summaryrefslogtreecommitdiffstats
path: root/drivers/gpu/drm/i915/i915_cmd_parser.c
blob: 5ac4a999f05a6ebb58ff9a8092d2a474416f97c7 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
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
/*
 * Copyright © 2013 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:
 *    Brad Volkin <bradley.d.volkin@intel.com>
 *
 */

#include "gt/intel_engine.h"

#include "i915_drv.h"
#include "i915_memcpy.h"

/**
 * DOC: batch buffer command parser
 *
 * Motivation:
 * Certain OpenGL features (e.g. transform feedback, performance monitoring)
 * require userspace code to submit batches containing commands such as
 * MI_LOAD_REGISTER_IMM to access various registers. Unfortunately, some
 * generations of the hardware will noop these commands in "unsecure" batches
 * (which includes all userspace batches submitted via i915) even though the
 * commands may be safe and represent the intended programming model of the
 * device.
 *
 * The software command parser is similar in operation to the command parsing
 * done in hardware for unsecure batches. However, the software parser allows
 * some operations that would be noop'd by hardware, if the parser determines
 * the operation is safe, and submits the batch as "secure" to prevent hardware
 * parsing.
 *
 * Threats:
 * At a high level, the hardware (and software) checks attempt to prevent
 * granting userspace undue privileges. There are three categories of privilege.
 *
 * First, commands which are explicitly defined as privileged or which should
 * only be used by the kernel driver. The parser rejects such commands
 *
 * Second, commands which access registers. To support correct/enhanced
 * userspace functionality, particularly certain OpenGL extensions, the parser
 * provides a whitelist of registers which userspace may safely access
 *
 * Third, commands which access privileged memory (i.e. GGTT, HWS page, etc).
 * The parser always rejects such commands.
 *
 * The majority of the problematic commands fall in the MI_* range, with only a
 * few specific commands on each engine (e.g. PIPE_CONTROL and MI_FLUSH_DW).
 *
 * Implementation:
 * Each engine maintains tables of commands and registers which the parser
 * uses in scanning batch buffers submitted to that engine.
 *
 * Since the set of commands that the parser must check for is significantly
 * smaller than the number of commands supported, the parser tables contain only
 * those commands required by the parser. This generally works because command
 * opcode ranges have standard command length encodings. So for commands that
 * the parser does not need to check, it can easily skip them. This is
 * implemented via a per-engine length decoding vfunc.
 *
 * Unfortunately, there are a number of commands that do not follow the standard
 * length encoding for their opcode range, primarily amongst the MI_* commands.
 * To handle this, the parser provides a way to define explicit "skip" entries
 * in the per-engine command tables.
 *
 * Other command table entries map fairly directly to high level categories
 * mentioned above: rejected, register whitelist. The parser implements a number
 * of checks, including the privileged memory checks, via a general bitmasking
 * mechanism.
 */

/*
 * A command that requires special handling by the command parser.
 */
struct drm_i915_cmd_descriptor {
	/*
	 * Flags describing how the command parser processes the command.
	 *
	 * CMD_DESC_FIXED: The command has a fixed length if this is set,
	 *                 a length mask if not set
	 * CMD_DESC_SKIP: The command is allowed but does not follow the
	 *                standard length encoding for the opcode range in
	 *                which it falls
	 * CMD_DESC_REJECT: The command is never allowed
	 * CMD_DESC_REGISTER: The command should be checked against the
	 *                    register whitelist for the appropriate ring
	 */
	u32 flags;
#define CMD_DESC_FIXED    (1<<0)
#define CMD_DESC_SKIP     (1<<1)
#define CMD_DESC_REJECT   (1<<2)
#define CMD_DESC_REGISTER (1<<3)
#define CMD_DESC_BITMASK  (1<<4)

	/*
	 * The command's unique identification bits and the bitmask to get them.
	 * This isn't strictly the opcode field as defined in the spec and may
	 * also include type, subtype, and/or subop fields.
	 */
	struct {
		u32 value;
		u32 mask;
	} cmd;

	/*
	 * The command's length. The command is either fixed length (i.e. does
	 * not include a length field) or has a length field mask. The flag
	 * CMD_DESC_FIXED indicates a fixed length. Otherwise, the command has
	 * a length mask. All command entries in a command table must include
	 * length information.
	 */
	union {
		u32 fixed;
		u32 mask;
	} length;

	/*
	 * Describes where to find a register address in the command to check
	 * against the ring's register whitelist. Only valid if flags has the
	 * CMD_DESC_REGISTER bit set.
	 *
	 * A non-zero step value implies that the command may access multiple
	 * registers in sequence (e.g. LRI), in that case step gives the
	 * distance in dwords between individual offset fields.
	 */
	struct {
		u32 offset;
		u32 mask;
		u32 step;
	} reg;

#define MAX_CMD_DESC_BITMASKS 3
	/*
	 * Describes command checks where a particular dword is masked and
	 * compared against an expected value. If the command does not match
	 * the expected value, the parser rejects it. Only valid if flags has
	 * the CMD_DESC_BITMASK bit set. Only entries where mask is non-zero
	 * are valid.
	 *
	 * If the check specifies a non-zero condition_mask then the parser
	 * only performs the check when the bits specified by condition_mask
	 * are non-zero.
	 */
	struct {
		u32 offset;
		u32 mask;
		u32 expected;
		u32 condition_offset;
		u32 condition_mask;
	} bits[MAX_CMD_DESC_BITMASKS];
};

/*
 * A table of commands requiring special handling by the command parser.
 *
 * Each engine has an array of tables. Each table consists of an array of
 * command descriptors, which must be sorted with command opcodes in
 * ascending order.
 */
struct drm_i915_cmd_table {
	const struct drm_i915_cmd_descriptor *table;
	int count;
};

#define STD_MI_OPCODE_SHIFT  (32 - 9)
#define STD_3D_OPCODE_SHIFT  (32 - 16)
#define STD_2D_OPCODE_SHIFT  (32 - 10)
#define STD_MFX_OPCODE_SHIFT (32 - 16)
#define MIN_OPCODE_SHIFT 16

#define CMD(op, opm, f, lm, fl, ...)				\
	{							\
		.flags = (fl) | ((f) ? CMD_DESC_FIXED : 0),	\
		.cmd = { (op & ~0u << (opm)), ~0u << (opm) },	\
		.length = { (lm) },				\
		__VA_ARGS__					\
	}

/* Convenience macros to compress the tables */
#define SMI STD_MI_OPCODE_SHIFT
#define S3D STD_3D_OPCODE_SHIFT
#define S2D STD_2D_OPCODE_SHIFT
#define SMFX STD_MFX_OPCODE_SHIFT
#define F true
#define S CMD_DESC_SKIP
#define R CMD_DESC_REJECT
#define W CMD_DESC_REGISTER
#define B CMD_DESC_BITMASK

