summaryrefslogtreecommitdiff
path: root/src/mesa/main/ffvertex_prog.c
blob: 5f3def257de1aee1339420b18c27818759b14975 (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
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
/**************************************************************************
 * 
 * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas.
 * 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, sub license, 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 NON-INFRINGEMENT.
 * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS 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 ffvertex_prog.
 *
 * Create a vertex program to execute the current fixed function T&L pipeline.
 * \author Keith Whitwell
 */


#include "main/glheader.h"
#include "main/mtypes.h"
#include "main/macros.h"
#include "main/enums.h"
#include "main/ffvertex_prog.h"
#include "shader/program.h"
#include "shader/prog_cache.h"
#include "shader/prog_instruction.h"
#include "shader/prog_parameter.h"
#include "shader/prog_print.h"
#include "shader/prog_statevars.h"


struct state_key {
   unsigned light_global_enabled:1;
   unsigned light_local_viewer:1;
   unsigned light_twoside:1;
   unsigned light_color_material:1;
   unsigned light_color_material_mask:12;
   unsigned light_material_mask:12;
   unsigned material_shininess_is_zero:1;

   unsigned need_eye_coords:1;
   unsigned normalize:1;
   unsigned rescale_normals:1;
   unsigned fog_source_is_depth:1;
   unsigned tnl_do_vertex_fog:1;
   unsigned separate_specular:1;
   unsigned fog_mode:2;
   unsigned point_attenuated:1;
   unsigned point_array:1;
   unsigned texture_enabled_global:1;
   unsigned fragprog_inputs_read:12;

   struct {
      unsigned light_enabled:1;
      unsigned light_eyepos3_is_zero:1;
      unsigned light_spotcutoff_is_180:1;
      unsigned light_attenuated:1;      
      unsigned texunit_really_enabled:1;
      unsigned texmat_enabled:1;
      unsigned texgen_enabled:4;
      unsigned texgen_mode0:4;
      unsigned texgen_mode1:4;
      unsigned texgen_mode2:4;
      unsigned texgen_mode3:4;
   } unit[8];
};



#define FOG_NONE   0
#define FOG_LINEAR 1
#define FOG_EXP    2
#define FOG_EXP2   3

static GLuint translate_fog_mode( GLenum mode )
{
   switch (mode) {
   case GL_LINEAR: return FOG_LINEAR;
   case GL_EXP: return FOG_EXP;
   case GL_EXP2: return FOG_EXP2;
   default: return FOG_NONE;
   }
}

#define TXG_NONE           0
#define TXG_OBJ_LINEAR     1
#define TXG_EYE_LINEAR     2
#define TXG_SPHERE_MAP     3
#define TXG_REFLECTION_MAP 4
#define TXG_NORMAL_MAP     5

static GLuint translate_texgen( GLboolean enabled, GLenum mode )
{
   if (!enabled)
      return TXG_NONE;

   switch (mode) {
   case GL_OBJECT_LINEAR: return TXG_OBJ_LINEAR;
   case GL_EYE_LINEAR: return TXG_EYE_LINEAR;
   case GL_SPHERE_MAP: return TXG_SPHERE_MAP;
   case GL_REFLECTION_MAP_NV: return TXG_REFLECTION_MAP;
   case GL_NORMAL_MAP_NV: return TXG_NORMAL_MAP;
   default: return TXG_NONE;
   }
}


/**
 * Returns bitmask of flags indicating which materials are set per-vertex
 * in the current VB.
 * XXX get these from the VBO...
 */
static GLbitfield
tnl_get_per_vertex_materials(GLcontext *ctx)
{
   GLbitfield mask = 0x0;
#if 0
   TNLcontext *tnl = TNL_CONTEXT(ctx);
   struct vertex_buffer *VB = &tnl->vb;
   GLuint i;

   for (i = _TNL_FIRST_MAT; i <= _TNL_LAST_MAT; i++) 
      if (VB->AttribPtr[i] && VB->AttribPtr[i]->stride) 
         mask |= 1 << (i - _TNL_FIRST_MAT);
#endif
   return mask;
}

/**
 * Should fog be computed per-vertex?
 */
static GLboolean
tnl_get_per_vertex_fog(GLcontext *ctx)
{
#if 0
   TNLcontext *tnl = TNL_CONTEXT(ctx);
   return tnl->_DoVertexFog;
#else
   return GL_FALSE;
#endif
}

static GLboolean check_active_shininess( GLcontext *ctx,
                                         const struct state_key *key,
                                         GLuint side )
{
   GLuint bit = 1 << (MAT_ATTRIB_FRONT_SHININESS + side);

   if (key->light_color_material_mask & bit)
      return GL_TRUE;

   if (key->light_material_mask & bit)
      return GL_TRUE;

   if (ctx->Light.Material.Attrib[MAT_ATTRIB_FRONT_SHININESS + side][0] != 0.0F)
      return GL_TRUE;

   return GL_FALSE;
}
     



static struct state_key *make_state_key( GLcontext *ctx )
{
   const struct gl_fragment_program *fp;
   struct state_key *key = CALLOC_STRUCT(state_key);
   GLuint i;

   fp = ctx->FragmentProgram._Current;

   /* This now relies on texenvprogram.c being active:
    */
   assert(fp);

   key->need_eye_coords = ctx->_NeedEyeCoords;

   key->fragprog_inputs_read = fp->Base.InputsRead;

   if (ctx->RenderMode == GL_FEEDBACK) {
      /* make sure the vertprog emits color and tex0 */
      key->fragprog_inputs_read |= (FRAG_BIT_COL0 | FRAG_BIT_TEX0);
   }

   key->separate_specular = (ctx->Light.Model.ColorControl ==
			     GL_SEPARATE_SPECULAR_COLOR);

   if (ctx->Light.Enabled) {
      key->light_global_enabled = 1;

      if (ctx->Light.Model.LocalViewer)
	 key->light_local_viewer = 1;

      if (ctx->Light.Model.TwoSide)
	 key->light_twoside = 1;

      if (ctx->Light.ColorMaterialEnabled) {
	 key->light_color_material = 1;
	 key->light_color_material_mask = ctx->Light.ColorMaterialBitmask;
      }

      key->light_material_mask = tnl_get_per_vertex_materials(ctx);

      for (i = 0; i < MAX_LIGHTS; i++) {
	 struct gl_light *light = &ctx->Light.Light[i];

	 if (light->Enabled) {
	    key->unit[i].light_enabled = 1;

	    if (light->EyePosition[3] == 0.0)
	       key->unit[i].light_eyepos3_is_zero = 1;
	    
	    if (light->SpotCutoff == 180.0)
	       key->unit[i].light_spotcutoff_is_180 = 1;

	    if (light->ConstantAttenuation != 1.0 ||
		light->LinearAttenuation != 0.0 ||
		light->QuadraticAttenuation != 0.0)
	       key->unit[i].light_attenuated = 1;
	 }
      }

      if (check_active_shininess(ctx, key, 0)) {
         key->material_shininess_is_zero = 0;
      }
      else if (key->light_twoside &&
               check_active_shininess(ctx, key, 1)) {
         key->material_shininess_is_zero = 0;
      }
      else {
         key->material_shininess_is_zero = 1;
      }
   }

   if (ctx->Transform.Normalize)
      key->normalize = 1;

   if (ctx->Transform.RescaleNormals)
      key->rescale_normals = 1;

   key->fog_mode = translate_fog_mode(fp->FogOption);
   
   if (ctx->Fog.FogCoordinateSource == GL_FRAGMENT_DEPTH_EXT)
      key->fog_source_is_depth = 1;
   
   key->tnl_do_vertex_fog = tnl_get_per_vertex_fog(ctx);

   if (ctx->Point._Attenuated)
      key->point_attenuated = 1;

