Skip to content

Commit

Permalink
Btrfs: lock the transition from dirty to writeback for an eb
Browse files Browse the repository at this point in the history
There is a small window where an eb can have no IO bits set on it, which
could potentially result in extent_buffer_under_io() returning false when we
want it to return true, which could result in not fun things happening.  So
in order to protect this case we need to hold the refs_lock when we make
this transition to make sure we get reliable results out of
extent_buffer_udner_io().  Thanks,

Signed-off-by: Josef Bacik <jbacik@fusionio.com>
  • Loading branch information
Josef Bacik authored and Chris Mason committed Jul 23, 2012
1 parent 594831c commit 51561ff
Showing 1 changed file with 9 additions and 0 deletions.
9 changes: 9 additions & 0 deletions fs/btrfs/extent_io.c
Original file line number Diff line number Diff line change
Expand Up @@ -3077,8 +3077,15 @@ static int lock_extent_buffer_for_io(struct extent_buffer *eb,
}
}

/*
* We need to do this to prevent races in people who check if the eb is
* under IO since we can end up having no IO bits set for a short period
* of time.
*/
spin_lock(&eb->refs_lock);
if (test_and_clear_bit(EXTENT_BUFFER_DIRTY, &eb->bflags)) {
set_bit(EXTENT_BUFFER_WRITEBACK, &eb->bflags);
spin_unlock(&eb->refs_lock);
btrfs_set_header_flag(eb, BTRFS_HEADER_FLAG_WRITTEN);
spin_lock(&fs_info->delalloc_lock);
if (fs_info->dirty_metadata_bytes >= eb->len)
Expand All @@ -3087,6 +3094,8 @@ static int lock_extent_buffer_for_io(struct extent_buffer *eb,
WARN_ON(1);
spin_unlock(&fs_info->delalloc_lock);
ret = 1;
} else {
spin_unlock(&eb->refs_lock);
}

btrfs_tree_unlock(eb);
Expand Down

0 comments on commit 51561ff

Please sign in to comment.