Merge branch 'bnxt_en-updates'

Michael Chan says:

====================
bnxt_en: Updates.

This patch series contains mainly NVRAM related features.  More
NVRAM error checking and logging are added when installing firmware
packages.  A new devlink hw health report is now added to report
and diagnose NVRAM issues.  Other miscellaneous patches include
reporting correctly cards that don't support link pause, adding
an internal unknown link state, and avoiding unnecessary link
toggle during firmware reset.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/broadcom/bnxt/bnxt.c

@@ -2061,6 +2061,22 @@ static void bnxt_event_error_report(struct bnxt *bp, u32 data1, u32 data2)
 	case ASYNC_EVENT_CMPL_ERROR_REPORT_BASE_EVENT_DATA1_ERROR_TYPE_DOORBELL_DROP_THRESHOLD:
 		netdev_warn(bp->dev, "One or more MMIO doorbells dropped by the device!\n");
 		break;
+	case ASYNC_EVENT_CMPL_ERROR_REPORT_BASE_EVENT_DATA1_ERROR_TYPE_NVM: {
+		struct bnxt_hw_health *hw_health = &bp->hw_health;
+
+		hw_health->nvm_err_address = EVENT_DATA2_NVM_ERR_ADDR(data2);
+		if (EVENT_DATA1_NVM_ERR_TYPE_WRITE(data1)) {
+			hw_health->synd = BNXT_HW_STATUS_NVM_WRITE_ERR;
+			hw_health->nvm_write_errors++;
+		} else if (EVENT_DATA1_NVM_ERR_TYPE_ERASE(data1)) {
+			hw_health->synd = BNXT_HW_STATUS_NVM_ERASE_ERR;
+			hw_health->nvm_erase_errors++;
+		} else {
+			hw_health->synd = BNXT_HW_STATUS_NVM_UNKNOWN_ERR;
+		}
+		set_bit(BNXT_FW_NVM_ERR_SP_EVENT, &bp->sp_event);
+		break;
+	}
 	default:
 		netdev_err(bp->dev, "FW reported unknown error type %u\n",
 			   err_type);
@@ -9300,7 +9316,7 @@ void bnxt_tx_enable(struct bnxt *bp)
 	/* Make sure napi polls see @dev_state change */
 	synchronize_net();
 	netif_tx_wake_all_queues(bp->dev);
-	if (bp->link_info.link_up)
+	if (BNXT_LINK_IS_UP(bp))
 		netif_carrier_on(bp->dev);
 }
 
@@ -9330,7 +9346,7 @@ static char *bnxt_report_fec(struct bnxt_link_info *link_info)
 
 void bnxt_report_link(struct bnxt *bp)
 {
-	if (bp->link_info.link_up) {
+	if (BNXT_LINK_IS_UP(bp)) {
 		const char *signal = "";
 		const char *flow_ctrl;
 		const char *duplex;
@@ -9416,7 +9432,7 @@ static int bnxt_hwrm_phy_qcaps(struct bnxt *bp)
 	if (rc)
 		goto hwrm_phy_qcaps_exit;
 
-	bp->phy_flags = resp->flags;
+	bp->phy_flags = resp->flags | (le16_to_cpu(resp->flags2) << 8);
 	if (resp->flags & PORT_PHY_QCAPS_RESP_FLAGS_EEE_SUPPORTED) {
 		struct ethtool_eee *eee = &bp->eee;
 		u16 fw_speeds = le16_to_cpu(resp->supported_speeds_eee_mode);
@@ -9466,7 +9482,7 @@ int bnxt_update_link(struct bnxt *bp, bool chng_link_state)
 	struct bnxt_link_info *link_info = &bp->link_info;
 	struct hwrm_port_phy_qcfg_output *resp;
 	struct hwrm_port_phy_qcfg_input *req;
-	u8 link_up = link_info->link_up;
+	u8 link_state = link_info->link_state;
 	bool support_changed = false;
 	int rc;
 