#if FEATURE_point_size_array
   if (ctx->Array.ArrayObj->PointSize.Enabled)
      key->point_array = 1;
#endif

   if (ctx->Texture._TexGenEnabled ||
       ctx->Texture._TexMatEnabled ||
       ctx->Texture._EnabledUnits)
      key->texture_enabled_global = 1;
      
   for (i = 0; i < MAX_TEXTURE_UNITS; i++) {
      struct gl_texture_unit *texUnit = &ctx->Texture.Unit[i];

      if (texUnit->_ReallyEnabled)
	 key->unit[i].texunit_really_enabled = 1;

      if (ctx->Texture._TexMatEnabled & ENABLE_TEXMAT(i))      
	 key->unit[i].texmat_enabled = 1;
      
      if (texUnit->TexGenEnabled) {
	 key->unit[i].texgen_enabled = 1;
      
	 key->unit[i].texgen_mode0 = 
	    translate_texgen( texUnit->TexGenEnabled & (1<<0),
			      texUnit->GenModeS );
	 key->unit[i].texgen_mode1 = 
	    translate_texgen( texUnit->TexGenEnabled & (1<<1),
			      texUnit->GenModeT );
	 key->unit[i].texgen_mode2 = 
	    translate_texgen( texUnit->TexGenEnabled & (1<<2),
			      texUnit->GenModeR );
	 key->unit[i].texgen_mode3 = 
	    translate_texgen( texUnit->TexGenEnabled & (1<<3),
			      texUnit->GenModeQ );
      }
   }
   
   return key;
}


   
/* Very useful debugging tool - produces annotated listing of
 * generated program with line/function references for each
 * instruction back into this file:
 */
#define DISASSEM 0

/* Should be tunable by the driver - do we want to do matrix
 * multiplications with DP4's or with MUL/MAD's?  SSE works better
 * with the latter, drivers may differ.
 */
#define PREFER_DP4 0

#define MAX_INSN 256

/* Use uregs to represent registers internally, translate to Mesa's
 * expected formats on emit.  
 *
 * NOTE: These are passed by value extensively in this file rather
 * than as usual by pointer reference.  If this disturbs you, try
 * remembering they are just 32bits in size.
 *
 * GCC is smart enough to deal with these dword-sized structures in
 * much the same way as if I had defined them as dwords and was using
 * macros to access and set the fields.  This is much nicer and easier
 * to evolve.
 */
struct ureg {
   GLuint file:4;
   GLint idx:8;      /* relative addressing may be negative */
   GLuint negate:1;
   GLuint swz:12;
   GLuint pad:7;
};


struct tnl_program {
   const struct state_key *state;
   struct gl_vertex_program *program;
   
   GLuint temp_in_use;
   GLuint temp_reserved;
   
   struct ureg eye_position;
   struct ureg eye_position_z;
   struct ureg eye_position_normalized;
   struct ureg transformed_normal;
   struct ureg identity;

   GLuint materials;
   GLuint color_materials;
};


static const struct ureg undef = { 
   PROGRAM_UNDEFINED,
   ~0,
   0,
   0,
   0
};

/* Local shorthand:
 */
#define X    SWIZZLE_X
#define Y    SWIZZLE_Y
#define Z    SWIZZLE_Z
#define W    SWIZZLE_W


/* Construct a ureg:
 */
static struct ureg make_ureg(GLuint file, GLint idx)
{
   struct ureg reg;
   reg.file = file;
   reg.idx = idx;
   reg.negate = 0;
   reg.swz = SWIZZLE_NOOP;
   reg.pad = 0;
   return reg;
}



static struct ureg negate( struct ureg reg )
{
   reg.negate ^= 1;
   return reg;
} 


static struct ureg swizzle( struct ureg reg, int x, int y, int z, int w )
{
   reg.swz = MAKE_SWIZZLE4(GET_SWZ(reg.swz, x),
			   GET_SWZ(reg.swz, y),
			   GET_SWZ(reg.swz, z),
			   GET_SWZ(reg.swz, w));

   return reg;
}

static struct ureg swizzle1( struct ureg reg, int x )
{
   return swizzle(reg, x, x, x, x);
}

static struct ureg get_temp( struct tnl_program *p )
{
   int bit = _mesa_ffs( ~p->temp_in_use );
   if (!bit) {
      _mesa_problem(NULL, "%s: out of temporaries\n", __FILE__);
      _mesa_exit(1);
   }

   if ((GLuint) bit > p->program->Base.NumTemporaries)
      p->program->Base.NumTemporaries = bit;

   p->temp_in_use |= 1<<(bit-1);
   return make_ureg(PROGRAM_TEMPORARY, bit-1);
}

static struct ureg reserve_temp( struct tnl_program *p )
{
   struct ureg temp = get_temp( p );
   p->temp_reserved |= 1<<temp.idx;
   return temp;
}

static void release_temp( struct tnl_program *p, struct ureg reg )
{
   if (reg.file == PROGRAM_TEMPORARY) {
      p->temp_in_use &= ~(1<<reg.idx);
      p->temp_in_use |= p->temp_reserved; /* can't release reserved temps */
   }
}

static void release_temps( struct tnl_program *p )
{
   p->temp_in_use = p->temp_reserved;
}



/**
 * \param input  one of VERT_ATTRIB_x tokens.
 */
static struct ureg register_input( struct tnl_program *p, GLuint input )
{
   p->program->Base.InputsRead |= (1<<input);
   return make_ureg(PROGRAM_INPUT, input);
}

/**
 * \param input  one of VERT_RESULT_x tokens.
 */
static struct ureg register_output( struct tnl_program *p, GLuint output )
{
   p->program->Base.OutputsWritten |= (1<<output);
   return make_ureg(PROGRAM_OUTPUT, output);
}

static struct ureg register_const4f( struct tnl_program *p, 
			      GLfloat s0,
			      GLfloat s1,
			      GLfloat s2,
			      GLfloat s3)
{
   GLfloat values[4];
   GLint idx;
   GLuint swizzle;
   values[0] = s0;
   values[1] = s1;
   values[2] = s2;
   values[3] = s3;
   idx = _mesa_add_unnamed_constant( p->program->Base.Parameters, values, 4,
                                     &swizzle );
   ASSERT(swizzle == SWIZZLE_NOOP);
   return make_ureg(PROGRAM_CONSTANT, idx);
}

#define register_const1f(p, s0)         register_const4f(p, s0, 0, 0, 1)
#define register_scalar_const(p, s0)    register_const4f(p, s0, s0, s0, s0)
#define register_const2f(p, s0, s1)     register_const4f(p, s0, s1, 0, 1)
#define register_const3f(p, s0, s1, s2) register_const4f(p, s0, s1, s2, 1)

static GLboolean is_undef( struct ureg reg )
{
   return reg.file == PROGRAM_UNDEFINED;
}

static struct ureg get_identity_param( struct tnl_program *p )
{
   if (is_undef(p->identity)) 
      p->identity = register_const4f(p, 0,0,0,1);

   return p->identity;
}

static struct ureg register_param5(struct tnl_program *p, 
				   GLint s0,
				   GLint s1,
				   GLint s2,
				   GLint s3,
                                   GLint s4)
{
   gl_state_index tokens[STATE_LENGTH];
   GLint idx;
   tokens[0] = s0;
   tokens[1] = s1;
   tokens[2] = s2;
   tokens[3] = s3;
   tokens[4] = s4;
   idx = _mesa_add_state_reference( p->program->Base.Parameters, tokens );
   return make_ureg(PROGRAM_STATE_VAR, idx);
}


