[SCSI] zfcp: Fix abort handler for completions in progress
authorChristof Schmitt <christof.schmitt@de.ibm.com>
Fri, 17 Apr 2009 13:08:11 +0000 (15:08 +0200)
committerJames Bottomley <James.Bottomley@HansenPartnership.com>
Mon, 27 Apr 2009 15:07:33 +0000 (10:07 -0500)
When the abort handler cannot find a pending FSF request, the request
completion could just be running. This means we cannot return SUCCESS,
since this would lead to call to scsi_done after exiting the SCSI
error handler which is not allowed.

Reviewed-by: Martin Petermann <martin@linux.vnet.ibm.com>
Signed-off-by: Christof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
drivers/s390/scsi/zfcp_scsi.c

index 5b11386d70bbc77d1e57289fef3be21c41cfd5a8..e7d065209065324f80e7b74f438ca23934662e26 100644 (file)
@@ -171,7 +171,7 @@ static int zfcp_scsi_eh_abort_handler(struct scsi_cmnd *scpnt)
                write_unlock_irqrestore(&adapter->abort_lock, flags);
                zfcp_scsi_dbf_event_abort("lte1", adapter, scpnt, NULL,
                                          old_req_id);
-               return SUCCESS;
+               return FAILED; /* completion could be in progress */
        }
        old_req->data = NULL;