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
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
|
/*
* Copyright © 2011-2012 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.
*
* Authors:
* Ben Widawsky <ben@bwidawsk.net>
*
*/
/*
* This file implements HW context support. On gen5+ a HW context consists of an
* opaque GPU object which is referenced at times of context saves and restores.
* With RC6 enabled, the context is also referenced as the GPU enters and exists
* from RC6 (GPU has it's own internal power context, except on gen5). Though
* something like a context does exist for the media ring, the code only
* supports contexts for the render ring.
*
* In software, there is a distinction between contexts created by the user,
* and the default HW context. The default HW context is used by GPU clients
* that do not request setup of their own hardware context. The default
* context's state is never restored to help prevent programming errors. This
* would happen if a client ran and piggy-backed off another clients GPU state.
* The default context only exists to give the GPU some offset to load as the
* current to invoke a save of the context we actually care about. In fact, the
* code could likely be constructed, albeit in a more complicated fashion, to
* never use the default context, though that limits the driver's ability to
* swap out, and/or destroy other contexts.
*
* All other contexts are created as a request by the GPU client. These contexts
* store GPU state, and thus allow GPU clients to not re-emit state (and
* potentially query certain state) at any time. The kernel driver makes
* certain that the appropriate commands are inserted.
*
* The context life cycle is semi-complicated in that context BOs may live
* longer than the context itself because of the way the hardware, and object
* tracking works. Below is a very crude representation of the state machine
* describing the context life.
* refcount pincount active
* S0: initial state 0 0 0
* S1: context created 1 0 0
* S2: context is currently running 2 1 X
* S3: GPU referenced, but not current 2 0 1
* S4: context is current, but destroyed 1 1 0
* S5: like S3, but destroyed 1 0 1
*
* The most common (but not all) transitions:
* S0->S1: client creates a context
* S1->S2: client submits execbuf with context
* S2->S3: other clients submits execbuf with context
* S3->S1: context object was retired
* S3->S2: clients submits another execbuf
* S2->S4: context destroy called with current context
* S3->S5->S0: destroy path
* S4->S5->S0: destroy path on current context
*
* There are two confusing terms used above:
* The "current context" means the context which is currently running on the
* GPU. The GPU has loaded its state already and has stored away the gtt
* offset of the BO. The GPU is not actively referencing the data at this
* offset, but it will on the next context switch. The only way to avoid this
* is to do a GPU reset.
*
* An "active context' is one which was previously the "current context" and is
* on the active list waiting for the next context switch to occur. Until this
* happens, the object must remain at the same gtt offset. It is therefore
* possible to destroy a context, but it is still active.
*
*/
#include <drm/drmP.h>
#include <drm/i915_drm.h>
#include "i915_drv.h"
#include "i915_trace.h"
#define ALL_L3_SLICES(dev) (1 << NUM_L3_SLICES(dev)) - 1
/* This is a HW constraint. The value below is the largest known requirement
* I've seen in a spec to date, and that was a workaround for a non-shipping
* part. It should be safe to decrease this, but it's more future proof as is.
*/
#define GEN6_CONTEXT_ALIGN (64<<10)
#define GEN7_CONTEXT_ALIGN 4096
static size_t get_context_alignment(struct drm_i915_private *dev_priv)
{
if (IS_GEN6(dev_priv))
return GEN6_CONTEXT_ALIGN;
return GEN7_CONTEXT_ALIGN;
}
static int get_context_size(struct drm_i915_private *dev_priv)
{
int ret;
u32 reg;
switch (INTEL_GEN(dev_priv)) {
case 6:
reg = I915_READ(CXT_SIZE);
ret = GEN6_CXT_TOTAL_SIZE(reg) * 64;
break;
case 7:
reg = I915_READ(GEN7_CXT_SIZE);
if (IS_HASWELL(dev_priv))
ret = HSW_CXT_TOTAL_SIZE;
else
ret = GEN7_CXT_TOTAL_SIZE(reg) * 64;
break;
case 8:
ret = GEN8_CXT_TOTAL_SIZE;
break;
default:
BUG();
}
return ret;
}
static void i915_gem_context_clean(struct intel_context *ctx)
{
struct i915_hw_ppgtt *ppgtt = ctx->ppgtt;
struct i915_vma *vma, *next;
if (!ppgtt)
return;
list_for_each_entry_safe(vma, next, &ppgtt->base.inactive_list,
vm_link) {
if (WARN_ON(__i915_vma_unbind_no_wait(vma)))
break;
}
}
void i915_gem_context_free(struct kref *ctx_ref)
{
struct intel_context *ctx = container_of(ctx_ref, typeof(*ctx), ref);
trace_i915_context_free(ctx);
if (i915.enable_execlists)
intel_lr_context_free(ctx);
/*
* This context is going away and we need to remove all VMAs still
* around. This is to handle imported shared objects for which
* destructor did not run when their handles were closed.
*/
i915_gem_context_clean(ctx);
i915_ppgtt_put(ctx->ppgtt);
if (ctx->legacy_hw_ctx.rcs_state)
drm_gem_object_unreference(&ctx->legacy_hw_ctx.rcs_state->base);
list_del(&ctx->link);
ida_simple_remove(&ctx->i915->context_hw_ida, ctx->hw_id);
kfree(ctx);
}
struct drm_i915_gem_object *
i915_gem_alloc_context_obj(struct drm_device *dev, size_t size)
{
struct drm_i915_gem_object *obj;
int ret;
obj = i915_gem_object_create(dev, size);
if (IS_ERR(obj))
return obj;
/*
* Try to make the context utilize L3 as well as LLC.
*
* On VLV we don't have L3 controls in the PTEs so we
* shouldn't touch the cache level, especially as that
* would make the object snooped which might have a
* negative performance impact.
*
* Snooping is required on non-llc platforms in execlist
* mode, but since all GGTT accesses use PAT entry 0 we
* get snooping anyway regardless of cache_level.
*
* This is only applicable for Ivy Bridge devices since
* later platforms don't have L3 control bits in the PTE.
*/
if (IS_IVYBRIDGE(dev)) {
ret = i915_gem_object_set_cache_level(obj, I915_CACHE_L3_LLC);
/* Failure shouldn't ever happen this early */
if (WARN_ON(ret)) {
drm_gem_object_unreference(&obj->base);
return ERR_PTR(ret);
}
}
return obj;
}
static int assign_hw_id(struct drm_i915_private *dev_priv, unsigned *out)
{
int ret;
ret = ida_simple_get(&dev_priv->context_hw_ida,
0, MAX_CONTEXT_HW_ID, GFP_KERNEL);
if (ret < 0) {
/* Contexts are only released when no longer active.
* Flush any pending retires to hopefully release some
* stale contexts and try again.
*/
i915_gem_retire_requests(dev_priv);
ret = ida_simple_get(&dev_priv->context_hw_ida,
0, MAX_CONTEXT_HW_ID, GFP_KERNEL);
if (ret < 0)
return ret;
}
*out = ret;
return 0;
}
static struct intel_context *
__create_hw_context(struct drm_device *dev,
struct drm_i915_file_private *file_priv)
{
struct drm_i915_private *dev_priv = dev->dev_private;
struct intel_context *ctx;
int ret;
ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
if (ctx == NULL)
return ERR_PTR(-ENOMEM);
ret = assign_hw_id(dev_priv, &ctx->hw_id);
if (ret) {
kfree(ctx);
return ERR_PTR(ret);
}
kref_init(&ctx->ref);
list_add_tail(&ctx->link, &dev_priv->context_list);
ctx->i915 = dev_priv;
if (dev_priv->hw_context_size) {
struct drm_i915_gem_object *obj =
i915_gem_alloc_context_obj(dev, dev_priv->hw_context_size);
if (IS_ERR(obj)) {
ret = PTR_ERR(obj);
goto err_out;
}
ctx->legacy_hw_ctx.rcs_state = obj;
}
/* Default context will never have a file_priv */
if (file_priv != NULL) {
ret = idr_alloc(&file_priv->context_idr, ctx,
DEFAULT_CONTEXT_HANDLE, 0, GFP_KERNEL);
if (ret < 0)
goto err_out;
} else
ret = DEFAULT_CONTEXT_HANDLE;
ctx->file_priv = file_priv;
ctx->user_handle = ret;
/* NB: Mark all slices as needing a remap so that when the context first
* loads it will restore whatever remap state already exists. If there
* is no remap info, it will be a NOP. */
ctx->remap_slice = ALL_L3_SLICES(dev_priv);
ctx->hang_stats.ban_period_seconds = DRM_I915_CTX_BAN_PERIOD;
return ctx;
err_out:
i915_gem_context_unreference(ctx);
return ERR_PTR(ret);
}
/**
* The default context needs to exist per ring that uses contexts. It stores the
* context state of the GPU for applications that don't utilize HW contexts, as
* well as an idle case.
