summaryrefslogtreecommitdiffstats
path: root/drivers/gpu/drm/radeon/radeon_object.c
blob: 983e8df5e000d392c26b0b711044055137acd761 (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
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
/*
 * Copyright 2009 Jerome Glisse.
 * 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.
 *
 */
/*
 * Authors:
 *    Jerome Glisse <glisse@freedesktop.org>
 *    Thomas Hellstrom <thomas-at-tungstengraphics-dot-com>
 *    Dave Airlie
 */
#include <linux/list.h>
#include <drm/drmP.h>
#include "radeon_drm.h"
#include "radeon.h"

struct radeon_object {
	struct ttm_buffer_object	tobj;
	struct list_head		list;
	struct radeon_device		*rdev;
	struct drm_gem_object		*gobj;
	struct ttm_bo_kmap_obj		kmap;
	unsigned			pin_count;
	uint64_t			gpu_addr;
	void				*kptr;
	bool				is_iomem;
};

int radeon_ttm_init(struct radeon_device *rdev);
void radeon_ttm_fini(struct radeon_device *rdev);

/*
 * To exclude mutual BO access we rely on bo_reserve exclusion, as all
 * function are calling it.
 */

static int radeon_object_reserve(struct radeon_object *robj, bool interruptible)
{
	return ttm_bo_reserve(&robj->tobj, interruptible, false, false, 0);
}

static void radeon_object_unreserve(struct radeon_object *robj)
{
	ttm_bo_unreserve(&robj->tobj);
}

static void radeon_ttm_object_object_destroy(struct ttm_buffer_object *tobj)
{
	struct radeon_object *robj;

	robj = container_of(tobj, struct radeon_object, tobj);
	list_del_init(&robj->list);
	kfree(robj);
}

static inline void radeon_object_gpu_addr(struct radeon_object *robj)
{
	/* Default gpu address */
	robj->gpu_addr = 0xFFFFFFFFFFFFFFFFULL;
	if (robj->tobj.mem.mm_node == NULL) {
		return;
	}
	robj->gpu_addr = ((u64)robj->tobj.mem.mm_node->start) << PAGE_SHIFT;
	switch (robj->tobj.mem.mem_type) {
	case TTM_PL_VRAM:
		robj->gpu_addr += (u64)robj->rdev->mc.vram_location;
		break;
	case TTM_PL_TT:
		robj->gpu_addr += (u64)robj->rdev->mc.gtt_location;
		break;
	default:
		DRM_ERROR("Unknown placement %d\n", robj->tobj.mem.mem_type);
		robj->gpu_addr = 0xFFFFFFFFFFFFFFFFULL;
		return;
	}
}

static inline uint32_t radeon_object_flags_from_domain(uint32_t domain)
{
	uint32_t flags = 0;
	if (domain & RADEON_GEM_DOMAIN_VRAM) {
		flags |= TTM_PL_FLAG_VRAM;
	}
	if (domain & RADEON_GEM_DOMAIN_GTT) {
		flags |= TTM_PL_FLAG_TT;
	}
	if (domain & RADEON_GEM_DOMAIN_CPU) {
		flags |= TTM_PL_FLAG_SYSTEM;
	}
	if (!flags) {
		flags |= TTM_PL_FLAG_SYSTEM;
	}
	return flags;
}

int radeon_object_create(struct radeon_device *rdev,
			 struct drm_gem_object *gobj,
			 unsigned long size,
			 bool kernel,
			 uint32_t domain,
			 bool interruptible,
			 struct radeon_object **robj_ptr)
{
	struct radeon_object *robj;
	enum ttm_bo_type type;
	uint32_t flags;
	int r;

	if (unlikely(rdev->mman.bdev.dev_mapping == NULL)) {
		rdev->mman.bdev.dev_mapping = rdev->ddev->dev_mapping;
	}
	if (kernel) {
		type = ttm_bo_type_kernel;
	} else {
		type = ttm_bo_type_device;
	}
	*robj_ptr = NULL;
	robj = kzalloc(sizeof(struct radeon_object), GFP_KERNEL);
	if (robj == NULL) {
		return -ENOMEM;
	}
	robj->rdev = rdev;
	robj->gobj = gobj;
	INIT_LIST_HEAD(&robj->list);

