From 8bb8ab2e93f9c3c9453e13be0f37d344a32a3a6d Mon Sep 17 00:00:00 2001
From: Josef Bacik <josef@redhat.com>
Date: Fri, 15 Oct 2010 16:52:49 -0400
Subject: Btrfs: rework how we reserve metadata bytes

With multi-threaded writes we were getting ENOSPC early because somebody would
come in, start flushing delalloc because they couldn't make their reservation,
and in the meantime other threads would come in and use the space that was
getting freed up, so when the original thread went to check to see if they had
space they didn't and they'd return ENOSPC.  So instead if we have some free
space but not enough for our reservation, take the reservation and then start
doing the flushing.  The only time we don't take reservations is when we've
already overcommitted our space, that way we don't have people who come late to
the party way overcommitting ourselves.  This also moves all of the retrying and
flushing code into reserve_metdata_bytes so it's all uniform.  This keeps my
fs_mark test from returning -ENOSPC as soon as it starts and actually lets me
fill up the disk.  Thanks,

Signed-off-by: Josef Bacik <josef@redhat.com>
---
 fs/btrfs/relocation.c | 14 +++-----------
 1 file changed, 3 insertions(+), 11 deletions(-)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index b37d723b9d4a..39adb68a653f 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -178,8 +178,6 @@ struct reloc_control {
 	u64 search_start;
 	u64 extents_found;
 
-	int block_rsv_retries;
-
 	unsigned int stage:8;
 	unsigned int create_reloc_tree:1;
 	unsigned int merge_reloc_tree:1;
@@ -2133,7 +2131,6 @@ int prepare_to_merge(struct reloc_control *rc, int err)
 	LIST_HEAD(reloc_roots);
 	u64 num_bytes = 0;
 	int ret;
-	int retries = 0;
 
 	mutex_lock(&root->fs_info->trans_mutex);
 	rc->merging_rsv_size += root->nodesize * (BTRFS_MAX_LEVEL - 1) * 2;
@@ -2143,7 +2140,7 @@ again:
 	if (!err) {
 		num_bytes = rc->merging_rsv_size;
 		ret = btrfs_block_rsv_add(NULL, root, rc->block_rsv,
-					  num_bytes, &retries);
+					  num_bytes);
 		if (ret)
 			err = ret;
 	}
@@ -2155,7 +2152,6 @@ again:
 			btrfs_end_transaction(trans, rc->extent_root);
 			btrfs_block_rsv_release(rc->extent_root,
 						rc->block_rsv, num_bytes);
-			retries = 0;
 			goto again;
 		}
 	}
@@ -2405,15 +2401,13 @@ static int reserve_metadata_space(struct btrfs_trans_handle *trans,
 	num_bytes = calcu_metadata_size(rc, node, 1) * 2;
 
 	trans->block_rsv = rc->block_rsv;
-	ret = btrfs_block_rsv_add(trans, root, rc->block_rsv, num_bytes,
-				  &rc->block_rsv_retries);
+	ret = btrfs_block_rsv_add(trans, root, rc->block_rsv, num_bytes);
 	if (ret) {
 		if (ret == -EAGAIN)
 			rc->commit_transaction = 1;
 		return ret;
 	}
 
-	rc->block_rsv_retries = 0;
 	return 0;
 }
 
@@ -3554,8 +3548,7 @@ int prepare_to_relocate(struct reloc_control *rc)
 	 * is no reservation in transaction handle.
 	 */
 	ret = btrfs_block_rsv_add(NULL, rc->extent_root, rc->block_rsv,
-				  rc->extent_root->nodesize * 256,
-				  &rc->block_rsv_retries);
+				  rc->extent_root->nodesize * 256);
 	if (ret)
 		return ret;
 
@@ -3567,7 +3560,6 @@ int prepare_to_relocate(struct reloc_control *rc)
 	rc->extents_found = 0;
 	rc->nodes_relocated = 0;
 	rc->merging_rsv_size = 0;
-	rc->block_rsv_retries = 0;
 
 	rc->create_reloc_tree = 1;
 	set_reloc_control(rc);
-- 
cgit v1.2.2


From 0af3d00bad38d3bb9912a60928ad0669f17bdb76 Mon Sep 17 00:00:00 2001
From: Josef Bacik <josef@redhat.com>
Date: Mon, 21 Jun 2010 14:48:16 -0400
Subject: Btrfs: create special free space cache inode

In order to save free space cache, we need an inode to hold the data, and we
need a special item to point at the right inode for the right block group.  So
first, create a special item that will point to the right inode, and the number
of extent entries we will have and the number of bitmaps we will have.  We
truncate and pre-allocate space everytime to make sure it's uptodate.

This feature will be turned on as soon as you mount with -o space_cache, however
it is safe to boot into old kernels, they will just generate the cache the old
fashion way.  When you boot back into a newer kernel we will notice that we
modified and not the cache and automatically discard the cache.

Signed-off-by: Josef Bacik <josef@redhat.com>
---
 fs/btrfs/relocation.c | 91 ++++++++++++++++++++++++++++++++++++++++++++++++---
 1 file changed, 87 insertions(+), 4 deletions(-)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index b37d723b9d4a..af339eee55b8 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -29,6 +29,7 @@
 #include "locking.h"
 #include "btrfs_inode.h"
 #include "async-thread.h"
+#include "free-space-cache.h"
 
 /*
  * backref_node, mapping_node and tree_block start with this
@@ -3191,6 +3192,54 @@ static int block_use_full_backref(struct reloc_control *rc,
 	return ret;
 }
 
+static int delete_block_group_cache(struct btrfs_fs_info *fs_info,
+				    struct inode *inode, u64 ino)
+{
+	struct btrfs_key key;
+	struct btrfs_path *path;
+	struct btrfs_root *root = fs_info->tree_root;
+	struct btrfs_trans_handle *trans;
+	unsigned long nr;
+	int ret = 0;
+
+	if (inode)
+		goto truncate;
+
+	key.objectid = ino;
+	key.type = BTRFS_INODE_ITEM_KEY;
+	key.offset = 0;
+
+	inode = btrfs_iget(fs_info->sb, &key, root, NULL);
+	if (!inode || IS_ERR(inode) || is_bad_inode(inode)) {
+		if (inode && !IS_ERR(inode))
+			iput(inode);
+		return -ENOENT;
+	}
+
+truncate:
+	path = btrfs_alloc_path();
+	if (!path) {
+		ret = -ENOMEM;
+		goto out;
+	}
+
+	trans = btrfs_join_transaction(root, 0);
+	if (IS_ERR(trans)) {
+		btrfs_free_path(path);
+		goto out;
+	}
+
+	ret = btrfs_truncate_free_space_cache(root, trans, path, inode);
+
+	btrfs_free_path(path);
+	nr = trans->blocks_used;
+	btrfs_end_transaction(trans, root);
+	btrfs_btree_balance_dirty(root, nr);
+out:
+	iput(inode);
+	return ret;
+}
+
 /*
  * helper to add tree blocks for backref of type BTRFS_EXTENT_DATA_REF_KEY
  * this function scans fs tree to find blocks reference the data extent
@@ -3217,15 +3266,27 @@ static int find_data_references(struct reloc_control *rc,
 	int counted;
 	int ret;
 
