summaryrefslogtreecommitdiff
path: root/include/drm/gpu_scheduler.h
blob: 9db9e5e504eeb32f5c527ccbbaa1773388af15ad (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
/*
 * Copyright 2015 Advanced Micro Devices, Inc.
 *
 * Permission is hereby granted, free of charge, to any person obtaining a
 * copy of this software and associated documentation files (the "Software"),
 * to deal in the Software without restriction, including without limitation
 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
 * and/or sell copies of the Software, and to permit persons to whom the
 * Software is furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in
 * all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
 * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
 * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
 * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
 * OTHER DEALINGS IN THE SOFTWARE.
 *
 */

#ifndef _DRM_GPU_SCHEDULER_H_
#define _DRM_GPU_SCHEDULER_H_

#include <drm/spsc_queue.h>
#include <linux/dma-fence.h>
#include <linux/completion.h>
#include <linux/xarray.h>
#include <linux/workqueue.h>

#define MAX_WAIT_SCHED_ENTITY_Q_EMPTY msecs_to_jiffies(1000)

/**
 * DRM_SCHED_FENCE_DONT_PIPELINE - Prefent dependency pipelining
 *
 * Setting this flag on a scheduler fence prevents pipelining of jobs depending
 * on this fence. In other words we always insert a full CPU round trip before
 * dependen jobs are pushed to the hw queue.
 */
#define DRM_SCHED_FENCE_DONT_PIPELINE	DMA_FENCE_FLAG_USER_BITS

enum dma_resv_usage;
struct dma_resv;
struct drm_gem_object;

struct drm_gpu_scheduler;
struct drm_sched_rq;

/* These are often used as an (initial) index
 * to an array, and as such should start at 0.
 */
enum drm_sched_priority {
	DRM_SCHED_PRIORITY_MIN,
	DRM_SCHED_PRIORITY_NORMAL,
	DRM_SCHED_PRIORITY_HIGH,
	DRM_SCHED_PRIORITY_KERNEL,

	DRM_SCHED_PRIORITY_COUNT,
	DRM_SCHED_PRIORITY_UNSET = -2
};

/* Used to chose between FIFO and RR jobs scheduling */
extern int drm_sched_policy;

#define DRM_SCHED_POLICY_RR    0
#define DRM_SCHED_POLICY_FIFO  1

/**
 * struct drm_sched_entity - A wrapper around a job queue (typically
 * attached to the DRM file_priv).
 *
 * Entities will emit jobs in order to their corresponding hardware
 * ring, and the scheduler will alternate between entities based on
 * scheduling policy.
 */
struct drm_sched_entity {
	/**
	 * @list:
	 *
	 * Used to append this struct to the list of entities in the runqueue
	 * @rq under &drm_sched_rq.entities.
	 *
	 * Protected by &drm_sched_rq.lock of @rq.
	 */
	struct list_head		list;

	/**
	 * @rq:
	 *
	 * Runqueue on which this entity is currently scheduled.
	 *
	 * FIXME: Locking is very unclear for this. Writers are protected by
	 * @rq_lock, but readers are generally lockless and seem to just race
	 * with not even a READ_ONCE.
	 */
	struct drm_sched_rq		*rq;

	/**
	 * @sched_list:
	 *
	 * A list of schedulers (struct drm_gpu_scheduler).  Jobs from this entity can
	 * be scheduled on any scheduler on this list.
	 *
	 * This can be modified by calling drm_sched_entity_modify_sched().
	 * Locking is entirely up to the driver, see the above function for more
	 * details.
	 *
	 * This will be set to NULL if &num_sched_list equals 1 and @rq has been
	 * set already.
	 *
	 * FIXME: This means priority changes through
	 * drm_sched_entity_set_priority() will be lost henceforth in this case.
	 */
	struct drm_gpu_scheduler        **sched_list;

	/**
	 * @num_sched_list:
	 *
	 * Number of drm_gpu_schedulers in the @sched_list.
	 */
	unsigned int                    num_sched_list;

	/**
	 * @priority:
	 *
	 * Priority of the entity. This can be modified by calling
	 * drm_sched_entity_set_priority(). Protected by &rq_lock.
	 */
	enum drm_sched_priority         priority;

