net: hns3: log detail error info of ROCEE ECC and AXI errors

driver inclusion
category: bugfix
bugzilla: NA
CVE: NA

This patch logs detail error info of ROCEE ECC and AXI errors for
debug purpose, and remove unnecessary resetting for ROCEE overflow
errors.

Feature or Bugfix: Feature
Signed-off-by: Ntanxiaofei <tanxiaofei@huawei.com>
Reviewed-by: Nliuyixian <liuyixian@huawei.com>
Reviewed-by: NYang Yingliang <yangyingliang@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

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

@@ -282,6 +282,8 @@ enum hclge_opcode_type {
 	HCLGE_CONFIG_ROCEE_RAS_INT_EN	= 0x1580,
 	HCLGE_QUERY_CLEAR_ROCEE_RAS_INT = 0x1581,
 	HCLGE_ROCEE_PF_RAS_INT_CMD	= 0x1584,
+	HCLGE_QUERY_ROCEE_ECC_RAS_INFO_CMD	= 0x1585,
+	HCLGE_QUERY_ROCEE_AXI_RAS_INFO_CMD	= 0x1586,
 	HCLGE_IGU_EGU_TNL_INT_EN	= 0x1803,
 	HCLGE_IGU_COMMON_INT_EN		= 0x1806,
 	HCLGE_TM_QCN_MEM_INT_CFG	= 0x1A14,

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

@@ -1344,6 +1344,64 @@ int hclge_handle_all_ras_errors(struct hclge_dev *hdev)
 	return ret;
 }
 
+static int hclge_log_rocee_axi_error(struct hclge_dev *hdev)
+{
+	struct device *dev = &hdev->pdev->dev;
+	struct hclge_desc desc[3];
+	int ret;
+
+	hclge_cmd_setup_basic_desc(&desc[0], HCLGE_QUERY_ROCEE_AXI_RAS_INFO_CMD,
+				   true);
+	hclge_cmd_setup_basic_desc(&desc[1], HCLGE_QUERY_ROCEE_AXI_RAS_INFO_CMD,
+				   true);
+	hclge_cmd_setup_basic_desc(&desc[2], HCLGE_QUERY_ROCEE_AXI_RAS_INFO_CMD,
+				   true);
+	desc[0].flag |= cpu_to_le16(HCLGE_CMD_FLAG_NEXT);
+	desc[1].flag |= cpu_to_le16(HCLGE_CMD_FLAG_NEXT);
+
+	ret = hclge_cmd_send(&hdev->hw, &desc[0], 3);
+	if (ret) {
+		dev_err(dev, "failed(%d) to query ROCEE AXI error sts\n", ret);
+		return ret;
+	}
+
+	dev_info(dev, "AXI1: %08X %08X %08X %08X %08X %08X\n",
+		 le32_to_cpu(desc[0].data[0]), le32_to_cpu(desc[0].data[1]),
+		 le32_to_cpu(desc[0].data[2]), le32_to_cpu(desc[0].data[3]),
+		 le32_to_cpu(desc[0].data[4]), le32_to_cpu(desc[0].data[5]));
+	dev_info(dev, "AXI2: %08X %08X %08X %08X %08X %08X\n",
+		 le32_to_cpu(desc[1].data[0]), le32_to_cpu(desc[1].data[1]),
+		 le32_to_cpu(desc[1].data[2]), le32_to_cpu(desc[1].data[3]),
+		 le32_to_cpu(desc[1].data[4]), le32_to_cpu(desc[1].data[5]));
+	dev_info(dev, "AXI3: %08X\n", le32_to_cpu(desc[2].data[0]));
+
+	return 0;
+}
+
+static int hclge_log_rocee_ecc_error(struct hclge_dev *hdev)
+{
+	struct device *dev = &hdev->pdev->dev;
+	struct hclge_desc desc[2];
+	int ret;
+
+	ret = hclge_cmd_query_error(hdev, &desc[0],
+				    HCLGE_QUERY_ROCEE_ECC_RAS_INFO_CMD,
+				    HCLGE_CMD_FLAG_NEXT, 0, 0);
+	if (ret) {
+		dev_err(dev, "failed(%d) to query ROCEE ECC error sts\n", ret);
+		return ret;
+	}
+
+	dev_info(dev, "ECC1: %08X %08X %08X %08X %08X %08X\n",
+		 le32_to_cpu(desc[0].data[0]), le32_to_cpu(desc[0].data[1]),
+		 le32_to_cpu(desc[0].data[2]), le32_to_cpu(desc[0].data[3]),
+		 le32_to_cpu(desc[0].data[4]), le32_to_cpu(desc[0].data[5]));
+	dev_info(dev, "ECC2: %08X %08X %08X\n", le32_to_cpu(desc[1].data[0]),
+		 le32_to_cpu(desc[1].data[1]), le32_to_cpu(desc[1].data[2]));
+
+	return 0;
+}
+
 static int hclge_log_rocee_ovf_error(struct hclge_dev *hdev)
 {
 	struct device *dev = &hdev->pdev->dev;
@@ -1411,20 +1469,18 @@ hclge_log_and_clear_rocee_ras_error(struct hclge_dev *hdev)
 
