diff options
Diffstat (limited to 'drivers/gpu/drm/scheduler/sched_main.c')
| -rw-r--r-- | drivers/gpu/drm/scheduler/sched_main.c | 510 |
1 files changed, 313 insertions, 197 deletions
diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c index 7e90c9f95611..1d4f1b822e7b 100644 --- a/drivers/gpu/drm/scheduler/sched_main.c +++ b/drivers/gpu/drm/scheduler/sched_main.c @@ -41,7 +41,7 @@ * 4. Entities themselves maintain a queue of jobs that will be scheduled on * the hardware. * - * The jobs in a entity are always scheduled in the order that they were pushed. + * The jobs in an entity are always scheduled in the order in which they were pushed. * * Note that once a job was taken from the entities queue and pushed to the * hardware, i.e. the pending queue, the entity must not be referenced anymore @@ -64,14 +64,9 @@ * credit limit, the job won't be executed. Instead, the scheduler will wait * until the credit count has decreased enough to not overflow its credit limit. * This implies waiting for previously executed jobs. - * - * Optionally, drivers may register a callback (update_job_credits) provided by - * struct drm_sched_backend_ops to update the job's credits dynamically. The - * scheduler executes this callback every time the scheduler considers a job for - * execution and subsequently checks whether the job fits the scheduler's credit - * limit. */ +#include <linux/export.h> #include <linux/wait.h> #include <linux/sched.h> #include <linux/completion.h> @@ -84,12 +79,11 @@ #include <drm/gpu_scheduler.h> #include <drm/spsc_queue.h> +#include "sched_internal.h" + #define CREATE_TRACE_POINTS #include "gpu_scheduler_trace.h" -#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; /** @@ -103,9 +97,9 @@ static u32 drm_sched_available_credits(struct drm_gpu_scheduler *sched) { u32 credits; - drm_WARN_ON(sched, check_sub_overflow(sched->credit_limit, - atomic_read(&sched->credit_count), - &credits)); + WARN_ON(check_sub_overflow(sched->credit_limit, + atomic_read(&sched->credit_count), + &credits)); return credits; } @@ -123,23 +117,18 @@ 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; - if (sched->ops->update_job_credits) { - s_job->credits = sched->ops->update_job_credits(s_job); - - drm_WARN(sched, !s_job->credits, - "Jobs with zero credits bypass job-flow control.\n"); - } - /* If a job exceeds the credit limit, truncate it to the credit limit * itself to guarantee forward progress. */ - if (drm_WARN(sched, s_job->credits > sched->credit_limit, - "Jobs may not exceed the credit limit, truncate.\n")) + if (s_job->credits > sched->credit_limit) { + dev_WARN(sched->dev, + "Jobs may not exceed the credit limit, truncate.\n"); s_job->credits = sched->credit_limit; + } return drm_sched_available_credits(sched) >= s_job->credits; } @@ -153,35 +142,33 @@ 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(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 * for entity from within concurrent drm_sched_entity_select_rq and the * other to update the rb tree structure. */ - spin_lock(&entity->rq_lock); - spin_lock(&entity->rq->lock); + lockdep_assert_held(&entity->lock); + lockdep_assert_held(&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); - spin_unlock(&entity->rq_lock); } /** @@ -213,15 +200,14 @@ 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 +221,8 @@ 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 +235,7 @@ 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); } @@ -275,38 +263,14 @@ drm_sched_rq_select_entity_rr(struct drm_gpu_scheduler *sched, entity = rq->current_entity; if (entity) { list_for_each_entry_continue(entity, &rq->entities, list) { - if (drm_sched_entity_is_ready(entity)) { - /* If we can't queue yet, preserve the current - * entity in terms of fairness. - */ - if (!drm_sched_can_queue(sched, entity)) { - spin_unlock(&rq->lock); - return ERR_PTR(-ENOSPC); - } - - rq->current_entity = entity; - reinit_completion(&entity->entity_idle); - spin_unlock(&rq->lock); - return entity; - } + if (drm_sched_entity_is_ready(entity)) + goto found; } } list_for_each_entry(entity, &rq->entities, list) { - if (drm_sched_entity_is_ready(entity)) { - /* If we can't queue yet, preserve the current entity in - * terms of fairness. - */ - if (!drm_sched_can_queue(sched, entity)) { - spin_unlock(&rq->lock); - return ERR_PTR(-ENOSPC); - } - - rq->current_entity = entity; - reinit_completion(&entity->entity_idle); - spin_unlock(&rq->lock); - return entity; - } + if (drm_sched_entity_is_ready(entity)) + goto found; if (entity == rq->current_entity) break; @@ -315,6 +279,22 @@ drm_sched_rq_select_entity_rr(struct drm_gpu_scheduler *sched, spin_unlock(&rq->lock); return NULL; + +found: + if (!drm_sched_can_queue(sched, entity)) { + /* + * If scheduler cannot take more jobs signal the caller to not + * consider lower priority queues. + */ + entity = ERR_PTR(-ENOSPC); + } else { + rq->current_entity = entity; + reinit_completion(&entity->entity_idle); + } + + spin_unlock(&rq->lock); + + return entity; } /** @@ -349,7 +329,6 @@ drm_sched_rq_select_entity_fifo(struct drm_gpu_scheduler *sched, return ERR_PTR(-ENOSPC); } - rq->current_entity = entity; reinit_completion(&entity->entity_idle); break; } @@ -370,36 +349,20 @@ static void drm_sched_run_job_queue(struct drm_gpu_scheduler *sched) } /** - * __drm_sched_run_free_queue - enqueue free-job work + * drm_sched_run_free_queue - enqueue free-job work * @sched: scheduler instance */ -static void __drm_sched_run_free_queue(struct drm_gpu_scheduler *sched) +static void drm_sched_run_free_queue(struct drm_gpu_scheduler *sched) { if (!READ_ONCE(sched->pause_submit)) queue_work(sched->submit_wq, &sched->work_free_job); } /** - * drm_sched_run_free_queue - enqueue free-job work if ready - * @sched: scheduler instance - */ -static void drm_sched_run_free_queue(struct drm_gpu_scheduler *sched) -{ - struct drm_sched_job *job; - - spin_lock(&sched->job_list_lock); - job = list_first_entry_or_null(&sched->pending_list, - struct drm_sched_job, list); - if (job && dma_fence_is_signaled(&job->s_fence->finished)) - __drm_sched_run_free_queue(sched); - spin_unlock(&sched->job_list_lock); -} - -/** * drm_sched_job_done - complete a job * @s_job: pointer to the job which is done * - * Finish the job's fence and wake up the worker thread. + * Finish the job's fence and resubmit the work items. */ static void drm_sched_job_done(struct drm_sched_job *s_job, int result) { @@ -409,12 +372,12 @@ static void drm_sched_job_done(struct drm_sched_job *s_job, int result) atomic_sub(s_job->credits, &sched->credit_count); atomic_dec(sched->score); - trace_drm_sched_process_job(s_fence); + trace_drm_sched_job_done(s_fence); dma_fence_get(&s_fence->finished); drm_sched_fence_finished(s_fence, result); dma_fence_put(&s_fence->finished); - __drm_sched_run_free_queue(sched); + drm_sched_run_free_queue(sched); } /** @@ -544,11 +507,37 @@ static void drm_sched_job_begin(struct drm_sched_job *s_job) spin_unlock(&sched->job_list_lock); } +/** + * drm_sched_job_reinsert_on_false_timeout - reinsert the job on a false timeout + * @sched: scheduler instance + * @job: job to be reinserted on the pending list + * + * In the case of a "false timeout" - when a timeout occurs but the GPU isn't + * hung and is making progress, the scheduler must reinsert the job back into + * @sched->pending_list. Otherwise, the job and its resources won't be freed + * through the &struct drm_sched_backend_ops.free_job callback. + * + * This function must be used in "false timeout" cases only. + */ +static void drm_sched_job_reinsert_on_false_timeout(struct drm_gpu_scheduler *sched, + struct drm_sched_job *job) +{ + spin_lock(&sched->job_list_lock); + list_add(&job->list, &sched->pending_list); + + /* After reinserting the job, the scheduler