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
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
|
/*
* Copyright © 2015 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.
*/
#ifndef ANV_PRIVATE_H
#define ANV_PRIVATE_H
#include <stdlib.h>
#include <stdio.h>
#include <stdbool.h>
#include <pthread.h>
#include <assert.h>
#include <stdint.h>
#include <i915_drm.h>
#ifdef HAVE_VALGRIND
#include <valgrind.h>
#include <memcheck.h>
#define VG(x) x
#define __gen_validate_value(x) VALGRIND_CHECK_MEM_IS_DEFINED(&(x), sizeof(x))
#else
#define VG(x)
#endif
#include "common/gen_device_info.h"
#include "blorp/blorp.h"
#include "brw_compiler.h"
#include "util/macros.h"
#include "util/list.h"
/* Pre-declarations needed for WSI entrypoints */
struct wl_surface;
struct wl_display;
typedef struct xcb_connection_t xcb_connection_t;
typedef uint32_t xcb_visualid_t;
typedef uint32_t xcb_window_t;
struct gen_l3_config;
#include <vulkan/vulkan.h>
#include <vulkan/vulkan_intel.h>
#include <vulkan/vk_icd.h>
#include "anv_entrypoints.h"
#include "brw_context.h"
#include "isl/isl.h"
#ifdef __cplusplus
extern "C" {
#endif
#define MAX_VBS 32
#define MAX_SETS 8
#define MAX_RTS 8
#define MAX_VIEWPORTS 16
#define MAX_SCISSORS 16
#define MAX_PUSH_CONSTANTS_SIZE 128
#define MAX_DYNAMIC_BUFFERS 16
#define MAX_IMAGES 8
#define MAX_SAMPLES_LOG2 4 /* SKL supports 16 samples */
#define anv_noreturn __attribute__((__noreturn__))
#define anv_printflike(a, b) __attribute__((__format__(__printf__, a, b)))
#define MIN(a, b) ((a) < (b) ? (a) : (b))
#define MAX(a, b) ((a) > (b) ? (a) : (b))
static inline uint32_t
align_down_npot_u32(uint32_t v, uint32_t a)
{
return v - (v % a);
}
static inline uint32_t
align_u32(uint32_t v, uint32_t a)
{
assert(a != 0 && a == (a & -a));
return (v + a - 1) & ~(a - 1);
}
static inline uint64_t
align_u64(uint64_t v, uint64_t a)
{
assert(a != 0 && a == (a & -a));
return (v + a - 1) & ~(a - 1);
}
static inline int32_t
align_i32(int32_t v, int32_t a)
{
assert(a != 0 && a == (a & -a));
return (v + a - 1) & ~(a - 1);
}
/** Alignment must be a power of 2. */
static inline bool
anv_is_aligned(uintmax_t n, uintmax_t a)
{
assert(a == (a & -a));
return (n & (a - 1)) == 0;
}
static inline uint32_t
anv_minify(uint32_t n, uint32_t levels)
{
if (unlikely(n == 0))
return 0;
else
return MAX(n >> levels, 1);
}
static inline float
anv_clamp_f(float f, float min, float max)
{
assert(min < max);
if (f > max)
return max;
else if (f < min)
return min;
else
return f;
}
static inline bool
anv_clear_mask(uint32_t *inout_mask, uint32_t clear_mask)
{
if (*inout_mask & clear_mask) {
*inout_mask &= ~clear_mask;
return true;
} else {
return false;
}
}
#define for_each_bit(b, dword) \
for (uint32_t __dword = (dword); \
(b) = __builtin_ffs(__dword) - 1, __dword; \
__dword &= ~(1 << (b)))
#define typed_memcpy(dest, src, count) ({ \
static_assert(sizeof(*src) == sizeof(*dest), ""); \
memcpy((dest), (src), (count) * sizeof(*(src))); \
})
#define zero(x) (memset(&(x), 0, sizeof(x)))
/* Define no kernel as 1, since that's an illegal offset for a kernel */
#define NO_KERNEL 1
struct anv_common {
VkStructureType sType;
const void* pNext;
};
/* Whenever we generate an error, pass it through this function. Useful for
* debugging, where we can break on it. Only call at error site, not when
* propagating errors. Might be useful to plug in a stack trace here.
*/
VkResult __vk_errorf(VkResult error, const char *file, int line, const char *format, ...);
#ifdef DEBUG
#define vk_error(error) __vk_errorf(error, __FILE__, __LINE__, NULL);
#define vk_errorf(error, format, ...) __vk_errorf(error, __FILE__, __LINE__, format, ## __VA_ARGS__);
#else
#define vk_error(error) error
#define vk_errorf(error, format, ...) error
#endif
void __anv_finishme(const char *file, int line, const char *format, ...)
anv_printflike(3, 4);
void anv_loge(const char *format, ...) anv_printflike(1, 2);
void anv_loge_v(const char *format, va_list va);
/**
* Print a FINISHME message, including its source location.
*/
#define anv_finishme(format, ...) \
__anv_finishme(__FILE__, __LINE__, format, ##__VA_ARGS__);
/* A non-fatal assert. Useful for debugging. */
#ifdef DEBUG
#define anv_assert(x) ({ \
if (unlikely(!(x))) \
fprintf(stderr, "%s:%d ASSERT: %s\n", __FILE__, __LINE__, #x); \
})
#else
#define anv_assert(x)
#endif
/**
* If a block of code is annotated with anv_validate, then the block runs only
* in debug builds.
*/
#ifdef DEBUG
#define anv_validate if (1)
#else
#define anv_validate if (0)
#endif
void anv_abortf(const char *format, ...) anv_noreturn anv_printflike(1, 2);
void anv_abortfv(const char *format, va_list va) anv_noreturn;
#define stub_return(v) \
do { \
anv_finishme("stub %s", __func__); \
return (v); \
} while (0)
#define stub() \
do { \
anv_finishme("stub %s", __func__); \
return; \
} while (0)
/**
* A dynamically growable, circular buffer. Elements are added at head and
* removed from tail. head and tail are free-running uint32_t indices and we
* only compute the modulo with size when accessing the array. This way,
* number of bytes in the queue is always head - tail, even in case of
* wraparound.
*/
struct anv_vector {
uint32_t head;
uint32_t tail;
uint32_t element_size;
uint32_t size;
void *data;
};
int anv_vector_init(struct anv_vector *queue, uint32_t element_size, uint32_t size);
void *anv_vector_add(struct anv_vector *queue);
void *anv_vector_remove(struct anv_vector *queue);
static inline int
anv_vector_length(struct anv_vector *queue)
{
return (queue->head - queue->tail) / queue->element_size;
}
static inline void *
anv_vector_head(struct anv_vector *vector)
{
assert(vector->tail < vector->head);
return (void *)((char *)vector->data +
((vector->head - vector->element_size) &
(vector->size - 1)));
}
static inline void *
anv_vector_tail(struct anv_vector *vector)
{
return (void *)((char *)vector->data + (vector->tail & (vector->size - 1)));
}
static inline void
anv_vector_finish(struct anv_vector *queue)
{
free(queue->data);
}
#define anv_vector_foreach(elem, queue) \
static_assert(__builtin_types_compatible_p(__typeof__(queue), struct anv_vector *), ""); \
for (uint32_t __anv_vector_offset = (queue)->tail; \
elem = (queue)->data + (__anv_vector_offset & ((queue)->size - 1)), __anv_vector_offset < (queue)->head; \
__anv_vector_offset += (queue)->element_size)
struct anv_bo {
uint32_t gem_handle;
/* Index into the current validation list. This is used by the
* validation list building alrogithm to track which buffers are already
* in the validation list so that we can ensure uniqueness.
*/
uint32_t index;
/* Last known offset. This value is provided by the kernel when we
* execbuf and is used as the presumed offset for the next bunch of
* relocations.