-	path = btrfs_alloc_path();
-	if (!path)
-		return -ENOMEM;
-
 	ref_root = btrfs_extent_data_ref_root(leaf, ref);
 	ref_objectid = btrfs_extent_data_ref_objectid(leaf, ref);
 	ref_offset = btrfs_extent_data_ref_offset(leaf, ref);
 	ref_count = btrfs_extent_data_ref_count(leaf, ref);
 
+	/*
+	 * This is an extent belonging to the free space cache, lets just delete
+	 * it and redo the search.
+	 */
+	if (ref_root == BTRFS_ROOT_TREE_OBJECTID) {
+		ret = delete_block_group_cache(rc->extent_root->fs_info,
+					       NULL, ref_objectid);
+		if (ret != -ENOENT)
+			return ret;
+		ret = 0;
+	}
+
+	path = btrfs_alloc_path();
+	if (!path)
+		return -ENOMEM;
+
 	root = read_fs_root(rc->extent_root->fs_info, ref_root);
 	if (IS_ERR(root)) {
 		err = PTR_ERR(root);
@@ -3860,6 +3921,8 @@ int btrfs_relocate_block_group(struct btrfs_root *extent_root, u64 group_start)
 {
 	struct btrfs_fs_info *fs_info = extent_root->fs_info;
 	struct reloc_control *rc;
+	struct inode *inode;
+	struct btrfs_path *path;
 	int ret;
 	int rw = 0;
 	int err = 0;
@@ -3882,6 +3945,26 @@ int btrfs_relocate_block_group(struct btrfs_root *extent_root, u64 group_start)
 		rw = 1;
 	}
 
+	path = btrfs_alloc_path();
+	if (!path) {
+		err = -ENOMEM;
+		goto out;
+	}
+
+	inode = lookup_free_space_inode(fs_info->tree_root, rc->block_group,
+					path);
+	btrfs_free_path(path);
+
+	if (!IS_ERR(inode))
+		ret = delete_block_group_cache(fs_info, inode, 0);
+	else
+		ret = PTR_ERR(inode);
+
+	if (ret && ret != -ENOENT) {
+		err = ret;
+		goto out;
+	}
+
 	rc->data_inode = create_reloc_inode(fs_info, rc->block_group);
 	if (IS_ERR(rc->data_inode)) {
 		err = PTR_ERR(rc->data_inode);
-- 
cgit v1.2.2


From 411fc6bcef54f828a5458f4730c68abdf13c6bf0 Mon Sep 17 00:00:00 2001
From: Andi Kleen <andi@firstfloor.org>
Date: Fri, 29 Oct 2010 15:14:31 -0400
Subject: Btrfs: Fix variables set but not read (bugs found by gcc 4.6)

These are all the cases where a variable is set, but not
read which are really bugs.

- Couple of incorrect error handling fixed.
- One incorrect use of a allocation policy
- Some other things

Still needs more review.

Found by gcc 4.6's new warnings.

[akpm@linux-foundation.org: fix build.  Might have been bitrot]
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Cc: Chris Mason <chris.mason@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/relocation.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index fd0714475db7..045c9c2b2d7e 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -3094,6 +3094,8 @@ static int add_tree_block(struct reloc_control *rc,
 		BUG_ON(item_size != sizeof(struct btrfs_extent_item_v0));
 		ret = get_ref_objectid_v0(rc, path, extent_key,
 					  &ref_owner, NULL);
+		if (ret < 0)
+			return ret;
 		BUG_ON(ref_owner >= BTRFS_MAX_LEVEL);
 		level = (int)ref_owner;
 		/* FIXME: get real generation */
@@ -4218,7 +4220,7 @@ int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len)
 		btrfs_add_ordered_sum(inode, ordered, sums);
 	}
 	btrfs_put_ordered_extent(ordered);
-	return 0;
+	return ret;
 }
 
 void btrfs_reloc_cow_block(struct btrfs_trans_handle *trans,
-- 
cgit v1.2.2


From 3612b49598c303cfb22a4b609427f829828e2427 Mon Sep 17 00:00:00 2001
From: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Date: Tue, 25 Jan 2011 02:51:38 +0000
Subject: btrfs: fix return value check of btrfs_join_transaction()

The error check of btrfs_join_transaction()/btrfs_join_transaction_nolock()
is added, and the mistake of the error check in several places is
corrected.

For more stable Btrfs, I think that we should reduce BUG_ON().
But, I think that long time is necessary for this.
So, I propose this patch as a short-term solution.

With this patch:
 - To more stable Btrfs, the part that should be corrected is clarified.
 - The panic isn't done by the NULL pointer reference etc. (even if
   BUG_ON() is increased temporarily)
 - The error code is returned in the place where the error can be easily
   returned.

As a long-term plan:
 - BUG_ON() is reduced by using the forced-readonly framework, etc.

Signed-off-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/relocation.c | 26 +++++++++++++++++++++++---
 1 file changed, 23 insertions(+), 3 deletions(-)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 045c9c2b2d7e..ea9965430241 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -2147,6 +2147,12 @@ again:
 	}
 
 	trans = btrfs_join_transaction(rc->extent_root, 1);
+	if (IS_ERR(trans)) {
+		if (!err)
+			btrfs_block_rsv_release(rc->extent_root,
+						rc->block_rsv, num_bytes);
+		return PTR_ERR(trans);
+	}
 
 	if (!err) {
 		if (num_bytes != rc->merging_rsv_size) {
@@ -3222,6 +3228,7 @@ truncate:
 	trans = btrfs_join_transaction(root, 0);
 	if (IS_ERR(trans)) {
 		btrfs_free_path(path);
+		ret = PTR_ERR(trans);
 		goto out;
 	}
 
@@ -3628,6 +3635,7 @@ int prepare_to_relocate(struct reloc_control *rc)
 	set_reloc_control(rc);
 
 	trans = btrfs_join_transaction(rc->extent_root, 1);
+	BUG_ON(IS_ERR(trans));
 	btrfs_commit_transaction(trans, rc->extent_root);
 	return 0;
 }
@@ -3804,7 +3812,10 @@ static noinline_for_stack int relocate_block_group(struct reloc_control *rc)
 
 	/* get rid of pinned extents */
 	trans = btrfs_join_transaction(rc->extent_root, 1);
-	btrfs_commit_transaction(trans, rc->extent_root);
+	if (IS_ERR(trans))
+		err = PTR_ERR(trans);
+	else
+		btrfs_commit_transaction(trans, rc->extent_root);
 out_free:
 	btrfs_free_block_rsv(rc->extent_root, rc->block_rsv);
 	btrfs_free_path(path);
@@ -4125,6 +4136,11 @@ int btrfs_recover_relocation(struct btrfs_root *root)
 	set_reloc_control(rc);
 
 	trans = btrfs_join_transaction(rc->extent_root, 1);
+	if (IS_ERR(trans)) {
+		unset_reloc_control(rc);
+		err = PTR_ERR(trans);
+		goto out_free;
+	}
 
 	rc->merge_reloc_tree = 1;
 
@@ -4154,9 +4170,13 @@ int btrfs_recover_relocation(struct btrfs_root *root)
 	unset_reloc_control(rc);
 
 	trans = btrfs_join_transaction(rc->extent_root, 1);
-	btrfs_commit_transaction(trans, rc->extent_root);
-out:
+	if (IS_ERR(trans))
+		err = PTR_ERR(trans);
+	else
+		btrfs_commit_transaction(trans, rc->extent_root);
+out_free:
 	kfree(rc);
+out:
 	while (!list_empty(&reloc_roots)) {
 		reloc_root = list_entry(reloc_roots.next,
 					struct btrfs_root, root_list);
-- 
cgit v1.2.2


From 98d5dc13e7e74b77ca3b4c3cbded9f48d2dbbbb7 Mon Sep 17 00:00:00 2001
From: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Date: Thu, 20 Jan 2011 06:19:37 +0000
Subject: btrfs: fix return value check of btrfs_start_transaction()

The error check of btrfs_start_transaction() is added, and the mistake
of the error check on several places is corrected.