enqueues the free-job work + * again if ready. Otherwise, a signaled job could be added to the + * pending list, but never freed. + */ + drm_sched_run_free_queue(sched); + spin_unlock(&sched->job_list_lock); +} + static void drm_sched_job_timedout(struct work_struct *work) { struct drm_gpu_scheduler *sched; struct drm_sched_job *job; - enum drm_gpu_sched_stat status = DRM_GPU_SCHED_STAT_NOMINAL; + enum drm_gpu_sched_stat status = DRM_GPU_SCHED_STAT_RESET; sched = container_of(work, struct drm_gpu_scheduler, work_tdr.work); @@ -559,9 +548,10 @@ static void drm_sched_job_timedout(struct work_struct *work) if (job) { /* - * Remove the bad job so it cannot be freed by concurrent - * drm_sched_cleanup_jobs. It will be reinserted back after sched->thread - * is parked at which point it's safe. + * Remove the bad job so it cannot be freed by a concurrent + * &struct drm_sched_backend_ops.free_job. It will be + * reinserted after the scheduler's work items have been + * cancelled, at which point it's safe. */ list_del_init(&job->list); spin_unlock(&sched->job_list_lock); @@ -576,6 +566,9 @@ static void drm_sched_job_timedout(struct work_struct *work) job->sched->ops->free_job(job); sched->free_guilty = false; } + + if (status == DRM_GPU_SCHED_STAT_NO_HANG) + drm_sched_job_reinsert_on_false_timeout(sched, job); } else { spin_unlock(&sched->job_list_lock); } @@ -595,6 +588,13 @@ static void drm_sched_job_timedout(struct work_struct *work) * callers responsibility to release it manually if it's not part of the * pending list any more. * + * This function is typically used for reset recovery (see the docu of + * drm_sched_backend_ops.timedout_job() for details). Do not call it for + * scheduler teardown, i.e., before calling drm_sched_fini(). + * + * As it's only used for reset recovery, drivers must not call this function + * in their &struct drm_sched_backend_ops.timedout_job callback when they + * skip a reset using &enum drm_gpu_sched_stat.DRM_GPU_SCHED_STAT_NO_HANG. */ void drm_sched_stop(struct drm_gpu_scheduler *sched, struct drm_sched_job *bad) { @@ -604,10 +604,10 @@ void drm_sched_stop(struct drm_gpu_scheduler *sched, struct drm_sched_job *bad) /* * Reinsert back the bad job here - now it's safe as - * drm_sched_get_finished_job cannot race against us and release the + * drm_sched_get_finished_job() cannot race against us and release the * bad job at this point - we parked (waited for) any in progress - * (earlier) cleanups and drm_sched_get_finished_job will not be called - * now until the scheduler thread is unparked. + * (earlier) cleanups and drm_sched_get_finished_job() will not be + * called now until the scheduler's work items are submitted again. */ if (bad && bad->sched == sched) /* @@ -620,7 +620,8 @@ void drm_sched_stop(struct drm_gpu_scheduler *sched, struct drm_sched_job *bad) * Iterate the job list from later to earlier one and either deactive * their HW callbacks or remove them from pending list if they already * signaled. - * This iteration is thread safe as sched thread is stopped. + * This iteration is thread safe as the scheduler's work items have been + * cancelled. */ list_for_each_entry_safe_reverse(s_job, tmp, &sched->pending_list, list) { @@ -667,49 +668,48 @@ void drm_sched_stop(struct drm_gpu_scheduler *sched, struct drm_sched_job *bad) */ cancel_delayed_work(&sched->work_tdr); } - EXPORT_SYMBOL(drm_sched_stop); /** * drm_sched_start - recover jobs after a reset * * @sched: scheduler instance - * @full_recovery: proceed with complete sched restart + * @errno: error to set on the pending fences + * + * This function is typically used for reset recovery (see the docu of + * drm_sched_backend_ops.timedout_job() for details). Do not call it for + * scheduler startup. The scheduler itself is fully operational after + * drm_sched_init() succeeded. * + * As it's only used for reset recovery, drivers must not