#define register_param1(p,s0)          register_param5(p,s0,0,0,0,0)
#define register_param2(p,s0,s1)       register_param5(p,s0,s1,0,0,0)
#define register_param3(p,s0,s1,s2)    register_param5(p,s0,s1,s2,0,0)
#define register_param4(p,s0,s1,s2,s3) register_param5(p,s0,s1,s2,s3,0)


static void register_matrix_param5( struct tnl_program *p,
				    GLint s0, /* modelview, projection, etc */
				    GLint s1, /* texture matrix number */
				    GLint s2, /* first row */
				    GLint s3, /* last row */
				    GLint s4, /* inverse, transpose, etc */
				    struct ureg *matrix )
{
   GLint i;

   /* This is a bit sad as the support is there to pull the whole
    * matrix out in one go:
    */
   for (i = 0; i <= s3 - s2; i++) 
      matrix[i] = register_param5( p, s0, s1, i, i, s4 );
}


static void emit_arg( struct prog_src_register *src,
		      struct ureg reg )
{
   src->File = reg.file;
   src->Index = reg.idx;
   src->Swizzle = reg.swz;
   src->NegateBase = reg.negate ? NEGATE_XYZW : 0;
   src->Abs = 0;
   src->NegateAbs = 0;
   src->RelAddr = 0;
}

static void emit_dst( struct prog_dst_register *dst,
		      struct ureg reg, GLuint mask )
{
   dst->File = reg.file;
   dst->Index = reg.idx;
   /* allow zero as a shorthand for xyzw */
   dst->WriteMask = mask ? mask : WRITEMASK_XYZW; 
   dst->CondMask = COND_TR;  /* always pass cond test */
   dst->CondSwizzle = SWIZZLE_NOOP;
   dst->CondSrc = 0;
   dst->pad = 0;
}

static void debug_insn( struct prog_instruction *inst, const char *fn,
			GLuint line )
{
   if (DISASSEM) {
      static const char *last_fn;
   
      if (fn != last_fn) {
	 last_fn = fn;
	 _mesa_printf("%s:\n", fn);
      }
	 
      _mesa_printf("%d:\t", line);
      _mesa_print_instruction(inst);
   }
}


static void emit_op3fn(struct tnl_program *p,
                       enum prog_opcode op,
		       struct ureg dest,
		       GLuint mask,
		       struct ureg src0,
		       struct ureg src1,
		       struct ureg src2,
		       const char *fn,
		       GLuint line)
{
   GLuint nr = p->program->Base.NumInstructions++;
   struct prog_instruction *inst = &p->program->Base.Instructions[nr];
      
   if (p->program->Base.NumInstructions > MAX_INSN) {
      _mesa_problem(0, "Out of instructions in emit_op3fn\n");
      return;
   }
      
   inst->Opcode = (enum prog_opcode) op; 
   inst->StringPos = 0;
   inst->Data = 0;
   
   emit_arg( &inst->SrcReg[0], src0 );
   emit_arg( &inst->SrcReg[1], src1 );
   emit_arg( &inst->SrcReg[2], src2 );   

   emit_dst( &inst->DstReg, dest, mask );

   debug_insn(inst, fn, line);
}


#define emit_op3(p, op, dst, mask, src0, src1, src2) \
   emit_op3fn(p, op, dst, mask, src0, src1, src2, __FUNCTION__, __LINE__)

#define emit_op2(p, op, dst, mask, src0, src1) \
    emit_op3fn(p, op, dst, mask, src0, src1, undef, __FUNCTION__, __LINE__)

#define emit_op1(p, op, dst, mask, src0) \
    emit_op3fn(p, op, dst, mask, src0, undef, undef, __FUNCTION__, __LINE__)


static struct ureg make_temp( struct tnl_program *p, struct ureg reg )
{
   if (reg.file == PROGRAM_TEMPORARY && 
       !(p->temp_reserved & (1<<reg.idx)))
      return reg;
   else {
      struct ureg temp = get_temp(p);
      emit_op1(p, OPCODE_MOV, temp, 0, reg);
      return temp;
   }
}


/* Currently no tracking performed of input/output/register size or
 * active elements.  Could be used to reduce these operations, as
 * could the matrix type.
 */
static void emit_matrix_transform_vec4( struct tnl_program *p,
					struct ureg dest,
					const struct ureg *mat,
					struct ureg src)
{
   emit_op2(p, OPCODE_DP4, dest, WRITEMASK_X, src, mat[0]);
   emit_op2(p, OPCODE_DP4, dest, WRITEMASK_Y, src, mat[1]);
   emit_op2(p, OPCODE_DP4, dest, WRITEMASK_Z, src, mat[2]);
   emit_op2(p, OPCODE_DP4, dest, WRITEMASK_W, src, mat[3]);
}

/* This version is much easier to implement if writemasks are not
 * supported natively on the target or (like SSE), the target doesn't
 * have a clean/obvious dotproduct implementation.
 */
static void emit_transpose_matrix_transform_vec4( struct tnl_program *p,
						  struct ureg dest,
						  const struct ureg *mat,
						  struct ureg src)
{
   struct ureg tmp;

   if (dest.file != PROGRAM_TEMPORARY)
      tmp = get_temp(p);
   else
      tmp = dest;

   emit_op2(p, OPCODE_MUL, tmp, 0, swizzle1(src,X), mat[0]);
   emit_op3(p, OPCODE_MAD, tmp, 0, swizzle1(src,Y), mat[1], tmp);
   emit_op3(p, OPCODE_MAD, tmp, 0, swizzle1(src,Z), mat[2], tmp);
   emit_op3(p, OPCODE_MAD, dest, 0, swizzle1(src,W), mat[3], tmp);

   if (dest.file != PROGRAM_TEMPORARY)
      release_temp(p, tmp);
}

static void emit_matrix_transform_vec3( struct tnl_program *p,
					struct ureg dest,
					const struct ureg *mat,
					struct ureg src)
{
   emit_op2(p, OPCODE_DP3, dest, WRITEMASK_X, src, mat[0]);
   emit_op2(p, OPCODE_DP3, dest, WRITEMASK_Y, src, mat[1]);
   emit_op2(p, OPCODE_DP3, dest, WRITEMASK_Z, src, mat[2]);
}


static void emit_normalize_vec3( struct tnl_program *p,
				 struct ureg dest,
				 struct ureg src )
{
   struct ureg tmp = get_temp(p);
   emit_op2(p, OPCODE_DP3, tmp, WRITEMASK_X, src, src);
   emit_op1(p, OPCODE_RSQ, tmp, WRITEMASK_X, tmp);
   emit_op2(p, OPCODE_MUL, dest, 0, src, swizzle1(tmp, X));
   release_temp(p, tmp);
}

