[PATCH v5 net-next 00/10] qed: introduce devlink health support

[PATCH v5 net-next 00/10] qed: introduce devlink health support

[Igor Russkikh]

This is a followup implementation after series

https://patchwork.ozlabs.org/project/netdev/cover/20200514095727.1361-1-irusskikh@marvell.com/

This is an implementation of devlink health infrastructure.

With this we are now able to report HW errors to devlink, and it'll take
its own actions depending on user configuration to capture and store the
dump at the bad moment, and to request the driver to recover the device.

So far we do not differentiate global device failures or specific PCI
function failures. This means that some errors specific to one physical
function will affect an entire device. This is not yet fully designed
and verified, will followup in future.

Solution was verified with artificial HW errors generated, existing
tools for dump analysis could be used.

v5: improved patch 4 description
v4:
 - commit message and other fixes after Jiri's comments
 - removed one patch (will send to net)
v3: fix uninit var usage in patch 11
v2: fix #include issue from kbuild test robot.

Igor Russkikh (10):
  qed: move out devlink logic into a new file
  qed/qede: make devlink survive recovery
  qed: fix kconfig help entries
  qed: implement devlink info request
  qed: health reporter init deinit seq
  qed: use devlink logic to report errors
  qed*: make use of devlink recovery infrastructure
  qed: implement devlink dump
  qed: align adjacent indent
  qede: make driver reliable on unload after failures

 drivers/net/ethernet/qlogic/Kconfig           |   5 +-
 drivers/net/ethernet/qlogic/qed/Makefile      |   1 +
 drivers/net/ethernet/qlogic/qed/qed.h         |   3 +-
 drivers/net/ethernet/qlogic/qed/qed_dev.c     |   9 +
 drivers/net/ethernet/qlogic/qed/qed_devlink.c | 259 ++++++++++++++++++
 drivers/net/ethernet/qlogic/qed/qed_devlink.h |  20 ++
 drivers/net/ethernet/qlogic/qed/qed_main.c    | 116 +-------
 drivers/net/ethernet/qlogic/qede/qede.h       |   2 +
 drivers/net/ethernet/qlogic/qede/qede_main.c  |  38 ++-
 include/linux/qed/qed_if.h                    |  23 +-
 10 files changed, 347 insertions(+), 129 deletions(-)
 create mode 100644 drivers/net/ethernet/qlogic/qed/qed_devlink.c
 create mode 100644 drivers/net/ethernet/qlogic/qed/qed_devlink.h

-- 
2.17.1

[PATCH v5 net-next 01/10] qed: move out devlink logic into a new file

[Igor Russkikh]

We are extending devlink infrastructure, thus move the existing
stuff into a new file qed_devlink.c

Signed-off-by: Igor Russkikh <irusskikh@marvell.com>
Signed-off-by: Alexander Lobakin <alobakin@marvell.com>
Signed-off-by: Michal Kalderon <michal.kalderon@marvell.com>
---
 drivers/net/ethernet/qlogic/qed/Makefile      |   1 +
 drivers/net/ethernet/qlogic/qed/qed_devlink.c | 110 ++++++++++++++++++
 drivers/net/ethernet/qlogic/qed/qed_devlink.h |  15 +++
 drivers/net/ethernet/qlogic/qed/qed_main.c    | 102 +---------------
 4 files changed, 127 insertions(+), 101 deletions(-)
 create mode 100644 drivers/net/ethernet/qlogic/qed/qed_devlink.c
 create mode 100644 drivers/net/ethernet/qlogic/qed/qed_devlink.h

diff --git a/drivers/net/ethernet/qlogic/qed/Makefile b/drivers/net/ethernet/qlogic/qed/Makefile
index f947b105cf14..8251755ec18c 100644
--- a/drivers/net/ethernet/qlogic/qed/Makefile
+++ b/drivers/net/ethernet/qlogic/qed/Makefile
@@ -9,6 +9,7 @@ qed-y :=			\
 	qed_dcbx.o		\
 	qed_debug.o		\
 	qed_dev.o		\
+	qed_devlink.o		\
 	qed_hw.o		\
 	qed_init_fw_funcs.o	\
 	qed_init_ops.o		\
diff --git a/drivers/net/ethernet/qlogic/qed/qed_devlink.c b/drivers/net/ethernet/qlogic/qed/qed_devlink.c
new file mode 100644
index 000000000000..eb693787c99e
--- /dev/null
+++ b/drivers/net/ethernet/qlogic/qed/qed_devlink.c
@@ -0,0 +1,110 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* Marvell/Qlogic FastLinQ NIC driver
+ *
+ * Copyright (C) 2020 Marvell International Ltd.
+ */
+
+#include <linux/kernel.h>
+#include "qed.h"
+#include "qed_devlink.h"
+
+enum qed_devlink_param_id {
+	QED_DEVLINK_PARAM_ID_BASE = DEVLINK_PARAM_GENERIC_ID_MAX,
+	QED_DEVLINK_PARAM_ID_IWARP_CMT,
+};
+
+struct qed_devlink {
+	struct qed_dev *cdev;
+};
+
+static int qed_dl_param_get(struct devlink *dl, u32 id,
+			    struct devlink_param_gset_ctx *ctx)
+{
+	struct qed_devlink *qed_dl;
+	struct qed_dev *cdev;
+
+	qed_dl = devlink_priv(dl);
+	cdev = qed_dl->cdev;
+	ctx->val.vbool = cdev->iwarp_cmt;
+
+	return 0;
+}
+
+static int qed_dl_param_set(struct devlink *dl, u32 id,
+			    struct devlink_param_gset_ctx *ctx)
+{
+	struct qed_devlink *qed_dl;
+	struct qed_dev *cdev;
+
+	qed_dl = devlink_priv(dl);
+	cdev = qed_dl->cdev;
+	cdev->iwarp_cmt = ctx->val.vbool;
+
+	return 0;
+}
+
+static const struct devlink_param qed_devlink_params[] = {
+	DEVLINK_PARAM_DRIVER(QED_DEVLINK_PARAM_ID_IWARP_CMT,
+			     "iwarp_cmt", DEVLINK_PARAM_TYPE_BOOL,
+			     BIT(DEVLINK_PARAM_CMODE_RUNTIME),
+			     qed_dl_param_get, qed_dl_param_set, NULL),
+};
+
+static const struct devlink_ops qed_dl_ops;
+
+int qed_devlink_register(struct qed_dev *cdev)
+{
+	union devlink_param_value value;
+	struct qed_devlink *qed_dl;
+	struct devlink *dl;
+	int rc;
+
+	dl = devlink_alloc(&qed_dl_ops, sizeof(*qed_dl));
+	if (!dl)
+		return -ENOMEM;
+
+	qed_dl = devlink_priv(dl);
+
+	cdev->dl = dl;
+	qed_dl->cdev = cdev;
+
+	rc = devlink_register(dl, &cdev->pdev->dev);
+	if (rc)
+		goto err_free;
+
+	rc = devlink_params_register(dl, qed_devlink_params,
+				     ARRAY_SIZE(qed_devlink_params));
+	if (rc)
+		goto err_unregister;
+
+	value.vbool = false;
+	devlink_param_driverinit_value_set(dl,
+					   QED_DEVLINK_PARAM_ID_IWARP_CMT,
+					   value);
+
+	devlink_params_publish(dl);
+	cdev->iwarp_cmt = false;
+
+	return 0;
+
+err_unregister:
+	devlink_unregister(dl);
+
+err_free:
+	cdev->dl = NULL;
+	devlink_free(dl);
+
+	return rc;
+}
+
+void qed_devlink_unregister(struct qed_dev *cdev)
+{
+	if (!cdev->dl)
+		return;
+
+	devlink_params_unregister(cdev->dl, qed_devlink_params,
+				  ARRAY_SIZE(qed_devlink_params));
+
+	devlink_unregister(cdev->dl);
+	devlink_free(cdev->dl);
+}
diff --git a/drivers/net/ethernet/qlogic/qed/qed_devlink.h b/drivers/net/ethernet/qlogic/qed/qed_devlink.h
new file mode 100644
index 000000000000..b94c40e9b7c1
--- /dev/null
+++ b/drivers/net/ethernet/qlogic/qed/qed_devlink.h
@@ -0,0 +1,15 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/* Marvell/Qlogic FastLinQ NIC driver
+ *
+ * Copyright (C) 2020 Marvell International Ltd.
+ */
+#ifndef _QED_DEVLINK_H
+#define _QED_DEVLINK_H
+
+#include <linux/qed/qed_if.h>
+#include <net/devlink.h>
+
+int qed_devlink_register(struct qed_dev *cdev);
+void qed_devlink_unregister(struct qed_dev *cdev);
+
+#endif
diff --git a/drivers/net/ethernet/qlogic/qed/qed_main.c b/drivers/net/ethernet/qlogic/qed/qed_main.c
index 2558cb680db3..8751355d9ef7 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_main.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_main.c
@@ -39,6 +39,7 @@
 #include "qed_hw.h"
 #include "qed_selftest.h"
 #include "qed_debug.h"
+#include "qed_devlink.h"
 