call this function + * in their &struct drm_sched_backend_ops.timedout_job callback when they + * skip a reset using &enum drm_gpu_sched_stat.DRM_GPU_SCHED_STAT_NO_HANG. */ -void drm_sched_start(struct drm_gpu_scheduler *sched, bool full_recovery) +void drm_sched_start(struct drm_gpu_scheduler *sched, int errno) { struct drm_sched_job *s_job, *tmp; - int r; /* - * Locking the list is not required here as the sched thread is parked - * so no new jobs are being inserted or removed. Also concurrent - * GPU recovers can't run in parallel. + * Locking the list is not required here as the scheduler's work items + * are currently not running, so no new jobs are being inserted or + * removed. Also concurrent GPU recovers can't run in parallel. */ list_for_each_entry_safe(s_job, tmp, &sched->pending_list, list) { struct dma_fence *fence = s_job->s_fence->parent; atomic_add(s_job->credits, &sched->credit_count); - if (!full_recovery) + if (!fence) { + drm_sched_job_done(s_job, errno ?: -ECANCELED); continue; + } - if (fence) { - r = dma_fence_add_callback(fence, &s_job->cb, - drm_sched_job_done_cb); - if (r == -ENOENT) - drm_sched_job_done(s_job, fence->error); - else if (r) - DRM_DEV_ERROR(sched->dev, "fence add callback failed (%d)\n", - r); - } else - drm_sched_job_done(s_job, -ECANCELED); + if (dma_fence_add_callback(fence, &s_job->cb, + drm_sched_job_done_cb)) + drm_sched_job_done(s_job, fence->error ?: errno); } - if (full_recovery) - drm_sched_start_timeout_unlocked(sched); - + drm_sched_start_timeout_unlocked(sched); drm_sched_wqueue_start(sched); } EXPORT_SYMBOL(drm_sched_start); @@ -774,6 +774,8 @@ EXPORT_SYMBOL(drm_sched_resubmit_jobs); * @credits: the number of credits this job contributes to the schedulers * credit limit * @owner: job owner for debugging + * @drm_client_id: &struct drm_file.client_id of the owner (used by trace + * events) * * Refer to drm_sched_entity_push_job() documentation * for locking considerations. @@ -781,6 +783,10 @@ EXPORT_SYMBOL(drm_sched_resubmit_jobs); * Drivers must make sure drm_sched_job_cleanup() if this function returns * successfully, even when @job is aborted before drm_sched_job_arm() is called. * + * Note that this function does not assign a valid value to each struct member + * of struct drm_sched_job. Take a look at that struct's documentation to see + * who sets which struct member with what lifetime. + * * WARNING: amdgpu abuses &drm_sched.ready to signal when the hardware * has died, which can mean that there's no valid runqueue for a @entity. * This function returns -ENOENT in this case (which probably should be -EIO as @@ -790,14 +796,15 @@ EXPORT_SYMBOL(drm_sched_resubmit_jobs); */ int drm_sched_job_init(struct drm_sched_job *job, struct drm_sched_entity *entity, - u32 credits, void *owner) + u32 credits, void *owner, + uint64_t drm_client_id) { if (!entity->rq) { /* This will most likely be followed by missing frames * or worse--a blank screen--leave a trail in the * logs, so this can be debugged easier. */ - drm_err(job->sched, "%s: entity has no rq!\n", __func__); + dev_err(job->sched->dev, "%s: entity has no rq!\n", __func__); return -ENOENT; } @@ -806,9 +813,17 @@ int drm_sched_job_init(struct drm_sched_job *job, return -EINVAL; } + /* + * We don't know for sure how the user has allocated. Thus, zero the + * struct so that unallowed (i.e., too early) usage of pointers that + * this function does not set is guaranteed to lead to a NULL pointer + * exception instead of UB. + */ + memset(job, 0, sizeof(*job)); + job->entity = entity; job->credits = credits; - job->s_fence = drm_sched_fence_alloc(entity, owner); + job->s_fence = drm_sched_fence_alloc(entity, owner, drm_client_id); if (!job->s_fence) return -ENOMEM; @@ -826,11 +841,15 @@ EXPORT_SYMBOL(drm_sched_job_init); * * This arms a scheduler job for execution. Specifically it initializes the * &drm_sched_job.s_fence