#include <linux/kthread.h>
#include <linux/slab.h>
#include <linux/completion.h>
#include <drm/drm_print.h>
#include <drm/gpu_scheduler.h>
#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_entity_init(struct drm_sched_entity *entity,
enum drm_sched_priority priority,
struct drm_gpu_scheduler **sched_list,
unsigned int num_sched_list,
atomic_t *guilty)
{
if (!(entity && sched_list && (num_sched_list == 0 || sched_list[0])))
return -EINVAL;
memset(entity, 0, sizeof(struct drm_sched_entity));
INIT_LIST_HEAD(&entity->list);
entity->rq = NULL;
entity->guilty = guilty;
entity->num_sched_list = num_sched_list;
entity->priority = priority;
entity->sched_list = num_sched_list > 1 ? sched_list : NULL;
RCU_INIT_POINTER(entity->last_scheduled, NULL);
RB_CLEAR_NODE(&entity->rb_tree_node);
if(num_sched_list)
entity->rq = &sched_list[0]->sched_rq[entity->priority];
init_completion(&entity->entity_idle);
complete_all(&entity->entity_idle);
spin_lock_init(&entity->rq_lock);
spsc_queue_init(&entity->job_queue);
atomic_set(&entity->fence_seq, 0);
entity->fence_context = dma_fence_context_alloc(2);
return 0;
}
EXPORT_SYMBOL(drm_sched_entity_init);
void drm_sched_entity_modify_sched(struct drm_sched_entity *entity,
struct drm_gpu_scheduler **sched_list,
unsigned int num_sched_list)
{
WARN_ON(!num_sched_list || !sched_list);
entity->sched_list = sched_list;
entity->num_sched_list = num_sched_list;
}
EXPORT_SYMBOL(drm_sched_entity_modify_sched);
static bool drm_sched_entity_is_idle(struct drm_sched_entity *entity)
{
rmb();
if (list_empty(&entity->list) ||
spsc_queue_count(&entity->job_queue) == 0 ||
entity->stopped)
return true;
return false;
}
bool drm_sched_entity_is_ready(struct drm_sched_entity *entity)
{
if (spsc_queue_peek(&entity->job_queue) == NULL)
return false;
if (READ_ONCE(entity->dependency))
return false;
return true;
}
int drm_sched_entity_error(struct drm_sched_entity *entity)
{
struct dma_fence *fence;
int r;
rcu_read_lock();
fence = rcu_dereference(entity->last_scheduled);
r = fence ? fence->error : 0;
rcu_read_unlock();
return r;
}
EXPORT_SYMBOL(drm_sched_entity_error);
static void drm_sched_entity_kill_jobs_work(struct work_struct *wrk)
{
struct drm_sched_job *job = container_of(wrk, typeof(*job), work);
drm_sched_fence_finished(job->s_fence, -ESRCH);
WARN_ON(job->s_fence->parent);
job->sched->ops->free_job(job);
}
static void drm_sched_entity_kill_jobs_cb(struct dma_fence *f,
struct dma_fence_cb *cb)
{
struct drm_sched_job *job = container_of(cb, struct drm_sched_job,
finish_cb);
unsigned long index;
dma_fence_put(f);
xa_for_each(&job->dependencies, index, f) {
struct drm_sched_fence *s_fence = to_drm_sched_fence(f);
if (s_fence && f == &s_fence->scheduled) {
f = dma_fence_get_rcu(&s_fence->finished);
dma_fence_put(&s_fence->scheduled);
}
xa_erase(&job->dependencies, index);
if (f && !dma_fence_add_callback(f, &job->finish_cb,
drm_sched_entity_kill_jobs_cb))
return;
dma_fence_put(f);
}
INIT_WORK(&job->work, drm_sched_entity_kill_jobs_work);
schedule_work(&job->work);
}
static void drm_sched_entity_kill(struct drm_sched_entity *entity)
{
struct drm_sched_job *job;