 #define QED_ROCE_QPS			(8192)
 #define QED_ROCE_DPIS			(8)
@@ -510,107 +511,6 @@ static int qed_set_power_state(struct qed_dev *cdev, pci_power_t state)
 	return 0;
 }
 
-struct qed_devlink {
-	struct qed_dev *cdev;
-};
-
-enum qed_devlink_param_id {
-	QED_DEVLINK_PARAM_ID_BASE = DEVLINK_PARAM_GENERIC_ID_MAX,
-	QED_DEVLINK_PARAM_ID_IWARP_CMT,
-};
-
-static int qed_dl_param_get(struct devlink *dl, u32 id,
-			    struct devlink_param_gset_ctx *ctx)
-{
-	struct qed_devlink *qed_dl;
-	struct qed_dev *cdev;
-
-	qed_dl = devlink_priv(dl);
-	cdev = qed_dl->cdev;
-	ctx->val.vbool = cdev->iwarp_cmt;
-
-	return 0;
-}
-
-static int qed_dl_param_set(struct devlink *dl, u32 id,
-			    struct devlink_param_gset_ctx *ctx)
-{
-	struct qed_devlink *qed_dl;
-	struct qed_dev *cdev;
-
-	qed_dl = devlink_priv(dl);
-	cdev = qed_dl->cdev;
-	cdev->iwarp_cmt = ctx->val.vbool;
-
-	return 0;
-}
-
-static const struct devlink_param qed_devlink_params[] = {
-	DEVLINK_PARAM_DRIVER(QED_DEVLINK_PARAM_ID_IWARP_CMT,
-			     "iwarp_cmt", DEVLINK_PARAM_TYPE_BOOL,
-			     BIT(DEVLINK_PARAM_CMODE_RUNTIME),
-			     qed_dl_param_get, qed_dl_param_set, NULL),
-};
-
-static const struct devlink_ops qed_dl_ops;
-
-static int qed_devlink_register(struct qed_dev *cdev)
-{
-	union devlink_param_value value;
-	struct qed_devlink *qed_dl;
-	struct devlink *dl;
-	int rc;
-
-	dl = devlink_alloc(&qed_dl_ops, sizeof(*qed_dl));
-	if (!dl)
-		return -ENOMEM;
-
-	qed_dl = devlink_priv(dl);
-
-	cdev->dl = dl;
-	qed_dl->cdev = cdev;
-
-	rc = devlink_register(dl, &cdev->pdev->dev);
-	if (rc)
-		goto err_free;
-
-	rc = devlink_params_register(dl, qed_devlink_params,
-				     ARRAY_SIZE(qed_devlink_params));
-	if (rc)
-		goto err_unregister;
-
-	value.vbool = false;
-	devlink_param_driverinit_value_set(dl,
-					   QED_DEVLINK_PARAM_ID_IWARP_CMT,
-					   value);
-
-	devlink_params_publish(dl);
-	cdev->iwarp_cmt = false;
-
-	return 0;
-
-err_unregister:
-	devlink_unregister(dl);
-
-err_free:
-	cdev->dl = NULL;
-	devlink_free(dl);
-
-	return rc;
-}
-
-static void qed_devlink_unregister(struct qed_dev *cdev)
-{
-	if (!cdev->dl)
-		return;
-
-	devlink_params_unregister(cdev->dl, qed_devlink_params,
-				  ARRAY_SIZE(qed_devlink_params));
-
-	devlink_unregister(cdev->dl);
-	devlink_free(cdev->dl);
-}
-
 /* probing */
 static struct qed_dev *qed_probe(struct pci_dev *pdev,
 				 struct qed_probe_params *params)
-- 
2.17.1

[PATCH v5 net-next 02/10] qed/qede: make devlink survive recovery

[Igor Russkikh]

Devlink instance lifecycle was linked to qed_dev object,
that caused devlink to be recreated on each recovery.

Changing it by making higher level driver (qede) responsible for its
life. This way devlink now survives recoveries.

qede now stores devlink structure pointer as a part of its device
object, devlink private data contains a linkage structure,
qed_devlink.

Signed-off-by: Igor Russkikh <irusskikh@marvell.com>
Signed-off-by: Alexander Lobakin <alobakin@marvell.com>
Signed-off-by: Michal Kalderon <michal.kalderon@marvell.com>
---
 drivers/net/ethernet/qlogic/qed/qed.h         |  1 -
 drivers/net/ethernet/qlogic/qed/qed_devlink.c | 40 ++++++++-----------
 drivers/net/ethernet/qlogic/qed/qed_devlink.h |  4 +-
 drivers/net/ethernet/qlogic/qed/qed_main.c    | 10 +----
 drivers/net/ethernet/qlogic/qede/qede.h       |  1 +
 drivers/net/ethernet/qlogic/qede/qede_main.c  | 18 +++++++++
 include/linux/qed/qed_if.h                    |  9 +++++
 7 files changed, 48 insertions(+), 35 deletions(-)

diff --git a/drivers/net/ethernet/qlogic/qed/qed.h b/drivers/net/ethernet/qlogic/qed/qed.h
index b2a7b53ee760..b6ce1488abcc 100644
--- a/drivers/net/ethernet/qlogic/qed/qed.h
+++ b/drivers/net/ethernet/qlogic/qed/qed.h
@@ -849,7 +849,6 @@ struct qed_dev {
 	u32 rdma_max_srq_sge;
 	u16 tunn_feature_mask;
 
