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
|
/*
* Mesa 3-D graphics library
* Version: 6.3
*
* Copyright (C) 1999-2004 Brian Paul All Rights Reserved.
*
* Permission is hereby granted, free of charge, to any person obtaining a
* copy of this software and associated documentation files (the "Software"),
* to deal in the Software without restriction, including without limitation
* the rights to use, copy, modify, merge, publish, distribute, 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 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
* BRIAN PAUL 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.
*/
/**
* \file t_arb_program.c
* Compile vertex programs to an intermediate representation.
* Execute vertex programs over a buffer of vertices.
* \author Keith Whitwell, Brian Paul
*/
#include "glheader.h"
#include "context.h"
#include "imports.h"
#include "macros.h"
#include "mtypes.h"
#include "arbprogparse.h"
#include "program.h"
#include "math/m_matrix.h"
#include "math/m_translate.h"
#include "t_context.h"
#include "t_pipeline.h"
#include "t_vp_build.h"
#include "t_vb_arbprogram.h"
#define DISASSEM 0
/*--------------------------------------------------------------------------- */
struct opcode_info {
GLuint nr_args;
const char *string;
void (*print)( union instruction , const struct opcode_info * );
};
struct compilation {
GLuint reg_active;
union instruction *csr;
};
#define ARB_VP_MACHINE(stage) ((struct arb_vp_machine *)(stage->privatePtr))
/**
* Set x to positive or negative infinity.
*
* XXX: FIXME - type punning.
*/
#if defined(USE_IEEE) || defined(_WIN32)
#define SET_POS_INFINITY(x) ( *((GLuint *) (void *)&x) = 0x7F800000 )
#define SET_NEG_INFINITY(x) ( *((GLuint *) (void *)&x) = 0xFF800000 )
#elif defined(VMS)
#define SET_POS_INFINITY(x) x = __MAXFLOAT
#define SET_NEG_INFINITY(x) x = -__MAXFLOAT
#define IS_INF_OR_NAN(t) ((t) == __MAXFLOAT)
#else
#define SET_POS_INFINITY(x) x = (GLfloat) HUGE_VAL
#define SET_NEG_INFINITY(x) x = (GLfloat) -HUGE_VAL
#endif
#define FREXPF(a,b) frexpf(a,b)
#define PUFF(x) ((x)[1] = (x)[2] = (x)[3] = (x)[0])
/* FIXME: more type punning (despite use of fi_type...)
*/
#define SET_FLOAT_BITS(x, bits) ((fi_type *) (void *) &(x))->i = bits
static GLfloat RoughApproxLog2(GLfloat t)
{
return LOG2(t);
}
static GLfloat RoughApproxPow2(GLfloat t)
{
GLfloat q;
#ifdef USE_IEEE
GLint ii = (GLint) t;
ii = (ii << 23) + 0x3f800000;
SET_FLOAT_BITS(q, ii);
q = *((GLfloat *) (void *)&ii);
#else
q = (GLfloat) pow(2.0, floor_t0);
#endif
return q;
}
static GLfloat RoughApproxPower(GLfloat x, GLfloat y)
{
#if 0
return RoughApproxPow2(y * RoughApproxLog2(x));
#else
return (GLfloat) _mesa_pow(x, y);
#endif
}
/**
* Perform a reduced swizzle:
*/
static void do_RSW( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.rsw.dst];
const GLfloat *arg0 = m->File[op.rsw.file0][op.rsw.idx0];
GLuint swz = op.rsw.swz;
GLuint neg = op.rsw.neg;
result[0] = arg0[GET_RSW(swz, 0)];
result[1] = arg0[GET_RSW(swz, 1)];
result[2] = arg0[GET_RSW(swz, 2)];
result[3] = arg0[GET_RSW(swz, 3)];
if (neg) {
if (neg & 0x1) result[0] = -result[0];
if (neg & 0x2) result[1] = -result[1];
if (neg & 0x4) result[2] = -result[2];
if (neg & 0x8) result[3] = -result[3];
}
}
/* Used to implement write masking. To make things easier for the sse
* generator I've gone back to a 1 argument version of this function
* (dst.msk = arg), rather than the semantically cleaner (dst = SEL
* arg0, arg1, msk)
*
* That means this is the only instruction which doesn't write a full
* 4 dwords out. This would make such a program harder to analyse,
* but it looks like analysis is going to take place on a higher level
* anyway.
*/
static void do_MSK( struct arb_vp_machine *m, union instruction op )
{
GLfloat *dst = m->File[0][op.msk.dst];
const GLfloat *arg = m->File[op.msk.file][op.msk.idx];
if (op.msk.mask & 0x1) dst[0] = arg[0];
if (op.msk.mask & 0x2) dst[1] = arg[1];
if (op.msk.mask & 0x4) dst[2] = arg[2];
if (op.msk.mask & 0x8) dst[3] = arg[3];
}
static void do_PRT( struct arb_vp_machine *m, union instruction op )
{
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
_mesa_printf("%d: %f %f %f %f\n", m->vtx_nr,
arg0[0], arg0[1], arg0[2], arg0[3]);
}
/**
* The traditional ALU and texturing instructions. All operate on
* internal registers and ignore write masks and swizzling issues.
