drm/amdgpu: add get_retire_flip_bits for UMC

Add the general interface to get flip bits for RAS bad page retirement.

Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This commit is contained in:
Tao Zhou 2025-04-10 19:29:57 +08:00 committed by Alex Deucher
parent 80f66ca7a4
commit 699bff37a5
2 changed files with 19 additions and 0 deletions

View File

@ -3473,6 +3473,10 @@ int amdgpu_ras_init_badpage_info(struct amdgpu_device *adev)
if (!adev->umc.ras || !adev->umc.ras->convert_ras_err_addr)
control->ras_num_pa_recs = control->ras_num_recs;
if (adev->umc.ras &&
adev->umc.ras->get_retire_flip_bits)
adev->umc.ras->get_retire_flip_bits(adev);
if (control->ras_num_recs) {
ret = amdgpu_ras_load_bad_pages(adev);
if (ret)

View File

@ -78,6 +78,18 @@
#define UMC_NPS_SHIFT 40
#define UMC_NPS_MASK 0xffULL
/* three column bits and one row bit in MCA address flip
* in bad page retirement
*/
#define RETIRE_FLIP_BITS_NUM 4
struct amdgpu_umc_flip_bits {
uint32_t flip_bits_in_pa[RETIRE_FLIP_BITS_NUM];
uint32_t flip_row_bit;
uint32_t r13_in_pa;
uint32_t bit_num;
};
typedef int (*umc_func)(struct amdgpu_device *adev, uint32_t node_inst,
uint32_t umc_inst, uint32_t ch_inst, void *data);
@ -100,6 +112,7 @@ struct amdgpu_umc_ras {
bool dump_addr);
uint32_t (*get_die_id_from_pa)(struct amdgpu_device *adev,
uint64_t mca_addr, uint64_t retired_page);
void (*get_retire_flip_bits)(struct amdgpu_device *adev);
};
struct amdgpu_umc_funcs {
@ -130,6 +143,8 @@ struct amdgpu_umc {
/* active mask for umc node instance */
unsigned long active_mask;
struct amdgpu_umc_flip_bits flip_bits;
};
int amdgpu_umc_ras_sw_init(struct amdgpu_device *adev);