-	struct devlink			*dl;
 	bool				iwarp_cmt;
 };
 
diff --git a/drivers/net/ethernet/qlogic/qed/qed_devlink.c b/drivers/net/ethernet/qlogic/qed/qed_devlink.c
index eb693787c99e..a62c47c61edf 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_devlink.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_devlink.c
@@ -5,6 +5,7 @@
  */
 
 #include <linux/kernel.h>
+#include <linux/qed/qed_if.h>
 #include "qed.h"
 #include "qed_devlink.h"
 
@@ -13,17 +14,12 @@ enum qed_devlink_param_id {
 	QED_DEVLINK_PARAM_ID_IWARP_CMT,
 };
 
-struct qed_devlink {
-	struct qed_dev *cdev;
-};
-
 static int qed_dl_param_get(struct devlink *dl, u32 id,
 			    struct devlink_param_gset_ctx *ctx)
 {
-	struct qed_devlink *qed_dl;
+	struct qed_devlink *qed_dl = devlink_priv(dl);
 	struct qed_dev *cdev;
 
-	qed_dl = devlink_priv(dl);
 	cdev = qed_dl->cdev;
 	ctx->val.vbool = cdev->iwarp_cmt;
 
@@ -33,10 +29,9 @@ static int qed_dl_param_get(struct devlink *dl, u32 id,
 static int qed_dl_param_set(struct devlink *dl, u32 id,
 			    struct devlink_param_gset_ctx *ctx)
 {
-	struct qed_devlink *qed_dl;
+	struct qed_devlink *qed_dl = devlink_priv(dl);
 	struct qed_dev *cdev;
 
-	qed_dl = devlink_priv(dl);
 	cdev = qed_dl->cdev;
 	cdev->iwarp_cmt = ctx->val.vbool;
 
@@ -52,21 +47,19 @@ static const struct devlink_param qed_devlink_params[] = {
 
 static const struct devlink_ops qed_dl_ops;
 
-int qed_devlink_register(struct qed_dev *cdev)
+struct devlink *qed_devlink_register(struct qed_dev *cdev)
 {
 	union devlink_param_value value;
-	struct qed_devlink *qed_dl;
+	struct qed_devlink *qdevlink;
 	struct devlink *dl;
 	int rc;
 
-	dl = devlink_alloc(&qed_dl_ops, sizeof(*qed_dl));
+	dl = devlink_alloc(&qed_dl_ops, sizeof(struct qed_devlink));
 	if (!dl)
-		return -ENOMEM;
+		return ERR_PTR(-ENOMEM);
 
-	qed_dl = devlink_priv(dl);
-
-	cdev->dl = dl;
-	qed_dl->cdev = cdev;
+	qdevlink = devlink_priv(dl);
+	qdevlink->cdev = cdev;
 
 	rc = devlink_register(dl, &cdev->pdev->dev);
 	if (rc)
@@ -85,26 +78,25 @@ int qed_devlink_register(struct qed_dev *cdev)
 	devlink_params_publish(dl);
 	cdev->iwarp_cmt = false;
 
-	return 0;
+	return dl;
 
 err_unregister:
 	devlink_unregister(dl);
 
 err_free:
-	cdev->dl = NULL;
 	devlink_free(dl);
 
-	return rc;
+	return ERR_PTR(rc);
 }
 
-void qed_devlink_unregister(struct qed_dev *cdev)
+void qed_devlink_unregister(struct devlink *devlink)
 {
-	if (!cdev->dl)
+	if (!devlink)
 		return;
 
-	devlink_params_unregister(cdev->dl, qed_devlink_params,
+	devlink_params_unregister(devlink, qed_devlink_params,
 				  ARRAY_SIZE(qed_devlink_params));
 
-	devlink_unregister(cdev->dl);
-	devlink_free(cdev->dl);
+	devlink_unregister(devlink);
+	devlink_free(devlink);
 }
diff --git a/drivers/net/ethernet/qlogic/qed/qed_devlink.h b/drivers/net/ethernet/qlogic/qed/qed_devlink.h
index b94c40e9b7c1..c79dc6bfa194 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_devlink.h
+++ b/drivers/net/ethernet/qlogic/qed/qed_devlink.h
@@ -9,7 +9,7 @@
 #include <linux/qed/qed_if.h>
 #include <net/devlink.h>
 
-int qed_devlink_register(struct qed_dev *cdev);
-void qed_devlink_unregister(struct qed_dev *cdev);
+struct devlink *qed_devlink_register(struct qed_dev *cdev);
+void qed_devlink_unregister(struct devlink *devlink);
 
 #endif
diff --git a/drivers/net/ethernet/qlogic/qed/qed_main.c b/drivers/net/ethernet/qlogic/qed/qed_main.c
index 8751355d9ef7..d6f76421379b 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_main.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_main.c
@@ -539,12 +539,6 @@ static struct qed_dev *qed_probe(struct pci_dev *pdev,
 	}
 	DP_INFO(cdev, "PCI init completed successfully\n");
 
-	rc = qed_devlink_register(cdev);
-	if (rc) {
-		DP_INFO(cdev, "Failed to register devlink.\n");
-		goto err2;
-	}
-
 	rc = qed_hw_prepare(cdev, QED_PCI_DEFAULT);
 	if (rc) {
 		DP_ERR(cdev, "hw prepare failed\n");
@@ -574,8 +568,6 @@ static void qed_remove(struct qed_dev *cdev)
 
 	qed_set_power_state(cdev, PCI_D3hot);
 
-	qed_devlink_unregister(cdev);
-
 	qed_free_cdev(cdev);
 }
 
@@ -3012,6 +3004,8 @@ const struct qed_common_ops qed_common_ops_pass = {
 	.get_link = &qed_get_current_link,
 	.drain = &qed_drain,
 	.update_msglvl = &qed_init_dp,
+	.devlink_register = qed_devlink_register,
+	.devlink_unregister = qed_devlink_unregister,
 	.dbg_all_data = &qed_dbg_all_data,
 	.dbg_all_data_size = &qed_dbg_all_data_size,
 	.chain_alloc = &qed_chain_alloc,
diff --git a/drivers/net/ethernet/qlogic/qede/qede.h b/drivers/net/ethernet/qlogic/qede/qede.h
index 803c1fcca8ad..1f0e7505a973 100644
--- a/drivers/net/ethernet/qlogic/qede/qede.h
+++ b/drivers/net/ethernet/qlogic/qede/qede.h
@@ -172,6 +172,7 @@ struct qede_dev {
 	struct qed_dev			*cdev;
 	struct net_device		*ndev;
 	struct pci_dev			*pdev;
+	struct devlink			*devlink;
 
 	u32				dp_module;
 	u8				dp_level;
diff --git a/drivers/net/ethernet/qlogic/qede/qede_main.c b/drivers/net/ethernet/qlogic/qede/qede_main.c
index 1aaae3203f5a..7c2d948b2035 100644
--- a/drivers/net/ethernet/qlogic/qede/qede_main.c
+++ b/drivers/net/ethernet/qlogic/qede/qede_main.c
@@ -1172,10 +1172,23 @@ static int __qede_probe(struct pci_dev *pdev, u32 dp_module, u8 dp_level,
 			rc = -ENOMEM;
 			goto err2;
 		}
+
+		edev->devlink = qed_ops->common->devlink_register(cdev);
+		if (IS_ERR(edev->devlink)) {
+			DP_NOTICE(edev, "Cannot register devlink\n");
+			edev->devlink = NULL;
+			/* Go on, we can live without devlink */
+		}
 	} else {
 		struct net_device *ndev = pci_get_drvdata(pdev);
 