/*            Command                          Mask   Fixed Len   Action
	      ---------------------------------------------------------- */
static const struct drm_i915_cmd_descriptor gen7_common_cmds[] = {
	CMD(  MI_NOOP,                          SMI,    F,  1,      S  ),
	CMD(  MI_USER_INTERRUPT,                SMI,    F,  1,      R  ),
	CMD(  MI_WAIT_FOR_EVENT,                SMI,    F,  1,      R  ),
	CMD(  MI_ARB_CHECK,                     SMI,    F,  1,      S  ),
	CMD(  MI_REPORT_HEAD,                   SMI,    F,  1,      S  ),
	CMD(  MI_SUSPEND_FLUSH,                 SMI,    F,  1,      S  ),
	CMD(  MI_SEMAPHORE_MBOX,                SMI,   !F,  0xFF,   R  ),
	CMD(  MI_STORE_DWORD_INDEX,             SMI,   !F,  0xFF,   R  ),
	CMD(  MI_LOAD_REGISTER_IMM(1),          SMI,   !F,  0xFF,   W,
	      .reg = { .offset = 1, .mask = 0x007FFFFC, .step = 2 }    ),
	CMD(  MI_STORE_REGISTER_MEM,            SMI,    F,  3,     W | B,
	      .reg = { .offset = 1, .mask = 0x007FFFFC },
	      .bits = {{
			.offset = 0,
			.mask = MI_GLOBAL_GTT,
			.expected = 0,
	      }},						       ),
	CMD(  MI_LOAD_REGISTER_MEM,             SMI,    F,  3,     W | B,
	      .reg = { .offset = 1, .mask = 0x007FFFFC },
	      .bits = {{
			.offset = 0,
			.mask = MI_GLOBAL_GTT,
			.expected = 0,
	      }},						       ),
	/*
	 * MI_BATCH_BUFFER_START requires some special handling. It's not
	 * really a 'skip' action but it doesn't seem like it's worth adding
	 * a new action. See intel_engine_cmd_parser().
	 */
	CMD(  MI_BATCH_BUFFER_START,            SMI,   !F,  0xFF,   S  ),
};

static const struct drm_i915_cmd_descriptor gen7_render_cmds[] = {
	CMD(  MI_FLUSH,                         SMI,    F,  1,      S  ),
	CMD(  MI_ARB_ON_OFF,                    SMI,    F,  1,      R  ),
	CMD(  MI_PREDICATE,                     SMI,    F,  1,      S  ),
	CMD(  MI_TOPOLOGY_FILTER,               SMI,    F,  1,      S  ),
	CMD(  MI_SET_APPID,                     SMI,    F,  1,      S  ),
	CMD(  MI_DISPLAY_FLIP,                  SMI,   !F,  0xFF,   R  ),
	CMD(  MI_SET_CONTEXT,                   SMI,   !F,  0xFF,   R  ),
	CMD(  MI_URB_CLEAR,                     SMI,   !F,  0xFF,   S  ),
	CMD(  MI_STORE_DWORD_IMM,               SMI,   !F,  0x3F,   B,
	      .bits = {{
			.offset = 0,
			.mask = MI_GLOBAL_GTT,
			.expected = 0,
	      }},						       ),
	CMD(  MI_UPDATE_GTT,                    SMI,   !F,  0xFF,   R  ),
	CMD(  MI_CLFLUSH,                       SMI,   !F,  0x3FF,  B,
	      .bits = {{
			.offset = 0,
			.mask = MI_GLOBAL_GTT,
			.expected = 0,
	      }},						       ),
	CMD(  MI_REPORT_PERF_COUNT,             SMI,   !F,  0x3F,   B,
	      .bits = {{
			.offset = 1,
			.mask = MI_REPORT_PERF_COUNT_GGTT,
			.expected = 0,
	      }},						       ),
	CMD(  MI_CONDITIONAL_BATCH_BUFFER_END,  SMI,   !F,  0xFF,   B,
	      .bits = {{
			.offset = 0,
			.mask = MI_GLOBAL_GTT,
			.expected = 0,
	      }},						       ),
	CMD(  GFX_OP_3DSTATE_VF_STATISTICS,     S3D,    F,  1,      S  ),
	CMD(  PIPELINE_SELECT,                  S3D,    F,  1,      S  ),
	CMD(  MEDIA_VFE_STATE,			S3D,   !F,  0xFFFF, B,
	      .bits = {{
			.offset = 2,
			.mask = MEDIA_VFE_STATE_MMIO_ACCESS_MASK,
			.expected = 0,
	      }},						       ),
	CMD(  GPGPU_OBJECT,                     S3D,   !F,  0xFF,   S  ),
	CMD(  GPGPU_WALKER,                     S3D,   !F,  0xFF,   S  ),
	CMD(  GFX_OP_3DSTATE_SO_DECL_LIST,      S3D,   !F,  0x1FF,  S  ),
	CMD(  GFX_OP_PIPE_CONTROL(5),           S3D,   !F,  0xFF,   B,
	      .bits = {{
			.offset = 1,
			.mask = (PIPE_CONTROL_MMIO_WRITE | PIPE_CONTROL_NOTIFY),
			.expected = 0,
	      },
	      {
			.offset = 1,
		        .mask = (PIPE_CONTROL_GLOBAL_GTT_IVB |
				 PIPE_CONTROL_STORE_DATA_INDEX),
			.expected = 0,
			.condition_offset = 1,
			.condition_mask = PIPE_CONTROL_POST_SYNC_OP_MASK,
	      }},						       ),
};

static const struct drm_i915_cmd_descriptor hsw_render_cmds[] = {
	CMD(  MI_SET_PREDICATE,                 SMI,    F,  1,      S  ),
	CMD(  MI_RS_CONTROL,                    SMI,    F,  1,      S  ),
	CMD(  MI_URB_ATOMIC_ALLOC,              SMI,    F,  1,      S  ),
	CMD(  MI_SET_APPID,                     SMI,    F,  1,      S  ),
	CMD(  MI_RS_CONTEXT,                    SMI,    F,  1,      S  ),
	CMD(  MI_LOAD_SCAN_LINES_INCL,          SMI,   !F,  0x3F,   R  ),
	CMD(  MI_LOAD_SCAN_LINES_EXCL,          SMI,   !F,  0x3F,   R  ),
	CMD(  MI_LOAD_REGISTER_REG,             SMI,   !F,  0xFF,   W,
	      .reg = { .offset = 1, .mask = 0x007FFFFC, .step = 1 }    ),
	CMD(  MI_RS_STORE_DATA_IMM,             SMI,   !F,  0xFF,   S  ),
	CMD(  MI_LOAD_URB_MEM,                  SMI,   !F,  0xFF,   S  ),
	CMD(  MI_STORE_URB_MEM,                 SMI,   !F,  0xFF,   S  ),
	CMD(  GFX_OP_3DSTATE_DX9_CONSTANTF_VS,  S3D,   !F,  0x7FF,  S  ),
	CMD(  GFX_OP_3DSTATE_DX9_CONSTANTF_PS,  S3D,   !F,  0x7FF,  S  ),

	CMD(  GFX_OP_3DSTATE_BINDING_TABLE_EDIT_VS,  S3D,   !F,  0x1FF,  S  ),
	CMD(  GFX_OP_3DSTATE_BINDING_TABLE_EDIT_GS,  S3D,   !F,  0x1FF,  S  ),
	CMD(  GFX_OP_3DSTATE_BINDING_TABLE_EDIT_HS,  S3D,   !F,  0x1FF,  S  ),
	CMD(  GFX_OP_3DSTATE_BINDING_TABLE_EDIT_DS,  S3D,   !F,  0x1FF,  S  ),
	CMD(  GFX_OP_3DSTATE_BINDING_TABLE_EDIT_PS,  S3D,   !F,  0x1FF,  S  ),
};

static const struct drm_i915_cmd_descriptor gen7_video_cmds[] = {
	CMD(  MI_ARB_ON_OFF,                    SMI,    F,  1,      R  ),
	CMD(  MI_SET_APPID,                     SMI,    F,  1,      S  ),
	CMD(  MI_STORE_DWORD_IMM,               SMI,   !F,  0xFF,   B,
	      .bits = {{
			.offset = 0,
			.mask = MI_GLOBAL_GTT,
			.expected = 0,
	      }},						       ),
	CMD(  MI_UPDATE_GTT,                    SMI,   !F,  0x3F,   R  ),
	CMD(  MI_FLUSH_DW,                      SMI,   !F,  0x3F,   B,
	      .bits = {{
			.offset = 0,
			.mask = MI_FLUSH_DW_NOTIFY,
			.expected = 0,
	      },
	      {
			.offset = 1,
			.mask = MI_FLUSH_DW_USE_GTT,
			.expected = 0,
			.condition_offset = 0,
			.condition_mask = MI_FLUSH_DW_OP_MASK,
	      },
	      {
			.offset = 0,
			.mask = MI_FLUSH_DW_STORE_INDEX,
			.expected = 0,
			.condition_offset = 0,
			.condition_mask = MI_FLUSH_DW_OP_MASK,
	      }},						       ),
	CMD(  MI_CONDITIONAL_BATCH_BUFFER_END,  SMI,   !F,  0xFF,   B,
	      .bits = {{
			.offset = 0,
			.mask = MI_GLOBAL_GTT,
			.expected = 0,
	      }},						       ),
	/*
	 * MFX_WAIT doesn't fit the way we handle length for most commands.
	 * It has a length field but it uses a non-standard length bias.
	 * It is always 1 dword though, so just treat it as fixed length.
	 */
	CMD(  MFX_WAIT,                         SMFX,   F,  1,      S  ),
};