static void emit_passthrough( struct tnl_program *p, 
			      GLuint input,
			      GLuint output )
{
   struct ureg out = register_output(p, output);
   emit_op1(p, OPCODE_MOV, out, 0, register_input(p, input)); 
}

static struct ureg get_eye_position( struct tnl_program *p )
{
   if (is_undef(p->eye_position)) {
      struct ureg pos = register_input( p, VERT_ATTRIB_POS ); 
      struct ureg modelview[4];

      p->eye_position = reserve_temp(p);

      if (PREFER_DP4) {
	 register_matrix_param5( p, STATE_MODELVIEW_MATRIX, 0, 0, 3,
                                 0, modelview );

	 emit_matrix_transform_vec4(p, p->eye_position, modelview, pos);
      }
      else {
	 register_matrix_param5( p, STATE_MODELVIEW_MATRIX, 0, 0, 3,
				 STATE_MATRIX_TRANSPOSE, modelview );

	 emit_transpose_matrix_transform_vec4(p, p->eye_position, modelview, pos);
      }
   }
   
   return p->eye_position;
}


static struct ureg get_eye_position_z( struct tnl_program *p )
{
   if (!is_undef(p->eye_position)) 
      return swizzle1(p->eye_position, Z);

   if (is_undef(p->eye_position_z)) {
      struct ureg pos = register_input( p, VERT_ATTRIB_POS ); 
      struct ureg modelview[4];

      p->eye_position_z = reserve_temp(p);

      register_matrix_param5( p, STATE_MODELVIEW_MATRIX, 0, 0, 3,
                              0, modelview );

      emit_op2(p, OPCODE_DP4, p->eye_position_z, 0, pos, modelview[2]);
   }
   
   return p->eye_position_z;
}
   


static struct ureg get_eye_position_normalized( struct tnl_program *p )
{
   if (is_undef(p->eye_position_normalized)) {
      struct ureg eye = get_eye_position(p);
      p->eye_position_normalized = reserve_temp(p);
      emit_normalize_vec3(p, p->eye_position_normalized, eye);
   }
   
   return p->eye_position_normalized;
}


static struct ureg get_transformed_normal( struct tnl_program *p )
{
   if (is_undef(p->transformed_normal) &&
       !p->state->need_eye_coords &&
       !p->state->normalize &&
       !(p->state->need_eye_coords == p->state->rescale_normals))
   {
      p->transformed_normal = register_input(p, VERT_ATTRIB_NORMAL );
   }
   else if (is_undef(p->transformed_normal)) 
   {
      struct ureg normal = register_input(p, VERT_ATTRIB_NORMAL );
      struct ureg mvinv[3];
      struct ureg transformed_normal = reserve_temp(p);

      if (p->state->need_eye_coords) {
         register_matrix_param5( p, STATE_MODELVIEW_MATRIX, 0, 0, 2,
                                 STATE_MATRIX_INVTRANS, mvinv );

         /* Transform to eye space:
          */
         emit_matrix_transform_vec3( p, transformed_normal, mvinv, normal );
         normal = transformed_normal;
      }

      /* Normalize/Rescale:
       */
      if (p->state->normalize) {
	 emit_normalize_vec3( p, transformed_normal, normal );
         normal = transformed_normal;
      }
      else if (p->state->need_eye_coords == p->state->rescale_normals) {
         /* This is already adjusted for eye/non-eye rendering:
          */
	 struct ureg rescale = register_param2(p, STATE_INTERNAL,
                                               STATE_NORMAL_SCALE);

	 emit_op2( p, OPCODE_MUL, transformed_normal, 0, normal, rescale );
         normal = transformed_normal;
      }
      
      assert(normal.file == PROGRAM_TEMPORARY);
      p->transformed_normal = normal;
   }

   return p->transformed_normal;
}



static void build_hpos( struct tnl_program *p )
{
   struct ureg pos = register_input( p, VERT_ATTRIB_POS ); 
   struct ureg hpos = register_output( p, VERT_RESULT_HPOS );
   struct ureg mvp[4];

   if (PREFER_DP4) {
      register_matrix_param5( p, STATE_MVP_MATRIX, 0, 0, 3, 
			      0, mvp );
      emit_matrix_transform_vec4( p, hpos, mvp, pos );
   }
   else {
      register_matrix_param5( p, STATE_MVP_MATRIX, 0, 0, 3, 
			      STATE_MATRIX_TRANSPOSE, mvp );
      emit_transpose_matrix_transform_vec4( p, hpos, mvp, pos );
   }
}


static GLuint material_attrib( GLuint side, GLuint property )
{
   return ((property - STATE_AMBIENT) * 2 + 
	   side);
}

/* Get a bitmask of which material values vary on a per-vertex basis.
 */
static void set_material_flags( struct tnl_program *p )
{
   p->color_materials = 0;
   p->materials = 0;

   if (p->state->light_color_material) {
      p->materials = 
	 p->color_materials = p->state->light_color_material_mask;
   }

   p->materials |= p->state->light_material_mask;
}


/* XXX temporary!!! */
#define _TNL_ATTRIB_MAT_FRONT_AMBIENT 32

static struct ureg get_material( struct tnl_program *p, GLuint side, 
				 GLuint property )
{
   GLuint attrib = material_attrib(side, property);

   if (p->color_materials & (1<<attrib))
      return register_input(p, VERT_ATTRIB_COLOR0);
   else if (p->materials & (1<<attrib)) 
      return register_input( p, attrib + _TNL_ATTRIB_MAT_FRONT_AMBIENT );
   else
      return register_param3( p, STATE_MATERIAL, side, property );
}

#define SCENE_COLOR_BITS(side) (( MAT_BIT_FRONT_EMISSION | \
				   MAT_BIT_FRONT_AMBIENT | \
				   MAT_BIT_FRONT_DIFFUSE) << (side))

/* Either return a precalculated constant value or emit code to
 * calculate these values dynamically in the case where material calls
 * are present between begin/end pairs.
 *
 * Probably want to shift this to the program compilation phase - if
 * we always emitted the calculation here, a smart compiler could
 * detect that it was constant (given a certain set of inputs), and
 * lift it out of the main loop.  That way the programs created here
 * would be independent of the vertex_buffer details.
 */
static struct ureg get_scenecolor( struct tnl_program *p, GLuint side )
{
   if (p->materials & SCENE_COLOR_BITS(side)) {
      struct ureg lm_ambient = register_param1(p, STATE_LIGHTMODEL_AMBIENT);
      struct ureg material_emission = get_material(p, side, STATE_EMISSION);
      struct ureg material_ambient = get_material(p, side, STATE_AMBIENT);
      struct ureg material_diffuse = get_material(p, side, STATE_DIFFUSE);
      struct ureg tmp = make_temp(p, material_diffuse);
      emit_op3(p, OPCODE_MAD, tmp,  WRITEMASK_XYZ, lm_ambient, 
	       material_ambient, material_emission);
      return tmp;
   }
   else
      return register_param2( p, STATE_LIGHTMODEL_SCENECOLOR, side );
}


static struct ureg get_lightprod( struct tnl_program *p, GLuint light, 
				  GLuint side, GLuint property )
{
   GLuint attrib = material_attrib(side, property);
   if (p->materials & (1<<attrib)) {
      struct ureg light_value = 
	 register_param3(p, STATE_LIGHT, light, property);
      struct ureg material_value = get_material(p, side, property);
      struct ureg tmp = get_temp(p);
      emit_op2(p, OPCODE_MUL, tmp,  0, light_value, material_value);
      return tmp;
   }
   else
      return register_param4(p, STATE_LIGHTPROD, light, side, property);
}

static struct ureg calculate_light_attenuation( struct tnl_program *p,
						GLuint i, 
						struct ureg VPpli,
						struct ureg dist )
{
   struct ureg attenuation = register_param3(p, STATE_LIGHT, i,
					     STATE_ATTENUATION);
   struct ureg att = get_temp(p);

