[PATCH 07/13] drm/amdgpu/userq: enable support for queue priorities

Khatri, Sunil sukhatri at amd.com
Fri Apr 11 17:07:26 UTC 2025


A small comment otherwise it looks great.
Reviewed-by: Sunil Khatri <sunil.khatri at amd.com>

On 4/11/2025 12:23 AM, Alex Deucher wrote:
> Enable users to create queues at different priority levels.
> The highest level is restricted to drm master.
>
> Signed-off-by: Alex Deucher <alexander.deucher at amd.com>
> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_userqueue.c | 26 ++++++++++++++++++-
>   1 file changed, 25 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_userqueue.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_userqueue.c
> index 57a4ef64e0b8b..b8b13b6ab4631 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_userqueue.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_userqueue.c
> @@ -22,6 +22,7 @@
>    *
>    */
>   
> +#include <drm/drm_auth.h>
>   #include <drm/drm_exec.h>
>   #include "amdgpu.h"
>   #include "amdgpu_vm.h"
> @@ -260,6 +261,21 @@ amdgpu_userqueue_destroy(struct drm_file *filp, int queue_id)
>   	return r;
>   }
>   
> +static int amdgpu_userq_priority_permit(struct drm_file *filp,
> +					int priority)
> Do we want this value of priority to be unsigned as we only want values >=0.
> +{
> +	if (priority < AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_HIGH)
> +		return 0;
> +
> +	if (capable(CAP_SYS_NICE))
> +		return 0;
> +
> +	if (drm_is_current_master(filp))
> +		return 0;
> +
> +	return -EACCES;
> +}
> +
>   static int
>   amdgpu_userqueue_create(struct drm_file *filp, union drm_amdgpu_userq *args)
>   {
> @@ -271,6 +287,9 @@ amdgpu_userqueue_create(struct drm_file *filp, union drm_amdgpu_userq *args)
>   	struct amdgpu_db_info db_info;
>   	uint64_t index;
>   	int qid, r = 0;
> +	int priority =
> +		(args->in.flags & AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_MASK) >>
> +		AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_SHIFT;
>   
>   	/* Usermode queues are only supported for GFX IP as of now */
>   	if (args->in.ip_type != AMDGPU_HW_IP_GFX &&
> @@ -280,6 +299,10 @@ amdgpu_userqueue_create(struct drm_file *filp, union drm_amdgpu_userq *args)
>   		return -EINVAL;
>   	}
>   
> +	r = amdgpu_userq_priority_permit(filp, priority);
> +	if (r)
> +		return r;
> +
>   	/*
>   	 * There could be a situation that we are creating a new queue while
>   	 * the other queues under this UQ_mgr are suspended. So if there is any
> @@ -305,6 +328,7 @@ amdgpu_userqueue_create(struct drm_file *filp, union drm_amdgpu_userq *args)
>   	queue->doorbell_handle = args->in.doorbell_handle;
>   	queue->queue_type = args->in.ip_type;
>   	queue->vm = &fpriv->vm;
> +	queue->priority = priority;
>   
>   	db_info.queue_type = queue->queue_type;
>   	db_info.doorbell_handle = queue->doorbell_handle;
> @@ -377,7 +401,7 @@ int amdgpu_userq_ioctl(struct drm_device *dev, void *data,
>   
>   	switch (args->in.op) {
>   	case AMDGPU_USERQ_OP_CREATE:
> -		if (args->in.flags)
> +		if (args->in.flags & ~AMDGPU_USERQ_CREATE_FLAGS_QUEUE_PRIORITY_MASK)
>   			return -EINVAL;
>   		r = amdgpu_userqueue_create(filp, args);
>   		if (r)


More information about the amd-gfx mailing list