 		edev = netdev_priv(ndev);
+
+		if (edev && edev->devlink) {
+			struct qed_devlink *qdl = devlink_priv(edev->devlink);
+
+			qdl->cdev = cdev;
+		}
 		edev->cdev = cdev;
 		memset(&edev->stats, 0, sizeof(edev->stats));
 		memcpy(&edev->dev_info, &dev_info, sizeof(dev_info));
@@ -1298,6 +1311,11 @@ static void __qede_remove(struct pci_dev *pdev, enum qede_remove_mode mode)
 	qed_ops->common->slowpath_stop(cdev);
 	if (system_state == SYSTEM_POWER_OFF)
 		return;
+
+	if (mode != QEDE_REMOVE_RECOVERY && edev->devlink) {
+		qed_ops->common->devlink_unregister(edev->devlink);
+		edev->devlink = NULL;
+	}
 	qed_ops->common->remove(cdev);
 	edev->cdev = NULL;
 
diff --git a/include/linux/qed/qed_if.h b/include/linux/qed/qed_if.h
index cd6a5c7e56eb..d8368e1770df 100644
--- a/include/linux/qed/qed_if.h
+++ b/include/linux/qed/qed_if.h
@@ -21,6 +21,7 @@
 #include <linux/qed/common_hsi.h>
 #include <linux/qed/qed_chain.h>
 #include <linux/io-64-nonatomic-lo-hi.h>
+#include <net/devlink.h>
 
 enum dcbx_protocol_type {
 	DCBX_PROTOCOL_ISCSI,
@@ -779,6 +780,10 @@ enum qed_nvm_flash_cmd {
 	QED_NVM_FLASH_CMD_NVM_MAX,
 };
 
+struct qed_devlink {
+	struct qed_dev *cdev;
+};
+
 struct qed_common_cb_ops {
 	void (*arfs_filter_op)(void *dev, void *fltr, u8 fw_rc);
 	void (*link_update)(void *dev, struct qed_link_output *link);
@@ -1137,6 +1142,10 @@ struct qed_common_ops {
  *
  */
 	int (*set_grc_config)(struct qed_dev *cdev, u32 cfg_id, u32 val);
+
+	struct devlink* (*devlink_register)(struct qed_dev *cdev);
+
+	void (*devlink_unregister)(struct devlink *devlink);
 };
 
 #define MASK_FIELD(_name, _value) \
-- 
2.17.1

[PATCH v5 net-next 03/10] qed: fix kconfig help entries

[Igor Russkikh]

This patch replaces stubs in kconfig help entries with an actual description.

Signed-off-by: Igor Russkikh <irusskikh@marvell.com>
Signed-off-by: Alexander Lobakin <alobakin@marvell.com>
Signed-off-by: Michal Kalderon <michal.kalderon@marvell.com>
---
 drivers/net/ethernet/qlogic/Kconfig | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/qlogic/Kconfig b/drivers/net/ethernet/qlogic/Kconfig
index 8f743d80760b..4366c7a8de95 100644
--- a/drivers/net/ethernet/qlogic/Kconfig
+++ b/drivers/net/ethernet/qlogic/Kconfig
@@ -80,7 +80,7 @@ config QED
 	select CRC8
 	select NET_DEVLINK
 	help
-	  This enables the support for ...
+	  This enables the support for Marvell FastLinQ adapters family.
 
 config QED_LL2
 	bool
@@ -100,7 +100,8 @@ config QEDE
 	depends on QED
 	imply PTP_1588_CLOCK
 	help
-	  This enables the support for ...
+	  This enables the support for Marvell FastLinQ adapters family,
+	  ethernet driver.
 
 config QED_RDMA
 	bool
-- 
2.17.1

[PATCH v5 net-next 04/10] qed: implement devlink info request

[Igor Russkikh]

Here we return existing fw & mfw versions, we also fetch device's
serial number:

~$ sudo ~/iproute2/devlink/devlink  dev info
pci/0000:01:00.0:
  driver qed
  serial_number REE1915E44552
  versions:
      running:
        fw 8.42.2.0
      stored:
        fw.mgmt 8.52.10.0
pci/0000:01:00.1:
  driver qed
  serial_number REE1915E44552
  versions:
      running:
        fw 8.42.2.0
      stored:
        fw.mgmt 8.52.10.0

MFW and FW are different firmwares on device.
Management is a firmware responsible for link configuration and
various control plane features. Its permanent and resides in NVM.

Running FW (or fastpath FW) is an embedded microprogram implementing
all the packet processing, offloads, etc. This FW is being loaded
on each start by the driver from FW binary blob.

The base device specific structure (qed_dev_info) was not directly
available to the base driver before. Thus, here we create and store
a private copy of this structure in qed_dev root object to
access the data.

Signed-off-by: Igor Russkikh <irusskikh@marvell.com>
Signed-off-by: Alexander Lobakin <alobakin@marvell.com>
Signed-off-by: Michal Kalderon <michal.kalderon@marvell.com>
---
 drivers/net/ethernet/qlogic/qed/qed.h         |  1 +
 drivers/net/ethernet/qlogic/qed/qed_dev.c     |  9 ++++
 drivers/net/ethernet/qlogic/qed/qed_devlink.c | 50 ++++++++++++++++++-
 drivers/net/ethernet/qlogic/qed/qed_main.c    |  1 +
 4 files changed, 60 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/qlogic/qed/qed.h b/drivers/net/ethernet/qlogic/qed/qed.h
index b6ce1488abcc..ccd789eeda3e 100644
--- a/drivers/net/ethernet/qlogic/qed/qed.h
+++ b/drivers/net/ethernet/qlogic/qed/qed.h
@@ -807,6 +807,7 @@ struct qed_dev {
 	struct qed_llh_info *p_llh_info;
 
 	/* Linux specific here */
+	struct qed_dev_info		common_dev_info;
 	struct  qede_dev		*edev;
 	struct  pci_dev			*pdev;
 	u32 flags;
diff --git a/drivers/net/ethernet/qlogic/qed/qed_dev.c b/drivers/net/ethernet/qlogic/qed/qed_dev.c
index b3c9ebaf2280..00f2d7f13de6 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_dev.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_dev.c
@@ -3973,6 +3973,7 @@ static int qed_hw_get_nvm_info(struct qed_hwfn *p_hwfn, struct qed_ptt *p_ptt)
 	struct qed_mcp_link_speed_params *ext_speed;
 	struct qed_mcp_link_capabilities *p_caps;
 	struct qed_mcp_link_params *link;
+	int i;
 
 	/* Read global nvm_cfg address */
 	nvm_cfg_addr = qed_rd(p_hwfn, p_ptt, MISC_REG_GEN_PURP_CR0);
@@ -4290,6 +4291,14 @@ static int qed_hw_get_nvm_info(struct qed_hwfn *p_hwfn, struct qed_ptt *p_ptt)
 		__set_bit(QED_DEV_CAP_ROCE,
 			  &p_hwfn->hw_info.device_capabilities);
 