*/
static void do_ABS( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
result[0] = (arg0[0] < 0.0) ? -arg0[0] : arg0[0];
result[1] = (arg0[1] < 0.0) ? -arg0[1] : arg0[1];
result[2] = (arg0[2] < 0.0) ? -arg0[2] : arg0[2];
result[3] = (arg0[3] < 0.0) ? -arg0[3] : arg0[3];
}
static void do_ADD( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
const GLfloat *arg1 = m->File[op.alu.file1][op.alu.idx1];
result[0] = arg0[0] + arg1[0];
result[1] = arg0[1] + arg1[1];
result[2] = arg0[2] + arg1[2];
result[3] = arg0[3] + arg1[3];
}
static void do_DP3( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
const GLfloat *arg1 = m->File[op.alu.file1][op.alu.idx1];
result[0] = (arg0[0] * arg1[0] +
arg0[1] * arg1[1] +
arg0[2] * arg1[2]);
PUFF(result);
}
static void do_DP4( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
const GLfloat *arg1 = m->File[op.alu.file1][op.alu.idx1];
result[0] = (arg0[0] * arg1[0] +
arg0[1] * arg1[1] +
arg0[2] * arg1[2] +
arg0[3] * arg1[3]);
PUFF(result);
}
static void do_DPH( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
const GLfloat *arg1 = m->File[op.alu.file1][op.alu.idx1];
result[0] = (arg0[0] * arg1[0] +
arg0[1] * arg1[1] +
arg0[2] * arg1[2] +
1.0 * arg1[3]);
PUFF(result);
}
static void do_DST( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
const GLfloat *arg1 = m->File[op.alu.file1][op.alu.idx1];
result[0] = 1.0F;
result[1] = arg0[1] * arg1[1];
result[2] = arg0[2];
result[3] = arg1[3];
}
static void do_EX2( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
result[0] = (GLfloat)RoughApproxPow2(arg0[0]);
PUFF(result);
}
static void do_EXP( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
GLfloat tmp = arg0[0];
GLfloat flr_tmp = FLOORF(tmp);
/* KW: nvvertexec has an optimized version of this which is pretty
* hard to understand/validate, but avoids the RoughApproxPow2.
*/
result[0] = (GLfloat) (1 << (int)flr_tmp);
result[1] = tmp - flr_tmp;
result[2] = RoughApproxPow2(tmp);
result[3] = 1.0F;
}
static void do_FLR( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
result[0] = FLOORF(arg0[0]);
result[1] = FLOORF(arg0[1]);
result[2] = FLOORF(arg0[2]);
result[3] = FLOORF(arg0[3]);
}
static void do_FRC( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
result[0] = arg0[0] - FLOORF(arg0[0]);
result[1] = arg0[1] - FLOORF(arg0[1]);
result[2] = arg0[2] - FLOORF(arg0[2]);
result[3] = arg0[3] - FLOORF(arg0[3]);
}
static void do_LG2( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
result[0] = RoughApproxLog2(arg0[0]);
PUFF(result);
}
static void do_LIT( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
const GLfloat epsilon = 1.0F / 256.0F; /* per NV spec */
GLfloat tmp[4];
tmp[0] = MAX2(arg0[0], 0.0F);
tmp[1] = MAX2(arg0[1], 0.0F);
tmp[3] = CLAMP(arg0[3], -(128.0F - epsilon), (128.0F - epsilon));
result[0] = 1.0;
result[1] = tmp[0];
result[2] = (tmp[0] > 0.0) ? RoughApproxPower(tmp[1], tmp[3]) : 0.0F;
result[3] = 1.0;
}
static void do_LOG( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
GLfloat tmp = FABSF(arg0[0]);
int exponent;
GLfloat mantissa = FREXPF(tmp, &exponent);
result[0] = (GLfloat) (exponent - 1);
result[1] = 2.0 * mantissa; /* map [.5, 1) -> [1, 2) */
result[2] = result[0] + LOG2(result[1]);
result[3] = 1.0;
}
static void do_MAX( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
const GLfloat *arg1 = m->File[op.alu.file1][op.alu.idx1];
result[0] = (arg0[0] > arg1[0]) ? arg0[0] : arg1[0];
result[1] = (arg0[1] > arg1[1]) ? arg0[1] : arg1[1];
result[2] = (arg0[2] > arg1[2]) ? arg0[2] : arg1[2];
result[3] = (arg0[3] > arg1[3]) ? arg0[3] : arg1[3];
}
static void do_MIN( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
const GLfloat *arg1 = m->File[op.alu.file1][op.alu.idx1];
result[0] = (arg0[0] < arg1[0]) ? arg0[0] : arg1[0];
result[1] = (arg0[1] < arg1[1]) ? arg0[1] : arg1[1];
result[2] = (arg0[2] < arg1[2]) ? arg0[2] : arg1[2];