of @job, so that it can be attached to struct dma_resv - * or other places that need to track the completion of this job. + * or other places that need to track the completion of this job. It also + * initializes sequence numbers, which are fundamental for fence ordering. * * Refer to drm_sched_entity_push_job() documentation for locking * considerations. * + * Once this function was called, you *must* submit @job with + * drm_sched_entity_push_job(). + * * This can only be called if drm_sched_job_init() succeeded. */ void drm_sched_job_arm(struct drm_sched_job *job) @@ -844,7 +863,6 @@ void drm_sched_job_arm(struct drm_sched_job *job) job->sched = sched; job->s_priority = entity->priority; - job->id = atomic64_inc_return(&sched->job_id_count); drm_sched_fence_init(job->s_fence, job->entity); } @@ -947,13 +965,14 @@ int drm_sched_job_add_resv_dependencies(struct drm_sched_job *job, dma_resv_assert_held(resv); dma_resv_for_each_fence(&cursor, resv, usage, fence) { - /* Make sure to grab an additional ref on the added fence */ - dma_fence_get(fence); - ret = drm_sched_job_add_dependency(job, fence); - if (ret) { - dma_fence_put(fence); + /* + * As drm_sched_job_add_dependency always consumes the fence + * reference (even when it fails), and dma_resv_for_each_fence + * is not obtaining one, we need to grab one before calling. + */ + ret = drm_sched_job_add_dependency(job, dma_fence_get(fence)); + if (ret) return ret; - } } return 0; } @@ -984,6 +1003,29 @@ int drm_sched_job_add_implicit_dependencies(struct drm_sched_job *job, EXPORT_SYMBOL(drm_sched_job_add_implicit_dependencies); /** + * drm_sched_job_has_dependency - check whether fence is the job's dependency + * @job: scheduler job to check + * @fence: fence to look for + * + * Returns: + * True if @fence is found within the job's dependencies, or otherwise false. + */ +bool drm_sched_job_has_dependency(struct drm_sched_job *job, + struct dma_fence *fence) +{ + struct dma_fence *f; + unsigned long index; + + xa_for_each(&job->dependencies, index, f) { + if (f == fence) + return true; + } + + return false; +} +EXPORT_SYMBOL(drm_sched_job_has_dependency); + +/** * drm_sched_job_cleanup - clean up scheduler job resources * @job: scheduler job to clean up * @@ -992,9 +1034,12 @@ EXPORT_SYMBOL(drm_sched_job_add_implicit_dependencies); * Drivers should call this from their error unwind code if @job is aborted * before drm_sched_job_arm() is called. * - * After that point of no return @job is committed to be executed by the - * scheduler, and this function should be called from the - * &drm_sched_backend_ops.free_job callback. + * drm_sched_job_arm() is a point of no return since it initializes the fences + * and their sequence number etc. Once that function has been called, you *must* + * submit it with drm_sched_entity_push_job() and cannot simply abort it by + * calling drm_sched_job_cleanup(). + * + * This function should be called in the &drm_sched_backend_ops.free_job callback. */ void drm_sched_job_cleanup(struct drm_sched_job *job) { @@ -1002,10 +1047,15 @@ void drm_sched_job_cleanup(struct drm_sched_job *job) unsigned long index; if (kref_read(&job->s_fence->finished.refcount)) { - /* drm_sched_job_arm() has been called */ + /* The job has been processed by the scheduler, i.e., + * drm_sched_job_arm() and drm_sched_entity_push_job() have + * been called. + */ dma_fence_put(&job->s_fence->finished); } else { - /* aborted job before committing to run it */ + /* The job was aborted before it has been committed to be run; + * notably, drm_sched_job_arm() has not been called. + */ drm_sched_fence_free(job->s_fence); } @@ -1022,15 +1072,12 @@ EXPORT_SYMBOL(drm_sched_job_cleanup); /** * drm_sched_wakeup - Wake up the scheduler if it is ready to queue * @sched: scheduler instance - * @entity: the scheduler entity * * Wake up the scheduler if we can queue jobs. */ -void