@@ -9567,14 +9583,14 @@ int bnxt_update_link(struct bnxt *bp, bool chng_link_state)
 	/* TODO: need to add more logic to report VF link */
 	if (chng_link_state) {
 		if (link_info->phy_link_status == BNXT_LINK_LINK)
-			link_info->link_up = 1;
+			link_info->link_state = BNXT_LINK_STATE_UP;
 		else
-			link_info->link_up = 0;
-		if (link_up != link_info->link_up)
+			link_info->link_state = BNXT_LINK_STATE_DOWN;
+		if (link_state != link_info->link_state)
 			bnxt_report_link(bp);
 	} else {
-		/* alwasy link down if not require to update link state */
-		link_info->link_up = 0;
+		/* always link down if not require to update link state */
+		link_info->link_state = BNXT_LINK_STATE_DOWN;
 	}
 	hwrm_req_drop(bp, req);
 
@@ -9774,7 +9790,18 @@ static int bnxt_hwrm_shutdown_link(struct bnxt *bp)
 		return rc;
 
 	req->flags = cpu_to_le32(PORT_PHY_CFG_REQ_FLAGS_FORCE_LINK_DWN);
-	return hwrm_req_send(bp, req);
+	rc = hwrm_req_send(bp, req);
+	if (!rc) {
+		mutex_lock(&bp->link_lock);
+		/* Device is not obliged link down in certain scenarios, even
+		 * when forced. Setting the state unknown is consistent with
+		 * driver startup and will force link state to be reported
+		 * during subsequent open based on PORT_PHY_QCFG.
+		 */
+		bp->link_info.link_state = BNXT_LINK_STATE_UNKNOWN;
+		mutex_unlock(&bp->link_lock);
+	}
+	return rc;
 }
 
 static int bnxt_fw_reset_via_optee(struct bnxt *bp)
@@ -10205,7 +10232,7 @@ static int bnxt_update_phy_setting(struct bnxt *bp)
 	/* The last close may have shutdown the link, so need to call
 	 * PHY_CFG to bring it back up.
 	 */
-	if (!bp->link_info.link_up)
+	if (!BNXT_LINK_IS_UP(bp))
 		update_link = true;
 
 	if (!bnxt_eee_config_ok(bp))
@@ -11437,7 +11464,7 @@ static void bnxt_timer(struct timer_list *t)
 	if (bp->fw_cap & BNXT_FW_CAP_ERROR_RECOVERY)
 		bnxt_fw_health_check(bp);
 
-	if (bp->link_info.link_up && bp->stats_coal_ticks) {
+	if (BNXT_LINK_IS_UP(bp) && bp->stats_coal_ticks) {
 		set_bit(BNXT_PERIODIC_STATS_SP_EVENT, &bp->sp_event);
 		bnxt_queue_sp_work(bp);
 	}
@@ -11876,6 +11903,9 @@ static void bnxt_sp_task(struct work_struct *work)
 	if (test_and_clear_bit(BNXT_FW_ECHO_REQUEST_SP_EVENT, &bp->sp_event))
 		bnxt_fw_echo_reply(bp);
 
+	if (test_and_clear_bit(BNXT_FW_NVM_ERR_SP_EVENT, &bp->sp_event))
+		bnxt_devlink_health_hw_report(bp);
+
 	/* These functions below will clear BNXT_STATE_IN_SP_TASK.  They
 	 * must be the last functions to be called before exiting.
 	 */
@@ -12138,11 +12168,6 @@ int bnxt_fw_init_one(struct bnxt *bp)
 	if (rc)
 		return rc;
 
-	/* In case fw capabilities have changed, destroy the unneeded
-	 * reporters and create newly capable ones.
-	 */
-	bnxt_dl_fw_reporters_destroy(bp, false);
-	bnxt_dl_fw_reporters_create(bp);
 	bnxt_fw_init_one_p3(bp);
 	return 0;
 }
@@ -12971,7 +12996,7 @@ static void bnxt_remove_one(struct pci_dev *pdev)
 	cancel_delayed_work_sync(&bp->fw_reset_task);
 	bp->sp_event = 0;
 
