[PATCH v2] drm/sched: Further optimise drm_sched_entity_push_job

Christian König christian.koenig at amd.com
Thu Sep 26 09:15:50 UTC 2024


Am 26.09.24 um 10:18 schrieb Philipp Stanner:
> On Tue, 2024-09-24 at 14:02 +0200, Christian König wrote:
>> Am 24.09.24 um 11:58 schrieb Tvrtko Ursulin:
>>> On 24/09/2024 10:45, Tvrtko Ursulin wrote:
>>>> On 24/09/2024 09:20, Christian König wrote:
>>>>> Am 16.09.24 um 19:30 schrieb Tvrtko Ursulin:
>>>>>> From: Tvrtko Ursulin <tvrtko.ursulin at igalia.com>
>>>>>>
>>>>>> Having removed one re-lock cycle on the entity->lock in a
>>>>>> patch titled
>>>>>> "drm/sched: Optimise drm_sched_entity_push_job", with only a
>>>>>> tiny bit
>>>>>> larger refactoring we can do the same optimisation on the rq-
>>>>>>> lock.
>>>>>> (Currently both drm_sched_rq_add_entity() and
>>>>>> drm_sched_rq_update_fifo_locked() take and release the same
>>>>>> lock.)
>>>>>>
>>>>>> To achieve this we make drm_sched_rq_update_fifo_locked() and
>>>>>> drm_sched_rq_add_entity() expect the rq->lock to be held.
>>>>>>
>>>>>> We also align drm_sched_rq_update_fifo_locked(),
>>>>>> drm_sched_rq_add_entity() and
>>>>>> drm_sched_rq_remove_fifo_locked() function signatures, by
>>>>>> adding rq
>>>>>> as a
>>>>>> parameter to the latter.
>>>>>>
>>>>>> v2:
>>>>>>    * Fix after rebase of the series.
>>>>>>    * Avoid naming incosistency between
>>>>>> drm_sched_rq_add/remove.
>>>>>> (Christian)
>>>>>>
>>>>>> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin at igalia.com>
>>>>>> Cc: Christian König <christian.koenig at amd.com>
>>>>>> Cc: Alex Deucher <alexander.deucher at amd.com>
>>>>>> Cc: Luben Tuikov <ltuikov89 at gmail.com>
>>>>>> Cc: Matthew Brost <matthew.brost at intel.com>
>>>>>> Cc: Philipp Stanner <pstanner at redhat.com>
>>>>> Reviewed-by: Christian König <christian.koenig at amd.com>
>>>> Thanks!
>>>>
>>>> Are you okay to pull into drm-misc-next or we should do some more
>>>> testing on this?
>>>>
>>>> And/or should I resend the series once more in it's entirety so
>>>> this
>>>> v2 is not a reply-to to the original?
>>> I have to respin for the drm_sched_wakeup fix that landed.
>> When I should push the series to drm-misc-next then please send it to
>> me
>> once more.
>>
>> On the other hand we should now have to maintainers for that.
> Yup, will pick up this responsibilities soonish. Danilo and I have been
> on conference recently and I'm out of office soon for a bit, but you
> can expect me / us to take over that work soonish in early autumn.

But please don't push this set. I have messed up the push because I 
didn't realized that the first three patches were supposed to land in 
-fixes instead of -next.

Should be cleaned up my me by the end of the day.

Sorry for the noise,
Christian.

