[SCSI] Fix command result state propagation
authorAlan Stern <stern@rowland.harvard.edu>
Fri, 28 Mar 2014 17:51:15 +0000 (10:51 -0700)
committerJames Bottomley <JBottomley@Parallels.com>
Mon, 21 Apr 2014 21:27:26 +0000 (14:27 -0700)
commit644373a4219add42123df69c8b7ce6a918475ccd
treebdf8825070f03d622ff3fa715f3ad5f19513bb4f
parentd555a2abf3481f81303d835046a5ec2c4fb3ca8e
[SCSI] Fix command result state propagation

We're seeing a case where the contents of scmd->result isn't being reset after
a SCSI command encounters an error, is resubmitted, times out and then gets
handled.  The error handler acts on the stale result of the previous error
instead of the timeout.  Fix this by properly zeroing the scmd->status before
the command is resubmitted.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
drivers/scsi/scsi_error.c
drivers/scsi/scsi_lib.c