   /* Calculate spot attenuation:
    */
   if (!p->state->unit[i].light_spotcutoff_is_180) {
      struct ureg spot_dir_norm = register_param3(p, STATE_INTERNAL,
						  STATE_LIGHT_SPOT_DIR_NORMALIZED, i);
      struct ureg spot = get_temp(p);
      struct ureg slt = get_temp(p);

      emit_op2(p, OPCODE_DP3, spot, 0, negate(VPpli), spot_dir_norm);
      emit_op2(p, OPCODE_SLT, slt, 0, swizzle1(spot_dir_norm,W), spot);
      emit_op2(p, OPCODE_POW, spot, 0, spot, swizzle1(attenuation, W));
      emit_op2(p, OPCODE_MUL, att, 0, slt, spot);

      release_temp(p, spot);
      release_temp(p, slt);
   }

   /* Calculate distance attenuation:
    */
   if (p->state->unit[i].light_attenuated) {

      /* 1/d,d,d,1/d */
      emit_op1(p, OPCODE_RCP, dist, WRITEMASK_YZ, dist); 
      /* 1,d,d*d,1/d */
      emit_op2(p, OPCODE_MUL, dist, WRITEMASK_XZ, dist, swizzle1(dist,Y)); 
      /* 1/dist-atten */
      emit_op2(p, OPCODE_DP3, dist, 0, attenuation, dist); 

      if (!p->state->unit[i].light_spotcutoff_is_180) {
	 /* dist-atten */
	 emit_op1(p, OPCODE_RCP, dist, 0, dist); 
	 /* spot-atten * dist-atten */
	 emit_op2(p, OPCODE_MUL, att, 0, dist, att);	
      } else {
	 /* dist-atten */
	 emit_op1(p, OPCODE_RCP, att, 0, dist); 
      }
   }

   return att;
}
						

/**
 * Compute:
 *   lit.y = MAX(0, dots.x)
 *   lit.z = SLT(0, dots.x)
 */
static void emit_degenerate_lit( struct tnl_program *p,
                                 struct ureg lit,
                                 struct ureg dots )
{
   struct ureg id = get_identity_param(p);  /* id = {0,0,0,1} */

   /* Note that lit.x & lit.w will not be examined.  Note also that
    * dots.xyzw == dots.xxxx.
    */

   /* MAX lit, id, dots;
    */
   emit_op2(p, OPCODE_MAX, lit, WRITEMASK_XYZW, id, dots); 

   /* result[2] = (in > 0 ? 1 : 0)
    * SLT lit.z, id.z, dots;   # lit.z = (0 < dots.z) ? 1 : 0
    */
   emit_op2(p, OPCODE_SLT, lit, WRITEMASK_Z, swizzle1(id,Z), dots);
}


/* Need to add some addtional parameters to allow lighting in object
 * space - STATE_SPOT_DIRECTION and STATE_HALF_VECTOR implicitly assume eye
 * space lighting.
 */
static void build_lighting( struct tnl_program *p )
{
   const GLboolean twoside = p->state->light_twoside;
   const GLboolean separate = p->state->separate_specular;
   GLuint nr_lights = 0, count = 0;
   struct ureg normal = get_transformed_normal(p);
   struct ureg lit = get_temp(p);
   struct ureg dots = get_temp(p);
   struct ureg _col0 = undef, _col1 = undef;
   struct ureg _bfc0 = undef, _bfc1 = undef;
   GLuint i;

   /*
    * NOTE:
    * dot.x = dot(normal, VPpli)
    * dot.y = dot(normal, halfAngle)
    * dot.z = back.shininess
    * dot.w = front.shininess
    */

   for (i = 0; i < MAX_LIGHTS; i++) 
      if (p->state->unit[i].light_enabled)
	 nr_lights++;
   
   set_material_flags(p);

   {
      if (!p->state->material_shininess_is_zero) {
         struct ureg shininess = get_material(p, 0, STATE_SHININESS);
         emit_op1(p, OPCODE_MOV, dots,  WRITEMASK_W, swizzle1(shininess,X));
         release_temp(p, shininess);
      }

      _col0 = make_temp(p, get_scenecolor(p, 0));
      if (separate)
	 _col1 = make_temp(p, get_identity_param(p));
      else
	 _col1 = _col0;

   }

   if (twoside) {
      if (!p->state->material_shininess_is_zero) {
         struct ureg shininess = get_material(p, 1, STATE_SHININESS);
         emit_op1(p, OPCODE_MOV, dots, WRITEMASK_Z, 
                  negate(swizzle1(shininess,X)));
         release_temp(p, shininess);
      }

      _bfc0 = make_temp(p, get_scenecolor(p, 1));
      if (separate)
	 _bfc1 = make_temp(p, get_identity_param(p));
      else
	 _bfc1 = _bfc0;
   }

   /* If no lights, still need to emit the scenecolor.
    */
   {
      struct ureg res0 = register_output( p, VERT_RESULT_COL0 );
      emit_op1(p, OPCODE_MOV, res0, 0, _col0);
   }

   if (separate) {
      struct ureg res1 = register_output( p, VERT_RESULT_COL1 );
      emit_op1(p, OPCODE_MOV, res1, 0, _col1);
   }

   if (twoside) {
      struct ureg res0 = register_output( p, VERT_RESULT_BFC0 );
      emit_op1(p, OPCODE_MOV, res0, 0, _bfc0);
   }
      
   if (twoside && separate) {
      struct ureg res1 = register_output( p, VERT_RESULT_BFC1 );
      emit_op1(p, OPCODE_MOV, res1, 0, _bfc1);
   }
      
   if (nr_lights == 0) {
      release_temps(p);
      return;
   }

   for (i = 0; i < MAX_LIGHTS; i++) {
      if (p->state->unit[i].light_enabled) {
	 struct ureg half = undef;
	 struct ureg att = undef, VPpli = undef;
	  
	 count++;

	 if (p->state->unit[i].light_eyepos3_is_zero) {
	    /* Can used precomputed constants in this case.
	     * Attenuation never applies to infinite lights.
	     */
	    VPpli = register_param3(p, STATE_INTERNAL, 
				    STATE_LIGHT_POSITION_NORMALIZED, i); 
            
            if (!p->state->material_shininess_is_zero) {
               if (p->state->light_local_viewer) {
                  struct ureg eye_hat = get_eye_position_normalized(p);
                  half = get_temp(p);
                  emit_op2(p, OPCODE_SUB, half, 0, VPpli, eye_hat);
                  emit_normalize_vec3(p, half, half);
               } else {
                  half = register_param3(p, STATE_INTERNAL, 
                                         STATE_LIGHT_HALF_VECTOR, i);
               }
            }
	 } 
	 else {
	    struct ureg Ppli = register_param3(p, STATE_INTERNAL, 
					       STATE_LIGHT_POSITION, i); 
	    struct ureg V = get_eye_position(p);
	    struct ureg dist = get_temp(p);

