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#include "drmP.h"
28#include "radeon_drm.h"
29#include "radeon_reg.h"
30#include "radeon.h"
31
32void r100_cs_dump_packet(struct radeon_cs_parser *p,
33 struct radeon_cs_packet *pkt);
34
35int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
36{
37 struct drm_device *ddev = p->rdev->ddev;
38 struct radeon_cs_chunk *chunk;
39 unsigned i, j;
40 bool duplicate;
41
42 if (p->chunk_relocs_idx == -1) {
43 return 0;
44 }
45 chunk = &p->chunks[p->chunk_relocs_idx];
46
47 p->nrelocs = chunk->length_dw / 4;
48 p->relocs_ptr = kcalloc(p->nrelocs, sizeof(void *), GFP_KERNEL);
49 if (p->relocs_ptr == NULL) {
50 return -ENOMEM;
51 }
52 p->relocs = kcalloc(p->nrelocs, sizeof(struct radeon_cs_reloc), GFP_KERNEL);
53 if (p->relocs == NULL) {
54 return -ENOMEM;
55 }
56 for (i = 0; i < p->nrelocs; i++) {
57 struct drm_radeon_cs_reloc *r;
58
59 duplicate = false;
60 r = (struct drm_radeon_cs_reloc *)&chunk->kdata[i*4];
61 for (j = 0; j < i; j++) {
62 if (r->handle == p->relocs[j].handle) {
63 p->relocs_ptr[i] = &p->relocs[j];
64 duplicate = true;
65 break;
66 }
67 }
68 if (!duplicate) {
69 p->relocs[i].gobj = drm_gem_object_lookup(ddev,
70 p->filp,
71 r->handle);
72 if (p->relocs[i].gobj == NULL) {
73 DRM_ERROR("gem object lookup failed 0x%x\n",
74 r->handle);
75 return -ENOENT;
76 }
77 p->relocs_ptr[i] = &p->relocs[i];
78 p->relocs[i].robj = gem_to_radeon_bo(p->relocs[i].gobj);
79 p->relocs[i].lobj.bo = p->relocs[i].robj;
80 p->relocs[i].lobj.wdomain = r->write_domain;
81 p->relocs[i].lobj.rdomain = r->read_domains;
82 p->relocs[i].lobj.tv.bo = &p->relocs[i].robj->tbo;
83 p->relocs[i].handle = r->handle;
84 p->relocs[i].flags = r->flags;
85 radeon_bo_list_add_object(&p->relocs[i].lobj,
86 &p->validated);
87
88 } else
89 p->relocs[i].handle = 0;
90 }
91 return radeon_bo_list_validate(&p->validated);
92}
93
94static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority)
95{
96 p->priority = priority;
97
98 switch (ring) {
99 default:
100 DRM_ERROR("unknown ring id: %d\n", ring);
101 return -EINVAL;
102 case RADEON_CS_RING_GFX:
103 p->ring = RADEON_RING_TYPE_GFX_INDEX;
104 break;
105 case RADEON_CS_RING_COMPUTE:
106 if (p->rdev->family >= CHIP_TAHITI) {
107 if (p->priority > 0)
108 p->ring = CAYMAN_RING_TYPE_CP1_INDEX;
109 else
110 p->ring = CAYMAN_RING_TYPE_CP2_INDEX;
111 } else
112 p->ring = RADEON_RING_TYPE_GFX_INDEX;
113 break;
114 }
115 return 0;
116}
117
118static void radeon_cs_sync_rings(struct radeon_cs_parser *p)
119{
120 int i;
121
122 for (i = 0; i < p->nrelocs; i++) {
123 struct radeon_fence *a, *b;
124
125 if (!p->relocs[i].robj || !p->relocs[i].robj->tbo.sync_obj)
126 continue;
127
128 a = p->relocs[i].robj->tbo.sync_obj;
129 b = p->ib.sync_to[a->ring];
130 p->ib.sync_to[a->ring] = radeon_fence_later(a, b);
131 }
132}