-	bnxt_dl_fw_reporters_destroy(bp, true);
+	bnxt_dl_fw_reporters_destroy(bp);
 	bnxt_dl_unregister(bp);
 	bnxt_shutdown_tc(bp);
 

drivers/net/ethernet/broadcom/bnxt/bnxt.h

@@ -516,6 +516,21 @@ struct rx_tpa_end_cmp_ext {
 	  ASYNC_EVENT_CMPL_ERROR_REPORT_INVALID_SIGNAL_EVENT_DATA2_PIN_ID_MASK) >>\
 	 ASYNC_EVENT_CMPL_ERROR_REPORT_INVALID_SIGNAL_EVENT_DATA2_PIN_ID_SFT)
 
+#define EVENT_DATA2_NVM_ERR_ADDR(data2)					\
+	(((data2) &							\
+	  ASYNC_EVENT_CMPL_ERROR_REPORT_NVM_EVENT_DATA2_ERR_ADDR_MASK) >>\
+	 ASYNC_EVENT_CMPL_ERROR_REPORT_NVM_EVENT_DATA2_ERR_ADDR_SFT)
+
+#define EVENT_DATA1_NVM_ERR_TYPE_WRITE(data1)				\
+	(((data1) &							\
+	  ASYNC_EVENT_CMPL_ERROR_REPORT_NVM_EVENT_DATA1_NVM_ERR_TYPE_MASK) ==\
+	 ASYNC_EVENT_CMPL_ERROR_REPORT_NVM_EVENT_DATA1_NVM_ERR_TYPE_WRITE)
+
+#define EVENT_DATA1_NVM_ERR_TYPE_ERASE(data1)				\
+	(((data1) &							\
+	  ASYNC_EVENT_CMPL_ERROR_REPORT_NVM_EVENT_DATA1_NVM_ERR_TYPE_MASK) ==\
+	 ASYNC_EVENT_CMPL_ERROR_REPORT_NVM_EVENT_DATA1_NVM_ERR_TYPE_ERASE)
+
 struct nqe_cn {
 	__le16	type;
 	#define NQ_CN_TYPE_MASK           0x3fUL
@@ -1175,7 +1190,11 @@ struct bnxt_link_info {
 #define BNXT_PHY_STATE_ENABLED		0
 #define BNXT_PHY_STATE_DISABLED		1
 
-	u8			link_up;
+	u8			link_state;
+#define BNXT_LINK_STATE_UNKNOWN	0
+#define BNXT_LINK_STATE_DOWN	1
+#define BNXT_LINK_STATE_UP	2
+#define BNXT_LINK_IS_UP(bp)	((bp)->link_info.link_state == BNXT_LINK_STATE_UP)
 	u8			duplex;
 #define BNXT_LINK_DUPLEX_HALF	PORT_PHY_QCFG_RESP_DUPLEX_STATE_HALF
 #define BNXT_LINK_DUPLEX_FULL	PORT_PHY_QCFG_RESP_DUPLEX_STATE_FULL
@@ -1524,6 +1543,33 @@ struct bnxt_ctx_mem_info {
 	struct bnxt_mem_init	mem_init[BNXT_CTX_MEM_INIT_MAX];
 };
 
