[GIT PULL] Btrfs updates

[Chris Mason <chris.mason@oracle.com>]

Hi everyone,

The for-linus branch of the btrfs unstable tree:

git://git.kernel.org/pub/scm/linux/kernel/git/mason/btrfs-unstable.git for-linus

Has our current set of fixes.  Linus will probably notice the head
commit was from this morning.  I reordered a refcount bump inside the
lock that protects the delayed inode tree and did a commit --amend.

This is just for clarity, the delayed_inode can't go away because it is
pinned by the inode itself.  But, I don't want that code copy & pasted
elsewhere to create bugs later on.

These fix a regression in the new delayed_inode code where inodes that
are dropped from ram need to check for a delayed inode before we load
them back in.

Josef's commit avoids oopsen if the free extent cache had
to be discarded, and my commit makes sure a newly created empty file
gets logged properly if someone calls fsync on it.

Miao Xie (1) commits (+91/-26):
    btrfs: fix inconsonant inode information

Josef Bacik (1) commits (+6/-3):
    Btrfs: make sure to update total_bitmaps when freeing cache V3

Ilya Dryomov (1) commits (+2/-2):
    Btrfs: fix type mismatch in find_free_extent()

Chris Mason (1) commits (+1/-0):
    Btrfs: make sure to record the transid in new inodes

Total: (4) commits (+100/-31)

 fs/btrfs/delayed-inode.c    |  104 +++++++++++++++++++++++++++++++++----------
 fs/btrfs/delayed-inode.h    |    1 +
 fs/btrfs/extent-tree.c      |    4 +-
 fs/btrfs/free-space-cache.c |    9 +++-
 fs/btrfs/inode.c            |   13 +++++-
 5 files changed, 100 insertions(+), 31 deletions(-)