Skip to contentSkip to conversation

bnxt_en: Error recovery fixes.

PATCH In review Michael Chan 2 participants 9 messages Dec 10, 2019 – Dec 11, 2019

Subsystems (3)
lore
Share & tools
lore

Michael Chan

author maintainer
#0 lore raw

cover bnxt_en: Error recovery fixes.

This patch series contains fixes mostly for the error recovery feature and related areas. Please queue the series for -stable also. Thanks.

Michael Chan (2):
bnxt_en: Fix MSIX request logic for RDMA driver.
bnxt_en: Free context memory in the open path if firmware has been
reset.

Vasundhara Volam (5):
bnxt_en: Return error if FW returns more data than dump length
bnxt_en: Fix bp->fw_health allocation and free logic.
bnxt_en: Remove unnecessary NULL checks for fw_health
bnxt_en: Fix the logic that creates the health reporters.
bnxt_en: Add missing devlink health reporters for VFs.

 drivers/net/ethernet/broadcom/bnxt/bnxt.c         | 61 ++++++++++-----
 drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.c | 93 +++++++++++++++--------
 drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.h |  2 +
 drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c | 38 +++++++--
 drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.h |  4 +
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c     |  8 +-
 6 files changed, 146 insertions(+), 60 deletions(-)
Signature · 1 line
2.5.1
author maintainer
#1 lore raw

PATCH 1/7 bnxt_en: Fix MSIX request logic for RDMA driver.

The logic needs to check both bp->total_irqs and the reserved IRQs in hw_resc->resv_irqs if applicable and see if both are enough to cover the L2 and RDMA requested vectors. The current code is only checking bp->total_irqs and can fail in some code paths, such as the TX timeout code path with the RDMA driver requesting vectors after recovery. In this code path, we have not reserved enough MSIX resources for the RDMA driver yet.

Fixes 75720e6323a1 ("bnxt_en: Keep track of reserved IRQs.")
---
 drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
drivers/net/ethernet/broadcom/bnxt/bnxt_ulp.c +6 -2
index c601ff7..4a316c4 100644
@@ -113,8 +113,10 @@ static int bnxt_req_msix_vecs(struct bnxt_en_dev *edev, int ulp_id,
 {
 	struct net_device *dev = edev->net;
 	struct bnxt *bp = netdev_priv(dev);
+	struct bnxt_hw_resc *hw_resc;
 	int max_idx, max_cp_rings;
 	int avail_msix, idx;
+	int total_vecs;
 	int rc = 0;
 
 	ASSERT_RTNL();
@@ -142,7 +144,10 @@ static int bnxt_req_msix_vecs(struct bnxt_en_dev *edev, int ulp_id,
 	}
 	edev->ulp_tbl[ulp_id].msix_base = idx;
 	edev->ulp_tbl[ulp_id].msix_requested = avail_msix;
-	if (bp->total_irqs < (idx + avail_msix)) {
+	hw_resc = &bp->hw_resc;
+	total_vecs = idx + avail_msix;
+	if (bp->total_irqs < total_vecs ||
+	    (BNXT_NEW_RM(bp) && hw_resc->resv_irqs < total_vecs)) {
 		if (netif_running(dev)) {
 			bnxt_close_nic(bp, true, false);
 			rc = bnxt_open_nic(bp, true, false);
@@ -156,7 +161,6 @@ static int bnxt_req_msix_vecs(struct bnxt_en_dev *edev, int ulp_id,
 	}
 
 	if (BNXT_NEW_RM(bp)) {
-		struct bnxt_hw_resc *hw_resc = &bp->hw_resc;
 		int resv_msix;
 
 		resv_msix = hw_resc->resv_irqs - bp->cp_nr_rings;
Signature · 1 line
2.5.1
author maintainer
#2 lore raw

PATCH 2/7 bnxt_en: Free context memory in the open path if firmware has been reset.

