c5d60bd8 100644 --- a/fs/gfs2/util.h +++ b/fs/gfs2/util.h @@ -164,6 +164,7 @@ gfs2_tune_get_i(&(sdp)->sd_tune, &(sdp)->sd_tune.field) void gfs2_icbit_munge(struct gfs2_sbd *sdp, unsigned char **bitmap, unsigned int bit, int new_value); +int gfs2_lm_withdraw(struct gfs2_sbd *sdp, char *fmt, ...); #endif /* __UTIL_DOT_H__ */ -- cgit v1.2.3 From cf45b752c9f23939e40d823b0600bf876e97b0e0 Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Thu, 31 Jan 2008 10:31:39 -0600 Subject: [GFS2] Remove rgrp and glock version numbers This patch further reduces GFS2's memory requirements by eliminating the 64-bit version number fields in lieu of a couple bits. Signed-off-by: Bob Peterson Signed-off-by: Steven Whitehouse --- fs/gfs2/glock.c | 3 +-- fs/gfs2/glops.c | 10 ++++++++-- fs/gfs2/incore.h | 5 ++--- fs/gfs2/ops_fstype.c | 4 ++-- fs/gfs2/rgrp.c | 14 +++++++------- 5 files changed, 20 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index 611f84d2257..d00dc37e3d5 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -1,6 +1,6 @@ /* * Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved. - * Copyright (C) 2004-2007 Red Hat, Inc. All rights reserved. + * Copyright (C) 2004-2008 Red Hat, Inc. All rights reserved. * * This copyrighted material is made available to anyone wishing to use, * modify, copy, or redistribute it subject to the terms and conditions @@ -348,7 +348,6 @@ int gfs2_glock_get(struct gfs2_sbd *sdp, u64 number, gl->gl_ip = 0; gl->gl_ops = glops; gl->gl_req_gh = NULL; - gl->gl_vn = 0; gl->gl_stamp = jiffies; gl->gl_tchange = jiffies; gl->gl_object = NULL; diff --git a/fs/gfs2/glops.c b/fs/gfs2/glops.c index c663b7a0f41..d31badadef8 100644 --- a/fs/gfs2/glops.c +++ b/fs/gfs2/glops.c @@ -1,6 +1,6 @@ /* * Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved. - * Copyright (C) 2004-2006 Red Hat, Inc. All rights reserved. + * Copyright (C) 2004-2008 Red Hat, Inc. All rights reserved. * * This copyrighted material is made available to anyone wishing to use, * modify, copy, or redistribute it subject to the terms and conditions @@ -126,7 +126,13 @@ static void meta_go_inval(struct gfs2_glock *gl, int flags) return; gfs2_meta_inval(gl); - gl->gl_vn++; + if (gl->gl_object == GFS2_I(gl->gl_sbd->sd_rindex)) + gl->gl_sbd->sd_rindex_uptodate = 0; + else if (gl->gl_ops == &gfs2_rgrp_glops && gl->gl_object) { + struct gfs2_rgrpd *rgd = (struct gfs2_rgrpd *)gl->gl_object; + + rgd->rd_flags &= ~GFS2_RDF_UPTODATE; + } } /** diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index 8dee4672c3d..e9c58dc7686 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -85,7 +85,6 @@ struct gfs2_rgrpd { u32 rd_data; /* num of data blocks in rgrp */ u32 rd_bitbytes; /* number of bytes in data bitmaps */ struct gfs2_rgrp_host rd_rg; - u64 rd_rg_vn; struct gfs2_bitmap *rd_bits; unsigned int rd_bh_count; struct mutex rd_mutex; @@ -97,6 +96,7 @@ struct gfs2_rgrpd { unsigned char rd_flags; #define GFS2_RDF_CHECK 0x01 /* Need to check for unlinked inodes */ #define GFS2_RDF_NOALLOC 0x02 /* rg prohibits allocation */ +#define GFS2_RDF_UPTODATE 0x04 /* rg is up to date */ }; enum gfs2_state_bits { @@ -196,7 +196,6 @@ struct gfs2_glock { char *gl_lvb; atomic_t gl_lvb_count; - u64 gl_vn; unsigned long gl_stamp; unsigned long gl_tchange; void *gl_object; @@ -533,7 +532,7 @@ struct gfs2_sbd { /* Resource group stuff */ - u64 sd_rindex_vn; + int sd_rindex_uptodate; spinlock_t sd_rindex_spin; struct mutex sd_rindex_mutex; struct list_head sd_rindex_list; diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c index 5b6a3451716..c4b7a210c0c 100644 --- a/fs/gfs2/ops_fstype.c +++ b/fs/gfs2/ops_fstype.c @@ -1,6 +1,6 @@ /* * Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved. - * Copyright (C) 2004-2007 Red Hat, Inc. All rights reserved. + * Copyright (C) 2004-2008 Red Hat, Inc. All rights reserved. * * This copyrighted material is made available to anyone wishing to use, * modify, copy, or redistribute it subject to the terms and conditions @@ -548,7 +548,7 @@ static int init_inodes(struct gfs2_sbd *sdp, int undo) } ip = GFS2_I(sdp->sd_rindex); set_bit(GLF_STICKY, &ip->i_gl->gl_flags); - sdp->sd_rindex_vn = ip->i_gl->gl_vn - 1; + sdp->sd_rindex_uptodate = 0; /* Read in the quota inode */ sdp->sd_quota_inode = gfs2_lookup_simple(sdp->sd_master_dir, "quota"); diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index 5fd87104e59..3f10b1fafd6 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -540,7 +540,7 @@ static int read_rindex_entry(struct gfs2_inode *ip, return error; rgd->rd_gl->gl_object = rgd; - rgd->rd_rg_vn = rgd->rd_gl->gl_vn - 1; + rgd->rd_flags &= ~GFS2_RDF_UPTODATE; rgd->rd_flags |= GFS2_RDF_CHECK; return error; } @@ -576,7 +576,7 @@ static int gfs2_ri_update(struct gfs2_inode *ip) } } - sdp->sd_rindex_vn = ip->i_gl->gl_vn; + sdp->sd_rindex_uptodate = 1; return 0; } @@ -610,7 +610,7 @@ static int gfs2_ri_update_special(struct gfs2_inode *ip) } } - sdp->sd_rindex_vn = ip->i_gl->gl_vn; + sdp->sd_rindex_uptodate = 1; return 0; } @@ -643,9 +643,9 @@ int gfs2_rindex_hold(struct gfs2_sbd *sdp, struct gfs2_holder *ri_gh) return error; /* Read new copy from disk if we don't have the latest */ - if (sdp->sd_rindex_vn != gl->gl_vn) { + if (!sdp->sd_rindex_uptodate) { mutex_lock(&sdp->sd_rindex_mutex); - if (sdp->sd_rindex_vn != gl->gl_vn) { + if (!sdp->sd_rindex_uptodate) { error = gfs2_ri_update(ip); if (error) gfs2_glock_dq_uninit(ri_gh); @@ -737,9 +737,9 @@ int gfs2_rgrp_bh_get(struct gfs2_rgrpd *rgd) } } - if (rgd->rd_rg_vn != gl->gl_vn) { + if (!(rgd->rd_flags & GFS2_RDF_UPTODATE)) { gfs2_rgrp_in(rgd, (rgd->rd_bits[0].bi_bh)->b_data); - rgd->rd_rg_vn = gl->gl_vn; + rgd->rd_flags |= GFS2_RDF_UPTODATE; } spin_lock(&sdp->sd_rindex_spin); -- cgit v1.2.3 From 9a0045088d888c9c539c8c626a366cb52c0fbdab Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Fri, 1 Feb 2008 09:23:44 +0000 Subject: [GFS2] Shrink & rename di_depth This patch forms a pair with the previous patch which shrunk di_height. Like that patch di_depth is renamed i_depth and moved into struct gfs2_inode directly. Also the field goes from 16 bits to 8 bits since it is also limited to a max value which is rather small (17 in this case). In addition we also now validate the field against this maximum value when its read in. Signed-off-by: Steven Whitehouse --- fs/gfs2/dir.c | 34 +++++++++++++++++----------------- fs/gfs2/incore.h | 2 +- fs/gfs2/inode.c | 11 +++++++---- 3 files changed, 25 insertions(+), 22 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/dir.c b/fs/gfs2/dir.c index 78c236fb34e..081daa96a9d 100644 --- a/fs/gfs2/dir.c +++ b/fs/gfs2/dir.c @@ -757,7 +757,7 @@ static struct gfs2_dirent *gfs2_dirent_search(struct inode *inode, if (ip->i_di.di_flags & GFS2_DIF_EXHASH) { struct gfs2_leaf *leaf; - unsigned hsize = 1 << ip->i_di.di_depth; + unsigned hsize = 1 << ip->i_depth; unsigned index; u64 ln; if (hsize * sizeof(u64) != ip->i_di.di_size) { @@ -765,7 +765,7 @@ static struct gfs2_dirent *gfs2_dirent_search(struct inode *inode, return ERR_PTR(-EIO); } - index = name->hash >> (32 - ip->i_di.di_depth); + index = name->hash >> (32 - ip->i_depth); error = get_first_leaf(ip, index, &bh); if (error) return ERR_PTR(error); @@ -910,7 +910,7 @@ static int dir_make_exhash(struct inode *inode) dip->i_di.di_flags |= GFS2_DIF_EXHASH; for (x = sdp->sd_hash_ptrs, y = -1; x; x >>= 1, y++) ; - dip->i_di.di_depth = y; + dip->i_depth = y; gfs2_dinode_out(dip, dibh->b_data); @@ -941,7 +941,7 @@ static int dir_split_leaf(struct inode *inode, const struct qstr *name) int x, moved = 0; int error; - index = name->hash >> (32 - dip->i_di.di_depth); + index = name->hash >> (32 - dip->i_depth); error = get_leaf_nr(dip, index, &leaf_no); if (error) return error; @@ -952,7 +952,7 @@ static int dir_split_leaf(struct inode *inode, const struct qstr *name) return error; oleaf = (struct gfs2_leaf *)obh->b_data; - if (dip->i_di.di_depth == be16_to_cpu(oleaf->lf_depth)) { + if (dip->i_depth == be16_to_cpu(oleaf->lf_depth)) { brelse(obh); return 1; /* can't split */ } @@ -967,10 +967,10 @@ static int dir_split_leaf(struct inode *inode, const struct qstr *name) bn = nbh->b_blocknr; /* Compute the start and len of leaf pointers in the hash table. */ - len = 1 << (dip->i_di.di_depth - be16_to_cpu(oleaf->lf_depth)); + len = 1 << (dip->i_depth - be16_to_cpu(oleaf->lf_depth)); half_len = len >> 1; if (!half_len) { - printk(KERN_WARNING "di_depth %u lf_depth %u index %u\n", dip->i_di.di_depth, be16_to_cpu(oleaf->lf_depth), index); + printk(KERN_WARNING "i_depth %u lf_depth %u index %u\n", dip->i_depth, be16_to_cpu(oleaf->lf_depth), index); gfs2_consist_inode(dip); error = -EIO; goto fail_brelse; @@ -997,7 +997,7 @@ static int dir_split_leaf(struct inode *inode, const struct qstr *name) kfree(lp); /* Compute the divider */ - divider = (start + half_len) << (32 - dip->i_di.di_depth); + divider = (start + half_len) << (32 - dip->i_depth); /* Copy the entries */ dirent_first(dip, obh, &dent); @@ -1082,7 +1082,7 @@ static int dir_double_exhash(struct gfs2_inode *dip) int x; int error = 0; - hsize = 1 << dip->i_di.di_depth; + hsize = 1 << dip->i_depth; if (hsize * sizeof(u64) != dip->i_di.di_size) { gfs2_consist_inode(dip); return -EIO; @@ -1125,7 +1125,7 @@ static int dir_double_exhash(struct gfs2_inode *dip) error = gfs2_meta_inode_buffer(dip, &dibh); if (!gfs2_assert_withdraw(sdp, !error)) { - dip->i_di.di_depth++; + dip->i_depth++; gfs2_dinode_out(dip, dibh->b_data); brelse(dibh); } @@ -1370,14 +1370,14 @@ static int dir_e_read(struct inode *inode, u64 *offset, void *opaque, int error = 0; unsigned depth = 0; - hsize = 1 << dip->i_di.di_depth; + hsize = 1 << dip->i_depth; if (hsize * sizeof(u64) != dip->i_di.di_size) { gfs2_consist_inode(dip); return -EIO; } hash = gfs2_dir_offset2hash(*offset); - index = hash >> (32 - dip->i_di.di_depth); + index = hash >> (32 - dip->i_depth); lp = kmalloc(sdp->sd_hash_bsize, GFP_KERNEL); if (!lp) @@ -1405,7 +1405,7 @@ static int dir_e_read(struct inode *inode, u64 *offset, void *opaque, if (error) break; - len = 1 << (dip->i_di.di_depth - depth); + len = 1 << (dip->i_depth - depth); index = (index & ~(len - 1)) + len; } @@ -1549,7 +1549,7 @@ static int dir_new_leaf(struct inode *inode, const struct qstr *name) u32 index; u64 bn; - index = name->hash >> (32 - ip->i_di.di_depth); + index = name->hash >> (32 - ip->i_depth); error = get_first_leaf(ip, index, &obh); if (error) return error; @@ -1641,7 +1641,7 @@ int gfs2_dir_add(struct inode *inode, const struct qstr *name, continue; if (error < 0) break; - if (ip->i_di.di_depth < GFS2_DIR_MAX_DEPTH) { + if (ip->i_depth < GFS2_DIR_MAX_DEPTH) { error = dir_double_exhash(ip); if (error) break; @@ -1785,7 +1785,7 @@ static int foreach_leaf(struct gfs2_inode *dip, leaf_call_t lc, void *data) u64 leaf_no; int error = 0; - hsize = 1 << dip->i_di.di_depth; + hsize = 1 << dip->i_depth; if (hsize * sizeof(u64) != dip->i_di.di_size) { gfs2_consist_inode(dip); return -EIO; @@ -1817,7 +1817,7 @@ static int foreach_leaf(struct gfs2_inode *dip, leaf_call_t lc, void *data) if (error) goto out; leaf = (struct gfs2_leaf *)bh->b_data; - len = 1 << (dip->i_di.di_depth - be16_to_cpu(leaf->lf_depth)); + len = 1 << (dip->i_depth - be16_to_cpu(leaf->lf_depth)); brelse(bh); error = lc(dip, index, len, leaf_no, data); diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index e9c58dc7686..9dfdde3612a 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -246,7 +246,6 @@ struct gfs2_dinode_host { u64 di_generation; /* generation number for NFS */ u32 di_flags; /* GFS2_DIF_... */ /* These only apply to directories */ - u16 di_depth; /* Number of bits in the table */ u32 di_entries; /* The number of entries in the directory */ u64 di_eattr; /* extended attribute block number */ }; @@ -267,6 +266,7 @@ struct gfs2_inode { struct rw_semaphore i_rw_mutex; u8 i_height; + u8 i_depth; }; /* diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c index db5961a9aa5..65fdfee9ca9 100644 --- a/fs/gfs2/inode.c +++ b/fs/gfs2/inode.c @@ -248,7 +248,7 @@ static int gfs2_dinode_in(struct gfs2_inode *ip, const void *buf) { struct gfs2_dinode_host *di = &ip->i_di; const struct gfs2_dinode *str = buf; - u16 height; + u16 height, depth; if (unlikely(ip->i_no_addr != be64_to_cpu(str->di_num.no_addr))) goto corrupt; @@ -293,7 +293,10 @@ static int gfs2_dinode_in(struct gfs2_inode *ip, const void *buf) goto corrupt; ip->i_height = (u8)height; - di->di_depth = be16_to_cpu(str->di_depth); + depth = be16_to_cpu(str->di_depth); + if (unlikely(depth > GFS2_DIR_MAX_DEPTH)) + goto corrupt; + ip->i_depth = (u8)depth; di->di_entries = be32_to_cpu(str->di_entries); di->di_eattr = be64_to_cpu(str->di_eattr); @@ -1410,7 +1413,7 @@ void gfs2_dinode_out(const struct gfs2_inode *ip, void *buf) str->di_payload_format = cpu_to_be32(S_ISDIR(ip->i_inode.i_mode) && !(ip->i_di.di_flags & GFS2_DIF_EXHASH) ? GFS2_FORMAT_DE : 0); - str->di_depth = cpu_to_be16(di->di_depth); + str->di_depth = cpu_to_be16(ip->i_depth); str->di_entries = cpu_to_be32(di->di_entries); str->di_eattr = cpu_to_be64(di->di_eattr); @@ -1436,7 +1439,7 @@ void gfs2_dinode_print(const struct gfs2_inode *ip) (unsigned long long)di->di_goal_data); printk(KERN_INFO " di_flags = 0x%.8X\n", di->di_flags); printk(KERN_INFO " i_height = %u\n", ip->i_height); - printk(KERN_INFO " di_depth = %u\n", di->di_depth); + printk(KERN_INFO " i_depth = %u\n", ip->i_depth); printk(KERN_INFO " di_entries = %u\n", di->di_entries); printk(KERN_INFO " di_eattr = %llu\n", (unsigned long long)di->di_eattr); -- cgit v1.2.3 From 9feb7c889f2a3b088a7f6583e609bd39997c0f47 Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Tue, 5 Feb 2008 17:11:40 -0600 Subject: [GFS2] Remove unused counters This is kind of trivial in the greater scheme of things, but this removes three counters that AFAICT are never used. Signed-off-by: Bob Peterson Signed-off-by: Steven Whitehouse --- fs/gfs2/incore.h | 3 --- fs/gfs2/sys.c | 6 ------ 2 files changed, 9 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index 9dfdde3612a..b67e44baa73 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -636,9 +636,6 @@ struct gfs2_sbd { /* Counters */ - atomic_t sd_glock_count; - atomic_t sd_glock_held_count; - atomic_t sd_inode_count; atomic_t sd_reclaimed; char sd_fsname[GFS2_FSNAME_LEN]; diff --git a/fs/gfs2/sys.c b/fs/gfs2/sys.c index cc35ec862ee..9ab9fc85ecd 100644 --- a/fs/gfs2/sys.c +++ b/fs/gfs2/sys.c @@ -327,15 +327,9 @@ static ssize_t name##_show(struct gfs2_sbd *sdp, char *buf) \ } \ static struct counters_attr counters_attr_##name = __ATTR_RO(name) -COUNTERS_ATTR(glock_count, "%u\n"); -COUNTERS_ATTR(glock_held_count, "%u\n"); -COUNTERS_ATTR(inode_count, "%u\n"); COUNTERS_ATTR(reclaimed, "%u\n"); static struct attribute *counters_attrs[] = { - &counters_attr_glock_count.attr, - &counters_attr_glock_held_count.attr, - &counters_attr_inode_count.attr, &counters_attr_reclaimed.attr, NULL, }; -- cgit v1.2.3 From ce276b06e8b81845926387e93f77bf81e14b5cc2 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Wed, 6 Feb 2008 09:25:45 +0000 Subject: [GFS2] Reduce inode size by merging fields There were three fields being used to keep track of the location of the most recently allocated block for each inode. These have been merged into a single field in order to better keep the data and metadata for an inode close on disk, and also to reduce the space required for storage. Signed-off-by: Steven Whitehouse --- fs/gfs2/bmap.c | 4 +--- fs/gfs2/incore.h | 5 +---- fs/gfs2/inode.c | 13 +++++-------- fs/gfs2/rgrp.c | 40 +++++++++++++++++----------------------- 4 files changed, 24 insertions(+), 38 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index 2011dd27f8d..30d718b3438 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -1020,9 +1020,7 @@ static int trunc_end(struct gfs2_inode *ip) if (!ip->i_di.di_size) { ip->i_height = 0; - ip->i_di.di_goal_meta = - ip->i_di.di_goal_data = - ip->i_no_addr; + ip->i_goal = ip->i_no_addr; gfs2_buffer_clear_tail(dibh, sizeof(struct gfs2_dinode)); } ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME; diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index b67e44baa73..c50dcdf7992 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -241,8 +241,6 @@ enum { struct gfs2_dinode_host { u64 di_size; /* number of bytes in file */ u64 di_blocks; /* number of blocks in file */ - u64 di_goal_meta; /* rgrp to alloc from next */ - u64 di_goal_data; /* data block goal */ u64 di_generation; /* generation number for NFS */ u32 di_flags; /* GFS2_DIF_... */ /* These only apply to directories */ @@ -262,8 +260,7 @@ struct gfs2_inode { struct gfs2_holder i_iopen_gh; struct gfs2_holder i_gh; /* for prepare/commit_write only */ struct gfs2_alloc *i_alloc; - u64 i_last_rg_alloc; - + u64 i_goal; /* goal block for allocations */ struct rw_semaphore i_rw_mutex; u8 i_height; u8 i_depth; diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c index 65fdfee9ca9..c3fe8aa03c4 100644 --- a/fs/gfs2/inode.c +++ b/fs/gfs2/inode.c @@ -282,8 +282,7 @@ static int gfs2_dinode_in(struct gfs2_inode *ip, const void *buf) ip->i_inode.i_ctime.tv_sec = be64_to_cpu(str->di_ctime); ip->i_inode.i_ctime.tv_nsec = be32_to_cpu(str->di_ctime_nsec); - di->di_goal_meta = be64_to_cpu(str->di_goal_meta); - di->di_goal_data = be64_to_cpu(str->di_goal_data); + ip->i_goal = be64_to_cpu(str->di_goal_meta); di->di_generation = be64_to_cpu(str->di_generation); di->di_flags = be32_to_cpu(str->di_flags); @@ -1404,8 +1403,8 @@ void gfs2_dinode_out(const struct gfs2_inode *ip, void *buf) str->di_mtime = cpu_to_be64(ip->i_inode.i_mtime.tv_sec); str->di_ctime = cpu_to_be64(ip->i_inode.i_ctime.tv_sec); - str->di_goal_meta = cpu_to_be64(di->di_goal_meta); - str->di_goal_data = cpu_to_be64(di->di_goal_data); + str->di_goal_meta = cpu_to_be64(ip->i_goal); + str->di_goal_data = cpu_to_be64(ip->i_goal); str->di_generation = cpu_to_be64(di->di_generation); str->di_flags = cpu_to_be32(di->di_flags); @@ -1433,10 +1432,8 @@ void gfs2_dinode_print(const struct gfs2_inode *ip) printk(KERN_INFO " di_size = %llu\n", (unsigned long long)di->di_size); printk(KERN_INFO " di_blocks = %llu\n", (unsigned long long)di->di_blocks); - printk(KERN_INFO " di_goal_meta = %llu\n", - (unsigned long long)di->di_goal_meta); - printk(KERN_INFO " di_goal_data = %llu\n", - (unsigned long long)di->di_goal_data); + printk(KERN_INFO " i_goal = %llu\n", + (unsigned long long)ip->i_goal); printk(KERN_INFO " di_flags = 0x%.8X\n", di->di_flags); printk(KERN_INFO " i_height = %u\n", ip->i_height); printk(KERN_INFO " i_depth = %u\n", ip->i_depth); diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index 3f10b1fafd6..66193b45e50 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -915,24 +915,20 @@ static struct inode *try_rgrp_unlink(struct gfs2_rgrpd *rgd, u64 *last_unlinked) static struct gfs2_rgrpd *recent_rgrp_first(struct gfs2_sbd *sdp, u64 rglast) { - struct gfs2_rgrpd *rgd = NULL; + struct gfs2_rgrpd *rgd; spin_lock(&sdp->sd_rindex_spin); - if (list_empty(&sdp->sd_rindex_recent_list)) - goto out; - - if (!rglast) - goto first; - - list_for_each_entry(rgd, &sdp->sd_rindex_recent_list, rd_recent) { - if (rgd->rd_addr == rglast) - goto out; + if (rglast) { + list_for_each_entry(rgd, &sdp->sd_rindex_recent_list, rd_recent) { + if (rgrp_contains_block(rgd, rglast)) + goto out; + } } - -first: - rgd = list_entry(sdp->sd_rindex_recent_list.next, struct gfs2_rgrpd, - rd_recent); + rgd = NULL; + if (!list_empty(&sdp->sd_rindex_recent_list)) + rgd = list_entry(sdp->sd_rindex_recent_list.next, + struct gfs2_rgrpd, rd_recent); out: spin_unlock(&sdp->sd_rindex_spin); return rgd; @@ -1078,7 +1074,7 @@ static struct inode *get_local_rgrp(struct gfs2_inode *ip, u64 *last_unlinked) /* Try recently successful rgrps */ - rgd = recent_rgrp_first(sdp, ip->i_last_rg_alloc); + rgd = recent_rgrp_first(sdp, ip->i_goal); while (rgd) { rg_locked = 0; @@ -1162,8 +1158,6 @@ static struct inode *get_local_rgrp(struct gfs2_inode *ip, u64 *last_unlinked) } out: - ip->i_last_rg_alloc = rgd->rd_addr; - if (begin) { recent_rgrp_add(rgd); rgd = gfs2_rgrpd_get_next(rgd); @@ -1425,8 +1419,8 @@ u64 gfs2_alloc_data(struct gfs2_inode *ip) u32 goal, blk; u64 block; - if (rgrp_contains_block(rgd, ip->i_di.di_goal_data)) - goal = ip->i_di.di_goal_data - rgd->rd_data0; + if (rgrp_contains_block(rgd, ip->i_goal)) + goal = ip->i_goal - rgd->rd_data0; else goal = rgd->rd_last_alloc_data; @@ -1435,7 +1429,7 @@ u64 gfs2_alloc_data(struct gfs2_inode *ip) rgd->rd_last_alloc_data = blk; block = rgd->rd_data0 + blk; - ip->i_di.di_goal_data = block; + ip->i_goal = block; gfs2_assert_withdraw(sdp, rgd->rd_rg.rg_free); rgd->rd_rg.rg_free--; @@ -1470,8 +1464,8 @@ u64 gfs2_alloc_meta(struct gfs2_inode *ip) u32 goal, blk; u64 block; - if (rgrp_contains_block(rgd, ip->i_di.di_goal_meta)) - goal = ip->i_di.di_goal_meta - rgd->rd_data0; + if (rgrp_contains_block(rgd, ip->i_goal)) + goal = ip->i_goal - rgd->rd_data0; else goal = rgd->rd_last_alloc_meta; @@ -1480,7 +1474,7 @@ u64 gfs2_alloc_meta(struct gfs2_inode *ip) rgd->rd_last_alloc_meta = blk; block = rgd->rd_data0 + blk; - ip->i_di.di_goal_meta = block; + ip->i_goal = block; gfs2_assert_withdraw(sdp, rgd->rd_rg.rg_free); rgd->rd_rg.rg_free--; -- cgit v1.2.3 From ac576cc5bed0dd7759e2b196468c7df93d6aeeee Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Fri, 1 Feb 2008 10:34:15 +0000 Subject: [GFS2] Merge the rd_last_alloc_meta and rd_last_alloc_data fields We don't need to keep track of when we last allocated data and metadata separately since the only thing thats important when searching for a free block is whether its free or not, which is independent from what type of block it is. Signed-off-by: Steven Whitehouse --- fs/gfs2/incore.h | 3 +-- fs/gfs2/rgrp.c | 12 ++++++------ 2 files changed, 7 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index c50dcdf7992..898b456b386 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -90,8 +90,7 @@ struct gfs2_rgrpd { struct mutex rd_mutex; u32 rd_free_clone; struct gfs2_log_element rd_le; - u32 rd_last_alloc_data; - u32 rd_last_alloc_meta; + u32 rd_last_alloc; struct gfs2_sbd *rd_sbd; unsigned char rd_flags; #define GFS2_RDF_CHECK 0x01 /* Need to check for unlinked inodes */ diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index 66193b45e50..cc28845ba6f 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -1422,11 +1422,11 @@ u64 gfs2_alloc_data(struct gfs2_inode *ip) if (rgrp_contains_block(rgd, ip->i_goal)) goal = ip->i_goal - rgd->rd_data0; else - goal = rgd->rd_last_alloc_data; + goal = rgd->rd_last_alloc; blk = rgblk_search(rgd, goal, GFS2_BLKST_FREE, GFS2_BLKST_USED); BUG_ON(blk == BFITNOENT); - rgd->rd_last_alloc_data = blk; + rgd->rd_last_alloc = blk; block = rgd->rd_data0 + blk; ip->i_goal = block; @@ -1467,11 +1467,11 @@ u64 gfs2_alloc_meta(struct gfs2_inode *ip) if (rgrp_contains_block(rgd, ip->i_goal)) goal = ip->i_goal - rgd->rd_data0; else - goal = rgd->rd_last_alloc_meta; + goal = rgd->rd_last_alloc; blk = rgblk_search(rgd, goal, GFS2_BLKST_FREE, GFS2_BLKST_USED); BUG_ON(blk == BFITNOENT); - rgd->rd_last_alloc_meta = blk; + rgd->rd_last_alloc = blk; block = rgd->rd_data0 + blk; ip->i_goal = block; @@ -1510,11 +1510,11 @@ u64 gfs2_alloc_di(struct gfs2_inode *dip, u64 *generation) u32 blk; u64 block; - blk = rgblk_search(rgd, rgd->rd_last_alloc_meta, + blk = rgblk_search(rgd, rgd->rd_last_alloc, GFS2_BLKST_FREE, GFS2_BLKST_DINODE); BUG_ON(blk == BFITNOENT); - rgd->rd_last_alloc_meta = blk; + rgd->rd_last_alloc = blk; block = rgd->rd_data0 + blk; -- cgit v1.2.3 From 5731be53e3d82aedd06e02574f833a57b07a08d2 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Fri, 1 Feb 2008 13:16:55 +0000 Subject: [GFS2] Update gfs2_trans_add_unrevoke to accept extents By adding an extra argument to gfs2_trans_add_unrevoke we can now specify an extent length of blocks to unrevoke. This means that we only need to make one pass through the list for each extent rather than each block. Currently the only extent length which is used is 1, but that will change in the future. Also gfs2_trans_add_unrevoke is removed from gfs2_alloc_meta since its the only difference between this and gfs2_alloc_data which is left. This will allow a future patch to merge these two functions into one (i.e. one call to allocate both data and metadata in a single extent in the future). Signed-off-by: Steven Whitehouse --- fs/gfs2/bmap.c | 7 +++++-- fs/gfs2/dir.c | 2 +- fs/gfs2/eattr.c | 6 +++--- fs/gfs2/rgrp.c | 3 +-- fs/gfs2/trans.c | 25 ++++++++++--------------- fs/gfs2/trans.h | 2 +- 6 files changed, 21 insertions(+), 24 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index 30d718b3438..651e5320bb4 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -138,7 +138,7 @@ int gfs2_unstuff_dinode(struct gfs2_inode *ip, struct page *page) if (isdir) { block = gfs2_alloc_meta(ip); - + gfs2_trans_add_unrevoke(GFS2_SB(&ip->i_inode), block, 1); error = gfs2_dir_get_new_buffer(ip, block, &bh); if (error) goto out_brelse; @@ -206,6 +206,7 @@ static int build_height(struct inode *inode, struct metapath *mp, unsigned heigh for(n = 0; n < new_height; n++) { bn = gfs2_alloc_meta(ip); + gfs2_trans_add_unrevoke(GFS2_SB(inode), bn, 1); mp->mp_bh[n] = gfs2_meta_new(ip->i_gl, bn); gfs2_trans_add_bh(ip->i_gl, mp->mp_bh[n], 1); } @@ -370,8 +371,10 @@ static int lookup_block(struct gfs2_inode *ip, unsigned int height, if (height == ip->i_height - 1 && !gfs2_is_dir(ip)) *block = gfs2_alloc_data(ip); - else + else { *block = gfs2_alloc_meta(ip); + gfs2_trans_add_unrevoke(GFS2_SB(&ip->i_inode), *block, 1); + } gfs2_trans_add_bh(ip->i_gl, mp->mp_bh[height], 1); diff --git a/fs/gfs2/dir.c b/fs/gfs2/dir.c index 081daa96a9d..55514ee06dd 100644 --- a/fs/gfs2/dir.c +++ b/fs/gfs2/dir.c @@ -810,7 +810,7 @@ static struct gfs2_leaf *new_leaf(struct inode *inode, struct buffer_head **pbh, struct qstr name = { .name = "", .len = 0, .hash = 0 }; if (!bh) return NULL; - + gfs2_trans_add_unrevoke(GFS2_SB(inode), bn, 1); gfs2_trans_add_bh(ip->i_gl, bh, 1); gfs2_metatype_set(bh, GFS2_METATYPE_LF, GFS2_FORMAT_LF); leaf = (struct gfs2_leaf *)bh->b_data; diff --git a/fs/gfs2/eattr.c b/fs/gfs2/eattr.c index 04febbc17a1..c7fa0a8b164 100644 --- a/fs/gfs2/eattr.c +++ b/fs/gfs2/eattr.c @@ -585,7 +585,7 @@ static int ea_alloc_blk(struct gfs2_inode *ip, struct buffer_head **bhp) u64 block; block = gfs2_alloc_meta(ip); - + gfs2_trans_add_unrevoke(sdp, block, 1); *bhp = gfs2_meta_new(ip->i_gl, block); gfs2_trans_add_bh(ip->i_gl, *bhp, 1); gfs2_metatype_set(*bhp, GFS2_METATYPE_EA, GFS2_FORMAT_EA); @@ -644,7 +644,7 @@ static int ea_write(struct gfs2_inode *ip, struct gfs2_ea_header *ea, int mh_size = sizeof(struct gfs2_meta_header); block = gfs2_alloc_meta(ip); - + gfs2_trans_add_unrevoke(sdp, block, 1); bh = gfs2_meta_new(ip->i_gl, block); gfs2_trans_add_bh(ip->i_gl, bh, 1); gfs2_metatype_set(bh, GFS2_METATYPE_ED, GFS2_FORMAT_ED); @@ -968,7 +968,7 @@ static int ea_set_block(struct gfs2_inode *ip, struct gfs2_ea_request *er, u64 blk; blk = gfs2_alloc_meta(ip); - + gfs2_trans_add_unrevoke(sdp, blk, 1); indbh = gfs2_meta_new(ip->i_gl, blk); gfs2_trans_add_bh(ip->i_gl, indbh, 1); gfs2_metatype_set(indbh, GFS2_METATYPE_IN, GFS2_FORMAT_IN); diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index cc28845ba6f..9f28463e62e 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -1486,7 +1486,6 @@ u64 gfs2_alloc_meta(struct gfs2_inode *ip) gfs2_statfs_change(sdp, 0, -1, 0); gfs2_quota_change(ip, +1, ip->i_inode.i_uid, ip->i_inode.i_gid); - gfs2_trans_add_unrevoke(sdp, block); spin_lock(&sdp->sd_rindex_spin); rgd->rd_free_clone--; @@ -1528,7 +1527,7 @@ u64 gfs2_alloc_di(struct gfs2_inode *dip, u64 *generation) al->al_alloced++; gfs2_statfs_change(sdp, 0, -1, +1); - gfs2_trans_add_unrevoke(sdp, block); + gfs2_trans_add_unrevoke(sdp, block, 1); spin_lock(&sdp->sd_rindex_spin); rgd->rd_free_clone--; diff --git a/fs/gfs2/trans.c b/fs/gfs2/trans.c index 73e5d92a657..f677b8a83f0 100644 --- a/fs/gfs2/trans.c +++ b/fs/gfs2/trans.c @@ -146,30 +146,25 @@ void gfs2_trans_add_revoke(struct gfs2_sbd *sdp, struct gfs2_bufdata *bd) lops_add(sdp, &bd->bd_le); } -void gfs2_trans_add_unrevoke(struct gfs2_sbd *sdp, u64 blkno) +void gfs2_trans_add_unrevoke(struct gfs2_sbd *sdp, u64 blkno, unsigned int len) { - struct gfs2_bufdata *bd; - int found = 0; + struct gfs2_bufdata *bd, *tmp; + struct gfs2_trans *tr = current->journal_info; + unsigned int n = len; gfs2_log_lock(sdp); - - list_for_each_entry(bd, &sdp->sd_log_le_revoke, bd_le.le_list) { - if (bd->bd_blkno == blkno) { + list_for_each_entry_safe(bd, tmp, &sdp->sd_log_le_revoke, bd_le.le_list) { + if ((bd->bd_blkno >= blkno) && (bd->bd_blkno < (blkno + len))) { list_del_init(&bd->bd_le.le_list); gfs2_assert_withdraw(sdp, sdp->sd_log_num_revoke); sdp->sd_log_num_revoke--; - found = 1; - break; + kmem_cache_free(gfs2_bufdata_cachep, bd); + tr->tr_num_revoke_rm++; + if (--n == 0) + break; } } - gfs2_log_unlock(sdp); - - if (found) { - struct gfs2_trans *tr = current->journal_info; - kmem_cache_free(gfs2_bufdata_cachep, bd); - tr->tr_num_revoke_rm++; - } } void gfs2_trans_add_rg(struct gfs2_rgrpd *rgd) diff --git a/fs/gfs2/trans.h b/fs/gfs2/trans.h index e826f0dab80..edf9d4bd908 100644 --- a/fs/gfs2/trans.h +++ b/fs/gfs2/trans.h @@ -32,7 +32,7 @@ void gfs2_trans_end(struct gfs2_sbd *sdp); void gfs2_trans_add_bh(struct gfs2_glock *gl, struct buffer_head *bh, int meta); void gfs2_trans_add_revoke(struct gfs2_sbd *sdp, struct gfs2_bufdata *bd); -void gfs2_trans_add_unrevoke(struct gfs2_sbd *sdp, u64 blkno); +void gfs2_trans_add_unrevoke(struct gfs2_sbd *sdp, u64 blkno, unsigned int len); void gfs2_trans_add_rg(struct gfs2_rgrpd *rgd); #endif /* __TRANS_DOT_H__ */ -- cgit v1.2.3 From 1639431a3f57b43da1e15e9268a1d691ac01ba26 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Fri, 1 Feb 2008 14:52:30 +0000 Subject: [GFS2] Merge gfs2_alloc_meta and gfs2_alloc_data Thanks to the preceeding patches, the only difference between these two functions is their name. We can thus merge them and call the new function gfs2_alloc_block to reflect the fact that it can allocate either kind of block. Signed-off-by: Steven Whitehouse --- fs/gfs2/bmap.c | 13 +++++-------- fs/gfs2/dir.c | 2 +- fs/gfs2/eattr.c | 6 +++--- fs/gfs2/rgrp.c | 51 +++------------------------------------------------ fs/gfs2/rgrp.h | 3 +-- 5 files changed, 13 insertions(+), 62 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index 651e5320bb4..e3a75a27cee 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -137,7 +137,7 @@ int gfs2_unstuff_dinode(struct gfs2_inode *ip, struct page *page) and write it out to disk */ if (isdir) { - block = gfs2_alloc_meta(ip); + block = gfs2_alloc_block(ip); gfs2_trans_add_unrevoke(GFS2_SB(&ip->i_inode), block, 1); error = gfs2_dir_get_new_buffer(ip, block, &bh); if (error) @@ -146,7 +146,7 @@ int gfs2_unstuff_dinode(struct gfs2_inode *ip, struct page *page) dibh, sizeof(struct gfs2_dinode)); brelse(bh); } else { - block = gfs2_alloc_data(ip); + block = gfs2_alloc_block(ip); error = gfs2_unstuffer_page(ip, dibh, block, page); if (error) @@ -205,7 +205,7 @@ static int build_height(struct inode *inode, struct metapath *mp, unsigned heigh return error; for(n = 0; n < new_height; n++) { - bn = gfs2_alloc_meta(ip); + bn = gfs2_alloc_block(ip); gfs2_trans_add_unrevoke(GFS2_SB(inode), bn, 1); mp->mp_bh[n] = gfs2_meta_new(ip->i_gl, bn); gfs2_trans_add_bh(ip->i_gl, mp->mp_bh[n], 1); @@ -369,12 +369,9 @@ static int lookup_block(struct gfs2_inode *ip, unsigned int height, if (!create) return 0; - if (height == ip->i_height - 1 && !gfs2_is_dir(ip)) - *block = gfs2_alloc_data(ip); - else { - *block = gfs2_alloc_meta(ip); + *block = gfs2_alloc_block(ip); + if (height != ip->i_height - 1 || gfs2_is_dir(ip)) gfs2_trans_add_unrevoke(GFS2_SB(&ip->i_inode), *block, 1); - } gfs2_trans_add_bh(ip->i_gl, mp->mp_bh[height], 1); diff --git a/fs/gfs2/dir.c b/fs/gfs2/dir.c index 55514ee06dd..fbdf31957cb 100644 --- a/fs/gfs2/dir.c +++ b/fs/gfs2/dir.c @@ -803,7 +803,7 @@ got_dent: static struct gfs2_leaf *new_leaf(struct inode *inode, struct buffer_head **pbh, u16 depth) { struct gfs2_inode *ip = GFS2_I(inode); - u64 bn = gfs2_alloc_meta(ip); + u64 bn = gfs2_alloc_block(ip); struct buffer_head *bh = gfs2_meta_new(ip->i_gl, bn); struct gfs2_leaf *leaf; struct gfs2_dirent *dent; diff --git a/fs/gfs2/eattr.c b/fs/gfs2/eattr.c index c7fa0a8b164..f9f63bc21cd 100644 --- a/fs/gfs2/eattr.c +++ b/fs/gfs2/eattr.c @@ -584,7 +584,7 @@ static int ea_alloc_blk(struct gfs2_inode *ip, struct buffer_head **bhp) struct gfs2_ea_header *ea; u64 block; - block = gfs2_alloc_meta(ip); + block = gfs2_alloc_block(ip); gfs2_trans_add_unrevoke(sdp, block, 1); *bhp = gfs2_meta_new(ip->i_gl, block); gfs2_trans_add_bh(ip->i_gl, *bhp, 1); @@ -643,7 +643,7 @@ static int ea_write(struct gfs2_inode *ip, struct gfs2_ea_header *ea, u64 block; int mh_size = sizeof(struct gfs2_meta_header); - block = gfs2_alloc_meta(ip); + block = gfs2_alloc_block(ip); gfs2_trans_add_unrevoke(sdp, block, 1); bh = gfs2_meta_new(ip->i_gl, block); gfs2_trans_add_bh(ip->i_gl, bh, 1); @@ -967,7 +967,7 @@ static int ea_set_block(struct gfs2_inode *ip, struct gfs2_ea_request *er, } else { u64 blk; - blk = gfs2_alloc_meta(ip); + blk = gfs2_alloc_block(ip); gfs2_trans_add_unrevoke(sdp, blk, 1); indbh = gfs2_meta_new(ip->i_gl, blk); gfs2_trans_add_bh(ip->i_gl, indbh, 1); diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index 9f28463e62e..274a2df13f0 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -1405,58 +1405,13 @@ static struct gfs2_rgrpd *rgblk_free(struct gfs2_sbd *sdp, u64 bstart, } /** - * gfs2_alloc_data - Allocate a data block - * @ip: the inode to allocate the data block for + * gfs2_alloc_block - Allocate a block + * @ip: the inode to allocate the block for * * Returns: the allocated block */ -u64 gfs2_alloc_data(struct gfs2_inode *ip) -{ - struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); - struct gfs2_alloc *al = ip->i_alloc; - struct gfs2_rgrpd *rgd = al->al_rgd; - u32 goal, blk; - u64 block; - - if (rgrp_contains_block(rgd, ip->i_goal)) - goal = ip->i_goal - rgd->rd_data0; - else - goal = rgd->rd_last_alloc; - - blk = rgblk_search(rgd, goal, GFS2_BLKST_FREE, GFS2_BLKST_USED); - BUG_ON(blk == BFITNOENT); - rgd->rd_last_alloc = blk; - - block = rgd->rd_data0 + blk; - ip->i_goal = block; - - gfs2_assert_withdraw(sdp, rgd->rd_rg.rg_free); - rgd->rd_rg.rg_free--; - - gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1); - gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data); - - al->al_alloced++; - - gfs2_statfs_change(sdp, 0, -1, 0); - gfs2_quota_change(ip, +1, ip->i_inode.i_uid, ip->i_inode.i_gid); - - spin_lock(&sdp->sd_rindex_spin); - rgd->rd_free_clone--; - spin_unlock(&sdp->sd_rindex_spin); - - return block; -} - -/** - * gfs2_alloc_meta - Allocate a metadata block - * @ip: the inode to allocate the metadata block for - * - * Returns: the allocated block - */ - -u64 gfs2_alloc_meta(struct gfs2_inode *ip) +u64 gfs2_alloc_block(struct gfs2_inode *ip) { struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); struct gfs2_alloc *al = ip->i_alloc; diff --git a/fs/gfs2/rgrp.h b/fs/gfs2/rgrp.h index 5683605695f..5e66613b33d 100644 --- a/fs/gfs2/rgrp.h +++ b/fs/gfs2/rgrp.h @@ -46,8 +46,7 @@ void gfs2_inplace_release(struct gfs2_inode *ip); unsigned char gfs2_get_block_type(struct gfs2_rgrpd *rgd, u64 block); -u64 gfs2_alloc_data(struct gfs2_inode *ip); -u64 gfs2_alloc_meta(struct gfs2_inode *ip); +u64 gfs2_alloc_block(struct gfs2_inode *ip); u64 gfs2_alloc_di(struct gfs2_inode *ip, u64 *generation); void gfs2_free_data(struct gfs2_inode *ip, u64 bstart, u32 blen); -- cgit v1.2.3 From b45e41d7d56dfef1ae9e02e6c59990066ba82e5c Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Wed, 6 Feb 2008 10:11:15 +0000 Subject: [GFS2] Add extent allocation to block allocator Rather than having to allocate a single block at a time, this patch allows the block allocator to allocate an extent. Since there is no difference (so far as the block allocator is concerned) between data blocks and indirect blocks, it is posible to allocate a single extent and for the caller to unrevoke just the blocks required for indirect blocks. Currently the only bit of GFS2 to make use of this feature is the build height function. The intention is that gfs2_block_map will be changed to make use of this feature in future patches. Signed-off-by: Steven Whitehouse --- fs/gfs2/bmap.c | 26 ++++++++------- fs/gfs2/dir.c | 3 +- fs/gfs2/eattr.c | 10 +++--- fs/gfs2/rgrp.c | 100 +++++++++++++++++++++++++++++++++++--------------------- fs/gfs2/rgrp.h | 2 +- 5 files changed, 87 insertions(+), 54 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index e3a75a27cee..1fda731c074 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -136,8 +136,9 @@ int gfs2_unstuff_dinode(struct gfs2_inode *ip, struct page *page) /* Get a free block, fill it with the stuffed data, and write it out to disk */ + unsigned int n = 1; + block = gfs2_alloc_block(ip, &n); if (isdir) { - block = gfs2_alloc_block(ip); gfs2_trans_add_unrevoke(GFS2_SB(&ip->i_inode), block, 1); error = gfs2_dir_get_new_buffer(ip, block, &bh); if (error) @@ -146,8 +147,6 @@ int gfs2_unstuff_dinode(struct gfs2_inode *ip, struct page *page) dibh, sizeof(struct gfs2_dinode)); brelse(bh); } else { - block = gfs2_alloc_block(ip); - error = gfs2_unstuffer_page(ip, dibh, block, page); if (error) goto out_brelse; @@ -195,7 +194,7 @@ static int build_height(struct inode *inode, struct metapath *mp, unsigned heigh int error; __be64 *bp; u64 bn; - unsigned n; + unsigned n, i = 0; if (height <= ip->i_height) return 0; @@ -204,12 +203,16 @@ static int build_height(struct inode *inode, struct metapath *mp, unsigned heigh if (error) return error; - for(n = 0; n < new_height; n++) { - bn = gfs2_alloc_block(ip); - gfs2_trans_add_unrevoke(GFS2_SB(inode), bn, 1); - mp->mp_bh[n] = gfs2_meta_new(ip->i_gl, bn); - gfs2_trans_add_bh(ip->i_gl, mp->mp_bh[n], 1); - } + do { + n = new_height - i; + bn = gfs2_alloc_block(ip, &n); + gfs2_trans_add_unrevoke(GFS2_SB(inode), bn, n); + do { + mp->mp_bh[i] = gfs2_meta_new(ip->i_gl, bn++); + gfs2_trans_add_bh(ip->i_gl, mp->mp_bh[i], 1); + i++; + } while(i < n); + } while(i < new_height); n = 0; bn = mp->mp_bh[0]->b_blocknr; @@ -358,6 +361,7 @@ static int lookup_block(struct gfs2_inode *ip, unsigned int height, { int boundary; __be64 *ptr = metapointer(&boundary, height, mp); + unsigned int n = 1; if (*ptr) { *block = be64_to_cpu(*ptr); @@ -369,7 +373,7 @@ static int lookup_block(struct gfs2_inode *ip, unsigned int height, if (!create) return 0; - *block = gfs2_alloc_block(ip); + *block = gfs2_alloc_block(ip, &n); if (height != ip->i_height - 1 || gfs2_is_dir(ip)) gfs2_trans_add_unrevoke(GFS2_SB(&ip->i_inode), *block, 1); diff --git a/fs/gfs2/dir.c b/fs/gfs2/dir.c index fbdf31957cb..93a2e6afbd8 100644 --- a/fs/gfs2/dir.c +++ b/fs/gfs2/dir.c @@ -803,7 +803,8 @@ got_dent: static struct gfs2_leaf *new_leaf(struct inode *inode, struct buffer_head **pbh, u16 depth) { struct gfs2_inode *ip = GFS2_I(inode); - u64 bn = gfs2_alloc_block(ip); + unsigned int n = 1; + u64 bn = gfs2_alloc_block(ip, &n); struct buffer_head *bh = gfs2_meta_new(ip->i_gl, bn); struct gfs2_leaf *leaf; struct gfs2_dirent *dent; diff --git a/fs/gfs2/eattr.c b/fs/gfs2/eattr.c index f9f63bc21cd..0e79cd54349 100644 --- a/fs/gfs2/eattr.c +++ b/fs/gfs2/eattr.c @@ -582,9 +582,10 @@ static int ea_alloc_blk(struct gfs2_inode *ip, struct buffer_head **bhp) { struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); struct gfs2_ea_header *ea; + unsigned int n = 1; u64 block; - block = gfs2_alloc_block(ip); + block = gfs2_alloc_block(ip, &n); gfs2_trans_add_unrevoke(sdp, block, 1); *bhp = gfs2_meta_new(ip->i_gl, block); gfs2_trans_add_bh(ip->i_gl, *bhp, 1); @@ -642,8 +643,9 @@ static int ea_write(struct gfs2_inode *ip, struct gfs2_ea_header *ea, struct buffer_head *bh; u64 block; int mh_size = sizeof(struct gfs2_meta_header); + unsigned int n = 1; - block = gfs2_alloc_block(ip); + block = gfs2_alloc_block(ip, &n); gfs2_trans_add_unrevoke(sdp, block, 1); bh = gfs2_meta_new(ip->i_gl, block); gfs2_trans_add_bh(ip->i_gl, bh, 1); @@ -966,8 +968,8 @@ static int ea_set_block(struct gfs2_inode *ip, struct gfs2_ea_request *er, gfs2_trans_add_bh(ip->i_gl, indbh, 1); } else { u64 blk; - - blk = gfs2_alloc_block(ip); + unsigned int n = 1; + blk = gfs2_alloc_block(ip, &n); gfs2_trans_add_unrevoke(sdp, blk, 1); indbh = gfs2_meta_new(ip->i_gl, blk); gfs2_trans_add_bh(ip->i_gl, indbh, 1); diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index 274a2df13f0..77eba0a3804 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -53,7 +53,8 @@ static const char valid_change[16] = { }; static u32 rgblk_search(struct gfs2_rgrpd *rgd, u32 goal, - unsigned char old_state, unsigned char new_state); + unsigned char old_state, unsigned char new_state, + unsigned int *n); /** * gfs2_setbit - Set a bit in the bitmaps @@ -64,26 +65,32 @@ static u32 rgblk_search(struct gfs2_rgrpd *rgd, u32 goal, * */ -static void gfs2_setbit(struct gfs2_rgrpd *rgd, unsigned char *buffer, - unsigned int buflen, u32 block, - unsigned char new_state) +static inline void gfs2_setbit(struct gfs2_rgrpd *rgd, unsigned char *buf1, + unsigned char *buf2, unsigned int offset, + unsigned int buflen, u32 block, + unsigned char new_state) { - unsigned char *byte, *end, cur_state; - unsigned int bit; + unsigned char *byte1, *byte2, *end, cur_state; + const unsigned int bit = (block % GFS2_NBBY) * GFS2_BIT_SIZE; - byte = buffer + (block / GFS2_NBBY); - bit = (block % GFS2_NBBY) * GFS2_BIT_SIZE; - end = buffer + buflen; + byte1 = buf1 + offset + (block / GFS2_NBBY); + end = buf1 + offset + buflen; - gfs2_assert(rgd->rd_sbd, byte < end); + BUG_ON(byte1 >= end); - cur_state = (*byte >> bit) & GFS2_BIT_MASK; + cur_state = (*byte1 >> bit) & GFS2_BIT_MASK; - if (valid_change[new_state * 4 + cur_state]) { - *byte ^= cur_state << bit; - *byte |= new_state << bit; - } else + if (unlikely(!valid_change[new_state * 4 + cur_state])) { gfs2_consist_rgrpd(rgd); + return; + } + *byte1 ^= (cur_state ^ new_state) << bit; + + if (buf2) { + byte2 = buf2 + offset + (block / GFS2_NBBY); + cur_state = (*byte2 >> bit) & GFS2_BIT_MASK; + *byte2 ^= (cur_state ^ new_state) << bit; + } } /** @@ -94,10 +101,12 @@ static void gfs2_setbit(struct gfs2_rgrpd *rgd, unsigned char *buffer, * */ -static unsigned char gfs2_testbit(struct gfs2_rgrpd *rgd, unsigned char *buffer, - unsigned int buflen, u32 block) +static inline unsigned char gfs2_testbit(struct gfs2_rgrpd *rgd, + const unsigned char *buffer, + unsigned int buflen, u32 block) { - unsigned char *byte, *end, cur_state; + const unsigned char *byte, *end; + unsigned char cur_state; unsigned int bit; byte = buffer + (block / GFS2_NBBY); @@ -877,13 +886,15 @@ static struct inode *try_rgrp_unlink(struct gfs2_rgrpd *rgd, u64 *last_unlinked) u32 goal = 0, block; u64 no_addr; struct gfs2_sbd *sdp = rgd->rd_sbd; + unsigned int n; for(;;) { if (goal >= rgd->rd_data) break; down_write(&sdp->sd_log_flush_lock); + n = 1; block = rgblk_search(rgd, goal, GFS2_BLKST_UNLINKED, - GFS2_BLKST_UNLINKED); + GFS2_BLKST_UNLINKED, &n); up_write(&sdp->sd_log_flush_lock); if (block == BFITNOENT) break; @@ -1280,6 +1291,7 @@ unsigned char gfs2_get_block_type(struct gfs2_rgrpd *rgd, u64 block) * @goal: the goal block within the RG (start here to search for avail block) * @old_state: GFS2_BLKST_XXX the before-allocation state to find * @new_state: GFS2_BLKST_XXX the after-allocation block state + * @n: The extent length * * Walk rgrp's bitmap to find bits that represent a block in @old_state. * Add the found bitmap buffer to the transaction. @@ -1295,13 +1307,17 @@ unsigned char gfs2_get_block_type(struct gfs2_rgrpd *rgd, u64 block) */ static u32 rgblk_search(struct gfs2_rgrpd *rgd, u32 goal, - unsigned char old_state, unsigned char new_state) + unsigned char old_state, unsigned char new_state, + unsigned int *n) { struct gfs2_bitmap *bi = NULL; - u32 length = rgd->rd_length; + const u32 length = rgd->rd_length; u32 blk = 0; unsigned int buf, x; + const unsigned int elen = *n; + const u8 *buffer; + *n = 0; /* Find bitmap block that contains bits for goal block */ for (buf = 0; buf < length; buf++) { bi = rgd->rd_bits + buf; @@ -1322,7 +1338,7 @@ static u32 rgblk_search(struct gfs2_rgrpd *rgd, u32 goal, for (x = 0; x <= length; x++) { /* The GFS2_BLKST_UNLINKED state doesn't apply to the clone bitmaps, so we must search the originals for that. */ - const u8 *buffer = bi->bi_bh->b_data + bi->bi_offset; + buffer = bi->bi_bh->b_data + bi->bi_offset; if (old_state != GFS2_BLKST_UNLINKED && bi->bi_clone) buffer = bi->bi_clone + bi->bi_offset; @@ -1337,12 +1353,21 @@ static u32 rgblk_search(struct gfs2_rgrpd *rgd, u32 goal, } if (blk != BFITNOENT && old_state != new_state) { + *n = 1; gfs2_trans_add_bh(rgd->rd_gl, bi->bi_bh, 1); - gfs2_setbit(rgd, bi->bi_bh->b_data + bi->bi_offset, + gfs2_setbit(rgd, bi->bi_bh->b_data, bi->bi_clone, bi->bi_offset, bi->bi_len, blk, new_state); - if (bi->bi_clone) - gfs2_setbit(rgd, bi->bi_clone + bi->bi_offset, - bi->bi_len, blk, new_state); + while(*n < elen) { + goal++; + if (goal >= (bi->bi_len / GFS2_NBBY)) + break; + if (gfs2_testbit(rgd, buffer, bi->bi_len, goal) != + GFS2_BLKST_FREE) + break; + (*n)++; + gfs2_setbit(rgd, bi->bi_bh->b_data, bi->bi_clone, + bi->bi_offset, bi->bi_len, blk, new_state); + } } return (blk == BFITNOENT) ? blk : (bi->bi_start * GFS2_NBBY) + blk; @@ -1397,7 +1422,7 @@ static struct gfs2_rgrpd *rgblk_free(struct gfs2_sbd *sdp, u64 bstart, bi->bi_len); } gfs2_trans_add_bh(rgd->rd_gl, bi->bi_bh, 1); - gfs2_setbit(rgd, bi->bi_bh->b_data + bi->bi_offset, + gfs2_setbit(rgd, bi->bi_bh->b_data, NULL, bi->bi_offset, bi->bi_len, buf_blk, new_state); } @@ -1411,7 +1436,7 @@ static struct gfs2_rgrpd *rgblk_free(struct gfs2_sbd *sdp, u64 bstart, * Returns: the allocated block */ -u64 gfs2_alloc_block(struct gfs2_inode *ip) +u64 gfs2_alloc_block(struct gfs2_inode *ip, unsigned int *n) { struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); struct gfs2_alloc *al = ip->i_alloc; @@ -1424,26 +1449,26 @@ u64 gfs2_alloc_block(struct gfs2_inode *ip) else goal = rgd->rd_last_alloc; - blk = rgblk_search(rgd, goal, GFS2_BLKST_FREE, GFS2_BLKST_USED); + blk = rgblk_search(rgd, goal, GFS2_BLKST_FREE, GFS2_BLKST_USED, n); BUG_ON(blk == BFITNOENT); - rgd->rd_last_alloc = blk; + rgd->rd_last_alloc = blk; block = rgd->rd_data0 + blk; ip->i_goal = block; - gfs2_assert_withdraw(sdp, rgd->rd_rg.rg_free); - rgd->rd_rg.rg_free--; + gfs2_assert_withdraw(sdp, rgd->rd_rg.rg_free >= *n); + rgd->rd_rg.rg_free -= *n; gfs2_trans_add_bh(rgd->rd_gl, rgd->rd_bits[0].bi_bh, 1); gfs2_rgrp_out(rgd, rgd->rd_bits[0].bi_bh->b_data); - al->al_alloced++; + al->al_alloced += *n; - gfs2_statfs_change(sdp, 0, -1, 0); - gfs2_quota_change(ip, +1, ip->i_inode.i_uid, ip->i_inode.i_gid); + gfs2_statfs_change(sdp, 0, -*n, 0); + gfs2_quota_change(ip, *n, ip->i_inode.i_uid, ip->i_inode.i_gid); spin_lock(&sdp->sd_rindex_spin); - rgd->rd_free_clone--; + rgd->rd_free_clone -= *n; spin_unlock(&sdp->sd_rindex_spin); return block; @@ -1463,9 +1488,10 @@ u64 gfs2_alloc_di(struct gfs2_inode *dip, u64 *generation) struct gfs2_rgrpd *rgd = al->al_rgd; u32 blk; u64 block; + unsigned int n = 1; blk = rgblk_search(rgd, rgd->rd_last_alloc, - GFS2_BLKST_FREE, GFS2_BLKST_DINODE); + GFS2_BLKST_FREE, GFS2_BLKST_DINODE, &n); BUG_ON(blk == BFITNOENT); rgd->rd_last_alloc = blk; diff --git a/fs/gfs2/rgrp.h b/fs/gfs2/rgrp.h index 5e66613b33d..3181c7e624b 100644 --- a/fs/gfs2/rgrp.h +++ b/fs/gfs2/rgrp.h @@ -46,7 +46,7 @@ void gfs2_inplace_release(struct gfs2_inode *ip); unsigned char gfs2_get_block_type(struct gfs2_rgrpd *rgd, u64 block); -u64 gfs2_alloc_block(struct gfs2_inode *ip); +u64 gfs2_alloc_block(struct gfs2_inode *ip, unsigned int *n); u64 gfs2_alloc_di(struct gfs2_inode *ip, u64 *generation); void gfs2_free_data(struct gfs2_inode *ip, u64 bstart, u32 blen); -- cgit v1.2.3 From c85a665f064863cc8a2fe88e5f1eb4def5446e90 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Tue, 12 Feb 2008 12:14:59 +0000 Subject: [GFS2] The case of the missing asterisk A dereference was forgotten. This adds it back correctly. Signed-off-by: Steven Whitehouse --- fs/gfs2/bmap.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index 1fda731c074..7f72564e059 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -402,7 +402,7 @@ static int lookup_metapath(struct inode *inode, struct metapath *mp, for (x = 0; x < end_of_metadata; x++) { lookup_block(ip, x, mp, create, new, dblock); - if (!dblock) + if (!*dblock) return 0; ret = gfs2_meta_indirect_buffer(ip, x+1, *dblock, *new, &mp->mp_bh[x+1]); -- cgit v1.2.3 From 30cbf189cd2a1ba13ff3c8c8ee2103dbdb18578a Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Fri, 8 Feb 2008 13:18:11 +0000 Subject: [GFS2] Add a function to interate over an extent This adds a function (currently the only use is during mapping of already allocated blocks, but watch this space) which iterates over a number of pointers in a block and returns the extent length. If the initial pointer is 0 (i.e. unallocated) it will return the number of unallocated blocks in the extent. If the initial pointer is allocated, then it returns the number of contiguously allocated blocks in the extent. Signed-off-by: Steven Whitehouse --- fs/gfs2/bmap.c | 68 ++++++++++++++++++++++++++++++++++++++++++++-------------- 1 file changed, 52 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index 7f72564e059..6780aa5841b 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -422,6 +422,42 @@ static void release_metapath(struct metapath *mp) brelse(mp->mp_bh[i]); } +/** + * gfs2_extent_length - Returns length of an extent of blocks + * @start: Start of the buffer + * @len: Length of the buffer in bytes + * @ptr: Current position in the buffer + * @limit: Max extent length to return (0 = unlimited) + * @eob: Set to 1 if we hit "end of block" + * + * If the first block is zero (unallocated) it will return the number of + * unallocated blocks in the extent, otherwise it will return the number + * of contiguous blocks in the extent. + * + * Returns: The length of the extent (minimum of one block) + */ + +static inline unsigned int gfs2_extent_length(void *start, unsigned int len, __be64 *ptr, unsigned limit, int *eob) +{ + const __be64 *end = (start + len); + const __be64 *first = ptr; + u64 d = be64_to_cpu(*ptr); + + *eob = 0; + do { + ptr++; + if (ptr >= end) + break; + if (limit && --limit == 0) + break; + if (d) + d++; + } while(be64_to_cpu(*ptr) == d); + if (ptr >= end) + *eob = 1; + return (ptr - first); +} + static inline void bmap_lock(struct inode *inode, int create) { struct gfs2_inode *ip = GFS2_I(inode); @@ -499,26 +535,26 @@ int gfs2_block_map(struct inode *inode, sector_t lblock, goto out_fail; boundary = error; + if (new) { + map_bh(bh_map, inode->i_sb, dblock); + if (boundary) + set_buffer_boundary(bh_map); + gfs2_trans_add_bh(ip->i_gl, mp.mp_bh[0], 1); + gfs2_dinode_out(ip, mp.mp_bh[0]->b_data); + set_buffer_new(bh_map); + goto out_ok; + } + if (dblock) { + unsigned int len; + struct buffer_head *bh = mp.mp_bh[ip->i_height - 1]; + __be64 *ptr = metapointer(&boundary, ip->i_height - 1, &mp); map_bh(bh_map, inode->i_sb, dblock); + len = gfs2_extent_length(bh->b_data, bh->b_size, ptr, maxlen, + &boundary); + bh_map->b_size = (len << inode->i_blkbits); if (boundary) set_buffer_boundary(bh_map); - if (new) { - gfs2_trans_add_bh(ip->i_gl, mp.mp_bh[0], 1); - gfs2_dinode_out(ip, mp.mp_bh[0]->b_data); - set_buffer_new(bh_map); - goto out_ok; - } - while(--maxlen && !buffer_boundary(bh_map)) { - u64 eblock; - mp.mp_list[ip->i_height - 1]++; - boundary = lookup_block(ip, ip->i_height - 1, &mp, 0, &new, &eblock); - if (eblock != ++dblock) - break; - bh_map->b_size += (1 << inode->i_blkbits); - if (boundary) - set_buffer_boundary(bh_map); - } } out_ok: error = 0; -- cgit v1.2.3 From 77658aad226866fb94097236d14d41a88aaab2ec Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Tue, 12 Feb 2008 14:17:27 +0000 Subject: [GFS2] Eliminate (almost) duplicate field from gfs2_inode The blocks counter is almost a duplicate of the i_blocks field in the VFS inode. The only difference is that i_blocks can be only 32bits long for 32bit arch without large single file support. Since GFS2 doesn't handle the non-large single file case (for 32 bit anyway) this adds a new config dependency on 64BIT || LSF. This has always been the case, however we've never explicitly said so before. Even if we do add support for the non-LSF case, we will still not require this field to be duplicated since we will not be able to access oversized files anyway. So the net result of all this is that we shave 8 bytes from a gfs2_inode and get our config deps correct. Signed-off-by: Steven Whitehouse --- fs/gfs2/Kconfig | 2 +- fs/gfs2/bmap.c | 18 ++++++------------ fs/gfs2/dir.c | 15 ++++----------- fs/gfs2/eattr.c | 24 ++++++------------------ fs/gfs2/incore.h | 1 - fs/gfs2/inode.c | 11 +++++------ fs/gfs2/inode.h | 20 +++++++++++++++++--- fs/gfs2/ops_inode.c | 5 +++-- 8 files changed, 42 insertions(+), 54 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/Kconfig b/fs/gfs2/Kconfig index de8e64c03f7..d147b53ef52 100644 --- a/fs/gfs2/Kconfig +++ b/fs/gfs2/Kconfig @@ -1,6 +1,6 @@ config GFS2_FS tristate "GFS2 file system support" - depends on EXPERIMENTAL + depends on EXPERIMENTAL && (64BIT || LSF) select FS_POSIX_ACL select CRC32 help diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index 6780aa5841b..e27e66046f0 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -161,9 +161,8 @@ int gfs2_unstuff_dinode(struct gfs2_inode *ip, struct page *page) if (ip->i_di.di_size) { *(__be64 *)(di + 1) = cpu_to_be64(block); - ip->i_di.di_blocks++; - gfs2_set_inode_blocks(&ip->i_inode); - di->di_blocks = cpu_to_be64(ip->i_di.di_blocks); + gfs2_add_inode_blocks(&ip->i_inode, 1); + di->di_blocks = cpu_to_be64(gfs2_get_inode_blocks(&ip->i_inode)); } ip->i_height = 1; @@ -238,10 +237,9 @@ static int build_height(struct inode *inode, struct metapath *mp, unsigned heigh gfs2_buffer_clear_tail(dibh, sizeof(struct gfs2_dinode)); *(__be64 *)(di + 1) = cpu_to_be64(bn); ip->i_height += new_height; - ip->i_di.di_blocks += new_height; - gfs2_set_inode_blocks(&ip->i_inode); + gfs2_add_inode_blocks(&ip->i_inode, new_height); di->di_height = cpu_to_be16(ip->i_height); - di->di_blocks = cpu_to_be64(ip->i_di.di_blocks); + di->di_blocks = cpu_to_be64(gfs2_get_inode_blocks(&ip->i_inode)); brelse(dibh); return error; } @@ -380,8 +378,7 @@ static int lookup_block(struct gfs2_inode *ip, unsigned int height, gfs2_trans_add_bh(ip->i_gl, mp->mp_bh[height], 1); *ptr = cpu_to_be64(*block); - ip->i_di.di_blocks++; - gfs2_set_inode_blocks(&ip->i_inode); + gfs2_add_inode_blocks(&ip->i_inode, 1); *new = 1; return 0; @@ -779,10 +776,7 @@ static int do_strip(struct gfs2_inode *ip, struct buffer_head *dibh, } *p = 0; - if (!ip->i_di.di_blocks) - gfs2_consist_inode(ip); - ip->i_di.di_blocks--; - gfs2_set_inode_blocks(&ip->i_inode); + gfs2_add_inode_blocks(&ip->i_inode, -1); } if (bstart) { if (metadata) diff --git a/fs/gfs2/dir.c b/fs/gfs2/dir.c index 93a2e6afbd8..862aa3228f7 100644 --- a/fs/gfs2/dir.c +++ b/fs/gfs2/dir.c @@ -906,8 +906,7 @@ static int dir_make_exhash(struct inode *inode) *lp = cpu_to_be64(bn); dip->i_di.di_size = sdp->sd_sb.sb_bsize / 2; - dip->i_di.di_blocks++; - gfs2_set_inode_blocks(&dip->i_inode); + gfs2_add_inode_blocks(&dip->i_inode, 1); dip->i_di.di_flags |= GFS2_DIF_EXHASH; for (x = sdp->sd_hash_ptrs, y = -1; x; x >>= 1, y++) ; @@ -1045,8 +1044,7 @@ static int dir_split_leaf(struct inode *inode, const struct qstr *name) error = gfs2_meta_inode_buffer(dip, &dibh); if (!gfs2_assert_withdraw(GFS2_SB(&dip->i_inode), !error)) { gfs2_trans_add_bh(dip->i_gl, dibh, 1); - dip->i_di.di_blocks++; - gfs2_set_inode_blocks(&dip->i_inode); + gfs2_add_inode_blocks(&dip->i_inode, 1); gfs2_dinode_out(dip, dibh->b_data); brelse(dibh); } @@ -1580,8 +1578,7 @@ static int dir_new_leaf(struct inode *inode, const struct qstr *name) if (error) return error; gfs2_trans_add_bh(ip->i_gl, bh, 1); - ip->i_di.di_blocks++; - gfs2_set_inode_blocks(&ip->i_inode); + gfs2_add_inode_blocks(&ip->i_inode, 1); gfs2_dinode_out(ip, bh->b_data); brelse(bh); return 0; @@ -1922,11 +1919,7 @@ static int leaf_dealloc(struct gfs2_inode *dip, u32 index, u32 len, brelse(bh); gfs2_free_meta(dip, blk, 1); - - if (!dip->i_di.di_blocks) - gfs2_consist_inode(dip); - dip->i_di.di_blocks--; - gfs2_set_inode_blocks(&dip->i_inode); + gfs2_add_inode_blocks(&dip->i_inode, -1); } error = gfs2_dir_write_data(dip, ht, index * sizeof(u64), size); diff --git a/fs/gfs2/eattr.c b/fs/gfs2/eattr.c index 0e79cd54349..76ead1acfcc 100644 --- a/fs/gfs2/eattr.c +++ b/fs/gfs2/eattr.c @@ -277,10 +277,7 @@ static int ea_dealloc_unstuffed(struct gfs2_inode *ip, struct buffer_head *bh, } *dataptrs = 0; - if (!ip->i_di.di_blocks) - gfs2_consist_inode(ip); - ip->i_di.di_blocks--; - gfs2_set_inode_blocks(&ip->i_inode); + gfs2_add_inode_blocks(&ip->i_inode, -1); } if (bstart) gfs2_free_meta(ip, bstart, blen); @@ -598,8 +595,7 @@ static int ea_alloc_blk(struct gfs2_inode *ip, struct buffer_head **bhp) ea->ea_flags = GFS2_EAFLAG_LAST; ea->ea_num_ptrs = 0; - ip->i_di.di_blocks++; - gfs2_set_inode_blocks(&ip->i_inode); + gfs2_add_inode_blocks(&ip->i_inode, 1); return 0; } @@ -651,8 +647,7 @@ static int ea_write(struct gfs2_inode *ip, struct gfs2_ea_header *ea, gfs2_trans_add_bh(ip->i_gl, bh, 1); gfs2_metatype_set(bh, GFS2_METATYPE_ED, GFS2_FORMAT_ED); - ip->i_di.di_blocks++; - gfs2_set_inode_blocks(&ip->i_inode); + gfs2_add_inode_blocks(&ip->i_inode, 1); copy = data_len > sdp->sd_jbsize ? sdp->sd_jbsize : data_len; @@ -980,8 +975,7 @@ static int ea_set_block(struct gfs2_inode *ip, struct gfs2_ea_request *er, *eablk = cpu_to_be64(ip->i_di.di_eattr); ip->i_di.di_eattr = blk; ip->i_di.di_flags |= GFS2_DIF_EA_INDIRECT; - ip->i_di.di_blocks++; - gfs2_set_inode_blocks(&ip->i_inode); + gfs2_add_inode_blocks(&ip->i_inode, 1); eablk++; } @@ -1389,10 +1383,7 @@ static int ea_dealloc_indirect(struct gfs2_inode *ip) } *eablk = 0; - if (!ip->i_di.di_blocks) - gfs2_consist_inode(ip); - ip->i_di.di_blocks--; - gfs2_set_inode_blocks(&ip->i_inode); + gfs2_add_inode_blocks(&ip->i_inode, -1); } if (bstart) gfs2_free_meta(ip, bstart, blen); @@ -1444,10 +1435,7 @@ static int ea_dealloc_block(struct gfs2_inode *ip) gfs2_free_meta(ip, ip->i_di.di_eattr, 1); ip->i_di.di_eattr = 0; - if (!ip->i_di.di_blocks) - gfs2_consist_inode(ip); - ip->i_di.di_blocks--; - gfs2_set_inode_blocks(&ip->i_inode); + gfs2_add_inode_blocks(&ip->i_inode, -1); error = gfs2_meta_inode_buffer(ip, &dibh); if (!error) { diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index 898b456b386..4ba2ea63119 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -239,7 +239,6 @@ enum { struct gfs2_dinode_host { u64 di_size; /* number of bytes in file */ - u64 di_blocks; /* number of blocks in file */ u64 di_generation; /* generation number for NFS */ u32 di_flags; /* GFS2_DIF_... */ /* These only apply to directories */ diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c index c3fe8aa03c4..5f50dd53bf6 100644 --- a/fs/gfs2/inode.c +++ b/fs/gfs2/inode.c @@ -273,8 +273,7 @@ static int gfs2_dinode_in(struct gfs2_inode *ip, const void *buf) ip->i_inode.i_nlink = be32_to_cpu(str->di_nlink); di->di_size = be64_to_cpu(str->di_size); i_size_write(&ip->i_inode, di->di_size); - di->di_blocks = be64_to_cpu(str->di_blocks); - gfs2_set_inode_blocks(&ip->i_inode); + gfs2_set_inode_blocks(&ip->i_inode, be64_to_cpu(str->di_blocks)); ip->i_inode.i_atime.tv_sec = be64_to_cpu(str->di_atime); ip->i_inode.i_atime.tv_nsec = be32_to_cpu(str->di_atime_nsec); ip->i_inode.i_mtime.tv_sec = be64_to_cpu(str->di_mtime); @@ -344,7 +343,7 @@ int gfs2_dinode_dealloc(struct gfs2_inode *ip) struct gfs2_rgrpd *rgd; int error; - if (ip->i_di.di_blocks != 1) { + if (gfs2_get_inode_blocks(&ip->i_inode) != 1) { if (gfs2_consist_inode(ip)) gfs2_dinode_print(ip); return -EIO; @@ -1398,7 +1397,7 @@ void gfs2_dinode_out(const struct gfs2_inode *ip, void *buf) str->di_gid = cpu_to_be32(ip->i_inode.i_gid); str->di_nlink = cpu_to_be32(ip->i_inode.i_nlink); str->di_size = cpu_to_be64(di->di_size); - str->di_blocks = cpu_to_be64(di->di_blocks); + str->di_blocks = cpu_to_be64(gfs2_get_inode_blocks(&ip->i_inode)); str->di_atime = cpu_to_be64(ip->i_inode.i_atime.tv_sec); str->di_mtime = cpu_to_be64(ip->i_inode.i_mtime.tv_sec); str->di_ctime = cpu_to_be64(ip->i_inode.i_ctime.tv_sec); @@ -1430,8 +1429,8 @@ void gfs2_dinode_print(const struct gfs2_inode *ip) printk(KERN_INFO " no_addr = %llu\n", (unsigned long long)ip->i_no_addr); printk(KERN_INFO " di_size = %llu\n", (unsigned long long)di->di_size); - printk(KERN_INFO " di_blocks = %llu\n", - (unsigned long long)di->di_blocks); + printk(KERN_INFO " blocks = %llu\n", + (unsigned long long)gfs2_get_inode_blocks(&ip->i_inode)); printk(KERN_INFO " i_goal = %llu\n", (unsigned long long)ip->i_goal); printk(KERN_INFO " di_flags = 0x%.8X\n", di->di_flags); diff --git a/fs/gfs2/inode.h b/fs/gfs2/inode.h index db738686ca1..580da454b38 100644 --- a/fs/gfs2/inode.h +++ b/fs/gfs2/inode.h @@ -10,6 +10,8 @@ #ifndef __INODE_DOT_H__ #define __INODE_DOT_H__ +#include "util.h" + static inline int gfs2_is_stuffed(const struct gfs2_inode *ip) { return !ip->i_height; @@ -37,13 +39,25 @@ static inline int gfs2_is_dir(const struct gfs2_inode *ip) return S_ISDIR(ip->i_inode.i_mode); } -static inline void gfs2_set_inode_blocks(struct inode *inode) +static inline void gfs2_set_inode_blocks(struct inode *inode, u64 blocks) +{ + inode->i_blocks = blocks << + (GFS2_SB(inode)->sd_sb.sb_bsize_shift - GFS2_BASIC_BLOCK_SHIFT); +} + +static inline u64 gfs2_get_inode_blocks(const struct inode *inode) { - struct gfs2_inode *ip = GFS2_I(inode); - inode->i_blocks = ip->i_di.di_blocks << + return inode->i_blocks >> (GFS2_SB(inode)->sd_sb.sb_bsize_shift - GFS2_BASIC_BLOCK_SHIFT); } +static inline void gfs2_add_inode_blocks(struct inode *inode, s64 change) +{ + gfs2_assert(GFS2_SB(inode), (change >= 0 || inode->i_blocks > -change)); + change *= (GFS2_SB(inode)->sd_sb.sb_bsize/GFS2_BASIC_BLOCK); + inode->i_blocks += change; +} + static inline int gfs2_check_inum(const struct gfs2_inode *ip, u64 no_addr, u64 no_formal_ino) { diff --git a/fs/gfs2/ops_inode.c b/fs/gfs2/ops_inode.c index e87412902be..301c9459667 100644 --- a/fs/gfs2/ops_inode.c +++ b/fs/gfs2/ops_inode.c @@ -981,8 +981,9 @@ static int setattr_chown(struct inode *inode, struct iattr *attr) brelse(dibh); if (ouid != NO_QUOTA_CHANGE || ogid != NO_QUOTA_CHANGE) { - gfs2_quota_change(ip, -ip->i_di.di_blocks, ouid, ogid); - gfs2_quota_change(ip, ip->i_di.di_blocks, nuid, ngid); + u64 blocks = gfs2_get_inode_blocks(&ip->i_inode); + gfs2_quota_change(ip, -blocks, ouid, ogid); + gfs2_quota_change(ip, blocks, nuid, ngid); } out_end_trans: -- cgit v1.2.3 From e23159d2a7b2df5bce5f0ee8d57d3292243abf66 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Tue, 12 Feb 2008 14:48:39 +0000 Subject: [GFS2] Get inode buffer only once per block map call In the case that we needed to grow the height of the metadata tree we were looking up the inode buffer and then brelse()ing it despite the fact that it is needed later in the block map process. This patch ensures that we look up the inode's buffer once and only once during the block map process. Signed-off-by: Steven Whitehouse --- fs/gfs2/bmap.c | 48 +++++++++++++++++++++--------------------------- 1 file changed, 21 insertions(+), 27 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index e27e66046f0..f1f38ca77a5 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -188,50 +188,45 @@ static int build_height(struct inode *inode, struct metapath *mp, unsigned heigh { struct gfs2_inode *ip = GFS2_I(inode); unsigned new_height = height - ip->i_height; - struct buffer_head *dibh; + struct buffer_head *dibh = mp->mp_bh[0]; struct gfs2_dinode *di; - int error; __be64 *bp; u64 bn; unsigned n, i = 0; - if (height <= ip->i_height) - return 0; - - error = gfs2_meta_inode_buffer(ip, &dibh); - if (error) - return error; + BUG_ON(height <= ip->i_height); do { n = new_height - i; bn = gfs2_alloc_block(ip, &n); gfs2_trans_add_unrevoke(GFS2_SB(inode), bn, n); do { - mp->mp_bh[i] = gfs2_meta_new(ip->i_gl, bn++); - gfs2_trans_add_bh(ip->i_gl, mp->mp_bh[i], 1); + mp->mp_bh[i + 1] = gfs2_meta_new(ip->i_gl, bn++); + gfs2_trans_add_bh(ip->i_gl, mp->mp_bh[i + 1], 1); i++; } while(i < n); } while(i < new_height); n = 0; - bn = mp->mp_bh[0]->b_blocknr; + bn = mp->mp_bh[1]->b_blocknr; if (new_height > 1) { for(; n < new_height-1; n++) { - gfs2_metatype_set(mp->mp_bh[n], GFS2_METATYPE_IN, + gfs2_metatype_set(mp->mp_bh[n + 1], GFS2_METATYPE_IN, GFS2_FORMAT_IN); - gfs2_buffer_clear_tail(mp->mp_bh[n], + gfs2_buffer_clear_tail(mp->mp_bh[n + 1], sizeof(struct gfs2_meta_header)); - bp = (__be64 *)(mp->mp_bh[n]->b_data + + bp = (__be64 *)(mp->mp_bh[n + 1]->b_data + sizeof(struct gfs2_meta_header)); - *bp = cpu_to_be64(mp->mp_bh[n+1]->b_blocknr); - brelse(mp->mp_bh[n]); - mp->mp_bh[n] = NULL; + *bp = cpu_to_be64(mp->mp_bh[n+2]->b_blocknr); + brelse(mp->mp_bh[n+1]); + mp->mp_bh[n+1] = NULL; } } - gfs2_metatype_set(mp->mp_bh[n], GFS2_METATYPE_IN, GFS2_FORMAT_IN); - gfs2_buffer_copy_tail(mp->mp_bh[n], sizeof(struct gfs2_meta_header), + gfs2_metatype_set(mp->mp_bh[n+1], GFS2_METATYPE_IN, GFS2_FORMAT_IN); + gfs2_buffer_copy_tail(mp->mp_bh[n+1], sizeof(struct gfs2_meta_header), dibh, sizeof(struct gfs2_dinode)); - brelse(mp->mp_bh[n]); + brelse(mp->mp_bh[n+1]); + mp->mp_bh[n+1] = NULL; gfs2_trans_add_bh(ip->i_gl, dibh, 1); di = (struct gfs2_dinode *)dibh->b_data; gfs2_buffer_clear_tail(dibh, sizeof(struct gfs2_dinode)); @@ -240,8 +235,7 @@ static int build_height(struct inode *inode, struct metapath *mp, unsigned heigh gfs2_add_inode_blocks(&ip->i_inode, new_height); di->di_height = cpu_to_be16(ip->i_height); di->di_blocks = cpu_to_be64(gfs2_get_inode_blocks(&ip->i_inode)); - brelse(dibh); - return error; + return 0; } /** @@ -391,11 +385,7 @@ static int lookup_metapath(struct inode *inode, struct metapath *mp, struct gfs2_inode *ip = GFS2_I(inode); unsigned int end_of_metadata = ip->i_height - 1; unsigned int x; - int ret = gfs2_meta_inode_buffer(ip, &bh); - if (ret) - return ret; - - mp->mp_bh[0] = bh; + int ret; for (x = 0; x < end_of_metadata; x++) { lookup_block(ip, x, mp, create, new, dblock); @@ -515,6 +505,10 @@ int gfs2_block_map(struct inode *inode, sector_t lblock, } size = (lblock + 1) * bsize; + error = gfs2_meta_inode_buffer(ip, &mp.mp_bh[0]); + if (error) + goto out_fail; + if (size > arr[ip->i_height]) { u8 height = ip->i_height; if (!create) -- cgit v1.2.3 From 840ca0ec70903ce8e0fba1596460876c796e4f60 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Tue, 12 Feb 2008 15:28:21 +0000 Subject: [GFS2] Fix bug where we called drop_bh incorrectly As a result of an earlier patch, drop_bh was being called in cases when it shouldn't have been. Since we never have a gh in the drop case and we always have a gh in the promote case, we can use that extra information to tell which case has been seen. Signed-off-by: Steven Whitehouse Cc: Bob Peterson --- fs/gfs2/glock.c | 15 +-------------- 1 file changed, 1 insertion(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index d00dc37e3d5..63981e2fb83 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -765,7 +765,6 @@ static void drop_bh(struct gfs2_glock *gl, unsigned int ret) { struct gfs2_sbd *sdp = gl->gl_sbd; const struct gfs2_glock_operations *glops = gl->gl_ops; - struct gfs2_holder *gh = gl->gl_req_gh; gfs2_assert_warn(sdp, test_bit(GLF_LOCK, &gl->gl_flags)); gfs2_assert_warn(sdp, list_empty(&gl->gl_holders)); @@ -776,23 +775,11 @@ static void drop_bh(struct gfs2_glock *gl, unsigned int ret) if (glops->go_inval) glops->go_inval(gl, DIO_METADATA); - if (gh) { - spin_lock(&gl->gl_spin); - list_del_init(&gh->gh_list); - gh->gh_error = 0; - spin_unlock(&gl->gl_spin); - } - spin_lock(&gl->gl_spin); gfs2_demote_wake(gl); - gl->gl_req_gh = NULL; clear_bit(GLF_LOCK, &gl->gl_flags); spin_unlock(&gl->gl_spin); - gfs2_glock_put(gl); - - if (gh) - gfs2_holder_wake(gh); } /** @@ -810,7 +797,7 @@ static void xmote_bh(struct gfs2_glock *gl, unsigned int ret) int prev_state = gl->gl_state; int op_done = 1; - if ((ret & LM_OUT_ST_MASK) == LM_ST_UNLOCKED) { + if (!gh && (ret & LM_OUT_ST_MASK) == LM_ST_UNLOCKED) { drop_bh(gl, ret); return; } -- cgit v1.2.3 From bb16b342b2e2c83fa47dbb042400db91b748ded7 Mon Sep 17 00:00:00 2001 From: Marcin Slusarz Date: Wed, 13 Feb 2008 00:06:10 +0100 Subject: [GFS2] be*_add_cpu conversion replace all: big_endian_variable = cpu_to_beX(beX_to_cpu(big_endian_variable) + expression_in_cpu_byteorder); with: beX_add_cpu(&big_endian_variable, expression_in_cpu_byteorder); generated with semantic patch Signed-off-by: Marcin Slusarz Signed-off-by: Steven Whitehouse --- fs/gfs2/dir.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/dir.c b/fs/gfs2/dir.c index 862aa3228f7..34dc8dfaba1 100644 --- a/fs/gfs2/dir.c +++ b/fs/gfs2/dir.c @@ -1021,13 +1021,13 @@ static int dir_split_leaf(struct inode *inode, const struct qstr *name) new->de_inum = dent->de_inum; /* No endian worries */ new->de_type = dent->de_type; /* No endian worries */ - nleaf->lf_entries = cpu_to_be16(be16_to_cpu(nleaf->lf_entries)+1); + be16_add_cpu(&nleaf->lf_entries, 1); dirent_del(dip, obh, prev, dent); if (!oleaf->lf_entries) gfs2_consist_inode(dip); - oleaf->lf_entries = cpu_to_be16(be16_to_cpu(oleaf->lf_entries)-1); + be16_add_cpu(&oleaf->lf_entries, -1); if (!prev) prev = dent; @@ -1614,7 +1614,7 @@ int gfs2_dir_add(struct inode *inode, const struct qstr *name, dent->de_type = cpu_to_be16(type); if (ip->i_di.di_flags & GFS2_DIF_EXHASH) { leaf = (struct gfs2_leaf *)bh->b_data; - leaf->lf_entries = cpu_to_be16(be16_to_cpu(leaf->lf_entries) + 1); + be16_add_cpu(&leaf->lf_entries, 1); } brelse(bh); error = gfs2_meta_inode_buffer(ip, &bh); -- cgit v1.2.3 From 8af4c72f7df2442230fca3ff49a97f978cfb4a04 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Sun, 17 Feb 2008 10:17:12 +0200 Subject: [GFS2] gfs2/ops_file.c should #include "ops_inode.h" Every file should include the headers containing the prototypes for its global functions (in this case for gfs2_set_inode_flags()). Signed-off-by: Adrian Bunk Signed-off-by: Steven Whitehouse --- fs/gfs2/ops_file.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/gfs2/ops_file.c b/fs/gfs2/ops_file.c index f97a8b86c48..2b25a5f7a1c 100644 --- a/fs/gfs2/ops_file.c +++ b/fs/gfs2/ops_file.c @@ -38,6 +38,7 @@ #include "util.h" #include "eaops.h" #include "ops_address.h" +#include "ops_inode.h" /** * gfs2_llseek - seek to a location in a file -- cgit v1.2.3 From 60b779cfc1fa52034a996ee12a23b62d32e86000 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Sun, 17 Feb 2008 10:20:34 +0200 Subject: [GFS2] proper extern for gfs2/locking/dlm/mount.c:gdlm_ops This patch adds a proper extern declaration for gdlm_ops in fs/gfs2/locking/dlm/lock_dlm.h Signed-off-by: Adrian Bunk Signed-off-by: Steven Whitehouse --- fs/gfs2/locking/dlm/lock_dlm.h | 5 +++++ fs/gfs2/locking/dlm/main.c | 2 -- fs/gfs2/locking/dlm/sysfs.c | 2 -- 3 files changed, 5 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/locking/dlm/lock_dlm.h b/fs/gfs2/locking/dlm/lock_dlm.h index 9e8265d2837..58fcf8c5bf3 100644 --- a/fs/gfs2/locking/dlm/lock_dlm.h +++ b/fs/gfs2/locking/dlm/lock_dlm.h @@ -183,5 +183,10 @@ int gdlm_plock_get(void *, struct lm_lockname *, struct file *, struct file_lock *); int gdlm_punlock(void *, struct lm_lockname *, struct file *, struct file_lock *); + +/* mount.c */ + +extern const struct lm_lockops gdlm_ops; + #endif diff --git a/fs/gfs2/locking/dlm/main.c b/fs/gfs2/locking/dlm/main.c index a0e7eda643e..36a225850bd 100644 --- a/fs/gfs2/locking/dlm/main.c +++ b/fs/gfs2/locking/dlm/main.c @@ -11,8 +11,6 @@ #include "lock_dlm.h" -extern struct lm_lockops gdlm_ops; - static int __init init_lock_dlm(void) { int error; diff --git a/fs/gfs2/locking/dlm/sysfs.c b/fs/gfs2/locking/dlm/sysfs.c index a87b0983976..8479da47049 100644 --- a/fs/gfs2/locking/dlm/sysfs.c +++ b/fs/gfs2/locking/dlm/sysfs.c @@ -12,8 +12,6 @@ #include "lock_dlm.h" -extern struct lm_lockops gdlm_ops; - static ssize_t proto_name_show(struct gdlm_ls *ls, char *buf) { return sprintf(buf, "%s\n", gdlm_ops.lm_proto_name); -- cgit v1.2.3 From 7afd88d9166a752b52517648bcbe923e05d393fc Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Fri, 22 Feb 2008 16:07:18 +0000 Subject: [GFS2] Fix a page lock / glock deadlock We've previously been using a "try lock" in readpage on the basis that it would prevent deadlocks due to the inverted lock ordering (our normal lock ordering is glock first and then page lock). Unfortunately tests have shown that this isn't enough. If the glock has a demote request queued such that run_queue() in the glock code tries to do a demote when its called under readpage then it will try and write out all the dirty pages which requires locking them. This then deadlocks with the page locked by readpage. The solution is to always require two calls into readpage. The first unlocks the page, gets the glock and returns AOP_TRUNCATED_PAGE, the second does the actual readpage and unlocks the glock & page as required. Signed-off-by: Steven Whitehouse --- fs/gfs2/glock.h | 13 ++++++------- fs/gfs2/inode.c | 2 +- fs/gfs2/ops_address.c | 27 +++++++++++++++------------ fs/gfs2/ops_dentry.c | 4 ++-- fs/gfs2/ops_inode.c | 4 ++-- 5 files changed, 26 insertions(+), 24 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/glock.h b/fs/gfs2/glock.h index ace5770760c..cdad3e6f815 100644 --- a/fs/gfs2/glock.h +++ b/fs/gfs2/glock.h @@ -32,24 +32,23 @@ #define GLR_TRYFAILED 13 #define GLR_CANCELED 14 -static inline int gfs2_glock_is_locked_by_me(struct gfs2_glock *gl) +static inline struct gfs2_holder *gfs2_glock_is_locked_by_me(struct gfs2_glock *gl) { struct gfs2_holder *gh; - int locked = 0; struct pid *pid; /* Look in glock's list of holders for one with current task as owner */ spin_lock(&gl->gl_spin); pid = task_pid(current); list_for_each_entry(gh, &gl->gl_holders, gh_list) { - if (gh->gh_owner_pid == pid) { - locked = 1; - break; - } + if (gh->gh_owner_pid == pid) + goto out; } + gh = NULL; +out: spin_unlock(&gl->gl_spin); - return locked; + return gh; } static inline int gfs2_glock_is_held_excl(struct gfs2_glock *gl) diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c index 5f50dd53bf6..810ff023fb1 100644 --- a/fs/gfs2/inode.c +++ b/fs/gfs2/inode.c @@ -493,7 +493,7 @@ struct inode *gfs2_lookupi(struct inode *dir, const struct qstr *name, return dir; } - if (gfs2_glock_is_locked_by_me(dip->i_gl) == 0) { + if (gfs2_glock_is_locked_by_me(dip->i_gl) == NULL) { error = gfs2_glock_nq_init(dip->i_gl, LM_ST_SHARED, 0, &d_gh); if (error) return ERR_PTR(error); diff --git a/fs/gfs2/ops_address.c b/fs/gfs2/ops_address.c index 7523999afc5..fbb4a6aa158 100644 --- a/fs/gfs2/ops_address.c +++ b/fs/gfs2/ops_address.c @@ -508,23 +508,26 @@ static int __gfs2_readpage(void *file, struct page *page) static int gfs2_readpage(struct file *file, struct page *page) { struct gfs2_inode *ip = GFS2_I(page->mapping->host); - struct gfs2_holder gh; + struct gfs2_holder *gh; int error; - gfs2_holder_init(ip->i_gl, LM_ST_SHARED, GL_ATIME|LM_FLAG_TRY_1CB, &gh); - error = gfs2_glock_nq_atime(&gh); - if (unlikely(error)) { + gh = gfs2_glock_is_locked_by_me(ip->i_gl); + if (!gh) { + gh = kmalloc(sizeof(struct gfs2_holder), GFP_NOFS); + if (!gh) + return -ENOBUFS; + gfs2_holder_init(ip->i_gl, LM_ST_SHARED, GL_ATIME, gh); unlock_page(page); - goto out; + error = gfs2_glock_nq_atime(gh); + if (likely(error != 0)) + goto out; + return AOP_TRUNCATED_PAGE; } error = __gfs2_readpage(file, page); - gfs2_glock_dq(&gh); + gfs2_glock_dq(gh); out: - gfs2_holder_uninit(&gh); - if (error == GLR_TRYFAILED) { - yield(); - return AOP_TRUNCATED_PAGE; - } + gfs2_holder_uninit(gh); + kfree(gh); return error; } @@ -826,7 +829,7 @@ static int gfs2_write_end(struct file *file, struct address_space *mapping, unsigned int to = from + len; int ret; - BUG_ON(gfs2_glock_is_locked_by_me(ip->i_gl) == 0); + BUG_ON(gfs2_glock_is_locked_by_me(ip->i_gl) == NULL); ret = gfs2_meta_inode_buffer(ip, &dibh); if (unlikely(ret)) { diff --git a/fs/gfs2/ops_dentry.c b/fs/gfs2/ops_dentry.c index 793e334d098..4a5e676b442 100644 --- a/fs/gfs2/ops_dentry.c +++ b/fs/gfs2/ops_dentry.c @@ -43,7 +43,7 @@ static int gfs2_drevalidate(struct dentry *dentry, struct nameidata *nd) struct gfs2_holder d_gh; struct gfs2_inode *ip = NULL; int error; - int had_lock=0; + int had_lock = 0; if (inode) { if (is_bad_inode(inode)) @@ -54,7 +54,7 @@ static int gfs2_drevalidate(struct dentry *dentry, struct nameidata *nd) if (sdp->sd_args.ar_localcaching) goto valid; - had_lock = gfs2_glock_is_locked_by_me(dip->i_gl); + had_lock = (gfs2_glock_is_locked_by_me(dip->i_gl) != NULL); if (!had_lock) { error = gfs2_glock_nq_init(dip->i_gl, LM_ST_SHARED, 0, &d_gh); if (error) diff --git a/fs/gfs2/ops_inode.c b/fs/gfs2/ops_inode.c index 301c9459667..af7097a514c 100644 --- a/fs/gfs2/ops_inode.c +++ b/fs/gfs2/ops_inode.c @@ -898,7 +898,7 @@ static int gfs2_permission(struct inode *inode, int mask, struct nameidata *nd) int error; int unlock = 0; - if (gfs2_glock_is_locked_by_me(ip->i_gl) == 0) { + if (gfs2_glock_is_locked_by_me(ip->i_gl) == NULL) { error = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, LM_FLAG_ANY, &i_gh); if (error) return error; @@ -1065,7 +1065,7 @@ static int gfs2_getattr(struct vfsmount *mnt, struct dentry *dentry, int error; int unlock = 0; - if (gfs2_glock_is_locked_by_me(ip->i_gl) == 0) { + if (gfs2_glock_is_locked_by_me(ip->i_gl) == NULL) { error = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, LM_FLAG_ANY, &gh); if (error) return error; -- cgit v1.2.3 From 9b8c81d1de49943ec69d157234b8981008c30d31 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Fri, 22 Feb 2008 16:09:31 +0000 Subject: [GFS2] Allow bmap to allocate extents We've supported mapping of extents when no block allocation is required for some time. This patch extends that to mapping of extents when an allocation has been requested. In that case we try to allocate as many blocks as are requested, but we might return fewer in case there is something preventing us from returning the complete amount (e.g. an already allocated block is in the way). Currently the only code path which can actually request multiple data blocks in a single bmap call is the page_mkwrite path and even then it only happens if there are multiple blocks per page. What this patch does do however, is merge the allocation requests for metadata (growing the metadata tree in either height or depth) with the allocation of the data blocks in the case that both are needed. This results in lower overheads even in the single block allocation case. The one thing which we can't handle here at the moment is unstuffing. I would like to be able to do that, but the problem which arises is that in order to unstuff one has to get a locked page from the page cache which results in locking problems in the (usual) case that the caller is holding the page lock on the page it wishes to map. So that case will have to be addressed in future patches. Signed-off-by: Steven Whitehouse --- fs/gfs2/bmap.c | 502 +++++++++++++++++++++++++++++++++------------------------ fs/gfs2/dir.c | 2 +- fs/gfs2/rgrp.c | 10 +- 3 files changed, 295 insertions(+), 219 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index f1f38ca77a5..c1ee6355ced 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -175,74 +175,13 @@ out: return error; } -/** - * build_height - Build a metadata tree of the requested height - * @ip: The GFS2 inode - * @height: The height to build to - * - * - * Returns: errno - */ - -static int build_height(struct inode *inode, struct metapath *mp, unsigned height) -{ - struct gfs2_inode *ip = GFS2_I(inode); - unsigned new_height = height - ip->i_height; - struct buffer_head *dibh = mp->mp_bh[0]; - struct gfs2_dinode *di; - __be64 *bp; - u64 bn; - unsigned n, i = 0; - - BUG_ON(height <= ip->i_height); - - do { - n = new_height - i; - bn = gfs2_alloc_block(ip, &n); - gfs2_trans_add_unrevoke(GFS2_SB(inode), bn, n); - do { - mp->mp_bh[i + 1] = gfs2_meta_new(ip->i_gl, bn++); - gfs2_trans_add_bh(ip->i_gl, mp->mp_bh[i + 1], 1); - i++; - } while(i < n); - } while(i < new_height); - - n = 0; - bn = mp->mp_bh[1]->b_blocknr; - if (new_height > 1) { - for(; n < new_height-1; n++) { - gfs2_metatype_set(mp->mp_bh[n + 1], GFS2_METATYPE_IN, - GFS2_FORMAT_IN); - gfs2_buffer_clear_tail(mp->mp_bh[n + 1], - sizeof(struct gfs2_meta_header)); - bp = (__be64 *)(mp->mp_bh[n + 1]->b_data + - sizeof(struct gfs2_meta_header)); - *bp = cpu_to_be64(mp->mp_bh[n+2]->b_blocknr); - brelse(mp->mp_bh[n+1]); - mp->mp_bh[n+1] = NULL; - } - } - gfs2_metatype_set(mp->mp_bh[n+1], GFS2_METATYPE_IN, GFS2_FORMAT_IN); - gfs2_buffer_copy_tail(mp->mp_bh[n+1], sizeof(struct gfs2_meta_header), - dibh, sizeof(struct gfs2_dinode)); - brelse(mp->mp_bh[n+1]); - mp->mp_bh[n+1] = NULL; - gfs2_trans_add_bh(ip->i_gl, dibh, 1); - di = (struct gfs2_dinode *)dibh->b_data; - gfs2_buffer_clear_tail(dibh, sizeof(struct gfs2_dinode)); - *(__be64 *)(di + 1) = cpu_to_be64(bn); - ip->i_height += new_height; - gfs2_add_inode_blocks(&ip->i_inode, new_height); - di->di_height = cpu_to_be16(ip->i_height); - di->di_blocks = cpu_to_be64(gfs2_get_inode_blocks(&ip->i_inode)); - return 0; -} /** * find_metapath - Find path through the metadata tree - * @ip: The inode pointer + * @sdp: The superblock * @mp: The metapath to return the result in * @block: The disk block to look up + * @height: The pre-calculated height of the metadata tree * * This routine returns a struct metapath structure that defines a path * through the metadata of inode "ip" to get to block "block". @@ -297,17 +236,27 @@ static int build_height(struct inode *inode, struct metapath *mp, unsigned heigh * */ -static void find_metapath(struct gfs2_inode *ip, u64 block, - struct metapath *mp) +static void find_metapath(const struct gfs2_sbd *sdp, u64 block, + struct metapath *mp, unsigned int height) { - struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); unsigned int i; - for (i = ip->i_height; i--;) + for (i = height; i--;) mp->mp_list[i] = do_div(block, sdp->sd_inptrs); } +static inline unsigned int zero_metapath_length(const struct metapath *mp, + unsigned height) +{ + unsigned int i; + for (i = 0; i < height - 1; i++) { + if (mp->mp_list[i] != 0) + return i; + } + return height; +} + /** * metapointer - Return pointer to start of metadata in a buffer * @height: The metadata height (0 = dinode) @@ -318,95 +267,62 @@ static void find_metapath(struct gfs2_inode *ip, u64 block, * metadata tree. */ -static inline __be64 *metapointer(int *boundary, unsigned int height, - const struct metapath *mp) +static inline __be64 *metapointer(unsigned int height, const struct metapath *mp) { struct buffer_head *bh = mp->mp_bh[height]; unsigned int head_size = (height > 0) ? sizeof(struct gfs2_meta_header) : sizeof(struct gfs2_dinode); - __be64 *ptr; - *boundary = 0; - ptr = ((__be64 *)(bh->b_data + head_size)) + mp->mp_list[height]; - if (ptr + 1 == (__be64 *)(bh->b_data + bh->b_size)) - *boundary = 1; - return ptr; + return ((__be64 *)(bh->b_data + head_size)) + mp->mp_list[height]; } /** - * lookup_block - Get the next metadata block in metadata tree - * @ip: The GFS2 inode - * @height: The height of the tree (0 = dinode) + * lookup_metapath - Walk the metadata tree to a specific point + * @ip: The inode * @mp: The metapath - * @create: Non-zero if we may create a new meatdata block - * @new: Used to indicate if we did create a new metadata block - * @block: the returned disk block number * - * Given a metatree, complete to a particular height, checks to see if the next - * height of the tree exists. If not the next height of the tree is created. - * The block number of the next height of the metadata tree is returned. + * Assumes that the inode's buffer has already been looked up and + * hooked onto mp->mp_bh[0] and that the metapath has been initialised + * by find_metapath(). * + * If this function encounters part of the tree which has not been + * allocated, it returns the current height of the tree at the point + * at which it found the unallocated block. Blocks which are found are + * added to the mp->mp_bh[] list. + * + * Returns: error or height of metadata tree */ -static int lookup_block(struct gfs2_inode *ip, unsigned int height, - struct metapath *mp, int create, - int *new, u64 *block) +static int lookup_metapath(struct gfs2_inode *ip, struct metapath *mp) { - int boundary; - __be64 *ptr = metapointer(&boundary, height, mp); - unsigned int n = 1; - - if (*ptr) { - *block = be64_to_cpu(*ptr); - return boundary; - } - - *block = 0; - - if (!create) - return 0; - - *block = gfs2_alloc_block(ip, &n); - if (height != ip->i_height - 1 || gfs2_is_dir(ip)) - gfs2_trans_add_unrevoke(GFS2_SB(&ip->i_inode), *block, 1); - - gfs2_trans_add_bh(ip->i_gl, mp->mp_bh[height], 1); - - *ptr = cpu_to_be64(*block); - gfs2_add_inode_blocks(&ip->i_inode, 1); - - *new = 1; - return 0; -} - -static int lookup_metapath(struct inode *inode, struct metapath *mp, - int create, int *new, u64 *dblock) -{ - struct buffer_head *bh; - struct gfs2_inode *ip = GFS2_I(inode); unsigned int end_of_metadata = ip->i_height - 1; unsigned int x; + __be64 *ptr; + u64 dblock; int ret; for (x = 0; x < end_of_metadata; x++) { - lookup_block(ip, x, mp, create, new, dblock); - if (!*dblock) - return 0; + ptr = metapointer(x, mp); + dblock = be64_to_cpu(*ptr); + if (!dblock) + return x + 1; - ret = gfs2_meta_indirect_buffer(ip, x+1, *dblock, *new, &mp->mp_bh[x+1]); + ret = gfs2_meta_indirect_buffer(ip, x+1, dblock, 0, &mp->mp_bh[x+1]); if (ret) return ret; } - return lookup_block(ip, end_of_metadata, mp, create, new, dblock); + return ip->i_height; } -static void release_metapath(struct metapath *mp) +static inline void release_metapath(struct metapath *mp) { int i; - for (i = 0; i < GFS2_MAX_META_HEIGHT; i++) - if (mp->mp_bh[i]) - brelse(mp->mp_bh[i]); + for (i = 0; i < GFS2_MAX_META_HEIGHT; i++) { + if (mp->mp_bh[i] == NULL) + break; + brelse(mp->mp_bh[i]); + } } /** @@ -445,32 +361,208 @@ static inline unsigned int gfs2_extent_length(void *start, unsigned int len, __b return (ptr - first); } -static inline void bmap_lock(struct inode *inode, int create) +static inline void bmap_lock(struct gfs2_inode *ip, int create) { - struct gfs2_inode *ip = GFS2_I(inode); if (create) down_write(&ip->i_rw_mutex); else down_read(&ip->i_rw_mutex); } -static inline void bmap_unlock(struct inode *inode, int create) +static inline void bmap_unlock(struct gfs2_inode *ip, int create) { - struct gfs2_inode *ip = GFS2_I(inode); if (create) up_write(&ip->i_rw_mutex); else up_read(&ip->i_rw_mutex); } +static inline __be64 *gfs2_indirect_init(struct metapath *mp, + struct gfs2_glock *gl, unsigned int i, + unsigned offset, u64 bn) +{ + __be64 *ptr = (__be64 *)(mp->mp_bh[i - 1]->b_data + + ((i > 1) ? sizeof(struct gfs2_meta_header) : + sizeof(struct gfs2_dinode))); + BUG_ON(i < 1); + BUG_ON(mp->mp_bh[i] != NULL); + mp->mp_bh[i] = gfs2_meta_new(gl, bn); + gfs2_trans_add_bh(gl, mp->mp_bh[i], 1); + gfs2_metatype_set(mp->mp_bh[i], GFS2_METATYPE_IN, GFS2_FORMAT_IN); + gfs2_buffer_clear_tail(mp->mp_bh[i], sizeof(struct gfs2_meta_header)); + ptr += offset; + *ptr = cpu_to_be64(bn); + return ptr; +} + +enum alloc_state { + ALLOC_DATA = 0, + ALLOC_GROW_DEPTH = 1, + ALLOC_GROW_HEIGHT = 2, + /* ALLOC_UNSTUFF = 3, TBD and rather complicated */ +}; + +/** + * gfs2_bmap_alloc - Build a metadata tree of the requested height + * @inode: The GFS2 inode + * @lblock: The logical starting block of the extent + * @bh_map: This is used to return the mapping details + * @mp: The metapath + * @sheight: The starting height (i.e. whats already mapped) + * @height: The height to build to + * @maxlen: The max number of data blocks to alloc + * + * In this routine we may have to alloc: + * i) Indirect blocks to grow the metadata tree height + * ii) Indirect blocks to fill in lower part of the metadata tree + * iii) Data blocks + * + * The function is in two parts. The first part works out the total + * number of blocks which we need. The second part does the actual + * allocation asking for an extent at a time (if enough contiguous free + * blocks are available, there will only be one request per bmap call) + * and uses the state machine to initialise the blocks in order. + * + * Returns: errno on error + */ + +static int gfs2_bmap_alloc(struct inode *inode, const sector_t lblock, + struct buffer_head *bh_map, struct metapath *mp, + const unsigned int sheight, + const unsigned int height, + const unsigned int maxlen) +{ + struct gfs2_inode *ip = GFS2_I(inode); + struct gfs2_sbd *sdp = GFS2_SB(inode); + struct buffer_head *dibh = mp->mp_bh[0]; + u64 bn, dblock = 0; + unsigned n, i, blks, alloced = 0, iblks = 0, zmpl = 0; + unsigned dblks = 0; + unsigned ptrs_per_blk; + const unsigned end_of_metadata = height - 1; + int eob = 0; + enum alloc_state state; + __be64 *ptr; + __be64 zero_bn = 0; + + BUG_ON(sheight < 1); + BUG_ON(dibh == NULL); + + gfs2_trans_add_bh(ip->i_gl, dibh, 1); + + if (height == sheight) { + struct buffer_head *bh; + /* Bottom indirect block exists, find unalloced extent size */ + ptr = metapointer(end_of_metadata, mp); + bh = mp->mp_bh[end_of_metadata]; + dblks = gfs2_extent_length(bh->b_data, bh->b_size, ptr, maxlen, + &eob); + BUG_ON(dblks < 1); + state = ALLOC_DATA; + } else { + /* Need to allocate indirect blocks */ + ptrs_per_blk = height > 1 ? sdp->sd_inptrs : sdp->sd_diptrs; + dblks = min(maxlen, ptrs_per_blk - mp->mp_list[end_of_metadata]); + if (height == ip->i_height) { + /* Writing into existing tree, extend tree down */ + iblks = height - sheight; + state = ALLOC_GROW_DEPTH; + } else { + /* Building up tree height */ + state = ALLOC_GROW_HEIGHT; + iblks = height - ip->i_height; + zmpl = zero_metapath_length(mp, height); + iblks -= zmpl; + iblks += height; + } + } + + /* start of the second part of the function (state machine) */ + + blks = dblks + iblks; + i = sheight; + do { + n = blks - alloced; + bn = gfs2_alloc_block(ip, &n); + alloced += n; + if (state != ALLOC_DATA || gfs2_is_jdata(ip)) + gfs2_trans_add_unrevoke(sdp, bn, n); + switch (state) { + /* Growing height of tree */ + case ALLOC_GROW_HEIGHT: + if (i == 1) { + ptr = (__be64 *)(dibh->b_data + + sizeof(struct gfs2_dinode)); + zero_bn = *ptr; + } + for (; i - 1 < height - ip->i_height && n > 0; i++, n--) + gfs2_indirect_init(mp, ip->i_gl, i, 0, bn++); + if (i - 1 == height - ip->i_height) { + i--; + gfs2_buffer_copy_tail(mp->mp_bh[i], + sizeof(struct gfs2_meta_header), + dibh, sizeof(struct gfs2_dinode)); + gfs2_buffer_clear_tail(dibh, + sizeof(struct gfs2_dinode) + + sizeof(__be64)); + ptr = (__be64 *)(mp->mp_bh[i]->b_data + + sizeof(struct gfs2_meta_header)); + *ptr = zero_bn; + state = ALLOC_GROW_DEPTH; + for(i = zmpl; i < height; i++) { + if (mp->mp_bh[i] == NULL) + break; + brelse(mp->mp_bh[i]); + mp->mp_bh[i] = NULL; + } + i = zmpl; + } + if (n == 0) + break; + /* Branching from existing tree */ + case ALLOC_GROW_DEPTH: + if (i > 1 && i < height) + gfs2_trans_add_bh(ip->i_gl, mp->mp_bh[i-1], 1); + for (; i < height && n > 0; i++, n--) + gfs2_indirect_init(mp, ip->i_gl, i, + mp->mp_list[i-1], bn++); + if (i == height) + state = ALLOC_DATA; + if (n == 0) + break; + /* Tree complete, adding data blocks */ + case ALLOC_DATA: + BUG_ON(n > dblks); + BUG_ON(mp->mp_bh[end_of_metadata] == NULL); + gfs2_trans_add_bh(ip->i_gl, mp->mp_bh[end_of_metadata], 1); + dblks = n; + ptr = metapointer(end_of_metadata, mp); + dblock = bn; + while (n-- > 0) + *ptr++ = cpu_to_be64(bn++); + break; + } + } while (state != ALLOC_DATA); + + ip->i_height = height; + gfs2_add_inode_blocks(&ip->i_inode, alloced); + gfs2_dinode_out(ip, mp->mp_bh[0]->b_data); + map_bh(bh_map, inode->i_sb, dblock); + bh_map->b_size = dblks << inode->i_blkbits; + set_buffer_new(bh_map); + return 0; +} + /** * gfs2_block_map - Map a block from an inode to a disk block * @inode: The inode * @lblock: The logical block number * @bh_map: The bh to be mapped + * @create: True if its ok to alloc blocks to satify the request * - * Find the block number on the current device which corresponds to an - * inode's block. If the block had to be created, "new" will be set. + * Sets buffer_mapped() if successful, sets buffer_boundary() if a + * read of metadata will be required before the next block can be + * mapped. Sets buffer_new() if new blocks were allocated. * * Returns: errno */ @@ -481,21 +573,21 @@ int gfs2_block_map(struct inode *inode, sector_t lblock, struct gfs2_inode *ip = GFS2_I(inode); struct gfs2_sbd *sdp = GFS2_SB(inode); unsigned int bsize = sdp->sd_sb.sb_bsize; - int error = 0; - int new = 0; - u64 dblock = 0; - int boundary; - unsigned int maxlen = bh_map->b_size >> inode->i_blkbits; - struct metapath mp; - u64 size; + const unsigned int maxlen = bh_map->b_size >> inode->i_blkbits; const u64 *arr = sdp->sd_heightsize; - BUG_ON(maxlen == 0); + __be64 *ptr; + u64 size; + struct metapath mp; + int ret; + int eob; + unsigned int len; + struct buffer_head *bh; + u8 height; - if (gfs2_assert_warn(sdp, !gfs2_is_stuffed(ip))) - return 0; + BUG_ON(maxlen == 0); memset(mp.mp_bh, 0, sizeof(mp.mp_bh)); - bmap_lock(inode, create); + bmap_lock(ip, create); clear_buffer_mapped(bh_map); clear_buffer_new(bh_map); clear_buffer_boundary(bh_map); @@ -503,56 +595,50 @@ int gfs2_block_map(struct inode *inode, sector_t lblock, bsize = sdp->sd_jbsize; arr = sdp->sd_jheightsize; } - size = (lblock + 1) * bsize; - error = gfs2_meta_inode_buffer(ip, &mp.mp_bh[0]); - if (error) - goto out_fail; - - if (size > arr[ip->i_height]) { - u8 height = ip->i_height; - if (!create) - goto out_ok; - while (size > arr[height]) - height++; - error = build_height(inode, &mp, height); - if (error) - goto out_fail; - } + ret = gfs2_meta_inode_buffer(ip, &mp.mp_bh[0]); + if (ret) + goto out; - find_metapath(ip, lblock, &mp); - error = lookup_metapath(inode, &mp, create, &new, &dblock); - if (error < 0) - goto out_fail; - boundary = error; - - if (new) { - map_bh(bh_map, inode->i_sb, dblock); - if (boundary) - set_buffer_boundary(bh_map); - gfs2_trans_add_bh(ip->i_gl, mp.mp_bh[0], 1); - gfs2_dinode_out(ip, mp.mp_bh[0]->b_data); - set_buffer_new(bh_map); - goto out_ok; - } + height = ip->i_height; + size = (lblock + 1) * bsize; + while (size > arr[height]) + height++; + find_metapath(sdp, lblock, &mp, height); + ret = 1; + if (height > ip->i_height || gfs2_is_stuffed(ip)) + goto do_alloc; + ret = lookup_metapath(ip, &mp); + if (ret < 0) + goto out; + if (ret != ip->i_height) + goto do_alloc; + ptr = metapointer(ip->i_height - 1, &mp); + if (*ptr == 0) + goto do_alloc; + map_bh(bh_map, inode->i_sb, be64_to_cpu(*ptr)); + bh = mp.mp_bh[ip->i_height - 1]; + len = gfs2_extent_length(bh->b_data, bh->b_size, ptr, maxlen, &eob); + bh_map->b_size = (len << inode->i_blkbits); + if (eob) + set_buffer_boundary(bh_map); + ret = 0; +out: + release_metapath(&mp); + bmap_unlock(ip, create); + return ret; - if (dblock) { - unsigned int len; - struct buffer_head *bh = mp.mp_bh[ip->i_height - 1]; - __be64 *ptr = metapointer(&boundary, ip->i_height - 1, &mp); - map_bh(bh_map, inode->i_sb, dblock); - len = gfs2_extent_length(bh->b_data, bh->b_size, ptr, maxlen, - &boundary); - bh_map->b_size = (len << inode->i_blkbits); - if (boundary) - set_buffer_boundary(bh_map); +do_alloc: + /* All allocations are done here, firstly check create flag */ + if (!create) { + BUG_ON(gfs2_is_stuffed(ip)); + ret = 0; + goto out; } -out_ok: - error = 0; -out_fail: - release_metapath(&mp); - bmap_unlock(inode, create); - return error; + + /* At this point ret is the tree depth of already allocated blocks */ + ret = gfs2_bmap_alloc(inode, lblock, bh_map, &mp, ret, height, maxlen); + goto out; } /* @@ -568,7 +654,7 @@ int gfs2_extent_map(struct inode *inode, u64 lblock, int *new, u64 *dblock, unsi BUG_ON(!dblock); BUG_ON(!new); - bh.b_size = 1 << (inode->i_blkbits + 5); + bh.b_size = 1 << (inode->i_blkbits + (create ? 0 : 5)); ret = gfs2_block_map(inode, lblock, &bh, create); *extlen = bh.b_size >> inode->i_blkbits; *dblock = bh.b_blocknr; @@ -835,38 +921,25 @@ static int do_grow(struct gfs2_inode *ip, u64 size) if (error) goto out_ipres; + error = gfs2_meta_inode_buffer(ip, &dibh); + if (error) + goto out_end_trans; + if (size > sdp->sd_sb.sb_bsize - sizeof(struct gfs2_dinode)) { - const u64 *arr = sdp->sd_heightsize; if (gfs2_is_stuffed(ip)) { error = gfs2_unstuff_dinode(ip, NULL); if (error) - goto out_end_trans; - } - - down_write(&ip->i_rw_mutex); - if (size > arr[ip->i_height]) { - struct metapath mp; - u8 height = ip->i_height; - while(size > arr[height]) - height++; - error = build_height(&ip->i_inode, &mp, height); + goto out_brelse; } - up_write(&ip->i_rw_mutex); - if (error) - goto out_end_trans; } ip->i_di.di_size = size; ip->i_inode.i_mtime = ip->i_inode.i_ctime = CURRENT_TIME; - - error = gfs2_meta_inode_buffer(ip, &dibh); - if (error) - goto out_end_trans; - gfs2_trans_add_bh(ip->i_gl, dibh, 1); gfs2_dinode_out(ip, dibh->b_data); - brelse(dibh); +out_brelse: + brelse(dibh); out_end_trans: gfs2_trans_end(sdp); out_ipres: @@ -996,6 +1069,7 @@ out: static int trunc_dealloc(struct gfs2_inode *ip, u64 size) { + struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); unsigned int height = ip->i_height; u64 lblock; struct metapath mp; @@ -1004,9 +1078,9 @@ static int trunc_dealloc(struct gfs2_inode *ip, u64 size) if (!size) lblock = 0; else - lblock = (size - 1) >> GFS2_SB(&ip->i_inode)->sd_sb.sb_bsize_shift; + lblock = (size - 1) >> sdp->sd_sb.sb_bsize_shift; - find_metapath(ip, lblock, &mp); + find_metapath(sdp, lblock, &mp, ip->i_height); gfs2_alloc_get(ip); error = gfs2_quota_hold(ip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE); diff --git a/fs/gfs2/dir.c b/fs/gfs2/dir.c index 34dc8dfaba1..a3753c7989f 100644 --- a/fs/gfs2/dir.c +++ b/fs/gfs2/dir.c @@ -159,6 +159,7 @@ static int gfs2_dir_write_data(struct gfs2_inode *ip, const char *buf, unsigned int o; int copied = 0; int error = 0; + int new = 0; if (!size) return 0; @@ -183,7 +184,6 @@ static int gfs2_dir_write_data(struct gfs2_inode *ip, const char *buf, while (copied < size) { unsigned int amount; struct buffer_head *bh; - int new = 0; amount = size - copied; if (amount > sdp->sd_sb.sb_bsize - o) diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index 77eba0a3804..4291375cecc 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -1357,16 +1357,18 @@ static u32 rgblk_search(struct gfs2_rgrpd *rgd, u32 goal, gfs2_trans_add_bh(rgd->rd_gl, bi->bi_bh, 1); gfs2_setbit(rgd, bi->bi_bh->b_data, bi->bi_clone, bi->bi_offset, bi->bi_len, blk, new_state); - while(*n < elen) { + goal = blk; + while (*n < elen) { goal++; - if (goal >= (bi->bi_len / GFS2_NBBY)) + if (goal >= (bi->bi_len * GFS2_NBBY)) break; if (gfs2_testbit(rgd, buffer, bi->bi_len, goal) != GFS2_BLKST_FREE) break; - (*n)++; gfs2_setbit(rgd, bi->bi_bh->b_data, bi->bi_clone, - bi->bi_offset, bi->bi_len, blk, new_state); + bi->bi_offset, bi->bi_len, goal, + new_state); + (*n)++; } } -- cgit v1.2.3 From 7dc2cf1c8ffbd471722f1aa479bc68d4df1c9edc Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Mon, 25 Feb 2008 18:58:36 +0100 Subject: [GFS2] fix file_system_type leak on gfs2meta mount get_gfs2_sb does a get_fs_type without doing a put_filesystem and thus leaking a file_system_type reference everytime it's called. Just use gfs2_fs_type directly instead of doing the lookup and thus fix the problem. Signed-off-by: Christoph Hellwig Signed-off-by: Steven Whitehouse --- fs/gfs2/ops_fstype.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c index c4b7a210c0c..63d5fd203d3 100644 --- a/fs/gfs2/ops_fstype.c +++ b/fs/gfs2/ops_fstype.c @@ -943,7 +943,6 @@ static struct super_block* get_gfs2_sb(const char *dev_name) { struct kstat stat; struct nameidata nd; - struct file_system_type *fstype; struct super_block *sb = NULL, *s; int error; @@ -955,8 +954,7 @@ static struct super_block* get_gfs2_sb(const char *dev_name) } error = vfs_getattr(nd.path.mnt, nd.path.dentry, &stat); - fstype = get_fs_type("gfs2"); - list_for_each_entry(s, &fstype->fs_supers, s_instances) { + list_for_each_entry(s, &gfs2_fs_type.fs_supers, s_instances) { if ((S_ISBLK(stat.mode) && s->s_dev == stat.rdev) || (S_ISDIR(stat.mode) && s == nd.path.dentry->d_inode->i_sb)) { -- cgit v1.2.3 From d83225d45d2b76175279abb2a3d7ee325a09aba8 Mon Sep 17 00:00:00 2001 From: Denis Cheng Date: Tue, 26 Feb 2008 15:25:03 +0800 Subject: [GFS2] remove gfs2_dev_iops struct inode_operations gfs2_dev_iops is always the same as gfs2_file_iops, since Jan 2006, when GFS2 merged into mainstream kernel. So one of them could be removed. Signed-off-by: Denis Cheng Signed-off-by: Steven Whitehouse --- fs/gfs2/inode.c | 2 +- fs/gfs2/ops_inode.c | 10 ---------- fs/gfs2/ops_inode.h | 1 - 3 files changed, 1 insertion(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c index 810ff023fb1..7b9f31e5d6d 100644 --- a/fs/gfs2/inode.c +++ b/fs/gfs2/inode.c @@ -149,7 +149,7 @@ void gfs2_set_iop(struct inode *inode) } else if (S_ISLNK(mode)) { inode->i_op = &gfs2_symlink_iops; } else { - inode->i_op = &gfs2_dev_iops; + inode->i_op = &gfs2_file_iops; } unlock_new_inode(inode); diff --git a/fs/gfs2/ops_inode.c b/fs/gfs2/ops_inode.c index af7097a514c..6cbbb5b46ce 100644 --- a/fs/gfs2/ops_inode.c +++ b/fs/gfs2/ops_inode.c @@ -1149,16 +1149,6 @@ const struct inode_operations gfs2_file_iops = { .removexattr = gfs2_removexattr, }; -const struct inode_operations gfs2_dev_iops = { - .permission = gfs2_permission, - .setattr = gfs2_setattr, - .getattr = gfs2_getattr, - .setxattr = gfs2_setxattr, - .getxattr = gfs2_getxattr, - .listxattr = gfs2_listxattr, - .removexattr = gfs2_removexattr, -}; - const struct inode_operations gfs2_dir_iops = { .create = gfs2_create, .lookup = gfs2_lookup, diff --git a/fs/gfs2/ops_inode.h b/fs/gfs2/ops_inode.h index fd8cee231e1..14b4b797622 100644 --- a/fs/gfs2/ops_inode.h +++ b/fs/gfs2/ops_inode.h @@ -15,7 +15,6 @@ extern const struct inode_operations gfs2_file_iops; extern const struct inode_operations gfs2_dir_iops; extern const struct inode_operations gfs2_symlink_iops; -extern const struct inode_operations gfs2_dev_iops; extern const struct file_operations gfs2_file_fops; extern const struct file_operations gfs2_dir_fops; extern const struct file_operations gfs2_file_fops_nolock; -- cgit v1.2.3 From 43a33c53cc9131a537522ab9736c6e4c03ddf57a Mon Sep 17 00:00:00 2001 From: Denis Cheng Date: Tue, 26 Feb 2008 15:25:04 +0800 Subject: [GFS2] re-support special inode a previous commit removed call to init_special_inode from inode lookuping, this cause problems as: # mknod /mnt/gfs2/dev/null c 1 3 # cat /mnt/gfs2/dev/null cat: /mnt/gfs2/dev/null: Invalid argument without special inode, GFS2 cannot support char device file, block device file, fifo pipe, and socket file, lose many important features as a common file system. this one line patch re add special inode support. Signed-off-by: Denis Cheng Signed-off-by: Steven Whitehouse --- fs/gfs2/inode.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c index 7b9f31e5d6d..92ea9afacb1 100644 --- a/fs/gfs2/inode.c +++ b/fs/gfs2/inode.c @@ -150,6 +150,7 @@ void gfs2_set_iop(struct inode *inode) inode->i_op = &gfs2_symlink_iops; } else { inode->i_op = &gfs2_file_iops; + init_special_inode(inode, inode->i_mode, inode->i_rdev); } unlock_new_inode(inode); -- cgit v1.2.3 From 105284970ba7d0d0ff4b97e57728eac7adf6a42a Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Wed, 27 Feb 2008 17:56:27 +0000 Subject: [GFS2] Need to ensure that sector_t is 64bits for GFS2 We need to ensure that sector_t is 64bits for GFS2, so that we need to depend on LBD as well as LSF. Signed-off-by: Steven Whitehouse --- fs/gfs2/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/gfs2/Kconfig b/fs/gfs2/Kconfig index d147b53ef52..7f7947e3dfb 100644 --- a/fs/gfs2/Kconfig +++ b/fs/gfs2/Kconfig @@ -1,6 +1,6 @@ config GFS2_FS tristate "GFS2 file system support" - depends on EXPERIMENTAL && (64BIT || LSF) + depends on EXPERIMENTAL && (64BIT || (LSF && LBD)) select FS_POSIX_ACL select CRC32 help -- cgit v1.2.3 From 182fe5abd8ebbb3a00c1be91f44e4783e139918c Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Mon, 3 Mar 2008 21:54:21 +0300 Subject: [GFS2] possible null pointer dereference fixup gfs2_alloc_get may fail so we have to check it to prevent NULL pointer dereference. Signed-off-by: Cyrill Gorcunov Signed-off-by: Steven Whitehouse --- fs/gfs2/bmap.c | 5 ++++- fs/gfs2/dir.c | 10 +++++++--- fs/gfs2/eattr.c | 6 ++++++ fs/gfs2/inode.c | 7 ++++++- fs/gfs2/ops_address.c | 4 ++++ fs/gfs2/ops_inode.c | 11 ++++++++++- fs/gfs2/quota.c | 9 +++++++-- 7 files changed, 44 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index c1ee6355ced..f7093aa69aa 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -900,6 +900,8 @@ static int do_grow(struct gfs2_inode *ip, u64 size) int error; al = gfs2_alloc_get(ip); + if (!al) + return -ENOMEM; error = gfs2_quota_lock(ip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE); if (error) @@ -1081,7 +1083,8 @@ static int trunc_dealloc(struct gfs2_inode *ip, u64 size) lblock = (size - 1) >> sdp->sd_sb.sb_bsize_shift; find_metapath(sdp, lblock, &mp, ip->i_height); - gfs2_alloc_get(ip); + if (!gfs2_alloc_get(ip)) + return -ENOMEM; error = gfs2_quota_hold(ip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE); if (error) diff --git a/fs/gfs2/dir.c b/fs/gfs2/dir.c index a3753c7989f..94070ad8826 100644 --- a/fs/gfs2/dir.c +++ b/fs/gfs2/dir.c @@ -1868,11 +1868,14 @@ static int leaf_dealloc(struct gfs2_inode *dip, u32 index, u32 len, if (!ht) return -ENOMEM; - gfs2_alloc_get(dip); + if (!gfs2_alloc_get(dip)) { + error = -ENOMEM; + goto out; + } error = gfs2_quota_hold(dip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE); if (error) - goto out; + goto out_put; error = gfs2_rindex_hold(sdp, &dip->i_alloc->al_ri_gh); if (error) @@ -1946,8 +1949,9 @@ out_rlist: gfs2_glock_dq_uninit(&dip->i_alloc->al_ri_gh); out_qs: gfs2_quota_unhold(dip); -out: +out_put: gfs2_alloc_put(dip); +out: kfree(ht); return error; } diff --git a/fs/gfs2/eattr.c b/fs/gfs2/eattr.c index 76ead1acfcc..288d5e6ad93 100644 --- a/fs/gfs2/eattr.c +++ b/fs/gfs2/eattr.c @@ -318,6 +318,8 @@ static int ea_remove_unstuffed(struct gfs2_inode *ip, struct buffer_head *bh, int error; al = gfs2_alloc_get(ip); + if (!al) + return -ENOMEM; error = gfs2_quota_hold(ip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE); if (error) @@ -681,6 +683,8 @@ static int ea_alloc_skeleton(struct gfs2_inode *ip, struct gfs2_ea_request *er, int error; al = gfs2_alloc_get(ip); + if (!al) + return -ENOMEM; error = gfs2_quota_lock(ip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE); if (error) @@ -1464,6 +1468,8 @@ int gfs2_ea_dealloc(struct gfs2_inode *ip) int error; al = gfs2_alloc_get(ip); + if (!al) + return -ENOMEM; error = gfs2_quota_hold(ip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE); if (error) diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c index 92ea9afacb1..dcae2aa83f1 100644 --- a/fs/gfs2/inode.c +++ b/fs/gfs2/inode.c @@ -351,6 +351,8 @@ int gfs2_dinode_dealloc(struct gfs2_inode *ip) } al = gfs2_alloc_get(ip); + if (!al) + return -ENOMEM; error = gfs2_quota_hold(ip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE); if (error) @@ -825,7 +827,8 @@ static int make_dinode(struct gfs2_inode *dip, struct gfs2_glock *gl, int error; munge_mode_uid_gid(dip, &mode, &uid, &gid); - gfs2_alloc_get(dip); + if (!gfs2_alloc_get(dip)) + return -ENOMEM; error = gfs2_quota_lock(dip, uid, gid); if (error) @@ -860,6 +863,8 @@ static int link_dinode(struct gfs2_inode *dip, const struct qstr *name, int error; al = gfs2_alloc_get(dip); + if (!al) + return -ENOMEM; error = gfs2_quota_lock(dip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE); if (error) diff --git a/fs/gfs2/ops_address.c b/fs/gfs2/ops_address.c index fbb4a6aa158..2483d874106 100644 --- a/fs/gfs2/ops_address.c +++ b/fs/gfs2/ops_address.c @@ -649,6 +649,10 @@ static int gfs2_write_begin(struct file *file, struct address_space *mapping, if (alloc_required) { al = gfs2_alloc_get(ip); + if (!al) { + error = -ENOMEM; + goto out_unlock; + } error = gfs2_quota_lock(ip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE); if (error) diff --git a/fs/gfs2/ops_inode.c b/fs/gfs2/ops_inode.c index 6cbbb5b46ce..34fe571e15e 100644 --- a/fs/gfs2/ops_inode.c +++ b/fs/gfs2/ops_inode.c @@ -200,6 +200,10 @@ static int gfs2_link(struct dentry *old_dentry, struct inode *dir, if (alloc_required) { struct gfs2_alloc *al = gfs2_alloc_get(dip); + if (!al) { + error = -ENOMEM; + goto out_gunlock; + } error = gfs2_quota_lock(dip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE); if (error) @@ -716,6 +720,10 @@ static int gfs2_rename(struct inode *odir, struct dentry *odentry, if (alloc_required) { struct gfs2_alloc *al = gfs2_alloc_get(ndip); + if (!al) { + error = -ENOMEM; + goto out_gunlock; + } error = gfs2_quota_lock(ndip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE); if (error) @@ -953,7 +961,8 @@ static int setattr_chown(struct inode *inode, struct iattr *attr) if (!(attr->ia_valid & ATTR_GID) || ogid == ngid) ogid = ngid = NO_QUOTA_CHANGE; - gfs2_alloc_get(ip); + if (!gfs2_alloc_get(ip)) + return -ENOMEM; error = gfs2_quota_lock(ip, nuid, ngid); if (error) diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c index a08dabd6ce9..636bccfd2bc 100644 --- a/fs/gfs2/quota.c +++ b/fs/gfs2/quota.c @@ -617,8 +617,9 @@ static int gfs2_adjust_quota(struct gfs2_inode *ip, loff_t loc, int err = -EIO; if (gfs2_is_stuffed(ip)) { - struct gfs2_alloc *al = NULL; - al = gfs2_alloc_get(ip); + struct gfs2_alloc *al = gfs2_alloc_get(ip); + if (!al) + return -ENOMEM; /* just request 1 blk */ al->al_requested = 1; gfs2_inplace_reserve(ip); @@ -729,6 +730,10 @@ static int do_sync(unsigned int num_qd, struct gfs2_quota_data **qda) if (nalloc) { al = gfs2_alloc_get(ip); + if (!al) { + error = -ENOMEM; + goto out_gunlock; + } al->al_requested = nalloc * (data_blocks + ind_blocks); -- cgit v1.2.3 From 20b95bf2c4c5c28e093aa42699e67829b6cd7fd0 Mon Sep 17 00:00:00 2001 From: Abhijith Das Date: Thu, 6 Mar 2008 17:43:52 -0600 Subject: [GFS2] gfs2_adjust_quota has broken unstuffing code This patch combines the 2 patches in bug 434736 to correct the lock ordering in the unstuffing of the quota inode in gfs2_adjust_quota and adjusting the number of revokes in gfs2_write_jdata_pagevec Signed-off-by: Abhijith Das Signed-off-by: Steven Whitehouse --- fs/gfs2/ops_address.c | 3 +-- fs/gfs2/quota.c | 71 ++++++++++++++++++++++----------------------------- 2 files changed, 31 insertions(+), 43 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/ops_address.c b/fs/gfs2/ops_address.c index 2483d874106..e72fd47d71e 100644 --- a/fs/gfs2/ops_address.c +++ b/fs/gfs2/ops_address.c @@ -21,7 +21,6 @@ #include #include #include -#include #include "gfs2.h" #include "incore.h" @@ -278,7 +277,7 @@ static int gfs2_write_jdata_pagevec(struct address_space *mapping, int i; int ret; - ret = gfs2_trans_begin(sdp, nrblocks, 0); + ret = gfs2_trans_begin(sdp, nrblocks, nrblocks); if (ret < 0) return ret; diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c index 636bccfd2bc..c71f781db5d 100644 --- a/fs/gfs2/quota.c +++ b/fs/gfs2/quota.c @@ -616,17 +616,9 @@ static int gfs2_adjust_quota(struct gfs2_inode *ip, loff_t loc, s64 value; int err = -EIO; - if (gfs2_is_stuffed(ip)) { - struct gfs2_alloc *al = gfs2_alloc_get(ip); - if (!al) - return -ENOMEM; - /* just request 1 blk */ - al->al_requested = 1; - gfs2_inplace_reserve(ip); + if (gfs2_is_stuffed(ip)) gfs2_unstuff_dinode(ip, NULL); - gfs2_inplace_release(ip); - gfs2_alloc_put(ip); - } + page = grab_cache_page(mapping, index); if (!page) return -ENOMEM; @@ -691,7 +683,7 @@ static int do_sync(unsigned int num_qd, struct gfs2_quota_data **qda) unsigned int qx, x; struct gfs2_quota_data *qd; loff_t offset; - unsigned int nalloc = 0; + unsigned int nalloc = 0, blocks; struct gfs2_alloc *al = NULL; int error; @@ -728,34 +720,33 @@ static int do_sync(unsigned int num_qd, struct gfs2_quota_data **qda) nalloc++; } - if (nalloc) { - al = gfs2_alloc_get(ip); - if (!al) { - error = -ENOMEM; - goto out_gunlock; - } + al = gfs2_alloc_get(ip); + if (!al) { + error = -ENOMEM; + goto out_gunlock; + } + /* + * 1 blk for unstuffing inode if stuffed. We add this extra + * block to the reservation unconditionally. If the inode + * doesn't need unstuffing, the block will be released to the + * rgrp since it won't be allocated during the transaction + */ + al->al_requested = 1; + /* +1 in the end for block requested above for unstuffing */ + blocks = num_qd * data_blocks + RES_DINODE + num_qd + 1; - al->al_requested = nalloc * (data_blocks + ind_blocks); + if (nalloc) + al->al_requested += nalloc * (data_blocks + ind_blocks); + error = gfs2_inplace_reserve(ip); + if (error) + goto out_alloc; - error = gfs2_inplace_reserve(ip); - if (error) - goto out_alloc; - - error = gfs2_trans_begin(sdp, - al->al_rgd->rd_length + - num_qd * data_blocks + - nalloc * ind_blocks + - RES_DINODE + num_qd + - RES_STATFS, 0); - if (error) - goto out_ipres; - } else { - error = gfs2_trans_begin(sdp, - num_qd * data_blocks + - RES_DINODE + num_qd, 0); - if (error) - goto out_gunlock; - } + if (nalloc) + blocks += al->al_rgd->rd_length + nalloc * ind_blocks + RES_STATFS; + + error = gfs2_trans_begin(sdp, blocks, 0); + if (error) + goto out_ipres; for (x = 0; x < num_qd; x++) { qd = qda[x]; @@ -774,11 +765,9 @@ static int do_sync(unsigned int num_qd, struct gfs2_quota_data **qda) out_end_trans: gfs2_trans_end(sdp); out_ipres: - if (nalloc) - gfs2_inplace_release(ip); + gfs2_inplace_release(ip); out_alloc: - if (nalloc) - gfs2_alloc_put(ip); + gfs2_alloc_put(ip); out_gunlock: gfs2_glock_dq_uninit(&i_gh); out: -- cgit v1.2.3 From 860b25d4a913a00331d333f8e207a088c7a1b84a Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Mon, 10 Mar 2008 10:13:31 +0000 Subject: [GFS2] Remove drop of module ref where not needed In an earlier patch "[GFS2] fix file_system_type leak on gfs2meta mount" we removed the code to grab a ref to the module which was not needed (since we know that the module cannot be unloaded at that time) so this patch removes the code to drop that reference. Signed-off-by: Steven Whitehouse --- fs/gfs2/ops_fstype.c | 1 - 1 file changed, 1 deletion(-) (limited to 'fs') diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c index 63d5fd203d3..5b518f73497 100644 --- a/fs/gfs2/ops_fstype.c +++ b/fs/gfs2/ops_fstype.c @@ -998,7 +998,6 @@ static int gfs2_get_sb_meta(struct file_system_type *fs_type, int flags, error = PTR_ERR(new); goto error; } - module_put(fs_type->owner); new->s_flags = flags; strlcpy(new->s_id, sb->s_id, sizeof(new->s_id)); sb_set_blocksize(new, sb->s_blocksize); -- cgit v1.2.3 From d82661d96993ac4efc1d54259ea85ffcd9b8bec6 Mon Sep 17 00:00:00 2001 From: Steven Whitehouse Date: Mon, 10 Mar 2008 15:34:50 +0000 Subject: [GFS2] Streamline quota lock/check for no-quota case This patch streamlines the quota checking in the "no quota" case by making the check inline in the calling function, thus reducing the number of function calls. Eventually we might be able to remove the checks from the gfs2_quota_lock() and gfs2_quota_check() functions, but currently we can't as there are a very few places in the code which need to call these functions directly still. Signed-off-by: Steven Whitehouse Cc: Abhijith Das --- fs/gfs2/bmap.c | 6 +----- fs/gfs2/eattr.c | 6 +----- fs/gfs2/ops_address.c | 6 +----- fs/gfs2/ops_file.c | 5 +---- fs/gfs2/ops_inode.c | 12 ++---------- fs/gfs2/quota.h | 17 +++++++++++++++++ 6 files changed, 23 insertions(+), 29 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/bmap.c b/fs/gfs2/bmap.c index f7093aa69aa..c19184f2e70 100644 --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -903,14 +903,10 @@ static int do_grow(struct gfs2_inode *ip, u64 size) if (!al) return -ENOMEM; - error = gfs2_quota_lock(ip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE); + error = gfs2_quota_lock_check(ip); if (error) goto out; - error = gfs2_quota_check(ip, ip->i_inode.i_uid, ip->i_inode.i_gid); - if (error) - goto out_gunlock_q; - al->al_requested = sdp->sd_max_height + RES_DATA; error = gfs2_inplace_reserve(ip); diff --git a/fs/gfs2/eattr.c b/fs/gfs2/eattr.c index 288d5e6ad93..81755925a75 100644 --- a/fs/gfs2/eattr.c +++ b/fs/gfs2/eattr.c @@ -686,14 +686,10 @@ static int ea_alloc_skeleton(struct gfs2_inode *ip, struct gfs2_ea_request *er, if (!al) return -ENOMEM; - error = gfs2_quota_lock(ip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE); + error = gfs2_quota_lock_check(ip); if (error) goto out; - error = gfs2_quota_check(ip, ip->i_inode.i_uid, ip->i_inode.i_gid); - if (error) - goto out_gunlock_q; - al->al_requested = blks; error = gfs2_inplace_reserve(ip); diff --git a/fs/gfs2/ops_address.c b/fs/gfs2/ops_address.c index e72fd47d71e..90a04a6e378 100644 --- a/fs/gfs2/ops_address.c +++ b/fs/gfs2/ops_address.c @@ -653,14 +653,10 @@ static int gfs2_write_begin(struct file *file, struct address_space *mapping, goto out_unlock; } - error = gfs2_quota_lock(ip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE); + error = gfs2_quota_lock_check(ip); if (error) goto out_alloc_put; - error = gfs2_quota_check(ip, ip->i_inode.i_uid, ip->i_inode.i_gid); - if (error) - goto out_qunlock; - al->al_requested = data_blocks + ind_blocks; error = gfs2_inplace_reserve(ip); if (error) diff --git a/fs/gfs2/ops_file.c b/fs/gfs2/ops_file.c index 2b25a5f7a1c..e1b7d525a06 100644 --- a/fs/gfs2/ops_file.c +++ b/fs/gfs2/ops_file.c @@ -369,12 +369,9 @@ static int gfs2_page_mkwrite(struct vm_area_struct *vma, struct page *page) if (al == NULL) goto out_unlock; - ret = gfs2_quota_lock(ip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE); + ret = gfs2_quota_lock_check(ip); if (ret) goto out_alloc_put; - ret = gfs2_quota_check(ip, ip->i_inode.i_uid, ip->i_inode.i_gid); - if (ret) - goto out_quota_unlock; al->al_requested = data_blocks + ind_blocks; ret = gfs2_inplace_reserve(ip); if (ret) diff --git a/fs/gfs2/ops_inode.c b/fs/gfs2/ops_inode.c index 34fe571e15e..2686ad4c002 100644 --- a/fs/gfs2/ops_inode.c +++ b/fs/gfs2/ops_inode.c @@ -205,14 +205,10 @@ static int gfs2_link(struct dentry *old_dentry, struct inode *dir, goto out_gunlock; } - error = gfs2_quota_lock(dip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE); + error = gfs2_quota_lock_check(dip); if (error) goto out_alloc; - error = gfs2_quota_check(dip, dip->i_inode.i_uid, dip->i_inode.i_gid); - if (error) - goto out_gunlock_q; - al->al_requested = sdp->sd_max_dirres; error = gfs2_inplace_reserve(dip); @@ -725,14 +721,10 @@ static int gfs2_rename(struct inode *odir, struct dentry *odentry, goto out_gunlock; } - error = gfs2_quota_lock(ndip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE); + error = gfs2_quota_lock_check(ndip); if (error) goto out_alloc; - error = gfs2_quota_check(ndip, ndip->i_inode.i_uid, ndip->i_inode.i_gid); - if (error) - goto out_gunlock_q; - al->al_requested = sdp->sd_max_dirres; error = gfs2_inplace_reserve(ndip); diff --git a/fs/gfs2/quota.h b/fs/gfs2/quota.h index a8be1417051..3b7f4b0e5df 100644 --- a/fs/gfs2/quota.h +++ b/fs/gfs2/quota.h @@ -32,4 +32,21 @@ int gfs2_quota_init(struct gfs2_sbd *sdp); void gfs2_quota_scan(struct gfs2_sbd *sdp); void gfs2_quota_cleanup(struct gfs2_sbd *sdp); +static inline int gfs2_quota_lock_check(struct gfs2_inode *ip) +{ + struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode); + int ret; + if (sdp->sd_args.ar_quota == GFS2_QUOTA_OFF) + return 0; + ret = gfs2_quota_lock(ip, NO_QUOTA_CHANGE, NO_QUOTA_CHANGE); + if (ret) + return ret; + if (sdp->sd_args.ar_quota != GFS2_QUOTA_ON) + return 0; + ret = gfs2_quota_check(ip, ip->i_inode.i_uid, ip->i_inode.i_gid); + if (ret) + gfs2_quota_unlock(ip); + return ret; +} + #endif /* __QUOTA_DOT_H__ */ -- cgit v1.2.3 From 1f466a47e8a3a3e3b527b3285c7b9c8a837fb7ec Mon Sep 17 00:00:00 2001 From: Bob Peterson Date: Mon, 10 Mar 2008 18:17:47 -0500 Subject: [GFS2] Faster gfs2_bitfit algorithm This version of the gfs2_bitfit algorithm includes the latest suggestions from Steve Whitehouse. It is typically eight to ten times faster than the version we're using today. If there is a lot of metadata mixed in (lots of small files) the algorithm is often 15 times faster, and given the right conditions, I've seen peaks of 20 times faster. Signed-off-by: Bob Peterson Signed-off-by: Steven Whitehouse --- fs/gfs2/rgrp.c | 93 ++++++++++++++++++++++++++++++++++++++-------------------- 1 file changed, 61 insertions(+), 32 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c index 4291375cecc..7e8f0b1d6c6 100644 --- a/fs/gfs2/rgrp.c +++ b/fs/gfs2/rgrp.c @@ -14,6 +14,7 @@ #include #include #include +#include #include "gfs2.h" #include "incore.h" @@ -33,6 +34,16 @@ #define BFITNOENT ((u32)~0) #define NO_BLOCK ((u64)~0) +#if BITS_PER_LONG == 32 +#define LBITMASK (0x55555555UL) +#define LBITSKIP55 (0x55555555UL) +#define LBITSKIP00 (0x00000000UL) +#else +#define LBITMASK (0x5555555555555555UL) +#define LBITSKIP55 (0x5555555555555555UL) +#define LBITSKIP00 (0x0000000000000000UL) +#endif + /* * These routines are used by the resource group routines (rgrp.c) * to keep track of block allocation. Each block is represented by two @@ -138,45 +149,63 @@ static inline unsigned char gfs2_testbit(struct gfs2_rgrpd *rgd, static u32 gfs2_bitfit(const u8 *buffer, unsigned int buflen, u32 goal, u8 old_state) { - const u8 *byte; - u32 blk = goal; - unsigned int bit, bitlong; - const unsigned long *plong; -#if BITS_PER_LONG == 32 - const unsigned long plong55 = 0x55555555; -#else - const unsigned long plong55 = 0x5555555555555555; -#endif - - byte = buffer + (goal / GFS2_NBBY); - plong = (const unsigned long *)(buffer + (goal / GFS2_NBBY)); - bit = (goal % GFS2_NBBY) * GFS2_BIT_SIZE; - bitlong = bit; - - while (byte < buffer + buflen) { - - if (bitlong == 0 && old_state == 0 && *plong == plong55) { - plong++; - byte += sizeof(unsigned long); - blk += sizeof(unsigned long) * GFS2_NBBY; - continue; + const u8 *byte, *start, *end; + int bit, startbit; + u32 g1, g2, misaligned; + unsigned long *plong; + unsigned long lskipval; + + lskipval = (old_state & GFS2_BLKST_USED) ? LBITSKIP00 : LBITSKIP55; + g1 = (goal / GFS2_NBBY); + start = buffer + g1; + byte = start; + end = buffer + buflen; + g2 = ALIGN(g1, sizeof(unsigned long)); + plong = (unsigned long *)(buffer + g2); + startbit = bit = (goal % GFS2_NBBY) * GFS2_BIT_SIZE; + misaligned = g2 - g1; + if (!misaligned) + goto ulong_aligned; +/* parse the bitmap a byte at a time */ +misaligned: + while (byte < end) { + if (((*byte >> bit) & GFS2_BIT_MASK) == old_state) { + return goal + + (((byte - start) * GFS2_NBBY) + + ((bit - startbit) >> 1)); } - if (((*byte >> bit) & GFS2_BIT_MASK) == old_state) - return blk; bit += GFS2_BIT_SIZE; - if (bit >= 8) { + if (bit >= GFS2_NBBY * GFS2_BIT_SIZE) { bit = 0; byte++; + misaligned--; + if (!misaligned) { + plong = (unsigned long *)byte; + goto ulong_aligned; + } } - bitlong += GFS2_BIT_SIZE; - if (bitlong >= sizeof(unsigned long) * 8) { - bitlong = 0; - plong++; - } - - blk++; } + return BFITNOENT; +/* parse the bitmap a unsigned long at a time */ +ulong_aligned: + /* Stop at "end - 1" or else prefetch can go past the end and segfault. + We could "if" it but we'd lose some of the performance gained. + This way will only slow down searching the very last 4/8 bytes + depending on architecture. I've experimented with several ways + of writing this section such as using an else before the goto + but this one seems to be the fastest. */ + while ((unsigned char *)plong < end - 1) { + prefetch(plong + 1); + if (((*plong) & LBITMASK) != lskipval) + break; + plong++; + } + if ((unsigned char *)plong < end) { + byte = (const u8 *)plong; + misaligned += sizeof(unsigned long) - 1; + goto misaligned; + } return BFITNOENT; } -- cgit v1.2.3 From f5a8cd020173c455705fc0095b7d299da6f8f87b Mon Sep 17 00:00:00 2001 From: "akpm@linux-foundation.org" Date: Wed, 12 Mar 2008 14:01:29 -0700 Subject: [GFS2] fs/gfs2/recovery.c: suppress warnings fs/gfs2/recovery.c: In function 'get_log_header': fs/gfs2/recovery.c:152: warning: 'lh.lh_sequence' may be used uninitialized in this function fs/gfs2/recovery.c:152: warning: 'lh.lh_flags' may be used uninitialized in this function fs/gfs2/recovery.c:152: warning: 'lh.lh_tail' may be used uninitialized in this function fs/gfs2/recovery.c:152: warning: 'lh.lh_blkno' may be used uninitialized in this function fs/gfs2/recovery.c:152: warning: 'lh.lh_hash' may be used uninitialized in this function Cc: David Teigland Cc: Bob Peterson Signed-off-by: Andrew Morton Signed-off-by: Steven Whitehouse --- fs/gfs2/recovery.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/gfs2/recovery.c b/fs/gfs2/recovery.c index b17d3b8b232..06dcdc04627 100644 --- a/fs/gfs2/recovery.c +++ b/fs/gfs2/recovery.c @@ -149,7 +149,7 @@ static int get_log_header(struct gfs2_jdesc *jd, unsigned int blk, struct gfs2_log_header_host *head) { struct buffer_head *bh; - struct gfs2_log_header_host lh; + struct gfs2_log_header_host uninitialized_var(lh); const u32 nothing = 0; u32 hash; int error; -- cgit v1.2.3 From 58e9fee13e579df44922172dbe3c9e3ba3edf7a3 Mon Sep 17 00:00:00 2001 From: Benjamin Marzinski Date: Fri, 14 Mar 2008 13:52:52 -0500 Subject: [GFS2] Invalidate cache at correct point GFS2 wasn't invalidating its cache before it called into the lock manager with a request that could potentially drop a lock. This was leaving a window where the lock could be actually be held by another node, but the file's page cache would still appear valid, causing coherency problems. This patch moves the cache invalidation to before the lock manager call when dropping a lock. It also adds the option to the lock_dlm lock manager to not use conversion mode deadlock avoidance, which, on a conversion from shared to exclusive, could internally drop the lock, and then reacquire in. GFS2 now asks lock_dlm to not do this. Instead, GFS2 manually drops the lock and reacquires it. Signed-off-by: Benjamin Marzinski Signed-off-by: Steven Whitehouse --- fs/gfs2/glock.c | 35 +++++++++++++++++++++-------------- fs/gfs2/incore.h | 1 + fs/gfs2/locking/dlm/lock.c | 3 ++- fs/gfs2/locking/dlm/thread.c | 10 +++++++++- fs/gfs2/ops_fstype.c | 2 +- 5 files changed, 34 insertions(+), 17 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/glock.c b/fs/gfs2/glock.c index 63981e2fb83..d636b3e80f5 100644 --- a/fs/gfs2/glock.c +++ b/fs/gfs2/glock.c @@ -764,7 +764,7 @@ static void state_change(struct gfs2_glock *gl, unsigned int new_state) static void drop_bh(struct gfs2_glock *gl, unsigned int ret) { struct gfs2_sbd *sdp = gl->gl_sbd; - const struct gfs2_glock_operations *glops = gl->gl_ops; + struct gfs2_holder *gh = gl->gl_req_gh; gfs2_assert_warn(sdp, test_bit(GLF_LOCK, &gl->gl_flags)); gfs2_assert_warn(sdp, list_empty(&gl->gl_holders)); @@ -772,8 +772,14 @@ static void drop_bh(struct gfs2_glock *gl, unsigned int ret) state_change(gl, LM_ST_UNLOCKED); - if (glops->go_inval) - glops->go_inval(gl, DIO_METADATA); + if (test_and_clear_bit(GLF_CONV_DEADLK, &gl->gl_flags)) { + spin_lock(&gl->gl_spin); + gh->gh_error = 0; + spin_unlock(&gl->gl_spin); + gfs2_glock_xmote_th(gl, gl->gl_req_gh); + gfs2_glock_put(gl); + return; + } spin_lock(&gl->gl_spin); gfs2_demote_wake(gl); @@ -794,7 +800,6 @@ static void xmote_bh(struct gfs2_glock *gl, unsigned int ret) struct gfs2_sbd *sdp = gl->gl_sbd; const struct gfs2_glock_operations *glops = gl->gl_ops; struct gfs2_holder *gh = gl->gl_req_gh; - int prev_state = gl->gl_state; int op_done = 1; if (!gh && (ret & LM_OUT_ST_MASK) == LM_ST_UNLOCKED) { @@ -808,16 +813,6 @@ static void xmote_bh(struct gfs2_glock *gl, unsigned int ret) state_change(gl, ret & LM_OUT_ST_MASK); - if (prev_state != LM_ST_UNLOCKED && !(ret & LM_OUT_CACHEABLE)) { - if (glops->go_inval) - glops->go_inval(gl, DIO_METADATA); - } else if (gl->gl_state == LM_ST_DEFERRED) { - /* We might not want to do this here. - Look at moving to the inode glops. */ - if (glops->go_inval) - glops->go_inval(gl, 0); - } - /* Deal with each possible exit condition */ if (!gh) { @@ -837,6 +832,14 @@ static void xmote_bh(struct gfs2_glock *gl, unsigned int ret) } } else { spin_lock(&gl->gl_spin); + if (ret & LM_OUT_CONV_DEADLK) { + gh->gh_error = 0; + set_bit(GLF_CONV_DEADLK, &gl->gl_flags); + spin_unlock(&gl->gl_spin); + gfs2_glock_drop_th(gl); + gfs2_glock_put(gl); + return; + } list_del_init(&gh->gh_list); gh->gh_error = -EIO; if (unlikely(test_bit(SDF_SHUTDOWN, &sdp->sd_flags))) @@ -910,6 +913,8 @@ static void gfs2_glock_xmote_th(struct gfs2_glock *gl, struct gfs2_holder *gh) if (glops->go_xmote_th) glops->go_xmote_th(gl); + if (state == LM_ST_DEFERRED && glops->go_inval) + glops->go_inval(gl, DIO_METADATA); gfs2_assert_warn(sdp, test_bit(GLF_LOCK, &gl->gl_flags)); gfs2_assert_warn(sdp, list_empty(&gl->gl_holders)); @@ -952,6 +957,8 @@ static void gfs2_glock_drop_th(struct gfs2_glock *gl) if (glops->go_xmote_th) glops->go_xmote_th(gl); + if (glops->go_inval) + glops->go_inval(gl, DIO_METADATA); gfs2_assert_warn(sdp, test_bit(GLF_LOCK, &gl->gl_flags)); gfs2_assert_warn(sdp, list_empty(&gl->gl_holders)); diff --git a/fs/gfs2/incore.h b/fs/gfs2/incore.h index 4ba2ea63119..9c2c0b90b22 100644 --- a/fs/gfs2/incore.h +++ b/fs/gfs2/incore.h @@ -167,6 +167,7 @@ enum { GLF_DEMOTE_IN_PROGRESS = 6, GLF_LFLUSH = 7, GLF_WAITERS2 = 8, + GLF_CONV_DEADLK = 9, }; struct gfs2_glock { diff --git a/fs/gfs2/locking/dlm/lock.c b/fs/gfs2/locking/dlm/lock.c index 542a797ac89..53a6ab3c091 100644 --- a/fs/gfs2/locking/dlm/lock.c +++ b/fs/gfs2/locking/dlm/lock.c @@ -137,7 +137,8 @@ static inline unsigned int make_flags(struct gdlm_lock *lp, /* Conversion deadlock avoidance by DLM */ - if (!test_bit(LFL_FORCE_PROMOTE, &lp->flags) && + if (!(lp->ls->fsflags & LM_MFLAG_CONV_NODROP) && + !test_bit(LFL_FORCE_PROMOTE, &lp->flags) && !(lkf & DLM_LKF_NOQUEUE) && cur > DLM_LOCK_NL && req > DLM_LOCK_NL && cur != req) lkf |= DLM_LKF_CONVDEADLK; diff --git a/fs/gfs2/locking/dlm/thread.c b/fs/gfs2/locking/dlm/thread.c index 521694fc19d..e53db6fd28a 100644 --- a/fs/gfs2/locking/dlm/thread.c +++ b/fs/gfs2/locking/dlm/thread.c @@ -135,7 +135,15 @@ static void process_complete(struct gdlm_lock *lp) lp->lksb.sb_status, lp->lockname.ln_type, (unsigned long long)lp->lockname.ln_number, lp->flags); - return; + if (lp->lksb.sb_status == -EDEADLOCK && + lp->ls->fsflags & LM_MFLAG_CONV_NODROP) { + lp->req = lp->cur; + acb.lc_ret |= LM_OUT_CONV_DEADLK; + if (lp->cur == DLM_LOCK_IV) + lp->lksb.sb_lkid = 0; + goto out; + } else + return; } /* diff --git a/fs/gfs2/ops_fstype.c b/fs/gfs2/ops_fstype.c index 5b518f73497..ef9c6c4f80f 100644 --- a/fs/gfs2/ops_fstype.c +++ b/fs/gfs2/ops_fstype.c @@ -723,7 +723,7 @@ static int gfs2_lm_mount(struct gfs2_sbd *sdp, int silent) { char *proto = sdp->sd_proto_name; char *table = sdp->sd_table_name; - int flags = 0; + int flags = LM_MFLAG_CONV_NODROP; int error; if (sdp->sd_args.ar_spectator) -- cgit v1.2.3 From 773adff8e983cba1f5844c3be3be224ca6645f26 Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Mon, 24 Mar 2008 19:08:48 +0100 Subject: [GFS2] test for IS_ERR rather than 0 The function gfs2_inode_lookup always returns either a valid pointer or a value made with ERR_PTR, so its result should be tested with IS_ERR, not with a test for 0. The problem was found using the following semantic match. (http://www.emn.fr/x-info/coccinelle/) // @a@ expression E, E1; statement S,S1; position p; @@ E = gfs2_inode_lookup(...) ... when != E = E1 if@p (E) S else S1 @n@ position a.p; expression E,E1; statement S,S1; @@ E = NULL ... when != E = E1 if@p (E) S else S1 @depends on !n@ expression E; statement S,S1; position a.p; @@ * if@p (E) S else S1 // Signed-off-by: Julia Lawall Signed-off-by: Steven Whitehouse --- fs/gfs2/ops_export.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/ops_export.c b/fs/gfs2/ops_export.c index 334c7f85351..990d9f4bc46 100644 --- a/fs/gfs2/ops_export.c +++ b/fs/gfs2/ops_export.c @@ -204,8 +204,6 @@ static struct dentry *gfs2_get_dentry(struct super_block *sb, inode = gfs2_inode_lookup(sb, DT_UNKNOWN, inum->no_addr, 0, 0); - if (!inode) - goto fail; if (IS_ERR(inode)) { error = PTR_ERR(inode); goto fail; -- cgit v1.2.3 From 856f6ff7a3132c8e412b23a7b9157b68ac9a2baf Mon Sep 17 00:00:00 2001 From: "Denis V. Lunev" Date: Wed, 2 Apr 2008 00:10:04 -0700 Subject: [NETNS]: Remove ifdef CONFIG_NET braces in fs/proc/proc_net.c. They are redundant as this file is linked in iff CONFIG_NET is turned on. Signed-off-by: Denis V. Lunev Signed-off-by: David S. Miller --- fs/proc/proc_net.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'fs') diff --git a/fs/proc/proc_net.c b/fs/proc/proc_net.c index 7034facf8b8..13cd7835d0d 100644 --- a/fs/proc/proc_net.c +++ b/fs/proc/proc_net.c @@ -51,7 +51,6 @@ int seq_open_net(struct inode *ino, struct file *f, } EXPORT_SYMBOL_GPL(seq_open_net); -#ifdef CONFIG_NET int seq_release_net(struct inode *ino, struct file *f) { struct seq_file *seq; @@ -219,4 +218,3 @@ int __init proc_net_init(void) return register_pernet_subsys(&proc_net_ns_ops); } -#endif /* CONFIG_NET */ -- cgit v1.2.3 From 932e2d23c8529c39fe74549dcbe4b2c8b2d66ba4 Mon Sep 17 00:00:00 2001 From: Steve French Date: Fri, 4 Apr 2008 21:59:35 +0000 Subject: [CIFS] minor update to change log Signed-off-by: Steve French --- fs/cifs/CHANGES | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/cifs/CHANGES b/fs/cifs/CHANGES index dbd91461853..05c9da6181c 100644 --- a/fs/cifs/CHANGES +++ b/fs/cifs/CHANGES @@ -8,7 +8,8 @@ of second share to disconnected server session (autoreconnect on this). Add ability to modify cifs acls for handling chmod (when mounted with cifsacl flag). Fix prefixpath path separator so we can handle mounts with prefixpaths longer than one directory (one path component) when -mounted to Windows servers. +mounted to Windows servers. Fix slow file open when cifsacl +enabled. Version 1.51 ------------ -- cgit v1.2.3 From 35028d71116926008f5c19b8fcb00aacaabf5eab Mon Sep 17 00:00:00 2001 From: Steve French Date: Wed, 9 Apr 2008 20:32:42 +0000 Subject: [CIFS] Fix looping on reconnect to Samba when unexpected tree connect fail on reconnect Signed-off-by: Shirish Pargaonkar Signed-off-by: Steve French --- fs/cifs/cifssmb.c | 32 ++++++++++++++++++-------------- 1 file changed, 18 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifssmb.c b/fs/cifs/cifssmb.c index 30bbe448e26..4728fa982a4 100644 --- a/fs/cifs/cifssmb.c +++ b/fs/cifs/cifssmb.c @@ -165,17 +165,19 @@ small_smb_init(int smb_command, int wct, struct cifsTconInfo *tcon, rc = CIFSTCon(0, tcon->ses, tcon->treeName, tcon, nls_codepage); up(&tcon->ses->sesSem); - /* tell server which Unix caps we support */ - if (tcon->ses->capabilities & CAP_UNIX) - reset_cifs_unix_caps(0 /* no xid */, - tcon, - NULL /* we do not know sb */, - NULL /* no vol info */); /* BB FIXME add code to check if wsize needs update due to negotiated smb buffer size shrinking */ - if (rc == 0) + if (rc == 0) { atomic_inc(&tconInfoReconnectCount); + /* tell server Unix caps we support */ + if (tcon->ses->capabilities & CAP_UNIX) + reset_cifs_unix_caps( + 0 /* no xid */, + tcon, + NULL /* we do not know sb */, + NULL /* no vol info */); + } cFYI(1, ("reconnect tcon rc = %d", rc)); /* Removed call to reopen open files here. @@ -310,17 +312,19 @@ smb_init(int smb_command, int wct, struct cifsTconInfo *tcon, rc = CIFSTCon(0, tcon->ses, tcon->treeName, tcon, nls_codepage); up(&tcon->ses->sesSem); - /* tell server which Unix caps we support */ - if (tcon->ses->capabilities & CAP_UNIX) - reset_cifs_unix_caps(0 /* no xid */, - tcon, - NULL /* do not know sb */, - NULL /* no vol info */); /* BB FIXME add code to check if wsize needs update due to negotiated smb buffer size shrinking */ - if (rc == 0) + if (rc == 0) { atomic_inc(&tconInfoReconnectCount); + /* tell server Unix caps we support */ + if (tcon->ses->capabilities & CAP_UNIX) + reset_cifs_unix_caps( + 0 /* no xid */, + tcon, + NULL /* do not know sb */, + NULL /* no vol info */); + } cFYI(1, ("reconnect tcon rc = %d", rc)); /* Removed call to reopen open files here. -- cgit v1.2.3 From cce246ee5f3c7f4d3539ea41d13feb7a07859145 Mon Sep 17 00:00:00 2001 From: Steve French Date: Wed, 9 Apr 2008 20:55:31 +0000 Subject: [CIFS] Fix acl length when very short ACL being modified by chmod Signed-off-by: Shirish Pargaonkar Signed-off-by: Steve French --- fs/cifs/cifsacl.c | 14 ++++++++------ fs/cifs/cifsacl.h | 1 + 2 files changed, 9 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifsacl.c b/fs/cifs/cifsacl.c index 1cb5b0a9f2a..e99d4faf5f0 100644 --- a/fs/cifs/cifsacl.c +++ b/fs/cifs/cifsacl.c @@ -516,7 +516,7 @@ static int parse_sec_desc(struct cifs_ntsd *pntsd, int acl_len, /* Convert permission bits from mode to equivalent CIFS ACL */ static int build_sec_desc(struct cifs_ntsd *pntsd, struct cifs_ntsd *pnntsd, - int acl_len, struct inode *inode, __u64 nmode) + struct inode *inode, __u64 nmode) { int rc = 0; __u32 dacloffset; @@ -692,14 +692,14 @@ void acl_to_uid_mode(struct inode *inode, const char *path, const __u16 *pfid) int mode_to_acl(struct inode *inode, const char *path, __u64 nmode) { int rc = 0; - __u32 acllen = 0; + __u32 secdesclen = 0; struct cifs_ntsd *pntsd = NULL; /* acl obtained from server */ struct cifs_ntsd *pnntsd = NULL; /* modified acl to be sent to server */ cFYI(DBG2, ("set ACL from mode for %s", path)); /* Get the security descriptor */ - pntsd = get_cifs_acl(&acllen, inode, path, NULL); + pntsd = get_cifs_acl(&secdesclen, inode, path, NULL); /* Add three ACEs for owner, group, everyone getting rid of other ACEs as chmod disables ACEs and set the security descriptor */ @@ -709,20 +709,22 @@ int mode_to_acl(struct inode *inode, const char *path, __u64 nmode) set security descriptor request security descriptor parameters, and secuirty descriptor itself */ - pnntsd = kmalloc(acllen, GFP_KERNEL); + secdesclen = secdesclen < DEFSECDESCLEN ? + DEFSECDESCLEN : secdesclen; + pnntsd = kmalloc(secdesclen, GFP_KERNEL); if (!pnntsd) { cERROR(1, ("Unable to allocate security descriptor")); kfree(pntsd); return (-ENOMEM); } - rc = build_sec_desc(pntsd, pnntsd, acllen, inode, nmode); + rc = build_sec_desc(pntsd, pnntsd, inode, nmode); cFYI(DBG2, ("build_sec_desc rc: %d", rc)); if (!rc) { /* Set the security descriptor */ - rc = set_cifs_acl(pnntsd, acllen, inode, path); + rc = set_cifs_acl(pnntsd, secdesclen, inode, path); cFYI(DBG2, ("set_cifs_acl rc: %d", rc)); } diff --git a/fs/cifs/cifsacl.h b/fs/cifs/cifsacl.h index 93a7c3462ea..6c8096cf515 100644 --- a/fs/cifs/cifsacl.h +++ b/fs/cifs/cifsacl.h @@ -27,6 +27,7 @@ #define NUM_SUBAUTHS 5 /* number of sub authority fields */ #define NUM_WK_SIDS 7 /* number of well known sids */ #define SIDNAMELENGTH 20 /* long enough for the ones we care about */ +#define DEFSECDESCLEN 192 /* sec desc len contaiting a dacl with three aces */ #define READ_BIT 0x4 #define WRITE_BIT 0x2 -- cgit v1.2.3 From 16c5f06f15ad4e5a5d6e90b78ffb1ac14319e445 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Wed, 9 Apr 2008 09:33:41 -0400 Subject: [GFS2] fix GFP_KERNEL misuses There are several places where GFP_KERNEL allocations happen under a glock, which will result in hangs if we're under memory pressure and go to re-enter the fs in order to flush stuff out. This patch changes the culprits to GFS_NOFS to keep this problem from happening. Thank you, Signed-off-by: Josef Bacik Signed-off-by: Steven Whitehouse --- fs/gfs2/acl.c | 6 +++--- fs/gfs2/dir.c | 10 +++++----- fs/gfs2/eattr.c | 4 ++-- fs/gfs2/inode.c | 2 +- fs/gfs2/locking/dlm/lock.c | 4 ++-- fs/gfs2/locking/nolock/main.c | 2 +- fs/gfs2/quota.c | 8 ++++---- fs/gfs2/recovery.c | 2 +- fs/gfs2/super.c | 4 ++-- 9 files changed, 21 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/gfs2/acl.c b/fs/gfs2/acl.c index 1047a8c7226..3e9bd46f27e 100644 --- a/fs/gfs2/acl.c +++ b/fs/gfs2/acl.c @@ -116,7 +116,7 @@ static int acl_get(struct gfs2_inode *ip, int access, struct posix_acl **acl, goto out; er.er_data_len = GFS2_EA_DATA_LEN(el->el_ea); - er.er_data = kmalloc(er.er_data_len, GFP_KERNEL); + er.er_data = kmalloc(er.er_data_len, GFP_NOFS); error = -ENOMEM; if (!er.er_data) goto out; @@ -222,7 +222,7 @@ int gfs2_acl_create(struct gfs2_inode *dip, struct gfs2_inode *ip) return error; } - clone = posix_acl_clone(acl, GFP_KERNEL); + clone = posix_acl_clone(acl, GFP_NOFS); error = -ENOMEM; if (!clone) goto out; @@ -272,7 +272,7 @@ int gfs2_acl_chmod(struct gfs2_inode *ip, struct iattr *attr) if (!acl) return gfs2_setattr_simple(ip, attr); - clone = posix_acl_clone(acl, GFP_KERNEL); + clone = posix_acl_clone(acl, GFP_NOFS); error = -ENOMEM; if (!clone) goto out; diff --git a/fs/gfs2/dir.c b/fs/gfs2/dir.c index 94070ad8826..eed040d8ba3 100644 --- a/fs/gfs2/dir.c +++ b/fs/gfs2/dir.c @@ -1089,7 +1089,7 @@ static int dir_double_exhash(struct gfs2_inode *dip) /* Allocate both the "from" and "to" buffers in one big chunk */ - buf = kcalloc(3, sdp->sd_hash_bsize, GFP_KERNEL | __GFP_NOFAIL); + buf = kcalloc(3, sdp->sd_hash_bsize, GFP_NOFS | __GFP_NOFAIL); for (block = dip->i_di.di_size >> sdp->sd_hash_bsize_shift; block--;) { error = gfs2_dir_read_data(dip, (char *)buf, @@ -1378,7 +1378,7 @@ static int dir_e_read(struct inode *inode, u64 *offset, void *opaque, hash = gfs2_dir_offset2hash(*offset); index = hash >> (32 - dip->i_depth); - lp = kmalloc(sdp->sd_hash_bsize, GFP_KERNEL); + lp = kmalloc(sdp->sd_hash_bsize, GFP_NOFS); if (!lp) return -ENOMEM; @@ -1443,7 +1443,7 @@ int gfs2_dir_read(struct inode *inode, u64 *offset, void *opaque, error = -ENOMEM; /* 96 is max number of dirents which can be stuffed into an inode */ - darr = kmalloc(96 * sizeof(struct gfs2_dirent *), GFP_KERNEL); + darr = kmalloc(96 * sizeof(struct gfs2_dirent *), GFP_NOFS); if (darr) { g.pdent = darr; g.offset = 0; @@ -1789,7 +1789,7 @@ static int foreach_leaf(struct gfs2_inode *dip, leaf_call_t lc, void *data) return -EIO; } - lp = kmalloc(sdp->sd_hash_bsize, GFP_KERNEL); + lp = kmalloc(sdp->sd_hash_bsize, GFP_NOFS); if (!lp) return -ENOMEM; @@ -1864,7 +1864,7 @@ static int leaf_dealloc(struct gfs2_inode *dip, u32 index, u32 len, memset(&rlist, 0, sizeof(struct gfs2_rgrp_list)); - ht = kzalloc(size, GFP_KERNEL); + ht = kzalloc(size, GFP_NOFS); if (!ht) return -ENOMEM; diff --git a/fs/gfs2/eattr.c b/fs/gfs2/eattr.c index 81755925a75..e3f76f451b0 100644 --- a/fs/gfs2/eattr.c +++ b/fs/gfs2/eattr.c @@ -448,7 +448,7 @@ static int ea_get_unstuffed(struct gfs2_inode *ip, struct gfs2_ea_header *ea, unsigned int x; int error = 0; - bh = kcalloc(nptrs, sizeof(struct buffer_head *), GFP_KERNEL); + bh = kcalloc(nptrs, sizeof(struct buffer_head *), GFP_NOFS); if (!bh) return -ENOMEM; @@ -1206,7 +1206,7 @@ static int ea_acl_chmod_unstuffed(struct gfs2_inode *ip, unsigned int x; int error; - bh = kcalloc(nptrs, sizeof(struct buffer_head *), GFP_KERNEL); + bh = kcalloc(nptrs, sizeof(struct buffer_head *), GFP_NOFS); if (!bh) return -ENOMEM; diff --git a/fs/gfs2/inode.c b/fs/gfs2/inode.c index dcae2aa83f1..3a9ef526c30 100644 --- a/fs/gfs2/inode.c +++ b/fs/gfs2/inode.c @@ -1231,7 +1231,7 @@ int gfs2_readlinki(struct gfs2_inode *ip, char **buf, unsigned int *len) x = ip->i_di.di_size + 1; if (x > *len) { - *buf = kmalloc(x, GFP_KERNEL); + *buf = kmalloc(x, GFP_NOFS); if (!*buf) { error = -ENOMEM; goto out_brelse; diff --git a/fs/gfs2/locking/dlm/lock.c b/fs/gfs2/locking/dlm/lock.c index 53a6ab3c091..cf7ea8abec8 100644 --- a/fs/gfs2/locking/dlm/lock.c +++ b/fs/gfs2/locking/dlm/lock.c @@ -165,7 +165,7 @@ static int gdlm_create_lp(struct gdlm_ls *ls, struct lm_lockname *name, { struct gdlm_lock *lp; - lp = kzalloc(sizeof(struct gdlm_lock), GFP_KERNEL); + lp = kzalloc(sizeof(struct gdlm_lock), GFP_NOFS); if (!lp) return -ENOMEM; @@ -383,7 +383,7 @@ static int gdlm_add_lvb(struct gdlm_lock *lp) { char *lvb; - lvb = kzalloc(GDLM_LVB_SIZE, GFP_KERNEL); + lvb = kzalloc(GDLM_LVB_SIZE, GFP_NOFS); if (!lvb) return -ENOMEM; diff --git a/fs/gfs2/locking/nolock/main.c b/fs/gfs2/locking/nolock/main.c index d3b8ce6fbbe..284a5ece8d9 100644 --- a/fs/gfs2/locking/nolock/main.c +++ b/fs/gfs2/locking/nolock/main.c @@ -140,7 +140,7 @@ static int nolock_hold_lvb(void *lock, char **lvbp) struct nolock_lockspace *nl = lock; int error = 0; - *lvbp = kzalloc(nl->nl_lvb_size, GFP_KERNEL); + *lvbp = kzalloc(nl->nl_lvb_size, GFP_NOFS); if (!*lvbp) error = -ENOMEM; diff --git a/fs/gfs2/quota.c b/fs/gfs2/quota.c index c71f781db5d..56aaf915c59 100644 --- a/fs/gfs2/quota.c +++ b/fs/gfs2/quota.c @@ -94,7 +94,7 @@ static int qd_alloc(struct gfs2_sbd *sdp, int user, u32 id, struct gfs2_quota_data *qd; int error; - qd = kzalloc(sizeof(struct gfs2_quota_data), GFP_KERNEL); + qd = kzalloc(sizeof(struct gfs2_quota_data), GFP_NOFS); if (!qd) return -ENOMEM; @@ -690,7 +690,7 @@ static int do_sync(unsigned int num_qd, struct gfs2_quota_data **qda) gfs2_write_calc_reserv(ip, sizeof(struct gfs2_quota), &data_blocks, &ind_blocks); - ghs = kcalloc(num_qd, sizeof(struct gfs2_holder), GFP_KERNEL); + ghs = kcalloc(num_qd, sizeof(struct gfs2_holder), GFP_NOFS); if (!ghs) return -ENOMEM; @@ -1118,12 +1118,12 @@ int gfs2_quota_init(struct gfs2_sbd *sdp) error = -ENOMEM; sdp->sd_quota_bitmap = kcalloc(sdp->sd_quota_chunks, - sizeof(unsigned char *), GFP_KERNEL); + sizeof(unsigned char *), GFP_NOFS); if (!sdp->sd_quota_bitmap) return error; for (x = 0; x < sdp->sd_quota_chunks; x++) { - sdp->sd_quota_bitmap[x] = kzalloc(PAGE_SIZE, GFP_KERNEL); + sdp->sd_quota_bitmap[x] = kzalloc(PAGE_SIZE, GFP_NOFS); if (!sdp->sd_quota_bitmap[x]) goto fail; } diff --git a/fs/gfs2/recovery.c b/fs/gfs2/recovery.c index 06dcdc04627..2888e4b4b1c 100644 --- a/fs/gfs2/recovery.c +++ b/fs/gfs2/recovery.c @@ -68,7 +68,7 @@ int gfs2_revoke_add(struct gfs2_sbd *sdp, u64 blkno, unsigned int where) return 0; } - rr = kmalloc(sizeof(struct gfs2_revoke_replay), GFP_KERNEL); + rr = kmalloc(sizeof(struct gfs2_revoke_replay), GFP_NOFS); if (!rr) return -ENOMEM; diff --git a/fs/gfs2/super.c b/fs/gfs2/super.c index 88497b0b733..7aeacbc65f3 100644 --- a/fs/gfs2/super.c +++ b/fs/gfs2/super.c @@ -210,7 +210,7 @@ int gfs2_read_super(struct gfs2_sbd *sdp, sector_t sector) struct page *page; struct bio *bio; - page = alloc_page(GFP_KERNEL); + page = alloc_page(GFP_NOFS); if (unlikely(!page)) return -ENOBUFS; @@ -218,7 +218,7 @@ int gfs2_read_super(struct gfs2_sbd *sdp, sector_t sector) ClearPageDirty(page); lock_page(page); - bio = bio_alloc(GFP_KERNEL, 1); + bio = bio_alloc(GFP_NOFS, 1); if (unlikely(!bio)) { __free_page(page); return -ENOBUFS; -- cgit v1.2.3 From 5d941ca628a142f44d7a2440fe919f8e8691f590 Mon Sep 17 00:00:00 2001 From: Steve French Date: Tue, 15 Apr 2008 18:40:48 +0000 Subject: [CIFS] Fix oops when slow oplock process races with unmount If a tcon is being freed in call tconInfoFree, clean up any entries that may exist in global oplock queue as the tcon structure hanging off of those entries will be invalid and can cause oops while accesing any elements in the tcon structure. Signed-off-by: Shirish Pargaonkar Signed-off-by: Steve French --- fs/cifs/cifsproto.h | 1 + fs/cifs/connect.c | 1 + fs/cifs/transport.c | 18 ++++++++++++++++++ 3 files changed, 20 insertions(+) (limited to 'fs') diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h index 7e5e0e78cd7..0c83da4a7da 100644 --- a/fs/cifs/cifsproto.h +++ b/fs/cifs/cifsproto.h @@ -84,6 +84,7 @@ extern __u16 GetNextMid(struct TCP_Server_Info *server); extern struct oplock_q_entry *AllocOplockQEntry(struct inode *, u16, struct cifsTconInfo *); extern void DeleteOplockQEntry(struct oplock_q_entry *); +extern void DeleteTconOplockQEntries(struct cifsTconInfo *); extern struct timespec cifs_NTtimeToUnix(u64 utc_nanoseconds_since_1601); extern u64 cifs_UnixTimeToNT(struct timespec); extern __le64 cnvrtDosCifsTm(__u16 date, __u16 time); diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c index 8dbfa97cd18..e1710673016 100644 --- a/fs/cifs/connect.c +++ b/fs/cifs/connect.c @@ -3527,6 +3527,7 @@ cifs_umount(struct super_block *sb, struct cifs_sb_info *cifs_sb) FreeXid(xid); return 0; } + DeleteTconOplockQEntries(cifs_sb->tcon); tconInfoFree(cifs_sb->tcon); if ((ses) && (ses->server)) { /* save off task so we do not refer to ses later */ diff --git a/fs/cifs/transport.c b/fs/cifs/transport.c index 3612d6c0a0b..000ac509c98 100644 --- a/fs/cifs/transport.c +++ b/fs/cifs/transport.c @@ -142,6 +142,24 @@ void DeleteOplockQEntry(struct oplock_q_entry *oplockEntry) kmem_cache_free(cifs_oplock_cachep, oplockEntry); } + +void DeleteTconOplockQEntries(struct cifsTconInfo *tcon) +{ + struct oplock_q_entry *temp; + + if (tcon == NULL) + return; + + spin_lock(&GlobalMid_Lock); + list_for_each_entry(temp, &GlobalOplock_Q, qhead) { + if ((temp->tcon) && (temp->tcon == tcon)) { + list_del(&temp->qhead); + kmem_cache_free(cifs_oplock_cachep, temp); + } + } + spin_unlock(&GlobalMid_Lock); +} + int smb_send(struct socket *ssocket, struct smb_hdr *smb_buffer, unsigned int smb_buf_length, struct sockaddr *sin) -- cgit v1.2.3 From 8d142137b4fe87188f211042b16a5993964226f9 Mon Sep 17 00:00:00 2001 From: Steve French Date: Wed, 16 Apr 2008 03:56:51 +0000 Subject: [CIFS] make cifs_dfs_automount_list_static This patch makes the needlessly global cifs_dfs_automount_list static. Signed-off-by: Adrian Bunk Signed-off-by: Steve French --- fs/cifs/cifs_dfs_ref.c | 2 +- fs/cifs/cifsfs.h | 2 -- 2 files changed, 1 insertion(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifs_dfs_ref.c b/fs/cifs/cifs_dfs_ref.c index 56c924033b7..f53f41ff166 100644 --- a/fs/cifs/cifs_dfs_ref.c +++ b/fs/cifs/cifs_dfs_ref.c @@ -23,7 +23,7 @@ #include "dns_resolve.h" #include "cifs_debug.h" -LIST_HEAD(cifs_dfs_automount_list); +static LIST_HEAD(cifs_dfs_automount_list); /* * DFS functions diff --git a/fs/cifs/cifsfs.h b/fs/cifs/cifsfs.h index 68978306c3c..e1dd9f32e1d 100644 --- a/fs/cifs/cifsfs.h +++ b/fs/cifs/cifsfs.h @@ -62,11 +62,9 @@ extern int cifs_setattr(struct dentry *, struct iattr *); extern const struct inode_operations cifs_file_inode_ops; extern const struct inode_operations cifs_symlink_inode_ops; -extern struct list_head cifs_dfs_automount_list; extern struct inode_operations cifs_dfs_referral_inode_operations; - /* Functions related to files and directories */ extern const struct file_operations cifs_file_ops; extern const struct file_operations cifs_file_direct_ops; /* if directio mnt */ -- cgit v1.2.3 From f1f73ba8e9b0eb97f90c6256b94afeb035d97562 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 22 Feb 2008 12:38:02 +0100 Subject: udf: kill udf_set_blocksize This helper has been quite useless since sb_min_blocksize was introduced and is misnamed while we're at it. Just opencode the few lines in the caller instead. Signed-off-by: Christoph Hellwig Signed-off-by: Jan Kara --- fs/udf/super.c | 43 ++++--------------------------------------- 1 file changed, 4 insertions(+), 39 deletions(-) (limited to 'fs') diff --git a/fs/udf/super.c b/fs/udf/super.c index f3ac4abfc94..3723f04c079 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -587,44 +587,6 @@ static int udf_remount_fs(struct super_block *sb, int *flags, char *options) return 0; } -/* - * udf_set_blocksize - * - * PURPOSE - * Set the block size to be used in all transfers. - * - * DESCRIPTION - * To allow room for a DMA transfer, it is best to guess big when unsure. - * This routine picks 2048 bytes as the blocksize when guessing. This - * should be adequate until devices with larger block sizes become common. - * - * Note that the Linux kernel can currently only deal with blocksizes of - * 512, 1024, 2048, 4096, and 8192 bytes. - * - * PRE-CONDITIONS - * sb Pointer to _locked_ superblock. - * - * POST-CONDITIONS - * sb->s_blocksize Blocksize. - * sb->s_blocksize_bits log2 of blocksize. - * 0 Blocksize is valid. - * 1 Blocksize is invalid. - * - * HISTORY - * July 1, 1997 - Andrew E. Mileski - * Written, tested, and released. - */ -static int udf_set_blocksize(struct super_block *sb, int bsize) -{ - if (!sb_min_blocksize(sb, bsize)) { - udf_debug("Bad block size (%d)\n", bsize); - printk(KERN_ERR "udf: bad block size (%d)\n", bsize); - return 0; - } - - return sb->s_blocksize; -} - static int udf_vrs(struct super_block *sb, int silent) { struct volStructDesc *vsd = NULL; @@ -1776,8 +1738,11 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent) sbi->s_nls_map = uopt.nls_map; /* Set the block size for all transfers */ - if (!udf_set_blocksize(sb, uopt.blocksize)) + if (!sb_min_blocksize(sb, uopt.blocksize)) { + udf_debug("Bad block size (%d)\n", uopt.blocksize); + printk(KERN_ERR "udf: bad block size (%d)\n", uopt.blocksize); goto error_out; + } if (uopt.session == 0xFFFFFFFF) sbi->s_session = udf_get_last_session(sb); -- cgit v1.2.3 From b1e321266d8797b21eac433b11458ac65b098938 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 22 Feb 2008 12:38:48 +0100 Subject: udf: kill useless file header comments for vfs method implementations There's not need to document vfs method invocation rules, we have Documentation/filesystems/vfs.txt and Documentation/filesystems/Locking for that. Also a lot of these comments where either plain wrong or horrible out of date. Signed-off-by: Christoph Hellwig Signed-off-by: Jan Kara --- fs/udf/dir.c | 26 -------------------------- fs/udf/file.c | 46 ---------------------------------------------- fs/udf/inode.c | 32 +------------------------------- fs/udf/namei.c | 33 --------------------------------- fs/udf/super.c | 42 ------------------------------------------ 5 files changed, 1 insertion(+), 178 deletions(-) (limited to 'fs') diff --git a/fs/udf/dir.c b/fs/udf/dir.c index 8d8643ada19..31feef916fb 100644 --- a/fs/udf/dir.c +++ b/fs/udf/dir.c @@ -188,32 +188,6 @@ static int do_udf_readdir(struct inode *dir, struct file *filp, return 0; } -/* - * udf_readdir - * - * PURPOSE - * Read a directory entry. - * - * DESCRIPTION - * Optional - sys_getdents() will return -ENOTDIR if this routine is not - * available. - * - * Refer to sys_getdents() in fs/readdir.c - * sys_getdents() -> . - * - * PRE-CONDITIONS - * filp Pointer to directory file. - * buf Pointer to directory entry buffer. - * filldir Pointer to filldir function. - * - * POST-CONDITIONS - * >=0 on success. - * - * HISTORY - * July 1, 1997 - Andrew E. Mileski - * Written, tested, and released. - */ - static int udf_readdir(struct file *filp, void *dirent, filldir_t filldir) { struct inode *dir = filp->f_path.dentry->d_inode; diff --git a/fs/udf/file.c b/fs/udf/file.c index 97c71ae7c68..822cc407152 100644 --- a/fs/udf/file.c +++ b/fs/udf/file.c @@ -144,40 +144,6 @@ static ssize_t udf_file_aio_write(struct kiocb *iocb, const struct iovec *iov, return retval; } -/* - * udf_ioctl - * - * PURPOSE - * Issue an ioctl. - * - * DESCRIPTION - * Optional - sys_ioctl() will return -ENOTTY if this routine is not - * available, and the ioctl cannot be handled without filesystem help. - * - * sys_ioctl() handles these ioctls that apply only to regular files: - * FIBMAP [requires udf_block_map()], FIGETBSZ, FIONREAD - * These ioctls are also handled by sys_ioctl(): - * FIOCLEX, FIONCLEX, FIONBIO, FIOASYNC - * All other ioctls are passed to the filesystem. - * - * Refer to sys_ioctl() in fs/ioctl.c - * sys_ioctl() -> . - * - * PRE-CONDITIONS - * inode Pointer to inode that ioctl was issued on. - * filp Pointer to file that ioctl was issued on. - * cmd The ioctl command. - * arg The ioctl argument [can be interpreted as a - * user-space pointer if desired]. - * - * POST-CONDITIONS - * Success (>=0) or an error code (<=0) that - * sys_ioctl() will return. - * - * HISTORY - * July 1, 1997 - Andrew E. Mileski - * Written, tested, and released. - */ int udf_ioctl(struct inode *inode, struct file *filp, unsigned int cmd, unsigned long arg) { @@ -225,18 +191,6 @@ int udf_ioctl(struct inode *inode, struct file *filp, unsigned int cmd, return result; } -/* - * udf_release_file - * - * PURPOSE - * Called when all references to the file are closed - * - * DESCRIPTION - * Discard prealloced blocks - * - * HISTORY - * - */ static int udf_release_file(struct inode *inode, struct file *filp) { if (filp->f_mode & FMODE_WRITE) { diff --git a/fs/udf/inode.c b/fs/udf/inode.c index 24cfa55d0fd..dc2f946dfca 100644 --- a/fs/udf/inode.c +++ b/fs/udf/inode.c @@ -66,22 +66,7 @@ static void udf_update_extents(struct inode *, struct extent_position *); static int udf_get_block(struct inode *, sector_t, struct buffer_head *, int); -/* - * udf_delete_inode - * - * PURPOSE - * Clean-up before the specified inode is destroyed. - * - * DESCRIPTION - * This routine is called when the kernel destroys an inode structure - * ie. when iput() finds i_count == 0. - * - * HISTORY - * July 1, 1997 - Andrew E. Mileski - * Written, tested, and released. - * - * Called at the last iput() if i_nlink is zero. - */ + void udf_delete_inode(struct inode *inode) { truncate_inode_pages(&inode->i_data, 0); @@ -1416,21 +1401,6 @@ static mode_t udf_convert_permissions(struct fileEntry *fe) return mode; } -/* - * udf_write_inode - * - * PURPOSE - * Write out the specified inode. - * - * DESCRIPTION - * This routine is called whenever an inode is synced. - * Currently this routine is just a placeholder. - * - * HISTORY - * July 1, 1997 - Andrew E. Mileski - * Written, tested, and released. - */ - int udf_write_inode(struct inode *inode, int sync) { int ret; diff --git a/fs/udf/namei.c b/fs/udf/namei.c index 112a5fb0b27..ac7f72779e9 100644 --- a/fs/udf/namei.c +++ b/fs/udf/namei.c @@ -251,39 +251,6 @@ static struct fileIdentDesc *udf_find_entry(struct inode *dir, return NULL; } -/* - * udf_lookup - * - * PURPOSE - * Look-up the inode for a given name. - * - * DESCRIPTION - * Required - lookup_dentry() will return -ENOTDIR if this routine is not - * available for a directory. The filesystem is useless if this routine is - * not available for at least the filesystem's root directory. - * - * This routine is passed an incomplete dentry - it must be completed by - * calling d_add(dentry, inode). If the name does not exist, then the - * specified inode must be set to null. An error should only be returned - * when the lookup fails for a reason other than the name not existing. - * Note that the directory inode semaphore is held during the call. - * - * Refer to lookup_dentry() in fs/namei.c - * lookup_dentry() -> lookup() -> real_lookup() -> . - * - * PRE-CONDITIONS - * dir Pointer to inode of parent directory. - * dentry Pointer to dentry to complete. - * nd Pointer to lookup nameidata - * - * POST-CONDITIONS - * Zero on success. - * - * HISTORY - * July 1, 1997 - Andrew E. Mileski - * Written, tested, and released. - */ - static struct dentry *udf_lookup(struct inode *dir, struct dentry *dentry, struct nameidata *nd) { diff --git a/fs/udf/super.c b/fs/udf/super.c index 3723f04c079..53e1d6e6dd3 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -1670,22 +1670,6 @@ static void udf_sb_free_bitmap(struct udf_bitmap *bitmap) vfree(bitmap); } -/* - * udf_read_super - * - * PURPOSE - * Complete the specified super block. - * - * PRE-CONDITIONS - * sb Pointer to superblock to complete - never NULL. - * sb->s_dev Device to read suberblock from. - * options Pointer to mount options. - * silent Silent flag. - * - * HISTORY - * July 1, 1997 - Andrew E. Mileski - * Written, tested, and released. - */ static int udf_fill_super(struct super_block *sb, void *options, int silent) { int i; @@ -1913,19 +1897,6 @@ void udf_warning(struct super_block *sb, const char *function, sb->s_id, function, error_buf); } -/* - * udf_put_super - * - * PURPOSE - * Prepare for destruction of the superblock. - * - * DESCRIPTION - * Called before the filesystem is unmounted. - * - * HISTORY - * July 1, 1997 - Andrew E. Mileski - * Written, tested, and released. - */ static void udf_put_super(struct super_block *sb) { int i; @@ -1961,19 +1932,6 @@ static void udf_put_super(struct super_block *sb) sb->s_fs_info = NULL; } -/* - * udf_stat_fs - * - * PURPOSE - * Return info about the filesystem. - * - * DESCRIPTION - * Called by sys_statfs() - * - * HISTORY - * July 1, 1997 - Andrew E. Mileski - * Written, tested, and released. - */ static int udf_statfs(struct dentry *dentry, struct kstatfs *buf) { struct super_block *sb = dentry->d_sb; -- cgit v1.2.3 From 15aebd2866b21a568d8defec134bf29f9aea9088 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 22 Feb 2008 12:39:12 +0100 Subject: udf: move headers out include/linux/ There's really no reason to keep udf headers in include/linux as they're not used by anything but fs/udf/. This patch merges most of include/linux/udf_fs_i.h into fs/udf/udf_i.h, include/linux/udf_fs_sb.h into fs/udf/udf_sb.h and include/linux/udf_fs.h into fs/udf/udfdecl.h. The only thing remaining in include/linux/ is a stub of udf_fs_i.h defining the four user-visible udf ioctls. It's also moved from unifdef-y to headers-y because it can be included unconditionally now. Signed-off-by: Christoph Hellwig Signed-off-by: Jan Kara --- fs/udf/file.c | 1 - fs/udf/ialloc.c | 1 - fs/udf/lowlevel.c | 1 - fs/udf/misc.c | 1 - fs/udf/partition.c | 1 - fs/udf/super.c | 1 - fs/udf/symlink.c | 1 - fs/udf/truncate.c | 1 - fs/udf/udf_i.h | 30 +++++++++++++++--- fs/udf/udf_sb.h | 93 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ fs/udf/udfdecl.h | 25 +++++++++++++-- fs/udf/unicode.c | 1 - 12 files changed, 141 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/udf/file.c b/fs/udf/file.c index 822cc407152..0ed6e146a0d 100644 --- a/fs/udf/file.c +++ b/fs/udf/file.c @@ -27,7 +27,6 @@ #include "udfdecl.h" #include -#include #include #include #include /* memset */ diff --git a/fs/udf/ialloc.c b/fs/udf/ialloc.c index 84360315aca..c3fb28eee64 100644 --- a/fs/udf/ialloc.c +++ b/fs/udf/ialloc.c @@ -21,7 +21,6 @@ #include "udfdecl.h" #include #include -#include #include #include diff --git a/fs/udf/lowlevel.c b/fs/udf/lowlevel.c index 579bae71e67..703843f30ff 100644 --- a/fs/udf/lowlevel.c +++ b/fs/udf/lowlevel.c @@ -23,7 +23,6 @@ #include #include -#include #include "udf_sb.h" unsigned int udf_get_last_session(struct super_block *sb) diff --git a/fs/udf/misc.c b/fs/udf/misc.c index a1d6da0caf7..581b6e4cc59 100644 --- a/fs/udf/misc.c +++ b/fs/udf/misc.c @@ -23,7 +23,6 @@ #include #include -#include #include #include "udf_i.h" diff --git a/fs/udf/partition.c b/fs/udf/partition.c index fc533345ab8..307c9c33d18 100644 --- a/fs/udf/partition.c +++ b/fs/udf/partition.c @@ -24,7 +24,6 @@ #include #include -#include #include #include diff --git a/fs/udf/super.c b/fs/udf/super.c index 53e1d6e6dd3..02815e92553 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -57,7 +57,6 @@ #include #include -#include #include "udf_sb.h" #include "udf_i.h" diff --git a/fs/udf/symlink.c b/fs/udf/symlink.c index 6ec99221e50..c3265e1385d 100644 --- a/fs/udf/symlink.c +++ b/fs/udf/symlink.c @@ -23,7 +23,6 @@ #include #include #include -#include #include #include #include diff --git a/fs/udf/truncate.c b/fs/udf/truncate.c index fe61be17cda..6111d97902d 100644 --- a/fs/udf/truncate.c +++ b/fs/udf/truncate.c @@ -22,7 +22,6 @@ #include "udfdecl.h" #include #include -#include #include #include "udf_i.h" diff --git a/fs/udf/udf_i.h b/fs/udf/udf_i.h index ccc52f16bf7..d6d9a774a1c 100644 --- a/fs/udf/udf_i.h +++ b/fs/udf/udf_i.h @@ -1,10 +1,32 @@ -#ifndef __LINUX_UDF_I_H -#define __LINUX_UDF_I_H +#ifndef _UDF_I_H +#define _UDF_I_H + +struct udf_inode_info { + struct timespec i_crtime; + /* Physical address of inode */ + kernel_lb_addr i_location; + __u64 i_unique; + __u32 i_lenEAttr; + __u32 i_lenAlloc; + __u64 i_lenExtents; + __u32 i_next_alloc_block; + __u32 i_next_alloc_goal; + unsigned i_alloc_type : 3; + unsigned i_efe : 1; + unsigned i_use : 1; + unsigned i_strat4096 : 1; + unsigned reserved : 26; + union { + short_ad *i_sad; + long_ad *i_lad; + __u8 *i_data; + } i_ext; + struct inode vfs_inode; +}; -#include static inline struct udf_inode_info *UDF_I(struct inode *inode) { return list_entry(inode, struct udf_inode_info, vfs_inode); } -#endif /* !defined(_LINUX_UDF_I_H) */ +#endif /* _UDF_I_H) */ diff --git a/fs/udf/udf_sb.h b/fs/udf/udf_sb.h index 737d1c604ee..fb2a6e9f8da 100644 --- a/fs/udf/udf_sb.h +++ b/fs/udf/udf_sb.h @@ -1,6 +1,8 @@ #ifndef __LINUX_UDF_SB_H #define __LINUX_UDF_SB_H +#include + /* Since UDF 2.01 is ISO 13346 based... */ #define UDF_SUPER_MAGIC 0x15013346 @@ -38,6 +40,97 @@ #define UDF_PART_FLAG_REWRITABLE 0x0040 #define UDF_PART_FLAG_OVERWRITABLE 0x0080 +#define UDF_MAX_BLOCK_LOADED 8 + +#define UDF_TYPE1_MAP15 0x1511U +#define UDF_VIRTUAL_MAP15 0x1512U +#define UDF_VIRTUAL_MAP20 0x2012U +#define UDF_SPARABLE_MAP15 0x1522U + +#pragma pack(1) /* XXX(hch): Why? This file just defines in-core structures */ + +struct udf_sparing_data { + __u16 s_packet_len; + struct buffer_head *s_spar_map[4]; +}; + +struct udf_virtual_data { + __u32 s_num_entries; + __u16 s_start_offset; +}; + +struct udf_bitmap { + __u32 s_extLength; + __u32 s_extPosition; + __u16 s_nr_groups; + struct buffer_head **s_block_bitmap; +}; + +struct udf_part_map { + union { + struct udf_bitmap *s_bitmap; + struct inode *s_table; + } s_uspace; + union { + struct udf_bitmap *s_bitmap; + struct inode *s_table; + } s_fspace; + __u32 s_partition_root; + __u32 s_partition_len; + __u16 s_partition_type; + __u16 s_partition_num; + union { + struct udf_sparing_data s_sparing; + struct udf_virtual_data s_virtual; + } s_type_specific; + __u32 (*s_partition_func)(struct super_block *, __u32, __u16, __u32); + __u16 s_volumeseqnum; + __u16 s_partition_flags; +}; + +#pragma pack() + +struct udf_sb_info { + struct udf_part_map *s_partmaps; + __u8 s_volume_ident[32]; + + /* Overall info */ + __u16 s_partitions; + __u16 s_partition; + + /* Sector headers */ + __s32 s_session; + __u32 s_anchor[4]; + __u32 s_last_block; + + struct buffer_head *s_lvid_bh; + + /* Default permissions */ + mode_t s_umask; + gid_t s_gid; + uid_t s_uid; + + /* Root Info */ + struct timespec s_record_time; + + /* Fileset Info */ + __u16 s_serial_number; + + /* highest UDF revision we have recorded to this media */ + __u16 s_udfrev; + + /* Miscellaneous flags */ + __u32 s_flags; + + /* Encoding info */ + struct nls_table *s_nls_map; + + /* VAT inode */ + struct inode *s_vat_inode; + + struct mutex s_alloc_mutex; +}; + static inline struct udf_sb_info *UDF_SB(struct super_block *sb) { return sb->s_fs_info; diff --git a/fs/udf/udfdecl.h b/fs/udf/udfdecl.h index 681dc2b66cd..c9c75856af8 100644 --- a/fs/udf/udfdecl.h +++ b/fs/udf/udfdecl.h @@ -1,18 +1,37 @@ #ifndef __UDF_DECL_H #define __UDF_DECL_H -#include #include "ecma_167.h" #include "osta_udf.h" #include #include -#include -#include #include +#include +#include "udf_sb.h" #include "udfend.h" +#define UDF_PREALLOCATE +#define UDF_DEFAULT_PREALLOC_BLOCKS 8 + +#undef UDFFS_DEBUG + +#ifdef UDFFS_DEBUG +#define udf_debug(f, a...) \ +do { \ + printk(KERN_DEBUG "UDF-fs DEBUG %s:%d:%s: ", \ + __FILE__, __LINE__, __func__); \ + printk(f, ##a); \ +} while (0) +#else +#define udf_debug(f, a...) /**/ +#endif + +#define udf_info(f, a...) \ + printk(KERN_INFO "UDF-fs INFO " f, ##a); + + #define udf_fixed_to_variable(x) ( ( ( (x) >> 5 ) * 39 ) + ( (x) & 0x0000001F ) ) #define udf_variable_to_fixed(x) ( ( ( (x) / 39 ) << 5 ) + ( (x) % 39 ) ) diff --git a/fs/udf/unicode.c b/fs/udf/unicode.c index e533b11703b..b9de050ad83 100644 --- a/fs/udf/unicode.c +++ b/fs/udf/unicode.c @@ -23,7 +23,6 @@ #include #include /* for memset */ #include -#include #include "udf_sb.h" -- cgit v1.2.3 From 8dee00bb758f9e1d7fac9f5d2463d09444d4f255 Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Thu, 14 Feb 2008 16:15:45 +0100 Subject: fs/udf: Use DIV_ROUND_UP The kernel.h macro DIV_ROUND_UP performs the computation (((n) + (d) - 1) / (d)) but is perhaps more readable. An extract of the semantic patch that makes this change is as follows: (http://www.emn.fr/x-info/coccinelle/) // @haskernel@ @@ #include @depends on haskernel@ expression n,d; @@ ( - (n + d - 1) / d + DIV_ROUND_UP(n,d) | - (n + (d - 1)) / d + DIV_ROUND_UP(n,d) ) @depends on haskernel@ expression n,d; @@ - DIV_ROUND_UP((n),d) + DIV_ROUND_UP(n,d) @depends on haskernel@ expression n,d; @@ - DIV_ROUND_UP(n,(d)) + DIV_ROUND_UP(n,d) // Signed-off-by: Julia Lawall Signed-off-by: Jan Kara --- fs/udf/super.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/udf/super.c b/fs/udf/super.c index 02815e92553..5f0d400385e 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -986,10 +986,9 @@ static void udf_load_fileset(struct super_block *sb, struct buffer_head *bh, int udf_compute_nr_groups(struct super_block *sb, u32 partition) { struct udf_part_map *map = &UDF_SB(sb)->s_partmaps[partition]; - return (map->s_partition_len + - (sizeof(struct spaceBitmapDesc) << 3) + - (sb->s_blocksize * 8) - 1) / - (sb->s_blocksize * 8); + return DIV_ROUND_UP(map->s_partition_len + + (sizeof(struct spaceBitmapDesc) << 3), + sb->s_blocksize * 8); } static struct udf_bitmap *udf_sb_alloc_bitmap(struct super_block *sb, u32 index) -- cgit v1.2.3 From b8145a769765ce9688eb84080d4c48b22a3553f2 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Sun, 17 Feb 2008 10:19:55 +0200 Subject: make udf_error() static This patch makes the needlessly global udf_error() static. Signed-off-by: Adrian Bunk Signed-off-by: Jan Kara --- fs/udf/super.c | 6 ++++-- fs/udf/udfdecl.h | 1 - 2 files changed, 4 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/udf/super.c b/fs/udf/super.c index 5f0d400385e..2666e5bc69f 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -99,6 +99,8 @@ static void udf_close_lvid(struct super_block *); static unsigned int udf_count_free(struct super_block *); static int udf_statfs(struct dentry *, struct kstatfs *); static int udf_show_options(struct seq_file *, struct vfsmount *); +static void udf_error(struct super_block *sb, const char *function, + const char *fmt, ...); struct logicalVolIntegrityDescImpUse *udf_sb_lvidiu(struct udf_sb_info *sbi) { @@ -1867,8 +1869,8 @@ error_out: return -EINVAL; } -void udf_error(struct super_block *sb, const char *function, - const char *fmt, ...) +static void udf_error(struct super_block *sb, const char *function, + const char *fmt, ...) { va_list args; diff --git a/fs/udf/udfdecl.h b/fs/udf/udfdecl.h index c9c75856af8..d3aae563372 100644 --- a/fs/udf/udfdecl.h +++ b/fs/udf/udfdecl.h @@ -102,7 +102,6 @@ struct extent_position { }; /* super.c */ -extern void udf_error(struct super_block *, const char *, const char *, ...); extern void udf_warning(struct super_block *, const char *, const char *, ...); /* namei.c */ -- cgit v1.2.3 From 79cfe0ff5fb585b92126365a2881262945ac0ee9 Mon Sep 17 00:00:00 2001 From: "marcin.slusarz@gmail.com" Date: Wed, 30 Jan 2008 22:03:51 +0100 Subject: udf: udf_CS0toUTF8 cleanup - fix error handling - always zero output variable - don't zero explicitely fields zeroed by memset - mark "in" paramater as const - remove outdated comment Signed-off-by: Marcin Slusarz Signed-off-by: Jan Kara --- fs/udf/udfdecl.h | 2 +- fs/udf/unicode.c | 27 ++++++++++----------------- 2 files changed, 11 insertions(+), 18 deletions(-) (limited to 'fs') diff --git a/fs/udf/udfdecl.h b/fs/udf/udfdecl.h index d3aae563372..2310b5c8fd8 100644 --- a/fs/udf/udfdecl.h +++ b/fs/udf/udfdecl.h @@ -175,7 +175,7 @@ extern int udf_get_filename(struct super_block *, uint8_t *, uint8_t *, int); extern int udf_put_filename(struct super_block *, const uint8_t *, uint8_t *, int); extern int udf_build_ustr(struct ustr *, dstring *, int); -extern int udf_CS0toUTF8(struct ustr *, struct ustr *); +extern int udf_CS0toUTF8(struct ustr *, const struct ustr *); /* ialloc.c */ extern void udf_free_inode(struct inode *); diff --git a/fs/udf/unicode.c b/fs/udf/unicode.c index b9de050ad83..05bc505ec01 100644 --- a/fs/udf/unicode.c +++ b/fs/udf/unicode.c @@ -82,9 +82,6 @@ static int udf_build_ustr_exact(struct ustr *dest, dstring *ptr, int exactsize) * PURPOSE * Convert OSTA Compressed Unicode to the UTF-8 equivalent. * - * DESCRIPTION - * This routine is only called by udf_filldir(). - * * PRE-CONDITIONS * utf Pointer to UTF-8 output buffer. * ocu Pointer to OSTA Compressed Unicode input buffer @@ -98,43 +95,39 @@ static int udf_build_ustr_exact(struct ustr *dest, dstring *ptr, int exactsize) * November 12, 1997 - Andrew E. Mileski * Written, tested, and released. */ -int udf_CS0toUTF8(struct ustr *utf_o, struct ustr *ocu_i) +int udf_CS0toUTF8(struct ustr *utf_o, const struct ustr *ocu_i) { - uint8_t *ocu; - uint32_t c; + const uint8_t *ocu; uint8_t cmp_id, ocu_len; int i; - ocu = ocu_i->u_name; - ocu_len = ocu_i->u_len; - cmp_id = ocu_i->u_cmpID; - utf_o->u_len = 0; - if (ocu_len == 0) { memset(utf_o, 0, sizeof(struct ustr)); - utf_o->u_cmpID = 0; - utf_o->u_len = 0; return 0; } - if ((cmp_id != 8) && (cmp_id != 16)) { + cmp_id = ocu_i->u_cmpID; + if (cmp_id != 8 && cmp_id != 16) { + memset(utf_o, 0, sizeof(struct ustr)); printk(KERN_ERR "udf: unknown compression code (%d) stri=%s\n", cmp_id, ocu_i->u_name); return 0; } + ocu = ocu_i->u_name; + utf_o->u_len = 0; for (i = 0; (i < ocu_len) && (utf_o->u_len <= (UDF_NAME_LEN - 3));) { /* Expand OSTA compressed Unicode to Unicode */ - c = ocu[i++]; + uint32_t c = ocu[i++]; if (cmp_id == 16) c = (c << 8) | ocu[i++]; /* Compress Unicode to UTF-8 */ - if (c < 0x80U) { + if (c < 0x80U) utf_o->u_name[utf_o->u_len++] = (uint8_t)c; - } else if (c < 0x800U) { + else if (c < 0x800U) { utf_o->u_name[utf_o->u_len++] = (uint8_t)(0xc0 | (c >> 6)); utf_o->u_name[utf_o->u_len++] = -- cgit v1.2.3 From 6305a0a9d559e97807b8bc6d5250fd525dc571a7 Mon Sep 17 00:00:00 2001 From: Marcin Slusarz Date: Mon, 4 Feb 2008 22:27:39 +0100 Subject: udf: fix udf_build_ustr udf_build_ustr was broken: - size == 1: dest->u_len = ptr[1 - 1], but at ptr[0] there's cmpID, so we created string with wrong length it should not happen, so we BUG() it - size > 1 and size < UDF_NAME_LEN: we set u_len correctly, but memcpy copied one needless byte - size == UDF_NAME_LEN - 1: memcpy overwrited u_len - with correct value, but... - size >= UDF_NAME_LEN: we copied UDF_NAME_LEN - 1 bytes, but dest->u_name is array of UDF_NAME_LEN - 2 bytes, so we were overwriting u_len with character from input string nobody noticed because all callers set size to acceptable values (constants within range) Signed-off-by: Marcin Slusarz Signed-off-by: Jan Kara --- fs/udf/unicode.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/udf/unicode.c b/fs/udf/unicode.c index 05bc505ec01..24d6165d21a 100644 --- a/fs/udf/unicode.c +++ b/fs/udf/unicode.c @@ -48,14 +48,16 @@ int udf_build_ustr(struct ustr *dest, dstring *ptr, int size) { int usesize; - if ((!dest) || (!ptr) || (!size)) + if (!dest || !ptr || !size) return -1; + BUG_ON(size < 2); - memset(dest, 0, sizeof(struct ustr)); - usesize = (size > UDF_NAME_LEN) ? UDF_NAME_LEN : size; + usesize = min_t(size_t, ptr[size - 1], sizeof(dest->u_name)); + usesize = min(usesize, size - 2); dest->u_cmpID = ptr[0]; - dest->u_len = ptr[size - 1]; - memcpy(dest->u_name, ptr + 1, usesize - 1); + dest->u_len = usesize; + memcpy(dest->u_name, ptr + 1, usesize); + memset(dest->u_name + usesize, 0, sizeof(dest->u_name) - usesize); return 0; } -- cgit v1.2.3 From 34f953ddfd15da8feb5b8383c93c35dc57202b66 Mon Sep 17 00:00:00 2001 From: "marcin.slusarz@gmail.com" Date: Wed, 27 Feb 2008 22:38:36 +0100 Subject: udf: udf_CS0toNLS cleanup - fix error handling - always zero output variable - don't zero explicitely fields zeroed by memset - mark "in" paramater as const Signed-off-by: Marcin Slusarz Signed-off-by: Jan Kara --- fs/udf/unicode.c | 19 ++++++++----------- 1 file changed, 8 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/udf/unicode.c b/fs/udf/unicode.c index 24d6165d21a..f5872ae325a 100644 --- a/fs/udf/unicode.c +++ b/fs/udf/unicode.c @@ -249,35 +249,32 @@ error_out: } static int udf_CS0toNLS(struct nls_table *nls, struct ustr *utf_o, - struct ustr *ocu_i) + const struct ustr *ocu_i) { - uint8_t *ocu; - uint32_t c; + const uint8_t *ocu; uint8_t cmp_id, ocu_len; int i; - ocu = ocu_i->u_name; ocu_len = ocu_i->u_len; - cmp_id = ocu_i->u_cmpID; - utf_o->u_len = 0; - if (ocu_len == 0) { memset(utf_o, 0, sizeof(struct ustr)); - utf_o->u_cmpID = 0; - utf_o->u_len = 0; return 0; } - if ((cmp_id != 8) && (cmp_id != 16)) { + cmp_id = ocu_i->u_cmpID; + if (cmp_id != 8 && cmp_id != 16) { + memset(utf_o, 0, sizeof(struct ustr)); printk(KERN_ERR "udf: unknown compression code (%d) stri=%s\n", cmp_id, ocu_i->u_name); return 0; } + ocu = ocu_i->u_name; + utf_o->u_len = 0; for (i = 0; (i < ocu_len) && (utf_o->u_len <= (UDF_NAME_LEN - 3));) { /* Expand OSTA compressed Unicode to Unicode */ - c = ocu[i++]; + uint32_t c = ocu[i++]; if (cmp_id == 16) c = (c << 8) | ocu[i++]; -- cgit v1.2.3 From c8ed837d371c24b678182a30e9f0b1f61dee212c Mon Sep 17 00:00:00 2001 From: "marcin.slusarz@gmail.com" Date: Wed, 27 Feb 2008 22:38:38 +0100 Subject: udf: constify crc - constify internal crc table - mark udf_crc "in" parameter as const Signed-off-by: Marcin Slusarz Signed-off-by: Jan Kara --- fs/udf/crc.c | 4 ++-- fs/udf/udfdecl.h | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/udf/crc.c b/fs/udf/crc.c index b1661296e78..f178c63686e 100644 --- a/fs/udf/crc.c +++ b/fs/udf/crc.c @@ -23,7 +23,7 @@ #include "udfdecl.h" -static uint16_t crc_table[256] = { +static const uint16_t crc_table[256] = { 0x0000U, 0x1021U, 0x2042U, 0x3063U, 0x4084U, 0x50a5U, 0x60c6U, 0x70e7U, 0x8108U, 0x9129U, 0xa14aU, 0xb16bU, 0xc18cU, 0xd1adU, 0xe1ceU, 0xf1efU, 0x1231U, 0x0210U, 0x3273U, 0x2252U, 0x52b5U, 0x4294U, 0x72f7U, 0x62d6U, @@ -79,7 +79,7 @@ static uint16_t crc_table[256] = { * July 21, 1997 - Andrew E. Mileski * Adapted from OSTA-UDF(tm) 1.50 standard. */ -uint16_t udf_crc(uint8_t *data, uint32_t size, uint16_t crc) +uint16_t udf_crc(const uint8_t *data, uint32_t size, uint16_t crc) { while (size--) crc = crc_table[(crc >> 8 ^ *(data++)) & 0xffU] ^ (crc << 8); diff --git a/fs/udf/udfdecl.h b/fs/udf/udfdecl.h index 2310b5c8fd8..26bc9c237cb 100644 --- a/fs/udf/udfdecl.h +++ b/fs/udf/udfdecl.h @@ -210,7 +210,7 @@ extern long_ad *udf_get_filelongad(uint8_t *, int, uint32_t *, int); extern short_ad *udf_get_fileshortad(uint8_t *, int, uint32_t *, int); /* crc.c */ -extern uint16_t udf_crc(uint8_t *, uint32_t, uint16_t); +extern uint16_t udf_crc(const uint8_t *, uint32_t, uint16_t); /* udftime.c */ extern time_t *udf_stamp_to_time(time_t *, long *, kernel_timestamp); -- cgit v1.2.3 From 9de90b76eb96e7cdeac8b8dbe7d3db948b070f4d Mon Sep 17 00:00:00 2001 From: "marcin.slusarz@gmail.com" Date: Wed, 30 Jan 2008 22:03:55 +0100 Subject: udf: simple cleanup of truncate.c - remove one indentation level by little code reorganization - convert "if (smth) BUG();" to "BUG_ON(smth);" Signed-off-by: Marcin Slusarz Signed-off-by: Jan Kara --- fs/udf/truncate.c | 76 +++++++++++++++++++++++++------------------------------ 1 file changed, 34 insertions(+), 42 deletions(-) (limited to 'fs') diff --git a/fs/udf/truncate.c b/fs/udf/truncate.c index 6111d97902d..cde328f1677 100644 --- a/fs/udf/truncate.c +++ b/fs/udf/truncate.c @@ -223,34 +223,29 @@ void udf_truncate_extents(struct inode *inode) if (indirect_ext_len) { /* We managed to free all extents in the * indirect extent - free it too */ - if (!epos.bh) - BUG(); + BUG_ON(!epos.bh); udf_free_blocks(sb, inode, epos.block, 0, indirect_ext_len); + } else if (!epos.bh) { + iinfo->i_lenAlloc = lenalloc; + mark_inode_dirty(inode); } else { - if (!epos.bh) { - iinfo->i_lenAlloc = - lenalloc; - mark_inode_dirty(inode); - } else { - struct allocExtDesc *aed = - (struct allocExtDesc *) - (epos.bh->b_data); - int len = - sizeof(struct allocExtDesc); + struct allocExtDesc *aed = + (struct allocExtDesc *) + (epos.bh->b_data); + int len = sizeof(struct allocExtDesc); - aed->lengthAllocDescs = - cpu_to_le32(lenalloc); - if (!UDF_QUERY_FLAG(sb, - UDF_FLAG_STRICT) || - sbi->s_udfrev >= 0x0201) - len += lenalloc; + aed->lengthAllocDescs = + cpu_to_le32(lenalloc); + if (!UDF_QUERY_FLAG(sb, + UDF_FLAG_STRICT) || + sbi->s_udfrev >= 0x0201) + len += lenalloc; - udf_update_tag(epos.bh->b_data, - len); - mark_buffer_dirty_inode( - epos.bh, inode); - } + udf_update_tag(epos.bh->b_data, + len); + mark_buffer_dirty_inode( + epos.bh, inode); } brelse(epos.bh); epos.offset = sizeof(struct allocExtDesc); @@ -271,28 +266,25 @@ void udf_truncate_extents(struct inode *inode) } if (indirect_ext_len) { - if (!epos.bh) - BUG(); + BUG_ON(!epos.bh); udf_free_blocks(sb, inode, epos.block, 0, indirect_ext_len); + } else if (!epos.bh) { + iinfo->i_lenAlloc = lenalloc; + mark_inode_dirty(inode); } else { - if (!epos.bh) { - iinfo->i_lenAlloc = lenalloc; - mark_inode_dirty(inode); - } else { - struct allocExtDesc *aed = - (struct allocExtDesc *)(epos.bh->b_data); - aed->lengthAllocDescs = cpu_to_le32(lenalloc); - if (!UDF_QUERY_FLAG(sb, UDF_FLAG_STRICT) || - sbi->s_udfrev >= 0x0201) - udf_update_tag(epos.bh->b_data, - lenalloc + - sizeof(struct allocExtDesc)); - else - udf_update_tag(epos.bh->b_data, - sizeof(struct allocExtDesc)); - mark_buffer_dirty_inode(epos.bh, inode); - } + struct allocExtDesc *aed = + (struct allocExtDesc *)(epos.bh->b_data); + aed->lengthAllocDescs = cpu_to_le32(lenalloc); + if (!UDF_QUERY_FLAG(sb, UDF_FLAG_STRICT) || + sbi->s_udfrev >= 0x0201) + udf_update_tag(epos.bh->b_data, + lenalloc + + sizeof(struct allocExtDesc)); + else + udf_update_tag(epos.bh->b_data, + sizeof(struct allocExtDesc)); + mark_buffer_dirty_inode(epos.bh, inode); } } else if (inode->i_size) { if (byte_offset) { -- cgit v1.2.3 From 456390de465e5a19c84bca5d78e2550971ab5a96 Mon Sep 17 00:00:00 2001 From: "marcin.slusarz@gmail.com" Date: Wed, 30 Jan 2008 22:03:56 +0100 Subject: udf: truncate: create function for updating of Allocation Ext Descriptor Signed-off-by: Marcin Slusarz Cc: Jan Kara Signed-off-by: Jan Kara --- fs/udf/truncate.c | 56 +++++++++++++++++++++++-------------------------------- 1 file changed, 23 insertions(+), 33 deletions(-) (limited to 'fs') diff --git a/fs/udf/truncate.c b/fs/udf/truncate.c index cde328f1677..65e19b4f942 100644 --- a/fs/udf/truncate.c +++ b/fs/udf/truncate.c @@ -179,6 +179,24 @@ void udf_discard_prealloc(struct inode *inode) brelse(epos.bh); } +static void udf_update_alloc_ext_desc(struct inode *inode, + struct extent_position *epos, + u32 lenalloc) +{ + struct super_block *sb = inode->i_sb; + struct udf_sb_info *sbi = UDF_SB(sb); + + struct allocExtDesc *aed = (struct allocExtDesc *) (epos->bh->b_data); + int len = sizeof(struct allocExtDesc); + + aed->lengthAllocDescs = cpu_to_le32(lenalloc); + if (!UDF_QUERY_FLAG(sb, UDF_FLAG_STRICT) || sbi->s_udfrev >= 0x0201) + len += lenalloc; + + udf_update_tag(epos->bh->b_data, len); + mark_buffer_dirty_inode(epos->bh, inode); +} + void udf_truncate_extents(struct inode *inode) { struct extent_position epos; @@ -186,7 +204,6 @@ void udf_truncate_extents(struct inode *inode) uint32_t elen, nelen = 0, indirect_ext_len = 0, lenalloc; int8_t etype; struct super_block *sb = inode->i_sb; - struct udf_sb_info *sbi = UDF_SB(sb); sector_t first_block = inode->i_size >> sb->s_blocksize_bits, offset; loff_t byte_offset; int adsize; @@ -229,24 +246,9 @@ void udf_truncate_extents(struct inode *inode) } else if (!epos.bh) { iinfo->i_lenAlloc = lenalloc; mark_inode_dirty(inode); - } else { - struct allocExtDesc *aed = - (struct allocExtDesc *) - (epos.bh->b_data); - int len = sizeof(struct allocExtDesc); - - aed->lengthAllocDescs = - cpu_to_le32(lenalloc); - if (!UDF_QUERY_FLAG(sb, - UDF_FLAG_STRICT) || - sbi->s_udfrev >= 0x0201) - len += lenalloc; - - udf_update_tag(epos.bh->b_data, - len); - mark_buffer_dirty_inode( - epos.bh, inode); - } + } else + udf_update_alloc_ext_desc(inode, + &epos, lenalloc); brelse(epos.bh); epos.offset = sizeof(struct allocExtDesc); epos.block = eloc; @@ -272,20 +274,8 @@ void udf_truncate_extents(struct inode *inode) } else if (!epos.bh) { iinfo->i_lenAlloc = lenalloc; mark_inode_dirty(inode); - } else { - struct allocExtDesc *aed = - (struct allocExtDesc *)(epos.bh->b_data); - aed->lengthAllocDescs = cpu_to_le32(lenalloc); - if (!UDF_QUERY_FLAG(sb, UDF_FLAG_STRICT) || - sbi->s_udfrev >= 0x0201) - udf_update_tag(epos.bh->b_data, - lenalloc + - sizeof(struct allocExtDesc)); - else - udf_update_tag(epos.bh->b_data, - sizeof(struct allocExtDesc)); - mark_buffer_dirty_inode(epos.bh, inode); - } + } else + udf_update_alloc_ext_desc(inode, &epos, lenalloc); } else if (inode->i_size) { if (byte_offset) { kernel_long_ad extent; -- cgit v1.2.3 From c2104fda5e6a6981e385b2d11c5c591ab06d82a2 Mon Sep 17 00:00:00 2001 From: "marcin.slusarz@gmail.com" Date: Wed, 30 Jan 2008 22:03:57 +0100 Subject: udf: replace all adds to little endians variables with le*_add_cpu replace all: little_endian_variable = cpu_to_leX(leX_to_cpu(little_endian_variable) + expression_in_cpu_byteorder); with: leX_add_cpu(&little_endian_variable, expression_in_cpu_byteorder); sparse didn't generate any new warning with this patch Signed-off-by: Marcin Slusarz --- fs/udf/balloc.c | 13 ++++--------- fs/udf/ialloc.c | 12 ++++-------- fs/udf/inode.c | 16 ++++------------ 3 files changed, 12 insertions(+), 29 deletions(-) (limited to 'fs') diff --git a/fs/udf/balloc.c b/fs/udf/balloc.c index f855dcbbdfb..1b809bd494b 100644 --- a/fs/udf/balloc.c +++ b/fs/udf/balloc.c @@ -149,8 +149,7 @@ static bool udf_add_free_space(struct udf_sb_info *sbi, return false; lvid = (struct logicalVolIntegrityDesc *)sbi->s_lvid_bh->b_data; - lvid->freeSpaceTable[partition] = cpu_to_le32(le32_to_cpu( - lvid->freeSpaceTable[partition]) + cnt); + le32_add_cpu(&lvid->freeSpaceTable[partition], cnt); return true; } @@ -589,10 +588,8 @@ static void udf_table_free_blocks(struct super_block *sb, sptr = oepos.bh->b_data + epos.offset; aed = (struct allocExtDesc *) oepos.bh->b_data; - aed->lengthAllocDescs = - cpu_to_le32(le32_to_cpu( - aed->lengthAllocDescs) + - adsize); + le32_add_cpu(&aed->lengthAllocDescs, + adsize); } else { sptr = iinfo->i_ext.i_data + epos.offset; @@ -645,9 +642,7 @@ static void udf_table_free_blocks(struct super_block *sb, mark_inode_dirty(table); } else { aed = (struct allocExtDesc *)epos.bh->b_data; - aed->lengthAllocDescs = - cpu_to_le32(le32_to_cpu( - aed->lengthAllocDescs) + adsize); + le32_add_cpu(&aed->lengthAllocDescs, adsize); udf_update_tag(epos.bh->b_data, epos.offset); mark_buffer_dirty(epos.bh); } diff --git a/fs/udf/ialloc.c b/fs/udf/ialloc.c index c3fb28eee64..eb9cfa23dc3 100644 --- a/fs/udf/ialloc.c +++ b/fs/udf/ialloc.c @@ -46,11 +46,9 @@ void udf_free_inode(struct inode *inode) struct logicalVolIntegrityDescImpUse *lvidiu = udf_sb_lvidiu(sbi); if (S_ISDIR(inode->i_mode)) - lvidiu->numDirs = - cpu_to_le32(le32_to_cpu(lvidiu->numDirs) - 1); + le32_add_cpu(&lvidiu->numDirs, -1); else - lvidiu->numFiles = - cpu_to_le32(le32_to_cpu(lvidiu->numFiles) - 1); + le32_add_cpu(&lvidiu->numFiles, -1); mark_buffer_dirty(sbi->s_lvid_bh); } @@ -104,11 +102,9 @@ struct inode *udf_new_inode(struct inode *dir, int mode, int *err) lvhd = (struct logicalVolHeaderDesc *) (lvid->logicalVolContentsUse); if (S_ISDIR(mode)) - lvidiu->numDirs = - cpu_to_le32(le32_to_cpu(lvidiu->numDirs) + 1); + le32_add_cpu(&lvidiu->numDirs, 1); else - lvidiu->numFiles = - cpu_to_le32(le32_to_cpu(lvidiu->numFiles) + 1); + le32_add_cpu(&lvidiu->numFiles, 1); iinfo->i_unique = uniqueID = le64_to_cpu(lvhd->uniqueID); if (!(++uniqueID & 0x00000000FFFFFFFFUL)) uniqueID += 16; diff --git a/fs/udf/inode.c b/fs/udf/inode.c index dc2f946dfca..91d1f1d3d7a 100644 --- a/fs/udf/inode.c +++ b/fs/udf/inode.c @@ -1748,9 +1748,7 @@ int8_t udf_add_aext(struct inode *inode, struct extent_position *epos, if (epos->bh) { aed = (struct allocExtDesc *)epos->bh->b_data; - aed->lengthAllocDescs = - cpu_to_le32(le32_to_cpu( - aed->lengthAllocDescs) + adsize); + le32_add_cpu(&aed->lengthAllocDescs, adsize); } else { iinfo->i_lenAlloc += adsize; mark_inode_dirty(inode); @@ -1800,9 +1798,7 @@ int8_t udf_add_aext(struct inode *inode, struct extent_position *epos, mark_inode_dirty(inode); } else { aed = (struct allocExtDesc *)epos->bh->b_data; - aed->lengthAllocDescs = - cpu_to_le32(le32_to_cpu(aed->lengthAllocDescs) + - adsize); + le32_add_cpu(&aed->lengthAllocDescs, adsize); if (!UDF_QUERY_FLAG(inode->i_sb, UDF_FLAG_STRICT) || UDF_SB(inode->i_sb)->s_udfrev >= 0x0201) udf_update_tag(epos->bh->b_data, @@ -2016,9 +2012,7 @@ int8_t udf_delete_aext(struct inode *inode, struct extent_position epos, mark_inode_dirty(inode); } else { aed = (struct allocExtDesc *)oepos.bh->b_data; - aed->lengthAllocDescs = - cpu_to_le32(le32_to_cpu(aed->lengthAllocDescs) - - (2 * adsize)); + le32_add_cpu(&aed->lengthAllocDescs, -(2 * adsize)); if (!UDF_QUERY_FLAG(inode->i_sb, UDF_FLAG_STRICT) || UDF_SB(inode->i_sb)->s_udfrev >= 0x0201) udf_update_tag(oepos.bh->b_data, @@ -2035,9 +2029,7 @@ int8_t udf_delete_aext(struct inode *inode, struct extent_position epos, mark_inode_dirty(inode); } else { aed = (struct allocExtDesc *)oepos.bh->b_data; - aed->lengthAllocDescs = - cpu_to_le32(le32_to_cpu(aed->lengthAllocDescs) - - adsize); + le32_add_cpu(&aed->lengthAllocDescs, -adsize); if (!UDF_QUERY_FLAG(inode->i_sb, UDF_FLAG_STRICT) || UDF_SB(inode->i_sb)->s_udfrev >= 0x0201) udf_update_tag(oepos.bh->b_data, -- cgit v1.2.3 From 1ab9278570077101d1e367399686be62b22c4019 Mon Sep 17 00:00:00 2001 From: "marcin.slusarz@gmail.com" Date: Wed, 30 Jan 2008 22:03:58 +0100 Subject: udf: simplify __udf_read_inode - move all brelse(ibh) after main if, because it's called on every path except one where ibh is null - move variables to the most inner blocks Signed-off-by: Marcin Slusarz Signed-off-by: Jan Kara --- fs/udf/inode.c | 52 +++++++++++++++++++++++----------------------------- 1 file changed, 23 insertions(+), 29 deletions(-) (limited to 'fs') diff --git a/fs/udf/inode.c b/fs/udf/inode.c index 91d1f1d3d7a..ddd7780d1ab 100644 --- a/fs/udf/inode.c +++ b/fs/udf/inode.c @@ -1101,42 +1101,36 @@ static void __udf_read_inode(struct inode *inode) fe = (struct fileEntry *)bh->b_data; if (fe->icbTag.strategyType == cpu_to_le16(4096)) { - struct buffer_head *ibh = NULL, *nbh = NULL; - struct indirectEntry *ie; + struct buffer_head *ibh; ibh = udf_read_ptagged(inode->i_sb, iinfo->i_location, 1, &ident); - if (ident == TAG_IDENT_IE) { - if (ibh) { - kernel_lb_addr loc; - ie = (struct indirectEntry *)ibh->b_data; - - loc = lelb_to_cpu(ie->indirectICB.extLocation); - - if (ie->indirectICB.extLength && - (nbh = udf_read_ptagged(inode->i_sb, loc, 0, - &ident))) { - if (ident == TAG_IDENT_FE || - ident == TAG_IDENT_EFE) { - memcpy(&iinfo->i_location, - &loc, - sizeof(kernel_lb_addr)); - brelse(bh); - brelse(ibh); - brelse(nbh); - __udf_read_inode(inode); - return; - } else { - brelse(nbh); - brelse(ibh); - } - } else { + if (ident == TAG_IDENT_IE && ibh) { + struct buffer_head *nbh = NULL; + kernel_lb_addr loc; + struct indirectEntry *ie; + + ie = (struct indirectEntry *)ibh->b_data; + loc = lelb_to_cpu(ie->indirectICB.extLocation); + + if (ie->indirectICB.extLength && + (nbh = udf_read_ptagged(inode->i_sb, loc, 0, + &ident))) { + if (ident == TAG_IDENT_FE || + ident == TAG_IDENT_EFE) { + memcpy(&iinfo->i_location, + &loc, + sizeof(kernel_lb_addr)); + brelse(bh); brelse(ibh); + brelse(nbh); + __udf_read_inode(inode); + return; } + brelse(nbh); } - } else { - brelse(ibh); } + brelse(ibh); } else if (fe->icbTag.strategyType != cpu_to_le16(4)) { printk(KERN_ERR "udf: unsupported strategy type: %d\n", le16_to_cpu(fe->icbTag.strategyType)); -- cgit v1.2.3 From d652eefb70142c64495f4188883f9dfc0430a96b Mon Sep 17 00:00:00 2001 From: "marcin.slusarz@gmail.com" Date: Wed, 30 Jan 2008 22:03:59 +0100 Subject: udf: replace udf_*_offset macros with functions - translate udf_file_entry_alloc_offset macro into function - translate udf_ext0_offset macro into function - add comment about crypticly named fields in struct udf_inode_info Signed-off-by: Marcin Slusarz Signed-off-by: Jan Kara --- fs/udf/udf_i.h | 4 ++-- fs/udf/udfdecl.h | 29 +++++++++++++++++++---------- 2 files changed, 21 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/udf/udf_i.h b/fs/udf/udf_i.h index d6d9a774a1c..4f86b1d98a5 100644 --- a/fs/udf/udf_i.h +++ b/fs/udf/udf_i.h @@ -12,8 +12,8 @@ struct udf_inode_info { __u32 i_next_alloc_block; __u32 i_next_alloc_goal; unsigned i_alloc_type : 3; - unsigned i_efe : 1; - unsigned i_use : 1; + unsigned i_efe : 1; /* extendedFileEntry */ + unsigned i_use : 1; /* unallocSpaceEntry */ unsigned i_strat4096 : 1; unsigned reserved : 26; union { diff --git a/fs/udf/udfdecl.h b/fs/udf/udfdecl.h index 26bc9c237cb..cc15f58d497 100644 --- a/fs/udf/udfdecl.h +++ b/fs/udf/udfdecl.h @@ -11,6 +11,7 @@ #include "udf_sb.h" #include "udfend.h" +#include "udf_i.h" #define UDF_PREALLOCATE #define UDF_DEFAULT_PREALLOC_BLOCKS 8 @@ -42,16 +43,24 @@ do { \ #define UDF_NAME_LEN 256 #define UDF_PATH_LEN 1023 -#define udf_file_entry_alloc_offset(inode)\ - (UDF_I(inode)->i_use ?\ - sizeof(struct unallocSpaceEntry) :\ - ((UDF_I(inode)->i_efe ?\ - sizeof(struct extendedFileEntry) :\ - sizeof(struct fileEntry)) + UDF_I(inode)->i_lenEAttr)) - -#define udf_ext0_offset(inode)\ - (UDF_I(inode)->i_alloc_type == ICBTAG_FLAG_AD_IN_ICB ?\ - udf_file_entry_alloc_offset(inode) : 0) +static inline size_t udf_file_entry_alloc_offset(struct inode *inode) +{ + struct udf_inode_info *iinfo = UDF_I(inode); + if (iinfo->i_use) + return sizeof(struct unallocSpaceEntry); + else if (iinfo->i_efe) + return sizeof(struct extendedFileEntry) + iinfo->i_lenEAttr; + else + return sizeof(struct fileEntry) + iinfo->i_lenEAttr; +} + +static inline size_t udf_ext0_offset(struct inode *inode) +{ + if (UDF_I(inode)->i_alloc_type == ICBTAG_FLAG_AD_IN_ICB) + return udf_file_entry_alloc_offset(inode); + else + return 0; +} #define udf_get_lb_pblock(sb,loc,offset) udf_get_pblock((sb), (loc).logicalBlockNum, (loc).partitionReferenceNum, (offset)) -- cgit v1.2.3 From 01b954a36a03d90a66c9dd1fc13e4cb51269caf7 Mon Sep 17 00:00:00 2001 From: Marcin Slusarz Date: Sat, 2 Feb 2008 22:37:07 +0100 Subject: udf: convert udf_count_free_bitmap to use bitmap_weight replace handwritten bits counting with bitmap_weight Signed-off-by: Marcin Slusarz Signed-off-by: Jan Kara --- fs/udf/super.c | 17 +++++------------ 1 file changed, 5 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/udf/super.c b/fs/udf/super.c index 2666e5bc69f..be0aa424b8f 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -55,6 +55,7 @@ #include #include #include +#include #include #include "udf_sb.h" @@ -1958,10 +1959,6 @@ static int udf_statfs(struct dentry *dentry, struct kstatfs *buf) return 0; } -static unsigned char udf_bitmap_lookup[16] = { - 0, 1, 1, 2, 1, 2, 2, 3, 1, 2, 2, 3, 2, 3, 3, 4 -}; - static unsigned int udf_count_free_bitmap(struct super_block *sb, struct udf_bitmap *bitmap) { @@ -1971,7 +1968,6 @@ static unsigned int udf_count_free_bitmap(struct super_block *sb, int block = 0, newblock; kernel_lb_addr loc; uint32_t bytes; - uint8_t value; uint8_t *ptr; uint16_t ident; struct spaceBitmapDesc *bm; @@ -1997,13 +1993,10 @@ static unsigned int udf_count_free_bitmap(struct super_block *sb, ptr = (uint8_t *)bh->b_data; while (bytes > 0) { - while ((bytes > 0) && (index < sb->s_blocksize)) { - value = ptr[index]; - accum += udf_bitmap_lookup[value & 0x0f]; - accum += udf_bitmap_lookup[value >> 4]; - index++; - bytes--; - } + u32 cur_bytes = min_t(u32, bytes, sb->s_blocksize - index); + accum += bitmap_weight((const unsigned long *)(ptr + index), + cur_bytes * 8); + bytes -= cur_bytes; if (bytes) { brelse(bh); newblock = udf_get_lb_pblock(sb, loc, ++block); -- cgit v1.2.3 From f18f17b0338a388ad87a2f8078dbbfb83798bdfd Mon Sep 17 00:00:00 2001 From: "marcin.slusarz@gmail.com" Date: Sun, 3 Feb 2008 19:36:06 +0100 Subject: udf: udf_get_block, inode_bmap - remove unneeded checks block cannot be less than 0, because it's sector_t, so remove unneeded checks Signed-off-by: Marcin Slusarz Signed-off-by: Jan Kara --- fs/udf/inode.c | 12 ------------ 1 file changed, 12 deletions(-) (limited to 'fs') diff --git a/fs/udf/inode.c b/fs/udf/inode.c index ddd7780d1ab..fb92476fcdd 100644 --- a/fs/udf/inode.c +++ b/fs/udf/inode.c @@ -308,9 +308,6 @@ static int udf_get_block(struct inode *inode, sector_t block, lock_kernel(); - if (block < 0) - goto abort_negative; - iinfo = UDF_I(inode); if (block == iinfo->i_next_alloc_block + 1) { iinfo->i_next_alloc_block++; @@ -332,10 +329,6 @@ static int udf_get_block(struct inode *inode, sector_t block, abort: unlock_kernel(); return err; - -abort_negative: - udf_warning(inode->i_sb, "udf_get_block", "block < 0"); - goto abort; } static struct buffer_head *udf_getblk(struct inode *inode, long block, @@ -2051,11 +2044,6 @@ int8_t inode_bmap(struct inode *inode, sector_t block, int8_t etype; struct udf_inode_info *iinfo; - if (block < 0) { - printk(KERN_ERR "udf: inode_bmap: block < 0\n"); - return -1; - } - iinfo = UDF_I(inode); pos->offset = 0; pos->block = iinfo->i_location; -- cgit v1.2.3 From c87e8e90d0da1134e42c89dc89559f4bfe282ef9 Mon Sep 17 00:00:00 2001 From: "marcin.slusarz@gmail.com" Date: Sun, 3 Feb 2008 19:36:07 +0100 Subject: udf: create function for conversion from timestamp to timespec Signed-off-by: Marcin Slusarz Signed-off-by: Jan Kara --- fs/udf/inode.c | 79 +++++++++++++++++----------------------------------------- 1 file changed, 23 insertions(+), 56 deletions(-) (limited to 'fs') diff --git a/fs/udf/inode.c b/fs/udf/inode.c index fb92476fcdd..3c0a60dad47 100644 --- a/fs/udf/inode.c +++ b/fs/udf/inode.c @@ -1136,12 +1136,24 @@ static void __udf_read_inode(struct inode *inode) brelse(bh); } +static void udf_fill_inode_time(struct timespec *tspec, + const timestamp *tstamp, + struct udf_sb_info *sbi) +{ + time_t convtime; + long convtime_usec; + if (udf_stamp_to_time(&convtime, &convtime_usec, + lets_to_cpu(*tstamp))) { + tspec->tv_sec = convtime; + tspec->tv_nsec = convtime_usec * 1000; + } else + *tspec = sbi->s_record_time; +} + static void udf_fill_inode(struct inode *inode, struct buffer_head *bh) { struct fileEntry *fe; struct extendedFileEntry *efe; - time_t convtime; - long convtime_usec; int offset; struct udf_sb_info *sbi = UDF_SB(inode->i_sb); struct udf_inode_info *iinfo = UDF_I(inode); @@ -1229,29 +1241,10 @@ static void udf_fill_inode(struct inode *inode, struct buffer_head *bh) inode->i_blocks = le64_to_cpu(fe->logicalBlocksRecorded) << (inode->i_sb->s_blocksize_bits - 9); - if (udf_stamp_to_time(&convtime, &convtime_usec, - lets_to_cpu(fe->accessTime))) { - inode->i_atime.tv_sec = convtime; - inode->i_atime.tv_nsec = convtime_usec * 1000; - } else { - inode->i_atime = sbi->s_record_time; - } - - if (udf_stamp_to_time(&convtime, &convtime_usec, - lets_to_cpu(fe->modificationTime))) { - inode->i_mtime.tv_sec = convtime; - inode->i_mtime.tv_nsec = convtime_usec * 1000; - } else { - inode->i_mtime = sbi->s_record_time; - } - - if (udf_stamp_to_time(&convtime, &convtime_usec, - lets_to_cpu(fe->attrTime))) { - inode->i_ctime.tv_sec = convtime; - inode->i_ctime.tv_nsec = convtime_usec * 1000; - } else { - inode->i_ctime = sbi->s_record_time; - } + udf_fill_inode_time(&inode->i_atime, &fe->accessTime, sbi); + udf_fill_inode_time(&inode->i_mtime, &fe->modificationTime, + sbi); + udf_fill_inode_time(&inode->i_ctime, &fe->attrTime, sbi); iinfo->i_unique = le64_to_cpu(fe->uniqueID); iinfo->i_lenEAttr = le32_to_cpu(fe->lengthExtendedAttr); @@ -1261,37 +1254,11 @@ static void udf_fill_inode(struct inode *inode, struct buffer_head *bh) inode->i_blocks = le64_to_cpu(efe->logicalBlocksRecorded) << (inode->i_sb->s_blocksize_bits - 9); - if (udf_stamp_to_time(&convtime, &convtime_usec, - lets_to_cpu(efe->accessTime))) { - inode->i_atime.tv_sec = convtime; - inode->i_atime.tv_nsec = convtime_usec * 1000; - } else { - inode->i_atime = sbi->s_record_time; - } - - if (udf_stamp_to_time(&convtime, &convtime_usec, - lets_to_cpu(efe->modificationTime))) { - inode->i_mtime.tv_sec = convtime; - inode->i_mtime.tv_nsec = convtime_usec * 1000; - } else { - inode->i_mtime = sbi->s_record_time; - } - - if (udf_stamp_to_time(&convtime, &convtime_usec, - lets_to_cpu(efe->createTime))) { - iinfo->i_crtime.tv_sec = convtime; - iinfo->i_crtime.tv_nsec = convtime_usec * 1000; - } else { - iinfo->i_crtime = sbi->s_record_time; - } - - if (udf_stamp_to_time(&convtime, &convtime_usec, - lets_to_cpu(efe->attrTime))) { - inode->i_ctime.tv_sec = convtime; - inode->i_ctime.tv_nsec = convtime_usec * 1000; - } else { - inode->i_ctime = sbi->s_record_time; - } + udf_fill_inode_time(&inode->i_atime, &efe->accessTime, sbi); + udf_fill_inode_time(&inode->i_mtime, &efe->modificationTime, + sbi); + udf_fill_inode_time(&iinfo->i_crtime, &efe->createTime, sbi); + udf_fill_inode_time(&inode->i_ctime, &efe->attrTime, sbi); iinfo->i_unique = le64_to_cpu(efe->uniqueID); iinfo->i_lenEAttr = le32_to_cpu(efe->lengthExtendedAttr); -- cgit v1.2.3 From cbf5676a0e0463f05e5073589f3194846dfb02e7 Mon Sep 17 00:00:00 2001 From: "marcin.slusarz@gmail.com" Date: Wed, 27 Feb 2008 22:50:14 +0100 Subject: udf: convert udf_stamp_to_time to return struct timespec Signed-off-by: Marcin Slusarz Signed-off-by: Jan Kara --- fs/udf/inode.c | 49 ++++++++++++++++++++++++++----------------------- fs/udf/super.c | 9 ++------- fs/udf/udfdecl.h | 4 ++-- fs/udf/udftime.c | 17 +++++++---------- 4 files changed, 37 insertions(+), 42 deletions(-) (limited to 'fs') diff --git a/fs/udf/inode.c b/fs/udf/inode.c index 3c0a60dad47..a7646e9bdbd 100644 --- a/fs/udf/inode.c +++ b/fs/udf/inode.c @@ -1136,20 +1136,6 @@ static void __udf_read_inode(struct inode *inode) brelse(bh); } -static void udf_fill_inode_time(struct timespec *tspec, - const timestamp *tstamp, - struct udf_sb_info *sbi) -{ - time_t convtime; - long convtime_usec; - if (udf_stamp_to_time(&convtime, &convtime_usec, - lets_to_cpu(*tstamp))) { - tspec->tv_sec = convtime; - tspec->tv_nsec = convtime_usec * 1000; - } else - *tspec = sbi->s_record_time; -} - static void udf_fill_inode(struct inode *inode, struct buffer_head *bh) { struct fileEntry *fe; @@ -1241,10 +1227,17 @@ static void udf_fill_inode(struct inode *inode, struct buffer_head *bh) inode->i_blocks = le64_to_cpu(fe->logicalBlocksRecorded) << (inode->i_sb->s_blocksize_bits - 9); - udf_fill_inode_time(&inode->i_atime, &fe->accessTime, sbi); - udf_fill_inode_time(&inode->i_mtime, &fe->modificationTime, - sbi); - udf_fill_inode_time(&inode->i_ctime, &fe->attrTime, sbi); + if (!udf_stamp_to_time(&inode->i_atime, + lets_to_cpu(fe->accessTime))) + inode->i_atime = sbi->s_record_time; + + if (!udf_stamp_to_time(&inode->i_mtime, + lets_to_cpu(fe->modificationTime))) + inode->i_mtime = sbi->s_record_time; + + if (!udf_stamp_to_time(&inode->i_ctime, + lets_to_cpu(fe->attrTime))) + inode->i_ctime = sbi->s_record_time; iinfo->i_unique = le64_to_cpu(fe->uniqueID); iinfo->i_lenEAttr = le32_to_cpu(fe->lengthExtendedAttr); @@ -1254,11 +1247,21 @@ static void udf_fill_inode(struct inode *inode, struct buffer_head *bh) inode->i_blocks = le64_to_cpu(efe->logicalBlocksRecorded) << (inode->i_sb->s_blocksize_bits - 9); - udf_fill_inode_time(&inode->i_atime, &efe->accessTime, sbi); - udf_fill_inode_time(&inode->i_mtime, &efe->modificationTime, - sbi); - udf_fill_inode_time(&iinfo->i_crtime, &efe->createTime, sbi); - udf_fill_inode_time(&inode->i_ctime, &efe->attrTime, sbi); + if (!udf_stamp_to_time(&inode->i_atime, + lets_to_cpu(efe->accessTime))) + inode->i_atime = sbi->s_record_time; + + if (!udf_stamp_to_time(&inode->i_mtime, + lets_to_cpu(efe->modificationTime))) + inode->i_mtime = sbi->s_record_time; + + if (!udf_stamp_to_time(&iinfo->i_crtime, + lets_to_cpu(efe->createTime))) + iinfo->i_crtime = sbi->s_record_time; + + if (!udf_stamp_to_time(&inode->i_ctime, + lets_to_cpu(efe->attrTime))) + inode->i_ctime = sbi->s_record_time; iinfo->i_unique = le64_to_cpu(efe->uniqueID); iinfo->i_lenEAttr = le32_to_cpu(efe->lengthExtendedAttr); diff --git a/fs/udf/super.c b/fs/udf/super.c index be0aa424b8f..f4cdd530c65 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -938,24 +938,19 @@ static int udf_find_fileset(struct super_block *sb, static void udf_load_pvoldesc(struct super_block *sb, struct buffer_head *bh) { struct primaryVolDesc *pvoldesc; - time_t recording; - long recording_usec; struct ustr instr; struct ustr outstr; pvoldesc = (struct primaryVolDesc *)bh->b_data; - if (udf_stamp_to_time(&recording, &recording_usec, + if (udf_stamp_to_time(&UDF_SB(sb)->s_record_time, lets_to_cpu(pvoldesc->recordingDateAndTime))) { kernel_timestamp ts; ts = lets_to_cpu(pvoldesc->recordingDateAndTime); - udf_debug("recording time %ld/%ld, %04u/%02u/%02u" + udf_debug("recording time %04u/%02u/%02u" " %02u:%02u (%x)\n", - recording, recording_usec, ts.year, ts.month, ts.day, ts.hour, ts.minute, ts.typeAndTimezone); - UDF_SB(sb)->s_record_time.tv_sec = recording; - UDF_SB(sb)->s_record_time.tv_nsec = recording_usec * 1000; } if (!udf_build_ustr(&instr, pvoldesc->volIdent, 32)) diff --git a/fs/udf/udfdecl.h b/fs/udf/udfdecl.h index cc15f58d497..b277524fe60 100644 --- a/fs/udf/udfdecl.h +++ b/fs/udf/udfdecl.h @@ -222,7 +222,7 @@ extern short_ad *udf_get_fileshortad(uint8_t *, int, uint32_t *, int); extern uint16_t udf_crc(const uint8_t *, uint32_t, uint16_t); /* udftime.c */ -extern time_t *udf_stamp_to_time(time_t *, long *, kernel_timestamp); -extern kernel_timestamp *udf_time_to_stamp(kernel_timestamp *, struct timespec); +extern struct timespec *udf_stamp_to_time(struct timespec *dest, kernel_timestamp src); +extern kernel_timestamp *udf_time_to_stamp(kernel_timestamp *dest, struct timespec src); #endif /* __UDF_DECL_H */ diff --git a/fs/udf/udftime.c b/fs/udf/udftime.c index ce595732ba6..12fae6cd444 100644 --- a/fs/udf/udftime.c +++ b/fs/udf/udftime.c @@ -85,7 +85,7 @@ extern struct timezone sys_tz; #define SECS_PER_HOUR (60 * 60) #define SECS_PER_DAY (SECS_PER_HOUR * 24) -time_t *udf_stamp_to_time(time_t *dest, long *dest_usec, kernel_timestamp src) +struct timespec *udf_stamp_to_time(struct timespec *dest, kernel_timestamp src) { int yday; uint8_t type = src.typeAndTimezone >> 12; @@ -97,23 +97,20 @@ time_t *udf_stamp_to_time(time_t *dest, long *dest_usec, kernel_timestamp src) offset = (offset >> 4); if (offset == -2047) /* unspecified offset */ offset = 0; - } else { + } else offset = 0; - } if ((src.year < EPOCH_YEAR) || (src.year >= EPOCH_YEAR + MAX_YEAR_SECONDS)) { - *dest = -1; - *dest_usec = -1; return NULL; } - *dest = year_seconds[src.year - EPOCH_YEAR]; - *dest -= offset * 60; + dest->tv_sec = year_seconds[src.year - EPOCH_YEAR]; + dest->tv_sec -= offset * 60; yday = ((__mon_yday[__isleap(src.year)][src.month - 1]) + src.day - 1); - *dest += (((yday * 24) + src.hour) * 60 + src.minute) * 60 + src.second; - *dest_usec = src.centiseconds * 10000 + - src.hundredsOfMicroseconds * 100 + src.microseconds; + dest->tv_sec += (((yday * 24) + src.hour) * 60 + src.minute) * 60 + src.second; + dest->tv_nsec = 1000 * (src.centiseconds * 10000 + + src.hundredsOfMicroseconds * 100 + src.microseconds); return dest; } -- cgit v1.2.3 From 56774805d5eeecd3f1fb700603e593a35dc523c8 Mon Sep 17 00:00:00 2001 From: Marcin Slusarz Date: Sun, 10 Feb 2008 11:25:31 +0100 Subject: udf: convert udf_stamp_to_time and udf_time_to_stamp to use timestamps * kernel_timestamp type was almost unused - only callers of udf_stamp_to_time and udf_time_to_stamp used it, so let these functions handle endianness internally and don't clutter code with conversions * rename udf_stamp_to_time to udf_disk_stamp_to_time and udf_time_to_stamp to udf_time_to_disk_stamp Signed-off-by: Marcin Slusarz Signed-off-by: Jan Kara --- fs/udf/inode.c | 45 ++++++++++++++++----------------------------- fs/udf/super.c | 22 ++++++++++------------ fs/udf/udfdecl.h | 5 +++-- fs/udf/udftime.c | 22 ++++++++++++---------- 4 files changed, 41 insertions(+), 53 deletions(-) (limited to 'fs') diff --git a/fs/udf/inode.c b/fs/udf/inode.c index a7646e9bdbd..2362bf0c690 100644 --- a/fs/udf/inode.c +++ b/fs/udf/inode.c @@ -1227,16 +1227,14 @@ static void udf_fill_inode(struct inode *inode, struct buffer_head *bh) inode->i_blocks = le64_to_cpu(fe->logicalBlocksRecorded) << (inode->i_sb->s_blocksize_bits - 9); - if (!udf_stamp_to_time(&inode->i_atime, - lets_to_cpu(fe->accessTime))) + if (!udf_disk_stamp_to_time(&inode->i_atime, fe->accessTime)) inode->i_atime = sbi->s_record_time; - if (!udf_stamp_to_time(&inode->i_mtime, - lets_to_cpu(fe->modificationTime))) + if (!udf_disk_stamp_to_time(&inode->i_mtime, + fe->modificationTime)) inode->i_mtime = sbi->s_record_time; - if (!udf_stamp_to_time(&inode->i_ctime, - lets_to_cpu(fe->attrTime))) + if (!udf_disk_stamp_to_time(&inode->i_ctime, fe->attrTime)) inode->i_ctime = sbi->s_record_time; iinfo->i_unique = le64_to_cpu(fe->uniqueID); @@ -1247,20 +1245,17 @@ static void udf_fill_inode(struct inode *inode, struct buffer_head *bh) inode->i_blocks = le64_to_cpu(efe->logicalBlocksRecorded) << (inode->i_sb->s_blocksize_bits - 9); - if (!udf_stamp_to_time(&inode->i_atime, - lets_to_cpu(efe->accessTime))) + if (!udf_disk_stamp_to_time(&inode->i_atime, efe->accessTime)) inode->i_atime = sbi->s_record_time; - if (!udf_stamp_to_time(&inode->i_mtime, - lets_to_cpu(efe->modificationTime))) + if (!udf_disk_stamp_to_time(&inode->i_mtime, + efe->modificationTime)) inode->i_mtime = sbi->s_record_time; - if (!udf_stamp_to_time(&iinfo->i_crtime, - lets_to_cpu(efe->createTime))) + if (!udf_disk_stamp_to_time(&iinfo->i_crtime, efe->createTime)) iinfo->i_crtime = sbi->s_record_time; - if (!udf_stamp_to_time(&inode->i_ctime, - lets_to_cpu(efe->attrTime))) + if (!udf_disk_stamp_to_time(&inode->i_ctime, efe->attrTime)) inode->i_ctime = sbi->s_record_time; iinfo->i_unique = le64_to_cpu(efe->uniqueID); @@ -1382,7 +1377,6 @@ static int udf_update_inode(struct inode *inode, int do_sync) uint32_t udfperms; uint16_t icbflags; uint16_t crclen; - kernel_timestamp cpu_time; int err = 0; struct udf_sb_info *sbi = UDF_SB(inode->i_sb); unsigned char blocksize_bits = inode->i_sb->s_blocksize_bits; @@ -1485,12 +1479,9 @@ static int udf_update_inode(struct inode *inode, int do_sync) (inode->i_blocks + (1 << (blocksize_bits - 9)) - 1) >> (blocksize_bits - 9)); - if (udf_time_to_stamp(&cpu_time, inode->i_atime)) - fe->accessTime = cpu_to_lets(cpu_time); - if (udf_time_to_stamp(&cpu_time, inode->i_mtime)) - fe->modificationTime = cpu_to_lets(cpu_time); - if (udf_time_to_stamp(&cpu_time, inode->i_ctime)) - fe->attrTime = cpu_to_lets(cpu_time); + udf_time_to_disk_stamp(&fe->accessTime, inode->i_atime); + udf_time_to_disk_stamp(&fe->modificationTime, inode->i_mtime); + udf_time_to_disk_stamp(&fe->attrTime, inode->i_ctime); memset(&(fe->impIdent), 0, sizeof(regid)); strcpy(fe->impIdent.ident, UDF_ID_DEVELOPER); fe->impIdent.identSuffix[0] = UDF_OS_CLASS_UNIX; @@ -1525,14 +1516,10 @@ static int udf_update_inode(struct inode *inode, int do_sync) iinfo->i_crtime.tv_nsec > inode->i_ctime.tv_nsec)) iinfo->i_crtime = inode->i_ctime; - if (udf_time_to_stamp(&cpu_time, inode->i_atime)) - efe->accessTime = cpu_to_lets(cpu_time); - if (udf_time_to_stamp(&cpu_time, inode->i_mtime)) - efe->modificationTime = cpu_to_lets(cpu_time); - if (udf_time_to_stamp(&cpu_time, iinfo->i_crtime)) - efe->createTime = cpu_to_lets(cpu_time); - if (udf_time_to_stamp(&cpu_time, inode->i_ctime)) - efe->attrTime = cpu_to_lets(cpu_time); + udf_time_to_disk_stamp(&efe->accessTime, inode->i_atime); + udf_time_to_disk_stamp(&efe->modificationTime, inode->i_mtime); + udf_time_to_disk_stamp(&efe->createTime, iinfo->i_crtime); + udf_time_to_disk_stamp(&efe->attrTime, inode->i_ctime); memset(&(efe->impIdent), 0, sizeof(regid)); strcpy(efe->impIdent.ident, UDF_ID_DEVELOPER); diff --git a/fs/udf/super.c b/fs/udf/super.c index f4cdd530c65..4d2ecee1970 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -943,8 +943,8 @@ static void udf_load_pvoldesc(struct super_block *sb, struct buffer_head *bh) pvoldesc = (struct primaryVolDesc *)bh->b_data; - if (udf_stamp_to_time(&UDF_SB(sb)->s_record_time, - lets_to_cpu(pvoldesc->recordingDateAndTime))) { + if (udf_disk_stamp_to_time(&UDF_SB(sb)->s_record_time, + pvoldesc->recordingDateAndTime)) { kernel_timestamp ts; ts = lets_to_cpu(pvoldesc->recordingDateAndTime); udf_debug("recording time %04u/%02u/%02u" @@ -1589,7 +1589,6 @@ static void udf_open_lvid(struct super_block *sb) struct udf_sb_info *sbi = UDF_SB(sb); struct buffer_head *bh = sbi->s_lvid_bh; if (bh) { - kernel_timestamp cpu_time; struct logicalVolIntegrityDesc *lvid = (struct logicalVolIntegrityDesc *)bh->b_data; struct logicalVolIntegrityDescImpUse *lvidiu = @@ -1597,8 +1596,8 @@ static void udf_open_lvid(struct super_block *sb) lvidiu->impIdent.identSuffix[0] = UDF_OS_CLASS_UNIX; lvidiu->impIdent.identSuffix[1] = UDF_OS_ID_LINUX; - if (udf_time_to_stamp(&cpu_time, CURRENT_TIME)) - lvid->recordingDateAndTime = cpu_to_lets(cpu_time); + udf_time_to_disk_stamp(&lvid->recordingDateAndTime, + CURRENT_TIME); lvid->integrityType = LVID_INTEGRITY_TYPE_OPEN; lvid->descTag.descCRC = cpu_to_le16( @@ -1613,7 +1612,6 @@ static void udf_open_lvid(struct super_block *sb) static void udf_close_lvid(struct super_block *sb) { - kernel_timestamp cpu_time; struct udf_sb_info *sbi = UDF_SB(sb); struct buffer_head *bh = sbi->s_lvid_bh; struct logicalVolIntegrityDesc *lvid; @@ -1628,8 +1626,8 @@ static void udf_close_lvid(struct super_block *sb) udf_sb_lvidiu(sbi); lvidiu->impIdent.identSuffix[0] = UDF_OS_CLASS_UNIX; lvidiu->impIdent.identSuffix[1] = UDF_OS_ID_LINUX; - if (udf_time_to_stamp(&cpu_time, CURRENT_TIME)) - lvid->recordingDateAndTime = cpu_to_lets(cpu_time); + udf_time_to_disk_stamp(&lvid->recordingDateAndTime, + CURRENT_TIME); if (UDF_MAX_WRITE_VERSION > le16_to_cpu(lvidiu->maxUDFWriteRev)) lvidiu->maxUDFWriteRev = cpu_to_le16(UDF_MAX_WRITE_VERSION); @@ -1801,12 +1799,12 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent) } if (!silent) { - kernel_timestamp ts; - udf_time_to_stamp(&ts, sbi->s_record_time); + timestamp ts; + udf_time_to_disk_stamp(&ts, sbi->s_record_time); udf_info("UDF: Mounting volume '%s', " "timestamp %04u/%02u/%02u %02u:%02u (%x)\n", - sbi->s_volume_ident, ts.year, ts.month, ts.day, - ts.hour, ts.minute, ts.typeAndTimezone); + sbi->s_volume_ident, le16_to_cpu(ts.year), ts.month, ts.day, + ts.hour, ts.minute, le16_to_cpu(ts.typeAndTimezone)); } if (!