linux.git/drivers/md, branch v4.12-rc4

Merge tag 'for-4.12/dm-fixes-3' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm

2017-06-02T18:50:37+00:00

Pull device mapper fixes from Mike Snitzer:

 - a DM verity fix for a mode when no salt is used

 - a fix to DM to account for the possibility that PREFLUSH or FUA are
   used without the SYNC flag if the underlying storage doesn't have a
   volatile write-cache

 - a DM ioctl memory allocation flag fix to use __GFP_HIGH to allow
   emergency forward progress (by using memory reserves as last resort)

 - a small DM integrity cleanup to use kvmalloc() instead of duplicating
   the same

* tag 'for-4.12/dm-fixes-3' of git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm:
  dm: make flush bios explicitly sync
  dm ioctl: restore __GFP_HIGH in copy_params()
  dm integrity: use kvmalloc() instead of dm_integrity_kvmalloc()
  dm verity: fix no salt use case

md: Make flush bios explicitely sync

2017-05-31T16:25:53+00:00

Commit b685d3d65ac7 "block: treat REQ_FUA and REQ_PREFLUSH as
synchronous" removed REQ_SYNC flag from WRITE_{FUA|PREFLUSH|...}
definitions.  generic_make_request_checks() however strips REQ_FUA and
REQ_PREFLUSH flags from a bio when the storage doesn't report volatile
write cache and thus write effectively becomes asynchronous which can
lead to performance regressions

Fix the problem by making sure all bios which are synchronous are
properly marked with REQ_SYNC.

CC: linux-raid@vger.kernel.org
CC: Shaohua Li 
Fixes: b685d3d65ac791406e0dfd8779cc9b3707fea5a3
CC: stable@vger.kernel.org
Signed-off-by: Jan Kara 
Signed-off-by: Shaohua Li

dm: make flush bios explicitly sync

2017-05-31T14:50:23+00:00

Commit b685d3d65ac7 ("block: treat REQ_FUA and REQ_PREFLUSH as
synchronous") removed REQ_SYNC flag from WRITE_{FUA|PREFLUSH|...}
definitions.  generic_make_request_checks() however strips REQ_FUA and
REQ_PREFLUSH flags from a bio when the storage doesn't report volatile
write cache and thus write effectively becomes asynchronous which can
lead to performance regressions.

Fix the problem by making sure all bios which are synchronous are
properly marked with REQ_SYNC.

Fixes: b685d3d65ac7 ("block: treat REQ_FUA and REQ_PREFLUSH as synchronous")
Cc: stable@vger.kernel.org
Signed-off-by: Jan Kara 
Signed-off-by: Mike Snitzer

md: report sector of stripes with check mismatches

2017-05-24T23:31:13+00:00

This makes it possible, with appropriate filesystem support, for a
sysadmin to tell what is affected by the mismatch, and whether
it should be ignored (if it's inside a swap partition, for
instance).

We ratelimit to prevent log flooding: if there are so many
mismatches that ratelimiting is necessary, the individual messages
are relatively unlikely to be important (either the machine is
swapping like crazy or something is very wrong with the disk).

Signed-off-by: Nick Alcock 
Signed-off-by: Shaohua Li

md: uuid debug statement now in processor byte order.

2017-05-24T22:58:43+00:00

Previously, the uuid debug statements were printed in little-endian
format, which wasn't consistent in machines that might not be in
little-endian byte order. With this change, the output will be
consistent for all machines with different byte-ordering.

Signed-off-by: Kyungchan Koh 
Signed-off-by: Shaohua Li

dm ioctl: restore __GFP_HIGH in copy_params()

2017-05-22T23:30:03+00:00

Commit d224e9381897 ("drivers/md/dm-ioctl.c: use kvmalloc rather than
opencoded variant") left out the __GFP_HIGH flag when converting from
__vmalloc to kvmalloc.  This can cause the DM ioctl to fail in some low
memory situations where it wouldn't have failed earlier.  Add __GFP_HIGH
back to avoid any potential regression.

Fixes: d224e9381897 ("drivers/md/dm-ioctl.c: use kvmalloc rather than opencoded variant")
Signed-off-by: Junaid Shahid 
Signed-off-by: Mikulas Patocka 
Acked-by: David Rientjes 
Signed-off-by: Mike Snitzer

dm integrity: use kvmalloc() instead of dm_integrity_kvmalloc()

2017-05-22T18:09:52+00:00

Signed-off-by: Mikulas Patocka 
Signed-off-by: Mike Snitzer

dm verity: fix no salt use case

2017-05-22T17:49:03+00:00

DM-Verity has an (undocumented) mode where no salt is used.  This was
never handled directly by the DM-Verity code, instead working due to the
fact that calling crypto_shash_update() with a zero length data is an
implicit noop.

This is no longer the case now that we have switched to
crypto_ahash_update().  Fix the issue by introducing explicit handling
of the no salt use case to DM-Verity.

Signed-off-by: Gilad Ben-Yossef 
Reported-by: Marian Csontos 
Fixes: d1ac3ff ("dm verity: switch to using asynchronous hash crypto API")
Tested-by: Milan Broz 
Signed-off-by: Mike Snitzer

md-cluster: fix potential lock issue in add_new_disk

2017-05-22T03:37:09+00:00

The add_new_disk returns with communication locked if
__sendmsg returns failure, fix it with call unlock_comm
before return.

Reported-by: Dan Carpenter 
CC: Goldwyn Rodrigues 
Signed-off-by: Guoqing Jiang 
Signed-off-by: Shaohua Li

Merge tag 'md/4.12-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/shli/md

2017-05-18T19:04:41+00:00

Pull MD fixes from Shaohua Li:

 - Several bug fixes for raid5-cache from Song Liu, mainly handle
   journal disk error

 - Fix bad block handling in choosing raid1 disk from Tomasz Majchrzak

 - Simplify external metadata array sysfs handling from Artur
   Paszkiewicz

 - Optimize raid0 discard handling from me, now raid0 will dispatch
   large discard IO directly to underlayer disks.

* tag 'md/4.12-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/shli/md:
  raid1: prefer disk without bad blocks
  md/r5cache: handle sync with data in write back cache
  md/r5cache: gracefully handle journal device errors for writeback mode
  md/raid1/10: avoid unnecessary locking
  md/raid5-cache: in r5l_do_submit_io(), submit io->split_bio first
  md/md0: optimize raid0 discard handling
  md: don't return -EAGAIN in md_allow_write for external metadata arrays
  md/raid5: make use of spin_lock_irq over local_irq_disable + spin_lock