summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/drm_suballoc.c
blob: 38cc7a1238198f72bafcfbbb33b0388c40bfeb9d (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
// SPDX-License-Identifier: GPL-2.0 OR MIT
/*
 * Copyright 2011 Red Hat Inc.
 * Copyright 2023 Intel Corporation.
 * All Rights Reserved.
 *
 * 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, sub license, 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 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 NON-INFRINGEMENT. IN NO EVENT SHALL
 * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS 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.
 *
 * The above copyright notice and this permission notice (including the
 * next paragraph) shall be included in all copies or substantial portions
 * of the Software.
 *
 */
/* Algorithm:
 *
 * We store the last allocated bo in "hole", we always try to allocate
 * after the last allocated bo. Principle is that in a linear GPU ring
 * progression was is after last is the oldest bo we allocated and thus
 * the first one that should no longer be in use by the GPU.
 *
 * If it's not the case we skip over the bo after last to the closest
 * done bo if such one exist. If none exist and we are not asked to
 * block we report failure to allocate.
 *
 * If we are asked to block we wait on all the oldest fence of all
 * rings. We just wait for any of those fence to complete.
 */

#include <drm/drm_suballoc.h>
#include <drm/drm_print.h>
#include <linux/slab.h>
#include <linux/sched.h>
#include <linux/wait.h>
#include <linux/dma-fence.h>

static void drm_suballoc_remove_locked(struct drm_suballoc *sa);
static void drm_suballoc_try_free(struct drm_suballoc_manager *sa_manager);

/**
 * drm_suballoc_manager_init() - Initialise the drm_suballoc_manager
 * @sa_manager: pointer to the sa_manager
 * @size: number of bytes we want to suballocate
 * @align: alignment for each suballocated chunk
 *
 * Prepares the suballocation manager for suballocations.
 */
void drm_suballoc_manager_init(struct drm_suballoc_manager *sa_manager,
			       size_t size, size_t align)
{
	unsigned int i;

	BUILD_BUG_ON(!is_power_of_2(DRM_SUBALLOC_MAX_QUEUES));

	if (!align)
		align = 1;

	/* alignment must be a power of 2 */
	if (WARN_ON_ONCE(align & (align - 1)))
		align = roundup_pow_of_two(align);

	init_waitqueue_head(&sa_manager->wq);
	sa_manager->size = size;
	sa_manager->align = align;
	sa_manager->hole = &sa_manager->olist;
	INIT_LIST_HEAD(&sa_manager->olist);
	for (i = 0; i < DRM_SUBALLOC_MAX_QUEUES; ++i)
		INIT_LIST_HEAD(&sa_manager->flist[i]);
}
EXPORT_SYMBOL(drm_suballoc_manager_init);

/**
 * drm_suballoc_manager_fini() - Destroy the drm_suballoc_manager
 * @sa_manager: pointer to the sa_manager
 *
 * Cleans up the suballocation manager after use. All fences added
 * with drm_suballoc_free() must be signaled, or we cannot clean up
 * the entire manager.
 */
void drm_suballoc_manager_fini(struct drm_suballoc_manager *sa_manager)
{
	struct drm_suballoc *sa, *tmp;

	if (!sa_manager->size)
		return;

	if (!list_empty(&sa_manager->olist)) {
		sa_manager->hole = &sa_manager->olist;
		drm_suballoc_try_free(sa_manager);
		if (!list_empty(&sa_manager->olist))
			DRM_ERROR("sa_manager is not empty, clearing anyway\n");
	}
	list_for_each_entry_safe(sa, tmp, &sa_manager->olist, olist) {
		drm_suballoc_remove_locked(sa);
	}

	sa_manager->size = 0;
}
EXPORT_SYMBOL(drm_suballoc_manager_fini);

static void drm_suballoc_remove_locked(struct drm_suballoc *sa)
{
	struct drm_suballoc_manager *sa_manager = sa->manager;

	if (sa_manager->hole == &sa->olist)
		sa_manager->hole = sa->olist.prev;

	list_del_init(&sa->olist);
	list_del_init(&sa->flist);
	dma_fence_put(sa->fence);
	kfree(sa);
}

static void drm_suballoc_try_free(struct drm_suballoc_manager *sa_manager)
{
	struct drm_suballoc *sa, *tmp;

	if (sa_manager->hole->next == &sa_manager->olist)
		return;

	sa = list_entry(sa_manager->hole->next, struct drm_suballoc, olist);
	list_for_each_entry_safe_from(sa, tmp, &sa_manager->olist, olist) {
		if (!sa->fence || !dma_fence_is_signaled(sa->fence))
			return;

		drm_suballoc_remove_locked(sa);
	}
}

static size_t drm_suballoc_hole_soffset(struct drm_suballoc_manager *sa_manager)
{
	struct list_head *hole = sa_manager->hole;

	if (hole != &sa_manager->olist)
		return list_entry(hole, struct drm_suballoc, olist)->eoffset;

	return 0;
}

static size_t drm_suballoc_hole_eoffset(struct drm_suballoc_manager *sa_manager)
{
	struct list_head *hole = sa_manager->hole;