*/
static struct intel_context *
i915_gem_create_context(struct drm_device *dev,
struct drm_i915_file_private *file_priv)
{
const bool is_global_default_ctx = file_priv == NULL;
struct intel_context *ctx;
int ret = 0;
BUG_ON(!mutex_is_locked(&dev->struct_mutex));
ctx = __create_hw_context(dev, file_priv);
if (IS_ERR(ctx))
return ctx;
if (is_global_default_ctx && ctx->legacy_hw_ctx.rcs_state) {
/* We may need to do things with the shrinker which
* require us to immediately switch back to the default
* context. This can cause a problem as pinning the
* default context also requires GTT space which may not
* be available. To avoid this we always pin the default
* context.
*/
ret = i915_gem_obj_ggtt_pin(ctx->legacy_hw_ctx.rcs_state,
get_context_alignment(to_i915(dev)), 0);
if (ret) {
DRM_DEBUG_DRIVER("Couldn't pin %d\n", ret);
goto err_destroy;
}
}
if (USES_FULL_PPGTT(dev)) {
struct i915_hw_ppgtt *ppgtt = i915_ppgtt_create(dev, file_priv);
if (IS_ERR_OR_NULL(ppgtt)) {
DRM_DEBUG_DRIVER("PPGTT setup failed (%ld)\n",
PTR_ERR(ppgtt));
ret = PTR_ERR(ppgtt);
goto err_unpin;
}
ctx->ppgtt = ppgtt;
}
trace_i915_context_create(ctx);
return ctx;
err_unpin:
if (is_global_default_ctx && ctx->legacy_hw_ctx.rcs_state)
i915_gem_object_ggtt_unpin(ctx->legacy_hw_ctx.rcs_state);
err_destroy:
idr_remove(&file_priv->context_idr, ctx->user_handle);
i915_gem_context_unreference(ctx);
return ERR_PTR(ret);
}
static void i915_gem_context_unpin(struct intel_context *ctx,
struct intel_engine_cs *engine)
{
if (i915.enable_execlists) {
intel_lr_context_unpin(ctx, engine);
} else {
if (engine->id == RCS && ctx->legacy_hw_ctx.rcs_state)
i915_gem_object_ggtt_unpin(ctx->legacy_hw_ctx.rcs_state);
i915_gem_context_unreference(ctx);
}
}
void i915_gem_context_reset(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
if (i915.enable_execlists) {
struct intel_context *ctx;
list_for_each_entry(ctx, &dev_priv->context_list, link)
intel_lr_context_reset(dev_priv, ctx);
}
i915_gem_context_lost(dev_priv);
}
int i915_gem_context_init(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
struct intel_context *ctx;
/* Init should only be called once per module load. Eventually the
* restriction on the context_disabled check can be loosened. */
if (WARN_ON(dev_priv->kernel_context))
return 0;
if (intel_vgpu_active(dev_priv) &&
HAS_LOGICAL_RING_CONTEXTS(dev_priv)) {
if (!i915.enable_execlists) {
DRM_INFO("Only EXECLIST mode is supported in vgpu.\n");
return -EINVAL;
}
}
/* Using the simple ida interface, the max is limited by sizeof(int) */
BUILD_BUG_ON(MAX_CONTEXT_HW_ID > INT_MAX);
ida_init(&dev_priv->context_hw_ida);
if (i915.enable_execlists) {
/* NB: intentionally left blank. We will allocate our own
* backing objects as we need them, thank you very much */
dev_priv->hw_context_size = 0;
} else if (HAS_HW_CONTEXTS(dev_priv)) {
dev_priv->hw_context_size =
round_up(get_context_size(dev_priv), 4096);
if (dev_priv->hw_context_size > (1<<20)) {
DRM_DEBUG_DRIVER("Disabling HW Contexts; invalid size %d\n",
dev_priv->hw_context_size);
dev_priv->hw_context_size = 0;
}
}
ctx = i915_gem_create_context(dev, NULL);
if (IS_ERR(ctx)) {
DRM_ERROR("Failed to create default global context (error %ld)\n",
PTR_ERR(ctx));
return PTR_ERR(ctx);
}
dev_priv->kernel_context = ctx;
DRM_DEBUG_DRIVER("%s context support initialized\n",
i915.enable_execlists ? "LR" :
dev_priv->hw_context_size ? "HW" : "fake");
return 0;
}
void i915_gem_context_lost(struct drm_i915_private *dev_priv)
{
struct intel_engine_cs *engine;
for_each_engine(engine, dev_priv) {
if (engine->last_context == NULL)
continue;
i915_gem_context_unpin(engine->last_context, engine);
engine->last_context = NULL;
}
/* Force the GPU state to be reinitialised on enabling */
dev_priv->kernel_context->legacy_hw_ctx.initialized = false;
dev_priv->kernel_context->remap_slice = ALL_L3_SLICES(dev_priv);
}
void i915_gem_context_fini(struct drm_device *dev)
{
struct drm_i915_private *dev_priv = dev->dev_private;
struct intel_context *dctx = dev_priv->kernel_context;
if (dctx->legacy_hw_ctx.rcs_state)
i915_gem_object_ggtt_unpin(dctx->legacy_hw_ctx.rcs_state);
i915_gem_context_unreference(dctx);
dev_priv->kernel_context = NULL;
ida_destroy(&dev_priv->context_hw_ida);
}
static int context_idr_cleanup(int id, void *p, void *data)
{
struct intel_context *ctx = p;
i915_gem_context_unreference(ctx);
return 0;
}
int i915_gem_context_open(struct drm_device *dev, struct drm_file *file)
{
struct drm_i915_file_private *file_priv = file->driver_priv;
struct intel_context *ctx;
idr_init(&file_priv->context_idr);
mutex_lock(&dev->struct_mutex);
ctx = i915_gem_create_context(dev, file_priv);
mutex_unlock(&dev->struct_mutex);
if (IS_ERR(ctx)) {
idr_destroy(&file_priv->context_idr);
return PTR_ERR(ctx);
}
return 0;
}
void i915_gem_context_close(struct drm_device *dev, struct drm_file *file)
{
struct drm_i915_file_private *file_priv = file->driver_priv;
idr_for_each(&file_priv->context_idr, context_idr_cleanup, NULL);
idr_destroy(&file_priv->context_idr);
}
struct intel_context *
i915_gem_context_get(struct drm_i915_file_private *file_priv, u32 id)
{
struct intel_context *ctx;
ctx = (struct intel_context *)idr_find(&file_priv->context_idr, id);
if (!ctx)
return ERR_PTR(-ENOENT);
return ctx;
}
static inline int
mi_set_context(struct drm_i915_gem_request *req, u32 hw_flags)
{
struct drm_i915_private *dev_priv = req->i915;
struct intel_engine_cs *engine = req->engine;
u32 flags = hw_flags | MI_MM_SPACE_GTT;
const int num_rings =
/* Use an extended w/a on ivb+ if signalling from other rings */
i915_semaphore_is_enabled(dev_priv) ?
hweight32(INTEL_INFO(dev_priv)->ring_mask) - 1 :
0;
int len, ret;
/* w/a: If Flush TLB Invalidation Mode is enabled, driver must do a TLB
* invalidation prior to MI_SET_CONTEXT. On GEN6 we don't set the value
* explicitly, so we rely on the value at ring init, stored in
* itlb_before_ctx_switch.
