summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c')
-rw-r--r--drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c33
1 files changed, 24 insertions, 9 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c
index 59fafb8392e0..24ad4b97177b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c
@@ -27,6 +27,16 @@
#include "umc/umc_6_7_0_offset.h"
#include "umc/umc_6_7_0_sh_mask.h"
+static bool amdgpu_mca_is_deferred_error(struct amdgpu_device *adev,
+ uint64_t mc_status)
+{
+ if (adev->umc.ras->check_ecc_err_status)
+ return adev->umc.ras->check_ecc_err_status(adev,
+ AMDGPU_MCA_ERROR_TYPE_DE, &mc_status);
+
+ return false;
+}
+
void amdgpu_mca_query_correctable_error_count(struct amdgpu_device *adev,
uint64_t mc_status_addr,
unsigned long *error_count)
@@ -202,16 +212,16 @@ int amdgpu_mca_smu_set_debug_mode(struct amdgpu_device *adev, bool enable)
static void amdgpu_mca_smu_mca_bank_dump(struct amdgpu_device *adev, int idx, struct mca_bank_entry *entry)
{
- dev_info(adev->dev, "[Hardware error] Accelerator Check Architecture events logged\n");
- dev_info(adev->dev, "[Hardware error] aca entry[%02d].STATUS=0x%016llx\n",
+ dev_info(adev->dev, HW_ERR "Accelerator Check Architecture events logged\n");
+ dev_info(adev->dev, HW_ERR "aca entry[%02d].STATUS=0x%016llx\n",
idx, entry->regs[MCA_REG_IDX_STATUS]);
- dev_info(adev->dev, "[Hardware error] aca entry[%02d].ADDR=0x%016llx\n",
+ dev_info(adev->dev, HW_ERR "aca entry[%02d].ADDR=0x%016llx\n",
idx, entry->regs[MCA_REG_IDX_ADDR]);
- dev_info(adev->dev, "[Hardware error] aca entry[%02d].MISC0=0x%016llx\n",
+ dev_info(adev->dev, HW_ERR "aca entry[%02d].MISC0=0x%016llx\n",
idx, entry->regs[MCA_REG_IDX_MISC0]);
- dev_info(adev->dev, "[Hardware error] aca entry[%02d].IPID=0x%016llx\n",
+ dev_info(adev->dev, HW_ERR "aca entry[%02d].IPID=0x%016llx\n",
idx, entry->regs[MCA_REG_IDX_IPID]);
- dev_info(adev->dev, "[Hardware error] aca entry[%02d].SYND=0x%016llx\n",
+ dev_info(adev->dev, HW_ERR "aca entry[%02d].SYND=0x%016llx\n",
idx, entry->regs[MCA_REG_IDX_SYND]);
}
@@ -256,9 +266,14 @@ int amdgpu_mca_smu_log_ras_error(struct amdgpu_device *adev, enum amdgpu_ras_blo
if (type == AMDGPU_MCA_ERROR_TYPE_UE)
amdgpu_ras_error_statistic_ue_count(err_data,
&mcm_info, &err_addr, (uint64_t)count);
- else
- amdgpu_ras_error_statistic_ce_count(err_data,
- &mcm_info, &err_addr, (uint64_t)count);
+ else {
+ if (amdgpu_mca_is_deferred_error(adev, entry->regs[MCA_REG_IDX_STATUS]))
+ amdgpu_ras_error_statistic_de_count(err_data,
+ &mcm_info, &err_addr, (uint64_t)count);
+ else
+ amdgpu_ras_error_statistic_ce_count(err_data,
+ &mcm_info, &err_addr, (uint64_t)count);
+ }
}
out_mca_release: