Skip to contentSkip to conversation

drm/amdgpu: Effective health check before reset

PATCH In review Ce Sun 2 participants 2 messages Jul 29, 2025

Subsystems (2)
lore
Share & tools
lore

Ce Sun

author
#0 lore raw

PATCH 1/1 drm/amdgpu: Effective health check before reset

Move amdgpu_device_health_check into amdgpu_device_gpu_recover to ensure that if the device is present can be checked before reset

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 25 +++++++---------------
 1 file changed, 8 insertions(+), 17 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +8 -17
index 2659e3ebbe49..176712225037 100644
@@ -6129,12 +6129,11 @@ static int amdgpu_device_health_check(struct list_head *device_list_handle)
 	return ret;
 }
 
-static int amdgpu_device_recovery_prepare(struct amdgpu_device *adev,
+static void amdgpu_device_recovery_prepare(struct amdgpu_device *adev,
 					  struct list_head *device_list,
 					  struct amdgpu_hive_info *hive)
 {
 	struct amdgpu_device *tmp_adev = NULL;
-	int r;
 
 	/*
 	 * Build list of devices to reset.
@@ -6155,13 +6154,6 @@ static int amdgpu_device_recovery_prepare(struct amdgpu_device *adev,
 		list_add_tail(&adev->reset_list, device_list);
 	}
 
-	if (!amdgpu_sriov_vf(adev) && (!adev->pcie_reset_ctx.occurs_dpc)) {
-		r = amdgpu_device_health_check(device_list);
-		if (r)
-			return r;
-	}
-
-	return 0;
 }
 
 static void amdgpu_device_recovery_get_reset_lock(struct amdgpu_device *adev,
@@ -6449,8 +6441,13 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 	reset_context->hive = hive;
 	INIT_LIST_HEAD(&device_list);
 
-	if (amdgpu_device_recovery_prepare(adev, &device_list, hive))
-		goto end_reset;
+	amdgpu_device_recovery_prepare(adev, &device_list, hive);
+
+	if (!amdgpu_sriov_vf(adev)) {
+		r = amdgpu_device_health_check(&device_list);
+		if (r)
+			goto end_reset;
+	}
 
 	/* We need to lock reset domain only once both for XGMI and single device */
 	amdgpu_device_recovery_get_reset_lock(adev, &device_list);
@@ -6956,12 +6953,6 @@ pci_ers_result_t amdgpu_pci_slot_reset(struct pci_dev *pdev)
 	int r = 0, i;
 	u32 memsize;
 
-	/* PCI error slot reset should be skipped During RAS recovery */
-	if ((amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
-	    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4)) &&
-	    amdgpu_ras_in_recovery(adev))
-		return PCI_ERS_RESULT_RECOVERED;
-
 	dev_info(adev->dev, "PCI error: slot reset callback!!\n");
 
 	memset(&reset_context, 0, sizeof(reset_context));
Signature · 1 line
2.34.1
#1 lore raw

[AMD Official Use Only - AMD Internal Distribution Only]

Quoted 52 lines

-----Original Message-----
From: Sun, Ce(Overlord) <[email protected]>
Sent: Tuesday, July 29, 2025 12:16 PM
To: [email protected]
Cc: Zhang, Hawking <[email protected]>; Zhou1, Tao <[email protected]>; Sun, Ce(Overlord) <[email protected]>
Subject: [PATCH] drm/amdgpu: Effective health check before reset

Move amdgpu_device_health_check into amdgpu_device_gpu_recover to ensure that if the device is present can be checked before reset

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 25 +++++++---------------
 1 file changed, 8 insertions(+), 17 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +1 -2
index 2659e3ebbe49..176712225037 100644
@@ -6129,12 +6129,11 @@ static int amdgpu_device_health_check(struct
 list_head *device_list_handle)
      return ret;
 }
 
-static int amdgpu_device_recovery_prepare(struct amdgpu_device *adev,
+static void amdgpu_device_recovery_prepare(struct amdgpu_device *adev,
                                        struct list_head *device_list,
                                        struct amdgpu_hive_info *hive)
 {
      struct amdgpu_device *tmp_adev = NULL;
-     int r;
 
      /*

* Build list of devices to reset.

drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +0 -7
@@ -6155,13 +6154,6 @@ static int amdgpu_device_recovery_prepare(struct
 amdgpu_device *adev,
              list_add_tail(&adev->reset_list, device_list);
      }
 
-     if (!amdgpu_sriov_vf(adev) && (!adev->pcie_reset_ctx.occurs_dpc)) {
-             r = amdgpu_device_health_check(device_list);
-             if (r)
-                     return r;
-     }
-
-     return 0;
 }
 

static void amdgpu_device_recovery_get_reset_lock(struct amdgpu_device *adev,

drivers/gpu/drm/amd/amdgpu/amdgpu_device.c int amdgpu_device_gpu_recover(struct amdgpu_device · line 6447

Quoted 9 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +3 -2
@@ -6449,8 +6441,13 @@ int amdgpu_device_gpu_recover(struct amdgpu_device
 *adev,
      reset_context->hive = hive;
      INIT_LIST_HEAD(&device_list);
 
-     if (amdgpu_device_recovery_prepare(adev, &device_list, hive))
-             goto end_reset;
+     amdgpu_device_recovery_prepare(adev, &device_list, hive);
+
+     if (!amdgpu_sriov_vf(adev)) {

[Tao] so the condition is also changed.
Please describe the issue you'd like to fix in more detail.

Quoted 22 lines
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +4 -0
+             r = amdgpu_device_health_check(&device_list);
+             if (r)
+                     goto end_reset;
+     }
 
      /* We need to lock reset domain only once both for XGMI and single device */
      amdgpu_device_recovery_get_reset_lock(adev, &device_list); @@ -6956,12
+6953,6 @@ pci_ers_result_t amdgpu_pci_slot_reset(struct pci_dev *pdev)
      int r = 0, i;
      u32 memsize;
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +0 -7
-     /* PCI error slot reset should be skipped During RAS recovery */
-     if ((amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
-         amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 4)) &&
-         amdgpu_ras_in_recovery(adev))
-             return PCI_ERS_RESULT_RECOVERED;
-
      dev_info(adev->dev, "PCI error: slot reset callback!!\n");
 
      memset(&reset_context, 0, sizeof(reset_context));
--

2.34.1