In reply to Michael Chan

This will trigger new context memory to be rediscovered and allocated during the re-probe process after a firmware reset. Without this, the newly reset firmware does not have valid context memory and the driver will eventually fail to allocate some resources.

Fixes ec5d31e3c15d ("bnxt_en: Handle firmware reset status during IF_UP.")
---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/net/ethernet/broadcom/bnxt/bnxt.c +3 -0
index 85983f0..65c1c4e 100644
@@ -8796,6 +8796,9 @@ static int bnxt_hwrm_if_change(struct bnxt *bp, bool up)
 		if (fw_reset) {
 			if (!test_bit(BNXT_STATE_IN_FW_RESET, &bp->state))
 				bnxt_ulp_stop(bp);
+			bnxt_free_ctx_mem(bp);
+			kfree(bp->ctx);
+			bp->ctx = NULL;
 			rc = bnxt_fw_init_one(bp);
 			if (rc) {
 				set_bit(BNXT_STATE_ABORT_ERR, &bp->state);
Signature · 1 line
2.5.1
author maintainer
#3 lore raw

PATCH 3/7 bnxt_en: Return error if FW returns more data than dump length

In reply to Michael Chan

From: Vasundhara Volam <[email protected]>

If any change happened in the configuration of VF in VM while collecting live dump, there could be a race and firmware can return more data than allocated dump length. Fix it by keeping track of the accumulated core dump length copied so far and abort the copy with error code if the next chunk of core dump will exceed the original dump length.

Fixes 6c5657d085ae ("bnxt_en: Add support for ethtool get dump.")
---
 drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c | 38 ++++++++++++++++++-----
 drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.h |  4 +++
 2 files changed, 34 insertions(+), 8 deletions(-)
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c +30 -8
index 2ccf79c..08d56ec 100644
@@ -3071,8 +3071,15 @@ static int bnxt_hwrm_dbg_dma_data(struct bnxt *bp, void *msg, int msg_len,
 			}
 		}
 
-		if (info->dest_buf)
-			memcpy(info->dest_buf + off, dma_buf, len);
+		if (info->dest_buf) {
+			if ((info->seg_start + off + len) <=
+			    BNXT_COREDUMP_BUF_LEN(info->buf_len)) {
+				memcpy(info->dest_buf + off, dma_buf, len);
+			} else {
+				rc = -ENOBUFS;
+				break;
+			}
+		}
 
 		if (cmn_req->req_type ==
 				cpu_to_le16(HWRM_DBG_COREDUMP_RETRIEVE))
@@ -3126,7 +3133,7 @@ static int bnxt_hwrm_dbg_coredump_initiate(struct bnxt *bp, u16 component_id,
 
 static int bnxt_hwrm_dbg_coredump_retrieve(struct bnxt *bp, u16 component_id,
 					   u16 segment_id, u32 *seg_len,
-					   void *buf, u32 offset)
+					   void *buf, u32 buf_len, u32 offset)
 {
 	struct hwrm_dbg_coredump_retrieve_input req = {0};
 	struct bnxt_hwrm_dbg_dma_info info = {NULL};
@@ -3141,8 +3148,11 @@ static int bnxt_hwrm_dbg_coredump_retrieve(struct bnxt *bp, u16 component_id,
 				seq_no);
 	info.data_len_off = offsetof(struct hwrm_dbg_coredump_retrieve_output,
 				     data_len);
-	if (buf)
+	if (buf) {
 		info.dest_buf = buf + offset;
+		info.buf_len = buf_len;
+		info.seg_start = offset;
+	}
 
 	rc = bnxt_hwrm_dbg_dma_data(bp, &req, sizeof(req), &info);
 	if (!rc)
@@ -3232,14 +3242,17 @@ bnxt_fill_coredump_record(struct bnxt *bp, struct bnxt_coredump_record *record,
 static int bnxt_get_coredump(struct bnxt *bp, void *buf, u32 *dump_len)
 {
 	u32 ver_get_resp_len = sizeof(struct hwrm_ver_get_output);
+	u32 offset = 0, seg_hdr_len, seg_record_len, buf_len = 0;
 	struct coredump_segment_record *seg_record = NULL;
-	u32 offset = 0, seg_hdr_len, seg_record_len;
 	struct bnxt_coredump_segment_hdr seg_hdr;
 	struct bnxt_coredump coredump = {NULL};
 	time64_t start_time;
 	u16 start_utc;
 	int rc = 0, i;
 