*/
uint64_t offset;
uint64_t size;
void *map;
/* We need to set the WRITE flag on winsys bos so GEM will know we're
* writing to them and synchronize uses on other rings (eg if the display
* server uses the blitter ring).
*/
bool is_winsys_bo;
};
/* Represents a lock-free linked list of "free" things. This is used by
* both the block pool and the state pools. Unfortunately, in order to
* solve the ABA problem, we can't use a single uint32_t head.
*/
union anv_free_list {
struct {
int32_t offset;
/* A simple count that is incremented every time the head changes. */
uint32_t count;
};
uint64_t u64;
};
#define ANV_FREE_LIST_EMPTY ((union anv_free_list) { { 1, 0 } })
struct anv_block_state {
union {
struct {
uint32_t next;
uint32_t end;
};
uint64_t u64;
};
};
struct anv_block_pool {
struct anv_device *device;
struct anv_bo bo;
/* The offset from the start of the bo to the "center" of the block
* pool. Pointers to allocated blocks are given by
* bo.map + center_bo_offset + offsets.
*/
uint32_t center_bo_offset;
/* Current memory map of the block pool. This pointer may or may not
* point to the actual beginning of the block pool memory. If
* anv_block_pool_alloc_back has ever been called, then this pointer
* will point to the "center" position of the buffer and all offsets
* (negative or positive) given out by the block pool alloc functions
* will be valid relative to this pointer.
*
* In particular, map == bo.map + center_offset
*/
void *map;
int fd;
/**
* Array of mmaps and gem handles owned by the block pool, reclaimed when
* the block pool is destroyed.
*/
struct anv_vector mmap_cleanups;
uint32_t block_size;
union anv_free_list free_list;
struct anv_block_state state;
union anv_free_list back_free_list;
struct anv_block_state back_state;
};
/* Block pools are backed by a fixed-size 2GB memfd */
#define BLOCK_POOL_MEMFD_SIZE (1ull << 32)
/* The center of the block pool is also the middle of the memfd. This may
* change in the future if we decide differently for some reason.
*/
#define BLOCK_POOL_MEMFD_CENTER (BLOCK_POOL_MEMFD_SIZE / 2)
static inline uint32_t
anv_block_pool_size(struct anv_block_pool *pool)
{
return pool->state.end + pool->back_state.end;
}
struct anv_state {
int32_t offset;
uint32_t alloc_size;
void *map;
};
struct anv_fixed_size_state_pool {
size_t state_size;
union anv_free_list free_list;
struct anv_block_state block;
};
#define ANV_MIN_STATE_SIZE_LOG2 6
#define ANV_MAX_STATE_SIZE_LOG2 17
#define ANV_STATE_BUCKETS (ANV_MAX_STATE_SIZE_LOG2 - ANV_MIN_STATE_SIZE_LOG2 + 1)
struct anv_state_pool {
struct anv_block_pool *block_pool;
struct anv_fixed_size_state_pool buckets[ANV_STATE_BUCKETS];
};
struct anv_state_stream_block;
struct anv_state_stream {
struct anv_block_pool *block_pool;
/* The current working block */
struct anv_state_stream_block *block;
/* Offset at which the current block starts */
uint32_t start;
/* Offset at which to allocate the next state */
uint32_t next;
/* Offset at which the current block ends */
uint32_t end;
};
#define CACHELINE_SIZE 64
#define CACHELINE_MASK 63
static inline void
anv_clflush_range(void *start, size_t size)
{
void *p = (void *) (((uintptr_t) start) & ~CACHELINE_MASK);
void *end = start + size;
__builtin_ia32_mfence();
while (p < end) {
__builtin_ia32_clflush(p);
p += CACHELINE_SIZE;
}
}
static void inline
anv_state_clflush(struct anv_state state)
{
anv_clflush_range(state.map, state.alloc_size);
}
void anv_block_pool_init(struct anv_block_pool *pool,
struct anv_device *device, uint32_t block_size);
void anv_block_pool_finish(struct anv_block_pool *pool);
int32_t anv_block_pool_alloc(struct anv_block_pool *pool);
int32_t anv_block_pool_alloc_back(struct anv_block_pool *pool);
void anv_block_pool_free(struct anv_block_pool *pool, int32_t offset);
void anv_state_pool_init(struct anv_state_pool *pool,
struct anv_block_pool *block_pool);
void anv_state_pool_finish(struct anv_state_pool *pool);
struct anv_state anv_state_pool_alloc(struct anv_state_pool *pool,
size_t state_size, size_t alignment);
void anv_state_pool_free(struct anv_state_pool *pool, struct anv_state state);
void anv_state_stream_init(struct anv_state_stream *stream,
struct anv_block_pool *block_pool);
void anv_state_stream_finish(struct anv_state_stream *stream);
struct anv_state anv_state_stream_alloc(struct anv_state_stream *stream,
uint32_t size, uint32_t alignment);
/**
* Implements a pool of re-usable BOs. The interface is identical to that
* of block_pool except that each block is its own BO.
*/
struct anv_bo_pool {
struct anv_device *device;
void *free_list[16];
};
void anv_bo_pool_init(struct anv_bo_pool *pool, struct anv_device *device);
void anv_bo_pool_finish(struct anv_bo_pool *pool);
VkResult anv_bo_pool_alloc(struct anv_bo_pool *pool, struct anv_bo *bo,
uint32_t size);
void anv_bo_pool_free(struct anv_bo_pool *pool, const struct anv_bo *bo);
struct anv_scratch_pool {
/* Indexed by Per-Thread Scratch Space number (the hardware value) and stage */
struct anv_bo bos[16][MESA_SHADER_STAGES];
};
void anv_scratch_pool_init(struct anv_device *device,
struct anv_scratch_pool *pool);
void anv_scratch_pool_finish(struct anv_device *device,
struct anv_scratch_pool *pool);
struct anv_bo *anv_scratch_pool_alloc(struct anv_device *device,
struct anv_scratch_pool *pool,
gl_shader_stage stage,
unsigned per_thread_scratch);
extern struct anv_dispatch_table dtable;
static inline void *
anv_alloc(const VkAllocationCallbacks *alloc,
size_t size, size_t align,
VkSystemAllocationScope scope)
{
return alloc->pfnAllocation(alloc->pUserData, size, align, scope);
}
static inline void *
anv_realloc(const VkAllocationCallbacks *alloc,
void *ptr, size_t size, size_t align,
VkSystemAllocationScope scope)
{
return alloc->pfnReallocation(alloc->pUserData, ptr, size, align, scope);
}
static inline void
anv_free(const VkAllocationCallbacks *alloc, void *data)
{
alloc->pfnFree(alloc->pUserData, data);
}
static inline void *
anv_alloc2(const VkAllocationCallbacks *parent_alloc,
const VkAllocationCallbacks *alloc,
size_t size, size_t align,
VkSystemAllocationScope scope)
{
if (alloc)
return anv_alloc(alloc, size, align, scope);
else
return anv_alloc(parent_alloc, size, align, scope);
}
static inline void
anv_free2(const VkAllocationCallbacks *parent_alloc,
const VkAllocationCallbacks *alloc,
void *data)
{
