[PATCH 11/12] drm/amdgpu: Add mode-2 reset in SMU v13.0.6

Alex Deucher alexander.deucher at amd.com
Tue May 9 22:07:13 UTC 2023


From: Lijo Lazar <lijo.lazar at amd.com>

Modifications to mode-2 reset flow for SMU v13.0.6 ASICs.

Signed-off-by: Lijo Lazar <lijo.lazar at amd.com>
Reviewed-by: Le Ma <le.ma at amd.com>
Reviewed-by: Asad Kamal <asad.kamal at amd.com>
Signed-off-by: Alex Deucher <alexander.deucher at amd.com>
---
 drivers/gpu/drm/amd/amdgpu/soc15.c               |  9 +++++++++
 .../gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c | 16 +++++++---------
 2 files changed, 16 insertions(+), 9 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c
index 9fbfd0811d06..082c1e9308d0 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc15.c
+++ b/drivers/gpu/drm/amd/amdgpu/soc15.c
@@ -556,6 +556,15 @@ soc15_asic_reset_method(struct amdgpu_device *adev)
 		if (connected_to_cpu)
 			return AMD_RESET_METHOD_MODE2;
 		break;
+	case IP_VERSION(13, 0, 6):
+		/* Use gpu_recovery param to target a reset method.
+		 * Enable triggering of GPU reset only if specified
+		 * by module parameter.
+		 */
+		if (amdgpu_gpu_recovery == 4 || amdgpu_gpu_recovery == 5)
+			return AMD_RESET_METHOD_MODE2;
+		else
+			return AMD_RESET_METHOD_NONE;
 	default:
 		break;
 	}
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c
index 6dcafd04c98d..4b808c0addc2 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c
@@ -2024,27 +2024,27 @@ static ssize_t smu_v13_0_6_get_gpu_metrics(struct smu_context *smu, void **table
 
 static int smu_v13_0_6_mode2_reset(struct smu_context *smu)
 {
-	u32 smu_version;
 	int ret = 0, index;
 	struct amdgpu_device *adev = smu->adev;
 	int timeout = 10;
 
-	smu_cmn_get_smc_version(smu, NULL, &smu_version);
-
 	index = smu_cmn_to_asic_specific_index(smu, CMN2ASIC_MAPPING_MSG,
 					       SMU_MSG_GfxDeviceDriverReset);
 
 	mutex_lock(&smu->message_lock);
+
 	ret = smu_cmn_send_msg_without_waiting(smu, (uint16_t)index,
 					       SMU_RESET_MODE_2);
+
 	/* This is similar to FLR, wait till max FLR timeout */
 	msleep(100);
+
 	dev_dbg(smu->adev->dev, "restore config space...\n");
 	/* Restore the config space saved during init */
 	amdgpu_device_load_pci_state(adev->pdev);
 
 	dev_dbg(smu->adev->dev, "wait for reset ack\n");
-	while (ret == -ETIME && timeout) {
+	do {
 		ret = smu_cmn_wait_for_response(smu);
 		/* Wait a bit more time for getting ACK */
 		if (ret == -ETIME) {
@@ -2053,16 +2053,14 @@ static int smu_v13_0_6_mode2_reset(struct smu_context *smu)
 			continue;
 		}
 
-		if (ret != 1) {
+		if (ret) {
 			dev_err(adev->dev,
-				"failed to send mode2 message \tparam: 0x%08x response %#x\n",
+				"failed to send mode2 message \tparam: 0x%08x error code %d\n",
 				SMU_RESET_MODE_2, ret);
 			goto out;
 		}
-	}
+	} while (ret == -ETIME && timeout);
 
-	if (ret == 1)
-		ret = 0;
 out:
 	mutex_unlock(&smu->message_lock);
 
-- 
2.40.1



More information about the amd-gfx mailing list