static const struct drm_i915_cmd_descriptor gen7_vecs_cmds[] = {
	CMD(  MI_ARB_ON_OFF,                    SMI,    F,  1,      R  ),
	CMD(  MI_SET_APPID,                     SMI,    F,  1,      S  ),
	CMD(  MI_STORE_DWORD_IMM,               SMI,   !F,  0xFF,   B,
	      .bits = {{
			.offset = 0,
			.mask = MI_GLOBAL_GTT,
			.expected = 0,
	      }},						       ),
	CMD(  MI_UPDATE_GTT,                    SMI,   !F,  0x3F,   R  ),
	CMD(  MI_FLUSH_DW,                      SMI,   !F,  0x3F,   B,
	      .bits = {{
			.offset = 0,
			.mask = MI_FLUSH_DW_NOTIFY,
			.expected = 0,
	      },
	      {
			.offset = 1,
			.mask = MI_FLUSH_DW_USE_GTT,
			.expected = 0,
			.condition_offset = 0,
			.condition_mask = MI_FLUSH_DW_OP_MASK,
	      },
	      {
			.offset = 0,
			.mask = MI_FLUSH_DW_STORE_INDEX,
			.expected = 0,
			.condition_offset = 0,
			.condition_mask = MI_FLUSH_DW_OP_MASK,
	      }},						       ),
	CMD(  MI_CONDITIONAL_BATCH_BUFFER_END,  SMI,   !F,  0xFF,   B,
	      .bits = {{
			.offset = 0,
			.mask = MI_GLOBAL_GTT,
			.expected = 0,
	      }},						       ),
};

static const struct drm_i915_cmd_descriptor gen7_blt_cmds[] = {
	CMD(  MI_DISPLAY_FLIP,                  SMI,   !F,  0xFF,   R  ),
	CMD(  MI_STORE_DWORD_IMM,               SMI,   !F,  0x3FF,  B,
	      .bits = {{
			.offset = 0,
			.mask = MI_GLOBAL_GTT,
			.expected = 0,
	      }},						       ),
	CMD(  MI_UPDATE_GTT,                    SMI,   !F,  0x3F,   R  ),
	CMD(  MI_FLUSH_DW,                      SMI,   !F,  0x3F,   B,
	      .bits = {{
			.offset = 0,
			.mask = MI_FLUSH_DW_NOTIFY,
			.expected = 0,
	      },
	      {
			.offset = 1,
			.mask = MI_FLUSH_DW_USE_GTT,
			.expected = 0,
			.condition_offset = 0,
			.condition_mask = MI_FLUSH_DW_OP_MASK,
	      },
	      {
			.offset = 0,
			.mask = MI_FLUSH_DW_STORE_INDEX,
			.expected = 0,
			.condition_offset = 0,
			.condition_mask = MI_FLUSH_DW_OP_MASK,
	      }},						       ),
	CMD(  COLOR_BLT,                        S2D,   !F,  0x3F,   S  ),
	CMD(  SRC_COPY_BLT,                     S2D,   !F,  0x3F,   S  ),
};

static const struct drm_i915_cmd_descriptor hsw_blt_cmds[] = {
	CMD(  MI_LOAD_SCAN_LINES_INCL,          SMI,   !F,  0x3F,   R  ),
	CMD(  MI_LOAD_SCAN_LINES_EXCL,          SMI,   !F,  0x3F,   R  ),
};

/*
 * For Gen9 we can still rely on the h/w to enforce cmd security, and only
 * need to re-enforce the register access checks. We therefore only need to
 * teach the cmdparser how to find the end of each command, and identify
 * register accesses. The table doesn't need to reject any commands, and so
 * the only commands listed here are:
 *   1) Those that touch registers
 *   2) Those that do not have the default 8-bit length
 *
 * Note that the default MI length mask chosen for this table is 0xFF, not
 * the 0x3F used on older devices. This is because the vast majority of MI
 * cmds on Gen9 use a standard 8-bit Length field.
 * All the Gen9 blitter instructions are standard 0xFF length mask, and
 * none allow access to non-general registers, so in fact no BLT cmds are
 * included in the table at all.
 *
 */
static const struct drm_i915_cmd_descriptor gen9_blt_cmds[] = {
	CMD(  MI_NOOP,                          SMI,    F,  1,      S  ),
	CMD(  MI_USER_INTERRUPT,                SMI,    F,  1,      S  ),
	CMD(  MI_WAIT_FOR_EVENT,                SMI,    F,  1,      S  ),
	CMD(  MI_FLUSH,                         SMI,    F,  1,      S  ),
	CMD(  MI_ARB_CHECK,                     SMI,    F,  1,      S  ),
	CMD(  MI_REPORT_HEAD,                   SMI,    F,  1,      S  ),
	CMD(  MI_ARB_ON_OFF,                    SMI,    F,  1,      S  ),
	CMD(  MI_SUSPEND_FLUSH,                 SMI,    F,  1,      S  ),
	CMD(  MI_LOAD_SCAN_LINES_INCL,          SMI,   !F,  0x3F,   S  ),
	CMD(  MI_LOAD_SCAN_LINES_EXCL,          SMI,   !F,  0x3F,   S  ),
	CMD(  MI_STORE_DWORD_IMM,               SMI,   !F,  0x3FF,  S  ),
	CMD(  MI_LOAD_REGISTER_IMM(1),          SMI,   !F,  0xFF,   W,
	      .reg = { .offset = 1, .mask = 0x007FFFFC, .step = 2 }    ),
	CMD(  MI_UPDATE_GTT,                    SMI,   !F,  0x3FF,  S  ),
	CMD(  MI_STORE_REGISTER_MEM_GEN8,       SMI,    F,  4,      W,
	      .reg = { .offset = 1, .mask = 0x007FFFFC }               ),
	CMD(  MI_FLUSH_DW,                      SMI,   !F,  0x3F,   S  ),
	CMD(  MI_LOAD_REGISTER_MEM_GEN8,        SMI,    F,  4,      W,
	      .reg = { .offset = 1, .mask = 0x007FFFFC }               ),
	CMD(  MI_LOAD_REGISTER_REG,             SMI,    !F,  0xFF,  W,
	      .reg = { .offset = 1, .mask = 0x007FFFFC, .step = 1 }    ),

	/*
	 * We allow BB_START but apply further checks. We just sanitize the
	 * basic fields here.
	 */
#define MI_BB_START_OPERAND_MASK   GENMASK(SMI-1, 0)
#define MI_BB_START_OPERAND_EXPECT (MI_BATCH_PPGTT_HSW | 1)
	CMD(  MI_BATCH_BUFFER_START_GEN8,       SMI,    !F,  0xFF,  B,
	      .bits = {{
			.offset = 0,
			.mask = MI_BB_START_OPERAND_MASK,
			.expected = MI_BB_START_OPERAND_EXPECT,
	      }},						       ),
};

static const struct drm_i915_cmd_descriptor noop_desc =
	CMD(MI_NOOP, SMI, F, 1, S);