if (alloc)
anv_free(alloc, data);
else
anv_free(parent_alloc, data);
}
struct anv_wsi_interaface;
#define VK_ICD_WSI_PLATFORM_MAX 5
struct anv_physical_device {
VK_LOADER_DATA _loader_data;
struct anv_instance * instance;
uint32_t chipset_id;
char path[20];
const char * name;
struct gen_device_info info;
uint64_t aperture_size;
struct brw_compiler * compiler;
struct isl_device isl_dev;
int cmd_parser_version;
uint32_t eu_total;
uint32_t subslice_total;
struct anv_wsi_interface * wsi[VK_ICD_WSI_PLATFORM_MAX];
};
struct anv_instance {
VK_LOADER_DATA _loader_data;
VkAllocationCallbacks alloc;
uint32_t apiVersion;
int physicalDeviceCount;
struct anv_physical_device physicalDevice;
};
VkResult anv_init_wsi(struct anv_physical_device *physical_device);
void anv_finish_wsi(struct anv_physical_device *physical_device);
struct anv_queue {
VK_LOADER_DATA _loader_data;
struct anv_device * device;
struct anv_state_pool * pool;
};
struct anv_pipeline_cache {
struct anv_device * device;
pthread_mutex_t mutex;
struct hash_table * cache;
};
struct anv_pipeline_bind_map;
void anv_pipeline_cache_init(struct anv_pipeline_cache *cache,
struct anv_device *device,
bool cache_enabled);
void anv_pipeline_cache_finish(struct anv_pipeline_cache *cache);
struct anv_shader_bin *
anv_pipeline_cache_search(struct anv_pipeline_cache *cache,
const void *key, uint32_t key_size);
struct anv_shader_bin *
anv_pipeline_cache_upload_kernel(struct anv_pipeline_cache *cache,
const void *key_data, uint32_t key_size,
const void *kernel_data, uint32_t kernel_size,
const void *prog_data, uint32_t prog_data_size,
const struct anv_pipeline_bind_map *bind_map);
struct anv_device {
VK_LOADER_DATA _loader_data;
VkAllocationCallbacks alloc;
struct anv_instance * instance;
uint32_t chipset_id;
struct gen_device_info info;
struct isl_device isl_dev;
int context_id;
int fd;
bool can_chain_batches;
bool robust_buffer_access;
struct anv_bo_pool batch_bo_pool;
struct anv_block_pool dynamic_state_block_pool;
struct anv_state_pool dynamic_state_pool;
struct anv_block_pool instruction_block_pool;
struct anv_state_pool instruction_state_pool;
struct anv_block_pool surface_state_block_pool;
struct anv_state_pool surface_state_pool;
struct anv_bo workaround_bo;
struct anv_pipeline_cache blorp_shader_cache;
struct blorp_context blorp;
struct anv_state border_colors;
struct anv_queue queue;
struct anv_scratch_pool scratch_pool;
uint32_t default_mocs;
pthread_mutex_t mutex;
};
void anv_device_get_cache_uuid(void *uuid);
void anv_device_init_blorp(struct anv_device *device);
void anv_device_finish_blorp(struct anv_device *device);
void* anv_gem_mmap(struct anv_device *device,
uint32_t gem_handle, uint64_t offset, uint64_t size, uint32_t flags);
void anv_gem_munmap(void *p, uint64_t size);
uint32_t anv_gem_create(struct anv_device *device, size_t size);
void anv_gem_close(struct anv_device *device, uint32_t gem_handle);
uint32_t anv_gem_userptr(struct anv_device *device, void *mem, size_t size);
int anv_gem_wait(struct anv_device *device, uint32_t gem_handle, int64_t *timeout_ns);
int anv_gem_execbuffer(struct anv_device *device,
struct drm_i915_gem_execbuffer2 *execbuf);
int anv_gem_set_tiling(struct anv_device *device, uint32_t gem_handle,
uint32_t stride, uint32_t tiling);
int anv_gem_create_context(struct anv_device *device);
int anv_gem_destroy_context(struct anv_device *device, int context);
int anv_gem_get_param(int fd, uint32_t param);
bool anv_gem_get_bit6_swizzle(int fd, uint32_t tiling);
int anv_gem_get_aperture(int fd, uint64_t *size);
int anv_gem_handle_to_fd(struct anv_device *device, uint32_t gem_handle);
uint32_t anv_gem_fd_to_handle(struct anv_device *device, int fd);
int anv_gem_set_caching(struct anv_device *device, uint32_t gem_handle, uint32_t caching);
int anv_gem_set_domain(struct anv_device *device, uint32_t gem_handle,
uint32_t read_domains, uint32_t write_domain);
VkResult anv_bo_init_new(struct anv_bo *bo, struct anv_device *device, uint64_t size);
struct anv_reloc_list {
size_t num_relocs;
size_t array_length;
struct drm_i915_gem_relocation_entry * relocs;
struct anv_bo ** reloc_bos;
};
VkResult anv_reloc_list_init(struct anv_reloc_list *list,
const VkAllocationCallbacks *alloc);
void anv_reloc_list_finish(struct anv_reloc_list *list,
const VkAllocationCallbacks *alloc);
uint64_t anv_reloc_list_add(struct anv_reloc_list *list,
const VkAllocationCallbacks *alloc,
uint32_t offset, struct anv_bo *target_bo,
uint32_t delta);
struct anv_batch_bo {
/* Link in the anv_cmd_buffer.owned_batch_bos list */
struct list_head link;
struct anv_bo bo;
/* Bytes actually consumed in this batch BO */
size_t length;
/* Last seen surface state block pool bo offset */
uint32_t last_ss_pool_bo_offset;
struct anv_reloc_list relocs;
};
struct anv_batch {
const VkAllocationCallbacks * alloc;
void * start;
void * end;
void * next;
struct anv_reloc_list * relocs;
/* This callback is called (with the associated user data) in the event
* that the batch runs out of space.
*/
VkResult (*extend_cb)(struct anv_batch *, void *);
void * user_data;
};
void *anv_batch_emit_dwords(struct anv_batch *batch, int num_dwords);
void anv_batch_emit_batch(struct anv_batch *batch, struct anv_batch *other);
uint64_t anv_batch_emit_reloc(struct anv_batch *batch,
void *location, struct anv_bo *bo, uint32_t offset);
VkResult anv_device_submit_simple_batch(struct anv_device *device,
struct anv_batch *batch);
struct anv_address {
struct anv_bo *bo;
uint32_t offset;
};
static inline uint64_t
_anv_combine_address(struct anv_batch *batch, void *location,
const struct anv_address address, uint32_t delta)
{
if (address.bo == NULL) {
return address.offset + delta;
} else {
assert(batch->start <= location && location < batch->end);
return anv_batch_emit_reloc(batch, location, address.bo, address.offset + delta);
}
}
#define __gen_address_type struct anv_address
#define __gen_user_data struct anv_batch
#define __gen_combine_address _anv_combine_address
/* Wrapper macros needed to work around preprocessor argument issues. In
* particular, arguments don't get pre-evaluated if they are concatenated.
* This means that, if you pass GENX(3DSTATE_PS) into the emit macro, the
* GENX macro won't get evaluated if the emit macro contains "cmd ## foo".
* We can work around this easily enough with these helpers.
