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
|
/*
* Copyright © 2008-2018 Intel Corporation
*
* 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 (including the next
* paragraph) 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 AUTHORS OR COPYRIGHT HOLDERS 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 I915_REQUEST_H
#define I915_REQUEST_H
#include <linux/dma-fence.h>
#include <linux/irq_work.h>
#include <linux/lockdep.h>
#include "gem/i915_gem_context_types.h"
#include "gt/intel_context_types.h"
#include "gt/intel_engine_types.h"
#include "gt/intel_timeline_types.h"
#include "i915_gem.h"
#include "i915_scheduler.h"
#include "i915_selftest.h"
#include "i915_sw_fence.h"
#include <uapi/drm/i915_drm.h>
struct drm_file;
struct drm_i915_gem_object;
struct i915_request;
struct i915_capture_list {
struct i915_capture_list *next;
struct i915_vma *vma;
};
#define RQ_TRACE(rq, fmt, ...) do { \
const struct i915_request *rq__ = (rq); \
ENGINE_TRACE(rq__->engine, "fence %llx:%lld, current %d " fmt, \
rq__->fence.context, rq__->fence.seqno, \
hwsp_seqno(rq__), ##__VA_ARGS__); \
} while (0)
enum {
/*
* I915_FENCE_FLAG_ACTIVE - this request is currently submitted to HW.
*
* Set by __i915_request_submit() on handing over to HW, and cleared
* by __i915_request_unsubmit() if we preempt this request.
*
* Finally cleared for consistency on retiring the request, when
* we know the HW is no longer running this request.
*
* See i915_request_is_active()
*/
I915_FENCE_FLAG_ACTIVE = DMA_FENCE_FLAG_USER_BITS,
/*
* I915_FENCE_FLAG_PQUEUE - this request is ready for execution
*
* Using the scheduler, when a request is ready for execution it is put
* into the priority queue, and removed from that queue when transferred
* to the HW runlists. We want to track its membership within the
* priority queue so that we can easily check before rescheduling.
*
* See i915_request_in_priority_queue()
*/
I915_FENCE_FLAG_PQUEUE,
/*
* I915_FENCE_FLAG_SIGNAL - this request is currently on signal_list
*
* Internal bookkeeping used by the breadcrumb code to track when
* a request is on the various signal_list.
*/
I915_FENCE_FLAG_SIGNAL,
/*
* I915_FENCE_FLAG_HOLD - this request is currently on hold
*
* This request has been suspended, pending an ongoing investigation.
*/
I915_FENCE_FLAG_HOLD,
/*
* I915_FENCE_FLAG_NOPREEMPT - this request should not be preempted
*
* The execution of some requests should not be interrupted. This is
* a sensitive operation as it makes the request super important,
* blocking other higher priority work. Abuse of this flag will
* lead to quality of service issues.
*/
I915_FENCE_FLAG_NOPREEMPT,
/*
* I915_FENCE_FLAG_SENTINEL - this request should be last in the queue
*
* A high priority sentinel request may be submitted to clear the
* submission queue. As it will be the only request in-flight, upon
* execution all other active requests will have been preempted and
* unsubmitted. This preemptive pulse is used to re-evaluate the
* in-flight requests, particularly in cases where an active context
* is banned and those active requests need to be cancelled.
*/
I915_FENCE_FLAG_SENTINEL,
/*
* I915_FENCE_FLAG_BOOST - upclock the gpu for this request
*
* Some requests are more important than others! In particular, a
* request that the user is waiting on is typically required for
* interactive latency, for which we want to minimise by upclocking
* the GPU. Here we track such boost requests on a per-request basis.
*/
I915_FENCE_FLAG_BOOST,
};
/**
* Request queue structure.
*
* The request queue allows us to note sequence numbers that have been emitted
* and may be associated with active buffers to be retired.
*
* By keeping this list, we can avoid having to do questionable sequence
* number comparisons on buffer last_read|write_seqno. It also allows an
* emission time to be associated with the request for tracking how far ahead
* of the GPU the submission is.
*
* When modifying this structure be very aware that we perform a lockless
* RCU lookup of it that may race against reallocation of the struct
* from the slab freelist. We intentionally do not zero the structure on
* allocation so that the lookup can use the dangling pointers (and is
* cogniscent that those pointers may be wrong). Instead, everything that
* needs to be initialised must be done so explicitly.
*
* The requests are reference counted.
*/
struct i915_request {
struct dma_fence fence;
spinlock_t lock;
/** On Which ring this request was generated */
struct drm_i915_private *i915;
/**
* Context and ring buffer related to this request
* Contexts are refcounted, so when this request is associated with a
* context, we must increment the context's refcount, to guarantee that
* it persists while any request is linked to it. Requests themselves
* are also refcounted, so the request will only be freed when the last
* reference to it is dismissed, and the code in
* i915_request_free() will then decrement the refcount on the
* context.
*/
struct intel_engine_cs *engine;
struct intel_context *context;
struct intel_ring *ring;
struct intel_timeline __rcu *timeline;
struct list_head signal_link;
/*
* The rcu epoch of when this request was allocated. Used to judiciously
* apply backpressure on future allocations to ensure that under
* mempressure there is sufficient RCU ticks for us to reclaim our
* RCU protected slabs.
*/
unsigned long rcustate;
/*
* We pin the timeline->mutex while constructing the request to
* ensure that no caller accidentally drops it during construction.