#undef CMD
#undef SMI
#undef S3D
#undef S2D
#undef SMFX
#undef F
#undef S
#undef R
#undef W
#undef B

static const struct drm_i915_cmd_table gen7_render_cmd_table[] = {
	{ gen7_common_cmds, ARRAY_SIZE(gen7_common_cmds) },
	{ gen7_render_cmds, ARRAY_SIZE(gen7_render_cmds) },
};

static const struct drm_i915_cmd_table hsw_render_ring_cmd_table[] = {
	{ gen7_common_cmds, ARRAY_SIZE(gen7_common_cmds) },
	{ gen7_render_cmds, ARRAY_SIZE(gen7_render_cmds) },
	{ hsw_render_cmds, ARRAY_SIZE(hsw_render_cmds) },
};

static const struct drm_i915_cmd_table gen7_video_cmd_table[] = {
	{ gen7_common_cmds, ARRAY_SIZE(gen7_common_cmds) },
	{ gen7_video_cmds, ARRAY_SIZE(gen7_video_cmds) },
};

static const struct drm_i915_cmd_table hsw_vebox_cmd_table[] = {
	{ gen7_common_cmds, ARRAY_SIZE(gen7_common_cmds) },
	{ gen7_vecs_cmds, ARRAY_SIZE(gen7_vecs_cmds) },
};

static const struct drm_i915_cmd_table gen7_blt_cmd_table[] = {
	{ gen7_common_cmds, ARRAY_SIZE(gen7_common_cmds) },
	{ gen7_blt_cmds, ARRAY_SIZE(gen7_blt_cmds) },
};

static const struct drm_i915_cmd_table hsw_blt_ring_cmd_table[] = {
	{ gen7_common_cmds, ARRAY_SIZE(gen7_common_cmds) },
	{ gen7_blt_cmds, ARRAY_SIZE(gen7_blt_cmds) },
	{ hsw_blt_cmds, ARRAY_SIZE(hsw_blt_cmds) },
};

static const struct drm_i915_cmd_table gen9_blt_cmd_table[] = {
	{ gen9_blt_cmds, ARRAY_SIZE(gen9_blt_cmds) },
};


/*
 * Register whitelists, sorted by increasing register offset.
 */

/*
 * An individual whitelist entry granting access to register addr.  If
 * mask is non-zero the argument of immediate register writes will be
 * AND-ed with mask, and the command will be rejected if the result
 * doesn't match value.
 *
 * Registers with non-zero mask are only allowed to be written using
 * LRI.
 */
struct drm_i915_reg_descriptor {
	i915_reg_t addr;
	u32 mask;
	u32 value;
};

/* Convenience macro for adding 32-bit registers. */
#define REG32(_reg, ...) \
	{ .addr = (_reg), __VA_ARGS__ }

#define REG32_IDX(_reg, idx) \
	{ .addr = _reg(idx) }

/*
 * Convenience macro for adding 64-bit registers.
 *
 * Some registers that userspace accesses are 64 bits. The register
 * access commands only allow 32-bit accesses. Hence, we have to include
 * entries for both halves of the 64-bit registers.
 */
#define REG64(_reg) \
	{ .addr = _reg }, \
	{ .addr = _reg ## _UDW }

#define REG64_IDX(_reg, idx) \
	{ .addr = _reg(idx) }, \
	{ .addr = _reg ## _UDW(idx) }

static const struct drm_i915_reg_descriptor gen7_render_regs[] = {
	REG64(GPGPU_THREADS_DISPATCHED),
	REG64(HS_INVOCATION_COUNT),
	REG64(DS_INVOCATION_COUNT),
	REG64(IA_VERTICES_COUNT),
	REG64(IA_PRIMITIVES_COUNT),
	REG64(VS_INVOCATION_COUNT),
	REG64(GS_INVOCATION_COUNT),
	REG64(GS_PRIMITIVES_COUNT),
	REG64(CL_INVOCATION_COUNT),
	REG64(CL_PRIMITIVES_COUNT),
	REG64(PS_INVOCATION_COUNT),
	REG64(PS_DEPTH_COUNT),
	REG64_IDX(RING_TIMESTAMP, RENDER_RING_BASE),
	REG64(MI_PREDICATE_SRC0),
	REG64(MI_PREDICATE_SRC1),
	REG32(GEN7_3DPRIM_END_OFFSET),
	REG32(GEN7_3DPRIM_START_VERTEX),
	REG32(GEN7_3DPRIM_VERTEX_COUNT),
	REG32(GEN7_3DPRIM_INSTANCE_COUNT),
	REG32(GEN7_3DPRIM_START_INSTANCE),
	REG32(GEN7_3DPRIM_BASE_VERTEX),
	REG32(GEN7_GPGPU_DISPATCHDIMX),
	REG32(GEN7_GPGPU_DISPATCHDIMY),
	REG32(GEN7_GPGPU_DISPATCHDIMZ),
	REG64_IDX(RING_TIMESTAMP, BSD_RING_BASE),
	REG64_IDX(GEN7_SO_NUM_PRIMS_WRITTEN, 0),
	REG64_IDX(GEN7_SO_NUM_PRIMS_WRITTEN, 1),
	REG64_IDX(GEN7_SO_NUM_PRIMS_WRITTEN, 2),
	REG64_IDX(GEN7_SO_NUM_PRIMS_WRITTEN, 3),
	REG64_IDX(GEN7_SO_PRIM_STORAGE_NEEDED, 0),
	REG64_IDX(GEN7_SO_PRIM_STORAGE_NEEDED, 1),
	REG64_IDX(GEN7_SO_PRIM_STORAGE_NEEDED, 2),
	REG64_IDX(GEN7_SO_PRIM_STORAGE_NEEDED, 3),
	REG32(GEN7_SO_WRITE_OFFSET(0)),
	REG32(GEN7_SO_WRITE_OFFSET(1)),
	REG32(GEN7_SO_WRITE_OFFSET(2)),
	REG32(GEN7_SO_WRITE_OFFSET(3)),
	REG32(GEN7_L3SQCREG1),
	REG32(GEN7_L3CNTLREG2),
	REG32(GEN7_L3CNTLREG3),
	REG64_IDX(RING_TIMESTAMP, BLT_RING_BASE),
};

static const struct drm_i915_reg_descriptor hsw_render_regs[] = {
	REG64_IDX(HSW_CS_GPR, 0),
	REG64_IDX(HSW_CS_GPR, 1),
	REG64_IDX(HSW_CS_GPR, 2),
	REG64_IDX(HSW_CS_GPR, 3),
	REG64_IDX(HSW_CS_GPR, 4),
	REG64_IDX(HSW_CS_GPR, 5),
	REG64_IDX(HSW_CS_GPR, 6),
	REG64_IDX(HSW_CS_GPR, 7),
	REG64_IDX(HSW_CS_GPR, 8),
	REG64_IDX(HSW_CS_GPR, 9),
	REG64_IDX(HSW_CS_GPR, 10),
	REG64_IDX(HSW_CS_GPR, 11),
	REG64_IDX(HSW_CS_GPR, 12),
	REG64_IDX(HSW_CS_GPR, 13),
	REG64_IDX(HSW_CS_GPR, 14),
	REG64_IDX(HSW_CS_GPR, 15),
	REG32(HSW_SCRATCH1,
	      .mask = ~HSW_SCRATCH1_L3_DATA_ATOMICS_DISABLE,
	      .value = 0),
	REG32(HSW_ROW_CHICKEN3,
	      .mask = ~(HSW_ROW_CHICKEN3_L3_GLOBAL_ATOMICS_DISABLE << 16 |
                        HSW_ROW_CHICKEN3_L3_GLOBAL_ATOMICS_DISABLE),
	      .value = 0),
};

static const struct drm_i915_reg_descriptor gen7_blt_regs[] = {
	REG64_IDX(RING_TIMESTAMP, RENDER_RING_BASE),
	REG64_IDX(RING_TIMESTAMP, BSD_RING_BASE),
	REG32(BCS_SWCTRL),
	REG64_IDX(RING_TIMESTAMP, BLT_RING_BASE),
};

static const struct drm_i915_reg_descriptor gen9_blt_regs[] = {
	REG64_IDX(RING_TIMESTAMP, RENDER_RING_BASE),
	REG64_IDX(RING_TIMESTAMP, BSD_RING_BASE),
	REG32(BCS_SWCTRL),
	REG64_IDX(RING_TIMESTAMP, BLT_RING_BASE),
	REG32_IDX(RING_CTX_TIMESTAMP, BLT_RING_BASE),
	REG64_IDX(BCS_GPR, 0),
	REG64_IDX(BCS_GPR, 1),
	REG64_IDX(BCS_GPR, 2),
	REG64_IDX(BCS_GPR, 3),
	REG64_IDX(BCS_GPR, 4),
	REG64_IDX(BCS_GPR, 5),
	REG64_IDX(BCS_GPR, 6),
	REG64_IDX(BCS_GPR, 7),
	REG64_IDX(BCS_GPR, 8),
	REG64_IDX(BCS_GPR, 9),
	REG64_IDX(BCS_GPR, 10),
	REG64_IDX(BCS_GPR, 11),
	REG64_IDX(BCS_GPR, 12),
	REG64_IDX(BCS_GPR, 13),
	REG64_IDX(BCS_GPR, 14),
	REG64_IDX(BCS_GPR, 15),
};

