[SCSI] qla2xxx: Use 'proper' DID_* status code for dropped-frame scenarios.
authorAndrew Vasquez <andrew.vasquez@qlogic.com>
Wed, 3 Jun 2009 16:55:18 +0000 (09:55 -0700)
committerJames Bottomley <James.Bottomley@HansenPartnership.com>
Mon, 8 Jun 2009 19:46:43 +0000 (14:46 -0500)
The SCSI-midlayer's fast-fail codes consider an DID_ERROR status
as a driver-error and the failed I/O would then be retried in the
midlayer without being fast-failed to dm-multipath.  DID_BUS_BUSY
status returns would induce unneeded path-failures events being
propagated to the DM/MD.

Signed-off-by: Andrew Vasquez <andrew.vasquez@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
drivers/scsi/qla2xxx/qla_isr.c

index eb35d2050f7af7d41746e07b302d1de8d8e65115..c8d0a176fea4a1287853716a7dd1d6f20c910060 100644 (file)
@@ -1211,7 +1211,7 @@ qla2x00_status_entry(scsi_qla_host_t *vha, struct rsp_que *rsp, void *pkt)
                        /*
                         * If RISC reports underrun and target does not report
                         * it then we must have a lost frame, so tell upper
-                        * layer to retry it by reporting a bus busy.
+                        * layer to retry it by reporting an error.
                         */
                        if (!(scsi_status & SS_RESIDUAL_UNDER)) {
                                DEBUG2(printk("scsi(%ld:%d:%d:%d) Dropped "
@@ -1221,7 +1221,7 @@ qla2x00_status_entry(scsi_qla_host_t *vha, struct rsp_que *rsp, void *pkt)
                                        cp->device->id, cp->device->lun, resid,
                                        scsi_bufflen(cp)));
 
-                               cp->result = DID_BUS_BUSY << 16;
+                               cp->result = DID_ERROR << 16;
                                break;
                        }