*/
if (IS_GEN6(dev_priv)) {
ret = engine->flush(req, I915_GEM_GPU_DOMAINS, 0);
if (ret)
return ret;
}
/* These flags are for resource streamer on HSW+ */
if (IS_HASWELL(dev_priv) || INTEL_GEN(dev_priv) >= 8)
flags |= (HSW_MI_RS_SAVE_STATE_EN | HSW_MI_RS_RESTORE_STATE_EN);
else if (INTEL_GEN(dev_priv) < 8)
flags |= (MI_SAVE_EXT_STATE_EN | MI_RESTORE_EXT_STATE_EN);
len = 4;
if (INTEL_GEN(dev_priv) >= 7)
len += 2 + (num_rings ? 4*num_rings + 6 : 0);
ret = intel_ring_begin(req, len);
if (ret)
return ret;
/* WaProgramMiArbOnOffAroundMiSetContext:ivb,vlv,hsw,bdw,chv */
if (INTEL_GEN(dev_priv) >= 7) {
intel_ring_emit(engine, MI_ARB_ON_OFF | MI_ARB_DISABLE);
if (num_rings) {
struct intel_engine_cs *signaller;
intel_ring_emit(engine,
MI_LOAD_REGISTER_IMM(num_rings));
for_each_engine(signaller, dev_priv) {
if (signaller == engine)
continue;
intel_ring_emit_reg(engine,
RING_PSMI_CTL(signaller->mmio_base));
intel_ring_emit(engine,
_MASKED_BIT_ENABLE(GEN6_PSMI_SLEEP_MSG_DISABLE));
}
}
}
intel_ring_emit(engine, MI_NOOP);
intel_ring_emit(engine, MI_SET_CONTEXT);
intel_ring_emit(engine,
i915_gem_obj_ggtt_offset(req->ctx->legacy_hw_ctx.rcs_state) |
flags);
/*
* w/a: MI_SET_CONTEXT must always be followed by MI_NOOP
* WaMiSetContext_Hang:snb,ivb,vlv
*/
intel_ring_emit(engine, MI_NOOP);
if (INTEL_GEN(dev_priv) >= 7) {
if (num_rings) {
struct intel_engine_cs *signaller;
i915_reg_t last_reg = {}; /* keep gcc quiet */
intel_ring_emit(engine,
MI_LOAD_REGISTER_IMM(num_rings));
for_each_engine(signaller, dev_priv) {
if (signaller == engine)
continue;
last_reg = RING_PSMI_CTL(signaller->mmio_base);
intel_ring_emit_reg(engine, last_reg);
intel_ring_emit(engine,
_MASKED_BIT_DISABLE(GEN6_PSMI_SLEEP_MSG_DISABLE));
}
/* Insert a delay before the next switch! */
intel_ring_emit(engine,
MI_STORE_REGISTER_MEM |
MI_SRM_LRM_GLOBAL_GTT);
intel_ring_emit_reg(engine, last_reg);
intel_ring_emit(engine, engine->scratch.gtt_offset);
intel_ring_emit(engine, MI_NOOP);
}
intel_ring_emit(engine, MI_ARB_ON_OFF | MI_ARB_ENABLE);
}
intel_ring_advance(engine);
return ret;
}
static int remap_l3(struct drm_i915_gem_request *req, int slice)
{
u32 *remap_info = req->i915->l3_parity.remap_info[slice];
struct intel_engine_cs *engine = req->engine;
int i, ret;
if (!remap_info)
return 0;
ret = intel_ring_begin(req, GEN7_L3LOG_SIZE/4 * 2 + 2);
if (ret)
return ret;
/*
* Note: We do not worry about the concurrent register cacheline hang
* here because no other code should access these registers other than
* at initialization time.
