mirrors/linux
2
0
Fork 0
mirror of https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git synced 2026-08-30 18:03:10 -04:00
Code Issues Packages Projects Releases Wiki Activity

drm/amd/ras: add device lost check to early exit psp cmd wait loop

Browse Source 
Add device lost status check in PSP fence wait loop to exit
polling early when GPU device lost

Signed-off-by: Ce Sun <cesun102@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This commit is contained in:
Ce Sun
2026-07-24 12:03:30 +08:00
committed by Alex Deucher
parent cba4928cdf
commit 9ba8e75d2e
4 changed files with 22 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
drivers/gpu/drm/amd/ras/ras_mgr/amdgpu_ras_sys.c
View File

@@ -171,6 +171,9 @@ static int amdgpu_ras_sys_check_gpu_status(struct ras_core_context *ras_core,
if (amdgpu_in_reset(adev) || amdgpu_ras_in_recovery(adev))
gpu_status |= RAS_GPU_STATUS__IN_RESET;
if (amdgpu_device_bus_status_check(adev))
gpu_status |= RAS_GPU_STATUS__DEVICE_LOST;
if (amdgpu_sriov_vf(adev))
gpu_status |= RAS_GPU_STATUS__IS_VF;

2
drivers/gpu/drm/amd/ras/rascore/ras.h
View File

@@ -137,6 +137,7 @@ enum ras_gpu_status {
RAS_GPU_STATUS__IN_RESET = 0x2,
RAS_GPU_STATUS__IS_RMA = 0x4,
RAS_GPU_STATUS__IS_VF = 0x8,
RAS_GPU_STATUS__DEVICE_LOST = 0x10,
};
enum ras_fw_eeprom_cmd {
@@ -374,6 +375,7 @@ int ras_core_query_block_ecc_data(struct ras_core_context *ras_core,
bool ras_core_gpu_in_reset(struct ras_core_context *ras_core);
bool ras_core_gpu_is_rma(struct ras_core_context *ras_core);
bool ras_core_gpu_is_vf(struct ras_core_context *ras_core);
bool ras_core_gpu_device_lost(struct ras_core_context *ras_core);
bool ras_core_handle_nbio_irq(struct ras_core_context *ras_core, void *data);
int ras_core_handle_fatal_error(struct ras_core_context *ras_core);

14
drivers/gpu/drm/amd/ras/rascore/ras_core.c
View File

@@ -147,6 +147,20 @@ bool ras_core_gpu_is_vf(struct ras_core_context *ras_core)
return (status & RAS_GPU_STATUS__IS_VF) ? true : false;
}
bool ras_core_gpu_device_lost(struct ras_core_context *ras_core)
{
uint32_t status = 0;
if (!ras_core)
return false;
if (ras_core->sys_fn &&
ras_core->sys_fn->check_gpu_status)
ras_core->sys_fn->check_gpu_status(ras_core, &status);
return (status & RAS_GPU_STATUS__DEVICE_LOST) ? true : false;
}
bool ras_core_gpu_is_rma(struct ras_core_context *ras_core)
{
if (!ras_core)

3
drivers/gpu/drm/amd/ras/rascore/ras_psp.c
View File

@@ -294,6 +294,9 @@ static int send_psp_cmd(struct ras_core_context *ras_core,
psp_ctx->in_fence_value) {
if (--timeout == 0)
break;
if (ras_core_gpu_device_lost(ras_core))
break;
/*
* Shouldn't wait for timeout when err_event_athub occurs,
* because gpu reset thread triggered and lock resource should