 	status = le32_to_cpu(desc[0].data[0]);
 
-	if (status & HCLGE_ROCEE_RERR_INT_MASK) {
+	if (status & HCLGE_ROCEE_AXI_ERR_INT_MASK) {
+		if (status & HCLGE_ROCEE_RERR_INT_MASK)
 			dev_warn(dev, "ROCEE RAS AXI rresp error\n");
-		reset_type = HNAE3_FUNC_RESET;
-	}
 
-	if (status & HCLGE_ROCEE_BERR_INT_MASK) {
+		if (status & HCLGE_ROCEE_BERR_INT_MASK)
 			dev_warn(dev, "ROCEE RAS AXI bresp error\n");
+
 		reset_type = HNAE3_FUNC_RESET;
-	}
 
-	if (status & HCLGE_ROCEE_OVF_INT_MASK) {
-		ret = hclge_log_rocee_ovf_error(hdev);
+		ret = hclge_log_rocee_axi_error(hdev);
 		if (ret) {
-			dev_err(dev, "failed(%d) to process ovf error\n", ret);
+			dev_err(dev, "failed(%d) to process axi error\n", ret);
 			/* reset everything for now */
 			return HNAE3_GLOBAL_RESET;
 		}
@@ -1433,6 +1489,22 @@ hclge_log_and_clear_rocee_ras_error(struct hclge_dev *hdev)
 	if (status & HCLGE_ROCEE_ECC_INT_MASK) {
 		dev_warn(dev, "ROCEE RAS 2bit ECC error\n");
 		reset_type = HNAE3_GLOBAL_RESET;
+
+		ret = hclge_log_rocee_ecc_error(hdev);
+		if (ret) {
+			dev_err(dev, "failed(%d) to process ecc error\n", ret);
+			/* reset everything for now */
+			return HNAE3_GLOBAL_RESET;
+		}
+	}
+
+	if (status & HCLGE_ROCEE_OVF_INT_MASK) {
+		ret = hclge_log_rocee_ovf_error(hdev);
+		if (ret) {
+			dev_err(dev, "failed(%d) to process ovf error\n", ret);
+			/* reset everything for now */
+			return HNAE3_GLOBAL_RESET;
+		}
 	}
 
 	/* clear error status */

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

@@ -94,6 +94,7 @@
 #define HCLGE_ROCEE_RAS_CE_INT_EN_MASK		0x1
 #define HCLGE_ROCEE_RERR_INT_MASK		BIT(0)
 #define HCLGE_ROCEE_BERR_INT_MASK		BIT(1)
+#define HCLGE_ROCEE_AXI_ERR_INT_MASK		GENMASK(1, 0)
 #define HCLGE_ROCEE_ECC_INT_MASK		BIT(2)
 #define HCLGE_ROCEE_OVF_INT_MASK		BIT(3)
 #define HCLGE_ROCEE_OVF_ERR_INT_MASK		0x10000