summaryrefslogtreecommitdiff
path: root/lib/igt_dummyload.c
blob: ddd434518cb0cf22ef839b1e668801fc43e4c247 (plain)
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
/*
 * Copyright © 2016 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.
 *
 */

#include <time.h>
#include <signal.h>
#include <pthread.h>

#include <i915_drm.h>

#include "igt_core.h"
#include "drmtest.h"
#include "igt_dummyload.h"
#include "igt_gt.h"
#include "intel_chipset.h"
#include "intel_reg.h"
#include "ioctl_wrappers.h"
#include "sw_sync.h"
#include "igt_vgem.h"

/**
 * SECTION:igt_dummyload
 * @short_description: Library for submitting GPU workloads
 * @title: Dummyload
 * @include: igt.h
 *
 * A lot of igt testcases need some GPU workload to make sure a race window is
 * big enough. Unfortunately having a fixed amount of workload leads to
 * spurious test failures or overly long runtimes on some fast/slow platforms.
 * This library contains functionality to submit GPU workloads that should
 * consume exactly a specific amount of time.
 */

#define LOCAL_I915_EXEC_BSD_SHIFT      (13)
#define LOCAL_I915_EXEC_BSD_MASK       (3 << LOCAL_I915_EXEC_BSD_SHIFT)

#define ENGINE_MASK  (I915_EXEC_RING_MASK | LOCAL_I915_EXEC_BSD_MASK)

#define MI_ARB_CHK (0x5 << 23)

static const int BATCH_SIZE = 4096;
static IGT_LIST(spin_list);
static pthread_mutex_t list_lock = PTHREAD_MUTEX_INITIALIZER;

static void
fill_reloc(struct drm_i915_gem_relocation_entry *reloc,
	   uint32_t gem_handle, uint32_t offset,
	   uint32_t read_domains, uint32_t write_domains)
{
	reloc->target_handle = gem_handle;
	reloc->offset = offset * sizeof(uint32_t);
	reloc->read_domains = read_domains;
	reloc->write_domain = write_domains;
}

static int emit_recursive_batch(igt_spin_t *spin,
				int fd, uint32_t ctx, unsigned engine,
				uint32_t dep, bool out_fence)
{
#define SCRATCH 0
#define BATCH 1
	const int gen = intel_gen(intel_get_drm_devid(fd));
	struct drm_i915_gem_exec_object2 obj[2];
	struct drm_i915_gem_relocation_entry relocs[2];
	struct drm_i915_gem_execbuffer2 execbuf;
	unsigned int engines[16];
	unsigned int nengine;
	int fence_fd = -1;
	uint32_t *batch;
	int i;

	nengine = 0;
	if (engine == -1) {
		for_each_engine(fd, engine)
			if (engine)
				engines[nengine++] = engine;
	} else {
		gem_require_ring(fd, engine);
		engines[nengine++] = engine;
	}
	igt_require(nengine);

	memset(&execbuf, 0, sizeof(execbuf));
	memset(obj, 0, sizeof(obj));
	memset(relocs, 0, sizeof(relocs));

	obj[BATCH].handle = gem_create(fd, BATCH_SIZE);
	batch = __gem_mmap__wc(fd, obj[BATCH].handle,
			       0, BATCH_SIZE, PROT_WRITE);
	if (!batch)
		batch = __gem_mmap__gtt(fd, obj[BATCH].handle,
				       	BATCH_SIZE, PROT_WRITE);
	gem_set_domain(fd, obj[BATCH].handle,
			I915_GEM_DOMAIN_GTT, I915_GEM_DOMAIN_GTT);
	execbuf.buffer_count++;

	if (dep) {
		/* dummy write to dependency */
		obj[SCRATCH].handle = dep;
		fill_reloc(&relocs[obj[BATCH].relocation_count++],
			   dep, 1020,
			   I915_GEM_DOMAIN_RENDER,
			   I915_GEM_DOMAIN_RENDER);
		execbuf.buffer_count++;
	}

	spin->batch = batch;
	spin->handle = obj[BATCH].handle;

