scsi: mpi3mr: Add support for internal watchdog thread

The watchdog thread is the driver's internal thread which does a few things
such as detecting firmware faults, resetting the controller, performing
timestamp sync, etc.

Link: https://lore.kernel.org/r/20210520152545.2710479-6-kashyap.desai@broadcom.com
Cc: sathya.prakash@broadcom.com
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Kashyap Desai <kashyap.desai@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

3 files changed, 140 insertions, 1 deletions

diff --git a/drivers/scsi/mpi3mr/mpi3mr.h b/drivers/scsi/mpi3mr/mpi3mr.h
index 4b3b76c7c6ff..1151a9435f3f 100644
--- a/drivers/scsi/mpi3mr/mpi3mr.h
+++ b/drivers/scsi/mpi3mr/mpi3mr.h
@@ -481,6 +481,10 @@ struct scmd_priv {
  * @sense_buf_q_dma: Sense buffer queue DMA address
  * @sbq_lock: Sense buffer queue lock
  * @sbq_host_index: Sense buffer queuehost index
+ * @watchdog_work_q_name: Fault watchdog worker thread name
+ * @watchdog_work_q: Fault watchdog worker thread
+ * @watchdog_work: Fault watchdog work
+ * @watchdog_lock: Fault watchdog lock
  * @is_driver_loading: Is driver still loading
  * @scan_started: Async scan started
  * @scan_failed: Asycn scan failed
@@ -494,6 +498,7 @@ struct scmd_priv {
  * @chain_buf_lock: Chain buffer list lock
  * @reset_in_progress: Reset in progress flag
  * @unrecoverable: Controller unrecoverable flag
+ * @diagsave_timeout: Diagnostic information save timeout
  * @logging_level: Controller debug logging level
  * @current_event: Firmware event currently in process
  * @driver_info: Driver, Kernel, OS information to firmware
@@ -575,6 +580,11 @@ struct mpi3mr_ioc {
 	spinlock_t sbq_lock;
 	u32 sbq_host_index;
 
+	char watchdog_work_q_name[20];
+	struct workqueue_struct *watchdog_work_q;
+	struct delayed_work watchdog_work;
+	spinlock_t watchdog_lock;
+
 	u8 is_driver_loading;
 	u8 scan_started;
 	u16 scan_failed;
@@ -592,6 +602,7 @@ struct mpi3mr_ioc {
 	u8 reset_in_progress;
 	u8 unrecoverable;
 
+	u16 diagsave_timeout;
 	int logging_level;
 
 	struct mpi3mr_fwevt *current_event;
diff --git a/drivers/scsi/mpi3mr/mpi3mr_fw.c b/drivers/scsi/mpi3mr/mpi3mr_fw.c
index dad7ef191586..daed0a0ed243 100644
--- a/drivers/scsi/mpi3mr/mpi3mr_fw.c
+++ b/drivers/scsi/mpi3mr/mpi3mr_fw.c
@@ -1464,6 +1464,129 @@ out:
 }
 