133
134
135int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
136{
137 struct drm_radeon_cs *cs = data;
138 uint64_t *chunk_array_ptr;
139 unsigned size, i;
140 u32 ring = RADEON_CS_RING_GFX;
141 s32 priority = 0;
142
143 if (!cs->num_chunks) {
144 return 0;
145 }
146
147 INIT_LIST_HEAD(&p->validated);
148 p->idx = 0;
149 p->ib.sa_bo = NULL;
150 p->ib.semaphore = NULL;
151 p->const_ib.sa_bo = NULL;
152 p->const_ib.semaphore = NULL;
153 p->chunk_ib_idx = -1;
154 p->chunk_relocs_idx = -1;
155 p->chunk_flags_idx = -1;
156 p->chunk_const_ib_idx = -1;
157 p->chunks_array = kcalloc(cs->num_chunks, sizeof(uint64_t), GFP_KERNEL);
158 if (p->chunks_array == NULL) {
159 return -ENOMEM;
160 }
161 chunk_array_ptr = (uint64_t *)(unsigned long)(cs->chunks);
162 if (DRM_COPY_FROM_USER(p->chunks_array, chunk_array_ptr,
163 sizeof(uint64_t)*cs->num_chunks)) {
164 return -EFAULT;
165 }
166 p->cs_flags = 0;
167 p->nchunks = cs->num_chunks;
168 p->chunks = kcalloc(p->nchunks, sizeof(struct radeon_cs_chunk), GFP_KERNEL);
169 if (p->chunks == NULL) {
170 return -ENOMEM;
171 }
172 for (i = 0; i < p->nchunks; i++) {
173 struct drm_radeon_cs_chunk __user **chunk_ptr = NULL;
174 struct drm_radeon_cs_chunk user_chunk;
175 uint32_t __user *cdata;
176
177 chunk_ptr = (void __user*)(unsigned long)p->chunks_array[i];
178 if (DRM_COPY_FROM_USER(&user_chunk, chunk_ptr,
179 sizeof(struct drm_radeon_cs_chunk))) {
180 return -EFAULT;
181 }
182 p->chunks[i].length_dw = user_chunk.length_dw;
183 p->chunks[i].kdata = NULL;
184 p->chunks[i].chunk_id = user_chunk.chunk_id;
185
186 if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_RELOCS) {
187 p->chunk_relocs_idx = i;
188 }
189 if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_IB) {
190 p->chunk_ib_idx = i;
191
192 if (p->chunks[i].length_dw == 0)
193 return -EINVAL;
194 }
195 if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_CONST_IB) {
196 p->chunk_const_ib_idx = i;
197
198 if (p->chunks[i].length_dw == 0)
199 return -EINVAL;
200 }
201 if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS) {
202 p->chunk_flags_idx = i;
203
204 if (p->chunks[i].length_dw == 0)
205 return -EINVAL;
206 }
207
208 p->chunks[i].length_dw = user_chunk.length_dw;
209 p->chunks[i].user_ptr = (void __user *)(unsigned long)user_chunk.chunk_data;
210
211 cdata = (uint32_t *)(unsigned long)user_chunk.chunk_data;
212 if ((p->chunks[i].chunk_id == RADEON_CHUNK_ID_RELOCS) ||
213 (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS)) {
214 size = p->chunks[i].length_dw * sizeof(uint32_t);
215 p->chunks[i].kdata = kmalloc(size, GFP_KERNEL);
216 if (p->chunks[i].kdata == NULL) {
217 return -ENOMEM;
218 }
219 if (DRM_COPY_FROM_USER(p->chunks[i].kdata,
220 p->chunks[i].user_ptr, size)) {
221 return -EFAULT;
222 }
223 if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS) {