	    VPpli = get_temp(p); 
 
	    /* Calculate VPpli vector
	     */
	    emit_op2(p, OPCODE_SUB, VPpli, 0, Ppli, V); 

	    /* Normalize VPpli.  The dist value also used in
	     * attenuation below.
	     */
	    emit_op2(p, OPCODE_DP3, dist, 0, VPpli, VPpli);
	    emit_op1(p, OPCODE_RSQ, dist, 0, dist);
	    emit_op2(p, OPCODE_MUL, VPpli, 0, VPpli, dist);

	    /* Calculate attenuation:
	     */ 
	    if (!p->state->unit[i].light_spotcutoff_is_180 ||
		p->state->unit[i].light_attenuated) {
	       att = calculate_light_attenuation(p, i, VPpli, dist);
	    }

	    /* Calculate viewer direction, or use infinite viewer:
	     */
            if (!p->state->material_shininess_is_zero) {
               half = get_temp(p);

               if (p->state->light_local_viewer) {
                  struct ureg eye_hat = get_eye_position_normalized(p);
                  emit_op2(p, OPCODE_SUB, half, 0, VPpli, eye_hat);
               }
               else {
                  struct ureg z_dir = swizzle(get_identity_param(p),X,Y,W,Z); 
                  emit_op2(p, OPCODE_ADD, half, 0, VPpli, z_dir);
               }

               emit_normalize_vec3(p, half, half);
            }

	    release_temp(p, dist);
	 }

	 /* Calculate dot products:
	  */
         if (p->state->material_shininess_is_zero) {
            emit_op2(p, OPCODE_DP3, dots, 0, normal, VPpli);
         }
         else {
            emit_op2(p, OPCODE_DP3, dots, WRITEMASK_X, normal, VPpli);
            emit_op2(p, OPCODE_DP3, dots, WRITEMASK_Y, normal, half);
         }

	 /* Front face lighting:
	  */
	 {
	    struct ureg ambient = get_lightprod(p, i, 0, STATE_AMBIENT);
	    struct ureg diffuse = get_lightprod(p, i, 0, STATE_DIFFUSE);
	    struct ureg specular = get_lightprod(p, i, 0, STATE_SPECULAR);
	    struct ureg res0, res1;
	    GLuint mask0, mask1;

	    
	    if (count == nr_lights) {
	       if (separate) {
		  mask0 = WRITEMASK_XYZ;
		  mask1 = WRITEMASK_XYZ;
		  res0 = register_output( p, VERT_RESULT_COL0 );
		  res1 = register_output( p, VERT_RESULT_COL1 );
	       }
	       else {
		  mask0 = 0;
		  mask1 = WRITEMASK_XYZ;
		  res0 = _col0;
		  res1 = register_output( p, VERT_RESULT_COL0 );
	       }
	    } else {
	       mask0 = 0;
	       mask1 = 0;
	       res0 = _col0;
	       res1 = _col1;
	    }


	    if (!is_undef(att)) {
               /* light is attenuated by distance */
               emit_op1(p, OPCODE_LIT, lit, 0, dots);
               emit_op2(p, OPCODE_MUL, lit, 0, lit, att);
               emit_op3(p, OPCODE_MAD, _col0, 0, swizzle1(lit,X), ambient, _col0);
            } 
            else if (!p->state->material_shininess_is_zero) {
               /* there's a non-zero specular term */
               emit_op1(p, OPCODE_LIT, lit, 0, dots);
               emit_op2(p, OPCODE_ADD, _col0, 0, ambient, _col0);
            } 
            else {
               /* no attenutation, no specular */
               emit_degenerate_lit(p, lit, dots);
               emit_op2(p, OPCODE_ADD, _col0, 0, ambient, _col0);
            }

	    emit_op3(p, OPCODE_MAD, res0, mask0, swizzle1(lit,Y), diffuse, _col0);
	    emit_op3(p, OPCODE_MAD, res1, mask1, swizzle1(lit,Z), specular, _col1);
      
	    release_temp(p, ambient);
	    release_temp(p, diffuse);
	    release_temp(p, specular);
	 }

	 /* Back face lighting:
	  */
	 if (twoside) {
	    struct ureg ambient = get_lightprod(p, i, 1, STATE_AMBIENT);
	    struct ureg diffuse = get_lightprod(p, i, 1, STATE_DIFFUSE);
	    struct ureg specular = get_lightprod(p, i, 1, STATE_SPECULAR);
	    struct ureg res0, res1;
	    GLuint mask0, mask1;
	       
	    if (count == nr_lights) {
	       if (separate) {
		  mask0 = WRITEMASK_XYZ;
		  mask1 = WRITEMASK_XYZ;
		  res0 = register_output( p, VERT_RESULT_BFC0 );
		  res1 = register_output( p, VERT_RESULT_BFC1 );
	       }
	       else {
		  mask0 = 0;
		  mask1 = WRITEMASK_XYZ;
		  res0 = _bfc0;
		  res1 = register_output( p, VERT_RESULT_BFC0 );
	       }
	    } else {
	       res0 = _bfc0;
	       res1 = _bfc1;
	       mask0 = 0;
	       mask1 = 0;
	    }

            dots = negate(swizzle(dots,X,Y,W,Z));

	    if (!is_undef(att)) {
               emit_op1(p, OPCODE_LIT, lit, 0, dots);
	       emit_op2(p, OPCODE_MUL, lit, 0, lit, att);
               emit_op3(p, OPCODE_MAD, _bfc0, 0, swizzle1(lit,X), ambient, _bfc0);
            }
            else if (!p->state->material_shininess_is_zero) {
               emit_op1(p, OPCODE_LIT, lit, 0, dots);
               emit_op2(p, OPCODE_ADD, _col0, 0, ambient, _col0);
            } 
            else {
               emit_degenerate_lit(p, lit, dots);
               emit_op2(p, OPCODE_ADD, _col0, 0, ambient, _col0);
            }

	    emit_op2(p, OPCODE_ADD, _bfc0, 0, ambient, _bfc0);
	    emit_op3(p, OPCODE_MAD, res0, mask0, swizzle1(lit,Y), diffuse, _bfc0);
	    emit_op3(p, OPCODE_MAD, res1, mask1, swizzle1(lit,Z), specular, _bfc1);

	    release_temp(p, ambient);
	    release_temp(p, diffuse);
	    release_temp(p, specular);
	 }

	 release_temp(p, half);
	 release_temp(p, VPpli);
	 release_temp(p, att);
      }
   }

   release_temps( p );
}


static void build_fog( struct tnl_program *p )
{
   struct ureg fog = register_output(p, VERT_RESULT_FOGC);
   struct ureg input;

   if (p->state->fog_source_is_depth) {
      input = get_eye_position_z(p);
   }
   else {
      input = swizzle1(register_input(p, VERT_ATTRIB_FOG), X);
   }

   if (p->state->fog_mode && p->state->tnl_do_vertex_fog) {
      struct ureg params = register_param2(p, STATE_INTERNAL,
					   STATE_FOG_PARAMS_OPTIMIZED);
      struct ureg tmp = get_temp(p);
      GLboolean useabs = (p->state->fog_mode != FOG_EXP2);