	/**
	 * @rq_lock:
	 *
	 * Lock to modify the runqueue to which this entity belongs.
	 */
	spinlock_t			rq_lock;

	/**
	 * @job_queue: the list of jobs of this entity.
	 */
	struct spsc_queue		job_queue;

	/**
	 * @fence_seq:
	 *
	 * A linearly increasing seqno incremented with each new
	 * &drm_sched_fence which is part of the entity.
	 *
	 * FIXME: Callers of drm_sched_job_arm() need to ensure correct locking,
	 * this doesn't need to be atomic.
	 */
	atomic_t			fence_seq;

	/**
	 * @fence_context:
	 *
	 * A unique context for all the fences which belong to this entity.  The
	 * &drm_sched_fence.scheduled uses the fence_context but
	 * &drm_sched_fence.finished uses fence_context + 1.
	 */
	uint64_t			fence_context;

	/**
	 * @dependency:
	 *
	 * The dependency fence of the job which is on the top of the job queue.
	 */
	struct dma_fence		*dependency;

	/**
	 * @cb:
	 *
	 * Callback for the dependency fence above.
	 */
	struct dma_fence_cb		cb;

	/**
	 * @guilty:
	 *
	 * Points to entities' guilty.
	 */
	atomic_t			*guilty;

	/**
	 * @last_scheduled:
	 *
	 * Points to the finished fence of the last scheduled job. Only written
	 * by the scheduler thread, can be accessed locklessly from
	 * drm_sched_job_arm() iff the queue is empty.
	 */
	struct dma_fence                *last_scheduled;

	/**
	 * @last_user: last group leader pushing a job into the entity.
	 */
	struct task_struct		*last_user;

	/**
	 * @stopped:
	 *
	 * Marks the enity as removed from rq and destined for
	 * termination. This is set by calling drm_sched_entity_flush() and by
	 * drm_sched_fini().
	 */
	bool 				stopped;

	/**
	 * @entity_idle:
	 *
	 * Signals when entity is not in use, used to sequence entity cleanup in
	 * drm_sched_entity_fini().
	 */
	struct completion		entity_idle;

	/**
	 * @oldest_job_waiting:
	 *
	 * Marks earliest job waiting in SW queue
	 */
	ktime_t				oldest_job_waiting;

	/**
	 * @rb_tree_node:
	 *
	 * The node used to insert this entity into time based priority queue
	 */
	struct rb_node			rb_tree_node;

	/**
	 * @elapsed_ns:
	 *
	 * Records the amount of time where jobs from this entity were active
	 * on the GPU.
	 */
	uint64_t elapsed_ns;
};

/**
 * struct drm_sched_rq - queue of entities to be scheduled.
 *
 * @lock: to modify the entities list.
 * @sched: the scheduler to which this rq belongs to.
 * @entities: list of the entities to be scheduled.
 * @current_entity: the entity which is to be scheduled.
 * @rb_tree_root: root of time based priory queue of entities for FIFO scheduling
 *
 * Run queue is a set of entities scheduling command submissions for
 * one specific ring. It implements the scheduling policy that selects
 * the next entity to emit commands from.
 */
struct drm_sched_rq {
	spinlock_t			lock;
	struct drm_gpu_scheduler	*sched;
	struct list_head		entities;
	struct drm_sched_entity		*current_entity;
	struct rb_root_cached		rb_tree_root;
};

/**
 * struct drm_sched_fence - fences corresponding to the scheduling of a job.
 */
struct drm_sched_fence {
        /**
         * @scheduled: this fence is what will be signaled by the scheduler
         * when the job is scheduled.
         */
	struct dma_fence		scheduled;

        /**
         * @finished: this fence is what will be signaled by the scheduler
         * when the job is completed.
         *
         * When setting up an out fence for the job, you should use
         * this, since it's available immediately upon
         * drm_sched_job_init(), and the fence returned by the driver
         * from run_job() won't be created until the dependencies have
         * resolved.
         */
	struct dma_fence		finished;