drm_sched_wakeup(struct drm_gpu_scheduler *sched, - struct drm_sched_entity *entity) +void drm_sched_wakeup(struct drm_gpu_scheduler *sched) { - if (drm_sched_can_queue(sched, entity)) - drm_sched_run_job_queue(sched); + drm_sched_run_job_queue(sched); } /** @@ -1067,12 +1114,16 @@ drm_sched_select_entity(struct drm_gpu_scheduler *sched) * drm_sched_get_finished_job - fetch the next finished job to be destroyed * * @sched: scheduler instance + * @have_more: are there more finished jobs on the list + * + * Informs the caller through @have_more whether there are more finished jobs + * besides the returned one. * * Returns the next finished job from the pending list (if there is one) * ready for it to be destroyed. */ static struct drm_sched_job * -drm_sched_get_finished_job(struct drm_gpu_scheduler *sched) +drm_sched_get_finished_job(struct drm_gpu_scheduler *sched, bool *have_more) { struct drm_sched_job *job, *next; @@ -1080,22 +1131,25 @@ drm_sched_get_finished_job(struct drm_gpu_scheduler *sched) job = list_first_entry_or_null(&sched->pending_list, struct drm_sched_job, list); - if (job && dma_fence_is_signaled(&job->s_fence->finished)) { /* remove job from pending_list */ list_del_init(&job->list); /* cancel this job's TO timer */ cancel_delayed_work(&sched->work_tdr); - /* make the scheduled timestamp more accurate */ + + *have_more = false; next = list_first_entry_or_null(&sched->pending_list, typeof(*next), list); - if (next) { + /* make the scheduled timestamp more accurate */ if (test_bit(DMA_FENCE_FLAG_TIMESTAMP_BIT, &next->s_fence->scheduled.flags)) next->s_fence->scheduled.timestamp = dma_fence_timestamp(&job->s_fence->finished); + + *have_more = dma_fence_is_signaled(&next->s_fence->finished); + /* start TO timer for next job */ drm_sched_start_timeout(sched); } @@ -1154,15 +1208,15 @@ static void drm_sched_free_job_work(struct work_struct *w) struct drm_gpu_scheduler *sched = container_of(w, struct drm_gpu_scheduler, work_free_job); struct drm_sched_job *job; + bool have_more; - if (READ_ONCE(sched->pause_submit)) - return; - - job = drm_sched_get_finished_job(sched); - if (job) + job = drm_sched_get_finished_job(sched, &have_more); + if (job) { sched->ops->free_job(job); + if (have_more) + drm_sched_run_free_queue(sched); + } - drm_sched_run_free_queue(sched); drm_sched_run_job_queue(sched); } @@ -1181,13 +1235,15 @@ static void drm_sched_run_job_work(struct work_struct *w) struct drm_sched_job *sched_job; int r; - if (READ_ONCE(sched->pause_submit)) - return; - /* Find entity with a ready job */ entity = drm_sched_select_entity(sched); - if (!entity) - return; /* No more work */ + if (!entity) { + /* + * Either no more work to do, or the next ready job needs more + * credits than the scheduler has currently available. + */ + return; + } sched_job = drm_sched_entity_pop_job(entity); if (!sched_job) { @@ -1201,21 +1257,24 @@ static void drm_sched_run_job_work(struct work_struct *w) atomic_add(sched_job->credits, &sched->credit_count); drm_sched_job_begin(sched_job); - trace_drm_run_job(sched_job, entity); + trace_drm_sched_job_run(sched_job, entity); + /* + * The run_job() callback must by definition return a fence whose + * refcount has been incremented for the scheduler already. + */ fence = sched->ops->run_job(sched_job); complete_all(&entity->entity_idle); drm_sched_fence_scheduled(s_fence, fence); if (!IS_ERR_OR_NULL(fence)) { - /* Drop for original kref_init of the fence */ - dma_fence_put(fence); - r = dma_fence_add_callback(fence, &sched_job->cb, drm_sched_job_done_cb); if (r == -ENOENT) drm_sched_job_done(sched_job, fence->error); else if (r) DRM_DEV_ERROR(sched->dev, "fence add callback failed (%d)\n", r); + + dma_fence_put(fence); } else { drm_sched_job_done(sched_job, IS_ERR(fence) ? PTR_ERR(fence) : 0); @@ -1225,47 +1284,50 @@ static void drm_sched_run_job_work(struct work_struct *w) drm_sched_run_job_queue(sched); } +static