+	/* Read device serial number information from shmem */
+	addr = MCP_REG_SCRATCH + nvm_cfg1_offset +
+		offsetof(struct nvm_cfg1, glob) +
+		offsetof(struct nvm_cfg1_glob, serial_number);
+
+	for (i = 0; i < 4; i++)
+		p_hwfn->hw_info.part_num[i] = qed_rd(p_hwfn, p_ptt, addr + i * 4);
+
 	return qed_mcp_fill_shmem_func_info(p_hwfn, p_ptt);
 }
 
diff --git a/drivers/net/ethernet/qlogic/qed/qed_devlink.c b/drivers/net/ethernet/qlogic/qed/qed_devlink.c
index a62c47c61edf..57ef2c56c884 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_devlink.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_devlink.c
@@ -45,7 +45,55 @@ static const struct devlink_param qed_devlink_params[] = {
 			     qed_dl_param_get, qed_dl_param_set, NULL),
 };
 
-static const struct devlink_ops qed_dl_ops;
+static int qed_devlink_info_get(struct devlink *devlink,
+				struct devlink_info_req *req,
+				struct netlink_ext_ack *extack)
+{
+	struct qed_devlink *qed_dl = devlink_priv(devlink);
+	struct qed_dev *cdev = qed_dl->cdev;
+	struct qed_dev_info *dev_info;
+	char buf[100];
+	int err;
+
+	dev_info = &cdev->common_dev_info;
+
+	err = devlink_info_driver_name_put(req, KBUILD_MODNAME);
+	if (err)
+		return err;
+
+	memcpy(buf, cdev->hwfns[0].hw_info.part_num, sizeof(cdev->hwfns[0].hw_info.part_num));
+	buf[sizeof(cdev->hwfns[0].hw_info.part_num)] = 0;
+
+	if (buf[0]) {
+		err = devlink_info_serial_number_put(req, buf);
+		if (err)
+			return err;
+	}
+
+	snprintf(buf, sizeof(buf), "%d.%d.%d.%d",
+		 GET_MFW_FIELD(dev_info->mfw_rev, QED_MFW_VERSION_3),
+		 GET_MFW_FIELD(dev_info->mfw_rev, QED_MFW_VERSION_2),
+		 GET_MFW_FIELD(dev_info->mfw_rev, QED_MFW_VERSION_1),
+		 GET_MFW_FIELD(dev_info->mfw_rev, QED_MFW_VERSION_0));
+
+	err = devlink_info_version_stored_put(req,
+					      DEVLINK_INFO_VERSION_GENERIC_FW_MGMT, buf);
+	if (err)
+		return err;
+
+	snprintf(buf, sizeof(buf), "%d.%d.%d.%d",
+		 dev_info->fw_major,
+		 dev_info->fw_minor,
+		 dev_info->fw_rev,
+		 dev_info->fw_eng);
+
+	return devlink_info_version_running_put(req,
+						DEVLINK_INFO_VERSION_GENERIC_FW, buf);
+}
+
+static const struct devlink_ops qed_dl_ops = {
+	.info_get = qed_devlink_info_get,
+};
 
 struct devlink *qed_devlink_register(struct qed_dev *cdev)
 {
diff --git a/drivers/net/ethernet/qlogic/qed/qed_main.c b/drivers/net/ethernet/qlogic/qed/qed_main.c
index d6f76421379b..d1a559ccf516 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_main.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_main.c
@@ -479,6 +479,7 @@ int qed_fill_dev_info(struct qed_dev *cdev,
 	}
 
 	dev_info->mtu = hw_info->mtu;
+	cdev->common_dev_info = *dev_info;
 
 	return 0;
 }
-- 
2.17.1

[PATCH v5 net-next 05/10] qed: health reporter init deinit seq

[Igor Russkikh]

Here we declare health reporter ops (empty for now)
and register these in qed probe and remove callbacks.

This way we get devlink attached to all kind of qed* PCI
device entities: networking or storage offload entity.

Signed-off-by: Igor Russkikh <irusskikh@marvell.com>
Signed-off-by: Alexander Lobakin <alobakin@marvell.com>
Signed-off-by: Michal Kalderon <michal.kalderon@marvell.com>
---
 drivers/net/ethernet/qlogic/qed/qed_devlink.c | 34 +++++++++++++++++++
 drivers/net/ethernet/qlogic/qed/qed_devlink.h |  3 ++
 include/linux/qed/qed_if.h                    |  1 +
 3 files changed, 38 insertions(+)

diff --git a/drivers/net/ethernet/qlogic/qed/qed_devlink.c b/drivers/net/ethernet/qlogic/qed/qed_devlink.c
index 57ef2c56c884..8d9150ecb580 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_devlink.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_devlink.c
@@ -14,6 +14,36 @@ enum qed_devlink_param_id {
 	QED_DEVLINK_PARAM_ID_IWARP_CMT,
 };
 
+static const struct devlink_health_reporter_ops qed_fw_fatal_reporter_ops = {
+		.name = "fw_fatal",
+};
+
+#define QED_REPORTER_FW_GRACEFUL_PERIOD 1200000
+
+void qed_fw_reporters_create(struct devlink *devlink)
+{
+	struct qed_devlink *dl = devlink_priv(devlink);
+
+	dl->fw_reporter = devlink_health_reporter_create(devlink, &qed_fw_fatal_reporter_ops,
+							 QED_REPORTER_FW_GRACEFUL_PERIOD, dl);
+	if (IS_ERR(dl->fw_reporter)) {
+		DP_NOTICE(dl->cdev, "Failed to create fw reporter, err = %ld\n",
+			  PTR_ERR(dl->fw_reporter));
+		dl->fw_reporter = NULL;
+	}
+}
+
+void qed_fw_reporters_destroy(struct devlink *devlink)
+{
+	struct qed_devlink *dl = devlink_priv(devlink);
+	struct devlink_health_reporter *rep;
+
+	rep = dl->fw_reporter;
+
+	if (!IS_ERR_OR_NULL(rep))
+		devlink_health_reporter_destroy(rep);
+}
+
 static int qed_dl_param_get(struct devlink *dl, u32 id,
 			    struct devlink_param_gset_ctx *ctx)
 {
@@ -126,6 +156,8 @@ struct devlink *qed_devlink_register(struct qed_dev *cdev)
 	devlink_params_publish(dl);
 	cdev->iwarp_cmt = false;
 
+	qed_fw_reporters_create(dl);
+
 	return dl;
 
 err_unregister:
@@ -142,6 +174,8 @@ void qed_devlink_unregister(struct devlink *devlink)
 	if (!devlink)
 		return;
 
+	qed_fw_reporters_destroy(devlink);
+
 	devlink_params_unregister(devlink, qed_devlink_params,
 				  ARRAY_SIZE(qed_devlink_params));
 
diff --git a/drivers/net/ethernet/qlogic/qed/qed_devlink.h b/drivers/net/ethernet/qlogic/qed/qed_devlink.h
index c79dc6bfa194..c68ecf778826 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_devlink.h
+++ b/drivers/net/ethernet/qlogic/qed/qed_devlink.h
@@ -12,4 +12,7 @@
 struct devlink *qed_devlink_register(struct qed_dev *cdev);
 void qed_devlink_unregister(struct devlink *devlink);
 
