net: hns3: add a separate error handling task
authorJiaran Zhang <zhangjiaran@huawei.com>
Mon, 7 Jun 2021 11:18:10 +0000 (19:18 +0800)
committerDavid S. Miller <davem@davemloft.net>
Mon, 7 Jun 2021 21:00:37 +0000 (14:00 -0700)
Error handling and recovery logic are intertwined. Error handling (i.e.
error identification, clearing error sources and initiation of recovery)
is done in context of reset task. If certain hardware errors get
delivered during driver init time, which can cause driver init/loading
to fail.

Introduce a separate error handling task to ensure below:

1. Reset logic remains independent of the error handling logic.
2. Add the hclge_errhand_task_schedule to schedule error recovery
tasks, This will ensure that common misellaneous MSI-X interrupt are
re-enabled quickly.

Signed-off-by: Jiaran Zhang <zhangjiaran@huawei.com>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: Yufeng Mo <moyufeng@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_err.c
drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.h

index 8223d69..f125aa4 100644 (file)
@@ -1940,8 +1940,8 @@ int hclge_handle_hw_msix_error(struct hclge_dev *hdev,
 
        if (!test_bit(HCLGE_STATE_SERVICE_INITED, &hdev->state)) {
                dev_err(dev,
-                       "Can't handle - MSIx error reported during dev init\n");
-               return 0;
+                       "failed to handle msix error during dev init\n");
+               return -EAGAIN;
        }
 
        return hclge_handle_all_hw_msix_error(hdev, reset_requests);
index 6ecc106..8a431e1 100644 (file)
@@ -2843,6 +2843,14 @@ static void hclge_reset_task_schedule(struct hclge_dev *hdev)
                                    hclge_wq, &hdev->service_task, 0);
 }
 
+static void hclge_errhand_task_schedule(struct hclge_dev *hdev)
+{
+       if (!test_bit(HCLGE_STATE_REMOVING, &hdev->state) &&
+           !test_and_set_bit(HCLGE_STATE_ERR_SERVICE_SCHED, &hdev->state))
+               mod_delayed_work_on(cpumask_first(&hdev->affinity_mask),
+                                   hclge_wq, &hdev->service_task, 0);
+}
+
 void hclge_task_schedule(struct hclge_dev *hdev, unsigned long delay_time)
 {
        if (!test_bit(HCLGE_STATE_REMOVING, &hdev->state) &&
@@ -4264,6 +4272,36 @@ static void hclge_reset_subtask(struct hclge_dev *hdev)
        hdev->reset_type = HNAE3_NONE_RESET;
 }
 
+static void hclge_misc_err_recovery(struct hclge_dev *hdev)
+{
+       struct hnae3_ae_dev *ae_dev = pci_get_drvdata(hdev->pdev);
+       struct device *dev = &hdev->pdev->dev;
+       u32 msix_sts_reg;
+
+       msix_sts_reg = hclge_read_dev(&hdev->hw, HCLGE_MISC_VECTOR_INT_STS);
+
+       if (msix_sts_reg & HCLGE_VECTOR0_REG_MSIX_MASK) {
+               if (hclge_handle_hw_msix_error(hdev,
+                                              &hdev->default_reset_request))
+                       dev_info(dev, "received msix interrupt 0x%x\n",
+                                msix_sts_reg);
+
+               if (hdev->default_reset_request)
+                       if (ae_dev->ops->reset_event)
+                               ae_dev->ops->reset_event(hdev->pdev, NULL);
+       }
+
+       hclge_enable_vector(&hdev->misc_vector, true);
+}
+
+static void hclge_errhand_service_task(struct hclge_dev *hdev)
+{
+       if (!test_and_clear_bit(HCLGE_STATE_ERR_SERVICE_SCHED, &hdev->state))
+               return;
+
+       hclge_misc_err_recovery(hdev);
+}
+
 static void hclge_reset_service_task(struct hclge_dev *hdev)
 {
        if (!test_and_clear_bit(HCLGE_STATE_RST_SERVICE_SCHED, &hdev->state))
index 7595f84..9b8abb5 100644 (file)
@@ -221,6 +221,7 @@ enum HCLGE_DEV_STATE {
        HCLGE_STATE_RST_HANDLING,
        HCLGE_STATE_MBX_SERVICE_SCHED,
        HCLGE_STATE_MBX_HANDLING,
+       HCLGE_STATE_ERR_SERVICE_SCHED,
        HCLGE_STATE_STATISTICS_UPDATING,
        HCLGE_STATE_CMD_DISABLE,
        HCLGE_STATE_LINK_UPDATING,