Signed-off-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/relocation.c | 3 +++
 1 file changed, 3 insertions(+)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index ea9965430241..1f5556acb530 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -2028,6 +2028,7 @@ static noinline_for_stack int merge_reloc_root(struct reloc_control *rc,
 
 	while (1) {
 		trans = btrfs_start_transaction(root, 0);
+		BUG_ON(IS_ERR(trans));
 		trans->block_rsv = rc->block_rsv;
 
 		ret = btrfs_block_rsv_check(trans, root, rc->block_rsv,
@@ -3665,6 +3666,7 @@ static noinline_for_stack int relocate_block_group(struct reloc_control *rc)
 
 	while (1) {
 		trans = btrfs_start_transaction(rc->extent_root, 0);
+		BUG_ON(IS_ERR(trans));
 
 		if (update_backref_cache(trans, &rc->backref_cache)) {
 			btrfs_end_transaction(trans, rc->extent_root);
@@ -4033,6 +4035,7 @@ static noinline_for_stack int mark_garbage_root(struct btrfs_root *root)
 	int ret;
 
 	trans = btrfs_start_transaction(root->fs_info->tree_root, 0);
+	BUG_ON(IS_ERR(trans));
 
 	memset(&root->root_item.drop_progress, 0,
 		sizeof(root->root_item.drop_progress));
-- 
cgit v1.2.2


From 6848ad6461e551849ba3c32d945d4f45e96453a6 Mon Sep 17 00:00:00 2001
From: "Yan, Zheng" <zheng.z.yan@linux.intel.com>
Date: Mon, 14 Feb 2011 16:00:03 -0500
Subject: Btrfs: Fix balance panic

Mark the cloned backref_node as checked in clone_backref_node()

Signed-off-by: Yan, Zheng <zheng.z.yan@intel.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/relocation.c | 1 +
 1 file changed, 1 insertion(+)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 1f5556acb530..0825e4ed9447 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -1157,6 +1157,7 @@ static int clone_backref_node(struct btrfs_trans_handle *trans,
 	new_node->bytenr = dest->node->start;
 	new_node->level = node->level;
 	new_node->lowest = node->lowest;
+	new_node->checked = 1;
 	new_node->root = dest;
 
 	if (!node->lowest) {
-- 
cgit v1.2.2


From c87f08ca44e83b2c8d28f63f9c33f3a270a04bbe Mon Sep 17 00:00:00 2001
From: Chris Mason <chris.mason@oracle.com>
Date: Wed, 16 Feb 2011 13:57:04 -0500
Subject: Btrfs: allow balance to explicitly allocate chunks as it relocates

Btrfs device shrinking and balancing ends up reallocating all the blocks
in order to allow COW to move them to new destinations.  It is somewhat
awkward in terms of ENOSPC because most of the enospc code is built
around the idea that some operation on a reference counted tree triggers
allocations in the non-reference counted trees.

This commit changes the balancing code to deal with enospc by trying to
allocate a new chunk.  If that allocation succeeds, we go ahead and
retry whatever failed due to enospc.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/relocation.c | 13 ++++++++++++-
 1 file changed, 12 insertions(+), 1 deletion(-)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 0825e4ed9447..31ade5802ae8 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -3654,6 +3654,7 @@ static noinline_for_stack int relocate_block_group(struct reloc_control *rc)
 	u32 item_size;
 	int ret;
 	int err = 0;
+	int progress = 0;
 
 	path = btrfs_alloc_path();
 	if (!path)
@@ -3666,9 +3667,10 @@ static noinline_for_stack int relocate_block_group(struct reloc_control *rc)
 	}
 
 	while (1) {
+		progress++;
 		trans = btrfs_start_transaction(rc->extent_root, 0);
 		BUG_ON(IS_ERR(trans));
-
+restart:
 		if (update_backref_cache(trans, &rc->backref_cache)) {
 			btrfs_end_transaction(trans, rc->extent_root);
 			continue;
@@ -3781,6 +3783,15 @@ static noinline_for_stack int relocate_block_group(struct reloc_control *rc)
 			}
 		}
 	}
+	if (trans && progress && err == -ENOSPC) {
+		ret = btrfs_force_chunk_alloc(trans, rc->extent_root,
+					      rc->block_group->flags);
+		if (ret == 0) {
+			err = 0;
+			progress = 0;
+			goto restart;
+		}
+	}
 
 	btrfs_release_path(rc->extent_root, path);
 	clear_extent_bits(&rc->processed_blocks, 0, (u64)-1, EXTENT_DIRTY,
-- 
cgit v1.2.2


From 66b4ffd110f9b48b8d8c1319ee446b53b8d073bf Mon Sep 17 00:00:00 2001
From: Josef Bacik <josef@redhat.com>
Date: Mon, 31 Jan 2011 16:22:42 -0500
Subject: Btrfs: handle errors in btrfs_orphan_cleanup

If we cannot truncate an inode for some reason we will never delete the orphan
item associated with that inode, which means that we will loop forever in
btrfs_orphan_cleanup.  Instead of doing this just return error so we fail to
mount.  It sucks, but hey it's better than hanging.  Thanks,

Signed-off-by: Josef Bacik <josef@redhat.com>
---
 fs/btrfs/relocation.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 31ade5802ae8..c863c8447015 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -4209,7 +4209,7 @@ out:
 		if (IS_ERR(fs_root))
 			err = PTR_ERR(fs_root);
 		else
-			btrfs_orphan_cleanup(fs_root);
+			err = btrfs_orphan_cleanup(fs_root);
 	}
 	return err;
 }
-- 
cgit v1.2.2


From 97d9a8a420444eb5b5c071d4b3b9c4100a7ae015 Mon Sep 17 00:00:00 2001
From: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Date: Thu, 24 Mar 2011 06:33:21 +0000
Subject: Btrfs: check return value of read_tree_block()

This patch is checking return value of read_tree_block(),
and if it is NULL, error processing.