224 p->cs_flags = p->chunks[i].kdata[0];
225 if (p->chunks[i].length_dw > 1)
226 ring = p->chunks[i].kdata[1];
227 if (p->chunks[i].length_dw > 2)
228 priority = (s32)p->chunks[i].kdata[2];
229 }
230 }
231 }
232
233
234 if (p->rdev) {
235 if ((p->cs_flags & RADEON_CS_USE_VM) &&
236 !p->rdev->vm_manager.enabled) {
237 DRM_ERROR("VM not active on asic!\n");
238 return -EINVAL;
239 }
240
241
242 if ((p->rdev->family >= CHIP_TAHITI) &&
243 ((p->cs_flags & RADEON_CS_USE_VM) == 0)) {
244 DRM_ERROR("VM required on SI+!\n");
245 return -EINVAL;
246 }
247
248 if (radeon_cs_get_ring(p, ring, priority))
249 return -EINVAL;
250 }
251
252
253 if ((p->chunk_ib_idx != -1) &&
254 ((p->cs_flags & RADEON_CS_USE_VM) == 0) &&
255 (p->chunks[p->chunk_ib_idx].chunk_id == RADEON_CHUNK_ID_IB)) {
256 if (p->chunks[p->chunk_ib_idx].length_dw > (16 * 1024)) {
257 DRM_ERROR("cs IB too big: %d\n",
258 p->chunks[p->chunk_ib_idx].length_dw);
259 return -EINVAL;
260 }
261 if ((p->rdev->flags & RADEON_IS_AGP)) {
262 p->chunks[p->chunk_ib_idx].kpage[0] = kmalloc(PAGE_SIZE, GFP_KERNEL);
263 p->chunks[p->chunk_ib_idx].kpage[1] = kmalloc(PAGE_SIZE, GFP_KERNEL);
264 if (p->chunks[p->chunk_ib_idx].kpage[0] == NULL ||
265 p->chunks[p->chunk_ib_idx].kpage[1] == NULL) {
266 kfree(p->chunks[i].kpage[0]);
267 kfree(p->chunks[i].kpage[1]);
268 return -ENOMEM;
269 }
270 }
271 p->chunks[p->chunk_ib_idx].kpage_idx[0] = -1;
272 p->chunks[p->chunk_ib_idx].kpage_idx[1] = -1;
273 p->chunks[p->chunk_ib_idx].last_copied_page = -1;
274 p->chunks[p->chunk_ib_idx].last_page_index =
275 ((p->chunks[p->chunk_ib_idx].length_dw * 4) - 1) / PAGE_SIZE;
276 }
277
278 return 0;
279}
280
281static void radeon_bo_vm_fence_va(struct radeon_cs_parser *parser,
282 struct radeon_fence *fence)
283{
284 struct radeon_fpriv *fpriv = parser->filp->driver_priv;
285 struct radeon_vm *vm = &fpriv->vm;
286 struct radeon_bo_list *lobj;
287
288 if (parser->chunk_ib_idx == -1) {
289 return;
290 }
291 if ((parser->cs_flags & RADEON_CS_USE_VM) == 0) {
292 return;
293 }
294
295 list_for_each_entry(lobj, &parser->validated, tv.head) {
296 struct radeon_bo_va *bo_va;
297 struct radeon_bo *rbo = lobj->bo;
298
299 bo_va = radeon_bo_va(rbo, vm);
300 radeon_fence_unref(&bo_va->fence);
301 bo_va->fence = radeon_fence_ref(fence);
302 }
303}
304
305
306
307
308
309
310
311
312
313static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error)
314{
315 unsigned i;
316
317 if (!error) {
318
319 radeon_bo_vm_fence_va(parser, parser->ib.fence);
320 ttm_eu_fence_buffer_objects(&parser->validated,
321 parser->ib.fence);
322 } else {
323 ttm_eu_backoff_reservation(&parser->validated);
324 }
325
326 if (parser->relocs != NULL) {
327 for (i = 0; i < parser->nrelocs; i++) {
328 if (parser->relocs[i].gobj)
329 drm_gem_object_unreference_unlocked(parser->relocs[i].gobj);