+void qed_fw_reporters_create(struct devlink *devlink);
+void qed_fw_reporters_destroy(struct devlink *devlink);
+
 #endif
diff --git a/include/linux/qed/qed_if.h b/include/linux/qed/qed_if.h
index d8368e1770df..30fe06fe06a0 100644
--- a/include/linux/qed/qed_if.h
+++ b/include/linux/qed/qed_if.h
@@ -782,6 +782,7 @@ enum qed_nvm_flash_cmd {
 
 struct qed_devlink {
 	struct qed_dev *cdev;
+	struct devlink_health_reporter *fw_reporter;
 };
 
 struct qed_common_cb_ops {
-- 
2.17.1

[PATCH v5 net-next 06/10] qed: use devlink logic to report errors

[Igor Russkikh]

Use devlink_health_report to push error indications.
We implement this in qede via callback function to make it possible
to reuse the same for other drivers sitting on top of qed in future.

Signed-off-by: Igor Russkikh <irusskikh@marvell.com>
Signed-off-by: Alexander Lobakin <alobakin@marvell.com>
Signed-off-by: Michal Kalderon <michal.kalderon@marvell.com>
---
 drivers/net/ethernet/qlogic/qed/qed_devlink.c | 18 ++++++++++++++++++
 drivers/net/ethernet/qlogic/qed/qed_devlink.h |  2 ++
 drivers/net/ethernet/qlogic/qed/qed_main.c    |  1 +
 drivers/net/ethernet/qlogic/qede/qede.h       |  1 +
 drivers/net/ethernet/qlogic/qede/qede_main.c  |  4 ++++
 include/linux/qed/qed_if.h                    |  3 +++
 6 files changed, 29 insertions(+)

diff --git a/drivers/net/ethernet/qlogic/qed/qed_devlink.c b/drivers/net/ethernet/qlogic/qed/qed_devlink.c
index 8d9150ecb580..e81bd3b39149 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_devlink.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_devlink.c
@@ -14,6 +14,24 @@ enum qed_devlink_param_id {
 	QED_DEVLINK_PARAM_ID_IWARP_CMT,
 };
 
+struct qed_fw_fatal_ctx {
+	enum qed_hw_err_type err_type;
+};
+
+int qed_report_fatal_error(struct devlink *devlink, enum qed_hw_err_type err_type)
+{
+	struct qed_devlink *qdl = devlink_priv(devlink);
+	struct qed_fw_fatal_ctx fw_fatal_ctx = {
+		.err_type = err_type,
+	};
+
+	if (qdl->fw_reporter)
+		devlink_health_report(qdl->fw_reporter,
+				      "Fatal error occurred", &fw_fatal_ctx);
+
+	return 0;
+}
+
 static const struct devlink_health_reporter_ops qed_fw_fatal_reporter_ops = {
 		.name = "fw_fatal",
 };
diff --git a/drivers/net/ethernet/qlogic/qed/qed_devlink.h b/drivers/net/ethernet/qlogic/qed/qed_devlink.h
index c68ecf778826..ccc7d1d1bfd4 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_devlink.h
+++ b/drivers/net/ethernet/qlogic/qed/qed_devlink.h
@@ -15,4 +15,6 @@ void qed_devlink_unregister(struct devlink *devlink);
 void qed_fw_reporters_create(struct devlink *devlink);
 void qed_fw_reporters_destroy(struct devlink *devlink);
 
+int qed_report_fatal_error(struct devlink *dl, enum qed_hw_err_type err_type);
+
 #endif
diff --git a/drivers/net/ethernet/qlogic/qed/qed_main.c b/drivers/net/ethernet/qlogic/qed/qed_main.c
index d1a559ccf516..a64d594f9294 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_main.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_main.c
@@ -3007,6 +3007,7 @@ const struct qed_common_ops qed_common_ops_pass = {
 	.update_msglvl = &qed_init_dp,
 	.devlink_register = qed_devlink_register,
 	.devlink_unregister = qed_devlink_unregister,
+	.report_fatal_error = qed_report_fatal_error,
 	.dbg_all_data = &qed_dbg_all_data,
 	.dbg_all_data_size = &qed_dbg_all_data_size,
 	.chain_alloc = &qed_chain_alloc,
diff --git a/drivers/net/ethernet/qlogic/qede/qede.h b/drivers/net/ethernet/qlogic/qede/qede.h
index 1f0e7505a973..3efc5899f656 100644
--- a/drivers/net/ethernet/qlogic/qede/qede.h
+++ b/drivers/net/ethernet/qlogic/qede/qede.h
@@ -264,6 +264,7 @@ struct qede_dev {
 
 	struct bpf_prog			*xdp_prog;
 
+	enum qed_hw_err_type		last_err_type;
 	unsigned long			err_flags;
 #define QEDE_ERR_IS_HANDLED		31
 #define QEDE_ERR_ATTN_CLR_EN		0
diff --git a/drivers/net/ethernet/qlogic/qede/qede_main.c b/drivers/net/ethernet/qlogic/qede/qede_main.c
index 7c2d948b2035..9895affa5064 100644
--- a/drivers/net/ethernet/qlogic/qede/qede_main.c
+++ b/drivers/net/ethernet/qlogic/qede/qede_main.c
@@ -2603,6 +2603,9 @@ static void qede_generic_hw_err_handler(struct qede_dev *edev)
 		  "Generic sleepable HW error handling started - err_flags 0x%lx\n",
 		  edev->err_flags);
 
+	if (edev->devlink)
+		edev->ops->common->report_fatal_error(edev->devlink, edev->last_err_type);
+
 	/* Trigger a recovery process.
 	 * This is placed in the sleep requiring section just to make
 	 * sure it is the last one, and that all the other operations
@@ -2663,6 +2666,7 @@ static void qede_schedule_hw_err_handler(void *dev,
 		return;
 	}
 
+	edev->last_err_type = err_type;
 	qede_set_hw_err_flags(edev, err_type);
 	qede_atomic_hw_err_handler(edev);
 	set_bit(QEDE_SP_HW_ERR, &edev->sp_flags);
diff --git a/include/linux/qed/qed_if.h b/include/linux/qed/qed_if.h
index 30fe06fe06a0..1297726f2b25 100644
--- a/include/linux/qed/qed_if.h
+++ b/include/linux/qed/qed_if.h
@@ -906,6 +906,9 @@ struct qed_common_ops {
 
 	int (*dbg_all_data_size) (struct qed_dev *cdev);
 
+	int		(*report_fatal_error)(struct devlink *devlink,
+					      enum qed_hw_err_type err_type);
+
 /**
  * @brief can_link_change - can the instance change the link or not
  *
-- 
2.17.1

[PATCH v5 net-next 07/10] qed*: make use of devlink recovery infrastructure

[Igor Russkikh]

Remove forcible recovery trigger and put it as a normal devlink
callback.