Signed-off-by: Tsutomu Itoh <t-itoh@jp.fujitsu.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/relocation.c | 6 ++++++
 1 file changed, 6 insertions(+)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index c863c8447015..58250e09eb05 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -1724,6 +1724,7 @@ again:
 
 			eb = read_tree_block(dest, old_bytenr, blocksize,
 					     old_ptr_gen);
+			BUG_ON(!eb);
 			btrfs_tree_lock(eb);
 			if (cow) {
 				ret = btrfs_cow_block(trans, dest, eb, parent,
@@ -2513,6 +2514,10 @@ static int do_relocation(struct btrfs_trans_handle *trans,
 		blocksize = btrfs_level_size(root, node->level);
 		generation = btrfs_node_ptr_generation(upper->eb, slot);
 		eb = read_tree_block(root, bytenr, blocksize, generation);
+		if (!eb) {
+			err = -EIO;
+			goto next;
+		}
 		btrfs_tree_lock(eb);
 		btrfs_set_lock_blocking(eb);
 
@@ -2670,6 +2675,7 @@ static int get_tree_block_key(struct reloc_control *rc,
 	BUG_ON(block->key_ready);
 	eb = read_tree_block(rc->extent_root, block->bytenr,
 			     block->key.objectid, block->key.offset);
+	BUG_ON(!eb);
 	WARN_ON(btrfs_header_level(eb) != block->level);
 	if (block->level == 0)
 		btrfs_item_key_to_cpu(eb, &block->key, 0);
-- 
cgit v1.2.2


From 25985edcedea6396277003854657b5f3cb31a628 Mon Sep 17 00:00:00 2001
From: Lucas De Marchi <lucas.demarchi@profusion.mobi>
Date: Wed, 30 Mar 2011 22:57:33 -0300
Subject: Fix common misspellings

Fixes generated by 'codespell' and manually reviewed.

Signed-off-by: Lucas De Marchi <lucas.demarchi@profusion.mobi>
---
 fs/btrfs/relocation.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 58250e09eb05..199a80134312 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -2346,7 +2346,7 @@ struct btrfs_root *select_one_root(struct btrfs_trans_handle *trans,
 		root = next->root;
 		BUG_ON(!root);
 
-		/* no other choice for non-refernce counted tree */
+		/* no other choice for non-references counted tree */
 		if (!root->ref_cows)
 			return root;
 
-- 
cgit v1.2.2


From 581bb050941b4f220f84d3e5ed6dace3d42dd382 Mon Sep 17 00:00:00 2001
From: Li Zefan <lizf@cn.fujitsu.com>
Date: Wed, 20 Apr 2011 10:06:11 +0800
Subject: Btrfs: Cache free inode numbers in memory

Currently btrfs stores the highest objectid of the fs tree, and it always
returns (highest+1) inode number when we create a file, so inode numbers
won't be reclaimed when we delete files, so we'll run out of inode numbers
as we keep create/delete files in 32bits machines.

This fixes it, and it works similarly to how we cache free space in block
cgroups.

We start a kernel thread to read the file tree. By scanning inode items,
we know which chunks of inode numbers are free, and we cache them in
an rb-tree.

Because we are searching the commit root, we have to carefully handle the
cross-transaction case.

The rb-tree is a hybrid extent+bitmap tree, so if we have too many small
chunks of inode numbers, we'll use bitmaps. Initially we allow 16K ram
of extents, and a bitmap will be used if we exceed this threshold. The
extents threshold is adjusted in runtime.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
 fs/btrfs/relocation.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 58250e09eb05..e6cb89357256 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -30,6 +30,7 @@
 #include "btrfs_inode.h"
 #include "async-thread.h"
 #include "free-space-cache.h"
+#include "inode-map.h"
 
 /*
  * backref_node, mapping_node and tree_block start with this
@@ -3897,7 +3898,7 @@ struct inode *create_reloc_inode(struct btrfs_fs_info *fs_info,
 	if (IS_ERR(trans))
 		return ERR_CAST(trans);
 
-	err = btrfs_find_free_objectid(trans, root, objectid, &objectid);
+	err = btrfs_find_free_objectid(root, &objectid);
 	if (err)
 		goto out;
 
-- 
cgit v1.2.2


From 33345d01522f8152f99dc84a3e7a1a45707f387f Mon Sep 17 00:00:00 2001
From: Li Zefan <lizf@cn.fujitsu.com>
Date: Wed, 20 Apr 2011 10:31:50 +0800
Subject: Btrfs: Always use 64bit inode number

There's a potential problem in 32bit system when we exhaust 32bit inode
numbers and start to allocate big inode numbers, because btrfs uses
inode->i_ino in many places.

So here we always use BTRFS_I(inode)->location.objectid, which is an
u64 variable.

There are 2 exceptions that BTRFS_I(inode)->location.objectid !=
inode->i_ino: the btree inode (0 vs 1) and empty subvol dirs (256 vs 2),
and inode->i_ino will be used in those cases.

Another reason to make this change is I'm going to use a special inode
to save free ino cache, and the inode number must be > (u64)-256.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
---
 fs/btrfs/relocation.c | 24 +++++++++++++-----------
 1 file changed, 13 insertions(+), 11 deletions(-)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index e6cb89357256..7b75e0c8ef8d 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -1410,9 +1410,9 @@ again:
 		prev = node;
 		entry = rb_entry(node, struct btrfs_inode, rb_node);
 
-		if (objectid < entry->vfs_inode.i_ino)
+		if (objectid < btrfs_ino(&entry->vfs_inode))
 			node = node->rb_left;
-		else if (objectid > entry->vfs_inode.i_ino)
+		else if (objectid > btrfs_ino(&entry->vfs_inode))
 			node = node->rb_right;
 		else
 			break;
@@ -1420,7 +1420,7 @@ again:
 	if (!node) {
 		while (prev) {
 			entry = rb_entry(prev, struct btrfs_inode, rb_node);
-			if (objectid <= entry->vfs_inode.i_ino) {
+			if (objectid <= btrfs_ino(&entry->vfs_inode)) {
 				node = prev;
 				break;
 			}
@@ -1435,7 +1435,7 @@ again:
 			return inode;
 		}
 
-		objectid = entry->vfs_inode.i_ino + 1;
+		objectid = btrfs_ino(&entry->vfs_inode) + 1;
 		if (cond_resched_lock(&root->inode_lock))
 			goto again;
 
@@ -1471,7 +1471,7 @@ static int get_new_location(struct inode *reloc_inode, u64 *new_bytenr,
 		return -ENOMEM;
 
 	bytenr -= BTRFS_I(reloc_inode)->index_cnt;
-	ret = btrfs_lookup_file_extent(NULL, root, path, reloc_inode->i_ino,
+	ret = btrfs_lookup_file_extent(NULL, root, path, btrfs_ino(reloc_inode),
 				       bytenr, 0);
 	if (ret < 0)
 		goto out;
@@ -1559,11 +1559,11 @@ int replace_file_extents(struct btrfs_trans_handle *trans,
 			if (first) {
 				inode = find_next_inode(root, key.objectid);
 				first = 0;
-			} else if (inode && inode->i_ino < key.objectid) {
+			} else if (inode && btrfs_ino(inode) < key.objectid) {
 				btrfs_add_delayed_iput(inode);
 				inode = find_next_inode(root, key.objectid);
 			}
-			if (inode && inode->i_ino == key.objectid) {
+			if (inode && btrfs_ino(inode) == key.objectid) {
 				end = key.offset +
 				      btrfs_file_extent_num_bytes(leaf, fi);
 				WARN_ON(!IS_ALIGNED(key.offset,
@@ -1894,6 +1894,7 @@ static int invalidate_extent_cache(struct btrfs_root *root,
 	struct inode *inode = NULL;
 	u64 objectid;
 	u64 start, end;
+	u64 ino;
 
 	objectid = min_key->objectid;
 	while (1) {
@@ -1906,17 +1907,18 @@ static int invalidate_extent_cache(struct btrfs_root *root,
 		inode = find_next_inode(root, objectid);
 		if (!inode)
 			break;
+		ino = btrfs_ino(inode);
 
