summaryrefslogtreecommitdiffstats
path: root/src/gallium/drivers/svga/svga_tgsi_vgpu10.c
blob: 1223e4460558af455b5b3556abd65bde51008a28 (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
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
2959
2960
2961
2962
2963
2964
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
3015
3016
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
3329
3330
3331
3332
3333
3334
3335
3336
3337
3338
3339
3340
3341
3342
3343
3344
3345
3346
3347
3348
3349
3350
3351
3352
3353
3354
3355
3356
3357
3358
3359
3360
3361
3362
3363
3364
3365
3366
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
3378
3379
3380
3381
3382
3383
3384
3385
3386
3387
3388
3389
3390
3391
3392
3393
3394
3395
3396
3397
3398
3399
3400
3401
3402
3403
3404
3405
3406
3407
3408
3409
3410
3411
3412
3413
3414
3415
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437
3438
3439
3440
3441
3442
3443
3444
3445
3446
3447
3448
3449
3450
3451
3452
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
3489
3490
3491
3492
3493
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
3522
3523
3524
3525
3526
3527
3528
3529
3530
3531
3532
3533
3534
3535
3536
3537
3538
3539
3540
3541
3542
3543
3544
3545
3546
3547
3548
3549
3550
3551
3552
3553
3554
3555
3556
3557
3558
3559
3560
3561
3562
3563
3564
3565
3566
3567
3568
3569
3570
3571
3572
3573
3574
3575
3576
3577
3578
3579
3580
3581
3582
3583
3584
3585
3586
3587
3588
3589
3590
3591
3592
3593
3594
3595
3596
3597
3598
3599
3600
3601
3602
3603
3604
3605
3606
3607
3608
3609
3610
3611
3612
3613
3614
3615
3616
3617
3618
3619
3620
3621
3622
3623
3624
3625
3626
3627
3628
3629
3630
3631
3632
3633
3634
3635
3636
3637
3638
3639
3640
3641
3642
3643
3644
3645
3646
3647
3648
3649
3650
3651
3652
3653
3654
3655
3656
3657
3658
3659
3660
3661
3662
3663
3664
3665
3666
3667
3668
3669
3670
3671
3672
3673
3674
3675
3676
3677
3678
3679
3680
3681
3682
3683
3684
3685
3686
3687
3688
3689
3690
3691
3692
3693
3694
3695
3696
3697
3698
3699
3700
3701
3702
3703
3704
3705
3706
3707
3708
3709
3710
3711
3712
3713
3714
3715
3716
3717
3718
3719
3720
3721
3722
3723
3724
3725
3726
3727
3728
3729
3730
3731
3732
3733
3734
3735
3736
3737
3738
3739
3740
3741
3742
3743
3744
3745
3746
3747
3748
3749
3750
3751
3752
3753
3754
3755
3756
3757
3758
3759
3760
3761
3762
3763
3764
3765
3766
3767
3768
3769
3770
3771
3772
3773
3774
3775
3776
3777
3778
3779
3780
3781
3782
3783
3784
3785
3786
3787
3788
3789
3790
3791
3792
3793
3794
3795
3796
3797
3798
3799
3800
3801
3802
3803
3804
3805
3806
3807
3808
3809
3810
3811
3812
3813
3814
3815
3816
3817
3818
3819
3820
3821
3822
3823
3824
3825
3826
3827
3828
3829
3830
3831
3832
3833
3834
3835
3836
3837
3838
3839
3840
3841
3842
3843
3844
3845
3846
3847
3848
3849
3850
3851
3852
3853
3854
3855
3856
3857
3858
3859
3860
3861
3862
3863
3864
3865
3866
3867
3868
3869
3870
3871
3872
3873
3874
3875
3876
3877
3878
3879
3880
3881
3882
3883
3884
3885
3886
3887
3888
3889
3890
3891
3892
3893
3894
3895
3896
3897
3898
3899
3900
3901
3902
3903
3904
3905
3906
3907
3908
3909
3910
3911
3912
3913
3914
3915
3916
3917
3918
3919
3920
3921
3922
3923
3924
3925
3926
3927
3928
3929
3930
3931
3932
3933
3934
3935
3936
3937
3938
3939
3940
3941
3942
3943
3944
3945
3946
3947
3948
3949
3950
3951
3952
3953
3954
3955
3956
3957
3958
3959
3960
3961
3962
3963
3964
3965
3966
3967
3968
3969
3970
3971
3972
3973
3974
3975
3976
3977
3978
3979
3980
3981
3982
3983
3984
3985
3986
3987
3988
3989
3990
3991
3992
3993
3994
3995
3996
3997
3998
3999
4000
4001
4002
4003
4004
4005
4006
4007
4008
4009
4010
4011
4012
4013
4014
4015
4016
4017
4018
4019
4020
4021
4022
4023
4024
4025
4026
4027
4028
4029
4030
4031
4032
4033
4034
4035
4036
4037
4038
4039
4040
4041
4042
4043
4044
4045
4046
4047
4048
4049
4050
4051
4052
4053
4054
4055
4056
4057
4058
4059
4060
4061
4062
4063
4064
4065
4066
4067
4068
4069
4070
4071
4072
4073
4074
4075
4076
4077
4078
4079
4080
4081
4082
4083
4084
4085
4086
4087
4088
4089
4090
4091
4092
4093
4094
4095
4096
4097
4098
4099
4100
4101
4102
4103
4104
4105
4106
4107
4108
4109
4110
4111
4112
4113
4114
4115
4116
4117
4118
4119
4120
4121
4122
4123
4124
4125
4126
4127
4128
4129
4130
4131
4132
4133
4134
4135
4136
4137
4138
4139
4140
4141
4142
4143
4144
4145
4146
4147
4148
4149
4150
4151
4152
4153
4154
4155
4156
4157
4158
4159
4160
4161
4162
4163
4164
4165
4166
4167
4168
4169
4170
4171
4172
4173
4174
4175
4176
4177
4178
4179
4180
4181
4182
4183
4184
4185
4186
4187
4188
4189
4190
4191
4192
4193
4194
4195
4196
4197
4198
4199
4200
4201
4202
4203
4204
4205
4206
4207
4208
4209
4210
4211
4212
4213
4214
4215
4216
4217
4218
4219
4220
4221
4222
4223
4224
4225
4226
4227
4228
4229
4230
4231
4232
4233
4234
4235
4236
4237
4238
4239
4240
4241
4242
4243
4244
4245
4246
4247
4248
4249
4250
4251
4252
4253
4254
4255
4256
4257
4258
4259
4260
4261
4262
4263
4264
4265
4266
4267
4268
4269
4270
4271
4272
4273
4274
4275
4276
4277
4278
4279
4280
4281
4282
4283
4284
4285
4286
4287
4288
4289
4290
4291
4292
4293
4294
4295
4296
4297
4298
4299
4300
4301
4302
4303
4304
4305
4306
4307
4308
4309
4310
4311
4312
4313
4314
4315
4316
4317
4318
4319
4320
4321
4322
4323
4324
4325
4326
4327
4328
4329
4330
4331
4332
4333
4334
4335
4336
4337
4338
4339
4340
4341
4342
4343
4344
4345
4346
4347
4348
4349
4350
4351
4352
4353
4354
4355
4356
4357
4358
4359
4360
4361
4362
4363
4364
4365
4366
4367
4368
4369
4370
4371
4372
4373
4374
4375
4376
4377
4378
4379
4380
4381
4382
4383
4384
4385
4386
4387
4388
4389
4390
4391
4392
4393
4394
4395
4396
4397
4398
4399
4400
4401
4402
4403
4404
4405
4406
4407
4408
4409
4410
4411
4412
4413
4414
4415
4416
4417
4418
4419
4420
4421
4422
4423
4424
4425
4426
4427
4428
4429
4430
4431
4432
4433
4434
4435
4436
4437
4438
4439
4440
4441
4442
4443
4444
4445
4446
4447
4448
4449
4450
4451
4452
4453
4454
4455
4456
4457
4458
4459
4460
4461
4462
4463
4464
4465
4466
4467
4468
4469
4470
4471
4472
4473
4474
4475
4476
4477
4478
4479
4480
4481
4482
4483
4484
4485
4486
4487
4488
4489
4490
4491
4492
4493
4494
4495
4496
4497
4498
4499
4500
4501
4502
4503
4504
4505
4506
4507
4508
4509
4510
4511
4512
4513
4514
4515
4516
4517
4518
4519
4520
4521
4522
4523
4524
4525
4526
4527
4528
4529
4530
4531
4532
4533
4534
4535
4536
4537
4538
4539
4540
4541
4542
4543
4544
4545
4546
4547
4548
4549
4550
4551
4552
4553
4554
4555
4556
4557
4558
4559
4560
4561
4562
4563
4564
4565
4566
4567
4568
4569
4570
4571
4572
4573
4574
4575
4576
4577
4578
4579
4580
4581
4582
4583
4584
4585
4586
4587
4588
4589
4590
4591
4592
4593
4594
4595
4596
4597
4598
4599
4600
4601
4602
4603
4604
4605
4606
4607
4608
4609
4610
4611
4612
4613
4614
4615
4616
4617
4618
4619
4620
4621
4622
4623
4624
4625
4626
4627
4628
4629
4630
4631
4632
4633
4634
4635
4636
4637
4638
4639
4640
4641
4642
4643
4644
4645
4646
4647
4648
4649
4650
4651
4652
4653
4654
4655
4656
4657
4658
4659
4660
4661
4662
4663
4664
4665
4666
4667
4668
4669
4670
4671
4672
4673
4674
4675
4676
4677
4678
4679
4680
4681
4682
4683
4684
4685
4686
4687
4688
4689
4690
4691
4692
4693
4694
4695
4696
4697
4698
4699
4700
4701
4702
4703
4704
4705
4706
4707
4708
4709
4710
4711
4712
4713
4714
4715
4716
4717
4718
4719
4720
4721
4722
4723
4724
4725
4726
4727
4728
4729
4730
4731
4732
4733
4734
4735
4736
4737
4738
4739
4740
4741
4742
4743
4744
4745
4746
4747
4748
4749
4750
4751
4752
4753
4754
4755
4756
4757
4758
4759
4760
4761
4762
4763
4764
4765
4766
4767
4768
4769
4770
4771
4772
4773
4774
4775
4776
4777
4778
4779
4780
4781
4782
4783
4784
4785
4786
4787
4788
4789
4790
4791
4792
4793
4794
4795
4796
4797
4798
4799
4800
4801
4802
4803
4804
4805
4806
4807
4808
4809
4810
4811
4812
4813
4814
4815
4816
4817
4818
4819
4820
4821
4822
4823
4824
4825
4826
4827
4828
4829
4830
4831
4832
4833
4834
4835
4836
4837
4838
4839
4840
4841
4842
4843
4844
4845
4846
4847
4848
4849
4850
4851
4852
4853
4854
4855
4856
4857
4858
4859
4860
4861
4862
4863
4864
4865
4866
4867
4868
4869
4870
4871
4872
4873
4874
4875
4876
4877
4878
4879
4880
4881
4882
4883
4884
4885
4886
4887
4888
4889
4890
4891
4892
4893
4894
4895
4896
4897
4898
4899
4900
4901
4902
4903
4904
4905
4906
4907
4908
4909
4910
4911
4912
4913
4914
4915
4916
4917
4918
4919
4920
4921
4922
4923
4924
4925
4926
4927
4928
4929
4930
4931
4932
4933
4934
4935
4936
4937
4938
4939
4940
4941
4942
4943
4944
4945
4946
4947
4948
4949
4950
4951
4952
4953
4954
4955
4956
4957
4958
4959
4960
4961
4962
4963
4964
4965
4966
4967
4968
4969
4970
4971
4972
4973
4974
4975
4976
4977
4978
4979
4980
4981
4982
4983
4984
4985
4986
4987
4988
4989
4990
4991
4992
4993
4994
4995
4996
4997
4998
4999
5000
5001
5002
5003
5004
5005
5006
5007
5008
5009
5010
5011
5012
5013
5014
5015
5016
5017
5018
5019
5020
5021
5022
5023
5024
5025
5026
5027
5028
5029
5030
5031
5032
5033
5034
5035
5036
5037
5038
5039
5040
5041
5042
5043
5044
5045
5046
5047
5048
5049
5050
5051
5052
5053
5054
5055
5056
5057
5058
5059
5060
5061
5062
5063
5064
5065
5066
5067
5068
5069
5070
5071
5072
5073
5074
5075
5076
5077
5078
5079
5080
5081
5082
5083
5084
5085
5086
5087
5088
5089
5090
5091
5092
5093
5094
5095
5096
5097
5098
5099
5100
5101
5102
5103
5104
5105
5106
5107
5108
5109
5110
5111
5112
5113
5114
5115
5116
5117
5118
5119
5120
5121
5122
5123
5124
5125
5126
5127
5128
5129
5130
5131
5132
5133
5134
5135
5136
5137
5138
5139
5140
5141
5142
5143
5144
5145
5146
5147
5148
5149
5150
5151
5152
5153
5154
5155
5156
5157
5158
5159
5160
5161
5162
5163
5164
5165
5166
5167
5168
5169
5170
5171
5172
5173
5174
5175
5176
5177
5178
5179
5180
5181
5182
5183
5184
5185
5186
5187
5188
5189
5190
5191
5192
5193
5194
5195
5196
5197
5198
5199
5200
5201
5202
5203
5204
5205
5206
5207
5208
5209
5210
5211
5212
5213
5214
5215
5216
5217
5218
5219
5220
5221
5222
5223
5224
5225
5226
5227
5228
5229
5230
5231
5232
5233
5234
5235
5236
5237
5238
5239
5240
5241
5242
5243
5244
5245
5246
5247
5248
5249
5250
5251
5252
5253
5254
5255
5256
5257
5258
5259
5260
5261
5262
5263
5264
5265
5266
5267
5268
5269
5270
5271
5272
5273
5274
5275
5276
5277
5278
5279
5280
5281
5282
5283
5284
5285
5286
5287
5288
5289
5290
5291
5292
5293
5294
5295
5296
5297
5298
5299
5300
5301
5302
5303
5304
5305
5306
5307
5308
5309
5310
5311
5312
5313
5314
5315
5316
5317
5318
5319
5320
5321
5322
5323
5324
5325
5326
5327
5328
5329
5330
5331
5332
5333
5334
5335
5336
5337
5338
5339
5340
5341
5342
5343
5344
5345
5346
5347
5348
5349
5350
5351
5352
5353
5354
5355
5356
5357
5358
5359
5360
5361
5362
5363
5364
5365
5366
5367
5368
5369
5370
5371
5372
5373
5374
5375
5376
5377
5378
5379
5380
5381
5382
5383
5384
5385
5386
5387
5388
5389
5390
5391
5392
5393
5394
5395
5396
5397
5398
5399
5400
5401
5402
5403
5404
5405
5406
5407
5408
5409
5410
5411
5412
5413
5414
5415
5416
5417
5418
5419
5420
5421
5422
5423
5424
5425
5426
5427
5428
5429
5430
5431
5432
5433
5434
5435
5436
5437
5438
5439
5440
5441
5442
5443
5444
5445
5446
5447
5448
5449
5450
5451
5452
5453
5454
5455
5456
5457
5458
5459
5460
5461
5462
5463
5464
5465
5466
5467
5468
5469
5470
5471
5472
5473
5474
5475
5476
5477
5478
5479
5480
5481
5482
5483
5484
5485
5486
5487
5488
5489
5490
5491
5492
5493
5494
5495
5496
5497
5498
5499
5500
5501
5502
5503
5504
5505
5506
5507
5508
5509
5510
5511
5512
5513
5514
5515
5516
5517
5518
5519
5520
5521
5522
5523
5524
5525
5526
5527
5528
5529
5530
5531
5532
5533
5534
5535
5536
5537
5538
5539
5540
5541
5542
5543
5544
5545
5546
5547
5548
5549
5550
5551
5552
5553
5554
5555
5556
5557
5558
5559
5560
5561
5562
5563
5564
5565
5566
5567
5568
5569
5570
5571
5572
5573
5574
5575
5576
5577
5578
5579
5580
5581
5582
5583
5584
5585
5586
5587
5588
5589
5590
5591
5592
5593
5594
5595
5596
5597
5598
5599
5600
5601
5602
5603
5604
5605
5606
5607
5608
5609
5610
5611
5612
5613
5614
5615
5616
5617
5618
5619
5620
5621
5622
5623
5624
5625
5626
5627
5628
5629
5630
5631
5632
5633
5634
5635
5636
5637
5638
5639
5640
5641
5642
5643
5644
5645
5646
5647
5648
5649
5650
5651
5652
5653
5654
5655
5656
5657
5658
5659
5660
5661
5662
5663
5664
5665
5666
5667
5668
5669
5670
5671
5672
5673
5674
5675
5676
5677
5678
5679
5680
5681
5682
5683
5684
5685
5686
5687
5688
5689
5690
5691
5692
5693
5694
5695
5696
5697
5698
5699
5700
5701
5702
5703
5704
5705
5706
5707
5708
5709
5710
5711
5712
5713
5714
5715
5716
5717
5718
5719
5720
5721
5722
5723
5724
5725
5726
5727
5728
5729
5730
5731
5732
5733
5734
5735
5736
5737
5738
5739
5740
5741
5742
5743
5744
5745
5746
5747
5748
5749
5750
5751
5752
5753
5754
5755
5756
5757
5758
5759
5760
5761
5762
5763
5764
5765
5766
5767
5768
5769
5770
5771
5772
5773
5774
5775
5776
5777
5778
5779
5780
5781
5782
5783
5784
5785
5786
5787
5788
5789
5790
5791
5792
5793
5794
5795
5796
5797
5798
5799
5800
5801
5802
5803
5804
5805
5806
5807
5808
5809
5810
5811
5812
5813
5814
5815
5816
5817
5818
5819
5820
5821
5822
5823
5824
5825
5826
5827
5828
5829
5830
5831
5832
5833
5834
5835
5836
5837
5838
5839
5840
5841
5842
5843
5844
5845
5846
5847
5848
5849
5850
5851
5852
5853
5854
5855
5856
5857
5858
5859
5860
5861
5862
5863
5864
5865
5866
5867
5868
5869
5870
5871
5872
5873
5874
5875
5876
5877
5878
5879
5880
5881
5882
5883
5884
5885
5886
5887
5888
5889
5890
5891
5892
5893
5894
5895
5896
5897
5898
5899
5900
5901
5902
5903
5904
5905
5906
5907
5908
5909
5910
5911
5912
5913
5914
5915
5916
5917
5918
5919
5920
5921
5922
5923
5924
5925
5926
5927
5928
5929
5930
5931
5932
5933
5934
5935
5936
5937
5938
5939
5940
5941
5942
5943
5944
5945
5946
5947
5948
5949
5950
5951
5952
5953
5954
5955
5956
5957
5958
5959
5960
5961
5962
5963
5964
5965
5966
5967
5968
5969
5970
5971
5972
5973
5974
5975
5976
5977
5978
5979
5980
5981
5982
5983
5984
5985
5986
5987
5988
5989
5990
5991
5992
5993
5994
5995
5996
5997
5998
5999
6000
6001
6002
6003
6004
6005
6006
6007
6008
6009
6010
6011
6012
6013
6014
6015
6016
6017
6018
6019
6020
6021
6022
6023
6024
6025
6026
6027
6028
6029
6030
6031
6032
6033
6034
6035
6036
6037
6038
6039
6040
6041
6042
6043
6044
6045
6046
6047
6048
6049
6050
6051
6052
6053
6054
6055
6056
6057
6058
6059
6060
6061
6062
6063
6064
6065
6066
6067
6068
6069
6070
6071
6072
6073
6074
6075
6076
6077
6078
6079
6080
6081
6082
6083
6084
6085
6086
6087
6088
6089
6090
6091
6092
6093
6094
6095
6096
6097
6098
6099
6100
6101
6102
6103
6104
6105
6106
6107
6108
6109
6110
6111
6112
6113
6114
6115
6116
6117
6118
6119
6120
6121
6122
6123
6124
6125
6126
6127
6128
6129
6130
6131
6132
6133
6134
6135
6136
6137
6138
6139
6140
6141
6142
6143
6144
6145
6146
6147
6148
6149
6150
6151
6152
6153
6154
6155
6156
6157
6158
6159
6160
6161
6162
6163
6164
6165
6166
6167
6168
6169
6170
6171
6172
6173
6174
6175
6176
6177
6178
6179
6180
6181
6182
6183
6184
6185
6186
6187
6188
6189
6190
6191
6192
6193
6194
6195
6196
6197
6198
6199
6200
6201
6202
6203
6204
6205
6206
6207
6208
6209
6210
6211
6212
6213
6214
6215
6216
6217
6218
6219
6220
6221
6222
6223
6224
6225
6226
6227
6228
6229
6230
6231
6232
6233
6234
6235
6236
6237
6238
6239
6240
6241
6242
6243
6244
6245
6246
6247
6248
6249
6250
6251
6252
6253
6254
6255
6256
6257
6258
6259
6260
6261
6262
6263
6264
6265
6266
6267
6268
6269
6270
6271
6272
6273
6274
6275
6276
6277
6278
6279
6280
6281
6282
6283
6284
6285
6286
6287
6288
6289
6290
6291
6292
6293
6294
6295
6296
6297
6298
6299
6300
6301
6302
6303
6304
6305
6306
6307
6308
6309
6310
6311
6312
6313
6314
6315
6316
6317
6318
6319
6320
6321
6322
6323
6324
6325
6326
6327
6328
6329
6330
6331
6332
6333
6334
6335
6336
6337
6338
6339
6340
6341
6342
6343
6344
6345
6346
6347
6348
6349
6350
6351
6352
6353
6354
6355
6356
6357
6358
6359
6360
6361
6362
6363
6364
6365
6366
6367
6368
6369
6370
6371
6372
6373
6374
6375
6376
6377
6378
6379
6380
6381
6382
6383
6384
6385
6386
6387
6388
6389
6390
6391
6392
6393
6394
6395
6396
6397
6398
6399
6400
6401
6402
6403
6404
6405
6406
6407
6408
6409
6410
6411
6412
6413
6414
6415
6416
6417
6418
6419
6420
6421
6422
6423
6424
6425
6426
6427
6428
6429
6430
6431
6432
6433
6434
6435
6436
6437
6438
6439
6440
6441
6442
6443
6444
6445
6446
6447
6448
6449
6450
6451
6452
6453
6454
6455
6456
6457
6458
6459
6460
6461
6462
6463
6464
6465
6466
6467
6468
6469
6470
6471
6472
6473
6474
6475
6476
6477
6478
6479
6480
6481
6482
6483
6484
6485
6486
6487
6488
6489
6490
6491
6492
6493
6494
6495
6496
6497
6498
6499
6500
6501
6502
6503
6504
6505
6506
6507
6508
6509
6510
6511
6512
6513
6514
6515
6516
6517
6518
6519
6520
6521
6522
6523
6524
6525
6526
6527
6528
6529
6530
6531
6532
6533
6534
6535
6536
6537
6538
6539
6540
6541
6542
6543
6544
6545
6546
6547
6548
6549
6550
6551
6552
6553
6554
6555
6556
6557
6558
6559
6560
6561
6562
6563
6564
6565
6566
6567
6568
6569
6570
6571
6572
6573
6574
6575
6576
6577
6578
6579
6580
6581
6582
6583
6584
6585
6586
6587
6588
6589
6590
6591
6592
6593
6594
6595
6596
6597
6598
6599
6600
6601
6602
6603
6604
6605
6606
6607
6608
6609
6610
6611
6612
6613
6614
6615
6616
6617
6618
6619
6620
6621
6622
6623
6624
6625
6626
6627
6628
6629
6630
6631
6632
6633
6634
6635
6636
6637
6638
6639
6640
6641
6642
6643
6644
6645
6646
6647
6648
6649
6650
6651
6652
6653
6654
6655
6656
6657
6658
6659
6660
6661
6662
6663
6664
6665
6666
6667
6668
6669
6670
6671
6672
6673
6674
6675
6676
6677
6678
6679
6680
6681
6682
6683
6684
6685
6686
6687
6688
6689
6690
6691
6692
6693
6694
6695
6696
6697
6698
6699
6700
6701
6702
6703
6704
6705
6706
6707
6708
6709
6710
6711
6712
6713
6714
6715
6716
6717
6718
6719
6720
6721
6722
6723
6724
6725
6726
6727
6728
6729
6730
6731
6732
6733
6734
6735
6736
6737
6738
6739
6740
6741
6742
6743
6744
6745
6746
6747
6748
6749
6750
6751
6752
6753
6754
6755
6756
6757
6758
6759
6760
6761
6762
6763
6764
6765
6766
6767
6768
6769
6770
6771
6772
6773
6774
6775
6776
6777
6778
6779
6780
6781
6782
6783
6784
6785
6786
6787
6788
6789
6790
6791
6792
6793
6794
6795
6796
6797
6798
6799
6800
6801
6802
6803
6804
6805
6806
6807
6808
6809
6810
6811
6812
6813
6814
6815
6816
6817
6818
6819
6820
6821
6822
6823
6824
6825
6826
6827
6828
6829
6830
6831
6832
6833
6834
6835
6836
6837
6838
6839
6840
6841
6842
6843
6844
6845
6846
/**********************************************************
 * Copyright 1998-2013 VMware, Inc.  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 THE AUTHORS OR COPYRIGHT HOLDERS
 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
 *
 **********************************************************/

/**
 * @file svga_tgsi_vgpu10.c
 *
 * TGSI -> VGPU10 shader translation.
 *
 * \author Mingcheng Chen
 * \author Brian Paul
 */

#include "pipe/p_compiler.h"
#include "pipe/p_shader_tokens.h"
#include "pipe/p_defines.h"
#include "tgsi/tgsi_build.h"
#include "tgsi/tgsi_dump.h"
#include "tgsi/tgsi_info.h"
#include "tgsi/tgsi_parse.h"
#include "tgsi/tgsi_scan.h"
#include "tgsi/tgsi_two_side.h"
#include "tgsi/tgsi_aa_point.h"
#include "tgsi/tgsi_util.h"
#include "util/u_math.h"
#include "util/u_memory.h"
#include "util/u_bitmask.h"
#include "util/u_debug.h"
#include "util/u_pstipple.h"

#include "svga_context.h"
#include "svga_debug.h"
#include "svga_link.h"
#include "svga_shader.h"
#include "svga_tgsi.h"

#include "VGPU10ShaderTokens.h"


#define INVALID_INDEX 99999
#define MAX_INTERNAL_TEMPS 3
#define MAX_SYSTEM_VALUES 4
#define MAX_IMMEDIATE_COUNT \
        (VGPU10_MAX_IMMEDIATE_CONSTANT_BUFFER_ELEMENT_COUNT/4)
#define MAX_TEMP_ARRAYS 64  /* Enough? */


/**
 * Clipping is complicated.  There's four different cases which we
 * handle during VS/GS shader translation:
 */
enum clipping_mode
{
   CLIP_NONE,     /**< No clipping enabled */
   CLIP_LEGACY,   /**< The shader has no clipping declarations or code but
                   * one or more user-defined clip planes are enabled.  We
                   * generate extra code to emit clip distances.
                   */
   CLIP_DISTANCE, /**< The shader already declares clip distance output
                   * registers and has code to write to them.
                   */
   CLIP_VERTEX    /**< The shader declares a clip vertex output register and
                  * has code that writes to the register.  We convert the
                  * clipvertex position into one or more clip distances.
                  */
};


struct svga_shader_emitter_v10
{
   /* The token output buffer */
   unsigned size;
   char *buf;
   char *ptr;

   /* Information about the shader and state (does not change) */
   struct svga_compile_key key;
   struct tgsi_shader_info info;
   unsigned unit;

   unsigned inst_start_token;
   boolean discard_instruction; /**< throw away current instruction? */

   union tgsi_immediate_data immediates[MAX_IMMEDIATE_COUNT][4];
   unsigned num_immediates;      /**< Number of immediates emitted */
   unsigned common_immediate_pos[8];  /**< literals for common immediates */
   unsigned num_common_immediates;
   boolean immediates_emitted;

   unsigned num_outputs;      /**< include any extra outputs */
                              /**  The first extra output is reserved for
                               *   non-adjusted vertex position for
                               *   stream output purpose
                               */

   /* Temporary Registers */
   unsigned num_shader_temps; /**< num of temps used by original shader */
   unsigned internal_temp_count;  /**< currently allocated internal temps */
   struct {
      unsigned start, size;
   } temp_arrays[MAX_TEMP_ARRAYS];
   unsigned num_temp_arrays;

   /** Map TGSI temp registers to VGPU10 temp array IDs and indexes */
   struct {
      unsigned arrayId, index;
   } temp_map[VGPU10_MAX_TEMPS]; /**< arrayId, element */

   /** Number of constants used by original shader for each constant buffer.
    * The size should probably always match with that of svga_state.constbufs.
    */
   unsigned num_shader_consts[SVGA_MAX_CONST_BUFS];

   /* Samplers */
   unsigned num_samplers;

   /* Address regs (really implemented with temps) */
   unsigned num_address_regs;
   unsigned address_reg_index[MAX_VGPU10_ADDR_REGS];

