net: hns3: dump more reg info based on ras mod

When the driver received an interrupte for hardware error,
it will try to restore by resetting. But the hardware registers
will also be reset at this case, which make it hard to analysis
why the hardware error occurs.

This patch dumps these registers before resetting to help
analyze the hardware error occurs.
Signed-off-by: Peiyang Wang <wangpeiyang1@huawei.com>
Signed-off-by: Jijie Shao <shaojijie@huawei.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://lore.kernel.org/r/20240410125354.2177067-4-shaojijie@huawei.comSigned-off-by: Jakub Kicinski <kuba@kernel.org>

drivers/net/ethernet/hisilicon/hns3/hnae3.h

@@ -104,6 +104,7 @@ enum HNAE3_DEV_CAP_BITS {
 	HNAE3_DEV_SUPPORT_WOL_B,
 	HNAE3_DEV_SUPPORT_TM_FLUSH_B,
 	HNAE3_DEV_SUPPORT_VF_FAULT_B,
+	HNAE3_DEV_SUPPORT_ERR_MOD_GEN_REG_B,
 };
 
 #define hnae3_ae_dev_fd_supported(ae_dev) \
@@ -181,6 +182,9 @@ enum HNAE3_DEV_CAP_BITS {
 #define hnae3_ae_dev_vf_fault_supported(ae_dev) \
 	test_bit(HNAE3_DEV_SUPPORT_VF_FAULT_B, (ae_dev)->caps)
 
+#define hnae3_ae_dev_gen_reg_dfx_supported(hdev) \
+	test_bit(HNAE3_DEV_SUPPORT_ERR_MOD_GEN_REG_B, (hdev)->ae_dev->caps)
+
 enum HNAE3_PF_CAP_BITS {
 	HNAE3_PF_SUPPORT_VLAN_FLTR_MDF_B = 0,
 };

drivers/net/ethernet/hisilicon/hns3/hns3_common/hclge_comm_cmd.c

@@ -158,6 +158,7 @@ static const struct hclge_comm_caps_bit_map hclge_pf_cmd_caps[] = {
 	{HCLGE_COMM_CAP_WOL_B, HNAE3_DEV_SUPPORT_WOL_B},
 	{HCLGE_COMM_CAP_TM_FLUSH_B, HNAE3_DEV_SUPPORT_TM_FLUSH_B},
 	{HCLGE_COMM_CAP_VF_FAULT_B, HNAE3_DEV_SUPPORT_VF_FAULT_B},
+	{HCLGE_COMM_CAP_ERR_MOD_GEN_REG_B, HNAE3_DEV_SUPPORT_ERR_MOD_GEN_REG_B},
 };
 
 static const struct hclge_comm_caps_bit_map hclge_vf_cmd_caps[] = {

drivers/net/ethernet/hisilicon/hns3/hns3_common/hclge_comm_cmd.h

@@ -91,6 +91,7 @@ enum hclge_opcode_type {
 	HCLGE_OPC_DFX_RCB_REG		= 0x004D,
 	HCLGE_OPC_DFX_TQP_REG		= 0x004E,
 	HCLGE_OPC_DFX_SSU_REG_2		= 0x004F,
+	HCLGE_OPC_DFX_GEN_REG		= 0x7038,
 
 	HCLGE_OPC_QUERY_DEV_SPECS	= 0x0050,
 	HCLGE_OPC_GET_QUEUE_ERR_VF      = 0x0067,
@@ -353,6 +354,7 @@ enum HCLGE_COMM_CAP_BITS {
 	HCLGE_COMM_CAP_LANE_NUM_B = 27,
 	HCLGE_COMM_CAP_WOL_B = 28,
 	HCLGE_COMM_CAP_TM_FLUSH_B = 31,
+	HCLGE_COMM_CAP_ERR_MOD_GEN_REG_B = 32,
 };
 
 enum HCLGE_COMM_API_CAP_BITS {

drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_debugfs.c

@@ -801,10 +801,8 @@ static int hclge_dbg_get_dfx_bd_num(struct hclge_dev *hdev, int offset,
 	return 0;
 }
 