struct dma_fence *prev;
if (!entity->rq)
return;
spin_lock(&entity->rq_lock);
entity->stopped = true;
drm_sched_rq_remove_entity(entity->rq, entity);
spin_unlock(&entity->rq_lock);
wait_for_completion(&entity->entity_idle);
prev = rcu_dereference_check(entity->last_scheduled, true);
dma_fence_get(prev);
while ((job = to_drm_sched_job(spsc_queue_pop(&entity->job_queue)))) {
struct drm_sched_fence *s_fence = job->s_fence;
dma_fence_get(&s_fence->finished);
if (!prev || dma_fence_add_callback(prev, &job->finish_cb,
drm_sched_entity_kill_jobs_cb))
drm_sched_entity_kill_jobs_cb(NULL, &job->finish_cb);
prev = &s_fence->finished;
}
dma_fence_put(prev);
}
long drm_sched_entity_flush(struct drm_sched_entity *entity, long timeout)
{
struct drm_gpu_scheduler *sched;
struct task_struct *last_user;
long ret = timeout;
if (!entity->rq)
return 0;
sched = entity->rq->sched;
if (current->flags & PF_EXITING) {
if (timeout)
ret = wait_event_timeout(
sched->job_scheduled,
drm_sched_entity_is_idle(entity),
timeout);
} else {
wait_event_killable(sched->job_scheduled,
drm_sched_entity_is_idle(entity));
}
last_user = cmpxchg(&entity->last_user, current->group_leader, NULL);
if ((!last_user || last_user == current->group_leader) &&
(current->flags & PF_EXITING) && (current->exit_code == SIGKILL))
drm_sched_entity_kill(entity);
return ret;
}
EXPORT_SYMBOL(drm_sched_entity_flush);
void drm_sched_entity_fini(struct drm_sched_entity *entity)
{
drm_sched_entity_kill(entity);
if (entity->dependency) {
dma_fence_remove_callback(entity->dependency, &entity->cb);
dma_fence_put(entity->dependency);
entity->dependency = NULL;
}
dma_fence_put(rcu_dereference_check(entity->last_scheduled, true));
RCU_INIT_POINTER(entity->last_scheduled, NULL);
}
EXPORT_SYMBOL(drm_sched_entity_fini);
void drm_sched_entity_destroy(struct drm_sched_entity *entity)
{
drm_sched_entity_flush(entity, MAX_WAIT_SCHED_ENTITY_Q_EMPTY);
drm_sched_entity_fini(entity);
}
EXPORT_SYMBOL(drm_sched_entity_destroy);
static void drm_sched_entity_clear_dep(struct dma_fence *f,
struct dma_fence_cb *cb)
{
struct drm_sched_entity *entity =
container_of(cb, struct drm_sched_entity, cb);
entity->dependency = NULL;
dma_fence_put(f);
}
static void drm_sched_entity_wakeup(struct dma_fence *f,
struct dma_fence_cb *cb)
{
struct drm_sched_entity *entity =
container_of(cb, struct drm_sched_entity, cb);
drm_sched_entity_clear_dep(f, cb);
drm_sched_wakeup_if_can_queue(entity->rq->sched);
}
void drm_sched_entity_set_priority(struct drm_sched_entity *entity,
enum drm_sched_priority priority)
{
spin_lock(&entity->rq_lock);
entity->priority = priority;
spin_unlock(&entity->rq_lock);
}
EXPORT_SYMBOL(drm_sched_entity_set_priority);
static bool drm_sched_entity_add_dependency_cb(struct drm_sched_entity *entity)
{
struct drm_gpu_scheduler *sched = entity->rq->sched;
struct dma_fence *fence = entity->dependency;
struct drm_sched_fence *s_fence;
if (fence->context == entity->fence_context ||
fence->context == entity->fence_context + 1) {
dma_fence_put(entity->dependency);
return false;
}
s_fence = to_drm_sched_fence(fence);
if (!fence->error && s_fence && s_fence->sched == sched &&