   /* Output register usage masks */
   ubyte output_usage_mask[PIPE_MAX_SHADER_OUTPUTS];

   /* To map TGSI system value index to VGPU shader input indexes */
   ubyte system_value_indexes[MAX_SYSTEM_VALUES];

   struct {
      /* vertex position scale/translation */
      unsigned out_index;  /**< the real position output reg */
      unsigned tmp_index;  /**< the fake/temp position output reg */
      unsigned so_index;   /**< the non-adjusted position output reg */
      unsigned prescale_scale_index, prescale_trans_index;
      boolean  need_prescale;
   } vposition;

   /* For vertex shaders only */
   struct {
      /* viewport constant */
      unsigned viewport_index;

      /* temp index of adjusted vertex attributes */
      unsigned adjusted_input[PIPE_MAX_SHADER_INPUTS];
   } vs;

   /* For fragment shaders only */
   struct {
      /* apha test */
      unsigned color_out_index[PIPE_MAX_COLOR_BUFS];  /**< the real color output regs */
      unsigned color_tmp_index;  /**< fake/temp color output reg */
      unsigned alpha_ref_index;  /**< immediate constant for alpha ref */

      /* front-face */
      unsigned face_input_index; /**< real fragment shader face reg (bool) */
      unsigned face_tmp_index;   /**< temp face reg converted to -1 / +1 */

      unsigned pstipple_sampler_unit;

      unsigned fragcoord_input_index;  /**< real fragment position input reg */
      unsigned fragcoord_tmp_index;    /**< 1/w modified position temp reg */
   } fs;

   /* For geometry shaders only */
   struct {
      VGPU10_PRIMITIVE prim_type;/**< VGPU10 primitive type */
      VGPU10_PRIMITIVE_TOPOLOGY prim_topology; /**< VGPU10 primitive topology */
      unsigned input_size;       /**< size of input arrays */
      unsigned prim_id_index;    /**< primitive id register index */
      unsigned max_out_vertices; /**< maximum number of output vertices */
   } gs;

   /* For vertex or geometry shaders */
   enum clipping_mode clip_mode;
   unsigned clip_dist_out_index; /**< clip distance output register index */
   unsigned clip_dist_tmp_index; /**< clip distance temporary register */
   unsigned clip_dist_so_index;  /**< clip distance shadow copy */

   /** Index of temporary holding the clipvertex coordinate */
   unsigned clip_vertex_out_index; /**< clip vertex output register index */
   unsigned clip_vertex_tmp_index; /**< clip vertex temporary index */

   /* user clip plane constant slot indexes */
   unsigned clip_plane_const[PIPE_MAX_CLIP_PLANES];

   unsigned num_output_writes;
   boolean constant_color_output;

   boolean uses_flat_interp;

   /* For all shaders: const reg index for RECT coord scaling */
   unsigned texcoord_scale_index[PIPE_MAX_SAMPLERS];

   /* For all shaders: const reg index for texture buffer size */
   unsigned texture_buffer_size_index[PIPE_MAX_SAMPLERS];

   /* VS/GS/FS Linkage info */
   struct shader_linkage linkage;

   bool register_overflow;  /**< Set if we exceed a VGPU10 register limit */
};


static boolean
emit_post_helpers(struct svga_shader_emitter_v10 *emit);

static boolean
emit_vertex(struct svga_shader_emitter_v10 *emit,
            const struct tgsi_full_instruction *inst);

static char err_buf[128];

static boolean
expand(struct svga_shader_emitter_v10 *emit)
{
   char *new_buf;
   unsigned newsize = emit->size * 2;

   if (emit->buf != err_buf)
      new_buf = REALLOC(emit->buf, emit->size, newsize);
   else
      new_buf = NULL;

   if (!new_buf) {
      emit->ptr = err_buf;
      emit->buf = err_buf;
      emit->size = sizeof(err_buf);
      return FALSE;
   }

   emit->size = newsize;
   emit->ptr = new_buf + (emit->ptr - emit->buf);
   emit->buf = new_buf;
   return TRUE;
}

/**
 * Create and initialize a new svga_shader_emitter_v10 object.
 */
static struct svga_shader_emitter_v10 *
alloc_emitter(void)
{
   struct svga_shader_emitter_v10 *emit = CALLOC(1, sizeof(*emit));

   if (!emit)
      return NULL;

   /* to initialize the output buffer */
   emit->size = 512;
   if (!expand(emit)) {
      FREE(emit);
      return NULL;
   }
   return emit;
}

/**
 * Free an svga_shader_emitter_v10 object.
 */
static void
free_emitter(struct svga_shader_emitter_v10 *emit)
{
   assert(emit);
   FREE(emit->buf);    /* will be NULL if translation succeeded */
   FREE(emit);
}

static inline boolean
reserve(struct svga_shader_emitter_v10 *emit,
        unsigned nr_dwords)
{
   while (emit->ptr - emit->buf + nr_dwords * sizeof(uint32) >= emit->size) {
      if (!expand(emit))
         return FALSE;
   }

   return TRUE;
}

static boolean
emit_dword(struct svga_shader_emitter_v10 *emit, uint32 dword)
{
   if (!reserve(emit, 1))
      return FALSE;

   *(uint32 *)emit->ptr = dword;
   emit->ptr += sizeof dword;
   return TRUE;
}

static boolean
emit_dwords(struct svga_shader_emitter_v10 *emit,
            const uint32 *dwords,
            unsigned nr)
{
   if (!reserve(emit, nr))
      return FALSE;

   memcpy(emit->ptr, dwords, nr * sizeof *dwords);
   emit->ptr += nr * sizeof *dwords;
   return TRUE;
}

/** Return the number of tokens in the emitter's buffer */
static unsigned
emit_get_num_tokens(const struct svga_shader_emitter_v10 *emit)
{
   return (emit->ptr - emit->buf) / sizeof(unsigned);
}


/**
 * Check for register overflow.  If we overflow we'll set an
 * error flag.  This function can be called for register declarations
 * or use as src/dst instruction operands.
 * \param type  register type.  One of VGPU10_OPERAND_TYPE_x
                or VGPU10_OPCODE_DCL_x
 * \param index  the register index
 */
static void
check_register_index(struct svga_shader_emitter_v10 *emit,
                     unsigned operandType, unsigned index)
{
   bool overflow_before = emit->register_overflow;

   switch (operandType) {
   case VGPU10_OPERAND_TYPE_TEMP:
   case VGPU10_OPERAND_TYPE_INDEXABLE_TEMP:
   case VGPU10_OPCODE_DCL_TEMPS:
      if (index >= VGPU10_MAX_TEMPS) {
         emit->register_overflow = TRUE;
      }
      break;
   case VGPU10_OPERAND_TYPE_CONSTANT_BUFFER:
   case VGPU10_OPCODE_DCL_CONSTANT_BUFFER:
      if (index >= VGPU10_MAX_CONSTANT_BUFFER_ELEMENT_COUNT) {
         emit->register_overflow = TRUE;
      }
      break;
   case VGPU10_OPERAND_TYPE_INPUT:
   case VGPU10_OPERAND_TYPE_INPUT_PRIMITIVEID:
   case VGPU10_OPCODE_DCL_INPUT:
   case VGPU10_OPCODE_DCL_INPUT_SGV:
   case VGPU10_OPCODE_DCL_INPUT_SIV:
   case VGPU10_OPCODE_DCL_INPUT_PS:
   case VGPU10_OPCODE_DCL_INPUT_PS_SGV:
   case VGPU10_OPCODE_DCL_INPUT_PS_SIV:
      if ((emit->unit == PIPE_SHADER_VERTEX &&
           index >= VGPU10_MAX_VS_INPUTS) ||
          (emit->unit == PIPE_SHADER_GEOMETRY &&
           index >= VGPU10_MAX_GS_INPUTS) ||
          (emit->unit == PIPE_SHADER_FRAGMENT &&
           index >= VGPU10_MAX_FS_INPUTS)) {
         emit->register_overflow = TRUE;
      }
      break;
   case VGPU10_OPERAND_TYPE_OUTPUT:
   case VGPU10_OPCODE_DCL_OUTPUT:
   case VGPU10_OPCODE_DCL_OUTPUT_SGV:
   case VGPU10_OPCODE_DCL_OUTPUT_SIV:
      if ((emit->unit == PIPE_SHADER_VERTEX &&
           index >= VGPU10_MAX_VS_OUTPUTS) ||
          (emit->unit == PIPE_SHADER_GEOMETRY &&
           index >= VGPU10_MAX_GS_OUTPUTS) ||
          (emit->unit == PIPE_SHADER_FRAGMENT &&
           index >= VGPU10_MAX_FS_OUTPUTS)) {
         emit->register_overflow = TRUE;
      }
      break;
   case VGPU10_OPERAND_TYPE_SAMPLER:
   case VGPU10_OPCODE_DCL_SAMPLER:
      if (index >= VGPU10_MAX_SAMPLERS) {
         emit->register_overflow = TRUE;
      }
      break;
   case VGPU10_OPERAND_TYPE_RESOURCE:
   case VGPU10_OPCODE_DCL_RESOURCE:
      if (index >= VGPU10_MAX_RESOURCES) {
         emit->register_overflow = TRUE;
      }
      break;
   case VGPU10_OPERAND_TYPE_IMMEDIATE_CONSTANT_BUFFER:
      if (index >= MAX_IMMEDIATE_COUNT) {
         emit->register_overflow = TRUE;
      }
      break;
   default:
      assert(0);
      ; /* nothing */
   }

   if (emit->register_overflow && !overflow_before) {
      debug_printf("svga: vgpu10 register overflow (reg %u, index %u)\n",
                   operandType, index);
   }
}


/**
 * Examine misc state to determine the clipping mode.
 */
static void
determine_clipping_mode(struct svga_shader_emitter_v10 *emit)
{
   if (emit->info.num_written_clipdistance > 0) {
      emit->clip_mode = CLIP_DISTANCE;
   }
   else if (emit->info.writes_clipvertex) {
      emit->clip_mode = CLIP_VERTEX;
   }
   else if (emit->key.clip_plane_enable) {
      emit->clip_mode = CLIP_LEGACY;
   }
   else {
      emit->clip_mode = CLIP_NONE;
   }
}


/**
 * For clip distance register declarations and clip distance register
 * writes we need to mask the declaration usage or instruction writemask
 * (respectively) against the set of the really-enabled clipping planes.
 *
 * The piglit test spec/glsl-1.30/execution/clipping/vs-clip-distance-enables
 * has a VS that writes to all 8 clip distance registers, but the plane enable
 * flags are a subset of that.
 *
 * This function is used to apply the plane enable flags to the register
 * declaration or instruction writemask.
 *
 * \param writemask  the declaration usage mask or instruction writemask
 * \param clip_reg_index  which clip plane register is being declared/written.
 *                        The legal values are 0 and 1 (two clip planes per
 *                        register, for a total of 8 clip planes)
 */
static unsigned
apply_clip_plane_mask(struct svga_shader_emitter_v10 *emit,
                      unsigned writemask, unsigned clip_reg_index)
{
   unsigned shift;

   assert(clip_reg_index < 2);

   /* four clip planes per clip register: */
   shift = clip_reg_index * 4;
   writemask &= ((emit->key.clip_plane_enable >> shift) & 0xf);

   return writemask;
}


/**
 * Translate gallium shader type into VGPU10 type.
 */
static VGPU10_PROGRAM_TYPE
translate_shader_type(unsigned type)
{
   switch (type) {
   case PIPE_SHADER_VERTEX:
      return VGPU10_VERTEX_SHADER;
   case PIPE_SHADER_GEOMETRY:
      return VGPU10_GEOMETRY_SHADER;
   case PIPE_SHADER_FRAGMENT:
      return VGPU10_PIXEL_SHADER;
   default:
      assert(!"Unexpected shader type");
      return VGPU10_VERTEX_SHADER;
   }
}


/**
 * Translate a TGSI_OPCODE_x into a VGPU10_OPCODE_x
 * Note: we only need to translate the opcodes for "simple" instructions,
 * as seen below.  All other opcodes are handled/translated specially.
 */
static VGPU10_OPCODE_TYPE
translate_opcode(unsigned opcode)
{
   switch (opcode) {
   case TGSI_OPCODE_MOV:
      return VGPU10_OPCODE_MOV;
   case TGSI_OPCODE_MUL:
      return VGPU10_OPCODE_MUL;
   case TGSI_OPCODE_ADD:
      return VGPU10_OPCODE_ADD;
   case TGSI_OPCODE_DP3:
      return VGPU10_OPCODE_DP3;
   case TGSI_OPCODE_DP4:
      return VGPU10_OPCODE_DP4;
   case TGSI_OPCODE_MIN:
      return VGPU10_OPCODE_MIN;
   case TGSI_OPCODE_MAX:
      return VGPU10_OPCODE_MAX;
   case TGSI_OPCODE_MAD:
      return VGPU10_OPCODE_MAD;
   case TGSI_OPCODE_SQRT:
      return VGPU10_OPCODE_SQRT;
   case TGSI_OPCODE_FRC:
      return VGPU10_OPCODE_FRC;
   case TGSI_OPCODE_FLR:
      return VGPU10_OPCODE_ROUND_NI;
   case TGSI_OPCODE_FSEQ:
      return VGPU10_OPCODE_EQ;
   case TGSI_OPCODE_FSGE:
      return VGPU10_OPCODE_GE;
   case TGSI_OPCODE_FSNE:
      return VGPU10_OPCODE_NE;
   case TGSI_OPCODE_DDX:
      return VGPU10_OPCODE_DERIV_RTX;
   case TGSI_OPCODE_DDY:
      return VGPU10_OPCODE_DERIV_RTY;
   case TGSI_OPCODE_RET:
      return VGPU10_OPCODE_RET;
   case TGSI_OPCODE_DIV:
      return VGPU10_OPCODE_DIV;
   case TGSI_OPCODE_IDIV:
      return VGPU10_OPCODE_IDIV;
   case TGSI_OPCODE_DP2:
      return VGPU10_OPCODE_DP2;
   case TGSI_OPCODE_BRK:
      return VGPU10_OPCODE_BREAK;
   case TGSI_OPCODE_IF:
      return VGPU10_OPCODE_IF;
   case TGSI_OPCODE_ELSE:
      return VGPU10_OPCODE_ELSE;
   case TGSI_OPCODE_ENDIF:
      return VGPU10_OPCODE_ENDIF;
   case TGSI_OPCODE_CEIL:
      return VGPU10_OPCODE_ROUND_PI;
   case TGSI_OPCODE_I2F:
      return VGPU10_OPCODE_ITOF;
   case TGSI_OPCODE_NOT:
      return VGPU10_OPCODE_NOT;
   case TGSI_OPCODE_TRUNC:
      return VGPU10_OPCODE_ROUND_Z;
   case TGSI_OPCODE_SHL:
      return VGPU10_OPCODE_ISHL;
   case TGSI_OPCODE_AND:
      return VGPU10_OPCODE_AND;
   case TGSI_OPCODE_OR:
      return VGPU10_OPCODE_OR;
   case TGSI_OPCODE_XOR:
      return VGPU10_OPCODE_XOR;
   case TGSI_OPCODE_CONT:
      return VGPU10_OPCODE_CONTINUE;
   case TGSI_OPCODE_EMIT:
      return VGPU10_OPCODE_EMIT;
   case TGSI_OPCODE_ENDPRIM:
      return VGPU10_OPCODE_CUT;
   case TGSI_OPCODE_BGNLOOP:
      return VGPU10_OPCODE_LOOP;
   case TGSI_OPCODE_ENDLOOP:
      return VGPU10_OPCODE_ENDLOOP;
   case TGSI_OPCODE_ENDSUB:
      return VGPU10_OPCODE_RET;
   case TGSI_OPCODE_NOP:
      return VGPU10_OPCODE_NOP;
   case TGSI_OPCODE_BREAKC:
      return VGPU10_OPCODE_BREAKC;
   case TGSI_OPCODE_END:
      return VGPU10_OPCODE_RET;
   case TGSI_OPCODE_F2I:
      return VGPU10_OPCODE_FTOI;
   case TGSI_OPCODE_IMAX:
      return VGPU10_OPCODE_IMAX;
   case TGSI_OPCODE_IMIN:
      return VGPU10_OPCODE_IMIN;
   case TGSI_OPCODE_UDIV:
   case TGSI_OPCODE_UMOD:
   case TGSI_OPCODE_MOD:
      return VGPU10_OPCODE_UDIV;
   case TGSI_OPCODE_IMUL_HI:
      return VGPU10_OPCODE_IMUL;
   case TGSI_OPCODE_INEG:
      return VGPU10_OPCODE_INEG;
   case TGSI_OPCODE_ISHR:
      return VGPU10_OPCODE_ISHR;
   case TGSI_OPCODE_ISGE:
      return VGPU10_OPCODE_IGE;
   case TGSI_OPCODE_ISLT:
      return VGPU10_OPCODE_ILT;
   case TGSI_OPCODE_F2U:
      return VGPU10_OPCODE_FTOU;
   case TGSI_OPCODE_UADD:
      return VGPU10_OPCODE_IADD;
   case TGSI_OPCODE_U2F:
      return VGPU10_OPCODE_UTOF;
   case TGSI_OPCODE_UCMP:
      return VGPU10_OPCODE_MOVC;
   case TGSI_OPCODE_UMAD:
      return VGPU10_OPCODE_UMAD;
   case TGSI_OPCODE_UMAX:
      return VGPU10_OPCODE_UMAX;
   case TGSI_OPCODE_UMIN:
      return VGPU10_OPCODE_UMIN;
   case TGSI_OPCODE_UMUL:
   case TGSI_OPCODE_UMUL_HI:
      return VGPU10_OPCODE_UMUL;
   case TGSI_OPCODE_USEQ:
      return VGPU10_OPCODE_IEQ;
   case TGSI_OPCODE_USGE:
      return VGPU10_OPCODE_UGE;
   case TGSI_OPCODE_USHR:
      return VGPU10_OPCODE_USHR;
   case TGSI_OPCODE_USLT:
      return VGPU10_OPCODE_ULT;
   case TGSI_OPCODE_USNE:
      return VGPU10_OPCODE_INE;
   case TGSI_OPCODE_SWITCH:
      return VGPU10_OPCODE_SWITCH;
   case TGSI_OPCODE_CASE:
      return VGPU10_OPCODE_CASE;
   case TGSI_OPCODE_DEFAULT:
      return VGPU10_OPCODE_DEFAULT;
   case TGSI_OPCODE_ENDSWITCH:
      return VGPU10_OPCODE_ENDSWITCH;
   case TGSI_OPCODE_FSLT:
      return VGPU10_OPCODE_LT;
   case TGSI_OPCODE_ROUND:
      return VGPU10_OPCODE_ROUND_NE;
   default:
      assert(!"Unexpected TGSI opcode in translate_opcode()");
      return VGPU10_OPCODE_NOP;
   }
}


/**
 * Translate a TGSI register file type into a VGPU10 operand type.
 * \param array  is the TGSI_FILE_TEMPORARY register an array?
 */
static VGPU10_OPERAND_TYPE
translate_register_file(enum tgsi_file_type file, boolean array)
{
   switch (file) {
   case TGSI_FILE_CONSTANT:
      return VGPU10_OPERAND_TYPE_CONSTANT_BUFFER;
   case TGSI_FILE_INPUT:
      return VGPU10_OPERAND_TYPE_INPUT;
   case TGSI_FILE_OUTPUT:
      return VGPU10_OPERAND_TYPE_OUTPUT;
   case TGSI_FILE_TEMPORARY:
      return array ? VGPU10_OPERAND_TYPE_INDEXABLE_TEMP
                   : VGPU10_OPERAND_TYPE_TEMP;
   case TGSI_FILE_IMMEDIATE:
      /* all immediates are 32-bit values at this time so
       * VGPU10_OPERAND_TYPE_IMMEDIATE64 is not possible at this time.
       */
      return VGPU10_OPERAND_TYPE_IMMEDIATE_CONSTANT_BUFFER;
   case TGSI_FILE_SAMPLER:
      return VGPU10_OPERAND_TYPE_SAMPLER;
   case TGSI_FILE_SYSTEM_VALUE:
      return VGPU10_OPERAND_TYPE_INPUT;

   /* XXX TODO more cases to finish */

   default:
      assert(!"Bad tgsi register file!");
      return VGPU10_OPERAND_TYPE_NULL;
   }
}


/**
 * Emit a null dst register
 */
static void
emit_null_dst_register(struct svga_shader_emitter_v10 *emit)
{
   VGPU10OperandToken0 operand;

   operand.value = 0;
   operand.operandType = VGPU10_OPERAND_TYPE_NULL;
   operand.numComponents = VGPU10_OPERAND_0_COMPONENT;

   emit_dword(emit, operand.value);
}


/**
 * If the given register is a temporary, return the array ID.
 * Else return zero.
 */
static unsigned
get_temp_array_id(const struct svga_shader_emitter_v10 *emit,
                  unsigned file, unsigned index)
{
   if (file == TGSI_FILE_TEMPORARY) {
      return emit->temp_map[index].arrayId;
   }
   else {
      return 0;
   }
}


/**
 * If the given register is a temporary, convert the index from a TGSI
 * TEMPORARY index to a VGPU10 temp index.
 */
static unsigned
remap_temp_index(const struct svga_shader_emitter_v10 *emit,
                 unsigned file, unsigned index)
{
   if (file == TGSI_FILE_TEMPORARY) {
      return emit->temp_map[index].index;
   }
   else {
      return index;
   }
}


/**
 * Setup the operand0 fields related to indexing (1D, 2D, relative, etc).
 * Note: the operandType field must already be initialized.
 */
static VGPU10OperandToken0
setup_operand0_indexing(struct svga_shader_emitter_v10 *emit,
                        VGPU10OperandToken0 operand0,
                        unsigned file,
                        boolean indirect, boolean index2D,
                        unsigned tempArrayID)
{
   unsigned indexDim, index0Rep, index1Rep = VGPU10_OPERAND_INDEX_0D;

   /*
    * Compute index dimensions
    */
   if (operand0.operandType == VGPU10_OPERAND_TYPE_IMMEDIATE32 ||
       operand0.operandType == VGPU10_OPERAND_TYPE_INPUT_PRIMITIVEID) {
      /* there's no swizzle for in-line immediates */
      indexDim = VGPU10_OPERAND_INDEX_0D;
      assert(operand0.selectionMode == 0);
   }
   else {
      if (index2D ||
          tempArrayID > 0 ||
          operand0.operandType == VGPU10_OPERAND_TYPE_CONSTANT_BUFFER) {
         indexDim = VGPU10_OPERAND_INDEX_2D;
      }
      else {
         indexDim = VGPU10_OPERAND_INDEX_1D;
      }
   }

   /*
    * Compute index representations (immediate, relative, etc).
    */
   if (tempArrayID > 0) {
      assert(file == TGSI_FILE_TEMPORARY);
      /* First index is the array ID, second index is the array element */
      index0Rep = VGPU10_OPERAND_INDEX_IMMEDIATE32;
      if (indirect) {
         index1Rep = VGPU10_OPERAND_INDEX_IMMEDIATE32_PLUS_RELATIVE;
      }
      else {
         index1Rep = VGPU10_OPERAND_INDEX_IMMEDIATE32;
      }
   }
   else if (indirect) {
      if (file == TGSI_FILE_CONSTANT) {
         /* index[0] indicates which constant buffer while index[1] indicates
          * the position in the constant buffer.
          */
         index0Rep = VGPU10_OPERAND_INDEX_IMMEDIATE32;
         index1Rep = VGPU10_OPERAND_INDEX_IMMEDIATE32_PLUS_RELATIVE;
      }
      else {
         /* All other register files are 1-dimensional */
         index0Rep = VGPU10_OPERAND_INDEX_IMMEDIATE32_PLUS_RELATIVE;
      }
   }
   else {
      index0Rep = VGPU10_OPERAND_INDEX_IMMEDIATE32;
      index1Rep = VGPU10_OPERAND_INDEX_IMMEDIATE32;
   }

   operand0.indexDimension = indexDim;
   operand0.index0Representation = index0Rep;
   operand0.index1Representation = index1Rep;

   return operand0;
}


/**
 * Emit the operand for expressing an address register for indirect indexing.
 * Note that the address register is really just a temp register.
 * \param addr_reg_index  which address register to use
 */
static void
emit_indirect_register(struct svga_shader_emitter_v10 *emit,
                       unsigned addr_reg_index)
{
   unsigned tmp_reg_index;
   VGPU10OperandToken0 operand0;

   assert(addr_reg_index < MAX_VGPU10_ADDR_REGS);

   tmp_reg_index = emit->address_reg_index[addr_reg_index];

   /* operand0 is a simple temporary register, selecting one component */
   operand0.value = 0;
   operand0.operandType = VGPU10_OPERAND_TYPE_TEMP;
   operand0.numComponents = VGPU10_OPERAND_4_COMPONENT;
   operand0.indexDimension = VGPU10_OPERAND_INDEX_1D;
   operand0.index0Representation = VGPU10_OPERAND_INDEX_IMMEDIATE32;
   operand0.selectionMode = VGPU10_OPERAND_4_COMPONENT_SELECT_1_MODE;
   operand0.swizzleX = 0;
   operand0.swizzleY = 1;
   operand0.swizzleZ = 2;
   operand0.swizzleW = 3;

   emit_dword(emit, operand0.value);
   emit_dword(emit, remap_temp_index(emit, TGSI_FILE_TEMPORARY, tmp_reg_index));
}


/**
 * Translate the dst register of a TGSI instruction and emit VGPU10 tokens.
 * \param emit  the emitter context
 * \param reg  the TGSI dst register to translate
 */
static void
emit_dst_register(struct svga_shader_emitter_v10 *emit,
                  const struct tgsi_full_dst_register *reg)
{
   unsigned file = reg->Register.File;
   unsigned index = reg->Register.Index;
   const unsigned sem_name = emit->info.output_semantic_name[index];
   const unsigned sem_index = emit->info.output_semantic_index[index];
   unsigned writemask = reg->Register.WriteMask;
   const unsigned indirect = reg->Register.Indirect;
   const unsigned tempArrayId = get_temp_array_id(emit, file, index);
   const unsigned index2d = reg->Register.Dimension;
   VGPU10OperandToken0 operand0;

   if (file == TGSI_FILE_OUTPUT) {
      if (emit->unit == PIPE_SHADER_VERTEX ||
          emit->unit == PIPE_SHADER_GEOMETRY) {
         if (index == emit->vposition.out_index &&
             emit->vposition.tmp_index != INVALID_INDEX) {
            /* replace OUTPUT[POS] with TEMP[POS].  We need to store the
             * vertex position result in a temporary so that we can modify
             * it in the post_helper() code.
             */
            file = TGSI_FILE_TEMPORARY;
            index = emit->vposition.tmp_index;
         }
         else if (sem_name == TGSI_SEMANTIC_CLIPDIST &&
                  emit->clip_dist_tmp_index != INVALID_INDEX) {
            /* replace OUTPUT[CLIPDIST] with TEMP[CLIPDIST].
             * We store the clip distance in a temporary first, then
             * we'll copy it to the shadow copy and to CLIPDIST with the
             * enabled planes mask in emit_clip_distance_instructions().
             */
            file = TGSI_FILE_TEMPORARY;
            index = emit->clip_dist_tmp_index + sem_index;
         }
         else if (sem_name == TGSI_SEMANTIC_CLIPVERTEX &&
                  emit->clip_vertex_tmp_index != INVALID_INDEX) {
            /* replace the CLIPVERTEX output register with a temporary */
            assert(emit->clip_mode == CLIP_VERTEX);
            assert(sem_index == 0);
            file = TGSI_FILE_TEMPORARY;
            index = emit->clip_vertex_tmp_index;
         }
      }
      else if (emit->unit == PIPE_SHADER_FRAGMENT) {
         if (sem_name == TGSI_SEMANTIC_POSITION) {
            /* Fragment depth output register */
            operand0.value = 0;
            operand0.operandType = VGPU10_OPERAND_TYPE_OUTPUT_DEPTH;
            operand0.indexDimension = VGPU10_OPERAND_INDEX_0D;
            operand0.numComponents = VGPU10_OPERAND_1_COMPONENT;
            emit_dword(emit, operand0.value);
            return;
         }
         else if (index == emit->fs.color_out_index[0] &&
             emit->fs.color_tmp_index != INVALID_INDEX) {
            /* replace OUTPUT[COLOR] with TEMP[COLOR].  We need to store the
             * fragment color result in a temporary so that we can read it
             * it in the post_helper() code.
             */
            file = TGSI_FILE_TEMPORARY;
            index = emit->fs.color_tmp_index;
         }
         else {
            /* Typically, for fragment shaders, the output register index
             * matches the color semantic index.  But not when we write to
             * the fragment depth register.  In that case, OUT[0] will be
             * fragdepth and OUT[1] will be the 0th color output.  We need
             * to use the semantic index for color outputs.
             */
            assert(sem_name == TGSI_SEMANTIC_COLOR);
            index = emit->info.output_semantic_index[index];

            emit->num_output_writes++;
         }
      }
   }

   /* init operand tokens to all zero */
   operand0.value = 0;

   operand0.numComponents = VGPU10_OPERAND_4_COMPONENT;