#undef REG64
#undef REG32

struct drm_i915_reg_table {
	const struct drm_i915_reg_descriptor *regs;
	int num_regs;
};

static const struct drm_i915_reg_table ivb_render_reg_tables[] = {
	{ gen7_render_regs, ARRAY_SIZE(gen7_render_regs) },
};

static const struct drm_i915_reg_table ivb_blt_reg_tables[] = {
	{ gen7_blt_regs, ARRAY_SIZE(gen7_blt_regs) },
};

static const struct drm_i915_reg_table hsw_render_reg_tables[] = {
	{ gen7_render_regs, ARRAY_SIZE(gen7_render_regs) },
	{ hsw_render_regs, ARRAY_SIZE(hsw_render_regs) },
};

static const struct drm_i915_reg_table hsw_blt_reg_tables[] = {
	{ gen7_blt_regs, ARRAY_SIZE(gen7_blt_regs) },
};

static const struct drm_i915_reg_table gen9_blt_reg_tables[] = {
	{ gen9_blt_regs, ARRAY_SIZE(gen9_blt_regs) },
};

static u32 gen7_render_get_cmd_length_mask(u32 cmd_header)
{
	u32 client = cmd_header >> INSTR_CLIENT_SHIFT;
	u32 subclient =
		(cmd_header & INSTR_SUBCLIENT_MASK) >> INSTR_SUBCLIENT_SHIFT;

	if (client == INSTR_MI_CLIENT)
		return 0x3F;
	else if (client == INSTR_RC_CLIENT) {
		if (subclient == INSTR_MEDIA_SUBCLIENT)
			return 0xFFFF;
		else
			return 0xFF;
	}

	DRM_DEBUG("CMD: Abnormal rcs cmd length! 0x%08X\n", cmd_header);
	return 0;
}

static u32 gen7_bsd_get_cmd_length_mask(u32 cmd_header)
{
	u32 client = cmd_header >> INSTR_CLIENT_SHIFT;
	u32 subclient =
		(cmd_header & INSTR_SUBCLIENT_MASK) >> INSTR_SUBCLIENT_SHIFT;
	u32 op = (cmd_header & INSTR_26_TO_24_MASK) >> INSTR_26_TO_24_SHIFT;

	if (client == INSTR_MI_CLIENT)
		return 0x3F;
	else if (client == INSTR_RC_CLIENT) {
		if (subclient == INSTR_MEDIA_SUBCLIENT) {
			if (op == 6)
				return 0xFFFF;
			else
				return 0xFFF;
		} else
			return 0xFF;
	}

	DRM_DEBUG("CMD: Abnormal bsd cmd length! 0x%08X\n", cmd_header);
	return 0;
}

static u32 gen7_blt_get_cmd_length_mask(u32 cmd_header)
{
	u32 client = cmd_header >> INSTR_CLIENT_SHIFT;

	if (client == INSTR_MI_CLIENT)
		return 0x3F;
	else if (client == INSTR_BC_CLIENT)
		return 0xFF;

	DRM_DEBUG("CMD: Abnormal blt cmd length! 0x%08X\n", cmd_header);
	return 0;
}

static u32 gen9_blt_get_cmd_length_mask(u32 cmd_header)
{
	u32 client = cmd_header >> INSTR_CLIENT_SHIFT;

	if (client == INSTR_MI_CLIENT || client == INSTR_BC_CLIENT)
		return 0xFF;

	DRM_DEBUG("CMD: Abnormal blt cmd length! 0x%08X\n", cmd_header);
	return 0;
}

static bool validate_cmds_sorted(const struct intel_engine_cs *engine,
				 const struct drm_i915_cmd_table *cmd_tables,
				 int cmd_table_count)
{
	int i;
	bool ret = true;

	if (!cmd_tables || cmd_table_count == 0)
		return true;

	for (i = 0; i < cmd_table_count; i++) {
		const struct drm_i915_cmd_table *table = &cmd_tables[i];
		u32 previous = 0;
		int j;

		for (j = 0; j < table->count; j++) {
			const struct drm_i915_cmd_descriptor *desc =
				&table->table[j];
			u32 curr = desc->cmd.value & desc->cmd.mask;

			if (curr < previous) {
				drm_err(&engine->i915->drm,
					"CMD: %s [%d] command table not sorted: "
					"table=%d entry=%d cmd=0x%08X prev=0x%08X\n",
					engine->name, engine->id,
					i, j, curr, previous);
				ret = false;
			}

			previous = curr;
		}
	}

	return ret;
}

static bool check_sorted(const struct intel_engine_cs *engine,
			 const struct drm_i915_reg_descriptor *reg_table,
			 int reg_count)
{
	int i;
	u32 previous = 0;
	bool ret = true;

	for (i = 0; i < reg_count; i++) {
		u32 curr = i915_mmio_reg_offset(reg_table[i].addr);

		if (curr < previous) {
			drm_err(&engine->i915->drm,
				"CMD: %s [%d] register table not sorted: "
				"entry=%d reg=0x%08X prev=0x%08X\n",
				engine->name, engine->id,
				i, curr, previous);
			ret = false;
		}

		previous = curr;
	}

	return ret;
}

static bool validate_regs_sorted(struct intel_engine_cs *engine)
{
	int i;
	const struct drm_i915_reg_table *table;

	for (i = 0; i < engine->reg_table_count; i++) {
		table = &engine->reg_tables[i];
		if (!check_sorted(engine, table->regs, table->num_regs))
			return false;
	}

	return true;
}

struct cmd_node {
	const struct drm_i915_cmd_descriptor *desc;
	struct hlist_node node;
};

/*
 * Different command ranges have different numbers of bits for the opcode. For
 * example, MI commands use bits 31:23 while 3D commands use bits 31:16. The
 * problem is that, for example, MI commands use bits 22:16 for other fields
 * such as GGTT vs PPGTT bits. If we include those bits in the mask then when
 * we mask a command from a batch it could hash to the wrong bucket due to
 * non-opcode bits being set. But if we don't include those bits, some 3D
 * commands may hash to the same bucket due to not including opcode bits that
 * make the command unique. For now, we will risk hashing to the same bucket.
 */
static inline u32 cmd_header_key(u32 x)
{
	switch (x >> INSTR_CLIENT_SHIFT) {
	default:
	case INSTR_MI_CLIENT:
		return x >> STD_MI_OPCODE_SHIFT;
	case INSTR_RC_CLIENT:
		return x >> STD_3D_OPCODE_SHIFT;
	case INSTR_BC_CLIENT:
		return x >> STD_2D_OPCODE_SHIFT;
	}
}

static int init_hash_table(struct intel_engine_cs *engine,
			   const struct drm_i915_cmd_table *cmd_tables,
			   int cmd_table_count)
{
	int i, j;

	hash_init(engine->cmd_hash);

	for (i = 0; i < cmd_table_count; i++) {
		const struct drm_i915_cmd_table *table = &cmd_tables[i];

		for (j = 0; j < table->count; j++) {
			const struct drm_i915_cmd_descriptor *desc =
				&table->table[j];
			struct cmd_node *desc_node =
				kmalloc(sizeof(*desc_node), GFP_KERNEL);

			if (!desc_node)
				return -ENOMEM;

			desc_node->desc = desc;
			hash_add(engine->cmd_hash, &desc_node->node,
				 cmd_header_key(desc->cmd.value));
		}
	}

	return 0;
}

static void fini_hash_table(struct intel_engine_cs *engine)
{
	struct hlist_node *tmp;
	struct cmd_node *desc_node;
	int i;

	hash_for_each_safe(engine->cmd_hash, i, tmp, desc_node, node) {
		hash_del(&desc_node->node);
		kfree(desc_node);
	}
}