(sb->s_flags & MS_RDONLY)) udf_open_lvid(sb); diff --git a/fs/udf/udfdecl.h b/fs/udf/udfdecl.h index b277524fe60..2cb2f5de424 100644 --- a/fs/udf/udfdecl.h +++ b/fs/udf/udfdecl.h @@ -222,7 +222,8 @@ extern short_ad *udf_get_fileshortad(uint8_t *, int, uint32_t *, int); extern uint16_t udf_crc(const uint8_t *, uint32_t, uint16_t); /* udftime.c */ -extern struct timespec *udf_stamp_to_time(struct timespec *dest, kernel_timestamp src); -extern kernel_timestamp *udf_time_to_stamp(kernel_timestamp *dest, struct timespec src); +extern struct timespec *udf_disk_stamp_to_time(struct timespec *dest, + timestamp src); +extern timestamp *udf_time_to_disk_stamp(timestamp *dest, struct timespec src); #endif /* __UDF_DECL_H */ diff --git a/fs/udf/udftime.c b/fs/udf/udftime.c index 12fae6cd444..5f811655c9b 100644 --- a/fs/udf/udftime.c +++ b/fs/udf/udftime.c @@ -85,14 +85,16 @@ extern struct timezone sys_tz; #define SECS_PER_HOUR (60 * 60) #define SECS_PER_DAY (SECS_PER_HOUR * 24) -struct timespec *udf_stamp_to_time(struct timespec *dest, kernel_timestamp src) +struct timespec *udf_disk_stamp_to_time(struct timespec *dest, timestamp src) { int yday; - uint8_t type = src.typeAndTimezone >> 12; + u16 typeAndTimezone = le16_to_cpu(src.typeAndTimezone); + u16 year = le16_to_cpu(src.year); + uint8_t type = typeAndTimezone >> 12; int16_t offset; if (type == 1) { - offset = src.typeAndTimezone << 4; + offset = typeAndTimezone << 4; /* sign extent offset */ offset = (offset >> 4); if (offset == -2047) /* unspecified offset */ @@ -100,21 +102,21 @@ struct timespec *udf_stamp_to_time(struct timespec *dest, kernel_timestamp src) } else offset = 0; - if ((src.year < EPOCH_YEAR) || - (src.year >= EPOCH_YEAR + MAX_YEAR_SECONDS)) { + if ((year < EPOCH_YEAR) || + (year >= EPOCH_YEAR + MAX_YEAR_SECONDS)) { return NULL; } - dest->tv_sec = year_seconds[src.year - EPOCH_YEAR]; + dest->tv_sec = year_seconds[year - EPOCH_YEAR]; dest->tv_sec -= offset * 60; - yday = ((__mon_yday[__isleap(src.year)][src.month - 1]) + src.day - 1); + yday = ((__mon_yday[__isleap(year)][src.month - 1]) + src.day - 1); dest->tv_sec += (((yday * 24) + src.hour) * 60 + src.minute) * 60 + src.second; dest->tv_nsec = 1000 * (src.centiseconds * 10000 + src.hundredsOfMicroseconds * 100 + src.microseconds); return dest; } -kernel_timestamp *udf_time_to_stamp(kernel_timestamp *dest, struct timespec ts) +timestamp *udf_time_to_disk_stamp(timestamp *dest, struct timespec ts) { long int days, rem, y; const unsigned short int *ip; @@ -125,7 +127,7 @@ kernel_timestamp *udf_time_to_stamp(kernel_timestamp *dest, struct timespec ts) if (!dest) return NULL; - dest->typeAndTimezone = 0x1000 | (offset & 0x0FFF); + dest->typeAndTimezone = cpu_to_le16(0x1000 | (offset & 0x0FFF)); ts.tv_sec += offset * 60; days = ts.tv_sec / SECS_PER_DAY; @@ -148,7 +150,7 @@ kernel_timestamp *udf_time_to_stamp(kernel_timestamp *dest, struct timespec ts) - LEAPS_THRU_END_OF(y - 1)); y = yg; } - dest->year = y; + dest->year = cpu_to_le16(y); ip = __mon_yday[__isleap(y)]; for (y = 11; days < (long int)ip[y]; --y) continue; -- cgit v1.2.3 From af15a298a49c9b5844cdaf70e10eb808e54ead2c Mon Sep 17 00:00:00 2001 From: Marcin Slusarz Date: Sun, 10 Feb 2008 11:29:10 +0100 Subject: udf: remove unneeded kernel_timestamp type remove now unneeded kernel_timestamp type with conversion functions Signed-off-by: Marcin Slusarz Signed-off-by: Jan Kara --- fs/udf/ecma_167.h | 13 ------------- fs/udf/super.c | 9 +++++---- fs/udf/udfend.h | 22 ---------------------- 3 files changed, 5 insertions(+), 39 deletions(-) (limited to 'fs') diff --git a/fs/udf/ecma_167.h b/fs/udf/ecma_167.h index 56387711589..a0974df82b3 100644 --- a/fs/udf/ecma_167.h +++ b/fs/udf/ecma_167.h @@ -70,19 +70,6 @@ typedef struct { uint8_t microseconds; } __attribute__ ((packed)) timestamp; -typedef struct { - uint16_t typeAndTimezone; - int16_t year; - uint8_t month; - uint8_t day; - uint8_t hour; - uint8_t minute; - uint8_t second; - uint8_t centiseconds; - uint8_t hundredsOfMicroseconds; - uint8_t microseconds; -} __attribute__ ((packed)) kernel_timestamp; - /* Type and Time Zone (ECMA 167r3 1/7.3.1) */ #define TIMESTAMP_TYPE_MASK 0xF000 #define TIMESTAMP_TYPE_CUT 0x0000 diff --git a/fs/udf/super.c b/fs/udf/super.c index 4d2ecee1970..829ddfa5a14 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -945,12 +945,13 @@ static void udf_load_pvoldesc(struct super_block *sb, struct buffer_head *bh) if (udf_disk_stamp_to_time(&UDF_SB(sb)->s_record_time, pvoldesc->recordingDateAndTime)) { - kernel_timestamp ts; - ts = lets_to_cpu(pvoldesc->recordingDateAndTime); +#ifdef UDFFS_DEBUG + timestamp *ts = &pvoldesc->recordingDateAndTime; udf_debug("recording time %04u/%02u/%02u" " %02u:%02u (%x)\n", - ts.year, ts.month, ts.day, ts.hour, - ts.minute, ts.typeAndTimezone); + le16_to_cpu(ts->year), ts->month, ts->day, ts->hour, + ts->minute, le16_to_cpu(ts->typeAndTimezone)); +#endif } if (!udf_build_ustr(&instr, pvoldesc->volIdent, 32)) diff --git a/fs/udf/udfend.h b/fs/udf/udfend.h index c4bd1203f85..489f52fb428 100644 --- a/fs/udf/udfend.h +++ b/fs/udf/udfend.h @@ -24,17 +24,6 @@ static inline lb_addr cpu_to_lelb(kernel_lb_addr in) return out; } -static inline kernel_timestamp lets_to_cpu(timestamp in) -{ - kernel_timestamp out; - - memcpy(&out, &in, sizeof(timestamp)); - out.typeAndTimezone = le16_to_cpu(in.typeAndTimezone); - out.year = le16_to_cpu(in.year); - - return out; -} - static inline short_ad lesa_to_cpu(short_ad in) { short_ad out; @@ -85,15 +74,4 @@ static inline kernel_extent_ad leea_to_cpu(extent_ad in) return out; } -static inline timestamp cpu_to_lets(kernel_timestamp in) -{ - timestamp out; - - memcpy(&out, &in, sizeof(timestamp)); - out.typeAndTimezone = cpu_to_le16(in.typeAndTimezone); - out.year = cpu_to_le16(in.year); - - return out; -} - #endif /* __UDF_ENDIAN_H */ -- cgit v1.2.3 From 165923fa4590b0eb77bec31af383ea16b2d5868f Mon Sep 17 00:00:00 2001 From: Marcin Slusarz Date: Sun, 10 Feb 2008 11:33:08 +0100 Subject: udf: super.c reorganization reorganize few code blocks in super.c which were needlessly indented (and hard to read): so change from: rettype fun() { init; if (sth) { long block of code; } } to: rettype fun() { init; if (!sth) return; long block of code; } or from: rettype fun2() { init; while (sth) { init2(); if (sth2) { long block of code; } } } to: rettype fun2() { init; while (sth) { init2(); if (!sth2) continue; long block of code; } } Signed-off-by: Marcin Slusarz Signed-off-by: Jan Kara --- fs/udf/super.c | 552 ++++++++++++++++++++++++++++----------------------------- 1 file changed, 270 insertions(+), 282 deletions(-) (limited to 'fs') diff --git a/fs/udf/super.c b/fs/udf/super.c index 829ddfa5a14..17ba054faaa 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -827,18 +827,18 @@ static void udf_find_anchor(struct super_block *sb) } for (i = 0; i < ARRAY_SIZE(sbi->s_anchor); i++) { - if (sbi->s_anchor[i]) { - bh = udf_read_tagged(sb, sbi->s_anchor[i], - sbi->s_anchor[i], &ident); - if (!bh) + if (!sbi->s_anchor[i]) + continue; + bh = udf_read_tagged(sb, sbi->s_anchor[i], + sbi->s_anchor[i], &ident); + if (!bh) + sbi->s_anchor[i] = 0; + else { + brelse(bh); + if ((ident != TAG_IDENT_AVDP) && + (i || (ident != TAG_IDENT_FE && + ident != TAG_IDENT_EFE))) sbi->s_anchor[i] = 0; - else { - brelse(bh); - if ((ident != TAG_IDENT_AVDP) && - (i || (ident != TAG_IDENT_FE && - ident != TAG_IDENT_EFE))) - sbi->s_anchor[i] = 0; - } } } @@ -1020,128 +1020,118 @@ static struct udf_bitmap *udf_sb_alloc_bitmap(struct super_block *sb, u32 index) static int udf_load_partdesc(struct super_block *sb, struct buffer_head *bh) { - struct partitionDesc *p; - int i; + struct partitionHeaderDesc *phd; + struct partitionDesc *p = (struct partitionDesc *)bh->b_data; struct udf_part_map *map; - struct udf_sb_info *sbi; - - p = (struct partitionDesc *)bh->b_data; - sbi = UDF_SB(sb); + struct udf_sb_info *sbi = UDF_SB(sb); + bool found = false; + int i; + u16 partitionNumber = le16_to_cpu(p->partitionNumber); for (i = 0; i < sbi->s_partitions; i++) { map = &sbi->s_partmaps[i]; udf_debug("Searching map: (%d == %d)\n", - map->s_partition_num, - le16_to_cpu(p->partitionNumber)); - if (map->s_partition_num == - le16_to_cpu(p->partitionNumber)) { - map->s_partition_len = - le32_to_cpu(p->partitionLength); /* blocks */ - map->s_partition_root = - le32_to_cpu(p->partitionStartingLocation); - if (p->accessType == - cpu_to_le32(PD_ACCESS_TYPE_READ_ONLY)) - map->s_partition_flags |= - UDF_PART_FLAG_READ_ONLY; - if (p->accessType == - cpu_to_le32(PD_ACCESS_TYPE_WRITE_ONCE)) - map->s_partition_flags |= - UDF_PART_FLAG_WRITE_ONCE; - if (p->accessType == - cpu_to_le32(PD_ACCESS_TYPE_REWRITABLE)) - map->s_partition_flags |= - UDF_PART_FLAG_REWRITABLE; - if (p->accessType == - cpu_to_le32(PD_ACCESS_TYPE_OVERWRITABLE)) - map->s_partition_flags |= - UDF_PART_FLAG_OVERWRITABLE; - - if (!strcmp(p->partitionContents.ident, - PD_PARTITION_CONTENTS_NSR02) || - !strcmp(p->partitionContents.ident, - PD_PARTITION_CONTENTS_NSR03)) { - struct partitionHeaderDesc *phd; - - phd = (struct partitionHeaderDesc *) - (p->partitionContentsUse); - if (phd->unallocSpaceTable.extLength) { - kernel_lb_addr loc = { - .logicalBlockNum = le32_to_cpu(phd->unallocSpaceTable.extPosition), - .partitionReferenceNum = i, - }; - - map->s_uspace.s_table = - udf_iget(sb, loc); - if (!map->s_uspace.s_table) { - udf_debug("cannot load unallocSpaceTable (part %d)\n", i); - return 1; - } - map->s_partition_flags |= - UDF_PART_FLAG_UNALLOC_TABLE; - udf_debug("unallocSpaceTable (part %d) @ %ld\n", - i, map->s_uspace.s_table->i_ino); - } - if (phd->unallocSpaceBitmap.extLength) { - struct udf_bitmap *bitmap = - udf_sb_alloc_bitmap(sb, i); - map->s_uspace.s_bitmap = bitmap; - if (bitmap != NULL) { - bitmap->s_extLength = - le32_to_cpu(phd->unallocSpaceBitmap.extLength); - bitmap->s_extPosition = - le32_to_cpu(phd->unallocSpaceBitmap.extPosition); - map->s_partition_flags |= UDF_PART_FLAG_UNALLOC_BITMAP; - udf_debug("unallocSpaceBitmap (part %d) @ %d\n", - i, bitmap->s_extPosition); - } - } - if (phd->partitionIntegrityTable.extLength) - udf_debug("partitionIntegrityTable (part %d)\n", i); - if (phd->freedSpaceTable.extLength) { - kernel_lb_addr loc = { - .logicalBlockNum = le32_to_cpu(phd->freedSpaceTable.extPosition), - .partitionReferenceNum = i, - }; - - map->s_fspace.s_table = - udf_iget(sb, loc); - if (!map->s_fspace.s_table) { - udf_debug("cannot load freedSpaceTable (part %d)\n", i); - return 1; - } - map->s_partition_flags |= - UDF_PART_FLAG_FREED_TABLE; - udf_debug("freedSpaceTable (part %d) @ %ld\n", - i, map->s_fspace.s_table->i_ino); - } - if (phd->freedSpaceBitmap.extLength) { - struct udf_bitmap *bitmap = - udf_sb_alloc_bitmap(sb, i); - map->s_fspace.s_bitmap = bitmap; - if (bitmap != NULL) { - bitmap->s_extLength = - le32_to_cpu(phd->freedSpaceBitmap.extLength); - bitmap->s_extPosition = - le32_to_cpu(phd->freedSpaceBitmap.extPosition); - map->s_partition_flags |= UDF_PART_FLAG_FREED_BITMAP; - udf_debug("freedSpaceBitmap (part %d) @ %d\n", - i, bitmap->s_extPosition); - } - } - } + map->s_partition_num, partitionNumber); + found = map->s_partition_num == partitionNumber; + if (found) break; - } } - if (i == sbi->s_partitions) + + if (!found) { udf_debug("Partition (%d) not found in partition map\n", - le16_to_cpu(p->partitionNumber)); - else - udf_debug("Partition (%d:%d type %x) starts at physical %d, " - "block length %d\n", - le16_to_cpu(p->partitionNumber), i, - map->s_partition_type, - map->s_partition_root, - map->s_partition_len); + partitionNumber); + return 0; + } + + map->s_partition_len = le32_to_cpu(p->partitionLength); /* blocks */ + map->s_partition_root = le32_to_cpu(p->partitionStartingLocation); + + if (p->accessType == cpu_to_le32(PD_ACCESS_TYPE_READ_ONLY)) + map->s_partition_flags |= UDF_PART_FLAG_READ_ONLY; + if (p->accessType == cpu_to_le32(PD_ACCESS_TYPE_WRITE_ONCE)) + map->s_partition_flags |= UDF_PART_FLAG_WRITE_ONCE; + if (p->accessType == cpu_to_le32(PD_ACCESS_TYPE_REWRITABLE)) + map->s_partition_flags |= UDF_PART_FLAG_REWRITABLE; + if (p->accessType == cpu_to_le32(PD_ACCESS_TYPE_OVERWRITABLE)) + map->s_partition_flags |= UDF_PART_FLAG_OVERWRITABLE; + + udf_debug("Partition (%d:%d type %x) starts at physical %d, " + "block length %d\n", partitionNumber, i, + map->s_partition_type, map->s_partition_root, + map->s_partition_len); + + if (strcmp(p->partitionContents.ident, PD_PARTITION_CONTENTS_NSR02) && + strcmp(p->partitionContents.ident, PD_PARTITION_CONTENTS_NSR03)) + return 0; + + phd = (struct partitionHeaderDesc *)p->partitionContentsUse; + if (phd->unallocSpaceTable.extLength) { + kernel_lb_addr loc = { + .logicalBlockNum = le32_to_cpu( + phd->unallocSpaceTable.extPosition), + .partitionReferenceNum = i, + }; + + map->s_uspace.s_table = udf_iget(sb, loc); + if (!map->s_uspace.s_table) { + udf_debug("cannot load unallocSpaceTable (part %d)\n", + i); + return 1; + } + map->s_partition_flags |= UDF_PART_FLAG_UNALLOC_TABLE; + udf_debug("unallocSpaceTable (part %d) @ %ld\n", + i, map->s_uspace.s_table->i_ino); + } + + if (phd->unallocSpaceBitmap.extLength) { + struct udf_bitmap *bitmap = udf_sb_alloc_bitmap(sb, i); + map->s_uspace.s_bitmap = bitmap; + if (bitmap != NULL) { + bitmap->s_extLength = le32_to_cpu( + phd->unallocSpaceBitmap.extLength); + bitmap->s_extPosition = le32_to_cpu( + phd->unallocSpaceBitmap.extPosition); + map->s_partition_flags |= UDF_PART_FLAG_UNALLOC_BITMAP; + udf_debug("unallocSpaceBitmap (part %d) @ %d\n", + i, bitmap->s_extPosition); + } + } + + if (phd->partitionIntegrityTable.extLength) + udf_debug("partitionIntegrityTable (part %d)\n", i); + + if (phd->freedSpaceTable.extLength) { + kernel_lb_addr loc = { + .logicalBlockNum = le32_to_cpu( + phd->freedSpaceTable.extPosition), + .partitionReferenceNum = i, + }; + + map->s_fspace.s_table = udf_iget(sb, loc); + if (!map->s_fspace.s_table) { + udf_debug("cannot load freedSpaceTable (part %d)\n", i); + return 1; + } + + map->s_partition_flags |= UDF_PART_FLAG_FREED_TABLE; + udf_debug("freedSpaceTable (part %d) @ %ld\n", + i, map->s_fspace.s_table->i_ino); + } + + if (phd->freedSpaceBitmap.extLength) { + struct udf_bitmap *bitmap = udf_sb_alloc_bitmap(sb, i); + map->s_fspace.s_bitmap = bitmap; + if (bitmap != NULL) { + bitmap->s_extLength = le32_to_cpu( + phd->freedSpaceBitmap.extLength); + bitmap->s_extPosition = le32_to_cpu( + phd->freedSpaceBitmap.extPosition); + map->s_partition_flags |= UDF_PART_FLAG_FREED_BITMAP; + udf_debug("freedSpaceBitmap (part %d) @ %d\n", + i, bitmap->s_extPosition); + } + } + return 0; } @@ -1215,19 +1205,17 @@ static int udf_load_logicalvol(struct super_block *sb, struct buffer_head *bh, map->s_type_specific.s_sparing. s_spar_map[j] = bh2; - if (bh2 != NULL) { - st = (struct sparingTable *) - bh2->b_data; - if (ident != 0 || strncmp( - st->sparingIdent.ident, - UDF_ID_SPARING, - strlen(UDF_ID_SPARING))) { - brelse(bh2); - map->s_type_specific. - s_sparing. - s_spar_map[j] = - NULL; - } + if (bh2 == NULL) + continue; + + st = (struct sparingTable *)bh2->b_data; + if (ident != 0 || strncmp( + st->sparingIdent.ident, + UDF_ID_SPARING, + strlen(UDF_ID_SPARING))) { + brelse(bh2); + map->s_type_specific.s_sparing. + s_spar_map[j] = NULL; } } map->s_partition_func = udf_get_pblock_spar15; @@ -1392,40 +1380,40 @@ static int udf_process_sequence(struct super_block *sb, long block, brelse(bh); } for (i = 0; i < VDS_POS_LENGTH; i++) { - if (vds[i].block) { - bh = udf_read_tagged(sb, vds[i].block, vds[i].block, - &ident); - - if (i == VDS_POS_PRIMARY_VOL_DESC) { - udf_load_pvoldesc(sb, bh); - } else if (i == VDS_POS_LOGICAL_VOL_DESC) { - if (udf_load_logicalvol(sb, bh, fileset)) { - brelse(bh); - return 1; - } - } else if (i == VDS_POS_PARTITION_DESC) { - struct buffer_head *bh2 = NULL; - if (udf_load_partdesc(sb, bh)) { - brelse(bh); - return 1; - } - for (j = vds[i].block + 1; - j < vds[VDS_POS_TERMINATING_DESC].block; - j++) { - bh2 = udf_read_tagged(sb, j, j, &ident); - gd = (struct generic_desc *)bh2->b_data; - if (ident == TAG_IDENT_PD) - if (udf_load_partdesc(sb, - bh2)) { - brelse(bh); - brelse(bh2); - return 1; - } - brelse(bh2); - } + if (!vds[i].block) + continue; + + bh = udf_read_tagged(sb, vds[i].block, vds[i].block, + &ident); + + if (i == VDS_POS_PRIMARY_VOL_DESC) + udf_load_pvoldesc(sb, bh); + else if (i == VDS_POS_LOGICAL_VOL_DESC) { + if (udf_load_logicalvol(sb, bh, fileset)) { + brelse(bh); + return 1; + } + } else if (i == VDS_POS_PARTITION_DESC) { + struct buffer_head *bh2 = NULL; + if (udf_load_partdesc(sb, bh)) { + brelse(bh); + return 1; + } + for (j = vds[i].block + 1; + j < vds[VDS_POS_TERMINATING_DESC].block; + j++) { + bh2 = udf_read_tagged(sb, j, j, &ident); + gd = (struct generic_desc *)bh2->b_data; + if (ident == TAG_IDENT_PD) + if (udf_load_partdesc(sb, bh2)) { + brelse(bh); + brelse(bh2); + return 1; + } + brelse(bh2); } - brelse(bh); } + brelse(bh); } return 0; @@ -1437,6 +1425,7 @@ static int udf_process_sequence(struct super_block *sb, long block, static int udf_check_valid(struct super_block *sb, int novrs, int silent) { long block; + struct udf_sb_info *sbi; if (novrs) { udf_debug("Validity check skipped because of novrs option\n"); @@ -1444,18 +1433,16 @@ static int udf_check_valid(struct super_block *sb, int novrs, int silent) } /* Check that it is NSR02 compliant */ /* Process any "CD-ROM Volume Descriptor Set" (ECMA 167 2/8.3.1) */ - else { - block = udf_vrs(sb, silent); - if (block == -1) { - struct udf_sb_info *sbi = UDF_SB(sb); - udf_debug("Failed to read byte 32768. Assuming open " - "disc. Skipping validity check\n"); - if (!sbi->s_last_block) - sbi->s_last_block = udf_get_last_block(sb); - return 0; - } else - return !block; - } + block = udf_vrs(sb, silent); + if (block != -1) + return !block; + + sbi = UDF_SB(sb); + udf_debug("Failed to read byte 32768. Assuming open " + "disc. Skipping validity check\n"); + if (!sbi->s_last_block) + sbi->s_last_block = udf_get_last_block(sb); + return 0; } static int udf_load_partition(struct super_block *sb, kernel_lb_addr *fileset) @@ -1474,6 +1461,7 @@ static int udf_load_partition(struct super_block *sb, kernel_lb_addr *fileset) for (i = 0; i < ARRAY_SIZE(sbi->s_anchor); i++) { if (!sbi->s_anchor[i]) continue; + bh = udf_read_tagged(sb, sbi->s_anchor[i], sbi->s_anchor[i], &ident); if (!bh) @@ -1515,72 +1503,73 @@ static int udf_load_partition(struct super_block *sb, kernel_lb_addr *fileset) for (i = 0; i < sbi->s_partitions; i++) { kernel_lb_addr uninitialized_var(ino); struct udf_part_map *map = &sbi->s_partmaps[i]; - switch (map->s_partition_type) { - case UDF_VIRTUAL_MAP15: - case UDF_VIRTUAL_MAP20: - if (!sbi->s_last_block) { - sbi->s_last_block = udf_get_last_block(sb); - udf_find_anchor(sb); - } - if (!sbi->s_last_block) { - udf_debug("Unable to determine Lastblock (For " - "Virtual Partition)\n"); - return 1; - } + if (map->s_partition_type != UDF_VIRTUAL_MAP15 && + map->s_partition_type != UDF_VIRTUAL_MAP20) + continue; - for (j = 0; j < sbi->s_partitions; j++) { - struct udf_part_map *map2 = &sbi->s_partmaps[j]; - if (j != i && - map->s_volumeseqnum == - map2->s_volumeseqnum && - map->s_partition_num == - map2->s_partition_num) { - ino.partitionReferenceNum = j; - ino.logicalBlockNum = - sbi->s_last_block - - map2->s_partition_root; - break; - } + if (!sbi->s_last_block) { + sbi->s_last_block = udf_get_last_block(sb); + udf_find_anchor(sb); + } + + if (!sbi->s_last_block) { + udf_debug("Unable to determine Lastblock (For " + "Virtual Partition)\n"); + return 1; + } + + for (j = 0; j < sbi->s_partitions; j++) { + struct udf_part_map *map2 = &sbi->s_partmaps[j]; + if (j != i && + map->s_volumeseqnum == + map2->s_volumeseqnum && + map->s_partition_num == + map2->s_partition_num) { + ino.partitionReferenceNum = j; + ino.logicalBlockNum = + sbi->s_last_block - + map2->s_partition_root; + break; } + } - if (j == sbi->s_partitions) - return 1; + if (j == sbi->s_partitions) + return 1; - sbi->s_vat_inode = udf_iget(sb, ino); - if (!sbi->s_vat_inode) - return 1; + sbi->s_vat_inode = udf_iget(sb, ino); + if (!sbi->s_vat_inode) + return 1; - if (map->s_partition_type == UDF_VIRTUAL_MAP15) { - map->s_type_specific.s_virtual.s_start_offset = - udf_ext0_offset(sbi->s_vat_inode); - map->s_type_specific.s_virtual.s_num_entries = - (sbi->s_vat_inode->i_size - 36) >> 2; - } else if (map->s_partition_type == UDF_VIRTUAL_MAP20) { - uint32_t pos; - struct virtualAllocationTable20 *vat20; - - pos = udf_block_map(sbi->s_vat_inode, 0); - bh = sb_bread(sb, pos); - if (!bh) - return 1; - vat20 = (struct virtualAllocationTable20 *) - bh->b_data + - udf_ext0_offset(sbi->s_vat_inode); - map->s_type_specific.s_virtual.s_start_offset = - le16_to_cpu(vat20->lengthHeader) + - udf_ext0_offset(sbi->s_vat_inode); - map->s_type_specific.s_virtual.s_num_entries = - (sbi->s_vat_inode->i_size - - map->s_type_specific.s_virtual. - s_start_offset) >> 2; - brelse(bh); - } - map->s_partition_root = udf_get_pblock(sb, 0, i, 0); - map->s_partition_len = - sbi->s_partmaps[ino.partitionReferenceNum]. - s_partition_len; + if (map->s_partition_type == UDF_VIRTUAL_MAP15) { + map->s_type_specific.s_virtual.s_start_offset = + udf_ext0_offset(sbi->s_vat_inode); + map->s_type_specific.s_virtual.s_num_entries = + (sbi->s_vat_inode->i_size - 36) >> 2; + } else if (map->s_partition_type == UDF_VIRTUAL_MAP20) { + uint32_t pos; + struct virtualAllocationTable20 *vat20; + + pos = udf_block_map(sbi->s_vat_inode, 0); + bh = sb_bread(sb, pos); + if (!bh) + return 1; + vat20 = (struct virtualAllocationTable20 *) + bh->b_data + + udf_ext0_offset(sbi->s_vat_inode); + map->s_type_specific.s_virtual.s_start_offset = + le16_to_cpu(vat20->lengthHeader) + + udf_ext0_offset(sbi->s_vat_inode); + map->s_type_specific.s_virtual.s_num_entries = + (sbi->s_vat_inode->i_size - + map->s_type_specific.s_virtual. + s_start_offset) >> 2; + brelse(bh); } + map->s_partition_root = udf_get_pblock(sb, 0, i, 0); + map->s_partition_len = + sbi->s_partmaps[ino.partitionReferenceNum]. + s_partition_len; } return 0; } @@ -1589,26 +1578,26 @@ static void udf_open_lvid(struct super_block *sb) { struct udf_sb_info *sbi = UDF_SB(sb); struct buffer_head *bh = sbi->s_lvid_bh; - if (bh) { - struct logicalVolIntegrityDesc *lvid = - (struct logicalVolIntegrityDesc *)bh->b_data; - struct logicalVolIntegrityDescImpUse *lvidiu = - udf_sb_lvidiu(sbi); + struct logicalVolIntegrityDesc *lvid; + struct logicalVolIntegrityDescImpUse *lvidiu; + if (!bh) + return; - lvidiu->impIdent.identSuffix[0] = UDF_OS_CLASS_UNIX; - lvidiu->impIdent.identSuffix[1] = UDF_OS_ID_LINUX; - udf_time_to_disk_stamp(&lvid->recordingDateAndTime, - CURRENT_TIME); - lvid->integrityType = LVID_INTEGRITY_TYPE_OPEN; + lvid = (struct logicalVolIntegrityDesc *)bh->b_data; + lvidiu = udf_sb_lvidiu(sbi); - lvid->descTag.descCRC = cpu_to_le16( - udf_crc((char *)lvid + sizeof(tag), - le16_to_cpu(lvid->descTag.descCRCLength), - 0)); + lvidiu->impIdent.identSuffix[0] = UDF_OS_CLASS_UNIX; + lvidiu->impIdent.identSuffix[1] = UDF_OS_ID_LINUX; + udf_time_to_disk_stamp(&lvid->recordingDateAndTime, + CURRENT_TIME); + lvid->integrityType = LVID_INTEGRITY_TYPE_OPEN; - lvid->descTag.tagChecksum = udf_tag_checksum(&lvid->descTag); - mark_buffer_dirty(bh); - } + lvid->descTag.descCRC = cpu_to_le16( + udf_crc((char *)lvid + sizeof(tag), + le16_to_cpu(lvid->descTag.descCRCLength), 0)); + + lvid->descTag.tagChecksum = udf_tag_checksum(&lvid->descTag); + mark_buffer_dirty(bh); } static void udf_close_lvid(struct super_block *sb) @@ -1616,36 +1605,35 @@ static void udf_close_lvid(struct super_block *sb) struct udf_sb_info *sbi = UDF_SB(sb); struct buffer_head *bh = sbi->s_lvid_bh; struct logicalVolIntegrityDesc *lvid; + struct logicalVolIntegrityDescImpUse *lvidiu; if (!bh) return; lvid = (struct logicalVolIntegrityDesc *)bh->b_data; - if (lvid->integrityType == LVID_INTEGRITY_TYPE_OPEN) { - struct logicalVolIntegrityDescImpUse *lvidiu = - udf_sb_lvidiu(sbi); - lvidiu->impIdent.identSuffix[0] = UDF_OS_CLASS_UNIX; - lvidiu->impIdent.identSuffix[1] = UDF_OS_ID_LINUX; - udf_time_to_disk_stamp(&lvid->recordingDateAndTime, - CURRENT_TIME); - if (UDF_MAX_WRITE_VERSION > le16_to_cpu(lvidiu->maxUDFWriteRev)) - lvidiu->maxUDFWriteRev = - cpu_to_le16(UDF_MAX_WRITE_VERSION); - if (sbi->s_udfrev > le16_to_cpu(lvidiu->minUDFReadRev)) - lvidiu->minUDFReadRev = cpu_to_le16(sbi->s_udfrev); - if (sbi->s_udfrev > le16_to_cpu(lvidiu->minUDFWriteRev)) - lvidiu->minUDFWriteRev = cpu_to_le16(sbi->s_udfrev); - lvid->integrityType = cpu_to_le32(LVID_INTEGRITY_TYPE_CLOSE); - - lvid->descTag.descCRC = cpu_to_le16( - udf_crc((char *)lvid + sizeof(tag), - le16_to_cpu(lvid->descTag.descCRCLength), - 0)); - - lvid->descTag.tagChecksum = udf_tag_checksum(&lvid->descTag); - mark_buffer_dirty(bh); - } + if (lvid->integrityType != LVID_INTEGRITY_TYPE_OPEN) + return; + + lvidiu = udf_sb_lvidiu(sbi); + lvidiu->impIdent.identSuffix[0] = UDF_OS_CLASS_UNIX; + lvidiu->impIdent.identSuffix[1] = UDF_OS_ID_LINUX; + udf_time_to_disk_stamp(&lvid->recordingDateAndTime, CURRENT_TIME); + if (UDF_MAX_WRITE_VERSION > le16_to_cpu(lvidiu->maxUDFWriteRev)) + lvidiu->maxUDFWriteRev = cpu_to_le16(UDF_MAX_WRITE_VERSION); + if (sbi->s_udfrev > le16_to_cpu(lvidiu->minUDFReadRev)) + lvidiu->minUDFReadRev = cpu_to_le16(sbi->s_udfrev); + if (sbi->s_udfrev > le16_to_cpu(lvidiu->minUDFWriteRev)) + lvidiu->minUDFWriteRev = cpu_to_le16(sbi->s_udfrev); + lvid->integrityType = cpu_to_le32(LVID_INTEGRITY_TYPE_CLOSE); + + lvid->descTag.descCRC = cpu_to_le16( + udf_crc((char *)lvid + sizeof(tag), + le16_to_cpu(lvid->descTag.descCRCLength), + 0)); + + lvid->descTag.tagChecksum = udf_tag_checksum(&lvid->descTag); + mark_buffer_dirty(bh); } static void udf_sb_free_bitmap(struct udf_bitmap *bitmap) -- cgit v1.2.3 From 200a3592cda7bfc010d30c0b1eb8eff3791e9ba9 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Tue, 4 Mar 2008 13:03:09 +0100 Subject: udf: Mark udf_process_sequence() as noinline Mark udf_process_sequence() as noinline since stack usage is terrible otherwise. Signed-off-by: Jan Kara --- fs/udf/super.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/udf/super.c b/fs/udf/super.c index 17ba054faaa..bb8d866deb3 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -1292,7 +1292,7 @@ static void udf_load_logicalvolint(struct super_block *sb, kernel_extent_ad loc) * July 1, 1997 - Andrew E. Mileski * Written, tested, and released. */ -static int udf_process_sequence(struct super_block *sb, long block, +static noinline int udf_process_sequence(struct super_block *sb, long block, long lastblock, kernel_lb_addr *fileset) { struct buffer_head *bh = NULL; -- cgit v1.2.3 From 9bf2c6b834f4caad82b3e2d962c266153d39e411 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Tue, 4 Mar 2008 13:10:29 +0100 Subject: udf: Remove checking of existence of filename in udf_add_entry() We don't have to check whether a directory entry already exists in a directory when creating a new one since we've already checked that earlier by lookup and we are holding directory i_mutex all the time. Signed-off-by: Jan Kara --- fs/udf/namei.c | 38 +------------------------------------- 1 file changed, 1 insertion(+), 37 deletions(-) (limited to 'fs') diff --git a/fs/udf/namei.c b/fs/udf/namei.c index ac7f72779e9..e9f587201e1 100644 --- a/fs/udf/namei.c +++ b/fs/udf/namei.c @@ -303,11 +303,9 @@ static struct fileIdentDesc *udf_add_entry(struct inode *dir, { struct super_block *sb = dir->i_sb; struct fileIdentDesc *fi = NULL; - char name[UDF_NAME_LEN], fname[UDF_NAME_LEN]; + char name[UDF_NAME_LEN]; int namelen; loff_t f_pos; - int flen; - char *nameptr; loff_t size = udf_ext0_offset(dir) + dir->i_size; int nfidlen; uint8_t lfi; @@ -385,26 +383,6 @@ static struct fileIdentDesc *udf_add_entry(struct inode *dir, liu = le16_to_cpu(cfi->lengthOfImpUse); lfi = cfi->lengthFileIdent; - if (fibh->sbh == fibh->ebh) - nameptr = fi->fileIdent + liu; - else { - int poffset; /* Unpaded ending offset */ - - poffset = fibh->soffset + sizeof(struct fileIdentDesc) + - liu + lfi; - - if (poffset >= lfi) - nameptr = (char *)(fibh->ebh->b_data + - poffset - lfi); - else { - nameptr = fname; - memcpy(nameptr, fi->fileIdent + liu, - lfi - poffset); - memcpy(nameptr + lfi - poffset, - fibh->ebh->b_data, poffset); - } - } - if ((cfi->fileCharacteristics & FID_FILE_CHAR_DELETED) != 0) { if (((sizeof(struct fileIdentDesc) + liu + lfi + 3) & ~3) == nfidlen) { @@ -423,20 +401,6 @@ static struct fileIdentDesc *udf_add_entry(struct inode *dir, } } } - - if (!lfi || !dentry) - continue; - - flen = udf_get_filename(dir->i_sb, nameptr, fname, lfi); - if (flen && udf_match(flen, fname, dentry->d_name.len, - dentry->d_name.name)) { - if (fibh->sbh != fibh->ebh) - brelse(fibh->ebh); - brelse(fibh->sbh); - brelse(epos.bh); - *err = -EEXIST; - return NULL; - } } add: -- cgit v1.2.3 From b80697c14dcacd83ed1b78e26ad93b25ecc52c5e Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Tue, 4 Mar 2008 14:14:05 +0100 Subject: udf: Remove declarations of arrays of size UDF_NAME_LEN (256 bytes) There are several places in UDF where we declared temporary arrays of UDF_NAME_LEN bytes on stack. This is not nice to stack usage so this patch changes those places to use kmalloc() instead. Also clean up bail-out paths in those functions when we are changing them. Signed-off-by: Jan Kara --- fs/udf/dir.c | 57 ++++++++++++------------- fs/udf/namei.c | 128 +++++++++++++++++++++++++++++---------------------------- 2 files changed, 93 insertions(+), 92 deletions(-) (limited to 'fs') diff --git a/fs/udf/dir.c b/fs/udf/dir.c index 31feef916fb..62dc270c69d 100644 --- a/fs/udf/dir.c +++ b/fs/udf/dir.c @@ -39,13 +39,13 @@ static int do_udf_readdir(struct inode *dir, struct file *filp, filldir_t filldir, void *dirent) { - struct udf_fileident_bh fibh; + struct udf_fileident_bh fibh = { .sbh = NULL, .ebh = NULL}; struct fileIdentDesc *fi = NULL; struct fileIdentDesc cfi; int block, iblock; loff_t nf_pos = (filp->f_pos - 1) << 2; int flen; - char fname[UDF_NAME_LEN]; + char *fname = NULL; char *nameptr; uint16_t liu; uint8_t lfi; @@ -54,23 +54,32 @@ static int do_udf_readdir(struct inode *dir, struct file *filp, kernel_lb_addr eloc; uint32_t elen; sector_t offset; - int i, num; + int i, num, ret = 0; unsigned int dt_type; struct extent_position epos = { NULL, 0, {0, 0} }; struct udf_inode_info *iinfo; if (nf_pos >= size) - return 0; + goto out; + + fname = kmalloc(UDF_NAME_LEN, GFP_NOFS); + if (!fname) { + ret = -ENOMEM; + goto out; + } if (nf_pos == 0) nf_pos = udf_ext0_offset(dir); fibh.soffset = fibh.eoffset = nf_pos & (dir->i_sb->s_blocksize - 1); iinfo = UDF_I(dir); - if (iinfo->i_alloc_type == ICBTAG_FLAG_AD_IN_ICB) { - fibh.sbh = fibh.ebh = NULL; - } else if (inode_bmap(dir, nf_pos >> dir->i_sb->s_blocksize_bits, - &epos, &eloc, &elen, &offset) == (EXT_RECORDED_ALLOCATED >> 30)) { + if (iinfo->i_alloc_type != ICBTAG_FLAG_AD_IN_ICB) { + if (inode_bmap(dir, nf_pos >> dir->i_sb->s_blocksize_bits, + &epos, &eloc, &elen, &offset) + != (EXT_RECORDED_ALLOCATED >> 30)) { + ret = -ENOENT; + goto out; + } block = udf_get_lb_pblock(dir->i_sb, eloc, offset); if ((++offset << dir->i_sb->s_blocksize_bits) < elen) { if (iinfo->i_alloc_type == ICBTAG_FLAG_AD_SHORT) @@ -83,8 +92,8 @@ static int do_udf_readdir(struct inode *dir, struct file *filp, } if (!(fibh.sbh = fibh.ebh = udf_tread(dir->i_sb, block))) { - brelse(epos.bh); - return -EIO; + ret = -EIO; + goto out; } if (!(offset & ((16 >> (dir->i_sb->s_blocksize_bits - 9)) - 1))) { @@ -105,9 +114,6 @@ static int do_udf_readdir(struct inode *dir, struct file *filp, brelse(bha[i]); } } - } else { - brelse(epos.bh); - return -ENOENT; } while (nf_pos < size) { @@ -115,13 +121,8 @@ static int do_udf_readdir(struct inode *dir, struct file *filp, fi = udf_fileident_read(dir, &nf_pos, &fibh, &cfi, &epos, &eloc, &elen, &offset); - if (!fi) { - if (fibh.sbh != fibh.ebh) - brelse(fibh.ebh); - brelse(fibh.sbh); - brelse(epos.bh); - return 0; - } + if (!fi) + goto out; liu = le16_to_cpu(cfi.lengthOfImpUse); lfi = cfi.lengthFileIdent; @@ -167,25 +168,21 @@ static int do_udf_readdir(struct inode *dir, struct file *filp, dt_type = DT_UNKNOWN; } - if (flen) { - if (filldir(dirent, fname, flen, filp->f_pos, iblock, dt_type) < 0) { - if (fibh.sbh != fibh.ebh) - brelse(fibh.ebh); - brelse(fibh.sbh); - brelse(epos.bh); - return 0; - } - } + if (flen && filldir(dirent, fname, flen, filp->f_pos, + iblock, dt_type) < 0) + goto out; } /* end while */ filp->f_pos = (nf_pos >> 2) + 1; +out: if (fibh.sbh != fibh.ebh) brelse(fibh.ebh); brelse(fibh.sbh); brelse(epos.bh); + kfree(fname); - return 0; + return ret; } static int udf_readdir(struct file *filp, void *dirent, filldir_t filldir) diff --git a/fs/udf/namei.c b/fs/udf/namei.c index e9f587201e1..68686b79650 100644 --- a/fs/udf/namei.c +++ b/fs/udf/namei.c @@ -149,7 +149,7 @@ static struct fileIdentDesc *udf_find_entry(struct inode *dir, struct fileIdentDesc *fi = NULL; loff_t f_pos; int block, flen; - char fname[UDF_NAME_LEN]; + char *fname = NULL; char *nameptr; uint8_t lfi; uint16_t liu; @@ -163,12 +163,12 @@ static struct fileIdentDesc *udf_find_entry(struct inode *dir, size = udf_ext0_offset(dir) + dir->i_size; f_pos = udf_ext0_offset(dir); + fibh->sbh = fibh->ebh = NULL; fibh->soffset = fibh->eoffset = f_pos & (dir->i_sb->s_blocksize - 1); - if (dinfo->i_alloc_type == ICBTAG_FLAG_AD_IN_ICB) - fibh->sbh = fibh->ebh = NULL; - else if (inode_bmap(dir, f_pos >> dir->i_sb->s_blocksize_bits, - &epos, &eloc, &elen, &offset) == - (EXT_RECORDED_ALLOCATED >> 30)) { + if (dinfo->i_alloc_type != ICBTAG_FLAG_AD_IN_ICB) { + if (inode_bmap(dir, f_pos >> dir->i_sb->s_blocksize_bits, &epos, + &eloc, &elen, &offset) != (EXT_RECORDED_ALLOCATED >> 30)) + goto out_err; block = udf_get_lb_pblock(dir->i_sb, eloc, offset); if ((++offset << dir->i_sb->s_blocksize_bits) < elen) { if (dinfo->i_alloc_type == ICBTAG_FLAG_AD_SHORT) @@ -179,25 +179,19 @@ static struct fileIdentDesc *udf_find_entry(struct inode *dir, offset = 0; fibh->sbh = fibh->ebh = udf_tread(dir->i_sb, block); - if (!fibh->sbh) { - brelse(epos.bh); - return NULL; - } - } else { - brelse(epos.bh); - return NULL; + if (!fibh->sbh) + goto out_err; } + fname = kmalloc(UDF_NAME_LEN, GFP_NOFS); + if (!fname) + goto out_err; + while (f_pos < size) { fi = udf_fileident_read(dir, &f_pos, fibh, cfi, &epos, &eloc, &elen, &offset); - if (!fi) { - if (fibh->sbh != fibh->ebh) - brelse(fibh->ebh); - brelse(fibh->sbh); - brelse(epos.bh); - return NULL; - } + if (!fi) + goto out_err; liu = le16_to_cpu(cfi->lengthOfImpUse); lfi = cfi->lengthFileIdent; @@ -237,18 +231,20 @@ static struct fileIdentDesc *udf_find_entry(struct inode *dir, flen = udf_get_filename(dir->i_sb, nameptr, fname, lfi); if (flen && udf_match(flen, fname, dentry->d_name.len, - dentry->d_name.name)) { - brelse(epos.bh); - return fi; - } + dentry->d_name.name)) + goto out_ok; } +out_err: + fi = NULL; if (fibh->sbh != fibh->ebh) brelse(fibh->ebh); brelse(fibh->sbh); +out_ok: brelse(epos.bh); + kfree(fname); - return NULL; + return fi; } static struct dentry *udf_lookup(struct inode *dir, struct dentry *dentry, @@ -303,7 +299,7 @@ static struct fileIdentDesc *udf_add_entry(struct inode *dir, { struct super_block *sb = dir->i_sb; struct fileIdentDesc *fi = NULL; - char name[UDF_NAME_LEN]; + char *name = NULL; int namelen; loff_t f_pos; loff_t size = udf_ext0_offset(dir) + dir->i_size; @@ -317,16 +313,23 @@ static struct fileIdentDesc *udf_add_entry(struct inode *dir, struct extent_position epos = {}; struct udf_inode_info *dinfo; + fibh->sbh = fibh->ebh = NULL; + name = kmalloc(UDF_NAME_LEN, GFP_NOFS); + if (!name) { + *err = -ENOMEM; + goto out_err; + } + if (dentry) { if (!dentry->d_name.len) { *err = -EINVAL; - return NULL; + goto out_err; } namelen = udf_put_filename(sb, dentry->d_name.name, name, dentry->d_name.len); if (!namelen) { *err = -ENAMETOOLONG; - return NULL; + goto out_err; } } else { namelen = 0; @@ -338,11 +341,14 @@ static struct fileIdentDesc *udf_add_entry(struct inode *dir, fibh->soffset = fibh->eoffset = f_pos & (dir->i_sb->s_blocksize - 1); dinfo = UDF_I(dir); - if (dinfo->i_alloc_type == ICBTAG_FLAG_AD_IN_ICB) - fibh->sbh = fibh->ebh = NULL; - else if (inode_bmap(dir, f_pos >> dir->i_sb->s_blocksize_bits, - &epos, &eloc, &elen, &offset) == - (EXT_RECORDED_ALLOCATED >> 30)) { + if (dinfo->i_alloc_type != ICBTAG_FLAG_AD_IN_ICB) { + if (inode_bmap(dir, f_pos >> dir->i_sb->s_blocksize_bits, &epos, + &eloc, &elen, &offset) != (EXT_RECORDED_ALLOCATED >> 30)) { + block = udf_get_lb_pblock(dir->i_sb, + dinfo->i_location, 0); + fibh->soffset = fibh->eoffset = sb->s_blocksize; + goto add; + } block = udf_get_lb_pblock(dir->i_sb, eloc, offset); if ((++offset << dir->i_sb->s_blocksize_bits) < elen) { if (dinfo->i_alloc_type == ICBTAG_FLAG_AD_SHORT) @@ -354,17 +360,11 @@ static struct fileIdentDesc *udf_add_entry(struct inode *dir, fibh->sbh = fibh->ebh = udf_tread(dir->i_sb, block); if (!fibh->sbh) { - brelse(epos.bh); *err = -EIO; - return NULL; + goto out_err; } block = dinfo->i_location.logicalBlockNum; - } else { - block = udf_get_lb_pblock(dir->i_sb, dinfo->i_location, 0); - fibh->sbh = fibh->ebh = NULL; - fibh->soffset = fibh->eoffset = sb->s_blocksize; - goto add; } while (f_pos < size) { @@ -372,12 +372,8 @@ static struct fileIdentDesc *udf_add_entry(struct inode *dir, &elen, &offset); if (!fi) { - if (fibh->sbh != fibh->ebh) - brelse(fibh->ebh); - brelse(fibh->sbh); - brelse(epos.bh); *err = -EIO; - return NULL; + goto out_err; } liu = le16_to_cpu(cfi->lengthOfImpUse); @@ -386,7 +382,6 @@ static struct fileIdentDesc *udf_add_entry(struct inode *dir, if ((cfi->fileCharacteristics & FID_FILE_CHAR_DELETED) != 0) { if (((sizeof(struct fileIdentDesc) + liu + lfi + 3) & ~3) == nfidlen) { - brelse(epos.bh); cfi->descTag.tagSerialNum = cpu_to_le16(1); cfi->fileVersionNum = cpu_to_le16(1); cfi->fileCharacteristics = 0; @@ -394,10 +389,10 @@ static struct fileIdentDesc *udf_add_entry(struct inode *dir, cfi->lengthOfImpUse = cpu_to_le16(0); if (!udf_write_fi(dir, cfi, fi, fibh, NULL, name)) - return fi; + goto out_ok; else { *err = -EIO; - return NULL; + goto out_err; } } } @@ -427,7 +422,7 @@ add: fibh->sbh = fibh->ebh = udf_expand_dir_adinicb(dir, &block, err); if (!fibh->sbh) - return NULL; + goto out_err; epos.block = dinfo->i_location; epos.offset = udf_file_entry_alloc_offset(dir); /* Load extent udf_expand_dir_adinicb() has created */ @@ -468,11 +463,8 @@ add: dir->i_sb->s_blocksize_bits); fibh->ebh = udf_bread(dir, f_pos >> dir->i_sb->s_blocksize_bits, 1, err); - if (!fibh->ebh) { - brelse(epos.bh); - brelse(fibh->sbh); - return NULL; - } + if (!fibh->ebh) + goto out_err; if (!fibh->soffset) { if (udf_next_aext(dir, &epos, &eloc, &elen, 1) == @@ -503,20 +495,25 @@ add: cfi->lengthFileIdent = namelen; cfi->lengthOfImpUse = cpu_to_le16(0); if (!udf_write_fi(dir, cfi, fi, fibh, NULL, name)) { - brelse(epos.bh); dir->i_size += nfidlen; if (dinfo->i_alloc_type == ICBTAG_FLAG_AD_IN_ICB) dinfo->i_lenAlloc += nfidlen; mark_inode_dirty(dir); - return fi; + goto out_ok; } else { - brelse(epos.bh); - if (fibh->sbh != fibh->ebh) - brelse(fibh->ebh); - brelse(fibh->sbh); *err = -EIO; - return NULL; + goto out_err; } + +out_err: + fi = NULL; + if (fibh->sbh != fibh->ebh) + brelse(fibh->ebh); + brelse(fibh->sbh); +out_ok: + brelse(epos.bh); + kfree(name); + return fi; } static int udf_delete_entry(struct inode *inode, struct fileIdentDesc *fi, @@ -871,7 +868,7 @@ static int udf_symlink(struct inode *dir, struct dentry *dentry, char *ea; int err; int block; - char name[UDF_NAME_LEN]; + char *name = NULL; int namelen; struct buffer_head *bh; struct udf_inode_info *iinfo; @@ -881,6 +878,12 @@ static int udf_symlink(struct inode *dir, struct dentry *dentry, if (!inode) goto out; + name = kmalloc(UDF_NAME_LEN, GFP_NOFS); + if (!name) { + err = -ENOMEM; + goto out_no_entry; + } + iinfo = UDF_I(inode); inode->i_mode = S_IFLNK | S_IRWXUGO; inode->i_data.a_ops = &udf_symlink_aops; @@ -1020,6 +1023,7 @@ static int udf_symlink(struct inode *dir, struct dentry *dentry, err = 0; out: + kfree(name); unlock_kernel(); return err; -- cgit v1.2.3 From d0db181c072259c21a375b290a4574e5ce6d2b5f Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Wed, 5 Mar 2008 00:03:36 +0100 Subject: udf: fix anchor point detection According to ECMA 167 rev. 3 (see 3/8.4.2.1), Anchor Volume Descriptor Pointer should be recorded at two or more anchor points located at sectors 256, N, N - 256, where N - is a largest logical sector number at volume space. So we should always try to detect N on UDF volume before trying to find Anchor Volume Descriptor (i.e. calling to udf_find_anchor()). That said, all this patch does is updates the s_last_block even if the udf_vrs() returns positive value. Originally written and tested by Yuri Per, ported on latest mainline by me. Signed-off-by: Yuri Per Signed-off-by: Pavel Emelyanov Cc: Max Lyadvinsky Cc: Vladimir Simonov Cc: Andrew Neporada Cc: Kirill Korotaev Signed-off-by: Jan Kara --- fs/udf/super.c | 15 ++++++--------- 1 file changed, 6 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/udf/super.c b/fs/udf/super.c index bb8d866deb3..137153c5005 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -1425,7 +1425,7 @@ static noinline int udf_process_sequence(struct super_block *sb, long block, static int udf_check_valid(struct super_block *sb, int novrs, int silent) { long block; - struct udf_sb_info *sbi; + struct udf_sb_info *sbi = UDF_SB(sb); if (novrs) { udf_debug("Validity check skipped because of novrs option\n"); @@ -1434,15 +1434,12 @@ static int udf_check_valid(struct super_block *sb, int novrs, int silent) /* Check that it is NSR02 compliant */ /* Process any "CD-ROM Volume Descriptor Set" (ECMA 167 2/8.3.1) */ block = udf_vrs(sb, silent); - if (block != -1) - return !block; - - sbi = UDF_SB(sb); - udf_debug("Failed to read byte 32768. Assuming open " - "disc. Skipping validity check\n"); - if (!sbi->s_last_block) + if (block == -1) + udf_debug("Failed to read byte 32768. Assuming open " + "disc. Skipping validity check\n"); + if (block && !sbi->s_last_block) sbi->s_last_block = udf_get_last_block(sb); - return 0; + return !block; } static int udf_load_partition(struct super_block *sb, kernel_lb_addr *fileset) -- cgit v1.2.3 From c0eb31ed130ab18e1858179a644e39aee2355a13 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Tue, 1 Apr 2008 16:50:35 +0200 Subject: udf: Cleanup volume descriptor sequence processing Cleanup processing of volume descriptor sequence so that it is more readable, make code handle errors (e.g. media problems) better. Signed-off-by: Jan Kara --- fs/udf/super.c | 135 ++++++++++++++++++++++++++++++++++----------------------- 1 file changed, 81 insertions(+), 54 deletions(-) (limited to 'fs') diff --git a/fs/udf/super.c b/fs/udf/super.c index 137153c5005..500d1e4cc1c 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -85,16 +85,12 @@ static int udf_remount_fs(struct super_block *, int *, char *); static int udf_check_valid(struct super_block *, int, int); static int udf_vrs(struct super_block *sb, int silent); static int udf_load_partition(struct super_block *, kernel_lb_addr *); -static int udf_load_logicalvol(struct super_block *, struct buffer_head *, - kernel_lb_addr *); static void udf_load_logicalvolint(struct super_block *, kernel_extent_ad); static void udf_find_anchor(struct super_block *); static int udf_find_fileset(struct super_block *, kernel_lb_addr *, kernel_lb_addr *); -static void udf_load_pvoldesc(struct super_block *, struct buffer_head *); static void udf_load_fileset(struct super_block *, struct buffer_head *, kernel_lb_addr *); -static int udf_load_partdesc(struct super_block *, struct buffer_head *); static void udf_open_lvid(struct super_block *); static void udf_close_lvid(struct super_block *); static unsigned int udf_count_free(struct super_block *); @@ -935,11 +931,18 @@ static int udf_find_fileset(struct super_block *sb, return 1; } -static void udf_load_pvoldesc(struct super_block *sb, struct buffer_head *bh) +static int udf_load_pvoldesc(struct super_block *sb, sector_t block) { struct primaryVolDesc *pvoldesc; struct ustr instr; struct ustr outstr; + struct buffer_head *bh; + uint16_t ident; + + bh = udf_read_tagged(sb, block, block, &ident); + if (!bh) + return 1; + BUG_ON(ident != TAG_IDENT_PVD); pvoldesc = (struct primaryVolDesc *)bh->b_data; @@ -965,6 +968,9 @@ static void udf_load_pvoldesc(struct super_block *sb, struct buffer_head *bh) if (!udf_build_ustr(&instr, pvoldesc->volSetIdent, 128)) if (udf_CS0toUTF8(&outstr, &instr)) udf_debug("volSetIdent[] = '%s'\n", outstr.u_name); + + brelse(bh); + return 0; } static void udf_load_fileset(struct super_block *sb, struct buffer_head *bh, @@ -1018,16 +1024,27 @@ static struct udf_bitmap *udf_sb_alloc_bitmap(struct super_block *sb, u32 index) return bitmap; } -static int udf_load_partdesc(struct super_block *sb, struct buffer_head *bh) +static int udf_load_partdesc(struct super_block *sb, sector_t block) { + struct buffer_head *bh; struct partitionHeaderDesc *phd; - struct partitionDesc *p = (struct partitionDesc *)bh->b_data; + struct partitionDesc *p; struct udf_part_map *map; struct udf_sb_info *sbi = UDF_SB(sb); bool found = false; int i; - u16 partitionNumber = le16_to_cpu(p->partitionNumber); + uint16_t partitionNumber; + uint16_t ident; + int ret = 0; + + bh = udf_read_tagged(sb, block, block, &ident); + if (!bh) + return 1; + if (ident != TAG_IDENT_PD) + goto out_bh; + p = (struct partitionDesc *)bh->b_data; + partitionNumber = le16_to_cpu(p->partitionNumber); for (i = 0; i < sbi->s_partitions; i++) { map = &sbi->s_partmaps[i]; udf_debug("Searching map: (%d == %d)\n", @@ -1040,7 +1057,7 @@ static int udf_load_partdesc(struct super_block *sb, struct buffer_head *bh) if (!found) { udf_debug("Partition (%d) not found in partition map\n", partitionNumber); - return 0; + goto out_bh; } map->s_partition_len = le32_to_cpu(p->partitionLength); /* blocks */ @@ -1062,7 +1079,7 @@ static int udf_load_partdesc(struct super_block *sb, struct buffer_head *bh) if (strcmp(p->partitionContents.ident, PD_PARTITION_CONTENTS_NSR02) && strcmp(p->partitionContents.ident, PD_PARTITION_CONTENTS_NSR03)) - return 0; + goto out_bh; phd = (struct partitionHeaderDesc *)p->partitionContentsUse; if (phd->unallocSpaceTable.extLength) { @@ -1076,7 +1093,8 @@ static int udf_load_partdesc(struct super_block *sb, struct buffer_head *bh) if (!map->s_uspace.s_table) { udf_debug("cannot load unallocSpaceTable (part %d)\n", i); - return 1; + ret = 1; + goto out_bh; } map->s_partition_flags |= UDF_PART_FLAG_UNALLOC_TABLE; udf_debug("unallocSpaceTable (part %d) @ %ld\n", @@ -1110,7 +1128,8 @@ static int udf_load_partdesc(struct super_block *sb, struct buffer_head *bh) map->s_fspace.s_table = udf_iget(sb, loc); if (!map->s_fspace.s_table) { udf_debug("cannot load freedSpaceTable (part %d)\n", i); - return 1; + ret = 1; + goto out_bh; } map->s_partition_flags |= UDF_PART_FLAG_FREED_TABLE; @@ -1132,10 +1151,12 @@ static int udf_load_partdesc(struct super_block *sb, struct buffer_head *bh) } } - return 0; +out_bh: + brelse(bh); + return ret; } -static int udf_load_logicalvol(struct super_block *sb, struct buffer_head *bh, +static int udf_load_logicalvol(struct super_block *sb, sector_t block, kernel_lb_addr *fileset) { struct logicalVolDesc *lvd; @@ -1143,12 +1164,21 @@ static int udf_load_logicalvol(struct super_block *sb, struct buffer_head *bh, uint8_t type; struct udf_sb_info *sbi = UDF_SB(sb); struct genericPartitionMap *gpm; + uint16_t ident; + struct buffer_head *bh; + int ret = 0; + bh = udf_read_tagged(sb, block, block, &ident); + if (!bh) + return 1; + BUG_ON(ident != TAG_IDENT_LVD); lvd = (struct logicalVolDesc *)bh->b_data; i = udf_sb_alloc_partition_maps(sb, le32_to_cpu(lvd->numPartitionMaps)); - if (i != 0) - return i; + if (i != 0) { + ret = i; + goto out_bh; + } for (i = 0, offset = 0; i < sbi->s_partitions && offset < le32_to_cpu(lvd->mapTableLength); @@ -1187,7 +1217,6 @@ static int udf_load_logicalvol(struct super_block *sb, struct buffer_head *bh, UDF_ID_SPARABLE, strlen(UDF_ID_SPARABLE))) { uint32_t loc; - uint16_t ident; struct sparingTable *st; struct sparablePartitionMap *spm = (struct sparablePartitionMap *)gpm; @@ -1243,7 +1272,9 @@ static int udf_load_logicalvol(struct super_block *sb, struct buffer_head *bh, if (lvd->integritySeqExt.extLength) udf_load_logicalvolint(sb, leea_to_cpu(lvd->integritySeqExt)); - return 0; +out_bh: + brelse(bh); + return ret; } /* @@ -1301,19 +1332,25 @@ static noinline int udf_process_sequence(struct super_block *sb, long block, struct generic_desc *gd; struct volDescPtr *vdp; int done = 0; - int i, j; uint32_t vdsn; uint16_t ident; long next_s = 0, next_e = 0; memset(vds, 0, sizeof(struct udf_vds_record) * VDS_POS_LENGTH); - /* Read the main descriptor sequence */ + /* + * Read the main descriptor sequence and find which descriptors + * are in it. + */ for (; (!done && block <= lastblock); block++) { bh = udf_read_tagged(sb, block, block, &ident); - if (!bh) - break; + if (!bh) { + printk(KERN_ERR "udf: Block %Lu of volume descriptor " + "sequence is corrupted or we could not read " + "it.\n", (unsigned long long)block); + return 1; + } /* Process each descriptor (ISO 13346 3/8.3-8.4) */ gd = (struct generic_desc *)bh->b_data; @@ -1379,41 +1416,31 @@ static noinline int udf_process_sequence(struct super_block *sb, long block, } brelse(bh); } - for (i = 0; i < VDS_POS_LENGTH; i++) { - if (!vds[i].block) - continue; + /* + * Now read interesting descriptors again and process them + * in a suitable order + */ + if (!vds[VDS_POS_PRIMARY_VOL_DESC].block) { + printk(KERN_ERR "udf: Primary Volume Descriptor not found!\n"); + return 1; + } + if (udf_load_pvoldesc(sb, vds[VDS_POS_PRIMARY_VOL_DESC].block)) + return 1; - bh = udf_read_tagged(sb, vds[i].block, vds[i].block, - &ident); + if (vds[VDS_POS_LOGICAL_VOL_DESC].block && udf_load_logicalvol(sb, + vds[VDS_POS_LOGICAL_VOL_DESC].block, fileset)) + return 1; - if (i == VDS_POS_PRIMARY_VOL_DESC) - udf_load_pvoldesc(sb, bh); - else if (i == VDS_POS_LOGICAL_VOL_DESC) { - if (udf_load_logicalvol(sb, bh, fileset)) { - brelse(bh); + if (vds[VDS_POS_PARTITION_DESC].block) { + /* + * We rescan the whole descriptor sequence to find + * partition descriptor blocks and process them. + */ + for (block = vds[VDS_POS_PARTITION_DESC].block; + block < vds[VDS_POS_TERMINATING_DESC].block; + block++) + if (udf_load_partdesc(sb, block)) return 1; - } - } else if (i == VDS_POS_PARTITION_DESC) { - struct buffer_head *bh2 = NULL; - if (udf_load_partdesc(sb, bh)) { - brelse(bh); - return 1; - } - for (j = vds[i].block + 1; - j < vds[VDS_POS_TERMINATING_DESC].block; - j++) { - bh2 = udf_read_tagged(sb, j, j, &ident); - gd = (struct generic_desc *)bh2->b_data; - if (ident == TAG_IDENT_PD) - if (udf_load_partdesc(sb, bh2)) { - brelse(bh); - brelse(bh2); - return 1; - } - brelse(bh2); - } - } - brelse(bh); } return 0; -- cgit v1.2.3 From 2e0838fd0c0b8f0753a4be9af9f9712c4be881e1 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Tue, 1 Apr 2008 18:08:51 +0200 Subject: udf: Improve error recovery on mount Report error when we fail to allocate memory for a bitmap and properly release allocated memory and inodes for all the partitions in case of mount failure and umount. Signed-off-by: Jan Kara --- fs/udf/super.c | 86 ++++++++++++++++++++++++++++------------------------------ 1 file changed, 42 insertions(+), 44 deletions(-) (limited to 'fs') diff --git a/fs/udf/super.c b/fs/udf/super.c index 500d1e4cc1c..d50e3f5c46e 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -1103,16 +1103,18 @@ static int udf_load_partdesc(struct super_block *sb, sector_t block) if (phd->unallocSpaceBitmap.extLength) { struct udf_bitmap *bitmap = udf_sb_alloc_bitmap(sb, i); + if (!bitmap) { + ret = 1; + goto out_bh; + } map->s_uspace.s_bitmap = bitmap; - if (bitmap != NULL) { - bitmap->s_extLength = le32_to_cpu( + bitmap->s_extLength = le32_to_cpu( phd->unallocSpaceBitmap.extLength); - bitmap->s_extPosition = le32_to_cpu( + bitmap->s_extPosition = le32_to_cpu( phd->unallocSpaceBitmap.extPosition); - map->s_partition_flags |= UDF_PART_FLAG_UNALLOC_BITMAP; - udf_debug("unallocSpaceBitmap (part %d) @ %d\n", - i, bitmap->s_extPosition); - } + map->s_partition_flags |= UDF_PART_FLAG_UNALLOC_BITMAP; + udf_debug("unallocSpaceBitmap (part %d) @ %d\n", i, + bitmap->s_extPosition); } if (phd->partitionIntegrityTable.extLength) @@ -1139,19 +1141,22 @@ static int udf_load_partdesc(struct super_block *sb, sector_t block) if (phd->freedSpaceBitmap.extLength) { struct udf_bitmap *bitmap = udf_sb_alloc_bitmap(sb, i); + if (!bitmap) { + ret = 1; + goto out_bh; + } map->s_fspace.s_bitmap = bitmap; - if (bitmap != NULL) { - bitmap->s_extLength = le32_to_cpu( + bitmap->s_extLength = le32_to_cpu( phd->freedSpaceBitmap.extLength); - bitmap->s_extPosition = le32_to_cpu( + bitmap->s_extPosition = le32_to_cpu( phd->freedSpaceBitmap.extPosition); - map->s_partition_flags |= UDF_PART_FLAG_FREED_BITMAP; - udf_debug("freedSpaceBitmap (part %d) @ %d\n", - i, bitmap->s_extPosition); - } + map->s_partition_flags |= UDF_PART_FLAG_FREED_BITMAP; + udf_debug("freedSpaceBitmap (part %d) @ %d\n", i, + bitmap->s_extPosition); } out_bh: + /* In case loading failed, we handle cleanup in udf_fill_super */ brelse(bh); return ret; } @@ -1677,6 +1682,23 @@ static void udf_sb_free_bitmap(struct udf_bitmap *bitmap) vfree(bitmap); } +static void udf_free_partition(struct udf_part_map *map) +{ + int i; + + if (map->s_partition_flags & UDF_PART_FLAG_UNALLOC_TABLE) + iput(map->s_uspace.s_table); + if (map->s_partition_flags & UDF_PART_FLAG_FREED_TABLE) + iput(map->s_fspace.s_table); + if (map->s_partition_flags & UDF_PART_FLAG_UNALLOC_BITMAP) + udf_sb_free_bitmap(map->s_uspace.s_bitmap); + if (map->s_partition_flags & UDF_PART_FLAG_FREED_BITMAP) + udf_sb_free_bitmap(map->s_fspace.s_bitmap); + if (map->s_partition_type == UDF_SPARABLE_MAP15) + for (i = 0; i < 4; i++) + brelse(map->s_type_specific.s_sparing.s_spar_map[i]); +} + static int udf_fill_super(struct super_block *sb, void *options, int silent) { int i; @@ -1846,21 +1868,9 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent) error_out: if (sbi->s_vat_inode) iput(sbi->s_vat_inode); - if (sbi->s_partitions) { - struct udf_part_map *map = &sbi->s_partmaps[sbi->s_partition]; - if (map->s_partition_flags & UDF_PART_FLAG_UNALLOC_TABLE) - iput(map->s_uspace.s_table); - if (map->s_partition_flags & UDF_PART_FLAG_FREED_TABLE) - iput(map->s_fspace.s_table); - if (map->s_partition_flags & UDF_PART_FLAG_UNALLOC_BITMAP) - udf_sb_free_bitmap(map->s_uspace.s_bitmap); - if (map->s_partition_flags & UDF_PART_FLAG_FREED_BITMAP) - udf_sb_free_bitmap(map->s_fspace.s_bitmap); - if (map->s_partition_type == UDF_SPARABLE_MAP15) - for (i = 0; i < 4; i++) - brelse(map->s_type_specific.s_sparing. - s_spar_map[i]); - } + if (sbi->s_partitions) + for (i = 0; i < sbi->s_partitions; i++) + udf_free_partition(&sbi->s_partmaps[i]); #ifdef CONFIG_UDF_NLS if (UDF_QUERY_FLAG(sb, UDF_FLAG_NLS_MAP)) unload_nls(sbi->s_nls_map); @@ -1912,21 +1922,9 @@ static void udf_put_super(struct super_block *sb) sbi = UDF_SB(sb); if (sbi->s_vat_inode) iput(sbi->s_vat_inode); - if (sbi->s_partitions) { - struct udf_part_map *map = &sbi->s_partmaps[sbi->s_partition]; - if (map->s_partition_flags & UDF_PART_FLAG_UNALLOC_TABLE) - iput(map->s_uspace.s_table); - if (map->s_partition_flags & UDF_PART_FLAG_FREED_TABLE) - iput(map->s_fspace.s_table); - if (map->s_partition_flags & UDF_PART_FLAG_UNALLOC_BITMAP) - udf_sb_free_bitmap(map->s_uspace.s_bitmap); - if (map->s_partition_flags & UDF_PART_FLAG_FREED_BITMAP) - udf_sb_free_bitmap(map->s_fspace.s_bitmap); - if (map->s_partition_type == UDF_SPARABLE_MAP15) - for (i = 0; i < 4; i++) - brelse(map->s_type_specific.s_sparing. - s_spar_map[i]); - } + if (sbi->s_partitions) + for (i = 0; i < sbi->s_partitions; i++) + udf_free_partition(&sbi->s_partmaps[i]); #ifdef CONFIG_UDF_NLS if (UDF_QUERY_FLAG(sb, UDF_FLAG_NLS_MAP)) unload_nls(sbi->s_nls_map); -- cgit v1.2.3 From 3fb38dfa0e28575170119c70cb2ab70fdb8478fb Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Wed, 2 Apr 2008 12:26:36 +0200 Subject: udf: Move filling of partition descriptor info into a separate function Signed-off-by: Jan Kara --- fs/udf/super.c | 118 ++++++++++++++++++++++++++++++--------------------------- 1 file changed, 62 insertions(+), 56 deletions(-) (limited to 'fs') diff --git a/fs/udf/super.c b/fs/udf/super.c index d50e3f5c46e..c5fef85a9c1 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -1024,41 +1024,14 @@ static struct udf_bitmap *udf_sb_alloc_bitmap(struct super_block *sb, u32 index) return bitmap; } -static int udf_load_partdesc(struct super_block *sb, sector_t block) +static int udf_fill_partdesc_info(struct super_block *sb, + struct partitionDesc *p, int p_index) { - struct buffer_head *bh; - struct partitionHeaderDesc *phd; - struct partitionDesc *p; struct udf_part_map *map; struct udf_sb_info *sbi = UDF_SB(sb); - bool found = false; - int i; - uint16_t partitionNumber; - uint16_t ident; - int ret = 0; - - bh = udf_read_tagged(sb, block, block, &ident); - if (!bh) - return 1; - if (ident != TAG_IDENT_PD) - goto out_bh; - - p = (struct partitionDesc *)bh->b_data; - partitionNumber = le16_to_cpu(p->partitionNumber); - for (i = 0; i < sbi->s_partitions; i++) { - map = &sbi->s_partmaps[i]; - udf_debug("Searching map: (%d == %d)\n", - map->s_partition_num, partitionNumber); - found = map->s_partition_num == partitionNumber; - if (found) - break; - } + struct partitionHeaderDesc *phd; - if (!found) { - udf_debug("Partition (%d) not found in partition map\n", - partitionNumber); - goto out_bh; - } + map = &sbi->s_partmaps[p_index]; map->s_partition_len = le32_to_cpu(p->partitionLength); /* blocks */ map->s_partition_root = le32_to_cpu(p->partitionStartingLocation); @@ -1073,88 +1046,121 @@ static int udf_load_partdesc(struct super_block *sb, sector_t block) map->s_partition_flags |= UDF_PART_FLAG_OVERWRITABLE; udf_debug("Partition (%d:%d type %x) starts at physical %d, " - "block length %d\n", partitionNumber, i, + "block length %d\n", partitionNumber, p_index, map->s_partition_type, map->s_partition_root, map->s_partition_len); if (strcmp(p->partitionContents.ident, PD_PARTITION_CONTENTS_NSR02) && strcmp(p->partitionContents.ident, PD_PARTITION_CONTENTS_NSR03)) - goto out_bh; + return 0; phd = (struct partitionHeaderDesc *)p->partitionContentsUse; if (phd->unallocSpaceTable.extLength) { kernel_lb_addr loc = { .logicalBlockNum = le32_to_cpu( phd->unallocSpaceTable.extPosition), - .partitionReferenceNum = i, + .partitionReferenceNum = p_index, }; map->s_uspace.s_table = udf_iget(sb, loc); if (!map->s_uspace.s_table) { udf_debug("cannot load unallocSpaceTable (part %d)\n", - i); - ret = 1; - goto out_bh; + p_index); + return 1; } map->s_partition_flags |= UDF_PART_FLAG_UNALLOC_TABLE; udf_debug("unallocSpaceTable (part %d) @ %ld\n", - i, map->s_uspace.s_table->i_ino); + p_index, map->s_uspace.s_table->i_ino); } if (phd->unallocSpaceBitmap.extLength) { - struct udf_bitmap *bitmap = udf_sb_alloc_bitmap(sb, i); - if (!bitmap) { - ret = 1; - goto out_bh; - } + struct udf_bitmap *bitmap = udf_sb_alloc_bitmap(sb, p_index); + if (!bitmap) + return 1; map->s_uspace.s_bitmap = bitmap; bitmap->s_extLength = le32_to_cpu( phd->unallocSpaceBitmap.extLength); bitmap->s_extPosition = le32_to_cpu( phd->unallocSpaceBitmap.extPosition); map->s_partition_flags |= UDF_PART_FLAG_UNALLOC_BITMAP; - udf_debug("unallocSpaceBitmap (part %d) @ %d\n", i, + udf_debug("unallocSpaceBitmap (part %d) @ %d\n", p_index, bitmap->s_extPosition); } if (phd->partitionIntegrityTable.extLength) - udf_debug("partitionIntegrityTable (part %d)\n", i); + udf_debug("partitionIntegrityTable (part %d)\n", p_index); if (phd->freedSpaceTable.extLength) { kernel_lb_addr loc = { .logicalBlockNum = le32_to_cpu( phd->freedSpaceTable.extPosition), - .partitionReferenceNum = i, + .partitionReferenceNum = p_index, }; map->s_fspace.s_table = udf_iget(sb, loc); if (!map->s_fspace.s_table) { - udf_debug("cannot load freedSpaceTable (part %d)\n", i); - ret = 1; - goto out_bh; + udf_debug("cannot load freedSpaceTable (part %d)\n", + p_index); + return 1; } map->s_partition_flags |= UDF_PART_FLAG_FREED_TABLE; udf_debug("freedSpaceTable (part %d) @ %ld\n", - i, map->s_fspace.s_table->i_ino); + p_index, map->s_fspace.s_table->i_ino); } if (phd->freedSpaceBitmap.extLength) { - struct udf_bitmap *bitmap = udf_sb_alloc_bitmap(sb, i); - if (!bitmap) { - ret = 1; - goto out_bh; - } + struct udf_bitmap *bitmap = udf_sb_alloc_bitmap(sb, p_index); + if (!bitmap) + return 1; map->s_fspace.s_bitmap = bitmap; bitmap->s_extLength = le32_to_cpu( phd->freedSpaceBitmap.extLength); bitmap->s_extPosition = le32_to_cpu( phd->freedSpaceBitmap.extPosition); map->s_partition_flags |= UDF_PART_FLAG_FREED_BITMAP; - udf_debug("freedSpaceBitmap (part %d) @ %d\n", i, + udf_debug("freedSpaceBitmap (part %d) @ %d\n", p_index, bitmap->s_extPosition); } + return 0; +} + +static int udf_load_partdesc(struct super_block *sb, sector_t block) +{ + struct buffer_head *bh; + struct partitionDesc *p; + struct udf_part_map *map; + struct udf_sb_info *sbi = UDF_SB(sb); + bool found = false; + int i; + uint16_t partitionNumber; + uint16_t ident; + int ret = 0; + + bh = udf_read_tagged(sb, block, block, &ident); + if (!bh) + return 1; + if (ident != TAG_IDENT_PD) + goto out_bh; + + p = (struct partitionDesc *)bh->b_data; + partitionNumber = le16_to_cpu(p->partitionNumber); + for (i = 0; i < sbi->s_partitions; i++) { + map = &sbi->s_partmaps[i]; + udf_debug("Searching map: (%d == %d)\n", + map->s_partition_num, partitionNumber); + found = map->s_partition_num == partitionNumber; + if (found) + break; + } + + if (!found) { + udf_debug("Partition (%d) not found in partition map\n", + partitionNumber); + goto out_bh; + } + ret = udf_fill_partdesc_info(sb, p, i); out_bh: /* In case loading failed, we handle cleanup in udf_fill_super */ brelse(bh); -- cgit v1.2.3 From 38b74a53e5625b7bbbd08918294b86f1db2f0565 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Wed, 2 Apr 2008 16:01:35 +0200 Subject: udf: Move processing of virtual partitions This patch move processing of UDF virtual partitions close to the place where other partition types are processed. As a result we now also properly fill in partition access type. Signed-off-by: Jan Kara --- fs/udf/super.c | 165 +++++++++++++++++++++++++++++---------------------------- 1 file changed, 85 insertions(+), 80 deletions(-) (limited to 'fs') diff --git a/fs/udf/super.c b/fs/udf/super.c index c5fef85a9c1..7ec828566df 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -84,7 +84,6 @@ static void udf_write_super(struct super_block *); static int udf_remount_fs(struct super_block *, int *, char *); static int udf_check_valid(struct super_block *, int, int); static int udf_vrs(struct super_block *sb, int silent); -static int udf_load_partition(struct super_block *, kernel_lb_addr *); static void udf_load_logicalvolint(struct super_block *, kernel_extent_ad); static void udf_find_anchor(struct super_block *); static int udf_find_fileset(struct super_block *, kernel_lb_addr *, @@ -1125,14 +1124,54 @@ static int udf_fill_partdesc_info(struct super_block *sb, return 0; } +static int udf_load_vat(struct super_block *sb, int p_index, int type1_index) +{ + struct udf_sb_info *sbi = UDF_SB(sb); + struct udf_part_map *map = &sbi->s_partmaps[p_index]; + kernel_lb_addr ino; + struct buffer_head *bh; + + /* VAT file entry is in the last recorded block */ + ino.partitionReferenceNum = type1_index; + ino.logicalBlockNum = sbi->s_last_block - map->s_partition_root; + sbi->s_vat_inode = udf_iget(sb, ino); + if (!sbi->s_vat_inode) + return 1; + + if (map->s_partition_type == UDF_VIRTUAL_MAP15) { + map->s_type_specific.s_virtual.s_start_offset = + udf_ext0_offset(sbi->s_vat_inode); + map->s_type_specific.s_virtual.s_num_entries = + (sbi->s_vat_inode->i_size - 36) >> 2; + } else if (map->s_partition_type == UDF_VIRTUAL_MAP20) { + uint32_t pos; + struct virtualAllocationTable20 *vat20; + + pos = udf_block_map(sbi->s_vat_inode, 0); + bh = sb_bread(sb, pos); + if (!bh) + return 1; + vat20 = (struct virtualAllocationTable20 *)bh->b_data + + udf_ext0_offset(sbi->s_vat_inode); + map->s_type_specific.s_virtual.s_start_offset = + le16_to_cpu(vat20->lengthHeader) + + udf_ext0_offset(sbi->s_vat_inode); + map->s_type_specific.s_virtual.s_num_entries = + (sbi->s_vat_inode->i_size - + map->s_type_specific.s_virtual. + s_start_offset) >> 2; + brelse(bh); + } + return 0; +} + static int udf_load_partdesc(struct super_block *sb, sector_t block) { struct buffer_head *bh; struct partitionDesc *p; struct udf_part_map *map; struct udf_sb_info *sbi = UDF_SB(sb); - bool found = false; - int i; + int i, type1_idx; uint16_t partitionNumber; uint16_t ident; int ret = 0; @@ -1145,22 +1184,59 @@ static int udf_load_partdesc(struct super_block *sb, sector_t block) p = (struct partitionDesc *)bh->b_data; partitionNumber = le16_to_cpu(p->partitionNumber); + + /* First scan for TYPE1 and SPARABLE partitions */ for (i = 0; i < sbi->s_partitions; i++) { map = &sbi->s_partmaps[i]; udf_debug("Searching map: (%d == %d)\n", map->s_partition_num, partitionNumber); - found = map->s_partition_num == partitionNumber; - if (found) + if (map->s_partition_num == partitionNumber && + (map->s_partition_type == UDF_TYPE1_MAP15 || + map->s_partition_type == UDF_SPARABLE_MAP15)) break; } - if (!found) { + if (i >= sbi->s_partitions) { udf_debug("Partition (%d) not found in partition map\n", partitionNumber); goto out_bh; } ret = udf_fill_partdesc_info(sb, p, i); + + /* + * Now rescan for VIRTUAL partitions when TYPE1 partitions are + * already set up + */ + type1_idx = i; + for (i = 0; i < sbi->s_partitions; i++) { + map = &sbi->s_partmaps[i]; + + if (map->s_partition_num == partitionNumber && + (map->s_partition_type == UDF_VIRTUAL_MAP15 || + map->s_partition_type == UDF_VIRTUAL_MAP20)) + break; + } + + if (i >= sbi->s_partitions) + goto out_bh; + + ret = udf_fill_partdesc_info(sb, p, i); + if (ret) + goto out_bh; + + if (!sbi->s_last_block) { + sbi->s_last_block = udf_get_last_block(sb); + udf_find_anchor(sb); + if (!sbi->s_last_block) { + udf_debug("Unable to determine Lastblock (For " + "Virtual Partition)\n"); + ret = 1; + goto out_bh; + } + } + + ret = udf_load_vat(sb, i, type1_idx); out_bh: /* In case loading failed, we handle cleanup in udf_fill_super */ brelse(bh); @@ -1480,13 +1556,13 @@ static int udf_check_valid(struct super_block *sb, int novrs, int silent) return !block; } -static int udf_load_partition(struct super_block *sb, kernel_lb_addr *fileset) +static int udf_load_sequence(struct super_block *sb, kernel_lb_addr *fileset) { struct anchorVolDescPtr *anchor; uint16_t ident; struct buffer_head *bh; long main_s, main_e, reserve_s, reserve_e; - int i, j; + int i; struct udf_sb_info *sbi; if (!sb) @@ -1535,77 +1611,6 @@ static int udf_load_partition(struct super_block *sb, kernel_lb_addr *fileset) } udf_debug("Using anchor in block %d\n", sbi->s_anchor[i]); - for (i = 0; i < sbi->s_partitions; i++) { - kernel_lb_addr uninitialized_var(ino); - struct udf_part_map *map = &sbi->s_partmaps[i]; - - if (map->s_partition_type != UDF_VIRTUAL_MAP15 && - map->s_partition_type != UDF_VIRTUAL_MAP20) - continue; - - if (!sbi->s_last_block) { - sbi->s_last_block = udf_get_last_block(sb); - udf_find_anchor(sb); - } - - if (!sbi->s_last_block) { - udf_debug("Unable to determine Lastblock (For " - "Virtual Partition)\n"); - return 1; - } - - for (j = 0; j < sbi->s_partitions; j++) { - struct udf_part_map *map2 = &sbi->s_partmaps[j]; - if (j != i && - map->s_volumeseqnum == - map2->s_volumeseqnum && - map->s_partition_num == - map2->s_partition_num) { - ino.partitionReferenceNum = j; - ino.logicalBlockNum = - sbi->s_last_block - - map2->s_partition_root; - break; - } - } - - if (j == sbi->s_partitions) - return 1; - - sbi->s_vat_inode = udf_iget(sb, ino); - if (!sbi->s_vat_inode) - return 1; - - if (map->s_partition_type == UDF_VIRTUAL_MAP15) { - map->s_type_specific.s_virtual.s_start_offset = - udf_ext0_offset(sbi->s_vat_inode); - map->s_type_specific.s_virtual.s_num_entries = - (sbi->s_vat_inode->i_size - 36) >> 2; - } else if (map->s_partition_type == UDF_VIRTUAL_MAP20) { - uint32_t pos; - struct virtualAllocationTable20 *vat20; - - pos = udf_block_map(sbi->s_vat_inode, 0); - bh = sb_bread(sb, pos); - if (!bh) - return 1; - vat20 = (struct virtualAllocationTable20 *) - bh->b_data + - udf_ext0_offset(sbi->s_vat_inode); - map->s_type_specific.s_virtual.s_start_offset = - le16_to_cpu(vat20->lengthHeader) + - udf_ext0_offset(sbi->s_vat_inode); - map->s_type_specific.s_virtual.s_num_entries = - (sbi->s_vat_inode->i_size - - map->s_type_specific.s_virtual. - s_start_offset) >> 2; - brelse(bh); - } - map->s_partition_root = udf_get_pblock(sb, 0, i, 0); - map->s_partition_len = - sbi->s_partmaps[ino.partitionReferenceNum]. - s_partition_len; - } return 0; } @@ -1790,7 +1795,7 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent) sb->s_magic = UDF_SUPER_MAGIC; sb->s_time_gran = 1000; - if (udf_load_partition(sb, &fileset)) { + if (udf_load_sequence(sb, &fileset)) { printk(KERN_WARNING "UDF-fs: No partition found (1)\n"); goto error_out; } -- cgit v1.2.3 From 423cf6dc04eb79d441bfda2b127bc4b57134b41d Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 7 Apr 2008 15:59:23 +0200 Subject: udf: Cleanup anchor block detection. UDF anchor block detection is complicated by several things - there are several places where the anchor point can be, some of them relative to the last recorded block which some devices report wrongly. Moreover some devices on some media seem to have 7 spare blocks sectors for every 32 blocks (at least as far as I understand the old code) so we have to count also with that possibility. This patch splits anchor block detection into several functions so that it is clearer what we actually try to do. We fix several bugs of the type "for such and such media, we fail to check block blah" as a result of the cleanup. Signed-off-by: Jan Kara --- fs/udf/super.c | 239 +++++++++++++++++++++++--------------------------------- fs/udf/udf_sb.h | 2 +- 2 files changed, 99 insertions(+), 142 deletions(-) (limited to 'fs') diff --git a/fs/udf/super.c b/fs/udf/super.c index 7ec828566df..fe0dafebde7 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -678,149 +678,120 @@ static int udf_vrs(struct super_block *sb, int silent) } /* - * udf_find_anchor - * - * PURPOSE - * Find an anchor volume descriptor. - * - * PRE-CONDITIONS - * sb Pointer to _locked_ superblock. - * lastblock Last block on media. - * - * POST-CONDITIONS - * 1 if not found, 0 if ok - * - * HISTORY - * July 1, 1997 - Andrew E. Mileski - * Written, tested, and released. + * Check whether there is an anchor block in the given block */ -static void udf_find_anchor(struct super_block *sb) +static int udf_check_anchor_block(struct super_block *sb, sector_t block, + bool varconv) { - int lastblock; struct buffer_head *bh = NULL; + tag *t; uint16_t ident; uint32_t location; + + if (varconv) + bh = sb_bread(sb, udf_fixed_to_variable(block)); + else + bh = sb_bread(sb, block); + + if (!bh) + return 0; + + t = (tag *)bh->b_data; + ident = le16_to_cpu(t->tagIdent); + location = le32_to_cpu(t->tagLocation); + brelse(bh); + if (ident != TAG_IDENT_AVDP) + return 0; + return location == block; +} + +/* Search for an anchor volume descriptor pointer */ +static sector_t udf_scan_anchors(struct super_block *sb, bool varconv, + sector_t lastblock) +{ + sector_t last[4]; int i; - struct udf_sb_info *sbi; + struct udf_sb_info *sbi = UDF_SB(sb); - sbi = UDF_SB(sb); - lastblock = sbi->s_last_block; + last[0] = lastblock; + last[1] = last[0] - 2; + last[2] = last[0] - 150; + last[3] = last[0] - 152; - if (lastblock) { - int varlastblock = udf_variable_to_fixed(lastblock); - int last[] = { lastblock, lastblock - 2, - lastblock - 150, lastblock - 152, - varlastblock, varlastblock - 2, - varlastblock - 150, varlastblock - 152 }; - - lastblock = 0; - - /* Search for an anchor volume descriptor pointer */ - - /* according to spec, anchor is in either: - * block 256 - * lastblock-256 - * lastblock - * however, if the disc isn't closed, it could be 512 */ - - for (i = 0; !lastblock && i < ARRAY_SIZE(last); i++) { - ident = location = 0; - if (last[i] >= 0) { - bh = sb_bread(sb, last[i]); - if (bh) { - tag *t = (tag *)bh->b_data; - ident = le16_to_cpu(t->tagIdent); - location = le32_to_cpu(t->tagLocation); - brelse(bh); - } - } + /* according to spec, anchor is in either: + * block 256 + * lastblock-256 + * lastblock + * however, if the disc isn't closed, it could be 512 */ - if (ident == TAG_IDENT_AVDP) { - if (location == last[i] - sbi->s_session) { - lastblock = last[i] - sbi->s_session; - sbi->s_anchor[0] = lastblock; - sbi->s_anchor[1] = lastblock - 256; - } else if (location == - udf_variable_to_fixed(last[i]) - - sbi->s_session) { - UDF_SET_FLAG(sb, UDF_FLAG_VARCONV); - lastblock = - udf_variable_to_fixed(last[i]) - - sbi->s_session; - sbi->s_anchor[0] = lastblock; - sbi->s_anchor[1] = lastblock - 256 - - sbi->s_session; - } else { - udf_debug("Anchor found at block %d, " - "location mismatch %d.\n", - last[i], location); - } - } else if (ident == TAG_IDENT_FE || - ident == TAG_IDENT_EFE) { - lastblock = last[i]; - sbi->s_anchor[3] = 512; - } else { - ident = location = 0; - if (last[i] >= 256) { - bh = sb_bread(sb, last[i] - 256); - if (bh) { - tag *t = (tag *)bh->b_data; - ident = le16_to_cpu( - t->tagIdent); - location = le32_to_cpu( - t->tagLocation); - brelse(bh); - } - } + for (i = 0; i < ARRAY_SIZE(last); i++) { + if (last[i] < 0) + continue; - if (ident == TAG_IDENT_AVDP && - location == last[i] - 256 - - sbi->s_session) { - lastblock = last[i]; - sbi->s_anchor[1] = last[i] - 256; - } else { - ident = location = 0; - if (last[i] >= 312 + sbi->s_session) { - bh = sb_bread(sb, - last[i] - 312 - - sbi->s_session); - if (bh) { - tag *t = (tag *) - bh->b_data; - ident = le16_to_cpu( - t->tagIdent); - location = le32_to_cpu( - t->tagLocation); - brelse(bh); - } - } + if (udf_check_anchor_block(sb, last[i], varconv)) { + sbi->s_anchor[0] = last[i]; + sbi->s_anchor[1] = last[i] - 256; + return last[i]; + } - if (ident == TAG_IDENT_AVDP && - location == udf_variable_to_fixed(last[i]) - 256) { - UDF_SET_FLAG(sb, - UDF_FLAG_VARCONV); - lastblock = udf_variable_to_fixed(last[i]); - sbi->s_anchor[1] = lastblock - 256; - } - } - } + if (last[i] < 256) + continue; + + if (udf_check_anchor_block(sb, last[i] - 256, varconv)) { + sbi->s_anchor[1] = last[i] - 256; + return last[i]; } } - if (!lastblock) { - /* We haven't found the lastblock. check 312 */ - bh = sb_bread(sb, 312 + sbi->s_session); - if (bh) { - tag *t = (tag *)bh->b_data; - ident = le16_to_cpu(t->tagIdent); - location = le32_to_cpu(t->tagLocation); - brelse(bh); + if (udf_check_anchor_block(sb, sbi->s_session + 256, varconv)) { + sbi->s_anchor[0] = sbi->s_session + 256; + return last[0]; + } + if (udf_check_anchor_block(sb, sbi->s_session + 512, varconv)) { + sbi->s_anchor[0] = sbi->s_session + 512; + return last[0]; + } + return 0; +} - if (ident == TAG_IDENT_AVDP && location == 256) - UDF_SET_FLAG(sb, UDF_FLAG_VARCONV); - } +/* + * Find an anchor volume descriptor. The function expects sbi->s_lastblock to + * be the last block on the media. + * + * Return 1 if not found, 0 if ok + * + */ +static void udf_find_anchor(struct super_block *sb) +{ + sector_t lastblock; + struct buffer_head *bh = NULL; + uint16_t ident; + int i; + struct udf_sb_info *sbi = UDF_SB(sb); + + lastblock = udf_scan_anchors(sb, 0, sbi->s_last_block); + if (lastblock) + goto check_anchor; + + /* No anchor found? Try VARCONV conversion of block numbers */ + /* Firstly, we try to not convert number of the last block */ + lastblock = udf_scan_anchors(sb, 1, + udf_variable_to_fixed(sbi->s_last_block)); + if (lastblock) { + UDF_SET_FLAG(sb, UDF_FLAG_VARCONV); + goto check_anchor; } + /* Secondly, we try with converted number of the last block */ + lastblock = udf_scan_anchors(sb, 1, sbi->s_last_block); + if (lastblock) + UDF_SET_FLAG(sb, UDF_FLAG_VARCONV); + +check_anchor: + /* + * Check located anchors and the anchor block supplied via + * mount options + */ for (i = 0; i < ARRAY_SIZE(sbi->s_anchor); i++) { if (!sbi->s_anchor[i]) continue; @@ -830,9 +801,7 @@ static void udf_find_anchor(struct super_block *sb) sbi->s_anchor[i] = 0; else { brelse(bh); - if ((ident != TAG_IDENT_AVDP) && - (i || (ident != TAG_IDENT_FE && - ident != TAG_IDENT_EFE))) + if (ident != TAG_IDENT_AVDP) sbi->s_anchor[i] = 0; } } @@ -1225,17 +1194,6 @@ static int udf_load_partdesc(struct super_block *sb, sector_t block) if (ret) goto out_bh; - if (!sbi->s_last_block) { - sbi->s_last_block = udf_get_last_block(sb); - udf_find_anchor(sb); - if (!sbi->s_last_block) { - udf_debug("Unable to determine Lastblock (For " - "Virtual Partition)\n"); - ret = 1; - goto out_bh; - } - } - ret = udf_load_vat(sb, i, type1_idx); out_bh: /* In case loading failed, we handle cleanup in udf_fill_super */ @@ -1778,7 +1736,6 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent) sbi->s_last_block = uopt.lastblock; sbi->s_anchor[0] = sbi->s_anchor[1] = 0; sbi->s_anchor[2] = uopt.anchor; - sbi->s_anchor[3] = 256; if (udf_check_valid(sb, uopt.novrs, silent)) { /* read volume recognition sequences */ diff --git a/fs/udf/udf_sb.h b/fs/udf/udf_sb.h index fb2a6e9f8da..8308a12e123 100644 --- a/fs/udf/udf_sb.h +++ b/fs/udf/udf_sb.h @@ -100,7 +100,7 @@ struct udf_sb_info { /* Sector headers */ __s32 s_session; - __u32 s_anchor[4]; + __u32 s_anchor[3]; __u32 s_last_block; struct buffer_head *s_lvid_bh; -- cgit v1.2.3 From 5fb28aa25ab0b71af2e441d68e63ad257e610a04 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 7 Apr 2008 16:15:04 +0200 Subject: udf: Improve anchor block detection Add +1 and -1 to a list of blocks which can be the real last recorded block on a UDF media. Sebastian Manciulea claims this helps some drive + media combinations he is able to test. Signed-off-by: Jan Kara --- fs/udf/super.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/udf/super.c b/fs/udf/super.c index fe0dafebde7..b9e719de070 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -709,14 +709,16 @@ static int udf_check_anchor_block(struct super_block *sb, sector_t block, static sector_t udf_scan_anchors(struct super_block *sb, bool varconv, sector_t lastblock) { - sector_t last[4]; + sector_t last[6]; int i; struct udf_sb_info *sbi = UDF_SB(sb); last[0] = lastblock; - last[1] = last[0] - 2; - last[2] = last[0] - 150; - last[3] = last[0] - 152; + last[1] = last[0] - 1; + last[2] = last[0] + 1; + last[3] = last[0] - 2; + last[4] = last[0] - 150; + last[5] = last[0] - 152; /* according to spec, anchor is in either: * block 256 @@ -727,6 +729,9 @@ static sector_t udf_scan_anchors(struct super_block *sb, bool varconv, for (i = 0; i < ARRAY_SIZE(last); i++) { if (last[i] < 0) continue; + if (last[i] >= sb->s_bdev->bd_inode->i_size >> + sb->s_blocksize_bits) + continue; if (udf_check_anchor_block(sb, last[i], varconv)) { sbi->s_anchor[0] = last[i]; -- cgit v1.2.3 From 4f7874c868eaedd0e64b2f6c800bc852bdc7f38b Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 7 Apr 2008 23:16:38 +0200 Subject: udf: Silence warning about accesses beyond end of device Some of the computed positions of anchor block could be beyond the end of device. Skip reading such blocks. Signed-off-by: Jan Kara --- fs/udf/super.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/udf/super.c b/fs/udf/super.c index b9e719de070..28ed3f5ebd7 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -688,8 +688,12 @@ static int udf_check_anchor_block(struct super_block *sb, sector_t block, uint16_t ident; uint32_t location; - if (varconv) + if (varconv) { + if (udf_fixed_to_variable(block) >= + sb->s_bdev->bd_inode->i_size >> sb->s_blocksize_bits) + return 0; bh = sb_bread(sb, udf_fixed_to_variable(block)); + } else bh = sb_bread(sb, block); -- cgit v1.2.3 From c82a127505d39fa81c886eceed6fdf8c1ff4a06b Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Tue, 8 Apr 2008 01:16:32 +0200 Subject: udf: Fix detection of VAT version We incorrectly (way to strictly) checked version of VAT on loading and thus refuse to mount correct media. There are just two format versions - below 2.0 and above 2.0 and we understand both. So update the version check accordingly. Signed-off-by: Jan Kara --- fs/udf/super.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/udf/super.c b/fs/udf/super.c index 28ed3f5ebd7..14f965e8a73 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -1256,12 +1256,12 @@ static int udf_load_logicalvol(struct super_block *sb, sector_t block, u16 suf = le16_to_cpu(((__le16 *)upm2->partIdent. identSuffix)[0]); - if (suf == 0x0150) { + if (suf < 0x0200) { map->s_partition_type = UDF_VIRTUAL_MAP15; map->s_partition_func = udf_get_pblock_virt15; - } else if (suf == 0x0200) { + } else { map->s_partition_type = UDF_VIRTUAL_MAP20; map->s_partition_func = -- cgit v1.2.3 From 742e1795e2d9dc99657742e5bbbb7907596bf6c3 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Tue, 8 Apr 2008 01:17:52 +0200 Subject: udf: Allow loading of VAT inode UDF media with VAT could have never worked because udf_fill_inode() didn't know about case FILE_TYPE_VAT20. Fix this. Signed-off-by: Jan Kara --- fs/udf/inode.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/udf/inode.c b/fs/udf/inode.c index 2362bf0c690..c150b6df626 100644 --- a/fs/udf/inode.c +++ b/fs/udf/inode.c @@ -1275,6 +1275,7 @@ static void udf_fill_inode(struct inode *inode, struct buffer_head *bh) case ICBTAG_FILE_TYPE_REALTIME: case ICBTAG_FILE_TYPE_REGULAR: case ICBTAG_FILE_TYPE_UNDEF: + case ICBTAG_FILE_TYPE_VAT20: if (iinfo->i_alloc_type == ICBTAG_FLAG_AD_IN_ICB) inode->i_data.a_ops = &udf_adinicb_aops; else -- cgit v1.2.3 From fa5e08156335d0687c85b4e724db9448fb166601 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Tue, 8 Apr 2008 02:08:53 +0200 Subject: udf: Handle VAT packed inside inode properly We didn't handle VAT packed inside the inode - we tried to call udf_block_map() on such file which lead to strange results at best. Add proper handling of packed VAT as we do it with other packed files. Signed-off-by: Jan Kara --- fs/udf/partition.c | 10 +++++++--- fs/udf/super.c | 24 +++++++++++++++--------- 2 files changed, 22 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/udf/partition.c b/fs/udf/partition.c index 307c9c33d18..b2e6e1eddb9 100644 --- a/fs/udf/partition.c +++ b/fs/udf/partition.c @@ -54,11 +54,10 @@ uint32_t udf_get_pblock_virt15(struct super_block *sb, uint32_t block, struct udf_sb_info *sbi = UDF_SB(sb); struct udf_part_map *map; struct udf_virtual_data *vdata; - struct udf_inode_info *iinfo; + struct udf_inode_info *iinfo = UDF_I(sbi->s_vat_inode); map = &sbi->s_partmaps[partition]; vdata = &map->s_type_specific.s_virtual; - index = (sb->s_blocksize - vdata->s_start_offset) / sizeof(uint32_t); if (block > vdata->s_num_entries) { udf_debug("Trying to access block beyond end of VAT " @@ -66,6 +65,11 @@ uint32_t udf_get_pblock_virt15(struct super_block *sb, uint32_t block, return 0xFFFFFFFF; } + if (iinfo->i_alloc_type == ICBTAG_FLAG_AD_IN_ICB) { + loc = le32_to_cpu(((__le32 *)iinfo->i_ext.i_data)[block]); + goto translate; + } + index = (sb->s_blocksize - vdata->s_start_offset) / sizeof(uint32_t); if (block >= index) { block -= index; newblock = 1 + (block / (sb->s_blocksize / sizeof(uint32_t))); @@ -88,7 +92,7 @@ uint32_t udf_get_pblock_virt15(struct super_block *sb, uint32_t block, brelse(bh); - iinfo = UDF_I(sbi->s_vat_inode); +translate: if (iinfo->i_location.partitionReferenceNum == partition) { udf_debug("recursive call to udf_get_pblock!\n"); return 0xFFFFFFFF; diff --git a/fs/udf/super.c b/fs/udf/super.c index 14f965e8a73..abdb9b31da4 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -1107,7 +1107,10 @@ static int udf_load_vat(struct super_block *sb, int p_index, int type1_index) struct udf_sb_info *sbi = UDF_SB(sb); struct udf_part_map *map = &sbi->s_partmaps[p_index]; kernel_lb_addr ino; - struct buffer_head *bh; + struct buffer_head *bh = NULL; + struct udf_inode_info *vati; + uint32_t pos; + struct virtualAllocationTable20 *vat20; /* VAT file entry is in the last recorded block */ ino.partitionReferenceNum = type1_index; @@ -1122,15 +1125,18 @@ static int udf_load_vat(struct super_block *sb, int p_index, int type1_index) map->s_type_specific.s_virtual.s_num_entries = (sbi->s_vat_inode->i_size - 36) >> 2; } else if (map->s_partition_type == UDF_VIRTUAL_MAP20) { - uint32_t pos; - struct virtualAllocationTable20 *vat20; + vati = UDF_I(sbi->s_vat_inode); + if (vati->i_alloc_type != ICBTAG_FLAG_AD_IN_ICB) { + pos = udf_block_map(sbi->s_vat_inode, 0); + bh = sb_bread(sb, pos); + if (!bh) + return 1; + vat20 = (struct virtualAllocationTable20 *)bh->b_data; + } else { + vat20 = (struct virtualAllocationTable20 *) + vati->i_ext.i_data; + } - pos = udf_block_map(sbi->s_vat_inode, 0); - bh = sb_bread(sb, pos); - if (!bh) - return 1; - vat20 = (struct virtualAllocationTable20 *)bh->b_data + - udf_ext0_offset(sbi->s_vat_inode); map->s_type_specific.s_virtual.s_start_offset = le16_to_cpu(vat20->lengthHeader) + udf_ext0_offset(sbi->s_vat_inode); -- cgit v1.2.3 From 96200be3077c5ede16a90b33aca815b444e66043 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Tue, 8 Apr 2008 13:29:20 +0200 Subject: udf: Mount filesystem read-only if it has pseudooverwrite partition As we don't properly support writing to pseudooverwrite partition (we should add entries to VAT and relocate blocks instead of just writing them), mount filesystems with such partition as read-only. Signed-off-by: Jan Kara --- fs/udf/super.c | 8 ++++++++ 1 file changed, 8 insertions(+) (limited to 'fs') diff --git a/fs/udf/super.c b/fs/udf/super.c index abdb9b31da4..650f20fe9d6 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -1208,6 +1208,14 @@ static int udf_load_partdesc(struct super_block *sb, sector_t block) ret = udf_fill_partdesc_info(sb, p, i); if (ret) goto out_bh; + /* + * Mark filesystem read-only if we have a partition with virtual map + * since we don't handle writing to it (we overwrite blocks instead of + * relocating them). + */ + sb->s_flags |= MS_RDONLY; + printk(KERN_NOTICE "UDF-fs: Filesystem marked read-only because " + "writing to pseudooverwrite partition is not implemented.\n"); ret = udf_load_vat(sb, i, type1_idx); out_bh: -- cgit v1.2.3 From f4bcbbd92ebda971f7c2cd1132b399808ed6cf9b Mon Sep 17 00:00:00 2001 From: Sebastian Manciulea Date: Tue, 8 Apr 2008 14:02:11 +0200 Subject: udf: Fix handling of multisession media According to OSTA UDF specification, only anchor blocks and primary volume descriptors are placed on media relative to the last session. All other block numbers are absolute (in the partition or the whole media). This seems to be confirmed by multisession media created by other systems. Signed-off-by: Sebastian Manciulea Signed-off-by: Jan Kara --- fs/udf/misc.c | 13 +++++-------- fs/udf/super.c | 2 +- 2 files changed, 6 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/udf/misc.c b/fs/udf/misc.c index 581b6e4cc59..96996204d92 100644 --- a/fs/udf/misc.c +++ b/fs/udf/misc.c @@ -203,16 +203,15 @@ struct buffer_head *udf_read_tagged(struct super_block *sb, uint32_t block, { tag *tag_p; struct buffer_head *bh = NULL; - struct udf_sb_info *sbi = UDF_SB(sb); /* Read the block */ if (block == 0xFFFFFFFF) return NULL; - bh = udf_tread(sb, block + sbi->s_session); + bh = udf_tread(sb, block); if (!bh) { udf_debug("block=%d, location=%d: read failed\n", - block + sbi->s_session, location); + block, location); return NULL; } @@ -222,8 +221,7 @@ struct buffer_head *udf_read_tagged(struct super_block *sb, uint32_t block, if (location != le32_to_cpu(tag_p->tagLocation)) { udf_debug("location mismatch block %u, tag %u != %u\n", - block + sbi->s_session, - le32_to_cpu(tag_p->tagLocation), location); + block, le32_to_cpu(tag_p->tagLocation), location); goto error_out; } @@ -247,9 +245,8 @@ struct buffer_head *udf_read_tagged(struct super_block *sb, uint32_t block, le16_to_cpu(tag_p->descCRCLength), 0)) return bh; - udf_debug("Crc failure block %d: crc = %d, crclen = %d\n", - block + sbi->s_session, le16_to_cpu(tag_p->descCRC), - le16_to_cpu(tag_p->descCRCLength)); + udf_debug("Crc failure block %d: crc = %d, crclen = %d\n", block, + le16_to_cpu(tag_p->descCRC), le16_to_cpu(tag_p->descCRCLength)); error_out: brelse(bh); diff --git a/fs/udf/super.c b/fs/udf/super.c index 650f20fe9d6..787cedf6cc0 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -587,7 +587,7 @@ static int udf_remount_fs(struct super_block *sb, int *flags, char *options) static int udf_vrs(struct super_block *sb, int silent) { struct volStructDesc *vsd = NULL; - int sector = 32768; + loff_t sector = 32768; int sectorsize; struct buffer_head *bh = NULL; int iso9660 = 0; -- cgit v1.2.3 From bfb257a5981af805a9394f00f75d3d9f7b611cc0 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Tue, 8 Apr 2008 20:37:21 +0200 Subject: udf: Add read-only support for 2.50 UDF media This patch implements parsing of metadata partitions and reading of Metadata File thus allowing to read UDF 2.50 media. Error resilience is implemented through accessing the Metadata Mirror File in case the data the Metadata File cannot be read. The patch is based on the original patch by Sebastian Manciulea and Mircea Fedoreanu . Signed-off-by: Sebastian Manciulea Signed-off-by: Mircea Fedoreanu Signed-off-by: Jan Kara --- fs/udf/inode.c | 9 +++ fs/udf/partition.c | 55 ++++++++++++++++ fs/udf/super.c | 190 +++++++++++++++++++++++++++++++++++++++++++++++++---- fs/udf/udf_sb.h | 16 ++++- fs/udf/udfdecl.h | 2 + 5 files changed, 258 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/udf/inode.c b/fs/udf/inode.c index c150b6df626..6e151f170c0 100644 --- a/fs/udf/inode.c +++ b/fs/udf/inode.c @@ -1301,6 +1301,15 @@ static void udf_fill_inode(struct inode *inode, struct buffer_head *bh) inode->i_op = &page_symlink_inode_operations; inode->i_mode = S_IFLNK | S_IRWXUGO; break; + case ICBTAG_FILE_TYPE_MAIN: + udf_debug("METADATA FILE-----\n"); + break; + case ICBTAG_FILE_TYPE_MIRROR: + udf_debug("METADATA MIRROR FILE-----\n"); + break; + case ICBTAG_FILE_TYPE_BITMAP: + udf_debug("METADATA BITMAP FILE-----\n"); + break; default: printk(KERN_ERR "udf: udf_fill_inode(ino %ld) failed unknown " "file type=%d\n", inode->i_ino, diff --git a/fs/udf/partition.c b/fs/udf/partition.c index b2e6e1eddb9..2dfe4be2eeb 100644 --- a/fs/udf/partition.c +++ b/fs/udf/partition.c @@ -266,3 +266,58 @@ int udf_relocate_blocks(struct super_block *sb, long old_block, long *new_block) return 0; } + +static uint32_t udf_try_read_meta(struct inode *inode, uint32_t block, + uint16_t partition, uint32_t offset) +{ + struct super_block *sb = inode->i_sb; + struct udf_part_map *map; + kernel_lb_addr eloc; + uint32_t elen; + sector_t ext_offset; + struct extent_position epos = {}; + uint32_t phyblock; + + if (inode_bmap(inode, block, &epos, &eloc, &elen, &ext_offset) != + (EXT_RECORDED_ALLOCATED >> 30)) + phyblock = 0xFFFFFFFF; + else { + map = &UDF_SB(sb)->s_partmaps[partition]; + /* map to sparable/physical partition desc */ + phyblock = udf_get_pblock(sb, eloc.logicalBlockNum, + map->s_partition_num, ext_offset + offset); + } + + brelse(epos.bh); + return phyblock; +} + +uint32_t udf_get_pblock_meta25(struct super_block *sb, uint32_t block, + uint16_t partition, uint32_t offset) +{ + struct udf_sb_info *sbi = UDF_SB(sb); + struct udf_part_map *map; + struct udf_meta_data *mdata; + uint32_t retblk; + struct inode *inode; + + udf_debug("READING from METADATA\n"); + + map = &sbi->s_partmaps[partition]; + mdata = &map->s_type_specific.s_metadata; + inode = mdata->s_metadata_fe ? : mdata->s_mirror_fe; + + /* We shouldn't mount such media... */ + BUG_ON(!inode); + retblk = udf_try_read_meta(inode, block, partition, offset); + if (retblk == 0xFFFFFFFF) { + udf_warning(sb, __func__, "error reading from METADATA, " + "trying to read from MIRROR"); + inode = mdata->s_mirror_fe; + if (!inode) + return 0xFFFFFFFF; + retblk = udf_try_read_meta(inode, block, partition, offset); + } + + return retblk; +} diff --git a/fs/udf/super.c b/fs/udf/super.c index 787cedf6cc0..29b19678327 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -950,6 +950,101 @@ static int udf_load_pvoldesc(struct super_block *sb, sector_t block) return 0; } +static int udf_load_metadata_files(struct super_block *sb, int partition) +{ + struct udf_sb_info *sbi = UDF_SB(sb); + struct udf_part_map *map; + struct udf_meta_data *mdata; + kernel_lb_addr addr; + int fe_error = 0; + + map = &sbi->s_partmaps[partition]; + mdata = &map->s_type_specific.s_metadata; + + /* metadata address */ + addr.logicalBlockNum = mdata->s_meta_file_loc; + addr.partitionReferenceNum = map->s_partition_num; + + udf_debug("Metadata file location: block = %d part = %d\n", + addr.logicalBlockNum, addr.partitionReferenceNum); + + mdata->s_metadata_fe = udf_iget(sb, addr); + + if (mdata->s_metadata_fe == NULL) { + udf_warning(sb, __func__, "metadata inode efe not found, " + "will try mirror inode."); + fe_error = 1; + } else if (UDF_I(mdata->s_metadata_fe)->i_alloc_type != + ICBTAG_FLAG_AD_SHORT) { + udf_warning(sb, __func__, "metadata inode efe does not have " + "short allocation descriptors!"); + fe_error = 1; + iput(mdata->s_metadata_fe); + mdata->s_metadata_fe = NULL; + } + + /* mirror file entry */ + addr.logicalBlockNum = mdata->s_mirror_file_loc; + addr.partitionReferenceNum = map->s_partition_num; + + udf_debug("Mirror metadata file location: block = %d part = %d\n", + addr.logicalBlockNum, addr.partitionReferenceNum); + + mdata->s_mirror_fe = udf_iget(sb, addr); + + if (mdata->s_mirror_fe == NULL) { + if (fe_error) { + udf_error(sb, __func__, "mirror inode efe not found " + "and metadata inode is missing too, exiting..."); + goto error_exit; + } else + udf_warning(sb, __func__, "mirror inode efe not found," + " but metadata inode is OK"); + } else if (UDF_I(mdata->s_mirror_fe)->i_alloc_type != + ICBTAG_FLAG_AD_SHORT) { + udf_warning(sb, __func__, "mirror inode efe does not have " + "short allocation descriptors!"); + iput(mdata->s_mirror_fe); + mdata->s_mirror_fe = NULL; + if (fe_error) + goto error_exit; + } + + /* + * bitmap file entry + * Note: + * Load only if bitmap file location differs from 0xFFFFFFFF (DCN-5102) + */ + if (mdata->s_bitmap_file_loc != 0xFFFFFFFF) { + addr.logicalBlockNum = mdata->s_bitmap_file_loc; + addr.partitionReferenceNum = map->s_partition_num; + + udf_debug("Bitmap file location: block = %d part = %d\n", + addr.logicalBlockNum, addr.partitionReferenceNum); + + mdata->s_bitmap_fe = udf_iget(sb, addr); + + if (mdata->s_bitmap_fe == NULL) { + if (sb->s_flags & MS_RDONLY) + udf_warning(sb, __func__, "bitmap inode efe " + "not found but it's ok since the disc" + " is mounted read-only"); + else { + udf_error(sb, __func__, "bitmap inode efe not " + "found and attempted read-write mount"); + goto error_exit; + } + } + } + + udf_debug("udf_load_metadata_files Ok\n"); + + return 0; + +error_exit: + return 1; +} + static void udf_load_fileset(struct super_block *sb, struct buffer_head *bh, kernel_lb_addr *root) { @@ -1169,7 +1264,7 @@ static int udf_load_partdesc(struct super_block *sb, sector_t block) p = (struct partitionDesc *)bh->b_data; partitionNumber = le16_to_cpu(p->partitionNumber); - /* First scan for TYPE1 and SPARABLE partitions */ + /* First scan for TYPE1, SPARABLE and METADATA partitions */ for (i = 0; i < sbi->s_partitions; i++) { map = &sbi->s_partmaps[i]; udf_debug("Searching map: (%d == %d)\n", @@ -1189,8 +1284,8 @@ static int udf_load_partdesc(struct super_block *sb, sector_t block) ret = udf_fill_partdesc_info(sb, p, i); /* - * Now rescan for VIRTUAL partitions when TYPE1 partitions are - * already set up + * Now rescan for VIRTUAL or METADATA partitions when SPARABLE and + * PHYSICAL partitions are already set up */ type1_idx = i; for (i = 0; i < sbi->s_partitions; i++) { @@ -1198,7 +1293,8 @@ static int udf_load_partdesc(struct super_block *sb, sector_t block) if (map->s_partition_num == partitionNumber && (map->s_partition_type == UDF_VIRTUAL_MAP15 || - map->s_partition_type == UDF_VIRTUAL_MAP20)) + map->s_partition_type == UDF_VIRTUAL_MAP20 || + map->s_partition_type == UDF_METADATA_MAP25)) break; } @@ -1208,16 +1304,28 @@ static int udf_load_partdesc(struct super_block *sb, sector_t block) ret = udf_fill_partdesc_info(sb, p, i); if (ret) goto out_bh; - /* - * Mark filesystem read-only if we have a partition with virtual map - * since we don't handle writing to it (we overwrite blocks instead of - * relocating them). - */ - sb->s_flags |= MS_RDONLY; - printk(KERN_NOTICE "UDF-fs: Filesystem marked read-only because " - "writing to pseudooverwrite partition is not implemented.\n"); - ret = udf_load_vat(sb, i, type1_idx); + if (map->s_partition_type == UDF_METADATA_MAP25) { + ret = udf_load_metadata_files(sb, i); + if (ret) { + printk(KERN_ERR "UDF-fs: error loading MetaData " + "partition map %d\n", i); + goto out_bh; + } + } else { + ret = udf_load_vat(sb, i, type1_idx); + if (ret) + goto out_bh; + /* + * Mark filesystem read-only if we have a partition with + * virtual map since we don't handle writing to it (we + * overwrite blocks instead of relocating them). + */ + sb->s_flags |= MS_RDONLY; + printk(KERN_NOTICE "UDF-fs: Filesystem marked read-only " + "because writing to pseudooverwrite partition is " + "not implemented.\n"); + } out_bh: /* In case loading failed, we handle cleanup in udf_fill_super */ brelse(bh); @@ -1316,6 +1424,50 @@ static int udf_load_logicalvol(struct super_block *sb, sector_t block, } } map->s_partition_func = udf_get_pblock_spar15; + } else if (!strncmp(upm2->partIdent.ident, + UDF_ID_METADATA, + strlen(UDF_ID_METADATA))) { + struct udf_meta_data *mdata = + &map->s_type_specific.s_metadata; + struct metadataPartitionMap *mdm = + (struct metadataPartitionMap *) + &(lvd->partitionMaps[offset]); + udf_debug("Parsing Logical vol part %d " + "type %d id=%s\n", i, type, + UDF_ID_METADATA); + + map->s_partition_type = UDF_METADATA_MAP25; + map->s_partition_func = udf_get_pblock_meta25; + + mdata->s_meta_file_loc = + le32_to_cpu(mdm->metadataFileLoc); + mdata->s_mirror_file_loc = + le32_to_cpu(mdm->metadataMirrorFileLoc); + mdata->s_bitmap_file_loc = + le32_to_cpu(mdm->metadataBitmapFileLoc); + mdata->s_alloc_unit_size = + le32_to_cpu(mdm->allocUnitSize); + mdata->s_align_unit_size = + le16_to_cpu(mdm->alignUnitSize); + mdata->s_dup_md_flag = + mdm->flags & 0x01; + + udf_debug("Metadata Ident suffix=0x%x\n", + (le16_to_cpu( + ((__le16 *) + mdm->partIdent.identSuffix)[0]))); + udf_debug("Metadata part num=%d\n", + le16_to_cpu(mdm->partitionNum)); + udf_debug("Metadata part alloc unit size=%d\n", + le32_to_cpu(mdm->allocUnitSize)); + udf_debug("Metadata file loc=%d\n", + le32_to_cpu(mdm->metadataFileLoc)); + udf_debug("Mirror file loc=%d\n", + le32_to_cpu(mdm->metadataMirrorFileLoc)); + udf_debug("Bitmap file loc=%d\n", + le32_to_cpu(mdm->metadataBitmapFileLoc)); + udf_debug("Duplicate Flag: %d %d\n", + mdata->s_dup_md_flag, mdm->flags); } else { udf_debug("Unknown ident: %s\n", upm2->partIdent.ident); @@ -1677,6 +1829,7 @@ static void udf_sb_free_bitmap(struct udf_bitmap *bitmap) static void udf_free_partition(struct udf_part_map *map) { int i; + struct udf_meta_data *mdata; if (map->s_partition_flags & UDF_PART_FLAG_UNALLOC_TABLE) iput(map->s_uspace.s_table); @@ -1689,6 +1842,17 @@ static void udf_free_partition(struct udf_part_map *map) if (map->s_partition_type == UDF_SPARABLE_MAP15) for (i = 0; i < 4; i++) brelse(map->s_type_specific.s_sparing.s_spar_map[i]); + else if (map->s_partition_type == UDF_METADATA_MAP25) { + mdata = &map->s_type_specific.s_metadata; + iput(mdata->s_metadata_fe); + mdata->s_metadata_fe = NULL; + + iput(mdata->s_mirror_fe); + mdata->s_mirror_fe = NULL; + + iput(mdata->s_bitmap_fe); + mdata->s_bitmap_fe = NULL; + } } static int udf_fill_super(struct super_block *sb, void *options, int silent) diff --git a/fs/udf/udf_sb.h b/fs/udf/udf_sb.h index 8308a12e123..1c1c514a972 100644 --- a/fs/udf/udf_sb.h +++ b/fs/udf/udf_sb.h @@ -6,7 +6,7 @@ /* Since UDF 2.01 is ISO 13346 based... */ #define UDF_SUPER_MAGIC 0x15013346 -#define UDF_MAX_READ_VERSION 0x0201 +#define UDF_MAX_READ_VERSION 0x0250 #define UDF_MAX_WRITE_VERSION 0x0201 #define UDF_FLAG_USE_EXTENDED_FE 0 @@ -46,9 +46,22 @@ #define UDF_VIRTUAL_MAP15 0x1512U #define UDF_VIRTUAL_MAP20 0x2012U #define UDF_SPARABLE_MAP15 0x1522U +#define UDF_METADATA_MAP25 0x2511U #pragma pack(1) /* XXX(hch): Why? This file just defines in-core structures */ +struct udf_meta_data { + __u32 s_meta_file_loc; + __u32 s_mirror_file_loc; + __u32 s_bitmap_file_loc; + __u32 s_alloc_unit_size; + __u16 s_align_unit_size; + __u8 s_dup_md_flag; + struct inode *s_metadata_fe; + struct inode *s_mirror_fe; + struct inode *s_bitmap_fe; +}; + struct udf_sparing_data { __u16 s_packet_len; struct buffer_head *s_spar_map[4]; @@ -82,6 +95,7 @@ struct udf_part_map { union { struct udf_sparing_data s_sparing; struct udf_virtual_data s_virtual; + struct udf_meta_data s_metadata; } s_type_specific; __u32 (*s_partition_func)(struct super_block *, __u32, __u16, __u32); __u16 s_volumeseqnum; diff --git a/fs/udf/udfdecl.h b/fs/udf/udfdecl.h index 2cb2f5de424..30f664ab0cd 100644 --- a/fs/udf/udfdecl.h +++ b/fs/udf/udfdecl.h @@ -177,6 +177,8 @@ extern uint32_t udf_get_pblock_virt20(struct super_block *, uint32_t, uint16_t, uint32_t); extern uint32_t udf_get_pblock_spar15(struct super_block *, uint32_t, uint16_t, uint32_t); +extern uint32_t udf_get_pblock_meta25(struct super_block *, uint32_t, uint16_t, + uint32_t); extern int udf_relocate_blocks(struct super_block *, long, long *); /* unicode.c */ -- cgit v1.2.3 From 47c9358a015199ed37c66235a2238271a7c8041f Mon Sep 17 00:00:00 2001 From: Sebastian Manciulea Date: Mon, 14 Apr 2008 17:06:36 +0200 Subject: udf: Fix bug in VAT mapping code Fix mapping of blocks using VAT when it is stored in an inode. UDF_I(inode)->i_data already points to the beginning of VAT header so there's no need to add udf_ext0_offset(inode). Signed-off-by: Sebastian Manciulea Signed-off-by: Jan Kara --- fs/udf/partition.c | 3 ++- fs/udf/super.c | 6 ++---- 2 files changed, 4 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/udf/partition.c b/fs/udf/partition.c index 2dfe4be2eeb..63610f026ae 100644 --- a/fs/udf/partition.c +++ b/fs/udf/partition.c @@ -66,7 +66,8 @@ uint32_t udf_get_pblock_virt15(struct super_block *sb, uint32_t block, } if (iinfo->i_alloc_type == ICBTAG_FLAG_AD_IN_ICB) { - loc = le32_to_cpu(((__le32 *)iinfo->i_ext.i_data)[block]); + loc = le32_to_cpu(((__le32 *)(iinfo->i_ext.i_data + + vdata->s_start_offset))[block]); goto translate; } index = (sb->s_blocksize - vdata->s_start_offset) / sizeof(uint32_t); diff --git a/fs/udf/super.c b/fs/udf/super.c index 29b19678327..8f02b30c22e 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -1215,8 +1215,7 @@ static int udf_load_vat(struct super_block *sb, int p_index, int type1_index) return 1; if (map->s_partition_type == UDF_VIRTUAL_MAP15) { - map->s_type_specific.s_virtual.s_start_offset = - udf_ext0_offset(sbi->s_vat_inode); + map->s_type_specific.s_virtual.s_start_offset = 0; map->s_type_specific.s_virtual.s_num_entries = (sbi->s_vat_inode->i_size - 36) >> 2; } else if (map->s_partition_type == UDF_VIRTUAL_MAP20) { @@ -1233,8 +1232,7 @@ static int udf_load_vat(struct super_block *sb, int p_index, int type1_index) } map->s_type_specific.s_virtual.s_start_offset = - le16_to_cpu(vat20->lengthHeader) + - udf_ext0_offset(sbi->s_vat_inode); + le16_to_cpu(vat20->lengthHeader); map->s_type_specific.s_virtual.s_num_entries = (sbi->s_vat_inode->i_size - map->s_type_specific.s_virtual. -- cgit v1.2.3 From 706047a79725b585cf272fdefc234b31b6545c72 Mon Sep 17 00:00:00 2001 From: Sebastian Manciulea Date: Mon, 14 Apr 2008 17:13:01 +0200 Subject: udf: Fix compilation warnings when UDF debug is on Fix two compilation warnings (and actual bugs in message formatting) when UDF debugging is turned on. Signed-off-by: Sebastian Manciulea Signed-off-by: Jan Kara --- fs/udf/super.c | 7 ++++--- fs/udf/udfdecl.h | 2 +- 2 files changed, 5 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/udf/super.c b/fs/udf/super.c index 8f02b30c22e..6823733c012 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -608,7 +608,8 @@ static int udf_vrs(struct super_block *sb, int silent) sector += (sbi->s_session << sb->s_blocksize_bits); udf_debug("Starting at sector %u (%ld byte sectors)\n", - (sector >> sb->s_blocksize_bits), sb->s_blocksize); + (unsigned int)(sector >> sb->s_blocksize_bits), + sb->s_blocksize); /* Process the sequence (if applicable) */ for (; !nsr02 && !nsr03; sector += sectorsize) { /* Read a block */ @@ -1117,8 +1118,8 @@ static int udf_fill_partdesc_info(struct super_block *sb, if (p->accessType == cpu_to_le32(PD_ACCESS_TYPE_OVERWRITABLE)) map->s_partition_flags |= UDF_PART_FLAG_OVERWRITABLE; - udf_debug("Partition (%d:%d type %x) starts at physical %d, " - "block length %d\n", partitionNumber, p_index, + udf_debug("Partition (%d type %x) starts at physical %d, " + "block length %d\n", p_index, map->s_partition_type, map->s_partition_root, map->s_partition_len); diff --git a/fs/udf/udfdecl.h b/fs/udf/udfdecl.h index 30f664ab0cd..61897d312a2 100644 --- a/fs/udf/udfdecl.h +++ b/fs/udf/udfdecl.h @@ -16,7 +16,7 @@ #define UDF_PREALLOCATE #define UDF_DEFAULT_PREALLOC_BLOCKS 8 -#undef UDFFS_DEBUG +#define UDFFS_DEBUG #ifdef UDFFS_DEBUG #define udf_debug(f, a...) \ -- cgit v1.2.3 From f845fced913b1437659bb5baf187698547697afe Mon Sep 17 00:00:00 2001 From: Bob Copeland Date: Thu, 17 Apr 2008 09:47:48 +0200 Subject: udf: use crc_itu_t from lib instead of udf_crc As pointed out by Sergey Vlasov, UDF implements its own version of the CRC ITU-T V.41. Convert it to use the one in the library. Signed-off-by: Bob Copeland Cc: Sergey Vlasov Signed-off-by: Jan Kara --- fs/Kconfig | 1 + fs/udf/Makefile | 2 +- fs/udf/crc.c | 172 ------------------------------------------------------- fs/udf/inode.c | 11 ++-- fs/udf/misc.c | 12 ++-- fs/udf/namei.c | 21 ++++--- fs/udf/super.c | 10 ++-- fs/udf/udfdecl.h | 3 - fs/udf/unicode.c | 3 +- 9 files changed, 32 insertions(+), 203 deletions(-) delete mode 100644 fs/udf/crc.c (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index d7312825592..86229c913ac 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -663,6 +663,7 @@ config ZISOFS config UDF_FS tristate "UDF file system support" + select CRC_ITU_T help This is the new file system used on some CD-ROMs and DVDs. Say Y if you intend to mount DVD discs or CDRW's written in packet mode, or diff --git a/fs/udf/Makefile b/fs/udf/Makefile index be845e7540e..0d4503f7446 100644 --- a/fs/udf/Makefile +++ b/fs/udf/Makefile @@ -6,4 +6,4 @@ obj-$(CONFIG_UDF_FS) += udf.o udf-objs := balloc.o dir.o file.o ialloc.o inode.o lowlevel.o namei.o \ partition.o super.o truncate.o symlink.o fsync.o \ - crc.o directory.o misc.o udftime.o unicode.o + directory.o misc.o udftime.o unicode.o diff --git a/fs/udf/crc.c b/fs/udf/crc.c deleted file mode 100644 index f178c63686e..00000000000 --- a/fs/udf/crc.c +++ /dev/null @@ -1,172 +0,0 @@ -/* - * crc.c - * - * PURPOSE - * Routines to generate, calculate, and test a 16-bit CRC. - * - * DESCRIPTION - * The CRC code was devised by Don P. Mitchell of AT&T Bell Laboratories - * and Ned W. Rhodes of Software Systems Group. It has been published in - * "Design and Validation of Computer Protocols", Prentice Hall, - * Englewood Cliffs, NJ, 1991, Chapter 3, ISBN 0-13-539925-4. - * - * Copyright is held by AT&T. - * - * AT&T gives permission for the free use of the CRC source code. - * - * COPYRIGHT - * This file is distributed under the terms of the GNU General Public - * License (GPL). Copies of the GPL can be obtained from: - * ftp://prep.ai.mit.edu/pub/gnu/GPL - * Each contributing author retains all rights to their own work. - */ - -#include "udfdecl.h" - -static const uint16_t crc_table[256] = { - 0x0000U, 0x1021U, 0x2042U, 0x3063U, 0x4084U, 0x50a5U, 0x60c6U, 0x70e7U, - 0x8108U, 0x9129U, 0xa14aU, 0xb16bU, 0xc18cU, 0xd1adU, 0xe1ceU, 0xf1efU, - 0x1231U, 0x0210U, 0x3273U, 0x2252U, 0x52b5U, 0x4294U, 0x72f7U, 0x62d6U, - 0x9339U, 0x8318U, 0xb37bU, 0xa35aU, 0xd3bdU, 0xc39cU, 0xf3ffU, 0xe3deU, - 0x2462U, 0x3443U, 0x0420U, 0x1401U, 0x64e6U, 0x74c7U, 0x44a4U, 0x5485U, - 0xa56aU, 0xb54bU, 0x8528U, 0x9509U, 0xe5eeU, 0xf5cfU, 0xc5acU, 0xd58dU, - 0x3653U, 0x2672U, 0x1611U, 0x0630U, 0x76d7U, 0x66f6U, 0x5695U, 0x46b4U, - 0xb75bU, 0xa77aU, 0x9719U, 0x8738U, 0xf7dfU, 0xe7feU, 0xd79dU, 0xc7bcU, - 0x48c4U, 0x58e5U, 0x6886U, 0x78a7U, 0x0840U, 0x1861U, 0x2802U, 0x3823U, - 0xc9ccU, 0xd9edU, 0xe98eU, 0xf9afU, 0x8948U, 0x9969U, 0xa90aU, 0xb92bU, - 0x5af5U, 0x4ad4U, 0x7ab7U, 0x6a96U, 0x1a71U, 0x0a50U, 0x3a33U, 0x2a12U, - 0xdbfdU, 0xcbdcU, 0xfbbfU, 0xeb9eU, 0x9b79U, 0x8b58U, 0xbb3bU, 0xab1aU, - 0x6ca6U, 0x7c87U, 0x4ce4U, 0x5cc5U, 0x2c22U, 0x3c03U, 0x0c60U, 0x1c41U, - 0xedaeU, 0xfd8fU, 0xcdecU, 0xddcdU, 0xad2aU, 0xbd0bU, 0x8d68U, 0x9d49U, - 0x7e97U, 0x6eb6U, 0x5ed5U, 0x4ef4U, 0x3e13U, 0x2e32U, 0x1e51U, 0x0e70U, - 0xff9fU, 0xefbeU, 0xdfddU, 0xcffcU, 0xbf1bU, 0xaf3aU, 0x9f59U, 0x8f78U, - 0x9188U, 0x81a9U, 0xb1caU, 0xa1ebU, 0xd10cU, 0xc12dU, 0xf14eU, 0xe16fU, - 0x1080U, 0x00a1U, 0x30c2U, 0x20e3U, 0x5004U, 0x4025U, 0x7046U, 0x6067U, - 0x83b9U, 0x9398U, 0xa3fbU, 0xb3daU, 0xc33dU, 0xd31cU, 0xe37fU, 0xf35eU, - 0x02b1U, 0x1290U, 0x22f3U, 0x32d2U, 0x4235U, 0x5214U, 0x6277U, 0x7256U, - 0xb5eaU, 0xa5cbU, 0x95a8U, 0x8589U, 0xf56eU, 0xe54fU, 0xd52cU, 0xc50dU, - 0x34e2U, 0x24c3U, 0x14a0U, 0x0481U, 0x7466U, 0x6447U, 0x5424U, 0x4405U, - 0xa7dbU, 0xb7faU, 0x8799U, 0x97b8U, 0xe75fU, 0xf77eU, 0xc71dU, 0xd73cU, - 0x26d3U, 0x36f2U, 0x0691U, 0x16b0U, 0x6657U, 0x7676U, 0x4615U, 0x5634U, - 0xd94cU, 0xc96dU, 0xf90eU, 0xe92fU, 0x99c8U, 0x89e9U, 0xb98aU, 0xa9abU, - 0x5844U, 0x4865U, 0x7806U, 0x6827U, 0x18c0U, 0x08e1U, 0x3882U, 0x28a3U, - 0xcb7dU, 0xdb5cU, 0xeb3fU, 0xfb1eU, 0x8bf9U, 0x9bd8U, 0xabbbU, 0xbb9aU, - 0x4a75U, 0x5a54U, 0x6a37U, 0x7a16U, 0x0af1U, 0x1ad0U, 0x2ab3U, 0x3a92U, - 0xfd2eU, 0xed0fU, 0xdd6cU, 0xcd4dU, 0xbdaaU, 0xad8bU, 0x9de8U, 0x8dc9U, - 0x7c26U, 0x6c07U, 0x5c64U, 0x4c45U, 0x3ca2U, 0x2c83U, 0x1ce0U, 0x0cc1U, - 0xef1fU, 0xff3eU, 0xcf5dU, 0xdf7cU, 0xaf9bU, 0xbfbaU, 0x8fd9U, 0x9ff8U, - 0x6e17U, 0x7e36U, 0x4e55U, 0x5e74U, 0x2e93U, 0x3eb2U, 0x0ed1U, 0x1ef0U -}; - -/* - * udf_crc - * - * PURPOSE - * Calculate a 16-bit CRC checksum using ITU-T V.41 polynomial. - * - * DESCRIPTION - * The OSTA-UDF(tm) 1.50 standard states that using CRCs is mandatory. - * The polynomial used is: x^16 + x^12 + x^15 + 1 - * - * PRE-CONDITIONS - * data Pointer to the data block. - * size Size of the data block. - * - * POST-CONDITIONS - * CRC of the data block. - * - * HISTORY - * July 21, 1997 - Andrew E. Mileski - * Adapted from OSTA-UDF(tm) 1.50 standard. - */ -uint16_t udf_crc(const uint8_t *data, uint32_t size, uint16_t crc) -{ - while (size--) - crc = crc_table[(crc >> 8 ^ *(data++)) & 0xffU] ^ (crc << 8); - - return crc; -} - -/****************************************************************************/ -#if defined(TEST) - -/* - * PURPOSE - * Test udf_crc() - * - * HISTORY - * July 21, 1997 - Andrew E. Mileski - * Adapted from OSTA-UDF(tm) 1.50 standard. - */ - -unsigned char bytes[] = { 0x70U, 0x6AU, 0x77U }; - -int main(void) -{ - unsigned short x; - - x = udf_crc(bytes, sizeof bytes); - printf("udf_crc: calculated = %4.4x, correct = %4.4x\n", x, 0x3299U); - - return 0; -} - -#endif /* defined(TEST) */ - -/****************************************************************************/ -#if defined(GENERATE) - -/* - * PURPOSE - * Generate a table for fast 16-bit CRC calculations (any polynomial). - * - * DESCRIPTION - * The ITU-T V.41 polynomial is 010041. - * - * HISTORY - * July 21, 1997 - Andrew E. Mileski - * Adapted from OSTA-UDF(tm) 1.50 standard. - */ - -#include - -int main(int argc, char **argv) -{ - unsigned long crc, poly; - int n, i; - - /* Get the polynomial */ - sscanf(argv[1], "%lo", &poly); - if (poly & 0xffff0000U) { - fprintf(stderr, "polynomial is too large\en"); - exit(1); - } - - printf("/* CRC 0%o */\n", poly); - - /* Create a table */ - printf("static unsigned short crc_table[256] = {\n"); - for (n = 0; n < 256; n++) { - if (n % 8 == 0) - printf("\t"); - crc = n << 8; - for (i = 0; i < 8; i++) { - if (crc & 0x8000U) - crc = (crc << 1) ^ poly; - else - crc <<= 1; - crc &= 0xFFFFU; - } - if (n == 255) - printf("0x%04xU ", crc); - else - printf("0x%04xU, ", crc); - if (n % 8 == 7) - printf("\n"); - } - printf("};\n"); - - return 0; -} - -#endif /* defined(GENERATE) */ diff --git a/fs/udf/inode.c b/fs/udf/inode.c index 6e151f170c0..6e74b117aaf 100644 --- a/fs/udf/inode.c +++ b/fs/udf/inode.c @@ -37,6 +37,7 @@ #include #include #include +#include #include "udf_i.h" #include "udf_sb.h" @@ -1419,9 +1420,9 @@ static int udf_update_inode(struct inode *inode, int do_sync) iinfo->i_location. logicalBlockNum); use->descTag.descCRCLength = cpu_to_le16(crclen); - use->descTag.descCRC = cpu_to_le16(udf_crc((char *)use + - sizeof(tag), crclen, - 0)); + use->descTag.descCRC = cpu_to_le16(crc_itu_t(0, (char *)use + + sizeof(tag), + crclen)); use->descTag.tagChecksum = udf_tag_checksum(&use->descTag); mark_buffer_dirty(bh); @@ -1584,8 +1585,8 @@ static int udf_update_inode(struct inode *inode, int do_sync) crclen += iinfo->i_lenEAttr + iinfo->i_lenAlloc - sizeof(tag); fe->descTag.descCRCLength = cpu_to_le16(crclen); - fe->descTag.descCRC = cpu_to_le16(udf_crc((char *)fe + sizeof(tag), - crclen, 0)); + fe->descTag.descCRC = cpu_to_le16(crc_itu_t(0, (char *)fe + sizeof(tag), + crclen)); fe->descTag.tagChecksum = udf_tag_checksum(&fe->descTag); /* write the data blocks */ diff --git a/fs/udf/misc.c b/fs/udf/misc.c index 96996204d92..84bf0fd4a4f 100644 --- a/fs/udf/misc.c +++ b/fs/udf/misc.c @@ -24,6 +24,7 @@ #include #include #include +#include #include "udf_i.h" #include "udf_sb.h" @@ -135,8 +136,8 @@ struct genericFormat *udf_add_extendedattr(struct inode *inode, uint32_t size, /* rewrite CRC + checksum of eahd */ crclen = sizeof(struct extendedAttrHeaderDesc) - sizeof(tag); eahd->descTag.descCRCLength = cpu_to_le16(crclen); - eahd->descTag.descCRC = cpu_to_le16(udf_crc((char *)eahd + - sizeof(tag), crclen, 0)); + eahd->descTag.descCRC = cpu_to_le16(crc_itu_t(0, (char *)eahd + + sizeof(tag), crclen)); eahd->descTag.tagChecksum = udf_tag_checksum(&eahd->descTag); iinfo->i_lenEAttr += size; return (struct genericFormat *)&ea[offset]; @@ -241,8 +242,9 @@ struct buffer_head *udf_read_tagged(struct super_block *sb, uint32_t block, /* Verify the descriptor CRC */ if (le16_to_cpu(tag_p->descCRCLength) + sizeof(tag) > sb->s_blocksize || - le16_to_cpu(tag_p->descCRC) == udf_crc(bh->b_data + sizeof(tag), - le16_to_cpu(tag_p->descCRCLength), 0)) + le16_to_cpu(tag_p->descCRC) == crc_itu_t(0, + bh->b_data + sizeof(tag), + le16_to_cpu(tag_p->descCRCLength))) return bh; udf_debug("Crc failure block %d: crc = %d, crclen = %d\n", block, @@ -266,7 +268,7 @@ void udf_update_tag(char *data, int length) length -= sizeof(tag); tptr->descCRCLength = cpu_to_le16(length); - tptr->descCRC = cpu_to_le16(udf_crc(data + sizeof(tag), length, 0)); + tptr->descCRC = cpu_to_le16(crc_itu_t(0, data + sizeof(tag), length)); tptr->tagChecksum = udf_tag_checksum(tptr); } diff --git a/fs/udf/namei.c b/fs/udf/namei.c index 68686b79650..ba5537d4bc1 100644 --- a/fs/udf/namei.c +++ b/fs/udf/namei.c @@ -31,6 +31,7 @@ #include #include #include +#include static inline int udf_match(int len1, const char *name1, int len2, const char *name2) @@ -97,25 +98,23 @@ int udf_write_fi(struct inode *inode, struct fileIdentDesc *cfi, memset(fibh->ebh->b_data, 0x00, padlen + offset); } - crc = udf_crc((uint8_t *)cfi + sizeof(tag), - sizeof(struct fileIdentDesc) - sizeof(tag), 0); + crc = crc_itu_t(0, (uint8_t *)cfi + sizeof(tag), + sizeof(struct fileIdentDesc) - sizeof(tag)); if (fibh->sbh == fibh->ebh) { - crc = udf_crc((uint8_t *)sfi->impUse, + crc = crc_itu_t(crc, (uint8_t *)sfi->impUse, crclen + sizeof(tag) - - sizeof(struct fileIdentDesc), crc); + sizeof(struct fileIdentDesc)); } else if (sizeof(struct fileIdentDesc) >= -fibh->soffset) { - crc = udf_crc(fibh->ebh->b_data + + crc = crc_itu_t(crc, fibh->ebh->b_data + sizeof(struct fileIdentDesc) + fibh->soffset, crclen + sizeof(tag) - - sizeof(struct fileIdentDesc), - crc); + sizeof(struct fileIdentDesc)); } else { - crc = udf_crc((uint8_t *)sfi->impUse, - -fibh->soffset - sizeof(struct fileIdentDesc), - crc); - crc = udf_crc(fibh->ebh->b_data, fibh->eoffset, crc); + crc = crc_itu_t(crc, (uint8_t *)sfi->impUse, + -fibh->soffset - sizeof(struct fileIdentDesc)); + crc = crc_itu_t(crc, fibh->ebh->b_data, fibh->eoffset); } cfi->descTag.descCRC = cpu_to_le16(crc); diff --git a/fs/udf/super.c b/fs/udf/super.c index 6823733c012..b564fc140fe 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -56,6 +56,7 @@ #include #include #include +#include #include #include "udf_sb.h" @@ -1765,8 +1766,8 @@ static void udf_open_lvid(struct super_block *sb) lvid->integrityType = LVID_INTEGRITY_TYPE_OPEN; lvid->descTag.descCRC = cpu_to_le16( - udf_crc((char *)lvid + sizeof(tag), - le16_to_cpu(lvid->descTag.descCRCLength), 0)); + crc_itu_t(0, (char *)lvid + sizeof(tag), + le16_to_cpu(lvid->descTag.descCRCLength))); lvid->descTag.tagChecksum = udf_tag_checksum(&lvid->descTag); mark_buffer_dirty(bh); @@ -1800,9 +1801,8 @@ static void udf_close_lvid(struct super_block *sb) lvid->integrityType = cpu_to_le32(LVID_INTEGRITY_TYPE_CLOSE); lvid->descTag.descCRC = cpu_to_le16( - udf_crc((char *)lvid + sizeof(tag), - le16_to_cpu(lvid->descTag.descCRCLength), - 0)); + crc_itu_t(0, (char *)lvid + sizeof(tag), + le16_to_cpu(lvid->descTag.descCRCLength))); lvid->descTag.tagChecksum = udf_tag_checksum(&lvid->descTag); mark_buffer_dirty(bh); diff --git a/fs/udf/udfdecl.h b/fs/udf/udfdecl.h index 61897d312a2..f3f45d02927 100644 --- a/fs/udf/udfdecl.h +++ b/fs/udf/udfdecl.h @@ -220,9 +220,6 @@ extern struct fileIdentDesc *udf_get_fileident(void *buffer, int bufsize, extern long_ad *udf_get_filelongad(uint8_t *, int, uint32_t *, int); extern short_ad *udf_get_fileshortad(uint8_t *, int, uint32_t *, int); -/* crc.c */ -extern uint16_t udf_crc(const uint8_t *, uint32_t, uint16_t); - /* udftime.c */ extern struct timespec *udf_disk_stamp_to_time(struct timespec *dest, timestamp src); diff --git a/fs/udf/unicode.c b/fs/udf/unicode.c index f5872ae325a..9fdf8c93c58 100644 --- a/fs/udf/unicode.c +++ b/fs/udf/unicode.c @@ -23,6 +23,7 @@ #include #include /* for memset */ #include +#include #include "udf_sb.h" @@ -454,7 +455,7 @@ static int udf_translate_to_linux(uint8_t *newName, uint8_t *udfName, } else if (newIndex > 250) newIndex = 250; newName[newIndex++] = CRC_MARK; - valueCRC = udf_crc(fidName, fidNameLen, 0); + valueCRC = crc_itu_t(0, fidName, fidNameLen); newName[newIndex++] = hexChar[(valueCRC & 0xf000) >> 12]; newName[newIndex++] = hexChar[(valueCRC & 0x0f00) >> 8]; newName[newIndex++] = hexChar[(valueCRC & 0x00f0) >> 4]; -- cgit v1.2.3 From 675c46796d601f63f2d47c6dd07332bca77ddae5 Mon Sep 17 00:00:00 2001 From: Steve French Date: Thu, 17 Apr 2008 23:41:01 +0000 Subject: [CIFS] Add various missing flags and defintions Signed-off-by: Steve French --- fs/cifs/README | 9 ++++- fs/cifs/cifspdu.h | 110 +++++++++++++++++++++++++++++++++++++++++++++++------- 2 files changed, 105 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/cifs/README b/fs/cifs/README index 50306229b0f..621aa1a8597 100644 --- a/fs/cifs/README +++ b/fs/cifs/README @@ -3,7 +3,14 @@ features such as hierarchical dfs like namespace, hardlinks, locking and more. It was designed to comply with the SNIA CIFS Technical Reference (which supersedes the 1992 X/Open SMB Standard) as well as to perform best practice practical interoperability with Windows 2000, Windows XP, Samba and equivalent -servers. +servers. This code was developed in participation with the Protocol Freedom +Information Foundation. + +Please see + http://protocolfreedom.org/ and + http://samba.org/samba/PFIF/ +for more details. + For questions or bug reports please contact: sfrench@samba.org (sfrench@us.ibm.com) diff --git a/fs/cifs/cifspdu.h b/fs/cifs/cifspdu.h index 47f79504f57..64bd3ed6803 100644 --- a/fs/cifs/cifspdu.h +++ b/fs/cifs/cifspdu.h @@ -1,7 +1,7 @@ /* * fs/cifs/cifspdu.h * - * Copyright (c) International Business Machines Corp., 2002,2007 + * Copyright (c) International Business Machines Corp., 2002,2008 * Author(s): Steve French (sfrench@us.ibm.com) * * This library is free software; you can redistribute it and/or modify @@ -163,7 +163,10 @@ path names in response */ #define SMBFLG2_KNOWS_EAS cpu_to_le16(2) #define SMBFLG2_SECURITY_SIGNATURE cpu_to_le16(4) +#define SMBFLG2_COMPRESSED (8) +#define SMBFLG2_SECURITY_SIGNATURE_REQUIRED (0x10) #define SMBFLG2_IS_LONG_NAME cpu_to_le16(0x40) +#define SMBFLG2_REPARSE_PATH (0x400) #define SMBFLG2_EXT_SEC cpu_to_le16(0x800) #define SMBFLG2_DFS cpu_to_le16(0x1000) #define SMBFLG2_PAGING_IO cpu_to_le16(0x2000) @@ -305,7 +308,7 @@ #define FILE_SHARE_DELETE 0x00000004 #define FILE_SHARE_ALL 0x00000007 -/* CreateDisposition flags */ +/* CreateDisposition flags, similar to CreateAction as well */ #define FILE_SUPERSEDE 0x00000000 #define FILE_OPEN 0x00000001 #define FILE_CREATE 0x00000002 @@ -317,15 +320,25 @@ #define CREATE_NOT_FILE 0x00000001 /* if set must not be file */ #define CREATE_WRITE_THROUGH 0x00000002 #define CREATE_SEQUENTIAL 0x00000004 -#define CREATE_SYNC_ALERT 0x00000010 -#define CREATE_ASYNC_ALERT 0x00000020 +#define CREATE_NO_BUFFER 0x00000008 /* should not buffer on srv */ +#define CREATE_SYNC_ALERT 0x00000010 /* MBZ */ +#define CREATE_ASYNC_ALERT 0x00000020 /* MBZ */ #define CREATE_NOT_DIR 0x00000040 /* if set must not be directory */ +#define CREATE_TREE_CONNECTION 0x00000080 /* should be zero */ +#define CREATE_COMPLETE_IF_OPLK 0x00000100 /* should be zero */ #define CREATE_NO_EA_KNOWLEDGE 0x00000200 -#define CREATE_EIGHT_DOT_THREE 0x00000400 +#define CREATE_EIGHT_DOT_THREE 0x00000400 /* doc says this is obsolete + open for recovery flag - should + be zero */ #define CREATE_RANDOM_ACCESS 0x00000800 #define CREATE_DELETE_ON_CLOSE 0x00001000 #define CREATE_OPEN_BY_ID 0x00002000 +#define CREATE_OPEN_BACKUP_INTN 0x00004000 +#define CREATE_NO_COMPRESSION 0x00008000 +#define CREATE_RESERVE_OPFILTER 0x00100000 /* should be zero */ #define OPEN_REPARSE_POINT 0x00200000 +#define OPEN_NO_RECALL 0x00400000 +#define OPEN_FREE_SPACE_QUERY 0x00800000 /* should be zero */ #define CREATE_OPTIONS_MASK 0x007FFFFF #define CREATE_OPTION_SPECIAL 0x20000000 /* system. NB not sent over wire */ @@ -470,7 +483,7 @@ typedef struct lanman_neg_rsp { typedef struct negotiate_rsp { struct smb_hdr hdr; /* wct = 17 */ - __le16 DialectIndex; + __le16 DialectIndex; /* 0xFFFF = no dialect acceptable */ __u8 SecurityMode; __le16 MaxMpxCount; __le16 MaxNumberVcs; @@ -516,10 +529,11 @@ typedef struct negotiate_rsp { #define CAP_INFOLEVEL_PASSTHRU 0x00002000 #define CAP_LARGE_READ_X 0x00004000 #define CAP_LARGE_WRITE_X 0x00008000 +#define CAP_LWIO 0x00010000 /* support fctl_srv_req_resume_key */ #define CAP_UNIX 0x00800000 -#define CAP_RESERVED 0x02000000 -#define CAP_BULK_TRANSFER 0x20000000 -#define CAP_COMPRESSED_DATA 0x40000000 +#define CAP_COMPRESSED_DATA 0x02000000 +#define CAP_DYNAMIC_REAUTH 0x20000000 +#define CAP_PERSISTENT_HANDLES 0x40000000 #define CAP_EXTENDED_SECURITY 0x80000000 typedef union smb_com_session_setup_andx { @@ -668,9 +682,7 @@ typedef struct smb_com_tconx_req { } __attribute__((packed)) TCONX_REQ; typedef struct smb_com_tconx_rsp { - struct smb_hdr hdr; /* wct = 3 note that Win2000 has sent wct = 7 - in some cases on responses. Four unspecified - words followed OptionalSupport */ + struct smb_hdr hdr; /* wct = 3 , not extended response */ __u8 AndXCommand; __u8 AndXReserved; __le16 AndXOffset; @@ -680,13 +692,48 @@ typedef struct smb_com_tconx_rsp { /* STRING NativeFileSystem */ } __attribute__((packed)) TCONX_RSP; +typedef struct smb_com_tconx_rsp_ext { + struct smb_hdr hdr; /* wct = 7, extended response */ + __u8 AndXCommand; + __u8 AndXReserved; + __le16 AndXOffset; + __le16 OptionalSupport; /* see below */ + __le32 MaximalShareAccessRights; + __le32 GuestMaximalShareAccessRights; + __u16 ByteCount; + unsigned char Service[1]; /* always ASCII, not Unicode */ + /* STRING NativeFileSystem */ +} __attribute__((packed)) TCONX_RSP_EXT; + + /* tree connect Flags */ #define DISCONNECT_TID 0x0001 +#define TCON_EXTENDED_SIGNATURES 0x0004 #define TCON_EXTENDED_SECINFO 0x0008 + /* OptionalSupport bits */ #define SMB_SUPPORT_SEARCH_BITS 0x0001 /* "must have" directory search bits (exclusive searches supported) */ #define SMB_SHARE_IS_IN_DFS 0x0002 +#define SMB_CSC_MASK 0x000C +/* CSC flags defined as follows */ +#define SMB_CSC_CACHE_MANUAL_REINT 0x0000 +#define SMB_CSC_CACHE_AUTO_REINT 0x0004 +#define SMB_CSC_CACHE_VDO 0x0008 +#define SMB_CSC_NO_CACHING 0x000C + +#define SMB_UNIQUE_FILE_NAME 0x0010 +#define SMB_EXTENDED_SIGNATURES 0x0020 + +/* services + * + * A: ie disk + * LPT1: ie printer + * IPC ie named pipe + * COMM + * ????? ie any type + * + */ typedef struct smb_com_logoff_andx_req { struct smb_hdr hdr; /* wct = 2 */ @@ -750,6 +797,17 @@ typedef struct smb_com_findclose_req { #define COMM_DEV_TYPE 0x0004 #define UNKNOWN_TYPE 0xFFFF +/* Device Type or File Status Flags */ +#define NO_EAS 0x0001 +#define NO_SUBSTREAMS 0x0002 +#define NO_REPARSETAG 0x0004 +/* following flags can apply if pipe */ +#define ICOUNT_MASK 0x00FF +#define PIPE_READ_MODE 0x0100 +#define NAMED_PIPE_TYPE 0x0400 +#define PIPE_END_POINT 0x0800 +#define BLOCKING_NAMED_PIPE 0x8000 + typedef struct smb_com_open_req { /* also handles create */ struct smb_hdr hdr; /* wct = 24 */ __u8 AndXCommand; @@ -758,7 +816,7 @@ typedef struct smb_com_open_req { /* also handles create */ __u8 Reserved; /* Must Be Zero */ __le16 NameLength; __le32 OpenFlags; - __le32 RootDirectoryFid; + __u32 RootDirectoryFid; __le32 DesiredAccess; __le64 AllocationSize; __le32 FileAttributes; @@ -801,6 +859,32 @@ typedef struct smb_com_open_rsp { __u16 ByteCount; /* bct = 0 */ } __attribute__((packed)) OPEN_RSP; +typedef struct smb_com_open_rsp_ext { + struct smb_hdr hdr; /* wct = 42 but meaningless due to MS bug? */ + __u8 AndXCommand; + __u8 AndXReserved; + __le16 AndXOffset; + __u8 OplockLevel; + __u16 Fid; + __le32 CreateAction; + __le64 CreationTime; + __le64 LastAccessTime; + __le64 LastWriteTime; + __le64 ChangeTime; + __le32 FileAttributes; + __le64 AllocationSize; + __le64 EndOfFile; + __le16 FileType; + __le16 DeviceState; + __u8 DirectoryFlag; + __u8 VolumeGUID[16]; + __u64 FileId; /* note no endian conversion - is opaque UniqueID */ + __le32 MaximalAccessRights; + __le32 GuestMaximalAccessRights; + __u16 ByteCount; /* bct = 0 */ +} __attribute__((packed)) OPEN_RSP_EXT; + + /* format of legacy open request */ typedef struct smb_com_openx_req { struct smb_hdr hdr; /* wct = 15 */ -- cgit v1.2.3 From e12070a5dca8bfeee352e9655ae79772a96b32f8 Mon Sep 17 00:00:00 2001 From: Niv Sardi Date: Thu, 6 Mar 2008 13:43:03 +1100 Subject: [XFS] actually check error returned by xfs_flush_pages, clean up and bailout if fails. SGI-PV: 973041 SGI-Modid: xfs-linux-melb:xfs-kern:30462a Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_bmap.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 2def273855a..87f64674981 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -5869,6 +5869,10 @@ xfs_getbmap( /* xfs_fsize_t last_byte = xfs_file_last_byte(ip); */ error = xfs_flush_pages(ip, (xfs_off_t)0, -1, 0, FI_REMAPF); + if (error) { + xfs_iunlock(ip, XFS_IOLOCK_SHARED); + return error; + } } ASSERT(whichfork == XFS_ATTR_FORK || ip->i_delayed_blks == 0); -- cgit v1.2.3 From 461aa8a22595e3bd3e6f4dc2894d7c4315ea2bb9 Mon Sep 17 00:00:00 2001 From: Lachlan McIlroy Date: Thu, 6 Mar 2008 13:43:11 +1100 Subject: [XFS] make inode reclaim synchronise with xfs_iflush_done() On a forced shutdown, xfs_finish_reclaim() will skip flushing the inode. If the inode flush lock is not already held and there is an outstanding xfs_iflush_done() then we might free the inode prematurely. By acquiring and releasing the flush lock we will synchronise with xfs_iflush_done(). SGI-PV: 909874 SGI-Modid: xfs-linux-melb:xfs-kern:30468a Signed-off-by: Lachlan McIlroy Signed-off-by: David Chinner --- fs/xfs/xfs_vnodeops.c | 22 ++++++++-------------- 1 file changed, 8 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index 64c5953feca..ce82a2d0acb 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -3694,12 +3694,12 @@ xfs_finish_reclaim( * We get the flush lock regardless, though, just to make sure * we don't free it while it is being flushed. */ - if (!XFS_FORCED_SHUTDOWN(ip->i_mount)) { - if (!locked) { - xfs_ilock(ip, XFS_ILOCK_EXCL); - xfs_iflock(ip); - } + if (!locked) { + xfs_ilock(ip, XFS_ILOCK_EXCL); + xfs_iflock(ip); + } + if (!XFS_FORCED_SHUTDOWN(ip->i_mount)) { if (ip->i_update_core || ((ip->i_itemp != NULL) && (ip->i_itemp->ili_format.ilf_fields != 0))) { @@ -3719,17 +3719,11 @@ xfs_finish_reclaim( ASSERT(ip->i_update_core == 0); ASSERT(ip->i_itemp == NULL || ip->i_itemp->ili_format.ilf_fields == 0); - xfs_iunlock(ip, XFS_ILOCK_EXCL); - } else if (locked) { - /* - * We are not interested in doing an iflush if we're - * in the process of shutting down the filesystem forcibly. - * So, just reclaim the inode. - */ - xfs_ifunlock(ip); - xfs_iunlock(ip, XFS_ILOCK_EXCL); } + xfs_ifunlock(ip); + xfs_iunlock(ip, XFS_ILOCK_EXCL); + reclaim: xfs_ireclaim(ip); return 0; -- cgit v1.2.3 From 163d3686bb09d88e2120bffe780a3f2d7cc4c948 Mon Sep 17 00:00:00 2001 From: Donald Douwsma Date: Thu, 6 Mar 2008 13:43:20 +1100 Subject: [XFS] Remove the xfs_refcache Remove the xfs_refcache, it was only needed while we were still building for 2.4 kernels. SGI-PV: 971186 SGI-Modid: xfs-linux-melb:xfs-kern:30472a Signed-off-by: Donald Douwsma Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_linux.h | 1 - fs/xfs/xfs_inode.h | 4 ---- fs/xfs/xfs_rename.c | 5 +---- fs/xfs/xfs_vfsops.c | 20 -------------------- fs/xfs/xfs_vnodeops.c | 34 +--------------------------------- 5 files changed, 2 insertions(+), 62 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_linux.h b/fs/xfs/linux-2.6/xfs_linux.h index 3ca39c4e5d2..e5143323e71 100644 --- a/fs/xfs/linux-2.6/xfs_linux.h +++ b/fs/xfs/linux-2.6/xfs_linux.h @@ -99,7 +99,6 @@ /* * Feature macros (disable/enable) */ -#undef HAVE_REFCACHE /* reference cache not needed for NFS in 2.6 */ #define HAVE_SPLICE /* a splice(2) exists in 2.6, but not in 2.4 */ #ifdef CONFIG_SMP #define HAVE_PERCPU_SB /* per cpu superblock counters are a 2.6 feature */ diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h index bfcd72cbaee..eaa01895ff9 100644 --- a/fs/xfs/xfs_inode.h +++ b/fs/xfs/xfs_inode.h @@ -240,10 +240,6 @@ typedef struct xfs_inode { atomic_t i_pincount; /* inode pin count */ wait_queue_head_t i_ipin_wait; /* inode pinning wait queue */ spinlock_t i_flags_lock; /* inode i_flags lock */ -#ifdef HAVE_REFCACHE - struct xfs_inode **i_refcache; /* ptr to entry in ref cache */ - struct xfs_inode *i_release; /* inode to unref */ -#endif /* Miscellaneous state. */ unsigned short i_flags; /* see defined flags below */ unsigned char i_update_core; /* timestamps/size is dirty */ diff --git a/fs/xfs/xfs_rename.c b/fs/xfs/xfs_rename.c index 7eb157a59f9..1c6d40ed681 100644 --- a/fs/xfs/xfs_rename.c +++ b/fs/xfs/xfs_rename.c @@ -36,7 +36,6 @@ #include "xfs_bmap.h" #include "xfs_error.h" #include "xfs_quota.h" -#include "xfs_refcache.h" #include "xfs_utils.h" #include "xfs_trans_space.h" #include "xfs_vnodeops.h" @@ -580,10 +579,8 @@ xfs_rename( * the vnode references. */ error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES); - if (target_ip != NULL) { - xfs_refcache_purge_ip(target_ip); + if (target_ip != NULL) IRELE(target_ip); - } /* * Let interposed file systems know about removed links. */ diff --git a/fs/xfs/xfs_vfsops.c b/fs/xfs/xfs_vfsops.c index 7094caff13c..79bdfb3d608 100644 --- a/fs/xfs/xfs_vfsops.c +++ b/fs/xfs/xfs_vfsops.c @@ -43,7 +43,6 @@ #include "xfs_error.h" #include "xfs_bmap.h" #include "xfs_rw.h" -#include "xfs_refcache.h" #include "xfs_buf_item.h" #include "xfs_log_priv.h" #include "xfs_dir2_trace.h" @@ -157,7 +156,6 @@ xfs_cleanup(void) xfs_cleanup_procfs(); xfs_sysctl_unregister(); - xfs_refcache_destroy(); xfs_filestream_uninit(); xfs_mru_cache_uninit(); xfs_acl_zone_destroy(xfs_acl_zone); @@ -584,11 +582,6 @@ xfs_unmount( 0 : DM_FLAGS_UNWANTED; } #endif - /* - * First blow any referenced inode from this file system - * out of the reference cache, and delete the timer. - */ - xfs_refcache_purge_mp(mp); /* * Blow away any referenced inode in the filestreams cache. @@ -652,7 +645,6 @@ xfs_quiesce_fs( { int count = 0, pincount; - xfs_refcache_purge_mp(mp); xfs_flush_buftarg(mp->m_ddev_targp, 0); xfs_finish_reclaim_all(mp, 0); @@ -1322,18 +1314,6 @@ xfs_syncsub( } } - /* - * If this is the periodic sync, then kick some entries out of - * the reference cache. This ensures that idle entries are - * eventually kicked out of the cache. - */ - if (flags & SYNC_REFCACHE) { - if (flags & SYNC_WAIT) - xfs_refcache_purge_mp(mp); - else - xfs_refcache_purge_some(mp); - } - /* * If asked, update the disk superblock with incore counter values if we * are using non-persistent counters so that they don't get too far out diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index ce82a2d0acb..35ac59dabf8 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -48,7 +48,6 @@ #include "xfs_quota.h" #include "xfs_utils.h" #include "xfs_rtalloc.h" -#include "xfs_refcache.h" #include "xfs_trans_space.h" #include "xfs_log_priv.h" #include "xfs_filestream.h" @@ -1520,12 +1519,6 @@ xfs_release( xfs_flush_pages(ip, 0, -1, XFS_B_ASYNC, FI_NONE); } -#ifdef HAVE_REFCACHE - /* If we are in the NFS reference cache then don't do this now */ - if (ip->i_refcache) - return 0; -#endif - if (ip->i_d.di_nlink != 0) { if ((((ip->i_d.di_mode & S_IFMT) == S_IFREG) && ((ip->i_size > 0) || (VN_CACHED(vp) > 0 || @@ -2448,14 +2441,6 @@ xfs_remove( goto std_return; } - /* - * Before we drop our extra reference to the inode, purge it - * from the refcache if it is there. By waiting until afterwards - * to do the IRELE, we ensure that we won't go inactive in the - * xfs_refcache_purge_ip routine (although that would be OK). - */ - xfs_refcache_purge_ip(ip); - /* * If we are using filestreams, kill the stream association. * If the file is still open it may get a new one but that @@ -2495,14 +2480,6 @@ xfs_remove( cancel_flags |= XFS_TRANS_ABORT; xfs_trans_cancel(tp, cancel_flags); - /* - * Before we drop our extra reference to the inode, purge it - * from the refcache if it is there. By waiting until afterwards - * to do the IRELE, we ensure that we won't go inactive in the - * xfs_refcache_purge_ip routine (although that would be OK). - */ - xfs_refcache_purge_ip(ip); - IRELE(ip); goto std_return; @@ -3460,16 +3437,7 @@ xfs_rwunlock( { if (S_ISDIR(ip->i_d.di_mode)) return; - if (locktype == VRWLOCK_WRITE) { - /* - * In the write case, we may have added a new entry to - * the reference cache. This might store a pointer to - * an inode to be released in this inode. If it is there, - * clear the pointer and release the inode after unlocking - * this one. - */ - xfs_refcache_iunlock(ip, XFS_IOLOCK_EXCL); - } else { + if (locktype != VRWLOCK_WRITE) { ASSERT((locktype == VRWLOCK_READ) || (locktype == VRWLOCK_WRITE_DIRECT)); xfs_iunlock(ip, XFS_IOLOCK_SHARED); -- cgit v1.2.3 From e9a56b7cdaf6129892fd7c8d950b71a1a4304bb0 Mon Sep 17 00:00:00 2001 From: Lachlan McIlroy Date: Thu, 6 Mar 2008 13:43:27 +1100 Subject: [XFS] Fix regression due to refcache removal SGI-PV: 971186 SGI-Modid: xfs-linux-melb:xfs-kern:30490a Signed-off-by: Lachlan McIlroy Signed-off-by: Donald Douwsma --- fs/xfs/xfs_vnodeops.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index 35ac59dabf8..40b95e3a88b 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -3437,7 +3437,9 @@ xfs_rwunlock( { if (S_ISDIR(ip->i_d.di_mode)) return; - if (locktype != VRWLOCK_WRITE) { + if (locktype == VRWLOCK_WRITE) { + xfs_iunlock(ip, XFS_IOLOCK_EXCL); + } else { ASSERT((locktype == VRWLOCK_READ) || (locktype == VRWLOCK_WRITE_DIRECT)); xfs_iunlock(ip, XFS_IOLOCK_SHARED); -- cgit v1.2.3 From 4ae29b4321b99b711bcfde5527c4fbf249eac60f Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 6 Mar 2008 13:43:34 +1100 Subject: [XFS] Factor xfs_itobp() and xfs_inotobp(). The only difference between the functions is one passes an inode for the lookup, the other passes an inode number. However, they don't do the same validity checking or set all the same state on the buffer that is returned yet they should. Factor the functions into a common implementation. SGI-PV: 970925 SGI-Modid: xfs-linux-melb:xfs-kern:30500a Signed-off-by: David Chinner Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_inode.c | 261 ++++++++++++++++++++++------------------------------- 1 file changed, 106 insertions(+), 155 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index f43a6e01d68..6f156faf9d4 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -125,6 +125,85 @@ xfs_inobp_check( } #endif +/* + * Find the buffer associated with the given inode map + * We do basic validation checks on the buffer once it has been + * retrieved from disk. + */ +STATIC int +xfs_imap_to_bp( + xfs_mount_t *mp, + xfs_trans_t *tp, + xfs_imap_t *imap, + xfs_buf_t **bpp, + uint buf_flags, + uint imap_flags) +{ + int error; + int i; + int ni; + xfs_buf_t *bp; + + error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, imap->im_blkno, + (int)imap->im_len, XFS_BUF_LOCK, &bp); + if (error) { + cmn_err(CE_WARN, "xfs_imap_to_bp: xfs_trans_read_buf()returned " + "an error %d on %s. Returning error.", + error, mp->m_fsname); + return error; + } + + /* + * Validate the magic number and version of every inode in the buffer + * (if DEBUG kernel) or the first inode in the buffer, otherwise. + */ +#ifdef DEBUG + ni = BBTOB(imap->im_len) >> mp->m_sb.sb_inodelog; +#else /* usual case */ + ni = 1; +#endif + + for (i = 0; i < ni; i++) { + int di_ok; + xfs_dinode_t *dip; + + dip = (xfs_dinode_t *)xfs_buf_offset(bp, + (i << mp->m_sb.sb_inodelog)); + di_ok = be16_to_cpu(dip->di_core.di_magic) == XFS_DINODE_MAGIC && + XFS_DINODE_GOOD_VERSION(dip->di_core.di_version); + if (unlikely(XFS_TEST_ERROR(!di_ok, mp, + XFS_ERRTAG_ITOBP_INOTOBP, + XFS_RANDOM_ITOBP_INOTOBP))) { + if (imap_flags & XFS_IMAP_BULKSTAT) { + xfs_trans_brelse(tp, bp); + return XFS_ERROR(EINVAL); + } + XFS_CORRUPTION_ERROR("xfs_imap_to_bp", + XFS_ERRLEVEL_HIGH, mp, dip); +#ifdef DEBUG + cmn_err(CE_PANIC, + "Device %s - bad inode magic/vsn " + "daddr %lld #%d (magic=%x)", + XFS_BUFTARG_NAME(mp->m_ddev_targp), + (unsigned long long)imap->im_blkno, i, + be16_to_cpu(dip->di_core.di_magic)); +#endif + xfs_trans_brelse(tp, bp); + return XFS_ERROR(EFSCORRUPTED); + } + } + + xfs_inobp_check(mp, bp); + + /* + * Mark the buffer as an inode buffer now that it looks good + */ + XFS_BUF_SET_VTYPE(bp, B_FS_INO); + + *bpp = bp; + return 0; +} + /* * This routine is called to map an inode number within a file * system to the buffer containing the on-disk version of the @@ -147,72 +226,19 @@ xfs_inotobp( xfs_buf_t **bpp, int *offset) { - int di_ok; xfs_imap_t imap; xfs_buf_t *bp; int error; - xfs_dinode_t *dip; - /* - * Call the space management code to find the location of the - * inode on disk. - */ imap.im_blkno = 0; error = xfs_imap(mp, tp, ino, &imap, XFS_IMAP_LOOKUP); - if (error != 0) { - cmn_err(CE_WARN, - "xfs_inotobp: xfs_imap() returned an " - "error %d on %s. Returning error.", error, mp->m_fsname); + if (error) return error; - } - - /* - * If the inode number maps to a block outside the bounds of the - * file system then return NULL rather than calling read_buf - * and panicing when we get an error from the driver. - */ - if ((imap.im_blkno + imap.im_len) > - XFS_FSB_TO_BB(mp, mp->m_sb.sb_dblocks)) { - cmn_err(CE_WARN, - "xfs_inotobp: inode number (%llu + %d) maps to a block outside the bounds " - "of the file system %s. Returning EINVAL.", - (unsigned long long)imap.im_blkno, - imap.im_len, mp->m_fsname); - return XFS_ERROR(EINVAL); - } - - /* - * Read in the buffer. If tp is NULL, xfs_trans_read_buf() will - * default to just a read_buf() call. - */ - error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, imap.im_blkno, - (int)imap.im_len, XFS_BUF_LOCK, &bp); - if (error) { - cmn_err(CE_WARN, - "xfs_inotobp: xfs_trans_read_buf() returned an " - "error %d on %s. Returning error.", error, mp->m_fsname); + error = xfs_imap_to_bp(mp, tp, &imap, &bp, XFS_BUF_LOCK, 0); + if (error) return error; - } - dip = (xfs_dinode_t *)xfs_buf_offset(bp, 0); - di_ok = - be16_to_cpu(dip->di_core.di_magic) == XFS_DINODE_MAGIC && - XFS_DINODE_GOOD_VERSION(dip->di_core.di_version); - if (unlikely(XFS_TEST_ERROR(!di_ok, mp, XFS_ERRTAG_ITOBP_INOTOBP, - XFS_RANDOM_ITOBP_INOTOBP))) { - XFS_CORRUPTION_ERROR("xfs_inotobp", XFS_ERRLEVEL_LOW, mp, dip); - xfs_trans_brelse(tp, bp); - cmn_err(CE_WARN, - "xfs_inotobp: XFS_TEST_ERROR() returned an " - "error on %s. Returning EFSCORRUPTED.", mp->m_fsname); - return XFS_ERROR(EFSCORRUPTED); - } - xfs_inobp_check(mp, bp); - - /* - * Set *dipp to point to the on-disk inode in the buffer. - */ *dipp = (xfs_dinode_t *)xfs_buf_offset(bp, imap.im_boffset); *bpp = bp; *offset = imap.im_boffset; @@ -253,40 +279,14 @@ xfs_itobp( xfs_imap_t imap; xfs_buf_t *bp; int error; - int i; - int ni; if (ip->i_blkno == (xfs_daddr_t)0) { - /* - * Call the space management code to find the location of the - * inode on disk. - */ imap.im_blkno = bno; - if ((error = xfs_imap(mp, tp, ip->i_ino, &imap, - XFS_IMAP_LOOKUP | imap_flags))) + error = xfs_imap(mp, tp, ip->i_ino, &imap, + XFS_IMAP_LOOKUP | imap_flags); + if (error) return error; - /* - * If the inode number maps to a block outside the bounds - * of the file system then return NULL rather than calling - * read_buf and panicing when we get an error from the - * driver. - */ - if ((imap.im_blkno + imap.im_len) > - XFS_FSB_TO_BB(mp, mp->m_sb.sb_dblocks)) { -#ifdef DEBUG - xfs_fs_cmn_err(CE_ALERT, mp, "xfs_itobp: " - "(imap.im_blkno (0x%llx) " - "+ imap.im_len (0x%llx)) > " - " XFS_FSB_TO_BB(mp, " - "mp->m_sb.sb_dblocks) (0x%llx)", - (unsigned long long) imap.im_blkno, - (unsigned long long) imap.im_len, - XFS_FSB_TO_BB(mp, mp->m_sb.sb_dblocks)); -#endif /* DEBUG */ - return XFS_ERROR(EINVAL); - } - /* * Fill in the fields in the inode that will be used to * map the inode to its buffer from now on. @@ -305,76 +305,10 @@ xfs_itobp( } ASSERT(bno == 0 || bno == imap.im_blkno); - /* - * Read in the buffer. If tp is NULL, xfs_trans_read_buf() will - * default to just a read_buf() call. - */ - error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, imap.im_blkno, - (int)imap.im_len, XFS_BUF_LOCK, &bp); - if (error) { -#ifdef DEBUG - xfs_fs_cmn_err(CE_ALERT, mp, "xfs_itobp: " - "xfs_trans_read_buf() returned error %d, " - "imap.im_blkno 0x%llx, imap.im_len 0x%llx", - error, (unsigned long long) imap.im_blkno, - (unsigned long long) imap.im_len); -#endif /* DEBUG */ + error = xfs_imap_to_bp(mp, tp, &imap, &bp, XFS_BUF_LOCK, imap_flags); + if (error) return error; - } - - /* - * Validate the magic number and version of every inode in the buffer - * (if DEBUG kernel) or the first inode in the buffer, otherwise. - * No validation is done here in userspace (xfs_repair). - */ -#if !defined(__KERNEL__) - ni = 0; -#elif defined(DEBUG) - ni = BBTOB(imap.im_len) >> mp->m_sb.sb_inodelog; -#else /* usual case */ - ni = 1; -#endif - - for (i = 0; i < ni; i++) { - int di_ok; - xfs_dinode_t *dip; - dip = (xfs_dinode_t *)xfs_buf_offset(bp, - (i << mp->m_sb.sb_inodelog)); - di_ok = be16_to_cpu(dip->di_core.di_magic) == XFS_DINODE_MAGIC && - XFS_DINODE_GOOD_VERSION(dip->di_core.di_version); - if (unlikely(XFS_TEST_ERROR(!di_ok, mp, - XFS_ERRTAG_ITOBP_INOTOBP, - XFS_RANDOM_ITOBP_INOTOBP))) { - if (imap_flags & XFS_IMAP_BULKSTAT) { - xfs_trans_brelse(tp, bp); - return XFS_ERROR(EINVAL); - } -#ifdef DEBUG - cmn_err(CE_ALERT, - "Device %s - bad inode magic/vsn " - "daddr %lld #%d (magic=%x)", - XFS_BUFTARG_NAME(mp->m_ddev_targp), - (unsigned long long)imap.im_blkno, i, - be16_to_cpu(dip->di_core.di_magic)); -#endif - XFS_CORRUPTION_ERROR("xfs_itobp", XFS_ERRLEVEL_HIGH, - mp, dip); - xfs_trans_brelse(tp, bp); - return XFS_ERROR(EFSCORRUPTED); - } - } - - xfs_inobp_check(mp, bp); - - /* - * Mark the buffer as an inode buffer now that it looks good - */ - XFS_BUF_SET_VTYPE(bp, B_FS_INO); - - /* - * Set *dipp to point to the on-disk inode in the buffer. - */ *dipp = (xfs_dinode_t *)xfs_buf_offset(bp, imap.im_boffset); *bpp = bp; return 0; @@ -2678,14 +2612,31 @@ xfs_imap( fsbno = imap->im_blkno ? XFS_DADDR_TO_FSB(mp, imap->im_blkno) : NULLFSBLOCK; error = xfs_dilocate(mp, tp, ino, &fsbno, &len, &off, flags); - if (error != 0) { + if (error) return error; - } + imap->im_blkno = XFS_FSB_TO_DADDR(mp, fsbno); imap->im_len = XFS_FSB_TO_BB(mp, len); imap->im_agblkno = XFS_FSB_TO_AGBNO(mp, fsbno); imap->im_ioffset = (ushort)off; imap->im_boffset = (ushort)(off << mp->m_sb.sb_inodelog); + + /* + * If the inode number maps to a block outside the bounds + * of the file system then return NULL rather than calling + * read_buf and panicing when we get an error from the + * driver. + */ + if ((imap->im_blkno + imap->im_len) > + XFS_FSB_TO_BB(mp, mp->m_sb.sb_dblocks)) { + xfs_fs_cmn_err(CE_ALERT, mp, "xfs_imap: " + "(imap->im_blkno (0x%llx) + imap->im_len (0x%llx)) > " + " XFS_FSB_TO_BB(mp, mp->m_sb.sb_dblocks) (0x%llx)", + (unsigned long long) imap->im_blkno, + (unsigned long long) imap->im_len, + XFS_FSB_TO_BB(mp, mp->m_sb.sb_dblocks)); + return EINVAL; + } return 0; } -- cgit v1.2.3 From a3f74ffb6d1448d9a8f482e593b80ec15f1695d4 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 6 Mar 2008 13:43:42 +1100 Subject: [XFS] Don't block pdflush when writing back inodes When pdflush is writing back inodes, it can get stuck on inode cluster buffers that are currently under I/O. This occurs when we write data to multiple inodes in the same inode cluster at the same time. Effectively, delayed allocation marks the inode dirty during the data writeback. Hence if the inode cluster was flushed during the writeback of the first inode, the writeback of the second inode will block waiting for the inode cluster write to complete before writing it again for the newly dirtied inode. Basically, we want to avoid this from happening so we don't block pdflush and slow down all of writeback. Hence we introduce a non-blocking async inode flush flag that pdflush uses. If this flag is set, we use non-blocking operations (e.g. try locks) whereever we can to avoid blocking or extra I/O being issued. SGI-PV: 970925 SGI-Modid: xfs-linux-melb:xfs-kern:30501a Signed-off-by: David Chinner Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_super.c | 3 +- fs/xfs/linux-2.6/xfs_vnode.h | 5 +- fs/xfs/xfs_inode.c | 135 ++++++++++++++++++++++++++----------------- fs/xfs/xfs_inode.h | 3 +- fs/xfs/xfs_itable.c | 3 +- fs/xfs/xfs_log_recover.c | 3 +- fs/xfs/xfs_trans_buf.c | 3 +- fs/xfs/xfs_vnodeops.c | 55 ++++-------------- 8 files changed, 105 insertions(+), 105 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c index 8831d951879..cb9ce90d1de 100644 --- a/fs/xfs/linux-2.6/xfs_super.c +++ b/fs/xfs/linux-2.6/xfs_super.c @@ -896,7 +896,8 @@ xfs_fs_write_inode( struct inode *inode, int sync) { - int error = 0, flags = FLUSH_INODE; + int error = 0; + int flags = 0; xfs_itrace_entry(XFS_I(inode)); if (sync) { diff --git a/fs/xfs/linux-2.6/xfs_vnode.h b/fs/xfs/linux-2.6/xfs_vnode.h index b5ea418693b..f200e024408 100644 --- a/fs/xfs/linux-2.6/xfs_vnode.h +++ b/fs/xfs/linux-2.6/xfs_vnode.h @@ -73,12 +73,9 @@ typedef enum bhv_vrwlock { #define IO_INVIS 0x00020 /* don't update inode timestamps */ /* - * Flags for vop_iflush call + * Flags for xfs_inode_flush */ #define FLUSH_SYNC 1 /* wait for flush to complete */ -#define FLUSH_INODE 2 /* flush the inode itself */ -#define FLUSH_LOG 4 /* force the last log entry for - * this inode out to disk */ /* * Flush/Invalidate options for vop_toss/flush/flushinval_pages. diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 6f156faf9d4..3c3e9e3c1da 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -145,11 +145,16 @@ xfs_imap_to_bp( xfs_buf_t *bp; error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, imap->im_blkno, - (int)imap->im_len, XFS_BUF_LOCK, &bp); + (int)imap->im_len, buf_flags, &bp); if (error) { - cmn_err(CE_WARN, "xfs_imap_to_bp: xfs_trans_read_buf()returned " + if (error != EAGAIN) { + cmn_err(CE_WARN, + "xfs_imap_to_bp: xfs_trans_read_buf()returned " "an error %d on %s. Returning error.", error, mp->m_fsname); + } else { + ASSERT(buf_flags & XFS_BUF_TRYLOCK); + } return error; } @@ -274,7 +279,8 @@ xfs_itobp( xfs_dinode_t **dipp, xfs_buf_t **bpp, xfs_daddr_t bno, - uint imap_flags) + uint imap_flags, + uint buf_flags) { xfs_imap_t imap; xfs_buf_t *bp; @@ -305,10 +311,17 @@ xfs_itobp( } ASSERT(bno == 0 || bno == imap.im_blkno); - error = xfs_imap_to_bp(mp, tp, &imap, &bp, XFS_BUF_LOCK, imap_flags); + error = xfs_imap_to_bp(mp, tp, &imap, &bp, buf_flags, imap_flags); if (error) return error; + if (!bp) { + ASSERT(buf_flags & XFS_BUF_TRYLOCK); + ASSERT(tp == NULL); + *bpp = NULL; + return EAGAIN; + } + *dipp = (xfs_dinode_t *)xfs_buf_offset(bp, imap.im_boffset); *bpp = bp; return 0; @@ -812,7 +825,7 @@ xfs_iread( * return NULL as well. Set i_blkno to 0 so that xfs_itobp() will * know that this is a new incore inode. */ - error = xfs_itobp(mp, tp, ip, &dip, &bp, bno, imap_flags); + error = xfs_itobp(mp, tp, ip, &dip, &bp, bno, imap_flags, XFS_BUF_LOCK); if (error) { kmem_zone_free(xfs_inode_zone, ip); return error; @@ -1901,7 +1914,7 @@ xfs_iunlink( * Here we put the head pointer into our next pointer, * and then we fall through to point the head at us. */ - error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0, 0); + error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0, 0, XFS_BUF_LOCK); if (error) return error; @@ -2009,7 +2022,7 @@ xfs_iunlink_remove( * of dealing with the buffer when there is no need to * change it. */ - error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0, 0); + error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0, 0, XFS_BUF_LOCK); if (error) { cmn_err(CE_WARN, "xfs_iunlink_remove: xfs_itobp() returned an error %d on %s. Returning error.", @@ -2071,7 +2084,7 @@ xfs_iunlink_remove( * Now last_ibp points to the buffer previous to us on * the unlinked list. Pull us from the list. */ - error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0, 0); + error = xfs_itobp(mp, tp, ip, &dip, &ibp, 0, 0, XFS_BUF_LOCK); if (error) { cmn_err(CE_WARN, "xfs_iunlink_remove: xfs_itobp() returned an error %d on %s. Returning error.", @@ -2334,7 +2347,7 @@ xfs_ifree( xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); - error = xfs_itobp(ip->i_mount, tp, ip, &dip, &ibp, 0, 0); + error = xfs_itobp(ip->i_mount, tp, ip, &dip, &ibp, 0, 0, XFS_BUF_LOCK); if (error) return error; @@ -2777,38 +2790,41 @@ xfs_iunpin( } /* - * This is called to wait for the given inode to be unpinned. - * It will sleep until this happens. The caller must have the - * inode locked in at least shared mode so that the buffer cannot - * be subsequently pinned once someone is waiting for it to be - * unpinned. + * This is called to unpin an inode. It can be directed to wait or to return + * immediately without waiting for the inode to be unpinned. The caller must + * have the inode locked in at least shared mode so that the buffer cannot be + * subsequently pinned once someone is waiting for it to be unpinned. */ STATIC void -xfs_iunpin_wait( - xfs_inode_t *ip) +__xfs_iunpin_wait( + xfs_inode_t *ip, + int wait) { - xfs_inode_log_item_t *iip; - xfs_lsn_t lsn; + xfs_inode_log_item_t *iip = ip->i_itemp; ASSERT(ismrlocked(&ip->i_lock, MR_UPDATE | MR_ACCESS)); - - if (atomic_read(&ip->i_pincount) == 0) { + if (atomic_read(&ip->i_pincount) == 0) return; - } - iip = ip->i_itemp; - if (iip && iip->ili_last_lsn) { - lsn = iip->ili_last_lsn; - } else { - lsn = (xfs_lsn_t)0; - } + /* Give the log a push to start the unpinning I/O */ + xfs_log_force(ip->i_mount, (iip && iip->ili_last_lsn) ? + iip->ili_last_lsn : 0, XFS_LOG_FORCE); + if (wait) + wait_event(ip->i_ipin_wait, (atomic_read(&ip->i_pincount) == 0)); +} - /* - * Give the log a push so we don't wait here too long. - */ - xfs_log_force(ip->i_mount, lsn, XFS_LOG_FORCE); +static inline void +xfs_iunpin_wait( + xfs_inode_t *ip) +{ + __xfs_iunpin_wait(ip, 1); +} - wait_event(ip->i_ipin_wait, (atomic_read(&ip->i_pincount) == 0)); +static inline void +xfs_iunpin_nowait( + xfs_inode_t *ip) +{ + __xfs_iunpin_wait(ip, 0); } @@ -3003,6 +3019,7 @@ xfs_iflush( int bufwasdelwri; struct hlist_node *entry; enum { INT_DELWRI = (1 << 0), INT_ASYNC = (1 << 1) }; + int noblock = (flags == XFS_IFLUSH_ASYNC_NOBLOCK); XFS_STATS_INC(xs_iflush_count); @@ -3027,11 +3044,21 @@ xfs_iflush( } /* - * We can't flush the inode until it is unpinned, so - * wait for it. We know noone new can pin it, because - * we are holding the inode lock shared and you need - * to hold it exclusively to pin the inode. + * We can't flush the inode until it is unpinned, so wait for it if we + * are allowed to block. We know noone new can pin it, because we are + * holding the inode lock shared and you need to hold it exclusively to + * pin the inode. + * + * If we are not allowed to block, force the log out asynchronously so + * that when we come back the inode will be unpinned. If other inodes + * in the same cluster are dirty, they will probably write the inode + * out for us if they occur after the log force completes. */ + if (noblock && xfs_ipincount(ip)) { + xfs_iunpin_nowait(ip); + xfs_ifunlock(ip); + return EAGAIN; + } xfs_iunpin_wait(ip); /* @@ -3047,15 +3074,6 @@ xfs_iflush( return XFS_ERROR(EIO); } - /* - * Get the buffer containing the on-disk inode. - */ - error = xfs_itobp(mp, NULL, ip, &dip, &bp, 0, 0); - if (error) { - xfs_ifunlock(ip); - return error; - } - /* * Decide how buffer will be flushed out. This is done before * the call to xfs_iflush_int because this field is zeroed by it. @@ -3072,6 +3090,7 @@ xfs_iflush( case XFS_IFLUSH_DELWRI_ELSE_SYNC: flags = 0; break; + case XFS_IFLUSH_ASYNC_NOBLOCK: case XFS_IFLUSH_ASYNC: case XFS_IFLUSH_DELWRI_ELSE_ASYNC: flags = INT_ASYNC; @@ -3091,6 +3110,7 @@ xfs_iflush( case XFS_IFLUSH_DELWRI: flags = INT_DELWRI; break; + case XFS_IFLUSH_ASYNC_NOBLOCK: case XFS_IFLUSH_ASYNC: flags = INT_ASYNC; break; @@ -3104,6 +3124,16 @@ xfs_iflush( } } + /* + * Get the buffer containing the on-disk inode. + */ + error = xfs_itobp(mp, NULL, ip, &dip, &bp, 0, 0, + noblock ? XFS_BUF_TRYLOCK : XFS_BUF_LOCK); + if (error || !bp) { + xfs_ifunlock(ip); + return error; + } + /* * First flush out the inode that xfs_iflush was called with. */ @@ -3112,6 +3142,13 @@ xfs_iflush( goto corrupt_out; } + /* + * If the buffer is pinned then push on the log now so we won't + * get stuck waiting in the write for too long. + */ + if (XFS_BUF_ISPINNED(bp)) + xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE); + /* * inode clustering: * see if other inodes can be gathered into this write @@ -3181,14 +3218,6 @@ xfs_iflush( XFS_STATS_ADD(xs_icluster_flushinode, clcount); } - /* - * If the buffer is pinned then push on the log so we won't - * get stuck waiting in the write for too long. - */ - if (XFS_BUF_ISPINNED(bp)){ - xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE); - } - if (flags & INT_DELWRI) { xfs_bdwrite(mp, bp); } else if (flags & INT_ASYNC) { diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h index eaa01895ff9..c3bfffca921 100644 --- a/fs/xfs/xfs_inode.h +++ b/fs/xfs/xfs_inode.h @@ -457,6 +457,7 @@ xfs_iflags_test_and_clear(xfs_inode_t *ip, unsigned short flags) #define XFS_IFLUSH_SYNC 3 #define XFS_IFLUSH_ASYNC 4 #define XFS_IFLUSH_DELWRI 5 +#define XFS_IFLUSH_ASYNC_NOBLOCK 6 /* * Flags for xfs_itruncate_start(). @@ -511,7 +512,7 @@ int xfs_finish_reclaim_all(struct xfs_mount *, int); */ int xfs_itobp(struct xfs_mount *, struct xfs_trans *, xfs_inode_t *, struct xfs_dinode **, struct xfs_buf **, - xfs_daddr_t, uint); + xfs_daddr_t, uint, uint); int xfs_iread(struct xfs_mount *, struct xfs_trans *, xfs_ino_t, xfs_inode_t **, xfs_daddr_t, uint); int xfs_iread_extents(struct xfs_trans *, xfs_inode_t *, int); diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c index f615e04364f..45d8776408e 100644 --- a/fs/xfs/xfs_itable.c +++ b/fs/xfs/xfs_itable.c @@ -614,7 +614,8 @@ xfs_bulkstat( xfs_buf_relse(bp); error = xfs_itobp(mp, NULL, ip, &dip, &bp, bno, - XFS_IMAP_BULKSTAT); + XFS_IMAP_BULKSTAT, + XFS_BUF_LOCK); if (!error) clustidx = ip->i_boffset / mp->m_sb.sb_inodesize; kmem_zone_free(xfs_inode_zone, ip); diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index b2b70eba282..cd24711ae27 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -3214,7 +3214,8 @@ xlog_recover_process_iunlinks( * next inode in the bucket. */ error = xfs_itobp(mp, NULL, ip, &dip, - &ibp, 0, 0); + &ibp, 0, 0, + XFS_BUF_LOCK); ASSERT(error || (dip != NULL)); } diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c index 60b6b898022..4e5c010f504 100644 --- a/fs/xfs/xfs_trans_buf.c +++ b/fs/xfs/xfs_trans_buf.c @@ -304,7 +304,8 @@ xfs_trans_read_buf( if (tp == NULL) { bp = xfs_buf_read_flags(target, blkno, len, flags | BUF_BUSY); if (!bp) - return XFS_ERROR(ENOMEM); + return (flags & XFS_BUF_TRYLOCK) ? + EAGAIN : XFS_ERROR(ENOMEM); if ((bp != NULL) && (XFS_BUF_GETERROR(bp) != 0)) { xfs_ioerror_alert("xfs_trans_read_buf", mp, diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index 40b95e3a88b..14140f6225b 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -3468,29 +3468,6 @@ xfs_inode_flush( ((iip == NULL) || !(iip->ili_format.ilf_fields & XFS_ILOG_ALL))) return 0; - if (flags & FLUSH_LOG) { - if (iip && iip->ili_last_lsn) { - xlog_t *log = mp->m_log; - xfs_lsn_t sync_lsn; - int log_flags = XFS_LOG_FORCE; - - spin_lock(&log->l_grant_lock); - sync_lsn = log->l_last_sync_lsn; - spin_unlock(&log->l_grant_lock); - - if ((XFS_LSN_CMP(iip->ili_last_lsn, sync_lsn) > 0)) { - if (flags & FLUSH_SYNC) - log_flags |= XFS_LOG_SYNC; - error = xfs_log_force(mp, iip->ili_last_lsn, log_flags); - if (error) - return error; - } - - if (ip->i_update_core == 0) - return 0; - } - } - /* * We make this non-blocking if the inode is contended, * return EAGAIN to indicate to the caller that they @@ -3498,30 +3475,22 @@ xfs_inode_flush( * blocking on inodes inside another operation right * now, they get caught later by xfs_sync. */ - if (flags & FLUSH_INODE) { - int flush_flags; - - if (flags & FLUSH_SYNC) { - xfs_ilock(ip, XFS_ILOCK_SHARED); - xfs_iflock(ip); - } else if (xfs_ilock_nowait(ip, XFS_ILOCK_SHARED)) { - if (xfs_ipincount(ip) || !xfs_iflock_nowait(ip)) { - xfs_iunlock(ip, XFS_ILOCK_SHARED); - return EAGAIN; - } - } else { + if (flags & FLUSH_SYNC) { + xfs_ilock(ip, XFS_ILOCK_SHARED); + xfs_iflock(ip); + } else if (xfs_ilock_nowait(ip, XFS_ILOCK_SHARED)) { + if (xfs_ipincount(ip) || !xfs_iflock_nowait(ip)) { + xfs_iunlock(ip, XFS_ILOCK_SHARED); return EAGAIN; } - - if (flags & FLUSH_SYNC) - flush_flags = XFS_IFLUSH_SYNC; - else - flush_flags = XFS_IFLUSH_ASYNC; - - error = xfs_iflush(ip, flush_flags); - xfs_iunlock(ip, XFS_ILOCK_SHARED); + } else { + return EAGAIN; } + error = xfs_iflush(ip, (flags & FLUSH_SYNC) ? XFS_IFLUSH_SYNC + : XFS_IFLUSH_ASYNC_NOBLOCK); + xfs_iunlock(ip, XFS_ILOCK_SHARED); + return error; } -- cgit v1.2.3 From bad5584332e888ac40ca13584e8c114149ddb01e Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 6 Mar 2008 13:43:49 +1100 Subject: [XFS] Remove the xfs_icluster structure Remove the xfs_icluster structure and replace with a radix tree lookup. We don't need to keep a list of inodes in each cluster around anymore as we can look them up quickly when we need to. The only time we need to do this now is during inode writeback. Factor the inode cluster writeback code out of xfs_iflush and convert it to use radix_tree_gang_lookup() instead of walking a list of inodes built when we first read in the inodes. This remove 3 pointers from each xfs_inode structure and the xfs_icluster structure per inode cluster. Hence we reduce the cache footprint of the xfs_inodes by between 5-10% depending on cluster sparseness. To be truly efficient we need a radix_tree_gang_lookup_range() call to stop searching once we are past the end of the cluster instead of trying to find a full cluster's worth of inodes. Before (ia64): $ cat /sys/slab/xfs_inode/object_size 536 After: $ cat /sys/slab/xfs_inode/object_size 512 SGI-PV: 977460 SGI-Modid: xfs-linux-melb:xfs-kern:30502a Signed-off-by: David Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_iget.c | 49 +--------- fs/xfs/xfs_inode.c | 268 ++++++++++++++++++++++++++++++---------------------- fs/xfs/xfs_inode.h | 16 ---- fs/xfs/xfs_vfsops.c | 5 - 4 files changed, 156 insertions(+), 182 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_iget.c b/fs/xfs/xfs_iget.c index 8e09b71f410..e657c512846 100644 --- a/fs/xfs/xfs_iget.c +++ b/fs/xfs/xfs_iget.c @@ -78,7 +78,6 @@ xfs_iget_core( xfs_inode_t *ip; xfs_inode_t *iq; int error; - xfs_icluster_t *icl, *new_icl = NULL; unsigned long first_index, mask; xfs_perag_t *pag; xfs_agino_t agino; @@ -229,11 +228,9 @@ finish_inode: } /* - * This is a bit messy - we preallocate everything we _might_ - * need before we pick up the ici lock. That way we don't have to - * juggle locks and go all the way back to the start. + * Preload the radix tree so we can insert safely under the + * write spinlock. */ - new_icl = kmem_zone_alloc(xfs_icluster_zone, KM_SLEEP); if (radix_tree_preload(GFP_KERNEL)) { xfs_idestroy(ip); delay(1); @@ -242,17 +239,6 @@ finish_inode: mask = ~(((XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_inodelog)) - 1); first_index = agino & mask; write_lock(&pag->pag_ici_lock); - - /* - * Find the cluster if it exists - */ - icl = NULL; - if (radix_tree_gang_lookup(&pag->pag_ici_root, (void**)&iq, - first_index, 1)) { - if ((XFS_INO_TO_AGINO(mp, iq->i_ino) & mask) == first_index) - icl = iq->i_cluster; - } - /* * insert the new inode */ @@ -267,30 +253,13 @@ finish_inode: } /* - * These values _must_ be set before releasing ihlock! + * These values _must_ be set before releasing the radix tree lock! */ ip->i_udquot = ip->i_gdquot = NULL; xfs_iflags_set(ip, XFS_INEW); - ASSERT(ip->i_cluster == NULL); - - if (!icl) { - spin_lock_init(&new_icl->icl_lock); - INIT_HLIST_HEAD(&new_icl->icl_inodes); - icl = new_icl; - new_icl = NULL; - } else { - ASSERT(!hlist_empty(&icl->icl_inodes)); - } - spin_lock(&icl->icl_lock); - hlist_add_head(&ip->i_cnode, &icl->icl_inodes); - ip->i_cluster = icl; - spin_unlock(&icl->icl_lock); - write_unlock(&pag->pag_ici_lock); radix_tree_preload_end(); - if (new_icl) - kmem_zone_free(xfs_icluster_zone, new_icl); /* * Link ip to its mount and thread it on the mount's inode list. @@ -528,18 +497,6 @@ xfs_iextract( write_unlock(&pag->pag_ici_lock); xfs_put_perag(mp, pag); - /* - * Remove from cluster list - */ - mp = ip->i_mount; - spin_lock(&ip->i_cluster->icl_lock); - hlist_del(&ip->i_cnode); - spin_unlock(&ip->i_cluster->icl_lock); - - /* was last inode in cluster? */ - if (hlist_empty(&ip->i_cluster->icl_inodes)) - kmem_zone_free(xfs_icluster_zone, ip->i_cluster); - /* * Remove from mount's inode list. */ diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 3c3e9e3c1da..040c0e41729 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -55,7 +55,6 @@ kmem_zone_t *xfs_ifork_zone; kmem_zone_t *xfs_inode_zone; -kmem_zone_t *xfs_icluster_zone; /* * Used in xfs_itruncate(). This is the maximum number of extents @@ -2994,6 +2993,153 @@ xfs_iflush_fork( return 0; } +STATIC int +xfs_iflush_cluster( + xfs_inode_t *ip, + xfs_buf_t *bp) +{ + xfs_mount_t *mp = ip->i_mount; + xfs_perag_t *pag = xfs_get_perag(mp, ip->i_ino); + unsigned long first_index, mask; + int ilist_size; + xfs_inode_t **ilist; + xfs_inode_t *iq; + xfs_inode_log_item_t *iip; + int nr_found; + int clcount = 0; + int bufwasdelwri; + int i; + + ASSERT(pag->pagi_inodeok); + ASSERT(pag->pag_ici_init); + + ilist_size = XFS_INODE_CLUSTER_SIZE(mp) * sizeof(xfs_inode_t *); + ilist = kmem_alloc(ilist_size, KM_MAYFAIL); + if (!ilist) + return 0; + + mask = ~(((XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_inodelog)) - 1); + first_index = XFS_INO_TO_AGINO(mp, ip->i_ino) & mask; + read_lock(&pag->pag_ici_lock); + /* really need a gang lookup range call here */ + nr_found = radix_tree_gang_lookup(&pag->pag_ici_root, (void**)ilist, + first_index, + XFS_INODE_CLUSTER_SIZE(mp)); + if (nr_found == 0) + goto out_free; + + for (i = 0; i < nr_found; i++) { + iq = ilist[i]; + if (iq == ip) + continue; + /* if the inode lies outside this cluster, we're done. */ + if ((XFS_INO_TO_AGINO(mp, iq->i_ino) & mask) != first_index) + break; + /* + * Do an un-protected check to see if the inode is dirty and + * is a candidate for flushing. These checks will be repeated + * later after the appropriate locks are acquired. + */ + iip = iq->i_itemp; + if ((iq->i_update_core == 0) && + ((iip == NULL) || + !(iip->ili_format.ilf_fields & XFS_ILOG_ALL)) && + xfs_ipincount(iq) == 0) { + continue; + } + + /* + * Try to get locks. If any are unavailable or it is pinned, + * then this inode cannot be flushed and is skipped. + */ + + if (!xfs_ilock_nowait(iq, XFS_ILOCK_SHARED)) + continue; + if (!xfs_iflock_nowait(iq)) { + xfs_iunlock(iq, XFS_ILOCK_SHARED); + continue; + } + if (xfs_ipincount(iq)) { + xfs_ifunlock(iq); + xfs_iunlock(iq, XFS_ILOCK_SHARED); + continue; + } + + /* + * arriving here means that this inode can be flushed. First + * re-check that it's dirty before flushing. + */ + iip = iq->i_itemp; + if ((iq->i_update_core != 0) || ((iip != NULL) && + (iip->ili_format.ilf_fields & XFS_ILOG_ALL))) { + int error; + error = xfs_iflush_int(iq, bp); + if (error) { + xfs_iunlock(iq, XFS_ILOCK_SHARED); + goto cluster_corrupt_out; + } + clcount++; + } else { + xfs_ifunlock(iq); + } + xfs_iunlock(iq, XFS_ILOCK_SHARED); + } + + if (clcount) { + XFS_STATS_INC(xs_icluster_flushcnt); + XFS_STATS_ADD(xs_icluster_flushinode, clcount); + } + +out_free: + read_unlock(&pag->pag_ici_lock); + kmem_free(ilist, ilist_size); + return 0; + + +cluster_corrupt_out: + /* + * Corruption detected in the clustering loop. Invalidate the + * inode buffer and shut down the filesystem. + */ + read_unlock(&pag->pag_ici_lock); + /* + * Clean up the buffer. If it was B_DELWRI, just release it -- + * brelse can handle it with no problems. If not, shut down the + * filesystem before releasing the buffer. + */ + bufwasdelwri = XFS_BUF_ISDELAYWRITE(bp); + if (bufwasdelwri) + xfs_buf_relse(bp); + + xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE); + + if (!bufwasdelwri) { + /* + * Just like incore_relse: if we have b_iodone functions, + * mark the buffer as an error and call them. Otherwise + * mark it as stale and brelse. + */ + if (XFS_BUF_IODONE_FUNC(bp)) { + XFS_BUF_CLR_BDSTRAT_FUNC(bp); + XFS_BUF_UNDONE(bp); + XFS_BUF_STALE(bp); + XFS_BUF_SHUT(bp); + XFS_BUF_ERROR(bp,EIO); + xfs_biodone(bp); + } else { + XFS_BUF_STALE(bp); + xfs_buf_relse(bp); + } + } + + /* + * Unlocks the flush lock + */ + xfs_iflush_abort(iq); + kmem_free(ilist, ilist_size); + return XFS_ERROR(EFSCORRUPTED); +} + /* * xfs_iflush() will write a modified inode's changes out to the * inode's on disk home. The caller must have the inode lock held @@ -3013,13 +3159,8 @@ xfs_iflush( xfs_dinode_t *dip; xfs_mount_t *mp; int error; - /* REFERENCED */ - xfs_inode_t *iq; - int clcount; /* count of inodes clustered */ - int bufwasdelwri; - struct hlist_node *entry; - enum { INT_DELWRI = (1 << 0), INT_ASYNC = (1 << 1) }; int noblock = (flags == XFS_IFLUSH_ASYNC_NOBLOCK); + enum { INT_DELWRI = (1 << 0), INT_ASYNC = (1 << 1) }; XFS_STATS_INC(xs_iflush_count); @@ -3138,9 +3279,8 @@ xfs_iflush( * First flush out the inode that xfs_iflush was called with. */ error = xfs_iflush_int(ip, bp); - if (error) { + if (error) goto corrupt_out; - } /* * If the buffer is pinned then push on the log now so we won't @@ -3153,70 +3293,9 @@ xfs_iflush( * inode clustering: * see if other inodes can be gathered into this write */ - spin_lock(&ip->i_cluster->icl_lock); - ip->i_cluster->icl_buf = bp; - - clcount = 0; - hlist_for_each_entry(iq, entry, &ip->i_cluster->icl_inodes, i_cnode) { - if (iq == ip) - continue; - - /* - * Do an un-protected check to see if the inode is dirty and - * is a candidate for flushing. These checks will be repeated - * later after the appropriate locks are acquired. - */ - iip = iq->i_itemp; - if ((iq->i_update_core == 0) && - ((iip == NULL) || - !(iip->ili_format.ilf_fields & XFS_ILOG_ALL)) && - xfs_ipincount(iq) == 0) { - continue; - } - - /* - * Try to get locks. If any are unavailable, - * then this inode cannot be flushed and is skipped. - */ - - /* get inode locks (just i_lock) */ - if (xfs_ilock_nowait(iq, XFS_ILOCK_SHARED)) { - /* get inode flush lock */ - if (xfs_iflock_nowait(iq)) { - /* check if pinned */ - if (xfs_ipincount(iq) == 0) { - /* arriving here means that - * this inode can be flushed. - * first re-check that it's - * dirty - */ - iip = iq->i_itemp; - if ((iq->i_update_core != 0)|| - ((iip != NULL) && - (iip->ili_format.ilf_fields & XFS_ILOG_ALL))) { - clcount++; - error = xfs_iflush_int(iq, bp); - if (error) { - xfs_iunlock(iq, - XFS_ILOCK_SHARED); - goto cluster_corrupt_out; - } - } else { - xfs_ifunlock(iq); - } - } else { - xfs_ifunlock(iq); - } - } - xfs_iunlock(iq, XFS_ILOCK_SHARED); - } - } - spin_unlock(&ip->i_cluster->icl_lock); - - if (clcount) { - XFS_STATS_INC(xs_icluster_flushcnt); - XFS_STATS_ADD(xs_icluster_flushinode, clcount); - } + error = xfs_iflush_cluster(ip, bp); + if (error) + goto cluster_corrupt_out; if (flags & INT_DELWRI) { xfs_bdwrite(mp, bp); @@ -3230,52 +3309,11 @@ xfs_iflush( corrupt_out: xfs_buf_relse(bp); xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE); - xfs_iflush_abort(ip); - /* - * Unlocks the flush lock - */ - return XFS_ERROR(EFSCORRUPTED); - cluster_corrupt_out: - /* Corruption detected in the clustering loop. Invalidate the - * inode buffer and shut down the filesystem. - */ - spin_unlock(&ip->i_cluster->icl_lock); - - /* - * Clean up the buffer. If it was B_DELWRI, just release it -- - * brelse can handle it with no problems. If not, shut down the - * filesystem before releasing the buffer. - */ - if ((bufwasdelwri= XFS_BUF_ISDELAYWRITE(bp))) { - xfs_buf_relse(bp); - } - - xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE); - - if(!bufwasdelwri) { - /* - * Just like incore_relse: if we have b_iodone functions, - * mark the buffer as an error and call them. Otherwise - * mark it as stale and brelse. - */ - if (XFS_BUF_IODONE_FUNC(bp)) { - XFS_BUF_CLR_BDSTRAT_FUNC(bp); - XFS_BUF_UNDONE(bp); - XFS_BUF_STALE(bp); - XFS_BUF_SHUT(bp); - XFS_BUF_ERROR(bp,EIO); - xfs_biodone(bp); - } else { - XFS_BUF_STALE(bp); - xfs_buf_relse(bp); - } - } - - xfs_iflush_abort(iq); /* * Unlocks the flush lock */ + xfs_iflush_abort(ip); return XFS_ERROR(EFSCORRUPTED); } diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h index c3bfffca921..93c37697a72 100644 --- a/fs/xfs/xfs_inode.h +++ b/fs/xfs/xfs_inode.h @@ -132,19 +132,6 @@ typedef struct dm_attrs_s { __uint16_t da_pad; /* DMIG extra padding */ } dm_attrs_t; -/* - * This is the xfs inode cluster structure. This structure is used by - * xfs_iflush to find inodes that share a cluster and can be flushed to disk at - * the same time. - */ -typedef struct xfs_icluster { - struct hlist_head icl_inodes; /* list of inodes on cluster */ - xfs_daddr_t icl_blkno; /* starting block number of - * the cluster */ - struct xfs_buf *icl_buf; /* the inode buffer */ - spinlock_t icl_lock; /* inode list lock */ -} xfs_icluster_t; - /* * This is the xfs in-core inode structure. * Most of the on-disk inode is embedded in the i_d field. @@ -248,8 +235,6 @@ typedef struct xfs_inode { unsigned int i_delayed_blks; /* count of delay alloc blks */ xfs_icdinode_t i_d; /* most of ondisk inode */ - xfs_icluster_t *i_cluster; /* cluster list header */ - struct hlist_node i_cnode; /* cluster link node */ xfs_fsize_t i_size; /* in-memory size */ xfs_fsize_t i_new_size; /* size when write completes */ @@ -594,7 +579,6 @@ void xfs_inobp_check(struct xfs_mount *, struct xfs_buf *); #define xfs_inobp_check(mp, bp) #endif /* DEBUG */ -extern struct kmem_zone *xfs_icluster_zone; extern struct kmem_zone *xfs_ifork_zone; extern struct kmem_zone *xfs_inode_zone; extern struct kmem_zone *xfs_ili_zone; diff --git a/fs/xfs/xfs_vfsops.c b/fs/xfs/xfs_vfsops.c index 79bdfb3d608..3ec27bf8531 100644 --- a/fs/xfs/xfs_vfsops.c +++ b/fs/xfs/xfs_vfsops.c @@ -112,9 +112,6 @@ xfs_init(void) xfs_ili_zone = kmem_zone_init_flags(sizeof(xfs_inode_log_item_t), "xfs_ili", KM_ZONE_SPREAD, NULL); - xfs_icluster_zone = - kmem_zone_init_flags(sizeof(xfs_icluster_t), "xfs_icluster", - KM_ZONE_SPREAD, NULL); /* * Allocate global trace buffers. @@ -152,7 +149,6 @@ xfs_cleanup(void) extern kmem_zone_t *xfs_inode_zone; extern kmem_zone_t *xfs_efd_zone; extern kmem_zone_t *xfs_efi_zone; - extern kmem_zone_t *xfs_icluster_zone; xfs_cleanup_procfs(); xfs_sysctl_unregister(); @@ -187,7 +183,6 @@ xfs_cleanup(void) kmem_zone_destroy(xfs_efi_zone); kmem_zone_destroy(xfs_ifork_zone); kmem_zone_destroy(xfs_ili_zone); - kmem_zone_destroy(xfs_icluster_zone); } /* -- cgit v1.2.3 From 3354040897f828644be6ca5783588e9f64a53b8e Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 6 Mar 2008 13:43:59 +1100 Subject: [XFS] Use xfs_inode_clean() in more places Remove open coded checks for the whether the inode is clean and replace them with an inlined function. SGI-PV: 977461 SGI-Modid: xfs-linux-melb:xfs-kern:30503a Signed-off-by: David Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_inode.c | 27 +++++---------------------- fs/xfs/xfs_inode_item.h | 8 ++++++++ fs/xfs/xfs_vnodeops.c | 4 +--- 3 files changed, 14 insertions(+), 25 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 040c0e41729..d7514f8317d 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -2118,13 +2118,6 @@ xfs_iunlink_remove( return 0; } -STATIC_INLINE int xfs_inode_clean(xfs_inode_t *ip) -{ - return (((ip->i_itemp == NULL) || - !(ip->i_itemp->ili_format.ilf_fields & XFS_ILOG_ALL)) && - (ip->i_update_core == 0)); -} - STATIC void xfs_ifree_cluster( xfs_inode_t *free_ip, @@ -3004,7 +2997,6 @@ xfs_iflush_cluster( int ilist_size; xfs_inode_t **ilist; xfs_inode_t *iq; - xfs_inode_log_item_t *iip; int nr_found; int clcount = 0; int bufwasdelwri; @@ -3040,13 +3032,8 @@ xfs_iflush_cluster( * is a candidate for flushing. These checks will be repeated * later after the appropriate locks are acquired. */ - iip = iq->i_itemp; - if ((iq->i_update_core == 0) && - ((iip == NULL) || - !(iip->ili_format.ilf_fields & XFS_ILOG_ALL)) && - xfs_ipincount(iq) == 0) { + if (xfs_inode_clean(iq) && xfs_ipincount(iq) == 0) continue; - } /* * Try to get locks. If any are unavailable or it is pinned, @@ -3069,10 +3056,8 @@ xfs_iflush_cluster( * arriving here means that this inode can be flushed. First * re-check that it's dirty before flushing. */ - iip = iq->i_itemp; - if ((iq->i_update_core != 0) || ((iip != NULL) && - (iip->ili_format.ilf_fields & XFS_ILOG_ALL))) { - int error; + if (!xfs_inode_clean(iq)) { + int error; error = xfs_iflush_int(iq, bp); if (error) { xfs_iunlock(iq, XFS_ILOCK_SHARED); @@ -3176,8 +3161,7 @@ xfs_iflush( * If the inode isn't dirty, then just release the inode * flush lock and do nothing. */ - if ((ip->i_update_core == 0) && - ((iip == NULL) || !(iip->ili_format.ilf_fields & XFS_ILOG_ALL))) { + if (xfs_inode_clean(ip)) { ASSERT((iip != NULL) ? !(iip->ili_item.li_flags & XFS_LI_IN_AIL) : 1); xfs_ifunlock(ip); @@ -3343,8 +3327,7 @@ xfs_iflush_int( * If the inode isn't dirty, then just release the inode * flush lock and do nothing. */ - if ((ip->i_update_core == 0) && - ((iip == NULL) || !(iip->ili_format.ilf_fields & XFS_ILOG_ALL))) { + if (xfs_inode_clean(ip)) { xfs_ifunlock(ip); return 0; } diff --git a/fs/xfs/xfs_inode_item.h b/fs/xfs/xfs_inode_item.h index bfe92ea1795..40513077ab3 100644 --- a/fs/xfs/xfs_inode_item.h +++ b/fs/xfs/xfs_inode_item.h @@ -168,6 +168,14 @@ static inline int xfs_ilog_fext(int w) return (w == XFS_DATA_FORK ? XFS_ILOG_DEXT : XFS_ILOG_AEXT); } +static inline int xfs_inode_clean(xfs_inode_t *ip) +{ + return (!ip->i_itemp || + !(ip->i_itemp->ili_format.ilf_fields & XFS_ILOG_ALL)) && + !ip->i_update_core; +} + + #ifdef __KERNEL__ extern void xfs_inode_item_init(struct xfs_inode *, struct xfs_mount *); diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index 14140f6225b..5390d124ad3 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -3454,7 +3454,6 @@ xfs_inode_flush( int flags) { xfs_mount_t *mp = ip->i_mount; - xfs_inode_log_item_t *iip = ip->i_itemp; int error = 0; if (XFS_FORCED_SHUTDOWN(mp)) @@ -3464,8 +3463,7 @@ xfs_inode_flush( * Bypass inodes which have already been cleaned by * the inode flush clustering code inside xfs_iflush */ - if ((ip->i_update_core == 0) && - ((iip == NULL) || !(iip->ili_format.ilf_fields & XFS_ILOG_ALL))) + if (xfs_inode_clean(ip)) return 0; /* -- cgit v1.2.3 From b589334c7a1fff85d2f009d5db4c34fad48925e9 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 6 Mar 2008 13:44:06 +1100 Subject: [XFS] Prevent AIL lock contention during transaction completion When hundreds of processors attempt to commit transactions at the same time, they can contend on the AIL lock when updating the tail LSN held in the in-core log structure. At the moment, the tail LSN is only needed when actually writing out an iclog, so it really does not need to be updated on every single transaction completion - only those that result in switching iclogs and flushing them to disk. The result is that we reduce the number of times we need to grab the AIL lock and the log grant lock by up to two orders of magnitude on large processor count machines. The problem has previously been hidden by AIL lock contention walking the AIL list which was recently solved and uncovered this issue. SGI-PV: 975671 SGI-Modid: xfs-linux-melb:xfs-kern:30504a Signed-off-by: David Chinner Signed-off-by: Tim Shimmin Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_log.c | 15 ++++++--------- 1 file changed, 6 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index 31f2b04f2c9..2e35077ff6b 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -2813,15 +2813,13 @@ xlog_state_put_ticket(xlog_t *log, * */ STATIC int -xlog_state_release_iclog(xlog_t *log, - xlog_in_core_t *iclog) +xlog_state_release_iclog( + xlog_t *log, + xlog_in_core_t *iclog) { int sync = 0; /* do we sync? */ - xlog_assign_tail_lsn(log->l_mp); - spin_lock(&log->l_icloglock); - if (iclog->ic_state & XLOG_STATE_IOERROR) { spin_unlock(&log->l_icloglock); return XFS_ERROR(EIO); @@ -2833,13 +2831,14 @@ xlog_state_release_iclog(xlog_t *log, if (--iclog->ic_refcnt == 0 && iclog->ic_state == XLOG_STATE_WANT_SYNC) { + /* update tail before writing to iclog */ + xlog_assign_tail_lsn(log->l_mp); sync++; iclog->ic_state = XLOG_STATE_SYNCING; iclog->ic_header.h_tail_lsn = cpu_to_be64(log->l_tail_lsn); xlog_verify_tail_lsn(log, iclog, log->l_tail_lsn); /* cycle incremented when incrementing curr_block */ } - spin_unlock(&log->l_icloglock); /* @@ -2849,11 +2848,9 @@ xlog_state_release_iclog(xlog_t *log, * this iclog has consistent data, so we ignore IOERROR * flags after this point. */ - if (sync) { + if (sync) return xlog_sync(log, iclog); - } return 0; - } /* xlog_state_release_iclog */ -- cgit v1.2.3 From 155cc6b784a959ed456fe46dca522e1d28b3b718 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 6 Mar 2008 13:44:14 +1100 Subject: [XFS] Use atomics for iclog reference counting MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Now that we update the log tail LSN less frequently on transaction completion, we pass the contention straight to the global log state lock (l_iclog_lock) during transaction completion. We currently have to take this lock to decrement the iclog reference count. there is a reference count on each iclog, so we need to take þhe global lock for all refcount changes. When large numbers of processes are all doing small trnasctions, the iclog reference counts will be quite high, and the state change that absolutely requires the l_iclog_lock is the except rather than the norm. Change the reference counting on the iclogs to use atomic_inc/dec so that we can use atomic_dec_and_lock during transaction completion and avoid the need for grabbing the l_iclog_lock for every reference count decrement except the one that matters - the last. SGI-PV: 975671 SGI-Modid: xfs-linux-melb:xfs-kern:30505a Signed-off-by: David Chinner Signed-off-by: Tim Shimmin Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_log.c | 36 ++++++++++++++++++++---------------- fs/xfs/xfs_log_priv.h | 2 +- 2 files changed, 21 insertions(+), 17 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index 2e35077ff6b..1fa98093389 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -675,7 +675,7 @@ xfs_log_unmount_write(xfs_mount_t *mp) spin_lock(&log->l_icloglock); iclog = log->l_iclog; - iclog->ic_refcnt++; + atomic_inc(&iclog->ic_refcnt); spin_unlock(&log->l_icloglock); xlog_state_want_sync(log, iclog); (void) xlog_state_release_iclog(log, iclog); @@ -713,7 +713,7 @@ xfs_log_unmount_write(xfs_mount_t *mp) */ spin_lock(&log->l_icloglock); iclog = log->l_iclog; - iclog->ic_refcnt++; + atomic_inc(&iclog->ic_refcnt); spin_unlock(&log->l_icloglock); xlog_state_want_sync(log, iclog); @@ -1405,7 +1405,7 @@ xlog_sync(xlog_t *log, int v2 = xfs_sb_version_haslogv2(&log->l_mp->m_sb); XFS_STATS_INC(xs_log_writes); - ASSERT(iclog->ic_refcnt == 0); + ASSERT(atomic_read(&iclog->ic_refcnt) == 0); /* Add for LR header */ count_init = log->l_iclog_hsize + iclog->ic_offset; @@ -2309,7 +2309,7 @@ xlog_state_done_syncing( ASSERT(iclog->ic_state == XLOG_STATE_SYNCING || iclog->ic_state == XLOG_STATE_IOERROR); - ASSERT(iclog->ic_refcnt == 0); + ASSERT(atomic_read(&iclog->ic_refcnt) == 0); ASSERT(iclog->ic_bwritecnt == 1 || iclog->ic_bwritecnt == 2); @@ -2391,7 +2391,7 @@ restart: ASSERT(iclog->ic_state == XLOG_STATE_ACTIVE); head = &iclog->ic_header; - iclog->ic_refcnt++; /* prevents sync */ + atomic_inc(&iclog->ic_refcnt); /* prevents sync */ log_offset = iclog->ic_offset; /* On the 1st write to an iclog, figure out lsn. This works @@ -2423,12 +2423,12 @@ restart: xlog_state_switch_iclogs(log, iclog, iclog->ic_size); /* If I'm the only one writing to this iclog, sync it to disk */ - if (iclog->ic_refcnt == 1) { + if (atomic_read(&iclog->ic_refcnt) == 1) { spin_unlock(&log->l_icloglock); if ((error = xlog_state_release_iclog(log, iclog))) return error; } else { - iclog->ic_refcnt--; + atomic_dec(&iclog->ic_refcnt); spin_unlock(&log->l_icloglock); } goto restart; @@ -2819,18 +2819,21 @@ xlog_state_release_iclog( { int sync = 0; /* do we sync? */ - spin_lock(&log->l_icloglock); + if (iclog->ic_state & XLOG_STATE_IOERROR) + return XFS_ERROR(EIO); + + ASSERT(atomic_read(&iclog->ic_refcnt) > 0); + if (!atomic_dec_and_lock(&iclog->ic_refcnt, &log->l_icloglock)) + return 0; + if (iclog->ic_state & XLOG_STATE_IOERROR) { spin_unlock(&log->l_icloglock); return XFS_ERROR(EIO); } - - ASSERT(iclog->ic_refcnt > 0); ASSERT(iclog->ic_state == XLOG_STATE_ACTIVE || iclog->ic_state == XLOG_STATE_WANT_SYNC); - if (--iclog->ic_refcnt == 0 && - iclog->ic_state == XLOG_STATE_WANT_SYNC) { + if (iclog->ic_state == XLOG_STATE_WANT_SYNC) { /* update tail before writing to iclog */ xlog_assign_tail_lsn(log->l_mp); sync++; @@ -2950,7 +2953,8 @@ xlog_state_sync_all(xlog_t *log, uint flags, int *log_flushed) * previous iclog and go to sleep. */ if (iclog->ic_state == XLOG_STATE_DIRTY || - (iclog->ic_refcnt == 0 && iclog->ic_offset == 0)) { + (atomic_read(&iclog->ic_refcnt) == 0 + && iclog->ic_offset == 0)) { iclog = iclog->ic_prev; if (iclog->ic_state == XLOG_STATE_ACTIVE || iclog->ic_state == XLOG_STATE_DIRTY) @@ -2958,14 +2962,14 @@ xlog_state_sync_all(xlog_t *log, uint flags, int *log_flushed) else goto maybe_sleep; } else { - if (iclog->ic_refcnt == 0) { + if (atomic_read(&iclog->ic_refcnt) == 0) { /* We are the only one with access to this * iclog. Flush it out now. There should * be a roundoff of zero to show that someone * has already taken care of the roundoff from * the previous sync. */ - iclog->ic_refcnt++; + atomic_inc(&iclog->ic_refcnt); lsn = be64_to_cpu(iclog->ic_header.h_lsn); xlog_state_switch_iclogs(log, iclog, 0); spin_unlock(&log->l_icloglock); @@ -3097,7 +3101,7 @@ try_again: already_slept = 1; goto try_again; } else { - iclog->ic_refcnt++; + atomic_inc(&iclog->ic_refcnt); xlog_state_switch_iclogs(log, iclog, 0); spin_unlock(&log->l_icloglock); if (xlog_state_release_iclog(log, iclog)) diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h index c6244cc733c..01c63db25a1 100644 --- a/fs/xfs/xfs_log_priv.h +++ b/fs/xfs/xfs_log_priv.h @@ -339,7 +339,7 @@ typedef struct xlog_iclog_fields { #endif int ic_size; int ic_offset; - int ic_refcnt; + atomic_t ic_refcnt; int ic_bwritecnt; ushort_t ic_state; char *ic_datap; /* pointer to iclog data */ -- cgit v1.2.3 From db0bb7baa1533db156d8af3ebeda1f0473a0197a Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 6 Mar 2008 13:44:35 +1100 Subject: [XFS] cleanup xfs_vn_mknod - use proper goto based unwinding instead of the current mess of multiple conditionals - rename ip to inode because that's the normal convention for Linux inodes while ip is the convention for xfs_inodes - remove unlikely checks for the default_acl - branches marked unlikely might lead to extreme branch bredictor slowdons if taken and for some workloads a default acl is quite common - properly indent the switch statements - remove xfs_has_fs_struct as nfsd has a fs_struct in any semi-recent kernel SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30529a Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_iops.c | 67 ++++++++++++++++++++------------------------- 1 file changed, 30 insertions(+), 37 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c index cc4abd3daa4..34670118331 100644 --- a/fs/xfs/linux-2.6/xfs_iops.c +++ b/fs/xfs/linux-2.6/xfs_iops.c @@ -241,18 +241,6 @@ xfs_init_security( return error; } -/* - * Determine whether a process has a valid fs_struct (kernel daemons - * like knfsd don't have an fs_struct). - * - * XXX(hch): nfsd is broken, better fix it instead. - */ -STATIC_INLINE int -xfs_has_fs_struct(struct task_struct *task) -{ - return (task->fs != init_task.fs); -} - STATIC void xfs_cleanup_inode( struct inode *dir, @@ -284,7 +272,7 @@ xfs_vn_mknod( int mode, dev_t rdev) { - struct inode *ip; + struct inode *inode; bhv_vnode_t *vp = NULL, *dvp = vn_from_inode(dir); xfs_acl_t *default_acl = NULL; attrexists_t test_default_acl = _ACL_DEFAULT_EXISTS; @@ -297,7 +285,7 @@ xfs_vn_mknod( if (unlikely(!sysv_valid_dev(rdev) || MAJOR(rdev) & ~0x1ff)) return -EINVAL; - if (unlikely(test_default_acl && test_default_acl(dvp))) { + if (test_default_acl && test_default_acl(dvp)) { if (!_ACL_ALLOC(default_acl)) { return -ENOMEM; } @@ -307,11 +295,14 @@ xfs_vn_mknod( } } - if (IS_POSIXACL(dir) && !default_acl && xfs_has_fs_struct(current)) + if (IS_POSIXACL(dir) && !default_acl) mode &= ~current->fs->umask; switch (mode & S_IFMT) { - case S_IFCHR: case S_IFBLK: case S_IFIFO: case S_IFSOCK: + case S_IFCHR: + case S_IFBLK: + case S_IFIFO: + case S_IFSOCK: rdev = sysv_encode_dev(rdev); case S_IFREG: error = xfs_create(XFS_I(dir), dentry, mode, rdev, &vp, NULL); @@ -324,32 +315,34 @@ xfs_vn_mknod( break; } - if (unlikely(!error)) { - error = xfs_init_security(vp, dir); - if (error) - xfs_cleanup_inode(dir, vp, dentry, mode); - } + if (unlikely(error)) + goto out_free_acl; - if (unlikely(default_acl)) { - if (!error) { - error = _ACL_INHERIT(vp, mode, default_acl); - if (!error) - xfs_iflags_set(XFS_I(vp), XFS_IMODIFIED); - else - xfs_cleanup_inode(dir, vp, dentry, mode); - } + error = xfs_init_security(vp, dir); + if (unlikely(error)) + goto out_cleanup_inode; + + if (default_acl) { + error = _ACL_INHERIT(vp, mode, default_acl); + if (unlikely(error)) + goto out_cleanup_inode; + xfs_iflags_set(XFS_I(vp), XFS_IMODIFIED); _ACL_FREE(default_acl); } - if (likely(!error)) { - ASSERT(vp); - ip = vn_to_inode(vp); + inode = vn_to_inode(vp); - if (S_ISDIR(mode)) - xfs_validate_fields(ip); - d_instantiate(dentry, ip); - xfs_validate_fields(dir); - } + if (S_ISDIR(mode)) + xfs_validate_fields(inode); + d_instantiate(dentry, inode); + xfs_validate_fields(dir); + return -error; + + out_cleanup_inode: + xfs_cleanup_inode(dir, vp, dentry, mode); + out_free_acl: + if (default_acl) + _ACL_FREE(default_acl); return -error; } -- cgit v1.2.3 From a8b3acd57e3aaaf73a863a28e0e9f6cca37cd8e3 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 6 Mar 2008 13:44:41 +1100 Subject: [XFS] vnode cleanup in xfs_fs_subr.c Cleanup the unneeded intermediate vnode step in the flushing helpers and go directly from the xfs_inode to the struct address_space. SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30530a Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_fs_subr.c | 36 +++++++++++------------------------- 1 file changed, 11 insertions(+), 25 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_fs_subr.c b/fs/xfs/linux-2.6/xfs_fs_subr.c index ac6d34cc355..1eefe61f0e1 100644 --- a/fs/xfs/linux-2.6/xfs_fs_subr.c +++ b/fs/xfs/linux-2.6/xfs_fs_subr.c @@ -17,18 +17,7 @@ */ #include "xfs.h" #include "xfs_vnodeops.h" - -/* - * The following six includes are needed so that we can include - * xfs_inode.h. What a mess.. - */ #include "xfs_bmap_btree.h" -#include "xfs_inum.h" -#include "xfs_dir2.h" -#include "xfs_dir2_sf.h" -#include "xfs_attr_sf.h" -#include "xfs_dinode.h" - #include "xfs_inode.h" int fs_noerr(void) { return 0; } @@ -42,11 +31,10 @@ xfs_tosspages( xfs_off_t last, int fiopt) { - bhv_vnode_t *vp = XFS_ITOV(ip); - struct inode *inode = vn_to_inode(vp); + struct address_space *mapping = ip->i_vnode->i_mapping; - if (VN_CACHED(vp)) - truncate_inode_pages(inode->i_mapping, first); + if (mapping->nrpages) + truncate_inode_pages(mapping, first); } int @@ -56,15 +44,14 @@ xfs_flushinval_pages( xfs_off_t last, int fiopt) { - bhv_vnode_t *vp = XFS_ITOV(ip); - struct inode *inode = vn_to_inode(vp); + struct address_space *mapping = ip->i_vnode->i_mapping; int ret = 0; - if (VN_CACHED(vp)) { + if (mapping->nrpages) { xfs_iflags_clear(ip, XFS_ITRUNCATED); - ret = filemap_write_and_wait(inode->i_mapping); + ret = filemap_write_and_wait(mapping); if (!ret) - truncate_inode_pages(inode->i_mapping, first); + truncate_inode_pages(mapping, first); } return ret; } @@ -77,17 +64,16 @@ xfs_flush_pages( uint64_t flags, int fiopt) { - bhv_vnode_t *vp = XFS_ITOV(ip); - struct inode *inode = vn_to_inode(vp); + struct address_space *mapping = ip->i_vnode->i_mapping; int ret = 0; int ret2; - if (VN_DIRTY(vp)) { + if (mapping_tagged(mapping, PAGECACHE_TAG_DIRTY)) { xfs_iflags_clear(ip, XFS_ITRUNCATED); - ret = filemap_fdatawrite(inode->i_mapping); + ret = filemap_fdatawrite(mapping); if (flags & XFS_B_ASYNC) return ret; - ret2 = filemap_fdatawait(inode->i_mapping); + ret2 = filemap_fdatawait(mapping); if (!ret) ret = ret2; } -- cgit v1.2.3 From 43973964a386348af0a392266f008ba24170aa30 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 6 Mar 2008 13:44:50 +1100 Subject: [XFS] kill xfs_get_dir_entry Instead of of xfs_get_dir_entry use a macro to get the xfs_inode from the dentry in the callers and grab the reference manually. Only grab the reference once as it's fine to keep it over the dmapi calls. (And even that reference is actually superflous in Linux but I'll leave that for another patch) SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30531a Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_vnode.h | 2 +- fs/xfs/xfs_rename.c | 12 ++++------- fs/xfs/xfs_utils.c | 22 ------------------- fs/xfs/xfs_utils.h | 1 - fs/xfs/xfs_vnodeops.c | 51 ++++++++------------------------------------ 5 files changed, 14 insertions(+), 74 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_vnode.h b/fs/xfs/linux-2.6/xfs_vnode.h index f200e024408..20223182828 100644 --- a/fs/xfs/linux-2.6/xfs_vnode.h +++ b/fs/xfs/linux-2.6/xfs_vnode.h @@ -227,7 +227,7 @@ static inline bhv_vnode_t *vn_grab(bhv_vnode_t *vp) */ #define VNAME(dentry) ((char *) (dentry)->d_name.name) #define VNAMELEN(dentry) ((dentry)->d_name.len) -#define VNAME_TO_VNODE(dentry) (vn_from_inode((dentry)->d_inode)) +#define VNAME_TO_INODE(dentry) (XFS_I((dentry)->d_inode)) /* * Dealing with bad inodes diff --git a/fs/xfs/xfs_rename.c b/fs/xfs/xfs_rename.c index 1c6d40ed681..fd1244cf50a 100644 --- a/fs/xfs/xfs_rename.c +++ b/fs/xfs/xfs_rename.c @@ -93,7 +93,8 @@ xfs_lock_for_rename( xfs_inode_t **i_tab,/* array of inode returned, sorted */ int *num_inodes) /* number of inodes in array */ { - xfs_inode_t *ip1, *ip2, *temp; + xfs_inode_t *ip1 = VNAME_TO_INODE(vname1); + xfs_inode_t *ip2, *temp; xfs_ino_t inum1, inum2; int error; int i, j; @@ -109,16 +110,11 @@ xfs_lock_for_rename( * to see if we still have the right inodes, directories, etc. */ lock_mode = xfs_ilock_map_shared(dp1); - error = xfs_get_dir_entry(vname1, &ip1); - if (error) { - xfs_iunlock_map_shared(dp1, lock_mode); - return error; - } + IHOLD(ip1); + xfs_itrace_ref(ip1); inum1 = ip1->i_ino; - ASSERT(ip1); - xfs_itrace_ref(ip1); /* * Unlock dp1 and lock dp2 if they are different. diff --git a/fs/xfs/xfs_utils.c b/fs/xfs/xfs_utils.c index 18a85e74668..47c45ff4a06 100644 --- a/fs/xfs/xfs_utils.c +++ b/fs/xfs/xfs_utils.c @@ -40,28 +40,6 @@ #include "xfs_itable.h" #include "xfs_utils.h" -/* - * xfs_get_dir_entry is used to get a reference to an inode given - * its parent directory inode and the name of the file. It does - * not lock the child inode, and it unlocks the directory before - * returning. The directory's generation number is returned for - * use by a later call to xfs_lock_dir_and_entry. - */ -int -xfs_get_dir_entry( - bhv_vname_t *dentry, - xfs_inode_t **ipp) -{ - bhv_vnode_t *vp; - - vp = VNAME_TO_VNODE(dentry); - - *ipp = xfs_vtoi(vp); - if (!*ipp) - return XFS_ERROR(ENOENT); - VN_HOLD(vp); - return 0; -} int xfs_dir_lookup_int( diff --git a/fs/xfs/xfs_utils.h b/fs/xfs/xfs_utils.h index f857fcccb72..c4c4a6aa654 100644 --- a/fs/xfs/xfs_utils.h +++ b/fs/xfs/xfs_utils.h @@ -21,7 +21,6 @@ #define IRELE(ip) VN_RELE(XFS_ITOV(ip)) #define IHOLD(ip) VN_HOLD(XFS_ITOV(ip)) -extern int xfs_get_dir_entry (bhv_vname_t *, xfs_inode_t **); extern int xfs_dir_lookup_int (xfs_inode_t *, uint, bhv_vname_t *, xfs_ino_t *, xfs_inode_t **); extern int xfs_truncate_file (xfs_mount_t *, xfs_inode_t *); diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index 5390d124ad3..4765e7c4b75 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -2270,41 +2270,30 @@ xfs_remove( bhv_vnode_t *dir_vp = XFS_ITOV(dp); char *name = VNAME(dentry); xfs_mount_t *mp = dp->i_mount; - xfs_inode_t *ip; + xfs_inode_t *ip = VNAME_TO_INODE(dentry); + int namelen = VNAMELEN(dentry); xfs_trans_t *tp = NULL; int error = 0; xfs_bmap_free_t free_list; xfs_fsblock_t first_block; int cancel_flags; int committed; - int dm_di_mode = 0; int link_zero; uint resblks; - int namelen; xfs_itrace_entry(dp); if (XFS_FORCED_SHUTDOWN(mp)) return XFS_ERROR(EIO); - namelen = VNAMELEN(dentry); - - if (!xfs_get_dir_entry(dentry, &ip)) { - dm_di_mode = ip->i_d.di_mode; - IRELE(ip); - } - if (DM_EVENT_ENABLED(dp, DM_EVENT_REMOVE)) { error = XFS_SEND_NAMESP(mp, DM_EVENT_REMOVE, dir_vp, DM_RIGHT_NULL, NULL, DM_RIGHT_NULL, - name, NULL, dm_di_mode, 0, 0); + name, NULL, ip->i_d.di_mode, 0, 0); if (error) return error; } - /* From this point on, return through std_return */ - ip = NULL; - /* * We need to get a reference to ip before we get our log * reservation. The reason for this is that we cannot call @@ -2317,13 +2306,7 @@ xfs_remove( * when we call xfs_iget. Instead we get an unlocked reference * to the inode before getting our log reservation. */ - error = xfs_get_dir_entry(dentry, &ip); - if (error) { - REMOVE_DEBUG_TRACE(__LINE__); - goto std_return; - } - - dm_di_mode = ip->i_d.di_mode; + IHOLD(ip); xfs_itrace_entry(ip); xfs_itrace_ref(ip); @@ -2459,7 +2442,7 @@ xfs_remove( (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTREMOVE, dir_vp, DM_RIGHT_NULL, NULL, DM_RIGHT_NULL, - name, NULL, dm_di_mode, error, 0); + name, NULL, ip->i_d.di_mode, error, 0); } return error; @@ -2868,14 +2851,13 @@ xfs_rmdir( char *name = VNAME(dentry); int namelen = VNAMELEN(dentry); xfs_mount_t *mp = dp->i_mount; - xfs_inode_t *cdp; /* child directory */ + xfs_inode_t *cdp = VNAME_TO_INODE(dentry); xfs_trans_t *tp; int error; xfs_bmap_free_t free_list; xfs_fsblock_t first_block; int cancel_flags; int committed; - int dm_di_mode = S_IFDIR; int last_cdp_link; uint resblks; @@ -2884,24 +2866,15 @@ xfs_rmdir( if (XFS_FORCED_SHUTDOWN(mp)) return XFS_ERROR(EIO); - if (!xfs_get_dir_entry(dentry, &cdp)) { - dm_di_mode = cdp->i_d.di_mode; - IRELE(cdp); - } - if (DM_EVENT_ENABLED(dp, DM_EVENT_REMOVE)) { error = XFS_SEND_NAMESP(mp, DM_EVENT_REMOVE, dir_vp, DM_RIGHT_NULL, NULL, DM_RIGHT_NULL, - name, NULL, dm_di_mode, 0, 0); + name, NULL, cdp->i_d.di_mode, 0, 0); if (error) return XFS_ERROR(error); } - /* Return through std_return after this point. */ - - cdp = NULL; - /* * We need to get a reference to cdp before we get our log * reservation. The reason for this is that we cannot call @@ -2914,13 +2887,7 @@ xfs_rmdir( * when we call xfs_iget. Instead we get an unlocked reference * to the inode before getting our log reservation. */ - error = xfs_get_dir_entry(dentry, &cdp); - if (error) { - REMOVE_DEBUG_TRACE(__LINE__); - goto std_return; - } - mp = dp->i_mount; - dm_di_mode = cdp->i_d.di_mode; + IHOLD(cdp); /* * Get the dquots for the inodes. @@ -3077,7 +3044,7 @@ xfs_rmdir( (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTREMOVE, dir_vp, DM_RIGHT_NULL, NULL, DM_RIGHT_NULL, - name, NULL, dm_di_mode, + name, NULL, cdp->i_d.di_mode, error, 0); } return error; -- cgit v1.2.3 From 126468b1156211e26d97f74b2f1767acd141005a Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 6 Mar 2008 13:44:57 +1100 Subject: [XFS] kill xfs_rwlock/xfs_rwunlock We can just use xfs_ilock/xfs_iunlock instead and get rid of the ugly bhv_vrwlock_t. SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30533a Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_aops.c | 4 ++-- fs/xfs/linux-2.6/xfs_lrw.c | 31 +++++++++++++------------------ fs/xfs/linux-2.6/xfs_vnode.h | 12 ------------ fs/xfs/xfs_mount.h | 2 +- fs/xfs/xfs_vnodeops.c | 41 ----------------------------------------- fs/xfs/xfs_vnodeops.h | 2 -- 6 files changed, 16 insertions(+), 76 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_aops.c b/fs/xfs/linux-2.6/xfs_aops.c index e0519529c26..169e6c06279 100644 --- a/fs/xfs/linux-2.6/xfs_aops.c +++ b/fs/xfs/linux-2.6/xfs_aops.c @@ -1532,9 +1532,9 @@ xfs_vm_bmap( struct xfs_inode *ip = XFS_I(inode); xfs_itrace_entry(XFS_I(inode)); - xfs_rwlock(ip, VRWLOCK_READ); + xfs_ilock(ip, XFS_IOLOCK_SHARED); xfs_flush_pages(ip, (xfs_off_t)0, -1, 0, FI_REMAPF); - xfs_rwunlock(ip, VRWLOCK_READ); + xfs_iunlock(ip, XFS_IOLOCK_SHARED); return generic_block_bmap(mapping, block, xfs_get_blocks); } diff --git a/fs/xfs/linux-2.6/xfs_lrw.c b/fs/xfs/linux-2.6/xfs_lrw.c index 16635338849..3c20007ab48 100644 --- a/fs/xfs/linux-2.6/xfs_lrw.c +++ b/fs/xfs/linux-2.6/xfs_lrw.c @@ -228,11 +228,11 @@ xfs_read( xfs_ilock(ip, XFS_IOLOCK_SHARED); if (DM_EVENT_ENABLED(ip, DM_EVENT_READ) && !(ioflags & IO_INVIS)) { - bhv_vrwlock_t locktype = VRWLOCK_READ; int dmflags = FILP_DELAY_FLAG(file) | DM_SEM_FLAG_RD(ioflags); + int iolock = XFS_IOLOCK_SHARED; ret = -XFS_SEND_DATA(mp, DM_EVENT_READ, vp, *offset, size, - dmflags, &locktype); + dmflags, &iolock); if (ret) { xfs_iunlock(ip, XFS_IOLOCK_SHARED); if (unlikely(ioflags & IO_ISDIRECT)) @@ -287,11 +287,11 @@ xfs_splice_read( xfs_ilock(ip, XFS_IOLOCK_SHARED); if (DM_EVENT_ENABLED(ip, DM_EVENT_READ) && !(ioflags & IO_INVIS)) { - bhv_vrwlock_t locktype = VRWLOCK_READ; + int iolock = XFS_IOLOCK_SHARED; int error; error = XFS_SEND_DATA(mp, DM_EVENT_READ, vp, *ppos, count, - FILP_DELAY_FLAG(infilp), &locktype); + FILP_DELAY_FLAG(infilp), &iolock); if (error) { xfs_iunlock(ip, XFS_IOLOCK_SHARED); return -error; @@ -330,11 +330,11 @@ xfs_splice_write( xfs_ilock(ip, XFS_IOLOCK_EXCL); if (DM_EVENT_ENABLED(ip, DM_EVENT_WRITE) && !(ioflags & IO_INVIS)) { - bhv_vrwlock_t locktype = VRWLOCK_WRITE; + int iolock = XFS_IOLOCK_EXCL; int error; error = XFS_SEND_DATA(mp, DM_EVENT_WRITE, vp, *ppos, count, - FILP_DELAY_FLAG(outfilp), &locktype); + FILP_DELAY_FLAG(outfilp), &iolock); if (error) { xfs_iunlock(ip, XFS_IOLOCK_EXCL); return -error; @@ -580,7 +580,6 @@ xfs_write( xfs_fsize_t isize, new_size; int iolock; int eventsent = 0; - bhv_vrwlock_t locktype; size_t ocount = 0, count; loff_t pos; int need_i_mutex; @@ -607,11 +606,9 @@ xfs_write( relock: if (ioflags & IO_ISDIRECT) { iolock = XFS_IOLOCK_SHARED; - locktype = VRWLOCK_WRITE_DIRECT; need_i_mutex = 0; } else { iolock = XFS_IOLOCK_EXCL; - locktype = VRWLOCK_WRITE; need_i_mutex = 1; mutex_lock(&inode->i_mutex); } @@ -635,8 +632,7 @@ start: xfs_iunlock(xip, XFS_ILOCK_EXCL); error = XFS_SEND_DATA(xip->i_mount, DM_EVENT_WRITE, vp, - pos, count, - dmflags, &locktype); + pos, count, dmflags, &iolock); if (error) { goto out_unlock_internal; } @@ -667,7 +663,6 @@ start: if (!need_i_mutex && (VN_CACHED(vp) || pos > xip->i_size)) { xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock); iolock = XFS_IOLOCK_EXCL; - locktype = VRWLOCK_WRITE; need_i_mutex = 1; mutex_lock(&inode->i_mutex); xfs_ilock(xip, XFS_ILOCK_EXCL|iolock); @@ -744,7 +739,6 @@ retry: mutex_unlock(&inode->i_mutex); iolock = XFS_IOLOCK_SHARED; - locktype = VRWLOCK_WRITE_DIRECT; need_i_mutex = 0; } @@ -781,7 +775,7 @@ retry: if (ret == -ENOSPC && DM_EVENT_ENABLED(xip, DM_EVENT_NOSPACE) && !(ioflags & IO_INVIS)) { - xfs_rwunlock(xip, locktype); + xfs_iunlock(xip, iolock); if (need_i_mutex) mutex_unlock(&inode->i_mutex); error = XFS_SEND_NAMESP(xip->i_mount, DM_EVENT_NOSPACE, vp, @@ -789,7 +783,7 @@ retry: 0, 0, 0); /* Delay flag intentionally unused */ if (need_i_mutex) mutex_lock(&inode->i_mutex); - xfs_rwlock(xip, locktype); + xfs_ilock(xip, iolock); if (error) goto out_unlock_internal; pos = xip->i_size; @@ -817,7 +811,8 @@ retry: /* Handle various SYNC-type writes */ if ((file->f_flags & O_SYNC) || IS_SYNC(inode)) { int error2; - xfs_rwunlock(xip, locktype); + + xfs_iunlock(xip, iolock); if (need_i_mutex) mutex_unlock(&inode->i_mutex); error2 = sync_page_range(inode, mapping, pos, ret); @@ -825,7 +820,7 @@ retry: error = error2; if (need_i_mutex) mutex_lock(&inode->i_mutex); - xfs_rwlock(xip, locktype); + xfs_ilock(xip, iolock); error2 = xfs_write_sync_logforce(mp, xip); if (!error) error = error2; @@ -846,7 +841,7 @@ retry: xip->i_d.di_size = xip->i_size; xfs_iunlock(xip, XFS_ILOCK_EXCL); } - xfs_rwunlock(xip, locktype); + xfs_iunlock(xip, iolock); out_unlock_mutex: if (need_i_mutex) mutex_unlock(&inode->i_mutex); diff --git a/fs/xfs/linux-2.6/xfs_vnode.h b/fs/xfs/linux-2.6/xfs_vnode.h index 20223182828..4ed5914adef 100644 --- a/fs/xfs/linux-2.6/xfs_vnode.h +++ b/fs/xfs/linux-2.6/xfs_vnode.h @@ -45,18 +45,6 @@ static inline struct inode *vn_to_inode(bhv_vnode_t *vnode) return vnode; } -/* - * Values for the vop_rwlock/rwunlock flags parameter. - */ -typedef enum bhv_vrwlock { - VRWLOCK_NONE, - VRWLOCK_READ, - VRWLOCK_WRITE, - VRWLOCK_WRITE_DIRECT, - VRWLOCK_TRY_READ, - VRWLOCK_TRY_WRITE -} bhv_vrwlock_t; - /* * Return values for xfs_inactive. A return value of * VN_INACTIVE_NOCACHE implies that the file system behavior diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index 1d8a4728d84..110ee83fcbe 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -67,7 +67,7 @@ struct xfs_mru_cache; */ typedef int (*xfs_send_data_t)(int, bhv_vnode_t *, - xfs_off_t, size_t, int, bhv_vrwlock_t *); + xfs_off_t, size_t, int, int *); typedef int (*xfs_send_mmap_t)(struct vm_area_struct *, uint); typedef int (*xfs_send_destroy_t)(bhv_vnode_t *, dm_right_t); typedef int (*xfs_send_namesp_t)(dm_eventtype_t, struct xfs_mount *, diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index 4765e7c4b75..811ee874d86 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -3374,47 +3374,6 @@ std_return: goto std_return; } -int -xfs_rwlock( - xfs_inode_t *ip, - bhv_vrwlock_t locktype) -{ - if (S_ISDIR(ip->i_d.di_mode)) - return 1; - if (locktype == VRWLOCK_WRITE) { - xfs_ilock(ip, XFS_IOLOCK_EXCL); - } else if (locktype == VRWLOCK_TRY_READ) { - return xfs_ilock_nowait(ip, XFS_IOLOCK_SHARED); - } else if (locktype == VRWLOCK_TRY_WRITE) { - return xfs_ilock_nowait(ip, XFS_IOLOCK_EXCL); - } else { - ASSERT((locktype == VRWLOCK_READ) || - (locktype == VRWLOCK_WRITE_DIRECT)); - xfs_ilock(ip, XFS_IOLOCK_SHARED); - } - - return 1; -} - - -void -xfs_rwunlock( - xfs_inode_t *ip, - bhv_vrwlock_t locktype) -{ - if (S_ISDIR(ip->i_d.di_mode)) - return; - if (locktype == VRWLOCK_WRITE) { - xfs_iunlock(ip, XFS_IOLOCK_EXCL); - } else { - ASSERT((locktype == VRWLOCK_READ) || - (locktype == VRWLOCK_WRITE_DIRECT)); - xfs_iunlock(ip, XFS_IOLOCK_SHARED); - } - return; -} - - int xfs_inode_flush( xfs_inode_t *ip, diff --git a/fs/xfs/xfs_vnodeops.h b/fs/xfs/xfs_vnodeops.h index 4e3970f0e5e..85340bafd42 100644 --- a/fs/xfs/xfs_vnodeops.h +++ b/fs/xfs/xfs_vnodeops.h @@ -38,8 +38,6 @@ int xfs_readdir(struct xfs_inode *dp, void *dirent, size_t bufsize, int xfs_symlink(struct xfs_inode *dp, bhv_vname_t *dentry, char *target_path, mode_t mode, bhv_vnode_t **vpp, struct cred *credp); -int xfs_rwlock(struct xfs_inode *ip, bhv_vrwlock_t locktype); -void xfs_rwunlock(struct xfs_inode *ip, bhv_vrwlock_t locktype); int xfs_inode_flush(struct xfs_inode *ip, int flags); int xfs_set_dmattrs(struct xfs_inode *ip, u_int evmask, u_int16_t state); int xfs_reclaim(struct xfs_inode *ip); -- cgit v1.2.3 From 24bd861d1c3fff5248de7ba3bdddb3369087ad46 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 6 Mar 2008 13:45:16 +1100 Subject: [XFS] don't encode parent in nfs filehandles unless nessecary As Dave pointed out after the export ops changes we now always encode the parent into the filehandle for regular files, but it's not actually needed when the filesystem is export with no_subtree_check. This one-liner fixes xfs_fs_encode_fh to skip encoding the parent unless nessecary. SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30535a Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_export.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_export.c b/fs/xfs/linux-2.6/xfs_export.c index ca4f66c4de1..21f0e825759 100644 --- a/fs/xfs/linux-2.6/xfs_export.c +++ b/fs/xfs/linux-2.6/xfs_export.c @@ -66,7 +66,7 @@ xfs_fs_encode_fh( int len; /* Directories don't need their parent encoded, they have ".." */ - if (S_ISDIR(inode->i_mode)) + if (S_ISDIR(inode->i_mode) || !connectable) fileid_type = FILEID_INO32_GEN; else fileid_type = FILEID_INO32_GEN_PARENT; -- cgit v1.2.3 From 44d814ced4cffbfe6a775c5bb8b941a6e734e7d9 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 6 Mar 2008 13:45:29 +1100 Subject: [XFS] Update c/mtime correctly on truncates XFS changes the c/mtime of an inode when truncating it to the same size. The c/mtime is only supposed to change if the size is changed. Not to be confused with ftruncate, where the c/mtime is supposed to be changed even if the size is not changed. The Linux VFS encodes this semantic difference in the flags it sends down to ->setattr, which XFS currently ignores. We need to make XFS pay attention to the VFS flags and hence Do The Right Thing. SGI-PV: 977547 SGI-Modid: xfs-linux-melb:xfs-kern:30536a Signed-off-by: David Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_vnodeops.c | 13 +++++++++---- 1 file changed, 9 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index 811ee874d86..b77dede91b7 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -633,6 +633,15 @@ xfs_setattr( * Truncate file. Must have write permission and not be a directory. */ if (mask & XFS_AT_SIZE) { + /* + * Only change the c/mtime if we are changing the size + * or we are explicitly asked to change it. This handles + * the semantic difference between truncate() and ftruncate() + * as implemented in the VFS. + */ + if (vap->va_size != ip->i_size || (mask & XFS_AT_CTIME)) + timeflags |= XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG; + if (vap->va_size > ip->i_size) { xfs_igrow_finish(tp, ip, vap->va_size, !(flags & ATTR_DMI)); @@ -661,10 +670,6 @@ xfs_setattr( */ xfs_iflags_set(ip, XFS_ITRUNCATED); } - /* - * Have to do this even if the file's size doesn't change. - */ - timeflags |= XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG; } /* -- cgit v1.2.3 From 6ee4752ffe782be6e86bea1403a2fe0f682aa71a Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 6 Mar 2008 13:45:35 +1100 Subject: [XFS] Use atomic counters for ktrace buffer indexes ktrace_enter() is consuming vast amounts of CPU time due to the use of a single global lock for protecting buffer index increments. Change it to use per-buffer atomic counters - this reduces ktrace_enter() overhead during a trace intensive test on a 4p machine from 58% of all CPU time to 12% and halves test runtime. SGI-PV: 977546 SGI-Modid: xfs-linux-melb:xfs-kern:30537a Signed-off-by: David Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/support/ktrace.c | 21 ++++++++------------- fs/xfs/support/ktrace.h | 2 +- 2 files changed, 9 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/xfs/support/ktrace.c b/fs/xfs/support/ktrace.c index 129067cfcb8..4e0444c0aca 100644 --- a/fs/xfs/support/ktrace.c +++ b/fs/xfs/support/ktrace.c @@ -92,7 +92,7 @@ ktrace_alloc(int nentries, unsigned int __nocast sleep) ktp->kt_entries = ktep; ktp->kt_nentries = nentries; - ktp->kt_index = 0; + atomic_set(&ktp->kt_index, 0); ktp->kt_rollover = 0; return ktp; } @@ -151,8 +151,6 @@ ktrace_enter( void *val14, void *val15) { - static DEFINE_SPINLOCK(wrap_lock); - unsigned long flags; int index; ktrace_entry_t *ktep; @@ -161,12 +159,8 @@ ktrace_enter( /* * Grab an entry by pushing the index up to the next one. */ - spin_lock_irqsave(&wrap_lock, flags); - index = ktp->kt_index; - if (++ktp->kt_index == ktp->kt_nentries) - ktp->kt_index = 0; - spin_unlock_irqrestore(&wrap_lock, flags); - + index = atomic_add_return(1, &ktp->kt_index); + index = (index - 1) % ktp->kt_nentries; if (!ktp->kt_rollover && index == ktp->kt_nentries - 1) ktp->kt_rollover = 1; @@ -199,11 +193,12 @@ int ktrace_nentries( ktrace_t *ktp) { - if (ktp == NULL) { + int index; + if (ktp == NULL) return 0; - } - return (ktp->kt_rollover ? ktp->kt_nentries : ktp->kt_index); + index = atomic_read(&ktp->kt_index) % ktp->kt_nentries; + return (ktp->kt_rollover ? ktp->kt_nentries : index); } /* @@ -228,7 +223,7 @@ ktrace_first(ktrace_t *ktp, ktrace_snap_t *ktsp) int nentries; if (ktp->kt_rollover) - index = ktp->kt_index; + index = atomic_read(&ktp->kt_index) % ktp->kt_nentries; else index = 0; diff --git a/fs/xfs/support/ktrace.h b/fs/xfs/support/ktrace.h index 56e72b40a85..782dbbb6a9d 100644 --- a/fs/xfs/support/ktrace.h +++ b/fs/xfs/support/ktrace.h @@ -30,7 +30,7 @@ typedef struct ktrace_entry { */ typedef struct ktrace { int kt_nentries; /* number of entries in trace buf */ - int kt_index; /* current index in entries */ + atomic_t kt_index; /* current index in entries */ int kt_rollover; ktrace_entry_t *kt_entries; /* buffer of entries */ } ktrace_t; -- cgit v1.2.3 From d234154125197053d5215711b5df867979e55ebd Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 6 Mar 2008 13:45:43 +1100 Subject: [XFS] Use power-of-2 sized buffers to reduce overhead Now that the ktrace_enter() code is using atomics, the non-power-of-2 buffer sizes - which require modulus operations to get the index - are showing up as using substantial CPU in the profiles. Force the buffer sizes to be rounded up to the nearest power of two and use masking rather than modulus operations to convert the index counter to the buffer index. This reduces ktrace_enter overhead to 8% of a CPU time, and again almost halves the trace intensive test runtime. SGI-PV: 977546 SGI-Modid: xfs-linux-melb:xfs-kern:30538a Signed-off-by: David Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/support/ktrace.c | 22 ++++++++++++++-------- fs/xfs/support/ktrace.h | 1 + 2 files changed, 15 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/xfs/support/ktrace.c b/fs/xfs/support/ktrace.c index 4e0444c0aca..0b75d302508 100644 --- a/fs/xfs/support/ktrace.c +++ b/fs/xfs/support/ktrace.c @@ -24,7 +24,7 @@ static int ktrace_zentries; void __init ktrace_init(int zentries) { - ktrace_zentries = zentries; + ktrace_zentries = roundup_pow_of_two(zentries); ktrace_hdr_zone = kmem_zone_init(sizeof(ktrace_t), "ktrace_hdr"); @@ -47,13 +47,16 @@ ktrace_uninit(void) * ktrace_alloc() * * Allocate a ktrace header and enough buffering for the given - * number of entries. + * number of entries. Round the number of entries up to a + * power of 2 so we can do fast masking to get the index from + * the atomic index counter. */ ktrace_t * ktrace_alloc(int nentries, unsigned int __nocast sleep) { ktrace_t *ktp; ktrace_entry_t *ktep; + int entries; ktp = (ktrace_t*)kmem_zone_alloc(ktrace_hdr_zone, sleep); @@ -70,11 +73,12 @@ ktrace_alloc(int nentries, unsigned int __nocast sleep) /* * Special treatment for buffers with the ktrace_zentries entries */ - if (nentries == ktrace_zentries) { + entries = roundup_pow_of_two(nentries); + if (entries == ktrace_zentries) { ktep = (ktrace_entry_t*)kmem_zone_zalloc(ktrace_ent_zone, sleep); } else { - ktep = (ktrace_entry_t*)kmem_zalloc((nentries * sizeof(*ktep)), + ktep = (ktrace_entry_t*)kmem_zalloc((entries * sizeof(*ktep)), sleep | KM_LARGE); } @@ -91,7 +95,9 @@ ktrace_alloc(int nentries, unsigned int __nocast sleep) } ktp->kt_entries = ktep; - ktp->kt_nentries = nentries; + ktp->kt_nentries = entries; + ASSERT(is_power_of_2(entries)); + ktp->kt_index_mask = entries - 1; atomic_set(&ktp->kt_index, 0); ktp->kt_rollover = 0; return ktp; @@ -160,7 +166,7 @@ ktrace_enter( * Grab an entry by pushing the index up to the next one. */ index = atomic_add_return(1, &ktp->kt_index); - index = (index - 1) % ktp->kt_nentries; + index = (index - 1) & ktp->kt_index_mask; if (!ktp->kt_rollover && index == ktp->kt_nentries - 1) ktp->kt_rollover = 1; @@ -197,7 +203,7 @@ ktrace_nentries( if (ktp == NULL) return 0; - index = atomic_read(&ktp->kt_index) % ktp->kt_nentries; + index = atomic_read(&ktp->kt_index) & ktp->kt_index_mask; return (ktp->kt_rollover ? ktp->kt_nentries : index); } @@ -223,7 +229,7 @@ ktrace_first(ktrace_t *ktp, ktrace_snap_t *ktsp) int nentries; if (ktp->kt_rollover) - index = atomic_read(&ktp->kt_index) % ktp->kt_nentries; + index = atomic_read(&ktp->kt_index) & ktp->kt_index_mask; else index = 0; diff --git a/fs/xfs/support/ktrace.h b/fs/xfs/support/ktrace.h index 782dbbb6a9d..741d6947ca6 100644 --- a/fs/xfs/support/ktrace.h +++ b/fs/xfs/support/ktrace.h @@ -31,6 +31,7 @@ typedef struct ktrace_entry { typedef struct ktrace { int kt_nentries; /* number of entries in trace buf */ atomic_t kt_index; /* current index in entries */ + unsigned int kt_index_mask; int kt_rollover; ktrace_entry_t *kt_entries; /* buffer of entries */ } ktrace_t; -- cgit v1.2.3 From bc4ac74a4e5bd7db02976eb1b681e1d11f81c9ce Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 6 Mar 2008 13:45:58 +1100 Subject: [XFS] cleanup vnode use in dmapi calls SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30545a Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_file.c | 13 ++++------- fs/xfs/linux-2.6/xfs_lrw.c | 14 +++++------- fs/xfs/xfs_bmap.c | 2 +- fs/xfs/xfs_mount.h | 22 +++++++++--------- fs/xfs/xfs_rename.c | 8 +++---- fs/xfs/xfs_vfsops.c | 4 ++-- fs/xfs/xfs_vnodeops.c | 55 +++++++++++++++++++++------------------------ 7 files changed, 54 insertions(+), 64 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_file.c b/fs/xfs/linux-2.6/xfs_file.c index edab1ffbb16..05905246434 100644 --- a/fs/xfs/linux-2.6/xfs_file.c +++ b/fs/xfs/linux-2.6/xfs_file.c @@ -469,16 +469,11 @@ xfs_file_open_exec( struct inode *inode) { struct xfs_mount *mp = XFS_M(inode->i_sb); + struct xfs_inode *ip = XFS_I(inode); - if (unlikely(mp->m_flags & XFS_MOUNT_DMAPI)) { - if (DM_EVENT_ENABLED(XFS_I(inode), DM_EVENT_READ)) { - bhv_vnode_t *vp = vn_from_inode(inode); - - return -XFS_SEND_DATA(mp, DM_EVENT_READ, - vp, 0, 0, 0, NULL); - } - } - + if (unlikely(mp->m_flags & XFS_MOUNT_DMAPI) && + DM_EVENT_ENABLED(ip, DM_EVENT_READ)) + return -XFS_SEND_DATA(mp, DM_EVENT_READ, ip, 0, 0, 0, NULL); return 0; } #endif /* HAVE_FOP_OPEN_EXEC */ diff --git a/fs/xfs/linux-2.6/xfs_lrw.c b/fs/xfs/linux-2.6/xfs_lrw.c index 3c20007ab48..01a8f26e1b1 100644 --- a/fs/xfs/linux-2.6/xfs_lrw.c +++ b/fs/xfs/linux-2.6/xfs_lrw.c @@ -231,7 +231,7 @@ xfs_read( int dmflags = FILP_DELAY_FLAG(file) | DM_SEM_FLAG_RD(ioflags); int iolock = XFS_IOLOCK_SHARED; - ret = -XFS_SEND_DATA(mp, DM_EVENT_READ, vp, *offset, size, + ret = -XFS_SEND_DATA(mp, DM_EVENT_READ, ip, *offset, size, dmflags, &iolock); if (ret) { xfs_iunlock(ip, XFS_IOLOCK_SHARED); @@ -276,7 +276,6 @@ xfs_splice_read( int flags, int ioflags) { - bhv_vnode_t *vp = XFS_ITOV(ip); xfs_mount_t *mp = ip->i_mount; ssize_t ret; @@ -290,7 +289,7 @@ xfs_splice_read( int iolock = XFS_IOLOCK_SHARED; int error; - error = XFS_SEND_DATA(mp, DM_EVENT_READ, vp, *ppos, count, + error = XFS_SEND_DATA(mp, DM_EVENT_READ, ip, *ppos, count, FILP_DELAY_FLAG(infilp), &iolock); if (error) { xfs_iunlock(ip, XFS_IOLOCK_SHARED); @@ -317,7 +316,6 @@ xfs_splice_write( int flags, int ioflags) { - bhv_vnode_t *vp = XFS_ITOV(ip); xfs_mount_t *mp = ip->i_mount; ssize_t ret; struct inode *inode = outfilp->f_mapping->host; @@ -333,7 +331,7 @@ xfs_splice_write( int iolock = XFS_IOLOCK_EXCL; int error; - error = XFS_SEND_DATA(mp, DM_EVENT_WRITE, vp, *ppos, count, + error = XFS_SEND_DATA(mp, DM_EVENT_WRITE, ip, *ppos, count, FILP_DELAY_FLAG(outfilp), &iolock); if (error) { xfs_iunlock(ip, XFS_IOLOCK_EXCL); @@ -631,7 +629,7 @@ start: dmflags |= DM_FLAGS_IMUX; xfs_iunlock(xip, XFS_ILOCK_EXCL); - error = XFS_SEND_DATA(xip->i_mount, DM_EVENT_WRITE, vp, + error = XFS_SEND_DATA(xip->i_mount, DM_EVENT_WRITE, xip, pos, count, dmflags, &iolock); if (error) { goto out_unlock_internal; @@ -778,8 +776,8 @@ retry: xfs_iunlock(xip, iolock); if (need_i_mutex) mutex_unlock(&inode->i_mutex); - error = XFS_SEND_NAMESP(xip->i_mount, DM_EVENT_NOSPACE, vp, - DM_RIGHT_NULL, vp, DM_RIGHT_NULL, NULL, NULL, + error = XFS_SEND_NAMESP(xip->i_mount, DM_EVENT_NOSPACE, xip, + DM_RIGHT_NULL, xip, DM_RIGHT_NULL, NULL, NULL, 0, 0, 0); /* Delay flag intentionally unused */ if (need_i_mutex) mutex_lock(&inode->i_mutex); diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 87f64674981..19aae13b7f9 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -5811,7 +5811,7 @@ xfs_getbmap( if ((interface & BMV_IF_NO_DMAPI_READ) == 0 && DM_EVENT_ENABLED(ip, DM_EVENT_READ) && whichfork == XFS_DATA_FORK) { - error = XFS_SEND_DATA(mp, DM_EVENT_READ, vp, 0, 0, 0, NULL); + error = XFS_SEND_DATA(mp, DM_EVENT_READ, ip, 0, 0, 0, NULL); if (error) return XFS_ERROR(error); } diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index 110ee83fcbe..7b37fa00929 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -66,17 +66,17 @@ struct xfs_mru_cache; * Prototypes and functions for the Data Migration subsystem. */ -typedef int (*xfs_send_data_t)(int, bhv_vnode_t *, +typedef int (*xfs_send_data_t)(int, struct xfs_inode *, xfs_off_t, size_t, int, int *); typedef int (*xfs_send_mmap_t)(struct vm_area_struct *, uint); -typedef int (*xfs_send_destroy_t)(bhv_vnode_t *, dm_right_t); +typedef int (*xfs_send_destroy_t)(struct xfs_inode *, dm_right_t); typedef int (*xfs_send_namesp_t)(dm_eventtype_t, struct xfs_mount *, - bhv_vnode_t *, - dm_right_t, bhv_vnode_t *, dm_right_t, + struct xfs_inode *, dm_right_t, + struct xfs_inode *, dm_right_t, char *, char *, mode_t, int, int); typedef int (*xfs_send_mount_t)(struct xfs_mount *, dm_right_t, char *, char *); -typedef void (*xfs_send_unmount_t)(struct xfs_mount *, bhv_vnode_t *, +typedef void (*xfs_send_unmount_t)(struct xfs_mount *, struct xfs_inode *, dm_right_t, mode_t, int, int); typedef struct xfs_dmops { @@ -88,20 +88,20 @@ typedef struct xfs_dmops { xfs_send_unmount_t xfs_send_unmount; } xfs_dmops_t; -#define XFS_SEND_DATA(mp, ev,vp,off,len,fl,lock) \ - (*(mp)->m_dm_ops->xfs_send_data)(ev,vp,off,len,fl,lock) +#define XFS_SEND_DATA(mp, ev,ip,off,len,fl,lock) \ + (*(mp)->m_dm_ops->xfs_send_data)(ev,ip,off,len,fl,lock) #define XFS_SEND_MMAP(mp, vma,fl) \ (*(mp)->m_dm_ops->xfs_send_mmap)(vma,fl) -#define XFS_SEND_DESTROY(mp, vp,right) \ - (*(mp)->m_dm_ops->xfs_send_destroy)(vp,right) +#define XFS_SEND_DESTROY(mp, ip,right) \ + (*(mp)->m_dm_ops->xfs_send_destroy)(ip,right) #define XFS_SEND_NAMESP(mp, ev,b1,r1,b2,r2,n1,n2,mode,rval,fl) \ (*(mp)->m_dm_ops->xfs_send_namesp)(ev,NULL,b1,r1,b2,r2,n1,n2,mode,rval,fl) #define XFS_SEND_PREUNMOUNT(mp,b1,r1,b2,r2,n1,n2,mode,rval,fl) \ (*(mp)->m_dm_ops->xfs_send_namesp)(DM_EVENT_PREUNMOUNT,mp,b1,r1,b2,r2,n1,n2,mode,rval,fl) #define XFS_SEND_MOUNT(mp,right,path,name) \ (*(mp)->m_dm_ops->xfs_send_mount)(mp,right,path,name) -#define XFS_SEND_UNMOUNT(mp, vp,right,mode,rval,fl) \ - (*(mp)->m_dm_ops->xfs_send_unmount)(mp,vp,right,mode,rval,fl) +#define XFS_SEND_UNMOUNT(mp, ip,right,mode,rval,fl) \ + (*(mp)->m_dm_ops->xfs_send_unmount)(mp,ip,right,mode,rval,fl) /* diff --git a/fs/xfs/xfs_rename.c b/fs/xfs/xfs_rename.c index fd1244cf50a..6f80cfdfbd8 100644 --- a/fs/xfs/xfs_rename.c +++ b/fs/xfs/xfs_rename.c @@ -258,8 +258,8 @@ xfs_rename( if (DM_EVENT_ENABLED(src_dp, DM_EVENT_RENAME) || DM_EVENT_ENABLED(target_dp, DM_EVENT_RENAME)) { error = XFS_SEND_NAMESP(mp, DM_EVENT_RENAME, - src_dir_vp, DM_RIGHT_NULL, - target_dir_vp, DM_RIGHT_NULL, + src_dp, DM_RIGHT_NULL, + target_dp, DM_RIGHT_NULL, src_name, target_name, 0, 0, 0); if (error) { @@ -591,8 +591,8 @@ std_return: if (DM_EVENT_ENABLED(src_dp, DM_EVENT_POSTRENAME) || DM_EVENT_ENABLED(target_dp, DM_EVENT_POSTRENAME)) { (void) XFS_SEND_NAMESP (mp, DM_EVENT_POSTRENAME, - src_dir_vp, DM_RIGHT_NULL, - target_dir_vp, DM_RIGHT_NULL, + src_dp, DM_RIGHT_NULL, + target_dp, DM_RIGHT_NULL, src_name, target_name, 0, error, 0); } diff --git a/fs/xfs/xfs_vfsops.c b/fs/xfs/xfs_vfsops.c index 3ec27bf8531..4c132a87d43 100644 --- a/fs/xfs/xfs_vfsops.c +++ b/fs/xfs/xfs_vfsops.c @@ -566,7 +566,7 @@ xfs_unmount( #ifdef HAVE_DMAPI if (mp->m_flags & XFS_MOUNT_DMAPI) { error = XFS_SEND_PREUNMOUNT(mp, - rvp, DM_RIGHT_NULL, rvp, DM_RIGHT_NULL, + rip, DM_RIGHT_NULL, rip, DM_RIGHT_NULL, NULL, NULL, 0, 0, (mp->m_dmevmask & (1<va_size, 0, dmflags, NULL); if (code) { lock_flags = 0; @@ -881,7 +881,7 @@ xfs_setattr( if (DM_EVENT_ENABLED(ip, DM_EVENT_ATTRIBUTE) && !(flags & ATTR_DMI)) { - (void) XFS_SEND_NAMESP(mp, DM_EVENT_ATTRIBUTE, vp, DM_RIGHT_NULL, + (void) XFS_SEND_NAMESP(mp, DM_EVENT_ATTRIBUTE, ip, DM_RIGHT_NULL, NULL, DM_RIGHT_NULL, NULL, NULL, 0, 0, AT_DELAY_FLAG(flags)); } @@ -1586,9 +1586,8 @@ xfs_inactive( mp = ip->i_mount; - if (ip->i_d.di_nlink == 0 && DM_EVENT_ENABLED(ip, DM_EVENT_DESTROY)) { - (void) XFS_SEND_DESTROY(mp, vp, DM_RIGHT_NULL); - } + if (ip->i_d.di_nlink == 0 && DM_EVENT_ENABLED(ip, DM_EVENT_DESTROY)) + XFS_SEND_DESTROY(mp, ip, DM_RIGHT_NULL); error = 0; @@ -1820,7 +1819,7 @@ xfs_create( if (DM_EVENT_ENABLED(dp, DM_EVENT_CREATE)) { error = XFS_SEND_NAMESP(mp, DM_EVENT_CREATE, - dir_vp, DM_RIGHT_NULL, NULL, + dp, DM_RIGHT_NULL, NULL, DM_RIGHT_NULL, name, NULL, mode, 0, 0); @@ -1976,8 +1975,8 @@ std_return: if ((*vpp || (error != 0 && dm_event_sent != 0)) && DM_EVENT_ENABLED(dp, DM_EVENT_POSTCREATE)) { (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTCREATE, - dir_vp, DM_RIGHT_NULL, - *vpp ? vp:NULL, + dp, DM_RIGHT_NULL, + *vpp ? ip : NULL, DM_RIGHT_NULL, name, NULL, mode, error, 0); } @@ -2272,7 +2271,6 @@ xfs_remove( xfs_inode_t *dp, bhv_vname_t *dentry) { - bhv_vnode_t *dir_vp = XFS_ITOV(dp); char *name = VNAME(dentry); xfs_mount_t *mp = dp->i_mount; xfs_inode_t *ip = VNAME_TO_INODE(dentry); @@ -2292,7 +2290,7 @@ xfs_remove( return XFS_ERROR(EIO); if (DM_EVENT_ENABLED(dp, DM_EVENT_REMOVE)) { - error = XFS_SEND_NAMESP(mp, DM_EVENT_REMOVE, dir_vp, + error = XFS_SEND_NAMESP(mp, DM_EVENT_REMOVE, dp, DM_RIGHT_NULL, NULL, DM_RIGHT_NULL, name, NULL, ip->i_d.di_mode, 0, 0); if (error) @@ -2445,7 +2443,7 @@ xfs_remove( std_return: if (DM_EVENT_ENABLED(dp, DM_EVENT_POSTREMOVE)) { (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTREMOVE, - dir_vp, DM_RIGHT_NULL, + dp, DM_RIGHT_NULL, NULL, DM_RIGHT_NULL, name, NULL, ip->i_d.di_mode, error, 0); } @@ -2504,8 +2502,8 @@ xfs_link( if (DM_EVENT_ENABLED(tdp, DM_EVENT_LINK)) { error = XFS_SEND_NAMESP(mp, DM_EVENT_LINK, - target_dir_vp, DM_RIGHT_NULL, - src_vp, DM_RIGHT_NULL, + tdp, DM_RIGHT_NULL, + sip, DM_RIGHT_NULL, target_name, NULL, 0, 0, 0); if (error) return error; @@ -2615,8 +2613,8 @@ xfs_link( std_return: if (DM_EVENT_ENABLED(sip, DM_EVENT_POSTLINK)) { (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTLINK, - target_dir_vp, DM_RIGHT_NULL, - src_vp, DM_RIGHT_NULL, + tdp, DM_RIGHT_NULL, + sip, DM_RIGHT_NULL, target_name, NULL, 0, error, 0); } return error; @@ -2665,7 +2663,7 @@ xfs_mkdir( if (DM_EVENT_ENABLED(dp, DM_EVENT_CREATE)) { error = XFS_SEND_NAMESP(mp, DM_EVENT_CREATE, - dir_vp, DM_RIGHT_NULL, NULL, + dp, DM_RIGHT_NULL, NULL, DM_RIGHT_NULL, dir_name, NULL, mode, 0, 0); if (error) @@ -2823,8 +2821,8 @@ std_return: if ((created || (error != 0 && dm_event_sent != 0)) && DM_EVENT_ENABLED(dp, DM_EVENT_POSTCREATE)) { (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTCREATE, - dir_vp, DM_RIGHT_NULL, - created ? XFS_ITOV(cdp):NULL, + dp, DM_RIGHT_NULL, + created ? cdp : NULL, DM_RIGHT_NULL, dir_name, NULL, mode, error, 0); @@ -2873,7 +2871,7 @@ xfs_rmdir( if (DM_EVENT_ENABLED(dp, DM_EVENT_REMOVE)) { error = XFS_SEND_NAMESP(mp, DM_EVENT_REMOVE, - dir_vp, DM_RIGHT_NULL, + dp, DM_RIGHT_NULL, NULL, DM_RIGHT_NULL, name, NULL, cdp->i_d.di_mode, 0, 0); if (error) @@ -3047,7 +3045,7 @@ xfs_rmdir( std_return: if (DM_EVENT_ENABLED(dp, DM_EVENT_POSTREMOVE)) { (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTREMOVE, - dir_vp, DM_RIGHT_NULL, + dp, DM_RIGHT_NULL, NULL, DM_RIGHT_NULL, name, NULL, cdp->i_d.di_mode, error, 0); @@ -3144,7 +3142,7 @@ xfs_symlink( } if (DM_EVENT_ENABLED(dp, DM_EVENT_SYMLINK)) { - error = XFS_SEND_NAMESP(mp, DM_EVENT_SYMLINK, dir_vp, + error = XFS_SEND_NAMESP(mp, DM_EVENT_SYMLINK, dp, DM_RIGHT_NULL, NULL, DM_RIGHT_NULL, link_name, target_path, 0, 0, 0); if (error) @@ -3348,8 +3346,8 @@ xfs_symlink( std_return: if (DM_EVENT_ENABLED(dp, DM_EVENT_POSTSYMLINK)) { (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTSYMLINK, - dir_vp, DM_RIGHT_NULL, - error ? NULL : XFS_ITOV(ip), + dp, DM_RIGHT_NULL, + error ? NULL : ip, DM_RIGHT_NULL, link_name, target_path, 0, error, 0); } @@ -3707,9 +3705,8 @@ xfs_alloc_file_space( end_dmi_offset = offset+len; if (end_dmi_offset > ip->i_size) end_dmi_offset = ip->i_size; - error = XFS_SEND_DATA(mp, DM_EVENT_WRITE, XFS_ITOV(ip), - offset, end_dmi_offset - offset, - 0, NULL); + error = XFS_SEND_DATA(mp, DM_EVENT_WRITE, ip, offset, + end_dmi_offset - offset, 0, NULL); if (error) return error; } @@ -3818,8 +3815,8 @@ dmapi_enospc_check: if (error == ENOSPC && (attr_flags & ATTR_DMI) == 0 && DM_EVENT_ENABLED(ip, DM_EVENT_NOSPACE)) { error = XFS_SEND_NAMESP(mp, DM_EVENT_NOSPACE, - XFS_ITOV(ip), DM_RIGHT_NULL, - XFS_ITOV(ip), DM_RIGHT_NULL, + ip, DM_RIGHT_NULL, + ip, DM_RIGHT_NULL, NULL, NULL, 0, 0, 0); /* Delay flag intentionally unused */ if (error == 0) goto retry; /* Maybe DMAPI app. has made space */ @@ -3964,7 +3961,7 @@ xfs_free_file_space( DM_EVENT_ENABLED(ip, DM_EVENT_WRITE)) { if (end_dmi_offset > ip->i_size) end_dmi_offset = ip->i_size; - error = XFS_SEND_DATA(mp, DM_EVENT_WRITE, vp, + error = XFS_SEND_DATA(mp, DM_EVENT_WRITE, ip, offset, end_dmi_offset - offset, AT_DELAY_FLAG(attr_flags), NULL); if (error) -- cgit v1.2.3 From 979ebab11623894528d4d37b947533ea4e8649d1 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 6 Mar 2008 13:46:05 +1100 Subject: [XFS] cleanup vnode use in xfs_create/mknod/mkdir SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30546a Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_iops.c | 21 +++++++++++---------- fs/xfs/xfs_vnodeops.c | 25 +++++++++---------------- fs/xfs/xfs_vnodeops.h | 4 ++-- 3 files changed, 22 insertions(+), 28 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c index 34670118331..62899a1ec7f 100644 --- a/fs/xfs/linux-2.6/xfs_iops.c +++ b/fs/xfs/linux-2.6/xfs_iops.c @@ -273,7 +273,7 @@ xfs_vn_mknod( dev_t rdev) { struct inode *inode; - bhv_vnode_t *vp = NULL, *dvp = vn_from_inode(dir); + struct xfs_inode *ip = NULL; xfs_acl_t *default_acl = NULL; attrexists_t test_default_acl = _ACL_DEFAULT_EXISTS; int error; @@ -285,11 +285,11 @@ xfs_vn_mknod( if (unlikely(!sysv_valid_dev(rdev) || MAJOR(rdev) & ~0x1ff)) return -EINVAL; - if (test_default_acl && test_default_acl(dvp)) { + if (test_default_acl && test_default_acl(dir)) { if (!_ACL_ALLOC(default_acl)) { return -ENOMEM; } - if (!_ACL_GET_DEFAULT(dvp, default_acl)) { + if (!_ACL_GET_DEFAULT(dir, default_acl)) { _ACL_FREE(default_acl); default_acl = NULL; } @@ -305,10 +305,10 @@ xfs_vn_mknod( case S_IFSOCK: rdev = sysv_encode_dev(rdev); case S_IFREG: - error = xfs_create(XFS_I(dir), dentry, mode, rdev, &vp, NULL); + error = xfs_create(XFS_I(dir), dentry, mode, rdev, &ip, NULL); break; case S_IFDIR: - error = xfs_mkdir(XFS_I(dir), dentry, mode, &vp, NULL); + error = xfs_mkdir(XFS_I(dir), dentry, mode, &ip, NULL); break; default: error = EINVAL; @@ -318,19 +318,20 @@ xfs_vn_mknod( if (unlikely(error)) goto out_free_acl; - error = xfs_init_security(vp, dir); + inode = ip->i_vnode; + + error = xfs_init_security(inode, dir); if (unlikely(error)) goto out_cleanup_inode; if (default_acl) { - error = _ACL_INHERIT(vp, mode, default_acl); + error = _ACL_INHERIT(inode, mode, default_acl); if (unlikely(error)) goto out_cleanup_inode; - xfs_iflags_set(XFS_I(vp), XFS_IMODIFIED); + xfs_iflags_set(ip, XFS_IMODIFIED); _ACL_FREE(default_acl); } - inode = vn_to_inode(vp); if (S_ISDIR(mode)) xfs_validate_fields(inode); @@ -339,7 +340,7 @@ xfs_vn_mknod( return -error; out_cleanup_inode: - xfs_cleanup_inode(dir, vp, dentry, mode); + xfs_cleanup_inode(dir, inode, dentry, mode); out_free_acl: if (default_acl) _ACL_FREE(default_acl); diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index 7e124b55c26..a42d7fe6a5e 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -1791,14 +1791,12 @@ xfs_create( bhv_vname_t *dentry, mode_t mode, xfs_dev_t rdev, - bhv_vnode_t **vpp, + xfs_inode_t **ipp, cred_t *credp) { char *name = VNAME(dentry); xfs_mount_t *mp = dp->i_mount; - bhv_vnode_t *dir_vp = XFS_ITOV(dp); xfs_inode_t *ip; - bhv_vnode_t *vp = NULL; xfs_trans_t *tp; int error; xfs_bmap_free_t free_list; @@ -1812,7 +1810,7 @@ xfs_create( uint resblks; int namelen; - ASSERT(!*vpp); + ASSERT(!