scsi: libfc: quarantine timed out xids

When a sequence times out we have no idea what happened to the
frame. And we do not know if we will ever receive the frame.
Hence we cannot re-use the xid as we would risk data corruption
if the xid had been re-used and the timed out frame would be
received after that.
So we need to quarantine the xid until the lport is reset.
Yes, I know this will (eventually) deplete the xid pool.
But for now it's the safest method.

Signed-off-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

1 files changed, 7 insertions, 6 deletions

diff --git a/drivers/scsi/libfc/fc_fcp.c b/drivers/scsi/libfc/fc_fcp.c
index f7700cccf793..780d9f09a267 100644
--- a/drivers/scsi/libfc/fc_fcp.c
+++ b/drivers/scsi/libfc/fc_fcp.c
@@ -1529,13 +1529,14 @@ static void fc_fcp_rec_resp(struct fc_seq *seq, struct fc_frame *fp, void *arg)
 				   fsp->rport->port_id, rjt->er_reason,
 				   rjt->er_explan);
 			/*
-			 * If no data transfer, the command frame got dropped
-			 * so we just retry.  If data was transferred, we
-			 * lost the response but the target has no record,
-			 * so we abort and retry.
+			 * If response got lost or is stuck in the
+			 * queue somewhere we have no idea if and when
+			 * the response will be received. So quarantine
+			 * the xid and retry the command.
 			 */
-			if (rjt->er_explan == ELS_EXPL_OXID_RXID &&
-			    fsp->xfer_len == 0) {
+			if (rjt->er_explan == ELS_EXPL_OXID_RXID) {
+				struct fc_exch *ep = fc_seq_exch(fsp->seq_ptr);
+				ep->state |= FC_EX_QUARANTINE;
 				fsp->state |= FC_SRB_ABORTED;
 				fc_fcp_retry_cmd(fsp, FC_TRANS_RESET);
 				break;