!test_bit(DRM_SCHED_FENCE_DONT_PIPELINE, &fence->flags)) {
fence = dma_fence_get(&s_fence->scheduled);
dma_fence_put(entity->dependency);
entity->dependency = fence;
if (!dma_fence_add_callback(fence, &entity->cb,
drm_sched_entity_clear_dep))
return true;
dma_fence_put(fence);
return false;
}
if (!dma_fence_add_callback(entity->dependency, &entity->cb,
drm_sched_entity_wakeup))
return true;
dma_fence_put(entity->dependency);
return false;
}
static struct dma_fence *
drm_sched_job_dependency(struct drm_sched_job *job,
struct drm_sched_entity *entity)
{
struct dma_fence *f;
f = xa_load(&job->dependencies, job->last_dependency);
if (f) {
job->last_dependency++;
return dma_fence_get(f);
}
if (job->sched->ops->prepare_job)
return job->sched->ops->prepare_job(job, entity);
return NULL;
}
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));
if (!sched_job)
return NULL;
while ((entity->dependency =
drm_sched_job_dependency(sched_job, entity))) {
trace_drm_sched_job_wait_dep(sched_job, entity->dependency);
if (drm_sched_entity_add_dependency_cb(entity))
return NULL;
}
if (entity->guilty && atomic_read(entity->guilty))
dma_fence_set_error(&sched_job->s_fence->finished, -ECANCELED);
dma_fence_put(rcu_dereference_check(entity->last_scheduled, true));
rcu_assign_pointer(entity->last_scheduled,
dma_fence_get(&sched_job->s_fence->finished));
smp_wmb();
spsc_queue_pop(&entity->job_queue);
if (drm_sched_policy == DRM_SCHED_POLICY_FIFO) {
struct drm_sched_job *next;
next = to_drm_sched_job(spsc_queue_peek(&entity->job_queue));
if (next)
drm_sched_rq_update_fifo(entity, next->submit_ts);
}
sched_job->entity = NULL;
return sched_job;
}
void drm_sched_entity_select_rq(struct drm_sched_entity *entity)
{
struct dma_fence *fence;
struct drm_gpu_scheduler *sched;
struct drm_sched_rq *rq;
if (!entity->sched_list)
return;
if (spsc_queue_count(&entity->job_queue))
return;
smp_rmb();
fence = rcu_dereference_check(entity->last_scheduled, true);
if (fence && !dma_fence_is_signaled(fence))
return;
spin_lock(&entity->rq_lock);
sched = drm_sched_pick_best(entity->sched_list, entity->num_sched_list);
rq = sched ? &sched->sched_rq[entity->priority] : NULL;
if (rq != entity->rq) {
drm_sched_rq_remove_entity(entity->rq, entity);
entity->rq = rq;
}
spin_unlock(&entity->rq_lock);
if (entity->num_sched_list == 1)
entity->sched_list = NULL;
}
void drm_sched_entity_push_job(struct drm_sched_job *sched_job)
{
struct drm_sched_entity *entity = sched_job->entity;
bool first;
ktime_t submit_ts;
trace_drm_sched_job(sched_job, entity);
atomic_inc(entity->rq->sched->score);
WRITE_ONCE(entity->last_user, current->group_leader);
sched_job->submit_ts = submit_ts = ktime_get();
first = spsc_queue_push(&entity->job_queue, &sched_job->queue_node);
if (first) {
spin_lock(&entity->rq_lock);
if (entity->stopped) {
spin_unlock(&entity->rq_lock);
DRM_ERROR("Trying to push to a killed entity\n");
return;
}
drm_sched_rq_add_entity(entity->rq, entity);
spin_unlock(&entity->rq_lock);
if (drm_sched_policy == DRM_SCHED_POLICY_FIFO)
drm_sched_rq_update_fifo(entity, submit_ts);
drm_sched_wakeup_if_can_queue(entity->rq->sched);
}
}
EXPORT_SYMBOL