*ipp); xfs_itrace_entry(dp); namelen = VNAMELEN(dentry); @@ -1911,7 +1909,7 @@ xfs_create( * the transaction cancel unlocking dp so don't do it explicitly in the * error path. */ - VN_HOLD(dir_vp); + IHOLD(dp); xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL); unlock_dp_on_error = B_FALSE; @@ -1949,7 +1947,6 @@ xfs_create( * vnode to the caller, we bump the vnode ref count now. */ IHOLD(ip); - vp = XFS_ITOV(ip); error = xfs_bmap_finish(&tp, &free_list, &committed); if (error) { @@ -1967,16 +1964,16 @@ xfs_create( XFS_QM_DQRELE(mp, udqp); XFS_QM_DQRELE(mp, gdqp); - *vpp = vp; + *ipp = ip; /* Fallthrough to std_return with error = 0 */ std_return: - if ((*vpp || (error != 0 && dm_event_sent != 0)) && + if ((*ipp || (error != 0 && dm_event_sent != 0)) && DM_EVENT_ENABLED(dp, DM_EVENT_POSTCREATE)) { (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTCREATE, dp, DM_RIGHT_NULL, - *vpp ? ip : NULL, + *ipp ? ip : NULL, DM_RIGHT_NULL, name, NULL, mode, error, 0); } @@ -2634,15 +2631,13 @@ xfs_mkdir( xfs_inode_t *dp, bhv_vname_t *dentry, mode_t mode, - bhv_vnode_t **vpp, + xfs_inode_t **ipp, cred_t *credp) { - bhv_vnode_t *dir_vp = XFS_ITOV(dp); char *dir_name = VNAME(dentry); int dir_namelen = VNAMELEN(dentry); xfs_mount_t *mp = dp->i_mount; xfs_inode_t *cdp; /* inode of created dir */ - bhv_vnode_t *cvp; /* vnode of created dir */ xfs_trans_t *tp; int cancel_flags; int error; @@ -2749,7 +2744,7 @@ xfs_mkdir( * from here on will result in the transaction cancel * unlocking dp so don't do it explicitly in the error path. */ - VN_HOLD(dir_vp); + IHOLD(dp); xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL); unlock_dp_on_error = B_FALSE; @@ -2780,11 +2775,9 @@ xfs_mkdir( if (error) goto error2; - cvp = XFS_ITOV(cdp); - created = B_TRUE; - *vpp = cvp; + *ipp = cdp; IHOLD(cdp); /* diff --git a/fs/xfs/xfs_vnodeops.h b/fs/xfs/xfs_vnodeops.h index 85340bafd42..0acef123141 100644 --- a/fs/xfs/xfs_vnodeops.h +++ b/fs/xfs/xfs_vnodeops.h @@ -26,12 +26,12 @@ int xfs_inactive(struct xfs_inode *ip); int xfs_lookup(struct xfs_inode *dp, bhv_vname_t *dentry, bhv_vnode_t **vpp); int xfs_create(struct xfs_inode *dp, bhv_vname_t *dentry, mode_t mode, - xfs_dev_t rdev, bhv_vnode_t **vpp, struct cred *credp); + xfs_dev_t rdev, struct xfs_inode **ipp, struct cred *credp); int xfs_remove(struct xfs_inode *dp, bhv_vname_t *dentry); int xfs_link(struct xfs_inode *tdp, bhv_vnode_t *src_vp, bhv_vname_t *dentry); int xfs_mkdir(struct xfs_inode *dp, bhv_vname_t *dentry, - mode_t mode, bhv_vnode_t **vpp, struct cred *credp); + mode_t mode, struct xfs_inode **ipp, struct cred *credp); int xfs_rmdir(struct xfs_inode *dp, bhv_vname_t *dentry); int xfs_readdir(struct xfs_inode *dp, void *dirent, size_t bufsize, xfs_off_t *offset, filldir_t filldir); -- cgit v1.2.3 From a3da789640871c897901c5f766e33be78d56f35a Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 6 Mar 2008 13:46:12 +1100 Subject: [XFS] cleanup vnode use in xfs_link SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30547a Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_iops.c | 23 +++++++++++------------ fs/xfs/xfs_vnodeops.c | 12 +++++------- fs/xfs/xfs_vnodeops.h | 2 +- 3 files changed, 17 insertions(+), 20 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c index 62899a1ec7f..1df48209d60 100644 --- a/fs/xfs/linux-2.6/xfs_iops.c +++ b/fs/xfs/linux-2.6/xfs_iops.c @@ -395,23 +395,22 @@ xfs_vn_link( struct inode *dir, struct dentry *dentry) { - struct inode *ip; /* inode of guy being linked to */ - bhv_vnode_t *vp; /* vp of name being linked */ + struct inode *inode; /* inode of guy being linked to */ int error; - ip = old_dentry->d_inode; /* inode being linked to */ - vp = vn_from_inode(ip); + inode = old_dentry->d_inode; - VN_HOLD(vp); - error = xfs_link(XFS_I(dir), vp, dentry); + igrab(inode); + error = xfs_link(XFS_I(dir), XFS_I(inode), dentry); if (unlikely(error)) { - VN_RELE(vp); - } else { - xfs_iflags_set(XFS_I(dir), XFS_IMODIFIED); - xfs_validate_fields(ip); - d_instantiate(dentry, ip); + iput(inode); + return -error; } - return -error; + + xfs_iflags_set(XFS_I(dir), XFS_IMODIFIED); + xfs_validate_fields(inode); + d_instantiate(dentry, inode); + return 0; } STATIC int diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index a42d7fe6a5e..10d2d22eb03 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -2471,12 +2471,10 @@ xfs_remove( int xfs_link( xfs_inode_t *tdp, - bhv_vnode_t *src_vp, + xfs_inode_t *sip, bhv_vname_t *dentry) { - bhv_vnode_t *target_dir_vp = XFS_ITOV(tdp); xfs_mount_t *mp = tdp->i_mount; - xfs_inode_t *sip = xfs_vtoi(src_vp); xfs_trans_t *tp; xfs_inode_t *ips[2]; int error; @@ -2489,10 +2487,10 @@ xfs_link( int target_namelen; xfs_itrace_entry(tdp); - xfs_itrace_entry(xfs_vtoi(src_vp)); + xfs_itrace_entry(sip); target_namelen = VNAMELEN(dentry); - ASSERT(!VN_ISDIR(src_vp)); + ASSERT(!S_ISDIR(sip->i_d.di_mode)); if (XFS_FORCED_SHUTDOWN(mp)) return XFS_ERROR(EIO); @@ -2544,8 +2542,8 @@ xfs_link( * xfs_trans_cancel will both unlock the inodes and * decrement the associated ref counts. */ - VN_HOLD(src_vp); - VN_HOLD(target_dir_vp); + IHOLD(sip); + IHOLD(tdp); xfs_trans_ijoin(tp, sip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, tdp, XFS_ILOCK_EXCL); diff --git a/fs/xfs/xfs_vnodeops.h b/fs/xfs/xfs_vnodeops.h index 0acef123141..79c13f57a81 100644 --- a/fs/xfs/xfs_vnodeops.h +++ b/fs/xfs/xfs_vnodeops.h @@ -28,7 +28,7 @@ int xfs_lookup(struct xfs_inode *dp, bhv_vname_t *dentry, int xfs_create(struct xfs_inode *dp, bhv_vname_t *dentry, mode_t mode, xfs_dev_t rdev, struct xfs_inode **ipp, struct cred *credp); int xfs_remove(struct xfs_inode *dp, bhv_vname_t *dentry); -int xfs_link(struct xfs_inode *tdp, bhv_vnode_t *src_vp, +int xfs_link(struct xfs_inode *tdp, struct xfs_inode *sip, bhv_vname_t *dentry); int xfs_mkdir(struct xfs_inode *dp, bhv_vname_t *dentry, mode_t mode, struct xfs_inode **ipp, struct cred *credp); -- cgit v1.2.3 From 3937be5ba836a204d3d1df96b518eecd6cdacbb9 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 6 Mar 2008 13:46:19 +1100 Subject: [XFS] cleanup vnode use in xfs_symlink and xfs_rename SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30548a Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_iops.c | 41 +++++++++++++++++++++-------------------- fs/xfs/xfs_rename.c | 20 +++++--------------- fs/xfs/xfs_vnodeops.c | 16 +++++----------- fs/xfs/xfs_vnodeops.h | 4 ++-- 4 files changed, 33 insertions(+), 48 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c index 1df48209d60..215158cbac4 100644 --- a/fs/xfs/linux-2.6/xfs_iops.c +++ b/fs/xfs/linux-2.6/xfs_iops.c @@ -437,29 +437,33 @@ xfs_vn_symlink( struct dentry *dentry, const char *symname) { - struct inode *ip; - bhv_vnode_t *cvp; /* used to lookup symlink to put in dentry */ + struct inode *inode; + struct xfs_inode *cip = NULL; int error; mode_t mode; - cvp = NULL; - mode = S_IFLNK | (irix_symlink_mode ? 0777 & ~current->fs->umask : S_IRWXUGO); error = xfs_symlink(XFS_I(dir), dentry, (char *)symname, mode, - &cvp, NULL); - if (likely(!error && cvp)) { - error = xfs_init_security(cvp, dir); - if (likely(!error)) { - ip = vn_to_inode(cvp); - d_instantiate(dentry, ip); - xfs_validate_fields(dir); - xfs_validate_fields(ip); - } else { - xfs_cleanup_inode(dir, cvp, dentry, 0); - } - } + &cip, NULL); + if (unlikely(error)) + goto out; + + inode = cip->i_vnode; + + error = xfs_init_security(inode, dir); + if (unlikely(error)) + goto out_cleanup_inode; + + d_instantiate(dentry, inode); + xfs_validate_fields(dir); + xfs_validate_fields(inode); + return 0; + + out_cleanup_inode: + xfs_cleanup_inode(dir, inode, dentry, 0); + out: return -error; } @@ -487,12 +491,9 @@ xfs_vn_rename( struct dentry *ndentry) { struct inode *new_inode = ndentry->d_inode; - bhv_vnode_t *tvp; /* target directory */ int error; - tvp = vn_from_inode(ndir); - - error = xfs_rename(XFS_I(odir), odentry, tvp, ndentry); + error = xfs_rename(XFS_I(odir), odentry, XFS_I(ndir), ndentry); if (likely(!error)) { if (new_inode) xfs_validate_fields(new_inode); diff --git a/fs/xfs/xfs_rename.c b/fs/xfs/xfs_rename.c index 6f80cfdfbd8..c4d0bac56a5 100644 --- a/fs/xfs/xfs_rename.c +++ b/fs/xfs/xfs_rename.c @@ -219,12 +219,11 @@ int xfs_rename( xfs_inode_t *src_dp, bhv_vname_t *src_vname, - bhv_vnode_t *target_dir_vp, + xfs_inode_t *target_dp, bhv_vname_t *target_vname) { - bhv_vnode_t *src_dir_vp = XFS_ITOV(src_dp); xfs_trans_t *tp; - xfs_inode_t *target_dp, *src_ip, *target_ip; + xfs_inode_t *src_ip, *target_ip; xfs_mount_t *mp = src_dp->i_mount; int new_parent; /* moving to a new dir */ int src_is_directory; /* src_name is a directory */ @@ -244,16 +243,7 @@ xfs_rename( int target_namelen = VNAMELEN(target_vname); xfs_itrace_entry(src_dp); - xfs_itrace_entry(xfs_vtoi(target_dir_vp)); - - /* - * Find the XFS behavior descriptor for the target directory - * vnode since it was not handed to us. - */ - target_dp = xfs_vtoi(target_dir_vp); - if (target_dp == NULL) { - return XFS_ERROR(EXDEV); - } + xfs_itrace_entry(target_dp); if (DM_EVENT_ENABLED(src_dp, DM_EVENT_RENAME) || DM_EVENT_ENABLED(target_dp, DM_EVENT_RENAME)) { @@ -360,10 +350,10 @@ xfs_rename( * them when they unlock the inodes. Also, we need to be careful * not to add an inode to the transaction more than once. */ - VN_HOLD(src_dir_vp); + IHOLD(src_dp); xfs_trans_ijoin(tp, src_dp, XFS_ILOCK_EXCL); if (new_parent) { - VN_HOLD(target_dir_vp); + IHOLD(target_dp); xfs_trans_ijoin(tp, target_dp, XFS_ILOCK_EXCL); } if ((src_ip != src_dp) && (src_ip != target_dp)) { diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index 10d2d22eb03..fa694dc5d30 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -3059,10 +3059,9 @@ xfs_symlink( bhv_vname_t *dentry, char *target_path, mode_t mode, - bhv_vnode_t **vpp, + xfs_inode_t **ipp, cred_t *credp) { - bhv_vnode_t *dir_vp = XFS_ITOV(dp); xfs_mount_t *mp = dp->i_mount; xfs_trans_t *tp; xfs_inode_t *ip; @@ -3088,7 +3087,7 @@ xfs_symlink( char *link_name = VNAME(dentry); int link_namelen; - *vpp = NULL; + *ipp = NULL; error = 0; ip = NULL; tp = NULL; @@ -3227,7 +3226,7 @@ xfs_symlink( * transaction cancel unlocking dp so don't do it explicitly in the * error path. */ - VN_HOLD(dir_vp); + IHOLD(dp); xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL); unlock_dp_on_error = B_FALSE; @@ -3343,13 +3342,8 @@ std_return: 0, error, 0); } - if (!error) { - bhv_vnode_t *vp; - - ASSERT(ip); - vp = XFS_ITOV(ip); - *vpp = vp; - } + if (!error) + *ipp = ip; return error; error2: diff --git a/fs/xfs/xfs_vnodeops.h b/fs/xfs/xfs_vnodeops.h index 79c13f57a81..71e9b15276f 100644 --- a/fs/xfs/xfs_vnodeops.h +++ b/fs/xfs/xfs_vnodeops.h @@ -36,7 +36,7 @@ int xfs_rmdir(struct xfs_inode *dp, bhv_vname_t *dentry); int xfs_readdir(struct xfs_inode *dp, void *dirent, size_t bufsize, xfs_off_t *offset, filldir_t filldir); int xfs_symlink(struct xfs_inode *dp, bhv_vname_t *dentry, - char *target_path, mode_t mode, bhv_vnode_t **vpp, + char *target_path, mode_t mode, struct xfs_inode **ipp, struct cred *credp); int xfs_inode_flush(struct xfs_inode *ip, int flags); int xfs_set_dmattrs(struct xfs_inode *ip, u_int evmask, u_int16_t state); @@ -45,7 +45,7 @@ int xfs_change_file_space(struct xfs_inode *ip, int cmd, xfs_flock64_t *bf, xfs_off_t offset, struct cred *credp, int attr_flags); int xfs_rename(struct xfs_inode *src_dp, bhv_vname_t *src_vname, - bhv_vnode_t *target_dir_vp, bhv_vname_t *target_vname); + struct xfs_inode *target_dp, bhv_vname_t *target_vname); int xfs_attr_get(struct xfs_inode *ip, const char *name, char *value, int *valuelenp, int flags, cred_t *cred); int xfs_attr_set(struct xfs_inode *dp, const char *name, char *value, -- cgit v1.2.3 From ef1f5e7ad38e5414d016983a8cc5a8db7654a61d Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 6 Mar 2008 13:46:25 +1100 Subject: [XFS] cleanup vnode use in xfs_lookup SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30550a Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_export.c | 9 ++++----- fs/xfs/linux-2.6/xfs_iops.c | 6 +++--- fs/xfs/xfs_vnodeops.c | 4 ++-- fs/xfs/xfs_vnodeops.h | 2 +- 4 files changed, 10 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_export.c b/fs/xfs/linux-2.6/xfs_export.c index 21f0e825759..66a9a9e76cb 100644 --- a/fs/xfs/linux-2.6/xfs_export.c +++ b/fs/xfs/linux-2.6/xfs_export.c @@ -213,17 +213,16 @@ xfs_fs_get_parent( struct dentry *child) { int error; - bhv_vnode_t *cvp; + struct xfs_inode *cip; struct dentry *parent; - cvp = NULL; - error = xfs_lookup(XFS_I(child->d_inode), &dotdot, &cvp); + error = xfs_lookup(XFS_I(child->d_inode), &dotdot, &cip); if (unlikely(error)) return ERR_PTR(-error); - parent = d_alloc_anon(vn_to_inode(cvp)); + parent = d_alloc_anon(cip->i_vnode); if (unlikely(!parent)) { - VN_RELE(cvp); + iput(cip->i_vnode); return ERR_PTR(-ENOMEM); } return parent; diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c index 215158cbac4..01d9b3f1e04 100644 --- a/fs/xfs/linux-2.6/xfs_iops.c +++ b/fs/xfs/linux-2.6/xfs_iops.c @@ -372,13 +372,13 @@ xfs_vn_lookup( struct dentry *dentry, struct nameidata *nd) { - bhv_vnode_t *cvp; + struct xfs_inode *cip; int error; if (dentry->d_name.len >= MAXNAMELEN) return ERR_PTR(-ENAMETOOLONG); - error = xfs_lookup(XFS_I(dir), dentry, &cvp); + error = xfs_lookup(XFS_I(dir), dentry, &cip); if (unlikely(error)) { if (unlikely(error != ENOENT)) return ERR_PTR(-error); @@ -386,7 +386,7 @@ xfs_vn_lookup( return NULL; } - return d_splice_alias(vn_to_inode(cvp), dentry); + return d_splice_alias(cip->i_vnode, dentry); } STATIC int diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index fa694dc5d30..3418c94bcf1 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -1763,7 +1763,7 @@ int xfs_lookup( xfs_inode_t *dp, bhv_vname_t *dentry, - bhv_vnode_t **vpp) + xfs_inode_t **ipp) { xfs_inode_t *ip; xfs_ino_t e_inum; @@ -1778,7 +1778,7 @@ xfs_lookup( lock_mode = xfs_ilock_map_shared(dp); error = xfs_dir_lookup_int(dp, lock_mode, dentry, &e_inum, &ip); if (!error) { - *vpp = XFS_ITOV(ip); + *ipp = ip; xfs_itrace_ref(ip); } xfs_iunlock_map_shared(dp, lock_mode); diff --git a/fs/xfs/xfs_vnodeops.h b/fs/xfs/xfs_vnodeops.h index 71e9b15276f..12e581865bd 100644 --- a/fs/xfs/xfs_vnodeops.h +++ b/fs/xfs/xfs_vnodeops.h @@ -24,7 +24,7 @@ int xfs_fsync(struct xfs_inode *ip, int flag, xfs_off_t start, int xfs_release(struct xfs_inode *ip); int xfs_inactive(struct xfs_inode *ip); int xfs_lookup(struct xfs_inode *dp, bhv_vname_t *dentry, - bhv_vnode_t **vpp); + struct xfs_inode **ipp); int xfs_create(struct xfs_inode *dp, bhv_vname_t *dentry, mode_t mode, xfs_dev_t rdev, struct xfs_inode **ipp, struct cred *credp); int xfs_remove(struct xfs_inode *dp, bhv_vname_t *dentry); -- cgit v1.2.3 From dcf49cc5cfbbc0070ad4307428f8282dc7e04e58 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 6 Mar 2008 13:46:37 +1100 Subject: [XFS] cleanup vnode use in xfs_lrw.c SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30551a Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_lrw.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_lrw.c b/fs/xfs/linux-2.6/xfs_lrw.c index 01a8f26e1b1..1d95dca96cf 100644 --- a/fs/xfs/linux-2.6/xfs_lrw.c +++ b/fs/xfs/linux-2.6/xfs_lrw.c @@ -176,7 +176,6 @@ xfs_read( { struct file *file = iocb->ki_filp; struct inode *inode = file->f_mapping->host; - bhv_vnode_t *vp = XFS_ITOV(ip); xfs_mount_t *mp = ip->i_mount; size_t size = 0; ssize_t ret = 0; @@ -242,7 +241,7 @@ xfs_read( } if (unlikely(ioflags & IO_ISDIRECT)) { - if (VN_CACHED(vp)) + if (inode->i_mapping->nrpages) ret = xfs_flushinval_pages(ip, (*offset & PAGE_CACHE_MASK), -1, FI_REMAPF_LOCKED); mutex_unlock(&inode->i_mutex); @@ -571,7 +570,6 @@ xfs_write( struct file *file = iocb->ki_filp; struct address_space *mapping = file->f_mapping; struct inode *inode = mapping->host; - bhv_vnode_t *vp = XFS_ITOV(xip); unsigned long segs = nsegs; xfs_mount_t *mp; ssize_t ret = 0, error = 0; @@ -658,7 +656,7 @@ start: return XFS_ERROR(-EINVAL); } - if (!need_i_mutex && (VN_CACHED(vp) || pos > xip->i_size)) { + if (!need_i_mutex && (mapping->nrpages || pos > xip->i_size)) { xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock); iolock = XFS_IOLOCK_EXCL; need_i_mutex = 1; @@ -720,7 +718,7 @@ retry: current->backing_dev_info = mapping->backing_dev_info; if ((ioflags & IO_ISDIRECT)) { - if (VN_CACHED(vp)) { + if (mapping->nrpages) { WARN_ON(need_i_mutex == 0); xfs_inval_cached_trace(xip, pos, -1, (pos & PAGE_CACHE_MASK), -1); -- cgit v1.2.3 From af048193fcfe2650e7ed3b1ab3d48b1ed0efb467 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 6 Mar 2008 13:46:43 +1100 Subject: [XFS] cleanup vnode use in xfs_iops.c SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30552a Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_iops.c | 34 ++++++++++++++++------------------ 1 file changed, 16 insertions(+), 18 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c index 01d9b3f1e04..53f8feb28e5 100644 --- a/fs/xfs/linux-2.6/xfs_iops.c +++ b/fs/xfs/linux-2.6/xfs_iops.c @@ -62,12 +62,11 @@ void xfs_synchronize_atime( xfs_inode_t *ip) { - bhv_vnode_t *vp; + struct inode *inode = ip->i_vnode; - vp = XFS_ITOV_NULL(ip); - if (vp) { - ip->i_d.di_atime.t_sec = (__int32_t)vp->i_atime.tv_sec; - ip->i_d.di_atime.t_nsec = (__int32_t)vp->i_atime.tv_nsec; + if (inode) { + ip->i_d.di_atime.t_sec = (__int32_t)inode->i_atime.tv_sec; + ip->i_d.di_atime.t_nsec = (__int32_t)inode->i_atime.tv_nsec; } } @@ -80,11 +79,10 @@ void xfs_mark_inode_dirty_sync( xfs_inode_t *ip) { - bhv_vnode_t *vp; + struct inode *inode = ip->i_vnode; - vp = XFS_ITOV_NULL(ip); - if (vp) - mark_inode_dirty_sync(vn_to_inode(vp)); + if (inode) + mark_inode_dirty_sync(inode); } /* @@ -215,26 +213,26 @@ xfs_validate_fields( */ STATIC int xfs_init_security( - bhv_vnode_t *vp, + struct inode *inode, struct inode *dir) { - struct inode *ip = vn_to_inode(vp); + struct xfs_inode *ip = XFS_I(inode); size_t length; void *value; char *name; int error; - error = security_inode_init_security(ip, dir, &name, &value, &length); + error = security_inode_init_security(inode, dir, &name, + &value, &length); if (error) { if (error == -EOPNOTSUPP) return 0; return -error; } - error = xfs_attr_set(XFS_I(ip), name, value, - length, ATTR_SECURE); + error = xfs_attr_set(ip, name, value, length, ATTR_SECURE); if (!error) - xfs_iflags_set(XFS_I(ip), XFS_IMODIFIED); + xfs_iflags_set(ip, XFS_IMODIFIED); kfree(name); kfree(value); @@ -244,7 +242,7 @@ xfs_init_security( STATIC void xfs_cleanup_inode( struct inode *dir, - bhv_vnode_t *vp, + struct inode *inode, struct dentry *dentry, int mode) { @@ -255,14 +253,14 @@ xfs_cleanup_inode( * xfs_init_security we must back out. * ENOSPC can hit here, among other things. */ - teardown.d_inode = vn_to_inode(vp); + teardown.d_inode = inode; teardown.d_name = dentry->d_name; if (S_ISDIR(mode)) xfs_rmdir(XFS_I(dir), &teardown); else xfs_remove(XFS_I(dir), &teardown); - VN_RELE(vp); + iput(inode); } STATIC int -- cgit v1.2.3 From 5f90150abad61b49dbb4a6ca1087fe0a75001ef9 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 6 Mar 2008 13:46:49 +1100 Subject: [XFS] cleanup vnode use in xfs_bmap.c SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30553a Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_bmap.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 19aae13b7f9..bce8e3bd8ad 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -5772,7 +5772,6 @@ xfs_getbmap( int error; /* return value */ __int64_t fixlen; /* length for -1 case */ int i; /* extent number */ - bhv_vnode_t *vp; /* corresponding vnode */ int lock; /* lock state */ xfs_bmbt_irec_t *map; /* buffer for user's data */ xfs_mount_t *mp; /* file system mount point */ @@ -5789,7 +5788,6 @@ xfs_getbmap( int bmapi_flags; /* flags for xfs_bmapi */ __int32_t oflags; /* getbmapx bmv_oflags field */ - vp = XFS_ITOV(ip); mp = ip->i_mount; whichfork = interface & BMV_IF_ATTRFORK ? XFS_ATTR_FORK : XFS_DATA_FORK; -- cgit v1.2.3 From dfa18b117974d7667a2d5b941853fac3f2e256db Mon Sep 17 00:00:00 2001 From: Niv Sardi Date: Thu, 6 Mar 2008 13:49:26 +1100 Subject: [XFS] kill t_sema member of struct xfs_trans It's completely unused so we might aswell kill it. Note that there is another t_sema in struct xlog_ticket, which is used and actually an sv_t despite the name. That one is left untouched by this patch. SGI-PV: 971186 SGI-Modid: xfs-linux-melb:xfs-kern:30591a Signed-off-by: Niv Sardi Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_trans.h | 1 - 1 file changed, 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h index 7f40628d85c..b5effce0008 100644 --- a/fs/xfs/xfs_trans.h +++ b/fs/xfs/xfs_trans.h @@ -341,7 +341,6 @@ typedef struct xfs_trans { unsigned int t_rtx_res; /* # of rt extents resvd */ unsigned int t_rtx_res_used; /* # of resvd rt extents used */ xfs_log_ticket_t t_ticket; /* log mgr ticket */ - sema_t t_sema; /* sema for commit completion */ xfs_lsn_t t_lsn; /* log seq num of start of * transaction. */ xfs_lsn_t t_commit_lsn; /* log seq num of end of -- cgit v1.2.3 From a45c796867df8dabc8eed6e72898d7ba1609bd7e Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 6 Mar 2008 13:49:36 +1100 Subject: [XFS] Remove superflous xfs_readsb call in xfs_mountfs. When xfs_mountfs is called by xfs_mount xfs_readsb was called 35 lines above unconditionally, so there is no need to try to read the superblock if it's not present. If any other port doesn't have the superblock read at this point it should just call it directly from it's xfs_mount equivalent. SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30603a Signed-off-by: Christoph Hellwig Signed-off-by: Donald Douwsma Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_mount.c | 5 ----- 1 file changed, 5 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 8ed164eb954..41b690e6896 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -964,11 +964,6 @@ xfs_mountfs( int uuid_mounted = 0; int error = 0; - if (mp->m_sb_bp == NULL) { - error = xfs_readsb(mp, mfsi_flags); - if (error) - return error; - } xfs_mount_common(mp, sbp); /* -- cgit v1.2.3 From 535f6b3735db6ef6026537bfe55ae00c3d9cc1ee Mon Sep 17 00:00:00 2001 From: Josef 'Jeff' Sipek Date: Thu, 27 Mar 2008 17:58:27 +1100 Subject: [XFS] Replace custom AIL linked-list code with struct list_head Replace the xfs_ail_entry_t with a struct list_head and clean the surrounding code up. Also fixes a livelock in xfs_trans_first_push_ail() by terminating the loop at the head of the list correctly. SGI-PV: 978682 SGI-Modid: xfs-linux-melb:xfs-kern:30636a Signed-off-by: Josef 'Jeff' Sipek Signed-off-by: David Chinner Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_mount.h | 2 +- fs/xfs/xfs_trans.h | 7 +-- fs/xfs/xfs_trans_ail.c | 149 ++++++++++++++++++++----------------------------- 3 files changed, 62 insertions(+), 96 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index 7b37fa00929..77b39f66cea 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -220,7 +220,7 @@ extern void xfs_icsb_sync_counters_flags(struct xfs_mount *, int); #endif typedef struct xfs_ail { - xfs_ail_entry_t xa_ail; + struct list_head xa_ail; uint xa_gen; struct task_struct *xa_task; xfs_lsn_t xa_target; diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h index b5effce0008..0804207c739 100644 --- a/fs/xfs/xfs_trans.h +++ b/fs/xfs/xfs_trans.h @@ -113,13 +113,8 @@ struct xfs_mount; struct xfs_trans; struct xfs_dquot_acct; -typedef struct xfs_ail_entry { - struct xfs_log_item *ail_forw; /* AIL forw pointer */ - struct xfs_log_item *ail_back; /* AIL back pointer */ -} xfs_ail_entry_t; - typedef struct xfs_log_item { - xfs_ail_entry_t li_ail; /* AIL pointers */ + struct list_head li_ail; /* AIL pointers */ xfs_lsn_t li_lsn; /* last on-disk lsn */ struct xfs_log_item_desc *li_desc; /* ptr to current desc*/ struct xfs_mount *li_mountp; /* ptr to fs mount */ diff --git a/fs/xfs/xfs_trans_ail.c b/fs/xfs/xfs_trans_ail.c index 76d470d8a1e..13235ae9a58 100644 --- a/fs/xfs/xfs_trans_ail.c +++ b/fs/xfs/xfs_trans_ail.c @@ -28,13 +28,13 @@ #include "xfs_trans_priv.h" #include "xfs_error.h" -STATIC void xfs_ail_insert(xfs_ail_entry_t *, xfs_log_item_t *); -STATIC xfs_log_item_t * xfs_ail_delete(xfs_ail_entry_t *, xfs_log_item_t *); -STATIC xfs_log_item_t * xfs_ail_min(xfs_ail_entry_t *); -STATIC xfs_log_item_t * xfs_ail_next(xfs_ail_entry_t *, xfs_log_item_t *); +STATIC void xfs_ail_insert(xfs_ail_t *, xfs_log_item_t *); +STATIC xfs_log_item_t * xfs_ail_delete(xfs_ail_t *, xfs_log_item_t *); +STATIC xfs_log_item_t * xfs_ail_min(xfs_ail_t *); +STATIC xfs_log_item_t * xfs_ail_next(xfs_ail_t *, xfs_log_item_t *); #ifdef DEBUG -STATIC void xfs_ail_check(xfs_ail_entry_t *, xfs_log_item_t *); +STATIC void xfs_ail_check(xfs_ail_t *, xfs_log_item_t *); #else #define xfs_ail_check(a,l) #endif /* DEBUG */ @@ -57,7 +57,7 @@ xfs_trans_tail_ail( xfs_log_item_t *lip; spin_lock(&mp->m_ail_lock); - lip = xfs_ail_min(&(mp->m_ail.xa_ail)); + lip = xfs_ail_min(&mp->m_ail); if (lip == NULL) { lsn = (xfs_lsn_t)0; } else { @@ -91,7 +91,7 @@ xfs_trans_push_ail( { xfs_log_item_t *lip; - lip = xfs_ail_min(&mp->m_ail.xa_ail); + lip = xfs_ail_min(&mp->m_ail); if (lip && !XFS_FORCED_SHUTDOWN(mp)) { if (XFS_LSN_CMP(threshold_lsn, mp->m_ail.xa_target) > 0) xfsaild_wakeup(mp, threshold_lsn); @@ -111,15 +111,17 @@ xfs_trans_first_push_ail( { xfs_log_item_t *lip; - lip = xfs_ail_min(&(mp->m_ail.xa_ail)); + lip = xfs_ail_min(&mp->m_ail); *gen = (int)mp->m_ail.xa_gen; if (lsn == 0) return lip; - while (lip && (XFS_LSN_CMP(lip->li_lsn, lsn) < 0)) - lip = lip->li_ail.ail_forw; + list_for_each_entry(lip, &mp->m_ail.xa_ail, li_ail) { + if (XFS_LSN_CMP(lip->li_lsn, lsn) >= 0) + return lip; + } - return lip; + return NULL; } /* @@ -329,7 +331,7 @@ xfs_trans_unlocked_item( * the call to xfs_log_move_tail() doesn't do anything if there's * not enough free space to wake people up so we're safe calling it. */ - min_lip = xfs_ail_min(&mp->m_ail.xa_ail); + min_lip = xfs_ail_min(&mp->m_ail); if (min_lip == lip) xfs_log_move_tail(mp, 1); @@ -357,15 +359,13 @@ xfs_trans_update_ail( xfs_log_item_t *lip, xfs_lsn_t lsn) __releases(mp->m_ail_lock) { - xfs_ail_entry_t *ailp; xfs_log_item_t *dlip=NULL; xfs_log_item_t *mlip; /* ptr to minimum lip */ - ailp = &(mp->m_ail.xa_ail); - mlip = xfs_ail_min(ailp); + mlip = xfs_ail_min(&mp->m_ail); if (lip->li_flags & XFS_LI_IN_AIL) { - dlip = xfs_ail_delete(ailp, lip); + dlip = xfs_ail_delete(&mp->m_ail, lip); ASSERT(dlip == lip); } else { lip->li_flags |= XFS_LI_IN_AIL; @@ -373,11 +373,11 @@ xfs_trans_update_ail( lip->li_lsn = lsn; - xfs_ail_insert(ailp, lip); + xfs_ail_insert(&mp->m_ail, lip); mp->m_ail.xa_gen++; if (mlip == dlip) { - mlip = xfs_ail_min(&(mp->m_ail.xa_ail)); + mlip = xfs_ail_min(&mp->m_ail); spin_unlock(&mp->m_ail_lock); xfs_log_move_tail(mp, mlip->li_lsn); } else { @@ -407,14 +407,12 @@ xfs_trans_delete_ail( xfs_mount_t *mp, xfs_log_item_t *lip) __releases(mp->m_ail_lock) { - xfs_ail_entry_t *ailp; xfs_log_item_t *dlip; xfs_log_item_t *mlip; if (lip->li_flags & XFS_LI_IN_AIL) { - ailp = &(mp->m_ail.xa_ail); - mlip = xfs_ail_min(ailp); - dlip = xfs_ail_delete(ailp, lip); + mlip = xfs_ail_min(&mp->m_ail); + dlip = xfs_ail_delete(&mp->m_ail, lip); ASSERT(dlip == lip); @@ -423,7 +421,7 @@ xfs_trans_delete_ail( mp->m_ail.xa_gen++; if (mlip == dlip) { - mlip = xfs_ail_min(&(mp->m_ail.xa_ail)); + mlip = xfs_ail_min(&mp->m_ail); spin_unlock(&mp->m_ail_lock); xfs_log_move_tail(mp, (mlip ? mlip->li_lsn : 0)); } else { @@ -461,7 +459,7 @@ xfs_trans_first_ail( { xfs_log_item_t *lip; - lip = xfs_ail_min(&(mp->m_ail.xa_ail)); + lip = xfs_ail_min(&mp->m_ail); *gen = (int)mp->m_ail.xa_gen; return lip; @@ -485,9 +483,9 @@ xfs_trans_next_ail( ASSERT(mp && lip && gen); if (mp->m_ail.xa_gen == *gen) { - nlip = xfs_ail_next(&(mp->m_ail.xa_ail), lip); + nlip = xfs_ail_next(&mp->m_ail, lip); } else { - nlip = xfs_ail_min(&(mp->m_ail).xa_ail); + nlip = xfs_ail_min(&mp->m_ail); *gen = (int)mp->m_ail.xa_gen; if (restarts != NULL) { XFS_STATS_INC(xs_push_ail_restarts); @@ -517,8 +515,7 @@ int xfs_trans_ail_init( xfs_mount_t *mp) { - mp->m_ail.xa_ail.ail_forw = (xfs_log_item_t*)&mp->m_ail.xa_ail; - mp->m_ail.xa_ail.ail_back = (xfs_log_item_t*)&mp->m_ail.xa_ail; + INIT_LIST_HEAD(&mp->m_ail.xa_ail); return xfsaild_start(mp); } @@ -537,7 +534,7 @@ xfs_trans_ail_destroy( */ STATIC void xfs_ail_insert( - xfs_ail_entry_t *base, + xfs_ail_t *ailp, xfs_log_item_t *lip) /* ARGSUSED */ { @@ -546,27 +543,22 @@ xfs_ail_insert( /* * If the list is empty, just insert the item. */ - if (base->ail_back == (xfs_log_item_t*)base) { - base->ail_forw = lip; - base->ail_back = lip; - lip->li_ail.ail_forw = (xfs_log_item_t*)base; - lip->li_ail.ail_back = (xfs_log_item_t*)base; + if (list_empty(&ailp->xa_ail)) { + list_add(&lip->li_ail, &ailp->xa_ail); return; } - next_lip = base->ail_back; - while ((next_lip != (xfs_log_item_t*)base) && - (XFS_LSN_CMP(next_lip->li_lsn, lip->li_lsn) > 0)) { - next_lip = next_lip->li_ail.ail_back; + list_for_each_entry_reverse(next_lip, &ailp->xa_ail, li_ail) { + if (XFS_LSN_CMP(next_lip->li_lsn, lip->li_lsn) <= 0) + break; } - ASSERT((next_lip == (xfs_log_item_t*)base) || + + ASSERT((&next_lip->li_ail == &ailp->xa_ail) || (XFS_LSN_CMP(next_lip->li_lsn, lip->li_lsn) <= 0)); - lip->li_ail.ail_forw = next_lip->li_ail.ail_forw; - lip->li_ail.ail_back = next_lip; - next_lip->li_ail.ail_forw = lip; - lip->li_ail.ail_forw->li_ail.ail_back = lip; - xfs_ail_check(base, lip); + list_add(&lip->li_ail, &next_lip->li_ail); + + xfs_ail_check(ailp, lip); return; } @@ -576,15 +568,13 @@ xfs_ail_insert( /*ARGSUSED*/ STATIC xfs_log_item_t * xfs_ail_delete( - xfs_ail_entry_t *base, + xfs_ail_t *ailp, xfs_log_item_t *lip) /* ARGSUSED */ { - xfs_ail_check(base, lip); - lip->li_ail.ail_forw->li_ail.ail_back = lip->li_ail.ail_back; - lip->li_ail.ail_back->li_ail.ail_forw = lip->li_ail.ail_forw; - lip->li_ail.ail_forw = NULL; - lip->li_ail.ail_back = NULL; + xfs_ail_check(ailp, lip); + + list_del(&lip->li_ail); return lip; } @@ -595,14 +585,13 @@ xfs_ail_delete( */ STATIC xfs_log_item_t * xfs_ail_min( - xfs_ail_entry_t *base) + xfs_ail_t *ailp) /* ARGSUSED */ { - register xfs_log_item_t *forw = base->ail_forw; - if (forw == (xfs_log_item_t*)base) { + if (list_empty(&ailp->xa_ail)) return NULL; - } - return forw; + + return list_first_entry(&ailp->xa_ail, xfs_log_item_t, li_ail); } /* @@ -612,15 +601,14 @@ xfs_ail_min( */ STATIC xfs_log_item_t * xfs_ail_next( - xfs_ail_entry_t *base, + xfs_ail_t *ailp, xfs_log_item_t *lip) /* ARGSUSED */ { - if (lip->li_ail.ail_forw == (xfs_log_item_t*)base) { + if (lip->li_ail.next == &ailp->xa_ail) return NULL; - } - return lip->li_ail.ail_forw; + return list_first_entry(&lip->li_ail, xfs_log_item_t, li_ail); } #ifdef DEBUG @@ -629,57 +617,40 @@ xfs_ail_next( */ STATIC void xfs_ail_check( - xfs_ail_entry_t *base, + xfs_ail_t *ailp, xfs_log_item_t *lip) { xfs_log_item_t *prev_lip; - prev_lip = base->ail_forw; - if (prev_lip == (xfs_log_item_t*)base) { - /* - * Make sure the pointers are correct when the list - * is empty. - */ - ASSERT(base->ail_back == (xfs_log_item_t*)base); + if (list_empty(&ailp->xa_ail)) return; - } /* * Check the next and previous entries are valid. */ ASSERT((lip->li_flags & XFS_LI_IN_AIL) != 0); - prev_lip = lip->li_ail.ail_back; - if (prev_lip != (xfs_log_item_t*)base) { - ASSERT(prev_lip->li_ail.ail_forw == lip); + prev_lip = list_entry(lip->li_ail.prev, xfs_log_item_t, li_ail); + if (&prev_lip->li_ail != &ailp->xa_ail) ASSERT(XFS_LSN_CMP(prev_lip->li_lsn, lip->li_lsn) <= 0); - } - prev_lip = lip->li_ail.ail_forw; - if (prev_lip != (xfs_log_item_t*)base) { - ASSERT(prev_lip->li_ail.ail_back == lip); + + prev_lip = list_entry(lip->li_ail.next, xfs_log_item_t, li_ail); + if (&prev_lip->li_ail != &ailp->xa_ail) ASSERT(XFS_LSN_CMP(prev_lip->li_lsn, lip->li_lsn) >= 0); - } #ifdef XFS_TRANS_DEBUG /* - * Walk the list checking forward and backward pointers, - * lsn ordering, and that every entry has the XFS_LI_IN_AIL - * flag set. This is really expensive, so only do it when - * specifically debugging the transaction subsystem. + * Walk the list checking lsn ordering, and that every entry has the + * XFS_LI_IN_AIL flag set. This is really expensive, so only do it + * when specifically debugging the transaction subsystem. */ - prev_lip = (xfs_log_item_t*)base; - while (lip != (xfs_log_item_t*)base) { - if (prev_lip != (xfs_log_item_t*)base) { - ASSERT(prev_lip->li_ail.ail_forw == lip); + prev_lip = list_entry(&ailp->xa_ail, xfs_log_item_t, li_ail); + list_for_each_entry(lip, &ailp->xa_ail, li_ail) { + if (&prev_lip->li_ail != &ailp->xa_ail) ASSERT(XFS_LSN_CMP(prev_lip->li_lsn, lip->li_lsn) <= 0); - } - ASSERT(lip->li_ail.ail_back == prev_lip); ASSERT((lip->li_flags & XFS_LI_IN_AIL) != 0); prev_lip = lip; - lip = lip->li_ail.ail_forw; } - ASSERT(lip == (xfs_log_item_t*)base); - ASSERT(base->ail_back == prev_lip); #endif /* XFS_TRANS_DEBUG */ } #endif /* DEBUG */ -- cgit v1.2.3 From 75de2a91c98a6f486f261c1367fe59f5583e15a3 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 27 Mar 2008 18:00:38 +1100 Subject: [XFS] Account for inode cluster alignment in all allocations At ENOSPC, we can get a filesystem shutdown due to a cancelling a dirty transaction in xfs_mkdir or xfs_create. This is due to the initial allocation attempt not taking into account inode alignment and hence we can prepare the AGF freelist for allocation when it's not actually possible to do an allocation. This results in inode allocation returning ENOSPC with a dirty transaction, and hence we shut down the filesystem. Because the first allocation is an exact allocation attempt, we must tell the allocator that the alignment does not affect the allocation attempt. i.e. we will accept any extent alignment as long as the extent starts at the block we want. Unfortunately, this means that if the longest free extent is less than the length + alignment necessary for fallback allocation attempts but is long enough to attempt a non-aligned allocation, we will modify the free list. If we then have the exact allocation fail, all other allocation attempts will also fail due to the alignment constraint being taken into account. Hence the initial attempt needs to set the "alignment slop" field so that alignment, while not required, must be taken into account when determining if there is enough space left in the AG to do the allocation. That means if the exact allocation fails, we will not dirty the freelist if there is not enough space available fo a subsequent allocation to succeed. Hence we get an ENOSPC error back to userspace without shutting down the filesystem. SGI-PV: 978886 SGI-Modid: xfs-linux-melb:xfs-kern:30699a Signed-off-by: David Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_ialloc.c | 44 +++++++++++++++++++++++++++++--------------- 1 file changed, 29 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_ialloc.c b/fs/xfs/xfs_ialloc.c index 5a146cb2298..a64dfbd565a 100644 --- a/fs/xfs/xfs_ialloc.c +++ b/fs/xfs/xfs_ialloc.c @@ -107,6 +107,16 @@ xfs_ialloc_log_di( /* * Allocation group level functions. */ +static inline int +xfs_ialloc_cluster_alignment( + xfs_alloc_arg_t *args) +{ + if (xfs_sb_version_hasalign(&args->mp->m_sb) && + args->mp->m_sb.sb_inoalignmt >= + XFS_B_TO_FSBT(args->mp, XFS_INODE_CLUSTER_SIZE(args->mp))) + return args->mp->m_sb.sb_inoalignmt; + return 1; +} /* * Allocate new inodes in the allocation group specified by agbp. @@ -167,10 +177,24 @@ xfs_ialloc_ag_alloc( args.mod = args.total = args.wasdel = args.isfl = args.userdata = args.minalignslop = 0; args.prod = 1; - args.alignment = 1; + /* - * Allow space for the inode btree to split. + * We need to take into account alignment here to ensure that + * we don't modify the free list if we fail to have an exact + * block. If we don't have an exact match, and every oher + * attempt allocation attempt fails, we'll end up cancelling + * a dirty transaction and shutting down. + * + * For an exact allocation, alignment must be 1, + * however we need to take cluster alignment into account when + * fixing up the freelist. Use the minalignslop field to + * indicate that extra blocks might be required for alignment, + * but not to use them in the actual exact allocation. */ + args.alignment = 1; + args.minalignslop = xfs_ialloc_cluster_alignment(&args) - 1; + + /* Allow space for the inode btree to split. */ args.minleft = XFS_IN_MAXLEVELS(args.mp) - 1; if ((error = xfs_alloc_vextent(&args))) return error; @@ -191,13 +215,8 @@ xfs_ialloc_ag_alloc( ASSERT(!(args.mp->m_flags & XFS_MOUNT_NOALIGN)); args.alignment = args.mp->m_dalign; isaligned = 1; - } else if (xfs_sb_version_hasalign(&args.mp->m_sb) && - args.mp->m_sb.sb_inoalignmt >= - XFS_B_TO_FSBT(args.mp, - XFS_INODE_CLUSTER_SIZE(args.mp))) - args.alignment = args.mp->m_sb.sb_inoalignmt; - else - args.alignment = 1; + } else + args.alignment = xfs_ialloc_cluster_alignment(&args); /* * Need to figure out where to allocate the inode blocks. * Ideally they should be spaced out through the a.g. @@ -230,12 +249,7 @@ xfs_ialloc_ag_alloc( args.agbno = be32_to_cpu(agi->agi_root); args.fsbno = XFS_AGB_TO_FSB(args.mp, be32_to_cpu(agi->agi_seqno), args.agbno); - if (xfs_sb_version_hasalign(&args.mp->m_sb) && - args.mp->m_sb.sb_inoalignmt >= - XFS_B_TO_FSBT(args.mp, XFS_INODE_CLUSTER_SIZE(args.mp))) - args.alignment = args.mp->m_sb.sb_inoalignmt; - else - args.alignment = 1; + args.alignment = xfs_ialloc_cluster_alignment(&args); if ((error = xfs_alloc_vextent(&args))) return error; } -- cgit v1.2.3 From 59a33f9f776b051018ec98af95bd9fe8ba9d0f3e Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 27 Mar 2008 18:00:45 +1100 Subject: [XFS] Ensure a btree insert returns a valid cursor. When writing into preallocated regions there is a case where XFS can oops or hang doing the unwritten extent conversion on I/O completion. It turns out that the problem is related to the btree cursor being invalid. When we do an insert into the tree, we may need to split blocks in the tree. When we only split at the leaf level (i.e. level 0), everything works just fine. However, if we have a multi-level split in the btreee, the cursor passed to the insert function is no longer valid once the insert is complete. The leaf level split is handled correctly because all the operations at level 0 are done using the original cursor, hence it is updated correctly. However, when we need to update the next level up the tree, we don't use that cursor - we use a cloned cursor that points to the index in the next level up where we need to do the insert. Hence if we need to split a second level, the changes to the tree are reflected in the cloned cursor and not the original cursor. This clone-and-move-up-a-level-on-split behaviour recurses all the way to the top of the tree. The complexity here is that these cloned cursors do not point to the original index that was inserted - they point to the newly allocated block (the right block) and the original cursor pointer to that level may still point to the left block. Hence, without deep examination of the cloned cursor and buffers, we cannot update the original cursor with the new path from the cloned cursor. In these cases the original cursor could be pointing to the wrong block(s) and hence a subsequent modification to the tree using that cursor will lead to corruption of the tree. The crash case occurs when the tree changes height - we insert a new level in the tree, and the cursor does not have a buffer in it's path for that level. Hence any attempt to walk back up the cursor to the root block will result in a null pointer dereference. To make matters even more complex, the BMAP BT is rooted in an inode, so we can have a change of height in the btree *without a root split*. That is, if the root block in the inode is full when we split a leaf node, we cannot fit the pointer to the new block in the root, so we allocate a new block, migrate all the ptrs out of the inode into the new block and point the inode root block at the newly allocated block. This changes the height of the tree without a root split having occurred and hence invalidates the path in the original cursor. The patch below prevents xfs_bmbt_insert() from returning with an invalid cursor by detecting the cases that invalidate the original cursor and refresh it by do a lookup into the btree for the original index we were inserting at. Note that the INOBT, AGFBNO and AGFCNT btree implementations also have this bug, but the cursor is currently always destroyed or revalidated after an insert for those trees. Hence this patch only address the problem in the BMBT code. SGI-PV: 979339 SGI-Modid: xfs-linux-melb:xfs-kern:30701a Signed-off-by: David Chinner Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_bmap_btree.c | 38 ++++++++++++++++++++++++++++++++++++-- 1 file changed, 36 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_bmap_btree.c b/fs/xfs/xfs_bmap_btree.c index bd18987326a..93470b728dd 100644 --- a/fs/xfs/xfs_bmap_btree.c +++ b/fs/xfs/xfs_bmap_btree.c @@ -2027,6 +2027,24 @@ xfs_bmbt_increment( /* * Insert the current record at the point referenced by cur. + * + * A multi-level split of the tree on insert will invalidate the original + * cursor. It appears, however, that some callers assume that the cursor is + * always valid. Hence if we do a multi-level split we need to revalidate the + * cursor. + * + * When a split occurs, we will see a new cursor returned. Use that as a + * trigger to determine if we need to revalidate the original cursor. If we get + * a split, then use the original irec to lookup up the path of the record we + * just inserted. + * + * Note that the fact that the btree root is in the inode means that we can + * have the level of the tree change without a "split" occurring at the root + * level. What happens is that the root is migrated to an allocated block and + * the inode root is pointed to it. This means a single split can change the + * level of the tree (level 2 -> level 3) and invalidate the old cursor. Hence + * the level change should be accounted as a split so as to correctly trigger a + * revalidation of the old cursor. */ int /* error */ xfs_bmbt_insert( @@ -2039,11 +2057,14 @@ xfs_bmbt_insert( xfs_fsblock_t nbno; xfs_btree_cur_t *ncur; xfs_bmbt_rec_t nrec; + xfs_bmbt_irec_t oirec; /* original irec */ xfs_btree_cur_t *pcur; + int splits = 0; XFS_BMBT_TRACE_CURSOR(cur, ENTRY); level = 0; nbno = NULLFSBLOCK; + oirec = cur->bc_rec.b; xfs_bmbt_disk_set_all(&nrec, &cur->bc_rec.b); ncur = NULL; pcur = cur; @@ -2052,11 +2073,13 @@ xfs_bmbt_insert( &i))) { if (pcur != cur) xfs_btree_del_cursor(pcur, XFS_BTREE_ERROR); - XFS_BMBT_TRACE_CURSOR(cur, ERROR); - return error; + goto error0; } XFS_WANT_CORRUPTED_GOTO(i == 1, error0); if (pcur != cur && (ncur || nbno == NULLFSBLOCK)) { + /* allocating a new root is effectively a split */ + if (cur->bc_nlevels != pcur->bc_nlevels) + splits++; cur->bc_nlevels = pcur->bc_nlevels; cur->bc_private.b.allocated += pcur->bc_private.b.allocated; @@ -2070,10 +2093,21 @@ xfs_bmbt_insert( xfs_btree_del_cursor(pcur, XFS_BTREE_NOERROR); } if (ncur) { + splits++; pcur = ncur; ncur = NULL; } } while (nbno != NULLFSBLOCK); + + if (splits > 1) { + /* revalidate the old cursor as we had a multi-level split */ + error = xfs_bmbt_lookup_eq(cur, oirec.br_startoff, + oirec.br_startblock, oirec.br_blockcount, &i); + if (error) + goto error0; + ASSERT(i == 1); + } + XFS_BMBT_TRACE_CURSOR(cur, EXIT); *stat = i; return 0; -- cgit v1.2.3 From f3dcc13f6fa20af1171eac7a537a4b89b1a84849 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 27 Mar 2008 18:00:54 +1100 Subject: [XFS] cleanup root inode handling in xfs_fs_fill_super - rename rootvp to root for clarify - remove useless vn_to_inode call - check is_bad_inode before calling d_alloc_root - use iput instead of VN_RELE in the error case SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30708a Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_super.c | 19 +++++++++---------- 1 file changed, 9 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c index cb9ce90d1de..72e55db948d 100644 --- a/fs/xfs/linux-2.6/xfs_super.c +++ b/fs/xfs/linux-2.6/xfs_super.c @@ -1307,7 +1307,7 @@ xfs_fs_fill_super( void *data, int silent) { - struct inode *rootvp; + struct inode *root; struct xfs_mount *mp = NULL; struct xfs_mount_args *args = xfs_args_allocate(sb, silent); int error; @@ -1345,19 +1345,18 @@ xfs_fs_fill_super( sb->s_time_gran = 1; set_posix_acl_flag(sb); - rootvp = igrab(mp->m_rootip->i_vnode); - if (!rootvp) { + root = igrab(mp->m_rootip->i_vnode); + if (!root) { error = ENOENT; goto fail_unmount; } - - sb->s_root = d_alloc_root(vn_to_inode(rootvp)); - if (!sb->s_root) { - error = ENOMEM; + if (is_bad_inode(root)) { + error = EINVAL; goto fail_vnrele; } - if (is_bad_inode(sb->s_root->d_inode)) { - error = EINVAL; + sb->s_root = d_alloc_root(root); + if (!sb->s_root) { + error = ENOMEM; goto fail_vnrele; } @@ -1379,7 +1378,7 @@ fail_vnrele: dput(sb->s_root); sb->s_root = NULL; } else { - VN_RELE(rootvp); + iput(root); } fail_unmount: -- cgit v1.2.3 From df26cfe849d8fd767b26fcd4bfebfff67bda9f3a Mon Sep 17 00:00:00 2001 From: Lachlan McIlroy Date: Fri, 18 Apr 2008 11:44:03 +1000 Subject: [XFS] split xfs_ioc_xattr The three subcases of xfs_ioc_xattr don't share any semantics and almost no code, so split it into three separate helpers. SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30709a Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_ioctl.c | 368 ++++++++++++++++++++++++++++++++++++------- 1 file changed, 312 insertions(+), 56 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_ioctl.c b/fs/xfs/linux-2.6/xfs_ioctl.c index f34bd010eb5..c6399b2cf17 100644 --- a/fs/xfs/linux-2.6/xfs_ioctl.c +++ b/fs/xfs/linux-2.6/xfs_ioctl.c @@ -1179,85 +1179,85 @@ xfs_ioc_fsgetxattr( } STATIC int -xfs_ioc_xattr( +xfs_ioc_fssetxattr( xfs_inode_t *ip, struct file *filp, - unsigned int cmd, void __user *arg) { struct fsxattr fa; struct bhv_vattr *vattr; - int error = 0; + int error; int attr_flags; - unsigned int flags; + + if (copy_from_user(&fa, arg, sizeof(fa))) + return -EFAULT; vattr = kmalloc(sizeof(*vattr), GFP_KERNEL); if (unlikely(!vattr)) return -ENOMEM; - switch (cmd) { - case XFS_IOC_FSSETXATTR: { - if (copy_from_user(&fa, arg, sizeof(fa))) { - error = -EFAULT; - break; - } + attr_flags = 0; + if (filp->f_flags & (O_NDELAY|O_NONBLOCK)) + attr_flags |= ATTR_NONBLOCK; - attr_flags = 0; - if (filp->f_flags & (O_NDELAY|O_NONBLOCK)) - attr_flags |= ATTR_NONBLOCK; + vattr->va_mask = XFS_AT_XFLAGS | XFS_AT_EXTSIZE | XFS_AT_PROJID; + vattr->va_xflags = fa.fsx_xflags; + vattr->va_extsize = fa.fsx_extsize; + vattr->va_projid = fa.fsx_projid; - vattr->va_mask = XFS_AT_XFLAGS | XFS_AT_EXTSIZE | XFS_AT_PROJID; - vattr->va_xflags = fa.fsx_xflags; - vattr->va_extsize = fa.fsx_extsize; - vattr->va_projid = fa.fsx_projid; + error = -xfs_setattr(ip, vattr, attr_flags, NULL); + if (!error) + vn_revalidate(XFS_ITOV(ip)); /* update flags */ + kfree(vattr); + return 0; +} - error = xfs_setattr(ip, vattr, attr_flags, NULL); - if (likely(!error)) - vn_revalidate(XFS_ITOV(ip)); /* update flags */ - error = -error; - break; - } +STATIC int +xfs_ioc_getxflags( + xfs_inode_t *ip, + void __user *arg) +{ + unsigned int flags; - case XFS_IOC_GETXFLAGS: { - flags = xfs_di2lxflags(ip->i_d.di_flags); - if (copy_to_user(arg, &flags, sizeof(flags))) - error = -EFAULT; - break; - } + flags = xfs_di2lxflags(ip->i_d.di_flags); + if (copy_to_user(arg, &flags, sizeof(flags))) + return -EFAULT; + return 0; +} - case XFS_IOC_SETXFLAGS: { - if (copy_from_user(&flags, arg, sizeof(flags))) { - error = -EFAULT; - break; - } +STATIC int +xfs_ioc_setxflags( + xfs_inode_t *ip, + struct file *filp, + void __user *arg) +{ + struct bhv_vattr *vattr; + unsigned int flags; + int attr_flags; + int error; - if (flags & ~(FS_IMMUTABLE_FL | FS_APPEND_FL | \ - FS_NOATIME_FL | FS_NODUMP_FL | \ - FS_SYNC_FL)) { - error = -EOPNOTSUPP; - break; - } + if (copy_from_user(&flags, arg, sizeof(flags))) + return -EFAULT; - attr_flags = 0; - if (filp->f_flags & (O_NDELAY|O_NONBLOCK)) - attr_flags |= ATTR_NONBLOCK; + if (flags & ~(FS_IMMUTABLE_FL | FS_APPEND_FL | \ + FS_NOATIME_FL | FS_NODUMP_FL | \ + FS_SYNC_FL)) + return -EOPNOTSUPP; - vattr->va_mask = XFS_AT_XFLAGS; - vattr->va_xflags = xfs_merge_ioc_xflags(flags, - xfs_ip2xflags(ip)); + vattr = kmalloc(sizeof(*vattr), GFP_KERNEL); + if (unlikely(!vattr)) + return -ENOMEM; - error = xfs_setattr(ip, vattr, attr_flags, NULL); - if (likely(!error)) - vn_revalidate(XFS_ITOV(ip)); /* update flags */ - error = -error; - break; - } + attr_flags = 0; + if (filp->f_flags & (O_NDELAY|O_NONBLOCK)) + attr_flags |= ATTR_NONBLOCK; - default: - error = -ENOTTY; - break; - } + vattr->va_mask = XFS_AT_XFLAGS; + vattr->va_xflags = xfs_merge_ioc_xflags(flags, xfs_ip2xflags(ip)); + error = -xfs_setattr(ip, vattr, attr_flags, NULL); + if (likely(!error)) + vn_revalidate(XFS_ITOV(ip)); /* update flags */ kfree(vattr); return error; } @@ -1332,3 +1332,259 @@ xfs_ioc_getbmapx( return 0; } + +int +xfs_ioctl( + xfs_inode_t *ip, + struct file *filp, + int ioflags, + unsigned int cmd, + void __user *arg) +{ + struct inode *inode = filp->f_path.dentry->d_inode; + xfs_mount_t *mp = ip->i_mount; + int error; + + xfs_itrace_entry(XFS_I(inode)); + switch (cmd) { + + case XFS_IOC_ALLOCSP: + case XFS_IOC_FREESP: + case XFS_IOC_RESVSP: + case XFS_IOC_UNRESVSP: + case XFS_IOC_ALLOCSP64: + case XFS_IOC_FREESP64: + case XFS_IOC_RESVSP64: + case XFS_IOC_UNRESVSP64: + /* + * Only allow the sys admin to reserve space unless + * unwritten extents are enabled. + */ + if (!xfs_sb_version_hasextflgbit(&mp->m_sb) && + !capable(CAP_SYS_ADMIN)) + return -EPERM; + + return xfs_ioc_space(ip, inode, filp, ioflags, cmd, arg); + + case XFS_IOC_DIOINFO: { + struct dioattr da; + xfs_buftarg_t *target = + XFS_IS_REALTIME_INODE(ip) ? + mp->m_rtdev_targp : mp->m_ddev_targp; + + da.d_mem = da.d_miniosz = 1 << target->bt_sshift; + da.d_maxiosz = INT_MAX & ~(da.d_miniosz - 1); + + if (copy_to_user(arg, &da, sizeof(da))) + return -XFS_ERROR(EFAULT); + return 0; + } + + case XFS_IOC_FSBULKSTAT_SINGLE: + case XFS_IOC_FSBULKSTAT: + case XFS_IOC_FSINUMBERS: + return xfs_ioc_bulkstat(mp, cmd, arg); + + case XFS_IOC_FSGEOMETRY_V1: + return xfs_ioc_fsgeometry_v1(mp, arg); + + case XFS_IOC_FSGEOMETRY: + return xfs_ioc_fsgeometry(mp, arg); + + case XFS_IOC_GETVERSION: + return put_user(inode->i_generation, (int __user *)arg); + + case XFS_IOC_FSGETXATTR: + return xfs_ioc_fsgetxattr(ip, 0, arg); + case XFS_IOC_FSGETXATTRA: + return xfs_ioc_fsgetxattr(ip, 1, arg); + case XFS_IOC_FSSETXATTR: + return xfs_ioc_fssetxattr(ip, filp, arg); + case XFS_IOC_GETXFLAGS: + return xfs_ioc_getxflags(ip, arg); + case XFS_IOC_SETXFLAGS: + return xfs_ioc_setxflags(ip, filp, arg); + + case XFS_IOC_FSSETDM: { + struct fsdmidata dmi; + + if (copy_from_user(&dmi, arg, sizeof(dmi))) + return -XFS_ERROR(EFAULT); + + error = xfs_set_dmattrs(ip, dmi.fsd_dmevmask, + dmi.fsd_dmstate); + return -error; + } + + case XFS_IOC_GETBMAP: + case XFS_IOC_GETBMAPA: + return xfs_ioc_getbmap(ip, ioflags, cmd, arg); + + case XFS_IOC_GETBMAPX: + return xfs_ioc_getbmapx(ip, arg); + + case XFS_IOC_FD_TO_HANDLE: + case XFS_IOC_PATH_TO_HANDLE: + case XFS_IOC_PATH_TO_FSHANDLE: + return xfs_find_handle(cmd, arg); + + case XFS_IOC_OPEN_BY_HANDLE: + return xfs_open_by_handle(mp, arg, filp, inode); + + case XFS_IOC_FSSETDM_BY_HANDLE: + return xfs_fssetdm_by_handle(mp, arg, inode); + + case XFS_IOC_READLINK_BY_HANDLE: + return xfs_readlink_by_handle(mp, arg, inode); + + case XFS_IOC_ATTRLIST_BY_HANDLE: + return xfs_attrlist_by_handle(mp, arg, inode); + + case XFS_IOC_ATTRMULTI_BY_HANDLE: + return xfs_attrmulti_by_handle(mp, arg, inode); + + case XFS_IOC_SWAPEXT: { + error = xfs_swapext((struct xfs_swapext __user *)arg); + return -error; + } + + case XFS_IOC_FSCOUNTS: { + xfs_fsop_counts_t out; + + error = xfs_fs_counts(mp, &out); + if (error) + return -error; + + if (copy_to_user(arg, &out, sizeof(out))) + return -XFS_ERROR(EFAULT); + return 0; + } + + case XFS_IOC_SET_RESBLKS: { + xfs_fsop_resblks_t inout; + __uint64_t in; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + if (copy_from_user(&inout, arg, sizeof(inout))) + return -XFS_ERROR(EFAULT); + + /* input parameter is passed in resblks field of structure */ + in = inout.resblks; + error = xfs_reserve_blocks(mp, &in, &inout); + if (error) + return -error; + + if (copy_to_user(arg, &inout, sizeof(inout))) + return -XFS_ERROR(EFAULT); + return 0; + } + + case XFS_IOC_GET_RESBLKS: { + xfs_fsop_resblks_t out; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + error = xfs_reserve_blocks(mp, NULL, &out); + if (error) + return -error; + + if (copy_to_user(arg, &out, sizeof(out))) + return -XFS_ERROR(EFAULT); + + return 0; + } + + case XFS_IOC_FSGROWFSDATA: { + xfs_growfs_data_t in; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + if (copy_from_user(&in, arg, sizeof(in))) + return -XFS_ERROR(EFAULT); + + error = xfs_growfs_data(mp, &in); + return -error; + } + + case XFS_IOC_FSGROWFSLOG: { + xfs_growfs_log_t in; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + if (copy_from_user(&in, arg, sizeof(in))) + return -XFS_ERROR(EFAULT); + + error = xfs_growfs_log(mp, &in); + return -error; + } + + case XFS_IOC_FSGROWFSRT: { + xfs_growfs_rt_t in; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + if (copy_from_user(&in, arg, sizeof(in))) + return -XFS_ERROR(EFAULT); + + error = xfs_growfs_rt(mp, &in); + return -error; + } + + case XFS_IOC_FREEZE: + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + if (inode->i_sb->s_frozen == SB_UNFROZEN) + freeze_bdev(inode->i_sb->s_bdev); + return 0; + + case XFS_IOC_THAW: + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + if (inode->i_sb->s_frozen != SB_UNFROZEN) + thaw_bdev(inode->i_sb->s_bdev, inode->i_sb); + return 0; + + case XFS_IOC_GOINGDOWN: { + __uint32_t in; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + if (get_user(in, (__uint32_t __user *)arg)) + return -XFS_ERROR(EFAULT); + + error = xfs_fs_goingdown(mp, in); + return -error; + } + + case XFS_IOC_ERROR_INJECTION: { + xfs_error_injection_t in; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + if (copy_from_user(&in, arg, sizeof(in))) + return -XFS_ERROR(EFAULT); + + error = xfs_errortag_add(in.errtag, mp); + return -error; + } + + case XFS_IOC_ERROR_CLEARALL: + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + error = xfs_errortag_clearall(mp, 1); + return -error; + + default: + return -ENOTTY; + } +} -- cgit v1.2.3 From 433550990e6c2e94995239bac6a52b4df454cae0 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 27 Mar 2008 18:01:08 +1100 Subject: [XFS] remove most calls to VN_RELE Most VN_RELE calls either directly contain a XFS_ITOV or have the corresponding xfs_inode already in scope. Use the IRELE helper instead of VN_RELE to clarify the code. With a little more work we can kill VN_RELE altogether and define IRELE in terms of iput directly. SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30710a Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/quota/xfs_qm.c | 6 +++--- fs/xfs/quota/xfs_qm_syscalls.c | 10 +++++----- fs/xfs/xfs_log_recover.c | 3 ++- fs/xfs/xfs_mount.c | 5 ++--- fs/xfs/xfs_rtalloc.c | 3 ++- fs/xfs/xfs_vfsops.c | 13 +++++++------ 6 files changed, 21 insertions(+), 19 deletions(-) (limited to 'fs') diff --git a/fs/xfs/quota/xfs_qm.c b/fs/xfs/quota/xfs_qm.c index 8e9c5ae6504..adbc7bb9fba 100644 --- a/fs/xfs/quota/xfs_qm.c +++ b/fs/xfs/quota/xfs_qm.c @@ -1810,7 +1810,7 @@ xfs_qm_dqusage_adjust( * Now release the inode. This will send it to 'inactive', and * possibly even free blocks. */ - VN_RELE(XFS_ITOV(ip)); + IRELE(ip); /* * Goto next inode. @@ -1968,7 +1968,7 @@ xfs_qm_init_quotainos( if ((error = xfs_iget(mp, NULL, mp->m_sb.sb_gquotino, 0, 0, &gip, 0))) { if (uip) - VN_RELE(XFS_ITOV(uip)); + IRELE(uip); return XFS_ERROR(error); } } @@ -1999,7 +1999,7 @@ xfs_qm_init_quotainos( sbflags | XFS_SB_GQUOTINO, flags); if (error) { if (uip) - VN_RELE(XFS_ITOV(uip)); + IRELE(uip); return XFS_ERROR(error); } diff --git a/fs/xfs/quota/xfs_qm_syscalls.c b/fs/xfs/quota/xfs_qm_syscalls.c index d2b8be7e75f..3dc161f39d1 100644 --- a/fs/xfs/quota/xfs_qm_syscalls.c +++ b/fs/xfs/quota/xfs_qm_syscalls.c @@ -386,7 +386,7 @@ xfs_qm_scall_trunc_qfiles( error = xfs_iget(mp, NULL, mp->m_sb.sb_uquotino, 0, 0, &qip, 0); if (! error) { (void) xfs_truncate_file(mp, qip); - VN_RELE(XFS_ITOV(qip)); + IRELE(qip); } } @@ -395,7 +395,7 @@ xfs_qm_scall_trunc_qfiles( error = xfs_iget(mp, NULL, mp->m_sb.sb_gquotino, 0, 0, &qip, 0); if (! error) { (void) xfs_truncate_file(mp, qip); - VN_RELE(XFS_ITOV(qip)); + IRELE(qip); } } @@ -552,13 +552,13 @@ xfs_qm_scall_getqstat( out->qs_uquota.qfs_nblks = uip->i_d.di_nblocks; out->qs_uquota.qfs_nextents = uip->i_d.di_nextents; if (tempuqip) - VN_RELE(XFS_ITOV(uip)); + IRELE(uip); } if (gip) { out->qs_gquota.qfs_nblks = gip->i_d.di_nblocks; out->qs_gquota.qfs_nextents = gip->i_d.di_nextents; if (tempgqip) - VN_RELE(XFS_ITOV(gip)); + IRELE(gip); } if (mp->m_quotainfo) { out->qs_incoredqs = XFS_QI_MPLNDQUOTS(mp); @@ -1095,7 +1095,7 @@ again: * inactive code in hell. */ if (vnode_refd) - VN_RELE(vp); + IRELE(ip); XFS_MOUNT_ILOCK(mp); /* * If an inode was inserted or removed, we gotta diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index cd24711ae27..962d74a9ea7 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -46,6 +46,7 @@ #include "xfs_trans_priv.h" #include "xfs_quota.h" #include "xfs_rw.h" +#include "xfs_utils.h" STATIC int xlog_find_zeroed(xlog_t *, xfs_daddr_t *); STATIC int xlog_clear_stale_blocks(xlog_t *, xfs_lsn_t); @@ -3248,7 +3249,7 @@ xlog_recover_process_iunlinks( if (ip->i_d.di_mode == 0) xfs_iput_new(ip, 0); else - VN_RELE(XFS_ITOV(ip)); + IRELE(ip); } else { /* * We can't read in the inode diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 41b690e6896..c2aafeb8c6c 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -43,6 +43,7 @@ #include "xfs_rw.h" #include "xfs_quota.h" #include "xfs_fsops.h" +#include "xfs_utils.h" STATIC void xfs_mount_log_sb(xfs_mount_t *, __int64_t); STATIC int xfs_uuid_mount(xfs_mount_t *); @@ -956,7 +957,6 @@ xfs_mountfs( { xfs_sb_t *sbp = &(mp->m_sb); xfs_inode_t *rip; - bhv_vnode_t *rvp = NULL; __uint64_t resblks; __int64_t update_flags = 0LL; uint quotamount, quotaflags; @@ -1158,7 +1158,6 @@ xfs_mountfs( } ASSERT(rip != NULL); - rvp = XFS_ITOV(rip); if (unlikely((rip->i_d.di_mode & S_IFMT) != S_IFDIR)) { cmn_err(CE_WARN, "XFS: corrupted root inode"); @@ -1241,7 +1240,7 @@ xfs_mountfs( /* * Free up the root inode. */ - VN_RELE(rvp); + IRELE(rip); error3: xfs_log_unmount_dealloc(mp); error2: diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c index 47082c01872..9cd6471cd60 100644 --- a/fs/xfs/xfs_rtalloc.c +++ b/fs/xfs/xfs_rtalloc.c @@ -44,6 +44,7 @@ #include "xfs_rw.h" #include "xfs_inode_item.h" #include "xfs_trans_space.h" +#include "xfs_utils.h" /* @@ -2278,7 +2279,7 @@ xfs_rtmount_inodes( ASSERT(sbp->sb_rsumino != NULLFSINO); error = xfs_iget(mp, NULL, sbp->sb_rsumino, 0, 0, &mp->m_rsumip, 0); if (error) { - VN_RELE(XFS_ITOV(mp->m_rbmip)); + IRELE(mp->m_rbmip); return error; } ASSERT(mp->m_rsumip != NULL); diff --git a/fs/xfs/xfs_vfsops.c b/fs/xfs/xfs_vfsops.c index 4c132a87d43..c21e4d16829 100644 --- a/fs/xfs/xfs_vfsops.c +++ b/fs/xfs/xfs_vfsops.c @@ -55,6 +55,7 @@ #include "xfs_fsops.h" #include "xfs_vnodeops.h" #include "xfs_vfsops.h" +#include "xfs_utils.h" int __init @@ -595,7 +596,7 @@ xfs_unmount( /* * Drop the reference count */ - VN_RELE(rvp); + IRELE(rip); /* * If we're forcing a shutdown, typically because of a media error, @@ -777,8 +778,8 @@ xfs_unmount_flush( goto fscorrupt_out2; if (rbmip) { - VN_RELE(XFS_ITOV(rbmip)); - VN_RELE(XFS_ITOV(rsumip)); + IRELE(rbmip); + IRELE(rsumip); } xfs_iunlock(rip, XFS_ILOCK_EXCL); @@ -1156,10 +1157,10 @@ xfs_sync_inodes( * above, then wait until after we've unlocked * the inode to release the reference. This is * because we can be already holding the inode - * lock when VN_RELE() calls xfs_inactive(). + * lock when IRELE() calls xfs_inactive(). * * Make sure to drop the mount lock before calling - * VN_RELE() so that we don't trip over ourselves if + * IRELE() so that we don't trip over ourselves if * we have to go for the mount lock again in the * inactive code. */ @@ -1167,7 +1168,7 @@ xfs_sync_inodes( IPOINTER_INSERT(ip, mp); } - VN_RELE(vp); + IRELE(ip); vnode_refed = B_FALSE; } -- cgit v1.2.3 From 2abdb8c88110bab78bfe17e51346e735560daa02 Mon Sep 17 00:00:00 2001 From: Lachlan McIlroy Date: Thu, 27 Mar 2008 18:01:14 +1100 Subject: [XFS] Prevent xfs_bmap_check_leaf_extents() referencing unmapped memory. While investigating the extent corruption bug I ran into this bug in debug only code. xfs_bmap_check_leaf_extents() loops through the leaf blocks of the extent btree checking that every extent is entirely before the next extent. It also compares the last extent in the previous block to the first extent in the current block when the previous block has been released and potentially unmapped. So take a copy of the last extent instead of a pointer. Also move the last extent check out of the loop because we only need to do it once. SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30718a Signed-off-by: Lachlan McIlroy Signed-off-by: Christoph Hellwig --- fs/xfs/xfs_bmap.c | 15 ++++++--------- 1 file changed, 6 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index bce8e3bd8ad..7d683e0b8ef 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -6194,7 +6194,7 @@ xfs_bmap_check_leaf_extents( xfs_mount_t *mp; /* file system mount structure */ __be64 *pp; /* pointer to block address */ xfs_bmbt_rec_t *ep; /* pointer to current extent */ - xfs_bmbt_rec_t *lastp; /* pointer to previous extent */ + xfs_bmbt_rec_t last = {0, 0}; /* last extent in prev block */ xfs_bmbt_rec_t *nextp; /* pointer to next extent */ int bp_release = 0; @@ -6264,7 +6264,6 @@ xfs_bmap_check_leaf_extents( /* * Loop over all leaf nodes checking that all extents are in the right order. */ - lastp = NULL; for (;;) { xfs_fsblock_t nextbno; xfs_extnum_t num_recs; @@ -6285,18 +6284,16 @@ xfs_bmap_check_leaf_extents( */ ep = XFS_BTREE_REC_ADDR(xfs_bmbt, block, 1); + if (i) { + xfs_btree_check_rec(XFS_BTNUM_BMAP, &last, ep); + } for (j = 1; j < num_recs; j++) { nextp = XFS_BTREE_REC_ADDR(xfs_bmbt, block, j + 1); - if (lastp) { - xfs_btree_check_rec(XFS_BTNUM_BMAP, - (void *)lastp, (void *)ep); - } - xfs_btree_check_rec(XFS_BTNUM_BMAP, (void *)ep, - (void *)(nextp)); - lastp = ep; + xfs_btree_check_rec(XFS_BTNUM_BMAP, ep, nextp); ep = nextp; } + last = *ep; i += num_recs; if (bp_release) { bp_release = 0; -- cgit v1.2.3 From 114d23aae51233b2bc62d8e2a632bcb55de1953d Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:18:39 +1000 Subject: [XFS] Per iclog callback chain lock Rather than use the icloglock for protecting the iclog completion callback chain, use a new per-iclog lock so that walking the callback chain doesn't require holding a global lock. This reduces contention on the icloglock during transaction commit and log I/O completion by reducing the number of times we need to hold the global icloglock during these operations. SGI-PV: 978729 SGI-Modid: xfs-linux-melb:xfs-kern:30770a Signed-off-by: David Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_log.c | 35 +++++++++++++++++++---------------- fs/xfs/xfs_log_priv.h | 33 ++++++++++++++++++++++++++------- 2 files changed, 45 insertions(+), 23 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index 1fa98093389..7a5b12d9353 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -397,12 +397,10 @@ xfs_log_notify(xfs_mount_t *mp, /* mount of partition */ void *iclog_hndl, /* iclog to hang callback off */ xfs_log_callback_t *cb) { - xlog_t *log = mp->m_log; xlog_in_core_t *iclog = (xlog_in_core_t *)iclog_hndl; int abortflg; - cb->cb_next = NULL; - spin_lock(&log->l_icloglock); + spin_lock(&iclog->ic_callback_lock); abortflg = (iclog->ic_state & XLOG_STATE_IOERROR); if (!abortflg) { ASSERT_ALWAYS((iclog->ic_state == XLOG_STATE_ACTIVE) || @@ -411,7 +409,7 @@ xfs_log_notify(xfs_mount_t *mp, /* mount of partition */ *(iclog->ic_callback_tail) = cb; iclog->ic_callback_tail = &(cb->cb_next); } - spin_unlock(&log->l_icloglock); + spin_unlock(&iclog->ic_callback_lock); return abortflg; } /* xfs_log_notify */ @@ -1257,6 +1255,8 @@ xlog_alloc_log(xfs_mount_t *mp, iclog->ic_size = XFS_BUF_SIZE(bp) - log->l_iclog_hsize; iclog->ic_state = XLOG_STATE_ACTIVE; iclog->ic_log = log; + atomic_set(&iclog->ic_refcnt, 0); + spin_lock_init(&iclog->ic_callback_lock); iclog->ic_callback_tail = &(iclog->ic_callback); iclog->ic_datap = (char *)iclog->hic_data + log->l_iclog_hsize; @@ -1987,7 +1987,7 @@ xlog_state_clean_log(xlog_t *log) if (iclog->ic_state == XLOG_STATE_DIRTY) { iclog->ic_state = XLOG_STATE_ACTIVE; iclog->ic_offset = 0; - iclog->ic_callback = NULL; /* don't need to free */ + ASSERT(iclog->ic_callback == NULL); /* * If the number of ops in this iclog indicate it just * contains the dummy transaction, we can @@ -2190,37 +2190,40 @@ xlog_state_do_callback( be64_to_cpu(iclog->ic_header.h_lsn); spin_unlock(&log->l_grant_lock); - /* - * Keep processing entries in the callback list - * until we come around and it is empty. We - * need to atomically see that the list is - * empty and change the state to DIRTY so that - * we don't miss any more callbacks being added. - */ - spin_lock(&log->l_icloglock); } else { + spin_unlock(&log->l_icloglock); ioerrors++; } - cb = iclog->ic_callback; + /* + * Keep processing entries in the callback list until + * we come around and it is empty. We need to + * atomically see that the list is empty and change the + * state to DIRTY so that we don't miss any more + * callbacks being added. + */ + spin_lock(&iclog->ic_callback_lock); + cb = iclog->ic_callback; while (cb) { iclog->ic_callback_tail = &(iclog->ic_callback); iclog->ic_callback = NULL; - spin_unlock(&log->l_icloglock); + spin_unlock(&iclog->ic_callback_lock); /* perform callbacks in the order given */ for (; cb; cb = cb_next) { cb_next = cb->cb_next; cb->cb_func(cb->cb_arg, aborted); } - spin_lock(&log->l_icloglock); + spin_lock(&iclog->ic_callback_lock); cb = iclog->ic_callback; } loopdidcallbacks++; funcdidcallbacks++; + spin_lock(&log->l_icloglock); ASSERT(iclog->ic_callback == NULL); + spin_unlock(&iclog->ic_callback_lock); if (!(iclog->ic_state & XLOG_STATE_IOERROR)) iclog->ic_state = XLOG_STATE_DIRTY; diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h index 01c63db25a1..104b623aa08 100644 --- a/fs/xfs/xfs_log_priv.h +++ b/fs/xfs/xfs_log_priv.h @@ -324,6 +324,19 @@ typedef struct xlog_rec_ext_header { * - ic_offset is the current number of bytes written to in this iclog. * - ic_refcnt is bumped when someone is writing to the log. * - ic_state is the state of the iclog. + * + * Because of cacheline contention on large machines, we need to separate + * various resources onto different cachelines. To start with, make the + * structure cacheline aligned. The following fields can be contended on + * by independent processes: + * + * - ic_callback_* + * - ic_refcnt + * - fields protected by the global l_icloglock + * + * so we need to ensure that these fields are located in separate cachelines. + * We'll put all the read-only and l_icloglock fields in the first cacheline, + * and move everything else out to subsequent cachelines. */ typedef struct xlog_iclog_fields { sv_t ic_forcesema; @@ -332,18 +345,23 @@ typedef struct xlog_iclog_fields { struct xlog_in_core *ic_prev; struct xfs_buf *ic_bp; struct log *ic_log; - xfs_log_callback_t *ic_callback; - xfs_log_callback_t **ic_callback_tail; -#ifdef XFS_LOG_TRACE - struct ktrace *ic_trace; -#endif int ic_size; int ic_offset; - atomic_t ic_refcnt; int ic_bwritecnt; ushort_t ic_state; char *ic_datap; /* pointer to iclog data */ -} xlog_iclog_fields_t; +#ifdef XFS_LOG_TRACE + struct ktrace *ic_trace; +#endif + + /* Callback structures need their own cacheline */ + spinlock_t ic_callback_lock ____cacheline_aligned_in_smp; + xfs_log_callback_t *ic_callback; + xfs_log_callback_t **ic_callback_tail; + + /* reference counts need their own cacheline */ + atomic_t ic_refcnt ____cacheline_aligned_in_smp; +} xlog_iclog_fields_t ____cacheline_aligned_in_smp; typedef union xlog_in_core2 { xlog_rec_header_t hic_header; @@ -366,6 +384,7 @@ typedef struct xlog_in_core { #define ic_bp hic_fields.ic_bp #define ic_log hic_fields.ic_log #define ic_callback hic_fields.ic_callback +#define ic_callback_lock hic_fields.ic_callback_lock #define ic_callback_tail hic_fields.ic_callback_tail #define ic_trace hic_fields.ic_trace #define ic_size hic_fields.ic_size -- cgit v1.2.3 From eb01c9cd87c7a9998c2edf209721ea069e3e3652 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:18:46 +1000 Subject: [XFS] Remove the xlog_ticket allocator The ticket allocator is just a simple slab implementation internal to the log. It requires the icloglock to be held when manipulating it and this contributes to contention on that lock. Just kill the entire allocator and use a memory zone instead. While there, allow us to gracefully fail allocation with ENOMEM. SGI-PV: 978729 SGI-Modid: xfs-linux-melb:xfs-kern:30771a Signed-off-by: David Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_log.c | 137 ++++---------------------------------------------- fs/xfs/xfs_log_priv.h | 9 ++-- fs/xfs/xfs_vfsops.c | 12 +++-- 3 files changed, 21 insertions(+), 137 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index 7a5b12d9353..3cf115d8de7 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -41,6 +41,7 @@ #include "xfs_inode.h" #include "xfs_rw.h" +kmem_zone_t *xfs_log_ticket_zone; #define xlog_write_adv_cnt(ptr, len, off, bytes) \ { (ptr) += (bytes); \ @@ -73,8 +74,6 @@ STATIC int xlog_state_get_iclog_space(xlog_t *log, xlog_ticket_t *ticket, int *continued_write, int *logoffsetp); -STATIC void xlog_state_put_ticket(xlog_t *log, - xlog_ticket_t *tic); STATIC int xlog_state_release_iclog(xlog_t *log, xlog_in_core_t *iclog); STATIC void xlog_state_switch_iclogs(xlog_t *log, @@ -101,7 +100,6 @@ STATIC void xlog_ungrant_log_space(xlog_t *log, /* local ticket functions */ -STATIC void xlog_state_ticket_alloc(xlog_t *log); STATIC xlog_ticket_t *xlog_ticket_get(xlog_t *log, int unit_bytes, int count, @@ -330,7 +328,7 @@ xfs_log_done(xfs_mount_t *mp, */ xlog_trace_loggrant(log, ticket, "xfs_log_done: (non-permanent)"); xlog_ungrant_log_space(log, ticket); - xlog_state_put_ticket(log, ticket); + xlog_ticket_put(log, ticket); } else { xlog_trace_loggrant(log, ticket, "xfs_log_done: (permanent)"); xlog_regrant_reserve_log_space(log, ticket); @@ -469,6 +467,8 @@ xfs_log_reserve(xfs_mount_t *mp, /* may sleep if need to allocate more tickets */ internal_ticket = xlog_ticket_get(log, unit_bytes, cnt, client, flags); + if (!internal_ticket) + return XFS_ERROR(ENOMEM); internal_ticket->t_trans_type = t_type; *ticket = internal_ticket; xlog_trace_loggrant(log, internal_ticket, @@ -693,7 +693,7 @@ xfs_log_unmount_write(xfs_mount_t *mp) if (tic) { xlog_trace_loggrant(log, tic, "unmount rec"); xlog_ungrant_log_space(log, tic); - xlog_state_put_ticket(log, tic); + xlog_ticket_put(log, tic); } } else { /* @@ -1208,7 +1208,6 @@ xlog_alloc_log(xfs_mount_t *mp, spin_lock_init(&log->l_icloglock); spin_lock_init(&log->l_grant_lock); initnsema(&log->l_flushsema, 0, "ic-flush"); - xlog_state_ticket_alloc(log); /* wait until after icloglock inited */ /* log record size must be multiple of BBSIZE; see xlog_rec_header_t */ ASSERT((XFS_BUF_SIZE(bp) & BBMASK) == 0); @@ -1538,7 +1537,6 @@ STATIC void xlog_dealloc_log(xlog_t *log) { xlog_in_core_t *iclog, *next_iclog; - xlog_ticket_t *tic, *next_tic; int i; iclog = log->l_iclog; @@ -1559,22 +1557,6 @@ xlog_dealloc_log(xlog_t *log) spinlock_destroy(&log->l_icloglock); spinlock_destroy(&log->l_grant_lock); - /* XXXsup take a look at this again. */ - if ((log->l_ticket_cnt != log->l_ticket_tcnt) && - !XLOG_FORCED_SHUTDOWN(log)) { - xfs_fs_cmn_err(CE_WARN, log->l_mp, - "xlog_dealloc_log: (cnt: %d, total: %d)", - log->l_ticket_cnt, log->l_ticket_tcnt); - /* ASSERT(log->l_ticket_cnt == log->l_ticket_tcnt); */ - - } else { - tic = log->l_unmount_free; - while (tic) { - next_tic = tic->t_next; - kmem_free(tic, PAGE_SIZE); - tic = next_tic; - } - } xfs_buf_free(log->l_xbuf); #ifdef XFS_LOG_TRACE if (log->l_trace != NULL) { @@ -2794,18 +2776,6 @@ xlog_ungrant_log_space(xlog_t *log, } /* xlog_ungrant_log_space */ -/* - * Atomically put back used ticket. - */ -STATIC void -xlog_state_put_ticket(xlog_t *log, - xlog_ticket_t *tic) -{ - spin_lock(&log->l_icloglock); - xlog_ticket_put(log, tic); - spin_unlock(&log->l_icloglock); -} /* xlog_state_put_ticket */ - /* * Flush iclog to disk if this is the last reference to the given iclog and * the WANT_SYNC bit is set. @@ -3176,92 +3146,19 @@ xlog_state_want_sync(xlog_t *log, xlog_in_core_t *iclog) */ /* - * Algorithm doesn't take into account page size. ;-( - */ -STATIC void -xlog_state_ticket_alloc(xlog_t *log) -{ - xlog_ticket_t *t_list; - xlog_ticket_t *next; - xfs_caddr_t buf; - uint i = (PAGE_SIZE / sizeof(xlog_ticket_t)) - 2; - - /* - * The kmem_zalloc may sleep, so we shouldn't be holding the - * global lock. XXXmiken: may want to use zone allocator. - */ - buf = (xfs_caddr_t) kmem_zalloc(PAGE_SIZE, KM_SLEEP); - - spin_lock(&log->l_icloglock); - - /* Attach 1st ticket to Q, so we can keep track of allocated memory */ - t_list = (xlog_ticket_t *)buf; - t_list->t_next = log->l_unmount_free; - log->l_unmount_free = t_list++; - log->l_ticket_cnt++; - log->l_ticket_tcnt++; - - /* Next ticket becomes first ticket attached to ticket free list */ - if (log->l_freelist != NULL) { - ASSERT(log->l_tail != NULL); - log->l_tail->t_next = t_list; - } else { - log->l_freelist = t_list; - } - log->l_ticket_cnt++; - log->l_ticket_tcnt++; - - /* Cycle through rest of alloc'ed memory, building up free Q */ - for ( ; i > 0; i--) { - next = t_list + 1; - t_list->t_next = next; - t_list = next; - log->l_ticket_cnt++; - log->l_ticket_tcnt++; - } - t_list->t_next = NULL; - log->l_tail = t_list; - spin_unlock(&log->l_icloglock); -} /* xlog_state_ticket_alloc */ - - -/* - * Put ticket into free list - * - * Assumption: log lock is held around this call. + * Free a used ticket. */ STATIC void xlog_ticket_put(xlog_t *log, xlog_ticket_t *ticket) { sv_destroy(&ticket->t_sema); - - /* - * Don't think caching will make that much difference. It's - * more important to make debug easier. - */ -#if 0 - /* real code will want to use LIFO for caching */ - ticket->t_next = log->l_freelist; - log->l_freelist = ticket; - /* no need to clear fields */ -#else - /* When we debug, it is easier if tickets are cycled */ - ticket->t_next = NULL; - if (log->l_tail) { - log->l_tail->t_next = ticket; - } else { - ASSERT(log->l_freelist == NULL); - log->l_freelist = ticket; - } - log->l_tail = ticket; -#endif /* DEBUG */ - log->l_ticket_cnt++; + kmem_zone_free(xfs_log_ticket_zone, ticket); } /* xlog_ticket_put */ /* - * Grab ticket off freelist or allocation some more + * Allocate and initialise a new log ticket. */ STATIC xlog_ticket_t * xlog_ticket_get(xlog_t *log, @@ -3273,21 +3170,9 @@ xlog_ticket_get(xlog_t *log, xlog_ticket_t *tic; uint num_headers; - alloc: - if (log->l_freelist == NULL) - xlog_state_ticket_alloc(log); /* potentially sleep */ - - spin_lock(&log->l_icloglock); - if (log->l_freelist == NULL) { - spin_unlock(&log->l_icloglock); - goto alloc; - } - tic = log->l_freelist; - log->l_freelist = tic->t_next; - if (log->l_freelist == NULL) - log->l_tail = NULL; - log->l_ticket_cnt--; - spin_unlock(&log->l_icloglock); + tic = kmem_zone_zalloc(xfs_log_ticket_zone, KM_SLEEP|KM_MAYFAIL); + if (!tic) + return NULL; /* * Permanent reservations have up to 'cnt'-1 active log operations diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h index 104b623aa08..c1583960009 100644 --- a/fs/xfs/xfs_log_priv.h +++ b/fs/xfs/xfs_log_priv.h @@ -242,7 +242,7 @@ typedef struct xlog_res { typedef struct xlog_ticket { sv_t t_sema; /* sleep on this semaphore : 20 */ - struct xlog_ticket *t_next; /* :4|8 */ + struct xlog_ticket *t_next; /* :4|8 */ struct xlog_ticket *t_prev; /* :4|8 */ xlog_tid_t t_tid; /* transaction identifier : 4 */ int t_curr_res; /* current reservation in bytes : 4 */ @@ -406,13 +406,8 @@ typedef struct log { sema_t l_flushsema; /* iclog flushing semaphore */ int l_flushcnt; /* # of procs waiting on this * sema */ - int l_ticket_cnt; /* free ticket count */ - int l_ticket_tcnt; /* total ticket count */ int l_covered_state;/* state of "covering disk * log entries" */ - xlog_ticket_t *l_freelist; /* free list of tickets */ - xlog_ticket_t *l_unmount_free;/* kmem_free these addresses */ - xlog_ticket_t *l_tail; /* free list of tickets */ xlog_in_core_t *l_iclog; /* head log queue */ spinlock_t l_icloglock; /* grab to change iclog state */ xfs_lsn_t l_tail_lsn; /* lsn of 1st LR with unflushed @@ -478,6 +473,8 @@ extern struct xfs_buf *xlog_get_bp(xlog_t *, int); extern void xlog_put_bp(struct xfs_buf *); extern int xlog_bread(xlog_t *, xfs_daddr_t, int, struct xfs_buf *); +extern kmem_zone_t *xfs_log_ticket_zone; + /* iclog tracing */ #define XLOG_TRACE_GRAB_FLUSH 1 #define XLOG_TRACE_REL_FLUSH 2 diff --git a/fs/xfs/xfs_vfsops.c b/fs/xfs/xfs_vfsops.c index c21e4d16829..ea94593b531 100644 --- a/fs/xfs/xfs_vfsops.c +++ b/fs/xfs/xfs_vfsops.c @@ -69,15 +69,17 @@ xfs_init(void) /* * Initialize all of the zone allocators we use. */ + xfs_log_ticket_zone = kmem_zone_init(sizeof(xlog_ticket_t), + "xfs_log_ticket"); xfs_bmap_free_item_zone = kmem_zone_init(sizeof(xfs_bmap_free_item_t), - "xfs_bmap_free_item"); + "xfs_bmap_free_item"); xfs_btree_cur_zone = kmem_zone_init(sizeof(xfs_btree_cur_t), - "xfs_btree_cur"); - xfs_trans_zone = kmem_zone_init(sizeof(xfs_trans_t), "xfs_trans"); - xfs_da_state_zone = - kmem_zone_init(sizeof(xfs_da_state_t), "xfs_da_state"); + "xfs_btree_cur"); + xfs_da_state_zone = kmem_zone_init(sizeof(xfs_da_state_t), + "xfs_da_state"); xfs_dabuf_zone = kmem_zone_init(sizeof(xfs_dabuf_t), "xfs_dabuf"); xfs_ifork_zone = kmem_zone_init(sizeof(xfs_ifork_t), "xfs_ifork"); + xfs_trans_zone = kmem_zone_init(sizeof(xfs_trans_t), "xfs_trans"); xfs_acl_zone_init(xfs_acl_zone, "xfs_acl"); xfs_mru_cache_init(); xfs_filestream_init(); -- cgit v1.2.3 From 4679b2d36d53ed508c956337972fbbea8db99a77 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:18:54 +1000 Subject: [XFS] Reorganise xlog_t for better cacheline isolation of contention To reduce contention on the log in large CPU count, separate out different parts of the xlog_t structure onto different cachelines. Move each lock onto a different cacheline along with all the members that are accessed/modified while that lock is held. Also, move the debugging code into debug code. SGI-PV: 978729 SGI-Modid: xfs-linux-melb:xfs-kern:30772a Signed-off-by: David Chinner Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_log.c | 5 ++--- fs/xfs/xfs_log_priv.h | 55 ++++++++++++++++++++++++++++----------------------- 2 files changed, 32 insertions(+), 28 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index 3cf115d8de7..319b98eb410 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -1237,9 +1237,9 @@ xlog_alloc_log(xfs_mount_t *mp, XFS_BUF_SET_FSPRIVATE2(bp, (unsigned long)1); iclog->ic_bp = bp; iclog->hic_data = bp->b_addr; - +#ifdef DEBUG log->l_iclog_bak[i] = (xfs_caddr_t)&(iclog->ic_header); - +#endif head = &iclog->ic_header; memset(head, 0, sizeof(xlog_rec_header_t)); head->h_magicno = cpu_to_be32(XLOG_HEADER_MAGIC_NUM); @@ -1250,7 +1250,6 @@ xlog_alloc_log(xfs_mount_t *mp, head->h_fmt = cpu_to_be32(XLOG_FMT); memcpy(&head->h_fs_uuid, &mp->m_sb.sb_uuid, sizeof(uuid_t)); - iclog->ic_size = XFS_BUF_SIZE(bp) - log->l_iclog_hsize; iclog->ic_state = XLOG_STATE_ACTIVE; iclog->ic_log = log; diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h index c1583960009..8952a392b5f 100644 --- a/fs/xfs/xfs_log_priv.h +++ b/fs/xfs/xfs_log_priv.h @@ -361,7 +361,7 @@ typedef struct xlog_iclog_fields { /* reference counts need their own cacheline */ atomic_t ic_refcnt ____cacheline_aligned_in_smp; -} xlog_iclog_fields_t ____cacheline_aligned_in_smp; +} xlog_iclog_fields_t; typedef union xlog_in_core2 { xlog_rec_header_t hic_header; @@ -402,8 +402,29 @@ typedef struct xlog_in_core { * that round off problems won't occur when releasing partial reservations. */ typedef struct log { + /* The following fields don't need locking */ + struct xfs_mount *l_mp; /* mount point */ + struct xfs_buf *l_xbuf; /* extra buffer for log + * wrapping */ + struct xfs_buftarg *l_targ; /* buftarg of log */ + uint l_flags; + uint l_quotaoffs_flag; /* XFS_DQ_*, for QUOTAOFFs */ + struct xfs_buf_cancel **l_buf_cancel_table; + int l_iclog_hsize; /* size of iclog header */ + int l_iclog_heads; /* # of iclog header sectors */ + uint l_sectbb_log; /* log2 of sector size in BBs */ + uint l_sectbb_mask; /* sector size (in BBs) + * alignment mask */ + int l_iclog_size; /* size of log in bytes */ + int l_iclog_size_log; /* log power size of log */ + int l_iclog_bufs; /* number of iclog buffers */ + xfs_daddr_t l_logBBstart; /* start block of log */ + int l_logsize; /* size of log in bytes */ + int l_logBBsize; /* size of log in BB chunks */ + /* The following block of fields are changed while holding icloglock */ - sema_t l_flushsema; /* iclog flushing semaphore */ + sema_t l_flushsema ____cacheline_aligned_in_smp; + /* iclog flushing semaphore */ int l_flushcnt; /* # of procs waiting on this * sema */ int l_covered_state;/* state of "covering disk @@ -413,27 +434,14 @@ typedef struct log { xfs_lsn_t l_tail_lsn; /* lsn of 1st LR with unflushed * buffers */ xfs_lsn_t l_last_sync_lsn;/* lsn of last LR on disk */ - struct xfs_mount *l_mp; /* mount point */ - struct xfs_buf *l_xbuf; /* extra buffer for log - * wrapping */ - struct xfs_buftarg *l_targ; /* buftarg of log */ - xfs_daddr_t l_logBBstart; /* start block of log */ - int l_logsize; /* size of log in bytes */ - int l_logBBsize; /* size of log in BB chunks */ int l_curr_cycle; /* Cycle number of log writes */ int l_prev_cycle; /* Cycle number before last * block increment */ int l_curr_block; /* current logical log block */ int l_prev_block; /* previous logical log block */ - int l_iclog_size; /* size of log in bytes */ - int l_iclog_size_log; /* log power size of log */ - int l_iclog_bufs; /* number of iclog buffers */ - - /* The following field are used for debugging; need to hold icloglock */ - char *l_iclog_bak[XLOG_MAX_ICLOGS]; /* The following block of fields are changed while holding grant_lock */ - spinlock_t l_grant_lock; + spinlock_t l_grant_lock ____cacheline_aligned_in_smp; xlog_ticket_t *l_reserve_headq; xlog_ticket_t *l_write_headq; int l_grant_reserve_cycle; @@ -441,19 +449,16 @@ typedef struct log { int l_grant_write_cycle; int l_grant_write_bytes; - /* The following fields don't need locking */ #ifdef XFS_LOG_TRACE struct ktrace *l_trace; struct ktrace *l_grant_trace; #endif - uint l_flags; - uint l_quotaoffs_flag; /* XFS_DQ_*, for QUOTAOFFs */ - struct xfs_buf_cancel **l_buf_cancel_table; - int l_iclog_hsize; /* size of iclog header */ - int l_iclog_heads; /* # of iclog header sectors */ - uint l_sectbb_log; /* log2 of sector size in BBs */ - uint l_sectbb_mask; /* sector size (in BBs) - * alignment mask */ + + /* The following field are used for debugging; need to hold icloglock */ +#ifdef DEBUG + char *l_iclog_bak[XLOG_MAX_ICLOGS]; +#endif + } xlog_t; #define XLOG_FORCED_SHUTDOWN(log) ((log)->l_flags & XLOG_IO_ERROR) -- cgit v1.2.3 From 6b1d1a732f886936fe515d911b1a01d9cc50e179 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:19:02 +1000 Subject: [XFS] Fix lock inversion in forced shutdown. Recent changes to xlog_state_release_iclog() placed the grant_lock inside the icloglock. forced unmount of the log does this the opposite way around, but does not depend on the order for correct working. Fix the inversion by changing the order locks are gained in xfs_log_force_umount(). SGI-PV: 979661 SGI-Modid: xfs-linux-melb:xfs-kern:30773a Signed-off-by: David Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_log.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index 319b98eb410..4a6f7c5d145 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -3499,8 +3499,8 @@ xfs_log_force_umount( * before we mark the filesystem SHUTDOWN and wake * everybody up to tell the bad news. */ - spin_lock(&log->l_grant_lock); spin_lock(&log->l_icloglock); + spin_lock(&log->l_grant_lock); mp->m_flags |= XFS_MOUNT_FS_SHUTDOWN; XFS_BUF_DONE(mp->m_sb_bp); /* -- cgit v1.2.3 From 0225da1f35df46c67785eb08526995d7cdb4e3b0 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Thu, 10 Apr 2008 12:19:10 +1000 Subject: [XFS] Replace __inline with inline Remove the remaining uses of __inline in the XFS code base. SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30774a Signed-off-by: Harvey Harrison Signed-off-by: David Chinner Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_cred.h | 2 +- fs/xfs/linux-2.6/xfs_stats.h | 4 ++-- fs/xfs/quota/xfs_qm_stats.h | 4 ++-- 3 files changed, 5 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_cred.h b/fs/xfs/linux-2.6/xfs_cred.h index e7f3da61c6c..652721ce0ea 100644 --- a/fs/xfs/linux-2.6/xfs_cred.h +++ b/fs/xfs/linux-2.6/xfs_cred.h @@ -30,7 +30,7 @@ typedef struct cred { extern struct cred *sys_cred; /* this is a hack.. (assumes sys_cred is the only cred_t in the system) */ -static __inline int capable_cred(cred_t *cr, int cid) +static inline int capable_cred(cred_t *cr, int cid) { return (cr == sys_cred) ? 1 : capable(cid); } diff --git a/fs/xfs/linux-2.6/xfs_stats.h b/fs/xfs/linux-2.6/xfs_stats.h index 8ba7a2fa6c1..afd0b0d5fdb 100644 --- a/fs/xfs/linux-2.6/xfs_stats.h +++ b/fs/xfs/linux-2.6/xfs_stats.h @@ -144,8 +144,8 @@ extern void xfs_cleanup_procfs(void); # define XFS_STATS_DEC(count) # define XFS_STATS_ADD(count, inc) -static __inline void xfs_init_procfs(void) { }; -static __inline void xfs_cleanup_procfs(void) { }; +static inline void xfs_init_procfs(void) { }; +static inline void xfs_cleanup_procfs(void) { }; #endif /* !CONFIG_PROC_FS */ diff --git a/fs/xfs/quota/xfs_qm_stats.h b/fs/xfs/quota/xfs_qm_stats.h index a50ffabcf55..5b964fc0dc0 100644 --- a/fs/xfs/quota/xfs_qm_stats.h +++ b/fs/xfs/quota/xfs_qm_stats.h @@ -45,8 +45,8 @@ extern void xfs_qm_cleanup_procfs(void); # define XQM_STATS_INC(count) do { } while (0) -static __inline void xfs_qm_init_procfs(void) { }; -static __inline void xfs_qm_cleanup_procfs(void) { }; +static inline void xfs_qm_init_procfs(void) { }; +static inline void xfs_qm_cleanup_procfs(void) { }; #endif -- cgit v1.2.3 From 34a622b2e1c8e11c8990184634f101c1aad42fec Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Thu, 10 Apr 2008 12:19:21 +1000 Subject: [XFS] replace remaining __FUNCTION__ occurrences __FUNCTION__ is gcc-specific, use __func__ SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30775a Signed-off-by: Harvey Harrison Signed-off-by: David Chinner Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/kmem.c | 6 +++--- fs/xfs/linux-2.6/xfs_buf.c | 6 +++--- fs/xfs/linux-2.6/xfs_super.c | 2 +- fs/xfs/linux-2.6/xfs_vnode.h | 4 ++-- fs/xfs/xfs_alloc.c | 18 +++++++++--------- fs/xfs/xfs_bmap.c | 18 +++++++++--------- fs/xfs/xfs_bmap.h | 2 +- fs/xfs/xfs_bmap_btree.c | 16 ++++++++-------- fs/xfs/xfs_filestream.c | 2 +- fs/xfs/xfs_log.c | 2 +- fs/xfs/xfs_log_recover.c | 4 ++-- fs/xfs/xfs_trans_ail.c | 2 +- 12 files changed, 41 insertions(+), 41 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/kmem.c b/fs/xfs/linux-2.6/kmem.c index e040f1ce1b6..9b1bb17a050 100644 --- a/fs/xfs/linux-2.6/kmem.c +++ b/fs/xfs/linux-2.6/kmem.c @@ -37,7 +37,7 @@ kmem_alloc(size_t size, unsigned int __nocast flags) #ifdef DEBUG if (unlikely(!(flags & KM_LARGE) && (size > PAGE_SIZE))) { printk(KERN_WARNING "Large %s attempt, size=%ld\n", - __FUNCTION__, (long)size); + __func__, (long)size); dump_stack(); } #endif @@ -52,7 +52,7 @@ kmem_alloc(size_t size, unsigned int __nocast flags) if (!(++retries % 100)) printk(KERN_ERR "XFS: possible memory allocation " "deadlock in %s (mode:0x%x)\n", - __FUNCTION__, lflags); + __func__, lflags); congestion_wait(WRITE, HZ/50); } while (1); } @@ -129,7 +129,7 @@ kmem_zone_alloc(kmem_zone_t *zone, unsigned int __nocast flags) if (!(++retries % 100)) printk(KERN_ERR "XFS: possible memory allocation " "deadlock in %s (mode:0x%x)\n", - __FUNCTION__, lflags); + __func__, lflags); congestion_wait(WRITE, HZ/50); } while (1); } diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c index e347bfd47c9..142ddbece37 100644 --- a/fs/xfs/linux-2.6/xfs_buf.c +++ b/fs/xfs/linux-2.6/xfs_buf.c @@ -400,7 +400,7 @@ _xfs_buf_lookup_pages( printk(KERN_ERR "XFS: possible memory allocation " "deadlock in %s (mode:0x%x)\n", - __FUNCTION__, gfp_mask); + __func__, gfp_mask); XFS_STATS_INC(xb_page_retries); xfsbufd_wakeup(0, gfp_mask); @@ -598,7 +598,7 @@ xfs_buf_get_flags( error = _xfs_buf_map_pages(bp, flags); if (unlikely(error)) { printk(KERN_WARNING "%s: failed to map pages\n", - __FUNCTION__); + __func__); goto no_buffer; } } @@ -778,7 +778,7 @@ xfs_buf_get_noaddr( error = _xfs_buf_map_pages(bp, XBF_MAPPED); if (unlikely(error)) { printk(KERN_WARNING "%s: failed to map pages\n", - __FUNCTION__); + __func__); goto fail_free_mem; } diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c index 72e55db948d..fb561beea37 100644 --- a/fs/xfs/linux-2.6/xfs_super.c +++ b/fs/xfs/linux-2.6/xfs_super.c @@ -935,7 +935,7 @@ xfs_fs_clear_inode( xfs_inactive(ip); xfs_iflags_clear(ip, XFS_IMODIFIED); if (xfs_reclaim(ip)) - panic("%s: cannot reclaim 0x%p\n", __FUNCTION__, inode); + panic("%s: cannot reclaim 0x%p\n", __func__, inode); } ASSERT(XFS_I(inode) == NULL); diff --git a/fs/xfs/linux-2.6/xfs_vnode.h b/fs/xfs/linux-2.6/xfs_vnode.h index 4ed5914adef..dbb8a5d27f7 100644 --- a/fs/xfs/linux-2.6/xfs_vnode.h +++ b/fs/xfs/linux-2.6/xfs_vnode.h @@ -288,9 +288,9 @@ extern void xfs_itrace_hold(struct xfs_inode *, char *, int, inst_t *); extern void _xfs_itrace_ref(struct xfs_inode *, char *, int, inst_t *); extern void xfs_itrace_rele(struct xfs_inode *, char *, int, inst_t *); #define xfs_itrace_entry(ip) \ - _xfs_itrace_entry(ip, __FUNCTION__, (inst_t *)__return_address) + _xfs_itrace_entry(ip, __func__, (inst_t *)__return_address) #define xfs_itrace_exit(ip) \ - _xfs_itrace_exit(ip, __FUNCTION__, (inst_t *)__return_address) + _xfs_itrace_exit(ip, __func__, (inst_t *)__return_address) #define xfs_itrace_exit_tag(ip, tag) \ _xfs_itrace_exit(ip, tag, (inst_t *)__return_address) #define xfs_itrace_ref(ip) \ diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c index bdbfbbee495..bd5c01788ef 100644 --- a/fs/xfs/xfs_alloc.c +++ b/fs/xfs/xfs_alloc.c @@ -55,17 +55,17 @@ xfs_alloc_search_busy(xfs_trans_t *tp, ktrace_t *xfs_alloc_trace_buf; #define TRACE_ALLOC(s,a) \ - xfs_alloc_trace_alloc(__FUNCTION__, s, a, __LINE__) + xfs_alloc_trace_alloc(__func__, s, a, __LINE__) #define TRACE_FREE(s,a,b,x,f) \ - xfs_alloc_trace_free(__FUNCTION__, s, mp, a, b, x, f, __LINE__) + xfs_alloc_trace_free(__func__, s, mp, a, b, x, f, __LINE__) #define TRACE_MODAGF(s,a,f) \ - xfs_alloc_trace_modagf(__FUNCTION__, s, mp, a, f, __LINE__) -#define TRACE_BUSY(__FUNCTION__,s,ag,agb,l,sl,tp) \ - xfs_alloc_trace_busy(__FUNCTION__, s, mp, ag, agb, l, sl, tp, XFS_ALLOC_KTRACE_BUSY, __LINE__) -#define TRACE_UNBUSY(__FUNCTION__,s,ag,sl,tp) \ - xfs_alloc_trace_busy(__FUNCTION__, s, mp, ag, -1, -1, sl, tp, XFS_ALLOC_KTRACE_UNBUSY, __LINE__) -#define TRACE_BUSYSEARCH(__FUNCTION__,s,ag,agb,l,sl,tp) \ - xfs_alloc_trace_busy(__FUNCTION__, s, mp, ag, agb, l, sl, tp, XFS_ALLOC_KTRACE_BUSYSEARCH, __LINE__) + xfs_alloc_trace_modagf(__func__, s, mp, a, f, __LINE__) +#define TRACE_BUSY(__func__,s,ag,agb,l,sl,tp) \ + xfs_alloc_trace_busy(__func__, s, mp, ag, agb, l, sl, tp, XFS_ALLOC_KTRACE_BUSY, __LINE__) +#define TRACE_UNBUSY(__func__,s,ag,sl,tp) \ + xfs_alloc_trace_busy(__func__, s, mp, ag, -1, -1, sl, tp, XFS_ALLOC_KTRACE_UNBUSY, __LINE__) +#define TRACE_BUSYSEARCH(__func__,s,ag,agb,l,sl,tp) \ + xfs_alloc_trace_busy(__func__, s, mp, ag, agb, l, sl, tp, XFS_ALLOC_KTRACE_BUSYSEARCH, __LINE__) #else #define TRACE_ALLOC(s,a) #define TRACE_FREE(s,a,b,x,f) diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 7d683e0b8ef..65b8fa83e07 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -323,13 +323,13 @@ xfs_bmap_trace_pre_update( int whichfork); /* data or attr fork */ #define XFS_BMAP_TRACE_DELETE(d,ip,i,c,w) \ - xfs_bmap_trace_delete(__FUNCTION__,d,ip,i,c,w) + xfs_bmap_trace_delete(__func__,d,ip,i,c,w) #define XFS_BMAP_TRACE_INSERT(d,ip,i,c,r1,r2,w) \ - xfs_bmap_trace_insert(__FUNCTION__,d,ip,i,c,r1,r2,w) + xfs_bmap_trace_insert(__func__,d,ip,i,c,r1,r2,w) #define XFS_BMAP_TRACE_POST_UPDATE(d,ip,i,w) \ - xfs_bmap_trace_post_update(__FUNCTION__,d,ip,i,w) + xfs_bmap_trace_post_update(__func__,d,ip,i,w) #define XFS_BMAP_TRACE_PRE_UPDATE(d,ip,i,w) \ - xfs_bmap_trace_pre_update(__FUNCTION__,d,ip,i,w) + xfs_bmap_trace_pre_update(__func__,d,ip,i,w) #else #define XFS_BMAP_TRACE_DELETE(d,ip,i,c,w) #define XFS_BMAP_TRACE_INSERT(d,ip,i,c,r1,r2,w) @@ -6164,10 +6164,10 @@ xfs_check_block( } if (*thispa == *pp) { cmn_err(CE_WARN, "%s: thispa(%d) == pp(%d) %Ld", - __FUNCTION__, j, i, + __func__, j, i, (unsigned long long)be64_to_cpu(*thispa)); panic("%s: ptrs are equal in node\n", - __FUNCTION__); + __func__); } } } @@ -6324,13 +6324,13 @@ xfs_bmap_check_leaf_extents( return; error0: - cmn_err(CE_WARN, "%s: at error0", __FUNCTION__); + cmn_err(CE_WARN, "%s: at error0", __func__); if (bp_release) xfs_trans_brelse(NULL, bp); error_norelse: cmn_err(CE_WARN, "%s: BAD after btree leaves for %d extents", - __FUNCTION__, i); - panic("%s: CORRUPTED BTREE OR SOMETHING", __FUNCTION__); + __func__, i); + panic("%s: CORRUPTED BTREE OR SOMETHING", __func__); return; } #endif diff --git a/fs/xfs/xfs_bmap.h b/fs/xfs/xfs_bmap.h index 87224b7d798..6ff70cda451 100644 --- a/fs/xfs/xfs_bmap.h +++ b/fs/xfs/xfs_bmap.h @@ -151,7 +151,7 @@ xfs_bmap_trace_exlist( xfs_extnum_t cnt, /* count of entries in list */ int whichfork); /* data or attr fork */ #define XFS_BMAP_TRACE_EXLIST(ip,c,w) \ - xfs_bmap_trace_exlist(__FUNCTION__,ip,c,w) + xfs_bmap_trace_exlist(__func__,ip,c,w) #else #define XFS_BMAP_TRACE_EXLIST(ip,c,w) #endif diff --git a/fs/xfs/xfs_bmap_btree.c b/fs/xfs/xfs_bmap_btree.c index 93470b728dd..4f0e849d973 100644 --- a/fs/xfs/xfs_bmap_btree.c +++ b/fs/xfs/xfs_bmap_btree.c @@ -275,21 +275,21 @@ xfs_bmbt_trace_cursor( } #define XFS_BMBT_TRACE_ARGBI(c,b,i) \ - xfs_bmbt_trace_argbi(__FUNCTION__, c, b, i, __LINE__) + xfs_bmbt_trace_argbi(__func__, c, b, i, __LINE__) #define XFS_BMBT_TRACE_ARGBII(c,b,i,j) \ - xfs_bmbt_trace_argbii(__FUNCTION__, c, b, i, j, __LINE__) + xfs_bmbt_trace_argbii(__func__, c, b, i, j, __LINE__) #define XFS_BMBT_TRACE_ARGFFFI(c,o,b,i,j) \ - xfs_bmbt_trace_argfffi(__FUNCTION__, c, o, b, i, j, __LINE__) + xfs_bmbt_trace_argfffi(__func__, c, o, b, i, j, __LINE__) #define XFS_BMBT_TRACE_ARGI(c,i) \ - xfs_bmbt_trace_argi(__FUNCTION__, c, i, __LINE__) + xfs_bmbt_trace_argi(__func__, c, i, __LINE__) #define XFS_BMBT_TRACE_ARGIFK(c,i,f,s) \ - xfs_bmbt_trace_argifk(__FUNCTION__, c, i, f, s, __LINE__) + xfs_bmbt_trace_argifk(__func__, c, i, f, s, __LINE__) #define XFS_BMBT_TRACE_ARGIFR(c,i,f,r) \ - xfs_bmbt_trace_argifr(__FUNCTION__, c, i, f, r, __LINE__) + xfs_bmbt_trace_argifr(__func__, c, i, f, r, __LINE__) #define XFS_BMBT_TRACE_ARGIK(c,i,k) \ - xfs_bmbt_trace_argik(__FUNCTION__, c, i, k, __LINE__) + xfs_bmbt_trace_argik(__func__, c, i, k, __LINE__) #define XFS_BMBT_TRACE_CURSOR(c,s) \ - xfs_bmbt_trace_cursor(__FUNCTION__, c, s, __LINE__) + xfs_bmbt_trace_cursor(__func__, c, s, __LINE__) #else #define XFS_BMBT_TRACE_ARGBI(c,b,i) #define XFS_BMBT_TRACE_ARGBII(c,b,i,j) diff --git a/fs/xfs/xfs_filestream.c b/fs/xfs/xfs_filestream.c index eb03eab5ca5..3f3785b1080 100644 --- a/fs/xfs/xfs_filestream.c +++ b/fs/xfs/xfs_filestream.c @@ -73,7 +73,7 @@ xfs_filestreams_trace( #define TRACE4(mp,t,a0,a1,a2,a3) TRACE6(mp,t,a0,a1,a2,a3,0,0) #define TRACE5(mp,t,a0,a1,a2,a3,a4) TRACE6(mp,t,a0,a1,a2,a3,a4,0) #define TRACE6(mp,t,a0,a1,a2,a3,a4,a5) \ - xfs_filestreams_trace(mp, t, __FUNCTION__, __LINE__, \ + xfs_filestreams_trace(mp, t, __func__, __LINE__, \ (__psunsigned_t)a0, (__psunsigned_t)a1, \ (__psunsigned_t)a2, (__psunsigned_t)a3, \ (__psunsigned_t)a4, (__psunsigned_t)a5) diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index 4a6f7c5d145..bece882f99e 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -2225,7 +2225,7 @@ xlog_state_do_callback( repeats = 0; xfs_fs_cmn_err(CE_WARN, log->l_mp, "%s: possible infinite loop (%d iterations)", - __FUNCTION__, flushcnt); + __func__, flushcnt); } } while (!ioerrors && loopdidcallbacks); diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 962d74a9ea7..c37521467fd 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -192,7 +192,7 @@ xlog_header_check_dump( { int b; - cmn_err(CE_DEBUG, "%s: SB : uuid = ", __FUNCTION__); + cmn_err(CE_DEBUG, "%s: SB : uuid = ", __func__); for (b = 0; b < 16; b++) cmn_err(CE_DEBUG, "%02x", ((uchar_t *)&mp->m_sb.sb_uuid)[b]); cmn_err(CE_DEBUG, ", fmt = %d\n", XLOG_FMT); @@ -3447,7 +3447,7 @@ xlog_valid_rec_header( (!rhead->h_version || (be32_to_cpu(rhead->h_version) & (~XLOG_VERSION_OKBITS))))) { xlog_warn("XFS: %s: unrecognised log version (%d).", - __FUNCTION__, be32_to_cpu(rhead->h_version)); + __func__, be32_to_cpu(rhead->h_version)); return XFS_ERROR(EIO); } diff --git a/fs/xfs/xfs_trans_ail.c b/fs/xfs/xfs_trans_ail.c index 13235ae9a58..1f77c00af56 100644 --- a/fs/xfs/xfs_trans_ail.c +++ b/fs/xfs/xfs_trans_ail.c @@ -438,7 +438,7 @@ xfs_trans_delete_ail( else { xfs_cmn_err(XFS_PTAG_AILDELETE, CE_ALERT, mp, "%s: attempting to delete a log item that is not in the AIL", - __FUNCTION__); + __func__); spin_unlock(&mp->m_ail_lock); xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE); } -- cgit v1.2.3 From b6ddc4e6fed9c6f4adb273c8b36e1731f90ec17e Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 10 Apr 2008 12:19:27 +1000 Subject: [XFS] Don't validate symlink target component length This target component validation is not POSIX conformant and it is not done by any other Linux filesystem so remove it from XFS. SGI-PV: 980080 SGI-Modid: xfs-linux-melb:xfs-kern:30776a Signed-off-by: Christoph Hellwig Signed-off-by: David Chinner Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_vnodeops.c | 25 ------------------------- 1 file changed, 25 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index 3418c94bcf1..d46f24c6849 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -3105,31 +3105,6 @@ xfs_symlink( pathlen = strlen(target_path); if (pathlen >= MAXPATHLEN) /* total string too long */ return XFS_ERROR(ENAMETOOLONG); - if (pathlen >= MAXNAMELEN) { /* is any component too long? */ - int len, total; - char *path; - - for (total = 0, path = target_path; total < pathlen;) { - /* - * Skip any slashes. - */ - while(*path == '/') { - total++; - path++; - } - - /* - * Count up to the next slash or end of path. - * Error out if the component is bigger than MAXNAMELEN. - */ - for(len = 0; *path != '/' && total < pathlen;total++, path++) { - if (++len >= MAXNAMELEN) { - error = ENAMETOOLONG; - return error; - } - } - } - } if (DM_EVENT_ENABLED(dp, DM_EVENT_SYMLINK)) { error = XFS_SEND_NAMESP(mp, DM_EVENT_SYMLINK, dp, -- cgit v1.2.3 From 3c85c36cc2e87018d38fcd033f41bbdf1360c07a Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:19:40 +1000 Subject: [XFS] xfs_quiesce_fs() never returns an error. Mark it void. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30780a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_vfsops.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_vfsops.c b/fs/xfs/xfs_vfsops.c index ea94593b531..6351efb569c 100644 --- a/fs/xfs/xfs_vfsops.c +++ b/fs/xfs/xfs_vfsops.c @@ -637,7 +637,7 @@ out: return XFS_ERROR(error); } -STATIC int +STATIC void xfs_quiesce_fs( xfs_mount_t *mp) { @@ -661,8 +661,6 @@ xfs_quiesce_fs( count++; } } while (count < 2); - - return 0; } /* -- cgit v1.2.3 From a414047fc97aea7db6237176ce00013117839cd5 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:19:47 +1000 Subject: [XFS] Remove useless whitespace in function prototypes Makes it simpler to annotate function prototypes with __must_check via sed scripts. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30781a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_utils.h | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_utils.h b/fs/xfs/xfs_utils.h index c4c4a6aa654..701accbbaea 100644 --- a/fs/xfs/xfs_utils.h +++ b/fs/xfs/xfs_utils.h @@ -21,14 +21,14 @@ #define IRELE(ip) VN_RELE(XFS_ITOV(ip)) #define IHOLD(ip) VN_HOLD(XFS_ITOV(ip)) -extern int xfs_dir_lookup_int (xfs_inode_t *, uint, bhv_vname_t *, xfs_ino_t *, +extern int xfs_dir_lookup_int(xfs_inode_t *, uint, bhv_vname_t *, xfs_ino_t *, xfs_inode_t **); -extern int xfs_truncate_file (xfs_mount_t *, xfs_inode_t *); -extern int xfs_dir_ialloc (xfs_trans_t **, xfs_inode_t *, mode_t, xfs_nlink_t, +extern int xfs_truncate_file(xfs_mount_t *, xfs_inode_t *); +extern int xfs_dir_ialloc(xfs_trans_t **, xfs_inode_t *, mode_t, xfs_nlink_t, xfs_dev_t, cred_t *, prid_t, int, xfs_inode_t **, int *); -extern int xfs_droplink (xfs_trans_t *, xfs_inode_t *); -extern int xfs_bumplink (xfs_trans_t *, xfs_inode_t *); -extern void xfs_bump_ino_vers2 (xfs_trans_t *, xfs_inode_t *); +extern int xfs_droplink(xfs_trans_t *, xfs_inode_t *); +extern int xfs_bumplink(xfs_trans_t *, xfs_inode_t *); +extern void xfs_bump_ino_vers2(xfs_trans_t *, xfs_inode_t *); #endif /* __XFS_UTILS_H__ */ -- cgit v1.2.3 From 36fbe6e6bd5408b09341043dfece978b4a7a0f34 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:19:56 +1000 Subject: [XFS] xfs_icsb_counter_disabled() never returns an error. Mark it void. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30782a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_mount.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index c2aafeb8c6c..eb348c16850 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -58,7 +58,7 @@ STATIC void xfs_icsb_balance_counter(xfs_mount_t *, xfs_sb_field_t, STATIC void xfs_icsb_sync_counters(xfs_mount_t *); STATIC int xfs_icsb_modify_counters(xfs_mount_t *, xfs_sb_field_t, int64_t, int); -STATIC int xfs_icsb_disable_counter(xfs_mount_t *, xfs_sb_field_t); +STATIC void xfs_icsb_disable_counter(xfs_mount_t *, xfs_sb_field_t); #else @@ -2183,7 +2183,7 @@ xfs_icsb_counter_disabled( return test_bit(field, &mp->m_icsb_counters); } -STATIC int +STATIC void xfs_icsb_disable_counter( xfs_mount_t *mp, xfs_sb_field_t field) @@ -2201,7 +2201,7 @@ xfs_icsb_disable_counter( * the m_icsb_mutex. */ if (xfs_icsb_counter_disabled(mp, field)) - return 0; + return; xfs_icsb_lock_all_counters(mp); if (!test_and_set_bit(field, &mp->m_icsb_counters)) { @@ -2224,8 +2224,6 @@ xfs_icsb_disable_counter( } xfs_icsb_unlock_all_counters(mp); - - return 0; } STATIC void -- cgit v1.2.3 From 714082bc12b6c305f825411df02177efcb0085f1 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:20:03 +1000 Subject: [XFS] Report errors from xfs_reserve_blocks(). xfs_reserve_blocks() can fail in interesting ways. In neither case is it a fatal error, but the result can lead to sub-optimal behaviour. Warn to the syslog if the call fails but otherwise continue. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30784a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_mount.c | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index eb348c16850..244aa1b9f13 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -1227,12 +1227,15 @@ xfs_mountfs( * * We default to 5% or 1024 fsbs of space reserved, whichever is smaller. * This may drive us straight to ENOSPC on mount, but that implies - * we were already there on the last unmount. + * we were already there on the last unmount. Warn if this occurs. */ resblks = mp->m_sb.sb_dblocks; do_div(resblks, 20); resblks = min_t(__uint64_t, resblks, 1024); - xfs_reserve_blocks(mp, &resblks, NULL); + error = xfs_reserve_blocks(mp, &resblks, NULL); + if (error) + cmn_err(CE_WARN, "XFS: Unable to allocate reserve blocks. " + "Continuing without a reserve pool."); return 0; @@ -1268,6 +1271,7 @@ int xfs_unmountfs(xfs_mount_t *mp, struct cred *cr) { __uint64_t resblks; + int error = 0; /* * We can potentially deadlock here if we have an inode cluster @@ -1311,7 +1315,11 @@ xfs_unmountfs(xfs_mount_t *mp, struct cred *cr) * value does not matter.... */ resblks = 0; - xfs_reserve_blocks(mp, &resblks, NULL); + error = xfs_reserve_blocks(mp, &resblks, NULL); + if (error) + cmn_err(CE_WARN, "XFS: Unable to free reserved block pool. " + "Freespace may not be correct on next mount."); + xfs_log_sbcount(mp, 1); xfs_unmountfs_writesb(mp); -- cgit v1.2.3 From 5b1397385bf536cbdb60f3362f44079d15d5f5ee Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:20:10 +1000 Subject: [XFS] xfs_qm_reset_dqcounts() does not return errors. Declare it void. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30785a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/quota/xfs_qm.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/xfs/quota/xfs_qm.c b/fs/xfs/quota/xfs_qm.c index adbc7bb9fba..dec5f95e847 100644 --- a/fs/xfs/quota/xfs_qm.c +++ b/fs/xfs/quota/xfs_qm.c @@ -1438,7 +1438,7 @@ xfs_qm_qino_alloc( } -STATIC int +STATIC void xfs_qm_reset_dqcounts( xfs_mount_t *mp, xfs_buf_t *bp, @@ -1478,8 +1478,6 @@ xfs_qm_reset_dqcounts( ddq->d_rtbwarns = 0; ddq = (xfs_disk_dquot_t *) ((xfs_dqblk_t *)ddq + 1); } - - return 0; } STATIC int @@ -1520,7 +1518,7 @@ xfs_qm_dqiter_bufs( if (error) break; - (void) xfs_qm_reset_dqcounts(mp, bp, firstid, type); + xfs_qm_reset_dqcounts(mp, bp, firstid, type); xfs_bdwrite(mp, bp); /* * goto the next block. -- cgit v1.2.3 From 4b8879df8c21bed3efd1eb2da5d72501199aba29 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:20:17 +1000 Subject: [XFS] Propagate xfs_qm_dqflush_all() errors. xfs_qm_dqflush_all() can return flush errors. Ensure they are propagated into the quotacheck code to determine if the quotacheck succeeded or not. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30786a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/quota/xfs_qm.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/xfs/quota/xfs_qm.c b/fs/xfs/quota/xfs_qm.c index dec5f95e847..04b29c67214 100644 --- a/fs/xfs/quota/xfs_qm.c +++ b/fs/xfs/quota/xfs_qm.c @@ -1877,6 +1877,14 @@ xfs_qm_quotacheck( } while (! done); + /* + * We've made all the changes that we need to make incore. + * Flush them down to disk buffers if everything was updated + * successfully. + */ + if (!error) + error = xfs_qm_dqflush_all(mp, XFS_QMOPT_DELWRI); + /* * We can get this error if we couldn't do a dquot allocation inside * xfs_qm_dqusage_adjust (via bulkstat). We don't care about the @@ -1888,11 +1896,6 @@ xfs_qm_quotacheck( xfs_qm_dqpurge_all(mp, XFS_QMOPT_QUOTALL | XFS_QMOPT_QUOTAOFF); goto error_return; } - /* - * We've made all the changes that we need to make incore. - * Now flush_them down to disk buffers. - */ - xfs_qm_dqflush_all(mp, XFS_QMOPT_DELWRI); /* * We didn't log anything, because if we crashed, we'll have to -- cgit v1.2.3 From 3c56836f92683cb871ebbf44c512069b0d48a08f Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:20:24 +1000 Subject: [XFS] Check for dquot flush errors xfs_qm_dqflush() can fail, but the return is not checked anywhere. Hence we never know if we've failed to flush a dquot to disk. Propagate the error and warn to the syslog if a flush ever fails. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30787a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/quota/xfs_dquot.c | 10 ++++++---- fs/xfs/quota/xfs_dquot_item.c | 7 ++++++- fs/xfs/quota/xfs_qm.c | 14 ++++++++++++-- 3 files changed, 24 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/xfs/quota/xfs_dquot.c b/fs/xfs/quota/xfs_dquot.c index 665babcca6a..15214fbb9aa 100644 --- a/fs/xfs/quota/xfs_dquot.c +++ b/fs/xfs/quota/xfs_dquot.c @@ -1439,9 +1439,7 @@ xfs_qm_dqpurge( uint flags) { xfs_dqhash_t *thishash; - xfs_mount_t *mp; - - mp = dqp->q_mount; + xfs_mount_t *mp = dqp->q_mount; ASSERT(XFS_QM_IS_MPLIST_LOCKED(mp)); ASSERT(XFS_DQ_IS_HASH_LOCKED(dqp->q_hash)); @@ -1485,6 +1483,7 @@ xfs_qm_dqpurge( * we're unmounting, we do care, so we flush it and wait. */ if (XFS_DQ_IS_DIRTY(dqp)) { + int error; xfs_dqtrace_entry(dqp, "DQPURGE ->DQFLUSH: DQDIRTY"); /* dqflush unlocks dqflock */ /* @@ -1495,7 +1494,10 @@ xfs_qm_dqpurge( * We don't care about getting disk errors here. We need * to purge this dquot anyway, so we go ahead regardless. */ - (void) xfs_qm_dqflush(dqp, XFS_QMOPT_SYNC); + error = xfs_qm_dqflush(dqp, XFS_QMOPT_SYNC); + if (error) + xfs_fs_cmn_err(CE_WARN, mp, + "xfs_qm_dqpurge: dquot %p flush failed", dqp); xfs_dqflock(dqp); } ASSERT(dqp->q_pincount == 0); diff --git a/fs/xfs/quota/xfs_dquot_item.c b/fs/xfs/quota/xfs_dquot_item.c index 1800e8d1f64..3dedce1d9cd 100644 --- a/fs/xfs/quota/xfs_dquot_item.c +++ b/fs/xfs/quota/xfs_dquot_item.c @@ -146,6 +146,7 @@ xfs_qm_dquot_logitem_push( xfs_dq_logitem_t *logitem) { xfs_dquot_t *dqp; + int error; dqp = logitem->qli_dquot; @@ -161,7 +162,11 @@ xfs_qm_dquot_logitem_push( * lock without sleeping, then there must not have been * anyone in the process of flushing the dquot. */ - xfs_qm_dqflush(dqp, XFS_B_DELWRI); + error = xfs_qm_dqflush(dqp, XFS_QMOPT_DELWRI); + if (error) + xfs_fs_cmn_err(CE_WARN, dqp->q_mount, + "xfs_qm_dquot_logitem_push: push error %d on dqp %p", + error, dqp); xfs_dqunlock(dqp); } diff --git a/fs/xfs/quota/xfs_qm.c b/fs/xfs/quota/xfs_qm.c index 04b29c67214..0ed3c8277fc 100644 --- a/fs/xfs/quota/xfs_qm.c +++ b/fs/xfs/quota/xfs_qm.c @@ -2094,12 +2094,17 @@ xfs_qm_shake_freelist( * dirty dquots. */ if (XFS_DQ_IS_DIRTY(dqp)) { + int error; xfs_dqtrace_entry(dqp, "DQSHAKE: DQDIRTY"); /* * We flush it delayed write, so don't bother * releasing the mplock. */ - (void) xfs_qm_dqflush(dqp, XFS_QMOPT_DELWRI); + error = xfs_qm_dqflush(dqp, XFS_QMOPT_DELWRI); + if (error) { + xfs_fs_cmn_err(CE_WARN, dqp->q_mount, + "xfs_qm_dqflush_all: dquot %p flush failed", dqp); + } xfs_dqunlock(dqp); /* dqflush unlocks dqflock */ dqp = dqp->dq_flnext; continue; @@ -2266,12 +2271,17 @@ xfs_qm_dqreclaim_one(void) * dirty dquots. */ if (XFS_DQ_IS_DIRTY(dqp)) { + int error; xfs_dqtrace_entry(dqp, "DQRECLAIM: DQDIRTY"); /* * We flush it delayed write, so don't bother * releasing the freelist lock. */ - (void) xfs_qm_dqflush(dqp, XFS_QMOPT_DELWRI); + error = xfs_qm_dqflush(dqp, XFS_QMOPT_DELWRI); + if (error) { + xfs_fs_cmn_err(CE_WARN, dqp->q_mount, + "xfs_qm_dqreclaim: dquot %p flush failed", dqp); + } xfs_dqunlock(dqp); /* dqflush unlocks dqflock */ continue; } -- cgit v1.2.3 From 53aa7915d67b9d0f5986c9f08e76846fedc520d4 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:20:31 +1000 Subject: [XFS] Clean up quotamount error handling. xfs_qm_mount_quotas() returns an error status that is ignored. If we fail to mount quotas, we continue with quota's turned off, which is all handled inside xfs_qm_mount_quotas(). Mark it as void to indicate that errors need not be returned to the callers. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30788a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/quota/xfs_qm.c | 28 ++++++++++++++-------------- fs/xfs/quota/xfs_qm.h | 2 +- 2 files changed, 15 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/xfs/quota/xfs_qm.c b/fs/xfs/quota/xfs_qm.c index 0ed3c8277fc..e15ee7cf3cc 100644 --- a/fs/xfs/quota/xfs_qm.c +++ b/fs/xfs/quota/xfs_qm.c @@ -304,8 +304,11 @@ xfs_qm_unmount_quotadestroy( * necessary data structures like quotainfo. This is also responsible for * running a quotacheck as necessary. We are guaranteed that the superblock * is consistently read in at this point. + * + * If we fail here, the mount will continue with quota turned off. We don't + * need to inidicate success or failure at all. */ -int +void xfs_qm_mount_quotas( xfs_mount_t *mp, int mfsi_flags) @@ -313,7 +316,6 @@ xfs_qm_mount_quotas( int error = 0; uint sbf; - /* * If quotas on realtime volumes is not supported, we disable * quotas immediately. @@ -332,7 +334,8 @@ xfs_qm_mount_quotas( * Allocate the quotainfo structure inside the mount struct, and * create quotainode(s), and change/rev superblock if necessary. */ - if ((error = xfs_qm_init_quotainfo(mp))) { + error = xfs_qm_init_quotainfo(mp); + if (error) { /* * We must turn off quotas. */ @@ -344,12 +347,11 @@ xfs_qm_mount_quotas( * If any of the quotas are not consistent, do a quotacheck. */ if (XFS_QM_NEED_QUOTACHECK(mp) && - !(mfsi_flags & XFS_MFSI_NO_QUOTACHECK)) { - if ((error = xfs_qm_quotacheck(mp))) { - /* Quotacheck has failed and quotas have - * been disabled. - */ - return XFS_ERROR(error); + !(mfsi_flags & XFS_MFSI_NO_QUOTACHECK)) { + error = xfs_qm_quotacheck(mp); + if (error) { + /* Quotacheck failed and disabled quotas. */ + return; } } /* @@ -357,12 +359,10 @@ xfs_qm_mount_quotas( * quotachecked status, since we won't be doing accounting for * that type anymore. */ - if (!XFS_IS_UQUOTA_ON(mp)) { + if (!XFS_IS_UQUOTA_ON(mp)) mp->m_qflags &= ~XFS_UQUOTA_CHKD; - } - if (!(XFS_IS_GQUOTA_ON(mp) || XFS_IS_PQUOTA_ON(mp))) { + if (!(XFS_IS_GQUOTA_ON(mp) || XFS_IS_PQUOTA_ON(mp))) mp->m_qflags &= ~XFS_OQUOTA_CHKD; - } write_changes: /* @@ -392,7 +392,7 @@ xfs_qm_mount_quotas( xfs_fs_cmn_err(CE_WARN, mp, "Failed to initialize disk quotas."); } - return XFS_ERROR(error); + return; } /* diff --git a/fs/xfs/quota/xfs_qm.h b/fs/xfs/quota/xfs_qm.h index baf537c1c17..cd2300e374a 100644 --- a/fs/xfs/quota/xfs_qm.h +++ b/fs/xfs/quota/xfs_qm.h @@ -165,7 +165,7 @@ typedef struct xfs_dquot_acct { #define XFS_QM_RELE(xqm) ((xqm)->qm_nrefs--) extern void xfs_qm_destroy_quotainfo(xfs_mount_t *); -extern int xfs_qm_mount_quotas(xfs_mount_t *, int); +extern void xfs_qm_mount_quotas(xfs_mount_t *, int); extern int xfs_qm_quotacheck(xfs_mount_t *); extern void xfs_qm_unmount_quotadestroy(xfs_mount_t *); extern int xfs_qm_unmount_quotas(xfs_mount_t *); -- cgit v1.2.3 From 31d5577b35d8397dea19f2ba7550e9225605a785 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:20:38 +1000 Subject: [XFS] Catch errors resetting quota flags. Warn to the syslog if we fail to reset the quota flags in the superblock when a quota check fails. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30789a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/quota/xfs_qm.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/quota/xfs_qm.c b/fs/xfs/quota/xfs_qm.c index e15ee7cf3cc..6aa3445caba 100644 --- a/fs/xfs/quota/xfs_qm.c +++ b/fs/xfs/quota/xfs_qm.c @@ -1927,7 +1927,10 @@ xfs_qm_quotacheck( ASSERT(mp->m_quotainfo != NULL); ASSERT(xfs_Gqm != NULL); xfs_qm_destroy_quotainfo(mp); - (void)xfs_mount_reset_sbqflags(mp); + if (xfs_mount_reset_sbqflags(mp)) { + cmn_err(CE_WARN, "XFS quotacheck %s: " + "Failed to reset quota flags.", mp->m_fsname); + } } else { cmn_err(CE_NOTE, "XFS quotacheck %s: Done.", mp->m_fsname); } -- cgit v1.2.3 From cb6edc26c386d2268dcf61bcdec02b6fb50b6ba2 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:20:45 +1000 Subject: [XFS] Catch errors when turning off quotas. When turning off quota, we need to write various transactions to the log to ensure that they are cleanly removed in the case of a crash. We need to check that the transactions hit the disk correctly. If we fail to write the final quota off transaction, we are corrupt in memory and so the only option is to shut the filesystem down at this point. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30790a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/quota/xfs_qm_syscalls.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/xfs/quota/xfs_qm_syscalls.c b/fs/xfs/quota/xfs_qm_syscalls.c index 3dc161f39d1..61cf68df547 100644 --- a/fs/xfs/quota/xfs_qm_syscalls.c +++ b/fs/xfs/quota/xfs_qm_syscalls.c @@ -279,9 +279,12 @@ xfs_qm_scall_quotaoff( /* * Write the LI_QUOTAOFF log record, and do SB changes atomically, - * and synchronously. + * and synchronously. If we fail to write, we should abort the + * operation as it cannot be recovered safely if we crash. */ - xfs_qm_log_quotaoff(mp, &qoffstart, flags); + error = xfs_qm_log_quotaoff(mp, &qoffstart, flags); + if (error) + goto out_error; /* * Next we clear the XFS_MOUNT_*DQ_ACTIVE bit(s) in the mount struct @@ -337,7 +340,12 @@ xfs_qm_scall_quotaoff( * So, we have QUOTAOFF start and end logitems; the start * logitem won't get overwritten until the end logitem appears... */ - xfs_qm_log_quotaoff_end(mp, qoffstart, flags); + error = xfs_qm_log_quotaoff_end(mp, qoffstart, flags); + if (error) { + /* We're screwed now. Shutdown is the only option. */ + xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE); + goto out_error; + } /* * If quotas is completely disabled, close shop. @@ -361,6 +369,7 @@ xfs_qm_scall_quotaoff( XFS_PURGE_INODE(XFS_QI_GQIP(mp)); XFS_QI_GQIP(mp) = NULL; } +out_error: mutex_unlock(&(XFS_QI_QOFFLOCK(mp))); return (error); -- cgit v1.2.3 From 88ab02085363b7c45935d66ab3e969b4fec9a20c Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:20:51 +1000 Subject: [XFS] Propagate quota file truncation errors. Truncating the quota files can silently fail. Ensure that truncation errors are propagated to the callers. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30791a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/quota/xfs_qm_syscalls.c | 15 +++++++-------- 1 file changed, 7 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/xfs/quota/xfs_qm_syscalls.c b/fs/xfs/quota/xfs_qm_syscalls.c index 61cf68df547..556018d24ca 100644 --- a/fs/xfs/quota/xfs_qm_syscalls.c +++ b/fs/xfs/quota/xfs_qm_syscalls.c @@ -380,12 +380,11 @@ xfs_qm_scall_trunc_qfiles( xfs_mount_t *mp, uint flags) { - int error; + int error = 0, error2 = 0; xfs_inode_t *qip; if (!capable(CAP_SYS_ADMIN)) return XFS_ERROR(EPERM); - error = 0; if (!xfs_sb_version_hasquota(&mp->m_sb) || flags == 0) { qdprintk("qtrunc flags=%x m_qflags=%x\n", flags, mp->m_qflags); return XFS_ERROR(EINVAL); @@ -393,22 +392,22 @@ xfs_qm_scall_trunc_qfiles( if ((flags & XFS_DQ_USER) && mp->m_sb.sb_uquotino != NULLFSINO) { error = xfs_iget(mp, NULL, mp->m_sb.sb_uquotino, 0, 0, &qip, 0); - if (! error) { - (void) xfs_truncate_file(mp, qip); + if (!error) { + error = xfs_truncate_file(mp, qip); IRELE(qip); } } if ((flags & (XFS_DQ_GROUP|XFS_DQ_PROJ)) && mp->m_sb.sb_gquotino != NULLFSINO) { - error = xfs_iget(mp, NULL, mp->m_sb.sb_gquotino, 0, 0, &qip, 0); - if (! error) { - (void) xfs_truncate_file(mp, qip); + error2 = xfs_iget(mp, NULL, mp->m_sb.sb_gquotino, 0, 0, &qip, 0); + if (!error2) { + error2 = xfs_truncate_file(mp, qip); IRELE(qip); } } - return (error); + return error ? error : error2; } -- cgit v1.2.3 From 0c928299676c8df2b00e75d5691cd4846e6c0868 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:20:58 +1000 Subject: [XFS] Catch errors from xfs_acl_setmode(). Propagate the error status from xfs_acl_setmode() so that callers know if the ACl was set correctly or not. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30792a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_acl.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_acl.c b/fs/xfs/xfs_acl.c index 7272fe39a92..98b515d3918 100644 --- a/fs/xfs/xfs_acl.c +++ b/fs/xfs/xfs_acl.c @@ -307,12 +307,13 @@ xfs_acl_vset( VN_HOLD(vp); error = xfs_acl_allow_set(vp, kind); - if (error) - goto out; /* Incoming ACL exists, set file mode based on its value */ - if (kind == _ACL_TYPE_ACCESS) - xfs_acl_setmode(vp, xfs_acl, &basicperms); + if (!error && kind == _ACL_TYPE_ACCESS) + error = xfs_acl_setmode(vp, xfs_acl, &basicperms); + + if (error) + goto out; /* * If we have more than std unix permissions, set up the actual attr. @@ -707,7 +708,9 @@ xfs_acl_inherit( memcpy(cacl, pdaclp, sizeof(xfs_acl_t)); xfs_acl_filter_mode(mode, cacl); - xfs_acl_setmode(vp, cacl, &basicperms); + error = xfs_acl_setmode(vp, cacl, &basicperms); + if (error) + goto out_error; /* * Set the Default and Access ACL on the file. The mode is already @@ -720,6 +723,7 @@ xfs_acl_inherit( xfs_acl_set_attr(vp, pdaclp, _ACL_TYPE_DEFAULT, &error); if (!error && !basicperms) xfs_acl_set_attr(vp, cacl, _ACL_TYPE_ACCESS, &error); +out_error: _ACL_FREE(cacl); return error; } -- cgit v1.2.3 From 5ca1f261a08d5cff5f29eaa0887b59baae2ae7f7 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:21:04 +1000 Subject: [XFS] Catch errors from xfs_acl_vremove(). Removing an ACL can return an error. Propagate it. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30793a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_acl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/xfs_acl.c b/fs/xfs/xfs_acl.c index 98b515d3918..8e130b9720a 100644 --- a/fs/xfs/xfs_acl.c +++ b/fs/xfs/xfs_acl.c @@ -324,7 +324,7 @@ xfs_acl_vset( if (!basicperms) { xfs_acl_set_attr(vp, xfs_acl, kind, &error); } else { - xfs_acl_vremove(vp, _ACL_TYPE_ACCESS); + error = -xfs_acl_vremove(vp, _ACL_TYPE_ACCESS); } out: -- cgit v1.2.3 From 3c1e2bbe5bcdcd435510a05eb121fa74b848e24f Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:21:11 +1000 Subject: [XFS] Propagate xfs_trans_reserve() errors. xfs_trans_reserve() reports errors that should not be ignored. For example, a shutdown filesystem will report errors through xfs_trans_reserve() to prevent further changes from being attempted on a damaged filesystem. Catch and propagate all error conditions from xfs_trans_reserve(). SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30794a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_log_recover.c | 35 ++++++++++++++++++++++++++--------- 1 file changed, 26 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index c37521467fd..957b8caddf1 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -2965,7 +2965,7 @@ xlog_recover_process_data( * Process an extent free intent item that was recovered from * the log. We need to free the extents that it describes. */ -STATIC void +STATIC int xlog_recover_process_efi( xfs_mount_t *mp, xfs_efi_log_item_t *efip) @@ -2973,6 +2973,7 @@ xlog_recover_process_efi( xfs_efd_log_item_t *efdp; xfs_trans_t *tp; int i; + int error = 0; xfs_extent_t *extp; xfs_fsblock_t startblock_fsb; @@ -2996,12 +2997,16 @@ xlog_recover_process_efi( * free the memory associated with it. */ xfs_efi_release(efip, efip->efi_format.efi_nextents); - return; + return XFS_ERROR(EIO); } } tp = xfs_trans_alloc(mp, 0); - xfs_trans_reserve(tp, 0, XFS_ITRUNCATE_LOG_RES(mp), 0, 0, 0); + error = xfs_trans_reserve(tp, 0, XFS_ITRUNCATE_LOG_RES(mp), 0, 0, 0); + if (error) { + xfs_trans_cancel(tp, XFS_TRANS_ABORT); + return error; + } efdp = xfs_trans_get_efd(tp, efip, efip->efi_format.efi_nextents); for (i = 0; i < efip->efi_format.efi_nextents; i++) { @@ -3013,6 +3018,7 @@ xlog_recover_process_efi( efip->efi_flags |= XFS_EFI_RECOVERED; xfs_trans_commit(tp, 0); + return error; } /* @@ -3060,7 +3066,7 @@ xlog_recover_check_ail( * everything already in the AIL, we stop processing as soon as * we see something other than an EFI in the AIL. */ -STATIC void +STATIC int xlog_recover_process_efis( xlog_t *log) { @@ -3068,6 +3074,7 @@ xlog_recover_process_efis( xfs_efi_log_item_t *efip; int gen; xfs_mount_t *mp; + int error = 0; mp = log->l_mp; spin_lock(&mp->m_ail_lock); @@ -3092,11 +3099,14 @@ xlog_recover_process_efis( } spin_unlock(&mp->m_ail_lock); - xlog_recover_process_efi(mp, efip); + error = xlog_recover_process_efi(mp, efip); + if (error) + return error; spin_lock(&mp->m_ail_lock); lip = xfs_trans_next_ail(mp, lip, &gen, NULL); } spin_unlock(&mp->m_ail_lock); + return error; } /* @@ -3116,9 +3126,9 @@ xlog_recover_clear_agi_bucket( int error; tp = xfs_trans_alloc(mp, XFS_TRANS_CLEAR_AGI_BUCKET); - xfs_trans_reserve(tp, 0, XFS_CLEAR_AGI_BUCKET_LOG_RES(mp), 0, 0, 0); - - error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, + error = xfs_trans_reserve(tp, 0, XFS_CLEAR_AGI_BUCKET_LOG_RES(mp), 0, 0, 0); + if (!error) + error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp)), XFS_FSS_TO_BB(mp, 1), 0, &agibp); if (error) { @@ -3919,7 +3929,14 @@ xlog_recover_finish( * rather than accepting new requests. */ if (log->l_flags & XLOG_RECOVERY_NEEDED) { - xlog_recover_process_efis(log); + int error; + error = xlog_recover_process_efis(log); + if (error) { + cmn_err(CE_ALERT, + "Failed to recover EFIs on filesystem: %s", + log->l_mp->m_fsname); + return error; + } /* * Sync the log to get all the EFIs out of the AIL. * This isn't absolutely necessary, but it helps in -- cgit v1.2.3 From e5720eec0548c08943d759e39db0388d8fe59287 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:21:18 +1000 Subject: [XFS] Propagate errors from xfs_trans_commit(). xfs_trans_commit() can return errors when there are problems in the transaction subsystem. They are indicative that the entire transaction may be incomplete, and hence the error should be propagated as there is a good possibility that there is something fatally wrong in the filesystem. Catch and propagate or warn about commit errors in the places where they are currently ignored. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30795a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/quota/xfs_qm.c | 4 ++-- fs/xfs/quota/xfs_qm_syscalls.c | 4 ++-- fs/xfs/xfs_inode.c | 52 ++++++++++++++++++------------------------ fs/xfs/xfs_log_recover.c | 27 ++++++++++++++-------- fs/xfs/xfs_mount.c | 35 +++++++++++++++++----------- fs/xfs/xfs_rtalloc.c | 38 ++++++++++++++++++------------ fs/xfs/xfs_vfsops.c | 15 +++++++++--- fs/xfs/xfs_vnodeops.c | 28 ++++++++++++----------- 8 files changed, 115 insertions(+), 88 deletions(-) (limited to 'fs') diff --git a/fs/xfs/quota/xfs_qm.c b/fs/xfs/quota/xfs_qm.c index 6aa3445caba..40ea5640956 100644 --- a/fs/xfs/quota/xfs_qm.c +++ b/fs/xfs/quota/xfs_qm.c @@ -2392,9 +2392,9 @@ xfs_qm_write_sb_changes( } xfs_mod_sb(tp, flags); - (void) xfs_trans_commit(tp, 0); + error = xfs_trans_commit(tp, 0); - return 0; + return error; } diff --git a/fs/xfs/quota/xfs_qm_syscalls.c b/fs/xfs/quota/xfs_qm_syscalls.c index 556018d24ca..8342823dbdc 100644 --- a/fs/xfs/quota/xfs_qm_syscalls.c +++ b/fs/xfs/quota/xfs_qm_syscalls.c @@ -734,12 +734,12 @@ xfs_qm_scall_setqlim( xfs_trans_log_dquot(tp, dqp); xfs_dqtrace_entry(dqp, "Q_SETQLIM: COMMIT"); - xfs_trans_commit(tp, 0); + error = xfs_trans_commit(tp, 0); xfs_qm_dqprint(dqp); xfs_qm_dqrele(dqp); mutex_unlock(&(XFS_QI_QOFFLOCK(mp))); - return (0); + return error; } STATIC int diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index d7514f8317d..63e66890f06 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -1699,33 +1699,16 @@ xfs_itruncate_finish( * blocks in the file system, but oh well. */ xfs_bmap_cancel(&free_list); - if (committed) { - /* - * If the passed in transaction committed - * in xfs_bmap_finish(), then we want to - * add the inode to this one before returning. - * This keeps things simple for the higher - * level code, because it always knows that - * the inode is locked and held in the - * transaction that returns to it whether - * errors occur or not. We don't mark the - * inode dirty so that this transaction can - * be easily aborted if possible. - */ - xfs_trans_ijoin(ntp, ip, - XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL); - xfs_trans_ihold(ntp, ip); - } + if (committed) + goto error_join; return error; } if (committed) { /* - * The first xact was committed, - * so add the inode to the new one. - * Mark it dirty so it will be logged - * and moved forward in the log as - * part of every commit. + * The first xact was committed, so add the inode to + * the new one. Mark it dirty so it will be logged and + * moved forward in the log as part of every commit. */ xfs_trans_ijoin(ntp, ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL); @@ -1733,19 +1716,16 @@ xfs_itruncate_finish( xfs_trans_log_inode(ntp, ip, XFS_ILOG_CORE); } ntp = xfs_trans_dup(ntp); - (void) xfs_trans_commit(*tp, 0); + error = xfs_trans_commit(*tp, 0); *tp = ntp; + if (error) + goto error_join; error = xfs_trans_reserve(ntp, 0, XFS_ITRUNCATE_LOG_RES(mp), 0, XFS_TRANS_PERM_LOG_RES, XFS_ITRUNCATE_LOG_COUNT); - /* - * Add the inode being truncated to the next chained - * transaction. - */ - xfs_trans_ijoin(ntp, ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL); - xfs_trans_ihold(ntp, ip); if (error) - return (error); + goto error_join; + } /* * Only update the size in the case of the data fork, but @@ -1777,6 +1757,18 @@ xfs_itruncate_finish( (ip->i_d.di_nextents == 0)); xfs_itrunc_trace(XFS_ITRUNC_FINISH2, ip, 0, new_size, 0, 0); return 0; + +error_join: + /* + * Add the inode being truncated to the next chained transaction. This + * keeps things simple for the higher level code, because it always + * knows that the inode is locked and held in the transaction that + * returns to it whether errors occur or not. We don't mark the inode + * dirty so that this transaction can be easily aborted if possible. + */ + xfs_trans_ijoin(ntp, ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL); + xfs_trans_ihold(ntp, ip); + return error; } diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 957b8caddf1..418582b709e 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -3017,7 +3017,7 @@ xlog_recover_process_efi( } efip->efi_flags |= XFS_EFI_RECOVERED; - xfs_trans_commit(tp, 0); + error = xfs_trans_commit(tp, 0); return error; } @@ -3131,16 +3131,13 @@ xlog_recover_clear_agi_bucket( error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, XFS_AG_DADDR(mp, agno, XFS_AGI_DADDR(mp)), XFS_FSS_TO_BB(mp, 1), 0, &agibp); - if (error) { - xfs_trans_cancel(tp, XFS_TRANS_ABORT); - return; - } + if (error) + goto out_abort; + error = EINVAL; agi = XFS_BUF_TO_AGI(agibp); - if (be32_to_cpu(agi->agi_magicnum) != XFS_AGI_MAGIC) { - xfs_trans_cancel(tp, XFS_TRANS_ABORT); - return; - } + if (be32_to_cpu(agi->agi_magicnum) != XFS_AGI_MAGIC) + goto out_abort; agi->agi_unlinked[bucket] = cpu_to_be32(NULLAGINO); offset = offsetof(xfs_agi_t, agi_unlinked) + @@ -3148,7 +3145,17 @@ xlog_recover_clear_agi_bucket( xfs_trans_log_buf(tp, agibp, offset, (offset + sizeof(xfs_agino_t) - 1)); - (void) xfs_trans_commit(tp, 0); + error = xfs_trans_commit(tp, 0); + if (error) + goto out_error; + return; + +out_abort: + xfs_trans_cancel(tp, XFS_TRANS_ABORT); +out_error: + xfs_fs_cmn_err(CE_WARN, mp, "xlog_recover_clear_agi_bucket: " + "failed to clear agi %d. Continuing.", agno); + return; } /* diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 244aa1b9f13..2d03fe194c2 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -45,7 +45,7 @@ #include "xfs_fsops.h" #include "xfs_utils.h" -STATIC void xfs_mount_log_sb(xfs_mount_t *, __int64_t); +STATIC int xfs_mount_log_sb(xfs_mount_t *, __int64_t); STATIC int xfs_uuid_mount(xfs_mount_t *); STATIC void xfs_uuid_unmount(xfs_mount_t *mp); STATIC void xfs_unmountfs_wait(xfs_mount_t *); @@ -1189,8 +1189,13 @@ xfs_mountfs( /* * If fs is not mounted readonly, then update the superblock changes. */ - if (update_flags && !(mp->m_flags & XFS_MOUNT_RDONLY)) - xfs_mount_log_sb(mp, update_flags); + if (update_flags && !(mp->m_flags & XFS_MOUNT_RDONLY)) { + error = xfs_mount_log_sb(mp, update_flags); + if (error) { + cmn_err(CE_WARN, "XFS: failed to write sb changes"); + goto error4; + } + } /* * Initialise the XFS quota management subsystem for this mount @@ -1320,8 +1325,10 @@ xfs_unmountfs(xfs_mount_t *mp, struct cred *cr) cmn_err(CE_WARN, "XFS: Unable to free reserved block pool. " "Freespace may not be correct on next mount."); - - xfs_log_sbcount(mp, 1); + error = xfs_log_sbcount(mp, 1); + if (error) + cmn_err(CE_WARN, "XFS: Unable to update superblock counters. " + "Freespace may not be correct on next mount."); xfs_unmountfs_writesb(mp); xfs_unmountfs_wait(mp); /* wait for async bufs */ xfs_log_unmount(mp); /* Done! No more fs ops. */ @@ -1413,9 +1420,8 @@ xfs_log_sbcount( xfs_mod_sb(tp, XFS_SB_IFREE | XFS_SB_ICOUNT | XFS_SB_FDBLOCKS); if (sync) xfs_trans_set_sync(tp); - xfs_trans_commit(tp, 0); - - return 0; + error = xfs_trans_commit(tp, 0); + return error; } STATIC void @@ -1913,24 +1919,27 @@ xfs_uuid_unmount( * be altered by the mount options, as well as any potential sb_features2 * fixup. Only the first superblock is updated. */ -STATIC void +STATIC int xfs_mount_log_sb( xfs_mount_t *mp, __int64_t fields) { xfs_trans_t *tp; + int error; ASSERT(fields & (XFS_SB_UNIT | XFS_SB_WIDTH | XFS_SB_UUID | XFS_SB_FEATURES2 | XFS_SB_BAD_FEATURES2)); tp = xfs_trans_alloc(mp, XFS_TRANS_SB_UNIT); - if (xfs_trans_reserve(tp, 0, mp->m_sb.sb_sectsize + 128, 0, 0, - XFS_DEFAULT_LOG_COUNT)) { + error = xfs_trans_reserve(tp, 0, mp->m_sb.sb_sectsize + 128, 0, 0, + XFS_DEFAULT_LOG_COUNT); + if (error) { xfs_trans_cancel(tp, 0); - return; + return error; } xfs_mod_sb(tp, fields); - xfs_trans_commit(tp, 0); + error = xfs_trans_commit(tp, 0); + return error; } diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c index 9cd6471cd60..a0dc6e5bc5b 100644 --- a/fs/xfs/xfs_rtalloc.c +++ b/fs/xfs/xfs_rtalloc.c @@ -124,14 +124,14 @@ xfs_growfs_rt_alloc( XFS_GROWRTALLOC_LOG_RES(mp), 0, XFS_TRANS_PERM_LOG_RES, XFS_DEFAULT_PERM_LOG_COUNT))) - goto error_exit; + goto error_cancel; cancelflags = XFS_TRANS_RELEASE_LOG_RES; /* * Lock the inode. */ if ((error = xfs_trans_iget(mp, tp, ino, 0, XFS_ILOCK_EXCL, &ip))) - goto error_exit; + goto error_cancel; XFS_BMAP_INIT(&flist, &firstblock); /* * Allocate blocks to the bitmap file. @@ -144,14 +144,16 @@ xfs_growfs_rt_alloc( if (!error && nmap < 1) error = XFS_ERROR(ENOSPC); if (error) - goto error_exit; + goto error_cancel; /* * Free any blocks freed up in the transaction, then commit. */ error = xfs_bmap_finish(&tp, &flist, &committed); if (error) - goto error_exit; - xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES); + goto error_cancel; + error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES); + if (error) + goto error; /* * Now we need to clear the allocated blocks. * Do this one block per transaction, to keep it simple. @@ -166,13 +168,13 @@ xfs_growfs_rt_alloc( */ if ((error = xfs_trans_reserve(tp, 0, XFS_GROWRTZERO_LOG_RES(mp), 0, 0, 0))) - goto error_exit; + goto error_cancel; /* * Lock the bitmap inode. */ if ((error = xfs_trans_iget(mp, tp, ino, 0, XFS_ILOCK_EXCL, &ip))) - goto error_exit; + goto error_cancel; /* * Get a buffer for the block. */ @@ -181,14 +183,16 @@ xfs_growfs_rt_alloc( mp->m_bsize, 0); if (bp == NULL) { error = XFS_ERROR(EIO); - goto error_exit; + goto error_cancel; } memset(XFS_BUF_PTR(bp), 0, mp->m_sb.sb_blocksize); xfs_trans_log_buf(tp, bp, 0, mp->m_sb.sb_blocksize - 1); /* * Commit the transaction. */ - xfs_trans_commit(tp, 0); + error = xfs_trans_commit(tp, 0); + if (error) + goto error; } /* * Go on to the next extent, if any. @@ -196,8 +200,9 @@ xfs_growfs_rt_alloc( oblocks = map.br_startoff + map.br_blockcount; } return 0; -error_exit: +error_cancel: xfs_trans_cancel(tp, cancelflags); +error: return error; } @@ -1876,6 +1881,7 @@ xfs_growfs_rt( xfs_trans_t *tp; /* transaction pointer */ sbp = &mp->m_sb; + cancelflags = 0; /* * Initial error checking. */ @@ -2042,13 +2048,15 @@ xfs_growfs_rt( */ mp->m_rsumlevels = nrsumlevels; mp->m_rsumsize = nrsumsize; - /* - * Commit the transaction. - */ - xfs_trans_commit(tp, 0); + + error = xfs_trans_commit(tp, 0); + if (error) { + tp = NULL; + break; + } } - if (error) + if (error && tp) xfs_trans_cancel(tp, cancelflags); /* diff --git a/fs/xfs/xfs_vfsops.c b/fs/xfs/xfs_vfsops.c index 6351efb569c..09e186d02c1 100644 --- a/fs/xfs/xfs_vfsops.c +++ b/fs/xfs/xfs_vfsops.c @@ -672,6 +672,8 @@ void xfs_attr_quiesce( xfs_mount_t *mp) { + int error = 0; + /* wait for all modifications to complete */ while (atomic_read(&mp->m_active_trans) > 0) delay(100); @@ -682,7 +684,11 @@ xfs_attr_quiesce( ASSERT_ALWAYS(atomic_read(&mp->m_active_trans) == 0); /* Push the superblock and write an unmount record */ - xfs_log_sbcount(mp, 1); + error = xfs_log_sbcount(mp, 1); + if (error) + xfs_fs_cmn_err(CE_WARN, mp, + "xfs_attr_quiesce: failed to log sb changes. " + "Frozen image may not be consistent."); xfs_log_unmount_write(mp); xfs_unmountfs_writesb(mp); } @@ -1316,8 +1322,11 @@ xfs_syncsub( * of sync if we crash or get a forced shutdown. We don't want to force * this to disk, just get a transaction into the iclogs.... */ - if (flags & SYNC_SUPER) - xfs_log_sbcount(mp, 0); + if (flags & SYNC_SUPER) { + error = xfs_log_sbcount(mp, 0); + if (error) + last_error = error; + } /* * Now check to see if the log needs a "dummy" transaction. diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index d46f24c6849..bc0a4707189 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -1447,28 +1447,22 @@ xfs_inactive_attrs( tp = *tpp; mp = ip->i_mount; ASSERT(ip->i_d.di_forkoff != 0); - xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES); + error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES); xfs_iunlock(ip, XFS_ILOCK_EXCL); + if (error) + goto error_unlock; error = xfs_attr_inactive(ip); - if (error) { - *tpp = NULL; - xfs_iunlock(ip, XFS_IOLOCK_EXCL); - return error; /* goto out */ - } + if (error) + goto error_unlock; tp = xfs_trans_alloc(mp, XFS_TRANS_INACTIVE); error = xfs_trans_reserve(tp, 0, XFS_IFREE_LOG_RES(mp), 0, XFS_TRANS_PERM_LOG_RES, XFS_INACTIVE_LOG_COUNT); - if (error) { - ASSERT(XFS_FORCED_SHUTDOWN(mp)); - xfs_trans_cancel(tp, 0); - *tpp = NULL; - xfs_iunlock(ip, XFS_IOLOCK_EXCL); - return error; - } + if (error) + goto error_cancel; xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL); @@ -1479,6 +1473,14 @@ xfs_inactive_attrs( *tpp = tp; return 0; + +error_cancel: + ASSERT(XFS_FORCED_SHUTDOWN(mp)); + xfs_trans_cancel(tp, 0); +error_unlock: + *tpp = NULL; + xfs_iunlock(ip, XFS_IOLOCK_EXCL); + return error; } int -- cgit v1.2.3 From f4586e40613a9f8bb9f7f9c8a796062a9ab1614c Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:21:25 +1000 Subject: [XFS] Clean up xfs_alloc_search_busy() return values. xfs_alloc_search_busy() returns an index into the busy array if the extent was found in the array. This is never checked, and the xfs_alloc_search_busy() does a log force to prevent reuse of the extent before the free transaction hits the disk. Hence the return value is useless. Declare the function void and remove the slot number from the tracing as well. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30796a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_alloc.c | 30 ++++++++++++------------------ 1 file changed, 12 insertions(+), 18 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c index bd5c01788ef..bd43f77daac 100644 --- a/fs/xfs/xfs_alloc.c +++ b/fs/xfs/xfs_alloc.c @@ -45,7 +45,7 @@ #define XFSA_FIXUP_BNO_OK 1 #define XFSA_FIXUP_CNT_OK 2 -STATIC int +STATIC void xfs_alloc_search_busy(xfs_trans_t *tp, xfs_agnumber_t agno, xfs_agblock_t bno, @@ -64,15 +64,15 @@ ktrace_t *xfs_alloc_trace_buf; xfs_alloc_trace_busy(__func__, s, mp, ag, agb, l, sl, tp, XFS_ALLOC_KTRACE_BUSY, __LINE__) #define TRACE_UNBUSY(__func__,s,ag,sl,tp) \ xfs_alloc_trace_busy(__func__, s, mp, ag, -1, -1, sl, tp, XFS_ALLOC_KTRACE_UNBUSY, __LINE__) -#define TRACE_BUSYSEARCH(__func__,s,ag,agb,l,sl,tp) \ - xfs_alloc_trace_busy(__func__, s, mp, ag, agb, l, sl, tp, XFS_ALLOC_KTRACE_BUSYSEARCH, __LINE__) +#define TRACE_BUSYSEARCH(__func__,s,ag,agb,l,tp) \ + xfs_alloc_trace_busy(__func__, s, mp, ag, agb, l, 0, tp, XFS_ALLOC_KTRACE_BUSYSEARCH, __LINE__) #else #define TRACE_ALLOC(s,a) #define TRACE_FREE(s,a,b,x,f) #define TRACE_MODAGF(s,a,f) #define TRACE_BUSY(s,a,ag,agb,l,sl,tp) #define TRACE_UNBUSY(fname,s,ag,sl,tp) -#define TRACE_BUSYSEARCH(fname,s,ag,agb,l,sl,tp) +#define TRACE_BUSYSEARCH(fname,s,ag,agb,l,tp) #endif /* XFS_ALLOC_TRACE */ /* @@ -2562,9 +2562,10 @@ xfs_alloc_clear_busy(xfs_trans_t *tp, /* - * returns non-zero if any of (agno,bno):len is in a busy list + * If we find the extent in the busy list, force the log out to get the + * extent out of the busy list so the caller can use it straight away. */ -STATIC int +STATIC void xfs_alloc_search_busy(xfs_trans_t *tp, xfs_agnumber_t agno, xfs_agblock_t bno, @@ -2572,7 +2573,6 @@ xfs_alloc_search_busy(xfs_trans_t *tp, { xfs_mount_t *mp; xfs_perag_busy_t *bsy; - int n; xfs_agblock_t uend, bend; xfs_lsn_t lsn; int cnt; @@ -2585,21 +2585,18 @@ xfs_alloc_search_busy(xfs_trans_t *tp, uend = bno + len - 1; /* search pagb_list for this slot, skipping open slots */ - for (bsy = mp->m_perag[agno].pagb_list, n = 0; - cnt; bsy++, n++) { + for (bsy = mp->m_perag[agno].pagb_list; cnt; bsy++) { /* * (start1,length1) within (start2, length2) */ if (bsy->busy_tp != NULL) { bend = bsy->busy_start + bsy->busy_length - 1; - if ((bno > bend) || - (uend < bsy->busy_start)) { + if ((bno > bend) || (uend < bsy->busy_start)) { cnt--; } else { TRACE_BUSYSEARCH("xfs_alloc_search_busy", - "found1", agno, bno, len, n, - tp); + "found1", agno, bno, len, tp); break; } } @@ -2610,15 +2607,12 @@ xfs_alloc_search_busy(xfs_trans_t *tp, * transaction that freed the block */ if (cnt) { - TRACE_BUSYSEARCH("xfs_alloc_search_busy", "found", agno, bno, len, n, tp); + TRACE_BUSYSEARCH("xfs_alloc_search_busy", "found", agno, bno, len, tp); lsn = bsy->busy_tp->t_commit_lsn; spin_unlock(&mp->m_perag[agno].pagb_lock); xfs_log_force(mp, lsn, XFS_LOG_FORCE|XFS_LOG_SYNC); } else { - TRACE_BUSYSEARCH("xfs_alloc_search_busy", "not-found", agno, bno, len, n, tp); - n = -1; + TRACE_BUSYSEARCH("xfs_alloc_search_busy", "not-found", agno, bno, len, tp); spin_unlock(&mp->m_perag[agno].pagb_lock); } - - return n; } -- cgit v1.2.3 From 12375c82375ec39ec948a3ad62e5e77533515e83 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:21:32 +1000 Subject: [XFS] Make xfs_alloc_compute_aligned() void. xfs_alloc_compute_aligned() returns a value based on a comparison of the computed extent length and the minimum length allowed. This is only used by some callers - the other four return parameters are used more often. Hence move the comparison to the code that actually needs to do it and make xfs_alloc_compute_aligned() a void function. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30797a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_alloc.c | 21 ++++++++++----------- 1 file changed, 10 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c index bd43f77daac..facdae14edd 100644 --- a/fs/xfs/xfs_alloc.c +++ b/fs/xfs/xfs_alloc.c @@ -93,7 +93,7 @@ STATIC int xfs_alloc_ag_vextent_small(xfs_alloc_arg_t *, * Compute aligned version of the found extent. * Takes alignment and min length into account. */ -STATIC int /* success (>= minlen) */ +STATIC void xfs_alloc_compute_aligned( xfs_agblock_t foundbno, /* starting block in found extent */ xfs_extlen_t foundlen, /* length in found extent */ @@ -116,7 +116,6 @@ xfs_alloc_compute_aligned( } *resbno = bno; *reslen = len; - return len >= minlen; } /* @@ -837,9 +836,9 @@ xfs_alloc_ag_vextent_near( if ((error = xfs_alloc_get_rec(cnt_cur, <bno, <len, &i))) goto error0; XFS_WANT_CORRUPTED_GOTO(i == 1, error0); - if (!xfs_alloc_compute_aligned(ltbno, ltlen, - args->alignment, args->minlen, - <bnoa, <lena)) + xfs_alloc_compute_aligned(ltbno, ltlen, args->alignment, + args->minlen, <bnoa, <lena); + if (ltlena >= args->minlen) continue; args->len = XFS_EXTLEN_MIN(ltlena, args->maxlen); xfs_alloc_fix_len(args); @@ -958,9 +957,9 @@ xfs_alloc_ag_vextent_near( if ((error = xfs_alloc_get_rec(bno_cur_lt, <bno, <len, &i))) goto error0; XFS_WANT_CORRUPTED_GOTO(i == 1, error0); - if (xfs_alloc_compute_aligned(ltbno, ltlen, - args->alignment, args->minlen, - <bnoa, <lena)) + xfs_alloc_compute_aligned(ltbno, ltlen, args->alignment, + args->minlen, <bnoa, <lena); + if (ltlena >= args->minlen) break; if ((error = xfs_alloc_decrement(bno_cur_lt, 0, &i))) goto error0; @@ -974,9 +973,9 @@ xfs_alloc_ag_vextent_near( if ((error = xfs_alloc_get_rec(bno_cur_gt, >bno, >len, &i))) goto error0; XFS_WANT_CORRUPTED_GOTO(i == 1, error0); - if (xfs_alloc_compute_aligned(gtbno, gtlen, - args->alignment, args->minlen, - >bnoa, >lena)) + xfs_alloc_compute_aligned(gtbno, gtlen, args->alignment, + args->minlen, >bnoa, >lena); + if (gtlena >= args->minlen) break; if ((error = xfs_alloc_increment(bno_cur_gt, 0, &i))) goto error0; -- cgit v1.2.3 From c2b1cba6833da77b1b478ac144f9cf5144d276ec Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:21:40 +1000 Subject: [XFS] xfs_bmap_adjacent() never returns an error. Mark it void. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30798a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_bmap.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 65b8fa83e07..6d9b5448deb 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -2402,7 +2402,7 @@ xfs_bmap_extsize_align( #define XFS_ALLOC_GAP_UNITS 4 -STATIC int +STATIC void xfs_bmap_adjacent( xfs_bmalloca_t *ap) /* bmap alloc argument struct */ { @@ -2548,7 +2548,6 @@ xfs_bmap_adjacent( ap->rval = gotbno; } #undef ISVALID - return 0; } STATIC int -- cgit v1.2.3 From d87dd6360dce86cad9099aed74f14b4dd0143301 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:21:46 +1000 Subject: [XFS] Warn if errors come from block_truncate_page(). block_truncate_page() can return errors that we currently ignore and silently discard. We should not ever get errors reported here - an error indicates a bug somewhere else. Hence catch the error and issue a stack dump to the syslog because we cannot propagate the error any further up the call chain. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30800a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_iops.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c index 53f8feb28e5..41e7baabfd9 100644 --- a/fs/xfs/linux-2.6/xfs_iops.c +++ b/fs/xfs/linux-2.6/xfs_iops.c @@ -692,11 +692,19 @@ xfs_vn_setattr( return -error; } +/* + * block_truncate_page can return an error, but we can't propagate it + * at all here. Leave a complaint + stack trace in the syslog because + * this could be bad. If it is bad, we need to propagate the error further. + */ STATIC void xfs_vn_truncate( struct inode *inode) { - block_truncate_page(inode->i_mapping, inode->i_size, xfs_get_blocks); + int error; + error = block_truncate_page(inode->i_mapping, inode->i_size, + xfs_get_blocks); + WARN_ON(error); } STATIC int -- cgit v1.2.3 From fc6149d8d9634814cdcd9283b8f2efd3359181df Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:21:53 +1000 Subject: [XFS] Check for xfs_free_extent() failing. xfs_free_extent() can fail, but log recovery never bothers to check if it successfully free the extent it was supposed to. This could lead to silent corruption during log recovery. Abort log recovery if we fail to free an extent. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30801a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_log_recover.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 418582b709e..3a8fe7bfa2a 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -3003,15 +3003,15 @@ xlog_recover_process_efi( tp = xfs_trans_alloc(mp, 0); error = xfs_trans_reserve(tp, 0, XFS_ITRUNCATE_LOG_RES(mp), 0, 0, 0); - if (error) { - xfs_trans_cancel(tp, XFS_TRANS_ABORT); - return error; - } + if (error) + goto abort_error; efdp = xfs_trans_get_efd(tp, efip, efip->efi_format.efi_nextents); for (i = 0; i < efip->efi_format.efi_nextents; i++) { extp = &(efip->efi_format.efi_extents[i]); - xfs_free_extent(tp, extp->ext_start, extp->ext_len); + error = xfs_free_extent(tp, extp->ext_start, extp->ext_len); + if (error) + goto abort_error; xfs_trans_log_efd_extent(tp, efdp, extp->ext_start, extp->ext_len); } @@ -3019,6 +3019,10 @@ xlog_recover_process_efi( efip->efi_flags |= XFS_EFI_RECOVERED; error = xfs_trans_commit(tp, 0); return error; + +abort_error: + xfs_trans_cancel(tp, XFS_TRANS_ABORT); + return error; } /* -- cgit v1.2.3 From 7c9ef85c5672ae316aafd7bbe0bbadebe90301e6 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:21:59 +1000 Subject: [XFS] Catch errors returned from xfs_bmap_last_offset(). xfs_bmap_last_offset() can fail and return an error. xfs_iomap_write_allocate() fails to detect and propagate the error. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30802a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_iomap.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c index fde37f87d52..fb3cf119141 100644 --- a/fs/xfs/xfs_iomap.c +++ b/fs/xfs/xfs_iomap.c @@ -802,8 +802,11 @@ xfs_iomap_write_allocate( */ nimaps = 1; end_fsb = XFS_B_TO_FSB(mp, ip->i_size); - xfs_bmap_last_offset(NULL, ip, &last_block, - XFS_DATA_FORK); + error = xfs_bmap_last_offset(NULL, ip, &last_block, + XFS_DATA_FORK); + if (error) + goto trans_cancel; + last_block = XFS_FILEOFF_MAX(last_block, end_fsb); if ((map_start_fsb + count_fsb) > last_block) { count_fsb = last_block - map_start_fsb; -- cgit v1.2.3 From 556b8b166c9514b5f940047a41dad8fe8cd9a778 Mon Sep 17 00:00:00 2001 From: Barry Naujok Date: Thu, 10 Apr 2008 12:22:07 +1000 Subject: [XFS] remove bhv_vname_t and xfs_rename code SGI-PV: 976035 SGI-Modid: xfs-linux-melb:xfs-kern:30804a Signed-off-by: Barry Naujok Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_export.c | 5 +- fs/xfs/linux-2.6/xfs_iops.c | 55 ++++++++++++++----- fs/xfs/linux-2.6/xfs_vnode.h | 9 ---- fs/xfs/xfs_dir2.c | 62 ++++++++++------------ fs/xfs/xfs_dir2.h | 12 +++-- fs/xfs/xfs_mount.h | 4 +- fs/xfs/xfs_rename.c | 82 ++++++++++------------------ fs/xfs/xfs_types.h | 5 ++ fs/xfs/xfs_utils.c | 4 +- fs/xfs/xfs_utils.h | 4 +- fs/xfs/xfs_vnodeops.c | 121 ++++++++++++++++++------------------------ fs/xfs/xfs_vnodeops.h | 23 ++++---- 12 files changed, 183 insertions(+), 203 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_export.c b/fs/xfs/linux-2.6/xfs_export.c index 66a9a9e76cb..265f0168ab7 100644 --- a/fs/xfs/linux-2.6/xfs_export.c +++ b/fs/xfs/linux-2.6/xfs_export.c @@ -22,6 +22,7 @@ #include "xfs_trans.h" #include "xfs_sb.h" #include "xfs_ag.h" +#include "xfs_dir2.h" #include "xfs_dmapi.h" #include "xfs_mount.h" #include "xfs_export.h" @@ -30,8 +31,6 @@ #include "xfs_inode.h" #include "xfs_vfsops.h" -static struct dentry dotdot = { .d_name.name = "..", .d_name.len = 2, }; - /* * Note that we only accept fileids which are long enough rather than allow * the parent generation number to default to zero. XFS considers zero a @@ -216,7 +215,7 @@ xfs_fs_get_parent( struct xfs_inode *cip; struct dentry *parent; - error = xfs_lookup(XFS_I(child->d_inode), &dotdot, &cip); + error = xfs_lookup(XFS_I(child->d_inode), &xfs_name_dotdot, &cip); if (unlikely(error)) return ERR_PTR(-error); diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c index 41e7baabfd9..0c958cf7775 100644 --- a/fs/xfs/linux-2.6/xfs_iops.c +++ b/fs/xfs/linux-2.6/xfs_iops.c @@ -239,6 +239,15 @@ xfs_init_security( return error; } +static void +xfs_dentry_to_name( + struct xfs_name *namep, + struct dentry *dentry) +{ + namep->name = dentry->d_name.name; + namep->len = dentry->d_name.len; +} + STATIC void xfs_cleanup_inode( struct inode *dir, @@ -246,20 +255,19 @@ xfs_cleanup_inode( struct dentry *dentry, int mode) { - struct dentry teardown = {}; + struct xfs_name teardown; /* Oh, the horror. * If we can't add the ACL or we fail in * xfs_init_security we must back out. * ENOSPC can hit here, among other things. */ - teardown.d_inode = inode; - teardown.d_name = dentry->d_name; + xfs_dentry_to_name(&teardown, dentry); if (S_ISDIR(mode)) - xfs_rmdir(XFS_I(dir), &teardown); + xfs_rmdir(XFS_I(dir), &teardown, XFS_I(inode)); else - xfs_remove(XFS_I(dir), &teardown); + xfs_remove(XFS_I(dir), &teardown, XFS_I(inode)); iput(inode); } @@ -273,6 +281,7 @@ xfs_vn_mknod( struct inode *inode; struct xfs_inode *ip = NULL; xfs_acl_t *default_acl = NULL; + struct xfs_name name; attrexists_t test_default_acl = _ACL_DEFAULT_EXISTS; int error; @@ -293,6 +302,8 @@ xfs_vn_mknod( } } + xfs_dentry_to_name(&name, dentry); + if (IS_POSIXACL(dir) && !default_acl) mode &= ~current->fs->umask; @@ -303,10 +314,10 @@ xfs_vn_mknod( case S_IFSOCK: rdev = sysv_encode_dev(rdev); case S_IFREG: - error = xfs_create(XFS_I(dir), dentry, mode, rdev, &ip, NULL); + error = xfs_create(XFS_I(dir), &name, mode, rdev, &ip, NULL); break; case S_IFDIR: - error = xfs_mkdir(XFS_I(dir), dentry, mode, &ip, NULL); + error = xfs_mkdir(XFS_I(dir), &name, mode, &ip, NULL); break; default: error = EINVAL; @@ -371,12 +382,14 @@ xfs_vn_lookup( struct nameidata *nd) { struct xfs_inode *cip; + struct xfs_name name; int error; if (dentry->d_name.len >= MAXNAMELEN) return ERR_PTR(-ENAMETOOLONG); - error = xfs_lookup(XFS_I(dir), dentry, &cip); + xfs_dentry_to_name(&name, dentry); + error = xfs_lookup(XFS_I(dir), &name, &cip); if (unlikely(error)) { if (unlikely(error != ENOENT)) return ERR_PTR(-error); @@ -394,12 +407,14 @@ xfs_vn_link( struct dentry *dentry) { struct inode *inode; /* inode of guy being linked to */ + struct xfs_name name; int error; inode = old_dentry->d_inode; + xfs_dentry_to_name(&name, dentry); igrab(inode); - error = xfs_link(XFS_I(dir), XFS_I(inode), dentry); + error = xfs_link(XFS_I(dir), XFS_I(inode), &name); if (unlikely(error)) { iput(inode); return -error; @@ -417,11 +432,13 @@ xfs_vn_unlink( struct dentry *dentry) { struct inode *inode; + struct xfs_name name; int error; inode = dentry->d_inode; + xfs_dentry_to_name(&name, dentry); - error = xfs_remove(XFS_I(dir), dentry); + error = xfs_remove(XFS_I(dir), &name, XFS_I(inode)); if (likely(!error)) { xfs_validate_fields(dir); /* size needs update */ xfs_validate_fields(inode); @@ -437,14 +454,15 @@ xfs_vn_symlink( { struct inode *inode; struct xfs_inode *cip = NULL; + struct xfs_name name; int error; mode_t mode; mode = S_IFLNK | (irix_symlink_mode ? 0777 & ~current->fs->umask : S_IRWXUGO); + xfs_dentry_to_name(&name, dentry); - error = xfs_symlink(XFS_I(dir), dentry, (char *)symname, mode, - &cip, NULL); + error = xfs_symlink(XFS_I(dir), &name, symname, mode, &cip, NULL); if (unlikely(error)) goto out; @@ -471,9 +489,12 @@ xfs_vn_rmdir( struct dentry *dentry) { struct inode *inode = dentry->d_inode; + struct xfs_name name; int error; - error = xfs_rmdir(XFS_I(dir), dentry); + xfs_dentry_to_name(&name, dentry); + + error = xfs_rmdir(XFS_I(dir), &name, XFS_I(inode)); if (likely(!error)) { xfs_validate_fields(inode); xfs_validate_fields(dir); @@ -489,9 +510,15 @@ xfs_vn_rename( struct dentry *ndentry) { struct inode *new_inode = ndentry->d_inode; + struct xfs_name oname; + struct xfs_name nname; int error; - error = xfs_rename(XFS_I(odir), odentry, XFS_I(ndir), ndentry); + xfs_dentry_to_name(&oname, odentry); + xfs_dentry_to_name(&nname, ndentry); + + error = xfs_rename(XFS_I(odir), &oname, XFS_I(odentry->d_inode), + XFS_I(ndir), &nname); if (likely(!error)) { if (new_inode) xfs_validate_fields(new_inode); diff --git a/fs/xfs/linux-2.6/xfs_vnode.h b/fs/xfs/linux-2.6/xfs_vnode.h index dbb8a5d27f7..8b4d63ce869 100644 --- a/fs/xfs/linux-2.6/xfs_vnode.h +++ b/fs/xfs/linux-2.6/xfs_vnode.h @@ -23,8 +23,6 @@ struct bhv_vattr; struct xfs_iomap; struct attrlist_cursor_kern; -typedef struct dentry bhv_vname_t; -typedef __u64 bhv_vnumber_t; typedef struct inode bhv_vnode_t; #define VN_ISLNK(vp) S_ISLNK((vp)->i_mode) @@ -210,13 +208,6 @@ static inline bhv_vnode_t *vn_grab(bhv_vnode_t *vp) return inode ? vn_from_inode(inode) : NULL; } -/* - * Vname handling macros. - */ -#define VNAME(dentry) ((char *) (dentry)->d_name.name) -#define VNAMELEN(dentry) ((dentry)->d_name.len) -#define VNAME_TO_INODE(dentry) (XFS_I((dentry)->d_inode)) - /* * Dealing with bad inodes */ diff --git a/fs/xfs/xfs_dir2.c b/fs/xfs/xfs_dir2.c index e92e73f0e6a..7cb26529766 100644 --- a/fs/xfs/xfs_dir2.c +++ b/fs/xfs/xfs_dir2.c @@ -44,6 +44,7 @@ #include "xfs_error.h" #include "xfs_vnodeops.h" +struct xfs_name xfs_name_dotdot = {"..", 2}; void xfs_dir_mount( @@ -146,8 +147,7 @@ int xfs_dir_createname( xfs_trans_t *tp, xfs_inode_t *dp, - char *name, - int namelen, + struct xfs_name *name, xfs_ino_t inum, /* new entry inode number */ xfs_fsblock_t *first, /* bmap's firstblock */ xfs_bmap_free_t *flist, /* bmap's freeblock list */ @@ -162,9 +162,9 @@ xfs_dir_createname( return rval; XFS_STATS_INC(xs_dir_create); - args.name = name; - args.namelen = namelen; - args.hashval = xfs_da_hashname(name, namelen); + args.name = name->name; + args.namelen = name->len; + args.hashval = xfs_da_hashname(name->name, name->len); args.inumber = inum; args.dp = dp; args.firstblock = first; @@ -197,8 +197,7 @@ int xfs_dir_lookup( xfs_trans_t *tp, xfs_inode_t *dp, - char *name, - int namelen, + struct xfs_name *name, xfs_ino_t *inum) /* out: inode number */ { xfs_da_args_t args; @@ -207,18 +206,14 @@ xfs_dir_lookup( ASSERT((dp->i_d.di_mode & S_IFMT) == S_IFDIR); XFS_STATS_INC(xs_dir_lookup); + memset(&args, 0, sizeof(xfs_da_args_t)); - args.name = name; - args.namelen = namelen; - args.hashval = xfs_da_hashname(name, namelen); - args.inumber = 0; + args.name = name->name; + args.namelen = name->len; + args.hashval = xfs_da_hashname(name->name, name->len); args.dp = dp; - args.firstblock = NULL; - args.flist = NULL; - args.total = 0; args.whichfork = XFS_DATA_FORK; args.trans = tp; - args.justcheck = args.addname = 0; args.oknoent = 1; if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL) @@ -247,8 +242,7 @@ int xfs_dir_removename( xfs_trans_t *tp, xfs_inode_t *dp, - char *name, - int namelen, + struct xfs_name *name, xfs_ino_t ino, xfs_fsblock_t *first, /* bmap's firstblock */ xfs_bmap_free_t *flist, /* bmap's freeblock list */ @@ -261,9 +255,9 @@ xfs_dir_removename( ASSERT((dp->i_d.di_mode & S_IFMT) == S_IFDIR); XFS_STATS_INC(xs_dir_remove); - args.name = name; - args.namelen = namelen; - args.hashval = xfs_da_hashname(name, namelen); + args.name = name->name; + args.namelen = name->len; + args.hashval = xfs_da_hashname(name->name, name->len); args.inumber = ino; args.dp = dp; args.firstblock = first; @@ -329,8 +323,7 @@ int xfs_dir_replace( xfs_trans_t *tp, xfs_inode_t *dp, - char *name, /* name of entry to replace */ - int namelen, + struct xfs_name *name, /* name of entry to replace */ xfs_ino_t inum, /* new inode number */ xfs_fsblock_t *first, /* bmap's firstblock */ xfs_bmap_free_t *flist, /* bmap's freeblock list */ @@ -345,9 +338,9 @@ xfs_dir_replace( if ((rval = xfs_dir_ino_validate(tp->t_mountp, inum))) return rval; - args.name = name; - args.namelen = namelen; - args.hashval = xfs_da_hashname(name, namelen); + args.name = name->name; + args.namelen = name->len; + args.hashval = xfs_da_hashname(name->name, name->len); args.inumber = inum; args.dp = dp; args.firstblock = first; @@ -374,28 +367,29 @@ xfs_dir_replace( /* * See if this entry can be added to the directory without allocating space. + * First checks that the caller couldn't reserve enough space (resblks = 0). */ int xfs_dir_canenter( xfs_trans_t *tp, xfs_inode_t *dp, - char *name, /* name of entry to add */ - int namelen) + struct xfs_name *name, /* name of entry to add */ + uint resblks) { xfs_da_args_t args; int rval; int v; /* type-checking value */ + if (resblks) + return 0; + ASSERT((dp->i_d.di_mode & S_IFMT) == S_IFDIR); + memset(&args, 0, sizeof(xfs_da_args_t)); - args.name = name; - args.namelen = namelen; - args.hashval = xfs_da_hashname(name, namelen); - args.inumber = 0; + args.name = name->name; + args.namelen = name->len; + args.hashval = xfs_da_hashname(name->name, name->len); args.dp = dp; - args.firstblock = NULL; - args.flist = NULL; - args.total = 0; args.whichfork = XFS_DATA_FORK; args.trans = tp; args.justcheck = args.addname = args.oknoent = 1; diff --git a/fs/xfs/xfs_dir2.h b/fs/xfs/xfs_dir2.h index b265197e74c..6392f939029 100644 --- a/fs/xfs/xfs_dir2.h +++ b/fs/xfs/xfs_dir2.h @@ -59,6 +59,8 @@ typedef __uint32_t xfs_dir2_db_t; */ typedef xfs_off_t xfs_dir2_off_t; +extern struct xfs_name xfs_name_dotdot; + /* * Generic directory interface routines */ @@ -68,21 +70,21 @@ extern int xfs_dir_isempty(struct xfs_inode *dp); extern int xfs_dir_init(struct xfs_trans *tp, struct xfs_inode *dp, struct xfs_inode *pdp); extern int xfs_dir_createname(struct xfs_trans *tp, struct xfs_inode *dp, - char *name, int namelen, xfs_ino_t inum, + struct xfs_name *name, xfs_ino_t inum, xfs_fsblock_t *first, struct xfs_bmap_free *flist, xfs_extlen_t tot); extern int xfs_dir_lookup(struct xfs_trans *tp, struct xfs_inode *dp, - char *name, int namelen, xfs_ino_t *inum); + struct xfs_name *name, xfs_ino_t *inum); extern int xfs_dir_removename(struct xfs_trans *tp, struct xfs_inode *dp, - char *name, int namelen, xfs_ino_t ino, + struct xfs_name *name, xfs_ino_t ino, xfs_fsblock_t *first, struct xfs_bmap_free *flist, xfs_extlen_t tot); extern int xfs_dir_replace(struct xfs_trans *tp, struct xfs_inode *dp, - char *name, int namelen, xfs_ino_t inum, + struct xfs_name *name, xfs_ino_t inum, xfs_fsblock_t *first, struct xfs_bmap_free *flist, xfs_extlen_t tot); extern int xfs_dir_canenter(struct xfs_trans *tp, struct xfs_inode *dp, - char *name, int namelen); + struct xfs_name *name, uint resblks); extern int xfs_dir_ino_validate(struct xfs_mount *mp, xfs_ino_t ino); /* diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index 77b39f66cea..1ed575110ff 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -73,7 +73,7 @@ typedef int (*xfs_send_destroy_t)(struct xfs_inode *, dm_right_t); typedef int (*xfs_send_namesp_t)(dm_eventtype_t, struct xfs_mount *, struct xfs_inode *, dm_right_t, struct xfs_inode *, dm_right_t, - char *, char *, mode_t, int, int); + const char *, const char *, mode_t, int, int); typedef int (*xfs_send_mount_t)(struct xfs_mount *, dm_right_t, char *, char *); typedef void (*xfs_send_unmount_t)(struct xfs_mount *, struct xfs_inode *, @@ -401,7 +401,7 @@ typedef struct xfs_mount { /* * Allow large block sizes to be reported to userspace programs if the - * "largeio" mount option is used. + * "largeio" mount option is used. * * If compatibility mode is specified, simply return the basic unit of caching * so that we don't get inefficient read/modify/write I/O from user apps. diff --git a/fs/xfs/xfs_rename.c b/fs/xfs/xfs_rename.c index c4d0bac56a5..ee371890d85 100644 --- a/fs/xfs/xfs_rename.c +++ b/fs/xfs/xfs_rename.c @@ -83,26 +83,23 @@ int xfs_rename_skip, xfs_rename_nskip; */ STATIC int xfs_lock_for_rename( - xfs_inode_t *dp1, /* old (source) directory inode */ - xfs_inode_t *dp2, /* new (target) directory inode */ - bhv_vname_t *vname1,/* old entry name */ - bhv_vname_t *vname2,/* new entry name */ - xfs_inode_t **ipp1, /* inode of old entry */ - xfs_inode_t **ipp2, /* inode of new entry, if it + xfs_inode_t *dp1, /* in: old (source) directory inode */ + xfs_inode_t *dp2, /* in: new (target) directory inode */ + xfs_inode_t *ip1, /* in: inode of old entry */ + struct xfs_name *name2, /* in: new entry name */ + xfs_inode_t **ipp2, /* out: inode of new entry, if it already exists, NULL otherwise. */ - xfs_inode_t **i_tab,/* array of inode returned, sorted */ - int *num_inodes) /* number of inodes in array */ + xfs_inode_t **i_tab,/* out: array of inode returned, sorted */ + int *num_inodes) /* out: number of inodes in array */ { - xfs_inode_t *ip1 = VNAME_TO_INODE(vname1); - xfs_inode_t *ip2, *temp; + xfs_inode_t *ip2 = NULL; + xfs_inode_t *temp; xfs_ino_t inum1, inum2; int error; int i, j; uint lock_mode; int diff_dirs = (dp1 != dp2); - ip2 = NULL; - /* * First, find out the current inums of the entries so that we * can determine the initial locking order. We'll have to @@ -115,17 +112,15 @@ xfs_lock_for_rename( inum1 = ip1->i_ino; - /* * Unlock dp1 and lock dp2 if they are different. */ - if (diff_dirs) { xfs_iunlock_map_shared(dp1, lock_mode); lock_mode = xfs_ilock_map_shared(dp2); } - error = xfs_dir_lookup_int(dp2, lock_mode, vname2, &inum2, &ip2); + error = xfs_dir_lookup_int(dp2, lock_mode, name2, &inum2, &ip2); if (error == ENOENT) { /* target does not need to exist. */ inum2 = 0; } else if (error) { @@ -157,6 +152,7 @@ xfs_lock_for_rename( *num_inodes = 4; i_tab[3] = ip2; } + *ipp2 = i_tab[3]; /* * Sort the elements via bubble sort. (Remember, there are at @@ -194,21 +190,6 @@ xfs_lock_for_rename( xfs_lock_inodes(i_tab, *num_inodes, 0, XFS_ILOCK_SHARED); } - /* - * Set the return value. Null out any unused entries in i_tab. - */ - *ipp1 = *ipp2 = NULL; - for (i=0; i < *num_inodes; i++) { - if (i_tab[i]->i_ino == inum1) { - *ipp1 = i_tab[i]; - } - if (i_tab[i]->i_ino == inum2) { - *ipp2 = i_tab[i]; - } - } - for (;i < 4; i++) { - i_tab[i] = NULL; - } return 0; } @@ -218,12 +199,13 @@ xfs_lock_for_rename( int xfs_rename( xfs_inode_t *src_dp, - bhv_vname_t *src_vname, + struct xfs_name *src_name, + xfs_inode_t *src_ip, xfs_inode_t *target_dp, - bhv_vname_t *target_vname) + struct xfs_name *target_name) { xfs_trans_t *tp; - xfs_inode_t *src_ip, *target_ip; + xfs_inode_t *target_ip; xfs_mount_t *mp = src_dp->i_mount; int new_parent; /* moving to a new dir */ int src_is_directory; /* src_name is a directory */ @@ -237,10 +219,6 @@ xfs_rename( int spaceres; int target_link_zero = 0; int num_inodes; - char *src_name = VNAME(src_vname); - char *target_name = VNAME(target_vname); - int src_namelen = VNAMELEN(src_vname); - int target_namelen = VNAMELEN(target_vname); xfs_itrace_entry(src_dp); xfs_itrace_entry(target_dp); @@ -250,7 +228,7 @@ xfs_rename( error = XFS_SEND_NAMESP(mp, DM_EVENT_RENAME, src_dp, DM_RIGHT_NULL, target_dp, DM_RIGHT_NULL, - src_name, target_name, + src_name->name, target_name->name, 0, 0, 0); if (error) { return error; @@ -267,10 +245,8 @@ xfs_rename( * does not exist in the source directory. */ tp = NULL; - error = xfs_lock_for_rename(src_dp, target_dp, src_vname, - target_vname, &src_ip, &target_ip, inodes, - &num_inodes); - + error = xfs_lock_for_rename(src_dp, target_dp, src_ip, target_name, + &target_ip, inodes, &num_inodes); if (error) { /* * We have nothing locked, no inode references, and @@ -316,7 +292,7 @@ xfs_rename( XFS_BMAP_INIT(&free_list, &first_block); tp = xfs_trans_alloc(mp, XFS_TRANS_RENAME); cancel_flags = XFS_TRANS_RELEASE_LOG_RES; - spaceres = XFS_RENAME_SPACE_RES(mp, target_namelen); + spaceres = XFS_RENAME_SPACE_RES(mp, target_name->len); error = xfs_trans_reserve(tp, spaceres, XFS_RENAME_LOG_RES(mp), 0, XFS_TRANS_PERM_LOG_RES, XFS_RENAME_LOG_COUNT); if (error == ENOSPC) { @@ -374,9 +350,8 @@ xfs_rename( * If there's no space reservation, check the entry will * fit before actually inserting it. */ - if (spaceres == 0 && - (error = xfs_dir_canenter(tp, target_dp, target_name, - target_namelen))) + error = xfs_dir_canenter(tp, target_dp, target_name, spaceres); + if (error) goto error_return; /* * If target does not exist and the rename crosses @@ -384,8 +359,8 @@ xfs_rename( * to account for the ".." reference from the new entry. */ error = xfs_dir_createname(tp, target_dp, target_name, - target_namelen, src_ip->i_ino, - &first_block, &free_list, spaceres); + src_ip->i_ino, &first_block, + &free_list, spaceres); if (error == ENOSPC) goto error_return; if (error) @@ -424,7 +399,7 @@ xfs_rename( * name at the destination directory, remove it first. */ error = xfs_dir_replace(tp, target_dp, target_name, - target_namelen, src_ip->i_ino, + src_ip->i_ino, &first_block, &free_list, spaceres); if (error) goto abort_return; @@ -461,7 +436,8 @@ xfs_rename( * Rewrite the ".." entry to point to the new * directory. */ - error = xfs_dir_replace(tp, src_ip, "..", 2, target_dp->i_ino, + error = xfs_dir_replace(tp, src_ip, &xfs_name_dotdot, + target_dp->i_ino, &first_block, &free_list, spaceres); ASSERT(error != EEXIST); if (error) @@ -497,8 +473,8 @@ xfs_rename( goto abort_return; } - error = xfs_dir_removename(tp, src_dp, src_name, src_namelen, - src_ip->i_ino, &first_block, &free_list, spaceres); + error = xfs_dir_removename(tp, src_dp, src_name, src_ip->i_ino, + &first_block, &free_list, spaceres); if (error) goto abort_return; xfs_ichgtime(src_dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); @@ -583,7 +559,7 @@ std_return: (void) XFS_SEND_NAMESP (mp, DM_EVENT_POSTRENAME, src_dp, DM_RIGHT_NULL, target_dp, DM_RIGHT_NULL, - src_name, target_name, + src_name->name, target_name->name, 0, error, 0); } return error; diff --git a/fs/xfs/xfs_types.h b/fs/xfs/xfs_types.h index 5c89be47546..0f5191644ab 100644 --- a/fs/xfs/xfs_types.h +++ b/fs/xfs/xfs_types.h @@ -160,4 +160,9 @@ typedef enum { XFS_BTNUM_MAX } xfs_btnum_t; +struct xfs_name { + const char *name; + int len; +}; + #endif /* __XFS_TYPES_H__ */ diff --git a/fs/xfs/xfs_utils.c b/fs/xfs/xfs_utils.c index 47c45ff4a06..2b8dc7e4077 100644 --- a/fs/xfs/xfs_utils.c +++ b/fs/xfs/xfs_utils.c @@ -45,7 +45,7 @@ int xfs_dir_lookup_int( xfs_inode_t *dp, uint lock_mode, - bhv_vname_t *dentry, + struct xfs_name *name, xfs_ino_t *inum, xfs_inode_t **ipp) { @@ -53,7 +53,7 @@ xfs_dir_lookup_int( xfs_itrace_entry(dp); - error = xfs_dir_lookup(NULL, dp, VNAME(dentry), VNAMELEN(dentry), inum); + error = xfs_dir_lookup(NULL, dp, name, inum); if (!error) { /* * Unlock the directory. We do this because we can't diff --git a/fs/xfs/xfs_utils.h b/fs/xfs/xfs_utils.h index 701accbbaea..175b126d2ca 100644 --- a/fs/xfs/xfs_utils.h +++ b/fs/xfs/xfs_utils.h @@ -21,8 +21,8 @@ #define IRELE(ip) VN_RELE(XFS_ITOV(ip)) #define IHOLD(ip) VN_HOLD(XFS_ITOV(ip)) -extern int xfs_dir_lookup_int(xfs_inode_t *, uint, bhv_vname_t *, xfs_ino_t *, - xfs_inode_t **); +extern int xfs_dir_lookup_int(xfs_inode_t *, uint, struct xfs_name *, + xfs_ino_t *, xfs_inode_t **); extern int xfs_truncate_file(xfs_mount_t *, xfs_inode_t *); extern int xfs_dir_ialloc(xfs_trans_t **, xfs_inode_t *, mode_t, xfs_nlink_t, xfs_dev_t, cred_t *, prid_t, int, diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index bc0a4707189..ca38fb9a993 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -1764,7 +1764,7 @@ xfs_inactive( int xfs_lookup( xfs_inode_t *dp, - bhv_vname_t *dentry, + struct xfs_name *name, xfs_inode_t **ipp) { xfs_inode_t *ip; @@ -1778,7 +1778,7 @@ xfs_lookup( return XFS_ERROR(EIO); lock_mode = xfs_ilock_map_shared(dp); - error = xfs_dir_lookup_int(dp, lock_mode, dentry, &e_inum, &ip); + error = xfs_dir_lookup_int(dp, lock_mode, name, &e_inum, &ip); if (!error) { *ipp = ip; xfs_itrace_ref(ip); @@ -1790,17 +1790,16 @@ xfs_lookup( int xfs_create( xfs_inode_t *dp, - bhv_vname_t *dentry, + struct xfs_name *name, mode_t mode, xfs_dev_t rdev, xfs_inode_t **ipp, cred_t *credp) { - char *name = VNAME(dentry); - xfs_mount_t *mp = dp->i_mount; + xfs_mount_t *mp = dp->i_mount; xfs_inode_t *ip; xfs_trans_t *tp; - int error; + int error; xfs_bmap_free_t free_list; xfs_fsblock_t first_block; boolean_t unlock_dp_on_error = B_FALSE; @@ -1810,17 +1809,14 @@ xfs_create( xfs_prid_t prid; struct xfs_dquot *udqp, *gdqp; uint resblks; - int namelen; ASSERT(!*ipp); xfs_itrace_entry(dp); - namelen = VNAMELEN(dentry); - if (DM_EVENT_ENABLED(dp, DM_EVENT_CREATE)) { error = XFS_SEND_NAMESP(mp, DM_EVENT_CREATE, dp, DM_RIGHT_NULL, NULL, - DM_RIGHT_NULL, name, NULL, + DM_RIGHT_NULL, name->name, NULL, mode, 0, 0); if (error) @@ -1852,7 +1848,7 @@ xfs_create( tp = xfs_trans_alloc(mp, XFS_TRANS_CREATE); cancel_flags = XFS_TRANS_RELEASE_LOG_RES; - resblks = XFS_CREATE_SPACE_RES(mp, namelen); + resblks = XFS_CREATE_SPACE_RES(mp, name->len); /* * Initially assume that the file does not exist and * reserve the resources for that case. If that is not @@ -1885,7 +1881,8 @@ xfs_create( if (error) goto error_return; - if (resblks == 0 && (error = xfs_dir_canenter(tp, dp, name, namelen))) + error = xfs_dir_canenter(tp, dp, name, resblks); + if (error) goto error_return; error = xfs_dir_ialloc(&tp, dp, mode, 1, rdev, credp, prid, resblks > 0, @@ -1915,7 +1912,7 @@ xfs_create( xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL); unlock_dp_on_error = B_FALSE; - error = xfs_dir_createname(tp, dp, name, namelen, ip->i_ino, + error = xfs_dir_createname(tp, dp, name, ip->i_ino, &first_block, &free_list, resblks ? resblks - XFS_IALLOC_SPACE_RES(mp) : 0); if (error) { @@ -1976,7 +1973,7 @@ std_return: (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTCREATE, dp, DM_RIGHT_NULL, *ipp ? ip : NULL, - DM_RIGHT_NULL, name, NULL, + DM_RIGHT_NULL, name->name, NULL, mode, error, 0); } return error; @@ -2268,12 +2265,10 @@ int remove_which_error_return = 0; int xfs_remove( xfs_inode_t *dp, - bhv_vname_t *dentry) + struct xfs_name *name, + xfs_inode_t *ip) { - char *name = VNAME(dentry); xfs_mount_t *mp = dp->i_mount; - xfs_inode_t *ip = VNAME_TO_INODE(dentry); - int namelen = VNAMELEN(dentry); xfs_trans_t *tp = NULL; int error = 0; xfs_bmap_free_t free_list; @@ -2289,9 +2284,9 @@ xfs_remove( return XFS_ERROR(EIO); if (DM_EVENT_ENABLED(dp, DM_EVENT_REMOVE)) { - error = XFS_SEND_NAMESP(mp, DM_EVENT_REMOVE, dp, - DM_RIGHT_NULL, NULL, DM_RIGHT_NULL, - name, NULL, ip->i_d.di_mode, 0, 0); + error = XFS_SEND_NAMESP(mp, DM_EVENT_REMOVE, dp, DM_RIGHT_NULL, + NULL, DM_RIGHT_NULL, name->name, NULL, + ip->i_d.di_mode, 0, 0); if (error) return error; } @@ -2376,7 +2371,7 @@ xfs_remove( * Entry must exist since we did a lookup in xfs_lock_dir_and_entry. */ XFS_BMAP_INIT(&free_list, &first_block); - error = xfs_dir_removename(tp, dp, name, namelen, ip->i_ino, + error = xfs_dir_removename(tp, dp, name, ip->i_ino, &first_block, &free_list, 0); if (error) { ASSERT(error != ENOENT); @@ -2444,7 +2439,7 @@ xfs_remove( (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTREMOVE, dp, DM_RIGHT_NULL, NULL, DM_RIGHT_NULL, - name, NULL, ip->i_d.di_mode, error, 0); + name->name, NULL, ip->i_d.di_mode, error, 0); } return error; @@ -2474,7 +2469,7 @@ int xfs_link( xfs_inode_t *tdp, xfs_inode_t *sip, - bhv_vname_t *dentry) + struct xfs_name *target_name) { xfs_mount_t *mp = tdp->i_mount; xfs_trans_t *tp; @@ -2485,13 +2480,10 @@ xfs_link( int cancel_flags; int committed; int resblks; - char *target_name = VNAME(dentry); - int target_namelen; xfs_itrace_entry(tdp); xfs_itrace_entry(sip); - target_namelen = VNAMELEN(dentry); ASSERT(!S_ISDIR(sip->i_d.di_mode)); if (XFS_FORCED_SHUTDOWN(mp)) @@ -2501,7 +2493,7 @@ xfs_link( error = XFS_SEND_NAMESP(mp, DM_EVENT_LINK, tdp, DM_RIGHT_NULL, sip, DM_RIGHT_NULL, - target_name, NULL, 0, 0, 0); + target_name->name, NULL, 0, 0, 0); if (error) return error; } @@ -2516,7 +2508,7 @@ xfs_link( tp = xfs_trans_alloc(mp, XFS_TRANS_LINK); cancel_flags = XFS_TRANS_RELEASE_LOG_RES; - resblks = XFS_LINK_SPACE_RES(mp, target_namelen); + resblks = XFS_LINK_SPACE_RES(mp, target_name->len); error = xfs_trans_reserve(tp, resblks, XFS_LINK_LOG_RES(mp), 0, XFS_TRANS_PERM_LOG_RES, XFS_LINK_LOG_COUNT); if (error == ENOSPC) { @@ -2568,15 +2560,14 @@ xfs_link( goto error_return; } - if (resblks == 0 && - (error = xfs_dir_canenter(tp, tdp, target_name, target_namelen))) + error = xfs_dir_canenter(tp, tdp, target_name, resblks); + if (error) goto error_return; XFS_BMAP_INIT(&free_list, &first_block); - error = xfs_dir_createname(tp, tdp, target_name, target_namelen, - sip->i_ino, &first_block, &free_list, - resblks); + error = xfs_dir_createname(tp, tdp, target_name, sip->i_ino, + &first_block, &free_list, resblks); if (error) goto abort_return; xfs_ichgtime(tdp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); @@ -2612,7 +2603,7 @@ std_return: (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTLINK, tdp, DM_RIGHT_NULL, sip, DM_RIGHT_NULL, - target_name, NULL, 0, error, 0); + target_name->name, NULL, 0, error, 0); } return error; @@ -2629,13 +2620,11 @@ std_return: int xfs_mkdir( xfs_inode_t *dp, - bhv_vname_t *dentry, + struct xfs_name *dir_name, mode_t mode, xfs_inode_t **ipp, cred_t *credp) { - char *dir_name = VNAME(dentry); - int dir_namelen = VNAMELEN(dentry); xfs_mount_t *mp = dp->i_mount; xfs_inode_t *cdp; /* inode of created dir */ xfs_trans_t *tp; @@ -2659,7 +2648,7 @@ xfs_mkdir( if (DM_EVENT_ENABLED(dp, DM_EVENT_CREATE)) { error = XFS_SEND_NAMESP(mp, DM_EVENT_CREATE, dp, DM_RIGHT_NULL, NULL, - DM_RIGHT_NULL, dir_name, NULL, + DM_RIGHT_NULL, dir_name->name, NULL, mode, 0, 0); if (error) return error; @@ -2688,7 +2677,7 @@ xfs_mkdir( tp = xfs_trans_alloc(mp, XFS_TRANS_MKDIR); cancel_flags = XFS_TRANS_RELEASE_LOG_RES; - resblks = XFS_MKDIR_SPACE_RES(mp, dir_namelen); + resblks = XFS_MKDIR_SPACE_RES(mp, dir_name->len); error = xfs_trans_reserve(tp, resblks, XFS_MKDIR_LOG_RES(mp), 0, XFS_TRANS_PERM_LOG_RES, XFS_MKDIR_LOG_COUNT); if (error == ENOSPC) { @@ -2720,8 +2709,8 @@ xfs_mkdir( if (error) goto error_return; - if (resblks == 0 && - (error = xfs_dir_canenter(tp, dp, dir_name, dir_namelen))) + error = xfs_dir_canenter(tp, dp, dir_name, resblks); + if (error) goto error_return; /* * create the directory inode. @@ -2750,9 +2739,9 @@ xfs_mkdir( XFS_BMAP_INIT(&free_list, &first_block); - error = xfs_dir_createname(tp, dp, dir_name, dir_namelen, cdp->i_ino, - &first_block, &free_list, resblks ? - resblks - XFS_IALLOC_SPACE_RES(mp) : 0); + error = xfs_dir_createname(tp, dp, dir_name, cdp->i_ino, + &first_block, &free_list, resblks ? + resblks - XFS_IALLOC_SPACE_RES(mp) : 0); if (error) { ASSERT(error != ENOSPC); goto error1; @@ -2817,7 +2806,7 @@ std_return: dp, DM_RIGHT_NULL, created ? cdp : NULL, DM_RIGHT_NULL, - dir_name, NULL, + dir_name->name, NULL, mode, error, 0); } return error; @@ -2841,13 +2830,11 @@ std_return: int xfs_rmdir( xfs_inode_t *dp, - bhv_vname_t *dentry) + struct xfs_name *name, + xfs_inode_t *cdp) { bhv_vnode_t *dir_vp = XFS_ITOV(dp); - char *name = VNAME(dentry); - int namelen = VNAMELEN(dentry); xfs_mount_t *mp = dp->i_mount; - xfs_inode_t *cdp = VNAME_TO_INODE(dentry); xfs_trans_t *tp; int error; xfs_bmap_free_t free_list; @@ -2865,8 +2852,8 @@ xfs_rmdir( if (DM_EVENT_ENABLED(dp, DM_EVENT_REMOVE)) { error = XFS_SEND_NAMESP(mp, DM_EVENT_REMOVE, dp, DM_RIGHT_NULL, - NULL, DM_RIGHT_NULL, - name, NULL, cdp->i_d.di_mode, 0, 0); + NULL, DM_RIGHT_NULL, name->name, + NULL, cdp->i_d.di_mode, 0, 0); if (error) return XFS_ERROR(error); } @@ -2960,7 +2947,7 @@ xfs_rmdir( goto error_return; } - error = xfs_dir_removename(tp, dp, name, namelen, cdp->i_ino, + error = xfs_dir_removename(tp, dp, name, cdp->i_ino, &first_block, &free_list, resblks); if (error) goto error1; @@ -3040,7 +3027,7 @@ xfs_rmdir( (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTREMOVE, dp, DM_RIGHT_NULL, NULL, DM_RIGHT_NULL, - name, NULL, cdp->i_d.di_mode, + name->name, NULL, cdp->i_d.di_mode, error, 0); } return error; @@ -3058,8 +3045,8 @@ xfs_rmdir( int xfs_symlink( xfs_inode_t *dp, - bhv_vname_t *dentry, - char *target_path, + struct xfs_name *link_name, + const char *target_path, mode_t mode, xfs_inode_t **ipp, cred_t *credp) @@ -3079,15 +3066,13 @@ xfs_symlink( int nmaps; xfs_bmbt_irec_t mval[SYMLINK_MAPS]; xfs_daddr_t d; - char *cur_chunk; + const char *cur_chunk; int byte_cnt; int n; xfs_buf_t *bp; xfs_prid_t prid; struct xfs_dquot *udqp, *gdqp; uint resblks; - char *link_name = VNAME(dentry); - int link_namelen; *ipp = NULL; error = 0; @@ -3099,8 +3084,6 @@ xfs_symlink( if (XFS_FORCED_SHUTDOWN(mp)) return XFS_ERROR(EIO); - link_namelen = VNAMELEN(dentry); - /* * Check component lengths of the target path name. */ @@ -3111,7 +3094,7 @@ xfs_symlink( if (DM_EVENT_ENABLED(dp, DM_EVENT_SYMLINK)) { error = XFS_SEND_NAMESP(mp, DM_EVENT_SYMLINK, dp, DM_RIGHT_NULL, NULL, DM_RIGHT_NULL, - link_name, target_path, 0, 0, 0); + link_name->name, target_path, 0, 0, 0); if (error) return error; } @@ -3143,7 +3126,7 @@ xfs_symlink( fs_blocks = 0; else fs_blocks = XFS_B_TO_FSB(mp, pathlen); - resblks = XFS_SYMLINK_SPACE_RES(mp, link_namelen, fs_blocks); + resblks = XFS_SYMLINK_SPACE_RES(mp, link_name->len, fs_blocks); error = xfs_trans_reserve(tp, resblks, XFS_SYMLINK_LOG_RES(mp), 0, XFS_TRANS_PERM_LOG_RES, XFS_SYMLINK_LOG_COUNT); if (error == ENOSPC && fs_blocks == 0) { @@ -3177,8 +3160,8 @@ xfs_symlink( /* * Check for ability to enter directory entry, if no space reserved. */ - if (resblks == 0 && - (error = xfs_dir_canenter(tp, dp, link_name, link_namelen))) + error = xfs_dir_canenter(tp, dp, link_name, resblks); + if (error) goto error_return; /* * Initialize the bmap freelist prior to calling either @@ -3270,8 +3253,8 @@ xfs_symlink( /* * Create the directory entry for the symlink. */ - error = xfs_dir_createname(tp, dp, link_name, link_namelen, ip->i_ino, - &first_block, &free_list, resblks); + error = xfs_dir_createname(tp, dp, link_name, ip->i_ino, + &first_block, &free_list, resblks); if (error) goto error1; xfs_ichgtime(dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); @@ -3315,8 +3298,8 @@ std_return: (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTSYMLINK, dp, DM_RIGHT_NULL, error ? NULL : ip, - DM_RIGHT_NULL, link_name, target_path, - 0, error, 0); + DM_RIGHT_NULL, link_name->name, + target_path, 0, error, 0); } if (!error) diff --git a/fs/xfs/xfs_vnodeops.h b/fs/xfs/xfs_vnodeops.h index 12e581865bd..24c53923dc2 100644 --- a/fs/xfs/xfs_vnodeops.h +++ b/fs/xfs/xfs_vnodeops.h @@ -23,20 +23,22 @@ int xfs_fsync(struct xfs_inode *ip, int flag, xfs_off_t start, xfs_off_t stop); int xfs_release(struct xfs_inode *ip); int xfs_inactive(struct xfs_inode *ip); -int xfs_lookup(struct xfs_inode *dp, bhv_vname_t *dentry, +int xfs_lookup(struct xfs_inode *dp, struct xfs_name *name, struct xfs_inode **ipp); -int xfs_create(struct xfs_inode *dp, bhv_vname_t *dentry, mode_t mode, +int xfs_create(struct xfs_inode *dp, struct xfs_name *name, mode_t mode, xfs_dev_t rdev, struct xfs_inode **ipp, struct cred *credp); -int xfs_remove(struct xfs_inode *dp, bhv_vname_t *dentry); +int xfs_remove(struct xfs_inode *dp, struct xfs_name *name, + struct xfs_inode *ip); int xfs_link(struct xfs_inode *tdp, struct xfs_inode *sip, - bhv_vname_t *dentry); -int xfs_mkdir(struct xfs_inode *dp, bhv_vname_t *dentry, + struct xfs_name *target_name); +int xfs_mkdir(struct xfs_inode *dp, struct xfs_name *dir_name, mode_t mode, struct xfs_inode **ipp, struct cred *credp); -int xfs_rmdir(struct xfs_inode *dp, bhv_vname_t *dentry); +int xfs_rmdir(struct xfs_inode *dp, struct xfs_name *name, + struct xfs_inode *cdp); int xfs_readdir(struct xfs_inode *dp, void *dirent, size_t bufsize, xfs_off_t *offset, filldir_t filldir); -int xfs_symlink(struct xfs_inode *dp, bhv_vname_t *dentry, - char *target_path, mode_t mode, struct xfs_inode **ipp, +int xfs_symlink(struct xfs_inode *dp, struct xfs_name *link_name, + const char *target_path, mode_t mode, struct xfs_inode **ipp, struct cred *credp); int xfs_inode_flush(struct xfs_inode *ip, int flags); int xfs_set_dmattrs(struct xfs_inode *ip, u_int evmask, u_int16_t state); @@ -44,8 +46,9 @@ int xfs_reclaim(struct xfs_inode *ip); int xfs_change_file_space(struct xfs_inode *ip, int cmd, xfs_flock64_t *bf, xfs_off_t offset, struct cred *credp, int attr_flags); -int xfs_rename(struct xfs_inode *src_dp, bhv_vname_t *src_vname, - struct xfs_inode *target_dp, bhv_vname_t *target_vname); +int xfs_rename(struct xfs_inode *src_dp, struct xfs_name *src_name, + struct xfs_inode *src_ip, struct xfs_inode *target_dp, + struct xfs_name *target_name); int xfs_attr_get(struct xfs_inode *ip, const char *name, char *value, int *valuelenp, int flags, cred_t *cred); int xfs_attr_set(struct xfs_inode *dp, const char *name, char *value, -- cgit v1.2.3 From d64e31a2f53cdcb2f95b782196faacb0995ca0c0 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:22:17 +1000 Subject: [XFS] Ensure errors from xfs_bdstrat() are correctly checked. xfsbdstrat() is declared to return an error. That is never checked because the error is propagated by the xfs_buf_t that is passed through the function. Mark xfsbdstrat() as returning void and comment the prototype on the methods needed for error checking. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30823a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_lrw.c | 19 ++++++------------- fs/xfs/linux-2.6/xfs_lrw.h | 3 ++- fs/xfs/xfs_log_recover.c | 6 ++++-- fs/xfs/xfs_mount.c | 1 - fs/xfs/xfs_trans_buf.c | 12 +++++------- fs/xfs/xfs_vnodeops.c | 6 ++++-- 6 files changed, 21 insertions(+), 26 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_lrw.c b/fs/xfs/linux-2.6/xfs_lrw.c index 1d95dca96cf..f6dab5d8944 100644 --- a/fs/xfs/linux-2.6/xfs_lrw.c +++ b/fs/xfs/linux-2.6/xfs_lrw.c @@ -875,28 +875,21 @@ xfs_bdstrat_cb(struct xfs_buf *bp) } /* - * Wrapper around bdstrat so that we can stop data - * from going to disk in case we are shutting down the filesystem. - * Typically user data goes thru this path; one of the exceptions - * is the superblock. + * Wrapper around bdstrat so that we can stop data from going to disk in case + * we are shutting down the filesystem. Typically user data goes thru this + * path; one of the exceptions is the superblock. */ -int +void xfsbdstrat( struct xfs_mount *mp, struct xfs_buf *bp) { ASSERT(mp); - if (!XFS_FORCED_SHUTDOWN(mp)) { - /* Grio redirection would go here - * if (XFS_BUF_IS_GRIO(bp)) { - */ - + if (!XFS_FORCED_SHUTDOWN(mp)) xfs_buf_iorequest(bp); - return 0; - } xfs_buftrace("XFSBDSTRAT IOERROR", bp); - return (xfs_bioerror_relse(bp)); + xfs_bioerror_relse(bp); } /* diff --git a/fs/xfs/linux-2.6/xfs_lrw.h b/fs/xfs/linux-2.6/xfs_lrw.h index e200253139c..e1d498b4ba7 100644 --- a/fs/xfs/linux-2.6/xfs_lrw.h +++ b/fs/xfs/linux-2.6/xfs_lrw.h @@ -68,7 +68,8 @@ extern void xfs_inval_cached_trace(struct xfs_inode *, #define xfs_inval_cached_trace(ip, offset, len, first, last) #endif -extern int xfsbdstrat(struct xfs_mount *, struct xfs_buf *); +/* errors from xfsbdstrat() must be extracted from the buffer */ +extern void xfsbdstrat(struct xfs_mount *, struct xfs_buf *); extern int xfs_bdstrat_cb(struct xfs_buf *); extern int xfs_dev_is_read_only(struct xfs_mount *, char *); diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 3a8fe7bfa2a..1f83298f90a 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -121,7 +121,8 @@ xlog_bread( XFS_BUF_SET_TARGET(bp, log->l_mp->m_logdev_targp); xfsbdstrat(log->l_mp, bp); - if ((error = xfs_iowait(bp))) + error = xfs_iowait(bp); + if (error) xfs_ioerror_alert("xlog_bread", log->l_mp, bp, XFS_BUF_ADDR(bp)); return error; @@ -3849,7 +3850,8 @@ xlog_do_recover( XFS_BUF_READ(bp); XFS_BUF_UNASYNC(bp); xfsbdstrat(log->l_mp, bp); - if ((error = xfs_iowait(bp))) { + error = xfs_iowait(bp); + if (error) { xfs_ioerror_alert("xlog_do_recover", log->l_mp, bp, XFS_BUF_ADDR(bp)); ASSERT(0); diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 2d03fe194c2..2fec452afbc 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -1470,7 +1470,6 @@ xfs_unmountfs_writesb(xfs_mount_t *mp) XFS_BUF_UNASYNC(sbp); ASSERT(XFS_BUF_TARGET(sbp) == mp->m_ddev_targp); xfsbdstrat(mp, sbp); - /* Nevermind errors we might get here. */ error = xfs_iowait(sbp); if (error) xfs_ioerror_alert("xfs_unmountfs_writesb", diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c index 4e5c010f504..cb0c5839154 100644 --- a/fs/xfs/xfs_trans_buf.c +++ b/fs/xfs/xfs_trans_buf.c @@ -354,17 +354,15 @@ xfs_trans_read_buf( ASSERT(!XFS_BUF_ISASYNC(bp)); XFS_BUF_READ(bp); xfsbdstrat(tp->t_mountp, bp); - xfs_iowait(bp); - if (XFS_BUF_GETERROR(bp) != 0) { + error = xfs_iowait(bp); + if (error) { xfs_ioerror_alert("xfs_trans_read_buf", mp, bp, blkno); - error = XFS_BUF_GETERROR(bp); xfs_buf_relse(bp); /* - * We can gracefully recover from most - * read errors. Ones we can't are those - * that happen after the transaction's - * already dirty. + * We can gracefully recover from most read + * errors. Ones we can't are those that happen + * after the transaction's already dirty. */ if (tp->t_flags & XFS_TRANS_DIRTY) xfs_force_shutdown(tp->t_mountp, diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index ca38fb9a993..dd4621e0ab3 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -3825,7 +3825,8 @@ xfs_zero_remaining_bytes( XFS_BUF_READ(bp); XFS_BUF_SET_ADDR(bp, XFS_FSB_TO_DB(ip, imap.br_startblock)); xfsbdstrat(mp, bp); - if ((error = xfs_iowait(bp))) { + error = xfs_iowait(bp); + if (error) { xfs_ioerror_alert("xfs_zero_remaining_bytes(read)", mp, bp, XFS_BUF_ADDR(bp)); break; @@ -3837,7 +3838,8 @@ xfs_zero_remaining_bytes( XFS_BUF_UNREAD(bp); XFS_BUF_WRITE(bp); xfsbdstrat(mp, bp); - if ((error = xfs_iowait(bp))) { + error = xfs_iowait(bp); + if (error) { xfs_ioerror_alert("xfs_zero_remaining_bytes(write)", mp, bp, XFS_BUF_ADDR(bp)); break; -- cgit v1.2.3 From db7a19f2c89d99b66874a7e0c0dc681ff1f37b4e Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:22:24 +1000 Subject: [XFS] Ensure xfs_bawrite() errors are checked. xfs_bawrite() can return immediate error status on async writes. Unlike xfsbdstrat() we don't ever check the error on the buffer after the call, so we currently do not catch errors at all here. Ensure we catch and propagate or warn to the syslog about up-front async write errors. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30824a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/quota/xfs_dquot.c | 10 ++++++++-- fs/xfs/quota/xfs_dquot_item.c | 7 ++++++- fs/xfs/xfs_buf_item.c | 7 ++++++- fs/xfs/xfs_inode.c | 2 +- fs/xfs/xfs_inode_item.c | 8 +++++++- 5 files changed, 28 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/xfs/quota/xfs_dquot.c b/fs/xfs/quota/xfs_dquot.c index 15214fbb9aa..631ebb31b29 100644 --- a/fs/xfs/quota/xfs_dquot.c +++ b/fs/xfs/quota/xfs_dquot.c @@ -1291,7 +1291,7 @@ xfs_qm_dqflush( if (flags & XFS_QMOPT_DELWRI) { xfs_bdwrite(mp, bp); } else if (flags & XFS_QMOPT_ASYNC) { - xfs_bawrite(mp, bp); + error = xfs_bawrite(mp, bp); } else { error = xfs_bwrite(mp, bp); } @@ -1582,12 +1582,18 @@ xfs_qm_dqflock_pushbuf_wait( XFS_INCORE_TRYLOCK); if (bp != NULL) { if (XFS_BUF_ISDELAYWRITE(bp)) { + int error; if (XFS_BUF_ISPINNED(bp)) { xfs_log_force(dqp->q_mount, (xfs_lsn_t)0, XFS_LOG_FORCE); } - xfs_bawrite(dqp->q_mount, bp); + error = xfs_bawrite(dqp->q_mount, bp); + if (error) + xfs_fs_cmn_err(CE_WARN, dqp->q_mount, + "xfs_qm_dqflock_pushbuf_wait: " + "pushbuf error %d on dqp %p, bp %p", + error, dqp, bp); } else { xfs_buf_relse(bp); } diff --git a/fs/xfs/quota/xfs_dquot_item.c b/fs/xfs/quota/xfs_dquot_item.c index 3dedce1d9cd..36e05ca7841 100644 --- a/fs/xfs/quota/xfs_dquot_item.c +++ b/fs/xfs/quota/xfs_dquot_item.c @@ -267,11 +267,16 @@ xfs_qm_dquot_logitem_pushbuf( XFS_LOG_FORCE); } if (dopush) { + int error; #ifdef XFSRACEDEBUG delay_for_intr(); delay(300); #endif - xfs_bawrite(mp, bp); + error = xfs_bawrite(mp, bp); + if (error) + xfs_fs_cmn_err(CE_WARN, mp, + "xfs_qm_dquot_logitem_pushbuf: pushbuf error %d on qip %p, bp %p", + error, qip, bp); } else { xfs_buf_relse(bp); } diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c index 63debd147eb..53a71c62025 100644 --- a/fs/xfs/xfs_buf_item.c +++ b/fs/xfs/xfs_buf_item.c @@ -645,7 +645,12 @@ xfs_buf_item_push( bp = bip->bli_buf; if (XFS_BUF_ISDELAYWRITE(bp)) { - xfs_bawrite(bip->bli_item.li_mountp, bp); + int error; + error = xfs_bawrite(bip->bli_item.li_mountp, bp); + if (error) + xfs_fs_cmn_err(CE_WARN, bip->bli_item.li_mountp, + "xfs_buf_item_push: pushbuf error %d on bip %p, bp %p", + error, bip, bp); } else { xfs_buf_relse(bp); } diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 63e66890f06..ca074ee01d0 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -3276,7 +3276,7 @@ xfs_iflush( if (flags & INT_DELWRI) { xfs_bdwrite(mp, bp); } else if (flags & INT_ASYNC) { - xfs_bawrite(mp, bp); + error = xfs_bawrite(mp, bp); } else { error = xfs_bwrite(mp, bp); } diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c index 2c775b4ae9e..93b5db453ea 100644 --- a/fs/xfs/xfs_inode_item.c +++ b/fs/xfs/xfs_inode_item.c @@ -40,6 +40,7 @@ #include "xfs_btree.h" #include "xfs_ialloc.h" #include "xfs_rw.h" +#include "xfs_error.h" kmem_zone_t *xfs_ili_zone; /* inode log item zone */ @@ -813,7 +814,12 @@ xfs_inode_item_pushbuf( XFS_LOG_FORCE); } if (dopush) { - xfs_bawrite(mp, bp); + int error; + error = xfs_bawrite(mp, bp); + if (error) + xfs_fs_cmn_err(CE_WARN, mp, + "xfs_inode_item_pushbuf: pushbuf error %d on iip %p, bp %p", + error, iip, bp); } else { xfs_buf_relse(bp); } -- cgit v1.2.3 From 958d4ec606d4af590f86a601a238613f21e878ee Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:23:46 +1000 Subject: [XFS] xfs_bdwrite() does not return errors. xfs_bdwrite() cannot return an error; it only queues buffers to the delayed write list and as such never encounters anything that can fail. Mark it void. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30825a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_buf.c | 2 +- fs/xfs/linux-2.6/xfs_buf.h | 8 ++++++-- 2 files changed, 7 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_buf.c b/fs/xfs/linux-2.6/xfs_buf.c index 142ddbece37..52f6846101d 100644 --- a/fs/xfs/linux-2.6/xfs_buf.c +++ b/fs/xfs/linux-2.6/xfs_buf.c @@ -1060,7 +1060,7 @@ xfs_buf_iostart( bp->b_flags &= ~(XBF_READ | XBF_WRITE | XBF_ASYNC); bp->b_flags |= flags & (XBF_DELWRI | XBF_ASYNC); xfs_buf_delwri_queue(bp, 1); - return status; + return 0; } bp->b_flags &= ~(XBF_READ | XBF_WRITE | XBF_ASYNC | XBF_DELWRI | \ diff --git a/fs/xfs/linux-2.6/xfs_buf.h b/fs/xfs/linux-2.6/xfs_buf.h index a3d207de48b..841d7883528 100644 --- a/fs/xfs/linux-2.6/xfs_buf.h +++ b/fs/xfs/linux-2.6/xfs_buf.h @@ -387,11 +387,15 @@ static inline int XFS_bwrite(xfs_buf_t *bp) return error; } -static inline int xfs_bdwrite(void *mp, xfs_buf_t *bp) +/* + * No error can be returned from xfs_buf_iostart for delwri + * buffers as they are queued and no I/O is issued. + */ +static inline void xfs_bdwrite(void *mp, xfs_buf_t *bp) { bp->b_strat = xfs_bdstrat_cb; bp->b_fspriv3 = mp; - return xfs_buf_iostart(bp, XBF_DELWRI | XBF_ASYNC); + (void)xfs_buf_iostart(bp, XBF_DELWRI | XBF_ASYNC); } #define XFS_bdstrat(bp) xfs_buf_iorequest(bp) -- cgit v1.2.3 From cc88466f3f67bb16fc91b0b974e51c2a43a9e597 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:23:52 +1000 Subject: [XFS] Catch unwritten extent conversion errors. On unwritten I/O completion, we fail to propagate an error when converting the extent to a written extent. This means that the I/O silently fails. propagate the error onto the ioend so that the inode is marked with an error appropriately. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30826a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_aops.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_aops.c b/fs/xfs/linux-2.6/xfs_aops.c index 169e6c06279..a55c3b26d84 100644 --- a/fs/xfs/linux-2.6/xfs_aops.c +++ b/fs/xfs/linux-2.6/xfs_aops.c @@ -243,8 +243,12 @@ xfs_end_bio_unwritten( size_t size = ioend->io_size; if (likely(!ioend->io_error)) { - if (!XFS_FORCED_SHUTDOWN(ip->i_mount)) - xfs_iomap_write_unwritten(ip, offset, size); + if (!XFS_FORCED_SHUTDOWN(ip->i_mount)) { + int error; + error = xfs_iomap_write_unwritten(ip, offset, size); + if (error) + ioend->io_error = error; + } xfs_setfilesize(ioend); } xfs_destroy_ioend(ioend); -- cgit v1.2.3 From e4ac967b117c5780760abbd9ae996210c31cb398 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:23:58 +1000 Subject: [XFS] xfs_iflush_fork() never returns an error. xfs_iflush_fork() never returns an error. Mark it void and clean up the code calling it that checks for errors. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30827a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_inode.c | 25 ++++++++----------------- 1 file changed, 8 insertions(+), 17 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index ca074ee01d0..2bc22790d65 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -2883,7 +2883,7 @@ xfs_iextents_copy( * format indicates the current state of the fork. */ /*ARGSUSED*/ -STATIC int +STATIC void xfs_iflush_fork( xfs_inode_t *ip, xfs_dinode_t *dip, @@ -2904,16 +2904,16 @@ xfs_iflush_fork( static const short extflag[2] = { XFS_ILOG_DEXT, XFS_ILOG_AEXT }; - if (iip == NULL) - return 0; + if (!iip) + return; ifp = XFS_IFORK_PTR(ip, whichfork); /* * This can happen if we gave up in iformat in an error path, * for the attribute fork. */ - if (ifp == NULL) { + if (!ifp) { ASSERT(whichfork == XFS_ATTR_FORK); - return 0; + return; } cp = XFS_DFORK_PTR(dip, whichfork); mp = ip->i_mount; @@ -2974,8 +2974,6 @@ xfs_iflush_fork( ASSERT(0); break; } - - return 0; } STATIC int @@ -3452,16 +3450,9 @@ xfs_iflush_int( } } - if (xfs_iflush_fork(ip, dip, iip, XFS_DATA_FORK, bp) == EFSCORRUPTED) { - goto corrupt_out; - } - - if (XFS_IFORK_Q(ip)) { - /* - * The only error from xfs_iflush_fork is on the data fork. - */ - (void) xfs_iflush_fork(ip, dip, iip, XFS_ATTR_FORK, bp); - } + xfs_iflush_fork(ip, dip, iip, XFS_DATA_FORK, bp); + if (XFS_IFORK_Q(ip)) + xfs_iflush_fork(ip, dip, iip, XFS_ATTR_FORK, bp); xfs_inobp_check(mp, bp); /* -- cgit v1.2.3 From 7b07339048f7b020575706b492c004b5664b67ab Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:24:04 +1000 Subject: [XFS] xfs_bulkstat_one_dinode() never returns an error. Mark it void. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30828a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_itable.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c index 45d8776408e..eb85bdedad0 100644 --- a/fs/xfs/xfs_itable.c +++ b/fs/xfs/xfs_itable.c @@ -129,7 +129,7 @@ xfs_bulkstat_one_iget( return error; } -STATIC int +STATIC void xfs_bulkstat_one_dinode( xfs_mount_t *mp, /* mount point for filesystem */ xfs_ino_t ino, /* inode number to get data for */ @@ -198,8 +198,6 @@ xfs_bulkstat_one_dinode( buf->bs_blocks = be64_to_cpu(dic->di_nblocks); break; } - - return 0; } STATIC int -- cgit v1.2.3 From 64bfe1bfae833e89ed77f72c61ded19f4b1976f8 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:24:10 +1000 Subject: [XFS] Catch errors from xfs_imap(). Catch errors from xfs_imap() in log recovery when we might be trying to map an invalid inode number due to a corrupted log. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30829a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_log_recover.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 1f83298f90a..a8039431b86 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -2282,7 +2282,9 @@ xlog_recover_do_inode_trans( * invalidate the buffer when we write it out below. */ imap.im_blkno = 0; - xfs_imap(log->l_mp, NULL, ino, &imap, 0); + error = xfs_imap(log->l_mp, NULL, ino, &imap, 0); + if (error) + goto error; } /* -- cgit v1.2.3 From 78e9da77f1bf265fe750b9223ec15707473fb6e8 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:24:17 +1000 Subject: [XFS] Don't allow silent errors in xfs_inactive(). xfs_inactive() fails to report errors when committing the inactive transaction. Hence we can get silent failures either finishing off the truncation or committing the transaction. Even if we get errors, we need to continue, so simply warn loudly to the system if we get errors here. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30830a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_vnodeops.c | 15 +++++++++++---- 1 file changed, 11 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_vnodeops.c b/fs/xfs/xfs_vnodeops.c index dd4621e0ab3..6650601c64f 100644 --- a/fs/xfs/xfs_vnodeops.c +++ b/fs/xfs/xfs_vnodeops.c @@ -1743,11 +1743,18 @@ xfs_inactive( XFS_TRANS_MOD_DQUOT_BYINO(mp, tp, ip, XFS_TRANS_DQ_ICOUNT, -1); /* - * Just ignore errors at this point. There is - * nothing we can do except to try to keep going. + * Just ignore errors at this point. There is nothing we can + * do except to try to keep going. Make sure it's not a silent + * error. */ - (void) xfs_bmap_finish(&tp, &free_list, &committed); - (void) xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES); + error = xfs_bmap_finish(&tp, &free_list, &committed); + if (error) + xfs_fs_cmn_err(CE_NOTE, mp, "xfs_inactive: " + "xfs_bmap_finish() returned error %d", error); + error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES); + if (error) + xfs_fs_cmn_err(CE_NOTE, mp, "xfs_inactive: " + "xfs_trans_commit() returned error %d", error); } /* * Release the dquots held by inode, if any. -- cgit v1.2.3 From 234f56aca20a4f66b6ba3d3bf2787634dd9e0999 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:24:24 +1000 Subject: [XFS] Check for errors when changing buffer pointers. xfs_buf_associate_memory() can fail, but the return is never checked. Propagate the error through XFS_BUF_SET_PTR() so that failures are detected. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30831a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_log_recover.c | 35 ++++++++++++++++++++++++----------- 1 file changed, 24 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index a8039431b86..e65ab4af095 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -1162,10 +1162,14 @@ xlog_write_log_records( if (j == 0 && (start_block + endcount > ealign)) { offset = XFS_BUF_PTR(bp); balign = BBTOB(ealign - start_block); - XFS_BUF_SET_PTR(bp, offset + balign, BBTOB(sectbb)); - if ((error = xlog_bread(log, ealign, sectbb, bp))) + error = XFS_BUF_SET_PTR(bp, offset + balign, + BBTOB(sectbb)); + if (!error) + error = xlog_bread(log, ealign, sectbb, bp); + if (!error) + error = XFS_BUF_SET_PTR(bp, offset, bufblks); + if (error) break; - XFS_BUF_SET_PTR(bp, offset, bufblks); } offset = xlog_align(log, start_block, endcount, bp); @@ -3630,15 +3634,19 @@ xlog_do_recovery_pass( * _first_, then the log start (LR header end) * - order is important. */ + wrapped_hblks = hblks - split_hblks; bufaddr = XFS_BUF_PTR(hbp); - XFS_BUF_SET_PTR(hbp, + error = XFS_BUF_SET_PTR(hbp, bufaddr + BBTOB(split_hblks), BBTOB(hblks - split_hblks)); - wrapped_hblks = hblks - split_hblks; - error = xlog_bread(log, 0, wrapped_hblks, hbp); + if (!error) + error = xlog_bread(log, 0, + wrapped_hblks, hbp); + if (!error) + error = XFS_BUF_SET_PTR(hbp, bufaddr, + BBTOB(hblks)); if (error) goto bread_err2; - XFS_BUF_SET_PTR(hbp, bufaddr, BBTOB(hblks)); if (!offset) offset = xlog_align(log, 0, wrapped_hblks, hbp); @@ -3690,13 +3698,18 @@ xlog_do_recovery_pass( * - order is important. */ bufaddr = XFS_BUF_PTR(dbp); - XFS_BUF_SET_PTR(dbp, + error = XFS_BUF_SET_PTR(dbp, bufaddr + BBTOB(split_bblks), BBTOB(bblks - split_bblks)); - if ((error = xlog_bread(log, wrapped_hblks, - bblks - split_bblks, dbp))) + if (!error) + error = xlog_bread(log, wrapped_hblks, + bblks - split_bblks, + dbp); + if (!error) + error = XFS_BUF_SET_PTR(dbp, bufaddr, + h_size); + if (error) goto bread_err2; - XFS_BUF_SET_PTR(dbp, bufaddr, h_size); if (!offset) offset = xlog_align(log, wrapped_hblks, bblks - split_bblks, dbp); -- cgit v1.2.3 From b911ca0472c3762d2bafc4d21e432a9056844064 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:24:30 +1000 Subject: [XFS] Sanitise xfs_log_force error checking. xfs_log_force() is declared to return an error, but we almost never check it. We don't need to check it in most cases; if there's a log I/O error then we'll be shutting down the filesystem anyway and that means we'll catch the error somewhere else. However, on certain calls we should be returning an error - sync transactions, fsync, sync writes, etc. so this isn't a pure black and white distinction. Hence make xfs_log_force() a void function that issues a warning to the syslog on error, and call _xfs_log_force() in all the places where we actually care about the error status returned. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30832a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_log.c | 25 +++++++++++++++++++++++-- fs/xfs/xfs_log.h | 5 +++-- fs/xfs/xfs_rw.c | 8 ++++---- 3 files changed, 30 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index bece882f99e..e29ea0a6d76 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -382,7 +382,27 @@ _xfs_log_force( return xlog_state_sync_all(log, flags, log_flushed); else return xlog_state_sync(log, lsn, flags, log_flushed); -} /* xfs_log_force */ +} /* _xfs_log_force */ + +/* + * Wrapper for _xfs_log_force(), to be used when caller doesn't care + * about errors or whether the log was flushed or not. This is the normal + * interface to use when trying to unpin items or move the log forward. + */ +void +xfs_log_force( + xfs_mount_t *mp, + xfs_lsn_t lsn, + uint flags) +{ + int error; + error = _xfs_log_force(mp, lsn, flags, NULL); + if (error) { + xfs_fs_cmn_err(CE_WARN, mp, "xfs_log_force: " + "error %d returned.", error); + } +} + /* * Attaches a new iclog I/O completion callback routine during @@ -634,7 +654,8 @@ xfs_log_unmount_write(xfs_mount_t *mp) if (mp->m_flags & XFS_MOUNT_RDONLY) return 0; - xfs_log_force(mp, 0, XFS_LOG_FORCE|XFS_LOG_SYNC); + error = _xfs_log_force(mp, 0, XFS_LOG_FORCE|XFS_LOG_SYNC, NULL); + ASSERT(error || !(XLOG_FORCED_SHUTDOWN(log))); #ifdef DEBUG first_iclog = iclog = log->l_iclog; diff --git a/fs/xfs/xfs_log.h b/fs/xfs/xfs_log.h index 4cdac048df5..d1d678ecb63 100644 --- a/fs/xfs/xfs_log.h +++ b/fs/xfs/xfs_log.h @@ -142,8 +142,9 @@ int _xfs_log_force(struct xfs_mount *mp, xfs_lsn_t lsn, uint flags, int *log_forced); -#define xfs_log_force(mp, lsn, flags) \ - _xfs_log_force(mp, lsn, flags, NULL); +void xfs_log_force(struct xfs_mount *mp, + xfs_lsn_t lsn, + uint flags); int xfs_log_mount(struct xfs_mount *mp, struct xfs_buftarg *log_target, xfs_daddr_t start_block, diff --git a/fs/xfs/xfs_rw.c b/fs/xfs/xfs_rw.c index cd3ece6cc91..b0f31c09a76 100644 --- a/fs/xfs/xfs_rw.c +++ b/fs/xfs/xfs_rw.c @@ -126,11 +126,11 @@ xfs_write_sync_logforce( * when we return. */ if (iip && iip->ili_last_lsn) { - xfs_log_force(mp, iip->ili_last_lsn, - XFS_LOG_FORCE | XFS_LOG_SYNC); + error = _xfs_log_force(mp, iip->ili_last_lsn, + XFS_LOG_FORCE | XFS_LOG_SYNC, NULL); } else if (xfs_ipincount(ip) > 0) { - xfs_log_force(mp, (xfs_lsn_t)0, - XFS_LOG_FORCE | XFS_LOG_SYNC); + error = _xfs_log_force(mp, (xfs_lsn_t)0, + XFS_LOG_FORCE | XFS_LOG_SYNC, NULL); } } else { -- cgit v1.2.3 From 1bb7d6b5a82f1d9487fd44415484a368f7c87bed Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 10 Apr 2008 12:24:38 +1000 Subject: [XFS] Catch log unmount failures. Unmounting the log can fail. unlikely, but it can. Catch all the error conditions an make sure it's propagated upwards. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30833a Signed-off-by: David Chinner Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_log.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index e29ea0a6d76..afaee301b0e 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -697,7 +697,7 @@ xfs_log_unmount_write(xfs_mount_t *mp) atomic_inc(&iclog->ic_refcnt); spin_unlock(&log->l_icloglock); xlog_state_want_sync(log, iclog); - (void) xlog_state_release_iclog(log, iclog); + error = xlog_state_release_iclog(log, iclog); spin_lock(&log->l_icloglock); if (!(iclog->ic_state == XLOG_STATE_ACTIVE || @@ -736,7 +736,7 @@ xfs_log_unmount_write(xfs_mount_t *mp) spin_unlock(&log->l_icloglock); xlog_state_want_sync(log, iclog); - (void) xlog_state_release_iclog(log, iclog); + error = xlog_state_release_iclog(log, iclog); spin_lock(&log->l_icloglock); @@ -751,7 +751,7 @@ xfs_log_unmount_write(xfs_mount_t *mp) } } - return 0; + return error; } /* xfs_log_unmount_write */ /* -- cgit v1.2.3 From d4055947bd0913864f4d8ac96bf1197338071622 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 17 Apr 2008 16:49:35 +1000 Subject: [XFS] Don't error out on good I/Os. xfsbdstrat() made all I/Os error out, good or bad. Fix it. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30836a Signed-off-by: David Chinner Signed-off-by: Donald Douwsma Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_lrw.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_lrw.c b/fs/xfs/linux-2.6/xfs_lrw.c index f6dab5d8944..21c0dbc7409 100644 --- a/fs/xfs/linux-2.6/xfs_lrw.c +++ b/fs/xfs/linux-2.6/xfs_lrw.c @@ -885,8 +885,10 @@ xfsbdstrat( struct xfs_buf *bp) { ASSERT(mp); - if (!XFS_FORCED_SHUTDOWN(mp)) + if (!XFS_FORCED_SHUTDOWN(mp)) { xfs_buf_iorequest(bp); + return; + } xfs_buftrace("XFSBDSTRAT IOERROR", bp); xfs_bioerror_relse(bp); -- cgit v1.2.3 From e6430037e9fd0b3d02ceaf5ab99bfe3ccb763be7 Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 17 Apr 2008 16:49:49 +1000 Subject: [XFS] fix logic error in xfs_alloc_ag_vextent_near() Fix a logic error in xfs_alloc_ag_vextent_near(). This is a regression introduced by the error handling changes. SGI-PV: 890084 SGI-Modid: xfs-linux-melb:xfs-kern:30838a Signed-off-by: David Chinner Signed-off-by: Barry Naujok Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_alloc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/xfs_alloc.c b/fs/xfs/xfs_alloc.c index facdae14edd..1956f83489f 100644 --- a/fs/xfs/xfs_alloc.c +++ b/fs/xfs/xfs_alloc.c @@ -838,7 +838,7 @@ xfs_alloc_ag_vextent_near( XFS_WANT_CORRUPTED_GOTO(i == 1, error0); xfs_alloc_compute_aligned(ltbno, ltlen, args->alignment, args->minlen, <bnoa, <lena); - if (ltlena >= args->minlen) + if (ltlena < args->minlen) continue; args->len = XFS_EXTLEN_MIN(ltlena, args->maxlen); xfs_alloc_fix_len(args); -- cgit v1.2.3 From 7e20694d91f817f8e9f62404aca793ae0df4d98a Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 17 Apr 2008 16:49:55 +1000 Subject: [XFS] Remove periodic logging of in-core superblock counters. xfssyncd triggers the logging of superblock counters every 30s if the filesystem is made with lazy-count=1. This will prevent disks from idling and spinning down as there will be a log write every 30s. With the way counter recovery works for lazy-count=1, this code is unnecessary and provides no real benefit, so just remove it. SGI-PV: 980145 SGI-Modid: xfs-linux-melb:xfs-kern:30840a Signed-off-by: David Chinner Signed-off-by: Barry Naujok Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_super.c | 3 +-- fs/xfs/linux-2.6/xfs_vfs.h | 1 - fs/xfs/xfs_vfsops.c | 13 ------------- 3 files changed, 1 insertion(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c index fb561beea37..865eb708aa9 100644 --- a/fs/xfs/linux-2.6/xfs_super.c +++ b/fs/xfs/linux-2.6/xfs_super.c @@ -1028,8 +1028,7 @@ xfs_sync_worker( int error; if (!(mp->m_flags & XFS_MOUNT_RDONLY)) - error = xfs_sync(mp, SYNC_FSDATA | SYNC_BDFLUSH | SYNC_ATTR | - SYNC_REFCACHE | SYNC_SUPER); + error = xfs_sync(mp, SYNC_FSDATA | SYNC_BDFLUSH | SYNC_ATTR); mp->m_sync_seq++; wake_up(&mp->m_wait_single_sync_task); } diff --git a/fs/xfs/linux-2.6/xfs_vfs.h b/fs/xfs/linux-2.6/xfs_vfs.h index 4da03a4e352..7e60c7776b1 100644 --- a/fs/xfs/linux-2.6/xfs_vfs.h +++ b/fs/xfs/linux-2.6/xfs_vfs.h @@ -49,7 +49,6 @@ typedef struct bhv_vfs_sync_work { #define SYNC_REFCACHE 0x0040 /* prune some of the nfs ref cache */ #define SYNC_REMOUNT 0x0080 /* remount readonly, no dummy LRs */ #define SYNC_IOWAIT 0x0100 /* wait for all I/O to complete */ -#define SYNC_SUPER 0x0200 /* flush superblock to disk */ /* * When remounting a filesystem read-only or freezing the filesystem, diff --git a/fs/xfs/xfs_vfsops.c b/fs/xfs/xfs_vfsops.c index 09e186d02c1..fc48158fe47 100644 --- a/fs/xfs/xfs_vfsops.c +++ b/fs/xfs/xfs_vfsops.c @@ -1316,22 +1316,9 @@ xfs_syncsub( } } - /* - * If asked, update the disk superblock with incore counter values if we - * are using non-persistent counters so that they don't get too far out - * of sync if we crash or get a forced shutdown. We don't want to force - * this to disk, just get a transaction into the iclogs.... - */ - if (flags & SYNC_SUPER) { - error = xfs_log_sbcount(mp, 0); - if (error) - last_error = error; - } - /* * Now check to see if the log needs a "dummy" transaction. */ - if (!(flags & SYNC_REMOUNT) && xfs_log_need_covered(mp)) { xfs_trans_t *tp; xfs_inode_t *ip; -- cgit v1.2.3 From f6485057c5cfbc84e5eff639ddea1ce0d668607b Mon Sep 17 00:00:00 2001 From: David Chinner Date: Thu, 17 Apr 2008 16:50:04 +1000 Subject: [XFS] Ensure the inode is joined in xfs_itruncate_finish On success, we still need to join the inode to the current transaction in xfs_itruncate_finish(). Fixes regression from error handling changes. SGI-PV: 980084 SGI-Modid: xfs-linux-melb:xfs-kern:30845a Signed-off-by: David Chinner Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_inode.c | 137 +++++++++++++++++++++++++---------------------------- 1 file changed, 65 insertions(+), 72 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 2bc22790d65..ca12acb9039 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -1464,51 +1464,50 @@ xfs_itruncate_start( } /* - * Shrink the file to the given new_size. The new - * size must be smaller than the current size. - * This will free up the underlying blocks - * in the removed range after a call to xfs_itruncate_start() - * or xfs_atruncate_start(). + * Shrink the file to the given new_size. The new size must be smaller than + * the current size. This will free up the underlying blocks in the removed + * range after a call to xfs_itruncate_start() or xfs_atruncate_start(). * - * The transaction passed to this routine must have made - * a permanent log reservation of at least XFS_ITRUNCATE_LOG_RES. - * This routine may commit the given transaction and - * start new ones, so make sure everything involved in - * the transaction is tidy before calling here. - * Some transaction will be returned to the caller to be - * committed. The incoming transaction must already include - * the inode, and both inode locks must be held exclusively. - * The inode must also be "held" within the transaction. On - * return the inode will be "held" within the returned transaction. - * This routine does NOT require any disk space to be reserved - * for it within the transaction. + * The transaction passed to this routine must have made a permanent log + * reservation of at least XFS_ITRUNCATE_LOG_RES. This routine may commit the + * given transaction and start new ones, so make sure everything involved in + * the transaction is tidy before calling here. Some transaction will be + * returned to the caller to be committed. The incoming transaction must + * already include the inode, and both inode locks must be held exclusively. + * The inode must also be "held" within the transaction. On return the inode + * will be "held" within the returned transaction. This routine does NOT + * require any disk space to be reserved for it within the transaction. * - * The fork parameter must be either xfs_attr_fork or xfs_data_fork, - * and it indicates the fork which is to be truncated. For the - * attribute fork we only support truncation to size 0. + * The fork parameter must be either xfs_attr_fork or xfs_data_fork, and it + * indicates the fork which is to be truncated. For the attribute fork we only + * support truncation to size 0. * - * We use the sync parameter to indicate whether or not the first - * transaction we perform might have to be synchronous. For the attr fork, - * it needs to be so if the unlink of the inode is not yet known to be - * permanent in the log. This keeps us from freeing and reusing the - * blocks of the attribute fork before the unlink of the inode becomes - * permanent. + * We use the sync parameter to indicate whether or not the first transaction + * we perform might have to be synchronous. For the attr fork, it needs to be + * so if the unlink of the inode is not yet known to be permanent in the log. + * This keeps us from freeing and reusing the blocks of the attribute fork + * before the unlink of the inode becomes permanent. * - * For the data fork, we normally have to run synchronously if we're - * being called out of the inactive path or we're being called - * out of the create path where we're truncating an existing file. - * Either way, the truncate needs to be sync so blocks don't reappear - * in the file with altered data in case of a crash. wsync filesystems - * can run the first case async because anything that shrinks the inode - * has to run sync so by the time we're called here from inactive, the - * inode size is permanently set to 0. + * For the data fork, we normally have to run synchronously if we're being + * called out of the inactive path or we're being called out of the create path + * where we're truncating an existing file. Either way, the truncate needs to + * be sync so blocks don't reappear in the file with altered data in case of a + * crash. wsync filesystems can run the first case async because anything that + * shrinks the inode has to run sync so by the time we're called here from + * inactive, the inode size is permanently set to 0. * - * Calls from the truncate path always need to be sync unless we're - * in a wsync filesystem and the file has already been unlinked. + * Calls from the truncate path always need to be sync unless we're in a wsync + * filesystem and the file has already been unlinked. * - * The caller is responsible for correctly setting the sync parameter. - * It gets too hard for us to guess here which path we're being called - * out of just based on inode state. + * The caller is responsible for correctly setting the sync parameter. It gets + * too hard for us to guess here which path we're being called out of just + * based on inode state. + * + * If we get an error, we must return with the inode locked and linked into the + * current transaction. This keeps things simple for the higher level code, + * because it always knows that the inode is locked and held in the transaction + * that returns to it whether errors occur or not. We don't mark the inode + * dirty on error so that transactions can be easily aborted if possible. */ int xfs_itruncate_finish( @@ -1687,45 +1686,51 @@ xfs_itruncate_finish( */ error = xfs_bmap_finish(tp, &free_list, &committed); ntp = *tp; + if (committed) { + /* link the inode into the next xact in the chain */ + xfs_trans_ijoin(ntp, ip, + XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL); + xfs_trans_ihold(ntp, ip); + } + if (error) { /* - * If the bmap finish call encounters an error, - * return to the caller where the transaction - * can be properly aborted. We just need to - * make sure we're not holding any resources - * that we were not when we came in. + * If the bmap finish call encounters an error, return + * to the caller where the transaction can be properly + * aborted. We just need to make sure we're not + * holding any resources that we were not when we came + * in. * - * Aborting from this point might lose some - * blocks in the file system, but oh well. + * Aborting from this point might lose some blocks in + * the file system, but oh well. */ xfs_bmap_cancel(&free_list); - if (committed) - goto error_join; return error; } if (committed) { /* - * The first xact was committed, so add the inode to - * the new one. Mark it dirty so it will be logged and + * Mark the inode dirty so it will be logged and * moved forward in the log as part of every commit. */ - xfs_trans_ijoin(ntp, ip, - XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL); - xfs_trans_ihold(ntp, ip); xfs_trans_log_inode(ntp, ip, XFS_ILOG_CORE); } + ntp = xfs_trans_dup(ntp); error = xfs_trans_commit(*tp, 0); *tp = ntp; - if (error) - goto error_join; - error = xfs_trans_reserve(ntp, 0, XFS_ITRUNCATE_LOG_RES(mp), 0, - XFS_TRANS_PERM_LOG_RES, - XFS_ITRUNCATE_LOG_COUNT); - if (error) - goto error_join; + /* link the inode into the next transaction in the chain */ + xfs_trans_ijoin(ntp, ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL); + xfs_trans_ihold(ntp, ip); + + if (!error) + error = xfs_trans_reserve(ntp, 0, + XFS_ITRUNCATE_LOG_RES(mp), 0, + XFS_TRANS_PERM_LOG_RES, + XFS_ITRUNCATE_LOG_COUNT); + if (error) + return error; } /* * Only update the size in the case of the data fork, but @@ -1757,18 +1762,6 @@ xfs_itruncate_finish( (ip->i_d.di_nextents == 0)); xfs_itrunc_trace(XFS_ITRUNC_FINISH2, ip, 0, new_size, 0, 0); return 0; - -error_join: - /* - * Add the inode being truncated to the next chained transaction. This - * keeps things simple for the higher level code, because it always - * knows that the inode is locked and held in the transaction that - * returns to it whether errors occur or not. We don't mark the inode - * dirty so that this transaction can be easily aborted if possible. - */ - xfs_trans_ijoin(ntp, ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL); - xfs_trans_ihold(ntp, ip); - return error; } -- cgit v1.2.3 From cb49dbb130e17a6f9af4cb4714cf6976cf09afdf Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Thu, 17 Apr 2008 16:50:09 +1000 Subject: [XFS] Always use di_forkoff when checking for attr space. In the case where we mount a filesystem which was previously using the attr2 format as attr1, returning the default mp->m_attroffset instead of the per-inode di_forkoff for inline attribute fit calculations, may result in corruption, if for example, the data fork is already taking more space than the default fork offset and we try to add an extended attribute. Fix tested by xfstests/186. SGI-PV: 979606 SGI-Modid: xfs-linux-melb:xfs-kern:30861a Signed-off-by: Eric Sandeen Signed-off-by: Tim Shimmin Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_attr_leaf.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/xfs_attr_leaf.c b/fs/xfs/xfs_attr_leaf.c index 96ba6aa4ed8..303d41e4217 100644 --- a/fs/xfs/xfs_attr_leaf.c +++ b/fs/xfs/xfs_attr_leaf.c @@ -166,7 +166,7 @@ xfs_attr_shortform_bytesfit(xfs_inode_t *dp, int bytes) if (!(mp->m_flags & XFS_MOUNT_ATTR2)) { if (bytes <= XFS_IFORK_ASIZE(dp)) - return mp->m_attroffset >> 3; + return dp->i_d.di_forkoff; return 0; } -- cgit v1.2.3 From 6d1337b29bf09a97682d39db36ac2d0dfc6659c0 Mon Sep 17 00:00:00 2001 From: Tim Shimmin Date: Thu, 17 Apr 2008 16:50:16 +1000 Subject: [XFS] xfs_bmap_compute_maxlevels should be based on di_forkoff Fix up xfs_bmap_compute_maxlevels() to account for the case when we go from using attr2 to using attr1. In that case attr1 will no longer necessarily be at m_attr_offset>>3, but could be at a different value for di_forkoff. Therefore, we return the worst case scenario using MINDBTPTRS and MINABTPTRS, as this function is used for determining the maximum log space. SGI-PV: 979606 SGI-Modid: xfs-linux-melb:xfs-kern:30862a Signed-off-by: Tim Shimmin Signed-off-by: Eric Sandeen Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs_bmap.c | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/xfs/xfs_bmap.c b/fs/xfs/xfs_bmap.c index 6d9b5448deb..eb198c01c35 100644 --- a/fs/xfs/xfs_bmap.c +++ b/fs/xfs/xfs_bmap.c @@ -4153,16 +4153,21 @@ xfs_bmap_compute_maxlevels( * number of leaf entries, is controlled by the type of di_nextents * (a signed 32-bit number, xfs_extnum_t), or by di_anextents * (a signed 16-bit number, xfs_aextnum_t). + * + * Note that we can no longer assume that if we are in ATTR1 that + * the fork offset of all the inodes will be (m_attroffset >> 3) + * because we could have mounted with ATTR2 and then mounted back + * with ATTR1, keeping the di_forkoff's fixed but probably at + * various positions. Therefore, for both ATTR1 and ATTR2 + * we have to assume the worst case scenario of a minimum size + * available. */ if (whichfork == XFS_DATA_FORK) { maxleafents = MAXEXTNUM; - sz = (mp->m_flags & XFS_MOUNT_ATTR2) ? - XFS_BMDR_SPACE_CALC(MINDBTPTRS) : mp->m_attroffset; + sz = XFS_BMDR_SPACE_CALC(MINDBTPTRS); } else { maxleafents = MAXAEXTNUM; - sz = (mp->m_flags & XFS_MOUNT_ATTR2) ? - XFS_BMDR_SPACE_CALC(MINABTPTRS) : - mp->m_sb.sb_inodesize - mp->m_attroffset; + sz = XFS_BMDR_SPACE_CALC(MINABTPTRS); } maxrootrecs = (int)XFS_BTREE_BLOCK_MAXRECS(sz, xfs_bmdr, 0); minleafrecs = mp->m_bmap_dmnr[0]; -- cgit v1.2.3 From f7d3c34788696f5ba9ac9fa414ad80e2a91d4b2e Mon Sep 17 00:00:00 2001 From: Eric Sandeen Date: Thu, 17 Apr 2008 16:50:22 +1000 Subject: [XFS] Remove CONFIG_XFS_SECURITY. There is no point to the CONFIG_XFS_SECURITY option; it disables the ability to set security attributes at runtime, but it does not actually slim down or remove any code for runtime. Just remove it and always allow security attributes to be set. SGI-PV: 980310 SGI-Modid: xfs-linux-melb:xfs-kern:30877a Signed-off-by: Eric Sandeen Signed-off-by: Tim Shimmin Signed-off-by: Lachlan McIlroy --- fs/xfs/Kconfig | 12 ------------ fs/xfs/linux-2.6/xfs_super.h | 8 +------- fs/xfs/xfs_attr.c | 10 +--------- 3 files changed, 2 insertions(+), 28 deletions(-) (limited to 'fs') diff --git a/fs/xfs/Kconfig b/fs/xfs/Kconfig index 35115bca036..524021ff543 100644 --- a/fs/xfs/Kconfig +++ b/fs/xfs/Kconfig @@ -35,18 +35,6 @@ config XFS_QUOTA with or without the generic quota support enabled (CONFIG_QUOTA) - they are completely independent subsystems. -config XFS_SECURITY - bool "XFS Security Label support" - depends on XFS_FS - help - Security labels support alternative access control models - implemented by security modules like SELinux. This option - enables an extended attribute namespace for inode security - labels in the XFS filesystem. - - If you are not using a security module that requires using - extended attributes for inode security labels, say N. - config XFS_POSIX_ACL bool "XFS POSIX ACL support" depends on XFS_FS diff --git a/fs/xfs/linux-2.6/xfs_super.h b/fs/xfs/linux-2.6/xfs_super.h index 3efcf45b14a..3efb7c6d330 100644 --- a/fs/xfs/linux-2.6/xfs_super.h +++ b/fs/xfs/linux-2.6/xfs_super.h @@ -50,13 +50,7 @@ extern void xfs_qm_exit(void); # define set_posix_acl_flag(sb) do { } while (0) #endif -#ifdef CONFIG_XFS_SECURITY -# define XFS_SECURITY_STRING "security attributes, " -# define ENOSECURITY 0 -#else -# define XFS_SECURITY_STRING -# define ENOSECURITY EOPNOTSUPP -#endif +#define XFS_SECURITY_STRING "security attributes, " #ifdef CONFIG_XFS_RT # define XFS_REALTIME_STRING "realtime, " diff --git a/fs/xfs/xfs_attr.c b/fs/xfs/xfs_attr.c index e58f321fdae..36d781ee5fc 100644 --- a/fs/xfs/xfs_attr.c +++ b/fs/xfs/xfs_attr.c @@ -2646,14 +2646,6 @@ attr_trusted_capable( return 0; } -STATIC int -attr_secure_capable( - bhv_vnode_t *vp, - cred_t *cred) -{ - return -ENOSECURITY; -} - STATIC int attr_system_set( bhv_vnode_t *vp, char *name, void *data, size_t size, int xflags) @@ -2724,7 +2716,7 @@ struct attrnames attr_secure = { .attr_get = attr_generic_get, .attr_set = attr_generic_set, .attr_remove = attr_generic_remove, - .attr_capable = attr_secure_capable, + .attr_capable = (attrcapable_t)fs_noerr, }; struct attrnames attr_user = { -- cgit v1.2.3 From e687330b5ed1ea899fdaf0dea50aba196b6e019a Mon Sep 17 00:00:00 2001 From: Donald Douwsma Date: Thu, 17 Apr 2008 16:50:28 +1000 Subject: [XFS] Remove unused HAVE_SPLICE macro. HAVE_SPLICE was part of the infrastructure for building 2.4 and 2.6 kernels out of the same tree. Now we don't build 2.4 kernels this SGI-PV: 971046 SGI-Modid: xfs-linux-melb:xfs-kern:30878a Signed-off-by: Donald Douwsma Signed-off-by: Christoph Hellwig Signed-off-by: Lachlan McIlroy --- fs/xfs/xfs.