Lines Matching full:scheduler

38  * Setting this flag on a scheduler fence prevents pipelining of jobs depending
79  * ring, and the scheduler will alternate between entities based on
116 	 * be scheduled on any scheduler on this list.
241  * @sched: the scheduler to which this rq belongs to.
266          * @scheduled: this fence is what will be signaled by the scheduler
272          * @finished: this fence is what will be signaled by the scheduler
296          * @sched: the scheduler instance to which the job having this struct
324  * @sched: the scheduler instance on which this job is scheduled.
327  * @credits: the number of credits this job contributes to the scheduler
330  *         limit of the scheduler then the job is marked guilty and will not
337  * should call drm_sched_entity_push_job() once it wants the scheduler
351 	 * The scheduler this job is or will be scheduled on. Gets set by
391  * enum drm_gpu_sched_stat - the scheduler's status
396  * @DRM_GPU_SCHED_STAT_NO_HANG: Contrary to scheduler's assumption, the GPU
408  *	called by the scheduler
416 	 * Called when the scheduler is considering scheduling this job next, to
444 	 * driver passed NULL, a separate, ordered workqueue the scheduler
447 	 * Note that the scheduler expects to 'inherit' its own reference to
450 	 * reference for the scheduler, and additional ones for the driver's
468 	 * or a hardware scheduler is being used.
470 	 * For a FIRMWARE SCHEDULER, each ring has one scheduler, and each
471 	 * scheduler has one entity. Hence, the steps taken typically look as
474 	 * 1. Stop the scheduler using drm_sched_stop(). This will pause the
475 	 *    scheduler workqueues and cancel the timeout work, guaranteeing
480 	 * 3. Kill the entity and the associated scheduler.
483 	 * For a HARDWARE SCHEDULER, a scheduler instance schedules jobs from
485 	 * associated with the affected scheduler cannot be torn down, because
489 	 * Consequently, the procedure to recover with a hardware scheduler
508 	 * Return: The scheduler's status, defined by &enum drm_gpu_sched_stat
520 	 * @cancel_job: Used by the scheduler to guarantee remaining jobs' fences
523 	 * Used by the scheduler to cancel all jobs that have not been executed
531 	 * The scheduler will only call this callback once it stopped calling
539  * struct drm_gpu_scheduler - scheduler instance-specific data
542  * @credit_limit: the credit limit of this scheduler
543  * @credit_count: the current credit count of this scheduler
544  * @timeout: the time after which a job is removed from the scheduler.
545  * @name: name of the ring for which this scheduler is being used.
549  * @job_scheduled: once @drm_sched_entity_do_release is called the scheduler
555  * @work_run_job: work which calls run_job op of each scheduler.
556  * @work_free_job: work which calls free_job op of each scheduler.
568  * @own_submit_wq: scheduler owns allocation of @submit_wq
571  * One scheduler is implemented for each hardware ring.
601  * struct drm_sched_init_args - parameters for initializing a DRM GPU scheduler
608  * @credit_limit: the number of credits this scheduler can hold from all jobs
630 /* Scheduler operations */