From 97e728d4353f38c87bf0804cdfd79a9b13fc2c3e Mon Sep 17 00:00:00 2001
From: Josef Bacik <jbacik@redhat.com>
Date: Tue, 21 Apr 2009 17:40:57 -0400
Subject: Btrfs: try to keep a healthy ratio of metadata vs data block groups

This patch makes the chunk allocator keep a good ratio of metadata vs data
block groups.  By default for every 8 data block groups, we'll allocate 1
metadata chunk, or about 12% of the disk will be allocated for metadata.  This
can be changed by specifying the metadata_ratio mount option.

This is simply the number of data block groups that have to be allocated to
force a metadata chunk allocation.  By making sure we allocate metadata chunks
more often, we are less likely to get into situations where the whole disk
has been allocated as data block groups.

Signed-off-by: Josef Bacik <jbacik@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/ctree.h       |  3 +++
 fs/btrfs/disk-io.c     |  1 +
 fs/btrfs/extent-tree.c | 28 +++++++++++++++++++++++++++-
 fs/btrfs/super.c       | 12 +++++++++++-
 4 files changed, 42 insertions(+), 2 deletions(-)

diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index ad96495dedc..213535f45da 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -881,6 +881,9 @@ struct btrfs_fs_info {
 	u64 metadata_alloc_profile;
 	u64 system_alloc_profile;
 
+	unsigned data_chunk_allocations;
+	unsigned metadata_ratio;
+
 	void *bdev_holder;
 };
 
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index a6b83744b05..44c94d808e2 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -1604,6 +1604,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	fs_info->btree_inode = new_inode(sb);
 	fs_info->btree_inode->i_ino = 1;
 	fs_info->btree_inode->i_nlink = 1;
+	fs_info->metadata_ratio = 8;
 
 	fs_info->thread_pool_size = min_t(unsigned long,
 					  num_online_cpus() + 2, 8);
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 178df4c67de..2895a837323 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -1918,15 +1918,29 @@ void btrfs_delalloc_free_space(struct btrfs_root *root, struct inode *inode,
 	spin_unlock(&info->lock);
 }
 
+static void force_metadata_allocation(struct btrfs_fs_info *info)
+{
+	struct list_head *head = &info->space_info;
+	struct btrfs_space_info *found;
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(found, head, list) {
+		if (found->flags & BTRFS_BLOCK_GROUP_METADATA)
+			found->force_alloc = 1;
+	}
+	rcu_read_unlock();
+}
+
 static int do_chunk_alloc(struct btrfs_trans_handle *trans,
 			  struct btrfs_root *extent_root, u64 alloc_bytes,
 			  u64 flags, int force)
 {
 	struct btrfs_space_info *space_info;
+	struct btrfs_fs_info *fs_info = extent_root->fs_info;
 	u64 thresh;
 	int ret = 0;
 
-	mutex_lock(&extent_root->fs_info->chunk_mutex);
+	mutex_lock(&fs_info->chunk_mutex);
 
 	flags = btrfs_reduce_alloc_profile(extent_root, flags);
 
@@ -1958,6 +1972,18 @@ static int do_chunk_alloc(struct btrfs_trans_handle *trans,
 	}
 	spin_unlock(&space_info->lock);
 
+	/*
+	 * if we're doing a data chunk, go ahead and make sure that
+	 * we keep a reasonable number of metadata chunks allocated in the
+	 * FS as well.
+	 */
+	if (flags & BTRFS_BLOCK_GROUP_DATA) {
+		fs_info->data_chunk_allocations++;
+		if (!(fs_info->data_chunk_allocations %
+		      fs_info->metadata_ratio))
+			force_metadata_allocation(fs_info);
+	}
+
 	ret = btrfs_alloc_chunk(trans, extent_root, flags);
 	if (ret)
 		space_info->full = 1;
diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index 9744af9d71e..30c9a8ca2a5 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -68,7 +68,7 @@ enum {
 	Opt_degraded, Opt_subvol, Opt_device, Opt_nodatasum, Opt_nodatacow,
 	Opt_max_extent, Opt_max_inline, Opt_alloc_start, Opt_nobarrier,
 	Opt_ssd, Opt_thread_pool, Opt_noacl,  Opt_compress, Opt_notreelog,
-	Opt_flushoncommit, Opt_err,
+	Opt_ratio, Opt_flushoncommit, Opt_err,
 };
 
 static match_table_t tokens = {
@@ -87,6 +87,7 @@ static match_table_t tokens = {
 	{Opt_noacl, "noacl"},
 	{Opt_notreelog, "notreelog"},
 	{Opt_flushoncommit, "flushoncommit"},
+	{Opt_ratio, "metadata_ratio=%d"},
 	{Opt_err, NULL},
 };
 
@@ -234,6 +235,15 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
 			printk(KERN_INFO "btrfs: turning on flush-on-commit\n");
 			btrfs_set_opt(info->mount_opt, FLUSHONCOMMIT);
 			break;
+		case Opt_ratio:
+			intarg = 0;
+			match_int(&args[0], &intarg);
+			if (intarg) {
+				info->metadata_ratio = intarg;
+				printk(KERN_INFO "btrfs: metadata ratio %d\n",
+				       info->metadata_ratio);
+			}
+			break;
 		default:
 			break;
 		}
-- 
cgit v1.2.3


From 2ea2544ef5dad5cac52f1e4c7b812631274fc1cb Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@lst.de>
Date: Mon, 13 Apr 2009 15:32:28 +0200
Subject: Btrfs: simplify makefile

Get rid of the hacks for building out of tree, and always use += for
assigning to the object lists.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/Makefile | 19 ++-----------------
 1 file changed, 2 insertions(+), 17 deletions(-)

diff --git a/fs/btrfs/Makefile b/fs/btrfs/Makefile
index 9adf5e4f7e9..94212844a9b 100644
--- a/fs/btrfs/Makefile
+++ b/fs/btrfs/Makefile
@@ -1,25 +1,10 @@
-ifneq ($(KERNELRELEASE),)
-# kbuild part of makefile
 
 obj-$(CONFIG_BTRFS_FS) := btrfs.o
-btrfs-y := super.o ctree.o extent-tree.o print-tree.o root-tree.o dir-item.o \
+
+btrfs-y += super.o ctree.o extent-tree.o print-tree.o root-tree.o dir-item.o \
 	   file-item.o inode-item.o inode-map.o disk-io.o \
 	   transaction.o inode.o file.o tree-defrag.o \
 	   extent_map.o sysfs.o struct-funcs.o xattr.o ordered-data.o \
 	   extent_io.o volumes.o async-thread.o ioctl.o locking.o orphan.o \
 	   ref-cache.o export.o tree-log.o acl.o free-space-cache.o zlib.o \
 	   compression.o delayed-ref.o
-else
-
-# Normal Makefile
-
-KERNELDIR := /lib/modules/`uname -r`/build
-all:
-	$(MAKE) -C $(KERNELDIR) M=`pwd` CONFIG_BTRFS_FS=m modules
-
-modules_install:
-	$(MAKE) -C $(KERNELDIR) M=`pwd` modules_install
-clean:
-	$(MAKE) -C $(KERNELDIR) M=`pwd` clean
-
-endif
-- 
cgit v1.2.3


From 0d4bf11e5309eff64272a49e1ea55658372abc56 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@lst.de>
Date: Mon, 13 Apr 2009 15:33:56 +0200
Subject: Btrfs: don't export symbols

Currently the extent_map code is only for btrfs so don't export it's
symbols.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/extent_map.c | 6 ------
 1 file changed, 6 deletions(-)

diff --git a/fs/btrfs/extent_map.c b/fs/btrfs/extent_map.c
index b187917b36f..9827fa1de4e 100644
--- a/fs/btrfs/extent_map.c
+++ b/fs/btrfs/extent_map.c
@@ -43,7 +43,6 @@ void extent_map_tree_init(struct extent_map_tree *tree, gfp_t mask)
 	tree->map.rb_node = NULL;
 	spin_lock_init(&tree->lock);
 }
-EXPORT_SYMBOL(extent_map_tree_init);
 
