Re: [PATCH 1/2] Btrfs: cleanup extents after we finish logging inode
From: Josef Bacik <hidden>
Date: 2012-08-27 17:05:31
From: Josef Bacik <hidden>
Date: 2012-08-27 17:05:31
On Mon, Aug 27, 2012 at 10:52:19AM -0600, Liu Bo wrote:
This is based on Josef's "Btrfs: turbo charge fsync". We should cleanup those extents after we've finished logging inode, otherwise we may do redundant work on them. Signed-off-by: Liu Bo <redacted> --- fs/btrfs/tree-log.c | 6 ++++++ 1 files changed, 6 insertions(+), 0 deletions(-)diff --git a/fs/btrfs/tree-log.c b/fs/btrfs/tree-log.c index 5190cd6..e7365d7 100644 --- a/fs/btrfs/tree-log.c +++ b/fs/btrfs/tree-log.c@@ -3167,6 +3167,12 @@ next_slot: err = ret; goto out_unlock; } + } else { + struct extent_map_tree *tree = &BTRFS_I(inode)->extent_tree; + struct extent_map *em, *n; + + list_for_each_entry_safe(em, n, &tree->modified_extents, list) + list_del_init(&em->list); } if (inode_only == LOG_INODE_ALL && S_ISDIR(inode->i_mode)) {--
Ah thanks I had been meaning to do this but I kept forgetting. I'll add it to btrfs-next. Thanks, Josef