-		if (inode->i_ino > max_key->objectid) {
+		if (ino > max_key->objectid) {
 			iput(inode);
 			break;
 		}
 
-		objectid = inode->i_ino + 1;
+		objectid = ino + 1;
 		if (!S_ISREG(inode->i_mode))
 			continue;
 
-		if (unlikely(min_key->objectid == inode->i_ino)) {
+		if (unlikely(min_key->objectid == ino)) {
 			if (min_key->type > BTRFS_EXTENT_DATA_KEY)
 				continue;
 			if (min_key->type < BTRFS_EXTENT_DATA_KEY)
@@ -1929,7 +1931,7 @@ static int invalidate_extent_cache(struct btrfs_root *root,
 			start = 0;
 		}
 
-		if (unlikely(max_key->objectid == inode->i_ino)) {
+		if (unlikely(max_key->objectid == ino)) {
 			if (max_key->type < BTRFS_EXTENT_DATA_KEY)
 				continue;
 			if (max_key->type > BTRFS_EXTENT_DATA_KEY) {
-- 
cgit v1.2.2


From c704005d886cf0bc9bc3974eb009b22fe0da32c7 Mon Sep 17 00:00:00 2001
From: David Sterba <dsterba@suse.cz>
Date: Tue, 19 Apr 2011 18:00:01 +0200
Subject: btrfs: unify checking of IS_ERR and null

use IS_ERR_OR_NULL when possible, done by this coccinelle script:

@ match @
identifier id;
@@
(
- BUG_ON(IS_ERR(id) || !id);
+ BUG_ON(IS_ERR_OR_NULL(id));
|
- IS_ERR(id) || !id
+ IS_ERR_OR_NULL(id)
|
- !id || IS_ERR(id)
+ IS_ERR_OR_NULL(id)
)

Signed-off-by: David Sterba <dsterba@suse.cz>
---
 fs/btrfs/relocation.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 199a80134312..fed0aaec0753 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -3220,7 +3220,7 @@ static int delete_block_group_cache(struct btrfs_fs_info *fs_info,
 	key.offset = 0;
 
 	inode = btrfs_iget(fs_info->sb, &key, root, NULL);
-	if (!inode || IS_ERR(inode) || is_bad_inode(inode)) {
+	if (IS_ERR_OR_NULL(inode) || is_bad_inode(inode)) {
 		if (inode && !IS_ERR(inode))
 			iput(inode);
 		return -ENOENT;
-- 
cgit v1.2.2


From f993c883ad8e111fb9e9ae603540acbe94f7246c Mon Sep 17 00:00:00 2001
From: David Sterba <dsterba@suse.cz>
Date: Wed, 20 Apr 2011 23:35:57 +0200
Subject: btrfs: drop unused argument from extent_io_tree_init

all callers pass GFP_NOFS, but the GFP mask argument is not used in the
function; GFP_ATOMIC is passed to radix tree initialization and it's the
only correct one, since we're using the preload/insert mechanism of
radix tree.
Let's drop the gfp mask from btrfs function, this will not change
behaviour.

Signed-off-by: David Sterba <dsterba@suse.cz>
---
 fs/btrfs/relocation.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index fed0aaec0753..f3edf45317bc 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -3935,7 +3935,7 @@ static struct reloc_control *alloc_reloc_control(void)
 	INIT_LIST_HEAD(&rc->reloc_roots);
 	backref_cache_init(&rc->backref_cache);
 	mapping_tree_init(&rc->reloc_root_tree);
-	extent_io_tree_init(&rc->processed_blocks, NULL, GFP_NOFS);
+	extent_io_tree_init(&rc->processed_blocks, NULL);
 	return rc;
 }
 
-- 
cgit v1.2.2


From 172ddd60a662c4d8bf2809462866ddddd6431ea5 Mon Sep 17 00:00:00 2001
From: David Sterba <dsterba@suse.cz>
Date: Thu, 21 Apr 2011 00:48:27 +0200
Subject: btrfs: drop gfp parameter from alloc_extent_map

pass GFP_NOFS directly to kmem_cache_alloc

Signed-off-by: David Sterba <dsterba@suse.cz>
---
 fs/btrfs/relocation.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index f3edf45317bc..2097a88f60aa 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -2870,7 +2870,7 @@ int setup_extent_mapping(struct inode *inode, u64 start, u64 end,
 	struct extent_map *em;
 	int ret = 0;
 
-	em = alloc_extent_map(GFP_NOFS);
+	em = alloc_extent_map();
 	if (!em)
 		return -ENOMEM;
 
-- 
cgit v1.2.2


From b3b4aa74b58bded927f579fff787fb6fa1c0393c Mon Sep 17 00:00:00 2001
From: David Sterba <dsterba@suse.cz>
Date: Thu, 21 Apr 2011 01:20:15 +0200
Subject: btrfs: drop unused parameter from btrfs_release_path

parameter tree root it's not used since commit
5f39d397dfbe140a14edecd4e73c34ce23c4f9ee ("Btrfs: Create extent_buffer
interface for large blocksizes")

Signed-off-by: David Sterba <dsterba@suse.cz>
---
 fs/btrfs/relocation.c | 30 +++++++++++++++---------------
 1 file changed, 15 insertions(+), 15 deletions(-)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 2097a88f60aa..f7b799b151aa 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -961,7 +961,7 @@ again:
 			lower = upper;
 			upper = NULL;
 		}
-		btrfs_release_path(root, path2);
+		btrfs_release_path(path2);
 next:
 		if (ptr < end) {
 			ptr += btrfs_extent_inline_ref_size(key.type);
@@ -974,7 +974,7 @@ next:
 		if (ptr >= end)
 			path1->slots[0]++;
 	}
-	btrfs_release_path(rc->extent_root, path1);
+	btrfs_release_path(path1);
 
 	cur->checked = 1;
 	WARN_ON(exist);
@@ -1749,7 +1749,7 @@ again:
 
 		btrfs_node_key_to_cpu(path->nodes[level], &key,
 				      path->slots[level]);
-		btrfs_release_path(src, path);
+		btrfs_release_path(path);
 
 		path->lowest_level = level;
 		ret = btrfs_search_slot(trans, src, &key, path, 0, 1);
@@ -2496,7 +2496,7 @@ static int do_relocation(struct btrfs_trans_handle *trans,
 			path->locks[upper->level] = 0;
 
 			slot = path->slots[upper->level];
-			btrfs_release_path(NULL, path);
+			btrfs_release_path(path);
 		} else {
 			ret = btrfs_bin_search(upper->eb, key, upper->level,
 					       &slot);
@@ -2737,7 +2737,7 @@ static int relocate_tree_block(struct btrfs_trans_handle *trans,
 		} else {
 			path->lowest_level = node->level;
 			ret = btrfs_search_slot(trans, root, key, path, 0, 1);
-			btrfs_release_path(root, path);
+			btrfs_release_path(path);
 			if (ret > 0)
 				ret = 0;
 		}
@@ -3119,7 +3119,7 @@ static int add_tree_block(struct reloc_control *rc,
 #endif
 	}
 