result[3] = (arg0[3] < arg1[3]) ? arg0[3] : arg1[3];
}
static void do_MOV( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
result[0] = arg0[0];
result[1] = arg0[1];
result[2] = arg0[2];
result[3] = arg0[3];
}
static void do_MUL( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
const GLfloat *arg1 = m->File[op.alu.file1][op.alu.idx1];
result[0] = arg0[0] * arg1[0];
result[1] = arg0[1] * arg1[1];
result[2] = arg0[2] * arg1[2];
result[3] = arg0[3] * arg1[3];
}
static void do_POW( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
const GLfloat *arg1 = m->File[op.alu.file1][op.alu.idx1];
result[0] = (GLfloat)RoughApproxPower(arg0[0], arg1[0]);
PUFF(result);
}
static void do_REL( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
GLuint idx = (op.alu.idx0 + (GLint)m->File[0][REG_ADDR][0]) & (MAX_NV_VERTEX_PROGRAM_PARAMS-1);
const GLfloat *arg0 = m->File[op.alu.file0][idx];
result[0] = arg0[0];
result[1] = arg0[1];
result[2] = arg0[2];
result[3] = arg0[3];
}
static void do_RCP( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
result[0] = 1.0F / arg0[0];
PUFF(result);
}
static void do_RSQ( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
result[0] = INV_SQRTF(FABSF(arg0[0]));
PUFF(result);
}
static void do_SGE( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
const GLfloat *arg1 = m->File[op.alu.file1][op.alu.idx1];
result[0] = (arg0[0] >= arg1[0]) ? 1.0F : 0.0F;
result[1] = (arg0[1] >= arg1[1]) ? 1.0F : 0.0F;
result[2] = (arg0[2] >= arg1[2]) ? 1.0F : 0.0F;
result[3] = (arg0[3] >= arg1[3]) ? 1.0F : 0.0F;
}
static void do_SLT( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
const GLfloat *arg1 = m->File[op.alu.file1][op.alu.idx1];
result[0] = (arg0[0] < arg1[0]) ? 1.0F : 0.0F;
result[1] = (arg0[1] < arg1[1]) ? 1.0F : 0.0F;
result[2] = (arg0[2] < arg1[2]) ? 1.0F : 0.0F;
result[3] = (arg0[3] < arg1[3]) ? 1.0F : 0.0F;
}
static void do_SUB( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
const GLfloat *arg1 = m->File[op.alu.file1][op.alu.idx1];
result[0] = arg0[0] - arg1[0];
result[1] = arg0[1] - arg1[1];
result[2] = arg0[2] - arg1[2];
result[3] = arg0[3] - arg1[3];
}
static void do_XPD( struct arb_vp_machine *m, union instruction op )
{
GLfloat *result = m->File[0][op.alu.dst];
const GLfloat *arg0 = m->File[op.alu.file0][op.alu.idx0];
const GLfloat *arg1 = m->File[op.alu.file1][op.alu.idx1];
result[0] = arg0[1] * arg1[2] - arg0[2] * arg1[1];
result[1] = arg0[2] * arg1[0] - arg0[0] * arg1[2];
result[2] = arg0[0] * arg1[1] - arg0[1] * arg1[0];
}
static void do_NOP( struct arb_vp_machine *m, union instruction op )
{
}
/* Some useful debugging functions:
*/
static void print_mask( GLuint mask )
{
_mesa_printf(".");
if (mask&0x1) _mesa_printf("x");
if (mask&0x2) _mesa_printf("y");
if (mask&0x4) _mesa_printf("z");
if (mask&0x8) _mesa_printf("w");
}
static void print_reg( GLuint file, GLuint reg )
{
static const char *reg_file[] = {
"REG",
"LOCAL_PARAM",
"ENV_PARAM",
"STATE_VAR",
};
if (file == 0) {
if (reg == REG_RES)
_mesa_printf("RES");
else if (reg >= REG_ARG0 && reg <= REG_ARG1)
_mesa_printf("ARG%d", reg - REG_ARG0);
else if (reg >= REG_TMP0 && reg <= REG_TMP11)
_mesa_printf("TMP%d", reg - REG_TMP0);
else if (reg >= REG_IN0 && reg <= REG_IN31)
_mesa_printf("IN%d", reg - REG_IN0);
else if (reg >= REG_OUT0 && reg <= REG_OUT14)
_mesa_printf("OUT%d", reg - REG_OUT0);
else if (reg == REG_ADDR)
_mesa_printf("ADDR");
else if (reg == REG_ID)
_mesa_printf("ID");
else
_mesa_printf("REG%d", reg);
}
else
_mesa_printf("%s:%d", reg_file[file], reg);
}
static void print_RSW( union instruction op, const struct opcode_info *info )
{
GLuint swz = op.rsw.swz;
GLuint neg = op.rsw.neg;
GLuint i;
_mesa_printf("%s ", info->string);
print_reg(0, op.rsw.dst);
_mesa_printf(", ");
print_reg(op.rsw.file0, op.rsw.idx0);
_mesa_printf(".");
for (i = 0; i < 4; i++, swz >>= 2) {
const char *cswz = "xyzw";
if (neg & (1<<i))