*/
intel_ring_emit(engine, MI_LOAD_REGISTER_IMM(GEN7_L3LOG_SIZE/4));
for (i = 0; i < GEN7_L3LOG_SIZE/4; i++) {
intel_ring_emit_reg(engine, GEN7_L3LOG(slice, i));
intel_ring_emit(engine, remap_info[i]);
}
intel_ring_emit(engine, MI_NOOP);
intel_ring_advance(engine);
return 0;
}
static inline bool skip_rcs_switch(struct i915_hw_ppgtt *ppgtt,
struct intel_engine_cs *engine,
struct intel_context *to)
{
if (to->remap_slice)
return false;
if (!to->legacy_hw_ctx.initialized)
return false;
if (ppgtt && (intel_engine_flag(engine) & ppgtt->pd_dirty_rings))
return false;
return to == engine->last_context;
}
static bool
needs_pd_load_pre(struct i915_hw_ppgtt *ppgtt,
struct intel_engine_cs *engine,
struct intel_context *to)
{
if (!ppgtt)
return false;
/* Always load the ppgtt on first use */
if (!engine->last_context)
return true;
/* Same context without new entries, skip */
if (engine->last_context == to &&
!(intel_engine_flag(engine) & ppgtt->pd_dirty_rings))
return false;
if (engine->id != RCS)
return true;
if (INTEL_GEN(engine->i915) < 8)
return true;
return false;
}
static bool
needs_pd_load_post(struct i915_hw_ppgtt *ppgtt,
struct intel_context *to,
u32 hw_flags)
{
if (!ppgtt)
return false;
if (!IS_GEN8(to->i915))
return false;
if (hw_flags & MI_RESTORE_INHIBIT)
return true;
return false;
}
static int do_rcs_switch(struct drm_i915_gem_request *req)
{
struct intel_context *to = req->ctx;
struct intel_engine_cs *engine = req->engine;
struct i915_hw_ppgtt *ppgtt = to->ppgtt ?: req->i915->mm.aliasing_ppgtt;
struct intel_context *from;
u32 hw_flags;
int ret, i;
if (skip_rcs_switch(ppgtt, engine, to))
return 0;
/* Trying to pin first makes error handling easier. */
ret = i915_gem_obj_ggtt_pin(to->legacy_hw_ctx.rcs_state,
get_context_alignment(engine->i915),
0);
if (ret)
return ret;
/*
* Pin can switch back to the default context if we end up calling into
* evict_everything - as a last ditch gtt defrag effort that also
* switches to the default context. Hence we need to reload from here.
*
* XXX: Doing so is painfully broken!
*/
from = engine->last_context;
/*
* Clear this page out of any CPU caches for coherent swap-in/out. Note
* that thanks to write = false in this call and us not setting any gpu
* write domains when putting a context object onto the active list
* (when switching away from it), this won't block.
*
* XXX: We need a real interface to do this instead of trickery.
*/
ret = i915_gem_object_set_to_gtt_domain(to->legacy_hw_ctx.rcs_state, false);
if (ret)
goto unpin_out;
if (needs_pd_load_pre(ppgtt, engine, to)) {
/* Older GENs and non render rings still want the load first,
* "PP_DCLV followed by PP_DIR_BASE register through Load
* Register Immediate commands in Ring Buffer before submitting
* a context."*/
trace_switch_mm(engine, to);
ret = ppgtt->switch_mm(ppgtt, req);
if (ret)
goto unpin_out;
}
if (!to->legacy_hw_ctx.initialized || i915_gem_context_is_default(to))
/* NB: If we inhibit the restore, the context is not allowed to
* die because future work may end up depending on valid address
* space. This means we must enforce that a page table load
* occur when this occurs. */
hw_flags = MI_RESTORE_INHIBIT;
else if (ppgtt && intel_engine_flag(engine) & ppgtt->pd_dirty_rings)
hw_flags = MI_FORCE_RESTORE;
else
hw_flags = 0;
if (to != from || (hw_flags & MI_FORCE_RESTORE)) {
ret = mi_set_context(req, hw_flags);
if (ret)
goto unpin_out;
}
/* The backing object for the context is done after switching to the
* *next* context. Therefore we cannot retire the previous context until
* the next context has already started running. In fact, the below code
* is a bit suboptimal because the retiring can occur simply after the
* MI_SET_CONTEXT instead of when the next seqno has completed.