	if (hole->next != &sa_manager->olist)
		return list_entry(hole->next, struct drm_suballoc, olist)->soffset;
	return sa_manager->size;
}

static bool drm_suballoc_try_alloc(struct drm_suballoc_manager *sa_manager,
				   struct drm_suballoc *sa,
				   size_t size, size_t align)
{
	size_t soffset, eoffset, wasted;

	soffset = drm_suballoc_hole_soffset(sa_manager);
	eoffset = drm_suballoc_hole_eoffset(sa_manager);
	wasted = round_up(soffset, align) - soffset;

	if ((eoffset - soffset) >= (size + wasted)) {
		soffset += wasted;

		sa->manager = sa_manager;
		sa->soffset = soffset;
		sa->eoffset = soffset + size;
		list_add(&sa->olist, sa_manager->hole);
		INIT_LIST_HEAD(&sa->flist);
		sa_manager->hole = &sa->olist;
		return true;
	}
	return false;
}

static bool __drm_suballoc_event(struct drm_suballoc_manager *sa_manager,
				 size_t size, size_t align)
{
	size_t soffset, eoffset, wasted;
	unsigned int i;

	for (i = 0; i < DRM_SUBALLOC_MAX_QUEUES; ++i)
		if (!list_empty(&sa_manager->flist[i]))
			return true;

	soffset = drm_suballoc_hole_soffset(sa_manager);
	eoffset = drm_suballoc_hole_eoffset(sa_manager);
	wasted = round_up(soffset, align) - soffset;

	return ((eoffset - soffset) >= (size + wasted));
}

/**
 * drm_suballoc_event() - Check if we can stop waiting
 * @sa_manager: pointer to the sa_manager
 * @size: number of bytes we want to allocate
 * @align: alignment we need to match
 *
 * Return: true if either there is a fence we can wait for or
 * enough free memory to satisfy the allocation directly.
 * false otherwise.
 */
static bool drm_suballoc_event(struct drm_suballoc_manager *sa_manager,
			       size_t size, size_t align)
{
	bool ret;

	spin_lock(&sa_manager->wq.lock);
	ret = __drm_suballoc_event(sa_manager, size, align);
	spin_unlock(&sa_manager->wq.lock);
	return ret;
}

static bool drm_suballoc_next_hole(struct drm_suballoc_manager *sa_manager,
				   struct dma_fence **fences,
				   unsigned int *tries)
{
	struct drm_suballoc *best_bo = NULL;
	unsigned int i, best_idx;
	size_t soffset, best, tmp;

	/* if hole points to the end of the buffer */
	if (sa_manager->hole->next == &sa_manager->olist) {
		/* try again with its beginning */
		sa_manager->hole = &sa_manager->olist;
		return true;
	}

	soffset = drm_suballoc_hole_soffset(sa_manager);
	/* to handle wrap around we add sa_manager->size */
	best = sa_manager->size * 2;
	/* go over all fence list and try to find the closest sa
	 * of the current last
	 */
	for (i = 0; i < DRM_SUBALLOC_MAX_QUEUES; ++i) {
		struct drm_suballoc *sa;

		fences[i] = NULL;

		if (list_empty(&sa_manager->flist[i]))
			continue;

		sa = list_first_entry(&sa_manager->flist[i],
				      struct drm_suballoc, flist);

		if (!dma_fence_is_signaled(sa->fence)) {
			fences[i] = sa->fence;
			continue;
		}

		/* limit the number of tries each freelist gets */
		if (tries[i] > 2)
			continue;

		tmp = sa->soffset;
		if (tmp < soffset) {
			/* wrap around, pretend it's after */
			tmp += sa_manager->size;
		}
		tmp -= soffset;
		if (tmp < best) {
			/* this sa bo is the closest one */
			best = tmp;
			best_idx = i;
			best_bo = sa;
		}
	}

	if (best_bo) {
		++tries[best_idx];
		sa_manager->hole = best_bo->olist.prev;

		/*
		 * We know that this one is signaled,
		 * so it's safe to remove it.
		 */
		drm_suballoc_remove_locked(best_bo);
		return true;
	}
	return false;
}

/**
 * drm_suballoc_new() - Make a suballocation.
 * @sa_manager: pointer to the sa_manager
 * @size: number of bytes we want to suballocate.
 * @gfp: gfp flags used for memory allocation. Typically GFP_KERNEL but
 *       the argument is provided for suballocations from reclaim context or
 *       where the caller wants to avoid pipelining rather than wait for
 *       reclaim.
 * @intr: Whether to perform waits interruptible. This should typically
 *        always be true, unless the caller needs to propagate a
 *        non-interruptible context from above layers.
 * @align: Alignment. Must not exceed the default manager alignment.
 *         If @align is zero, then the manager alignment is used.
 *
 * Try to make a suballocation of size @size, which will be rounded
 * up to the alignment specified in specified in drm_suballoc_manager_init().
 *
 * Return: a new suballocated bo, or an ERR_PTR.
 */
struct drm_suballoc *
drm_suballoc_new(struct drm_suballoc_manager *sa_manager, size_t size,
		 gfp_t gfp, bool intr, size_t align)
{
	struct dma_fence *fences[DRM_SUBALLOC_MAX_QUEUES];
	unsigned int tries[DRM_SUBALLOC_MAX_QUEUES];
	unsigned int count;
	int i, r;
	struct drm_suballoc *sa;