	/* Allow ourselves to be preempted */
	*batch++ = MI_ARB_CHK;

	/* Pad with a few nops so that we do not completely hog the system.
	 *
	 * Part of the attraction of using a recursive batch is that it is
	 * hard on the system (executing the "function" call is apparently
	 * quite expensive). However, the GPU may hog the entire system for
	 * a few minutes, preventing even NMI. Quite why this is so is unclear,
	 * but presumably it relates to the PM_INTRMSK workaround on gen6/gen7.
	 * If we give the system a break by having the GPU execute a few nops
	 * between function calls, that appears enough to keep SNB out of
	 * trouble. See https://bugs.freedesktop.org/show_bug.cgi?id=102262
	 */
	batch += 1000;

	/* recurse */
	fill_reloc(&relocs[obj[BATCH].relocation_count],
		   obj[BATCH].handle, (batch - spin->batch) + 1,
		   I915_GEM_DOMAIN_COMMAND, 0);
	if (gen >= 8) {
		*batch++ = MI_BATCH_BUFFER_START | 1 << 8 | 1;
		*batch++ = 0;
		*batch++ = 0;
	} else if (gen >= 6) {
		*batch++ = MI_BATCH_BUFFER_START | 1 << 8;
		*batch++ = 0;
	} else {
		*batch++ = MI_BATCH_BUFFER_START | 2 << 6;
		*batch = 0;
		if (gen < 4) {
			*batch |= 1;
			relocs[obj[BATCH].relocation_count].delta = 1;
		}
		batch++;
	}
	obj[BATCH].relocation_count++;
	obj[BATCH].relocs_ptr = to_user_pointer(relocs);

	execbuf.buffers_ptr = to_user_pointer(obj + (2 - execbuf.buffer_count));
	execbuf.rsvd1 = ctx;

	if (out_fence)
		execbuf.flags |= I915_EXEC_FENCE_OUT;

	for (i = 0; i < nengine; i++) {
		execbuf.flags &= ~ENGINE_MASK;
		execbuf.flags |= engines[i];
		gem_execbuf_wr(fd, &execbuf);
		if (out_fence) {
			int _fd = execbuf.rsvd2 >> 32;

			igt_assert(_fd >= 0);
			if (fence_fd == -1) {
				fence_fd = _fd;
			} else {
				int old_fd = fence_fd;

				fence_fd = sync_fence_merge(old_fd, _fd);
				close(old_fd);
				close(_fd);
			}
			igt_assert(fence_fd >= 0);
		}
	}

	return fence_fd;
}

static igt_spin_t *
___igt_spin_batch_new(int fd, uint32_t ctx, unsigned engine, uint32_t dep,
		      int out_fence)
{
	igt_spin_t *spin;

	spin = calloc(1, sizeof(struct igt_spin));
	igt_assert(spin);

	spin->out_fence = emit_recursive_batch(spin, fd, ctx, engine, dep,
					       out_fence);
	igt_assert(gem_bo_busy(fd, spin->handle));

	pthread_mutex_lock(&list_lock);
	igt_list_add(&spin->link, &spin_list);
	pthread_mutex_unlock(&list_lock);

	return spin;
}

igt_spin_t *
__igt_spin_batch_new(int fd, uint32_t ctx, unsigned engine, uint32_t dep)
{
	return ___igt_spin_batch_new(fd, ctx, engine, dep, false);
}