   /* the operand has a writemask */
   operand0.selectionMode = VGPU10_OPERAND_4_COMPONENT_MASK_MODE;

   /* Which of the four dest components to write to. Note that we can use a
    * simple assignment here since TGSI writemasks match VGPU10 writemasks.
    */
   STATIC_ASSERT(TGSI_WRITEMASK_X == VGPU10_OPERAND_4_COMPONENT_MASK_X);
   operand0.mask = writemask;

   /* translate TGSI register file type to VGPU10 operand type */
   operand0.operandType = translate_register_file(file, tempArrayId > 0);

   check_register_index(emit, operand0.operandType, index);

   operand0 = setup_operand0_indexing(emit, operand0, file, indirect,
                                      index2d, tempArrayId);

   /* Emit tokens */
   emit_dword(emit, operand0.value);
   if (tempArrayId > 0) {
      emit_dword(emit, tempArrayId);
   }

   emit_dword(emit, remap_temp_index(emit, file, index));

   if (indirect) {
      emit_indirect_register(emit, reg->Indirect.Index);
   }
}


/**
 * Translate a src register of a TGSI instruction and emit VGPU10 tokens.
 */
static void
emit_src_register(struct svga_shader_emitter_v10 *emit,
                  const struct tgsi_full_src_register *reg)
{
   unsigned file = reg->Register.File;
   unsigned index = reg->Register.Index;
   const unsigned indirect = reg->Register.Indirect;
   const unsigned tempArrayId = get_temp_array_id(emit, file, index);
   const unsigned index2d = reg->Register.Dimension;
   const unsigned swizzleX = reg->Register.SwizzleX;
   const unsigned swizzleY = reg->Register.SwizzleY;
   const unsigned swizzleZ = reg->Register.SwizzleZ;
   const unsigned swizzleW = reg->Register.SwizzleW;
   const unsigned absolute = reg->Register.Absolute;
   const unsigned negate = reg->Register.Negate;
   bool is_prim_id = FALSE;

   VGPU10OperandToken0 operand0;
   VGPU10OperandToken1 operand1;

   if (emit->unit == PIPE_SHADER_FRAGMENT &&
      file == TGSI_FILE_INPUT) {
      if (index == emit->fs.face_input_index) {
         /* Replace INPUT[FACE] with TEMP[FACE] */
         file = TGSI_FILE_TEMPORARY;
         index = emit->fs.face_tmp_index;
      }
      else if (index == emit->fs.fragcoord_input_index) {
         /* Replace INPUT[POSITION] with TEMP[POSITION] */
         file = TGSI_FILE_TEMPORARY;
         index = emit->fs.fragcoord_tmp_index;
      }
      else {
         /* We remap fragment shader inputs to that FS input indexes
          * match up with VS/GS output indexes.
          */
         index = emit->linkage.input_map[index];
      }
   }
   else if (emit->unit == PIPE_SHADER_GEOMETRY &&
            file == TGSI_FILE_INPUT) {
      is_prim_id = (index == emit->gs.prim_id_index);
      index = emit->linkage.input_map[index];
   }
   else if (emit->unit == PIPE_SHADER_VERTEX) {
      if (file == TGSI_FILE_INPUT) {
         /* if input is adjusted... */
         if ((emit->key.vs.adjust_attrib_w_1 |
              emit->key.vs.adjust_attrib_itof |
              emit->key.vs.adjust_attrib_utof |
              emit->key.vs.attrib_is_bgra |
              emit->key.vs.attrib_puint_to_snorm |
              emit->key.vs.attrib_puint_to_uscaled |
              emit->key.vs.attrib_puint_to_sscaled) & (1 << index)) {
            file = TGSI_FILE_TEMPORARY;
            index = emit->vs.adjusted_input[index];
         }
      }
      else if (file == TGSI_FILE_SYSTEM_VALUE) {
         assert(index < Elements(emit->system_value_indexes));
         index = emit->system_value_indexes[index];
      }
   }

   operand0.value = operand1.value = 0;

   if (is_prim_id) {
      operand0.numComponents = VGPU10_OPERAND_0_COMPONENT;
      operand0.operandType = VGPU10_OPERAND_TYPE_INPUT_PRIMITIVEID;
   }
   else {
      operand0.numComponents = VGPU10_OPERAND_4_COMPONENT;
      operand0.operandType = translate_register_file(file, tempArrayId > 0);
   }

   operand0 = setup_operand0_indexing(emit, operand0, file, indirect,
                                      index2d, tempArrayId);

   if (operand0.operandType != VGPU10_OPERAND_TYPE_IMMEDIATE32 &&
       operand0.operandType != VGPU10_OPERAND_TYPE_INPUT_PRIMITIVEID) {
      /* there's no swizzle for in-line immediates */
      if (swizzleX == swizzleY &&
          swizzleX == swizzleZ &&
          swizzleX == swizzleW) {
         operand0.selectionMode = VGPU10_OPERAND_4_COMPONENT_SELECT_1_MODE;
      }
      else {
         operand0.selectionMode = VGPU10_OPERAND_4_COMPONENT_SWIZZLE_MODE;
      }

      operand0.swizzleX = swizzleX;
      operand0.swizzleY = swizzleY;
      operand0.swizzleZ = swizzleZ;
      operand0.swizzleW = swizzleW;

      if (absolute || negate) {
         operand0.extended = 1;
         operand1.extendedOperandType = VGPU10_EXTENDED_OPERAND_MODIFIER;
         if (absolute && !negate)
            operand1.operandModifier = VGPU10_OPERAND_MODIFIER_ABS;
         if (!absolute && negate)
            operand1.operandModifier = VGPU10_OPERAND_MODIFIER_NEG;
         if (absolute && negate)
            operand1.operandModifier = VGPU10_OPERAND_MODIFIER_ABSNEG;
      }
   }

   /* Emit the operand tokens */
   emit_dword(emit, operand0.value);
   if (operand0.extended)
      emit_dword(emit, operand1.value);

   if (operand0.operandType == VGPU10_OPERAND_TYPE_IMMEDIATE32) {
      /* Emit the four float/int in-line immediate values */
      unsigned *c;
      assert(index < Elements(emit->immediates));
      assert(file == TGSI_FILE_IMMEDIATE);
      assert(swizzleX < 4);
      assert(swizzleY < 4);
      assert(swizzleZ < 4);
      assert(swizzleW < 4);
      c = (unsigned *) emit->immediates[index];
      emit_dword(emit, c[swizzleX]);
      emit_dword(emit, c[swizzleY]);
      emit_dword(emit, c[swizzleZ]);
      emit_dword(emit, c[swizzleW]);
   }
   else if (operand0.indexDimension >= VGPU10_OPERAND_INDEX_1D) {
      /* Emit the register index(es) */
      if (index2d ||
          operand0.operandType == VGPU10_OPERAND_TYPE_CONSTANT_BUFFER) {
         emit_dword(emit, reg->Dimension.Index);
      }

      if (tempArrayId > 0) {
         emit_dword(emit, tempArrayId);
      }

      emit_dword(emit, remap_temp_index(emit, file, index));

      if (indirect) {
         emit_indirect_register(emit, reg->Indirect.Index);
      }
   }
}


/**
 * Emit a resource operand (for use with a SAMPLE instruction).
 */
static void
emit_resource_register(struct svga_shader_emitter_v10 *emit,
                       unsigned resource_number)
{
   VGPU10OperandToken0 operand0;

   check_register_index(emit, VGPU10_OPERAND_TYPE_RESOURCE, resource_number);

   /* init */
   operand0.value = 0;

   operand0.operandType = VGPU10_OPERAND_TYPE_RESOURCE;
   operand0.indexDimension = VGPU10_OPERAND_INDEX_1D;
   operand0.numComponents = VGPU10_OPERAND_4_COMPONENT;
   operand0.selectionMode = VGPU10_OPERAND_4_COMPONENT_SWIZZLE_MODE;
   operand0.swizzleX = VGPU10_COMPONENT_X;
   operand0.swizzleY = VGPU10_COMPONENT_Y;
   operand0.swizzleZ = VGPU10_COMPONENT_Z;
   operand0.swizzleW = VGPU10_COMPONENT_W;

   emit_dword(emit, operand0.value);
   emit_dword(emit, resource_number);
}


/**
 * Emit a sampler operand (for use with a SAMPLE instruction).
 */
static void
emit_sampler_register(struct svga_shader_emitter_v10 *emit,
                      unsigned sampler_number)
{
   VGPU10OperandToken0 operand0;

   check_register_index(emit, VGPU10_OPERAND_TYPE_SAMPLER, sampler_number);

   /* init */
   operand0.value = 0;

   operand0.operandType = VGPU10_OPERAND_TYPE_SAMPLER;
   operand0.indexDimension = VGPU10_OPERAND_INDEX_1D;

   emit_dword(emit, operand0.value);
   emit_dword(emit, sampler_number);
}


/**
 * Emit an operand which reads the IS_FRONT_FACING register.
 */
static void
emit_face_register(struct svga_shader_emitter_v10 *emit)
{
   VGPU10OperandToken0 operand0;
   unsigned index = emit->linkage.input_map[emit->fs.face_input_index];

   /* init */
   operand0.value = 0;

   operand0.operandType = VGPU10_OPERAND_TYPE_INPUT;
   operand0.indexDimension = VGPU10_OPERAND_INDEX_1D;
   operand0.selectionMode = VGPU10_OPERAND_4_COMPONENT_SELECT_1_MODE;
   operand0.numComponents = VGPU10_OPERAND_4_COMPONENT;

   operand0.swizzleX = VGPU10_COMPONENT_X;
   operand0.swizzleY = VGPU10_COMPONENT_X;
   operand0.swizzleZ = VGPU10_COMPONENT_X;
   operand0.swizzleW = VGPU10_COMPONENT_X;

   emit_dword(emit, operand0.value);
   emit_dword(emit, index);
}


/**
 * Emit the token for a VGPU10 opcode.
 * \param saturate   clamp result to [0,1]?
 */
static void
emit_opcode(struct svga_shader_emitter_v10 *emit,
            unsigned vgpu10_opcode, boolean saturate)
{
   VGPU10OpcodeToken0 token0;

   token0.value = 0;  /* init all fields to zero */
   token0.opcodeType = vgpu10_opcode;
   token0.instructionLength = 0; /* Filled in by end_emit_instruction() */
   token0.saturate = saturate;

   emit_dword(emit, token0.value);
}


/**
 * Emit the token for a VGPU10 resinfo instruction.
 * \param modifier   return type modifier, _uint or _rcpFloat.
 *                   TODO: We may want to remove this parameter if it will
 *                   only ever be used as _uint.
 */
static void
emit_opcode_resinfo(struct svga_shader_emitter_v10 *emit,
                    VGPU10_RESINFO_RETURN_TYPE modifier)
{
   VGPU10OpcodeToken0 token0;

   token0.value = 0;  /* init all fields to zero */
   token0.opcodeType = VGPU10_OPCODE_RESINFO;
   token0.instructionLength = 0; /* Filled in by end_emit_instruction() */
   token0.resinfoReturnType = modifier;

   emit_dword(emit, token0.value);
}


/**
 * Emit opcode tokens for a texture sample instruction.  Texture instructions
 * can be rather complicated (texel offsets, etc) so we have this specialized
 * function.
 */
static void
emit_sample_opcode(struct svga_shader_emitter_v10 *emit,
                   unsigned vgpu10_opcode, boolean saturate,
                   const int offsets[3])
{
   VGPU10OpcodeToken0 token0;
   VGPU10OpcodeToken1 token1;

   token0.value = 0;  /* init all fields to zero */
   token0.opcodeType = vgpu10_opcode;
   token0.instructionLength = 0; /* Filled in by end_emit_instruction() */
   token0.saturate = saturate;

   if (offsets[0] || offsets[1] || offsets[2]) {
      assert(offsets[0] >= VGPU10_MIN_TEXEL_FETCH_OFFSET);
      assert(offsets[1] >= VGPU10_MIN_TEXEL_FETCH_OFFSET);
      assert(offsets[2] >= VGPU10_MIN_TEXEL_FETCH_OFFSET);
      assert(offsets[0] <= VGPU10_MAX_TEXEL_FETCH_OFFSET);
      assert(offsets[1] <= VGPU10_MAX_TEXEL_FETCH_OFFSET);
      assert(offsets[2] <= VGPU10_MAX_TEXEL_FETCH_OFFSET);

      token0.extended = 1;
      token1.value = 0;
      token1.opcodeType = VGPU10_EXTENDED_OPCODE_SAMPLE_CONTROLS;
      token1.offsetU = offsets[0];
      token1.offsetV = offsets[1];
      token1.offsetW = offsets[2];
   }

   emit_dword(emit, token0.value);
   if (token0.extended) {
      emit_dword(emit, token1.value);
   }
}


/**
 * Emit a DISCARD opcode token.
 * If nonzero is set, we'll discard the fragment if the X component is not 0.
 * Otherwise, we'll discard the fragment if the X component is 0.
 */
static void
emit_discard_opcode(struct svga_shader_emitter_v10 *emit, boolean nonzero)
{
   VGPU10OpcodeToken0 opcode0;

   opcode0.value = 0;
   opcode0.opcodeType = VGPU10_OPCODE_DISCARD;
   if (nonzero)
      opcode0.testBoolean = VGPU10_INSTRUCTION_TEST_NONZERO;

   emit_dword(emit, opcode0.value);
}


/**
 * We need to call this before we begin emitting a VGPU10 instruction.
 */
static void
begin_emit_instruction(struct svga_shader_emitter_v10 *emit)
{
   assert(emit->inst_start_token == 0);
   /* Save location of the instruction's VGPU10OpcodeToken0 token.
    * Note, we can't save a pointer because it would become invalid if
    * we have to realloc the output buffer.
    */
   emit->inst_start_token = emit_get_num_tokens(emit);
}


/**
 * We need to call this after we emit the last token of a VGPU10 instruction.
 * This function patches in the opcode token's instructionLength field.
 */
static void
end_emit_instruction(struct svga_shader_emitter_v10 *emit)
{
   VGPU10OpcodeToken0 *tokens = (VGPU10OpcodeToken0 *) emit->buf;
   unsigned inst_length;

   assert(emit->inst_start_token > 0);

   if (emit->discard_instruction) {
      /* Back up the emit->ptr to where this instruction started so
       * that we discard the current instruction.
       */
      emit->ptr = (char *) (tokens + emit->inst_start_token);
   }
   else {
      /* Compute instruction length and patch that into the start of
       * the instruction.
       */
      inst_length = emit_get_num_tokens(emit) - emit->inst_start_token;

      assert(inst_length > 0);

      tokens[emit->inst_start_token].instructionLength = inst_length;
   }

   emit->inst_start_token = 0; /* reset to zero for error checking */
   emit->discard_instruction = FALSE;
}


/**
 * Return index for a free temporary register.
 */
static unsigned
get_temp_index(struct svga_shader_emitter_v10 *emit)
{
   assert(emit->internal_temp_count < MAX_INTERNAL_TEMPS);
   return emit->num_shader_temps + emit->internal_temp_count++;
}


/**
 * Release the temporaries which were generated by get_temp_index().
 */
static void
free_temp_indexes(struct svga_shader_emitter_v10 *emit)
{
   emit->internal_temp_count = 0;
}


/**
 * Create a tgsi_full_src_register.
 */
static struct tgsi_full_src_register
make_src_reg(unsigned file, unsigned index)
{
   struct tgsi_full_src_register reg;

   memset(&reg, 0, sizeof(reg));
   reg.Register.File = file;
   reg.Register.Index = index;
   reg.Register.SwizzleX = TGSI_SWIZZLE_X;
   reg.Register.SwizzleY = TGSI_SWIZZLE_Y;
   reg.Register.SwizzleZ = TGSI_SWIZZLE_Z;
   reg.Register.SwizzleW = TGSI_SWIZZLE_W;
   return reg;
}


/**
 * Create a tgsi_full_src_register for a temporary.
 */
static struct tgsi_full_src_register
make_src_temp_reg(unsigned index)
{
   return make_src_reg(TGSI_FILE_TEMPORARY, index);
}


/**
 * Create a tgsi_full_src_register for a constant.
 */
static struct tgsi_full_src_register
make_src_const_reg(unsigned index)
{
   return make_src_reg(TGSI_FILE_CONSTANT, index);
}


/**
 * Create a tgsi_full_src_register for an immediate constant.
 */
static struct tgsi_full_src_register
make_src_immediate_reg(unsigned index)
{
   return make_src_reg(TGSI_FILE_IMMEDIATE, index);
}


/**
 * Create a tgsi_full_dst_register.
 */
static struct tgsi_full_dst_register
make_dst_reg(unsigned file, unsigned index)
{
   struct tgsi_full_dst_register reg;

   memset(&reg, 0, sizeof(reg));
   reg.Register.File = file;
   reg.Register.Index = index;
   reg.Register.WriteMask = TGSI_WRITEMASK_XYZW;
   return reg;
}


/**
 * Create a tgsi_full_dst_register for a temporary.
 */
static struct tgsi_full_dst_register
make_dst_temp_reg(unsigned index)
{
   return make_dst_reg(TGSI_FILE_TEMPORARY, index);
}


/**
 * Create a tgsi_full_dst_register for an output.
 */
static struct tgsi_full_dst_register
make_dst_output_reg(unsigned index)
{
   return make_dst_reg(TGSI_FILE_OUTPUT, index);
}


/**
 * Create negated tgsi_full_src_register.
 */
static struct tgsi_full_src_register
negate_src(const struct tgsi_full_src_register *reg)
{
   struct tgsi_full_src_register neg = *reg;
   neg.Register.Negate = !reg->Register.Negate;
   return neg;
}

/**
 * Create absolute value of a tgsi_full_src_register.
 */
static struct tgsi_full_src_register
absolute_src(const struct tgsi_full_src_register *reg)
{
   struct tgsi_full_src_register absolute = *reg;
   absolute.Register.Absolute = 1;
   return absolute;
}


/** Return the named swizzle term from the src register */
static inline unsigned
get_swizzle(const struct tgsi_full_src_register *reg, unsigned term)
{
   switch (term) {
   case TGSI_SWIZZLE_X:
      return reg->Register.SwizzleX;
   case TGSI_SWIZZLE_Y:
      return reg->Register.SwizzleY;
   case TGSI_SWIZZLE_Z:
      return reg->Register.SwizzleZ;
   case TGSI_SWIZZLE_W:
      return reg->Register.SwizzleW;
   default:
      assert(!"Bad swizzle");
      return TGSI_SWIZZLE_X;
   }
}


/**
 * Create swizzled tgsi_full_src_register.
 */
static struct tgsi_full_src_register
swizzle_src(const struct tgsi_full_src_register *reg,
            unsigned swizzleX, unsigned swizzleY,
            unsigned swizzleZ, unsigned swizzleW)
{
   struct tgsi_full_src_register swizzled = *reg;
   /* Note: we swizzle the current swizzle */
   swizzled.Register.SwizzleX = get_swizzle(reg, swizzleX);
   swizzled.Register.SwizzleY = get_swizzle(reg, swizzleY);
   swizzled.Register.SwizzleZ = get_swizzle(reg, swizzleZ);
   swizzled.Register.SwizzleW = get_swizzle(reg, swizzleW);
   return swizzled;
}


/**
 * Create swizzled tgsi_full_src_register where all the swizzle
 * terms are the same.
 */
static struct tgsi_full_src_register
scalar_src(const struct tgsi_full_src_register *reg, unsigned swizzle)
{
   struct tgsi_full_src_register swizzled = *reg;
   /* Note: we swizzle the current swizzle */
   swizzled.Register.SwizzleX =
   swizzled.Register.SwizzleY =
   swizzled.Register.SwizzleZ =
   swizzled.Register.SwizzleW = get_swizzle(reg, swizzle);
   return swizzled;
}


/**
 * Create new tgsi_full_dst_register with writemask.
 * \param mask  bitmask of TGSI_WRITEMASK_[XYZW]
 */
static struct tgsi_full_dst_register
writemask_dst(const struct tgsi_full_dst_register *reg, unsigned mask)
{
   struct tgsi_full_dst_register masked = *reg;
   masked.Register.WriteMask = mask;
   return masked;
}


/**
 * Check if the register's swizzle is XXXX, YYYY, ZZZZ, or WWWW.
 */
static boolean
same_swizzle_terms(const struct tgsi_full_src_register *reg)
{
   return (reg->Register.SwizzleX == reg->Register.SwizzleY &&
           reg->Register.SwizzleY == reg->Register.SwizzleZ &&
           reg->Register.SwizzleZ == reg->Register.SwizzleW);
}


/**
 * Search the vector for the value 'x' and return its position.
 */
static int
find_imm_in_vec4(const union tgsi_immediate_data vec[4],
                 union tgsi_immediate_data x)
{
   unsigned i;
   for (i = 0; i < 4; i++) {
      if (vec[i].Int == x.Int)
         return i;
   }
   return -1;
}


/**
 * Helper used by make_immediate_reg(), make_immediate_reg_4().
 */
static int
find_immediate(struct svga_shader_emitter_v10 *emit,
               union tgsi_immediate_data x, unsigned startIndex)
{
   const unsigned endIndex = emit->num_immediates;
   unsigned i;

   assert(emit->immediates_emitted);

   /* Search immediates for x, y, z, w */
   for (i = startIndex; i < endIndex; i++) {
      if (x.Int == emit->immediates[i][0].Int ||
          x.Int == emit->immediates[i][1].Int ||
          x.Int == emit->immediates[i][2].Int ||
          x.Int == emit->immediates[i][3].Int) {
         return i;
      }
   }
   /* Should never try to use an immediate value that wasn't pre-declared */
   assert(!"find_immediate() failed!");
   return -1;
}


/**
 * Return a tgsi_full_src_register for an immediate/literal
 * union tgsi_immediate_data[4] value.
 * Note: the values must have been previously declared/allocated in
 * emit_pre_helpers().  And, all of x,y,z,w must be located in the same
 * vec4 immediate.
 */
static struct tgsi_full_src_register
make_immediate_reg_4(struct svga_shader_emitter_v10 *emit,
                     const union tgsi_immediate_data imm[4])
{
   struct tgsi_full_src_register reg;
   unsigned i;

   for (i = 0; i < emit->num_common_immediates; i++) {
      /* search for first component value */
      int immpos = find_immediate(emit, imm[0], i);
      int x, y, z, w;

      assert(immpos >= 0);

      /* find remaining components within the immediate vector */
      x = find_imm_in_vec4(emit->immediates[immpos], imm[0]);
      y = find_imm_in_vec4(emit->immediates[immpos], imm[1]);
      z = find_imm_in_vec4(emit->immediates[immpos], imm[2]);
      w = find_imm_in_vec4(emit->immediates[immpos], imm[3]);

      if (x >=0 &&  y >= 0 && z >= 0 && w >= 0) {
         /* found them all */
         memset(&reg, 0, sizeof(reg));
         reg.Register.File = TGSI_FILE_IMMEDIATE;
         reg.Register.Index = immpos;
         reg.Register.SwizzleX = x;
         reg.Register.SwizzleY = y;
         reg.Register.SwizzleZ = z;
         reg.Register.SwizzleW = w;
         return reg;
      }
      /* else, keep searching */
   }

   assert(!"Failed to find immediate register!");

   /* Just return IMM[0].xxxx */
   memset(&reg, 0, sizeof(reg));
   reg.Register.File = TGSI_FILE_IMMEDIATE;
   return reg;
}


/**
 * Return a tgsi_full_src_register for an immediate/literal
 * union tgsi_immediate_data value of the form {value, value, value, value}.
 * \sa make_immediate_reg_4() regarding allowed values.
 */
static struct tgsi_full_src_register
make_immediate_reg(struct svga_shader_emitter_v10 *emit,
                   union tgsi_immediate_data value)
{
   struct tgsi_full_src_register reg;
   int immpos = find_immediate(emit, value, 0);

   assert(immpos >= 0);

   memset(&reg, 0, sizeof(reg));
   reg.Register.File = TGSI_FILE_IMMEDIATE;
   reg.Register.Index = immpos;
   reg.Register.SwizzleX =
   reg.Register.SwizzleY =
   reg.Register.SwizzleZ =
   reg.Register.SwizzleW = find_imm_in_vec4(emit->immediates[immpos], value);

   return reg;
}


/**
 * Return a tgsi_full_src_register for an immediate/literal float[4] value.
 * \sa make_immediate_reg_4() regarding allowed values.
 */
static struct tgsi_full_src_register
make_immediate_reg_float4(struct svga_shader_emitter_v10 *emit,
                          float x, float y, float z, float w)
{
   union tgsi_immediate_data imm[4];
   imm[0].Float = x;
   imm[1].Float = y;
   imm[2].Float = z;
   imm[3].Float = w;
   return make_immediate_reg_4(emit, imm);
}


/**
 * Return a tgsi_full_src_register for an immediate/literal float value
 * of the form {value, value, value, value}.
 * \sa make_immediate_reg_4() regarding allowed values.
 */
static struct tgsi_full_src_register
make_immediate_reg_float(struct svga_shader_emitter_v10 *emit, float value)
{
   union tgsi_immediate_data imm;
   imm.Float = value;
   return make_immediate_reg(emit, imm);
}


/**
 * Return a tgsi_full_src_register for an immediate/literal int[4] vector.
 */
static struct tgsi_full_src_register
make_immediate_reg_int4(struct svga_shader_emitter_v10 *emit,
                        int x, int y, int z, int w)
{
   union tgsi_immediate_data imm[4];
   imm[0].Int = x;
   imm[1].Int = y;
   imm[2].Int = z;
   imm[3].Int = w;
   return make_immediate_reg_4(emit, imm);
}


/**
 * Return a tgsi_full_src_register for an immediate/literal int value
 * of the form {value, value, value, value}.
 * \sa make_immediate_reg_4() regarding allowed values.
 */
static struct tgsi_full_src_register
make_immediate_reg_int(struct svga_shader_emitter_v10 *emit, int value)
{
   union tgsi_immediate_data imm;
   imm.Int = value;
   return make_immediate_reg(emit, imm);
}


/**
 * Allocate space for a union tgsi_immediate_data[4] immediate.
 * \return  the index/position of the immediate.
 */
static unsigned
alloc_immediate_4(struct svga_shader_emitter_v10 *emit,
                  const union tgsi_immediate_data imm[4])
{
   unsigned n = emit->num_immediates++;
   assert(!emit->immediates_emitted);
   assert(n < Elements(emit->immediates));
   emit->immediates[n][0] = imm[0];
   emit->immediates[n][1] = imm[1];
   emit->immediates[n][2] = imm[2];
   emit->immediates[n][3] = imm[3];
   return n;
}


/**
 * Allocate space for a float[4] immediate.
 * \return  the index/position of the immediate.
 */
static unsigned
alloc_immediate_float4(struct svga_shader_emitter_v10 *emit,
                       float x, float y, float z, float w)
{
   union tgsi_immediate_data imm[4];
   imm[0].Float = x;
   imm[1].Float = y;
   imm[2].Float = z;
   imm[3].Float = w;
   return alloc_immediate_4(emit, imm);
}


/**
 * Allocate space for a int[4] immediate.
 * \return  the index/position of the immediate.
 */
static unsigned
alloc_immediate_int4(struct svga_shader_emitter_v10 *emit,
                       int x, int y, int z, int w)
{
   union tgsi_immediate_data imm[4];
   imm[0].Int = x;
   imm[1].Int = y;
   imm[2].Int = z;
   imm[3].Int = w;
   return alloc_immediate_4(emit, imm);
}


/**
 * Allocate a shader input to store a system value.
 */
static unsigned
alloc_system_value_index(struct svga_shader_emitter_v10 *emit, unsigned index)
{
   const unsigned n = emit->info.num_inputs + index;
   assert(index < Elements(emit->system_value_indexes));
   emit->system_value_indexes[index] = n;
   return n;
}


/**
 * Translate a TGSI immediate value (union tgsi_immediate_data[4]) to VGPU10.
 */
static boolean
emit_vgpu10_immediate(struct svga_shader_emitter_v10 *emit,
                      const struct tgsi_full_immediate *imm)
{
   /* We don't actually emit any code here.  We just save the
    * immediate values and emit them later.
    */
   alloc_immediate_4(emit, imm->u);
   return TRUE;
}


/**
 * Emit a VGPU10_CUSTOMDATA_DCL_IMMEDIATE_CONSTANT_BUFFER block
 * containing all the immediate values previously allocated
 * with alloc_immediate_4().
 */
static boolean
emit_vgpu10_immediates_block(struct svga_shader_emitter_v10 *emit)
{
   VGPU10OpcodeToken0 token;

   assert(!emit->immediates_emitted);

   token.value = 0;
   token.opcodeType = VGPU10_OPCODE_CUSTOMDATA;
   token.customDataClass = VGPU10_CUSTOMDATA_DCL_IMMEDIATE_CONSTANT_BUFFER;