	flags = radeon_object_flags_from_domain(domain);
	r = ttm_buffer_object_init(&rdev->mman.bdev, &robj->tobj, size, type, flags,
				   0, 0, false, NULL, size,
				   &radeon_ttm_object_object_destroy);
	if (unlikely(r != 0)) {
		/* ttm call radeon_ttm_object_object_destroy if error happen */
		DRM_ERROR("Failed to allocate TTM object (%ld, 0x%08X, %u)\n",
			  size, flags, 0);
		return r;
	}
	*robj_ptr = robj;
	if (gobj) {
		list_add_tail(&robj->list, &rdev->gem.objects);
	}
	return 0;
}

int radeon_object_kmap(struct radeon_object *robj, void **ptr)
{
	int r;

	spin_lock(&robj->tobj.lock);
	if (robj->kptr) {
		if (ptr) {
			*ptr = robj->kptr;
		}
		spin_unlock(&robj->tobj.lock);
		return 0;
	}
	spin_unlock(&robj->tobj.lock);
	r = ttm_bo_kmap(&robj->tobj, 0, robj->tobj.num_pages, &robj->kmap);
	if (r) {
		return r;
	}
	spin_lock(&robj->tobj.lock);
	robj->kptr = ttm_kmap_obj_virtual(&robj->kmap, &robj->is_iomem);
	spin_unlock(&robj->tobj.lock);
	if (ptr) {
		*ptr = robj->kptr;
	}
	return 0;
}

void radeon_object_kunmap(struct radeon_object *robj)
{
	spin_lock(&robj->tobj.lock);
	if (robj->kptr == NULL) {
		spin_unlock(&robj->tobj.lock);
		return;
	}
	robj->kptr = NULL;
	spin_unlock(&robj->tobj.lock);
	ttm_bo_kunmap(&robj->kmap);
}

void radeon_object_unref(struct radeon_object **robj)
{
	struct ttm_buffer_object *tobj;

	if ((*robj) == NULL) {
		return;
	}
	tobj = &((*robj)->tobj);
	ttm_bo_unref(&tobj);
	if (tobj == NULL) {
		*robj = NULL;
	}
}

int radeon_object_mmap(struct radeon_object *robj, uint64_t *offset)
{
	*offset = robj->tobj.addr_space_offset;
	return 0;
}

int radeon_object_pin(struct radeon_object *robj, uint32_t domain,
		      uint64_t *gpu_addr)
{
	uint32_t flags;
	uint32_t tmp;
	void *fbptr;
	int r;

	flags = radeon_object_flags_from_domain(domain);
	spin_lock(&robj->tobj.lock);
	if (robj->pin_count) {
		robj->pin_count++;
		if (gpu_addr != NULL) {
			*gpu_addr = robj->gpu_addr;
		}
		spin_unlock(&robj->tobj.lock);
		return 0;
	}
	spin_unlock(&robj->tobj.lock);
	r = radeon_object_reserve(robj, false);
	if (unlikely(r != 0)) {
		DRM_ERROR("radeon: failed to reserve object for pinning it.\n");
		return r;
	}
	if (robj->rdev->fbdev_robj == robj) {
		mutex_lock(&robj->rdev->fbdev_info->lock);
		radeon_object_kunmap(robj);
	}
	tmp = robj->tobj.mem.placement;
	ttm_flag_masked(&tmp, flags, TTM_PL_MASK_MEM);
	robj->tobj.proposed_placement = tmp | TTM_PL_FLAG_NO_EVICT | TTM_PL_MASK_CACHING;
	r = ttm_buffer_object_validate(&robj->tobj,
				       robj->tobj.proposed_placement,
				       false, false);
	radeon_object_gpu_addr(robj);
	if (gpu_addr != NULL) {
		*gpu_addr = robj->gpu_addr;
	}
	robj->pin_count = 1;
	if (unlikely(r != 0)) {
		DRM_ERROR("radeon: failed to pin object.\n");
	}
	radeon_object_unreserve(robj);
	if (robj->rdev->fbdev_robj == robj) {
		if (!r) {
			r = radeon_object_kmap(robj, &fbptr);
		}
		if (!r) {
			robj->rdev->fbdev_info->screen_base = fbptr;
			robj->rdev->fbdev_info->fix.smem_start = (unsigned long)fbptr;
		}
		mutex_unlock(&robj->rdev->fbdev_info->lock);
	}
	return r;
}