 /**
  * alloc_extent_map - allocate new extent map structure
@@ -64,7 +63,6 @@ struct extent_map *alloc_extent_map(gfp_t mask)
 	atomic_set(&em->refs, 1);
 	return em;
 }
-EXPORT_SYMBOL(alloc_extent_map);
 
 /**
  * free_extent_map - drop reference count of an extent_map
@@ -83,7 +81,6 @@ void free_extent_map(struct extent_map *em)
 		kmem_cache_free(extent_map_cache, em);
 	}
 }
-EXPORT_SYMBOL(free_extent_map);
 
 static struct rb_node *tree_insert(struct rb_root *root, u64 offset,
 				   struct rb_node *node)
@@ -264,7 +261,6 @@ int add_extent_mapping(struct extent_map_tree *tree,
 out:
 	return ret;
 }
-EXPORT_SYMBOL(add_extent_mapping);
 
 /* simple helper to do math around the end of an extent, handling wrap */
 static u64 range_end(u64 start, u64 len)
@@ -326,7 +322,6 @@ found:
 out:
 	return em;
 }
-EXPORT_SYMBOL(lookup_extent_mapping);
 
 /**
  * remove_extent_mapping - removes an extent_map from the extent tree
@@ -346,4 +341,3 @@ int remove_extent_mapping(struct extent_map_tree *tree, struct extent_map *em)
 	em->in_tree = 0;
 	return ret;
 }
-EXPORT_SYMBOL(remove_extent_mapping);
-- 
cgit v1.2.3


From 9601e3f6336f6ca66929f451b1f66085e68e36e3 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@lst.de>
Date: Mon, 13 Apr 2009 15:33:09 +0200
Subject: Btrfs: kill btrfs_cache_create

Just use kmem_cache_create directly.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/extent_io.c  | 18 ++++++------------
 fs/btrfs/extent_map.c | 11 +++--------
 fs/btrfs/inode.c      | 42 +++++++++++++++++++-----------------------
 3 files changed, 28 insertions(+), 43 deletions(-)

diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index 05a1c42e25b..c33b54029d7 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -17,12 +17,6 @@
 #include "ctree.h"
 #include "btrfs_inode.h"
 
-/* temporary define until extent_map moves out of btrfs */
-struct kmem_cache *btrfs_cache_create(const char *name, size_t size,
-				       unsigned long extra_flags,
-				       void (*ctor)(void *, struct kmem_cache *,
-						    unsigned long));
-
 static struct kmem_cache *extent_state_cache;
 static struct kmem_cache *extent_buffer_cache;
 
@@ -58,15 +52,15 @@ struct extent_page_data {
 
 int __init extent_io_init(void)
 {
-	extent_state_cache = btrfs_cache_create("extent_state",
-					    sizeof(struct extent_state), 0,
-					    NULL);
+	extent_state_cache = kmem_cache_create("extent_state",
+			sizeof(struct extent_state), 0,
+			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
 	if (!extent_state_cache)
 		return -ENOMEM;
 
-	extent_buffer_cache = btrfs_cache_create("extent_buffers",
-					    sizeof(struct extent_buffer), 0,
-					    NULL);
+	extent_buffer_cache = kmem_cache_create("extent_buffers",
+			sizeof(struct extent_buffer), 0,
+			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
 	if (!extent_buffer_cache)
 		goto free_state_cache;
 	return 0;
diff --git a/fs/btrfs/extent_map.c b/fs/btrfs/extent_map.c
index 9827fa1de4e..30c9365861e 100644
--- a/fs/btrfs/extent_map.c
+++ b/fs/btrfs/extent_map.c
@@ -6,19 +6,14 @@
 #include <linux/hardirq.h>
 #include "extent_map.h"
 
-/* temporary define until extent_map moves out of btrfs */
-struct kmem_cache *btrfs_cache_create(const char *name, size_t size,
-				       unsigned long extra_flags,
-				       void (*ctor)(void *, struct kmem_cache *,
-						    unsigned long));
 
 static struct kmem_cache *extent_map_cache;
 
 int __init extent_map_init(void)
 {
-	extent_map_cache = btrfs_cache_create("extent_map",
-					    sizeof(struct extent_map), 0,
-					    NULL);
+	extent_map_cache = kmem_cache_create("extent_map",
+			sizeof(struct extent_map), 0,
+			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
 	if (!extent_map_cache)
 		return -ENOMEM;
 	return 0;
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 65219f6a16a..176b6cc28b1 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -4640,39 +4640,35 @@ void btrfs_destroy_cachep(void)
 		kmem_cache_destroy(btrfs_path_cachep);
 }
 
-struct kmem_cache *btrfs_cache_create(const char *name, size_t size,
-				       unsigned long extra_flags,
-				       void (*ctor)(void *))
-{
-	return kmem_cache_create(name, size, 0, (SLAB_RECLAIM_ACCOUNT |
-				 SLAB_MEM_SPREAD | extra_flags), ctor);
-}
-
 int btrfs_init_cachep(void)
 {
-	btrfs_inode_cachep = btrfs_cache_create("btrfs_inode_cache",
-					  sizeof(struct btrfs_inode),
-					  0, init_once);
+	btrfs_inode_cachep = kmem_cache_create("btrfs_inode_cache",
+			sizeof(struct btrfs_inode), 0,
+			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, init_once);
 	if (!btrfs_inode_cachep)
 		goto fail;
-	btrfs_trans_handle_cachep =
-			btrfs_cache_create("btrfs_trans_handle_cache",
-					   sizeof(struct btrfs_trans_handle),
-					   0, NULL);
+
+	btrfs_trans_handle_cachep = kmem_cache_create("btrfs_trans_handle_cache",
+			sizeof(struct btrfs_trans_handle), 0,
+			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
 	if (!btrfs_trans_handle_cachep)
 		goto fail;
-	btrfs_transaction_cachep = btrfs_cache_create("btrfs_transaction_cache",
-					     sizeof(struct btrfs_transaction),
-					     0, NULL);
+
+	btrfs_transaction_cachep = kmem_cache_create("btrfs_transaction_cache",
+			sizeof(struct btrfs_transaction), 0,
+			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
 	if (!btrfs_transaction_cachep)
 		goto fail;
-	btrfs_path_cachep = btrfs_cache_create("btrfs_path_cache",
-					 sizeof(struct btrfs_path),
-					 0, NULL);
+
+	btrfs_path_cachep = kmem_cache_create("btrfs_path_cache",
+			sizeof(struct btrfs_path), 0,
+			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD, NULL);
 	if (!btrfs_path_cachep)
 		goto fail;
-	btrfs_bit_radix_cachep = btrfs_cache_create("btrfs_radix", 256,
-					      SLAB_DESTROY_BY_RCU, NULL);
+
+	btrfs_bit_radix_cachep = kmem_cache_create("btrfs_radix", 256, 0,
+			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD |
+			SLAB_DESTROY_BY_RCU, NULL);
 	if (!btrfs_bit_radix_cachep)
 		goto fail;
 	return 0;
-- 
cgit v1.2.3


From e980b50cda1610f1c17978d9b7fd311a9dd93877 Mon Sep 17 00:00:00 2001
From: Chris Mason <chris.mason@oracle.com>
Date: Fri, 24 Apr 2009 14:39:24 -0400
Subject: Btrfs: fix fallocate deadlock on inode extent lock

The btrfs fallocate call takes an extent lock on the entire range
being fallocated, and then runs through insert_reserved_extent on each
extent as they are allocated.

The problem with this is that btrfs_drop_extents may decide to try
and take the same extent lock fallocate was already holding.  The solution
used here is to push down knowledge of the range that is already locked
going into btrfs_drop_extents.