_mesa_printf("-");
_mesa_printf("%c", cswz[swz&0x3]);
}
_mesa_printf("\n");
}
static void print_ALU( union instruction op, const struct opcode_info *info )
{
_mesa_printf("%s ", info->string);
print_reg(0, op.alu.dst);
_mesa_printf(", ");
print_reg(op.alu.file0, op.alu.idx0);
if (info->nr_args > 1) {
_mesa_printf(", ");
print_reg(op.alu.file1, op.alu.idx1);
}
_mesa_printf("\n");
}
static void print_MSK( union instruction op, const struct opcode_info *info )
{
_mesa_printf("%s ", info->string);
print_reg(0, op.msk.dst);
print_mask(op.msk.mask);
_mesa_printf(", ");
print_reg(op.msk.file, op.msk.idx);
_mesa_printf("\n");
}
static void print_NOP( union instruction op, const struct opcode_info *info )
{
}
#define NOP 0
#define ALU 1
#define SWZ 2
static const struct opcode_info opcode_info[] =
{
{ 1, "ABS", print_ALU },
{ 2, "ADD", print_ALU },
{ 1, "ARL", print_NOP },
{ 2, "DP3", print_ALU },
{ 2, "DP4", print_ALU },
{ 2, "DPH", print_ALU },
{ 2, "DST", print_ALU },
{ 0, "END", print_NOP },
{ 1, "EX2", print_ALU },
{ 1, "EXP", print_ALU },
{ 1, "FLR", print_ALU },
{ 1, "FRC", print_ALU },
{ 1, "LG2", print_ALU },
{ 1, "LIT", print_ALU },
{ 1, "LOG", print_ALU },
{ 3, "MAD", print_NOP },
{ 2, "MAX", print_ALU },
{ 2, "MIN", print_ALU },
{ 1, "MOV", print_ALU },
{ 2, "MUL", print_ALU },
{ 2, "POW", print_ALU },
{ 1, "PRT", print_ALU }, /* PRINT */
{ 1, "RCC", print_NOP },
{ 1, "RCP", print_ALU },
{ 1, "RSQ", print_ALU },
{ 2, "SGE", print_ALU },
{ 2, "SLT", print_ALU },
{ 2, "SUB", print_ALU },
{ 1, "SWZ", print_NOP },
{ 2, "XPD", print_ALU },
{ 1, "RSW", print_RSW },
{ 2, "MSK", print_MSK },
{ 1, "REL", print_ALU },
};
void _tnl_disassem_vba_insn( union instruction op )
{
const struct opcode_info *info = &opcode_info[op.alu.opcode];
info->print( op, info );
}
static void (* const opcode_func[])(struct arb_vp_machine *, union instruction) =
{
do_ABS,
do_ADD,
do_NOP,
do_DP3,
do_DP4,
do_DPH,
do_DST,
do_NOP,
do_EX2,
do_EXP,
do_FLR,
do_FRC,
do_LG2,
do_LIT,
do_LOG,
do_NOP,
do_MAX,
do_MIN,
do_MOV,
do_MUL,
do_POW,
do_PRT,
do_NOP,
do_RCP,
do_RSQ,
do_SGE,
do_SLT,
do_SUB,
do_RSW,
do_XPD,
do_RSW,
do_MSK,
do_REL,
};
static union instruction *cvp_next_instruction( struct compilation *cp )
{
union instruction *op = cp->csr++;
op->dword = 0;
return op;
}
static struct reg cvp_make_reg( GLuint file, GLuint idx )
{
struct reg reg;
reg.file = file;
reg.idx = idx;
return reg;
}
static struct reg cvp_emit_rel( struct compilation *cp,
struct reg reg,
struct reg tmpreg )
{
union instruction *op = cvp_next_instruction(cp);
op->alu.opcode = REL;
op->alu.file0 = reg.file;
op->alu.idx0 = reg.idx;
op->alu.dst = tmpreg.idx;
return tmpreg;
}
static struct reg cvp_load_reg( struct compilation *cp,
GLuint file,
GLuint index,
GLuint rel,
GLuint tmpidx )
{
struct reg tmpreg = cvp_make_reg(FILE_REG, tmpidx);
struct reg reg;
switch (file) {
case PROGRAM_TEMPORARY:
return cvp_make_reg(FILE_REG, REG_TMP0 + index);
case PROGRAM_INPUT:
return cvp_make_reg(FILE_REG, REG_IN0 + index);
case PROGRAM_OUTPUT:
return cvp_make_reg(FILE_REG, REG_OUT0 + index);
/* These two aren't populated by the parser?
*/
case PROGRAM_LOCAL_PARAM:
reg = cvp_make_reg(FILE_LOCAL_PARAM, index);
if (rel)
return cvp_emit_rel(cp, reg, tmpreg);
else
return reg;
case PROGRAM_ENV_PARAM:
reg = cvp_make_reg(FILE_ENV_PARAM, index);
if (rel)
return cvp_emit_rel(cp, reg, tmpreg);
else
return reg;
case PROGRAM_STATE_VAR:
reg = cvp_make_reg(FILE_STATE_PARAM, index);
if (rel)
return cvp_emit_rel(cp, reg, tmpreg);
else
return reg;
/* Invalid values:
*/
case PROGRAM_WRITE_ONLY:
case PROGRAM_ADDRESS:
default:
assert(0);
return tmpreg; /* can't happen */
}
}
static struct reg cvp_emit_arg( struct compilation *cp,
const struct vp_src_register *src,
GLuint arg )
{
struct reg reg = cvp_load_reg( cp, src->File, src->Index, src->RelAddr, arg );
union instruction rsw, noop;
/* Emit any necessary swizzling.