   /* Note: no begin/end_emit_instruction() calls */
   emit_dword(emit, token.value);
   emit_dword(emit, 2 + 4 * emit->num_immediates);
   emit_dwords(emit, (unsigned *) emit->immediates, 4 * emit->num_immediates);

   emit->immediates_emitted = TRUE;

   return TRUE;
}


/**
 * Translate a fragment shader's TGSI_INTERPOLATE_x mode to a vgpu10
 * interpolation mode.
 * \return a VGPU10_INTERPOLATION_x value
 */
static unsigned
translate_interpolation(const struct svga_shader_emitter_v10 *emit,
                        unsigned interp, unsigned interpolate_loc)
{
   if (interp == TGSI_INTERPOLATE_COLOR) {
      interp = emit->key.fs.flatshade ?
         TGSI_INTERPOLATE_CONSTANT : TGSI_INTERPOLATE_PERSPECTIVE;
   }

   switch (interp) {
   case TGSI_INTERPOLATE_CONSTANT:
      return VGPU10_INTERPOLATION_CONSTANT;
   case TGSI_INTERPOLATE_LINEAR:
      return interpolate_loc == TGSI_INTERPOLATE_LOC_CENTROID ?
             VGPU10_INTERPOLATION_LINEAR_NOPERSPECTIVE_CENTROID :
             VGPU10_INTERPOLATION_LINEAR_NOPERSPECTIVE;
   case TGSI_INTERPOLATE_PERSPECTIVE:
      return interpolate_loc == TGSI_INTERPOLATE_LOC_CENTROID ?
             VGPU10_INTERPOLATION_LINEAR_CENTROID :
             VGPU10_INTERPOLATION_LINEAR;
   default:
      assert(!"Unexpected interpolation mode");
      return VGPU10_INTERPOLATION_CONSTANT;
   }
}


/**
 * Translate a TGSI property to VGPU10.
 * Don't emit any instructions yet, only need to gather the primitive property information.
 * The output primitive topology might be changed later. The final property instructions
 * will be emitted as part of the pre-helper code.
 */
static boolean
emit_vgpu10_property(struct svga_shader_emitter_v10 *emit,
                     const struct tgsi_full_property *prop)
{
   static const VGPU10_PRIMITIVE primType[] = {
      VGPU10_PRIMITIVE_POINT,           /* PIPE_PRIM_POINTS */
      VGPU10_PRIMITIVE_LINE,            /* PIPE_PRIM_LINES */
      VGPU10_PRIMITIVE_LINE,            /* PIPE_PRIM_LINE_LOOP */
      VGPU10_PRIMITIVE_LINE,            /* PIPE_PRIM_LINE_STRIP */
      VGPU10_PRIMITIVE_TRIANGLE,        /* PIPE_PRIM_TRIANGLES */
      VGPU10_PRIMITIVE_TRIANGLE,        /* PIPE_PRIM_TRIANGLE_STRIP */
      VGPU10_PRIMITIVE_TRIANGLE,        /* PIPE_PRIM_TRIANGLE_FAN */
      VGPU10_PRIMITIVE_UNDEFINED,       /* PIPE_PRIM_QUADS */
      VGPU10_PRIMITIVE_UNDEFINED,       /* PIPE_PRIM_QUAD_STRIP */
      VGPU10_PRIMITIVE_UNDEFINED,       /* PIPE_PRIM_POLYGON */
      VGPU10_PRIMITIVE_LINE_ADJ,        /* PIPE_PRIM_LINES_ADJACENCY */
      VGPU10_PRIMITIVE_LINE_ADJ,        /* PIPE_PRIM_LINE_STRIP_ADJACENCY */
      VGPU10_PRIMITIVE_TRIANGLE_ADJ,    /* PIPE_PRIM_TRIANGLES_ADJACENCY */
      VGPU10_PRIMITIVE_TRIANGLE_ADJ     /* PIPE_PRIM_TRIANGLE_STRIP_ADJACENCY */
   };

   static const VGPU10_PRIMITIVE_TOPOLOGY primTopology[] = {
      VGPU10_PRIMITIVE_TOPOLOGY_POINTLIST,     /* PIPE_PRIM_POINTS */
      VGPU10_PRIMITIVE_TOPOLOGY_LINELIST,      /* PIPE_PRIM_LINES */
      VGPU10_PRIMITIVE_TOPOLOGY_LINELIST,      /* PIPE_PRIM_LINE_LOOP */
      VGPU10_PRIMITIVE_TOPOLOGY_LINESTRIP,     /* PIPE_PRIM_LINE_STRIP */
      VGPU10_PRIMITIVE_TOPOLOGY_TRIANGLELIST,  /* PIPE_PRIM_TRIANGLES */
      VGPU10_PRIMITIVE_TOPOLOGY_TRIANGLESTRIP, /* PIPE_PRIM_TRIANGLE_STRIP */
      VGPU10_PRIMITIVE_TOPOLOGY_TRIANGLESTRIP, /* PIPE_PRIM_TRIANGLE_FAN */
      VGPU10_PRIMITIVE_TOPOLOGY_UNDEFINED,     /* PIPE_PRIM_QUADS */
      VGPU10_PRIMITIVE_TOPOLOGY_UNDEFINED,     /* PIPE_PRIM_QUAD_STRIP */
      VGPU10_PRIMITIVE_TOPOLOGY_UNDEFINED,     /* PIPE_PRIM_POLYGON */
      VGPU10_PRIMITIVE_TOPOLOGY_LINELIST_ADJ,  /* PIPE_PRIM_LINES_ADJACENCY */
      VGPU10_PRIMITIVE_TOPOLOGY_LINELIST_ADJ,  /* PIPE_PRIM_LINE_STRIP_ADJACENCY */
      VGPU10_PRIMITIVE_TOPOLOGY_TRIANGLELIST_ADJ, /* PIPE_PRIM_TRIANGLES_ADJACENCY */
      VGPU10_PRIMITIVE_TOPOLOGY_TRIANGLESTRIP_ADJ /* PIPE_PRIM_TRIANGLE_STRIP_ADJACENCY */
   };

   static const unsigned inputArraySize[] = {
      0,       /* VGPU10_PRIMITIVE_UNDEFINED */
      1,       /* VGPU10_PRIMITIVE_POINT */
      2,       /* VGPU10_PRIMITIVE_LINE */
      3,       /* VGPU10_PRIMITIVE_TRIANGLE */
      0,
      0,
      4,       /* VGPU10_PRIMITIVE_LINE_ADJ */
      6        /* VGPU10_PRIMITIVE_TRIANGLE_ADJ */
   };

   switch (prop->Property.PropertyName) {
   case TGSI_PROPERTY_GS_INPUT_PRIM:
      assert(prop->u[0].Data < Elements(primType));
      emit->gs.prim_type = primType[prop->u[0].Data];
      assert(emit->gs.prim_type != VGPU10_PRIMITIVE_UNDEFINED);
      emit->gs.input_size = inputArraySize[emit->gs.prim_type];
      break;

   case TGSI_PROPERTY_GS_OUTPUT_PRIM:
      assert(prop->u[0].Data < Elements(primTopology));
      emit->gs.prim_topology = primTopology[prop->u[0].Data];
      assert(emit->gs.prim_topology != VGPU10_PRIMITIVE_TOPOLOGY_UNDEFINED);
      break;

   case TGSI_PROPERTY_GS_MAX_OUTPUT_VERTICES:
      emit->gs.max_out_vertices = prop->u[0].Data;
      break;

   default:
      break;
   }

   return TRUE;
}


static void
emit_property_instruction(struct svga_shader_emitter_v10 *emit,
                          VGPU10OpcodeToken0 opcode0, unsigned nData,
                          unsigned data)
{
   begin_emit_instruction(emit);
   emit_dword(emit, opcode0.value);
   if (nData)
      emit_dword(emit, data);
   end_emit_instruction(emit);
}


/**
 * Emit property instructions
 */
static void
emit_property_instructions(struct svga_shader_emitter_v10 *emit)
{
   VGPU10OpcodeToken0 opcode0;

   assert(emit->unit == PIPE_SHADER_GEOMETRY);

   /* emit input primitive type declaration */
   opcode0.value = 0;
   opcode0.opcodeType = VGPU10_OPCODE_DCL_GS_INPUT_PRIMITIVE;
   opcode0.primitive = emit->gs.prim_type;
   emit_property_instruction(emit, opcode0, 0, 0);

   /* emit output primitive topology declaration */
   opcode0.value = 0;
   opcode0.opcodeType = VGPU10_OPCODE_DCL_GS_OUTPUT_PRIMITIVE_TOPOLOGY;
   opcode0.primitiveTopology = emit->gs.prim_topology;
   emit_property_instruction(emit, opcode0, 0, 0);

   /* emit max output vertices */
   opcode0.value = 0;
   opcode0.opcodeType = VGPU10_OPCODE_DCL_MAX_OUTPUT_VERTEX_COUNT;
   emit_property_instruction(emit, opcode0, 1, emit->gs.max_out_vertices);
}


/**
 * Emit a vgpu10 declaration "instruction".
 * \param index  the register index
 * \param size   array size of the operand. In most cases, it is 1,
 *               but for inputs to geometry shader, the array size varies
 *               depending on the primitive type.
 */
static void
emit_decl_instruction(struct svga_shader_emitter_v10 *emit,
                      VGPU10OpcodeToken0 opcode0,
                      VGPU10OperandToken0 operand0,
                      VGPU10NameToken name_token,
                      unsigned index, unsigned size)
{
   assert(opcode0.opcodeType);
   assert(operand0.mask);

   begin_emit_instruction(emit);
   emit_dword(emit, opcode0.value);

   emit_dword(emit, operand0.value);

   if (operand0.indexDimension == VGPU10_OPERAND_INDEX_1D) {
      /* Next token is the index of the register to declare */
      emit_dword(emit, index);
   }
   else if (operand0.indexDimension >= VGPU10_OPERAND_INDEX_2D) {
      /* Next token is the size of the register */
      emit_dword(emit, size);

      /* Followed by the index of the register */
      emit_dword(emit, index);
   }

   if (name_token.value) {
      emit_dword(emit, name_token.value);
   }

   end_emit_instruction(emit);
}


/**
 * Emit the declaration for a shader input.
 * \param opcodeType  opcode type, one of VGPU10_OPCODE_DCL_INPUTx
 * \param operandType operand type, one of VGPU10_OPERAND_TYPE_INPUT_x
 * \param dim         index dimension
 * \param index       the input register index
 * \param size        array size of the operand. In most cases, it is 1,
 *                    but for inputs to geometry shader, the array size varies
 *                    depending on the primitive type.
 * \param name        one of VGPU10_NAME_x
 * \parma numComp     number of components
 * \param selMode     component selection mode
 * \param usageMask   bitfield of VGPU10_OPERAND_4_COMPONENT_MASK_x values
 * \param interpMode  interpolation mode
 */
static void
emit_input_declaration(struct svga_shader_emitter_v10 *emit,
                       unsigned opcodeType, unsigned operandType,
                       unsigned dim, unsigned index, unsigned size,
                       unsigned name, unsigned numComp,
                       unsigned selMode, unsigned usageMask,
                       unsigned interpMode)
{
   VGPU10OpcodeToken0 opcode0;
   VGPU10OperandToken0 operand0;
   VGPU10NameToken name_token;

   assert(usageMask <= VGPU10_OPERAND_4_COMPONENT_MASK_ALL);
   assert(opcodeType == VGPU10_OPCODE_DCL_INPUT ||
          opcodeType == VGPU10_OPCODE_DCL_INPUT_SIV ||
          opcodeType == VGPU10_OPCODE_DCL_INPUT_PS ||
          opcodeType == VGPU10_OPCODE_DCL_INPUT_PS_SGV);
   assert(operandType == VGPU10_OPERAND_TYPE_INPUT ||
          operandType == VGPU10_OPERAND_TYPE_INPUT_PRIMITIVEID);
   assert(numComp <= VGPU10_OPERAND_4_COMPONENT);
   assert(selMode <= VGPU10_OPERAND_4_COMPONENT_MASK_MODE);
   assert(dim <= VGPU10_OPERAND_INDEX_3D);
   assert(name == VGPU10_NAME_UNDEFINED ||
          name == VGPU10_NAME_POSITION ||
          name == VGPU10_NAME_INSTANCE_ID ||
          name == VGPU10_NAME_VERTEX_ID ||
          name == VGPU10_NAME_PRIMITIVE_ID ||
          name == VGPU10_NAME_IS_FRONT_FACE);
   assert(interpMode == VGPU10_INTERPOLATION_UNDEFINED ||
          interpMode == VGPU10_INTERPOLATION_CONSTANT ||
          interpMode == VGPU10_INTERPOLATION_LINEAR ||
          interpMode == VGPU10_INTERPOLATION_LINEAR_CENTROID ||
          interpMode == VGPU10_INTERPOLATION_LINEAR_NOPERSPECTIVE ||
          interpMode == VGPU10_INTERPOLATION_LINEAR_NOPERSPECTIVE_CENTROID);

   check_register_index(emit, opcodeType, index);

   opcode0.value = operand0.value = name_token.value = 0;

   opcode0.opcodeType = opcodeType;
   opcode0.interpolationMode = interpMode;

   operand0.operandType = operandType;
   operand0.numComponents = numComp;
   operand0.selectionMode = selMode;
   operand0.mask = usageMask;
   operand0.indexDimension = dim;
   operand0.index0Representation = VGPU10_OPERAND_INDEX_IMMEDIATE32;
   if (dim == VGPU10_OPERAND_INDEX_2D)
      operand0.index1Representation = VGPU10_OPERAND_INDEX_IMMEDIATE32;

   name_token.name = name;

   emit_decl_instruction(emit, opcode0, operand0, name_token, index, size);
}


/**
 * Emit the declaration for a shader output.
 * \param type  one of VGPU10_OPCODE_DCL_OUTPUTx
 * \param index  the output register index
 * \param name  one of VGPU10_NAME_x
 * \param usageMask  bitfield of VGPU10_OPERAND_4_COMPONENT_MASK_x values
 */
static void
emit_output_declaration(struct svga_shader_emitter_v10 *emit,
                        unsigned type, unsigned index,
                        unsigned name, unsigned usageMask)
{
   VGPU10OpcodeToken0 opcode0;
   VGPU10OperandToken0 operand0;
   VGPU10NameToken name_token;

   assert(usageMask <= VGPU10_OPERAND_4_COMPONENT_MASK_ALL);
   assert(type == VGPU10_OPCODE_DCL_OUTPUT ||
          type == VGPU10_OPCODE_DCL_OUTPUT_SGV ||
          type == VGPU10_OPCODE_DCL_OUTPUT_SIV);
   assert(name == VGPU10_NAME_UNDEFINED ||
          name == VGPU10_NAME_POSITION ||
          name == VGPU10_NAME_PRIMITIVE_ID ||
          name == VGPU10_NAME_RENDER_TARGET_ARRAY_INDEX ||
          name == VGPU10_NAME_CLIP_DISTANCE);

   check_register_index(emit, type, index);

   opcode0.value = operand0.value = name_token.value = 0;

   opcode0.opcodeType = type;
   operand0.operandType = VGPU10_OPERAND_TYPE_OUTPUT;
   operand0.numComponents = VGPU10_OPERAND_4_COMPONENT;
   operand0.selectionMode = VGPU10_OPERAND_4_COMPONENT_MASK_MODE;
   operand0.mask = usageMask;
   operand0.indexDimension = VGPU10_OPERAND_INDEX_1D;
   operand0.index0Representation = VGPU10_OPERAND_INDEX_IMMEDIATE32;

   name_token.name = name;

   emit_decl_instruction(emit, opcode0, operand0, name_token, index, 1);
}


/**
 * Emit the declaration for the fragment depth output.
 */
static void
emit_fragdepth_output_declaration(struct svga_shader_emitter_v10 *emit)
{
   VGPU10OpcodeToken0 opcode0;
   VGPU10OperandToken0 operand0;
   VGPU10NameToken name_token;

   assert(emit->unit == PIPE_SHADER_FRAGMENT);

   opcode0.value = operand0.value = name_token.value = 0;

   opcode0.opcodeType = VGPU10_OPCODE_DCL_OUTPUT;
   operand0.operandType = VGPU10_OPERAND_TYPE_OUTPUT_DEPTH;
   operand0.numComponents = VGPU10_OPERAND_1_COMPONENT;
   operand0.indexDimension = VGPU10_OPERAND_INDEX_0D;
   operand0.mask = VGPU10_OPERAND_4_COMPONENT_MASK_ALL;

   emit_decl_instruction(emit, opcode0, operand0, name_token, 0, 1);
}


/**
 * Emit the declaration for a system value input/output.
 */
static void
emit_system_value_declaration(struct svga_shader_emitter_v10 *emit,
                              unsigned semantic_name, unsigned index)
{
   switch (semantic_name) {
   case TGSI_SEMANTIC_INSTANCEID:
      index = alloc_system_value_index(emit, index);
      emit_input_declaration(emit, VGPU10_OPCODE_DCL_INPUT_SIV,
                             VGPU10_OPERAND_TYPE_INPUT,
                             VGPU10_OPERAND_INDEX_1D,
                             index, 1,
                             VGPU10_NAME_INSTANCE_ID,
                             VGPU10_OPERAND_4_COMPONENT,
                             VGPU10_OPERAND_4_COMPONENT_MASK_MODE,
                             VGPU10_OPERAND_4_COMPONENT_MASK_X,
                             VGPU10_INTERPOLATION_UNDEFINED);
      break;
   case TGSI_SEMANTIC_VERTEXID:
      index = alloc_system_value_index(emit, index);
      emit_input_declaration(emit, VGPU10_OPCODE_DCL_INPUT_SIV,
                             VGPU10_OPERAND_TYPE_INPUT,
                             VGPU10_OPERAND_INDEX_1D,
                             index, 1,
                             VGPU10_NAME_VERTEX_ID,
                             VGPU10_OPERAND_4_COMPONENT,
                             VGPU10_OPERAND_4_COMPONENT_MASK_MODE,
                             VGPU10_OPERAND_4_COMPONENT_MASK_X,
                             VGPU10_INTERPOLATION_UNDEFINED);
      break;
   default:
      ; /* XXX */
   }
}

/**
 * Translate a TGSI declaration to VGPU10.
 */
static boolean
emit_vgpu10_declaration(struct svga_shader_emitter_v10 *emit,
                        const struct tgsi_full_declaration *decl)
{
   switch (decl->Declaration.File) {
   case TGSI_FILE_INPUT:
      /* do nothing - see emit_input_declarations() */
      return TRUE;

   case TGSI_FILE_OUTPUT:
      assert(decl->Range.First == decl->Range.Last);
      emit->output_usage_mask[decl->Range.First] = decl->Declaration.UsageMask;
      return TRUE;

   case TGSI_FILE_TEMPORARY:
      /* Don't declare the temps here.  Just keep track of how many
       * and emit the declaration later.
       */
      if (decl->Declaration.Array) {
         /* Indexed temporary array.  Save the start index of the array
          * and the size of the array.
          */
         const unsigned arrayID = MIN2(decl->Array.ArrayID, MAX_TEMP_ARRAYS);
         unsigned i;

         assert(arrayID < ARRAY_SIZE(emit->temp_arrays));

         /* Save this array so we can emit the declaration for it later */
         emit->temp_arrays[arrayID].start = decl->Range.First;
         emit->temp_arrays[arrayID].size =
            decl->Range.Last - decl->Range.First + 1;

         emit->num_temp_arrays = MAX2(emit->num_temp_arrays, arrayID + 1);
         assert(emit->num_temp_arrays <= MAX_TEMP_ARRAYS);
         emit->num_temp_arrays = MIN2(emit->num_temp_arrays, MAX_TEMP_ARRAYS);

         /* Fill in the temp_map entries for this array */
         for (i = decl->Range.First; i <= decl->Range.Last; i++) {
            emit->temp_map[i].arrayId = arrayID;
            emit->temp_map[i].index = i - decl->Range.First;
         }
      }

      /* for all temps, indexed or not, keep track of highest index */
      emit->num_shader_temps = MAX2(emit->num_shader_temps,
                                    decl->Range.Last + 1);
      return TRUE;

   case TGSI_FILE_CONSTANT:
      /* Don't declare constants here.  Just keep track and emit later. */
      {
         unsigned constbuf = 0, num_consts;
         if (decl->Declaration.Dimension) {
            constbuf = decl->Dim.Index2D;
         }
         /* We throw an assertion here when, in fact, the shader should never
          * have linked due to constbuf index out of bounds, so we shouldn't
          * have reached here.
          */
         assert(constbuf < Elements(emit->num_shader_consts));

         num_consts = MAX2(emit->num_shader_consts[constbuf],
                           decl->Range.Last + 1);

         if (num_consts > VGPU10_MAX_CONSTANT_BUFFER_ELEMENT_COUNT) {
            debug_printf("Warning: constant buffer is declared to size [%u]"
                         " but [%u] is the limit.\n",
                         num_consts,
                         VGPU10_MAX_CONSTANT_BUFFER_ELEMENT_COUNT);
         }
         /* The linker doesn't enforce the max UBO size so we clamp here */
         emit->num_shader_consts[constbuf] =
            MIN2(num_consts, VGPU10_MAX_CONSTANT_BUFFER_ELEMENT_COUNT);
      }
      return TRUE;

   case TGSI_FILE_IMMEDIATE:
      assert(!"TGSI_FILE_IMMEDIATE not handled yet!");
      return FALSE;

   case TGSI_FILE_SYSTEM_VALUE:
      emit_system_value_declaration(emit, decl->Semantic.Name,
                                    decl->Range.First);
      return TRUE;

   case TGSI_FILE_SAMPLER:
      /* Don't declare samplers here.  Just keep track and emit later. */
      emit->num_samplers = MAX2(emit->num_samplers, decl->Range.Last + 1);
      return TRUE;

#if 0
   case TGSI_FILE_RESOURCE:
      /*opcode0.opcodeType = VGPU10_OPCODE_DCL_RESOURCE;*/
      /* XXX more, VGPU10_RETURN_TYPE_FLOAT */
      assert(!"TGSI_FILE_RESOURCE not handled yet");
      return FALSE;
#endif

   case TGSI_FILE_ADDRESS:
      emit->num_address_regs = MAX2(emit->num_address_regs,
                                    decl->Range.Last + 1);
      return TRUE;

   case TGSI_FILE_SAMPLER_VIEW:
      /* Not used at this time, but maybe in the future.
       * See emit_resource_declarations().
       */
      return TRUE;

   default:
      assert(!"Unexpected type of declaration");
      return FALSE;
   }
}



/**
 * Emit all input declarations.
 */
static boolean
emit_input_declarations(struct svga_shader_emitter_v10 *emit)
{
   unsigned i;

   if (emit->unit == PIPE_SHADER_FRAGMENT) {

      for (i = 0; i < emit->linkage.num_inputs; i++) {
         unsigned semantic_name = emit->info.input_semantic_name[i];
         unsigned usage_mask = emit->info.input_usage_mask[i];
         unsigned index = emit->linkage.input_map[i];
         unsigned type, interpolationMode, name;

         if (usage_mask == 0)
            continue;  /* register is not actually used */

         if (semantic_name == TGSI_SEMANTIC_POSITION) {
            /* fragment position input */
            type = VGPU10_OPCODE_DCL_INPUT_PS_SGV;
            interpolationMode = VGPU10_INTERPOLATION_LINEAR;
            name = VGPU10_NAME_POSITION;
            if (usage_mask & TGSI_WRITEMASK_W) {
               /* we need to replace use of 'w' with '1/w' */
               emit->fs.fragcoord_input_index = i;
            }
         }
         else if (semantic_name == TGSI_SEMANTIC_FACE) {
            /* fragment front-facing input */
            type = VGPU10_OPCODE_DCL_INPUT_PS_SGV;
            interpolationMode = VGPU10_INTERPOLATION_CONSTANT;
            name = VGPU10_NAME_IS_FRONT_FACE;
            emit->fs.face_input_index = i;
         }
         else if (semantic_name == TGSI_SEMANTIC_PRIMID) {
            /* primitive ID */
            type = VGPU10_OPCODE_DCL_INPUT_PS_SGV;
            interpolationMode = VGPU10_INTERPOLATION_CONSTANT;
            name = VGPU10_NAME_PRIMITIVE_ID;
         }
         else {
            /* general fragment input */
            type = VGPU10_OPCODE_DCL_INPUT_PS;
            interpolationMode =
               translate_interpolation(emit,
                                       emit->info.input_interpolate[i],
                                       emit->info.input_interpolate_loc[i]);

            /* keeps track if flat interpolation mode is being used */
            emit->uses_flat_interp = emit->uses_flat_interp ||
               (interpolationMode == VGPU10_INTERPOLATION_CONSTANT);

            name = VGPU10_NAME_UNDEFINED;
         }

         emit_input_declaration(emit, type,
                                VGPU10_OPERAND_TYPE_INPUT,
                                VGPU10_OPERAND_INDEX_1D, index, 1,
                                name,
                                VGPU10_OPERAND_4_COMPONENT,
                                VGPU10_OPERAND_4_COMPONENT_MASK_MODE,
                                VGPU10_OPERAND_4_COMPONENT_MASK_ALL,
                                interpolationMode);
      }
   }
   else if (emit->unit == PIPE_SHADER_GEOMETRY) {

      for (i = 0; i < emit->info.num_inputs; i++) {
         unsigned semantic_name = emit->info.input_semantic_name[i];
         unsigned usage_mask = emit->info.input_usage_mask[i];
         unsigned index = emit->linkage.input_map[i];
         unsigned opcodeType, operandType;
         unsigned numComp, selMode;
         unsigned name;
         unsigned dim;

         if (usage_mask == 0)
            continue;  /* register is not actually used */

         opcodeType = VGPU10_OPCODE_DCL_INPUT;
         operandType = VGPU10_OPERAND_TYPE_INPUT;
         numComp = VGPU10_OPERAND_4_COMPONENT;
         selMode = VGPU10_OPERAND_4_COMPONENT_MASK_MODE;
         name = VGPU10_NAME_UNDEFINED;

         /* all geometry shader inputs are two dimensional except gl_PrimitiveID */
         dim = VGPU10_OPERAND_INDEX_2D;

         if (semantic_name == TGSI_SEMANTIC_PRIMID) {
            /* Primitive ID */
            operandType = VGPU10_OPERAND_TYPE_INPUT_PRIMITIVEID;
            dim = VGPU10_OPERAND_INDEX_0D;
            numComp = VGPU10_OPERAND_0_COMPONENT;
            selMode = 0;

            /* also save the register index so we can check for
             * primitive id when emit src register. We need to modify the
             * operand type, index dimension when emit primitive id src reg.
             */
            emit->gs.prim_id_index = i;
         }
         else if (semantic_name == TGSI_SEMANTIC_POSITION) {
            /* vertex position input */
            opcodeType = VGPU10_OPCODE_DCL_INPUT_SIV;
            name = VGPU10_NAME_POSITION;
         }

         emit_input_declaration(emit, opcodeType, operandType,
                                dim, index,
                                emit->gs.input_size,
                                name,
                                numComp, selMode,
                                VGPU10_OPERAND_4_COMPONENT_MASK_ALL,
                                VGPU10_INTERPOLATION_UNDEFINED);
      }
   }
   else {
      assert(emit->unit == PIPE_SHADER_VERTEX);

      for (i = 0; i < emit->info.num_inputs; i++) {
         unsigned usage_mask = emit->info.input_usage_mask[i];
         unsigned index = i;

         if (usage_mask == 0)
            continue;  /* register is not actually used */

         emit_input_declaration(emit, VGPU10_OPCODE_DCL_INPUT,
                                VGPU10_OPERAND_TYPE_INPUT,
                                VGPU10_OPERAND_INDEX_1D, index, 1,
                                VGPU10_NAME_UNDEFINED,
                                VGPU10_OPERAND_4_COMPONENT,
                                VGPU10_OPERAND_4_COMPONENT_MASK_MODE,
                                VGPU10_OPERAND_4_COMPONENT_MASK_ALL,
                                VGPU10_INTERPOLATION_UNDEFINED);
      }
   }

   return TRUE;
}


/**
 * Emit all output declarations.
 */
static boolean
emit_output_declarations(struct svga_shader_emitter_v10 *emit)
{
   unsigned i;

   for (i = 0; i < emit->info.num_outputs; i++) {
      /*const unsigned usage_mask = emit->info.output_usage_mask[i];*/
      const unsigned semantic_name = emit->info.output_semantic_name[i];
      const unsigned semantic_index = emit->info.output_semantic_index[i];
      unsigned index = i;

      if (emit->unit == PIPE_SHADER_FRAGMENT) {
         if (semantic_name == TGSI_SEMANTIC_COLOR) {
            assert(semantic_index < Elements(emit->fs.color_out_index));

            emit->fs.color_out_index[semantic_index] = index;