-static int hclge_dbg_cmd_send(struct hclge_dev *hdev,
-			      struct hclge_desc *desc_src,
-			      int index, int bd_num,
-			      enum hclge_opcode_type cmd)
+int hclge_dbg_cmd_send(struct hclge_dev *hdev, struct hclge_desc *desc_src,
+		       int index, int bd_num, enum hclge_opcode_type cmd)
 {
 	struct hclge_desc *desc = desc_src;
 	int ret, i;

drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_debugfs.h

@@ -131,4 +131,7 @@ struct hclge_dbg_vlan_cfg {
 	u8 pri_only2;
 };
 
+int hclge_dbg_cmd_send(struct hclge_dev *hdev, struct hclge_desc *desc_src,
+		       int index, int bd_num, enum hclge_opcode_type cmd);
+
 #endif

drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_err.h

@@ -5,6 +5,7 @@
 #define __HCLGE_ERR_H
 
 #include "hclge_main.h"
+#include "hclge_debugfs.h"
 #include "hnae3.h"
 
 #define HCLGE_MPF_RAS_INT_MIN_BD_NUM	10
@@ -115,6 +116,18 @@
 #define HCLGE_REG_NUM_MAX			256
 #define HCLGE_DESC_NO_DATA_LEN			8
 
+#define HCLGE_BD_NUM_SSU_REG_0		10
+#define HCLGE_BD_NUM_SSU_REG_1		15
+#define HCLGE_BD_NUM_RPU_REG_0		1
+#define HCLGE_BD_NUM_RPU_REG_1		2
+#define HCLGE_BD_NUM_IGU_EGU_REG	9
+#define HCLGE_BD_NUM_GEN_REG		8
+#define HCLGE_MOD_REG_INFO_LEN_MAX	256
+#define HCLGE_MOD_REG_EXTRA_LEN		11
+#define HCLGE_MOD_REG_VALUE_LEN		9
+#define HCLGE_MOD_REG_GROUP_MAX_SIZE	6
+#define HCLGE_MOD_MSG_PARA_ARRAY_MAX_SIZE	8
+
 enum hclge_err_int_type {
 	HCLGE_ERR_INT_MSIX = 0,
 	HCLGE_ERR_INT_RAS_CE = 1,
@@ -191,6 +204,7 @@ struct hclge_hw_error {
 struct hclge_hw_module_id {
 	enum hclge_mod_name_list module_id;
 	const char *msg;
+	void (*query_reg_info)(struct hclge_dev *hdev);
 };
 
 struct hclge_hw_type_id {
@@ -218,6 +232,28 @@ struct hclge_type_reg_err_info {
 	u32 hclge_reg[HCLGE_REG_NUM_MAX];
 };
 
+struct hclge_mod_reg_info {
+	const char *reg_name;
+	bool has_suffix; /* add suffix for register name */
+	/* the positions of reg values in hclge_desc.data */
+	u8 reg_offset_group[HCLGE_MOD_REG_GROUP_MAX_SIZE];
+	u8 group_size;
+};
+
+/* This structure defines cmdq used to query the hardware module debug
+ * regisgers.
+ */
+struct hclge_mod_reg_common_msg {
+	enum hclge_opcode_type cmd;
+	struct hclge_desc *desc;
+	u8 bd_num; /* the bd number of hclge_desc used */
+	bool need_para; /* whether this cmdq needs to add para */
+
+	/* the regs need to print */
+	const struct hclge_mod_reg_info *result_regs;
+	u16 result_regs_size;
+};
+
 int hclge_config_mac_tnl_int(struct hclge_dev *hdev, bool en);
 int hclge_config_nic_hw_error(struct hclge_dev *hdev, bool state);
 int hclge_config_rocee_ras_interrupt(struct hclge_dev *hdev, bool en);