*/
rsw.dword = 0;
rsw.rsw.neg = src->Negate ? WRITEMASK_XYZW : 0;
rsw.rsw.swz = ((GET_SWZ(src->Swizzle, 0) << 0) |
(GET_SWZ(src->Swizzle, 1) << 2) |
(GET_SWZ(src->Swizzle, 2) << 4) |
(GET_SWZ(src->Swizzle, 3) << 6));
noop.dword = 0;
noop.rsw.neg = 0;
noop.rsw.swz = RSW_NOOP;
if (rsw.dword != noop.dword) {
union instruction *op = cvp_next_instruction(cp);
struct reg rsw_reg = cvp_make_reg(FILE_REG, REG_ARG0 + arg);
op->dword = rsw.dword;
op->rsw.opcode = RSW;
op->rsw.file0 = reg.file;
op->rsw.idx0 = reg.idx;
op->rsw.dst = rsw_reg.idx;
return rsw_reg;
}
else
return reg;
}
static GLuint cvp_choose_result( struct compilation *cp,
const struct vp_dst_register *dst,
union instruction *fixup )
{
GLuint mask = dst->WriteMask;
GLuint idx;
switch (dst->File) {
case PROGRAM_TEMPORARY:
idx = REG_TMP0 + dst->Index;
break;
case PROGRAM_OUTPUT:
idx = REG_OUT0 + dst->Index;
break;
default:
assert(0);
return REG_RES; /* can't happen */
}
/* Optimization: When writing (with a writemask) to an undefined
* value for the first time, the writemask may be ignored.
*/
if (mask != WRITEMASK_XYZW && (cp->reg_active & (1 << idx))) {
fixup->msk.opcode = MSK;
fixup->msk.dst = idx;
fixup->msk.file = FILE_REG;
fixup->msk.idx = REG_RES;
fixup->msk.mask = mask;
cp->reg_active |= 1 << idx;
return REG_RES;
}
else {
fixup->dword = 0;
cp->reg_active |= 1 << idx;
return idx;
}
}
static struct reg cvp_emit_rsw( struct compilation *cp,
GLuint dst,
struct reg src,
GLuint neg,
GLuint swz,
GLboolean force)
{
struct reg retval;
if (swz != RSW_NOOP || neg != 0) {
union instruction *op = cvp_next_instruction(cp);
op->rsw.opcode = RSW;
op->rsw.dst = dst;
op->rsw.file0 = src.file;
op->rsw.idx0 = src.idx;
op->rsw.neg = neg;
op->rsw.swz = swz;
retval.file = FILE_REG;
retval.idx = dst;
return retval;
}
else if (force) {
/* Oops. Degenerate case:
*/
union instruction *op = cvp_next_instruction(cp);
op->alu.opcode = VP_OPCODE_MOV;
op->alu.dst = dst;
op->alu.file0 = src.file;
op->alu.idx0 = src.idx;
retval.file = FILE_REG;
retval.idx = dst;
return retval;
}
else {
return src;
}
}
static void cvp_emit_inst( struct compilation *cp,
const struct vp_instruction *inst )
{
const struct opcode_info *info = &opcode_info[inst->Opcode];
union instruction *op;
union instruction fixup;
struct reg reg[3];
GLuint result, i;
assert(sizeof(*op) == sizeof(GLuint));
/* Need to handle SWZ, ARL specially.
*/
switch (inst->Opcode) {
/* Split into mul and add:
*/
case VP_OPCODE_MAD:
result = cvp_choose_result( cp, &inst->DstReg, &fixup );
for (i = 0; i < 3; i++)
reg[i] = cvp_emit_arg( cp, &inst->SrcReg[i], REG_ARG0+i );
op = cvp_next_instruction(cp);
op->alu.opcode = VP_OPCODE_MUL;
op->alu.file0 = reg[0].file;
op->alu.idx0 = reg[0].idx;
op->alu.file1 = reg[1].file;
op->alu.idx1 = reg[1].idx;
op->alu.dst = REG_ARG0;
op = cvp_next_instruction(cp);
op->alu.opcode = VP_OPCODE_ADD;
op->alu.file0 = FILE_REG;
op->alu.idx0 = REG_ARG0;
op->alu.file1 = reg[2].file;
op->alu.idx1 = reg[2].idx;
op->alu.dst = result;
break;
case VP_OPCODE_ARL:
reg[0] = cvp_emit_arg( cp, &inst->SrcReg[0], REG_ARG0 );
op = cvp_next_instruction(cp);
op->alu.opcode = VP_OPCODE_FLR;
op->alu.dst = REG_ADDR;
op->alu.file0 = reg[0].file;
op->alu.idx0 = reg[0].idx;
break;
case VP_OPCODE_SWZ: {
GLuint swz0 = 0, swz1 = 0;
GLuint neg0 = 0, neg1 = 0;
GLuint mask = 0;
/* Translate 3-bit-per-element swizzle into two 2-bit swizzles,
* one from the source register the other from a constant
* {0,0,0,1}.