            /* The semantic index is the shader's color output/buffer index */
            emit_output_declaration(emit,
                                    VGPU10_OPCODE_DCL_OUTPUT, semantic_index,
                                    VGPU10_NAME_UNDEFINED,
                                    VGPU10_OPERAND_4_COMPONENT_MASK_ALL);

            if (semantic_index == 0) {
               if (emit->key.fs.write_color0_to_n_cbufs > 1) {
                  /* Emit declarations for the additional color outputs
                   * for broadcasting.
                   */
                  unsigned j;
                  for (j = 1; j < emit->key.fs.write_color0_to_n_cbufs; j++) {
                     /* Allocate a new output index */
                     unsigned idx = emit->info.num_outputs + j - 1;
                     emit->fs.color_out_index[j] = idx;
                     emit_output_declaration(emit,
                                        VGPU10_OPCODE_DCL_OUTPUT, idx,
                                        VGPU10_NAME_UNDEFINED,
                                        VGPU10_OPERAND_4_COMPONENT_MASK_ALL);
                     emit->info.output_semantic_index[idx] = j;
                  }
               }
            }
            else {
               assert(!emit->key.fs.write_color0_to_n_cbufs);
            }
         }
         else if (semantic_name == TGSI_SEMANTIC_POSITION) {
            /* Fragment depth output */
            emit_fragdepth_output_declaration(emit);
         }
         else {
            assert(!"Bad output semantic name");
         }
      }
      else {
         /* VS or GS */
         unsigned name, type;
         unsigned writemask = VGPU10_OPERAND_4_COMPONENT_MASK_ALL;

         switch (semantic_name) {
         case TGSI_SEMANTIC_POSITION:
            assert(emit->unit != PIPE_SHADER_FRAGMENT);
            type = VGPU10_OPCODE_DCL_OUTPUT_SIV;
            name = VGPU10_NAME_POSITION;
            /* Save the index of the vertex position output register */
            emit->vposition.out_index = index;
            break;
         case TGSI_SEMANTIC_CLIPDIST:
            type = VGPU10_OPCODE_DCL_OUTPUT_SIV;
            name = VGPU10_NAME_CLIP_DISTANCE;
            /* save the starting index of the clip distance output register */
            if (semantic_index == 0)
               emit->clip_dist_out_index = index;
            writemask = emit->output_usage_mask[index];
            writemask = apply_clip_plane_mask(emit, writemask, semantic_index);
            if (writemask == 0x0) {
               continue; /* discard this do-nothing declaration */
            }
            break;
         case TGSI_SEMANTIC_PRIMID:
            assert(emit->unit == PIPE_SHADER_GEOMETRY);
            type = VGPU10_OPCODE_DCL_OUTPUT_SGV;
            name = VGPU10_NAME_PRIMITIVE_ID;
            break;
         case TGSI_SEMANTIC_LAYER:
            assert(emit->unit == PIPE_SHADER_GEOMETRY);
            type = VGPU10_OPCODE_DCL_OUTPUT_SGV;
            name = VGPU10_NAME_RENDER_TARGET_ARRAY_INDEX;
            break;
         case TGSI_SEMANTIC_CLIPVERTEX:
            type = VGPU10_OPCODE_DCL_OUTPUT;
            name = VGPU10_NAME_UNDEFINED;
            emit->clip_vertex_out_index = index;
            break;
         default:
            /* generic output */
            type = VGPU10_OPCODE_DCL_OUTPUT;
            name = VGPU10_NAME_UNDEFINED;
         }

         emit_output_declaration(emit, type, index, name, writemask);
      }
   }

   if (emit->vposition.so_index != INVALID_INDEX &&
       emit->vposition.out_index != INVALID_INDEX) {

      assert(emit->unit != PIPE_SHADER_FRAGMENT);

      /* Emit the declaration for the non-adjusted vertex position
       * for stream output purpose
       */
      emit_output_declaration(emit, VGPU10_OPCODE_DCL_OUTPUT,
                              emit->vposition.so_index,
                              VGPU10_NAME_UNDEFINED,
                              VGPU10_OPERAND_4_COMPONENT_MASK_ALL);
   }

   if (emit->clip_dist_so_index != INVALID_INDEX &&
       emit->clip_dist_out_index != INVALID_INDEX) {

      assert(emit->unit != PIPE_SHADER_FRAGMENT);

      /* Emit the declaration for the clip distance shadow copy which
       * will be used for stream output purpose and for clip distance
       * varying variable
       */
      emit_output_declaration(emit, VGPU10_OPCODE_DCL_OUTPUT,
                              emit->clip_dist_so_index,
                              VGPU10_NAME_UNDEFINED,
                              emit->output_usage_mask[emit->clip_dist_out_index]);

      if (emit->info.num_written_clipdistance > 4) {
         /* for the second clip distance register, each handles 4 planes */
         emit_output_declaration(emit, VGPU10_OPCODE_DCL_OUTPUT,
                                 emit->clip_dist_so_index + 1,
                                 VGPU10_NAME_UNDEFINED,
                                 emit->output_usage_mask[emit->clip_dist_out_index+1]);
      }
   }

   return TRUE;
}


/**
 * Emit the declaration for the temporary registers.
 */
static boolean
emit_temporaries_declaration(struct svga_shader_emitter_v10 *emit)
{
   unsigned total_temps, reg, i;

   total_temps = emit->num_shader_temps;

   /* Allocate extra temps for specially-implemented instructions,
    * such as LIT.
    */
   total_temps += MAX_INTERNAL_TEMPS;

   if (emit->unit == PIPE_SHADER_VERTEX || emit->unit == PIPE_SHADER_GEOMETRY) {
      if (emit->vposition.need_prescale || emit->key.vs.undo_viewport ||
          emit->key.clip_plane_enable ||
          emit->vposition.so_index != INVALID_INDEX) {
         emit->vposition.tmp_index = total_temps;
         total_temps += 1;
      }

      if (emit->unit == PIPE_SHADER_VERTEX) {
         unsigned attrib_mask = (emit->key.vs.adjust_attrib_w_1 |
                                 emit->key.vs.adjust_attrib_itof |
                                 emit->key.vs.adjust_attrib_utof |
                                 emit->key.vs.attrib_is_bgra |
                                 emit->key.vs.attrib_puint_to_snorm |
                                 emit->key.vs.attrib_puint_to_uscaled |
                                 emit->key.vs.attrib_puint_to_sscaled);
         while (attrib_mask) {
            unsigned index = u_bit_scan(&attrib_mask);
            emit->vs.adjusted_input[index] = total_temps++;
         }
      }

      if (emit->clip_mode == CLIP_DISTANCE) {
         /* We need to write the clip distance to a temporary register
          * first. Then it will be copied to the shadow copy for
          * the clip distance varying variable and stream output purpose.
          * It will also be copied to the actual CLIPDIST register
          * according to the enabled clip planes
          */
         emit->clip_dist_tmp_index = total_temps++;
         if (emit->info.num_written_clipdistance > 4)
            total_temps++; /* second clip register */
      }
      else if (emit->clip_mode == CLIP_VERTEX) {
         /* We need to convert the TGSI CLIPVERTEX output to one or more
          * clip distances.  Allocate a temp reg for the clipvertex here.
          */
         assert(emit->info.writes_clipvertex > 0);
         emit->clip_vertex_tmp_index = total_temps;
         total_temps++;
      }
   }
   else if (emit->unit == PIPE_SHADER_FRAGMENT) {
      if (emit->key.fs.alpha_func != SVGA3D_CMP_ALWAYS ||
          emit->key.fs.white_fragments ||
          emit->key.fs.write_color0_to_n_cbufs > 1) {
         /* Allocate a temp to hold the output color */
         emit->fs.color_tmp_index = total_temps;
         total_temps += 1;
      }

      if (emit->fs.face_input_index != INVALID_INDEX) {
         /* Allocate a temp for the +/-1 face register */
         emit->fs.face_tmp_index = total_temps;
         total_temps += 1;
      }

      if (emit->fs.fragcoord_input_index != INVALID_INDEX) {
         /* Allocate a temp for modified fragment position register */
         emit->fs.fragcoord_tmp_index = total_temps;
         total_temps += 1;
      }
   }

   for (i = 0; i < emit->num_address_regs; i++) {
      emit->address_reg_index[i] = total_temps++;
   }

   /* Initialize the temp_map array which maps TGSI temp indexes to VGPU10
    * temp indexes.  Basically, we compact all the non-array temp register
    * indexes into a consecutive series.
    *
    * Before, we may have some TGSI declarations like:
    *   DCL TEMP[0..1], LOCAL
    *   DCL TEMP[2..4], ARRAY(1), LOCAL
    *   DCL TEMP[5..7], ARRAY(2), LOCAL
    *   plus, some extra temps, like TEMP[8], TEMP[9] for misc things
    *
    * After, we'll have a map like this:
    *   temp_map[0] = { array 0, index 0 }
    *   temp_map[1] = { array 0, index 1 }
    *   temp_map[2] = { array 1, index 0 }
    *   temp_map[3] = { array 1, index 1 }
    *   temp_map[4] = { array 1, index 2 }
    *   temp_map[5] = { array 2, index 0 }
    *   temp_map[6] = { array 2, index 1 }
    *   temp_map[7] = { array 2, index 2 }
    *   temp_map[8] = { array 0, index 2 }
    *   temp_map[9] = { array 0, index 3 }
    *
    * We'll declare two arrays of 3 elements, plus a set of four non-indexed
    * temps numbered 0..3
    *
    * Any time we emit a temporary register index, we'll have to use the
    * temp_map[] table to convert the TGSI index to the VGPU10 index.
    *
    * Finally, we recompute the total_temps value here.
    */
   reg = 0;
   for (i = 0; i < total_temps; i++) {
      if (emit->temp_map[i].arrayId == 0) {
         emit->temp_map[i].index = reg++;
      }
   }
   total_temps = reg;

   if (0) {
      debug_printf("total_temps %u\n", total_temps);
      for (i = 0; i < 30; i++) {
         debug_printf("temp %u ->  array %u  index %u\n",
                      i, emit->temp_map[i].arrayId, emit->temp_map[i].index);
      }
   }

   /* Emit declaration of ordinary temp registers */
   if (total_temps > 0) {
      VGPU10OpcodeToken0 opcode0;

      opcode0.value = 0;
      opcode0.opcodeType = VGPU10_OPCODE_DCL_TEMPS;

      begin_emit_instruction(emit);
      emit_dword(emit, opcode0.value);
      emit_dword(emit, total_temps);
      end_emit_instruction(emit);
   }

   /* Emit declarations for indexable temp arrays.  Skip 0th entry since
    * it's unused.
    */
   for (i = 1; i < emit->num_temp_arrays; i++) {
      unsigned num_temps = emit->temp_arrays[i].size;

      if (num_temps > 0) {
         VGPU10OpcodeToken0 opcode0;

         opcode0.value = 0;
         opcode0.opcodeType = VGPU10_OPCODE_DCL_INDEXABLE_TEMP;

         begin_emit_instruction(emit);
         emit_dword(emit, opcode0.value);
         emit_dword(emit, i); /* which array */
         emit_dword(emit, num_temps);
         emit_dword(emit, 4); /* num components */
         end_emit_instruction(emit);

         total_temps += num_temps;
      }
   }

   /* Check that the grand total of all regular and indexed temps is
    * under the limit.
    */
   check_register_index(emit, VGPU10_OPCODE_DCL_TEMPS, total_temps - 1);

   return TRUE;
}


static boolean
emit_constant_declaration(struct svga_shader_emitter_v10 *emit)
{
   VGPU10OpcodeToken0 opcode0;
   VGPU10OperandToken0 operand0;
   unsigned total_consts, i;

   opcode0.value = 0;
   opcode0.opcodeType = VGPU10_OPCODE_DCL_CONSTANT_BUFFER;
   opcode0.accessPattern = VGPU10_CB_IMMEDIATE_INDEXED;
   /* XXX or, access pattern = VGPU10_CB_DYNAMIC_INDEXED */

   operand0.value = 0;
   operand0.numComponents = VGPU10_OPERAND_4_COMPONENT;
   operand0.indexDimension = VGPU10_OPERAND_INDEX_2D;
   operand0.index0Representation = VGPU10_OPERAND_INDEX_IMMEDIATE32;
   operand0.index1Representation = VGPU10_OPERAND_INDEX_IMMEDIATE32;
   operand0.operandType = VGPU10_OPERAND_TYPE_CONSTANT_BUFFER;
   operand0.selectionMode = VGPU10_OPERAND_4_COMPONENT_SWIZZLE_MODE;
   operand0.swizzleX = 0;
   operand0.swizzleY = 1;
   operand0.swizzleZ = 2;
   operand0.swizzleW = 3;

   /**
    * Emit declaration for constant buffer [0].  We also allocate
    * room for the extra constants here.
    */
   total_consts = emit->num_shader_consts[0];

   /* Now, allocate constant slots for the "extra" constants */

   /* Vertex position scale/translation */
   if (emit->vposition.need_prescale) {
      emit->vposition.prescale_scale_index = total_consts++;
      emit->vposition.prescale_trans_index = total_consts++;
   }

   if (emit->unit == PIPE_SHADER_VERTEX) {
      if (emit->key.vs.undo_viewport) {
         emit->vs.viewport_index = total_consts++;
      }
   }

   /* user-defined clip planes */
   if (emit->key.clip_plane_enable) {
      unsigned n = util_bitcount(emit->key.clip_plane_enable);
      assert(emit->unit == PIPE_SHADER_VERTEX ||
             emit->unit == PIPE_SHADER_GEOMETRY);
      for (i = 0; i < n; i++) {
         emit->clip_plane_const[i] = total_consts++;
      }
   }

   /* Texcoord scale factors for RECT textures */
   {
      for (i = 0; i < emit->num_samplers; i++) {
         if (emit->key.tex[i].unnormalized) {
            emit->texcoord_scale_index[i] = total_consts++;
         }
      }
   }

   /* Texture buffer sizes */
   for (i = 0; i < emit->num_samplers; i++) {
      if (emit->key.tex[i].texture_target == PIPE_BUFFER) {
         emit->texture_buffer_size_index[i] = total_consts++;
      }
   }

   if (total_consts > 0) {
      begin_emit_instruction(emit);
      emit_dword(emit, opcode0.value);
      emit_dword(emit, operand0.value);
      emit_dword(emit, 0);  /* which const buffer slot */
      emit_dword(emit, total_consts);
      end_emit_instruction(emit);
   }

   /* Declare remaining constant buffers (UBOs) */
   for (i = 1; i < Elements(emit->num_shader_consts); i++) {
      if (emit->num_shader_consts[i] > 0) {
         begin_emit_instruction(emit);
         emit_dword(emit, opcode0.value);
         emit_dword(emit, operand0.value);
         emit_dword(emit, i);  /* which const buffer slot */
         emit_dword(emit, emit->num_shader_consts[i]);
         end_emit_instruction(emit);
      }
   }

   return TRUE;
}


/**
 * Emit declarations for samplers.
 */
static boolean
emit_sampler_declarations(struct svga_shader_emitter_v10 *emit)
{
   unsigned i;

   for (i = 0; i < emit->num_samplers; i++) {
      VGPU10OpcodeToken0 opcode0;
      VGPU10OperandToken0 operand0;

      opcode0.value = 0;
      opcode0.opcodeType = VGPU10_OPCODE_DCL_SAMPLER;
      opcode0.samplerMode = VGPU10_SAMPLER_MODE_DEFAULT;

      operand0.value = 0;
      operand0.numComponents = VGPU10_OPERAND_0_COMPONENT;
      operand0.operandType = VGPU10_OPERAND_TYPE_SAMPLER;
      operand0.indexDimension = VGPU10_OPERAND_INDEX_1D;
      operand0.index0Representation = VGPU10_OPERAND_INDEX_IMMEDIATE32;

      begin_emit_instruction(emit);
      emit_dword(emit, opcode0.value);
      emit_dword(emit, operand0.value);
      emit_dword(emit, i);
      end_emit_instruction(emit);
   }

   return TRUE;
}


/**
 * Translate PIPE_TEXTURE_x to VGAPU10_RESOURCE_DIMENSION_x.
 */
static unsigned
pipe_texture_to_resource_dimension(unsigned target, bool msaa)
{
   switch (target) {
   case PIPE_BUFFER:
      return VGPU10_RESOURCE_DIMENSION_BUFFER;
   case PIPE_TEXTURE_1D:
      return VGPU10_RESOURCE_DIMENSION_TEXTURE1D;
   case PIPE_TEXTURE_2D:
   case PIPE_TEXTURE_RECT:
      return msaa ? VGPU10_RESOURCE_DIMENSION_TEXTURE2DMS
         : VGPU10_RESOURCE_DIMENSION_TEXTURE2D;
   case PIPE_TEXTURE_3D:
      return VGPU10_RESOURCE_DIMENSION_TEXTURE3D;
   case PIPE_TEXTURE_CUBE:
      return VGPU10_RESOURCE_DIMENSION_TEXTURECUBE;
   case PIPE_TEXTURE_1D_ARRAY:
      return VGPU10_RESOURCE_DIMENSION_TEXTURE1DARRAY;
   case PIPE_TEXTURE_2D_ARRAY:
      return msaa ? VGPU10_RESOURCE_DIMENSION_TEXTURE2DMSARRAY
         : VGPU10_RESOURCE_DIMENSION_TEXTURE2DARRAY;
   case PIPE_TEXTURE_CUBE_ARRAY:
      return VGPU10_RESOURCE_DIMENSION_TEXTURECUBEARRAY;
   default:
      assert(!"Unexpected resource type");
      return VGPU10_RESOURCE_DIMENSION_TEXTURE2D;
   }
}


/**
 * Given a tgsi_return_type, return true iff it is an integer type.
 */
static boolean
is_integer_type(enum tgsi_return_type type)
{
   switch (type) {
      case TGSI_RETURN_TYPE_SINT:
      case TGSI_RETURN_TYPE_UINT:
         return TRUE;
      case TGSI_RETURN_TYPE_FLOAT:
      case TGSI_RETURN_TYPE_UNORM:
      case TGSI_RETURN_TYPE_SNORM:
         return FALSE;
      case TGSI_RETURN_TYPE_COUNT:
      default:
         assert(!"is_integer_type: Unknown tgsi_return_type");
         return FALSE;
   }
}


/**
 * Emit declarations for resources.
 * XXX When we're sure that all TGSI shaders will be generated with
 * sampler view declarations (Ex: DCL SVIEW[n], 2D, UINT) we may
 * rework this code.
 */
static boolean
emit_resource_declarations(struct svga_shader_emitter_v10 *emit)
{
   unsigned i;

   /* Emit resource decl for each sampler */
   for (i = 0; i < emit->num_samplers; i++) {
      VGPU10OpcodeToken0 opcode0;
      VGPU10OperandToken0 operand0;
      VGPU10ResourceReturnTypeToken return_type;
      VGPU10_RESOURCE_RETURN_TYPE rt;

      opcode0.value = 0;
      opcode0.opcodeType = VGPU10_OPCODE_DCL_RESOURCE;
      opcode0.resourceDimension =
         pipe_texture_to_resource_dimension(emit->key.tex[i].texture_target,
                                            emit->key.tex[i].texture_msaa);
      operand0.value = 0;
      operand0.numComponents = VGPU10_OPERAND_0_COMPONENT;
      operand0.operandType = VGPU10_OPERAND_TYPE_RESOURCE;
      operand0.indexDimension = VGPU10_OPERAND_INDEX_1D;
      operand0.index0Representation = VGPU10_OPERAND_INDEX_IMMEDIATE32;

#if 1
      /* convert TGSI_RETURN_TYPE_x to VGPU10_RETURN_TYPE_x */
      STATIC_ASSERT(VGPU10_RETURN_TYPE_UNORM == TGSI_RETURN_TYPE_UNORM + 1);
      STATIC_ASSERT(VGPU10_RETURN_TYPE_SNORM == TGSI_RETURN_TYPE_SNORM + 1);
      STATIC_ASSERT(VGPU10_RETURN_TYPE_SINT == TGSI_RETURN_TYPE_SINT + 1);
      STATIC_ASSERT(VGPU10_RETURN_TYPE_UINT == TGSI_RETURN_TYPE_UINT + 1);
      STATIC_ASSERT(VGPU10_RETURN_TYPE_FLOAT == TGSI_RETURN_TYPE_FLOAT + 1);
      assert(emit->key.tex[i].return_type <= TGSI_RETURN_TYPE_FLOAT);
      rt = emit->key.tex[i].return_type + 1;
#else
      switch (emit->key.tex[i].return_type) {
         case TGSI_RETURN_TYPE_UNORM: rt = VGPU10_RETURN_TYPE_UNORM; break;
         case TGSI_RETURN_TYPE_SNORM: rt = VGPU10_RETURN_TYPE_SNORM; break;
         case TGSI_RETURN_TYPE_SINT:  rt = VGPU10_RETURN_TYPE_SINT;  break;
         case TGSI_RETURN_TYPE_UINT:  rt = VGPU10_RETURN_TYPE_UINT;  break;
         case TGSI_RETURN_TYPE_FLOAT: rt = VGPU10_RETURN_TYPE_FLOAT; break;
         case TGSI_RETURN_TYPE_COUNT:
         default:
            rt = VGPU10_RETURN_TYPE_FLOAT;
            assert(!"emit_resource_declarations: Unknown tgsi_return_type");
      }
#endif

      return_type.value = 0;
      return_type.component0 = rt;
      return_type.component1 = rt;
      return_type.component2 = rt;
      return_type.component3 = rt;

      begin_emit_instruction(emit);
      emit_dword(emit, opcode0.value);
      emit_dword(emit, operand0.value);
      emit_dword(emit, i);
      emit_dword(emit, return_type.value);
      end_emit_instruction(emit);
   }

   return TRUE;
}

static void
emit_instruction_op1(struct svga_shader_emitter_v10 *emit,
                     unsigned opcode,
                     const struct tgsi_full_dst_register *dst,
                     const struct tgsi_full_src_register *src,
                     boolean saturate)
{
   begin_emit_instruction(emit);
   emit_opcode(emit, opcode, saturate);
   emit_dst_register(emit, dst);
   emit_src_register(emit, src);
   end_emit_instruction(emit);
}

static void
emit_instruction_op2(struct svga_shader_emitter_v10 *emit,
                     unsigned opcode,
                     const struct tgsi_full_dst_register *dst,
                     const struct tgsi_full_src_register *src1,
                     const struct tgsi_full_src_register *src2,
                     boolean saturate)
{
   begin_emit_instruction(emit);
   emit_opcode(emit, opcode, saturate);
   emit_dst_register(emit, dst);
   emit_src_register(emit, src1);
   emit_src_register(emit, src2);
   end_emit_instruction(emit);
}

static void
emit_instruction_op3(struct svga_shader_emitter_v10 *emit,
                     unsigned opcode,
                     const struct tgsi_full_dst_register *dst,
                     const struct tgsi_full_src_register *src1,
                     const struct tgsi_full_src_register *src2,
                     const struct tgsi_full_src_register *src3,
                     boolean saturate)
{
   begin_emit_instruction(emit);
   emit_opcode(emit, opcode, saturate);
   emit_dst_register(emit, dst);
   emit_src_register(emit, src1);
   emit_src_register(emit, src2);
   emit_src_register(emit, src3);
   end_emit_instruction(emit);
}

/**
 * Emit the actual clip distance instructions to be used for clipping
 * by copying the clip distance from the temporary registers to the
 * CLIPDIST registers written with the enabled planes mask.
 * Also copy the clip distance from the temporary to the clip distance
 * shadow copy register which will be referenced by the input shader
 */
static void
emit_clip_distance_instructions(struct svga_shader_emitter_v10 *emit)
{
   struct tgsi_full_src_register tmp_clip_dist_src;
   struct tgsi_full_dst_register clip_dist_dst;

   unsigned i;
   unsigned clip_plane_enable = emit->key.clip_plane_enable;
   unsigned clip_dist_tmp_index = emit->clip_dist_tmp_index;
   int num_written_clipdist = emit->info.num_written_clipdistance;

   assert(emit->clip_dist_out_index != INVALID_INDEX);
   assert(emit->clip_dist_tmp_index != INVALID_INDEX);

   /**
    * Temporary reset the temporary clip dist register index so
    * that the copy to the real clip dist register will not
    * attempt to copy to the temporary register again
    */
   emit->clip_dist_tmp_index = INVALID_INDEX;

   for (i = 0; i < 2 && num_written_clipdist > 0; i++, num_written_clipdist-=4) {

      tmp_clip_dist_src = make_src_temp_reg(clip_dist_tmp_index + i);

      /**
       * copy to the shadow copy for use by varying variable and
       * stream output. All clip distances
       * will be written regardless of the enabled clipping planes.
       */
      clip_dist_dst = make_dst_reg(TGSI_FILE_OUTPUT,
                                   emit->clip_dist_so_index + i);

      /* MOV clip_dist_so, tmp_clip_dist */
      emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &clip_dist_dst,
                           &tmp_clip_dist_src, FALSE);

      /**
       * copy those clip distances to enabled clipping planes
       * to CLIPDIST registers for clipping
       */
      if (clip_plane_enable & 0xf) {
         clip_dist_dst = make_dst_reg(TGSI_FILE_OUTPUT,
                                      emit->clip_dist_out_index + i);
         clip_dist_dst = writemask_dst(&clip_dist_dst, clip_plane_enable & 0xf);

         /* MOV CLIPDIST, tmp_clip_dist */
         emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &clip_dist_dst,
                              &tmp_clip_dist_src, FALSE);
      }
      /* four clip planes per clip register */
      clip_plane_enable >>= 4;
   }
   /**
    * set the temporary clip dist register index back to the
    * temporary index for the next vertex
    */
   emit->clip_dist_tmp_index = clip_dist_tmp_index;
}

/* Declare clip distance output registers for user-defined clip planes
 * or the TGSI_CLIPVERTEX output.
 */
static void
emit_clip_distance_declarations(struct svga_shader_emitter_v10 *emit)
{
   unsigned num_clip_planes = util_bitcount(emit->key.clip_plane_enable);
   unsigned index = emit->num_outputs;
   unsigned plane_mask;

   assert(emit->unit == PIPE_SHADER_VERTEX ||
          emit->unit == PIPE_SHADER_GEOMETRY);
   assert(num_clip_planes <= 8);

   if (emit->clip_mode != CLIP_LEGACY &&
       emit->clip_mode != CLIP_VERTEX) {
      return;
   }

   if (num_clip_planes == 0)
      return;

   /* Declare one or two clip output registers.  The number of components
    * in the mask reflects the number of clip planes.  For example, if 5
    * clip planes are needed, we'll declare outputs similar to:
    * dcl_output_siv o2.xyzw, clip_distance
    * dcl_output_siv o3.x, clip_distance
    */
   emit->clip_dist_out_index = index; /* save the starting clip dist reg index */

   plane_mask = (1 << num_clip_planes) - 1;
   if (plane_mask & 0xf) {
      unsigned cmask = plane_mask & VGPU10_OPERAND_4_COMPONENT_MASK_ALL;
      emit_output_declaration(emit, VGPU10_OPCODE_DCL_OUTPUT_SIV, index,
                              VGPU10_NAME_CLIP_DISTANCE, cmask);
      emit->num_outputs++;
   }
   if (plane_mask & 0xf0) {
      unsigned cmask = (plane_mask >> 4) & VGPU10_OPERAND_4_COMPONENT_MASK_ALL;
      emit_output_declaration(emit, VGPU10_OPCODE_DCL_OUTPUT_SIV, index + 1,
                              VGPU10_NAME_CLIP_DISTANCE, cmask);
      emit->num_outputs++;
   }
}


/**
 * Emit the instructions for writing to the clip distance registers
 * to handle legacy/automatic clip planes.
 * For each clip plane, the distance is the dot product of the vertex
 * position (found in TEMP[vpos_tmp_index]) and the clip plane coefficients.
 * This is not used when the shader has an explicit CLIPVERTEX or CLIPDISTANCE
 * output registers already declared.
 */
static void
emit_clip_distance_from_vpos(struct svga_shader_emitter_v10 *emit,
                             unsigned vpos_tmp_index)
{
   unsigned i, num_clip_planes = util_bitcount(emit->key.clip_plane_enable);

   assert(emit->clip_mode == CLIP_LEGACY);
   assert(num_clip_planes <= 8);

   assert(emit->unit == PIPE_SHADER_VERTEX ||
          emit->unit == PIPE_SHADER_GEOMETRY);

   for (i = 0; i < num_clip_planes; i++) {
      struct tgsi_full_dst_register dst;
      struct tgsi_full_src_register plane_src, vpos_src;
      unsigned reg_index = emit->clip_dist_out_index + i / 4;
      unsigned comp = i % 4;
      unsigned writemask = VGPU10_OPERAND_4_COMPONENT_MASK_X << comp;

      /* create dst, src regs */
      dst = make_dst_reg(TGSI_FILE_OUTPUT, reg_index);
      dst = writemask_dst(&dst, writemask);

      plane_src = make_src_const_reg(emit->clip_plane_const[i]);
      vpos_src = make_src_temp_reg(vpos_tmp_index);

      /* DP4 clip_dist, plane, vpos */
      emit_instruction_op2(emit, VGPU10_OPCODE_DP4, &dst,
                           &plane_src, &vpos_src, FALSE);
   }
}