*/
#define __anv_cmd_length(cmd) cmd ## _length
#define __anv_cmd_length_bias(cmd) cmd ## _length_bias
#define __anv_cmd_header(cmd) cmd ## _header
#define __anv_cmd_pack(cmd) cmd ## _pack
#define __anv_reg_num(reg) reg ## _num
#define anv_pack_struct(dst, struc, ...) do { \
struct struc __template = { \
__VA_ARGS__ \
}; \
__anv_cmd_pack(struc)(NULL, dst, &__template); \
VG(VALGRIND_CHECK_MEM_IS_DEFINED(dst, __anv_cmd_length(struc) * 4)); \
} while (0)
#define anv_batch_emitn(batch, n, cmd, ...) ({ \
void *__dst = anv_batch_emit_dwords(batch, n); \
struct cmd __template = { \
__anv_cmd_header(cmd), \
.DWordLength = n - __anv_cmd_length_bias(cmd), \
__VA_ARGS__ \
}; \
__anv_cmd_pack(cmd)(batch, __dst, &__template); \
__dst; \
})
#define anv_batch_emit_merge(batch, dwords0, dwords1) \
do { \
uint32_t *dw; \
\
static_assert(ARRAY_SIZE(dwords0) == ARRAY_SIZE(dwords1), "mismatch merge"); \
dw = anv_batch_emit_dwords((batch), ARRAY_SIZE(dwords0)); \
for (uint32_t i = 0; i < ARRAY_SIZE(dwords0); i++) \
dw[i] = (dwords0)[i] | (dwords1)[i]; \
VG(VALGRIND_CHECK_MEM_IS_DEFINED(dw, ARRAY_SIZE(dwords0) * 4));\
} while (0)
#define anv_batch_emit(batch, cmd, name) \
for (struct cmd name = { __anv_cmd_header(cmd) }, \
*_dst = anv_batch_emit_dwords(batch, __anv_cmd_length(cmd)); \
__builtin_expect(_dst != NULL, 1); \
({ __anv_cmd_pack(cmd)(batch, _dst, &name); \
VG(VALGRIND_CHECK_MEM_IS_DEFINED(_dst, __anv_cmd_length(cmd) * 4)); \
_dst = NULL; \
}))
#define anv_state_pool_emit(pool, cmd, align, ...) ({ \
const uint32_t __size = __anv_cmd_length(cmd) * 4; \
struct anv_state __state = \
anv_state_pool_alloc((pool), __size, align); \
struct cmd __template = { \
__VA_ARGS__ \
}; \
__anv_cmd_pack(cmd)(NULL, __state.map, &__template); \
VG(VALGRIND_CHECK_MEM_IS_DEFINED(__state.map, __anv_cmd_length(cmd) * 4)); \
if (!(pool)->block_pool->device->info.has_llc) \
anv_state_clflush(__state); \
__state; \
})
#define GEN7_MOCS (struct GEN7_MEMORY_OBJECT_CONTROL_STATE) { \
.GraphicsDataTypeGFDT = 0, \
.LLCCacheabilityControlLLCCC = 0, \
.L3CacheabilityControlL3CC = 1, \
}
#define GEN75_MOCS (struct GEN75_MEMORY_OBJECT_CONTROL_STATE) { \
.LLCeLLCCacheabilityControlLLCCC = 0, \
.L3CacheabilityControlL3CC = 1, \
}
#define GEN8_MOCS (struct GEN8_MEMORY_OBJECT_CONTROL_STATE) { \
.MemoryTypeLLCeLLCCacheabilityControl = WB, \
.TargetCache = L3DefertoPATforLLCeLLCselection, \
.AgeforQUADLRU = 0 \
}
/* Skylake: MOCS is now an index into an array of 62 different caching
* configurations programmed by the kernel.
*/
#define GEN9_MOCS (struct GEN9_MEMORY_OBJECT_CONTROL_STATE) { \
/* TC=LLC/eLLC, LeCC=WB, LRUM=3, L3CC=WB */ \
.IndextoMOCSTables = 2 \
}
#define GEN9_MOCS_PTE { \
/* TC=LLC/eLLC, LeCC=WB, LRUM=3, L3CC=WB */ \
.IndextoMOCSTables = 1 \
}
struct anv_device_memory {
struct anv_bo bo;
uint32_t type_index;
VkDeviceSize map_size;
void * map;
};
/**
* Header for Vertex URB Entry (VUE)
*/
struct anv_vue_header {
uint32_t Reserved;
uint32_t RTAIndex; /* RenderTargetArrayIndex */
uint32_t ViewportIndex;
float PointWidth;
};
struct anv_descriptor_set_binding_layout {
#ifndef NDEBUG
/* The type of the descriptors in this binding */
VkDescriptorType type;
#endif
/* Number of array elements in this binding */
uint16_t array_size;
/* Index into the flattend descriptor set */
uint16_t descriptor_index;
/* Index into the dynamic state array for a dynamic buffer */
int16_t dynamic_offset_index;
/* Index into the descriptor set buffer views */
int16_t buffer_index;
struct {
/* Index into the binding table for the associated surface */
int16_t surface_index;
/* Index into the sampler table for the associated sampler */
int16_t sampler_index;
/* Index into the image table for the associated image */
int16_t image_index;
} stage[MESA_SHADER_STAGES];
/* Immutable samplers (or NULL if no immutable samplers) */
struct anv_sampler **immutable_samplers;
};
struct anv_descriptor_set_layout {
/* Number of bindings in this descriptor set */
uint16_t binding_count;
/* Total size of the descriptor set with room for all array entries */
uint16_t size;
/* Shader stages affected by this descriptor set */
uint16_t shader_stages;
/* Number of buffers in this descriptor set */
uint16_t buffer_count;
/* Number of dynamic offsets used by this descriptor set */
uint16_t dynamic_offset_count;
/* Bindings in this descriptor set */
struct anv_descriptor_set_binding_layout binding[0];
};
struct anv_descriptor {
VkDescriptorType type;
union {
struct {
struct anv_image_view *image_view;
struct anv_sampler *sampler;
};
struct anv_buffer_view *buffer_view;
};
};
struct anv_descriptor_set {
const struct anv_descriptor_set_layout *layout;
uint32_t size;
uint32_t buffer_count;
struct anv_buffer_view *buffer_views;
struct anv_descriptor descriptors[0];
};
struct anv_descriptor_pool {
uint32_t size;
uint32_t next;
uint32_t free_list;
struct anv_state_stream surface_state_stream;
void *surface_state_free_list;
char data[0];
};
VkResult
anv_descriptor_set_create(struct anv_device *device,
struct anv_descriptor_pool *pool,
const struct anv_descriptor_set_layout *layout,
struct anv_descriptor_set **out_set);
void
anv_descriptor_set_destroy(struct anv_device *device,
struct anv_descriptor_pool *pool,
struct anv_descriptor_set *set);
#define ANV_DESCRIPTOR_SET_COLOR_ATTACHMENTS UINT8_MAX
struct anv_pipeline_binding {
/* The descriptor set this surface corresponds to. The special value of
* ANV_DESCRIPTOR_SET_COLOR_ATTACHMENTS indicates that the offset refers
* to a color attachment and not a regular descriptor.