*/
for (i = 0; i < 4; i++) {
GLuint swzelt = GET_SWZ(inst->SrcReg[0].Swizzle, i);
if (swzelt >= SWIZZLE_ZERO) {
neg0 |= inst->SrcReg[0].Negate & (1<<i);
if (swzelt == SWIZZLE_ONE)
swz0 |= SWIZZLE_W << (i*2);
else if (i < SWIZZLE_W)
swz0 |= i << (i*2);
}
else {
mask |= 1<<i;
neg1 |= inst->SrcReg[0].Negate & (1<<i);
swz1 |= swzelt << (i*2);
}
}
result = cvp_choose_result( cp, &inst->DstReg, &fixup );
reg[0].file = FILE_REG;
reg[0].idx = REG_ID;
reg[1] = cvp_emit_arg( cp, &inst->SrcReg[0], REG_ARG0 );
if (mask == WRITEMASK_XYZW) {
cvp_emit_rsw(cp, result, reg[0], neg0, swz0, GL_TRUE);
}
else if (mask == 0) {
cvp_emit_rsw(cp, result, reg[1], neg1, swz1, GL_TRUE);
}
else {
cvp_emit_rsw(cp, result, reg[0], neg0, swz0, GL_TRUE);
reg[1] = cvp_emit_rsw(cp, REG_ARG0, reg[1], neg1, swz1, GL_FALSE);
op = cvp_next_instruction(cp);
op->msk.opcode = MSK;
op->msk.dst = result;
op->msk.file = reg[1].file;
op->msk.idx = reg[1].idx;
op->msk.mask = mask;
}
if (result == REG_RES) {
op = cvp_next_instruction(cp);
op->dword = fixup.dword;
}
break;
}
case VP_OPCODE_PRINT:
case VP_OPCODE_END:
break;
default:
result = cvp_choose_result( cp, &inst->DstReg, &fixup );
for (i = 0; i < info->nr_args; i++)
reg[i] = cvp_emit_arg( cp, &inst->SrcReg[i], REG_ARG0 + i );
op = cvp_next_instruction(cp);
op->alu.opcode = inst->Opcode;
op->alu.file0 = reg[0].file;
op->alu.idx0 = reg[0].idx;
op->alu.file1 = reg[1].file;
op->alu.idx1 = reg[1].idx;
op->alu.dst = result;
if (result == REG_RES) {
op = cvp_next_instruction(cp);
op->dword = fixup.dword;
}
break;
}
}
static void free_tnl_data( struct vertex_program *program )
{
struct tnl_compiled_program *p = program->TnlData;
if (p->compiled_func) free((void *)p->compiled_func);
free(p);
program->TnlData = NULL;
}
static void compile_vertex_program( struct vertex_program *program,
GLboolean try_codegen )
{
struct compilation cp;
struct tnl_compiled_program *p = CALLOC_STRUCT(tnl_compiled_program);
GLuint i;
_mesa_printf("%s\n", __FUNCTION__);
if (program->TnlData)
free_tnl_data( program );
program->TnlData = p;
/* Initialize cp. Note that ctx and VB aren't used in compilation
* so we don't have to worry about statechanges:
*/
memset(&cp, 0, sizeof(cp));
cp.csr = p->instructions;
/* Compile instructions:
*/
for (i = 0; i < program->Base.NumInstructions; i++) {
cvp_emit_inst(&cp, &program->Instructions[i]);
}
/* Finish up:
*/
p->nr_instructions = cp.csr - p->instructions;
/* Print/disassemble:
*/
if (DISASSEM) {
for (i = 0; i < p->nr_instructions; i++) {
_tnl_disassem_vba_insn(p->instructions[i]);
}
_mesa_printf("\n\n");
}
#ifdef USE_SSE_ASM
if (try_codegen)
_tnl_sse_codegen_vertex_program(p);
#endif
}
/* ----------------------------------------------------------------------
* Execution
*/
static void userclip( GLcontext *ctx,
GLvector4f *clip,
GLubyte *clipmask,
GLubyte *clipormask,
GLubyte *clipandmask )
{
GLuint p;
for (p = 0; p < ctx->Const.MaxClipPlanes; p++) {
if (ctx->Transform.ClipPlanesEnabled & (1 << p)) {
GLuint nr, i;
const GLfloat a = ctx->Transform._ClipUserPlane[p][0];
const GLfloat b = ctx->Transform._ClipUserPlane[p][1];
const GLfloat c = ctx->Transform._ClipUserPlane[p][2];
const GLfloat d = ctx->Transform._ClipUserPlane[p][3];
GLfloat *coord = (GLfloat *)clip->data;
GLuint stride = clip->stride;
GLuint count = clip->count;
for (nr = 0, i = 0 ; i < count ; i++) {
GLfloat dp = (coord[0] * a +
coord[1] * b +
coord[2] * c +
coord[3] * d);
if (dp < 0) {
nr++;
clipmask[i] |= CLIP_USER_BIT;
}
STRIDE_F(coord, stride);
}
if (nr > 0) {
*clipormask |= CLIP_USER_BIT;
if (nr == count) {
*clipandmask |= CLIP_USER_BIT;
return;
}
}
}
}
}
static GLboolean do_ndc_cliptest( struct arb_vp_machine *m )
{
GLcontext *ctx = m->ctx;
TNLcontext *tnl = TNL_CONTEXT(ctx);
struct vertex_buffer *VB = m->VB;
/* Cliptest and perspective divide. Clip functions must clear
* the clipmask.