/**
 * igt_spin_batch_new:
 * @fd: open i915 drm file descriptor
 * @engine: Ring to execute batch OR'd with execbuf flags. If value is less
 *          than 0, execute on all available rings.
 * @dep: handle to a buffer object dependency. If greater than 0, add a
 *              relocation entry to this buffer within the batch.
 *
 * Start a recursive batch on a ring. Immediately returns a #igt_spin_t that
 * contains the batch's handle that can be waited upon. The returned structure
 * must be passed to igt_spin_batch_free() for post-processing.
 *
 * Returns:
 * Structure with helper internal state for igt_spin_batch_free().
 */
igt_spin_t *
igt_spin_batch_new(int fd, uint32_t ctx, unsigned engine, uint32_t dep)
{
	igt_require_gem(fd);

	return __igt_spin_batch_new(fd, ctx, engine, dep);
}

igt_spin_t *
__igt_spin_batch_new_fence(int fd, uint32_t ctx, unsigned engine)
{
	return ___igt_spin_batch_new(fd, ctx, engine, 0, true);
}

/**
 * igt_spin_batch_new_fence:
 * @fd: open i915 drm file descriptor
 * @engine: Ring to execute batch OR'd with execbuf flags. If value is less
 *          than 0, execute on all available rings.
 *
 * Start a recursive batch on a ring. Immediately returns a #igt_spin_t that
 * contains the batch's handle that can be waited upon. The returned structure
 * must be passed to igt_spin_batch_free() for post-processing.
 *
 * igt_spin_t will contain an output fence associtated with this batch.
 *
 * Returns:
 * Structure with helper internal state for igt_spin_batch_free().
 */
igt_spin_t *
igt_spin_batch_new_fence(int fd, uint32_t ctx, unsigned engine)
{
	igt_require_gem(fd);
	igt_require(gem_has_exec_fence(fd));

	return __igt_spin_batch_new_fence(fd, ctx, engine);
}

static void notify(union sigval arg)
{
	igt_spin_t *spin = arg.sival_ptr;

	igt_spin_batch_end(spin);
}

/**
 * igt_spin_batch_set_timeout:
 * @spin: spin batch state from igt_spin_batch_new()
 * @ns: amount of time in nanoseconds the batch continues to execute
 *      before finishing.
 *
 * Specify a timeout. This ends the recursive batch associated with @spin after
 * the timeout has elapsed.
 */
void igt_spin_batch_set_timeout(igt_spin_t *spin, int64_t ns)
{
	timer_t timer;
	struct sigevent sev;
	struct itimerspec its;

	igt_assert(ns > 0);
	if (!spin)
		return;

	igt_assert(!spin->timer);

	memset(&sev, 0, sizeof(sev));
	sev.sigev_notify = SIGEV_THREAD;
	sev.sigev_value.sival_ptr = spin;
	sev.sigev_notify_function = notify;
	igt_assert(timer_create(CLOCK_MONOTONIC, &sev, &timer) == 0);
	igt_assert(timer);

	memset(&its, 0, sizeof(its));
	its.it_value.tv_sec = ns / NSEC_PER_SEC;
	its.it_value.tv_nsec = ns % NSEC_PER_SEC;
	igt_assert(timer_settime(timer, 0, &its, NULL) == 0);

	spin->timer = timer;
}

/**
 * igt_spin_batch_end:
 * @spin: spin batch state from igt_spin_batch_new()
 *
 * End the recursive batch associated with @spin manually.
 */
void igt_spin_batch_end(igt_spin_t *spin)
{
	if (!spin)
		return;

	*spin->batch = MI_BATCH_BUFFER_END;
	__sync_synchronize();
}

/**
 * igt_spin_batch_free:
 * @fd: open i915 drm file descriptor
 * @spin: spin batch state from igt_spin_batch_new()
 *
 * This function does the necessary post-processing after starting a recursive
 * batch with igt_spin_batch_new().
 */
void igt_spin_batch_free(int fd, igt_spin_t *spin)
{
	if (!spin)
		return;

	pthread_mutex_lock(&list_lock);
	igt_list_del(&spin->link);
	pthread_mutex_unlock(&list_lock);

	if (spin->timer)
		timer_delete(spin->timer);

	igt_spin_batch_end(spin);
	gem_munmap(spin->batch, BATCH_SIZE);

