[PATCH] drm/amdgpu: change the default timeout for kernel compute queues

[PATCH] drm/amdgpu: change the default timeout for kernel compute queues

[Alex Deucher]

Change to 60s.  This matches what we already do in virtualization.
Infinite timeout can lead to deadlocks in the kernel.

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 8 +++-----
 drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c    | 4 ++--
 2 files changed, 5 insertions(+), 7 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 357f9405f1aa..1b60f8205f15 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -3167,8 +3167,8 @@ static int amdgpu_device_get_job_timeout_settings(struct amdgpu_device *adev)
 	int ret = 0;
 
 	/*
-	 * By default timeout for non compute jobs is 10000.
-	 * And there is no timeout enforced on compute jobs.
+	 * By default timeout for non compute jobs is 10000
+	 * and 60000 for compute jobs.
 	 * In SR-IOV or passthrough mode, timeout for compute
 	 * jobs are 60000 by default.
 	 */
@@ -3177,10 +3177,8 @@ static int amdgpu_device_get_job_timeout_settings(struct amdgpu_device *adev)
 	if (amdgpu_sriov_vf(adev))
 		adev->compute_timeout = amdgpu_sriov_is_pp_one_vf(adev) ?
 					msecs_to_jiffies(60000) : msecs_to_jiffies(10000);
-	else if (amdgpu_passthrough(adev))
-		adev->compute_timeout =  msecs_to_jiffies(60000);
 	else
-		adev->compute_timeout = MAX_SCHEDULE_TIMEOUT;
+		adev->compute_timeout =  msecs_to_jiffies(60000);
 
 	if (strnlen(input, AMDGPU_MAX_TIMEOUT_PARAM_LENGTH)) {
 		while ((timeout_setting = strsep(&input, ",")) &&
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index ad8a2c391bbd..d3d756585db4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -287,9 +287,9 @@ module_param_named(msi, amdgpu_msi, int, 0444);
  *   for SDMA and Video.
  *
  * By default(with no lockup_timeout settings), the timeout for all non-compute(GFX, SDMA and Video)
- * jobs is 10000. And there is no timeout enforced on compute jobs.
+ * jobs is 10000. The timeout for compute is 60000.
  */
-MODULE_PARM_DESC(lockup_timeout, "GPU lockup timeout in ms (default: for bare metal 10000 for non-compute jobs and infinity timeout for compute jobs; "
+MODULE_PARM_DESC(lockup_timeout, "GPU lockup timeout in ms (default: for bare metal 10000 for non-compute jobs and 60000 for compute jobs; "
 		"for passthrough or sriov, 10000 for all jobs."
 		" 0: keep default value. negative: infinity timeout), "
 		"format: for bare metal [Non-Compute] or [GFX,Compute,SDMA,Video]; "
-- 
2.30.2

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx

Re: [PATCH] drm/amdgpu: change the default timeout for kernel compute queues

[Christian König]

Am 04.05.21 um 17:03 schrieb Alex Deucher:
> Change to 60s.  This matches what we already do in virtualization.
> Infinite timeout can lead to deadlocks in the kernel.
>
> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Reviewed-by: Christian König <christian.koenig@amd.com>

> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 8 +++-----
>   drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c    | 4 ++--
>   2 files changed, 5 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> index 357f9405f1aa..1b60f8205f15 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> @@ -3167,8 +3167,8 @@ static int amdgpu_device_get_job_timeout_settings(struct amdgpu_device *adev)
>   	int ret = 0;
>   
>   	/*
> -	 * By default timeout for non compute jobs is 10000.
> -	 * And there is no timeout enforced on compute jobs.
> +	 * By default timeout for non compute jobs is 10000
> +	 * and 60000 for compute jobs.
>   	 * In SR-IOV or passthrough mode, timeout for compute
>   	 * jobs are 60000 by default.
>   	 */
> @@ -3177,10 +3177,8 @@ static int amdgpu_device_get_job_timeout_settings(struct amdgpu_device *adev)
>   	if (amdgpu_sriov_vf(adev))
>   		adev->compute_timeout = amdgpu_sriov_is_pp_one_vf(adev) ?
>   					msecs_to_jiffies(60000) : msecs_to_jiffies(10000);
> -	else if (amdgpu_passthrough(adev))
> -		adev->compute_timeout =  msecs_to_jiffies(60000);
>   	else
> -		adev->compute_timeout = MAX_SCHEDULE_TIMEOUT;
> +		adev->compute_timeout =  msecs_to_jiffies(60000);
>   
>   	if (strnlen(input, AMDGPU_MAX_TIMEOUT_PARAM_LENGTH)) {
>   		while ((timeout_setting = strsep(&input, ",")) &&
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> index ad8a2c391bbd..d3d756585db4 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> @@ -287,9 +287,9 @@ module_param_named(msi, amdgpu_msi, int, 0444);
>    *   for SDMA and Video.
>    *
>    * By default(with no lockup_timeout settings), the timeout for all non-compute(GFX, SDMA and Video)
> - * jobs is 10000. And there is no timeout enforced on compute jobs.
> + * jobs is 10000. The timeout for compute is 60000.
>    */
> -MODULE_PARM_DESC(lockup_timeout, "GPU lockup timeout in ms (default: for bare metal 10000 for non-compute jobs and infinity timeout for compute jobs; "
> +MODULE_PARM_DESC(lockup_timeout, "GPU lockup timeout in ms (default: for bare metal 10000 for non-compute jobs and 60000 for compute jobs; "
>   		"for passthrough or sriov, 10000 for all jobs."
>   		" 0: keep default value. negative: infinity timeout), "
>   		"format: for bare metal [Non-Compute] or [GFX,Compute,SDMA,Video]; "

_______________________________________________
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/amd-gfx