void radeon_object_unpin(struct radeon_object *robj)
{
	uint32_t flags;
	void *fbptr;
	int r;

	spin_lock(&robj->tobj.lock);
	if (!robj->pin_count) {
		spin_unlock(&robj->tobj.lock);
		printk(KERN_WARNING "Unpin not necessary for %p !\n", robj);
		return;
	}
	robj->pin_count--;
	if (robj->pin_count) {
		spin_unlock(&robj->tobj.lock);
		return;
	}
	spin_unlock(&robj->tobj.lock);
	r = radeon_object_reserve(robj, false);
	if (unlikely(r != 0)) {
		DRM_ERROR("radeon: failed to reserve object for unpinning it.\n");
		return;
	}
	if (robj->rdev->fbdev_robj == robj) {
		mutex_lock(&robj->rdev->fbdev_info->lock);
		radeon_object_kunmap(robj);
	}
	flags = robj->tobj.mem.placement;
	robj->tobj.proposed_placement = flags & ~TTM_PL_FLAG_NO_EVICT;
	r = ttm_buffer_object_validate(&robj->tobj,
				       robj->tobj.proposed_placement,
				       false, false);
	if (unlikely(r != 0)) {
		DRM_ERROR("radeon: failed to unpin buffer.\n");
	}
	radeon_object_unreserve(robj);
	if (robj->rdev->fbdev_robj == robj) {
		if (!r) {
			r = radeon_object_kmap(robj, &fbptr);
		}
		if (!r) {
			robj->rdev->fbdev_info->screen_base = fbptr;
			robj->rdev->fbdev_info->fix.smem_start = (unsigned long)fbptr;
		}
		mutex_unlock(&robj->rdev->fbdev_info->lock);
	}
}

int radeon_object_wait(struct radeon_object *robj)
{
	int r = 0;

	/* FIXME: should use block reservation instead */
	r = radeon_object_reserve(robj, true);
	if (unlikely(r != 0)) {
		DRM_ERROR("radeon: failed to reserve object for waiting.\n");
		return r;
	}
	spin_lock(&robj->tobj.lock);
	if (robj->tobj.sync_obj) {
		r = ttm_bo_wait(&robj->tobj, true, false, false);
	}
	spin_unlock(&robj->tobj.lock);
	radeon_object_unreserve(robj);
	return r;
}

int radeon_object_evict_vram(struct radeon_device *rdev)
{
	if (rdev->flags & RADEON_IS_IGP) {
		/* Useless to evict on IGP chips */
		return 0;
	}
	return ttm_bo_evict_mm(&rdev->mman.bdev, TTM_PL_VRAM);
}

void radeon_object_force_delete(struct radeon_device *rdev)
{
	struct radeon_object *robj, *n;
	struct drm_gem_object *gobj;

	if (list_empty(&rdev->gem.objects)) {
		return;
	}
	DRM_ERROR("Userspace still has active objects !\n");
	list_for_each_entry_safe(robj, n, &rdev->gem.objects, list) {
		mutex_lock(&rdev->ddev->struct_mutex);
		gobj = robj->gobj;
		DRM_ERROR("Force free for (%p,%p,%lu,%lu)\n",
			  gobj, robj, (unsigned long)gobj->size,
			  *((unsigned long *)&gobj->refcount));
		list_del_init(&robj->list);
		radeon_object_unref(&robj);
		gobj->driver_private = NULL;
		drm_gem_object_unreference(gobj);
		mutex_unlock(&rdev->ddev->struct_mutex);
	}
}

