scsi_lib: correctly retry failed zero length REQ_TYPE_FS commands
authorJames Bottomley <James.Bottomley@HansenPartnership.com>
Fri, 13 May 2016 19:04:06 +0000 (12:04 -0700)
committerMartin K. Petersen <martin.petersen@oracle.com>
Sun, 22 May 2016 18:52:45 +0000 (14:52 -0400)
commita621bac3044ed6f7ec5fa0326491b2d4838bfa93
treee8c0c7dc229c83c07bf6edb61733ff1f68ffb240
parent97f00905ec98472050d65c46629237b299f29035
scsi_lib: correctly retry failed zero length REQ_TYPE_FS commands

When SCSI was written, all commands coming from the filesystem
(REQ_TYPE_FS commands) had data.  This meant that our signal for needing
to complete the command was the number of bytes completed being equal to
the number of bytes in the request.  Unfortunately, with the advent of
flush barriers, we can now get zero length REQ_TYPE_FS commands, which
confuse this logic because they satisfy the condition every time.  This
means they never get retried even for retryable conditions, like UNIT
ATTENTION because we complete them early assuming they're done.  Fix
this by special casing the early completion condition to recognise zero
length commands with errors and let them drop through to the retry code.

Cc: stable@vger.kernel.org
Reported-by: Sebastian Parschauer <s.parschauer@gmx.de>
Signed-off-by: James E.J. Bottomley <jejb@linux.vnet.ibm.com>
Tested-by: Jack Wang <jinpu.wang@profitbricks.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
drivers/scsi/scsi_lib.c