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/xfs/xfs.h b/fs/xfs/xfs.h index 540e4c98982..765aaf65e2d 100644 --- a/fs/xfs/xfs.h +++ b/fs/xfs/xfs.h @@ -22,7 +22,7 @@ #define STATIC #define DEBUG 1 #define XFS_BUF_LOCK_TRACKING 1 -/* #define QUOTADEBUG 1 */ +#define QUOTADEBUG 1 #endif #ifdef CONFIG_XFS_TRACE -- cgit v1.2.3 From 3b2816be271b8b364294a5b48721a3e68af46cfa Mon Sep 17 00:00:00 2001 From: Lachlan McIlroy Date: Fri, 18 Apr 2008 12:43:35 +1000 Subject: [XFS] The forward declarations for the xfs_ioctl() helpers and the associated comment about gcc behavior really aren't needed; all of these functions are marked STATIC which includes noinline, and the stack usage won't be a problem. This effectively just removes the forward declarations and moves xfs_ioctl() back to the end of the file. SGI-PV: 971186 SGI-Modid: xfs-linux-melb:xfs-kern:30534a Signed-off-by: Eric Sandeen Signed-off-by: Niv Sardi Signed-off-by: Lachlan McIlroy --- fs/xfs/linux-2.6/xfs_ioctl.c | 314 +------------------------------------------ 1 file changed, 2 insertions(+), 312 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_ioctl.c b/fs/xfs/linux-2.6/xfs_ioctl.c index c6399b2cf17..640c8b6b195 100644 --- a/fs/xfs/linux-2.6/xfs_ioctl.c +++ b/fs/xfs/linux-2.6/xfs_ioctl.c @@ -651,314 +651,6 @@ xfs_attrmulti_by_handle( return -error; } -/* prototypes for a few of the stack-hungry cases that have - * their own functions. Functions are defined after their use - * so gcc doesn't get fancy and inline them with -03 */ - -STATIC int -xfs_ioc_space( - struct xfs_inode *ip, - struct inode *inode, - struct file *filp, - int flags, - unsigned int cmd, - void __user *arg); - -STATIC int -xfs_ioc_bulkstat( - xfs_mount_t *mp, - unsigned int cmd, - void __user *arg); - -STATIC int -xfs_ioc_fsgeometry_v1( - xfs_mount_t *mp, - void __user *arg); - -STATIC int -xfs_ioc_fsgeometry( - xfs_mount_t *mp, - void __user *arg); - -STATIC int -xfs_ioc_xattr( - xfs_inode_t *ip, - struct file *filp, - unsigned int cmd, - void __user *arg); - -STATIC int -xfs_ioc_fsgetxattr( - xfs_inode_t *ip, - int attr, - void __user *arg); - -STATIC int -xfs_ioc_getbmap( - struct xfs_inode *ip, - int flags, - unsigned int cmd, - void __user *arg); - -STATIC int -xfs_ioc_getbmapx( - struct xfs_inode *ip, - void __user *arg); - -int -xfs_ioctl( - xfs_inode_t *ip, - struct file *filp, - int ioflags, - unsigned int cmd, - void __user *arg) -{ - struct inode *inode = filp->f_path.dentry->d_inode; - xfs_mount_t *mp = ip->i_mount; - int error; - - xfs_itrace_entry(XFS_I(inode)); - switch (cmd) { - - case XFS_IOC_ALLOCSP: - case XFS_IOC_FREESP: - case XFS_IOC_RESVSP: - case XFS_IOC_UNRESVSP: - case XFS_IOC_ALLOCSP64: - case XFS_IOC_FREESP64: - case XFS_IOC_RESVSP64: - case XFS_IOC_UNRESVSP64: - /* - * Only allow the sys admin to reserve space unless - * unwritten extents are enabled. - */ - if (!xfs_sb_version_hasextflgbit(&mp->m_sb) && - !capable(CAP_SYS_ADMIN)) - return -EPERM; - - return xfs_ioc_space(ip, inode, filp, ioflags, cmd, arg); - - case XFS_IOC_DIOINFO: { - struct dioattr da; - xfs_buftarg_t *target = - XFS_IS_REALTIME_INODE(ip) ? - mp->m_rtdev_targp : mp->m_ddev_targp; - - da.d_mem = da.d_miniosz = 1 << target->bt_sshift; - da.d_maxiosz = INT_MAX & ~(da.d_miniosz - 1); - - if (copy_to_user(arg, &da, sizeof(da))) - return -XFS_ERROR(EFAULT); - return 0; - } - - case XFS_IOC_FSBULKSTAT_SINGLE: - case XFS_IOC_FSBULKSTAT: - case XFS_IOC_FSINUMBERS: - return xfs_ioc_bulkstat(mp, cmd, arg); - - case XFS_IOC_FSGEOMETRY_V1: - return xfs_ioc_fsgeometry_v1(mp, arg); - - case XFS_IOC_FSGEOMETRY: - return xfs_ioc_fsgeometry(mp, arg); - - case XFS_IOC_GETVERSION: - return put_user(inode->i_generation, (int __user *)arg); - - case XFS_IOC_FSGETXATTR: - return xfs_ioc_fsgetxattr(ip, 0, arg); - case XFS_IOC_FSGETXATTRA: - return xfs_ioc_fsgetxattr(ip, 1, arg); - case XFS_IOC_GETXFLAGS: - case XFS_IOC_SETXFLAGS: - case XFS_IOC_FSSETXATTR: - return xfs_ioc_xattr(ip, filp, cmd, arg); - - case XFS_IOC_FSSETDM: { - struct fsdmidata dmi; - - if (copy_from_user(&dmi, arg, sizeof(dmi))) - return -XFS_ERROR(EFAULT); - - error = xfs_set_dmattrs(ip, dmi.fsd_dmevmask, - dmi.fsd_dmstate); - return -error; - } - - case XFS_IOC_GETBMAP: - case XFS_IOC_GETBMAPA: - return xfs_ioc_getbmap(ip, ioflags, cmd, arg); - - case XFS_IOC_GETBMAPX: - return xfs_ioc_getbmapx(ip, arg); - - case XFS_IOC_FD_TO_HANDLE: - case XFS_IOC_PATH_TO_HANDLE: - case XFS_IOC_PATH_TO_FSHANDLE: - return xfs_find_handle(cmd, arg); - - case XFS_IOC_OPEN_BY_HANDLE: - return xfs_open_by_handle(mp, arg, filp, inode); - - case XFS_IOC_FSSETDM_BY_HANDLE: - return xfs_fssetdm_by_handle(mp, arg, inode); - - case XFS_IOC_READLINK_BY_HANDLE: - return xfs_readlink_by_handle(mp, arg, inode); - - case XFS_IOC_ATTRLIST_BY_HANDLE: - return xfs_attrlist_by_handle(mp, arg, inode); - - case XFS_IOC_ATTRMULTI_BY_HANDLE: - return xfs_attrmulti_by_handle(mp, arg, inode); - - case XFS_IOC_SWAPEXT: { - error = xfs_swapext((struct xfs_swapext __user *)arg); - return -error; - } - - case XFS_IOC_FSCOUNTS: { - xfs_fsop_counts_t out; - - error = xfs_fs_counts(mp, &out); - if (error) - return -error; - - if (copy_to_user(arg, &out, sizeof(out))) - return -XFS_ERROR(EFAULT); - return 0; - } - - case XFS_IOC_SET_RESBLKS: { - xfs_fsop_resblks_t inout; - __uint64_t in; - - if (!capable(CAP_SYS_ADMIN)) - return -EPERM; - - if (copy_from_user(&inout, arg, sizeof(inout))) - return -XFS_ERROR(EFAULT); - - /* input parameter is passed in resblks field of structure */ - in = inout.resblks; - error = xfs_reserve_blocks(mp, &in, &inout); - if (error) - return -error; - - if (copy_to_user(arg, &inout, sizeof(inout))) - return -XFS_ERROR(EFAULT); - return 0; - } - - case XFS_IOC_GET_RESBLKS: { - xfs_fsop_resblks_t out; - - if (!capable(CAP_SYS_ADMIN)) - return -EPERM; - - error = xfs_reserve_blocks(mp, NULL, &out); - if (error) - return -error; - - if (copy_to_user(arg, &out, sizeof(out))) - return -XFS_ERROR(EFAULT); - - return 0; - } - - case XFS_IOC_FSGROWFSDATA: { - xfs_growfs_data_t in; - - if (!capable(CAP_SYS_ADMIN)) - return -EPERM; - - if (copy_from_user(&in, arg, sizeof(in))) - return -XFS_ERROR(EFAULT); - - error = xfs_growfs_data(mp, &in); - return -error; - } - - case XFS_IOC_FSGROWFSLOG: { - xfs_growfs_log_t in; - - if (!capable(CAP_SYS_ADMIN)) - return -EPERM; - - if (copy_from_user(&in, arg, sizeof(in))) - return -XFS_ERROR(EFAULT); - - error = xfs_growfs_log(mp, &in); - return -error; - } - - case XFS_IOC_FSGROWFSRT: { - xfs_growfs_rt_t in; - - if (!capable(CAP_SYS_ADMIN)) - return -EPERM; - - if (copy_from_user(&in, arg, sizeof(in))) - return -XFS_ERROR(EFAULT); - - error = xfs_growfs_rt(mp, &in); - return -error; - } - - case XFS_IOC_FREEZE: - if (!capable(CAP_SYS_ADMIN)) - return -EPERM; - - if (inode->i_sb->s_frozen == SB_UNFROZEN) - freeze_bdev(inode->i_sb->s_bdev); - return 0; - - case XFS_IOC_THAW: - if (!capable(CAP_SYS_ADMIN)) - return -EPERM; - if (inode->i_sb->s_frozen != SB_UNFROZEN) - thaw_bdev(inode->i_sb->s_bdev, inode->i_sb); - return 0; - - case XFS_IOC_GOINGDOWN: { - __uint32_t in; - - if (!capable(CAP_SYS_ADMIN)) - return -EPERM; - - if (get_user(in, (__uint32_t __user *)arg)) - return -XFS_ERROR(EFAULT); - - error = xfs_fs_goingdown(mp, in); - return -error; - } - - case XFS_IOC_ERROR_INJECTION: { - xfs_error_injection_t in; - - if (!capable(CAP_SYS_ADMIN)) - return -EPERM; - - if (copy_from_user(&in, arg, sizeof(in))) - return -XFS_ERROR(EFAULT); - - error = xfs_errortag_add(in.errtag, mp); - return -error; - } - - case XFS_IOC_ERROR_CLEARALL: - if (!capable(CAP_SYS_ADMIN)) - return -EPERM; - - error = xfs_errortag_clearall(mp, 1); - return -error; - - default: - return -ENOTTY; - } -} - STATIC int xfs_ioc_space( struct xfs_inode *ip, @@ -1398,12 +1090,10 @@ xfs_ioctl( return xfs_ioc_fsgetxattr(ip, 0, arg); case XFS_IOC_FSGETXATTRA: return xfs_ioc_fsgetxattr(ip, 1, arg); - case XFS_IOC_FSSETXATTR: - return xfs_ioc_fssetxattr(ip, filp, arg); case XFS_IOC_GETXFLAGS: - return xfs_ioc_getxflags(ip, arg); case XFS_IOC_SETXFLAGS: - return xfs_ioc_setxflags(ip, filp, arg); + case XFS_IOC_FSSETXATTR: + return xfs_ioc_xattr(ip, filp, cmd, arg); case XFS_IOC_FSSETDM: { struct fsdmidata dmi; -- cgit v1.2.3 From 65e67f5165c8a156b34ee7adf65d5ed3b16a910d Mon Sep 17 00:00:00 2001 From: Lachlan McIlroy Date: Fri, 18 Apr 2008 12:59:45 +1000 Subject: [XFS] Fix merge failure --- fs/xfs/linux-2.6/xfs_ioctl.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_ioctl.c b/fs/xfs/linux-2.6/xfs_ioctl.c index 640c8b6b195..bf775979385 100644 --- a/fs/xfs/linux-2.6/xfs_ioctl.c +++ b/fs/xfs/linux-2.6/xfs_ioctl.c @@ -1090,10 +1090,12 @@ xfs_ioctl( return xfs_ioc_fsgetxattr(ip, 0, arg); case XFS_IOC_FSGETXATTRA: return xfs_ioc_fsgetxattr(ip, 1, arg); + case XFS_IOC_FSSETXATTR: + return xfs_ioc_fssetxattr(ip, filp, arg); case XFS_IOC_GETXFLAGS: + return xfs_ioc_getxflags(ip, arg); case XFS_IOC_SETXFLAGS: - case XFS_IOC_FSSETXATTR: - return xfs_ioc_xattr(ip, filp, cmd, arg); + return xfs_ioc_setxflags(ip, filp, arg); case XFS_IOC_FSSETDM: { struct fsdmidata dmi; -- cgit v1.2.3 From 62be1f71677c53d5e51223807a06ac9052f49b0f Mon Sep 17 00:00:00 2001 From: Roel Kluin <12o3l@tiscali.nl> Date: Thu, 17 Apr 2008 17:25:37 +0200 Subject: [GFS2] fix assertion in log_refund() since unsigned, unused >= 0 is always true. Signed-off-by: Roel Kluin <12o3l@tiscali.nl> Signed-off-by: Steven Whitehouse --- fs/gfs2/log.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/gfs2/log.c b/fs/gfs2/log.c index b335304fc5d..548264b1836 100644 --- a/fs/gfs2/log.c +++ b/fs/gfs2/log.c @@ -769,8 +769,8 @@ static void log_refund(struct gfs2_sbd *sdp, struct gfs2_trans *tr) sdp->sd_log_commited_revoke += tr->tr_num_revoke - tr->tr_num_revoke_rm; gfs2_assert_withdraw(sdp, ((int)sdp->sd_log_commited_revoke) >= 0); reserved = calc_reserved(sdp); + gfs2_assert_withdraw(sdp, sdp->sd_log_blks_reserved + tr->tr_reserved >= reserved); unused = sdp->sd_log_blks_reserved - reserved + tr->tr_reserved; - gfs2_assert_withdraw(sdp, unused >= 0); atomic_add(unused, &sdp->sd_log_blks_free); gfs2_assert_withdraw(sdp, atomic_read(&sdp->sd_log_blks_free) <= sdp->sd_jdesc->jd_blocks); -- cgit v1.2.3 From 8e8a4603b5422c9145880e73b23bc4c2c4de0098 Mon Sep 17 00:00:00 2001 From: Mark Fasheh Date: Fri, 1 Feb 2008 11:59:09 -0800 Subject: ocfs2: Move slot map access into slot_map.c journal.c and dlmglue.c would refresh the slot map by hand. Instead, have the update and clear functions do the work inside slot_map.c. The eventual result is to make ocfs2_slot_info defined privately in slot_map.c Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlmglue.c | 8 +------ fs/ocfs2/journal.c | 3 +-- fs/ocfs2/slot_map.c | 62 +++++++++++++++++++++++++++++++++++++++++------------ fs/ocfs2/slot_map.h | 11 ++++------ fs/ocfs2/super.c | 3 +-- 5 files changed, 55 insertions(+), 32 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index 1f1873bf41f..1a80fa9e7c9 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -2132,8 +2132,6 @@ int ocfs2_super_lock(struct ocfs2_super *osb, int status = 0; int level = ex ? LKM_EXMODE : LKM_PRMODE; struct ocfs2_lock_res *lockres = &osb->osb_super_lockres; - struct buffer_head *bh; - struct ocfs2_slot_info *si = osb->slot_info; mlog_entry_void(); @@ -2159,11 +2157,7 @@ int ocfs2_super_lock(struct ocfs2_super *osb, goto bail; } if (status) { - bh = si->si_bh; - status = ocfs2_read_block(osb, bh->b_blocknr, &bh, 0, - si->si_inode); - if (status == 0) - ocfs2_update_slot_info(si); + status = ocfs2_refresh_slot_info(osb); ocfs2_complete_lock_res_refresh(lockres, status); diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c index f31c7e8c19c..c2e654ee703 100644 --- a/fs/ocfs2/journal.c +++ b/fs/ocfs2/journal.c @@ -1123,8 +1123,7 @@ static int ocfs2_recover_node(struct ocfs2_super *osb, /* Likewise, this would be a strange but ultimately not so * harmful place to get an error... */ - ocfs2_clear_slot(si, slot_num); - status = ocfs2_update_disk_slots(osb, si); + status = ocfs2_clear_slot(osb, slot_num); if (status < 0) mlog_errno(status); diff --git a/fs/ocfs2/slot_map.c b/fs/ocfs2/slot_map.c index 3a50ce555e6..f5727b8cc91 100644 --- a/fs/ocfs2/slot_map.c +++ b/fs/ocfs2/slot_map.c @@ -49,7 +49,7 @@ static void __ocfs2_fill_slot(struct ocfs2_slot_info *si, s16 node_num); /* post the slot information on disk into our slot_info struct. */ -void ocfs2_update_slot_info(struct ocfs2_slot_info *si) +static void ocfs2_update_slot_info(struct ocfs2_slot_info *si) { int i; __le16 *disk_info; @@ -65,10 +65,27 @@ void ocfs2_update_slot_info(struct ocfs2_slot_info *si) spin_unlock(&si->si_lock); } +int ocfs2_refresh_slot_info(struct ocfs2_super *osb) +{ + int ret; + struct ocfs2_slot_info *si = osb->slot_info; + struct buffer_head *bh; + + if (si == NULL) + return 0; + + bh = si->si_bh; + ret = ocfs2_read_block(osb, bh->b_blocknr, &bh, 0, si->si_inode); + if (ret == 0) + ocfs2_update_slot_info(si); + + return ret; +} + /* post the our slot info stuff into it's destination bh and write it * out. */ -int ocfs2_update_disk_slots(struct ocfs2_super *osb, - struct ocfs2_slot_info *si) +static int ocfs2_update_disk_slots(struct ocfs2_super *osb, + struct ocfs2_slot_info *si) { int status, i; __le16 *disk_info = (__le16 *) si->si_bh->b_data; @@ -135,6 +152,19 @@ s16 ocfs2_node_num_to_slot(struct ocfs2_slot_info *si, return ret; } +static void __ocfs2_free_slot_info(struct ocfs2_slot_info *si) +{ + if (si == NULL) + return; + + if (si->si_inode) + iput(si->si_inode); + if (si->si_bh) + brelse(si->si_bh); + + kfree(si); +} + static void __ocfs2_fill_slot(struct ocfs2_slot_info *si, s16 slot_num, s16 node_num) @@ -147,12 +177,18 @@ static void __ocfs2_fill_slot(struct ocfs2_slot_info *si, si->si_global_node_nums[slot_num] = node_num; } -void ocfs2_clear_slot(struct ocfs2_slot_info *si, - s16 slot_num) +int ocfs2_clear_slot(struct ocfs2_super *osb, s16 slot_num) { + struct ocfs2_slot_info *si = osb->slot_info; + + if (si == NULL) + return 0; + spin_lock(&si->si_lock); __ocfs2_fill_slot(si, slot_num, OCFS2_INVALID_SLOT); spin_unlock(&si->si_lock); + + return ocfs2_update_disk_slots(osb, osb->slot_info); } int ocfs2_init_slot_info(struct ocfs2_super *osb) @@ -202,18 +238,17 @@ int ocfs2_init_slot_info(struct ocfs2_super *osb) osb->slot_info = si; bail: if (status < 0 && si) - ocfs2_free_slot_info(si); + __ocfs2_free_slot_info(si); return status; } -void ocfs2_free_slot_info(struct ocfs2_slot_info *si) +void ocfs2_free_slot_info(struct ocfs2_super *osb) { - if (si->si_inode) - iput(si->si_inode); - if (si->si_bh) - brelse(si->si_bh); - kfree(si); + struct ocfs2_slot_info *si = osb->slot_info; + + osb->slot_info = NULL; + __ocfs2_free_slot_info(si); } int ocfs2_find_slot(struct ocfs2_super *osb) @@ -285,7 +320,6 @@ void ocfs2_put_slot(struct ocfs2_super *osb) } bail: - osb->slot_info = NULL; - ocfs2_free_slot_info(si); + ocfs2_free_slot_info(osb); } diff --git a/fs/ocfs2/slot_map.h b/fs/ocfs2/slot_map.h index 1025872aaad..b029ffdc8ea 100644 --- a/fs/ocfs2/slot_map.h +++ b/fs/ocfs2/slot_map.h @@ -30,7 +30,7 @@ struct ocfs2_slot_info { spinlock_t si_lock; - struct inode *si_inode; + struct inode *si_inode; struct buffer_head *si_bh; unsigned int si_num_slots; unsigned int si_size; @@ -38,19 +38,16 @@ struct ocfs2_slot_info { }; int ocfs2_init_slot_info(struct ocfs2_super *osb); -void ocfs2_free_slot_info(struct ocfs2_slot_info *si); +void ocfs2_free_slot_info(struct ocfs2_super *osb); int ocfs2_find_slot(struct ocfs2_super *osb); void ocfs2_put_slot(struct ocfs2_super *osb); -void ocfs2_update_slot_info(struct ocfs2_slot_info *si); -int ocfs2_update_disk_slots(struct ocfs2_super *osb, - struct ocfs2_slot_info *si); +int ocfs2_refresh_slot_info(struct ocfs2_super *osb); s16 ocfs2_node_num_to_slot(struct ocfs2_slot_info *si, s16 global); -void ocfs2_clear_slot(struct ocfs2_slot_info *si, - s16 slot_num); +int ocfs2_clear_slot(struct ocfs2_super *osb, s16 slot_num); static inline int ocfs2_is_empty_slot(struct ocfs2_slot_info *si, int slot_num) diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index bec75aff3d9..fad37af2af9 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -1724,8 +1724,7 @@ static void ocfs2_delete_osb(struct ocfs2_super *osb) /* This function assumes that the caller has the main osb resource */ - if (osb->slot_info) - ocfs2_free_slot_info(osb->slot_info); + ocfs2_free_slot_info(osb); kfree(osb->osb_orphan_wipes); /* FIXME -- cgit v1.2.3 From d85b20e4b300edfd290f21fc2d790ba16d2f225b Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Fri, 1 Feb 2008 12:01:05 -0800 Subject: ocfs2: Make ocfs2_slot_info private. Just use osb_lock around the ocfs2_slot_info data. This allows us to take the ocfs2_slot_info structure private in slot_info.c. All access is now via accessors. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/journal.c | 24 ++++++++-------- fs/ocfs2/ocfs2.h | 1 + fs/ocfs2/slot_map.c | 81 +++++++++++++++++++++++++++++++++++++---------------- fs/ocfs2/slot_map.h | 25 +++-------------- 4 files changed, 74 insertions(+), 57 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c index c2e654ee703..ed0c6d0850d 100644 --- a/fs/ocfs2/journal.c +++ b/fs/ocfs2/journal.c @@ -1079,7 +1079,6 @@ static int ocfs2_recover_node(struct ocfs2_super *osb, { int status = 0; int slot_num; - struct ocfs2_slot_info *si = osb->slot_info; struct ocfs2_dinode *la_copy = NULL; struct ocfs2_dinode *tl_copy = NULL; @@ -1092,8 +1091,8 @@ static int ocfs2_recover_node(struct ocfs2_super *osb, * case we should've called ocfs2_journal_load instead. */ BUG_ON(osb->node_num == node_num); - slot_num = ocfs2_node_num_to_slot(si, node_num); - if (slot_num == OCFS2_INVALID_SLOT) { + slot_num = ocfs2_node_num_to_slot(osb, node_num); + if (slot_num == -ENOENT) { status = 0; mlog(0, "no slot for this node, so no recovery required.\n"); goto done; @@ -1183,23 +1182,24 @@ bail: * slot info struct has been updated from disk. */ int ocfs2_mark_dead_nodes(struct ocfs2_super *osb) { - int status, i, node_num; - struct ocfs2_slot_info *si = osb->slot_info; + unsigned int node_num; + int status, i; /* This is called with the super block cluster lock, so we * know that the slot map can't change underneath us. */ - spin_lock(&si->si_lock); - for(i = 0; i < si->si_num_slots; i++) { + spin_lock(&osb->osb_lock); + for (i = 0; i < osb->max_slots; i++) { if (i == osb->slot_num) continue; - if (ocfs2_is_empty_slot(si, i)) + + status = ocfs2_slot_to_node_num_locked(osb, i, &node_num); + if (status == -ENOENT) continue; - node_num = si->si_global_node_nums[i]; if (ocfs2_node_map_test_bit(osb, &osb->recovery_map, node_num)) continue; - spin_unlock(&si->si_lock); + spin_unlock(&osb->osb_lock); /* Ok, we have a slot occupied by another node which * is not in the recovery map. We trylock his journal @@ -1215,9 +1215,9 @@ int ocfs2_mark_dead_nodes(struct ocfs2_super *osb) goto bail; } - spin_lock(&si->si_lock); + spin_lock(&osb->osb_lock); } - spin_unlock(&si->si_lock); + spin_unlock(&osb->osb_lock); status = 0; bail: diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h index 6546cef212e..ee3f675a421 100644 --- a/fs/ocfs2/ocfs2.h +++ b/fs/ocfs2/ocfs2.h @@ -179,6 +179,7 @@ enum ocfs2_mount_options #define OCFS2_DEFAULT_ATIME_QUANTUM 60 struct ocfs2_journal; +struct ocfs2_slot_info; struct ocfs2_super { struct task_struct *commit_task; diff --git a/fs/ocfs2/slot_map.c b/fs/ocfs2/slot_map.c index f5727b8cc91..762360d95e9 100644 --- a/fs/ocfs2/slot_map.c +++ b/fs/ocfs2/slot_map.c @@ -42,13 +42,25 @@ #include "buffer_head_io.h" +struct ocfs2_slot_info { + struct inode *si_inode; + struct buffer_head *si_bh; + unsigned int si_num_slots; + unsigned int si_size; + s16 si_global_node_nums[OCFS2_MAX_SLOTS]; +}; + + static s16 __ocfs2_node_num_to_slot(struct ocfs2_slot_info *si, s16 global); static void __ocfs2_fill_slot(struct ocfs2_slot_info *si, s16 slot_num, s16 node_num); -/* post the slot information on disk into our slot_info struct. */ +/* + * Post the slot information on disk into our slot_info struct. + * Must be protected by osb_lock. + */ static void ocfs2_update_slot_info(struct ocfs2_slot_info *si) { int i; @@ -56,13 +68,10 @@ static void ocfs2_update_slot_info(struct ocfs2_slot_info *si) /* we don't read the slot block here as ocfs2_super_lock * should've made sure we have the most recent copy. */ - spin_lock(&si->si_lock); disk_info = (__le16 *) si->si_bh->b_data; for (i = 0; i < si->si_size; i++) si->si_global_node_nums[i] = le16_to_cpu(disk_info[i]); - - spin_unlock(&si->si_lock); } int ocfs2_refresh_slot_info(struct ocfs2_super *osb) @@ -76,8 +85,11 @@ int ocfs2_refresh_slot_info(struct ocfs2_super *osb) bh = si->si_bh; ret = ocfs2_read_block(osb, bh->b_blocknr, &bh, 0, si->si_inode); - if (ret == 0) + if (ret == 0) { + spin_lock(&osb->osb_lock); ocfs2_update_slot_info(si); + spin_unlock(&osb->osb_lock); + } return ret; } @@ -90,10 +102,10 @@ static int ocfs2_update_disk_slots(struct ocfs2_super *osb, int status, i; __le16 *disk_info = (__le16 *) si->si_bh->b_data; - spin_lock(&si->si_lock); + spin_lock(&osb->osb_lock); for (i = 0; i < si->si_size; i++) disk_info[i] = cpu_to_le16(si->si_global_node_nums[i]); - spin_unlock(&si->si_lock); + spin_unlock(&osb->osb_lock); status = ocfs2_write_block(osb, si->si_bh, si->si_inode); if (status < 0) @@ -119,7 +131,8 @@ static s16 __ocfs2_node_num_to_slot(struct ocfs2_slot_info *si, return ret; } -static s16 __ocfs2_find_empty_slot(struct ocfs2_slot_info *si, s16 preferred) +static s16 __ocfs2_find_empty_slot(struct ocfs2_slot_info *si, + s16 preferred) { int i; s16 ret = OCFS2_INVALID_SLOT; @@ -141,15 +154,36 @@ out: return ret; } -s16 ocfs2_node_num_to_slot(struct ocfs2_slot_info *si, - s16 global) +int ocfs2_node_num_to_slot(struct ocfs2_super *osb, unsigned int node_num) { - s16 ret; + s16 slot; + struct ocfs2_slot_info *si = osb->slot_info; - spin_lock(&si->si_lock); - ret = __ocfs2_node_num_to_slot(si, global); - spin_unlock(&si->si_lock); - return ret; + spin_lock(&osb->osb_lock); + slot = __ocfs2_node_num_to_slot(si, node_num); + spin_unlock(&osb->osb_lock); + + if (slot == OCFS2_INVALID_SLOT) + return -ENOENT; + + return slot; +} + +int ocfs2_slot_to_node_num_locked(struct ocfs2_super *osb, int slot_num, + unsigned int *node_num) +{ + struct ocfs2_slot_info *si = osb->slot_info; + + assert_spin_locked(&osb->osb_lock); + + BUG_ON(slot_num < 0); + BUG_ON(slot_num > osb->max_slots); + + if (si->si_global_node_nums[slot_num] == OCFS2_INVALID_SLOT) + return -ENOENT; + + *node_num = si->si_global_node_nums[slot_num]; + return 0; } static void __ocfs2_free_slot_info(struct ocfs2_slot_info *si) @@ -184,9 +218,9 @@ int ocfs2_clear_slot(struct ocfs2_super *osb, s16 slot_num) if (si == NULL) return 0; - spin_lock(&si->si_lock); + spin_lock(&osb->osb_lock); __ocfs2_fill_slot(si, slot_num, OCFS2_INVALID_SLOT); - spin_unlock(&si->si_lock); + spin_unlock(&osb->osb_lock); return ocfs2_update_disk_slots(osb, osb->slot_info); } @@ -206,7 +240,6 @@ int ocfs2_init_slot_info(struct ocfs2_super *osb) goto bail; } - spin_lock_init(&si->si_lock); si->si_num_slots = osb->max_slots; si->si_size = OCFS2_MAX_SLOTS; @@ -235,7 +268,7 @@ int ocfs2_init_slot_info(struct ocfs2_super *osb) si->si_inode = inode; si->si_bh = bh; - osb->slot_info = si; + osb->slot_info = (struct ocfs2_slot_info *)si; bail: if (status < 0 && si) __ocfs2_free_slot_info(si); @@ -261,9 +294,9 @@ int ocfs2_find_slot(struct ocfs2_super *osb) si = osb->slot_info; + spin_lock(&osb->osb_lock); ocfs2_update_slot_info(si); - spin_lock(&si->si_lock); /* search for ourselves first and take the slot if it already * exists. Perhaps we need to mark this in a variable for our * own journal recovery? Possibly not, though we certainly @@ -274,7 +307,7 @@ int ocfs2_find_slot(struct ocfs2_super *osb) * one. */ slot = __ocfs2_find_empty_slot(si, osb->preferred_slot); if (slot == OCFS2_INVALID_SLOT) { - spin_unlock(&si->si_lock); + spin_unlock(&osb->osb_lock); mlog(ML_ERROR, "no free slots available!\n"); status = -EINVAL; goto bail; @@ -285,7 +318,7 @@ int ocfs2_find_slot(struct ocfs2_super *osb) __ocfs2_fill_slot(si, slot, osb->node_num); osb->slot_num = slot; - spin_unlock(&si->si_lock); + spin_unlock(&osb->osb_lock); mlog(0, "taking node slot %d\n", osb->slot_num); @@ -306,12 +339,12 @@ void ocfs2_put_slot(struct ocfs2_super *osb) if (!si) return; + spin_lock(&osb->osb_lock); ocfs2_update_slot_info(si); - spin_lock(&si->si_lock); __ocfs2_fill_slot(si, osb->slot_num, OCFS2_INVALID_SLOT); osb->slot_num = OCFS2_INVALID_SLOT; - spin_unlock(&si->si_lock); + spin_unlock(&osb->osb_lock); status = ocfs2_update_disk_slots(osb, si); if (status < 0) { diff --git a/fs/ocfs2/slot_map.h b/fs/ocfs2/slot_map.h index b029ffdc8ea..5118e89c84e 100644 --- a/fs/ocfs2/slot_map.h +++ b/fs/ocfs2/slot_map.h @@ -27,16 +27,6 @@ #ifndef SLOTMAP_H #define SLOTMAP_H -struct ocfs2_slot_info { - spinlock_t si_lock; - - struct inode *si_inode; - struct buffer_head *si_bh; - unsigned int si_num_slots; - unsigned int si_size; - s16 si_global_node_nums[OCFS2_MAX_SLOTS]; -}; - int ocfs2_init_slot_info(struct ocfs2_super *osb); void ocfs2_free_slot_info(struct ocfs2_super *osb); @@ -45,17 +35,10 @@ void ocfs2_put_slot(struct ocfs2_super *osb); int ocfs2_refresh_slot_info(struct ocfs2_super *osb); -s16 ocfs2_node_num_to_slot(struct ocfs2_slot_info *si, - s16 global); -int ocfs2_clear_slot(struct ocfs2_super *osb, s16 slot_num); +int ocfs2_node_num_to_slot(struct ocfs2_super *osb, unsigned int node_num); +int ocfs2_slot_to_node_num_locked(struct ocfs2_super *osb, int slot_num, + unsigned int *node_num); -static inline int ocfs2_is_empty_slot(struct ocfs2_slot_info *si, - int slot_num) -{ - BUG_ON(slot_num == OCFS2_INVALID_SLOT); - assert_spin_locked(&si->si_lock); - - return si->si_global_node_nums[slot_num] == OCFS2_INVALID_SLOT; -} +int ocfs2_clear_slot(struct ocfs2_super *osb, s16 slot_num); #endif -- cgit v1.2.3 From 553abd046af609191a91af7289d87d477adc659f Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Fri, 1 Feb 2008 12:03:57 -0800 Subject: ocfs2: Change the recovery map to an array of node numbers. The old recovery map was a bitmap of node numbers. This was sufficient for the maximum node number of 254. Going forward, we want node numbers to be UINT32. Thus, we need a new recovery map. Note that we can't keep track of slots here. We must write down the node number to recovery *before* we get the locks needed to convert a node number into a slot number. The recovery map is now an array of unsigned ints, max_slots in size. It moves to journal.c with the rest of recovery. Because it needs to be initialized, we move all of recovery initialization into a new function, ocfs2_recovery_init(). This actually cleans up ocfs2_initialize_super() a little as well. Following on, recovery cleaup becomes part of ocfs2_recovery_exit(). A number of node map functions are rendered obsolete and are removed. Finally, waiting on recovery is wrapped in a function rather than naked checks on the recovery_event. This is a cleanup from Mark. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlmglue.c | 6 +- fs/ocfs2/heartbeat.c | 111 ------------------------------- fs/ocfs2/heartbeat.h | 14 ---- fs/ocfs2/journal.c | 181 ++++++++++++++++++++++++++++++++++++++++++++++----- fs/ocfs2/journal.h | 4 ++ fs/ocfs2/ocfs2.h | 3 +- fs/ocfs2/super.c | 33 +++------- 7 files changed, 182 insertions(+), 170 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index 1a80fa9e7c9..15a5167e051 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -1950,8 +1950,7 @@ int ocfs2_inode_lock_full(struct inode *inode, goto local; if (!(arg_flags & OCFS2_META_LOCK_RECOVERY)) - wait_event(osb->recovery_event, - ocfs2_node_map_is_empty(osb, &osb->recovery_map)); + ocfs2_wait_for_recovery(osb); lockres = &OCFS2_I(inode)->ip_inode_lockres; level = ex ? LKM_EXMODE : LKM_PRMODE; @@ -1974,8 +1973,7 @@ int ocfs2_inode_lock_full(struct inode *inode, * committed to owning this lock so we don't allow signals to * abort the operation. */ if (!(arg_flags & OCFS2_META_LOCK_RECOVERY)) - wait_event(osb->recovery_event, - ocfs2_node_map_is_empty(osb, &osb->recovery_map)); + ocfs2_wait_for_recovery(osb); local: /* diff --git a/fs/ocfs2/heartbeat.c b/fs/ocfs2/heartbeat.c index 0758daf64da..80de2397c16 100644 --- a/fs/ocfs2/heartbeat.c +++ b/fs/ocfs2/heartbeat.c @@ -48,7 +48,6 @@ static inline void __ocfs2_node_map_set_bit(struct ocfs2_node_map *map, int bit); static inline void __ocfs2_node_map_clear_bit(struct ocfs2_node_map *map, int bit); -static inline int __ocfs2_node_map_is_empty(struct ocfs2_node_map *map); /* special case -1 for now * TODO: should *really* make sure the calling func never passes -1!! */ @@ -62,7 +61,6 @@ static void ocfs2_node_map_init(struct ocfs2_node_map *map) void ocfs2_init_node_maps(struct ocfs2_super *osb) { spin_lock_init(&osb->node_map_lock); - ocfs2_node_map_init(&osb->recovery_map); ocfs2_node_map_init(&osb->osb_recovering_orphan_dirs); } @@ -192,112 +190,3 @@ int ocfs2_node_map_test_bit(struct ocfs2_super *osb, return ret; } -static inline int __ocfs2_node_map_is_empty(struct ocfs2_node_map *map) -{ - int bit; - bit = find_next_bit(map->map, map->num_nodes, 0); - if (bit < map->num_nodes) - return 0; - return 1; -} - -int ocfs2_node_map_is_empty(struct ocfs2_super *osb, - struct ocfs2_node_map *map) -{ - int ret; - BUG_ON(map->num_nodes == 0); - spin_lock(&osb->node_map_lock); - ret = __ocfs2_node_map_is_empty(map); - spin_unlock(&osb->node_map_lock); - return ret; -} - -#if 0 - -static void __ocfs2_node_map_dup(struct ocfs2_node_map *target, - struct ocfs2_node_map *from) -{ - BUG_ON(from->num_nodes == 0); - ocfs2_node_map_init(target); - __ocfs2_node_map_set(target, from); -} - -/* returns 1 if bit is the only bit set in target, 0 otherwise */ -int ocfs2_node_map_is_only(struct ocfs2_super *osb, - struct ocfs2_node_map *target, - int bit) -{ - struct ocfs2_node_map temp; - int ret; - - spin_lock(&osb->node_map_lock); - __ocfs2_node_map_dup(&temp, target); - __ocfs2_node_map_clear_bit(&temp, bit); - ret = __ocfs2_node_map_is_empty(&temp); - spin_unlock(&osb->node_map_lock); - - return ret; -} - -static void __ocfs2_node_map_set(struct ocfs2_node_map *target, - struct ocfs2_node_map *from) -{ - int num_longs, i; - - BUG_ON(target->num_nodes != from->num_nodes); - BUG_ON(target->num_nodes == 0); - - num_longs = BITS_TO_LONGS(target->num_nodes); - for (i = 0; i < num_longs; i++) - target->map[i] = from->map[i]; -} - -#endif /* 0 */ - -/* Returns whether the recovery bit was actually set - it may not be - * if a node is still marked as needing recovery */ -int ocfs2_recovery_map_set(struct ocfs2_super *osb, - int num) -{ - int set = 0; - - spin_lock(&osb->node_map_lock); - - if (!test_bit(num, osb->recovery_map.map)) { - __ocfs2_node_map_set_bit(&osb->recovery_map, num); - set = 1; - } - - spin_unlock(&osb->node_map_lock); - - return set; -} - -void ocfs2_recovery_map_clear(struct ocfs2_super *osb, - int num) -{ - ocfs2_node_map_clear_bit(osb, &osb->recovery_map, num); -} - -int ocfs2_node_map_iterate(struct ocfs2_super *osb, - struct ocfs2_node_map *map, - int idx) -{ - int i = idx; - - idx = O2NM_INVALID_NODE_NUM; - spin_lock(&osb->node_map_lock); - if ((i != O2NM_INVALID_NODE_NUM) && - (i >= 0) && - (i < map->num_nodes)) { - while(i < map->num_nodes) { - if (test_bit(i, map->map)) { - idx = i; - break; - } - i++; - } - } - spin_unlock(&osb->node_map_lock); - return idx; -} diff --git a/fs/ocfs2/heartbeat.h b/fs/ocfs2/heartbeat.h index eac63aed761..98d8ffc995b 100644 --- a/fs/ocfs2/heartbeat.h +++ b/fs/ocfs2/heartbeat.h @@ -33,8 +33,6 @@ void ocfs2_stop_heartbeat(struct ocfs2_super *osb); /* node map functions - used to keep track of mounted and in-recovery * nodes. */ -int ocfs2_node_map_is_empty(struct ocfs2_super *osb, - struct ocfs2_node_map *map); void ocfs2_node_map_set_bit(struct ocfs2_super *osb, struct ocfs2_node_map *map, int bit); @@ -44,17 +42,5 @@ void ocfs2_node_map_clear_bit(struct ocfs2_super *osb, int ocfs2_node_map_test_bit(struct ocfs2_super *osb, struct ocfs2_node_map *map, int bit); -int ocfs2_node_map_iterate(struct ocfs2_super *osb, - struct ocfs2_node_map *map, - int idx); -static inline int ocfs2_node_map_first_set_bit(struct ocfs2_super *osb, - struct ocfs2_node_map *map) -{ - return ocfs2_node_map_iterate(osb, map, 0); -} -int ocfs2_recovery_map_set(struct ocfs2_super *osb, - int num); -void ocfs2_recovery_map_clear(struct ocfs2_super *osb, - int num); #endif /* OCFS2_HEARTBEAT_H */ diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c index ed0c6d0850d..ca4c0ea5a4c 100644 --- a/fs/ocfs2/journal.c +++ b/fs/ocfs2/journal.c @@ -64,6 +64,137 @@ static int ocfs2_recover_orphans(struct ocfs2_super *osb, int slot); static int ocfs2_commit_thread(void *arg); + +/* + * The recovery_list is a simple linked list of node numbers to recover. + * It is protected by the recovery_lock. + */ + +struct ocfs2_recovery_map { + int rm_used; + unsigned int *rm_entries; +}; + +int ocfs2_recovery_init(struct ocfs2_super *osb) +{ + struct ocfs2_recovery_map *rm; + + mutex_init(&osb->recovery_lock); + osb->disable_recovery = 0; + osb->recovery_thread_task = NULL; + init_waitqueue_head(&osb->recovery_event); + + rm = kzalloc(sizeof(struct ocfs2_recovery_map) + + osb->max_slots * sizeof(unsigned int), + GFP_KERNEL); + if (!rm) { + mlog_errno(-ENOMEM); + return -ENOMEM; + } + + rm->rm_entries = (unsigned int *)((char *)rm + + sizeof(struct ocfs2_recovery_map)); + osb->recovery_map = rm; + + return 0; +} + +/* we can't grab the goofy sem lock from inside wait_event, so we use + * memory barriers to make sure that we'll see the null task before + * being woken up */ +static int ocfs2_recovery_thread_running(struct ocfs2_super *osb) +{ + mb(); + return osb->recovery_thread_task != NULL; +} + +void ocfs2_recovery_exit(struct ocfs2_super *osb) +{ + struct ocfs2_recovery_map *rm; + + /* disable any new recovery threads and wait for any currently + * running ones to exit. Do this before setting the vol_state. */ + mutex_lock(&osb->recovery_lock); + osb->disable_recovery = 1; + mutex_unlock(&osb->recovery_lock); + wait_event(osb->recovery_event, !ocfs2_recovery_thread_running(osb)); + + /* At this point, we know that no more recovery threads can be + * launched, so wait for any recovery completion work to + * complete. */ + flush_workqueue(ocfs2_wq); + + /* + * Now that recovery is shut down, and the osb is about to be + * freed, the osb_lock is not taken here. + */ + rm = osb->recovery_map; + /* XXX: Should we bug if there are dirty entries? */ + + kfree(rm); +} + +static int __ocfs2_recovery_map_test(struct ocfs2_super *osb, + unsigned int node_num) +{ + int i; + struct ocfs2_recovery_map *rm = osb->recovery_map; + + assert_spin_locked(&osb->osb_lock); + + for (i = 0; i < rm->rm_used; i++) { + if (rm->rm_entries[i] == node_num) + return 1; + } + + return 0; +} + +/* Behaves like test-and-set. Returns the previous value */ +static int ocfs2_recovery_map_set(struct ocfs2_super *osb, + unsigned int node_num) +{ + struct ocfs2_recovery_map *rm = osb->recovery_map; + + spin_lock(&osb->osb_lock); + if (__ocfs2_recovery_map_test(osb, node_num)) { + spin_unlock(&osb->osb_lock); + return 1; + } + + /* XXX: Can this be exploited? Not from o2dlm... */ + BUG_ON(rm->rm_used >= osb->max_slots); + + rm->rm_entries[rm->rm_used] = node_num; + rm->rm_used++; + spin_unlock(&osb->osb_lock); + + return 0; +} + +static void ocfs2_recovery_map_clear(struct ocfs2_super *osb, + unsigned int node_num) +{ + int i; + struct ocfs2_recovery_map *rm = osb->recovery_map; + + spin_lock(&osb->osb_lock); + + for (i = 0; i < rm->rm_used; i++) { + if (rm->rm_entries[i] == node_num) + break; + } + + if (i < rm->rm_used) { + /* XXX: be careful with the pointer math */ + memmove(&(rm->rm_entries[i]), &(rm->rm_entries[i + 1]), + (rm->rm_used - i - 1) * sizeof(unsigned int)); + rm->rm_used--; + } + + spin_unlock(&osb->osb_lock); +} + static int ocfs2_commit_cache(struct ocfs2_super *osb) { int status = 0; @@ -650,6 +781,23 @@ bail: return status; } +static int ocfs2_recovery_completed(struct ocfs2_super *osb) +{ + int empty; + struct ocfs2_recovery_map *rm = osb->recovery_map; + + spin_lock(&osb->osb_lock); + empty = (rm->rm_used == 0); + spin_unlock(&osb->osb_lock); + + return empty; +} + +void ocfs2_wait_for_recovery(struct ocfs2_super *osb) +{ + wait_event(osb->recovery_event, ocfs2_recovery_completed(osb)); +} + /* * JBD Might read a cached version of another nodes journal file. We * don't want this as this file changes often and we get no @@ -848,6 +996,7 @@ static int __ocfs2_recovery_thread(void *arg) { int status, node_num; struct ocfs2_super *osb = arg; + struct ocfs2_recovery_map *rm = osb->recovery_map; mlog_entry_void(); @@ -863,26 +1012,29 @@ restart: goto bail; } - while(!ocfs2_node_map_is_empty(osb, &osb->recovery_map)) { - node_num = ocfs2_node_map_first_set_bit(osb, - &osb->recovery_map); - if (node_num == O2NM_INVALID_NODE_NUM) { - mlog(0, "Out of nodes to recover.\n"); - break; - } + spin_lock(&osb->osb_lock); + while (rm->rm_used) { + /* It's always safe to remove entry zero, as we won't + * clear it until ocfs2_recover_node() has succeeded. */ + node_num = rm->rm_entries[0]; + spin_unlock(&osb->osb_lock); status = ocfs2_recover_node(osb, node_num); - if (status < 0) { + if (!status) { + ocfs2_recovery_map_clear(osb, node_num); + } else { mlog(ML_ERROR, "Error %d recovering node %d on device (%u,%u)!\n", status, node_num, MAJOR(osb->sb->s_dev), MINOR(osb->sb->s_dev)); mlog(ML_ERROR, "Volume requires unmount.\n"); - continue; } - ocfs2_recovery_map_clear(osb, node_num); + spin_lock(&osb->osb_lock); } + spin_unlock(&osb->osb_lock); + mlog(0, "All nodes recovered\n"); + ocfs2_super_unlock(osb, 1); /* We always run recovery on our own orphan dir - the dead @@ -893,8 +1045,7 @@ restart: bail: mutex_lock(&osb->recovery_lock); - if (!status && - !ocfs2_node_map_is_empty(osb, &osb->recovery_map)) { + if (!status && !ocfs2_recovery_completed(osb)) { mutex_unlock(&osb->recovery_lock); goto restart; } @@ -924,8 +1075,8 @@ void ocfs2_recovery_thread(struct ocfs2_super *osb, int node_num) /* People waiting on recovery will wait on * the recovery map to empty. */ - if (!ocfs2_recovery_map_set(osb, node_num)) - mlog(0, "node %d already be in recovery.\n", node_num); + if (ocfs2_recovery_map_set(osb, node_num)) + mlog(0, "node %d already in recovery map.\n", node_num); mlog(0, "starting recovery thread...\n"); @@ -1197,7 +1348,7 @@ int ocfs2_mark_dead_nodes(struct ocfs2_super *osb) if (status == -ENOENT) continue; - if (ocfs2_node_map_test_bit(osb, &osb->recovery_map, node_num)) + if (__ocfs2_recovery_map_test(osb, node_num)) continue; spin_unlock(&osb->osb_lock); diff --git a/fs/ocfs2/journal.h b/fs/ocfs2/journal.h index 220f3e818e7..db82be2532e 100644 --- a/fs/ocfs2/journal.h +++ b/fs/ocfs2/journal.h @@ -134,6 +134,10 @@ static inline void ocfs2_inode_set_new(struct ocfs2_super *osb, /* Exported only for the journal struct init code in super.c. Do not call. */ void ocfs2_complete_recovery(struct work_struct *work); +void ocfs2_wait_for_recovery(struct ocfs2_super *osb); + +int ocfs2_recovery_init(struct ocfs2_super *osb); +void ocfs2_recovery_exit(struct ocfs2_super *osb); /* * Journal Control: diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h index ee3f675a421..c6ed8c35de0 100644 --- a/fs/ocfs2/ocfs2.h +++ b/fs/ocfs2/ocfs2.h @@ -180,6 +180,7 @@ enum ocfs2_mount_options struct ocfs2_journal; struct ocfs2_slot_info; +struct ocfs2_recovery_map; struct ocfs2_super { struct task_struct *commit_task; @@ -191,7 +192,6 @@ struct ocfs2_super struct ocfs2_slot_info *slot_info; spinlock_t node_map_lock; - struct ocfs2_node_map recovery_map; u64 root_blkno; u64 system_dir_blkno; @@ -226,6 +226,7 @@ struct ocfs2_super atomic_t vol_state; struct mutex recovery_lock; + struct ocfs2_recovery_map *recovery_map; struct task_struct *recovery_thread_task; int disable_recovery; wait_queue_head_t checkpoint_event; diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index fad37af2af9..1a4c7c7850f 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -1224,15 +1224,6 @@ leave: return status; } -/* we can't grab the goofy sem lock from inside wait_event, so we use - * memory barriers to make sure that we'll see the null task before - * being woken up */ -static int ocfs2_recovery_thread_running(struct ocfs2_super *osb) -{ - mb(); - return osb->recovery_thread_task != NULL; -} - static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err) { int tmp; @@ -1249,17 +1240,8 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err) ocfs2_truncate_log_shutdown(osb); - /* disable any new recovery threads and wait for any currently - * running ones to exit. Do this before setting the vol_state. */ - mutex_lock(&osb->recovery_lock); - osb->disable_recovery = 1; - mutex_unlock(&osb->recovery_lock); - wait_event(osb->recovery_event, !ocfs2_recovery_thread_running(osb)); - - /* At this point, we know that no more recovery threads can be - * launched, so wait for any recovery completion work to - * complete. */ - flush_workqueue(ocfs2_wq); + /* This will disable recovery and flush any recovery work. */ + ocfs2_recovery_exit(osb); ocfs2_journal_shutdown(osb); @@ -1368,7 +1350,6 @@ static int ocfs2_initialize_super(struct super_block *sb, osb->s_sectsize_bits = blksize_bits(sector_size); BUG_ON(!osb->s_sectsize_bits); - init_waitqueue_head(&osb->recovery_event); spin_lock_init(&osb->dc_task_lock); init_waitqueue_head(&osb->dc_event); osb->dc_work_sequence = 0; @@ -1388,10 +1369,12 @@ static int ocfs2_initialize_super(struct super_block *sb, snprintf(osb->dev_str, sizeof(osb->dev_str), "%u,%u", MAJOR(osb->sb->s_dev), MINOR(osb->sb->s_dev)); - mutex_init(&osb->recovery_lock); - - osb->disable_recovery = 0; - osb->recovery_thread_task = NULL; + status = ocfs2_recovery_init(osb); + if (status) { + mlog(ML_ERROR, "Unable to initialize recovery state\n"); + mlog_errno(status); + goto bail; + } init_waitqueue_head(&osb->checkpoint_event); atomic_set(&osb->needs_checkpoint, 0); -- cgit v1.2.3 From 1c8d9a6a330f46b3a6ddd204a2580131d5f0d6b7 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Fri, 1 Feb 2008 11:59:07 -0800 Subject: ocfs2: slot_map I/O based on max_slots. The slot map code assumed a slot_map file has one block allocated. This changes the code to I/O as many blocks as will cover max_slots. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/slot_map.c | 128 ++++++++++++++++++++++++++++++++++++++++++++-------- 1 file changed, 108 insertions(+), 20 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/slot_map.c b/fs/ocfs2/slot_map.c index 762360d95e9..5bddee11009 100644 --- a/fs/ocfs2/slot_map.c +++ b/fs/ocfs2/slot_map.c @@ -44,7 +44,8 @@ struct ocfs2_slot_info { struct inode *si_inode; - struct buffer_head *si_bh; + unsigned int si_blocks; + struct buffer_head **si_bh; unsigned int si_num_slots; unsigned int si_size; s16 si_global_node_nums[OCFS2_MAX_SLOTS]; @@ -68,7 +69,7 @@ static void ocfs2_update_slot_info(struct ocfs2_slot_info *si) /* we don't read the slot block here as ocfs2_super_lock * should've made sure we have the most recent copy. */ - disk_info = (__le16 *) si->si_bh->b_data; + disk_info = (__le16 *) si->si_bh[0]->b_data; for (i = 0; i < si->si_size; i++) si->si_global_node_nums[i] = le16_to_cpu(disk_info[i]); @@ -78,13 +79,23 @@ int ocfs2_refresh_slot_info(struct ocfs2_super *osb) { int ret; struct ocfs2_slot_info *si = osb->slot_info; - struct buffer_head *bh; if (si == NULL) return 0; - bh = si->si_bh; - ret = ocfs2_read_block(osb, bh->b_blocknr, &bh, 0, si->si_inode); + BUG_ON(si->si_blocks == 0); + BUG_ON(si->si_bh == NULL); + + mlog(0, "Refreshing slot map, reading %u block(s)\n", + si->si_blocks); + + /* + * We pass -1 as blocknr because we expect all of si->si_bh to + * be !NULL. Thus, ocfs2_read_blocks() will ignore blocknr. If + * this is not true, the read of -1 (UINT64_MAX) will fail. + */ + ret = ocfs2_read_blocks(osb, -1, si->si_blocks, si->si_bh, 0, + si->si_inode); if (ret == 0) { spin_lock(&osb->osb_lock); ocfs2_update_slot_info(si); @@ -100,20 +111,42 @@ static int ocfs2_update_disk_slots(struct ocfs2_super *osb, struct ocfs2_slot_info *si) { int status, i; - __le16 *disk_info = (__le16 *) si->si_bh->b_data; + __le16 *disk_info = (__le16 *) si->si_bh[0]->b_data; spin_lock(&osb->osb_lock); for (i = 0; i < si->si_size; i++) disk_info[i] = cpu_to_le16(si->si_global_node_nums[i]); spin_unlock(&osb->osb_lock); - status = ocfs2_write_block(osb, si->si_bh, si->si_inode); + status = ocfs2_write_block(osb, si->si_bh[0], si->si_inode); if (status < 0) mlog_errno(status); return status; } +/* + * Calculate how many bytes are needed by the slot map. Returns + * an error if the slot map file is too small. + */ +static int ocfs2_slot_map_physical_size(struct ocfs2_super *osb, + struct inode *inode, + unsigned long long *bytes) +{ + unsigned long long bytes_needed; + + bytes_needed = osb->max_slots * sizeof(__le16); + if (bytes_needed > i_size_read(inode)) { + mlog(ML_ERROR, + "Slot map file is too small! (size %llu, needed %llu)\n", + i_size_read(inode), bytes_needed); + return -ENOSPC; + } + + *bytes = bytes_needed; + return 0; +} + /* try to find global node in the slot info. Returns * OCFS2_INVALID_SLOT if nothing is found. */ static s16 __ocfs2_node_num_to_slot(struct ocfs2_slot_info *si, @@ -188,13 +221,22 @@ int ocfs2_slot_to_node_num_locked(struct ocfs2_super *osb, int slot_num, static void __ocfs2_free_slot_info(struct ocfs2_slot_info *si) { + unsigned int i; + if (si == NULL) return; if (si->si_inode) iput(si->si_inode); - if (si->si_bh) - brelse(si->si_bh); + if (si->si_bh) { + for (i = 0; i < si->si_blocks; i++) { + if (si->si_bh[i]) { + brelse(si->si_bh[i]); + si->si_bh[i] = NULL; + } + } + kfree(si->si_bh); + } kfree(si); } @@ -225,12 +267,65 @@ int ocfs2_clear_slot(struct ocfs2_super *osb, s16 slot_num) return ocfs2_update_disk_slots(osb, osb->slot_info); } +static int ocfs2_map_slot_buffers(struct ocfs2_super *osb, + struct ocfs2_slot_info *si) +{ + int status = 0; + u64 blkno; + unsigned long long blocks, bytes; + unsigned int i; + struct buffer_head *bh; + + status = ocfs2_slot_map_physical_size(osb, si->si_inode, &bytes); + if (status) + goto bail; + + blocks = ocfs2_blocks_for_bytes(si->si_inode->i_sb, bytes); + BUG_ON(blocks > UINT_MAX); + si->si_blocks = blocks; + if (!si->si_blocks) + goto bail; + + mlog(0, "Slot map needs %u buffers for %llu bytes\n", + si->si_blocks, bytes); + + si->si_bh = kzalloc(sizeof(struct buffer_head *) * si->si_blocks, + GFP_KERNEL); + if (!si->si_bh) { + status = -ENOMEM; + mlog_errno(status); + goto bail; + } + + for (i = 0; i < si->si_blocks; i++) { + status = ocfs2_extent_map_get_blocks(si->si_inode, i, + &blkno, NULL, NULL); + if (status < 0) { + mlog_errno(status); + goto bail; + } + + mlog(0, "Reading slot map block %u at %llu\n", i, + (unsigned long long)blkno); + + bh = NULL; /* Acquire a fresh bh */ + status = ocfs2_read_block(osb, blkno, &bh, 0, si->si_inode); + if (status < 0) { + mlog_errno(status); + goto bail; + } + + si->si_bh[i] = bh; + } + +bail: + return status; +} + int ocfs2_init_slot_info(struct ocfs2_super *osb) { int status, i; - u64 blkno; struct inode *inode = NULL; - struct buffer_head *bh = NULL; struct ocfs2_slot_info *si; si = kzalloc(sizeof(struct ocfs2_slot_info), GFP_KERNEL); @@ -254,20 +349,13 @@ int ocfs2_init_slot_info(struct ocfs2_super *osb) goto bail; } - status = ocfs2_extent_map_get_blocks(inode, 0ULL, &blkno, NULL, NULL); - if (status < 0) { - mlog_errno(status); - goto bail; - } - - status = ocfs2_read_block(osb, blkno, &bh, 0, inode); + si->si_inode = inode; + status = ocfs2_map_slot_buffers(osb, si); if (status < 0) { mlog_errno(status); goto bail; } - si->si_inode = inode; - si->si_bh = bh; osb->slot_info = (struct ocfs2_slot_info *)si; bail: if (status < 0 && si) -- cgit v1.2.3 From fc881fa0d59596c02f8707b5572567c369d4789a Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Fri, 1 Feb 2008 12:04:48 -0800 Subject: ocfs2: De-magic the in-memory slot map. The in-memory slot map uses the same magic as the on-disk one. There is a special value to mark a slot as invalid. It relies on the size of certain types and so on. Write a new in-memory map that keeps validity as a separate field. Outside of the I/O functions, OCFS2_INVALID_SLOT now means what it is supposed to. It also is no longer tied to the type size. This also means that only the I/O functions refer to 16bit quantities. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/journal.c | 2 +- fs/ocfs2/ocfs2.h | 6 +-- fs/ocfs2/slot_map.c | 130 +++++++++++++++++++++++++++++----------------------- fs/ocfs2/slot_map.h | 2 +- 4 files changed, 77 insertions(+), 63 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c index ca4c0ea5a4c..bffd2d714f6 100644 --- a/fs/ocfs2/journal.c +++ b/fs/ocfs2/journal.c @@ -71,7 +71,7 @@ static int ocfs2_commit_thread(void *arg); */ struct ocfs2_recovery_map { - int rm_used; + unsigned int rm_used; unsigned int *rm_entries; }; diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h index c6ed8c35de0..95f783dbe8b 100644 --- a/fs/ocfs2/ocfs2.h +++ b/fs/ocfs2/ocfs2.h @@ -216,10 +216,10 @@ struct ocfs2_super unsigned long s_mount_opt; unsigned int s_atime_quantum; - u16 max_slots; + unsigned int max_slots; s16 node_num; - s16 slot_num; - s16 preferred_slot; + int slot_num; + int preferred_slot; int s_sectsize_bits; int s_clustersize; int s_clustersize_bits; diff --git a/fs/ocfs2/slot_map.c b/fs/ocfs2/slot_map.c index 5bddee11009..65a61bfa3f2 100644 --- a/fs/ocfs2/slot_map.c +++ b/fs/ocfs2/slot_map.c @@ -42,21 +42,41 @@ #include "buffer_head_io.h" + +struct ocfs2_slot { + int sl_valid; + unsigned int sl_node_num; +}; + struct ocfs2_slot_info { struct inode *si_inode; unsigned int si_blocks; struct buffer_head **si_bh; unsigned int si_num_slots; - unsigned int si_size; - s16 si_global_node_nums[OCFS2_MAX_SLOTS]; + struct ocfs2_slot *si_slots; }; -static s16 __ocfs2_node_num_to_slot(struct ocfs2_slot_info *si, - s16 global); -static void __ocfs2_fill_slot(struct ocfs2_slot_info *si, - s16 slot_num, - s16 node_num); +static int __ocfs2_node_num_to_slot(struct ocfs2_slot_info *si, + unsigned int node_num); + +static void ocfs2_invalidate_slot(struct ocfs2_slot_info *si, + int slot_num) +{ + BUG_ON((slot_num < 0) || (slot_num >= si->si_num_slots)); + si->si_slots[slot_num].sl_valid = 0; +} + +static void ocfs2_set_slot(struct ocfs2_slot_info *si, + int slot_num, unsigned int node_num) +{ + BUG_ON((slot_num < 0) || (slot_num >= si->si_num_slots)); + BUG_ON((node_num == O2NM_INVALID_NODE_NUM) || + (node_num >= O2NM_MAX_NODES)); + + si->si_slots[slot_num].sl_valid = 1; + si->si_slots[slot_num].sl_node_num = node_num; +} /* * Post the slot information on disk into our slot_info struct. @@ -71,8 +91,12 @@ static void ocfs2_update_slot_info(struct ocfs2_slot_info *si) * should've made sure we have the most recent copy. */ disk_info = (__le16 *) si->si_bh[0]->b_data; - for (i = 0; i < si->si_size; i++) - si->si_global_node_nums[i] = le16_to_cpu(disk_info[i]); + for (i = 0; i < si->si_num_slots; i++) { + if (le16_to_cpu(disk_info[i]) == (u16)OCFS2_INVALID_SLOT) + ocfs2_invalidate_slot(si, i); + else + ocfs2_set_slot(si, i, le16_to_cpu(disk_info[i])); + } } int ocfs2_refresh_slot_info(struct ocfs2_super *osb) @@ -114,8 +138,13 @@ static int ocfs2_update_disk_slots(struct ocfs2_super *osb, __le16 *disk_info = (__le16 *) si->si_bh[0]->b_data; spin_lock(&osb->osb_lock); - for (i = 0; i < si->si_size; i++) - disk_info[i] = cpu_to_le16(si->si_global_node_nums[i]); + for (i = 0; i < si->si_num_slots; i++) { + if (si->si_slots[i].sl_valid) + disk_info[i] = + cpu_to_le16(si->si_slots[i].sl_node_num); + else + disk_info[i] = cpu_to_le16(OCFS2_INVALID_SLOT); + } spin_unlock(&osb->osb_lock); status = ocfs2_write_block(osb, si->si_bh[0], si->si_inode); @@ -147,39 +176,39 @@ static int ocfs2_slot_map_physical_size(struct ocfs2_super *osb, return 0; } -/* try to find global node in the slot info. Returns - * OCFS2_INVALID_SLOT if nothing is found. */ -static s16 __ocfs2_node_num_to_slot(struct ocfs2_slot_info *si, - s16 global) +/* try to find global node in the slot info. Returns -ENOENT + * if nothing is found. */ +static int __ocfs2_node_num_to_slot(struct ocfs2_slot_info *si, + unsigned int node_num) { - int i; - s16 ret = OCFS2_INVALID_SLOT; + int i, ret = -ENOENT; for(i = 0; i < si->si_num_slots; i++) { - if (global == si->si_global_node_nums[i]) { - ret = (s16) i; + if (si->si_slots[i].sl_valid && + (node_num == si->si_slots[i].sl_node_num)) { + ret = i; break; } } + return ret; } -static s16 __ocfs2_find_empty_slot(struct ocfs2_slot_info *si, - s16 preferred) +static int __ocfs2_find_empty_slot(struct ocfs2_slot_info *si, + int preferred) { - int i; - s16 ret = OCFS2_INVALID_SLOT; + int i, ret = -ENOSPC; - if (preferred >= 0 && preferred < si->si_num_slots) { - if (OCFS2_INVALID_SLOT == si->si_global_node_nums[preferred]) { + if ((preferred >= 0) && (preferred < si->si_num_slots)) { + if (!si->si_slots[preferred].sl_valid) { ret = preferred; goto out; } } for(i = 0; i < si->si_num_slots; i++) { - if (OCFS2_INVALID_SLOT == si->si_global_node_nums[i]) { - ret = (s16) i; + if (!si->si_slots[i].sl_valid) { + ret = i; break; } } @@ -189,16 +218,13 @@ out: int ocfs2_node_num_to_slot(struct ocfs2_super *osb, unsigned int node_num) { - s16 slot; + int slot; struct ocfs2_slot_info *si = osb->slot_info; spin_lock(&osb->osb_lock); slot = __ocfs2_node_num_to_slot(si, node_num); spin_unlock(&osb->osb_lock); - if (slot == OCFS2_INVALID_SLOT) - return -ENOENT; - return slot; } @@ -212,10 +238,10 @@ int ocfs2_slot_to_node_num_locked(struct ocfs2_super *osb, int slot_num, BUG_ON(slot_num < 0); BUG_ON(slot_num > osb->max_slots); - if (si->si_global_node_nums[slot_num] == OCFS2_INVALID_SLOT) + if (!si->si_slots[slot_num].sl_valid) return -ENOENT; - *node_num = si->si_global_node_nums[slot_num]; + *node_num = si->si_slots[slot_num].sl_node_num; return 0; } @@ -241,19 +267,7 @@ static void __ocfs2_free_slot_info(struct ocfs2_slot_info *si) kfree(si); } -static void __ocfs2_fill_slot(struct ocfs2_slot_info *si, - s16 slot_num, - s16 node_num) -{ - BUG_ON(slot_num == OCFS2_INVALID_SLOT); - BUG_ON(slot_num >= si->si_num_slots); - BUG_ON((node_num != O2NM_INVALID_NODE_NUM) && - (node_num >= O2NM_MAX_NODES)); - - si->si_global_node_nums[slot_num] = node_num; -} - -int ocfs2_clear_slot(struct ocfs2_super *osb, s16 slot_num) +int ocfs2_clear_slot(struct ocfs2_super *osb, int slot_num) { struct ocfs2_slot_info *si = osb->slot_info; @@ -261,7 +275,7 @@ int ocfs2_clear_slot(struct ocfs2_super *osb, s16 slot_num) return 0; spin_lock(&osb->osb_lock); - __ocfs2_fill_slot(si, slot_num, OCFS2_INVALID_SLOT); + ocfs2_invalidate_slot(si, slot_num); spin_unlock(&osb->osb_lock); return ocfs2_update_disk_slots(osb, osb->slot_info); @@ -324,11 +338,13 @@ bail: int ocfs2_init_slot_info(struct ocfs2_super *osb) { - int status, i; + int status; struct inode *inode = NULL; struct ocfs2_slot_info *si; - si = kzalloc(sizeof(struct ocfs2_slot_info), GFP_KERNEL); + si = kzalloc(sizeof(struct ocfs2_slot_info) + + (sizeof(struct ocfs2_slot) * osb->max_slots), + GFP_KERNEL); if (!si) { status = -ENOMEM; mlog_errno(status); @@ -336,10 +352,8 @@ int ocfs2_init_slot_info(struct ocfs2_super *osb) } si->si_num_slots = osb->max_slots; - si->si_size = OCFS2_MAX_SLOTS; - - for(i = 0; i < si->si_num_slots; i++) - si->si_global_node_nums[i] = OCFS2_INVALID_SLOT; + si->si_slots = (struct ocfs2_slot *)((char *)si + + sizeof(struct ocfs2_slot_info)); inode = ocfs2_get_system_file_inode(osb, SLOT_MAP_SYSTEM_INODE, OCFS2_INVALID_SLOT); @@ -375,7 +389,7 @@ void ocfs2_free_slot_info(struct ocfs2_super *osb) int ocfs2_find_slot(struct ocfs2_super *osb) { int status; - s16 slot; + int slot; struct ocfs2_slot_info *si; mlog_entry_void(); @@ -390,11 +404,11 @@ int ocfs2_find_slot(struct ocfs2_super *osb) * own journal recovery? Possibly not, though we certainly * need to warn to the user */ slot = __ocfs2_node_num_to_slot(si, osb->node_num); - if (slot == OCFS2_INVALID_SLOT) { + if (slot < 0) { /* if no slot yet, then just take 1st available * one. */ slot = __ocfs2_find_empty_slot(si, osb->preferred_slot); - if (slot == OCFS2_INVALID_SLOT) { + if (slot < 0) { spin_unlock(&osb->osb_lock); mlog(ML_ERROR, "no free slots available!\n"); status = -EINVAL; @@ -404,7 +418,7 @@ int ocfs2_find_slot(struct ocfs2_super *osb) mlog(ML_NOTICE, "slot %d is already allocated to this node!\n", slot); - __ocfs2_fill_slot(si, slot, osb->node_num); + ocfs2_set_slot(si, slot, osb->node_num); osb->slot_num = slot; spin_unlock(&osb->osb_lock); @@ -430,7 +444,7 @@ void ocfs2_put_slot(struct ocfs2_super *osb) spin_lock(&osb->osb_lock); ocfs2_update_slot_info(si); - __ocfs2_fill_slot(si, osb->slot_num, OCFS2_INVALID_SLOT); + ocfs2_invalidate_slot(si, osb->slot_num); osb->slot_num = OCFS2_INVALID_SLOT; spin_unlock(&osb->osb_lock); diff --git a/fs/ocfs2/slot_map.h b/fs/ocfs2/slot_map.h index 5118e89c84e..601c95fd700 100644 --- a/fs/ocfs2/slot_map.h +++ b/fs/ocfs2/slot_map.h @@ -39,6 +39,6 @@ int ocfs2_node_num_to_slot(struct ocfs2_super *osb, unsigned int node_num); int ocfs2_slot_to_node_num_locked(struct ocfs2_super *osb, int slot_num, unsigned int *node_num); -int ocfs2_clear_slot(struct ocfs2_super *osb, s16 slot_num); +int ocfs2_clear_slot(struct ocfs2_super *osb, int slot_num); #endif -- cgit v1.2.3 From fb86b1f07120b66769a39c445da5c4300069dd44 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Fri, 1 Feb 2008 11:59:05 -0800 Subject: ocfs2: Define the contents of the slot_map file. The slot map file is merely an array of __le16. Wrap it in a structure for cleaner reference. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/ocfs2_fs.h | 12 ++++++++++++ fs/ocfs2/slot_map.c | 15 ++++++++------- 2 files changed, 20 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/ocfs2_fs.h b/fs/ocfs2/ocfs2_fs.h index 3633edd3982..3299116b802 100644 --- a/fs/ocfs2/ocfs2_fs.h +++ b/fs/ocfs2/ocfs2_fs.h @@ -474,6 +474,18 @@ struct ocfs2_extent_block /* Actual on-disk size is one block */ }; +/* + * On disk slot map for OCFS2. This defines the contents of the "slot_map" + * system file. + */ +struct ocfs2_slot_map { +/*00*/ __le16 sm_slots[0]; +/* + * Actual on-disk size is one block. OCFS2_MAX_SLOTS is 255, + * 255 * sizeof(__le16) == 512B, within the 512B block minimum blocksize. + */ +}; + /* * On disk superblock for OCFS2 * Note that it is contained inside an ocfs2_dinode, so all offsets diff --git a/fs/ocfs2/slot_map.c b/fs/ocfs2/slot_map.c index 65a61bfa3f2..e7e7a74156b 100644 --- a/fs/ocfs2/slot_map.c +++ b/fs/ocfs2/slot_map.c @@ -85,17 +85,17 @@ static void ocfs2_set_slot(struct ocfs2_slot_info *si, static void ocfs2_update_slot_info(struct ocfs2_slot_info *si) { int i; - __le16 *disk_info; + struct ocfs2_slot_map *sm; /* we don't read the slot block here as ocfs2_super_lock * should've made sure we have the most recent copy. */ - disk_info = (__le16 *) si->si_bh[0]->b_data; + sm = (struct ocfs2_slot_map *)si->si_bh[0]->b_data; for (i = 0; i < si->si_num_slots; i++) { - if (le16_to_cpu(disk_info[i]) == (u16)OCFS2_INVALID_SLOT) + if (le16_to_cpu(sm->sm_slots[i]) == (u16)OCFS2_INVALID_SLOT) ocfs2_invalidate_slot(si, i); else - ocfs2_set_slot(si, i, le16_to_cpu(disk_info[i])); + ocfs2_set_slot(si, i, le16_to_cpu(sm->sm_slots[i])); } } @@ -135,15 +135,16 @@ static int ocfs2_update_disk_slots(struct ocfs2_super *osb, struct ocfs2_slot_info *si) { int status, i; - __le16 *disk_info = (__le16 *) si->si_bh[0]->b_data; + struct ocfs2_slot_map *sm; spin_lock(&osb->osb_lock); + sm = (struct ocfs2_slot_map *)si->si_bh[0]->b_data; for (i = 0; i < si->si_num_slots; i++) { if (si->si_slots[i].sl_valid) - disk_info[i] = + sm->sm_slots[i] = cpu_to_le16(si->si_slots[i].sl_node_num); else - disk_info[i] = cpu_to_le16(OCFS2_INVALID_SLOT); + sm->sm_slots[i] = cpu_to_le16(OCFS2_INVALID_SLOT); } spin_unlock(&osb->osb_lock); -- cgit v1.2.3 From 386a2ef8576e966076c293f6496b9e3d7e3d9035 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Fri, 1 Feb 2008 12:06:54 -0800 Subject: ocfs2: New slot map format The old slot map had a few limitations: - It was limited to one block, so the maximum slot count was 255. - Each slot was signed 16bits, limiting node numbers to INT16_MAX. - An empty slot was marked by the magic 0xFFFF (-1). The new slot map format provides 32bit node numbers (UINT32_MAX), a separate space to mark a slot in use, and extra room to grow. The slot map is now bounded by i_size, not a block. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/ocfs2.h | 7 ++++ fs/ocfs2/ocfs2_fs.h | 31 ++++++++++++++- fs/ocfs2/slot_map.c | 110 +++++++++++++++++++++++++++++++++++++++++++++------- 3 files changed, 133 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h index 95f783dbe8b..f78e9ed5324 100644 --- a/fs/ocfs2/ocfs2.h +++ b/fs/ocfs2/ocfs2.h @@ -374,6 +374,13 @@ static inline int ocfs2_mount_local(struct ocfs2_super *osb) return (osb->s_feature_incompat & OCFS2_FEATURE_INCOMPAT_LOCAL_MOUNT); } +static inline int ocfs2_uses_extended_slot_map(struct ocfs2_super *osb) +{ + return (osb->s_feature_incompat & + OCFS2_FEATURE_INCOMPAT_EXTENDED_SLOT_MAP); +} + + #define OCFS2_IS_VALID_DINODE(ptr) \ (!strcmp((ptr)->i_signature, OCFS2_INODE_SIGNATURE)) diff --git a/fs/ocfs2/ocfs2_fs.h b/fs/ocfs2/ocfs2_fs.h index 3299116b802..c49502329ab 100644 --- a/fs/ocfs2/ocfs2_fs.h +++ b/fs/ocfs2/ocfs2_fs.h @@ -88,7 +88,8 @@ #define OCFS2_FEATURE_COMPAT_SUPP OCFS2_FEATURE_COMPAT_BACKUP_SB #define OCFS2_FEATURE_INCOMPAT_SUPP (OCFS2_FEATURE_INCOMPAT_LOCAL_MOUNT \ | OCFS2_FEATURE_INCOMPAT_SPARSE_ALLOC \ - | OCFS2_FEATURE_INCOMPAT_INLINE_DATA) + | OCFS2_FEATURE_INCOMPAT_INLINE_DATA \ + | OCFS2_FEATURE_INCOMPAT_EXTENDED_SLOT_MAP) #define OCFS2_FEATURE_RO_COMPAT_SUPP OCFS2_FEATURE_RO_COMPAT_UNWRITTEN /* @@ -125,6 +126,10 @@ /* Support for data packed into inode blocks */ #define OCFS2_FEATURE_INCOMPAT_INLINE_DATA 0x0040 +/* Support for the extended slot map */ +#define OCFS2_FEATURE_INCOMPAT_EXTENDED_SLOT_MAP 0x100 + + /* * backup superblock flag is used to indicate that this volume * has backup superblocks. @@ -476,7 +481,8 @@ struct ocfs2_extent_block /* * On disk slot map for OCFS2. This defines the contents of the "slot_map" - * system file. + * system file. A slot is valid if it contains a node number >= 0. The + * value -1 (0xFFFF) is OCFS2_INVALID_SLOT. This marks a slot empty. */ struct ocfs2_slot_map { /*00*/ __le16 sm_slots[0]; @@ -486,6 +492,27 @@ struct ocfs2_slot_map { */ }; +struct ocfs2_extended_slot { +/*00*/ __u8 es_valid; + __u8 es_reserved1[3]; + __le32 es_node_num; +/*10*/ +}; + +/* + * The extended slot map, used when OCFS2_FEATURE_INCOMPAT_EXTENDED_SLOT_MAP + * is set. It separates out the valid marker from the node number, and + * has room to grow. Unlike the old slot map, this format is defined by + * i_size. + */ +struct ocfs2_slot_map_extended { +/*00*/ struct ocfs2_extended_slot se_slots[0]; +/* + * Actual size is i_size of the slot_map system file. It should + * match s_max_slots * sizeof(struct ocfs2_extended_slot) + */ +}; + /* * On disk superblock for OCFS2 * Note that it is contained inside an ocfs2_dinode, so all offsets diff --git a/fs/ocfs2/slot_map.c b/fs/ocfs2/slot_map.c index e7e7a74156b..63fb1b26d96 100644 --- a/fs/ocfs2/slot_map.c +++ b/fs/ocfs2/slot_map.c @@ -49,6 +49,8 @@ struct ocfs2_slot { }; struct ocfs2_slot_info { + int si_extended; + int si_slots_per_block; struct inode *si_inode; unsigned int si_blocks; struct buffer_head **si_bh; @@ -78,17 +80,37 @@ static void ocfs2_set_slot(struct ocfs2_slot_info *si, si->si_slots[slot_num].sl_node_num = node_num; } +/* This version is for the extended slot map */ +static void ocfs2_update_slot_info_extended(struct ocfs2_slot_info *si) +{ + int b, i, slotno; + struct ocfs2_slot_map_extended *se; + + slotno = 0; + for (b = 0; b < si->si_blocks; b++) { + se = (struct ocfs2_slot_map_extended *)si->si_bh[b]->b_data; + for (i = 0; + (i < si->si_slots_per_block) && + (slotno < si->si_num_slots); + i++, slotno++) { + if (se->se_slots[i].es_valid) + ocfs2_set_slot(si, slotno, + le32_to_cpu(se->se_slots[i].es_node_num)); + else + ocfs2_invalidate_slot(si, slotno); + } + } +} + /* * Post the slot information on disk into our slot_info struct. * Must be protected by osb_lock. */ -static void ocfs2_update_slot_info(struct ocfs2_slot_info *si) +static void ocfs2_update_slot_info_old(struct ocfs2_slot_info *si) { int i; struct ocfs2_slot_map *sm; - /* we don't read the slot block here as ocfs2_super_lock - * should've made sure we have the most recent copy. */ sm = (struct ocfs2_slot_map *)si->si_bh[0]->b_data; for (i = 0; i < si->si_num_slots; i++) { @@ -99,6 +121,18 @@ static void ocfs2_update_slot_info(struct ocfs2_slot_info *si) } } +static void ocfs2_update_slot_info(struct ocfs2_slot_info *si) +{ + /* + * The slot data will have been refreshed when ocfs2_super_lock + * was taken. + */ + if (si->si_extended) + ocfs2_update_slot_info_extended(si); + else + ocfs2_update_slot_info_old(si); +} + int ocfs2_refresh_slot_info(struct ocfs2_super *osb) { int ret; @@ -131,13 +165,31 @@ int ocfs2_refresh_slot_info(struct ocfs2_super *osb) /* post the our slot info stuff into it's destination bh and write it * out. */ -static int ocfs2_update_disk_slots(struct ocfs2_super *osb, - struct ocfs2_slot_info *si) +static void ocfs2_update_disk_slot_extended(struct ocfs2_slot_info *si, + int slot_num, + struct buffer_head **bh) +{ + int blkind = slot_num / si->si_slots_per_block; + int slotno = slot_num % si->si_slots_per_block; + struct ocfs2_slot_map_extended *se; + + BUG_ON(blkind >= si->si_blocks); + + se = (struct ocfs2_slot_map_extended *)si->si_bh[blkind]->b_data; + se->se_slots[slotno].es_valid = si->si_slots[slot_num].sl_valid; + if (si->si_slots[slot_num].sl_valid) + se->se_slots[slotno].es_node_num = + cpu_to_le32(si->si_slots[slot_num].sl_node_num); + *bh = si->si_bh[blkind]; +} + +static void ocfs2_update_disk_slot_old(struct ocfs2_slot_info *si, + int slot_num, + struct buffer_head **bh) { - int status, i; + int i; struct ocfs2_slot_map *sm; - spin_lock(&osb->osb_lock); sm = (struct ocfs2_slot_map *)si->si_bh[0]->b_data; for (i = 0; i < si->si_num_slots; i++) { if (si->si_slots[i].sl_valid) @@ -146,9 +198,24 @@ static int ocfs2_update_disk_slots(struct ocfs2_super *osb, else sm->sm_slots[i] = cpu_to_le16(OCFS2_INVALID_SLOT); } + *bh = si->si_bh[0]; +} + +static int ocfs2_update_disk_slot(struct ocfs2_super *osb, + struct ocfs2_slot_info *si, + int slot_num) +{ + int status; + struct buffer_head *bh; + + spin_lock(&osb->osb_lock); + if (si->si_extended) + ocfs2_update_disk_slot_extended(si, slot_num, &bh); + else + ocfs2_update_disk_slot_old(si, slot_num, &bh); spin_unlock(&osb->osb_lock); - status = ocfs2_write_block(osb, si->si_bh[0], si->si_inode); + status = ocfs2_write_block(osb, bh, si->si_inode); if (status < 0) mlog_errno(status); @@ -165,7 +232,12 @@ static int ocfs2_slot_map_physical_size(struct ocfs2_super *osb, { unsigned long long bytes_needed; - bytes_needed = osb->max_slots * sizeof(__le16); + if (ocfs2_uses_extended_slot_map(osb)) { + bytes_needed = osb->max_slots * + sizeof(struct ocfs2_extended_slot); + } else { + bytes_needed = osb->max_slots * sizeof(__le16); + } if (bytes_needed > i_size_read(inode)) { mlog(ML_ERROR, "Slot map file is too small! (size %llu, needed %llu)\n", @@ -279,7 +351,7 @@ int ocfs2_clear_slot(struct ocfs2_super *osb, int slot_num) ocfs2_invalidate_slot(si, slot_num); spin_unlock(&osb->osb_lock); - return ocfs2_update_disk_slots(osb, osb->slot_info); + return ocfs2_update_disk_slot(osb, osb->slot_info, slot_num); } static int ocfs2_map_slot_buffers(struct ocfs2_super *osb, @@ -301,6 +373,16 @@ static int ocfs2_map_slot_buffers(struct ocfs2_super *osb, if (!si->si_blocks) goto bail; + if (si->si_extended) + si->si_slots_per_block = + (osb->sb->s_blocksize / + sizeof(struct ocfs2_extended_slot)); + else + si->si_slots_per_block = osb->sb->s_blocksize / sizeof(__le16); + + /* The size checks above should ensure this */ + BUG_ON((osb->max_slots / si->si_slots_per_block) > blocks); + mlog(0, "Slot map needs %u buffers for %llu bytes\n", si->si_blocks, bytes); @@ -352,6 +434,7 @@ int ocfs2_init_slot_info(struct ocfs2_super *osb) goto bail; } + si->si_extended = ocfs2_uses_extended_slot_map(osb); si->si_num_slots = osb->max_slots; si->si_slots = (struct ocfs2_slot *)((char *)si + sizeof(struct ocfs2_slot_info)); @@ -425,7 +508,7 @@ int ocfs2_find_slot(struct ocfs2_super *osb) mlog(0, "taking node slot %d\n", osb->slot_num); - status = ocfs2_update_disk_slots(osb, si); + status = ocfs2_update_disk_slot(osb, si, osb->slot_num); if (status < 0) mlog_errno(status); @@ -436,7 +519,7 @@ bail: void ocfs2_put_slot(struct ocfs2_super *osb) { - int status; + int status, slot_num; struct ocfs2_slot_info *si = osb->slot_info; if (!si) @@ -445,11 +528,12 @@ void ocfs2_put_slot(struct ocfs2_super *osb) spin_lock(&osb->osb_lock); ocfs2_update_slot_info(si); + slot_num = osb->slot_num; ocfs2_invalidate_slot(si, osb->slot_num); osb->slot_num = OCFS2_INVALID_SLOT; spin_unlock(&osb->osb_lock); - status = ocfs2_update_disk_slots(osb, si); + status = ocfs2_update_disk_slot(osb, si, slot_num); if (status < 0) { mlog_errno(status); goto bail; -- cgit v1.2.3 From 24ef1815e5e13e50196eb1ab8ddc0d783443bdf8 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Tue, 29 Jan 2008 17:37:32 -0800 Subject: ocfs2: Separate out dlm lock functions. This is the first in a series of patches to isolate ocfs2 from the underlying cluster stack. Here we wrap the dlm locking functions with ocfs2-specific calls. Because ocfs2 always uses the same dlm lock status callbacks, we can eliminate the callbacks from the filesystem visible functions. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/Makefile | 1 + fs/ocfs2/dlmglue.c | 110 ++++++++++++++++++++++++++++----------------------- fs/ocfs2/dlmglue.h | 3 ++ fs/ocfs2/stackglue.c | 65 ++++++++++++++++++++++++++++++ fs/ocfs2/stackglue.h | 45 +++++++++++++++++++++ fs/ocfs2/super.c | 4 ++ 6 files changed, 179 insertions(+), 49 deletions(-) create mode 100644 fs/ocfs2/stackglue.c create mode 100644 fs/ocfs2/stackglue.h (limited to 'fs') diff --git a/fs/ocfs2/Makefile b/fs/ocfs2/Makefile index 4d4ce48bb42..3ba64af2695 100644 --- a/fs/ocfs2/Makefile +++ b/fs/ocfs2/Makefile @@ -24,6 +24,7 @@ ocfs2-objs := \ namei.o \ resize.o \ slot_map.o \ + stackglue.o \ suballoc.o \ super.o \ symlink.o \ diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index 15a5167e051..aea3bef1917 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -53,6 +53,7 @@ #include "heartbeat.h" #include "inode.h" #include "journal.h" +#include "stackglue.h" #include "slot_map.h" #include "super.h" #include "uptodate.h" @@ -888,22 +889,21 @@ static int ocfs2_lock_create(struct ocfs2_super *osb, lockres_or_flags(lockres, OCFS2_LOCK_BUSY); spin_unlock_irqrestore(&lockres->l_lock, flags); - status = dlmlock(osb->dlm, - level, - &lockres->l_lksb, - dlm_flags, - lockres->l_name, - OCFS2_LOCK_ID_MAX_LEN - 1, - ocfs2_locking_ast, - lockres, - ocfs2_blocking_ast); + status = ocfs2_dlm_lock(osb->dlm, + level, + &lockres->l_lksb, + dlm_flags, + lockres->l_name, + OCFS2_LOCK_ID_MAX_LEN - 1, + lockres); if (status != DLM_NORMAL) { - ocfs2_log_dlm_error("dlmlock", status, lockres); + ocfs2_log_dlm_error("ocfs2_dlm_lock", status, lockres); ret = -EINVAL; ocfs2_recover_from_dlm_error(lockres, 1); } - mlog(0, "lock %s, successfull return from dlmlock\n", lockres->l_name); + mlog(0, "lock %s, successfull return from ocfs2_dlm_lock\n", + lockres->l_name); bail: mlog_exit(ret); @@ -1091,29 +1091,27 @@ again: lockres->l_name, lockres->l_level, level); /* call dlm_lock to upgrade lock now */ - status = dlmlock(osb->dlm, - level, - &lockres->l_lksb, - lkm_flags, - lockres->l_name, - OCFS2_LOCK_ID_MAX_LEN - 1, - ocfs2_locking_ast, - lockres, - ocfs2_blocking_ast); + status = ocfs2_dlm_lock(osb->dlm, + level, + &lockres->l_lksb, + lkm_flags, + lockres->l_name, + OCFS2_LOCK_ID_MAX_LEN - 1, + lockres); if (status != DLM_NORMAL) { if ((lkm_flags & LKM_NOQUEUE) && (status == DLM_NOTQUEUED)) ret = -EAGAIN; else { - ocfs2_log_dlm_error("dlmlock", status, - lockres); + ocfs2_log_dlm_error("ocfs2_dlm_lock", + status, lockres); ret = -EINVAL; } ocfs2_recover_from_dlm_error(lockres, 1); goto out; } - mlog(0, "lock %s, successfull return from dlmlock\n", + mlog(0, "lock %s, successfull return from ocfs2_dlm_lock\n", lockres->l_name); /* At this point we've gone inside the dlm and need to @@ -1503,14 +1501,14 @@ int ocfs2_file_lock(struct file *file, int ex, int trylock) lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_BUSY, 0); spin_unlock_irqrestore(&lockres->l_lock, flags); - ret = dlmlock(osb->dlm, level, &lockres->l_lksb, lkm_flags, - lockres->l_name, OCFS2_LOCK_ID_MAX_LEN - 1, - ocfs2_locking_ast, lockres, ocfs2_blocking_ast); + ret = ocfs2_dlm_lock(osb->dlm, level, &lockres->l_lksb, lkm_flags, + lockres->l_name, OCFS2_LOCK_ID_MAX_LEN - 1, + lockres); if (ret != DLM_NORMAL) { if (trylock && ret == DLM_NOTQUEUED) ret = -EAGAIN; else { - ocfs2_log_dlm_error("dlmlock", ret, lockres); + ocfs2_log_dlm_error("ocfs2_dlm_lock", ret, lockres); ret = -EINVAL; } @@ -2699,15 +2697,15 @@ static int ocfs2_drop_lock(struct ocfs2_super *osb, mlog(0, "lock %s\n", lockres->l_name); - status = dlmunlock(osb->dlm, &lockres->l_lksb, lkm_flags, - ocfs2_unlock_ast, lockres); + status = ocfs2_dlm_unlock(osb->dlm, &lockres->l_lksb, lkm_flags, + lockres); if (status != DLM_NORMAL) { - ocfs2_log_dlm_error("dlmunlock", status, lockres); + ocfs2_log_dlm_error("ocfs2_dlm_unlock", status, lockres); mlog(ML_ERROR, "lockres flags: %lu\n", lockres->l_flags); dlm_print_one_lock(lockres->l_lksb.lockid); BUG(); } - mlog(0, "lock %s, successfull return from dlmunlock\n", + mlog(0, "lock %s, successfull return from ocfs2_dlm_unlock\n", lockres->l_name); ocfs2_wait_on_busy_lock(lockres); @@ -2832,17 +2830,15 @@ static int ocfs2_downconvert_lock(struct ocfs2_super *osb, if (lvb) dlm_flags |= LKM_VALBLK; - status = dlmlock(osb->dlm, - new_level, - &lockres->l_lksb, - dlm_flags, - lockres->l_name, - OCFS2_LOCK_ID_MAX_LEN - 1, - ocfs2_locking_ast, - lockres, - ocfs2_blocking_ast); + status = ocfs2_dlm_lock(osb->dlm, + new_level, + &lockres->l_lksb, + dlm_flags, + lockres->l_name, + OCFS2_LOCK_ID_MAX_LEN - 1, + lockres); if (status != DLM_NORMAL) { - ocfs2_log_dlm_error("dlmlock", status, lockres); + ocfs2_log_dlm_error("ocfs2_dlm_lock", status, lockres); ret = -EINVAL; ocfs2_recover_from_dlm_error(lockres, 1); goto bail; @@ -2854,7 +2850,7 @@ bail: return ret; } -/* returns 1 when the caller should unlock and call dlmunlock */ +/* returns 1 when the caller should unlock and call ocfs2_dlm_unlock */ static int ocfs2_prepare_cancel_convert(struct ocfs2_super *osb, struct ocfs2_lock_res *lockres) { @@ -2896,18 +2892,17 @@ static int ocfs2_cancel_convert(struct ocfs2_super *osb, mlog(0, "lock %s\n", lockres->l_name); ret = 0; - status = dlmunlock(osb->dlm, - &lockres->l_lksb, - LKM_CANCEL, - ocfs2_unlock_ast, - lockres); + status = ocfs2_dlm_unlock(osb->dlm, + &lockres->l_lksb, + LKM_CANCEL, + lockres); if (status != DLM_NORMAL) { - ocfs2_log_dlm_error("dlmunlock", status, lockres); + ocfs2_log_dlm_error("ocfs2_dlm_unlock", status, lockres); ret = -EINVAL; ocfs2_recover_from_dlm_error(lockres, 0); } - mlog(0, "lock %s return from dlmunlock\n", lockres->l_name); + mlog(0, "lock %s return from ocfs2_dlm_unlock\n", lockres->l_name); mlog_exit(ret); return ret; @@ -3211,6 +3206,23 @@ static int ocfs2_dentry_convert_worker(struct ocfs2_lock_res *lockres, return UNBLOCK_CONTINUE_POST; } +static struct ocfs2_locking_protocol lproto = { + .lp_lock_ast = ocfs2_locking_ast, + .lp_blocking_ast = ocfs2_blocking_ast, + .lp_unlock_ast = ocfs2_unlock_ast, +}; + +/* This interface isn't the final one, hence the less-than-perfect names */ +void dlmglue_init_stack(void) +{ + o2cb_get_stack(&lproto); +} + +void dlmglue_exit_stack(void) +{ + o2cb_put_stack(); +} + static void ocfs2_process_blocked_lock(struct ocfs2_super *osb, struct ocfs2_lock_res *lockres) { diff --git a/fs/ocfs2/dlmglue.h b/fs/ocfs2/dlmglue.h index e3cf902404b..32380439401 100644 --- a/fs/ocfs2/dlmglue.h +++ b/fs/ocfs2/dlmglue.h @@ -114,5 +114,8 @@ void ocfs2_wake_downconvert_thread(struct ocfs2_super *osb); struct ocfs2_dlm_debug *ocfs2_new_dlm_debug(void); void ocfs2_put_dlm_debug(struct ocfs2_dlm_debug *dlm_debug); +void dlmglue_init_stack(void); +void dlmglue_exit_stack(void); + extern const struct dlm_protocol_version ocfs2_locking_protocol; #endif /* DLMGLUE_H */ diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c new file mode 100644 index 00000000000..4f44f23795f --- /dev/null +++ b/fs/ocfs2/stackglue.c @@ -0,0 +1,65 @@ +/* -*- mode: c; c-basic-offset: 8; -*- + * vim: noexpandtab sw=8 ts=8 sts=0: + * + * stackglue.c + * + * Code which implements an OCFS2 specific interface to underlying + * cluster stacks. + * + * Copyright (C) 2007 Oracle. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License as published by the Free Software Foundation, version 2. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + */ + +#include +#include + +#include "dlm/dlmapi.h" + +#include "stackglue.h" + +static struct ocfs2_locking_protocol *lproto; + +enum dlm_status ocfs2_dlm_lock(struct dlm_ctxt *dlm, + int mode, + struct dlm_lockstatus *lksb, + u32 flags, + void *name, + unsigned int namelen, + void *astarg) +{ + BUG_ON(lproto == NULL); + return dlmlock(dlm, mode, lksb, flags, name, namelen, + lproto->lp_lock_ast, astarg, + lproto->lp_blocking_ast); +} + +enum dlm_status ocfs2_dlm_unlock(struct dlm_ctxt *dlm, + struct dlm_lockstatus *lksb, + u32 flags, + void *astarg) +{ + BUG_ON(lproto == NULL); + + return dlmunlock(dlm, lksb, flags, lproto->lp_unlock_ast, astarg); +} + + +void o2cb_get_stack(struct ocfs2_locking_protocol *proto) +{ + BUG_ON(proto == NULL); + + lproto = proto; +} + +void o2cb_put_stack(void) +{ + lproto = NULL; +} diff --git a/fs/ocfs2/stackglue.h b/fs/ocfs2/stackglue.h new file mode 100644 index 00000000000..40a00241340 --- /dev/null +++ b/fs/ocfs2/stackglue.h @@ -0,0 +1,45 @@ +/* -*- mode: c; c-basic-offset: 8; -*- + * vim: noexpandtab sw=8 ts=8 sts=0: + * + * stackglue.h + * + * Glue to the underlying cluster stack. + * + * Copyright (C) 2007 Oracle. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License as published by the Free Software Foundation, version 2. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + */ + + +#ifndef STACKGLUE_H +#define STACKGLUE_H + +struct ocfs2_locking_protocol { + void (*lp_lock_ast)(void *astarg); + void (*lp_blocking_ast)(void *astarg, int level); + void (*lp_unlock_ast)(void *astarg, enum dlm_status status); +}; + +enum dlm_status ocfs2_dlm_lock(struct dlm_ctxt *dlm, + int mode, + struct dlm_lockstatus *lksb, + u32 flags, + void *name, + unsigned int namelen, + void *astarg); +enum dlm_status ocfs2_dlm_unlock(struct dlm_ctxt *dlm, + struct dlm_lockstatus *lksb, + u32 flags, + void *astarg); + +void o2cb_get_stack(struct ocfs2_locking_protocol *proto); +void o2cb_put_stack(void); + +#endif /* STACKGLUE_H */ diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index 1a4c7c7850f..c8675464e29 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -933,6 +933,8 @@ static int __init ocfs2_init(void) ocfs2_print_version(); + dlmglue_init_stack(); + status = init_ocfs2_uptodate_cache(); if (status < 0) { mlog_errno(status); @@ -988,6 +990,8 @@ static void __exit ocfs2_exit(void) exit_ocfs2_uptodate_cache(); + dlmglue_exit_stack(); + mlog_exit_void(); } -- cgit v1.2.3 From bd3e76105d4478ab89951a52d1a35250d24a9f16 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Fri, 1 Feb 2008 12:14:57 -0800 Subject: ocfs2: Use global DLM_ constants in generic code. The ocfs2 generic code should use the values in . stackglue.c will convert them to o2dlm values. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlmglue.c | 140 +++++++++++++++++++++++++-------------------------- fs/ocfs2/stackglue.c | 71 +++++++++++++++++++++++--- fs/ocfs2/stackglue.h | 13 +++++ 3 files changed, 147 insertions(+), 77 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index aea3bef1917..b8ac903ef8e 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -37,8 +37,6 @@ #include #include -#include - #define MLOG_MASK_PREFIX ML_DLM_GLUE #include @@ -317,7 +315,7 @@ static inline struct ocfs2_super *ocfs2_get_lockres_osb(struct ocfs2_lock_res *l static int ocfs2_lock_create(struct ocfs2_super *osb, struct ocfs2_lock_res *lockres, int level, - int dlm_flags); + u32 dlm_flags); static inline int ocfs2_may_continue_on_blocked_lock(struct ocfs2_lock_res *lockres, int wanted); static void ocfs2_cluster_unlock(struct ocfs2_super *osb, @@ -407,9 +405,9 @@ static void ocfs2_lock_res_init_common(struct ocfs2_super *osb, res->l_ops = ops; res->l_priv = priv; - res->l_level = LKM_IVMODE; - res->l_requested = LKM_IVMODE; - res->l_blocking = LKM_IVMODE; + res->l_level = DLM_LOCK_IV; + res->l_requested = DLM_LOCK_IV; + res->l_blocking = DLM_LOCK_IV; res->l_action = OCFS2_AST_INVALID; res->l_unlock_action = OCFS2_UNLOCK_INVALID; @@ -605,10 +603,10 @@ static inline void ocfs2_inc_holders(struct ocfs2_lock_res *lockres, BUG_ON(!lockres); switch(level) { - case LKM_EXMODE: + case DLM_LOCK_EX: lockres->l_ex_holders++; break; - case LKM_PRMODE: + case DLM_LOCK_PR: lockres->l_ro_holders++; break; default: @@ -626,11 +624,11 @@ static inline void ocfs2_dec_holders(struct ocfs2_lock_res *lockres, BUG_ON(!lockres); switch(level) { - case LKM_EXMODE: + case DLM_LOCK_EX: BUG_ON(!lockres->l_ex_holders); lockres->l_ex_holders--; break; - case LKM_PRMODE: + case DLM_LOCK_PR: BUG_ON(!lockres->l_ro_holders); lockres->l_ro_holders--; break; @@ -645,12 +643,12 @@ static inline void ocfs2_dec_holders(struct ocfs2_lock_res *lockres, * lock types are added. */ static inline int ocfs2_highest_compat_lock_level(int level) { - int new_level = LKM_EXMODE; + int new_level = DLM_LOCK_EX; - if (level == LKM_EXMODE) - new_level = LKM_NLMODE; - else if (level == LKM_PRMODE) - new_level = LKM_PRMODE; + if (level == DLM_LOCK_EX) + new_level = DLM_LOCK_NL; + else if (level == DLM_LOCK_PR) + new_level = DLM_LOCK_PR; return new_level; } @@ -689,12 +687,12 @@ static inline void ocfs2_generic_handle_downconvert_action(struct ocfs2_lock_res BUG_ON(!(lockres->l_flags & OCFS2_LOCK_BUSY)); BUG_ON(!(lockres->l_flags & OCFS2_LOCK_ATTACHED)); BUG_ON(!(lockres->l_flags & OCFS2_LOCK_BLOCKED)); - BUG_ON(lockres->l_blocking <= LKM_NLMODE); + BUG_ON(lockres->l_blocking <= DLM_LOCK_NL); lockres->l_level = lockres->l_requested; if (lockres->l_level <= ocfs2_highest_compat_lock_level(lockres->l_blocking)) { - lockres->l_blocking = LKM_NLMODE; + lockres->l_blocking = DLM_LOCK_NL; lockres_clear_flags(lockres, OCFS2_LOCK_BLOCKED); } lockres_clear_flags(lockres, OCFS2_LOCK_BUSY); @@ -713,7 +711,7 @@ static inline void ocfs2_generic_handle_convert_action(struct ocfs2_lock_res *lo * information is already up to data. Convert from NL to * *anything* however should mark ourselves as needing an * update */ - if (lockres->l_level == LKM_NLMODE && + if (lockres->l_level == DLM_LOCK_NL && lockres->l_ops->flags & LOCK_TYPE_REQUIRES_REFRESH) lockres_or_flags(lockres, OCFS2_LOCK_NEEDS_REFRESH); @@ -730,7 +728,7 @@ static inline void ocfs2_generic_handle_attach_action(struct ocfs2_lock_res *loc BUG_ON((!(lockres->l_flags & OCFS2_LOCK_BUSY))); BUG_ON(lockres->l_flags & OCFS2_LOCK_ATTACHED); - if (lockres->l_requested > LKM_NLMODE && + if (lockres->l_requested > DLM_LOCK_NL && !(lockres->l_flags & OCFS2_LOCK_LOCAL) && lockres->l_ops->flags & LOCK_TYPE_REQUIRES_REFRESH) lockres_or_flags(lockres, OCFS2_LOCK_NEEDS_REFRESH); @@ -775,7 +773,7 @@ static void ocfs2_blocking_ast(void *opaque, int level) int needs_downconvert; unsigned long flags; - BUG_ON(level <= LKM_NLMODE); + BUG_ON(level <= DLM_LOCK_NL); mlog(0, "BAST fired for lockres %s, blocking %d, level %d type %s\n", lockres->l_name, level, lockres->l_level, @@ -866,7 +864,7 @@ static inline void ocfs2_recover_from_dlm_error(struct ocfs2_lock_res *lockres, static int ocfs2_lock_create(struct ocfs2_super *osb, struct ocfs2_lock_res *lockres, int level, - int dlm_flags) + u32 dlm_flags) { int ret = 0; enum dlm_status status = DLM_NORMAL; @@ -874,7 +872,7 @@ static int ocfs2_lock_create(struct ocfs2_super *osb, mlog_entry_void(); - mlog(0, "lock %s, level = %d, flags = %d\n", lockres->l_name, level, + mlog(0, "lock %s, level = %d, flags = %u\n", lockres->l_name, level, dlm_flags); spin_lock_irqsave(&lockres->l_lock, flags); @@ -1016,7 +1014,7 @@ static int ocfs2_wait_for_mask_interruptible(struct ocfs2_mask_waiter *mw, static int ocfs2_cluster_lock(struct ocfs2_super *osb, struct ocfs2_lock_res *lockres, int level, - int lkm_flags, + u32 lkm_flags, int arg_flags) { struct ocfs2_mask_waiter mw; @@ -1030,7 +1028,7 @@ static int ocfs2_cluster_lock(struct ocfs2_super *osb, ocfs2_init_mask_waiter(&mw); if (lockres->l_ops->flags & LOCK_TYPE_USES_LVB) - lkm_flags |= LKM_VALBLK; + lkm_flags |= DLM_LKF_VALBLK; again: wait = 0; @@ -1074,18 +1072,18 @@ again: if (!(lockres->l_flags & OCFS2_LOCK_ATTACHED)) { lockres->l_action = OCFS2_AST_ATTACH; - lkm_flags &= ~LKM_CONVERT; + lkm_flags &= ~DLM_LKF_CONVERT; } else { lockres->l_action = OCFS2_AST_CONVERT; - lkm_flags |= LKM_CONVERT; + lkm_flags |= DLM_LKF_CONVERT; } lockres->l_requested = level; lockres_or_flags(lockres, OCFS2_LOCK_BUSY); spin_unlock_irqrestore(&lockres->l_lock, flags); - BUG_ON(level == LKM_IVMODE); - BUG_ON(level == LKM_NLMODE); + BUG_ON(level == DLM_LOCK_IV); + BUG_ON(level == DLM_LOCK_NL); mlog(0, "lock %s, convert from %d to level = %d\n", lockres->l_name, lockres->l_level, level); @@ -1099,7 +1097,7 @@ again: OCFS2_LOCK_ID_MAX_LEN - 1, lockres); if (status != DLM_NORMAL) { - if ((lkm_flags & LKM_NOQUEUE) && + if ((lkm_flags & DLM_LKF_NOQUEUE) && (status == DLM_NOTQUEUED)) ret = -EAGAIN; else { @@ -1175,9 +1173,9 @@ static int ocfs2_create_new_lock(struct ocfs2_super *osb, int ex, int local) { - int level = ex ? LKM_EXMODE : LKM_PRMODE; + int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR; unsigned long flags; - int lkm_flags = local ? LKM_LOCAL : 0; + u32 lkm_flags = local ? DLM_LKF_LOCAL : 0; spin_lock_irqsave(&lockres->l_lock, flags); BUG_ON(lockres->l_flags & OCFS2_LOCK_ATTACHED); @@ -1220,7 +1218,7 @@ int ocfs2_create_new_inode_locks(struct inode *inode) } /* - * We don't want to use LKM_LOCAL on a meta data lock as they + * We don't want to use DLM_LKF_LOCAL on a meta data lock as they * don't use a generation in their lock names. */ ret = ocfs2_create_new_lock(osb, &OCFS2_I(inode)->ip_inode_lockres, 1, 0); @@ -1259,7 +1257,7 @@ int ocfs2_rw_lock(struct inode *inode, int write) lockres = &OCFS2_I(inode)->ip_rw_lockres; - level = write ? LKM_EXMODE : LKM_PRMODE; + level = write ? DLM_LOCK_EX : DLM_LOCK_PR; status = ocfs2_cluster_lock(OCFS2_SB(inode->i_sb), lockres, level, 0, 0); @@ -1272,7 +1270,7 @@ int ocfs2_rw_lock(struct inode *inode, int write) void ocfs2_rw_unlock(struct inode *inode, int write) { - int level = write ? LKM_EXMODE : LKM_PRMODE; + int level = write ? DLM_LOCK_EX : DLM_LOCK_PR; struct ocfs2_lock_res *lockres = &OCFS2_I(inode)->ip_rw_lockres; struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); @@ -1310,7 +1308,7 @@ int ocfs2_open_lock(struct inode *inode) lockres = &OCFS2_I(inode)->ip_open_lockres; status = ocfs2_cluster_lock(OCFS2_SB(inode->i_sb), lockres, - LKM_PRMODE, 0, 0); + DLM_LOCK_PR, 0, 0); if (status < 0) mlog_errno(status); @@ -1338,16 +1336,16 @@ int ocfs2_try_open_lock(struct inode *inode, int write) lockres = &OCFS2_I(inode)->ip_open_lockres; - level = write ? LKM_EXMODE : LKM_PRMODE; + level = write ? DLM_LOCK_EX : DLM_LOCK_PR; /* * The file system may already holding a PRMODE/EXMODE open lock. - * Since we pass LKM_NOQUEUE, the request won't block waiting on + * Since we pass DLM_LKF_NOQUEUE, the request won't block waiting on * other nodes and the -EAGAIN will indicate to the caller that * this inode is still in use. */ status = ocfs2_cluster_lock(OCFS2_SB(inode->i_sb), lockres, - level, LKM_NOQUEUE, 0); + level, DLM_LKF_NOQUEUE, 0); out: mlog_exit(status); @@ -1372,10 +1370,10 @@ void ocfs2_open_unlock(struct inode *inode) if(lockres->l_ro_holders) ocfs2_cluster_unlock(OCFS2_SB(inode->i_sb), lockres, - LKM_PRMODE); + DLM_LOCK_PR); if(lockres->l_ex_holders) ocfs2_cluster_unlock(OCFS2_SB(inode->i_sb), lockres, - LKM_EXMODE); + DLM_LOCK_EX); out: mlog_exit_void(); @@ -1462,7 +1460,7 @@ int ocfs2_file_lock(struct file *file, int ex, int trylock) ocfs2_init_mask_waiter(&mw); if ((lockres->l_flags & OCFS2_LOCK_BUSY) || - (lockres->l_level > LKM_NLMODE)) { + (lockres->l_level > DLM_LOCK_NL)) { mlog(ML_ERROR, "File lock \"%s\" has busy or locked state: flags: 0x%lx, " "level: %u\n", lockres->l_name, lockres->l_flags, @@ -1570,7 +1568,7 @@ void ocfs2_file_unlock(struct file *file) * Fake a blocking ast for the downconvert code. */ lockres_or_flags(lockres, OCFS2_LOCK_BLOCKED); - lockres->l_blocking = LKM_EXMODE; + lockres->l_blocking = DLM_LOCK_EX; ocfs2_prepare_downconvert(lockres, LKM_NLMODE); lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_BUSY, 0); @@ -1599,11 +1597,11 @@ static void ocfs2_downconvert_on_unlock(struct ocfs2_super *osb, * condition. */ if (lockres->l_flags & OCFS2_LOCK_BLOCKED) { switch(lockres->l_blocking) { - case LKM_EXMODE: + case DLM_LOCK_EX: if (!lockres->l_ex_holders && !lockres->l_ro_holders) kick = 1; break; - case LKM_PRMODE: + case DLM_LOCK_PR: if (!lockres->l_ex_holders) kick = 1; break; @@ -1921,7 +1919,8 @@ int ocfs2_inode_lock_full(struct inode *inode, int ex, int arg_flags) { - int status, level, dlm_flags, acquired; + int status, level, acquired; + u32 dlm_flags; struct ocfs2_lock_res *lockres = NULL; struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); struct buffer_head *local_bh = NULL; @@ -1951,10 +1950,10 @@ int ocfs2_inode_lock_full(struct inode *inode, ocfs2_wait_for_recovery(osb); lockres = &OCFS2_I(inode)->ip_inode_lockres; - level = ex ? LKM_EXMODE : LKM_PRMODE; + level = ex ? DLM_LOCK_EX : DLM_LOCK_PR; dlm_flags = 0; if (arg_flags & OCFS2_META_LOCK_NOQUEUE) - dlm_flags |= LKM_NOQUEUE; + dlm_flags |= DLM_LKF_NOQUEUE; status = ocfs2_cluster_lock(osb, lockres, level, dlm_flags, arg_flags); if (status < 0) { @@ -2105,7 +2104,7 @@ int ocfs2_inode_lock_atime(struct inode *inode, void ocfs2_inode_unlock(struct inode *inode, int ex) { - int level = ex ? LKM_EXMODE : LKM_PRMODE; + int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR; struct ocfs2_lock_res *lockres = &OCFS2_I(inode)->ip_inode_lockres; struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); @@ -2126,7 +2125,7 @@ int ocfs2_super_lock(struct ocfs2_super *osb, int ex) { int status = 0; - int level = ex ? LKM_EXMODE : LKM_PRMODE; + int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR; struct ocfs2_lock_res *lockres = &osb->osb_super_lockres; mlog_entry_void(); @@ -2168,7 +2167,7 @@ bail: void ocfs2_super_unlock(struct ocfs2_super *osb, int ex) { - int level = ex ? LKM_EXMODE : LKM_PRMODE; + int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR; struct ocfs2_lock_res *lockres = &osb->osb_super_lockres; if (!ocfs2_mount_local(osb)) @@ -2186,7 +2185,7 @@ int ocfs2_rename_lock(struct ocfs2_super *osb) if (ocfs2_mount_local(osb)) return 0; - status = ocfs2_cluster_lock(osb, lockres, LKM_EXMODE, 0, 0); + status = ocfs2_cluster_lock(osb, lockres, DLM_LOCK_EX, 0, 0); if (status < 0) mlog_errno(status); @@ -2198,13 +2197,13 @@ void ocfs2_rename_unlock(struct ocfs2_super *osb) struct ocfs2_lock_res *lockres = &osb->osb_rename_lockres; if (!ocfs2_mount_local(osb)) - ocfs2_cluster_unlock(osb, lockres, LKM_EXMODE); + ocfs2_cluster_unlock(osb, lockres, DLM_LOCK_EX); } int ocfs2_dentry_lock(struct dentry *dentry, int ex) { int ret; - int level = ex ? LKM_EXMODE : LKM_PRMODE; + int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR; struct ocfs2_dentry_lock *dl = dentry->d_fsdata; struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb); @@ -2225,7 +2224,7 @@ int ocfs2_dentry_lock(struct dentry *dentry, int ex) void ocfs2_dentry_unlock(struct dentry *dentry, int ex) { - int level = ex ? LKM_EXMODE : LKM_PRMODE; + int level = ex ? DLM_LOCK_EX : DLM_LOCK_PR; struct ocfs2_dentry_lock *dl = dentry->d_fsdata; struct ocfs2_super *osb = OCFS2_SB(dentry->d_sb); @@ -2614,7 +2613,7 @@ static void ocfs2_unlock_ast(void *opaque, enum dlm_status status) lockres->l_action = OCFS2_AST_INVALID; break; case OCFS2_UNLOCK_DROP_LOCK: - lockres->l_level = LKM_IVMODE; + lockres->l_level = DLM_LOCK_IV; break; default: BUG(); @@ -2635,14 +2634,14 @@ static int ocfs2_drop_lock(struct ocfs2_super *osb, { enum dlm_status status; unsigned long flags; - int lkm_flags = 0; + u32 lkm_flags = 0; /* We didn't get anywhere near actually using this lockres. */ if (!(lockres->l_flags & OCFS2_LOCK_INITIALIZED)) goto out; if (lockres->l_ops->flags & LOCK_TYPE_USES_LVB) - lkm_flags |= LKM_VALBLK; + lkm_flags |= DLM_LKF_VALBLK; spin_lock_irqsave(&lockres->l_lock, flags); @@ -2668,7 +2667,7 @@ static int ocfs2_drop_lock(struct ocfs2_super *osb, if (lockres->l_ops->flags & LOCK_TYPE_USES_LVB) { if (lockres->l_flags & OCFS2_LOCK_ATTACHED && - lockres->l_level == LKM_EXMODE && + lockres->l_level == DLM_LOCK_EX && !(lockres->l_flags & OCFS2_LOCK_NEEDS_REFRESH)) lockres->l_ops->set_lvb(lockres); } @@ -2801,10 +2800,10 @@ static void ocfs2_prepare_downconvert(struct ocfs2_lock_res *lockres, { assert_spin_locked(&lockres->l_lock); - BUG_ON(lockres->l_blocking <= LKM_NLMODE); + BUG_ON(lockres->l_blocking <= DLM_LOCK_NL); if (lockres->l_level <= new_level) { - mlog(ML_ERROR, "lockres->l_level (%u) <= new_level (%u)\n", + mlog(ML_ERROR, "lockres->l_level (%d) <= new_level (%d)\n", lockres->l_level, new_level); BUG(); } @@ -2822,13 +2821,14 @@ static int ocfs2_downconvert_lock(struct ocfs2_super *osb, int new_level, int lvb) { - int ret, dlm_flags = LKM_CONVERT; + int ret; + u32 dlm_flags = DLM_LKF_CONVERT; enum dlm_status status; mlog_entry_void(); if (lvb) - dlm_flags |= LKM_VALBLK; + dlm_flags |= DLM_LKF_VALBLK; status = ocfs2_dlm_lock(osb->dlm, new_level, @@ -2894,7 +2894,7 @@ static int ocfs2_cancel_convert(struct ocfs2_super *osb, ret = 0; status = ocfs2_dlm_unlock(osb->dlm, &lockres->l_lksb, - LKM_CANCEL, + DLM_LKF_CANCEL, lockres); if (status != DLM_NORMAL) { ocfs2_log_dlm_error("ocfs2_dlm_unlock", status, lockres); @@ -2939,13 +2939,13 @@ recheck: /* if we're blocking an exclusive and we have *any* holders, * then requeue. */ - if ((lockres->l_blocking == LKM_EXMODE) + if ((lockres->l_blocking == DLM_LOCK_EX) && (lockres->l_ex_holders || lockres->l_ro_holders)) goto leave_requeue; /* If it's a PR we're blocking, then only * requeue if we've got any EX holders */ - if (lockres->l_blocking == LKM_PRMODE && + if (lockres->l_blocking == DLM_LOCK_PR && lockres->l_ex_holders) goto leave_requeue; @@ -2992,7 +2992,7 @@ downconvert: ctl->requeue = 0; if (lockres->l_ops->flags & LOCK_TYPE_USES_LVB) { - if (lockres->l_level == LKM_EXMODE) + if (lockres->l_level == DLM_LOCK_EX) set_lvb = 1; /* @@ -3046,7 +3046,7 @@ static int ocfs2_data_convert_worker(struct ocfs2_lock_res *lockres, (unsigned long long)OCFS2_I(inode)->ip_blkno); } sync_mapping_buffers(mapping); - if (blocking == LKM_EXMODE) { + if (blocking == DLM_LOCK_EX) { truncate_inode_pages(mapping, 0); } else { /* We only need to wait on the I/O if we're not also @@ -3067,8 +3067,8 @@ static int ocfs2_check_meta_downconvert(struct ocfs2_lock_res *lockres, struct inode *inode = ocfs2_lock_res_inode(lockres); int checkpointed = ocfs2_inode_fully_checkpointed(inode); - BUG_ON(new_level != LKM_NLMODE && new_level != LKM_PRMODE); - BUG_ON(lockres->l_level != LKM_EXMODE && !checkpointed); + BUG_ON(new_level != DLM_LOCK_NL && new_level != DLM_LOCK_PR); + BUG_ON(lockres->l_level != DLM_LOCK_EX && !checkpointed); if (checkpointed) return 1; @@ -3132,7 +3132,7 @@ static int ocfs2_dentry_convert_worker(struct ocfs2_lock_res *lockres, * valid. The downconvert code will retain a PR for this node, * so there's no further work to do. */ - if (blocking == LKM_PRMODE) + if (blocking == DLM_LOCK_PR) return UNBLOCK_CONTINUE; /* diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c index 4f44f23795f..99538043fc1 100644 --- a/fs/ocfs2/stackglue.c +++ b/fs/ocfs2/stackglue.c @@ -18,15 +18,65 @@ * General Public License for more details. */ -#include -#include - -#include "dlm/dlmapi.h" - #include "stackglue.h" static struct ocfs2_locking_protocol *lproto; +/* These should be identical */ +#if (DLM_LOCK_IV != LKM_IVMODE) +# error Lock modes do not match +#endif +#if (DLM_LOCK_NL != LKM_NLMODE) +# error Lock modes do not match +#endif +#if (DLM_LOCK_CR != LKM_CRMODE) +# error Lock modes do not match +#endif +#if (DLM_LOCK_CW != LKM_CWMODE) +# error Lock modes do not match +#endif +#if (DLM_LOCK_PR != LKM_PRMODE) +# error Lock modes do not match +#endif +#if (DLM_LOCK_PW != LKM_PWMODE) +# error Lock modes do not match +#endif +#if (DLM_LOCK_EX != LKM_EXMODE) +# error Lock modes do not match +#endif +static inline int mode_to_o2dlm(int mode) +{ + BUG_ON(mode > LKM_MAXMODE); + + return mode; +} + +#define map_flag(_generic, _o2dlm) \ + if (flags & (_generic)) { \ + flags &= ~(_generic); \ + o2dlm_flags |= (_o2dlm); \ + } +static int flags_to_o2dlm(u32 flags) +{ + int o2dlm_flags = 0; + + map_flag(DLM_LKF_NOQUEUE, LKM_NOQUEUE); + map_flag(DLM_LKF_CANCEL, LKM_CANCEL); + map_flag(DLM_LKF_CONVERT, LKM_CONVERT); + map_flag(DLM_LKF_VALBLK, LKM_VALBLK); + map_flag(DLM_LKF_IVVALBLK, LKM_INVVALBLK); + map_flag(DLM_LKF_ORPHAN, LKM_ORPHAN); + map_flag(DLM_LKF_FORCEUNLOCK, LKM_FORCE); + map_flag(DLM_LKF_TIMEOUT, LKM_TIMEOUT); + map_flag(DLM_LKF_LOCAL, LKM_LOCAL); + + /* map_flag() should have cleared every flag passed in */ + BUG_ON(flags != 0); + + return o2dlm_flags; +} +#undef map_flag + enum dlm_status ocfs2_dlm_lock(struct dlm_ctxt *dlm, int mode, struct dlm_lockstatus *lksb, @@ -35,8 +85,12 @@ enum dlm_status ocfs2_dlm_lock(struct dlm_ctxt *dlm, unsigned int namelen, void *astarg) { + int o2dlm_mode = mode_to_o2dlm(mode); + int o2dlm_flags = flags_to_o2dlm(flags); + BUG_ON(lproto == NULL); - return dlmlock(dlm, mode, lksb, flags, name, namelen, + + return dlmlock(dlm, o2dlm_mode, lksb, o2dlm_flags, name, namelen, lproto->lp_lock_ast, astarg, lproto->lp_blocking_ast); } @@ -46,9 +100,12 @@ enum dlm_status ocfs2_dlm_unlock(struct dlm_ctxt *dlm, u32 flags, void *astarg) { + int o2dlm_flags = flags_to_o2dlm(flags); + BUG_ON(lproto == NULL); - return dlmunlock(dlm, lksb, flags, lproto->lp_unlock_ast, astarg); + return dlmunlock(dlm, lksb, o2dlm_flags, + lproto->lp_unlock_ast, astarg); } diff --git a/fs/ocfs2/stackglue.h b/fs/ocfs2/stackglue.h index 40a00241340..986d059ed1e 100644 --- a/fs/ocfs2/stackglue.h +++ b/fs/ocfs2/stackglue.h @@ -21,6 +21,19 @@ #ifndef STACKGLUE_H #define STACKGLUE_H +#include +#include +#include + +/* + * dlmconstants.h does not have a LOCAL flag. We hope to remove it + * some day, but right now we need it. Let's fake it. This value is larger + * than any flag in dlmconstants.h. + */ +#define DLM_LKF_LOCAL 0x00100000 + +#include "dlm/dlmapi.h" + struct ocfs2_locking_protocol { void (*lp_lock_ast)(void *astarg); void (*lp_blocking_ast)(void *astarg, int level); -- cgit v1.2.3 From 7431cd7e8dd0e46e9b12bd6a1ac1286f4b420371 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Fri, 1 Feb 2008 12:15:37 -0800 Subject: ocfs2: Use -errno instead of dlm_status for ocfs2_dlm_lock/unlock() API. Change the ocfs2_dlm_lock/unlock() functions to return -errno values. This is the first step towards elminiating dlm_status in fs/ocfs2/dlmglue.c. The change also passes -errno values to ->unlock_ast(). [ Fix a return code in dlmglue.c and change the error translation table into an array of ints. --Mark ] Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlmglue.c | 116 ++++++++++++++++++----------------------- fs/ocfs2/stackglue.c | 142 ++++++++++++++++++++++++++++++++++++++++++++++++--- fs/ocfs2/stackglue.h | 6 +-- 3 files changed, 188 insertions(+), 76 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index b8ac903ef8e..6a222a5c81d 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -329,10 +329,9 @@ static void ocfs2_schedule_blocked_lock(struct ocfs2_super *osb, struct ocfs2_lock_res *lockres); static inline void ocfs2_recover_from_dlm_error(struct ocfs2_lock_res *lockres, int convert); -#define ocfs2_log_dlm_error(_func, _stat, _lockres) do { \ - mlog(ML_ERROR, "Dlm error \"%s\" while calling %s on " \ - "resource %s: %s\n", dlm_errname(_stat), _func, \ - _lockres->l_name, dlm_errmsg(_stat)); \ +#define ocfs2_log_dlm_error(_func, _err, _lockres) do { \ + mlog(ML_ERROR, "DLM error %d while calling %s on resource %s\n", \ + _err, _func, _lockres->l_name); \ } while (0) static int ocfs2_downconvert_thread(void *arg); static void ocfs2_downconvert_on_unlock(struct ocfs2_super *osb, @@ -867,7 +866,6 @@ static int ocfs2_lock_create(struct ocfs2_super *osb, u32 dlm_flags) { int ret = 0; - enum dlm_status status = DLM_NORMAL; unsigned long flags; mlog_entry_void(); @@ -887,21 +885,19 @@ static int ocfs2_lock_create(struct ocfs2_super *osb, lockres_or_flags(lockres, OCFS2_LOCK_BUSY); spin_unlock_irqrestore(&lockres->l_lock, flags); - status = ocfs2_dlm_lock(osb->dlm, - level, - &lockres->l_lksb, - dlm_flags, - lockres->l_name, - OCFS2_LOCK_ID_MAX_LEN - 1, - lockres); - if (status != DLM_NORMAL) { - ocfs2_log_dlm_error("ocfs2_dlm_lock", status, lockres); - ret = -EINVAL; + ret = ocfs2_dlm_lock(osb->dlm, + level, + &lockres->l_lksb, + dlm_flags, + lockres->l_name, + OCFS2_LOCK_ID_MAX_LEN - 1, + lockres); + if (ret) { + ocfs2_log_dlm_error("ocfs2_dlm_lock", ret, lockres); ocfs2_recover_from_dlm_error(lockres, 1); } - mlog(0, "lock %s, successfull return from ocfs2_dlm_lock\n", - lockres->l_name); + mlog(0, "lock %s, return from ocfs2_dlm_lock\n", lockres->l_name); bail: mlog_exit(ret); @@ -1018,7 +1014,6 @@ static int ocfs2_cluster_lock(struct ocfs2_super *osb, int arg_flags) { struct ocfs2_mask_waiter mw; - enum dlm_status status; int wait, catch_signals = !(osb->s_mount_opt & OCFS2_MOUNT_NOINTR); int ret = 0; /* gcc doesn't realize wait = 1 guarantees ret is set */ unsigned long flags; @@ -1089,21 +1084,18 @@ again: lockres->l_name, lockres->l_level, level); /* call dlm_lock to upgrade lock now */ - status = ocfs2_dlm_lock(osb->dlm, - level, - &lockres->l_lksb, - lkm_flags, - lockres->l_name, - OCFS2_LOCK_ID_MAX_LEN - 1, - lockres); - if (status != DLM_NORMAL) { - if ((lkm_flags & DLM_LKF_NOQUEUE) && - (status == DLM_NOTQUEUED)) - ret = -EAGAIN; - else { + ret = ocfs2_dlm_lock(osb->dlm, + level, + &lockres->l_lksb, + lkm_flags, + lockres->l_name, + OCFS2_LOCK_ID_MAX_LEN - 1, + lockres); + if (ret) { + if (!(lkm_flags & DLM_LKF_NOQUEUE) || + (ret != -EAGAIN)) { ocfs2_log_dlm_error("ocfs2_dlm_lock", - status, lockres); - ret = -EINVAL; + ret, lockres); } ocfs2_recover_from_dlm_error(lockres, 1); goto out; @@ -1502,10 +1494,8 @@ int ocfs2_file_lock(struct file *file, int ex, int trylock) ret = ocfs2_dlm_lock(osb->dlm, level, &lockres->l_lksb, lkm_flags, lockres->l_name, OCFS2_LOCK_ID_MAX_LEN - 1, lockres); - if (ret != DLM_NORMAL) { - if (trylock && ret == DLM_NOTQUEUED) - ret = -EAGAIN; - else { + if (ret) { + if (!trylock || (ret != -EAGAIN)) { ocfs2_log_dlm_error("ocfs2_dlm_lock", ret, lockres); ret = -EINVAL; } @@ -2573,7 +2563,7 @@ void ocfs2_dlm_shutdown(struct ocfs2_super *osb) mlog_exit_void(); } -static void ocfs2_unlock_ast(void *opaque, enum dlm_status status) +static void ocfs2_unlock_ast(void *opaque, int error) { struct ocfs2_lock_res *lockres = opaque; unsigned long flags; @@ -2589,7 +2579,7 @@ static void ocfs2_unlock_ast(void *opaque, enum dlm_status status) * state. The wake_up call done at the bottom is redundant * (ocfs2_prepare_cancel_convert doesn't sleep on this) but doesn't * hurt anything anyway */ - if (status == DLM_CANCELGRANT && + if (error == -DLM_ECANCEL && lockres->l_unlock_action == OCFS2_UNLOCK_CANCEL_CONVERT) { mlog(0, "Got cancelgrant for %s\n", lockres->l_name); @@ -2599,9 +2589,10 @@ static void ocfs2_unlock_ast(void *opaque, enum dlm_status status) goto complete_unlock; } - if (status != DLM_NORMAL) { - mlog(ML_ERROR, "Dlm passes status %d for lock %s, " - "unlock_action %d\n", status, lockres->l_name, + /* DLM_EUNLOCK is the success code for unlock */ + if (error != -DLM_EUNLOCK) { + mlog(ML_ERROR, "Dlm passes error %d for lock %s, " + "unlock_action %d\n", error, lockres->l_name, lockres->l_unlock_action); spin_unlock_irqrestore(&lockres->l_lock, flags); return; @@ -2632,7 +2623,7 @@ complete_unlock: static int ocfs2_drop_lock(struct ocfs2_super *osb, struct ocfs2_lock_res *lockres) { - enum dlm_status status; + int ret; unsigned long flags; u32 lkm_flags = 0; @@ -2696,10 +2687,10 @@ static int ocfs2_drop_lock(struct ocfs2_super *osb, mlog(0, "lock %s\n", lockres->l_name); - status = ocfs2_dlm_unlock(osb->dlm, &lockres->l_lksb, lkm_flags, - lockres); - if (status != DLM_NORMAL) { - ocfs2_log_dlm_error("ocfs2_dlm_unlock", status, lockres); + ret = ocfs2_dlm_unlock(osb->dlm, &lockres->l_lksb, lkm_flags, + lockres); + if (ret) { + ocfs2_log_dlm_error("ocfs2_dlm_unlock", ret, lockres); mlog(ML_ERROR, "lockres flags: %lu\n", lockres->l_flags); dlm_print_one_lock(lockres->l_lksb.lockid); BUG(); @@ -2823,23 +2814,21 @@ static int ocfs2_downconvert_lock(struct ocfs2_super *osb, { int ret; u32 dlm_flags = DLM_LKF_CONVERT; - enum dlm_status status; mlog_entry_void(); if (lvb) dlm_flags |= DLM_LKF_VALBLK; - status = ocfs2_dlm_lock(osb->dlm, - new_level, - &lockres->l_lksb, - dlm_flags, - lockres->l_name, - OCFS2_LOCK_ID_MAX_LEN - 1, - lockres); - if (status != DLM_NORMAL) { - ocfs2_log_dlm_error("ocfs2_dlm_lock", status, lockres); - ret = -EINVAL; + ret = ocfs2_dlm_lock(osb->dlm, + new_level, + &lockres->l_lksb, + dlm_flags, + lockres->l_name, + OCFS2_LOCK_ID_MAX_LEN - 1, + lockres); + if (ret) { + ocfs2_log_dlm_error("ocfs2_dlm_lock", ret, lockres); ocfs2_recover_from_dlm_error(lockres, 1); goto bail; } @@ -2886,19 +2875,14 @@ static int ocfs2_cancel_convert(struct ocfs2_super *osb, struct ocfs2_lock_res *lockres) { int ret; - enum dlm_status status; mlog_entry_void(); mlog(0, "lock %s\n", lockres->l_name); - ret = 0; - status = ocfs2_dlm_unlock(osb->dlm, - &lockres->l_lksb, - DLM_LKF_CANCEL, - lockres); - if (status != DLM_NORMAL) { - ocfs2_log_dlm_error("ocfs2_dlm_unlock", status, lockres); - ret = -EINVAL; + ret = ocfs2_dlm_unlock(osb->dlm, &lockres->l_lksb, + DLM_LKF_CANCEL, lockres); + if (ret) { + ocfs2_log_dlm_error("ocfs2_dlm_unlock", ret, lockres); ocfs2_recover_from_dlm_error(lockres, 0); } diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c index 99538043fc1..0aec2fcf217 100644 --- a/fs/ocfs2/stackglue.c +++ b/fs/ocfs2/stackglue.c @@ -18,6 +18,7 @@ * General Public License for more details. */ +#include "cluster/masklog.h" #include "stackglue.h" static struct ocfs2_locking_protocol *lproto; @@ -77,7 +78,126 @@ static int flags_to_o2dlm(u32 flags) } #undef map_flag -enum dlm_status ocfs2_dlm_lock(struct dlm_ctxt *dlm, +/* + * Map an o2dlm status to standard errno values. + * + * o2dlm only uses a handful of these, and returns even fewer to the + * caller. Still, we try to assign sane values to each error. + * + * The following value pairs have special meanings to dlmglue, thus + * the right hand side needs to stay unique - never duplicate the + * mapping elsewhere in the table! + * + * DLM_NORMAL: 0 + * DLM_NOTQUEUED: -EAGAIN + * DLM_CANCELGRANT: -DLM_ECANCEL + * DLM_CANCEL: -DLM_EUNLOCK + */ +/* Keep in sync with dlmapi.h */ +static int status_map[] = { + [DLM_NORMAL] = 0, /* Success */ + [DLM_GRANTED] = -EINVAL, + [DLM_DENIED] = -EACCES, + [DLM_DENIED_NOLOCKS] = -EACCES, + [DLM_WORKING] = -EBUSY, + [DLM_BLOCKED] = -EINVAL, + [DLM_BLOCKED_ORPHAN] = -EINVAL, + [DLM_DENIED_GRACE_PERIOD] = -EACCES, + [DLM_SYSERR] = -ENOMEM, /* It is what it is */ + [DLM_NOSUPPORT] = -EPROTO, + [DLM_CANCELGRANT] = -DLM_ECANCEL, /* Cancel after grant */ + [DLM_IVLOCKID] = -EINVAL, + [DLM_SYNC] = -EINVAL, + [DLM_BADTYPE] = -EINVAL, + [DLM_BADRESOURCE] = -EINVAL, + [DLM_MAXHANDLES] = -ENOMEM, + [DLM_NOCLINFO] = -EINVAL, + [DLM_NOLOCKMGR] = -EINVAL, + [DLM_NOPURGED] = -EINVAL, + [DLM_BADARGS] = -EINVAL, + [DLM_VOID] = -EINVAL, + [DLM_NOTQUEUED] = -EAGAIN, /* Trylock failed */ + [DLM_IVBUFLEN] = -EINVAL, + [DLM_CVTUNGRANT] = -EPERM, + [DLM_BADPARAM] = -EINVAL, + [DLM_VALNOTVALID] = -EINVAL, + [DLM_REJECTED] = -EPERM, + [DLM_ABORT] = -EINVAL, + [DLM_CANCEL] = -DLM_EUNLOCK, /* Successful cancel */ + [DLM_IVRESHANDLE] = -EINVAL, + [DLM_DEADLOCK] = -EDEADLK, + [DLM_DENIED_NOASTS] = -EINVAL, + [DLM_FORWARD] = -EINVAL, + [DLM_TIMEOUT] = -ETIMEDOUT, + [DLM_IVGROUPID] = -EINVAL, + [DLM_VERS_CONFLICT] = -EOPNOTSUPP, + [DLM_BAD_DEVICE_PATH] = -ENOENT, + [DLM_NO_DEVICE_PERMISSION] = -EPERM, + [DLM_NO_CONTROL_DEVICE] = -ENOENT, + [DLM_RECOVERING] = -ENOTCONN, + [DLM_MIGRATING] = -ERESTART, + [DLM_MAXSTATS] = -EINVAL, +}; +static int dlm_status_to_errno(enum dlm_status status) +{ + BUG_ON(status > (sizeof(status_map) / sizeof(status_map[0]))); + + return status_map[status]; +} + +static void o2dlm_lock_ast_wrapper(void *astarg) +{ + BUG_ON(lproto == NULL); + + lproto->lp_lock_ast(astarg); +} + +static void o2dlm_blocking_ast_wrapper(void *astarg, int level) +{ + BUG_ON(lproto == NULL); + + lproto->lp_blocking_ast(astarg, level); +} + +static void o2dlm_unlock_ast_wrapper(void *astarg, enum dlm_status status) +{ + int error; + + BUG_ON(lproto == NULL); + + /* + * XXX: CANCEL values are sketchy. + * + * Currently we have preserved the o2dlm paradigm. You can get + * unlock_ast() whether the cancel succeded or not. + * + * First, we're going to pass DLM_EUNLOCK just like fs/dlm does for + * successful unlocks. That is a clean behavior. + * + * In o2dlm, you can get both the lock_ast() for the lock being + * granted and the unlock_ast() for the CANCEL failing. A + * successful cancel sends DLM_NORMAL here. If the + * lock grant happened before the cancel arrived, you get + * DLM_CANCELGRANT. For now, we'll use DLM_ECANCEL to signify + * CANCELGRANT - the CANCEL was supposed to happen but didn't. We + * can then use DLM_EUNLOCK to signify a successful CANCEL - + * effectively, the CANCEL caused the lock to roll back. + * + * In the future, we will likely move the o2dlm to send only one + * ast - either unlock_ast() for a successful CANCEL or lock_ast() + * when the grant succeeds. At that point, we'll send DLM_ECANCEL + * for all cancel results (CANCELGRANT will no longer exist). + */ + error = dlm_status_to_errno(status); + + /* Successful unlock is DLM_EUNLOCK */ + if (!error) + error = -DLM_EUNLOCK; + + lproto->lp_unlock_ast(astarg, error); +} + +int ocfs2_dlm_lock(struct dlm_ctxt *dlm, int mode, struct dlm_lockstatus *lksb, u32 flags, @@ -85,27 +205,35 @@ enum dlm_status ocfs2_dlm_lock(struct dlm_ctxt *dlm, unsigned int namelen, void *astarg) { + enum dlm_status status; int o2dlm_mode = mode_to_o2dlm(mode); int o2dlm_flags = flags_to_o2dlm(flags); + int ret; BUG_ON(lproto == NULL); - return dlmlock(dlm, o2dlm_mode, lksb, o2dlm_flags, name, namelen, - lproto->lp_lock_ast, astarg, - lproto->lp_blocking_ast); + status = dlmlock(dlm, o2dlm_mode, lksb, o2dlm_flags, name, namelen, + o2dlm_lock_ast_wrapper, astarg, + o2dlm_blocking_ast_wrapper); + ret = dlm_status_to_errno(status); + return ret; } -enum dlm_status ocfs2_dlm_unlock(struct dlm_ctxt *dlm, +int ocfs2_dlm_unlock(struct dlm_ctxt *dlm, struct dlm_lockstatus *lksb, u32 flags, void *astarg) { + enum dlm_status status; int o2dlm_flags = flags_to_o2dlm(flags); + int ret; BUG_ON(lproto == NULL); - return dlmunlock(dlm, lksb, o2dlm_flags, - lproto->lp_unlock_ast, astarg); + status = dlmunlock(dlm, lksb, o2dlm_flags, + o2dlm_unlock_ast_wrapper, astarg); + ret = dlm_status_to_errno(status); + return ret; } diff --git a/fs/ocfs2/stackglue.h b/fs/ocfs2/stackglue.h index 986d059ed1e..8ebcfba62c7 100644 --- a/fs/ocfs2/stackglue.h +++ b/fs/ocfs2/stackglue.h @@ -37,17 +37,17 @@ struct ocfs2_locking_protocol { void (*lp_lock_ast)(void *astarg); void (*lp_blocking_ast)(void *astarg, int level); - void (*lp_unlock_ast)(void *astarg, enum dlm_status status); + void (*lp_unlock_ast)(void *astarg, int error); }; -enum dlm_status ocfs2_dlm_lock(struct dlm_ctxt *dlm, +int ocfs2_dlm_lock(struct dlm_ctxt *dlm, int mode, struct dlm_lockstatus *lksb, u32 flags, void *name, unsigned int namelen, void *astarg); -enum dlm_status ocfs2_dlm_unlock(struct dlm_ctxt *dlm, +int ocfs2_dlm_unlock(struct dlm_ctxt *dlm, struct dlm_lockstatus *lksb, u32 flags, void *astarg); -- cgit v1.2.3 From 8f2c9c1b16bf6ed0903b29c49d56fa0109a390e4 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Fri, 1 Feb 2008 12:16:57 -0800 Subject: ocfs2: Create the lock status block union. Wrap the lock status block (lksb) in a union. Later we will add a union element for the fs/dlm lksb. Create accessors for the status and lvb fields. Other than a debugging function, dlmglue.c does not directly reference the o2dlm locking path anymore. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlmglue.c | 23 +++++++++++++---------- fs/ocfs2/ocfs2.h | 5 +++-- fs/ocfs2/stackglue.c | 29 ++++++++++++++++++++++------- fs/ocfs2/stackglue.h | 11 +++++++++-- 4 files changed, 47 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index 6a222a5c81d..459037653e5 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -112,7 +112,8 @@ static void ocfs2_dump_meta_lvb_info(u64 level, unsigned int line, struct ocfs2_lock_res *lockres) { - struct ocfs2_meta_lvb *lvb = (struct ocfs2_meta_lvb *) lockres->l_lksb.lvb; + struct ocfs2_meta_lvb *lvb = + (struct ocfs2_meta_lvb *)ocfs2_dlm_lvb(&lockres->l_lksb); mlog(level, "LVB information for %s (called from %s:%u):\n", lockres->l_name, function, line); @@ -799,14 +800,14 @@ static void ocfs2_blocking_ast(void *opaque, int level) static void ocfs2_locking_ast(void *opaque) { struct ocfs2_lock_res *lockres = opaque; - struct dlm_lockstatus *lksb = &lockres->l_lksb; unsigned long flags; spin_lock_irqsave(&lockres->l_lock, flags); - if (lksb->status != DLM_NORMAL) { - mlog(ML_ERROR, "lockres %s: lksb status value of %u!\n", - lockres->l_name, lksb->status); + if (ocfs2_dlm_lock_status(&lockres->l_lksb)) { + mlog(ML_ERROR, "lockres %s: lksb status value of %d!\n", + lockres->l_name, + ocfs2_dlm_lock_status(&lockres->l_lksb)); spin_unlock_irqrestore(&lockres->l_lock, flags); return; } @@ -1634,7 +1635,7 @@ static void __ocfs2_stuff_meta_lvb(struct inode *inode) mlog_entry_void(); - lvb = (struct ocfs2_meta_lvb *) lockres->l_lksb.lvb; + lvb = (struct ocfs2_meta_lvb *)ocfs2_dlm_lvb(&lockres->l_lksb); /* * Invalidate the LVB of a deleted inode - this way other @@ -1686,7 +1687,7 @@ static void ocfs2_refresh_inode_from_lvb(struct inode *inode) mlog_meta_lvb(0, lockres); - lvb = (struct ocfs2_meta_lvb *) lockres->l_lksb.lvb; + lvb = (struct ocfs2_meta_lvb *)ocfs2_dlm_lvb(&lockres->l_lksb); /* We're safe here without the lockres lock... */ spin_lock(&oi->ip_lock); @@ -1721,7 +1722,8 @@ static void ocfs2_refresh_inode_from_lvb(struct inode *inode) static inline int ocfs2_meta_lvb_is_trustable(struct inode *inode, struct ocfs2_lock_res *lockres) { - struct ocfs2_meta_lvb *lvb = (struct ocfs2_meta_lvb *) lockres->l_lksb.lvb; + struct ocfs2_meta_lvb *lvb = + (struct ocfs2_meta_lvb *)ocfs2_dlm_lvb(&lockres->l_lksb); if (lvb->lvb_version == OCFS2_LVB_VERSION && be32_to_cpu(lvb->lvb_igeneration) == inode->i_generation) @@ -2379,7 +2381,7 @@ static int ocfs2_dlm_seq_show(struct seq_file *m, void *v) lockres->l_blocking); /* Dump the raw LVB */ - lvb = lockres->l_lksb.lvb; + lvb = ocfs2_dlm_lvb(&lockres->l_lksb); for(i = 0; i < DLM_LVB_LEN; i++) seq_printf(m, "0x%x\t", lvb[i]); @@ -2692,7 +2694,8 @@ static int ocfs2_drop_lock(struct ocfs2_super *osb, if (ret) { ocfs2_log_dlm_error("ocfs2_dlm_unlock", ret, lockres); mlog(ML_ERROR, "lockres flags: %lu\n", lockres->l_flags); - dlm_print_one_lock(lockres->l_lksb.lockid); + /* XXX Need to abstract this */ + dlm_print_one_lock(lockres->l_lksb.lksb_o2dlm.lockid); BUG(); } mlog(0, "lock %s, successfull return from ocfs2_dlm_unlock\n", diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h index f78e9ed5324..6d7c6d2d0c2 100644 --- a/fs/ocfs2/ocfs2.h +++ b/fs/ocfs2/ocfs2.h @@ -40,7 +40,8 @@ #include "cluster/heartbeat.h" #include "cluster/tcp.h" -#include "dlm/dlmapi.h" +/* For union ocfs2_dlm_lksb */ +#include "stackglue.h" #include "ocfs2_fs.h" #include "ocfs2_lockid.h" @@ -120,7 +121,7 @@ struct ocfs2_lock_res { int l_level; unsigned int l_ro_holders; unsigned int l_ex_holders; - struct dlm_lockstatus l_lksb; + union ocfs2_dlm_lksb l_lksb; /* used from AST/BAST funcs. */ enum ocfs2_ast_action l_action; diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c index 0aec2fcf217..eb88854cb97 100644 --- a/fs/ocfs2/stackglue.c +++ b/fs/ocfs2/stackglue.c @@ -199,7 +199,7 @@ static void o2dlm_unlock_ast_wrapper(void *astarg, enum dlm_status status) int ocfs2_dlm_lock(struct dlm_ctxt *dlm, int mode, - struct dlm_lockstatus *lksb, + union ocfs2_dlm_lksb *lksb, u32 flags, void *name, unsigned int namelen, @@ -212,15 +212,16 @@ int ocfs2_dlm_lock(struct dlm_ctxt *dlm, BUG_ON(lproto == NULL); - status = dlmlock(dlm, o2dlm_mode, lksb, o2dlm_flags, name, namelen, - o2dlm_lock_ast_wrapper, astarg, - o2dlm_blocking_ast_wrapper); + status = dlmlock(dlm, o2dlm_mode, &lksb->lksb_o2dlm, o2dlm_flags, + name, namelen, + o2dlm_lock_ast_wrapper, astarg, + o2dlm_blocking_ast_wrapper); ret = dlm_status_to_errno(status); return ret; } int ocfs2_dlm_unlock(struct dlm_ctxt *dlm, - struct dlm_lockstatus *lksb, + union ocfs2_dlm_lksb *lksb, u32 flags, void *astarg) { @@ -230,12 +231,26 @@ int ocfs2_dlm_unlock(struct dlm_ctxt *dlm, BUG_ON(lproto == NULL); - status = dlmunlock(dlm, lksb, o2dlm_flags, - o2dlm_unlock_ast_wrapper, astarg); + status = dlmunlock(dlm, &lksb->lksb_o2dlm, o2dlm_flags, + o2dlm_unlock_ast_wrapper, astarg); ret = dlm_status_to_errno(status); return ret; } +int ocfs2_dlm_lock_status(union ocfs2_dlm_lksb *lksb) +{ + return dlm_status_to_errno(lksb->lksb_o2dlm.status); +} + +/* + * Why don't we cast to ocfs2_meta_lvb? The "clean" answer is that we + * don't cast at the glue level. The real answer is that the header + * ordering is nigh impossible. + */ +void *ocfs2_dlm_lvb(union ocfs2_dlm_lksb *lksb) +{ + return (void *)(lksb->lksb_o2dlm.lvb); +} void o2cb_get_stack(struct ocfs2_locking_protocol *proto) { diff --git a/fs/ocfs2/stackglue.h b/fs/ocfs2/stackglue.h index 8ebcfba62c7..3c91e241892 100644 --- a/fs/ocfs2/stackglue.h +++ b/fs/ocfs2/stackglue.h @@ -40,18 +40,25 @@ struct ocfs2_locking_protocol { void (*lp_unlock_ast)(void *astarg, int error); }; +union ocfs2_dlm_lksb { + struct dlm_lockstatus lksb_o2dlm; +}; + int ocfs2_dlm_lock(struct dlm_ctxt *dlm, int mode, - struct dlm_lockstatus *lksb, + union ocfs2_dlm_lksb *lksb, u32 flags, void *name, unsigned int namelen, void *astarg); int ocfs2_dlm_unlock(struct dlm_ctxt *dlm, - struct dlm_lockstatus *lksb, + union ocfs2_dlm_lksb *lksb, u32 flags, void *astarg); +int ocfs2_dlm_lock_status(union ocfs2_dlm_lksb *lksb); +void *ocfs2_dlm_lvb(union ocfs2_dlm_lksb *lksb); + void o2cb_get_stack(struct ocfs2_locking_protocol *proto); void o2cb_put_stack(void); -- cgit v1.2.3 From 4670c46ded9a18268d1265417ff4ac72145a7917 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Fri, 1 Feb 2008 14:39:35 -0800 Subject: ocfs2: Introduce the new ocfs2_cluster_connect/disconnect() API. This step introduces a cluster stack agnostic API for initializing and exiting. fs/ocfs2/dlmglue.c no longer uses o2cb/o2dlm knowledge to connect to the stack. It is all handled in stackglue.c. heartbeat.c no longer needs to know how it gets called. ocfs2_do_node_down() is now a clean recovery trigger. The big gotcha is the ordering of initializations and de-initializations done underneath ocfs2_cluster_connect(). ocfs2_dlm_init() used to do all o2dlm initialization in one block. Thus, the o2dlm functionality of ocfs2_cluster_connect() is very straightforward. ocfs2_dlm_shutdown(), however, did a few things between de-registration of the eviction callback and actually shutting down the domain. Now de-registration and shutdown of the domain are wrapped within the single ocfs2_cluster_disconnect() call. I've checked the code paths to make sure we can safely tear down things in ocfs2_dlm_shutdown() before calling ocfs2_cluster_disconnect(). The filesystem has already set itself to ignore the callback. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlmglue.c | 97 +++++++++++++++++++------------------- fs/ocfs2/dlmglue.h | 1 - fs/ocfs2/heartbeat.c | 40 ++++------------ fs/ocfs2/heartbeat.h | 2 +- fs/ocfs2/ocfs2.h | 4 +- fs/ocfs2/stackglue.c | 131 ++++++++++++++++++++++++++++++++++++++++++++++++--- fs/ocfs2/stackglue.h | 35 +++++++++++++- fs/ocfs2/super.c | 13 ++--- 8 files changed, 221 insertions(+), 102 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index 459037653e5..6652a48cbe0 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -27,7 +27,6 @@ #include #include #include -#include #include #include #include @@ -259,31 +258,6 @@ static struct ocfs2_lock_res_ops ocfs2_flock_lops = { .flags = 0, }; -/* - * This is the filesystem locking protocol version. - * - * Whenever the filesystem does new things with locks (adds or removes a - * lock, orders them differently, does different things underneath a lock), - * the version must be changed. The protocol is negotiated when joining - * the dlm domain. A node may join the domain if its major version is - * identical to all other nodes and its minor version is greater than - * or equal to all other nodes. When its minor version is greater than - * the other nodes, it will run at the minor version specified by the - * other nodes. - * - * If a locking change is made that will not be compatible with older - * versions, the major number must be increased and the minor version set - * to zero. If a change merely adds a behavior that can be disabled when - * speaking to older versions, the minor version must be increased. If a - * change adds a fully backwards compatible change (eg, LVB changes that - * are just ignored by older versions), the version does not need to be - * updated. - */ -const struct dlm_protocol_version ocfs2_locking_protocol = { - .pv_major = OCFS2_LOCKING_PROTOCOL_MAJOR, - .pv_minor = OCFS2_LOCKING_PROTOCOL_MINOR, -}; - static inline int ocfs2_is_inode_lock(struct ocfs2_lock_res *lockres) { return lockres->l_type == OCFS2_LOCK_TYPE_META || @@ -886,7 +860,7 @@ static int ocfs2_lock_create(struct ocfs2_super *osb, lockres_or_flags(lockres, OCFS2_LOCK_BUSY); spin_unlock_irqrestore(&lockres->l_lock, flags); - ret = ocfs2_dlm_lock(osb->dlm, + ret = ocfs2_dlm_lock(osb->cconn, level, &lockres->l_lksb, dlm_flags, @@ -1085,7 +1059,7 @@ again: lockres->l_name, lockres->l_level, level); /* call dlm_lock to upgrade lock now */ - ret = ocfs2_dlm_lock(osb->dlm, + ret = ocfs2_dlm_lock(osb->cconn, level, &lockres->l_lksb, lkm_flags, @@ -1492,7 +1466,7 @@ int ocfs2_file_lock(struct file *file, int ex, int trylock) lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_BUSY, 0); spin_unlock_irqrestore(&lockres->l_lock, flags); - ret = ocfs2_dlm_lock(osb->dlm, level, &lockres->l_lksb, lkm_flags, + ret = ocfs2_dlm_lock(osb->cconn, level, &lockres->l_lksb, lkm_flags, lockres->l_name, OCFS2_LOCK_ID_MAX_LEN - 1, lockres); if (ret) { @@ -2485,8 +2459,7 @@ static void ocfs2_dlm_shutdown_debug(struct ocfs2_super *osb) int ocfs2_dlm_init(struct ocfs2_super *osb) { int status = 0; - u32 dlm_key; - struct dlm_ctxt *dlm = NULL; + struct ocfs2_cluster_connection *conn = NULL; mlog_entry_void(); @@ -2508,26 +2481,21 @@ int ocfs2_dlm_init(struct ocfs2_super *osb) goto bail; } - /* used by the dlm code to make message headers unique, each - * node in this domain must agree on this. */ - dlm_key = crc32_le(0, osb->uuid_str, strlen(osb->uuid_str)); - /* for now, uuid == domain */ - dlm = dlm_register_domain(osb->uuid_str, dlm_key, - &osb->osb_locking_proto); - if (IS_ERR(dlm)) { - status = PTR_ERR(dlm); + status = ocfs2_cluster_connect(osb->uuid_str, + strlen(osb->uuid_str), + ocfs2_do_node_down, osb, + &conn); + if (status) { mlog_errno(status); goto bail; } - dlm_register_eviction_cb(dlm, &osb->osb_eviction_cb); - local: ocfs2_super_lock_res_init(&osb->osb_super_lockres, osb); ocfs2_rename_lock_res_init(&osb->osb_rename_lockres, osb); - osb->dlm = dlm; + osb->cconn = conn; status = 0; bail: @@ -2545,10 +2513,14 @@ void ocfs2_dlm_shutdown(struct ocfs2_super *osb) { mlog_entry_void(); - dlm_unregister_eviction_cb(&osb->osb_eviction_cb); - ocfs2_drop_osb_locks(osb); + /* + * Now that we have dropped all locks and ocfs2_dismount_volume() + * has disabled recovery, the DLM won't be talking to us. It's + * safe to tear things down before disconnecting the cluster. + */ + if (osb->dc_task) { kthread_stop(osb->dc_task); osb->dc_task = NULL; @@ -2557,8 +2529,8 @@ void ocfs2_dlm_shutdown(struct ocfs2_super *osb) ocfs2_lock_res_free(&osb->osb_super_lockres); ocfs2_lock_res_free(&osb->osb_rename_lockres); - dlm_unregister_domain(osb->dlm); - osb->dlm = NULL; + ocfs2_cluster_disconnect(osb->cconn); + osb->cconn = NULL; ocfs2_dlm_shutdown_debug(osb); @@ -2689,7 +2661,7 @@ static int ocfs2_drop_lock(struct ocfs2_super *osb, mlog(0, "lock %s\n", lockres->l_name); - ret = ocfs2_dlm_unlock(osb->dlm, &lockres->l_lksb, lkm_flags, + ret = ocfs2_dlm_unlock(osb->cconn, &lockres->l_lksb, lkm_flags, lockres); if (ret) { ocfs2_log_dlm_error("ocfs2_dlm_unlock", ret, lockres); @@ -2823,7 +2795,7 @@ static int ocfs2_downconvert_lock(struct ocfs2_super *osb, if (lvb) dlm_flags |= DLM_LKF_VALBLK; - ret = ocfs2_dlm_lock(osb->dlm, + ret = ocfs2_dlm_lock(osb->cconn, new_level, &lockres->l_lksb, dlm_flags, @@ -2882,7 +2854,7 @@ static int ocfs2_cancel_convert(struct ocfs2_super *osb, mlog_entry_void(); mlog(0, "lock %s\n", lockres->l_name); - ret = ocfs2_dlm_unlock(osb->dlm, &lockres->l_lksb, + ret = ocfs2_dlm_unlock(osb->cconn, &lockres->l_lksb, DLM_LKF_CANCEL, lockres); if (ret) { ocfs2_log_dlm_error("ocfs2_dlm_unlock", ret, lockres); @@ -3193,7 +3165,34 @@ static int ocfs2_dentry_convert_worker(struct ocfs2_lock_res *lockres, return UNBLOCK_CONTINUE_POST; } +/* + * This is the filesystem locking protocol. It provides the lock handling + * hooks for the underlying DLM. It has a maximum version number. + * The version number allows interoperability with systems running at + * the same major number and an equal or smaller minor number. + * + * Whenever the filesystem does new things with locks (adds or removes a + * lock, orders them differently, does different things underneath a lock), + * the version must be changed. The protocol is negotiated when joining + * the dlm domain. A node may join the domain if its major version is + * identical to all other nodes and its minor version is greater than + * or equal to all other nodes. When its minor version is greater than + * the other nodes, it will run at the minor version specified by the + * other nodes. + * + * If a locking change is made that will not be compatible with older + * versions, the major number must be increased and the minor version set + * to zero. If a change merely adds a behavior that can be disabled when + * speaking to older versions, the minor version must be increased. If a + * change adds a fully backwards compatible change (eg, LVB changes that + * are just ignored by older versions), the version does not need to be + * updated. + */ static struct ocfs2_locking_protocol lproto = { + .lp_max_version = { + .pv_major = OCFS2_LOCKING_PROTOCOL_MAJOR, + .pv_minor = OCFS2_LOCKING_PROTOCOL_MINOR, + }, .lp_lock_ast = ocfs2_locking_ast, .lp_blocking_ast = ocfs2_blocking_ast, .lp_unlock_ast = ocfs2_unlock_ast, diff --git a/fs/ocfs2/dlmglue.h b/fs/ocfs2/dlmglue.h index 32380439401..2d0a8a03c43 100644 --- a/fs/ocfs2/dlmglue.h +++ b/fs/ocfs2/dlmglue.h @@ -117,5 +117,4 @@ void ocfs2_put_dlm_debug(struct ocfs2_dlm_debug *dlm_debug); void dlmglue_init_stack(void); void dlmglue_exit_stack(void); -extern const struct dlm_protocol_version ocfs2_locking_protocol; #endif /* DLMGLUE_H */ diff --git a/fs/ocfs2/heartbeat.c b/fs/ocfs2/heartbeat.c index 80de2397c16..dcac1a48728 100644 --- a/fs/ocfs2/heartbeat.c +++ b/fs/ocfs2/heartbeat.c @@ -30,8 +30,6 @@ #include #include -#include - #define MLOG_MASK_PREFIX ML_SUPER #include @@ -64,19 +62,20 @@ void ocfs2_init_node_maps(struct ocfs2_super *osb) ocfs2_node_map_init(&osb->osb_recovering_orphan_dirs); } -static void ocfs2_do_node_down(int node_num, - struct ocfs2_super *osb) +void ocfs2_do_node_down(int node_num, void *data) { + struct ocfs2_super *osb = data; + BUG_ON(osb->node_num == node_num); mlog(0, "ocfs2: node down event for %d\n", node_num); - if (!osb->dlm) { + if (!osb->cconn) { /* - * No DLM means we're not even ready to participate yet. - * We check the slots after the DLM comes up, so we will - * notice the node death then. We can safely ignore it - * here. + * No cluster connection means we're not even ready to + * participate yet. We check the slots after the cluster + * comes up, so we will notice the node death then. We + * can safely ignore it here. */ return; } @@ -84,29 +83,6 @@ static void ocfs2_do_node_down(int node_num, ocfs2_recovery_thread(osb, node_num); } -/* Called from the dlm when it's about to evict a node. We may also - * get a heartbeat callback later. */ -static void ocfs2_dlm_eviction_cb(int node_num, - void *data) -{ - struct ocfs2_super *osb = (struct ocfs2_super *) data; - struct super_block *sb = osb->sb; - - mlog(ML_NOTICE, "device (%u,%u): dlm has evicted node %d\n", - MAJOR(sb->s_dev), MINOR(sb->s_dev), node_num); - - ocfs2_do_node_down(node_num, osb); -} - -void ocfs2_setup_hb_callbacks(struct ocfs2_super *osb) -{ - /* Not exactly a heartbeat callback, but leads to essentially - * the same path so we set it up here. */ - dlm_setup_eviction_cb(&osb->osb_eviction_cb, - ocfs2_dlm_eviction_cb, - osb); -} - void ocfs2_stop_heartbeat(struct ocfs2_super *osb) { int ret; diff --git a/fs/ocfs2/heartbeat.h b/fs/ocfs2/heartbeat.h index 98d8ffc995b..38e2450bf14 100644 --- a/fs/ocfs2/heartbeat.h +++ b/fs/ocfs2/heartbeat.h @@ -28,7 +28,7 @@ void ocfs2_init_node_maps(struct ocfs2_super *osb); -void ocfs2_setup_hb_callbacks(struct ocfs2_super *osb); +void ocfs2_do_node_down(int node_num, void *data); void ocfs2_stop_heartbeat(struct ocfs2_super *osb); /* node map functions - used to keep track of mounted and in-recovery diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h index 6d7c6d2d0c2..664e4fe3eab 100644 --- a/fs/ocfs2/ocfs2.h +++ b/fs/ocfs2/ocfs2.h @@ -248,12 +248,10 @@ struct ocfs2_super struct ocfs2_alloc_stats alloc_stats; char dev_str[20]; /* "major,minor" of the device */ - struct dlm_ctxt *dlm; + struct ocfs2_cluster_connection *cconn; struct ocfs2_lock_res osb_super_lockres; struct ocfs2_lock_res osb_rename_lockres; - struct dlm_eviction_cb osb_eviction_cb; struct ocfs2_dlm_debug *osb_dlm_debug; - struct dlm_protocol_version osb_locking_proto; struct dentry *osb_debug_root; diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c index eb88854cb97..f6f309a0834 100644 --- a/fs/ocfs2/stackglue.c +++ b/fs/ocfs2/stackglue.c @@ -18,11 +18,21 @@ * General Public License for more details. */ +#include +#include + +/* Needed for AOP_TRUNCATED_PAGE in mlog_errno() */ +#include + #include "cluster/masklog.h" #include "stackglue.h" static struct ocfs2_locking_protocol *lproto; +struct o2dlm_private { + struct dlm_eviction_cb op_eviction_cb; +}; + /* These should be identical */ #if (DLM_LOCK_IV != LKM_IVMODE) # error Lock modes do not match @@ -197,7 +207,7 @@ static void o2dlm_unlock_ast_wrapper(void *astarg, enum dlm_status status) lproto->lp_unlock_ast(astarg, error); } -int ocfs2_dlm_lock(struct dlm_ctxt *dlm, +int ocfs2_dlm_lock(struct ocfs2_cluster_connection *conn, int mode, union ocfs2_dlm_lksb *lksb, u32 flags, @@ -212,15 +222,15 @@ int ocfs2_dlm_lock(struct dlm_ctxt *dlm, BUG_ON(lproto == NULL); - status = dlmlock(dlm, o2dlm_mode, &lksb->lksb_o2dlm, o2dlm_flags, - name, namelen, + status = dlmlock(conn->cc_lockspace, o2dlm_mode, &lksb->lksb_o2dlm, + o2dlm_flags, name, namelen, o2dlm_lock_ast_wrapper, astarg, o2dlm_blocking_ast_wrapper); ret = dlm_status_to_errno(status); return ret; } -int ocfs2_dlm_unlock(struct dlm_ctxt *dlm, +int ocfs2_dlm_unlock(struct ocfs2_cluster_connection *conn, union ocfs2_dlm_lksb *lksb, u32 flags, void *astarg) @@ -231,8 +241,8 @@ int ocfs2_dlm_unlock(struct dlm_ctxt *dlm, BUG_ON(lproto == NULL); - status = dlmunlock(dlm, &lksb->lksb_o2dlm, o2dlm_flags, - o2dlm_unlock_ast_wrapper, astarg); + status = dlmunlock(conn->cc_lockspace, &lksb->lksb_o2dlm, + o2dlm_flags, o2dlm_unlock_ast_wrapper, astarg); ret = dlm_status_to_errno(status); return ret; } @@ -252,6 +262,115 @@ void *ocfs2_dlm_lvb(union ocfs2_dlm_lksb *lksb) return (void *)(lksb->lksb_o2dlm.lvb); } +/* + * Called from the dlm when it's about to evict a node. This is how the + * classic stack signals node death. + */ +static void o2dlm_eviction_cb(int node_num, void *data) +{ + struct ocfs2_cluster_connection *conn = data; + + mlog(ML_NOTICE, "o2dlm has evicted node %d from group %.*s\n", + node_num, conn->cc_namelen, conn->cc_name); + + conn->cc_recovery_handler(node_num, conn->cc_recovery_data); +} + +int ocfs2_cluster_connect(const char *group, + int grouplen, + void (*recovery_handler)(int node_num, + void *recovery_data), + void *recovery_data, + struct ocfs2_cluster_connection **conn) +{ + int rc = 0; + struct ocfs2_cluster_connection *new_conn; + u32 dlm_key; + struct dlm_ctxt *dlm; + struct o2dlm_private *priv; + struct dlm_protocol_version dlm_version; + + BUG_ON(group == NULL); + BUG_ON(conn == NULL); + BUG_ON(recovery_handler == NULL); + + if (grouplen > GROUP_NAME_MAX) { + rc = -EINVAL; + goto out; + } + + new_conn = kzalloc(sizeof(struct ocfs2_cluster_connection), + GFP_KERNEL); + if (!new_conn) { + rc = -ENOMEM; + goto out; + } + + memcpy(new_conn->cc_name, group, grouplen); + new_conn->cc_namelen = grouplen; + new_conn->cc_recovery_handler = recovery_handler; + new_conn->cc_recovery_data = recovery_data; + + /* Start the new connection at our maximum compatibility level */ + new_conn->cc_version = lproto->lp_max_version; + + priv = kzalloc(sizeof(struct o2dlm_private), GFP_KERNEL); + if (!priv) { + rc = -ENOMEM; + goto out_free; + } + + /* This just fills the structure in. It is safe to use new_conn. */ + dlm_setup_eviction_cb(&priv->op_eviction_cb, o2dlm_eviction_cb, + new_conn); + + new_conn->cc_private = priv; + + /* used by the dlm code to make message headers unique, each + * node in this domain must agree on this. */ + dlm_key = crc32_le(0, group, grouplen); + dlm_version.pv_major = new_conn->cc_version.pv_major; + dlm_version.pv_minor = new_conn->cc_version.pv_minor; + + dlm = dlm_register_domain(group, dlm_key, &dlm_version); + if (IS_ERR(dlm)) { + rc = PTR_ERR(dlm); + mlog_errno(rc); + goto out_free; + } + + new_conn->cc_version.pv_major = dlm_version.pv_major; + new_conn->cc_version.pv_minor = dlm_version.pv_minor; + new_conn->cc_lockspace = dlm; + + dlm_register_eviction_cb(dlm, &priv->op_eviction_cb); + + *conn = new_conn; + +out_free: + if (rc) { + kfree(new_conn->cc_private); + kfree(new_conn); + } + +out: + return rc; +} + +int ocfs2_cluster_disconnect(struct ocfs2_cluster_connection *conn) +{ + struct dlm_ctxt *dlm = conn->cc_lockspace; + struct o2dlm_private *priv = conn->cc_private; + + dlm_unregister_eviction_cb(&priv->op_eviction_cb); + dlm_unregister_domain(dlm); + + kfree(priv); + kfree(conn); + + return 0; +} + void o2cb_get_stack(struct ocfs2_locking_protocol *proto) { BUG_ON(proto == NULL); diff --git a/fs/ocfs2/stackglue.h b/fs/ocfs2/stackglue.h index 3c91e241892..3900b5c3933 100644 --- a/fs/ocfs2/stackglue.h +++ b/fs/ocfs2/stackglue.h @@ -32,9 +32,22 @@ */ #define DLM_LKF_LOCAL 0x00100000 +/* + * This shadows DLM_LOCKSPACE_LEN in fs/dlm/dlm_internal.h. That probably + * wants to be in a public header. + */ +#define GROUP_NAME_MAX 64 + + #include "dlm/dlmapi.h" +struct ocfs2_protocol_version { + u8 pv_major; + u8 pv_minor; +}; + struct ocfs2_locking_protocol { + struct ocfs2_protocol_version lp_max_version; void (*lp_lock_ast)(void *astarg); void (*lp_blocking_ast)(void *astarg, int level); void (*lp_unlock_ast)(void *astarg, int error); @@ -44,14 +57,32 @@ union ocfs2_dlm_lksb { struct dlm_lockstatus lksb_o2dlm; }; -int ocfs2_dlm_lock(struct dlm_ctxt *dlm, +struct ocfs2_cluster_connection { + char cc_name[GROUP_NAME_MAX]; + int cc_namelen; + struct ocfs2_protocol_version cc_version; + void (*cc_recovery_handler)(int node_num, void *recovery_data); + void *cc_recovery_data; + void *cc_lockspace; + void *cc_private; +}; + +int ocfs2_cluster_connect(const char *group, + int grouplen, + void (*recovery_handler)(int node_num, + void *recovery_data), + void *recovery_data, + struct ocfs2_cluster_connection **conn); +int ocfs2_cluster_disconnect(struct ocfs2_cluster_connection *conn); + +int ocfs2_dlm_lock(struct ocfs2_cluster_connection *conn, int mode, union ocfs2_dlm_lksb *lksb, u32 flags, void *name, unsigned int namelen, void *astarg); -int ocfs2_dlm_unlock(struct dlm_ctxt *dlm, +int ocfs2_dlm_unlock(struct ocfs2_cluster_connection *conn, union ocfs2_dlm_lksb *lksb, u32 flags, void *astarg); diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index c8675464e29..0ee49757467 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -1251,9 +1251,9 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err) ocfs2_sync_blockdev(sb); - /* No dlm means we've failed during mount, so skip all the - * steps which depended on that to complete. */ - if (osb->dlm) { + /* No cluster connection means we've failed during mount, so skip + * all the steps which depended on that to complete. */ + if (osb->cconn) { tmp = ocfs2_super_lock(osb, 1); if (tmp < 0) { mlog_errno(tmp); @@ -1264,12 +1264,12 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err) if (osb->slot_num != OCFS2_INVALID_SLOT) ocfs2_put_slot(osb); - if (osb->dlm) + if (osb->cconn) ocfs2_super_unlock(osb, 1); ocfs2_release_system_inodes(osb); - if (osb->dlm) + if (osb->cconn) ocfs2_dlm_shutdown(osb); debugfs_remove(osb->osb_debug_root); @@ -1341,7 +1341,6 @@ static int ocfs2_initialize_super(struct super_block *sb, sb->s_fs_info = osb; sb->s_op = &ocfs2_sops; sb->s_export_op = &ocfs2_export_ops; - osb->osb_locking_proto = ocfs2_locking_protocol; sb->s_time_gran = 1; sb->s_flags |= MS_NOATIME; /* this is needed to support O_LARGEFILE */ @@ -1391,8 +1390,6 @@ static int ocfs2_initialize_super(struct super_block *sb, osb->local_alloc_state = OCFS2_LA_UNUSED; osb->local_alloc_bh = NULL; - ocfs2_setup_hb_callbacks(osb); - init_waitqueue_head(&osb->osb_mount_event); osb->vol_label = kmalloc(OCFS2_MAX_VOL_LABEL_LEN, GFP_KERNEL); -- cgit v1.2.3 From 19fdb624dc8ccb663f6e48b3a3a3fa4e4e567fc1 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Wed, 30 Jan 2008 15:38:24 -0800 Subject: ocfs2: Abstract out node number queries. ocfs2 asks the cluster stack for the local node's node number for two reasons; to fill the slot map and to print it. While the slot map isn't necessary for userspace cluster stacks, the printing is very nice for debugging. Thus we add ocfs2_cluster_this_node() as a generic API to get this value. It is anticipated that the slot map will not be used under a userspace cluster stack, so validity checks of the node num only need to exist in the slot map code. Otherwise, it just gets used and printed as an opaque value. [ Fixed up some "int" versus "unsigned int" issues and made osb->node_num truly opaque. --Mark ] Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/ocfs2.h | 2 +- fs/ocfs2/slot_map.c | 2 -- fs/ocfs2/stackglue.c | 17 +++++++++++++++++ fs/ocfs2/stackglue.h | 1 + fs/ocfs2/super.c | 22 +++++++++++----------- 5 files changed, 30 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h index 664e4fe3eab..7006abaffd4 100644 --- a/fs/ocfs2/ocfs2.h +++ b/fs/ocfs2/ocfs2.h @@ -218,7 +218,7 @@ struct ocfs2_super unsigned int s_atime_quantum; unsigned int max_slots; - s16 node_num; + unsigned int node_num; int slot_num; int preferred_slot; int s_sectsize_bits; diff --git a/fs/ocfs2/slot_map.c b/fs/ocfs2/slot_map.c index 63fb1b26d96..bb5ff8939bf 100644 --- a/fs/ocfs2/slot_map.c +++ b/fs/ocfs2/slot_map.c @@ -73,8 +73,6 @@ static void ocfs2_set_slot(struct ocfs2_slot_info *si, int slot_num, unsigned int node_num) { BUG_ON((slot_num < 0) || (slot_num >= si->si_num_slots)); - BUG_ON((node_num == O2NM_INVALID_NODE_NUM) || - (node_num >= O2NM_MAX_NODES)); si->si_slots[slot_num].sl_valid = 1; si->si_slots[slot_num].sl_node_num = node_num; diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c index f6f309a0834..814686356cc 100644 --- a/fs/ocfs2/stackglue.c +++ b/fs/ocfs2/stackglue.c @@ -25,6 +25,8 @@ #include #include "cluster/masklog.h" +#include "cluster/nodemanager.h" + #include "stackglue.h" static struct ocfs2_locking_protocol *lproto; @@ -371,6 +373,21 @@ int ocfs2_cluster_disconnect(struct ocfs2_cluster_connection *conn) return 0; } +int ocfs2_cluster_this_node(unsigned int *node) +{ + int node_num; + + node_num = o2nm_this_node(); + if (node_num == O2NM_INVALID_NODE_NUM) + return -ENOENT; + + if (node_num >= O2NM_MAX_NODES) + return -EOVERFLOW; + + *node = node_num; + return 0; +} + void o2cb_get_stack(struct ocfs2_locking_protocol *proto) { BUG_ON(proto == NULL); diff --git a/fs/ocfs2/stackglue.h b/fs/ocfs2/stackglue.h index 3900b5c3933..ccb03991b51 100644 --- a/fs/ocfs2/stackglue.h +++ b/fs/ocfs2/stackglue.h @@ -74,6 +74,7 @@ int ocfs2_cluster_connect(const char *group, void *recovery_data, struct ocfs2_cluster_connection **conn); int ocfs2_cluster_disconnect(struct ocfs2_cluster_connection *conn); +int ocfs2_cluster_this_node(unsigned int *node); int ocfs2_dlm_lock(struct ocfs2_cluster_connection *conn, int mode, diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index 0ee49757467..d3c4d323fab 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -694,7 +694,7 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent) if (ocfs2_mount_local(osb)) snprintf(nodestr, sizeof(nodestr), "local"); else - snprintf(nodestr, sizeof(nodestr), "%d", osb->node_num); + snprintf(nodestr, sizeof(nodestr), "%u", osb->node_num); printk(KERN_INFO "ocfs2: Mounting device (%s) on (node %s, slot %d) " "with %s data mode.\n", @@ -1145,16 +1145,17 @@ static int ocfs2_fill_local_node_info(struct ocfs2_super *osb) * desirable. */ if (ocfs2_mount_local(osb)) osb->node_num = 0; - else - osb->node_num = o2nm_this_node(); - - if (osb->node_num == O2NM_MAX_NODES) { - mlog(ML_ERROR, "could not find this host's node number\n"); - status = -ENOENT; - goto bail; + else { + status = ocfs2_cluster_this_node(&osb->node_num); + if (status < 0) { + mlog_errno(status); + mlog(ML_ERROR, + "could not find this host's node number\n"); + goto bail; + } } - mlog(0, "I am node %d\n", osb->node_num); + mlog(0, "I am node %u\n", osb->node_num); status = 0; bail: @@ -1282,7 +1283,7 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err) if (ocfs2_mount_local(osb)) snprintf(nodestr, sizeof(nodestr), "local"); else - snprintf(nodestr, sizeof(nodestr), "%d", osb->node_num); + snprintf(nodestr, sizeof(nodestr), "%u", osb->node_num); printk(KERN_INFO "ocfs2: Unmounting device (%s) on (node %s)\n", osb->dev_str, nodestr); @@ -1384,7 +1385,6 @@ static int ocfs2_initialize_super(struct super_block *sb, osb->s_atime_quantum = OCFS2_DEFAULT_ATIME_QUANTUM; - osb->node_num = O2NM_INVALID_NODE_NUM; osb->slot_num = OCFS2_INVALID_SLOT; osb->local_alloc_state = OCFS2_LA_UNUSED; -- cgit v1.2.3 From 6953b4c008628b945bfe0cee97f6e78a98773859 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Tue, 29 Jan 2008 16:59:56 -0800 Subject: ocfs2: Move o2hb functionality into the stack glue. The last bit of classic stack used directly in ocfs2 code is o2hb. Specifically, the check for heartbeat during mount and the call to ocfs2_hb_ctl during unmount. We create an extra API, ocfs2_cluster_hangup(), to encapsulate the call to ocfs2_hb_ctl. Other stacks will just leave hangup() empty. The check for heartbeat is moved into ocfs2_cluster_connect(). It will be matched by a similar check for other stacks. With this change, only stackglue.c includes cluster/ headers. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlmglue.c | 4 ---- fs/ocfs2/heartbeat.c | 33 --------------------------------- fs/ocfs2/heartbeat.h | 1 - fs/ocfs2/ioctl.c | 1 + fs/ocfs2/ocfs2.h | 4 ---- fs/ocfs2/stackglue.c | 50 ++++++++++++++++++++++++++++++++++++++++++++++++++ fs/ocfs2/stackglue.h | 1 + fs/ocfs2/super.c | 23 ++++++++++------------- 8 files changed, 62 insertions(+), 55 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index 6652a48cbe0..05fd016ba4b 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -32,10 +32,6 @@ #include #include -#include -#include -#include - #define MLOG_MASK_PREFIX ML_DLM_GLUE #include diff --git a/fs/ocfs2/heartbeat.c b/fs/ocfs2/heartbeat.c index dcac1a48728..c6e7213db86 100644 --- a/fs/ocfs2/heartbeat.c +++ b/fs/ocfs2/heartbeat.c @@ -28,7 +28,6 @@ #include #include #include -#include #define MLOG_MASK_PREFIX ML_SUPER #include @@ -83,38 +82,6 @@ void ocfs2_do_node_down(int node_num, void *data) ocfs2_recovery_thread(osb, node_num); } -void ocfs2_stop_heartbeat(struct ocfs2_super *osb) -{ - int ret; - char *argv[5], *envp[3]; - - if (ocfs2_mount_local(osb)) - return; - - if (!osb->uuid_str) { - /* This can happen if we don't get far enough in mount... */ - mlog(0, "No UUID with which to stop heartbeat!\n\n"); - return; - } - - argv[0] = (char *)o2nm_get_hb_ctl_path(); - argv[1] = "-K"; - argv[2] = "-u"; - argv[3] = osb->uuid_str; - argv[4] = NULL; - - mlog(0, "Run: %s %s %s %s\n", argv[0], argv[1], argv[2], argv[3]); - - /* minimal command environment taken from cpu_run_sbin_hotplug */ - envp[0] = "HOME=/"; - envp[1] = "PATH=/sbin:/bin:/usr/sbin:/usr/bin"; - envp[2] = NULL; - - ret = call_usermodehelper(argv[0], argv, envp, UMH_WAIT_PROC); - if (ret < 0) - mlog_errno(ret); -} - static inline void __ocfs2_node_map_set_bit(struct ocfs2_node_map *map, int bit) { diff --git a/fs/ocfs2/heartbeat.h b/fs/ocfs2/heartbeat.h index 38e2450bf14..74b9c5dda28 100644 --- a/fs/ocfs2/heartbeat.h +++ b/fs/ocfs2/heartbeat.h @@ -29,7 +29,6 @@ void ocfs2_init_node_maps(struct ocfs2_super *osb); void ocfs2_do_node_down(int node_num, void *data); -void ocfs2_stop_heartbeat(struct ocfs2_super *osb); /* node map functions - used to keep track of mounted and in-recovery * nodes. */ diff --git a/fs/ocfs2/ioctl.c b/fs/ocfs2/ioctl.c index 5177fba5162..ab1c2167d7f 100644 --- a/fs/ocfs2/ioctl.c +++ b/fs/ocfs2/ioctl.c @@ -7,6 +7,7 @@ #include #include +#include #define MLOG_MASK_PREFIX ML_INODE #include diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h index 7006abaffd4..31dc28b4839 100644 --- a/fs/ocfs2/ocfs2.h +++ b/fs/ocfs2/ocfs2.h @@ -36,10 +36,6 @@ #include #include -#include "cluster/nodemanager.h" -#include "cluster/heartbeat.h" -#include "cluster/tcp.h" - /* For union ocfs2_dlm_lksb */ #include "stackglue.h" diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c index 814686356cc..670fa945c21 100644 --- a/fs/ocfs2/stackglue.c +++ b/fs/ocfs2/stackglue.c @@ -20,12 +20,14 @@ #include #include +#include /* Needed for AOP_TRUNCATED_PAGE in mlog_errno() */ #include #include "cluster/masklog.h" #include "cluster/nodemanager.h" +#include "cluster/heartbeat.h" #include "stackglue.h" @@ -301,6 +303,13 @@ int ocfs2_cluster_connect(const char *group, goto out; } + /* for now we only have one cluster/node, make sure we see it + * in the heartbeat universe */ + if (!o2hb_check_local_node_heartbeating()) { + rc = -EINVAL; + goto out; + } + new_conn = kzalloc(sizeof(struct ocfs2_cluster_connection), GFP_KERNEL); if (!new_conn) { @@ -359,6 +368,7 @@ out: return rc; } + int ocfs2_cluster_disconnect(struct ocfs2_cluster_connection *conn) { struct dlm_ctxt *dlm = conn->cc_lockspace; @@ -373,6 +383,46 @@ int ocfs2_cluster_disconnect(struct ocfs2_cluster_connection *conn) return 0; } +static void o2hb_stop(const char *group) +{ + int ret; + char *argv[5], *envp[3]; + + argv[0] = (char *)o2nm_get_hb_ctl_path(); + argv[1] = "-K"; + argv[2] = "-u"; + argv[3] = (char *)group; + argv[4] = NULL; + + mlog(0, "Run: %s %s %s %s\n", argv[0], argv[1], argv[2], argv[3]); + + /* minimal command environment taken from cpu_run_sbin_hotplug */ + envp[0] = "HOME=/"; + envp[1] = "PATH=/sbin:/bin:/usr/sbin:/usr/bin"; + envp[2] = NULL; + + ret = call_usermodehelper(argv[0], argv, envp, UMH_WAIT_PROC); + if (ret < 0) + mlog_errno(ret); +} + +/* + * Hangup is a hack for tools compatibility. Older ocfs2-tools software + * expects the filesystem to call "ocfs2_hb_ctl" during unmount. This + * happens regardless of whether the DLM got started, so we can't do it + * in ocfs2_cluster_disconnect(). We bring the o2hb_stop() function into + * the glue and provide a "hangup" API for super.c to call. + * + * Other stacks will eventually provide a NULL ->hangup() pointer. + */ +void ocfs2_cluster_hangup(const char *group, int grouplen) +{ + BUG_ON(group == NULL); + BUG_ON(group[grouplen] != '\0'); + + o2hb_stop(group); +} + int ocfs2_cluster_this_node(unsigned int *node) { int node_num; diff --git a/fs/ocfs2/stackglue.h b/fs/ocfs2/stackglue.h index ccb03991b51..22af77b9a69 100644 --- a/fs/ocfs2/stackglue.h +++ b/fs/ocfs2/stackglue.h @@ -74,6 +74,7 @@ int ocfs2_cluster_connect(const char *group, void *recovery_data, struct ocfs2_cluster_connection **conn); int ocfs2_cluster_disconnect(struct ocfs2_cluster_connection *conn); +void ocfs2_cluster_hangup(const char *group, int grouplen); int ocfs2_cluster_this_node(unsigned int *node); int ocfs2_dlm_lock(struct ocfs2_cluster_connection *conn, diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index d3c4d323fab..8f536b39ce5 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -40,8 +40,7 @@ #include #include #include - -#include +#include #define MLOG_MASK_PREFIX ML_SUPER #include @@ -579,15 +578,6 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent) goto read_super_error; } - /* for now we only have one cluster/node, make sure we see it - * in the heartbeat universe */ - if (parsed_options.mount_opt & OCFS2_MOUNT_HB_LOCAL) { - if (!o2hb_check_local_node_heartbeating()) { - status = -EINVAL; - goto read_super_error; - } - } - /* probe for superblock */ status = ocfs2_sb_probe(sb, &bh, §or_size); if (status < 0) { @@ -1275,8 +1265,15 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err) debugfs_remove(osb->osb_debug_root); - if (!mnt_err) - ocfs2_stop_heartbeat(osb); + /* + * This is a small hack to move ocfs2_hb_ctl into stackglue. + * If we're dismounting due to mount error, mount.ocfs2 will clean + * up heartbeat. If we're a local mount, there is no heartbeat. + * If we failed before we got a uuid_str yet, we can't stop + * heartbeat. Otherwise, do it. + */ + if (!mnt_err && !ocfs2_mount_local(osb) && osb->uuid_str) + ocfs2_cluster_hangup(osb->uuid_str, strlen(osb->uuid_str)); atomic_set(&osb->vol_state, VOLUME_DISMOUNTED); -- cgit v1.2.3 From 0abd6d1803b01c741430af270026d1d95a103d9c Mon Sep 17 00:00:00 2001 From: Mark Fasheh Date: Tue, 29 Jan 2008 16:59:56 -0800 Subject: ocfs2: Fill node number during cluster stack init It doesn't make sense to query for a node number before connecting to the cluster stack. This should be safe to do because node_num is only just printed, and we're actually only moving the setting of node num a small amount further in the mount process. [ Disconnect when node query fails -- Joel ] Reviewed-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlmglue.c | 13 ++++++++++++- fs/ocfs2/super.c | 33 --------------------------------- 2 files changed, 12 insertions(+), 34 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index 05fd016ba4b..c7653bb343e 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -2459,8 +2459,10 @@ int ocfs2_dlm_init(struct ocfs2_super *osb) mlog_entry_void(); - if (ocfs2_mount_local(osb)) + if (ocfs2_mount_local(osb)) { + osb->node_num = 0; goto local; + } status = ocfs2_dlm_init_debug(osb); if (status < 0) { @@ -2487,6 +2489,15 @@ int ocfs2_dlm_init(struct ocfs2_super *osb) goto bail; } + status = ocfs2_cluster_this_node(&osb->node_num); + if (status < 0) { + mlog_errno(status); + mlog(ML_ERROR, + "could not find this host's node number\n"); + ocfs2_cluster_disconnect(conn); + goto bail; + } + local: ocfs2_super_lock_res_init(&osb->osb_super_lockres, osb); ocfs2_rename_lock_res_init(&osb->osb_rename_lockres, osb); diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index 8f536b39ce5..fa9c46e2eab 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -108,7 +108,6 @@ static int ocfs2_sync_fs(struct super_block *sb, int wait); static int ocfs2_init_global_system_inodes(struct ocfs2_super *osb); static int ocfs2_init_local_system_inodes(struct ocfs2_super *osb); static void ocfs2_release_system_inodes(struct ocfs2_super *osb); -static int ocfs2_fill_local_node_info(struct ocfs2_super *osb); static int ocfs2_check_volume(struct ocfs2_super *osb); static int ocfs2_verify_volume(struct ocfs2_dinode *di, struct buffer_head *bh, @@ -1126,32 +1125,6 @@ static int ocfs2_get_sector(struct super_block *sb, return 0; } -/* ocfs2 1.0 only allows one cluster and node identity per kernel image. */ -static int ocfs2_fill_local_node_info(struct ocfs2_super *osb) -{ - int status; - - /* XXX hold a ref on the node while mounte? easy enough, if - * desirable. */ - if (ocfs2_mount_local(osb)) - osb->node_num = 0; - else { - status = ocfs2_cluster_this_node(&osb->node_num); - if (status < 0) { - mlog_errno(status); - mlog(ML_ERROR, - "could not find this host's node number\n"); - goto bail; - } - } - - mlog(0, "I am node %u\n", osb->node_num); - - status = 0; -bail: - return status; -} - static int ocfs2_mount_volume(struct super_block *sb) { int status = 0; @@ -1163,12 +1136,6 @@ static int ocfs2_mount_volume(struct super_block *sb) if (ocfs2_is_hard_readonly(osb)) goto leave; - status = ocfs2_fill_local_node_info(osb); - if (status < 0) { - mlog_errno(status); - goto leave; - } - status = ocfs2_dlm_init(osb); if (status < 0) { mlog_errno(status); -- cgit v1.2.3 From de551246e7bc5558371c3427889a8db1b8cc60f4 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Fri, 1 Feb 2008 14:45:08 -0800 Subject: ocfs2: Remove CANCELGRANT from the view of dlmglue. o2dlm has the non-standard behavior of providing a cancel callback (unlock_ast) even when the cancel has failed (the locking operation succeeded without canceling). This is called CANCELGRANT after the status code sent to the callback. fs/dlm does not provide this callback, so dlmglue must be changed to live without it. o2dlm_unlock_ast_wrapper() in stackglue now ignores CANCELGRANT calls. Because dlmglue no longer sees CANCELGRANT, ocfs2_unlock_ast() no longer needs to check for it. ocfs2_locking_ast() must catch that a cancel was tried and clear the cancel state. Making these changes opens up a locking race. dlmglue uses the the OCFS2_LOCK_BUSY flag to ensure only one thread is calling the dlm at any one time. But dlmglue must unlock the lockres before calling into the dlm. In the small window of time between unlocking the lockres and calling the dlm, the downconvert thread can try to cancel the lock. The downconvert thread is checking the OCFS2_LOCK_BUSY flag - it doesn't know that ocfs2_dlm_lock() has not yet been called. Because ocfs2_dlm_lock() has not yet been called, the cancel operation will just be a no-op. There's nothing to cancel. With CANCELGRANT, dlmglue uses the CANCELGRANT callback to clear up the cancel state. When it comes around again, it will retry the cancel. Eventually, the first thread will have called into ocfs2_dlm_lock(), and either the lock or the cancel will succeed. The downconvert thread can then do its downconvert. Without CANCELGRANT, there is nothing to clean up the cancellation state. The downconvert thread does not know to retry its operations. More importantly, the original lock may be blocking on the other node that is trying to cancel us. With neither able to make progress, the ast is never called and the cancellation state is never cleaned up that way. dlmglue is deadlocked. The OCFS2_LOCK_PENDING flag is introduced to remedy this window. It is set at the same time OCFS2_LOCK_BUSY is. Thus, the downconvert thread can check whether the lock is cancelable. If not, it just loops around to try again. Once ocfs2_dlm_lock() is called, the thread then clears OCFS2_LOCK_PENDING and wakes the downconvert thread. Now, if the downconvert thread finds the lock BUSY, it can safely try to cancel it. Whether the cancel works or not, the state will be properly set and the lock processing can continue. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlmglue.c | 199 +++++++++++++++++++++++++++++++++++++++++++-------- fs/ocfs2/ocfs2.h | 4 ++ fs/ocfs2/stackglue.c | 40 ++++------- 3 files changed, 188 insertions(+), 55 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index c7653bb343e..295c47f7aba 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -311,12 +311,13 @@ static int ocfs2_inode_lock_update(struct inode *inode, struct buffer_head **bh); static void ocfs2_drop_osb_locks(struct ocfs2_super *osb); static inline int ocfs2_highest_compat_lock_level(int level); -static void ocfs2_prepare_downconvert(struct ocfs2_lock_res *lockres, - int new_level); +static unsigned int ocfs2_prepare_downconvert(struct ocfs2_lock_res *lockres, + int new_level); static int ocfs2_downconvert_lock(struct ocfs2_super *osb, struct ocfs2_lock_res *lockres, int new_level, - int lvb); + int lvb, + unsigned int generation); static int ocfs2_prepare_cancel_convert(struct ocfs2_super *osb, struct ocfs2_lock_res *lockres); static int ocfs2_cancel_convert(struct ocfs2_super *osb, @@ -736,6 +737,113 @@ static int ocfs2_generic_handle_bast(struct ocfs2_lock_res *lockres, return needs_downconvert; } +/* + * OCFS2_LOCK_PENDING and l_pending_gen. + * + * Why does OCFS2_LOCK_PENDING exist? To close a race between setting + * OCFS2_LOCK_BUSY and calling ocfs2_dlm_lock(). See ocfs2_unblock_lock() + * for more details on the race. + * + * OCFS2_LOCK_PENDING closes the race quite nicely. However, it introduces + * a race on itself. In o2dlm, we can get the ast before ocfs2_dlm_lock() + * returns. The ast clears OCFS2_LOCK_BUSY, and must therefore clear + * OCFS2_LOCK_PENDING at the same time. When ocfs2_dlm_lock() returns, + * the caller is going to try to clear PENDING again. If nothing else is + * happening, __lockres_clear_pending() sees PENDING is unset and does + * nothing. + * + * But what if another path (eg downconvert thread) has just started a + * new locking action? The other path has re-set PENDING. Our path + * cannot clear PENDING, because that will re-open the original race + * window. + * + * [Example] + * + * ocfs2_meta_lock() + * ocfs2_cluster_lock() + * set BUSY + * set PENDING + * drop l_lock + * ocfs2_dlm_lock() + * ocfs2_locking_ast() ocfs2_downconvert_thread() + * clear PENDING ocfs2_unblock_lock() + * take_l_lock + * !BUSY + * ocfs2_prepare_downconvert() + * set BUSY + * set PENDING + * drop l_lock + * take l_lock + * clear PENDING + * drop l_lock + * + * ocfs2_dlm_lock() + * + * So as you can see, we now have a window where l_lock is not held, + * PENDING is not set, and ocfs2_dlm_lock() has not been called. + * + * The core problem is that ocfs2_cluster_lock() has cleared the PENDING + * set by ocfs2_prepare_downconvert(). That wasn't nice. + * + * To solve this we introduce l_pending_gen. A call to + * lockres_clear_pending() will only do so when it is passed a generation + * number that matches the lockres. lockres_set_pending() will return the + * current generation number. When ocfs2_cluster_lock() goes to clear + * PENDING, it passes the generation it got from set_pending(). In our + * example above, the generation numbers will *not* match. Thus, + * ocfs2_cluster_lock() will not clear the PENDING set by + * ocfs2_prepare_downconvert(). + */ + +/* Unlocked version for ocfs2_locking_ast() */ +static void __lockres_clear_pending(struct ocfs2_lock_res *lockres, + unsigned int generation, + struct ocfs2_super *osb) +{ + assert_spin_locked(&lockres->l_lock); + + /* + * The ast and locking functions can race us here. The winner + * will clear pending, the loser will not. + */ + if (!(lockres->l_flags & OCFS2_LOCK_PENDING) || + (lockres->l_pending_gen != generation)) + return; + + lockres_clear_flags(lockres, OCFS2_LOCK_PENDING); + lockres->l_pending_gen++; + + /* + * The downconvert thread may have skipped us because we + * were PENDING. Wake it up. + */ + if (lockres->l_flags & OCFS2_LOCK_BLOCKED) + ocfs2_wake_downconvert_thread(osb); +} + +/* Locked version for callers of ocfs2_dlm_lock() */ +static void lockres_clear_pending(struct ocfs2_lock_res *lockres, + unsigned int generation, + struct ocfs2_super *osb) +{ + unsigned long flags; + + spin_lock_irqsave(&lockres->l_lock, flags); + __lockres_clear_pending(lockres, generation, osb); + spin_unlock_irqrestore(&lockres->l_lock, flags); +} + +static unsigned int lockres_set_pending(struct ocfs2_lock_res *lockres) +{ + assert_spin_locked(&lockres->l_lock); + BUG_ON(!(lockres->l_flags & OCFS2_LOCK_BUSY)); + + lockres_or_flags(lockres, OCFS2_LOCK_PENDING); + + return lockres->l_pending_gen; +} + + static void ocfs2_blocking_ast(void *opaque, int level) { struct ocfs2_lock_res *lockres = opaque; @@ -770,6 +878,7 @@ static void ocfs2_blocking_ast(void *opaque, int level) static void ocfs2_locking_ast(void *opaque) { struct ocfs2_lock_res *lockres = opaque; + struct ocfs2_super *osb = ocfs2_get_lockres_osb(lockres); unsigned long flags; spin_lock_irqsave(&lockres->l_lock, flags); @@ -805,6 +914,18 @@ static void ocfs2_locking_ast(void *opaque) * can catch it. */ lockres->l_action = OCFS2_AST_INVALID; + /* Did we try to cancel this lock? Clear that state */ + if (lockres->l_unlock_action == OCFS2_UNLOCK_CANCEL_CONVERT) + lockres->l_unlock_action = OCFS2_UNLOCK_INVALID; + + /* + * We may have beaten the locking functions here. We certainly + * know that dlm_lock() has been called :-) + * Because we can't have two lock calls in flight at once, we + * can use lockres->l_pending_gen. + */ + __lockres_clear_pending(lockres, lockres->l_pending_gen, osb); + wake_up(&lockres->l_event); spin_unlock_irqrestore(&lockres->l_lock, flags); } @@ -838,6 +959,7 @@ static int ocfs2_lock_create(struct ocfs2_super *osb, { int ret = 0; unsigned long flags; + unsigned int gen; mlog_entry_void(); @@ -854,6 +976,7 @@ static int ocfs2_lock_create(struct ocfs2_super *osb, lockres->l_action = OCFS2_AST_ATTACH; lockres->l_requested = level; lockres_or_flags(lockres, OCFS2_LOCK_BUSY); + gen = lockres_set_pending(lockres); spin_unlock_irqrestore(&lockres->l_lock, flags); ret = ocfs2_dlm_lock(osb->cconn, @@ -863,6 +986,7 @@ static int ocfs2_lock_create(struct ocfs2_super *osb, lockres->l_name, OCFS2_LOCK_ID_MAX_LEN - 1, lockres); + lockres_clear_pending(lockres, gen, osb); if (ret) { ocfs2_log_dlm_error("ocfs2_dlm_lock", ret, lockres); ocfs2_recover_from_dlm_error(lockres, 1); @@ -988,6 +1112,7 @@ static int ocfs2_cluster_lock(struct ocfs2_super *osb, int wait, catch_signals = !(osb->s_mount_opt & OCFS2_MOUNT_NOINTR); int ret = 0; /* gcc doesn't realize wait = 1 guarantees ret is set */ unsigned long flags; + unsigned int gen; mlog_entry_void(); @@ -1046,6 +1171,7 @@ again: lockres->l_requested = level; lockres_or_flags(lockres, OCFS2_LOCK_BUSY); + gen = lockres_set_pending(lockres); spin_unlock_irqrestore(&lockres->l_lock, flags); BUG_ON(level == DLM_LOCK_IV); @@ -1062,6 +1188,7 @@ again: lockres->l_name, OCFS2_LOCK_ID_MAX_LEN - 1, lockres); + lockres_clear_pending(lockres, gen, osb); if (ret) { if (!(lkm_flags & DLM_LKF_NOQUEUE) || (ret != -EAGAIN)) { @@ -1506,6 +1633,7 @@ out: void ocfs2_file_unlock(struct file *file) { int ret; + unsigned int gen; unsigned long flags; struct ocfs2_file_private *fp = file->private_data; struct ocfs2_lock_res *lockres = &fp->fp_flock; @@ -1531,11 +1659,11 @@ void ocfs2_file_unlock(struct file *file) lockres_or_flags(lockres, OCFS2_LOCK_BLOCKED); lockres->l_blocking = DLM_LOCK_EX; - ocfs2_prepare_downconvert(lockres, LKM_NLMODE); + gen = ocfs2_prepare_downconvert(lockres, LKM_NLMODE); lockres_add_mask_waiter(lockres, &mw, OCFS2_LOCK_BUSY, 0); spin_unlock_irqrestore(&lockres->l_lock, flags); - ret = ocfs2_downconvert_lock(osb, lockres, LKM_NLMODE, 0); + ret = ocfs2_downconvert_lock(osb, lockres, LKM_NLMODE, 0, gen); if (ret) { mlog_errno(ret); return; @@ -2555,23 +2683,7 @@ static void ocfs2_unlock_ast(void *opaque, int error) lockres->l_unlock_action); spin_lock_irqsave(&lockres->l_lock, flags); - /* We tried to cancel a convert request, but it was already - * granted. All we want to do here is clear our unlock - * state. The wake_up call done at the bottom is redundant - * (ocfs2_prepare_cancel_convert doesn't sleep on this) but doesn't - * hurt anything anyway */ - if (error == -DLM_ECANCEL && - lockres->l_unlock_action == OCFS2_UNLOCK_CANCEL_CONVERT) { - mlog(0, "Got cancelgrant for %s\n", lockres->l_name); - - /* We don't clear the busy flag in this case as it - * should have been cleared by the ast which the dlm - * has called. */ - goto complete_unlock; - } - - /* DLM_EUNLOCK is the success code for unlock */ - if (error != -DLM_EUNLOCK) { + if (error) { mlog(ML_ERROR, "Dlm passes error %d for lock %s, " "unlock_action %d\n", error, lockres->l_name, lockres->l_unlock_action); @@ -2592,7 +2704,6 @@ static void ocfs2_unlock_ast(void *opaque, int error) } lockres_clear_flags(lockres, OCFS2_LOCK_BUSY); -complete_unlock: lockres->l_unlock_action = OCFS2_UNLOCK_INVALID; spin_unlock_irqrestore(&lockres->l_lock, flags); @@ -2768,8 +2879,8 @@ int ocfs2_drop_inode_locks(struct inode *inode) return status; } -static void ocfs2_prepare_downconvert(struct ocfs2_lock_res *lockres, - int new_level) +static unsigned int ocfs2_prepare_downconvert(struct ocfs2_lock_res *lockres, + int new_level) { assert_spin_locked(&lockres->l_lock); @@ -2787,12 +2898,14 @@ static void ocfs2_prepare_downconvert(struct ocfs2_lock_res *lockres, lockres->l_action = OCFS2_AST_DOWNCONVERT; lockres->l_requested = new_level; lockres_or_flags(lockres, OCFS2_LOCK_BUSY); + return lockres_set_pending(lockres); } static int ocfs2_downconvert_lock(struct ocfs2_super *osb, struct ocfs2_lock_res *lockres, int new_level, - int lvb) + int lvb, + unsigned int generation) { int ret; u32 dlm_flags = DLM_LKF_CONVERT; @@ -2809,6 +2922,7 @@ static int ocfs2_downconvert_lock(struct ocfs2_super *osb, lockres->l_name, OCFS2_LOCK_ID_MAX_LEN - 1, lockres); + lockres_clear_pending(lockres, generation, osb); if (ret) { ocfs2_log_dlm_error("ocfs2_dlm_lock", ret, lockres); ocfs2_recover_from_dlm_error(lockres, 1); @@ -2883,6 +2997,7 @@ static int ocfs2_unblock_lock(struct ocfs2_super *osb, int new_level; int ret = 0; int set_lvb = 0; + unsigned int gen; mlog_entry_void(); @@ -2892,6 +3007,32 @@ static int ocfs2_unblock_lock(struct ocfs2_super *osb, recheck: if (lockres->l_flags & OCFS2_LOCK_BUSY) { + /* XXX + * This is a *big* race. The OCFS2_LOCK_PENDING flag + * exists entirely for one reason - another thread has set + * OCFS2_LOCK_BUSY, but has *NOT* yet called dlm_lock(). + * + * If we do ocfs2_cancel_convert() before the other thread + * calls dlm_lock(), our cancel will do nothing. We will + * get no ast, and we will have no way of knowing the + * cancel failed. Meanwhile, the other thread will call + * into dlm_lock() and wait...forever. + * + * Why forever? Because another node has asked for the + * lock first; that's why we're here in unblock_lock(). + * + * The solution is OCFS2_LOCK_PENDING. When PENDING is + * set, we just requeue the unblock. Only when the other + * thread has called dlm_lock() and cleared PENDING will + * we then cancel their request. + * + * All callers of dlm_lock() must set OCFS2_DLM_PENDING + * at the same time they set OCFS2_DLM_BUSY. They must + * clear OCFS2_DLM_PENDING after dlm_lock() returns. + */ + if (lockres->l_flags & OCFS2_LOCK_PENDING) + goto leave_requeue; + ctl->requeue = 1; ret = ocfs2_prepare_cancel_convert(osb, lockres); spin_unlock_irqrestore(&lockres->l_lock, flags); @@ -2971,9 +3112,11 @@ downconvert: lockres->l_ops->set_lvb(lockres); } - ocfs2_prepare_downconvert(lockres, new_level); + gen = ocfs2_prepare_downconvert(lockres, new_level); spin_unlock_irqrestore(&lockres->l_lock, flags); - ret = ocfs2_downconvert_lock(osb, lockres, new_level, set_lvb); + ret = ocfs2_downconvert_lock(osb, lockres, new_level, set_lvb, + gen); + leave: mlog_exit(ret); return ret; diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h index 31dc28b4839..af929eca541 100644 --- a/fs/ocfs2/ocfs2.h +++ b/fs/ocfs2/ocfs2.h @@ -98,6 +98,9 @@ enum ocfs2_unlock_action { * dropped. */ #define OCFS2_LOCK_QUEUED (0x00000100) /* queued for downconvert */ #define OCFS2_LOCK_NOCACHE (0x00000200) /* don't use a holder count */ +#define OCFS2_LOCK_PENDING (0x00000400) /* This lockres is pending a + call to dlm_lock. Only + exists with BUSY set. */ struct ocfs2_lock_res_ops; @@ -124,6 +127,7 @@ struct ocfs2_lock_res { enum ocfs2_unlock_action l_unlock_action; int l_requested; int l_blocking; + unsigned int l_pending_gen; wait_queue_head_t l_event; diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c index 670fa945c21..abdb9f6f4cc 100644 --- a/fs/ocfs2/stackglue.c +++ b/fs/ocfs2/stackglue.c @@ -104,8 +104,8 @@ static int flags_to_o2dlm(u32 flags) * * DLM_NORMAL: 0 * DLM_NOTQUEUED: -EAGAIN - * DLM_CANCELGRANT: -DLM_ECANCEL - * DLM_CANCEL: -DLM_EUNLOCK + * DLM_CANCELGRANT: -EBUSY + * DLM_CANCEL: -DLM_ECANCEL */ /* Keep in sync with dlmapi.h */ static int status_map[] = { @@ -113,13 +113,13 @@ static int status_map[] = { [DLM_GRANTED] = -EINVAL, [DLM_DENIED] = -EACCES, [DLM_DENIED_NOLOCKS] = -EACCES, - [DLM_WORKING] = -EBUSY, + [DLM_WORKING] = -EACCES, [DLM_BLOCKED] = -EINVAL, [DLM_BLOCKED_ORPHAN] = -EINVAL, [DLM_DENIED_GRACE_PERIOD] = -EACCES, [DLM_SYSERR] = -ENOMEM, /* It is what it is */ [DLM_NOSUPPORT] = -EPROTO, - [DLM_CANCELGRANT] = -DLM_ECANCEL, /* Cancel after grant */ + [DLM_CANCELGRANT] = -EBUSY, /* Cancel after grant */ [DLM_IVLOCKID] = -EINVAL, [DLM_SYNC] = -EINVAL, [DLM_BADTYPE] = -EINVAL, @@ -137,7 +137,7 @@ static int status_map[] = { [DLM_VALNOTVALID] = -EINVAL, [DLM_REJECTED] = -EPERM, [DLM_ABORT] = -EINVAL, - [DLM_CANCEL] = -DLM_EUNLOCK, /* Successful cancel */ + [DLM_CANCEL] = -DLM_ECANCEL, /* Successful cancel */ [DLM_IVRESHANDLE] = -EINVAL, [DLM_DEADLOCK] = -EDEADLK, [DLM_DENIED_NOASTS] = -EINVAL, @@ -152,6 +152,7 @@ static int status_map[] = { [DLM_MIGRATING] = -ERESTART, [DLM_MAXSTATS] = -EINVAL, }; + static int dlm_status_to_errno(enum dlm_status status) { BUG_ON(status > (sizeof(status_map) / sizeof(status_map[0]))); @@ -175,38 +176,23 @@ static void o2dlm_blocking_ast_wrapper(void *astarg, int level) static void o2dlm_unlock_ast_wrapper(void *astarg, enum dlm_status status) { - int error; + int error = dlm_status_to_errno(status); BUG_ON(lproto == NULL); /* - * XXX: CANCEL values are sketchy. - * - * Currently we have preserved the o2dlm paradigm. You can get - * unlock_ast() whether the cancel succeded or not. - * - * First, we're going to pass DLM_EUNLOCK just like fs/dlm does for - * successful unlocks. That is a clean behavior. - * * In o2dlm, you can get both the lock_ast() for the lock being * granted and the unlock_ast() for the CANCEL failing. A * successful cancel sends DLM_NORMAL here. If the * lock grant happened before the cancel arrived, you get - * DLM_CANCELGRANT. For now, we'll use DLM_ECANCEL to signify - * CANCELGRANT - the CANCEL was supposed to happen but didn't. We - * can then use DLM_EUNLOCK to signify a successful CANCEL - - * effectively, the CANCEL caused the lock to roll back. + * DLM_CANCELGRANT. * - * In the future, we will likely move the o2dlm to send only one - * ast - either unlock_ast() for a successful CANCEL or lock_ast() - * when the grant succeeds. At that point, we'll send DLM_ECANCEL - * for all cancel results (CANCELGRANT will no longer exist). + * There's no need for the double-ast. If we see DLM_CANCELGRANT, + * we just ignore it. We expect the lock_ast() to handle the + * granted lock. */ - error = dlm_status_to_errno(status); - - /* Successful unlock is DLM_EUNLOCK */ - if (!error) - error = -DLM_EUNLOCK; + if (status == DLM_CANCELGRANT) + return; lproto->lp_unlock_ast(astarg, error); } -- cgit v1.2.3 From 1693a5c0117f8ccd010a666f97aaf0f14fb0a0e4 Mon Sep 17 00:00:00 2001 From: David Teigland Date: Wed, 30 Jan 2008 16:52:53 -0800 Subject: ocfs2: handle async EAGAIN from NOQUEUE request When using fsdlm, -EAGAIN is returned in the async callback for NOQUEUE requests. Fix up dlmglue to expect this. Signed-off-by: David Teigland Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlmglue.c | 27 +++++++++++++++++++++++---- 1 file changed, 23 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index 295c47f7aba..b640423b936 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -880,13 +880,20 @@ static void ocfs2_locking_ast(void *opaque) struct ocfs2_lock_res *lockres = opaque; struct ocfs2_super *osb = ocfs2_get_lockres_osb(lockres); unsigned long flags; + int status; spin_lock_irqsave(&lockres->l_lock, flags); - if (ocfs2_dlm_lock_status(&lockres->l_lksb)) { + status = ocfs2_dlm_lock_status(&lockres->l_lksb); + + if (status == -EAGAIN) { + lockres_clear_flags(lockres, OCFS2_LOCK_BUSY); + goto out; + } + + if (status) { mlog(ML_ERROR, "lockres %s: lksb status value of %d!\n", - lockres->l_name, - ocfs2_dlm_lock_status(&lockres->l_lksb)); + lockres->l_name, status); spin_unlock_irqrestore(&lockres->l_lock, flags); return; } @@ -909,7 +916,7 @@ static void ocfs2_locking_ast(void *opaque) lockres->l_unlock_action); BUG(); } - +out: /* set it to something invalid so if we get called again we * can catch it. */ lockres->l_action = OCFS2_AST_INVALID; @@ -1113,6 +1120,7 @@ static int ocfs2_cluster_lock(struct ocfs2_super *osb, int ret = 0; /* gcc doesn't realize wait = 1 guarantees ret is set */ unsigned long flags; unsigned int gen; + int noqueue_attempted = 0; mlog_entry_void(); @@ -1157,6 +1165,13 @@ again: } if (level > lockres->l_level) { + if (noqueue_attempted > 0) { + ret = -EAGAIN; + goto unlock; + } + if (lkm_flags & DLM_LKF_NOQUEUE) + noqueue_attempted = 1; + if (lockres->l_action != OCFS2_AST_INVALID) mlog(ML_ERROR, "lockres %s has action %u pending\n", lockres->l_name, lockres->l_action); @@ -1621,6 +1636,10 @@ int ocfs2_file_lock(struct file *file, int ex, int trylock) * to just bubble sucess back up to the user. */ ret = ocfs2_flock_handle_signal(lockres, level); + } else if (!ret && (level > lockres->l_level)) { + /* Trylock failed asynchronously */ + BUG_ON(!trylock); + ret = -EAGAIN; } out: -- cgit v1.2.3 From cf0acdcd640e9466059e69951c557e90b4bee45a Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Tue, 29 Jan 2008 16:59:55 -0800 Subject: ocfs2: Abstract out a debugging function for underlying dlms. dlmglue.c was still referencing a raw o2dlm lksb in one instance. Let's create a generic ocfs2_dlm_dump_lksb() function. This allows underlying DLMs to print whatever they want about their lock. We then move the o2dlm dump into stackglue.c where it belongs. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlmglue.c | 3 +-- fs/ocfs2/stackglue.c | 5 +++++ fs/ocfs2/stackglue.h | 1 + 3 files changed, 7 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index b640423b936..f41ff1c10ae 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -2803,8 +2803,7 @@ static int ocfs2_drop_lock(struct ocfs2_super *osb, if (ret) { ocfs2_log_dlm_error("ocfs2_dlm_unlock", ret, lockres); mlog(ML_ERROR, "lockres flags: %lu\n", lockres->l_flags); - /* XXX Need to abstract this */ - dlm_print_one_lock(lockres->l_lksb.lksb_o2dlm.lockid); + ocfs2_dlm_dump_lksb(&lockres->l_lksb); BUG(); } mlog(0, "lock %s, successfull return from ocfs2_dlm_unlock\n", diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c index abdb9f6f4cc..bd805411a85 100644 --- a/fs/ocfs2/stackglue.c +++ b/fs/ocfs2/stackglue.c @@ -252,6 +252,11 @@ void *ocfs2_dlm_lvb(union ocfs2_dlm_lksb *lksb) return (void *)(lksb->lksb_o2dlm.lvb); } +void ocfs2_dlm_dump_lksb(union ocfs2_dlm_lksb *lksb) +{ + dlm_print_one_lock(lksb->lksb_o2dlm.lockid); +} + /* * Called from the dlm when it's about to evict a node. This is how the * classic stack signals node death. diff --git a/fs/ocfs2/stackglue.h b/fs/ocfs2/stackglue.h index 22af77b9a69..01e3c9b9192 100644 --- a/fs/ocfs2/stackglue.h +++ b/fs/ocfs2/stackglue.h @@ -91,6 +91,7 @@ int ocfs2_dlm_unlock(struct ocfs2_cluster_connection *conn, int ocfs2_dlm_lock_status(union ocfs2_dlm_lksb *lksb); void *ocfs2_dlm_lvb(union ocfs2_dlm_lksb *lksb); +void ocfs2_dlm_dump_lksb(union ocfs2_dlm_lksb *lksb); void o2cb_get_stack(struct ocfs2_locking_protocol *proto); void o2cb_put_stack(void); -- cgit v1.2.3 From 63e0c48ae6986a5bbb8e8dd9210c0e6ca79f2e50 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Wed, 30 Jan 2008 16:58:36 -0800 Subject: ocfs2: Clean up stackglue initialization The stack glue initialization function needs a better name so that it can be used cleanly when stackglue becomes a module. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlmglue.c | 9 ++------- fs/ocfs2/dlmglue.h | 5 ++--- fs/ocfs2/stackglue.c | 8 ++------ fs/ocfs2/stackglue.h | 3 +-- fs/ocfs2/super.c | 6 ++---- 5 files changed, 9 insertions(+), 22 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index f41ff1c10ae..8a9c84909be 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -3366,16 +3366,11 @@ static struct ocfs2_locking_protocol lproto = { .lp_unlock_ast = ocfs2_unlock_ast, }; -/* This interface isn't the final one, hence the less-than-perfect names */ -void dlmglue_init_stack(void) +void ocfs2_set_locking_protocol(void) { - o2cb_get_stack(&lproto); + ocfs2_stack_glue_set_locking_protocol(&lproto); } -void dlmglue_exit_stack(void) -{ - o2cb_put_stack(); -} static void ocfs2_process_blocked_lock(struct ocfs2_super *osb, struct ocfs2_lock_res *lockres) diff --git a/fs/ocfs2/dlmglue.h b/fs/ocfs2/dlmglue.h index 2d0a8a03c43..34b7598a0dc 100644 --- a/fs/ocfs2/dlmglue.h +++ b/fs/ocfs2/dlmglue.h @@ -114,7 +114,6 @@ void ocfs2_wake_downconvert_thread(struct ocfs2_super *osb); struct ocfs2_dlm_debug *ocfs2_new_dlm_debug(void); void ocfs2_put_dlm_debug(struct ocfs2_dlm_debug *dlm_debug); -void dlmglue_init_stack(void); -void dlmglue_exit_stack(void); - +/* To set the locking protocol on module initialization */ +void ocfs2_set_locking_protocol(void); #endif /* DLMGLUE_H */ diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c index bd805411a85..51c2546b328 100644 --- a/fs/ocfs2/stackglue.c +++ b/fs/ocfs2/stackglue.c @@ -429,14 +429,10 @@ int ocfs2_cluster_this_node(unsigned int *node) return 0; } -void o2cb_get_stack(struct ocfs2_locking_protocol *proto) +void ocfs2_stack_glue_set_locking_protocol(struct ocfs2_locking_protocol *proto) { - BUG_ON(proto == NULL); + BUG_ON(proto != NULL); lproto = proto; } -void o2cb_put_stack(void) -{ - lproto = NULL; -} diff --git a/fs/ocfs2/stackglue.h b/fs/ocfs2/stackglue.h index 01e3c9b9192..decb147106f 100644 --- a/fs/ocfs2/stackglue.h +++ b/fs/ocfs2/stackglue.h @@ -93,7 +93,6 @@ int ocfs2_dlm_lock_status(union ocfs2_dlm_lksb *lksb); void *ocfs2_dlm_lvb(union ocfs2_dlm_lksb *lksb); void ocfs2_dlm_dump_lksb(union ocfs2_dlm_lksb *lksb); -void o2cb_get_stack(struct ocfs2_locking_protocol *proto); -void o2cb_put_stack(void); +void ocfs2_stack_glue_set_locking_protocol(struct ocfs2_locking_protocol *proto); #endif /* STACKGLUE_H */ diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index fa9c46e2eab..b4a02a00665 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -922,8 +922,6 @@ static int __init ocfs2_init(void) ocfs2_print_version(); - dlmglue_init_stack(); - status = init_ocfs2_uptodate_cache(); if (status < 0) { mlog_errno(status); @@ -948,6 +946,8 @@ static int __init ocfs2_init(void) mlog(ML_ERROR, "Unable to create ocfs2 debugfs root.\n"); } + ocfs2_set_locking_protocol(); + leave: if (status < 0) { ocfs2_free_mem_caches(); @@ -979,8 +979,6 @@ static void __exit ocfs2_exit(void) exit_ocfs2_uptodate_cache(); - dlmglue_exit_stack(); - mlog_exit_void(); } -- cgit v1.2.3 From 553aa7e408eac402c00b67ddfa7aec13fe1f3a33 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Fri, 1 Feb 2008 14:51:03 -0800 Subject: ocfs2: Split o2cb code from generic stack functions. Split off the o2cb-specific funtionality from the generic stack glue calls. This is a precurser to wrapping the o2cb functionality in an operations vector. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/stackglue.c | 209 +++++++++++++++++++++++++++++++++++---------------- 1 file changed, 144 insertions(+), 65 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c index 51c2546b328..e35dde6217f 100644 --- a/fs/ocfs2/stackglue.c +++ b/fs/ocfs2/stackglue.c @@ -197,21 +197,19 @@ static void o2dlm_unlock_ast_wrapper(void *astarg, enum dlm_status status) lproto->lp_unlock_ast(astarg, error); } -int ocfs2_dlm_lock(struct ocfs2_cluster_connection *conn, - int mode, - union ocfs2_dlm_lksb *lksb, - u32 flags, - void *name, - unsigned int namelen, - void *astarg) +static int o2cb_dlm_lock(struct ocfs2_cluster_connection *conn, + int mode, + union ocfs2_dlm_lksb *lksb, + u32 flags, + void *name, + unsigned int namelen, + void *astarg) { enum dlm_status status; int o2dlm_mode = mode_to_o2dlm(mode); int o2dlm_flags = flags_to_o2dlm(flags); int ret; - BUG_ON(lproto == NULL); - status = dlmlock(conn->cc_lockspace, o2dlm_mode, &lksb->lksb_o2dlm, o2dlm_flags, name, namelen, o2dlm_lock_ast_wrapper, astarg, @@ -220,43 +218,80 @@ int ocfs2_dlm_lock(struct ocfs2_cluster_connection *conn, return ret; } -int ocfs2_dlm_unlock(struct ocfs2_cluster_connection *conn, - union ocfs2_dlm_lksb *lksb, - u32 flags, - void *astarg) +int ocfs2_dlm_lock(struct ocfs2_cluster_connection *conn, + int mode, + union ocfs2_dlm_lksb *lksb, + u32 flags, + void *name, + unsigned int namelen, + void *astarg) +{ + BUG_ON(lproto == NULL); + + return o2cb_dlm_lock(conn, mode, lksb, flags, + name, namelen, astarg); +} + +static int o2cb_dlm_unlock(struct ocfs2_cluster_connection *conn, + union ocfs2_dlm_lksb *lksb, + u32 flags, + void *astarg) { enum dlm_status status; int o2dlm_flags = flags_to_o2dlm(flags); int ret; - BUG_ON(lproto == NULL); - status = dlmunlock(conn->cc_lockspace, &lksb->lksb_o2dlm, o2dlm_flags, o2dlm_unlock_ast_wrapper, astarg); ret = dlm_status_to_errno(status); return ret; } -int ocfs2_dlm_lock_status(union ocfs2_dlm_lksb *lksb) +int ocfs2_dlm_unlock(struct ocfs2_cluster_connection *conn, + union ocfs2_dlm_lksb *lksb, + u32 flags, + void *astarg) +{ + BUG_ON(lproto == NULL); + + return o2cb_dlm_unlock(conn, lksb, flags, astarg); +} + +static int o2cb_dlm_lock_status(union ocfs2_dlm_lksb *lksb) { return dlm_status_to_errno(lksb->lksb_o2dlm.status); } +int ocfs2_dlm_lock_status(union ocfs2_dlm_lksb *lksb) +{ + return o2cb_dlm_lock_status(lksb); +} + /* * Why don't we cast to ocfs2_meta_lvb? The "clean" answer is that we * don't cast at the glue level. The real answer is that the header * ordering is nigh impossible. */ -void *ocfs2_dlm_lvb(union ocfs2_dlm_lksb *lksb) +static void *o2cb_dlm_lvb(union ocfs2_dlm_lksb *lksb) { return (void *)(lksb->lksb_o2dlm.lvb); } -void ocfs2_dlm_dump_lksb(union ocfs2_dlm_lksb *lksb) +void *ocfs2_dlm_lvb(union ocfs2_dlm_lksb *lksb) +{ + return o2cb_dlm_lvb(lksb); +} + +static void o2cb_dlm_dump_lksb(union ocfs2_dlm_lksb *lksb) { dlm_print_one_lock(lksb->lksb_o2dlm.lockid); } +void ocfs2_dlm_dump_lksb(union ocfs2_dlm_lksb *lksb) +{ + o2cb_dlm_dump_lksb(lksb); +} + /* * Called from the dlm when it's about to evict a node. This is how the * classic stack signals node death. @@ -271,6 +306,62 @@ static void o2dlm_eviction_cb(int node_num, void *data) conn->cc_recovery_handler(node_num, conn->cc_recovery_data); } +static int o2cb_cluster_connect(struct ocfs2_cluster_connection *conn) +{ + int rc = 0; + u32 dlm_key; + struct dlm_ctxt *dlm; + struct o2dlm_private *priv; + struct dlm_protocol_version dlm_version; + + BUG_ON(conn == NULL); + + /* for now we only have one cluster/node, make sure we see it + * in the heartbeat universe */ + if (!o2hb_check_local_node_heartbeating()) { + rc = -EINVAL; + goto out; + } + + priv = kzalloc(sizeof(struct o2dlm_private), GFP_KERNEL); + if (!priv) { + rc = -ENOMEM; + goto out_free; + } + + /* This just fills the structure in. It is safe to pass conn. */ + dlm_setup_eviction_cb(&priv->op_eviction_cb, o2dlm_eviction_cb, + conn); + + conn->cc_private = priv; + + /* used by the dlm code to make message headers unique, each + * node in this domain must agree on this. */ + dlm_key = crc32_le(0, conn->cc_name, conn->cc_namelen); + dlm_version.pv_major = conn->cc_version.pv_major; + dlm_version.pv_minor = conn->cc_version.pv_minor; + + dlm = dlm_register_domain(conn->cc_name, dlm_key, &dlm_version); + if (IS_ERR(dlm)) { + rc = PTR_ERR(dlm); + mlog_errno(rc); + goto out_free; + } + + conn->cc_version.pv_major = dlm_version.pv_major; + conn->cc_version.pv_minor = dlm_version.pv_minor; + conn->cc_lockspace = dlm; + + dlm_register_eviction_cb(dlm, &priv->op_eviction_cb); + +out_free: + if (rc && conn->cc_private) + kfree(conn->cc_private); + +out: + return rc; +} + int ocfs2_cluster_connect(const char *group, int grouplen, void (*recovery_handler)(int node_num, @@ -280,10 +371,6 @@ int ocfs2_cluster_connect(const char *group, { int rc = 0; struct ocfs2_cluster_connection *new_conn; - u32 dlm_key; - struct dlm_ctxt *dlm; - struct o2dlm_private *priv; - struct dlm_protocol_version dlm_version; BUG_ON(group == NULL); BUG_ON(conn == NULL); @@ -294,13 +381,6 @@ int ocfs2_cluster_connect(const char *group, goto out; } - /* for now we only have one cluster/node, make sure we see it - * in the heartbeat universe */ - if (!o2hb_check_local_node_heartbeating()) { - rc = -EINVAL; - goto out; - } - new_conn = kzalloc(sizeof(struct ocfs2_cluster_connection), GFP_KERNEL); if (!new_conn) { @@ -316,64 +396,53 @@ int ocfs2_cluster_connect(const char *group, /* Start the new connection at our maximum compatibility level */ new_conn->cc_version = lproto->lp_max_version; - priv = kzalloc(sizeof(struct o2dlm_private), GFP_KERNEL); - if (!priv) { - rc = -ENOMEM; - goto out_free; - } - - /* This just fills the structure in. It is safe to use new_conn. */ - dlm_setup_eviction_cb(&priv->op_eviction_cb, o2dlm_eviction_cb, - new_conn); - - new_conn->cc_private = priv; - - /* used by the dlm code to make message headers unique, each - * node in this domain must agree on this. */ - dlm_key = crc32_le(0, group, grouplen); - dlm_version.pv_major = new_conn->cc_version.pv_major; - dlm_version.pv_minor = new_conn->cc_version.pv_minor; - - dlm = dlm_register_domain(group, dlm_key, &dlm_version); - if (IS_ERR(dlm)) { - rc = PTR_ERR(dlm); + rc = o2cb_cluster_connect(new_conn); + if (rc) { mlog_errno(rc); goto out_free; } - new_conn->cc_version.pv_major = dlm_version.pv_major; - new_conn->cc_version.pv_minor = dlm_version.pv_minor; - new_conn->cc_lockspace = dlm; - - dlm_register_eviction_cb(dlm, &priv->op_eviction_cb); - *conn = new_conn; out_free: - if (rc) { - kfree(new_conn->cc_private); + if (rc) kfree(new_conn); - } out: return rc; } -int ocfs2_cluster_disconnect(struct ocfs2_cluster_connection *conn) +static int o2cb_cluster_disconnect(struct ocfs2_cluster_connection *conn) { struct dlm_ctxt *dlm = conn->cc_lockspace; struct o2dlm_private *priv = conn->cc_private; dlm_unregister_eviction_cb(&priv->op_eviction_cb); - dlm_unregister_domain(dlm); - + conn->cc_private = NULL; kfree(priv); - kfree(conn); + + dlm_unregister_domain(dlm); + conn->cc_lockspace = NULL; return 0; } +int ocfs2_cluster_disconnect(struct ocfs2_cluster_connection *conn) +{ + int ret; + + BUG_ON(conn == NULL); + + ret = o2cb_cluster_disconnect(conn); + + /* XXX Should we free it anyway? */ + if (!ret) + kfree(conn); + + return ret; +} + static void o2hb_stop(const char *group) { int ret; @@ -406,15 +475,20 @@ static void o2hb_stop(const char *group) * * Other stacks will eventually provide a NULL ->hangup() pointer. */ +static void o2cb_cluster_hangup(const char *group, int grouplen) +{ + o2hb_stop(group); +} + void ocfs2_cluster_hangup(const char *group, int grouplen) { BUG_ON(group == NULL); BUG_ON(group[grouplen] != '\0'); - o2hb_stop(group); + o2cb_cluster_hangup(group, grouplen); } -int ocfs2_cluster_this_node(unsigned int *node) +static int o2cb_cluster_this_node(unsigned int *node) { int node_num; @@ -429,6 +503,11 @@ int ocfs2_cluster_this_node(unsigned int *node) return 0; } +int ocfs2_cluster_this_node(unsigned int *node) +{ + return o2cb_cluster_this_node(node); +} + void ocfs2_stack_glue_set_locking_protocol(struct ocfs2_locking_protocol *proto) { BUG_ON(proto != NULL); -- cgit v1.2.3 From e3dad42bf993a0f24eb6e46152356c9b119c15e8 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Fri, 1 Feb 2008 15:02:36 -0800 Subject: ocfs2: Create ocfs2_stack_operations and split out the o2cb stack. Define the ocfs2_stack_operations structure. Build o2cb_stack_ops from all of the o2cb-specific stack functions. Change the generic stack glue functions to call the stack_ops instead of the o2cb functions directly. The o2cb functions are moved to stack_o2cb.c. The headers are cleaned up to where only needed headers are included. In this code, stackglue.c and stack_o2cb.c refer to some shared extern variables. When they become modules, that will change. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/Makefile | 1 + fs/ocfs2/stack_o2cb.c | 395 ++++++++++++++++++++++++++++++++++++++++++++++++++ fs/ocfs2/stackglue.c | 385 ++---------------------------------------------- fs/ocfs2/stackglue.h | 123 +++++++++++++++- 4 files changed, 532 insertions(+), 372 deletions(-) create mode 100644 fs/ocfs2/stack_o2cb.c (limited to 'fs') diff --git a/fs/ocfs2/Makefile b/fs/ocfs2/Makefile index 3ba64af2695..8e86195fedb 100644 --- a/fs/ocfs2/Makefile +++ b/fs/ocfs2/Makefile @@ -25,6 +25,7 @@ ocfs2-objs := \ resize.o \ slot_map.o \ stackglue.o \ + stack_o2cb.o \ suballoc.o \ super.o \ symlink.o \ diff --git a/fs/ocfs2/stack_o2cb.c b/fs/ocfs2/stack_o2cb.c new file mode 100644 index 00000000000..c9bc3541a33 --- /dev/null +++ b/fs/ocfs2/stack_o2cb.c @@ -0,0 +1,395 @@ +/* -*- mode: c; c-basic-offset: 8; -*- + * vim: noexpandtab sw=8 ts=8 sts=0: + * + * stack_o2cb.c + * + * Code which interfaces ocfs2 with the o2cb stack. + * + * Copyright (C) 2007 Oracle. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License as published by the Free Software Foundation, version 2. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + */ + +#include +#include + +/* Needed for AOP_TRUNCATED_PAGE in mlog_errno() */ +#include + +#include "cluster/masklog.h" +#include "cluster/nodemanager.h" +#include "cluster/heartbeat.h" + +#include "stackglue.h" + +struct o2dlm_private { + struct dlm_eviction_cb op_eviction_cb; +}; + +/* These should be identical */ +#if (DLM_LOCK_IV != LKM_IVMODE) +# error Lock modes do not match +#endif +#if (DLM_LOCK_NL != LKM_NLMODE) +# error Lock modes do not match +#endif +#if (DLM_LOCK_CR != LKM_CRMODE) +# error Lock modes do not match +#endif +#if (DLM_LOCK_CW != LKM_CWMODE) +# error Lock modes do not match +#endif +#if (DLM_LOCK_PR != LKM_PRMODE) +# error Lock modes do not match +#endif +#if (DLM_LOCK_PW != LKM_PWMODE) +# error Lock modes do not match +#endif +#if (DLM_LOCK_EX != LKM_EXMODE) +# error Lock modes do not match +#endif +static inline int mode_to_o2dlm(int mode) +{ + BUG_ON(mode > LKM_MAXMODE); + + return mode; +} + +#define map_flag(_generic, _o2dlm) \ + if (flags & (_generic)) { \ + flags &= ~(_generic); \ + o2dlm_flags |= (_o2dlm); \ + } +static int flags_to_o2dlm(u32 flags) +{ + int o2dlm_flags = 0; + + map_flag(DLM_LKF_NOQUEUE, LKM_NOQUEUE); + map_flag(DLM_LKF_CANCEL, LKM_CANCEL); + map_flag(DLM_LKF_CONVERT, LKM_CONVERT); + map_flag(DLM_LKF_VALBLK, LKM_VALBLK); + map_flag(DLM_LKF_IVVALBLK, LKM_INVVALBLK); + map_flag(DLM_LKF_ORPHAN, LKM_ORPHAN); + map_flag(DLM_LKF_FORCEUNLOCK, LKM_FORCE); + map_flag(DLM_LKF_TIMEOUT, LKM_TIMEOUT); + map_flag(DLM_LKF_LOCAL, LKM_LOCAL); + + /* map_flag() should have cleared every flag passed in */ + BUG_ON(flags != 0); + + return o2dlm_flags; +} +#undef map_flag + +/* + * Map an o2dlm status to standard errno values. + * + * o2dlm only uses a handful of these, and returns even fewer to the + * caller. Still, we try to assign sane values to each error. + * + * The following value pairs have special meanings to dlmglue, thus + * the right hand side needs to stay unique - never duplicate the + * mapping elsewhere in the table! + * + * DLM_NORMAL: 0 + * DLM_NOTQUEUED: -EAGAIN + * DLM_CANCELGRANT: -EBUSY + * DLM_CANCEL: -DLM_ECANCEL + */ +/* Keep in sync with dlmapi.h */ +static int status_map[] = { + [DLM_NORMAL] = 0, /* Success */ + [DLM_GRANTED] = -EINVAL, + [DLM_DENIED] = -EACCES, + [DLM_DENIED_NOLOCKS] = -EACCES, + [DLM_WORKING] = -EACCES, + [DLM_BLOCKED] = -EINVAL, + [DLM_BLOCKED_ORPHAN] = -EINVAL, + [DLM_DENIED_GRACE_PERIOD] = -EACCES, + [DLM_SYSERR] = -ENOMEM, /* It is what it is */ + [DLM_NOSUPPORT] = -EPROTO, + [DLM_CANCELGRANT] = -EBUSY, /* Cancel after grant */ + [DLM_IVLOCKID] = -EINVAL, + [DLM_SYNC] = -EINVAL, + [DLM_BADTYPE] = -EINVAL, + [DLM_BADRESOURCE] = -EINVAL, + [DLM_MAXHANDLES] = -ENOMEM, + [DLM_NOCLINFO] = -EINVAL, + [DLM_NOLOCKMGR] = -EINVAL, + [DLM_NOPURGED] = -EINVAL, + [DLM_BADARGS] = -EINVAL, + [DLM_VOID] = -EINVAL, + [DLM_NOTQUEUED] = -EAGAIN, /* Trylock failed */ + [DLM_IVBUFLEN] = -EINVAL, + [DLM_CVTUNGRANT] = -EPERM, + [DLM_BADPARAM] = -EINVAL, + [DLM_VALNOTVALID] = -EINVAL, + [DLM_REJECTED] = -EPERM, + [DLM_ABORT] = -EINVAL, + [DLM_CANCEL] = -DLM_ECANCEL, /* Successful cancel */ + [DLM_IVRESHANDLE] = -EINVAL, + [DLM_DEADLOCK] = -EDEADLK, + [DLM_DENIED_NOASTS] = -EINVAL, + [DLM_FORWARD] = -EINVAL, + [DLM_TIMEOUT] = -ETIMEDOUT, + [DLM_IVGROUPID] = -EINVAL, + [DLM_VERS_CONFLICT] = -EOPNOTSUPP, + [DLM_BAD_DEVICE_PATH] = -ENOENT, + [DLM_NO_DEVICE_PERMISSION] = -EPERM, + [DLM_NO_CONTROL_DEVICE] = -ENOENT, + [DLM_RECOVERING] = -ENOTCONN, + [DLM_MIGRATING] = -ERESTART, + [DLM_MAXSTATS] = -EINVAL, +}; + +static int dlm_status_to_errno(enum dlm_status status) +{ + BUG_ON(status > (sizeof(status_map) / sizeof(status_map[0]))); + + return status_map[status]; +} + +static void o2dlm_lock_ast_wrapper(void *astarg) +{ + BUG_ON(stack_glue_lproto == NULL); + + stack_glue_lproto->lp_lock_ast(astarg); +} + +static void o2dlm_blocking_ast_wrapper(void *astarg, int level) +{ + BUG_ON(stack_glue_lproto == NULL); + + stack_glue_lproto->lp_blocking_ast(astarg, level); +} + +static void o2dlm_unlock_ast_wrapper(void *astarg, enum dlm_status status) +{ + int error = dlm_status_to_errno(status); + + BUG_ON(stack_glue_lproto == NULL); + + /* + * In o2dlm, you can get both the lock_ast() for the lock being + * granted and the unlock_ast() for the CANCEL failing. A + * successful cancel sends DLM_NORMAL here. If the + * lock grant happened before the cancel arrived, you get + * DLM_CANCELGRANT. + * + * There's no need for the double-ast. If we see DLM_CANCELGRANT, + * we just ignore it. We expect the lock_ast() to handle the + * granted lock. + */ + if (status == DLM_CANCELGRANT) + return; + + stack_glue_lproto->lp_unlock_ast(astarg, error); +} + +static int o2cb_dlm_lock(struct ocfs2_cluster_connection *conn, + int mode, + union ocfs2_dlm_lksb *lksb, + u32 flags, + void *name, + unsigned int namelen, + void *astarg) +{ + enum dlm_status status; + int o2dlm_mode = mode_to_o2dlm(mode); + int o2dlm_flags = flags_to_o2dlm(flags); + int ret; + + status = dlmlock(conn->cc_lockspace, o2dlm_mode, &lksb->lksb_o2dlm, + o2dlm_flags, name, namelen, + o2dlm_lock_ast_wrapper, astarg, + o2dlm_blocking_ast_wrapper); + ret = dlm_status_to_errno(status); + return ret; +} + +static int o2cb_dlm_unlock(struct ocfs2_cluster_connection *conn, + union ocfs2_dlm_lksb *lksb, + u32 flags, + void *astarg) +{ + enum dlm_status status; + int o2dlm_flags = flags_to_o2dlm(flags); + int ret; + + status = dlmunlock(conn->cc_lockspace, &lksb->lksb_o2dlm, + o2dlm_flags, o2dlm_unlock_ast_wrapper, astarg); + ret = dlm_status_to_errno(status); + return ret; +} + +static int o2cb_dlm_lock_status(union ocfs2_dlm_lksb *lksb) +{ + return dlm_status_to_errno(lksb->lksb_o2dlm.status); +} + +static void *o2cb_dlm_lvb(union ocfs2_dlm_lksb *lksb) +{ + return (void *)(lksb->lksb_o2dlm.lvb); +} + +static void o2cb_dump_lksb(union ocfs2_dlm_lksb *lksb) +{ + dlm_print_one_lock(lksb->lksb_o2dlm.lockid); +} + +/* + * Called from the dlm when it's about to evict a node. This is how the + * classic stack signals node death. + */ +static void o2dlm_eviction_cb(int node_num, void *data) +{ + struct ocfs2_cluster_connection *conn = data; + + mlog(ML_NOTICE, "o2dlm has evicted node %d from group %.*s\n", + node_num, conn->cc_namelen, conn->cc_name); + + conn->cc_recovery_handler(node_num, conn->cc_recovery_data); +} + +static int o2cb_cluster_connect(struct ocfs2_cluster_connection *conn) +{ + int rc = 0; + u32 dlm_key; + struct dlm_ctxt *dlm; + struct o2dlm_private *priv; + struct dlm_protocol_version dlm_version; + + BUG_ON(conn == NULL); + + /* for now we only have one cluster/node, make sure we see it + * in the heartbeat universe */ + if (!o2hb_check_local_node_heartbeating()) { + rc = -EINVAL; + goto out; + } + + priv = kzalloc(sizeof(struct o2dlm_private), GFP_KERNEL); + if (!priv) { + rc = -ENOMEM; + goto out_free; + } + + /* This just fills the structure in. It is safe to pass conn. */ + dlm_setup_eviction_cb(&priv->op_eviction_cb, o2dlm_eviction_cb, + conn); + + conn->cc_private = priv; + + /* used by the dlm code to make message headers unique, each + * node in this domain must agree on this. */ + dlm_key = crc32_le(0, conn->cc_name, conn->cc_namelen); + dlm_version.pv_major = conn->cc_version.pv_major; + dlm_version.pv_minor = conn->cc_version.pv_minor; + + dlm = dlm_register_domain(conn->cc_name, dlm_key, &dlm_version); + if (IS_ERR(dlm)) { + rc = PTR_ERR(dlm); + mlog_errno(rc); + goto out_free; + } + + conn->cc_version.pv_major = dlm_version.pv_major; + conn->cc_version.pv_minor = dlm_version.pv_minor; + conn->cc_lockspace = dlm; + + dlm_register_eviction_cb(dlm, &priv->op_eviction_cb); + +out_free: + if (rc && conn->cc_private) + kfree(conn->cc_private); + +out: + return rc; +} + +static int o2cb_cluster_disconnect(struct ocfs2_cluster_connection *conn) +{ + struct dlm_ctxt *dlm = conn->cc_lockspace; + struct o2dlm_private *priv = conn->cc_private; + + dlm_unregister_eviction_cb(&priv->op_eviction_cb); + conn->cc_private = NULL; + kfree(priv); + + dlm_unregister_domain(dlm); + conn->cc_lockspace = NULL; + + return 0; +} + +static void o2hb_stop(const char *group) +{ + int ret; + char *argv[5], *envp[3]; + + argv[0] = (char *)o2nm_get_hb_ctl_path(); + argv[1] = "-K"; + argv[2] = "-u"; + argv[3] = (char *)group; + argv[4] = NULL; + + mlog(0, "Run: %s %s %s %s\n", argv[0], argv[1], argv[2], argv[3]); + + /* minimal command environment taken from cpu_run_sbin_hotplug */ + envp[0] = "HOME=/"; + envp[1] = "PATH=/sbin:/bin:/usr/sbin:/usr/bin"; + envp[2] = NULL; + + ret = call_usermodehelper(argv[0], argv, envp, UMH_WAIT_PROC); + if (ret < 0) + mlog_errno(ret); +} + +/* + * Hangup is a hack for tools compatibility. Older ocfs2-tools software + * expects the filesystem to call "ocfs2_hb_ctl" during unmount. This + * happens regardless of whether the DLM got started, so we can't do it + * in ocfs2_cluster_disconnect(). We bring the o2hb_stop() function into + * the glue and provide a "hangup" API for super.c to call. + * + * Other stacks will eventually provide a NULL ->hangup() pointer. + */ +static void o2cb_cluster_hangup(const char *group, int grouplen) +{ + o2hb_stop(group); +} + +static int o2cb_cluster_this_node(unsigned int *node) +{ + int node_num; + + node_num = o2nm_this_node(); + if (node_num == O2NM_INVALID_NODE_NUM) + return -ENOENT; + + if (node_num >= O2NM_MAX_NODES) + return -EOVERFLOW; + + *node = node_num; + return 0; +} + +struct ocfs2_stack_operations o2cb_stack_ops = { + .connect = o2cb_cluster_connect, + .disconnect = o2cb_cluster_disconnect, + .hangup = o2cb_cluster_hangup, + .this_node = o2cb_cluster_this_node, + .dlm_lock = o2cb_dlm_lock, + .dlm_unlock = o2cb_dlm_unlock, + .lock_status = o2cb_dlm_lock_status, + .lock_lvb = o2cb_dlm_lvb, + .dump_lksb = o2cb_dump_lksb, +}; + diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c index e35dde6217f..e197367b6bd 100644 --- a/fs/ocfs2/stackglue.c +++ b/fs/ocfs2/stackglue.c @@ -19,204 +19,17 @@ */ #include -#include #include /* Needed for AOP_TRUNCATED_PAGE in mlog_errno() */ #include #include "cluster/masklog.h" -#include "cluster/nodemanager.h" -#include "cluster/heartbeat.h" #include "stackglue.h" -static struct ocfs2_locking_protocol *lproto; - -struct o2dlm_private { - struct dlm_eviction_cb op_eviction_cb; -}; - -/* These should be identical */ -#if (DLM_LOCK_IV != LKM_IVMODE) -# error Lock modes do not match -#endif -#if (DLM_LOCK_NL != LKM_NLMODE) -# error Lock modes do not match -#endif -#if (DLM_LOCK_CR != LKM_CRMODE) -# error Lock modes do not match -#endif -#if (DLM_LOCK_CW != LKM_CWMODE) -# error Lock modes do not match -#endif -#if (DLM_LOCK_PR != LKM_PRMODE) -# error Lock modes do not match -#endif -#if (DLM_LOCK_PW != LKM_PWMODE) -# error Lock modes do not match -#endif -#if (DLM_LOCK_EX != LKM_EXMODE) -# error Lock modes do not match -#endif -static inline int mode_to_o2dlm(int mode) -{ - BUG_ON(mode > LKM_MAXMODE); - - return mode; -} - -#define map_flag(_generic, _o2dlm) \ - if (flags & (_generic)) { \ - flags &= ~(_generic); \ - o2dlm_flags |= (_o2dlm); \ - } -static int flags_to_o2dlm(u32 flags) -{ - int o2dlm_flags = 0; - - map_flag(DLM_LKF_NOQUEUE, LKM_NOQUEUE); - map_flag(DLM_LKF_CANCEL, LKM_CANCEL); - map_flag(DLM_LKF_CONVERT, LKM_CONVERT); - map_flag(DLM_LKF_VALBLK, LKM_VALBLK); - map_flag(DLM_LKF_IVVALBLK, LKM_INVVALBLK); - map_flag(DLM_LKF_ORPHAN, LKM_ORPHAN); - map_flag(DLM_LKF_FORCEUNLOCK, LKM_FORCE); - map_flag(DLM_LKF_TIMEOUT, LKM_TIMEOUT); - map_flag(DLM_LKF_LOCAL, LKM_LOCAL); - - /* map_flag() should have cleared every flag passed in */ - BUG_ON(flags != 0); - - return o2dlm_flags; -} -#undef map_flag - -/* - * Map an o2dlm status to standard errno values. - * - * o2dlm only uses a handful of these, and returns even fewer to the - * caller. Still, we try to assign sane values to each error. - * - * The following value pairs have special meanings to dlmglue, thus - * the right hand side needs to stay unique - never duplicate the - * mapping elsewhere in the table! - * - * DLM_NORMAL: 0 - * DLM_NOTQUEUED: -EAGAIN - * DLM_CANCELGRANT: -EBUSY - * DLM_CANCEL: -DLM_ECANCEL - */ -/* Keep in sync with dlmapi.h */ -static int status_map[] = { - [DLM_NORMAL] = 0, /* Success */ - [DLM_GRANTED] = -EINVAL, - [DLM_DENIED] = -EACCES, - [DLM_DENIED_NOLOCKS] = -EACCES, - [DLM_WORKING] = -EACCES, - [DLM_BLOCKED] = -EINVAL, - [DLM_BLOCKED_ORPHAN] = -EINVAL, - [DLM_DENIED_GRACE_PERIOD] = -EACCES, - [DLM_SYSERR] = -ENOMEM, /* It is what it is */ - [DLM_NOSUPPORT] = -EPROTO, - [DLM_CANCELGRANT] = -EBUSY, /* Cancel after grant */ - [DLM_IVLOCKID] = -EINVAL, - [DLM_SYNC] = -EINVAL, - [DLM_BADTYPE] = -EINVAL, - [DLM_BADRESOURCE] = -EINVAL, - [DLM_MAXHANDLES] = -ENOMEM, - [DLM_NOCLINFO] = -EINVAL, - [DLM_NOLOCKMGR] = -EINVAL, - [DLM_NOPURGED] = -EINVAL, - [DLM_BADARGS] = -EINVAL, - [DLM_VOID] = -EINVAL, - [DLM_NOTQUEUED] = -EAGAIN, /* Trylock failed */ - [DLM_IVBUFLEN] = -EINVAL, - [DLM_CVTUNGRANT] = -EPERM, - [DLM_BADPARAM] = -EINVAL, - [DLM_VALNOTVALID] = -EINVAL, - [DLM_REJECTED] = -EPERM, - [DLM_ABORT] = -EINVAL, - [DLM_CANCEL] = -DLM_ECANCEL, /* Successful cancel */ - [DLM_IVRESHANDLE] = -EINVAL, - [DLM_DEADLOCK] = -EDEADLK, - [DLM_DENIED_NOASTS] = -EINVAL, - [DLM_FORWARD] = -EINVAL, - [DLM_TIMEOUT] = -ETIMEDOUT, - [DLM_IVGROUPID] = -EINVAL, - [DLM_VERS_CONFLICT] = -EOPNOTSUPP, - [DLM_BAD_DEVICE_PATH] = -ENOENT, - [DLM_NO_DEVICE_PERMISSION] = -EPERM, - [DLM_NO_CONTROL_DEVICE] = -ENOENT, - [DLM_RECOVERING] = -ENOTCONN, - [DLM_MIGRATING] = -ERESTART, - [DLM_MAXSTATS] = -EINVAL, -}; - -static int dlm_status_to_errno(enum dlm_status status) -{ - BUG_ON(status > (sizeof(status_map) / sizeof(status_map[0]))); +struct ocfs2_locking_protocol *stack_glue_lproto; - return status_map[status]; -} - -static void o2dlm_lock_ast_wrapper(void *astarg) -{ - BUG_ON(lproto == NULL); - - lproto->lp_lock_ast(astarg); -} - -static void o2dlm_blocking_ast_wrapper(void *astarg, int level) -{ - BUG_ON(lproto == NULL); - - lproto->lp_blocking_ast(astarg, level); -} - -static void o2dlm_unlock_ast_wrapper(void *astarg, enum dlm_status status) -{ - int error = dlm_status_to_errno(status); - - BUG_ON(lproto == NULL); - - /* - * In o2dlm, you can get both the lock_ast() for the lock being - * granted and the unlock_ast() for the CANCEL failing. A - * successful cancel sends DLM_NORMAL here. If the - * lock grant happened before the cancel arrived, you get - * DLM_CANCELGRANT. - * - * There's no need for the double-ast. If we see DLM_CANCELGRANT, - * we just ignore it. We expect the lock_ast() to handle the - * granted lock. - */ - if (status == DLM_CANCELGRANT) - return; - - lproto->lp_unlock_ast(astarg, error); -} - -static int o2cb_dlm_lock(struct ocfs2_cluster_connection *conn, - int mode, - union ocfs2_dlm_lksb *lksb, - u32 flags, - void *name, - unsigned int namelen, - void *astarg) -{ - enum dlm_status status; - int o2dlm_mode = mode_to_o2dlm(mode); - int o2dlm_flags = flags_to_o2dlm(flags); - int ret; - - status = dlmlock(conn->cc_lockspace, o2dlm_mode, &lksb->lksb_o2dlm, - o2dlm_flags, name, namelen, - o2dlm_lock_ast_wrapper, astarg, - o2dlm_blocking_ast_wrapper); - ret = dlm_status_to_errno(status); - return ret; -} int ocfs2_dlm_lock(struct ocfs2_cluster_connection *conn, int mode, @@ -226,25 +39,10 @@ int ocfs2_dlm_lock(struct ocfs2_cluster_connection *conn, unsigned int namelen, void *astarg) { - BUG_ON(lproto == NULL); - - return o2cb_dlm_lock(conn, mode, lksb, flags, - name, namelen, astarg); -} - -static int o2cb_dlm_unlock(struct ocfs2_cluster_connection *conn, - union ocfs2_dlm_lksb *lksb, - u32 flags, - void *astarg) -{ - enum dlm_status status; - int o2dlm_flags = flags_to_o2dlm(flags); - int ret; + BUG_ON(stack_glue_lproto == NULL); - status = dlmunlock(conn->cc_lockspace, &lksb->lksb_o2dlm, - o2dlm_flags, o2dlm_unlock_ast_wrapper, astarg); - ret = dlm_status_to_errno(status); - return ret; + return o2cb_stack_ops.dlm_lock(conn, mode, lksb, flags, + name, namelen, astarg); } int ocfs2_dlm_unlock(struct ocfs2_cluster_connection *conn, @@ -252,19 +50,14 @@ int ocfs2_dlm_unlock(struct ocfs2_cluster_connection *conn, u32 flags, void *astarg) { - BUG_ON(lproto == NULL); + BUG_ON(stack_glue_lproto == NULL); - return o2cb_dlm_unlock(conn, lksb, flags, astarg); -} - -static int o2cb_dlm_lock_status(union ocfs2_dlm_lksb *lksb) -{ - return dlm_status_to_errno(lksb->lksb_o2dlm.status); + return o2cb_stack_ops.dlm_unlock(conn, lksb, flags, astarg); } int ocfs2_dlm_lock_status(union ocfs2_dlm_lksb *lksb) { - return o2cb_dlm_lock_status(lksb); + return o2cb_stack_ops.lock_status(lksb); } /* @@ -272,94 +65,14 @@ int ocfs2_dlm_lock_status(union ocfs2_dlm_lksb *lksb) * don't cast at the glue level. The real answer is that the header * ordering is nigh impossible. */ -static void *o2cb_dlm_lvb(union ocfs2_dlm_lksb *lksb) -{ - return (void *)(lksb->lksb_o2dlm.lvb); -} - void *ocfs2_dlm_lvb(union ocfs2_dlm_lksb *lksb) { - return o2cb_dlm_lvb(lksb); -} - -static void o2cb_dlm_dump_lksb(union ocfs2_dlm_lksb *lksb) -{ - dlm_print_one_lock(lksb->lksb_o2dlm.lockid); + return o2cb_stack_ops.lock_lvb(lksb); } void ocfs2_dlm_dump_lksb(union ocfs2_dlm_lksb *lksb) { - o2cb_dlm_dump_lksb(lksb); -} - -/* - * Called from the dlm when it's about to evict a node. This is how the - * classic stack signals node death. - */ -static void o2dlm_eviction_cb(int node_num, void *data) -{ - struct ocfs2_cluster_connection *conn = data; - - mlog(ML_NOTICE, "o2dlm has evicted node %d from group %.