*/
uint8_t set;
/* Binding in the descriptor set */
uint8_t binding;
/* Index in the binding */
uint8_t index;
};
struct anv_pipeline_layout {
struct {
struct anv_descriptor_set_layout *layout;
uint32_t dynamic_offset_start;
} set[MAX_SETS];
uint32_t num_sets;
struct {
bool has_dynamic_offsets;
} stage[MESA_SHADER_STAGES];
unsigned char sha1[20];
};
struct anv_buffer {
struct anv_device * device;
VkDeviceSize size;
VkBufferUsageFlags usage;
/* Set when bound */
struct anv_bo * bo;
VkDeviceSize offset;
};
enum anv_cmd_dirty_bits {
ANV_CMD_DIRTY_DYNAMIC_VIEWPORT = 1 << 0, /* VK_DYNAMIC_STATE_VIEWPORT */
ANV_CMD_DIRTY_DYNAMIC_SCISSOR = 1 << 1, /* VK_DYNAMIC_STATE_SCISSOR */
ANV_CMD_DIRTY_DYNAMIC_LINE_WIDTH = 1 << 2, /* VK_DYNAMIC_STATE_LINE_WIDTH */
ANV_CMD_DIRTY_DYNAMIC_DEPTH_BIAS = 1 << 3, /* VK_DYNAMIC_STATE_DEPTH_BIAS */
ANV_CMD_DIRTY_DYNAMIC_BLEND_CONSTANTS = 1 << 4, /* VK_DYNAMIC_STATE_BLEND_CONSTANTS */
ANV_CMD_DIRTY_DYNAMIC_DEPTH_BOUNDS = 1 << 5, /* VK_DYNAMIC_STATE_DEPTH_BOUNDS */
ANV_CMD_DIRTY_DYNAMIC_STENCIL_COMPARE_MASK = 1 << 6, /* VK_DYNAMIC_STATE_STENCIL_COMPARE_MASK */
ANV_CMD_DIRTY_DYNAMIC_STENCIL_WRITE_MASK = 1 << 7, /* VK_DYNAMIC_STATE_STENCIL_WRITE_MASK */
ANV_CMD_DIRTY_DYNAMIC_STENCIL_REFERENCE = 1 << 8, /* VK_DYNAMIC_STATE_STENCIL_REFERENCE */
ANV_CMD_DIRTY_DYNAMIC_ALL = (1 << 9) - 1,
ANV_CMD_DIRTY_PIPELINE = 1 << 9,
ANV_CMD_DIRTY_INDEX_BUFFER = 1 << 10,
ANV_CMD_DIRTY_RENDER_TARGETS = 1 << 11,
};
typedef uint32_t anv_cmd_dirty_mask_t;
enum anv_pipe_bits {
ANV_PIPE_DEPTH_CACHE_FLUSH_BIT = (1 << 0),
ANV_PIPE_STALL_AT_SCOREBOARD_BIT = (1 << 1),
ANV_PIPE_STATE_CACHE_INVALIDATE_BIT = (1 << 2),
ANV_PIPE_CONSTANT_CACHE_INVALIDATE_BIT = (1 << 3),
ANV_PIPE_VF_CACHE_INVALIDATE_BIT = (1 << 4),
ANV_PIPE_DATA_CACHE_FLUSH_BIT = (1 << 5),
ANV_PIPE_TEXTURE_CACHE_INVALIDATE_BIT = (1 << 10),
ANV_PIPE_INSTRUCTION_CACHE_INVALIDATE_BIT = (1 << 11),
ANV_PIPE_RENDER_TARGET_CACHE_FLUSH_BIT = (1 << 12),
ANV_PIPE_DEPTH_STALL_BIT = (1 << 13),
ANV_PIPE_CS_STALL_BIT = (1 << 20),
/* This bit does not exist directly in PIPE_CONTROL. Instead it means that
* a flush has happened but not a CS stall. The next time we do any sort
* of invalidation we need to insert a CS stall at that time. Otherwise,
* we would have to CS stall on every flush which could be bad.
*/
ANV_PIPE_NEEDS_CS_STALL_BIT = (1 << 21),
};
#define ANV_PIPE_FLUSH_BITS ( \
ANV_PIPE_DEPTH_CACHE_FLUSH_BIT | \
ANV_PIPE_DATA_CACHE_FLUSH_BIT | \
ANV_PIPE_RENDER_TARGET_CACHE_FLUSH_BIT)
#define ANV_PIPE_STALL_BITS ( \
ANV_PIPE_STALL_AT_SCOREBOARD_BIT | \
ANV_PIPE_DEPTH_STALL_BIT | \
ANV_PIPE_CS_STALL_BIT)
#define ANV_PIPE_INVALIDATE_BITS ( \
ANV_PIPE_STATE_CACHE_INVALIDATE_BIT | \
ANV_PIPE_CONSTANT_CACHE_INVALIDATE_BIT | \
ANV_PIPE_VF_CACHE_INVALIDATE_BIT | \
ANV_PIPE_DATA_CACHE_FLUSH_BIT | \
ANV_PIPE_TEXTURE_CACHE_INVALIDATE_BIT | \
ANV_PIPE_INSTRUCTION_CACHE_INVALIDATE_BIT)
struct anv_vertex_binding {
struct anv_buffer * buffer;
VkDeviceSize offset;
};
struct anv_push_constants {
/* Current allocated size of this push constants data structure.
* Because a decent chunk of it may not be used (images on SKL, for
* instance), we won't actually allocate the entire structure up-front.
*/
uint32_t size;
/* Push constant data provided by the client through vkPushConstants */
uint8_t client_data[MAX_PUSH_CONSTANTS_SIZE];
/* Our hardware only provides zero-based vertex and instance id so, in
* order to satisfy the vulkan requirements, we may have to push one or
* both of these into the shader.
*/
uint32_t base_vertex;
uint32_t base_instance;
/* Offsets and ranges for dynamically bound buffers */
struct {
uint32_t offset;
uint32_t range;
} dynamic[MAX_DYNAMIC_BUFFERS];
/* Image data for image_load_store on pre-SKL */
struct brw_image_param images[MAX_IMAGES];
};
struct anv_dynamic_state {
struct {
uint32_t count;
VkViewport viewports[MAX_VIEWPORTS];
} viewport;
struct {
uint32_t count;
VkRect2D scissors[MAX_SCISSORS];
} scissor;
float line_width;
struct {
float bias;
float clamp;
float slope;
} depth_bias;
float blend_constants[4];
struct {
float min;
float max;
} depth_bounds;
struct {
uint32_t front;
uint32_t back;
} stencil_compare_mask;
struct {
uint32_t front;
uint32_t back;
} stencil_write_mask;
struct {
uint32_t front;
uint32_t back;
} stencil_reference;
};
extern const struct anv_dynamic_state default_dynamic_state;
void anv_dynamic_state_copy(struct anv_dynamic_state *dest,
const struct anv_dynamic_state *src,
uint32_t copy_mask);
/**
* Attachment state when recording a renderpass instance.
*
* The clear value is valid only if there exists a pending clear.
*/
struct anv_attachment_state {
VkImageAspectFlags pending_clear_aspects;
VkClearValue clear_value;
};
/** State required while building cmd buffer */
struct anv_cmd_state {
/* PIPELINE_SELECT.PipelineSelection */
uint32_t current_pipeline;
const struct gen_l3_config * current_l3_config;
uint32_t vb_dirty;
anv_cmd_dirty_mask_t dirty;
anv_cmd_dirty_mask_t compute_dirty;
enum anv_pipe_bits pending_pipe_bits;
uint32_t num_workgroups_offset;
struct anv_bo *num_workgroups_bo;
VkShaderStageFlags descriptors_dirty;
VkShaderStageFlags push_constants_dirty;
uint32_t scratch_size;
struct anv_pipeline * pipeline;
struct anv_pipeline * compute_pipeline;
struct anv_framebuffer * framebuffer;
struct anv_render_pass * pass;
struct anv_subpass * subpass;
VkRect2D render_area;
uint32_t restart_index;
struct anv_vertex_binding vertex_bindings[MAX_VBS];
struct anv_descriptor_set * descriptors[MAX_SETS];
VkShaderStageFlags push_constant_stages;
struct anv_push_constants * push_constants[MESA_SHADER_STAGES];
struct anv_state binding_tables[MESA_SHADER_STAGES];
struct anv_state samplers[MESA_SHADER_STAGES];
struct anv_dynamic_state dynamic;
bool need_query_wa;
/**
* Array length is anv_cmd_state::pass::attachment_count. Array content is
* valid only when recording a render pass instance.
*/
struct anv_attachment_state * attachments;
struct {
struct anv_buffer * index_buffer;
uint32_t index_type; /**< 3DSTATE_INDEX_BUFFER.IndexFormat */
uint32_t index_offset;
} gen7;
};
struct anv_cmd_pool {
VkAllocationCallbacks alloc;
struct list_head cmd_buffers;
};
#define ANV_CMD_BUFFER_BATCH_SIZE 8192
enum anv_cmd_buffer_exec_mode {
ANV_CMD_BUFFER_EXEC_MODE_PRIMARY,
ANV_CMD_BUFFER_EXEC_MODE_EMIT,
ANV_CMD_BUFFER_EXEC_MODE_GROW_AND_EMIT,
ANV_CMD_BUFFER_EXEC_MODE_CHAIN,
ANV_CMD_BUFFER_EXEC_MODE_COPY_AND_CHAIN,
};
struct anv_cmd_buffer {
VK_LOADER_DATA _loader_data;
struct anv_device * device;
struct anv_cmd_pool * pool;
struct list_head pool_link;
struct anv_batch batch;
/* Fields required for the actual chain of anv_batch_bo's.
*
* These fields are initialized by anv_cmd_buffer_init_batch_bo_chain().