        /**
         * @parent: the fence returned by &drm_sched_backend_ops.run_job
         * when scheduling the job on hardware. We signal the
         * &drm_sched_fence.finished fence once parent is signalled.
         */
	struct dma_fence		*parent;
        /**
         * @sched: the scheduler instance to which the job having this struct
         * belongs to.
         */
	struct drm_gpu_scheduler	*sched;
        /**
         * @lock: the lock used by the scheduled and the finished fences.
         */
	spinlock_t			lock;
        /**
         * @owner: job owner for debugging
         */
	void				*owner;
};

struct drm_sched_fence *to_drm_sched_fence(struct dma_fence *f);

/**
 * struct drm_sched_job - A job to be run by an entity.
 *
 * @queue_node: used to append this struct to the queue of jobs in an entity.
 * @list: a job participates in a "pending" and "done" lists.
 * @sched: the scheduler instance on which this job is scheduled.
 * @s_fence: contains the fences for the scheduling of job.
 * @finish_cb: the callback for the finished fence.
 * @work: Helper to reschdeule job kill to different context.
 * @id: a unique id assigned to each job scheduled on the scheduler.
 * @karma: increment on every hang caused by this job. If this exceeds the hang
 *         limit of the scheduler then the job is marked guilty and will not
 *         be scheduled further.
 * @s_priority: the priority of the job.
 * @entity: the entity to which this job belongs.
 * @cb: the callback for the parent fence in s_fence.
 *
 * A job is created by the driver using drm_sched_job_init(), and
 * should call drm_sched_entity_push_job() once it wants the scheduler
 * to schedule the job.
 */
struct drm_sched_job {
	struct spsc_node		queue_node;
	struct list_head		list;
	struct drm_gpu_scheduler	*sched;
	struct drm_sched_fence		*s_fence;

	/*
	 * work is used only after finish_cb has been used and will not be
	 * accessed anymore.
	 */
	union {
		struct dma_fence_cb		finish_cb;
		struct work_struct		work;
	};

	uint64_t			id;
	atomic_t			karma;
	enum drm_sched_priority		s_priority;
	struct drm_sched_entity         *entity;
	struct dma_fence_cb		cb;
	/**
	 * @dependencies:
	 *
	 * Contains the dependencies as struct dma_fence for this job, see
	 * drm_sched_job_add_dependency() and
	 * drm_sched_job_add_implicit_dependencies().
	 */
	struct xarray			dependencies;

	/** @last_dependency: tracks @dependencies as they signal */
	unsigned long			last_dependency;

	/**
	 * @submit_ts:
	 *
	 * When the job was pushed into the entity queue.
	 */
	ktime_t                         submit_ts;
};

static inline bool drm_sched_invalidate_job(struct drm_sched_job *s_job,
					    int threshold)
{
	return s_job && atomic_inc_return(&s_job->karma) > threshold;
}

enum drm_gpu_sched_stat {
	DRM_GPU_SCHED_STAT_NONE, /* Reserve 0 */
	DRM_GPU_SCHED_STAT_NOMINAL,
	DRM_GPU_SCHED_STAT_ENODEV,
};

/**
 * struct drm_sched_backend_ops - Define the backend operations
 *	called by the scheduler
 *
 * These functions should be implemented in the driver side.
 */
struct drm_sched_backend_ops {
	/**
	 * @prepare_job:
	 *
	 * Called when the scheduler is considering scheduling this job next, to
	 * get another struct dma_fence for this job to block on.  Once it
	 * returns NULL, run_job() may be called.
	 *
	 * Can be NULL if no additional preparation to the dependencies are
	 * necessary. Skipped when jobs are killed instead of run.
	 */
	struct dma_fence *(*prepare_job)(struct drm_sched_job *sched_job,
					 struct drm_sched_entity *s_entity);

	/**
         * @run_job: Called to execute the job once all of the dependencies
         * have been resolved.  This may be called multiple times, if
	 * timedout_job() has happened and drm_sched_job_recovery()
	 * decides to try it again.
	 */
	struct dma_fence *(*run_job)(struct drm_sched_job *sched_job);