*s\n", - node_num, conn->cc_namelen, conn->cc_name); - - conn->cc_recovery_handler(node_num, conn->cc_recovery_data); -} - -static int o2cb_cluster_connect(struct ocfs2_cluster_connection *conn) -{ - int rc = 0; - u32 dlm_key; - struct dlm_ctxt *dlm; - struct o2dlm_private *priv; - struct dlm_protocol_version dlm_version; - - BUG_ON(conn == NULL); - - /* for now we only have one cluster/node, make sure we see it - * in the heartbeat universe */ - if (!o2hb_check_local_node_heartbeating()) { - rc = -EINVAL; - goto out; - } - - priv = kzalloc(sizeof(struct o2dlm_private), GFP_KERNEL); - if (!priv) { - rc = -ENOMEM; - goto out_free; - } - - /* This just fills the structure in. It is safe to pass conn. */ - dlm_setup_eviction_cb(&priv->op_eviction_cb, o2dlm_eviction_cb, - conn); - - conn->cc_private = priv; - - /* used by the dlm code to make message headers unique, each - * node in this domain must agree on this. */ - dlm_key = crc32_le(0, conn->cc_name, conn->cc_namelen); - dlm_version.pv_major = conn->cc_version.pv_major; - dlm_version.pv_minor = conn->cc_version.pv_minor; - - dlm = dlm_register_domain(conn->cc_name, dlm_key, &dlm_version); - if (IS_ERR(dlm)) { - rc = PTR_ERR(dlm); - mlog_errno(rc); - goto out_free; - } - - conn->cc_version.pv_major = dlm_version.pv_major; - conn->cc_version.pv_minor = dlm_version.pv_minor; - conn->cc_lockspace = dlm; - - dlm_register_eviction_cb(dlm, &priv->op_eviction_cb); - -out_free: - if (rc && conn->cc_private) - kfree(conn->cc_private); - -out: - return rc; + o2cb_stack_ops.dump_lksb(lksb); } int ocfs2_cluster_connect(const char *group, @@ -394,9 +107,9 @@ int ocfs2_cluster_connect(const char *group, new_conn->cc_recovery_data = recovery_data; /* Start the new connection at our maximum compatibility level */ - new_conn->cc_version = lproto->lp_max_version; + new_conn->cc_version = stack_glue_lproto->lp_max_version; - rc = o2cb_cluster_connect(new_conn); + rc = o2cb_stack_ops.connect(new_conn); if (rc) { mlog_errno(rc); goto out_free; @@ -412,29 +125,13 @@ out: return rc; } - -static int o2cb_cluster_disconnect(struct ocfs2_cluster_connection *conn) -{ - struct dlm_ctxt *dlm = conn->cc_lockspace; - struct o2dlm_private *priv = conn->cc_private; - - dlm_unregister_eviction_cb(&priv->op_eviction_cb); - conn->cc_private = NULL; - kfree(priv); - - dlm_unregister_domain(dlm); - conn->cc_lockspace = NULL; - - return 0; -} - int ocfs2_cluster_disconnect(struct ocfs2_cluster_connection *conn) { int ret; BUG_ON(conn == NULL); - ret = o2cb_cluster_disconnect(conn); + ret = o2cb_stack_ops.disconnect(conn); /* XXX Should we free it anyway? */ if (!ret) @@ -443,75 +140,23 @@ int ocfs2_cluster_disconnect(struct ocfs2_cluster_connection *conn) return ret; } -static void o2hb_stop(const char *group) -{ - int ret; - char *argv[5], *envp[3]; - - argv[0] = (char *)o2nm_get_hb_ctl_path(); - argv[1] = "-K"; - argv[2] = "-u"; - argv[3] = (char *)group; - argv[4] = NULL; - - mlog(0, "Run: %s %s %s %s\n", argv[0], argv[1], argv[2], argv[3]); - - /* minimal command environment taken from cpu_run_sbin_hotplug */ - envp[0] = "HOME=/"; - envp[1] = "PATH=/sbin:/bin:/usr/sbin:/usr/bin"; - envp[2] = NULL; - - ret = call_usermodehelper(argv[0], argv, envp, UMH_WAIT_PROC); - if (ret < 0) - mlog_errno(ret); -} - -/* - * Hangup is a hack for tools compatibility. Older ocfs2-tools software - * expects the filesystem to call "ocfs2_hb_ctl" during unmount. This - * happens regardless of whether the DLM got started, so we can't do it - * in ocfs2_cluster_disconnect(). We bring the o2hb_stop() function into - * the glue and provide a "hangup" API for super.c to call. - * - * Other stacks will eventually provide a NULL ->hangup() pointer. - */ -static void o2cb_cluster_hangup(const char *group, int grouplen) -{ - o2hb_stop(group); -} - void ocfs2_cluster_hangup(const char *group, int grouplen) { BUG_ON(group == NULL); BUG_ON(group[grouplen] != '\0'); - o2cb_cluster_hangup(group, grouplen); -} - -static int o2cb_cluster_this_node(unsigned int *node) -{ - int node_num; - - node_num = o2nm_this_node(); - if (node_num == O2NM_INVALID_NODE_NUM) - return -ENOENT; - - if (node_num >= O2NM_MAX_NODES) - return -EOVERFLOW; - - *node = node_num; - return 0; + o2cb_stack_ops.hangup(group, grouplen); } int ocfs2_cluster_this_node(unsigned int *node) { - return o2cb_cluster_this_node(node); + return o2cb_stack_ops.this_node(node); } void ocfs2_stack_glue_set_locking_protocol(struct ocfs2_locking_protocol *proto) { BUG_ON(proto != NULL); - lproto = proto; + stack_glue_lproto = proto; } diff --git a/fs/ocfs2/stackglue.h b/fs/ocfs2/stackglue.h index decb147106f..083632215dc 100644 --- a/fs/ocfs2/stackglue.h +++ b/fs/ocfs2/stackglue.h @@ -25,6 +25,8 @@ #include #include +#include "dlm/dlmapi.h" + /* * dlmconstants.h does not have a LOCAL flag. We hope to remove it * some day, but right now we need it. Let's fake it. This value is larger @@ -39,13 +41,18 @@ #define GROUP_NAME_MAX 64 -#include "dlm/dlmapi.h" - +/* + * ocfs2_protocol_version changes when ocfs2 does something different in + * its inter-node behavior. See dlmglue.c for more information. + */ struct ocfs2_protocol_version { u8 pv_major; u8 pv_minor; }; +/* + * The ocfs2_locking_protocol defines the handlers called on ocfs2's behalf. + */ struct ocfs2_locking_protocol { struct ocfs2_protocol_version lp_max_version; void (*lp_lock_ast)(void *astarg); @@ -53,10 +60,20 @@ struct ocfs2_locking_protocol { void (*lp_unlock_ast)(void *astarg, int error); }; +/* + * A union of all lock status structures. We define it here so that the + * size of the union is known. Lock status structures are embedded in + * ocfs2 inodes. + */ union ocfs2_dlm_lksb { struct dlm_lockstatus lksb_o2dlm; }; +/* + * A cluster connection. Mostly opaque to ocfs2, the connection holds + * state for the underlying stack. ocfs2 does use cc_version to determine + * locking compatibility. + */ struct ocfs2_cluster_connection { char cc_name[GROUP_NAME_MAX]; int cc_namelen; @@ -67,6 +84,106 @@ struct ocfs2_cluster_connection { void *cc_private; }; +/* + * Each cluster stack implements the stack operations structure. Not used + * in the ocfs2 code, the stackglue code translates generic cluster calls + * into stack operations. + */ +struct ocfs2_stack_operations { + /* + * The fs code calls ocfs2_cluster_connect() to attach a new + * filesystem to the cluster stack. The ->connect() op is passed + * an ocfs2_cluster_connection with the name and recovery field + * filled in. + * + * The stack must set up any notification mechanisms and create + * the filesystem lockspace in the DLM. The lockspace should be + * stored on cc_lockspace. Any other information can be stored on + * cc_private. + * + * ->connect() must not return until it is guaranteed that + * + * - Node down notifications for the filesystem will be recieved + * and passed to conn->cc_recovery_handler(). + * - Locking requests for the filesystem will be processed. + */ + int (*connect)(struct ocfs2_cluster_connection *conn); + + /* + * The fs code calls ocfs2_cluster_disconnect() when a filesystem + * no longer needs cluster services. All DLM locks have been + * dropped, and recovery notification is being ignored by the + * fs code. The stack must disengage from the DLM and discontinue + * recovery notification. + * + * Once ->disconnect() has returned, the connection structure will + * be freed. Thus, a stack must not return from ->disconnect() + * until it will no longer reference the conn pointer. + */ + int (*disconnect)(struct ocfs2_cluster_connection *conn); + + /* + * ocfs2_cluster_hangup() exists for compatibility with older + * ocfs2 tools. Only the classic stack really needs it. As such + * ->hangup() is not required of all stacks. See the comment by + * ocfs2_cluster_hangup() for more details. + */ + void (*hangup)(const char *group, int grouplen); + + /* + * ->this_node() returns the cluster's unique identifier for the + * local node. + */ + int (*this_node)(unsigned int *node); + + /* + * Call the underlying dlm lock function. The ->dlm_lock() + * callback should convert the flags and mode as appropriate. + * + * ast and bast functions are not part of the call because the + * stack will likely want to wrap ast and bast calls before passing + * them to stack->sp_proto. + */ + int (*dlm_lock)(struct ocfs2_cluster_connection *conn, + int mode, + union ocfs2_dlm_lksb *lksb, + u32 flags, + void *name, + unsigned int namelen, + void *astarg); + + /* + * Call the underlying dlm unlock function. The ->dlm_unlock() + * function should convert the flags as appropriate. + * + * The unlock ast is not passed, as the stack will want to wrap + * it before calling stack->sp_proto->lp_unlock_ast(). + */ + int (*dlm_unlock)(struct ocfs2_cluster_connection *conn, + union ocfs2_dlm_lksb *lksb, + u32 flags, + void *astarg); + + /* + * Return the status of the current lock status block. The fs + * code should never dereference the union. The ->lock_status() + * callback pulls out the stack-specific lksb, converts the status + * to a proper errno, and returns it. + */ + int (*lock_status)(union ocfs2_dlm_lksb *lksb); + + /* + * Pull the lvb pointer off of the stack-specific lksb. + */ + void *(*lock_lvb)(union ocfs2_dlm_lksb *lksb); + + /* + * This is an optoinal debugging hook. If provided, the + * stack can dump debugging information about this lock. + */ + void (*dump_lksb)(union ocfs2_dlm_lksb *lksb); +}; + int ocfs2_cluster_connect(const char *group, int grouplen, void (*recovery_handler)(int node_num, @@ -95,4 +212,6 @@ void ocfs2_dlm_dump_lksb(union ocfs2_dlm_lksb *lksb); void ocfs2_stack_glue_set_locking_protocol(struct ocfs2_locking_protocol *proto); +extern struct ocfs2_locking_protocol *stack_glue_lproto; +extern struct ocfs2_stack_operations o2cb_stack_ops; #endif /* STACKGLUE_H */ -- cgit v1.2.3 From 286eaa95c5c5915a6b72cc3f0a2534161fd7928b Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Fri, 1 Feb 2008 15:03:57 -0800 Subject: ocfs2: Break out stackglue into modules. We define the ocfs2_stack_plugin structure to represent a stack driver. The o2cb stack code is split into stack_o2cb.c. This becomes the ocfs2_stack_o2cb.ko module. The stackglue generic functions are similarly split into the ocfs2_stackglue.ko module. This module now provides an interface to register drivers. The ocfs2_stack_o2cb driver registers itself. As part of this interface, ocfs2_stackglue can load drivers on demand. This is accomplished in ocfs2_cluster_connect(). ocfs2_cluster_disconnect() is now notified when a _hangup() is pending. If a hangup is pending, it will not release the driver module and will let _hangup() do that. Signed-off-by: Joel Becker --- fs/ocfs2/Makefile | 7 +- fs/ocfs2/dlmglue.c | 7 +- fs/ocfs2/dlmglue.h | 2 +- fs/ocfs2/stack_o2cb.c | 41 +++++++-- fs/ocfs2/stackglue.c | 238 +++++++++++++++++++++++++++++++++++++++++++++----- fs/ocfs2/stackglue.h | 36 +++++++- fs/ocfs2/super.c | 16 ++-- 7 files changed, 297 insertions(+), 50 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/Makefile b/fs/ocfs2/Makefile index 8e86195fedb..b734254329b 100644 --- a/fs/ocfs2/Makefile +++ b/fs/ocfs2/Makefile @@ -2,7 +2,7 @@ EXTRA_CFLAGS += -Ifs/ocfs2 EXTRA_CFLAGS += -DCATCH_BH_JBD_RACES -obj-$(CONFIG_OCFS2_FS) += ocfs2.o +obj-$(CONFIG_OCFS2_FS) += ocfs2.o ocfs2_stackglue.o ocfs2_stack_o2cb.o ocfs2-objs := \ alloc.o \ @@ -24,8 +24,6 @@ ocfs2-objs := \ namei.o \ resize.o \ slot_map.o \ - stackglue.o \ - stack_o2cb.o \ suballoc.o \ super.o \ symlink.o \ @@ -33,5 +31,8 @@ ocfs2-objs := \ uptodate.o \ ver.o +ocfs2_stackglue-objs := stackglue.o +ocfs2_stack_o2cb-objs := stack_o2cb.o + obj-$(CONFIG_OCFS2_FS) += cluster/ obj-$(CONFIG_OCFS2_FS) += dlm/ diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index 8a9c84909be..f62a9e4fc31 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -2641,7 +2641,7 @@ int ocfs2_dlm_init(struct ocfs2_super *osb) mlog_errno(status); mlog(ML_ERROR, "could not find this host's node number\n"); - ocfs2_cluster_disconnect(conn); + ocfs2_cluster_disconnect(conn, 0); goto bail; } @@ -2663,7 +2663,8 @@ bail: return status; } -void ocfs2_dlm_shutdown(struct ocfs2_super *osb) +void ocfs2_dlm_shutdown(struct ocfs2_super *osb, + int hangup_pending) { mlog_entry_void(); @@ -2683,7 +2684,7 @@ void ocfs2_dlm_shutdown(struct ocfs2_super *osb) ocfs2_lock_res_free(&osb->osb_super_lockres); ocfs2_lock_res_free(&osb->osb_rename_lockres); - ocfs2_cluster_disconnect(osb->cconn); + ocfs2_cluster_disconnect(osb->cconn, hangup_pending); osb->cconn = NULL; ocfs2_dlm_shutdown_debug(osb); diff --git a/fs/ocfs2/dlmglue.h b/fs/ocfs2/dlmglue.h index 34b7598a0dc..2bb01f09c1b 100644 --- a/fs/ocfs2/dlmglue.h +++ b/fs/ocfs2/dlmglue.h @@ -58,7 +58,7 @@ struct ocfs2_meta_lvb { #define OCFS2_LOCK_NONBLOCK (0x04) int ocfs2_dlm_init(struct ocfs2_super *osb); -void ocfs2_dlm_shutdown(struct ocfs2_super *osb); +void ocfs2_dlm_shutdown(struct ocfs2_super *osb, int hangup_pending); void ocfs2_lock_res_init_once(struct ocfs2_lock_res *res); void ocfs2_inode_lock_res_init(struct ocfs2_lock_res *res, enum ocfs2_lock_type type, diff --git a/fs/ocfs2/stack_o2cb.c b/fs/ocfs2/stack_o2cb.c index c9bc3541a33..ac1d74c63bf 100644 --- a/fs/ocfs2/stack_o2cb.c +++ b/fs/ocfs2/stack_o2cb.c @@ -18,7 +18,7 @@ */ #include -#include +#include /* Needed for AOP_TRUNCATED_PAGE in mlog_errno() */ #include @@ -33,6 +33,8 @@ struct o2dlm_private { struct dlm_eviction_cb op_eviction_cb; }; +static struct ocfs2_stack_plugin o2cb_stack; + /* These should be identical */ #if (DLM_LOCK_IV != LKM_IVMODE) # error Lock modes do not match @@ -158,23 +160,23 @@ static int dlm_status_to_errno(enum dlm_status status) static void o2dlm_lock_ast_wrapper(void *astarg) { - BUG_ON(stack_glue_lproto == NULL); + BUG_ON(o2cb_stack.sp_proto == NULL); - stack_glue_lproto->lp_lock_ast(astarg); + o2cb_stack.sp_proto->lp_lock_ast(astarg); } static void o2dlm_blocking_ast_wrapper(void *astarg, int level) { - BUG_ON(stack_glue_lproto == NULL); + BUG_ON(o2cb_stack.sp_proto == NULL); - stack_glue_lproto->lp_blocking_ast(astarg, level); + o2cb_stack.sp_proto->lp_blocking_ast(astarg, level); } static void o2dlm_unlock_ast_wrapper(void *astarg, enum dlm_status status) { int error = dlm_status_to_errno(status); - BUG_ON(stack_glue_lproto == NULL); + BUG_ON(o2cb_stack.sp_proto == NULL); /* * In o2dlm, you can get both the lock_ast() for the lock being @@ -190,7 +192,7 @@ static void o2dlm_unlock_ast_wrapper(void *astarg, enum dlm_status status) if (status == DLM_CANCELGRANT) return; - stack_glue_lproto->lp_unlock_ast(astarg, error); + o2cb_stack.sp_proto->lp_unlock_ast(astarg, error); } static int o2cb_dlm_lock(struct ocfs2_cluster_connection *conn, @@ -267,6 +269,7 @@ static int o2cb_cluster_connect(struct ocfs2_cluster_connection *conn) struct dlm_protocol_version dlm_version; BUG_ON(conn == NULL); + BUG_ON(o2cb_stack.sp_proto == NULL); /* for now we only have one cluster/node, make sure we see it * in the heartbeat universe */ @@ -314,7 +317,8 @@ out: return rc; } -static int o2cb_cluster_disconnect(struct ocfs2_cluster_connection *conn) +static int o2cb_cluster_disconnect(struct ocfs2_cluster_connection *conn, + int hangup_pending) { struct dlm_ctxt *dlm = conn->cc_lockspace; struct o2dlm_private *priv = conn->cc_private; @@ -393,3 +397,24 @@ struct ocfs2_stack_operations o2cb_stack_ops = { .dump_lksb = o2cb_dump_lksb, }; +static struct ocfs2_stack_plugin o2cb_stack = { + .sp_name = "o2cb", + .sp_ops = &o2cb_stack_ops, + .sp_owner = THIS_MODULE, +}; + +static int __init o2cb_stack_init(void) +{ + return ocfs2_stack_glue_register(&o2cb_stack); +} + +static void __exit o2cb_stack_exit(void) +{ + ocfs2_stack_glue_unregister(&o2cb_stack); +} + +MODULE_AUTHOR("Oracle"); +MODULE_DESCRIPTION("ocfs2 driver for the classic o2cb stack"); +MODULE_LICENSE("GPL"); +module_init(o2cb_stack_init); +module_exit(o2cb_stack_exit); diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c index e197367b6bd..1978c9cff0e 100644 --- a/fs/ocfs2/stackglue.c +++ b/fs/ocfs2/stackglue.c @@ -18,17 +18,176 @@ * General Public License for more details. */ +#include +#include +#include #include #include -/* Needed for AOP_TRUNCATED_PAGE in mlog_errno() */ -#include +#include "stackglue.h" -#include "cluster/masklog.h" +static struct ocfs2_locking_protocol *lproto; +static DEFINE_SPINLOCK(ocfs2_stack_lock); +static LIST_HEAD(ocfs2_stack_list); -#include "stackglue.h" +/* + * The stack currently in use. If not null, active_stack->sp_count > 0, + * the module is pinned, and the locking protocol cannot be changed. + */ +static struct ocfs2_stack_plugin *active_stack; + +static struct ocfs2_stack_plugin *ocfs2_stack_lookup(const char *name) +{ + struct ocfs2_stack_plugin *p; + + assert_spin_locked(&ocfs2_stack_lock); + + list_for_each_entry(p, &ocfs2_stack_list, sp_list) { + if (!strcmp(p->sp_name, name)) + return p; + } + + return NULL; +} + +static int ocfs2_stack_driver_request(const char *name) +{ + int rc; + struct ocfs2_stack_plugin *p; + + spin_lock(&ocfs2_stack_lock); + + if (active_stack) { + /* + * If the active stack isn't the one we want, it cannot + * be selected right now. + */ + if (!strcmp(active_stack->sp_name, name)) + rc = 0; + else + rc = -EBUSY; + goto out; + } + + p = ocfs2_stack_lookup(name); + if (!p || !try_module_get(p->sp_owner)) { + rc = -ENOENT; + goto out; + } + + /* Ok, the stack is pinned */ + p->sp_count++; + active_stack = p; + + rc = 0; + +out: + spin_unlock(&ocfs2_stack_lock); + return rc; +} + +/* + * This function looks up the appropriate stack and makes it active. If + * there is no stack, it tries to load it. It will fail if the stack still + * cannot be found. It will also fail if a different stack is in use. + */ +static int ocfs2_stack_driver_get(const char *name) +{ + int rc; + + rc = ocfs2_stack_driver_request(name); + if (rc == -ENOENT) { + request_module("ocfs2_stack_%s", name); + rc = ocfs2_stack_driver_request(name); + } + + if (rc == -ENOENT) { + printk(KERN_ERR + "ocfs2: Cluster stack driver \"%s\" cannot be found\n", + name); + } else if (rc == -EBUSY) { + printk(KERN_ERR + "ocfs2: A different cluster stack driver is in use\n"); + } + + return rc; +} -struct ocfs2_locking_protocol *stack_glue_lproto; +static void ocfs2_stack_driver_put(void) +{ + spin_lock(&ocfs2_stack_lock); + BUG_ON(active_stack == NULL); + BUG_ON(active_stack->sp_count == 0); + + active_stack->sp_count--; + if (!active_stack->sp_count) { + module_put(active_stack->sp_owner); + active_stack = NULL; + } + spin_unlock(&ocfs2_stack_lock); +} + +int ocfs2_stack_glue_register(struct ocfs2_stack_plugin *plugin) +{ + int rc; + + spin_lock(&ocfs2_stack_lock); + if (!ocfs2_stack_lookup(plugin->sp_name)) { + plugin->sp_count = 0; + plugin->sp_proto = lproto; + list_add(&plugin->sp_list, &ocfs2_stack_list); + printk(KERN_INFO "ocfs2: Registered cluster interface %s\n", + plugin->sp_name); + rc = 0; + } else { + printk(KERN_ERR "ocfs2: Stack \"%s\" already registered\n", + plugin->sp_name); + rc = -EEXIST; + } + spin_unlock(&ocfs2_stack_lock); + + return rc; +} +EXPORT_SYMBOL_GPL(ocfs2_stack_glue_register); + +void ocfs2_stack_glue_unregister(struct ocfs2_stack_plugin *plugin) +{ + struct ocfs2_stack_plugin *p; + + spin_lock(&ocfs2_stack_lock); + p = ocfs2_stack_lookup(plugin->sp_name); + if (p) { + BUG_ON(p != plugin); + BUG_ON(plugin == active_stack); + BUG_ON(plugin->sp_count != 0); + list_del_init(&plugin->sp_list); + printk(KERN_INFO "ocfs2: Unregistered cluster interface %s\n", + plugin->sp_name); + } else { + printk(KERN_ERR "Stack \"%s\" is not registered\n", + plugin->sp_name); + } + spin_unlock(&ocfs2_stack_lock); +} +EXPORT_SYMBOL_GPL(ocfs2_stack_glue_unregister); + +void ocfs2_stack_glue_set_locking_protocol(struct ocfs2_locking_protocol *proto) +{ + struct ocfs2_stack_plugin *p; + + BUG_ON(proto == NULL); + + spin_lock(&ocfs2_stack_lock); + BUG_ON(active_stack != NULL); + + lproto = proto; + list_for_each_entry(p, &ocfs2_stack_list, sp_list) { + p->sp_proto = lproto; + } + + spin_unlock(&ocfs2_stack_lock); +} +EXPORT_SYMBOL_GPL(ocfs2_stack_glue_set_locking_protocol); int ocfs2_dlm_lock(struct ocfs2_cluster_connection *conn, @@ -39,26 +198,29 @@ int ocfs2_dlm_lock(struct ocfs2_cluster_connection *conn, unsigned int namelen, void *astarg) { - BUG_ON(stack_glue_lproto == NULL); + BUG_ON(lproto == NULL); - return o2cb_stack_ops.dlm_lock(conn, mode, lksb, flags, - name, namelen, astarg); + return active_stack->sp_ops->dlm_lock(conn, mode, lksb, flags, + name, namelen, astarg); } +EXPORT_SYMBOL_GPL(ocfs2_dlm_lock); int ocfs2_dlm_unlock(struct ocfs2_cluster_connection *conn, union ocfs2_dlm_lksb *lksb, u32 flags, void *astarg) { - BUG_ON(stack_glue_lproto == NULL); + BUG_ON(lproto == NULL); - return o2cb_stack_ops.dlm_unlock(conn, lksb, flags, astarg); + return active_stack->sp_ops->dlm_unlock(conn, lksb, flags, astarg); } +EXPORT_SYMBOL_GPL(ocfs2_dlm_unlock); int ocfs2_dlm_lock_status(union ocfs2_dlm_lksb *lksb) { - return o2cb_stack_ops.lock_status(lksb); + return active_stack->sp_ops->lock_status(lksb); } +EXPORT_SYMBOL_GPL(ocfs2_dlm_lock_status); /* * Why don't we cast to ocfs2_meta_lvb? The "clean" answer is that we @@ -67,13 +229,15 @@ int ocfs2_dlm_lock_status(union ocfs2_dlm_lksb *lksb) */ void *ocfs2_dlm_lvb(union ocfs2_dlm_lksb *lksb) { - return o2cb_stack_ops.lock_lvb(lksb); + return active_stack->sp_ops->lock_lvb(lksb); } +EXPORT_SYMBOL_GPL(ocfs2_dlm_lvb); void ocfs2_dlm_dump_lksb(union ocfs2_dlm_lksb *lksb) { - o2cb_stack_ops.dump_lksb(lksb); + active_stack->sp_ops->dump_lksb(lksb); } +EXPORT_SYMBOL_GPL(ocfs2_dlm_dump_lksb); int ocfs2_cluster_connect(const char *group, int grouplen, @@ -107,11 +271,16 @@ int ocfs2_cluster_connect(const char *group, new_conn->cc_recovery_data = recovery_data; /* Start the new connection at our maximum compatibility level */ - new_conn->cc_version = stack_glue_lproto->lp_max_version; + new_conn->cc_version = lproto->lp_max_version; + + /* This will pin the stack driver if successful */ + rc = ocfs2_stack_driver_get("o2cb"); + if (rc) + goto out_free; - rc = o2cb_stack_ops.connect(new_conn); + rc = active_stack->sp_ops->connect(new_conn); if (rc) { - mlog_errno(rc); + ocfs2_stack_driver_put(); goto out_free; } @@ -124,39 +293,60 @@ out_free: out: return rc; } +EXPORT_SYMBOL_GPL(ocfs2_cluster_connect); -int ocfs2_cluster_disconnect(struct ocfs2_cluster_connection *conn) +/* If hangup_pending is 0, the stack driver will be dropped */ +int ocfs2_cluster_disconnect(struct ocfs2_cluster_connection *conn, + int hangup_pending) { int ret; BUG_ON(conn == NULL); - ret = o2cb_stack_ops.disconnect(conn); + ret = active_stack->sp_ops->disconnect(conn, hangup_pending); /* XXX Should we free it anyway? */ - if (!ret) + if (!ret) { kfree(conn); + if (!hangup_pending) + ocfs2_stack_driver_put(); + } return ret; } +EXPORT_SYMBOL_GPL(ocfs2_cluster_disconnect); void ocfs2_cluster_hangup(const char *group, int grouplen) { BUG_ON(group == NULL); BUG_ON(group[grouplen] != '\0'); - o2cb_stack_ops.hangup(group, grouplen); + active_stack->sp_ops->hangup(group, grouplen); + + /* cluster_disconnect() was called with hangup_pending==1 */ + ocfs2_stack_driver_put(); } +EXPORT_SYMBOL_GPL(ocfs2_cluster_hangup); int ocfs2_cluster_this_node(unsigned int *node) { - return o2cb_stack_ops.this_node(node); + return active_stack->sp_ops->this_node(node); } +EXPORT_SYMBOL_GPL(ocfs2_cluster_this_node); -void ocfs2_stack_glue_set_locking_protocol(struct ocfs2_locking_protocol *proto) + +static int __init ocfs2_stack_glue_init(void) { - BUG_ON(proto != NULL); + return 0; +} - stack_glue_lproto = proto; +static void __exit ocfs2_stack_glue_exit(void) +{ + lproto = NULL; } +MODULE_AUTHOR("Oracle"); +MODULE_DESCRIPTION("ocfs2 cluter stack glue layer"); +MODULE_LICENSE("GPL"); +module_init(ocfs2_stack_glue_init); +module_exit(ocfs2_stack_glue_exit); diff --git a/fs/ocfs2/stackglue.h b/fs/ocfs2/stackglue.h index 083632215dc..c96c8bb7686 100644 --- a/fs/ocfs2/stackglue.h +++ b/fs/ocfs2/stackglue.h @@ -119,14 +119,21 @@ struct ocfs2_stack_operations { * Once ->disconnect() has returned, the connection structure will * be freed. Thus, a stack must not return from ->disconnect() * until it will no longer reference the conn pointer. + * + * If hangup_pending is zero, ocfs2_cluster_disconnect() will also + * be dropping the reference on the module. */ - int (*disconnect)(struct ocfs2_cluster_connection *conn); + int (*disconnect)(struct ocfs2_cluster_connection *conn, + int hangup_pending); /* * ocfs2_cluster_hangup() exists for compatibility with older * ocfs2 tools. Only the classic stack really needs it. As such * ->hangup() is not required of all stacks. See the comment by * ocfs2_cluster_hangup() for more details. + * + * Note that ocfs2_cluster_hangup() can only be called if + * hangup_pending was passed to ocfs2_cluster_disconnect(). */ void (*hangup)(const char *group, int grouplen); @@ -184,13 +191,32 @@ struct ocfs2_stack_operations { void (*dump_lksb)(union ocfs2_dlm_lksb *lksb); }; +/* + * Each stack plugin must describe itself by registering a + * ocfs2_stack_plugin structure. This is only seen by stackglue and the + * stack driver. + */ +struct ocfs2_stack_plugin { + char *sp_name; + struct ocfs2_stack_operations *sp_ops; + struct module *sp_owner; + + /* These are managed by the stackglue code. */ + struct list_head sp_list; + unsigned int sp_count; + struct ocfs2_locking_protocol *sp_proto; +}; + + +/* Used by the filesystem */ int ocfs2_cluster_connect(const char *group, int grouplen, void (*recovery_handler)(int node_num, void *recovery_data), void *recovery_data, struct ocfs2_cluster_connection **conn); -int ocfs2_cluster_disconnect(struct ocfs2_cluster_connection *conn); +int ocfs2_cluster_disconnect(struct ocfs2_cluster_connection *conn, + int hangup_pending); void ocfs2_cluster_hangup(const char *group, int grouplen); int ocfs2_cluster_this_node(unsigned int *node); @@ -212,6 +238,8 @@ void ocfs2_dlm_dump_lksb(union ocfs2_dlm_lksb *lksb); void ocfs2_stack_glue_set_locking_protocol(struct ocfs2_locking_protocol *proto); -extern struct ocfs2_locking_protocol *stack_glue_lproto; -extern struct ocfs2_stack_operations o2cb_stack_ops; + +/* Used by stack plugins */ +int ocfs2_stack_glue_register(struct ocfs2_stack_plugin *plugin); +void ocfs2_stack_glue_unregister(struct ocfs2_stack_plugin *plugin); #endif /* STACKGLUE_H */ diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index b4a02a00665..e27a0d47ea2 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -1186,7 +1186,7 @@ leave: static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err) { - int tmp; + int tmp, hangup_needed = 0; struct ocfs2_super *osb = NULL; char nodestr[8]; @@ -1225,19 +1225,21 @@ static void ocfs2_dismount_volume(struct super_block *sb, int mnt_err) ocfs2_release_system_inodes(osb); - if (osb->cconn) - ocfs2_dlm_shutdown(osb); - - debugfs_remove(osb->osb_debug_root); - /* - * This is a small hack to move ocfs2_hb_ctl into stackglue. * If we're dismounting due to mount error, mount.ocfs2 will clean * up heartbeat. If we're a local mount, there is no heartbeat. * If we failed before we got a uuid_str yet, we can't stop * heartbeat. Otherwise, do it. */ if (!mnt_err && !ocfs2_mount_local(osb) && osb->uuid_str) + hangup_needed = 1; + + if (osb->cconn) + ocfs2_dlm_shutdown(osb, hangup_needed); + + debugfs_remove(osb->osb_debug_root); + + if (hangup_needed) ocfs2_cluster_hangup(osb->uuid_str, strlen(osb->uuid_str)); atomic_set(&osb->vol_state, VOLUME_DISMOUNTED); -- cgit v1.2.3 From 74ae4e104dfc57017783fc07d5f2f9129062207f Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Thu, 31 Jan 2008 23:56:17 -0800 Subject: ocfs2: Create stack glue sysfs files. Introduce a set of sysfs files that describe the current stack glue state. The files live under /sys/fs/ocfs2. The locking_protocol file displays the version of ocfs2's locking code. The loaded_cluster_plugins file displays all of the currently loaded stack plugins. When filesystems are mounted, the active_cluster_plugin file will display the plugin in use. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/stackglue.c | 121 ++++++++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 120 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c index 1978c9cff0e..76ae4fcebcb 100644 --- a/fs/ocfs2/stackglue.c +++ b/fs/ocfs2/stackglue.c @@ -23,6 +23,9 @@ #include #include #include +#include +#include +#include #include "stackglue.h" @@ -335,14 +338,130 @@ int ocfs2_cluster_this_node(unsigned int *node) EXPORT_SYMBOL_GPL(ocfs2_cluster_this_node); -static int __init ocfs2_stack_glue_init(void) +/* + * Sysfs bits + */ + +static ssize_t ocfs2_max_locking_protocol_show(struct kobject *kobj, + struct kobj_attribute *attr, + char *buf) +{ + ssize_t ret = 0; + + spin_lock(&ocfs2_stack_lock); + if (lproto) + ret = snprintf(buf, PAGE_SIZE, "%u.%u\n", + lproto->lp_max_version.pv_major, + lproto->lp_max_version.pv_minor); + spin_unlock(&ocfs2_stack_lock); + + return ret; +} + +static struct kobj_attribute ocfs2_attr_max_locking_protocol = + __ATTR(max_locking_protocol, S_IFREG | S_IRUGO, + ocfs2_max_locking_protocol_show, NULL); + +static ssize_t ocfs2_loaded_cluster_plugins_show(struct kobject *kobj, + struct kobj_attribute *attr, + char *buf) { + ssize_t ret = 0, total = 0, remain = PAGE_SIZE; + struct ocfs2_stack_plugin *p; + + spin_lock(&ocfs2_stack_lock); + list_for_each_entry(p, &ocfs2_stack_list, sp_list) { + ret = snprintf(buf, remain, "%s\n", + p->sp_name); + if (ret < 0) { + total = ret; + break; + } + if (ret == remain) { + /* snprintf() didn't fit */ + total = -E2BIG; + break; + } + total += ret; + remain -= ret; + } + spin_unlock(&ocfs2_stack_lock); + + return total; +} + +static struct kobj_attribute ocfs2_attr_loaded_cluster_plugins = + __ATTR(loaded_cluster_plugins, S_IFREG | S_IRUGO, + ocfs2_loaded_cluster_plugins_show, NULL); + +static ssize_t ocfs2_active_cluster_plugin_show(struct kobject *kobj, + struct kobj_attribute *attr, + char *buf) +{ + ssize_t ret = 0; + + spin_lock(&ocfs2_stack_lock); + if (active_stack) { + ret = snprintf(buf, PAGE_SIZE, "%s\n", + active_stack->sp_name); + if (ret == PAGE_SIZE) + ret = -E2BIG; + } + spin_unlock(&ocfs2_stack_lock); + + return ret; +} + +static struct kobj_attribute ocfs2_attr_active_cluster_plugin = + __ATTR(active_cluster_plugin, S_IFREG | S_IRUGO, + ocfs2_active_cluster_plugin_show, NULL); + +static struct attribute *ocfs2_attrs[] = { + &ocfs2_attr_max_locking_protocol.attr, + &ocfs2_attr_loaded_cluster_plugins.attr, + &ocfs2_attr_active_cluster_plugin.attr, + NULL, +}; + +static struct attribute_group ocfs2_attr_group = { + .attrs = ocfs2_attrs, +}; + +static struct kset *ocfs2_kset; + +static void ocfs2_sysfs_exit(void) +{ + kset_unregister(ocfs2_kset); +} + +static int ocfs2_sysfs_init(void) +{ + int ret; + + ocfs2_kset = kset_create_and_add("ocfs2", NULL, fs_kobj); + if (!ocfs2_kset) + return -ENOMEM; + + ret = sysfs_create_group(&ocfs2_kset->kobj, &ocfs2_attr_group); + if (ret) + goto error; + return 0; + +error: + kset_unregister(ocfs2_kset); + return ret; +} + +static int __init ocfs2_stack_glue_init(void) +{ + return ocfs2_sysfs_init(); } static void __exit ocfs2_stack_glue_exit(void) { lproto = NULL; + ocfs2_sysfs_exit(); } MODULE_AUTHOR("Oracle"); -- cgit v1.2.3 From b61817e1166c5e19c08baf05196477cc345e1b1a Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Fri, 1 Feb 2008 15:08:23 -0800 Subject: ocfs2: Add the USERSPACE_STACK incompat bit. The filesystem gains the USERSPACE_STACK incomat bit and the s_cluster_info field on the superblock. When a userspace stack is in use, the name of the stack is stored on-disk for mount-time verification. The "cluster_stack" option is added to mount(2) processing. The mount process needs to pass the matching stack name. If the passed name and the on-disk name do not match, the mount is failed. When using the classic o2cb stack, the incompat bit is *not* set and no mount option is used other than the usual heartbeat=local. Thus, the filesystem is compatible with older tools. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/ocfs2.h | 7 +++++ fs/ocfs2/ocfs2_fs.h | 40 ++++++++++++++++++++++-- fs/ocfs2/super.c | 90 ++++++++++++++++++++++++++++++++++++++++++++++++++++- 3 files changed, 134 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h index af929eca541..9ff5811345a 100644 --- a/fs/ocfs2/ocfs2.h +++ b/fs/ocfs2/ocfs2.h @@ -248,6 +248,7 @@ struct ocfs2_super struct ocfs2_alloc_stats alloc_stats; char dev_str[20]; /* "major,minor" of the device */ + char osb_cluster_stack[OCFS2_STACK_LABEL_LEN + 1]; struct ocfs2_cluster_connection *cconn; struct ocfs2_lock_res osb_super_lockres; struct ocfs2_lock_res osb_rename_lockres; @@ -368,6 +369,12 @@ static inline int ocfs2_is_soft_readonly(struct ocfs2_super *osb) return ret; } +static inline int ocfs2_userspace_stack(struct ocfs2_super *osb) +{ + return (osb->s_feature_incompat & + OCFS2_FEATURE_INCOMPAT_USERSPACE_STACK); +} + static inline int ocfs2_mount_local(struct ocfs2_super *osb) { return (osb->s_feature_incompat & OCFS2_FEATURE_INCOMPAT_LOCAL_MOUNT); diff --git a/fs/ocfs2/ocfs2_fs.h b/fs/ocfs2/ocfs2_fs.h index c49502329ab..52c42666515 100644 --- a/fs/ocfs2/ocfs2_fs.h +++ b/fs/ocfs2/ocfs2_fs.h @@ -89,7 +89,8 @@ #define OCFS2_FEATURE_INCOMPAT_SUPP (OCFS2_FEATURE_INCOMPAT_LOCAL_MOUNT \ | OCFS2_FEATURE_INCOMPAT_SPARSE_ALLOC \ | OCFS2_FEATURE_INCOMPAT_INLINE_DATA \ - | OCFS2_FEATURE_INCOMPAT_EXTENDED_SLOT_MAP) + | OCFS2_FEATURE_INCOMPAT_EXTENDED_SLOT_MAP \ + | OCFS2_FEATURE_INCOMPAT_USERSPACE_STACK) #define OCFS2_FEATURE_RO_COMPAT_SUPP OCFS2_FEATURE_RO_COMPAT_UNWRITTEN /* @@ -130,6 +131,17 @@ #define OCFS2_FEATURE_INCOMPAT_EXTENDED_SLOT_MAP 0x100 +/* + * Support for alternate, userspace cluster stacks. If set, the superblock + * field s_cluster_info contains a tag for the alternate stack in use as + * well as the name of the cluster being joined. + * mount.ocfs2 must pass in a matching stack name. + * + * If not set, the classic stack will be used. This is compatbile with + * all older versions. + */ +#define OCFS2_FEATURE_INCOMPAT_USERSPACE_STACK 0x0080 + /* * backup superblock flag is used to indicate that this volume * has backup superblocks. @@ -272,6 +284,10 @@ struct ocfs2_new_group_input { #define OCFS2_VOL_UUID_LEN 16 #define OCFS2_MAX_VOL_LABEL_LEN 64 +/* The alternate, userspace stack fields */ +#define OCFS2_STACK_LABEL_LEN 4 +#define OCFS2_CLUSTER_NAME_LEN 16 + /* Journal limits (in bytes) */ #define OCFS2_MIN_JOURNAL_SIZE (4 * 1024 * 1024) @@ -513,6 +529,13 @@ struct ocfs2_slot_map_extended { */ }; +struct ocfs2_cluster_info { +/*00*/ __u8 ci_stack[OCFS2_STACK_LABEL_LEN]; + __le32 ci_reserved; +/*08*/ __u8 ci_cluster[OCFS2_CLUSTER_NAME_LEN]; +/*18*/ +}; + /* * On disk superblock for OCFS2 * Note that it is contained inside an ocfs2_dinode, so all offsets @@ -545,7 +568,20 @@ struct ocfs2_super_block { * group header */ /*50*/ __u8 s_label[OCFS2_MAX_VOL_LABEL_LEN]; /* Label for mounting, etc. */ /*90*/ __u8 s_uuid[OCFS2_VOL_UUID_LEN]; /* 128-bit uuid */ -/*A0*/ +/*A0*/ struct ocfs2_cluster_info s_cluster_info; /* Selected userspace + stack. Only valid + with INCOMPAT flag. */ +/*B8*/ __le64 s_reserved2[17]; /* Fill out superblock */ +/*140*/ + + /* + * NOTE: As stated above, all offsets are relative to + * ocfs2_dinode.id2, which is at 0xC0 in the inode. + * 0xC0 + 0x140 = 0x200 or 512 bytes. A superblock must fit within + * our smallest blocksize, which is 512 bytes. To ensure this, + * we reserve the space in s_reserved2. Anything past s_reserved2 + * will not be available on the smallest blocksize. + */ }; /* diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index e27a0d47ea2..96ebe36d5d7 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -87,6 +87,7 @@ struct mount_options unsigned int atime_quantum; signed short slot; unsigned int localalloc_opt; + char cluster_stack[OCFS2_STACK_LABEL_LEN + 1]; }; static int ocfs2_parse_options(struct super_block *sb, char *options, @@ -152,6 +153,7 @@ enum { Opt_commit, Opt_localalloc, Opt_localflocks, + Opt_stack, Opt_err, }; @@ -170,6 +172,7 @@ static match_table_t tokens = { {Opt_commit, "commit=%u"}, {Opt_localalloc, "localalloc=%d"}, {Opt_localflocks, "localflocks"}, + {Opt_stack, "cluster_stack=%s"}, {Opt_err, NULL} }; @@ -549,8 +552,17 @@ static int ocfs2_verify_heartbeat(struct ocfs2_super *osb) } } + if (ocfs2_userspace_stack(osb)) { + if (osb->s_mount_opt & OCFS2_MOUNT_HB_LOCAL) { + mlog(ML_ERROR, "Userspace stack expected, but " + "o2cb heartbeat arguments passed to mount\n"); + return -EINVAL; + } + } + if (!(osb->s_mount_opt & OCFS2_MOUNT_HB_LOCAL)) { - if (!ocfs2_mount_local(osb) && !ocfs2_is_hard_readonly(osb)) { + if (!ocfs2_mount_local(osb) && !ocfs2_is_hard_readonly(osb) && + !ocfs2_userspace_stack(osb)) { mlog(ML_ERROR, "Heartbeat has to be started to mount " "a read-write clustered device.\n"); return -EINVAL; @@ -560,6 +572,35 @@ static int ocfs2_verify_heartbeat(struct ocfs2_super *osb) return 0; } +/* + * If we're using a userspace stack, mount should have passed + * a name that matches the disk. If not, mount should not + * have passed a stack. + */ +static int ocfs2_verify_userspace_stack(struct ocfs2_super *osb, + struct mount_options *mopt) +{ + if (!ocfs2_userspace_stack(osb) && mopt->cluster_stack[0]) { + mlog(ML_ERROR, + "cluster stack passed to mount, but this filesystem " + "does not support it\n"); + return -EINVAL; + } + + if (ocfs2_userspace_stack(osb) && + strncmp(osb->osb_cluster_stack, mopt->cluster_stack, + OCFS2_STACK_LABEL_LEN)) { + mlog(ML_ERROR, + "cluster stack passed to mount (\"%s\") does not " + "match the filesystem (\"%s\")\n", + mopt->cluster_stack, + osb->osb_cluster_stack); + return -EINVAL; + } + + return 0; +} + static int ocfs2_fill_super(struct super_block *sb, void *data, int silent) { struct dentry *root; @@ -598,6 +639,10 @@ static int ocfs2_fill_super(struct super_block *sb, void *data, int silent) osb->osb_commit_interval = parsed_options.commit_interval; osb->local_alloc_size = parsed_options.localalloc_opt; + status = ocfs2_verify_userspace_stack(osb, &parsed_options); + if (status) + goto read_super_error; + sb->s_magic = OCFS2_SUPER_MAGIC; /* Hard readonly mode only if: bdev_read_only, MS_RDONLY, @@ -752,6 +797,7 @@ static int ocfs2_parse_options(struct super_block *sb, mopt->atime_quantum = OCFS2_DEFAULT_ATIME_QUANTUM; mopt->slot = OCFS2_INVALID_SLOT; mopt->localalloc_opt = OCFS2_DEFAULT_LOCAL_ALLOC_SIZE; + mopt->cluster_stack[0] = '\0'; if (!options) { status = 1; @@ -853,6 +899,25 @@ static int ocfs2_parse_options(struct super_block *sb, if (!is_remount) mopt->mount_opt |= OCFS2_MOUNT_LOCALFLOCKS; break; + case Opt_stack: + /* Check both that the option we were passed + * is of the right length and that it is a proper + * string of the right length. + */ + if (((args[0].to - args[0].from) != + OCFS2_STACK_LABEL_LEN) || + (strnlen(args[0].from, + OCFS2_STACK_LABEL_LEN) != + OCFS2_STACK_LABEL_LEN)) { + mlog(ML_ERROR, + "Invalid cluster_stack option\n"); + status = 0; + goto bail; + } + memcpy(mopt->cluster_stack, args[0].from, + OCFS2_STACK_LABEL_LEN); + mopt->cluster_stack[OCFS2_STACK_LABEL_LEN] = '\0'; + break; default: mlog(ML_ERROR, "Unrecognized mount option \"%s\" " @@ -911,6 +976,10 @@ static int ocfs2_show_options(struct seq_file *s, struct vfsmount *mnt) if (opts & OCFS2_MOUNT_LOCALFLOCKS) seq_printf(s, ",localflocks,"); + if (osb->osb_cluster_stack[0]) + seq_printf(s, ",cluster_stack=%.*s", OCFS2_STACK_LABEL_LEN, + osb->osb_cluster_stack); + return 0; } @@ -1403,6 +1472,25 @@ static int ocfs2_initialize_super(struct super_block *sb, goto bail; } + if (ocfs2_userspace_stack(osb)) { + memcpy(osb->osb_cluster_stack, + OCFS2_RAW_SB(di)->s_cluster_info.ci_stack, + OCFS2_STACK_LABEL_LEN); + osb->osb_cluster_stack[OCFS2_STACK_LABEL_LEN] = '\0'; + if (strlen(osb->osb_cluster_stack) != OCFS2_STACK_LABEL_LEN) { + mlog(ML_ERROR, + "couldn't mount because of an invalid " + "cluster stack label (%s) \n", + osb->osb_cluster_stack); + status = -EINVAL; + goto bail; + } + } else { + /* The empty string is identical with classic tools that + * don't know about s_cluster_info. */ + osb->osb_cluster_stack[0] = '\0'; + } + get_random_bytes(&osb->s_next_generation, sizeof(u32)); /* FIXME -- cgit v1.2.3 From 9c6c877c04ce17d76a35d2173d3a3840d6b796a2 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Fri, 1 Feb 2008 15:17:30 -0800 Subject: ocfs2: Add the 'cluster_stack' sysfs file. Userspace can now query and specify the cluster stack in use via the /sys/fs/ocfs2/cluster_stack file. By default, it is 'o2cb', which is the classic stack. Thus, old tools that do not know how to modify this file will work just fine. The stack cannot be modified if there is a live filesystem. ocfs2_cluster_connect() now takes the expected cluster stack as an argument. This way, the filesystem and the stack glue ensure they are speaking to the same backend. If the stack is 'o2cb', the o2cb stack plugin is used. For any other value, the fsdlm stack plugin is selected. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlmglue.c | 3 +- fs/ocfs2/stackglue.c | 111 ++++++++++++++++++++++++++++++++++++++++++++++----- fs/ocfs2/stackglue.h | 3 +- 3 files changed, 104 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index f62a9e4fc31..394d25a131a 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -2627,7 +2627,8 @@ int ocfs2_dlm_init(struct ocfs2_super *osb) } /* for now, uuid == domain */ - status = ocfs2_cluster_connect(osb->uuid_str, + status = ocfs2_cluster_connect(osb->osb_cluster_stack, + osb->uuid_str, strlen(osb->uuid_str), ocfs2_do_node_down, osb, &conn); diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c index 76ae4fcebcb..bf45d9bff8a 100644 --- a/fs/ocfs2/stackglue.c +++ b/fs/ocfs2/stackglue.c @@ -27,11 +27,17 @@ #include #include +#include "ocfs2_fs.h" + #include "stackglue.h" +#define OCFS2_STACK_PLUGIN_O2CB "o2cb" +#define OCFS2_STACK_PLUGIN_USER "user" + static struct ocfs2_locking_protocol *lproto; static DEFINE_SPINLOCK(ocfs2_stack_lock); static LIST_HEAD(ocfs2_stack_list); +static char cluster_stack_name[OCFS2_STACK_LABEL_LEN + 1]; /* * The stack currently in use. If not null, active_stack->sp_count > 0, @@ -53,26 +59,36 @@ static struct ocfs2_stack_plugin *ocfs2_stack_lookup(const char *name) return NULL; } -static int ocfs2_stack_driver_request(const char *name) +static int ocfs2_stack_driver_request(const char *stack_name, + const char *plugin_name) { int rc; struct ocfs2_stack_plugin *p; spin_lock(&ocfs2_stack_lock); + /* + * If the stack passed by the filesystem isn't the selected one, + * we can't continue. + */ + if (strcmp(stack_name, cluster_stack_name)) { + rc = -EBUSY; + goto out; + } + if (active_stack) { /* * If the active stack isn't the one we want, it cannot * be selected right now. */ - if (!strcmp(active_stack->sp_name, name)) + if (!strcmp(active_stack->sp_name, plugin_name)) rc = 0; else rc = -EBUSY; goto out; } - p = ocfs2_stack_lookup(name); + p = ocfs2_stack_lookup(plugin_name); if (!p || !try_module_get(p->sp_owner)) { rc = -ENOENT; goto out; @@ -94,23 +110,42 @@ out: * there is no stack, it tries to load it. It will fail if the stack still * cannot be found. It will also fail if a different stack is in use. */ -static int ocfs2_stack_driver_get(const char *name) +static int ocfs2_stack_driver_get(const char *stack_name) { int rc; + char *plugin_name = OCFS2_STACK_PLUGIN_O2CB; + + /* + * Classic stack does not pass in a stack name. This is + * compatible with older tools as well. + */ + if (!stack_name || !*stack_name) + stack_name = OCFS2_STACK_PLUGIN_O2CB; + + if (strlen(stack_name) != OCFS2_STACK_LABEL_LEN) { + printk(KERN_ERR + "ocfs2 passed an invalid cluster stack label: \"%s\"\n", + stack_name); + return -EINVAL; + } - rc = ocfs2_stack_driver_request(name); + /* Anything that isn't the classic stack is a user stack */ + if (strcmp(stack_name, OCFS2_STACK_PLUGIN_O2CB)) + plugin_name = OCFS2_STACK_PLUGIN_USER; + + rc = ocfs2_stack_driver_request(stack_name, plugin_name); if (rc == -ENOENT) { - request_module("ocfs2_stack_%s", name); - rc = ocfs2_stack_driver_request(name); + request_module("ocfs2_stack_%s", plugin_name); + rc = ocfs2_stack_driver_request(stack_name, plugin_name); } if (rc == -ENOENT) { printk(KERN_ERR "ocfs2: Cluster stack driver \"%s\" cannot be found\n", - name); + plugin_name); } else if (rc == -EBUSY) { printk(KERN_ERR - "ocfs2: A different cluster stack driver is in use\n"); + "ocfs2: A different cluster stack is in use\n"); } return rc; @@ -242,7 +277,8 @@ void ocfs2_dlm_dump_lksb(union ocfs2_dlm_lksb *lksb) } EXPORT_SYMBOL_GPL(ocfs2_dlm_dump_lksb); -int ocfs2_cluster_connect(const char *group, +int ocfs2_cluster_connect(const char *stack_name, + const char *group, int grouplen, void (*recovery_handler)(int node_num, void *recovery_data), @@ -277,7 +313,7 @@ int ocfs2_cluster_connect(const char *group, new_conn->cc_version = lproto->lp_max_version; /* This will pin the stack driver if successful */ - rc = ocfs2_stack_driver_get("o2cb"); + rc = ocfs2_stack_driver_get(stack_name); if (rc) goto out_free; @@ -416,10 +452,61 @@ static struct kobj_attribute ocfs2_attr_active_cluster_plugin = __ATTR(active_cluster_plugin, S_IFREG | S_IRUGO, ocfs2_active_cluster_plugin_show, NULL); +static ssize_t ocfs2_cluster_stack_show(struct kobject *kobj, + struct kobj_attribute *attr, + char *buf) +{ + ssize_t ret; + spin_lock(&ocfs2_stack_lock); + ret = snprintf(buf, PAGE_SIZE, "%s\n", cluster_stack_name); + spin_unlock(&ocfs2_stack_lock); + + return ret; +} + +static ssize_t ocfs2_cluster_stack_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t count) +{ + size_t len = count; + ssize_t ret; + + if (len == 0) + return len; + + if (buf[len - 1] == '\n') + len--; + + if ((len != OCFS2_STACK_LABEL_LEN) || + (strnlen(buf, len) != len)) + return -EINVAL; + + spin_lock(&ocfs2_stack_lock); + if (active_stack) { + if (!strncmp(buf, cluster_stack_name, len)) + ret = count; + else + ret = -EBUSY; + } else { + memcpy(cluster_stack_name, buf, len); + ret = count; + } + spin_unlock(&ocfs2_stack_lock); + + return ret; +} + + +static struct kobj_attribute ocfs2_attr_cluster_stack = + __ATTR(cluster_stack, S_IFREG | S_IRUGO | S_IWUSR, + ocfs2_cluster_stack_show, + ocfs2_cluster_stack_store); + static struct attribute *ocfs2_attrs[] = { &ocfs2_attr_max_locking_protocol.attr, &ocfs2_attr_loaded_cluster_plugins.attr, &ocfs2_attr_active_cluster_plugin.attr, + &ocfs2_attr_cluster_stack.attr, NULL, }; @@ -455,6 +542,8 @@ error: static int __init ocfs2_stack_glue_init(void) { + strcpy(cluster_stack_name, OCFS2_STACK_PLUGIN_O2CB); + return ocfs2_sysfs_init(); } diff --git a/fs/ocfs2/stackglue.h b/fs/ocfs2/stackglue.h index c96c8bb7686..d88bc655644 100644 --- a/fs/ocfs2/stackglue.h +++ b/fs/ocfs2/stackglue.h @@ -209,7 +209,8 @@ struct ocfs2_stack_plugin { /* Used by the filesystem */ -int ocfs2_cluster_connect(const char *group, +int ocfs2_cluster_connect(const char *stack_name, + const char *group, int grouplen, void (*recovery_handler)(int node_num, void *recovery_data), -- cgit v1.2.3 From 8adf0536c9fb578a8542dcf81104d3438a5287e4 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Wed, 28 Nov 2007 14:38:40 -0800 Subject: ocfs2: Add the user stack module. Add a skeleton for the stack_user module. It's just the barebones module code. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/stack_user.c | 38 ++++++++++++++++++++++++++++++++++++++ 1 file changed, 38 insertions(+) create mode 100644 fs/ocfs2/stack_user.c (limited to 'fs') diff --git a/fs/ocfs2/stack_user.c b/fs/ocfs2/stack_user.c new file mode 100644 index 00000000000..920eb1111b6 --- /dev/null +++ b/fs/ocfs2/stack_user.c @@ -0,0 +1,38 @@ +/* -*- mode: c; c-basic-offset: 8; -*- + * vim: noexpandtab sw=8 ts=8 sts=0: + * + * stack_user.c + * + * Code which interfaces ocfs2 with fs/dlm and a userspace stack. + * + * Copyright (C) 2007 Oracle. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License as published by the Free Software Foundation, version 2. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + */ + +#include + +#include "stackglue.h" + + +static int __init user_stack_init(void) +{ + return 0; +} + +static void __exit user_stack_exit(void) +{ +} + +MODULE_AUTHOR("Oracle"); +MODULE_DESCRIPTION("ocfs2 driver for userspace cluster stacks"); +MODULE_LICENSE("GPL"); +module_init(user_stack_init); +module_exit(user_stack_exit); -- cgit v1.2.3 From 6427a727557d9c964b7b162ae11bb156e2c501d5 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Mon, 18 Feb 2008 19:23:28 -0800 Subject: ocfs2: Add the ocfs2_control misc device. The ocfs2_control misc device is how a userspace control daemon (controld) talks to the filesystem. Introduce the bare-bones filesystem ops. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/stack_user.c | 184 +++++++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 183 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ocfs2/stack_user.c b/fs/ocfs2/stack_user.c index 920eb1111b6..fdca5d3c766 100644 --- a/fs/ocfs2/stack_user.c +++ b/fs/ocfs2/stack_user.c @@ -18,17 +18,199 @@ */ #include +#include +#include +#include +#include #include "stackglue.h" -static int __init user_stack_init(void) +/* + * The control protocol starts with a handshake. Until the handshake + * is complete, the control device will fail all write(2)s. + * + * The handshake is simple. First, the client reads until EOF. Each line + * of output is a supported protocol tag. All protocol tags are a single + * character followed by a two hex digit version number. Currently the + * only things supported is T01, for "Text-base version 0x01". Next, the + * client writes the version they would like to use. If the version tag + * written is unknown, -EINVAL is returned. Once the negotiation is + * complete, the client can start sending messages. + */ + +/* + * ocfs2_live_connection is refcounted because the filesystem and + * miscdevice sides can detach in different order. Let's just be safe. + */ +struct ocfs2_live_connection { + struct list_head oc_list; + struct ocfs2_cluster_connection *oc_conn; +}; + +static atomic_t ocfs2_control_opened; + +static LIST_HEAD(ocfs2_live_connection_list); +static DEFINE_MUTEX(ocfs2_control_lock); + +static struct ocfs2_live_connection *ocfs2_connection_find(const char *name) +{ + size_t len = strlen(name); + struct ocfs2_live_connection *c; + + BUG_ON(!mutex_is_locked(&ocfs2_control_lock)); + + list_for_each_entry(c, &ocfs2_live_connection_list, oc_list) { + if ((c->oc_conn->cc_namelen == len) && + !strncmp(c->oc_conn->cc_name, name, len)) + return c; + } + + return c; +} + +/* + * ocfs2_live_connection structures are created underneath the ocfs2 + * mount path. Since the VFS prevents multiple calls to + * fill_super(), we can't get dupes here. + */ +static int ocfs2_live_connection_new(struct ocfs2_cluster_connection *conn, + struct ocfs2_live_connection **c_ret) +{ + int rc = 0; + struct ocfs2_live_connection *c; + + c = kzalloc(sizeof(struct ocfs2_live_connection), GFP_KERNEL); + if (!c) + return -ENOMEM; + + mutex_lock(&ocfs2_control_lock); + c->oc_conn = conn; + + if (atomic_read(&ocfs2_control_opened)) + list_add(&c->oc_list, &ocfs2_live_connection_list); + else { + printk(KERN_ERR + "ocfs2: Userspace control daemon is not present\n"); + rc = -ESRCH; + } + + mutex_unlock(&ocfs2_control_lock); + + if (!rc) + *c_ret = c; + else + kfree(c); + + return rc; +} + +/* + * This function disconnects the cluster connection from ocfs2_control. + * Afterwards, userspace can't affect the cluster connection. + */ +static void ocfs2_live_connection_drop(struct ocfs2_live_connection *c) +{ + mutex_lock(&ocfs2_control_lock); + list_del_init(&c->oc_list); + c->oc_conn = NULL; + mutex_unlock(&ocfs2_control_lock); + + kfree(c); +} + + +static ssize_t ocfs2_control_write(struct file *file, + const char __user *buf, + size_t count, + loff_t *ppos) { return 0; } +static ssize_t ocfs2_control_read(struct file *file, + char __user *buf, + size_t count, + loff_t *ppos) +{ + return 0; +} + +static int ocfs2_control_release(struct inode *inode, struct file *file) +{ + if (atomic_dec_and_test(&ocfs2_control_opened)) { + mutex_lock(&ocfs2_control_lock); + if (!list_empty(&ocfs2_live_connection_list)) { + /* XXX: Do bad things! */ + printk(KERN_ERR + "ocfs2: Unexpected release of ocfs2_control!\n" + " Loss of cluster connection requires " + "an emergency restart!\n"); + emergency_restart(); + } + mutex_unlock(&ocfs2_control_lock); + } + + return 0; +} + +static int ocfs2_control_open(struct inode *inode, struct file *file) +{ + atomic_inc(&ocfs2_control_opened); + + return 0; +} + +static const struct file_operations ocfs2_control_fops = { + .open = ocfs2_control_open, + .release = ocfs2_control_release, + .read = ocfs2_control_read, + .write = ocfs2_control_write, + .owner = THIS_MODULE, +}; + +struct miscdevice ocfs2_control_device = { + .minor = MISC_DYNAMIC_MINOR, + .name = "ocfs2_control", + .fops = &ocfs2_control_fops, +}; + +static int ocfs2_control_init(void) +{ + int rc; + + atomic_set(&ocfs2_control_opened, 0); + + rc = misc_register(&ocfs2_control_device); + if (rc) + printk(KERN_ERR + "ocfs2: Unable to register ocfs2_control device " + "(errno %d)\n", + -rc); + + return rc; +} + +static void ocfs2_control_exit(void) +{ + int rc; + + rc = misc_deregister(&ocfs2_control_device); + if (rc) + printk(KERN_ERR + "ocfs2: Unable to deregister ocfs2_control device " + "(errno %d)\n", + -rc); +} + +static int __init user_stack_init(void) +{ + return ocfs2_control_init(); +} + static void __exit user_stack_exit(void) { + ocfs2_control_exit(); } MODULE_AUTHOR("Oracle"); -- cgit v1.2.3 From 462c7e6a257e547eebe1648396cf7c45e684091b Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Mon, 18 Feb 2008 19:40:12 -0800 Subject: ocfs2: Start the ocfs2_control handshake. When a control daemon opens the ocfs2_control device, it must perform a handshake to tell the filesystem it is something capable of monitoring cluster status. Only after the handshake is complete will the filesystem allow mounts. This is the first part of the handshake. The daemon reads all supported ocfs2_control protocols, then writes in the protocol it will use. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/stack_user.c | 144 ++++++++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 139 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/stack_user.c b/fs/ocfs2/stack_user.c index fdca5d3c766..ff8d3075792 100644 --- a/fs/ocfs2/stack_user.c +++ b/fs/ocfs2/stack_user.c @@ -22,6 +22,7 @@ #include #include #include +#include #include "stackglue.h" @@ -39,6 +40,16 @@ * complete, the client can start sending messages. */ +/* + * Whether or not the client has done the handshake. + * For now, we have just one protocol version. + */ +#define OCFS2_CONTROL_PROTO "T01\n" +#define OCFS2_CONTROL_PROTO_LEN 4 +#define OCFS2_CONTROL_HANDSHAKE_INVALID (0) +#define OCFS2_CONTROL_HANDSHAKE_READ (1) +#define OCFS2_CONTROL_HANDSHAKE_VALID (2) + /* * ocfs2_live_connection is refcounted because the filesystem and * miscdevice sides can detach in different order. Let's just be safe. @@ -48,11 +59,30 @@ struct ocfs2_live_connection { struct ocfs2_cluster_connection *oc_conn; }; +struct ocfs2_control_private { + struct list_head op_list; + int op_state; +}; + static atomic_t ocfs2_control_opened; static LIST_HEAD(ocfs2_live_connection_list); +static LIST_HEAD(ocfs2_control_private_list); static DEFINE_MUTEX(ocfs2_control_lock); +static inline void ocfs2_control_set_handshake_state(struct file *file, + int state) +{ + struct ocfs2_control_private *p = file->private_data; + p->op_state = state; +} + +static inline int ocfs2_control_get_handshake_state(struct file *file) +{ + struct ocfs2_control_private *p = file->private_data; + return p->op_state; +} + static struct ocfs2_live_connection *ocfs2_connection_find(const char *name) { size_t len = strlen(name); @@ -119,27 +149,115 @@ static void ocfs2_live_connection_drop(struct ocfs2_live_connection *c) kfree(c); } +static ssize_t ocfs2_control_cfu(char *target, size_t target_len, + const char __user *buf, size_t count) +{ + /* The T01 expects write(2) calls to have exactly one command */ + if (count != target_len) + return -EINVAL; + + if (copy_from_user(target, buf, target_len)) + return -EFAULT; + + return count; +} + +static ssize_t ocfs2_control_validate_handshake(struct file *file, + const char __user *buf, + size_t count) +{ + ssize_t ret; + char kbuf[OCFS2_CONTROL_PROTO_LEN]; + + ret = ocfs2_control_cfu(kbuf, OCFS2_CONTROL_PROTO_LEN, + buf, count); + if (ret != count) + return ret; + + if (strncmp(kbuf, OCFS2_CONTROL_PROTO, OCFS2_CONTROL_PROTO_LEN)) + return -EINVAL; + + atomic_inc(&ocfs2_control_opened); + ocfs2_control_set_handshake_state(file, + OCFS2_CONTROL_HANDSHAKE_VALID); + + + return count; +} + static ssize_t ocfs2_control_write(struct file *file, const char __user *buf, size_t count, loff_t *ppos) { - return 0; + ssize_t ret; + + switch (ocfs2_control_get_handshake_state(file)) { + case OCFS2_CONTROL_HANDSHAKE_INVALID: + ret = -EINVAL; + break; + + case OCFS2_CONTROL_HANDSHAKE_READ: + ret = ocfs2_control_validate_handshake(file, buf, + count); + break; + + case OCFS2_CONTROL_HANDSHAKE_VALID: + ret = count; /* XXX */ + break; + + default: + BUG(); + ret = -EIO; + break; + } + + return ret; } +/* + * This is a naive version. If we ever have a new protocol, we'll expand + * it. Probably using seq_file. + */ static ssize_t ocfs2_control_read(struct file *file, char __user *buf, size_t count, loff_t *ppos) { - return 0; + char *proto_string = OCFS2_CONTROL_PROTO; + size_t to_write = 0; + + if (*ppos >= OCFS2_CONTROL_PROTO_LEN) + return 0; + + to_write = OCFS2_CONTROL_PROTO_LEN - *ppos; + if (to_write > count) + to_write = count; + if (copy_to_user(buf, proto_string + *ppos, to_write)) + return -EFAULT; + + *ppos += to_write; + + /* Have we read the whole protocol list? */ + if (*ppos >= OCFS2_CONTROL_PROTO_LEN) + ocfs2_control_set_handshake_state(file, + OCFS2_CONTROL_HANDSHAKE_READ); + + return to_write; } static int ocfs2_control_release(struct inode *inode, struct file *file) { + struct ocfs2_control_private *p = file->private_data; + + mutex_lock(&ocfs2_control_lock); + + if (ocfs2_control_get_handshake_state(file) != + OCFS2_CONTROL_HANDSHAKE_VALID) + goto out; + if (atomic_dec_and_test(&ocfs2_control_opened)) { - mutex_lock(&ocfs2_control_lock); if (!list_empty(&ocfs2_live_connection_list)) { /* XXX: Do bad things! */ printk(KERN_ERR @@ -148,15 +266,31 @@ static int ocfs2_control_release(struct inode *inode, struct file *file) "an emergency restart!\n"); emergency_restart(); } - mutex_unlock(&ocfs2_control_lock); } +out: + list_del_init(&p->op_list); + file->private_data = NULL; + + mutex_unlock(&ocfs2_control_lock); + + kfree(p); + return 0; } static int ocfs2_control_open(struct inode *inode, struct file *file) { - atomic_inc(&ocfs2_control_opened); + struct ocfs2_control_private *p; + + p = kzalloc(sizeof(struct ocfs2_control_private), GFP_KERNEL); + if (!p) + return -ENOMEM; + + mutex_lock(&ocfs2_control_lock); + file->private_data = p; + list_add(&p->op_list, &ocfs2_control_private_list); + mutex_unlock(&ocfs2_control_lock); return 0; } -- cgit v1.2.3 From de870ef02295c9f5601dbf2efdc1be6df44b187b Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Mon, 18 Feb 2008 17:07:09 -0800 Subject: ocfs2: Introduce the DOWN message to ocfs2_control When the control daemon sees a node go down, it sends a DOWN message through the ocfs2_control device. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/stack_user.c | 94 ++++++++++++++++++++++++++++++++++++++++++++++++--- 1 file changed, 89 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/stack_user.c b/fs/ocfs2/stack_user.c index ff8d3075792..a5e58e2b9d5 100644 --- a/fs/ocfs2/stack_user.c +++ b/fs/ocfs2/stack_user.c @@ -35,9 +35,21 @@ * of output is a supported protocol tag. All protocol tags are a single * character followed by a two hex digit version number. Currently the * only things supported is T01, for "Text-base version 0x01". Next, the - * client writes the version they would like to use. If the version tag - * written is unknown, -EINVAL is returned. Once the negotiation is - * complete, the client can start sending messages. + * client writes the version they would like to use, including the newline. + * Thus, the protocol tag is 'T01\n'. If the version tag written is + * unknown, -EINVAL is returned. Once the negotiation is complete, the + * client can start sending messages. + * + * The T01 protocol only has one message, "DOWN". It has the following + * syntax: + * + * DOWN<32-char-cap-hex-uuid><8-char-hex-nodenum> + * + * eg: + * + * DOWN 632A924FDD844190BDA93C0DF6B94899 00000001\n + * + * This is 47 characters. */ /* @@ -49,6 +61,11 @@ #define OCFS2_CONTROL_HANDSHAKE_INVALID (0) #define OCFS2_CONTROL_HANDSHAKE_READ (1) #define OCFS2_CONTROL_HANDSHAKE_VALID (2) +#define OCFS2_CONTROL_MESSAGE_DOWN "DOWN" +#define OCFS2_CONTROL_MESSAGE_DOWN_LEN 4 +#define OCFS2_CONTROL_MESSAGE_DOWN_TOTAL_LEN 47 +#define OCFS2_TEXT_UUID_LEN 32 +#define OCFS2_CONTROL_MESSAGE_NODENUM_LEN 8 /* * ocfs2_live_connection is refcounted because the filesystem and @@ -149,7 +166,7 @@ static void ocfs2_live_connection_drop(struct ocfs2_live_connection *c) kfree(c); } -static ssize_t ocfs2_control_cfu(char *target, size_t target_len, +static ssize_t ocfs2_control_cfu(void *target, size_t target_len, const char __user *buf, size_t count) { /* The T01 expects write(2) calls to have exactly one command */ @@ -185,6 +202,73 @@ static ssize_t ocfs2_control_validate_handshake(struct file *file, return count; } +static void ocfs2_control_send_down(const char *uuid, + int nodenum) +{ + struct ocfs2_live_connection *c; + + mutex_lock(&ocfs2_control_lock); + + c = ocfs2_connection_find(uuid); + if (c) { + BUG_ON(c->oc_conn == NULL); + c->oc_conn->cc_recovery_handler(nodenum, + c->oc_conn->cc_recovery_data); + } + + mutex_unlock(&ocfs2_control_lock); +} + +/* DOWN<32-char-cap-hex-uuid><8-char-hex-nodenum> */ +struct ocfs2_control_message_down { + char tag[OCFS2_CONTROL_MESSAGE_DOWN_LEN]; + char space1; + char uuid[OCFS2_TEXT_UUID_LEN]; + char space2; + char nodestr[OCFS2_CONTROL_MESSAGE_NODENUM_LEN]; + char newline; +}; + +static ssize_t ocfs2_control_message(struct file *file, + const char __user *buf, + size_t count) +{ + ssize_t ret; + char *p = NULL; + long nodenum; + struct ocfs2_control_message_down msg; + + /* Try to catch padding issues */ + WARN_ON(offsetof(struct ocfs2_control_message_down, uuid) != + (sizeof(msg.tag) + sizeof(msg.space1))); + + memset(&msg, 0, sizeof(struct ocfs2_control_message_down)); + ret = ocfs2_control_cfu(&msg, OCFS2_CONTROL_MESSAGE_DOWN_TOTAL_LEN, + buf, count); + if (ret != count) + return ret; + + if (strncmp(msg.tag, OCFS2_CONTROL_MESSAGE_DOWN, + strlen(OCFS2_CONTROL_MESSAGE_DOWN))) + return -EINVAL; + + if ((msg.space1 != ' ') || (msg.space2 != ' ') || + (msg.newline != '\n')) + return -EINVAL; + msg.space1 = msg.space2 = msg.newline = '\0'; + + nodenum = simple_strtol(msg.nodestr, &p, 16); + if (!p || *p) + return -EINVAL; + + if ((nodenum == LONG_MIN) || (nodenum == LONG_MAX) || + (nodenum > INT_MAX) || (nodenum < 0)) + return -ERANGE; + + ocfs2_control_send_down(msg.uuid, nodenum); + + return count; +} static ssize_t ocfs2_control_write(struct file *file, const char __user *buf, @@ -204,7 +288,7 @@ static ssize_t ocfs2_control_write(struct file *file, break; case OCFS2_CONTROL_HANDSHAKE_VALID: - ret = count; /* XXX */ + ret = ocfs2_control_message(file, buf, count); break; default: -- cgit v1.2.3 From 3cfd4ab6b6b4bee2035b62e1c293801c3d257502 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Wed, 20 Feb 2008 14:44:34 -0800 Subject: ocfs2: Add the local node id to the handshake. This is the second part of the ocfs2_control handshake. After negotiating the ocfs2_control protocol, the daemon tells the filesystem what the local node id is via the SETN message. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/stack_user.c | 222 +++++++++++++++++++++++++++++++++++++++----------- 1 file changed, 173 insertions(+), 49 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/stack_user.c b/fs/ocfs2/stack_user.c index a5e58e2b9d5..43e6105369c 100644 --- a/fs/ocfs2/stack_user.c +++ b/fs/ocfs2/stack_user.c @@ -40,8 +40,18 @@ * unknown, -EINVAL is returned. Once the negotiation is complete, the * client can start sending messages. * - * The T01 protocol only has one message, "DOWN". It has the following - * syntax: + * The T01 protocol only has two messages. First is the "SETN" message. + * It has the following syntax: + * + * SETN<8-char-hex-nodenum> + * + * This is 14 characters. + * + * The "SETN" message must be the first message following the protocol. + * It tells ocfs2_control the local node number. + * + * Once the local node number has been set, the "DOWN" message can be + * sent for node down notification. It has the following syntax: * * DOWN<32-char-cap-hex-uuid><8-char-hex-nodenum> * @@ -58,11 +68,18 @@ */ #define OCFS2_CONTROL_PROTO "T01\n" #define OCFS2_CONTROL_PROTO_LEN 4 + +/* Handshake states */ #define OCFS2_CONTROL_HANDSHAKE_INVALID (0) #define OCFS2_CONTROL_HANDSHAKE_READ (1) -#define OCFS2_CONTROL_HANDSHAKE_VALID (2) -#define OCFS2_CONTROL_MESSAGE_DOWN "DOWN" -#define OCFS2_CONTROL_MESSAGE_DOWN_LEN 4 +#define OCFS2_CONTROL_HANDSHAKE_PROTOCOL (2) +#define OCFS2_CONTROL_HANDSHAKE_VALID (3) + +/* Messages */ +#define OCFS2_CONTROL_MESSAGE_OP_LEN 4 +#define OCFS2_CONTROL_MESSAGE_SETNODE_OP "SETN" +#define OCFS2_CONTROL_MESSAGE_SETNODE_TOTAL_LEN 14 +#define OCFS2_CONTROL_MESSAGE_DOWN_OP "DOWN" #define OCFS2_CONTROL_MESSAGE_DOWN_TOTAL_LEN 47 #define OCFS2_TEXT_UUID_LEN 32 #define OCFS2_CONTROL_MESSAGE_NODENUM_LEN 8 @@ -79,9 +96,35 @@ struct ocfs2_live_connection { struct ocfs2_control_private { struct list_head op_list; int op_state; + int op_this_node; +}; + +/* SETN<8-char-hex-nodenum> */ +struct ocfs2_control_message_setn { + char tag[OCFS2_CONTROL_MESSAGE_OP_LEN]; + char space; + char nodestr[OCFS2_CONTROL_MESSAGE_NODENUM_LEN]; + char newline; +}; + +/* DOWN<32-char-cap-hex-uuid><8-char-hex-nodenum> */ +struct ocfs2_control_message_down { + char tag[OCFS2_CONTROL_MESSAGE_OP_LEN]; + char space1; + char uuid[OCFS2_TEXT_UUID_LEN]; + char space2; + char nodestr[OCFS2_CONTROL_MESSAGE_NODENUM_LEN]; + char newline; +}; + +union ocfs2_control_message { + char tag[OCFS2_CONTROL_MESSAGE_OP_LEN]; + struct ocfs2_control_message_setn u_setn; + struct ocfs2_control_message_down u_down; }; static atomic_t ocfs2_control_opened; +static int ocfs2_control_this_node = -1; static LIST_HEAD(ocfs2_live_connection_list); static LIST_HEAD(ocfs2_control_private_list); @@ -166,38 +209,37 @@ static void ocfs2_live_connection_drop(struct ocfs2_live_connection *c) kfree(c); } -static ssize_t ocfs2_control_cfu(void *target, size_t target_len, - const char __user *buf, size_t count) +static int ocfs2_control_cfu(void *target, size_t target_len, + const char __user *buf, size_t count) { /* The T01 expects write(2) calls to have exactly one command */ - if (count != target_len) + if ((count != target_len) || + (count > sizeof(union ocfs2_control_message))) return -EINVAL; if (copy_from_user(target, buf, target_len)) return -EFAULT; - return count; + return 0; } -static ssize_t ocfs2_control_validate_handshake(struct file *file, - const char __user *buf, - size_t count) +static ssize_t ocfs2_control_validate_protocol(struct file *file, + const char __user *buf, + size_t count) { ssize_t ret; char kbuf[OCFS2_CONTROL_PROTO_LEN]; ret = ocfs2_control_cfu(kbuf, OCFS2_CONTROL_PROTO_LEN, buf, count); - if (ret != count) + if (ret) return ret; if (strncmp(kbuf, OCFS2_CONTROL_PROTO, OCFS2_CONTROL_PROTO_LEN)) return -EINVAL; - atomic_inc(&ocfs2_control_opened); ocfs2_control_set_handshake_state(file, - OCFS2_CONTROL_HANDSHAKE_VALID); - + OCFS2_CONTROL_HANDSHAKE_PROTOCOL); return count; } @@ -219,45 +261,92 @@ static void ocfs2_control_send_down(const char *uuid, mutex_unlock(&ocfs2_control_lock); } -/* DOWN<32-char-cap-hex-uuid><8-char-hex-nodenum> */ -struct ocfs2_control_message_down { - char tag[OCFS2_CONTROL_MESSAGE_DOWN_LEN]; - char space1; - char uuid[OCFS2_TEXT_UUID_LEN]; - char space2; - char nodestr[OCFS2_CONTROL_MESSAGE_NODENUM_LEN]; - char newline; -}; +/* + * Called whenever configuration elements are sent to /dev/ocfs2_control. + * If all configuration elements are present, try to set the global + * values. If not, return -EAGAIN. If there is a problem, return a + * different error. + */ +static int ocfs2_control_install_private(struct file *file) +{ + int rc = 0; + int set_p = 1; + struct ocfs2_control_private *p = file->private_data; -static ssize_t ocfs2_control_message(struct file *file, - const char __user *buf, - size_t count) + BUG_ON(p->op_state != OCFS2_CONTROL_HANDSHAKE_PROTOCOL); + + if (p->op_this_node < 0) + set_p = 0; + + mutex_lock(&ocfs2_control_lock); + if (ocfs2_control_this_node < 0) { + if (set_p) + ocfs2_control_this_node = p->op_this_node; + } else if (ocfs2_control_this_node != p->op_this_node) + rc = -EINVAL; + mutex_unlock(&ocfs2_control_lock); + + if (!rc && set_p) { + /* We set the global values successfully */ + atomic_inc(&ocfs2_control_opened); + ocfs2_control_set_handshake_state(file, + OCFS2_CONTROL_HANDSHAKE_VALID); + } + + return rc; +} + +static int ocfs2_control_do_setnode_msg(struct file *file, + struct ocfs2_control_message_setn *msg) { - ssize_t ret; - char *p = NULL; long nodenum; - struct ocfs2_control_message_down msg; + char *ptr = NULL; + struct ocfs2_control_private *p = file->private_data; - /* Try to catch padding issues */ - WARN_ON(offsetof(struct ocfs2_control_message_down, uuid) != - (sizeof(msg.tag) + sizeof(msg.space1))); + if (ocfs2_control_get_handshake_state(file) != + OCFS2_CONTROL_HANDSHAKE_PROTOCOL) + return -EINVAL; - memset(&msg, 0, sizeof(struct ocfs2_control_message_down)); - ret = ocfs2_control_cfu(&msg, OCFS2_CONTROL_MESSAGE_DOWN_TOTAL_LEN, - buf, count); - if (ret != count) - return ret; + if (strncmp(msg->tag, OCFS2_CONTROL_MESSAGE_SETNODE_OP, + OCFS2_CONTROL_MESSAGE_OP_LEN)) + return -EINVAL; + + if ((msg->space != ' ') || (msg->newline != '\n')) + return -EINVAL; + msg->space = msg->newline = '\0'; + + nodenum = simple_strtol(msg->nodestr, &ptr, 16); + if (!ptr || *ptr) + return -EINVAL; + + if ((nodenum == LONG_MIN) || (nodenum == LONG_MAX) || + (nodenum > INT_MAX) || (nodenum < 0)) + return -ERANGE; + p->op_this_node = nodenum; + + return ocfs2_control_install_private(file); +} + +static int ocfs2_control_do_down_msg(struct file *file, + struct ocfs2_control_message_down *msg) +{ + long nodenum; + char *p = NULL; + + if (ocfs2_control_get_handshake_state(file) != + OCFS2_CONTROL_HANDSHAKE_VALID) + return -EINVAL; - if (strncmp(msg.tag, OCFS2_CONTROL_MESSAGE_DOWN, - strlen(OCFS2_CONTROL_MESSAGE_DOWN))) + if (strncmp(msg->tag, OCFS2_CONTROL_MESSAGE_DOWN_OP, + OCFS2_CONTROL_MESSAGE_OP_LEN)) return -EINVAL; - if ((msg.space1 != ' ') || (msg.space2 != ' ') || - (msg.newline != '\n')) + if ((msg->space1 != ' ') || (msg->space2 != ' ') || + (msg->newline != '\n')) return -EINVAL; - msg.space1 = msg.space2 = msg.newline = '\0'; + msg->space1 = msg->space2 = msg->newline = '\0'; - nodenum = simple_strtol(msg.nodestr, &p, 16); + nodenum = simple_strtol(msg->nodestr, &p, 16); if (!p || *p) return -EINVAL; @@ -265,9 +354,40 @@ static ssize_t ocfs2_control_message(struct file *file, (nodenum > INT_MAX) || (nodenum < 0)) return -ERANGE; - ocfs2_control_send_down(msg.uuid, nodenum); + ocfs2_control_send_down(msg->uuid, nodenum); - return count; + return 0; +} + +static ssize_t ocfs2_control_message(struct file *file, + const char __user *buf, + size_t count) +{ + ssize_t ret; + union ocfs2_control_message msg; + + /* Try to catch padding issues */ + WARN_ON(offsetof(struct ocfs2_control_message_down, uuid) != + (sizeof(msg.u_down.tag) + sizeof(msg.u_down.space1))); + + memset(&msg, 0, sizeof(union ocfs2_control_message)); + ret = ocfs2_control_cfu(&msg, count, buf, count); + if (ret) + goto out; + + if ((count == OCFS2_CONTROL_MESSAGE_SETNODE_TOTAL_LEN) && + !strncmp(msg.tag, OCFS2_CONTROL_MESSAGE_SETNODE_OP, + OCFS2_CONTROL_MESSAGE_OP_LEN)) + ret = ocfs2_control_do_setnode_msg(file, &msg.u_setn); + else if ((count == OCFS2_CONTROL_MESSAGE_DOWN_TOTAL_LEN) && + !strncmp(msg.tag, OCFS2_CONTROL_MESSAGE_DOWN_OP, + OCFS2_CONTROL_MESSAGE_OP_LEN)) + ret = ocfs2_control_do_down_msg(file, &msg.u_down); + else + ret = -EINVAL; + +out: + return ret ? ret : count; } static ssize_t ocfs2_control_write(struct file *file, @@ -283,10 +403,11 @@ static ssize_t ocfs2_control_write(struct file *file, break; case OCFS2_CONTROL_HANDSHAKE_READ: - ret = ocfs2_control_validate_handshake(file, buf, - count); + ret = ocfs2_control_validate_protocol(file, buf, + count); break; + case OCFS2_CONTROL_HANDSHAKE_PROTOCOL: case OCFS2_CONTROL_HANDSHAKE_VALID: ret = ocfs2_control_message(file, buf, count); break; @@ -350,6 +471,8 @@ static int ocfs2_control_release(struct inode *inode, struct file *file) "an emergency restart!\n"); emergency_restart(); } + /* Last valid close clears the node number */ + ocfs2_control_this_node = -1; } out: @@ -370,6 +493,7 @@ static int ocfs2_control_open(struct inode *inode, struct file *file) p = kzalloc(sizeof(struct ocfs2_control_private), GFP_KERNEL); if (!p) return -ENOMEM; + p->op_this_node = -1; mutex_lock(&ocfs2_control_lock); file->private_data = p; -- cgit v1.2.3 From d4b95eef4dc4a59bcd42bdf783638a2eaa57b4c8 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Wed, 20 Feb 2008 15:39:44 -0800 Subject: ocfs2: Add the 'set version' message to the ocfs2_control device. The "SETV" message sets the filesystem locking protocol version as negotiated by the client. The client negotiates based on the maximum version advertised in /sys/fs/ocfs2/max_locking_protocol. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/stack_user.c | 131 +++++++++++++++++++++++++++++++++++++++++++++----- 1 file changed, 119 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/stack_user.c b/fs/ocfs2/stack_user.c index 43e6105369c..de982c11e69 100644 --- a/fs/ocfs2/stack_user.c +++ b/fs/ocfs2/stack_user.c @@ -40,7 +40,7 @@ * unknown, -EINVAL is returned. Once the negotiation is complete, the * client can start sending messages. * - * The T01 protocol only has two messages. First is the "SETN" message. + * The T01 protocol has three messages. First is the "SETN" message. * It has the following syntax: * * SETN<8-char-hex-nodenum> @@ -50,8 +50,22 @@ * The "SETN" message must be the first message following the protocol. * It tells ocfs2_control the local node number. * - * Once the local node number has been set, the "DOWN" message can be - * sent for node down notification. It has the following syntax: + * Next comes the "SETV" message. It has the following syntax: + * + * SETV<2-char-hex-major><2-char-hex-minor> + * + * This is 11 characters. + * + * The "SETV" message sets the filesystem locking protocol version as + * negotiated by the client. The client negotiates based on the maximum + * version advertised in /sys/fs/ocfs2/max_locking_protocol. The major + * number from the "SETV" message must match + * user_stack.sp_proto->lp_max_version.pv_major, and the minor number + * must be less than or equal to ...->lp_max_version.pv_minor. + * + * Once this information has been set, mounts will be allowed. From this + * point on, the "DOWN" message can be sent for node down notification. + * It has the following syntax: * * DOWN<32-char-cap-hex-uuid><8-char-hex-nodenum> * @@ -79,9 +93,12 @@ #define OCFS2_CONTROL_MESSAGE_OP_LEN 4 #define OCFS2_CONTROL_MESSAGE_SETNODE_OP "SETN" #define OCFS2_CONTROL_MESSAGE_SETNODE_TOTAL_LEN 14 +#define OCFS2_CONTROL_MESSAGE_SETVERSION_OP "SETV" +#define OCFS2_CONTROL_MESSAGE_SETVERSION_TOTAL_LEN 11 #define OCFS2_CONTROL_MESSAGE_DOWN_OP "DOWN" #define OCFS2_CONTROL_MESSAGE_DOWN_TOTAL_LEN 47 #define OCFS2_TEXT_UUID_LEN 32 +#define OCFS2_CONTROL_MESSAGE_VERNUM_LEN 2 #define OCFS2_CONTROL_MESSAGE_NODENUM_LEN 8 /* @@ -97,6 +114,7 @@ struct ocfs2_control_private { struct list_head op_list; int op_state; int op_this_node; + struct ocfs2_protocol_version op_proto; }; /* SETN<8-char-hex-nodenum> */ @@ -107,6 +125,16 @@ struct ocfs2_control_message_setn { char newline; }; +/* SETV<2-char-hex-major><2-char-hex-minor> */ +struct ocfs2_control_message_setv { + char tag[OCFS2_CONTROL_MESSAGE_OP_LEN]; + char space1; + char major[OCFS2_CONTROL_MESSAGE_VERNUM_LEN]; + char space2; + char minor[OCFS2_CONTROL_MESSAGE_VERNUM_LEN]; + char newline; +}; + /* DOWN<32-char-cap-hex-uuid><8-char-hex-nodenum> */ struct ocfs2_control_message_down { char tag[OCFS2_CONTROL_MESSAGE_OP_LEN]; @@ -120,11 +148,13 @@ struct ocfs2_control_message_down { union ocfs2_control_message { char tag[OCFS2_CONTROL_MESSAGE_OP_LEN]; struct ocfs2_control_message_setn u_setn; + struct ocfs2_control_message_setv u_setv; struct ocfs2_control_message_down u_down; }; static atomic_t ocfs2_control_opened; static int ocfs2_control_this_node = -1; +static struct ocfs2_protocol_version running_proto; static LIST_HEAD(ocfs2_live_connection_list); static LIST_HEAD(ocfs2_control_private_list); @@ -264,8 +294,9 @@ static void ocfs2_control_send_down(const char *uuid, /* * Called whenever configuration elements are sent to /dev/ocfs2_control. * If all configuration elements are present, try to set the global - * values. If not, return -EAGAIN. If there is a problem, return a - * different error. + * values. If there is a problem, return an error. Skip any missing + * elements, and only bump ocfs2_control_opened when we have all elements + * and are successful. */ static int ocfs2_control_install_private(struct file *file) { @@ -275,15 +306,32 @@ static int ocfs2_control_install_private(struct file *file) BUG_ON(p->op_state != OCFS2_CONTROL_HANDSHAKE_PROTOCOL); - if (p->op_this_node < 0) + mutex_lock(&ocfs2_control_lock); + + if (p->op_this_node < 0) { set_p = 0; + } else if ((ocfs2_control_this_node >= 0) && + (ocfs2_control_this_node != p->op_this_node)) { + rc = -EINVAL; + goto out_unlock; + } - mutex_lock(&ocfs2_control_lock); - if (ocfs2_control_this_node < 0) { - if (set_p) - ocfs2_control_this_node = p->op_this_node; - } else if (ocfs2_control_this_node != p->op_this_node) + if (!p->op_proto.pv_major) { + set_p = 0; + } else if (!list_empty(&ocfs2_live_connection_list) && + ((running_proto.pv_major != p->op_proto.pv_major) || + (running_proto.pv_minor != p->op_proto.pv_minor))) { rc = -EINVAL; + goto out_unlock; + } + + if (set_p) { + ocfs2_control_this_node = p->op_this_node; + running_proto.pv_major = p->op_proto.pv_major; + running_proto.pv_minor = p->op_proto.pv_minor; + } + +out_unlock: mutex_unlock(&ocfs2_control_lock); if (!rc && set_p) { @@ -327,6 +375,56 @@ static int ocfs2_control_do_setnode_msg(struct file *file, return ocfs2_control_install_private(file); } +static int ocfs2_control_do_setversion_msg(struct file *file, + struct ocfs2_control_message_setv *msg) + { + long major, minor; + char *ptr = NULL; + struct ocfs2_control_private *p = file->private_data; + struct ocfs2_protocol_version *max = + &user_stack.sp_proto->lp_max_version; + + if (ocfs2_control_get_handshake_state(file) != + OCFS2_CONTROL_HANDSHAKE_PROTOCOL) + return -EINVAL; + + if (strncmp(msg->tag, OCFS2_CONTROL_MESSAGE_SETVERSION_OP, + OCFS2_CONTROL_MESSAGE_OP_LEN)) + return -EINVAL; + + if ((msg->space1 != ' ') || (msg->space2 != ' ') || + (msg->newline != '\n')) + return -EINVAL; + msg->space1 = msg->space2 = msg->newline = '\0'; + + major = simple_strtol(msg->major, &ptr, 16); + if (!ptr || *ptr) + return -EINVAL; + minor = simple_strtol(msg->minor, &ptr, 16); + if (!ptr || *ptr) + return -EINVAL; + + /* + * The major must be between 1 and 255, inclusive. The minor + * must be between 0 and 255, inclusive. The version passed in + * must be within the maximum version supported by the filesystem. + */ + if ((major == LONG_MIN) || (major == LONG_MAX) || + (major > (u8)-1) || (major < 1)) + return -ERANGE; + if ((minor == LONG_MIN) || (minor == LONG_MAX) || + (minor > (u8)-1) || (minor < 0)) + return -ERANGE; + if ((major != max->pv_major) || + (minor > max->pv_minor)) + return -EINVAL; + + p->op_proto.pv_major = major; + p->op_proto.pv_minor = minor; + + return ocfs2_control_install_private(file); +} + static int ocfs2_control_do_down_msg(struct file *file, struct ocfs2_control_message_down *msg) { @@ -379,6 +477,10 @@ static ssize_t ocfs2_control_message(struct file *file, !strncmp(msg.tag, OCFS2_CONTROL_MESSAGE_SETNODE_OP, OCFS2_CONTROL_MESSAGE_OP_LEN)) ret = ocfs2_control_do_setnode_msg(file, &msg.u_setn); + else if ((count == OCFS2_CONTROL_MESSAGE_SETVERSION_TOTAL_LEN) && + !strncmp(msg.tag, OCFS2_CONTROL_MESSAGE_SETVERSION_OP, + OCFS2_CONTROL_MESSAGE_OP_LEN)) + ret = ocfs2_control_do_setversion_msg(file, &msg.u_setv); else if ((count == OCFS2_CONTROL_MESSAGE_DOWN_TOTAL_LEN) && !strncmp(msg.tag, OCFS2_CONTROL_MESSAGE_DOWN_OP, OCFS2_CONTROL_MESSAGE_OP_LEN)) @@ -471,8 +573,13 @@ static int ocfs2_control_release(struct inode *inode, struct file *file) "an emergency restart!\n"); emergency_restart(); } - /* Last valid close clears the node number */ + /* + * Last valid close clears the node number and resets + * the locking protocol version + */ ocfs2_control_this_node = -1; + running_proto.pv_major = 0; + running_proto.pv_major = 0; } out: -- cgit v1.2.3 From cf4d8d75d8aba537a19b313a9364fd08ddbd5622 Mon Sep 17 00:00:00 2001 From: David Teigland Date: Wed, 20 Feb 2008 14:29:27 -0800 Subject: ocfs2: add fsdlm to stackglue Add code to use fs/dlm. [ Modified to be part of the stack_user module -- Joel ] Signed-off-by: David Teigland Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/stack_user.c | 216 +++++++++++++++++++++++++++++++++++++++++++++++++- fs/ocfs2/stackglue.c | 14 +++- fs/ocfs2/stackglue.h | 19 ++++- 3 files changed, 243 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/stack_user.c b/fs/ocfs2/stack_user.c index de982c11e69..7428663f9cb 100644 --- a/fs/ocfs2/stack_user.c +++ b/fs/ocfs2/stack_user.c @@ -24,6 +24,7 @@ #include #include +#include "ocfs2.h" /* For struct ocfs2_lock_res */ #include "stackglue.h" @@ -152,6 +153,8 @@ union ocfs2_control_message { struct ocfs2_control_message_down u_down; }; +static struct ocfs2_stack_plugin user_stack; + static atomic_t ocfs2_control_opened; static int ocfs2_control_this_node = -1; static struct ocfs2_protocol_version running_proto; @@ -344,6 +347,20 @@ out_unlock: return rc; } +static int ocfs2_control_get_this_node(void) +{ + int rc; + + mutex_lock(&ocfs2_control_lock); + if (ocfs2_control_this_node < 0) + rc = -EINVAL; + else + rc = ocfs2_control_this_node; + mutex_unlock(&ocfs2_control_lock); + + return rc; +} + static int ocfs2_control_do_setnode_msg(struct file *file, struct ocfs2_control_message_setn *msg) { @@ -652,13 +669,210 @@ static void ocfs2_control_exit(void) -rc); } +static struct dlm_lksb *fsdlm_astarg_to_lksb(void *astarg) +{ + struct ocfs2_lock_res *res = astarg; + return &res->l_lksb.lksb_fsdlm; +} + +static void fsdlm_lock_ast_wrapper(void *astarg) +{ + struct dlm_lksb *lksb = fsdlm_astarg_to_lksb(astarg); + int status = lksb->sb_status; + + BUG_ON(user_stack.sp_proto == NULL); + + /* + * For now we're punting on the issue of other non-standard errors + * where we can't tell if the unlock_ast or lock_ast should be called. + * The main "other error" that's possible is EINVAL which means the + * function was called with invalid args, which shouldn't be possible + * since the caller here is under our control. Other non-standard + * errors probably fall into the same category, or otherwise are fatal + * which means we can't carry on anyway. + */ + + if (status == -DLM_EUNLOCK || status == -DLM_ECANCEL) + user_stack.sp_proto->lp_unlock_ast(astarg, 0); + else + user_stack.sp_proto->lp_lock_ast(astarg); +} + +static void fsdlm_blocking_ast_wrapper(void *astarg, int level) +{ + BUG_ON(user_stack.sp_proto == NULL); + + user_stack.sp_proto->lp_blocking_ast(astarg, level); +} + +static int user_dlm_lock(struct ocfs2_cluster_connection *conn, + int mode, + union ocfs2_dlm_lksb *lksb, + u32 flags, + void *name, + unsigned int namelen, + void *astarg) +{ + int ret; + + if (!lksb->lksb_fsdlm.sb_lvbptr) + lksb->lksb_fsdlm.sb_lvbptr = (char *)lksb + + sizeof(struct dlm_lksb); + + ret = dlm_lock(conn->cc_lockspace, mode, &lksb->lksb_fsdlm, + flags|DLM_LKF_NODLCKWT, name, namelen, 0, + fsdlm_lock_ast_wrapper, astarg, + fsdlm_blocking_ast_wrapper); + return ret; +} + +static int user_dlm_unlock(struct ocfs2_cluster_connection *conn, + union ocfs2_dlm_lksb *lksb, + u32 flags, + void *astarg) +{ + int ret; + + ret = dlm_unlock(conn->cc_lockspace, lksb->lksb_fsdlm.sb_lkid, + flags, &lksb->lksb_fsdlm, astarg); + return ret; +} + +static int user_dlm_lock_status(union ocfs2_dlm_lksb *lksb) +{ + return lksb->lksb_fsdlm.sb_status; +} + +static void *user_dlm_lvb(union ocfs2_dlm_lksb *lksb) +{ + return (void *)(lksb->lksb_fsdlm.sb_lvbptr); +} + +static void user_dlm_dump_lksb(union ocfs2_dlm_lksb *lksb) +{ +} + +/* + * Compare a requested locking protocol version against the current one. + * + * If the major numbers are different, they are incompatible. + * If the current minor is greater than the request, they are incompatible. + * If the current minor is less than or equal to the request, they are + * compatible, and the requester should run at the current minor version. + */ +static int fs_protocol_compare(struct ocfs2_protocol_version *existing, + struct ocfs2_protocol_version *request) +{ + if (existing->pv_major != request->pv_major) + return 1; + + if (existing->pv_minor > request->pv_minor) + return 1; + + if (existing->pv_minor < request->pv_minor) + request->pv_minor = existing->pv_minor; + + return 0; +} + +static int user_cluster_connect(struct ocfs2_cluster_connection *conn) +{ + dlm_lockspace_t *fsdlm; + struct ocfs2_live_connection *control; + int rc = 0; + + BUG_ON(conn == NULL); + + rc = ocfs2_live_connection_new(conn, &control); + if (rc) + goto out; + + /* + * running_proto must have been set before we allowed any mounts + * to proceed. + */ + if (fs_protocol_compare(&running_proto, &conn->cc_version)) { + printk(KERN_ERR + "Unable to mount with fs locking protocol version " + "%u.%u because the userspace control daemon has " + "negotiated %u.%u\n", + conn->cc_version.pv_major, conn->cc_version.pv_minor, + running_proto.pv_major, running_proto.pv_minor); + rc = -EPROTO; + ocfs2_live_connection_drop(control); + goto out; + } + + rc = dlm_new_lockspace(conn->cc_name, strlen(conn->cc_name), + &fsdlm, DLM_LSFL_FS, DLM_LVB_LEN); + if (rc) { + ocfs2_live_connection_drop(control); + goto out; + } + + conn->cc_private = control; + conn->cc_lockspace = fsdlm; +out: + return rc; +} + +static int user_cluster_disconnect(struct ocfs2_cluster_connection *conn, + int hangup_pending) +{ + dlm_release_lockspace(conn->cc_lockspace, 2); + conn->cc_lockspace = NULL; + ocfs2_live_connection_drop(conn->cc_private); + conn->cc_private = NULL; + return 0; +} + +static int user_cluster_this_node(unsigned int *this_node) +{ + int rc; + + rc = ocfs2_control_get_this_node(); + if (rc < 0) + return rc; + + *this_node = rc; + return 0; +} + +static struct ocfs2_stack_operations user_stack_ops = { + .connect = user_cluster_connect, + .disconnect = user_cluster_disconnect, + .this_node = user_cluster_this_node, + .dlm_lock = user_dlm_lock, + .dlm_unlock = user_dlm_unlock, + .lock_status = user_dlm_lock_status, + .lock_lvb = user_dlm_lvb, + .dump_lksb = user_dlm_dump_lksb, +}; + +static struct ocfs2_stack_plugin user_stack = { + .sp_name = "user", + .sp_ops = &user_stack_ops, + .sp_owner = THIS_MODULE, +}; + + static int __init user_stack_init(void) { - return ocfs2_control_init(); + int rc; + + rc = ocfs2_control_init(); + if (!rc) { + rc = ocfs2_stack_glue_register(&user_stack); + if (rc) + ocfs2_control_exit(); + } + + return rc; } static void __exit user_stack_exit(void) { + ocfs2_stack_glue_unregister(&user_stack); ocfs2_control_exit(); } diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c index bf45d9bff8a..119f60cea9c 100644 --- a/fs/ocfs2/stackglue.c +++ b/fs/ocfs2/stackglue.c @@ -228,13 +228,20 @@ void ocfs2_stack_glue_set_locking_protocol(struct ocfs2_locking_protocol *proto) EXPORT_SYMBOL_GPL(ocfs2_stack_glue_set_locking_protocol); +/* + * The ocfs2_dlm_lock() and ocfs2_dlm_unlock() functions take + * "struct ocfs2_lock_res *astarg" instead of "void *astarg" because the + * underlying stack plugins need to pilfer the lksb off of the lock_res. + * If some other structure needs to be passed as an astarg, the plugins + * will need to be given a different avenue to the lksb. + */ int ocfs2_dlm_lock(struct ocfs2_cluster_connection *conn, int mode, union ocfs2_dlm_lksb *lksb, u32 flags, void *name, unsigned int namelen, - void *astarg) + struct ocfs2_lock_res *astarg) { BUG_ON(lproto == NULL); @@ -246,7 +253,7 @@ EXPORT_SYMBOL_GPL(ocfs2_dlm_lock); int ocfs2_dlm_unlock(struct ocfs2_cluster_connection *conn, union ocfs2_dlm_lksb *lksb, u32 flags, - void *astarg) + struct ocfs2_lock_res *astarg) { BUG_ON(lproto == NULL); @@ -360,7 +367,8 @@ void ocfs2_cluster_hangup(const char *group, int grouplen) BUG_ON(group == NULL); BUG_ON(group[grouplen] != '\0'); - active_stack->sp_ops->hangup(group, grouplen); + if (active_stack->sp_ops->hangup) + active_stack->sp_ops->hangup(group, grouplen); /* cluster_disconnect() was called with hangup_pending==1 */ ocfs2_stack_driver_put(); diff --git a/fs/ocfs2/stackglue.h b/fs/ocfs2/stackglue.h index d88bc655644..005e4f170e0 100644 --- a/fs/ocfs2/stackglue.h +++ b/fs/ocfs2/stackglue.h @@ -26,6 +26,7 @@ #include #include "dlm/dlmapi.h" +#include /* * dlmconstants.h does not have a LOCAL flag. We hope to remove it @@ -60,6 +61,17 @@ struct ocfs2_locking_protocol { void (*lp_unlock_ast)(void *astarg, int error); }; + +/* + * The dlm_lockstatus struct includes lvb space, but the dlm_lksb struct only + * has a pointer to separately allocated lvb space. This struct exists only to + * include in the lksb union to make space for a combined dlm_lksb and lvb. + */ +struct fsdlm_lksb_plus_lvb { + struct dlm_lksb lksb; + char lvb[DLM_LVB_LEN]; +}; + /* * A union of all lock status structures. We define it here so that the * size of the union is known. Lock status structures are embedded in @@ -67,6 +79,8 @@ struct ocfs2_locking_protocol { */ union ocfs2_dlm_lksb { struct dlm_lockstatus lksb_o2dlm; + struct dlm_lksb lksb_fsdlm; + struct fsdlm_lksb_plus_lvb padding; }; /* @@ -221,17 +235,18 @@ int ocfs2_cluster_disconnect(struct ocfs2_cluster_connection *conn, void ocfs2_cluster_hangup(const char *group, int grouplen); int ocfs2_cluster_this_node(unsigned int *node); +struct ocfs2_lock_res; int ocfs2_dlm_lock(struct ocfs2_cluster_connection *conn, int mode, union ocfs2_dlm_lksb *lksb, u32 flags, void *name, unsigned int namelen, - void *astarg); + struct ocfs2_lock_res *astarg); int ocfs2_dlm_unlock(struct ocfs2_cluster_connection *conn, union ocfs2_dlm_lksb *lksb, u32 flags, - void *astarg); + struct ocfs2_lock_res *astarg); int ocfs2_dlm_lock_status(union ocfs2_dlm_lksb *lksb); void *ocfs2_dlm_lvb(union ocfs2_dlm_lksb *lksb); -- cgit v1.2.3 From 8f318311faf57481452895448e6ffaec7c38a146 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Tue, 4 Mar 2008 16:09:39 -0800 Subject: ocfs2: Change mlog_bug_on to BUG_ON in ocfs2_lockid.h The masklog code is in the o2cb stack, but ocfs2_lockid.h now needs to be included by the user stack. The BUG() in ocfs2_lock_type_string() does not need masklog support, so change it to a regular BUG_ON(). Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/ocfs2_lockid.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ocfs2/ocfs2_lockid.h b/fs/ocfs2/ocfs2_lockid.h index 86f3e3799c2..82c200f7a8f 100644 --- a/fs/ocfs2/ocfs2_lockid.h +++ b/fs/ocfs2/ocfs2_lockid.h @@ -100,7 +100,7 @@ static char *ocfs2_lock_type_strings[] = { static inline const char *ocfs2_lock_type_string(enum ocfs2_lock_type type) { #ifdef __KERNEL__ - mlog_bug_on_msg(type >= OCFS2_NUM_LOCK_TYPES, "%d\n", type); + BUG_ON(type >= OCFS2_NUM_LOCK_TYPES); #endif return ocfs2_lock_type_strings[type]; } -- cgit v1.2.3 From b92eccdd28e1e3870a5b2aa625282c9ae8e35cec Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Wed, 28 Nov 2007 14:53:30 -0800 Subject: ocfs2: Add kbuild for ocfs2_stack_user.ko Add ocfs2_stack_user.ko to the Makefile so that it builds. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/Makefile | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ocfs2/Makefile b/fs/ocfs2/Makefile index b734254329b..b8d6d022573 100644 --- a/fs/ocfs2/Makefile +++ b/fs/ocfs2/Makefile @@ -2,7 +2,11 @@ EXTRA_CFLAGS += -Ifs/ocfs2 EXTRA_CFLAGS += -DCATCH_BH_JBD_RACES -obj-$(CONFIG_OCFS2_FS) += ocfs2.o ocfs2_stackglue.o ocfs2_stack_o2cb.o +obj-$(CONFIG_OCFS2_FS) += \ + ocfs2.o \ + ocfs2_stackglue.o \ + ocfs2_stack_o2cb.o \ + ocfs2_stack_user.o ocfs2-objs := \ alloc.o \ @@ -33,6 +37,7 @@ ocfs2-objs := \ ocfs2_stackglue-objs := stackglue.o ocfs2_stack_o2cb-objs := stack_o2cb.o +ocfs2_stack_user-objs := stack_user.o obj-$(CONFIG_OCFS2_FS) += cluster/ obj-$(CONFIG_OCFS2_FS) += dlm/ -- cgit v1.2.3 From 9341d22942d63d6a1e4cc90f246980dbb7e1ca94 Mon Sep 17 00:00:00 2001 From: Joel Becker Date: Tue, 4 Mar 2008 17:58:56 -0800 Subject: ocfs2: Allow selection of cluster plug-ins. ocfs2 now supports plug-ins for the classic O2CB stack as well as userspace cluster stacks in conjunction with fs/dlm. This allows zero, one, or both of the plug-ins to be selected in Kconfig. For local mounts (non-clustered), neither plug-in is needed. Both plugins can be loaded at one time, the runtime will select the one needed for the cluster systme in use. Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/Kconfig | 26 ++++++++++++++++++++++++++ fs/ocfs2/Makefile | 10 ++++++---- 2 files changed, 32 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index c509123bea4..028ae38ecc5 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -444,6 +444,32 @@ config OCFS2_FS For more information on OCFS2, see the file . +config OCFS2_FS_O2CB + tristate "O2CB Kernelspace Clustering" + depends on OCFS2_FS + default y + help + OCFS2 includes a simple kernelspace clustering package, the OCFS2 + Cluster Base. It only requires a very small userspace component + to configure it. This comes with the standard ocfs2-tools package. + O2CB is limited to maintaining a cluster for OCFS2 file systems. + It cannot manage any other cluster applications. + + It is always safe to say Y here, as the clustering method is + run-time selectable. + +config OCFS2_FS_USERSPACE_CLUSTER + tristate "OCFS2 Userspace Clustering" + depends on OCFS2_FS && DLM + default y + help + This option will allow OCFS2 to use userspace clustering services + in conjunction with the DLM in fs/dlm. If you are using a + userspace cluster manager, say Y here. + + It is safe to say Y, as the clustering method is run-time + selectable. + config OCFS2_DEBUG_MASKLOG bool "OCFS2 logging support" depends on OCFS2_FS diff --git a/fs/ocfs2/Makefile b/fs/ocfs2/Makefile index b8d6d022573..f6956de56fd 100644 --- a/fs/ocfs2/Makefile +++ b/fs/ocfs2/Makefile @@ -4,9 +4,10 @@ EXTRA_CFLAGS += -DCATCH_BH_JBD_RACES obj-$(CONFIG_OCFS2_FS) += \ ocfs2.o \ - ocfs2_stackglue.o \ - ocfs2_stack_o2cb.o \ - ocfs2_stack_user.o + ocfs2_stackglue.o + +obj-$(CONFIG_OCFS2_FS_O2CB) += ocfs2_stack_o2cb.o +obj-$(CONFIG_OCFS2_FS_USERSPACE_CLUSTER) += ocfs2_stack_user.o ocfs2-objs := \ alloc.o \ @@ -39,5 +40,6 @@ ocfs2_stackglue-objs := stackglue.o ocfs2_stack_o2cb-objs := stack_o2cb.o ocfs2_stack_user-objs := stack_user.o +# cluster/ is always needed when OCFS2_FS for masklog support obj-$(CONFIG_OCFS2_FS) += cluster/ -obj-$(CONFIG_OCFS2_FS) += dlm/ +obj-$(CONFIG_OCFS2_FS_O2CB) += dlm/ -- cgit v1.2.3 From 12eb0035d6f0466038ef2c6e5f6f9296b9b74d91 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Mon, 10 Mar 2008 15:16:19 -0700 Subject: ocfs2/dlm: Rename slabcache dlm_mle_cache to o2dlm_mle This patch renames dlm_mle_slabcache to prevent namespace clashes with fs/dlm. Signed-off-by: Sunil Mushran Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlm/dlmdomain.c | 4 +++- fs/ocfs2/dlm/dlmmaster.c | 2 +- 2 files changed, 4 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlm/dlmdomain.c b/fs/ocfs2/dlm/dlmdomain.c index 0879d86113e..2ce620742f9 100644 --- a/fs/ocfs2/dlm/dlmdomain.c +++ b/fs/ocfs2/dlm/dlmdomain.c @@ -1816,8 +1816,10 @@ static int __init dlm_init(void) dlm_print_version(); status = dlm_init_mle_cache(); - if (status) + if (status) { + mlog(ML_ERROR, "Could not create o2dlm_mle slabcache\n"); return -1; + } status = dlm_register_net_handlers(); if (status) { diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c index ea6b8957786..90797c59101 100644 --- a/fs/ocfs2/dlm/dlmmaster.c +++ b/fs/ocfs2/dlm/dlmmaster.c @@ -507,7 +507,7 @@ static void dlm_mle_node_up(struct dlm_ctxt *dlm, int dlm_init_mle_cache(void) { - dlm_mle_cache = kmem_cache_create("dlm_mle_cache", + dlm_mle_cache = kmem_cache_create("o2dlm_mle", sizeof(struct dlm_master_list_entry), 0, SLAB_HWCACHE_ALIGN, NULL); -- cgit v1.2.3 From 724bdca9b8449d9ee5f779dc27ee3d906a04508c Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Mon, 10 Mar 2008 15:16:20 -0700 Subject: ocfs2/dlm: Create slabcaches for lock and lockres This patch makes the o2dlm allocate memory for lockres, lockname and lock structures from slabcaches rather than kmalloc. This allows us to not only make these allocs more efficient but also allows us to track the memory being consumed by these structures. Signed-off-by: Sunil Mushran Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlm/dlmcommon.h | 7 ++++++ fs/ocfs2/dlm/dlmdomain.c | 26 ++++++++++++++++++--- fs/ocfs2/dlm/dlmlock.c | 22 +++++++++++++++-- fs/ocfs2/dlm/dlmmaster.c | 61 ++++++++++++++++++++++++++++++++++++++---------- 4 files changed, 99 insertions(+), 17 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlm/dlmcommon.h b/fs/ocfs2/dlm/dlmcommon.h index dc8ea666efd..7525a8ae394 100644 --- a/fs/ocfs2/dlm/dlmcommon.h +++ b/fs/ocfs2/dlm/dlmcommon.h @@ -963,9 +963,16 @@ static inline void __dlm_wait_on_lockres(struct dlm_lock_resource *res) DLM_LOCK_RES_MIGRATING)); } +/* create/destroy slab caches */ +int dlm_init_master_caches(void); +void dlm_destroy_master_caches(void); + +int dlm_init_lock_cache(void); +void dlm_destroy_lock_cache(void); int dlm_init_mle_cache(void); void dlm_destroy_mle_cache(void); + void dlm_hb_event_notify_attached(struct dlm_ctxt *dlm, int idx, int node_up); int dlm_drop_lockres_ref(struct dlm_ctxt *dlm, struct dlm_lock_resource *res); diff --git a/fs/ocfs2/dlm/dlmdomain.c b/fs/ocfs2/dlm/dlmdomain.c index 2ce620742f9..b092364d0d5 100644 --- a/fs/ocfs2/dlm/dlmdomain.c +++ b/fs/ocfs2/dlm/dlmdomain.c @@ -1818,21 +1818,41 @@ static int __init dlm_init(void) status = dlm_init_mle_cache(); if (status) { mlog(ML_ERROR, "Could not create o2dlm_mle slabcache\n"); - return -1; + goto error; + } + + status = dlm_init_master_caches(); + if (status) { + mlog(ML_ERROR, "Could not create o2dlm_lockres and " + "o2dlm_lockname slabcaches\n"); + goto error; + } + + status = dlm_init_lock_cache(); + if (status) { + mlog(ML_ERROR, "Count not create o2dlm_lock slabcache\n"); + goto error; } status = dlm_register_net_handlers(); if (status) { - dlm_destroy_mle_cache(); - return -1; + mlog(ML_ERROR, "Unable to register network handlers\n"); + goto error; } return 0; +error: + dlm_destroy_lock_cache(); + dlm_destroy_master_caches(); + dlm_destroy_mle_cache(); + return -1; } static void __exit dlm_exit (void) { dlm_unregister_net_handlers(); + dlm_destroy_lock_cache(); + dlm_destroy_master_caches(); dlm_destroy_mle_cache(); } diff --git a/fs/ocfs2/dlm/dlmlock.c b/fs/ocfs2/dlm/dlmlock.c index 52578d907d9..83a9f2972ac 100644 --- a/fs/ocfs2/dlm/dlmlock.c +++ b/fs/ocfs2/dlm/dlmlock.c @@ -53,6 +53,8 @@ #define MLOG_MASK_PREFIX ML_DLM #include "cluster/masklog.h" +static struct kmem_cache *dlm_lock_cache = NULL; + static DEFINE_SPINLOCK(dlm_cookie_lock); static u64 dlm_next_cookie = 1; @@ -64,6 +66,22 @@ static void dlm_init_lock(struct dlm_lock *newlock, int type, static void dlm_lock_release(struct kref *kref); static void dlm_lock_detach_lockres(struct dlm_lock *lock); +int dlm_init_lock_cache(void) +{ + dlm_lock_cache = kmem_cache_create("o2dlm_lock", + sizeof(struct dlm_lock), + 0, SLAB_HWCACHE_ALIGN, NULL); + if (dlm_lock_cache == NULL) + return -ENOMEM; + return 0; +} + +void dlm_destroy_lock_cache(void) +{ + if (dlm_lock_cache) + kmem_cache_destroy(dlm_lock_cache); +} + /* Tell us whether we can grant a new lock request. * locking: * caller needs: res->spinlock @@ -353,7 +371,7 @@ static void dlm_lock_release(struct kref *kref) mlog(0, "freeing kernel-allocated lksb\n"); kfree(lock->lksb); } - kfree(lock); + kmem_cache_free(dlm_lock_cache, lock); } /* associate a lock with it's lockres, getting a ref on the lockres */ @@ -412,7 +430,7 @@ struct dlm_lock * dlm_new_lock(int type, u8 node, u64 cookie, struct dlm_lock *lock; int kernel_allocated = 0; - lock = kzalloc(sizeof(*lock), GFP_NOFS); + lock = (struct dlm_lock *) kmem_cache_zalloc(dlm_lock_cache, GFP_NOFS); if (!lock) return NULL; diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c index 90797c59101..ac9ed31e544 100644 --- a/fs/ocfs2/dlm/dlmmaster.c +++ b/fs/ocfs2/dlm/dlmmaster.c @@ -216,10 +216,10 @@ EXPORT_SYMBOL_GPL(dlm_dump_all_mles); #endif /* 0 */ - +static struct kmem_cache *dlm_lockres_cache = NULL; +static struct kmem_cache *dlm_lockname_cache = NULL; static struct kmem_cache *dlm_mle_cache = NULL; - static void dlm_mle_release(struct kref *kref); static void dlm_init_mle(struct dlm_master_list_entry *mle, enum dlm_mle_type type, @@ -560,6 +560,35 @@ static void dlm_mle_release(struct kref *kref) * LOCK RESOURCE FUNCTIONS */ +int dlm_init_master_caches(void) +{ + dlm_lockres_cache = kmem_cache_create("o2dlm_lockres", + sizeof(struct dlm_lock_resource), + 0, SLAB_HWCACHE_ALIGN, NULL); + if (!dlm_lockres_cache) + goto bail; + + dlm_lockname_cache = kmem_cache_create("o2dlm_lockname", + DLM_LOCKID_NAME_MAX, 0, + SLAB_HWCACHE_ALIGN, NULL); + if (!dlm_lockname_cache) + goto bail; + + return 0; +bail: + dlm_destroy_master_caches(); + return -ENOMEM; +} + +void dlm_destroy_master_caches(void) +{ + if (dlm_lockname_cache) + kmem_cache_destroy(dlm_lockname_cache); + + if (dlm_lockres_cache) + kmem_cache_destroy(dlm_lockres_cache); +} + static void dlm_set_lockres_owner(struct dlm_ctxt *dlm, struct dlm_lock_resource *res, u8 owner) @@ -642,9 +671,9 @@ static void dlm_lockres_release(struct kref *kref) BUG_ON(!list_empty(&res->recovering)); BUG_ON(!list_empty(&res->purge)); - kfree(res->lockname.name); + kmem_cache_free(dlm_lockname_cache, (void *)res->lockname.name); - kfree(res); + kmem_cache_free(dlm_lockres_cache, res); } void dlm_lockres_put(struct dlm_lock_resource *res) @@ -700,20 +729,28 @@ struct dlm_lock_resource *dlm_new_lockres(struct dlm_ctxt *dlm, const char *name, unsigned int namelen) { - struct dlm_lock_resource *res; + struct dlm_lock_resource *res = NULL; - res = kmalloc(sizeof(struct dlm_lock_resource), GFP_NOFS); + res = (struct dlm_lock_resource *) + kmem_cache_zalloc(dlm_lockres_cache, GFP_NOFS); if (!res) - return NULL; + goto error; - res->lockname.name = kmalloc(namelen, GFP_NOFS); - if (!res->lockname.name) { - kfree(res); - return NULL; - } + res->lockname.name = (char *) + kmem_cache_zalloc(dlm_lockname_cache, GFP_NOFS); + if (!res->lockname.name) + goto error; dlm_init_lockres(dlm, res, name, namelen); return res; + +error: + if (res && res->lockname.name) + kmem_cache_free(dlm_lockname_cache, (void *)res->lockname.name); + + if (res) + kmem_cache_free(dlm_lockres_cache, res); + return NULL; } void __dlm_lockres_grab_inflight_ref(struct dlm_ctxt *dlm, -- cgit v1.2.3 From 29576f8bb54045be944ba809d4fca1ad77c94165 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Mon, 10 Mar 2008 15:16:21 -0700 Subject: ocfs2/dlm: Link all lockres' to a tracking list This patch links all the lockres' to a tracking list in dlm_ctxt. We will use this in an upcoming patch that will walk the entire list and to dump the lockres states to a debugfs file. Signed-off-by: Sunil Mushran Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlm/dlmcommon.h | 4 ++++ fs/ocfs2/dlm/dlmdomain.c | 11 +++++++++++ fs/ocfs2/dlm/dlmmaster.c | 11 +++++++++++ 3 files changed, 26 insertions(+) (limited to 'fs') diff --git a/fs/ocfs2/dlm/dlmcommon.h b/fs/ocfs2/dlm/dlmcommon.h index 7525a8ae394..cc31abeadb8 100644 --- a/fs/ocfs2/dlm/dlmcommon.h +++ b/fs/ocfs2/dlm/dlmcommon.h @@ -101,6 +101,7 @@ struct dlm_ctxt struct list_head purge_list; struct list_head pending_asts; struct list_head pending_basts; + struct list_head tracking_list; unsigned int purge_count; spinlock_t spinlock; spinlock_t ast_lock; @@ -270,6 +271,9 @@ struct dlm_lock_resource struct list_head dirty; struct list_head recovering; // dlm_recovery_ctxt.resources list + /* Added during init and removed during release */ + struct list_head tracking; /* dlm->tracking_list */ + /* unused lock resources have their last_used stamped and are * put on a list for the dlm thread to run. */ unsigned long last_used; diff --git a/fs/ocfs2/dlm/dlmdomain.c b/fs/ocfs2/dlm/dlmdomain.c index b092364d0d5..4f7695c851e 100644 --- a/fs/ocfs2/dlm/dlmdomain.c +++ b/fs/ocfs2/dlm/dlmdomain.c @@ -644,6 +644,7 @@ int dlm_shutting_down(struct dlm_ctxt *dlm) void dlm_unregister_domain(struct dlm_ctxt *dlm) { int leave = 0; + struct dlm_lock_resource *res; spin_lock(&dlm_domain_lock); BUG_ON(dlm->dlm_state != DLM_CTXT_JOINED); @@ -673,6 +674,15 @@ void dlm_unregister_domain(struct dlm_ctxt *dlm) msleep(500); mlog(0, "%s: more migration to do\n", dlm->name); } + + /* This list should be empty. If not, print remaining lockres */ + if (!list_empty(&dlm->tracking_list)) { + mlog(ML_ERROR, "Following lockres' are still on the " + "tracking list:\n"); + list_for_each_entry(res, &dlm->tracking_list, tracking) + dlm_print_one_lock_resource(res); + } + dlm_mark_domain_leaving(dlm); dlm_leave_domain(dlm); dlm_complete_dlm_shutdown(dlm); @@ -1526,6 +1536,7 @@ static struct dlm_ctxt *dlm_alloc_ctxt(const char *domain, INIT_LIST_HEAD(&dlm->reco.node_data); INIT_LIST_HEAD(&dlm->purge_list); INIT_LIST_HEAD(&dlm->dlm_domain_handlers); + INIT_LIST_HEAD(&dlm->tracking_list); dlm->reco.state = 0; INIT_LIST_HEAD(&dlm->pending_asts); diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c index ac9ed31e544..97133465891 100644 --- a/fs/ocfs2/dlm/dlmmaster.c +++ b/fs/ocfs2/dlm/dlmmaster.c @@ -639,6 +639,14 @@ static void dlm_lockres_release(struct kref *kref) mlog(0, "destroying lockres %.*s\n", res->lockname.len, res->lockname.name); + if (!list_empty(&res->tracking)) + list_del_init(&res->tracking); + else { + mlog(ML_ERROR, "Resource %.*s not on the Tracking list\n", + res->lockname.len, res->lockname.name); + dlm_print_one_lock_resource(res); + } + if (!hlist_unhashed(&res->hash_node) || !list_empty(&res->granted) || !list_empty(&res->converting) || @@ -706,6 +714,7 @@ static void dlm_init_lockres(struct dlm_ctxt *dlm, INIT_LIST_HEAD(&res->dirty); INIT_LIST_HEAD(&res->recovering); INIT_LIST_HEAD(&res->purge); + INIT_LIST_HEAD(&res->tracking); atomic_set(&res->asts_reserved, 0); res->migration_pending = 0; res->inflight_locks = 0; @@ -721,6 +730,8 @@ static void dlm_init_lockres(struct dlm_ctxt *dlm, res->last_used = 0; + list_add_tail(&res->tracking, &dlm->tracking_list); + memset(res->lvb, 0, DLM_LVB_LEN); memset(res->refmap, 0, sizeof(res->refmap)); } -- cgit v1.2.3 From 6325b4a22b8f5e40ea9353288b3d6a32181f9718 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Mon, 10 Mar 2008 15:16:22 -0700 Subject: ocfs2/dlm: Create debugfs dirs This patch creates the debugfs directories that will hold the files to be used to dump the dlm state. Signed-off-by: Sunil Mushran Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlm/dlmcommon.h | 2 ++ fs/ocfs2/dlm/dlmdebug.c | 50 +++++++++++++++++++++++++++++++++++++++++++- fs/ocfs2/dlm/dlmdebug.h | 54 ++++++++++++++++++++++++++++++++++++++++++++++++ fs/ocfs2/dlm/dlmdomain.c | 21 ++++++++++++++++++- 4 files changed, 125 insertions(+), 2 deletions(-) create mode 100644 fs/ocfs2/dlm/dlmdebug.h (limited to 'fs') diff --git a/fs/ocfs2/dlm/dlmcommon.h b/fs/ocfs2/dlm/dlmcommon.h index cc31abeadb8..6a491403bbf 100644 --- a/fs/ocfs2/dlm/dlmcommon.h +++ b/fs/ocfs2/dlm/dlmcommon.h @@ -123,6 +123,8 @@ struct dlm_ctxt atomic_t remote_resources; atomic_t unknown_resources; + struct dentry *dlm_debugfs_subroot; + /* NOTE: Next three are protected by dlm_domain_lock */ struct kref dlm_refs; enum dlm_ctxt_state dlm_state; diff --git a/fs/ocfs2/dlm/dlmdebug.c b/fs/ocfs2/dlm/dlmdebug.c index 64239b37e5d..62e2a4cbf60 100644 --- a/fs/ocfs2/dlm/dlmdebug.c +++ b/fs/ocfs2/dlm/dlmdebug.c @@ -30,6 +30,7 @@ #include #include #include +#include #include "cluster/heartbeat.h" #include "cluster/nodemanager.h" @@ -37,8 +38,8 @@ #include "dlmapi.h" #include "dlmcommon.h" - #include "dlmdomain.h" +#include "dlmdebug.h" #define MLOG_MASK_PREFIX ML_DLM #include "cluster/masklog.h" @@ -266,3 +267,50 @@ const char *dlm_errname(enum dlm_status err) return dlm_errnames[err]; } EXPORT_SYMBOL_GPL(dlm_errname); + + +#ifdef CONFIG_DEBUG_FS + +static struct dentry *dlm_debugfs_root = NULL; + +#define DLM_DEBUGFS_DIR "o2dlm" + +/* subroot - domain dir */ +int dlm_create_debugfs_subroot(struct dlm_ctxt *dlm) +{ + dlm->dlm_debugfs_subroot = debugfs_create_dir(dlm->name, + dlm_debugfs_root); + if (!dlm->dlm_debugfs_subroot) { + mlog_errno(-ENOMEM); + goto bail; + } + + return 0; +bail: + dlm_destroy_debugfs_subroot(dlm); + return -ENOMEM; +} + +void dlm_destroy_debugfs_subroot(struct dlm_ctxt *dlm) +{ + if (dlm->dlm_debugfs_subroot) + debugfs_remove(dlm->dlm_debugfs_subroot); +} + +/* debugfs root */ +int dlm_create_debugfs_root(void) +{ + dlm_debugfs_root = debugfs_create_dir(DLM_DEBUGFS_DIR, NULL); + if (!dlm_debugfs_root) { + mlog_errno(-ENOMEM); + return -ENOMEM; + } + return 0; +} + +void dlm_destroy_debugfs_root(void) +{ + if (dlm_debugfs_root) + debugfs_remove(dlm_debugfs_root); +} +#endif /* CONFIG_DEBUG_FS */ diff --git a/fs/ocfs2/dlm/dlmdebug.h b/fs/ocfs2/dlm/dlmdebug.h new file mode 100644 index 00000000000..b96959512d9 --- /dev/null +++ b/fs/ocfs2/dlm/dlmdebug.h @@ -0,0 +1,54 @@ +/* -*- mode: c; c-basic-offset: 8; -*- + * vim: noexpandtab sw=8 ts=8 sts=0: + * + * dlmdebug.h + * + * Copyright (C) 2008 Oracle. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License as published by the Free Software Foundation; either + * version 2 of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public + * License along with this program; if not, write to the + * Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 021110-1307, USA. + * + */ + +#ifndef DLMDEBUG_H +#define DLMDEBUG_H + +#ifdef CONFIG_DEBUG_FS + +int dlm_create_debugfs_subroot(struct dlm_ctxt *dlm); +void dlm_destroy_debugfs_subroot(struct dlm_ctxt *dlm); + +int dlm_create_debugfs_root(void); +void dlm_destroy_debugfs_root(void); + +#else + +static int dlm_create_debugfs_subroot(struct dlm_ctxt *dlm) +{ + return 0; +} +static void dlm_destroy_debugfs_subroot(struct dlm_ctxt *dlm) +{ +} +static int dlm_create_debugfs_root(void) +{ + return 0; +} +static void dlm_destroy_debugfs_root(void) +{ +} + +#endif /* CONFIG_DEBUG_FS */ +#endif /* DLMDEBUG_H */ diff --git a/fs/ocfs2/dlm/dlmdomain.c b/fs/ocfs2/dlm/dlmdomain.c index 4f7695c851e..c137d693cef 100644 --- a/fs/ocfs2/dlm/dlmdomain.c +++ b/fs/ocfs2/dlm/dlmdomain.c @@ -33,6 +33,7 @@ #include #include #include +#include #include "cluster/heartbeat.h" #include "cluster/nodemanager.h" @@ -40,8 +41,8 @@ #include "dlmapi.h" #include "dlmcommon.h" - #include "dlmdomain.h" +#include "dlmdebug.h" #include "dlmver.h" @@ -298,6 +299,8 @@ static int dlm_wait_on_domain_helper(const char *domain) static void dlm_free_ctxt_mem(struct dlm_ctxt *dlm) { + dlm_destroy_debugfs_subroot(dlm); + if (dlm->lockres_hash) dlm_free_pagevec((void **)dlm->lockres_hash, DLM_HASH_PAGES); @@ -1494,6 +1497,7 @@ static struct dlm_ctxt *dlm_alloc_ctxt(const char *domain, u32 key) { int i; + int ret; struct dlm_ctxt *dlm = NULL; dlm = kzalloc(sizeof(*dlm), GFP_KERNEL); @@ -1526,6 +1530,15 @@ static struct dlm_ctxt *dlm_alloc_ctxt(const char *domain, dlm->key = key; dlm->node_num = o2nm_this_node(); + ret = dlm_create_debugfs_subroot(dlm); + if (ret < 0) { + dlm_free_pagevec((void **)dlm->lockres_hash, DLM_HASH_PAGES); + kfree(dlm->name); + kfree(dlm); + dlm = NULL; + goto leave; + } + spin_lock_init(&dlm->spinlock); spin_lock_init(&dlm->master_lock); spin_lock_init(&dlm->ast_lock); @@ -1851,8 +1864,13 @@ static int __init dlm_init(void) goto error; } + status = dlm_create_debugfs_root(); + if (status) + goto error; + return 0; error: + dlm_unregister_net_handlers(); dlm_destroy_lock_cache(); dlm_destroy_master_caches(); dlm_destroy_mle_cache(); @@ -1861,6 +1879,7 @@ error: static void __exit dlm_exit (void) { + dlm_destroy_debugfs_root(); dlm_unregister_net_handlers(); dlm_destroy_lock_cache(); dlm_destroy_master_caches(); -- cgit v1.2.3 From 007dce53a29ccffc000ab5373d188f73881390fd Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Mon, 10 Mar 2008 15:16:23 -0700 Subject: ocfs2/dlm: Dump the dlm state in a debugfs file This patch dumps the dlm state (dlm_ctxt) into a debugfs file. Useful for debugging. Signed-off-by: Sunil Mushran Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlm/dlmcommon.h | 1 + fs/ocfs2/dlm/dlmdebug.c | 296 +++++++++++++++++++++++++++++++++++++++++++++++ fs/ocfs2/dlm/dlmdebug.h | 20 ++++ fs/ocfs2/dlm/dlmdomain.c | 8 ++ 4 files changed, 325 insertions(+) (limited to 'fs') diff --git a/fs/ocfs2/dlm/dlmcommon.h b/fs/ocfs2/dlm/dlmcommon.h index 6a491403bbf..f7a51ca64a2 100644 --- a/fs/ocfs2/dlm/dlmcommon.h +++ b/fs/ocfs2/dlm/dlmcommon.h @@ -123,6 +123,7 @@ struct dlm_ctxt atomic_t remote_resources; atomic_t unknown_resources; + struct dlm_debug_ctxt *dlm_debug_ctxt; struct dentry *dlm_debugfs_subroot; /* NOTE: Next three are protected by dlm_domain_lock */ diff --git a/fs/ocfs2/dlm/dlmdebug.c b/fs/ocfs2/dlm/dlmdebug.c index 62e2a4cbf60..e3354038001 100644 --- a/fs/ocfs2/dlm/dlmdebug.c +++ b/fs/ocfs2/dlm/dlmdebug.c @@ -274,6 +274,294 @@ EXPORT_SYMBOL_GPL(dlm_errname); static struct dentry *dlm_debugfs_root = NULL; #define DLM_DEBUGFS_DIR "o2dlm" +#define DLM_DEBUGFS_DLM_STATE "dlm_state" + +/* begin - utils funcs */ +static void dlm_debug_free(struct kref *kref) +{ + struct dlm_debug_ctxt *dc; + + dc = container_of(kref, struct dlm_debug_ctxt, debug_refcnt); + + kfree(dc); +} + +void dlm_debug_put(struct dlm_debug_ctxt *dc) +{ + if (dc) + kref_put(&dc->debug_refcnt, dlm_debug_free); +} + +static void dlm_debug_get(struct dlm_debug_ctxt *dc) +{ + kref_get(&dc->debug_refcnt); +} + +static int stringify_nodemap(unsigned long *nodemap, int maxnodes, + char *buf, int len) +{ + int out = 0; + int i = -1; + + while ((i = find_next_bit(nodemap, maxnodes, i + 1)) < maxnodes) + out += snprintf(buf + out, len - out, "%d ", i); + + return out; +} + +static struct debug_buffer *debug_buffer_allocate(void) +{ + struct debug_buffer *db = NULL; + + db = kzalloc(sizeof(struct debug_buffer), GFP_KERNEL); + if (!db) + goto bail; + + db->len = PAGE_SIZE; + db->buf = kmalloc(db->len, GFP_KERNEL); + if (!db->buf) + goto bail; + + return db; +bail: + kfree(db); + return NULL; +} + +static ssize_t debug_buffer_read(struct file *file, char __user *buf, + size_t nbytes, loff_t *ppos) +{ + struct debug_buffer *db = file->private_data; + + return simple_read_from_buffer(buf, nbytes, ppos, db->buf, db->len); +} + +static loff_t debug_buffer_llseek(struct file *file, loff_t off, int whence) +{ + struct debug_buffer *db = file->private_data; + loff_t new = -1; + + switch (whence) { + case 0: + new = off; + break; + case 1: + new = file->f_pos + off; + break; + } + + if (new < 0 || new > db->len) + return -EINVAL; + + return (file->f_pos = new); +} + +static int debug_buffer_release(struct inode *inode, struct file *file) +{ + struct debug_buffer *db = (struct debug_buffer *)file->private_data; + + if (db) + kfree(db->buf); + kfree(db); + + return 0; +} +/* end - util funcs */ + +/* begin - debug state funcs */ +static int debug_state_print(struct dlm_ctxt *dlm, struct debug_buffer *db) +{ + int out = 0; + struct dlm_reco_node_data *node; + char *state; + int lres, rres, ures, tres; + + lres = atomic_read(&dlm->local_resources); + rres = atomic_read(&dlm->remote_resources); + ures = atomic_read(&dlm->unknown_resources); + tres = lres + rres + ures; + + spin_lock(&dlm->spinlock); + + switch (dlm->dlm_state) { + case DLM_CTXT_NEW: + state = "NEW"; break; + case DLM_CTXT_JOINED: + state = "JOINED"; break; + case DLM_CTXT_IN_SHUTDOWN: + state = "SHUTDOWN"; break; + case DLM_CTXT_LEAVING: + state = "LEAVING"; break; + default: + state = "UNKNOWN"; break; + } + + /* Domain: xxxxxxxxxx Key: 0xdfbac769 */ + out += snprintf(db->buf + out, db->len - out, + "Domain: %s Key: 0x%08x\n", dlm->name, dlm->key); + + /* Thread Pid: xxx Node: xxx State: xxxxx */ + out += snprintf(db->buf + out, db->len - out, + "Thread Pid: %d Node: %d State: %s\n", + dlm->dlm_thread_task->pid, dlm->node_num, state); + + /* Number of Joins: xxx Joining Node: xxx */ + out += snprintf(db->buf + out, db->len - out, + "Number of Joins: %d Joining Node: %d\n", + dlm->num_joins, dlm->joining_node); + + /* Domain Map: xx xx xx */ + out += snprintf(db->buf + out, db->len - out, "Domain Map: "); + out += stringify_nodemap(dlm->domain_map, O2NM_MAX_NODES, + db->buf + out, db->len - out); + out += snprintf(db->buf + out, db->len - out, "\n"); + + /* Live Map: xx xx xx */ + out += snprintf(db->buf + out, db->len - out, "Live Map: "); + out += stringify_nodemap(dlm->live_nodes_map, O2NM_MAX_NODES, + db->buf + out, db->len - out); + out += snprintf(db->buf + out, db->len - out, "\n"); + + /* Mastered Resources Total: xxx Locally: xxx Remotely: ... */ + out += snprintf(db->buf + out, db->len - out, + "Mastered Resources Total: %d Locally: %d " + "Remotely: %d Unknown: %d\n", + tres, lres, rres, ures); + + /* Lists: Dirty=Empty Purge=InUse PendingASTs=Empty ... */ + out += snprintf(db->buf + out, db->len - out, + "Lists: Dirty=%s Purge=%s PendingASTs=%s " + "PendingBASTs=%s Master=%s\n", + (list_empty(&dlm->dirty_list) ? "Empty" : "InUse"), + (list_empty(&dlm->purge_list) ? "Empty" : "InUse"), + (list_empty(&dlm->pending_asts) ? "Empty" : "InUse"), + (list_empty(&dlm->pending_basts) ? "Empty" : "InUse"), + (list_empty(&dlm->master_list) ? "Empty" : "InUse")); + + /* Purge Count: xxx Refs: xxx */ + out += snprintf(db->buf + out, db->len - out, + "Purge Count: %d Refs: %d\n", dlm->purge_count, + atomic_read(&dlm->dlm_refs.refcount)); + + /* Dead Node: xxx */ + out += snprintf(db->buf + out, db->len - out, + "Dead Node: %d\n", dlm->reco.dead_node); + + /* What about DLM_RECO_STATE_FINALIZE? */ + if (dlm->reco.state == DLM_RECO_STATE_ACTIVE) + state = "ACTIVE"; + else + state = "INACTIVE"; + + /* Recovery Pid: xxxx Master: xxx State: xxxx */ + out += snprintf(db->buf + out, db->len - out, + "Recovery Pid: %d Master: %d State: %s\n", + dlm->dlm_reco_thread_task->pid, + dlm->reco.new_master, state); + + /* Recovery Map: xx xx */ + out += snprintf(db->buf + out, db->len - out, "Recovery Map: "); + out += stringify_nodemap(dlm->recovery_map, O2NM_MAX_NODES, + db->buf + out, db->len - out); + out += snprintf(db->buf + out, db->len - out, "\n"); + + /* Recovery Node State: */ + out += snprintf(db->buf + out, db->len - out, "Recovery Node State:\n"); + list_for_each_entry(node, &dlm->reco.node_data, list) { + switch (node->state) { + case DLM_RECO_NODE_DATA_INIT: + state = "INIT"; + break; + case DLM_RECO_NODE_DATA_REQUESTING: + state = "REQUESTING"; + break; + case DLM_RECO_NODE_DATA_DEAD: + state = "DEAD"; + break; + case DLM_RECO_NODE_DATA_RECEIVING: + state = "RECEIVING"; + break; + case DLM_RECO_NODE_DATA_REQUESTED: + state = "REQUESTED"; + break; + case DLM_RECO_NODE_DATA_DONE: + state = "DONE"; + break; + case DLM_RECO_NODE_DATA_FINALIZE_SENT: + state = "FINALIZE-SENT"; + break; + default: + state = "BAD"; + break; + } + out += snprintf(db->buf + out, db->len - out, "\t%u - %s\n", + node->node_num, state); + } + + spin_unlock(&dlm->spinlock); + + return out; +} + +static int debug_state_open(struct inode *inode, struct file *file) +{ + struct dlm_ctxt *dlm = inode->i_private; + struct debug_buffer *db = NULL; + + db = debug_buffer_allocate(); + if (!db) + goto bail; + + db->len = debug_state_print(dlm, db); + + file->private_data = db; + + return 0; +bail: + return -ENOMEM; +} + +static struct file_operations debug_state_fops = { + .open = debug_state_open, + .release = debug_buffer_release, + .read = debug_buffer_read, + .llseek = debug_buffer_llseek, +}; +/* end - debug state funcs */ + +/* files in subroot */ +int dlm_debug_init(struct dlm_ctxt *dlm) +{ + struct dlm_debug_ctxt *dc = dlm->dlm_debug_ctxt; + + /* for dumping dlm_ctxt */ + dc->debug_state_dentry = debugfs_create_file(DLM_DEBUGFS_DLM_STATE, + S_IFREG|S_IRUSR, + dlm->dlm_debugfs_subroot, + dlm, &debug_state_fops); + if (!dc->debug_state_dentry) { + mlog_errno(-ENOMEM); + goto bail; + } + + dlm_debug_get(dc); + return 0; + +bail: + dlm_debug_shutdown(dlm); + return -ENOMEM; +} + +void dlm_debug_shutdown(struct dlm_ctxt *dlm) +{ + struct dlm_debug_ctxt *dc = dlm->dlm_debug_ctxt; + + if (dc) { + if (dc->debug_state_dentry) + debugfs_remove(dc->debug_state_dentry); + dlm_debug_put(dc); + } +} /* subroot - domain dir */ int dlm_create_debugfs_subroot(struct dlm_ctxt *dlm) @@ -285,6 +573,14 @@ int dlm_create_debugfs_subroot(struct dlm_ctxt *dlm) goto bail; } + dlm->dlm_debug_ctxt = kzalloc(sizeof(struct dlm_debug_ctxt), + GFP_KERNEL); + if (!dlm->dlm_debug_ctxt) { + mlog_errno(-ENOMEM); + goto bail; + } + kref_init(&dlm->dlm_debug_ctxt->debug_refcnt); + return 0; bail: dlm_destroy_debugfs_subroot(dlm); diff --git a/fs/ocfs2/dlm/dlmdebug.h b/fs/ocfs2/dlm/dlmdebug.h index b96959512d9..94cc10a4e19 100644 --- a/fs/ocfs2/dlm/dlmdebug.h +++ b/fs/ocfs2/dlm/dlmdebug.h @@ -27,6 +27,19 @@ #ifdef CONFIG_DEBUG_FS +struct dlm_debug_ctxt { + struct kref debug_refcnt; + struct dentry *debug_state_dentry; +}; + +struct debug_buffer { + int len; + char *buf; +}; + +int dlm_debug_init(struct dlm_ctxt *dlm); +void dlm_debug_shutdown(struct dlm_ctxt *dlm); + int dlm_create_debugfs_subroot(struct dlm_ctxt *dlm); void dlm_destroy_debugfs_subroot(struct dlm_ctxt *dlm); @@ -35,6 +48,13 @@ void dlm_destroy_debugfs_root(void); #else +static int dlm_debug_init(struct dlm_ctxt *dlm) +{ + return 0; +} +static void dlm_debug_shutdown(struct dlm_ctxt *dlm) +{ +} static int dlm_create_debugfs_subroot(struct dlm_ctxt *dlm) { return 0; diff --git a/fs/ocfs2/dlm/dlmdomain.c b/fs/ocfs2/dlm/dlmdomain.c index c137d693cef..63f8125824e 100644 --- a/fs/ocfs2/dlm/dlmdomain.c +++ b/fs/ocfs2/dlm/dlmdomain.c @@ -398,6 +398,7 @@ static void dlm_destroy_dlm_worker(struct dlm_ctxt *dlm) static void dlm_complete_dlm_shutdown(struct dlm_ctxt *dlm) { dlm_unregister_domain_handlers(dlm); + dlm_debug_shutdown(dlm); dlm_complete_thread(dlm); dlm_complete_recovery_thread(dlm); dlm_destroy_dlm_worker(dlm); @@ -1418,6 +1419,12 @@ static int dlm_join_domain(struct dlm_ctxt *dlm) goto bail; } + status = dlm_debug_init(dlm); + if (status < 0) { + mlog_errno(status); + goto bail; + } + status = dlm_launch_thread(dlm); if (status < 0) { mlog_errno(status); @@ -1485,6 +1492,7 @@ bail: if (status) { dlm_unregister_domain_handlers(dlm); + dlm_debug_shutdown(dlm); dlm_complete_thread(dlm); dlm_complete_recovery_thread(dlm); dlm_destroy_dlm_worker(dlm); -- cgit v1.2.3 From 4e3d24ed1a1285fe3289653aacc965642706bacb Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Mon, 10 Mar 2008 15:16:24 -0700 Subject: ocfs2/dlm: Dumps the lockres' into a debugfs file This patch dumps all the lockres' alongwith all the locks into a debugfs file. Useful for debugging. Signed-off-by: Sunil Mushran Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlm/dlmdebug.c | 247 ++++++++++++++++++++++++++++++++++++++++++++++++ fs/ocfs2/dlm/dlmdebug.h | 8 ++ 2 files changed, 255 insertions(+) (limited to 'fs') diff --git a/fs/ocfs2/dlm/dlmdebug.c b/fs/ocfs2/dlm/dlmdebug.c index e3354038001..cccb1ce3372 100644 --- a/fs/ocfs2/dlm/dlmdebug.c +++ b/fs/ocfs2/dlm/dlmdebug.c @@ -273,8 +273,35 @@ EXPORT_SYMBOL_GPL(dlm_errname); static struct dentry *dlm_debugfs_root = NULL; +/* NOTE: This function converts a lockname into a string. It uses knowledge + * of the format of the lockname that should be outside the purview of the dlm. + * We are adding only to make dlm debugging slightly easier. + * + * For more on lockname formats, please refer to dlmglue.c and ocfs2_lockid.h. + */ +static int stringify_lockname(const char *lockname, int locklen, + char *buf, int len) +{ + int out = 0; + __be64 inode_blkno_be; + +#define OCFS2_DENTRY_LOCK_INO_START 18 + if (*lockname == 'N') { + memcpy((__be64 *)&inode_blkno_be, + (char *)&lockname[OCFS2_DENTRY_LOCK_INO_START], + sizeof(__be64)); + out += snprintf(buf + out, len - out, "%.*s%08x", + OCFS2_DENTRY_LOCK_INO_START - 1, lockname, + (unsigned int)be64_to_cpu(inode_blkno_be)); + } else + out += snprintf(buf + out, len - out, "%.*s", + locklen, lockname); + return out; +} + #define DLM_DEBUGFS_DIR "o2dlm" #define DLM_DEBUGFS_DLM_STATE "dlm_state" +#define DLM_DEBUGFS_LOCKING_STATE "locking_state" /* begin - utils funcs */ static void dlm_debug_free(struct kref *kref) @@ -368,6 +395,213 @@ static int debug_buffer_release(struct inode *inode, struct file *file) } /* end - util funcs */ +/* begin - debug lockres funcs */ +static int dump_lock(struct dlm_lock *lock, int list_type, char *buf, int len) +{ + int out; + +#define DEBUG_LOCK_VERSION 1 + spin_lock(&lock->spinlock); + out = snprintf(buf, len, "LOCK:%d,%d,%d,%d,%d,%d:%lld,%d,%d,%d,%d,%d," + "%d,%d,%d,%d\n", + DEBUG_LOCK_VERSION, + list_type, lock->ml.type, lock->ml.convert_type, + lock->ml.node, + dlm_get_lock_cookie_node(be64_to_cpu(lock->ml.cookie)), + dlm_get_lock_cookie_seq(be64_to_cpu(lock->ml.cookie)), + !list_empty(&lock->ast_list), + !list_empty(&lock->bast_list), + lock->ast_pending, lock->bast_pending, + lock->convert_pending, lock->lock_pending, + lock->cancel_pending, lock->unlock_pending, + atomic_read(&lock->lock_refs.refcount)); + spin_unlock(&lock->spinlock); + + return out; +} + +static int dump_lockres(struct dlm_lock_resource *res, char *buf, int len) +{ + struct dlm_lock *lock; + int i; + int out = 0; + + out += snprintf(buf + out, len - out, "NAME:"); + out += stringify_lockname(res->lockname.name, res->lockname.len, + buf + out, len - out); + out += snprintf(buf + out, len - out, "\n"); + +#define DEBUG_LRES_VERSION 1 + out += snprintf(buf + out, len - out, + "LRES:%d,%d,%d,%ld,%d,%d,%d,%d,%d,%d,%d\n", + DEBUG_LRES_VERSION, + res->owner, res->state, res->last_used, + !list_empty(&res->purge), + !list_empty(&res->dirty), + !list_empty(&res->recovering), + res->inflight_locks, res->migration_pending, + atomic_read(&res->asts_reserved), + atomic_read(&res->refs.refcount)); + + /* refmap */ + out += snprintf(buf + out, len - out, "RMAP:"); + out += stringify_nodemap(res->refmap, O2NM_MAX_NODES, + buf + out, len - out); + out += snprintf(buf + out, len - out, "\n"); + + /* lvb */ + out += snprintf(buf + out, len - out, "LVBX:"); + for (i = 0; i < DLM_LVB_LEN; i++) + out += snprintf(buf + out, len - out, + "%02x", (unsigned char)res->lvb[i]); + out += snprintf(buf + out, len - out, "\n"); + + /* granted */ + list_for_each_entry(lock, &res->granted, list) + out += dump_lock(lock, 0, buf + out, len - out); + + /* converting */ + list_for_each_entry(lock, &res->converting, list) + out += dump_lock(lock, 1, buf + out, len - out); + + /* blocked */ + list_for_each_entry(lock, &res->blocked, list) + out += dump_lock(lock, 2, buf + out, len - out); + + out += snprintf(buf + out, len - out, "\n"); + + return out; +} + +static void *lockres_seq_start(struct seq_file *m, loff_t *pos) +{ + struct debug_lockres *dl = m->private; + struct dlm_ctxt *dlm = dl->dl_ctxt; + struct dlm_lock_resource *res = NULL; + + spin_lock(&dlm->spinlock); + + if (dl->dl_res) { + list_for_each_entry(res, &dl->dl_res->tracking, tracking) { + if (dl->dl_res) { + dlm_lockres_put(dl->dl_res); + dl->dl_res = NULL; + } + if (&res->tracking == &dlm->tracking_list) { + mlog(0, "End of list found, %p\n", res); + dl = NULL; + break; + } + dlm_lockres_get(res); + dl->dl_res = res; + break; + } + } else { + if (!list_empty(&dlm->tracking_list)) { + list_for_each_entry(res, &dlm->tracking_list, tracking) + break; + dlm_lockres_get(res); + dl->dl_res = res; + } else + dl = NULL; + } + + if (dl) { + spin_lock(&dl->dl_res->spinlock); + dump_lockres(dl->dl_res, dl->dl_buf, dl->dl_len - 1); + spin_unlock(&dl->dl_res->spinlock); + } + + spin_unlock(&dlm->spinlock); + + return dl; +} + +static void lockres_seq_stop(struct seq_file *m, void *v) +{ +} + +static void *lockres_seq_next(struct seq_file *m, void *v, loff_t *pos) +{ + return NULL; +} + +static int lockres_seq_show(struct seq_file *s, void *v) +{ + struct debug_lockres *dl = (struct debug_lockres *)v; + + seq_printf(s, "%s", dl->dl_buf); + + return 0; +} + +static struct seq_operations debug_lockres_ops = { + .start = lockres_seq_start, + .stop = lockres_seq_stop, + .next = lockres_seq_next, + .show = lockres_seq_show, +}; + +static int debug_lockres_open(struct inode *inode, struct file *file) +{ + struct dlm_ctxt *dlm = inode->i_private; + int ret = -ENOMEM; + struct seq_file *seq; + struct debug_lockres *dl = NULL; + + dl = kzalloc(sizeof(struct debug_lockres), GFP_KERNEL); + if (!dl) { + mlog_errno(ret); + goto bail; + } + + dl->dl_len = PAGE_SIZE; + dl->dl_buf = kmalloc(dl->dl_len, GFP_KERNEL); + if (!dl->dl_buf) { + mlog_errno(ret); + goto bail; + } + + ret = seq_open(file, &debug_lockres_ops); + if (ret) { + mlog_errno(ret); + goto bail; + } + + seq = (struct seq_file *) file->private_data; + seq->private = dl; + + dlm_grab(dlm); + dl->dl_ctxt = dlm; + + return 0; +bail: + if (dl) + kfree(dl->dl_buf); + kfree(dl); + return ret; +} + +static int debug_lockres_release(struct inode *inode, struct file *file) +{ + struct seq_file *seq = (struct seq_file *)file->private_data; + struct debug_lockres *dl = (struct debug_lockres *)seq->private; + + if (dl->dl_res) + dlm_lockres_put(dl->dl_res); + dlm_put(dl->dl_ctxt); + kfree(dl->dl_buf); + return seq_release_private(inode, file); +} + +static struct file_operations debug_lockres_fops = { + .open = debug_lockres_open, + .release = debug_lockres_release, + .read = seq_read, + .llseek = seq_lseek, +}; +/* end - debug lockres funcs */ + /* begin - debug state funcs */ static int debug_state_print(struct dlm_ctxt *dlm, struct debug_buffer *db) { @@ -544,6 +778,17 @@ int dlm_debug_init(struct dlm_ctxt *dlm) goto bail; } + /* for dumping lockres */ + dc->debug_lockres_dentry = + debugfs_create_file(DLM_DEBUGFS_LOCKING_STATE, + S_IFREG|S_IRUSR, + dlm->dlm_debugfs_subroot, + dlm, &debug_lockres_fops); + if (!dc->debug_lockres_dentry) { + mlog_errno(-ENOMEM); + goto bail; + } + dlm_debug_get(dc); return 0; @@ -557,6 +802,8 @@ void dlm_debug_shutdown(struct dlm_ctxt *dlm) struct dlm_debug_ctxt *dc = dlm->dlm_debug_ctxt; if (dc) { + if (dc->debug_lockres_dentry) + debugfs_remove(dc->debug_lockres_dentry); if (dc->debug_state_dentry) debugfs_remove(dc->debug_state_dentry); dlm_debug_put(dc); diff --git a/fs/ocfs2/dlm/dlmdebug.h b/fs/ocfs2/dlm/dlmdebug.h index 94cc10a4e19..7c5b2b0a05e 100644 --- a/fs/ocfs2/dlm/dlmdebug.h +++ b/fs/ocfs2/dlm/dlmdebug.h @@ -30,6 +30,7 @@ struct dlm_debug_ctxt { struct kref debug_refcnt; struct dentry *debug_state_dentry; + struct dentry *debug_lockres_dentry; }; struct debug_buffer { @@ -37,6 +38,13 @@ struct debug_buffer { char *buf; }; +struct debug_lockres { + int dl_len; + char *dl_buf; + struct dlm_ctxt *dl_ctxt; + struct dlm_lock_resource *dl_res; +}; + int dlm_debug_init(struct dlm_ctxt *dlm); void dlm_debug_shutdown(struct dlm_ctxt *dlm); -- cgit v1.2.3 From 751155a953e1fe558d3d3c3db7087712ffc15c3e Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Mon, 10 Mar 2008 15:16:25 -0700 Subject: ocfs2/dlm: Move struct dlm_master_list_entry to dlmcommon.h This patch moves some mle related definitions from dlmmaster.c to dlmcommon.h. Future patches need these definitions to dump mle debugging information. Signed-off-by: Sunil Mushran Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlm/dlmcommon.h | 35 +++++++++++++++++++++++++++++++++++ fs/ocfs2/dlm/dlmmaster.c | 37 ------------------------------------- 2 files changed, 35 insertions(+), 37 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlm/dlmcommon.h b/fs/ocfs2/dlm/dlmcommon.h index f7a51ca64a2..d5a86fb81a4 100644 --- a/fs/ocfs2/dlm/dlmcommon.h +++ b/fs/ocfs2/dlm/dlmcommon.h @@ -49,6 +49,41 @@ /* Intended to make it easier for us to switch out hash functions */ #define dlm_lockid_hash(_n, _l) full_name_hash(_n, _l) +enum dlm_mle_type { + DLM_MLE_BLOCK, + DLM_MLE_MASTER, + DLM_MLE_MIGRATION +}; + +struct dlm_lock_name { + u8 len; + u8 name[DLM_LOCKID_NAME_MAX]; +}; + +struct dlm_master_list_entry { + struct list_head list; + struct list_head hb_events; + struct dlm_ctxt *dlm; + spinlock_t spinlock; + wait_queue_head_t wq; + atomic_t woken; + struct kref mle_refs; + int inuse; + unsigned long maybe_map[BITS_TO_LONGS(O2NM_MAX_NODES)]; + unsigned long vote_map[BITS_TO_LONGS(O2NM_MAX_NODES)]; + unsigned long response_map[BITS_TO_LONGS(O2NM_MAX_NODES)]; + unsigned long node_map[BITS_TO_LONGS(O2NM_MAX_NODES)]; + u8 master; + u8 new_master; + enum dlm_mle_type type; + struct o2hb_callback_func mle_hb_up; + struct o2hb_callback_func mle_hb_down; + union { + struct dlm_lock_resource *res; + struct dlm_lock_name name; + } u; +}; + enum dlm_ast_type { DLM_AST = 0, DLM_BAST, diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c index 97133465891..94cadcb0cba 100644 --- a/fs/ocfs2/dlm/dlmmaster.c +++ b/fs/ocfs2/dlm/dlmmaster.c @@ -52,43 +52,6 @@ #define MLOG_MASK_PREFIX (ML_DLM|ML_DLM_MASTER) #include "cluster/masklog.h" -enum dlm_mle_type { - DLM_MLE_BLOCK, - DLM_MLE_MASTER, - DLM_MLE_MIGRATION -}; - -struct dlm_lock_name -{ - u8 len; - u8 name[DLM_LOCKID_NAME_MAX]; -}; - -struct dlm_master_list_entry -{ - struct list_head list; - struct list_head hb_events; - struct dlm_ctxt *dlm; - spinlock_t spinlock; - wait_queue_head_t wq; - atomic_t woken; - struct kref mle_refs; - int inuse; - unsigned long maybe_map[BITS_TO_LONGS(O2NM_MAX_NODES)]; - unsigned long vote_map[BITS_TO_LONGS(O2NM_MAX_NODES)]; - unsigned long response_map[BITS_TO_LONGS(O2NM_MAX_NODES)]; - unsigned long node_map[BITS_TO_LONGS(O2NM_MAX_NODES)]; - u8 master; - u8 new_master; - enum dlm_mle_type type; - struct o2hb_callback_func mle_hb_up; - struct o2hb_callback_func mle_hb_down; - union { - struct dlm_lock_resource *res; - struct dlm_lock_name name; - } u; -}; - static void dlm_mle_node_down(struct dlm_ctxt *dlm, struct dlm_master_list_entry *mle, struct o2nm_node *node, -- cgit v1.2.3 From d0129aceaecc2b1f5171b8e8036eb469b6e0fe81 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Mon, 10 Mar 2008 15:16:26 -0700 Subject: ocfs2/dlm: Dumps the mles into a debugfs file This patch dumps all mles it can fit in one page into a debugfs file. Useful for debugging. Signed-off-by: Sunil Mushran Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlm/dlmdebug.c | 119 ++++++++++++++++++++++++++++++++++++++++++++++++ fs/ocfs2/dlm/dlmdebug.h | 1 + 2 files changed, 120 insertions(+) (limited to 'fs') diff --git a/fs/ocfs2/dlm/dlmdebug.c b/fs/ocfs2/dlm/dlmdebug.c index cccb1ce3372..6de326bf603 100644 --- a/fs/ocfs2/dlm/dlmdebug.c +++ b/fs/ocfs2/dlm/dlmdebug.c @@ -302,6 +302,7 @@ static int stringify_lockname(const char *lockname, int locklen, #define DLM_DEBUGFS_DIR "o2dlm" #define DLM_DEBUGFS_DLM_STATE "dlm_state" #define DLM_DEBUGFS_LOCKING_STATE "locking_state" +#define DLM_DEBUGFS_MLE_STATE "mle_state" /* begin - utils funcs */ static void dlm_debug_free(struct kref *kref) @@ -395,6 +396,112 @@ static int debug_buffer_release(struct inode *inode, struct file *file) } /* end - util funcs */ +/* begin - debug mle funcs */ +static int dump_mle(struct dlm_master_list_entry *mle, char *buf, int len) +{ + int out = 0; + unsigned int namelen; + const char *name; + char *mle_type; + + if (mle->type != DLM_MLE_MASTER) { + namelen = mle->u.name.len; + name = mle->u.name.name; + } else { + namelen = mle->u.res->lockname.len; + name = mle->u.res->lockname.name; + } + + if (mle->type == DLM_MLE_BLOCK) + mle_type = "BLK"; + else if (mle->type == DLM_MLE_MASTER) + mle_type = "MAS"; + else + mle_type = "MIG"; + + out += stringify_lockname(name, namelen, buf + out, len - out); + out += snprintf(buf + out, len - out, + "\t%3s\tmas=%3u\tnew=%3u\tevt=%1d\tuse=%1d\tref=%3d\n", + mle_type, mle->master, mle->new_master, + !list_empty(&mle->hb_events), + !!mle->inuse, + atomic_read(&mle->mle_refs.refcount)); + + out += snprintf(buf + out, len - out, "Maybe="); + out += stringify_nodemap(mle->maybe_map, O2NM_MAX_NODES, + buf + out, len - out); + out += snprintf(buf + out, len - out, "\n"); + + out += snprintf(buf + out, len - out, "Vote="); + out += stringify_nodemap(mle->vote_map, O2NM_MAX_NODES, + buf + out, len - out); + out += snprintf(buf + out, len - out, "\n"); + + out += snprintf(buf + out, len - out, "Response="); + out += stringify_nodemap(mle->response_map, O2NM_MAX_NODES, + buf + out, len - out); + out += snprintf(buf + out, len - out, "\n"); + + out += snprintf(buf + out, len - out, "Node="); + out += stringify_nodemap(mle->node_map, O2NM_MAX_NODES, + buf + out, len - out); + out += snprintf(buf + out, len - out, "\n"); + + out += snprintf(buf + out, len - out, "\n"); + + return out; +} + +static int debug_mle_print(struct dlm_ctxt *dlm, struct debug_buffer *db) +{ + struct dlm_master_list_entry *mle; + int out = 0; + unsigned long total = 0; + + out += snprintf(db->buf + out, db->len - out, + "Dumping MLEs for Domain: %s\n", dlm->name); + + spin_lock(&dlm->master_lock); + list_for_each_entry(mle, &dlm->master_list, list) { + ++total; + if (db->len - out < 200) + continue; + out += dump_mle(mle, db->buf + out, db->len - out); + } + spin_unlock(&dlm->master_lock); + + out += snprintf(db->buf + out, db->len - out, + "Total on list: %ld\n", total); + return out; +} + +static int debug_mle_open(struct inode *inode, struct file *file) +{ + struct dlm_ctxt *dlm = inode->i_private; + struct debug_buffer *db; + + db = debug_buffer_allocate(); + if (!db) + goto bail; + + db->len = debug_mle_print(dlm, db); + + file->private_data = db; + + return 0; +bail: + return -ENOMEM; +} + +static struct file_operations debug_mle_fops = { + .open = debug_mle_open, + .release = debug_buffer_release, + .read = debug_buffer_read, + .llseek = debug_buffer_llseek, +}; + +/* end - debug mle funcs */ + /* begin - debug lockres funcs */ static int dump_lock(struct dlm_lock *lock, int list_type, char *buf, int len) { @@ -789,6 +896,16 @@ int dlm_debug_init(struct dlm_ctxt *dlm) goto bail; } + /* for dumping mles */ + dc->debug_mle_dentry = debugfs_create_file(DLM_DEBUGFS_MLE_STATE, + S_IFREG|S_IRUSR, + dlm->dlm_debugfs_subroot, + dlm, &debug_mle_fops); + if (!dc->debug_mle_dentry) { + mlog_errno(-ENOMEM); + goto bail; + } + dlm_debug_get(dc); return 0; @@ -802,6 +919,8 @@ void dlm_debug_shutdown(struct dlm_ctxt *dlm) struct dlm_debug_ctxt *dc = dlm->dlm_debug_ctxt; if (dc) { + if (dc->debug_mle_dentry) + debugfs_remove(dc->debug_mle_dentry); if (dc->debug_lockres_dentry) debugfs_remove(dc->debug_lockres_dentry); if (dc->debug_state_dentry) diff --git a/fs/ocfs2/dlm/dlmdebug.h b/fs/ocfs2/dlm/dlmdebug.h index 7c5b2b0a05e..cbc69f29578 100644 --- a/fs/ocfs2/dlm/dlmdebug.h +++ b/fs/ocfs2/dlm/dlmdebug.h @@ -31,6 +31,7 @@ struct dlm_debug_ctxt { struct kref debug_refcnt; struct dentry *debug_state_dentry; struct dentry *debug_lockres_dentry; + struct dentry *debug_mle_dentry; }; struct debug_buffer { -- cgit v1.2.3 From 7209300a9b987e017cae2ef9d7ef55b0fdd71869 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Mon, 10 Mar 2008 15:16:27 -0700 Subject: ocfs2/dlm: Dumps the purgelist into a debugfs file This patch dumps all the lockres' on the purgelist it can fit in one page into a debugfs file. Useful for debugging. Signed-off-by: Sunil Mushran Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlm/dlmdebug.c | 71 +++++++++++++++++++++++++++++++++++++++++++++++++ fs/ocfs2/dlm/dlmdebug.h | 1 + 2 files changed, 72 insertions(+) (limited to 'fs') diff --git a/fs/ocfs2/dlm/dlmdebug.c b/fs/ocfs2/dlm/dlmdebug.c index 6de326bf603..a109005f70d 100644 --- a/fs/ocfs2/dlm/dlmdebug.c +++ b/fs/ocfs2/dlm/dlmdebug.c @@ -303,6 +303,7 @@ static int stringify_lockname(const char *lockname, int locklen, #define DLM_DEBUGFS_DLM_STATE "dlm_state" #define DLM_DEBUGFS_LOCKING_STATE "locking_state" #define DLM_DEBUGFS_MLE_STATE "mle_state" +#define DLM_DEBUGFS_PURGE_LIST "purge_list" /* begin - utils funcs */ static void dlm_debug_free(struct kref *kref) @@ -396,6 +397,63 @@ static int debug_buffer_release(struct inode *inode, struct file *file) } /* end - util funcs */ +/* begin - purge list funcs */ +static int debug_purgelist_print(struct dlm_ctxt *dlm, struct debug_buffer *db) +{ + struct dlm_lock_resource *res; + int out = 0; + unsigned long total = 0; + + out += snprintf(db->buf + out, db->len - out, + "Dumping Purgelist for Domain: %s\n", dlm->name); + + spin_lock(&dlm->spinlock); + list_for_each_entry(res, &dlm->purge_list, purge) { + ++total; + if (db->len - out < 100) + continue; + spin_lock(&res->spinlock); + out += stringify_lockname(res->lockname.name, + res->lockname.len, + db->buf + out, db->len - out); + out += snprintf(db->buf + out, db->len - out, "\t%ld\n", + (jiffies - res->last_used)/HZ); + spin_unlock(&res->spinlock); + } + spin_unlock(&dlm->spinlock); + + out += snprintf(db->buf + out, db->len - out, + "Total on list: %ld\n", total); + + return out; +} + +static int debug_purgelist_open(struct inode *inode, struct file *file) +{ + struct dlm_ctxt *dlm = inode->i_private; + struct debug_buffer *db; + + db = debug_buffer_allocate(); + if (!db) + goto bail; + + db->len = debug_purgelist_print(dlm, db); + + file->private_data = db; + + return 0; +bail: + return -ENOMEM; +} + +static struct file_operations debug_purgelist_fops = { + .open = debug_purgelist_open, + .release = debug_buffer_release, + .read = debug_buffer_read, + .llseek = debug_buffer_llseek, +}; +/* end - purge list funcs */ + /* begin - debug mle funcs */ static int dump_mle(struct dlm_master_list_entry *mle, char *buf, int len) { @@ -906,6 +964,17 @@ int dlm_debug_init(struct dlm_ctxt *dlm) goto bail; } + /* for dumping lockres on the purge list */ + dc->debug_purgelist_dentry = + debugfs_create_file(DLM_DEBUGFS_PURGE_LIST, + S_IFREG|S_IRUSR, + dlm->dlm_debugfs_subroot, + dlm, &debug_purgelist_fops); + if (!dc->debug_purgelist_dentry) { + mlog_errno(-ENOMEM); + goto bail; + } + dlm_debug_get(dc); return 0; @@ -919,6 +988,8 @@ void dlm_debug_shutdown(struct dlm_ctxt *dlm) struct dlm_debug_ctxt *dc = dlm->dlm_debug_ctxt; if (dc) { + if (dc->debug_purgelist_dentry) + debugfs_remove(dc->debug_purgelist_dentry); if (dc->debug_mle_dentry) debugfs_remove(dc->debug_mle_dentry); if (dc->debug_lockres_dentry) diff --git a/fs/ocfs2/dlm/dlmdebug.h b/fs/ocfs2/dlm/dlmdebug.h index cbc69f29578..8857743e8f3 100644 --- a/fs/ocfs2/dlm/dlmdebug.h +++ b/fs/ocfs2/dlm/dlmdebug.h @@ -32,6 +32,7 @@ struct dlm_debug_ctxt { struct dentry *debug_state_dentry; struct dentry *debug_lockres_dentry; struct dentry *debug_mle_dentry; + struct dentry *debug_purgelist_dentry; }; struct debug_buffer { -- cgit v1.2.3 From e5a0334cbd65e27f8dfd9985aa805874fe59e879 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Mon, 10 Mar 2008 15:16:28 -0700 Subject: ocfs2/dlm: Move dlm_print_one_mle() from dlmmaster.c to dlmdebug.c This patch helps in consolidating debugging related functions in dlmdebug.c. Signed-off-by: Sunil Mushran Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlm/dlmdebug.c | 154 +++++++++++++++++++++++++---------------------- fs/ocfs2/dlm/dlmdebug.h | 2 + fs/ocfs2/dlm/dlmmaster.c | 89 +-------------------------- 3 files changed, 85 insertions(+), 160 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlm/dlmdebug.c b/fs/ocfs2/dlm/dlmdebug.c index a109005f70d..58e45795a6c 100644 --- a/fs/ocfs2/dlm/dlmdebug.c +++ b/fs/ocfs2/dlm/dlmdebug.c @@ -268,11 +268,6 @@ const char *dlm_errname(enum dlm_status err) } EXPORT_SYMBOL_GPL(dlm_errname); - -#ifdef CONFIG_DEBUG_FS - -static struct dentry *dlm_debugfs_root = NULL; - /* NOTE: This function converts a lockname into a string. It uses knowledge * of the format of the lockname that should be outside the purview of the dlm. * We are adding only to make dlm debugging slightly easier. @@ -299,6 +294,88 @@ static int stringify_lockname(const char *lockname, int locklen, return out; } +static int stringify_nodemap(unsigned long *nodemap, int maxnodes, + char *buf, int len) +{ + int out = 0; + int i = -1; + + while ((i = find_next_bit(nodemap, maxnodes, i + 1)) < maxnodes) + out += snprintf(buf + out, len - out, "%d ", i); + + return out; +} + +static int dump_mle(struct dlm_master_list_entry *mle, char *buf, int len) +{ + int out = 0; + unsigned int namelen; + const char *name; + char *mle_type; + + if (mle->type != DLM_MLE_MASTER) { + namelen = mle->u.name.len; + name = mle->u.name.name; + } else { + namelen = mle->u.res->lockname.len; + name = mle->u.res->lockname.name; + } + + if (mle->type == DLM_MLE_BLOCK) + mle_type = "BLK"; + else if (mle->type == DLM_MLE_MASTER) + mle_type = "MAS"; + else + mle_type = "MIG"; + + out += stringify_lockname(name, namelen, buf + out, len - out); + out += snprintf(buf + out, len - out, + "\t%3s\tmas=%3u\tnew=%3u\tevt=%1d\tuse=%1d\tref=%3d\n", + mle_type, mle->master, mle->new_master, + !list_empty(&mle->hb_events), + !!mle->inuse, + atomic_read(&mle->mle_refs.refcount)); + + out += snprintf(buf + out, len - out, "Maybe="); + out += stringify_nodemap(mle->maybe_map, O2NM_MAX_NODES, + buf + out, len - out); + out += snprintf(buf + out, len - out, "\n"); + + out += snprintf(buf + out, len - out, "Vote="); + out += stringify_nodemap(mle->vote_map, O2NM_MAX_NODES, + buf + out, len - out); + out += snprintf(buf + out, len - out, "\n"); + + out += snprintf(buf + out, len - out, "Response="); + out += stringify_nodemap(mle->response_map, O2NM_MAX_NODES, + buf + out, len - out); + out += snprintf(buf + out, len - out, "\n"); + + out += snprintf(buf + out, len - out, "Node="); + out += stringify_nodemap(mle->node_map, O2NM_MAX_NODES, + buf + out, len - out); + out += snprintf(buf + out, len - out, "\n"); + + out += snprintf(buf + out, len - out, "\n"); + + return out; +} + +void dlm_print_one_mle(struct dlm_master_list_entry *mle) +{ + char *buf; + + buf = (char *) get_zeroed_page(GFP_NOFS); + if (buf) { + dump_mle(mle, buf, PAGE_SIZE - 1); + free_page((unsigned long)buf); + } +} + +#ifdef CONFIG_DEBUG_FS + +static struct dentry *dlm_debugfs_root = NULL; + #define DLM_DEBUGFS_DIR "o2dlm" #define DLM_DEBUGFS_DLM_STATE "dlm_state" #define DLM_DEBUGFS_LOCKING_STATE "locking_state" @@ -326,18 +403,6 @@ static void dlm_debug_get(struct dlm_debug_ctxt *dc) kref_get(&dc->debug_refcnt); } -static int stringify_nodemap(unsigned long *nodemap, int maxnodes, - char *buf, int len) -{ - int out = 0; - int i = -1; - - while ((i = find_next_bit(nodemap, maxnodes, i + 1)) < maxnodes) - out += snprintf(buf + out, len - out, "%d ", i); - - return out; -} - static struct debug_buffer *debug_buffer_allocate(void) { struct debug_buffer *db = NULL; @@ -455,61 +520,6 @@ static struct file_operations debug_purgelist_fops = { /* end - purge list funcs */ /* begin - debug mle funcs */ -static int dump_mle(struct dlm_master_list_entry *mle, char *buf, int len) -{ - int out = 0; - unsigned int namelen; - const char *name; - char *mle_type; - - if (mle->type != DLM_MLE_MASTER) { - namelen = mle->u.name.len; - name = mle->u.name.name; - } else { - namelen = mle->u.res->lockname.len; - name = mle->u.res->lockname.name; - } - - if (mle->type == DLM_MLE_BLOCK) - mle_type = "BLK"; - else if (mle->type == DLM_MLE_MASTER) - mle_type = "MAS"; - else - mle_type = "MIG"; - - out += stringify_lockname(name, namelen, buf + out, len - out); - out += snprintf(buf + out, len - out, - "\t%3s\tmas=%3u\tnew=%3u\tevt=%1d\tuse=%1d\tref=%3d\n", - mle_type, mle->master, mle->new_master, - !list_empty(&mle->hb_events), - !!mle->inuse, - atomic_read(&mle->mle_refs.refcount)); - - out += snprintf(buf + out, len - out, "Maybe="); - out += stringify_nodemap(mle->maybe_map, O2NM_MAX_NODES, - buf + out, len - out); - out += snprintf(buf + out, len - out, "\n"); - - out += snprintf(buf + out, len - out, "Vote="); - out += stringify_nodemap(mle->vote_map, O2NM_MAX_NODES, - buf + out, len - out); - out += snprintf(buf + out, len - out, "\n"); - - out += snprintf(buf + out, len - out, "Response="); - out += stringify_nodemap(mle->response_map, O2NM_MAX_NODES, - buf + out, len - out); - out += snprintf(buf + out, len - out, "\n"); - - out += snprintf(buf + out, len - out, "Node="); - out += stringify_nodemap(mle->node_map, O2NM_MAX_NODES, - buf + out, len - out); - out += snprintf(buf + out, len - out, "\n"); - - out += snprintf(buf + out, len - out, "\n"); - - return out; -} - static int debug_mle_print(struct dlm_ctxt *dlm, struct debug_buffer *db) { struct dlm_master_list_entry *mle; diff --git a/fs/ocfs2/dlm/dlmdebug.h b/fs/ocfs2/dlm/dlmdebug.h index 8857743e8f3..d34a62a3a62 100644 --- a/fs/ocfs2/dlm/dlmdebug.h +++ b/fs/ocfs2/dlm/dlmdebug.h @@ -25,6 +25,8 @@ #ifndef DLMDEBUG_H #define DLMDEBUG_H +void dlm_print_one_mle(struct dlm_master_list_entry *mle); + #ifdef CONFIG_DEBUG_FS struct dlm_debug_ctxt { diff --git a/fs/ocfs2/dlm/dlmmaster.c b/fs/ocfs2/dlm/dlmmaster.c index 94cadcb0cba..efc015c6128 100644 --- a/fs/ocfs2/dlm/dlmmaster.c +++ b/fs/ocfs2/dlm/dlmmaster.c @@ -48,6 +48,7 @@ #include "dlmapi.h" #include "dlmcommon.h" #include "dlmdomain.h" +#include "dlmdebug.h" #define MLOG_MASK_PREFIX (ML_DLM|ML_DLM_MASTER) #include "cluster/masklog.h" @@ -91,94 +92,6 @@ static inline int dlm_mle_equal(struct dlm_ctxt *dlm, return 1; } -#define dlm_print_nodemap(m) _dlm_print_nodemap(m,#m) -static void _dlm_print_nodemap(unsigned long *map, const char *mapname) -{ - int i; - printk("%s=[ ", mapname); - for (i=0; imaybe_map, - *vote = mle->vote_map, - *resp = mle->response_map, - *node = mle->node_map; - - k = &mle->mle_refs; - if (mle->type == DLM_MLE_BLOCK) - type = "BLK"; - else if (mle->type == DLM_MLE_MASTER) - type = "MAS"; - else - type = "MIG"; - refs = atomic_read(&k->refcount); - master = mle->master; - attached = (list_empty(&mle->hb_events) ? 'N' : 'Y'); - - if (mle->type != DLM_MLE_MASTER) { - namelen = mle->u.name.len; - name = mle->u.name.name; - } else { - namelen = mle->u.res->lockname.len; - name = mle->u.res->lockname.name; - } - - mlog(ML_NOTICE, "%.*s: %3s refs=%3d mas=%3u new=%3u evt=%c inuse=%d ", - namelen, name, type, refs, master, mle->new_master, attached, - mle->inuse); - dlm_print_nodemap(maybe); - printk(", "); - dlm_print_nodemap(vote); - printk(", "); - dlm_print_nodemap(resp); - printk(", "); - dlm_print_nodemap(node); - printk(", "); - printk("\n"); -} - -#if 0 -/* Code here is included but defined out as it aids debugging */ - -static void dlm_dump_mles(struct dlm_ctxt *dlm) -{ - struct dlm_master_list_entry *mle; - - mlog(ML_NOTICE, "dumping all mles for domain %s:\n", dlm->name); - spin_lock(&dlm->master_lock); - list_for_each_entry(mle, &dlm->master_list, list) - dlm_print_one_mle(mle); - spin_unlock(&dlm->master_lock); -} - -int dlm_dump_all_mles(const char __user *data, unsigned int len) -{ - struct dlm_ctxt *dlm; - - spin_lock(&dlm_domain_lock); - list_for_each_entry(dlm, &dlm_domains, list) { - mlog(ML_NOTICE, "found dlm: %p, name=%s\n", dlm, dlm->name); - dlm_dump_mles(dlm); - } - spin_unlock(&dlm_domain_lock); - return len; -} -EXPORT_SYMBOL_GPL(dlm_dump_all_mles); - -#endif /* 0 */ - static struct kmem_cache *dlm_lockres_cache = NULL; static struct kmem_cache *dlm_lockname_cache = NULL; static struct kmem_cache *dlm_mle_cache = NULL; -- cgit v1.2.3 From c834cdb15702dd0147875b352cc7d4df93d7d900 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Mon, 10 Mar 2008 15:16:29 -0700 Subject: ocfs2/dlm: Fix lockname in lockres print function __dlm_print_one_lock_resource was printing lockname incorrectly. Also, we now use printk directly instead of mlog as the latter prints the line context which is not useful for this print. Signed-off-by: Sunil Mushran Signed-off-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/dlm/dlmdebug.c | 126 ++++++++++++++++++++---------------------------- 1 file changed, 51 insertions(+), 75 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlm/dlmdebug.c b/fs/ocfs2/dlm/dlmdebug.c index 58e45795a6c..53a9e6093e6 100644 --- a/fs/ocfs2/dlm/dlmdebug.c +++ b/fs/ocfs2/dlm/dlmdebug.c @@ -5,7 +5,7 @@ * * debug functionality for the dlm * - * Copyright (C) 2004 Oracle. All rights reserved. + * Copyright (C) 2004, 2008 Oracle. All rights reserved. * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public @@ -44,11 +44,10 @@ #define MLOG_MASK_PREFIX ML_DLM #include "cluster/masklog.h" +int stringify_lockname(const char *lockname, int locklen, char *buf, int len); + void dlm_print_one_lock_resource(struct dlm_lock_resource *res) { - mlog(ML_NOTICE, "lockres: %.*s, owner=%u, state=%u\n", - res->lockname.len, res->lockname.name, - res->owner, res->state); spin_lock(&res->spinlock); __dlm_print_one_lock_resource(res); spin_unlock(&res->spinlock); @@ -59,75 +58,78 @@ static void dlm_print_lockres_refmap(struct dlm_lock_resource *res) int bit; assert_spin_locked(&res->spinlock); - mlog(ML_NOTICE, " refmap nodes: [ "); + printk(KERN_NOTICE " refmap nodes: [ "); bit = 0; while (1) { bit = find_next_bit(res->refmap, O2NM_MAX_NODES, bit); if (bit >= O2NM_MAX_NODES) break; - printk("%u ", bit); + printk(KERN_NOTICE "%u ", bit); bit++; } - printk("], inflight=%u\n", res->inflight_locks); + printk(KERN_NOTICE "], inflight=%u\n", res->inflight_locks); +} + +static void __dlm_print_lock(struct dlm_lock *lock) +{ + spin_lock(&lock->spinlock); + + printk(KERN_NOTICE " type=%d, conv=%d, node=%u, cookie=%u:%llu, " + "ref=%u, ast=(empty=%c,pend=%c), bast=(empty=%c,pend=%c), " + "pending=(conv=%c,lock=%c,cancel=%c,unlock=%c)\n", + lock->ml.type, lock->ml.convert_type, lock->ml.node, + dlm_get_lock_cookie_node(be64_to_cpu(lock->ml.cookie)), + dlm_get_lock_cookie_seq(be64_to_cpu(lock->ml.cookie)), + atomic_read(&lock->lock_refs.refcount), + (list_empty(&lock->ast_list) ? 'y' : 'n'), + (lock->ast_pending ? 'y' : 'n'), + (list_empty(&lock->bast_list) ? 'y' : 'n'), + (lock->bast_pending ? 'y' : 'n'), + (lock->convert_pending ? 'y' : 'n'), + (lock->lock_pending ? 'y' : 'n'), + (lock->cancel_pending ? 'y' : 'n'), + (lock->unlock_pending ? 'y' : 'n')); + + spin_unlock(&lock->spinlock); } void __dlm_print_one_lock_resource(struct dlm_lock_resource *res) { struct list_head *iter2; struct dlm_lock *lock; + char buf[DLM_LOCKID_NAME_MAX]; assert_spin_locked(&res->spinlock); - mlog(ML_NOTICE, "lockres: %.*s, owner=%u, state=%u\n", - res->lockname.len, res->lockname.name, - res->owner, res->state); - mlog(ML_NOTICE, " last used: %lu, on purge list: %s\n", - res->last_used, list_empty(&res->purge) ? "no" : "yes"); + stringify_lockname(res->lockname.name, res->lockname.len, + buf, sizeof(buf) - 1); + printk(KERN_NOTICE "lockres: %s, owner=%u, state=%u\n", + buf, res->owner, res->state); + printk(KERN_NOTICE " last used: %lu, refcnt: %u, on purge list: %s\n", + res->last_used, atomic_read(&res->refs.refcount), + list_empty(&res->purge) ? "no" : "yes"); + printk(KERN_NOTICE " on dirty list: %s, on reco list: %s, " + "migrating pending: %s\n", + list_empty(&res->dirty) ? "no" : "yes", + list_empty(&res->recovering) ? "no" : "yes", + res->migration_pending ? "yes" : "no"); + printk(KERN_NOTICE " inflight locks: %d, asts reserved: %d\n", + res->inflight_locks, atomic_read(&res->asts_reserved)); dlm_print_lockres_refmap(res); - mlog(ML_NOTICE, " granted queue: \n"); + printk(KERN_NOTICE " granted queue:\n"); list_for_each(iter2, &res->granted) { lock = list_entry(iter2, struct dlm_lock, list); - spin_lock(&lock->spinlock); - mlog(ML_NOTICE, " type=%d, conv=%d, node=%u, " - "cookie=%u:%llu, ast=(empty=%c,pend=%c), bast=(empty=%c,pend=%c)\n", - lock->ml.type, lock->ml.convert_type, lock->ml.node, - dlm_get_lock_cookie_node(be64_to_cpu(lock->ml.cookie)), - dlm_get_lock_cookie_seq(be64_to_cpu(lock->ml.cookie)), - list_empty(&lock->ast_list) ? 'y' : 'n', - lock->ast_pending ? 'y' : 'n', - list_empty(&lock->bast_list) ? 'y' : 'n', - lock->bast_pending ? 'y' : 'n'); - spin_unlock(&lock->spinlock); + __dlm_print_lock(lock); } - mlog(ML_NOTICE, " converting queue: \n"); + printk(KERN_NOTICE " converting queue:\n"); list_for_each(iter2, &res->converting) { lock = list_entry(iter2, struct dlm_lock, list); - spin_lock(&lock->spinlock); - mlog(ML_NOTICE, " type=%d, conv=%d, node=%u, " - "cookie=%u:%llu, ast=(empty=%c,pend=%c), bast=(empty=%c,pend=%c)\n", - lock->ml.type, lock->ml.convert_type, lock->ml.node, - dlm_get_lock_cookie_node(be64_to_cpu(lock->ml.cookie)), - dlm_get_lock_cookie_seq(be64_to_cpu(lock->ml.cookie)), - list_empty(&lock->ast_list) ? 'y' : 'n', - lock->ast_pending ? 'y' : 'n', - list_empty(&lock->bast_list) ? 'y' : 'n', - lock->bast_pending ? 'y' : 'n'); - spin_unlock(&lock->spinlock); + __dlm_print_lock(lock); } - mlog(ML_NOTICE, " blocked queue: \n"); + printk(KERN_NOTICE " blocked queue:\n"); list_for_each(iter2, &res->blocked) { lock = list_entry(iter2, struct dlm_lock, list); - spin_lock(&lock->spinlock); - mlog(ML_NOTICE, " type=%d, conv=%d, node=%u, " - "cookie=%u:%llu, ast=(empty=%c,pend=%c), bast=(empty=%c,pend=%c)\n", - lock->ml.type, lock->ml.convert_type, lock->ml.node, - dlm_get_lock_cookie_node(be64_to_cpu(lock->ml.cookie)), - dlm_get_lock_cookie_seq(be64_to_cpu(lock->ml.cookie)), - list_empty(&lock->ast_list) ? 'y' : 'n', - lock->ast_pending ? 'y' : 'n', - list_empty(&lock->bast_list) ? 'y' : 'n', - lock->bast_pending ? 'y' : 'n'); - spin_unlock(&lock->spinlock); + __dlm_print_lock(lock); } } @@ -137,31 +139,6 @@ void dlm_print_one_lock(struct dlm_lock *lockid) } EXPORT_SYMBOL_GPL(dlm_print_one_lock); -#if 0 -void dlm_dump_lock_resources(struct dlm_ctxt *dlm) -{ - struct dlm_lock_resource *res; - struct hlist_node *iter; - struct hlist_head *bucket; - int i; - - mlog(ML_NOTICE, "struct dlm_ctxt: %s, node=%u, key=%u\n", - dlm->name, dlm->node_num, dlm->key); - if (!dlm || !dlm->name) { - mlog(ML_ERROR, "dlm=%p\n", dlm); - return; - } - - spin_lock(&dlm->spinlock); - for (i=0; ispinlock); -} -#endif /* 0 */ - static const char *dlm_errnames[] = { [DLM_NORMAL] = "DLM_NORMAL", [DLM_GRANTED] = "DLM_GRANTED", @@ -274,8 +251,7 @@ EXPORT_SYMBOL_GPL(dlm_errname); * * For more on lockname formats, please refer to dlmglue.c and ocfs2_lockid.h. */ -static int stringify_lockname(const char *lockname, int locklen, - char *buf, int len) +int stringify_lockname(const char *lockname, int locklen, char *buf, int len) { int out = 0; __be64 inode_blkno_be; -- cgit v1.2.3 From 8f50eb978935431ccbf89b0344efd4ce6a924875 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Fri, 14 Mar 2008 11:18:24 -0700 Subject: ocfs2/dlm: Cleanup lockres print A previous patch added KERN_NOTICE to printks printing the lockres that cluttered the output. This patch removes the log level. For people concerned with syslog clutter, please note we now use this facility to print lockres only during an error. Signed-off-by: Sunil Mushran Signed-off-by: Mark Fasheh --- fs/ocfs2/dlm/dlmdebug.c | 22 +++++++++++----------- 1 file changed, 11 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/dlm/dlmdebug.c b/fs/ocfs2/dlm/dlmdebug.c index 53a9e6093e6..5f6d858770a 100644 --- a/fs/ocfs2/dlm/dlmdebug.c +++ b/fs/ocfs2/dlm/dlmdebug.c @@ -58,23 +58,23 @@ static void dlm_print_lockres_refmap(struct dlm_lock_resource *res) int bit; assert_spin_locked(&res->spinlock); - printk(KERN_NOTICE " refmap nodes: [ "); + printk(" refmap nodes: [ "); bit = 0; while (1) { bit = find_next_bit(res->refmap, O2NM_MAX_NODES, bit); if (bit >= O2NM_MAX_NODES) break; - printk(KERN_NOTICE "%u ", bit); + printk("%u ", bit); bit++; } - printk(KERN_NOTICE "], inflight=%u\n", res->inflight_locks); + printk("], inflight=%u\n", res->inflight_locks); } static void __dlm_print_lock(struct dlm_lock *lock) { spin_lock(&lock->spinlock); - printk(KERN_NOTICE " type=%d, conv=%d, node=%u, cookie=%u:%llu, " + printk(" type=%d, conv=%d, node=%u, cookie=%u:%llu, " "ref=%u, ast=(empty=%c,pend=%c), bast=(empty=%c,pend=%c), " "pending=(conv=%c,lock=%c,cancel=%c,unlock=%c)\n", lock->ml.type, lock->ml.convert_type, lock->ml.node, @@ -103,30 +103,30 @@ void __dlm_print_one_lock_resource(struct dlm_lock_resource *res) stringify_lockname(res->lockname.name, res->lockname.len, buf, sizeof(buf) - 1); - printk(KERN_NOTICE "lockres: %s, owner=%u, state=%u\n", + printk("lockres: %s, owner=%u, state=%u\n", buf, res->owner, res->state); - printk(KERN_NOTICE " last used: %lu, refcnt: %u, on purge list: %s\n", + printk(" last used: %lu, refcnt: %u, on purge list: %s\n", res->last_used, atomic_read(&res->refs.refcount), list_empty(&res->purge) ? "no" : "yes"); - printk(KERN_NOTICE " on dirty list: %s, on reco list: %s, " + printk(" on dirty list: %s, on reco list: %s, " "migrating pending: %s\n", list_empty(&res->dirty) ? "no" : "yes", list_empty(&res->recovering) ? "no" : "yes", res->migration_pending ? "yes" : "no"); - printk(KERN_NOTICE " inflight locks: %d, asts reserved: %d\n", + printk(" inflight locks: %d, asts reserved: %d\n", res->inflight_locks, atomic_read(&res->asts_reserved)); dlm_print_lockres_refmap(res); - printk(KERN_NOTICE " granted queue:\n"); + printk(" granted queue:\n"); list_for_each(iter2, &res->granted) { lock = list_entry(iter2, struct dlm_lock, list); __dlm_print_lock(lock); } - printk(KERN_NOTICE " converting queue:\n"); + printk(" converting queue:\n"); list_for_each(iter2, &res->converting) { lock = list_entry(iter2, struct dlm_lock, list); __dlm_print_lock(lock); } - printk(KERN_NOTICE " blocked queue:\n"); + printk(" blocked queue:\n"); list_for_each(iter2, &res->blocked) { lock = list_entry(iter2, struct dlm_lock, list); __dlm_print_lock(lock); -- cgit v1.2.3 From 5cc3bf2786f63cceb191c3c02ddd83c6f38a7d64 Mon Sep 17 00:00:00 2001 From: Tao Ma Date: Wed, 5 Mar 2008 15:50:12 +0800 Subject: ocfs2: Reconnect after idle time out. Currently, o2net connects to a node on hb_up and disconnects on hb_down and net timeout. It disconnects on net timeout is ok, but it should attempt to reconnect back. This is because sometimes nodes get overloaded enough that the network connection breaks but the disk hb does not. And if we get into that situation, we either fence (unnecessarily) or wait for its disk hb to die (and sometimes hang in the process). So in this updated scheme, when the network disconnects, we keep attempting to reconnect till we succeed or we get a disk hb down event. If the other node is really dead, then we will eventually get a node down event. If not, we should be able to connect again and continue. Signed-off-by: Tao Ma Signed-off-by: Mark Fasheh --- fs/ocfs2/cluster/tcp.c | 51 +++++++++++++++++++++++++++++------------ fs/ocfs2/cluster/tcp_internal.h | 2 ++ 2 files changed, 38 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c index b8057c51b20..4ea4b0a2697 100644 --- a/fs/ocfs2/cluster/tcp.c +++ b/fs/ocfs2/cluster/tcp.c @@ -399,8 +399,6 @@ static void o2net_set_nn_state(struct o2net_node *nn, mlog_bug_on_msg(err && valid, "err %d valid %u\n", err, valid); mlog_bug_on_msg(valid && !sc, "valid %u sc %p\n", valid, sc); - /* we won't reconnect after our valid conn goes away for - * this hb iteration.. here so it shows up in the logs */ if (was_valid && !valid && err == 0) err = -ENOTCONN; @@ -430,11 +428,6 @@ static void o2net_set_nn_state(struct o2net_node *nn, if (!was_valid && valid) { o2quo_conn_up(o2net_num_from_nn(nn)); - /* this is a bit of a hack. we only try reconnecting - * when heartbeating starts until we get a connection. - * if that connection then dies we don't try reconnecting. - * the only way to start connecting again is to down - * heartbeat and bring it back up. */ cancel_delayed_work(&nn->nn_connect_expired); printk(KERN_INFO "o2net: %s " SC_NODEF_FMT "\n", o2nm_this_node() > sc->sc_node->nd_num ? @@ -457,6 +450,18 @@ static void o2net_set_nn_state(struct o2net_node *nn, delay = 0; mlog(ML_CONN, "queueing conn attempt in %lu jiffies\n", delay); queue_delayed_work(o2net_wq, &nn->nn_connect_work, delay); + + /* + * Delay the expired work after idle timeout. + * + * We might have lots of failed connection attempts that run + * through here but we only cancel the connect_expired work when + * a connection attempt succeeds. So only the first enqueue of + * the connect_expired work will do anything. The rest will see + * that it's already queued and do nothing. + */ + delay += msecs_to_jiffies(o2net_idle_timeout(NULL)); + queue_delayed_work(o2net_wq, &nn->nn_connect_expired, delay); } /* keep track of the nn's sc ref for the caller */ @@ -1193,6 +1198,7 @@ static int o2net_check_handshake(struct o2net_sock_container *sc) * shut down already */ if (nn->nn_sc == sc) { o2net_sc_reset_idle_timer(sc); + atomic_set(&nn->nn_timeout, 0); o2net_set_nn_state(nn, sc, 1, 0); } spin_unlock(&nn->nn_lock); @@ -1391,6 +1397,7 @@ static void o2net_sc_send_keep_req(struct work_struct *work) static void o2net_idle_timer(unsigned long data) { struct o2net_sock_container *sc = (struct o2net_sock_container *)data; + struct o2net_node *nn = o2net_nn_from_num(sc->sc_node->nd_num); struct timeval now; do_gettimeofday(&now); @@ -1413,6 +1420,12 @@ static void o2net_idle_timer(unsigned long data) sc->sc_tv_func_start.tv_sec, (long) sc->sc_tv_func_start.tv_usec, sc->sc_tv_func_stop.tv_sec, (long) sc->sc_tv_func_stop.tv_usec); + /* + * Initialize the nn_timeout so that the next connection attempt + * will continue in o2net_start_connect. + */ + atomic_set(&nn->nn_timeout, 1); + o2net_sc_queue_work(sc, &sc->sc_shutdown_work); } @@ -1447,6 +1460,7 @@ static void o2net_start_connect(struct work_struct *work) struct socket *sock = NULL; struct sockaddr_in myaddr = {0, }, remoteaddr = {0, }; int ret = 0, stop; + unsigned int timeout; /* if we're greater we initiate tx, otherwise we accept */ if (o2nm_this_node() <= o2net_num_from_nn(nn)) @@ -1466,8 +1480,17 @@ static void o2net_start_connect(struct work_struct *work) } spin_lock(&nn->nn_lock); - /* see if we already have one pending or have given up */ - stop = (nn->nn_sc || nn->nn_persistent_error); + /* + * see if we already have one pending or have given up. + * For nn_timeout, it is set when we close the connection + * because of the idle time out. So it means that we have + * at least connected to that node successfully once, + * now try to connect to it again. + */ + timeout = atomic_read(&nn->nn_timeout); + stop = (nn->nn_sc || + (nn->nn_persistent_error && + (nn->nn_persistent_error != -ENOTCONN || timeout == 0))); spin_unlock(&nn->nn_lock); if (stop) goto out; @@ -1579,6 +1602,7 @@ void o2net_disconnect_node(struct o2nm_node *node) /* don't reconnect until it's heartbeating again */ spin_lock(&nn->nn_lock); + atomic_set(&nn->nn_timeout, 0); o2net_set_nn_state(nn, NULL, 0, -ENOTCONN); spin_unlock(&nn->nn_lock); @@ -1613,17 +1637,12 @@ static void o2net_hb_node_up_cb(struct o2nm_node *node, int node_num, (msecs_to_jiffies(o2net_reconnect_delay(node)) + 1); if (node_num != o2nm_this_node()) { - /* heartbeat doesn't work unless a local node number is - * configured and doing so brings up the o2net_wq, so we can - * use it.. */ - queue_delayed_work(o2net_wq, &nn->nn_connect_expired, - msecs_to_jiffies(o2net_idle_timeout(node))); - /* believe it or not, accept and node hearbeating testing * can succeed for this node before we got here.. so * only use set_nn_state to clear the persistent error * if that hasn't already happened */ spin_lock(&nn->nn_lock); + atomic_set(&nn->nn_timeout, 0); if (nn->nn_persistent_error) o2net_set_nn_state(nn, NULL, 0, 0); spin_unlock(&nn->nn_lock); @@ -1747,6 +1766,7 @@ static int o2net_accept_one(struct socket *sock) new_sock = NULL; spin_lock(&nn->nn_lock); + atomic_set(&nn->nn_timeout, 0); o2net_set_nn_state(nn, sc, 0, 0); spin_unlock(&nn->nn_lock); @@ -1941,6 +1961,7 @@ int o2net_init(void) for (i = 0; i < ARRAY_SIZE(o2net_nodes); i++) { struct o2net_node *nn = o2net_nn_from_num(i); + atomic_set(&nn->nn_timeout, 0); spin_lock_init(&nn->nn_lock); INIT_DELAYED_WORK(&nn->nn_connect_work, o2net_start_connect); INIT_DELAYED_WORK(&nn->nn_connect_expired, diff --git a/fs/ocfs2/cluster/tcp_internal.h b/fs/ocfs2/cluster/tcp_internal.h index d25b9af2850..b4c5586f46e 100644 --- a/fs/ocfs2/cluster/tcp_internal.h +++ b/fs/ocfs2/cluster/tcp_internal.h @@ -95,6 +95,8 @@ struct o2net_node { unsigned nn_sc_valid:1; /* if this is set tx just returns it */ int nn_persistent_error; + /* It is only set to 1 after the idle time out. */ + atomic_t nn_timeout; /* threads waiting for an sc to arrive wait on the wq for generation * to increase. it is increased when a connecting socket succeeds -- cgit v1.2.3 From a839c5afcdc345aecb35b0d3bcd0e09b571329c3 Mon Sep 17 00:00:00 2001 From: Mark Fasheh Date: Tue, 29 Jan 2008 14:35:18 -0800 Subject: sysfs: Allow removal of symlinks in the sysfs root Allow callers of sysfs_remove_link() to pass a NULL kobj, in which case sysfs_root will be used as the parent directory. This allows us to tear down top level symlinks created via sysfs_create_link(), which already has similar handling of a NULL parent object. Signed-off-by: Mark Fasheh Acked-by: Greg Kroah-Hartman --- fs/sysfs/symlink.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/sysfs/symlink.c b/fs/sysfs/symlink.c index 5f66c446615..817f5966edc 100644 --- a/fs/sysfs/symlink.c +++ b/fs/sysfs/symlink.c @@ -87,7 +87,14 @@ int sysfs_create_link(struct kobject * kobj, struct kobject * target, const char void sysfs_remove_link(struct kobject * kobj, const char * name) { - sysfs_hash_and_remove(kobj->sd, name); + struct sysfs_dirent *parent_sd = NULL; + + if (!kobj) + parent_sd = &sysfs_root; + else + parent_sd = kobj->sd; + + sysfs_hash_and_remove(parent_sd, name); } static int sysfs_get_target_path(struct sysfs_dirent *parent_sd, -- cgit v1.2.3 From 52f7c21b613f80cb425d115c9e5b4ed958a133c0 Mon Sep 17 00:00:00 2001 From: Mark Fasheh Date: Tue, 29 Jan 2008 17:08:26 -0800 Subject: ocfs2: Move /sys/o2cb to /sys/fs/o2cb /sys/fs is where we really want file system specific sysfs objects. Ocfs2-tools has been updated to look in /sys/fs/o2cb. We can maintain backwards compatibility with old ocfs2-tools by using a sysfs symlink. After some time (2 years), the symlink can be safely removed. This patch also adds documentation to make it easier for people to figure out what /sys/fs/o2cb is used for. Signed-off-by: Mark Fasheh --- fs/ocfs2/cluster/sys.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'fs') diff --git a/fs/ocfs2/cluster/sys.c b/fs/ocfs2/cluster/sys.c index 0c095ce7723..98429fd6849 100644 --- a/fs/ocfs2/cluster/sys.c +++ b/fs/ocfs2/cluster/sys.c @@ -57,6 +57,7 @@ static struct kset *o2cb_kset; void o2cb_sys_shutdown(void) { mlog_sys_shutdown(); + sysfs_remove_link(NULL, "o2cb"); kset_unregister(o2cb_kset); } @@ -68,6 +69,14 @@ int o2cb_sys_init(void) if (!o2cb_kset) return -ENOMEM; + /* + * Create this symlink for backwards compatibility with old + * versions of ocfs2-tools which look for things in /sys/o2cb. + */ + ret = sysfs_create_link(NULL, &o2cb_kset->kobj, "o2cb"); + if (ret) + goto error; + ret = sysfs_create_group(&o2cb_kset->kobj, &o2cb_attr_group); if (ret) goto error; -- cgit v1.2.3 From 677b975282e48d1818df4181336307377d56b04e Mon Sep 17 00:00:00 2001 From: Tao Ma Date: Wed, 30 Jan 2008 14:21:05 +0800 Subject: ocfs2: Add support for cross extent block In ocfs2_merge_rec_left, when we find the merge extent is "CONTIG_RIGHT" with the first extent record of the next extent block, we will merge it to the next extent block and change all the related extent blocks accordingly. In ocfs2_merge_rec_right, when we find the merge extent is "CONTIG_LEFT" with the last extent record of the previous extent block, we will merge it to the prevoius extent block and change all the related extent blocks accordingly. As for CONTIG_LEFTRIGHT, we will handle CONTIG_RIGHT first so that when the index is zero, the merge process will be more efficient and easier. Signed-off-by: Tao Ma Signed-off-by: Mark Fasheh --- fs/ocfs2/alloc.c | 366 ++++++++++++++++++++++++++++++++++++++++++++++++------- 1 file changed, 325 insertions(+), 41 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c index 447206eb5c2..f63cb32e224 100644 --- a/fs/ocfs2/alloc.c +++ b/fs/ocfs2/alloc.c @@ -1450,6 +1450,8 @@ static void ocfs2_adjust_root_records(struct ocfs2_extent_list *root_el, * - When our insert into the right path leaf is at the leftmost edge * and requires an update of the path immediately to it's left. This * can occur at the end of some types of rotation and appending inserts. + * - When we've adjusted the last extent record in the left path leaf and the + * 1st extent record in the right path leaf during cross extent block merge. */ static void ocfs2_complete_edge_insert(struct inode *inode, handle_t *handle, struct ocfs2_path *left_path, @@ -2712,24 +2714,147 @@ static void ocfs2_cleanup_merge(struct ocfs2_extent_list *el, } } +static int ocfs2_get_right_path(struct inode *inode, + struct ocfs2_path *left_path, + struct ocfs2_path **ret_right_path) +{ + int ret; + u32 right_cpos; + struct ocfs2_path *right_path = NULL; + struct ocfs2_extent_list *left_el; + + *ret_right_path = NULL; + + /* This function shouldn't be called for non-trees. */ + BUG_ON(left_path->p_tree_depth == 0); + + left_el = path_leaf_el(left_path); + BUG_ON(left_el->l_next_free_rec != left_el->l_count); + + ret = ocfs2_find_cpos_for_right_leaf(inode->i_sb, left_path, + &right_cpos); + if (ret) { + mlog_errno(ret); + goto out; + } + + /* This function shouldn't be called for the rightmost leaf. */ + BUG_ON(right_cpos == 0); + + right_path = ocfs2_new_path(path_root_bh(left_path), + path_root_el(left_path)); + if (!right_path) { + ret = -ENOMEM; + mlog_errno(ret); + goto out; + } + + ret = ocfs2_find_path(inode, right_path, right_cpos); + if (ret) { + mlog_errno(ret); + goto out; + } + + *ret_right_path = right_path; +out: + if (ret) + ocfs2_free_path(right_path); + return ret; +} + /* * Remove split_rec clusters from the record at index and merge them - * onto the beginning of the record at index + 1. + * onto the beginning of the record "next" to it. + * For index < l_count - 1, the next means the extent rec at index + 1. + * For index == l_count - 1, the "next" means the 1st extent rec of the + * next extent block. */ -static int ocfs2_merge_rec_right(struct inode *inode, struct buffer_head *bh, - handle_t *handle, - struct ocfs2_extent_rec *split_rec, - struct ocfs2_extent_list *el, int index) +static int ocfs2_merge_rec_right(struct inode *inode, + struct ocfs2_path *left_path, + handle_t *handle, + struct ocfs2_extent_rec *split_rec, + int index) { - int ret; + int ret, next_free, i; unsigned int split_clusters = le16_to_cpu(split_rec->e_leaf_clusters); struct ocfs2_extent_rec *left_rec; struct ocfs2_extent_rec *right_rec; + struct ocfs2_extent_list *right_el; + struct ocfs2_path *right_path = NULL; + int subtree_index = 0; + struct ocfs2_extent_list *el = path_leaf_el(left_path); + struct buffer_head *bh = path_leaf_bh(left_path); + struct buffer_head *root_bh = NULL; BUG_ON(index >= le16_to_cpu(el->l_next_free_rec)); - left_rec = &el->l_recs[index]; - right_rec = &el->l_recs[index + 1]; + + if (index == le16_to_cpu(el->l_next_free_rec - 1) && + le16_to_cpu(el->l_next_free_rec) == le16_to_cpu(el->l_count)) { + /* we meet with a cross extent block merge. */ + ret = ocfs2_get_right_path(inode, left_path, &right_path); + if (ret) { + mlog_errno(ret); + goto out; + } + + right_el = path_leaf_el(right_path); + next_free = le16_to_cpu(right_el->l_next_free_rec); + BUG_ON(next_free <= 0); + right_rec = &right_el->l_recs[0]; + if (ocfs2_is_empty_extent(right_rec)) { + BUG_ON(le16_to_cpu(next_free) <= 1); + right_rec = &right_el->l_recs[1]; + } + + BUG_ON(le32_to_cpu(left_rec->e_cpos) + + le16_to_cpu(left_rec->e_leaf_clusters) != + le32_to_cpu(right_rec->e_cpos)); + + subtree_index = ocfs2_find_subtree_root(inode, + left_path, right_path); + + ret = ocfs2_extend_rotate_transaction(handle, subtree_index, + handle->h_buffer_credits, + right_path); + if (ret) { + mlog_errno(ret); + goto out; + } + + root_bh = left_path->p_node[subtree_index].bh; + BUG_ON(root_bh != right_path->p_node[subtree_index].bh); + + ret = ocfs2_journal_access(handle, inode, root_bh, + OCFS2_JOURNAL_ACCESS_WRITE); + if (ret) { + mlog_errno(ret); + goto out; + } + + for (i = subtree_index + 1; + i < path_num_items(right_path); i++) { + ret = ocfs2_journal_access(handle, inode, + right_path->p_node[i].bh, + OCFS2_JOURNAL_ACCESS_WRITE); + if (ret) { + mlog_errno(ret); + goto out; + } + + ret = ocfs2_journal_access(handle, inode, + left_path->p_node[i].bh, + OCFS2_JOURNAL_ACCESS_WRITE); + if (ret) { + mlog_errno(ret); + goto out; + } + } + + } else { + BUG_ON(index == le16_to_cpu(el->l_next_free_rec) - 1); + right_rec = &el->l_recs[index + 1]; + } ret = ocfs2_journal_access(handle, inode, bh, OCFS2_JOURNAL_ACCESS_WRITE); @@ -2751,30 +2876,156 @@ static int ocfs2_merge_rec_right(struct inode *inode, struct buffer_head *bh, if (ret) mlog_errno(ret); + if (right_path) { + ret = ocfs2_journal_dirty(handle, path_leaf_bh(right_path)); + if (ret) + mlog_errno(ret); + + ocfs2_complete_edge_insert(inode, handle, left_path, + right_path, subtree_index); + } +out: + if (right_path) + ocfs2_free_path(right_path); + return ret; +} + +static int ocfs2_get_left_path(struct inode *inode, + struct ocfs2_path *right_path, + struct ocfs2_path **ret_left_path) +{ + int ret; + u32 left_cpos; + struct ocfs2_path *left_path = NULL; + + *ret_left_path = NULL; + + /* This function shouldn't be called for non-trees. */ + BUG_ON(right_path->p_tree_depth == 0); + + ret = ocfs2_find_cpos_for_left_leaf(inode->i_sb, + right_path, &left_cpos); + if (ret) { + mlog_errno(ret); + goto out; + } + + /* This function shouldn't be called for the leftmost leaf. */ + BUG_ON(left_cpos == 0); + + left_path = ocfs2_new_path(path_root_bh(right_path), + path_root_el(right_path)); + if (!left_path) { + ret = -ENOMEM; + mlog_errno(ret); + goto out; + } + + ret = ocfs2_find_path(inode, left_path, left_cpos); + if (ret) { + mlog_errno(ret); + goto out; + } + + *ret_left_path = left_path; out: + if (ret) + ocfs2_free_path(left_path); return ret; } /* * Remove split_rec clusters from the record at index and merge them - * onto the tail of the record at index - 1. + * onto the tail of the record "before" it. + * For index > 0, the "before" means the extent rec at index - 1. + * + * For index == 0, the "before" means the last record of the previous + * extent block. And there is also a situation that we may need to + * remove the rightmost leaf extent block in the right_path and change + * the right path to indicate the new rightmost path. */ -static int ocfs2_merge_rec_left(struct inode *inode, struct buffer_head *bh, +static int ocfs2_merge_rec_left(struct inode *inode, + struct ocfs2_path *right_path, handle_t *handle, struct ocfs2_extent_rec *split_rec, - struct ocfs2_extent_list *el, int index) + struct ocfs2_cached_dealloc_ctxt *dealloc, + int index) { - int ret, has_empty_extent = 0; + int ret, i, subtree_index = 0, has_empty_extent = 0; unsigned int split_clusters = le16_to_cpu(split_rec->e_leaf_clusters); struct ocfs2_extent_rec *left_rec; struct ocfs2_extent_rec *right_rec; + struct ocfs2_extent_list *el = path_leaf_el(right_path); + struct buffer_head *bh = path_leaf_bh(right_path); + struct buffer_head *root_bh = NULL; + struct ocfs2_path *left_path = NULL; + struct ocfs2_extent_list *left_el; - BUG_ON(index <= 0); + BUG_ON(index < 0); - left_rec = &el->l_recs[index - 1]; right_rec = &el->l_recs[index]; - if (ocfs2_is_empty_extent(&el->l_recs[0])) - has_empty_extent = 1; + if (index == 0) { + /* we meet with a cross extent block merge. */ + ret = ocfs2_get_left_path(inode, right_path, &left_path); + if (ret) { + mlog_errno(ret); + goto out; + } + + left_el = path_leaf_el(left_path); + BUG_ON(le16_to_cpu(left_el->l_next_free_rec) != + le16_to_cpu(left_el->l_count)); + + left_rec = &left_el->l_recs[ + le16_to_cpu(left_el->l_next_free_rec) - 1]; + BUG_ON(le32_to_cpu(left_rec->e_cpos) + + le16_to_cpu(left_rec->e_leaf_clusters) != + le32_to_cpu(split_rec->e_cpos)); + + subtree_index = ocfs2_find_subtree_root(inode, + left_path, right_path); + + ret = ocfs2_extend_rotate_transaction(handle, subtree_index, + handle->h_buffer_credits, + left_path); + if (ret) { + mlog_errno(ret); + goto out; + } + + root_bh = left_path->p_node[subtree_index].bh; + BUG_ON(root_bh != right_path->p_node[subtree_index].bh); + + ret = ocfs2_journal_access(handle, inode, root_bh, + OCFS2_JOURNAL_ACCESS_WRITE); + if (ret) { + mlog_errno(ret); + goto out; + } + + for (i = subtree_index + 1; + i < path_num_items(right_path); i++) { + ret = ocfs2_journal_access(handle, inode, + right_path->p_node[i].bh, + OCFS2_JOURNAL_ACCESS_WRITE); + if (ret) { + mlog_errno(ret); + goto out; + } + + ret = ocfs2_journal_access(handle, inode, + left_path->p_node[i].bh, + OCFS2_JOURNAL_ACCESS_WRITE); + if (ret) { + mlog_errno(ret); + goto out; + } + } + } else { + left_rec = &el->l_recs[index - 1]; + if (ocfs2_is_empty_extent(&el->l_recs[0])) + has_empty_extent = 1; + } ret = ocfs2_journal_access(handle, inode, bh, OCFS2_JOURNAL_ACCESS_WRITE); @@ -2790,9 +3041,8 @@ static int ocfs2_merge_rec_left(struct inode *inode, struct buffer_head *bh, *left_rec = *split_rec; has_empty_extent = 0; - } else { + } else le16_add_cpu(&left_rec->e_leaf_clusters, split_clusters); - } le32_add_cpu(&right_rec->e_cpos, split_clusters); le64_add_cpu(&right_rec->e_blkno, @@ -2805,13 +3055,44 @@ static int ocfs2_merge_rec_left(struct inode *inode, struct buffer_head *bh, if (ret) mlog_errno(ret); + if (left_path) { + ret = ocfs2_journal_dirty(handle, path_leaf_bh(left_path)); + if (ret) + mlog_errno(ret); + + /* + * In the situation that the right_rec is empty and the extent + * block is empty also, ocfs2_complete_edge_insert can't handle + * it and we need to delete the right extent block. + */ + if (le16_to_cpu(right_rec->e_leaf_clusters) == 0 && + le16_to_cpu(el->l_next_free_rec) == 1) { + + ret = ocfs2_remove_rightmost_path(inode, handle, + right_path, dealloc); + if (ret) { + mlog_errno(ret); + goto out; + } + + /* Now the rightmost extent block has been deleted. + * So we use the new rightmost path. + */ + ocfs2_mv_path(right_path, left_path); + left_path = NULL; + } else + ocfs2_complete_edge_insert(inode, handle, left_path, + right_path, subtree_index); + } out: + if (left_path) + ocfs2_free_path(left_path); return ret; } static int ocfs2_try_to_merge_extent(struct inode *inode, handle_t *handle, - struct ocfs2_path *left_path, + struct ocfs2_path *path, int split_index, struct ocfs2_extent_rec *split_rec, struct ocfs2_cached_dealloc_ctxt *dealloc, @@ -2819,7 +3100,7 @@ static int ocfs2_try_to_merge_extent(struct inode *inode, { int ret = 0; - struct ocfs2_extent_list *el = path_leaf_el(left_path); + struct ocfs2_extent_list *el = path_leaf_el(path); struct ocfs2_extent_rec *rec = &el->l_recs[split_index]; BUG_ON(ctxt->c_contig_type == CONTIG_NONE); @@ -2832,7 +3113,7 @@ static int ocfs2_try_to_merge_extent(struct inode *inode, * extents - having more than one in a leaf is * illegal. */ - ret = ocfs2_rotate_tree_left(inode, handle, left_path, + ret = ocfs2_rotate_tree_left(inode, handle, path, dealloc); if (ret) { mlog_errno(ret); @@ -2847,7 +3128,6 @@ static int ocfs2_try_to_merge_extent(struct inode *inode, * Left-right contig implies this. */ BUG_ON(!ctxt->c_split_covers_rec); - BUG_ON(split_index == 0); /* * Since the leftright insert always covers the entire @@ -2858,9 +3138,14 @@ static int ocfs2_try_to_merge_extent(struct inode *inode, * Since the adding of an empty extent shifts * everything back to the right, there's no need to * update split_index here. + * + * When the split_index is zero, we need to merge it to the + * prevoius extent block. It is more efficient and easier + * if we do merge_right first and merge_left later. */ - ret = ocfs2_merge_rec_left(inode, path_leaf_bh(left_path), - handle, split_rec, el, split_index); + ret = ocfs2_merge_rec_right(inode, path, + handle, split_rec, + split_index); if (ret) { mlog_errno(ret); goto out; @@ -2871,32 +3156,30 @@ static int ocfs2_try_to_merge_extent(struct inode *inode, */ BUG_ON(!ocfs2_is_empty_extent(&el->l_recs[0])); - /* - * The left merge left us with an empty extent, remove - * it. - */ - ret = ocfs2_rotate_tree_left(inode, handle, left_path, dealloc); + /* The merge left us with an empty extent, remove it. */ + ret = ocfs2_rotate_tree_left(inode, handle, path, dealloc); if (ret) { mlog_errno(ret); goto out; } - split_index--; + rec = &el->l_recs[split_index]; /* * Note that we don't pass split_rec here on purpose - - * we've merged it into the left side. + * we've merged it into the rec already. */ - ret = ocfs2_merge_rec_right(inode, path_leaf_bh(left_path), - handle, rec, el, split_index); + ret = ocfs2_merge_rec_left(inode, path, + handle, rec, + dealloc, + split_index); + if (ret) { mlog_errno(ret); goto out; } - BUG_ON(!ocfs2_is_empty_extent(&el->l_recs[0])); - - ret = ocfs2_rotate_tree_left(inode, handle, left_path, + ret = ocfs2_rotate_tree_left(inode, handle, path, dealloc); /* * Error from this last rotate is not critical, so @@ -2915,8 +3198,9 @@ static int ocfs2_try_to_merge_extent(struct inode *inode, */ if (ctxt->c_contig_type == CONTIG_RIGHT) { ret = ocfs2_merge_rec_left(inode, - path_leaf_bh(left_path), - handle, split_rec, el, + path, + handle, split_rec, + dealloc, split_index); if (ret) { mlog_errno(ret); @@ -2924,8 +3208,8 @@ static int ocfs2_try_to_merge_extent(struct inode *inode, } } else { ret = ocfs2_merge_rec_right(inode, - path_leaf_bh(left_path), - handle, split_rec, el, + path, + handle, split_rec, split_index); if (ret) { mlog_errno(ret); @@ -2938,7 +3222,7 @@ static int ocfs2_try_to_merge_extent(struct inode *inode, * The merge may have left an empty extent in * our leaf. Try to rotate it away. */ - ret = ocfs2_rotate_tree_left(inode, handle, left_path, + ret = ocfs2_rotate_tree_left(inode, handle, path, dealloc); if (ret) mlog_errno(ret); -- cgit v1.2.3 From ad5a4d7093a76fa245e277e6f0f0e168a08aeff7 Mon Sep 17 00:00:00 2001 From: Tao Ma Date: Wed, 30 Jan 2008 14:21:32 +0800 Subject: ocfs2: Enable cross extent block merge. In ocfs2_figure_merge_contig_type, we judge whether there exists a cross extent block merge and enable it by setting CONTIG_LEFT and CONTIG_RIGHT accordingly. Signed-off-by: Tao Ma Signed-off-by: Mark Fasheh --- fs/ocfs2/alloc.c | 94 +++++++++++++++++++++++++++++++++++++++++++++++++++----- 1 file changed, 86 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c index f63cb32e224..7d81aa6f567 100644 --- a/fs/ocfs2/alloc.c +++ b/fs/ocfs2/alloc.c @@ -3782,20 +3782,57 @@ out: } static enum ocfs2_contig_type -ocfs2_figure_merge_contig_type(struct inode *inode, +ocfs2_figure_merge_contig_type(struct inode *inode, struct ocfs2_path *path, struct ocfs2_extent_list *el, int index, struct ocfs2_extent_rec *split_rec) { - struct ocfs2_extent_rec *rec; + int status; enum ocfs2_contig_type ret = CONTIG_NONE; + u32 left_cpos, right_cpos; + struct ocfs2_extent_rec *rec = NULL; + struct ocfs2_extent_list *new_el; + struct ocfs2_path *left_path = NULL, *right_path = NULL; + struct buffer_head *bh; + struct ocfs2_extent_block *eb; + + if (index > 0) { + rec = &el->l_recs[index - 1]; + } else if (path->p_tree_depth > 0) { + status = ocfs2_find_cpos_for_left_leaf(inode->i_sb, + path, &left_cpos); + if (status) + goto out; + + if (left_cpos != 0) { + left_path = ocfs2_new_path(path_root_bh(path), + path_root_el(path)); + if (!left_path) + goto out; + + status = ocfs2_find_path(inode, left_path, left_cpos); + if (status) + goto out; + + new_el = path_leaf_el(left_path); + + if (le16_to_cpu(new_el->l_next_free_rec) != + le16_to_cpu(new_el->l_count)) { + bh = path_leaf_bh(left_path); + eb = (struct ocfs2_extent_block *)bh->b_data; + OCFS2_RO_ON_INVALID_EXTENT_BLOCK(inode->i_sb, + eb); + goto out; + } + rec = &new_el->l_recs[ + le16_to_cpu(new_el->l_next_free_rec) - 1]; + } + } /* * We're careful to check for an empty extent record here - * the merge code will know what to do if it sees one. */ - - if (index > 0) { - rec = &el->l_recs[index - 1]; + if (rec) { if (index == 1 && ocfs2_is_empty_extent(rec)) { if (split_rec->e_cpos == el->l_recs[index].e_cpos) ret = CONTIG_RIGHT; @@ -3804,10 +3841,45 @@ ocfs2_figure_merge_contig_type(struct inode *inode, } } - if (index < (le16_to_cpu(el->l_next_free_rec) - 1)) { + rec = NULL; + if (index < (le16_to_cpu(el->l_next_free_rec) - 1)) + rec = &el->l_recs[index + 1]; + else if (le16_to_cpu(el->l_next_free_rec) == le16_to_cpu(el->l_count) && + path->p_tree_depth > 0) { + status = ocfs2_find_cpos_for_right_leaf(inode->i_sb, + path, &right_cpos); + if (status) + goto out; + + if (right_cpos == 0) + goto out; + + right_path = ocfs2_new_path(path_root_bh(path), + path_root_el(path)); + if (!right_path) + goto out; + + status = ocfs2_find_path(inode, right_path, right_cpos); + if (status) + goto out; + + new_el = path_leaf_el(right_path); + rec = &new_el->l_recs[0]; + if (ocfs2_is_empty_extent(rec)) { + if (le16_to_cpu(new_el->l_next_free_rec) <= 1) { + bh = path_leaf_bh(right_path); + eb = (struct ocfs2_extent_block *)bh->b_data; + OCFS2_RO_ON_INVALID_EXTENT_BLOCK(inode->i_sb, + eb); + goto out; + } + rec = &new_el->l_recs[1]; + } + } + + if (rec) { enum ocfs2_contig_type contig_type; - rec = &el->l_recs[index + 1]; contig_type = ocfs2_extent_contig(inode, rec, split_rec); if (contig_type == CONTIG_LEFT && ret == CONTIG_RIGHT) @@ -3816,6 +3888,12 @@ ocfs2_figure_merge_contig_type(struct inode *inode, ret = contig_type; } +out: + if (left_path) + ocfs2_free_path(left_path); + if (right_path) + ocfs2_free_path(right_path); + return ret; } @@ -4278,7 +4356,7 @@ static int __ocfs2_mark_extent_written(struct inode *inode, goto out; } - ctxt.c_contig_type = ocfs2_figure_merge_contig_type(inode, el, + ctxt.c_contig_type = ocfs2_figure_merge_contig_type(inode, path, el, split_index, split_rec); -- cgit v1.2.3 From ffda89a3bf3b968bdc268584c6bc1da5c173cf12 Mon Sep 17 00:00:00 2001 From: Tao Ma Date: Mon, 3 Mar 2008 17:12:09 +0800 Subject: ocfs2: Add a new parameter for ocfs2_reserve_suballoc_bits In some cases(Inode stealing from other nodes), we may not want ocfs2_reserve_suballoc_bits to allocate new groups from the global_bitmap since it may already be full. So add a new parameter for this. Signed-off-by: Tao Ma Signed-off-by: Sunil Mushran Signed-off-by: Mark Fasheh --- fs/ocfs2/suballoc.c | 22 ++++++++++++++++++---- 1 file changed, 18 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c index 72c198a004d..3be4e73e8b1 100644 --- a/fs/ocfs2/suballoc.c +++ b/fs/ocfs2/suballoc.c @@ -46,6 +46,9 @@ #include "buffer_head_io.h" +#define NOT_ALLOC_NEW_GROUP 0 +#define ALLOC_NEW_GROUP 1 + static inline void ocfs2_debug_bg(struct ocfs2_group_desc *bg); static inline void ocfs2_debug_suballoc_inode(struct ocfs2_dinode *fe); static inline u16 ocfs2_find_victim_chain(struct ocfs2_chain_list *cl); @@ -391,7 +394,8 @@ bail: static int ocfs2_reserve_suballoc_bits(struct ocfs2_super *osb, struct ocfs2_alloc_context *ac, int type, - u32 slot) + u32 slot, + int alloc_new_group) { int status; u32 bits_wanted = ac->ac_bits_wanted; @@ -446,6 +450,14 @@ static int ocfs2_reserve_suballoc_bits(struct ocfs2_super *osb, goto bail; } + if (alloc_new_group != ALLOC_NEW_GROUP) { + mlog(0, "Alloc File %u Full: wanted=%u, free_bits=%u, " + "and we don't alloc a new group for it.\n", + slot, bits_wanted, free_bits); + status = -ENOSPC; + goto bail; + } + status = ocfs2_block_group_alloc(osb, alloc_inode, bh); if (status < 0) { if (status != -ENOSPC) @@ -490,7 +502,8 @@ int ocfs2_reserve_new_metadata(struct ocfs2_super *osb, (*ac)->ac_group_search = ocfs2_block_group_search; status = ocfs2_reserve_suballoc_bits(osb, (*ac), - EXTENT_ALLOC_SYSTEM_INODE, slot); + EXTENT_ALLOC_SYSTEM_INODE, + slot, ALLOC_NEW_GROUP); if (status < 0) { if (status != -ENOSPC) mlog_errno(status); @@ -527,7 +540,7 @@ int ocfs2_reserve_new_inode(struct ocfs2_super *osb, status = ocfs2_reserve_suballoc_bits(osb, *ac, INODE_ALLOC_SYSTEM_INODE, - osb->slot_num); + osb->slot_num, ALLOC_NEW_GROUP); if (status < 0) { if (status != -ENOSPC) mlog_errno(status); @@ -557,7 +570,8 @@ int ocfs2_reserve_cluster_bitmap_bits(struct ocfs2_super *osb, status = ocfs2_reserve_suballoc_bits(osb, ac, GLOBAL_BITMAP_SYSTEM_INODE, - OCFS2_INVALID_SLOT); + OCFS2_INVALID_SLOT, + ALLOC_NEW_GROUP); if (status < 0 && status != -ENOSPC) { mlog_errno(status); goto bail; -- cgit v1.2.3 From a4a4891164d4f6f383cc17e7c90828a7ca6a1146 Mon Sep 17 00:00:00 2001 From: Tao Ma Date: Mon, 3 Mar 2008 17:12:30 +0800 Subject: ocfs2: Add ac_alloc_slot in ocfs2_alloc_context In inode stealing, we no longer restrict the allocation to happen in the local node. So it is neccessary for us to add a new member in ocfs2_alloc_context to indicate which slot we are using for allocation. We also modify the process of local alloc so that this member can be used there also. Signed-off-by: Tao Ma Signed-off-by: Sunil Mushran Signed-off-by: Mark Fasheh --- fs/ocfs2/localalloc.c | 2 ++ fs/ocfs2/suballoc.c | 1 + fs/ocfs2/suballoc.h | 1 + 3 files changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/ocfs2/localalloc.c b/fs/ocfs2/localalloc.c index ab83fd56242..b6d07198118 100644 --- a/fs/ocfs2/localalloc.c +++ b/fs/ocfs2/localalloc.c @@ -523,6 +523,8 @@ int ocfs2_reserve_local_alloc_bits(struct ocfs2_super *osb, } ac->ac_inode = local_alloc_inode; + /* We should never use localalloc from another slot */ + ac->ac_alloc_slot = osb->slot_num; ac->ac_which = OCFS2_AC_USE_LOCAL; get_bh(osb->local_alloc_bh); ac->ac_bh = osb->local_alloc_bh; diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c index 3be4e73e8b1..33d55734c51 100644 --- a/fs/ocfs2/suballoc.c +++ b/fs/ocfs2/suballoc.c @@ -424,6 +424,7 @@ static int ocfs2_reserve_suballoc_bits(struct ocfs2_super *osb, } ac->ac_inode = alloc_inode; + ac->ac_alloc_slot = slot; fe = (struct ocfs2_dinode *) bh->b_data; if (!OCFS2_IS_VALID_DINODE(fe)) { diff --git a/fs/ocfs2/suballoc.h b/fs/ocfs2/suballoc.h index 8799033bb45..544c600662b 100644 --- a/fs/ocfs2/suballoc.h +++ b/fs/ocfs2/suballoc.h @@ -36,6 +36,7 @@ typedef int (group_search_t)(struct inode *, struct ocfs2_alloc_context { struct inode *ac_inode; /* which bitmap are we allocating from? */ struct buffer_head *ac_bh; /* file entry bh */ + u32 ac_alloc_slot; /* which slot are we allocating from? */ u32 ac_bits_wanted; u32 ac_bits_given; #define OCFS2_AC_USE_LOCAL 1 -- cgit v1.2.3 From 4d0ddb2ce25db2254d468233d942276ecf40bff8 Mon Sep 17 00:00:00 2001 From: Tao Ma Date: Wed, 5 Mar 2008 16:11:46 +0800 Subject: ocfs2: Add inode stealing for ocfs2_reserve_new_inode Inode allocation is modified to look in other nodes allocators during extreme out of space situations. We retry our own slot when space is freed back to the global bitmap, or whenever we've allocated more than 1024 inodes from another slot. Signed-off-by: Tao Ma Signed-off-by: Mark Fasheh --- fs/ocfs2/alloc.c | 2 ++ fs/ocfs2/localalloc.c | 2 ++ fs/ocfs2/namei.c | 2 +- fs/ocfs2/ocfs2.h | 34 ++++++++++++++++++++-- fs/ocfs2/suballoc.c | 80 +++++++++++++++++++++++++++++++++++++++++++++++++-- fs/ocfs2/super.c | 1 + 6 files changed, 116 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c index 7d81aa6f567..a26882144e8 100644 --- a/fs/ocfs2/alloc.c +++ b/fs/ocfs2/alloc.c @@ -5150,6 +5150,8 @@ static void ocfs2_truncate_log_worker(struct work_struct *work) status = ocfs2_flush_truncate_log(osb); if (status < 0) mlog_errno(status); + else + ocfs2_init_inode_steal_slot(osb); mlog_exit(status); } diff --git a/fs/ocfs2/localalloc.c b/fs/ocfs2/localalloc.c index b6d07198118..ce0dc147602 100644 --- a/fs/ocfs2/localalloc.c +++ b/fs/ocfs2/localalloc.c @@ -447,6 +447,8 @@ out_mutex: iput(main_bm_inode); out: + if (!status) + ocfs2_init_inode_steal_slot(osb); mlog_exit(status); return status; } diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c index ae9ad958751..ab5a2272d0e 100644 --- a/fs/ocfs2/namei.c +++ b/fs/ocfs2/namei.c @@ -424,7 +424,7 @@ static int ocfs2_mknod_locked(struct ocfs2_super *osb, fe->i_fs_generation = cpu_to_le32(osb->fs_generation); fe->i_blkno = cpu_to_le64(fe_blkno); fe->i_suballoc_bit = cpu_to_le16(suballoc_bit); - fe->i_suballoc_slot = cpu_to_le16(osb->slot_num); + fe->i_suballoc_slot = cpu_to_le16(inode_ac->ac_alloc_slot); fe->i_uid = cpu_to_le32(current->fsuid); if (dir->i_mode & S_ISGID) { fe->i_gid = cpu_to_le32(dir->i_gid); diff --git a/fs/ocfs2/ocfs2.h b/fs/ocfs2/ocfs2.h index 9ff5811345a..31692379c17 100644 --- a/fs/ocfs2/ocfs2.h +++ b/fs/ocfs2/ocfs2.h @@ -208,11 +208,14 @@ struct ocfs2_super u32 s_feature_incompat; u32 s_feature_ro_compat; - /* Protects s_next_generaion, osb_flags. Could protect more on - * osb as it's very short lived. */ + /* Protects s_next_generation, osb_flags and s_inode_steal_slot. + * Could protect more on osb as it's very short lived. + */ spinlock_t osb_lock; u32 s_next_generation; unsigned long osb_flags; + s16 s_inode_steal_slot; + atomic_t s_num_inodes_stolen; unsigned long s_mount_opt; unsigned int s_atime_quantum; @@ -537,6 +540,33 @@ static inline unsigned int ocfs2_pages_per_cluster(struct super_block *sb) return pages_per_cluster; } +static inline void ocfs2_init_inode_steal_slot(struct ocfs2_super *osb) +{ + spin_lock(&osb->osb_lock); + osb->s_inode_steal_slot = OCFS2_INVALID_SLOT; + spin_unlock(&osb->osb_lock); + atomic_set(&osb->s_num_inodes_stolen, 0); +} + +static inline void ocfs2_set_inode_steal_slot(struct ocfs2_super *osb, + s16 slot) +{ + spin_lock(&osb->osb_lock); + osb->s_inode_steal_slot = slot; + spin_unlock(&osb->osb_lock); +} + +static inline s16 ocfs2_get_inode_steal_slot(struct ocfs2_super *osb) +{ + s16 slot; + + spin_lock(&osb->osb_lock); + slot = osb->s_inode_steal_slot; + spin_unlock(&osb->osb_lock); + + return slot; +} + #define ocfs2_set_bit ext2_set_bit #define ocfs2_clear_bit ext2_clear_bit #define ocfs2_test_bit ext2_test_bit diff --git a/fs/ocfs2/suballoc.c b/fs/ocfs2/suballoc.c index 33d55734c51..d2d278fb981 100644 --- a/fs/ocfs2/suballoc.c +++ b/fs/ocfs2/suballoc.c @@ -49,6 +49,8 @@ #define NOT_ALLOC_NEW_GROUP 0 #define ALLOC_NEW_GROUP 1 +#define OCFS2_MAX_INODES_TO_STEAL 1024 + static inline void ocfs2_debug_bg(struct ocfs2_group_desc *bg); static inline void ocfs2_debug_suballoc_inode(struct ocfs2_dinode *fe); static inline u16 ocfs2_find_victim_chain(struct ocfs2_chain_list *cl); @@ -109,7 +111,7 @@ static inline void ocfs2_block_to_cluster_group(struct inode *inode, u64 *bg_blkno, u16 *bg_bit_off); -void ocfs2_free_alloc_context(struct ocfs2_alloc_context *ac) +static void ocfs2_free_ac_resource(struct ocfs2_alloc_context *ac) { struct inode *inode = ac->ac_inode; @@ -120,9 +122,17 @@ void ocfs2_free_alloc_context(struct ocfs2_alloc_context *ac) mutex_unlock(&inode->i_mutex); iput(inode); + ac->ac_inode = NULL; } - if (ac->ac_bh) + if (ac->ac_bh) { brelse(ac->ac_bh); + ac->ac_bh = NULL; + } +} + +void ocfs2_free_alloc_context(struct ocfs2_alloc_context *ac) +{ + ocfs2_free_ac_resource(ac); kfree(ac); } @@ -522,10 +532,42 @@ bail: return status; } +static int ocfs2_steal_inode_from_other_nodes(struct ocfs2_super *osb, + struct ocfs2_alloc_context *ac) +{ + int i, status = -ENOSPC; + s16 slot = ocfs2_get_inode_steal_slot(osb); + + /* Start to steal inodes from the first slot after ours. */ + if (slot == OCFS2_INVALID_SLOT) + slot = osb->slot_num + 1; + + for (i = 0; i < osb->max_slots; i++, slot++) { + if (slot == osb->max_slots) + slot = 0; + + if (slot == osb->slot_num) + continue; + + status = ocfs2_reserve_suballoc_bits(osb, ac, + INODE_ALLOC_SYSTEM_INODE, + slot, NOT_ALLOC_NEW_GROUP); + if (status >= 0) { + ocfs2_set_inode_steal_slot(osb, slot); + break; + } + + ocfs2_free_ac_resource(ac); + } + + return status; +} + int ocfs2_reserve_new_inode(struct ocfs2_super *osb, struct ocfs2_alloc_context **ac) { int status; + s16 slot = ocfs2_get_inode_steal_slot(osb); *ac = kzalloc(sizeof(struct ocfs2_alloc_context), GFP_KERNEL); if (!(*ac)) { @@ -539,9 +581,43 @@ int ocfs2_reserve_new_inode(struct ocfs2_super *osb, (*ac)->ac_group_search = ocfs2_block_group_search; + /* + * slot is set when we successfully steal inode from other nodes. + * It is reset in 3 places: + * 1. when we flush the truncate log + * 2. when we complete local alloc recovery. + * 3. when we successfully allocate from our own slot. + * After it is set, we will go on stealing inodes until we find the + * need to check our slots to see whether there is some space for us. + */ + if (slot != OCFS2_INVALID_SLOT && + atomic_read(&osb->s_num_inodes_stolen) < OCFS2_MAX_INODES_TO_STEAL) + goto inode_steal; + + atomic_set(&osb->s_num_inodes_stolen, 0); status = ocfs2_reserve_suballoc_bits(osb, *ac, INODE_ALLOC_SYSTEM_INODE, osb->slot_num, ALLOC_NEW_GROUP); + if (status >= 0) { + status = 0; + + /* + * Some inodes must be freed by us, so try to allocate + * from our own next time. + */ + if (slot != OCFS2_INVALID_SLOT) + ocfs2_init_inode_steal_slot(osb); + goto bail; + } else if (status < 0 && status != -ENOSPC) { + mlog_errno(status); + goto bail; + } + + ocfs2_free_ac_resource(*ac); + +inode_steal: + status = ocfs2_steal_inode_from_other_nodes(osb, *ac); + atomic_inc(&osb->s_num_inodes_stolen); if (status < 0) { if (status != -ENOSPC) mlog_errno(status); diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index 96ebe36d5d7..df63ba20ae9 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -1394,6 +1394,7 @@ static int ocfs2_initialize_super(struct super_block *sb, INIT_LIST_HEAD(&osb->blocked_lock_list); osb->blocked_lock_count = 0; spin_lock_init(&osb->osb_lock); + ocfs2_init_inode_steal_slot(osb); atomic_set(&osb->alloc_stats.moves, 0); atomic_set(&osb->alloc_stats.local_data, 0); -- cgit v1.2.3 From 58dadcdbc2584db050969f9781727fc5a3f618db Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Fri, 28 Mar 2008 14:43:10 -0700 Subject: fs/ocfs2/aops.c: test for IS_ERR rather than 0 The function ocfs2_start_trans always returns either a valid pointer or a value made with ERR_PTR, so its result should be tested with IS_ERR, not with a test for 0. Signed-off-by: Julia Lawall Signed-off-by: Andrew Morton Signed-off-by: Mark Fasheh --- fs/ocfs2/aops.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c index 90383ed6100..17964c0505a 100644 --- a/fs/ocfs2/aops.c +++ b/fs/ocfs2/aops.c @@ -467,11 +467,11 @@ handle_t *ocfs2_start_walk_page_trans(struct inode *inode, unsigned to) { struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); - handle_t *handle = NULL; + handle_t *handle; int ret = 0; handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS); - if (!handle) { + if (IS_ERR(handle)) { ret = -ENOMEM; mlog_errno(ret); goto out; @@ -487,7 +487,7 @@ handle_t *ocfs2_start_walk_page_trans(struct inode *inode, } out: if (ret) { - if (handle) + if (!IS_ERR(handle)) ocfs2_commit_trans(osb, handle); handle = ERR_PTR(ret); } -- cgit v1.2.3 From 5dabd69515765156605b09261abf969236a77803 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Thu, 21 Feb 2008 18:00:00 +0100 Subject: ocfs2: Improve rename locking ocfs2_rename() was being too aggressive with the rename lock - we only need it for certain forms of directory rename. Signed-off-by: Jan Kara Signed-off-by: Mark Fasheh --- fs/ocfs2/namei.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c index ab5a2272d0e..d5d808fe014 100644 --- a/fs/ocfs2/namei.c +++ b/fs/ocfs2/namei.c @@ -997,7 +997,7 @@ static int ocfs2_rename(struct inode *old_dir, * * And that's why, just like the VFS, we need a file system * rename lock. */ - if (old_dentry != new_dentry) { + if (old_dir != new_dir && S_ISDIR(old_inode->i_mode)) { status = ocfs2_rename_lock(osb); if (status < 0) { mlog_errno(status); -- cgit v1.2.3 From c9ec14884d69a303eef4faae42bd3c4e25b19941 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Sun, 27 Jan 2008 03:17:17 +0100 Subject: ocfs2: Convert ocfs2 over to unlocked_ioctl As far as I can see there is nothing in ocfs2_ioctl that requires the BKL, so use unlocked_ioctl Signed-off-by: Andi Kleen Signed-off-by: Mark Fasheh --- fs/ocfs2/file.c | 4 ++-- fs/ocfs2/ioctl.c | 12 +++--------- fs/ocfs2/ioctl.h | 3 +-- 3 files changed, 6 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index ed5d5232e85..9154c82d325 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -2242,7 +2242,7 @@ const struct file_operations ocfs2_fops = { .open = ocfs2_file_open, .aio_read = ocfs2_file_aio_read, .aio_write = ocfs2_file_aio_write, - .ioctl = ocfs2_ioctl, + .unlocked_ioctl = ocfs2_ioctl, #ifdef CONFIG_COMPAT .compat_ioctl = ocfs2_compat_ioctl, #endif @@ -2258,7 +2258,7 @@ const struct file_operations ocfs2_dops = { .fsync = ocfs2_sync_file, .release = ocfs2_dir_release, .open = ocfs2_dir_open, - .ioctl = ocfs2_ioctl, + .unlocked_ioctl = ocfs2_ioctl, #ifdef CONFIG_COMPAT .compat_ioctl = ocfs2_compat_ioctl, #endif diff --git a/fs/ocfs2/ioctl.c b/fs/ocfs2/ioctl.c index ab1c2167d7f..b413166dd16 100644 --- a/fs/ocfs2/ioctl.c +++ b/fs/ocfs2/ioctl.c @@ -113,9 +113,9 @@ bail: return status; } -int ocfs2_ioctl(struct inode * inode, struct file * filp, - unsigned int cmd, unsigned long arg) +long ocfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) { + struct inode *inode = filp->f_path.dentry->d_inode; unsigned int flags; int new_clusters; int status; @@ -169,9 +169,6 @@ int ocfs2_ioctl(struct inode * inode, struct file * filp, #ifdef CONFIG_COMPAT long ocfs2_compat_ioctl(struct file *file, unsigned cmd, unsigned long arg) { - struct inode *inode = file->f_path.dentry->d_inode; - int ret; - switch (cmd) { case OCFS2_IOC32_GETFLAGS: cmd = OCFS2_IOC_GETFLAGS; @@ -191,9 +188,6 @@ long ocfs2_compat_ioctl(struct file *file, unsigned cmd, unsigned long arg) return -ENOIOCTLCMD; } - lock_kernel(); - ret = ocfs2_ioctl(inode, file, cmd, arg); - unlock_kernel(); - return ret; + return ocfs2_ioctl(file, cmd, arg); } #endif diff --git a/fs/ocfs2/ioctl.h b/fs/ocfs2/ioctl.h index 4d6c4f430d0..cf9a5ee30fe 100644 --- a/fs/ocfs2/ioctl.h +++ b/fs/ocfs2/ioctl.h @@ -10,8 +10,7 @@ #ifndef OCFS2_IOCTL_H #define OCFS2_IOCTL_H -int ocfs2_ioctl(struct inode * inode, struct file * filp, - unsigned int cmd, unsigned long arg); +long ocfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg); long ocfs2_compat_ioctl(struct file *file, unsigned cmd, unsigned long arg); #endif /* OCFS2_IOCTL_H */ -- cgit v1.2.3 From b1f3550fa1471b691ad6c2f35b5b22e93eaa5855 Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Tue, 4 Mar 2008 15:21:05 -0800 Subject: ocfs2: Use BUG_ON if (...) BUG(); should be replaced with BUG_ON(...) when the test has no side-effects to allow a definition of BUG_ON that drops the code completely. The semantic patch that makes this change is as follows: (http://www.emn.fr/x-info/coccinelle/) // @ disable unlikely @ expression E,f; @@ ( if (<... f(...) ...>) { BUG(); } | - if (unlikely(E)) { BUG(); } + BUG_ON(E); ) @@ expression E,f; @@ ( if (<... f(...) ...>) { BUG(); } | - if (E) { BUG(); } + BUG_ON(E); ) // Signed-off-by: Julia Lawall Signed-off-by: Andrew Morton Signed-off-by: Mark Fasheh --- fs/ocfs2/alloc.c | 3 +-- fs/ocfs2/journal.c | 3 +-- 2 files changed, 2 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c index a26882144e8..41f84c92094 100644 --- a/fs/ocfs2/alloc.c +++ b/fs/ocfs2/alloc.c @@ -1029,8 +1029,7 @@ static void ocfs2_rotate_leaf(struct ocfs2_extent_list *el, BUG_ON(!next_free); /* The tree code before us didn't allow enough room in the leaf. */ - if (el->l_next_free_rec == el->l_count && !has_empty) - BUG(); + BUG_ON(el->l_next_free_rec == el->l_count && !has_empty); /* * The easiest way to approach this is to just remove the diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c index bffd2d714f6..9698338adc3 100644 --- a/fs/ocfs2/journal.c +++ b/fs/ocfs2/journal.c @@ -717,8 +717,7 @@ int ocfs2_journal_load(struct ocfs2_journal *journal, int local) mlog_entry_void(); - if (!journal) - BUG(); + BUG_ON(!journal); osb = journal->j_osb; -- cgit v1.2.3 From 409753bf6da4a2db038027471abaf324e063db2f Mon Sep 17 00:00:00 2001 From: Jeff Mahoney Date: Fri, 28 Mar 2008 16:44:13 -0700 Subject: ocfs2/cluster: Get rid of arguments to the timeout routines We keep seeing bug reports related to NULL pointer derefs in o2net_set_nn_state(). When I originally wrote up the configurable timeout patch, I had tried to plan for multiple clusters. This was silly. The timeout routines all use o2nm_single_cluster so there's no point in passing an argument at all. This patch removes the arguments and kills those bugs dead. Signed-off-by: Jeff Mahoney Signed-off-by: Mark Fasheh --- fs/ocfs2/cluster/tcp.c | 47 ++++++++++++++++++++--------------------------- 1 file changed, 20 insertions(+), 27 deletions(-) (limited to 'fs') diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c index 4ea4b0a2697..1170918a931 100644 --- a/fs/ocfs2/cluster/tcp.c +++ b/fs/ocfs2/cluster/tcp.c @@ -142,23 +142,17 @@ static void o2net_idle_timer(unsigned long data); static void o2net_sc_postpone_idle(struct o2net_sock_container *sc); static void o2net_sc_reset_idle_timer(struct o2net_sock_container *sc); -/* - * FIXME: These should use to_o2nm_cluster_from_node(), but we end up - * losing our parent link to the cluster during shutdown. This can be - * solved by adding a pre-removal callback to configfs, or passing - * around the cluster with the node. -jeffm - */ -static inline int o2net_reconnect_delay(struct o2nm_node *node) +static inline int o2net_reconnect_delay(void) { return o2nm_single_cluster->cl_reconnect_delay_ms; } -static inline int o2net_keepalive_delay(struct o2nm_node *node) +static inline int o2net_keepalive_delay(void) { return o2nm_single_cluster->cl_keepalive_delay_ms; } -static inline int o2net_idle_timeout(struct o2nm_node *node) +static inline int o2net_idle_timeout(void) { return o2nm_single_cluster->cl_idle_timeout_ms; } @@ -444,9 +438,9 @@ static void o2net_set_nn_state(struct o2net_node *nn, /* delay if we're withing a RECONNECT_DELAY of the * last attempt */ delay = (nn->nn_last_connect_attempt + - msecs_to_jiffies(o2net_reconnect_delay(NULL))) + msecs_to_jiffies(o2net_reconnect_delay())) - jiffies; - if (delay > msecs_to_jiffies(o2net_reconnect_delay(NULL))) + if (delay > msecs_to_jiffies(o2net_reconnect_delay())) delay = 0; mlog(ML_CONN, "queueing conn attempt in %lu jiffies\n", delay); queue_delayed_work(o2net_wq, &nn->nn_connect_work, delay); @@ -460,7 +454,7 @@ static void o2net_set_nn_state(struct o2net_node *nn, * the connect_expired work will do anything. The rest will see * that it's already queued and do nothing. */ - delay += msecs_to_jiffies(o2net_idle_timeout(NULL)); + delay += msecs_to_jiffies(o2net_idle_timeout()); queue_delayed_work(o2net_wq, &nn->nn_connect_expired, delay); } @@ -1159,23 +1153,23 @@ static int o2net_check_handshake(struct o2net_sock_container *sc) * but isn't. This can ultimately cause corruption. */ if (be32_to_cpu(hand->o2net_idle_timeout_ms) != - o2net_idle_timeout(sc->sc_node)) { + o2net_idle_timeout()) { mlog(ML_NOTICE, SC_NODEF_FMT " uses a network idle timeout of " "%u ms, but we use %u ms locally. disconnecting\n", SC_NODEF_ARGS(sc), be32_to_cpu(hand->o2net_idle_timeout_ms), - o2net_idle_timeout(sc->sc_node)); + o2net_idle_timeout()); o2net_ensure_shutdown(nn, sc, -ENOTCONN); return -1; } if (be32_to_cpu(hand->o2net_keepalive_delay_ms) != - o2net_keepalive_delay(sc->sc_node)) { + o2net_keepalive_delay()) { mlog(ML_NOTICE, SC_NODEF_FMT " uses a keepalive delay of " "%u ms, but we use %u ms locally. disconnecting\n", SC_NODEF_ARGS(sc), be32_to_cpu(hand->o2net_keepalive_delay_ms), - o2net_keepalive_delay(sc->sc_node)); + o2net_keepalive_delay()); o2net_ensure_shutdown(nn, sc, -ENOTCONN); return -1; } @@ -1353,12 +1347,11 @@ static void o2net_initialize_handshake(void) { o2net_hand->o2hb_heartbeat_timeout_ms = cpu_to_be32( O2HB_MAX_WRITE_TIMEOUT_MS); - o2net_hand->o2net_idle_timeout_ms = cpu_to_be32( - o2net_idle_timeout(NULL)); + o2net_hand->o2net_idle_timeout_ms = cpu_to_be32(o2net_idle_timeout()); o2net_hand->o2net_keepalive_delay_ms = cpu_to_be32( - o2net_keepalive_delay(NULL)); + o2net_keepalive_delay()); o2net_hand->o2net_reconnect_delay_ms = cpu_to_be32( - o2net_reconnect_delay(NULL)); + o2net_reconnect_delay()); } /* ------------------------------------------------------------ */ @@ -1404,8 +1397,8 @@ static void o2net_idle_timer(unsigned long data) printk(KERN_INFO "o2net: connection to " SC_NODEF_FMT " has been idle for %u.%u " "seconds, shutting it down.\n", SC_NODEF_ARGS(sc), - o2net_idle_timeout(sc->sc_node) / 1000, - o2net_idle_timeout(sc->sc_node) % 1000); + o2net_idle_timeout() / 1000, + o2net_idle_timeout() % 1000); mlog(ML_NOTICE, "here are some times that might help debug the " "situation: (tmr %ld.%ld now %ld.%ld dr %ld.%ld adv " "%ld.%ld:%ld.%ld func (%08x:%u) %ld.%ld:%ld.%ld)\n", @@ -1433,10 +1426,10 @@ static void o2net_sc_reset_idle_timer(struct o2net_sock_container *sc) { o2net_sc_cancel_delayed_work(sc, &sc->sc_keepalive_work); o2net_sc_queue_delayed_work(sc, &sc->sc_keepalive_work, - msecs_to_jiffies(o2net_keepalive_delay(sc->sc_node))); + msecs_to_jiffies(o2net_keepalive_delay())); do_gettimeofday(&sc->sc_tv_timer); mod_timer(&sc->sc_idle_timeout, - jiffies + msecs_to_jiffies(o2net_idle_timeout(sc->sc_node))); + jiffies + msecs_to_jiffies(o2net_idle_timeout())); } static void o2net_sc_postpone_idle(struct o2net_sock_container *sc) @@ -1578,8 +1571,8 @@ static void o2net_connect_expired(struct work_struct *work) mlog(ML_ERROR, "no connection established with node %u after " "%u.%u seconds, giving up and returning errors.\n", o2net_num_from_nn(nn), - o2net_idle_timeout(NULL) / 1000, - o2net_idle_timeout(NULL) % 1000); + o2net_idle_timeout() / 1000, + o2net_idle_timeout() % 1000); o2net_set_nn_state(nn, NULL, 0, -ENOTCONN); } @@ -1634,7 +1627,7 @@ static void o2net_hb_node_up_cb(struct o2nm_node *node, int node_num, /* ensure an immediate connect attempt */ nn->nn_last_connect_attempt = jiffies - - (msecs_to_jiffies(o2net_reconnect_delay(node)) + 1); + (msecs_to_jiffies(o2net_reconnect_delay()) + 1); if (node_num != o2nm_this_node()) { /* believe it or not, accept and node hearbeating testing -- cgit v1.2.3 From 93b06edb5127315473d87e075b2b1d1acf74659c Mon Sep 17 00:00:00 2001 From: Mark Fasheh Date: Fri, 4 Apr 2008 12:45:55 -0700 Subject: ocfs2: Only build ocfs2/dlm with the o2cb stack module fs/ocfs2/dlm/ocfs2_dlm.ko and fs/ocfs2/dlm/ocfs2_dlmfs.ko get built if CONFIG_FS_OCFS2 is specified. This isn't quite how it should happen any more - the "o2cb" dlm modules should only be built if CONFIG_FS_OCFS2_O2CB is set, so update the dlm Makefile accordingly. Signed-off-by: Mark Fasheh Acked-by: Randy Dunlap Acked-by: Joel Becker --- fs/ocfs2/dlm/Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ocfs2/dlm/Makefile b/fs/ocfs2/dlm/Makefile index ce3f7c29d27..19036137570 100644 --- a/fs/ocfs2/dlm/Makefile +++ b/fs/ocfs2/dlm/Makefile @@ -1,6 +1,6 @@ EXTRA_CFLAGS += -Ifs/ocfs2 -obj-$(CONFIG_OCFS2_FS) += ocfs2_dlm.o ocfs2_dlmfs.o +obj-$(CONFIG_OCFS2_FS_O2CB) += ocfs2_dlm.o ocfs2_dlmfs.o ocfs2_dlm-objs := dlmdomain.o dlmdebug.o dlmthread.o dlmrecovery.o \ dlmmaster.o dlmast.o dlmconvert.o dlmlock.o dlmunlock.o dlmver.o -- cgit v1.2.3 From 2309e9e040fe29469fb85a384636c455b62fe525 Mon Sep 17 00:00:00 2001 From: Sunil Mushran Date: Mon, 14 Apr 2008 10:46:19 -0700 Subject: ocfs2/net: Add debug interface to o2net This patch exposes o2net information via debugfs. The information includes the list of sockets (sock_containers) as well as the list of outstanding messages (send_tracking). Useful for o2dlm debugging. (This patch is derived from an earlier one written by Zach Brown that exposed the same information via /proc.) [Mark: checkpatch fixes] Signed-off-by: Sunil Mushran Reviewed-by: Joel Becker Signed-off-by: Mark Fasheh --- fs/ocfs2/cluster/Makefile | 2 +- fs/ocfs2/cluster/netdebug.c | 441 ++++++++++++++++++++++++++++++++++++++++ fs/ocfs2/cluster/nodemanager.c | 5 +- fs/ocfs2/cluster/tcp.c | 68 +++++++ fs/ocfs2/cluster/tcp.h | 32 +++ fs/ocfs2/cluster/tcp_internal.h | 24 ++- 6 files changed, 569 insertions(+), 3 deletions(-) create mode 100644 fs/ocfs2/cluster/netdebug.c (limited to 'fs') diff --git a/fs/ocfs2/cluster/Makefile b/fs/ocfs2/cluster/Makefile index cdd162f1365..bc8c5e7d860 100644 --- a/fs/ocfs2/cluster/Makefile +++ b/fs/ocfs2/cluster/Makefile @@ -1,4 +1,4 @@ obj-$(CONFIG_OCFS2_FS) += ocfs2_nodemanager.o ocfs2_nodemanager-objs := heartbeat.o masklog.o sys.o nodemanager.o \ - quorum.o tcp.o ver.o + quorum.o tcp.o netdebug.o ver.o diff --git a/fs/ocfs2/cluster/netdebug.c b/fs/ocfs2/cluster/netdebug.c new file mode 100644 index 00000000000..7bf3c0ea7bd --- /dev/null +++ b/fs/ocfs2/cluster/netdebug.c @@ -0,0 +1,441 @@ +/* -*- mode: c; c-basic-offset: 8; -*- + * vim: noexpandtab sw=8 ts=8 sts=0: + * + * netdebug.c + * + * debug functionality for o2net + * + * Copyright (C) 2005, 2008 Oracle. All rights reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public + * License as published by the Free Software Foundation; either + * version 2 of the License, or (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * General Public License for more details. + * + * You should have received a copy of the GNU General Public + * License along with this program; if not, write to the + * Free Software Foundation, Inc., 59 Temple Place - Suite 330, + * Boston, MA 021110-1307, USA. + * + */ + +#ifdef CONFIG_DEBUG_FS + +#include +#include +#include +#include +#include +#include +#include + +#include + +#include "tcp.h" +#include "nodemanager.h" +#define MLOG_MASK_PREFIX ML_TCP +#include "masklog.h" + +#include "tcp_internal.h" + +#define O2NET_DEBUG_DIR "o2net" +#define SC_DEBUG_NAME "sock_containers" +#define NST_DEBUG_NAME "send_tracking" + +static struct dentry *o2net_dentry; +static struct dentry *sc_dentry; +static struct dentry *nst_dentry; + +static DEFINE_SPINLOCK(o2net_debug_lock); + +static LIST_HEAD(sock_containers); +static LIST_HEAD(send_tracking); + +void o2net_debug_add_nst(struct o2net_send_tracking *nst) +{ + spin_lock(&o2net_debug_lock); + list_add(&nst->st_net_debug_item, &send_tracking); + spin_unlock(&o2net_debug_lock); +} + +void o2net_debug_del_nst(struct o2net_send_tracking *nst) +{ + spin_lock(&o2net_debug_lock); + if (!list_empty(&nst->st_net_debug_item)) + list_del_init(&nst->st_net_debug_item); + spin_unlock(&o2net_debug_lock); +} + +static struct o2net_send_tracking + *next_nst(struct o2net_send_tracking *nst_start) +{ + struct o2net_send_tracking *nst, *ret = NULL; + + assert_spin_locked(&o2net_debug_lock); + + list_for_each_entry(nst, &nst_start->st_net_debug_item, + st_net_debug_item) { + /* discover the head of the list */ + if (&nst->st_net_debug_item == &send_tracking) + break; + + /* use st_task to detect real nsts in the list */ + if (nst->st_task != NULL) { + ret = nst; + break; + } + } + + return ret; +} + +static void *nst_seq_start(struct seq_file *seq, loff_t *pos) +{ + struct o2net_send_tracking *nst, *dummy_nst = seq->private; + + spin_lock(&o2net_debug_lock); + nst = next_nst(dummy_nst); + spin_unlock(&o2net_debug_lock); + + return nst; +} + +static void *nst_seq_next(struct seq_file *seq, void *v, loff_t *pos) +{ + struct o2net_send_tracking *nst, *dummy_nst = seq->private; + + spin_lock(&o2net_debug_lock); + nst = next_nst(dummy_nst); + list_del_init(&dummy_nst->st_net_debug_item); + if (nst) + list_add(&dummy_nst->st_net_debug_item, + &nst->st_net_debug_item); + spin_unlock(&o2net_debug_lock); + + return nst; /* unused, just needs to be null when done */ +} + +static int nst_seq_show(struct seq_file *seq, void *v) +{ + struct o2net_send_tracking *nst, *dummy_nst = seq->private; + + spin_lock(&o2net_debug_lock); + nst = next_nst(dummy_nst); + + if (nst != NULL) { + /* get_task_comm isn't exported. oh well. */ + seq_printf(seq, "%p:\n" + " pid: %lu\n" + " tgid: %lu\n" + " process name: %s\n" + " node: %u\n" + " sc: %p\n" + " message id: %d\n" + " message type: %u\n" + " message key: 0x%08x\n" + " sock acquiry: %lu.%lu\n" + " send start: %lu.%lu\n" + " wait start: %lu.%lu\n", + nst, (unsigned long)nst->st_task->pid, + (unsigned long)nst->st_task->tgid, + nst->st_task->comm, nst->st_node, + nst->st_sc, nst->st_id, nst->st_msg_type, + nst->st_msg_key, + nst->st_sock_time.tv_sec, nst->st_sock_time.tv_usec, + nst->st_send_time.tv_sec, nst->st_send_time.tv_usec, + nst->st_status_time.tv_sec, + nst->st_status_time.tv_usec); + } + + spin_unlock(&o2net_debug_lock); + + return 0; +} + +static void nst_seq_stop(struct seq_file *seq, void *v) +{ +} + +static struct seq_operations nst_seq_ops = { + .start = nst_seq_start, + .next = nst_seq_next, + .stop = nst_seq_stop, + .show = nst_seq_show, +}; + +static int nst_fop_open(struct inode *inode, struct file *file) +{ + struct o2net_send_tracking *dummy_nst; + struct seq_file *seq; + int ret; + + dummy_nst = kmalloc(sizeof(struct o2net_send_tracking), GFP_KERNEL); + if (dummy_nst == NULL) { + ret = -ENOMEM; + goto out; + } + dummy_nst->st_task = NULL; + + ret = seq_open(file, &nst_seq_ops); + if (ret) + goto out; + + seq = file->private_data; + seq->private = dummy_nst; + o2net_debug_add_nst(dummy_nst); + + dummy_nst = NULL; + +out: + kfree(dummy_nst); + return ret; +} + +static int nst_fop_release(struct inode *inode, struct file *file) +{ + struct seq_file *seq = file->private_data; + struct o2net_send_tracking *dummy_nst = seq->private; + + o2net_debug_del_nst(dummy_nst); + return seq_release_private(inode, file); +} + +static struct file_operations nst_seq_fops = { + .open = nst_fop_open, + .read = seq_read, + .llseek = seq_lseek, + .release = nst_fop_release, +}; + +void o2net_debug_add_sc(struct o2net_sock_container *sc) +{ + spin_lock(&o2net_debug_lock); + list_add(&sc->sc_net_debug_item, &sock_containers); + spin_unlock(&o2net_debug_lock); +} + +void o2net_debug_del_sc(struct o2net_sock_container *sc) +{ + spin_lock(&o2net_debug_lock); + list_del_init(&sc->sc_net_debug_item); + spin_unlock(&o2net_debug_lock); +} + +static struct o2net_sock_container + *next_sc(struct o2net_sock_container *sc_start) +{ + struct o2net_sock_container *sc, *ret = NULL; + + assert_spin_locked(&o2net_debug_lock); + + list_for_each_entry(sc, &sc_start->sc_net_debug_item, + sc_net_debug_item) { + /* discover the head of the list miscast as a sc */ + if (&sc->sc_net_debug_item == &sock_containers) + break; + + /* use sc_page to detect real scs in the list */ + if (sc->sc_page != NULL) { + ret = sc; + break; + } + } + + return ret; +} + +static void *sc_seq_start(struct seq_file *seq, loff_t *pos) +{ + struct o2net_sock_container *sc, *dummy_sc = seq->private; + + spin_lock(&o2net_debug_lock); + sc = next_sc(dummy_sc); + spin_unlock(&o2net_debug_lock); + + return sc; +} + +static void *sc_seq_next(struct seq_file *seq, void *v, loff_t *pos) +{ + struct o2net_sock_container *sc, *dummy_sc = seq->private; + + spin_lock(&o2net_debug_lock); + sc = next_sc(dummy_sc); + list_del_init(&dummy_sc->sc_net_debug_item); + if (sc) + list_add(&dummy_sc->sc_net_debug_item, &sc->sc_net_debug_item); + spin_unlock(&o2net_debug_lock); + + return sc; /* unused, just needs to be null when done */ +} + +#define TV_SEC_USEC(TV) TV.tv_sec, TV.tv_usec + +static int sc_seq_show(struct seq_file *seq, void *v) +{ + struct o2net_sock_container *sc, *dummy_sc = seq->private; + + spin_lock(&o2net_debug_lock); + sc = next_sc(dummy_sc); + + if (sc != NULL) { + struct inet_sock *inet = NULL; + + __be32 saddr = 0, daddr = 0; + __be16 sport = 0, dport = 0; + + if (sc->sc_sock) { + inet = inet_sk(sc->sc_sock->sk); + /* the stack's structs aren't sparse endian clean */ + saddr = (__force __be32)inet->saddr; + daddr = (__force __be32)inet->daddr; + sport = (__force __be16)inet->sport; + dport = (__force __be16)inet->dport; + } + + /* XXX sigh, inet-> doesn't have sparse annotation so any + * use of it here generates a warning with -Wbitwise */ + seq_printf(seq, "%p:\n" + " krefs: %d\n" + " sock: %u.%u.%u.%u:%u -> " + "%u.%u.%u.%u:%u\n" + " remote node: %s\n" + " page off: %zu\n" + " handshake ok: %u\n" + " timer: %lu.%lu\n" + " data ready: %lu.%lu\n" + " advance start: %lu.%lu\n" + " advance stop: %lu.%lu\n" + " func start: %lu.%lu\n" + " func stop: %lu.%lu\n" + " func key: %u\n" + " func type: %u\n", + sc, + atomic_read(&sc->sc_kref.refcount), + NIPQUAD(saddr), inet ? ntohs(sport) : 0, + NIPQUAD(daddr), inet ? ntohs(dport) : 0, + sc->sc_node->nd_name, + sc->sc_page_off, + sc->sc_handshake_ok, + TV_SEC_USEC(sc->sc_tv_timer), + TV_SEC_USEC(sc->sc_tv_data_ready), + TV_SEC_USEC(sc->sc_tv_advance_start), + TV_SEC_USEC(sc->sc_tv_advance_stop), + TV_SEC_USEC(sc->sc_tv_func_start), + TV_SEC_USEC(sc->sc_tv_func_stop), + sc->sc_msg_key, + sc->sc_msg_type); + } + + + spin_unlock(&o2net_debug_lock); + + return 0; +} + +static void sc_seq_stop(struct seq_file *seq, void *v) +{ +} + +static struct seq_operations sc_seq_ops = { + .start = sc_seq_start, + .next = sc_seq_next, + .stop = sc_seq_stop, + .show = sc_seq_show, +}; + +static int sc_fop_open(struct inode *inode, struct file *file) +{ + struct o2net_sock_container *dummy_sc; + struct seq_file *seq; + int ret; + + dummy_sc = kmalloc(sizeof(struct o2net_sock_container), GFP_KERNEL); + if (dummy_sc == NULL) { + ret = -ENOMEM; + goto out; + } + dummy_sc->sc_page = NULL; + + ret = seq_open(file, &sc_seq_ops); + if (ret) + goto out; + + seq = file->private_data; + seq->private = dummy_sc; + o2net_debug_add_sc(dummy_sc); + + dummy_sc = NULL; + +out: + kfree(dummy_sc); + return ret; +} + +static int sc_fop_release(struct inode *inode, struct file *file) +{ + struct seq_file *seq = file->private_data; + struct o2net_sock_container *dummy_sc = seq->private; + + o2net_debug_del_sc(dummy_sc); + return seq_release_private(inode, file); +} + +static struct file_operations sc_seq_fops = { + .open = sc_fop_open, + .read = seq_read, + .llseek = seq_lseek, + .release = sc_fop_release, +}; + +int o2net_debugfs_init(void) +{ + o2net_dentry = debugfs_create_dir(O2NET_DEBUG_DIR, NULL); + if (!o2net_dentry) { + mlog_errno(-ENOMEM); + goto bail; + } + + nst_dentry = debugfs_create_file(NST_DEBUG_NAME, S_IFREG|S_IRUSR, + o2net_dentry, NULL, + &nst_seq_fops); + if (!nst_dentry) { + mlog_errno(-ENOMEM); + goto bail; + } + + sc_dentry = debugfs_create_file(SC_DEBUG_NAME, S_IFREG|S_IRUSR, + o2net_dentry, NULL, + &sc_seq_fops); + if (!sc_dentry) { + mlog_errno(-ENOMEM); + goto bail; + } + + return 0; +bail: + if (sc_dentry) + debugfs_remove(sc_dentry); + if (nst_dentry) + debugfs_remove(nst_dentry); + if (o2net_dentry) + debugfs_remove(o2net_dentry); + return -ENOMEM; +} + +void o2net_debugfs_exit(void) +{ + if (sc_dentry) + debugfs_remove(sc_dentry); + if (nst_dentry) + debugfs_remove(nst_dentry); + if (o2net_dentry) + debugfs_remove(o2net_dentry); +} + +#endif /* CONFIG_DEBUG_FS */ diff --git a/fs/ocfs2/cluster/nodemanager.c b/fs/ocfs2/cluster/nodemanager.c index 709fba25bf7..cf9401e8cd0 100644 --- a/fs/ocfs2/cluster/nodemanager.c +++ b/fs/ocfs2/cluster/nodemanager.c @@ -959,7 +959,10 @@ static int __init init_o2nm(void) cluster_print_version(); o2hb_init(); - o2net_init(); + + ret = o2net_init(); + if (ret) + goto out; ocfs2_table_header = register_sysctl_table(ocfs2_root_table); if (!ocfs2_table_header) { diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c index 1170918a931..1e44ad14881 100644 --- a/fs/ocfs2/cluster/tcp.c +++ b/fs/ocfs2/cluster/tcp.c @@ -142,6 +142,54 @@ static void o2net_idle_timer(unsigned long data); static void o2net_sc_postpone_idle(struct o2net_sock_container *sc); static void o2net_sc_reset_idle_timer(struct o2net_sock_container *sc); +static void o2net_init_nst(struct o2net_send_tracking *nst, u32 msgtype, + u32 msgkey, struct task_struct *task, u8 node) +{ +#ifdef CONFIG_DEBUG_FS + INIT_LIST_HEAD(&nst->st_net_debug_item); + nst->st_task = task; + nst->st_msg_type = msgtype; + nst->st_msg_key = msgkey; + nst->st_node = node; +#endif +} + +static void o2net_set_nst_sock_time(struct o2net_send_tracking *nst) +{ +#ifdef CONFIG_DEBUG_FS + do_gettimeofday(&nst->st_sock_time); +#endif +} + +static void o2net_set_nst_send_time(struct o2net_send_tracking *nst) +{ +#ifdef CONFIG_DEBUG_FS + do_gettimeofday(&nst->st_send_time); +#endif +} + +static void o2net_set_nst_status_time(struct o2net_send_tracking *nst) +{ +#ifdef CONFIG_DEBUG_FS + do_gettimeofday(&nst->st_status_time); +#endif +} + +static void o2net_set_nst_sock_container(struct o2net_send_tracking *nst, + struct o2net_sock_container *sc) +{ +#ifdef CONFIG_DEBUG_FS + nst->st_sc = sc; +#endif +} + +static void o2net_set_nst_msg_id(struct o2net_send_tracking *nst, u32 msg_id) +{ +#ifdef CONFIG_DEBUG_FS + nst->st_id = msg_id; +#endif +} + static inline int o2net_reconnect_delay(void) { return o2nm_single_cluster->cl_reconnect_delay_ms; @@ -290,6 +338,7 @@ static void sc_kref_release(struct kref *kref) o2nm_node_put(sc->sc_node); sc->sc_node = NULL; + o2net_debug_del_sc(sc); kfree(sc); } @@ -330,6 +379,7 @@ static struct o2net_sock_container *sc_alloc(struct o2nm_node *node) ret = sc; sc->sc_page = page; + o2net_debug_add_sc(sc); sc = NULL; page = NULL; @@ -913,6 +963,9 @@ int o2net_send_message_vec(u32 msg_type, u32 key, struct kvec *caller_vec, struct o2net_status_wait nsw = { .ns_node_item = LIST_HEAD_INIT(nsw.ns_node_item), }; + struct o2net_send_tracking nst; + + o2net_init_nst(&nst, msg_type, key, current, target_node); if (o2net_wq == NULL) { mlog(0, "attempt to tx without o2netd running\n"); @@ -938,6 +991,10 @@ int o2net_send_message_vec(u32 msg_type, u32 key, struct kvec *caller_vec, goto out; } + o2net_debug_add_nst(&nst); + + o2net_set_nst_sock_time(&nst); + ret = wait_event_interruptible(nn->nn_sc_wq, o2net_tx_can_proceed(nn, &sc, &error)); if (!ret && error) @@ -945,6 +1002,8 @@ int o2net_send_message_vec(u32 msg_type, u32 key, struct kvec *caller_vec, if (ret) goto out; + o2net_set_nst_sock_container(&nst, sc); + veclen = caller_veclen + 1; vec = kmalloc(sizeof(struct kvec) * veclen, GFP_ATOMIC); if (vec == NULL) { @@ -971,6 +1030,9 @@ int o2net_send_message_vec(u32 msg_type, u32 key, struct kvec *caller_vec, goto out; msg->msg_num = cpu_to_be32(nsw.ns_id); + o2net_set_nst_msg_id(&nst, nsw.ns_id); + + o2net_set_nst_send_time(&nst); /* finally, convert the message header to network byte-order * and send */ @@ -985,6 +1047,7 @@ int o2net_send_message_vec(u32 msg_type, u32 key, struct kvec *caller_vec, } /* wait on other node's handler */ + o2net_set_nst_status_time(&nst); wait_event(nsw.ns_wq, o2net_nsw_completed(nn, &nsw)); /* Note that we avoid overwriting the callers status return @@ -997,6 +1060,7 @@ int o2net_send_message_vec(u32 msg_type, u32 key, struct kvec *caller_vec, mlog(0, "woken, returning system status %d, user status %d\n", ret, nsw.ns_status); out: + o2net_debug_del_nst(&nst); /* must be before dropping sc and node */ if (sc) sc_put(sc); if (vec) @@ -1935,6 +1999,9 @@ int o2net_init(void) o2quo_init(); + if (o2net_debugfs_init()) + return -ENOMEM; + o2net_hand = kzalloc(sizeof(struct o2net_handshake), GFP_KERNEL); o2net_keep_req = kzalloc(sizeof(struct o2net_msg), GFP_KERNEL); o2net_keep_resp = kzalloc(sizeof(struct o2net_msg), GFP_KERNEL); @@ -1976,4 +2043,5 @@ void o2net_exit(void) kfree(o2net_hand); kfree(o2net_keep_req); kfree(o2net_keep_resp); + o2net_debugfs_exit(); } diff --git a/fs/ocfs2/cluster/tcp.h b/fs/ocfs2/cluster/tcp.h index f36f66aab3d..a705d5d1903 100644 --- a/fs/ocfs2/cluster/tcp.h +++ b/fs/ocfs2/cluster/tcp.h @@ -117,4 +117,36 @@ int o2net_num_connected_peers(void); int o2net_init(void); void o2net_exit(void); +struct o2net_send_tracking; +struct o2net_sock_container; + +#ifdef CONFIG_DEBUG_FS +int o2net_debugfs_init(void); +void o2net_debugfs_exit(void); +void o2net_debug_add_nst(struct o2net_send_tracking *nst); +void o2net_debug_del_nst(struct o2net_send_tracking *nst); +void o2net_debug_add_sc(struct o2net_sock_container *sc); +void o2net_debug_del_sc(struct o2net_sock_container *sc); +#else +static int o2net_debugfs_init(void) +{ + return 0; +} +static void o2net_debugfs_exit(void) +{ +} +static void o2net_debug_add_nst(struct o2net_send_tracking *nst) +{ +} +static void o2net_debug_del_nst(struct o2net_send_tracking *nst) +{ +} +static void o2net_debug_add_sc(struct o2net_sock_container *sc) +{ +} +static void o2net_debug_del_sc(struct o2net_sock_container *sc) +{ +} +#endif /* CONFIG_DEBUG_FS */ + #endif /* O2CLUSTER_TCP_H */ diff --git a/fs/ocfs2/cluster/tcp_internal.h b/fs/ocfs2/cluster/tcp_internal.h index b4c5586f46e..8d58cfe410b 100644 --- a/fs/ocfs2/cluster/tcp_internal.h +++ b/fs/ocfs2/cluster/tcp_internal.h @@ -166,7 +166,9 @@ struct o2net_sock_container { /* original handlers for the sockets */ void (*sc_state_change)(struct sock *sk); void (*sc_data_ready)(struct sock *sk, int bytes); - +#ifdef CONFIG_DEBUG_FS + struct list_head sc_net_debug_item; +#endif struct timeval sc_tv_timer; struct timeval sc_tv_data_ready; struct timeval sc_tv_advance_start; @@ -208,4 +210,24 @@ struct o2net_status_wait { struct list_head ns_node_item; }; +#ifdef CONFIG_DEBUG_FS +/* just for state dumps */ +struct o2net_send_tracking { + struct list_head st_net_debug_item; + struct task_struct *st_task; + struct o2net_sock_container *st_sc; + u32 st_id; + u32 st_msg_type; + u32 st_msg_key; + u8 st_node; + struct timeval st_sock_time; + struct timeval st_send_time; + struct timeval st_status_time; +}; +#else +struct o2net_send_tracking { + u32 dummy; +}; +#endif /* CONFIG_DEBUG_FS */ + #endif /* O2CLUSTER_TCP_INTERNAL_H */ -- cgit v1.2.3 From 2302aca8508ee727a0c5edde3a7518a4ee03da1b Mon Sep 17 00:00:00 2001 From: Steve French Date: Fri, 18 Apr 2008 16:40:32 +0000 Subject: [CIFS] Reserve new proxy cap for WAFS New WAFS filer uses ioctls which are shown to be available on a share by querying this info level Acked-by: Sam Liddicott Signed-off-by: Stevef French --- fs/cifs/cifspdu.h | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/cifs/cifspdu.h b/fs/cifs/cifspdu.h index 64bd3ed6803..b18c6d43bef 100644 --- a/fs/cifs/cifspdu.h +++ b/fs/cifs/cifspdu.h @@ -1787,6 +1787,11 @@ typedef struct smb_com_transaction2_fnext_rsp_parms { #define SMB_QUERY_CIFS_UNIX_INFO 0x200 #define SMB_QUERY_POSIX_FS_INFO 0x201 #define SMB_QUERY_POSIX_WHO_AM_I 0x202 +#define SMB_QUERY_FS_PROXY 0x203 /* WAFS enabled. Returns structure + FILE_SYSTEM__UNIX_INFO to tell + whether new NTIOCTL available + (0xACE) for WAN friendly SMB + operations to be carried */ #define SMB_QUERY_LABEL_INFO 0x3ea #define SMB_QUERY_FS_QUOTA_INFO 0x3ee #define SMB_QUERY_FS_FULL_SIZE_INFO 0x3ef @@ -2043,7 +2048,8 @@ typedef struct { #define CIFS_UNIX_LARGE_READ_CAP 0x00000040 /* support reads >128K (up to 0xFFFF00 */ #define CIFS_UNIX_LARGE_WRITE_CAP 0x00000080 - +#define CIFS_UNIX_PROXY_CAP 0x00000100 /* Proxy cap: 0xACE ioctl and + QFS PROXY call */ #ifdef CONFIG_CIFS_POSIX /* Can not set pathnames cap yet until we send new posix create SMB since otherwise server can treat such handles opened with older ntcreatex -- cgit v1.2.3 From 076d8423a98659a92837b07aa494cb74bfefe77c Mon Sep 17 00:00:00 2001 From: Steve French Date: Fri, 18 Apr 2008 23:26:26 +0000 Subject: [CIFS] Fix UNC path prefix on QueryUnixPathInfo to have correct slash When a share was in DFS and the server was Unix/Linux, we were sending paths of the form \\server\share/dir/file rather than //server/share/dir/file There was some discussion between me and jra over whether we should use /server/share/dir/file as MS sometimes says - but the documentation for this claims it should be doubleslash for this type of UNC-like path format and that works, so leaving it as doubleslash but converting the \ to / in the the //server/share portion. This gets Samba to now correctly return STATUS_PATH_NOT_COVERED when it is supposed to (Windows already did since the direction of the slash was not an issue for them). Still need another minor change to fully enable DFS (need to finish some chages to SMBGetDFSRefer Signed-off-by: Steve French --- fs/cifs/inode.c | 15 +++++++++++---- 1 file changed, 11 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/cifs/inode.c b/fs/cifs/inode.c index bc673c8c1e6..e1031b9e2c5 100644 --- a/fs/cifs/inode.c +++ b/fs/cifs/inode.c @@ -161,12 +161,14 @@ static void cifs_unix_info_to_inode(struct inode *inode, spin_unlock(&inode->i_lock); } -static const unsigned char *cifs_get_search_path(struct cifsTconInfo *pTcon, - const char *search_path) +static const unsigned char *cifs_get_search_path(struct cifs_sb_info *cifs_sb, + const char *search_path) { int tree_len; int path_len; + int i; char *tmp_path; + struct cifsTconInfo *pTcon = cifs_sb->tcon; if (!