+enum bnxt_hw_err {
+	BNXT_HW_STATUS_HEALTHY			= 0x0,
+	BNXT_HW_STATUS_NVM_WRITE_ERR		= 0x1,
+	BNXT_HW_STATUS_NVM_ERASE_ERR		= 0x2,
+	BNXT_HW_STATUS_NVM_UNKNOWN_ERR		= 0x3,
+	BNXT_HW_STATUS_NVM_TEST_VPD_ENT_ERR	= 0x4,
+	BNXT_HW_STATUS_NVM_TEST_VPD_READ_ERR	= 0x5,
+	BNXT_HW_STATUS_NVM_TEST_VPD_WRITE_ERR	= 0x6,
+	BNXT_HW_STATUS_NVM_TEST_INCMPL_ERR	= 0x7,
+};
+
+struct bnxt_hw_health {
+	u32 nvm_err_address;
+	u32 nvm_write_errors;
+	u32 nvm_erase_errors;
+	u32 nvm_test_vpd_ent_errors;
+	u32 nvm_test_vpd_read_errors;
+	u32 nvm_test_vpd_write_errors;
+	u32 nvm_test_incmpl_errors;
+	u8 synd;
+	/* max a test in a day if previous test was successful */
+#define HW_RETEST_MIN_TIME	(1000 * 3600 * 24)
+	u8 nvm_test_result;
+	unsigned long nvm_test_timestamp;
+	struct devlink_health_reporter *hw_reporter;
+};
+
 enum bnxt_health_severity {
 	SEVERITY_NORMAL = 0,
 	SEVERITY_WARNING,
@@ -2041,6 +2087,7 @@ struct bnxt {
 #define BNXT_FW_EXCEPTION_SP_EVENT	19
 #define BNXT_LINK_CFG_CHANGE_SP_EVENT	21
 #define BNXT_FW_ECHO_REQUEST_SP_EVENT	23
+#define BNXT_FW_NVM_ERR_SP_EVENT	25
 
 	struct delayed_work	fw_reset_task;
 	int			fw_reset_state;
@@ -2100,8 +2147,8 @@ struct bnxt {
 	u32			lpi_tmr_lo;
 	u32			lpi_tmr_hi;
 
-	/* copied from flags in hwrm_port_phy_qcaps_output */
-	u8			phy_flags;
+	/* copied from flags and flags2 in hwrm_port_phy_qcaps_output */
+	u32			phy_flags;
 #define BNXT_PHY_FL_EEE_CAP		PORT_PHY_QCAPS_RESP_FLAGS_EEE_SUPPORTED
 #define BNXT_PHY_FL_EXT_LPBK		PORT_PHY_QCAPS_RESP_FLAGS_EXTERNAL_LPBK_SUPPORTED
 #define BNXT_PHY_FL_AN_PHY_LPBK		PORT_PHY_QCAPS_RESP_FLAGS_AUTONEG_LPBK_SUPPORTED
@@ -2110,6 +2157,8 @@ struct bnxt {
 #define BNXT_PHY_FL_NO_PHY_LPBK		PORT_PHY_QCAPS_RESP_FLAGS_LOCAL_LPBK_NOT_SUPPORTED
 #define BNXT_PHY_FL_FW_MANAGED_LKDN	PORT_PHY_QCAPS_RESP_FLAGS_FW_MANAGED_LINK_DOWN
 #define BNXT_PHY_FL_NO_FCS		PORT_PHY_QCAPS_RESP_FLAGS_NO_FCS
+#define BNXT_PHY_FL_NO_PAUSE		(PORT_PHY_QCAPS_RESP_FLAGS2_PAUSE_UNSUPPORTED << 8)
+#define BNXT_PHY_FL_NO_PFC		(PORT_PHY_QCAPS_RESP_FLAGS2_PFC_UNSUPPORTED << 8)
 
 	u8			num_tests;
 	struct bnxt_test_info	*test_info;
@@ -2139,6 +2188,8 @@ struct bnxt {
 	struct dentry		*debugfs_pdev;
 	struct device		*hwmon_dev;
 	enum board_idx		board_idx;
+
+	struct bnxt_hw_health	hw_health;
 };
 
 #define BNXT_NUM_RX_RING_STATS			8

drivers/net/ethernet/broadcom/bnxt/bnxt_dcb.c

@@ -627,7 +627,8 @@ static int bnxt_dcbnl_ieee_setpfc(struct net_device *dev, struct ieee_pfc *pfc)
 	int rc;
 
 	if (!(bp->dcbx_cap & DCB_CAP_DCBX_VER_IEEE) ||
-	    !(bp->dcbx_cap & DCB_CAP_DCBX_HOST))
+	    !(bp->dcbx_cap & DCB_CAP_DCBX_HOST) ||
+	    (bp->phy_flags & BNXT_PHY_FL_NO_PAUSE))
 		return -EINVAL;
 
 	if (!my_pfc) {