/**
 * intel_engine_init_cmd_parser() - set cmd parser related fields for an engine
 * @engine: the engine to initialize
 *
 * Optionally initializes fields related to batch buffer command parsing in the
 * struct intel_engine_cs based on whether the platform requires software
 * command parsing.
 */
void intel_engine_init_cmd_parser(struct intel_engine_cs *engine)
{
	const struct drm_i915_cmd_table *cmd_tables;
	int cmd_table_count;
	int ret;

	if (!IS_GEN(engine->i915, 7) && !(IS_GEN(engine->i915, 9) &&
					  engine->class == COPY_ENGINE_CLASS))
		return;

	switch (engine->class) {
	case RENDER_CLASS:
		if (IS_HASWELL(engine->i915)) {
			cmd_tables = hsw_render_ring_cmd_table;
			cmd_table_count =
				ARRAY_SIZE(hsw_render_ring_cmd_table);
		} else {
			cmd_tables = gen7_render_cmd_table;
			cmd_table_count = ARRAY_SIZE(gen7_render_cmd_table);
		}

		if (IS_HASWELL(engine->i915)) {
			engine->reg_tables = hsw_render_reg_tables;
			engine->reg_table_count = ARRAY_SIZE(hsw_render_reg_tables);
		} else {
			engine->reg_tables = ivb_render_reg_tables;
			engine->reg_table_count = ARRAY_SIZE(ivb_render_reg_tables);
		}
		engine->get_cmd_length_mask = gen7_render_get_cmd_length_mask;
		break;
	case VIDEO_DECODE_CLASS:
		cmd_tables = gen7_video_cmd_table;
		cmd_table_count = ARRAY_SIZE(gen7_video_cmd_table);
		engine->get_cmd_length_mask = gen7_bsd_get_cmd_length_mask;
		break;
	case COPY_ENGINE_CLASS:
		engine->get_cmd_length_mask = gen7_blt_get_cmd_length_mask;
		if (IS_GEN(engine->i915, 9)) {
			cmd_tables = gen9_blt_cmd_table;
			cmd_table_count = ARRAY_SIZE(gen9_blt_cmd_table);
			engine->get_cmd_length_mask =
				gen9_blt_get_cmd_length_mask;

			/* BCS Engine unsafe without parser */
			engine->flags |= I915_ENGINE_REQUIRES_CMD_PARSER;
		} else if (IS_HASWELL(engine->i915)) {
			cmd_tables = hsw_blt_ring_cmd_table;
			cmd_table_count = ARRAY_SIZE(hsw_blt_ring_cmd_table);
		} else {
			cmd_tables = gen7_blt_cmd_table;
			cmd_table_count = ARRAY_SIZE(gen7_blt_cmd_table);
		}

		if (IS_GEN(engine->i915, 9)) {
			engine->reg_tables = gen9_blt_reg_tables;
			engine->reg_table_count =
				ARRAY_SIZE(gen9_blt_reg_tables);
		} else if (IS_HASWELL(engine->i915)) {
			engine->reg_tables = hsw_blt_reg_tables;
			engine->reg_table_count = ARRAY_SIZE(hsw_blt_reg_tables);
		} else {
			engine->reg_tables = ivb_blt_reg_tables;
			engine->reg_table_count = ARRAY_SIZE(ivb_blt_reg_tables);
		}
		break;
	case VIDEO_ENHANCEMENT_CLASS:
		cmd_tables = hsw_vebox_cmd_table;
		cmd_table_count = ARRAY_SIZE(hsw_vebox_cmd_table);
		/* VECS can use the same length_mask function as VCS */
		engine->get_cmd_length_mask = gen7_bsd_get_cmd_length_mask;
		break;
	default:
		MISSING_CASE(engine->class);
		return;
	}

	if (!validate_cmds_sorted(engine, cmd_tables, cmd_table_count)) {
		drm_err(&engine->i915->drm,
			"%s: command descriptions are not sorted\n",
			engine->name);
		return;
	}
	if (!validate_regs_sorted(engine)) {
		drm_err(&engine->i915->drm,
			"%s: registers are not sorted\n", engine->name);
		return;
	}

	ret = init_hash_table(engine, cmd_tables, cmd_table_count);
	if (ret) {
		drm_err(&engine->i915->drm,
			"%s: initialised failed!\n", engine->name);
		fini_hash_table(engine);
		return;
	}

	engine->flags |= I915_ENGINE_USING_CMD_PARSER;
}

/**
 * intel_engine_cleanup_cmd_parser() - clean up cmd parser related fields
 * @engine: the engine to clean up
 *
 * Releases any resources related to command parsing that may have been
 * initialized for the specified engine.
 */
void intel_engine_cleanup_cmd_parser(struct intel_engine_cs *engine)
{
	if (!intel_engine_using_cmd_parser(engine))
		return;

	fini_hash_table(engine);
}

static const struct drm_i915_cmd_descriptor*
find_cmd_in_table(struct intel_engine_cs *engine,
		  u32 cmd_header)
{
	struct cmd_node *desc_node;

	hash_for_each_possible(engine->cmd_hash, desc_node, node,
			       cmd_header_key(cmd_header)) {
		const struct drm_i915_cmd_descriptor *desc = desc_node->desc;
		if (((cmd_header ^ desc->cmd.value) & desc->cmd.mask) == 0)
			return desc;
	}

	return NULL;
}

/*
 * Returns a pointer to a descriptor for the command specified by cmd_header.
 *
 * The caller must supply space for a default descriptor via the default_desc
 * parameter. If no descriptor for the specified command exists in the engine's
 * command parser tables, this function fills in default_desc based on the
 * engine's default length encoding and returns default_desc.
 */
static const struct drm_i915_cmd_descriptor*
find_cmd(struct intel_engine_cs *engine,
	 u32 cmd_header,
	 const struct drm_i915_cmd_descriptor *desc,
	 struct drm_i915_cmd_descriptor *default_desc)
{
	u32 mask;

	if (((cmd_header ^ desc->cmd.value) & desc->cmd.mask) == 0)
		return desc;

	desc = find_cmd_in_table(engine, cmd_header);
	if (desc)
		return desc;

	mask = engine->get_cmd_length_mask(cmd_header);
	if (!mask)
		return NULL;

	default_desc->cmd.value = cmd_header;
	default_desc->cmd.mask = ~0u << MIN_OPCODE_SHIFT;
	default_desc->length.mask = mask;
	default_desc->flags = CMD_DESC_SKIP;
	return default_desc;
}

static const struct drm_i915_reg_descriptor *
__find_reg(const struct drm_i915_reg_descriptor *table, int count, u32 addr)
{
	int start = 0, end = count;
	while (start < end) {
		int mid = start + (end - start) / 2;
		int ret = addr - i915_mmio_reg_offset(table[mid].addr);
		if (ret < 0)
			end = mid;
		else if (ret > 0)
			start = mid + 1;
		else
			return &table[mid];
	}
	return NULL;
}

static const struct drm_i915_reg_descriptor *
find_reg(const struct intel_engine_cs *engine, u32 addr)
{
	const struct drm_i915_reg_table *table = engine->reg_tables;
	const struct drm_i915_reg_descriptor *reg = NULL;
	int count = engine->reg_table_count;