*/
struct list_head batch_bos;
enum anv_cmd_buffer_exec_mode exec_mode;
/* A vector of anv_batch_bo pointers for every batch or surface buffer
* referenced by this command buffer
*
* initialized by anv_cmd_buffer_init_batch_bo_chain()
*/
struct anv_vector seen_bbos;
/* A vector of int32_t's for every block of binding tables.
*
* initialized by anv_cmd_buffer_init_batch_bo_chain()
*/
struct anv_vector bt_blocks;
uint32_t bt_next;
struct anv_reloc_list surface_relocs;
/* Information needed for execbuf
*
* These fields are generated by anv_cmd_buffer_prepare_execbuf().
*/
struct {
struct drm_i915_gem_execbuffer2 execbuf;
struct drm_i915_gem_exec_object2 * objects;
uint32_t bo_count;
struct anv_bo ** bos;
/* Allocated length of the 'objects' and 'bos' arrays */
uint32_t array_length;
bool need_reloc;
} execbuf2;
/* Serial for tracking buffer completion */
uint32_t serial;
/* Stream objects for storing temporary data */
struct anv_state_stream surface_state_stream;
struct anv_state_stream dynamic_state_stream;
VkCommandBufferUsageFlags usage_flags;
VkCommandBufferLevel level;
struct anv_cmd_state state;
};
VkResult anv_cmd_buffer_init_batch_bo_chain(struct anv_cmd_buffer *cmd_buffer);
void anv_cmd_buffer_fini_batch_bo_chain(struct anv_cmd_buffer *cmd_buffer);
void anv_cmd_buffer_reset_batch_bo_chain(struct anv_cmd_buffer *cmd_buffer);
void anv_cmd_buffer_end_batch_buffer(struct anv_cmd_buffer *cmd_buffer);
void anv_cmd_buffer_add_secondary(struct anv_cmd_buffer *primary,
struct anv_cmd_buffer *secondary);
void anv_cmd_buffer_prepare_execbuf(struct anv_cmd_buffer *cmd_buffer);
VkResult anv_cmd_buffer_reset(struct anv_cmd_buffer *cmd_buffer);
VkResult
anv_cmd_buffer_ensure_push_constants_size(struct anv_cmd_buffer *cmd_buffer,
gl_shader_stage stage, uint32_t size);
#define anv_cmd_buffer_ensure_push_constant_field(cmd_buffer, stage, field) \
anv_cmd_buffer_ensure_push_constants_size(cmd_buffer, stage, \
(offsetof(struct anv_push_constants, field) + \
sizeof(cmd_buffer->state.push_constants[0]->field)))
struct anv_state anv_cmd_buffer_emit_dynamic(struct anv_cmd_buffer *cmd_buffer,
const void *data, uint32_t size, uint32_t alignment);
struct anv_state anv_cmd_buffer_merge_dynamic(struct anv_cmd_buffer *cmd_buffer,
uint32_t *a, uint32_t *b,
uint32_t dwords, uint32_t alignment);
struct anv_address
anv_cmd_buffer_surface_base_address(struct anv_cmd_buffer *cmd_buffer);
struct anv_state
anv_cmd_buffer_alloc_binding_table(struct anv_cmd_buffer *cmd_buffer,
uint32_t entries, uint32_t *state_offset);
struct anv_state
anv_cmd_buffer_alloc_surface_state(struct anv_cmd_buffer *cmd_buffer);
struct anv_state
anv_cmd_buffer_alloc_dynamic_state(struct anv_cmd_buffer *cmd_buffer,
uint32_t size, uint32_t alignment);
VkResult
anv_cmd_buffer_new_binding_table_block(struct anv_cmd_buffer *cmd_buffer);
void gen8_cmd_buffer_emit_viewport(struct anv_cmd_buffer *cmd_buffer);
void gen8_cmd_buffer_emit_depth_viewport(struct anv_cmd_buffer *cmd_buffer,
bool depth_clamp_enable);
void gen7_cmd_buffer_emit_scissor(struct anv_cmd_buffer *cmd_buffer);
void anv_cmd_buffer_emit_state_base_address(struct anv_cmd_buffer *cmd_buffer);
void anv_cmd_state_setup_attachments(struct anv_cmd_buffer *cmd_buffer,
const VkRenderPassBeginInfo *info);
struct anv_state
anv_cmd_buffer_push_constants(struct anv_cmd_buffer *cmd_buffer,
gl_shader_stage stage);
struct anv_state
anv_cmd_buffer_cs_push_constants(struct anv_cmd_buffer *cmd_buffer);
void anv_cmd_buffer_clear_subpass(struct anv_cmd_buffer *cmd_buffer);
void anv_cmd_buffer_resolve_subpass(struct anv_cmd_buffer *cmd_buffer);
const struct anv_image_view *
anv_cmd_buffer_get_depth_stencil_view(const struct anv_cmd_buffer *cmd_buffer);
void anv_cmd_buffer_dump(struct anv_cmd_buffer *cmd_buffer);
struct anv_fence {
struct anv_bo bo;
struct drm_i915_gem_execbuffer2 execbuf;
struct drm_i915_gem_exec_object2 exec2_objects[1];
bool ready;
};
struct anv_event {
uint64_t semaphore;
struct anv_state state;
};
struct anv_shader_module {
unsigned char sha1[20];
uint32_t size;
char data[0];
};
void anv_hash_shader(unsigned char *hash, const void *key, size_t key_size,
struct anv_shader_module *module,
const char *entrypoint,
const struct anv_pipeline_layout *pipeline_layout,
const VkSpecializationInfo *spec_info);
static inline gl_shader_stage
vk_to_mesa_shader_stage(VkShaderStageFlagBits vk_stage)
{
assert(__builtin_popcount(vk_stage) == 1);
return ffs(vk_stage) - 1;
}
static inline VkShaderStageFlagBits
mesa_to_vk_shader_stage(gl_shader_stage mesa_stage)
{
return (1 << mesa_stage);
}
#define ANV_STAGE_MASK ((1 << MESA_SHADER_STAGES) - 1)
#define anv_foreach_stage(stage, stage_bits) \
for (gl_shader_stage stage, \
__tmp = (gl_shader_stage)((stage_bits) & ANV_STAGE_MASK); \
stage = __builtin_ffs(__tmp) - 1, __tmp; \
__tmp &= ~(1 << (stage)))
struct anv_pipeline_bind_map {
uint32_t surface_count;
uint32_t sampler_count;
uint32_t image_count;
struct anv_pipeline_binding * surface_to_descriptor;
struct anv_pipeline_binding * sampler_to_descriptor;
};
struct anv_shader_bin {
uint32_t ref_cnt;
struct anv_state kernel;
uint32_t kernel_size;
struct anv_pipeline_bind_map bind_map;
uint32_t prog_data_size;
/* Prog data follows, then the key, both aligned to 8-bytes */
};
struct anv_shader_bin *
anv_shader_bin_create(struct anv_device *device,
const void *key, uint32_t key_size,
const void *kernel, uint32_t kernel_size,
const void *prog_data, uint32_t prog_data_size,
const struct anv_pipeline_bind_map *bind_map);
void
anv_shader_bin_destroy(struct anv_device *device, struct anv_shader_bin *shader);
static inline void
anv_shader_bin_ref(struct anv_shader_bin *shader)
{
assert(shader->ref_cnt >= 1);
__sync_fetch_and_add(&shader->ref_cnt, 1);
}
static inline void
anv_shader_bin_unref(struct anv_device *device, struct anv_shader_bin *shader)
{
assert(shader->ref_cnt >= 1);
if (__sync_fetch_and_add(&shader->ref_cnt, -1) == 1)
anv_shader_bin_destroy(device, shader);
}
