Btrfs: avoid ENOSPC errors in btrfs_dirty_inode
authorChris Mason <chris.mason@oracle.com>
Wed, 26 May 2010 15:02:00 +0000 (11:02 -0400)
committerChris Mason <chris.mason@oracle.com>
Wed, 26 May 2010 15:02:00 +0000 (11:02 -0400)
commit94b604429a6cd94ddc128fa2772c57209bb1318f
tree700c5b9cb431bde48f07813224e6a5165a9d6540
parent3f7c579c41a3d20af76fd6ff1f6b949edf105fd1
Btrfs: avoid ENOSPC errors in btrfs_dirty_inode

btrfs_dirty_inode tries to sneak in without much waiting or
space reservation, mostly for performance reasons.  This
usually works well but can cause problems when there are
many many writers.

When btrfs_update_inode fails with ENOSPC, we fallback
to a slower btrfs_start_transaction call that will reserve
some space.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
fs/btrfs/inode.c