aboutsummaryrefslogtreecommitdiff
path: root/fs/ocfs2
diff options
context:
space:
mode:
authorMark Fasheh <mark.fasheh@oracle.com>2007-07-03 13:34:11 -0700
committerMark Fasheh <mark.fasheh@oracle.com>2007-07-10 17:32:08 -0700
commit063c4561f52a74de686fe0ff2f96f4f54c9fecd2 (patch)
tree73a202c316df70bdfafa489d70e2863c5c5ea33a /fs/ocfs2
parent35edec1d52c075975991471d624b33b9336226f2 (diff)
ocfs2: support for removing file regions
Provide an internal interface for the removal of arbitrary file regions. ocfs2_remove_inode_range() takes a byte range within a file and will remove existing extents within that range. Partial clusters will be zeroed so that any read from within the region will return zeros. Signed-off-by: Mark Fasheh <mark.fasheh@oracle.com>
Diffstat (limited to 'fs/ocfs2')
-rw-r--r--fs/ocfs2/alloc.c20
-rw-r--r--fs/ocfs2/alloc.h10
-rw-r--r--fs/ocfs2/file.c242
-rw-r--r--fs/ocfs2/journal.h2
4 files changed, 262 insertions, 12 deletions
diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c
index df186d2e824..f5e11f4fa95 100644
--- a/fs/ocfs2/alloc.c
+++ b/fs/ocfs2/alloc.c
@@ -4373,10 +4373,10 @@ out:
return ret;
}
-static int ocfs2_remove_extent(struct inode *inode, struct buffer_head *di_bh,
- u32 cpos, u32 len, handle_t *handle,
- struct ocfs2_alloc_context *meta_ac,
- struct ocfs2_cached_dealloc_ctxt *dealloc)
+int ocfs2_remove_extent(struct inode *inode, struct buffer_head *di_bh,
+ u32 cpos, u32 len, handle_t *handle,
+ struct ocfs2_alloc_context *meta_ac,
+ struct ocfs2_cached_dealloc_ctxt *dealloc)
{
int ret, index;
u32 rec_range, trunc_range;
@@ -4506,7 +4506,7 @@ out:
return ret;
}
-static inline int ocfs2_truncate_log_needs_flush(struct ocfs2_super *osb)
+int ocfs2_truncate_log_needs_flush(struct ocfs2_super *osb)
{
struct buffer_head *tl_bh = osb->osb_tl_bh;
struct ocfs2_dinode *di;
@@ -4539,10 +4539,10 @@ static int ocfs2_truncate_log_can_coalesce(struct ocfs2_truncate_log *tl,
return current_tail == new_start;
}
-static int ocfs2_truncate_log_append(struct ocfs2_super *osb,
- handle_t *handle,
- u64 start_blk,
- unsigned int num_clusters)
+int ocfs2_truncate_log_append(struct ocfs2_super *osb,
+ handle_t *handle,
+ u64 start_blk,
+ unsigned int num_clusters)
{
int status, index;
unsigned int start_cluster, tl_count;
@@ -4698,7 +4698,7 @@ bail:
}
/* Expects you to already be holding tl_inode->i_mutex */
-static int __ocfs2_flush_truncate_log(struct ocfs2_super *osb)
+int __ocfs2_flush_truncate_log(struct ocfs2_super *osb)
{
int status;
unsigned int num_to_flush;
diff --git a/fs/ocfs2/alloc.h b/fs/ocfs2/alloc.h
index 752ef860873..990df48ae8d 100644
--- a/fs/ocfs2/alloc.h
+++ b/fs/ocfs2/alloc.h
@@ -41,6 +41,10 @@ int ocfs2_mark_extent_written(struct inode *inode, struct buffer_head *di_bh,
handle_t *handle, u32 cpos, u32 len, u32 phys,
struct ocfs2_alloc_context *meta_ac,
struct ocfs2_cached_dealloc_ctxt *dealloc);
+int ocfs2_remove_extent(struct inode *inode, struct buffer_head *di_bh,
+ u32 cpos, u32 len, handle_t *handle,
+ struct ocfs2_alloc_context *meta_ac,
+ struct ocfs2_cached_dealloc_ctxt *dealloc);
int ocfs2_num_free_extents(struct ocfs2_super *osb,
struct inode *inode,
struct ocfs2_dinode *fe);
@@ -68,6 +72,12 @@ int ocfs2_begin_truncate_log_recovery(struct ocfs2_super *osb,
struct ocfs2_dinode **tl_copy);
int ocfs2_complete_truncate_log_recovery(struct ocfs2_super *osb,
struct ocfs2_dinode *tl_copy);
+int ocfs2_truncate_log_needs_flush(struct ocfs2_super *osb);
+int ocfs2_truncate_log_append(struct ocfs2_super *osb,
+ handle_t *handle,
+ u64 start_blk,
+ unsigned int num_clusters);
+int __ocfs2_flush_truncate_log(struct ocfs2_super *osb);
/*
* Process local structure which describes the block unlinks done
diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
index f0a6b1330a6..11f7cf9f251 100644