      if (useabs) {
	 emit_op1(p, OPCODE_ABS, tmp, 0, input);
      }

      switch (p->state->fog_mode) {
      case FOG_LINEAR: {
	 struct ureg id = get_identity_param(p);
	 emit_op3(p, OPCODE_MAD, tmp, 0, useabs ? tmp : input,
			swizzle1(params,X), swizzle1(params,Y));
	 emit_op2(p, OPCODE_MAX, tmp, 0, tmp, swizzle1(id,X)); /* saturate */
	 emit_op2(p, OPCODE_MIN, fog, WRITEMASK_X, tmp, swizzle1(id,W));
	 break;
      }
      case FOG_EXP:
	 emit_op2(p, OPCODE_MUL, tmp, 0, useabs ? tmp : input,
			swizzle1(params,Z));
	 emit_op1(p, OPCODE_EX2, fog, WRITEMASK_X, negate(tmp));
	 break;
      case FOG_EXP2:
	 emit_op2(p, OPCODE_MUL, tmp, 0, input, swizzle1(params,W));
	 emit_op2(p, OPCODE_MUL, tmp, 0, tmp, tmp);
	 emit_op1(p, OPCODE_EX2, fog, WRITEMASK_X, negate(tmp));
	 break;
      }

      release_temp(p, tmp);
   }
   else {
      /* results = incoming fog coords (compute fog per-fragment later) 
       *
       * KW:  Is it really necessary to do anything in this case?
       * BP: Yes, we always need to compute the absolute value, unless
       * we want to push that down into the fragment program...
       */
      GLboolean useabs = GL_TRUE;
      emit_op1(p, useabs ? OPCODE_ABS : OPCODE_MOV, fog, WRITEMASK_X, input);
   }
}
 
static void build_reflect_texgen( struct tnl_program *p,
				  struct ureg dest,
				  GLuint writemask )
{
   struct ureg normal = get_transformed_normal(p);
   struct ureg eye_hat = get_eye_position_normalized(p);
   struct ureg tmp = get_temp(p);

   /* n.u */
   emit_op2(p, OPCODE_DP3, tmp, 0, normal, eye_hat); 
   /* 2n.u */
   emit_op2(p, OPCODE_ADD, tmp, 0, tmp, tmp); 
   /* (-2n.u)n + u */
   emit_op3(p, OPCODE_MAD, dest, writemask, negate(tmp), normal, eye_hat);

   release_temp(p, tmp);
}

static void build_sphere_texgen( struct tnl_program *p,
				 struct ureg dest,
				 GLuint writemask )
{
   struct ureg normal = get_transformed_normal(p);
   struct ureg eye_hat = get_eye_position_normalized(p);
   struct ureg tmp = get_temp(p);
   struct ureg half = register_scalar_const(p, .5);
   struct ureg r = get_temp(p);
   struct ureg inv_m = get_temp(p);
   struct ureg id = get_identity_param(p);

   /* Could share the above calculations, but it would be
    * a fairly odd state for someone to set (both sphere and
    * reflection active for different texture coordinate
    * components.  Of course - if two texture units enable
    * reflect and/or sphere, things start to tilt in favour
    * of seperating this out:
    */

   /* n.u */
   emit_op2(p, OPCODE_DP3, tmp, 0, normal, eye_hat); 
   /* 2n.u */
   emit_op2(p, OPCODE_ADD, tmp, 0, tmp, tmp); 
   /* (-2n.u)n + u */
   emit_op3(p, OPCODE_MAD, r, 0, negate(tmp), normal, eye_hat); 
   /* r + 0,0,1 */
   emit_op2(p, OPCODE_ADD, tmp, 0, r, swizzle(id,X,Y,W,Z)); 
   /* rx^2 + ry^2 + (rz+1)^2 */
   emit_op2(p, OPCODE_DP3, tmp, 0, tmp, tmp); 
   /* 2/m */
   emit_op1(p, OPCODE_RSQ, tmp, 0, tmp); 
   /* 1/m */
   emit_op2(p, OPCODE_MUL, inv_m, 0, tmp, half); 
   /* r/m + 1/2 */
   emit_op3(p, OPCODE_MAD, dest, writemask, r, inv_m, half); 
	       
   release_temp(p, tmp);
   release_temp(p, r);
   release_temp(p, inv_m);
}


static void build_texture_transform( struct tnl_program *p )
{
   GLuint i, j;

   for (i = 0; i < MAX_TEXTURE_UNITS; i++) {

      if (!(p->state->fragprog_inputs_read & FRAG_BIT_TEX(i)))
	 continue;
							     
      if (p->state->unit[i].texgen_enabled || 
	  p->state->unit[i].texmat_enabled) {
	 
	 GLuint texmat_enabled = p->state->unit[i].texmat_enabled;
	 struct ureg out = register_output(p, VERT_RESULT_TEX0 + i);
	 struct ureg out_texgen = undef;

	 if (p->state->unit[i].texgen_enabled) {
	    GLuint copy_mask = 0;
	    GLuint sphere_mask = 0;
	    GLuint reflect_mask = 0;
	    GLuint normal_mask = 0;
	    GLuint modes[4];
	 
	    if (texmat_enabled) 
	       out_texgen = get_temp(p);
	    else
	       out_texgen = out;

	    modes[0] = p->state->unit[i].texgen_mode0;
	    modes[1] = p->state->unit[i].texgen_mode1;
	    modes[2] = p->state->unit[i].texgen_mode2;
	    modes[3] = p->state->unit[i].texgen_mode3;

	    for (j = 0; j < 4; j++) {
	       switch (modes[j]) {
	       case TXG_OBJ_LINEAR: {
		  struct ureg obj = register_input(p, VERT_ATTRIB_POS);
		  struct ureg plane = 
		     register_param3(p, STATE_TEXGEN, i,
				     STATE_TEXGEN_OBJECT_S + j);

		  emit_op2(p, OPCODE_DP4, out_texgen, WRITEMASK_X << j, 
			   obj, plane );
		  break;
	       }
	       case TXG_EYE_LINEAR: {
		  struct ureg eye = get_eye_position(p);
		  struct ureg plane = 
		     register_param3(p, STATE_TEXGEN, i, 
				     STATE_TEXGEN_EYE_S + j);

		  emit_op2(p, OPCODE_DP4, out_texgen, WRITEMASK_X << j, 
			   eye, plane );
		  break;
	       }
	       case TXG_SPHERE_MAP: 
		  sphere_mask |= WRITEMASK_X << j;
		  break;
	       case TXG_REFLECTION_MAP:
		  reflect_mask |= WRITEMASK_X << j;
		  break;
	       case TXG_NORMAL_MAP: 
		  normal_mask |= WRITEMASK_X << j;
		  break;
	       case TXG_NONE:
		  copy_mask |= WRITEMASK_X << j;
	       }

