drm/amdgpu: RAS harvest on driver load

In event of RAS UE + warm reset, error counters shall be harvested and cleared on driver load

Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: John Clements <john.clements@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

1 files changed, 29 insertions, 0 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index 1d905bcbc1ac..b0fe5885e4c6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
@@ -2090,6 +2090,32 @@ release_con:
 	return r;
 }
 
+static int amdgpu_persistent_edc_harvesting_supported(struct amdgpu_device *adev)
+{
+	if (adev->gmc.xgmi.connected_to_cpu)
+		return 1;
+	return 0;
+}
+
+static int amdgpu_persistent_edc_harvesting(struct amdgpu_device *adev,
+					struct ras_common_if *ras_block)
+{
+	struct ras_query_if info = {
+		.head = *ras_block,
+	};
+
+	if (!amdgpu_persistent_edc_harvesting_supported(adev))
+		return 0;
+
+	if (amdgpu_ras_query_error_status(adev, &info) != 0)
+		DRM_WARN("RAS init harvest failure");
+
+	if (amdgpu_ras_reset_error_status(adev, ras_block->block) != 0)
+		DRM_WARN("RAS init harvest reset failure");
+
+	return 0;
+}
+
 /* helper function to handle common stuff in ip late init phase */
 int amdgpu_ras_late_init(struct amdgpu_device *adev,
 			 struct ras_common_if *ras_block,
@@ -2119,6 +2145,9 @@ int amdgpu_ras_late_init(struct amdgpu_device *adev,
 			return r;
 	}
 
+	/* check for errors on warm reset edc persisant supported ASIC */
+	amdgpu_persistent_edc_harvesting(adev, ras_block);
+
 	/* in resume phase, no need to create ras fs node */
 	if (adev->in_suspend || amdgpu_in_reset(adev))
 		return 0;