+	if (buf)
+		buf_len = *dump_len;
+
 	start_time = ktime_get_real_seconds();
 	start_utc = sys_tz.tz_minuteswest * 60;
 	seg_hdr_len = sizeof(seg_hdr);
@@ -3272,6 +3285,12 @@ static int bnxt_get_coredump(struct bnxt *bp, void *buf, u32 *dump_len)
 		u32 duration = 0, seg_len = 0;
 		unsigned long start, end;
 
+		if (buf && ((offset + seg_hdr_len) >
+			    BNXT_COREDUMP_BUF_LEN(buf_len))) {
+			rc = -ENOBUFS;
+			goto err;
+		}
+
 		start = jiffies;
 
 		rc = bnxt_hwrm_dbg_coredump_initiate(bp, comp_id, seg_id);
@@ -3284,9 +3303,11 @@ static int bnxt_get_coredump(struct bnxt *bp, void *buf, u32 *dump_len)
 
 		/* Write segment data into the buffer */
 		rc = bnxt_hwrm_dbg_coredump_retrieve(bp, comp_id, seg_id,
-						     &seg_len, buf,
+						     &seg_len, buf, buf_len,
 						     offset + seg_hdr_len);
-		if (rc)
+		if (rc && rc == -ENOBUFS)
+			goto err;
+		else if (rc)
 			netdev_err(bp->dev,
 				   "Failed to retrieve coredump for seg = %d\n",
 				   seg_record->segment_id);
@@ -3316,7 +3337,8 @@ static int bnxt_get_coredump(struct bnxt *bp, void *buf, u32 *dump_len)
 					  rc);
 	kfree(coredump.data);
 	*dump_len += sizeof(struct bnxt_coredump_record);
-
+	if (rc == -ENOBUFS)
+		netdev_err(bp->dev, "Firmware returned large coredump buffer");
 	return rc;
 }
 
drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.h +4 -0
index 4428d0a..3576d95 100644
@@ -31,6 +31,8 @@ struct bnxt_coredump {
 	u16		total_segs;
 };
 
+#define BNXT_COREDUMP_BUF_LEN(len) ((len) - sizeof(struct bnxt_coredump_record))
+
 struct bnxt_hwrm_dbg_dma_info {
 	void *dest_buf;
 	int dest_buf_size;
@@ -38,6 +40,8 @@ struct bnxt_hwrm_dbg_dma_info {
 	u16 seq_off;
 	u16 data_len_off;
 	u16 segs;
+	u32 seg_start;
+	u32 buf_len;
 };
 
 struct hwrm_dbg_cmn_input {
Signature · 1 line
2.5.1
author maintainer
#4 lore raw

PATCH 4/7 bnxt_en: Fix bp->fw_health allocation and free logic.

In reply to Michael Chan

From: Vasundhara Volam <[email protected]>

bp->fw_health needs to be allocated for either the firmware initiated reset feature or the driver initiated error recovery feature. The current code is not allocating bp->fw_health for all the necessary cases. This patch corrects the logic to allocate bp->fw_health correctly when needed. If allocation fails, we clear the feature flags.

We also add the the missing kfree(bp->fw_health) when the driver is unloaded. If we get an async reset message from the firmware, we also need to make sure that we have a valid bp->fw_health before proceeding.