--- a/fs/ocfs2/file.c
+++ b/fs/ocfs2/file.c
@@ -541,12 +541,15 @@ int ocfs2_lock_allocators(struct inode *inode, struct ocfs2_dinode *di,
struct ocfs2_alloc_context **data_ac,
struct ocfs2_alloc_context **meta_ac)
{
- int ret, num_free_extents;
+ int ret = 0, num_free_extents;
unsigned int max_recs_needed = clusters_to_add + 2 * extents_to_split;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
*meta_ac = NULL;
- *data_ac = NULL;
+ if (data_ac)
+ *data_ac = NULL;
+
+ BUG_ON(clusters_to_add != 0 && data_ac == NULL);
mlog(0, "extend inode %llu, i_size = %lld, di->i_clusters = %u, "
"clusters_to_add = %u, extents_to_split = %u\n",
@@ -583,6 +586,9 @@ int ocfs2_lock_allocators(struct inode *inode, struct ocfs2_dinode *di,
}
}
+ if (clusters_to_add == 0)
+ goto out;
+
ret = ocfs2_reserve_clusters(osb, clusters_to_add, data_ac);
if (ret < 0) {
if (ret != -ENOSPC)
@@ -1252,6 +1258,238 @@ out:
return ret;
}
+static int __ocfs2_remove_inode_range(struct inode *inode,
+ struct buffer_head *di_bh,
+ u32 cpos, u32 phys_cpos, u32 len,
+ struct ocfs2_cached_dealloc_ctxt *dealloc)
+{
+ int ret;
+ u64 phys_blkno = ocfs2_clusters_to_blocks(inode->i_sb, phys_cpos);
+ struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+ struct inode *tl_inode = osb->osb_tl_inode;
+ handle_t *handle;
+ struct ocfs2_alloc_context *meta_ac = NULL;
+ struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
+
+ ret = ocfs2_lock_allocators(inode, di, 0, 1, NULL, &meta_ac);
+ if (ret) {
+ mlog_errno(ret);
+ return ret;
+ }
+
+ mutex_lock(&tl_inode->i_mutex);
+
+ if (ocfs2_truncate_log_needs_flush(osb)) {
+ ret = __ocfs2_flush_truncate_log(osb);
+ if (ret < 0) {
+ mlog_errno(ret);
+ goto out;
+ }
+ }
+
+ handle = ocfs2_start_trans(osb, OCFS2_REMOVE_EXTENT_CREDITS);
+ if (handle == NULL) {
+ ret = -ENOMEM;
+ mlog_errno(ret);
+ goto out;
+ }
+
+ ret = ocfs2_journal_access(handle, inode, di_bh,
+ OCFS2_JOURNAL_ACCESS_WRITE);
+ if (ret) {
+ mlog_errno(ret);
+ goto out;
+ }
+
+ ret = ocfs2_remove_extent(inode, di_bh, cpos, len, handle, meta_ac,
+ dealloc);
+ if (ret) {
+ mlog_errno(ret);
+ goto out_commit;
+ }
+
+ OCFS2_I(inode)->ip_clusters -= len;
+ di->i_clusters = cpu_to_le32(OCFS2_I(inode)->ip_clusters);
+
+ ret = ocfs2_journal_dirty(handle, di_bh);
+ if (ret) {
+ mlog_errno(ret);
+ goto out_commit;
+ }
+
+ ret = ocfs2_truncate_log_append(osb, handle, phys_blkno, len);
+ if (ret)
+ mlog_errno(ret);
+
+out_commit:
+ ocfs2_commit_trans(osb, handle);
+out:
+ mutex_unlock(&tl_inode->i_mutex);
+
+ if (meta_ac)
+ ocfs2_free_alloc_context(meta_ac);
+
+ return ret;
+}
+
+/*
+ * Truncate a byte range, avoiding pages within partial clusters. This
+ * preserves those pages for the zeroing code to write to.
+ */
+static void ocfs2_truncate_cluster_pages(struct inode *inode, u64 byte_start,
+ u64 byte_len)
+{
+ struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+ loff_t start, end;
+ struct address_space *mapping = inode->i_mapping;
+
+ start = (loff_t)ocfs2_align_bytes_to_clusters(inode->i_sb, byte_start);
+ end = byte_start + byte_len;
+ end = end & ~(osb->s_clustersize - 1);
+
+ if (start < end) {
+ unmap_mapping_range(mapping, start, end - start, 0);
+ truncate_inode_pages_range(mapping, start, end - 1);
+ }
+}
+
+static int ocfs2_zero_partial_clusters(struct inode *inode,
+ u64 start, u64 len)
+{
+ int ret = 0;
+ u64 tmpend, end = start + len;
+ struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+ unsigned int csize = osb->s_clustersize;
+ handle_t *handle;
+
+ /*
+ * The "start" and "end" values are NOT necessarily part of
+ * the range whose allocation is being deleted. Rather, this
+ * is what the user passed in with the request. We must zero
+ * partial clusters here. There's no need to worry about
+ * physical allocation - the zeroing code knows to skip holes.