This allows user to enable/disable it via

    devlink health set pci/0000:03:00.0 reporter fw_fatal auto_recover false

Signed-off-by: Igor Russkikh <irusskikh@marvell.com>
Signed-off-by: Alexander Lobakin <alobakin@marvell.com>
Signed-off-by: Michal Kalderon <michal.kalderon@marvell.com>
---
 drivers/net/ethernet/qlogic/qed/qed.h         |  1 +
 drivers/net/ethernet/qlogic/qed/qed_devlink.c | 14 ++++++++++++++
 drivers/net/ethernet/qlogic/qed/qed_main.c    |  2 +-
 drivers/net/ethernet/qlogic/qede/qede_main.c  | 10 ----------
 4 files changed, 16 insertions(+), 11 deletions(-)

diff --git a/drivers/net/ethernet/qlogic/qed/qed.h b/drivers/net/ethernet/qlogic/qed/qed.h
index ccd789eeda3e..f34b25a79449 100644
--- a/drivers/net/ethernet/qlogic/qed/qed.h
+++ b/drivers/net/ethernet/qlogic/qed/qed.h
@@ -981,6 +981,7 @@ void qed_bw_update(struct qed_hwfn *hwfn, struct qed_ptt *ptt);
 u32 qed_unzip_data(struct qed_hwfn *p_hwfn,
 		   u32 input_len, u8 *input_buf,
 		   u32 max_size, u8 *unzip_buf);
+int qed_recovery_process(struct qed_dev *cdev);
 void qed_schedule_recovery_handler(struct qed_hwfn *p_hwfn);
 void qed_hw_error_occurred(struct qed_hwfn *p_hwfn,
 			   enum qed_hw_err_type err_type);
diff --git a/drivers/net/ethernet/qlogic/qed/qed_devlink.c b/drivers/net/ethernet/qlogic/qed/qed_devlink.c
index e81bd3b39149..1cf2199a3124 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_devlink.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_devlink.c
@@ -32,8 +32,22 @@ int qed_report_fatal_error(struct devlink *devlink, enum qed_hw_err_type err_typ
 	return 0;
 }
 
+static int
+qed_fw_fatal_reporter_recover(struct devlink_health_reporter *reporter,
+			      void *priv_ctx,
+			      struct netlink_ext_ack *extack)
+{
+	struct qed_devlink *qdl = devlink_health_reporter_priv(reporter);
+	struct qed_dev *cdev = qdl->cdev;
+
+	qed_recovery_process(cdev);
+
+	return 0;
+}
+
 static const struct devlink_health_reporter_ops qed_fw_fatal_reporter_ops = {
 		.name = "fw_fatal",
+		.recover = qed_fw_fatal_reporter_recover,
 };
 
 #define QED_REPORTER_FW_GRACEFUL_PERIOD 1200000
diff --git a/drivers/net/ethernet/qlogic/qed/qed_main.c b/drivers/net/ethernet/qlogic/qed/qed_main.c
index a64d594f9294..db5d003770ba 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_main.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_main.c
@@ -2817,7 +2817,7 @@ static int qed_set_led(struct qed_dev *cdev, enum qed_led_mode mode)
 	return status;
 }
 
-static int qed_recovery_process(struct qed_dev *cdev)
+int qed_recovery_process(struct qed_dev *cdev)
 {
 	struct qed_hwfn *p_hwfn = QED_LEADING_HWFN(cdev);
 	struct qed_ptt *p_ptt;
diff --git a/drivers/net/ethernet/qlogic/qede/qede_main.c b/drivers/net/ethernet/qlogic/qede/qede_main.c
index 9895affa5064..287e10effb49 100644
--- a/drivers/net/ethernet/qlogic/qede/qede_main.c
+++ b/drivers/net/ethernet/qlogic/qede/qede_main.c
@@ -2597,8 +2597,6 @@ static void qede_atomic_hw_err_handler(struct qede_dev *edev)
 
 static void qede_generic_hw_err_handler(struct qede_dev *edev)
 {
-	struct qed_dev *cdev = edev->cdev;
-
 	DP_NOTICE(edev,
 		  "Generic sleepable HW error handling started - err_flags 0x%lx\n",
 		  edev->err_flags);
@@ -2606,14 +2604,6 @@ static void qede_generic_hw_err_handler(struct qede_dev *edev)
 	if (edev->devlink)
 		edev->ops->common->report_fatal_error(edev->devlink, edev->last_err_type);
 
-	/* Trigger a recovery process.
-	 * This is placed in the sleep requiring section just to make
-	 * sure it is the last one, and that all the other operations
-	 * were completed.
-	 */
-	if (test_bit(QEDE_ERR_IS_RECOVERABLE, &edev->err_flags))
-		edev->ops->common->recovery_process(cdev);
-
 	clear_bit(QEDE_ERR_IS_HANDLED, &edev->err_flags);
 
 	DP_NOTICE(edev, "Generic sleepable HW error handling is done\n");
-- 
2.17.1

[PATCH v5 net-next 08/10] qed: implement devlink dump

[Igor Russkikh]

Gather and push out full device dump to devlink.
Device dump is the same as with `ethtool -d`, but now its generated
exactly at the moment bad thing happens.

Signed-off-by: Igor Russkikh <irusskikh@marvell.com>
Signed-off-by: Alexander Lobakin <alobakin@marvell.com>
Signed-off-by: Michal Kalderon <michal.kalderon@marvell.com>
---
 drivers/net/ethernet/qlogic/qed/qed_devlink.c | 43 +++++++++++++++++++
 1 file changed, 43 insertions(+)

diff --git a/drivers/net/ethernet/qlogic/qed/qed_devlink.c b/drivers/net/ethernet/qlogic/qed/qed_devlink.c
index 1cf2199a3124..4f6b1928edaf 100644
--- a/drivers/net/ethernet/qlogic/qed/qed_devlink.c
+++ b/drivers/net/ethernet/qlogic/qed/qed_devlink.c
@@ -6,6 +6,7 @@
 
 #include <linux/kernel.h>
 #include <linux/qed/qed_if.h>
+#include <linux/vmalloc.h>
 #include "qed.h"
 #include "qed_devlink.h"
 
@@ -32,6 +33,47 @@ int qed_report_fatal_error(struct devlink *devlink, enum qed_hw_err_type err_typ
 	return 0;
 }
 
+static int
+qed_fw_fatal_reporter_dump(struct devlink_health_reporter *reporter,
+			   struct devlink_fmsg *fmsg, void *priv_ctx,
+			   struct netlink_ext_ack *extack)
+{
+	struct qed_devlink *qdl = devlink_health_reporter_priv(reporter);
+	struct qed_fw_fatal_ctx *fw_fatal_ctx = priv_ctx;
+	struct qed_dev *cdev = qdl->cdev;
+	u32 dbg_data_buf_size;
+	u8 *p_dbg_data_buf;
+	int err;
+
+	/* Having context means that was a dump request after fatal,
+	 * so we enable extra debugging while gathering the dump,
+	 * just in case
+	 */
+	cdev->print_dbg_data = fw_fatal_ctx ? true : false;
+
+	dbg_data_buf_size = qed_dbg_all_data_size(cdev);
+	p_dbg_data_buf = vzalloc(dbg_data_buf_size);
+	if (!p_dbg_data_buf) {
+		DP_NOTICE(cdev,
+			  "Failed to allocate memory for a debug data buffer\n");
+		return -ENOMEM;
+	}
+
+	err = qed_dbg_all_data(cdev, p_dbg_data_buf);
+	if (err) {
+		DP_NOTICE(cdev, "Failed to obtain debug data\n");
+		vfree(p_dbg_data_buf);
+		return err;
+	}
+
+	err = devlink_fmsg_binary_pair_put(fmsg, "dump_data",
+					   p_dbg_data_buf, dbg_data_buf_size);
+
+	vfree(p_dbg_data_buf);
+
+	return err;
+}
+
 static int
 qed_fw_fatal_reporter_recover(struct devlink_health_reporter *reporter,
 			      void *priv_ctx,
@@ -48,6 +90,7 @@ qed_fw_fatal_reporter_recover(struct devlink_health_reporter *reporter,
 static const struct devlink_health_reporter_ops qed_fw_fatal_reporter_ops = {
 		.name = "fw_fatal",
 		.recover = qed_fw_fatal_reporter_recover,
+		.dump = qed_fw_fatal_reporter_dump,
 };
 