(pTcon->Flags & SMB_SHARE_IS_IN_DFS)) return search_path; @@ -180,6 +182,11 @@ static const unsigned char *cifs_get_search_path(struct cifsTconInfo *pTcon, return search_path; strncpy(tmp_path, pTcon->treeName, tree_len); + if (cifs_sb->mnt_cifs_flags & CIFS_MOUNT_POSIX_PATHS) + for (i = 0; i < tree_len; i++) { + if (tmp_path[i] == '\\') + tmp_path[i] = '/'; + } strncpy(tmp_path+tree_len, search_path, path_len); tmp_path[tree_len+path_len] = 0; return tmp_path; @@ -199,7 +206,7 @@ int cifs_get_inode_info_unix(struct inode **pinode, pTcon = cifs_sb->tcon; cFYI(1, ("Getting info on %s", search_path)); - full_path = cifs_get_search_path(pTcon, search_path); + full_path = cifs_get_search_path(cifs_sb, search_path); try_again_CIFSSMBUnixQPathInfo: /* could have done a find first instead but this returns more info */ @@ -402,7 +409,7 @@ int cifs_get_inode_info(struct inode **pinode, return -ENOMEM; pfindData = (FILE_ALL_INFO *)buf; - full_path = cifs_get_search_path(pTcon, search_path); + full_path = cifs_get_search_path(cifs_sb, search_path); try_again_CIFSSMBQPathInfo: /* could do find first instead but this returns more info */ -- cgit v1.2.3 From cb688371e27880d86c42323826846d1cd7caad8f Mon Sep 17 00:00:00 2001 From: Matthew Wilcox Date: Tue, 26 Feb 2008 09:59:26 -0500 Subject: fs: Remove unnecessary inclusions of asm/semaphore.h None of these files use any of the functionality promised by asm/semaphore.h. Signed-off-by: Matthew Wilcox --- fs/9p/fid.c | 1 - fs/cramfs/inode.c | 1 - fs/dlm/dlm_internal.h | 1 - fs/locks.c | 1 - fs/reiserfs/xattr.c | 1 - 5 files changed, 5 deletions(-) (limited to 'fs') diff --git a/fs/9p/fid.c b/fs/9p/fid.c index dfebdbe7440..3031e3233dd 100644 --- a/fs/9p/fid.c +++ b/fs/9p/fid.c @@ -26,7 +26,6 @@ #include #include #include -#include #include #include diff --git a/fs/cramfs/inode.c b/fs/cramfs/inode.c index 350680fd7da..0c3b618c15b 100644 --- a/fs/cramfs/inode.c +++ b/fs/cramfs/inode.c @@ -23,7 +23,6 @@ #include #include #include -#include #include diff --git a/fs/dlm/dlm_internal.h b/fs/dlm/dlm_internal.h index d30ea8b433a..7a8824f475f 100644 --- a/fs/dlm/dlm_internal.h +++ b/fs/dlm/dlm_internal.h @@ -37,7 +37,6 @@ #include #include #include -#include #include #include diff --git a/fs/locks.c b/fs/locks.c index 43c0af21a0c..592faadbcec 100644 --- a/fs/locks.c +++ b/fs/locks.c @@ -127,7 +127,6 @@ #include #include -#include #include #define IS_POSIX(fl) (fl->fl_flags & FL_POSIX) diff --git a/fs/reiserfs/xattr.c b/fs/reiserfs/xattr.c index 344b9b96cc5..d7c4935c103 100644 --- a/fs/reiserfs/xattr.c +++ b/fs/reiserfs/xattr.c @@ -44,7 +44,6 @@ #include #include #include -#include #define FL_READONLY 128 #define FL_DIR_SEM_HELD 256 -- cgit v1.2.3 From 6188e10d38b8d7244ee7776d5f1f88c837b4b93f Mon Sep 17 00:00:00 2001 From: Matthew Wilcox Date: Fri, 18 Apr 2008 22:21:05 -0400 Subject: Convert asm/semaphore.h users to linux/semaphore.h Signed-off-by: Matthew Wilcox --- fs/jffs2/jffs2_fs_i.h | 2 +- fs/jffs2/jffs2_fs_sb.h | 2 +- fs/reiserfs/journal.c | 2 +- fs/xfs/linux-2.6/sema.h | 2 +- 4 files changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/jffs2/jffs2_fs_i.h b/fs/jffs2/jffs2_fs_i.h index 0b78fdc9773..a841f4973a7 100644 --- a/fs/jffs2/jffs2_fs_i.h +++ b/fs/jffs2/jffs2_fs_i.h @@ -15,7 +15,7 @@ #include #include #include -#include +#include struct jffs2_inode_info { /* We need an internal mutex similar to inode->i_mutex. diff --git a/fs/jffs2/jffs2_fs_sb.h b/fs/jffs2/jffs2_fs_sb.h index 3a2197f3c81..18fca2b9e53 100644 --- a/fs/jffs2/jffs2_fs_sb.h +++ b/fs/jffs2/jffs2_fs_sb.h @@ -16,7 +16,7 @@ #include #include #include -#include +#include #include #include #include diff --git a/fs/reiserfs/journal.c b/fs/reiserfs/journal.c index bb05a3e51b9..060eb3f598e 100644 --- a/fs/reiserfs/journal.c +++ b/fs/reiserfs/journal.c @@ -38,7 +38,7 @@ #include #include -#include +#include #include #include diff --git a/fs/xfs/linux-2.6/sema.h b/fs/xfs/linux-2.6/sema.h index 2009e6d922c..3abe7e9ceb3 100644 --- a/fs/xfs/linux-2.6/sema.h +++ b/fs/xfs/linux-2.6/sema.h @@ -20,8 +20,8 @@ #include #include +#include #include -#include /* * sema_t structure just maps to struct semaphore in Linux kernel. -- cgit v1.2.3 From d57999e1527f0b0c818846dcba5a23015beb4823 Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:27 -0800 Subject: [PATCH] do namei_flags calculation inside open_namei() My end goal here is to make sure all users of may_open() return filps. This will ensure that we properly release mount write counts which were taken for the filp in may_open(). This patch moves the sys_open flags to namei flags calculation into fs/namei.c. We'll shortly be moving the nameidata_to_filp() calls into namei.c, and this gets the sys_open flags to a place where we can get at them when we need them. Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Dave Hansen Signed-off-by: Al Viro --- fs/namei.c | 43 ++++++++++++++++++++++++++++++++++--------- fs/open.c | 22 ++-------------------- 2 files changed, 36 insertions(+), 29 deletions(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index 8cf9bb9c2fc..c70dbf72010 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -1677,7 +1677,12 @@ int may_open(struct nameidata *nd, int acc_mode, int flag) return 0; } -static int open_namei_create(struct nameidata *nd, struct path *path, +/* + * Be careful about ever adding any more callers of this + * function. Its flags must be in the namei format, not + * what get passed to sys_open(). + */ +static int __open_namei_create(struct nameidata *nd, struct path *path, int flag, int mode) { int error; @@ -1695,27 +1700,47 @@ static int open_namei_create(struct nameidata *nd, struct path *path, return may_open(nd, 0, flag & ~O_TRUNC); } +/* + * Note that while the flag value (low two bits) for sys_open means: + * 00 - read-only + * 01 - write-only + * 10 - read-write + * 11 - special + * it is changed into + * 00 - no permissions needed + * 01 - read-permission + * 10 - write-permission + * 11 - read-write + * for the internal routines (ie open_namei()/follow_link() etc) + * This is more logical, and also allows the 00 "no perm needed" + * to be used for symlinks (where the permissions are checked + * later). + * +*/ +static inline int open_to_namei_flags(int flag) +{ + if ((flag+1) & O_ACCMODE) + flag++; + return flag; +} + /* * open_namei() * * namei for open - this is in fact almost the whole open-routine. * * Note that the low bits of "flag" aren't the same as in the open - * system call - they are 00 - no permissions needed - * 01 - read permission needed - * 10 - write permission needed - * 11 - read/write permissions needed - * which is a lot more logical, and also allows the "no perm" needed - * for symlinks (where the permissions are checked later). + * system call. See open_to_namei_flags(). * SMP-safe */ -int open_namei(int dfd, const char *pathname, int flag, +int open_namei(int dfd, const char *pathname, int open_flag, int mode, struct nameidata *nd) { int acc_mode, error; struct path path; struct dentry *dir; int count = 0; + int flag = open_to_namei_flags(open_flag); acc_mode = ACC_MODE(flag); @@ -1776,7 +1801,7 @@ do_last: /* Negative dentry, just create the file */ if (!path.dentry->d_inode) { - error = open_namei_create(nd, &path, flag, mode); + error = __open_namei_create(nd, &path, flag, mode); if (error) goto exit; return 0; diff --git a/fs/open.c b/fs/open.c index 3fa4e4ffce4..5ab3f3f079c 100644 --- a/fs/open.c +++ b/fs/open.c @@ -796,31 +796,13 @@ cleanup_file: return ERR_PTR(error); } -/* - * Note that while the flag value (low two bits) for sys_open means: - * 00 - read-only - * 01 - write-only - * 10 - read-write - * 11 - special - * it is changed into - * 00 - no permissions needed - * 01 - read-permission - * 10 - write-permission - * 11 - read-write - * for the internal routines (ie open_namei()/follow_link() etc). 00 is - * used by symlinks. - */ static struct file *do_filp_open(int dfd, const char *filename, int flags, int mode) { - int namei_flags, error; + int error; struct nameidata nd; - namei_flags = flags; - if ((namei_flags+1) & O_ACCMODE) - namei_flags++; - - error = open_namei(dfd, filename, namei_flags, mode, &nd); + error = open_namei(dfd, filename, flags, mode, &nd); if (!error) return nameidata_to_filp(&nd, flags); -- cgit v1.2.3 From a70e65df8812c52252fa07a2eb92a46451a4427f Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 15 Feb 2008 14:37:28 -0800 Subject: [PATCH] merge open_namei() and do_filp_open() open_namei() will, in the future, need to take mount write counts over its creation and truncation (via may_open()) operations. It needs to keep these write counts until any potential filp that is created gets __fput()'d. This gets complicated in the error handling and becomes very murky as to how far open_namei() actually got, and whether or not that mount write count was taken. That makes it a bad interface. All that the current do_filp_open() really does is allocate the nameidata on the stack, then call open_namei(). So, this merges those two functions and moves filp_open() over to namei.c so it can be close to its buddy: do_filp_open(). It also gets a kerneldoc comment in the process. Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Andrew Morton Signed-off-by: Dave Hansen Signed-off-by: Al Viro --- fs/namei.c | 100 +++++++++++++++++++++++++++++++++++-------------------------- fs/open.c | 19 ------------ 2 files changed, 57 insertions(+), 62 deletions(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index c70dbf72010..a1f8bbbd58e 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -1725,17 +1725,13 @@ static inline int open_to_namei_flags(int flag) } /* - * open_namei() - * - * namei for open - this is in fact almost the whole open-routine. - * * Note that the low bits of "flag" aren't the same as in the open * system call. See open_to_namei_flags(). - * SMP-safe */ -int open_namei(int dfd, const char *pathname, int open_flag, - int mode, struct nameidata *nd) +struct file *do_filp_open(int dfd, const char *pathname, + int open_flag, int mode) { + struct nameidata nd; int acc_mode, error; struct path path; struct dentry *dir; @@ -1758,18 +1754,19 @@ int open_namei(int dfd, const char *pathname, int open_flag, */ if (!(flag & O_CREAT)) { error = path_lookup_open(dfd, pathname, lookup_flags(flag), - nd, flag); + &nd, flag); if (error) - return error; + return ERR_PTR(error); goto ok; } /* * Create - we need to know the parent. */ - error = path_lookup_create(dfd,pathname,LOOKUP_PARENT,nd,flag,mode); + error = path_lookup_create(dfd, pathname, LOOKUP_PARENT, + &nd, flag, mode); if (error) - return error; + return ERR_PTR(error); /* * We have the parent and last component. First of all, check @@ -1777,14 +1774,14 @@ int open_namei(int dfd, const char *pathname, int open_flag, * will not do. */ error = -EISDIR; - if (nd->last_type != LAST_NORM || nd->last.name[nd->last.len]) + if (nd.last_type != LAST_NORM || nd.last.name[nd.last.len]) goto exit; - dir = nd->path.dentry; - nd->flags &= ~LOOKUP_PARENT; + dir = nd.path.dentry; + nd.flags &= ~LOOKUP_PARENT; mutex_lock(&dir->d_inode->i_mutex); - path.dentry = lookup_hash(nd); - path.mnt = nd->path.mnt; + path.dentry = lookup_hash(&nd); + path.mnt = nd.path.mnt; do_last: error = PTR_ERR(path.dentry); @@ -1793,18 +1790,18 @@ do_last: goto exit; } - if (IS_ERR(nd->intent.open.file)) { + if (IS_ERR(nd.intent.open.file)) { mutex_unlock(&dir->d_inode->i_mutex); - error = PTR_ERR(nd->intent.open.file); + error = PTR_ERR(nd.intent.open.file); goto exit_dput; } /* Negative dentry, just create the file */ if (!path.dentry->d_inode) { - error = __open_namei_create(nd, &path, flag, mode); + error = __open_namei_create(&nd, &path, flag, mode); if (error) goto exit; - return 0; + return nameidata_to_filp(&nd, open_flag); } /* @@ -1829,23 +1826,23 @@ do_last: if (path.dentry->d_inode->i_op && path.dentry->d_inode->i_op->follow_link) goto do_link; - path_to_nameidata(&path, nd); + path_to_nameidata(&path, &nd); error = -EISDIR; if (path.dentry->d_inode && S_ISDIR(path.dentry->d_inode->i_mode)) goto exit; ok: - error = may_open(nd, acc_mode, flag); + error = may_open(&nd, acc_mode, flag); if (error) goto exit; - return 0; + return nameidata_to_filp(&nd, open_flag); exit_dput: - path_put_conditional(&path, nd); + path_put_conditional(&path, &nd); exit: - if (!IS_ERR(nd->intent.open.file)) - release_open_intent(nd); - path_put(&nd->path); - return error; + if (!IS_ERR(nd.intent.open.file)) + release_open_intent(&nd); + path_put(&nd.path); + return ERR_PTR(error); do_link: error = -ELOOP; @@ -1861,42 +1858,59 @@ do_link: * stored in nd->last.name and we will have to putname() it when we * are done. Procfs-like symlinks just set LAST_BIND. */ - nd->flags |= LOOKUP_PARENT; - error = security_inode_follow_link(path.dentry, nd); + nd.flags |= LOOKUP_PARENT; + error = security_inode_follow_link(path.dentry, &nd); if (error) goto exit_dput; - error = __do_follow_link(&path, nd); + error = __do_follow_link(&path, &nd); if (error) { /* Does someone understand code flow here? Or it is only * me so stupid? Anathema to whoever designed this non-sense * with "intent.open". */ - release_open_intent(nd); - return error; + release_open_intent(&nd); + return ERR_PTR(error); } - nd->flags &= ~LOOKUP_PARENT; - if (nd->last_type == LAST_BIND) + nd.flags &= ~LOOKUP_PARENT; + if (nd.last_type == LAST_BIND) goto ok; error = -EISDIR; - if (nd->last_type != LAST_NORM) + if (nd.last_type != LAST_NORM) goto exit; - if (nd->last.name[nd->last.len]) { - __putname(nd->last.name); + if (nd.last.name[nd.last.len]) { + __putname(nd.last.name); goto exit; } error = -ELOOP; if (count++==32) { - __putname(nd->last.name); + __putname(nd.last.name); goto exit; } - dir = nd->path.dentry; + dir = nd.path.dentry; mutex_lock(&dir->d_inode->i_mutex); - path.dentry = lookup_hash(nd); - path.mnt = nd->path.mnt; - __putname(nd->last.name); + path.dentry = lookup_hash(&nd); + path.mnt = nd.path.mnt; + __putname(nd.last.name); goto do_last; } +/** + * filp_open - open file and return file pointer + * + * @filename: path to open + * @flags: open flags as per the open(2) second argument + * @mode: mode for the new file if O_CREAT is set, else ignored + * + * This is the helper to open a file from kernelspace if you really + * have to. But in generally you should not do this, so please move + * along, nothing to see here.. + */ +struct file *filp_open(const char *filename, int flags, int mode) +{ + return do_filp_open(AT_FDCWD, filename, flags, mode); +} +EXPORT_SYMBOL(filp_open); + /** * lookup_create - lookup a dentry, creating it if it doesn't exist * @nd: nameidata info diff --git a/fs/open.c b/fs/open.c index 5ab3f3f079c..8111947905d 100644 --- a/fs/open.c +++ b/fs/open.c @@ -796,25 +796,6 @@ cleanup_file: return ERR_PTR(error); } -static struct file *do_filp_open(int dfd, const char *filename, int flags, - int mode) -{ - int error; - struct nameidata nd; - - error = open_namei(dfd, filename, flags, mode, &nd); - if (!error) - return nameidata_to_filp(&nd, flags); - - return ERR_PTR(error); -} - -struct file *filp_open(const char *filename, int flags, int mode) -{ - return do_filp_open(AT_FDCWD, filename, flags, mode); -} -EXPORT_SYMBOL(filp_open); - /** * lookup_instantiate_filp - instantiates the open intent filp * @nd: pointer to nameidata -- cgit v1.2.3 From 8366025eb80dfa0d8d94b286d53027081c280ef1 Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:30 -0800 Subject: [PATCH] r/o bind mounts: stub functions This patch adds two function mnt_want_write() and mnt_drop_write(). These are used like a lock pair around and fs operations that might cause a write to the filesystem. Before these can become useful, we must first cover each place in the VFS where writes are performed with a want/drop pair. When that is complete, we can actually introduce code that will safely check the counts before allowing r/w<->r/o transitions to occur. Acked-by: Serge Hallyn Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Andrew Morton Signed-off-by: Dave Hansen Signed-off-by: Al Viro --- fs/namespace.c | 54 ++++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 54 insertions(+) (limited to 'fs') diff --git a/fs/namespace.c b/fs/namespace.c index 94f026ec990..066b393578c 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -80,6 +80,60 @@ struct vfsmount *alloc_vfsmnt(const char *name) return mnt; } +/* + * Most r/o checks on a fs are for operations that take + * discrete amounts of time, like a write() or unlink(). + * We must keep track of when those operations start + * (for permission checks) and when they end, so that + * we can determine when writes are able to occur to + * a filesystem. + */ +/** + * mnt_want_write - get write access to a mount + * @mnt: the mount on which to take a write + * + * This tells the low-level filesystem that a write is + * about to be performed to it, and makes sure that + * writes are allowed before returning success. When + * the write operation is finished, mnt_drop_write() + * must be called. This is effectively a refcount. + */ +int mnt_want_write(struct vfsmount *mnt) +{ + if (__mnt_is_readonly(mnt)) + return -EROFS; + return 0; +} +EXPORT_SYMBOL_GPL(mnt_want_write); + +/** + * mnt_drop_write - give up write access to a mount + * @mnt: the mount on which to give up write access + * + * Tells the low-level filesystem that we are done + * performing writes to it. Must be matched with + * mnt_want_write() call above. + */ +void mnt_drop_write(struct vfsmount *mnt) +{ +} +EXPORT_SYMBOL_GPL(mnt_drop_write); + +/* + * __mnt_is_readonly: check whether a mount is read-only + * @mnt: the mount to check for its write status + * + * This shouldn't be used directly ouside of the VFS. + * It does not guarantee that the filesystem will stay + * r/w, just that it is right *now*. This can not and + * should not be used in place of IS_RDONLY(inode). + */ +int __mnt_is_readonly(struct vfsmount *mnt) +{ + return (mnt->mnt_sb->s_flags & MS_RDONLY); +} +EXPORT_SYMBOL_GPL(__mnt_is_readonly); + int simple_set_mnt(struct vfsmount *mnt, struct super_block *sb) { mnt->mnt_sb = sb; -- cgit v1.2.3 From aceaf78da92a53f5e1b105649a1b8c0afdb2135c Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:31 -0800 Subject: [PATCH] r/o bind mounts: create helper to drop file write access If someone decides to demote a file from r/w to just r/o, they can use this same code as __fput(). NFS does just that, and will use this in the next patch. AV: drop write access in __fput() only after we evict from file list. Signed-off-by: Dave Hansen Cc: Erez Zadok Cc: Trond Myklebust Cc: "J Bruce Fields" Acked-by: Al Viro Signed-off-by: Andrew Morton Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/file_table.c | 21 +++++++++++++++++++-- fs/nfsd/nfs4state.c | 3 ++- 2 files changed, 21 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/file_table.c b/fs/file_table.c index 986ff4ed0a7..3f73eb1f195 100644 --- a/fs/file_table.c +++ b/fs/file_table.c @@ -211,6 +211,23 @@ void fput(struct file *file) EXPORT_SYMBOL(fput); +/** + * drop_file_write_access - give up ability to write to a file + * @file: the file to which we will stop writing + * + * This is a central place which will give up the ability + * to write to @file, along with access to write through + * its vfsmount. + */ +void drop_file_write_access(struct file *file) +{ + struct dentry *dentry = file->f_path.dentry; + struct inode *inode = dentry->d_inode; + + put_write_access(inode); +} +EXPORT_SYMBOL_GPL(drop_file_write_access); + /* __fput is called from task context when aio completion releases the last * last use of a struct file *. Do not use otherwise. */ @@ -236,10 +253,10 @@ void __fput(struct file *file) if (unlikely(S_ISCHR(inode->i_mode) && inode->i_cdev != NULL)) cdev_put(inode->i_cdev); fops_put(file->f_op); - if (file->f_mode & FMODE_WRITE) - put_write_access(inode); put_pid(file->f_owner.pid); file_kill(file); + if (file->f_mode & FMODE_WRITE) + drop_file_write_access(file); file->f_path.dentry = NULL; file->f_path.mnt = NULL; file_free(file); diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index bcb97d8e8b8..81a75f3081f 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -41,6 +41,7 @@ #include #include #include +#include #include #include #include @@ -1239,7 +1240,7 @@ static inline void nfs4_file_downgrade(struct file *filp, unsigned int share_access) { if (share_access & NFS4_SHARE_ACCESS_WRITE) { - put_write_access(filp->f_path.dentry->d_inode); + drop_file_write_access(filp); filp->f_mode = (filp->f_mode | FMODE_READ) & ~FMODE_WRITE; } } -- cgit v1.2.3 From 49e0d02cf018d4edf24bfc8531a816a26367e4ce Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:32 -0800 Subject: [PATCH] r/o bind mounts: drop write during emergency remount The emergency remount code forcibly removes FMODE_WRITE from filps. The r/o bind mount code notices that this was done without a proper mnt_drop_write() and properly gives a warning. This patch does a mnt_drop_write() to keep everything balanced. Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Dave Hansen Signed-off-by: Al Viro --- fs/super.c | 21 +++++++++++++++++++-- 1 file changed, 19 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/super.c b/fs/super.c index 09008dbd264..01d5c40e911 100644 --- a/fs/super.c +++ b/fs/super.c @@ -37,6 +37,7 @@ #include #include #include +#include #include @@ -567,10 +568,26 @@ static void mark_files_ro(struct super_block *sb) { struct file *f; +retry: file_list_lock(); list_for_each_entry(f, &sb->s_files, f_u.fu_list) { - if (S_ISREG(f->f_path.dentry->d_inode->i_mode) && file_count(f)) - f->f_mode &= ~FMODE_WRITE; + struct vfsmount *mnt; + if (!S_ISREG(f->f_path.dentry->d_inode->i_mode)) + continue; + if (!file_count(f)) + continue; + if (!(f->f_mode & FMODE_WRITE)) + continue; + f->f_mode &= ~FMODE_WRITE; + mnt = mntget(f->f_path.mnt); + file_list_unlock(); + /* + * This can sleep, so we can't hold + * the file_list_lock() spinlock. + */ + mnt_drop_write(mnt); + mntput(mnt); + goto retry; } file_list_unlock(); } -- cgit v1.2.3 From 0622753b800e4cc6cb9319b36b27658c72dd7cdc Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:34 -0800 Subject: [PATCH] r/o bind mounts: elevate write count for rmdir and unlink. Elevate the write count during the vfs_rmdir() and vfs_unlink(). [AV: merged rmdir and unlink parts, added missing pieces in nfsd] Acked-by: Serge Hallyn Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Dave Hansen Signed-off-by: Andrew Morton Signed-off-by: Al Viro --- fs/namei.c | 9 +++++++++ fs/nfsd/nfs4recover.c | 12 +++++++++++- fs/nfsd/vfs.c | 8 +++++++- 3 files changed, 27 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index a1f8bbbd58e..89ef3178eaa 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -2190,7 +2190,12 @@ static long do_rmdir(int dfd, const char __user *pathname) error = PTR_ERR(dentry); if (IS_ERR(dentry)) goto exit2; + error = mnt_want_write(nd.path.mnt); + if (error) + goto exit3; error = vfs_rmdir(nd.path.dentry->d_inode, dentry); + mnt_drop_write(nd.path.mnt); +exit3: dput(dentry); exit2: mutex_unlock(&nd.path.dentry->d_inode->i_mutex); @@ -2271,7 +2276,11 @@ static long do_unlinkat(int dfd, const char __user *pathname) inode = dentry->d_inode; if (inode) atomic_inc(&inode->i_count); + error = mnt_want_write(nd.path.mnt); + if (error) + goto exit2; error = vfs_unlink(nd.path.dentry->d_inode, dentry); + mnt_drop_write(nd.path.mnt); exit2: dput(dentry); } diff --git a/fs/nfsd/nfs4recover.c b/fs/nfsd/nfs4recover.c index 1ff90625860..4e77a1a3bd7 100644 --- a/fs/nfsd/nfs4recover.c +++ b/fs/nfsd/nfs4recover.c @@ -46,6 +46,7 @@ #include #include #include +#include #define NFSDDBG_FACILITY NFSDDBG_PROC @@ -313,12 +314,17 @@ nfsd4_remove_clid_dir(struct nfs4_client *clp) if (!rec_dir_init || !clp->cl_firststate) return; + status = mnt_want_write(rec_dir.path.mnt); + if (status) + goto out; clp->cl_firststate = 0; nfs4_save_user(&uid, &gid); status = nfsd4_unlink_clid_dir(clp->cl_recdir, HEXDIR_LEN-1); nfs4_reset_user(uid, gid); if (status == 0) nfsd4_sync_rec_dir(); + mnt_drop_write(rec_dir.path.mnt); +out: if (status) printk("NFSD: Failed to remove expired client state directory" " %.*s\n", HEXDIR_LEN, clp->cl_recdir); @@ -347,13 +353,17 @@ nfsd4_recdir_purge_old(void) { if (!rec_dir_init) return; + status = mnt_want_write(rec_dir.path.mnt); + if (status) + goto out; status = nfsd4_list_rec_dir(rec_dir.path.dentry, purge_old); if (status == 0) nfsd4_sync_rec_dir(); + mnt_drop_write(rec_dir.path.mnt); +out: if (status) printk("nfsd4: failed to purge old clients from recovery" " directory %s\n", rec_dir.path.dentry->d_name.name); - return; } static int diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index 46f59d5365a..efff58a5818 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1750,6 +1750,10 @@ nfsd_unlink(struct svc_rqst *rqstp, struct svc_fh *fhp, int type, if (!type) type = rdentry->d_inode->i_mode & S_IFMT; + host_err = mnt_want_write(fhp->fh_export->ex_path.mnt); + if (host_err) + goto out_nfserr; + if (type != S_IFDIR) { /* It's UNLINK */ #ifdef MSNFS if ((fhp->fh_export->ex_flags & NFSEXP_MSNFS) && @@ -1765,10 +1769,12 @@ nfsd_unlink(struct svc_rqst *rqstp, struct svc_fh *fhp, int type, dput(rdentry); if (host_err) - goto out_nfserr; + goto out_drop; if (EX_ISSYNC(fhp->fh_export)) host_err = nfsd_sync_dir(dentry); +out_drop: + mnt_drop_write(fhp->fh_export->ex_path.mnt); out_nfserr: err = nfserrno(host_err); out: -- cgit v1.2.3 From 463c3197263bd26ac59a00d2484990e17e35c50e Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:57 -0800 Subject: [PATCH] r/o bind mounts: get callers of vfs_mknod/create/mkdir() This takes care of all of the direct callers of vfs_mknod(). Since a few of these cases also handle normal file creation as well, this also covers some calls to vfs_create(). So that we don't have to make three mnt_want/drop_write() calls inside of the switch statement, we move some of its logic outside of the switch and into a helper function suggested by Christoph. This also encapsulates a fix for mknod(S_IFREG) that Miklos found. [AV: merged mkdir handling, added missing nfsd pieces] Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Dave Hansen Signed-off-by: Andrew Morton Signed-off-by: Al Viro --- fs/namei.c | 48 +++++++++++++++++++++++++++++++++++++----------- fs/nfsd/nfs4recover.c | 4 ++++ fs/nfsd/vfs.c | 24 ++++++++++++++++++++++-- 3 files changed, 63 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index 89ef3178eaa..3fbcf2021a2 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -1984,6 +1984,23 @@ int vfs_mknod(struct inode *dir, struct dentry *dentry, int mode, dev_t dev) return error; } +static int may_mknod(mode_t mode) +{ + switch (mode & S_IFMT) { + case S_IFREG: + case S_IFCHR: + case S_IFBLK: + case S_IFIFO: + case S_IFSOCK: + case 0: /* zero mode translates to S_IFREG */ + return 0; + case S_IFDIR: + return -EPERM; + default: + return -EINVAL; + } +} + asmlinkage long sys_mknodat(int dfd, const char __user *filename, int mode, unsigned dev) { @@ -2002,12 +2019,19 @@ asmlinkage long sys_mknodat(int dfd, const char __user *filename, int mode, if (error) goto out; dentry = lookup_create(&nd, 0); - error = PTR_ERR(dentry); - + if (IS_ERR(dentry)) { + error = PTR_ERR(dentry); + goto out_unlock; + } if (!IS_POSIXACL(nd.path.dentry->d_inode)) mode &= ~current->fs->umask; - if (!IS_ERR(dentry)) { - switch (mode & S_IFMT) { + error = may_mknod(mode); + if (error) + goto out_dput; + error = mnt_want_write(nd.path.mnt); + if (error) + goto out_dput; + switch (mode & S_IFMT) { case 0: case S_IFREG: error = vfs_create(nd.path.dentry->d_inode,dentry,mode,&nd); break; @@ -2018,14 +2042,11 @@ asmlinkage long sys_mknodat(int dfd, const char __user *filename, int mode, case S_IFIFO: case S_IFSOCK: error = vfs_mknod(nd.path.dentry->d_inode,dentry,mode,0); break; - case S_IFDIR: - error = -EPERM; - break; - default: - error = -EINVAL; - } - dput(dentry); } + mnt_drop_write(nd.path.mnt); +out_dput: + dput(dentry); +out_unlock: mutex_unlock(&nd.path.dentry->d_inode->i_mutex); path_put(&nd.path); out: @@ -2083,7 +2104,12 @@ asmlinkage long sys_mkdirat(int dfd, const char __user *pathname, int mode) if (!IS_POSIXACL(nd.path.dentry->d_inode)) mode &= ~current->fs->umask; + error = mnt_want_write(nd.path.mnt); + if (error) + goto out_dput; error = vfs_mkdir(nd.path.dentry->d_inode, dentry, mode); + mnt_drop_write(nd.path.mnt); +out_dput: dput(dentry); out_unlock: mutex_unlock(&nd.path.dentry->d_inode->i_mutex); diff --git a/fs/nfsd/nfs4recover.c b/fs/nfsd/nfs4recover.c index 4e77a1a3bd7..145b3c877a2 100644 --- a/fs/nfsd/nfs4recover.c +++ b/fs/nfsd/nfs4recover.c @@ -155,7 +155,11 @@ nfsd4_create_clid_dir(struct nfs4_client *clp) dprintk("NFSD: nfsd4_create_clid_dir: DIRECTORY EXISTS\n"); goto out_put; } + status = mnt_want_write(rec_dir.path.mnt); + if (status) + goto out_put; status = vfs_mkdir(rec_dir.path.dentry->d_inode, dentry, S_IRWXU); + mnt_drop_write(rec_dir.path.mnt); out_put: dput(dentry); out_unlock: diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index efff58a5818..14d58246702 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1255,23 +1255,35 @@ nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp, err = 0; switch (type) { case S_IFREG: + host_err = mnt_want_write(fhp->fh_export->ex_path.mnt); + if (host_err) + goto out_nfserr; host_err = vfs_create(dirp, dchild, iap->ia_mode, NULL); break; case S_IFDIR: + host_err = mnt_want_write(fhp->fh_export->ex_path.mnt); + if (host_err) + goto out_nfserr; host_err = vfs_mkdir(dirp, dchild, iap->ia_mode); break; case S_IFCHR: case S_IFBLK: case S_IFIFO: case S_IFSOCK: + host_err = mnt_want_write(fhp->fh_export->ex_path.mnt); + if (host_err) + goto out_nfserr; host_err = vfs_mknod(dirp, dchild, iap->ia_mode, rdev); break; default: printk("nfsd: bad file type %o in nfsd_create\n", type); host_err = -EINVAL; + goto out_nfserr; } - if (host_err < 0) + if (host_err < 0) { + mnt_drop_write(fhp->fh_export->ex_path.mnt); goto out_nfserr; + } if (EX_ISSYNC(fhp->fh_export)) { err = nfserrno(nfsd_sync_dir(dentry)); @@ -1282,6 +1294,7 @@ nfsd_create(struct svc_rqst *rqstp, struct svc_fh *fhp, err2 = nfsd_create_setattr(rqstp, resfhp, iap); if (err2) err = err2; + mnt_drop_write(fhp->fh_export->ex_path.mnt); /* * Update the file handle to get the new inode info. */ @@ -1359,6 +1372,9 @@ nfsd_create_v3(struct svc_rqst *rqstp, struct svc_fh *fhp, v_atime = verifier[1]&0x7fffffff; } + host_err = mnt_want_write(fhp->fh_export->ex_path.mnt); + if (host_err) + goto out_nfserr; if (dchild->d_inode) { err = 0; @@ -1390,12 +1406,15 @@ nfsd_create_v3(struct svc_rqst *rqstp, struct svc_fh *fhp, case NFS3_CREATE_GUARDED: err = nfserr_exist; } + mnt_drop_write(fhp->fh_export->ex_path.mnt); goto out; } host_err = vfs_create(dirp, dchild, iap->ia_mode, NULL); - if (host_err < 0) + if (host_err < 0) { + mnt_drop_write(fhp->fh_export->ex_path.mnt); goto out_nfserr; + } if (created) *created = 1; @@ -1420,6 +1439,7 @@ nfsd_create_v3(struct svc_rqst *rqstp, struct svc_fh *fhp, if (err2) err = err2; + mnt_drop_write(fhp->fh_export->ex_path.mnt); /* * Update the filehandle to get the new inode info. */ -- cgit v1.2.3 From 75c3f29de7451677c59580b0a959f694f36aac28 Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:45 -0800 Subject: [PATCH] r/o bind mounts: write counts for link/symlink [AV: add missing nfsd pieces] Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Dave Hansen Signed-off-by: Al Viro --- fs/namei.c | 10 ++++++++++ fs/nfsd/vfs.c | 14 +++++++++++++- 2 files changed, 23 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index 3fbcf2021a2..00df735fb50 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -2387,7 +2387,12 @@ asmlinkage long sys_symlinkat(const char __user *oldname, if (IS_ERR(dentry)) goto out_unlock; + error = mnt_want_write(nd.path.mnt); + if (error) + goto out_dput; error = vfs_symlink(nd.path.dentry->d_inode, dentry, from, S_IALLUGO); + mnt_drop_write(nd.path.mnt); +out_dput: dput(dentry); out_unlock: mutex_unlock(&nd.path.dentry->d_inode->i_mutex); @@ -2482,7 +2487,12 @@ asmlinkage long sys_linkat(int olddfd, const char __user *oldname, error = PTR_ERR(new_dentry); if (IS_ERR(new_dentry)) goto out_unlock; + error = mnt_want_write(nd.path.mnt); + if (error) + goto out_dput; error = vfs_link(old_nd.path.dentry, nd.path.dentry->d_inode, new_dentry); + mnt_drop_write(nd.path.mnt); +out_dput: dput(new_dentry); out_unlock: mutex_unlock(&nd.path.dentry->d_inode->i_mutex); diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index 14d58246702..71f89948230 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1542,6 +1542,10 @@ nfsd_symlink(struct svc_rqst *rqstp, struct svc_fh *fhp, if (iap && (iap->ia_valid & ATTR_MODE)) mode = iap->ia_mode & S_IALLUGO; + host_err = mnt_want_write(fhp->fh_export->ex_path.mnt); + if (host_err) + goto out_nfserr; + if (unlikely(path[plen] != 0)) { char *path_alloced = kmalloc(plen+1, GFP_KERNEL); if (path_alloced == NULL) @@ -1562,6 +1566,8 @@ nfsd_symlink(struct svc_rqst *rqstp, struct svc_fh *fhp, err = nfserrno(host_err); fh_unlock(fhp); + mnt_drop_write(fhp->fh_export->ex_path.mnt); + cerr = fh_compose(resfhp, fhp->fh_export, dnew, fhp); dput(dnew); if (err==0) err = cerr; @@ -1612,6 +1618,11 @@ nfsd_link(struct svc_rqst *rqstp, struct svc_fh *ffhp, dold = tfhp->fh_dentry; dest = dold->d_inode; + host_err = mnt_want_write(tfhp->fh_export->ex_path.mnt); + if (host_err) { + err = nfserrno(host_err); + goto out_dput; + } host_err = vfs_link(dold, dirp, dnew); if (!host_err) { if (EX_ISSYNC(ffhp->fh_export)) { @@ -1625,7 +1636,8 @@ nfsd_link(struct svc_rqst *rqstp, struct svc_fh *ffhp, else err = nfserrno(host_err); } - + mnt_drop_write(tfhp->fh_export->ex_path.mnt); +out_dput: dput(dnew); out_unlock: fh_unlock(ffhp); -- cgit v1.2.3 From 9079b1eb1753f217c3de9f1b7dd7fd549cc3f0cf Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:49 -0800 Subject: [PATCH] r/o bind mounts: get write access for vfs_rename() callers This also uses the little helper in the NFS code to make an if() a little bit less ugly. We introduced the helper at the beginning of the series. Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Dave Hansen Signed-off-by: Andrew Morton Signed-off-by: Al Viro --- fs/namei.c | 4 ++++ fs/nfsd/vfs.c | 17 +++++++++++++---- 2 files changed, 17 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/namei.c b/fs/namei.c index 00df735fb50..83c843b3fea 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -2718,8 +2718,12 @@ static int do_rename(int olddfd, const char *oldname, if (new_dentry == trap) goto exit5; + error = mnt_want_write(oldnd.path.mnt); + if (error) + goto exit5; error = vfs_rename(old_dir->d_inode, old_dentry, new_dir->d_inode, new_dentry); + mnt_drop_write(oldnd.path.mnt); exit5: dput(new_dentry); exit4: diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index 71f89948230..18a4cc9feeb 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1710,13 +1710,20 @@ nfsd_rename(struct svc_rqst *rqstp, struct svc_fh *ffhp, char *fname, int flen, if (ndentry == trap) goto out_dput_new; -#ifdef MSNFS - if ((ffhp->fh_export->ex_flags & NFSEXP_MSNFS) && + if (svc_msnfs(ffhp) && ((atomic_read(&odentry->d_count) > 1) || (atomic_read(&ndentry->d_count) > 1))) { host_err = -EPERM; - } else -#endif + goto out_dput_new; + } + + host_err = -EXDEV; + if (ffhp->fh_export->ex_path.mnt != tfhp->fh_export->ex_path.mnt) + goto out_dput_new; + host_err = mnt_want_write(ffhp->fh_export->ex_path.mnt); + if (host_err) + goto out_dput_new; + host_err = vfs_rename(fdir, odentry, tdir, ndentry); if (!host_err && EX_ISSYNC(tfhp->fh_export)) { host_err = nfsd_sync_dir(tdentry); @@ -1724,6 +1731,8 @@ nfsd_rename(struct svc_rqst *rqstp, struct svc_fh *ffhp, char *fname, int flen, host_err = nfsd_sync_dir(fdentry); } + mnt_drop_write(ffhp->fh_export->ex_path.mnt); + out_dput_new: dput(ndentry); out_dput_old: -- cgit v1.2.3 From 18f335aff86913de3c76f88d32c8135c1da62ce6 Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:38 -0800 Subject: [PATCH] r/o bind mounts: elevate write count for xattr_permission() callers This basically audits the callers of xattr_permission(), which calls permission() and can perform writes to the filesystem. [AV: add missing parts - removexattr() and nfsd posix acls, plug for a leak spotted by Miklos] Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Dave Hansen Signed-off-by: Andrew Morton Signed-off-by: Al Viro --- fs/nfsd/nfs4proc.c | 7 ++++++- fs/nfsd/vfs.c | 4 ++++ fs/xattr.c | 40 ++++++++++++++++++++++++++++++++-------- 3 files changed, 42 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4proc.c b/fs/nfsd/nfs4proc.c index c593db047d8..c309c881bd4 100644 --- a/fs/nfsd/nfs4proc.c +++ b/fs/nfsd/nfs4proc.c @@ -658,14 +658,19 @@ nfsd4_setattr(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, return status; } } + status = mnt_want_write(cstate->current_fh.fh_export->ex_path.mnt); + if (status) + return status; status = nfs_ok; if (setattr->sa_acl != NULL) status = nfsd4_set_nfs4_acl(rqstp, &cstate->current_fh, setattr->sa_acl); if (status) - return status; + goto out; status = nfsd_setattr(rqstp, &cstate->current_fh, &setattr->sa_iattr, 0, (time_t)0); +out: + mnt_drop_write(cstate->current_fh.fh_export->ex_path.mnt); return status; } diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index 18a4cc9feeb..626dfd38528 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -2086,6 +2086,9 @@ nfsd_set_posix_acl(struct svc_fh *fhp, int type, struct posix_acl *acl) } else size = 0; + error = mnt_want_write(fhp->fh_export->ex_path.mnt); + if (error) + goto getout; if (size) error = vfs_setxattr(fhp->fh_dentry, name, value, size, 0); else { @@ -2097,6 +2100,7 @@ nfsd_set_posix_acl(struct svc_fh *fhp, int type, struct posix_acl *acl) error = 0; } } + mnt_drop_write(fhp->fh_export->ex_path.mnt); getout: kfree(value); diff --git a/fs/xattr.c b/fs/xattr.c index 3acab161546..f7062da505d 100644 --- a/fs/xattr.c +++ b/fs/xattr.c @@ -11,6 +11,7 @@ #include #include #include +#include #include #include #include @@ -32,8 +33,6 @@ xattr_permission(struct inode *inode, const char *name, int mask) * filesystem or on an immutable / append-only inode. */ if (mask & MAY_WRITE) { - if (IS_RDONLY(inode)) - return -EROFS; if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) return -EPERM; } @@ -262,7 +261,11 @@ sys_setxattr(char __user *path, char __user *name, void __user *value, error = user_path_walk(path, &nd); if (error) return error; - error = setxattr(nd.path.dentry, name, value, size, flags); + error = mnt_want_write(nd.path.mnt); + if (!error) { + error = setxattr(nd.path.dentry, name, value, size, flags); + mnt_drop_write(nd.path.mnt); + } path_put(&nd.path); return error; } @@ -277,7 +280,11 @@ sys_lsetxattr(char __user *path, char __user *name, void __user *value, error = user_path_walk_link(path, &nd); if (error) return error; - error = setxattr(nd.path.dentry, name, value, size, flags); + error = mnt_want_write(nd.path.mnt); + if (!error) { + error = setxattr(nd.path.dentry, name, value, size, flags); + mnt_drop_write(nd.path.mnt); + } path_put(&nd.path); return error; } @@ -295,7 +302,12 @@ sys_fsetxattr(int fd, char __user *name, void __user *value, return error; dentry = f->f_path.dentry; audit_inode(NULL, dentry); - error = setxattr(dentry, name, value, size, flags); + error = mnt_want_write(f->f_path.mnt); + if (!error) { + error = setxattr(dentry, name, value, size, flags); + mnt_drop_write(f->f_path.mnt); + } +out_fput: fput(f); return error; } @@ -482,7 +494,11 @@ sys_removexattr(char __user *path, char __user *name) error = user_path_walk(path, &nd); if (error) return error; - error = removexattr(nd.path.dentry, name); + error = mnt_want_write(nd.path.mnt); + if (!error) { + error = removexattr(nd.path.dentry, name); + mnt_drop_write(nd.path.mnt); + } path_put(&nd.path); return error; } @@ -496,7 +512,11 @@ sys_lremovexattr(char __user *path, char __user *name) error = user_path_walk_link(path, &nd); if (error) return error; - error = removexattr(nd.path.dentry, name); + error = mnt_want_write(nd.path.mnt); + if (!error) { + error = removexattr(nd.path.dentry, name); + mnt_drop_write(nd.path.mnt); + } path_put(&nd.path); return error; } @@ -513,7 +533,11 @@ sys_fremovexattr(int fd, char __user *name) return error; dentry = f->f_path.dentry; audit_inode(NULL, dentry); - error = removexattr(dentry, name); + error = mnt_want_write(f->f_path.mnt); + if (!error) { + error = removexattr(dentry, name); + mnt_drop_write(f->f_path.mnt); + } fput(f); return error; } -- cgit v1.2.3 From a761a1c03a739f04afd6c8d37fd16405bbe754da Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:39 -0800 Subject: [PATCH] r/o bind mounts: elevate write count for ncp_ioctl() Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Andrew Morton Signed-off-by: Dave Hansen Signed-off-by: Al Viro --- fs/ncpfs/ioctl.c | 54 +++++++++++++++++++++++++++++++++++++++++++++++++++++- 1 file changed, 53 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ncpfs/ioctl.c b/fs/ncpfs/ioctl.c index c67b4bdcf71..ad8f167e54b 100644 --- a/fs/ncpfs/ioctl.c +++ b/fs/ncpfs/ioctl.c @@ -14,6 +14,7 @@ #include #include #include +#include #include #include #include @@ -261,7 +262,7 @@ ncp_get_charsets(struct ncp_server* server, struct ncp_nls_ioctl __user *arg) } #endif /* CONFIG_NCPFS_NLS */ -int ncp_ioctl(struct inode *inode, struct file *filp, +static int __ncp_ioctl(struct inode *inode, struct file *filp, unsigned int cmd, unsigned long arg) { struct ncp_server *server = NCP_SERVER(inode); @@ -822,6 +823,57 @@ outrel: return -EINVAL; } +static int ncp_ioctl_need_write(unsigned int cmd) +{ + switch (cmd) { + case NCP_IOC_GET_FS_INFO: + case NCP_IOC_GET_FS_INFO_V2: + case NCP_IOC_NCPREQUEST: + case NCP_IOC_SETDENTRYTTL: + case NCP_IOC_SIGN_INIT: + case NCP_IOC_LOCKUNLOCK: + case NCP_IOC_SET_SIGN_WANTED: + return 1; + case NCP_IOC_GETOBJECTNAME: + case NCP_IOC_SETOBJECTNAME: + case NCP_IOC_GETPRIVATEDATA: + case NCP_IOC_SETPRIVATEDATA: + case NCP_IOC_SETCHARSETS: + case NCP_IOC_GETCHARSETS: + case NCP_IOC_CONN_LOGGED_IN: + case NCP_IOC_GETDENTRYTTL: + case NCP_IOC_GETMOUNTUID2: + case NCP_IOC_SIGN_WANTED: + case NCP_IOC_GETROOT: + case NCP_IOC_SETROOT: + return 0; + default: + /* unkown IOCTL command, assume write */ + return 1; + } +} + +int ncp_ioctl(struct inode *inode, struct file *filp, + unsigned int cmd, unsigned long arg) +{ + int ret; + + if (ncp_ioctl_need_write(cmd)) { + /* + * inside the ioctl(), any failures which + * are because of file_permission() are + * -EACCESS, so it seems consistent to keep + * that here. + */ + if (mnt_want_write(filp->f_path.mnt)) + return -EACCES; + } + ret = __ncp_ioctl(inode, filp, cmd, arg); + if (ncp_ioctl_need_write(cmd)) + mnt_drop_write(filp->f_path.mnt); + return ret; +} + #ifdef CONFIG_COMPAT long ncp_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg) { -- cgit v1.2.3 From cdb70f3f74b31576cc4d707a3d3b00d159cab8bb Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:41 -0800 Subject: [PATCH] r/o bind mounts: write counts for touch_atime() Remove handling of NULL mnt while we are at it - that can't happen these days. Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Dave Hansen Signed-off-by: Al Viro --- fs/inode.c | 45 ++++++++++++++++++++------------------------- 1 file changed, 20 insertions(+), 25 deletions(-) (limited to 'fs') diff --git a/fs/inode.c b/fs/inode.c index 53245ffcf93..6f6250c08ce 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -1199,42 +1199,37 @@ void touch_atime(struct vfsmount *mnt, struct dentry *dentry) struct inode *inode = dentry->d_inode; struct timespec now; - if (inode->i_flags & S_NOATIME) + if (mnt_want_write(mnt)) return; + if (inode->i_flags & S_NOATIME) + goto out; if (IS_NOATIME(inode)) - return; + goto out; if ((inode->i_sb->s_flags & MS_NODIRATIME) && S_ISDIR(inode->i_mode)) - return; - - /* - * We may have a NULL vfsmount when coming from NFSD - */ - if (mnt) { - if (mnt->mnt_flags & MNT_NOATIME) - return; - if ((mnt->mnt_flags & MNT_NODIRATIME) && S_ISDIR(inode->i_mode)) - return; + goto out; - if (mnt->mnt_flags & MNT_RELATIME) { - /* - * With relative atime, only update atime if the - * previous atime is earlier than either the ctime or - * mtime. - */ - if (timespec_compare(&inode->i_mtime, - &inode->i_atime) < 0 && - timespec_compare(&inode->i_ctime, - &inode->i_atime) < 0) - return; - } + if (mnt->mnt_flags & MNT_NOATIME) + goto out; + if ((mnt->mnt_flags & MNT_NODIRATIME) && S_ISDIR(inode->i_mode)) + goto out; + if (mnt->mnt_flags & MNT_RELATIME) { + /* + * With relative atime, only update atime if the previous + * atime is earlier than either the ctime or mtime. + */ + if (timespec_compare(&inode->i_mtime, &inode->i_atime) < 0 && + timespec_compare(&inode->i_ctime, &inode->i_atime) < 0) + goto out; } now = current_fs_time(inode->i_sb); if (timespec_equal(&inode->i_atime, &now)) - return; + goto out; inode->i_atime = now; mark_inode_dirty_sync(inode); +out: + mnt_drop_write(mnt); } EXPORT_SYMBOL(touch_atime); -- cgit v1.2.3 From 74f9fdfa1f229284ee1ea58fa47f2cdeeb12f6fe Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:42 -0800 Subject: [PATCH] r/o bind mounts: elevate write count for do_utimes() Now includes fix for oops seen by akpm. "never let a libc developer write your kernel code" - hch "nor, apparently, a kernel developer" - akpm Acked-by: Al Viro Signed-off-by: Christoph Hellwig Cc: Valdis Kletnieks Cc: Balbir Singh Signed-off-by: Dave Hansen Signed-off-by: Al Viro --- fs/utimes.c | 18 ++++++++++++------ 1 file changed, 12 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/utimes.c b/fs/utimes.c index b18da9c0b97..a2bef77dc9c 100644 --- a/fs/utimes.c +++ b/fs/utimes.c @@ -2,6 +2,7 @@ #include #include #include +#include #include #include #include @@ -59,6 +60,7 @@ long do_utimes(int dfd, char __user *filename, struct timespec *times, int flags struct inode *inode; struct iattr newattrs; struct file *f = NULL; + struct vfsmount *mnt; error = -EINVAL; if (times && (!nsec_valid(times[0].tv_nsec) || @@ -79,18 +81,20 @@ long do_utimes(int dfd, char __user *filename, struct timespec *times, int flags if (!f) goto out; dentry = f->f_path.dentry; + mnt = f->f_path.mnt; } else { error = __user_walk_fd(dfd, filename, (flags & AT_SYMLINK_NOFOLLOW) ? 0 : LOOKUP_FOLLOW, &nd); if (error) goto out; dentry = nd.path.dentry; + mnt = nd.path.mnt; } inode = dentry->d_inode; - error = -EROFS; - if (IS_RDONLY(inode)) + error = mnt_want_write(mnt); + if (error) goto dput_and_out; /* Don't worry, the checks are done in inode_change_ok() */ @@ -98,7 +102,7 @@ long do_utimes(int dfd, char __user *filename, struct timespec *times, int flags if (times) { error = -EPERM; if (IS_APPEND(inode) || IS_IMMUTABLE(inode)) - goto dput_and_out; + goto mnt_drop_write_and_out; if (times[0].tv_nsec == UTIME_OMIT) newattrs.ia_valid &= ~ATTR_ATIME; @@ -118,22 +122,24 @@ long do_utimes(int dfd, char __user *filename, struct timespec *times, int flags } else { error = -EACCES; if (IS_IMMUTABLE(inode)) - goto dput_and_out; + goto mnt_drop_write_and_out; if (!is_owner_or_cap(inode)) { if (f) { if (!(f->f_mode & FMODE_WRITE)) - goto dput_and_out; + goto mnt_drop_write_and_out; } else { error = vfs_permission(&nd, MAY_WRITE); if (error) - goto dput_and_out; + goto mnt_drop_write_and_out; } } } mutex_lock(&inode->i_mutex); error = notify_change(dentry, &newattrs); mutex_unlock(&inode->i_mutex); +mnt_drop_write_and_out: + mnt_drop_write(mnt); dput_and_out: if (f) fput(f); -- cgit v1.2.3 From 20ddee2c75339cc095f6191c3115f81da8955e96 Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:43 -0800 Subject: [PATCH] r/o bind mounts: write count for file_update_time() Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Dave Hansen Signed-off-by: Al Viro --- fs/inode.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/inode.c b/fs/inode.c index 6f6250c08ce..27ee1af50d0 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -1250,10 +1250,13 @@ void file_update_time(struct file *file) struct inode *inode = file->f_path.dentry->d_inode; struct timespec now; int sync_it = 0; + int err; if (IS_NOCMTIME(inode)) return; - if (IS_RDONLY(inode)) + + err = mnt_want_write(file->f_path.mnt); + if (err) return; now = current_fs_time(inode->i_sb); @@ -1274,6 +1277,7 @@ void file_update_time(struct file *file) if (sync_it) mark_inode_dirty_sync(inode); + mnt_drop_write(file->f_path.mnt); } EXPORT_SYMBOL(file_update_time); -- cgit v1.2.3 From 42a74f206b914db13ee1f5ae932dcd91a77c8579 Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:46 -0800 Subject: [PATCH] r/o bind mounts: elevate write count for ioctls() Some ioctl()s can cause writes to the filesystem. Take these, and make them use mnt_want/drop_write() instead. [AV: updated] Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Dave Hansen Signed-off-by: Andrew Morton Signed-off-by: Al Viro --- fs/ext2/ioctl.c | 57 +++++++++++++++--------- fs/ext3/ioctl.c | 103 ++++++++++++++++++++++++++++--------------- fs/ext4/ioctl.c | 86 ++++++++++++++++++++---------------- fs/fat/file.c | 12 ++--- fs/hfsplus/ioctl.c | 40 ++++++++++------- fs/jfs/ioctl.c | 33 +++++++++----- fs/ocfs2/ioctl.c | 11 ++--- fs/reiserfs/ioctl.c | 63 +++++++++++++++++--------- fs/xfs/linux-2.6/xfs_ioctl.c | 15 ++++--- 9 files changed, 265 insertions(+), 155 deletions(-) (limited to 'fs') diff --git a/fs/ext2/ioctl.c b/fs/ext2/ioctl.c index b8ea11fee5c..de876fa793e 100644 --- a/fs/ext2/ioctl.c +++ b/fs/ext2/ioctl.c @@ -12,6 +12,7 @@ #include #include #include +#include #include #include #include @@ -23,6 +24,7 @@ long ext2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) struct ext2_inode_info *ei = EXT2_I(inode); unsigned int flags; unsigned short rsv_window_size; + int ret; ext2_debug ("cmd = %u, arg = %lu\n", cmd, arg); @@ -34,14 +36,19 @@ long ext2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) case EXT2_IOC_SETFLAGS: { unsigned int oldflags; - if (IS_RDONLY(inode)) - return -EROFS; + ret = mnt_want_write(filp->f_path.mnt); + if (ret) + return ret; - if (!is_owner_or_cap(inode)) - return -EACCES; + if (!is_owner_or_cap(inode)) { + ret = -EACCES; + goto setflags_out; + } - if (get_user(flags, (int __user *) arg)) - return -EFAULT; + if (get_user(flags, (int __user *) arg)) { + ret = -EFAULT; + goto setflags_out; + } if (!S_ISDIR(inode->i_mode)) flags &= ~EXT2_DIRSYNC_FL; @@ -50,7 +57,8 @@ long ext2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) /* Is it quota file? Do not allow user to mess with it */ if (IS_NOQUOTA(inode)) { mutex_unlock(&inode->i_mutex); - return -EPERM; + ret = -EPERM; + goto setflags_out; } oldflags = ei->i_flags; @@ -63,7 +71,8 @@ long ext2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) if ((flags ^ oldflags) & (EXT2_APPEND_FL | EXT2_IMMUTABLE_FL)) { if (!capable(CAP_LINUX_IMMUTABLE)) { mutex_unlock(&inode->i_mutex); - return -EPERM; + ret = -EPERM; + goto setflags_out; } } @@ -75,20 +84,26 @@ long ext2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) ext2_set_inode_flags(inode); inode->i_ctime = CURRENT_TIME_SEC; mark_inode_dirty(inode); - return 0; +setflags_out: + mnt_drop_write(filp->f_path.mnt); + return ret; } case EXT2_IOC_GETVERSION: return put_user(inode->i_generation, (int __user *) arg); case EXT2_IOC_SETVERSION: if (!is_owner_or_cap(inode)) return -EPERM; - if (IS_RDONLY(inode)) - return -EROFS; - if (get_user(inode->i_generation, (int __user *) arg)) - return -EFAULT; - inode->i_ctime = CURRENT_TIME_SEC; - mark_inode_dirty(inode); - return 0; + ret = mnt_want_write(filp->f_path.mnt); + if (ret) + return ret; + if (get_user(inode->i_generation, (int __user *) arg)) { + ret = -EFAULT; + } else { + inode->i_ctime = CURRENT_TIME_SEC; + mark_inode_dirty(inode); + } + mnt_drop_write(filp->f_path.mnt); + return ret; case EXT2_IOC_GETRSVSZ: if (test_opt(inode->i_sb, RESERVATION) && S_ISREG(inode->i_mode) @@ -102,15 +117,16 @@ long ext2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) if (!test_opt(inode->i_sb, RESERVATION) ||!S_ISREG(inode->i_mode)) return -ENOTTY; - if (IS_RDONLY(inode)) - return -EROFS; - - if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER)) + if (!is_owner_or_cap(inode)) return -EACCES; if (get_user(rsv_window_size, (int __user *)arg)) return -EFAULT; + ret = mnt_want_write(filp->f_path.mnt); + if (ret) + return ret; + if (rsv_window_size > EXT2_MAX_RESERVE_BLOCKS) rsv_window_size = EXT2_MAX_RESERVE_BLOCKS; @@ -131,6 +147,7 @@ long ext2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) rsv->rsv_goal_size = rsv_window_size; } mutex_unlock(&ei->truncate_mutex); + mnt_drop_write(filp->f_path.mnt); return 0; } default: diff --git a/fs/ext3/ioctl.c b/fs/ext3/ioctl.c index 023a070f55f..0d0c7015164 100644 --- a/fs/ext3/ioctl.c +++ b/fs/ext3/ioctl.c @@ -12,6 +12,7 @@ #include #include #include +#include #include #include #include @@ -38,14 +39,19 @@ int ext3_ioctl (struct inode * inode, struct file * filp, unsigned int cmd, unsigned int oldflags; unsigned int jflag; - if (IS_RDONLY(inode)) - return -EROFS; + err = mnt_want_write(filp->f_path.mnt); + if (err) + return err; - if (!is_owner_or_cap(inode)) - return -EACCES; + if (!is_owner_or_cap(inode)) { + err = -EACCES; + goto flags_out; + } - if (get_user(flags, (int __user *) arg)) - return -EFAULT; + if (get_user(flags, (int __user *) arg)) { + err = -EFAULT; + goto flags_out; + } if (!S_ISDIR(inode->i_mode)) flags &= ~EXT3_DIRSYNC_FL; @@ -54,7 +60,8 @@ int ext3_ioctl (struct inode * inode, struct file * filp, unsigned int cmd, /* Is it quota file? Do not allow user to mess with it */ if (IS_NOQUOTA(inode)) { mutex_unlock(&inode->i_mutex); - return -EPERM; + err = -EPERM; + goto flags_out; } oldflags = ei->i_flags; @@ -70,7 +77,8 @@ int ext3_ioctl (struct inode * inode, struct file * filp, unsigned int cmd, if ((flags ^ oldflags) & (EXT3_APPEND_FL | EXT3_IMMUTABLE_FL)) { if (!capable(CAP_LINUX_IMMUTABLE)) { mutex_unlock(&inode->i_mutex); - return -EPERM; + err = -EPERM; + goto flags_out; } } @@ -81,7 +89,8 @@ int ext3_ioctl (struct inode * inode, struct file * filp, unsigned int cmd, if ((jflag ^ oldflags) & (EXT3_JOURNAL_DATA_FL)) { if (!capable(CAP_SYS_RESOURCE)) { mutex_unlock(&inode->i_mutex); - return -EPERM; + err = -EPERM; + goto flags_out; } } @@ -89,7 +98,8 @@ int ext3_ioctl (struct inode * inode, struct file * filp, unsigned int cmd, handle = ext3_journal_start(inode, 1); if (IS_ERR(handle)) { mutex_unlock(&inode->i_mutex); - return PTR_ERR(handle); + err = PTR_ERR(handle); + goto flags_out; } if (IS_SYNC(inode)) handle->h_sync = 1; @@ -115,6 +125,8 @@ flags_err: if ((jflag ^ oldflags) & (EXT3_JOURNAL_DATA_FL)) err = ext3_change_inode_journal_flag(inode, jflag); mutex_unlock(&inode->i_mutex); +flags_out: + mnt_drop_write(filp->f_path.mnt); return err; } case EXT3_IOC_GETVERSION: @@ -129,14 +141,18 @@ flags_err: if (!is_owner_or_cap(inode)) return -EPERM; - if (IS_RDONLY(inode)) - return -EROFS; - if (get_user(generation, (int __user *) arg)) - return -EFAULT; - + err = mnt_want_write(filp->f_path.mnt); + if (err) + return err; + if (get_user(generation, (int __user *) arg)) { + err = -EFAULT; + goto setversion_out; + } handle = ext3_journal_start(inode, 1); - if (IS_ERR(handle)) - return PTR_ERR(handle); + if (IS_ERR(handle)) { + err = PTR_ERR(handle); + goto setversion_out; + } err = ext3_reserve_inode_write(handle, inode, &iloc); if (err == 0) { inode->i_ctime = CURRENT_TIME_SEC; @@ -144,6 +160,8 @@ flags_err: err = ext3_mark_iloc_dirty(handle, inode, &iloc); } ext3_journal_stop(handle); +setversion_out: + mnt_drop_write(filp->f_path.mnt); return err; } #ifdef CONFIG_JBD_DEBUG @@ -179,18 +197,24 @@ flags_err: } return -ENOTTY; case EXT3_IOC_SETRSVSZ: { + int err; if (!test_opt(inode->i_sb, RESERVATION) ||!S_ISREG(inode->i_mode)) return -ENOTTY; - if (IS_RDONLY(inode)) - return -EROFS; + err = mnt_want_write(filp->f_path.mnt); + if (err) + return err; - if (!is_owner_or_cap(inode)) - return -EACCES; + if (!is_owner_or_cap(inode)) { + err = -EACCES; + goto setrsvsz_out; + } - if (get_user(rsv_window_size, (int __user *)arg)) - return -EFAULT; + if (get_user(rsv_window_size, (int __user *)arg)) { + err = -EFAULT; + goto setrsvsz_out; + } if (rsv_window_size > EXT3_MAX_RESERVE_BLOCKS) rsv_window_size = EXT3_MAX_RESERVE_BLOCKS; @@ -208,7 +232,9 @@ flags_err: rsv->rsv_goal_size = rsv_window_size; } mutex_unlock(&ei->truncate_mutex); - return 0; +setrsvsz_out: + mnt_drop_write(filp->f_path.mnt); + return err; } case EXT3_IOC_GROUP_EXTEND: { ext3_fsblk_t n_blocks_count; @@ -218,17 +244,20 @@ flags_err: if (!capable(CAP_SYS_RESOURCE)) return -EPERM; - if (IS_RDONLY(inode)) - return -EROFS; - - if (get_user(n_blocks_count, (__u32 __user *)arg)) - return -EFAULT; + err = mnt_want_write(filp->f_path.mnt); + if (err) + return err; + if (get_user(n_blocks_count, (__u32 __user *)arg)) { + err = -EFAULT; + goto group_extend_out; + } err = ext3_group_extend(sb, EXT3_SB(sb)->s_es, n_blocks_count); journal_lock_updates(EXT3_SB(sb)->s_journal); journal_flush(EXT3_SB(sb)->s_journal); journal_unlock_updates(EXT3_SB(sb)->s_journal); - +group_extend_out: + mnt_drop_write(filp->f_path.mnt); return err; } case EXT3_IOC_GROUP_ADD: { @@ -239,18 +268,22 @@ flags_err: if (!capable(CAP_SYS_RESOURCE)) return -EPERM; - if (IS_RDONLY(inode)) - return -EROFS; + err = mnt_want_write(filp->f_path.mnt); + if (err) + return err; if (copy_from_user(&input, (struct ext3_new_group_input __user *)arg, - sizeof(input))) - return -EFAULT; + sizeof(input))) { + err = -EFAULT; + goto group_add_out; + } err = ext3_group_add(sb, &input); journal_lock_updates(EXT3_SB(sb)->s_journal); journal_flush(EXT3_SB(sb)->s_journal); journal_unlock_updates(EXT3_SB(sb)->s_journal); - +group_add_out: + mnt_drop_write(filp->f_path.mnt); return err; } diff --git a/fs/ext4/ioctl.c b/fs/ext4/ioctl.c index 2ed7c37f897..25b13ede808 100644 --- a/fs/ext4/ioctl.c +++ b/fs/ext4/ioctl.c @@ -15,6 +15,7 @@ #include #include #include +#include #include int ext4_ioctl (struct inode * inode, struct file * filp, unsigned int cmd, @@ -38,24 +39,25 @@ int ext4_ioctl (struct inode * inode, struct file * filp, unsigned int cmd, unsigned int oldflags; unsigned int jflag; - if (IS_RDONLY(inode)) - return -EROFS; - if (!is_owner_or_cap(inode)) return -EACCES; if (get_user(flags, (int __user *) arg)) return -EFAULT; + err = mnt_want_write(filp->f_path.mnt); + if (err) + return err; + if (!S_ISDIR(inode->i_mode)) flags &= ~EXT4_DIRSYNC_FL; + err = -EPERM; mutex_lock(&inode->i_mutex); /* Is it quota file? Do not allow user to mess with it */ - if (IS_NOQUOTA(inode)) { - mutex_unlock(&inode->i_mutex); - return -EPERM; - } + if (IS_NOQUOTA(inode)) + goto flags_out; + oldflags = ei->i_flags; /* The JOURNAL_DATA flag is modifiable only by root */ @@ -68,10 +70,8 @@ int ext4_ioctl (struct inode * inode, struct file * filp, unsigned int cmd, * This test looks nicer. Thanks to Pauline Middelink */ if ((flags ^ oldflags) & (EXT4_APPEND_FL | EXT4_IMMUTABLE_FL)) { - if (!capable(CAP_LINUX_IMMUTABLE)) { - mutex_unlock(&inode->i_mutex); - return -EPERM; - } + if (!capable(CAP_LINUX_IMMUTABLE)) + goto flags_out; } /* @@ -79,17 +79,14 @@ int ext4_ioctl (struct inode * inode, struct file * filp, unsigned int cmd, * the relevant capability. */ if ((jflag ^ oldflags) & (EXT4_JOURNAL_DATA_FL)) { - if (!capable(CAP_SYS_RESOURCE)) { - mutex_unlock(&inode->i_mutex); - return -EPERM; - } + if (!capable(CAP_SYS_RESOURCE)) + goto flags_out; } - handle = ext4_journal_start(inode, 1); if (IS_ERR(handle)) { - mutex_unlock(&inode->i_mutex); - return PTR_ERR(handle); + err = PTR_ERR(handle); + goto flags_out; } if (IS_SYNC(inode)) handle->h_sync = 1; @@ -107,14 +104,14 @@ int ext4_ioctl (struct inode * inode, struct file * filp, unsigned int cmd, err = ext4_mark_iloc_dirty(handle, inode, &iloc); flags_err: ext4_journal_stop(handle); - if (err) { - mutex_unlock(&inode->i_mutex); - return err; - } + if (err) + goto flags_out; if ((jflag ^ oldflags) & (EXT4_JOURNAL_DATA_FL)) err = ext4_change_inode_journal_flag(inode, jflag); +flags_out: mutex_unlock(&inode->i_mutex); + mnt_drop_write(filp->f_path.mnt); return err; } case EXT4_IOC_GETVERSION: @@ -129,14 +126,20 @@ flags_err: if (!is_owner_or_cap(inode)) return -EPERM; - if (IS_RDONLY(inode)) - return -EROFS; - if (get_user(generation, (int __user *) arg)) - return -EFAULT; + + err = mnt_want_write(filp->f_path.mnt); + if (err) + return err; + if (get_user(generation, (int __user *) arg)) { + err = -EFAULT; + goto setversion_out; + } handle = ext4_journal_start(inode, 1); - if (IS_ERR(handle)) - return PTR_ERR(handle); + if (IS_ERR(handle)) { + err = PTR_ERR(handle); + goto setversion_out; + } err = ext4_reserve_inode_write(handle, inode, &iloc); if (err == 0) { inode->i_ctime = ext4_current_time(inode); @@ -144,6 +147,8 @@ flags_err: err = ext4_mark_iloc_dirty(handle, inode, &iloc); } ext4_journal_stop(handle); +setversion_out: + mnt_drop_write(filp->f_path.mnt); return err; } #ifdef CONFIG_JBD2_DEBUG @@ -179,19 +184,21 @@ flags_err: } return -ENOTTY; case EXT4_IOC_SETRSVSZ: { + int err; if (!test_opt(inode->i_sb, RESERVATION) ||!S_ISREG(inode->i_mode)) return -ENOTTY; - if (IS_RDONLY(inode)) - return -EROFS; - if (!is_owner_or_cap(inode)) return -EACCES; if (get_user(rsv_window_size, (int __user *)arg)) return -EFAULT; + err = mnt_want_write(filp->f_path.mnt); + if (err) + return err; + if (rsv_window_size > EXT4_MAX_RESERVE_BLOCKS) rsv_window_size = EXT4_MAX_RESERVE_BLOCKS; @@ -208,6 +215,7 @@ flags_err: rsv->rsv_goal_size = rsv_window_size; } up_write(&ei->i_data_sem); + mnt_drop_write(filp->f_path.mnt); return 0; } case EXT4_IOC_GROUP_EXTEND: { @@ -218,16 +226,18 @@ flags_err: if (!capable(CAP_SYS_RESOURCE)) return -EPERM; - if (IS_RDONLY(inode)) - return -EROFS; - if (get_user(n_blocks_count, (__u32 __user *)arg)) return -EFAULT; + err = mnt_want_write(filp->f_path.mnt); + if (err) + return err; + err = ext4_group_extend(sb, EXT4_SB(sb)->s_es, n_blocks_count); jbd2_journal_lock_updates(EXT4_SB(sb)->s_journal); jbd2_journal_flush(EXT4_SB(sb)->s_journal); jbd2_journal_unlock_updates(EXT4_SB(sb)->s_journal); + mnt_drop_write(filp->f_path.mnt); return err; } @@ -239,17 +249,19 @@ flags_err: if (!capable(CAP_SYS_RESOURCE)) return -EPERM; - if (IS_RDONLY(inode)) - return -EROFS; - if (copy_from_user(&input, (struct ext4_new_group_input __user *)arg, sizeof(input))) return -EFAULT; + err = mnt_want_write(filp->f_path.mnt); + if (err) + return err; + err = ext4_group_add(sb, &input); jbd2_journal_lock_updates(EXT4_SB(sb)->s_journal); jbd2_journal_flush(EXT4_SB(sb)->s_journal); jbd2_journal_unlock_updates(EXT4_SB(sb)->s_journal); + mnt_drop_write(filp->f_path.mnt); return err; } diff --git a/fs/fat/file.c b/fs/fat/file.c index c614175876e..2a3bed96704 100644 --- a/fs/fat/file.c +++ b/fs/fat/file.c @@ -8,6 +8,7 @@ #include #include +#include #include #include #include @@ -46,10 +47,9 @@ int fat_generic_ioctl(struct inode *inode, struct file *filp, mutex_lock(&inode->i_mutex); - if (IS_RDONLY(inode)) { - err = -EROFS; - goto up; - } + err = mnt_want_write(filp->f_path.mnt); + if (err) + goto up_no_drop_write; /* * ATTR_VOLUME and ATTR_DIR cannot be changed; this also @@ -105,7 +105,9 @@ int fat_generic_ioctl(struct inode *inode, struct file *filp, MSDOS_I(inode)->i_attrs = attr & ATTR_UNUSED; mark_inode_dirty(inode); - up: +up: + mnt_drop_write(filp->f_path.mnt); +up_no_drop_write: mutex_unlock(&inode->i_mutex); return err; } diff --git a/fs/hfsplus/ioctl.c b/fs/hfsplus/ioctl.c index b60c0affbec..f457d2ca51a 100644 --- a/fs/hfsplus/ioctl.c +++ b/fs/hfsplus/ioctl.c @@ -14,6 +14,7 @@ #include #include +#include #include #include #include @@ -35,25 +36,32 @@ int hfsplus_ioctl(struct inode *inode, struct file *filp, unsigned int cmd, flags |= FS_NODUMP_FL; /* EXT2_NODUMP_FL */ return put_user(flags, (int __user *)arg); case HFSPLUS_IOC_EXT2_SETFLAGS: { - if (IS_RDONLY(inode)) - return -EROFS; - - if (!is_owner_or_cap(inode)) - return -EACCES; - - if (get_user(flags, (int __user *)arg)) - return -EFAULT; - + int err = 0; + err = mnt_want_write(filp->f_path.mnt); + if (err) + return err; + + if (!is_owner_or_cap(inode)) { + err = -EACCES; + goto setflags_out; + } + if (get_user(flags, (int __user *)arg)) { + err = -EFAULT; + goto setflags_out; + } if (flags & (FS_IMMUTABLE_FL|FS_APPEND_FL) || HFSPLUS_I(inode).rootflags & (HFSPLUS_FLG_IMMUTABLE|HFSPLUS_FLG_APPEND)) { - if (!capable(CAP_LINUX_IMMUTABLE)) - return -EPERM; + if (!capable(CAP_LINUX_IMMUTABLE)) { + err = -EPERM; + goto setflags_out; + } } /* don't silently ignore unsupported ext2 flags */ - if (flags & ~(FS_IMMUTABLE_FL|FS_APPEND_FL|FS_NODUMP_FL)) - return -EOPNOTSUPP; - + if (flags & ~(FS_IMMUTABLE_FL|FS_APPEND_FL|FS_NODUMP_FL)) { + err = -EOPNOTSUPP; + goto setflags_out; + } if (flags & FS_IMMUTABLE_FL) { /* EXT2_IMMUTABLE_FL */ inode->i_flags |= S_IMMUTABLE; HFSPLUS_I(inode).rootflags |= HFSPLUS_FLG_IMMUTABLE; @@ -75,7 +83,9 @@ int hfsplus_ioctl(struct inode *inode, struct file *filp, unsigned int cmd, inode->i_ctime = CURRENT_TIME_SEC; mark_inode_dirty(inode); - return 0; +setflags_out: + mnt_drop_write(filp->f_path.mnt); + return err; } default: return -ENOTTY; diff --git a/fs/jfs/ioctl.c b/fs/jfs/ioctl.c index a1f8e375ad2..afe222bf300 100644 --- a/fs/jfs/ioctl.c +++ b/fs/jfs/ioctl.c @@ -8,6 +8,7 @@ #include #include #include +#include #include #include #include @@ -65,23 +66,30 @@ long jfs_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) return put_user(flags, (int __user *) arg); case JFS_IOC_SETFLAGS: { unsigned int oldflags; + int err; - if (IS_RDONLY(inode)) - return -EROFS; + err = mnt_want_write(filp->f_path.mnt); + if (err) + return err; - if (!is_owner_or_cap(inode)) - return -EACCES; - - if (get_user(flags, (int __user *) arg)) - return -EFAULT; + if (!is_owner_or_cap(inode)) { + err = -EACCES; + goto setflags_out; + } + if (get_user(flags, (int __user *) arg)) { + err = -EFAULT; + goto setflags_out; + } flags = jfs_map_ext2(flags, 1); if (!S_ISDIR(inode->i_mode)) flags &= ~JFS_DIRSYNC_FL; /* Is it quota file? Do not allow user to mess with it */ - if (IS_NOQUOTA(inode)) - return -EPERM; + if (IS_NOQUOTA(inode)) { + err = -EPERM; + goto setflags_out; + } /* Lock against other parallel changes of flags */ mutex_lock(&inode->i_mutex); @@ -98,7 +106,8 @@ long jfs_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) (JFS_APPEND_FL | JFS_IMMUTABLE_FL))) { if (!capable(CAP_LINUX_IMMUTABLE)) { mutex_unlock(&inode->i_mutex); - return -EPERM; + err = -EPERM; + goto setflags_out; } } @@ -110,7 +119,9 @@ long jfs_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) mutex_unlock(&inode->i_mutex); inode->i_ctime = CURRENT_TIME_SEC; mark_inode_dirty(inode); - return 0; +setflags_out: + mnt_drop_write(filp->f_path.mnt); + return err; } default: return -ENOTTY; diff --git a/fs/ocfs2/ioctl.c b/fs/ocfs2/ioctl.c index b413166dd16..7b142f0ce99 100644 --- a/fs/ocfs2/ioctl.c +++ b/fs/ocfs2/ioctl.c @@ -60,10 +60,6 @@ static int ocfs2_set_inode_attr(struct inode *inode, unsigned flags, goto bail; } - status = -EROFS; - if (IS_RDONLY(inode)) - goto bail_unlock; - status = -EACCES; if (!is_owner_or_cap(inode)) goto bail_unlock; @@ -134,8 +130,13 @@ long ocfs2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) if (get_user(flags, (int __user *) arg)) return -EFAULT; - return ocfs2_set_inode_attr(inode, flags, + status = mnt_want_write(filp->f_path.mnt); + if (status) + return status; + status = ocfs2_set_inode_attr(inode, flags, OCFS2_FL_MODIFIABLE); + mnt_drop_write(filp->f_path.mnt); + return status; case OCFS2_IOC_RESVSP: case OCFS2_IOC_RESVSP64: case OCFS2_IOC_UNRESVSP: diff --git a/fs/reiserfs/ioctl.c b/fs/reiserfs/ioctl.c index e0f0f098a52..74363a7aacb 100644 --- a/fs/reiserfs/ioctl.c +++ b/fs/reiserfs/ioctl.c @@ -4,6 +4,7 @@ #include #include +#include #include #include #include @@ -25,6 +26,7 @@ int reiserfs_ioctl(struct inode *inode, struct file *filp, unsigned int cmd, unsigned long arg) { unsigned int flags; + int err = 0; switch (cmd) { case REISERFS_IOC_UNPACK: @@ -48,50 +50,67 @@ int reiserfs_ioctl(struct inode *inode, struct file *filp, unsigned int cmd, if (!reiserfs_attrs(inode->i_sb)) return -ENOTTY; - if (IS_RDONLY(inode)) - return -EROFS; + err = mnt_want_write(filp->f_path.mnt); + if (err) + return err; - if (!is_owner_or_cap(inode)) - return -EPERM; - - if (get_user(flags, (int __user *)arg)) - return -EFAULT; - - /* Is it quota file? Do not allow user to mess with it. */ - if (IS_NOQUOTA(inode)) - return -EPERM; + if (!is_owner_or_cap(inode)) { + err = -EPERM; + goto setflags_out; + } + if (get_user(flags, (int __user *)arg)) { + err = -EFAULT; + goto setflags_out; + } + /* + * Is it quota file? Do not allow user to mess with it + */ + if (IS_NOQUOTA(inode)) { + err = -EPERM; + goto setflags_out; + } if (((flags ^ REISERFS_I(inode)-> i_attrs) & (REISERFS_IMMUTABLE_FL | REISERFS_APPEND_FL)) - && !capable(CAP_LINUX_IMMUTABLE)) - return -EPERM; - + && !capable(CAP_LINUX_IMMUTABLE)) { + err = -EPERM; + goto setflags_out; + } if ((flags & REISERFS_NOTAIL_FL) && S_ISREG(inode->i_mode)) { int result; result = reiserfs_unpack(inode, filp); - if (result) - return result; + if (result) { + err = result; + goto setflags_out; + } } sd_attrs_to_i_attrs(flags, inode); REISERFS_I(inode)->i_attrs = flags; inode->i_ctime = CURRENT_TIME_SEC; mark_inode_dirty(inode); - return 0; +setflags_out: + mnt_drop_write(filp->f_path.mnt); + return err; } case REISERFS_IOC_GETVERSION: return put_user(inode->i_generation, (int __user *)arg); case REISERFS_IOC_SETVERSION: if (!is_owner_or_cap(inode)) return -EPERM; - if (IS_RDONLY(inode)) - return -EROFS; - if (get_user(inode->i_generation, (int __user *)arg)) - return -EFAULT; + err = mnt_want_write(filp->f_path.mnt); + if (err) + return err; + if (get_user(inode->i_generation, (int __user *)arg)) { + err = -EFAULT; + goto setversion_out; + } inode->i_ctime = CURRENT_TIME_SEC; mark_inode_dirty(inode); - return 0; +setversion_out: + mnt_drop_write(filp->f_path.mnt); + return err; default: return -ENOTTY; } diff --git a/fs/xfs/linux-2.6/xfs_ioctl.c b/fs/xfs/linux-2.6/xfs_ioctl.c index bf775979385..4ddb86b73c6 100644 --- a/fs/xfs/linux-2.6/xfs_ioctl.c +++ b/fs/xfs/linux-2.6/xfs_ioctl.c @@ -535,8 +535,6 @@ xfs_attrmulti_attr_set( char *kbuf; int error = EFAULT; - if (IS_RDONLY(inode)) - return -EROFS; if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) return EPERM; if (len > XATTR_SIZE_MAX) @@ -562,8 +560,6 @@ xfs_attrmulti_attr_remove( char *name, __uint32_t flags) { - if (IS_RDONLY(inode)) - return -EROFS; if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) return EPERM; return xfs_attr_remove(XFS_I(inode), name, flags); @@ -573,6 +569,7 @@ STATIC int xfs_attrmulti_by_handle( xfs_mount_t *mp, void __user *arg, + struct file *parfilp, struct inode *parinode) { int error; @@ -626,13 +623,21 @@ xfs_attrmulti_by_handle( &ops[i].am_length, ops[i].am_flags); break; case ATTR_OP_SET: + ops[i].am_error = mnt_want_write(parfilp->f_path.mnt); + if (ops[i].am_error) + break; ops[i].am_error = xfs_attrmulti_attr_set(inode, attr_name, ops[i].am_attrvalue, ops[i].am_length, ops[i].am_flags); + mnt_drop_write(parfilp->f_path.mnt); break; case ATTR_OP_REMOVE: + ops[i].am_error = mnt_want_write(parfilp->f_path.mnt); + if (ops[i].am_error) + break; ops[i].am_error = xfs_attrmulti_attr_remove(inode, attr_name, ops[i].am_flags); + mnt_drop_write(parfilp->f_path.mnt); break; default: ops[i].am_error = EINVAL; @@ -1133,7 +1138,7 @@ xfs_ioctl( return xfs_attrlist_by_handle(mp, arg, inode); case XFS_IOC_ATTRMULTI_BY_HANDLE: - return xfs_attrmulti_by_handle(mp, arg, inode); + return xfs_attrmulti_by_handle(mp, arg, filp, inode); case XFS_IOC_SWAPEXT: { error = xfs_swapext((struct xfs_swapext __user *)arg); -- cgit v1.2.3 From 4a3fd211ccfc08a88edc824300e25a87785c6a5f Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:48 -0800 Subject: [PATCH] r/o bind mounts: elevate write count for open()s This is the first really tricky patch in the series. It elevates the writer count on a mount each time a non-special file is opened for write. We used to do this in may_open(), but Miklos pointed out that __dentry_open() is used as well to create filps. This will cover even those cases, while a call in may_open() would not have. There is also an elevated count around the vfs_create() call in open_namei(). See the comments for more details, but we need this to fix a 'create, remount, fail r/w open()' race. Some filesystems forego the use of normal vfs calls to create struct files. Make sure that these users elevate the mnt writer count because they will get __fput(), and we need to make sure they're balanced. Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Dave Hansen Signed-off-by: Andrew Morton Signed-off-by: Al Viro --- fs/file_table.c | 14 +++++++++++ fs/namei.c | 75 +++++++++++++++++++++++++++++++++++++++++++++++++-------- fs/open.c | 36 +++++++++++++++++++++++++-- 3 files changed, 113 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/file_table.c b/fs/file_table.c index 3f73eb1f195..71efc700022 100644 --- a/fs/file_table.c +++ b/fs/file_table.c @@ -199,6 +199,17 @@ int init_file(struct file *file, struct vfsmount *mnt, struct dentry *dentry, file->f_mapping = dentry->d_inode->i_mapping; file->f_mode = mode; file->f_op = fop; + + /* + * These mounts don't really matter in practice + * for r/o bind mounts. They aren't userspace- + * visible. We do this for consistency, and so + * that we can do debugging checks at __fput() + */ + if ((mode & FMODE_WRITE) && !special_file(dentry->d_inode->i_mode)) { + error = mnt_want_write(mnt); + WARN_ON(error); + } return error; } EXPORT_SYMBOL(init_file); @@ -221,10 +232,13 @@ EXPORT_SYMBOL(fput); */ void drop_file_write_access(struct file *file) { + struct vfsmount *mnt = file->f_path.mnt; struct dentry *dentry = file->f_path.dentry; struct inode *inode = dentry->d_inode; put_write_access(inode); + if (!special_file(inode->i_mode)) + mnt_drop_write(mnt); } EXPORT_SYMBOL_GPL(drop_file_write_access); diff --git a/fs/namei.c b/fs/namei.c index 83c843b3fea..e179f71bfcb 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -1623,8 +1623,7 @@ int may_open(struct nameidata *nd, int acc_mode, int flag) return -EACCES; flag &= ~O_TRUNC; - } else if (IS_RDONLY(inode) && (acc_mode & MAY_WRITE)) - return -EROFS; + } error = vfs_permission(nd, acc_mode); if (error) @@ -1724,18 +1723,32 @@ static inline int open_to_namei_flags(int flag) return flag; } +static int open_will_write_to_fs(int flag, struct inode *inode) +{ + /* + * We'll never write to the fs underlying + * a device file. + */ + if (special_file(inode->i_mode)) + return 0; + return (flag & O_TRUNC); +} + /* - * Note that the low bits of "flag" aren't the same as in the open - * system call. See open_to_namei_flags(). + * Note that the low bits of the passed in "open_flag" + * are not the same as in the local variable "flag". See + * open_to_namei_flags() for more details. */ struct file *do_filp_open(int dfd, const char *pathname, int open_flag, int mode) { + struct file *filp; struct nameidata nd; int acc_mode, error; struct path path; struct dentry *dir; int count = 0; + int will_write; int flag = open_to_namei_flags(open_flag); acc_mode = ACC_MODE(flag); @@ -1791,17 +1804,30 @@ do_last: } if (IS_ERR(nd.intent.open.file)) { - mutex_unlock(&dir->d_inode->i_mutex); error = PTR_ERR(nd.intent.open.file); - goto exit_dput; + goto exit_mutex_unlock; } /* Negative dentry, just create the file */ if (!path.dentry->d_inode) { - error = __open_namei_create(&nd, &path, flag, mode); + /* + * This write is needed to ensure that a + * ro->rw transition does not occur between + * the time when the file is created and when + * a permanent write count is taken through + * the 'struct file' in nameidata_to_filp(). + */ + error = mnt_want_write(nd.path.mnt); if (error) + goto exit_mutex_unlock; + error = __open_namei_create(&nd, &path, flag, mode); + if (error) { + mnt_drop_write(nd.path.mnt); goto exit; - return nameidata_to_filp(&nd, open_flag); + } + filp = nameidata_to_filp(&nd, open_flag); + mnt_drop_write(nd.path.mnt); + return filp; } /* @@ -1831,11 +1857,40 @@ do_last: if (path.dentry->d_inode && S_ISDIR(path.dentry->d_inode->i_mode)) goto exit; ok: + /* + * Consider: + * 1. may_open() truncates a file + * 2. a rw->ro mount transition occurs + * 3. nameidata_to_filp() fails due to + * the ro mount. + * That would be inconsistent, and should + * be avoided. Taking this mnt write here + * ensures that (2) can not occur. + */ + will_write = open_will_write_to_fs(flag, nd.path.dentry->d_inode); + if (will_write) { + error = mnt_want_write(nd.path.mnt); + if (error) + goto exit; + } error = may_open(&nd, acc_mode, flag); - if (error) + if (error) { + if (will_write) + mnt_drop_write(nd.path.mnt); goto exit; - return nameidata_to_filp(&nd, open_flag); + } + filp = nameidata_to_filp(&nd, open_flag); + /* + * It is now safe to drop the mnt write + * because the filp has had a write taken + * on its behalf. + */ + if (will_write) + mnt_drop_write(nd.path.mnt); + return filp; +exit_mutex_unlock: + mutex_unlock(&dir->d_inode->i_mutex); exit_dput: path_put_conditional(&path, &nd); exit: diff --git a/fs/open.c b/fs/open.c index 8111947905d..e12f1701032 100644 --- a/fs/open.c +++ b/fs/open.c @@ -730,6 +730,35 @@ out: return error; } +/* + * You have to be very careful that these write + * counts get cleaned up in error cases and + * upon __fput(). This should probably never + * be called outside of __dentry_open(). + */ +static inline int __get_file_write_access(struct inode *inode, + struct vfsmount *mnt) +{ + int error; + error = get_write_access(inode); + if (error) + return error; + /* + * Do not take mount writer counts on + * special files since no writes to + * the mount itself will occur. + */ + if (!special_file(inode->i_mode)) { + /* + * Balanced in __fput() + */ + error = mnt_want_write(mnt); + if (error) + put_write_access(inode); + } + return error; +} + static struct file *__dentry_open(struct dentry *dentry, struct vfsmount *mnt, int flags, struct file *f, int (*open)(struct inode *, struct file *)) @@ -742,7 +771,7 @@ static struct file *__dentry_open(struct dentry *dentry, struct vfsmount *mnt, FMODE_PREAD | FMODE_PWRITE; inode = dentry->d_inode; if (f->f_mode & FMODE_WRITE) { - error = get_write_access(inode); + error = __get_file_write_access(inode, mnt); if (error) goto cleanup_file; } @@ -784,8 +813,11 @@ static struct file *__dentry_open(struct dentry *dentry, struct vfsmount *mnt, cleanup_all: fops_put(f->f_op); - if (f->f_mode & FMODE_WRITE) + if (f->f_mode & FMODE_WRITE) { put_write_access(inode); + if (!special_file(inode->i_mode)) + mnt_drop_write(mnt); + } file_kill(f); f->f_path.dentry = NULL; f->f_path.mnt = NULL; -- cgit v1.2.3 From 2af482a7edfb8810539cacc2fdd8242611ca43bb Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:50 -0800 Subject: [PATCH] r/o bind mounts: elevate write count for chmod/chown callers chown/chmod,etc... don't call permission in the same way that the normal "open for write" calls do. They still write to the filesystem, so bump the write count during these operations. Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Dave Hansen Signed-off-by: Andrew Morton Signed-off-by: Al Viro --- fs/open.c | 39 ++++++++++++++++++++++++++++++--------- 1 file changed, 30 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/open.c b/fs/open.c index e12f1701032..e2df8fd1eb0 100644 --- a/fs/open.c +++ b/fs/open.c @@ -567,12 +567,12 @@ asmlinkage long sys_fchmod(unsigned int fd, mode_t mode) audit_inode(NULL, dentry); - err = -EROFS; - if (IS_RDONLY(inode)) + err = mnt_want_write(file->f_path.mnt); + if (err) goto out_putf; err = -EPERM; if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) - goto out_putf; + goto out_drop_write; mutex_lock(&inode->i_mutex); if (mode == (mode_t) -1) mode = inode->i_mode; @@ -581,6 +581,8 @@ asmlinkage long sys_fchmod(unsigned int fd, mode_t mode) err = notify_change(dentry, &newattrs); mutex_unlock(&inode->i_mutex); +out_drop_write: + mnt_drop_write(file->f_path.mnt); out_putf: fput(file); out: @@ -600,13 +602,13 @@ asmlinkage long sys_fchmodat(int dfd, const char __user *filename, goto out; inode = nd.path.dentry->d_inode; - error = -EROFS; - if (IS_RDONLY(inode)) + error = mnt_want_write(nd.path.mnt); + if (error) goto dput_and_out; error = -EPERM; if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) - goto dput_and_out; + goto out_drop_write; mutex_lock(&inode->i_mutex); if (mode == (mode_t) -1) @@ -616,6 +618,8 @@ asmlinkage long sys_fchmodat(int dfd, const char __user *filename, error = notify_change(nd.path.dentry, &newattrs); mutex_unlock(&inode->i_mutex); +out_drop_write: + mnt_drop_write(nd.path.mnt); dput_and_out: path_put(&nd.path); out: @@ -638,9 +642,6 @@ static int chown_common(struct dentry * dentry, uid_t user, gid_t group) printk(KERN_ERR "chown_common: NULL inode\n"); goto out; } - error = -EROFS; - if (IS_RDONLY(inode)) - goto out; error = -EPERM; if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) goto out; @@ -671,7 +672,12 @@ asmlinkage long sys_chown(const char __user * filename, uid_t user, gid_t group) error = user_path_walk(filename, &nd); if (error) goto out; + error = mnt_want_write(nd.path.mnt); + if (error) + goto out_release; error = chown_common(nd.path.dentry, user, group); + mnt_drop_write(nd.path.mnt); +out_release: path_put(&nd.path); out: return error; @@ -691,7 +697,12 @@ asmlinkage long sys_fchownat(int dfd, const char __user *filename, uid_t user, error = __user_walk_fd(dfd, filename, follow, &nd); if (error) goto out; + error = mnt_want_write(nd.path.mnt); + if (error) + goto out_release; error = chown_common(nd.path.dentry, user, group); + mnt_drop_write(nd.path.mnt); +out_release: path_put(&nd.path); out: return error; @@ -705,7 +716,12 @@ asmlinkage long sys_lchown(const char __user * filename, uid_t user, gid_t group error = user_path_walk_link(filename, &nd); if (error) goto out; + error = mnt_want_write(nd.path.mnt); + if (error) + goto out_release; error = chown_common(nd.path.dentry, user, group); + mnt_drop_write(nd.path.mnt); +out_release: path_put(&nd.path); out: return error; @@ -722,9 +738,14 @@ asmlinkage long sys_fchown(unsigned int fd, uid_t user, gid_t group) if (!file) goto out; + error = mnt_want_write(file->f_path.mnt); + if (error) + goto out_fput; dentry = file->f_path.dentry; audit_inode(NULL, dentry); error = chown_common(dentry, user, group); + mnt_drop_write(file->f_path.mnt); +out_fput: fput(file); out: return error; -- cgit v1.2.3 From 9ac9b8474c39c3ae2c2b37d8e1f08db8a9146124 Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:52 -0800 Subject: [PATCH] r/o bind mounts: write counts for truncate() Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Dave Hansen Signed-off-by: Al Viro --- fs/open.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/open.c b/fs/open.c index e2df8fd1eb0..4d690e33446 100644 --- a/fs/open.c +++ b/fs/open.c @@ -244,21 +244,21 @@ static long do_sys_truncate(const char __user * path, loff_t length) if (!S_ISREG(inode->i_mode)) goto dput_and_out; - error = vfs_permission(&nd, MAY_WRITE); + error = mnt_want_write(nd.path.mnt); if (error) goto dput_and_out; - error = -EROFS; - if (IS_RDONLY(inode)) - goto dput_and_out; + error = vfs_permission(&nd, MAY_WRITE); + if (error) + goto mnt_drop_write_and_out; error = -EPERM; if (IS_IMMUTABLE(inode) || IS_APPEND(inode)) - goto dput_and_out; + goto mnt_drop_write_and_out; error = get_write_access(inode); if (error) - goto dput_and_out; + goto mnt_drop_write_and_out; /* * Make sure that there are no leases. get_write_access() protects @@ -276,6 +276,8 @@ static long do_sys_truncate(const char __user * path, loff_t length) put_write_and_out: put_write_access(inode); +mnt_drop_write_and_out: + mnt_drop_write(nd.path.mnt); dput_and_out: path_put(&nd.path); out: -- cgit v1.2.3 From 2f676cbc0d60ae806216c7a61c6971bd72dedde8 Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:55 -0800 Subject: [PATCH] r/o bind mounts: make access() use new r/o helper It is OK to let access() go without using a mnt_want/drop_write() pair because it doesn't actually do writes to the filesystem, and it is inherently racy anyway. This is a rare case when it is OK to use __mnt_is_readonly() directly. Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Dave Hansen Signed-off-by: Andrew Morton Signed-off-by: Al Viro --- fs/open.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/open.c b/fs/open.c index 4d690e33446..e58382d57e7 100644 --- a/fs/open.c +++ b/fs/open.c @@ -459,8 +459,17 @@ asmlinkage long sys_faccessat(int dfd, const char __user *filename, int mode) if(res || !(mode & S_IWOTH) || special_file(nd.path.dentry->d_inode->i_mode)) goto out_path_release; - - if(IS_RDONLY(nd.path.dentry->d_inode)) + /* + * This is a rare case where using __mnt_is_readonly() + * is OK without a mnt_want/drop_write() pair. Since + * no actual write to the fs is performed here, we do + * not need to telegraph to that to anyone. + * + * By doing this, we accept that this access is + * inherently racy and know that the fs may change + * state before we even see this result. + */ + if (__mnt_is_readonly(nd.path.mnt)) res = -EROFS; out_path_release: -- cgit v1.2.3 From ec82687f29127a954dd0da95dc1e0a4ce92b560c Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:53 -0800 Subject: [PATCH] r/o bind mounts: elevate count for xfs timestamp updates Elevate the write count during the xfs m/ctime updates. XFS has to do it's own timestamp updates due to an unfortunate VFS design limitation, so it will have to track writers by itself aswell. [hch: split out from the touch_atime patch as it's not related to it at all] Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Dave Hansen Signed-off-by: Al Viro --- fs/xfs/linux-2.6/xfs_iops.c | 7 ------- fs/xfs/linux-2.6/xfs_lrw.c | 9 ++++++++- 2 files changed, 8 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/xfs/linux-2.6/xfs_iops.c b/fs/xfs/linux-2.6/xfs_iops.c index 0c958cf7775..a1237dad643 100644 --- a/fs/xfs/linux-2.6/xfs_iops.c +++ b/fs/xfs/linux-2.6/xfs_iops.c @@ -155,13 +155,6 @@ xfs_ichgtime_fast( */ ASSERT((flags & XFS_ICHGTIME_ACC) == 0); - /* - * We're not supposed to change timestamps in readonly-mounted - * filesystems. Throw it away if anyone asks us. - */ - if (unlikely(IS_RDONLY(inode))) - return; - if (flags & XFS_ICHGTIME_MOD) { tvp = &inode->i_mtime; ip->i_d.di_mtime.t_sec = (__int32_t)tvp->tv_sec; diff --git a/fs/xfs/linux-2.6/xfs_lrw.c b/fs/xfs/linux-2.6/xfs_lrw.c index 21c0dbc7409..1ebd8004469 100644 --- a/fs/xfs/linux-2.6/xfs_lrw.c +++ b/fs/xfs/linux-2.6/xfs_lrw.c @@ -51,6 +51,7 @@ #include "xfs_vnodeops.h" #include +#include #include @@ -670,10 +671,16 @@ start: if (new_size > xip->i_size) xip->i_new_size = new_size; - if (likely(!(ioflags & IO_INVIS))) { + /* + * We're not supposed to change timestamps in readonly-mounted + * filesystems. Throw it away if anyone asks us. + */ + if (likely(!(ioflags & IO_INVIS) && + !mnt_want_write(file->f_path.mnt))) { file_update_time(file); xfs_ichgtime_fast(xip, inode, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); + mnt_drop_write(file->f_path.mnt); } /* -- cgit v1.2.3 From 2c463e95480829a2fe8f386589516e13b1289db6 Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:56 -0800 Subject: [PATCH] r/o bind mounts: check mnt instead of superblock directly If we depend on the inodes for writeability, we will not catch the r/o mounts when implemented. This patches uses __mnt_want_write(). It does not guarantee that the mount will stay writeable after the check. But, this is OK for one of the checks because it is just for a printk(). The other two are probably unnecessary and duplicate existing checks in the VFS. This won't make them better checks than before, but it will make them detect r/o mounts. Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Dave Hansen Signed-off-by: Andrew Morton Signed-off-by: Al Viro --- fs/nfs/dir.c | 3 ++- fs/nfsd/vfs.c | 5 +++-- 2 files changed, 5 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index 6cea7479c5b..d9e30ac2798 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -967,7 +967,8 @@ static int is_atomic_open(struct inode *dir, struct nameidata *nd) if (nd->flags & LOOKUP_DIRECTORY) return 0; /* Are we trying to write to a read only partition? */ - if (IS_RDONLY(dir) && (nd->intent.open.flags & (O_CREAT|O_TRUNC|FMODE_WRITE))) + if (__mnt_is_readonly(nd->path.mnt) && + (nd->intent.open.flags & (O_CREAT|O_TRUNC|FMODE_WRITE))) return 0; return 1; } diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index 626dfd38528..304bf5f643c 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1912,7 +1912,7 @@ nfsd_permission(struct svc_rqst *rqstp, struct svc_export *exp, inode->i_mode, IS_IMMUTABLE(inode)? " immut" : "", IS_APPEND(inode)? " append" : "", - IS_RDONLY(inode)? " ro" : ""); + __mnt_is_readonly(exp->ex_path.mnt)? " ro" : ""); dprintk(" owner %d/%d user %d/%d\n", inode->i_uid, inode->i_gid, current->fsuid, current->fsgid); #endif @@ -1923,7 +1923,8 @@ nfsd_permission(struct svc_rqst *rqstp, struct svc_export *exp, */ if (!(acc & MAY_LOCAL_ACCESS)) if (acc & (MAY_WRITE | MAY_SATTR | MAY_TRUNC)) { - if (exp_rdonly(rqstp, exp) || IS_RDONLY(inode)) + if (exp_rdonly(rqstp, exp) || + __mnt_is_readonly(exp->ex_path.mnt)) return nfserr_rofs; if (/* (acc & MAY_WRITE) && */ IS_IMMUTABLE(inode)) return nfserr_perm; -- cgit v1.2.3 From 3d733633a633065729c9e4e254b2e5442c00ef7e Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:37:59 -0800 Subject: [PATCH] r/o bind mounts: track numbers of writers to mounts This is the real meat of the entire series. It actually implements the tracking of the number of writers to a mount. However, it causes scalability problems because there can be hundreds of cpus doing open()/close() on files on the same mnt at the same time. Even an atomic_t in the mnt has massive scalaing problems because the cacheline gets so terribly contended. This uses a statically-allocated percpu variable. All want/drop operations are local to a cpu as long that cpu operates on the same mount, and there are no writer count imbalances. Writer count imbalances happen when a write is taken on one cpu, and released on another, like when an open/close pair is performed on two Upon a remount,ro request, all of the data from the percpu variables is collected (expensive, but very rare) and we determine if there are any outstanding writers to the mount. I've written a little benchmark to sit in a loop for a couple of seconds in several cpus in parallel doing open/write/close loops. http://sr71.net/~dave/linux/openbench.c The code in here is a a worst-possible case for this patch. It does opens on a _pair_ of files in two different mounts in parallel. This should cause my code to lose its "operate on the same mount" optimization completely. This worst-case scenario causes a 3% degredation in the benchmark. I could probably get rid of even this 3%, but it would be more complex than what I have here, and I think this is getting into acceptable territory. In practice, I expect writing more than 3 bytes to a file, as well as disk I/O to mask any effects that this has. (To get rid of that 3%, we could have an #defined number of mounts in the percpu variable. So, instead of a CPU getting operate only on percpu data when it accesses only one mount, it could stay on percpu data when it only accesses N or fewer mounts.) [AV] merged fix for __clear_mnt_mount() stepping on freed vfsmount Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Dave Hansen Signed-off-by: Andrew Morton Signed-off-by: Al Viro --- fs/namespace.c | 252 +++++++++++++++++++++++++++++++++++++++++++++++++++++---- 1 file changed, 237 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/namespace.c b/fs/namespace.c index 066b393578c..e3ce18d91aa 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -17,6 +17,7 @@ #include #include #include +#include #include #include #include @@ -55,6 +56,8 @@ static inline unsigned long hash(struct vfsmount *mnt, struct dentry *dentry) return tmp & (HASH_SIZE - 1); } +#define MNT_WRITER_UNDERFLOW_LIMIT -(1<<16) + struct vfsmount *alloc_vfsmnt(const char *name) { struct vfsmount *mnt = kmem_cache_zalloc(mnt_cache, GFP_KERNEL); @@ -68,6 +71,7 @@ struct vfsmount *alloc_vfsmnt(const char *name) INIT_LIST_HEAD(&mnt->mnt_share); INIT_LIST_HEAD(&mnt->mnt_slave_list); INIT_LIST_HEAD(&mnt->mnt_slave); + atomic_set(&mnt->__mnt_writers, 0); if (name) { int size = strlen(name) + 1; char *newname = kmalloc(size, GFP_KERNEL); @@ -80,6 +84,92 @@ struct vfsmount *alloc_vfsmnt(const char *name) return mnt; } +/* + * Most r/o checks on a fs are for operations that take + * discrete amounts of time, like a write() or unlink(). + * We must keep track of when those operations start + * (for permission checks) and when they end, so that + * we can determine when writes are able to occur to + * a filesystem. + */ +/* + * __mnt_is_readonly: check whether a mount is read-only + * @mnt: the mount to check for its write status + * + * This shouldn't be used directly ouside of the VFS. + * It does not guarantee that the filesystem will stay + * r/w, just that it is right *now*. This can not and + * should not be used in place of IS_RDONLY(inode). + * mnt_want/drop_write() will _keep_ the filesystem + * r/w. + */ +int __mnt_is_readonly(struct vfsmount *mnt) +{ + return (mnt->mnt_sb->s_flags & MS_RDONLY); +} +EXPORT_SYMBOL_GPL(__mnt_is_readonly); + +struct mnt_writer { + /* + * If holding multiple instances of this lock, they + * must be ordered by cpu number. + */ + spinlock_t lock; + struct lock_class_key lock_class; /* compiles out with !lockdep */ + unsigned long count; + struct vfsmount *mnt; +} ____cacheline_aligned_in_smp; +static DEFINE_PER_CPU(struct mnt_writer, mnt_writers); + +static int __init init_mnt_writers(void) +{ + int cpu; + for_each_possible_cpu(cpu) { + struct mnt_writer *writer = &per_cpu(mnt_writers, cpu); + spin_lock_init(&writer->lock); + lockdep_set_class(&writer->lock, &writer->lock_class); + writer->count = 0; + } + return 0; +} +fs_initcall(init_mnt_writers); + +static void unlock_mnt_writers(void) +{ + int cpu; + struct mnt_writer *cpu_writer; + + for_each_possible_cpu(cpu) { + cpu_writer = &per_cpu(mnt_writers, cpu); + spin_unlock(&cpu_writer->lock); + } +} + +static inline void __clear_mnt_count(struct mnt_writer *cpu_writer) +{ + if (!cpu_writer->mnt) + return; + /* + * This is in case anyone ever leaves an invalid, + * old ->mnt and a count of 0. + */ + if (!cpu_writer->count) + return; + atomic_add(cpu_writer->count, &cpu_writer->mnt->__mnt_writers); + cpu_writer->count = 0; +} + /* + * must hold cpu_writer->lock + */ +static inline void use_cpu_writer_for_mount(struct mnt_writer *cpu_writer, + struct vfsmount *mnt) +{ + if (cpu_writer->mnt == mnt) + return; + __clear_mnt_count(cpu_writer); + cpu_writer->mnt = mnt; +} + /* * Most r/o checks on a fs are for operations that take * discrete amounts of time, like a write() or unlink(). @@ -100,12 +190,77 @@ struct vfsmount *alloc_vfsmnt(const char *name) */ int mnt_want_write(struct vfsmount *mnt) { - if (__mnt_is_readonly(mnt)) - return -EROFS; - return 0; + int ret = 0; + struct mnt_writer *cpu_writer; + + cpu_writer = &get_cpu_var(mnt_writers); + spin_lock(&cpu_writer->lock); + if (__mnt_is_readonly(mnt)) { + ret = -EROFS; + goto out; + } + use_cpu_writer_for_mount(cpu_writer, mnt); + cpu_writer->count++; +out: + spin_unlock(&cpu_writer->lock); + put_cpu_var(mnt_writers); + return ret; } EXPORT_SYMBOL_GPL(mnt_want_write); +static void lock_mnt_writers(void) +{ + int cpu; + struct mnt_writer *cpu_writer; + + for_each_possible_cpu(cpu) { + cpu_writer = &per_cpu(mnt_writers, cpu); + spin_lock(&cpu_writer->lock); + __clear_mnt_count(cpu_writer); + cpu_writer->mnt = NULL; + } +} + +/* + * These per-cpu write counts are not guaranteed to have + * matched increments and decrements on any given cpu. + * A file open()ed for write on one cpu and close()d on + * another cpu will imbalance this count. Make sure it + * does not get too far out of whack. + */ +static void handle_write_count_underflow(struct vfsmount *mnt) +{ + if (atomic_read(&mnt->__mnt_writers) >= + MNT_WRITER_UNDERFLOW_LIMIT) + return; + /* + * It isn't necessary to hold all of the locks + * at the same time, but doing it this way makes + * us share a lot more code. + */ + lock_mnt_writers(); + /* + * vfsmount_lock is for mnt_flags. + */ + spin_lock(&vfsmount_lock); + /* + * If coalescing the per-cpu writer counts did not + * get us back to a positive writer count, we have + * a bug. + */ + if ((atomic_read(&mnt->__mnt_writers) < 0) && + !(mnt->mnt_flags & MNT_IMBALANCED_WRITE_COUNT)) { + printk(KERN_DEBUG "leak detected on mount(%p) writers " + "count: %d\n", + mnt, atomic_read(&mnt->__mnt_writers)); + WARN_ON(1); + /* use the flag to keep the dmesg spam down */ + mnt->mnt_flags |= MNT_IMBALANCED_WRITE_COUNT; + } + spin_unlock(&vfsmount_lock); + unlock_mnt_writers(); +} + /** * mnt_drop_write - give up write access to a mount * @mnt: the mount on which to give up write access @@ -116,23 +271,61 @@ EXPORT_SYMBOL_GPL(mnt_want_write); */ void mnt_drop_write(struct vfsmount *mnt) { + int must_check_underflow = 0; + struct mnt_writer *cpu_writer; + + cpu_writer = &get_cpu_var(mnt_writers); + spin_lock(&cpu_writer->lock); + + use_cpu_writer_for_mount(cpu_writer, mnt); + if (cpu_writer->count > 0) { + cpu_writer->count--; + } else { + must_check_underflow = 1; + atomic_dec(&mnt->__mnt_writers); + } + + spin_unlock(&cpu_writer->lock); + /* + * Logically, we could call this each time, + * but the __mnt_writers cacheline tends to + * be cold, and makes this expensive. + */ + if (must_check_underflow) + handle_write_count_underflow(mnt); + /* + * This could be done right after the spinlock + * is taken because the spinlock keeps us on + * the cpu, and disables preemption. However, + * putting it here bounds the amount that + * __mnt_writers can underflow. Without it, + * we could theoretically wrap __mnt_writers. + */ + put_cpu_var(mnt_writers); } EXPORT_SYMBOL_GPL(mnt_drop_write); -/* - * __mnt_is_readonly: check whether a mount is read-only - * @mnt: the mount to check for its write status - * - * This shouldn't be used directly ouside of the VFS. - * It does not guarantee that the filesystem will stay - * r/w, just that it is right *now*. This can not and - * should not be used in place of IS_RDONLY(inode). - */ -int __mnt_is_readonly(struct vfsmount *mnt) +int mnt_make_readonly(struct vfsmount *mnt) { - return (mnt->mnt_sb->s_flags & MS_RDONLY); + int ret = 0; + + lock_mnt_writers(); + /* + * With all the locks held, this value is stable + */ + if (atomic_read(&mnt->__mnt_writers) > 0) { + ret = -EBUSY; + goto out; + } + /* + * actually set mount's r/o flag here to make + * __mnt_is_readonly() true, which keeps anyone + * from doing a successful mnt_want_write(). + */ +out: + unlock_mnt_writers(); + return ret; } -EXPORT_SYMBOL_GPL(__mnt_is_readonly); int simple_set_mnt(struct vfsmount *mnt, struct super_block *sb) { @@ -325,7 +518,36 @@ static struct vfsmount *clone_mnt(struct vfsmount *old, struct dentry *root, static inline void __mntput(struct vfsmount *mnt) { + int cpu; struct super_block *sb = mnt->mnt_sb; + /* + * We don't have to hold all of the locks at the + * same time here because we know that we're the + * last reference to mnt and that no new writers + * can come in. + */ + for_each_possible_cpu(cpu) { + struct mnt_writer *cpu_writer = &per_cpu(mnt_writers, cpu); + if (cpu_writer->mnt != mnt) + continue; + spin_lock(&cpu_writer->lock); + atomic_add(cpu_writer->count, &mnt->__mnt_writers); + cpu_writer->count = 0; + /* + * Might as well do this so that no one + * ever sees the pointer and expects + * it to be valid. + */ + cpu_writer->mnt = NULL; + spin_unlock(&cpu_writer->lock); + } + /* + * This probably indicates that somebody messed + * up a mnt_want/drop_write() pair. If this + * happens, the filesystem was probably unable + * to make r/w->r/o transitions. + */ + WARN_ON(atomic_read(&mnt->__mnt_writers)); dput(mnt->mnt_root); free_vfsmnt(mnt); deactivate_super(sb); -- cgit v1.2.3 From 2e4b7fcd926006531935a4c79a5e9349fe51125b Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:38:00 -0800 Subject: [PATCH] r/o bind mounts: honor mount writer counts at remount Originally from: Herbert Poetzl This is the core of the read-only bind mount patch set. Note that this does _not_ add a "ro" option directly to the bind mount operation. If you require such a mount, you must first do the bind, then follow it up with a 'mount -o remount,ro' operation: If you wish to have a r/o bind mount of /foo on bar: mount --bind /foo /bar mount -o remount,ro /bar Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Dave Hansen Signed-off-by: Andrew Morton Signed-off-by: Al Viro --- fs/namespace.c | 50 +++++++++++++++++++++++++++++++++++++++++++------- 1 file changed, 43 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/namespace.c b/fs/namespace.c index e3ce18d91aa..678f7ce060f 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -105,7 +105,11 @@ struct vfsmount *alloc_vfsmnt(const char *name) */ int __mnt_is_readonly(struct vfsmount *mnt) { - return (mnt->mnt_sb->s_flags & MS_RDONLY); + if (mnt->mnt_flags & MNT_READONLY) + return 1; + if (mnt->mnt_sb->s_flags & MS_RDONLY) + return 1; + return 0; } EXPORT_SYMBOL_GPL(__mnt_is_readonly); @@ -305,7 +309,7 @@ void mnt_drop_write(struct vfsmount *mnt) } EXPORT_SYMBOL_GPL(mnt_drop_write); -int mnt_make_readonly(struct vfsmount *mnt) +static int mnt_make_readonly(struct vfsmount *mnt) { int ret = 0; @@ -318,15 +322,25 @@ int mnt_make_readonly(struct vfsmount *mnt) goto out; } /* - * actually set mount's r/o flag here to make - * __mnt_is_readonly() true, which keeps anyone - * from doing a successful mnt_want_write(). + * nobody can do a successful mnt_want_write() with all + * of the counts in MNT_DENIED_WRITE and the locks held. */ + spin_lock(&vfsmount_lock); + if (!ret) + mnt->mnt_flags |= MNT_READONLY; + spin_unlock(&vfsmount_lock); out: unlock_mnt_writers(); return ret; } +static void __mnt_unmake_readonly(struct vfsmount *mnt) +{ + spin_lock(&vfsmount_lock); + mnt->mnt_flags &= ~MNT_READONLY; + spin_unlock(&vfsmount_lock); +} + int simple_set_mnt(struct vfsmount *mnt, struct super_block *sb) { mnt->mnt_sb = sb; @@ -693,7 +707,7 @@ static int show_vfsmnt(struct seq_file *m, void *v) seq_putc(m, '.'); mangle(m, mnt->mnt_sb->s_subtype); } - seq_puts(m, mnt->mnt_sb->s_flags & MS_RDONLY ? " ro" : " rw"); + seq_puts(m, __mnt_is_readonly(mnt) ? " ro" : " rw"); for (fs_infop = fs_info; fs_infop->flag; fs_infop++) { if (mnt->mnt_sb->s_flags & fs_infop->flag) seq_puts(m, fs_infop->str); @@ -1295,6 +1309,23 @@ out: return err; } +static int change_mount_flags(struct vfsmount *mnt, int ms_flags) +{ + int error = 0; + int readonly_request = 0; + + if (ms_flags & MS_RDONLY) + readonly_request = 1; + if (readonly_request == __mnt_is_readonly(mnt)) + return 0; + + if (readonly_request) + error = mnt_make_readonly(mnt); + else + __mnt_unmake_readonly(mnt); + return error; +} + /* * change filesystem flags. dir should be a physical root of filesystem. * If you've mounted a non-root directory somewhere and want to do remount @@ -1317,7 +1348,10 @@ static noinline int do_remount(struct nameidata *nd, int flags, int mnt_flags, return -EINVAL; down_write(&sb->s_umount); - err = do_remount_sb(sb, flags, data, 0); + if (flags & MS_BIND) + err = change_mount_flags(nd->path.mnt, flags); + else + err = do_remount_sb(sb, flags, data, 0); if (!err) nd->path.mnt->mnt_flags = mnt_flags; up_write(&sb->s_umount); @@ -1701,6 +1735,8 @@ long do_mount(char *dev_name, char *dir_name, char *type_page, mnt_flags |= MNT_NODIRATIME; if (flags & MS_RELATIME) mnt_flags |= MNT_RELATIME; + if (flags & MS_RDONLY) + mnt_flags |= MNT_READONLY; flags &= ~(MS_NOSUID | MS_NOEXEC | MS_NODEV | MS_ACTIVE | MS_NOATIME | MS_NODIRATIME | MS_RELATIME| MS_KERNMOUNT); -- cgit v1.2.3 From ad775f5a8faa5845377f093ca11caf577404add9 Mon Sep 17 00:00:00 2001 From: Dave Hansen Date: Fri, 15 Feb 2008 14:38:01 -0800 Subject: [PATCH] r/o bind mounts: debugging for missed calls There have been a few oopses caused by 'struct file's with NULL f_vfsmnts. There was also a set of potentially missed mnt_want_write()s from dentry_open() calls. This patch provides a very simple debugging framework to catch these kinds of bugs. It will WARN_ON() them, but should stop us from having any oopses or mnt_writer count imbalances. I'm quite convinced that this is a good thing because it found bugs in the stuff I was working on as soon as I wrote it. [hch: made it conditional on a debug option. But it's still a little bit too ugly] [hch: merged forced remount r/o fix from Dave and akpm's fix for the fix] Signed-off-by: Dave Hansen Acked-by: Al Viro Signed-off-by: Christoph Hellwig Signed-off-by: Andrew Morton Signed-off-by: Al Viro --- fs/file_table.c | 11 +++++++++-- fs/open.c | 12 +++++++++++- fs/super.c | 3 +++ 3 files changed, 23 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/file_table.c b/fs/file_table.c index 71efc700022..7a0a9b87225 100644 --- a/fs/file_table.c +++ b/fs/file_table.c @@ -42,6 +42,7 @@ static inline void file_free_rcu(struct rcu_head *head) static inline void file_free(struct file *f) { percpu_counter_dec(&nr_files); + file_check_state(f); call_rcu(&f->f_u.fu_rcuhead, file_free_rcu); } @@ -207,6 +208,7 @@ int init_file(struct file *file, struct vfsmount *mnt, struct dentry *dentry, * that we can do debugging checks at __fput() */ if ((mode & FMODE_WRITE) && !special_file(dentry->d_inode->i_mode)) { + file_take_write(file); error = mnt_want_write(mnt); WARN_ON(error); } @@ -237,8 +239,13 @@ void drop_file_write_access(struct file *file) struct inode *inode = dentry->d_inode; put_write_access(inode); - if (!special_file(inode->i_mode)) - mnt_drop_write(mnt); + + if (special_file(inode->i_mode)) + return; + if (file_check_writeable(file) != 0) + return; + mnt_drop_write(mnt); + file_release_write(file); } EXPORT_SYMBOL_GPL(drop_file_write_access); diff --git a/fs/open.c b/fs/open.c index e58382d57e7..b70e7666bb2 100644 --- a/fs/open.c +++ b/fs/open.c @@ -806,6 +806,8 @@ static struct file *__dentry_open(struct dentry *dentry, struct vfsmount *mnt, error = __get_file_write_access(inode, mnt); if (error) goto cleanup_file; + if (!special_file(inode->i_mode)) + file_take_write(f); } f->f_mapping = inode->i_mapping; @@ -847,8 +849,16 @@ cleanup_all: fops_put(f->f_op); if (f->f_mode & FMODE_WRITE) { put_write_access(inode); - if (!special_file(inode->i_mode)) + if (!special_file(inode->i_mode)) { + /* + * We don't consider this a real + * mnt_want/drop_write() pair + * because it all happenend right + * here, so just reset the state. + */ + file_reset_write(f); mnt_drop_write(mnt); + } } file_kill(f); f->f_path.dentry = NULL; diff --git a/fs/super.c b/fs/super.c index 01d5c40e911..1f8f05ede43 100644 --- a/fs/super.c +++ b/fs/super.c @@ -579,6 +579,9 @@ retry: if (!(f->f_mode & FMODE_WRITE)) continue; f->f_mode &= ~FMODE_WRITE; + if (file_check_writeable(f) != 0) + continue; + file_release_write(f); mnt = mntget(f->f_path.mnt); file_list_unlock(); /* -- cgit v1.2.3 From 73e3302f60c0e11a0db0b34b903f591139c4f937 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 11 Apr 2008 16:03:54 -0400 Subject: NFS: Fix nfs_wb_page() to always exit with an error or a clean page It is possible for nfs_wb_page() to sometimes exit with 0 return value, yet the page is left in a dirty state. For instance in the case where the server rebooted, and the COMMIT request failed, then all the previously "clean" pages which were cached by the server, but were not guaranteed to have been writted out to disk, have to be redirtied and resent to the server. The fix is to have nfs_wb_page_priority() check that the page is clean before it exits... This fixes a condition that triggers the BUG_ON(PagePrivate(page)) in nfs_create_request() when we're in the nfs_readpage() path. Also eliminate a redundant BUG_ON(!PageLocked(page)) while we're at it. It turns out that clear_page_dirty_for_io() has the exact same test. Signed-off-by: Trond Myklebust --- fs/nfs/write.c | 23 ++++++++++++----------- 1 file changed, 12 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/nfs/write.c b/fs/nfs/write.c index ce40cadb15d..997b42aa370 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -1493,18 +1493,19 @@ static int nfs_wb_page_priority(struct inode *inode, struct page *page, }; int ret; - BUG_ON(!PageLocked(page)); - if (clear_page_dirty_for_io(page)) { - ret = nfs_writepage_locked(page, &wbc); + do { + if (clear_page_dirty_for_io(page)) { + ret = nfs_writepage_locked(page, &wbc); + if (ret < 0) + goto out_error; + } else if (!PagePrivate(page)) + break; + ret = nfs_sync_mapping_wait(page->mapping, &wbc, how); if (ret < 0) - goto out; - } - if (!PagePrivate(page)) - return 0; - ret = nfs_sync_mapping_wait(page->mapping, &wbc, how); - if (ret >= 0) - return 0; -out: + goto out_error; + } while (PagePrivate(page)); + return 0; +out_error: __mark_inode_dirty(inode, I_DIRTY_PAGES); return ret; } -- cgit v1.2.3 From fdd1e74c89fe39259a29c494209abad63ff76f82 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 15 Apr 2008 16:33:58 -0400 Subject: NFS: Ensure that the read code cleans up properly when rpc_run_task() fails In the case of readpage() we need to ensure that the pages get unlocked, and that the error is flagged. In the case of O_DIRECT, we need to ensure that the pages are all released. Signed-off-by: Trond Myklebust --- fs/nfs/direct.c | 19 +++++++++++++------ fs/nfs/read.c | 53 +++++++++++++++++++++++++++++++++-------------------- 2 files changed, 46 insertions(+), 26 deletions(-) (limited to 'fs') diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c index e44200579c8..9d9085b93a3 100644 --- a/fs/nfs/direct.c +++ b/fs/nfs/direct.c @@ -229,14 +229,20 @@ static void nfs_direct_complete(struct nfs_direct_req *dreq) static void nfs_direct_read_result(struct rpc_task *task, void *calldata) { struct nfs_read_data *data = calldata; - struct nfs_direct_req *dreq = (struct nfs_direct_req *) data->req; - if (nfs_readpage_result(task, data) != 0) - return; + nfs_readpage_result(task, data); +} + +static void nfs_direct_read_release(void *calldata) +{ + + struct nfs_read_data *data = calldata; + struct nfs_direct_req *dreq = (struct nfs_direct_req *) data->req; + int status = data->task.tk_status; spin_lock(&dreq->lock); - if (unlikely(task->tk_status < 0)) { - dreq->error = task->tk_status; + if (unlikely(status < 0)) { + dreq->error = status; spin_unlock(&dreq->lock); } else { dreq->count += data->res.count; @@ -249,11 +255,12 @@ static void nfs_direct_read_result(struct rpc_task *task, void *calldata) if (put_dreq(dreq)) nfs_direct_complete(dreq); + nfs_readdata_release(calldata); } static const struct rpc_call_ops nfs_read_direct_ops = { .rpc_call_done = nfs_direct_read_result, - .rpc_release = nfs_readdata_release, + .rpc_release = nfs_direct_read_release, }; /* diff --git a/fs/nfs/read.c b/fs/nfs/read.c index d333f5fedca..6f9208a549a 100644 --- a/fs/nfs/read.c +++ b/fs/nfs/read.c @@ -338,26 +338,25 @@ int nfs_readpage_result(struct rpc_task *task, struct nfs_read_data *data) return 0; } -static int nfs_readpage_retry(struct rpc_task *task, struct nfs_read_data *data) +static void nfs_readpage_retry(struct rpc_task *task, struct nfs_read_data *data) { struct nfs_readargs *argp = &data->args; struct nfs_readres *resp = &data->res; if (resp->eof || resp->count == argp->count) - return 0; + return; /* This is a short read! */ nfs_inc_stats(data->inode, NFSIOS_SHORTREAD); /* Has the server at least made some progress? */ if (resp->count == 0) - return 0; + return; /* Yes, so retry the read at the end of the data */ argp->offset += resp->count; argp->pgbase += resp->count; argp->count -= resp->count; rpc_restart_call(task); - return -EAGAIN; } /* @@ -366,29 +365,37 @@ static int nfs_readpage_retry(struct rpc_task *task, struct nfs_read_data *data) static void nfs_readpage_result_partial(struct rpc_task *task, void *calldata) { struct nfs_read_data *data = calldata; - struct nfs_page *req = data->req; - struct page *page = req->wb_page; if (nfs_readpage_result(task, data) != 0) return; + if (task->tk_status < 0) + return; - if (likely(task->tk_status >= 0)) { - nfs_readpage_truncate_uninitialised_page(data); - if (nfs_readpage_retry(task, data) != 0) - return; - } - if (unlikely(task->tk_status < 0)) + nfs_readpage_truncate_uninitialised_page(data); + nfs_readpage_retry(task, data); +} + +static void nfs_readpage_release_partial(void *calldata) +{ + struct nfs_read_data *data = calldata; + struct nfs_page *req = data->req; + struct page *page = req->wb_page; + int status = data->task.tk_status; + + if (status < 0) SetPageError(page); + if (atomic_dec_and_test(&req->wb_complete)) { if (!PageError(page)) SetPageUptodate(page); nfs_readpage_release(req); } + nfs_readdata_release(calldata); } static const struct rpc_call_ops nfs_read_partial_ops = { .rpc_call_done = nfs_readpage_result_partial, - .rpc_release = nfs_readdata_release, + .rpc_release = nfs_readpage_release_partial, }; static void nfs_readpage_set_pages_uptodate(struct nfs_read_data *data) @@ -423,29 +430,35 @@ static void nfs_readpage_result_full(struct rpc_task *task, void *calldata) if (nfs_readpage_result(task, data) != 0) return; + if (task->tk_status < 0) + return; /* * Note: nfs_readpage_retry may change the values of * data->args. In the multi-page case, we therefore need * to ensure that we call nfs_readpage_set_pages_uptodate() * first. */ - if (likely(task->tk_status >= 0)) { - nfs_readpage_truncate_uninitialised_page(data); - nfs_readpage_set_pages_uptodate(data); - if (nfs_readpage_retry(task, data) != 0) - return; - } + nfs_readpage_truncate_uninitialised_page(data); + nfs_readpage_set_pages_uptodate(data); + nfs_readpage_retry(task, data); +} + +static void nfs_readpage_release_full(void *calldata) +{ + struct nfs_read_data *data = calldata; + while (!list_empty(&data->pages)) { struct nfs_page *req = nfs_list_entry(data->pages.next); nfs_list_remove_request(req); nfs_readpage_release(req); } + nfs_readdata_release(calldata); } static const struct rpc_call_ops nfs_read_full_ops = { .rpc_call_done = nfs_readpage_result_full, - .rpc_release = nfs_readdata_release, + .rpc_release = nfs_readpage_release_full, }; /* -- cgit v1.2.3 From c9d8f89d9816c1d16ada492aa547a4d692508c0d Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 15 Apr 2008 16:56:39 -0400 Subject: NFS: Ensure that the write code cleans up properly when rpc_run_task() fails Signed-off-by: Trond Myklebust --- fs/nfs/direct.c | 50 +++++++++++++++++++++++------------------- fs/nfs/write.c | 67 ++++++++++++++++++++++++++++++++++++--------------------- 2 files changed, 71 insertions(+), 46 deletions(-) (limited to 'fs') diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c index 9d9085b93a3..abf8e0286e3 100644 --- a/fs/nfs/direct.c +++ b/fs/nfs/direct.c @@ -508,27 +508,34 @@ static void nfs_direct_write_reschedule(struct nfs_direct_req *dreq) static void nfs_direct_commit_result(struct rpc_task *task, void *calldata) { struct nfs_write_data *data = calldata; - struct nfs_direct_req *dreq = (struct nfs_direct_req *) data->req; /* Call the NFS version-specific code */ - if (NFS_PROTO(data->inode)->commit_done(task, data) != 0) - return; - if (unlikely(task->tk_status < 0)) { + NFS_PROTO(data->inode)->commit_done(task, data); +} + +static void nfs_direct_commit_release(void *calldata) +{ + struct nfs_write_data *data = calldata; + struct nfs_direct_req *dreq = (struct nfs_direct_req *) data->req; + int status = data->task.tk_status; + + if (status < 0) { dprintk("NFS: %5u commit failed with error %d.\n", - task->tk_pid, task->tk_status); + data->task.tk_pid, status); dreq->flags = NFS_ODIRECT_RESCHED_WRITES; } else if (memcmp(&dreq->verf, &data->verf, sizeof(data->verf))) { - dprintk("NFS: %5u commit verify failed\n", task->tk_pid); + dprintk("NFS: %5u commit verify failed\n", data->task.tk_pid); dreq->flags = NFS_ODIRECT_RESCHED_WRITES; } - dprintk("NFS: %5u commit returned %d\n", task->tk_pid, task->tk_status); + dprintk("NFS: %5u commit returned %d\n", data->task.tk_pid, status); nfs_direct_write_complete(dreq, data->inode); + nfs_commitdata_release(calldata); } static const struct rpc_call_ops nfs_commit_direct_ops = { .rpc_call_done = nfs_direct_commit_result, - .rpc_release = nfs_commit_release, + .rpc_release = nfs_direct_commit_release, }; static void nfs_direct_commit_schedule(struct nfs_direct_req *dreq) @@ -596,7 +603,7 @@ static void nfs_direct_write_complete(struct nfs_direct_req *dreq, struct inode static void nfs_alloc_commit_data(struct nfs_direct_req *dreq) { - dreq->commit_data = nfs_commit_alloc(); + dreq->commit_data = nfs_commitdata_alloc(); if (dreq->commit_data != NULL) dreq->commit_data->req = (struct nfs_page *) dreq; } @@ -617,11 +624,20 @@ static void nfs_direct_write_complete(struct nfs_direct_req *dreq, struct inode static void nfs_direct_write_result(struct rpc_task *task, void *calldata) { struct nfs_write_data *data = calldata; - struct nfs_direct_req *dreq = (struct nfs_direct_req *) data->req; - int status = task->tk_status; if (nfs_writeback_done(task, data) != 0) return; +} + +/* + * NB: Return the value of the first error return code. Subsequent + * errors after the first one are ignored. + */ +static void nfs_direct_write_release(void *calldata) +{ + struct nfs_write_data *data = calldata; + struct nfs_direct_req *dreq = (struct nfs_direct_req *) data->req; + int status = data->task.tk_status; spin_lock(&dreq->lock); @@ -643,23 +659,13 @@ static void nfs_direct_write_result(struct rpc_task *task, void *calldata) break; case NFS_ODIRECT_DO_COMMIT: if (memcmp(&dreq->verf, &data->verf, sizeof(dreq->verf))) { - dprintk("NFS: %5u write verify failed\n", task->tk_pid); + dprintk("NFS: %5u write verify failed\n", data->task.tk_pid); dreq->flags = NFS_ODIRECT_RESCHED_WRITES; } } } out_unlock: spin_unlock(&dreq->lock); -} - -/* - * NB: Return the value of the first error return code. Subsequent - * errors after the first one are ignored. - */ -static void nfs_direct_write_release(void *calldata) -{ - struct nfs_write_data *data = calldata; - struct nfs_direct_req *dreq = (struct nfs_direct_req *) data->req; if (put_dreq(dreq)) nfs_direct_write_complete(dreq, data->inode); diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 997b42aa370..31681bb5e4c 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -48,7 +48,7 @@ static struct kmem_cache *nfs_wdata_cachep; static mempool_t *nfs_wdata_mempool; static mempool_t *nfs_commit_mempool; -struct nfs_write_data *nfs_commit_alloc(void) +struct nfs_write_data *nfs_commitdata_alloc(void) { struct nfs_write_data *p = mempool_alloc(nfs_commit_mempool, GFP_NOFS); @@ -973,7 +973,6 @@ static void nfs_writeback_done_partial(struct rpc_task *task, void *calldata) { struct nfs_write_data *data = calldata; struct nfs_page *req = data->req; - struct page *page = req->wb_page; dprintk("NFS: write (%s/%Ld %d@%Ld)", req->wb_context->path.dentry->d_inode->i_sb->s_id, @@ -981,13 +980,20 @@ static void nfs_writeback_done_partial(struct rpc_task *task, void *calldata) req->wb_bytes, (long long)req_offset(req)); - if (nfs_writeback_done(task, data) != 0) - return; + nfs_writeback_done(task, data); +} - if (task->tk_status < 0) { +static void nfs_writeback_release_partial(void *calldata) +{ + struct nfs_write_data *data = calldata; + struct nfs_page *req = data->req; + struct page *page = req->wb_page; + int status = data->task.tk_status; + + if (status < 0) { nfs_set_pageerror(page); - nfs_context_set_write_error(req->wb_context, task->tk_status); - dprintk(", error = %d\n", task->tk_status); + nfs_context_set_write_error(req->wb_context, status); + dprintk(", error = %d\n", status); goto out; } @@ -1012,11 +1018,12 @@ static void nfs_writeback_done_partial(struct rpc_task *task, void *calldata) out: if (atomic_dec_and_test(&req->wb_complete)) nfs_writepage_release(req); + nfs_writedata_release(calldata); } static const struct rpc_call_ops nfs_write_partial_ops = { .rpc_call_done = nfs_writeback_done_partial, - .rpc_release = nfs_writedata_release, + .rpc_release = nfs_writeback_release_partial, }; /* @@ -1029,17 +1036,21 @@ static const struct rpc_call_ops nfs_write_partial_ops = { static void nfs_writeback_done_full(struct rpc_task *task, void *calldata) { struct nfs_write_data *data = calldata; - struct nfs_page *req; - struct page *page; - if (nfs_writeback_done(task, data) != 0) - return; + nfs_writeback_done(task, data); +} + +static void nfs_writeback_release_full(void *calldata) +{ + struct nfs_write_data *data = calldata; + int status = data->task.tk_status; /* Update attributes as result of writeback. */ while (!list_empty(&data->pages)) { - req = nfs_list_entry(data->pages.next); + struct nfs_page *req = nfs_list_entry(data->pages.next); + struct page *page = req->wb_page; + nfs_list_remove_request(req); - page = req->wb_page; dprintk("NFS: write (%s/%Ld %d@%Ld)", req->wb_context->path.dentry->d_inode->i_sb->s_id, @@ -1047,10 +1058,10 @@ static void nfs_writeback_done_full(struct rpc_task *task, void *calldata) req->wb_bytes, (long long)req_offset(req)); - if (task->tk_status < 0) { + if (status < 0) { nfs_set_pageerror(page); - nfs_context_set_write_error(req->wb_context, task->tk_status); - dprintk(", error = %d\n", task->tk_status); + nfs_context_set_write_error(req->wb_context, status); + dprintk(", error = %d\n", status); goto remove_request; } @@ -1070,11 +1081,12 @@ remove_request: next: nfs_clear_page_tag_locked(req); } + nfs_writedata_release(calldata); } static const struct rpc_call_ops nfs_write_full_ops = { .rpc_call_done = nfs_writeback_done_full, - .rpc_release = nfs_writedata_release, + .rpc_release = nfs_writeback_release_full, }; @@ -1160,7 +1172,7 @@ int nfs_writeback_done(struct rpc_task *task, struct nfs_write_data *data) #if defined(CONFIG_NFS_V3) || defined(CONFIG_NFS_V4) -void nfs_commit_release(void *data) +void nfs_commitdata_release(void *data) { struct nfs_write_data *wdata = data; @@ -1233,7 +1245,7 @@ nfs_commit_list(struct inode *inode, struct list_head *head, int how) struct nfs_write_data *data; struct nfs_page *req; - data = nfs_commit_alloc(); + data = nfs_commitdata_alloc(); if (!data) goto out_bad; @@ -1261,7 +1273,6 @@ nfs_commit_list(struct inode *inode, struct list_head *head, int how) static void nfs_commit_done(struct rpc_task *task, void *calldata) { struct nfs_write_data *data = calldata; - struct nfs_page *req; dprintk("NFS: %5u nfs_commit_done (status %d)\n", task->tk_pid, task->tk_status); @@ -1269,6 +1280,13 @@ static void nfs_commit_done(struct rpc_task *task, void *calldata) /* Call the NFS version-specific code */ if (NFS_PROTO(data->inode)->commit_done(task, data) != 0) return; +} + +static void nfs_commit_release(void *calldata) +{ + struct nfs_write_data *data = calldata; + struct nfs_page *req; + int status = data->task.tk_status; while (!list_empty(&data->pages)) { req = nfs_list_entry(data->pages.next); @@ -1283,10 +1301,10 @@ static void nfs_commit_done(struct rpc_task *task, void *calldata) (long long)NFS_FILEID(req->wb_context->path.dentry->d_inode), req->wb_bytes, (long long)req_offset(req)); - if (task->tk_status < 0) { - nfs_context_set_write_error(req->wb_context, task->tk_status); + if (status < 0) { + nfs_context_set_write_error(req->wb_context, status); nfs_inode_remove_request(req); - dprintk(", error = %d\n", task->tk_status); + dprintk(", error = %d\n", status); goto next; } @@ -1307,6 +1325,7 @@ static void nfs_commit_done(struct rpc_task *task, void *calldata) next: nfs_clear_page_tag_locked(req); } + nfs_commitdata_release(calldata); } static const struct rpc_call_ops nfs_commit_ops = { -- cgit v1.2.3 From dbae4c73f08b8a7980cc912954ade3d4c1fb6147 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 14 Apr 2008 14:54:53 -0400 Subject: NFS: Ensure that rpc_run_task() errors are propagated back to the caller Signed-off-by: Trond Myklebust --- fs/nfs/direct.c | 10 ++++++---- fs/nfs/read.c | 23 +++++++++++++++-------- fs/nfs/write.c | 33 +++++++++++++++++++-------------- 3 files changed, 40 insertions(+), 26 deletions(-) (limited to 'fs') diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c index abf8e0286e3..4757a2b326a 100644 --- a/fs/nfs/direct.c +++ b/fs/nfs/direct.c @@ -347,8 +347,9 @@ static ssize_t nfs_direct_read_schedule_segment(struct nfs_direct_req *dreq, NFS_PROTO(inode)->read_setup(data, &msg); task = rpc_run_task(&task_setup_data); - if (!IS_ERR(task)) - rpc_put_task(task); + if (IS_ERR(task)) + break; + rpc_put_task(task); dprintk("NFS: %5u initiated direct read call " "(req %s/%Ld, %zu bytes @ offset %Lu)\n", @@ -763,8 +764,9 @@ static ssize_t nfs_direct_write_schedule_segment(struct nfs_direct_req *dreq, NFS_PROTO(inode)->write_setup(data, &msg); task = rpc_run_task(&task_setup_data); - if (!IS_ERR(task)) - rpc_put_task(task); + if (IS_ERR(task)) + break; + rpc_put_task(task); dprintk("NFS: %5u initiated direct write call " "(req %s/%Ld, %zu bytes @ offset %Lu)\n", diff --git a/fs/nfs/read.c b/fs/nfs/read.c index 6f9208a549a..16f57e0af99 100644 --- a/fs/nfs/read.c +++ b/fs/nfs/read.c @@ -153,7 +153,7 @@ static void nfs_readpage_release(struct nfs_page *req) /* * Set up the NFS read request struct */ -static void nfs_read_rpcsetup(struct nfs_page *req, struct nfs_read_data *data, +static int nfs_read_rpcsetup(struct nfs_page *req, struct nfs_read_data *data, const struct rpc_call_ops *call_ops, unsigned int count, unsigned int offset) { @@ -202,8 +202,10 @@ static void nfs_read_rpcsetup(struct nfs_page *req, struct nfs_read_data *data, (unsigned long long)data->args.offset); task = rpc_run_task(&task_setup_data); - if (!IS_ERR(task)) - rpc_put_task(task); + if (IS_ERR(task)) + return PTR_ERR(task); + rpc_put_task(task); + return 0; } static void @@ -240,6 +242,7 @@ static int nfs_pagein_multi(struct inode *inode, struct list_head *head, unsigne size_t rsize = NFS_SERVER(inode)->rsize, nbytes; unsigned int offset; int requests = 0; + int ret = 0; LIST_HEAD(list); nfs_list_remove_request(req); @@ -261,6 +264,8 @@ static int nfs_pagein_multi(struct inode *inode, struct list_head *head, unsigne offset = 0; nbytes = count; do { + int ret2; + data = list_entry(list.next, struct nfs_read_data, pages); list_del_init(&data->pages); @@ -268,13 +273,15 @@ static int nfs_pagein_multi(struct inode *inode, struct list_head *head, unsigne if (nbytes < rsize) rsize = nbytes; - nfs_read_rpcsetup(req, data, &nfs_read_partial_ops, + ret2 = nfs_read_rpcsetup(req, data, &nfs_read_partial_ops, rsize, offset); + if (ret == 0) + ret = ret2; offset += rsize; nbytes -= rsize; } while (nbytes != 0); - return 0; + return ret; out_bad: while (!list_empty(&list)) { @@ -292,6 +299,7 @@ static int nfs_pagein_one(struct inode *inode, struct list_head *head, unsigned struct nfs_page *req; struct page **pages; struct nfs_read_data *data; + int ret = -ENOMEM; data = nfs_readdata_alloc(npages); if (!data) @@ -307,11 +315,10 @@ static int nfs_pagein_one(struct inode *inode, struct list_head *head, unsigned } req = nfs_list_entry(data->pages.next); - nfs_read_rpcsetup(req, data, &nfs_read_full_ops, count, 0); - return 0; + return nfs_read_rpcsetup(req, data, &nfs_read_full_ops, count, 0); out_bad: nfs_async_read_error(head); - return -ENOMEM; + return ret; } /* diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 31681bb5e4c..1ade11d1ba0 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -778,7 +778,7 @@ static int flush_task_priority(int how) /* * Set up the argument/result storage required for the RPC call. */ -static void nfs_write_rpcsetup(struct nfs_page *req, +static int nfs_write_rpcsetup(struct nfs_page *req, struct nfs_write_data *data, const struct rpc_call_ops *call_ops, unsigned int count, unsigned int offset, @@ -841,8 +841,10 @@ static void nfs_write_rpcsetup(struct nfs_page *req, (unsigned long long)data->args.offset); task = rpc_run_task(&task_setup_data); - if (!IS_ERR(task)) - rpc_put_task(task); + if (IS_ERR(task)) + return PTR_ERR(task); + rpc_put_task(task); + return 0; } /* If a nfs_flush_* function fails, it should remove reqs from @head and @@ -868,6 +870,7 @@ static int nfs_flush_multi(struct inode *inode, struct list_head *head, unsigned size_t wsize = NFS_SERVER(inode)->wsize, nbytes; unsigned int offset; int requests = 0; + int ret = 0; LIST_HEAD(list); nfs_list_remove_request(req); @@ -889,6 +892,8 @@ static int nfs_flush_multi(struct inode *inode, struct list_head *head, unsigned offset = 0; nbytes = count; do { + int ret2; + data = list_entry(list.next, struct nfs_write_data, pages); list_del_init(&data->pages); @@ -896,13 +901,15 @@ static int nfs_flush_multi(struct inode *inode, struct list_head *head, unsigned if (nbytes < wsize) wsize = nbytes; - nfs_write_rpcsetup(req, data, &nfs_write_partial_ops, + ret2 = nfs_write_rpcsetup(req, data, &nfs_write_partial_ops, wsize, offset, how); + if (ret == 0) + ret = ret2; offset += wsize; nbytes -= wsize; } while (nbytes != 0); - return 0; + return ret; out_bad: while (!list_empty(&list)) { @@ -943,9 +950,7 @@ static int nfs_flush_one(struct inode *inode, struct list_head *head, unsigned i req = nfs_list_entry(data->pages.next); /* Set up the argument struct */ - nfs_write_rpcsetup(req, data, &nfs_write_full_ops, count, 0, how); - - return 0; + return nfs_write_rpcsetup(req, data, &nfs_write_full_ops, count, 0, how); out_bad: while (!list_empty(head)) { req = nfs_list_entry(head->next); @@ -1183,7 +1188,7 @@ void nfs_commitdata_release(void *data) /* * Set up the argument/result storage required for the RPC call. */ -static void nfs_commit_rpcsetup(struct list_head *head, +static int nfs_commit_rpcsetup(struct list_head *head, struct nfs_write_data *data, int how) { @@ -1232,8 +1237,10 @@ static void nfs_commit_rpcsetup(struct list_head *head, dprintk("NFS: %5u initiated commit call\n", data->task.tk_pid); task = rpc_run_task(&task_setup_data); - if (!IS_ERR(task)) - rpc_put_task(task); + if (IS_ERR(task)) + return PTR_ERR(task); + rpc_put_task(task); + return 0; } /* @@ -1251,9 +1258,7 @@ nfs_commit_list(struct inode *inode, struct list_head *head, int how) goto out_bad; /* Set up the argument struct */ - nfs_commit_rpcsetup(head, data, how); - - return 0; + return nfs_commit_rpcsetup(head, data, how); out_bad: while (!list_empty(head)) { req = nfs_list_entry(head->next); -- cgit v1.2.3 From 35d05778e25ee16dbddb60331be0bc1309efba19 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Sat, 5 Apr 2008 15:54:17 -0400 Subject: NFSv4: Remove bogus call to nfs4_drop_state_owner() in _nfs4_open_expired() There should be no need to invalidate a perfectly good state owner just because of a stale filehandle. Doing so can cause the state recovery code to break, since nfs4_get_renew_cred() and nfs4_get_setclientid_cred() rely on finding active state owners. Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index f38d0573be1..424aa206470 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -982,11 +982,8 @@ static int _nfs4_open_expired(struct nfs_open_context *ctx, struct nfs4_state *s if (IS_ERR(opendata)) return PTR_ERR(opendata); ret = nfs4_open_recover(opendata, state); - if (ret == -ESTALE) { - /* Invalidate the state owner so we don't ever use it again */ - nfs4_drop_state_owner(state->owner); + if (ret == -ESTALE) d_drop(ctx->path.dentry); - } nfs4_opendata_put(opendata); return ret; } -- cgit v1.2.3 From c1d519312dcdf11532fed9f99a8ecc3547ffd9d6 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 7 Apr 2008 13:20:54 -0400 Subject: NFSv4: Only increment the sequence id if the server saw it It is quite possible that the OPEN, CLOSE, LOCK, LOCKU,... compounds fail before the actual stateful operation has been executed (for instance in the PUTFH call). There is no way to tell from the overall status result which operations were executed from the COMPOUND. The fix is to move incrementing of the sequence id into the XDR layer, so that we do it as we process the results from the stateful operation. Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 12 ++++++------ fs/nfs/nfs4xdr.c | 18 +++++++++++++++++- 2 files changed, 23 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 424aa206470..9f2759da74e 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -240,6 +240,8 @@ static void nfs4_init_opendata_res(struct nfs4_opendata *p) { p->o_res.f_attr = &p->f_attr; p->o_res.dir_attr = &p->dir_attr; + p->o_res.seqid = p->o_arg.seqid; + p->c_res.seqid = p->c_arg.seqid; p->o_res.server = p->o_arg.server; nfs_fattr_init(&p->f_attr); nfs_fattr_init(&p->dir_attr); @@ -730,7 +732,6 @@ static void nfs4_open_confirm_done(struct rpc_task *task, void *calldata) renew_lease(data->o_res.server, data->timestamp); data->rpc_done = 1; } - nfs_increment_open_seqid(data->rpc_status, data->c_arg.seqid); } static void nfs4_open_confirm_release(void *calldata) @@ -860,7 +861,6 @@ static void nfs4_open_done(struct rpc_task *task, void *calldata) if (!(data->o_res.rflags & NFS4_OPEN_RESULT_CONFIRM)) nfs_confirm_seqid(&data->owner->so_seqid, 0); } - nfs_increment_open_seqid(data->rpc_status, data->o_arg.seqid); data->rpc_done = 1; } @@ -1226,7 +1226,6 @@ static void nfs4_close_done(struct rpc_task *task, void *data) /* hmm. we are done with the inode, and in the process of freeing * the state_owner. we keep this around to process errors */ - nfs_increment_open_seqid(task->tk_status, calldata->arg.seqid); switch (task->tk_status) { case 0: nfs_set_open_stateid(state, &calldata->res.stateid, 0); @@ -1333,6 +1332,7 @@ int nfs4_do_close(struct path *path, struct nfs4_state *state, int wait) goto out_free_calldata; calldata->arg.bitmask = server->attr_bitmask; calldata->res.fattr = &calldata->fattr; + calldata->res.seqid = calldata->arg.seqid; calldata->res.server = server; calldata->path.mnt = mntget(path->mnt); calldata->path.dentry = dget(path->dentry); @@ -3159,6 +3159,7 @@ static struct nfs4_unlockdata *nfs4_alloc_unlockdata(struct file_lock *fl, p->arg.fh = NFS_FH(inode); p->arg.fl = &p->fl; p->arg.seqid = seqid; + p->res.seqid = seqid; p->arg.stateid = &lsp->ls_stateid; p->lsp = lsp; atomic_inc(&lsp->ls_count); @@ -3184,7 +3185,6 @@ static void nfs4_locku_done(struct rpc_task *task, void *data) if (RPC_ASSASSINATED(task)) return; - nfs_increment_lock_seqid(task->tk_status, calldata->arg.seqid); switch (task->tk_status) { case 0: memcpy(calldata->lsp->ls_stateid.data, @@ -3322,6 +3322,7 @@ static struct nfs4_lockdata *nfs4_alloc_lockdata(struct file_lock *fl, p->arg.lock_stateid = &lsp->ls_stateid; p->arg.lock_owner.clientid = server->nfs_client->cl_clientid; p->arg.lock_owner.id = lsp->ls_id.id; + p->res.lock_seqid = p->arg.lock_seqid; p->lsp = lsp; atomic_inc(&lsp->ls_count); p->ctx = get_nfs_open_context(ctx); @@ -3348,6 +3349,7 @@ static void nfs4_lock_prepare(struct rpc_task *task, void *calldata) return; data->arg.open_stateid = &state->stateid; data->arg.new_lock_owner = 1; + data->res.open_seqid = data->arg.open_seqid; } else data->arg.new_lock_owner = 0; data->timestamp = jiffies; @@ -3365,7 +3367,6 @@ static void nfs4_lock_done(struct rpc_task *task, void *calldata) if (RPC_ASSASSINATED(task)) goto out; if (data->arg.new_lock_owner != 0) { - nfs_increment_open_seqid(data->rpc_status, data->arg.open_seqid); if (data->rpc_status == 0) nfs_confirm_seqid(&data->lsp->ls_seqid, 0); else @@ -3377,7 +3378,6 @@ static void nfs4_lock_done(struct rpc_task *task, void *calldata) data->lsp->ls_flags |= NFS_LOCK_INITIALIZED; renew_lease(NFS_SERVER(data->ctx->path.dentry->d_inode), data->timestamp); } - nfs_increment_lock_seqid(data->rpc_status, data->arg.lock_seqid); out: dprintk("%s: done, ret = %d!\n", __FUNCTION__, data->rpc_status); } diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index 2b519f6325f..b8b2e391d18 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -3005,6 +3005,8 @@ static int decode_close(struct xdr_stream *xdr, struct nfs_closeres *res) int status; status = decode_op_hdr(xdr, OP_CLOSE); + if (status != -EIO) + nfs_increment_open_seqid(status, res->seqid); if (status) return status; READ_BUF(NFS4_STATEID_SIZE); @@ -3296,11 +3298,17 @@ static int decode_lock(struct xdr_stream *xdr, struct nfs_lock_res *res) int status; status = decode_op_hdr(xdr, OP_LOCK); + if (status == -EIO) + goto out; if (status == 0) { READ_BUF(NFS4_STATEID_SIZE); COPYMEM(res->stateid.data, NFS4_STATEID_SIZE); } else if (status == -NFS4ERR_DENIED) - return decode_lock_denied(xdr, NULL); + status = decode_lock_denied(xdr, NULL); + if (res->open_seqid != NULL) + nfs_increment_open_seqid(status, res->open_seqid); + nfs_increment_lock_seqid(status, res->lock_seqid); +out: return status; } @@ -3319,6 +3327,8 @@ static int decode_locku(struct xdr_stream *xdr, struct nfs_locku_res *res) int status; status = decode_op_hdr(xdr, OP_LOCKU); + if (status != -EIO) + nfs_increment_lock_seqid(status, res->seqid); if (status == 0) { READ_BUF(NFS4_STATEID_SIZE); COPYMEM(res->stateid.data, NFS4_STATEID_SIZE); @@ -3384,6 +3394,8 @@ static int decode_open(struct xdr_stream *xdr, struct nfs_openres *res) int status; status = decode_op_hdr(xdr, OP_OPEN); + if (status != -EIO) + nfs_increment_open_seqid(status, res->seqid); if (status) return status; READ_BUF(NFS4_STATEID_SIZE); @@ -3416,6 +3428,8 @@ static int decode_open_confirm(struct xdr_stream *xdr, struct nfs_open_confirmre int status; status = decode_op_hdr(xdr, OP_OPEN_CONFIRM); + if (status != -EIO) + nfs_increment_open_seqid(status, res->seqid); if (status) return status; READ_BUF(NFS4_STATEID_SIZE); @@ -3429,6 +3443,8 @@ static int decode_open_downgrade(struct xdr_stream *xdr, struct nfs_closeres *re int status; status = decode_op_hdr(xdr, OP_OPEN_DOWNGRADE); + if (status != -EIO) + nfs_increment_open_seqid(status, res->seqid); if (status) return status; READ_BUF(NFS4_STATEID_SIZE); -- cgit v1.2.3 From 4a9af59fee0701d9db99bc148d87b8852d6d6dd8 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 1 Apr 2008 18:57:06 -0400 Subject: NLM/lockd: Ensure we don't corrupt fl->fl_flags in nlmclnt_unlock() Also fix up nlmclnt_lock() so that it doesn't pass modified versions of fl->fl_flags to nlmclnt_cancel() and other helpers. Signed-off-by: Trond Myklebust --- fs/lockd/clntproc.c | 16 ++++++++++------ 1 file changed, 10 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/lockd/clntproc.c b/fs/lockd/clntproc.c index b6b74a60e1e..4e1c0123b45 100644 --- a/fs/lockd/clntproc.c +++ b/fs/lockd/clntproc.c @@ -493,6 +493,7 @@ nlmclnt_lock(struct nlm_rqst *req, struct file_lock *fl) } fl->fl_flags |= FL_ACCESS; status = do_vfs_lock(fl); + fl->fl_flags = fl_flags; if (status < 0) goto out; @@ -530,10 +531,11 @@ again: goto again; } /* Ensure the resulting lock will get added to granted list */ - fl->fl_flags = fl_flags | FL_SLEEP; + fl->fl_flags |= FL_SLEEP; if (do_vfs_lock(fl) < 0) printk(KERN_WARNING "%s: VFS is out of sync with lock manager!\n", __FUNCTION__); up_read(&host->h_rwsem); + fl->fl_flags = fl_flags; } status = nlm_stat_to_errno(resp->status); out_unblock: @@ -543,7 +545,6 @@ out_unblock: nlmclnt_cancel(host, req->a_args.block, fl); out: nlm_release_call(req); - fl->fl_flags = fl_flags; return status; } @@ -598,7 +599,8 @@ nlmclnt_unlock(struct nlm_rqst *req, struct file_lock *fl) { struct nlm_host *host = req->a_host; struct nlm_res *resp = &req->a_res; - int status = 0; + int status; + unsigned char fl_flags = fl->fl_flags; /* * Note: the server is supposed to either grant us the unlock @@ -607,11 +609,13 @@ nlmclnt_unlock(struct nlm_rqst *req, struct file_lock *fl) */ fl->fl_flags |= FL_EXISTS; down_read(&host->h_rwsem); - if (do_vfs_lock(fl) == -ENOENT) { - up_read(&host->h_rwsem); + status = do_vfs_lock(fl); + up_read(&host->h_rwsem); + fl->fl_flags = fl_flags; + if (status == -ENOENT) { + status = 0; goto out; } - up_read(&host->h_rwsem); if (req->a_flags & RPC_TASK_ASYNC) return nlm_async_call(req, NLMPROC_UNLOCK, &nlmclnt_unlock_ops); -- cgit v1.2.3 From 536ff0f809b0f4d56e1c41e66768d330668e0a55 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 4 Apr 2008 15:08:02 -0400 Subject: NFSv4: Ensure we don't corrupt fl->fl_flags in nfs4_proc_unlck Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 9f2759da74e..a1069325b87 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -3263,6 +3263,7 @@ static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock * struct nfs4_lock_state *lsp; struct rpc_task *task; int status = 0; + unsigned char fl_flags = request->fl_flags; status = nfs4_set_lock_state(state, request); /* Unlock _before_ we do the RPC call */ @@ -3286,6 +3287,7 @@ static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock * status = nfs4_wait_for_completion_rpc_task(task); rpc_put_task(task); out: + request->fl_flags = fl_flags; return status; } -- cgit v1.2.3 From 5e7f37a76fa5b604949020b7317962262812b2dd Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 1 Apr 2008 18:58:49 -0400 Subject: NLM/lockd: Add a reference counter to struct nlm_rqst When we replace the existing synchronous RPC calls with asynchronous calls, the reference count will be needed in order to allow us to examine the result of the RPC call. Signed-off-by: Trond Myklebust --- fs/lockd/clntproc.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'fs') diff --git a/fs/lockd/clntproc.c b/fs/lockd/clntproc.c index 4e1c0123b45..749eb5328cb 100644 --- a/fs/lockd/clntproc.c +++ b/fs/lockd/clntproc.c @@ -221,6 +221,7 @@ struct nlm_rqst *nlm_alloc_call(struct nlm_host *host) for(;;) { call = kzalloc(sizeof(*call), GFP_KERNEL); if (call != NULL) { + atomic_set(&call->a_count, 1); locks_init_lock(&call->a_args.lock.fl); locks_init_lock(&call->a_res.lock.fl); call->a_host = host; @@ -237,6 +238,8 @@ struct nlm_rqst *nlm_alloc_call(struct nlm_host *host) void nlm_release_call(struct nlm_rqst *call) { + if (!atomic_dec_and_test(&call->a_count)) + return; nlm_release_host(call->a_host); nlmclnt_release_lockargs(call); kfree(call); -- cgit v1.2.3 From dc9d8d048168ff61c458bec06b28996cb90b182a Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 28 Mar 2008 16:04:36 -0400 Subject: NLM/lockd: convert __nlm_async_call to use rpc_run_task() Peter Staubach comments: > In the course of investigating testing failures in the locking phase of > the Connectathon testsuite, I discovered a couple of things. One was > that one of the tests in the locking tests was racy when it didn't seem > to need to be and two, that the NFS client asynchronously releases locks > when a process is exiting. ... > The Single UNIX Specification Version 3 specifies that: "All locks > associated with a file for a given process shall be removed when a file > descriptor for that file is closed by that process or the process holding > that file descriptor terminates.". > > This does not specify whether those locks must be released prior to the > completion of the exit processing for the process or not. However, > general assumptions seem to be that those locks will be released. This > leads to more deterministic behavior under normal circumstances. The following patch converts the NFSv2/v3 locking code to use the same mechanism as NFSv4 for sending asynchronous RPC calls and then waiting for them to complete. This ensures that the UNLOCK and CANCEL RPC calls will complete even if the user interrupts the call, yet satisfies the above request for synchronous behaviour on process exit. Signed-off-by: Trond Myklebust --- fs/lockd/clntproc.c | 64 ++++++++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 54 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/lockd/clntproc.c b/fs/lockd/clntproc.c index 749eb5328cb..a34b709006a 100644 --- a/fs/lockd/clntproc.c +++ b/fs/lockd/clntproc.c @@ -346,10 +346,16 @@ in_grace_period: /* * Generic NLM call, async version. */ -static int __nlm_async_call(struct nlm_rqst *req, u32 proc, struct rpc_message *msg, const struct rpc_call_ops *tk_ops) +static struct rpc_task *__nlm_async_call(struct nlm_rqst *req, u32 proc, struct rpc_message *msg, const struct rpc_call_ops *tk_ops) { struct nlm_host *host = req->a_host; struct rpc_clnt *clnt; + struct rpc_task_setup task_setup_data = { + .rpc_message = msg, + .callback_ops = tk_ops, + .callback_data = req, + .flags = RPC_TASK_ASYNC, + }; dprintk("lockd: call procedure %d on %s (async)\n", (int)proc, host->h_name); @@ -359,21 +365,36 @@ static int __nlm_async_call(struct nlm_rqst *req, u32 proc, struct rpc_message * if (clnt == NULL) goto out_err; msg->rpc_proc = &clnt->cl_procinfo[proc]; + task_setup_data.rpc_client = clnt; /* bootstrap and kick off the async RPC call */ - return rpc_call_async(clnt, msg, RPC_TASK_ASYNC, tk_ops, req); + return rpc_run_task(&task_setup_data); out_err: tk_ops->rpc_release(req); - return -ENOLCK; + return ERR_PTR(-ENOLCK); +} + +static int nlm_do_async_call(struct nlm_rqst *req, u32 proc, struct rpc_message *msg, const struct rpc_call_ops *tk_ops) +{ + struct rpc_task *task; + + task = __nlm_async_call(req, proc, msg, tk_ops); + if (IS_ERR(task)) + return PTR_ERR(task); + rpc_put_task(task); + return 0; } +/* + * NLM asynchronous call. + */ int nlm_async_call(struct nlm_rqst *req, u32 proc, const struct rpc_call_ops *tk_ops) { struct rpc_message msg = { .rpc_argp = &req->a_args, .rpc_resp = &req->a_res, }; - return __nlm_async_call(req, proc, &msg, tk_ops); + return nlm_do_async_call(req, proc, &msg, tk_ops); } int nlm_async_reply(struct nlm_rqst *req, u32 proc, const struct rpc_call_ops *tk_ops) @@ -381,7 +402,32 @@ int nlm_async_reply(struct nlm_rqst *req, u32 proc, const struct rpc_call_ops *t struct rpc_message msg = { .rpc_argp = &req->a_res, }; - return __nlm_async_call(req, proc, &msg, tk_ops); + return nlm_do_async_call(req, proc, &msg, tk_ops); +} + +/* + * NLM client asynchronous call. + * + * Note that although the calls are asynchronous, and are therefore + * guaranteed to complete, we still always attempt to wait for + * completion in order to be able to correctly track the lock + * state. + */ +static int nlmclnt_async_call(struct nlm_rqst *req, u32 proc, const struct rpc_call_ops *tk_ops) +{ + struct rpc_message msg = { + .rpc_argp = &req->a_args, + .rpc_resp = &req->a_res, + }; + struct rpc_task *task; + int err; + + task = __nlm_async_call(req, proc, &msg, tk_ops); + if (IS_ERR(task)) + return PTR_ERR(task); + err = rpc_wait_for_completion_task(task); + rpc_put_task(task); + return err; } /* @@ -620,10 +666,8 @@ nlmclnt_unlock(struct nlm_rqst *req, struct file_lock *fl) goto out; } - if (req->a_flags & RPC_TASK_ASYNC) - return nlm_async_call(req, NLMPROC_UNLOCK, &nlmclnt_unlock_ops); - - status = nlmclnt_call(req, NLMPROC_UNLOCK); + atomic_inc(&req->a_count); + status = nlmclnt_async_call(req, NLMPROC_UNLOCK, &nlmclnt_unlock_ops); if (status < 0) goto out; @@ -697,7 +741,7 @@ static int nlmclnt_cancel(struct nlm_host *host, int block, struct file_lock *fl nlmclnt_setlockargs(req, fl); req->a_args.block = block; - status = nlm_async_call(req, NLMPROC_CANCEL, &nlmclnt_cancel_ops); + status = nlmclnt_async_call(req, NLMPROC_CANCEL, &nlmclnt_cancel_ops); spin_lock_irqsave(¤t->sighand->siglock, flags); current->blocked = oldset; -- cgit v1.2.3 From 8ec7ff74448f65ac963e330795d771ab14ec8408 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 28 Mar 2008 16:04:51 -0400 Subject: NLM: Remove the signal masking in nlmclnt_proc/nlmclnt_cancel The signal masks have been rendered obsolete by the preceding patch. Signed-off-by: Trond Myklebust --- fs/lockd/clntproc.c | 42 +----------------------------------------- 1 file changed, 1 insertion(+), 41 deletions(-) (limited to 'fs') diff --git a/fs/lockd/clntproc.c b/fs/lockd/clntproc.c index a34b709006a..5f13e0363b2 100644 --- a/fs/lockd/clntproc.c +++ b/fs/lockd/clntproc.c @@ -155,8 +155,6 @@ static void nlmclnt_release_lockargs(struct nlm_rqst *req) int nlmclnt_proc(struct nlm_host *host, int cmd, struct file_lock *fl) { struct nlm_rqst *call; - sigset_t oldset; - unsigned long flags; int status; nlm_get_host(host); @@ -168,22 +166,6 @@ int nlmclnt_proc(struct nlm_host *host, int cmd, struct file_lock *fl) /* Set up the argument struct */ nlmclnt_setlockargs(call, fl); - /* Keep the old signal mask */ - spin_lock_irqsave(¤t->sighand->siglock, flags); - oldset = current->blocked; - - /* If we're cleaning up locks because the process is exiting, - * perform the RPC call asynchronously. */ - if ((IS_SETLK(cmd) || IS_SETLKW(cmd)) - && fl->fl_type == F_UNLCK - && (current->flags & PF_EXITING)) { - sigfillset(¤t->blocked); /* Mask all signals */ - recalc_sigpending(); - - call->a_flags = RPC_TASK_ASYNC; - } - spin_unlock_irqrestore(¤t->sighand->siglock, flags); - if (IS_SETLK(cmd) || IS_SETLKW(cmd)) { if (fl->fl_type != F_UNLCK) { call->a_args.block = IS_SETLKW(cmd) ? 1 : 0; @@ -198,11 +180,6 @@ int nlmclnt_proc(struct nlm_host *host, int cmd, struct file_lock *fl) fl->fl_ops->fl_release_private(fl); fl->fl_ops = NULL; - spin_lock_irqsave(¤t->sighand->siglock, flags); - current->blocked = oldset; - recalc_sigpending(); - spin_unlock_irqrestore(¤t->sighand->siglock, flags); - dprintk("lockd: clnt proc returns %d\n", status); return status; } @@ -722,16 +699,6 @@ static const struct rpc_call_ops nlmclnt_unlock_ops = { static int nlmclnt_cancel(struct nlm_host *host, int block, struct file_lock *fl) { struct nlm_rqst *req; - unsigned long flags; - sigset_t oldset; - int status; - - /* Block all signals while setting up call */ - spin_lock_irqsave(¤t->sighand->siglock, flags); - oldset = current->blocked; - sigfillset(¤t->blocked); - recalc_sigpending(); - spin_unlock_irqrestore(¤t->sighand->siglock, flags); req = nlm_alloc_call(nlm_get_host(host)); if (!req) @@ -741,14 +708,7 @@ static int nlmclnt_cancel(struct nlm_host *host, int block, struct file_lock *fl nlmclnt_setlockargs(req, fl); req->a_args.block = block; - status = nlmclnt_async_call(req, NLMPROC_CANCEL, &nlmclnt_cancel_ops); - - spin_lock_irqsave(¤t->sighand->siglock, flags); - current->blocked = oldset; - recalc_sigpending(); - spin_unlock_irqrestore(¤t->sighand->siglock, flags); - - return status; + return nlmclnt_async_call(req, NLMPROC_CANCEL, &nlmclnt_cancel_ops); } static void nlmclnt_cancel_callback(struct rpc_task *task, void *data) -- cgit v1.2.3 From 6b4b3a752b3464f2fd9fe2837fb19270c23c1d6b Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 2 Apr 2008 14:40:53 -0400 Subject: NLM/lockd: Ensure that nlmclnt_cancel() returns results of the CANCEL call Currently, it returns success as long as the RPC call was sent. We'd like to know if the CANCEL operation succeeded on the server. Signed-off-by: Trond Myklebust --- fs/lockd/clntproc.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/lockd/clntproc.c b/fs/lockd/clntproc.c index 5f13e0363b2..ea1a6940af2 100644 --- a/fs/lockd/clntproc.c +++ b/fs/lockd/clntproc.c @@ -699,6 +699,10 @@ static const struct rpc_call_ops nlmclnt_unlock_ops = { static int nlmclnt_cancel(struct nlm_host *host, int block, struct file_lock *fl) { struct nlm_rqst *req; + int status; + + dprintk("lockd: blocking lock attempt was interrupted by a signal.\n" + " Attempting to cancel lock.\n"); req = nlm_alloc_call(nlm_get_host(host)); if (!req) @@ -708,7 +712,12 @@ static int nlmclnt_cancel(struct nlm_host *host, int block, struct file_lock *fl nlmclnt_setlockargs(req, fl); req->a_args.block = block; - return nlmclnt_async_call(req, NLMPROC_CANCEL, &nlmclnt_cancel_ops); + atomic_inc(&req->a_count); + status = nlmclnt_async_call(req, NLMPROC_CANCEL, &nlmclnt_cancel_ops); + if (status == 0 && req->a_res.status == nlm_lck_denied) + status = -ENOLCK; + nlm_release_call(req); + return status; } static void nlmclnt_cancel_callback(struct rpc_task *task, void *data) -- cgit v1.2.3 From 5f50c0c6d644d6c8180d9079c13c5d9de3adeb34 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 1 Apr 2008 20:26:22 -0400 Subject: NLM/lockd: Fix a race when cancelling a blocking lock We shouldn't remove the lock from the list of blocked locks until the CANCEL call has completed since we may be racing with a GRANTED callback. Also ensure that we send an UNLOCK if the CANCEL request failed. Normally that should only happen if the process gets hit with a fatal signal. Signed-off-by: Trond Myklebust --- fs/lockd/clntproc.c | 43 ++++++++++++++++++++++++++++++++++--------- 1 file changed, 34 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/lockd/clntproc.c b/fs/lockd/clntproc.c index ea1a6940af2..37d1aa20a60 100644 --- a/fs/lockd/clntproc.c +++ b/fs/lockd/clntproc.c @@ -510,6 +510,7 @@ nlmclnt_lock(struct nlm_rqst *req, struct file_lock *fl) struct nlm_res *resp = &req->a_res; struct nlm_wait *block = NULL; unsigned char fl_flags = fl->fl_flags; + unsigned char fl_type; int status = -ENOLCK; if (nsm_monitor(host) < 0) { @@ -525,13 +526,16 @@ nlmclnt_lock(struct nlm_rqst *req, struct file_lock *fl) block = nlmclnt_prepare_block(host, fl); again: + /* + * Initialise resp->status to a valid non-zero value, + * since 0 == nlm_lck_granted + */ + resp->status = nlm_lck_blocked; for(;;) { /* Reboot protection */ fl->fl_u.nfs_fl.state = host->h_state; status = nlmclnt_call(req, NLMPROC_LOCK); if (status < 0) - goto out_unblock; - if (!req->a_args.block) break; /* Did a reclaimer thread notify us of a server reboot? */ if (resp->status == nlm_lck_denied_grace_period) @@ -540,15 +544,22 @@ again: break; /* Wait on an NLM blocking lock */ status = nlmclnt_block(block, req, NLMCLNT_POLL_TIMEOUT); - /* if we were interrupted. Send a CANCEL request to the server - * and exit - */ if (status < 0) - goto out_unblock; + break; if (resp->status != nlm_lck_blocked) break; } + /* if we were interrupted while blocking, then cancel the lock request + * and exit + */ + if (resp->status == nlm_lck_blocked) { + if (!req->a_args.block) + goto out_unlock; + if (nlmclnt_cancel(host, req->a_args.block, fl) == 0) + goto out_unblock; + } + if (resp->status == nlm_granted) { down_read(&host->h_rwsem); /* Check whether or not the server has rebooted */ @@ -562,16 +573,30 @@ again: printk(KERN_WARNING "%s: VFS is out of sync with lock manager!\n", __FUNCTION__); up_read(&host->h_rwsem); fl->fl_flags = fl_flags; + status = 0; } + if (status < 0) + goto out_unlock; status = nlm_stat_to_errno(resp->status); out_unblock: nlmclnt_finish_block(block); - /* Cancel the blocked request if it is still pending */ - if (resp->status == nlm_lck_blocked) - nlmclnt_cancel(host, req->a_args.block, fl); out: nlm_release_call(req); return status; +out_unlock: + /* Fatal error: ensure that we remove the lock altogether */ + dprintk("lockd: lock attempt ended in fatal error.\n" + " Attempting to unlock.\n"); + nlmclnt_finish_block(block); + fl_type = fl->fl_type; + fl->fl_type = F_UNLCK; + down_read(&host->h_rwsem); + do_vfs_lock(fl); + up_read(&host->h_rwsem); + fl->fl_type = fl_type; + fl->fl_flags = fl_flags; + nlmclnt_async_call(req, NLMPROC_UNLOCK, &nlmclnt_unlock_ops); + return status; } /* -- cgit v1.2.3 From c4d7c402b788b73dc24f1e54a57f89d3dc5eb7bc Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 1 Apr 2008 20:26:52 -0400 Subject: NFS: Remove the buggy lock-if-signalled case from do_setlk() Both NLM and NFSv4 should be able to clean up adequately in the case where the user interrupts the RPC call... Signed-off-by: Trond Myklebust --- fs/nfs/file.c | 12 ++---------- 1 file changed, 2 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/nfs/file.c b/fs/nfs/file.c index 10e8b807e7f..742cb745cb4 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -566,17 +566,9 @@ static int do_setlk(struct file *filp, int cmd, struct file_lock *fl) lock_kernel(); /* Use local locking if mounted with "-onolock" */ - if (!(NFS_SERVER(inode)->flags & NFS_MOUNT_NONLM)) { + if (!(NFS_SERVER(inode)->flags & NFS_MOUNT_NONLM)) status = NFS_PROTO(inode)->lock(filp, cmd, fl); - /* If we were signalled we still need to ensure that - * we clean up any state on the server. We therefore - * record the lock call as having succeeded in order to - * ensure that locks_remove_posix() cleans it out when - * the process exits. - */ - if (status == -EINTR || status == -ERESTARTSYS) - do_vfs_lock(filp, fl); - } else + else status = do_vfs_lock(filp, fl); unlock_kernel(); if (status < 0) -- cgit v1.2.3 From d11d10cc05c94a32632d6928d15a1034200dd9a5 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 2 Apr 2008 14:44:05 -0400 Subject: NLM/lockd: Ensure client locking calls use correct credentials Now that we've added the 'generic' credentials (that are independent of the rpc_client) to the nfs_open_context, we can use those in the NLM client to ensure that the lock/unlock requests are authenticated to whoever originally opened the file. Signed-off-by: Trond Myklebust --- fs/lockd/clntproc.c | 23 ++++++++++++++--------- 1 file changed, 14 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/lockd/clntproc.c b/fs/lockd/clntproc.c index 37d1aa20a60..40b16f23e49 100644 --- a/fs/lockd/clntproc.c +++ b/fs/lockd/clntproc.c @@ -247,7 +247,7 @@ static int nlm_wait_on_grace(wait_queue_head_t *queue) * Generic NLM call */ static int -nlmclnt_call(struct nlm_rqst *req, u32 proc) +nlmclnt_call(struct rpc_cred *cred, struct nlm_rqst *req, u32 proc) { struct nlm_host *host = req->a_host; struct rpc_clnt *clnt; @@ -256,6 +256,7 @@ nlmclnt_call(struct nlm_rqst *req, u32 proc) struct rpc_message msg = { .rpc_argp = argp, .rpc_resp = resp, + .rpc_cred = cred, }; int status; @@ -390,11 +391,12 @@ int nlm_async_reply(struct nlm_rqst *req, u32 proc, const struct rpc_call_ops *t * completion in order to be able to correctly track the lock * state. */ -static int nlmclnt_async_call(struct nlm_rqst *req, u32 proc, const struct rpc_call_ops *tk_ops) +static int nlmclnt_async_call(struct rpc_cred *cred, struct nlm_rqst *req, u32 proc, const struct rpc_call_ops *tk_ops) { struct rpc_message msg = { .rpc_argp = &req->a_args, .rpc_resp = &req->a_res, + .rpc_cred = cred, }; struct rpc_task *task; int err; @@ -415,7 +417,7 @@ nlmclnt_test(struct nlm_rqst *req, struct file_lock *fl) { int status; - status = nlmclnt_call(req, NLMPROC_TEST); + status = nlmclnt_call(nfs_file_cred(fl->fl_file), req, NLMPROC_TEST); if (status < 0) goto out; @@ -506,6 +508,7 @@ static int do_vfs_lock(struct file_lock *fl) static int nlmclnt_lock(struct nlm_rqst *req, struct file_lock *fl) { + struct rpc_cred *cred = nfs_file_cred(fl->fl_file); struct nlm_host *host = req->a_host; struct nlm_res *resp = &req->a_res; struct nlm_wait *block = NULL; @@ -534,7 +537,7 @@ again: for(;;) { /* Reboot protection */ fl->fl_u.nfs_fl.state = host->h_state; - status = nlmclnt_call(req, NLMPROC_LOCK); + status = nlmclnt_call(cred, req, NLMPROC_LOCK); if (status < 0) break; /* Did a reclaimer thread notify us of a server reboot? */ @@ -595,7 +598,7 @@ out_unlock: up_read(&host->h_rwsem); fl->fl_type = fl_type; fl->fl_flags = fl_flags; - nlmclnt_async_call(req, NLMPROC_UNLOCK, &nlmclnt_unlock_ops); + nlmclnt_async_call(cred, req, NLMPROC_UNLOCK, &nlmclnt_unlock_ops); return status; } @@ -619,8 +622,8 @@ nlmclnt_reclaim(struct nlm_host *host, struct file_lock *fl) nlmclnt_setlockargs(req, fl); req->a_args.reclaim = 1; - if ((status = nlmclnt_call(req, NLMPROC_LOCK)) >= 0 - && req->a_res.status == nlm_granted) + status = nlmclnt_call(nfs_file_cred(fl->fl_file), req, NLMPROC_LOCK); + if (status >= 0 && req->a_res.status == nlm_granted) return 0; printk(KERN_WARNING "lockd: failed to reclaim lock for pid %d " @@ -669,7 +672,8 @@ nlmclnt_unlock(struct nlm_rqst *req, struct file_lock *fl) } atomic_inc(&req->a_count); - status = nlmclnt_async_call(req, NLMPROC_UNLOCK, &nlmclnt_unlock_ops); + status = nlmclnt_async_call(nfs_file_cred(fl->fl_file), req, + NLMPROC_UNLOCK, &nlmclnt_unlock_ops); if (status < 0) goto out; @@ -738,7 +742,8 @@ static int nlmclnt_cancel(struct nlm_host *host, int block, struct file_lock *fl req->a_args.block = block; atomic_inc(&req->a_count); - status = nlmclnt_async_call(req, NLMPROC_CANCEL, &nlmclnt_cancel_ops); + status = nlmclnt_async_call(nfs_file_cred(fl->fl_file), req, + NLMPROC_CANCEL, &nlmclnt_cancel_ops); if (status == 0 && req->a_res.status == nlm_lck_denied) status = -ENOLCK; nlm_release_call(req); -- cgit v1.2.3 From 856dff3d3875bdc8b88e4a65779873af76776a69 Mon Sep 17 00:00:00 2001 From: Benny Halevy Date: Mon, 31 Mar 2008 17:39:06 +0300 Subject: nfs: return negative error value from nfs{,4}_stat_to_errno All use sites for nfs{,4}_stat_to_errno negate their return value. It's more efficient to return a negative error from the stat_to_errno convertors rather than negating its return value everywhere. This also produces slightly smaller code. Signed-off-by: Benny Halevy Signed-off-by: Trond Myklebust --- fs/nfs/nfs2xdr.c | 76 ++++++++++++++++++++++++++--------------------------- fs/nfs/nfs3xdr.c | 34 ++++++++++++------------ fs/nfs/nfs4xdr.c | 80 ++++++++++++++++++++++++++++---------------------------- 3 files changed, 95 insertions(+), 95 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs2xdr.c b/fs/nfs/nfs2xdr.c index 86a80b33ec8..28bab67d151 100644 --- a/fs/nfs/nfs2xdr.c +++ b/fs/nfs/nfs2xdr.c @@ -267,7 +267,7 @@ nfs_xdr_readres(struct rpc_rqst *req, __be32 *p, struct nfs_readres *res) int status; if ((status = ntohl(*p++))) - return -nfs_stat_to_errno(status); + return nfs_stat_to_errno(status); p = xdr_decode_fattr(p, res->fattr); count = ntohl(*p++); @@ -432,7 +432,7 @@ nfs_xdr_readdirres(struct rpc_rqst *req, __be32 *p, void *dummy) __be32 *end, *entry, *kaddr; if ((status = ntohl(*p++))) - return -nfs_stat_to_errno(status); + return nfs_stat_to_errno(status); hdrlen = (u8 *) p - (u8 *) iov->iov_base; if (iov->iov_len < hdrlen) { @@ -537,7 +537,7 @@ nfs_xdr_stat(struct rpc_rqst *req, __be32 *p, void *dummy) int status; if ((status = ntohl(*p++)) != 0) - status = -nfs_stat_to_errno(status); + status = nfs_stat_to_errno(status); return status; } @@ -551,7 +551,7 @@ nfs_xdr_attrstat(struct rpc_rqst *req, __be32 *p, struct nfs_fattr *fattr) int status; if ((status = ntohl(*p++))) - return -nfs_stat_to_errno(status); + return nfs_stat_to_errno(status); xdr_decode_fattr(p, fattr); return 0; } @@ -566,7 +566,7 @@ nfs_xdr_diropres(struct rpc_rqst *req, __be32 *p, struct nfs_diropok *res) int status; if ((status = ntohl(*p++))) - return -nfs_stat_to_errno(status); + return nfs_stat_to_errno(status); p = xdr_decode_fhandle(p, res->fh); xdr_decode_fattr(p, res->fattr); return 0; @@ -604,7 +604,7 @@ nfs_xdr_readlinkres(struct rpc_rqst *req, __be32 *p, void *dummy) int status; if ((status = ntohl(*p++))) - return -nfs_stat_to_errno(status); + return nfs_stat_to_errno(status); /* Convert length of symlink */ len = ntohl(*p++); if (len >= rcvbuf->page_len) { @@ -653,7 +653,7 @@ nfs_xdr_statfsres(struct rpc_rqst *req, __be32 *p, struct nfs2_fsstat *res) int status; if ((status = ntohl(*p++))) - return -nfs_stat_to_errno(status); + return nfs_stat_to_errno(status); res->tsize = ntohl(*p++); res->bsize = ntohl(*p++); @@ -672,39 +672,39 @@ static struct { int errno; } nfs_errtbl[] = { { NFS_OK, 0 }, - { NFSERR_PERM, EPERM }, - { NFSERR_NOENT, ENOENT }, - { NFSERR_IO, errno_NFSERR_IO }, - { NFSERR_NXIO, ENXIO }, -/* { NFSERR_EAGAIN, EAGAIN }, */ - { NFSERR_ACCES, EACCES }, - { NFSERR_EXIST, EEXIST }, - { NFSERR_XDEV, EXDEV }, - { NFSERR_NODEV, ENODEV }, - { NFSERR_NOTDIR, ENOTDIR }, - { NFSERR_ISDIR, EISDIR }, - { NFSERR_INVAL, EINVAL }, - { NFSERR_FBIG, EFBIG }, - { NFSERR_NOSPC, ENOSPC }, - { NFSERR_ROFS, EROFS }, - { NFSERR_MLINK, EMLINK }, - { NFSERR_NAMETOOLONG, ENAMETOOLONG }, - { NFSERR_NOTEMPTY, ENOTEMPTY }, - { NFSERR_DQUOT, EDQUOT }, - { NFSERR_STALE, ESTALE }, - { NFSERR_REMOTE, EREMOTE }, + { NFSERR_PERM, -EPERM }, + { NFSERR_NOENT, -ENOENT }, + { NFSERR_IO, -errno_NFSERR_IO}, + { NFSERR_NXIO, -ENXIO }, +/* { NFSERR_EAGAIN, -EAGAIN }, */ + { NFSERR_ACCES, -EACCES }, + { NFSERR_EXIST, -EEXIST }, + { NFSERR_XDEV, -EXDEV }, + { NFSERR_NODEV, -ENODEV }, + { NFSERR_NOTDIR, -ENOTDIR }, + { NFSERR_ISDIR, -EISDIR }, + { NFSERR_INVAL, -EINVAL }, + { NFSERR_FBIG, -EFBIG }, + { NFSERR_NOSPC, -ENOSPC }, + { NFSERR_ROFS, -EROFS }, + { NFSERR_MLINK, -EMLINK }, + { NFSERR_NAMETOOLONG, -ENAMETOOLONG }, + { NFSERR_NOTEMPTY, -ENOTEMPTY }, + { NFSERR_DQUOT, -EDQUOT }, + { NFSERR_STALE, -ESTALE }, + { NFSERR_REMOTE, -EREMOTE }, #ifdef EWFLUSH - { NFSERR_WFLUSH, EWFLUSH }, + { NFSERR_WFLUSH, -EWFLUSH }, #endif - { NFSERR_BADHANDLE, EBADHANDLE }, - { NFSERR_NOT_SYNC, ENOTSYNC }, - { NFSERR_BAD_COOKIE, EBADCOOKIE }, - { NFSERR_NOTSUPP, ENOTSUPP }, - { NFSERR_TOOSMALL, ETOOSMALL }, - { NFSERR_SERVERFAULT, ESERVERFAULT }, - { NFSERR_BADTYPE, EBADTYPE }, - { NFSERR_JUKEBOX, EJUKEBOX }, - { -1, EIO } + { NFSERR_BADHANDLE, -EBADHANDLE }, + { NFSERR_NOT_SYNC, -ENOTSYNC }, + { NFSERR_BAD_COOKIE, -EBADCOOKIE }, + { NFSERR_NOTSUPP, -ENOTSUPP }, + { NFSERR_TOOSMALL, -ETOOSMALL }, + { NFSERR_SERVERFAULT, -ESERVERFAULT }, + { NFSERR_BADTYPE, -EBADTYPE }, + { NFSERR_JUKEBOX, -EJUKEBOX }, + { -1, -EIO } }; /* diff --git a/fs/nfs/nfs3xdr.c b/fs/nfs/nfs3xdr.c index fb03048ac65..11cdddec143 100644 --- a/fs/nfs/nfs3xdr.c +++ b/fs/nfs/nfs3xdr.c @@ -515,7 +515,7 @@ nfs3_xdr_readdirres(struct rpc_rqst *req, __be32 *p, struct nfs3_readdirres *res /* Decode post_op_attrs */ p = xdr_decode_post_op_attr(p, res->dir_attr); if (status) - return -nfs_stat_to_errno(status); + return nfs_stat_to_errno(status); /* Decode verifier cookie */ if (res->verf) { res->verf[0] = *p++; @@ -751,7 +751,7 @@ nfs3_xdr_attrstat(struct rpc_rqst *req, __be32 *p, struct nfs_fattr *fattr) int status; if ((status = ntohl(*p++))) - return -nfs_stat_to_errno(status); + return nfs_stat_to_errno(status); xdr_decode_fattr(p, fattr); return 0; } @@ -766,7 +766,7 @@ nfs3_xdr_wccstat(struct rpc_rqst *req, __be32 *p, struct nfs_fattr *fattr) int status; if ((status = ntohl(*p++))) - status = -nfs_stat_to_errno(status); + status = nfs_stat_to_errno(status); xdr_decode_wcc_data(p, fattr); return status; } @@ -786,7 +786,7 @@ nfs3_xdr_lookupres(struct rpc_rqst *req, __be32 *p, struct nfs3_diropres *res) int status; if ((status = ntohl(*p++))) { - status = -nfs_stat_to_errno(status); + status = nfs_stat_to_errno(status); } else { if (!(p = xdr_decode_fhandle(p, res->fh))) return -errno_NFSERR_IO; @@ -806,7 +806,7 @@ nfs3_xdr_accessres(struct rpc_rqst *req, __be32 *p, struct nfs3_accessres *res) p = xdr_decode_post_op_attr(p, res->fattr); if (status) - return -nfs_stat_to_errno(status); + return nfs_stat_to_errno(status); res->access = ntohl(*p++); return 0; } @@ -843,7 +843,7 @@ nfs3_xdr_readlinkres(struct rpc_rqst *req, __be32 *p, struct nfs_fattr *fattr) p = xdr_decode_post_op_attr(p, fattr); if (status != 0) - return -nfs_stat_to_errno(status); + return nfs_stat_to_errno(status); /* Convert length of symlink */ len = ntohl(*p++); @@ -891,7 +891,7 @@ nfs3_xdr_readres(struct rpc_rqst *req, __be32 *p, struct nfs_readres *res) p = xdr_decode_post_op_attr(p, res->fattr); if (status != 0) - return -nfs_stat_to_errno(status); + return nfs_stat_to_errno(status); /* Decode reply count and EOF flag. NFSv3 is somewhat redundant * in that it puts the count both in the res struct and in the @@ -941,7 +941,7 @@ nfs3_xdr_writeres(struct rpc_rqst *req, __be32 *p, struct nfs_writeres *res) p = xdr_decode_wcc_data(p, res->fattr); if (status != 0) - return -nfs_stat_to_errno(status); + return nfs_stat_to_errno(status); res->count = ntohl(*p++); res->verf->committed = (enum nfs3_stable_how)ntohl(*p++); @@ -972,7 +972,7 @@ nfs3_xdr_createres(struct rpc_rqst *req, __be32 *p, struct nfs3_diropres *res) res->fattr->valid = 0; } } else { - status = -nfs_stat_to_errno(status); + status = nfs_stat_to_errno(status); } p = xdr_decode_wcc_data(p, res->dir_attr); return status; @@ -987,7 +987,7 @@ nfs3_xdr_renameres(struct rpc_rqst *req, __be32 *p, struct nfs3_renameres *res) int status; if ((status = ntohl(*p++)) != 0) - status = -nfs_stat_to_errno(status); + status = nfs_stat_to_errno(status); p = xdr_decode_wcc_data(p, res->fromattr); p = xdr_decode_wcc_data(p, res->toattr); return status; @@ -1002,7 +1002,7 @@ nfs3_xdr_linkres(struct rpc_rqst *req, __be32 *p, struct nfs3_linkres *res) int status; if ((status = ntohl(*p++)) != 0) - status = -nfs_stat_to_errno(status); + status = nfs_stat_to_errno(status); p = xdr_decode_post_op_attr(p, res->fattr); p = xdr_decode_wcc_data(p, res->dir_attr); return status; @@ -1020,7 +1020,7 @@ nfs3_xdr_fsstatres(struct rpc_rqst *req, __be32 *p, struct nfs_fsstat *res) p = xdr_decode_post_op_attr(p, res->fattr); if (status != 0) - return -nfs_stat_to_errno(status); + return nfs_stat_to_errno(status); p = xdr_decode_hyper(p, &res->tbytes); p = xdr_decode_hyper(p, &res->fbytes); @@ -1045,7 +1045,7 @@ nfs3_xdr_fsinfores(struct rpc_rqst *req, __be32 *p, struct nfs_fsinfo *res) p = xdr_decode_post_op_attr(p, res->fattr); if (status != 0) - return -nfs_stat_to_errno(status); + return nfs_stat_to_errno(status); res->rtmax = ntohl(*p++); res->rtpref = ntohl(*p++); @@ -1073,7 +1073,7 @@ nfs3_xdr_pathconfres(struct rpc_rqst *req, __be32 *p, struct nfs_pathconf *res) p = xdr_decode_post_op_attr(p, res->fattr); if (status != 0) - return -nfs_stat_to_errno(status); + return nfs_stat_to_errno(status); res->max_link = ntohl(*p++); res->max_namelen = ntohl(*p++); @@ -1092,7 +1092,7 @@ nfs3_xdr_commitres(struct rpc_rqst *req, __be32 *p, struct nfs_writeres *res) status = ntohl(*p++); p = xdr_decode_wcc_data(p, res->fattr); if (status != 0) - return -nfs_stat_to_errno(status); + return nfs_stat_to_errno(status); res->verf->verifier[0] = *p++; res->verf->verifier[1] = *p++; @@ -1114,7 +1114,7 @@ nfs3_xdr_getaclres(struct rpc_rqst *req, __be32 *p, int err, base; if (status != 0) - return -nfs_stat_to_errno(status); + return nfs_stat_to_errno(status); p = xdr_decode_post_op_attr(p, res->fattr); res->mask = ntohl(*p++); if (res->mask & ~(NFS_ACL|NFS_ACLCNT|NFS_DFACL|NFS_DFACLCNT)) @@ -1141,7 +1141,7 @@ nfs3_xdr_setaclres(struct rpc_rqst *req, __be32 *p, struct nfs_fattr *fattr) int status = ntohl(*p++); if (status) - return -nfs_stat_to_errno(status); + return nfs_stat_to_errno(status); xdr_decode_post_op_attr(p, fattr); return 0; } diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index b8b2e391d18..809ef0d787e 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -2241,7 +2241,7 @@ static int decode_op_hdr(struct xdr_stream *xdr, enum nfs_opnum4 expected) } READ32(nfserr); if (nfserr != NFS_OK) - return -nfs4_stat_to_errno(nfserr); + return nfs4_stat_to_errno(nfserr); return 0; } @@ -3760,7 +3760,7 @@ static int decode_setclientid(struct xdr_stream *xdr, struct nfs_client *clp) READ_BUF(len); return -NFSERR_CLID_INUSE; } else - return -nfs4_stat_to_errno(nfserr); + return nfs4_stat_to_errno(nfserr); return 0; } @@ -4422,7 +4422,7 @@ static int nfs4_xdr_dec_fsinfo(struct rpc_rqst *req, __be32 *p, struct nfs_fsinf if (!status) status = decode_fsinfo(&xdr, fsinfo); if (!status) - status = -nfs4_stat_to_errno(hdr.status); + status = nfs4_stat_to_errno(hdr.status); return status; } @@ -4512,7 +4512,7 @@ static int nfs4_xdr_dec_setclientid(struct rpc_rqst *req, __be32 *p, if (!status) status = decode_setclientid(&xdr, clp); if (!status) - status = -nfs4_stat_to_errno(hdr.status); + status = nfs4_stat_to_errno(hdr.status); return status; } @@ -4534,7 +4534,7 @@ static int nfs4_xdr_dec_setclientid_confirm(struct rpc_rqst *req, __be32 *p, str if (!status) status = decode_fsinfo(&xdr, fsinfo); if (!status) - status = -nfs4_stat_to_errno(hdr.status); + status = nfs4_stat_to_errno(hdr.status); return status; } @@ -4644,42 +4644,42 @@ static struct { int errno; } nfs_errtbl[] = { { NFS4_OK, 0 }, - { NFS4ERR_PERM, EPERM }, - { NFS4ERR_NOENT, ENOENT }, - { NFS4ERR_IO, errno_NFSERR_IO }, - { NFS4ERR_NXIO, ENXIO }, - { NFS4ERR_ACCESS, EACCES }, - { NFS4ERR_EXIST, EEXIST }, - { NFS4ERR_XDEV, EXDEV }, - { NFS4ERR_NOTDIR, ENOTDIR }, - { NFS4ERR_ISDIR, EISDIR }, - { NFS4ERR_INVAL, EINVAL }, - { NFS4ERR_FBIG, EFBIG }, - { NFS4ERR_NOSPC, ENOSPC }, - { NFS4ERR_ROFS, EROFS }, - { NFS4ERR_MLINK, EMLINK }, - { NFS4ERR_NAMETOOLONG, ENAMETOOLONG }, - { NFS4ERR_NOTEMPTY, ENOTEMPTY }, - { NFS4ERR_DQUOT, EDQUOT }, - { NFS4ERR_STALE, ESTALE }, - { NFS4ERR_BADHANDLE, EBADHANDLE }, - { NFS4ERR_BADOWNER, EINVAL }, - { NFS4ERR_BADNAME, EINVAL }, - { NFS4ERR_BAD_COOKIE, EBADCOOKIE }, - { NFS4ERR_NOTSUPP, ENOTSUPP }, - { NFS4ERR_TOOSMALL, ETOOSMALL }, - { NFS4ERR_SERVERFAULT, ESERVERFAULT }, - { NFS4ERR_BADTYPE, EBADTYPE }, - { NFS4ERR_LOCKED, EAGAIN }, - { NFS4ERR_RESOURCE, EREMOTEIO }, - { NFS4ERR_SYMLINK, ELOOP }, - { NFS4ERR_OP_ILLEGAL, EOPNOTSUPP }, - { NFS4ERR_DEADLOCK, EDEADLK }, - { NFS4ERR_WRONGSEC, EPERM }, /* FIXME: this needs + { NFS4ERR_PERM, -EPERM }, + { NFS4ERR_NOENT, -ENOENT }, + { NFS4ERR_IO, -errno_NFSERR_IO}, + { NFS4ERR_NXIO, -ENXIO }, + { NFS4ERR_ACCESS, -EACCES }, + { NFS4ERR_EXIST, -EEXIST }, + { NFS4ERR_XDEV, -EXDEV }, + { NFS4ERR_NOTDIR, -ENOTDIR }, + { NFS4ERR_ISDIR, -EISDIR }, + { NFS4ERR_INVAL, -EINVAL }, + { NFS4ERR_FBIG, -EFBIG }, + { NFS4ERR_NOSPC, -ENOSPC }, + { NFS4ERR_ROFS, -EROFS }, + { NFS4ERR_MLINK, -EMLINK }, + { NFS4ERR_NAMETOOLONG, -ENAMETOOLONG }, + { NFS4ERR_NOTEMPTY, -ENOTEMPTY }, + { NFS4ERR_DQUOT, -EDQUOT }, + { NFS4ERR_STALE, -ESTALE }, + { NFS4ERR_BADHANDLE, -EBADHANDLE }, + { NFS4ERR_BADOWNER, -EINVAL }, + { NFS4ERR_BADNAME, -EINVAL }, + { NFS4ERR_BAD_COOKIE, -EBADCOOKIE }, + { NFS4ERR_NOTSUPP, -ENOTSUPP }, + { NFS4ERR_TOOSMALL, -ETOOSMALL }, + { NFS4ERR_SERVERFAULT, -ESERVERFAULT }, + { NFS4ERR_BADTYPE, -EBADTYPE }, + { NFS4ERR_LOCKED, -EAGAIN }, + { NFS4ERR_RESOURCE, -EREMOTEIO }, + { NFS4ERR_SYMLINK, -ELOOP }, + { NFS4ERR_OP_ILLEGAL, -EOPNOTSUPP }, + { NFS4ERR_DEADLOCK, -EDEADLK }, + { NFS4ERR_WRONGSEC, -EPERM }, /* FIXME: this needs * to be handled by a * middle-layer. */ - { -1, EIO } + { -1, -EIO } }; /* @@ -4696,14 +4696,14 @@ nfs4_stat_to_errno(int stat) } if (stat <= 10000 || stat > 10100) { /* The server is looney tunes. */ - return ESERVERFAULT; + return -ESERVERFAULT; } /* If we cannot translate the error, the recovery routines should * handle it. * Note: remaining NFSv4 error codes have values > 10000, so should * not conflict with native Linux error codes. */ - return stat; + return -stat; } #define PROC(proc, argtype, restype) \ -- cgit v1.2.3 From 441092415770ddec648800701895913c4bfd60c1 Mon Sep 17 00:00:00 2001 From: Fred Isaman Date: Wed, 2 Apr 2008 15:21:15 +0300 Subject: nfs: fix printout of multiword bitfields Benny points out that zero-padding of multiword bitfields is necessary, and that delimiting each word is nice to avoid endianess confusion. bhalevy: without zero padding output can be ambiguous. Also, since the printed array of two 32-bit unsigned integers is not a 64-bit number, delimiting the output with a semicolon makes more sense. Signed-off-by: Fred Isaman Signed-off-by: Benny Halevy Signed-off-by: Trond Myklebust --- fs/nfs/nfs4xdr.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index 809ef0d787e..5a2d64927b3 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -1191,8 +1191,8 @@ static int encode_readdir(struct xdr_stream *xdr, const struct nfs4_readdir_arg attrs[1] &= ~FATTR4_WORD1_MOUNTED_ON_FILEID; WRITE32(attrs[0] & readdir->bitmask[0]); WRITE32(attrs[1] & readdir->bitmask[1]); - dprintk("%s: cookie = %Lu, verifier = 0x%x%x, bitmap = 0x%x%x\n", - __FUNCTION__, + dprintk("%s: cookie = %Lu, verifier = %08x:%08x, bitmap = %08x:%08x\n", + __func__, (unsigned long long)readdir->cookie, ((u32 *)readdir->verifier.data)[0], ((u32 *)readdir->verifier.data)[1], @@ -2291,7 +2291,7 @@ static int decode_attr_supported(struct xdr_stream *xdr, uint32_t *bitmap, uint3 bitmap[0] &= ~FATTR4_WORD0_SUPPORTED_ATTRS; } else bitmask[0] = bitmask[1] = 0; - dprintk("%s: bitmask=0x%x%x\n", __FUNCTION__, bitmask[0], bitmask[1]); + dprintk("%s: bitmask=%08x:%08x\n", __func__, bitmask[0], bitmask[1]); return 0; } @@ -3505,8 +3505,8 @@ static int decode_readdir(struct xdr_stream *xdr, struct rpc_rqst *req, struct n return status; READ_BUF(8); COPYMEM(readdir->verifier.data, 8); - dprintk("%s: verifier = 0x%x%x\n", - __FUNCTION__, + dprintk("%s: verifier = %08x:%08x\n", + __func__, ((u32 *)readdir->verifier.data)[0], ((u32 *)readdir->verifier.data)[1]); -- cgit v1.2.3 From 78ea323be6380a9313e87fe241809e912e8ae401 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 7 Apr 2008 20:49:28 -0400 Subject: NFSv4: Don't use cred->cr_ops->cr_name in nfs4_proc_setclientid() With the recent change to generic creds, we can no longer use cred->cr_ops->cr_name to distinguish between RPCSEC_GSS principals and AUTH_SYS/AUTH_NULL identities. Replace it with the rpc_authops->au_name instead... Signed-off-by: Trond Myklebust --- fs/nfs/nfs4proc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index a1069325b87..dbc09271af0 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -2885,7 +2885,7 @@ int nfs4_proc_setclientid(struct nfs_client *clp, u32 program, unsigned short po RPC_DISPLAY_ADDR), rpc_peeraddr2str(clp->cl_rpcclient, RPC_DISPLAY_PROTO), - cred->cr_ops->cr_name, + clp->cl_rpcclient->cl_auth->au_ops->au_name, clp->cl_id_uniquifier); setclientid.sc_netid_len = scnprintf(setclientid.sc_netid, sizeof(setclientid.sc_netid), -- cgit v1.2.3 From 7c67db3a8a98045744f06fcd6d8f476d9df0ba5c Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Mon, 7 Apr 2008 20:50:11 -0400 Subject: NFSv4: Reintroduce machine creds We need to try to ensure that we always use the same credentials whenever we re-establish the clientid on the server. If not, the server won't recognise that we're the same client, and so may not allow us to recover state. Signed-off-by: Trond Myklebust --- fs/nfs/client.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'fs') diff --git a/fs/nfs/client.c b/fs/nfs/client.c index 93dfd75aba7..f2f3b284e6d 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -112,6 +112,7 @@ struct nfs_client_initdata { static struct nfs_client *nfs_alloc_client(const struct nfs_client_initdata *cl_init) { struct nfs_client *clp; + struct rpc_cred *cred; if ((clp = kzalloc(sizeof(*clp), GFP_KERNEL)) == NULL) goto error_0; @@ -150,6 +151,9 @@ static struct nfs_client *nfs_alloc_client(const struct nfs_client_initdata *cl_ clp->cl_boot_time = CURRENT_TIME; clp->cl_state = 1 << NFS4CLNT_LEASE_EXPIRED; #endif + cred = rpc_lookup_machine_cred(); + if (!IS_ERR(cred)) + clp->cl_machine_cred = cred; return clp; @@ -191,6 +195,9 @@ static void nfs_free_client(struct nfs_client *clp) if (__test_and_clear_bit(NFS_CS_CALLBACK, &clp->cl_res_state)) nfs_callback_down(); + if (clp->cl_machine_cred != NULL) + put_rpccred(clp->cl_machine_cred); + kfree(clp->cl_hostname); kfree(clp); -- cgit v1.2.3 From a2b2bb8822c78806930bbb4d4c5bb3ae69648fd0 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Tue, 8 Apr 2008 16:02:17 -0400 Subject: NFSv4: Attempt to use machine credentials in SETCLIENTID calls Signed-off-by: Trond Myklebust --- fs/nfs/nfs4state.c | 41 +++++++++++++++++++++++++++++++++++++---- 1 file changed, 37 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index 7775435ea7a..46eb624e4f1 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -71,6 +71,29 @@ static int nfs4_init_client(struct nfs_client *clp, struct rpc_cred *cred) return status; } +static struct rpc_cred *nfs4_get_machine_cred(struct nfs_client *clp) +{ + struct rpc_cred *cred = NULL; + + spin_lock(&clp->cl_lock); + if (clp->cl_machine_cred != NULL) + cred = get_rpccred(clp->cl_machine_cred); + spin_unlock(&clp->cl_lock); + return cred; +} + +static void nfs4_clear_machine_cred(struct nfs_client *clp) +{ + struct rpc_cred *cred; + + spin_lock(&clp->cl_lock); + cred = clp->cl_machine_cred; + clp->cl_machine_cred = NULL; + spin_unlock(&clp->cl_lock); + if (cred != NULL) + put_rpccred(cred); +} + struct rpc_cred *nfs4_get_renew_cred(struct nfs_client *clp) { struct nfs4_state_owner *sp; @@ -91,13 +114,18 @@ static struct rpc_cred *nfs4_get_setclientid_cred(struct nfs_client *clp) { struct nfs4_state_owner *sp; struct rb_node *pos; + struct rpc_cred *cred; + cred = nfs4_get_machine_cred(clp); + if (cred != NULL) + goto out; pos = rb_first(&clp->cl_state_owners); if (pos != NULL) { sp = rb_entry(pos, struct nfs4_state_owner, so_client_node); - return get_rpccred(sp->so_cred); + cred = get_rpccred(sp->so_cred); } - return NULL; +out: + return cred; } static void nfs_alloc_unique_id(struct rb_root *root, struct nfs_unique_id *new, @@ -924,10 +952,10 @@ restart_loop: if (cred != NULL) { /* Yes there are: try to renew the old lease */ status = nfs4_proc_renew(clp, cred); + put_rpccred(cred); switch (status) { case 0: case -NFS4ERR_CB_PATH_DOWN: - put_rpccred(cred); goto out; case -NFS4ERR_STALE_CLIENTID: case -NFS4ERR_LEASE_MOVED: @@ -936,14 +964,19 @@ restart_loop: } else { /* "reboot" to ensure we clear all state on the server */ clp->cl_boot_time = CURRENT_TIME; - cred = nfs4_get_setclientid_cred(clp); } /* We're going to have to re-establish a clientid */ nfs4_state_mark_reclaim(clp); status = -ENOENT; + cred = nfs4_get_setclientid_cred(clp); if (cred != NULL) { status = nfs4_init_client(clp, cred); put_rpccred(cred); + /* Handle case where the user hasn't set up machine creds */ + if (status == -EACCES && cred == clp->cl_machine_cred) { + nfs4_clear_machine_cred(clp); + goto restart_loop; + } } if (status) goto out_error; -- cgit v1.2.3 From 63649bd7080a6a50fabcb1935f4b7c4e64155066 Mon Sep 17 00:00:00 2001 From: Cyrill Gorcunov Date: Thu, 17 Apr 2008 20:42:09 +0400 Subject: NFS - fix potential NULL pointer dereference v2 There is possible NULL pointer dereference if kstr[n]dup failed. So fix them for safety. Signed-off-by: Cyrill Gorcunov Signed-off-by: Trond Myklebust --- fs/nfs/super.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) (limited to 'fs') diff --git a/fs/nfs/super.c b/fs/nfs/super.c index c99ca1f992c..2215bcd24bd 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -1297,6 +1297,8 @@ static int nfs_validate_mount_data(void *options, args->namlen = data->namlen; args->bsize = data->bsize; args->auth_flavors[0] = data->pseudoflavor; + if (!args->nfs_server.hostname) + goto out_nomem; /* * The legacy version 6 binary mount data from userspace has a @@ -1343,6 +1345,8 @@ static int nfs_validate_mount_data(void *options, len = c - dev_name; /* N.B. caller will free nfs_server.hostname in all cases */ args->nfs_server.hostname = kstrndup(dev_name, len, GFP_KERNEL); + if (!args->nfs_server.hostname) + goto out_nomem; c++; if (strlen(c) > NFS_MAXPATHLEN) @@ -1386,6 +1390,10 @@ out_v3_not_compiled: return -EPROTONOSUPPORT; #endif /* !CONFIG_NFS_V3 */ +out_nomem: + dfprintk(MOUNT, "NFS: not enough memory to handle mount options\n"); + return -ENOMEM; + out_no_address: dfprintk(MOUNT, "NFS: mount program didn't pass remote address\n"); return -EINVAL; @@ -1892,12 +1900,16 @@ static int nfs4_validate_mount_data(void *options, return -ENAMETOOLONG; /* N.B. caller will free nfs_server.hostname in all cases */ args->nfs_server.hostname = kstrndup(dev_name, len, GFP_KERNEL); + if (!args->nfs_server.hostname) + goto out_nomem; c++; /* step over the ':' */ len = strlen(c); if (len > NFS4_MAXPATHLEN) return -ENAMETOOLONG; args->nfs_server.export_path = kstrndup(c, len, GFP_KERNEL); + if (!args->nfs_server.export_path) + goto out_nomem; dprintk("NFS: MNTPATH: '%s'\n", args->nfs_server.export_path); @@ -1919,6 +1931,10 @@ out_inval_auth: data->auth_flavourlen); return -EINVAL; +out_nomem: + dfprintk(MOUNT, "NFS4: not enough memory to handle mount options\n"); + return -ENOMEM; + out_no_address: dfprintk(MOUNT, "NFS4: mount program didn't pass remote address\n"); return -EINVAL; -- cgit v1.2.3 From daa7da5fd3040e08e3d7878cd0667c1f4cfd338a Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Fri, 11 Apr 2008 11:50:58 -0400 Subject: NFS: remove duplicate flags assignment from nfs_validate_mount_data Signed-off-by: Jeff Layton Signed-off-by: Trond Myklebust --- fs/nfs/super.c | 1 - 1 file changed, 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 2215bcd24bd..88a0ecd84c6 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -1275,7 +1275,6 @@ static int nfs_validate_mount_data(void *options, args->flags = data->flags; args->rsize = data->rsize; args->wsize = data->wsize; - args->flags = data->flags; args->timeo = data->timeo; args->retrans = data->retrans; args->acregmin = data->acregmin; -- cgit v1.2.3 From a3dab293539031b0970585b9b355cebbc91ecbd4 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Mon, 14 Apr 2008 21:41:32 +0300 Subject: make nfs_automount_list static nfs_automount_list can now become static. Signed-off-by: Adrian Bunk Signed-off-by: Trond Myklebust --- fs/nfs/namespace.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/namespace.c b/fs/nfs/namespace.c index 607f6eb9cdb..af4d0f1e402 100644 --- a/fs/nfs/namespace.c +++ b/fs/nfs/namespace.c @@ -20,7 +20,7 @@ static void nfs_expire_automounts(struct work_struct *work); -LIST_HEAD(nfs_automount_list); +static LIST_HEAD(nfs_automount_list); static DECLARE_DELAYED_WORK(nfs_automount_task, nfs_expire_automounts); int nfs_mountpoint_expiry_timeout = 500 * HZ; -- cgit v1.2.3 From 6bcf19d02a5d7e627fa054f2f10e0a8d830df326 Mon Sep 17 00:00:00 2001 From: Kay Sievers Date: Sun, 9 Mar 2008 21:26:02 +0100 Subject: block: send disk "change" event for rescan_partitions() Userspace likes to get notified that the disk may have changed, when rescan_partitions() is called after partitioning or media change. It will make it possible to update the state of the disk with the "change" event, before the following partition "add" events are handled. Cc: David Zeuthen Signed-off-by: Kay Sievers Signed-off-by: Greg Kroah-Hartman --- fs/partitions/check.c | 4 ++++ 1 file changed, 4 insertions(+) (limited to 'fs') diff --git a/fs/partitions/check.c b/fs/partitions/check.c index 03f808c5b79..6149e4b58c8 100644 --- a/fs/partitions/check.c +++ b/fs/partitions/check.c @@ -473,6 +473,10 @@ int rescan_partitions(struct gendisk *disk, struct block_device *bdev) return 0; if (IS_ERR(state)) /* I/O error reading the partition table */ return -EIO; + + /* tell userspace that the media / partition table may have changed */ + kobject_uevent(&disk->dev.kobj, KOBJ_CHANGE); + for (p = 1; p < state->limit; p++) { sector_t size = state->parts[p].size; sector_t from = state->parts[p].from; -- cgit v1.2.3 From c6f87733823d69a8f12e391688ceeb1ff4922530 Mon Sep 17 00:00:00 2001 From: "Robert P. J. Day" Date: Thu, 13 Mar 2008 22:41:52 -0400 Subject: SYSFS: Explicitly include required header file slab.h. After an experimental deletion of the unnecessary inclusion of from the header file , the following files under fs/sysfs were exposed as needing to explicitly include . Signed-off-by: Robert P. J. Day Signed-off-by: Greg Kroah-Hartman --- fs/sysfs/dir.c | 1 + fs/sysfs/file.c | 1 + 2 files changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/sysfs/dir.c b/fs/sysfs/dir.c index 4948d9bc405..a1c3a1fab7f 100644 --- a/fs/sysfs/dir.c +++ b/fs/sysfs/dir.c @@ -20,6 +20,7 @@ #include #include #include +#include #include "sysfs.h" DEFINE_MUTEX(sysfs_mutex); diff --git a/fs/sysfs/file.c b/fs/sysfs/file.c index baa663e6938..a859c32ff93 100644 --- a/fs/sysfs/file.c +++ b/fs/sysfs/file.c @@ -13,6 +13,7 @@ #include #include #include +#include #include #include #include -- cgit v1.2.3 From 2424b5dd062cbe3e0578ae7b11a1b360ad22f451 Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Mon, 7 Apr 2008 15:35:01 -0700 Subject: sysfs: refill attribute buffer when reading from offset 0 Requiring userspace to close and re-open sysfs attributes has been the policy since before 2.6.12. It allows userspace to get a consistent snapshot of kernel state and consume it with incremental reads and seeks. Now, if the file position is zero the kernel assumes userspace wants to see the new value. The application for this change is to allow a userspace RAID metadata handler to check the state of an array without causing any memory allocations. Thus not causing writeback to a raid array that might be blocked waiting for userspace to take action. Cc: Neil Brown Acked-by: Tejun Heo Signed-off-by: Dan Williams Signed-off-by: Greg Kroah-Hartman --- fs/sysfs/file.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/sysfs/file.c b/fs/sysfs/file.c index a859c32ff93..ade9a7e6a75 100644 --- a/fs/sysfs/file.c +++ b/fs/sysfs/file.c @@ -129,7 +129,7 @@ sysfs_read_file(struct file *file, char __user *buf, size_t count, loff_t *ppos) ssize_t retval = 0; mutex_lock(&buffer->mutex); - if (buffer->needs_read_fill) { + if (buffer->needs_read_fill || *ppos == 0) { retval = fill_read_buffer(file->f_path.dentry,buffer); if (retval) goto out; @@ -410,8 +410,7 @@ static int sysfs_release(struct inode *inode, struct file *filp) * return POLLERR|POLLPRI, and select will return the fd whether * it is waiting for read, write, or exceptions. * Once poll/select indicates that the value has changed, you - * need to close and re-open the file, as simply seeking and reading - * again will not get new data, or reset the state of 'poll'. + * need to close and re-open the file, or seek to 0 and read again. * Reminder: this only works for attributes which actively support * it, and it is not possible to test an attribute from userspace * to see if it supports poll (Neither 'poll' nor 'select' return -- cgit v1.2.3 From c5dec1c3034f1ae3503efbf641ff3b0273b64797 Mon Sep 17 00:00:00 2001 From: FUJITA Tomonori Date: Fri, 11 Apr 2008 12:56:49 +0200 Subject: block: convert bio_copy_user to bio_copy_user_iov This patch enables bio_copy_user to take struct sg_iovec (renamed bio_copy_user_iov). bio_copy_user uses bio_copy_user_iov internally as bio_map_user uses bio_map_user_iov. The major changes are: - adds sg_iovec array to struct bio_map_data - adds __bio_copy_iov that copy data between bio and sg_iovec. bio_copy_user_iov and bio_uncopy_user use it. Signed-off-by: FUJITA Tomonori Cc: Tejun Heo Cc: Mike Christie Cc: James Bottomley Signed-off-by: Jens Axboe --- fs/bio.c | 158 ++++++++++++++++++++++++++++++++++++++++++++++----------------- 1 file changed, 117 insertions(+), 41 deletions(-) (limited to 'fs') diff --git a/fs/bio.c b/fs/bio.c index 553b5b7960a..6e0b6f66df0 100644 --- a/fs/bio.c +++ b/fs/bio.c @@ -444,22 +444,27 @@ int bio_add_page(struct bio *bio, struct page *page, unsigned int len, struct bio_map_data { struct bio_vec *iovecs; - void __user *userptr; + int nr_sgvecs; + struct sg_iovec *sgvecs; }; -static void bio_set_map_data(struct bio_map_data *bmd, struct bio *bio) +static void bio_set_map_data(struct bio_map_data *bmd, struct bio *bio, + struct sg_iovec *iov, int iov_count) { memcpy(bmd->iovecs, bio->bi_io_vec, sizeof(struct bio_vec) * bio->bi_vcnt); + memcpy(bmd->sgvecs, iov, sizeof(struct sg_iovec) * iov_count); + bmd->nr_sgvecs = iov_count; bio->bi_private = bmd; } static void bio_free_map_data(struct bio_map_data *bmd) { kfree(bmd->iovecs); + kfree(bmd->sgvecs); kfree(bmd); } -static struct bio_map_data *bio_alloc_map_data(int nr_segs) +static struct bio_map_data *bio_alloc_map_data(int nr_segs, int iov_count) { struct bio_map_data *bmd = kmalloc(sizeof(*bmd), GFP_KERNEL); @@ -467,13 +472,71 @@ static struct bio_map_data *bio_alloc_map_data(int nr_segs) return NULL; bmd->iovecs = kmalloc(sizeof(struct bio_vec) * nr_segs, GFP_KERNEL); - if (bmd->iovecs) + if (!bmd->iovecs) { + kfree(bmd); + return NULL; + } + + bmd->sgvecs = kmalloc(sizeof(struct sg_iovec) * iov_count, GFP_KERNEL); + if (bmd->sgvecs) return bmd; + kfree(bmd->iovecs); kfree(bmd); return NULL; } +static int __bio_copy_iov(struct bio *bio, struct sg_iovec *iov, int iov_count, + int uncopy) +{ + int ret = 0, i; + struct bio_vec *bvec; + int iov_idx = 0; + unsigned int iov_off = 0; + int read = bio_data_dir(bio) == READ; + + __bio_for_each_segment(bvec, bio, i, 0) { + char *bv_addr = page_address(bvec->bv_page); + unsigned int bv_len = bvec->bv_len; + + while (bv_len && iov_idx < iov_count) { + unsigned int bytes; + char *iov_addr; + + bytes = min_t(unsigned int, + iov[iov_idx].iov_len - iov_off, bv_len); + iov_addr = iov[iov_idx].iov_base + iov_off; + + if (!ret) { + if (!read && !uncopy) + ret = copy_from_user(bv_addr, iov_addr, + bytes); + if (read && uncopy) + ret = copy_to_user(iov_addr, bv_addr, + bytes); + + if (ret) + ret = -EFAULT; + } + + bv_len -= bytes; + bv_addr += bytes; + iov_addr += bytes; + iov_off += bytes; + + if (iov[iov_idx].iov_len == iov_off) { + iov_idx++; + iov_off = 0; + } + } + + if (uncopy) + __free_page(bvec->bv_page); + } + + return ret; +} + /** * bio_uncopy_user - finish previously mapped bio * @bio: bio being terminated @@ -484,55 +547,56 @@ static struct bio_map_data *bio_alloc_map_data(int nr_segs) int bio_uncopy_user(struct bio *bio) { struct bio_map_data *bmd = bio->bi_private; - const int read = bio_data_dir(bio) == READ; - struct bio_vec *bvec; - int i, ret = 0; + int ret; - __bio_for_each_segment(bvec, bio, i, 0) { - char *addr = page_address(bvec->bv_page); - unsigned int len = bmd->iovecs[i].bv_len; + ret = __bio_copy_iov(bio, bmd->sgvecs, bmd->nr_sgvecs, 1); - if (read && !ret && copy_to_user(bmd->userptr, addr, len)) - ret = -EFAULT; - - __free_page(bvec->bv_page); - bmd->userptr += len; - } bio_free_map_data(bmd); bio_put(bio); return ret; } /** - * bio_copy_user - copy user data to bio + * bio_copy_user_iov - copy user data to bio * @q: destination block queue - * @uaddr: start of user address - * @len: length in bytes + * @iov: the iovec. + * @iov_count: number of elements in the iovec * @write_to_vm: bool indicating writing to pages or not * * Prepares and returns a bio for indirect user io, bouncing data * to/from kernel pages as necessary. Must be paired with * call bio_uncopy_user() on io completion. */ -struct bio *bio_copy_user(struct request_queue *q, unsigned long uaddr, - unsigned int len, int write_to_vm) +struct bio *bio_copy_user_iov(struct request_queue *q, struct sg_iovec *iov, + int iov_count, int write_to_vm) { - unsigned long end = (uaddr + len + PAGE_SIZE - 1) >> PAGE_SHIFT; - unsigned long start = uaddr >> PAGE_SHIFT; struct bio_map_data *bmd; struct bio_vec *bvec; struct page *page; struct bio *bio; int i, ret; + int nr_pages = 0; + unsigned int len = 0; - bmd = bio_alloc_map_data(end - start); + for (i = 0; i < iov_count; i++) { + unsigned long uaddr; + unsigned long end; + unsigned long start; + + uaddr = (unsigned long)iov[i].iov_base; + end = (uaddr + iov[i].iov_len + PAGE_SIZE - 1) >> PAGE_SHIFT; + start = uaddr >> PAGE_SHIFT; + + nr_pages += end - start; + len += iov[i].iov_len; + } + + bmd = bio_alloc_map_data(nr_pages, iov_count); if (!bmd) return ERR_PTR(-ENOMEM); - bmd->userptr = (void __user *) uaddr; - ret = -ENOMEM; - bio = bio_alloc(GFP_KERNEL, end - start); + bio = bio_alloc(GFP_KERNEL, nr_pages); if (!bio) goto out_bmd; @@ -564,22 +628,12 @@ struct bio *bio_copy_user(struct request_queue *q, unsigned long uaddr, * success */ if (!write_to_vm) { - char __user *p = (char __user *) uaddr; - - /* - * for a write, copy in data to kernel pages - */ - ret = -EFAULT; - bio_for_each_segment(bvec, bio, i) { - char *addr = page_address(bvec->bv_page); - - if (copy_from_user(addr, p, bvec->bv_len)) - goto cleanup; - p += bvec->bv_len; - } + ret = __bio_copy_iov(bio, iov, iov_count, 0); + if (ret) + goto cleanup; } - bio_set_map_data(bmd, bio); + bio_set_map_data(bmd, bio, iov, iov_count); return bio; cleanup: bio_for_each_segment(bvec, bio, i) @@ -591,6 +645,28 @@ out_bmd: return ERR_PTR(ret); } +/** + * bio_copy_user - copy user data to bio + * @q: destination block queue + * @uaddr: start of user address + * @len: length in bytes + * @write_to_vm: bool indicating writing to pages or not + * + * Prepares and returns a bio for indirect user io, bouncing data + * to/from kernel pages as necessary. Must be paired with + * call bio_uncopy_user() on io completion. + */ +struct bio *bio_copy_user(struct request_queue *q, unsigned long uaddr, + unsigned int len, int write_to_vm) +{ + struct sg_iovec iov; + + iov.iov_base = (void __user *)uaddr; + iov.iov_len = len; + + return bio_copy_user_iov(q, &iov, 1, write_to_vm); +} + static struct bio *__bio_map_user_iov(struct request_queue *q, struct block_device *bdev, struct sg_iovec *iov, int iov_count, -- cgit v1.2.3 From 5416b704aef0b7350073421f4f6ac1a21bc213e7 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 13 Feb 2008 16:54:29 -0800 Subject: dlm: match signedness between dlm_config_info and cluster_set cluster_set is only called from the macro CLUSTER_ATTR which defines read/write access functions. Make the signedness match to avoid sparse warnings every time CLUSTER_ATTR is used (lines 149-159) all of the form: fs/dlm/config.c:149:1: warning: incorrect type in argument 3 (different signedness) fs/dlm/config.c:149:1: expected unsigned int *info_field fs/dlm/config.c:149:1: got int extern [toplevel] * Signed-off-by: Harvey Harrison Signed-off-by: David Teigland --- fs/dlm/config.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/dlm/config.c b/fs/dlm/config.c index c3ad1dff3b2..7ceaea3d983 100644 --- a/fs/dlm/config.c +++ b/fs/dlm/config.c @@ -114,7 +114,7 @@ struct cluster_attribute { }; static ssize_t cluster_set(struct cluster *cl, unsigned int *cl_field, - unsigned int *info_field, int check_zero, + int *info_field, int check_zero, const char *buf, size_t len) { unsigned int x; -- cgit v1.2.3 From 170e19ab2900b7c959d7a0e627fd12f383efcfa1 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Wed, 13 Feb 2008 23:29:38 +0200 Subject: dlm: make dlm_print_rsb() static dlm_print_rsb() can now become static. Signed-off-by: Adrian Bunk Signed-off-by: David Teigland --- fs/dlm/lock.c | 2 +- fs/dlm/lock.h | 1 - 2 files changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/dlm/lock.c b/fs/dlm/lock.c index 8f250ac8b92..1e9e8ebee25 100644 --- a/fs/dlm/lock.c +++ b/fs/dlm/lock.c @@ -165,7 +165,7 @@ void dlm_print_lkb(struct dlm_lkb *lkb) lkb->lkb_grmode, lkb->lkb_wait_type, lkb->lkb_ast_type); } -void dlm_print_rsb(struct dlm_rsb *r) +static void dlm_print_rsb(struct dlm_rsb *r) { printk(KERN_ERR "rsb: nodeid %d flags %lx first %x rlc %d name %s\n", r->res_nodeid, r->res_flags, r->res_first_lkid, diff --git a/fs/dlm/lock.h b/fs/dlm/lock.h index 05d9c82e646..88e93c80cc2 100644 --- a/fs/dlm/lock.h +++ b/fs/dlm/lock.h @@ -13,7 +13,6 @@ #ifndef __LOCK_DOT_H__ #define __LOCK_DOT_H__ -void dlm_print_rsb(struct dlm_rsb *r); void dlm_dump_rsb(struct dlm_rsb *r); void dlm_print_lkb(struct dlm_lkb *lkb); void dlm_receive_message_saved(struct dlm_ls *ls, struct dlm_message *ms); -- cgit v1.2.3 From 761b9d3ffc953c24ceb55d8e12ff7e02b17e0484 Mon Sep 17 00:00:00 2001 From: David Teigland Date: Thu, 21 Feb 2008 11:25:42 -0600 Subject: dlm: save master info after failed no-queue request When a NOQUEUE request fails, the rsb res_master field is unnecessarily reset to -1, instead of leaving the valid master setting in place. We want to save the looked-up master values while the rsb is on the "toss list" so that another lookup can be avoided if the rsb is soon reused. The fix is to simply leave res_master value alone. Signed-off-by: David Teigland --- fs/dlm/lock.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/dlm/lock.c b/fs/dlm/lock.c index 1e9e8ebee25..2d3d1027ce2 100644 --- a/fs/dlm/lock.c +++ b/fs/dlm/lock.c @@ -1956,8 +1956,7 @@ static void confirm_master(struct dlm_rsb *r, int error) list_del_init(&lkb->lkb_rsb_lookup); r->res_first_lkid = lkb->lkb_id; _request_lock(r, lkb); - } else - r->res_nodeid = -1; + } break; default: -- cgit v1.2.3 From d44e0fc704143624b3e88fbf8fbcfda7a83fd299 Mon Sep 17 00:00:00 2001 From: David Teigland Date: Tue, 18 Mar 2008 14:22:11 -0500 Subject: dlm: recover nodes that are removed and re-added If a node is removed from a lockspace, and then added back before the dlm is notified of the removal, the dlm will not detect the removal and won't clear the old state from the node. This is fixed by using a list of added nodes so the membership recovery can detect when a newly added node is already in the member list. Signed-off-by: David Teigland --- fs/dlm/config.c | 48 +++++++++++++++++++++++++++++++++++++++--------- fs/dlm/config.h | 3 ++- fs/dlm/dlm_internal.h | 4 +++- fs/dlm/member.c | 34 +++++++++++++++++++++++++++++----- fs/dlm/recoverd.c | 1 + 5 files changed, 74 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/dlm/config.c b/fs/dlm/config.c index 7ceaea3d983..eac23bd288b 100644 --- a/fs/dlm/config.c +++ b/fs/dlm/config.c @@ -284,6 +284,7 @@ struct node { struct list_head list; /* space->members */ int nodeid; int weight; + int new; }; static struct configfs_group_operations clusters_ops = { @@ -565,6 +566,7 @@ static struct config_item *make_node(struct config_group *g, const char *name) config_item_init_type_name(&nd->item, name, &node_type); nd->nodeid = -1; nd->weight = 1; /* default weight of 1 if none is set */ + nd->new = 1; /* set to 0 once it's been read by dlm_nodeid_list() */ mutex_lock(&sp->members_lock); list_add(&nd->list, &sp->members); @@ -805,12 +807,13 @@ static void put_comm(struct comm *cm) } /* caller must free mem */ -int dlm_nodeid_list(char *lsname, int **ids_out) +int dlm_nodeid_list(char *lsname, int **ids_out, int *ids_count_out, + int **new_out, int *new_count_out) { struct space *sp; struct node *nd; - int i = 0, rv = 0; - int *ids; + int i = 0, rv = 0, ids_count = 0, new_count = 0; + int *ids, *new; sp = get_space(lsname); if (!sp) @@ -818,23 +821,50 @@ int dlm_nodeid_list(char *lsname, int **ids_out) mutex_lock(&sp->members_lock); if (!sp->members_count) { - rv = 0; + rv = -EINVAL; + printk(KERN_ERR "dlm: zero members_count\n"); goto out; } - ids = kcalloc(sp->members_count, sizeof(int), GFP_KERNEL); + ids_count = sp->members_count; + + ids = kcalloc(ids_count, sizeof(int), GFP_KERNEL); if (!ids) { rv = -ENOMEM; goto out; } - rv = sp->members_count; - list_for_each_entry(nd, &sp->members, list) + list_for_each_entry(nd, &sp->members, list) { ids[i++] = nd->nodeid; + if (nd->new) + new_count++; + } + + if (ids_count != i) + printk(KERN_ERR "dlm: bad nodeid count %d %d\n", ids_count, i); + + if (!new_count) + goto out_ids; + + new = kcalloc(new_count, sizeof(int), GFP_KERNEL); + if (!new) { + kfree(ids); + rv = -ENOMEM; + goto out; + } - if (rv != i) - printk("bad nodeid count %d %d\n", rv, i); + i = 0; + list_for_each_entry(nd, &sp->members, list) { + if (nd->new) { + new[i++] = nd->nodeid; + nd->new = 0; + } + } + *new_count_out = new_count; + *new_out = new; + out_ids: + *ids_count_out = ids_count; *ids_out = ids; out: mutex_unlock(&sp->members_lock); diff --git a/fs/dlm/config.h b/fs/dlm/config.h index a3170fe2209..4f1d6fce58c 100644 --- a/fs/dlm/config.h +++ b/fs/dlm/config.h @@ -35,7 +35,8 @@ extern struct dlm_config_info dlm_config; int dlm_config_init(void); void dlm_config_exit(void); int dlm_node_weight(char *lsname, int nodeid); -int dlm_nodeid_list(char *lsname, int **ids_out); +int dlm_nodeid_list(char *lsname, int **ids_out, int *ids_count_out, + int **new_out, int *new_count_out); int dlm_nodeid_to_addr(int nodeid, struct sockaddr_storage *addr); int dlm_addr_to_nodeid(struct sockaddr_storage *addr, int *nodeid); int dlm_our_nodeid(void); diff --git a/fs/dlm/dlm_internal.h b/fs/dlm/dlm_internal.h index d30ea8b433a..c70c8e58358 100644 --- a/fs/dlm/dlm_internal.h +++ b/fs/dlm/dlm_internal.h @@ -133,8 +133,10 @@ struct dlm_member { struct dlm_recover { struct list_head list; - int *nodeids; + int *nodeids; /* nodeids of all members */ int node_count; + int *new; /* nodeids of new members */ + int new_count; uint64_t seq; }; diff --git a/fs/dlm/member.c b/fs/dlm/member.c index fa17f5a2788..26133f05ae3 100644 --- a/fs/dlm/member.c +++ b/fs/dlm/member.c @@ -210,6 +210,23 @@ int dlm_recover_members(struct dlm_ls *ls, struct dlm_recover *rv, int *neg_out) } } + /* Add an entry to ls_nodes_gone for members that were removed and + then added again, so that previous state for these nodes will be + cleared during recovery. */ + + for (i = 0; i < rv->new_count; i++) { + if (!dlm_is_member(ls, rv->new[i])) + continue; + log_debug(ls, "new nodeid %d is a re-added member", rv->new[i]); + + memb = kzalloc(sizeof(struct dlm_member), GFP_KERNEL); + if (!memb) + return -ENOMEM; + memb->nodeid = rv->new[i]; + list_add_tail(&memb->list, &ls->ls_nodes_gone); + neg++; + } + /* add new members to ls_nodes */ for (i = 0; i < rv->node_count; i++) { @@ -314,15 +331,16 @@ int dlm_ls_stop(struct dlm_ls *ls) int dlm_ls_start(struct dlm_ls *ls) { struct dlm_recover *rv = NULL, *rv_old; - int *ids = NULL; - int error, count; + int *ids = NULL, *new = NULL; + int error, ids_count = 0, new_count = 0; rv = kzalloc(sizeof(struct dlm_recover), GFP_KERNEL); if (!rv) return -ENOMEM; - error = count = dlm_nodeid_list(ls->ls_name, &ids); - if (error <= 0) + error = dlm_nodeid_list(ls->ls_name, &ids, &ids_count, + &new, &new_count); + if (error < 0) goto fail; spin_lock(&ls->ls_recover_lock); @@ -337,14 +355,19 @@ int dlm_ls_start(struct dlm_ls *ls) } rv->nodeids = ids; - rv->node_count = count; + rv->node_count = ids_count; + rv->new = new; + rv->new_count = new_count; rv->seq = ++ls->ls_recover_seq; rv_old = ls->ls_recover_args; ls->ls_recover_args = rv; spin_unlock(&ls->ls_recover_lock); if (rv_old) { + log_error(ls, "unused recovery %llx %d", + (unsigned long long)rv_old->seq, rv_old->node_count); kfree(rv_old->nodeids); + kfree(rv_old->new); kfree(rv_old); } @@ -354,6 +377,7 @@ int dlm_ls_start(struct dlm_ls *ls) fail: kfree(rv); kfree(ids); + kfree(new); return error; } diff --git a/fs/dlm/recoverd.c b/fs/dlm/recoverd.c index 997f9531d59..fd677c8c3d3 100644 --- a/fs/dlm/recoverd.c +++ b/fs/dlm/recoverd.c @@ -257,6 +257,7 @@ static void do_ls_recovery(struct dlm_ls *ls) if (rv) { ls_recover(ls, rv); kfree(rv->nodeids); + kfree(rv->new); kfree(rv); } } -- cgit v1.2.3 From 2402211a8389282fd2942fad4511f02c0eeeffc5 Mon Sep 17 00:00:00 2001 From: David Teigland Date: Fri, 14 Mar 2008 15:09:15 -0500 Subject: dlm: move plock code from gfs2 Move the code that handles cluster posix locks from gfs2 into the dlm so that it can be used by both gfs2 and ocfs2. Signed-off-by: David Teigland --- fs/dlm/Makefile | 1 + fs/dlm/dlm_internal.h | 2 + fs/dlm/main.c | 7 + fs/dlm/plock.c | 439 +++++++++++++++++++++++++++++++++++++++++ fs/gfs2/locking/dlm/Makefile | 2 +- fs/gfs2/locking/dlm/lock_dlm.h | 12 +- fs/gfs2/locking/dlm/main.c | 8 - fs/gfs2/locking/dlm/mount.c | 21 ++ fs/gfs2/locking/dlm/plock.c | 406 ------------------------------------- 9 files changed, 472 insertions(+), 426 deletions(-) create mode 100644 fs/dlm/plock.c delete mode 100644 fs/gfs2/locking/dlm/plock.c (limited to 'fs') diff --git a/fs/dlm/Makefile b/fs/dlm/Makefile index d248e60951b..ca1c9124c8c 100644 --- a/fs/dlm/Makefile +++ b/fs/dlm/Makefile @@ -10,6 +10,7 @@ dlm-y := ast.o \ midcomms.o \ netlink.o \ lowcomms.o \ + plock.o \ rcom.o \ recover.o \ recoverd.o \ diff --git a/fs/dlm/dlm_internal.h b/fs/dlm/dlm_internal.h index c70c8e58358..caa1581e158 100644 --- a/fs/dlm/dlm_internal.h +++ b/fs/dlm/dlm_internal.h @@ -582,6 +582,8 @@ static inline int dlm_no_directory(struct dlm_ls *ls) int dlm_netlink_init(void); void dlm_netlink_exit(void); void dlm_timeout_warn(struct dlm_lkb *lkb); +int dlm_plock_init(void); +void dlm_plock_exit(void); #ifdef CONFIG_DLM_DEBUG int dlm_register_debugfs(void); diff --git a/fs/dlm/main.c b/fs/dlm/main.c index 58487fb95a4..b80e0aa3cfa 100644 --- a/fs/dlm/main.c +++ b/fs/dlm/main.c @@ -46,10 +46,16 @@ static int __init init_dlm(void) if (error) goto out_user; + error = dlm_plock_init(); + if (error) + goto out_netlink; + printk("DLM (built %s %s) installed\n", __DATE__, __TIME__); return 0; + out_netlink: + dlm_netlink_exit(); out_user: dlm_user_exit(); out_debug: @@ -66,6 +72,7 @@ static int __init init_dlm(void) static void __exit exit_dlm(void) { + dlm_plock_exit(); dlm_netlink_exit(); dlm_user_exit(); dlm_config_exit(); diff --git a/fs/dlm/plock.c b/fs/dlm/plock.c new file mode 100644 index 00000000000..d6d6e370f89 --- /dev/null +++ b/fs/dlm/plock.c @@ -0,0 +1,439 @@ +/* + * Copyright (C) 2005-2008 Red Hat, Inc. All rights reserved. + * + * This copyrighted material is made available to anyone wishing to use, + * modify, copy, or redistribute it subject to the terms and conditions + * of the GNU General Public License version 2. + */ + +#include +#include +#include +#include +#include + +#include "dlm_internal.h" +#include "lockspace.h" + +static spinlock_t ops_lock; +static struct list_head send_list; +static struct list_head recv_list; +static wait_queue_head_t send_wq; +static wait_queue_head_t recv_wq; + +struct plock_op { + struct list_head list; + int done; + struct dlm_plock_info info; +}; + +struct plock_xop { + struct plock_op xop; + void *callback; + void *fl; + void *file; + struct file_lock flc; +}; + + +static inline void set_version(struct dlm_plock_info *info) +{ + info->version[0] = DLM_PLOCK_VERSION_MAJOR; + info->version[1] = DLM_PLOCK_VERSION_MINOR; + info->version[2] = DLM_PLOCK_VERSION_PATCH; +} + +static int check_version(struct dlm_plock_info *info) +{ + if ((DLM_PLOCK_VERSION_MAJOR != info->version[0]) || + (DLM_PLOCK_VERSION_MINOR < info->version[1])) { + log_print("plock device version mismatch: " + "kernel (%u.%u.%u), user (%u.%u.%u)", + DLM_PLOCK_VERSION_MAJOR, + DLM_PLOCK_VERSION_MINOR, + DLM_PLOCK_VERSION_PATCH, + info->version[0], + info->version[1], + info->version[2]); + return -EINVAL; + } + return 0; +} + +static void send_op(struct plock_op *op) +{ + set_version(&op->info); + INIT_LIST_HEAD(&op->list); + spin_lock(&ops_lock); + list_add_tail(&op->list, &send_list); + spin_unlock(&ops_lock); + wake_up(&send_wq); +} + +int dlm_posix_lock(dlm_lockspace_t *lockspace, u64 number, struct file *file, + int cmd, struct file_lock *fl) +{ + struct dlm_ls *ls; + struct plock_op *op; + struct plock_xop *xop; + int rv; + + ls = dlm_find_lockspace_local(lockspace); + if (!ls) + return -EINVAL; + + xop = kzalloc(sizeof(*xop), GFP_KERNEL); + if (!xop) { + rv = -ENOMEM; + goto out; + } + + op = &xop->xop; + op->info.optype = DLM_PLOCK_OP_LOCK; + op->info.pid = fl->fl_pid; + op->info.ex = (fl->fl_type == F_WRLCK); + op->info.wait = IS_SETLKW(cmd); + op->info.fsid = ls->ls_global_id; + op->info.number = number; + op->info.start = fl->fl_start; + op->info.end = fl->fl_end; + if (fl->fl_lmops && fl->fl_lmops->fl_grant) { + /* fl_owner is lockd which doesn't distinguish + processes on the nfs client */ + op->info.owner = (__u64) fl->fl_pid; + xop->callback = fl->fl_lmops->fl_grant; + locks_init_lock(&xop->flc); + locks_copy_lock(&xop->flc, fl); + xop->fl = fl; + xop->file = file; + } else { + op->info.owner = (__u64)(long) fl->fl_owner; + xop->callback = NULL; + } + + send_op(op); + + if (xop->callback == NULL) + wait_event(recv_wq, (op->done != 0)); + else { + rv = -EINPROGRESS; + goto out; + } + + spin_lock(&ops_lock); + if (!list_empty(&op->list)) { + log_error(ls, "dlm_posix_lock: op on list %llx", + (unsigned long long)number); + list_del(&op->list); + } + spin_unlock(&ops_lock); + + rv = op->info.rv; + + if (!rv) { + if (posix_lock_file_wait(file, fl) < 0) + log_error(ls, "dlm_posix_lock: vfs lock error %llx", + (unsigned long long)number); + } + + kfree(xop); +out: + dlm_put_lockspace(ls); + return rv; +} +EXPORT_SYMBOL_GPL(dlm_posix_lock); + +/* Returns failure iff a succesful lock operation should be canceled */ +static int dlm_plock_callback(struct plock_op *op) +{ + struct file *file; + struct file_lock *fl; + struct file_lock *flc; + int (*notify)(void *, void *, int) = NULL; + struct plock_xop *xop = (struct plock_xop *)op; + int rv = 0; + + spin_lock(&ops_lock); + if (!list_empty(&op->list)) { + log_print("dlm_plock_callback: op on list %llx", + (unsigned long long)op->info.number); + list_del(&op->list); + } + spin_unlock(&ops_lock); + + /* check if the following 2 are still valid or make a copy */ + file = xop->file; + flc = &xop->flc; + fl = xop->fl; + notify = xop->callback; + + if (op->info.rv) { + notify(flc, NULL, op->info.rv); + goto out; + } + + /* got fs lock; bookkeep locally as well: */ + flc->fl_flags &= ~FL_SLEEP; + if (posix_lock_file(file, flc, NULL)) { + /* + * This can only happen in the case of kmalloc() failure. + * The filesystem's own lock is the authoritative lock, + * so a failure to get the lock locally is not a disaster. + * As long as the fs cannot reliably cancel locks (especially + * in a low-memory situation), we're better off ignoring + * this failure than trying to recover. + */ + log_print("dlm_plock_callback: vfs lock error %llx file %p fl %p", + (unsigned long long)op->info.number, file, fl); + } + + rv = notify(flc, NULL, 0); + if (rv) { + /* XXX: We need to cancel the fs lock here: */ + log_print("dlm_plock_callback: lock granted after lock request " + "failed; dangling lock!\n"); + goto out; + } + +out: + kfree(xop); + return rv; +} + +int dlm_posix_unlock(dlm_lockspace_t *lockspace, u64 number, struct file *file, + struct file_lock *fl) +{ + struct dlm_ls *ls; + struct plock_op *op; + int rv; + + ls = dlm_find_lockspace_local(lockspace); + if (!ls) + return -EINVAL; + + op = kzalloc(sizeof(*op), GFP_KERNEL); + if (!op) { + rv = -ENOMEM; + goto out; + } + + if (posix_lock_file_wait(file, fl) < 0) + log_error(ls, "dlm_posix_unlock: vfs unlock error %llx", + (unsigned long long)number); + + op->info.optype = DLM_PLOCK_OP_UNLOCK; + op->info.pid = fl->fl_pid; + op->info.fsid = ls->ls_global_id; + op->info.number = number; + op->info.start = fl->fl_start; + op->info.end = fl->fl_end; + if (fl->fl_lmops && fl->fl_lmops->fl_grant) + op->info.owner = (__u64) fl->fl_pid; + else + op->info.owner = (__u64)(long) fl->fl_owner; + + send_op(op); + wait_event(recv_wq, (op->done != 0)); + + spin_lock(&ops_lock); + if (!list_empty(&op->list)) { + log_error(ls, "dlm_posix_unlock: op on list %llx", + (unsigned long long)number); + list_del(&op->list); + } + spin_unlock(&ops_lock); + + rv = op->info.rv; + + if (rv == -ENOENT) + rv = 0; + + kfree(op); +out: + dlm_put_lockspace(ls); + return rv; +} +EXPORT_SYMBOL_GPL(dlm_posix_unlock); + +int dlm_posix_get(dlm_lockspace_t *lockspace, u64 number, struct file *file, + struct file_lock *fl) +{ + struct dlm_ls *ls; + struct plock_op *op; + int rv; + + ls = dlm_find_lockspace_local(lockspace); + if (!ls) + return -EINVAL; + + op = kzalloc(sizeof(*op), GFP_KERNEL); + if (!op) { + rv = -ENOMEM; + goto out; + } + + op->info.optype = DLM_PLOCK_OP_GET; + op->info.pid = fl->fl_pid; + op->info.ex = (fl->fl_type == F_WRLCK); + op->info.fsid = ls->ls_global_id; + op->info.number = number; + op->info.start = fl->fl_start; + op->info.end = fl->fl_end; + if (fl->fl_lmops && fl->fl_lmops->fl_grant) + op->info.owner = (__u64) fl->fl_pid; + else + op->info.owner = (__u64)(long) fl->fl_owner; + + send_op(op); + wait_event(recv_wq, (op->done != 0)); + + spin_lock(&ops_lock); + if (!list_empty(&op->list)) { + log_error(ls, "dlm_posix_get: op on list %llx", + (unsigned long long)number); + list_del(&op->list); + } + spin_unlock(&ops_lock); + + /* info.rv from userspace is 1 for conflict, 0 for no-conflict, + -ENOENT if there are no locks on the file */ + + rv = op->info.rv; + + fl->fl_type = F_UNLCK; + if (rv == -ENOENT) + rv = 0; + else if (rv > 0) { + fl->fl_type = (op->info.ex) ? F_WRLCK : F_RDLCK; + fl->fl_pid = op->info.pid; + fl->fl_start = op->info.start; + fl->fl_end = op->info.end; + rv = 0; + } + + kfree(op); +out: + dlm_put_lockspace(ls); + return rv; +} +EXPORT_SYMBOL_GPL(dlm_posix_get); + +/* a read copies out one plock request from the send list */ +static ssize_t dev_read(struct file *file, char __user *u, size_t count, + loff_t *ppos) +{ + struct dlm_plock_info info; + struct plock_op *op = NULL; + + if (count < sizeof(info)) + return -EINVAL; + + spin_lock(&ops_lock); + if (!list_empty(&send_list)) { + op = list_entry(send_list.next, struct plock_op, list); + list_move(&op->list, &recv_list); + memcpy(&info, &op->info, sizeof(info)); + } + spin_unlock(&ops_lock); + + if (!op) + return -EAGAIN; + + if (copy_to_user(u, &info, sizeof(info))) + return -EFAULT; + return sizeof(info); +} + +/* a write copies in one plock result that should match a plock_op + on the recv list */ +static ssize_t dev_write(struct file *file, const char __user *u, size_t count, + loff_t *ppos) +{ + struct dlm_plock_info info; + struct plock_op *op; + int found = 0; + + if (count != sizeof(info)) + return -EINVAL; + + if (copy_from_user(&info, u, sizeof(info))) + return -EFAULT; + + if (check_version(&info)) + return -EINVAL; + + spin_lock(&ops_lock); + list_for_each_entry(op, &recv_list, list) { + if (op->info.fsid == info.fsid && op->info.number == info.number && + op->info.owner == info.owner) { + list_del_init(&op->list); + found = 1; + op->done = 1; + memcpy(&op->info, &info, sizeof(info)); + break; + } + } + spin_unlock(&ops_lock); + + if (found) { + struct plock_xop *xop; + xop = (struct plock_xop *)op; + if (xop->callback) + count = dlm_plock_callback(op); + else + wake_up(&recv_wq); + } else + log_print("dev_write no op %x %llx", info.fsid, + (unsigned long long)info.number); + return count; +} + +static unsigned int dev_poll(struct file *file, poll_table *wait) +{ + unsigned int mask = 0; + + poll_wait(file, &send_wq, wait); + + spin_lock(&ops_lock); + if (!list_empty(&send_list)) + mask = POLLIN | POLLRDNORM; + spin_unlock(&ops_lock); + + return mask; +} + +static const struct file_operations dev_fops = { + .read = dev_read, + .write = dev_write, + .poll = dev_poll, + .owner = THIS_MODULE +}; + +static struct miscdevice plock_dev_misc = { + .minor = MISC_DYNAMIC_MINOR, + .name = DLM_PLOCK_MISC_NAME, + .fops = &dev_fops +}; + +int dlm_plock_init(void) +{ + int rv; + + spin_lock_init(&ops_lock); + INIT_LIST_HEAD(&send_list); + INIT_LIST_HEAD(&recv_list); + init_waitqueue_head(&send_wq); + init_waitqueue_head(&recv_wq); + + rv = misc_register(&plock_dev_misc); + if (rv) + log_print("dlm_plock_init: misc_register failed %d", rv); + return rv; +} + +void dlm_plock_exit(void) +{ + if (misc_deregister(&plock_dev_misc) < 0) + log_print("dlm_plock_exit: misc_deregister failed"); +} + diff --git a/fs/gfs2/locking/dlm/Makefile b/fs/gfs2/locking/dlm/Makefile index 89b93b6b45c..2609bb6cd01 100644 --- a/fs/gfs2/locking/dlm/Makefile +++ b/fs/gfs2/locking/dlm/Makefile @@ -1,3 +1,3 @@ obj-$(CONFIG_GFS2_FS_LOCKING_DLM) += lock_dlm.o -lock_dlm-y := lock.o main.o mount.o sysfs.o thread.o plock.o +lock_dlm-y := lock.o main.o mount.o sysfs.o thread.o diff --git a/fs/gfs2/locking/dlm/lock_dlm.h b/fs/gfs2/locking/dlm/lock_dlm.h index 58fcf8c5bf3..a243cf69c54 100644 --- a/fs/gfs2/locking/dlm/lock_dlm.h +++ b/fs/gfs2/locking/dlm/lock_dlm.h @@ -25,6 +25,7 @@ #include #include +#include #include /* @@ -173,17 +174,6 @@ void gdlm_cancel(void *); int gdlm_hold_lvb(void *, char **); void gdlm_unhold_lvb(void *, char *); -/* plock.c */ - -int gdlm_plock_init(void); -void gdlm_plock_exit(void); -int gdlm_plock(void *, struct lm_lockname *, struct file *, int, - struct file_lock *); -int gdlm_plock_get(void *, struct lm_lockname *, struct file *, - struct file_lock *); -int gdlm_punlock(void *, struct lm_lockname *, struct file *, - struct file_lock *); - /* mount.c */ extern const struct lm_lockops gdlm_ops; diff --git a/fs/gfs2/locking/dlm/main.c b/fs/gfs2/locking/dlm/main.c index 36a225850bd..b9a03a7ff80 100644 --- a/fs/gfs2/locking/dlm/main.c +++ b/fs/gfs2/locking/dlm/main.c @@ -28,13 +28,6 @@ static int __init init_lock_dlm(void) return error; } - error = gdlm_plock_init(); - if (error) { - gdlm_sysfs_exit(); - gfs2_unregister_lockproto(&gdlm_ops); - return error; - } - printk(KERN_INFO "Lock_DLM (built %s %s) installed\n", __DATE__, __TIME__); return 0; @@ -42,7 +35,6 @@ static int __init init_lock_dlm(void) static void __exit exit_lock_dlm(void) { - gdlm_plock_exit(); gdlm_sysfs_exit(); gfs2_unregister_lockproto(&gdlm_ops); } diff --git a/fs/gfs2/locking/dlm/mount.c b/fs/gfs2/locking/dlm/mount.c index f2efff42422..470bdf650b5 100644 --- a/fs/gfs2/locking/dlm/mount.c +++ b/fs/gfs2/locking/dlm/mount.c @@ -236,6 +236,27 @@ static void gdlm_withdraw(void *lockspace) gdlm_kobject_release(ls); } +static int gdlm_plock(void *lockspace, struct lm_lockname *name, + struct file *file, int cmd, struct file_lock *fl) +{ + struct gdlm_ls *ls = lockspace; + return dlm_posix_lock(ls->dlm_lockspace, name->ln_number, file, cmd, fl); +} + +static int gdlm_punlock(void *lockspace, struct lm_lockname *name, + struct file *file, struct file_lock *fl) +{ + struct gdlm_ls *ls = lockspace; + return dlm_posix_unlock(ls->dlm_lockspace, name->ln_number, file, fl); +} + +static int gdlm_plock_get(void *lockspace, struct lm_lockname *name, + struct file *file, struct file_lock *fl) +{ + struct gdlm_ls *ls = lockspace; + return dlm_posix_get(ls->dlm_lockspace, name->ln_number, file, fl); +} + const struct lm_lockops gdlm_ops = { .lm_proto_name = "lock_dlm", .lm_mount = gdlm_mount, diff --git a/fs/gfs2/locking/dlm/plock.c b/fs/gfs2/locking/dlm/plock.c deleted file mode 100644 index 2ebd374b314..00000000000 --- a/fs/gfs2/locking/dlm/plock.c +++ /dev/null @@ -1,406 +0,0 @@ -/* - * Copyright (C) 2005 Red Hat, Inc. All rights reserved. - * - * This copyrighted material is made available to anyone wishing to use, - * modify, copy, or redistribute it subject to the terms and conditions - * of the GNU General Public License version 2. - */ - -#include -#include -#include - -#include "lock_dlm.h" - - -static spinlock_t ops_lock; -static struct list_head send_list; -static struct list_head recv_list; -static wait_queue_head_t send_wq; -static wait_queue_head_t recv_wq; - -struct plock_op { - struct list_head list; - int done; - struct gdlm_plock_info info; -}; - -struct plock_xop { - struct plock_op xop; - void *callback; - void *fl; - void *file; - struct file_lock flc; -}; - - -static inline void set_version(struct gdlm_plock_info *info) -{ - info->version[0] = GDLM_PLOCK_VERSION_MAJOR; - info->version[1] = GDLM_PLOCK_VERSION_MINOR; - info->version[2] = GDLM_PLOCK_VERSION_PATCH; -} - -static int check_version(struct gdlm_plock_info *info) -{ - if ((GDLM_PLOCK_VERSION_MAJOR != info->version[0]) || - (GDLM_PLOCK_VERSION_MINOR < info->version[1])) { - log_error("plock device version mismatch: " - "kernel (%u.