Fixes 07f83d72d238 ("bnxt_en: Discover firmware error recovery capabilities.")
---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c | 36 ++++++++++++++++++++++---------
 1 file changed, 26 insertions(+), 10 deletions(-)
drivers/net/ethernet/broadcom/bnxt/bnxt.c +26 -10
index 65c1c4e..d6a5fce 100644
@@ -2001,6 +2001,9 @@ static int bnxt_async_event_process(struct bnxt *bp,
 	case ASYNC_EVENT_CMPL_EVENT_ID_RESET_NOTIFY: {
 		u32 data1 = le32_to_cpu(cmpl->event_data1);
 
+		if (!bp->fw_health)
+			goto async_event_process_exit;
+
 		bp->fw_reset_timestamp = jiffies;
 		bp->fw_reset_min_dsecs = cmpl->timestamp_lo;
 		if (!bp->fw_reset_min_dsecs)
@@ -4421,8 +4424,9 @@ int bnxt_hwrm_func_drv_rgtr(struct bnxt *bp, unsigned long *bmap, int bmap_size,
 			    FUNC_DRV_RGTR_REQ_ENABLES_ASYNC_EVENT_FWD);
 
 	req.os_type = cpu_to_le16(FUNC_DRV_RGTR_REQ_OS_TYPE_LINUX);
-	flags = FUNC_DRV_RGTR_REQ_FLAGS_16BIT_VER_MODE |
-		FUNC_DRV_RGTR_REQ_FLAGS_HOT_RESET_SUPPORT;
+	flags = FUNC_DRV_RGTR_REQ_FLAGS_16BIT_VER_MODE;
+	if (bp->fw_cap & BNXT_FW_CAP_HOT_RESET)
+		flags |= FUNC_DRV_RGTR_REQ_FLAGS_HOT_RESET_SUPPORT;
 	if (bp->fw_cap & BNXT_FW_CAP_ERROR_RECOVERY)
 		flags |= FUNC_DRV_RGTR_REQ_FLAGS_ERROR_RECOVERY_SUPPORT |
 			 FUNC_DRV_RGTR_REQ_FLAGS_MASTER_SUPPORT;
@@ -7115,14 +7119,6 @@ static int bnxt_hwrm_error_recovery_qcfg(struct bnxt *bp)
 	rc = _hwrm_send_message(bp, &req, sizeof(req), HWRM_CMD_TIMEOUT);
 	if (rc)
 		goto err_recovery_out;
-	if (!fw_health) {
-		fw_health = kzalloc(sizeof(*fw_health), GFP_KERNEL);
-		bp->fw_health = fw_health;
-		if (!fw_health) {
-			rc = -ENOMEM;
-			goto err_recovery_out;
-		}
-	}
 	fw_health->flags = le32_to_cpu(resp->flags);
 	if ((fw_health->flags & ERROR_RECOVERY_QCFG_RESP_FLAGS_CO_CPU) &&
 	    !(bp->fw_cap & BNXT_FW_CAP_KONG_MB_CHNL)) {
@@ -10485,6 +10481,23 @@ static void bnxt_init_dflt_coal(struct bnxt *bp)
 	bp->stats_coal_ticks = BNXT_DEF_STATS_COAL_TICKS;
 }
 
+static void bnxt_alloc_fw_health(struct bnxt *bp)
+{
+	if (bp->fw_health)
+		return;
+
+	if (!(bp->fw_cap & BNXT_FW_CAP_HOT_RESET) &&
+	    !(bp->fw_cap & BNXT_FW_CAP_ERROR_RECOVERY))
+		return;
+
+	bp->fw_health = kzalloc(sizeof(*bp->fw_health), GFP_KERNEL);
+	if (!bp->fw_health) {
+		netdev_warn(bp->dev, "Failed to allocate fw_health\n");
+		bp->fw_cap &= ~BNXT_FW_CAP_HOT_RESET;
+		bp->fw_cap &= ~BNXT_FW_CAP_ERROR_RECOVERY;
+	}
+}
+
 static int bnxt_fw_init_one_p1(struct bnxt *bp)
 {
 	int rc;
@@ -10531,6 +10544,7 @@ static int bnxt_fw_init_one_p2(struct bnxt *bp)
 		netdev_warn(bp->dev, "hwrm query adv flow mgnt failure rc: %d\n",
 			    rc);
 