It turns out that at least one other caller had the same bug.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/ctree.h    |  3 ++-
 fs/btrfs/file.c     | 11 ++++++-----
 fs/btrfs/inode.c    | 27 ++++++++++++++++++---------
 fs/btrfs/ioctl.c    |  3 ++-
 fs/btrfs/tree-log.c |  2 +-
 5 files changed, 29 insertions(+), 17 deletions(-)

diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 213535f45da..4414a5d9983 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -2177,7 +2177,8 @@ int btrfs_check_file(struct btrfs_root *root, struct inode *inode);
 extern struct file_operations btrfs_file_operations;
 int btrfs_drop_extents(struct btrfs_trans_handle *trans,
 		       struct btrfs_root *root, struct inode *inode,
-		       u64 start, u64 end, u64 inline_limit, u64 *hint_block);
+		       u64 start, u64 end, u64 locked_end,
+		       u64 inline_limit, u64 *hint_block);
 int btrfs_mark_extent_written(struct btrfs_trans_handle *trans,
 			      struct btrfs_root *root,
 			      struct inode *inode, u64 start, u64 end);
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index 482f8db2cfd..da3ed965c95 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -363,15 +363,16 @@ out:
  */
 noinline int btrfs_drop_extents(struct btrfs_trans_handle *trans,
 		       struct btrfs_root *root, struct inode *inode,
-		       u64 start, u64 end, u64 inline_limit, u64 *hint_byte)
+		       u64 start, u64 end, u64 locked_end,
+		       u64 inline_limit, u64 *hint_byte)
 {
 	u64 extent_end = 0;
-	u64 locked_end = end;
 	u64 search_start = start;
 	u64 leaf_start;
 	u64 ram_bytes = 0;
 	u64 orig_parent = 0;
 	u64 disk_bytenr = 0;
+	u64 orig_locked_end = locked_end;
 	u8 compression;
 	u8 encryption;
 	u16 other_encoding = 0;
@@ -684,9 +685,9 @@ next_slot:
 	}
 out:
 	btrfs_free_path(path);
-	if (locked_end > end) {
-		unlock_extent(&BTRFS_I(inode)->io_tree, end, locked_end - 1,
-			      GFP_NOFS);
+	if (locked_end > orig_locked_end) {
+		unlock_extent(&BTRFS_I(inode)->io_tree, orig_locked_end,
+			      locked_end - 1, GFP_NOFS);
 	}
 	btrfs_check_file(root, inode);
 	return ret;
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 176b6cc28b1..2fdb2995be6 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -234,7 +234,7 @@ static noinline int cow_file_range_inline(struct btrfs_trans_handle *trans,
 	}
 
 	ret = btrfs_drop_extents(trans, root, inode, start,
-				 aligned_end, start, &hint_byte);
+				 aligned_end, aligned_end, start, &hint_byte);
 	BUG_ON(ret);
 
 	if (isize > actual_end)
@@ -1439,6 +1439,7 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans,
 				       struct inode *inode, u64 file_pos,
 				       u64 disk_bytenr, u64 disk_num_bytes,
 				       u64 num_bytes, u64 ram_bytes,
+				       u64 locked_end,
 				       u8 compression, u8 encryption,
 				       u16 other_encoding, int extent_type)
 {
@@ -1455,7 +1456,8 @@ static int insert_reserved_file_extent(struct btrfs_trans_handle *trans,
 
 	path->leave_spinning = 1;
 	ret = btrfs_drop_extents(trans, root, inode, file_pos,
-				 file_pos + num_bytes, file_pos, &hint);
+				 file_pos + num_bytes, locked_end,
+				 file_pos, &hint);
 	BUG_ON(ret);
 
 	ins.objectid = inode->i_ino;
@@ -1590,6 +1592,8 @@ static int btrfs_finish_ordered_io(struct inode *inode, u64 start, u64 end)
 						ordered_extent->disk_len,
 						ordered_extent->len,
 						ordered_extent->len,
+						ordered_extent->file_offset +
+						ordered_extent->len,
 						compressed, 0, 0,
 						BTRFS_FILE_EXTENT_REG);
 		BUG_ON(ret);
@@ -2877,6 +2881,7 @@ int btrfs_cont_expand(struct inode *inode, loff_t size)
 			err = btrfs_drop_extents(trans, root, inode,
 						 cur_offset,
 						 cur_offset + hole_size,
+						 block_end,
 						 cur_offset, &hint_byte);
 			if (err)
 				break;
@@ -4968,7 +4973,7 @@ out_fail:
 
 static int prealloc_file_range(struct btrfs_trans_handle *trans,
 			       struct inode *inode, u64 start, u64 end,
-			       u64 alloc_hint, int mode)
+			       u64 locked_end, u64 alloc_hint, int mode)
 {
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_key ins;
@@ -4989,7 +4994,8 @@ static int prealloc_file_range(struct btrfs_trans_handle *trans,
 		ret = insert_reserved_file_extent(trans, inode,
 						  cur_offset, ins.objectid,
 						  ins.offset, ins.offset,
-						  ins.offset, 0, 0, 0,
+						  ins.offset, locked_end,
+						  0, 0, 0,
 						  BTRFS_FILE_EXTENT_PREALLOC);
 		BUG_ON(ret);
 		num_bytes -= ins.offset;
@@ -5018,6 +5024,7 @@ static long btrfs_fallocate(struct inode *inode, int mode,
 	u64 alloc_start;
 	u64 alloc_end;
 	u64 alloc_hint = 0;
+	u64 locked_end;
 	u64 mask = BTRFS_I(inode)->root->sectorsize - 1;
 	struct extent_map *em;
 	struct btrfs_trans_handle *trans;
@@ -5039,6 +5046,7 @@ static long btrfs_fallocate(struct inode *inode, int mode,
 			goto out;
 	}
 
+	locked_end = alloc_end - 1;
 	while (1) {
 		struct btrfs_ordered_extent *ordered;
 
@@ -5051,8 +5059,8 @@ static long btrfs_fallocate(struct inode *inode, int mode,
 		/* the extent lock is ordered inside the running
 		 * transaction
 		 */
-		lock_extent(&BTRFS_I(inode)->io_tree, alloc_start,
-			    alloc_end - 1, GFP_NOFS);
+		lock_extent(&BTRFS_I(inode)->io_tree, alloc_start, locked_end,
+			    GFP_NOFS);
 		ordered = btrfs_lookup_first_ordered_extent(inode,
 							    alloc_end - 1);
 		if (ordered &&
@@ -5060,7 +5068,7 @@ static long btrfs_fallocate(struct inode *inode, int mode,
 		    ordered->file_offset < alloc_end) {
 			btrfs_put_ordered_extent(ordered);
 			unlock_extent(&BTRFS_I(inode)->io_tree,
-				      alloc_start, alloc_end - 1, GFP_NOFS);
+				      alloc_start, locked_end, GFP_NOFS);
 			btrfs_end_transaction(trans, BTRFS_I(inode)->root);
 
 			/*
@@ -5085,7 +5093,8 @@ static long btrfs_fallocate(struct inode *inode, int mode,
 		last_byte = (last_byte + mask) & ~mask;
 		if (em->block_start == EXTENT_MAP_HOLE) {
 			ret = prealloc_file_range(trans, inode, cur_offset,
-					last_byte, alloc_hint, mode);
+					last_byte, locked_end + 1,
+					alloc_hint, mode);
 			if (ret < 0) {
 				free_extent_map(em);
 				break;
@@ -5101,7 +5110,7 @@ static long btrfs_fallocate(struct inode *inode, int mode,
 			break;
 		}
 	}
-	unlock_extent(&BTRFS_I(inode)->io_tree, alloc_start, alloc_end - 1,
+	unlock_extent(&BTRFS_I(inode)->io_tree, alloc_start, locked_end,
 		      GFP_NOFS);
 
 	btrfs_end_transaction(trans, BTRFS_I(inode)->root);
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index 7594bec1be1..f4e5d2e5ece 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -830,7 +830,8 @@ static long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 	BUG_ON(!trans);
 
 	/* punch hole in destination first */
-	btrfs_drop_extents(trans, root, inode, off, off+len, 0, &hint_byte);
+	btrfs_drop_extents(trans, root, inode, off, off + len,
+			   off + len, 0, &hint_byte);
 
 	/* clone data */
 	key.objectid = src->i_ino;
diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c
index 25f20ea11f2..db5e212e844 100644
--- a/fs/btrfs/tree-log.c
+++ b/fs/btrfs/tree-log.c
@@ -536,7 +536,7 @@ static noinline int replay_one_extent(struct btrfs_trans_handle *trans,
 	saved_nbytes = inode_get_bytes(inode);
 	/* drop any overlapping extents */
 	ret = btrfs_drop_extents(trans, root, inode,
-			 start, extent_end, start, &alloc_hint);
+			 start, extent_end, extent_end, start, &alloc_hint);
 	BUG_ON(ret);
 
 	if (found_type == BTRFS_FILE_EXTENT_REG ||
-- 
cgit v1.2.3


From 59bc5c758ece00fb0b2a170dd8fbbf31f1856c8a Mon Sep 17 00:00:00 2001
From: Chris Mason <chris.mason@oracle.com>
Date: Fri, 24 Apr 2009 14:39:25 -0400
Subject: Btrfs: fix deadlocks and stalls on dead root removal

After a transaction commit, the old root of the subvol btrees are sent through
snapshot removal.  This is what actually frees up any blocks replaced by
COW, and anything the old blocks pointed to.