*/
m->ormask = 0;
m->andmask = CLIP_ALL_BITS;
if (tnl->NeedNdcCoords) {
VB->NdcPtr =
_mesa_clip_tab[VB->ClipPtr->size]( VB->ClipPtr,
&m->ndcCoords,
m->clipmask,
&m->ormask,
&m->andmask );
}
else {
VB->NdcPtr = NULL;
_mesa_clip_np_tab[VB->ClipPtr->size]( VB->ClipPtr,
NULL,
m->clipmask,
&m->ormask,
&m->andmask );
}
if (m->andmask) {
/* All vertices are outside the frustum */
return GL_FALSE;
}
/* Test userclip planes. This contributes to VB->ClipMask.
*/
if (ctx->Transform.ClipPlanesEnabled && !ctx->VertexProgram._Enabled) {
userclip( ctx,
VB->ClipPtr,
m->clipmask,
&m->ormask,
&m->andmask );
if (m->andmask) {
return GL_FALSE;
}
}
VB->ClipAndMask = m->andmask;
VB->ClipOrMask = m->ormask;
VB->ClipMask = m->clipmask;
return GL_TRUE;
}
static INLINE void call_func( struct tnl_compiled_program *p,
struct arb_vp_machine *m )
{
p->compiled_func(m);
}
/**
* Execute the given vertex program.
*
* TODO: Integrate the t_vertex.c code here, to build machine vertices
* directly at this point.
*
* TODO: Eliminate the VB struct entirely and just use
* struct arb_vertex_machine.
*/
static GLboolean
run_arb_vertex_program(GLcontext *ctx, struct tnl_pipeline_stage *stage)
{
struct vertex_program *program = (ctx->VertexProgram._Enabled ?
ctx->VertexProgram.Current :
ctx->_TnlProgram);
struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
struct arb_vp_machine *m = ARB_VP_MACHINE(stage);
struct tnl_compiled_program *p;
GLuint i, j, outputs;
if (!program || program->IsNVProgram)
return GL_TRUE;
if (program->Parameters) {
_mesa_load_state_parameters(ctx, program->Parameters);
}
p = (struct tnl_compiled_program *)program->TnlData;
assert(p);
/* Initialize regs where necessary:
*/
ASSIGN_4V(m->File[0][REG_ID], 0, 0, 0, 1);
m->nr_inputs = m->nr_outputs = 0;
for (i = 0; i < _TNL_ATTRIB_MAX; i++) {
if (program->InputsRead & (1<<i)) {
GLuint j = m->nr_inputs++;
m->input[j].idx = i;
m->input[j].data = (GLfloat *)m->VB->AttribPtr[i]->data;
m->input[j].stride = m->VB->AttribPtr[i]->stride;
m->input[j].size = m->VB->AttribPtr[i]->size;
ASSIGN_4V(m->File[0][REG_IN0 + i], 0, 0, 0, 1);
}
}
for (i = 0; i < 15; i++) {
if (program->OutputsWritten & (1<<i)) {
GLuint j = m->nr_outputs++;
m->output[j].idx = i;
m->output[j].data = (GLfloat *)m->attribs[i].data;
}
}
/* Run the actual program:
*/
for (m->vtx_nr = 0; m->vtx_nr < VB->Count; m->vtx_nr++) {
for (j = 0; j < m->nr_inputs; j++) {
GLuint idx = REG_IN0 + m->input[j].idx;
switch (m->input[j].size) {
case 4: m->File[0][idx][3] = m->input[j].data[3];
case 3: m->File[0][idx][2] = m->input[j].data[2];
case 2: m->File[0][idx][1] = m->input[j].data[1];
case 1: m->File[0][idx][0] = m->input[j].data[0];
}
STRIDE_F(m->input[j].data, m->input[j].stride);
}
if (p->compiled_func) {
call_func( p, m );
}
else {
for (j = 0; j < p->nr_instructions; j++) {
union instruction inst = p->instructions[j];
opcode_func[inst.alu.opcode]( m, inst );
}
}
for (j = 0; j < m->nr_outputs; j++) {
GLuint idx = REG_OUT0 + m->output[j].idx;
m->output[j].data[0] = m->File[0][idx][0];
m->output[j].data[1] = m->File[0][idx][1];
m->output[j].data[2] = m->File[0][idx][2];
m->output[j].data[3] = m->File[0][idx][3];
m->output[j].data += 4;
}
}
/* Setup the VB pointers so that the next pipeline stages get
* their data from the right place (the program output arrays).
*
* TODO: 1) Have tnl use these RESULT values for outputs rather
* than trying to shoe-horn inputs and outputs into one set of
* values.
*
* TODO: 2) Integrate t_vertex.c so that we just go straight ahead
* and build machine vertices here.