+	bnxt_alloc_fw_health(bp);
 	rc = bnxt_hwrm_error_recovery_qcfg(bp);
 	if (rc)
 		netdev_warn(bp->dev, "hwrm query error recovery failure rc: %d\n",
@@ -11418,6 +11432,8 @@ static void bnxt_remove_one(struct pci_dev *pdev)
 	bnxt_dcb_free(bp);
 	kfree(bp->edev);
 	bp->edev = NULL;
+	kfree(bp->fw_health);
+	bp->fw_health = NULL;
 	bnxt_cleanup_pci(bp);
 	bnxt_free_ctx_mem(bp);
 	kfree(bp->ctx);
Signature · 1 line
2.5.1
author maintainer
#5 lore raw

PATCH 5/7 bnxt_en: Remove unnecessary NULL checks for fw_health

In reply to Michael Chan

From: Vasundhara Volam <[email protected]>

After fixing the allocation of bp->fw_health in the previous patch, the driver will not go through the fw reset and recovery code paths if bp->fw_health allocation fails. So we can now remove the unnecessary NULL checks.

---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c         | 6 ++----
 drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.c | 6 +-----
 2 files changed, 3 insertions(+), 9 deletions(-)
drivers/net/ethernet/broadcom/bnxt/bnxt.c +2 -4
index d6a5fce..2a100ff 100644
@@ -9989,8 +9989,7 @@ static void bnxt_fw_health_check(struct bnxt *bp)
 	struct bnxt_fw_health *fw_health = bp->fw_health;
 	u32 val;
 
-	if (!fw_health || !fw_health->enabled ||
-	    test_bit(BNXT_STATE_IN_FW_RESET, &bp->state))
+	if (!fw_health->enabled || test_bit(BNXT_STATE_IN_FW_RESET, &bp->state))
 		return;
 
 	if (fw_health->tmr_counter) {
@@ -10768,8 +10767,7 @@ static void bnxt_fw_reset_task(struct work_struct *work)
 		bnxt_queue_fw_reset_work(bp, bp->fw_reset_min_dsecs * HZ / 10);
 		return;
 	case BNXT_FW_RESET_STATE_ENABLE_DEV:
-		if (test_bit(BNXT_STATE_FW_FATAL_COND, &bp->state) &&
-		    bp->fw_health) {
+		if (test_bit(BNXT_STATE_FW_FATAL_COND, &bp->state)) {
 			u32 val;
 
 			val = bnxt_fw_health_readl(bp,
drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.c +1 -5
index acb2dd6..1e7c7c3 100644
@@ -39,11 +39,10 @@ static int bnxt_fw_reporter_diagnose(struct devlink_health_reporter *reporter,
 				     struct netlink_ext_ack *extack)
 {
 	struct bnxt *bp = devlink_health_reporter_priv(reporter);
-	struct bnxt_fw_health *health = bp->fw_health;
 	u32 val, health_status;
 	int rc;
 
-	if (!health || test_bit(BNXT_STATE_IN_FW_RESET, &bp->state))
+	if (test_bit(BNXT_STATE_IN_FW_RESET, &bp->state))
 		return 0;
 
 	val = bnxt_fw_health_readl(bp, BNXT_FW_HEALTH_REG);
@@ -185,9 +184,6 @@ void bnxt_devlink_health_report(struct bnxt *bp, unsigned long event)
 	struct bnxt_fw_health *fw_health = bp->fw_health;
 	struct bnxt_fw_reporter_ctx fw_reporter_ctx;
 
-	if (!fw_health)
-		return;
-
 	fw_reporter_ctx.sp_event = event;
 	switch (event) {
 	case BNXT_FW_RESET_NOTIFY_SP_EVENT:
Signature · 1 line
2.5.1
author maintainer
#6 lore raw

PATCH 6/7 bnxt_en: Fix the logic that creates the health reporters.

