md/raid1: extend spinlock to protect raid1_end_read_request against inconsistencies
authorNeilBrown <neilb@suse.com>
Mon, 27 Jul 2015 01:48:52 +0000 (11:48 +1000)
committerNeilBrown <neilb@suse.com>
Mon, 3 Aug 2015 02:29:42 +0000 (12:29 +1000)
commit423f04d63cf421ea436bcc5be02543d549ce4b28
tree0d98d3b3de62c7689fbe58a6229c7d760d998c55
parent74d33293e467df61de1b1d8b2fbe29e550dec33b
md/raid1: extend spinlock to protect raid1_end_read_request against inconsistencies

raid1_end_read_request() assumes that the In_sync bits are consistent
with the ->degaded count.
raid1_spare_active updates the In_sync bit before the ->degraded count
and so exposes an inconsistency, as does error()
So extend the spinlock in raid1_spare_active() and error() to hide those
inconsistencies.

This should probably be part of
  Commit: 34cab6f42003 ("md/raid1: fix test for 'was read error from
  last working device'.")
as it addresses the same issue.  It fixes the same bug and should go
to -stable for same reasons.

Fixes: 76073054c95b ("md/raid1: clean up read_balance.")
Cc: stable@vger.kernel.org (v3.0+)
Signed-off-by: NeilBrown <neilb@suse.com>
drivers/md/raid1.c