	for (; !reg && (count > 0); ++table, --count)
		reg = __find_reg(table->regs, table->num_regs, addr);

	return reg;
}

/* Returns a vmap'd pointer to dst_obj, which the caller must unmap */
static u32 *copy_batch(struct drm_i915_gem_object *dst_obj,
		       struct drm_i915_gem_object *src_obj,
		       u32 offset, u32 length)
{
	bool needs_clflush;
	void *dst, *src;
	int ret;

	dst = i915_gem_object_pin_map(dst_obj, I915_MAP_FORCE_WB);
	if (IS_ERR(dst))
		return dst;

	ret = i915_gem_object_pin_pages(src_obj);
	if (ret) {
		i915_gem_object_unpin_map(dst_obj);
		return ERR_PTR(ret);
	}

	needs_clflush =
		!(src_obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_READ);

	src = ERR_PTR(-ENODEV);
	if (needs_clflush && i915_has_memcpy_from_wc()) {
		src = i915_gem_object_pin_map(src_obj, I915_MAP_WC);
		if (!IS_ERR(src)) {
			i915_unaligned_memcpy_from_wc(dst,
						      src + offset,
						      length);
			i915_gem_object_unpin_map(src_obj);
		}
	}
	if (IS_ERR(src)) {
		void *ptr;
		int x, n;

		/*
		 * We can avoid clflushing partial cachelines before the write
		 * if we only every write full cache-lines. Since we know that
		 * both the source and destination are in multiples of
		 * PAGE_SIZE, we can simply round up to the next cacheline.
		 * We don't care about copying too much here as we only
		 * validate up to the end of the batch.
		 */
		if (!(dst_obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_READ))
			length = round_up(length,
					  boot_cpu_data.x86_clflush_size);

		ptr = dst;
		x = offset_in_page(offset);
		for (n = offset >> PAGE_SHIFT; length; n++) {
			int len = min_t(int, length, PAGE_SIZE - x);

			src = kmap_atomic(i915_gem_object_get_page(src_obj, n));
			if (needs_clflush)
				drm_clflush_virt_range(src + x, len);
			memcpy(ptr, src + x, len);
			kunmap_atomic(src);

			ptr += len;
			length -= len;
			x = 0;
		}
	}

	i915_gem_object_unpin_pages(src_obj);

	/* dst_obj is returned with vmap pinned */
	return dst;
}

static inline bool cmd_desc_is(const struct drm_i915_cmd_descriptor * const desc,
			       const u32 cmd)
{
	return desc->cmd.value == (cmd & desc->cmd.mask);
}

static bool check_cmd(const struct intel_engine_cs *engine,
		      const struct drm_i915_cmd_descriptor *desc,
		      const u32 *cmd, u32 length)
{
	if (desc->flags & CMD_DESC_SKIP)
		return true;

	if (desc->flags & CMD_DESC_REJECT) {
		DRM_DEBUG("CMD: Rejected command: 0x%08X\n", *cmd);
		return false;
	}

	if (desc->flags & CMD_DESC_REGISTER) {
		/*
		 * Get the distance between individual register offset
		 * fields if the command can perform more than one
		 * access at a time.
		 */
		const u32 step = desc->reg.step ? desc->reg.step : length;
		u32 offset;

		for (offset = desc->reg.offset; offset < length;
		     offset += step) {
			const u32 reg_addr = cmd[offset] & desc->reg.mask;
			const struct drm_i915_reg_descriptor *reg =
				find_reg(engine, reg_addr);

			if (!reg) {
				DRM_DEBUG("CMD: Rejected register 0x%08X in command: 0x%08X (%s)\n",
					  reg_addr, *cmd, engine->name);
				return false;
			}

			/*
			 * Check the value written to the register against the
			 * allowed mask/value pair given in the whitelist entry.
			 */
			if (reg->mask) {
				if (cmd_desc_is(desc, MI_LOAD_REGISTER_MEM)) {
					DRM_DEBUG("CMD: Rejected LRM to masked register 0x%08X\n",
						  reg_addr);
					return false;
				}

				if (cmd_desc_is(desc, MI_LOAD_REGISTER_REG)) {
					DRM_DEBUG("CMD: Rejected LRR to masked register 0x%08X\n",
						  reg_addr);
					return false;
				}

				if (cmd_desc_is(desc, MI_LOAD_REGISTER_IMM(1)) &&
				    (offset + 2 > length ||
				     (cmd[offset + 1] & reg->mask) != reg->value)) {
					DRM_DEBUG("CMD: Rejected LRI to masked register 0x%08X\n",
						  reg_addr);
					return false;
				}
			}
		}
	}

	if (desc->flags & CMD_DESC_BITMASK) {
		int i;

		for (i = 0; i < MAX_CMD_DESC_BITMASKS; i++) {
			u32 dword;

			if (desc->bits[i].mask == 0)
				break;

			if (desc->bits[i].condition_mask != 0) {
				u32 offset =
					desc->bits[i].condition_offset;
				u32 condition = cmd[offset] &
					desc->bits[i].condition_mask;

				if (condition == 0)
					continue;
			}

			if (desc->bits[i].offset >= length) {
				DRM_DEBUG("CMD: Rejected command 0x%08X, too short to check bitmask (%s)\n",
					  *cmd, engine->name);
				return false;
			}

			dword = cmd[desc->bits[i].offset] &
				desc->bits[i].mask;

			if (dword != desc->bits[i].expected) {
				DRM_DEBUG("CMD: Rejected command 0x%08X for bitmask 0x%08X (exp=0x%08X act=0x%08X) (%s)\n",
					  *cmd,
					  desc->bits[i].mask,
					  desc->bits[i].expected,
					  dword, engine->name);
				return false;
			}
		}
	}

	return true;
}

static int check_bbstart(u32 *cmd, u32 offset, u32 length,
			 u32 batch_length,
			 u64 batch_addr,
			 u64 shadow_addr,
			 const unsigned long *jump_whitelist)
{
	u64 jump_offset, jump_target;
	u32 target_cmd_offset, target_cmd_index;

	/* For igt compatibility on older platforms */
	if (!jump_whitelist) {
		DRM_DEBUG("CMD: Rejecting BB_START for ggtt based submission\n");
		return -EACCES;
	}

	if (length != 3) {
		DRM_DEBUG("CMD: Recursive BB_START with bad length(%u)\n",
			  length);
		return -EINVAL;
	}

	jump_target = *(u64 *)(cmd + 1);
	jump_offset = jump_target - batch_addr;

	/*
	 * Any underflow of jump_target is guaranteed to be outside the range
	 * of a u32, so >= test catches both too large and too small
	 */
	if (jump_offset >= batch_length) {
		DRM_DEBUG("CMD: BB_START to 0x%llx jumps out of BB\n",
			  jump_target);
		return -EINVAL;
	}

	/*
	 * This cannot overflow a u32 because we already checked jump_offset
	 * is within the BB, and the batch_length is a u32
	 */
	target_cmd_offset = lower_32_bits(jump_offset);
	target_cmd_index = target_cmd_offset / sizeof(u32);

	*(u64 *)(cmd + 1) = shadow_addr + target_cmd_offset;

	if (target_cmd_index == offset)
		return 0;

	if (IS_ERR(jump_whitelist))
		return PTR_ERR(jump_whitelist);

	if (!test_bit(target_cmd_index, jump_whitelist)) {
		DRM_DEBUG("CMD: BB_START to 0x%llx not a previously executed cmd\n",
			  jump_target);
		return -EINVAL;
	}

	return 0;
}

static unsigned long *alloc_whitelist(u32 batch_length)
{
	unsigned long *jmp;

	/*
	 * We expect batch_length to be less than 256KiB for known users,
	 * i.e. we need at most an 8KiB bitmap allocation which should be
	 * reasonably cheap due to kmalloc caches.
	 */

