1 files changed, 54 insertions, 22 deletions

diff --git a/drivers/net/ethernet/meta/fbnic/fbnic_pci.c b/drivers/net/ethernet/meta/fbnic/fbnic_pci.c
index b70e4cadb37b..a7a6b4db8016 100644
--- a/drivers/net/ethernet/meta/fbnic/fbnic_pci.c
+++ b/drivers/net/ethernet/meta/fbnic/fbnic_pci.c
@@ -118,14 +118,12 @@ static void fbnic_service_task_start(struct fbnic_net *fbn)
 	struct fbnic_dev *fbd = fbn->fbd;
 
 	schedule_delayed_work(&fbd->service_task, HZ);
-	phylink_resume(fbn->phylink);
 }
 
 static void fbnic_service_task_stop(struct fbnic_net *fbn)
 {
 	struct fbnic_dev *fbd = fbn->fbd;
 
-	phylink_suspend(fbn->phylink, fbnic_bmc_present(fbd));
 	cancel_delayed_work(&fbd->service_task);
 }
 
@@ -137,7 +135,7 @@ void fbnic_up(struct fbnic_net *fbn)
 
 	fbnic_rss_reinit_hw(fbn->fbd, fbn);
 
-	__fbnic_set_rx_mode(fbn->netdev);
+	__fbnic_set_rx_mode(fbn->fbd);
 
 	/* Enable Tx/Rx processing */
 	fbnic_napi_enable(fbn);
@@ -154,7 +152,7 @@ void fbnic_down_noidle(struct fbnic_net *fbn)
 	fbnic_napi_disable(fbn);
 	netif_tx_disable(fbn->netdev);
 
-	fbnic_clear_rx_mode(fbn->netdev);
+	fbnic_clear_rx_mode(fbn->fbd);
 	fbnic_clear_rules(fbn->fbd);
 	fbnic_rss_disable_hw(fbn->fbd);
 	fbnic_disable(fbn);
@@ -169,6 +167,20 @@ void fbnic_down(struct fbnic_net *fbn)
 	fbnic_flush(fbn);
 }
 
+static int fbnic_fw_config_after_crash(struct fbnic_dev *fbd)
+{
+	if (fbnic_fw_xmit_ownership_msg(fbd, true)) {
+		dev_err(fbd->dev, "NIC failed to take ownership\n");
+
+		return -1;
+	}
+
+	fbnic_rpc_reset_valid_entries(fbd);
+	__fbnic_set_rx_mode(fbd);
+
+	return 0;
+}
+
 static void fbnic_health_check(struct fbnic_dev *fbd)
 {
 	struct fbnic_fw_mbx *tx_mbx = &fbd->mbx[FBNIC_IPC_MBX_TX_IDX];
@@ -184,13 +196,11 @@ static void fbnic_health_check(struct fbnic_dev *fbd)
 	if (tx_mbx->head != tx_mbx->tail)
 		return;
 
-	/* TBD: Need to add a more thorough recovery here.
-	 *	Specifically I need to verify what all the firmware will have
-	 *	changed since we had setup and it rebooted. May just need to
-	 *	perform a down/up. For now we will just reclaim ownership so
-	 *	the heartbeat can catch the next fault.
-	 */
-	fbnic_fw_xmit_ownership_msg(fbd, true);
+	fbnic_devlink_fw_report(fbd, "Firmware crashed detected!");
+	fbnic_devlink_otp_check(fbd, "error detected after firmware recovery");
+
+	if (fbnic_fw_config_after_crash(fbd))
+		dev_err(fbd->dev, "Firmware recovery failed after crash\n");
 }
 
 static void fbnic_service_task(struct work_struct *work)
@@ -206,8 +216,13 @@ static void fbnic_service_task(struct work_struct *work)
 
 	fbnic_health_check(fbd);
 
-	if (netif_carrier_ok(fbd->netdev))
+	fbnic_bmc_rpc_check(fbd);
+
+	if (netif_carrier_ok(fbd->netdev)) {
+		netdev_lock(fbd->netdev);
 		fbnic_napi_depletion_check(fbd->netdev);
+		netdev_unlock(fbd->netdev);
+	}
 
 	if (netif_running(fbd->netdev))
 		schedule_delayed_work(&fbd->service_task, HZ);
@@ -266,6 +281,10 @@ static int fbnic_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 		return -ENOMEM;
 	}
 