Snapshot deletion will pause when a transaction commit has started, which
helps to avoid a huge amount of delayed reference count updates piling up
as the transaction is trying to close.

But, this pause happens after the snapshot deletion process has asked other
procs on the system to throttle back a bit so that it can make progress.

We don't want to throttle everyone while we're waiting for the transaction
commit, it leads to deadlocks in the user transaction ioctls used by Ceph
and makes things slower in general.

This patch changes things to avoid the throttling while we sleep.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/transaction.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/fs/btrfs/transaction.c b/fs/btrfs/transaction.c
index 2869b3361eb..01b143605ec 100644
--- a/fs/btrfs/transaction.c
+++ b/fs/btrfs/transaction.c
@@ -687,7 +687,13 @@ static noinline int wait_transaction_pre_flush(struct btrfs_fs_info *info)
 		prepare_to_wait(&info->transaction_wait, &wait,
 				TASK_UNINTERRUPTIBLE);
 		mutex_unlock(&info->trans_mutex);
+
+		atomic_dec(&info->throttles);
+		wake_up(&info->transaction_throttle);
+
 		schedule();
+
+		atomic_inc(&info->throttles);
 		mutex_lock(&info->trans_mutex);
 		finish_wait(&info->transaction_wait, &wait);
 	}
-- 
cgit v1.2.3


From d6397baee468809ef311e763dfc6e9f73418f8a6 Mon Sep 17 00:00:00 2001
From: Chris Ball <cjb@laptop.org>
Date: Mon, 27 Apr 2009 07:29:03 -0400
Subject: Btrfs: When shrinking, only update disk size on success

Previously, we updated a device's size prior to attempting a shrink
operation.  This patch moves the device resizing logic to only happen if
the shrink completes successfully.  In the process, it introduces a new
field to btrfs_device -- disk_total_bytes -- to track the on-disk size.

Signed-off-by: Chris Ball <cjb@laptop.org>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/volumes.c | 35 ++++++++++++++++++++++++-----------
 fs/btrfs/volumes.h |  3 +++
 2 files changed, 27 insertions(+), 11 deletions(-)

diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index e53835b8859..5f01dad4b69 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -1543,7 +1543,7 @@ static noinline int btrfs_update_device(struct btrfs_trans_handle *trans,
 	btrfs_set_device_io_align(leaf, dev_item, device->io_align);
 	btrfs_set_device_io_width(leaf, dev_item, device->io_width);
 	btrfs_set_device_sector_size(leaf, dev_item, device->sector_size);
-	btrfs_set_device_total_bytes(leaf, dev_item, device->total_bytes);
+	btrfs_set_device_total_bytes(leaf, dev_item, device->disk_total_bytes);
 	btrfs_set_device_bytes_used(leaf, dev_item, device->bytes_used);
 	btrfs_mark_buffer_dirty(leaf);
 
@@ -1940,14 +1940,6 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size)
 	device->total_bytes = new_size;
 	if (device->writeable)
 		device->fs_devices->total_rw_bytes -= diff;
-	ret = btrfs_update_device(trans, device);
-	if (ret) {
-		unlock_chunks(root);
-		btrfs_end_transaction(trans, root);
-		goto done;
-	}
-	WARN_ON(diff > old_total);
-	btrfs_set_super_total_bytes(super_copy, old_total - diff);
 	unlock_chunks(root);
 	btrfs_end_transaction(trans, root);
 
@@ -1979,7 +1971,7 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size)
 		length = btrfs_dev_extent_length(l, dev_extent);
 
 		if (key.offset + length <= new_size)
-			goto done;
+			break;
 
 		chunk_tree = btrfs_dev_extent_chunk_tree(l, dev_extent);
 		chunk_objectid = btrfs_dev_extent_chunk_objectid(l, dev_extent);
@@ -1992,6 +1984,26 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size)
 			goto done;
 	}
 
+	/* Shrinking succeeded, else we would be at "done". */
+	trans = btrfs_start_transaction(root, 1);
+	if (!trans) {
+		ret = -ENOMEM;
+		goto done;
+	}
+	lock_chunks(root);
+
+	device->disk_total_bytes = new_size;
+	/* Now btrfs_update_device() will change the on-disk size. */
+	ret = btrfs_update_device(trans, device);
+	if (ret) {
+		unlock_chunks(root);
+		btrfs_end_transaction(trans, root);
+		goto done;
+	}
+	WARN_ON(diff > old_total);
+	btrfs_set_super_total_bytes(super_copy, old_total - diff);
+	unlock_chunks(root);
+	btrfs_end_transaction(trans, root);
 done:
 	btrfs_free_path(path);
 	return ret;
@@ -3076,7 +3088,8 @@ static int fill_device_from_item(struct extent_buffer *leaf,
 	unsigned long ptr;
 
 	device->devid = btrfs_device_id(leaf, dev_item);
-	device->total_bytes = btrfs_device_total_bytes(leaf, dev_item);
+	device->disk_total_bytes = btrfs_device_total_bytes(leaf, dev_item);
+	device->total_bytes = device->disk_total_bytes;
 	device->bytes_used = btrfs_device_bytes_used(leaf, dev_item);
 	device->type = btrfs_device_type(leaf, dev_item);
 	device->io_align = btrfs_device_io_align(leaf, dev_item);
diff --git a/fs/btrfs/volumes.h b/fs/btrfs/volumes.h
index 5836327ba5d..5c3ff6d02fd 100644
--- a/fs/btrfs/volumes.h
+++ b/fs/btrfs/volumes.h
@@ -61,6 +61,9 @@ struct btrfs_device {
 	/* size of the device */
 	u64 total_bytes;
 
+	/* size of the disk */
+	u64 disk_total_bytes;
+
 	/* bytes used */
 	u64 bytes_used;
 
-- 
cgit v1.2.3


From b7967db75a38df4891b22efe1b0969b9357eb946 Mon Sep 17 00:00:00 2001
From: Chris Mason <chris.mason@oracle.com>
Date: Mon, 27 Apr 2009 07:29:04 -0400
Subject: Btrfs: remove #if 0 code

Btrfs had some old code sitting around under #if 0, this drops it.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/disk-io.c   | 48 +-------------------------------
 fs/btrfs/extent_io.c | 63 ------------------------------------------
 fs/btrfs/file.c      | 78 ----------------------------------------------------
 3 files changed, 1 insertion(+), 188 deletions(-)

diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 44c94d808e2..77f9a3b824b 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -584,18 +584,7 @@ int btrfs_wq_submit_bio(struct btrfs_fs_info *fs_info, struct inode *inode,
 		btrfs_set_work_high_prio(&async->work);
 
 	btrfs_queue_worker(&fs_info->workers, &async->work);
-#if 0
-	int limit = btrfs_async_submit_limit(fs_info);
-	if (atomic_read(&fs_info->nr_async_submits) > limit) {
-		wait_event_timeout(fs_info->async_submit_wait,
-			   (atomic_read(&fs_info->nr_async_submits) < limit),
-			   HZ/10);
 