	if (WARN_ON_ONCE(align > sa_manager->align))
		return ERR_PTR(-EINVAL);
	if (WARN_ON_ONCE(size > sa_manager->size || !size))
		return ERR_PTR(-EINVAL);

	if (!align)
		align = sa_manager->align;

	sa = kmalloc(sizeof(*sa), gfp);
	if (!sa)
		return ERR_PTR(-ENOMEM);
	sa->manager = sa_manager;
	sa->fence = NULL;
	INIT_LIST_HEAD(&sa->olist);
	INIT_LIST_HEAD(&sa->flist);

	spin_lock(&sa_manager->wq.lock);
	do {
		for (i = 0; i < DRM_SUBALLOC_MAX_QUEUES; ++i)
			tries[i] = 0;

		do {
			drm_suballoc_try_free(sa_manager);

			if (drm_suballoc_try_alloc(sa_manager, sa,
						   size, align)) {
				spin_unlock(&sa_manager->wq.lock);
				return sa;
			}

			/* see if we can skip over some allocations */
		} while (drm_suballoc_next_hole(sa_manager, fences, tries));

		for (i = 0, count = 0; i < DRM_SUBALLOC_MAX_QUEUES; ++i)
			if (fences[i])
				fences[count++] = dma_fence_get(fences[i]);

		if (count) {
			long t;

			spin_unlock(&sa_manager->wq.lock);
			t = dma_fence_wait_any_timeout(fences, count, intr,
						       MAX_SCHEDULE_TIMEOUT,
						       NULL);
			for (i = 0; i < count; ++i)
				dma_fence_put(fences[i]);

			r = (t > 0) ? 0 : t;
			spin_lock(&sa_manager->wq.lock);
		} else if (intr) {
			/* if we have nothing to wait for block */
			r = wait_event_interruptible_locked
				(sa_manager->wq,
				 __drm_suballoc_event(sa_manager, size, align));
		} else {
			spin_unlock(&sa_manager->wq.lock);
			wait_event(sa_manager->wq,
				   drm_suballoc_event(sa_manager, size, align));
			r = 0;
			spin_lock(&sa_manager->wq.lock);
		}
	} while (!r);

	spin_unlock(&sa_manager->wq.lock);
	kfree(sa);
	return ERR_PTR(r);
}
EXPORT_SYMBOL(drm_suballoc_new);

/**
 * drm_suballoc_free - Free a suballocation
 * @suballoc: pointer to the suballocation
 * @fence: fence that signals when suballocation is idle
 *
 * Free the suballocation. The suballocation can be re-used after @fence signals.
 */
void drm_suballoc_free(struct drm_suballoc *suballoc,
		       struct dma_fence *fence)
{
	struct drm_suballoc_manager *sa_manager;

	if (!suballoc)
		return;

	sa_manager = suballoc->manager;

	spin_lock(&sa_manager->wq.lock);
	if (fence && !dma_fence_is_signaled(fence)) {
		u32 idx;

		suballoc->fence = dma_fence_get(fence);
		idx = fence->context & (DRM_SUBALLOC_MAX_QUEUES - 1);
		list_add_tail(&suballoc->flist, &sa_manager->flist[idx]);
	} else {
		drm_suballoc_remove_locked(suballoc);
	}
	wake_up_all_locked(&sa_manager->wq);
	spin_unlock(&sa_manager->wq.lock);
}
EXPORT_SYMBOL(drm_suballoc_free);

#ifdef CONFIG_DEBUG_FS
void drm_suballoc_dump_debug_info(struct drm_suballoc_manager *sa_manager,
				  struct drm_printer *p,
				  unsigned long long suballoc_base)
{
	struct drm_suballoc *i;

	spin_lock(&sa_manager->wq.lock);
	list_for_each_entry(i, &sa_manager->olist, olist) {
		unsigned long long soffset = i->soffset;
		unsigned long long eoffset = i->eoffset;

		if (&i->olist == sa_manager->hole)
			drm_puts(p, ">");
		else
			drm_puts(p, " ");

		drm_printf(p, "[0x%010llx 0x%010llx] size %8lld",
			   suballoc_base + soffset, suballoc_base + eoffset,
			   eoffset - soffset);

		if (i->fence)
			drm_printf(p, " protected by 0x%016llx on context %llu",
				   (unsigned long long)i->fence->seqno,
				   (unsigned long long)i->fence->context);

		drm_puts(p, "\n");
	}
	spin_unlock(&sa_manager->wq.lock);
}
EXPORT_SYMBOL(drm_suballoc_dump_debug_info);
#endif
MODULE_AUTHOR("Multiple");
MODULE_DESCRIPTION("Range suballocator helper");
MODULE_LICENSE("Dual MIT/GPL");