%u.%u), user (%u.%u.%u)", - GDLM_PLOCK_VERSION_MAJOR, - GDLM_PLOCK_VERSION_MINOR, - GDLM_PLOCK_VERSION_PATCH, - info->version[0], - info->version[1], - info->version[2]); - return -EINVAL; - } - return 0; -} - -static void send_op(struct plock_op *op) -{ - set_version(&op->info); - INIT_LIST_HEAD(&op->list); - spin_lock(&ops_lock); - list_add_tail(&op->list, &send_list); - spin_unlock(&ops_lock); - wake_up(&send_wq); -} - -int gdlm_plock(void *lockspace, struct lm_lockname *name, - struct file *file, int cmd, struct file_lock *fl) -{ - struct gdlm_ls *ls = lockspace; - struct plock_op *op; - struct plock_xop *xop; - int rv; - - xop = kzalloc(sizeof(*xop), GFP_KERNEL); - if (!xop) - return -ENOMEM; - - op = &xop->xop; - op->info.optype = GDLM_PLOCK_OP_LOCK; - op->info.pid = fl->fl_pid; - op->info.ex = (fl->fl_type == F_WRLCK); - op->info.wait = IS_SETLKW(cmd); - op->info.fsid = ls->id; - op->info.number = name->ln_number; - op->info.start = fl->fl_start; - op->info.end = fl->fl_end; - if (fl->fl_lmops && fl->fl_lmops->fl_grant) { - /* fl_owner is lockd which doesn't distinguish - processes on the nfs client */ - op->info.owner = (__u64) fl->fl_pid; - xop->callback = fl->fl_lmops->fl_grant; - locks_init_lock(&xop->flc); - locks_copy_lock(&xop->flc, fl); - xop->fl = fl; - xop->file = file; - } else { - op->info.owner = (__u64)(long) fl->fl_owner; - xop->callback = NULL; - } - - send_op(op); - - if (xop->callback == NULL) - wait_event(recv_wq, (op->done != 0)); - else - return -EINPROGRESS; - - spin_lock(&ops_lock); - if (!list_empty(&op->list)) { - printk(KERN_INFO "plock op on list\n"); - list_del(&op->list); - } - spin_unlock(&ops_lock); - - rv = op->info.rv; - - if (!rv) { - if (posix_lock_file_wait(file, fl) < 0) - log_error("gdlm_plock: vfs lock error %x,%llx", - name->ln_type, - (unsigned long long)name->ln_number); - } - - kfree(xop); - return rv; -} - -/* Returns failure iff a succesful lock operation should be canceled */ -static int gdlm_plock_callback(struct plock_op *op) -{ - struct file *file; - struct file_lock *fl; - struct file_lock *flc; - int (*notify)(void *, void *, int) = NULL; - struct plock_xop *xop = (struct plock_xop *)op; - int rv = 0; - - spin_lock(&ops_lock); - if (!list_empty(&op->list)) { - printk(KERN_INFO "plock op on list\n"); - list_del(&op->list); - } - spin_unlock(&ops_lock); - - /* check if the following 2 are still valid or make a copy */ - file = xop->file; - flc = &xop->flc; - fl = xop->fl; - notify = xop->callback; - - if (op->info.rv) { - notify(flc, NULL, op->info.rv); - goto out; - } - - /* got fs lock; bookkeep locally as well: */ - flc->fl_flags &= ~FL_SLEEP; - if (posix_lock_file(file, flc, NULL)) { - /* - * This can only happen in the case of kmalloc() failure. - * The filesystem's own lock is the authoritative lock, - * so a failure to get the lock locally is not a disaster. - * As long as GFS cannot reliably cancel locks (especially - * in a low-memory situation), we're better off ignoring - * this failure than trying to recover. - */ - log_error("gdlm_plock: vfs lock error file %p fl %p", - file, fl); - } - - rv = notify(flc, NULL, 0); - if (rv) { - /* XXX: We need to cancel the fs lock here: */ - printk("gfs2 lock granted after lock request failed;" - " dangling lock!\n"); - goto out; - } - -out: - kfree(xop); - return rv; -} - -int gdlm_punlock(void *lockspace, struct lm_lockname *name, - struct file *file, struct file_lock *fl) -{ - struct gdlm_ls *ls = lockspace; - struct plock_op *op; - int rv; - - op = kzalloc(sizeof(*op), GFP_KERNEL); - if (!op) - return -ENOMEM; - - if (posix_lock_file_wait(file, fl) < 0) - log_error("gdlm_punlock: vfs unlock error %x,%llx", - name->ln_type, (unsigned long long)name->ln_number); - - op->info.optype = GDLM_PLOCK_OP_UNLOCK; - op->info.pid = fl->fl_pid; - op->info.fsid = ls->id; - op->info.number = name->ln_number; - op->info.start = fl->fl_start; - op->info.end = fl->fl_end; - if (fl->fl_lmops && fl->fl_lmops->fl_grant) - op->info.owner = (__u64) fl->fl_pid; - else - op->info.owner = (__u64)(long) fl->fl_owner; - - send_op(op); - wait_event(recv_wq, (op->done != 0)); - - spin_lock(&ops_lock); - if (!list_empty(&op->list)) { - printk(KERN_INFO "punlock op on list\n"); - list_del(&op->list); - } - spin_unlock(&ops_lock); - - rv = op->info.rv; - - if (rv == -ENOENT) - rv = 0; - - kfree(op); - return rv; -} - -int gdlm_plock_get(void *lockspace, struct lm_lockname *name, - struct file *file, struct file_lock *fl) -{ - struct gdlm_ls *ls = lockspace; - struct plock_op *op; - int rv; - - op = kzalloc(sizeof(*op), GFP_KERNEL); - if (!op) - return -ENOMEM; - - op->info.optype = GDLM_PLOCK_OP_GET; - op->info.pid = fl->fl_pid; - op->info.ex = (fl->fl_type == F_WRLCK); - op->info.fsid = ls->id; - op->info.number = name->ln_number; - op->info.start = fl->fl_start; - op->info.end = fl->fl_end; - if (fl->fl_lmops && fl->fl_lmops->fl_grant) - op->info.owner = (__u64) fl->fl_pid; - else - op->info.owner = (__u64)(long) fl->fl_owner; - - send_op(op); - wait_event(recv_wq, (op->done != 0)); - - spin_lock(&ops_lock); - if (!list_empty(&op->list)) { - printk(KERN_INFO "plock_get op on list\n"); - list_del(&op->list); - } - spin_unlock(&ops_lock); - - /* info.rv from userspace is 1 for conflict, 0 for no-conflict, - -ENOENT if there are no locks on the file */ - - rv = op->info.rv; - - fl->fl_type = F_UNLCK; - if (rv == -ENOENT) - rv = 0; - else if (rv > 0) { - fl->fl_type = (op->info.ex) ? F_WRLCK : F_RDLCK; - fl->fl_pid = op->info.pid; - fl->fl_start = op->info.start; - fl->fl_end = op->info.end; - rv = 0; - } - - kfree(op); - return rv; -} - -/* a read copies out one plock request from the send list */ -static ssize_t dev_read(struct file *file, char __user *u, size_t count, - loff_t *ppos) -{ - struct gdlm_plock_info info; - struct plock_op *op = NULL; - - if (count < sizeof(info)) - return -EINVAL; - - spin_lock(&ops_lock); - if (!list_empty(&send_list)) { - op = list_entry(send_list.next, struct plock_op, list); - list_move(&op->list, &recv_list); - memcpy(&info, &op->info, sizeof(info)); - } - spin_unlock(&ops_lock); - - if (!op) - return -EAGAIN; - - if (copy_to_user(u, &info, sizeof(info))) - return -EFAULT; - return sizeof(info); -} - -/* a write copies in one plock result that should match a plock_op - on the recv list */ -static ssize_t dev_write(struct file *file, const char __user *u, size_t count, - loff_t *ppos) -{ - struct gdlm_plock_info info; - struct plock_op *op; - int found = 0; - - if (count != sizeof(info)) - return -EINVAL; - - if (copy_from_user(&info, u, sizeof(info))) - return -EFAULT; - - if (check_version(&info)) - return -EINVAL; - - spin_lock(&ops_lock); - list_for_each_entry(op, &recv_list, list) { - if (op->info.fsid == info.fsid && op->info.number == info.number && - op->info.owner == info.owner) { - list_del_init(&op->list); - found = 1; - op->done = 1; - memcpy(&op->info, &info, sizeof(info)); - break; - } - } - spin_unlock(&ops_lock); - - if (found) { - struct plock_xop *xop; - xop = (struct plock_xop *)op; - if (xop->callback) - count = gdlm_plock_callback(op); - else - wake_up(&recv_wq); - } else - printk(KERN_INFO "gdlm dev_write no op %x %llx\n", info.fsid, - (unsigned long long)info.number); - return count; -} - -static unsigned int dev_poll(struct file *file, poll_table *wait) -{ - unsigned int mask = 0; - - poll_wait(file, &send_wq, wait); - - spin_lock(&ops_lock); - if (!list_empty(&send_list)) - mask = POLLIN | POLLRDNORM; - spin_unlock(&ops_lock); - - return mask; -} - -static const struct file_operations dev_fops = { - .read = dev_read, - .write = dev_write, - .poll = dev_poll, - .owner = THIS_MODULE -}; - -static struct miscdevice plock_dev_misc = { - .minor = MISC_DYNAMIC_MINOR, - .name = GDLM_PLOCK_MISC_NAME, - .fops = &dev_fops -}; - -int gdlm_plock_init(void) -{ - int rv; - - spin_lock_init(&ops_lock); - INIT_LIST_HEAD(&send_list); - INIT_LIST_HEAD(&recv_list); - init_waitqueue_head(&send_wq); - init_waitqueue_head(&recv_wq); - - rv = misc_register(&plock_dev_misc); - if (rv) - printk(KERN_INFO "gdlm_plock_init: misc_register failed %d", - rv); - return rv; -} - -void gdlm_plock_exit(void) -{ - if (misc_deregister(&plock_dev_misc) < 0) - printk(KERN_INFO "gdlm_plock_exit: misc_deregister failed"); -} - -- cgit v1.2.3 From 3d564fa3472d36cd6aa70514c37b8bbbec5b17ab Mon Sep 17 00:00:00 2001 From: David Teigland Date: Mon, 14 Apr 2008 14:06:29 -0500 Subject: dlm: common max length definitions Add central definitions for max lockspace name length and max resource name length. The lack of central definitions has resulted in scattered private definitions which we can now clean up, including an unused one in dlm_device.h. Signed-off-by: David Teigland --- fs/dlm/dlm_internal.h | 2 -- 1 file changed, 2 deletions(-) (limited to 'fs') diff --git a/fs/dlm/dlm_internal.h b/fs/dlm/dlm_internal.h index caa1581e158..73c36f50170 100644 --- a/fs/dlm/dlm_internal.h +++ b/fs/dlm/dlm_internal.h @@ -43,8 +43,6 @@ #include #include "config.h" -#define DLM_LOCKSPACE_LEN 64 - /* Size of the temp buffer midcomms allocates on the stack. We try to make this large enough so most messages fit. FIXME: should sctp make this unnecessary? */ -- cgit v1.2.3 From ec98c6b9b47df6df1c1fa6cf3d427414f8c2cf16 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Sun, 20 Apr 2008 02:14:23 -0700 Subject: [SPARC]: Remove SunOS and Solaris binary support. As per Documentation/feature-removal-schedule.txt Signed-off-by: David S. Miller --- fs/Kconfig.binfmt | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/Kconfig.binfmt b/fs/Kconfig.binfmt index b5c3b6114ad..853845abcca 100644 --- a/fs/Kconfig.binfmt +++ b/fs/Kconfig.binfmt @@ -62,7 +62,7 @@ config BINFMT_SHARED_FLAT config BINFMT_AOUT tristate "Kernel support for a.out and ECOFF binaries" depends on ARCH_SUPPORTS_AOUT && \ - (X86_32 || ALPHA || ARM || M68K || SPARC32) + (X86_32 || ALPHA || ARM || M68K) ---help--- A.out (Assembler.OUTput) is a set of formats for libraries and executables used in the earliest versions of UNIX. Linux used -- cgit v1.2.3 From f5264481c8049673e2cc8c7aca410931f571ba2d Mon Sep 17 00:00:00 2001 From: Pavel Machek Date: Mon, 21 Apr 2008 22:15:06 +0000 Subject: trivial: small cleanups These are small cleanups all over the tree. Trivial style and comment changes to fs/select.c, kernel/signal.c, kernel/stop_machine.c & mm/pdflush.c Signed-off-by: Pavel Machek Signed-off-by: Jesper Juhl --- fs/select.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/select.c b/fs/select.c index 5633fe98078..00f58c5c7e0 100644 --- a/fs/select.c +++ b/fs/select.c @@ -260,7 +260,7 @@ int do_select(int n, fd_set_bits *fds, s64 *timeout) wait = NULL; if (retval || !*timeout || signal_pending(current)) break; - if(table.error) { + if (table.error) { retval = table.error; break; } -- cgit v1.2.3 From 1cc8dcf569a3fcefb7ae32652225f2bd3e85257e Mon Sep 17 00:00:00 2001 From: Benoit Boissinot Date: Mon, 21 Apr 2008 22:45:55 +0000 Subject: ext*: spelling fix prefered -> preferred Spelling fix: prefered -> preferred Signed-off-by: Benoit Boissinot Signed-off-by: Jesper Juhl --- fs/ext2/ialloc.c | 2 +- fs/ext2/inode.c | 4 ++-- fs/ext3/ialloc.c | 2 +- fs/ext3/inode.c | 6 +++--- fs/ext4/ialloc.c | 2 +- fs/ext4/inode.c | 6 +++--- 6 files changed, 11 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/ext2/ialloc.c b/fs/ext2/ialloc.c index 5deb8b74e64..08f647d8188 100644 --- a/fs/ext2/ialloc.c +++ b/fs/ext2/ialloc.c @@ -253,7 +253,7 @@ static int find_group_dir(struct super_block *sb, struct inode *parent) * it has too few free inodes left (min_inodes) or * it has too few free blocks left (min_blocks) or * it's already running too large debt (max_debt). - * Parent's group is prefered, if it doesn't satisfy these + * Parent's group is preferred, if it doesn't satisfy these * conditions we search cyclically through the rest. If none * of the groups look good we just look for a group with more * free inodes than average (starting at parent's group). diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c index c6200680542..b8a2990bab8 100644 --- a/fs/ext2/inode.c +++ b/fs/ext2/inode.c @@ -239,7 +239,7 @@ no_block: * @inode: owner * @ind: descriptor of indirect block. * - * This function returns the prefered place for block allocation. + * This function returns the preferred place for block allocation. * It is used when heuristic for sequential allocation fails. * Rules are: * + if there is a block to the left of our position - allocate near it. @@ -283,7 +283,7 @@ static unsigned long ext2_find_near(struct inode *inode, Indirect *ind) } /** - * ext2_find_goal - find a prefered place for allocation. + * ext2_find_goal - find a preferred place for allocation. * @inode: owner * @block: block we want * @partial: pointer to the last triple within a chain diff --git a/fs/ext3/ialloc.c b/fs/ext3/ialloc.c index 4f4020c5468..96dd5573e49 100644 --- a/fs/ext3/ialloc.c +++ b/fs/ext3/ialloc.c @@ -239,7 +239,7 @@ static int find_group_dir(struct super_block *sb, struct inode *parent) * it has too few free inodes left (min_inodes) or * it has too few free blocks left (min_blocks) or * it's already running too large debt (max_debt). - * Parent's group is prefered, if it doesn't satisfy these + * Parent's group is preferred, if it doesn't satisfy these * conditions we search cyclically through the rest. If none * of the groups look good we just look for a group with more * free inodes than average (starting at parent's group). diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c index eb95670a27e..c683609b0e3 100644 --- a/fs/ext3/inode.c +++ b/fs/ext3/inode.c @@ -392,7 +392,7 @@ no_block: * @inode: owner * @ind: descriptor of indirect block. * - * This function returns the prefered place for block allocation. + * This function returns the preferred place for block allocation. * It is used when heuristic for sequential allocation fails. * Rules are: * + if there is a block to the left of our position - allocate near it. @@ -436,12 +436,12 @@ static ext3_fsblk_t ext3_find_near(struct inode *inode, Indirect *ind) } /** - * ext3_find_goal - find a prefered place for allocation. + * ext3_find_goal - find a preferred place for allocation. * @inode: owner * @block: block we want * @partial: pointer to the last triple within a chain * - * Normally this function find the prefered place for block allocation, + * Normally this function find the preferred place for block allocation, * returns it. */ diff --git a/fs/ext4/ialloc.c b/fs/ext4/ialloc.c index 8036b9b5376..486e46a3918 100644 --- a/fs/ext4/ialloc.c +++ b/fs/ext4/ialloc.c @@ -305,7 +305,7 @@ static int find_group_dir(struct super_block *sb, struct inode *parent, * it has too few free inodes left (min_inodes) or * it has too few free blocks left (min_blocks) or * it's already running too large debt (max_debt). - * Parent's group is prefered, if it doesn't satisfy these + * Parent's group is preferred, if it doesn't satisfy these * conditions we search cyclically through the rest. If none * of the groups look good we just look for a group with more * free inodes than average (starting at parent's group). diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 945cbf6cb1f..8fab233cb05 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -382,7 +382,7 @@ no_block: * @inode: owner * @ind: descriptor of indirect block. * - * This function returns the prefered place for block allocation. + * This function returns the preferred place for block allocation. * It is used when heuristic for sequential allocation fails. * Rules are: * + if there is a block to the left of our position - allocate near it. @@ -432,12 +432,12 @@ static ext4_fsblk_t ext4_find_near(struct inode *inode, Indirect *ind) } /** - * ext4_find_goal - find a prefered place for allocation. + * ext4_find_goal - find a preferred place for allocation. * @inode: owner * @block: block we want * @partial: pointer to the last triple within a chain * - * Normally this function find the prefered place for block allocation, + * Normally this function find the preferred place for block allocation, * returns it. */ static ext4_fsblk_t ext4_find_goal(struct inode *inode, ext4_lblk_t block, -- cgit v1.2.3 From 6d59e7f582ef1c1988542d0fc3b36d0087b757ce Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 22 Mar 2008 15:48:17 -0400 Subject: [PATCH] move a bunch of declarations to fs/internal.h Signed-off-by: Al Viro --- fs/internal.h | 11 +++++++++++ fs/pnode.c | 1 + fs/super.c | 1 + 3 files changed, 13 insertions(+) (limited to 'fs') diff --git a/fs/internal.h b/fs/internal.h index 392e8ccd6fc..80aa9a02337 100644 --- a/fs/internal.h +++ b/fs/internal.h @@ -43,3 +43,14 @@ extern void __init chrdev_init(void); * namespace.c */ extern int copy_mount_options(const void __user *, unsigned long *); + +extern void free_vfsmnt(struct vfsmount *); +extern struct vfsmount *alloc_vfsmnt(const char *); +extern struct vfsmount *__lookup_mnt(struct vfsmount *, struct dentry *, int); +extern void mnt_set_mountpoint(struct vfsmount *, struct dentry *, + struct vfsmount *); +extern void release_mounts(struct list_head *); +extern void umount_tree(struct vfsmount *, int, struct list_head *); +extern struct vfsmount *copy_tree(struct vfsmount *, struct dentry *, int); + +extern void __init mnt_init(void); diff --git a/fs/pnode.c b/fs/pnode.c index 1d8f5447f3f..a9e0d6fadbc 100644 --- a/fs/pnode.c +++ b/fs/pnode.c @@ -9,6 +9,7 @@ #include #include #include +#include "internal.h" #include "pnode.h" /* return the next shared peer mount of @p */ diff --git a/fs/super.c b/fs/super.c index 1f8f05ede43..4798350b2bc 100644 --- a/fs/super.c +++ b/fs/super.c @@ -39,6 +39,7 @@ #include #include #include +#include "internal.h" LIST_HEAD(super_blocks); -- cgit v1.2.3 From 1a60a280778ff90270fc7390d9ec102f713a5a29 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 22 Mar 2008 16:19:49 -0400 Subject: [PATCH] lock exclusively in collect_mounts() and drop_collected_mounts() Taking namespace_sem shared there isn't worth the trouble, especially with vfsmount ID allocation about to be added. That way we know that umount_tree(), copy_tree() and clone_mnt() are _always_ serialized by namespace_sem. umount_tree() still needs vfsmount_lock (it manipulates hash chains, among other things), but that's a separate story. Signed-off-by: Al Viro --- fs/namespace.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/namespace.c b/fs/namespace.c index 678f7ce060f..af2fb3707d0 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1091,20 +1091,20 @@ Enomem: struct vfsmount *collect_mounts(struct vfsmount *mnt, struct dentry *dentry) { struct vfsmount *tree; - down_read(&namespace_sem); + down_write(&namespace_sem); tree = copy_tree(mnt, dentry, CL_COPY_ALL | CL_PRIVATE); - up_read(&namespace_sem); + up_write(&namespace_sem); return tree; } void drop_collected_mounts(struct vfsmount *mnt) { LIST_HEAD(umount_list); - down_read(&namespace_sem); + down_write(&namespace_sem); spin_lock(&vfsmount_lock); umount_tree(mnt, 0, &umount_list); spin_unlock(&vfsmount_lock); - up_read(&namespace_sem); + up_write(&namespace_sem); release_mounts(&umount_list); } -- cgit v1.2.3 From b5266eb4c8d1a2887a19aaec8144ee4ad1b054c3 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 22 Mar 2008 17:48:24 -0400 Subject: [PATCH] switch a bunch of LSM hooks from nameidata to path Namely, ones from namespace.c Signed-off-by: Al Viro --- fs/namespace.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/namespace.c b/fs/namespace.c index af2fb3707d0..87d2d82010b 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1220,7 +1220,7 @@ static int graft_tree(struct vfsmount *mnt, struct nameidata *nd) if (IS_DEADDIR(nd->path.dentry->d_inode)) goto out_unlock; - err = security_sb_check_sb(mnt, nd); + err = security_sb_check_sb(mnt, &nd->path); if (err) goto out_unlock; @@ -1230,7 +1230,7 @@ static int graft_tree(struct vfsmount *mnt, struct nameidata *nd) out_unlock: mutex_unlock(&nd->path.dentry->d_inode->i_mutex); if (!err) - security_sb_post_addmount(mnt, nd); + security_sb_post_addmount(mnt, &nd->path); return err; } @@ -1746,7 +1746,8 @@ long do_mount(char *dev_name, char *dir_name, char *type_page, if (retval) return retval; - retval = security_sb_mount(dev_name, &nd, type_page, flags, data_page); + retval = security_sb_mount(dev_name, &nd.path, + type_page, flags, data_page); if (retval) goto dput_out; @@ -2007,7 +2008,7 @@ asmlinkage long sys_pivot_root(const char __user * new_root, if (error) goto out1; - error = security_sb_pivotroot(&old_nd, &new_nd); + error = security_sb_pivotroot(&old_nd.path, &new_nd.path); if (error) { path_put(&old_nd.path); goto out1; @@ -2070,7 +2071,7 @@ asmlinkage long sys_pivot_root(const char __user * new_root, touch_mnt_namespace(current->nsproxy->mnt_ns); spin_unlock(&vfsmount_lock); chroot_fs_refs(&user_nd.path, &new_nd.path); - security_sb_post_pivotroot(&user_nd, &new_nd); + security_sb_post_pivotroot(&user_nd.path, &new_nd.path); error = 0; path_put(&root_parent); path_put(&parent_path); -- cgit v1.2.3 From 8c3ee42e80ccead805806b3cb50b9855ceb957a2 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sat, 22 Mar 2008 18:00:39 -0400 Subject: [PATCH] get rid of more nameidata passing in namespace.c Further reduction of stack footprint (sys_pivot_root()); lose useless BKL in there, while we are at it. Signed-off-by: Al Viro --- fs/namespace.c | 53 +++++++++++++++++++++++++---------------------------- 1 file changed, 25 insertions(+), 28 deletions(-) (limited to 'fs') diff --git a/fs/namespace.c b/fs/namespace.c index 87d2d82010b..1bf302d0478 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1205,32 +1205,32 @@ static int attach_recursive_mnt(struct vfsmount *source_mnt, return 0; } -static int graft_tree(struct vfsmount *mnt, struct nameidata *nd) +static int graft_tree(struct vfsmount *mnt, struct path *path) { int err; if (mnt->mnt_sb->s_flags & MS_NOUSER) return -EINVAL; - if (S_ISDIR(nd->path.dentry->d_inode->i_mode) != + if (S_ISDIR(path->dentry->d_inode->i_mode) != S_ISDIR(mnt->mnt_root->d_inode->i_mode)) return -ENOTDIR; err = -ENOENT; - mutex_lock(&nd->path.dentry->d_inode->i_mutex); - if (IS_DEADDIR(nd->path.dentry->d_inode)) + mutex_lock(&path->dentry->d_inode->i_mutex); + if (IS_DEADDIR(path->dentry->d_inode)) goto out_unlock; - err = security_sb_check_sb(mnt, &nd->path); + err = security_sb_check_sb(mnt, path); if (err) goto out_unlock; err = -ENOENT; - if (IS_ROOT(nd->path.dentry) || !d_unhashed(nd->path.dentry)) - err = attach_recursive_mnt(mnt, &nd->path, NULL); + if (IS_ROOT(path->dentry) || !d_unhashed(path->dentry)) + err = attach_recursive_mnt(mnt, path, NULL); out_unlock: - mutex_unlock(&nd->path.dentry->d_inode->i_mutex); + mutex_unlock(&path->dentry->d_inode->i_mutex); if (!err) - security_sb_post_addmount(mnt, &nd->path); + security_sb_post_addmount(mnt, path); return err; } @@ -1294,7 +1294,7 @@ static noinline int do_loopback(struct nameidata *nd, char *old_name, if (!mnt) goto out; - err = graft_tree(mnt, nd); + err = graft_tree(mnt, &nd->path); if (err) { LIST_HEAD(umount_list); spin_lock(&vfsmount_lock); @@ -1501,7 +1501,7 @@ int do_add_mount(struct vfsmount *newmnt, struct nameidata *nd, goto unlock; newmnt->mnt_flags = mnt_flags; - if ((err = graft_tree(newmnt, nd))) + if ((err = graft_tree(newmnt, &nd->path))) goto unlock; if (fslist) /* add to the specified expiration list */ @@ -1987,15 +1987,13 @@ asmlinkage long sys_pivot_root(const char __user * new_root, const char __user * put_old) { struct vfsmount *tmp; - struct nameidata new_nd, old_nd, user_nd; - struct path parent_path, root_parent; + struct nameidata new_nd, old_nd; + struct path parent_path, root_parent, root; int error; if (!capable(CAP_SYS_ADMIN)) return -EPERM; - lock_kernel(); - error = __user_walk(new_root, LOOKUP_FOLLOW | LOOKUP_DIRECTORY, &new_nd); if (error) @@ -2015,7 +2013,7 @@ asmlinkage long sys_pivot_root(const char __user * new_root, } read_lock(¤t->fs->lock); - user_nd.path = current->fs->root; + root = current->fs->root; path_get(¤t->fs->root); read_unlock(¤t->fs->lock); down_write(&namespace_sem); @@ -2023,9 +2021,9 @@ asmlinkage long sys_pivot_root(const char __user * new_root, error = -EINVAL; if (IS_MNT_SHARED(old_nd.path.mnt) || IS_MNT_SHARED(new_nd.path.mnt->mnt_parent) || - IS_MNT_SHARED(user_nd.path.mnt->mnt_parent)) + IS_MNT_SHARED(root.mnt->mnt_parent)) goto out2; - if (!check_mnt(user_nd.path.mnt)) + if (!check_mnt(root.mnt)) goto out2; error = -ENOENT; if (IS_DEADDIR(new_nd.path.dentry->d_inode)) @@ -2035,13 +2033,13 @@ asmlinkage long sys_pivot_root(const char __user * new_root, if (d_unhashed(old_nd.path.dentry) && !IS_ROOT(old_nd.path.dentry)) goto out2; error = -EBUSY; - if (new_nd.path.mnt == user_nd.path.mnt || - old_nd.path.mnt == user_nd.path.mnt) + if (new_nd.path.mnt == root.mnt || + old_nd.path.mnt == root.mnt) goto out2; /* loop, on the same file system */ error = -EINVAL; - if (user_nd.path.mnt->mnt_root != user_nd.path.dentry) + if (root.mnt->mnt_root != root.dentry) goto out2; /* not a mountpoint */ - if (user_nd.path.mnt->mnt_parent == user_nd.path.mnt) + if (root.mnt->mnt_parent == root.mnt) goto out2; /* not attached */ if (new_nd.path.mnt->mnt_root != new_nd.path.dentry) goto out2; /* not a mountpoint */ @@ -2063,27 +2061,26 @@ asmlinkage long sys_pivot_root(const char __user * new_root, } else if (!is_subdir(old_nd.path.dentry, new_nd.path.dentry)) goto out3; detach_mnt(new_nd.path.mnt, &parent_path); - detach_mnt(user_nd.path.mnt, &root_parent); + detach_mnt(root.mnt, &root_parent); /* mount old root on put_old */ - attach_mnt(user_nd.path.mnt, &old_nd.path); + attach_mnt(root.mnt, &old_nd.path); /* mount new_root on / */ attach_mnt(new_nd.path.mnt, &root_parent); touch_mnt_namespace(current->nsproxy->mnt_ns); spin_unlock(&vfsmount_lock); - chroot_fs_refs(&user_nd.path, &new_nd.path); - security_sb_post_pivotroot(&user_nd.path, &new_nd.path); + chroot_fs_refs(&root, &new_nd.path); + security_sb_post_pivotroot(&root, &new_nd.path); error = 0; path_put(&root_parent); path_put(&parent_path); out2: mutex_unlock(&old_nd.path.dentry->d_inode->i_mutex); up_write(&namespace_sem); - path_put(&user_nd.path); + path_put(&root); path_put(&old_nd.path); out1: path_put(&new_nd.path); out0: - unlock_kernel(); return error; out3: spin_unlock(&vfsmount_lock); -- cgit v1.2.3 From 4e1b36fb485dd81b0818ef1bc8fb5c0f2923a283 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Mon, 24 Mar 2008 00:16:03 -0400 Subject: [PATCH] umount_tree() will unhash everything itself Signed-off-by: Al Viro --- fs/pnode.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/pnode.c b/fs/pnode.c index a9e0d6fadbc..f968e35d978 100644 --- a/fs/pnode.c +++ b/fs/pnode.c @@ -212,8 +212,7 @@ int propagate_mnt(struct vfsmount *dest_mnt, struct dentry *dest_dentry, out: spin_lock(&vfsmount_lock); while (!list_empty(&tmp_list)) { - child = list_entry(tmp_list.next, struct vfsmount, mnt_hash); - list_del_init(&child->mnt_hash); + child = list_first_entry(&tmp_list, struct vfsmount, mnt_hash); umount_tree(child, 0, &umount_list); } spin_unlock(&vfsmount_lock); -- cgit v1.2.3 From 521b5d0c40386f4a9805cdec7bd979fc96a86aeb Mon Sep 17 00:00:00 2001 From: Al Viro Date: Fri, 28 Mar 2008 00:46:41 -0400 Subject: [PATCH] teach seq_file to discard entries Allow ->show() return SEQ_SKIP; that will discard all output from that element and move on. Signed-off-by: Al Viro --- fs/pnode.h | 1 + fs/seq_file.c | 16 ++++++++++++---- 2 files changed, 13 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/pnode.h b/fs/pnode.h index f249be2fee7..973c3f825e7 100644 --- a/fs/pnode.h +++ b/fs/pnode.h @@ -35,4 +35,5 @@ int propagate_mnt(struct vfsmount *, struct dentry *, struct vfsmount *, struct list_head *); int propagate_umount(struct list_head *); int propagate_mount_busy(struct vfsmount *, int); +void mnt_release_group_id(struct vfsmount *); #endif /* _LINUX_PNODE_H */ diff --git a/fs/seq_file.c b/fs/seq_file.c index 853770274f2..bf2bcfd4bcf 100644 --- a/fs/seq_file.c +++ b/fs/seq_file.c @@ -25,6 +25,7 @@ * into the buffer. In case of error ->start() and ->next() return * ERR_PTR(error). In the end of sequence they return %NULL. ->show() * returns 0 in case of success and negative number in case of error. + * Returning SEQ_SKIP means "discard this element and move on". */ int seq_open(struct file *file, const struct seq_operations *op) { @@ -114,8 +115,10 @@ ssize_t seq_read(struct file *file, char __user *buf, size_t size, loff_t *ppos) if (!p || IS_ERR(p)) break; err = m->op->show(m, p); - if (err) + if (err < 0) break; + if (unlikely(err)) + m->count = 0; if (m->count < m->size) goto Fill; m->op->stop(m, p); @@ -140,9 +143,10 @@ Fill: break; } err = m->op->show(m, p); - if (err || m->count == m->size) { + if (m->count == m->size || err) { m->count = offs; - break; + if (likely(err <= 0)) + break; } pos = next; } @@ -199,8 +203,12 @@ static int traverse(struct seq_file *m, loff_t offset) if (IS_ERR(p)) break; error = m->op->show(m, p); - if (error) + if (error < 0) break; + if (unlikely(error)) { + error = 0; + m->count = 0; + } if (m->count == m->size) goto Eoverflow; if (pos + m->count > offset) { -- cgit v1.2.3 From f876a59dae09a353444913bdf73b125bc124a848 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Fri, 18 Apr 2008 13:44:12 -0700 Subject: [JFFS2] include function prototype for jffs2_ioctl fs/jffs2/ioctl.c:14:5: warning: symbol 'jffs2_ioctl' was not declared. Signed-off-by: Harvey Harrison Signed-off-by: Andrew Morton Signed-off-by: David Woodhouse --- fs/jffs2/ioctl.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/jffs2/ioctl.c b/fs/jffs2/ioctl.c index f4d525b0ea5..e2177210f62 100644 --- a/fs/jffs2/ioctl.c +++ b/fs/jffs2/ioctl.c @@ -10,6 +10,7 @@ */ #include +#include "nodelist.h" int jffs2_ioctl(struct inode *inode, struct file *filp, unsigned int cmd, unsigned long arg) -- cgit v1.2.3 From 8ca646abb4503f39a7d235b89b9f8015e3ab4631 Mon Sep 17 00:00:00 2001 From: David Woodhouse Date: Tue, 22 Apr 2008 11:25:47 +0100 Subject: [JFFS2] Fix sparse warning in nodemgmt.c fs/jffs2/nodemgmt.c:60:8: warning: symbol 'ret' shadows an earlier one fs/jffs2/nodemgmt.c:45:6: originally declared here (reported by Harvey Harrison) Just remove the offending declaration of 'int ret' and use the earlier one. Signed-off-by: David Woodhouse --- fs/jffs2/nodemgmt.c | 1 - 1 file changed, 1 deletion(-) (limited to 'fs') diff --git a/fs/jffs2/nodemgmt.c b/fs/jffs2/nodemgmt.c index a0313fa8748..a3e67a4ed67 100644 --- a/fs/jffs2/nodemgmt.c +++ b/fs/jffs2/nodemgmt.c @@ -57,7 +57,6 @@ int jffs2_reserve_space(struct jffs2_sb_info *c, uint32_t minsize, /* this needs a little more thought (true :)) */ while(ret == -EAGAIN) { while(c->nr_free_blocks + c->nr_erasing_blocks < blocksneeded) { - int ret; uint32_t dirty, avail; /* calculate real dirty size -- cgit v1.2.3 From bf66737ca85c41442e99c9d380eb7807d88bac1f Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Fri, 18 Apr 2008 13:44:14 -0700 Subject: [JFFS2] fix sparse warning in write.c fs/jffs2/write.c:585:28: warning: symbol 'fd' shadows an earlier one fs/jffs2/write.c:536:27: originally declared here No need to redeclare fd, use the original one, after this point, fd is always reassigned before it used again. Signed-off-by: Harvey Harrison Signed-off-by: Andrew Morton Signed-off-by: David Woodhouse --- fs/jffs2/write.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/jffs2/write.c b/fs/jffs2/write.c index 776f13cbf2b..beade550909 100644 --- a/fs/jffs2/write.c +++ b/fs/jffs2/write.c @@ -582,9 +582,9 @@ int jffs2_do_unlink(struct jffs2_sb_info *c, struct jffs2_inode_info *dir_f, jffs2_add_fd_to_list(c, fd, &dir_f->dents); up(&dir_f->sem); } else { - struct jffs2_full_dirent *fd = dir_f->dents; uint32_t nhash = full_name_hash(name, namelen); + fd = dir_f->dents; /* We don't actually want to reserve any space, but we do want to be holding the alloc_sem when we write to flash */ down(&c->alloc_sem); -- cgit v1.2.3 From 25dc30b4cd68df1de8932fe77ca574227d42a259 Mon Sep 17 00:00:00 2001 From: David Woodhouse Date: Tue, 22 Apr 2008 12:12:25 +0100 Subject: [JFFS2] fix sparse warnings in gc.c fs/jffs2/gc.c:1147:29: warning: symbol 'jeb' shadows an earlier one fs/jffs2/gc.c:1084:89: originally declared here fs/jffs2/gc.c:1197:29: warning: symbol 'jeb' shadows an earlier one fs/jffs2/gc.c:1084:89: originally declared here Rename the unused 'jeb' argument to avoid this. We could potentially remove the argument, but GCC should be doing that anyway. Signed-off-by: David Woodhouse --- fs/jffs2/gc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/jffs2/gc.c b/fs/jffs2/gc.c index 32ff0373aa0..59aeb982043 100644 --- a/fs/jffs2/gc.c +++ b/fs/jffs2/gc.c @@ -1081,7 +1081,7 @@ static int jffs2_garbage_collect_hole(struct jffs2_sb_info *c, struct jffs2_eras return 0; } -static int jffs2_garbage_collect_dnode(struct jffs2_sb_info *c, struct jffs2_eraseblock *jeb, +static int jffs2_garbage_collect_dnode(struct jffs2_sb_info *c, struct jffs2_eraseblock *orig_jeb, struct jffs2_inode_info *f, struct jffs2_full_dnode *fn, uint32_t start, uint32_t end) { -- cgit v1.2.3 From cca15841710da39490abc17f56b55703e3140955 Mon Sep 17 00:00:00 2001 From: michael Date: Fri, 18 Apr 2008 13:44:17 -0700 Subject: [JFFS2] add write verify on dataflash. Add the write verification buffer to the dataflash. The mtd_dataflash has the CONFIG_DATAFLASH_WRITE_VERIFY so is better a change to Kconfig. Signed-off-by: Michael Trimarchi Signed-off-by: Andrew Morton Signed-off-by: David Woodhouse --- fs/jffs2/wbuf.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'fs') diff --git a/fs/jffs2/wbuf.c b/fs/jffs2/wbuf.c index d1d4f27464b..ba49f19cff8 100644 --- a/fs/jffs2/wbuf.c +++ b/fs/jffs2/wbuf.c @@ -1236,12 +1236,24 @@ int jffs2_dataflash_setup(struct jffs2_sb_info *c) { if (!c->wbuf) return -ENOMEM; +#ifdef CONFIG_JFFS2_FS_WBUF_VERIFY + c->wbuf_verify = kmalloc(c->wbuf_pagesize, GFP_KERNEL); + if (!c->wbuf_verify) { + kfree(c->oobbuf); + kfree(c->wbuf); + return -ENOMEM; + } +#endif + printk(KERN_INFO "JFFS2 write-buffering enabled buffer (%d) erasesize (%d)\n", c->wbuf_pagesize, c->sector_size); return 0; } void jffs2_dataflash_cleanup(struct jffs2_sb_info *c) { +#ifdef CONFIG_JFFS2_FS_WBUF_VERIFY + kfree(c->wbuf_verify); +#endif kfree(c->wbuf); } -- cgit v1.2.3 From ced22070363ef50e4a47aadd003a81ebeaa3f917 Mon Sep 17 00:00:00 2001 From: David Woodhouse Date: Tue, 22 Apr 2008 15:13:40 +0100 Subject: [JFFS2] semaphore->mutex conversion Signed-off-by: David Woodhouse --- fs/jffs2/README.Locking | 22 +++++++++---------- fs/jffs2/debug.c | 8 +++---- fs/jffs2/dir.c | 58 ++++++++++++++++++++++++------------------------- fs/jffs2/erase.c | 42 +++++++++++++++++------------------ fs/jffs2/file.c | 16 +++++++------- fs/jffs2/fs.c | 28 ++++++++++++------------ fs/jffs2/gc.c | 32 +++++++++++++-------------- fs/jffs2/jffs2_fs_i.h | 2 +- fs/jffs2/jffs2_fs_sb.h | 6 ++--- fs/jffs2/nodemgmt.c | 18 +++++++-------- fs/jffs2/readinode.c | 23 ++++++++++---------- fs/jffs2/super.c | 14 ++++++------ fs/jffs2/wbuf.c | 16 +++++++------- fs/jffs2/write.c | 50 +++++++++++++++++++++--------------------- 14 files changed, 168 insertions(+), 167 deletions(-) (limited to 'fs') diff --git a/fs/jffs2/README.Locking b/fs/jffs2/README.Locking index d14d5a4dc5a..3ea36554107 100644 --- a/fs/jffs2/README.Locking +++ b/fs/jffs2/README.Locking @@ -14,7 +14,7 @@ be fairly close. alloc_sem --------- -The alloc_sem is a per-filesystem semaphore, used primarily to ensure +The alloc_sem is a per-filesystem mutex, used primarily to ensure contiguous allocation of space on the medium. It is automatically obtained during space allocations (jffs2_reserve_space()) and freed upon write completion (jffs2_complete_reservation()). Note that @@ -41,10 +41,10 @@ if the wbuf is currently holding any data is permitted, though. Ordering constraints: See f->sem. - File Semaphore f->sem + File Mutex f->sem --------------------- -This is the JFFS2-internal equivalent of the inode semaphore i->i_sem. +This is the JFFS2-internal equivalent of the inode mutex i->i_sem. It protects the contents of the jffs2_inode_info private inode data, including the linked list of node fragments (but see the notes below on erase_completion_lock), etc. @@ -60,14 +60,14 @@ lead to deadlock, unless we played games with unlocking the i_sem before calling the space allocation functions. Instead of playing such games, we just have an extra internal -semaphore, which is obtained by the garbage collection code and also +mutex, which is obtained by the garbage collection code and also by the normal file system code _after_ allocation of space. Ordering constraints: 1. Never attempt to allocate space or lock alloc_sem with any f->sem held. - 2. Never attempt to lock two file semaphores in one thread. + 2. Never attempt to lock two file mutexes in one thread. No ordering rules have been made for doing so. @@ -86,8 +86,8 @@ a simple spin_lock() rather than spin_lock_bh(). Note that the per-inode list of physical nodes (f->nodes) is a special case. Any changes to _valid_ nodes (i.e. ->flash_offset & 1 == 0) in -the list are protected by the file semaphore f->sem. But the erase -code may remove _obsolete_ nodes from the list while holding only the +the list are protected by the file mutex f->sem. But the erase code +may remove _obsolete_ nodes from the list while holding only the erase_completion_lock. So you can walk the list only while holding the erase_completion_lock, and can drop the lock temporarily mid-walk as long as the pointer you're holding is to a _valid_ node, not an @@ -124,10 +124,10 @@ Ordering constraints: erase_free_sem -------------- -This semaphore is only used by the erase code which frees obsolete -node references and the jffs2_garbage_collect_deletion_dirent() -function. The latter function on NAND flash must read _obsolete_ nodes -to determine whether the 'deletion dirent' under consideration can be +This mutex is only used by the erase code which frees obsolete node +references and the jffs2_garbage_collect_deletion_dirent() function. +The latter function on NAND flash must read _obsolete_ nodes to +determine whether the 'deletion dirent' under consideration can be discarded or whether it is still required to show that an inode has been unlinked. Because reading from the flash may sleep, the erase_completion_lock cannot be held, so an alternative, more diff --git a/fs/jffs2/debug.c b/fs/jffs2/debug.c index 3a32c64ed49..660793107e9 100644 --- a/fs/jffs2/debug.c +++ b/fs/jffs2/debug.c @@ -62,9 +62,9 @@ __jffs2_dbg_acct_sanity_check(struct jffs2_sb_info *c, void __jffs2_dbg_fragtree_paranoia_check(struct jffs2_inode_info *f) { - down(&f->sem); + mutex_lock(&f->sem); __jffs2_dbg_fragtree_paranoia_check_nolock(f); - up(&f->sem); + mutex_unlock(&f->sem); } void @@ -532,9 +532,9 @@ __jffs2_dbg_dump_block_lists_nolock(struct jffs2_sb_info *c) void __jffs2_dbg_dump_fragtree(struct jffs2_inode_info *f) { - down(&f->sem); + mutex_lock(&f->sem); jffs2_dbg_dump_fragtree_nolock(f); - up(&f->sem); + mutex_unlock(&f->sem); } void diff --git a/fs/jffs2/dir.c b/fs/jffs2/dir.c index f948f7e6ec8..c63e7a96af0 100644 --- a/fs/jffs2/dir.c +++ b/fs/jffs2/dir.c @@ -86,7 +86,7 @@ static struct dentry *jffs2_lookup(struct inode *dir_i, struct dentry *target, dir_f = JFFS2_INODE_INFO(dir_i); c = JFFS2_SB_INFO(dir_i->i_sb); - down(&dir_f->sem); + mutex_lock(&dir_f->sem); /* NB: The 2.2 backport will need to explicitly check for '.' and '..' here */ for (fd_list = dir_f->dents; fd_list && fd_list->nhash <= target->d_name.hash; fd_list = fd_list->next) { @@ -99,7 +99,7 @@ static struct dentry *jffs2_lookup(struct inode *dir_i, struct dentry *target, } if (fd) ino = fd->ino; - up(&dir_f->sem); + mutex_unlock(&dir_f->sem); if (ino) { inode = jffs2_iget(dir_i->i_sb, ino); if (IS_ERR(inode)) { @@ -146,7 +146,7 @@ static int jffs2_readdir(struct file *filp, void *dirent, filldir_t filldir) } curofs=1; - down(&f->sem); + mutex_lock(&f->sem); for (fd = f->dents; fd; fd = fd->next) { curofs++; @@ -166,7 +166,7 @@ static int jffs2_readdir(struct file *filp, void *dirent, filldir_t filldir) break; offset++; } - up(&f->sem); + mutex_unlock(&f->sem); out: filp->f_pos = offset; return 0; @@ -275,9 +275,9 @@ static int jffs2_link (struct dentry *old_dentry, struct inode *dir_i, struct de ret = jffs2_do_link(c, dir_f, f->inocache->ino, type, dentry->d_name.name, dentry->d_name.len, now); if (!ret) { - down(&f->sem); + mutex_lock(&f->sem); old_dentry->d_inode->i_nlink = ++f->inocache->nlink; - up(&f->sem); + mutex_unlock(&f->sem); d_instantiate(dentry, old_dentry->d_inode); dir_i->i_mtime = dir_i->i_ctime = ITIME(now); atomic_inc(&old_dentry->d_inode->i_count); @@ -351,7 +351,7 @@ static int jffs2_symlink (struct inode *dir_i, struct dentry *dentry, const char if (IS_ERR(fn)) { /* Eeek. Wave bye bye */ - up(&f->sem); + mutex_unlock(&f->sem); jffs2_complete_reservation(c); jffs2_clear_inode(inode); return PTR_ERR(fn); @@ -361,7 +361,7 @@ static int jffs2_symlink (struct inode *dir_i, struct dentry *dentry, const char f->target = kmalloc(targetlen + 1, GFP_KERNEL); if (!f->target) { printk(KERN_WARNING "Can't allocate %d bytes of memory\n", targetlen + 1); - up(&f->sem); + mutex_unlock(&f->sem); jffs2_complete_reservation(c); jffs2_clear_inode(inode); return -ENOMEM; @@ -374,7 +374,7 @@ static int jffs2_symlink (struct inode *dir_i, struct dentry *dentry, const char obsoleted by the first data write */ f->metadata = fn; - up(&f->sem); + mutex_unlock(&f->sem); jffs2_complete_reservation(c); @@ -406,7 +406,7 @@ static int jffs2_symlink (struct inode *dir_i, struct dentry *dentry, const char } dir_f = JFFS2_INODE_INFO(dir_i); - down(&dir_f->sem); + mutex_lock(&dir_f->sem); rd->magic = cpu_to_je16(JFFS2_MAGIC_BITMASK); rd->nodetype = cpu_to_je16(JFFS2_NODETYPE_DIRENT); @@ -429,7 +429,7 @@ static int jffs2_symlink (struct inode *dir_i, struct dentry *dentry, const char as if it were the final unlink() */ jffs2_complete_reservation(c); jffs2_free_raw_dirent(rd); - up(&dir_f->sem); + mutex_unlock(&dir_f->sem); jffs2_clear_inode(inode); return PTR_ERR(fd); } @@ -442,7 +442,7 @@ static int jffs2_symlink (struct inode *dir_i, struct dentry *dentry, const char one if necessary. */ jffs2_add_fd_to_list(c, fd, &dir_f->dents); - up(&dir_f->sem); + mutex_unlock(&dir_f->sem); jffs2_complete_reservation(c); d_instantiate(dentry, inode); @@ -507,7 +507,7 @@ static int jffs2_mkdir (struct inode *dir_i, struct dentry *dentry, int mode) if (IS_ERR(fn)) { /* Eeek. Wave bye bye */ - up(&f->sem); + mutex_unlock(&f->sem); jffs2_complete_reservation(c); jffs2_clear_inode(inode); return PTR_ERR(fn); @@ -516,7 +516,7 @@ static int jffs2_mkdir (struct inode *dir_i, struct dentry *dentry, int mode) obsoleted by the first data write */ f->metadata = fn; - up(&f->sem); + mutex_unlock(&f->sem); jffs2_complete_reservation(c); @@ -548,7 +548,7 @@ static int jffs2_mkdir (struct inode *dir_i, struct dentry *dentry, int mode) } dir_f = JFFS2_INODE_INFO(dir_i); - down(&dir_f->sem); + mutex_lock(&dir_f->sem); rd->magic = cpu_to_je16(JFFS2_MAGIC_BITMASK); rd->nodetype = cpu_to_je16(JFFS2_NODETYPE_DIRENT); @@ -571,7 +571,7 @@ static int jffs2_mkdir (struct inode *dir_i, struct dentry *dentry, int mode) as if it were the final unlink() */ jffs2_complete_reservation(c); jffs2_free_raw_dirent(rd); - up(&dir_f->sem); + mutex_unlock(&dir_f->sem); jffs2_clear_inode(inode); return PTR_ERR(fd); } @@ -585,7 +585,7 @@ static int jffs2_mkdir (struct inode *dir_i, struct dentry *dentry, int mode) one if necessary. */ jffs2_add_fd_to_list(c, fd, &dir_f->dents); - up(&dir_f->sem); + mutex_unlock(&dir_f->sem); jffs2_complete_reservation(c); d_instantiate(dentry, inode); @@ -673,7 +673,7 @@ static int jffs2_mknod (struct inode *dir_i, struct dentry *dentry, int mode, de if (IS_ERR(fn)) { /* Eeek. Wave bye bye */ - up(&f->sem); + mutex_unlock(&f->sem); jffs2_complete_reservation(c); jffs2_clear_inode(inode); return PTR_ERR(fn); @@ -682,7 +682,7 @@ static int jffs2_mknod (struct inode *dir_i, struct dentry *dentry, int mode, de obsoleted by the first data write */ f->metadata = fn; - up(&f->sem); + mutex_unlock(&f->sem); jffs2_complete_reservation(c); @@ -714,7 +714,7 @@ static int jffs2_mknod (struct inode *dir_i, struct dentry *dentry, int mode, de } dir_f = JFFS2_INODE_INFO(dir_i); - down(&dir_f->sem); + mutex_lock(&dir_f->sem); rd->magic = cpu_to_je16(JFFS2_MAGIC_BITMASK); rd->nodetype = cpu_to_je16(JFFS2_NODETYPE_DIRENT); @@ -740,7 +740,7 @@ static int jffs2_mknod (struct inode *dir_i, struct dentry *dentry, int mode, de as if it were the final unlink() */ jffs2_complete_reservation(c); jffs2_free_raw_dirent(rd); - up(&dir_f->sem); + mutex_unlock(&dir_f->sem); jffs2_clear_inode(inode); return PTR_ERR(fd); } @@ -753,7 +753,7 @@ static int jffs2_mknod (struct inode *dir_i, struct dentry *dentry, int mode, de one if necessary. */ jffs2_add_fd_to_list(c, fd, &dir_f->dents); - up(&dir_f->sem); + mutex_unlock(&dir_f->sem); jffs2_complete_reservation(c); d_instantiate(dentry, inode); @@ -780,14 +780,14 @@ static int jffs2_rename (struct inode *old_dir_i, struct dentry *old_dentry, if (S_ISDIR(new_dentry->d_inode->i_mode)) { struct jffs2_full_dirent *fd; - down(&victim_f->sem); + mutex_lock(&victim_f->sem); for (fd = victim_f->dents; fd; fd = fd->next) { if (fd->ino) { - up(&victim_f->sem); + mutex_unlock(&victim_f->sem); return -ENOTEMPTY; } } - up(&victim_f->sem); + mutex_unlock(&victim_f->sem); } } @@ -816,9 +816,9 @@ static int jffs2_rename (struct inode *old_dir_i, struct dentry *old_dentry, /* Don't oops if the victim was a dirent pointing to an inode which didn't exist. */ if (victim_f->inocache) { - down(&victim_f->sem); + mutex_lock(&victim_f->sem); victim_f->inocache->nlink--; - up(&victim_f->sem); + mutex_unlock(&victim_f->sem); } } @@ -836,11 +836,11 @@ static int jffs2_rename (struct inode *old_dir_i, struct dentry *old_dentry, if (ret) { /* Oh shit. We really ought to make a single node which can do both atomically */ struct jffs2_inode_info *f = JFFS2_INODE_INFO(old_dentry->d_inode); - down(&f->sem); + mutex_lock(&f->sem); inc_nlink(old_dentry->d_inode); if (f->inocache) f->inocache->nlink++; - up(&f->sem); + mutex_unlock(&f->sem); printk(KERN_NOTICE "jffs2_rename(): Link succeeded, unlink failed (err %d). You now have a hard link\n", ret); /* Might as well let the VFS know */ diff --git a/fs/jffs2/erase.c b/fs/jffs2/erase.c index a1db9180633..96d9ad56e57 100644 --- a/fs/jffs2/erase.c +++ b/fs/jffs2/erase.c @@ -50,14 +50,14 @@ static void jffs2_erase_block(struct jffs2_sb_info *c, instr = kmalloc(sizeof(struct erase_info) + sizeof(struct erase_priv_struct), GFP_KERNEL); if (!instr) { printk(KERN_WARNING "kmalloc for struct erase_info in jffs2_erase_block failed. Refiling block for later\n"); - down(&c->erase_free_sem); + mutex_lock(&c->erase_free_sem); spin_lock(&c->erase_completion_lock); list_move(&jeb->list, &c->erase_pending_list); c->erasing_size -= c->sector_size; c->dirty_size += c->sector_size; jeb->dirty_size = c->sector_size; spin_unlock(&c->erase_completion_lock); - up(&c->erase_free_sem); + mutex_unlock(&c->erase_free_sem); return; } @@ -84,14 +84,14 @@ static void jffs2_erase_block(struct jffs2_sb_info *c, if (ret == -ENOMEM || ret == -EAGAIN) { /* Erase failed immediately. Refile it on the list */ D1(printk(KERN_DEBUG "Erase at 0x%08x failed: %d. Refiling on erase_pending_list\n", jeb->offset, ret)); - down(&c->erase_free_sem); + mutex_lock(&c->erase_free_sem); spin_lock(&c->erase_completion_lock); list_move(&jeb->list, &c->erase_pending_list); c->erasing_size -= c->sector_size; c->dirty_size += c->sector_size; jeb->dirty_size = c->sector_size; spin_unlock(&c->erase_completion_lock); - up(&c->erase_free_sem); + mutex_unlock(&c->erase_free_sem); return; } @@ -107,7 +107,7 @@ void jffs2_erase_pending_blocks(struct jffs2_sb_info *c, int count) { struct jffs2_eraseblock *jeb; - down(&c->erase_free_sem); + mutex_lock(&c->erase_free_sem); spin_lock(&c->erase_completion_lock); @@ -118,7 +118,7 @@ void jffs2_erase_pending_blocks(struct jffs2_sb_info *c, int count) jeb = list_entry(c->erase_complete_list.next, struct jffs2_eraseblock, list); list_del(&jeb->list); spin_unlock(&c->erase_completion_lock); - up(&c->erase_free_sem); + mutex_unlock(&c->erase_free_sem); jffs2_mark_erased_block(c, jeb); if (!--count) { @@ -139,7 +139,7 @@ void jffs2_erase_pending_blocks(struct jffs2_sb_info *c, int count) jffs2_free_jeb_node_refs(c, jeb); list_add(&jeb->list, &c->erasing_list); spin_unlock(&c->erase_completion_lock); - up(&c->erase_free_sem); + mutex_unlock(&c->erase_free_sem); jffs2_erase_block(c, jeb); @@ -149,12 +149,12 @@ void jffs2_erase_pending_blocks(struct jffs2_sb_info *c, int count) /* Be nice */ yield(); - down(&c->erase_free_sem); + mutex_lock(&c->erase_free_sem); spin_lock(&c->erase_completion_lock); } spin_unlock(&c->erase_completion_lock); - up(&c->erase_free_sem); + mutex_unlock(&c->erase_free_sem); done: D1(printk(KERN_DEBUG "jffs2_erase_pending_blocks completed\n")); } @@ -162,11 +162,11 @@ void jffs2_erase_pending_blocks(struct jffs2_sb_info *c, int count) static void jffs2_erase_succeeded(struct jffs2_sb_info *c, struct jffs2_eraseblock *jeb) { D1(printk(KERN_DEBUG "Erase completed successfully at 0x%08x\n", jeb->offset)); - down(&c->erase_free_sem); + mutex_lock(&c->erase_free_sem); spin_lock(&c->erase_completion_lock); list_move_tail(&jeb->list, &c->erase_complete_list); spin_unlock(&c->erase_completion_lock); - up(&c->erase_free_sem); + mutex_unlock(&c->erase_free_sem); /* Ensure that kupdated calls us again to mark them clean */ jffs2_erase_pending_trigger(c); } @@ -180,26 +180,26 @@ static void jffs2_erase_failed(struct jffs2_sb_info *c, struct jffs2_eraseblock failed too many times. */ if (!jffs2_write_nand_badblock(c, jeb, bad_offset)) { /* We'd like to give this block another try. */ - down(&c->erase_free_sem); + mutex_lock(&c->erase_free_sem); spin_lock(&c->erase_completion_lock); list_move(&jeb->list, &c->erase_pending_list); c->erasing_size -= c->sector_size; c->dirty_size += c->sector_size; jeb->dirty_size = c->sector_size; spin_unlock(&c->erase_completion_lock); - up(&c->erase_free_sem); + mutex_unlock(&c->erase_free_sem); return; } } - down(&c->erase_free_sem); + mutex_lock(&c->erase_free_sem); spin_lock(&c->erase_completion_lock); c->erasing_size -= c->sector_size; c->bad_size += c->sector_size; list_move(&jeb->list, &c->bad_list); c->nr_erasing_blocks--; spin_unlock(&c->erase_completion_lock); - up(&c->erase_free_sem); + mutex_unlock(&c->erase_free_sem); wake_up(&c->erase_wait); } @@ -456,7 +456,7 @@ static void jffs2_mark_erased_block(struct jffs2_sb_info *c, struct jffs2_eraseb jffs2_link_node_ref(c, jeb, jeb->offset | REF_NORMAL, c->cleanmarker_size, NULL); } - down(&c->erase_free_sem); + mutex_lock(&c->erase_free_sem); spin_lock(&c->erase_completion_lock); c->erasing_size -= c->sector_size; c->free_size += jeb->free_size; @@ -469,28 +469,28 @@ static void jffs2_mark_erased_block(struct jffs2_sb_info *c, struct jffs2_eraseb c->nr_erasing_blocks--; c->nr_free_blocks++; spin_unlock(&c->erase_completion_lock); - up(&c->erase_free_sem); + mutex_unlock(&c->erase_free_sem); wake_up(&c->erase_wait); return; filebad: - down(&c->erase_free_sem); + mutex_lock(&c->erase_free_sem); spin_lock(&c->erase_completion_lock); /* Stick it on a list (any list) so erase_failed can take it right off again. Silly, but shouldn't happen often. */ list_add(&jeb->list, &c->erasing_list); spin_unlock(&c->erase_completion_lock); - up(&c->erase_free_sem); + mutex_unlock(&c->erase_free_sem); jffs2_erase_failed(c, jeb, bad_offset); return; refile: /* Stick it back on the list from whence it came and come back later */ jffs2_erase_pending_trigger(c); - down(&c->erase_free_sem); + mutex_lock(&c->erase_free_sem); spin_lock(&c->erase_completion_lock); list_add(&jeb->list, &c->erase_complete_list); spin_unlock(&c->erase_completion_lock); - up(&c->erase_free_sem); + mutex_unlock(&c->erase_free_sem); return; } diff --git a/fs/jffs2/file.c b/fs/jffs2/file.c index dcc2734e0b5..5e920343b2c 100644 --- a/fs/jffs2/file.c +++ b/fs/jffs2/file.c @@ -115,9 +115,9 @@ static int jffs2_readpage (struct file *filp, struct page *pg) struct jffs2_inode_info *f = JFFS2_INODE_INFO(pg->mapping->host); int ret; - down(&f->sem); + mutex_lock(&f->sem); ret = jffs2_do_readpage_unlock(pg->mapping->host, pg); - up(&f->sem); + mutex_unlock(&f->sem); return ret; } @@ -154,7 +154,7 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping, if (ret) goto out_page; - down(&f->sem); + mutex_lock(&f->sem); memset(&ri, 0, sizeof(ri)); ri.magic = cpu_to_je16(JFFS2_MAGIC_BITMASK); @@ -181,7 +181,7 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping, if (IS_ERR(fn)) { ret = PTR_ERR(fn); jffs2_complete_reservation(c); - up(&f->sem); + mutex_unlock(&f->sem); goto out_page; } ret = jffs2_add_full_dnode_to_inode(c, f, fn); @@ -195,12 +195,12 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping, jffs2_mark_node_obsolete(c, fn->raw); jffs2_free_full_dnode(fn); jffs2_complete_reservation(c); - up(&f->sem); + mutex_unlock(&f->sem); goto out_page; } jffs2_complete_reservation(c); inode->i_size = pageofs; - up(&f->sem); + mutex_unlock(&f->sem); } /* @@ -209,9 +209,9 @@ static int jffs2_write_begin(struct file *filp, struct address_space *mapping, * case of a short-copy. */ if (!PageUptodate(pg)) { - down(&f->sem); + mutex_lock(&f->sem); ret = jffs2_do_readpage_nolock(inode, pg); - up(&f->sem); + mutex_unlock(&f->sem); if (ret) goto out_page; } diff --git a/fs/jffs2/fs.c b/fs/jffs2/fs.c index 9dafb53fb1d..3eb1c84b0a3 100644 --- a/fs/jffs2/fs.c +++ b/fs/jffs2/fs.c @@ -51,20 +51,20 @@ int jffs2_do_setattr (struct inode *inode, struct iattr *iattr) mdata = (char *)&dev; D1(printk(KERN_DEBUG "jffs2_setattr(): Writing %d bytes of kdev_t\n", mdatalen)); } else if (S_ISLNK(inode->i_mode)) { - down(&f->sem); + mutex_lock(&f->sem); mdatalen = f->metadata->size; mdata = kmalloc(f->metadata->size, GFP_USER); if (!mdata) { - up(&f->sem); + mutex_unlock(&f->sem); return -ENOMEM; } ret = jffs2_read_dnode(c, f, f->metadata, mdata, 0, mdatalen); if (ret) { - up(&f->sem); + mutex_unlock(&f->sem); kfree(mdata); return ret; } - up(&f->sem); + mutex_unlock(&f->sem); D1(printk(KERN_DEBUG "jffs2_setattr(): Writing %d bytes of symlink target\n", mdatalen)); } @@ -83,7 +83,7 @@ int jffs2_do_setattr (struct inode *inode, struct iattr *iattr) kfree(mdata); return ret; } - down(&f->sem); + mutex_lock(&f->sem); ivalid = iattr->ia_valid; ri->magic = cpu_to_je16(JFFS2_MAGIC_BITMASK); @@ -134,7 +134,7 @@ int jffs2_do_setattr (struct inode *inode, struct iattr *iattr) if (IS_ERR(new_metadata)) { jffs2_complete_reservation(c); jffs2_free_raw_inode(ri); - up(&f->sem); + mutex_unlock(&f->sem); return PTR_ERR(new_metadata); } /* It worked. Update the inode */ @@ -165,7 +165,7 @@ int jffs2_do_setattr (struct inode *inode, struct iattr *iattr) } jffs2_free_raw_inode(ri); - up(&f->sem); + mutex_unlock(&f->sem); jffs2_complete_reservation(c); /* We have to do the vmtruncate() without f->sem held, since @@ -256,12 +256,12 @@ struct inode *jffs2_iget(struct super_block *sb, unsigned long ino) c = JFFS2_SB_INFO(inode->i_sb); jffs2_init_inode_info(f); - down(&f->sem); + mutex_lock(&f->sem); ret = jffs2_do_read_inode(c, f, inode->i_ino, &latest_node); if (ret) { - up(&f->sem); + mutex_unlock(&f->sem); iget_failed(inode); return ERR_PTR(ret); } @@ -338,7 +338,7 @@ struct inode *jffs2_iget(struct super_block *sb, unsigned long ino) printk(KERN_WARNING "jffs2_read_inode(): Bogus imode %o for ino %lu\n", inode->i_mode, (unsigned long)inode->i_ino); } - up(&f->sem); + mutex_unlock(&f->sem); D1(printk(KERN_DEBUG "jffs2_read_inode() returning\n")); unlock_new_inode(inode); @@ -347,7 +347,7 @@ struct inode *jffs2_iget(struct super_block *sb, unsigned long ino) error_io: ret = -EIO; error: - up(&f->sem); + mutex_unlock(&f->sem); jffs2_do_clear_inode(c, f); iget_failed(inode); return ERR_PTR(ret); @@ -388,9 +388,9 @@ int jffs2_remount_fs (struct super_block *sb, int *flags, char *data) Flush the writebuffer, if neccecary, else we loose it */ if (!(sb->s_flags & MS_RDONLY)) { jffs2_stop_garbage_collect_thread(c); - down(&c->alloc_sem); + mutex_lock(&c->alloc_sem); jffs2_flush_wbuf_pad(c); - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); } if (!(*flags & MS_RDONLY)) @@ -437,7 +437,7 @@ struct inode *jffs2_new_inode (struct inode *dir_i, int mode, struct jffs2_raw_i f = JFFS2_INODE_INFO(inode); jffs2_init_inode_info(f); - down(&f->sem); + mutex_lock(&f->sem); memset(ri, 0, sizeof(*ri)); /* Set OS-specific defaults for new inodes */ diff --git a/fs/jffs2/gc.c b/fs/jffs2/gc.c index 59aeb982043..26c7992c45c 100644 --- a/fs/jffs2/gc.c +++ b/fs/jffs2/gc.c @@ -126,7 +126,7 @@ int jffs2_garbage_collect_pass(struct jffs2_sb_info *c) int ret = 0, inum, nlink; int xattr = 0; - if (down_interruptible(&c->alloc_sem)) + if (mutex_lock_interruptible(&c->alloc_sem)) return -EINTR; for (;;) { @@ -143,7 +143,7 @@ int jffs2_garbage_collect_pass(struct jffs2_sb_info *c) c->unchecked_size); jffs2_dbg_dump_block_lists_nolock(c); spin_unlock(&c->erase_completion_lock); - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); return -ENOSPC; } @@ -190,7 +190,7 @@ int jffs2_garbage_collect_pass(struct jffs2_sb_info *c) made no progress in this case, but that should be OK */ c->checked_ino--; - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); sleep_on_spinunlock(&c->inocache_wq, &c->inocache_lock); return 0; @@ -210,7 +210,7 @@ int jffs2_garbage_collect_pass(struct jffs2_sb_info *c) printk(KERN_WARNING "Returned error for crccheck of ino #%u. Expect badness...\n", ic->ino); jffs2_set_inocache_state(c, ic, INO_STATE_CHECKEDABSENT); - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); return ret; } @@ -223,7 +223,7 @@ int jffs2_garbage_collect_pass(struct jffs2_sb_info *c) if (!jeb) { D1 (printk(KERN_NOTICE "jffs2: Couldn't find erase block to garbage collect!\n")); spin_unlock(&c->erase_completion_lock); - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); return -EIO; } @@ -232,7 +232,7 @@ int jffs2_garbage_collect_pass(struct jffs2_sb_info *c) printk(KERN_DEBUG "Nextblock at %08x, used_size %08x, dirty_size %08x, wasted_size %08x, free_size %08x\n", c->nextblock->offset, c->nextblock->used_size, c->nextblock->dirty_size, c->nextblock->wasted_size, c->nextblock->free_size)); if (!jeb->used_size) { - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); goto eraseit; } @@ -248,7 +248,7 @@ int jffs2_garbage_collect_pass(struct jffs2_sb_info *c) jeb->offset, jeb->free_size, jeb->dirty_size, jeb->used_size); jeb->gc_node = raw; spin_unlock(&c->erase_completion_lock); - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); BUG(); } } @@ -266,7 +266,7 @@ int jffs2_garbage_collect_pass(struct jffs2_sb_info *c) /* Just mark it obsolete */ jffs2_mark_node_obsolete(c, raw); } - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); goto eraseit_lock; } @@ -334,7 +334,7 @@ int jffs2_garbage_collect_pass(struct jffs2_sb_info *c) */ printk(KERN_CRIT "Inode #%u already in state %d in jffs2_garbage_collect_pass()!\n", ic->ino, ic->state); - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); spin_unlock(&c->inocache_lock); BUG(); @@ -345,7 +345,7 @@ int jffs2_garbage_collect_pass(struct jffs2_sb_info *c) the alloc_sem() (for marking nodes invalid) so we must drop the alloc_sem before sleeping. */ - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); D1(printk(KERN_DEBUG "jffs2_garbage_collect_pass() waiting for ino #%u in state %d\n", ic->ino, ic->state)); sleep_on_spinunlock(&c->inocache_wq, &c->inocache_lock); @@ -416,7 +416,7 @@ int jffs2_garbage_collect_pass(struct jffs2_sb_info *c) ret = -ENOSPC; } release_sem: - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); eraseit_lock: /* If we've finished this block, start it erasing */ @@ -445,7 +445,7 @@ static int jffs2_garbage_collect_live(struct jffs2_sb_info *c, struct jffs2_era uint32_t start = 0, end = 0, nrfrags = 0; int ret = 0; - down(&f->sem); + mutex_lock(&f->sem); /* Now we have the lock for this inode. Check that it's still the one at the head of the list. */ @@ -525,7 +525,7 @@ static int jffs2_garbage_collect_live(struct jffs2_sb_info *c, struct jffs2_era } } upnout: - up(&f->sem); + mutex_unlock(&f->sem); return ret; } @@ -846,7 +846,7 @@ static int jffs2_garbage_collect_deletion_dirent(struct jffs2_sb_info *c, struct /* Prevent the erase code from nicking the obsolete node refs while we're looking at them. I really don't like this extra lock but can't see any alternative. Suggestions on a postcard to... */ - down(&c->erase_free_sem); + mutex_lock(&c->erase_free_sem); for (raw = f->inocache->nodes; raw != (void *)f->inocache; raw = raw->next_in_ino) { @@ -899,7 +899,7 @@ static int jffs2_garbage_collect_deletion_dirent(struct jffs2_sb_info *c, struct /* OK. The name really does match. There really is still an older node on the flash which our deletion dirent obsoletes. So we have to write out a new deletion dirent to replace it */ - up(&c->erase_free_sem); + mutex_unlock(&c->erase_free_sem); D1(printk(KERN_DEBUG "Deletion dirent at %08x still obsoletes real dirent \"%s\" at %08x for ino #%u\n", ref_offset(fd->raw), fd->name, ref_offset(raw), je32_to_cpu(rd->ino))); @@ -908,7 +908,7 @@ static int jffs2_garbage_collect_deletion_dirent(struct jffs2_sb_info *c, struct return jffs2_garbage_collect_dirent(c, jeb, f, fd); } - up(&c->erase_free_sem); + mutex_unlock(&c->erase_free_sem); kfree(rd); } diff --git a/fs/jffs2/jffs2_fs_i.h b/fs/jffs2/jffs2_fs_i.h index a841f4973a7..b5427b5c948 100644 --- a/fs/jffs2/jffs2_fs_i.h +++ b/fs/jffs2/jffs2_fs_i.h @@ -24,7 +24,7 @@ struct jffs2_inode_info { before letting GC proceed. Or we'd have to put ugliness into the GC code so it didn't attempt to obtain the i_mutex for the inode(s) which are already locked */ - struct semaphore sem; + struct mutex sem; /* The highest (datanode) version number used for this ino */ uint32_t highest_version; diff --git a/fs/jffs2/jffs2_fs_sb.h b/fs/jffs2/jffs2_fs_sb.h index 18fca2b9e53..d9c4b27575e 100644 --- a/fs/jffs2/jffs2_fs_sb.h +++ b/fs/jffs2/jffs2_fs_sb.h @@ -16,7 +16,7 @@ #include #include #include -#include +#include #include #include #include @@ -44,7 +44,7 @@ struct jffs2_sb_info { struct completion gc_thread_start; /* GC thread start completion */ struct completion gc_thread_exit; /* GC thread exit completion port */ - struct semaphore alloc_sem; /* Used to protect all the following + struct mutex alloc_sem; /* Used to protect all the following fields, and also to protect against out-of-order writing of nodes. And GC. */ uint32_t cleanmarker_size; /* Size of an _inline_ CLEANMARKER @@ -104,7 +104,7 @@ struct jffs2_sb_info { /* Sem to allow jffs2_garbage_collect_deletion_dirent to drop the erase_completion_lock while it's holding a pointer to an obsoleted node. I don't like this. Alternatives welcomed. */ - struct semaphore erase_free_sem; + struct mutex erase_free_sem; uint32_t wbuf_pagesize; /* 0 for NOR and other flashes with no wbuf */ diff --git a/fs/jffs2/nodemgmt.c b/fs/jffs2/nodemgmt.c index a3e67a4ed67..747a73f0aa4 100644 --- a/fs/jffs2/nodemgmt.c +++ b/fs/jffs2/nodemgmt.c @@ -48,7 +48,7 @@ int jffs2_reserve_space(struct jffs2_sb_info *c, uint32_t minsize, minsize = PAD(minsize); D1(printk(KERN_DEBUG "jffs2_reserve_space(): Requested 0x%x bytes\n", minsize)); - down(&c->alloc_sem); + mutex_lock(&c->alloc_sem); D1(printk(KERN_DEBUG "jffs2_reserve_space(): alloc sem got\n")); @@ -81,7 +81,7 @@ int jffs2_reserve_space(struct jffs2_sb_info *c, uint32_t minsize, dirty, c->unchecked_size, c->sector_size)); spin_unlock(&c->erase_completion_lock); - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); return -ENOSPC; } @@ -104,11 +104,11 @@ int jffs2_reserve_space(struct jffs2_sb_info *c, uint32_t minsize, D1(printk(KERN_DEBUG "max. available size 0x%08x < blocksneeded * sector_size 0x%08x, returning -ENOSPC\n", avail, blocksneeded * c->sector_size)); spin_unlock(&c->erase_completion_lock); - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); return -ENOSPC; } - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); D1(printk(KERN_DEBUG "Triggering GC pass. nr_free_blocks %d, nr_erasing_blocks %d, free_size 0x%08x, dirty_size 0x%08x, wasted_size 0x%08x, used_size 0x%08x, erasing_size 0x%08x, bad_size 0x%08x (total 0x%08x of 0x%08x)\n", c->nr_free_blocks, c->nr_erasing_blocks, c->free_size, c->dirty_size, c->wasted_size, c->used_size, c->erasing_size, c->bad_size, @@ -124,7 +124,7 @@ int jffs2_reserve_space(struct jffs2_sb_info *c, uint32_t minsize, if (signal_pending(current)) return -EINTR; - down(&c->alloc_sem); + mutex_lock(&c->alloc_sem); spin_lock(&c->erase_completion_lock); } @@ -137,7 +137,7 @@ int jffs2_reserve_space(struct jffs2_sb_info *c, uint32_t minsize, if (!ret) ret = jffs2_prealloc_raw_node_refs(c, c->nextblock, 1); if (ret) - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); return ret; } @@ -462,7 +462,7 @@ void jffs2_complete_reservation(struct jffs2_sb_info *c) { D1(printk(KERN_DEBUG "jffs2_complete_reservation()\n")); jffs2_garbage_collect_trigger(c); - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); } static inline int on_list(struct list_head *obj, struct list_head *head) @@ -511,7 +511,7 @@ void jffs2_mark_node_obsolete(struct jffs2_sb_info *c, struct jffs2_raw_node_ref any jffs2_raw_node_refs. So we don't need to stop erases from happening, or protect against people holding an obsolete jffs2_raw_node_ref without the erase_completion_lock. */ - down(&c->erase_free_sem); + mutex_lock(&c->erase_free_sem); } spin_lock(&c->erase_completion_lock); @@ -714,7 +714,7 @@ void jffs2_mark_node_obsolete(struct jffs2_sb_info *c, struct jffs2_raw_node_ref } out_erase_sem: - up(&c->erase_free_sem); + mutex_unlock(&c->erase_free_sem); } int jffs2_thread_should_wake(struct jffs2_sb_info *c) diff --git a/fs/jffs2/readinode.c b/fs/jffs2/readinode.c index e512a93d624..8a7cf1e8d68 100644 --- a/fs/jffs2/readinode.c +++ b/fs/jffs2/readinode.c @@ -1193,7 +1193,7 @@ static int jffs2_do_read_inode_internal(struct jffs2_sb_info *c, JFFS2_ERROR("failed to read from flash: error %d, %zd of %zd bytes read\n", ret, retlen, sizeof(*latest_node)); /* FIXME: If this fails, there seems to be a memory leak. Find it. */ - up(&f->sem); + mutex_unlock(&f->sem); jffs2_do_clear_inode(c, f); return ret?ret:-EIO; } @@ -1202,7 +1202,7 @@ static int jffs2_do_read_inode_internal(struct jffs2_sb_info *c, if (crc != je32_to_cpu(latest_node->node_crc)) { JFFS2_ERROR("CRC failed for read_inode of inode %u at physical location 0x%x\n", f->inocache->ino, ref_offset(rii.latest_ref)); - up(&f->sem); + mutex_unlock(&f->sem); jffs2_do_clear_inode(c, f); return -EIO; } @@ -1242,7 +1242,7 @@ static int jffs2_do_read_inode_internal(struct jffs2_sb_info *c, f->target = kmalloc(je32_to_cpu(latest_node->csize) + 1, GFP_KERNEL); if (!f->target) { JFFS2_ERROR("can't allocate %d bytes of memory for the symlink target path cache\n", je32_to_cpu(latest_node->csize)); - up(&f->sem); + mutex_unlock(&f->sem); jffs2_do_clear_inode(c, f); return -ENOMEM; } @@ -1255,7 +1255,7 @@ static int jffs2_do_read_inode_internal(struct jffs2_sb_info *c, ret = -EIO; kfree(f->target); f->target = NULL; - up(&f->sem); + mutex_unlock(&f->sem); jffs2_do_clear_inode(c, f); return -ret; } @@ -1273,14 +1273,14 @@ static int jffs2_do_read_inode_internal(struct jffs2_sb_info *c, if (f->metadata) { JFFS2_ERROR("Argh. Special inode #%u with mode 0%o had metadata node\n", f->inocache->ino, jemode_to_cpu(latest_node->mode)); - up(&f->sem); + mutex_unlock(&f->sem); jffs2_do_clear_inode(c, f); return -EIO; } if (!frag_first(&f->fragtree)) { JFFS2_ERROR("Argh. Special inode #%u with mode 0%o has no fragments\n", f->inocache->ino, jemode_to_cpu(latest_node->mode)); - up(&f->sem); + mutex_unlock(&f->sem); jffs2_do_clear_inode(c, f); return -EIO; } @@ -1289,7 +1289,7 @@ static int jffs2_do_read_inode_internal(struct jffs2_sb_info *c, JFFS2_ERROR("Argh. Special inode #%u with mode 0x%x had more than one node\n", f->inocache->ino, jemode_to_cpu(latest_node->mode)); /* FIXME: Deal with it - check crc32, check for duplicate node, check times and discard the older one */ - up(&f->sem); + mutex_unlock(&f->sem); jffs2_do_clear_inode(c, f); return -EIO; } @@ -1379,12 +1379,13 @@ int jffs2_do_crccheck_inode(struct jffs2_sb_info *c, struct jffs2_inode_cache *i if (!f) return -ENOMEM; - init_MUTEX_LOCKED(&f->sem); + mutex_init(&f->sem); + mutex_lock(&f->sem); f->inocache = ic; ret = jffs2_do_read_inode_internal(c, f, &n); if (!ret) { - up(&f->sem); + mutex_unlock(&f->sem); jffs2_do_clear_inode(c, f); } kfree (f); @@ -1398,7 +1399,7 @@ void jffs2_do_clear_inode(struct jffs2_sb_info *c, struct jffs2_inode_info *f) jffs2_clear_acl(f); jffs2_xattr_delete_inode(c, f->inocache); - down(&f->sem); + mutex_lock(&f->sem); deleted = f->inocache && !f->inocache->nlink; if (f->inocache && f->inocache->state != INO_STATE_CHECKING) @@ -1430,5 +1431,5 @@ void jffs2_do_clear_inode(struct jffs2_sb_info *c, struct jffs2_inode_info *f) jffs2_del_ino_cache(c, f->inocache); } - up(&f->sem); + mutex_unlock(&f->sem); } diff --git a/fs/jffs2/super.c b/fs/jffs2/super.c index 4677355996c..f3353df178e 100644 --- a/fs/jffs2/super.c +++ b/fs/jffs2/super.c @@ -47,7 +47,7 @@ static void jffs2_i_init_once(struct kmem_cache *cachep, void *foo) { struct jffs2_inode_info *ei = (struct jffs2_inode_info *) foo; - init_MUTEX(&ei->sem); + mutex_init(&ei->sem); inode_init_once(&ei->vfs_inode); } @@ -55,9 +55,9 @@ static int jffs2_sync_fs(struct super_block *sb, int wait) { struct jffs2_sb_info *c = JFFS2_SB_INFO(sb); - down(&c->alloc_sem); + mutex_lock(&c->alloc_sem); jffs2_flush_wbuf_pad(c); - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); return 0; } @@ -95,8 +95,8 @@ static int jffs2_fill_super(struct super_block *sb, void *data, int silent) /* Initialize JFFS2 superblock locks, the further initialization will * be done later */ - init_MUTEX(&c->alloc_sem); - init_MUTEX(&c->erase_free_sem); + mutex_init(&c->alloc_sem); + mutex_init(&c->erase_free_sem); init_waitqueue_head(&c->erase_wait); init_waitqueue_head(&c->inocache_wq); spin_lock_init(&c->erase_completion_lock); @@ -125,9 +125,9 @@ static void jffs2_put_super (struct super_block *sb) D2(printk(KERN_DEBUG "jffs2: jffs2_put_super()\n")); - down(&c->alloc_sem); + mutex_lock(&c->alloc_sem); jffs2_flush_wbuf_pad(c); - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); jffs2_sum_exit(c); diff --git a/fs/jffs2/wbuf.c b/fs/jffs2/wbuf.c index ba49f19cff8..8de52b60767 100644 --- a/fs/jffs2/wbuf.c +++ b/fs/jffs2/wbuf.c @@ -578,8 +578,8 @@ static int __jffs2_flush_wbuf(struct jffs2_sb_info *c, int pad) if (!jffs2_is_writebuffered(c)) return 0; - if (!down_trylock(&c->alloc_sem)) { - up(&c->alloc_sem); + if (mutex_trylock(&c->alloc_sem)) { + mutex_unlock(&c->alloc_sem); printk(KERN_CRIT "jffs2_flush_wbuf() called with alloc_sem not locked!\n"); BUG(); } @@ -702,10 +702,10 @@ int jffs2_flush_wbuf_gc(struct jffs2_sb_info *c, uint32_t ino) if (!c->wbuf) return 0; - down(&c->alloc_sem); + mutex_lock(&c->alloc_sem); if (!jffs2_wbuf_pending_for_ino(c, ino)) { D1(printk(KERN_DEBUG "Ino #%d not pending in wbuf. Returning\n", ino)); - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); return 0; } @@ -725,14 +725,14 @@ int jffs2_flush_wbuf_gc(struct jffs2_sb_info *c, uint32_t ino) } else while (old_wbuf_len && old_wbuf_ofs == c->wbuf_ofs) { - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); D1(printk(KERN_DEBUG "jffs2_flush_wbuf_gc() calls gc pass\n")); ret = jffs2_garbage_collect_pass(c); if (ret) { /* GC failed. Flush it with padding instead */ - down(&c->alloc_sem); + mutex_lock(&c->alloc_sem); down_write(&c->wbuf_sem); ret = __jffs2_flush_wbuf(c, PAD_ACCOUNTING); /* retry flushing wbuf in case jffs2_wbuf_recover @@ -742,12 +742,12 @@ int jffs2_flush_wbuf_gc(struct jffs2_sb_info *c, uint32_t ino) up_write(&c->wbuf_sem); break; } - down(&c->alloc_sem); + mutex_lock(&c->alloc_sem); } D1(printk(KERN_DEBUG "jffs2_flush_wbuf_gc() ends...\n")); - up(&c->alloc_sem); + mutex_unlock(&c->alloc_sem); return ret; } diff --git a/fs/jffs2/write.c b/fs/jffs2/write.c index beade550909..665fce9797d 100644 --- a/fs/jffs2/write.c +++ b/fs/jffs2/write.c @@ -137,12 +137,12 @@ struct jffs2_full_dnode *jffs2_write_dnode(struct jffs2_sb_info *c, struct jffs2 JFFS2_SUMMARY_INODE_SIZE); } else { /* Locking pain */ - up(&f->sem); + mutex_unlock(&f->sem); jffs2_complete_reservation(c); ret = jffs2_reserve_space(c, sizeof(*ri) + datalen, &dummy, alloc_mode, JFFS2_SUMMARY_INODE_SIZE); - down(&f->sem); + mutex_lock(&f->sem); } if (!ret) { @@ -285,12 +285,12 @@ struct jffs2_full_dirent *jffs2_write_dirent(struct jffs2_sb_info *c, struct jff JFFS2_SUMMARY_DIRENT_SIZE(namelen)); } else { /* Locking pain */ - up(&f->sem); + mutex_unlock(&f->sem); jffs2_complete_reservation(c); ret = jffs2_reserve_space(c, sizeof(*rd) + namelen, &dummy, alloc_mode, JFFS2_SUMMARY_DIRENT_SIZE(namelen)); - down(&f->sem); + mutex_lock(&f->sem); } if (!ret) { @@ -353,7 +353,7 @@ int jffs2_write_inode_range(struct jffs2_sb_info *c, struct jffs2_inode_info *f, D1(printk(KERN_DEBUG "jffs2_reserve_space returned %d\n", ret)); break; } - down(&f->sem); + mutex_lock(&f->sem); datalen = min_t(uint32_t, writelen, PAGE_CACHE_SIZE - (offset & (PAGE_CACHE_SIZE-1))); cdatalen = min_t(uint32_t, alloclen - sizeof(*ri), datalen); @@ -381,7 +381,7 @@ int jffs2_write_inode_range(struct jffs2_sb_info *c, struct jffs2_inode_info *f, if (IS_ERR(fn)) { ret = PTR_ERR(fn); - up(&f->sem); + mutex_unlock(&f->sem); jffs2_complete_reservation(c); if (!retried) { /* Write error to be retried */ @@ -403,11 +403,11 @@ int jffs2_write_inode_range(struct jffs2_sb_info *c, struct jffs2_inode_info *f, jffs2_mark_node_obsolete(c, fn->raw); jffs2_free_full_dnode(fn); - up(&f->sem); + mutex_unlock(&f->sem); jffs2_complete_reservation(c); break; } - up(&f->sem); + mutex_unlock(&f->sem); jffs2_complete_reservation(c); if (!datalen) { printk(KERN_WARNING "Eep. We didn't actually write any data in jffs2_write_inode_range()\n"); @@ -439,7 +439,7 @@ int jffs2_do_create(struct jffs2_sb_info *c, struct jffs2_inode_info *dir_f, str JFFS2_SUMMARY_INODE_SIZE); D1(printk(KERN_DEBUG "jffs2_do_create(): reserved 0x%x bytes\n", alloclen)); if (ret) { - up(&f->sem); + mutex_unlock(&f->sem); return ret; } @@ -454,7 +454,7 @@ int jffs2_do_create(struct jffs2_sb_info *c, struct jffs2_inode_info *dir_f, str if (IS_ERR(fn)) { D1(printk(KERN_DEBUG "jffs2_write_dnode() failed\n")); /* Eeek. Wave bye bye */ - up(&f->sem); + mutex_unlock(&f->sem); jffs2_complete_reservation(c); return PTR_ERR(fn); } @@ -463,7 +463,7 @@ int jffs2_do_create(struct jffs2_sb_info *c, struct jffs2_inode_info *dir_f, str */ f->metadata = fn; - up(&f->sem); + mutex_unlock(&f->sem); jffs2_complete_reservation(c); ret = jffs2_init_security(&f->vfs_inode, &dir_f->vfs_inode); @@ -489,7 +489,7 @@ int jffs2_do_create(struct jffs2_sb_info *c, struct jffs2_inode_info *dir_f, str return -ENOMEM; } - down(&dir_f->sem); + mutex_lock(&dir_f->sem); rd->magic = cpu_to_je16(JFFS2_MAGIC_BITMASK); rd->nodetype = cpu_to_je16(JFFS2_NODETYPE_DIRENT); @@ -513,7 +513,7 @@ int jffs2_do_create(struct jffs2_sb_info *c, struct jffs2_inode_info *dir_f, str /* dirent failed to write. Delete the inode normally as if it were the final unlink() */ jffs2_complete_reservation(c); - up(&dir_f->sem); + mutex_unlock(&dir_f->sem); return PTR_ERR(fd); } @@ -522,7 +522,7 @@ int jffs2_do_create(struct jffs2_sb_info *c, struct jffs2_inode_info *dir_f, str jffs2_add_fd_to_list(c, fd, &dir_f->dents); jffs2_complete_reservation(c); - up(&dir_f->sem); + mutex_unlock(&dir_f->sem); return 0; } @@ -551,7 +551,7 @@ int jffs2_do_unlink(struct jffs2_sb_info *c, struct jffs2_inode_info *dir_f, return ret; } - down(&dir_f->sem); + mutex_lock(&dir_f->sem); /* Build a deletion node */ rd->magic = cpu_to_je16(JFFS2_MAGIC_BITMASK); @@ -574,21 +574,21 @@ int jffs2_do_unlink(struct jffs2_sb_info *c, struct jffs2_inode_info *dir_f, if (IS_ERR(fd)) { jffs2_complete_reservation(c); - up(&dir_f->sem); + mutex_unlock(&dir_f->sem); return PTR_ERR(fd); } /* File it. This will mark the old one obsolete. */ jffs2_add_fd_to_list(c, fd, &dir_f->dents); - up(&dir_f->sem); + mutex_unlock(&dir_f->sem); } else { uint32_t nhash = full_name_hash(name, namelen); fd = dir_f->dents; /* We don't actually want to reserve any space, but we do want to be holding the alloc_sem when we write to flash */ - down(&c->alloc_sem); - down(&dir_f->sem); + mutex_lock(&c->alloc_sem); + mutex_lock(&dir_f->sem); for (fd = dir_f->dents; fd; fd = fd->next) { if (fd->nhash == nhash && @@ -607,7 +607,7 @@ int jffs2_do_unlink(struct jffs2_sb_info *c, struct jffs2_inode_info *dir_f, break; } } - up(&dir_f->sem); + mutex_unlock(&dir_f->sem); } /* dead_f is NULL if this was a rename not a real unlink */ @@ -615,7 +615,7 @@ int jffs2_do_unlink(struct jffs2_sb_info *c, struct jffs2_inode_info *dir_f, pointing to an inode which didn't exist. */ if (dead_f && dead_f->inocache) { - down(&dead_f->sem); + mutex_lock(&dead_f->sem); if (S_ISDIR(OFNI_EDONI_2SFFJ(dead_f)->i_mode)) { while (dead_f->dents) { @@ -639,7 +639,7 @@ int jffs2_do_unlink(struct jffs2_sb_info *c, struct jffs2_inode_info *dir_f, dead_f->inocache->nlink--; /* NB: Caller must set inode nlink if appropriate */ - up(&dead_f->sem); + mutex_unlock(&dead_f->sem); } jffs2_complete_reservation(c); @@ -666,7 +666,7 @@ int jffs2_do_link (struct jffs2_sb_info *c, struct jffs2_inode_info *dir_f, uint return ret; } - down(&dir_f->sem); + mutex_lock(&dir_f->sem); /* Build a deletion node */ rd->magic = cpu_to_je16(JFFS2_MAGIC_BITMASK); @@ -691,7 +691,7 @@ int jffs2_do_link (struct jffs2_sb_info *c, struct jffs2_inode_info *dir_f, uint if (IS_ERR(fd)) { jffs2_complete_reservation(c); - up(&dir_f->sem); + mutex_unlock(&dir_f->sem); return PTR_ERR(fd); } @@ -699,7 +699,7 @@ int jffs2_do_link (struct jffs2_sb_info *c, struct jffs2_inode_info *dir_f, uint jffs2_add_fd_to_list(c, fd, &dir_f->dents); jffs2_complete_reservation(c); - up(&dir_f->sem); + mutex_unlock(&dir_f->sem); return 0; } -- cgit v1.2.3 From 0f4238958d28044b335644b69df6071cdb04b5ce Mon Sep 17 00:00:00 2001 From: James Bottomley Date: Thu, 20 Mar 2008 20:47:52 -0500 Subject: [SCSI] sysfs: make group is_valid return a mode_t We have a problem in scsi_transport_spi in that we need to customise not only the visibility of the attributes, but also their mode. Fix this by making the is_visible() callback return a mode, with 0 indicating is not visible. Also add a sysfs_update_group() API to allow us to change either the visibility or mode of the files at any time on the fly. Acked-by: Kay Sievers Signed-off-by: James Bottomley --- fs/sysfs/file.c | 14 +++++++--- fs/sysfs/group.c | 83 +++++++++++++++++++++++++++++++++++++++++++++++--------- fs/sysfs/sysfs.h | 2 ++ 3 files changed, 82 insertions(+), 17 deletions(-) (limited to 'fs') diff --git a/fs/sysfs/file.c b/fs/sysfs/file.c index ade9a7e6a75..dbdfabbfd60 100644 --- a/fs/sysfs/file.c +++ b/fs/sysfs/file.c @@ -477,11 +477,10 @@ const struct file_operations sysfs_file_operations = { .poll = sysfs_poll, }; - -int sysfs_add_file(struct sysfs_dirent *dir_sd, const struct attribute *attr, - int type) +int sysfs_add_file_mode(struct sysfs_dirent *dir_sd, + const struct attribute *attr, int type, mode_t amode) { - umode_t mode = (attr->mode & S_IALLUGO) | S_IFREG; + umode_t mode = (amode & S_IALLUGO) | S_IFREG; struct sysfs_addrm_cxt acxt; struct sysfs_dirent *sd; int rc; @@ -502,6 +501,13 @@ int sysfs_add_file(struct sysfs_dirent *dir_sd, const struct attribute *attr, } +int sysfs_add_file(struct sysfs_dirent *dir_sd, const struct attribute *attr, + int type) +{ + return sysfs_add_file_mode(dir_sd, attr, type, attr->mode); +} + + /** * sysfs_create_file - create an attribute file for an object. * @kobj: object we're creating for. diff --git a/fs/sysfs/group.c b/fs/sysfs/group.c index 47790491503..eeba38417b1 100644 --- a/fs/sysfs/group.c +++ b/fs/sysfs/group.c @@ -23,35 +23,50 @@ static void remove_files(struct sysfs_dirent *dir_sd, struct kobject *kobj, int i; for (i = 0, attr = grp->attrs; *attr; i++, attr++) - if (!grp->is_visible || - grp->is_visible(kobj, *attr, i)) - sysfs_hash_and_remove(dir_sd, (*attr)->name); + sysfs_hash_and_remove(dir_sd, (*attr)->name); } static int create_files(struct sysfs_dirent *dir_sd, struct kobject *kobj, - const struct attribute_group *grp) + const struct attribute_group *grp, int update) { struct attribute *const* attr; int error = 0, i; - for (i = 0, attr = grp->attrs; *attr && !error; i++, attr++) - if (!grp->is_visible || - grp->is_visible(kobj, *attr, i)) - error |= - sysfs_add_file(dir_sd, *attr, SYSFS_KOBJ_ATTR); + for (i = 0, attr = grp->attrs; *attr && !error; i++, attr++) { + mode_t mode = 0; + + /* in update mode, we're changing the permissions or + * visibility. Do this by first removing then + * re-adding (if required) the file */ + if (update) + sysfs_hash_and_remove(dir_sd, (*attr)->name); + if (grp->is_visible) { + mode = grp->is_visible(kobj, *attr, i); + if (!mode) + continue; + } + error = sysfs_add_file_mode(dir_sd, *attr, SYSFS_KOBJ_ATTR, + (*attr)->mode | mode); + if (unlikely(error)) + break; + } if (error) remove_files(dir_sd, kobj, grp); return error; } -int sysfs_create_group(struct kobject * kobj, - const struct attribute_group * grp) +static int internal_create_group(struct kobject *kobj, int update, + const struct attribute_group *grp) { struct sysfs_dirent *sd; int error; - BUG_ON(!kobj || !kobj->sd); + BUG_ON(!kobj || (!update && !kobj->sd)); + + /* Updates may happen before the object has been instantiated */ + if (unlikely(update && !kobj->sd)) + return -EINVAL; if (grp->name) { error = sysfs_create_subdir(kobj, grp->name, &sd); @@ -60,7 +75,7 @@ int sysfs_create_group(struct kobject * kobj, } else sd = kobj->sd; sysfs_get(sd); - error = create_files(sd, kobj, grp); + error = create_files(sd, kobj, grp, update); if (error) { if (grp->name) sysfs_remove_subdir(sd); @@ -69,6 +84,47 @@ int sysfs_create_group(struct kobject * kobj, return error; } +/** + * sysfs_create_group - given a directory kobject, create an attribute group + * @kobj: The kobject to create the group on + * @grp: The attribute group to create + * + * This function creates a group for the first time. It will explicitly + * warn and error if any of the attribute files being created already exist. + * + * Returns 0 on success or error. + */ +int sysfs_create_group(struct kobject *kobj, + const struct attribute_group *grp) +{ + return internal_create_group(kobj, 0, grp); +} + +/** + * sysfs_update_group - given a directory kobject, create an attribute group + * @kobj: The kobject to create the group on + * @grp: The attribute group to create + * + * This function updates an attribute group. Unlike + * sysfs_create_group(), it will explicitly not warn or error if any + * of the attribute files being created already exist. Furthermore, + * if the visibility of the files has changed through the is_visible() + * callback, it will update the permissions and add or remove the + * relevant files. + * + * The primary use for this function is to call it after making a change + * that affects group visibility. + * + * Returns 0 on success or error. + */ +int sysfs_update_group(struct kobject *kobj, + const struct attribute_group *grp) +{ + return internal_create_group(kobj, 1, grp); +} + + + void sysfs_remove_group(struct kobject * kobj, const struct attribute_group * grp) { @@ -95,4 +151,5 @@ void sysfs_remove_group(struct kobject * kobj, EXPORT_SYMBOL_GPL(sysfs_create_group); +EXPORT_SYMBOL_GPL(sysfs_update_group); EXPORT_SYMBOL_GPL(sysfs_remove_group); diff --git a/fs/sysfs/sysfs.h b/fs/sysfs/sysfs.h index ff17f8da9b4..ce4e15f8aae 100644 --- a/fs/sysfs/sysfs.h +++ b/fs/sysfs/sysfs.h @@ -154,6 +154,8 @@ extern const struct file_operations sysfs_file_operations; int sysfs_add_file(struct sysfs_dirent *dir_sd, const struct attribute *attr, int type); +int sysfs_add_file_mode(struct sysfs_dirent *dir_sd, + const struct attribute *attr, int type, mode_t amode); /* * bin.c */ -- cgit v1.2.3 From 16abef0e9e79643827fd5a2a14a07bced851ae72 Mon Sep 17 00:00:00 2001 From: David Sterba Date: Tue, 22 Apr 2008 15:09:22 +0200 Subject: fs: use loff_t type instead of long long Use offset type consistently. Signed-off-by: David Sterba Signed-off-by: Linus Torvalds --- fs/read_write.c | 6 +++--- fs/seq_file.c | 2 +- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/read_write.c b/fs/read_write.c index 49a98718ecd..f0d1240a5c6 100644 --- a/fs/read_write.c +++ b/fs/read_write.c @@ -33,7 +33,7 @@ EXPORT_SYMBOL(generic_ro_fops); loff_t generic_file_llseek(struct file *file, loff_t offset, int origin) { - long long retval; + loff_t retval; struct inode *inode = file->f_mapping->host; mutex_lock(&inode->i_mutex); @@ -60,7 +60,7 @@ EXPORT_SYMBOL(generic_file_llseek); loff_t remote_llseek(struct file *file, loff_t offset, int origin) { - long long retval; + loff_t retval; lock_kernel(); switch (origin) { @@ -91,7 +91,7 @@ EXPORT_SYMBOL(no_llseek); loff_t default_llseek(struct file *file, loff_t offset, int origin) { - long long retval; + loff_t retval; lock_kernel(); switch (origin) { diff --git a/fs/seq_file.c b/fs/seq_file.c index 853770274f2..9943408b315 100644 --- a/fs/seq_file.c +++ b/fs/seq_file.c @@ -239,7 +239,7 @@ Eoverflow: loff_t seq_lseek(struct file *file, loff_t offset, int origin) { struct seq_file *m = (struct seq_file *)file->private_data; - long long retval = -EINVAL; + loff_t retval = -EINVAL; mutex_lock(&m->lock); m->version = file->f_version; -- cgit v1.2.3 From cf9d1e428cc28ef5798aeda0822a6ce64849a439 Mon Sep 17 00:00:00 2001 From: David Woodhouse Date: Tue, 22 Apr 2008 23:53:26 +0100 Subject: [JFFS2] Self-sufficient #includes in jffs2_fs_i.h: include ... instead of which we don't need any more anyway. Signed-off-by: David Woodhouse --- fs/jffs2/jffs2_fs_i.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/jffs2/jffs2_fs_i.h b/fs/jffs2/jffs2_fs_i.h index b5427b5c948..31559f45fdd 100644 --- a/fs/jffs2/jffs2_fs_i.h +++ b/fs/jffs2/jffs2_fs_i.h @@ -15,7 +15,7 @@ #include #include #include -#include +#include struct jffs2_inode_info { /* We need an internal mutex similar to inode->i_mutex. -- cgit v1.2.3 From 014b164e1392a166fe96e003d2f0e7ad2e2a0bb7 Mon Sep 17 00:00:00 2001 From: David Woodhouse Date: Tue, 22 Apr 2008 23:54:38 +0100 Subject: [JFFS2] Fix free space leak with in-band cleanmarkers We were accounting for the cleanmarker by calling jffs2_link_node_ref() (without locking!), which adjusted both superblock and per-eraseblock accounting, subtracting the size of the cleanmarker from {jeb,c}->free_size and adding it to {jeb,c}->used_size. But only _then_ were we adding the size of the newly-erased block back to the superblock counts, and we were adding each of jeb->{free,used}_size to the corresponding superblock counts. Thus, the size of the cleanmarker was effectively subtracted from the superblock's free_size _twice_. Fix this, by always adding a full eraseblock size to c->free_size when we've erased a block. And call jffs2_link_node_ref() under the proper lock, while we're at it. Thanks to Alexander Yurchenko and/or Damir Shayhutdinov for (almost) pinpointing the problem. Signed-off-by: David Woodhouse --- fs/jffs2/erase.c | 18 ++++++++---------- 1 file changed, 8 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/jffs2/erase.c b/fs/jffs2/erase.c index 96d9ad56e57..bdc6a7bec80 100644 --- a/fs/jffs2/erase.c +++ b/fs/jffs2/erase.c @@ -419,9 +419,6 @@ static void jffs2_mark_erased_block(struct jffs2_sb_info *c, struct jffs2_eraseb if (jffs2_write_nand_cleanmarker(c, jeb)) goto filebad; } - - /* Everything else got zeroed before the erase */ - jeb->free_size = c->sector_size; } else { struct kvec vecs[1]; @@ -449,18 +446,19 @@ static void jffs2_mark_erased_block(struct jffs2_sb_info *c, struct jffs2_eraseb goto filebad; } - - /* Everything else got zeroed before the erase */ - jeb->free_size = c->sector_size; - /* FIXME Special case for cleanmarker in empty block */ - jffs2_link_node_ref(c, jeb, jeb->offset | REF_NORMAL, c->cleanmarker_size, NULL); } + /* Everything else got zeroed before the erase */ + jeb->free_size = c->sector_size; mutex_lock(&c->erase_free_sem); spin_lock(&c->erase_completion_lock); + c->erasing_size -= c->sector_size; - c->free_size += jeb->free_size; - c->used_size += jeb->used_size; + c->free_size += c->sector_size; + + /* Account for cleanmarker now, if it's in-band */ + if (c->cleanmarker_size && !jffs2_cleanmarker_oob(c)) + jffs2_link_node_ref(c, jeb, jeb->offset | REF_NORMAL, c->cleanmarker_size, NULL); jffs2_dbg_acct_sanity_check_nolock(c,jeb); jffs2_dbg_acct_paranoia_check_nolock(c, jeb); -- cgit v1.2.3 From ed1524371716466e9c762808b02601d0d0276a92 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 22 Apr 2008 19:51:27 -0400 Subject: [PATCH] double-free of inode on alloc_file() failure exit in create_write_pipe() Duh... Fortunately, the bug is quite recent (post-2.6.25) and, embarrassingly, mine ;-/ Signed-off-by: Al Viro --- fs/pipe.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'fs') diff --git a/fs/pipe.c b/fs/pipe.c index 8be381bbcb5..f73492b6817 100644 --- a/fs/pipe.c +++ b/fs/pipe.c @@ -988,7 +988,10 @@ struct file *create_write_pipe(void) return f; err_dentry: + free_pipe_info(inode); dput(dentry); + return ERR_PTR(err); + err_inode: free_pipe_info(inode); iput(inode); -- cgit v1.2.3 From 9b4f526cdc0f95f635607dfba6ac788b3deca188 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 22 Apr 2008 01:32:44 -0400 Subject: [PATCH] proc_readfd_common() race fix Since we drop the rcu_read_lock inside the loop, we can't assume that files->fdt will remain unchanged (and not freed) between iterations. Signed-off-by: Al Viro --- fs/proc/base.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/proc/base.c b/fs/proc/base.c index 81d7d145292..7313c62e3e9 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -1626,7 +1626,6 @@ static int proc_readfd_common(struct file * filp, void * dirent, unsigned int fd, ino; int retval; struct files_struct * files; - struct fdtable *fdt; retval = -ENOENT; if (!p) @@ -1649,9 +1648,8 @@ static int proc_readfd_common(struct file * filp, void * dirent, if (!files) goto out; rcu_read_lock(); - fdt = files_fdtable(files); for (fd = filp->f_pos-2; - fd < fdt->max_fds; + fd < files_fdtable(files)->max_fds; fd++, filp->f_pos++) { char name[PROC_NUMBUF]; int len; -- cgit v1.2.3 From 85a62db6245a82f07a31b387915ee2180b9ea11a Mon Sep 17 00:00:00 2001 From: David Woodhouse Date: Wed, 23 Apr 2008 01:17:51 +0100 Subject: [JFFS2] Add paranoia debugging for superblock counts The problem fixed in commit 014b164e1392a166fe96e003d2f0e7ad2e2a0bb7 (space leak with in-band cleanmarkers) would have been caught a lot quicker if our paranoid debugging mode had included adding up the size counts from all the eraseblocks and comparing the totals with the counts in the superblock. Add that. Make jffs2_mark_erased_block() file the newly-erased block on the free_list before calling the debug function, to make it happy. Signed-off-by: David Woodhouse --- fs/jffs2/debug.c | 132 +++++++++++++++++++++++++++++++++++++++++++++++++++++++ fs/jffs2/erase.c | 7 +-- 2 files changed, 136 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/jffs2/debug.c b/fs/jffs2/debug.c index 660793107e9..590bdd6e014 100644 --- a/fs/jffs2/debug.c +++ b/fs/jffs2/debug.c @@ -153,6 +153,135 @@ __jffs2_dbg_prewrite_paranoia_check(struct jffs2_sb_info *c, kfree(buf); } +void __jffs2_dbg_superblock_counts(struct jffs2_sb_info *c) +{ + struct jffs2_eraseblock *jeb; + uint32_t free = 0, dirty = 0, used = 0, wasted = 0, + erasing = 0, bad = 0, unchecked = 0; + int nr_counted = 0; + int dump = 0; + + if (c->gcblock) { + nr_counted++; + free += c->gcblock->free_size; + dirty += c->gcblock->dirty_size; + used += c->gcblock->used_size; + wasted += c->gcblock->wasted_size; + unchecked += c->gcblock->unchecked_size; + } + if (c->nextblock) { + nr_counted++; + free += c->nextblock->free_size; + dirty += c->nextblock->dirty_size; + used += c->nextblock->used_size; + wasted += c->nextblock->wasted_size; + unchecked += c->nextblock->unchecked_size; + } + list_for_each_entry(jeb, &c->clean_list, list) { + nr_counted++; + free += jeb->free_size; + dirty += jeb->dirty_size; + used += jeb->used_size; + wasted += jeb->wasted_size; + unchecked += jeb->unchecked_size; + } + list_for_each_entry(jeb, &c->very_dirty_list, list) { + nr_counted++; + free += jeb->free_size; + dirty += jeb->dirty_size; + used += jeb->used_size; + wasted += jeb->wasted_size; + unchecked += jeb->unchecked_size; + } + list_for_each_entry(jeb, &c->dirty_list, list) { + nr_counted++; + free += jeb->free_size; + dirty += jeb->dirty_size; + used += jeb->used_size; + wasted += jeb->wasted_size; + unchecked += jeb->unchecked_size; + } + list_for_each_entry(jeb, &c->erasable_list, list) { + nr_counted++; + free += jeb->free_size; + dirty += jeb->dirty_size; + used += jeb->used_size; + wasted += jeb->wasted_size; + unchecked += jeb->unchecked_size; + } + list_for_each_entry(jeb, &c->erasable_pending_wbuf_list, list) { + nr_counted++; + free += jeb->free_size; + dirty += jeb->dirty_size; + used += jeb->used_size; + wasted += jeb->wasted_size; + unchecked += jeb->unchecked_size; + } + list_for_each_entry(jeb, &c->erase_pending_list, list) { + nr_counted++; + free += jeb->free_size; + dirty += jeb->dirty_size; + used += jeb->used_size; + wasted += jeb->wasted_size; + unchecked += jeb->unchecked_size; + } + list_for_each_entry(jeb, &c->free_list, list) { + nr_counted++; + free += jeb->free_size; + dirty += jeb->dirty_size; + used += jeb->used_size; + wasted += jeb->wasted_size; + unchecked += jeb->unchecked_size; + } + list_for_each_entry(jeb, &c->bad_used_list, list) { + nr_counted++; + free += jeb->free_size; + dirty += jeb->dirty_size; + used += jeb->used_size; + wasted += jeb->wasted_size; + unchecked += jeb->unchecked_size; + } + + list_for_each_entry(jeb, &c->erasing_list, list) { + nr_counted++; + erasing += c->sector_size; + } + list_for_each_entry(jeb, &c->erase_complete_list, list) { + nr_counted++; + erasing += c->sector_size; + } + list_for_each_entry(jeb, &c->bad_list, list) { + nr_counted++; + bad += c->sector_size; + } + +#define check(sz) \ + if (sz != c->sz##_size) { \ + printk(KERN_WARNING #sz "_size mismatch counted 0x%x, c->" #sz "_size 0x%x\n", \ + sz, c->sz##_size); \ + dump = 1; \ + } + check(free); + check(dirty); + check(used); + check(wasted); + check(unchecked); + check(bad); + check(erasing); +#undef check + + if (nr_counted != c->nr_blocks) { + printk(KERN_WARNING "%s counted only 0x%x blocks of 0x%x. Where are the others?\n", + __func__, nr_counted, c->nr_blocks); + dump = 1; + } + + if (dump) { + __jffs2_dbg_dump_block_lists_nolock(c); + BUG(); + } +} + /* * Check the space accounting and node_ref list correctness for the JFFS2 erasable block 'jeb'. */ @@ -229,6 +358,9 @@ __jffs2_dbg_acct_paranoia_check_nolock(struct jffs2_sb_info *c, } #endif + if (!(c->flags & (JFFS2_SB_FLAG_BUILDING|JFFS2_SB_FLAG_SCANNING))) + __jffs2_dbg_superblock_counts(c); + return; error: diff --git a/fs/jffs2/erase.c b/fs/jffs2/erase.c index bdc6a7bec80..65d91943fc2 100644 --- a/fs/jffs2/erase.c +++ b/fs/jffs2/erase.c @@ -460,12 +460,13 @@ static void jffs2_mark_erased_block(struct jffs2_sb_info *c, struct jffs2_eraseb if (c->cleanmarker_size && !jffs2_cleanmarker_oob(c)) jffs2_link_node_ref(c, jeb, jeb->offset | REF_NORMAL, c->cleanmarker_size, NULL); - jffs2_dbg_acct_sanity_check_nolock(c,jeb); - jffs2_dbg_acct_paranoia_check_nolock(c, jeb); - list_add_tail(&jeb->list, &c->free_list); c->nr_erasing_blocks--; c->nr_free_blocks++; + + jffs2_dbg_acct_sanity_check_nolock(c, jeb); + jffs2_dbg_acct_paranoia_check_nolock(c, jeb); + spin_unlock(&c->erase_completion_lock); mutex_unlock(&c->erase_free_sem); wake_up(&c->erase_wait); -- cgit v1.2.3 From 27e6b8e388fffb332476ddab00bbe05cd5da5f32 Mon Sep 17 00:00:00 2001 From: David Woodhouse Date: Wed, 23 Apr 2008 01:25:33 +0100 Subject: [JFFS2] Honour TEST_TOTLEN macro in debugging code. ref->__totlen is going! Signed-off-by: David Woodhouse --- fs/jffs2/debug.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/jffs2/debug.c b/fs/jffs2/debug.c index 590bdd6e014..e198468a8c6 100644 --- a/fs/jffs2/debug.c +++ b/fs/jffs2/debug.c @@ -400,7 +400,10 @@ __jffs2_dbg_dump_node_refs_nolock(struct jffs2_sb_info *c, printk(JFFS2_DBG); for (ref = jeb->first_node; ; ref = ref_next(ref)) { - printk("%#08x(%#x)", ref_offset(ref), ref->__totlen); + printk("%#08x", ref_offset(ref)); +#ifdef TEST_TOTLEN + printk("(%x)", ref->__totlen); +#endif if (ref_next(ref)) printk("->"); else -- cgit v1.2.3 From 19e56ceae7cb1833ffd806038c19477b2c265f9f Mon Sep 17 00:00:00 2001 From: David Woodhouse Date: Wed, 23 Apr 2008 01:26:12 +0100 Subject: [JFFS2] Finally remove redundant ref->__totlen field. Haven't had any complaints about it recently, despite having the test code enabled to verify that the calculated length is correct. Kill it off, just by #undef TEST_TOTLEN for now; removing it for real can come a little later. Signed-off-by: David Woodhouse --- fs/jffs2/nodelist.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/jffs2/nodelist.h b/fs/jffs2/nodelist.h index ec1aae9e695..8219df6eb6d 100644 --- a/fs/jffs2/nodelist.h +++ b/fs/jffs2/nodelist.h @@ -87,7 +87,7 @@ struct jffs2_raw_node_ref xattr_ref or xattr_datum instead. The common part of those structures has NULL in the first word. See jffs2_raw_ref_to_ic() below */ uint32_t flash_offset; -#define TEST_TOTLEN +#undef TEST_TOTLEN #ifdef TEST_TOTLEN uint32_t __totlen; /* This may die; use ref_totlen(c, jeb, ) below */ #endif -- cgit v1.2.3 From 934b25c597c0e98304a7eaec198a87e4633a42bb Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 23 Apr 2008 00:04:04 -0400 Subject: [PATCH] remove unused label in xattr.c (noise from ro-bind) Signed-off-by: Al Viro --- fs/xattr.c | 1 - 1 file changed, 1 deletion(-) (limited to 'fs') diff --git a/fs/xattr.c b/fs/xattr.c index f7062da505d..89a942f07e1 100644 --- a/fs/xattr.c +++ b/fs/xattr.c @@ -307,7 +307,6 @@ sys_fsetxattr(int fd, char __user *name, void __user *value, error = setxattr(dentry, name, value, size, flags); mnt_drop_write(f->f_path.mnt); } -out_fput: fput(f); return error; } -- cgit v1.2.3 From 6092d048183b76bfa3f84b32f8158dd8d10bd811 Mon Sep 17 00:00:00 2001 From: Ram Pai Date: Thu, 27 Mar 2008 13:06:20 +0100 Subject: [patch 1/7] vfs: mountinfo: add dentry_path() [mszeredi@suse.cz] split big patch into managable chunks Add the following functions: dentry_path() seq_dentry() These are similar to d_path() and seq_path(). But instead of calculating the path within a mount namespace, they calculate the path from the root of the filesystem to a given dentry, ignoring mounts completely. Signed-off-by: Ram Pai Signed-off-by: Miklos Szeredi Signed-off-by: Al Viro --- fs/dcache.c | 90 +++++++++++++++++++++++++++++++++++++++++++---------------- fs/seq_file.c | 65 +++++++++++++++++++++++++++++++----------- 2 files changed, 115 insertions(+), 40 deletions(-) (limited to 'fs') diff --git a/fs/dcache.c b/fs/dcache.c index 43455776711..635c2aa427e 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -1746,6 +1746,17 @@ shouldnt_be_hashed: goto shouldnt_be_hashed; } +static int prepend(char **buffer, int *buflen, const char *str, + int namelen) +{ + *buflen -= namelen; + if (*buflen < 0) + return -ENAMETOOLONG; + *buffer -= namelen; + memcpy(*buffer, str, namelen); + return 0; +} + /** * d_path - return the path of a dentry * @dentry: dentry to report @@ -1767,17 +1778,11 @@ static char *__d_path(struct dentry *dentry, struct vfsmount *vfsmnt, { char * end = buffer+buflen; char * retval; - int namelen; - - *--end = '\0'; - buflen--; - if (!IS_ROOT(dentry) && d_unhashed(dentry)) { - buflen -= 10; - end -= 10; - if (buflen < 0) + + prepend(&end, &buflen, "\0", 1); + if (!IS_ROOT(dentry) && d_unhashed(dentry) && + (prepend(&end, &buflen, " (deleted)", 10) != 0)) goto Elong; - memcpy(end, " (deleted)", 10); - } if (buflen < 1) goto Elong; @@ -1804,13 +1809,10 @@ static char *__d_path(struct dentry *dentry, struct vfsmount *vfsmnt, } parent = dentry->d_parent; prefetch(parent); - namelen = dentry->d_name.len; - buflen -= namelen + 1; - if (buflen < 0) + if ((prepend(&end, &buflen, dentry->d_name.name, + dentry->d_name.len) != 0) || + (prepend(&end, &buflen, "/", 1) != 0)) goto Elong; - end -= namelen; - memcpy(end, dentry->d_name.name, namelen); - *--end = '/'; retval = end; dentry = parent; } @@ -1818,12 +1820,10 @@ static char *__d_path(struct dentry *dentry, struct vfsmount *vfsmnt, return retval; global_root: - namelen = dentry->d_name.len; - buflen -= namelen; - if (buflen < 0) + retval += 1; /* hit the slash */ + if (prepend(&retval, &buflen, dentry->d_name.name, + dentry->d_name.len) != 0) goto Elong; - retval -= namelen-1; /* hit the slash */ - memcpy(retval, dentry->d_name.name, namelen); return retval; Elong: return ERR_PTR(-ENAMETOOLONG); @@ -1859,7 +1859,7 @@ char *d_path(struct path *path, char *buf, int buflen) read_lock(¤t->fs->lock); root = current->fs->root; - path_get(¤t->fs->root); + path_get(&root); read_unlock(¤t->fs->lock); spin_lock(&dcache_lock); res = __d_path(path->dentry, path->mnt, &root, buf, buflen); @@ -1889,6 +1889,48 @@ char *dynamic_dname(struct dentry *dentry, char *buffer, int buflen, return memcpy(buffer, temp, sz); } +/* + * Write full pathname from the root of the filesystem into the buffer. + */ +char *dentry_path(struct dentry *dentry, char *buf, int buflen) +{ + char *end = buf + buflen; + char *retval; + + spin_lock(&dcache_lock); + prepend(&end, &buflen, "\0", 1); + if (!IS_ROOT(dentry) && d_unhashed(dentry) && + (prepend(&end, &buflen, "//deleted", 9) != 0)) + goto Elong; + if (buflen < 1) + goto Elong; + /* Get '/' right */ + retval = end-1; + *retval = '/'; + + for (;;) { + struct dentry *parent; + if (IS_ROOT(dentry)) + break; + + parent = dentry->d_parent; + prefetch(parent); + + if ((prepend(&end, &buflen, dentry->d_name.name, + dentry->d_name.len) != 0) || + (prepend(&end, &buflen, "/", 1) != 0)) + goto Elong; + + retval = end; + dentry = parent; + } + spin_unlock(&dcache_lock); + return retval; +Elong: + spin_unlock(&dcache_lock); + return ERR_PTR(-ENAMETOOLONG); +} + /* * NOTE! The user-level library version returns a * character pointer. The kernel system call just @@ -1918,9 +1960,9 @@ asmlinkage long sys_getcwd(char __user *buf, unsigned long size) read_lock(¤t->fs->lock); pwd = current->fs->pwd; - path_get(¤t->fs->pwd); + path_get(&pwd); root = current->fs->root; - path_get(¤t->fs->root); + path_get(&root); read_unlock(¤t->fs->lock); error = -ENOENT; diff --git a/fs/seq_file.c b/fs/seq_file.c index cdfd996ca6e..ae59f5a6c5c 100644 --- a/fs/seq_file.c +++ b/fs/seq_file.c @@ -350,28 +350,40 @@ int seq_printf(struct seq_file *m, const char *f, ...) } EXPORT_SYMBOL(seq_printf); +static char *mangle_path(char *s, char *p, char *esc) +{ + while (s <= p) { + char c = *p++; + if (!c) { + return s; + } else if (!strchr(esc, c)) { + *s++ = c; + } else if (s + 4 > p) { + break; + } else { + *s++ = '\\'; + *s++ = '0' + ((c & 0300) >> 6); + *s++ = '0' + ((c & 070) >> 3); + *s++ = '0' + (c & 07); + } + } + return NULL; +} + +/* + * return the absolute path of 'dentry' residing in mount 'mnt'. + */ int seq_path(struct seq_file *m, struct path *path, char *esc) { if (m->count < m->size) { char *s = m->buf + m->count; char *p = d_path(path, s, m->size - m->count); if (!IS_ERR(p)) { - while (s <= p) { - char c = *p++; - if (!c) { - p = m->buf + m->count; - m->count = s - m->buf; - return s - p; - } else if (!strchr(esc, c)) { - *s++ = c; - } else if (s + 4 > p) { - break; - } else { - *s++ = '\\'; - *s++ = '0' + ((c & 0300) >> 6); - *s++ = '0' + ((c & 070) >> 3); - *s++ = '0' + (c & 07); - } + s = mangle_path(s, p, esc); + if (s) { + p = m->buf + m->count; + m->count = s - m->buf; + return s - p; } } } @@ -380,6 +392,27 @@ int seq_path(struct seq_file *m, struct path *path, char *esc) } EXPORT_SYMBOL(seq_path); +/* + * returns the path of the 'dentry' from the root of its filesystem. + */ +int seq_dentry(struct seq_file *m, struct dentry *dentry, char *esc) +{ + if (m->count < m->size) { + char *s = m->buf + m->count; + char *p = dentry_path(dentry, s, m->size - m->count); + if (!IS_ERR(p)) { + s = mangle_path(s, p, esc); + if (s) { + p = m->buf + m->count; + m->count = s - m->buf; + return s - p; + } + } + } + m->count = m->size; + return -1; +} + static void *single_start(struct seq_file *p, loff_t *pos) { return NULL + (*pos == 0); -- cgit v1.2.3 From 9d1bc60138977d9c79471b344a64f2df13b2ccef Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Thu, 27 Mar 2008 13:06:21 +0100 Subject: [patch 2/7] vfs: mountinfo: add seq_file_root() Add a new function: seq_file_root() This is similar to seq_path(), but calculates the path relative to the given root, instead of current->fs->root. If the path was unreachable from root, then modify the root parameter to reflect this. Signed-off-by: Miklos Szeredi Signed-off-by: Al Viro --- fs/dcache.c | 24 ++++++++++++++++-------- fs/seq_file.c | 30 ++++++++++++++++++++++++++++++ 2 files changed, 46 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/dcache.c b/fs/dcache.c index 635c2aa427e..3ee588d5f58 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -1759,10 +1759,8 @@ static int prepend(char **buffer, int *buflen, const char *str, /** * d_path - return the path of a dentry - * @dentry: dentry to report - * @vfsmnt: vfsmnt to which the dentry belongs - * @root: root dentry - * @rootmnt: vfsmnt to which the root dentry belongs + * @path: the dentry/vfsmount to report + * @root: root vfsmnt/dentry (may be modified by this function) * @buffer: buffer to return value in * @buflen: buffer length * @@ -1772,10 +1770,15 @@ static int prepend(char **buffer, int *buflen, const char *str, * Returns the buffer or an error code if the path was too long. * * "buflen" should be positive. Caller holds the dcache_lock. + * + * If path is not reachable from the supplied root, then the value of + * root is changed (without modifying refcounts). */ -static char *__d_path(struct dentry *dentry, struct vfsmount *vfsmnt, - struct path *root, char *buffer, int buflen) +char *__d_path(const struct path *path, struct path *root, + char *buffer, int buflen) { + struct dentry *dentry = path->dentry; + struct vfsmount *vfsmnt = path->mnt; char * end = buffer+buflen; char * retval; @@ -1824,6 +1827,8 @@ global_root: if (prepend(&retval, &buflen, dentry->d_name.name, dentry->d_name.len) != 0) goto Elong; + root->mnt = vfsmnt; + root->dentry = dentry; return retval; Elong: return ERR_PTR(-ENAMETOOLONG); @@ -1846,6 +1851,7 @@ char *d_path(struct path *path, char *buf, int buflen) { char *res; struct path root; + struct path tmp; /* * We have various synthetic filesystems that never get mounted. On @@ -1862,7 +1868,8 @@ char *d_path(struct path *path, char *buf, int buflen) path_get(&root); read_unlock(¤t->fs->lock); spin_lock(&dcache_lock); - res = __d_path(path->dentry, path->mnt, &root, buf, buflen); + tmp = root; + res = __d_path(path, &tmp, buf, buflen); spin_unlock(&dcache_lock); path_put(&root); return res; @@ -1970,9 +1977,10 @@ asmlinkage long sys_getcwd(char __user *buf, unsigned long size) spin_lock(&dcache_lock); if (pwd.dentry->d_parent == pwd.dentry || !d_unhashed(pwd.dentry)) { unsigned long len; + struct path tmp = root; char * cwd; - cwd = __d_path(pwd.dentry, pwd.mnt, &root, page, PAGE_SIZE); + cwd = __d_path(&pwd, &tmp, page, PAGE_SIZE); spin_unlock(&dcache_lock); error = PTR_ERR(cwd); diff --git a/fs/seq_file.c b/fs/seq_file.c index ae59f5a6c5c..3f54dbd6c49 100644 --- a/fs/seq_file.c +++ b/fs/seq_file.c @@ -392,6 +392,36 @@ int seq_path(struct seq_file *m, struct path *path, char *esc) } EXPORT_SYMBOL(seq_path); +/* + * Same as seq_path, but relative to supplied root. + * + * root may be changed, see __d_path(). + */ +int seq_path_root(struct seq_file *m, struct path *path, struct path *root, + char *esc) +{ + int err = -ENAMETOOLONG; + if (m->count < m->size) { + char *s = m->buf + m->count; + char *p; + + spin_lock(&dcache_lock); + p = __d_path(path, root, s, m->size - m->count); + spin_unlock(&dcache_lock); + err = PTR_ERR(p); + if (!IS_ERR(p)) { + s = mangle_path(s, p, esc); + if (s) { + p = m->buf + m->count; + m->count = s - m->buf; + return 0; + } + } + } + m->count = m->size; + return err; +} + /* * returns the path of the 'dentry' from the root of its filesystem. */ -- cgit v1.2.3 From 73cd49ecdde92fdce131938bdaff4993010d181b Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Wed, 26 Mar 2008 22:11:34 +0100 Subject: [patch 3/7] vfs: mountinfo: add mount ID Add a unique ID to each vfsmount using the IDR infrastructure. The identifiers are reused after the vfsmount is freed. Signed-off-by: Miklos Szeredi Signed-off-by: Al Viro --- fs/namespace.c | 34 ++++++++++++++++++++++++++++++++++ 1 file changed, 34 insertions(+) (limited to 'fs') diff --git a/fs/namespace.c b/fs/namespace.c index 1bf302d0478..8ca6317cb40 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -27,6 +27,7 @@ #include #include #include +#include #include #include #include "pnode.h" @@ -39,6 +40,7 @@ __cacheline_aligned_in_smp DEFINE_SPINLOCK(vfsmount_lock); static int event; +static DEFINE_IDA(mnt_id_ida); static struct list_head *mount_hashtable __read_mostly; static struct kmem_cache *mnt_cache __read_mostly; @@ -58,10 +60,41 @@ static inline unsigned long hash(struct vfsmount *mnt, struct dentry *dentry) #define MNT_WRITER_UNDERFLOW_LIMIT -(1<<16) +/* allocation is serialized by namespace_sem */ +static int mnt_alloc_id(struct vfsmount *mnt) +{ + int res; + +retry: + ida_pre_get(&mnt_id_ida, GFP_KERNEL); + spin_lock(&vfsmount_lock); + res = ida_get_new(&mnt_id_ida, &mnt->mnt_id); + spin_unlock(&vfsmount_lock); + if (res == -EAGAIN) + goto retry; + + return res; +} + +static void mnt_free_id(struct vfsmount *mnt) +{ + spin_lock(&vfsmount_lock); + ida_remove(&mnt_id_ida, mnt->mnt_id); + spin_unlock(&vfsmount_lock); +} + struct vfsmount *alloc_vfsmnt(const char *name) { struct vfsmount *mnt = kmem_cache_zalloc(mnt_cache, GFP_KERNEL); if (mnt) { + int err; + + err = mnt_alloc_id(mnt); + if (err) { + kmem_cache_free(mnt_cache, mnt); + return NULL; + } + atomic_set(&mnt->mnt_count, 1); INIT_LIST_HEAD(&mnt->mnt_hash); INIT_LIST_HEAD(&mnt->mnt_child); @@ -353,6 +386,7 @@ EXPORT_SYMBOL(simple_set_mnt); void free_vfsmnt(struct vfsmount *mnt) { kfree(mnt->mnt_devname); + mnt_free_id(mnt); kmem_cache_free(mnt_cache, mnt); } -- cgit v1.2.3 From 719f5d7f0b90ac2c8f8ca4232eb322b266fea01e Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Thu, 27 Mar 2008 13:06:23 +0100 Subject: [patch 4/7] vfs: mountinfo: add mount peer group ID Add a unique ID to each peer group using the IDR infrastructure. The identifiers are reused after the peer group dissolves. The IDR structures are protected by holding namepspace_sem for write while allocating or deallocating IDs. IDs are allocated when a previously unshared vfsmount becomes the first member of a peer group. When a new member is added to an existing group, the ID is copied from one of the old members. IDs are freed when the last member of a peer group is unshared. Setting the MNT_SHARED flag on members of a subtree is done as a separate step, after all the IDs have been allocated. This way an allocation failure can be cleaned up easilty, without affecting the propagation state. Based on design sketch by Al Viro. Signed-off-by: Miklos Szeredi Signed-off-by: Al Viro --- fs/namespace.c | 93 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++-- fs/pnode.c | 5 +++- 2 files changed, 94 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/namespace.c b/fs/namespace.c index 8ca6317cb40..cefa1d9939b 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -41,6 +41,7 @@ __cacheline_aligned_in_smp DEFINE_SPINLOCK(vfsmount_lock); static int event; static DEFINE_IDA(mnt_id_ida); +static DEFINE_IDA(mnt_group_ida); static struct list_head *mount_hashtable __read_mostly; static struct kmem_cache *mnt_cache __read_mostly; @@ -83,6 +84,28 @@ static void mnt_free_id(struct vfsmount *mnt) spin_unlock(&vfsmount_lock); } +/* + * Allocate a new peer group ID + * + * mnt_group_ida is protected by namespace_sem + */ +static int mnt_alloc_group_id(struct vfsmount *mnt) +{ + if (!ida_pre_get(&mnt_group_ida, GFP_KERNEL)) + return -ENOMEM; + + return ida_get_new_above(&mnt_group_ida, 1, &mnt->mnt_group_id); +} + +/* + * Release a peer group ID + */ +void mnt_release_group_id(struct vfsmount *mnt) +{ + ida_remove(&mnt_group_ida, mnt->mnt_group_id); + mnt->mnt_group_id = 0; +} + struct vfsmount *alloc_vfsmnt(const char *name) { struct vfsmount *mnt = kmem_cache_zalloc(mnt_cache, GFP_KERNEL); @@ -533,6 +556,17 @@ static struct vfsmount *clone_mnt(struct vfsmount *old, struct dentry *root, struct vfsmount *mnt = alloc_vfsmnt(old->mnt_devname); if (mnt) { + if (flag & (CL_SLAVE | CL_PRIVATE)) + mnt->mnt_group_id = 0; /* not a peer of original */ + else + mnt->mnt_group_id = old->mnt_group_id; + + if ((flag & CL_MAKE_SHARED) && !mnt->mnt_group_id) { + int err = mnt_alloc_group_id(mnt); + if (err) + goto out_free; + } + mnt->mnt_flags = old->mnt_flags; atomic_inc(&sb->s_active); mnt->mnt_sb = sb; @@ -562,6 +596,10 @@ static struct vfsmount *clone_mnt(struct vfsmount *old, struct dentry *root, } } return mnt; + + out_free: + free_vfsmnt(mnt); + return NULL; } static inline void __mntput(struct vfsmount *mnt) @@ -1142,6 +1180,33 @@ void drop_collected_mounts(struct vfsmount *mnt) release_mounts(&umount_list); } +static void cleanup_group_ids(struct vfsmount *mnt, struct vfsmount *end) +{ + struct vfsmount *p; + + for (p = mnt; p != end; p = next_mnt(p, mnt)) { + if (p->mnt_group_id && !IS_MNT_SHARED(p)) + mnt_release_group_id(p); + } +} + +static int invent_group_ids(struct vfsmount *mnt, bool recurse) +{ + struct vfsmount *p; + + for (p = mnt; p; p = recurse ? next_mnt(p, mnt) : NULL) { + if (!p->mnt_group_id && !IS_MNT_SHARED(p)) { + int err = mnt_alloc_group_id(p); + if (err) { + cleanup_group_ids(mnt, p); + return err; + } + } + } + + return 0; +} + /* * @source_mnt : mount tree to be attached * @nd : place the mount tree @source_mnt is attached @@ -1212,9 +1277,16 @@ static int attach_recursive_mnt(struct vfsmount *source_mnt, struct vfsmount *dest_mnt = path->mnt; struct dentry *dest_dentry = path->dentry; struct vfsmount *child, *p; + int err; - if (propagate_mnt(dest_mnt, dest_dentry, source_mnt, &tree_list)) - return -EINVAL; + if (IS_MNT_SHARED(dest_mnt)) { + err = invent_group_ids(source_mnt, true); + if (err) + goto out; + } + err = propagate_mnt(dest_mnt, dest_dentry, source_mnt, &tree_list); + if (err) + goto out_cleanup_ids; if (IS_MNT_SHARED(dest_mnt)) { for (p = source_mnt; p; p = next_mnt(p, source_mnt)) @@ -1237,6 +1309,12 @@ static int attach_recursive_mnt(struct vfsmount *source_mnt, } spin_unlock(&vfsmount_lock); return 0; + + out_cleanup_ids: + if (IS_MNT_SHARED(dest_mnt)) + cleanup_group_ids(source_mnt, NULL); + out: + return err; } static int graft_tree(struct vfsmount *mnt, struct path *path) @@ -1277,6 +1355,7 @@ static noinline int do_change_type(struct nameidata *nd, int flag) struct vfsmount *m, *mnt = nd->path.mnt; int recurse = flag & MS_REC; int type = flag & ~MS_REC; + int err = 0; if (!capable(CAP_SYS_ADMIN)) return -EPERM; @@ -1285,12 +1364,20 @@ static noinline int do_change_type(struct nameidata *nd, int flag) return -EINVAL; down_write(&namespace_sem); + if (type == MS_SHARED) { + err = invent_group_ids(mnt, recurse); + if (err) + goto out_unlock; + } + spin_lock(&vfsmount_lock); for (m = mnt; m; m = (recurse ? next_mnt(m, mnt) : NULL)) change_mnt_propagation(m, type); spin_unlock(&vfsmount_lock); + + out_unlock: up_write(&namespace_sem); - return 0; + return err; } /* diff --git a/fs/pnode.c b/fs/pnode.c index f968e35d978..d18d66491a0 100644 --- a/fs/pnode.c +++ b/fs/pnode.c @@ -46,7 +46,11 @@ static int do_make_slave(struct vfsmount *mnt) if (peer_mnt == mnt) peer_mnt = NULL; } + if (IS_MNT_SHARED(mnt) && list_empty(&mnt->mnt_share)) + mnt_release_group_id(mnt); + list_del_init(&mnt->mnt_share); + mnt->mnt_group_id = 0; if (peer_mnt) master = peer_mnt; @@ -68,7 +72,6 @@ static int do_make_slave(struct vfsmount *mnt) } mnt->mnt_master = master; CLEAR_MNT_SHARED(mnt); - INIT_LIST_HEAD(&mnt->mnt_slave_list); return 0; } -- cgit v1.2.3 From a1a2c409b666befc58c2db9c7fbddf200f153470 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Thu, 27 Mar 2008 13:06:24 +0100 Subject: [patch 5/7] vfs: mountinfo: allow using process root Allow /proc//mountinfo to use the root of to calculate mountpoints. - move definition of 'struct proc_mounts' to - add the process's namespace and root to this structure - pass a pointer to 'struct proc_mounts' into seq_operations In addition the following cleanups are made: - use a common open function for /proc//{mounts,mountstat} - surround namespace.c part of these proc files with #ifdef CONFIG_PROC_FS - make the seq_operations structures const Signed-off-by: Miklos Szeredi Signed-off-by: Al Viro --- fs/namespace.c | 14 ++++---- fs/proc/base.c | 108 +++++++++++++++++++++++++++------------------------------ 2 files changed, 59 insertions(+), 63 deletions(-) (limited to 'fs') diff --git a/fs/namespace.c b/fs/namespace.c index cefa1d9939b..dfdf51e81c1 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -724,20 +724,21 @@ void save_mount_options(struct super_block *sb, char *options) } EXPORT_SYMBOL(save_mount_options); +#ifdef CONFIG_PROC_FS /* iterator */ static void *m_start(struct seq_file *m, loff_t *pos) { - struct mnt_namespace *n = m->private; + struct proc_mounts *p = m->private; down_read(&namespace_sem); - return seq_list_start(&n->list, *pos); + return seq_list_start(&p->ns->list, *pos); } static void *m_next(struct seq_file *m, void *v, loff_t *pos) { - struct mnt_namespace *n = m->private; + struct proc_mounts *p = m->private; - return seq_list_next(v, &n->list, pos); + return seq_list_next(v, &p->ns->list, pos); } static void m_stop(struct seq_file *m, void *v) @@ -794,7 +795,7 @@ static int show_vfsmnt(struct seq_file *m, void *v) return err; } -struct seq_operations mounts_op = { +const struct seq_operations mounts_op = { .start = m_start, .next = m_next, .stop = m_stop, @@ -833,12 +834,13 @@ static int show_vfsstat(struct seq_file *m, void *v) return err; } -struct seq_operations mountstats_op = { +const struct seq_operations mountstats_op = { .start = m_start, .next = m_next, .stop = m_stop, .show = show_vfsstat, }; +#endif /* CONFIG_PROC_FS */ /** * may_umount_tree - check if a mount tree is busy diff --git a/fs/proc/base.c b/fs/proc/base.c index 7313c62e3e9..a04b3db7a29 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -502,17 +502,14 @@ static const struct inode_operations proc_def_inode_operations = { .setattr = proc_setattr, }; -extern const struct seq_operations mounts_op; -struct proc_mounts { - struct seq_file m; - int event; -}; - -static int mounts_open(struct inode *inode, struct file *file) +static int mounts_open_common(struct inode *inode, struct file *file, + const struct seq_operations *op) { struct task_struct *task = get_proc_task(inode); struct nsproxy *nsp; struct mnt_namespace *ns = NULL; + struct fs_struct *fs = NULL; + struct path root; struct proc_mounts *p; int ret = -EINVAL; @@ -525,40 +522,61 @@ static int mounts_open(struct inode *inode, struct file *file) get_mnt_ns(ns); } rcu_read_unlock(); - + if (ns) + fs = get_fs_struct(task); put_task_struct(task); } - if (ns) { - ret = -ENOMEM; - p = kmalloc(sizeof(struct proc_mounts), GFP_KERNEL); - if (p) { - file->private_data = &p->m; - ret = seq_open(file, &mounts_op); - if (!ret) { - p->m.private = ns; - p->event = ns->event; - return 0; - } - kfree(p); - } - put_mnt_ns(ns); - } + if (!ns) + goto err; + if (!fs) + goto err_put_ns; + + read_lock(&fs->lock); + root = fs->root; + path_get(&root); + read_unlock(&fs->lock); + put_fs_struct(fs); + + ret = -ENOMEM; + p = kmalloc(sizeof(struct proc_mounts), GFP_KERNEL); + if (!p) + goto err_put_path; + + file->private_data = &p->m; + ret = seq_open(file, op); + if (ret) + goto err_free; + + p->m.private = p; + p->ns = ns; + p->root = root; + p->event = ns->event; + + return 0; + + err_free: + kfree(p); + err_put_path: + path_put(&root); + err_put_ns: + put_mnt_ns(ns); + err: return ret; } static int mounts_release(struct inode *inode, struct file *file) { - struct seq_file *m = file->private_data; - struct mnt_namespace *ns = m->private; - put_mnt_ns(ns); + struct proc_mounts *p = file->private_data; + path_put(&p->root); + put_mnt_ns(p->ns); return seq_release(inode, file); } static unsigned mounts_poll(struct file *file, poll_table *wait) { struct proc_mounts *p = file->private_data; - struct mnt_namespace *ns = p->m.private; + struct mnt_namespace *ns = p->ns; unsigned res = 0; poll_wait(file, &ns->poll, wait); @@ -573,6 +591,11 @@ static unsigned mounts_poll(struct file *file, poll_table *wait) return res; } +static int mounts_open(struct inode *inode, struct file *file) +{ + return mounts_open_common(inode, file, &mounts_op); +} + static const struct file_operations proc_mounts_operations = { .open = mounts_open, .read = seq_read, @@ -581,38 +604,9 @@ static const struct file_operations proc_mounts_operations = { .poll = mounts_poll, }; -extern const struct seq_operations mountstats_op; static int mountstats_open(struct inode *inode, struct file *file) { - int ret = seq_open(file, &mountstats_op); - - if (!ret) { - struct seq_file *m = file->private_data; - struct nsproxy *nsp; - struct mnt_namespace *mnt_ns = NULL; - struct task_struct *task = get_proc_task(inode); - - if (task) { - rcu_read_lock(); - nsp = task_nsproxy(task); - if (nsp) { - mnt_ns = nsp->mnt_ns; - if (mnt_ns) - get_mnt_ns(mnt_ns); - } - rcu_read_unlock(); - - put_task_struct(task); - } - - if (mnt_ns) - m->private = mnt_ns; - else { - seq_release(inode, file); - ret = -EINVAL; - } - } - return ret; + return mounts_open_common(inode, file, &mountstats_op); } static const struct file_operations proc_mountstats_operations = { -- cgit v1.2.3 From 2d4d4864ac08caff5c204a752bd004eed4f08760 Mon Sep 17 00:00:00 2001 From: Ram Pai Date: Thu, 27 Mar 2008 13:06:25 +0100 Subject: [patch 6/7] vfs: mountinfo: add /proc//mountinfo [mszeredi@suse.cz] rewrite and split big patch into managable chunks /proc/mounts in its current form lacks important information: - propagation state - root of mount for bind mounts - the st_dev value used within the filesystem - identifier for each mount and it's parent It also suffers from the following problems: - not easily extendable - ambiguity of mountpoints within a chrooted environment - doesn't distinguish between filesystem dependent and independent options - doesn't distinguish between per mount and per super block options This patch introduces /proc//mountinfo which attempts to address all these deficiencies. Code shared between /proc//mounts and /proc//mountinfo is extracted into separate functions. Thanks to Al Viro for the help in getting the design right. Signed-off-by: Ram Pai Signed-off-by: Miklos Szeredi Signed-off-by: Al Viro --- fs/namespace.c | 119 ++++++++++++++++++++++++++++++++++++++++++++++----------- fs/proc/base.c | 15 ++++++++ 2 files changed, 111 insertions(+), 23 deletions(-) (limited to 'fs') diff --git a/fs/namespace.c b/fs/namespace.c index dfdf51e81c1..c807b8d5f89 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -746,20 +746,30 @@ static void m_stop(struct seq_file *m, void *v) up_read(&namespace_sem); } -static int show_vfsmnt(struct seq_file *m, void *v) +struct proc_fs_info { + int flag; + const char *str; +}; + +static void show_sb_opts(struct seq_file *m, struct super_block *sb) { - struct vfsmount *mnt = list_entry(v, struct vfsmount, mnt_list); - int err = 0; - static struct proc_fs_info { - int flag; - char *str; - } fs_info[] = { + static const struct proc_fs_info fs_info[] = { { MS_SYNCHRONOUS, ",sync" }, { MS_DIRSYNC, ",dirsync" }, { MS_MANDLOCK, ",mand" }, { 0, NULL } }; - static struct proc_fs_info mnt_info[] = { + const struct proc_fs_info *fs_infop; + + for (fs_infop = fs_info; fs_infop->flag; fs_infop++) { + if (sb->s_flags & fs_infop->flag) + seq_puts(m, fs_infop->str); + } +} + +static void show_mnt_opts(struct seq_file *m, struct vfsmount *mnt) +{ + static const struct proc_fs_info mnt_info[] = { { MNT_NOSUID, ",nosuid" }, { MNT_NODEV, ",nodev" }, { MNT_NOEXEC, ",noexec" }, @@ -768,27 +778,37 @@ static int show_vfsmnt(struct seq_file *m, void *v) { MNT_RELATIME, ",relatime" }, { 0, NULL } }; - struct proc_fs_info *fs_infop; + const struct proc_fs_info *fs_infop; + + for (fs_infop = mnt_info; fs_infop->flag; fs_infop++) { + if (mnt->mnt_flags & fs_infop->flag) + seq_puts(m, fs_infop->str); + } +} + +static void show_type(struct seq_file *m, struct super_block *sb) +{ + mangle(m, sb->s_type->name); + if (sb->s_subtype && sb->s_subtype[0]) { + seq_putc(m, '.'); + mangle(m, sb->s_subtype); + } +} + +static int show_vfsmnt(struct seq_file *m, void *v) +{ + struct vfsmount *mnt = list_entry(v, struct vfsmount, mnt_list); + int err = 0; struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt }; mangle(m, mnt->mnt_devname ? mnt->mnt_devname : "none"); seq_putc(m, ' '); seq_path(m, &mnt_path, " \t\n\\"); seq_putc(m, ' '); - mangle(m, mnt->mnt_sb->s_type->name); - if (mnt->mnt_sb->s_subtype && mnt->mnt_sb->s_subtype[0]) { - seq_putc(m, '.'); - mangle(m, mnt->mnt_sb->s_subtype); - } + show_type(m, mnt->mnt_sb); seq_puts(m, __mnt_is_readonly(mnt) ? " ro" : " rw"); - for (fs_infop = fs_info; fs_infop->flag; fs_infop++) { - if (mnt->mnt_sb->s_flags & fs_infop->flag) - seq_puts(m, fs_infop->str); - } - for (fs_infop = mnt_info; fs_infop->flag; fs_infop++) { - if (mnt->mnt_flags & fs_infop->flag) - seq_puts(m, fs_infop->str); - } + show_sb_opts(m, mnt->mnt_sb); + show_mnt_opts(m, mnt); if (mnt->mnt_sb->s_op->show_options) err = mnt->mnt_sb->s_op->show_options(m, mnt); seq_puts(m, " 0 0\n"); @@ -802,6 +822,59 @@ const struct seq_operations mounts_op = { .show = show_vfsmnt }; +static int show_mountinfo(struct seq_file *m, void *v) +{ + struct proc_mounts *p = m->private; + struct vfsmount *mnt = list_entry(v, struct vfsmount, mnt_list); + struct super_block *sb = mnt->mnt_sb; + struct path mnt_path = { .dentry = mnt->mnt_root, .mnt = mnt }; + struct path root = p->root; + int err = 0; + + seq_printf(m, "%i %i %u:%u ", mnt->mnt_id, mnt->mnt_parent->mnt_id, + MAJOR(sb->s_dev), MINOR(sb->s_dev)); + seq_dentry(m, mnt->mnt_root, " \t\n\\"); + seq_putc(m, ' '); + seq_path_root(m, &mnt_path, &root, " \t\n\\"); + if (root.mnt != p->root.mnt || root.dentry != p->root.dentry) { + /* + * Mountpoint is outside root, discard that one. Ugly, + * but less so than trying to do that in iterator in a + * race-free way (due to renames). + */ + return SEQ_SKIP; + } + seq_puts(m, mnt->mnt_flags & MNT_READONLY ? " ro" : " rw"); + show_mnt_opts(m, mnt); + + /* Tagged fields ("foo:X" or "bar") */ + if (IS_MNT_SHARED(mnt)) + seq_printf(m, " shared:%i", mnt->mnt_group_id); + if (IS_MNT_SLAVE(mnt)) + seq_printf(m, " master:%i", mnt->mnt_master->mnt_group_id); + if (IS_MNT_UNBINDABLE(mnt)) + seq_puts(m, " unbindable"); + + /* Filesystem specific data */ + seq_puts(m, " - "); + show_type(m, sb); + seq_putc(m, ' '); + mangle(m, mnt->mnt_devname ? mnt->mnt_devname : "none"); + seq_puts(m, sb->s_flags & MS_RDONLY ? " ro" : " rw"); + show_sb_opts(m, sb); + if (sb->s_op->show_options) + err = sb->s_op->show_options(m, mnt); + seq_putc(m, '\n'); + return err; +} + +const struct seq_operations mountinfo_op = { + .start = m_start, + .next = m_next, + .stop = m_stop, + .show = show_mountinfo, +}; + static int show_vfsstat(struct seq_file *m, void *v) { struct vfsmount *mnt = list_entry(v, struct vfsmount, mnt_list); @@ -822,7 +895,7 @@ static int show_vfsstat(struct seq_file *m, void *v) /* file system type */ seq_puts(m, "with fstype "); - mangle(m, mnt->mnt_sb->s_type->name); + show_type(m, mnt->mnt_sb); /* optional statistics */ if (mnt->mnt_sb->s_op->show_stats) { diff --git a/fs/proc/base.c b/fs/proc/base.c index a04b3db7a29..c5e412a00b1 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -604,6 +604,19 @@ static const struct file_operations proc_mounts_operations = { .poll = mounts_poll, }; +static int mountinfo_open(struct inode *inode, struct file *file) +{ + return mounts_open_common(inode, file, &mountinfo_op); +} + +static const struct file_operations proc_mountinfo_operations = { + .open = mountinfo_open, + .read = seq_read, + .llseek = seq_lseek, + .release = mounts_release, + .poll = mounts_poll, +}; + static int mountstats_open(struct inode *inode, struct file *file) { return mounts_open_common(inode, file, &mountstats_op); @@ -2303,6 +2316,7 @@ static const struct pid_entry tgid_base_stuff[] = { LNK("root", root), LNK("exe", exe), REG("mounts", S_IRUGO, mounts), + REG("mountinfo", S_IRUGO, mountinfo), REG("mountstats", S_IRUSR, mountstats), #ifdef CONFIG_PROC_PAGE_MONITOR REG("clear_refs", S_IWUSR, clear_refs), @@ -2635,6 +2649,7 @@ static const struct pid_entry tid_base_stuff[] = { LNK("root", root), LNK("exe", exe), REG("mounts", S_IRUGO, mounts), + REG("mountinfo", S_IRUGO, mountinfo), #ifdef CONFIG_PROC_PAGE_MONITOR REG("clear_refs", S_IWUSR, clear_refs), REG("smaps", S_IRUGO, smaps), -- cgit v1.2.3 From 97e7e0f71d6d948c25f11f0a33878d9356d9579e Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Thu, 27 Mar 2008 13:06:26 +0100 Subject: [patch 7/7] vfs: mountinfo: show dominating group id Show peer group ID of nearest dominating group that has intersection with the mount's namespace. Signed-off-by: Miklos Szeredi Signed-off-by: Al Viro --- fs/namespace.c | 9 +++++++-- fs/pnode.c | 51 +++++++++++++++++++++++++++++++++++++++++++++++++++ fs/pnode.h | 1 + 3 files changed, 59 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/namespace.c b/fs/namespace.c index c807b8d5f89..0505fb61aa7 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -850,8 +850,13 @@ static int show_mountinfo(struct seq_file *m, void *v) /* Tagged fields ("foo:X" or "bar") */ if (IS_MNT_SHARED(mnt)) seq_printf(m, " shared:%i", mnt->mnt_group_id); - if (IS_MNT_SLAVE(mnt)) - seq_printf(m, " master:%i", mnt->mnt_master->mnt_group_id); + if (IS_MNT_SLAVE(mnt)) { + int master = mnt->mnt_master->mnt_group_id; + int dom = get_dominating_id(mnt, &p->root); + seq_printf(m, " master:%i", master); + if (dom && dom != master) + seq_printf(m, " propagate_from:%i", dom); + } if (IS_MNT_UNBINDABLE(mnt)) seq_puts(m, " unbindable"); diff --git a/fs/pnode.c b/fs/pnode.c index d18d66491a0..8d5f392ec3d 100644 --- a/fs/pnode.c +++ b/fs/pnode.c @@ -28,6 +28,57 @@ static inline struct vfsmount *next_slave(struct vfsmount *p) return list_entry(p->mnt_slave.next, struct vfsmount, mnt_slave); } +/* + * Return true if path is reachable from root + * + * namespace_sem is held, and mnt is attached + */ +static bool is_path_reachable(struct vfsmount *mnt, struct dentry *dentry, + const struct path *root) +{ + while (mnt != root->mnt && mnt->mnt_parent != mnt) { + dentry = mnt->mnt_mountpoint; + mnt = mnt->mnt_parent; + } + return mnt == root->mnt && is_subdir(dentry, root->dentry); +} + +static struct vfsmount *get_peer_under_root(struct vfsmount *mnt, + struct mnt_namespace *ns, + const struct path *root) +{ + struct vfsmount *m = mnt; + + do { + /* Check the namespace first for optimization */ + if (m->mnt_ns == ns && is_path_reachable(m, m->mnt_root, root)) + return m; + + m = next_peer(m); + } while (m != mnt); + + return NULL; +} + +/* + * Get ID of closest dominating peer group having a representative + * under the given root. + * + * Caller must hold namespace_sem + */ +int get_dominating_id(struct vfsmount *mnt, const struct path *root) +{ + struct vfsmount *m; + + for (m = mnt->mnt_master; m != NULL; m = m->mnt_master) { + struct vfsmount *d = get_peer_under_root(m, mnt->mnt_ns, root); + if (d) + return d->mnt_group_id; + } + + return 0; +} + static int do_make_slave(struct vfsmount *mnt) { struct vfsmount *peer_mnt = mnt, *master = mnt->mnt_master; diff --git a/fs/pnode.h b/fs/pnode.h index 973c3f825e7..958665d662a 100644 --- a/fs/pnode.h +++ b/fs/pnode.h @@ -36,4 +36,5 @@ int propagate_mnt(struct vfsmount *, struct dentry *, struct vfsmount *, int propagate_umount(struct list_head *); int propagate_mount_busy(struct vfsmount *, int); void mnt_release_group_id(struct vfsmount *); +int get_dominating_id(struct vfsmount *mnt, const struct path *root); #endif /* _LINUX_PNODE_H */ -- cgit v1.2.3 From 8a0f572397ca0673c65c1662946014bb73b5cdc6 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Anders=20Grafstr=C3=B6m?= Date: Wed, 12 Mar 2008 20:29:23 +0100 Subject: [JFFS2] Return values of jffs2_block_check_erase error paths It looks the error paths in jffs2_block_check_erase() have wrong return values. A block that failed to be erased never gets marked as bad. Signed-off-by: David Woodhouse --- fs/jffs2/erase.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/jffs2/erase.c b/fs/jffs2/erase.c index 65d91943fc2..5e2719cb693 100644 --- a/fs/jffs2/erase.c +++ b/fs/jffs2/erase.c @@ -350,9 +350,11 @@ static int jffs2_block_check_erase(struct jffs2_sb_info *c, struct jffs2_erasebl break; } while(--retlen); c->mtd->unpoint(c->mtd, ebuf, jeb->offset, c->sector_size); - if (retlen) + if (retlen) { printk(KERN_WARNING "Newly-erased block contained word 0x%lx at offset 0x%08tx\n", *wordebuf, jeb->offset + c->sector_size-retlen*sizeof(*wordebuf)); + return -EIO; + } return 0; } do_flash_read: @@ -373,10 +375,12 @@ static int jffs2_block_check_erase(struct jffs2_sb_info *c, struct jffs2_erasebl ret = c->mtd->read(c->mtd, ofs, readlen, &retlen, ebuf); if (ret) { printk(KERN_WARNING "Read of newly-erased block at 0x%08x failed: %d. Putting on bad_list\n", ofs, ret); + ret = -EIO; goto fail; } if (retlen != readlen) { printk(KERN_WARNING "Short read from newly-erased block at 0x%08x. Wanted %d, got %zd\n", ofs, readlen, retlen); + ret = -EIO; goto fail; } for (i=0; i Date: Wed, 23 Apr 2008 14:15:24 +0100 Subject: [JFFS2] Add erase_checking_list to hold blocks being marked. Just to keep the debug code happy when it's adding all the blocks up. Otherwise, they disappear for a while while the locks are dropped to check them and write the cleanmarker. Signed-off-by: David Woodhouse --- fs/jffs2/build.c | 1 + fs/jffs2/debug.c | 19 +++++++++++++++++++ fs/jffs2/erase.c | 8 ++++---- fs/jffs2/jffs2_fs_sb.h | 1 + 4 files changed, 25 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/jffs2/build.c b/fs/jffs2/build.c index 722a6b68295..d58f845ccb8 100644 --- a/fs/jffs2/build.c +++ b/fs/jffs2/build.c @@ -345,6 +345,7 @@ int jffs2_do_mount_fs(struct jffs2_sb_info *c) INIT_LIST_HEAD(&c->dirty_list); INIT_LIST_HEAD(&c->erasable_list); INIT_LIST_HEAD(&c->erasing_list); + INIT_LIST_HEAD(&c->erase_checking_list); INIT_LIST_HEAD(&c->erase_pending_list); INIT_LIST_HEAD(&c->erasable_pending_wbuf_list); INIT_LIST_HEAD(&c->erase_complete_list); diff --git a/fs/jffs2/debug.c b/fs/jffs2/debug.c index e198468a8c6..5544d31c066 100644 --- a/fs/jffs2/debug.c +++ b/fs/jffs2/debug.c @@ -246,6 +246,10 @@ void __jffs2_dbg_superblock_counts(struct jffs2_sb_info *c) nr_counted++; erasing += c->sector_size; } + list_for_each_entry(jeb, &c->erase_checking_list, list) { + nr_counted++; + erasing += c->sector_size; + } list_for_each_entry(jeb, &c->erase_complete_list, list) { nr_counted++; erasing += c->sector_size; @@ -582,6 +586,21 @@ __jffs2_dbg_dump_block_lists_nolock(struct jffs2_sb_info *c) } } } + if (list_empty(&c->erase_checking_list)) { + printk(JFFS2_DBG "erase_checking_list: empty\n"); + } else { + struct list_head *this; + + list_for_each(this, &c->erase_checking_list) { + struct jffs2_eraseblock *jeb = list_entry(this, struct jffs2_eraseblock, list); + + if (!(jeb->used_size == 0 && jeb->dirty_size == 0 && jeb->wasted_size == 0)) { + printk(JFFS2_DBG "erase_checking_list: %#08x (used %#08x, dirty %#08x, wasted %#08x, unchecked %#08x, free %#08x)\n", + jeb->offset, jeb->used_size, jeb->dirty_size, jeb->wasted_size, + jeb->unchecked_size, jeb->free_size); + } + } + } if (list_empty(&c->erase_pending_list)) { printk(JFFS2_DBG "erase_pending_list: empty\n"); diff --git a/fs/jffs2/erase.c b/fs/jffs2/erase.c index 5e2719cb693..25a640e566d 100644 --- a/fs/jffs2/erase.c +++ b/fs/jffs2/erase.c @@ -116,7 +116,7 @@ void jffs2_erase_pending_blocks(struct jffs2_sb_info *c, int count) if (!list_empty(&c->erase_complete_list)) { jeb = list_entry(c->erase_complete_list.next, struct jffs2_eraseblock, list); - list_del(&jeb->list); + list_move(&jeb->list, &c->erase_checking_list); spin_unlock(&c->erase_completion_lock); mutex_unlock(&c->erase_free_sem); jffs2_mark_erased_block(c, jeb); @@ -465,7 +465,7 @@ static void jffs2_mark_erased_block(struct jffs2_sb_info *c, struct jffs2_eraseb if (c->cleanmarker_size && !jffs2_cleanmarker_oob(c)) jffs2_link_node_ref(c, jeb, jeb->offset | REF_NORMAL, c->cleanmarker_size, NULL); - list_add_tail(&jeb->list, &c->free_list); + list_move_tail(&jeb->list, &c->free_list); c->nr_erasing_blocks--; c->nr_free_blocks++; @@ -482,7 +482,7 @@ filebad: spin_lock(&c->erase_completion_lock); /* Stick it on a list (any list) so erase_failed can take it right off again. Silly, but shouldn't happen often. */ - list_add(&jeb->list, &c->erasing_list); + list_move(&jeb->list, &c->erasing_list); spin_unlock(&c->erase_completion_lock); mutex_unlock(&c->erase_free_sem); jffs2_erase_failed(c, jeb, bad_offset); @@ -493,7 +493,7 @@ refile: jffs2_erase_pending_trigger(c); mutex_lock(&c->erase_free_sem); spin_lock(&c->erase_completion_lock); - list_add(&jeb->list, &c->erase_complete_list); + list_move(&jeb->list, &c->erase_complete_list); spin_unlock(&c->erase_completion_lock); mutex_unlock(&c->erase_free_sem); return; diff --git a/fs/jffs2/jffs2_fs_sb.h b/fs/jffs2/jffs2_fs_sb.h index d9c4b27575e..85ef6dbb1be 100644 --- a/fs/jffs2/jffs2_fs_sb.h +++ b/fs/jffs2/jffs2_fs_sb.h @@ -87,6 +87,7 @@ struct jffs2_sb_info { struct list_head erasable_list; /* Blocks which are completely dirty, and need erasing */ struct list_head erasable_pending_wbuf_list; /* Blocks which need erasing but only after the current wbuf is flushed */ struct list_head erasing_list; /* Blocks which are currently erasing */ + struct list_head erase_checking_list; /* Blocks which are being checked and marked */ struct list_head erase_pending_list; /* Blocks which need erasing now */ struct list_head erase_complete_list; /* Blocks which are erased and need the clean marker written to them */ struct list_head free_list; /* Blocks which are free and ready to be used */ -- cgit v1.2.3 From 422b120238130307da64fa44c9fb722bfaf5f1af Mon Sep 17 00:00:00 2001 From: David Woodhouse Date: Wed, 23 Apr 2008 15:40:52 +0100 Subject: [JFFS2] Fix jffs2_reserve_space() when all blocks are pending erasure. When _all_ the blocks were on the erase_pending_list, we could't find a block to GC from but there was no _actually_ free space, and jffs2_reserve_space() would get a little unhappy. Handle this case by returning -EAGAIN from jffs2_garbage_collect_pass(). There are two callers of that function -- jffs2_flush_wbuf_gc(), which will interpret it as an error and flush the writebuffer by other means, and jffs2_reserve_space(), which we modify to respond to -EAGAIN with an immediate call to jffs2_erase_pending_blocks() and another run round the loop. Signed-off-by: David Woodhouse --- fs/jffs2/gc.c | 8 +++++++- fs/jffs2/nodemgmt.c | 5 ++++- 2 files changed, 11 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/jffs2/gc.c b/fs/jffs2/gc.c index 26c7992c45c..bad005664e3 100644 --- a/fs/jffs2/gc.c +++ b/fs/jffs2/gc.c @@ -221,7 +221,13 @@ int jffs2_garbage_collect_pass(struct jffs2_sb_info *c) jeb = jffs2_find_gc_block(c); if (!jeb) { - D1 (printk(KERN_NOTICE "jffs2: Couldn't find erase block to garbage collect!\n")); + /* Couldn't find a free block. But maybe we can just erase one and make 'progress'? */ + if (!list_empty(&c->erase_pending_list)) { + spin_unlock(&c->erase_completion_lock); + mutex_unlock(&c->alloc_sem); + return -EAGAIN; + } + D1(printk(KERN_NOTICE "jffs2: Couldn't find erase block to garbage collect!\n")); spin_unlock(&c->erase_completion_lock); mutex_unlock(&c->alloc_sem); return -EIO; diff --git a/fs/jffs2/nodemgmt.c b/fs/jffs2/nodemgmt.c index 747a73f0aa4..9df8f3ef20d 100644 --- a/fs/jffs2/nodemgmt.c +++ b/fs/jffs2/nodemgmt.c @@ -116,7 +116,10 @@ int jffs2_reserve_space(struct jffs2_sb_info *c, uint32_t minsize, spin_unlock(&c->erase_completion_lock); ret = jffs2_garbage_collect_pass(c); - if (ret) + + if (ret == -EAGAIN) + jffs2_erase_pending_blocks(c, 1); + else if (ret) return ret; cond_resched(); -- cgit v1.2.3 From 2c61cb250cf7e8cdd3b83b79b76d2ea0b3da010a Mon Sep 17 00:00:00 2001 From: David Woodhouse Date: Wed, 23 Apr 2008 16:43:15 +0100 Subject: [JFFS2] Introduce dbg_readinode2 log level, use it to shut read_dnode() up We haven't seen bugs in this for a while now, since the rewrite. No need to be _quite_ so verbose... Signed-off-by: David Woodhouse --- fs/jffs2/debug.h | 6 ++++++ fs/jffs2/readinode.c | 15 ++++++++------- 2 files changed, 14 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/jffs2/debug.h b/fs/jffs2/debug.h index 4130adabd76..9645275023e 100644 --- a/fs/jffs2/debug.h +++ b/fs/jffs2/debug.h @@ -38,6 +38,7 @@ #if CONFIG_JFFS2_FS_DEBUG > 1 #define JFFS2_DBG_FRAGTREE2_MESSAGES +#define JFFS2_DBG_READINODE2_MESSAGES #define JFFS2_DBG_MEMALLOC_MESSAGES #endif @@ -115,6 +116,11 @@ #else #define dbg_readinode(fmt, ...) #endif +#ifdef JFFS2_DBG_READINODE2_MESSAGES +#define dbg_readinode2(fmt, ...) JFFS2_DEBUG(fmt, ##__VA_ARGS__) +#else +#define dbg_readinode2(fmt, ...) +#endif /* Fragtree build debugging messages */ #ifdef JFFS2_DBG_FRAGTREE_MESSAGES diff --git a/fs/jffs2/readinode.c b/fs/jffs2/readinode.c index 8a7cf1e8d68..4cb4d76de07 100644 --- a/fs/jffs2/readinode.c +++ b/fs/jffs2/readinode.c @@ -825,8 +825,9 @@ static inline int read_dnode(struct jffs2_sb_info *c, struct jffs2_raw_node_ref else // normal case... tn->fn->size = je32_to_cpu(rd->dsize); - dbg_readinode("dnode @%08x: ver %u, offset %#04x, dsize %#04x, csize %#04x\n", - ref_offset(ref), je32_to_cpu(rd->version), je32_to_cpu(rd->offset), je32_to_cpu(rd->dsize), csize); + dbg_readinode2("dnode @%08x: ver %u, offset %#04x, dsize %#04x, csize %#04x\n", + ref_offset(ref), je32_to_cpu(rd->version), + je32_to_cpu(rd->offset), je32_to_cpu(rd->dsize), csize); ret = jffs2_add_tn_to_tree(c, rii, tn); @@ -836,13 +837,13 @@ static inline int read_dnode(struct jffs2_sb_info *c, struct jffs2_raw_node_ref jffs2_free_tmp_dnode_info(tn); return ret; } -#ifdef JFFS2_DBG_READINODE_MESSAGES - dbg_readinode("After adding ver %d:\n", je32_to_cpu(rd->version)); +#ifdef JFFS2_DBG_READINODE2_MESSAGES + dbg_readinode2("After adding ver %d:\n", je32_to_cpu(rd->version)); tn = tn_first(&rii->tn_root); while (tn) { - dbg_readinode("%p: v %d r 0x%x-0x%x ov %d\n", - tn, tn->version, tn->fn->ofs, - tn->fn->ofs+tn->fn->size, tn->overlapped); + dbg_readinode2("%p: v %d r 0x%x-0x%x ov %d\n", + tn, tn->version, tn->fn->ofs, + tn->fn->ofs+tn->fn->size, tn->overlapped); tn = tn_next(tn); } #endif -- cgit v1.2.3 From 1447d25eb3a7bbe5bf5e4e7489f09be13e1ec73a Mon Sep 17 00:00:00 2001 From: NeilBrown Date: Fri, 8 Feb 2008 13:03:37 +1100 Subject: knfsd: Remove NLM_HOST_MAX and associated logic. Lockd caches information about hosts that have recently held locks to expedite the taking of further locks. It periodically discards this information for hosts that have not been used for a few minutes. lockd currently has a value NLM_HOST_MAX, and changes the 'garbage collection' behaviour when the number of hosts exceeds this threshold. However its behaviour is strange, and likely not what was intended. When the number of hosts exceeds the max, it scans *less* often (every 2 minutes vs every minute) and allows unused host information to remain around longer (5 minutes instead of 2). Having this limit is of dubious value anyway, and we have not suffered from the code not getting the limit right, so remove the limit altogether. We go with the larger values (discard 5 minute old hosts every 2 minutes) as they are probably safer. Maybe the periodic garbage collection should be replace to with 'shrinker' handler so we just respond to memory pressure.... Acked-by: Jeff Layton Signed-off-by: Neil Brown Signed-off-by: J. Bruce Fields --- fs/lockd/host.c | 9 +++------ 1 file changed, 3 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/lockd/host.c b/fs/lockd/host.c index f1ef49fff11..c3f119426d8 100644 --- a/fs/lockd/host.c +++ b/fs/lockd/host.c @@ -19,12 +19,11 @@ #define NLMDBG_FACILITY NLMDBG_HOSTCACHE -#define NLM_HOST_MAX 64 #define NLM_HOST_NRHASH 32 #define NLM_ADDRHASH(addr) (ntohl(addr) & (NLM_HOST_NRHASH-1)) #define NLM_HOST_REBIND (60 * HZ) -#define NLM_HOST_EXPIRE ((nrhosts > NLM_HOST_MAX)? 300 * HZ : 120 * HZ) -#define NLM_HOST_COLLECT ((nrhosts > NLM_HOST_MAX)? 120 * HZ : 60 * HZ) +#define NLM_HOST_EXPIRE (300 * HZ) +#define NLM_HOST_COLLECT (120 * HZ) static struct hlist_head nlm_hosts[NLM_HOST_NRHASH]; static unsigned long next_gc; @@ -142,9 +141,7 @@ nlm_lookup_host(int server, const struct sockaddr_in *sin, INIT_LIST_HEAD(&host->h_granted); INIT_LIST_HEAD(&host->h_reclaim); - if (++nrhosts > NLM_HOST_MAX) - next_gc = 0; - + nrhosts++; out: mutex_unlock(&nlm_host_mutex); return host; -- cgit v1.2.3 From d751a7cd0695554498f25d3026ca6710dbb3698f Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Thu, 7 Feb 2008 16:34:55 -0500 Subject: NLM: Convert lockd to use kthreads Have lockd_up start lockd using kthread_run. With this change, lockd_down now blocks until lockd actually exits, so there's no longer need for the waitqueue code at the end of lockd_down. This also means that only one lockd can be running at a time which simplifies the code within lockd's main loop. This also adds a check for kthread_should_stop in the main loop of nlmsvc_retry_blocked and after that function returns. There's no sense continuing to retry blocks if lockd is coming down anyway. Signed-off-by: Jeff Layton Signed-off-by: J. Bruce Fields --- fs/lockd/svc.c | 132 +++++++++++++++++++++++------------------------------ fs/lockd/svclock.c | 3 +- 2 files changed, 60 insertions(+), 75 deletions(-) (limited to 'fs') diff --git a/fs/lockd/svc.c b/fs/lockd/svc.c index 1ed8bd4de94..66b5c98c7ff 100644 --- a/fs/lockd/svc.c +++ b/fs/lockd/svc.c @@ -25,6 +25,7 @@ #include #include #include +#include #include #include @@ -48,14 +49,11 @@ EXPORT_SYMBOL(nlmsvc_ops); static DEFINE_MUTEX(nlmsvc_mutex); static unsigned int nlmsvc_users; -static pid_t nlmsvc_pid; +static struct task_struct *nlmsvc_task; static struct svc_serv *nlmsvc_serv; int nlmsvc_grace_period; unsigned long nlmsvc_timeout; -static DECLARE_COMPLETION(lockd_start_done); -static DECLARE_WAIT_QUEUE_HEAD(lockd_exit); - /* * These can be set at insmod time (useful for NFS as root filesystem), * and also changed through the sysctl interface. -- Jamie Lokier, Aug 2003 @@ -111,35 +109,30 @@ static inline void clear_grace_period(void) /* * This is the lockd kernel thread */ -static void -lockd(struct svc_rqst *rqstp) +static int +lockd(void *vrqstp) { int err = 0; + struct svc_rqst *rqstp = vrqstp; unsigned long grace_period_expire; - /* Lock module and set up kernel thread */ - /* lockd_up is waiting for us to startup, so will - * be holding a reference to this module, so it - * is safe to just claim another reference - */ - __module_get(THIS_MODULE); - lock_kernel(); - - /* - * Let our maker know we're running. - */ - nlmsvc_pid = current->pid; - nlmsvc_serv = rqstp->rq_server; - complete(&lockd_start_done); - - daemonize("lockd"); + /* try_to_freeze() is called from svc_recv() */ set_freezable(); - /* Process request with signals blocked, but allow SIGKILL. */ + /* Allow SIGKILL to tell lockd to drop all of its locks */ allow_signal(SIGKILL); dprintk("NFS locking service started (ver " LOCKD_VERSION ").\n"); + /* + * FIXME: it would be nice if lockd didn't spend its entire life + * running under the BKL. At the very least, it would be good to + * have someone clarify what it's intended to protect here. I've + * seen some handwavy posts about posix locking needing to be + * done under the BKL, but it's far from clear. + */ + lock_kernel(); + if (!nlm_timeout) nlm_timeout = LOCKD_DFLT_TIMEO; nlmsvc_timeout = nlm_timeout * HZ; @@ -148,10 +141,9 @@ lockd(struct svc_rqst *rqstp) /* * The main request loop. We don't terminate until the last - * NFS mount or NFS daemon has gone away, and we've been sent a - * signal, or else another process has taken over our job. + * NFS mount or NFS daemon has gone away. */ - while ((nlmsvc_users || !signalled()) && nlmsvc_pid == current->pid) { + while (!kthread_should_stop()) { long timeout = MAX_SCHEDULE_TIMEOUT; RPC_IFDEBUG(char buf[RPC_MAX_ADDRBUFLEN]); @@ -161,6 +153,7 @@ lockd(struct svc_rqst *rqstp) nlmsvc_invalidate_all(); grace_period_expire = set_grace_period(); } + continue; } /* @@ -195,28 +188,19 @@ lockd(struct svc_rqst *rqstp) } flush_signals(current); + if (nlmsvc_ops) + nlmsvc_invalidate_all(); + nlm_shutdown_hosts(); - /* - * Check whether there's a new lockd process before - * shutting down the hosts and clearing the slot. - */ - if (!nlmsvc_pid || current->pid == nlmsvc_pid) { - if (nlmsvc_ops) - nlmsvc_invalidate_all(); - nlm_shutdown_hosts(); - nlmsvc_pid = 0; - nlmsvc_serv = NULL; - } else - printk(KERN_DEBUG - "lockd: new process, skipping host shutdown\n"); - wake_up(&lockd_exit); + unlock_kernel(); + + nlmsvc_task = NULL; + nlmsvc_serv = NULL; /* Exit the RPC thread */ svc_exit_thread(rqstp); - /* Release module */ - unlock_kernel(); - module_put_and_exit(0); + return 0; } /* @@ -261,14 +245,15 @@ static int make_socks(struct svc_serv *serv, int proto) int lockd_up(int proto) /* Maybe add a 'family' option when IPv6 is supported ?? */ { - struct svc_serv * serv; - int error = 0; + struct svc_serv *serv; + struct svc_rqst *rqstp; + int error = 0; mutex_lock(&nlmsvc_mutex); /* * Check whether we're already up and running. */ - if (nlmsvc_pid) { + if (nlmsvc_serv) { if (proto) error = make_socks(nlmsvc_serv, proto); goto out; @@ -295,13 +280,28 @@ lockd_up(int proto) /* Maybe add a 'family' option when IPv6 is supported ?? */ /* * Create the kernel thread and wait for it to start. */ - error = svc_create_thread(lockd, serv); - if (error) { + rqstp = svc_prepare_thread(serv, &serv->sv_pools[0]); + if (IS_ERR(rqstp)) { + error = PTR_ERR(rqstp); printk(KERN_WARNING - "lockd_up: create thread failed, error=%d\n", error); + "lockd_up: svc_rqst allocation failed, error=%d\n", + error); + goto destroy_and_out; + } + + svc_sock_update_bufs(serv); + nlmsvc_serv = rqstp->rq_server; + + nlmsvc_task = kthread_run(lockd, rqstp, serv->sv_name); + if (IS_ERR(nlmsvc_task)) { + error = PTR_ERR(nlmsvc_task); + nlmsvc_task = NULL; + nlmsvc_serv = NULL; + printk(KERN_WARNING + "lockd_up: kthread_run failed, error=%d\n", error); + svc_exit_thread(rqstp); goto destroy_and_out; } - wait_for_completion(&lockd_start_done); /* * Note: svc_serv structures have an initial use count of 1, @@ -323,37 +323,21 @@ EXPORT_SYMBOL(lockd_up); void lockd_down(void) { - static int warned; - mutex_lock(&nlmsvc_mutex); if (nlmsvc_users) { if (--nlmsvc_users) goto out; - } else - printk(KERN_WARNING "lockd_down: no users! pid=%d\n", nlmsvc_pid); - - if (!nlmsvc_pid) { - if (warned++ == 0) - printk(KERN_WARNING "lockd_down: no lockd running.\n"); - goto out; + } else { + printk(KERN_ERR "lockd_down: no users! task=%p\n", + nlmsvc_task); + BUG(); } - warned = 0; - kill_proc(nlmsvc_pid, SIGKILL, 1); - /* - * Wait for the lockd process to exit, but since we're holding - * the lockd semaphore, we can't wait around forever ... - */ - clear_thread_flag(TIF_SIGPENDING); - interruptible_sleep_on_timeout(&lockd_exit, HZ); - if (nlmsvc_pid) { - printk(KERN_WARNING - "lockd_down: lockd failed to exit, clearing pid\n"); - nlmsvc_pid = 0; + if (!nlmsvc_task) { + printk(KERN_ERR "lockd_down: no lockd running.\n"); + BUG(); } - spin_lock_irq(¤t->sighand->siglock); - recalc_sigpending(); - spin_unlock_irq(¤t->sighand->siglock); + kthread_stop(nlmsvc_task); out: mutex_unlock(&nlmsvc_mutex); } diff --git a/fs/lockd/svclock.c b/fs/lockd/svclock.c index fe9bdb4a220..4da7c4c2706 100644 --- a/fs/lockd/svclock.c +++ b/fs/lockd/svclock.c @@ -29,6 +29,7 @@ #include #include #include +#include #define NLMDBG_FACILITY NLMDBG_SVCLOCK @@ -887,7 +888,7 @@ nlmsvc_retry_blocked(void) unsigned long timeout = MAX_SCHEDULE_TIMEOUT; struct nlm_block *block; - while (!list_empty(&nlm_blocked)) { + while (!list_empty(&nlm_blocked) && !kthread_should_stop()) { block = list_entry(nlm_blocked.next, struct nlm_block, b_list); if (block->b_when == NLM_NEVER) -- cgit v1.2.3 From f15364bd4cf8799a7677b6daeed7b67d9139d974 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Aur=C3=A9lien=20Charbon?= Date: Fri, 18 Jan 2008 15:50:56 +0100 Subject: IPv6 support for NFS server export caches MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This adds IPv6 support to the interfaces that are used to express nfsd exports. All addressed are stored internally as IPv6; backwards compatibility is maintained using mapped addresses. Thanks to Bruce Fields, Brian Haley, Neil Brown and Hideaki Joshifuji for comments Signed-off-by: Aurelien Charbon Cc: Neil Brown Cc: Brian Haley Cc: YOSHIFUJI Hideaki / å‰è—¤è‹±æ˜Ž Signed-off-by: J. Bruce Fields --- fs/nfsd/export.c | 9 ++++++--- fs/nfsd/nfsctl.c | 15 +++++++++++++-- 2 files changed, 19 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/export.c b/fs/nfsd/export.c index 8a6f7c924c7..33bfcf09db4 100644 --- a/fs/nfsd/export.c +++ b/fs/nfsd/export.c @@ -35,6 +35,7 @@ #include #include #include +#include #define NFSDDBG_FACILITY NFSDDBG_EXPORT @@ -1548,6 +1549,7 @@ exp_addclient(struct nfsctl_client *ncp) { struct auth_domain *dom; int i, err; + struct in6_addr addr6; /* First, consistency check. */ err = -EINVAL; @@ -1566,9 +1568,10 @@ exp_addclient(struct nfsctl_client *ncp) goto out_unlock; /* Insert client into hashtable. */ - for (i = 0; i < ncp->cl_naddr; i++) - auth_unix_add_addr(ncp->cl_addrlist[i], dom); - + for (i = 0; i < ncp->cl_naddr; i++) { + ipv6_addr_set_v4mapped(ncp->cl_addrlist[i].s_addr, &addr6); + auth_unix_add_addr(&addr6, dom); + } auth_unix_forget_old(dom); auth_domain_put(dom); diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index 8516137cdbb..9f038a4a148 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -37,6 +37,7 @@ #include #include +#include /* * We have a single directory with 9 nodes in it. @@ -222,6 +223,7 @@ static ssize_t write_getfs(struct file *file, char *buf, size_t size) struct auth_domain *clp; int err = 0; struct knfsd_fh *res; + struct in6_addr in6; if (size < sizeof(*data)) return -EINVAL; @@ -236,7 +238,11 @@ static ssize_t write_getfs(struct file *file, char *buf, size_t size) res = (struct knfsd_fh*)buf; exp_readlock(); - if (!(clp = auth_unix_lookup(sin->sin_addr))) + + ipv6_addr_set_v4mapped(sin->sin_addr.s_addr, &in6); + + clp = auth_unix_lookup(&in6); + if (!clp) err = -EPERM; else { err = exp_rootfh(clp, data->gd_path, res, data->gd_maxlen); @@ -257,6 +263,7 @@ static ssize_t write_getfd(struct file *file, char *buf, size_t size) int err = 0; struct knfsd_fh fh; char *res; + struct in6_addr in6; if (size < sizeof(*data)) return -EINVAL; @@ -271,7 +278,11 @@ static ssize_t write_getfd(struct file *file, char *buf, size_t size) res = buf; sin = (struct sockaddr_in *)&data->gd_addr; exp_readlock(); - if (!(clp = auth_unix_lookup(sin->sin_addr))) + + ipv6_addr_set_v4mapped(sin->sin_addr.s_addr, &in6); + + clp = auth_unix_lookup(&in6); + if (!clp) err = -EPERM; else { err = exp_rootfh(clp, data->gd_path, &fh, NFS_FHSIZE); -- cgit v1.2.3 From 065f30ec14b1460c695b371bc44e068832a60d9b Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Sat, 19 Jan 2008 13:58:23 -0500 Subject: nfs: remove unnecessary NFS_NEED_* defines Thanks to Robert Day for pointing out that these two defines are unused. Signed-off-by: J. Bruce Fields Cc: Trond Myklebust Trond Myklebust Cc: Neil Brown Cc: "Robert P. J. Day" --- fs/nfs/symlink.c | 1 - 1 file changed, 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/symlink.c b/fs/nfs/symlink.c index 83e865a16ad..412738dbfbc 100644 --- a/fs/nfs/symlink.c +++ b/fs/nfs/symlink.c @@ -10,7 +10,6 @@ * nfs symlink handling code */ -#define NFS_NEED_XDR_TYPES #include #include #include -- cgit v1.2.3 From f3362737be14668f4e8f5c8d082eb131aafc1353 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Sat, 26 Jan 2008 14:58:45 -0500 Subject: nfsd4: remove unnecessary CHECK_FH check in preprocess_seqid_op Every caller sets this flag, so it's meaningless. Signed-off-by: J. Bruce Fields --- fs/nfsd/nfs4state.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index 81a75f3081f..b73e96db1f5 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -2127,7 +2127,7 @@ nfs4_preprocess_seqid_op(struct svc_fh *current_fh, u32 seqid, stateid_t *statei } } - if ((flags & CHECK_FH) && nfs4_check_fh(current_fh, stp)) { + if (nfs4_check_fh(current_fh, stp)) { dprintk("NFSD: preprocess_seqid_op: fh-stateid mismatch!\n"); return nfserr_bad_stateid; } @@ -2194,7 +2194,7 @@ nfsd4_open_confirm(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, if ((status = nfs4_preprocess_seqid_op(&cstate->current_fh, oc->oc_seqid, &oc->oc_req_stateid, - CHECK_FH | CONFIRM | OPEN_STATE, + CONFIRM | OPEN_STATE, &oc->oc_stateowner, &stp, NULL))) goto out; @@ -2265,7 +2265,7 @@ nfsd4_open_downgrade(struct svc_rqst *rqstp, if ((status = nfs4_preprocess_seqid_op(&cstate->current_fh, od->od_seqid, &od->od_stateid, - CHECK_FH | OPEN_STATE, + OPEN_STATE, &od->od_stateowner, &stp, NULL))) goto out; @@ -2318,7 +2318,7 @@ nfsd4_close(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, if ((status = nfs4_preprocess_seqid_op(&cstate->current_fh, close->cl_seqid, &close->cl_stateid, - CHECK_FH | OPEN_STATE | CLOSE_STATE, + OPEN_STATE | CLOSE_STATE, &close->cl_stateowner, &stp, NULL))) goto out; status = nfs_ok; @@ -2623,7 +2623,7 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, status = nfs4_preprocess_seqid_op(&cstate->current_fh, lock->lk_new_open_seqid, &lock->lk_new_open_stateid, - CHECK_FH | OPEN_STATE, + OPEN_STATE, &lock->lk_replay_owner, &open_stp, lock); if (status) @@ -2650,7 +2650,7 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, status = nfs4_preprocess_seqid_op(&cstate->current_fh, lock->lk_old_lock_seqid, &lock->lk_old_lock_stateid, - CHECK_FH | LOCK_STATE, + LOCK_STATE, &lock->lk_replay_owner, &lock_stp, lock); if (status) goto out; @@ -2847,7 +2847,7 @@ nfsd4_locku(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, if ((status = nfs4_preprocess_seqid_op(&cstate->current_fh, locku->lu_seqid, &locku->lu_stateid, - CHECK_FH | LOCK_STATE, + LOCK_STATE, &locku->lu_stateowner, &stp, NULL))) goto out; -- cgit v1.2.3 From 0836f587258c2a24bfdc8810ad2327e7f354b6c7 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Sat, 26 Jan 2008 19:08:12 -0500 Subject: nfsd4: simplify stateid sequencing checks Pull this common code into a separate function. Signed-off-by: J. Bruce Fields --- fs/nfsd/nfs4state.c | 41 ++++++++++++++++++++++++++--------------- 1 file changed, 26 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index b73e96db1f5..b83b58da0cc 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -1975,6 +1975,26 @@ io_during_grace_disallowed(struct inode *inode, int flags) && mandatory_lock(inode); } +static int check_stateid_generation(stateid_t *in, stateid_t *ref) +{ + /* If the client sends us a stateid from the future, it's buggy: */ + if (in->si_generation > ref->si_generation) + return nfserr_bad_stateid; + /* + * The following, however, can happen. For example, if the + * client sends an open and some IO at the same time, the open + * may bump si_generation while the IO is still in flight. + * Thanks to hard links and renames, the client never knows what + * file an open will affect. So it could avoid that situation + * only by serializing all opens and IO from the same open + * owner. To recover from the old_stateid error, the client + * will just have to retry the IO: + */ + if (in->si_generation < ref->si_generation) + return nfserr_old_stateid; + return nfs_ok; +} + /* * Checks for stateid operations */ @@ -2023,12 +2043,8 @@ nfs4_preprocess_stateid_op(struct svc_fh *current_fh, stateid_t *stateid, int fl goto out; stidp = &stp->st_stateid; } - if (stateid->si_generation > stidp->si_generation) - goto out; - - /* OLD STATEID */ - status = nfserr_old_stateid; - if (stateid->si_generation < stidp->si_generation) + status = check_stateid_generation(stateid, stidp); + if (status) goto out; if (stp) { if ((status = nfs4_check_openmode(stp,flags))) @@ -2065,6 +2081,7 @@ nfs4_preprocess_seqid_op(struct svc_fh *current_fh, u32 seqid, stateid_t *statei { struct nfs4_stateid *stp; struct nfs4_stateowner *sop; + __be32 status; dprintk("NFSD: preprocess_seqid_op: seqid=%d " "stateid = (%08x/%08x/%08x/%08x)\n", seqid, @@ -2150,15 +2167,9 @@ nfs4_preprocess_seqid_op(struct svc_fh *current_fh, u32 seqid, stateid_t *statei " confirmed yet!\n"); return nfserr_bad_stateid; } - if (stateid->si_generation > stp->st_stateid.si_generation) { - dprintk("NFSD: preprocess_seqid_op: future stateid?!\n"); - return nfserr_bad_stateid; - } - - if (stateid->si_generation < stp->st_stateid.si_generation) { - dprintk("NFSD: preprocess_seqid_op: old stateid!\n"); - return nfserr_old_stateid; - } + status = check_stateid_generation(stateid, &stp->st_stateid); + if (status) + return status; renew_client(sop->so_client); return nfs_ok; -- cgit v1.2.3 From 6a85fa3adddd3a74bd5b94c4b72668d307b88377 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Sat, 26 Jan 2008 23:36:48 -0500 Subject: nfsd4: kill unnecessary check in preprocess_stateid_op This condition is always true. Signed-off-by: J. Bruce Fields --- fs/nfsd/nfs4state.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index b83b58da0cc..a40d1ec52fe 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -2052,7 +2052,7 @@ nfs4_preprocess_stateid_op(struct svc_fh *current_fh, stateid_t *stateid, int fl renew_client(stp->st_stateowner->so_client); if (filpp) *filpp = stp->st_vfs_file; - } else if (dp) { + } else { if ((status = nfs4_check_delegmode(dp, flags))) goto out; renew_client(dp->dl_client); -- cgit v1.2.3 From c0ce6ec87c59d7a29438717b1f72f83fb408f416 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Mon, 11 Feb 2008 15:48:47 -0500 Subject: nfsd: clarify readdir/mountpoint-crossing code The code here is difficult to understand; attempt to clarify somewhat by pulling out one of the more mystifying conditionals into a separate function. While we're here, also add lease_time to the list of attributes that we don't really need to cross a mountpoint to fetch. Signed-off-by: J. Bruce Fields Cc: Peter Staubach --- fs/nfsd/nfs4xdr.c | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c index 0e6a179ecca..1ba7ad98193 100644 --- a/fs/nfsd/nfs4xdr.c +++ b/fs/nfsd/nfs4xdr.c @@ -1867,6 +1867,15 @@ out_serverfault: goto out; } +static inline int attributes_need_mount(u32 *bmval) +{ + if (bmval[0] & ~(FATTR4_WORD0_RDATTR_ERROR | FATTR4_WORD0_LEASE_TIME)) + return 1; + if (bmval[1] & ~FATTR4_WORD1_MOUNTED_ON_FILEID) + return 1; + return 0; +} + static __be32 nfsd4_encode_dirent_fattr(struct nfsd4_readdir *cd, const char *name, int namlen, __be32 *p, int *buflen) @@ -1888,9 +1897,7 @@ nfsd4_encode_dirent_fattr(struct nfsd4_readdir *cd, * we will not follow the cross mount and will fill the attribtutes * directly from the mountpoint dentry. */ - if (d_mountpoint(dentry) && - (cd->rd_bmval[0] & ~FATTR4_WORD0_RDATTR_ERROR) == 0 && - (cd->rd_bmval[1] & ~FATTR4_WORD1_MOUNTED_ON_FILEID) == 0) + if (d_mountpoint(dentry) && !attributes_need_mount(cd->rd_bmval)) ignore_crossmnt = 1; else if (d_mountpoint(dentry)) { int err; -- cgit v1.2.3 From 5ea0dd61f221ba2701314a85e998b8202412553d Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Mon, 11 Feb 2008 17:11:39 -0500 Subject: NFSD: Remove NFSD_TCP kernel build option Likewise, distros usually leave CONFIG_NFSD_TCP enabled. TCP support in the Linux NFS server is stable enough that we can leave it on always. CONFIG_NFSD_TCP adds about 10 lines of code, and defaults to "Y" anyway. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields --- fs/Kconfig | 10 ---------- fs/nfsd/nfssvc.c | 2 -- 2 files changed, 12 deletions(-) (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index 8b18a875867..12002775d29 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -1696,7 +1696,6 @@ config NFSD select EXPORTFS select NFSD_V2_ACL if NFSD_V3_ACL select NFS_ACL_SUPPORT if NFSD_V2_ACL - select NFSD_TCP if NFSD_V4 select CRYPTO_MD5 if NFSD_V4 select CRYPTO if NFSD_V4 select FS_POSIX_ACL if NFSD_V4 @@ -1755,15 +1754,6 @@ config NFSD_V4 should only be used if you are interested in helping to test NFSv4. If unsure, say N. -config NFSD_TCP - bool "Provide NFS server over TCP support" - depends on NFSD - default y - help - If you want your NFS server to support TCP connections, say Y here. - TCP connections usually perform better than the default UDP when - the network is lossy or congested. If unsure, say Y. - config ROOT_NFS bool "Root file system on NFS" depends on NFS_FS=y && IP_PNP diff --git a/fs/nfsd/nfssvc.c b/fs/nfsd/nfssvc.c index 9647b0f7bc0..941041f4b13 100644 --- a/fs/nfsd/nfssvc.c +++ b/fs/nfsd/nfssvc.c @@ -244,7 +244,6 @@ static int nfsd_init_socks(int port) if (error < 0) return error; -#ifdef CONFIG_NFSD_TCP error = lockd_up(IPPROTO_TCP); if (error >= 0) { error = svc_create_xprt(nfsd_serv, "tcp", port, @@ -254,7 +253,6 @@ static int nfsd_init_socks(int port) } if (error < 0) return error; -#endif return 0; } -- cgit v1.2.3 From d24455b5ffe02a652e8cb1ed2d3570a512c898f8 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Mon, 11 Feb 2008 17:11:54 -0500 Subject: NFSD: Update help text for CONFIG_NFSD Clean up: refresh the help text for Kconfig items related to the NFS server. Remove obsolete URLs, and make the language consistent among the options. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields --- fs/Kconfig | 76 ++++++++++++++++++++++++++++++++++++++------------------------ 1 file changed, 47 insertions(+), 29 deletions(-) (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index 12002775d29..a1dccb6a9c9 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -1702,56 +1702,74 @@ config NFSD select PROC_FS if NFSD_V4 select PROC_FS if SUNRPC_GSS help - If you want your Linux box to act as an NFS *server*, so that other - computers on your local network which support NFS can access certain - directories on your box transparently, you have two options: you can - use the self-contained user space program nfsd, in which case you - should say N here, or you can say Y and use the kernel based NFS - server. The advantage of the kernel based solution is that it is - faster. + Choose Y here if you want to allow other computers to access + files residing on this system using Sun's Network File System + protocol. To compile the NFS server support as a module, + choose M here: the module will be called nfsd. - In either case, you will need support software; the respective - locations are given in the file in the - NFS section. + You may choose to use a user-space NFS server instead, in which + case you can choose N here. - If you say Y here, you will get support for version 2 of the NFS - protocol (NFSv2). If you also want NFSv3, say Y to the next question - as well. + To export local file systems using NFS, you also need to install + user space programs which can be found in the Linux nfs-utils + package, available from http://linux-nfs.org/. More detail about + the Linux NFS server implementation is available via the + exports(5) man page. - Please read the NFS-HOWTO, available from - . + Below you can choose which versions of the NFS protocol are + available to clients mounting the NFS server on this system. + Support for NFS version 2 (RFC 1094) is always available when + CONFIG_NFSD is selected. - To compile the NFS server support as a module, choose M here: the - module will be called nfsd. If unsure, say N. + If unsure, say N. config NFSD_V2_ACL bool depends on NFSD config NFSD_V3 - bool "Provide NFSv3 server support" + bool "NFS server support for NFS version 3" depends on NFSD help - If you would like to include the NFSv3 server as well as the NFSv2 - server, say Y here. If unsure, say Y. + This option enables support in your system's NFS server for + version 3 of the NFS protocol (RFC 1813). + + If unsure, say Y. config NFSD_V3_ACL - bool "Provide server support for the NFSv3 ACL protocol extension" + bool "NFS server support for the NFSv3 ACL protocol extension" depends on NFSD_V3 help - Implement the NFSv3 ACL protocol extension for manipulating POSIX - Access Control Lists on exported file systems. NFS clients should - be compiled with the NFSv3 ACL protocol extension; see the - CONFIG_NFS_V3_ACL option. If unsure, say N. + Solaris NFS servers support an auxiliary NFSv3 ACL protocol that + never became an official part of the NFS version 3 protocol. + This protocol extension allows applications on NFS clients to + manipulate POSIX Access Control Lists on files residing on NFS + servers. NFS servers enforce POSIX ACLs on local files whether + this protocol is available or not. + + This option enables support in your system's NFS server for the + NFSv3 ACL protocol extension allowing NFS clients to manipulate + POSIX ACLs on files exported by your system's NFS server. NFS + clients which support the Solaris NFSv3 ACL protocol can then + access and modify ACLs on your NFS server. + + To store ACLs on your NFS server, you also need to enable ACL- + related CONFIG options for your local file systems of choice. + + If unsure, say N. config NFSD_V4 - bool "Provide NFSv4 server support (EXPERIMENTAL)" + bool "NFS server support for NFS version 4 (EXPERIMENTAL)" depends on NFSD && NFSD_V3 && EXPERIMENTAL select RPCSEC_GSS_KRB5 help - If you would like to include the NFSv4 server as well as the NFSv2 - and NFSv3 servers, say Y here. This feature is experimental, and - should only be used if you are interested in helping to test NFSv4. + This option enables support in your system's NFS server for + version 4 of the NFS protocol (RFC 3530). + + To export files using NFSv4, you need to install additional user + space programs which can be found in the Linux nfs-utils package, + available from http://linux-nfs.org/. + If unsure, say N. config ROOT_NFS -- cgit v1.2.3 From 892069552eedfa0d2fca41680d076a5bc0a3c26e Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Mon, 11 Feb 2008 17:12:24 -0500 Subject: NFSD: Move "select FS_POSIX_ACL if NFSD_V4" Clean up: since FS_POSIX_ACL is a non-visible boolean entry, it can be selected safely under the NFSD_V4 entry (also a boolean). Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields --- fs/Kconfig | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index a1dccb6a9c9..94250f0c0cc 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -411,7 +411,7 @@ config JFS_STATISTICS to be made available to the user in the /proc/fs/jfs/ directory. config FS_POSIX_ACL -# Posix ACL utility routines (for now, only ext2/ext3/jfs/reiserfs) +# Posix ACL utility routines (for now, only ext2/ext3/jfs/reiserfs/nfs4) # # NOTE: you can implement Posix ACLs without these helpers (XFS does). # Never use this symbol for ifdefs. @@ -1698,7 +1698,6 @@ config NFSD select NFS_ACL_SUPPORT if NFSD_V2_ACL select CRYPTO_MD5 if NFSD_V4 select CRYPTO if NFSD_V4 - select FS_POSIX_ACL if NFSD_V4 select PROC_FS if NFSD_V4 select PROC_FS if SUNRPC_GSS help @@ -1761,6 +1760,7 @@ config NFSD_V3_ACL config NFSD_V4 bool "NFS server support for NFS version 4 (EXPERIMENTAL)" depends on NFSD && NFSD_V3 && EXPERIMENTAL + select FS_POSIX_ACL select RPCSEC_GSS_KRB5 help This option enables support in your system's NFS server for -- cgit v1.2.3 From 78dd0992a3462e05138cf4d08df759bf1c7a08c9 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Mon, 11 Feb 2008 17:12:31 -0500 Subject: NFSD: Move "select NFSD_V2_ACL if NFSD_V3_ACL" Clean up: since NFSD_V2_ACL is a boolean, it can be selected safely under the NFSD_V3_ACL entry (also a boolean). Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields --- fs/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index 94250f0c0cc..b609843f8d6 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -1694,7 +1694,6 @@ config NFSD select LOCKD select SUNRPC select EXPORTFS - select NFSD_V2_ACL if NFSD_V3_ACL select NFS_ACL_SUPPORT if NFSD_V2_ACL select CRYPTO_MD5 if NFSD_V4 select CRYPTO if NFSD_V4 @@ -1738,6 +1737,7 @@ config NFSD_V3 config NFSD_V3_ACL bool "NFS server support for the NFSv3 ACL protocol extension" depends on NFSD_V3 + select NFSD_V2_ACL help Solaris NFS servers support an auxiliary NFSv3 ACL protocol that never became an official part of the NFS version 3 protocol. -- cgit v1.2.3 From 6aaa67b5f3b9fe24f0c76d0415cc72e5a1137bea Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Mon, 11 Feb 2008 17:12:38 -0500 Subject: NFSD: Remove redundant "select" clauses in fs/Kconfig As far as I can tell, selecting the CRYPTO and CRYPTO_MD5 entries under CONFIG_NFSD is redundant, since CONFIG_NFSD_V4 already selects RPCSEC_GSS_KRB5, which selects these entries. Testing with "make menuconfig" shows that the entries under CRYPTO still properly reflect "Y" or "M" based on the setting of CONFIG_NFSD after this change is applied. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields --- fs/Kconfig | 2 -- 1 file changed, 2 deletions(-) (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index b609843f8d6..1d81be3fef3 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -1695,8 +1695,6 @@ config NFSD select SUNRPC select EXPORTFS select NFS_ACL_SUPPORT if NFSD_V2_ACL - select CRYPTO_MD5 if NFSD_V4 - select CRYPTO if NFSD_V4 select PROC_FS if NFSD_V4 select PROC_FS if SUNRPC_GSS help -- cgit v1.2.3 From f2b0dee2ec8d562ad9ced2b7481be72d356c6cfc Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Wed, 13 Feb 2008 23:30:26 +0200 Subject: make nfsd_create_setattr() static This patch makes the needlessly global nfsd_create_setattr() static. Signed-off-by: Adrian Bunk Signed-off-by: J. Bruce Fields --- fs/nfsd/vfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index 304bf5f643c..d5a238e6e6e 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1152,7 +1152,7 @@ nfsd_commit(struct svc_rqst *rqstp, struct svc_fh *fhp, } #endif /* CONFIG_NFSD_V3 */ -__be32 +static __be32 nfsd_create_setattr(struct svc_rqst *rqstp, struct svc_fh *resfhp, struct iattr *iap) { -- cgit v1.2.3 From 93245d11fcaccdebccabe86a2b92db524f82d8b4 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Mon, 18 Feb 2008 02:01:49 -0800 Subject: lockd: fix sparse warning in svcshare.c fs/lockd/svcshare.c:74:50: warning: Using plain integer as NULL pointer Signed-off-by: Harvey Harrison Cc: Neil Brown Signed-off-by: Andrew Morton Signed-off-by: J. Bruce Fields --- fs/lockd/svcshare.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/lockd/svcshare.c b/fs/lockd/svcshare.c index 068886de4dd..b0ae0700870 100644 --- a/fs/lockd/svcshare.c +++ b/fs/lockd/svcshare.c @@ -71,7 +71,8 @@ nlmsvc_unshare_file(struct nlm_host *host, struct nlm_file *file, struct nlm_share *share, **shpp; struct xdr_netobj *oh = &argp->lock.oh; - for (shpp = &file->f_shares; (share = *shpp) != 0; shpp = &share->s_next) { + for (shpp = &file->f_shares; (share = *shpp) != NULL; + shpp = &share->s_next) { if (share->s_host == host && nlm_cmp_owner(share, oh)) { *shpp = share->s_next; kfree(share); -- cgit v1.2.3 From 164f98adbbd50c67177b096a59f55c1a56a45c82 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Wed, 20 Feb 2008 14:02:47 -0500 Subject: lockd: fix race in nlm_release() The sm_count is decremented to zero but left on the nsm_handles list. So in the space between decrementing sm_count and acquiring nsm_mutex, it is possible for another task to find this nsm_handle, increment the use count and then enter nsm_release itself. Thus there's nothing to prevent the nsm being freed before we acquire nsm_mutex here. Signed-off-by: J. Bruce Fields --- fs/lockd/host.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/lockd/host.c b/fs/lockd/host.c index c3f119426d8..960911c4a11 100644 --- a/fs/lockd/host.c +++ b/fs/lockd/host.c @@ -529,12 +529,10 @@ nsm_release(struct nsm_handle *nsm) { if (!nsm) return; + mutex_lock(&nsm_mutex); if (atomic_dec_and_test(&nsm->sm_count)) { - mutex_lock(&nsm_mutex); - if (atomic_read(&nsm->sm_count) == 0) { - list_del(&nsm->sm_link); - kfree(nsm); - } - mutex_unlock(&nsm_mutex); + list_del(&nsm->sm_link); + kfree(nsm); } + mutex_unlock(&nsm_mutex); } -- cgit v1.2.3 From a95e56e72c196970a8067cd515c658d064813170 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Wed, 20 Feb 2008 15:27:31 -0500 Subject: lockd: clean up __nsm_find() Use list_for_each_entry(). Also, in keeping with kernel style, make the normal case (kzalloc succeeds) unindented and handle the abnormal case with a goto. Signed-off-by: J. Bruce Fields --- fs/lockd/host.c | 32 ++++++++++++++++---------------- 1 file changed, 16 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/lockd/host.c b/fs/lockd/host.c index 960911c4a11..de0ffb6106c 100644 --- a/fs/lockd/host.c +++ b/fs/lockd/host.c @@ -465,7 +465,7 @@ __nsm_find(const struct sockaddr_in *sin, int create) { struct nsm_handle *nsm = NULL; - struct list_head *pos; + struct nsm_handle *pos; if (!sin) return NULL; @@ -480,16 +480,16 @@ __nsm_find(const struct sockaddr_in *sin, } mutex_lock(&nsm_mutex); - list_for_each(pos, &nsm_handles) { - nsm = list_entry(pos, struct nsm_handle, sm_link); + list_for_each_entry(pos, &nsm_handles, sm_link) { if (hostname && nsm_use_hostnames) { - if (strlen(nsm->sm_name) != hostname_len - || memcmp(nsm->sm_name, hostname, hostname_len)) + if (strlen(pos->sm_name) != hostname_len + || memcmp(pos->sm_name, hostname, hostname_len)) continue; - } else if (!nlm_cmp_addr(&nsm->sm_addr, sin)) + } else if (!nlm_cmp_addr(&pos->sm_addr, sin)) continue; - atomic_inc(&nsm->sm_count); + atomic_inc(&pos->sm_count); + nsm = pos; goto out; } @@ -499,15 +499,15 @@ __nsm_find(const struct sockaddr_in *sin, } nsm = kzalloc(sizeof(*nsm) + hostname_len + 1, GFP_KERNEL); - if (nsm != NULL) { - nsm->sm_addr = *sin; - nsm->sm_name = (char *) (nsm + 1); - memcpy(nsm->sm_name, hostname, hostname_len); - nsm->sm_name[hostname_len] = '\0'; - atomic_set(&nsm->sm_count, 1); - - list_add(&nsm->sm_link, &nsm_handles); - } + if (nsm == NULL) + goto out; + nsm->sm_addr = *sin; + nsm->sm_name = (char *) (nsm + 1); + memcpy(nsm->sm_name, hostname, hostname_len); + nsm->sm_name[hostname_len] = '\0'; + atomic_set(&nsm->sm_count, 1); + + list_add(&nsm->sm_link, &nsm_handles); out: mutex_unlock(&nsm_mutex); -- cgit v1.2.3 From d8421202121ce74daf4625ca9d1d825bbd7ce66a Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Wed, 20 Feb 2008 15:40:15 -0500 Subject: lockd: convert nsm_mutex to a spinlock There's no reason for a mutex here, except to allow an allocation under the lock, which we can avoid with the usual trick of preallocating memory for the new object and freeing it if it turns out to be unnecessary. Signed-off-by: J. Bruce Fields --- fs/lockd/host.c | 34 +++++++++++++++++++--------------- 1 file changed, 19 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/lockd/host.c b/fs/lockd/host.c index de0ffb6106c..c7854791898 100644 --- a/fs/lockd/host.c +++ b/fs/lockd/host.c @@ -457,7 +457,7 @@ nlm_gc_hosts(void) * Manage NSM handles */ static LIST_HEAD(nsm_handles); -static DEFINE_MUTEX(nsm_mutex); +static DEFINE_SPINLOCK(nsm_lock); static struct nsm_handle * __nsm_find(const struct sockaddr_in *sin, @@ -479,7 +479,8 @@ __nsm_find(const struct sockaddr_in *sin, return NULL; } - mutex_lock(&nsm_mutex); +retry: + spin_lock(&nsm_lock); list_for_each_entry(pos, &nsm_handles, sm_link) { if (hostname && nsm_use_hostnames) { @@ -489,28 +490,32 @@ __nsm_find(const struct sockaddr_in *sin, } else if (!nlm_cmp_addr(&pos->sm_addr, sin)) continue; atomic_inc(&pos->sm_count); + kfree(nsm); nsm = pos; - goto out; + goto found; } - - if (!create) { - nsm = NULL; - goto out; + if (nsm) { + list_add(&nsm->sm_link, &nsm_handles); + goto found; } + spin_unlock(&nsm_lock); + + if (!create) + return NULL; nsm = kzalloc(sizeof(*nsm) + hostname_len + 1, GFP_KERNEL); if (nsm == NULL) - goto out; + return NULL; + nsm->sm_addr = *sin; nsm->sm_name = (char *) (nsm + 1); memcpy(nsm->sm_name, hostname, hostname_len); nsm->sm_name[hostname_len] = '\0'; atomic_set(&nsm->sm_count, 1); + goto retry; - list_add(&nsm->sm_link, &nsm_handles); - -out: - mutex_unlock(&nsm_mutex); +found: + spin_unlock(&nsm_lock); return nsm; } @@ -529,10 +534,9 @@ nsm_release(struct nsm_handle *nsm) { if (!nsm) return; - mutex_lock(&nsm_mutex); - if (atomic_dec_and_test(&nsm->sm_count)) { + if (atomic_dec_and_lock(&nsm->sm_count, &nsm_lock)) { list_del(&nsm->sm_link); + spin_unlock(&nsm_lock); kfree(nsm); } - mutex_unlock(&nsm_mutex); } -- cgit v1.2.3 From a254b246ee238ab90e7b3fae1f76875b608b2213 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 20 Feb 2008 12:49:00 -0800 Subject: nfsd: fix sparse warnings Add extern to nfsd/nfsd.h fs/nfsd/nfssvc.c:146:5: warning: symbol 'nfsd_nrthreads' was not declared. Should it be static? fs/nfsd/nfssvc.c:261:5: warning: symbol 'nfsd_nrpools' was not declared. Should it be static? fs/nfsd/nfssvc.c:269:5: warning: symbol 'nfsd_get_nrthreads' was not declared. Should it be static? fs/nfsd/nfssvc.c:281:5: warning: symbol 'nfsd_set_nrthreads' was not declared. Should it be static? fs/nfsd/export.c:1534:23: warning: symbol 'nfs_exports_op' was not declared. Should it be static? Add include of auth.h fs/nfsd/auth.c:27:5: warning: symbol 'nfsd_setuser' was not declared. Should it be static? Make static, move forward declaration closer to where it's needed. fs/nfsd/nfs4state.c:1877:1: warning: symbol 'laundromat_main' was not declared. Should it be static? Make static, forward declaration was already marked static. fs/nfsd/nfs4idmap.c:206:1: warning: symbol 'idtoname_parse' was not declared. Should it be static? fs/nfsd/vfs.c:1156:1: warning: symbol 'nfsd_create_setattr' was not declared. Should it be static? Signed-off-by: Harvey Harrison Signed-off-by: J. Bruce Fields --- fs/nfsd/auth.c | 1 + fs/nfsd/nfs4idmap.c | 2 +- fs/nfsd/nfs4state.c | 10 +++++----- fs/nfsd/nfsctl.c | 7 ------- 4 files changed, 7 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/auth.c b/fs/nfsd/auth.c index d13403e3362..294992e9bf6 100644 --- a/fs/nfsd/auth.c +++ b/fs/nfsd/auth.c @@ -10,6 +10,7 @@ #include #include #include +#include "auth.h" int nfsexp_flags(struct svc_rqst *rqstp, struct svc_export *exp) { diff --git a/fs/nfsd/nfs4idmap.c b/fs/nfsd/nfs4idmap.c index 996bd88b75b..5b398421b05 100644 --- a/fs/nfsd/nfs4idmap.c +++ b/fs/nfsd/nfs4idmap.c @@ -202,7 +202,7 @@ static struct cache_detail idtoname_cache = { .alloc = ent_alloc, }; -int +static int idtoname_parse(struct cache_detail *cd, char *buf, int buflen) { struct ent ent, *res; diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index a40d1ec52fe..8235d315cd6 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -1763,10 +1763,6 @@ out: return status; } -static struct workqueue_struct *laundry_wq; -static void laundromat_main(struct work_struct *); -static DECLARE_DELAYED_WORK(laundromat_work, laundromat_main); - __be32 nfsd4_renew(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, clientid_t *clid) @@ -1874,7 +1870,11 @@ nfs4_laundromat(void) return clientid_val; } -void +static struct workqueue_struct *laundry_wq; +static void laundromat_main(struct work_struct *); +static DECLARE_DELAYED_WORK(laundromat_work, laundromat_main); + +static void laundromat_main(struct work_struct *not_used) { time_t t; diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index 9f038a4a148..613bcb8171a 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -150,7 +150,6 @@ static const struct file_operations transaction_ops = { .release = simple_transaction_release, }; -extern struct seq_operations nfs_exports_op; static int exports_open(struct inode *inode, struct file *file) { return seq_open(file, &nfs_exports_op); @@ -358,8 +357,6 @@ static ssize_t write_filehandle(struct file *file, char *buf, size_t size) return mesg - buf; } -extern int nfsd_nrthreads(void); - static ssize_t write_threads(struct file *file, char *buf, size_t size) { /* if size > 0, look for a number of threads and call nfsd_svc @@ -382,10 +379,6 @@ static ssize_t write_threads(struct file *file, char *buf, size_t size) return strlen(buf); } -extern int nfsd_nrpools(void); -extern int nfsd_get_nrthreads(int n, int *); -extern int nfsd_set_nrthreads(int n, int *); - static ssize_t write_pool_threads(struct file *file, char *buf, size_t size) { /* if size > 0, look for an array of number of threads per node -- cgit v1.2.3 From 3ba1514815817f93a4f09615726dd4bcd0ddbbc9 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Wed, 20 Feb 2008 12:49:02 -0800 Subject: nfsd: fix sparse warning in vfs.c fs/nfsd/vfs.c:991:27: warning: Using plain integer as NULL pointer Signed-off-by: Harvey Harrison Signed-off-by: J. Bruce Fields --- fs/nfsd/vfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index d5a238e6e6e..832e2b86c54 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -988,7 +988,7 @@ nfsd_vfs_write(struct svc_rqst *rqstp, struct svc_fh *fhp, struct file *file, * flushing the data to disk is handled separately below. */ - if (file->f_op->fsync == 0) {/* COMMIT3 cannot work */ + if (!file->f_op->fsync) {/* COMMIT3 cannot work */ stable = 2; *stablep = 2; /* FILE_SYNC */ } -- cgit v1.2.3 From a277e33cbe3fdfb9a77b448ea3043be22f000dfd Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Wed, 20 Feb 2008 08:55:30 -0500 Subject: NFS: convert nfs4 callback thread to kthread API There's a general push to convert kernel threads to use the (much cleaner) kthread API. This patch converts the NFSv4 callback kernel thread to the kthread API. In addition to being generally cleaner this also removes the dependency on signals when shutting down the thread. Note that this patch depends on the recent patches to svc_recv() to make it check kthread_should_stop() periodically. Those patches are in Bruce's tree at the moment and are slated for 2.6.26 along with the lockd conversion, so this conversion is probably also appropriate for 2.6.26. Signed-off-by: Jeff Layton Acked-by: Trond Myklebust Signed-off-by: J. Bruce Fields --- fs/nfs/callback.c | 73 ++++++++++++++++++++++++++----------------------------- 1 file changed, 35 insertions(+), 38 deletions(-) (limited to 'fs') diff --git a/fs/nfs/callback.c b/fs/nfs/callback.c index 66648dd92d9..2e5de77ff03 100644 --- a/fs/nfs/callback.c +++ b/fs/nfs/callback.c @@ -15,6 +15,7 @@ #include #include #include +#include #include @@ -27,9 +28,7 @@ struct nfs_callback_data { unsigned int users; struct svc_serv *serv; - pid_t pid; - struct completion started; - struct completion stopped; + struct task_struct *task; }; static struct nfs_callback_data nfs_callback_info; @@ -57,27 +56,20 @@ module_param_call(callback_tcpport, param_set_port, param_get_int, /* * This is the callback kernel thread. */ -static void nfs_callback_svc(struct svc_rqst *rqstp) +static int +nfs_callback_svc(void *vrqstp) { int err; + struct svc_rqst *rqstp = vrqstp; - __module_get(THIS_MODULE); - lock_kernel(); - - nfs_callback_info.pid = current->pid; - daemonize("nfsv4-svc"); - /* Process request with signals blocked, but allow SIGKILL. */ - allow_signal(SIGKILL); set_freezable(); - complete(&nfs_callback_info.started); - - for(;;) { - if (signalled()) { - if (nfs_callback_info.users == 0) - break; - flush_signals(current); - } + /* + * FIXME: do we really need to run this under the BKL? If so, please + * add a comment about what it's intended to protect. + */ + lock_kernel(); + while (!kthread_should_stop()) { /* * Listen for a request on the socket */ @@ -92,13 +84,10 @@ static void nfs_callback_svc(struct svc_rqst *rqstp) } svc_process(rqstp); } - - flush_signals(current); - svc_exit_thread(rqstp); - nfs_callback_info.pid = 0; - complete(&nfs_callback_info.stopped); unlock_kernel(); - module_put_and_exit(0); + nfs_callback_info.task = NULL; + svc_exit_thread(rqstp); + return 0; } /* @@ -107,14 +96,13 @@ static void nfs_callback_svc(struct svc_rqst *rqstp) int nfs_callback_up(void) { struct svc_serv *serv = NULL; + struct svc_rqst *rqstp; int ret = 0; lock_kernel(); mutex_lock(&nfs_callback_mutex); - if (nfs_callback_info.users++ || nfs_callback_info.pid != 0) + if (nfs_callback_info.users++ || nfs_callback_info.task != NULL) goto out; - init_completion(&nfs_callback_info.started); - init_completion(&nfs_callback_info.stopped); serv = svc_create(&nfs4_callback_program, NFS4_CALLBACK_BUFSIZE, NULL); ret = -ENOMEM; if (!serv) @@ -127,15 +115,28 @@ int nfs_callback_up(void) nfs_callback_tcpport = ret; dprintk("Callback port = 0x%x\n", nfs_callback_tcpport); - ret = svc_create_thread(nfs_callback_svc, serv); - if (ret < 0) + rqstp = svc_prepare_thread(serv, &serv->sv_pools[0]); + if (IS_ERR(rqstp)) { + ret = PTR_ERR(rqstp); goto out_err; + } + + svc_sock_update_bufs(serv); nfs_callback_info.serv = serv; - wait_for_completion(&nfs_callback_info.started); + + nfs_callback_info.task = kthread_run(nfs_callback_svc, rqstp, + "nfsv4-svc"); + if (IS_ERR(nfs_callback_info.task)) { + ret = PTR_ERR(nfs_callback_info.task); + nfs_callback_info.serv = NULL; + nfs_callback_info.task = NULL; + svc_exit_thread(rqstp); + goto out_err; + } out: /* * svc_create creates the svc_serv with sv_nrthreads == 1, and then - * svc_create_thread increments that. So we need to call svc_destroy + * svc_prepare_thread increments that. So we need to call svc_destroy * on both success and failure so that the refcount is 1 when the * thread exits. */ @@ -159,12 +160,8 @@ void nfs_callback_down(void) lock_kernel(); mutex_lock(&nfs_callback_mutex); nfs_callback_info.users--; - do { - if (nfs_callback_info.users != 0 || nfs_callback_info.pid == 0) - break; - if (kill_proc(nfs_callback_info.pid, SIGKILL, 1) < 0) - break; - } while (wait_for_completion_timeout(&nfs_callback_info.stopped, 5*HZ) == 0); + if (nfs_callback_info.users == 0 && nfs_callback_info.task != NULL) + kthread_stop(nfs_callback_info.task); mutex_unlock(&nfs_callback_mutex); unlock_kernel(); } -- cgit v1.2.3 From 9167f501c6b53492eb2566dd618ce7f55f2856d5 Mon Sep 17 00:00:00 2001 From: Felix Blyakher Date: Tue, 26 Feb 2008 10:54:36 -0800 Subject: nfsd: initialize lease type in nfs4_open_delegation() While lease is correctly checked by supplying the type argument to vfs_setlease(), it's stored with fl_type uninitialized. This breaks the logic when checking the type of the lease. The fix is to initialize fl_type. The old code still happened to function correctly since F_RDLCK is zero, and we only implement read delegations currently (nor write delegations). But that's no excuse for not fixing this. Signed-off-by: Felix Blyakher Signed-off-by: J. Bruce Fields --- fs/nfsd/nfs4state.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index 8235d315cd6..55dfdd71f1b 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -1639,6 +1639,7 @@ nfs4_open_delegation(struct svc_fh *fh, struct nfsd4_open *open, struct nfs4_sta locks_init_lock(&fl); fl.fl_lmops = &nfsd_lease_mng_ops; fl.fl_flags = FL_LEASE; + fl.fl_type = flag == NFS4_OPEN_DELEGATE_READ? F_RDLCK: F_WRLCK; fl.fl_end = OFFSET_MAX; fl.fl_owner = (fl_owner_t)dp; fl.fl_file = stp->st_vfs_file; @@ -1647,8 +1648,7 @@ nfs4_open_delegation(struct svc_fh *fh, struct nfsd4_open *open, struct nfs4_sta /* vfs_setlease checks to see if delegation should be handed out. * the lock_manager callbacks fl_mylease and fl_change are used */ - if ((status = vfs_setlease(stp->st_vfs_file, - flag == NFS4_OPEN_DELEGATE_READ? F_RDLCK: F_WRLCK, &flp))) { + if ((status = vfs_setlease(stp->st_vfs_file, fl.fl_type, &flp))) { dprintk("NFSD: setlease failed [%d], no delegation\n", status); unhash_delegation(dp); flag = NFS4_OPEN_DELEGATE_NONE; -- cgit v1.2.3 From 03550fac06c4f0c39a1885d46015c28794413c82 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Fri, 14 Mar 2008 17:51:12 -0400 Subject: nfsd: move most of fh_verify to separate function Move the code that actually parses the filehandle and looks up the dentry and export to a separate function. This simplifies the reference counting a little and moves fh_verify() a little closer to the kernel ideal of small, minimally-indentended functions. Clean up a few other minor style sins along the way. Signed-off-by: J. Bruce Fields Cc: Neil Brown --- fs/nfsd/nfsfh.c | 228 ++++++++++++++++++++++++++++++-------------------------- 1 file changed, 123 insertions(+), 105 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfsfh.c b/fs/nfsd/nfsfh.c index 3e6b3f41ee1..100ae564116 100644 --- a/fs/nfsd/nfsfh.c +++ b/fs/nfsd/nfsfh.c @@ -112,6 +112,124 @@ static __be32 nfsd_setuser_and_check_port(struct svc_rqst *rqstp, return nfserrno(nfsd_setuser(rqstp, exp)); } +/* + * Use the given filehandle to look up the corresponding export and + * dentry. On success, the results are used to set fh_export and + * fh_dentry. + */ +static __be32 nfsd_set_fh_dentry(struct svc_rqst *rqstp, struct svc_fh *fhp) +{ + struct knfsd_fh *fh = &fhp->fh_handle; + struct fid *fid = NULL, sfid; + struct svc_export *exp; + struct dentry *dentry; + int fileid_type; + int data_left = fh->fh_size/4; + __be32 error; + + error = nfserr_stale; + if (rqstp->rq_vers > 2) + error = nfserr_badhandle; + if (rqstp->rq_vers == 4 && fh->fh_size == 0) + return nfserr_nofilehandle; + + if (fh->fh_version == 1) { + int len; + + if (--data_left < 0) + return error; + if (fh->fh_auth_type != 0) + return error; + len = key_len(fh->fh_fsid_type) / 4; + if (len == 0) + return error; + if (fh->fh_fsid_type == FSID_MAJOR_MINOR) { + /* deprecated, convert to type 3 */ + len = key_len(FSID_ENCODE_DEV)/4; + fh->fh_fsid_type = FSID_ENCODE_DEV; + fh->fh_fsid[0] = new_encode_dev(MKDEV(ntohl(fh->fh_fsid[0]), ntohl(fh->fh_fsid[1]))); + fh->fh_fsid[1] = fh->fh_fsid[2]; + } + data_left -= len; + if (data_left < 0) + return error; + exp = rqst_exp_find(rqstp, fh->fh_fsid_type, fh->fh_auth); + fid = (struct fid *)(fh->fh_auth + len); + } else { + __u32 tfh[2]; + dev_t xdev; + ino_t xino; + + if (fh->fh_size != NFS_FHSIZE) + return error; + /* assume old filehandle format */ + xdev = old_decode_dev(fh->ofh_xdev); + xino = u32_to_ino_t(fh->ofh_xino); + mk_fsid(FSID_DEV, tfh, xdev, xino, 0, NULL); + exp = rqst_exp_find(rqstp, FSID_DEV, tfh); + } + + error = nfserr_stale; + if (PTR_ERR(exp) == -ENOENT) + return error; + + if (IS_ERR(exp)) + return nfserrno(PTR_ERR(exp)); + + error = nfsd_setuser_and_check_port(rqstp, exp); + if (error) + goto out; + + /* + * Look up the dentry using the NFS file handle. + */ + error = nfserr_stale; + if (rqstp->rq_vers > 2) + error = nfserr_badhandle; + + if (fh->fh_version != 1) { + sfid.i32.ino = fh->ofh_ino; + sfid.i32.gen = fh->ofh_generation; + sfid.i32.parent_ino = fh->ofh_dirino; + fid = &sfid; + data_left = 3; + if (fh->ofh_dirino == 0) + fileid_type = FILEID_INO32_GEN; + else + fileid_type = FILEID_INO32_GEN_PARENT; + } else + fileid_type = fh->fh_fileid_type; + + if (fileid_type == FILEID_ROOT) + dentry = dget(exp->ex_path.dentry); + else { + dentry = exportfs_decode_fh(exp->ex_path.mnt, fid, + data_left, fileid_type, + nfsd_acceptable, exp); + } + if (dentry == NULL) + goto out; + if (IS_ERR(dentry)) { + if (PTR_ERR(dentry) != -EINVAL) + error = nfserrno(PTR_ERR(dentry)); + goto out; + } + + if (S_ISDIR(dentry->d_inode->i_mode) && + (dentry->d_flags & DCACHE_DISCONNECTED)) { + printk("nfsd: find_fh_dentry returned a DISCONNECTED directory: %s/%s\n", + dentry->d_parent->d_name.name, dentry->d_name.name); + } + + fhp->fh_dentry = dentry; + fhp->fh_export = exp; + nfsd_nr_verified++; + return 0; +out: + exp_put(exp); + return error; +} + /* * Perform sanity checks on the dentry in a client's file handle. * @@ -124,115 +242,18 @@ static __be32 nfsd_setuser_and_check_port(struct svc_rqst *rqstp, __be32 fh_verify(struct svc_rqst *rqstp, struct svc_fh *fhp, int type, int access) { - struct knfsd_fh *fh = &fhp->fh_handle; - struct svc_export *exp = NULL; + struct svc_export *exp; struct dentry *dentry; - __be32 error = 0; + __be32 error; dprintk("nfsd: fh_verify(%s)\n", SVCFH_fmt(fhp)); if (!fhp->fh_dentry) { - struct fid *fid = NULL, sfid; - int fileid_type; - int data_left = fh->fh_size/4; - - error = nfserr_stale; - if (rqstp->rq_vers > 2) - error = nfserr_badhandle; - if (rqstp->rq_vers == 4 && fh->fh_size == 0) - return nfserr_nofilehandle; - - if (fh->fh_version == 1) { - int len; - if (--data_left<0) goto out; - switch (fh->fh_auth_type) { - case 0: break; - default: goto out; - } - len = key_len(fh->fh_fsid_type) / 4; - if (len == 0) goto out; - if (fh->fh_fsid_type == FSID_MAJOR_MINOR) { - /* deprecated, convert to type 3 */ - len = key_len(FSID_ENCODE_DEV)/4; - fh->fh_fsid_type = FSID_ENCODE_DEV; - fh->fh_fsid[0] = new_encode_dev(MKDEV(ntohl(fh->fh_fsid[0]), ntohl(fh->fh_fsid[1]))); - fh->fh_fsid[1] = fh->fh_fsid[2]; - } - if ((data_left -= len)<0) goto out; - exp = rqst_exp_find(rqstp, fh->fh_fsid_type, - fh->fh_auth); - fid = (struct fid *)(fh->fh_auth + len); - } else { - __u32 tfh[2]; - dev_t xdev; - ino_t xino; - if (fh->fh_size != NFS_FHSIZE) - goto out; - /* assume old filehandle format */ - xdev = old_decode_dev(fh->ofh_xdev); - xino = u32_to_ino_t(fh->ofh_xino); - mk_fsid(FSID_DEV, tfh, xdev, xino, 0, NULL); - exp = rqst_exp_find(rqstp, FSID_DEV, tfh); - } - - error = nfserr_stale; - if (PTR_ERR(exp) == -ENOENT) - goto out; - - if (IS_ERR(exp)) { - error = nfserrno(PTR_ERR(exp)); - goto out; - } - - error = nfsd_setuser_and_check_port(rqstp, exp); + error = nfsd_set_fh_dentry(rqstp, fhp); if (error) goto out; - - /* - * Look up the dentry using the NFS file handle. - */ - error = nfserr_stale; - if (rqstp->rq_vers > 2) - error = nfserr_badhandle; - - if (fh->fh_version != 1) { - sfid.i32.ino = fh->ofh_ino; - sfid.i32.gen = fh->ofh_generation; - sfid.i32.parent_ino = fh->ofh_dirino; - fid = &sfid; - data_left = 3; - if (fh->ofh_dirino == 0) - fileid_type = FILEID_INO32_GEN; - else - fileid_type = FILEID_INO32_GEN_PARENT; - } else - fileid_type = fh->fh_fileid_type; - - if (fileid_type == FILEID_ROOT) - dentry = dget(exp->ex_path.dentry); - else { - dentry = exportfs_decode_fh(exp->ex_path.mnt, fid, - data_left, fileid_type, - nfsd_acceptable, exp); - } - if (dentry == NULL) - goto out; - if (IS_ERR(dentry)) { - if (PTR_ERR(dentry) != -EINVAL) - error = nfserrno(PTR_ERR(dentry)); - goto out; - } - - if (S_ISDIR(dentry->d_inode->i_mode) && - (dentry->d_flags & DCACHE_DISCONNECTED)) { - printk("nfsd: find_fh_dentry returned a DISCONNECTED directory: %s/%s\n", - dentry->d_parent->d_name.name, dentry->d_name.name); - } - - fhp->fh_dentry = dentry; - fhp->fh_export = exp; - nfsd_nr_verified++; - cache_get(&exp->h); + dentry = fhp->fh_dentry; + exp = fhp->fh_export; } else { /* * just rechecking permissions @@ -242,7 +263,6 @@ fh_verify(struct svc_rqst *rqstp, struct svc_fh *fhp, int type, int access) dprintk("nfsd: fh_verify - just checking\n"); dentry = fhp->fh_dentry; exp = fhp->fh_export; - cache_get(&exp->h); /* * Set user creds for this exportpoint; necessary even * in the "just checking" case because this may be a @@ -281,8 +301,6 @@ fh_verify(struct svc_rqst *rqstp, struct svc_fh *fhp, int type, int access) access, ntohl(error)); } out: - if (exp && !IS_ERR(exp)) - exp_put(exp); if (error == nfserr_stale) nfsdstats.fh_stale++; return error; -- cgit v1.2.3 From 6ffd4be6336f9c3f1a10822099587544cd0a11d7 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 27 Mar 2008 16:34:40 -0400 Subject: NFSD: Use "depends on" for PROC_FS dependency Recently, commit 440bcc59 added a reverse dependency to fs/Kconfig to ensure that PROC_FS was enabled if NFSD_V4 was enabled. There is a guideline in Documentation/kbuild/kconfig-language.txt that states "In general use select only for non-visible symbols (no prompts anywhere) and for symbols with no dependencies." A quick grep around other Kconfig files reveals that no entry currently uses "select PROC_FS" -- every one uses "depends on". Thus CONFIG_NFSD_V4 should use "depends on PROC_FS" as well. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields --- fs/Kconfig | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index 1d81be3fef3..9539848b411 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -1695,7 +1695,6 @@ config NFSD select SUNRPC select EXPORTFS select NFS_ACL_SUPPORT if NFSD_V2_ACL - select PROC_FS if NFSD_V4 select PROC_FS if SUNRPC_GSS help Choose Y here if you want to allow other computers to access @@ -1757,7 +1756,7 @@ config NFSD_V3_ACL config NFSD_V4 bool "NFS server support for NFS version 4 (EXPERIMENTAL)" - depends on NFSD && NFSD_V3 && EXPERIMENTAL + depends on NFSD && NFSD_V3 && PROC_FS && EXPERIMENTAL select FS_POSIX_ACL select RPCSEC_GSS_KRB5 help -- cgit v1.2.3 From 3329ba05231808c96cf9fd0598b8f46afec9777c Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 27 Mar 2008 16:34:47 -0400 Subject: SUNRPC: Remove PROC_FS dependency Recently, commit 440bcc59 added a reverse dependency to fs/Kconfig to ensure that PROC_FS was enabled if SUNRPC_GSS was enabled. Apparently this isn't necessary because the auth_gss components under net/sunrpc will build correctly even if PROC_FS is disabled, though RPCSEC_GSS will not work without /proc. It also violates the guideline in Documentation/kbuild/kconfig-language.txt that states "In general use select only for non-visible symbols (no prompts anywhere) and for symbols with no dependencies." To address these issues, remove the dependency. Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields --- fs/Kconfig | 1 - 1 file changed, 1 deletion(-) (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index 9539848b411..157a88583b7 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -1695,7 +1695,6 @@ config NFSD select SUNRPC select EXPORTFS select NFS_ACL_SUPPORT if NFSD_V2_ACL - select PROC_FS if SUNRPC_GSS help Choose Y here if you want to allow other computers to access files residing on this system using Sun's Network File System -- cgit v1.2.3 From 1a448fdb3c5495405bc44d77ea676150f9195444 Mon Sep 17 00:00:00 2001 From: Chuck Lever Date: Thu, 27 Mar 2008 16:34:54 -0400 Subject: NFSD: Remove NFSv4 dependency on NFSv3 Clean up: Because NFSD_V4 "depends on" NFSD_V3, it appears as a child of the NFSD_V3 menu entry, and is not visible if NFSD_V3 is unselected. Replace the dependency on NFSD_V3 with a "select NFSD_V3". This makes NFSD_V4 look and work just like NFS_V3, while ensuring that NFSD_V3 is enabled if NFSD_V4 is. Sam Ravnborg adds: "This use of select is questionable. In general it is bad to select a symbol with dependencies. In this case the dependencies of NFSD_V3 are duplicated for NFSD_V4 so we will not se erratic configurations but do you remember to update NFSD_V4 when you add a depends on NFSD_V3? But I see no other clean way to do it right now." Later he said: "My comment was more to say we have things to address in kconfig. This is abuse in the acceptable range." Signed-off-by: Chuck Lever Signed-off-by: J. Bruce Fields --- fs/Kconfig | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/Kconfig b/fs/Kconfig index 157a88583b7..56c83f40cdb 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -1755,7 +1755,8 @@ config NFSD_V3_ACL config NFSD_V4 bool "NFS server support for NFS version 4 (EXPERIMENTAL)" - depends on NFSD && NFSD_V3 && PROC_FS && EXPERIMENTAL + depends on NFSD && PROC_FS && EXPERIMENTAL + select NFSD_V3 select FS_POSIX_ACL select RPCSEC_GSS_KRB5 help -- cgit v1.2.3 From 3c61eecb607dbc2777074b1a95b8a97e31a96a73 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Mon, 7 Apr 2008 13:05:27 -0400 Subject: lockd: Fix stale nlmsvc_unlink_block comment As of 5996a298da43a03081e9ba2116983d173001c862 ("NLM: don't unlock on cancel requests") we no longer unlock in this case, so the comment is no longer accurate. Thanks to Stuart Friedberg for pointing out the inconsistency. Cc: Stuart Friedberg Signed-off-by: J. Bruce Fields --- fs/lockd/svclock.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/lockd/svclock.c b/fs/lockd/svclock.c index 4da7c4c2706..1f122c1940a 100644 --- a/fs/lockd/svclock.c +++ b/fs/lockd/svclock.c @@ -227,8 +227,7 @@ failed: } /* - * Delete a block. If the lock was cancelled or the grant callback - * failed, unlock is set to 1. + * Delete a block. * It is the caller's responsibility to check whether the file * can be closed hereafter. */ -- cgit v1.2.3 From e1ba1ab76e68de9f4a93fae8406627924efaed99 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Mon, 7 Apr 2008 13:09:47 -0400 Subject: nfsd: fix comment Obvious comment nit. Signed-off-by: J. Bruce Fields --- fs/nfs/callback.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/nfs/callback.c b/fs/nfs/callback.c index 2e5de77ff03..a9f15386755 100644 --- a/fs/nfs/callback.c +++ b/fs/nfs/callback.c @@ -153,7 +153,7 @@ out_err: } /* - * Kill the server process if it is not already up. + * Kill the server process if it is not already down. */ void nfs_callback_down(void) { -- cgit v1.2.3 From ff7d9756b501744540be65e172d27ee321d86103 Mon Sep 17 00:00:00 2001 From: Olga Kornievskaia Date: Fri, 28 Mar 2008 16:04:56 -0400 Subject: nfsd: use static memory for callback program and stats There's no need to dynamically allocate this memory, and doing so may create the possibility of races on shutdown of the rpc client. (We've witnessed it only after adding rpcsec_gss support to the server, after which the rpc code can send destroys calls that expect to still be able to access the rpc_stats structure after it has been destroyed.) Such races are in theory possible if the module containing this "static" memory is removed very quickly after an rpc client is destroyed, but we haven't seen that happen. Signed-off-by: J. Bruce Fields --- fs/nfsd/nfs4callback.c | 28 +++++++++++++++++----------- 1 file changed, 17 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c index aae2b29ae2c..562abf3380d 100644 --- a/fs/nfsd/nfs4callback.c +++ b/fs/nfsd/nfs4callback.c @@ -344,6 +344,21 @@ static struct rpc_version * nfs_cb_version[] = { &nfs_cb_version4, }; +static struct rpc_program cb_program; + +static struct rpc_stat cb_stats = { + .program = &cb_program +}; + +#define NFS4_CALLBACK 0x40000000 +static struct rpc_program cb_program = { + .name = "nfs4_cb", + .number = NFS4_CALLBACK, + .nrvers = ARRAY_SIZE(nfs_cb_version), + .version = nfs_cb_version, + .stats = &cb_stats, +}; + /* Reference counting, callback cleanup, etc., all look racy as heck. * And why is cb_set an atomic? */ @@ -358,13 +373,12 @@ static int do_probe_callback(void *data) .to_maxval = (NFSD_LEASE_TIME/2) * HZ, .to_exponential = 1, }; - struct rpc_program * program = &cb->cb_program; struct rpc_create_args args = { .protocol = IPPROTO_TCP, .address = (struct sockaddr *)&addr, .addrsize = sizeof(addr), .timeout = &timeparms, - .program = program, + .program = &cb_program, .version = nfs_cb_version[1]->number, .authflavor = RPC_AUTH_UNIX, /* XXX: need AUTH_GSS... */ .flags = (RPC_CLNT_CREATE_NOPING), @@ -382,16 +396,8 @@ static int do_probe_callback(void *data) addr.sin_port = htons(cb->cb_port); addr.sin_addr.s_addr = htonl(cb->cb_addr); - /* Initialize rpc_program */ - program->name = "nfs4_cb"; - program->number = cb->cb_prog; - program->nrvers = ARRAY_SIZE(nfs_cb_version); - program->version = nfs_cb_version; - program->stats = &cb->cb_stat; - /* Initialize rpc_stat */ - memset(program->stats, 0, sizeof(cb->cb_stat)); - program->stats->program = program; + memset(args.program->stats, 0, sizeof(struct rpc_stat)); /* Create RPC client */ client = rpc_create(&args); -- cgit v1.2.3 From 06e02d66fa0055230efc2443c43ee4f3ab5eb0b6 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Tue, 8 Apr 2008 15:40:07 -0400 Subject: NFS: don't let nfs_callback_svc exit on unexpected svc_recv errors (try #2) When svc_recv returns an unexpected error, nfs_callback_svc will print a warning and exit. This problematic for several reasons. In particular, it will cause the reference counts for the thread to be wrong, and no new thread will be started until all nfs4 mounts are unmounted. Rather than exiting on error from svc_recv, have the thread do a 1s sleep and then retry the loop. This is unlikely to cause any harm, and if the error turns out to be something temporary then it may be able to recover. Signed-off-by: Jeff Layton Signed-off-by: J. Bruce Fields --- fs/nfs/callback.c | 18 ++++++++++++------ 1 file changed, 12 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/nfs/callback.c b/fs/nfs/callback.c index a9f15386755..5606ae3d72d 100644 --- a/fs/nfs/callback.c +++ b/fs/nfs/callback.c @@ -59,7 +59,7 @@ module_param_call(callback_tcpport, param_set_port, param_get_int, static int nfs_callback_svc(void *vrqstp) { - int err; + int err, preverr = 0; struct svc_rqst *rqstp = vrqstp; set_freezable(); @@ -74,14 +74,20 @@ nfs_callback_svc(void *vrqstp) * Listen for a request on the socket */ err = svc_recv(rqstp, MAX_SCHEDULE_TIMEOUT); - if (err == -EAGAIN || err == -EINTR) + if (err == -EAGAIN || err == -EINTR) { + preverr = err; continue; + } if (err < 0) { - printk(KERN_WARNING - "%s: terminating on error %d\n", - __FUNCTION__, -err); - break; + if (err != preverr) { + printk(KERN_WARNING "%s: unexpected error " + "from svc_recv (%d)\n", __func__, err); + preverr = err; + } + schedule_timeout_uninterruptible(HZ); + continue; } + preverr = err; svc_process(rqstp); } unlock_kernel(); -- cgit v1.2.3 From f97c650dda24e48405399aa0676e90da52408515 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Tue, 8 Apr 2008 15:40:08 -0400 Subject: NLM: don't let lockd exit on unexpected svc_recv errors (try #2) When svc_recv returns an unexpected error, lockd will print a warning and exit. This problematic for several reasons. In particular, it will cause the reference counts for the thread to be wrong, and can lead to a potential BUG() call. Rather than exiting on error from svc_recv, have the thread do a 1s sleep and then retry the loop. This is unlikely to cause any harm, and if the error turns out to be something temporary then it may be able to recover. Signed-off-by: Jeff Layton Signed-off-by: J. Bruce Fields --- fs/lockd/svc.c | 18 ++++++++++++------ 1 file changed, 12 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/lockd/svc.c b/fs/lockd/svc.c index 66b5c98c7ff..cf977bbcf30 100644 --- a/fs/lockd/svc.c +++ b/fs/lockd/svc.c @@ -112,7 +112,7 @@ static inline void clear_grace_period(void) static int lockd(void *vrqstp) { - int err = 0; + int err = 0, preverr = 0; struct svc_rqst *rqstp = vrqstp; unsigned long grace_period_expire; @@ -172,14 +172,20 @@ lockd(void *vrqstp) * recvfrom routine. */ err = svc_recv(rqstp, timeout); - if (err == -EAGAIN || err == -EINTR) + if (err == -EAGAIN || err == -EINTR) { + preverr = err; continue; + } if (err < 0) { - printk(KERN_WARNING - "lockd: terminating on error %d\n", - -err); - break; + if (err != preverr) { + printk(KERN_WARNING "%s: unexpected error " + "from svc_recv (%d)\n", __func__, err); + preverr = err; + } + schedule_timeout_interruptible(HZ); + continue; } + preverr = err; dprintk("lockd: request from %s\n", svc_print_addr(rqstp, buf, sizeof(buf))); -- cgit v1.2.3 From dee3209d993f17081d2c58d6470dfc8d6662078b Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Wed, 16 Apr 2008 16:28:46 -0400 Subject: knfsd: get rid of imode variable in nfsd_setattr ...it's not really needed. Signed-off-by: Jeff Layton Signed-off-by: J. Bruce Fields --- fs/nfsd/vfs.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index 832e2b86c54..1d0406c31a4 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -264,7 +264,6 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp, struct iattr *iap, struct inode *inode; int accmode = MAY_SATTR; int ftype = 0; - int imode; __be32 err; int host_err; int size_change = 0; @@ -360,10 +359,9 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp, struct iattr *iap, DQUOT_INIT(inode); } - imode = inode->i_mode; if (iap->ia_valid & ATTR_MODE) { iap->ia_mode &= S_IALLUGO; - imode = iap->ia_mode |= (imode & ~S_IALLUGO); + iap->ia_mode |= (inode->i_mode & ~S_IALLUGO); /* if changing uid/gid revoke setuid/setgid in mode */ if ((iap->ia_valid & ATTR_UID) && iap->ia_uid != inode->i_uid) { iap->ia_valid |= ATTR_KILL_PRIV; -- cgit v1.2.3 From ca456252db0521e5e88024fa2b67535e9739e030 Mon Sep 17 00:00:00 2001 From: Jeff Layton Date: Wed, 16 Apr 2008 16:28:47 -0400 Subject: knfsd: clear both setuid and setgid whenever a chown is done Currently, knfsd only clears the setuid bit if the owner of a file is changed on a SETATTR call, and only clears the setgid bit if the group is changed. POSIX says this in the spec for chown(): "If the specified file is a regular file, one or more of the S_IXUSR, S_IXGRP, or S_IXOTH bits of the file mode are set, and the process does not have appropriate privileges, the set-user-ID (S_ISUID) and set-group-ID (S_ISGID) bits of the file mode shall be cleared upon successful return from chown()." If I'm reading this correctly, then knfsd is doing this wrong. It should be clearing both the setuid and setgid bit on any SETATTR that changes the uid or gid. This wasn't really as noticable before, but now that the ATTR_KILL_S*ID bits are a no-op for the NFS client, it's more evident. This patch corrects the nfsd_setattr logic so that this occurs. It also does a bit of cleanup to the function. There is also one small behavioral change. If a SETATTR call comes in that changes the uid/gid and the mode, then we now only clear the setgid bit if the group execute bit isn't set. The setgid bit without a group execute bit signifies mandatory locking and we likely don't want to clear the bit in that case. Since there is no call in POSIX that should generate a SETATTR call like this, then this should rarely happen, but it's worth noting. Signed-off-by: Jeff Layton Signed-off-by: J. Bruce Fields --- fs/nfsd/vfs.c | 27 ++++++++++++++------------- 1 file changed, 14 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index 1d0406c31a4..a3a291f771f 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -359,24 +359,25 @@ nfsd_setattr(struct svc_rqst *rqstp, struct svc_fh *fhp, struct iattr *iap, DQUOT_INIT(inode); } + /* sanitize the mode change */ if (iap->ia_valid & ATTR_MODE) { iap->ia_mode &= S_IALLUGO; iap->ia_mode |= (inode->i_mode & ~S_IALLUGO); - /* if changing uid/gid revoke setuid/setgid in mode */ - if ((iap->ia_valid & ATTR_UID) && iap->ia_uid != inode->i_uid) { - iap->ia_valid |= ATTR_KILL_PRIV; + } + + /* Revoke setuid/setgid on chown */ + if (((iap->ia_valid & ATTR_UID) && iap->ia_uid != inode->i_uid) || + ((iap->ia_valid & ATTR_GID) && iap->ia_gid != inode->i_gid)) { + iap->ia_valid |= ATTR_KILL_PRIV; + if (iap->ia_valid & ATTR_MODE) { + /* we're setting mode too, just clear the s*id bits */ iap->ia_mode &= ~S_ISUID; + if (iap->ia_mode & S_IXGRP) + iap->ia_mode &= ~S_ISGID; + } else { + /* set ATTR_KILL_* bits and let VFS handle it */ + iap->ia_valid |= (ATTR_KILL_SUID | ATTR_KILL_SGID); } - if ((iap->ia_valid & ATTR_GID) && iap->ia_gid != inode->i_gid) - iap->ia_mode &= ~S_ISGID; - } else { - /* - * Revoke setuid/setgid bit on chown/chgrp - */ - if ((iap->ia_valid & ATTR_UID) && iap->ia_uid != inode->i_uid) - iap->ia_valid |= ATTR_KILL_SUID | ATTR_KILL_PRIV; - if ((iap->ia_valid & ATTR_GID) && iap->ia_gid != inode->i_gid) - iap->ia_valid |= ATTR_KILL_SGID; } /* Change the attributes. */ -- cgit v1.2.3 From ee4987ab5cc9d00be38cfeec90174229565211be Mon Sep 17 00:00:00 2001 From: Steve French Date: Thu, 24 Apr 2008 16:31:12 +0000 Subject: [CIFS] Fix define for new proxy cap to match documentation The transport encryption capability and new SetFSInfo level were missing, and the new proxy capability (which Samba server is implementing) and proxy setfsinfo needed to be moved down to not collide with Samba's transport encryption capability. CC: Jeremy Allison CC: Sam Liddicott Signed-off-by: Steve French --- fs/cifs/cifspdu.h | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifspdu.h b/fs/cifs/cifspdu.h index b18c6d43bef..3b889bc3e8d 100644 --- a/fs/cifs/cifspdu.h +++ b/fs/cifs/cifspdu.h @@ -1787,7 +1787,8 @@ typedef struct smb_com_transaction2_fnext_rsp_parms { #define SMB_QUERY_CIFS_UNIX_INFO 0x200 #define SMB_QUERY_POSIX_FS_INFO 0x201 #define SMB_QUERY_POSIX_WHO_AM_I 0x202 -#define SMB_QUERY_FS_PROXY 0x203 /* WAFS enabled. Returns structure +#define SMB_REQUEST_TRANSPORT_ENCRYPTION 0x203 +#define SMB_QUERY_FS_PROXY 0x204 /* WAFS enabled. Returns structure FILE_SYSTEM__UNIX_INFO to tell whether new NTIOCTL available (0xACE) for WAN friendly SMB @@ -2048,7 +2049,9 @@ typedef struct { #define CIFS_UNIX_LARGE_READ_CAP 0x00000040 /* support reads >128K (up to 0xFFFF00 */ #define CIFS_UNIX_LARGE_WRITE_CAP 0x00000080 -#define CIFS_UNIX_PROXY_CAP 0x00000100 /* Proxy cap: 0xACE ioctl and +#define CIFS_UNIX_TRANSPORT_ENCRYPTION_ 0x00000100 /* can do SPNEGO encrypt */ +#define CIFS_UNIX_TRANPSORT_ENCRYPTION 0x00000200 /* must do SPNEGO encrypt */ +#define CIFS_UNIX_PROXY_CAP 0x00000400 /* Proxy cap: 0xACE ioctl and QFS PROXY call */ #ifdef CONFIG_CIFS_POSIX /* Can not set pathnames cap yet until we send new posix create SMB since -- cgit v1.2.3 From a7f796a60bb5dcdc154c3ae04d484a395a60809f Mon Sep 17 00:00:00 2001 From: Steve French Date: Thu, 24 Apr 2008 16:39:07 +0000 Subject: [CIFS] Fix typo in previous commit Signed-off-by: Steve French --- fs/cifs/cifspdu.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifspdu.h b/fs/cifs/cifspdu.h index 3b889bc3e8d..9f49c2f3582 100644 --- a/fs/cifs/cifspdu.h +++ b/fs/cifs/cifspdu.h @@ -2049,8 +2049,8 @@ typedef struct { #define CIFS_UNIX_LARGE_READ_CAP 0x00000040 /* support reads >128K (up to 0xFFFF00 */ #define CIFS_UNIX_LARGE_WRITE_CAP 0x00000080 -#define CIFS_UNIX_TRANSPORT_ENCRYPTION_ 0x00000100 /* can do SPNEGO encrypt */ -#define CIFS_UNIX_TRANPSORT_ENCRYPTION 0x00000200 /* must do SPNEGO encrypt */ +#define CIFS_UNIX_TRANSPORT_ENCRYPTION_CAP 0x00000100 /* can do SPNEGO crypt */ +#define CIFS_UNIX_TRANPSORT_ENCRYPTION_MANDATORY_CAP 0x00000200 /* must do */ #define CIFS_UNIX_PROXY_CAP 0x00000400 /* Proxy cap: 0xACE ioctl and QFS PROXY call */ #ifdef CONFIG_CIFS_POSIX -- cgit v1.2.3 From 78d31a3a87f84cf56004b7bc154831f2ee1186e8 Mon Sep 17 00:00:00 2001 From: Igor Mammedov Date: Thu, 24 Apr 2008 12:56:07 +0400 Subject: cifs: timeout dfs automounts +little fix. Signed-off-by: Christoph Hellwig Signed-off-by: Al Viro --- fs/cifs/cifs_dfs_ref.c | 29 +++++++++++++++++++++-------- fs/cifs/cifsfs.c | 3 +-- fs/cifs/cifsproto.h | 8 +------- 3 files changed, 23 insertions(+), 17 deletions(-) (limited to 'fs') diff --git a/fs/cifs/cifs_dfs_ref.c b/fs/cifs/cifs_dfs_ref.c index f53f41ff166..95024c066d8 100644 --- a/fs/cifs/cifs_dfs_ref.c +++ b/fs/cifs/cifs_dfs_ref.c @@ -25,14 +25,26 @@ static LIST_HEAD(cifs_dfs_automount_list); -/* - * DFS functions -*/ +static void cifs_dfs_expire_automounts(struct work_struct *work); +static DECLARE_DELAYED_WORK(cifs_dfs_automount_task, + cifs_dfs_expire_automounts); +static int cifs_dfs_mountpoint_expiry_timeout = 500 * HZ; + +static void cifs_dfs_expire_automounts(struct work_struct *work) +{ + struct list_head *list = &cifs_dfs_automount_list; + + mark_mounts_for_expiry(list); + if (!list_empty(list)) + schedule_delayed_work(&cifs_dfs_automount_task, + cifs_dfs_mountpoint_expiry_timeout); +} -void dfs_shrink_umount_helper(struct vfsmount *vfsmnt) +void cifs_dfs_release_automount_timer(void) { - mark_mounts_for_expiry(&cifs_dfs_automount_list); - mark_mounts_for_expiry(&cifs_dfs_automount_list); + BUG_ON(!list_empty(&cifs_dfs_automount_list)); + cancel_delayed_work(&cifs_dfs_automount_task); + flush_scheduled_work(); } /** @@ -261,10 +273,11 @@ static int add_mount_helper(struct vfsmount *newmnt, struct nameidata *nd, err = do_add_mount(newmnt, nd, nd->path.mnt->mnt_flags, mntlist); switch (err) { case 0: - dput(nd->path.dentry); - mntput(nd->path.mnt); + path_put(&nd->path); nd->path.mnt = newmnt; nd->path.dentry = dget(newmnt->mnt_root); + schedule_delayed_work(&cifs_dfs_automount_task, + cifs_dfs_mountpoint_expiry_timeout); break; case -EBUSY: /* someone else made a mount here whilst we were busy */ diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index a04b17e5a9d..dbb2cd678bf 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -471,8 +471,6 @@ static void cifs_umount_begin(struct vfsmount *vfsmnt, int flags) struct cifs_sb_info *cifs_sb; struct cifsTconInfo *tcon; - dfs_shrink_umount_helper(vfsmnt); - if (!(flags & MNT_FORCE)) return; cifs_sb = CIFS_SB(vfsmnt->mnt_sb); @@ -1100,6 +1098,7 @@ exit_cifs(void) cFYI(DBG2, ("exit_cifs")); cifs_proc_clean(); #ifdef CONFIG_CIFS_DFS_UPCALL + cifs_dfs_release_automount_timer(); unregister_key_type(&key_type_dns_resolver); #endif #ifdef CONFIG_CIFS_UPCALL diff --git a/fs/cifs/cifsproto.h b/fs/cifs/cifsproto.h index 0c83da4a7da..50f9fdae19b 100644 --- a/fs/cifs/cifsproto.h +++ b/fs/cifs/cifsproto.h @@ -104,13 +104,7 @@ extern int mode_to_acl(struct inode *inode, const char *path, __u64); extern int cifs_mount(struct super_block *, struct cifs_sb_info *, char *, const char *); extern int cifs_umount(struct super_block *, struct cifs_sb_info *); -#ifdef CONFIG_CIFS_DFS_UPCALL -extern void dfs_shrink_umount_helper(struct vfsmount *vfsmnt); -#else -static inline void dfs_shrink_umount_helper(struct vfsmount *vfsmnt) -{ -} -#endif /* DFS_UPCALL */ +extern void cifs_dfs_release_automount_timer(void); void cifs_proc_init(void); void cifs_proc_clean(void); -- cgit v1.2.3 From 42faad99658eed7ca8bd328ffa4bcb7d78c9bcca Mon Sep 17 00:00:00 2001 From: Al Viro Date: Thu, 24 Apr 2008 07:21:56 -0400 Subject: [PATCH] restore sane ->umount_begin() API Signed-off-by: Al Viro --- fs/9p/vfs_super.c | 7 +++---- fs/cifs/cifsfs.c | 7 ++----- fs/fuse/inode.c | 5 ++--- fs/namespace.c | 9 +++++---- fs/nfs/super.c | 8 +++----- 5 files changed, 15 insertions(+), 21 deletions(-) (limited to 'fs') diff --git a/fs/9p/vfs_super.c b/fs/9p/vfs_super.c index 678c02f1ae2..a452ac67fc9 100644 --- a/fs/9p/vfs_super.c +++ b/fs/9p/vfs_super.c @@ -224,12 +224,11 @@ static int v9fs_show_options(struct seq_file *m, struct vfsmount *mnt) } static void -v9fs_umount_begin(struct vfsmount *vfsmnt, int flags) +v9fs_umount_begin(struct super_block *sb) { - struct v9fs_session_info *v9ses = vfsmnt->mnt_sb->s_fs_info; + struct v9fs_session_info *v9ses = sb->s_fs_info; - if (flags & MNT_FORCE) - v9fs_session_cancel(v9ses); + v9fs_session_cancel(v9ses); } static const struct super_operations v9fs_super_ops = { diff --git a/fs/cifs/cifsfs.c b/fs/cifs/cifsfs.c index dbb2cd678bf..39c2cbdface 100644 --- a/fs/cifs/cifsfs.c +++ b/fs/cifs/cifsfs.c @@ -466,14 +466,11 @@ static struct quotactl_ops cifs_quotactl_ops = { }; #endif -static void cifs_umount_begin(struct vfsmount *vfsmnt, int flags) +static void cifs_umount_begin(struct super_block *sb) { - struct cifs_sb_info *cifs_sb; + struct cifs_sb_info *cifs_sb = CIFS_SB(sb); struct cifsTconInfo *tcon; - if (!(flags & MNT_FORCE)) - return; - cifs_sb = CIFS_SB(vfsmnt->mnt_sb); if (cifs_sb == NULL) return; diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c index 033f7bdd47e..4df34da2284 100644 --- a/fs/fuse/inode.c +++ b/fs/fuse/inode.c @@ -242,10 +242,9 @@ struct inode *fuse_iget(struct super_block *sb, unsigned long nodeid, return inode; } -static void fuse_umount_begin(struct vfsmount *vfsmnt, int flags) +static void fuse_umount_begin(struct super_block *sb) { - if (flags & MNT_FORCE) - fuse_abort_conn(get_fuse_conn_super(vfsmnt->mnt_sb)); + fuse_abort_conn(get_fuse_conn_super(sb)); } static void fuse_send_destroy(struct fuse_conn *fc) diff --git a/fs/namespace.c b/fs/namespace.c index 0505fb61aa7..f48f98110c3 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -1061,10 +1061,11 @@ static int do_umount(struct vfsmount *mnt, int flags) * about for the moment. */ - lock_kernel(); - if (sb->s_op->umount_begin) - sb->s_op->umount_begin(mnt, flags); - unlock_kernel(); + if (flags & MNT_FORCE && sb->s_op->umount_begin) { + lock_kernel(); + sb->s_op->umount_begin(sb); + unlock_kernel(); + } /* * No sense to grab the lock for this test, but test itself looks diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 20a1cb1810f..fa220dc7460 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -198,7 +198,7 @@ static match_table_t nfs_secflavor_tokens = { }; -static void nfs_umount_begin(struct vfsmount *, int); +static void nfs_umount_begin(struct super_block *); static int nfs_statfs(struct dentry *, struct kstatfs *); static int nfs_show_options(struct seq_file *, struct vfsmount *); static int nfs_show_stats(struct seq_file *, struct vfsmount *); @@ -647,13 +647,11 @@ static int nfs_show_stats(struct seq_file *m, struct vfsmount *mnt) * Begin unmount by attempting to remove all automounted mountpoints we added * in response to xdev traversals and referrals */ -static void nfs_umount_begin(struct vfsmount *vfsmnt, int flags) +static void nfs_umount_begin(struct super_block *sb) { - struct nfs_server *server = NFS_SB(vfsmnt->mnt_sb); + struct nfs_server *server = NFS_SB(sb); struct rpc_clnt *rpc; - if (!(flags & MNT_FORCE)) - return; /* -EIO all pending I/O */ rpc = server->client_acl; if (!IS_ERR(rpc)) -- cgit v1.2.3 From fd8328be874f4190a811c58cd4778ec2c74d2c05 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 22 Apr 2008 05:11:59 -0400 Subject: [PATCH] sanitize handling of shared descriptor tables in failing execve() * unshare_files() can fail; doing it after irreversible actions is wrong and de_thread() is certainly irreversible. * since we do it unconditionally anyway, we might as well do it in do_execve() and save ourselves the PITA in binfmt handlers, etc. * while we are at it, binfmt_som actually leaked files_struct on failure. As a side benefit, unshare_files(), put_files_struct() and reset_files_struct() become unexported. Signed-off-by: Al Viro --- fs/binfmt_elf.c | 23 +---------------------- fs/binfmt_misc.c | 18 +----------------- fs/binfmt_som.c | 10 ---------- fs/exec.c | 34 ++++++++++++++++++---------------- 4 files changed, 20 insertions(+), 65 deletions(-) (limited to 'fs') diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c index 5e1a4fb5cac..9924581df6f 100644 --- a/fs/binfmt_elf.c +++ b/fs/binfmt_elf.c @@ -543,7 +543,6 @@ static int load_elf_binary(struct linux_binprm *bprm, struct pt_regs *regs) unsigned long interp_load_addr = 0; unsigned long start_code, end_code, start_data, end_data; unsigned long reloc_func_desc = 0; - struct files_struct *files; int executable_stack = EXSTACK_DEFAULT; unsigned long def_flags = 0; struct { @@ -593,20 +592,9 @@ static int load_elf_binary(struct linux_binprm *bprm, struct pt_regs *regs) goto out_free_ph; } - files = current->files; /* Refcounted so ok */ - retval = unshare_files(); - if (retval < 0) - goto out_free_ph; - if (files == current->files) { - put_files_struct(files); - files = NULL; - } - - /* exec will make our files private anyway, but for the a.out - loader stuff we need to do it earlier */ retval = get_unused_fd(); if (retval < 0) - goto out_free_fh; + goto out_free_ph; get_file(bprm->file); fd_install(elf_exec_fileno = retval, bprm->file); @@ -728,12 +716,6 @@ static int load_elf_binary(struct linux_binprm *bprm, struct pt_regs *regs) if (retval) goto out_free_dentry; - /* Discard our unneeded old files struct */ - if (files) { - put_files_struct(files); - files = NULL; - } - /* OK, This is the point of no return */ current->flags &= ~PF_FORKNOEXEC; current->mm->def_flags = def_flags; @@ -1016,9 +998,6 @@ out_free_interp: kfree(elf_interpreter); out_free_file: sys_close(elf_exec_fileno); -out_free_fh: - if (files) - reset_files_struct(current, files); out_free_ph: kfree(elf_phdata); goto out; diff --git a/fs/binfmt_misc.c b/fs/binfmt_misc.c index b53c7e5f41b..dbf0ac0523d 100644 --- a/fs/binfmt_misc.c +++ b/fs/binfmt_misc.c @@ -110,7 +110,6 @@ static int load_misc_binary(struct linux_binprm *bprm, struct pt_regs *regs) char *iname_addr = iname; int retval; int fd_binary = -1; - struct files_struct *files = NULL; retval = -ENOEXEC; if (!enabled) @@ -133,21 +132,13 @@ static int load_misc_binary(struct linux_binprm *bprm, struct pt_regs *regs) if (fmt->flags & MISC_FMT_OPEN_BINARY) { - files = current->files; - retval = unshare_files(); - if (retval < 0) - goto _ret; - if (files == current->files) { - put_files_struct(files); - files = NULL; - } /* if the binary should be opened on behalf of the * interpreter than keep it open and assign descriptor * to it */ fd_binary = get_unused_fd(); if (fd_binary < 0) { retval = fd_binary; - goto _unshare; + goto _ret; } fd_install(fd_binary, bprm->file); @@ -205,10 +196,6 @@ static int load_misc_binary(struct linux_binprm *bprm, struct pt_regs *regs) if (retval < 0) goto _error; - if (files) { - put_files_struct(files); - files = NULL; - } _ret: return retval; _error: @@ -216,9 +203,6 @@ _error: sys_close(fd_binary); bprm->interp_flags = 0; bprm->interp_data = 0; -_unshare: - if (files) - reset_files_struct(current, files); goto _ret; } diff --git a/fs/binfmt_som.c b/fs/binfmt_som.c index 14c63527c76..fdc36bfd6a7 100644 --- a/fs/binfmt_som.c +++ b/fs/binfmt_som.c @@ -194,7 +194,6 @@ load_som_binary(struct linux_binprm * bprm, struct pt_regs * regs) unsigned long som_entry; struct som_hdr *som_ex; struct som_exec_auxhdr *hpuxhdr; - struct files_struct *files; /* Get the exec-header */ som_ex = (struct som_hdr *) bprm->buf; @@ -221,15 +220,6 @@ load_som_binary(struct linux_binprm * bprm, struct pt_regs * regs) goto out_free; } - files = current->files; /* Refcounted so ok */ - retval = unshare_files(); - if (retval < 0) - goto out_free; - if (files == current->files) { - put_files_struct(files); - files = NULL; - } - retval = get_unused_fd(); if (retval < 0) goto out_free; diff --git a/fs/exec.c b/fs/exec.c index 54a0a557b67..475543002f1 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -953,7 +953,6 @@ int flush_old_exec(struct linux_binprm * bprm) { char * name; int i, ch, retval; - struct files_struct *files; char tcomm[sizeof(current->comm)]; /* @@ -964,27 +963,16 @@ int flush_old_exec(struct linux_binprm * bprm) if (retval) goto out; - /* - * Make sure we have private file handles. Ask the - * fork helper to do the work for us and the exit - * helper to do the cleanup of the old one. - */ - files = current->files; /* refcounted so safe to hold */ - retval = unshare_files(); - if (retval) - goto out; /* * Release all of the old mmap stuff */ retval = exec_mmap(bprm->mm); if (retval) - goto mmap_failed; + goto out; bprm->mm = NULL; /* We're using it now */ /* This is the point of no return */ - put_files_struct(files); - current->sas_ss_sp = current->sas_ss_size = 0; if (current->euid == current->uid && current->egid == current->gid) @@ -1034,8 +1022,6 @@ int flush_old_exec(struct linux_binprm * bprm) return 0; -mmap_failed: - reset_files_struct(current, files); out: return retval; } @@ -1283,12 +1269,23 @@ int do_execve(char * filename, struct linux_binprm *bprm; struct file *file; unsigned long env_p; + struct files_struct *files; int retval; + files = current->files; + retval = unshare_files(); + if (retval) + goto out_ret; + + if (files == current->files) { + put_files_struct(files); + files = NULL; + } + retval = -ENOMEM; bprm = kzalloc(sizeof(*bprm), GFP_KERNEL); if (!bprm) - goto out_ret; + goto out_files; file = open_exec(filename); retval = PTR_ERR(file); @@ -1343,6 +1340,8 @@ int do_execve(char * filename, security_bprm_free(bprm); acct_update_integrals(current); kfree(bprm); + if (files) + put_files_struct(files); return retval; } @@ -1363,6 +1362,9 @@ out_file: out_kfree: kfree(bprm); +out_files: + if (files) + reset_files_struct(current, files); out_ret: return retval; } -- cgit v1.2.3 From 3b1253880b7a9e6db54b943b2d40bcf2202f58ab Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 22 Apr 2008 05:31:30 -0400 Subject: [PATCH] sanitize unshare_files/reset_files_struct * let unshare_files() give caller the displaced files_struct * don't bother with grabbing reference only to drop it in the caller if it hadn't been shared in the first place * in that form unshare_files() is trivially implemented via unshare_fd(), so we eliminate the duplicate logics in fork.c * reset_files_struct() is not just only called for current; it will break the system if somebody ever calls it for anything else (we can't modify ->files of somebody else). Lose the task_struct * argument. Signed-off-by: Al Viro --- fs/exec.c | 18 ++++++------------ 1 file changed, 6 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/exec.c b/fs/exec.c index 475543002f1..b152029f18f 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -1269,19 +1269,13 @@ int do_execve(char * filename, struct linux_binprm *bprm; struct file *file; unsigned long env_p; - struct files_struct *files; + struct files_struct *displaced; int retval; - files = current->files; - retval = unshare_files(); + retval = unshare_files(&displaced); if (retval) goto out_ret; - if (files == current->files) { - put_files_struct(files); - files = NULL; - } - retval = -ENOMEM; bprm = kzalloc(sizeof(*bprm), GFP_KERNEL); if (!bprm) @@ -1340,8 +1334,8 @@ int do_execve(char * filename, security_bprm_free(bprm); acct_update_integrals(current); kfree(bprm); - if (files) - put_files_struct(files); + if (displaced) + put_files_struct(displaced); return retval; } @@ -1363,8 +1357,8 @@ out_kfree: kfree(bprm); out_files: - if (files) - reset_files_struct(current, files); + if (displaced) + reset_files_struct(displaced); out_ret: return retval; } -- cgit v1.2.3 From f8f95702f0c4529b0f59488f4509608f0c160e77 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Wed, 23 Apr 2008 20:38:10 -0400 Subject: [PATCH] sanitize locate_fd() * 'file' argument is unused; lose it. * move setting flags from the caller (dupfd()) to locate_fd(); pass cloexec flag as new argument. Note that files_fdtable() that used to be in dupfd() isn't needed in the place in locate_fd() where the moved code ends up - we know that ->file_lock hadn't been dropped since the last time we calculated fdt because we can get there only if expand_files() returns 0 and it doesn't drop/reacquire in that case. * move getting/dropping ->file_lock into locate_fd(). Now the caller doesn't need to do anything with files_struct *files anymore and we can move that inside locate_fd() as well, killing the struct files_struct * argument. At that point locate_fd() is extremely similar to get_unused_fd_flags() and the next patches will merge those two. Signed-off-by: Al Viro --- fs/fcntl.c | 40 ++++++++++++++-------------------------- 1 file changed, 14 insertions(+), 26 deletions(-) (limited to 'fs') diff --git a/fs/fcntl.c b/fs/fcntl.c index e632da761fc..3f3ac630ccd 100644 --- a/fs/fcntl.c +++ b/fs/fcntl.c @@ -55,14 +55,16 @@ static int get_close_on_exec(unsigned int fd) * file_lock held for write. */ -static int locate_fd(struct files_struct *files, - struct file *file, unsigned int orig_start) +static int locate_fd(unsigned int orig_start, int cloexec) { + struct files_struct *files = current->files; unsigned int newfd; unsigned int start; int error; struct fdtable *fdt; + spin_lock(&files->file_lock); + error = -EINVAL; if (orig_start >= current->signal->rlim[RLIMIT_NOFILE].rlim_cur) goto out; @@ -97,42 +99,28 @@ repeat: if (error) goto repeat; - /* - * We reacquired files_lock, so we are safe as long as - * we reacquire the fdtable pointer and use it while holding - * the lock, no one can free it during that time. - */ if (start <= files->next_fd) files->next_fd = newfd + 1; + FD_SET(newfd, fdt->open_fds); + if (cloexec) + FD_SET(newfd, fdt->close_on_exec); + else + FD_CLR(newfd, fdt->close_on_exec); error = newfd; - + out: + spin_unlock(&files->file_lock); return error; } static int dupfd(struct file *file, unsigned int start, int cloexec) { - struct files_struct * files = current->files; - struct fdtable *fdt; - int fd; - - spin_lock(&files->file_lock); - fd = locate_fd(files, file, start); - if (fd >= 0) { - /* locate_fd() may have expanded fdtable, load the ptr */ - fdt = files_fdtable(files); - FD_SET(fd, fdt->open_fds); - if (cloexec) - FD_SET(fd, fdt->close_on_exec); - else - FD_CLR(fd, fdt->close_on_exec); - spin_unlock(&files->file_lock); + int fd = locate_fd(start, cloexec); + if (fd >= 0) fd_install(fd, file); - } else { - spin_unlock(&files->file_lock); + else fput(file); - } return fd; } -- cgit v1.2.3 From 5fcc60c3a05bf417229fba715e7aec52bf6717fb Mon Sep 17 00:00:00 2001 From: "David M. Richter" Date: Wed, 23 Apr 2008 16:28:59 -0400 Subject: leases: fix a return-value mixup Fixes a return-value mixup from 85c59580b30c82aa771aa33b37217a6b6851bc14 "locks: Fix potential OOPS in generic_setlease()", in which -ENOMEM replaced what had been intended to stay -EAGAIN in the variable "error". Signed-off-by: David M. Richter Signed-off-by: J. Bruce Fields --- fs/locks.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/locks.c b/fs/locks.c index 592faadbcec..b9f3a0bed30 100644 --- a/fs/locks.c +++ b/fs/locks.c @@ -1404,6 +1404,7 @@ int generic_setlease(struct file *filp, long arg, struct file_lock **flp) rdlease_count++; } + error = -EAGAIN; if ((arg == F_RDLCK && (wrlease_count > 0)) || (arg == F_WRLCK && ((rdlease_count + wrlease_count) > 0))) goto out; -- cgit v1.2.3 From 288b2fd8251cb0bcb14b8a93755ef9c78de70e0f Mon Sep 17 00:00:00 2001 From: "David M. Richter" Date: Wed, 23 Apr 2008 16:29:00 -0400 Subject: leases: when unlocking, skip locking-related steps In generic_setlease(), we don't need to allocate a new struct file_lock or check for readers or writers when called with F_UNLCK. Signed-off-by: David M. Richter Signed-off-by: J. Bruce Fields --- fs/locks.c | 24 +++++++++++++----------- 1 file changed, 13 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/locks.c b/fs/locks.c index b9f3a0bed30..da1d0ddb4ab 100644 --- a/fs/locks.c +++ b/fs/locks.c @@ -1367,18 +1367,20 @@ int generic_setlease(struct file *filp, long arg, struct file_lock **flp) lease = *flp; - error = -EAGAIN; - if ((arg == F_RDLCK) && (atomic_read(&inode->i_writecount) > 0)) - goto out; - if ((arg == F_WRLCK) - && ((atomic_read(&dentry->d_count) > 1) - || (atomic_read(&inode->i_count) > 1))) - goto out; + if (arg != F_UNLCK) { + error = -EAGAIN; + if ((arg == F_RDLCK) && (atomic_read(&inode->i_writecount) > 0)) + goto out; + if ((arg == F_WRLCK) + && ((atomic_read(&dentry->d_count) > 1) + || (atomic_read(&inode->i_count) > 1))) + goto out; - error = -ENOMEM; - new_fl = locks_alloc_lock(); - if (new_fl == NULL) - goto out; + error = -ENOMEM; + new_fl = locks_alloc_lock(); + if (new_fl == NULL) + goto out; + } /* * At this point, we know that if there is an exclusive -- cgit v1.2.3 From 190855576743a510219fc67886dace29b825d8cb Mon Sep 17 00:00:00 2001 From: "David M. Richter" Date: Wed, 23 Apr 2008 16:29:01 -0400 Subject: leases: move lock allocation earlier in generic_setlease() In generic_setlease(), the struct file_lock is allocated after tests for the presence of conflicting readers/writers is done, despite the fact that the allocation might block; this patch moves the allocation earlier. A subsequent set of patches will rely on this behavior to properly serialize between a modified __break_lease() and generic_setlease(). Signed-off-by: David M. Richter Signed-off-by: J. Bruce Fields --- fs/locks.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/locks.c b/fs/locks.c index da1d0ddb4ab..6a132cd4fa5 100644 --- a/fs/locks.c +++ b/fs/locks.c @@ -1368,6 +1368,11 @@ int generic_setlease(struct file *filp, long arg, struct file_lock **flp) lease = *flp; if (arg != F_UNLCK) { + error = -ENOMEM; + new_fl = locks_alloc_lock(); + if (new_fl == NULL) + goto out; + error = -EAGAIN; if ((arg == F_RDLCK) && (atomic_read(&inode->i_writecount) > 0)) goto out; @@ -1375,11 +1380,6 @@ int generic_setlease(struct file *filp, long arg, struct file_lock **flp) && ((atomic_read(&dentry->d_count) > 1) || (atomic_read(&inode->i_count) > 1))) goto out; - - error = -ENOMEM; - new_fl = locks_alloc_lock(); - if (new_fl == NULL) - goto out; } /* -- cgit v1.2.3 From 9d91cdcc0cce3186742f38e7352459b2087fbb86 Mon Sep 17 00:00:00 2001 From: "David M. Richter" Date: Wed, 23 Apr 2008 16:29:02 -0400 Subject: leases: remove unneeded variable from fcntl_setlease(). fcntl_setlease() has a struct dentry* that is used only once; this patch removes it. Signed-off-by: David M. Richter Signed-off-by: J. Bruce Fields --- fs/locks.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/locks.c b/fs/locks.c index 6a132cd4fa5..2e0fa661e42 100644 --- a/fs/locks.c +++ b/fs/locks.c @@ -1493,8 +1493,7 @@ EXPORT_SYMBOL_GPL(vfs_setlease); int fcntl_setlease(unsigned int fd, struct file *filp, long arg) { struct file_lock fl, *flp = &fl; - struct dentry *dentry = filp->f_path.dentry; - struct inode *inode = dentry->d_inode; + struct inode *inode = filp->f_path.dentry->d_inode; int error; locks_init_lock(&fl); -- cgit v1.2.3 From 4373ea84c84d8a96e99d3da99e813d3e36d1bd11 Mon Sep 17 00:00:00 2001 From: Wendy Cheng Date: Thu, 17 Jan 2008 11:10:12 -0500 Subject: lockd: unlock lockd locks associated with a given server ip For high-availability NFS service, we generally need to be able to drop file locks held on the exported filesystem before moving clients to a new server. Currently the only way to do that is by shutting down lockd entirely, which is often undesireable (for example, if you want to continue exporting other filesystems). This patch allows the administrator to release all locks held by clients accessing the client through a given server ip address, by echoing that address to a new file, /proc/fs/nfsd/unlock_ip, as in: shell> echo 10.1.1.2 > /proc/fs/nfsd/unlock_ip The expected sequence of events can be: 1. Tear down the IP address 2. Unexport the path 3. Write IP to /proc/fs/nfsd/unlock_ip to unlock files 4. Signal peer to begin take-over. For now we only support IPv4 addresses and NFSv2/v3 (NFSv4 locks are not affected). Also, if unmounting the filesystem is required, we assume at step 3 that clients using the given server ip are the only clients holding locks on the given filesystem; otherwise, an additional patch is required to allow revoking all locks held by lockd on a given filesystem. Signed-off-by: S. Wendy Cheng Cc: Lon Hohberger Cc: Christoph Hellwig Signed-off-by: J. Bruce Fields fs/lockd/svcsubs.c | 66 +++++++++++++++++++++++++++++++++++++++----- fs/nfsd/nfsctl.c | 65 +++++++++++++++++++++++++++++++++++++++++++ include/linux/lockd/lockd.h | 7 ++++ 3 files changed, 131 insertions(+), 7 deletions(-) --- fs/lockd/svcsubs.c | 36 +++++++++++++++++++++++++++++++----- fs/nfsd/nfsctl.c | 33 +++++++++++++++++++++++++++++++++ 2 files changed, 64 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/lockd/svcsubs.c b/fs/lockd/svcsubs.c index dbbefbcd671..1adcf93cb6e 100644 --- a/fs/lockd/svcsubs.c +++ b/fs/lockd/svcsubs.c @@ -18,6 +18,8 @@ #include #include #include +#include +#include #define NLMDBG_FACILITY NLMDBG_SVCSUBS @@ -230,7 +232,7 @@ nlm_file_inuse(struct nlm_file *file) * Loop over all files in the file table. */ static int -nlm_traverse_files(struct nlm_host *host, nlm_host_match_fn_t match) +nlm_traverse_files(void *data, nlm_host_match_fn_t match) { struct hlist_node *pos, *next; struct nlm_file *file; @@ -244,7 +246,7 @@ nlm_traverse_files(struct nlm_host *host, nlm_host_match_fn_t match) /* Traverse locks, blocks and shares of this file * and update file->f_locks count */ - if (nlm_inspect_file(host, file, match)) + if (nlm_inspect_file(data, file, match)) ret = 1; mutex_lock(&nlm_file_mutex); @@ -303,21 +305,27 @@ nlm_release_file(struct nlm_file *file) * Used by nlmsvc_invalidate_all */ static int -nlmsvc_mark_host(struct nlm_host *host, struct nlm_host *dummy) +nlmsvc_mark_host(void *data, struct nlm_host *dummy) { + struct nlm_host *host = data; + host->h_inuse = 1; return 0; } static int -nlmsvc_same_host(struct nlm_host *host, struct nlm_host *other) +nlmsvc_same_host(void *data, struct nlm_host *other) { + struct nlm_host *host = data; + return host == other; } static int -nlmsvc_is_client(struct nlm_host *host, struct nlm_host *dummy) +nlmsvc_is_client(void *data, struct nlm_host *dummy) { + struct nlm_host *host = data; + if (host->h_server) { /* we are destroying locks even though the client * hasn't asked us too, so don't unmonitor the @@ -370,3 +378,21 @@ nlmsvc_invalidate_all(void) */ nlm_traverse_files(NULL, nlmsvc_is_client); } + +static int +nlmsvc_match_ip(void *datap, struct nlm_host *host) +{ + __be32 *server_addr = datap; + + return host->h_saddr.sin_addr.s_addr == *server_addr; +} + +int +nlmsvc_unlock_all_by_ip(__be32 server_addr) +{ + int ret; + ret = nlm_traverse_files(&server_addr, nlmsvc_match_ip); + return ret ? -EIO : 0; + +} +EXPORT_SYMBOL_GPL(nlmsvc_unlock_all_by_ip); diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index 613bcb8171a..7706f6c9ab8 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -22,6 +22,7 @@ #include #include #include +#include #include #include #include @@ -35,6 +36,7 @@ #include #include #include +#include #include #include @@ -53,6 +55,7 @@ enum { NFSD_Getfs, NFSD_List, NFSD_Fh, + NFSD_FO_UnlockIP, NFSD_Threads, NFSD_Pool_Threads, NFSD_Versions, @@ -89,6 +92,8 @@ static ssize_t write_leasetime(struct file *file, char *buf, size_t size); static ssize_t write_recoverydir(struct file *file, char *buf, size_t size); #endif +static ssize_t failover_unlock_ip(struct file *file, char *buf, size_t size); + static ssize_t (*write_op[])(struct file *, char *, size_t) = { [NFSD_Svc] = write_svc, [NFSD_Add] = write_add, @@ -98,6 +103,7 @@ static ssize_t (*write_op[])(struct file *, char *, size_t) = { [NFSD_Getfd] = write_getfd, [NFSD_Getfs] = write_getfs, [NFSD_Fh] = write_filehandle, + [NFSD_FO_UnlockIP] = failover_unlock_ip, [NFSD_Threads] = write_threads, [NFSD_Pool_Threads] = write_pool_threads, [NFSD_Versions] = write_versions, @@ -298,6 +304,31 @@ static ssize_t write_getfd(struct file *file, char *buf, size_t size) return err; } +static ssize_t failover_unlock_ip(struct file *file, char *buf, size_t size) +{ + __be32 server_ip; + char *fo_path, c; + int b1, b2, b3, b4; + + /* sanity check */ + if (size == 0) + return -EINVAL; + + if (buf[size-1] != '\n') + return -EINVAL; + + fo_path = buf; + if (qword_get(&buf, fo_path, size) < 0) + return -EINVAL; + + /* get ipv4 address */ + if (sscanf(fo_path, "%u.%u.%u.%u%c", &b1, &b2, &b3, &b4, &c) != 4) + return -EINVAL; + server_ip = htonl((((((b1<<8)|b2)<<8)|b3)<<8)|b4); + + return nlmsvc_unlock_all_by_ip(server_ip); +} + static ssize_t write_filehandle(struct file *file, char *buf, size_t size) { /* request is: @@ -700,6 +731,8 @@ static int nfsd_fill_super(struct super_block * sb, void * data, int silent) [NFSD_Getfd] = {".getfd", &transaction_ops, S_IWUSR|S_IRUSR}, [NFSD_Getfs] = {".getfs", &transaction_ops, S_IWUSR|S_IRUSR}, [NFSD_List] = {"exports", &exports_operations, S_IRUGO}, + [NFSD_FO_UnlockIP] = {"unlock_ip", + &transaction_ops, S_IWUSR|S_IRUSR}, [NFSD_Fh] = {"filehandle", &transaction_ops, S_IWUSR|S_IRUSR}, [NFSD_Threads] = {"threads", &transaction_ops, S_IWUSR|S_IRUSR}, [NFSD_Pool_Threads] = {"pool_threads", &transaction_ops, S_IWUSR|S_IRUSR}, -- cgit v1.2.3 From 17efa372cfe4d189705edf6cd4fbe283827a5dc7 Mon Sep 17 00:00:00 2001 From: Wendy Cheng Date: Thu, 17 Jan 2008 11:10:12 -0500 Subject: lockd: unlock lockd locks held for a certain filesystem Add /proc/fs/nfsd/unlock_filesystem, which allows e.g.: shell> echo /mnt/sfs1 > /proc/fs/nfsd/unlock_filesystem so that a filesystem can be unmounted before allowing a peer nfsd to take over nfs service for the filesystem. Signed-off-by: S. Wendy Cheng Cc: Lon Hohberger Cc: Christoph Hellwig Signed-off-by: J. Bruce Fields fs/lockd/svcsubs.c | 66 +++++++++++++++++++++++++++++++++++++++----- fs/nfsd/nfsctl.c | 65 +++++++++++++++++++++++++++++++++++++++++++ include/linux/lockd/lockd.h | 7 ++++ 3 files changed, 131 insertions(+), 7 deletions(-) --- fs/lockd/svcsubs.c | 37 ++++++++++++++++++++++++++++++++----- fs/nfsd/nfsctl.c | 32 ++++++++++++++++++++++++++++++++ 2 files changed, 64 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/lockd/svcsubs.c b/fs/lockd/svcsubs.c index 1adcf93cb6e..d1c48b539df 100644 --- a/fs/lockd/svcsubs.c +++ b/fs/lockd/svcsubs.c @@ -196,6 +196,12 @@ again: return 0; } +static int +nlmsvc_always_match(void *dummy1, struct nlm_host *dummy2) +{ + return 1; +} + /* * Inspect a single file */ @@ -232,7 +238,8 @@ nlm_file_inuse(struct nlm_file *file) * Loop over all files in the file table. */ static int -nlm_traverse_files(void *data, nlm_host_match_fn_t match) +nlm_traverse_files(void *data, nlm_host_match_fn_t match, + int (*is_failover_file)(void *data, struct nlm_file *file)) { struct hlist_node *pos, *next; struct nlm_file *file; @@ -241,6 +248,8 @@ nlm_traverse_files(void *data, nlm_host_match_fn_t match) mutex_lock(&nlm_file_mutex); for (i = 0; i < FILE_NRHASH; i++) { hlist_for_each_entry_safe(file, pos, next, &nlm_files[i], f_list) { + if (is_failover_file && !is_failover_file(data, file)) + continue; file->f_count++; mutex_unlock(&nlm_file_mutex); @@ -345,7 +354,7 @@ void nlmsvc_mark_resources(void) { dprintk("lockd: nlmsvc_mark_resources\n"); - nlm_traverse_files(NULL, nlmsvc_mark_host); + nlm_traverse_files(NULL, nlmsvc_mark_host, NULL); } /* @@ -356,7 +365,7 @@ nlmsvc_free_host_resources(struct nlm_host *host) { dprintk("lockd: nlmsvc_free_host_resources\n"); - if (nlm_traverse_files(host, nlmsvc_same_host)) { + if (nlm_traverse_files(host, nlmsvc_same_host, NULL)) { printk(KERN_WARNING "lockd: couldn't remove all locks held by %s\n", host->h_name); @@ -376,8 +385,26 @@ nlmsvc_invalidate_all(void) * turn, which is about as inefficient as it gets. * Now we just do it once in nlm_traverse_files. */ - nlm_traverse_files(NULL, nlmsvc_is_client); + nlm_traverse_files(NULL, nlmsvc_is_client, NULL); +} + +static int +nlmsvc_match_sb(void *datap, struct nlm_file *file) +{ + struct super_block *sb = datap; + + return sb == file->f_file->f_path.mnt->mnt_sb; +} + +int +nlmsvc_unlock_all_by_sb(struct super_block *sb) +{ + int ret; + + ret = nlm_traverse_files(sb, nlmsvc_always_match, nlmsvc_match_sb); + return ret ? -EIO : 0; } +EXPORT_SYMBOL_GPL(nlmsvc_unlock_all_by_sb); static int nlmsvc_match_ip(void *datap, struct nlm_host *host) @@ -391,7 +418,7 @@ int nlmsvc_unlock_all_by_ip(__be32 server_addr) { int ret; - ret = nlm_traverse_files(&server_addr, nlmsvc_match_ip); + ret = nlm_traverse_files(&server_addr, nlmsvc_match_ip, NULL); return ret ? -EIO : 0; } diff --git a/fs/nfsd/nfsctl.c b/fs/nfsd/nfsctl.c index 7706f6c9ab8..42f3820ee8f 100644 --- a/fs/nfsd/nfsctl.c +++ b/fs/nfsd/nfsctl.c @@ -56,6 +56,7 @@ enum { NFSD_List, NFSD_Fh, NFSD_FO_UnlockIP, + NFSD_FO_UnlockFS, NFSD_Threads, NFSD_Pool_Threads, NFSD_Versions, @@ -93,6 +94,7 @@ static ssize_t write_recoverydir(struct file *file, char *buf, size_t size); #endif static ssize_t failover_unlock_ip(struct file *file, char *buf, size_t size); +static ssize_t failover_unlock_fs(struct file *file, char *buf, size_t size); static ssize_t (*write_op[])(struct file *, char *, size_t) = { [NFSD_Svc] = write_svc, @@ -104,6 +106,7 @@ static ssize_t (*write_op[])(struct file *, char *, size_t) = { [NFSD_Getfs] = write_getfs, [NFSD_Fh] = write_filehandle, [NFSD_FO_UnlockIP] = failover_unlock_ip, + [NFSD_FO_UnlockFS] = failover_unlock_fs, [NFSD_Threads] = write_threads, [NFSD_Pool_Threads] = write_pool_threads, [NFSD_Versions] = write_versions, @@ -329,6 +332,33 @@ static ssize_t failover_unlock_ip(struct file *file, char *buf, size_t size) return nlmsvc_unlock_all_by_ip(server_ip); } +static ssize_t failover_unlock_fs(struct file *file, char *buf, size_t size) +{ + struct nameidata nd; + char *fo_path; + int error; + + /* sanity check */ + if (size == 0) + return -EINVAL; + + if (buf[size-1] != '\n') + return -EINVAL; + + fo_path = buf; + if (qword_get(&buf, fo_path, size) < 0) + return -EINVAL; + + error = path_lookup(fo_path, 0, &nd); + if (error) + return error; + + error = nlmsvc_unlock_all_by_sb(nd.path.mnt->mnt_sb); + + path_put(&nd.path); + return error; +} + static ssize_t write_filehandle(struct file *file, char *buf, size_t size) { /* request is: @@ -733,6 +763,8 @@ static int nfsd_fill_super(struct super_block * sb, void * data, int silent) [NFSD_List] = {"exports", &exports_operations, S_IRUGO}, [NFSD_FO_UnlockIP] = {"unlock_ip", &transaction_ops, S_IWUSR|S_IRUSR}, + [NFSD_FO_UnlockFS] = {"unlock_filesystem", + &transaction_ops, S_IWUSR|S_IRUSR}, [NFSD_Fh] = {"filehandle", &transaction_ops, S_IWUSR|S_IRUSR}, [NFSD_Threads] = {"threads", &transaction_ops, S_IWUSR|S_IRUSR}, [NFSD_Pool_Threads] = {"pool_threads", &transaction_ops, S_IWUSR|S_IRUSR}, -- cgit v1.2.3 From 1a747ee0cc11a198f9e2435add821bd0dfedb7c1 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Thu, 24 Apr 2008 10:08:22 -0400 Subject: locks: don't call ->copy_lock methods on return of conflicting locks The file_lock structure is used both as a heavy-weight representation of an active lock, with pointers to reference-counted structures, etc., and as a simple container for parameters that describe a file lock. The conflicting lock returned from __posix_lock_file is an example of the latter; so don't call the filesystem or lock manager callbacks when copying to it. This also saves the need for an unnecessary locks_init_lock in the nfsv4 server. Thanks to Trond for pointing out the error. Signed-off-by: J. Bruce Fields Cc: Trond Myklebust --- fs/lockd/svclock.c | 2 +- fs/locks.c | 4 ++-- fs/nfsd/nfs4state.c | 3 --- 3 files changed, 3 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/lockd/svclock.c b/fs/lockd/svclock.c index 1f122c1940a..4d81553d294 100644 --- a/fs/lockd/svclock.c +++ b/fs/lockd/svclock.c @@ -632,7 +632,7 @@ nlmsvc_update_deferred_block(struct nlm_block *block, struct file_lock *conf, block->b_flags |= B_TIMED_OUT; if (conf) { if (block->b_fl) - locks_copy_lock(block->b_fl, conf); + __locks_copy_lock(block->b_fl, conf); } } diff --git a/fs/locks.c b/fs/locks.c index 2e0fa661e42..e1ea2fe0368 100644 --- a/fs/locks.c +++ b/fs/locks.c @@ -224,7 +224,7 @@ static void locks_copy_private(struct file_lock *new, struct file_lock *fl) /* * Initialize a new lock from an existing file_lock structure. */ -static void __locks_copy_lock(struct file_lock *new, const struct file_lock *fl) +void __locks_copy_lock(struct file_lock *new, const struct file_lock *fl) { new->fl_owner = fl->fl_owner; new->fl_pid = fl->fl_pid; @@ -833,7 +833,7 @@ static int __posix_lock_file(struct inode *inode, struct file_lock *request, str if (!posix_locks_conflict(request, fl)) continue; if (conflock) - locks_copy_lock(conflock, fl); + __locks_copy_lock(conflock, fl); error = -EAGAIN; if (!(request->fl_flags & FL_SLEEP)) goto out; diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c index 55dfdd71f1b..8799b870818 100644 --- a/fs/nfsd/nfs4state.c +++ b/fs/nfsd/nfs4state.c @@ -2712,9 +2712,6 @@ nfsd4_lock(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate, * Note: locks.c uses the BKL to protect the inode's lock list. */ - /* XXX?: Just to divert the locks_release_private at the start of - * locks_copy_lock: */ - locks_init_lock(&conflock); err = vfs_lock_file(filp, cmd, &file_lock, &conflock); switch (-err) { case 0: /* success! */ -- cgit v1.2.3 From e36cd4a2873c398ba188f16e4087cce7f00a1506 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Thu, 24 Apr 2008 16:59:30 -0400 Subject: nfsd: don't allow setting ctime over v4 Presumably this is left over from earlier drafts of v4, which listed TIME_METADATA as writeable. It's read-only in rfc 3530, and shouldn't be modifiable anyway. Signed-off-by: J. Bruce Fields --- fs/nfsd/nfs4xdr.c | 14 -------------- 1 file changed, 14 deletions(-) (limited to 'fs') diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c index 1ba7ad98193..c513bbdf2d3 100644 --- a/fs/nfsd/nfs4xdr.c +++ b/fs/nfsd/nfs4xdr.c @@ -376,20 +376,6 @@ nfsd4_decode_fattr(struct nfsd4_compoundargs *argp, u32 *bmval, struct iattr *ia goto xdr_error; } } - if (bmval[1] & FATTR4_WORD1_TIME_METADATA) { - /* We require the high 32 bits of 'seconds' to be 0, and we ignore - all 32 bits of 'nseconds'. */ - READ_BUF(12); - len += 12; - READ32(dummy32); - if (dummy32) - return nfserr_inval; - READ32(iattr->ia_ctime.tv_sec); - READ32(iattr->ia_ctime.tv_nsec); - if (iattr->ia_ctime.tv_nsec >= (u32)1000000000) - return nfserr_inval; - iattr->ia_valid |= ATTR_CTIME; - } if (bmval[1] & FATTR4_WORD1_TIME_MODIFY_SET) { READ_BUF(4); len += 4; -- cgit v1.2.3 From 3dd7b71ca0f7ff8410a6b8cb15e5f670f90181e4 Mon Sep 17 00:00:00 2001 From: Roland Dreier Date: Fri, 25 Apr 2008 15:32:51 -0700 Subject: Export __locks_copy_lock() so modular lockd builds Commit 1a747ee0 ("locks: don't call ->copy_lock methods on return of conflicting locks") changed fs/lockd/svclock.c to call __locks_copy_lock() instead of locks_copy_lock(), but lockd can be built as a module and __locks_copy_lock() is not exported, which causes a build error ERROR: "__locks_copy_lock" [fs/lockd/lockd.ko] undefined! with CONFIG_LOCKD=m. Fix this by exporting __locks_copy_lock(). Signed-off-by: Roland Dreier Signed-off-by: Linus Torvalds --- fs/locks.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/locks.c b/fs/locks.c index e1ea2fe0368..44d9a6a7ec5 100644 --- a/fs/locks.c +++ b/fs/locks.c @@ -236,6 +236,7 @@ void __locks_copy_lock(struct file_lock *new, const struct file_lock *fl) new->fl_ops = NULL; new->fl_lmops = NULL; } +EXPORT_SYMBOL(__locks_copy_lock); void locks_copy_lock(struct file_lock *new, struct file_lock *fl) { -- cgit v1.2.3 From e92adcba261fd391591bb63c1703185a04a41554 Mon Sep 17 00:00:00 2001 From: Jeff Moyer Date: Mon, 28 Apr 2008 02:12:04 -0700 Subject: aio: io_getevents() should return if io_destroy() is invoked This patch wakes up a thread waiting in io_getevents if another thread destroys the context. This was tested using a small program that spawns a thread to wait in io_getevents while the parent thread destroys the io context and then waits for the getevents thread to exit. Without this patch, the program hangs indefinitely. With the patch, the program exits as expected. Signed-off-by: Jeff Moyer Cc: Zach Brown Cc: Christopher Smith Cc: Benjamin LaHaise Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/aio.c | 12 +++++++++++- 1 file changed, 11 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/aio.c b/fs/aio.c index 228368610df..ae94e1dea26 100644 --- a/fs/aio.c +++ b/fs/aio.c @@ -1166,7 +1166,10 @@ retry: break; if (min_nr <= i) break; - ret = 0; + if (unlikely(ctx->dead)) { + ret = -EINVAL; + break; + } if (to.timed_out) /* Only check after read evt */ break; /* Try to only show up in io wait if there are ops @@ -1231,6 +1234,13 @@ static void io_destroy(struct kioctx *ioctx) aio_cancel_all(ioctx); wait_for_all_aios(ioctx); + + /* + * Wake up any waiters. The setting of ctx->dead must be seen + * by other CPUs at this point. Right now, we rely on the + * locking done by the above calls to ensure this consistency. + */ + wake_up(&ioctx->wait); put_ioctx(ioctx); /* once for the lookup */ } -- cgit v1.2.3 From 488514d1798289f56f80ed018e246179fe500383 Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Mon, 28 Apr 2008 02:12:05 -0700 Subject: Remove set_migrateflags() Migrate flags must be set on slab creation as agreed upon when the antifrag logic was reviewed. Otherwise some slabs of a slabcache will end up in the unmovable and others in the reclaimable section depending on which flag was active when a new slab page was allocated. This likely slid in somehow when antifrag was merged. Remove it. The buffer_heads are always allocated with __GFP_RECLAIMABLE because the SLAB_RECLAIM_ACCOUNT option is set. The set_migrateflags() never had any effect there. Radix tree allocations are not directly reclaimable but they are allocated with __GFP_RECLAIMABLE set on each allocation. We now set SLAB_RECLAIM_ACCOUNT on radix tree slab creation making sure that radix tree slabs are consistently placed in the reclaimable section. Radix tree slabs will also be accounted as such. There is then no user left of set_migratepages. So remove it. Signed-off-by: Christoph Lameter Cc: Mel Gorman Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/buffer.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/buffer.c b/fs/buffer.c index 39ff14403d1..8b9807523ef 100644 --- a/fs/buffer.c +++ b/fs/buffer.c @@ -3180,8 +3180,7 @@ static void recalc_bh_state(void) struct buffer_head *alloc_buffer_head(gfp_t gfp_flags) { - struct buffer_head *ret = kmem_cache_alloc(bh_cachep, - set_migrateflags(gfp_flags, __GFP_RECLAIMABLE)); + struct buffer_head *ret = kmem_cache_alloc(bh_cachep, gfp_flags); if (ret) { INIT_LIST_HEAD(&ret->b_assoc_buffers); get_cpu_var(bh_accounting).nr++; -- cgit v1.2.3 From 3c18ddd160d1fcd46d1131d9ad6c594dd8e9af99 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Mon, 28 Apr 2008 02:12:10 -0700 Subject: mm: remove nopage Nothing in the tree uses nopage any more. Remove support for it in the core mm code and documentation (and a few stray references to it in comments). Signed-off-by: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/gfs2/ops_address.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/gfs2/ops_address.c b/fs/gfs2/ops_address.c index 90a04a6e378..f55394e57cb 100644 --- a/fs/gfs2/ops_address.c +++ b/fs/gfs2/ops_address.c @@ -438,7 +438,7 @@ static int stuffed_readpage(struct gfs2_inode *ip, struct page *page) int error; /* - * Due to the order of unstuffing files and ->nopage(), we can be + * Due to the order of unstuffing files and ->fault(), we can be * asked for a zero page in the case of a stuffed file being extended, * so we need to supply one here. It doesn't happen often. */ -- cgit v1.2.3 From 9d02dbc8137759e4c2f91db0b7f9c8a1ec2a9276 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Mon, 28 Apr 2008 02:12:11 -0700 Subject: make swap_pte_to_pagemap_entry() static Make the needlessly global swap_pte_to_pagemap_entry() static. Signed-off-by: Adrian Bunk Acked-by: Matt Mackall Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/task_mmu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c index 9dfb5ff2420..f4ab76c7c66 100644 --- a/fs/proc/task_mmu.c +++ b/fs/proc/task_mmu.c @@ -579,7 +579,7 @@ static int pagemap_pte_hole(unsigned long start, unsigned long end, return err; } -u64 swap_pte_to_pagemap_entry(pte_t pte) +static u64 swap_pte_to_pagemap_entry(pte_t pte) { swp_entry_t e = pte_to_swp_entry(pte); return swp_type(e) | (swp_offset(e) << MAX_SWAPFILES_SHIFT); -- cgit v1.2.3 From dac1d27bc8d5ca636d3014ecfdf94407031d1970 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Mon, 28 Apr 2008 02:12:12 -0700 Subject: mm: use zonelists instead of zones when direct reclaiming pages The following patches replace multiple zonelists per node with two zonelists that are filtered based on the GFP flags. The patches as a set fix a bug with regard to the use of MPOL_BIND and ZONE_MOVABLE. With this patchset, the MPOL_BIND will apply to the two highest zones when the highest zone is ZONE_MOVABLE. This should be considered as an alternative fix for the MPOL_BIND+ZONE_MOVABLE in 2.6.23 to the previously discussed hack that filters only custom zonelists. The first patch cleans up an inconsistency where direct reclaim uses zonelist->zones where other places use zonelist. The second patch introduces a helper function node_zonelist() for looking up the appropriate zonelist for a GFP mask which simplifies patches later in the set. The third patch defines/remembers the "preferred zone" for numa statistics, as it is no longer always the first zone in a zonelist. The forth patch replaces multiple zonelists with two zonelists that are filtered. The two zonelists are due to the fact that the memoryless patchset introduces a second set of zonelists for __GFP_THISNODE. The fifth patch introduces helper macros for retrieving the zone and node indices of entries in a zonelist. The final patch introduces filtering of the zonelists based on a nodemask. Two zonelists exist per node, one for normal allocations and one for __GFP_THISNODE. Performance results varied depending on the machine configuration. In real workloads the gain/loss will depend on how much the userspace portion of the benchmark benefits from having more cache available due to reduced referencing of zonelists. These are the range of performance losses/gains when running against 2.6.24-rc4-mm1. The set and these machines are a mix of i386, x86_64 and ppc64 both NUMA and non-NUMA. loss to gain Total CPU time on Kernbench: -0.86% to 1.13% Elapsed time on Kernbench: -0.79% to 0.76% page_test from aim9: -4.37% to 0.79% brk_test from aim9: -0.71% to 4.07% fork_test from aim9: -1.84% to 4.60% exec_test from aim9: -0.71% to 1.08% This patch: The allocator deals with zonelists which indicate the order in which zones should be targeted for an allocation. Similarly, direct reclaim of pages iterates over an array of zones. For consistency, this patch converts direct reclaim to use a zonelist. No functionality is changed by this patch. This simplifies zonelist iterators in the next patch. Signed-off-by: Mel Gorman Acked-by: Christoph Lameter Signed-off-by: Lee Schermerhorn Cc: KAMEZAWA Hiroyuki Cc: Mel Gorman Cc: Christoph Lameter Cc: Hugh Dickins Cc: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/buffer.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/buffer.c b/fs/buffer.c index 8b9807523ef..1dae94acb3f 100644 --- a/fs/buffer.c +++ b/fs/buffer.c @@ -360,16 +360,16 @@ void invalidate_bdev(struct block_device *bdev) */ static void free_more_memory(void) { - struct zone **zones; + struct zonelist *zonelist; pg_data_t *pgdat; wakeup_pdflush(1024); yield(); for_each_online_pgdat(pgdat) { - zones = pgdat->node_zonelists[gfp_zone(GFP_NOFS)].zones; - if (*zones) - try_to_free_pages(zones, 0, GFP_NOFS); + zonelist = &pgdat->node_zonelists[gfp_zone(GFP_NOFS)]; + if (zonelist->zones[0]) + try_to_free_pages(zonelist, 0, GFP_NOFS); } } -- cgit v1.2.3 From 0e88460da6ab7bb6a7ef83675412ed5b6315d741 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Mon, 28 Apr 2008 02:12:14 -0700 Subject: mm: introduce node_zonelist() for accessing the zonelist for a GFP mask Introduce a node_zonelist() helper function. It is used to lookup the appropriate zonelist given a node and a GFP mask. The patch on its own is a cleanup but it helps clarify parts of the two-zonelist-per-node patchset. If necessary, it can be merged with the next patch in this set without problems. Reviewed-by: Christoph Lameter Signed-off-by: Mel Gorman Signed-off-by: Lee Schermerhorn Cc: KAMEZAWA Hiroyuki Cc: Mel Gorman Cc: Christoph Lameter Cc: Hugh Dickins Cc: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/buffer.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/buffer.c b/fs/buffer.c index 1dae94acb3f..71358499bc5 100644 --- a/fs/buffer.c +++ b/fs/buffer.c @@ -361,13 +361,13 @@ void invalidate_bdev(struct block_device *bdev) static void free_more_memory(void) { struct zonelist *zonelist; - pg_data_t *pgdat; + int nid; wakeup_pdflush(1024); yield(); - for_each_online_pgdat(pgdat) { - zonelist = &pgdat->node_zonelists[gfp_zone(GFP_NOFS)]; + for_each_online_node(nid) { + zonelist = node_zonelist(nid, GFP_NOFS); if (zonelist->zones[0]) try_to_free_pages(zonelist, 0, GFP_NOFS); } -- cgit v1.2.3 From 54a6eb5c4765aa573a030ceeba2c14e3d2ea5706 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Mon, 28 Apr 2008 02:12:16 -0700 Subject: mm: use two zonelist that are filtered by GFP mask Currently a node has two sets of zonelists, one for each zone type in the system and a second set for GFP_THISNODE allocations. Based on the zones allowed by a gfp mask, one of these zonelists is selected. All of these zonelists consume memory and occupy cache lines. This patch replaces the multiple zonelists per-node with two zonelists. The first contains all populated zones in the system, ordered by distance, for fallback allocations when the target/preferred node has no free pages. The second contains all populated zones in the node suitable for GFP_THISNODE allocations. An iterator macro is introduced called for_each_zone_zonelist() that interates through each zone allowed by the GFP flags in the selected zonelist. Signed-off-by: Mel Gorman Acked-by: Christoph Lameter Signed-off-by: Lee Schermerhorn Cc: KAMEZAWA Hiroyuki Cc: Mel Gorman Cc: Christoph Lameter Cc: Hugh Dickins Cc: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/buffer.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/buffer.c b/fs/buffer.c index 71358499bc5..9b5434a8047 100644 --- a/fs/buffer.c +++ b/fs/buffer.c @@ -360,16 +360,18 @@ void invalidate_bdev(struct block_device *bdev) */ static void free_more_memory(void) { - struct zonelist *zonelist; + struct zone **zones; int nid; wakeup_pdflush(1024); yield(); for_each_online_node(nid) { - zonelist = node_zonelist(nid, GFP_NOFS); - if (zonelist->zones[0]) - try_to_free_pages(zonelist, 0, GFP_NOFS); + zones = first_zones_zonelist(node_zonelist(nid, GFP_NOFS), + gfp_zone(GFP_NOFS)); + if (*zones) + try_to_free_pages(node_zonelist(nid, GFP_NOFS), 0, + GFP_NOFS); } } -- cgit v1.2.3 From dd1a239f6f2d4d3eedd318583ec319aa145b324c Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Mon, 28 Apr 2008 02:12:17 -0700 Subject: mm: have zonelist contains structs with both a zone pointer and zone_idx Filtering zonelists requires very frequent use of zone_idx(). This is costly as it involves a lookup of another structure and a substraction operation. As the zone_idx is often required, it should be quickly accessible. The node idx could also be stored here if it was found that accessing zone->node is significant which may be the case on workloads where nodemasks are heavily used. This patch introduces a struct zoneref to store a zone pointer and a zone index. The zonelist then consists of an array of these struct zonerefs which are looked up as necessary. Helpers are given for accessing the zone index as well as the node index. [kamezawa.hiroyu@jp.fujitsu.com: Suggested struct zoneref instead of embedding information in pointers] [hugh@veritas.com: mm-have-zonelist: fix memcg ooms] [hugh@veritas.com: just return do_try_to_free_pages] [hugh@veritas.com: do_try_to_free_pages gfp_mask redundant] Signed-off-by: Mel Gorman Acked-by: Christoph Lameter Acked-by: David Rientjes Signed-off-by: Lee Schermerhorn Cc: KAMEZAWA Hiroyuki Cc: Mel Gorman Cc: Christoph Lameter Cc: Nick Piggin Signed-off-by: Hugh Dickins Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/buffer.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/buffer.c b/fs/buffer.c index 9b5434a8047..ac84cd13075 100644 --- a/fs/buffer.c +++ b/fs/buffer.c @@ -360,16 +360,16 @@ void invalidate_bdev(struct block_device *bdev) */ static void free_more_memory(void) { - struct zone **zones; + struct zoneref *zrefs; int nid; wakeup_pdflush(1024); yield(); for_each_online_node(nid) { - zones = first_zones_zonelist(node_zonelist(nid, GFP_NOFS), + zrefs = first_zones_zonelist(node_zonelist(nid, GFP_NOFS), gfp_zone(GFP_NOFS)); - if (*zones) + if (zrefs->zone) try_to_free_pages(node_zonelist(nid, GFP_NOFS), 0, GFP_NOFS); } -- cgit v1.2.3 From 19770b32609b6bf97a3dece2529089494cbfc549 Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Mon, 28 Apr 2008 02:12:18 -0700 Subject: mm: filter based on a nodemask as well as a gfp_mask The MPOL_BIND policy creates a zonelist that is used for allocations controlled by that mempolicy. As the per-node zonelist is already being filtered based on a zone id, this patch adds a version of __alloc_pages() that takes a nodemask for further filtering. This eliminates the need for MPOL_BIND to create a custom zonelist. A positive benefit of this is that allocations using MPOL_BIND now use the local node's distance-ordered zonelist instead of a custom node-id-ordered zonelist. I.e., pages will be allocated from the closest allowed node with available memory. [Lee.Schermerhorn@hp.com: Mempolicy: update stale documentation and comments] [Lee.Schermerhorn@hp.com: Mempolicy: make dequeue_huge_page_vma() obey MPOL_BIND nodemask] [Lee.Schermerhorn@hp.com: Mempolicy: make dequeue_huge_page_vma() obey MPOL_BIND nodemask rework] Signed-off-by: Mel Gorman Acked-by: Christoph Lameter Signed-off-by: Lee Schermerhorn Cc: KAMEZAWA Hiroyuki Cc: Mel Gorman Cc: Hugh Dickins Cc: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/buffer.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/buffer.c b/fs/buffer.c index ac84cd13075..7d51e649b19 100644 --- a/fs/buffer.c +++ b/fs/buffer.c @@ -360,16 +360,17 @@ void invalidate_bdev(struct block_device *bdev) */ static void free_more_memory(void) { - struct zoneref *zrefs; + struct zone *zone; int nid; wakeup_pdflush(1024); yield(); for_each_online_node(nid) { - zrefs = first_zones_zonelist(node_zonelist(nid, GFP_NOFS), - gfp_zone(GFP_NOFS)); - if (zrefs->zone) + (void)first_zones_zonelist(node_zonelist(nid, GFP_NOFS), + gfp_zone(GFP_NOFS), NULL, + &zone); + if (zone) try_to_free_pages(node_zonelist(nid, GFP_NOFS), 0, GFP_NOFS); } -- cgit v1.2.3 From 028fec414d803117eb4b2ed12acb4dd5da65b32d Mon Sep 17 00:00:00 2001 From: David Rientjes Date: Mon, 28 Apr 2008 02:12:25 -0700 Subject: mempolicy: support optional mode flags With the evolution of mempolicies, it is necessary to support mempolicy mode flags that specify how the policy shall behave in certain circumstances. The most immediate need for mode flag support is to suppress remapping the nodemask of a policy at the time of rebind. Both the mempolicy mode and flags are passed by the user in the 'int policy' formal of either the set_mempolicy() or mbind() syscall. A new constant, MPOL_MODE_FLAGS, represents the union of legal optional flags that may be passed as part of this int. Mempolicies that include illegal flags as part of their policy are rejected as invalid. An additional member to struct mempolicy is added to support the mode flags: struct mempolicy { ... unsigned short policy; unsigned short flags; } The splitting of the 'int' actual passed by the user is done in sys_set_mempolicy() and sys_mbind() for their respective syscalls. This is done by intersecting the actual with MPOL_MODE_FLAGS, rejecting the syscall of there are additional flags, and storing it in the new 'flags' member of struct mempolicy. The intersection of the actual with ~MPOL_MODE_FLAGS is stored in the 'policy' member of the struct and all current users of pol->policy remain unchanged. The union of the policy mode and optional mode flags is passed back to the user in get_mempolicy(). This combination of mode and flags within the same actual does not break userspace code that relies on get_mempolicy(&policy, ...) and either switch (policy) { case MPOL_BIND: ... case MPOL_INTERLEAVE: ... }; statements or if (policy == MPOL_INTERLEAVE) { ... } statements. Such applications would need to use optional mode flags when calling set_mempolicy() or mbind() for these previously implemented statements to stop working. If an application does start using optional mode flags, it will need to mask the optional flags off the policy in switch and conditional statements that only test mode. An additional member is also added to struct shmem_sb_info to store the optional mode flags. [hugh@veritas.com: shmem mpol: fix build warning] Cc: Paul Jackson Cc: Christoph Lameter Cc: Lee Schermerhorn Cc: Andi Kleen Signed-off-by: David Rientjes Signed-off-by: Hugh Dickins Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/hugetlbfs/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c index 6846785fe90..2e9e5bdd562 100644 --- a/fs/hugetlbfs/inode.c +++ b/fs/hugetlbfs/inode.c @@ -504,7 +504,7 @@ static struct inode *hugetlbfs_get_inode(struct super_block *sb, uid_t uid, inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME; INIT_LIST_HEAD(&inode->i_mapping->private_list); info = HUGETLBFS_I(inode); - mpol_shared_policy_init(&info->policy, MPOL_DEFAULT, NULL); + mpol_shared_policy_init(&info->policy, MPOL_DEFAULT, 0, NULL); switch (mode & S_IFMT) { default: init_special_inode(inode, mode, dev); -- cgit v1.2.3 From a10aa579878fc6f9cd17455067380bbdf1d53c91 Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Mon, 28 Apr 2008 02:12:40 -0700 Subject: vmalloc: show vmalloced areas via /proc/vmallocinfo Implement a new proc file that allows the display of the currently allocated vmalloc memory. It allows to see the users of vmalloc. That is important if vmalloc space is scarce (i386 for example). And it's going to be important for the compound page fallback to vmalloc. Many of the current users can be switched to use compound pages with fallback. This means that the number of users of vmalloc is reduced and page tables no longer necessary to access the memory. /proc/vmallocinfo allows to review how that reduction occurs. If memory becomes fragmented and larger order allocations are no longer possible then /proc/vmallocinfo allows to see which compound page allocations fell back to virtual compound pages. That is important for new users of virtual compound pages. Such as order 1 stack allocation etc that may fallback to virtual compound pages in the future. /proc/vmallocinfo permissions are made readable-only-by-root to avoid possible information leakage. [akpm@linux-foundation.org: coding-style fixes] [akpm@linux-foundation.org: CONFIG_MMU=n build fix] Signed-off-by: Christoph Lameter Reviewed-by: KOSAKI Motohiro Cc: Hugh Dickins Cc: Nick Piggin Cc: Arjan van de Ven Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/proc_misc.c | 17 +++++++++++++++++ 1 file changed, 17 insertions(+) (limited to 'fs') diff --git a/fs/proc/proc_misc.c b/fs/proc/proc_misc.c index 2d563979cb0..441a32f0e5f 100644 --- a/fs/proc/proc_misc.c +++ b/fs/proc/proc_misc.c @@ -456,6 +456,20 @@ static const struct file_operations proc_slabstats_operations = { #endif #endif +#ifdef CONFIG_MMU +static int vmalloc_open(struct inode *inode, struct file *file) +{ + return seq_open(file, &vmalloc_op); +} + +static const struct file_operations proc_vmalloc_operations = { + .open = vmalloc_open, + .read = seq_read, + .llseek = seq_lseek, + .release = seq_release, +}; +#endif + static int show_stat(struct seq_file *p, void *v) { int i; @@ -868,6 +882,9 @@ void __init proc_misc_init(void) #ifdef CONFIG_DEBUG_SLAB_LEAK create_seq_entry("slab_allocators", 0 ,&proc_slabstats_operations); #endif +#endif +#ifdef CONFIG_MMU + proc_create("vmallocinfo", S_IRUSR, NULL, &proc_vmalloc_operations); #endif create_seq_entry("buddyinfo",S_IRUGO, &fragmentation_file_operations); create_seq_entry("pagetypeinfo", S_IRUGO, &pagetypeinfo_file_ops); -- cgit v1.2.3 From 214e471ff99064726b2d8af3aa0e24a73c775531 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Mon, 28 Apr 2008 02:12:55 -0700 Subject: smaps: account swap entries Show the amount of swap for each vma. This can be used to see where all the swap goes. [akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Peter Zijlstra Acked-by: Matt Mackall Acked-by: KOSAKI Motohiro Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/proc/task_mmu.c | 16 ++++++++++++---- 1 file changed, 12 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c index f4ab76c7c66..7415eeb7cc3 100644 --- a/fs/proc/task_mmu.c +++ b/fs/proc/task_mmu.c @@ -338,8 +338,7 @@ const struct file_operations proc_maps_operations = { #define PSS_SHIFT 12 #ifdef CONFIG_PROC_PAGE_MONITOR -struct mem_size_stats -{ +struct mem_size_stats { struct vm_area_struct *vma; unsigned long resident; unsigned long shared_clean; @@ -347,6 +346,7 @@ struct mem_size_stats unsigned long private_clean; unsigned long private_dirty; unsigned long referenced; + unsigned long swap; u64 pss; }; @@ -363,6 +363,12 @@ static int smaps_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end, pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl); for (; addr != end; pte++, addr += PAGE_SIZE) { ptent = *pte; + + if (is_swap_pte(ptent)) { + mss->swap += PAGE_SIZE; + continue; + } + if (!pte_present(ptent)) continue; @@ -421,7 +427,8 @@ static int show_smap(struct seq_file *m, void *v) "Shared_Dirty: %8lu kB\n" "Private_Clean: %8lu kB\n" "Private_Dirty: %8lu kB\n" - "Referenced: %8lu kB\n", + "Referenced: %8lu kB\n" + "Swap: %8lu kB\n", (vma->vm_end - vma->vm_start) >> 10, mss.resident >> 10, (unsigned long)(mss.pss >> (10 + PSS_SHIFT)), @@ -429,7 +436,8 @@ static int show_smap(struct seq_file *m, void *v) mss.shared_dirty >> 10, mss.private_clean >> 10, mss.private_dirty >> 10, - mss.referenced >> 10); + mss.referenced >> 10, + mss.swap >> 10); return ret; } -- cgit v1.2.3 From 30afcb4bd2762fa4b87b17ada9500aa46dc10b1b Mon Sep 17 00:00:00 2001 From: Jared Hulbert Date: Mon, 28 Apr 2008 02:13:02 -0700 Subject: return pfn from direct_access, for XIP Alter the block device ->direct_access() API to work with the new get_xip_mem() API (that requires both kaddr and pfn are returned). Some architectures will not do the right thing in their virt_to_page() for use by XIP (to translate from the kernel virtual address returned by direct_access(), to a user mappable pfn in XIP's page fault handler. However, we can't switch it to just return the pfn and not the kaddr, because we have no good way to get a kva from a pfn, and XIP requires the kva for its read(2) and write(2) handlers. So we have to return both. Signed-off-by: Jared Hulbert Signed-off-by: Nick Piggin Cc: Carsten Otte Cc: Heiko Carstens Cc: linux-mm@kvack.org Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext2/xip.c | 24 ++++++++++++++---------- 1 file changed, 14 insertions(+), 10 deletions(-) (limited to 'fs') diff --git a/fs/ext2/xip.c b/fs/ext2/xip.c index ca7f0031238..430b4c8ee97 100644 --- a/fs/ext2/xip.c +++ b/fs/ext2/xip.c @@ -16,11 +16,13 @@ static inline int __inode_direct_access(struct inode *inode, sector_t sector, - unsigned long *data) + void **kaddr, unsigned long *pfn) { - BUG_ON(!inode->i_sb->s_bdev->bd_disk->fops->direct_access); - return inode->i_sb->s_bdev->bd_disk->fops - ->direct_access(inode->i_sb->s_bdev,sector,data); + struct block_device *bdev = inode->i_sb->s_bdev; + struct block_device_operations *ops = bdev->bd_disk->fops; + + BUG_ON(!ops->direct_access); + return ops->direct_access(bdev, sector, kaddr, pfn); } static inline int @@ -48,12 +50,13 @@ int ext2_clear_xip_target(struct inode *inode, int block) { sector_t sector = block * (PAGE_SIZE/512); - unsigned long data; + void *kaddr; + unsigned long pfn; int rc; - rc = __inode_direct_access(inode, sector, &data); + rc = __inode_direct_access(inode, sector, &kaddr, &pfn); if (!rc) - clear_page((void*)data); + clear_page(kaddr); return rc; } @@ -74,7 +77,8 @@ ext2_get_xip_page(struct address_space *mapping, sector_t offset, int create) { int rc; - unsigned long data; + void *kaddr; + unsigned long pfn; sector_t sector; /* first, retrieve the sector number */ @@ -84,9 +88,9 @@ ext2_get_xip_page(struct address_space *mapping, sector_t offset, /* retrieve address of the target data */ rc = __inode_direct_access - (mapping->host, sector * (PAGE_SIZE/512), &data); + (mapping->host, sector * (PAGE_SIZE/512), &kaddr, &pfn); if (!rc) - return virt_to_page(data); + return pfn_to_page(pfn); error: return ERR_PTR(rc); -- cgit v1.2.3 From 70688e4dd1647f0ceb502bbd5964fa344c5eb411 Mon Sep 17 00:00:00 2001 From: Nick Piggin Date: Mon, 28 Apr 2008 02:13:02 -0700 Subject: xip: support non-struct page backed memory Convert XIP to support non-struct page backed memory, using VM_MIXEDMAP for the user mappings. This requires the get_xip_page API to be changed to an address based one. Improve the API layering a little bit too, while we're here. This is required in order to support XIP filesystems on memory that isn't backed with struct page (but memory with struct page is still supported too). Signed-off-by: Nick Piggin Acked-by: Carsten Otte Cc: Jared Hulbert Cc: Martin Schwidefsky Cc: Heiko Carstens Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext2/inode.c | 2 +- fs/ext2/xip.c | 37 +++++++++++++++---------------------- fs/ext2/xip.h | 9 +++++---- fs/open.c | 2 +- 4 files changed, 22 insertions(+), 28 deletions(-) (limited to 'fs') diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c index b8a2990bab8..687023bdfd1 100644 --- a/fs/ext2/inode.c +++ b/fs/ext2/inode.c @@ -796,7 +796,7 @@ const struct address_space_operations ext2_aops = { const struct address_space_operations ext2_aops_xip = { .bmap = ext2_bmap, - .get_xip_page = ext2_get_xip_page, + .get_xip_mem = ext2_get_xip_mem, }; const struct address_space_operations ext2_nobh_aops = { diff --git a/fs/ext2/xip.c b/fs/ext2/xip.c index 430b4c8ee97..233f7fdbe31 100644 --- a/fs/ext2/xip.c +++ b/fs/ext2/xip.c @@ -15,26 +15,28 @@ #include "xip.h" static inline int -__inode_direct_access(struct inode *inode, sector_t sector, +__inode_direct_access(struct inode *inode, sector_t block, void **kaddr, unsigned long *pfn) { struct block_device *bdev = inode->i_sb->s_bdev; struct block_device_operations *ops = bdev->bd_disk->fops; + sector_t sector; + + sector = block * (PAGE_SIZE / 512); /* ext2 block to bdev sector */ BUG_ON(!ops->direct_access); return ops->direct_access(bdev, sector, kaddr, pfn); } static inline int -__ext2_get_sector(struct inode *inode, sector_t offset, int create, +__ext2_get_block(struct inode *inode, pgoff_t pgoff, int create, sector_t *result) { struct buffer_head tmp; int rc; memset(&tmp, 0, sizeof(struct buffer_head)); - rc = ext2_get_block(inode, offset/ (PAGE_SIZE/512), &tmp, - create); + rc = ext2_get_block(inode, pgoff, &tmp, create); *result = tmp.b_blocknr; /* did we get a sparse block (hole in the file)? */ @@ -47,14 +49,13 @@ __ext2_get_sector(struct inode *inode, sector_t offset, int create, } int -ext2_clear_xip_target(struct inode *inode, int block) +ext2_clear_xip_target(struct inode *inode, sector_t block) { - sector_t sector = block * (PAGE_SIZE/512); void *kaddr; unsigned long pfn; int rc; - rc = __inode_direct_access(inode, sector, &kaddr, &pfn); + rc = __inode_direct_access(inode, block, &kaddr, &pfn); if (!rc) clear_page(kaddr); return rc; @@ -72,26 +73,18 @@ void ext2_xip_verify_sb(struct super_block *sb) } } -struct page * -ext2_get_xip_page(struct address_space *mapping, sector_t offset, - int create) +int ext2_get_xip_mem(struct address_space *mapping, pgoff_t pgoff, int create, + void **kmem, unsigned long *pfn) { int rc; - void *kaddr; - unsigned long pfn; - sector_t sector; + sector_t block; /* first, retrieve the sector number */ - rc = __ext2_get_sector(mapping->host, offset, create, §or); + rc = __ext2_get_block(mapping->host, pgoff, create, &block); if (rc) - goto error; + return rc; /* retrieve address of the target data */ - rc = __inode_direct_access - (mapping->host, sector * (PAGE_SIZE/512), &kaddr, &pfn); - if (!rc) - return pfn_to_page(pfn); - - error: - return ERR_PTR(rc); + rc = __inode_direct_access(mapping->host, block, kmem, pfn); + return rc; } diff --git a/fs/ext2/xip.h b/fs/ext2/xip.h index aa85331d6c5..18b34d2f31b 100644 --- a/fs/ext2/xip.h +++ b/fs/ext2/xip.h @@ -7,19 +7,20 @@ #ifdef CONFIG_EXT2_FS_XIP extern void ext2_xip_verify_sb (struct super_block *); -extern int ext2_clear_xip_target (struct inode *, int); +extern int ext2_clear_xip_target (struct inode *, sector_t); static inline int ext2_use_xip (struct super_block *sb) { struct ext2_sb_info *sbi = EXT2_SB(sb); return (sbi->s_mount_opt & EXT2_MOUNT_XIP); } -struct page* ext2_get_xip_page (struct address_space *, sector_t, int); -#define mapping_is_xip(map) unlikely(map->a_ops->get_xip_page) +int ext2_get_xip_mem(struct address_space *, pgoff_t, int, + void **, unsigned long *); +#define mapping_is_xip(map) unlikely(map->a_ops->get_xip_mem) #else #define mapping_is_xip(map) 0 #define ext2_xip_verify_sb(sb) do { } while (0) #define ext2_use_xip(sb) 0 #define ext2_clear_xip_target(inode, chain) 0 -#define ext2_get_xip_page NULL +#define ext2_get_xip_mem NULL #endif diff --git a/fs/open.c b/fs/open.c index b70e7666bb2..7af1f05d597 100644 --- a/fs/open.c +++ b/fs/open.c @@ -837,7 +837,7 @@ static struct file *__dentry_open(struct dentry *dentry, struct vfsmount *mnt, if (f->f_flags & O_DIRECT) { if (!f->f_mapping->a_ops || ((!f->f_mapping->a_ops->direct_IO) && - (!f->f_mapping->a_ops->get_xip_page))) { + (!f->f_mapping->a_ops->get_xip_mem))) { fput(f); f = ERR_PTR(-EINVAL); } -- cgit v1.2.3 From 71fe804b6d56d6a7aed680e096901434cef6a2c3 Mon Sep 17 00:00:00 2001 From: Lee Schermerhorn Date: Mon, 28 Apr 2008 02:13:26 -0700 Subject: mempolicy: use struct mempolicy pointer in shmem_sb_info This patch replaces the mempolicy mode, mode_flags, and nodemask in the shmem_sb_info struct with a struct mempolicy pointer, initialized to NULL. This removes dependency on the details of mempolicy from shmem.c and hugetlbfs inode.c and simplifies the interfaces. mpol_parse_str() in mempolicy.c is changed to return, via a pointer to a pointer arg, a struct mempolicy pointer on success. For MPOL_DEFAULT, the returned pointer is NULL. Further, mpol_parse_str() now takes a 'no_context' argument that causes the input nodemask to be stored in the w.user_nodemask of the created mempolicy for use when the mempolicy is installed in a tmpfs inode shared policy tree. At that time, any cpuset contextualization is applied to the original input nodemask. This preserves the previous behavior where the input nodemask was stored in the superblock. We can think of the returned mempolicy as "context free". Because mpol_parse_str() is now calling mpol_new(), we can remove from mpol_to_str() the semantic checks that mpol_new() already performs. Add 'no_context' parameter to mpol_to_str() to specify that it should format the nodemask in w.user_nodemask for 'bind' and 'interleave' policies. Change mpol_shared_policy_init() to take a pointer to a "context free" struct mempolicy and to create a new, "contextualized" mempolicy using the mode, mode_flags and user_nodemask from the input mempolicy. Note: we know that the mempolicy passed to mpol_to_str() or mpol_shared_policy_init() from a tmpfs superblock is "context free". This is currently the only instance thereof. However, if we found more uses for this concept, and introduced any ambiguity as to whether a mempolicy was context free or not, we could add another internal mode flag to identify context free mempolicies. Then, we could remove the 'no_context' argument from mpol_to_str(). Added shmem_get_sbmpol() to return a reference counted superblock mempolicy, if one exists, to pass to mpol_shared_policy_init(). We must add the reference under the sb stat_lock to prevent races with replacement of the mpol by remount. This reference is removed in mpol_shared_policy_init(). [akpm@linux-foundation.org: build fix] [akpm@linux-foundation.org: another build fix] [akpm@linux-foundation.org: yet another build fix] Signed-off-by: Lee Schermerhorn Cc: Christoph Lameter Cc: David Rientjes Cc: Mel Gorman Cc: Andi Kleen Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/hugetlbfs/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c index 2e9e5bdd562..9783723e8ff 100644 --- a/fs/hugetlbfs/inode.c +++ b/fs/hugetlbfs/inode.c @@ -504,7 +504,7 @@ static struct inode *hugetlbfs_get_inode(struct super_block *sb, uid_t uid, inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME; INIT_LIST_HEAD(&inode->i_mapping->private_list); info = HUGETLBFS_I(inode); - mpol_shared_policy_init(&info->policy, MPOL_DEFAULT, 0, NULL); + mpol_shared_policy_init(&info->policy, NULL); switch (mode & S_IFMT) { default: init_special_inode(inode, mode, dev); -- cgit v1.2.3 From cdf8803768db6f652d498628fe1421a23c025253 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Mon, 28 Apr 2008 02:13:59 -0700 Subject: ncpfs: add prototypes to ncp_fs.h Removes some externs from C files, noticed from the sparse warnings: fs/ncpfs/dir.c:90:26: warning: symbol 'ncp_root_dentry_operations' was not declared. Should it be static? fs/ncpfs/symlink.c:107:5: warning: symbol 'ncp_symlink' was not declared. Should it be static? fs/ncpfs/symlink.c:101:39: warning: symbol 'ncp_symlink_aops' was not declared. Should it be static? Signed-off-by: Harvey Harrison Acked-by: Petr Vandrovec Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ncpfs/inode.c | 6 ------ 1 file changed, 6 deletions(-) (limited to 'fs') diff --git a/fs/ncpfs/inode.c b/fs/ncpfs/inode.c index fbbb9f7afa1..2e5ab1204de 100644 --- a/fs/ncpfs/inode.c +++ b/fs/ncpfs/inode.c @@ -107,12 +107,6 @@ static const struct super_operations ncp_sops = .show_options = ncp_show_options, }; -extern struct dentry_operations ncp_root_dentry_operations; -#if defined(CONFIG_NCPFS_EXTRAS) || defined(CONFIG_NCPFS_NFS_NS) -extern const struct address_space_operations ncp_symlink_aops; -extern int ncp_symlink(struct inode*, struct dentry*, const char*); -#endif - /* * Fill in the ncpfs-specific information in the inode. */ -- cgit v1.2.3 From 305787e44ebc21d87ab4d4949da5b97d4252aa9b Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Mon, 28 Apr 2008 02:14:00 -0700 Subject: ncpfs: fix sparse warnings in ioctl.c In both cases, these inode variables arebeing used to test the server's root inode against NULL. Change them to s_inode. fs/ncpfs/ioctl.c:391:18: warning: symbol 'inode' shadows an earlier one fs/ncpfs/ioctl.c:264:28: originally declared here fs/ncpfs/ioctl.c:441:17: warning: symbol 'inode' shadows an earlier one fs/ncpfs/ioctl.c:264:28: originally declared here In this case, we are about to return anyway, just reuse result. fs/ncpfs/ioctl.c:521:8: warning: symbol 'result' shadows an earlier one fs/ncpfs/ioctl.c:268:6: originally declared here Signed-off-by: Harvey Harrison Acked-by: Petr Vandrovec Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ncpfs/ioctl.c | 17 ++++++++--------- 1 file changed, 8 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/ncpfs/ioctl.c b/fs/ncpfs/ioctl.c index ad8f167e54b..3a97c95e1ca 100644 --- a/fs/ncpfs/ioctl.c +++ b/fs/ncpfs/ioctl.c @@ -389,11 +389,11 @@ static int __ncp_ioctl(struct inode *inode, struct file *filp, struct dentry* dentry = inode->i_sb->s_root; if (dentry) { - struct inode* inode = dentry->d_inode; + struct inode* s_inode = dentry->d_inode; - if (inode) { - sr.volNumber = NCP_FINFO(inode)->volNumber; - sr.dirEntNum = NCP_FINFO(inode)->dirEntNum; + if (s_inode) { + sr.volNumber = NCP_FINFO(s_inode)->volNumber; + sr.dirEntNum = NCP_FINFO(s_inode)->dirEntNum; sr.namespace = server->name_space[sr.volNumber]; } else DPRINTK("ncpfs: s_root->d_inode==NULL\n"); @@ -439,12 +439,12 @@ static int __ncp_ioctl(struct inode *inode, struct file *filp, dentry = inode->i_sb->s_root; server->root_setuped = 1; if (dentry) { - struct inode* inode = dentry->d_inode; + struct inode* s_inode = dentry->d_inode; if (inode) { - NCP_FINFO(inode)->volNumber = vnum; - NCP_FINFO(inode)->dirEntNum = de; - NCP_FINFO(inode)->DosDirNum = dosde; + NCP_FINFO(s_inode)->volNumber = vnum; + NCP_FINFO(s_inode)->dirEntNum = de; + NCP_FINFO(s_inode)->DosDirNum = dosde; } else DPRINTK("ncpfs: s_root->d_inode==NULL\n"); } else @@ -519,7 +519,6 @@ static int __ncp_ioctl(struct inode *inode, struct file *filp, } { struct ncp_lock_ioctl rqdata; - int result; if (copy_from_user(&rqdata, argp, sizeof(rqdata))) return -EFAULT; -- cgit v1.2.3 From eee3754f5e45bd27e001ea41823bdbcdd0d192d4 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Mon, 28 Apr 2008 02:14:01 -0700 Subject: ncpfs: fix sparse warning in ncpsign_kernel.c We're casting anyway, might as well cast to the correct sign. Specific to i386 (ifdef __i386__) fs/ncpfs/ncpsign_kernel.c:58:23: warning: incorrect type in initializer (different signedness) fs/ncpfs/ncpsign_kernel.c:58:23: expected unsigned int *data2 fs/ncpfs/ncpsign_kernel.c:58:23: got int * Signed-off-by: Harvey Harrison Acked-by: Petr Vandrovec Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ncpfs/ncpsign_kernel.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ncpfs/ncpsign_kernel.c b/fs/ncpfs/ncpsign_kernel.c index 749a18d3359..7c0b5c21e6c 100644 --- a/fs/ncpfs/ncpsign_kernel.c +++ b/fs/ncpfs/ncpsign_kernel.c @@ -55,7 +55,7 @@ static void nwsign(char *r_data1, char *r_data2, char *outdata) { unsigned int w0,w1,w2,w3; static int rbit[4]={0, 2, 1, 3}; #ifdef __i386__ - unsigned int *data2=(int *)r_data2; + unsigned int *data2=(unsigned int *)r_data2; #else unsigned int data2[16]; for (i=0;i<16;i++) -- cgit v1.2.3 From 338bf9afda91ec005a1e9a0de4af0271cc167d56 Mon Sep 17 00:00:00 2001 From: Andrew Perepechko Date: Mon, 28 Apr 2008 02:14:31 -0700 Subject: quota: do not allow setting of quota limits to too high values We should check whether quota limits set via Q_SETQUOTA are not exceeding limits which quota format is able to handle. Signed-off-by: Andrew Perepechko Signed-off-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/dquot.c | 22 +++++++++++++++++----- fs/quota_v1.c | 3 +++ fs/quota_v2.c | 3 +++ 3 files changed, 23 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/dquot.c b/fs/dquot.c index 41b9dbd68b0..24eef582d2a 100644 --- a/fs/dquot.c +++ b/fs/dquot.c @@ -1709,10 +1709,19 @@ int vfs_get_dqblk(struct super_block *sb, int type, qid_t id, struct if_dqblk *d } /* Generic routine for setting common part of quota structure */ -static void do_set_dqblk(struct dquot *dquot, struct if_dqblk *di) +static int do_set_dqblk(struct dquot *dquot, struct if_dqblk *di) { struct mem_dqblk *dm = &dquot->dq_dqb; int check_blim = 0, check_ilim = 0; + struct mem_dqinfo *dqi = &sb_dqopt(dquot->dq_sb)->info[dquot->dq_type]; + + if ((di->dqb_valid & QIF_BLIMITS && + (di->dqb_bhardlimit > dqi->dqi_maxblimit || + di->dqb_bsoftlimit > dqi->dqi_maxblimit)) || + (di->dqb_valid & QIF_ILIMITS && + (di->dqb_ihardlimit > dqi->dqi_maxilimit || + di->dqb_isoftlimit > dqi->dqi_maxilimit))) + return -ERANGE; spin_lock(&dq_data_lock); if (di->dqb_valid & QIF_SPACE) { @@ -1744,7 +1753,7 @@ static void do_set_dqblk(struct dquot *dquot, struct if_dqblk *di) clear_bit(DQ_BLKS_B, &dquot->dq_flags); } else if (!(di->dqb_valid & QIF_BTIME)) /* Set grace only if user hasn't provided his own... */ - dm->dqb_btime = get_seconds() + sb_dqopt(dquot->dq_sb)->info[dquot->dq_type].dqi_bgrace; + dm->dqb_btime = get_seconds() + dqi->dqi_bgrace; } if (check_ilim) { if (!dm->dqb_isoftlimit || dm->dqb_curinodes < dm->dqb_isoftlimit) { @@ -1752,7 +1761,7 @@ static void do_set_dqblk(struct dquot *dquot, struct if_dqblk *di) clear_bit(DQ_INODES_B, &dquot->dq_flags); } else if (!(di->dqb_valid & QIF_ITIME)) /* Set grace only if user hasn't provided his own... */ - dm->dqb_itime = get_seconds() + sb_dqopt(dquot->dq_sb)->info[dquot->dq_type].dqi_igrace; + dm->dqb_itime = get_seconds() + dqi->dqi_igrace; } if (dm->dqb_bhardlimit || dm->dqb_bsoftlimit || dm->dqb_ihardlimit || dm->dqb_isoftlimit) clear_bit(DQ_FAKE_B, &dquot->dq_flags); @@ -1760,21 +1769,24 @@ static void do_set_dqblk(struct dquot *dquot, struct if_dqblk *di) set_bit(DQ_FAKE_B, &dquot->dq_flags); spin_unlock(&dq_data_lock); mark_dquot_dirty(dquot); + + return 0; } int vfs_set_dqblk(struct super_block *sb, int type, qid_t id, struct if_dqblk *di) { struct dquot *dquot; + int rc; mutex_lock(&sb_dqopt(sb)->dqonoff_mutex); if (!(dquot = dqget(sb, id, type))) { mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex); return -ESRCH; } - do_set_dqblk(dquot, di); + rc = do_set_dqblk(dquot, di); dqput(dquot); mutex_unlock(&sb_dqopt(sb)->dqonoff_mutex); - return 0; + return rc; } /* Generic routine for getting common part of quota file information */ diff --git a/fs/quota_v1.c b/fs/quota_v1.c index f3841f23306..a6cf9269105 100644 --- a/fs/quota_v1.c +++ b/fs/quota_v1.c @@ -139,6 +139,9 @@ static int v1_read_file_info(struct super_block *sb, int type) goto out; } ret = 0; + /* limits are stored as unsigned 32-bit data */ + dqopt->info[type].dqi_maxblimit = 0xffffffff; + dqopt->info[type].dqi_maxilimit = 0xffffffff; dqopt->info[type].dqi_igrace = dqblk.dqb_itime ? dqblk.dqb_itime : MAX_IQ_TIME; dqopt->info[type].dqi_bgrace = dqblk.dqb_btime ? dqblk.dqb_btime : MAX_DQ_TIME; out: diff --git a/fs/quota_v2.c b/fs/quota_v2.c index c519a583e68..23b647f25d0 100644 --- a/fs/quota_v2.c +++ b/fs/quota_v2.c @@ -59,6 +59,9 @@ static int v2_read_file_info(struct super_block *sb, int type) sb->s_id); return -1; } + /* limits are stored as unsigned 32-bit data */ + info->dqi_maxblimit = 0xffffffff; + info->dqi_maxilimit = 0xffffffff; info->dqi_bgrace = le32_to_cpu(dinfo.dqi_bgrace); info->dqi_igrace = le32_to_cpu(dinfo.dqi_igrace); info->dqi_flags = le32_to_cpu(dinfo.dqi_flags); -- cgit v1.2.3 From 8794b5b246cf6f67baf57bd9db386e79ca5cac33 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 28 Apr 2008 02:14:32 -0700 Subject: quota: remove superfluous DQUOT_OFF() in fs/namespace.c We don't need to turn quotas off before remounting root ro, because do_remount_sb() already handles this. Signed-off-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/namespace.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'fs') diff --git a/fs/namespace.c b/fs/namespace.c index f48f98110c3..fe376805cf5 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -14,7 +14,6 @@ #include #include #include -#include #include #include #include @@ -1084,7 +1083,6 @@ static int do_umount(struct vfsmount *mnt, int flags) down_write(&sb->s_umount); if (!(sb->s_flags & MS_RDONLY)) { lock_kernel(); - DQUOT_OFF(sb); retval = do_remount_sb(sb, MS_RDONLY, NULL, 0); unlock_kernel(); } -- cgit v1.2.3 From 03f6e92bdd467aed9d7571a571868563ae6ad288 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 28 Apr 2008 02:14:32 -0700 Subject: quota: various style cleanups Cleanups in quota code: Change __inline__ to inline. Change some macros to inline functions. Remove vfs_quota_off_mount() macro. DQUOT_OFF() should be (0) is CONFIG_QUOTA is disabled. Move declaration of mark_dquot_dirty and dirty_dquot from quota.h to dquot.c [akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/dquot.c | 10 +++++++++- fs/reiserfs/super.c | 2 +- 2 files changed, 10 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/dquot.c b/fs/dquot.c index 24eef582d2a..fc26d1097d3 100644 --- a/fs/dquot.c +++ b/fs/dquot.c @@ -289,7 +289,15 @@ static void wait_on_dquot(struct dquot *dquot) mutex_unlock(&dquot->dq_lock); } -#define mark_dquot_dirty(dquot) ((dquot)->dq_sb->dq_op->mark_dirty(dquot)) +static inline int dquot_dirty(struct dquot *dquot) +{ + return test_bit(DQ_MOD_B, &dquot->dq_flags); +} + +static inline int mark_dquot_dirty(struct dquot *dquot) +{ + return dquot->dq_sb->dq_op->mark_dirty(dquot); +} int dquot_mark_dquot_dirty(struct dquot *dquot) { diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index 393cc22c171..3e1972d31e4 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -304,7 +304,7 @@ static int finish_unfinished(struct super_block *s) /* Turn quotas off */ for (i = 0; i < MAXQUOTAS; i++) { if (sb_dqopt(s)->files[i]) - vfs_quota_off_mount(s, i); + vfs_quota_off(s, i); } if (ms_active_set) /* Restore the flag back */ -- cgit v1.2.3 From 0ff5af8340aa6be44220d7237ef4a654314cf795 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 28 Apr 2008 02:14:33 -0700 Subject: quota: quota core changes for quotaon on remount Currently, we just turn quotas off on remount of filesystem to read-only state. The patch below adds necessary framework so that we can turn quotas off on remount RO but we are able to automatically reenable them again when filesystem is remounted to RW state. All we need to do is to keep references to inodes of quota files when remounting RO and using these references to reenable quotas when remounting RW. Signed-off-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/dquot.c | 77 ++++++++++++++++++++++++++++++++++++++++++++++++++++++-------- fs/quota.c | 5 ++-- fs/super.c | 10 ++++++-- 3 files changed, 78 insertions(+), 14 deletions(-) (limited to 'fs') diff --git a/fs/dquot.c b/fs/dquot.c index fc26d1097d3..dfba1623ccc 100644 --- a/fs/dquot.c +++ b/fs/dquot.c @@ -1449,31 +1449,43 @@ static inline void set_enable_flags(struct quota_info *dqopt, int type) switch (type) { case USRQUOTA: dqopt->flags |= DQUOT_USR_ENABLED; + dqopt->flags &= ~DQUOT_USR_SUSPENDED; break; case GRPQUOTA: dqopt->flags |= DQUOT_GRP_ENABLED; + dqopt->flags &= ~DQUOT_GRP_SUSPENDED; break; } } -static inline void reset_enable_flags(struct quota_info *dqopt, int type) +static inline void reset_enable_flags(struct quota_info *dqopt, int type, + int remount) { switch (type) { case USRQUOTA: dqopt->flags &= ~DQUOT_USR_ENABLED; + if (remount) + dqopt->flags |= DQUOT_USR_SUSPENDED; + else + dqopt->flags &= ~DQUOT_USR_SUSPENDED; break; case GRPQUOTA: dqopt->flags &= ~DQUOT_GRP_ENABLED; + if (remount) + dqopt->flags |= DQUOT_GRP_SUSPENDED; + else + dqopt->flags &= ~DQUOT_GRP_SUSPENDED; break; } } + /* * Turn quota off on a device. type == -1 ==> quotaoff for all types (umount) */ -int vfs_quota_off(struct super_block *sb, int type) +int vfs_quota_off(struct super_block *sb, int type, int remount) { - int cnt; + int cnt, ret = 0; struct quota_info *dqopt = sb_dqopt(sb); struct inode *toputinode[MAXQUOTAS]; @@ -1483,9 +1495,17 @@ int vfs_quota_off(struct super_block *sb, int type) toputinode[cnt] = NULL; if (type != -1 && cnt != type) continue; + /* If we keep inodes of quota files after remount and quotaoff + * is called, drop kept inodes. */ + if (!remount && sb_has_quota_suspended(sb, cnt)) { + iput(dqopt->files[cnt]); + dqopt->files[cnt] = NULL; + reset_enable_flags(dqopt, cnt, 0); + continue; + } if (!sb_has_quota_enabled(sb, cnt)) continue; - reset_enable_flags(dqopt, cnt); + reset_enable_flags(dqopt, cnt, remount); /* Note: these are blocking operations */ drop_dquot_ref(sb, cnt); @@ -1501,7 +1521,8 @@ int vfs_quota_off(struct super_block *sb, int type) put_quota_format(dqopt->info[cnt].dqi_format); toputinode[cnt] = dqopt->files[cnt]; - dqopt->files[cnt] = NULL; + if (!remount) + dqopt->files[cnt] = NULL; dqopt->info[cnt].dqi_flags = 0; dqopt->info[cnt].dqi_igrace = 0; dqopt->info[cnt].dqi_bgrace = 0; @@ -1531,12 +1552,19 @@ int vfs_quota_off(struct super_block *sb, int type) mutex_unlock(&toputinode[cnt]->i_mutex); mark_inode_dirty(toputinode[cnt]); } - iput(toputinode[cnt]); mutex_unlock(&dqopt->dqonoff_mutex); + /* On remount RO, we keep the inode pointer so that we + * can reenable quota on the subsequent remount RW. + * But we have better not keep inode pointer when there + * is pending delete on the quota file... */ + if (!remount) + iput(toputinode[cnt]); + else if (!toputinode[cnt]->i_nlink) + ret = -EBUSY; } if (sb->s_bdev) invalidate_bdev(sb->s_bdev); - return 0; + return ret; } /* @@ -1574,7 +1602,8 @@ static int vfs_quota_on_inode(struct inode *inode, int type, int format_id) invalidate_bdev(sb->s_bdev); mutex_lock(&inode->i_mutex); mutex_lock(&dqopt->dqonoff_mutex); - if (sb_has_quota_enabled(sb, type)) { + if (sb_has_quota_enabled(sb, type) || + sb_has_quota_suspended(sb, type)) { error = -EBUSY; goto out_lock; } @@ -1597,6 +1626,7 @@ static int vfs_quota_on_inode(struct inode *inode, int type, int format_id) dqopt->ops[type] = fmt->qf_ops; dqopt->info[type].dqi_format = fmt; + dqopt->info[type].dqi_fmt_id = format_id; INIT_LIST_HEAD(&dqopt->info[type].dqi_dirty_list); mutex_lock(&dqopt->dqio_mutex); if ((error = dqopt->ops[type]->read_file_info(sb, type)) < 0) { @@ -1632,12 +1662,41 @@ out_fmt: return error; } +/* Reenable quotas on remount RW */ +static int vfs_quota_on_remount(struct super_block *sb, int type) +{ + struct quota_info *dqopt = sb_dqopt(sb); + struct inode *inode; + int ret; + + mutex_lock(&dqopt->dqonoff_mutex); + if (!sb_has_quota_suspended(sb, type)) { + mutex_unlock(&dqopt->dqonoff_mutex); + return 0; + } + BUG_ON(sb_has_quota_enabled(sb, type)); + + inode = dqopt->files[type]; + dqopt->files[type] = NULL; + reset_enable_flags(dqopt, type, 0); + mutex_unlock(&dqopt->dqonoff_mutex); + + ret = vfs_quota_on_inode(inode, type, dqopt->info[type].dqi_fmt_id); + iput(inode); + + return ret; +} + /* Actual function called from quotactl() */ -int vfs_quota_on(struct super_block *sb, int type, int format_id, char *path) +int vfs_quota_on(struct super_block *sb, int type, int format_id, char *path, + int remount) { struct nameidata nd; int error; + if (remount) + return vfs_quota_on_remount(sb, type); + error = path_lookup(path, LOOKUP_FOLLOW, &nd); if (error < 0) return error; diff --git a/fs/quota.c b/fs/quota.c index 84f28dd7211..db1cc9f3c7a 100644 --- a/fs/quota.c +++ b/fs/quota.c @@ -69,7 +69,6 @@ static int generic_quotactl_valid(struct super_block *sb, int type, int cmd, qid switch (cmd) { case Q_GETFMT: case Q_GETINFO: - case Q_QUOTAOFF: case Q_SETINFO: case Q_SETQUOTA: case Q_GETQUOTA: @@ -229,12 +228,12 @@ static int do_quotactl(struct super_block *sb, int type, int cmd, qid_t id, void if (IS_ERR(pathname = getname(addr))) return PTR_ERR(pathname); - ret = sb->s_qcop->quota_on(sb, type, id, pathname); + ret = sb->s_qcop->quota_on(sb, type, id, pathname, 0); putname(pathname); return ret; } case Q_QUOTAOFF: - return sb->s_qcop->quota_off(sb, type); + return sb->s_qcop->quota_off(sb, type, 0); case Q_GETFMT: { __u32 fmt; diff --git a/fs/super.c b/fs/super.c index 4798350b2bc..a5a4aca7e22 100644 --- a/fs/super.c +++ b/fs/super.c @@ -179,7 +179,7 @@ void deactivate_super(struct super_block *s) if (atomic_dec_and_lock(&s->s_active, &sb_lock)) { s->s_count -= S_BIAS-1; spin_unlock(&sb_lock); - DQUOT_OFF(s); + DQUOT_OFF(s, 0); down_write(&s->s_umount); fs->kill_sb(s); put_filesystem(fs); @@ -608,6 +608,7 @@ retry: int do_remount_sb(struct super_block *sb, int flags, void *data, int force) { int retval; + int remount_rw; #ifdef CONFIG_BLOCK if (!(flags & MS_RDONLY) && bdev_read_only(sb->s_bdev)) @@ -625,8 +626,11 @@ int do_remount_sb(struct super_block *sb, int flags, void *data, int force) mark_files_ro(sb); else if (!fs_may_remount_ro(sb)) return -EBUSY; - DQUOT_OFF(sb); + retval = DQUOT_OFF(sb, 1); + if (retval < 0 && retval != -ENOSYS) + return -EBUSY; } + remount_rw = !(flags & MS_RDONLY) && (sb->s_flags & MS_RDONLY); if (sb->s_op->remount_fs) { lock_super(sb); @@ -636,6 +640,8 @@ int do_remount_sb(struct super_block *sb, int flags, void *data, int force) return retval; } sb->s_flags = (sb->s_flags & ~MS_RMT_MASK) | (flags & MS_RMT_MASK); + if (remount_rw) + DQUOT_ON_REMOUNT(sb); return 0; } -- cgit v1.2.3 From 2fd83a4f3cd5a725168e3a269746dfce2adfa56a Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 28 Apr 2008 02:14:34 -0700 Subject: quota: ext3: make ext3 handle quotaon on remount Update ext3 handle quotaon on remount RW. [akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/super.c | 19 ++++++++++--------- 1 file changed, 10 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/ext3/super.c b/fs/ext3/super.c index ad536066408..883ff965d98 100644 --- a/fs/ext3/super.c +++ b/fs/ext3/super.c @@ -685,7 +685,8 @@ static int ext3_acquire_dquot(struct dquot *dquot); static int ext3_release_dquot(struct dquot *dquot); static int ext3_mark_dquot_dirty(struct dquot *dquot); static int ext3_write_info(struct super_block *sb, int type); -static int ext3_quota_on(struct super_block *sb, int type, int format_id, char *path); +static int ext3_quota_on(struct super_block *sb, int type, int format_id, + char *path, int remount); static int ext3_quota_on_mount(struct super_block *sb, int type); static ssize_t ext3_quota_read(struct super_block *sb, int type, char *data, size_t len, loff_t off); @@ -1415,7 +1416,7 @@ static void ext3_orphan_cleanup (struct super_block * sb, /* Turn quotas off */ for (i = 0; i < MAXQUOTAS; i++) { if (sb_dqopt(sb)->files[i]) - vfs_quota_off(sb, i); + vfs_quota_off(sb, i, 0); } #endif sb->s_flags = s_flags; /* Restore MS_RDONLY status */ @@ -2743,17 +2744,17 @@ static int ext3_quota_on_mount(struct super_block *sb, int type) * Standard function to be called on quota_on */ static int ext3_quota_on(struct super_block *sb, int type, int format_id, - char *path) + char *path, int remount) { int err; struct nameidata nd; if (!test_opt(sb, QUOTA)) return -EINVAL; - /* Not journalling quota? */ - if (!EXT3_SB(sb)->s_qf_names[USRQUOTA] && - !EXT3_SB(sb)->s_qf_names[GRPQUOTA]) - return vfs_quota_on(sb, type, format_id, path); + /* Not journalling quota or remount? */ + if ((!EXT3_SB(sb)->s_qf_names[USRQUOTA] && + !EXT3_SB(sb)->s_qf_names[GRPQUOTA]) || remount) + return vfs_quota_on(sb, type, format_id, path, remount); err = path_lookup(path, LOOKUP_FOLLOW, &nd); if (err) return err; @@ -2762,13 +2763,13 @@ static int ext3_quota_on(struct super_block *sb, int type, int format_id, path_put(&nd.path); return -EXDEV; } - /* Quotafile not of fs root? */ + /* Quotafile not in fs root? */ if (nd.path.dentry->d_parent->d_inode != sb->s_root->d_inode) printk(KERN_WARNING "EXT3-fs: Quota file not on filesystem root. " "Journalled quota will not work.\n"); path_put(&nd.path); - return vfs_quota_on(sb, type, format_id, path); + return vfs_quota_on(sb, type, format_id, path, remount); } /* Read data from quotafile - avoid pagecache and such because we cannot afford -- cgit v1.2.3 From 6f28e08794749f3431e89302728e612343d9d9e4 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 28 Apr 2008 02:14:34 -0700 Subject: quota: ext4: make ext4 handle quotaon on remount Update ext4 to handle quotaon on remount RW. [akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext4/super.c | 15 ++++++++------- 1 file changed, 8 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 13383ba18f1..c81a8e759ba 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -813,7 +813,8 @@ static int ext4_acquire_dquot(struct dquot *dquot); static int ext4_release_dquot(struct dquot *dquot); static int ext4_mark_dquot_dirty(struct dquot *dquot); static int ext4_write_info(struct super_block *sb, int type); -static int ext4_quota_on(struct super_block *sb, int type, int format_id, char *path); +static int ext4_quota_on(struct super_block *sb, int type, int format_id, + char *path, int remount); static int ext4_quota_on_mount(struct super_block *sb, int type); static ssize_t ext4_quota_read(struct super_block *sb, int type, char *data, size_t len, loff_t off); @@ -1632,7 +1633,7 @@ static void ext4_orphan_cleanup (struct super_block * sb, /* Turn quotas off */ for (i = 0; i < MAXQUOTAS; i++) { if (sb_dqopt(sb)->files[i]) - vfs_quota_off(sb, i); + vfs_quota_off(sb, i, 0); } #endif sb->s_flags = s_flags; /* Restore MS_RDONLY status */ @@ -3143,7 +3144,7 @@ static int ext4_quota_on_mount(struct super_block *sb, int type) * Standard function to be called on quota_on */ static int ext4_quota_on(struct super_block *sb, int type, int format_id, - char *path) + char *path, int remount) { int err; struct nameidata nd; @@ -3151,9 +3152,9 @@ static int ext4_quota_on(struct super_block *sb, int type, int format_id, if (!test_opt(sb, QUOTA)) return -EINVAL; /* Not journalling quota? */ - if (!EXT4_SB(sb)->s_qf_names[USRQUOTA] && - !EXT4_SB(sb)->s_qf_names[GRPQUOTA]) - return vfs_quota_on(sb, type, format_id, path); + if ((!EXT4_SB(sb)->s_qf_names[USRQUOTA] && + !EXT4_SB(sb)->s_qf_names[GRPQUOTA]) || remount) + return vfs_quota_on(sb, type, format_id, path, remount); err = path_lookup(path, LOOKUP_FOLLOW, &nd); if (err) return err; @@ -3168,7 +3169,7 @@ static int ext4_quota_on(struct super_block *sb, int type, int format_id, "EXT4-fs: Quota file not on filesystem root. " "Journalled quota will not work.\n"); path_put(&nd.path); - return vfs_quota_on(sb, type, format_id, path); + return vfs_quota_on(sb, type, format_id, path, remount); } /* Read data from quotafile - avoid pagecache and such because we cannot afford -- cgit v1.2.3 From 1b445a9c21f593be9d3c4ab912359d2c51c371dd Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 28 Apr 2008 02:14:35 -0700 Subject: quota: reiserfs: make reiserfs handle quotaon on remount Update reiserfs to handle quotaon on remount RW. Signed-off-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/reiserfs/super.c | 13 ++++++++----- 1 file changed, 8 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index 3e1972d31e4..f8f9a473e67 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -304,7 +304,7 @@ static int finish_unfinished(struct super_block *s) /* Turn quotas off */ for (i = 0; i < MAXQUOTAS; i++) { if (sb_dqopt(s)->files[i]) - vfs_quota_off(s, i); + vfs_quota_off(s, i, 0); } if (ms_active_set) /* Restore the flag back */ @@ -634,7 +634,7 @@ static int reiserfs_acquire_dquot(struct dquot *); static int reiserfs_release_dquot(struct dquot *); static int reiserfs_mark_dquot_dirty(struct dquot *); static int reiserfs_write_info(struct super_block *, int); -static int reiserfs_quota_on(struct super_block *, int, int, char *); +static int reiserfs_quota_on(struct super_block *, int, int, char *, int); static struct dquot_operations reiserfs_quota_operations = { .initialize = reiserfs_dquot_initialize, @@ -2015,13 +2015,16 @@ static int reiserfs_quota_on_mount(struct super_block *sb, int type) * Standard function to be called on quota_on */ static int reiserfs_quota_on(struct super_block *sb, int type, int format_id, - char *path) + char *path, int remount) { int err; struct nameidata nd; if (!(REISERFS_SB(sb)->s_mount_opt & (1 << REISERFS_QUOTA))) return -EINVAL; + /* No more checks needed? Path and format_id are bogus anyway... */ + if (remount) + return vfs_quota_on(sb, type, format_id, path, 1); err = path_lookup(path, LOOKUP_FOLLOW, &nd); if (err) return err; @@ -2041,7 +2044,7 @@ static int reiserfs_quota_on(struct super_block *sb, int type, int format_id, if (!REISERFS_SB(sb)->s_qf_names[USRQUOTA] && !REISERFS_SB(sb)->s_qf_names[GRPQUOTA]) { path_put(&nd.path); - return vfs_quota_on(sb, type, format_id, path); + return vfs_quota_on(sb, type, format_id, path, 0); } /* Quotafile not of fs root? */ if (nd.path.dentry->d_parent->d_inode != sb->s_root->d_inode) @@ -2049,7 +2052,7 @@ static int reiserfs_quota_on(struct super_block *sb, int type, int format_id, "reiserfs: Quota file not on filesystem root. " "Journalled quota will not work."); path_put(&nd.path); - return vfs_quota_on(sb, type, format_id, path); + return vfs_quota_on(sb, type, format_id, path, 0); } /* Read data from quotafile - avoid pagecache and such because we cannot afford -- cgit v1.2.3 From fba4d3997f4c3dc6de927f7f516168580033ae1a Mon Sep 17 00:00:00 2001 From: Marcin Slusarz Date: Mon, 28 Apr 2008 02:15:59 -0700 Subject: ext2: le*_add_cpu conversion replace all: little_endian_variable = cpu_to_leX(leX_to_cpu(little_endian_variable) + expression_in_cpu_byteorder); with: leX_add_cpu(&little_endian_variable, expression_in_cpu_byteorder); generated with semantic patch Signed-off-by: Marcin Slusarz Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext2/ialloc.c | 12 ++++-------- fs/ext2/super.c | 2 +- fs/ext2/xattr.c | 9 +++------ 3 files changed, 8 insertions(+), 15 deletions(-) (limited to 'fs') diff --git a/fs/ext2/ialloc.c b/fs/ext2/ialloc.c index 08f647d8188..f5974134676 100644 --- a/fs/ext2/ialloc.c +++ b/fs/ext2/ialloc.c @@ -75,11 +75,9 @@ static void ext2_release_inode(struct super_block *sb, int group, int dir) } spin_lock(sb_bgl_lock(EXT2_SB(sb), group)); - desc->bg_free_inodes_count = - cpu_to_le16(le16_to_cpu(desc->bg_free_inodes_count) + 1); + le16_add_cpu(&desc->bg_free_inodes_count, 1); if (dir) - desc->bg_used_dirs_count = - cpu_to_le16(le16_to_cpu(desc->bg_used_dirs_count) - 1); + le16_add_cpu(&desc->bg_used_dirs_count, -1); spin_unlock(sb_bgl_lock(EXT2_SB(sb), group)); if (dir) percpu_counter_dec(&EXT2_SB(sb)->s_dirs_counter); @@ -539,13 +537,11 @@ got: percpu_counter_inc(&sbi->s_dirs_counter); spin_lock(sb_bgl_lock(sbi, group)); - gdp->bg_free_inodes_count = - cpu_to_le16(le16_to_cpu(gdp->bg_free_inodes_count) - 1); + le16_add_cpu(&gdp->bg_free_inodes_count, -1); if (S_ISDIR(mode)) { if (sbi->s_debts[group] < 255) sbi->s_debts[group]++; - gdp->bg_used_dirs_count = - cpu_to_le16(le16_to_cpu(gdp->bg_used_dirs_count) + 1); + le16_add_cpu(&gdp->bg_used_dirs_count, 1); } else { if (sbi->s_debts[group]) sbi->s_debts[group]--; diff --git a/fs/ext2/super.c b/fs/ext2/super.c index 088b011bb97..40d934c3e40 100644 --- a/fs/ext2/super.c +++ b/fs/ext2/super.c @@ -604,7 +604,7 @@ static int ext2_setup_super (struct super_block * sb, "running e2fsck is recommended\n"); if (!le16_to_cpu(es->s_max_mnt_count)) es->s_max_mnt_count = cpu_to_le16(EXT2_DFL_MAX_MNT_COUNT); - es->s_mnt_count=cpu_to_le16(le16_to_cpu(es->s_mnt_count) + 1); + le16_add_cpu(&es->s_mnt_count, 1); ext2_write_super(sb); if (test_opt (sb, DEBUG)) printk ("[EXT II FS %s, %s, bs=%lu, fs=%lu, gc=%lu, " diff --git a/fs/ext2/xattr.c b/fs/ext2/xattr.c index a99d46f3b26..b86f02747ea 100644 --- a/fs/ext2/xattr.c +++ b/fs/ext2/xattr.c @@ -646,8 +646,7 @@ ext2_xattr_set2(struct inode *inode, struct buffer_head *old_bh, unlock_buffer(new_bh); goto cleanup; } - HDR(new_bh)->h_refcount = cpu_to_le32(1 + - le32_to_cpu(HDR(new_bh)->h_refcount)); + le32_add_cpu(&HDR(new_bh)->h_refcount, 1); ea_bdebug(new_bh, "refcount now=%d", le32_to_cpu(HDR(new_bh)->h_refcount)); } @@ -731,8 +730,7 @@ ext2_xattr_set2(struct inode *inode, struct buffer_head *old_bh, bforget(old_bh); } else { /* Decrement the refcount only. */ - HDR(old_bh)->h_refcount = cpu_to_le32( - le32_to_cpu(HDR(old_bh)->h_refcount) - 1); + le32_add_cpu(&HDR(old_bh)->h_refcount, -1); if (ce) mb_cache_entry_release(ce); DQUOT_FREE_BLOCK(inode, 1); @@ -789,8 +787,7 @@ ext2_xattr_delete_inode(struct inode *inode) bforget(bh); unlock_buffer(bh); } else { - HDR(bh)->h_refcount = cpu_to_le32( - le32_to_cpu(HDR(bh)->h_refcount) - 1); + le32_add_cpu(&HDR(bh)->h_refcount, -1); if (ce) mb_cache_entry_release(ce); ea_bdebug(bh, "refcount now=%d", -- cgit v1.2.3 From 31f68e130149809c5d49583e1ff225dd5d404d83 Mon Sep 17 00:00:00 2001 From: Marcin Slusarz Date: Mon, 28 Apr 2008 02:16:00 -0700 Subject: ext2: convert byte order of constant instead of variable Convert byte order of constant instead of variable which can be done at compile time (vs run time). Signed-off-by: Marcin Slusarz Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext2/super.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/ext2/super.c b/fs/ext2/super.c index 40d934c3e40..7e686732922 100644 --- a/fs/ext2/super.c +++ b/fs/ext2/super.c @@ -51,8 +51,7 @@ void ext2_error (struct super_block * sb, const char * function, if (!(sb->s_flags & MS_RDONLY)) { sbi->s_mount_state |= EXT2_ERROR_FS; - es->s_state = - cpu_to_le16(le16_to_cpu(es->s_state) | EXT2_ERROR_FS); + es->s_state |= cpu_to_le16(EXT2_ERROR_FS); ext2_sync_super(sb, es); } @@ -1126,10 +1125,9 @@ void ext2_write_super (struct super_block * sb) if (!(sb->s_flags & MS_RDONLY)) { es = EXT2_SB(sb)->s_es; - if (le16_to_cpu(es->s_state) & EXT2_VALID_FS) { + if (es->s_state & cpu_to_le16(EXT2_VALID_FS)) { ext2_debug ("setting valid to 0\n"); - es->s_state = cpu_to_le16(le16_to_cpu(es->s_state) & - ~EXT2_VALID_FS); + es->s_state &= cpu_to_le16(~EXT2_VALID_FS); es->s_free_blocks_count = cpu_to_le32(ext2_count_free_blocks(sb)); es->s_free_inodes_count = cpu_to_le32(ext2_count_free_inodes(sb)); es->s_mtime = cpu_to_le32(get_seconds()); -- cgit v1.2.3 From bbff28602436cc7ca660757bba81f03b99e8586d Mon Sep 17 00:00:00 2001 From: Akinobu Mita Date: Mon, 28 Apr 2008 02:16:00 -0700 Subject: ext2: improve ext2_readdir() return value Improve ext2_readdir() return value for ext2_get_page() failure by using the actual result of ext2_get_page(). Signed-off-by: Akinobu Mita Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext2/dir.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ext2/dir.c b/fs/ext2/dir.c index 8dededd80fe..a0995eb2dc2 100644 --- a/fs/ext2/dir.c +++ b/fs/ext2/dir.c @@ -299,7 +299,7 @@ ext2_readdir (struct file * filp, void * dirent, filldir_t filldir) "bad page in #%lu", inode->i_ino); filp->f_pos += PAGE_CACHE_SIZE - offset; - return -EIO; + return PTR_ERR(page); } kaddr = page_address(page); if (unlikely(need_revalidate)) { -- cgit v1.2.3 From 24097d12efbb97bff14fb6e350508853db0f9595 Mon Sep 17 00:00:00 2001 From: Akinobu Mita Date: Mon, 28 Apr 2008 02:16:01 -0700 Subject: ext2: use ext2_group_first_block_no() Use ext2_group_first_block_no() and assign the return values to ext2_fsblk_t variables. Signed-off-by: Akinobu Mita Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext2/inode.c | 3 +-- fs/ext2/super.c | 11 +++++------ fs/ext2/xattr.c | 6 ++---- 3 files changed, 8 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c index 687023bdfd1..be0d02ac7ac 100644 --- a/fs/ext2/inode.c +++ b/fs/ext2/inode.c @@ -275,8 +275,7 @@ static unsigned long ext2_find_near(struct inode *inode, Indirect *ind) * It is going to be refered from inode itself? OK, just put it into * the same cylinder group then. */ - bg_start = (ei->i_block_group * EXT2_BLOCKS_PER_GROUP(inode->i_sb)) + - le32_to_cpu(EXT2_SB(inode->i_sb)->s_es->s_first_data_block); + bg_start = ext2_group_first_block_no(inode->i_sb, ei->i_block_group); colour = (current->pid % 16) * (EXT2_BLOCKS_PER_GROUP(inode->i_sb) / 16); return bg_start + colour; diff --git a/fs/ext2/super.c b/fs/ext2/super.c index 7e686732922..3f745fb5fff 100644 --- a/fs/ext2/super.c +++ b/fs/ext2/super.c @@ -621,13 +621,13 @@ static int ext2_check_descriptors(struct super_block *sb) { int i; struct ext2_sb_info *sbi = EXT2_SB(sb); - unsigned long first_block = le32_to_cpu(sbi->s_es->s_first_data_block); - unsigned long last_block; ext2_debug ("Checking group descriptors"); for (i = 0; i < sbi->s_groups_count; i++) { struct ext2_group_desc *gdp = ext2_get_group_desc(sb, i, NULL); + ext2_fsblk_t first_block = ext2_group_first_block_no(sb, i); + ext2_fsblk_t last_block; if (i == sbi->s_groups_count - 1) last_block = le32_to_cpu(sbi->s_es->s_blocks_count) - 1; @@ -663,7 +663,6 @@ static int ext2_check_descriptors(struct super_block *sb) i, (unsigned long) le32_to_cpu(gdp->bg_inode_table)); return 0; } - first_block += EXT2_BLOCKS_PER_GROUP(sb); } return 1; } @@ -720,10 +719,9 @@ static unsigned long descriptor_loc(struct super_block *sb, int nr) { struct ext2_sb_info *sbi = EXT2_SB(sb); - unsigned long bg, first_data_block, first_meta_bg; + unsigned long bg, first_meta_bg; int has_super = 0; - first_data_block = le32_to_cpu(sbi->s_es->s_first_data_block); first_meta_bg = le32_to_cpu(sbi->s_es->s_first_meta_bg); if (!EXT2_HAS_INCOMPAT_FEATURE(sb, EXT2_FEATURE_INCOMPAT_META_BG) || @@ -732,7 +730,8 @@ static unsigned long descriptor_loc(struct super_block *sb, bg = sbi->s_desc_per_block * nr; if (ext2_bg_has_super(sb, bg)) has_super = 1; - return (first_data_block + has_super + (bg * sbi->s_blocks_per_group)); + + return ext2_group_first_block_no(sb, bg) + has_super; } static int ext2_fill_super(struct super_block *sb, void *data, int silent) diff --git a/fs/ext2/xattr.c b/fs/ext2/xattr.c index b86f02747ea..987a5261cc2 100644 --- a/fs/ext2/xattr.c +++ b/fs/ext2/xattr.c @@ -659,10 +659,8 @@ ext2_xattr_set2(struct inode *inode, struct buffer_head *old_bh, ext2_xattr_cache_insert(new_bh); } else { /* We need to allocate a new block */ - int goal = le32_to_cpu(EXT2_SB(sb)->s_es-> - s_first_data_block) + - EXT2_I(inode)->i_block_group * - EXT2_BLOCKS_PER_GROUP(sb); + ext2_fsblk_t goal = ext2_group_first_block_no(sb, + EXT2_I(inode)->i_block_group); int block = ext2_new_block(inode, goal, &error); if (error) goto cleanup; -- cgit v1.2.3 From 4c8b3125f8dcefd1ee2990113fe69cfa9e9fa609 Mon Sep 17 00:00:00 2001 From: Akinobu Mita Date: Mon, 28 Apr 2008 02:16:02 -0700 Subject: ext2: use ext2_fsblk_t type Use ext2_fsblk_t type for filesystem-wide blocks number Signed-off-by: Akinobu Mita Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext2/inode.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c index be0d02ac7ac..384fc0d1dd7 100644 --- a/fs/ext2/inode.c +++ b/fs/ext2/inode.c @@ -254,13 +254,13 @@ no_block: * Caller must make sure that @ind is valid and will stay that way. */ -static unsigned long ext2_find_near(struct inode *inode, Indirect *ind) +static ext2_fsblk_t ext2_find_near(struct inode *inode, Indirect *ind) { struct ext2_inode_info *ei = EXT2_I(inode); __le32 *start = ind->bh ? (__le32 *) ind->bh->b_data : ei->i_data; __le32 *p; - unsigned long bg_start; - unsigned long colour; + ext2_fsblk_t bg_start; + ext2_fsblk_t colour; /* Try to find previous block */ for (p = ind->p - 1; p >= start; p--) @@ -290,8 +290,8 @@ static unsigned long ext2_find_near(struct inode *inode, Indirect *ind) * Returns preferred place for a block (the goal). */ -static inline int ext2_find_goal(struct inode *inode, long block, - Indirect *partial) +static inline ext2_fsblk_t ext2_find_goal(struct inode *inode, long block, + Indirect *partial) { struct ext2_block_alloc_info *block_i; -- cgit v1.2.3 From 2c11619a590e13fecc801c9c50859cba29740f38 Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Mon, 28 Apr 2008 02:16:02 -0700 Subject: fs/ext2: use BUG_ON if (...) BUG(); should be replaced with BUG_ON(...) when the test has no side-effects to allow a definition of BUG_ON that drops the code completely. The semantic patch that makes this change is as follows: (http://www.emn.fr/x-info/coccinelle/) // @ disable unlikely @ expression E,f; @@ ( if (<... f(...) ...>) { BUG(); } | - if (unlikely(E)) { BUG(); } + BUG_ON(E); ) @@ expression E,f; @@ ( if (<... f(...) ...>) { BUG(); } | - if (E) { BUG(); } + BUG_ON(E); ) // Signed-off-by: Julia Lawall Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext2/balloc.c | 3 +-- fs/ext2/dir.c | 4 ++-- 2 files changed, 3 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/ext2/balloc.c b/fs/ext2/balloc.c index e7b2bafa1dd..c834f512dce 100644 --- a/fs/ext2/balloc.c +++ b/fs/ext2/balloc.c @@ -245,8 +245,7 @@ restart: prev = rsv; } printk("Window map complete.\n"); - if (bad) - BUG(); + BUG_ON(bad); } #define rsv_window_dump(root, verbose) \ __rsv_window_dump((root), (verbose), __FUNCTION__) diff --git a/fs/ext2/dir.c b/fs/ext2/dir.c index a0995eb2dc2..dd9dbb25eab 100644 --- a/fs/ext2/dir.c +++ b/fs/ext2/dir.c @@ -41,8 +41,8 @@ static inline __le16 ext2_rec_len_to_disk(unsigned len) { if (len == (1 << 16)) return cpu_to_le16(EXT2_MAX_REC_LEN); - else if (len > (1 << 16)) - BUG(); + else + BUG_ON(len > (1 << 16)); return cpu_to_le16(len); } -- cgit v1.2.3 From 605afd60ef6dde254fc9b8667c5bccebade251c3 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Mon, 28 Apr 2008 02:16:03 -0700 Subject: ext2: replace remaining __FUNCTION__ occurrences __FUNCTION__ is gcc-specific, use __func__ Signed-off-by: Harvey Harrison Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext2/balloc.c | 10 +++++----- fs/ext2/dir.c | 14 +++++++------- fs/ext2/super.c | 6 +++--- fs/ext2/xip.c | 2 +- 4 files changed, 16 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/ext2/balloc.c b/fs/ext2/balloc.c index c834f512dce..25dec8633c9 100644 --- a/fs/ext2/balloc.c +++ b/fs/ext2/balloc.c @@ -106,7 +106,7 @@ static int ext2_valid_block_bitmap(struct super_block *sb, return 1; err_out: - ext2_error(sb, __FUNCTION__, + ext2_error(sb, __func__, "Invalid block bitmap - " "block_group = %d, block = %lu", block_group, bitmap_blk); @@ -132,7 +132,7 @@ read_block_bitmap(struct super_block *sb, unsigned int block_group) bitmap_blk = le32_to_cpu(desc->bg_block_bitmap); bh = sb_getblk(sb, bitmap_blk); if (unlikely(!bh)) { - ext2_error(sb, __FUNCTION__, + ext2_error(sb, __func__, "Cannot read block bitmap - " "block_group = %d, block_bitmap = %u", block_group, le32_to_cpu(desc->bg_block_bitmap)); @@ -143,7 +143,7 @@ read_block_bitmap(struct super_block *sb, unsigned int block_group) if (bh_submit_read(bh) < 0) { brelse(bh); - ext2_error(sb, __FUNCTION__, + ext2_error(sb, __func__, "Cannot read block bitmap - " "block_group = %d, block_bitmap = %u", block_group, le32_to_cpu(desc->bg_block_bitmap)); @@ -248,7 +248,7 @@ restart: BUG_ON(bad); } #define rsv_window_dump(root, verbose) \ - __rsv_window_dump((root), (verbose), __FUNCTION__) + __rsv_window_dump((root), (verbose), __func__) #else #define rsv_window_dump(root, verbose) do {} while (0) #endif @@ -547,7 +547,7 @@ do_more: for (i = 0, group_freed = 0; i < count; i++) { if (!ext2_clear_bit_atomic(sb_bgl_lock(sbi, block_group), bit + i, bitmap_bh->b_data)) { - ext2_error(sb, __FUNCTION__, + ext2_error(sb, __func__, "bit already cleared for block %lu", block + i); } else { group_freed++; diff --git a/fs/ext2/dir.c b/fs/ext2/dir.c index dd9dbb25eab..a78c6b4af06 100644 --- a/fs/ext2/dir.c +++ b/fs/ext2/dir.c @@ -295,7 +295,7 @@ ext2_readdir (struct file * filp, void * dirent, filldir_t filldir) struct page *page = ext2_get_page(inode, n); if (IS_ERR(page)) { - ext2_error(sb, __FUNCTION__, + ext2_error(sb, __func__, "bad page in #%lu", inode->i_ino); filp->f_pos += PAGE_CACHE_SIZE - offset; @@ -314,7 +314,7 @@ ext2_readdir (struct file * filp, void * dirent, filldir_t filldir) limit = kaddr + ext2_last_byte(inode, n) - EXT2_DIR_REC_LEN(1); for ( ;(char*)de <= limit; de = ext2_next_entry(de)) { if (de->rec_len == 0) { - ext2_error(sb, __FUNCTION__, + ext2_error(sb, __func__, "zero-length directory entry"); ext2_put_page(page); return -EIO; @@ -381,7 +381,7 @@ struct ext2_dir_entry_2 * ext2_find_entry (struct inode * dir, kaddr += ext2_last_byte(dir, n) - reclen; while ((char *) de <= kaddr) { if (de->rec_len == 0) { - ext2_error(dir->i_sb, __FUNCTION__, + ext2_error(dir->i_sb, __func__, "zero-length directory entry"); ext2_put_page(page); goto out; @@ -396,7 +396,7 @@ struct ext2_dir_entry_2 * ext2_find_entry (struct inode * dir, n = 0; /* next page is past the blocks we've got */ if (unlikely(n > (dir->i_blocks >> (PAGE_CACHE_SHIFT - 9)))) { - ext2_error(dir->i_sb, __FUNCTION__, + ext2_error(dir->i_sb, __func__, "dir %lu size %lld exceeds block count %llu", dir->i_ino, dir->i_size, (unsigned long long)dir->i_blocks); @@ -506,7 +506,7 @@ int ext2_add_link (struct dentry *dentry, struct inode *inode) goto got_it; } if (de->rec_len == 0) { - ext2_error(dir->i_sb, __FUNCTION__, + ext2_error(dir->i_sb, __func__, "zero-length directory entry"); err = -EIO; goto out_unlock; @@ -578,7 +578,7 @@ int ext2_delete_entry (struct ext2_dir_entry_2 * dir, struct page * page ) while ((char*)de < (char*)dir) { if (de->rec_len == 0) { - ext2_error(inode->i_sb, __FUNCTION__, + ext2_error(inode->i_sb, __func__, "zero-length directory entry"); err = -EIO; goto out; @@ -670,7 +670,7 @@ int ext2_empty_dir (struct inode * inode) while ((char *)de <= kaddr) { if (de->rec_len == 0) { - ext2_error(inode->i_sb, __FUNCTION__, + ext2_error(inode->i_sb, __func__, "zero-length directory entry"); printk("kaddr=%p, de=%p\n", kaddr, de); goto not_empty; diff --git a/fs/ext2/super.c b/fs/ext2/super.c index 3f745fb5fff..ef50cbc792d 100644 --- a/fs/ext2/super.c +++ b/fs/ext2/super.c @@ -89,7 +89,7 @@ void ext2_update_dynamic_rev(struct super_block *sb) if (le32_to_cpu(es->s_rev_level) > EXT2_GOOD_OLD_REV) return; - ext2_warning(sb, __FUNCTION__, + ext2_warning(sb, __func__, "updating to rev %d because of new feature flag, " "running e2fsck is recommended", EXT2_DYNAMIC_REV); @@ -1060,7 +1060,7 @@ static int ext2_fill_super(struct super_block *sb, void *data, int silent) goto failed_mount3; } if (EXT2_HAS_COMPAT_FEATURE(sb, EXT3_FEATURE_COMPAT_HAS_JOURNAL)) - ext2_warning(sb, __FUNCTION__, + ext2_warning(sb, __func__, "mounting ext3 filesystem as ext2"); ext2_setup_super (sb, es, sb->s_flags & MS_RDONLY); return 0; @@ -1177,7 +1177,7 @@ static int ext2_remount (struct super_block * sb, int * flags, char * data) if (((sbi->s_mount_opt & EXT2_MOUNT_XIP) != (old_mount_opt & EXT2_MOUNT_XIP)) && invalidate_inodes(sb)) - ext2_warning(sb, __FUNCTION__, "busy inodes while remounting "\ + ext2_warning(sb, __func__, "busy inodes while remounting "\ "xip remain in cache (no functional problem)"); if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY)) return 0; diff --git a/fs/ext2/xip.c b/fs/ext2/xip.c index 233f7fdbe31..4fb94c20041 100644 --- a/fs/ext2/xip.c +++ b/fs/ext2/xip.c @@ -68,7 +68,7 @@ void ext2_xip_verify_sb(struct super_block *sb) if ((sbi->s_mount_opt & EXT2_MOUNT_XIP) && !sb->s_bdev->bd_disk->fops->direct_access) { sbi->s_mount_opt &= (~EXT2_MOUNT_XIP); - ext2_warning(sb, __FUNCTION__, + ext2_warning(sb, __func__, "ignoring xip option - not supported by bdev"); } } -- cgit v1.2.3 From 8b91582500ae750db22bd515379616e5e9ad06ab Mon Sep 17 00:00:00 2001 From: "Aneesh Kumar K.V" Date: Mon, 28 Apr 2008 02:16:04 -0700 Subject: ext2: retry block allocation if new blocks are allocated from system zone If the block allocator gets blocks out of system zone ext2 calls ext2_error. But if the file system is mounted with errors=continue retry block allocation. We need to mark the system zone blocks as in use to make sure retry don't pick them again System zone is the block range mapping block bitmap, inode bitmap and inode table. [akpm@linux-foundation.org: fix typo in comment] Signed-off-by: Aneesh Kumar K.V Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext2/balloc.c | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/ext2/balloc.c b/fs/ext2/balloc.c index 25dec8633c9..10bb02c3f25 100644 --- a/fs/ext2/balloc.c +++ b/fs/ext2/balloc.c @@ -149,11 +149,12 @@ read_block_bitmap(struct super_block *sb, unsigned int block_group) block_group, le32_to_cpu(desc->bg_block_bitmap)); return NULL; } - if (!ext2_valid_block_bitmap(sb, desc, block_group, bh)) { - brelse(bh); - return NULL; - } + ext2_valid_block_bitmap(sb, desc, block_group, bh); + /* + * file system mounted not to panic on error, continue with corrupt + * bitmap + */ return bh; } @@ -1380,7 +1381,12 @@ allocated: "Allocating block in system zone - " "blocks from "E2FSBLK", length %lu", ret_block, num); - goto out; + /* + * ext2_try_to_allocate marked the blocks we allocated as in + * use. So we may want to selectively mark some of the blocks + * as free + */ + goto retry_alloc; } performed_allocation = 1; -- cgit v1.2.3 From 3d61f75eefedf75914ab4453c67aaa2ee64bcf93 Mon Sep 17 00:00:00 2001 From: Hisashi Hifumi Date: Mon, 28 Apr 2008 02:16:05 -0700 Subject: ext3: fdatasync should skip metadata writeout when overwriting Currently fdatasync is identical to fsync in ext3. I think fdatasync should skip journal flush in data=ordered and data=writeback mode when it overwrites to already-instantiated blocks on HDD. When I_DIRTY_DATASYNC flag is not set, fdatasync should skip journal writeout because this indicates only atime or/and mtime updates. Following patch is the same approach of ext2's fsync code(ext2_sync_file). I did a performance test using the sysbench. #sysbench --num-threads=128 --max-requests=50000 --test=fileio --file-total-size=128G --file-test-mode=rndwr --file-fsync-mode=fdatasync run The result on ext3 was: -2.6.24 Operations performed: 0 Read, 50080 Write, 59600 Other = 109680 Total Read 0b Written 782.5Mb Total transferred 782.5Mb (12.116Mb/sec) 775.45 Requests/sec executed Test execution summary: total time: 64.5814s total number of events: 50080 total time taken by event execution: 3713.9836 per-request statistics: min: 0.0000s avg: 0.0742s max: 0.9375s approx. 95 percentile: 0.2901s Threads fairness: events (avg/stddev): 391.2500/23.26 execution time (avg/stddev): 29.0155/1.99 -2.6.24-patched Operations performed: 0 Read, 50009 Write, 61596 Other = 111605 Total Read 0b Written 781.39Mb Total transferred 781.39Mb (16.419Mb/sec) 1050.83 Requests/sec executed Test execution summary: total time: 47.5900s total number of events: 50009 total time taken by event execution: 2934.5768 per-request statistics: min: 0.0000s avg: 0.0587s max: 0.8938s approx. 95 percentile: 0.1993s Threads fairness: events (avg/stddev): 390.6953/22.64 execution time (avg/stddev): 22.9264/1.17 Filesystem I/O throughput was improved. Signed-off-by :Hisashi Hifumi Acked-by: Jan Kara Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/fsync.c | 3 +++ 1 file changed, 3 insertions(+) (limited to 'fs') diff --git a/fs/ext3/fsync.c b/fs/ext3/fsync.c index a588e23841d..d33634119e1 100644 --- a/fs/ext3/fsync.c +++ b/fs/ext3/fsync.c @@ -72,6 +72,9 @@ int ext3_sync_file(struct file * file, struct dentry *dentry, int datasync) goto out; } + if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) + goto out; + /* * The VFS has written the file data. If the inode is unaltered * then we need not start a commit. -- cgit v1.2.3 From e7f23ebdef879226817ce94ae6e298afc8cd093d Mon Sep 17 00:00:00 2001 From: Marcin Slusarz Date: Mon, 28 Apr 2008 02:16:06 -0700 Subject: ext3: convert byte order of constant instead of variable Convert byte order of constant instead of variable which can be done at compile time (vs run time). Signed-off-by: Marcin Slusarz Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/super.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ext3/super.c b/fs/ext3/super.c index 883ff965d98..d7f242a93b7 100644 --- a/fs/ext3/super.c +++ b/fs/ext3/super.c @@ -1219,7 +1219,7 @@ static int ext3_setup_super(struct super_block *sb, struct ext3_super_block *es, inconsistencies, to force a fsck at reboot. But for a plain journaled filesystem we can keep it set as valid forever! :) */ - es->s_state = cpu_to_le16(le16_to_cpu(es->s_state) & ~EXT3_VALID_FS); + es->s_state &= cpu_to_le16(~EXT3_VALID_FS); #endif if (!(__s16) le16_to_cpu(es->s_max_mnt_count)) es->s_max_mnt_count = cpu_to_le16(EXT3_DFL_MAX_MNT_COUNT); -- cgit v1.2.3 From 15633005e07883b57c6c7ca539c32148c3a7f588 Mon Sep 17 00:00:00 2001 From: Adrian Bunk Date: Mon, 28 Apr 2008 02:16:06 -0700 Subject: make ext3_xattr_list() static Make the needlessly global ext3_xattr_list() static. Signed-off-by: Adrian Bunk Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/xattr.c | 4 +++- fs/ext3/xattr.h | 7 ------- 2 files changed, 3 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/ext3/xattr.c b/fs/ext3/xattr.c index 42856541e9a..8516831709b 100644 --- a/fs/ext3/xattr.c +++ b/fs/ext3/xattr.c @@ -99,6 +99,8 @@ static struct buffer_head *ext3_xattr_cache_find(struct inode *, struct mb_cache_entry **); static void ext3_xattr_rehash(struct ext3_xattr_header *, struct ext3_xattr_entry *); +static int ext3_xattr_list(struct inode *inode, char *buffer, + size_t buffer_size); static struct mb_cache *ext3_xattr_cache; @@ -427,7 +429,7 @@ cleanup: * Returns a negative error number on failure, or the number of bytes * used / required on success. */ -int +static int ext3_xattr_list(struct inode *inode, char *buffer, size_t buffer_size) { int i_error, b_error; diff --git a/fs/ext3/xattr.h b/fs/ext3/xattr.h index 6b1ae1c6182..148a4dfc82a 100644 --- a/fs/ext3/xattr.h +++ b/fs/ext3/xattr.h @@ -67,7 +67,6 @@ extern struct xattr_handler ext3_xattr_security_handler; extern ssize_t ext3_listxattr(struct dentry *, char *, size_t); extern int ext3_xattr_get(struct inode *, int, const char *, void *, size_t); -extern int ext3_xattr_list(struct inode *, char *, size_t); extern int ext3_xattr_set(struct inode *, int, const char *, const void *, size_t, int); extern int ext3_xattr_set_handle(handle_t *, struct inode *, int, const char *, const void *, size_t, int); @@ -88,12 +87,6 @@ ext3_xattr_get(struct inode *inode, int name_index, const char *name, return -EOPNOTSUPP; } -static inline int -ext3_xattr_list(struct inode *inode, void *buffer, size_t size) -{ - return -EOPNOTSUPP; -} - static inline int ext3_xattr_set(struct inode *inode, int name_index, const char *name, const void *value, size_t size, int flags) -- cgit v1.2.3 From 1eaafeae4b6f87eabdbabe3277826696f4ca196f Mon Sep 17 00:00:00 2001 From: Akinobu Mita Date: Mon, 28 Apr 2008 02:16:07 -0700 Subject: ext3: use ext3_group_first_block_no() Use ext3_group_first_block_no() Signed-off-by: Akinobu Mita Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/super.c | 5 ++--- fs/ext3/xattr.c | 6 ++---- 2 files changed, 4 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/ext3/super.c b/fs/ext3/super.c index d7f242a93b7..770bbb4fa58 100644 --- a/fs/ext3/super.c +++ b/fs/ext3/super.c @@ -1254,14 +1254,14 @@ static int ext3_setup_super(struct super_block *sb, struct ext3_super_block *es, static int ext3_check_descriptors(struct super_block *sb) { struct ext3_sb_info *sbi = EXT3_SB(sb); - ext3_fsblk_t first_block = le32_to_cpu(sbi->s_es->s_first_data_block); - ext3_fsblk_t last_block; int i; ext3_debug ("Checking group descriptors"); for (i = 0; i < sbi->s_groups_count; i++) { struct ext3_group_desc *gdp = ext3_get_group_desc(sb, i, NULL); + ext3_fsblk_t first_block = ext3_group_first_block_no(sb, i); + ext3_fsblk_t last_block; if (i == sbi->s_groups_count - 1) last_block = le32_to_cpu(sbi->s_es->s_blocks_count) - 1; @@ -1300,7 +1300,6 @@ static int ext3_check_descriptors(struct super_block *sb) le32_to_cpu(gdp->bg_inode_table)); return 0; } - first_block += EXT3_BLOCKS_PER_GROUP(sb); } sbi->s_es->s_free_blocks_count=cpu_to_le32(ext3_count_free_blocks(sb)); diff --git a/fs/ext3/xattr.c b/fs/ext3/xattr.c index 8516831709b..ca591ad1540 100644 --- a/fs/ext3/xattr.c +++ b/fs/ext3/xattr.c @@ -799,10 +799,8 @@ inserted: get_bh(new_bh); } else { /* We need to allocate a new block */ - ext3_fsblk_t goal = le32_to_cpu( - EXT3_SB(sb)->s_es->s_first_data_block) + - (ext3_fsblk_t)EXT3_I(inode)->i_block_group * - EXT3_BLOCKS_PER_GROUP(sb); + ext3_fsblk_t goal = ext3_group_first_block_no(sb, + EXT3_I(inode)->i_block_group); ext3_fsblk_t block = ext3_new_block(handle, inode, goal, &error); if (error) -- cgit v1.2.3 From 22a5daf5375a900e1a4efe8ffe2daef9be01e873 Mon Sep 17 00:00:00 2001 From: Akinobu Mita Date: Mon, 28 Apr 2008 02:16:07 -0700 Subject: ext3: add missing ext3_journal_stop() Add missing ext3_journal_stop() in error handling. Signed-off-by: Akinobu Mita Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/resize.c | 1 + 1 file changed, 1 insertion(+) (limited to 'fs') diff --git a/fs/ext3/resize.c b/fs/ext3/resize.c index 0e97b6e07cb..7a9dac1d365 100644 --- a/fs/ext3/resize.c +++ b/fs/ext3/resize.c @@ -1049,6 +1049,7 @@ int ext3_group_extend(struct super_block *sb, struct ext3_super_block *es, ext3_warning(sb, __FUNCTION__, "multiple resizers run on filesystem!"); unlock_super(sb); + ext3_journal_stop(handle); err = -EBUSY; goto exit_put; } -- cgit v1.2.3 From e0e369a7dd39894465b6501a9492173e8104a19b Mon Sep 17 00:00:00 2001 From: Akinobu Mita Date: Mon, 28 Apr 2008 02:16:08 -0700 Subject: ext3: use ext3_get_group_desc() Use ext3_get_group_desc() Signed-off-by: Akinobu Mita Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/inode.c | 23 +++++------------------ 1 file changed, 5 insertions(+), 18 deletions(-) (limited to 'fs') diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c index c683609b0e3..e2acbc7a3c3 100644 --- a/fs/ext3/inode.c +++ b/fs/ext3/inode.c @@ -2454,11 +2454,10 @@ out_stop: static ext3_fsblk_t ext3_get_inode_block(struct super_block *sb, unsigned long ino, struct ext3_iloc *iloc) { - unsigned long desc, group_desc, block_group; + unsigned long block_group; unsigned long offset; ext3_fsblk_t block; - struct buffer_head *bh; - struct ext3_group_desc * gdp; + struct ext3_group_desc *gdp; if (!ext3_valid_inum(sb, ino)) { /* @@ -2470,27 +2469,15 @@ static ext3_fsblk_t ext3_get_inode_block(struct super_block *sb, } block_group = (ino - 1) / EXT3_INODES_PER_GROUP(sb); - if (block_group >= EXT3_SB(sb)->s_groups_count) { - ext3_error(sb,"ext3_get_inode_block","group >= groups count"); - return 0; - } - smp_rmb(); - group_desc = block_group >> EXT3_DESC_PER_BLOCK_BITS(sb); - desc = block_group & (EXT3_DESC_PER_BLOCK(sb) - 1); - bh = EXT3_SB(sb)->s_group_desc[group_desc]; - if (!bh) { - ext3_error (sb, "ext3_get_inode_block", - "Descriptor not loaded"); + gdp = ext3_get_group_desc(sb, block_group, NULL); + if (!gdp) return 0; - } - - gdp = (struct ext3_group_desc *)bh->b_data; /* * Figure out the offset within the block group inode table */ offset = ((ino - 1) % EXT3_INODES_PER_GROUP(sb)) * EXT3_INODE_SIZE(sb); - block = le32_to_cpu(gdp[desc].bg_inode_table) + + block = le32_to_cpu(gdp->bg_inode_table) + (offset >> EXT3_BLOCK_SIZE_BITS(sb)); iloc->block_group = block_group; -- cgit v1.2.3 From 33575f8ffe99bf9e381161ccd76b39079c0aa92f Mon Sep 17 00:00:00 2001 From: Akinobu Mita Date: Mon, 28 Apr 2008 02:16:08 -0700 Subject: ext3: check ext3_journal_get_write_access() errors Check ext3_journal_get_write_access() errors. Signed-off-by: Akinobu Mita Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/namei.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ext3/namei.c b/fs/ext3/namei.c index dec3e0d88ab..ba55865caaa 100644 --- a/fs/ext3/namei.c +++ b/fs/ext3/namei.c @@ -57,10 +57,15 @@ static struct buffer_head *ext3_append(handle_t *handle, *block = inode->i_size >> inode->i_sb->s_blocksize_bits; - if ((bh = ext3_bread(handle, inode, *block, 1, err))) { + bh = ext3_bread(handle, inode, *block, 1, err); + if (bh) { inode->i_size += inode->i_sb->s_blocksize; EXT3_I(inode)->i_disksize = inode->i_size; - ext3_journal_get_write_access(handle,bh); + *err = ext3_journal_get_write_access(handle, bh); + if (*err) { + brelse(bh); + bh = NULL; + } } return bh; } -- cgit v1.2.3 From 269b26191650be46ce6c91dec24cf20f59650529 Mon Sep 17 00:00:00 2001 From: Julia Lawall Date: Mon, 28 Apr 2008 02:16:09 -0700 Subject: fs/ext3: use BUG_ON if (...) BUG(); should be replaced with BUG_ON(...) when the test has no side-effects to allow a definition of BUG_ON that drops the code completely. The semantic patch that makes this change is as follows: (http://www.emn.fr/x-info/coccinelle/) // @ disable unlikely @ expression E,f; @@ ( if (<... f(...) ...>) { BUG(); } | - if (unlikely(E)) { BUG(); } + BUG_ON(E); ) @@ expression E,f; @@ ( if (<... f(...) ...>) { BUG(); } | - if (E) { BUG(); } + BUG_ON(E); ) // Signed-off-by: Julia Lawall Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/balloc.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ext3/balloc.c b/fs/ext3/balloc.c index da0cb2c0e43..dd4237e5e9e 100644 --- a/fs/ext3/balloc.c +++ b/fs/ext3/balloc.c @@ -232,8 +232,7 @@ restart: prev = rsv; } printk("Window map complete.\n"); - if (bad) - BUG(); + BUG_ON(bad); } #define rsv_window_dump(root, verbose) \ __rsv_window_dump((root), (verbose), __FUNCTION__) -- cgit v1.2.3 From 5bc833feaa8b2236265764e7e81f44937be46eda Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Mon, 28 Apr 2008 02:16:10 -0700 Subject: jbd: fix the way the b_modified flag is cleared Currently at the start of a journal commit we loop through all of the buffers on the committing transaction and clear the b_modified flag (the flag that is set when a transaction modifies the buffer) under the j_list_lock. The problem is that everywhere else this flag is modified only under the jbd lock buffer flag, so it will race with a running transaction who could potentially set it, and have it unset by the committing transaction. This is also a big waste, you can have several thousands of buffers that you are clearing the modified flag on when you may not need to. This patch removes this code and instead clears the b_modified flag upon entering do_get_write_access/journal_get_create_access, so if that transaction does indeed use the buffer then it will be accounted for properly, and if it does not then we know we didn't use it. That will be important for the next patch in this series. Tested thoroughly by myself using postmark/iozone/bonnie++. Signed-off-by: Josef Bacik Cc: Acked-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/jbd/commit.c | 16 ---------------- fs/jbd/transaction.c | 13 +++++++++++++ 2 files changed, 13 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/jbd/commit.c b/fs/jbd/commit.c index a38c7186c57..d2f652e6e22 100644 --- a/fs/jbd/commit.c +++ b/fs/jbd/commit.c @@ -406,22 +406,6 @@ void journal_commit_transaction(journal_t *journal) jbd_debug (3, "JBD: commit phase 2\n"); - /* - * First, drop modified flag: all accesses to the buffers - * will be tracked for a new trasaction only -bzzz - */ - spin_lock(&journal->j_list_lock); - if (commit_transaction->t_buffers) { - new_jh = jh = commit_transaction->t_buffers->b_tnext; - do { - J_ASSERT_JH(new_jh, new_jh->b_modified == 1 || - new_jh->b_modified == 0); - new_jh->b_modified = 0; - new_jh = new_jh->b_tnext; - } while (new_jh != jh); - } - spin_unlock(&journal->j_list_lock); - /* * Now start flushing things to disk, in the order they appear * on the transaction lists. Data blocks go first. diff --git a/fs/jbd/transaction.c b/fs/jbd/transaction.c index 2c9e8f5d13a..2e9670e2237 100644 --- a/fs/jbd/transaction.c +++ b/fs/jbd/transaction.c @@ -608,6 +608,12 @@ repeat: jh->b_next_transaction == transaction) goto done; + /* + * this is the first time this transaction is touching this buffer, + * reset the modified flag + */ + jh->b_modified = 0; + /* * If there is already a copy-out version of this buffer, then we don't * need to make another one @@ -820,9 +826,16 @@ int journal_get_create_access(handle_t *handle, struct buffer_head *bh) if (jh->b_transaction == NULL) { jh->b_transaction = transaction; + + /* first access by this transaction */ + jh->b_modified = 0; + JBUFFER_TRACE(jh, "file as BJ_Reserved"); __journal_file_buffer(jh, transaction, BJ_Reserved); } else if (jh->b_transaction == journal->j_committing_transaction) { + /* first access by this transaction */ + jh->b_modified = 0; + JBUFFER_TRACE(jh, "set next transaction"); jh->b_next_transaction = transaction; } -- cgit v1.2.3 From 5b9a499d77e9dd39c9e6611ea10c56a31604f274 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Mon, 28 Apr 2008 02:16:12 -0700 Subject: jbd: fix possible journal overflow issues There are several cases where the running transaction can get buffers added to its BJ_Metadata list which it never dirtied, which makes its t_nr_buffers counter end up larger than its t_outstanding_credits counter. This will cause issues when starting new transactions as while we are logging buffers we decrement t_outstanding_buffers, so when t_outstanding_buffers goes negative, we will report that we need less space in the journal than we actually need, so transactions will be started even though there may not be enough room for them. In the worst case scenario (which admittedly is almost impossible to reproduce) this will result in the journal running out of space. The fix is to only refile buffers from the committing transaction to the running transactions BJ_Modified list when b_modified is set on that journal, which is the only way to be sure if the running transaction has modified that buffer. This patch also fixes an accounting error in journal_forget, it is possible that we can call journal_forget on a buffer without having modified it, only gotten write access to it, so instead of freeing a credit, we only do so if the buffer was modified. The assert will help catch if this problem occurs. Without these two patches I could hit this assert within minutes of running postmark, with them this issue no longer arises. Thank you, Signed-off-by: Josef Bacik Cc: Acked-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/jbd/commit.c | 3 +++ fs/jbd/transaction.c | 21 ++++++++++++++++++--- 2 files changed, 21 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/jbd/commit.c b/fs/jbd/commit.c index d2f652e6e22..cd931ef1f00 100644 --- a/fs/jbd/commit.c +++ b/fs/jbd/commit.c @@ -472,6 +472,9 @@ void journal_commit_transaction(journal_t *journal) */ commit_transaction->t_state = T_COMMIT; + J_ASSERT(commit_transaction->t_nr_buffers <= + commit_transaction->t_outstanding_credits); + descriptor = NULL; bufs = 0; while (commit_transaction->t_buffers) { diff --git a/fs/jbd/transaction.c b/fs/jbd/transaction.c index 2e9670e2237..d8dd3d915fd 100644 --- a/fs/jbd/transaction.c +++ b/fs/jbd/transaction.c @@ -1235,6 +1235,7 @@ int journal_forget (handle_t *handle, struct buffer_head *bh) struct journal_head *jh; int drop_reserve = 0; int err = 0; + int was_modified = 0; BUFFER_TRACE(bh, "entry"); @@ -1253,6 +1254,9 @@ int journal_forget (handle_t *handle, struct buffer_head *bh) goto not_jbd; } + /* keep track of wether or not this transaction modified us */ + was_modified = jh->b_modified; + /* * The buffer's going from the transaction, we must drop * all references -bzzz @@ -1270,7 +1274,12 @@ int journal_forget (handle_t *handle, struct buffer_head *bh) JBUFFER_TRACE(jh, "belongs to current transaction: unfile"); - drop_reserve = 1; + /* + * we only want to drop a reference if this transaction + * modified the buffer + */ + if (was_modified) + drop_reserve = 1; /* * We are no longer going to journal this buffer. @@ -1310,7 +1319,13 @@ int journal_forget (handle_t *handle, struct buffer_head *bh) if (jh->b_next_transaction) { J_ASSERT(jh->b_next_transaction == transaction); jh->b_next_transaction = NULL; - drop_reserve = 1; + + /* + * only drop a reference if this transaction modified + * the buffer + */ + if (was_modified) + drop_reserve = 1; } } @@ -2082,7 +2097,7 @@ void __journal_refile_buffer(struct journal_head *jh) jh->b_transaction = jh->b_next_transaction; jh->b_next_transaction = NULL; __journal_file_buffer(jh, jh->b_transaction, - was_dirty ? BJ_Metadata : BJ_Reserved); + jh->b_modified ? BJ_Metadata : BJ_Reserved); J_ASSERT_JH(jh, jh->b_transaction->t_state == T_RUNNING); if (was_dirty) -- cgit v1.2.3 From 0b23076988b44b2c165e060248345de6f2337387 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 28 Apr 2008 02:16:12 -0700 Subject: ext3: fix update of mtime and ctime on rename Make ext3 update mtime and ctime of the directory into which we move file even if the directory entry already exists. Signed-off-by: Jan Kara Cc: Al Viro Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/namei.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/ext3/namei.c b/fs/ext3/namei.c index ba55865caaa..1a106572d63 100644 --- a/fs/ext3/namei.c +++ b/fs/ext3/namei.c @@ -2323,6 +2323,8 @@ static int ext3_rename (struct inode * old_dir, struct dentry *old_dentry, EXT3_FEATURE_INCOMPAT_FILETYPE)) new_de->file_type = old_de->file_type; new_dir->i_version++; + new_dir->i_ctime = new_dir->i_mtime = CURRENT_TIME_SEC; + ext3_mark_inode_dirty(handle, new_dir); BUFFER_TRACE(new_bh, "call ext3_journal_dirty_metadata"); ext3_journal_dirty_metadata(handle, new_bh); brelse(new_bh); -- cgit v1.2.3 From 07c9938a4e2c92b796b163dc70e99d3d1870aaee Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 28 Apr 2008 02:16:13 -0700 Subject: ext3: fix hang on umount with quotas when journal is aborted Call dquot_drop() from ext3_dquot_drop() even if we fail to start a transaction. Otherwise we never get to dropping references to quota structures from the inode and umount will hang indefinitely. Thanks to Payphone LIOU for spotting the problem. Signed-off-by: Jan Kara Cc: Payphone LIOU Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/super.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ext3/super.c b/fs/ext3/super.c index 770bbb4fa58..03d5c626c1c 100644 --- a/fs/ext3/super.c +++ b/fs/ext3/super.c @@ -2639,8 +2639,14 @@ static int ext3_dquot_drop(struct inode *inode) /* We may delete quota structure so we need to reserve enough blocks */ handle = ext3_journal_start(inode, 2*EXT3_QUOTA_DEL_BLOCKS(inode->i_sb)); - if (IS_ERR(handle)) + if (IS_ERR(handle)) { + /* + * We call dquot_drop() anyway to at least release references + * to quota structures so that umount does not hang. + */ + dquot_drop(inode); return PTR_ERR(handle); + } ret = dquot_drop(inode); err = ext3_journal_stop(handle); if (!ret) -- cgit v1.2.3 From 2588ef83f7933d8ae42868d7bf68fc8a3001186b Mon Sep 17 00:00:00 2001 From: "Aneesh Kumar K.V" Date: Mon, 28 Apr 2008 02:16:14 -0700 Subject: ext3: retry block allocation if new blocks are allocated from system zone If the block allocator gets blocks out of system zone ext3 calls ext3_error. But if the file system is mounted with errors=continue retry block allocation. We need to mark the system zone blocks as in use to make sure retry don't pick them again System zone is the block range mapping block bitmap, inode bitmap and inode table. [akpm@linux-foundation.org: fix typo in comment] Signed-off-by: Aneesh Kumar K.V Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/balloc.c | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/ext3/balloc.c b/fs/ext3/balloc.c index dd4237e5e9e..731fbc0f129 100644 --- a/fs/ext3/balloc.c +++ b/fs/ext3/balloc.c @@ -164,10 +164,11 @@ read_block_bitmap(struct super_block *sb, unsigned int block_group) block_group, le32_to_cpu(desc->bg_block_bitmap)); return NULL; } - if (!ext3_valid_block_bitmap(sb, desc, block_group, bh)) { - brelse(bh); - return NULL; - } + ext3_valid_block_bitmap(sb, desc, block_group, bh); + /* + * file system mounted not to panic on error, continue with corrupt + * bitmap + */ return bh; } /* @@ -1641,7 +1642,11 @@ allocated: "Allocating block in system zone - " "blocks from "E3FSBLK", length %lu", ret_block, num); - goto out; + /* + * claim_block() marked the blocks we allocated as in use. So we + * may want to selectively mark some of the blocks as free. + */ + goto retry_alloc; } performed_allocation = 1; -- cgit v1.2.3 From fa1ff1e02fee908dfdc3f92902d39acc38041e4c Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 28 Apr 2008 02:16:14 -0700 Subject: ext3: fix mount messages when quota disabled When quota is disabled, we should not print 'journaled quota not supported' when user tried to mount non-journaled quota. Also fix typo in the message. Signed-off-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/super.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ext3/super.c b/fs/ext3/super.c index 03d5c626c1c..6303adfbdae 100644 --- a/fs/ext3/super.c +++ b/fs/ext3/super.c @@ -1097,6 +1097,9 @@ clear_qf_name: case Opt_quota: case Opt_usrquota: case Opt_grpquota: + printk(KERN_ERR + "EXT3-fs: quota options not supported.\n"); + break; case Opt_usrjquota: case Opt_grpjquota: case Opt_offusrjquota: @@ -1104,7 +1107,7 @@ clear_qf_name: case Opt_jqfmt_vfsold: case Opt_jqfmt_vfsv0: printk(KERN_ERR - "EXT3-fs: journalled quota options not " + "EXT3-fs: journaled quota options not " "supported.\n"); break; case Opt_noquota: -- cgit v1.2.3 From e05b6b524bd5c5c2bae1b64a7cbe08d46d57a6fe Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Mon, 28 Apr 2008 02:16:15 -0700 Subject: ext3: replace remaining __FUNCTION__ occurrences __FUNCTION__ is gcc-specific, use __func__ Signed-off-by: Harvey Harrison Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ext3/balloc.c | 12 +++++----- fs/ext3/ext3_jbd.c | 12 +++++----- fs/ext3/ialloc.c | 6 ++--- fs/ext3/inode.c | 6 ++--- fs/ext3/namei.c | 26 ++++++++++---------- fs/ext3/resize.c | 70 +++++++++++++++++++++++++++--------------------------- fs/ext3/super.c | 14 +++++------ fs/ext3/xattr.c | 14 +++++------ 8 files changed, 80 insertions(+), 80 deletions(-) (limited to 'fs') diff --git a/fs/ext3/balloc.c b/fs/ext3/balloc.c index 731fbc0f129..92fd0338a6e 100644 --- a/fs/ext3/balloc.c +++ b/fs/ext3/balloc.c @@ -117,7 +117,7 @@ static int ext3_valid_block_bitmap(struct super_block *sb, return 1; err_out: - ext3_error(sb, __FUNCTION__, + ext3_error(sb, __func__, "Invalid block bitmap - " "block_group = %d, block = %lu", block_group, bitmap_blk); @@ -147,7 +147,7 @@ read_block_bitmap(struct super_block *sb, unsigned int block_group) bitmap_blk = le32_to_cpu(desc->bg_block_bitmap); bh = sb_getblk(sb, bitmap_blk); if (unlikely(!bh)) { - ext3_error(sb, __FUNCTION__, + ext3_error(sb, __func__, "Cannot read block bitmap - " "block_group = %d, block_bitmap = %u", block_group, le32_to_cpu(desc->bg_block_bitmap)); @@ -158,7 +158,7 @@ read_block_bitmap(struct super_block *sb, unsigned int block_group) if (bh_submit_read(bh) < 0) { brelse(bh); - ext3_error(sb, __FUNCTION__, + ext3_error(sb, __func__, "Cannot read block bitmap - " "block_group = %d, block_bitmap = %u", block_group, le32_to_cpu(desc->bg_block_bitmap)); @@ -236,7 +236,7 @@ restart: BUG_ON(bad); } #define rsv_window_dump(root, verbose) \ - __rsv_window_dump((root), (verbose), __FUNCTION__) + __rsv_window_dump((root), (verbose), __func__) #else #define rsv_window_dump(root, verbose) do {} while (0) #endif @@ -618,7 +618,7 @@ do_more: if (!ext3_clear_bit_atomic(sb_bgl_lock(sbi, block_group), bit + i, bitmap_bh->b_data)) { jbd_unlock_bh_state(bitmap_bh); - ext3_error(sb, __FUNCTION__, + ext3_error(sb, __func__, "bit already cleared for block "E3FSBLK, block + i); jbd_lock_bh_state(bitmap_bh); @@ -1672,7 +1672,7 @@ allocated: if (ext3_test_bit(grp_alloc_blk+i, bh2jh(bitmap_bh)->b_committed_data)) { printk("%s: block was unexpectedly set in " - "b_committed_data\n", __FUNCTION__); + "b_committed_data\n", __func__); } } } diff --git a/fs/ext3/ext3_jbd.c b/fs/ext3/ext3_jbd.c index e1f91fd26a9..d401f148d74 100644 --- a/fs/ext3/ext3_jbd.c +++ b/fs/ext3/ext3_jbd.c @@ -9,7 +9,7 @@ int __ext3_journal_get_undo_access(const char *where, handle_t *handle, { int err = journal_get_undo_access(handle, bh); if (err) - ext3_journal_abort_handle(where, __FUNCTION__, bh, handle,err); + ext3_journal_abort_handle(where, __func__, bh, handle,err); return err; } @@ -18,7 +18,7 @@ int __ext3_journal_get_write_access(const char *where, handle_t *handle, { int err = journal_get_write_access(handle, bh); if (err) - ext3_journal_abort_handle(where, __FUNCTION__, bh, handle,err); + ext3_journal_abort_handle(where, __func__, bh, handle,err); return err; } @@ -27,7 +27,7 @@ int __ext3_journal_forget(const char *where, handle_t *handle, { int err = journal_forget(handle, bh); if (err) - ext3_journal_abort_handle(where, __FUNCTION__, bh, handle,err); + ext3_journal_abort_handle(where, __func__, bh, handle,err); return err; } @@ -36,7 +36,7 @@ int __ext3_journal_revoke(const char *where, handle_t *handle, { int err = journal_revoke(handle, blocknr, bh); if (err) - ext3_journal_abort_handle(where, __FUNCTION__, bh, handle,err); + ext3_journal_abort_handle(where, __func__, bh, handle,err); return err; } @@ -45,7 +45,7 @@ int __ext3_journal_get_create_access(const char *where, { int err = journal_get_create_access(handle, bh); if (err) - ext3_journal_abort_handle(where, __FUNCTION__, bh, handle,err); + ext3_journal_abort_handle(where, __func__, bh, handle,err); return err; } @@ -54,6 +54,6 @@ int __ext3_journal_dirty_metadata(const char *where, { int err = journal_dirty_metadata(handle, bh); if (err) - ext3_journal_abort_handle(where, __FUNCTION__, bh, handle,err); + ext3_journal_abort_handle(where, __func__, bh, handle,err); return err; } diff --git a/fs/ext3/ialloc.c b/fs/ext3/ialloc.c index 96dd5573e49..77126821b2e 100644 --- a/fs/ext3/ialloc.c +++ b/fs/ext3/ialloc.c @@ -644,7 +644,7 @@ struct inode *ext3_orphan_get(struct super_block *sb, unsigned long ino) /* Error cases - e2fsck has already cleaned up for us */ if (ino > max_ino) { - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "bad orphan ino %lu! e2fsck was run?", ino); goto error; } @@ -653,7 +653,7 @@ struct inode *ext3_orphan_get(struct super_block *sb, unsigned long ino) bit = (ino - 1) % EXT3_INODES_PER_GROUP(sb); bitmap_bh = read_inode_bitmap(sb, block_group); if (!bitmap_bh) { - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "inode bitmap error for orphan %lu", ino); goto error; } @@ -678,7 +678,7 @@ iget_failed: err = PTR_ERR(inode); inode = NULL; bad_orphan: - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "bad orphan inode %lu! e2fsck was run?", ino); printk(KERN_NOTICE "ext3_test_bit(bit=%d, block=%llu) = %d\n", bit, (unsigned long long)bitmap_bh->b_blocknr, diff --git a/fs/ext3/inode.c b/fs/ext3/inode.c index e2acbc7a3c3..cc47b76091b 100644 --- a/fs/ext3/inode.c +++ b/fs/ext3/inode.c @@ -95,7 +95,7 @@ int ext3_forget(handle_t *handle, int is_metadata, struct inode *inode, BUFFER_TRACE(bh, "call ext3_journal_revoke"); err = ext3_journal_revoke(handle, blocknr, bh); if (err) - ext3_abort(inode->i_sb, __FUNCTION__, + ext3_abort(inode->i_sb, __func__, "error %d when attempting revoke", err); BUFFER_TRACE(bh, "exit"); return err; @@ -1190,7 +1190,7 @@ int ext3_journal_dirty_data(handle_t *handle, struct buffer_head *bh) { int err = journal_dirty_data(handle, bh); if (err) - ext3_journal_abort_handle(__FUNCTION__, __FUNCTION__, + ext3_journal_abort_handle(__func__, __func__, bh, handle, err); return err; } @@ -3201,7 +3201,7 @@ void ext3_dirty_inode(struct inode *inode) current_handle->h_transaction != handle->h_transaction) { /* This task has a transaction open against a different fs */ printk(KERN_EMERG "%s: transactions do not match!\n", - __FUNCTION__); + __func__); } else { jbd_debug(5, "marking dirty. outer handle=%p\n", current_handle); diff --git a/fs/ext3/namei.c b/fs/ext3/namei.c index 1a106572d63..0b8cf80154f 100644 --- a/fs/ext3/namei.c +++ b/fs/ext3/namei.c @@ -361,7 +361,7 @@ dx_probe(struct dentry *dentry, struct inode *dir, if (root->info.hash_version != DX_HASH_TEA && root->info.hash_version != DX_HASH_HALF_MD4 && root->info.hash_version != DX_HASH_LEGACY) { - ext3_warning(dir->i_sb, __FUNCTION__, + ext3_warning(dir->i_sb, __func__, "Unrecognised inode hash code %d", root->info.hash_version); brelse(bh); @@ -375,7 +375,7 @@ dx_probe(struct dentry *dentry, struct inode *dir, hash = hinfo->hash; if (root->info.unused_flags & 1) { - ext3_warning(dir->i_sb, __FUNCTION__, + ext3_warning(dir->i_sb, __func__, "Unimplemented inode hash flags: %#06x", root->info.unused_flags); brelse(bh); @@ -384,7 +384,7 @@ dx_probe(struct dentry *dentry, struct inode *dir, } if ((indirect = root->info.indirect_levels) > 1) { - ext3_warning(dir->i_sb, __FUNCTION__, + ext3_warning(dir->i_sb, __func__, "Unimplemented inode hash depth: %#06x", root->info.indirect_levels); brelse(bh); @@ -397,7 +397,7 @@ dx_probe(struct dentry *dentry, struct inode *dir, if (dx_get_limit(entries) != dx_root_limit(dir, root->info.info_length)) { - ext3_warning(dir->i_sb, __FUNCTION__, + ext3_warning(dir->i_sb, __func__, "dx entry: limit != root limit"); brelse(bh); *err = ERR_BAD_DX_DIR; @@ -409,7 +409,7 @@ dx_probe(struct dentry *dentry, struct inode *dir, { count = dx_get_count(entries); if (!count || count > dx_get_limit(entries)) { - ext3_warning(dir->i_sb, __FUNCTION__, + ext3_warning(dir->i_sb, __func__, "dx entry: no count or count > limit"); brelse(bh); *err = ERR_BAD_DX_DIR; @@ -454,7 +454,7 @@ dx_probe(struct dentry *dentry, struct inode *dir, goto fail2; at = entries = ((struct dx_node *) bh->b_data)->entries; if (dx_get_limit(entries) != dx_node_limit (dir)) { - ext3_warning(dir->i_sb, __FUNCTION__, + ext3_warning(dir->i_sb, __func__, "dx entry: limit != node limit"); brelse(bh); *err = ERR_BAD_DX_DIR; @@ -470,7 +470,7 @@ fail2: } fail: if (*err == ERR_BAD_DX_DIR) - ext3_warning(dir->i_sb, __FUNCTION__, + ext3_warning(dir->i_sb, __func__, "Corrupt dir inode %ld, running e2fsck is " "recommended.", dir->i_ino); return NULL; @@ -918,7 +918,7 @@ restart: wait_on_buffer(bh); if (!buffer_uptodate(bh)) { /* read error, skip block & hope for the best */ - ext3_error(sb, __FUNCTION__, "reading directory #%lu " + ext3_error(sb, __func__, "reading directory #%lu " "offset %lu", dir->i_ino, block); brelse(bh); goto next; @@ -1010,7 +1010,7 @@ static struct buffer_head * ext3_dx_find_entry(struct dentry *dentry, retval = ext3_htree_next_block(dir, hash, frame, frames, NULL); if (retval < 0) { - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "error reading index page in directory #%lu", dir->i_ino); *err = retval; @@ -1535,7 +1535,7 @@ static int ext3_dx_add_entry(handle_t *handle, struct dentry *dentry, if (levels && (dx_get_count(frames->entries) == dx_get_limit(frames->entries))) { - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "Directory index full!"); err = -ENOSPC; goto cleanup; @@ -1837,11 +1837,11 @@ static int empty_dir (struct inode * inode) if (inode->i_size < EXT3_DIR_REC_LEN(1) + EXT3_DIR_REC_LEN(2) || !(bh = ext3_bread (NULL, inode, 0, 0, &err))) { if (err) - ext3_error(inode->i_sb, __FUNCTION__, + ext3_error(inode->i_sb, __func__, "error %d reading directory #%lu offset 0", err, inode->i_ino); else - ext3_warning(inode->i_sb, __FUNCTION__, + ext3_warning(inode->i_sb, __func__, "bad directory (dir #%lu) - no data block", inode->i_ino); return 1; @@ -1870,7 +1870,7 @@ static int empty_dir (struct inode * inode) offset >> EXT3_BLOCK_SIZE_BITS(sb), 0, &err); if (!bh) { if (err) - ext3_error(sb, __FUNCTION__, + ext3_error(sb, __func__, "error %d reading directory" " #%lu offset %lu", err, inode->i_ino, offset); diff --git a/fs/ext3/resize.c b/fs/ext3/resize.c index 7a9dac1d365..28cfd0b4052 100644 --- a/fs/ext3/resize.c +++ b/fs/ext3/resize.c @@ -48,60 +48,60 @@ static int verify_group_input(struct super_block *sb, free_blocks_count, input->reserved_blocks); if (group != sbi->s_groups_count) - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "Cannot add at group %u (only %lu groups)", input->group, sbi->s_groups_count); else if ((start - le32_to_cpu(es->s_first_data_block)) % EXT3_BLOCKS_PER_GROUP(sb)) - ext3_warning(sb, __FUNCTION__, "Last group not full"); + ext3_warning(sb, __func__, "Last group not full"); else if (input->reserved_blocks > input->blocks_count / 5) - ext3_warning(sb, __FUNCTION__, "Reserved blocks too high (%u)", + ext3_warning(sb, __func__, "Reserved blocks too high (%u)", input->reserved_blocks); else if (free_blocks_count < 0) - ext3_warning(sb, __FUNCTION__, "Bad blocks count %u", + ext3_warning(sb, __func__, "Bad blocks count %u", input->blocks_count); else if (!(bh = sb_bread(sb, end - 1))) - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "Cannot read last block ("E3FSBLK")", end - 1); else if (outside(input->block_bitmap, start, end)) - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "Block bitmap not in group (block %u)", input->block_bitmap); else if (outside(input->inode_bitmap, start, end)) - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "Inode bitmap not in group (block %u)", input->inode_bitmap); else if (outside(input->inode_table, start, end) || outside(itend - 1, start, end)) - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "Inode table not in group (blocks %u-"E3FSBLK")", input->inode_table, itend - 1); else if (input->inode_bitmap == input->block_bitmap) - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "Block bitmap same as inode bitmap (%u)", input->block_bitmap); else if (inside(input->block_bitmap, input->inode_table, itend)) - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "Block bitmap (%u) in inode table (%u-"E3FSBLK")", input->block_bitmap, input->inode_table, itend-1); else if (inside(input->inode_bitmap, input->inode_table, itend)) - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "Inode bitmap (%u) in inode table (%u-"E3FSBLK")", input->inode_bitmap, input->inode_table, itend-1); else if (inside(input->block_bitmap, start, metaend)) - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "Block bitmap (%u) in GDT table" " ("E3FSBLK"-"E3FSBLK")", input->block_bitmap, start, metaend - 1); else if (inside(input->inode_bitmap, start, metaend)) - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "Inode bitmap (%u) in GDT table" " ("E3FSBLK"-"E3FSBLK")", input->inode_bitmap, start, metaend - 1); else if (inside(input->inode_table, start, metaend) || inside(itend - 1, start, metaend)) - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "Inode table (%u-"E3FSBLK") overlaps" "GDT table ("E3FSBLK"-"E3FSBLK")", input->inode_table, itend - 1, start, metaend - 1); @@ -386,7 +386,7 @@ static int verify_reserved_gdb(struct super_block *sb, while ((grp = ext3_list_backups(sb, &three, &five, &seven)) < end) { if (le32_to_cpu(*p++) != grp * EXT3_BLOCKS_PER_GROUP(sb) + blk){ - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "reserved GDT "E3FSBLK " missing grp %d ("E3FSBLK")", blk, grp, @@ -440,7 +440,7 @@ static int add_new_gdb(handle_t *handle, struct inode *inode, */ if (EXT3_SB(sb)->s_sbh->b_blocknr != le32_to_cpu(EXT3_SB(sb)->s_es->s_first_data_block)) { - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "won't resize using backup superblock at %llu", (unsigned long long)EXT3_SB(sb)->s_sbh->b_blocknr); return -EPERM; @@ -464,7 +464,7 @@ static int add_new_gdb(handle_t *handle, struct inode *inode, data = (__le32 *)dind->b_data; if (le32_to_cpu(data[gdb_num % EXT3_ADDR_PER_BLOCK(sb)]) != gdblock) { - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "new group %u GDT block "E3FSBLK" not reserved", input->group, gdblock); err = -EINVAL; @@ -488,7 +488,7 @@ static int add_new_gdb(handle_t *handle, struct inode *inode, GFP_NOFS); if (!n_group_desc) { err = -ENOMEM; - ext3_warning (sb, __FUNCTION__, + ext3_warning (sb, __func__, "not enough memory for %lu groups", gdb_num + 1); goto exit_inode; } @@ -586,7 +586,7 @@ static int reserve_backup_gdb(handle_t *handle, struct inode *inode, /* Get each reserved primary GDT block and verify it holds backups */ for (res = 0; res < reserved_gdb; res++, blk++) { if (le32_to_cpu(*data) != blk) { - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "reserved block "E3FSBLK " not at offset %ld", blk, @@ -730,7 +730,7 @@ static void update_backups(struct super_block *sb, */ exit_err: if (err) { - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "can't update backup for group %d (err %d), " "forcing fsck on next reboot", group, err); sbi->s_mount_state &= ~EXT3_VALID_FS; @@ -770,33 +770,33 @@ int ext3_group_add(struct super_block *sb, struct ext3_new_group_data *input) if (gdb_off == 0 && !EXT3_HAS_RO_COMPAT_FEATURE(sb, EXT3_FEATURE_RO_COMPAT_SPARSE_SUPER)) { - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "Can't resize non-sparse filesystem further"); return -EPERM; } if (le32_to_cpu(es->s_blocks_count) + input->blocks_count < le32_to_cpu(es->s_blocks_count)) { - ext3_warning(sb, __FUNCTION__, "blocks_count overflow\n"); + ext3_warning(sb, __func__, "blocks_count overflow\n"); return -EINVAL; } if (le32_to_cpu(es->s_inodes_count) + EXT3_INODES_PER_GROUP(sb) < le32_to_cpu(es->s_inodes_count)) { - ext3_warning(sb, __FUNCTION__, "inodes_count overflow\n"); + ext3_warning(sb, __func__, "inodes_count overflow\n"); return -EINVAL; } if (reserved_gdb || gdb_off == 0) { if (!EXT3_HAS_COMPAT_FEATURE(sb, EXT3_FEATURE_COMPAT_RESIZE_INODE)){ - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "No reserved GDT blocks, can't resize"); return -EPERM; } inode = ext3_iget(sb, EXT3_RESIZE_INO); if (IS_ERR(inode)) { - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "Error opening resize inode"); return PTR_ERR(inode); } @@ -825,7 +825,7 @@ int ext3_group_add(struct super_block *sb, struct ext3_new_group_data *input) lock_super(sb); if (input->group != sbi->s_groups_count) { - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "multiple resizers run on filesystem!"); err = -EBUSY; goto exit_journal; @@ -988,13 +988,13 @@ int ext3_group_extend(struct super_block *sb, struct ext3_super_block *es, " too large to resize to %lu blocks safely\n", sb->s_id, n_blocks_count); if (sizeof(sector_t) < 8) - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "CONFIG_LBD not enabled\n"); return -EINVAL; } if (n_blocks_count < o_blocks_count) { - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "can't shrink FS - resize aborted"); return -EBUSY; } @@ -1004,7 +1004,7 @@ int ext3_group_extend(struct super_block *sb, struct ext3_super_block *es, EXT3_BLOCKS_PER_GROUP(sb); if (last == 0) { - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "need to use ext2online to resize further"); return -EPERM; } @@ -1012,7 +1012,7 @@ int ext3_group_extend(struct super_block *sb, struct ext3_super_block *es, add = EXT3_BLOCKS_PER_GROUP(sb) - last; if (o_blocks_count + add < o_blocks_count) { - ext3_warning(sb, __FUNCTION__, "blocks_count overflow"); + ext3_warning(sb, __func__, "blocks_count overflow"); return -EINVAL; } @@ -1020,7 +1020,7 @@ int ext3_group_extend(struct super_block *sb, struct ext3_super_block *es, add = n_blocks_count - o_blocks_count; if (o_blocks_count + add < n_blocks_count) - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "will only finish group ("E3FSBLK " blocks, %u new)", o_blocks_count + add, add); @@ -1028,7 +1028,7 @@ int ext3_group_extend(struct super_block *sb, struct ext3_super_block *es, /* See if the device is actually as big as what was requested */ bh = sb_bread(sb, o_blocks_count + add -1); if (!bh) { - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "can't read last block, resize aborted"); return -ENOSPC; } @@ -1040,13 +1040,13 @@ int ext3_group_extend(struct super_block *sb, struct ext3_super_block *es, handle = ext3_journal_start_sb(sb, 3); if (IS_ERR(handle)) { err = PTR_ERR(handle); - ext3_warning(sb, __FUNCTION__, "error %d on journal start",err); + ext3_warning(sb, __func__, "error %d on journal start",err); goto exit_put; } lock_super(sb); if (o_blocks_count != le32_to_cpu(es->s_blocks_count)) { - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "multiple resizers run on filesystem!"); unlock_super(sb); ext3_journal_stop(handle); @@ -1056,7 +1056,7 @@ int ext3_group_extend(struct super_block *sb, struct ext3_super_block *es, if ((err = ext3_journal_get_write_access(handle, EXT3_SB(sb)->s_sbh))) { - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "error %d on journal write access", err); unlock_super(sb); ext3_journal_stop(handle); diff --git a/fs/ext3/super.c b/fs/ext3/super.c index 6303adfbdae..fe3119a71ad 100644 --- a/fs/ext3/super.c +++ b/fs/ext3/super.c @@ -84,7 +84,7 @@ handle_t *ext3_journal_start_sb(struct super_block *sb, int nblocks) * take the FS itself readonly cleanly. */ journal = EXT3_SB(sb)->s_journal; if (is_journal_aborted(journal)) { - ext3_abort(sb, __FUNCTION__, + ext3_abort(sb, __func__, "Detected aborted journal"); return ERR_PTR(-EROFS); } @@ -304,7 +304,7 @@ void ext3_update_dynamic_rev(struct super_block *sb) if (le32_to_cpu(es->s_rev_level) > EXT3_GOOD_OLD_REV) return; - ext3_warning(sb, __FUNCTION__, + ext3_warning(sb, __func__, "updating to rev %d because of new feature flag, " "running e2fsck is recommended", EXT3_DYNAMIC_REV); @@ -1390,7 +1390,7 @@ static void ext3_orphan_cleanup (struct super_block * sb, if (inode->i_nlink) { printk(KERN_DEBUG "%s: truncating inode %lu to %Ld bytes\n", - __FUNCTION__, inode->i_ino, inode->i_size); + __func__, inode->i_ino, inode->i_size); jbd_debug(2, "truncating inode %lu to %Ld bytes\n", inode->i_ino, inode->i_size); ext3_truncate(inode); @@ -1398,7 +1398,7 @@ static void ext3_orphan_cleanup (struct super_block * sb, } else { printk(KERN_DEBUG "%s: deleting unreferenced inode %lu\n", - __FUNCTION__, inode->i_ino); + __func__, inode->i_ino); jbd_debug(2, "deleting unreferenced inode %lu\n", inode->i_ino); nr_orphans++; @@ -2301,9 +2301,9 @@ static void ext3_clear_journal_err(struct super_block * sb, char nbuf[16]; errstr = ext3_decode_error(sb, j_errno, nbuf); - ext3_warning(sb, __FUNCTION__, "Filesystem error recorded " + ext3_warning(sb, __func__, "Filesystem error recorded " "from previous mount: %s", errstr); - ext3_warning(sb, __FUNCTION__, "Marking fs in need of " + ext3_warning(sb, __func__, "Marking fs in need of " "filesystem check."); EXT3_SB(sb)->s_mount_state |= EXT3_ERROR_FS; @@ -2430,7 +2430,7 @@ static int ext3_remount (struct super_block * sb, int * flags, char * data) } if (sbi->s_mount_opt & EXT3_MOUNT_ABORT) - ext3_abort(sb, __FUNCTION__, "Abort forced by user"); + ext3_abort(sb, __func__, "Abort forced by user"); sb->s_flags = (sb->s_flags & ~MS_POSIXACL) | ((sbi->s_mount_opt & EXT3_MOUNT_POSIX_ACL) ? MS_POSIXACL : 0); diff --git a/fs/ext3/xattr.c b/fs/ext3/xattr.c index ca591ad1540..d4a4f0e9ff6 100644 --- a/fs/ext3/xattr.c +++ b/fs/ext3/xattr.c @@ -234,7 +234,7 @@ ext3_xattr_block_get(struct inode *inode, int name_index, const char *name, ea_bdebug(bh, "b_count=%d, refcount=%d", atomic_read(&(bh->b_count)), le32_to_cpu(BHDR(bh)->h_refcount)); if (ext3_xattr_check_block(bh)) { -bad_block: ext3_error(inode->i_sb, __FUNCTION__, +bad_block: ext3_error(inode->i_sb, __func__, "inode %lu: bad block "E3FSBLK, inode->i_ino, EXT3_I(inode)->i_file_acl); error = -EIO; @@ -376,7 +376,7 @@ ext3_xattr_block_list(struct inode *inode, char *buffer, size_t buffer_size) ea_bdebug(bh, "b_count=%d, refcount=%d", atomic_read(&(bh->b_count)), le32_to_cpu(BHDR(bh)->h_refcount)); if (ext3_xattr_check_block(bh)) { - ext3_error(inode->i_sb, __FUNCTION__, + ext3_error(inode->i_sb, __func__, "inode %lu: bad block "E3FSBLK, inode->i_ino, EXT3_I(inode)->i_file_acl); error = -EIO; @@ -651,7 +651,7 @@ ext3_xattr_block_find(struct inode *inode, struct ext3_xattr_info *i, atomic_read(&(bs->bh->b_count)), le32_to_cpu(BHDR(bs->bh)->h_refcount)); if (ext3_xattr_check_block(bs->bh)) { - ext3_error(sb, __FUNCTION__, + ext3_error(sb, __func__, "inode %lu: bad block "E3FSBLK, inode->i_ino, EXT3_I(inode)->i_file_acl); error = -EIO; @@ -852,7 +852,7 @@ cleanup_dquot: goto cleanup; bad_block: - ext3_error(inode->i_sb, __FUNCTION__, + ext3_error(inode->i_sb, __func__, "inode %lu: bad block "E3FSBLK, inode->i_ino, EXT3_I(inode)->i_file_acl); goto cleanup; @@ -1081,14 +1081,14 @@ ext3_xattr_delete_inode(handle_t *handle, struct inode *inode) goto cleanup; bh = sb_bread(inode->i_sb, EXT3_I(inode)->i_file_acl); if (!bh) { - ext3_error(inode->i_sb, __FUNCTION__, + ext3_error(inode->i_sb, __func__, "inode %lu: block "E3FSBLK" read error", inode->i_ino, EXT3_I(inode)->i_file_acl); goto cleanup; } if (BHDR(bh)->h_magic != cpu_to_le32(EXT3_XATTR_MAGIC) || BHDR(bh)->h_blocks != cpu_to_le32(1)) { - ext3_error(inode->i_sb, __FUNCTION__, + ext3_error(inode->i_sb, __func__, "inode %lu: bad block "E3FSBLK, inode->i_ino, EXT3_I(inode)->i_file_acl); goto cleanup; @@ -1215,7 +1215,7 @@ again: } bh = sb_bread(inode->i_sb, ce->e_block); if (!bh) { - ext3_error(inode->i_sb, __FUNCTION__, + ext3_error(inode->i_sb, __func__, "inode %lu: block %lu read error", inode->i_ino, (unsigned long) ce->e_block); } else if (le32_to_cpu(BHDR(bh)->h_refcount) >= -- cgit v1.2.3 From 08fc99bfc39d5b394aa27498ac5f14df0ef8e52a Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Mon, 28 Apr 2008 02:16:16 -0700 Subject: jbd: replace remaining __FUNCTION__ occurrences __FUNCTION__ is gcc-specific, use __func__ Signed-off-by: Harvey Harrison Cc: Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/jbd/journal.c | 18 +++++++++--------- fs/jbd/revoke.c | 2 +- fs/jbd/transaction.c | 4 ++-- 3 files changed, 12 insertions(+), 12 deletions(-) (limited to 'fs') diff --git a/fs/jbd/journal.c b/fs/jbd/journal.c index 0e081d5f32e..b99c3b3654c 100644 --- a/fs/jbd/journal.c +++ b/fs/jbd/journal.c @@ -534,7 +534,7 @@ int log_wait_commit(journal_t *journal, tid_t tid) if (!tid_geq(journal->j_commit_request, tid)) { printk(KERN_EMERG "%s: error: j_commit_request=%d, tid=%d\n", - __FUNCTION__, journal->j_commit_request, tid); + __func__, journal->j_commit_request, tid); } spin_unlock(&journal->j_state_lock); #endif @@ -599,7 +599,7 @@ int journal_bmap(journal_t *journal, unsigned long blocknr, printk(KERN_ALERT "%s: journal block not found " "at offset %lu on %s\n", - __FUNCTION__, + __func__, blocknr, bdevname(journal->j_dev, b)); err = -EIO; @@ -728,7 +728,7 @@ journal_t * journal_init_dev(struct block_device *bdev, journal->j_wbuf = kmalloc(n * sizeof(struct buffer_head*), GFP_KERNEL); if (!journal->j_wbuf) { printk(KERN_ERR "%s: Cant allocate bhs for commit thread\n", - __FUNCTION__); + __func__); kfree(journal); journal = NULL; goto out; @@ -782,7 +782,7 @@ journal_t * journal_init_inode (struct inode *inode) journal->j_wbuf = kmalloc(n * sizeof(struct buffer_head*), GFP_KERNEL); if (!journal->j_wbuf) { printk(KERN_ERR "%s: Cant allocate bhs for commit thread\n", - __FUNCTION__); + __func__); kfree(journal); return NULL; } @@ -791,7 +791,7 @@ journal_t * journal_init_inode (struct inode *inode) /* If that failed, give up */ if (err) { printk(KERN_ERR "%s: Cannnot locate journal superblock\n", - __FUNCTION__); + __func__); kfree(journal); return NULL; } @@ -877,7 +877,7 @@ int journal_create(journal_t *journal) */ printk(KERN_EMERG "%s: creation of journal on external device!\n", - __FUNCTION__); + __func__); BUG(); } @@ -1657,7 +1657,7 @@ static struct journal_head *journal_alloc_journal_head(void) jbd_debug(1, "out of memory for journal_head\n"); if (time_after(jiffies, last_warning + 5*HZ)) { printk(KERN_NOTICE "ENOMEM in %s, retrying.\n", - __FUNCTION__); + __func__); last_warning = jiffies; } while (ret == NULL) { @@ -1794,13 +1794,13 @@ static void __journal_remove_journal_head(struct buffer_head *bh) if (jh->b_frozen_data) { printk(KERN_WARNING "%s: freeing " "b_frozen_data\n", - __FUNCTION__); + __func__); jbd_free(jh->b_frozen_data, bh->b_size); } if (jh->b_committed_data) { printk(KERN_WARNING "%s: freeing " "b_committed_data\n", - __FUNCTION__); + __func__); jbd_free(jh->b_committed_data, bh->b_size); } bh->b_private = NULL; diff --git a/fs/jbd/revoke.c b/fs/jbd/revoke.c index d5f8eee7c88..1bb43e987f4 100644 --- a/fs/jbd/revoke.c +++ b/fs/jbd/revoke.c @@ -138,7 +138,7 @@ repeat: oom: if (!journal_oom_retry) return -ENOMEM; - jbd_debug(1, "ENOMEM in %s, retrying\n", __FUNCTION__); + jbd_debug(1, "ENOMEM in %s, retrying\n", __func__); yield(); goto repeat; } diff --git a/fs/jbd/transaction.c b/fs/jbd/transaction.c index d8dd3d915fd..67ff2024c23 100644 --- a/fs/jbd/transaction.c +++ b/fs/jbd/transaction.c @@ -687,7 +687,7 @@ repeat: if (!frozen_buffer) { printk(KERN_EMERG "%s: OOM for frozen_buffer\n", - __FUNCTION__); + __func__); JBUFFER_TRACE(jh, "oom!"); error = -ENOMEM; jbd_lock_bh_state(bh); @@ -904,7 +904,7 @@ repeat: committed_data = jbd_alloc(jh2bh(jh)->b_size, GFP_NOFS); if (!committed_data) { printk(KERN_EMERG "%s: No memory for committed data\n", - __FUNCTION__); + __func__); err = -ENOMEM; goto out; } -- cgit v1.2.3 From 3c5afae2bac974fe2c15cd9439f921526b3ea2d7 Mon Sep 17 00:00:00 2001 From: Marcin Slusarz Date: Mon, 28 Apr 2008 02:16:16 -0700 Subject: ufs: [bl]e*_add_cpu conversion replace all: big/little_endian_variable = cpu_to_[bl]eX([bl]eX_to_cpu(big/little_endian_variable) + expression_in_cpu_byteorder); with: [bl]eX_add_cpu(&big/little_endian_variable, expression_in_cpu_byteorder); generated with semantic patch Signed-off-by: Marcin Slusarz Cc: Evgeniy Dushistov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ufs/swab.h | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/ufs/swab.h b/fs/ufs/swab.h index 1683d2bee61..a4340d0267b 100644 --- a/fs/ufs/swab.h +++ b/fs/ufs/swab.h @@ -80,18 +80,18 @@ static inline void fs32_add(struct super_block *sbp, __fs32 *n, int d) { if (UFS_SB(sbp)->s_bytesex == BYTESEX_LE) - *(__le32 *)n = cpu_to_le32(le32_to_cpu(*(__le32 *)n)+d); + le32_add_cpu((__le32 *)n, d); else - *(__be32 *)n = cpu_to_be32(be32_to_cpu(*(__be32 *)n)+d); + be32_add_cpu((__be32 *)n, d); } static inline void fs32_sub(struct super_block *sbp, __fs32 *n, int d) { if (UFS_SB(sbp)->s_bytesex == BYTESEX_LE) - *(__le32 *)n = cpu_to_le32(le32_to_cpu(*(__le32 *)n)-d); + le32_add_cpu((__le32 *)n, -d); else - *(__be32 *)n = cpu_to_be32(be32_to_cpu(*(__be32 *)n)-d); + be32_add_cpu((__be32 *)n, -d); } static inline u16 @@ -116,18 +116,18 @@ static inline void fs16_add(struct super_block *sbp, __fs16 *n, int d) { if (UFS_SB(sbp)->s_bytesex == BYTESEX_LE) - *(__le16 *)n = cpu_to_le16(le16_to_cpu(*(__le16 *)n)+d); + le16_add_cpu((__le16 *)n, d); else - *(__be16 *)n = cpu_to_be16(be16_to_cpu(*(__be16 *)n)+d); + be16_add_cpu((__be16 *)n, d); } static inline void fs16_sub(struct super_block *sbp, __fs16 *n, int d) { if (UFS_SB(sbp)->s_bytesex == BYTESEX_LE) - *(__le16 *)n = cpu_to_le16(le16_to_cpu(*(__le16 *)n)-d); + le16_add_cpu((__le16 *)n, -d); else - *(__be16 *)n = cpu_to_be16(be16_to_cpu(*(__be16 *)n)-d); + be16_add_cpu((__be16 *)n, -d); } #endif /* _UFS_SWAB_H */ -- cgit v1.2.3 From 9746077a71d01b8d86f5e471a3a2c8f1fa91ef94 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Mon, 28 Apr 2008 02:16:17 -0700 Subject: ufs: replace remaining __FUNCTION__ occurrences __FUNCTION__ is gcc-specific, use __func__ Signed-off-by: Harvey Harrison Cc: Jan Engelhardt Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ufs/balloc.c | 4 ++-- fs/ufs/dir.c | 14 +++++++------- fs/ufs/inode.c | 2 +- fs/ufs/ufs.h | 2 +- 4 files changed, 11 insertions(+), 11 deletions(-) (limited to 'fs') diff --git a/fs/ufs/balloc.c b/fs/ufs/balloc.c index 1e7598fb978..0d9ada17373 100644 --- a/fs/ufs/balloc.c +++ b/fs/ufs/balloc.c @@ -277,7 +277,7 @@ static void ufs_change_blocknr(struct inode *inode, sector_t beg, if (!page)/* it was truncated */ continue; if (IS_ERR(page)) {/* or EIO */ - ufs_error(inode->i_sb, __FUNCTION__, + ufs_error(inode->i_sb, __func__, "read of page %llu failed\n", (unsigned long long)index); continue; @@ -308,7 +308,7 @@ static void ufs_change_blocknr(struct inode *inode, sector_t beg, ll_rw_block(READ, 1, &bh); wait_on_buffer(bh); if (!buffer_uptodate(bh)) { - ufs_error(inode->i_sb, __FUNCTION__, + ufs_error(inode->i_sb, __func__, "read of block failed\n"); break; } diff --git a/fs/ufs/dir.c b/fs/ufs/dir.c index ef563fc8d72..df0bef18742 100644 --- a/fs/ufs/dir.c +++ b/fs/ufs/dir.c @@ -179,7 +179,7 @@ bad_entry: goto fail; Eend: p = (struct ufs_dir_entry *)(kaddr + offs); - ufs_error(sb, __FUNCTION__, + ufs_error(sb, __func__, "entry in directory #%lu spans the page boundary" "offset=%lu", dir->i_ino, (page->index<d_reclen == 0) { - ufs_error(dir->i_sb, __FUNCTION__, + ufs_error(dir->i_sb, __func__, "zero-length directory entry"); ufs_put_page(page); goto out; @@ -356,7 +356,7 @@ int ufs_add_link(struct dentry *dentry, struct inode *inode) goto got_it; } if (de->d_reclen == 0) { - ufs_error(dir->i_sb, __FUNCTION__, + ufs_error(dir->i_sb, __func__, "zero-length directory entry"); err = -EIO; goto out_unlock; @@ -456,7 +456,7 @@ ufs_readdir(struct file *filp, void *dirent, filldir_t filldir) struct page *page = ufs_get_page(inode, n); if (IS_ERR(page)) { - ufs_error(sb, __FUNCTION__, + ufs_error(sb, __func__, "bad page in #%lu", inode->i_ino); filp->f_pos += PAGE_CACHE_SIZE - offset; @@ -475,7 +475,7 @@ ufs_readdir(struct file *filp, void *dirent, filldir_t filldir) limit = kaddr + ufs_last_byte(inode, n) - UFS_DIR_REC_LEN(1); for ( ;(char*)de <= limit; de = ufs_next_entry(sb, de)) { if (de->d_reclen == 0) { - ufs_error(sb, __FUNCTION__, + ufs_error(sb, __func__, "zero-length directory entry"); ufs_put_page(page); return -EIO; @@ -536,7 +536,7 @@ int ufs_delete_entry(struct inode *inode, struct ufs_dir_entry *dir, while ((char*)de < (char*)dir) { if (de->d_reclen == 0) { - ufs_error(inode->i_sb, __FUNCTION__, + ufs_error(inode->i_sb, __func__, "zero-length directory entry"); err = -EIO; goto out; @@ -633,7 +633,7 @@ int ufs_empty_dir(struct inode * inode) while ((char *)de <= kaddr) { if (de->d_reclen == 0) { - ufs_error(inode->i_sb, __FUNCTION__, + ufs_error(inode->i_sb, __func__, "zero-length directory entry: " "kaddr=%p, de=%p\n", kaddr, de); goto not_empty; diff --git a/fs/ufs/inode.c b/fs/ufs/inode.c index 5446b888fc8..39f87789856 100644 --- a/fs/ufs/inode.c +++ b/fs/ufs/inode.c @@ -929,7 +929,7 @@ void ufs_delete_inode (struct inode * inode) old_i_size = inode->i_size; inode->i_size = 0; if (inode->i_blocks && ufs_truncate(inode, old_i_size)) - ufs_warning(inode->i_sb, __FUNCTION__, "ufs_truncate failed\n"); + ufs_warning(inode->i_sb, __func__, "ufs_truncate failed\n"); ufs_free_inode (inode); unlock_kernel(); return; diff --git a/fs/ufs/ufs.h b/fs/ufs/ufs.h index fcb9231bb9e..244a1aaa940 100644 --- a/fs/ufs/ufs.h +++ b/fs/ufs/ufs.h @@ -66,7 +66,7 @@ struct ufs_inode_info { #ifdef CONFIG_UFS_DEBUG # define UFSD(f, a...) { \ printk ("UFSD (%s, %d): %s:", \ - __FILE__, __LINE__, __FUNCTION__); \ + __FILE__, __LINE__, __func__); \ printk (f, ## a); \ } #else -- cgit v1.2.3 From 0045edaaf980e577272eeb11832daebaeb5d63c4 Mon Sep 17 00:00:00 2001 From: Marcin Slusarz Date: Mon, 28 Apr 2008 02:16:18 -0700 Subject: ufs: remove unused fs64_add and fs64_sub remove fs64_add and fs64_sub - they probably weren't ever used because their prototypes used u32 instead of __fs64 Signed-off-by: Marcin Slusarz Cc: Evgeniy Dushistov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ufs/swab.h | 18 ------------------ 1 file changed, 18 deletions(-) (limited to 'fs') diff --git a/fs/ufs/swab.h b/fs/ufs/swab.h index a4340d0267b..c89e6ad891f 100644 --- a/fs/ufs/swab.h +++ b/fs/ufs/swab.h @@ -40,24 +40,6 @@ cpu_to_fs64(struct super_block *sbp, u64 n) return (__force __fs64)cpu_to_be64(n); } -static __inline u32 -fs64_add(struct super_block *sbp, u32 *n, int d) -{ - if (UFS_SB(sbp)->s_bytesex == BYTESEX_LE) - return *n = cpu_to_le64(le64_to_cpu(*n)+d); - else - return *n = cpu_to_be64(be64_to_cpu(*n)+d); -} - -static __inline u32 -fs64_sub(struct super_block *sbp, u32 *n, int d) -{ - if (UFS_SB(sbp)->s_bytesex == BYTESEX_LE) - return *n = cpu_to_le64(le64_to_cpu(*n)-d); - else - return *n = cpu_to_be64(be64_to_cpu(*n)-d); -} - static __inline u32 fs32_to_cpu(struct super_block *sbp, __fs32 n) { -- cgit v1.2.3 From 36a53ddf859b0a9aa56e2e72978997bf68eec37a Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Mon, 28 Apr 2008 02:16:18 -0700 Subject: ufs: replace __inline with inline Signed-off-by: Harvey Harrison Cc: Evgeniy Dushistov Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/ufs/swab.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ufs/swab.h b/fs/ufs/swab.h index c89e6ad891f..8d974c4fd18 100644 --- a/fs/ufs/swab.h +++ b/fs/ufs/swab.h @@ -40,7 +40,7 @@ cpu_to_fs64(struct super_block *sbp, u64 n) return (__force __fs64)cpu_to_be64(n); } -static __inline u32 +static inline u32 fs32_to_cpu(struct super_block *sbp, __fs32 n) { if (UFS_SB(sbp)->s_bytesex == BYTESEX_LE) -- cgit v1.2.3 From 78e917d59c30c3d5a4cda7f47e0f40f1c98f9b02 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Mon, 28 Apr 2008 02:16:19 -0700 Subject: udf: fix sparse warning in namei.c Let's use bsize instead. fs/udf/namei.c:960:12: warning: symbol 'elen' shadows an earlier one fs/udf/namei.c:937:15: originally declared here Signed-off-by: Harvey Harrison Cc: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/udf/namei.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'fs') diff --git a/fs/udf/namei.c b/fs/udf/namei.c index ba5537d4bc1..2b34c8ca6c8 100644 --- a/fs/udf/namei.c +++ b/fs/udf/namei.c @@ -890,7 +890,7 @@ static int udf_symlink(struct inode *dir, struct dentry *dentry, if (iinfo->i_alloc_type != ICBTAG_FLAG_AD_IN_ICB) { kernel_lb_addr eloc; - uint32_t elen; + uint32_t bsize; block = udf_new_block(inode->i_sb, inode, iinfo->i_location.partitionReferenceNum, @@ -903,9 +903,9 @@ static int udf_symlink(struct inode *dir, struct dentry *dentry, eloc.logicalBlockNum = block; eloc.partitionReferenceNum = iinfo->i_location.partitionReferenceNum; - elen = inode->i_sb->s_blocksize; - iinfo->i_lenExtents = elen; - udf_add_aext(inode, &epos, eloc, elen, 0); + bsize = inode->i_sb->s_blocksize; + iinfo->i_lenExtents = bsize; + udf_add_aext(inode, &epos, eloc, bsize, 0); brelse(epos.bh); block = udf_get_pblock(inode->i_sb, block, -- cgit v1.2.3 From 9e902df6be2cb7444e5a0f7e2e72bcbf3b978f3e Mon Sep 17 00:00:00 2001 From: Marcin Slusarz Date: Mon, 28 Apr 2008 02:16:20 -0700 Subject: reiserfs: le*_add_cpu conversion replace all: little_endian_variable = cpu_to_leX(leX_to_cpu(little_endian_variable) + expression_in_cpu_byteorder); with: leX_add_cpu(&little_endian_variable, expression_in_cpu_byteorder); generated with semantic patch Signed-off-by: Marcin Slusarz Cc: Jeff Mahoney Cc: Chris Mason Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/reiserfs/objectid.c | 5 ++--- fs/reiserfs/stree.c | 3 +-- 2 files changed, 3 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/reiserfs/objectid.c b/fs/reiserfs/objectid.c index 65feba4deb6..f0c154372f5 100644 --- a/fs/reiserfs/objectid.c +++ b/fs/reiserfs/objectid.c @@ -114,7 +114,7 @@ void reiserfs_release_objectid(struct reiserfs_transaction_handle *th, if (objectid_to_release == le32_to_cpu(map[i])) { /* This incrementation unallocates the objectid. */ //map[i]++; - map[i] = cpu_to_le32(le32_to_cpu(map[i]) + 1); + le32_add_cpu(&map[i], 1); /* Did we unallocate the last member of an odd sequence, and can shrink oids? */ if (map[i] == map[i + 1]) { @@ -138,8 +138,7 @@ void reiserfs_release_objectid(struct reiserfs_transaction_handle *th, /* size of objectid map is not changed */ if (objectid_to_release + 1 == le32_to_cpu(map[i + 1])) { //objectid_map[i+1]--; - map[i + 1] = - cpu_to_le32(le32_to_cpu(map[i + 1]) - 1); + le32_add_cpu(&map[i + 1], -1); return; } diff --git a/fs/reiserfs/stree.c b/fs/reiserfs/stree.c index d2db2417b2b..abbc64dcc8d 100644 --- a/fs/reiserfs/stree.c +++ b/fs/reiserfs/stree.c @@ -1419,8 +1419,7 @@ int reiserfs_delete_object(struct reiserfs_transaction_handle *th, inode_generation = &REISERFS_SB(th->t_super)->s_rs->s_inode_generation; - *inode_generation = - cpu_to_le32(le32_to_cpu(*inode_generation) + 1); + le32_add_cpu(inode_generation, 1); } /* USE_INODE_GENERATION_COUNTER */ #endif -- cgit v1.2.3 From e13601bc6aa427a6483e7f66fde55ac3b5fed335 Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Mon, 28 Apr 2008 02:16:21 -0700 Subject: reiserfs: fix sparse warning in journal.c fs/reiserfs/journal.c:4319:2: warning: returning void-valued expression Signed-off-by: Harvey Harrison Cc: Chris Mason Cc: Jeff Mahoney Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/reiserfs/journal.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/reiserfs/journal.c b/fs/reiserfs/journal.c index 060eb3f598e..2e87b6aaec7 100644 --- a/fs/reiserfs/journal.c +++ b/fs/reiserfs/journal.c @@ -4316,5 +4316,5 @@ static void __reiserfs_journal_abort_soft(struct super_block *sb, int errno) void reiserfs_journal_abort(struct super_block *sb, int errno) { - return __reiserfs_journal_abort_soft(sb, errno); + __reiserfs_journal_abort_soft(sb, errno); } -- cgit v1.2.3 From 8acc570fabc022d41baedc0c9bf976a4b9d16d6f Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Mon, 28 Apr 2008 02:16:21 -0700 Subject: reiserfs: fix more sparse warnings in do_balan.c fs/reiserfs/do_balan.c:1467:10: warning: symbol 'ret_val' shadows an earlier one fs/reiserfs/do_balan.c:275:6: originally declared here fs/reiserfs/do_balan.c:1471:23: warning: symbol 'ih' shadows an earlier one fs/reiserfs/do_balan.c:249:67: originally declared here Signed-off-by: Harvey Harrison Cc: Chris Mason Cc: Jeff Mahoney Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/reiserfs/do_balan.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/reiserfs/do_balan.c b/fs/reiserfs/do_balan.c index 7ee4208793b..2f87f5b1463 100644 --- a/fs/reiserfs/do_balan.c +++ b/fs/reiserfs/do_balan.c @@ -1464,29 +1464,29 @@ static int balance_leaf(struct tree_balance *tb, struct item_head *ih, /* item h } else /* item falls wholly into S_new[i] */ { - int ret_val; + int leaf_mi; struct item_head *pasted; #ifdef CONFIG_REISERFS_CHECK - struct item_head *ih = + struct item_head *ih_check = B_N_PITEM_HEAD(tbS0, item_pos); - if (!is_direntry_le_ih(ih) - && (pos_in_item != ih_item_len(ih) + if (!is_direntry_le_ih(ih_check) + && (pos_in_item != ih_item_len(ih_check) || tb->insert_size[0] <= 0)) reiserfs_panic(tb->tb_sb, "PAP-12235: balance_leaf: pos_in_item must be equal to ih_item_len"); #endif /* CONFIG_REISERFS_CHECK */ - ret_val = + leaf_mi = leaf_move_items(LEAF_FROM_S_TO_SNEW, tb, snum[i], sbytes[i], S_new[i]); - RFALSE(ret_val, + RFALSE(leaf_mi, "PAP-12240: unexpected value returned by leaf_move_items (%d)", - ret_val); + leaf_mi); /* paste into item */ bi.tb = tb; -- cgit v1.2.3 From fbe5498b3d5af44a8671800e85b2858c2ea236fa Mon Sep 17 00:00:00 2001 From: Harvey Harrison Date: Mon, 28 Apr 2008 02:16:22 -0700 Subject: reiserfs: replace remaining __FUNCTION__ occurrences __FUNCTION__ is gcc-specific, use __func__ Signed-off-by: Harvey Harrison Cc: Chris Mason Cc: Jeff Mahoney Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/reiserfs/bitmap.c | 8 ++++---- fs/reiserfs/journal.c | 8 ++++---- fs/reiserfs/namei.c | 8 ++++---- fs/reiserfs/objectid.c | 2 +- 4 files changed, 13 insertions(+), 13 deletions(-) (limited to 'fs') diff --git a/fs/reiserfs/bitmap.c b/fs/reiserfs/bitmap.c index f491ceb5af0..4646caa6045 100644 --- a/fs/reiserfs/bitmap.c +++ b/fs/reiserfs/bitmap.c @@ -479,7 +479,7 @@ static void __discard_prealloc(struct reiserfs_transaction_handle *th, if (ei->i_prealloc_count < 0) reiserfs_warning(th->t_super, "zam-4001:%s: inode has negative prealloc blocks count.", - __FUNCTION__); + __func__); #endif while (ei->i_prealloc_count > 0) { reiserfs_free_prealloc_block(th, inode, ei->i_prealloc_block); @@ -517,7 +517,7 @@ void reiserfs_discard_all_prealloc(struct reiserfs_transaction_handle *th) if (!ei->i_prealloc_count) { reiserfs_warning(th->t_super, "zam-4001:%s: inode is in prealloc list but has no preallocated blocks.", - __FUNCTION__); + __func__); } #endif __discard_prealloc(th, ei); @@ -632,7 +632,7 @@ int reiserfs_parse_alloc_options(struct super_block *s, char *options) } reiserfs_warning(s, "zam-4001: %s : unknown option - %s", - __FUNCTION__, this_char); + __func__, this_char); return 1; } @@ -1254,7 +1254,7 @@ struct buffer_head *reiserfs_read_bitmap_block(struct super_block *sb, bh = sb_bread(sb, block); if (bh == NULL) reiserfs_warning(sb, "sh-2029: %s: bitmap block (#%u) " - "reading failed", __FUNCTION__, block); + "reading failed", __func__, block); else { if (buffer_locked(bh)) { PROC_INFO_INC(sb, scan_bitmap.wait); diff --git a/fs/reiserfs/journal.c b/fs/reiserfs/journal.c index 2e87b6aaec7..da86042b3e0 100644 --- a/fs/reiserfs/journal.c +++ b/fs/reiserfs/journal.c @@ -1187,7 +1187,7 @@ static int flush_commit_list(struct super_block *s, if (retval) reiserfs_abort(s, retval, "Journal write error in %s", - __FUNCTION__); + __func__); put_fs_excl(); return retval; } @@ -1534,7 +1534,7 @@ static int flush_journal_list(struct super_block *s, reiserfs_warning(s, "clm-2082: Unable to flush buffer %llu in %s", (unsigned long long)saved_bh-> - b_blocknr, __FUNCTION__); + b_blocknr, __func__); } free_cnode: last = cn; @@ -1586,7 +1586,7 @@ static int flush_journal_list(struct super_block *s, if (err) reiserfs_abort(s, -EIO, "Write error while pushing transaction to disk in %s", - __FUNCTION__); + __func__); flush_older_and_return: /* before we can update the journal header block, we _must_ flush all @@ -1616,7 +1616,7 @@ static int flush_journal_list(struct super_block *s, if (err) reiserfs_abort(s, -EIO, "Write error while updating journal header in %s", - __FUNCTION__); + __func__); } remove_all_from_journal_list(s, jl, 0); list_del_init(&jl->j_list); diff --git a/fs/reiserfs/namei.c b/fs/reiserfs/namei.c index 8867533cb72..c1add28dd45 100644 --- a/fs/reiserfs/namei.c +++ b/fs/reiserfs/namei.c @@ -301,7 +301,7 @@ static int reiserfs_find_entry(struct inode *dir, const char *name, int namelen, path_to_entry, de); if (retval == IO_ERROR) { reiserfs_warning(dir->i_sb, "zam-7001: io error in %s", - __FUNCTION__); + __func__); return IO_ERROR; } @@ -496,7 +496,7 @@ static int reiserfs_add_entry(struct reiserfs_transaction_handle *th, reiserfs_warning(dir->i_sb, "zam-7002:%s: \"reiserfs_find_entry\" " "has returned unexpected value (%d)", - __FUNCTION__, retval); + __func__, retval); } return -EEXIST; @@ -907,7 +907,7 @@ static int reiserfs_rmdir(struct inode *dir, struct dentry *dentry) if (inode->i_nlink != 2 && inode->i_nlink != 1) reiserfs_warning(inode->i_sb, "%s: empty directory has nlink " - "!= 2 (%d)", __FUNCTION__, inode->i_nlink); + "!= 2 (%d)", __func__, inode->i_nlink); clear_nlink(inode); inode->i_ctime = dir->i_ctime = dir->i_mtime = CURRENT_TIME_SEC; @@ -984,7 +984,7 @@ static int reiserfs_unlink(struct inode *dir, struct dentry *dentry) if (!inode->i_nlink) { reiserfs_warning(inode->i_sb, "%s: deleting nonexistent file " - "(%s:%lu), %d", __FUNCTION__, + "(%s:%lu), %d", __func__, reiserfs_bdevname(inode->i_sb), inode->i_ino, inode->i_nlink); inode->i_nlink = 1; diff --git a/fs/reiserfs/objectid.c b/fs/reiserfs/objectid.c index f0c154372f5..ea0cf8c28a9 100644 --- a/fs/reiserfs/objectid.c +++ b/fs/reiserfs/objectid.c @@ -61,7 +61,7 @@ __u32 reiserfs_get_unused_objectid(struct reiserfs_transaction_handle *th) /* comment needed -Hans */ unused_objectid = le32_to_cpu(map[1]); if (unused_objectid == U32_MAX) { - reiserfs_warning(s, "%s: no more object ids", __FUNCTION__); + reiserfs_warning(s, "%s: no more object ids", __func__); reiserfs_restore_prepared_buffer(s, SB_BUFFER_WITH_SB(s)); return 0; } -- cgit v1.2.3 From a2fe594fa3732298947377409b5d8111b14f56ec Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 28 Apr 2008 02:16:23 -0700 Subject: reiserfs: fix hang on umount with quotas when journal is aborted Call dquot_drop() from reiserfs_dquot_drop() even if we fail to start a transaction. Otherwise we never get to dropping references to quota structures from the inode and umount will hang indefinitely. Signed-off-by: Jan Kara Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/reiserfs/super.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index f8f9a473e67..a4c9770a601 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -1890,8 +1890,14 @@ static int reiserfs_dquot_drop(struct inode *inode) ret = journal_begin(&th, inode->i_sb, 2 * REISERFS_QUOTA_DEL_BLOCKS(inode->i_sb)); - if (ret) + if (ret) { + /* + * We call dquot_drop() anyway to at least release references + * to quota structures so that umount does not hang. + */ + dquot_drop(inode); goto out; + } ret = dquot_drop(inode); err = journal_end(&th, inode->i_sb, -- cgit v1.2.3 From d5dee5c395062a55236318ac4eec1f4ebb9de6db Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Mon, 28 Apr 2008 02:16:23 -0700 Subject: reiserfs: unpack tails on quota files Quota files cannot have tails because quota_write and quota_read functions do not support them. So far when quota files did have tail, we just refused to turn quotas on it. Sadly this check has been wrong and so there are now plenty installations where quota files don't have NOTAIL flag set and so now after fixing the check, they suddently fail to turn quotas on. Since it's easy to unpack the tail from kernel, do this from reiserfs_quota_on() which solves the problem and is generally nicer to users anyway. Signed-off-by: Jan Kara Reported-by: Cc: Jeff Mahoney Cc: Chris Mason Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/reiserfs/ioctl.c | 4 +--- fs/reiserfs/super.c | 17 ++++++++++++----- 2 files changed, 13 insertions(+), 8 deletions(-) (limited to 'fs') diff --git a/fs/reiserfs/ioctl.c b/fs/reiserfs/ioctl.c index 74363a7aacb..830332021ed 100644 --- a/fs/reiserfs/ioctl.c +++ b/fs/reiserfs/ioctl.c @@ -12,8 +12,6 @@ #include #include -static int reiserfs_unpack(struct inode *inode, struct file *filp); - /* ** reiserfs_ioctl - handler for ioctl for inode ** supported commands: @@ -159,7 +157,7 @@ int reiserfs_prepare_write(struct file *f, struct page *page, ** Function try to convert tail from direct item into indirect. ** It set up nopack attribute in the REISERFS_I(inode)->nopack */ -static int reiserfs_unpack(struct inode *inode, struct file *filp) +int reiserfs_unpack(struct inode *inode, struct file *filp) { int retval = 0; int index; diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index a4c9770a601..ed424d708e6 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -2025,6 +2025,7 @@ static int reiserfs_quota_on(struct super_block *sb, int type, int format_id, { int err; struct nameidata nd; + struct inode *inode; if (!(REISERFS_SB(sb)->s_mount_opt & (1 << REISERFS_QUOTA))) return -EINVAL; @@ -2039,12 +2040,18 @@ static int reiserfs_quota_on(struct super_block *sb, int type, int format_id, path_put(&nd.path); return -EXDEV; } + inode = nd.path.dentry->d_inode; /* We must not pack tails for quota files on reiserfs for quota IO to work */ - if (!(REISERFS_I(nd.path.dentry->d_inode)->i_flags & i_nopack_mask)) { - reiserfs_warning(sb, - "reiserfs: Quota file must have tail packing disabled."); - path_put(&nd.path); - return -EINVAL; + if (!(REISERFS_I(inode)->i_flags & i_nopack_mask)) { + err = reiserfs_unpack(inode, NULL); + if (err) { + reiserfs_warning(sb, + "reiserfs: Unpacking tail of quota file failed" + " (%d). Cannot turn on quotas.", err); + path_put(&nd.path); + return -EINVAL; + } + mark_inode_dirty(inode); } /* Not journalling quota? No more tests needed... */ if (!REISERFS_SB(sb)->s_qf_names[USRQUOTA] && -- cgit v1.2.3 From 3754a544474b9fd4fdd32a96d98bc4ab558b2bb6 Mon Sep 17 00:00:00 2001 From: OGAWA Hirofumi Date: Mon, 28 Apr 2008 02:16:24 -0700 Subject: fat: kill is_bad_inode() check FAT doesn't need to check bad inode anymore. Signed-off-by: OGAWA Hirofumi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fat/inode.c | 9 ++------- 1 file changed, 2 insertions(+), 7 deletions(-) (limited to 'fs') diff --git a/fs/fat/inode.c b/fs/fat/inode.c index 53f3cf62b7c..f22f287e507 100644 --- a/fs/fat/inode.c +++ b/fs/fat/inode.c @@ -433,11 +433,8 @@ EXPORT_SYMBOL_GPL(fat_build_inode); static void fat_delete_inode(struct inode *inode) { truncate_inode_pages(&inode->i_data, 0); - - if (!is_bad_inode(inode)) { - inode->i_size = 0; - fat_truncate(inode); - } + inode->i_size = 0; + fat_truncate(inode); clear_inode(inode); } @@ -445,8 +442,6 @@ static void fat_clear_inode(struct inode *inode) { struct msdos_sb_info *sbi = MSDOS_SB(inode->i_sb); - if (is_bad_inode(inode)) - return; lock_kernel(); spin_lock(&sbi->inode_hash_lock); fat_cache_inval_inode(inode); -- cgit v1.2.3 From 1278fdd34b12214b5c2e91e64848a5e54e57ed96 Mon Sep 17 00:00:00 2001 From: OGAWA Hirofumi Date: Mon, 28 Apr 2008 02:16:25 -0700 Subject: fat: fat_notify_change() and check_mode() cleanup - Rename fat_notify_change() to fat_setattr() - check_mode() cleanup - Change layout of code Signed-off-by: OGAWA Hirofumi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fat/file.c | 183 +++++++++++++++++++++++++------------------------------ fs/msdos/namei.c | 2 +- fs/vfat/namei.c | 2 +- 3 files changed, 86 insertions(+), 101 deletions(-) (limited to 'fs') diff --git a/fs/fat/file.c b/fs/fat/file.c index 2a3bed96704..b61a98f5398 100644 --- a/fs/fat/file.c +++ b/fs/fat/file.c @@ -157,104 +157,6 @@ out: return err; } -static int check_mode(const struct msdos_sb_info *sbi, mode_t mode) -{ - mode_t req = mode & ~S_IFMT; - - /* - * Of the r and x bits, all (subject to umask) must be present. Of the - * w bits, either all (subject to umask) or none must be present. - */ - - if (S_ISREG(mode)) { - req &= ~sbi->options.fs_fmask; - - if ((req & (S_IRUGO | S_IXUGO)) != - ((S_IRUGO | S_IXUGO) & ~sbi->options.fs_fmask)) - return -EPERM; - - if ((req & S_IWUGO) != 0 && - (req & S_IWUGO) != (S_IWUGO & ~sbi->options.fs_fmask)) - return -EPERM; - } else if (S_ISDIR(mode)) { - req &= ~sbi->options.fs_dmask; - - if ((req & (S_IRUGO | S_IXUGO)) != - ((S_IRUGO | S_IXUGO) & ~sbi->options.fs_dmask)) - return -EPERM; - - if ((req & S_IWUGO) != 0 && - (req & S_IWUGO) != (S_IWUGO & ~sbi->options.fs_dmask)) - return -EPERM; - } else { - return -EPERM; - } - - return 0; -} - -int fat_notify_change(struct dentry *dentry, struct iattr *attr) -{ - struct msdos_sb_info *sbi = MSDOS_SB(dentry->d_sb); - struct inode *inode = dentry->d_inode; - int mask, error = 0; - - lock_kernel(); - - /* - * Expand the file. Since inode_setattr() updates ->i_size - * before calling the ->truncate(), but FAT needs to fill the - * hole before it. - */ - if (attr->ia_valid & ATTR_SIZE) { - if (attr->ia_size > inode->i_size) { - error = fat_cont_expand(inode, attr->ia_size); - if (error || attr->ia_valid == ATTR_SIZE) - goto out; - attr->ia_valid &= ~ATTR_SIZE; - } - } - - error = inode_change_ok(inode, attr); - if (error) { - if (sbi->options.quiet) - error = 0; - goto out; - } - if (((attr->ia_valid & ATTR_UID) && - (attr->ia_uid != sbi->options.fs_uid)) || - ((attr->ia_valid & ATTR_GID) && - (attr->ia_gid != sbi->options.fs_gid))) - error = -EPERM; - - if (error) { - if (sbi->options.quiet) - error = 0; - goto out; - } - - if (attr->ia_valid & ATTR_MODE) { - error = check_mode(sbi, attr->ia_mode); - if (error != 0 && !sbi->options.quiet) - goto out; - } - - error = inode_setattr(inode, attr); - if (error) - goto out; - - if (S_ISDIR(inode->i_mode)) - mask = sbi->options.fs_dmask; - else - mask = sbi->options.fs_fmask; - inode->i_mode &= S_IFMT | (S_IRWXUGO & ~mask); -out: - unlock_kernel(); - return error; -} - -EXPORT_SYMBOL_GPL(fat_notify_change); - /* Free all clusters after the skip'th cluster. */ static int fat_free(struct inode *inode, int skip) { @@ -355,8 +257,91 @@ int fat_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat) } EXPORT_SYMBOL_GPL(fat_getattr); +static int fat_check_mode(const struct msdos_sb_info *sbi, mode_t mode) +{ + mode_t mask, req = mode & ~S_IFMT; + + if (S_ISREG(mode)) + mask = sbi->options.fs_fmask; + else + mask = sbi->options.fs_dmask; + + /* + * Of the r and x bits, all (subject to umask) must be present. Of the + * w bits, either all (subject to umask) or none must be present. + */ + req &= ~mask; + if ((req & (S_IRUGO | S_IXUGO)) != ((S_IRUGO | S_IXUGO) & ~mask)) + return -EPERM; + if ((req & S_IWUGO) && ((req & S_IWUGO) != (S_IWUGO & ~mask))) + return -EPERM; + + return 0; +} + +int fat_setattr(struct dentry *dentry, struct iattr *attr) +{ + struct msdos_sb_info *sbi = MSDOS_SB(dentry->d_sb); + struct inode *inode = dentry->d_inode; + int mask, error = 0; + + lock_kernel(); + + /* + * Expand the file. Since inode_setattr() updates ->i_size + * before calling the ->truncate(), but FAT needs to fill the + * hole before it. + */ + if (attr->ia_valid & ATTR_SIZE) { + if (attr->ia_size > inode->i_size) { + error = fat_cont_expand(inode, attr->ia_size); + if (error || attr->ia_valid == ATTR_SIZE) + goto out; + attr->ia_valid &= ~ATTR_SIZE; + } + } + + error = inode_change_ok(inode, attr); + if (error) { + if (sbi->options.quiet) + error = 0; + goto out; + } + if (((attr->ia_valid & ATTR_UID) && + (attr->ia_uid != sbi->options.fs_uid)) || + ((attr->ia_valid & ATTR_GID) && + (attr->ia_gid != sbi->options.fs_gid))) + error = -EPERM; + + if (error) { + if (sbi->options.quiet) + error = 0; + goto out; + } + + if (attr->ia_valid & ATTR_MODE) { + error = fat_check_mode(sbi, attr->ia_mode); + if (error != 0 && !sbi->options.quiet) + goto out; + } + + error = inode_setattr(inode, attr); + if (error) + goto out; + + if (S_ISDIR(inode->i_mode)) + mask = sbi->options.fs_dmask; + else + mask = sbi->options.fs_fmask; + inode->i_mode &= S_IFMT | (S_IRWXUGO & ~mask); +out: + unlock_kernel(); + return error; +} +EXPORT_SYMBOL_GPL(fat_setattr); + const struct inode_operations fat_file_inode_operations = { .truncate = fat_truncate, - .setattr = fat_notify_change, + .setattr = fat_setattr, .getattr = fat_getattr, }; diff --git a/fs/msdos/namei.c b/fs/msdos/namei.c index 30f7d0ae221..2d4358c59f6 100644 --- a/fs/msdos/namei.c +++ b/fs/msdos/namei.c @@ -653,7 +653,7 @@ static const struct inode_operations msdos_dir_inode_operations = { .mkdir = msdos_mkdir, .rmdir = msdos_rmdir, .rename = msdos_rename, - .setattr = fat_notify_change, + .setattr = fat_setattr, .getattr = fat_getattr, }; diff --git a/fs/vfat/namei.c b/fs/vfat/namei.c index cd450bea9f1..efc70576e4b 100644 --- a/fs/vfat/namei.c +++ b/fs/vfat/namei.c @@ -1003,7 +1003,7 @@ static const struct inode_operations vfat_dir_inode_operations = { .mkdir = vfat_mkdir, .rmdir = vfat_rmdir, .rename = vfat_rename, - .setattr = fat_notify_change, + .setattr = fat_setattr, .getattr = fat_getattr, }; -- cgit v1.2.3 From e97e8de388723f9491514fa0434ddf1fd713a188 Mon Sep 17 00:00:00 2001 From: OGAWA Hirofumi Date: Mon, 28 Apr 2008 02:16:26 -0700 Subject: fat: fat_setattr() fix Fix fat_setattr() on the case of showexec option. If user specified showexec option, inode->i_mode may not have S_IXUGO. This just use inode->i_mode to fix it. And with this patch, we don't allow chmod() on memory inode, it's just bad behaviour. IOW, we allow changing S_IWUGO only which can be stored to disk. Signed-off-by: OGAWA Hirofumi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fat/file.c | 15 ++++++--------- 1 file changed, 6 insertions(+), 9 deletions(-) (limited to 'fs') diff --git a/fs/fat/file.c b/fs/fat/file.c index b61a98f5398..e73f13a1379 100644 --- a/fs/fat/file.c +++ b/fs/fat/file.c @@ -257,7 +257,8 @@ int fat_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat) } EXPORT_SYMBOL_GPL(fat_getattr); -static int fat_check_mode(const struct msdos_sb_info *sbi, mode_t mode) +static int fat_check_mode(const struct msdos_sb_info *sbi, struct inode *inode, + mode_t mode) { mode_t mask, req = mode & ~S_IFMT; @@ -271,7 +272,7 @@ static int fat_check_mode(const struct msdos_sb_info *sbi, mode_t mode) * w bits, either all (subject to umask) or none must be present. */ req &= ~mask; - if ((req & (S_IRUGO | S_IXUGO)) != ((S_IRUGO | S_IXUGO) & ~mask)) + if ((req & (S_IRUGO | S_IXUGO)) != (inode->i_mode & (S_IRUGO|S_IXUGO))) return -EPERM; if ((req & S_IWUGO) && ((req & S_IWUGO) != (S_IWUGO & ~mask))) return -EPERM; @@ -310,7 +311,9 @@ int fat_setattr(struct dentry *dentry, struct iattr *attr) if (((attr->ia_valid & ATTR_UID) && (attr->ia_uid != sbi->options.fs_uid)) || ((attr->ia_valid & ATTR_GID) && - (attr->ia_gid != sbi->options.fs_gid))) + (attr->ia_gid != sbi->options.fs_gid)) || + ((attr->ia_valid & ATTR_MODE) && + fat_check_mode(sbi, inode, attr->ia_mode) < 0)) error = -EPERM; if (error) { @@ -319,12 +322,6 @@ int fat_setattr(struct dentry *dentry, struct iattr *attr) goto out; } - if (attr->ia_valid & ATTR_MODE) { - error = fat_check_mode(sbi, attr->ia_mode); - if (error != 0 && !sbi->options.quiet) - goto out; - } - error = inode_setattr(inode, attr); if (error) goto out; -- cgit v1.2.3 From 1ae43f826b6cb951fc5b0f9c92372a8d5b63c7f9 Mon Sep 17 00:00:00 2001 From: OGAWA Hirofumi Date: Mon, 28 Apr 2008 02:16:26 -0700 Subject: fat: Add allow_utime option Normally utime(2) checks current process is owner of the file, or it has CAP_FOWNER capability. But FAT filesystem doesn't have uid/gid as on disk info, so normal check is too unflexible. With this option you can relax it. Signed-off-by: OGAWA Hirofumi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fat/file.c | 24 ++++++++++++++++++++++++ fs/fat/inode.c | 18 +++++++++++++++--- 2 files changed, 39 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/fat/file.c b/fs/fat/file.c index e73f13a1379..d604bb13242 100644 --- a/fs/fat/file.c +++ b/fs/fat/file.c @@ -280,11 +280,27 @@ static int fat_check_mode(const struct msdos_sb_info *sbi, struct inode *inode, return 0; } +static int fat_allow_set_time(struct msdos_sb_info *sbi, struct inode *inode) +{ + mode_t allow_utime = sbi->options.allow_utime; + + if (current->fsuid != inode->i_uid) { + if (in_group_p(inode->i_gid)) + allow_utime >>= 3; + if (allow_utime & MAY_WRITE) + return 1; + } + + /* use a default check */ + return 0; +} + int fat_setattr(struct dentry *dentry, struct iattr *attr) { struct msdos_sb_info *sbi = MSDOS_SB(dentry->d_sb); struct inode *inode = dentry->d_inode; int mask, error = 0; + unsigned int ia_valid; lock_kernel(); @@ -302,7 +318,15 @@ int fat_setattr(struct dentry *dentry, struct iattr *attr) } } + /* Check for setting the inode time. */ + ia_valid = attr->ia_valid; + if (ia_valid & (ATTR_MTIME_SET | ATTR_ATIME_SET)) { + if (fat_allow_set_time(sbi, inode)) + attr->ia_valid &= ~(ATTR_MTIME_SET | ATTR_ATIME_SET); + } + error = inode_change_ok(inode, attr); + attr->ia_valid = ia_valid; if (error) { if (sbi->options.quiet) error = 0; diff --git a/fs/fat/inode.c b/fs/fat/inode.c index f22f287e507..886f6095a37 100644 --- a/fs/fat/inode.c +++ b/fs/fat/inode.c @@ -785,6 +785,8 @@ static int fat_show_options(struct seq_file *m, struct vfsmount *mnt) seq_printf(m, ",gid=%u", opts->fs_gid); seq_printf(m, ",fmask=%04o", opts->fs_fmask); seq_printf(m, ",dmask=%04o", opts->fs_dmask); + if (opts->allow_utime) + seq_printf(m, ",allow_utime=%04o", opts->allow_utime); if (sbi->nls_disk) seq_printf(m, ",codepage=%s", sbi->nls_disk->charset); if (isvfat) { @@ -840,9 +842,9 @@ static int fat_show_options(struct seq_file *m, struct vfsmount *mnt) enum { Opt_check_n, Opt_check_r, Opt_check_s, Opt_uid, Opt_gid, - Opt_umask, Opt_dmask, Opt_fmask, Opt_codepage, Opt_usefree, Opt_nocase, - Opt_quiet, Opt_showexec, Opt_debug, Opt_immutable, - Opt_dots, Opt_nodots, + Opt_umask, Opt_dmask, Opt_fmask, Opt_allow_utime, Opt_codepage, + Opt_usefree, Opt_nocase, Opt_quiet, Opt_showexec, Opt_debug, + Opt_immutable, Opt_dots, Opt_nodots, Opt_charset, Opt_shortname_lower, Opt_shortname_win95, Opt_shortname_winnt, Opt_shortname_mixed, Opt_utf8_no, Opt_utf8_yes, Opt_uni_xl_no, Opt_uni_xl_yes, Opt_nonumtail_no, Opt_nonumtail_yes, @@ -861,6 +863,7 @@ static match_table_t fat_tokens = { {Opt_umask, "umask=%o"}, {Opt_dmask, "dmask=%o"}, {Opt_fmask, "fmask=%o"}, + {Opt_allow_utime, "allow_utime=%o"}, {Opt_codepage, "codepage=%u"}, {Opt_usefree, "usefree"}, {Opt_nocase, "nocase"}, @@ -932,6 +935,7 @@ static int parse_options(char *options, int is_vfat, int silent, int *debug, opts->fs_uid = current->uid; opts->fs_gid = current->gid; opts->fs_fmask = opts->fs_dmask = current->fs->umask; + opts->allow_utime = -1; opts->codepage = fat_default_codepage; opts->iocharset = fat_default_iocharset; if (is_vfat) @@ -1019,6 +1023,11 @@ static int parse_options(char *options, int is_vfat, int silent, int *debug, return 0; opts->fs_fmask = option; break; + case Opt_allow_utime: + if (match_octal(&args[0], &option)) + return 0; + opts->allow_utime = option & (S_IWGRP | S_IWOTH); + break; case Opt_codepage: if (match_int(&args[0], &option)) return 0; @@ -1101,6 +1110,9 @@ static int parse_options(char *options, int is_vfat, int silent, int *debug, " for FAT filesystems, filesystem will be case sensitive!\n"); } + /* If user doesn't specify allow_utime, it's initialized from dmask. */ + if (opts->allow_utime == (unsigned short)-1) + opts->allow_utime = ~opts->fs_dmask & (S_IWGRP | S_IWOTH); if (opts->unicode_xlate) opts->utf8 = 0; -- cgit v1.2.3 From 606e423e43bac0c2f7b85b682eb1ddd2a634586e Mon Sep 17 00:00:00 2001 From: OGAWA Hirofumi Date: Mon, 28 Apr 2008 02:16:27 -0700 Subject: fat: Update free_clusters even if it is untrusted Currently, free_clusters is not updated until it is trusted, because Windows doesn't update it correctly. But if user is using FAT driver of Linux, it updates free_clusters correctly. Instead, this updates it even if it's untrusted, so if free_clustes is correct, now keep correct value. Signed-off-by: OGAWA Hirofumi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fat/fatent.c | 7 +++++-- fs/fat/inode.c | 7 ++++--- 2 files changed, 9 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/fat/fatent.c b/fs/fat/fatent.c index 5fb366992b7..c7c8ab2209e 100644 --- a/fs/fat/fatent.c +++ b/fs/fat/fatent.c @@ -450,7 +450,8 @@ int fat_alloc_clusters(struct inode *inode, int *cluster, int nr_cluster) BUG_ON(nr_cluster > (MAX_BUF_PER_PAGE / 2)); /* fixed limit */ lock_fat(sbi); - if (sbi->free_clusters != -1 && sbi->free_clusters < nr_cluster) { + if (sbi->free_clusters != -1 && sbi->free_clus_valid && + sbi->free_clusters < nr_cluster) { unlock_fat(sbi); return -ENOSPC; } @@ -504,6 +505,7 @@ int fat_alloc_clusters(struct inode *inode, int *cluster, int nr_cluster) /* Couldn't allocate the free entries */ sbi->free_clusters = 0; + sbi->free_clus_valid = 1; sb->s_dirt = 1; err = -ENOSPC; @@ -615,7 +617,7 @@ int fat_count_free_clusters(struct super_block *sb) int err = 0, free; lock_fat(sbi); - if (sbi->free_clusters != -1) + if (sbi->free_clusters != -1 && sbi->free_clus_valid) goto out; reada_blocks = FAT_READA_SIZE >> sb->s_blocksize_bits; @@ -643,6 +645,7 @@ int fat_count_free_clusters(struct super_block *sb) } while (fat_ent_next(sbi, &fatent)); } sbi->free_clusters = free; + sbi->free_clus_valid = 1; sb->s_dirt = 1; fatent_brelse(&fatent); out: diff --git a/fs/fat/inode.c b/fs/fat/inode.c index 886f6095a37..1604a8a5c01 100644 --- a/fs/fat/inode.c +++ b/fs/fat/inode.c @@ -537,7 +537,7 @@ static int fat_statfs(struct dentry *dentry, struct kstatfs *buf) struct msdos_sb_info *sbi = MSDOS_SB(dentry->d_sb); /* If the count of free cluster is still unknown, counts it here. */ - if (sbi->free_clusters == -1) { + if (sbi->free_clusters == -1 || !sbi->free_clus_valid) { int err = fat_count_free_clusters(dentry->d_sb); if (err) return err; @@ -1274,6 +1274,7 @@ int fat_fill_super(struct super_block *sb, void *data, int silent, sbi->fat_length = le16_to_cpu(b->fat_length); sbi->root_cluster = 0; sbi->free_clusters = -1; /* Don't know yet */ + sbi->free_clus_valid = 0; sbi->prev_free = FAT_START_ENT; if (!sbi->fat_length && b->fat32_length) { @@ -1309,8 +1310,8 @@ int fat_fill_super(struct super_block *sb, void *data, int silent, sbi->fsinfo_sector); } else { if (sbi->options.usefree) - sbi->free_clusters = - le32_to_cpu(fsinfo->free_clusters); + sbi->free_clus_valid = 1; + sbi->free_clusters = le32_to_cpu(fsinfo->free_clusters); sbi->prev_free = le32_to_cpu(fsinfo->next_cluster); } -- cgit v1.2.3 From e69be4c9c4ac1b6bb6380512557da0afa6262e98 Mon Sep 17 00:00:00 2001 From: OGAWA Hirofumi Date: Mon, 28 Apr 2008 02:16:27 -0700 Subject: fat: Remove fat_clusters_flush() This removes unneeded fat_clusters_flush(). Signed-off-by: OGAWA Hirofumi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fat/fatent.c | 2 -- 1 file changed, 2 deletions(-) (limited to 'fs') diff --git a/fs/fat/fatent.c b/fs/fat/fatent.c index c7c8ab2209e..13ab763cc51 100644 --- a/fs/fat/fatent.c +++ b/fs/fat/fatent.c @@ -585,8 +585,6 @@ error: brelse(bhs[i]); unlock_fat(sbi); - fat_clusters_flush(sb); - return err; } -- cgit v1.2.3 From 061e97469f46f924cf14bbf1dd4805b46986691a Mon Sep 17 00:00:00 2001 From: OGAWA Hirofumi Date: Mon, 28 Apr 2008 02:16:28 -0700 Subject: Add balance_dirty_pages_ratelimited() to cont_expand_zero() On the systems, ftruncate() which expand size for FAT became the cause of OOM. The cont_expand_zero() filled all memory with dirty pages, and since disk is very slow, limit of page scanning was exceeded, then it triggered OOM. This adds balance_dirty_pages_ratelimited() to avoid filling memory with dirty pages. Signed-off-by: OGAWA Hirofumi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/buffer.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'fs') diff --git a/fs/buffer.c b/fs/buffer.c index 7d51e649b19..3db4a26adc4 100644 --- a/fs/buffer.c +++ b/fs/buffer.c @@ -2246,6 +2246,8 @@ int cont_expand_zero(struct file *file, struct address_space *mapping, goto out; BUG_ON(err != len); err = 0; + + balance_dirty_pages_ratelimited(mapping); } /* page covers the boundary, find the boundary offset */ -- cgit v1.2.3 From f22032ba8de2960a64a3dd9719fb5c99b1f1ae6e Mon Sep 17 00:00:00 2001 From: Keith Mok Date: Mon, 28 Apr 2008 02:16:29 -0700 Subject: vfat: bug fix for vfat cannot handle filename with 255 This patch fix the problem that the buffer allocated for convert of unicode to utf8 in fat/dir.c is too small. And cannot handle filename with 255 asian characters when mounted with utf8 options. Also it fix the filename length limitation checking in vfat/namei.c that the filename length should be checked against the number of converted unicode characters. Not the length before NLS/UTF8 converted. Signed-off-by: Keith Mok Signed-off-by: OGAWA Hirofumi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fat/dir.c | 44 ++++++++++++++++++++++++++++---------------- fs/vfat/namei.c | 25 +++++++++++-------------- 2 files changed, 39 insertions(+), 30 deletions(-) (limited to 'fs') diff --git a/fs/fat/dir.c b/fs/fat/dir.c index 72cbcd61bd9..7b62ffb663f 100644 --- a/fs/fat/dir.c +++ b/fs/fat/dir.c @@ -124,8 +124,8 @@ static inline int fat_get_entry(struct inode *dir, loff_t *pos, * but ignore that right now. * Ahem... Stack smashing in ring 0 isn't fun. Fixed. */ -static int uni16_to_x8(unsigned char *ascii, wchar_t *uni, int uni_xlate, - struct nls_table *nls) +static int uni16_to_x8(unsigned char *ascii, wchar_t *uni, int len, + int uni_xlate, struct nls_table *nls) { wchar_t *ip, ec; unsigned char *op, nc; @@ -135,10 +135,11 @@ static int uni16_to_x8(unsigned char *ascii, wchar_t *uni, int uni_xlate, ip = uni; op = ascii; - while (*ip) { + while (*ip && ((len - NLS_MAX_CHARSET_SIZE) > 0)) { ec = *ip++; if ( (charlen = nls->uni2char(ec, op, NLS_MAX_CHARSET_SIZE)) > 0) { op += charlen; + len -= charlen; } else { if (uni_xlate == 1) { *op = ':'; @@ -149,16 +150,19 @@ static int uni16_to_x8(unsigned char *ascii, wchar_t *uni, int uni_xlate, ec >>= 4; } op += 5; + len -= 5; } else { *op++ = '?'; + len--; } } - /* We have some slack there, so it's OK */ - if (op>ascii+256) { - op = ascii + 256; - break; - } } + + if (unlikely(*ip)) { + printk(KERN_WARNING "FAT: filename was truncated while " + "converting."); + } + *op = 0; return (op - ascii); } @@ -311,9 +315,11 @@ int fat_search_long(struct inode *inode, const unsigned char *name, struct nls_table *nls_io = sbi->nls_io; struct nls_table *nls_disk = sbi->nls_disk; wchar_t bufuname[14]; - unsigned char xlate_len, nr_slots; + unsigned char nr_slots; + int xlate_len; wchar_t *unicode = NULL; - unsigned char work[MSDOS_NAME], bufname[260]; /* 256 + 4 */ + unsigned char work[MSDOS_NAME]; + unsigned char *bufname = NULL; int uni_xlate = sbi->options.unicode_xlate; int utf8 = sbi->options.utf8; int anycase = (sbi->options.name_check != 's'); @@ -321,6 +327,10 @@ int fat_search_long(struct inode *inode, const unsigned char *name, loff_t cpos = 0; int chl, i, j, last_u, err; + bufname = (unsigned char*)__get_free_page(GFP_KERNEL); + if (!bufname) + return -ENOMEM; + err = -ENOENT; while(1) { if (fat_get_entry(inode, &cpos, &bh, &de) == -1) @@ -386,8 +396,8 @@ parse_record: bufuname[last_u] = 0x0000; xlate_len = utf8 - ?utf8_wcstombs(bufname, bufuname, sizeof(bufname)) - :uni16_to_x8(bufname, bufuname, uni_xlate, nls_io); + ?utf8_wcstombs(bufname, bufuname, PAGE_SIZE) + :uni16_to_x8(bufname, bufuname, PAGE_SIZE, uni_xlate, nls_io); if (xlate_len == name_len) if ((!anycase && !memcmp(name, bufname, xlate_len)) || (anycase && !nls_strnicmp(nls_io, name, bufname, @@ -396,8 +406,8 @@ parse_record: if (nr_slots) { xlate_len = utf8 - ?utf8_wcstombs(bufname, unicode, sizeof(bufname)) - :uni16_to_x8(bufname, unicode, uni_xlate, nls_io); + ?utf8_wcstombs(bufname, unicode, PAGE_SIZE) + :uni16_to_x8(bufname, unicode, PAGE_SIZE, uni_xlate, nls_io); if (xlate_len != name_len) continue; if ((!anycase && !memcmp(name, bufname, xlate_len)) || @@ -416,6 +426,8 @@ Found: sinfo->i_pos = fat_make_i_pos(sb, sinfo->bh, sinfo->de); err = 0; EODir: + if (bufname) + free_page((unsigned long)bufname); if (unicode) free_page((unsigned long)unicode); @@ -598,7 +610,7 @@ parse_record: if (isvfat) { bufuname[j] = 0x0000; i = utf8 ? utf8_wcstombs(bufname, bufuname, sizeof(bufname)) - : uni16_to_x8(bufname, bufuname, uni_xlate, nls_io); + : uni16_to_x8(bufname, bufuname, sizeof(bufname), uni_xlate, nls_io); } fill_name = bufname; @@ -610,7 +622,7 @@ parse_record: int buf_size = PAGE_SIZE - (261 * sizeof(unicode[0])); int long_len = utf8 ? utf8_wcstombs(longname, unicode, buf_size) - : uni16_to_x8(longname, unicode, uni_xlate, nls_io); + : uni16_to_x8(longname, unicode, buf_size, uni_xlate, nls_io); if (!both) { fill_name = longname; diff --git a/fs/vfat/namei.c b/fs/vfat/namei.c index efc70576e4b..ab4f3da770f 100644 --- a/fs/vfat/namei.c +++ b/fs/vfat/namei.c @@ -176,15 +176,10 @@ static inline int vfat_is_used_badchars(const wchar_t *s, int len) for (i = 0; i < len; i++) if (vfat_bad_char(s[i])) return -EINVAL; - return 0; -} -static int vfat_valid_longname(const unsigned char *name, unsigned int len) -{ - if (name[len - 1] == ' ') + if (s[i - 1] == ' ') /* last character cannot be space */ return -EINVAL; - if (len >= 256) - return -ENAMETOOLONG; + return 0; } @@ -485,11 +480,14 @@ xlate_to_uni(const unsigned char *name, int len, unsigned char *outname, */ *outlen -= (name_len - len); + if (*outlen > 255) + return -ENAMETOOLONG; + op = &outname[*outlen * sizeof(wchar_t)]; } else { if (nls) { for (i = 0, ip = name, op = outname, *outlen = 0; - i < len && *outlen <= 260; + i < len && *outlen <= 255; *outlen += 1) { if (escape && (*ip == ':')) { @@ -525,18 +523,20 @@ xlate_to_uni(const unsigned char *name, int len, unsigned char *outname, op += 2; } } + if (i < len) + return -ENAMETOOLONG; } else { for (i = 0, ip = name, op = outname, *outlen = 0; - i < len && *outlen <= 260; + i < len && *outlen <= 255; i++, *outlen += 1) { *op++ = *ip++; *op++ = 0; } + if (i < len) + return -ENAMETOOLONG; } } - if (*outlen > 260) - return -ENAMETOOLONG; *longlen = *outlen; if (*outlen % 13) { @@ -574,9 +574,6 @@ static int vfat_build_slots(struct inode *dir, const unsigned char *name, loff_t offset; *nr_slots = 0; - err = vfat_valid_longname(name, len); - if (err) - return err; page = __get_free_page(GFP_KERNEL); if (!page) -- cgit v1.2.3 From c7a6c4edc74ee5c3d990325fe93b6916f43499f4 Mon Sep 17 00:00:00 2001 From: OGAWA Hirofumi Date: Mon, 28 Apr 2008 02:16:29 -0700 Subject: fat: use __getname() __getname() is faster than __get_free_page(). Use it. Signed-off-by: OGAWA Hirofumi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fat/dir.c | 20 ++++++++++---------- fs/vfat/namei.c | 10 ++++------ 2 files changed, 14 insertions(+), 16 deletions(-) (limited to 'fs') diff --git a/fs/fat/dir.c b/fs/fat/dir.c index 7b62ffb663f..486725ee99a 100644 --- a/fs/fat/dir.c +++ b/fs/fat/dir.c @@ -247,7 +247,7 @@ static int fat_parse_long(struct inode *dir, loff_t *pos, unsigned char id, slot, slots, alias_checksum; if (!*unicode) { - *unicode = (wchar_t *)__get_free_page(GFP_KERNEL); + *unicode = __getname(); if (!*unicode) { brelse(*bh); return -ENOMEM; @@ -327,7 +327,7 @@ int fat_search_long(struct inode *inode, const unsigned char *name, loff_t cpos = 0; int chl, i, j, last_u, err; - bufname = (unsigned char*)__get_free_page(GFP_KERNEL); + bufname = __getname(); if (!bufname) return -ENOMEM; @@ -396,8 +396,8 @@ parse_record: bufuname[last_u] = 0x0000; xlate_len = utf8 - ?utf8_wcstombs(bufname, bufuname, PAGE_SIZE) - :uni16_to_x8(bufname, bufuname, PAGE_SIZE, uni_xlate, nls_io); + ?utf8_wcstombs(bufname, bufuname, PATH_MAX) + :uni16_to_x8(bufname, bufuname, PATH_MAX, uni_xlate, nls_io); if (xlate_len == name_len) if ((!anycase && !memcmp(name, bufname, xlate_len)) || (anycase && !nls_strnicmp(nls_io, name, bufname, @@ -406,8 +406,8 @@ parse_record: if (nr_slots) { xlate_len = utf8 - ?utf8_wcstombs(bufname, unicode, PAGE_SIZE) - :uni16_to_x8(bufname, unicode, PAGE_SIZE, uni_xlate, nls_io); + ?utf8_wcstombs(bufname, unicode, PATH_MAX) + :uni16_to_x8(bufname, unicode, PATH_MAX, uni_xlate, nls_io); if (xlate_len != name_len) continue; if ((!anycase && !memcmp(name, bufname, xlate_len)) || @@ -427,9 +427,9 @@ Found: err = 0; EODir: if (bufname) - free_page((unsigned long)bufname); + __putname(bufname); if (unicode) - free_page((unsigned long)unicode); + __putname(unicode); return err; } @@ -619,7 +619,7 @@ parse_record: /* convert the unicode long name. 261 is maximum size * of unicode buffer. (13 * slots + nul) */ void *longname = unicode + 261; - int buf_size = PAGE_SIZE - (261 * sizeof(unicode[0])); + int buf_size = PATH_MAX - (261 * sizeof(unicode[0])); int long_len = utf8 ? utf8_wcstombs(longname, unicode, buf_size) : uni16_to_x8(longname, unicode, buf_size, uni_xlate, nls_io); @@ -652,7 +652,7 @@ EODir: FillFailed: brelse(bh); if (unicode) - free_page((unsigned long)unicode); + __putname(unicode); out: unlock_kernel(); return ret; diff --git a/fs/vfat/namei.c b/fs/vfat/namei.c index ab4f3da770f..5b66162d074 100644 --- a/fs/vfat/namei.c +++ b/fs/vfat/namei.c @@ -472,7 +472,7 @@ xlate_to_uni(const unsigned char *name, int len, unsigned char *outname, if (utf8) { int name_len = strlen(name); - *outlen = utf8_mbstowcs((wchar_t *)outname, name, PAGE_SIZE); + *outlen = utf8_mbstowcs((wchar_t *)outname, name, PATH_MAX); /* * We stripped '.'s before and set len appropriately, @@ -565,7 +565,6 @@ static int vfat_build_slots(struct inode *dir, const unsigned char *name, struct fat_mount_options *opts = &sbi->options; struct msdos_dir_slot *ps; struct msdos_dir_entry *de; - unsigned long page; unsigned char cksum, lcase; unsigned char msdos_name[MSDOS_NAME]; wchar_t *uname; @@ -575,11 +574,10 @@ static int vfat_build_slots(struct inode *dir, const unsigned char *name, *nr_slots = 0; - page = __get_free_page(GFP_KERNEL); - if (!page) + uname = __getname(); + if (!uname) return -ENOMEM; - uname = (wchar_t *)page; err = xlate_to_uni(name, len, (unsigned char *)uname, &ulen, &usize, opts->unicode_xlate, opts->utf8, sbi->nls_io); if (err) @@ -631,7 +629,7 @@ shortname: de->starthi = cpu_to_le16(cluster >> 16); de->size = 0; out_free: - free_page(page); + __putname(uname); return err; } -- cgit v1.2.3 From 73f20e58b1d586e9f6d3ddc3aad872829aca7743 Mon Sep 17 00:00:00 2001 From: Andrew Morton Date: Mon, 28 Apr 2008 02:16:30 -0700 Subject: FAT_VALID_MEDIA(): remove pointless test The on-disk media specification field in FAT is only 8-bits, so testing for <=0xff is pointless, and can generate a "comparison is always true due to limited range of data type" warning. While we're there, convert FAT_VALID_MEDIA() into a C function - the present implementation is buggy: it generates either one or two references to its argument. Cc: Frank Seidel Acked-by: OGAWA Hirofumi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fat/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/fat/inode.c b/fs/fat/inode.c index 1604a8a5c01..bf918929e79 100644 --- a/fs/fat/inode.c +++ b/fs/fat/inode.c @@ -1215,7 +1215,7 @@ int fat_fill_super(struct super_block *sb, void *data, int silent, */ media = b->media; - if (!FAT_VALID_MEDIA(media)) { + if (!fat_valid_media(media)) { if (!silent) printk(KERN_ERR "FAT: invalid media value (0x%02x)\n", media); -- cgit v1.2.3 From 0607fd02587a6b4b086dc746d63123c1f284db68 Mon Sep 17 00:00:00 2001 From: Frank Seidel Date: Mon, 28 Apr 2008 02:16:31 -0700 Subject: fat: detect media without partition table correctly I received a complaint that some FAT formated medias (e.g. sd memory cards) trigger a "unknown partition table" message even though there is no partition table and they work correctly, while in general (when e.g. formated with mkdosfs or even Windows Vista) this message is not shown. Currently this seems only to happen when the medias get formatted with Windows XP (and possibly Win 2000). Then the boot indicator byte contains garbage (part of text message) and so do the other parts checked by msdos_paritition which then later triggers this message. References: novell bug #364365 Most fat formatted media without partition table contains zeros in the boot indication and the other tested bytes and so falls through the checks in msdos_partition, leading it to return with 1 (all is fine). But some (e.g. WinXP formatted) fat fomated medias don't use boot_ind and so the check fails and causes a "unkown partition table" warning eventhough there is none and everything would be fine. This additional check directly verifies if there is a fat formatted medium without a partition table. Signed-off-by: Frank Seidel Cc: Andreas Dilger Acked-by: OGAWA Hirofumi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/partitions/msdos.c | 20 +++++++++++++++++--- 1 file changed, 17 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/partitions/msdos.c b/fs/partitions/msdos.c index 5567ec0d03a..796511886f2 100644 --- a/fs/partitions/msdos.c +++ b/fs/partitions/msdos.c @@ -18,7 +18,7 @@ * * Re-organised Feb 1998 Russell King */ - +#include #include "check.h" #include "msdos.h" @@ -419,6 +419,7 @@ int msdos_partition(struct parsed_partitions *state, struct block_device *bdev) Sector sect; unsigned char *data; struct partition *p; + struct fat_boot_sector *fb; int slot; data = read_dev_sector(bdev, 0, §); @@ -444,8 +445,21 @@ int msdos_partition(struct parsed_partitions *state, struct block_device *bdev) p = (struct partition *) (data + 0x1be); for (slot = 1; slot <= 4; slot++, p++) { if (p->boot_ind != 0 && p->boot_ind != 0x80) { - put_dev_sector(sect); - return 0; + /* + * Even without a valid boot inidicator value + * its still possible this is valid FAT filesystem + * without a partition table. + */ + fb = (struct fat_boot_sector *) data; + if (slot == 1 && fb->reserved && fb->fats + && fat_valid_media(fb->media)) { + printk("\n"); + put_dev_sector(sect); + return 1; + } else { + put_dev_sector(sect); + return 0; + } } } -- cgit v1.2.3 From 9f354858b8ea29e969b82dd96caea376157d76ca Mon Sep 17 00:00:00 2001 From: Olof Johansson Date: Mon, 28 Apr 2008 02:16:32 -0700 Subject: fatfs: fix build warning with 64k PAGE_SIZE Annoying gcc warning: fs/fat/inode.c: In function 'fat_fill_super': fs/fat/inode.c:1222: warning: comparison is always false due to limited range of data type Change it to compare with 4K instead of PAGE_CACHE_SIZE, as suggested by OGAWA-san. [FAT spec says: logical_sector_size should be 512, 1024, 2048 4096] So, at least for now, we limit it to 4096. Signed-off-by: Olof Johansson Signed-off-by: OGAWA Hirofumi Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- fs/fat/inode.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/fat/inode.c b/fs/fat/inode.c index bf918929e79..5f522a55b59 100644 --- a/fs/fat/inode.c +++ b/fs/fat/inode.c @@ -1226,7 +1226,7 @@ int fat_fill_super(struct super_block *sb, void *data, int silent, le16_to_cpu(get_unaligned((__le16 *)&b->sector_size)); if (!is_power_of_2(logical_sector_size) || (logical_sector_size < 512) - || (PAGE_CACHE_SIZE < logical_sector_size)) { + || (logical_sector_size > 4096)) { if (!silent) printk(KERN_ERR "FAT: bogus logical sector size %u\n", logical_sector_size); -- cgit v1.2.3