static inline const struct brw_stage_prog_data *
anv_shader_bin_get_prog_data(const struct anv_shader_bin *shader)
{
const void *data = shader;
data += align_u32(sizeof(struct anv_shader_bin), 8);
return data;
}
struct anv_pipeline {
struct anv_device * device;
struct anv_batch batch;
uint32_t batch_data[512];
struct anv_reloc_list batch_relocs;
uint32_t dynamic_state_mask;
struct anv_dynamic_state dynamic_state;
struct anv_pipeline_layout * layout;
bool needs_data_cache;
struct anv_shader_bin * shaders[MESA_SHADER_STAGES];
struct {
const struct gen_l3_config * l3_config;
uint32_t total_size;
} urb;
VkShaderStageFlags active_stages;
struct anv_state blend_state;
uint32_t vs_simd8;
uint32_t vs_vec4;
uint32_t ps_ksp0;
uint32_t gs_kernel;
uint32_t cs_simd;
uint32_t vb_used;
uint32_t binding_stride[MAX_VBS];
bool instancing_enable[MAX_VBS];
bool primitive_restart;
uint32_t topology;
uint32_t cs_right_mask;
bool depth_clamp_enable;
struct {
uint32_t sf[7];
uint32_t depth_stencil_state[3];
} gen7;
struct {
uint32_t sf[4];
uint32_t raster[5];
uint32_t wm_depth_stencil[3];
} gen8;
struct {
uint32_t wm_depth_stencil[4];
} gen9;
};
static inline bool
anv_pipeline_has_stage(const struct anv_pipeline *pipeline,
gl_shader_stage stage)
{
return (pipeline->active_stages & mesa_to_vk_shader_stage(stage)) != 0;
}
#define ANV_DECL_GET_PROG_DATA_FUNC(prefix, stage) \
static inline const struct brw_##prefix##_prog_data * \
get_##prefix##_prog_data(struct anv_pipeline *pipeline) \
{ \
if (anv_pipeline_has_stage(pipeline, stage)) { \
return (const struct brw_##prefix##_prog_data *) \
anv_shader_bin_get_prog_data(pipeline->shaders[stage]); \
} else { \
return NULL; \
} \
}
ANV_DECL_GET_PROG_DATA_FUNC(vs, MESA_SHADER_VERTEX)
ANV_DECL_GET_PROG_DATA_FUNC(gs, MESA_SHADER_GEOMETRY)
ANV_DECL_GET_PROG_DATA_FUNC(wm, MESA_SHADER_FRAGMENT)
ANV_DECL_GET_PROG_DATA_FUNC(cs, MESA_SHADER_COMPUTE)
VkResult
anv_pipeline_init(struct anv_pipeline *pipeline, struct anv_device *device,
struct anv_pipeline_cache *cache,
const VkGraphicsPipelineCreateInfo *pCreateInfo,
const VkAllocationCallbacks *alloc);
VkResult
anv_pipeline_compile_cs(struct anv_pipeline *pipeline,
struct anv_pipeline_cache *cache,
const VkComputePipelineCreateInfo *info,
struct anv_shader_module *module,
const char *entrypoint,
const VkSpecializationInfo *spec_info);
struct anv_format {
enum isl_format isl_format:16;
struct isl_swizzle swizzle;
};
struct anv_format
anv_get_format(const struct gen_device_info *devinfo, VkFormat format,
VkImageAspectFlags aspect, VkImageTiling tiling);
static inline enum isl_format
anv_get_isl_format(const struct gen_device_info *devinfo, VkFormat vk_format,
VkImageAspectFlags aspect, VkImageTiling tiling)
{
return anv_get_format(devinfo, vk_format, aspect, tiling).isl_format;
}
void
anv_pipeline_setup_l3_config(struct anv_pipeline *pipeline, bool needs_slm);
/**
* Subsurface of an anv_image.
*/
struct anv_surface {
/** Valid only if isl_surf::size > 0. */
struct isl_surf isl;
/**
* Offset from VkImage's base address, as bound by vkBindImageMemory().
*/
uint32_t offset;
};
struct anv_image {
VkImageType type;
/* The original VkFormat provided by the client. This may not match any
* of the actual surface formats.
*/
VkFormat vk_format;
VkImageAspectFlags aspects;
VkExtent3D extent;
uint32_t levels;
uint32_t array_size;
uint32_t samples; /**< VkImageCreateInfo::samples */
VkImageUsageFlags usage; /**< Superset of VkImageCreateInfo::usage. */
VkImageTiling tiling; /** VkImageCreateInfo::tiling */
VkDeviceSize size;
uint32_t alignment;
/* Set when bound */
struct anv_bo *bo;
VkDeviceSize offset;
/**
* Image subsurfaces
*
* For each foo, anv_image::foo_surface is valid if and only if
* anv_image::aspects has a foo aspect.
*
* The hardware requires that the depth buffer and stencil buffer be
* separate surfaces. From Vulkan's perspective, though, depth and stencil
* reside in the same VkImage. To satisfy both the hardware and Vulkan, we
* allocate the depth and stencil buffers as separate surfaces in the same
* bo.
*/
union {
struct anv_surface color_surface;
struct {
struct anv_surface depth_surface;
struct anv_surface hiz_surface;
struct anv_surface stencil_surface;
};
};
};
static inline uint32_t
anv_get_layerCount(const struct anv_image *image,
const VkImageSubresourceRange *range)
{
return range->layerCount == VK_REMAINING_ARRAY_LAYERS ?
image->array_size - range->baseArrayLayer : range->layerCount;
}
static inline uint32_t
anv_get_levelCount(const struct anv_image *image,
const VkImageSubresourceRange *range)
{
return range->levelCount == VK_REMAINING_MIP_LEVELS ?
image->levels - range->baseMipLevel : range->levelCount;
}
struct anv_image_view {
const struct anv_image *image; /**< VkImageViewCreateInfo::image */
struct anv_bo *bo;
uint32_t offset; /**< Offset into bo. */
struct isl_view isl;
VkImageAspectFlags aspect_mask;
VkFormat vk_format;
VkExtent3D extent; /**< Extent of VkImageViewCreateInfo::baseMipLevel. */
/** RENDER_SURFACE_STATE when using image as a color render target. */
struct anv_state color_rt_surface_state;
/** RENDER_SURFACE_STATE when using image as a sampler surface. */
struct anv_state sampler_surface_state;
/** RENDER_SURFACE_STATE when using image as a storage image. */
struct anv_state storage_surface_state;
struct brw_image_param storage_image_param;
};
struct anv_image_create_info {
const VkImageCreateInfo *vk_info;
/** An opt-in bitmask which filters an ISL-mapping of the Vulkan tiling. */
isl_tiling_flags_t isl_tiling_flags;
uint32_t stride;
};
VkResult anv_image_create(VkDevice _device,
const struct anv_image_create_info *info,
const VkAllocationCallbacks* alloc,
VkImage *pImage);
const struct anv_surface *
anv_image_get_surface_for_aspect_mask(const struct anv_image *image,
VkImageAspectFlags aspect_mask);
static inline bool
anv_image_has_hiz(const struct anv_image *image)
{
/* We must check the aspect because anv_image::hiz_surface belongs to
* a union.