In reply to Michael Chan

From: Vasundhara Volam <[email protected]>

Fix the logic to properly check the fw capabilities and create the devlink health reporters only when needed. The current code creates the reporters unconditionally as long as bp->fw_health is valid, and that's not correct.

Call bnxt_dl_fw_reporters_create() directly from the init and reset code path instead of from bnxt_dl_register(). This allows the reporters to be adjusted when capabilities change. The same applies to bnxt_dl_fw_reporters_destroy().

Fixes 6763c779c2d8 ("bnxt_en: Add new FW devlink_health_reporter")
---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c         | 11 +++-
 drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.c | 64 ++++++++++++++++-------
 drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.h |  2 +
 3 files changed, 56 insertions(+), 21 deletions(-)
drivers/net/ethernet/broadcom/bnxt/bnxt.c +10 -1
index 2a100ff..819b7d7 100644
@@ -10625,6 +10625,12 @@ static int bnxt_fw_init_one(struct bnxt *bp)
 	rc = bnxt_approve_mac(bp, bp->dev->dev_addr, false);
 	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;
 }
@@ -11413,6 +11419,7 @@ static void bnxt_remove_one(struct pci_dev *pdev)
 
 	if (BNXT_PF(bp)) {
 		bnxt_sriov_disable(bp);
+		bnxt_dl_fw_reporters_destroy(bp, true);
 		bnxt_dl_unregister(bp);
 	}
 
@@ -11892,8 +11899,10 @@ static int bnxt_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 	if (rc)
 		goto init_err_cleanup_tc;
 
-	if (BNXT_PF(bp))
+	if (BNXT_PF(bp)) {
 		bnxt_dl_register(bp);
+		bnxt_dl_fw_reporters_create(bp);
+	}
 
 	netdev_info(dev, "%s found at mem %lx, node addr %pM\n",
 		    board_info[ent->driver_data].name,
drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.c +44 -20

112 lines not on this page · raw message

drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.h +2 -0
index 665d4bd..6db6c3d 100644
@@ -58,6 +58,8 @@ struct bnxt_dl_nvm_param {
 
 void bnxt_devlink_health_report(struct bnxt *bp, unsigned long event);
 void bnxt_dl_health_status_update(struct bnxt *bp, bool healthy);
+void bnxt_dl_fw_reporters_create(struct bnxt *bp);
+void bnxt_dl_fw_reporters_destroy(struct bnxt *bp, bool all);
 int bnxt_dl_register(struct bnxt *bp);
 void bnxt_dl_unregister(struct bnxt *bp);
 
Signature · 1 line
2.5.1
author maintainer
#7 lore raw

PATCH 7/7 bnxt_en: Add missing devlink health reporters for VFs.

In reply to Michael Chan

From: Vasundhara Volam <[email protected]>

The VF driver also needs to create the health reporters since VFs are also involved in firmware reset and recovery. Modify bnxt_dl_register() and bnxt_dl_unregister() so that they can be called by the VFs to register/unregister devlink. Only the PF will register the devlink parameters. With devlink registered, we can now create the health reporters on the VFs.

Fixes 6763c779c2d8 ("bnxt_en: Add new FW devlink_health_reporter")
---
 drivers/net/ethernet/broadcom/bnxt/bnxt.c         | 13 +++++--------
 drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.c | 23 +++++++++++++++++------
 2 files changed, 22 insertions(+), 14 deletions(-)
drivers/net/ethernet/broadcom/bnxt/bnxt.c +5 -8
index 819b7d7..a754903 100644
@@ -11417,12 +11417,11 @@ static void bnxt_remove_one(struct pci_dev *pdev)
 	struct net_device *dev = pci_get_drvdata(pdev);
 	struct bnxt *bp = netdev_priv(dev);
 