/**
 * Emit the instructions for computing the clip distance results from
 * the clip vertex temporary.
 * For each clip plane, the distance is the dot product of the clip vertex
 * position (found in a temp reg) and the clip plane coefficients.
 */
static void
emit_clip_vertex_instructions(struct svga_shader_emitter_v10 *emit)
{
   const unsigned num_clip = util_bitcount(emit->key.clip_plane_enable);
   unsigned i;
   struct tgsi_full_dst_register dst;
   struct tgsi_full_src_register clipvert_src;
   const unsigned clip_vertex_tmp = emit->clip_vertex_tmp_index;

   assert(emit->unit == PIPE_SHADER_VERTEX ||
          emit->unit == PIPE_SHADER_GEOMETRY);

   assert(emit->clip_mode == CLIP_VERTEX);

   clipvert_src = make_src_temp_reg(clip_vertex_tmp);

   for (i = 0; i < num_clip; i++) {
      struct tgsi_full_src_register plane_src;
      unsigned reg_index = emit->clip_dist_out_index + i / 4;
      unsigned comp = i % 4;
      unsigned writemask = VGPU10_OPERAND_4_COMPONENT_MASK_X << comp;

      /* create dst, src regs */
      dst = make_dst_reg(TGSI_FILE_OUTPUT, reg_index);
      dst = writemask_dst(&dst, writemask);

      plane_src = make_src_const_reg(emit->clip_plane_const[i]);

      /* DP4 clip_dist, plane, vpos */
      emit_instruction_op2(emit, VGPU10_OPCODE_DP4, &dst,
                           &plane_src, &clipvert_src, FALSE);
   }

   /* copy temporary clip vertex register to the clip vertex register */

   assert(emit->clip_vertex_out_index != INVALID_INDEX);

   /**
    * temporary reset the temporary clip vertex register index so
    * that copy to the clip vertex register will not attempt
    * to copy to the temporary register again
    */
   emit->clip_vertex_tmp_index = INVALID_INDEX;

   /* MOV clip_vertex, clip_vertex_tmp */
   dst = make_dst_reg(TGSI_FILE_OUTPUT, emit->clip_vertex_out_index);
   emit_instruction_op1(emit, VGPU10_OPCODE_MOV,
                        &dst, &clipvert_src, FALSE);

   /**
    * set the temporary clip vertex register index back to the
    * temporary index for the next vertex
    */
   emit->clip_vertex_tmp_index = clip_vertex_tmp;
}

/**
 * Emit code to convert RGBA to BGRA
 */
static void
emit_swap_r_b(struct svga_shader_emitter_v10 *emit,
                     const struct tgsi_full_dst_register *dst,
                     const struct tgsi_full_src_register *src)
{
   struct tgsi_full_src_register bgra_src =
      swizzle_src(src, TGSI_SWIZZLE_Z, TGSI_SWIZZLE_Y, TGSI_SWIZZLE_X, TGSI_SWIZZLE_W);

   begin_emit_instruction(emit);
   emit_opcode(emit, VGPU10_OPCODE_MOV, FALSE);
   emit_dst_register(emit, dst);
   emit_src_register(emit, &bgra_src);
   end_emit_instruction(emit);
}


/** Convert from 10_10_10_2 normalized to 10_10_10_2_snorm */
static void
emit_puint_to_snorm(struct svga_shader_emitter_v10 *emit,
                    const struct tgsi_full_dst_register *dst,
                    const struct tgsi_full_src_register *src)
{
   struct tgsi_full_src_register half = make_immediate_reg_float(emit, 0.5f);
   struct tgsi_full_src_register two =
      make_immediate_reg_float4(emit, 2.0f, 2.0f, 2.0f, 3.0f);
   struct tgsi_full_src_register neg_two =
      make_immediate_reg_float4(emit, -2.0f, -2.0f, -2.0f, -1.66666f);

   unsigned val_tmp = get_temp_index(emit);
   struct tgsi_full_dst_register val_dst = make_dst_temp_reg(val_tmp);
   struct tgsi_full_src_register val_src = make_src_temp_reg(val_tmp);

   unsigned bias_tmp = get_temp_index(emit);
   struct tgsi_full_dst_register bias_dst = make_dst_temp_reg(bias_tmp);
   struct tgsi_full_src_register bias_src = make_src_temp_reg(bias_tmp);

   /* val = src * 2.0 */
   emit_instruction_op2(emit, VGPU10_OPCODE_MUL, &val_dst,
                        src, &two, FALSE);

   /* bias = src > 0.5 */
   emit_instruction_op2(emit, VGPU10_OPCODE_GE, &bias_dst,
                        src, &half, FALSE);

   /* bias = bias & -2.0 */
   emit_instruction_op2(emit, VGPU10_OPCODE_AND, &bias_dst,
                        &bias_src, &neg_two, FALSE);

   /* dst = val + bias */
   emit_instruction_op2(emit, VGPU10_OPCODE_ADD, dst,
                        &val_src, &bias_src, FALSE);

   free_temp_indexes(emit);
}


/** Convert from 10_10_10_2_unorm to 10_10_10_2_uscaled */
static void
emit_puint_to_uscaled(struct svga_shader_emitter_v10 *emit,
                      const struct tgsi_full_dst_register *dst,
                      const struct tgsi_full_src_register *src)
{
   struct tgsi_full_src_register scale =
      make_immediate_reg_float4(emit, 1023.0f, 1023.0f, 1023.0f, 3.0f);

   /* dst = src * scale */
   emit_instruction_op2(emit, VGPU10_OPCODE_MUL, dst, src, &scale, FALSE);
}


/** Convert from R32_UINT to 10_10_10_2_sscaled */
static void
emit_puint_to_sscaled(struct svga_shader_emitter_v10 *emit,
                      const struct tgsi_full_dst_register *dst,
                      const struct tgsi_full_src_register *src)
{
   struct tgsi_full_src_register lshift =
      make_immediate_reg_int4(emit, 22, 12, 2, 0);
   struct tgsi_full_src_register rshift =
      make_immediate_reg_int4(emit, 22, 22, 22, 30);

   struct tgsi_full_src_register src_xxxx = scalar_src(src, TGSI_SWIZZLE_X);

   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);
   struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);

   /*
    * r = (pixel << 22) >> 22;   # signed int in [511, -512]
    * g = (pixel << 12) >> 22;   # signed int in [511, -512]
    * b = (pixel <<  2) >> 22;   # signed int in [511, -512]
    * a = (pixel <<  0) >> 30;   # signed int in [1, -2]
    * dst = i_to_f(r,g,b,a);     # convert to float
    */
   emit_instruction_op2(emit, VGPU10_OPCODE_ISHL, &tmp_dst,
                        &src_xxxx, &lshift, FALSE);
   emit_instruction_op2(emit, VGPU10_OPCODE_ISHR, &tmp_dst,
                        &tmp_src, &rshift, FALSE);
   emit_instruction_op1(emit, VGPU10_OPCODE_ITOF, dst, &tmp_src, FALSE);

   free_temp_indexes(emit);
}


/**
 * Emit code for TGSI_OPCODE_ABS instruction.
 */
static boolean
emit_abs(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /* dst = ABS(s0):
    *   dst = abs(s0)
    * Translates into:
    *   MOV dst, abs(s0)
    */
   struct tgsi_full_src_register abs_src0 = absolute_src(&inst->Src[0]);

   /* MOV dst, abs(s0) */
   emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &inst->Dst[0],
                        &abs_src0, inst->Instruction.Saturate);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_ARL or TGSI_OPCODE_UARL instruction.
 */
static boolean
emit_arl_uarl(struct svga_shader_emitter_v10 *emit,
              const struct tgsi_full_instruction *inst)
{
   unsigned index = inst->Dst[0].Register.Index;
   struct tgsi_full_dst_register dst;
   unsigned opcode;

   assert(index < MAX_VGPU10_ADDR_REGS);
   dst = make_dst_temp_reg(emit->address_reg_index[index]);

   /* ARL dst, s0
    * Translates into:
    * FTOI address_tmp, s0
    *
    * UARL dst, s0
    * Translates into:
    * MOV address_tmp, s0
    */
   if (inst->Instruction.Opcode == TGSI_OPCODE_ARL)
      opcode = VGPU10_OPCODE_FTOI;
   else
      opcode = VGPU10_OPCODE_MOV;

   emit_instruction_op1(emit, opcode, &dst, &inst->Src[0], FALSE);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_CAL instruction.
 */
static boolean
emit_cal(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   unsigned label = inst->Label.Label;
   VGPU10OperandToken0 operand;
   operand.value = 0;
   operand.operandType = VGPU10_OPERAND_TYPE_LABEL;

   begin_emit_instruction(emit);
   emit_dword(emit, operand.value);
   emit_dword(emit, label);
   end_emit_instruction(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_IABS instruction.
 */
static boolean
emit_iabs(struct svga_shader_emitter_v10 *emit,
          const struct tgsi_full_instruction *inst)
{
   /* dst.x = (src0.x < 0) ? -src0.x : src0.x
    * dst.y = (src0.y < 0) ? -src0.y : src0.y
    * dst.z = (src0.z < 0) ? -src0.z : src0.z
    * dst.w = (src0.w < 0) ? -src0.w : src0.w
    *
    * Translates into
    *   IMAX dst, src, neg(src)
    */
   struct tgsi_full_src_register neg_src = negate_src(&inst->Src[0]);
   emit_instruction_op2(emit, VGPU10_OPCODE_IMAX, &inst->Dst[0],
                        &inst->Src[0], &neg_src, FALSE);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_CMP instruction.
 */
static boolean
emit_cmp(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /* dst.x = (src0.x < 0) ? src1.x : src2.x
    * dst.y = (src0.y < 0) ? src1.y : src2.y
    * dst.z = (src0.z < 0) ? src1.z : src2.z
    * dst.w = (src0.w < 0) ? src1.w : src2.w
    *
    * Translates into
    *   LT tmp, src0, 0.0
    *   MOVC dst, tmp, src1, src2
    */
   struct tgsi_full_src_register zero = make_immediate_reg_float(emit, 0.0f);
   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
   struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);

   emit_instruction_op2(emit, VGPU10_OPCODE_LT, &tmp_dst,
                        &inst->Src[0], &zero, FALSE);
   emit_instruction_op3(emit, VGPU10_OPCODE_MOVC, &inst->Dst[0],
                        &tmp_src, &inst->Src[1], &inst->Src[2],
                        inst->Instruction.Saturate);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_DP2A instruction.
 */
static boolean
emit_dp2a(struct svga_shader_emitter_v10 *emit,
          const struct tgsi_full_instruction *inst)
{
   /* dst.x = src0.x * src1.x + src0.y * src1.y + src2.x
    * dst.y = src0.x * src1.x + src0.y * src1.y + src2.x
    * dst.z = src0.x * src1.x + src0.y * src1.y + src2.x
    * dst.w = src0.x * src1.x + src0.y * src1.y + src2.x
    * Translate into
    *   MAD tmp.x, s0.y, s1.y, s2.x
    *   MAD tmp.x, s0.x, s1.x, tmp.x
    *   MOV dst.xyzw, tmp.xxxx
    */
   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
   struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);

   struct tgsi_full_src_register tmp_src_xxxx =
      scalar_src(&tmp_src, TGSI_SWIZZLE_X);
   struct tgsi_full_dst_register tmp_dst_x =
      writemask_dst(&tmp_dst, TGSI_WRITEMASK_X);

   struct tgsi_full_src_register src0_xxxx =
      scalar_src(&inst->Src[0], TGSI_SWIZZLE_X);
   struct tgsi_full_src_register src0_yyyy =
      scalar_src(&inst->Src[0], TGSI_SWIZZLE_Y);
   struct tgsi_full_src_register src1_xxxx =
      scalar_src(&inst->Src[1], TGSI_SWIZZLE_X);
   struct tgsi_full_src_register src1_yyyy =
      scalar_src(&inst->Src[1], TGSI_SWIZZLE_Y);
   struct tgsi_full_src_register src2_xxxx =
      scalar_src(&inst->Src[2], TGSI_SWIZZLE_X);

   emit_instruction_op3(emit, VGPU10_OPCODE_MAD, &tmp_dst_x, &src0_yyyy,
                        &src1_yyyy, &src2_xxxx, FALSE);
   emit_instruction_op3(emit, VGPU10_OPCODE_MAD, &tmp_dst_x, &src0_xxxx,
                        &src1_xxxx, &tmp_src_xxxx, FALSE);
   emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &inst->Dst[0],
                        &tmp_src_xxxx, inst->Instruction.Saturate);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_DPH instruction.
 */
static boolean
emit_dph(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /*
    * DP3 tmp, s0, s1
    * ADD dst, tmp, s1.wwww
    */

   struct tgsi_full_src_register s1_wwww =
      swizzle_src(&inst->Src[1], TGSI_SWIZZLE_W, TGSI_SWIZZLE_W,
                  TGSI_SWIZZLE_W, TGSI_SWIZZLE_W);

   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
   struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);

   /* DP3 tmp, s0, s1 */
   emit_instruction_op2(emit, VGPU10_OPCODE_DP3, &tmp_dst, &inst->Src[0],
                        &inst->Src[1], FALSE);

   /* ADD dst, tmp, s1.wwww */
   emit_instruction_op2(emit, VGPU10_OPCODE_ADD, &inst->Dst[0], &tmp_src,
                        &s1_wwww, inst->Instruction.Saturate);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_DST instruction.
 */
static boolean
emit_dst(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /*
    * dst.x = 1
    * dst.y = src0.y * src1.y
    * dst.z = src0.z
    * dst.w = src1.w
    */

   struct tgsi_full_src_register s0_yyyy =
      scalar_src(&inst->Src[0], TGSI_SWIZZLE_Y);
   struct tgsi_full_src_register s0_zzzz =
      scalar_src(&inst->Src[0], TGSI_SWIZZLE_Z);
   struct tgsi_full_src_register s1_yyyy =
      scalar_src(&inst->Src[1], TGSI_SWIZZLE_Y);
   struct tgsi_full_src_register s1_wwww =
      scalar_src(&inst->Src[1], TGSI_SWIZZLE_W);

   /*
    * If dst and either src0 and src1 are the same we need
    * to create a temporary for it and insert a extra move.
    */
   unsigned tmp_move = get_temp_index(emit);
   struct tgsi_full_src_register move_src = make_src_temp_reg(tmp_move);
   struct tgsi_full_dst_register move_dst = make_dst_temp_reg(tmp_move);

   /* MOV dst.x, 1.0 */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_X) {
      struct tgsi_full_dst_register dst_x =
         writemask_dst(&move_dst, TGSI_WRITEMASK_X);
      struct tgsi_full_src_register one = make_immediate_reg_float(emit, 1.0f);

      emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &dst_x, &one, FALSE);
   }

   /* MUL dst.y, s0.y, s1.y */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_Y) {
      struct tgsi_full_dst_register dst_y =
         writemask_dst(&move_dst, TGSI_WRITEMASK_Y);

      emit_instruction_op2(emit, VGPU10_OPCODE_MUL, &dst_y, &s0_yyyy,
                           &s1_yyyy, inst->Instruction.Saturate);
   }

   /* MOV dst.z, s0.z */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_Z) {
      struct tgsi_full_dst_register dst_z =
         writemask_dst(&move_dst, TGSI_WRITEMASK_Z);

      emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &dst_z, &s0_zzzz,
                           inst->Instruction.Saturate);
  }

   /* MOV dst.w, s1.w */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_W) {
      struct tgsi_full_dst_register dst_w =
         writemask_dst(&move_dst, TGSI_WRITEMASK_W);

      emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &dst_w, &s1_wwww,
                           inst->Instruction.Saturate);
   }

   emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &inst->Dst[0], &move_src,
                        FALSE);
   free_temp_indexes(emit);

   return TRUE;
}



/**
 * Emit code for TGSI_OPCODE_ENDPRIM (GS only)
 */
static boolean
emit_endprim(struct svga_shader_emitter_v10 *emit,
             const struct tgsi_full_instruction *inst)
{
   assert(emit->unit == PIPE_SHADER_GEOMETRY);

   /* We can't use emit_simple() because the TGSI instruction has one
    * operand (vertex stream number) which we must ignore for VGPU10.
    */
   begin_emit_instruction(emit);
   emit_opcode(emit, VGPU10_OPCODE_CUT, FALSE);
   end_emit_instruction(emit);
   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_EX2 (2^x) instruction.
 */
static boolean
emit_ex2(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /* Note that TGSI_OPCODE_EX2 computes only one value from src.x
    * while VGPU10 computes four values.
    *
    * dst = EX2(src):
    *   dst.xyzw = 2.0 ^ src.x
    */

   struct tgsi_full_src_register src_xxxx =
      swizzle_src(&inst->Src[0], TGSI_SWIZZLE_X, TGSI_SWIZZLE_X,
                  TGSI_SWIZZLE_X, TGSI_SWIZZLE_X);

   /* EXP tmp, s0.xxxx */
   emit_instruction_op1(emit, VGPU10_OPCODE_EXP, &inst->Dst[0], &src_xxxx,
                        inst->Instruction.Saturate);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_EXP instruction.
 */
static boolean
emit_exp(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /*
    * dst.x = 2 ^ floor(s0.x)
    * dst.y = s0.x - floor(s0.x)
    * dst.z = 2 ^ s0.x
    * dst.w = 1.0
    */

   struct tgsi_full_src_register src_xxxx =
      scalar_src(&inst->Src[0], TGSI_SWIZZLE_X);
   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
   struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);

   /*
    * If dst and src are the same we need to create
    * a temporary for it and insert a extra move.
    */
   unsigned tmp_move = get_temp_index(emit);
   struct tgsi_full_src_register move_src = make_src_temp_reg(tmp_move);
   struct tgsi_full_dst_register move_dst = make_dst_temp_reg(tmp_move);

   /* only use X component of temp reg */
   tmp_dst = writemask_dst(&tmp_dst, TGSI_WRITEMASK_X);
   tmp_src = scalar_src(&tmp_src, TGSI_SWIZZLE_X);

   /* ROUND_NI tmp.x, s0.x */
   emit_instruction_op1(emit, VGPU10_OPCODE_ROUND_NI, &tmp_dst,
                        &src_xxxx, FALSE); /* round to -infinity */

   /* EXP dst.x, tmp.x */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_X) {
      struct tgsi_full_dst_register dst_x =
         writemask_dst(&move_dst, TGSI_WRITEMASK_X);

      emit_instruction_op1(emit, VGPU10_OPCODE_EXP, &dst_x, &tmp_src,
                           inst->Instruction.Saturate);
   }

   /* ADD dst.y, s0.x, -tmp */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_Y) {
      struct tgsi_full_dst_register dst_y =
         writemask_dst(&move_dst, TGSI_WRITEMASK_Y);
      struct tgsi_full_src_register neg_tmp_src = negate_src(&tmp_src);

      emit_instruction_op2(emit, VGPU10_OPCODE_ADD, &dst_y, &src_xxxx,
                           &neg_tmp_src, inst->Instruction.Saturate);
   }

   /* EXP dst.z, s0.x */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_Z) {
      struct tgsi_full_dst_register dst_z =
         writemask_dst(&move_dst, TGSI_WRITEMASK_Z);

      emit_instruction_op1(emit, VGPU10_OPCODE_EXP, &dst_z, &src_xxxx,
                           inst->Instruction.Saturate);
   }

   /* MOV dst.w, 1.0 */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_W) {
      struct tgsi_full_dst_register dst_w =
         writemask_dst(&move_dst, TGSI_WRITEMASK_W);
      struct tgsi_full_src_register one = make_immediate_reg_float(emit, 1.0f);

      emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &dst_w, &one,
                           FALSE);
   }

   emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &inst->Dst[0], &move_src,
                        FALSE);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_IF instruction.
 */
static boolean
emit_if(struct svga_shader_emitter_v10 *emit,
        const struct tgsi_full_instruction *inst)
{
   VGPU10OpcodeToken0 opcode0;

   /* The src register should be a scalar */
   assert(inst->Src[0].Register.SwizzleX == inst->Src[0].Register.SwizzleY &&
          inst->Src[0].Register.SwizzleX == inst->Src[0].Register.SwizzleZ &&
          inst->Src[0].Register.SwizzleX == inst->Src[0].Register.SwizzleW);

   /* The only special thing here is that we need to set the
    * VGPU10_INSTRUCTION_TEST_NONZERO flag since we want to test if
    * src.x is non-zero.
    */
   opcode0.value = 0;
   opcode0.opcodeType = VGPU10_OPCODE_IF;
   opcode0.testBoolean = VGPU10_INSTRUCTION_TEST_NONZERO;

   begin_emit_instruction(emit);
   emit_dword(emit, opcode0.value);
   emit_src_register(emit, &inst->Src[0]);
   end_emit_instruction(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_KILL_IF instruction (kill fragment if any of
 * the register components are negative).
 */
static boolean
emit_kill_if(struct svga_shader_emitter_v10 *emit,
             const struct tgsi_full_instruction *inst)
{
   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
   struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);

   struct tgsi_full_src_register zero = make_immediate_reg_float(emit, 0.0f);

   struct tgsi_full_dst_register tmp_dst_x =
      writemask_dst(&tmp_dst, TGSI_WRITEMASK_X);
   struct tgsi_full_src_register tmp_src_xxxx =
      scalar_src(&tmp_src, TGSI_SWIZZLE_X);

   /* tmp = src[0] < 0.0 */
   emit_instruction_op2(emit, VGPU10_OPCODE_LT, &tmp_dst, &inst->Src[0],
                        &zero, FALSE);

   if (!same_swizzle_terms(&inst->Src[0])) {
      /* If the swizzle is not XXXX, YYYY, ZZZZ or WWWW we need to
       * logically OR the swizzle terms.  Most uses of KILL_IF only
       * test one channel so it's good to avoid these extra steps.
       */
      struct tgsi_full_src_register tmp_src_yyyy =
         scalar_src(&tmp_src, TGSI_SWIZZLE_Y);
      struct tgsi_full_src_register tmp_src_zzzz =
         scalar_src(&tmp_src, TGSI_SWIZZLE_Z);
      struct tgsi_full_src_register tmp_src_wwww =
         scalar_src(&tmp_src, TGSI_SWIZZLE_W);

      emit_instruction_op2(emit, VGPU10_OPCODE_OR, &tmp_dst_x, &tmp_src_xxxx,
                           &tmp_src_yyyy, FALSE);
      emit_instruction_op2(emit, VGPU10_OPCODE_OR, &tmp_dst_x, &tmp_src_xxxx,
                           &tmp_src_zzzz, FALSE);
      emit_instruction_op2(emit, VGPU10_OPCODE_OR, &tmp_dst_x, &tmp_src_xxxx,
                           &tmp_src_wwww, FALSE);
   }

   begin_emit_instruction(emit);
   emit_discard_opcode(emit, TRUE); /* discard if src0.x is non-zero */
   emit_src_register(emit, &tmp_src_xxxx);
   end_emit_instruction(emit);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_KILL instruction (unconditional discard).
 */
static boolean
emit_kill(struct svga_shader_emitter_v10 *emit,
          const struct tgsi_full_instruction *inst)
{
   struct tgsi_full_src_register zero = make_immediate_reg_float(emit, 0.0f);

   /* DISCARD if 0.0 is zero */
   begin_emit_instruction(emit);
   emit_discard_opcode(emit, FALSE);
   emit_src_register(emit, &zero);
   end_emit_instruction(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_LG2 instruction.
 */
static boolean
emit_lg2(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /* Note that TGSI_OPCODE_LG2 computes only one value from src.x
    * while VGPU10 computes four values.
    *
    * dst = LG2(src):
    *   dst.xyzw = log2(src.x)
    */

   struct tgsi_full_src_register src_xxxx =
      swizzle_src(&inst->Src[0], TGSI_SWIZZLE_X, TGSI_SWIZZLE_X,
                  TGSI_SWIZZLE_X, TGSI_SWIZZLE_X);

   /* LOG tmp, s0.xxxx */
   emit_instruction_op1(emit, VGPU10_OPCODE_LOG, &inst->Dst[0], &src_xxxx,
                        inst->Instruction.Saturate);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_LIT instruction.
 */
static boolean
emit_lit(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   struct tgsi_full_src_register one = make_immediate_reg_float(emit, 1.0f);

   /*
    * If dst and src are the same we need to create
    * a temporary for it and insert a extra move.
    */
   unsigned tmp_move = get_temp_index(emit);
   struct tgsi_full_src_register move_src = make_src_temp_reg(tmp_move);
   struct tgsi_full_dst_register move_dst = make_dst_temp_reg(tmp_move);

   /*
    * dst.x = 1
    * dst.y = max(src.x, 0)
    * dst.z = (src.x > 0) ? max(src.y, 0)^{clamp(src.w, -128, 128))} : 0
    * dst.w = 1
    */

   /* MOV dst.x, 1.0 */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_X) {
      struct tgsi_full_dst_register dst_x =
         writemask_dst(&move_dst, TGSI_WRITEMASK_X);
      emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &dst_x, &one, FALSE);
   }

   /* MOV dst.w, 1.0 */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_W) {
      struct tgsi_full_dst_register dst_w =
         writemask_dst(&move_dst, TGSI_WRITEMASK_W);
      emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &dst_w, &one, FALSE);
   }

   /* MAX dst.y, src.x, 0.0 */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_Y) {
      struct tgsi_full_dst_register dst_y =
         writemask_dst(&move_dst, TGSI_WRITEMASK_Y);
      struct tgsi_full_src_register zero =
         make_immediate_reg_float(emit, 0.0f);
      struct tgsi_full_src_register src_xxxx =
         swizzle_src(&inst->Src[0], TGSI_SWIZZLE_X, TGSI_SWIZZLE_X,
                     TGSI_SWIZZLE_X, TGSI_SWIZZLE_X);

      emit_instruction_op2(emit, VGPU10_OPCODE_MAX, &dst_y, &src_xxxx,
                           &zero, inst->Instruction.Saturate);
   }

   /*
    * tmp1 = clamp(src.w, -128, 128);
    *   MAX tmp1, src.w, -128
    *   MIN tmp1, tmp1, 128
    *
    * tmp2 = max(tmp2, 0);
    *   MAX tmp2, src.y, 0
    *
    * tmp1 = pow(tmp2, tmp1);
    *   LOG tmp2, tmp2
    *   MUL tmp1, tmp2, tmp1
    *   EXP tmp1, tmp1
    *
    * tmp1 = (src.w == 0) ? 1 : tmp1;
    *   EQ tmp2, 0, src.w
    *   MOVC tmp1, tmp2, 1.0, tmp1
    *
    * dst.z = (0 < src.x) ? tmp1 : 0;
    *   LT tmp2, 0, src.x
    *   MOVC dst.z, tmp2, tmp1, 0.0
    */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_Z) {
      struct tgsi_full_dst_register dst_z =
         writemask_dst(&move_dst, TGSI_WRITEMASK_Z);

      unsigned tmp1 = get_temp_index(emit);
      struct tgsi_full_src_register tmp1_src = make_src_temp_reg(tmp1);
      struct tgsi_full_dst_register tmp1_dst = make_dst_temp_reg(tmp1);
      unsigned tmp2 = get_temp_index(emit);
      struct tgsi_full_src_register tmp2_src = make_src_temp_reg(tmp2);
      struct tgsi_full_dst_register tmp2_dst = make_dst_temp_reg(tmp2);

      struct tgsi_full_src_register src_xxxx =
         scalar_src(&inst->Src[0], TGSI_SWIZZLE_X);
      struct tgsi_full_src_register src_yyyy =
         scalar_src(&inst->Src[0], TGSI_SWIZZLE_Y);
      struct tgsi_full_src_register src_wwww =
         scalar_src(&inst->Src[0], TGSI_SWIZZLE_W);

      struct tgsi_full_src_register zero =
         make_immediate_reg_float(emit, 0.0f);
      struct tgsi_full_src_register lowerbound =
         make_immediate_reg_float(emit, -128.0f);
      struct tgsi_full_src_register upperbound =
         make_immediate_reg_float(emit, 128.0f);

      emit_instruction_op2(emit, VGPU10_OPCODE_MAX, &tmp1_dst, &src_wwww,
                           &lowerbound, FALSE);
      emit_instruction_op2(emit, VGPU10_OPCODE_MIN, &tmp1_dst, &tmp1_src,
                           &upperbound, FALSE);
      emit_instruction_op2(emit, VGPU10_OPCODE_MAX, &tmp2_dst, &src_yyyy,
                           &zero, FALSE);

      /* POW tmp1, tmp2, tmp1 */
      /* LOG tmp2, tmp2 */
      emit_instruction_op1(emit, VGPU10_OPCODE_LOG, &tmp2_dst, &tmp2_src,
                           FALSE);

      /* MUL tmp1, tmp2, tmp1 */
      emit_instruction_op2(emit, VGPU10_OPCODE_MUL, &tmp1_dst, &tmp2_src,
                           &tmp1_src, FALSE);

      /* EXP tmp1, tmp1 */
      emit_instruction_op1(emit, VGPU10_OPCODE_EXP, &tmp1_dst, &tmp1_src,
                           FALSE);

      /* EQ tmp2, 0, src.w */
      emit_instruction_op2(emit, VGPU10_OPCODE_EQ, &tmp2_dst, &zero,
                           &src_wwww, FALSE);
      /* MOVC tmp1.z, tmp2, tmp1, 1.0 */
      emit_instruction_op3(emit, VGPU10_OPCODE_MOVC, &tmp1_dst,
                           &tmp2_src, &one, &tmp1_src, FALSE);

