[PULL REQUEST] md updates for 3.7

[NeilBrown]

MD updates for 3.7.
Latish because I needed to wait for the block tree to get merged.

Thanks,
NeilBrown



The following changes since commit 02f3939e1a9357b7c370a4a69717cf9c02452737:

  block: makes bio_split support bio without data (2012-09-28 10:38:48 +0200)

are available in the git repository at:

  git://neil.brown.name/md/ tags/md-3.7

for you to fetch changes up to 72f36d5972a166197036c1281963f6863c429bf2:

  md: refine reporting of resync/reshape delays. (2012-10-11 14:25:57 +1100)

----------------------------------------------------------------
md updates for 3.7

"discard" support, some dm-raid improvements and other assorted
bits and pieces.

----------------------------------------------------------------
Alexander Lyakas (1):
      md: When RAID5 is dirty, force reconstruct-write instead of read-modify-write.

Dan Carpenter (1):
      md/raid10: use correct limit variable

Denis Efremov (1):
      md/linear: rcu_dereference outside read-lock section

Jan Beulich (1):
      add further __init annotations to crypto/xor.c

Jianpeng Ma (2):
      md/bitmap:Don't use IS_ERR to judge alloc_page().
      Subject: [PATCH] md:change resync_mismatches to atomic64_t to avoid races

Jonathan Brassow (5):
      MD RAID10: Prep for DM RAID10 device replacement capability
      DM RAID: Move 'rebuild' checking code to its own function
      DM RAID: Add rebuild capability for RAID10
      DM RAID: Fix comparison of index and quantity for "rebuild" parameter
      DM RAID: Fix for "sync" directive ineffectiveness

Michael Wang (1):
      raid: replace list_for_each_continue_rcu with new interface

NeilBrown (10):
      md/raid10: submit IO from originating thread instead of md thread.
      md: avoid taking the mutex on some ioctls.
      md/raid1: Don't release reference to device while handling read error.
      md/raid5: add some missing locking in handle_failed_stripe.
      md/raid5: protect debug message against NULL derefernce.
      md/raid5: make sure to_read and to_write never go negative.
      md: writing to sync_action should clear the read-auto state.
      md: make sure manual changes to recovery checkpoint are saved.
      md/raid5: be careful not to resize_stripes too big.
      md: refine reporting of resync/reshape delays.

Shaohua Li (7):
      md: linear supports TRIM
      md: raid 0 supports TRIM
      md: raid 1 supports TRIM
      md: raid 10 supports TRIM
      MD: change the parameter of md thread
      MD: raid5 trim support
      MD: raid5 avoid unnecessary zero page for trim

 Documentation/device-mapper/dm-raid.txt |   9 ++
 crypto/xor.c                            |   4 +-
 drivers/md/bitmap.c                     |  17 +--
 drivers/md/dm-raid.c                    | 124 ++++++++++++++----
 drivers/md/linear.c                     |  25 +++-
 drivers/md/md.c                         | 145 +++++++++++++++------
 drivers/md/md.h                         |   9 +-
 drivers/md/multipath.c                  |   3 +-
 drivers/md/raid0.c                      |  19 ++-
 drivers/md/raid1.c                      |  37 ++++--
 drivers/md/raid10.c                     |  95 ++++++++++++--
 drivers/md/raid5.c                      | 219 ++++++++++++++++++++++++++++----
 drivers/md/raid5.h                      |   1 +
 13 files changed, 578 insertions(+), 129 deletions(-)

Attachment: signature.asc
Description: PGP signature