*/
VB->ClipPtr = &m->attribs[VERT_RESULT_HPOS];
VB->ClipPtr->count = VB->Count;
outputs = program->OutputsWritten;
if (outputs & (1<<VERT_RESULT_COL0)) {
VB->ColorPtr[0] = &m->attribs[VERT_RESULT_COL0];
VB->AttribPtr[VERT_ATTRIB_COLOR0] = VB->ColorPtr[0];
}
if (outputs & (1<<VERT_RESULT_BFC0)) {
VB->ColorPtr[1] = &m->attribs[VERT_RESULT_BFC0];
}
if (outputs & (1<<VERT_RESULT_COL1)) {
VB->SecondaryColorPtr[0] = &m->attribs[VERT_RESULT_COL1];
VB->AttribPtr[VERT_ATTRIB_COLOR1] = VB->SecondaryColorPtr[0];
}
if (outputs & (1<<VERT_RESULT_BFC1)) {
VB->SecondaryColorPtr[1] = &m->attribs[VERT_RESULT_BFC1];
}
if (outputs & (1<<VERT_RESULT_FOGC)) {
VB->FogCoordPtr = &m->attribs[VERT_RESULT_FOGC];
VB->AttribPtr[VERT_ATTRIB_FOG] = VB->FogCoordPtr;
}
if (outputs & (1<<VERT_RESULT_PSIZ)) {
VB->PointSizePtr = &m->attribs[VERT_RESULT_PSIZ];
VB->AttribPtr[_TNL_ATTRIB_POINTSIZE] = &m->attribs[VERT_RESULT_PSIZ];
}
for (i = 0; i < ctx->Const.MaxTextureUnits; i++) {
if (outputs & (1<<(VERT_RESULT_TEX0+i))) {
VB->TexCoordPtr[i] = &m->attribs[VERT_RESULT_TEX0 + i];
VB->AttribPtr[VERT_ATTRIB_TEX0+i] = VB->TexCoordPtr[i];
}
}
#if 0
for (i = 0; i < VB->Count; i++) {
printf("Out %d: %f %f %f %f %f %f %f %f\n", i,
VEC_ELT(VB->ClipPtr, GLfloat, i)[0],
VEC_ELT(VB->ClipPtr, GLfloat, i)[1],
VEC_ELT(VB->ClipPtr, GLfloat, i)[2],
VEC_ELT(VB->ClipPtr, GLfloat, i)[3],
VEC_ELT(VB->TexCoordPtr[0], GLfloat, i)[0],
VEC_ELT(VB->TexCoordPtr[0], GLfloat, i)[1],
VEC_ELT(VB->TexCoordPtr[0], GLfloat, i)[2],
VEC_ELT(VB->TexCoordPtr[0], GLfloat, i)[3]);
}
#endif
/* Perform NDC and cliptest operations:
*/
return do_ndc_cliptest(m);
}
static void
validate_vertex_program( GLcontext *ctx, struct tnl_pipeline_stage *stage )
{
struct arb_vp_machine *m = ARB_VP_MACHINE(stage);
struct vertex_program *program =
(ctx->VertexProgram._Enabled ? ctx->VertexProgram.Current : 0);
if (!program && ctx->_MaintainTnlProgram) {
program = ctx->_TnlProgram;
}
if (program) {
if (!program->TnlData)
compile_vertex_program( program, m->try_codegen );
/* Grab the state GL state and put into registers:
*/
m->File[FILE_LOCAL_PARAM] = program->Base.LocalParams;
m->File[FILE_ENV_PARAM] = ctx->VertexProgram.Parameters;
m->File[FILE_STATE_PARAM] = program->Parameters->ParameterValues;
}
}
/**
* Called the first time stage->run is called. In effect, don't
* allocate data until the first time the stage is run.
*/
static GLboolean init_vertex_program( GLcontext *ctx,
struct tnl_pipeline_stage *stage )
{
TNLcontext *tnl = TNL_CONTEXT(ctx);
struct vertex_buffer *VB = &(tnl->vb);
struct arb_vp_machine *m;
const GLuint size = VB->Size;
GLuint i;
stage->privatePtr = MALLOC(sizeof(*m));
m = ARB_VP_MACHINE(stage);
if (!m)
return GL_FALSE;
/* arb_vertex_machine struct should subsume the VB:
*/
m->VB = VB;
m->ctx = ctx;
m->File[0] = ALIGN_MALLOC(REG_MAX * sizeof(GLfloat) * 4, 16);
if (_mesa_getenv("MESA_EXPERIMENTAL"))
m->try_codegen = 1;
/* Allocate arrays of vertex output values */
for (i = 0; i < VERT_RESULT_MAX; i++) {
_mesa_vector4f_alloc( &m->attribs[i], 0, size, 32 );
m->attribs[i].size = 4;
}
/* a few other misc allocations */
_mesa_vector4f_alloc( &m->ndcCoords, 0, size, 32 );
m->clipmask = (GLubyte *) ALIGN_MALLOC(sizeof(GLubyte)*size, 32 );
if (ctx->_MaintainTnlProgram)
_mesa_allow_light_in_model( ctx, GL_FALSE );
return GL_TRUE;
}
/**
* Destructor for this pipeline stage.
*/
static void dtr( struct tnl_pipeline_stage *stage )
{
struct arb_vp_machine *m = ARB_VP_MACHINE(stage);
if (m) {
GLuint i;
/* free the vertex program result arrays */
for (i = 0; i < VERT_RESULT_MAX; i++)
_mesa_vector4f_free( &m->attribs[i] );
/* free misc arrays */
_mesa_vector4f_free( &m->ndcCoords );
ALIGN_FREE( m->clipmask );
ALIGN_FREE( m->File[0] );
FREE( m );
stage->privatePtr = NULL;
}
}
/**
* Public description of this pipeline stage.
*/
const struct tnl_pipeline_stage _tnl_arb_vertex_program_stage =
{
"vertex-program",
NULL, /* private_data */
init_vertex_program, /* create */
dtr, /* destroy */
validate_vertex_program, /* validate */
run_arb_vertex_program /* run */
};
|