330 }
331 }
332 kfree(parser->track);
333 kfree(parser->relocs);
334 kfree(parser->relocs_ptr);
335 for (i = 0; i < parser->nchunks; i++) {
336 kfree(parser->chunks[i].kdata);
337 if ((parser->rdev->flags & RADEON_IS_AGP)) {
338 kfree(parser->chunks[i].kpage[0]);
339 kfree(parser->chunks[i].kpage[1]);
340 }
341 }
342 kfree(parser->chunks);
343 kfree(parser->chunks_array);
344 radeon_ib_free(parser->rdev, &parser->ib);
345 radeon_ib_free(parser->rdev, &parser->const_ib);
346}
347
348static int radeon_cs_ib_chunk(struct radeon_device *rdev,
349 struct radeon_cs_parser *parser)
350{
351 struct radeon_cs_chunk *ib_chunk;
352 int r;
353
354 if (parser->chunk_ib_idx == -1)
355 return 0;
356
357 if (parser->cs_flags & RADEON_CS_USE_VM)
358 return 0;
359
360 ib_chunk = &parser->chunks[parser->chunk_ib_idx];
361
362
363
364
365 r = radeon_ib_get(rdev, parser->ring, &parser->ib,
366 ib_chunk->length_dw * 4);
367 if (r) {
368 DRM_ERROR("Failed to get ib !\n");
369 return r;
370 }
371 parser->ib.length_dw = ib_chunk->length_dw;
372 r = radeon_cs_parse(rdev, parser->ring, parser);
373 if (r || parser->parser_error) {
374 DRM_ERROR("Invalid command stream !\n");
375 return r;
376 }
377 r = radeon_cs_finish_pages(parser);
378 if (r) {
379 DRM_ERROR("Invalid command stream !\n");
380 return r;
381 }
382 radeon_cs_sync_rings(parser);
383 parser->ib.vm_id = 0;
384 r = radeon_ib_schedule(rdev, &parser->ib, NULL);
385 if (r) {
386 DRM_ERROR("Failed to schedule IB !\n");
387 }
388 return r;
389}
390
391static int radeon_bo_vm_update_pte(struct radeon_cs_parser *parser,
392 struct radeon_vm *vm)
393{
394 struct radeon_bo_list *lobj;
395 struct radeon_bo *bo;
396 int r;
397
398 list_for_each_entry(lobj, &parser->validated, tv.head) {
399 bo = lobj->bo;
400 r = radeon_vm_bo_update_pte(parser->rdev, vm, bo, &bo->tbo.mem);
401 if (r) {
402 return r;
403 }
404 }
405 return 0;
406}
407
408static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev,
409 struct radeon_cs_parser *parser)
410{
411 struct radeon_cs_chunk *ib_chunk;
412 struct radeon_fpriv *fpriv = parser->filp->driver_priv;
413 struct radeon_vm *vm = &fpriv->vm;
414 int r;
415
416 if (parser->chunk_ib_idx == -1)
417 return 0;
418 if ((parser->cs_flags & RADEON_CS_USE_VM) == 0)
419 return 0;
420
421 if ((rdev->family >= CHIP_TAHITI) &&
422 (parser->chunk_const_ib_idx != -1)) {
423 ib_chunk = &parser->chunks[parser->chunk_const_ib_idx];
424 if (ib_chunk->length_dw > RADEON_IB_VM_MAX_SIZE) {
425 DRM_ERROR("cs IB CONST too big: %d\n", ib_chunk->length_dw);
426 return -EINVAL;
427 }
428 r = radeon_ib_get(rdev, parser->ring, &parser->const_ib,
429 ib_chunk->length_dw * 4);
430 if (r) {
431 DRM_ERROR("Failed to get const ib !\n");
432 return r;
433 }
434 parser->const_ib.is_const_ib = true;
435 parser->const_ib.length_dw = ib_chunk->length_dw;
