[PATCH 1/6] drm/sched: Add internal job peek/pop API

Tvrtko Ursulin tvrtko.ursulin at igalia.com
Wed Feb 12 12:30:08 UTC 2025


On 12/02/2025 10:40, Philipp Stanner wrote:
> On Wed, 2025-02-12 at 09:32 +0000, Tvrtko Ursulin wrote:
>>
>> On 12/02/2025 09:02, Philipp Stanner wrote:
>>> On Fri, 2025-02-07 at 14:50 +0000, Tvrtko Ursulin wrote:
>>>> Idea is to add helpers for peeking and popping jobs from entities
>>>> with
>>>> the goal of decoupling the hidden assumption in the code that
>>>> queue_node
>>>> is the first element in struct drm_sched_job.
>>>>
>>>> That assumption usually comes in the form of:
>>>>
>>>>     while ((job = to_drm_sched_job(spsc_queue_pop(&entity-
>>>>> job_queue))))
>>>>
>>>> Which breaks if the queue_node is re-positioned due
>>>> to_drm_sched_job
>>>> being implemented with a container_of.
>>>>
>>>> This also allows us to remove duplicate definitions of
>>>> to_drm_sched_job.
>>>>
>>>> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin at igalia.com>
>>>> Cc: Christian König <christian.koenig at amd.com>
>>>> Cc: Danilo Krummrich <dakr at kernel.org>
>>>> Cc: Matthew Brost <matthew.brost at intel.com>
>>>> Cc: Philipp Stanner <phasta at kernel.org>
>>>> ---
>>>>    drivers/gpu/drm/scheduler/sched_entity.c   | 11 +++---
>>>>    drivers/gpu/drm/scheduler/sched_internal.h | 46
>>>> ++++++++++++++++++++++
>>>>    drivers/gpu/drm/scheduler/sched_main.c     |  7 ++--
>>>>    3 files changed, 54 insertions(+), 10 deletions(-)
>>>>    create mode 100644 drivers/gpu/drm/scheduler/sched_internal.h
>>>>
>>>> diff --git a/drivers/gpu/drm/scheduler/sched_entity.c
>>>> b/drivers/gpu/drm/scheduler/sched_entity.c
>>>> index 69bcf0e99d57..a171f05ad761 100644
>>>> --- a/drivers/gpu/drm/scheduler/sched_entity.c
>>>> +++ b/drivers/gpu/drm/scheduler/sched_entity.c
>>>> @@ -28,11 +28,10 @@
>>>>    #include <drm/drm_print.h>
>>>>    #include <drm/gpu_scheduler.h>
>>>>    
>>>> +#include "sched_internal.h"
>>>> +
>>>>    #include "gpu_scheduler_trace.h"
>>>>    
>>>> -#define to_drm_sched_job(sched_job)		\
>>>> -		container_of((sched_job), struct drm_sched_job,
>>>> queue_node)
>>>> -
>>>>    /**
>>>>     * drm_sched_entity_init - Init a context entity used by
>>>> scheduler
>>>> when
>>>>     * submit to HW ring.
>>>> @@ -255,7 +254,7 @@ static void drm_sched_entity_kill(struct
>>>> drm_sched_entity *entity)
>>>>    	/* The entity is guaranteed to not be used by the
>>>> scheduler
>>>> */
>>>>    	prev = rcu_dereference_check(entity->last_scheduled,
>>>> true);
>>>>    	dma_fence_get(prev);
>>>> -	while ((job = to_drm_sched_job(spsc_queue_pop(&entity-
>>>>> job_queue)))) {
>>>> +	while ((job = drm_sched_entity_queue_pop(entity))) {
>>>>    		struct drm_sched_fence *s_fence = job->s_fence;
>>>>    
>>>>    		dma_fence_get(&s_fence->finished);
>>>> @@ -477,7 +476,7 @@ struct drm_sched_job
>>>> *drm_sched_entity_pop_job(struct drm_sched_entity *entity)
>>>>    {
>>>>    	struct drm_sched_job *sched_job;
>>>>    
>>>> -	sched_job = to_drm_sched_job(spsc_queue_peek(&entity-
>>>>> job_queue));
>>>> +	sched_job = drm_sched_entity_queue_peek(entity);
>>>>    	if (!sched_job)
>>>>    		return NULL;
>>>>    
>>>> @@ -513,7 +512,7 @@ struct drm_sched_job
>>>> *drm_sched_entity_pop_job(struct drm_sched_entity *entity)
>>>>    	if (drm_sched_policy == DRM_SCHED_POLICY_FIFO) {
>>>>    		struct drm_sched_job *next;
>>>>    
>>>> -		next = to_drm_sched_job(spsc_queue_peek(&entity-
>>>>> job_queue));
>>>> +		next = drm_sched_entity_queue_peek(entity);
>>>>    		if (next) {
>>>>    			struct drm_sched_rq *rq;
>>>>    
>>>> diff --git a/drivers/gpu/drm/scheduler/sched_internal.h
>>>> b/drivers/gpu/drm/scheduler/sched_internal.h
>>>> new file mode 100644
>>>> index 000000000000..25ac62ac2bf3
>>>> --- /dev/null
>>>> +++ b/drivers/gpu/drm/scheduler/sched_internal.h
>>>> @@ -0,0 +1,46 @@
>>>> +#ifndef _DRM_GPU_SCHEDULER_INTERNAL_H_
>>>> +#define _DRM_GPU_SCHEDULER_INTERNAL_H_
>>>> +
>>>> +/**
>>>> + * drm_sched_entity_queue_pop - Low level helper for popping
>>>> queued
>>>> jobs
>>>> + *
>>>> + * @entity: scheduler entity
>>>> + *
>>>> + * Low level helper for popping queued jobs.
>>>> + *
>>>> + * Returns the job dequeued or NULL.
>>>> + */
>>>> +static inline struct drm_sched_job *
>>>> +drm_sched_entity_queue_pop(struct drm_sched_entity *entity)
>>>> +{
>>>> +	struct spsc_node *node;
>>>> +
>>>> +	node = spsc_queue_pop(&entity->job_queue);
>>>> +	if (!node)
>>>> +		return NULL;
>>>> +
>>>> +	return container_of(node, struct drm_sched_job,
>>>> queue_node);
>>>> +}
>>>> +
>>>> +/**
>>>> + * drm_sched_entity_queue_peek - Low level helper for peeking at
>>>> the
>>>> job queue
>>>> + *
>>>> + * @entity: scheduler entity
>>>> + *
>>>> + * Low level helper for peeking at the job queue
>>>> + *
>>>> + * Returns the job at the head of the queue or NULL.
>>>
>>> I would like to (slowly) work towards a unified style regarding the
>>> docstrings. They're currently relatively inconsistent in drm/sched.
>>>
>>> I think we should do it that way:
>>>
>>> ""
>>> @entity: scheduler entity
>>>
>>> Returns: the job at the head of the queue or NULL.
>>>
>>> Low level helper for peeking at the the job queue.
>>> ""
>>
>> Returns before the description would be yet another new style, no?
>> I's
>> say that if we are churning lets follow
>> Documentation/doc-guide/kernel-doc.rst.
> 
> Oh yes, you are right – official guideline demands "Return:" at the
> end. So let's go for that for contributions.