	    }

	 
	    if (sphere_mask) {
	       build_sphere_texgen(p, out_texgen, sphere_mask);
	    }

	    if (reflect_mask) {
	       build_reflect_texgen(p, out_texgen, reflect_mask);
	    }

	    if (normal_mask) {
	       struct ureg normal = get_transformed_normal(p);
	       emit_op1(p, OPCODE_MOV, out_texgen, normal_mask, normal );
	    }

	    if (copy_mask) {
	       struct ureg in = register_input(p, VERT_ATTRIB_TEX0+i);
	       emit_op1(p, OPCODE_MOV, out_texgen, copy_mask, in );
	    }
	 }

	 if (texmat_enabled) {
	    struct ureg texmat[4];
	    struct ureg in = (!is_undef(out_texgen) ? 
			      out_texgen : 
			      register_input(p, VERT_ATTRIB_TEX0+i));
	    if (PREFER_DP4) {
	       register_matrix_param5( p, STATE_TEXTURE_MATRIX, i, 0, 3,
				       0, texmat );
	       emit_matrix_transform_vec4( p, out, texmat, in );
	    }
	    else {
	       register_matrix_param5( p, STATE_TEXTURE_MATRIX, i, 0, 3,
				       STATE_MATRIX_TRANSPOSE, texmat );
	       emit_transpose_matrix_transform_vec4( p, out, texmat, in );
	    }
	 }

	 release_temps(p);
      } 
      else {
	 emit_passthrough(p, VERT_ATTRIB_TEX0+i, VERT_RESULT_TEX0+i);
      }
   }
}


/**
 * Point size attenuation computation.
 */
static void build_atten_pointsize( struct tnl_program *p )
{
   struct ureg eye = get_eye_position_z(p);
   struct ureg state_size = register_param1(p, STATE_POINT_SIZE);
   struct ureg state_attenuation = register_param1(p, STATE_POINT_ATTENUATION);
   struct ureg out = register_output(p, VERT_RESULT_PSIZ);
   struct ureg ut = get_temp(p);

   /* dist = |eyez| */
   emit_op1(p, OPCODE_ABS, ut, WRITEMASK_Y, swizzle1(eye, Z));
   /* p1 + dist * (p2 + dist * p3); */
   emit_op3(p, OPCODE_MAD, ut, WRITEMASK_X, swizzle1(ut, Y),
		swizzle1(state_attenuation, Z), swizzle1(state_attenuation, Y));
   emit_op3(p, OPCODE_MAD, ut, WRITEMASK_X, swizzle1(ut, Y),
		ut, swizzle1(state_attenuation, X));

   /* 1 / sqrt(factor) */
   emit_op1(p, OPCODE_RSQ, ut, WRITEMASK_X, ut );

#if 0
   /* out = pointSize / sqrt(factor) */
   emit_op2(p, OPCODE_MUL, out, WRITEMASK_X, ut, state_size);
#else
   /* this is a good place to clamp the point size since there's likely
    * no hardware registers to clamp point size at rasterization time.
    */
   emit_op2(p, OPCODE_MUL, ut, WRITEMASK_X, ut, state_size);
   emit_op2(p, OPCODE_MAX, ut, WRITEMASK_X, ut, swizzle1(state_size, Y));
   emit_op2(p, OPCODE_MIN, out, WRITEMASK_X, ut, swizzle1(state_size, Z));
#endif

   release_temp(p, ut);
}

/**
 * Emit constant point size.
 */
static void build_constant_pointsize( struct tnl_program *p )
{
   struct ureg state_size = register_param1(p, STATE_POINT_SIZE);
   struct ureg out = register_output(p, VERT_RESULT_PSIZ);
   emit_op1(p, OPCODE_MOV, out, WRITEMASK_X, state_size);
}

/**
 * Pass-though per-vertex point size, from user's point size array.
 */
static void build_array_pointsize( struct tnl_program *p )
{
   struct ureg in = register_input(p, VERT_ATTRIB_POINT_SIZE);
   struct ureg out = register_output(p, VERT_RESULT_PSIZ);
   emit_op1(p, OPCODE_MOV, out, WRITEMASK_X, in);
}


static void build_tnl_program( struct tnl_program *p )
{   /* Emit the program, starting with modelviewproject:
    */
   build_hpos(p);

   /* Lighting calculations:
    */
   if (p->state->fragprog_inputs_read & (FRAG_BIT_COL0|FRAG_BIT_COL1)) {
      if (p->state->light_global_enabled)
	 build_lighting(p);
      else {
	 if (p->state->fragprog_inputs_read & FRAG_BIT_COL0)
	    emit_passthrough(p, VERT_ATTRIB_COLOR0, VERT_RESULT_COL0);

	 if (p->state->fragprog_inputs_read & FRAG_BIT_COL1)
	    emit_passthrough(p, VERT_ATTRIB_COLOR1, VERT_RESULT_COL1);
      }
   }

   if ((p->state->fragprog_inputs_read & FRAG_BIT_FOGC) ||
       p->state->fog_mode != FOG_NONE)
      build_fog(p);

   if (p->state->fragprog_inputs_read & FRAG_BITS_TEX_ANY)
      build_texture_transform(p);

   if (p->state->point_attenuated)
      build_atten_pointsize(p);
   else if (p->state->point_array)
      build_array_pointsize(p);
#if 0
   else
      build_constant_pointsize(p);
#endif

   /* Finish up:
    */
   emit_op1(p, OPCODE_END, undef, 0, undef);

   /* Disassemble:
    */
   if (DISASSEM) {
      _mesa_printf ("\n");
   }
}


static void
create_new_program( const struct state_key *key,
                    struct gl_vertex_program *program,
                    GLuint max_temps)
{
   struct tnl_program p;

   _mesa_memset(&p, 0, sizeof(p));
   p.state = key;
   p.program = program;
   p.eye_position = undef;
   p.eye_position_z = undef;
   p.eye_position_normalized = undef;
   p.transformed_normal = undef;
   p.identity = undef;
   p.temp_in_use = 0;
   
   if (max_temps >= sizeof(int) * 8)
      p.temp_reserved = 0;
   else
      p.temp_reserved = ~((1<<max_temps)-1);

   p.program->Base.Instructions = _mesa_alloc_instructions(MAX_INSN);
   p.program->Base.String = NULL;
   p.program->Base.NumInstructions =
   p.program->Base.NumTemporaries =
   p.program->Base.NumParameters =
   p.program->Base.NumAttributes = p.program->Base.NumAddressRegs = 0;
   p.program->Base.Parameters = _mesa_new_parameter_list();
   p.program->Base.InputsRead = 0;
   p.program->Base.OutputsWritten = 0;

   build_tnl_program( &p );
}


/**
 * Return a vertex program which implements the current fixed-function
 * transform/lighting/texgen operations.
 * XXX move this into core mesa (main/)
 */
struct gl_vertex_program *
_mesa_get_fixed_func_vertex_program(GLcontext *ctx)
{
   struct gl_vertex_program *prog;
   struct state_key *key;

   /* Grab all the relevent state and put it in a single structure:
    */
   key = make_state_key(ctx);

   /* Look for an already-prepared program for this state:
    */
   prog = (struct gl_vertex_program *)
      _mesa_search_program_cache(ctx->VertexProgram.Cache, key, sizeof(*key));
   
   if (!prog) {
      /* OK, we'll have to build a new one */
      if (0)
         _mesa_printf("Build new TNL program\n");
	 
      prog = (struct gl_vertex_program *)
         ctx->Driver.NewProgram(ctx, GL_VERTEX_PROGRAM_ARB, 0); 
      if (!prog)
         return NULL;

      create_new_program( key, prog,
                          ctx->Const.VertexProgram.MaxTemps );

#if 0
      if (ctx->Driver.ProgramStringNotify)
         ctx->Driver.ProgramStringNotify( ctx, GL_VERTEX_PROGRAM_ARB, 
                                          &prog->Base );
#endif
      _mesa_program_cache_insert(ctx, ctx->VertexProgram.Cache,
                                 key, sizeof(*key), &prog->Base);
   }

   _mesa_free(key);

   return prog;
}