* The timeline->mutex must be held to ensure that only this caller
* can use the ring and manipulate the associated timeline during
* construction.
*/
struct pin_cookie cookie;
/*
* Fences for the various phases in the request's lifetime.
*
* The submit fence is used to await upon all of the request's
* dependencies. When it is signaled, the request is ready to run.
* It is used by the driver to then queue the request for execution.
*/
struct i915_sw_fence submit;
union {
wait_queue_entry_t submitq;
struct i915_sw_dma_fence_cb dmaq;
struct i915_request_duration_cb {
struct dma_fence_cb cb;
ktime_t emitted;
} duration;
};
struct list_head execute_cb;
struct i915_sw_fence semaphore;
struct irq_work semaphore_work;
/*
* A list of everyone we wait upon, and everyone who waits upon us.
* Even though we will not be submitted to the hardware before the
* submit fence is signaled (it waits for all external events as well
* as our own requests), the scheduler still needs to know the
* dependency tree for the lifetime of the request (from execbuf
* to retirement), i.e. bidirectional dependency information for the
* request not tied to individual fences.
*/
struct i915_sched_node sched;
struct i915_dependency dep;
intel_engine_mask_t execution_mask;
/*
* A convenience pointer to the current breadcrumb value stored in
* the HW status page (or our timeline's local equivalent). The full
* path would be rq->hw_context->ring->timeline->hwsp_seqno.
*/
const u32 *hwsp_seqno;
/*
* If we need to access the timeline's seqno for this request in
* another request, we need to keep a read reference to this associated
* cacheline, so that we do not free and recycle it before the foreign
* observers have completed. Hence, we keep a pointer to the cacheline
* inside the timeline's HWSP vma, but it is only valid while this
* request has not completed and guarded by the timeline mutex.
*/
struct intel_timeline_cacheline __rcu *hwsp_cacheline;
/** Position in the ring of the start of the request */
u32 head;
/** Position in the ring of the start of the user packets */
u32 infix;
/**
* Position in the ring of the start of the postfix.
* This is required to calculate the maximum available ring space
* without overwriting the postfix.
*/
u32 postfix;
/** Position in the ring of the end of the whole request */
u32 tail;
/** Position in the ring of the end of any workarounds after the tail */
u32 wa_tail;
/** Preallocate space in the ring for the emitting the request */
u32 reserved_space;
/** Batch buffer related to this request if any (used for
* error state dump only).
*/
struct i915_vma *batch;
/**
* Additional buffers requested by userspace to be captured upon
* a GPU hang. The vma/obj on this list are protected by their
* active reference - all objects on this list must also be
* on the active_list (of their final request).
*/
struct i915_capture_list *capture_list;
/** Time at which this request was emitted, in jiffies. */
unsigned long emitted_jiffies;
/** timeline->request entry for this request */
struct list_head link;
struct drm_i915_file_private *file_priv;
/** file_priv list entry for this request */
struct list_head client_link;
I915_SELFTEST_DECLARE(struct {
struct list_head link;
unsigned long delay;
} mock;)
};
#define I915_FENCE_GFP (GFP_KERNEL | __GFP_RETRY_MAYFAIL | __GFP_NOWARN)
extern const struct dma_fence_ops i915_fence_ops;
static inline bool dma_fence_is_i915(const struct dma_fence *fence)
{
return fence->ops == &i915_fence_ops;
}
struct i915_request * __must_check
__i915_request_create(struct intel_context *ce, gfp_t gfp);
struct i915_request * __must_check
i915_request_create(struct intel_context *ce);
struct i915_request *__i915_request_commit(struct i915_request *request);
void __i915_request_queue(struct i915_request *rq,
const struct i915_sched_attr *attr);
bool i915_request_retire(struct i915_request *rq);
void i915_request_retire_upto(struct i915_request *rq);
static inline struct i915_request *
to_request(struct dma_fence *fence)
{
/* We assume that NULL fence/request are interoperable */
BUILD_BUG_ON(offsetof(struct i915_request, fence) != 0);
GEM_BUG_ON(fence && !dma_fence_is_i915(fence));
return container_of(fence, struct i915_request, fence);
}
static inline struct i915_request *
i915_request_get(struct i915_request *rq)
{
return to_request(dma_fence_get(&rq->fence));
}
static inline struct i915_request *
i915_request_get_rcu(struct i915_request *rq)
{
return to_request(dma_fence_get_rcu(&rq->fence));
}
static inline void
i915_request_put(struct i915_request *rq)
{
dma_fence_put(&rq->fence);
}
int i915_request_await_object(struct i915_request *to,
struct drm_i915_gem_object *obj,
bool write);
int i915_request_await_dma_fence(struct i915_request *rq,
struct dma_fence *fence);
int i915_request_await_execution(struct i915_request *rq,
struct dma_fence *fence,
void (*hook)(struct i915_request *rq,
struct dma_fence *signal));
void i915_request_add(struct i915_request *rq);
bool __i915_request_submit(struct i915_request *request);
void i915_request_submit(struct i915_request *request);
void i915_request_skip(struct i915_request *request, int error);
void __i915_request_unsubmit(struct i915_request *request);
void i915_request_unsubmit(struct i915_request *request);
/* Note: part of the intel_breadcrumbs family */
bool i915_request_enable_breadcrumb(struct i915_request *request);
void i915_request_cancel_breadcrumb(struct i915_request
|