436
437 if (DRM_COPY_FROM_USER(parser->const_ib.ptr, ib_chunk->user_ptr,
438 ib_chunk->length_dw * 4)) {
439 return -EFAULT;
440 }
441 r = radeon_ring_ib_parse(rdev, parser->ring, &parser->const_ib);
442 if (r) {
443 return r;
444 }
445 }
446
447 ib_chunk = &parser->chunks[parser->chunk_ib_idx];
448 if (ib_chunk->length_dw > RADEON_IB_VM_MAX_SIZE) {
449 DRM_ERROR("cs IB too big: %d\n", ib_chunk->length_dw);
450 return -EINVAL;
451 }
452 r = radeon_ib_get(rdev, parser->ring, &parser->ib,
453 ib_chunk->length_dw * 4);
454 if (r) {
455 DRM_ERROR("Failed to get ib !\n");
456 return r;
457 }
458 parser->ib.length_dw = ib_chunk->length_dw;
459
460 if (DRM_COPY_FROM_USER(parser->ib.ptr, ib_chunk->user_ptr,
461 ib_chunk->length_dw * 4)) {
462 return -EFAULT;
463 }
464 r = radeon_ring_ib_parse(rdev, parser->ring, &parser->ib);
465 if (r) {
466 return r;
467 }
468
469 mutex_lock(&rdev->vm_manager.lock);
470 mutex_lock(&vm->mutex);
471 r = radeon_vm_bind(rdev, vm);
472 if (r) {
473 goto out;
474 }
475 r = radeon_bo_vm_update_pte(parser, vm);
476 if (r) {
477 goto out;
478 }
479 radeon_cs_sync_rings(parser);
480
481 parser->ib.vm_id = vm->id;
482
483
484
485 parser->ib.gpu_addr = parser->ib.sa_bo->soffset;
486
487 if ((rdev->family >= CHIP_TAHITI) &&
488 (parser->chunk_const_ib_idx != -1)) {
489 parser->const_ib.vm_id = vm->id;
490
491
492
493 parser->const_ib.gpu_addr = parser->const_ib.sa_bo->soffset;
494 r = radeon_ib_schedule(rdev, &parser->ib, &parser->const_ib);
495 } else {
496 r = radeon_ib_schedule(rdev, &parser->ib, NULL);
497 }
498
499out:
500 if (!r) {
501 if (vm->fence) {
502 radeon_fence_unref(&vm->fence);
503 }
504 vm->fence = radeon_fence_ref(parser->ib.fence);
505 }
506 mutex_unlock(&vm->mutex);
507 mutex_unlock(&rdev->vm_manager.lock);
508 return r;
509}
510
511static int radeon_cs_handle_lockup(struct radeon_device *rdev, int r)
512{
513 if (r == -EDEADLK) {
514 r = radeon_gpu_reset(rdev);
515 if (!r)
516 r = -EAGAIN;
517 }
518 return r;
519}
520
521int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
522{
523 struct radeon_device *rdev = dev->dev_private;
524 struct radeon_cs_parser parser;
525 int r;
526
527 down_read(&rdev->exclusive_lock);
528 if (!rdev->accel_working) {
529 up_read(&rdev->exclusive_lock);
530 return -EBUSY;
531 }
532
533 memset(&parser, 0, sizeof(struct radeon_cs_parser));
534 parser.filp = filp;
535 parser.rdev = rdev;
536 parser.dev = rdev->dev;
537 parser.family = rdev->family;
538 r = radeon_cs_parser_init(&parser, data);
539 if (r) {
540 DRM_ERROR("Failed to initialize parser !\n");
541 radeon_cs_parser_fini(&parser, r);
542 up_read(&rdev->exclusive_lock);
543 r = radeon_cs_handle_lockup(rdev, r);
544 return r;
545 }
546 r = radeon_cs_parser_relocs(&parser);
547 if (r) {
548 if (r != -ERESTARTSYS)