*/
if (from != NULL) {
from->legacy_hw_ctx.rcs_state->base.read_domains = I915_GEM_DOMAIN_INSTRUCTION;
i915_vma_move_to_active(i915_gem_obj_to_ggtt(from->legacy_hw_ctx.rcs_state), req);
/* As long as MI_SET_CONTEXT is serializing, ie. it flushes the
* whole damn pipeline, we don't need to explicitly mark the
* object dirty. The only exception is that the context must be
* correct in case the object gets swapped out. Ideally we'd be
* able to defer doing this until we know the object would be
* swapped, but there is no way to do that yet.
*/
from->legacy_hw_ctx.rcs_state->dirty = 1;
/* obj is kept alive until the next request by its active ref */
i915_gem_object_ggtt_unpin(from->legacy_hw_ctx.rcs_state);
i915_gem_context_unreference(from);
}
i915_gem_context_reference(to);
engine->last_context = to;
/* GEN8 does *not* require an explicit reload if the PDPs have been
* setup, and we do not wish to move them.
*/
if (needs_pd_load_post(ppgtt, to, hw_flags)) {
trace_switch_mm(engine, to);
ret = ppgtt->switch_mm(ppgtt, req);
/* The hardware context switch is emitted, but we haven't
* actually changed the state - so it's probably safe to bail
* here. Still, let the user know something dangerous has
* happened.
*/
if (ret)
return ret;
}
if (ppgtt)
ppgtt->pd_dirty_rings &= ~intel_engine_flag(engine);
for (i = 0; i < MAX_L3_SLICES; i++) {
if (!(to->remap_slice & (1<<i)))
continue;
ret = remap_l3(req, i);
if (ret)
return ret;
to->remap_slice &= ~(1<<i);
}
if (!to->legacy_hw_ctx.initialized) {
if (engine->init_context) {
ret = engine->init_context(req);
if (ret)
return ret;
}
to->legacy_hw_ctx.initialized = true;
}
return 0;
unpin_out:
i915_gem_object_ggtt_unpin(to->legacy_hw_ctx.rcs_state);
return ret;
}
/**
* i915_switch_context() - perform a GPU context switch.
* @req: request for which we'll execute the context switch
*
* The context life cycle is simple. The context refcount is incremented and
* decremented by 1 and create and destroy. If the context is in use by the GPU,
* it will have a refcount > 1. This allows us to destroy the context abstract
* object while letting the normal object tracking destroy the backing BO.
*
* This function should not be used in execlists mode. Instead the context is
* switched by writing to the ELSP and requests keep a reference to their
* context.
*/
int i915_switch_context(struct drm_i915_gem_request *req)
{
struct intel_engine_cs *engine = req->engine;
struct drm_i915_private *dev_priv = req->i915;
WARN_ON(i915.enable_execlists);
WARN_ON(!mutex_is_locked(&dev_priv->dev->struct_mutex));
if (engine->id != RCS ||
req->ctx->legacy_hw_ctx.rcs_state == NULL) {
struct intel_context *to = req->ctx;
struct i915_hw_ppgtt *ppgtt =
to->ppgtt ?: req->i915->mm.aliasing_ppgtt;
if (needs_pd_load_pre(ppgtt, engine, to)) {
int ret;
trace_switch_mm(engine, to);
ret = ppgtt->switch_mm(ppgtt, req);
if (ret)
return ret;
ppgtt->pd_dirty_rings &= ~intel_engine_flag(engine);
}
if (to != engine->last_context) {
i915_gem_context_reference(to);
if (engine->last_context)
i915_gem_context_unreference(engine->last_context);
engine->last_context = to;
}
return 0;
}
return do_rcs_switch(req);
}
static bool contexts_enabled(struct drm_device *dev)
{
return i915.enable_execlists || to_i915(dev)->hw_context_size;