>
> Regards,
> P.
>
>> Regards,
>> Christian.
>>
>>> Regards,
>>>
>>> Tvrtko
>>>
>>>> Regards,
>>>>
>>>> Tvrtko
>>>>
>>>>>> ---
>>>>>>    drivers/gpu/drm/scheduler/sched_entity.c | 12 ++++++++--
>>>>>>    drivers/gpu/drm/scheduler/sched_main.c   | 29
>>>>>> ++++++++++++------------
>>>>>>    include/drm/gpu_scheduler.h              |  3 ++-
>>>>>>    3 files changed, 26 insertions(+), 18 deletions(-)
>>>>>>
>>>>>> diff --git a/drivers/gpu/drm/scheduler/sched_entity.c
>>>>>> b/drivers/gpu/drm/scheduler/sched_entity.c
>>>>>> index d982cebc6bee..8ace1f1ea66b 100644
>>>>>> --- a/drivers/gpu/drm/scheduler/sched_entity.c
>>>>>> +++ b/drivers/gpu/drm/scheduler/sched_entity.c
>>>>>> @@ -515,9 +515,14 @@ struct drm_sched_job
>>>>>> *drm_sched_entity_pop_job(struct drm_sched_entity *entity)
>>>>>>            next =
>>>>>> to_drm_sched_job(spsc_queue_peek(&entity->job_queue));
>>>>>>            if (next) {
>>>>>> +            struct drm_sched_rq *rq;
>>>>>> +
>>>>>>                spin_lock(&entity->lock);
>>>>>> -            drm_sched_rq_update_fifo_locked(entity,
>>>>>> +            rq = entity->rq;
>>>>>> +            spin_lock(&rq->lock);
>>>>>> +            drm_sched_rq_update_fifo_locked(entity, rq,
>>>>>>                                next->submit_ts);
>>>>>> +            spin_unlock(&rq->lock);
>>>>>>                spin_unlock(&entity->lock);
>>>>>>            }
>>>>>>        }
>>>>>> @@ -618,11 +623,14 @@ void drm_sched_entity_push_job(struct
>>>>>> drm_sched_job *sched_job)
>>>>>>            sched = rq->sched;
>>>>>>            atomic_inc(sched->score);
>>>>>> +
>>>>>> +        spin_lock(&rq->lock);
>>>>>>            drm_sched_rq_add_entity(rq, entity);
>>>>>>            if (drm_sched_policy == DRM_SCHED_POLICY_FIFO)
>>>>>> -            drm_sched_rq_update_fifo_locked(entity,
>>>>>> submit_ts);
>>>>>> +            drm_sched_rq_update_fifo_locked(entity, rq,
>>>>>> submit_ts);
>>>>>> +        spin_unlock(&rq->lock);
>>>>>>            spin_unlock(&entity->lock);
>>>>>>            drm_sched_wakeup(sched, entity);
>>>>>> diff --git a/drivers/gpu/drm/scheduler/sched_main.c
>>>>>> b/drivers/gpu/drm/scheduler/sched_main.c
>>>>>> index 18a952f73ecb..5c83fb92bb89 100644
>>>>>> --- a/drivers/gpu/drm/scheduler/sched_main.c
>>>>>> +++ b/drivers/gpu/drm/scheduler/sched_main.c
>>>>>> @@ -153,17 +153,18 @@ static __always_inline bool
>>>>>> drm_sched_entity_compare_before(struct rb_node *a,
>>>>>>        return ktime_before(ent_a->oldest_job_waiting,
>>>>>> ent_b->oldest_job_waiting);
>>>>>>    }
>>>>>> -static inline void drm_sched_rq_remove_fifo_locked(struct
>>>>>> drm_sched_entity *entity)
>>>>>> +static void drm_sched_rq_remove_fifo_locked(struct
>>>>>> drm_sched_entity *entity,
>>>>>> +                        struct drm_sched_rq *rq)
>>>>>>    {
>>>>>> -    struct drm_sched_rq *rq = entity->rq;
>>>>>> -
>>>>>>        if (!RB_EMPTY_NODE(&entity->rb_tree_node)) {
>>>>>>            rb_erase_cached(&entity->rb_tree_node, &rq-
>>>>>>> rb_tree_root);
>>>>>>            RB_CLEAR_NODE(&entity->rb_tree_node);
>>>>>>        }
>>>>>>    }
>>>>>> -void drm_sched_rq_update_fifo_locked(struct drm_sched_entity
>>>>>> *entity, ktime_t ts)
>>>>>> +void drm_sched_rq_update_fifo_locked(struct drm_sched_entity