-	if (BNXT_PF(bp)) {
+	if (BNXT_PF(bp))
 		bnxt_sriov_disable(bp);
-		bnxt_dl_fw_reporters_destroy(bp, true);
-		bnxt_dl_unregister(bp);
-	}
 
+	bnxt_dl_fw_reporters_destroy(bp, true);
+	bnxt_dl_unregister(bp);
 	pci_disable_pcie_error_reporting(pdev);
 	unregister_netdev(dev);
 	bnxt_shutdown_tc(bp);
@@ -11899,10 +11898,8 @@ static int bnxt_init_one(struct pci_dev *pdev, const struct pci_device_id *ent)
 	if (rc)
 		goto init_err_cleanup_tc;
 
-	if (BNXT_PF(bp)) {
-		bnxt_dl_register(bp);
-		bnxt_dl_fw_reporters_create(bp);
-	}
+	bnxt_dl_register(bp);
+	bnxt_dl_fw_reporters_create(bp);
 
 	netdev_info(dev, "%s found at mem %lx, node addr %pM\n",
 		    board_info[ent->driver_data].name,
drivers/net/ethernet/broadcom/bnxt/bnxt_devlink.c +17 -6
index 136953a..3eedd44 100644
@@ -270,6 +270,8 @@ static const struct devlink_ops bnxt_dl_ops = {
 	.flash_update	  = bnxt_dl_flash_update,
 };
 
+static const struct devlink_ops bnxt_vf_dl_ops;
+
 enum bnxt_dl_param_id {
 	BNXT_DEVLINK_PARAM_ID_BASE = DEVLINK_PARAM_GENERIC_ID_MAX,
 	BNXT_DEVLINK_PARAM_ID_GRE_VER_CHECK,
@@ -483,7 +485,10 @@ int bnxt_dl_register(struct bnxt *bp)
 		return -ENOTSUPP;
 	}
 
-	dl = devlink_alloc(&bnxt_dl_ops, sizeof(struct bnxt_dl));
+	if (BNXT_PF(bp))
+		dl = devlink_alloc(&bnxt_dl_ops, sizeof(struct bnxt_dl));
+	else
+		dl = devlink_alloc(&bnxt_vf_dl_ops, sizeof(struct bnxt_dl));
 	if (!dl) {
 		netdev_warn(bp->dev, "devlink_alloc failed");
 		return -ENOMEM;
@@ -502,6 +507,9 @@ int bnxt_dl_register(struct bnxt *bp)
 		goto err_dl_free;
 	}
 
+	if (!BNXT_PF(bp))
+		return 0;
+
 	rc = devlink_params_register(dl, bnxt_dl_params,
 				     ARRAY_SIZE(bnxt_dl_params));
 	if (rc) {
@@ -551,11 +559,14 @@ void bnxt_dl_unregister(struct bnxt *bp)
 	if (!dl)
 		return;
 
-	devlink_port_params_unregister(&bp->dl_port, bnxt_dl_port_params,
-				       ARRAY_SIZE(bnxt_dl_port_params));
-	devlink_port_unregister(&bp->dl_port);
-	devlink_params_unregister(dl, bnxt_dl_params,
-				  ARRAY_SIZE(bnxt_dl_params));
+	if (BNXT_PF(bp)) {
+		devlink_port_params_unregister(&bp->dl_port,
+					       bnxt_dl_port_params,
+					       ARRAY_SIZE(bnxt_dl_port_params));
+		devlink_port_unregister(&bp->dl_port);
+		devlink_params_unregister(dl, bnxt_dl_params,
+					  ARRAY_SIZE(bnxt_dl_params));
+	}
 	devlink_unregister(dl);
 	devlink_free(dl);
 }
Signature · 1 line
2.5.1
maintainer
#8 lore raw
In reply to Michael Chan

From: Michael Chan <[email protected]>
Date: Tue, 10 Dec 2019 02:49:06 -0500

Quoted 2 lines

This patch series contains fixes mostly for the error recovery feature and related areas. Please queue the series for -stable also. Thanks.

Series applied and queued up for -stable, thanks.