linux-btrfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH] Btrfs: fix race of using total_bytes_pinned
@ 2014-06-30  3:49 Liu Bo
  0 siblings, 0 replies; only message in thread
From: Liu Bo @ 2014-06-30  3:49 UTC (permalink / raw)
  To: linux-btrfs

This percpu counter @total_bytes_pinned is introduced to skip unnecessary
operations of 'commit transaction', it accounts for those space we may free
but are stuck in delayed refs.

And we zero out @space_info->total_bytes_pinned every transaction period so
we have a better idea of how much space we'll actually free up by committing
this transaction.  However, we do the 'zero out' part a little earlier, before
we actually unpin space, so we end up returning ENOSPC when we actually have
free space that's just unpinned from committing transaction.

xfstests/generic/074 complained then.

This fixes it by zeroing out the percpu pinned number after 'unpin' and when
finding space for writing, if it finds that pinned bytes is less than needed,
we first try again to check if we have space now, as someone may have
committed transaction, yes means we're good, while no means we really have
run out of space.

Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
---
 fs/btrfs/extent-tree.c | 22 ++++++++++++++--------
 1 file changed, 14 insertions(+), 8 deletions(-)

diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 99c2539..35c68cd 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -3685,7 +3685,7 @@ int btrfs_check_data_free_space(struct inode *inode, u64 bytes)
 	struct btrfs_root *root = BTRFS_I(inode)->root;
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	u64 used;
-	int ret = 0, committed = 0, alloc_chunk = 1;
+	int ret = 0, committed = 0, alloc_chunk = 1, check_pinned = 0;
 
 	/* make sure bytes are sectorsize aligned */
 	bytes = ALIGN(bytes, root->sectorsize);
@@ -3756,11 +3756,18 @@ alloc:
 		 * allocation don't bother committing the transaction.
 		 */
 		if (percpu_counter_compare(&data_sinfo->total_bytes_pinned,
-					   bytes) < 0)
-			committed = 1;
+					   bytes) < 0) {
+			if (check_pinned)
+				committed = 1;	/* really run out of space. */
+			else {
+				check_pinned = 1;
+				spin_unlock(&data_sinfo->lock);
+				goto again;
+			}
+		}
 		spin_unlock(&data_sinfo->lock);
 
-		/* commit the current transaction and try again */
+	/* commit the current transaction and try again */
 commit_trans:
 		if (!committed &&
 		    !atomic_read(&root->fs_info->open_ioctl_trans)) {
@@ -5678,7 +5685,6 @@ void btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans,
 	struct btrfs_caching_control *next;
 	struct btrfs_caching_control *caching_ctl;
 	struct btrfs_block_group_cache *cache;
-	struct btrfs_space_info *space_info;
 
 	down_write(&fs_info->commit_root_sem);
 
@@ -5701,9 +5707,6 @@ void btrfs_prepare_extent_commit(struct btrfs_trans_handle *trans,
 
 	up_write(&fs_info->commit_root_sem);
 
-	list_for_each_entry_rcu(space_info, &fs_info->space_info, list)
-		percpu_counter_set(&space_info->total_bytes_pinned, 0);
-
 	update_global_block_rsv(fs_info);
 }
 
@@ -5771,6 +5774,7 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
 {
 	struct btrfs_fs_info *fs_info = root->fs_info;
 	struct extent_io_tree *unpin;
+	struct btrfs_space_info *space_info;
 	u64 start;
 	u64 end;
 	int ret;
@@ -5798,6 +5802,8 @@ int btrfs_finish_extent_commit(struct btrfs_trans_handle *trans,
 		cond_resched();
 	}
 
+	list_for_each_entry_rcu(space_info, &fs_info->space_info, list)
+		percpu_counter_set(&space_info->total_bytes_pinned, 0);
 	return 0;
 }
 
-- 
1.8.1.4


^ permalink raw reply related	[flat|nested] only message in thread

only message in thread, other threads:[~2014-06-30  3:49 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2014-06-30  3:49 [PATCH] Btrfs: fix race of using total_bytes_pinned Liu Bo

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).