-	btrfs_release_path(rc->extent_root, path);
+	btrfs_release_path(path);
 
 	BUG_ON(level == -1);
 
@@ -3505,7 +3505,7 @@ int add_data_references(struct reloc_control *rc,
 		}
 		path->slots[0]++;
 	}
-	btrfs_release_path(rc->extent_root, path);
+	btrfs_release_path(path);
 	if (err)
 		free_block_list(blocks);
 	return err;
@@ -3568,7 +3568,7 @@ next:
 					    EXTENT_DIRTY);
 
 		if (ret == 0 && start <= key.objectid) {
-			btrfs_release_path(rc->extent_root, path);
+			btrfs_release_path(path);
 			rc->search_start = end + 1;
 		} else {
 			rc->search_start = key.objectid + key.offset;
@@ -3576,7 +3576,7 @@ next:
 			return 0;
 		}
 	}
-	btrfs_release_path(rc->extent_root, path);
+	btrfs_release_path(path);
 	return ret;
 }
 
@@ -3713,7 +3713,7 @@ restart:
 				flags = BTRFS_EXTENT_FLAG_DATA;
 
 			if (path_change) {
-				btrfs_release_path(rc->extent_root, path);
+				btrfs_release_path(path);
 
 				path->search_commit_root = 1;
 				path->skip_locking = 1;
@@ -3736,7 +3736,7 @@ restart:
 			   (flags & BTRFS_EXTENT_FLAG_DATA)) {
 			ret = add_data_references(rc, &key, path, &blocks);
 		} else {
-			btrfs_release_path(rc->extent_root, path);
+			btrfs_release_path(path);
 			ret = 0;
 		}
 		if (ret < 0) {
@@ -3799,7 +3799,7 @@ restart:
 		}
 	}
 
-	btrfs_release_path(rc->extent_root, path);
+	btrfs_release_path(path);
 	clear_extent_bits(&rc->processed_blocks, 0, (u64)-1, EXTENT_DIRTY,
 			  GFP_NOFS);
 
@@ -3867,7 +3867,7 @@ static int __insert_orphan_inode(struct btrfs_trans_handle *trans,
 	btrfs_set_inode_flags(leaf, item, BTRFS_INODE_NOCOMPRESS |
 					  BTRFS_INODE_PREALLOC);
 	btrfs_mark_buffer_dirty(leaf);
-	btrfs_release_path(root, path);
+	btrfs_release_path(path);
 out:
 	btrfs_free_path(path);
 	return ret;
@@ -4109,7 +4109,7 @@ int btrfs_recover_relocation(struct btrfs_root *root)
 		}
 		leaf = path->nodes[0];
 		btrfs_item_key_to_cpu(leaf, &key, path->slots[0]);
-		btrfs_release_path(root->fs_info->tree_root, path);
+		btrfs_release_path(path);
 
 		if (key.objectid != BTRFS_TREE_RELOC_OBJECTID ||
 		    key.type != BTRFS_ROOT_ITEM_KEY)
@@ -4141,7 +4141,7 @@ int btrfs_recover_relocation(struct btrfs_root *root)
 
 		key.offset--;
 	}
-	btrfs_release_path(root->fs_info->tree_root, path);
+	btrfs_release_path(path);
 
 	if (list_empty(&reloc_roots))
 		goto out;
-- 
cgit v1.2.2


From f2a97a9dbd86eb1ef956bdf20e05c507b32beb96 Mon Sep 17 00:00:00 2001
From: David Sterba <dsterba@suse.cz>
Date: Thu, 5 May 2011 12:44:41 +0200
Subject: btrfs: remove all unused functions

Remove static and global declarations and/or definitions. Reduces size
of btrfs.ko by ~3.4kB.

  text    data     bss     dec     hex filename
402081    7464     200  409745   64091 btrfs.ko.base
398620    7144     200  405964   631cc btrfs.ko.remove-all

Signed-off-by: David Sterba <dsterba@suse.cz>
---
 fs/btrfs/relocation.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index f7b799b151aa..f726e72dd362 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -507,6 +507,7 @@ static int update_backref_cache(struct btrfs_trans_handle *trans,
 	return 1;
 }
 
+
 static int should_ignore_root(struct btrfs_root *root)
 {
 	struct btrfs_root *reloc_root;
@@ -529,7 +530,6 @@ static int should_ignore_root(struct btrfs_root *root)
 	 */
 	return 1;
 }
-
 /*
  * find reloc tree by address of tree root
  */
-- 
cgit v1.2.2


From 70f23fd66bc821a0e99647f70a809e277cc93c4c Mon Sep 17 00:00:00 2001
From: "Justin P. Mattock" <justinmattock@gmail.com>
Date: Tue, 10 May 2011 10:16:21 +0200
Subject: treewide: fix a few typos in comments

- kenrel -> kernel
- whetehr -> whether
- ttt -> tt
- sss -> ss

Signed-off-by: Justin P. Mattock <justinmattock@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
---
 fs/btrfs/relocation.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 199a80134312..f340f7c99d09 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -709,7 +709,7 @@ again:
 	WARN_ON(cur->checked);
 	if (!list_empty(&cur->upper)) {
 		/*
-		 * the backref was added previously when processsing
+		 * the backref was added previously when processing
 		 * backref of type BTRFS_TREE_BLOCK_REF_KEY
 		 */
 		BUG_ON(!list_is_singular(&cur->upper));
-- 
cgit v1.2.2


From a2de733c78fa7af51ba9670482fa7d392aa67c57 Mon Sep 17 00:00:00 2001
From: Arne Jansen <sensille@gmx.net>
Date: Tue, 8 Mar 2011 14:14:00 +0100
Subject: btrfs: scrub

This adds an initial implementation for scrub. It works quite
straightforward. The usermode issues an ioctl for each device in the
fs. For each device, it enumerates the allocated device chunks. For
each chunk, the contained extents are enumerated and the data checksums
fetched. The extents are read sequentially and the checksums verified.
If an error occurs (checksum or EIO), a good copy is searched for. If
one is found, the bad copy will be rewritten.
All enumerations happen from the commit roots. During a transaction
commit, the scrubs get paused and afterwards continue from the new
roots.

This commit is based on the series originally posted to linux-btrfs
with some improvements that resulted from comments from David Sterba,
Ilya Dryomov and Jan Schmidt.