      /* LT tmp2, 0, src.x */
      emit_instruction_op2(emit, VGPU10_OPCODE_LT, &tmp2_dst, &zero,
                           &src_xxxx, FALSE);
      /* MOVC dst.z, tmp2, tmp1, 0.0 */
      emit_instruction_op3(emit, VGPU10_OPCODE_MOVC, &dst_z,
                           &tmp2_src, &tmp1_src, &zero, FALSE);
   }

   emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &inst->Dst[0], &move_src,
                        FALSE);
   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_LOG instruction.
 */
static boolean
emit_log(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /*
    * dst.x = floor(lg2(abs(s0.x)))
    * dst.y = abs(s0.x) / (2 ^ floor(lg2(abs(s0.x))))
    * dst.z = lg2(abs(s0.x))
    * dst.w = 1.0
    */

   struct tgsi_full_src_register src_xxxx =
      scalar_src(&inst->Src[0], TGSI_SWIZZLE_X);
   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
   struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);
   struct tgsi_full_src_register abs_src_xxxx = absolute_src(&src_xxxx);

   /* only use X component of temp reg */
   tmp_dst = writemask_dst(&tmp_dst, TGSI_WRITEMASK_X);
   tmp_src = scalar_src(&tmp_src, TGSI_SWIZZLE_X);

   /* LOG tmp.x, abs(s0.x) */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_XYZ) {
      emit_instruction_op1(emit, VGPU10_OPCODE_LOG, &tmp_dst,
                          &abs_src_xxxx, FALSE);
   }

   /* MOV dst.z, tmp.x */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_Z) {
      struct tgsi_full_dst_register dst_z =
         writemask_dst(&inst->Dst[0], TGSI_WRITEMASK_Z);

      emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &dst_z,
                           &tmp_src, inst->Instruction.Saturate);
   }

   /* FLR tmp.x, tmp.x */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_XY) {
      emit_instruction_op1(emit, VGPU10_OPCODE_ROUND_NI, &tmp_dst,
                           &tmp_src, FALSE);
   }

   /* MOV dst.x, tmp.x */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_X) {
      struct tgsi_full_dst_register dst_x =
         writemask_dst(&inst->Dst[0], TGSI_WRITEMASK_X);

      emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &dst_x, &tmp_src,
                           inst->Instruction.Saturate);
   }

   /* EXP tmp.x, tmp.x */
   /* DIV dst.y, abs(s0.x), tmp.x */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_Y) {
      struct tgsi_full_dst_register dst_y =
         writemask_dst(&inst->Dst[0], TGSI_WRITEMASK_Y);

      emit_instruction_op1(emit, VGPU10_OPCODE_EXP, &tmp_dst, &tmp_src,
                           FALSE);
      emit_instruction_op2(emit, VGPU10_OPCODE_DIV, &dst_y, &abs_src_xxxx,
                           &tmp_src, inst->Instruction.Saturate);
   }

   /* MOV dst.w, 1.0 */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_W) {
      struct tgsi_full_dst_register dst_w =
         writemask_dst(&inst->Dst[0], TGSI_WRITEMASK_W);
      struct tgsi_full_src_register one =
         make_immediate_reg_float(emit, 1.0f);

      emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &dst_w, &one, FALSE);
   }

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_LRP instruction.
 */
static boolean
emit_lrp(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /* dst = LRP(s0, s1, s2):
    *   dst = s0 * (s1 - s2) + s2
    * Translates into:
    *   SUB tmp, s1, s2;        tmp = s1 - s2
    *   MAD dst, s0, tmp, s2;   dst = s0 * t1 + s2
    */
   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_src_register src_tmp = make_src_temp_reg(tmp);
   struct tgsi_full_dst_register dst_tmp = make_dst_temp_reg(tmp);
   struct tgsi_full_src_register neg_src2 = negate_src(&inst->Src[2]);

   /* ADD tmp, s1, -s2 */
   emit_instruction_op2(emit, VGPU10_OPCODE_ADD, &dst_tmp,
                        &inst->Src[1], &neg_src2, FALSE);

   /* MAD dst, s1, tmp, s3 */
   emit_instruction_op3(emit, VGPU10_OPCODE_MAD, &inst->Dst[0],
                        &inst->Src[0], &src_tmp, &inst->Src[2],
                        inst->Instruction.Saturate);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_POW instruction.
 */
static boolean
emit_pow(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /* Note that TGSI_OPCODE_POW computes only one value from src0.x and
    * src1.x while VGPU10 computes four values.
    *
    * dst = POW(src0, src1):
    *   dst.xyzw = src0.x ^ src1.x
    */
   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
   struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);
   struct tgsi_full_src_register src0_xxxx =
      swizzle_src(&inst->Src[0], TGSI_SWIZZLE_X, TGSI_SWIZZLE_X,
                  TGSI_SWIZZLE_X, TGSI_SWIZZLE_X);
   struct tgsi_full_src_register src1_xxxx =
      swizzle_src(&inst->Src[1], TGSI_SWIZZLE_X, TGSI_SWIZZLE_X,
                  TGSI_SWIZZLE_X, TGSI_SWIZZLE_X);

   /* LOG tmp, s0.xxxx */
   emit_instruction_op1(emit, VGPU10_OPCODE_LOG, &tmp_dst, &src0_xxxx,
                        FALSE);

   /* MUL tmp, tmp, s1.xxxx */
   emit_instruction_op2(emit, VGPU10_OPCODE_MUL, &tmp_dst, &tmp_src,
                        &src1_xxxx, FALSE);

   /* EXP tmp, s0.xxxx */
   emit_instruction_op1(emit, VGPU10_OPCODE_EXP, &inst->Dst[0],
                        &tmp_src, inst->Instruction.Saturate);

   /* free tmp */
   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_RCP (reciprocal) instruction.
 */
static boolean
emit_rcp(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   struct tgsi_full_src_register one = make_immediate_reg_float(emit, 1.0f);

   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
   struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);

   struct tgsi_full_dst_register tmp_dst_x =
      writemask_dst(&tmp_dst, TGSI_WRITEMASK_X);
   struct tgsi_full_src_register tmp_src_xxxx =
      scalar_src(&tmp_src, TGSI_SWIZZLE_X);

   /* DIV tmp.x, 1.0, s0 */
   emit_instruction_op2(emit, VGPU10_OPCODE_DIV, &tmp_dst_x, &one,
                        &inst->Src[0], FALSE);

   /* MOV dst, tmp.xxxx */
   emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &inst->Dst[0],
                        &tmp_src_xxxx, inst->Instruction.Saturate);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_RSQ instruction.
 */
static boolean
emit_rsq(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /* dst = RSQ(src):
    *   dst.xyzw = 1 / sqrt(src.x)
    * Translates into:
    *   RSQ tmp, src.x
    *   MOV dst, tmp.xxxx
    */

   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
   struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);

   struct tgsi_full_dst_register tmp_dst_x =
      writemask_dst(&tmp_dst, TGSI_WRITEMASK_X);
   struct tgsi_full_src_register tmp_src_xxxx =
      scalar_src(&tmp_src, TGSI_SWIZZLE_X);

   /* RSQ tmp, src.x */
   emit_instruction_op1(emit, VGPU10_OPCODE_RSQ, &tmp_dst_x,
                        &inst->Src[0], FALSE);

   /* MOV dst, tmp.xxxx */
   emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &inst->Dst[0],
                        &tmp_src_xxxx, inst->Instruction.Saturate);

   /* free tmp */
   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_SCS instruction.
 */
static boolean
emit_scs(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /* dst.x = cos(src.x)
    * dst.y = sin(src.x)
    * dst.z = 0.0
    * dst.w = 1.0
    */
   struct tgsi_full_dst_register dst_x =
      writemask_dst(&inst->Dst[0], TGSI_WRITEMASK_X);
   struct tgsi_full_dst_register dst_y =
      writemask_dst(&inst->Dst[0], TGSI_WRITEMASK_Y);
   struct tgsi_full_dst_register dst_zw =
      writemask_dst(&inst->Dst[0], TGSI_WRITEMASK_ZW);

   struct tgsi_full_src_register zero_one =
      make_immediate_reg_float4(emit, 0.0f, 0.0f, 0.0f, 1.0f);

   begin_emit_instruction(emit);
   emit_opcode(emit, VGPU10_OPCODE_SINCOS, inst->Instruction.Saturate);
   emit_dst_register(emit, &dst_y);
   emit_dst_register(emit, &dst_x);
   emit_src_register(emit, &inst->Src[0]);
   end_emit_instruction(emit);

   emit_instruction_op1(emit, VGPU10_OPCODE_MOV,
                        &dst_zw, &zero_one, inst->Instruction.Saturate);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_SEQ (Set Equal) instruction.
 */
static boolean
emit_seq(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /* dst = SEQ(s0, s1):
    *   dst = s0 == s1 ? 1.0 : 0.0  (per component)
    * Translates into:
    *   EQ tmp, s0, s1;           tmp = s0 == s1 : 0xffffffff : 0 (per comp)
    *   MOVC dst, tmp, 1.0, 0.0;  dst = tmp ? 1.0 : 0.0 (per component)
    */
   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
   struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);
   struct tgsi_full_src_register zero = make_immediate_reg_float(emit, 0.0f);
   struct tgsi_full_src_register one = make_immediate_reg_float(emit, 1.0f);

   /* EQ tmp, s0, s1 */
   emit_instruction_op2(emit, VGPU10_OPCODE_EQ, &tmp_dst, &inst->Src[0],
                        &inst->Src[1], FALSE);

   /* MOVC dst, tmp, one, zero */
   emit_instruction_op3(emit, VGPU10_OPCODE_MOVC, &inst->Dst[0], &tmp_src,
                        &one, &zero, FALSE);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_SGE (Set Greater than or Equal) instruction.
 */
static boolean
emit_sge(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /* dst = SGE(s0, s1):
    *   dst = s0 >= s1 ? 1.0 : 0.0  (per component)
    * Translates into:
    *   GE tmp, s0, s1;           tmp = s0 >= s1 : 0xffffffff : 0 (per comp)
    *   MOVC dst, tmp, 1.0, 0.0;  dst = tmp ? 1.0 : 0.0 (per component)
    */
   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
   struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);
   struct tgsi_full_src_register zero = make_immediate_reg_float(emit, 0.0f);
   struct tgsi_full_src_register one = make_immediate_reg_float(emit, 1.0f);

   /* GE tmp, s0, s1 */
   emit_instruction_op2(emit, VGPU10_OPCODE_GE, &tmp_dst, &inst->Src[0],
                        &inst->Src[1], FALSE);

   /* MOVC dst, tmp, one, zero */
   emit_instruction_op3(emit, VGPU10_OPCODE_MOVC, &inst->Dst[0], &tmp_src,
                        &one, &zero, FALSE);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_SGT (Set Greater than) instruction.
 */
static boolean
emit_sgt(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /* dst = SGT(s0, s1):
    *   dst = s0 > s1 ? 1.0 : 0.0  (per component)
    * Translates into:
    *   LT tmp, s1, s0;           tmp = s1 < s0 ? 0xffffffff : 0 (per comp)
    *   MOVC dst, tmp, 1.0, 0.0;  dst = tmp ? 1.0 : 0.0 (per component)
    */
   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
   struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);
   struct tgsi_full_src_register zero = make_immediate_reg_float(emit, 0.0f);
   struct tgsi_full_src_register one = make_immediate_reg_float(emit, 1.0f);

   /* LT tmp, s1, s0 */
   emit_instruction_op2(emit, VGPU10_OPCODE_LT, &tmp_dst, &inst->Src[1],
                        &inst->Src[0], FALSE);

   /* MOVC dst, tmp, one, zero */
   emit_instruction_op3(emit, VGPU10_OPCODE_MOVC, &inst->Dst[0], &tmp_src,
                        &one, &zero, FALSE);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_SIN and TGSI_OPCODE_COS instructions.
 */
static boolean
emit_sincos(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
   struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);

   struct tgsi_full_src_register tmp_src_xxxx =
      scalar_src(&tmp_src, TGSI_SWIZZLE_X);
   struct tgsi_full_dst_register tmp_dst_x =
      writemask_dst(&tmp_dst, TGSI_WRITEMASK_X);

   begin_emit_instruction(emit);
   emit_opcode(emit, VGPU10_OPCODE_SINCOS, FALSE);

   if(inst->Instruction.Opcode == TGSI_OPCODE_SIN)
   {
      emit_dst_register(emit, &tmp_dst_x);  /* first destination register */
      emit_null_dst_register(emit);  /* second destination register */
   }
   else {
      emit_null_dst_register(emit);
      emit_dst_register(emit, &tmp_dst_x);
   }

   emit_src_register(emit, &inst->Src[0]);
   end_emit_instruction(emit);

   emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &inst->Dst[0],
                        &tmp_src_xxxx, inst->Instruction.Saturate);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_SLE (Set Less than or Equal) instruction.
 */
static boolean
emit_sle(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /* dst = SLE(s0, s1):
    *   dst = s0 <= s1 ? 1.0 : 0.0  (per component)
    * Translates into:
    *   GE tmp, s1, s0;           tmp = s1 >= s0 : 0xffffffff : 0 (per comp)
    *   MOVC dst, tmp, 1.0, 0.0;  dst = tmp ? 1.0 : 0.0 (per component)
    */
   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
   struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);
   struct tgsi_full_src_register zero = make_immediate_reg_float(emit, 0.0f);
   struct tgsi_full_src_register one = make_immediate_reg_float(emit, 1.0f);

   /* GE tmp, s1, s0 */
   emit_instruction_op2(emit, VGPU10_OPCODE_GE, &tmp_dst, &inst->Src[1],
                        &inst->Src[0], FALSE);

   /* MOVC dst, tmp, one, zero */
   emit_instruction_op3(emit, VGPU10_OPCODE_MOVC, &inst->Dst[0], &tmp_src,
                        &one, &zero, FALSE);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_SLT (Set Less than) instruction.
 */
static boolean
emit_slt(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /* dst = SLT(s0, s1):
    *   dst = s0 < s1 ? 1.0 : 0.0  (per component)
    * Translates into:
    *   LT tmp, s0, s1;           tmp = s0 < s1 ? 0xffffffff : 0 (per comp)
    *   MOVC dst, tmp, 1.0, 0.0;  dst = tmp ? 1.0 : 0.0 (per component)
    */
   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
   struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);
   struct tgsi_full_src_register zero = make_immediate_reg_float(emit, 0.0f);
   struct tgsi_full_src_register one = make_immediate_reg_float(emit, 1.0f);

   /* LT tmp, s0, s1 */
   emit_instruction_op2(emit, VGPU10_OPCODE_LT, &tmp_dst, &inst->Src[0],
                        &inst->Src[1], FALSE);

   /* MOVC dst, tmp, one, zero */
   emit_instruction_op3(emit, VGPU10_OPCODE_MOVC, &inst->Dst[0], &tmp_src,
                        &one, &zero, FALSE);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_SNE (Set Not Equal) instruction.
 */
static boolean
emit_sne(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /* dst = SNE(s0, s1):
    *   dst = s0 != s1 ? 1.0 : 0.0  (per component)
    * Translates into:
    *   EQ tmp, s0, s1;           tmp = s0 == s1 : 0xffffffff : 0 (per comp)
    *   MOVC dst, tmp, 1.0, 0.0;  dst = tmp ? 1.0 : 0.0 (per component)
    */
   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
   struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);
   struct tgsi_full_src_register zero = make_immediate_reg_float(emit, 0.0f);
   struct tgsi_full_src_register one = make_immediate_reg_float(emit, 1.0f);

   /* NE tmp, s0, s1 */
   emit_instruction_op2(emit, VGPU10_OPCODE_NE, &tmp_dst, &inst->Src[0],
                        &inst->Src[1], FALSE);

   /* MOVC dst, tmp, one, zero */
   emit_instruction_op3(emit, VGPU10_OPCODE_MOVC, &inst->Dst[0], &tmp_src,
                        &one, &zero, FALSE);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_SSG (Set Sign) instruction.
 */
static boolean
emit_ssg(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /* dst.x = (src.x > 0.0) ? 1.0 : (src.x < 0.0) ? -1.0 : 0.0
    * dst.y = (src.y > 0.0) ? 1.0 : (src.y < 0.0) ? -1.0 : 0.0
    * dst.z = (src.z > 0.0) ? 1.0 : (src.z < 0.0) ? -1.0 : 0.0
    * dst.w = (src.w > 0.0) ? 1.0 : (src.w < 0.0) ? -1.0 : 0.0
    * Translates into:
    *   LT tmp1, src, zero;           tmp1 = src < zero ? 0xffffffff : 0 (per comp)
    *   MOVC tmp2, tmp1, -1.0, 0.0;   tmp2 = tmp1 ? -1.0 : 0.0 (per component)
    *   LT tmp1, zero, src;           tmp1 = zero < src ? 0xffffffff : 0 (per comp)
    *   MOVC dst, tmp1, 1.0, tmp2;    dst = tmp1 ? 1.0 : tmp2 (per component)
    */
   struct tgsi_full_src_register zero =
      make_immediate_reg_float(emit, 0.0f);
   struct tgsi_full_src_register one =
      make_immediate_reg_float(emit, 1.0f);
   struct tgsi_full_src_register neg_one =
      make_immediate_reg_float(emit, -1.0f);

   unsigned tmp1 = get_temp_index(emit);
   struct tgsi_full_src_register tmp1_src = make_src_temp_reg(tmp1);
   struct tgsi_full_dst_register tmp1_dst = make_dst_temp_reg(tmp1);

   unsigned tmp2 = get_temp_index(emit);
   struct tgsi_full_src_register tmp2_src = make_src_temp_reg(tmp2);
   struct tgsi_full_dst_register tmp2_dst = make_dst_temp_reg(tmp2);

   emit_instruction_op2(emit, VGPU10_OPCODE_LT, &tmp1_dst, &inst->Src[0],
                        &zero, FALSE);
   emit_instruction_op3(emit, VGPU10_OPCODE_MOVC, &tmp2_dst, &tmp1_src,
                        &neg_one, &zero, FALSE);
   emit_instruction_op2(emit, VGPU10_OPCODE_LT, &tmp1_dst, &zero,
                        &inst->Src[0], FALSE);
   emit_instruction_op3(emit, VGPU10_OPCODE_MOVC, &inst->Dst[0], &tmp1_src,
                        &one, &tmp2_src, FALSE);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_ISSG (Integer Set Sign) instruction.
 */
static boolean
emit_issg(struct svga_shader_emitter_v10 *emit,
          const struct tgsi_full_instruction *inst)
{
   /* dst.x = (src.x > 0) ? 1 : (src.x < 0) ? -1 : 0
    * dst.y = (src.y > 0) ? 1 : (src.y < 0) ? -1 : 0
    * dst.z = (src.z > 0) ? 1 : (src.z < 0) ? -1 : 0
    * dst.w = (src.w > 0) ? 1 : (src.w < 0) ? -1 : 0
    * Translates into:
    *   ILT tmp1, src, 0              tmp1 = src < 0 ? -1 : 0 (per component)
    *   ILT tmp2, 0, src              tmp2 = 0 < src ? -1 : 0 (per component)
    *   IADD dst, tmp1, neg(tmp2)     dst  = tmp1 - tmp2      (per component)
    */
   struct tgsi_full_src_register zero = make_immediate_reg_float(emit, 0.0f);

   unsigned tmp1 = get_temp_index(emit);
   struct tgsi_full_src_register tmp1_src = make_src_temp_reg(tmp1);
   struct tgsi_full_dst_register tmp1_dst = make_dst_temp_reg(tmp1);

   unsigned tmp2 = get_temp_index(emit);
   struct tgsi_full_src_register tmp2_src = make_src_temp_reg(tmp2);
   struct tgsi_full_dst_register tmp2_dst = make_dst_temp_reg(tmp2);

   struct tgsi_full_src_register neg_tmp2 = negate_src(&tmp2_src);

   emit_instruction_op2(emit, VGPU10_OPCODE_ILT, &tmp1_dst,
                        &inst->Src[0], &zero, FALSE);
   emit_instruction_op2(emit, VGPU10_OPCODE_ILT, &tmp2_dst,
                        &zero, &inst->Src[0], FALSE);
   emit_instruction_op2(emit, VGPU10_OPCODE_IADD, &inst->Dst[0],
                        &tmp1_src, &neg_tmp2, FALSE);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_SUB instruction.
 */
static boolean
emit_sub(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /* dst = SUB(s0, s1):
    *   dst = s0 - s1
    * Translates into:
    *   ADD dst, s0, neg(s1)
    */
   struct tgsi_full_src_register neg_src1 = negate_src(&inst->Src[1]);

   /* ADD dst, s0, neg(s1) */
   emit_instruction_op2(emit, VGPU10_OPCODE_ADD, &inst->Dst[0],
                        &inst->Src[0], &neg_src1,
                        inst->Instruction.Saturate);

   return TRUE;
}


/**
 * Emit a comparison instruction.  The dest register will get
 * 0 or ~0 values depending on the outcome of comparing src0 to src1.
 */
static void
emit_comparison(struct svga_shader_emitter_v10 *emit,
                SVGA3dCmpFunc func,
                const struct tgsi_full_dst_register *dst,
                const struct tgsi_full_src_register *src0,
                const struct tgsi_full_src_register *src1)
{
   struct tgsi_full_src_register immediate;
   VGPU10OpcodeToken0 opcode0;
   boolean swapSrc = FALSE;

   /* Sanity checks for svga vs. gallium enums */
   STATIC_ASSERT(SVGA3D_CMP_LESS == (PIPE_FUNC_LESS + 1));
   STATIC_ASSERT(SVGA3D_CMP_GREATEREQUAL == (PIPE_FUNC_GEQUAL + 1));

   opcode0.value = 0;

   switch (func) {
   case SVGA3D_CMP_NEVER:
      immediate = make_immediate_reg_int(emit, 0);
      /* MOV dst, {0} */
      begin_emit_instruction(emit);
      emit_dword(emit, VGPU10_OPCODE_MOV);
      emit_dst_register(emit, dst);
      emit_src_register(emit, &immediate);
      end_emit_instruction(emit);
      return;
   case SVGA3D_CMP_ALWAYS:
      immediate = make_immediate_reg_int(emit, -1);
      /* MOV dst, {-1} */
      begin_emit_instruction(emit);
      emit_dword(emit, VGPU10_OPCODE_MOV);
      emit_dst_register(emit, dst);
      emit_src_register(emit, &immediate);
      end_emit_instruction(emit);
      return;
   case SVGA3D_CMP_LESS:
      opcode0.opcodeType = VGPU10_OPCODE_LT;
      break;
   case SVGA3D_CMP_EQUAL:
      opcode0.opcodeType = VGPU10_OPCODE_EQ;
      break;
   case SVGA3D_CMP_LESSEQUAL:
      opcode0.opcodeType = VGPU10_OPCODE_GE;
      swapSrc = TRUE;
      break;
   case SVGA3D_CMP_GREATER:
      opcode0.opcodeType = VGPU10_OPCODE_LT;
      swapSrc = TRUE;
      break;
   case SVGA3D_CMP_NOTEQUAL:
      opcode0.opcodeType = VGPU10_OPCODE_NE;
      break;
   case SVGA3D_CMP_GREATEREQUAL:
      opcode0.opcodeType = VGPU10_OPCODE_GE;
      break;
   default:
      assert(!"Unexpected comparison mode");
      opcode0.opcodeType = VGPU10_OPCODE_EQ;
   }

   begin_emit_instruction(emit);
   emit_dword(emit, opcode0.value);
   emit_dst_register(emit, dst);
   if (swapSrc) {
      emit_src_register(emit, src1);
      emit_src_register(emit, src0);
   }
   else {
      emit_src_register(emit, src0);
      emit_src_register(emit, src1);
   }
   end_emit_instruction(emit);
}


/**
 * Get texel/address offsets for a texture instruction.
 */
static void
get_texel_offsets(const struct svga_shader_emitter_v10 *emit,
                  const struct tgsi_full_instruction *inst, int offsets[3])
{
   if (inst->Texture.NumOffsets == 1) {
      /* According to OpenGL Shader Language spec the offsets are only
       * fetched from a previously-declared immediate/literal.
       */
      const struct tgsi_texture_offset *off = inst->TexOffsets;
      const unsigned index = off[0].Index;
      const unsigned swizzleX = off[0].SwizzleX;
      const unsigned swizzleY = off[0].SwizzleY;
      const unsigned swizzleZ = off[0].SwizzleZ;
      const union tgsi_immediate_data *imm = emit->immediates[index];

      assert(inst->TexOffsets[0].File == TGSI_FILE_IMMEDIATE);

      offsets[0] = imm[swizzleX].Int;
      offsets[1] = imm[swizzleY].Int;
      offsets[2] = imm[swizzleZ].Int;
   }
   else {
      offsets[0] = offsets[1] = offsets[2] = 0;
   }
}


/**
 * Set up the coordinate register for texture sampling.
 * When we're sampling from a RECT texture we have to scale the
 * unnormalized coordinate to a normalized coordinate.
 * We do that by multiplying the coordinate by an "extra" constant.
 * An alternative would be to use the RESINFO instruction to query the
 * texture's size.
 */
static struct tgsi_full_src_register
setup_texcoord(struct svga_shader_emitter_v10 *emit,
               unsigned unit,
               const struct tgsi_full_src_register *coord)
{
   if (emit->key.tex[unit].unnormalized) {
      unsigned scale_index = emit->texcoord_scale_index[unit];
      unsigned tmp = get_temp_index(emit);
      struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
      struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);
      struct tgsi_full_src_register scale_src = make_src_const_reg(scale_index);

      /* MUL tmp, coord, const[] */
      emit_instruction_op2(emit, VGPU10_OPCODE_MUL, &tmp_dst,
                           coord, &scale_src, FALSE);
      return tmp_src;
   }
   else {
      /* use texcoord as-is */
      return *coord;
   }
}


/**
 * For SAMPLE_C instructions, emit the extra src register which indicates
 * the reference/comparision value.
 */
static void
emit_tex_compare_refcoord(struct svga_shader_emitter_v10 *emit,
                          unsigned target,
                          const struct tgsi_full_src_register *coord)
{
   struct tgsi_full_src_register coord_src_ref;
   unsigned component;

   assert(tgsi_is_shadow_target(target));

   assert(target != TGSI_TEXTURE_SHADOWCUBE_ARRAY); /* XXX not implemented */
   if (target == TGSI_TEXTURE_SHADOW2D_ARRAY ||
       target == TGSI_TEXTURE_SHADOWCUBE)
      component = TGSI_SWIZZLE_W;
   else
      component = TGSI_SWIZZLE_Z;

   coord_src_ref = scalar_src(coord, component);

   emit_src_register(emit, &coord_src_ref);
}


/**
 * Info for implementing texture swizzles.
 * The begin_tex_swizzle(), get_tex_swizzle_dst() and end_tex_swizzle()
 * functions use this to encapsulate the extra steps needed to perform
 * a texture swizzle, or shadow/depth comparisons.
 * The shadow/depth comparison is only done here if for the cases where
 * there's no VGPU10 opcode (like texture bias lookup w/ shadow compare).
 */
struct tex_swizzle_info
{
   boolean swizzled;
   boolean shadow_compare;
   unsigned unit;
   unsigned texture_target;  /**< TGSI_TEXTURE_x */
   struct tgsi_full_src_register tmp_src;
   struct tgsi_full_dst_register tmp_dst;
   const struct tgsi_full_dst_register *inst_dst;
   const struct tgsi_full_src_register *coord_src;
};


/**
 * Do setup for handling texture swizzles or shadow compares.
 * \param unit  the texture unit
 * \param inst  the TGSI texture instruction
 * \param shadow_compare  do shadow/depth comparison?
 * \param swz  returns the swizzle info
 */
static void
begin_tex_swizzle(struct svga_shader_emitter_v10 *emit,
                  unsigned unit,
                  const struct tgsi_full_instruction *inst,
                  boolean shadow_compare,
                  struct tex_swizzle_info *swz)
{
   swz->swizzled = (emit->key.tex[unit].swizzle_r != TGSI_SWIZZLE_X ||
                    emit->key.tex[unit].swizzle_g != TGSI_SWIZZLE_Y ||
                    emit->key.tex[unit].swizzle_b != TGSI_SWIZZLE_Z ||
                    emit->key.tex[unit].swizzle_a != TGSI_SWIZZLE_W);

   swz->shadow_compare = shadow_compare;
   swz->texture_target = inst->Texture.Texture;

   if (swz->swizzled || shadow_compare) {
      /* Allocate temp register for the result of the SAMPLE instruction
       * and the source of the MOV/compare/swizzle instructions.
       */
      unsigned tmp = get_temp_index(emit);
      swz->tmp_src = make_src_temp_reg(tmp);
      swz->tmp_dst = make_dst_temp_reg(tmp);

      swz->unit = unit;
   }
   swz->inst_dst = &inst->Dst[0];
   swz->coord_src = &inst->Src[0];
}