>>>>>> *entity,
>>>>>> +                     struct drm_sched_rq *rq,
>>>>>> +                     ktime_t ts)
>>>>>>    {
>>>>>>        /*
>>>>>>         * Both locks need to be grabbed, one to protect from
>>>>>> entity->rq change
>>>>>> @@ -171,17 +172,14 @@ void
>>>>>> drm_sched_rq_update_fifo_locked(struct
>>>>>> drm_sched_entity *entity, ktime_t ts
>>>>>>         * other to update the rb tree structure.
>>>>>>         */
>>>>>>        lockdep_assert_held(&entity->lock);
>>>>>> +    lockdep_assert_held(&rq->lock);
>>>>>> -    spin_lock(&entity->rq->lock);
>>>>>> -
>>>>>> -    drm_sched_rq_remove_fifo_locked(entity);
>>>>>> +    drm_sched_rq_remove_fifo_locked(entity, rq);
>>>>>>        entity->oldest_job_waiting = ts;
>>>>>> -    rb_add_cached(&entity->rb_tree_node, &entity->rq-
>>>>>>> rb_tree_root,
>>>>>> +    rb_add_cached(&entity->rb_tree_node, &rq->rb_tree_root,
>>>>>>                  drm_sched_entity_compare_before);
>>>>>> -
>>>>>> -    spin_unlock(&entity->rq->lock);
>>>>>>    }
>>>>>>    /**
>>>>>> @@ -213,15 +211,14 @@ static void drm_sched_rq_init(struct
>>>>>> drm_gpu_scheduler *sched,
>>>>>>    void drm_sched_rq_add_entity(struct drm_sched_rq *rq,
>>>>>>                     struct drm_sched_entity *entity)
>>>>>>    {
>>>>>> +    lockdep_assert_held(&entity->lock);
>>>>>> +    lockdep_assert_held(&rq->lock);
>>>>>> +
>>>>>>        if (!list_empty(&entity->list))
>>>>>>            return;
>>>>>> -    spin_lock(&rq->lock);
>>>>>> -
>>>>>>        atomic_inc(rq->sched->score);
>>>>>>        list_add_tail(&entity->list, &rq->entities);
>>>>>> -
>>>>>> -    spin_unlock(&rq->lock);
>>>>>>    }
>>>>>>    /**
>>>>>> @@ -235,6 +232,8 @@ void drm_sched_rq_add_entity(struct
>>>>>> drm_sched_rq *rq,
>>>>>>    void drm_sched_rq_remove_entity(struct drm_sched_rq *rq,
>>>>>>                    struct drm_sched_entity *entity)
>>>>>>    {
>>>>>> +    lockdep_assert_held(&entity->lock);
>>>>>> +
>>>>>>        if (list_empty(&entity->list))
>>>>>>            return;
>>>>>> @@ -247,7 +246,7 @@ void drm_sched_rq_remove_entity(struct
>>>>>> drm_sched_rq *rq,
>>>>>>            rq->current_entity = NULL;
>>>>>>        if (drm_sched_policy == DRM_SCHED_POLICY_FIFO)
>>>>>> -        drm_sched_rq_remove_fifo_locked(entity);
>>>>>> +        drm_sched_rq_remove_fifo_locked(entity, rq);
>>>>>>        spin_unlock(&rq->lock);
>>>>>>    }
>>>>>> diff --git a/include/drm/gpu_scheduler.h
>>>>>> b/include/drm/gpu_scheduler.h
>>>>>> index 80198e6cf537..b21806d5a8eb 100644
>>>>>> --- a/include/drm/gpu_scheduler.h
>>>>>> +++ b/include/drm/gpu_scheduler.h
>>>>>> @@ -596,7 +596,8 @@ void drm_sched_rq_add_entity(struct
>>>>>> drm_sched_rq *rq,
>>>>>>    void drm_sched_rq_remove_entity(struct drm_sched_rq *rq,
>>>>>>                    struct drm_sched_entity *entity);
>>>>>> -void drm_sched_rq_update_fifo_locked(struct drm_sched_entity
>>>>>> *entity, ktime_t ts);
>>>>>> +void drm_sched_rq_update_fifo_locked(struct drm_sched_entity
>>>>>> *entity,
>>>>>> +                     struct drm_sched_rq *rq, ktime_t ts);
>>>>>>    int drm_sched_entity_init(struct drm_sched_entity *entity,
>>>>>>                  enum drm_sched_priority priority,



More information about the dri-devel mailing list