-		wait_event_timeout(fs_info->async_submit_wait,
-			   (atomic_read(&fs_info->nr_async_bios) < limit),
-			   HZ/10);
-	}
-#endif
 	while (atomic_read(&fs_info->async_submit_draining) &&
 	      atomic_read(&fs_info->nr_async_submits)) {
 		wait_event(fs_info->async_submit_wait,
@@ -770,27 +759,6 @@ static void btree_invalidatepage(struct page *page, unsigned long offset)
 	}
 }
 
-#if 0
-static int btree_writepage(struct page *page, struct writeback_control *wbc)
-{
-	struct buffer_head *bh;
-	struct btrfs_root *root = BTRFS_I(page->mapping->host)->root;
-	struct buffer_head *head;
-	if (!page_has_buffers(page)) {
-		create_empty_buffers(page, root->fs_info->sb->s_blocksize,
-					(1 << BH_Dirty)|(1 << BH_Uptodate));
-	}
-	head = page_buffers(page);
-	bh = head;
-	do {
-		if (buffer_dirty(bh))
-			csum_tree_block(root, bh, 0);
-		bh = bh->b_this_page;
-	} while (bh != head);
-	return block_write_full_page(page, btree_get_block, wbc);
-}
-#endif
-
 static struct address_space_operations btree_aops = {
 	.readpage	= btree_readpage,
 	.writepage	= btree_writepage,
@@ -1278,11 +1246,7 @@ static int btrfs_congested_fn(void *congested_data, int bdi_bits)
 	int ret = 0;
 	struct btrfs_device *device;
 	struct backing_dev_info *bdi;
-#if 0
-	if ((bdi_bits & (1 << BDI_write_congested)) &&
-	    btrfs_congested_async(info, 0))
-		return 1;
-#endif
+
 	list_for_each_entry(device, &info->fs_devices->devices, dev_list) {
 		if (!device->bdev)
 			continue;
@@ -2334,16 +2298,6 @@ int close_ctree(struct btrfs_root *root)
 	btrfs_stop_workers(&fs_info->endio_write_workers);
 	btrfs_stop_workers(&fs_info->submit_workers);
 
-#if 0
-	while (!list_empty(&fs_info->hashers)) {
-		struct btrfs_hasher *hasher;
-		hasher = list_entry(fs_info->hashers.next, struct btrfs_hasher,
-				    hashers);
-		list_del(&hasher->hashers);
-		crypto_free_hash(&fs_info->hash_tfm);
-		kfree(hasher);
-	}
-#endif
 	btrfs_close_devices(fs_info->fs_devices);
 	btrfs_mapping_tree_free(&fs_info->mapping_tree);
 
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index c33b54029d7..fe9eb990e44 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -1401,69 +1401,6 @@ out:
 	return total_bytes;
 }
 
-#if 0
-/*
- * helper function to lock both pages and extents in the tree.
- * pages must be locked first.
- */
-static int lock_range(struct extent_io_tree *tree, u64 start, u64 end)
-{
-	unsigned long index = start >> PAGE_CACHE_SHIFT;
-	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
-	struct page *page;
-	int err;
-
-	while (index <= end_index) {
-		page = grab_cache_page(tree->mapping, index);
-		if (!page) {
-			err = -ENOMEM;
-			goto failed;
-		}
-		if (IS_ERR(page)) {
-			err = PTR_ERR(page);
-			goto failed;
-		}
-		index++;
-	}
-	lock_extent(tree, start, end, GFP_NOFS);
-	return 0;
-
-failed:
-	/*
-	 * we failed above in getting the page at 'index', so we undo here
-	 * up to but not including the page at 'index'
-	 */
-	end_index = index;
-	index = start >> PAGE_CACHE_SHIFT;
-	while (index < end_index) {
-		page = find_get_page(tree->mapping, index);
-		unlock_page(page);
-		page_cache_release(page);
-		index++;
-	}
-	return err;
-}
-
-/*
- * helper function to unlock both pages and extents in the tree.
- */
-static int unlock_range(struct extent_io_tree *tree, u64 start, u64 end)
-{
-	unsigned long index = start >> PAGE_CACHE_SHIFT;
-	unsigned long end_index = end >> PAGE_CACHE_SHIFT;
-	struct page *page;
-
-	while (index <= end_index) {
-		page = find_get_page(tree->mapping, index);
-		unlock_page(page);
-		page_cache_release(page);
-		index++;
-	}
-	unlock_extent(tree, start, end, GFP_NOFS);
-	return 0;
-}
-#endif
-
 /*
  * set the private field for a given byte offset in the tree.  If there isn't
  * an extent_state there already, this does nothing.
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index da3ed965c95..1d51dc38bb4 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -272,83 +272,6 @@ int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end,
 	return 0;
 }
 
-int btrfs_check_file(struct btrfs_root *root, struct inode *inode)
-{
-	return 0;
-#if 0
-	struct btrfs_path *path;
-	struct btrfs_key found_key;
-	struct extent_buffer *leaf;
-	struct btrfs_file_extent_item *extent;
-	u64 last_offset = 0;
-	int nritems;
-	int slot;
-	int found_type;
-	int ret;
-	int err = 0;
-	u64 extent_end = 0;
-
-	path = btrfs_alloc_path();
-	ret = btrfs_lookup_file_extent(NULL, root, path, inode->i_ino,
-				       last_offset, 0);
-	while (1) {
-		nritems = btrfs_header_nritems(path->nodes[0]);
-		if (path->slots[0] >= nritems) {
-			ret = btrfs_next_leaf(root, path);
-			if (ret)
-				goto out;
-			nritems = btrfs_header_nritems(path->nodes[0]);
-		}
-		slot = path->slots[0];
-		leaf = path->nodes[0];
-		btrfs_item_key_to_cpu(leaf, &found_key, slot);
-		if (found_key.objectid != inode->i_ino)
-			break;
-		if (found_key.type != BTRFS_EXTENT_DATA_KEY)
-			goto out;
-
-		if (found_key.offset < last_offset) {
-			WARN_ON(1);
-			btrfs_print_leaf(root, leaf);
-			printk(KERN_ERR "inode %lu found offset %llu "
-			       "expected %llu\n", inode->i_ino,
-			       (unsigned long long)found_key.offset,
-			       (unsigned long long)last_offset);
-			err = 1;
-			goto out;
-		}
-		extent = btrfs_item_ptr(leaf, slot,
-					struct btrfs_file_extent_item);
-		found_type = btrfs_file_extent_type(leaf, extent);
-		if (found_type == BTRFS_FILE_EXTENT_REG) {
-			extent_end = found_key.offset +
-			     btrfs_file_extent_num_bytes(leaf, extent);
-		} else if (found_type == BTRFS_FILE_EXTENT_INLINE) {
-			struct btrfs_item *item;
-			item = btrfs_item_nr(leaf, slot);
-			extent_end = found_key.offset +
-			     btrfs_file_extent_inline_len(leaf, extent);
-			extent_end = (extent_end + root->sectorsize - 1) &
-				~((u64)root->sectorsize - 1);
-		}
-		last_offset = extent_end;
-		path->slots[0]++;
-	}
-	if (0 && last_offset < inode->i_size) {
-		WARN_ON(1);
-		btrfs_print_leaf(root, leaf);
-		printk(KERN_ERR "inode %lu found offset %llu size %llu\n",
-		       inode->i_ino, (unsigned long long)last_offset,
-		       (unsigned long long)inode->i_size);
-		err = 1;
-
-	}
-out:
-	btrfs_free_path(path);
-	return err;
-#endif
-}
-
 /*
  * this is very complex, but the basic idea is to drop all extents
  * in the range start - end.  hint_block is filled in with a block number
@@ -689,7 +612,6 @@ out:
 		unlock_extent(&BTRFS_I(inode)->io_tree, orig_locked_end,
 			      locked_end - 1, GFP_NOFS);
 	}
-	btrfs_check_file(root, inode);
 	return ret;
 }
 
-- 
cgit v1.2.3


From 193f284d4985db0370a8a1bbdfb20df548cf9ffb Mon Sep 17 00:00:00 2001
From: Chris Mason <chris.mason@oracle.com>
Date: Mon, 27 Apr 2009 07:29:05 -0400
Subject: Btrfs: ratelimit IO error printks

Btrfs has printks for various IO errors, including bad checksums and
mismatches between what we expect the block headers to contain and what
we actually find on the disk.