 #define QED_REPORTER_FW_GRACEFUL_PERIOD 1200000
-- 
2.17.1

[PATCH v5 net-next 09/10] qed: align adjacent indent

[Igor Russkikh]

Fix indent on some of adjacent declarations.

Signed-off-by: Igor Russkikh <irusskikh@marvell.com>
Signed-off-by: Alexander Lobakin <alobakin@marvell.com>
Signed-off-by: Michal Kalderon <michal.kalderon@marvell.com>
---
 include/linux/qed/qed_if.h | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/include/linux/qed/qed_if.h b/include/linux/qed/qed_if.h
index 1297726f2b25..b8fb80c9be80 100644
--- a/include/linux/qed/qed_if.h
+++ b/include/linux/qed/qed_if.h
@@ -897,14 +897,14 @@ struct qed_common_ops {
 
 	void		(*simd_handler_clean)(struct qed_dev *cdev,
 					      int index);
-	int (*dbg_grc)(struct qed_dev *cdev,
-		       void *buffer, u32 *num_dumped_bytes);
+	int		(*dbg_grc)(struct qed_dev *cdev,
+				   void *buffer, u32 *num_dumped_bytes);
 
-	int (*dbg_grc_size)(struct qed_dev *cdev);
+	int		(*dbg_grc_size)(struct qed_dev *cdev);
 
-	int (*dbg_all_data) (struct qed_dev *cdev, void *buffer);
+	int		(*dbg_all_data)(struct qed_dev *cdev, void *buffer);
 
-	int (*dbg_all_data_size) (struct qed_dev *cdev);
+	int		(*dbg_all_data_size)(struct qed_dev *cdev);
 
 	int		(*report_fatal_error)(struct devlink *devlink,
 					      enum qed_hw_err_type err_type);
-- 
2.17.1

[PATCH v5 net-next 10/10] qede: make driver reliable on unload after failures

[Igor Russkikh]

In case recovery was not successful, netdev still should be
present. But we should clear cdev if something bad happens
on recovery.

We also check cdev for null on dev close. That could be a case
if recovery was not successful.

Signed-off-by: Igor Russkikh <irusskikh@marvell.com>
Signed-off-by: Alexander Lobakin <alobakin@marvell.com>
Signed-off-by: Michal Kalderon <michal.kalderon@marvell.com>
---
 drivers/net/ethernet/qlogic/qede/qede_main.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/qlogic/qede/qede_main.c b/drivers/net/ethernet/qlogic/qede/qede_main.c
index 287e10effb49..01a7bff91d6c 100644
--- a/drivers/net/ethernet/qlogic/qede/qede_main.c
+++ b/drivers/net/ethernet/qlogic/qede/qede_main.c
@@ -1240,7 +1240,10 @@ static int __qede_probe(struct pci_dev *pdev, u32 dp_module, u8 dp_level,
 err4:
 	qede_rdma_dev_remove(edev, (mode == QEDE_PROBE_RECOVERY));
 err3:
-	free_netdev(edev->ndev);
+	if (mode != QEDE_PROBE_RECOVERY)
+		free_netdev(edev->ndev);
+	else
+		edev->cdev = NULL;
 err2:
 	qed_ops->common->slowpath_stop(cdev);
 err1:
@@ -2475,7 +2478,8 @@ static int qede_close(struct net_device *ndev)
 
 	qede_unload(edev, QEDE_UNLOAD_NORMAL, false);
 
-	edev->ops->common->update_drv_state(edev->cdev, false);
+	if (edev->cdev)
+		edev->ops->common->update_drv_state(edev->cdev, false);
 
 	return 0;
 }
-- 
2.17.1

Re: [PATCH v5 net-next 04/10] qed: implement devlink info request

[Jakub Kicinski]

On Sun, 2 Aug 2020 13:08:28 +0300 Igor Russkikh wrote:
> Here we return existing fw & mfw versions, we also fetch device's
> serial number:
> 
> ~$ sudo ~/iproute2/devlink/devlink  dev info
> pci/0000:01:00.0:
>   driver qed
>   serial_number REE1915E44552

Just to confirm - this is not:

   * - ``board.serial_number``
     - Board serial number of the device.

       This is usually the serial number of the board, often available in
       PCI *Vital Product Data*.

>   versions:
>       running:
>         fw 8.42.2.0
>       stored:
>         fw.mgmt 8.52.10.0
> pci/0000:01:00.1:
>   driver qed
>   serial_number REE1915E44552
>   versions:
>       running:
>         fw 8.42.2.0
>       stored:
>         fw.mgmt 8.52.10.0
> 
> MFW and FW are different firmwares on device.
> Management is a firmware responsible for link configuration and
> various control plane features. Its permanent and resides in NVM.
> 
> Running FW (or fastpath FW) is an embedded microprogram implementing
> all the packet processing, offloads, etc. This FW is being loaded
> on each start by the driver from FW binary blob.

Sounds like you should use:

fw.app - Data path microcode controlling high-speed packet processing.

Rather than:

fw - Overall firmware version, often representing the collection of fw.mgmt, fw.app, etc.

All quotes from devlink-info.rst in the kernel sources.

Re: [EXT] Re: [PATCH v5 net-next 04/10] qed: implement devlink info request

[Igor Russkikh]

Hi Jakub,

On 04/08/2020 12:29 am, Jakub Kicinski wrote:

>> ~$ sudo ~/iproute2/devlink/devlink  dev info
>> pci/0000:01:00.0:
>>   driver qed
>>   serial_number REE1915E44552
> 
> Just to confirm - this is not:
> 
>    * - ``board.serial_number``
>      - Board serial number of the device.
> 
>        This is usually the serial number of the board, often available in
>        PCI *Vital Product Data*.

I'm not sure on difference between board. or just serial_number, but lspci
shows this ID as a part of vital product data.

Considering this you are right, its better to change to "board.serial_number".

>> Running FW (or fastpath FW) is an embedded microprogram implementing
>> all the packet processing, offloads, etc. This FW is being loaded
>> on each start by the driver from FW binary blob.
> 
> Sounds like you should use:
> 
> fw.app - Data path microcode controlling high-speed packet processing.

Thanks, agreed.

Jakub, David, I'll prepare v6. Should I postpone posting v6 until the next
cycle then?

Regards,
  Igor