	gem_close(fd, spin->handle);

	if (spin->out_fence >= 0)
		close(spin->out_fence);

	free(spin);
}

void igt_terminate_spin_batches(void)
{
	struct igt_spin *iter;

	pthread_mutex_lock(&list_lock);
	igt_list_for_each(iter, &spin_list, link)
		igt_spin_batch_end(iter);
	pthread_mutex_unlock(&list_lock);
}

static uint32_t plug_vgem_handle(struct igt_cork *cork, int fd)
{
	struct vgem_bo bo;
	int dmabuf;
	uint32_t handle;

	cork->vgem.device = drm_open_driver(DRIVER_VGEM);
	igt_require(vgem_has_fences(cork->vgem.device));

	bo.width = bo.height = 1;
	bo.bpp = 4;
	vgem_create(cork->vgem.device, &bo);
	cork->vgem.fence = vgem_fence_attach(cork->vgem.device, &bo, VGEM_FENCE_WRITE);

	dmabuf = prime_handle_to_fd(cork->vgem.device, bo.handle);
	handle = prime_fd_to_handle(fd, dmabuf);
	close(dmabuf);

	return handle;
}

static void unplug_vgem_handle(struct igt_cork *cork)
{
	vgem_fence_signal(cork->vgem.device, cork->vgem.fence);
	close(cork->vgem.device);
}

static uint32_t plug_sync_fd(struct igt_cork *cork)
{
	int fence;

	igt_require_sw_sync();

	cork->sw_sync.timeline = sw_sync_timeline_create();
	fence = sw_sync_timeline_create_fence(cork->sw_sync.timeline, 1);

	return fence;
}

static void unplug_sync_fd(struct igt_cork *cork)
{
	sw_sync_timeline_inc(cork->sw_sync.timeline, 1);
	close(cork->sw_sync.timeline);
}

/**
 * igt_cork_plug:
 * @fd: open drm file descriptor
 * @method: method to utilize for corking.
 * @cork: structure that will be filled with the state of the cork bo.
 * Note: this has to match the corking method.
 *
 * This function provides a mechanism to stall submission. It provides two
 * blocking methods:
 *
 * VGEM_BO.
 * Imports a vgem bo with a fence attached to it. This bo can be used as a
 * dependency during submission to stall execution until the fence is signaled.
 *
 * SW_SYNC:
 * Creates a timeline and then a fence on that timeline. The fence can be used
 * as an input fence to a request, the request will be stalled until the fence
 * is signaled.
 *
 * The parameters required to unblock the execution and to cleanup are stored in
 * the provided cork structure.
 *
 * Returns:
 * Handle of the imported BO / Sw sync fence FD.
 */
uint32_t igt_cork_plug(struct igt_cork *cork, int fd)
{
	igt_assert(cork->fd == -1);

	switch (cork->type) {
	case CORK_SYNC_FD:
		return plug_sync_fd(cork);

	case CORK_VGEM_HANDLE:
		return plug_vgem_handle(cork, fd);

	default:
		igt_assert_f(0, "Invalid cork type!\n");
		return 0;
	}
}

/**
 * igt_cork_unplug:
 * @method: method to utilize for corking.
 * @cork: cork state from igt_cork_plug()
 *
 * This function unblocks the execution by signaling the fence attached to the
 * imported bo and does the necessary post-processing.
 *
 * NOTE: the handle returned by igt_cork_plug is not closed during this phase.
 */
void igt_cork_unplug(struct igt_cork *cork)
{
	igt_assert(cork->fd != -1);

	switch (cork->type) {
	case CORK_SYNC_FD:
		unplug_sync_fd(cork);
		break;

	case CORK_VGEM_HANDLE:
		unplug_vgem_handle(cork);
		break;

	default:
		igt_assert_f(0, "Invalid cork type!\n");
	}

	cork->fd = -1; /* Reset cork */
}