leandrof/linux
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

drm/amdgpu: add gpu reset control for umc page retirement

Browse Source 
Add a reset parameter for umc page retirement, let user decide whether
call gpu reset in umc page retirement.

Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Acked-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This commit is contained in:
Tao Zhou
2021-12-10 17:01:15 +08:00
committed by Alex Deucher
parent d764fb2af6
commit f4409ee846
2 changed files with 15 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
View File

@@ -23,6 +23,13 @@
#include "amdgpu_ras.h"
static int amdgpu_umc_process_ras_data_cb(struct amdgpu_device *adev,
void *ras_error_status,
struct amdgpu_iv_entry *entry)
{
return amdgpu_umc_do_page_retirement(adev, ras_error_status, entry, true);
}
int amdgpu_umc_ras_late_init(struct amdgpu_device *adev)
{
int r;
@@ -88,9 +95,10 @@ void amdgpu_umc_ras_fini(struct amdgpu_device *adev)
}
}
int amdgpu_umc_process_ras_data_cb(struct amdgpu_device *adev,
int amdgpu_umc_do_page_retirement(struct amdgpu_device *adev,
void *ras_error_status,
struct amdgpu_iv_entry *entry)
struct amdgpu_iv_entry *entry,
bool reset)
{
struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
@@ -164,7 +172,8 @@ int amdgpu_umc_process_ras_data_cb(struct amdgpu_device *adev,
adev->smu.ppt_funcs->send_hbm_bad_pages_num(&adev->smu, con->eeprom_control.ras_num_recs);
}
amdgpu_ras_reset_gpu(adev);
if (reset)
amdgpu_ras_reset_gpu(adev);
}
kfree(err_data->err_addr);

5
drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h
View File

@@ -78,9 +78,10 @@ struct amdgpu_umc {
int amdgpu_umc_ras_late_init(struct amdgpu_device *adev);
void amdgpu_umc_ras_fini(struct amdgpu_device *adev);
int amdgpu_umc_process_ras_data_cb(struct amdgpu_device *adev,
int amdgpu_umc_do_page_retirement(struct amdgpu_device *adev,
void *ras_error_status,
struct amdgpu_iv_entry *entry);
struct amdgpu_iv_entry *entry,
bool reset);
int amdgpu_umc_process_ecc_irq(struct amdgpu_device *adev,
struct amdgpu_irq_src *source,
struct amdgpu_iv_entry *entry);