+ */
+ mlog(0, "byte start: %llu, end: %llu\n",
+ (unsigned long long)start, (unsigned long long)end);
+
+ /*
+ * If both edges are on a cluster boundary then there's no
+ * zeroing required as the region is part of the allocation to
+ * be truncated.
+ */
+ if ((start & (csize - 1)) == 0 && (end & (csize - 1)) == 0)
+ goto out;
+
+ handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
+ if (handle == NULL) {
+ ret = -ENOMEM;
+ mlog_errno(ret);
+ goto out;
+ }
+
+ /*
+ * We want to get the byte offset of the end of the 1st cluster.
+ */
+ tmpend = (u64)osb->s_clustersize + (start & ~(osb->s_clustersize - 1));
+ if (tmpend > end)
+ tmpend = end;
+
+ mlog(0, "1st range: start: %llu, tmpend: %llu\n",
+ (unsigned long long)start, (unsigned long long)tmpend);
+
+ ret = ocfs2_zero_range_for_truncate(inode, handle, start, tmpend);
+ if (ret)
+ mlog_errno(ret);
+
+ if (tmpend < end) {
+ /*
+ * This may make start and end equal, but the zeroing
+ * code will skip any work in that case so there's no
+ * need to catch it up here.
+ */
+ start = end & ~(osb->s_clustersize - 1);
+
+ mlog(0, "2nd range: start: %llu, end: %llu\n",
+ (unsigned long long)start, (unsigned long long)end);
+
+ ret = ocfs2_zero_range_for_truncate(inode, handle, start, end);
+ if (ret)
+ mlog_errno(ret);
+ }
+
+ ocfs2_commit_trans(osb, handle);
+out:
+ return ret;
+}
+
+static int ocfs2_remove_inode_range(struct inode *inode,
+ struct buffer_head *di_bh, u64 byte_start,
+ u64 byte_len)
+{
+ int ret = 0;
+ u32 trunc_start, trunc_len, cpos, phys_cpos, alloc_size;
+ struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+ struct ocfs2_cached_dealloc_ctxt dealloc;
+
+ ocfs2_init_dealloc_ctxt(&dealloc);
+
+ if (byte_len == 0)
+ return 0;
+
+ trunc_start = ocfs2_clusters_for_bytes(osb->sb, byte_start);
+ trunc_len = (byte_start + byte_len) >> osb->s_clustersize_bits;
+ if (trunc_len >= trunc_start)
+ trunc_len -= trunc_start;
+ else
+ trunc_len = 0;
+
+ mlog(0, "Inode: %llu, start: %llu, len: %llu, cstart: %u, clen: %u\n",
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long long)byte_start,
+ (unsigned long long)byte_len, trunc_start, trunc_len);
+
+ ret = ocfs2_zero_partial_clusters(inode, byte_start, byte_len);
+ if (ret) {
+ mlog_errno(ret);
+ goto out;
+ }
+
+ cpos = trunc_start;
+ while (trunc_len) {
+ ret = ocfs2_get_clusters(inode, cpos, &phys_cpos,
+ &alloc_size, NULL);
+ if (ret) {
+ mlog_errno(ret);
+ goto out;
+ }
+
+ if (alloc_size > trunc_len)
+ alloc_size = trunc_len;
+
+ /* Only do work for non-holes */
+ if (phys_cpos != 0) {
+ ret = __ocfs2_remove_inode_range(inode, di_bh, cpos,
+ phys_cpos, alloc_size,
+ &dealloc);
+ if (ret) {
+ mlog_errno(ret);
+ goto out;
+ }
+ }
+
+ cpos += alloc_size;
+ trunc_len -= alloc_size;
+ }
+
+ ocfs2_truncate_cluster_pages(inode, byte_start, byte_len);
+
+out:
+ ocfs2_schedule_truncate_log_flush(osb, 1);
+ ocfs2_run_deallocs(osb, &dealloc);
+
+ return ret;
+}
+
static int ocfs2_prepare_inode_for_write(struct dentry *dentry,
loff_t *ppos,
size_t count,
diff --git a/fs/ocfs2/journal.h b/fs/ocfs2/journal.h
index 3db5de4506d..ce60aab013a 100644
--- a/fs/ocfs2/journal.h
+++ b/fs/ocfs2/journal.h
@@ -289,6 +289,8 @@ int ocfs2_journal_dirty_data(handle_t *handle,
#define OCFS2_TRUNCATE_LOG_FLUSH_ONE_REC (OCFS2_SUBALLOC_FREE \
+ OCFS2_TRUNCATE_LOG_UPDATE)
+#define OCFS2_REMOVE_EXTENT_CREDITS (OCFS2_TRUNCATE_LOG_UPDATE + OCFS2_INODE_UPDATE_CREDITS)
+
/* data block for new dir/symlink, 2 for bitmap updates (bitmap fe +
* bitmap block for the new bit) */
#define OCFS2_DIR_LINK_ADDITIONAL_CREDITS (1 + 2)