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

drm/amdgpu: add condition check for amdgpu_umc_fill_error_record

Browse Source 
Add condition check for amdgpu_umc_fill_error_record.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This commit is contained in:
YiPeng Chai 2024-03-22 12:06:01 +08:00 committed by Alex Deucher
parent 2cf8e50ec3
commit e74313be5a
3 changed files with 19 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
View File

@ -579,6 +579,7 @@ struct ras_err_data {
unsigned long de_count;
unsigned long err_addr_cnt;
struct eeprom_table_record *err_addr;
unsigned long err_addr_len;
u32 err_list_count;
struct list_head err_node_list;
};

20
drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
View File

@ -66,6 +66,8 @@ int amdgpu_umc_page_retirement_mca(struct amdgpu_device *adev,
goto out_fini_err_data;
}
err_data.err_addr_len = adev->umc.max_ras_err_cnt_per_query;
/*
* Translate UMC channel address to Physical address
*/
@ -121,6 +123,8 @@ void amdgpu_umc_handle_bad_pages(struct amdgpu_device *adev,
if(!err_data->err_addr)
dev_warn(adev->dev, "Failed to alloc memory for "
"umc error address record!\n");
else
err_data->err_addr_len = adev->umc.max_ras_err_cnt_per_query;
/* umc query_ras_error_address is also responsible for clearing
* error status
@ -146,6 +150,8 @@ void amdgpu_umc_handle_bad_pages(struct amdgpu_device *adev,
if(!err_data->err_addr)
dev_warn(adev->dev, "Failed to alloc memory for "
"umc error address record!\n");
else
err_data->err_addr_len = adev->umc.max_ras_err_cnt_per_query;
/* umc query_ras_error_address is also responsible for clearing
* error status
@ -389,14 +395,20 @@ int amdgpu_umc_process_ecc_irq(struct amdgpu_device *adev,
return 0;
}
void amdgpu_umc_fill_error_record(struct ras_err_data *err_data,
int amdgpu_umc_fill_error_record(struct ras_err_data *err_data,
uint64_t err_addr,
uint64_t retired_page,
uint32_t channel_index,
uint32_t umc_inst)
{
struct eeprom_table_record *err_rec =
&err_data->err_addr[err_data->err_addr_cnt];
struct eeprom_table_record *err_rec;
if (!err_data ||
!err_data->err_addr ||
(err_data->err_addr_cnt >= err_data->err_addr_len))
return -EINVAL;
err_rec = &err_data->err_addr[err_data->err_addr_cnt];
err_rec->address = err_addr;
/* page frame address is saved */
@ -408,6 +420,8 @@ void amdgpu_umc_fill_error_record(struct ras_err_data *err_data,
err_rec->mcumc_id = umc_inst;
err_data->err_addr_cnt++;
return 0;
}
int amdgpu_umc_loop_channels(struct amdgpu_device *adev,

2
drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h
View File

@ -109,7 +109,7 @@ int amdgpu_umc_poison_handler(struct amdgpu_device *adev,
int amdgpu_umc_process_ecc_irq(struct amdgpu_device *adev,
struct amdgpu_irq_src *source,
struct amdgpu_iv_entry *entry);
void amdgpu_umc_fill_error_record(struct ras_err_data *err_data,
int amdgpu_umc_fill_error_record(struct ras_err_data *err_data,
uint64_t err_addr,
uint64_t retired_page,
uint32_t channel_index,