	/**
	 * @timedout_job: Called when a job has taken too long to execute,
	 * to trigger GPU recovery.
	 *
	 * This method is called in a workqueue context.
	 *
	 * Drivers typically issue a reset to recover from GPU hangs, and this
	 * procedure usually follows the following workflow:
	 *
	 * 1. Stop the scheduler using drm_sched_stop(). This will park the
	 *    scheduler thread and cancel the timeout work, guaranteeing that
	 *    nothing is queued while we reset the hardware queue
	 * 2. Try to gracefully stop non-faulty jobs (optional)
	 * 3. Issue a GPU reset (driver-specific)
	 * 4. Re-submit jobs using drm_sched_resubmit_jobs()
	 * 5. Restart the scheduler using drm_sched_start(). At that point, new
	 *    jobs can be queued, and the scheduler thread is unblocked
	 *
	 * Note that some GPUs have distinct hardware queues but need to reset
	 * the GPU globally, which requires extra synchronization between the
	 * timeout handler of the different &drm_gpu_scheduler. One way to
	 * achieve this synchronization is to create an ordered workqueue
	 * (using alloc_ordered_workqueue()) at the driver level, and pass this
	 * queue to drm_sched_init(), to guarantee that timeout handlers are
	 * executed sequentially. The above workflow needs to be slightly
	 * adjusted in that case:
	 *
	 * 1. Stop all schedulers impacted by the reset using drm_sched_stop()
	 * 2. Try to gracefully stop non-faulty jobs on all queues impacted by
	 *    the reset (optional)
	 * 3. Issue a GPU reset on all faulty queues (driver-specific)
	 * 4. Re-submit jobs on all schedulers impacted by the reset using
	 *    drm_sched_resubmit_jobs()
	 * 5. Restart all schedulers that were stopped in step #1 using
	 *    drm_sched_start()
	 *
	 * Return DRM_GPU_SCHED_STAT_NOMINAL, when all is normal,
	 * and the underlying driver has started or completed recovery.
	 *
	 * Return DRM_GPU_SCHED_STAT_ENODEV, if the device is no longer
	 * available, i.e. has been unplugged.
	 */
	enum drm_gpu_sched_stat (*timedout_job)(struct drm_sched_job *sched_job);

	/**
         * @free_job: Called once the job's finished fence has been signaled
         * and it's time to clean it up.
	 */
	void (*free_job)(struct drm_sched_job *sched_job);
};

/**
 * struct drm_gpu_scheduler - scheduler instance-specific data
 *
 * @ops: backend operations provided by the driver.
 * @hw_submission_limit: the max size of the hardware queue.
 * @timeout: the time after which a job is removed from the scheduler.
 * @name: name of the ring for which this scheduler is being used.
 * @sched_rq: priority wise array of run queues.
 * @wake_up_worker: the wait queue on which the scheduler sleeps until a job
 *                  is ready to be scheduled.
 * @job_scheduled: once @drm_sched_entity_do_release is called the scheduler
 *                 waits on this wait queue until all the scheduled jobs are
 *                 finished.
 * @hw_rq_count: the number of jobs currently in the hardware queue.
 * @job_id_count: used to assign unique id to the each job.
 * @timeout_wq: workqueue used to queue @work_tdr
 * @work_tdr: schedules a delayed call to @drm_sched_job_timedout after the
 *            timeout interval is over.
 * @thread: the kthread on which the scheduler which run.
 * @pending_list: the list of jobs which are currently in the job queue.
 * @job_list_lock: lock to protect the pending_list.
 * @hang_limit: once the hangs by a job crosses this limit then it is marked
 *              guilty and it will no longer be considered for scheduling.
 * @score: score to help loadbalancer pick a idle sched
 * @_score: score used when the driver doesn't provide one
 * @ready: marks if the underlying HW is ready to work
 * @free_guilty: A hit to time out handler to free the guilty job.
 * @dev: system &struct device
 *
 * One scheduler is implemented for each hardware ring.
 */
struct drm_gpu_scheduler {
	const struct drm_sched_backend_ops	*ops;
	uint32_t			hw_submission_limit;
	long				timeout;
	const char			*name;
	struct drm_sched_rq		sched_rq[DRM_SCHED_PRIORITY_COUNT];
	wait_queue_head_t		wake_up_worker;
	wait_queue_head_t		job_scheduled;
	atomic_t			hw_rq_count;
	atomic64_t			job_id_count;
	struct workqueue_struct		*timeout_wq;
	struct delayed_work		work_tdr;
	struct task_struct		*thread;
	struct list_head		pending_list;
	spinlock_t			job_list_lock;
	int				hang_limit;
	atomic_t                        *score;
	atomic_t                        _score;
	bool				ready;
	bool				free_guilty;
	struct device			*dev;
};