Longer term we need a real reporting mechanism for this, but for now
printk is going to have to do.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/disk-io.c | 38 +++++++++++++++++++++++++-------------
 fs/btrfs/inode.c   | 10 ++++++----
 2 files changed, 31 insertions(+), 17 deletions(-)

diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 77f9a3b824b..aa0c259b9c2 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -232,10 +232,14 @@ static int csum_tree_block(struct btrfs_root *root, struct extent_buffer *buf,
 			memcpy(&found, result, csum_size);
 
 			read_extent_buffer(buf, &val, 0, csum_size);
-			printk(KERN_INFO "btrfs: %s checksum verify failed "
-			       "on %llu wanted %X found %X level %d\n",
-			       root->fs_info->sb->s_id,
-			       buf->start, val, found, btrfs_header_level(buf));
+			if (printk_ratelimit()) {
+				printk(KERN_INFO "btrfs: %s checksum verify "
+				       "failed on %llu wanted %X found %X "
+				       "level %d\n",
+				       root->fs_info->sb->s_id,
+				       (unsigned long long)buf->start, val, found,
+				       btrfs_header_level(buf));
+			}
 			if (result != (char *)&inline_result)
 				kfree(result);
 			return 1;
@@ -268,10 +272,13 @@ static int verify_parent_transid(struct extent_io_tree *io_tree,
 		ret = 0;
 		goto out;
 	}
-	printk("parent transid verify failed on %llu wanted %llu found %llu\n",
-	       (unsigned long long)eb->start,
-	       (unsigned long long)parent_transid,
-	       (unsigned long long)btrfs_header_generation(eb));
+	if (printk_ratelimit()) {
+		printk("parent transid verify failed on %llu wanted %llu "
+		       "found %llu\n",
+		       (unsigned long long)eb->start,
+		       (unsigned long long)parent_transid,
+		       (unsigned long long)btrfs_header_generation(eb));
+	}
 	ret = 1;
 	clear_extent_buffer_uptodate(io_tree, eb);
 out:
@@ -415,9 +422,12 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 
 	found_start = btrfs_header_bytenr(eb);
 	if (found_start != start) {
-		printk(KERN_INFO "btrfs bad tree block start %llu %llu\n",
-		       (unsigned long long)found_start,
-		       (unsigned long long)eb->start);
+		if (printk_ratelimit()) {
+			printk(KERN_INFO "btrfs bad tree block start "
+			       "%llu %llu\n",
+			       (unsigned long long)found_start,
+			       (unsigned long long)eb->start);
+		}
 		ret = -EIO;
 		goto err;
 	}
@@ -429,8 +439,10 @@ static int btree_readpage_end_io_hook(struct page *page, u64 start, u64 end,
 		goto err;
 	}
 	if (check_tree_block_fsid(root, eb)) {
-		printk(KERN_INFO "btrfs bad fsid on block %llu\n",
-		       (unsigned long long)eb->start);
+		if (printk_ratelimit()) {
+			printk(KERN_INFO "btrfs bad fsid on block %llu\n",
+			       (unsigned long long)eb->start);
+		}
 		ret = -EIO;
 		goto err;
 	}
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 2fdb2995be6..552e08afc7f 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -1823,10 +1823,12 @@ good:
 	return 0;
 
 zeroit:
-	printk(KERN_INFO "btrfs csum failed ino %lu off %llu csum %u "
-	       "private %llu\n", page->mapping->host->i_ino,
-	       (unsigned long long)start, csum,
-	       (unsigned long long)private);
+	if (printk_ratelimit()) {
+		printk(KERN_INFO "btrfs csum failed ino %lu off %llu csum %u "
+		       "private %llu\n", page->mapping->host->i_ino,
+		       (unsigned long long)start, csum,
+		       (unsigned long long)private);
+	}
 	memset(kaddr + offset, 1, end - start + 1);
 	flush_dcache_page(page);
 	kunmap_atomic(kaddr, KM_USER0);
-- 
cgit v1.2.3


From 45c06543afe2772c02f21efee0e2138b4e1c911e Mon Sep 17 00:00:00 2001
From: Chris Mason <chris.mason@oracle.com>
Date: Mon, 27 Apr 2009 07:49:10 -0400
Subject: Btrfs: remove unused btrfs_bit_radix slab

Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/inode.c | 8 --------
 1 file changed, 8 deletions(-)

diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 552e08afc7f..98bd5069d54 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -70,7 +70,6 @@ static struct extent_io_ops btrfs_extent_io_ops;
 static struct kmem_cache *btrfs_inode_cachep;
 struct kmem_cache *btrfs_trans_handle_cachep;
 struct kmem_cache *btrfs_transaction_cachep;
-struct kmem_cache *btrfs_bit_radix_cachep;
 struct kmem_cache *btrfs_path_cachep;
 
 #define S_SHIFT 12
@@ -4641,8 +4640,6 @@ void btrfs_destroy_cachep(void)
 		kmem_cache_destroy(btrfs_trans_handle_cachep);
 	if (btrfs_transaction_cachep)
 		kmem_cache_destroy(btrfs_transaction_cachep);
-	if (btrfs_bit_radix_cachep)
-		kmem_cache_destroy(btrfs_bit_radix_cachep);
 	if (btrfs_path_cachep)
 		kmem_cache_destroy(btrfs_path_cachep);
 }
@@ -4673,11 +4670,6 @@ int btrfs_init_cachep(void)
 	if (!btrfs_path_cachep)
 		goto fail;
 
-	btrfs_bit_radix_cachep = kmem_cache_create("btrfs_radix", 256, 0,
-			SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD |
-			SLAB_DESTROY_BY_RCU, NULL);
-	if (!btrfs_bit_radix_cachep)
-		goto fail;
 	return 0;
 fail:
 	btrfs_destroy_cachep();
-- 
cgit v1.2.3


From e63b6a6c0ffa2ebd8617cc1a10969000296831aa Mon Sep 17 00:00:00 2001
From: Joel Becker <joel.becker@oracle.com>
Date: Tue, 21 Apr 2009 12:38:30 -0700
Subject: Btrfs: Fix a trivial warning using max() of u64 vs ULL.

A small warning popped up on ia64 because inode-map.c was comparing a
u64 object id with the ULL FIRST_FREE_OBJECTID.  My first thought was
that all the OBJECTID constants should contain the u64 cast because
btrfs code deals entirely in u64s.  But then I saw how large that was,
and figured I'd just fix the max() call.

Signed-off-by: Joel Becker <joel.becker@oracle.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/inode-map.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/btrfs/inode-map.c b/fs/btrfs/inode-map.c
index cc7334d833c..9abbced1123 100644
--- a/fs/btrfs/inode-map.c
+++ b/fs/btrfs/inode-map.c
@@ -79,7 +79,7 @@ int btrfs_find_free_objectid(struct btrfs_trans_handle *trans,
 	}
 	path = btrfs_alloc_path();
 	BUG_ON(!path);
-	search_start = max(search_start, BTRFS_FIRST_FREE_OBJECTID);
+	search_start = max(search_start, (u64)BTRFS_FIRST_FREE_OBJECTID);
 	search_key.objectid = search_start;
 	search_key.type = 0;
 	search_key.offset = 0;
-- 
cgit v1.2.3


From 21380931eb4da4e29ac663d0221581282cbba208 Mon Sep 17 00:00:00 2001
From: Joel Becker <joel.becker@oracle.com>
Date: Tue, 21 Apr 2009 12:38:29 -0700
Subject: Btrfs: Fix a bunch of printk() warnings.