int radeon_object_init(struct radeon_device *rdev)
{
	return radeon_ttm_init(rdev);
}

void radeon_object_fini(struct radeon_device *rdev)
{
	radeon_ttm_fini(rdev);
}

void radeon_object_list_add_object(struct radeon_object_list *lobj,
				   struct list_head *head)
{
	if (lobj->wdomain) {
		list_add(&lobj->list, head);
	} else {
		list_add_tail(&lobj->list, head);
	}
}

int radeon_object_list_reserve(struct list_head *head)
{
	struct radeon_object_list *lobj;
	struct list_head *i;
	int r;

	list_for_each(i, head) {
		lobj = list_entry(i, struct radeon_object_list, list);
		if (!lobj->robj->pin_count) {
			r = radeon_object_reserve(lobj->robj, true);
			if (unlikely(r != 0)) {
				DRM_ERROR("radeon: failed to reserve object.\n");
				return r;
			}
		} else {
		}
	}
	return 0;
}

void radeon_object_list_unreserve(struct list_head *head)
{
	struct radeon_object_list *lobj;
	struct list_head *i;

	list_for_each(i, head) {
		lobj = list_entry(i, struct radeon_object_list, list);
		if (!lobj->robj->pin_count) {
			radeon_object_unreserve(lobj->robj);
		} else {
		}
	}
}

int radeon_object_list_validate(struct list_head *head, void *fence)
{
	struct radeon_object_list *lobj;
	struct radeon_object *robj;
	struct radeon_fence *old_fence = NULL;
	struct list_head *i;
	uint32_t flags;
	int r;

	r = radeon_object_list_reserve(head);
	if (unlikely(r != 0)) {
		radeon_object_list_unreserve(head);
		return r;
	}
	list_for_each(i, head) {
		lobj = list_entry(i, struct radeon_object_list, list);
		robj = lobj->robj;
		if (lobj->wdomain) {
			flags = radeon_object_flags_from_domain(lobj->wdomain);
			flags |= TTM_PL_FLAG_TT;
		} else {
			flags = radeon_object_flags_from_domain(lobj->rdomain);
			flags |= TTM_PL_FLAG_TT;
			flags |= TTM_PL_FLAG_VRAM;
		}
		if (!robj->pin_count) {
			robj->tobj.proposed_placement = flags | TTM_PL_MASK_CACHING;
			r = ttm_buffer_object_validate(&robj->tobj,
						       robj->tobj.proposed_placement,
						       true, false);
			if (unlikely(r)) {
				radeon_object_list_unreserve(head);
				DRM_ERROR("radeon: failed to validate.\n");
				return r;
			}
			radeon_object_gpu_addr(robj);
		}
		lobj->gpu_offset = robj->gpu_addr;
		if (fence) {
			old_fence = (struct radeon_fence *)robj->tobj.sync_obj;
			robj->tobj.sync_obj = radeon_fence_ref(fence);
			robj->tobj.sync_obj_arg = NULL;
		}
		if (old_fence) {
			radeon_fence_unref(&old_fence);
		}
	}
	return 0;
}

void radeon_object_list_unvalidate(struct list_head *head)
{
	struct radeon_object_list *lobj;
	struct radeon_fence *old_fence = NULL;
	struct list_head *i;

	list_for_each(i, head) {
		lobj = list_entry(i, struct radeon_object_list, list);
		old_fence = (struct radeon_fence *)lobj->robj->tobj.sync_obj;
		lobj->robj->tobj.sync_obj = NULL;
		if (old_fence) {
			radeon_fence_unref(&old_fence);
		}
	}
	radeon_object_list_unreserve(head);
}

void radeon_object_list_clean(struct list_head *head)
{
	radeon_object_list_unreserve(head);
}

int radeon_object_fbdev_mmap(struct radeon_object *robj,
			     struct vm_area_struct *vma)
{
	return ttm_fbdev_mmap(vma, &robj->tobj);
}

unsigned long radeon_object_size(struct radeon_object *robj)
{
	return robj->tobj.num_pages << PAGE_SHIFT;
}