int drm_sched_init(struct drm_gpu_scheduler *sched,
		   const struct drm_sched_backend_ops *ops,
		   uint32_t hw_submission, unsigned hang_limit,
		   long timeout, struct workqueue_struct *timeout_wq,
		   atomic_t *score, const char *name, struct device *dev);

void drm_sched_fini(struct drm_gpu_scheduler *sched);
int drm_sched_job_init(struct drm_sched_job *job,
		       struct drm_sched_entity *entity,
		       void *owner);
void drm_sched_job_arm(struct drm_sched_job *job);
int drm_sched_job_add_dependency(struct drm_sched_job *job,
				 struct dma_fence *fence);
int drm_sched_job_add_resv_dependencies(struct drm_sched_job *job,
					struct dma_resv *resv,
					enum dma_resv_usage usage);
int drm_sched_job_add_implicit_dependencies(struct drm_sched_job *job,
					    struct drm_gem_object *obj,
					    bool write);


void drm_sched_entity_modify_sched(struct drm_sched_entity *entity,
				    struct drm_gpu_scheduler **sched_list,
                                   unsigned int num_sched_list);

void drm_sched_job_cleanup(struct drm_sched_job *job);
void drm_sched_wakeup(struct drm_gpu_scheduler *sched);
void drm_sched_stop(struct drm_gpu_scheduler *sched, struct drm_sched_job *bad);
void drm_sched_start(struct drm_gpu_scheduler *sched, bool full_recovery);
void drm_sched_resubmit_jobs(struct drm_gpu_scheduler *sched);
void drm_sched_increase_karma(struct drm_sched_job *bad);
void drm_sched_reset_karma(struct drm_sched_job *bad);
void drm_sched_increase_karma_ext(struct drm_sched_job *bad, int type);
bool drm_sched_dependency_optimized(struct dma_fence* fence,
				    struct drm_sched_entity *entity);
void drm_sched_fault(struct drm_gpu_scheduler *sched);

void drm_sched_rq_add_entity(struct drm_sched_rq *rq,
			     struct drm_sched_entity *entity);
void drm_sched_rq_remove_entity(struct drm_sched_rq *rq,
				struct drm_sched_entity *entity);

void drm_sched_rq_update_fifo(struct drm_sched_entity *entity, ktime_t ts);

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);
long drm_sched_entity_flush(struct drm_sched_entity *entity, long timeout);
void drm_sched_entity_fini(struct drm_sched_entity *entity);
void drm_sched_entity_destroy(struct drm_sched_entity *entity);
void drm_sched_entity_select_rq(struct drm_sched_entity *entity);
struct drm_sched_job *drm_sched_entity_pop_job(struct drm_sched_entity *entity);
void drm_sched_entity_push_job(struct drm_sched_job *sched_job);
void drm_sched_entity_set_priority(struct drm_sched_entity *entity,
				   enum drm_sched_priority priority);
bool drm_sched_entity_is_ready(struct drm_sched_entity *entity);

struct drm_sched_fence *drm_sched_fence_alloc(
	struct drm_sched_entity *s_entity, void *owner);
void drm_sched_fence_init(struct drm_sched_fence *fence,
			  struct drm_sched_entity *entity);
void drm_sched_fence_free(struct drm_sched_fence *fence);

void drm_sched_fence_scheduled(struct drm_sched_fence *fence);
void drm_sched_fence_finished(struct drm_sched_fence *fence);

unsigned long drm_sched_suspend_timeout(struct drm_gpu_scheduler *sched);
void drm_sched_resume_timeout(struct drm_gpu_scheduler *sched,
		                unsigned long remaining);
struct drm_gpu_scheduler *
drm_sched_pick_best(struct drm_gpu_scheduler **sched_list,
		     unsigned int num_sched_list);

#endif