linux-raid.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PULL REQUEST] Several bug fixes for md in 3.7
@ 2012-11-23 19:40 NeilBrown
  0 siblings, 0 replies; only message in thread
From: NeilBrown @ 2012-11-23 19:40 UTC (permalink / raw)
  To: Linus Torvalds
  Cc: George Spelvin, Shaohua Li, majianpeng, zhuwenfeng, linux RAID,
	lkml

[-- Attachment #1: Type: text/plain, Size: 2161 bytes --]


Coming a bit late to the -rc party, but here are some bug fixes found
recently through testing, regular usages, and code inspection.

Most worrying is a data corruption risk when 'replacing' devices in a RAID10.
Maybe most interesting is that generic DISCARD support requires granularity to
be a power of two, so RAID5 cannot make the  granularity be one-stripe in
most cases.  This probably makes DISCARD on raid5 less than useful.
Hopefully the generic code will get fixed.

NeilBrown


The following changes since commit ed30be077e705e0dff53bfc51d23feb8aeeab78f:

  MD RAID10: Fix oops when creating RAID10 arrays via dm-raid.c (2012-10-31 11:42:30 +1100)

are available in the git repository at:

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

for you to fetch changes up to 884162df2aadd7414bef4935e1a54976fd4e3988:

  md/raid10: decrement correct pending counter when writing to replacement. (2012-11-22 15:12:42 +1100)

----------------------------------------------------------------
Several bug fixes for md in 3.7

 - raid5 discard has problems
 - raid10 replacement devices have problems
 - bad block lock seqlock usage has problems
 - dm-raid doesn't free everything

----------------------------------------------------------------
NeilBrown (6):
      md: make sure everything is freed when dm-raid stops an array.
      md/raid5: round discard alignment up to power of 2.
      md/raid5: move resolving of reconstruct_state earlier in     stripe_handle.
      md/raid5: Make sure we clear R5_Discard when discard is finished.
      md/raid10: close race that lose writes lost when replacement completes.
      md/raid10: decrement correct pending counter when writing to replacement.

majianpeng (2):
      md: Reassigned the parameters if read_seqretry returned true in func md_is_badblock.
      md: Avoid write invalid address if read_seqretry returned true.

 drivers/md/md.c     |  27 ++++++++---
 drivers/md/raid10.c | 131 +++++++++++++++++++++++++++-------------------------
 drivers/md/raid5.c  |  79 ++++++++++++++++---------------
 3 files changed, 132 insertions(+), 105 deletions(-)

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 828 bytes --]

^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2012-11-23 19:40 UTC | newest]

Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2012-11-23 19:40 [PULL REQUEST] Several bug fixes for md in 3.7 NeilBrown

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).