}
int i915_gem_context_create_ioctl(struct drm_device *dev, void *data,
struct drm_file *file)
{
struct drm_i915_gem_context_create *args = data;
struct drm_i915_file_private *file_priv = file->driver_priv;
struct intel_context *ctx;
int ret;
if (!contexts_enabled(dev))
return -ENODEV;
if (args->pad != 0)
return -EINVAL;
ret = i915_mutex_lock_interruptible(dev);
if (ret)
return ret;
ctx = i915_gem_create_context(dev, file_priv);
mutex_unlock(&dev->struct_mutex);
if (IS_ERR(ctx))
return PTR_ERR(ctx);
args->ctx_id = ctx->user_handle;
DRM_DEBUG_DRIVER("HW context %d created\n", args->ctx_id);
return 0;
}
int i915_gem_context_destroy_ioctl(struct drm_device *dev, void *data,
struct drm_file *file)
{
struct drm_i915_gem_context_destroy *args = data;
struct drm_i915_file_private *file_priv = file->driver_priv;
struct intel_context *ctx;
int ret;
if (args->pad != 0)
return -EINVAL;
if (args->ctx_id == DEFAULT_CONTEXT_HANDLE)
return -ENOENT;
ret = i915_mutex_lock_interruptible(dev);
if (ret)
return ret;
ctx = i915_gem_context_get(file_priv, args->ctx_id);
if (IS_ERR(ctx)) {
mutex_unlock(&dev->struct_mutex);
return PTR_ERR(ctx);
}
idr_remove(&ctx->file_priv->context_idr, ctx->user_handle);
i915_gem_context_unreference(ctx);
mutex_unlock(&dev->struct_mutex);
DRM_DEBUG_DRIVER("HW context %d destroyed\n", args->ctx_id);
return 0;
}
int i915_gem_context_getparam_ioctl(struct drm_device *dev, void *data,
struct drm_file *file)
{
struct drm_i915_file_private *file_priv = file->driver_priv;
struct drm_i915_gem_context_param *args = data;
struct intel_context *ctx;
int ret;
ret = i915_mutex_lock_interruptible(dev);
if (ret)
return ret;
ctx = i915_gem_context_get(file_priv, args->ctx_id);
if (IS_ERR(ctx)) {
mutex_unlock(&dev->struct_mutex);
return PTR_ERR(ctx);
}
args->size = 0;
switch (args->param) {
case I915_CONTEXT_PARAM_BAN_PERIOD:
args->value = ctx->hang_stats.ban_period_seconds;
break;
case I915_CONTEXT_PARAM_NO_ZEROMAP:
args->value = ctx->flags & CONTEXT_NO_ZEROMAP;
break;
case I915_CONTEXT_PARAM_GTT_SIZE:
if (ctx->ppgtt)
args->value = ctx->ppgtt->base.total;
else if (to_i915(dev)->mm.aliasing_ppgtt)
args->value = to_i915(dev)->mm.aliasing_ppgtt->base.total;
else
args->value = to_i915(dev)->ggtt.base.total;
break;
default:
ret = -EINVAL;
break;
}
mutex_unlock(&dev->struct_mutex);
return ret;
}
int i915_gem_context_setparam_ioctl(struct drm_device *dev, void *data,
struct drm_file *file)
{
struct drm_i915_file_private *file_priv = file->driver_priv;
struct drm_i915_gem_context_param *args = data;
struct intel_context *ctx;
int ret;
ret = i915_mutex_lock_interruptible(dev);
if (ret)
return ret;
ctx = i915_gem_context_get(file_priv, args->ctx_id);
if (IS_ERR(ctx)) {
mutex_unlock(&dev->struct_mutex);
return PTR_ERR(ctx);
}
switch (args->param) {
case I915_CONTEXT_PARAM_BAN_PERIOD:
if (args->size)
ret = -EINVAL;
else if (args->value < ctx->hang_stats.ban_period_seconds &&
!capable(CAP_SYS_ADMIN))
ret = -EPERM;
else
ctx->hang_stats.ban_period_seconds = args->value;
break;
case I915_CONTEXT_PARAM_NO_ZEROMAP:
if (args->size) {
ret = -EINVAL;
} else {
ctx->flags &= ~CONTEXT_NO_ZEROMAP;
ctx->flags |= args->value ? CONTEXT_NO_ZEROMAP : 0;
}
break;
default:
ret = -EINVAL;
break;
}
mutex_unlock(&dev->struct_mutex);
return ret;
}
|