+	err = fbnic_devlink_health_create(fbd);
+	if (err)
+		goto free_fbd;
+
 	/* Populate driver with hardware-specific info and handlers */
 	fbd->max_num_queues = info->max_num_queues;
 
@@ -276,7 +295,7 @@ static int fbnic_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 
 	err = fbnic_alloc_irqs(fbd);
 	if (err)
-		goto free_fbd;
+		goto err_destroy_health;
 
 	err = fbnic_mac_init(fbd);
 	if (err) {
@@ -303,11 +322,11 @@ static int fbnic_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 			 err);
 
 	fbnic_devlink_register(fbd);
+	fbnic_devlink_otp_check(fbd, "error detected during probe");
 	fbnic_dbg_fbd_init(fbd);
-	spin_lock_init(&fbd->hw_stats_lock);
 
 	/* Capture snapshot of hardware stats so netdev can calculate delta */
-	fbnic_reset_hw_stats(fbd);
+	fbnic_init_hw_stats(fbd);
 
 	fbnic_hwmon_register(fbd);
 
@@ -346,6 +365,8 @@ init_failure_mode:
 	return 0;
 free_irqs:
 	fbnic_free_irqs(fbd);
+err_destroy_health:
+	fbnic_devlink_health_destroy(fbd);
 free_fbd:
 	fbnic_devlink_free(fbd);
 
@@ -380,6 +401,7 @@ static void fbnic_remove(struct pci_dev *pdev)
 	fbnic_fw_free_mbx(fbd);
 	fbnic_free_irqs(fbd);
 
+	fbnic_devlink_health_destroy(fbd);
 	fbnic_devlink_free(fbd);
 }
 
@@ -392,12 +414,14 @@ static int fbnic_pm_suspend(struct device *dev)
 		goto null_uc_addr;
 
 	rtnl_lock();
+	netdev_lock(netdev);
 
 	netif_device_detach(netdev);
 
 	if (netif_running(netdev))
 		netdev->netdev_ops->ndo_stop(netdev);
 
+	netdev_unlock(netdev);
 	rtnl_unlock();
 
 null_uc_addr:
@@ -443,16 +467,18 @@ static int __fbnic_pm_resume(struct device *dev)
 
 	/* Re-enable mailbox */
 	err = fbnic_fw_request_mbx(fbd);
+	devl_unlock(priv_to_devlink(fbd));
 	if (err)
 		goto err_free_irqs;
 
-	devl_unlock(priv_to_devlink(fbd));
-
 	/* Only send log history if log buffer is empty to prevent duplicate
 	 * log entries.
 	 */
 	fbnic_fw_log_enable(fbd, list_empty(&fbd->fw_log.entries));
 
+	/* Since the FW should be up, check if it reported OTP errors */
+	fbnic_devlink_otp_check(fbd, "error detected after PM resume");
+
 	/* No netdev means there isn't a network interface to bring up */
 	if (fbnic_init_failure(fbd))
 		return 0;
@@ -463,21 +489,23 @@ static int __fbnic_pm_resume(struct device *dev)
 	fbnic_reset_queues(fbn, fbn->num_tx_queues, fbn->num_rx_queues);
 
 	rtnl_lock();
+	netdev_lock(netdev);
 
-	if (netif_running(netdev)) {
+	if (netif_running(netdev))
 		err = __fbnic_open(fbn);
-		if (err)
-			goto err_free_mbx;
-	}
 
+	netdev_unlock(netdev);
 	rtnl_unlock();
+	if (err)
+		goto err_free_mbx;
 
 	return 0;
 err_free_mbx:
 	fbnic_fw_log_disable(fbd);
 
-	rtnl_unlock();
+	devl_lock(priv_to_devlink(fbd));
 	fbnic_fw_free_mbx(fbd);
+	devl_unlock(priv_to_devlink(fbd));
 err_free_irqs:
 	fbnic_free_irqs(fbd);
 err_invalidate_uc_addr:
@@ -492,6 +520,10 @@ static void __fbnic_pm_attach(struct device *dev)
 	struct net_device *netdev = fbd->netdev;
 	struct fbnic_net *fbn;
 
+	rtnl_lock();
+	fbnic_reset_hw_stats(fbd);
+	rtnl_unlock();
+
 	if (fbnic_init_failure(fbd))
 		return;