Signed-off-by: Arne Jansen <sensille@gmx.net>
---
 fs/btrfs/relocation.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 58250e09eb05..db1dffa9952b 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -4242,7 +4242,7 @@ int btrfs_reloc_clone_csums(struct inode *inode, u64 file_pos, u64 len)
 
 	disk_bytenr = file_pos + BTRFS_I(inode)->index_cnt;
 	ret = btrfs_lookup_csums_range(root->fs_info->csum_root, disk_bytenr,
-				       disk_bytenr + len - 1, &list);
+				       disk_bytenr + len - 1, &list, 0);
 
 	while (!list_empty(&list)) {
 		sums = list_entry(list.next, struct btrfs_ordered_sum, list);
-- 
cgit v1.2.2


From 7a7eaa40a39bde4eefc91aadeb1ce3dc4e6a1252 Mon Sep 17 00:00:00 2001
From: Josef Bacik <josef@redhat.com>
Date: Wed, 13 Apr 2011 12:54:33 -0400
Subject: Btrfs: take away the num_items argument from btrfs_join_transaction

I keep forgetting that btrfs_join_transaction() just ignores the num_items
argument, which leads me to sending pointless patches and looking stupid :).  So
just kill the num_items argument from btrfs_join_transaction and
btrfs_start_ioctl_transaction, since neither of them use it.  Thanks,

Signed-off-by: Josef Bacik <josef@redhat.com>
---
 fs/btrfs/relocation.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 199a80134312..8bb256667f2d 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -2149,7 +2149,7 @@ again:
 			err = ret;
 	}
 
-	trans = btrfs_join_transaction(rc->extent_root, 1);
+	trans = btrfs_join_transaction(rc->extent_root);
 	if (IS_ERR(trans)) {
 		if (!err)
 			btrfs_block_rsv_release(rc->extent_root,
@@ -3233,7 +3233,7 @@ truncate:
 		goto out;
 	}
 
-	trans = btrfs_join_transaction(root, 0);
+	trans = btrfs_join_transaction(root);
 	if (IS_ERR(trans)) {
 		btrfs_free_path(path);
 		ret = PTR_ERR(trans);
@@ -3642,7 +3642,7 @@ int prepare_to_relocate(struct reloc_control *rc)
 	rc->create_reloc_tree = 1;
 	set_reloc_control(rc);
 
-	trans = btrfs_join_transaction(rc->extent_root, 1);
+	trans = btrfs_join_transaction(rc->extent_root);
 	BUG_ON(IS_ERR(trans));
 	btrfs_commit_transaction(trans, rc->extent_root);
 	return 0;
@@ -3831,7 +3831,7 @@ restart:
 	btrfs_block_rsv_release(rc->extent_root, rc->block_rsv, (u64)-1);
 
 	/* get rid of pinned extents */
-	trans = btrfs_join_transaction(rc->extent_root, 1);
+	trans = btrfs_join_transaction(rc->extent_root);
 	if (IS_ERR(trans))
 		err = PTR_ERR(trans);
 	else
@@ -4156,7 +4156,7 @@ int btrfs_recover_relocation(struct btrfs_root *root)
 
 	set_reloc_control(rc);
 
-	trans = btrfs_join_transaction(rc->extent_root, 1);
+	trans = btrfs_join_transaction(rc->extent_root);
 	if (IS_ERR(trans)) {
 		unset_reloc_control(rc);
 		err = PTR_ERR(trans);
@@ -4190,7 +4190,7 @@ int btrfs_recover_relocation(struct btrfs_root *root)
 
 	unset_reloc_control(rc);
 
-	trans = btrfs_join_transaction(rc->extent_root, 1);
+	trans = btrfs_join_transaction(rc->extent_root);
 	if (IS_ERR(trans))
 		err = PTR_ERR(trans);
 	else
-- 
cgit v1.2.2


From a4abeea41adfa3c143c289045f4625dfaeba2212 Mon Sep 17 00:00:00 2001
From: Josef Bacik <josef@redhat.com>
Date: Mon, 11 Apr 2011 17:25:13 -0400
Subject: Btrfs: kill trans_mutex

We use trans_mutex for lots of things, here's a basic list

1) To serialize trans_handles joining the currently running transaction
2) To make sure that no new trans handles are started while we are committing
3) To protect the dead_roots list and the transaction lists

Really the serializing trans_handles joining is not too hard, and can really get
bogged down in acquiring a reference to the transaction.  So replace the
trans_mutex with a trans_lock spinlock and use it to do the following

1) Protect fs_info->running_transaction.  All trans handles have to do is check
this, and then take a reference of the transaction and keep on going.
2) Protect the fs_info->trans_list.  This doesn't get used too much, basically
it just holds the current transactions, which will usually just be the currently
committing transaction and the currently running transaction at most.
3) Protect the dead roots list.  This is only ever processed by splicing the
list so this is relatively simple.
4) Protect the fs_info->reloc_ctl stuff.  This is very lightweight and was using
the trans_mutex before, so this is a pretty straightforward change.
5) Protect fs_info->no_trans_join.  Because we don't hold the trans_lock over
the entirety of the commit we need to have a way to block new people from
creating a new transaction while we're doing our work.  So we set no_trans_join
and in join_transaction we test to see if that is set, and if it is we do a
wait_on_commit.
6) Make the transaction use count atomic so we don't need to take locks to
modify it when we're dropping references.
7) Add a commit_lock to the transaction to make sure multiple people trying to
commit the same transaction don't race and commit at the same time.
8) Make open_ioctl_trans an atomic so we don't have to take any locks for ioctl
trans.

I have tested this with xfstests, but obviously it is a pretty hairy change so
lots of testing is greatly appreciated.  Thanks,

Signed-off-by: Josef Bacik <josef@redhat.com>
---
 fs/btrfs/relocation.c | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 8bb256667f2d..09c30d37d43e 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -2136,10 +2136,10 @@ int prepare_to_merge(struct reloc_control *rc, int err)
 	u64 num_bytes = 0;
 	int ret;
 
-	mutex_lock(&root->fs_info->trans_mutex);
+	spin_lock(&root->fs_info->trans_lock);
 	rc->merging_rsv_size += root->nodesize * (BTRFS_MAX_LEVEL - 1) * 2;
 	rc->merging_rsv_size += rc->nodes_relocated * 2;
-	mutex_unlock(&root->fs_info->trans_mutex);
+	spin_unlock(&root->fs_info->trans_lock);
 again:
 	if (!err) {
 		num_bytes = rc->merging_rsv_size;
@@ -2208,9 +2208,9 @@ int merge_reloc_roots(struct reloc_control *rc)
 	int ret;
 again:
 	root = rc->extent_root;
-	mutex_lock(&root->fs_info->trans_mutex);
+	spin_lock(&root->fs_info->trans_lock);
 	list_splice_init(&rc->reloc_roots, &reloc_roots);
-	mutex_unlock(&root->fs_info->trans_mutex);
+	spin_unlock(&root->fs_info->trans_lock);
 
 	while (!list_empty(&reloc_roots)) {
 		found = 1;
@@ -3583,17 +3583,17 @@ next:
 static void set_reloc_control(struct reloc_control *rc)
 {
 	struct btrfs_fs_info *fs_info = rc->extent_root->fs_info;
-	mutex_lock(&fs_info->trans_mutex);
+	spin_lock(&fs_info->trans_lock);
 	fs_info->reloc_ctl = rc;
-	mutex_unlock(&fs_info->trans_mutex);
+	spin_unlock(&fs_info->trans_lock);
 }
 
 static void unset_reloc_control(struct reloc_control *rc)
 {
 	struct btrfs_fs_info *fs_info = rc->extent_root->fs_info;
-	mutex_lock(&fs_info->trans_mutex);
+	spin_lock(&fs_info->trans_lock);
 	fs_info->reloc_ctl = NULL;
-	mutex_unlock(&fs_info->trans_mutex);
+	spin_unlock(&fs_info->trans_lock);
 }
 
 static int check_extent_flags(u64 flags)
-- 
cgit v1.2.2


From 026fd317828500524cdc7e5ff9e8e7923abb2868 Mon Sep 17 00:00:00 2001
From: Josef Bacik <josef@redhat.com>
Date: Fri, 13 May 2011 10:32:11 -0400
Subject: Btrfs: don't always do readahead