	/* Prefer to report transient allocation failure rather than hit oom */
	jmp = bitmap_zalloc(DIV_ROUND_UP(batch_length, sizeof(u32)),
			    GFP_KERNEL | __GFP_RETRY_MAYFAIL | __GFP_NOWARN);
	if (!jmp)
		return ERR_PTR(-ENOMEM);

	return jmp;
}

#define LENGTH_BIAS 2

static bool shadow_needs_clflush(struct drm_i915_gem_object *obj)
{
	return !(obj->cache_coherent & I915_BO_CACHE_COHERENT_FOR_WRITE);
}

/**
 * intel_engine_cmd_parser() - parse a batch buffer for privilege violations
 * @engine: the engine on which the batch is to execute
 * @batch: the batch buffer in question
 * @batch_offset: byte offset in the batch at which execution starts
 * @batch_length: length of the commands in batch_obj
 * @shadow: validated copy of the batch buffer in question
 * @trampoline: whether to emit a conditional trampoline at the end of the batch
 *
 * Parses the specified batch buffer looking for privilege violations as
 * described in the overview.
 *
 * Return: non-zero if the parser finds violations or otherwise fails; -EACCES
 * if the batch appears legal but should use hardware parsing
 */
int intel_engine_cmd_parser(struct intel_engine_cs *engine,
			    struct i915_vma *batch,
			    u32 batch_offset,
			    u32 batch_length,
			    struct i915_vma *shadow,
			    bool trampoline)
{
	u32 *cmd, *batch_end, offset = 0;
	struct drm_i915_cmd_descriptor default_desc = noop_desc;
	const struct drm_i915_cmd_descriptor *desc = &default_desc;
	unsigned long *jump_whitelist;
	u64 batch_addr, shadow_addr;
	int ret = 0;

	GEM_BUG_ON(!IS_ALIGNED(batch_offset, sizeof(*cmd)));
	GEM_BUG_ON(!IS_ALIGNED(batch_length, sizeof(*cmd)));
	GEM_BUG_ON(range_overflows_t(u64, batch_offset, batch_length,
				     batch->size));
	GEM_BUG_ON(!batch_length);

	cmd = copy_batch(shadow->obj, batch->obj, batch_offset, batch_length);
	if (IS_ERR(cmd)) {
		DRM_DEBUG("CMD: Failed to copy batch\n");
		return PTR_ERR(cmd);
	}

	jump_whitelist = NULL;
	if (!trampoline)
		/* Defer failure until attempted use */
		jump_whitelist = alloc_whitelist(batch_length);

	shadow_addr = gen8_canonical_addr(shadow->node.start);
	batch_addr = gen8_canonical_addr(batch->node.start + batch_offset);

	/*
	 * We use the batch length as size because the shadow object is as
	 * large or larger and copy_batch() will write MI_NOPs to the extra
	 * space. Parsing should be faster in some cases this way.
	 */
	batch_end = cmd + batch_length / sizeof(*batch_end);
	do {
		u32 length;

		if (*cmd == MI_BATCH_BUFFER_END)
			break;

		desc = find_cmd(engine, *cmd, desc, &default_desc);
		if (!desc) {
			DRM_DEBUG("CMD: Unrecognized command: 0x%08X\n", *cmd);
			ret = -EINVAL;
			break;
		}

		if (desc->flags & CMD_DESC_FIXED)
			length = desc->length.fixed;
		else
			length = (*cmd & desc->length.mask) + LENGTH_BIAS;

		if ((batch_end - cmd) < length) {
			DRM_DEBUG("CMD: Command length exceeds batch length: 0x%08X length=%u batchlen=%td\n",
				  *cmd,
				  length,
				  batch_end - cmd);
			ret = -EINVAL;
			break;
		}

		if (!check_cmd(engine, desc, cmd, length)) {
			ret = -EACCES;
			break;
		}

		if (cmd_desc_is(desc, MI_BATCH_BUFFER_START)) {
			ret = check_bbstart(cmd, offset, length, batch_length,
					    batch_addr, shadow_addr,
					    jump_whitelist);
			break;
		}

		if (!IS_ERR_OR_NULL(jump_whitelist))
			__set_bit(offset, jump_whitelist);

		cmd += length;
		offset += length;
		if  (cmd >= batch_end) {
			DRM_DEBUG("CMD: Got to the end of the buffer w/o a BBE cmd!\n");
			ret = -EINVAL;
			break;
		}
	} while (1);

	if (trampoline) {
		/*
		 * With the trampoline, the shadow is executed twice.
		 *
		 *   1 - starting at offset 0, in privileged mode
		 *   2 - starting at offset batch_len, as non-privileged
		 *
		 * Only if the batch is valid and safe to execute, do we
		 * allow the first privileged execution to proceed. If not,
		 * we terminate the first batch and use the second batchbuffer
		 * entry to chain to the original unsafe non-privileged batch,
		 * leaving it to the HW to validate.
		 */
		*batch_end = MI_BATCH_BUFFER_END;

		if (ret) {
			/* Batch unsafe to execute with privileges, cancel! */
			cmd = page_mask_bits(shadow->obj->mm.mapping);
			*cmd = MI_BATCH_BUFFER_END;

			/* If batch is unsafe but valid, jump to the original */
			if (ret == -EACCES) {
				unsigned int flags;

				flags = MI_BATCH_NON_SECURE_I965;
				if (IS_HASWELL(engine->i915))
					flags = MI_BATCH_NON_SECURE_HSW;

				GEM_BUG_ON(!IS_GEN_RANGE(engine->i915, 6, 7));
				__gen6_emit_bb_start(batch_end,
						     batch_addr,
						     flags);

				ret = 0; /* allow execution */
			}
		}

		if (shadow_needs_clflush(shadow->obj))
			drm_clflush_virt_range(batch_end, 8);
	}

	if (shadow_needs_clflush(shadow->obj)) {
		void *ptr = page_mask_bits(shadow->obj->mm.mapping);

		drm_clflush_virt_range(ptr, (void *)(cmd + 1) - ptr);
	}

	if (!IS_ERR_OR_NULL(jump_whitelist))
		kfree(jump_whitelist);
	i915_gem_object_unpin_map(shadow->obj);
	return ret;
}

/**
 * i915_cmd_parser_get_version() - get the cmd parser version number
 * @dev_priv: i915 device private
 *
 * The cmd parser maintains a simple increasing integer version number suitable
 * for passing to userspace clients to determine what operations are permitted.
 *
 * Return: the current version number of the cmd parser
 */
int i915_cmd_parser_get_version(struct drm_i915_private *dev_priv)
{
	struct intel_engine_cs *engine;
	bool active = false;

	/* If the command parser is not enabled, report 0 - unsupported */
	for_each_uabi_engine(engine, dev_priv) {
		if (intel_engine_using_cmd_parser(engine)) {
			active = true;
			break;
		}
	}
	if (!active)
		return 0;

	/*
	 * Command parser version history
	 *
	 * 1. Initial version. Checks batches and reports violations, but leaves
	 *    hardware parsing enabled (so does not allow new use cases).
	 * 2. Allow access to the MI_PREDICATE_SRC0 and
	 *    MI_PREDICATE_SRC1 registers.
	 * 3. Allow access to the GPGPU_THREADS_DISPATCHED register.
	 * 4. L3 atomic chicken bits of HSW_SCRATCH1 and HSW_ROW_CHICKEN3.
	 * 5. GPGPU dispatch compute indirect registers.
	 * 6. TIMESTAMP register and Haswell CS GPR registers
	 * 7. Allow MI_LOAD_REGISTER_REG between whitelisted registers.
	 * 8. Don't report cmd_check() failures as EINVAL errors to userspace;
	 *    rely on the HW to NOOP disallowed commands as it would without
	 *    the parser enabled.
	 * 9. Don't whitelist or handle oacontrol specially, as ownership
	 *    for oacontrol state is moving to i915-perf.
	 * 10. Support for Gen9 BCS Parsing
	 */
	return 10;
}