Just happened to notice a bunch of %llu vs u64 warnings.  Here's a patch
to cast them all.

Signed-off-by: Joel Becker <joel.becker@oracle.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/disk-io.c          |  6 +++---
 fs/btrfs/extent-tree.c      | 21 ++++++++++++++-------
 fs/btrfs/free-space-cache.c | 15 ++++++++++-----
 fs/btrfs/ioctl.c            |  6 ++++--
 fs/btrfs/super.c            | 15 +++++++++------
 5 files changed, 40 insertions(+), 23 deletions(-)

diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index aa0c259b9c2..0ff16d3331d 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -1671,7 +1671,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	if (features) {
 		printk(KERN_ERR "BTRFS: couldn't mount because of "
 		       "unsupported optional features (%Lx).\n",
-		       features);
+		       (unsigned long long)features);
 		err = -EINVAL;
 		goto fail_iput;
 	}
@@ -1681,7 +1681,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
 	if (!(sb->s_flags & MS_RDONLY) && features) {
 		printk(KERN_ERR "BTRFS: couldn't mount RDWR because of "
 		       "unsupported option features (%Lx).\n",
-		       features);
+		       (unsigned long long)features);
 		err = -EINVAL;
 		goto fail_iput;
 	}
@@ -2273,7 +2273,7 @@ int close_ctree(struct btrfs_root *root)
 
 	if (fs_info->delalloc_bytes) {
 		printk(KERN_INFO "btrfs: at unmount delalloc count %llu\n",
-		       fs_info->delalloc_bytes);
+		       (unsigned long long)fs_info->delalloc_bytes);
 	}
 	if (fs_info->total_ref_cache_size) {
 		printk(KERN_INFO "btrfs: at umount reference cache size %llu\n",
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 2895a837323..e4966444811 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -1844,10 +1844,14 @@ again:
 		printk(KERN_ERR "no space left, need %llu, %llu delalloc bytes"
 		       ", %llu bytes_used, %llu bytes_reserved, "
 		       "%llu bytes_pinned, %llu bytes_readonly, %llu may use"
-		       "%llu total\n", bytes, data_sinfo->bytes_delalloc,
-		       data_sinfo->bytes_used, data_sinfo->bytes_reserved,
-		       data_sinfo->bytes_pinned, data_sinfo->bytes_readonly,
-		       data_sinfo->bytes_may_use, data_sinfo->total_bytes);
+		       "%llu total\n", (unsigned long long)bytes,
+		       (unsigned long long)data_sinfo->bytes_delalloc,
+		       (unsigned long long)data_sinfo->bytes_used,
+		       (unsigned long long)data_sinfo->bytes_reserved,
+		       (unsigned long long)data_sinfo->bytes_pinned,
+		       (unsigned long long)data_sinfo->bytes_readonly,
+		       (unsigned long long)data_sinfo->bytes_may_use,
+		       (unsigned long long)data_sinfo->total_bytes);
 		return -ENOSPC;
 	}
 	data_sinfo->bytes_may_use += bytes;
@@ -2824,9 +2828,12 @@ static void dump_space_info(struct btrfs_space_info *info, u64 bytes)
 				    info->bytes_pinned - info->bytes_reserved),
 	       (info->full) ? "" : "not ");
 	printk(KERN_INFO "space_info total=%llu, pinned=%llu, delalloc=%llu,"
-	       " may_use=%llu, used=%llu\n", info->total_bytes,
-	       info->bytes_pinned, info->bytes_delalloc, info->bytes_may_use,
-	       info->bytes_used);
+	       " may_use=%llu, used=%llu\n",
+	       (unsigned long long)info->total_bytes,
+	       (unsigned long long)info->bytes_pinned,
+	       (unsigned long long)info->bytes_delalloc,
+	       (unsigned long long)info->bytes_may_use,
+	       (unsigned long long)info->bytes_used);
 
 	down_read(&info->groups_sem);
 	list_for_each_entry(cache, &info->block_groups, list) {
diff --git a/fs/btrfs/free-space-cache.c b/fs/btrfs/free-space-cache.c
index 768b9523662..0bc93657b46 100644
--- a/fs/btrfs/free-space-cache.c
+++ b/fs/btrfs/free-space-cache.c
@@ -332,13 +332,17 @@ int btrfs_remove_free_space(struct btrfs_block_group_cache *block_group,
 			printk(KERN_ERR "couldn't find space %llu to free\n",
 			       (unsigned long long)offset);
 			printk(KERN_ERR "cached is %d, offset %llu bytes %llu\n",
-			       block_group->cached, block_group->key.objectid,
-			       block_group->key.offset);
+			       block_group->cached,
+			       (unsigned long long)block_group->key.objectid,
+			       (unsigned long long)block_group->key.offset);
 			btrfs_dump_free_space(block_group, bytes);
 		} else if (info) {
 			printk(KERN_ERR "hmm, found offset=%llu bytes=%llu, "
 			       "but wanted offset=%llu bytes=%llu\n",
-			       info->offset, info->bytes, offset, bytes);
+			       (unsigned long long)info->offset,
+			       (unsigned long long)info->bytes,
+			       (unsigned long long)offset,
+			       (unsigned long long)bytes);
 		}
 		WARN_ON(1);
 	}
@@ -357,8 +361,9 @@ void btrfs_dump_free_space(struct btrfs_block_group_cache *block_group,
 		info = rb_entry(n, struct btrfs_free_space, offset_index);
 		if (info->bytes >= bytes)
 			count++;
-		printk(KERN_ERR "entry offset %llu, bytes %llu\n", info->offset,
-		       info->bytes);
+		printk(KERN_ERR "entry offset %llu, bytes %llu\n",
+		       (unsigned long long)info->offset,
+		       (unsigned long long)info->bytes);
 	}
 	printk(KERN_INFO "%d blocks of free space at or bigger than bytes is"
 	       "\n", count);
diff --git a/fs/btrfs/ioctl.c b/fs/btrfs/ioctl.c
index f4e5d2e5ece..48762aa1e94 100644
--- a/fs/btrfs/ioctl.c
+++ b/fs/btrfs/ioctl.c
@@ -483,11 +483,13 @@ static int btrfs_ioctl_resize(struct btrfs_root *root, void __user *arg)
 		*devstr = '\0';
 		devstr = vol_args->name;
 		devid = simple_strtoull(devstr, &end, 10);
-		printk(KERN_INFO "resizing devid %llu\n", devid);
+		printk(KERN_INFO "resizing devid %llu\n",
+		       (unsigned long long)devid);
 	}
 	device = btrfs_find_device(root, devid, NULL, NULL);
 	if (!device) {
-		printk(KERN_INFO "resizer unable to find device %llu\n", devid);
+		printk(KERN_INFO "resizer unable to find device %llu\n",
+		       (unsigned long long)devid);
 		ret = -EINVAL;
 		goto out_unlock;
 	}
diff --git a/fs/btrfs/super.c b/fs/btrfs/super.c
index 30c9a8ca2a5..bf0e84c7560 100644
--- a/fs/btrfs/super.c
+++ b/fs/btrfs/super.c
@@ -196,7 +196,7 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
 				info->max_extent = max_t(u64,
 					info->max_extent, root->sectorsize);
 				printk(KERN_INFO "btrfs: max_extent at %llu\n",
-				       info->max_extent);
+				       (unsigned long long)info->max_extent);
 			}
 			break;
 		case Opt_max_inline:
@@ -211,7 +211,7 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
 						root->sectorsize);
 				}
 				printk(KERN_INFO "btrfs: max_inline at %llu\n",
-					info->max_inline);
+					(unsigned long long)info->max_inline);
 			}
 			break;
 		case Opt_alloc_start:
@@ -221,7 +221,7 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
 				kfree(num);
 				printk(KERN_INFO
 					"btrfs: allocations start at %llu\n",
-					info->alloc_start);
+					(unsigned long long)info->alloc_start);
 			}
 			break;
 		case Opt_noacl:
@@ -420,11 +420,14 @@ static int btrfs_show_options(struct seq_file *seq, struct vfsmount *vfs)
 	if (btrfs_test_opt(root, NOBARRIER))
 		seq_puts(seq, ",nobarrier");
 	if (info->max_extent != (u64)-1)
-		seq_printf(seq, ",max_extent=%llu", info->max_extent);
+		seq_printf(seq, ",max_extent=%llu",
+			   (unsigned long long)info->max_extent);
 	if (info->max_inline != 8192 * 1024)
-		seq_printf(seq, ",max_inline=%llu", info->max_inline);
+		seq_printf(seq, ",max_inline=%llu",
+			   (unsigned long long)info->max_inline);
 	if (info->alloc_start != 0)
-		seq_printf(seq, ",alloc_start=%llu", info->alloc_start);
+		seq_printf(seq, ",alloc_start=%llu",
+			   (unsigned long long)info->alloc_start);
 	if (info->thread_pool_size !=  min_t(unsigned long,
 					     num_online_cpus() + 2, 8))
 		seq_printf(seq, ",thread_pool=%d", info->thread_pool_size);
-- 
cgit v1.2.3