549 DRM_ERROR("Failed to parse relocation %d!\n", r);
550 radeon_cs_parser_fini(&parser, r);
551 up_read(&rdev->exclusive_lock);
552 r = radeon_cs_handle_lockup(rdev, r);
553 return r;
554 }
555 r = radeon_cs_ib_chunk(rdev, &parser);
556 if (r) {
557 goto out;
558 }
559 r = radeon_cs_ib_vm_chunk(rdev, &parser);
560 if (r) {
561 goto out;
562 }
563out:
564 radeon_cs_parser_fini(&parser, r);
565 up_read(&rdev->exclusive_lock);
566 r = radeon_cs_handle_lockup(rdev, r);
567 return r;
568}
569
570int radeon_cs_finish_pages(struct radeon_cs_parser *p)
571{
572 struct radeon_cs_chunk *ibc = &p->chunks[p->chunk_ib_idx];
573 int i;
574 int size = PAGE_SIZE;
575
576 for (i = ibc->last_copied_page + 1; i <= ibc->last_page_index; i++) {
577 if (i == ibc->last_page_index) {
578 size = (ibc->length_dw * 4) % PAGE_SIZE;
579 if (size == 0)
580 size = PAGE_SIZE;
581 }
582
583 if (DRM_COPY_FROM_USER(p->ib.ptr + (i * (PAGE_SIZE/4)),
584 ibc->user_ptr + (i * PAGE_SIZE),
585 size))
586 return -EFAULT;
587 }
588 return 0;
589}
590
591static int radeon_cs_update_pages(struct radeon_cs_parser *p, int pg_idx)
592{
593 int new_page;
594 struct radeon_cs_chunk *ibc = &p->chunks[p->chunk_ib_idx];
595 int i;
596 int size = PAGE_SIZE;
597 bool copy1 = (p->rdev->flags & RADEON_IS_AGP) ? false : true;
598
599 for (i = ibc->last_copied_page + 1; i < pg_idx; i++) {
600 if (DRM_COPY_FROM_USER(p->ib.ptr + (i * (PAGE_SIZE/4)),
601 ibc->user_ptr + (i * PAGE_SIZE),
602 PAGE_SIZE)) {
603 p->parser_error = -EFAULT;
604 return 0;
605 }
606 }
607
608 if (pg_idx == ibc->last_page_index) {
609 size = (ibc->length_dw * 4) % PAGE_SIZE;
610 if (size == 0)
611 size = PAGE_SIZE;
612 }
613
614 new_page = ibc->kpage_idx[0] < ibc->kpage_idx[1] ? 0 : 1;
615 if (copy1)
616 ibc->kpage[new_page] = p->ib.ptr + (pg_idx * (PAGE_SIZE / 4));
617
618 if (DRM_COPY_FROM_USER(ibc->kpage[new_page],
619 ibc->user_ptr + (pg_idx * PAGE_SIZE),
620 size)) {
621 p->parser_error = -EFAULT;
622 return 0;
623 }
624
625
626 if (!copy1)
627 memcpy((void *)(p->ib.ptr+(pg_idx*(PAGE_SIZE/4))), ibc->kpage[new_page], size);
628
629 ibc->last_copied_page = pg_idx;
630 ibc->kpage_idx[new_page] = pg_idx;
631
632 return new_page;
633}
634
635u32 radeon_get_ib_value(struct radeon_cs_parser *p, int idx)
636{
637 struct radeon_cs_chunk *ibc = &p->chunks[p->chunk_ib_idx];
638 u32 pg_idx, pg_offset;
639 u32 idx_value = 0;
640 int new_page;
641
642 pg_idx = (idx * 4) / PAGE_SIZE;
643 pg_offset = (idx * 4) % PAGE_SIZE;
644
645 if (ibc->kpage_idx[0] == pg_idx)
646 return ibc->kpage[0][pg_offset/4];
647 if (ibc->kpage_idx[1] == pg_idx)
648 return ibc->kpage[1][pg_offset/4];
649
650 new_page = radeon_cs_update_pages(p, pg_idx);
651 if (new_page < 0) {
652 p->parser_error = new_page;
653 return 0;
654 }
655
656 idx_value = ibc->kpage[new_page][pg_offset/4];
657 return idx_value;
658}
659