struct workqueue_struct *drm_sched_alloc_wq(const char *name) +{ +#if (IS_ENABLED(CONFIG_LOCKDEP)) + static struct lockdep_map map = { + .name = "drm_sched_lockdep_map" + }; + + /* + * Avoid leaking a lockdep map on each drm sched creation and + * destruction by using a single lockdep map for all drm sched + * allocated submit_wq. + */ + + return alloc_ordered_workqueue_lockdep_map(name, WQ_MEM_RECLAIM, &map); +#else + return alloc_ordered_workqueue(name, WQ_MEM_RECLAIM); +#endif +} + /** * drm_sched_init - Init a gpu scheduler instance * * @sched: scheduler instance - * @ops: backend operations for this scheduler - * @submit_wq: workqueue to use for submission. If NULL, an ordered wq is - * allocated and used - * @num_rqs: number of runqueues, one for each priority, up to DRM_SCHED_PRIORITY_COUNT - * @credit_limit: the number of credits this scheduler can hold from all jobs - * @hang_limit: number of times to allow a job to hang before dropping it - * @timeout: timeout value in jiffies for the scheduler - * @timeout_wq: workqueue to use for timeout work. If NULL, the system_wq is - * used - * @score: optional score atomic shared with other schedulers - * @name: name used for debugging - * @dev: target &struct device + * @args: scheduler initialization arguments * * Return 0 on success, otherwise error code. */ -int drm_sched_init(struct drm_gpu_scheduler *sched, - const struct drm_sched_backend_ops *ops, - struct workqueue_struct *submit_wq, - u32 num_rqs, u32 credit_limit, unsigned int hang_limit, - long timeout, struct workqueue_struct *timeout_wq, - atomic_t *score, const char *name, struct device *dev) +int drm_sched_init(struct drm_gpu_scheduler *sched, const struct drm_sched_init_args *args) { int i; - sched->ops = ops; - sched->credit_limit = credit_limit; - sched->name = name; - sched->timeout = timeout; - sched->timeout_wq = timeout_wq ? : system_wq; - sched->hang_limit = hang_limit; - sched->score = score ? score : &sched->_score; - sched->dev = dev; + sched->ops = args->ops; + sched->credit_limit = args->credit_limit; + sched->name = args->name; + sched->timeout = args->timeout; + sched->hang_limit = args->hang_limit; + sched->timeout_wq = args->timeout_wq ? args->timeout_wq : system_percpu_wq; + sched->score = args->score ? args->score : &sched->_score; + sched->dev = args->dev; - if (num_rqs > DRM_SCHED_PRIORITY_COUNT) { + if (args->num_rqs > DRM_SCHED_PRIORITY_COUNT) { /* This is a gross violation--tell drivers what the problem is. */ - drm_err(sched, "%s: num_rqs cannot be greater than DRM_SCHED_PRIORITY_COUNT\n", + dev_err(sched->dev, "%s: num_rqs cannot be greater than DRM_SCHED_PRIORITY_COUNT\n", __func__); return -EINVAL; } else if (sched->sched_rq) { @@ -1273,26 +1335,26 @@ int drm_sched_init(struct drm_gpu_scheduler *sched, * fine-tune their DRM calling order, and return all * is good. */ - drm_warn(sched, "%s: scheduler already initialized!\n", __func__); + dev_warn(sched->dev, "%s: scheduler already initialized!\n", __func__); return 0; } - if (submit_wq) { - sched->submit_wq = submit_wq; + if (args->submit_wq) { + sched->submit_wq = args->submit_wq; sched->own_submit_wq = false; } else { - sched->submit_wq = alloc_ordered_workqueue(name, 0); + sched->submit_wq = drm_sched_alloc_wq(args->name); if (!sched->submit_wq) return -ENOMEM; sched->own_submit_wq = true; } - sched->sched_rq = kmalloc_array(num_rqs, sizeof(*sched->sched_rq), + sched->sched_rq = kmalloc_array(args->num_rqs, sizeof(*sched->sched_rq), GFP_KERNEL | __GFP_ZERO); if (!sched->sched_rq) goto Out_check_own; - sched->num_rqs = num_rqs; + sched->num_rqs = args->num_rqs; for (i = DRM_SCHED_PRIORITY_KERNEL; i < sched->num_rqs; i++) { sched->sched_rq[i] = kzalloc(sizeof(*sched->sched_rq[i]), GFP_KERNEL); if (!sched->sched_rq[i]) @@ -1322,17 +1384,35 @@ Out_unroll: Out_check_own: if (sched->own_submit_wq) destroy_workqueue(sched->submit_wq); - drm_err(sched, "%s: Failed to setup GPU scheduler--out of memory\n", __func__); + dev_err(sched->dev, "%s: Failed to setup GPU scheduler--out of memory\n", __func__); return -ENOMEM; } EXPORT_SYMBOL(drm_sched_init); +static void drm_sched_cancel_remaining_jobs(struct drm_gpu_scheduler *sched) +{ + struct drm_sched_job *job, *tmp; + + /* All other accessors are stopped. No locking necessary. */ + list_for_each_entry_safe_reverse(job, tmp, &sched->pending_list, list) { + sched->ops->cancel_job(job); + list_del(&job->list); + sched->ops->free_job(job); + } +} + /** * drm_sched_fini - Destroy a gpu scheduler * * @sched: scheduler instance * * Tears down and cleans up the scheduler. + * + * This stops submission of new jobs to the hardware through &struct + * drm_sched_backend_ops.run_job. If &struct drm_sched_backend_ops.cancel_job + * is implemented, all jobs will be canceled through it and afterwards cleaned + * up through &struct drm_sched_backend_ops.free_job. If cancel_job is not + * implemented, memory could leak. */ void drm_sched_fini(struct drm_gpu_scheduler *sched) { @@ -1345,13 +1425,36 @@ void drm_sched_fini(struct drm_gpu_scheduler *sched) struct drm_sched_rq *rq = sched->sched_rq[i]; spin_lock(&rq->lock); - list_for_each_entry(s_entity, &rq->entities, list) + list_for_each_entry(s_entity, &rq->entities, list) { /* * Prevents reinsertion and marks job_queue as idle, - * it will removed from rq in drm_sched_entity_fini + * it will be removed from the rq in drm_sched_entity_fini() * eventually + * + * FIXME: + * This lacks the proper spin_lock(&s_entity->lock) and + * is, therefore, a race condition. Most notably, it + * can race with drm_sched_entity_push_job(). The lock + * cannot be taken here, however, because this would + * lead to lock inversion -> deadlock. + * + * The best solution probably is to enforce the life + * time rule of all entities having to be torn down + * before their scheduler. Then, however, locking could + * be dropped alltogether from this function. + * + * For now, this remains a potential race in all + * drivers that keep entities alive for longer than + * the scheduler. + * + * The READ_ONCE() is there to make the lockless read + * (warning about the lockless write below) slightly + * less broken... */ + if (!READ_ONCE(s_entity->stopped)) + dev_warn(sched->dev, "Tearing down scheduler with active entities!\n"); s_entity->stopped = true; + } spin_unlock(&rq->lock); kfree(sched->sched_rq[i]); } @@ -1362,11 +1465,18 @@ void drm_sched_fini(struct drm_gpu_scheduler *sched) /* Confirm no work left behind accessing device structures */ cancel_delayed_work_sync(&sched->work_tdr); + /* Avoid memory leaks if supported by the driver. */ + if (sched->ops->cancel_job) + drm_sched_cancel_remaining_jobs(sched); + if (sched->own_submit_wq) destroy_workqueue(sched->submit_wq); sched->ready = false; kfree(sched->sched_rq); sched->sched_rq = NULL; + + if (!list_empty(&sched->pending_list)) + dev_warn(sched->dev, "Tearing down scheduler while jobs are pending!\n"); } EXPORT_SYMBOL(drm_sched_fini); @@ -1428,8 +1538,10 @@ EXPORT_SYMBOL(drm_sched_wqueue_ready); /** * drm_sched_wqueue_stop - stop scheduler submission - * * @sched: scheduler instance + * + * Stops the scheduler from pulling new jobs from entities. It also stops + * freeing jobs automatically through drm_sched_backend_ops.free_job(). */ void drm_sched_wqueue_stop(struct drm_gpu_scheduler *sched) { @@ -1441,8 +1553,12 @@ EXPORT_SYMBOL(drm_sched_wqueue_stop); /** * drm_sched_wqueue_start - start scheduler submission - * * @sched: scheduler instance + * + * Restarts the scheduler after drm_sched_wqueue_stop() has stopped it. + * + * This function is not necessary for 'conventional' startup. The scheduler is + * fully operational after drm_sched_init() succeeded. */ void drm_sched_wqueue_start(struct drm_gpu_scheduler *sched) { |