 /**
+ * mpi3mr_watchdog_work - watchdog thread to monitor faults
+ * @work: work struct
+ *
+ * Watch dog work periodically executed (1 second interval) to
+ * monitor firmware fault and to issue periodic timer sync to
+ * the firmware.
+ *
+ * Return: Nothing.
+ */
+static void mpi3mr_watchdog_work(struct work_struct *work)
+{
+	struct mpi3mr_ioc *mrioc =
+	    container_of(work, struct mpi3mr_ioc, watchdog_work.work);
+	unsigned long flags;
+	enum mpi3mr_iocstate ioc_state;
+	u32 fault, host_diagnostic;
+
+	/*Check for fault state every one second and issue Soft reset*/
+	ioc_state = mpi3mr_get_iocstate(mrioc);
+	if (ioc_state == MRIOC_STATE_FAULT) {
+		fault = readl(&mrioc->sysif_regs->fault) &
+		    MPI3_SYSIF_FAULT_CODE_MASK;
+		host_diagnostic = readl(&mrioc->sysif_regs->host_diagnostic);
+		if (host_diagnostic & MPI3_SYSIF_HOST_DIAG_SAVE_IN_PROGRESS) {
+			if (!mrioc->diagsave_timeout) {
+				mpi3mr_print_fault_info(mrioc);
+				ioc_warn(mrioc, "Diag save in progress\n");
+			}
+			if ((mrioc->diagsave_timeout++) <=
+			    MPI3_SYSIF_DIAG_SAVE_TIMEOUT)
+				goto schedule_work;
+		} else
+			mpi3mr_print_fault_info(mrioc);
+		mrioc->diagsave_timeout = 0;
+
+		if (fault == MPI3_SYSIF_FAULT_CODE_FACTORY_RESET) {
+			ioc_info(mrioc,
+			    "Factory Reset fault occurred marking controller as unrecoverable"
+			    );
+			mrioc->unrecoverable = 1;
+			goto out;
+		}
+
+		if ((fault == MPI3_SYSIF_FAULT_CODE_DIAG_FAULT_RESET) ||
+		    (fault == MPI3_SYSIF_FAULT_CODE_SOFT_RESET_IN_PROGRESS) ||
+		    (mrioc->reset_in_progress))
+			goto out;
+		if (fault == MPI3_SYSIF_FAULT_CODE_CI_ACTIVATION_RESET)
+			mpi3mr_soft_reset_handler(mrioc,
+			    MPI3MR_RESET_FROM_CIACTIV_FAULT, 0);
+		else
+			mpi3mr_soft_reset_handler(mrioc,
+			    MPI3MR_RESET_FROM_FAULT_WATCH, 0);
+	}
+
+schedule_work:
+	spin_lock_irqsave(&mrioc->watchdog_lock, flags);
+	if (mrioc->watchdog_work_q)
+		queue_delayed_work(mrioc->watchdog_work_q,
+		    &mrioc->watchdog_work,
+		    msecs_to_jiffies(MPI3MR_WATCHDOG_INTERVAL));
+	spin_unlock_irqrestore(&mrioc->watchdog_lock, flags);
+out:
+	return;
+}
+
+/**
+ * mpi3mr_start_watchdog - Start watchdog
+ * @mrioc: Adapter instance reference
+ *
+ * Create and start the watchdog thread to monitor controller
+ * faults.
+ *
+ * Return: Nothing.
+ */
+void mpi3mr_start_watchdog(struct mpi3mr_ioc *mrioc)
+{
+	if (mrioc->watchdog_work_q)
+		return;
+
+	INIT_DELAYED_WORK(&mrioc->watchdog_work, mpi3mr_watchdog_work);
+	snprintf(mrioc->watchdog_work_q_name,
+	    sizeof(mrioc->watchdog_work_q_name), "watchdog_%s%d", mrioc->name,
+	    mrioc->id);
+	mrioc->watchdog_work_q =
+	    create_singlethread_workqueue(mrioc->watchdog_work_q_name);
+	if (!mrioc->watchdog_work_q) {
+		ioc_err(mrioc, "%s: failed (line=%d)\n", __func__, __LINE__);
+		return;
+	}
+
+	if (mrioc->watchdog_work_q)
+		queue_delayed_work(mrioc->watchdog_work_q,
+		    &mrioc->watchdog_work,
+		    msecs_to_jiffies(MPI3MR_WATCHDOG_INTERVAL));
+}
+
+/**
+ * mpi3mr_stop_watchdog - Stop watchdog
+ * @mrioc: Adapter instance reference
+ *
+ * Stop the watchdog thread created to monitor controller
+ * faults.
+ *
+ * Return: Nothing.
+ */
+void mpi3mr_stop_watchdog(struct mpi3mr_ioc *mrioc)
+{
+	unsigned long flags;
+	struct workqueue_struct *wq;
+
+	spin_lock_irqsave(&mrioc->watchdog_lock, flags);
+	wq = mrioc->watchdog_work_q;
+	mrioc->watchdog_work_q = NULL;
+	spin_unlock_irqrestore(&mrioc->watchdog_lock, flags);
+	if (wq) {
+		if (!cancel_delayed_work_sync(&mrioc->watchdog_work))
+			flush_workqueue(wq);
+		destroy_workqueue(wq);
+	}
+}
+
+/**
  * mpi3mr_setup_admin_qpair - Setup admin queue pair
  * @mrioc: Adapter instance reference
  *
@@ -2609,6 +2732,8 @@ void mpi3mr_cleanup_ioc(struct mpi3mr_ioc *mrioc)
 {
 	enum mpi3mr_iocstate ioc_state;
 
+	mpi3mr_stop_watchdog(mrioc);
+
 	mpi3mr_ioc_disable_intr(mrioc);
 
 	ioc_state = mpi3mr_get_iocstate(mrioc);
diff --git a/drivers/scsi/mpi3mr/mpi3mr_os.c b/drivers/scsi/mpi3mr/mpi3mr_os.c
index 5afcd55684cf..26daf764c552 100644
--- a/drivers/scsi/mpi3mr/mpi3mr_os.c
+++ b/drivers/scsi/mpi3mr/mpi3mr_os.c
@@ -272,7 +272,7 @@ void mpi3mr_process_op_reply_desc(struct mpi3mr_ioc *mrioc,
 	case MPI3_IOCSTATUS_SUCCESS:
 		scmd->result = (DID_OK << 16) | scsi_status;
 		if ((scsi_state & (MPI3_SCSI_STATE_NO_SCSI_STATUS)) ||
-			(sense_state == MPI3_SCSI_STATE_SENSE_FAILED) ||
+		    (sense_state == MPI3_SCSI_STATE_SENSE_FAILED) ||
 			(sense_state == MPI3_SCSI_STATE_SENSE_BUFF_Q_EMPTY))
 			scmd->result = DID_SOFT_ERROR << 16;
 		else if (scsi_state & MPI3_SCSI_STATE_TERMINATED)
@@ -559,6 +559,7 @@ static int mpi3mr_scan_finished(struct Scsi_Host *shost,
 	if (mrioc->scan_started)
 		return 0;
 	ioc_info(mrioc, "%s :port enable: SUCCESS\n", __func__);
+	mpi3mr_start_watchdog(mrioc);
 	mrioc->is_driver_loading = 0;
 
 	return 1;
@@ -843,9 +844,11 @@ mpi3mr_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	spin_lock_init(&mrioc->admin_req_lock);
 	spin_lock_init(&mrioc->reply_free_queue_lock);
 	spin_lock_init(&mrioc->sbq_lock);
+	spin_lock_init(&mrioc->watchdog_lock);
 	spin_lock_init(&mrioc->chain_buf_lock);
 
 	mpi3mr_init_drv_cmd(&mrioc->init_cmds, MPI3MR_HOSTTAG_INITCMDS);
+
 	if (pdev->revision)
 		mrioc->enable_segqueue = true;