From 7b1a14bbb0e547aaa4d30cc376e6c8c12539ab0f Mon Sep 17 00:00:00 2001
From: Chris Mason <chris.mason@oracle.com>
Date: Mon, 27 Apr 2009 10:49:53 -0400
Subject: Btrfs: fix acl caching

Linus noticed the btrfs code to cache acls wasn't properly caching
a NULL acl when the inode didn't have any acls.  This meant the common
case of no acls resulted in expensive btree searches every time the
kernel checked permissions (which is quite often).

This is a modified version of Linus' original patch:

Properly set initial acl fields to BTRFS_ACL_NOT_CACHED in the inode.
This forces an acl lookup when permission checks are done.

Fix btrfs_get_acl to avoid lookups and locking when the inode acls fields
are set to null.

Fix btrfs_get_acl to use the right return value from __btrfs_getxattr
when deciding to cache a NULL acl.  It was storing a NULL acl when
__btrfs_getxattr return -ENOENT, but __btrfs_getxattr was actually returning
-ENODATA for this case.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/acl.c   | 18 +++++++++++++-----
 fs/btrfs/inode.c |  4 ++--
 2 files changed, 15 insertions(+), 7 deletions(-)

diff --git a/fs/btrfs/acl.c b/fs/btrfs/acl.c
index 7fdd184a528..cbba000dccb 100644
--- a/fs/btrfs/acl.c
+++ b/fs/btrfs/acl.c
@@ -60,15 +60,20 @@ static struct posix_acl *btrfs_get_acl(struct inode *inode, int type)
 		return ERR_PTR(-EINVAL);
 	}
 
+	/* Handle the cached NULL acl case without locking */
+	acl = ACCESS_ONCE(*p_acl);
+	if (!acl)
+		return acl;
+
 	spin_lock(&inode->i_lock);
-	if (*p_acl != BTRFS_ACL_NOT_CACHED)
-		acl = posix_acl_dup(*p_acl);
+	acl = *p_acl;
+	if (acl != BTRFS_ACL_NOT_CACHED)
+		acl = posix_acl_dup(acl);
 	spin_unlock(&inode->i_lock);
 
-	if (acl)
+	if (acl != BTRFS_ACL_NOT_CACHED)
 		return acl;
 
-
 	size = __btrfs_getxattr(inode, name, "", 0);
 	if (size > 0) {
 		value = kzalloc(size, GFP_NOFS);
@@ -80,9 +85,12 @@ static struct posix_acl *btrfs_get_acl(struct inode *inode, int type)
 			btrfs_update_cached_acl(inode, p_acl, acl);
 		}
 		kfree(value);
-	} else if (size == -ENOENT) {
+	} else if (size == -ENOENT || size == -ENODATA || size == 0) {
+		/* FIXME, who returns -ENOENT?  I think nobody */
 		acl = NULL;
 		btrfs_update_cached_acl(inode, p_acl, acl);
+	} else {
+		acl = ERR_PTR(-EIO);
 	}
 
 	return acl;
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 98bd5069d54..dc812ec551f 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -3047,8 +3047,8 @@ static noinline void init_btrfs_i(struct inode *inode)
 {
 	struct btrfs_inode *bi = BTRFS_I(inode);
 
-	bi->i_acl = NULL;
-	bi->i_default_acl = NULL;
+	bi->i_acl = BTRFS_ACL_NOT_CACHED;
+	bi->i_default_acl = BTRFS_ACL_NOT_CACHED;
 
 	bi->generation = 0;
 	bi->sequence = 0;
-- 
cgit v1.2.3


From 46a53cca826e71effe59e3cb4f383622c33ebdcb Mon Sep 17 00:00:00 2001
From: Chris Mason <chris.mason@oracle.com>
Date: Mon, 27 Apr 2009 11:47:50 -0400
Subject: Btrfs: look for acls during btrfs_read_locked_inode

This changes btrfs_read_locked_inode() to peek ahead in the btree for acl items.
If it is certain a given inode has no acls, it will set the in memory acl
fields to null to avoid acl lookups completely.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/inode.c | 62 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 62 insertions(+)

diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index dc812ec551f..90c23eb2882 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -2015,6 +2015,57 @@ void btrfs_orphan_cleanup(struct btrfs_root *root)
 	btrfs_free_path(path);
 }
 
+/*
+ * very simple check to peek ahead in the leaf looking for xattrs.  If we
+ * don't find any xattrs, we know there can't be any acls.
+ *
+ * slot is the slot the inode is in, objectid is the objectid of the inode
+ */
+static noinline int acls_after_inode_item(struct extent_buffer *leaf,
+					  int slot, u64 objectid)
+{
+	u32 nritems = btrfs_header_nritems(leaf);
+	struct btrfs_key found_key;
+	int scanned = 0;
+
+	slot++;
+	while (slot < nritems) {
+		btrfs_item_key_to_cpu(leaf, &found_key, slot);
+
+		/* we found a different objectid, there must not be acls */
+		if (found_key.objectid != objectid)
+			return 0;
+
+		/* we found an xattr, assume we've got an acl */
+		if (found_key.type == BTRFS_XATTR_ITEM_KEY)
+			return 1;
+
+		/*
+		 * we found a key greater than an xattr key, there can't
+		 * be any acls later on
+		 */
+		if (found_key.type > BTRFS_XATTR_ITEM_KEY)
+			return 0;
+
+		slot++;
+		scanned++;
+
+		/*
+		 * it goes inode, inode backrefs, xattrs, extents,
+		 * so if there are a ton of hard links to an inode there can
+		 * be a lot of backrefs.  Don't waste time searching too hard,
+		 * this is just an optimization
+		 */
+		if (scanned >= 8)
+			break;
+	}
+	/* we hit the end of the leaf before we found an xattr or
+	 * something larger than an xattr.  We have to assume the inode
+	 * has acls
+	 */
+	return 1;
+}
+
 /*
  * read an inode from the btree into the in-memory inode
  */
@@ -2026,6 +2077,7 @@ void btrfs_read_locked_inode(struct inode *inode)
 	struct btrfs_timespec *tspec;
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_key location;
+	int maybe_acls;
 	u64 alloc_group_block;
 	u32 rdev;
 	int ret;
@@ -2072,6 +2124,16 @@ void btrfs_read_locked_inode(struct inode *inode)
 
 	alloc_group_block = btrfs_inode_block_group(leaf, inode_item);
 
+	/*
+	 * try to precache a NULL acl entry for files that don't have
+	 * any xattrs or acls
+	 */
+	maybe_acls = acls_after_inode_item(leaf, path->slots[0], inode->i_ino);
+	if (!maybe_acls) {
+		BTRFS_I(inode)->i_acl = NULL;
+		BTRFS_I(inode)->i_default_acl = NULL;
+	}
+
 	BTRFS_I(inode)->block_group = btrfs_find_block_group(root, 0,
 						alloc_group_block, 0);
 	btrfs_free_path(path);
-- 
cgit v1.2.3