scsi: qla2xxx: Move heartbeat handling from DPC thread to workqueue

DPC thread gets restricted due to a no-op mailbox, which is a blocking call
and has a high execution frequency. To free up the DPC thread we move no-op
handling to the workqueue. Also, modified qla_do_heartbeat() to send no-op
MBC if we don’t have any active interrupts, but there are still I/Os
outstanding with firmware.

Link: https://lore.kernel.org/r/20210908164622.19240-9-njavali@marvell.com
Fixes: d94d8158e184 ("scsi: qla2xxx: Add heartbeat check")
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Manish Rangankar <mrangankar@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

1 files changed, 36 insertions, 42 deletions

diff --git a/drivers/scsi/qla2xxx/qla_os.c b/drivers/scsi/qla2xxx/qla_os.c
index a1e861ecfc01..0c446438dcd3 100644
--- a/drivers/scsi/qla2xxx/qla_os.c
+++ b/drivers/scsi/qla2xxx/qla_os.c
@@ -2794,6 +2794,16 @@ qla2xxx_scan_finished(struct Scsi_Host *shost, unsigned long time)
 	return atomic_read(&vha->loop_state) == LOOP_READY;
 }
 
+static void qla_heartbeat_work_fn(struct work_struct *work)
+{
+	struct qla_hw_data *ha = container_of(work,
+		struct qla_hw_data, heartbeat_work);
+	struct scsi_qla_host *base_vha = pci_get_drvdata(ha->pdev);
+
+	if (!ha->flags.mbox_busy && base_vha->flags.init_done)
+		qla_no_op_mb(base_vha);
+}
+
 static void qla2x00_iocb_work_fn(struct work_struct *work)
 {
 	struct scsi_qla_host *vha = container_of(work,
@@ -3232,6 +3242,7 @@ qla2x00_probe_one(struct pci_dev *pdev, const struct pci_device_id *id)
 	    host->transportt, sht->vendor_id);
 
 	INIT_WORK(&base_vha->iocb_work, qla2x00_iocb_work_fn);
+	INIT_WORK(&ha->heartbeat_work, qla_heartbeat_work_fn);
 
 	/* Set up the irqs */
 	ret = qla2x00_request_irqs(ha, rsp);
@@ -7118,17 +7129,6 @@ intr_on_check:
 			qla2x00_lip_reset(base_vha);
 		}
 
-		if (test_bit(HEARTBEAT_CHK, &base_vha->dpc_flags)) {
-			/*
-			 * if there is a mb in progress then that's
-			 * enough of a check to see if fw is still ticking.
-			 */
-			if (!ha->flags.mbox_busy && base_vha->flags.init_done)
-				qla_no_op_mb(base_vha);
-
-			clear_bit(HEARTBEAT_CHK, &base_vha->dpc_flags);
-		}
-
 		ha->dpc_active = 0;
 end_loop:
 		set_current_state(TASK_INTERRUPTIBLE);
@@ -7187,57 +7187,51 @@ qla2x00_rst_aen(scsi_qla_host_t *vha)
 
 static bool qla_do_heartbeat(struct scsi_qla_host *vha)
 {
-	u64 cmd_cnt, prev_cmd_cnt;
-	bool do_hb = false;
 	struct qla_hw_data *ha = vha->hw;
-	int i;
+	u32 cmpl_cnt;
+	u16 i;
+	bool do_heartbeat = false;
 
-	/* if cmds are still pending down in fw, then do hb */
-	if (ha->base_qpair->cmd_cnt != ha->base_qpair->cmd_completion_cnt) {
-		do_hb = true;
+	/*
+	 * Allow do_heartbeat only if we don’t have any active interrupts,
+	 * but there are still IOs outstanding with firmware.
+	 */
+	cmpl_cnt = ha->base_qpair->cmd_completion_cnt;
+	if (cmpl_cnt == ha->base_qpair->prev_completion_cnt &&
+	    cmpl_cnt != ha->base_qpair->cmd_cnt) {
+		do_heartbeat = true;
 		goto skip;
 	}
+	ha->base_qpair->prev_completion_cnt = cmpl_cnt;
 
 	for (i = 0; i < ha->max_qpairs; i++) {
-		if (ha->queue_pair_map[i] &&
-		    ha->queue_pair_map[i]->cmd_cnt !=
-		    ha->queue_pair_map[i]->cmd_completion_cnt) {
-			do_hb = true;
-			break;
+		if (ha->queue_pair_map[i]) {
+			cmpl_cnt = ha->queue_pair_map[i]->cmd_completion_cnt;
+			if (cmpl_cnt == ha->queue_pair_map[i]->prev_completion_cnt &&
+			    cmpl_cnt != ha->queue_pair_map[i]->cmd_cnt) {
+				do_heartbeat = true;
+				break;
+			}
+			ha->queue_pair_map[i]->prev_completion_cnt = cmpl_cnt;
 		}
 	}
 
 skip:
-	prev_cmd_cnt = ha->prev_cmd_cnt;
-	cmd_cnt = ha->base_qpair->cmd_cnt;
-	for (i = 0; i < ha->max_qpairs; i++) {
-		if (ha->queue_pair_map[i])
-			cmd_cnt += ha->queue_pair_map[i]->cmd_cnt;
-	}
-	ha->prev_cmd_cnt = cmd_cnt;
-
-	if (!do_hb && ((cmd_cnt - prev_cmd_cnt) > 50))
-		/*
-		 * IOs are completing before periodic hb check.
-		 * IOs seems to be running, do hb for sanity check.
-		 */
-		do_hb = true;
-
-	return do_hb;
+	return do_heartbeat;
 }
 
 static void qla_heart_beat(struct scsi_qla_host *vha)
 {
+	struct qla_hw_data *ha = vha->hw;
+
 	if (vha->vp_idx)
 		return;
 
 	if (vha->hw->flags.eeh_busy || qla2x00_chip_is_down(vha))
 		return;
 
-	if (qla_do_heartbeat(vha)) {
-		set_bit(HEARTBEAT_CHK, &vha->dpc_flags);
-		qla2xxx_wake_dpc(vha);
-	}
+	if (qla_do_heartbeat(vha))
+		queue_work(ha->wq, &ha->heartbeat_work);
 }
 
 /**************************************************************************