So you want me to respin or are you okay with doing all scheduler kernel 
doc in one patch afterwards?

Regards,

Tvrtko

>>   Or even consider sending a patch
>> which churns everything at once.
>>
>> Regards,
>>
>> Tvrtko
>>
>>>> + */
>>>> +static inline struct drm_sched_job *
>>>> +drm_sched_entity_queue_peek(struct drm_sched_entity *entity)
>>>> +{
>>>> +	struct spsc_node *node;
>>>> +
>>>> +	node = spsc_queue_peek(&entity->job_queue);
>>>> +	if (!node)
>>>> +		return NULL;
>>>> +
>>>> +	return container_of(node, struct drm_sched_job,
>>>> queue_node);
>>>> +}
>>>> +
>>>> +#endif
>>>> diff --git a/drivers/gpu/drm/scheduler/sched_main.c
>>>> b/drivers/gpu/drm/scheduler/sched_main.c
>>>> index a48be16ab84f..9f614a775c49 100644
>>>> --- a/drivers/gpu/drm/scheduler/sched_main.c
>>>> +++ b/drivers/gpu/drm/scheduler/sched_main.c
>>>> @@ -78,6 +78,8 @@
>>>>    #include <drm/gpu_scheduler.h>
>>>>    #include <drm/spsc_queue.h>
>>>>    
>>>> +#include "sched_internal.h"
>>>> +
>>>>    #define CREATE_TRACE_POINTS
>>>>    #include "gpu_scheduler_trace.h"
>>>>    
>>>> @@ -87,9 +89,6 @@ static struct lockdep_map drm_sched_lockdep_map
>>>> = {
>>>>    };
>>>>    #endif
>>>>    
>>>> -#define to_drm_sched_job(sched_job)		\
>>>> -		container_of((sched_job), struct drm_sched_job,
>>>> queue_node)
>>>> -
>>>>    int drm_sched_policy = DRM_SCHED_POLICY_FIFO;
>>>>    
>>>>    /**
>>>> @@ -123,7 +122,7 @@ static bool drm_sched_can_queue(struct
>>>> drm_gpu_scheduler *sched,
>>>>    {
>>>>    	struct drm_sched_job *s_job;
>>>>    
>>>> -	s_job = to_drm_sched_job(spsc_queue_peek(&entity-
>>>>> job_queue));
>>>> +	s_job = drm_sched_entity_queue_peek(entity);
>>>>    	if (!s_job)
>>>>    		return false;
>>>>    
>>>
>>
> 



More information about the dri-devel mailing list