Our readahead is sort of sloppy, and really isn't always needed.  For example if
ls is doing a stating ls (which is the default) it's going to stat in non-disk
order, so if say you have a directory with a stupid amount of files, readahead
is going to do nothing but waste time in the case of doing the stat.  Taking the
unconditional readahead out made my test go from 57 minutes to 36 minutes.  This
means that everywhere we do loop through the tree we want to make sure we do set
path->reada properly, so I went through and found all of the places where we
loop through the path and set reada to 1.  Thanks,

Signed-off-by: Josef Bacik <josef@redhat.com>
---
 fs/btrfs/relocation.c | 6 ++++++
 1 file changed, 6 insertions(+)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 09c30d37d43e..5872b41581f4 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -676,6 +676,8 @@ struct backref_node *build_backref_tree(struct reloc_control *rc,
 		err = -ENOMEM;
 		goto out;
 	}
+	path1->reada = 1;
+	path2->reada = 2;
 
 	node = alloc_backref_node(cache);
 	if (!node) {
@@ -1996,6 +1998,7 @@ static noinline_for_stack int merge_reloc_root(struct reloc_control *rc,
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
+	path->reada = 1;
 
 	reloc_root = root->reloc_root;
 	root_item = &reloc_root->root_item;
@@ -3297,6 +3300,7 @@ static int find_data_references(struct reloc_control *rc,
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
+	path->reada = 1;
 
 	root = read_fs_root(rc->extent_root->fs_info, ref_root);
 	if (IS_ERR(root)) {
@@ -3665,6 +3669,7 @@ static noinline_for_stack int relocate_block_group(struct reloc_control *rc)
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
+	path->reada = 1;
 
 	ret = prepare_to_relocate(rc);
 	if (ret) {
@@ -4090,6 +4095,7 @@ int btrfs_recover_relocation(struct btrfs_root *root)
 	path = btrfs_alloc_path();
 	if (!path)
 		return -ENOMEM;
+	path->reada = -1;
 
 	key.objectid = BTRFS_TREE_RELOC_OBJECTID;
 	key.type = BTRFS_ROOT_ITEM_KEY;
-- 
cgit v1.2.2


From 7585717f304f5ed005cc4ad933a69aab3efbd136 Mon Sep 17 00:00:00 2001
From: Chris Mason <chris.mason@oracle.com>
Date: Mon, 13 Jun 2011 20:00:16 -0400
Subject: Btrfs: fix relocation races

The recent commit to get rid of our trans_mutex introduced
some races with block group relocation.  The problem is that relocation
needs to do some record keeping about each root, and it was relying
on the transaction mutex to coordinate things in subtle ways.

This fix adds a mutex just for the relocation code and makes sure
it doesn't have a big impact on normal operations.  The race is
really fixed in btrfs_record_root_in_trans, which is where we
step back and wait for the relocation code to finish accounting
setup.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
---
 fs/btrfs/relocation.c | 30 +++++++++++++++++++++---------
 1 file changed, 21 insertions(+), 9 deletions(-)

(limited to 'fs/btrfs/relocation.c')

diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index f25b10a22a0a..086b1e6b8614 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -1368,7 +1368,7 @@ int btrfs_update_reloc_root(struct btrfs_trans_handle *trans,
 	int ret;
 
 	if (!root->reloc_root)
-		return 0;
+		goto out;
 
 	reloc_root = root->reloc_root;
 	root_item = &reloc_root->root_item;
@@ -1390,6 +1390,8 @@ int btrfs_update_reloc_root(struct btrfs_trans_handle *trans,
 	ret = btrfs_update_root(trans, root->fs_info->tree_root,
 				&reloc_root->root_key, root_item);
 	BUG_ON(ret);
+
+out:
 	return 0;
 }
 
@@ -2142,10 +2144,11 @@ int prepare_to_merge(struct reloc_control *rc, int err)
 	u64 num_bytes = 0;
 	int ret;
 
-	spin_lock(&root->fs_info->trans_lock);
+	mutex_lock(&root->fs_info->reloc_mutex);
 	rc->merging_rsv_size += root->nodesize * (BTRFS_MAX_LEVEL - 1) * 2;
 	rc->merging_rsv_size += rc->nodes_relocated * 2;
-	spin_unlock(&root->fs_info->trans_lock);
+	mutex_unlock(&root->fs_info->reloc_mutex);
+
 again:
 	if (!err) {
 		num_bytes = rc->merging_rsv_size;
@@ -2214,9 +2217,16 @@ int merge_reloc_roots(struct reloc_control *rc)
 	int ret;
 again:
 	root = rc->extent_root;
-	spin_lock(&root->fs_info->trans_lock);
+
+	/*
+	 * this serializes us with btrfs_record_root_in_transaction,
+	 * we have to make sure nobody is in the middle of
+	 * adding their roots to the list while we are
+	 * doing this splice
+	 */
+	mutex_lock(&root->fs_info->reloc_mutex);
 	list_splice_init(&rc->reloc_roots, &reloc_roots);
-	spin_unlock(&root->fs_info->trans_lock);
+	mutex_unlock(&root->fs_info->reloc_mutex);
 
 	while (!list_empty(&reloc_roots)) {
 		found = 1;
@@ -3590,17 +3600,19 @@ next:
 static void set_reloc_control(struct reloc_control *rc)
 {
 	struct btrfs_fs_info *fs_info = rc->extent_root->fs_info;
-	spin_lock(&fs_info->trans_lock);
+
+	mutex_lock(&fs_info->reloc_mutex);
 	fs_info->reloc_ctl = rc;
-	spin_unlock(&fs_info->trans_lock);
+	mutex_unlock(&fs_info->reloc_mutex);
 }
 
 static void unset_reloc_control(struct reloc_control *rc)
 {
 	struct btrfs_fs_info *fs_info = rc->extent_root->fs_info;
-	spin_lock(&fs_info->trans_lock);
+
+	mutex_lock(&fs_info->reloc_mutex);
 	fs_info->reloc_ctl = NULL;
-	spin_unlock(&fs_info->trans_lock);
+	mutex_unlock(&fs_info->reloc_mutex);
 }
 
 static int check_extent_flags(u64 flags)
-- 
cgit v1.2.2