drm/amdgpu: make reset method configurable for RAS poison

Each RAS block has different requirement for gpu reset in poison
consumption handling.
Add support for mmhub RAS poison consumption handling.

v2: remove the mmhub poison support for kfd int v10.

Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

3 files changed, 28 insertions, 14 deletions

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v10.c b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v10.c
index 650da18b0d87..740f89aafbc0 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v10.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v10.c
@@ -134,6 +134,7 @@ static void event_interrupt_poison_consumption(struct kfd_node *dev,
 {
 	enum amdgpu_ras_block block = 0;
 	int old_poison, ret = -EINVAL;
+	uint32_t reset = 0;
 	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid);
 
 	if (!p)
@@ -153,6 +154,8 @@ static void event_interrupt_poison_consumption(struct kfd_node *dev,
 	case SOC15_IH_CLIENTID_UTCL2:
 		ret = kfd_dqm_evict_pasid(dev->dqm, pasid);
 		block = AMDGPU_RAS_BLOCK__GFX;
+		if (ret)
+			reset = AMDGPU_RAS_GPU_RESET_MODE2_RESET;
 		break;
 	case SOC15_IH_CLIENTID_SDMA0:
 	case SOC15_IH_CLIENTID_SDMA1:
@@ -160,6 +163,7 @@ static void event_interrupt_poison_consumption(struct kfd_node *dev,
 	case SOC15_IH_CLIENTID_SDMA3:
 	case SOC15_IH_CLIENTID_SDMA4:
 		block = AMDGPU_RAS_BLOCK__SDMA;
+		reset = AMDGPU_RAS_GPU_RESET_MODE2_RESET;
 		break;
 	default:
 		break;
@@ -170,17 +174,16 @@ static void event_interrupt_poison_consumption(struct kfd_node *dev,
 	/* resetting queue passes, do page retirement without gpu reset
 	 * resetting queue fails, fallback to gpu reset solution
 	 */
-	if (!ret) {
+	if (!ret)
 		dev_warn(dev->adev->dev,
 			"RAS poison consumption, unmap queue flow succeeded: client id %d\n",
 			client_id);
-		amdgpu_amdkfd_ras_poison_consumption_handler(dev->adev, block, false);
-	} else {
+	else
 		dev_warn(dev->adev->dev,
 			"RAS poison consumption, fall back to gpu reset flow: client id %d\n",
 			client_id);
-		amdgpu_amdkfd_ras_poison_consumption_handler(dev->adev, block, true);
-	}
+
+	amdgpu_amdkfd_ras_poison_consumption_handler(dev->adev, block, reset);
 }
 
 static bool event_interrupt_isr_v10(struct kfd_node *dev,
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v11.c b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v11.c
index 7e2859736a55..d3d6b5c180b3 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v11.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v11.c
@@ -193,6 +193,7 @@ static void event_interrupt_poison_consumption_v11(struct kfd_node *dev,
 {
 	enum amdgpu_ras_block block = 0;
 	int ret = -EINVAL;
+	uint32_t reset = 0;
 	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid);
 
 	if (!p)
@@ -212,10 +213,13 @@ static void event_interrupt_poison_consumption_v11(struct kfd_node *dev,
 		if (dev->dqm->ops.reset_queues)
 			ret = dev->dqm->ops.reset_queues(dev->dqm, pasid);
 		block = AMDGPU_RAS_BLOCK__GFX;
+		if (ret)
+			reset = AMDGPU_RAS_GPU_RESET_MODE2_RESET;
 		break;
 	case SOC21_INTSRC_SDMA_ECC:
 	default:
 		block = AMDGPU_RAS_BLOCK__GFX;
+		reset = AMDGPU_RAS_GPU_RESET_MODE2_RESET;
 		break;
 	}
 
@@ -223,10 +227,7 @@ static void event_interrupt_poison_consumption_v11(struct kfd_node *dev,
 
 	/* resetting queue passes, do page retirement without gpu reset
 	   resetting queue fails, fallback to gpu reset solution */
-	if (!ret)
-		amdgpu_amdkfd_ras_poison_consumption_handler(dev->adev, block, false);
-	else
-		amdgpu_amdkfd_ras_poison_consumption_handler(dev->adev, block, true);
+	amdgpu_amdkfd_ras_poison_consumption_handler(dev->adev, block, reset);
 }
 
 static bool event_interrupt_isr_v11(struct kfd_node *dev,
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c
index 11641f4645e6..2a37ab7a7150 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c
@@ -145,6 +145,7 @@ static void event_interrupt_poison_consumption_v9(struct kfd_node *dev,
 {
 	enum amdgpu_ras_block block = 0;
 	int old_poison, ret = -EINVAL;
+	uint32_t reset = 0;
 	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid);
 
 	if (!p)
@@ -164,6 +165,15 @@ static void event_interrupt_poison_consumption_v9(struct kfd_node *dev,
 	case SOC15_IH_CLIENTID_UTCL2:
 		ret = kfd_dqm_evict_pasid(dev->dqm, pasid);
 		block = AMDGPU_RAS_BLOCK__GFX;
+		if (ret)
+			reset = AMDGPU_RAS_GPU_RESET_MODE2_RESET;
+		break;
+	case SOC15_IH_CLIENTID_VMC:
+	case SOC15_IH_CLIENTID_VMC1:
+		ret = kfd_dqm_evict_pasid(dev->dqm, pasid);
+		block = AMDGPU_RAS_BLOCK__MMHUB;
+		if (ret)
+			reset = AMDGPU_RAS_GPU_RESET_MODE1_RESET;
 		break;
 	case SOC15_IH_CLIENTID_SDMA0:
 	case SOC15_IH_CLIENTID_SDMA1:
@@ -171,6 +181,7 @@ static void event_interrupt_poison_consumption_v9(struct kfd_node *dev,
 	case SOC15_IH_CLIENTID_SDMA3:
 	case SOC15_IH_CLIENTID_SDMA4:
 		block = AMDGPU_RAS_BLOCK__SDMA;
+		reset = AMDGPU_RAS_GPU_RESET_MODE2_RESET;
 		break;
 	default:
 		break;
@@ -181,17 +192,16 @@ static void event_interrupt_poison_consumption_v9(struct kfd_node *dev,
 	/* resetting queue passes, do page retirement without gpu reset
 	 * resetting queue fails, fallback to gpu reset solution
 	 */
-	if (!ret) {
+	if (!ret)
 		dev_warn(dev->adev->dev,
 			"RAS poison consumption, unmap queue flow succeeded: client id %d\n",
 			client_id);
-		amdgpu_amdkfd_ras_poison_consumption_handler(dev->adev, block, false);
-	} else {
+	else
 		dev_warn(dev->adev->dev,
 			"RAS poison consumption, fall back to gpu reset flow: client id %d\n",
 			client_id);
-		amdgpu_amdkfd_ras_poison_consumption_handler(dev->adev, block, true);
-	}
+
+	amdgpu_amdkfd_ras_poison_consumption_handler(dev->adev, block, reset);
 }
 
 static bool context_id_expected(struct kfd_dev *dev)