*/
return (image->aspects & VK_IMAGE_ASPECT_DEPTH_BIT) &&
image->hiz_surface.isl.size > 0;
}
struct anv_buffer_view {
enum isl_format format; /**< VkBufferViewCreateInfo::format */
struct anv_bo *bo;
uint32_t offset; /**< Offset into bo. */
uint64_t range; /**< VkBufferViewCreateInfo::range */
struct anv_state surface_state;
struct anv_state storage_surface_state;
struct brw_image_param storage_image_param;
};
enum isl_format
anv_isl_format_for_descriptor_type(VkDescriptorType type);
static inline struct VkExtent3D
anv_sanitize_image_extent(const VkImageType imageType,
const struct VkExtent3D imageExtent)
{
switch (imageType) {
case VK_IMAGE_TYPE_1D:
return (VkExtent3D) { imageExtent.width, 1, 1 };
case VK_IMAGE_TYPE_2D:
return (VkExtent3D) { imageExtent.width, imageExtent.height, 1 };
case VK_IMAGE_TYPE_3D:
return imageExtent;
default:
unreachable("invalid image type");
}
}
static inline struct VkOffset3D
anv_sanitize_image_offset(const VkImageType imageType,
const struct VkOffset3D imageOffset)
{
switch (imageType) {
case VK_IMAGE_TYPE_1D:
return (VkOffset3D) { imageOffset.x, 0, 0 };
case VK_IMAGE_TYPE_2D:
return (VkOffset3D) { imageOffset.x, imageOffset.y, 0 };
case VK_IMAGE_TYPE_3D:
return imageOffset;
default:
unreachable("invalid image type");
}
}
void anv_fill_buffer_surface_state(struct anv_device *device,
struct anv_state state,
enum isl_format format,
uint32_t offset, uint32_t range,
uint32_t stride);
void anv_image_view_fill_image_param(struct anv_device *device,
struct anv_image_view *view,
struct brw_image_param *param);
void anv_buffer_view_fill_image_param(struct anv_device *device,
struct anv_buffer_view *view,
struct brw_image_param *param);
struct anv_sampler {
uint32_t state[4];
};
struct anv_framebuffer {
uint32_t width;
uint32_t height;
uint32_t layers;
uint32_t attachment_count;
struct anv_image_view * attachments[0];
};
struct anv_subpass {
uint32_t input_count;
uint32_t * input_attachments;
uint32_t color_count;
uint32_t * color_attachments;
uint32_t * resolve_attachments;
uint32_t depth_stencil_attachment;
/** Subpass has at least one resolve attachment */
bool has_resolve;
};
struct anv_render_pass_attachment {
VkFormat format;
uint32_t samples;
VkAttachmentLoadOp load_op;
VkAttachmentStoreOp store_op;
VkAttachmentLoadOp stencil_load_op;
};
struct anv_render_pass {
uint32_t attachment_count;
uint32_t subpass_count;
uint32_t * subpass_attachments;
struct anv_render_pass_attachment * attachments;
struct anv_subpass subpasses[0];
};
struct anv_query_pool_slot {
uint64_t begin;
uint64_t end;
uint64_t available;
};
struct anv_query_pool {
VkQueryType type;
uint32_t slots;
struct anv_bo bo;
};
void *anv_lookup_entrypoint(const struct gen_device_info *devinfo,
const char *name);
void anv_dump_image_to_ppm(struct anv_device *device,
struct anv_image *image, unsigned miplevel,
unsigned array_layer, VkImageAspectFlagBits aspect,
const char *filename);
enum anv_dump_action {
ANV_DUMP_FRAMEBUFFERS_BIT = 0x1,
};
void anv_dump_start(struct anv_device *device, enum anv_dump_action actions);
void anv_dump_finish(void);
void anv_dump_add_framebuffer(struct anv_cmd_buffer *cmd_buffer,
struct anv_framebuffer *fb);
#define ANV_DEFINE_HANDLE_CASTS(__anv_type, __VkType) \
\
static inline struct __anv_type * \
__anv_type ## _from_handle(__VkType _handle) \
{ \
return (struct __anv_type *) _handle; \
} \
\
static inline __VkType \
__anv_type ## _to_handle(struct __anv_type *_obj) \
{ \
return (__VkType) _obj; \
}
#define ANV_DEFINE_NONDISP_HANDLE_CASTS(__anv_type, __VkType) \
\
static inline struct __anv_type * \
__anv_type ## _from_handle(__VkType _handle) \
{ \
return (struct __anv_type *)(uintptr_t) _handle; \
} \
\
static inline __VkType \
__anv_type ## _to_handle(struct __anv_type *_obj) \
{ \
return (__VkType)(uintptr_t) _obj; \
}
#define ANV_FROM_HANDLE(__anv_type, __name, __handle) \
struct __anv_type *__name = __anv_type ## _from_handle(__handle)
ANV_DEFINE_HANDLE_CASTS(anv_cmd_buffer, VkCommandBuffer)
ANV_DEFINE_HANDLE_CASTS(anv_device, VkDevice)
ANV_DEFINE_HANDLE_CASTS(anv_instance, VkInstance)
ANV_DEFINE_HANDLE_CASTS(anv_physical_device, VkPhysicalDevice)
ANV_DEFINE_HANDLE_CASTS(anv_queue, VkQueue)
ANV_DEFINE_NONDISP_HANDLE_CASTS(anv_cmd_pool, VkCommandPool)
ANV_DEFINE_NONDISP_HANDLE_CASTS(anv_buffer, VkBuffer)
ANV_DEFINE_NONDISP_HANDLE_CASTS(anv_buffer_view, VkBufferView)
ANV_DEFINE_NONDISP_HANDLE_CASTS(anv_descriptor_pool, VkDescriptorPool)
ANV_DEFINE_NONDISP_HANDLE_CASTS(anv_descriptor_set, VkDescriptorSet)
ANV_DEFINE_NONDISP_HANDLE_CASTS(anv_descriptor_set_layout, VkDescriptorSetLayout)
ANV_DEFINE_NONDISP_HANDLE_CASTS(anv_device_memory, VkDeviceMemory)
ANV_DEFINE_NONDISP_HANDLE_CASTS(anv_fence, VkFence)
ANV_DEFINE_NONDISP_HANDLE_CASTS(anv_event, VkEvent)
ANV_DEFINE_NONDISP_HANDLE_CASTS(anv_framebuffer, VkFramebuffer)
ANV_DEFINE_NONDISP_HANDLE_CASTS(anv_image, VkImage)
ANV_DEFINE_NONDISP_HANDLE_CASTS(anv_image_view, VkImageView);
ANV_DEFINE_NONDISP_HANDLE_CASTS(anv_pipeline_cache, VkPipelineCache)
ANV_DEFINE_NONDISP_HANDLE_CASTS(anv_pipeline, VkPipeline)
ANV_DEFINE_NONDISP_HANDLE_CASTS(anv_pipeline_layout, VkPipelineLayout)
ANV_DEFINE_NONDISP_HANDLE_CASTS(anv_query_pool, VkQueryPool)
ANV_DEFINE_NONDISP_HANDLE_CASTS(anv_render_pass, VkRenderPass)
ANV_DEFINE_NONDISP_HANDLE_CASTS(anv_sampler, VkSampler)
ANV_DEFINE_NONDISP_HANDLE_CASTS(anv_shader_module, VkShaderModule)
#define ANV_DEFINE_STRUCT_CASTS(__anv_type, __VkType) \
\
static inline const __VkType * \
__anv_type ## _to_ ## __VkType(const struct __anv_type *__anv_obj) \
{ \
return (const __VkType *) __anv_obj; \
}
#define ANV_COMMON_TO_STRUCT(__VkType, __vk_name, __common_name) \
const __VkType *__vk_name = anv_common_to_ ## __VkType(__common_name)
ANV_DEFINE_STRUCT_CASTS(anv_common, VkMemoryBarrier)
ANV_DEFINE_STRUCT_CASTS(anv_common, VkBufferMemoryBarrier)
ANV_DEFINE_STRUCT_CASTS(anv_common, VkImageMemoryBarrier)
/* Gen-specific function declarations */
#ifdef genX
# include "anv_genX.h"
#else
# define genX(x) gen7_##x
# include "anv_genX.h"
# undef genX
# define genX(x) gen75_##x
# include "anv_genX.h"
# undef genX
# define genX(x) gen8_##x
# include "anv_genX.h"
# undef genX
# define genX(x) gen9_##x
# include "anv_genX.h"
# undef genX
#endif
#ifdef __cplusplus
}
#endif
#endif /* ANV_PRIVATE_H */
|