/**
 * Returns the register to put the SAMPLE instruction results into.
 * This will either be the original instruction dst reg (if no swizzle
 * and no shadow comparison) or a temporary reg if there is a swizzle.
 */
static const struct tgsi_full_dst_register *
get_tex_swizzle_dst(const struct tex_swizzle_info *swz)
{
   return (swz->swizzled || swz->shadow_compare)
      ? &swz->tmp_dst : swz->inst_dst;
}


/**
 * This emits the MOV instruction that actually implements a texture swizzle
 * and/or shadow comparison.
 */
static void
end_tex_swizzle(struct svga_shader_emitter_v10 *emit,
                const struct tex_swizzle_info *swz)
{
   if (swz->shadow_compare) {
      /* Emit extra instructions to compare the fetched texel value against
       * a texture coordinate component.  The result of the comparison
       * is 0.0 or 1.0.
       */
      struct tgsi_full_src_register coord_src;
      struct tgsi_full_src_register texel_src =
         scalar_src(&swz->tmp_src, TGSI_SWIZZLE_X);
      struct tgsi_full_src_register one =
         make_immediate_reg_float(emit, 1.0f);
      /* convert gallium comparison func to SVGA comparison func */
      SVGA3dCmpFunc compare_func = emit->key.tex[swz->unit].compare_func + 1;

      assert(emit->unit == PIPE_SHADER_FRAGMENT);

      switch (swz->texture_target) {
      case TGSI_TEXTURE_SHADOW2D:
      case TGSI_TEXTURE_SHADOWRECT:
      case TGSI_TEXTURE_SHADOW1D_ARRAY:
         coord_src = scalar_src(swz->coord_src, TGSI_SWIZZLE_Z);
         break;
      case TGSI_TEXTURE_SHADOW1D:
         coord_src = scalar_src(swz->coord_src, TGSI_SWIZZLE_Y);
         break;
      case TGSI_TEXTURE_SHADOWCUBE:
      case TGSI_TEXTURE_SHADOW2D_ARRAY:
         coord_src = scalar_src(swz->coord_src, TGSI_SWIZZLE_W);
         break;
      default:
         assert(!"Unexpected texture target in end_tex_swizzle()");
         coord_src = scalar_src(swz->coord_src, TGSI_SWIZZLE_Z);
      }

      /* COMPARE tmp, coord, texel */
      /* XXX it would seem that the texel and coord arguments should
       * be transposed here, but piglit tests indicate otherwise.
       */
      emit_comparison(emit, compare_func,
                      &swz->tmp_dst, &texel_src, &coord_src);

      /* AND dest, tmp, {1.0} */
      begin_emit_instruction(emit);
      emit_opcode(emit, VGPU10_OPCODE_AND, FALSE);
      if (swz->swizzled) {
         emit_dst_register(emit, &swz->tmp_dst);
      }
      else {
         emit_dst_register(emit, swz->inst_dst);
      }
      emit_src_register(emit, &swz->tmp_src);
      emit_src_register(emit, &one);
      end_emit_instruction(emit);
   }

   if (swz->swizzled) {
      unsigned swz_r = emit->key.tex[swz->unit].swizzle_r;
      unsigned swz_g = emit->key.tex[swz->unit].swizzle_g;
      unsigned swz_b = emit->key.tex[swz->unit].swizzle_b;
      unsigned swz_a = emit->key.tex[swz->unit].swizzle_a;
      unsigned writemask_0 = 0, writemask_1 = 0;
      boolean int_tex = is_integer_type(emit->key.tex[swz->unit].return_type);

      /* Swizzle w/out zero/one terms */
      struct tgsi_full_src_register src_swizzled =
         swizzle_src(&swz->tmp_src,
                     swz_r < PIPE_SWIZZLE_ZERO ? swz_r : PIPE_SWIZZLE_RED,
                     swz_g < PIPE_SWIZZLE_ZERO ? swz_g : PIPE_SWIZZLE_GREEN,
                     swz_b < PIPE_SWIZZLE_ZERO ? swz_b : PIPE_SWIZZLE_BLUE,
                     swz_a < PIPE_SWIZZLE_ZERO ? swz_a : PIPE_SWIZZLE_ALPHA);

      /* MOV dst, color(tmp).<swizzle> */
      emit_instruction_op1(emit, VGPU10_OPCODE_MOV,
                           swz->inst_dst, &src_swizzled, FALSE);

      /* handle swizzle zero terms */
      writemask_0 = (((swz_r == PIPE_SWIZZLE_ZERO) << 0) |
                     ((swz_g == PIPE_SWIZZLE_ZERO) << 1) |
                     ((swz_b == PIPE_SWIZZLE_ZERO) << 2) |
                     ((swz_a == PIPE_SWIZZLE_ZERO) << 3));

      if (writemask_0) {
         struct tgsi_full_src_register zero = int_tex ?
            make_immediate_reg_int(emit, 0) :
            make_immediate_reg_float(emit, 0.0f);
         struct tgsi_full_dst_register dst =
            writemask_dst(swz->inst_dst, writemask_0);

         /* MOV dst.writemask_0, {0,0,0,0} */
         emit_instruction_op1(emit, VGPU10_OPCODE_MOV,
                              &dst, &zero, FALSE);
      }

      /* handle swizzle one terms */
      writemask_1 = (((swz_r == PIPE_SWIZZLE_ONE) << 0) |
                     ((swz_g == PIPE_SWIZZLE_ONE) << 1) |
                     ((swz_b == PIPE_SWIZZLE_ONE) << 2) |
                     ((swz_a == PIPE_SWIZZLE_ONE) << 3));

      if (writemask_1) {
         struct tgsi_full_src_register one = int_tex ?
            make_immediate_reg_int(emit, 1) :
            make_immediate_reg_float(emit, 1.0f);
         struct tgsi_full_dst_register dst =
            writemask_dst(swz->inst_dst, writemask_1);

         /* MOV dst.writemask_1, {1,1,1,1} */
         emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &dst, &one, FALSE);
      }
   }
}


/**
 * Emit code for TGSI_OPCODE_SAMPLE instruction.
 */
static boolean
emit_sample(struct svga_shader_emitter_v10 *emit,
            const struct tgsi_full_instruction *inst)
{
   const unsigned resource_unit = inst->Src[1].Register.Index;
   const unsigned sampler_unit = inst->Src[2].Register.Index;
   struct tgsi_full_src_register coord;
   int offsets[3];
   struct tex_swizzle_info swz_info;

   begin_tex_swizzle(emit, sampler_unit, inst, FALSE, &swz_info);

   get_texel_offsets(emit, inst, offsets);

   coord = setup_texcoord(emit, resource_unit, &inst->Src[0]);

   /* SAMPLE dst, coord(s0), resource, sampler */
   begin_emit_instruction(emit);

   emit_sample_opcode(emit, VGPU10_OPCODE_SAMPLE,
                      inst->Instruction.Saturate, offsets);
   emit_dst_register(emit, get_tex_swizzle_dst(&swz_info));
   emit_src_register(emit, &coord);
   emit_resource_register(emit, resource_unit);
   emit_sampler_register(emit, sampler_unit);
   end_emit_instruction(emit);

   end_tex_swizzle(emit, &swz_info);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Check if a texture instruction is valid.
 * An example of an invalid texture instruction is doing shadow comparison
 * with an integer-valued texture.
 * If we detect an invalid texture instruction, we replace it with:
 *   MOV dst, {1,1,1,1};
 * \return TRUE if valid, FALSE if invalid.
 */
static boolean
is_valid_tex_instruction(struct svga_shader_emitter_v10 *emit,
                         const struct tgsi_full_instruction *inst)
{
   const unsigned unit = inst->Src[1].Register.Index;
   const unsigned target = inst->Texture.Texture;
   boolean valid = TRUE;

   if (tgsi_is_shadow_target(target) &&
       is_integer_type(emit->key.tex[unit].return_type)) {
      debug_printf("Invalid SAMPLE_C with an integer texture!\n");
      valid = FALSE;
   }
   /* XXX might check for other conditions in the future here */

   if (!valid) {
      /* emit a MOV dst, {1,1,1,1} instruction. */
      struct tgsi_full_src_register one = make_immediate_reg_float(emit, 1.0f);
      begin_emit_instruction(emit);
      emit_opcode(emit, VGPU10_OPCODE_MOV, FALSE);
      emit_dst_register(emit, &inst->Dst[0]);
      emit_src_register(emit, &one);
      end_emit_instruction(emit);
   }

   return valid;
}


/**
 * Emit code for TGSI_OPCODE_TEX (simple texture lookup)
 */
static boolean
emit_tex(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   const uint unit = inst->Src[1].Register.Index;
   unsigned target = inst->Texture.Texture;
   unsigned opcode;
   struct tgsi_full_src_register coord;
   int offsets[3];
   struct tex_swizzle_info swz_info;

   /* check that the sampler returns a float */
   if (!is_valid_tex_instruction(emit, inst))
      return TRUE;

   begin_tex_swizzle(emit, unit, inst, FALSE, &swz_info);

   get_texel_offsets(emit, inst, offsets);

   coord = setup_texcoord(emit, unit, &inst->Src[0]);

   /* SAMPLE dst, coord(s0), resource, sampler */
   begin_emit_instruction(emit);

   if (tgsi_is_shadow_target(target))
      opcode = VGPU10_OPCODE_SAMPLE_C;
   else
      opcode = VGPU10_OPCODE_SAMPLE;

   emit_sample_opcode(emit, opcode, inst->Instruction.Saturate, offsets);
   emit_dst_register(emit, get_tex_swizzle_dst(&swz_info));
   emit_src_register(emit, &coord);
   emit_resource_register(emit, unit);
   emit_sampler_register(emit, unit);
   if (opcode == VGPU10_OPCODE_SAMPLE_C) {
      emit_tex_compare_refcoord(emit, target, &coord);
   }
   end_emit_instruction(emit);

   end_tex_swizzle(emit, &swz_info);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_TXP (projective texture)
 */
static boolean
emit_txp(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   const uint unit = inst->Src[1].Register.Index;
   unsigned target = inst->Texture.Texture;
   unsigned opcode;
   int offsets[3];
   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
   struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);
   struct tgsi_full_src_register src0_wwww =
      scalar_src(&inst->Src[0], TGSI_SWIZZLE_W);
   struct tgsi_full_src_register coord;
   struct tex_swizzle_info swz_info;

   /* check that the sampler returns a float */
   if (!is_valid_tex_instruction(emit, inst))
      return TRUE;

   begin_tex_swizzle(emit, unit, inst, FALSE, &swz_info);

   get_texel_offsets(emit, inst, offsets);

   coord = setup_texcoord(emit, unit, &inst->Src[0]);

   /* DIV tmp, coord, coord.wwww */
   emit_instruction_op2(emit, VGPU10_OPCODE_DIV, &tmp_dst,
                        &coord, &src0_wwww, FALSE);

   /* SAMPLE dst, coord(tmp), resource, sampler */
   begin_emit_instruction(emit);

   if (tgsi_is_shadow_target(target))
      opcode = VGPU10_OPCODE_SAMPLE_C;
   else
      opcode = VGPU10_OPCODE_SAMPLE;

   emit_sample_opcode(emit, opcode, inst->Instruction.Saturate, offsets);
   emit_dst_register(emit, get_tex_swizzle_dst(&swz_info));
   emit_src_register(emit, &tmp_src);  /* projected coord */
   emit_resource_register(emit, unit);
   emit_sampler_register(emit, unit);
   if (opcode == VGPU10_OPCODE_SAMPLE_C) {
      emit_tex_compare_refcoord(emit, target, &tmp_src);
   }
   end_emit_instruction(emit);

   end_tex_swizzle(emit, &swz_info);

   free_temp_indexes(emit);

   return TRUE;
}


/*
 * Emit code for TGSI_OPCODE_XPD instruction.
 */
static boolean
emit_xpd(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   /* dst.x = src0.y * src1.z - src1.y * src0.z
    * dst.y = src0.z * src1.x - src1.z * src0.x
    * dst.z = src0.x * src1.y - src1.x * src0.y
    * dst.w = 1
    */
   struct tgsi_full_src_register s0_xxxx =
      scalar_src(&inst->Src[0], TGSI_SWIZZLE_X);
   struct tgsi_full_src_register s0_yyyy =
      scalar_src(&inst->Src[0], TGSI_SWIZZLE_Y);
   struct tgsi_full_src_register s0_zzzz =
      scalar_src(&inst->Src[0], TGSI_SWIZZLE_Z);

   struct tgsi_full_src_register s1_xxxx =
      scalar_src(&inst->Src[1], TGSI_SWIZZLE_X);
   struct tgsi_full_src_register s1_yyyy =
      scalar_src(&inst->Src[1], TGSI_SWIZZLE_Y);
   struct tgsi_full_src_register s1_zzzz =
      scalar_src(&inst->Src[1], TGSI_SWIZZLE_Z);

   unsigned tmp1 = get_temp_index(emit);
   struct tgsi_full_src_register tmp1_src = make_src_temp_reg(tmp1);
   struct tgsi_full_dst_register tmp1_dst = make_dst_temp_reg(tmp1);

   unsigned tmp2 = get_temp_index(emit);
   struct tgsi_full_src_register tmp2_src = make_src_temp_reg(tmp2);
   struct tgsi_full_dst_register tmp2_dst = make_dst_temp_reg(tmp2);
   struct tgsi_full_src_register neg_tmp2_src = negate_src(&tmp2_src);

   unsigned tmp3 = get_temp_index(emit);
   struct tgsi_full_src_register tmp3_src = make_src_temp_reg(tmp3);
   struct tgsi_full_dst_register tmp3_dst = make_dst_temp_reg(tmp3);
   struct tgsi_full_dst_register tmp3_dst_x =
      writemask_dst(&tmp3_dst, TGSI_WRITEMASK_X);
   struct tgsi_full_dst_register tmp3_dst_y =
      writemask_dst(&tmp3_dst, TGSI_WRITEMASK_Y);
   struct tgsi_full_dst_register tmp3_dst_z =
      writemask_dst(&tmp3_dst, TGSI_WRITEMASK_Z);
   struct tgsi_full_dst_register tmp3_dst_w =
      writemask_dst(&tmp3_dst, TGSI_WRITEMASK_W);

   /* Note: we put all the intermediate computations into tmp3 in case
    * the XPD dest register is that same as one of the src regs (in which
    * case we could clobber a src reg before we're done with it) .
    *
    * Note: we could get by with just one temp register instead of three
    * since we're doing scalar operations and there's enough room in one
    * temp for everything.
    */

   /* MUL tmp1, src0.y, src1.z */
   /* MUL tmp2, src1.y, src0.z */
   /* ADD tmp3.x, tmp1, -tmp2 */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_X) {
      emit_instruction_op2(emit, VGPU10_OPCODE_MUL, &tmp1_dst,
                           &s0_yyyy, &s1_zzzz, FALSE);
      emit_instruction_op2(emit, VGPU10_OPCODE_MUL, &tmp2_dst,
                           &s1_yyyy, &s0_zzzz, FALSE);
      emit_instruction_op2(emit, VGPU10_OPCODE_ADD, &tmp3_dst_x,
                           &tmp1_src, &neg_tmp2_src, FALSE);
   }

   /* MUL tmp1, src0.z, src1.x */
   /* MUL tmp2, src1.z, src0.x */
   /* ADD tmp3.y, tmp1, -tmp2 */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_Y) {
      emit_instruction_op2(emit, VGPU10_OPCODE_MUL, &tmp1_dst, &s0_zzzz,
                           &s1_xxxx, FALSE);
      emit_instruction_op2(emit, VGPU10_OPCODE_MUL, &tmp2_dst, &s1_zzzz,
                           &s0_xxxx, FALSE);
      emit_instruction_op2(emit, VGPU10_OPCODE_ADD, &tmp3_dst_y,
                           &tmp1_src, &neg_tmp2_src, FALSE);
   }

   /* MUL tmp1, src0.x, src1.y */
   /* MUL tmp2, src1.x, src0.y */
   /* ADD tmp3.z, tmp1, -tmp2 */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_Z) {
      emit_instruction_op2(emit, VGPU10_OPCODE_MUL, &tmp1_dst, &s0_xxxx,
                           &s1_yyyy, FALSE);
      emit_instruction_op2(emit, VGPU10_OPCODE_MUL, &tmp2_dst, &s1_xxxx,
                           &s0_yyyy, FALSE);
      emit_instruction_op2(emit, VGPU10_OPCODE_ADD, &tmp3_dst_z,
                           &tmp1_src, &neg_tmp2_src, FALSE);
   }

   /* MOV tmp3.w, 1.0 */
   if (inst->Dst[0].Register.WriteMask & TGSI_WRITEMASK_W) {
      struct tgsi_full_src_register one =
         make_immediate_reg_float(emit, 1.0f);

      emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &tmp3_dst_w, &one, FALSE);
   }

   /* MOV dst, tmp3 */
   emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &inst->Dst[0], &tmp3_src,
                        inst->Instruction.Saturate);


   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_TXD (explicit derivatives)
 */
static boolean
emit_txd(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   const uint unit = inst->Src[3].Register.Index;
   unsigned target = inst->Texture.Texture;
   int offsets[3];
   struct tgsi_full_src_register coord;
   struct tex_swizzle_info swz_info;

   begin_tex_swizzle(emit, unit, inst, tgsi_is_shadow_target(target),
                     &swz_info);

   get_texel_offsets(emit, inst, offsets);

   coord = setup_texcoord(emit, unit, &inst->Src[0]);

   /* SAMPLE_D dst, coord(s0), resource, sampler, Xderiv(s1), Yderiv(s2) */
   begin_emit_instruction(emit);
   emit_sample_opcode(emit, VGPU10_OPCODE_SAMPLE_D,
                      inst->Instruction.Saturate, offsets);
   emit_dst_register(emit, get_tex_swizzle_dst(&swz_info));
   emit_src_register(emit, &coord);
   emit_resource_register(emit, unit);
   emit_sampler_register(emit, unit);
   emit_src_register(emit, &inst->Src[1]);  /* Xderiv */
   emit_src_register(emit, &inst->Src[2]);  /* Yderiv */
   end_emit_instruction(emit);

   end_tex_swizzle(emit, &swz_info);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_TXF (texel fetch)
 */
static boolean
emit_txf(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   const uint unit = inst->Src[1].Register.Index;
   const unsigned msaa = emit->key.tex[unit].texture_msaa;
   int offsets[3];
   struct tex_swizzle_info swz_info;

   begin_tex_swizzle(emit, unit, inst, FALSE, &swz_info);

   get_texel_offsets(emit, inst, offsets);

   if (msaa) {
      /* Fetch one sample from an MSAA texture */
      struct tgsi_full_src_register sampleIndex =
         scalar_src(&inst->Src[0], TGSI_SWIZZLE_W);
      /* LD_MS dst, coord(s0), resource, sampleIndex */
      begin_emit_instruction(emit);
      emit_sample_opcode(emit, VGPU10_OPCODE_LD_MS,
                         inst->Instruction.Saturate, offsets);
      emit_dst_register(emit, get_tex_swizzle_dst(&swz_info));
      emit_src_register(emit, &inst->Src[0]);
      emit_resource_register(emit, unit);
      emit_src_register(emit, &sampleIndex);
      end_emit_instruction(emit);
   }
   else {
      /* Fetch one texel specified by integer coordinate */
      /* LD dst, coord(s0), resource */
      begin_emit_instruction(emit);
      emit_sample_opcode(emit, VGPU10_OPCODE_LD,
                         inst->Instruction.Saturate, offsets);
      emit_dst_register(emit, get_tex_swizzle_dst(&swz_info));
      emit_src_register(emit, &inst->Src[0]);
      emit_resource_register(emit, unit);
      end_emit_instruction(emit);
   }

   end_tex_swizzle(emit, &swz_info);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_TXL (explicit LOD) or TGSI_OPCODE_TXB (LOD bias)
 * or TGSI_OPCODE_TXB2 (for cube shadow maps).
 */
static boolean
emit_txl_txb(struct svga_shader_emitter_v10 *emit,
             const struct tgsi_full_instruction *inst)
{
   unsigned target = inst->Texture.Texture;
   unsigned opcode, unit;
   int offsets[3];
   struct tgsi_full_src_register coord, lod_bias;
   struct tex_swizzle_info swz_info;

   assert(inst->Instruction.Opcode == TGSI_OPCODE_TXL ||
          inst->Instruction.Opcode == TGSI_OPCODE_TXB ||
          inst->Instruction.Opcode == TGSI_OPCODE_TXB2);

   if (inst->Instruction.Opcode == TGSI_OPCODE_TXB2) {
      lod_bias = scalar_src(&inst->Src[1], TGSI_SWIZZLE_X);
      unit = inst->Src[2].Register.Index;
   }
   else {
      lod_bias = scalar_src(&inst->Src[0], TGSI_SWIZZLE_W);
      unit = inst->Src[1].Register.Index;
   }

   begin_tex_swizzle(emit, unit, inst, tgsi_is_shadow_target(target),
                     &swz_info);

   get_texel_offsets(emit, inst, offsets);

   coord = setup_texcoord(emit, unit, &inst->Src[0]);

   /* SAMPLE_L/B dst, coord(s0), resource, sampler, lod(s3) */
   begin_emit_instruction(emit);
   if (inst->Instruction.Opcode == TGSI_OPCODE_TXL) {
      opcode = VGPU10_OPCODE_SAMPLE_L;
   }
   else {
      opcode = VGPU10_OPCODE_SAMPLE_B;
   }
   emit_sample_opcode(emit, opcode, inst->Instruction.Saturate, offsets);
   emit_dst_register(emit, get_tex_swizzle_dst(&swz_info));
   emit_src_register(emit, &coord);
   emit_resource_register(emit, unit);
   emit_sampler_register(emit, unit);
   emit_src_register(emit, &lod_bias);
   end_emit_instruction(emit);

   end_tex_swizzle(emit, &swz_info);

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit code for TGSI_OPCODE_TXQ (texture query) instruction.
 */
static boolean
emit_txq(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   const uint unit = inst->Src[1].Register.Index;

   if (emit->key.tex[unit].texture_target == PIPE_BUFFER) {
      /* RESINFO does not support querying texture buffers, so we instead
       * store texture buffer sizes in shader constants, then copy them to
       * implement TXQ instead of emitting RESINFO.
       * MOV dst, const[texture_buffer_size_index[unit]]
       */
      struct tgsi_full_src_register size_src =
         make_src_const_reg(emit->texture_buffer_size_index[unit]);
      emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &inst->Dst[0], &size_src,
                           FALSE);
   } else {
      /* RESINFO dst, srcMipLevel, resource */
      begin_emit_instruction(emit);
      emit_opcode_resinfo(emit, VGPU10_RESINFO_RETURN_UINT);
      emit_dst_register(emit, &inst->Dst[0]);
      emit_src_register(emit, &inst->Src[0]);
      emit_resource_register(emit, unit);
      end_emit_instruction(emit);
   }

   free_temp_indexes(emit);

   return TRUE;
}


/**
 * Emit a simple instruction (like ADD, MUL, MIN, etc).
 */
static boolean
emit_simple(struct svga_shader_emitter_v10 *emit,
            const struct tgsi_full_instruction *inst)
{
   const unsigned opcode = inst->Instruction.Opcode;
   const struct tgsi_opcode_info *op = tgsi_get_opcode_info(opcode);
   unsigned i;

   begin_emit_instruction(emit);
   emit_opcode(emit, translate_opcode(inst->Instruction.Opcode),
               inst->Instruction.Saturate);
   for (i = 0; i < op->num_dst; i++) {
      emit_dst_register(emit, &inst->Dst[i]);
   }
   for (i = 0; i < op->num_src; i++) {
      emit_src_register(emit, &inst->Src[i]);
   }
   end_emit_instruction(emit);

   return TRUE;
}


/**
 * We only special case the MOV instruction to try to detect constant
 * color writes in the fragment shader.
 */
static boolean
emit_mov(struct svga_shader_emitter_v10 *emit,
         const struct tgsi_full_instruction *inst)
{
   const struct tgsi_full_src_register *src = &inst->Src[0];
   const struct tgsi_full_dst_register *dst = &inst->Dst[0];

   if (emit->unit == PIPE_SHADER_FRAGMENT &&
       dst->Register.File == TGSI_FILE_OUTPUT &&
       dst->Register.Index == 0 &&
       src->Register.File == TGSI_FILE_CONSTANT &&
       !src->Register.Indirect) {
      emit->constant_color_output = TRUE;
   }

   return emit_simple(emit, inst);
}


/**
 * Emit a simple VGPU10 instruction which writes to multiple dest registers,
 * where TGSI only uses one dest register.
 */
static boolean
emit_simple_1dst(struct svga_shader_emitter_v10 *emit,
                 const struct tgsi_full_instruction *inst,
                 unsigned dst_count,
                 unsigned dst_index)
{
   const unsigned opcode = inst->Instruction.Opcode;
   const struct tgsi_opcode_info *op = tgsi_get_opcode_info(opcode);
   unsigned i;

   begin_emit_instruction(emit);
   emit_opcode(emit, translate_opcode(inst->Instruction.Opcode),
               inst->Instruction.Saturate);

   for (i = 0; i < dst_count; i++) {
      if (i == dst_index) {
         emit_dst_register(emit, &inst->Dst[0]);
      } else {
         emit_null_dst_register(emit);
      }
   }

   for (i = 0; i < op->num_src; i++) {
      emit_src_register(emit, &inst->Src[i]);
   }
   end_emit_instruction(emit);

   return TRUE;
}


/**
 * Translate a single TGSI instruction to VGPU10.
 */
static boolean
emit_vgpu10_instruction(struct svga_shader_emitter_v10 *emit,
                        unsigned inst_number,
                        const struct tgsi_full_instruction *inst)
{
   const unsigned opcode = inst->Instruction.Opcode;

   switch (opcode) {
   case TGSI_OPCODE_ADD:
   case TGSI_OPCODE_AND:
   case TGSI_OPCODE_BGNLOOP:
   case TGSI_OPCODE_BRK:
   case TGSI_OPCODE_CEIL:
   case TGSI_OPCODE_CONT:
   case TGSI_OPCODE_DDX:
   case TGSI_OPCODE_DDY:
   case TGSI_OPCODE_DIV:
   case TGSI_OPCODE_DP2:
   case TGSI_OPCODE_DP3:
   case TGSI_OPCODE_DP4:
   case TGSI_OPCODE_ELSE:
   case TGSI_OPCODE_ENDIF:
   case TGSI_OPCODE_ENDLOOP:
   case TGSI_OPCODE_ENDSUB:
   case TGSI_OPCODE_F2I:
   case TGSI_OPCODE_F2U:
   case TGSI_OPCODE_FLR:
   case TGSI_OPCODE_FRC:
   case TGSI_OPCODE_FSEQ:
   case TGSI_OPCODE_FSGE:
   case TGSI_OPCODE_FSLT:
   case TGSI_OPCODE_FSNE:
   case TGSI_OPCODE_I2F:
   case TGSI_OPCODE_IMAX:
   case TGSI_OPCODE_IMIN:
   case TGSI_OPCODE_INEG:
   case TGSI_OPCODE_ISGE:
   case TGSI_OPCODE_ISHR:
   case TGSI_OPCODE_ISLT:
   case TGSI_OPCODE_MAD:
   case TGSI_OPCODE_MAX:
   case TGSI_OPCODE_MIN:
   case TGSI_OPCODE_MUL:
   case TGSI_OPCODE_NOP:
   case TGSI_OPCODE_NOT:
   case TGSI_OPCODE_OR:
   case TGSI_OPCODE_RET:
   case TGSI_OPCODE_UADD:
   case TGSI_OPCODE_USEQ:
   case TGSI_OPCODE_USGE:
   case TGSI_OPCODE_USLT:
   case TGSI_OPCODE_UMIN:
   case TGSI_OPCODE_UMAD:
   case TGSI_OPCODE_UMAX:
   case TGSI_OPCODE_ROUND:
   case TGSI_OPCODE_SQRT:
   case TGSI_OPCODE_SHL:
   case TGSI_OPCODE_TRUNC:
   case TGSI_OPCODE_U2F:
   case TGSI_OPCODE_UCMP:
   case TGSI_OPCODE_USHR:
   case TGSI_OPCODE_USNE:
   case TGSI_OPCODE_XOR:
      /* simple instructions */
      return emit_simple(emit, inst);

   case TGSI_OPCODE_MOV:
      return emit_mov(emit, inst);
   case TGSI_OPCODE_EMIT:
      return emit_vertex(emit, inst);
   case TGSI_OPCODE_ENDPRIM:
      return emit_endprim(emit, inst);
   case TGSI_OPCODE_ABS:
      return emit_abs(emit, inst);
   case TGSI_OPCODE_IABS:
      return emit_iabs(emit, inst);
   case TGSI_OPCODE_ARL:
      /* fall-through */
   case TGSI_OPCODE_UARL:
      return emit_arl_uarl(emit, inst);
   case TGSI_OPCODE_BGNSUB:
      /* no-op */
      return TRUE;
   case TGSI_OPCODE_CAL:
      return emit_cal(emit, inst);
   case TGSI_OPCODE_CMP:
      return emit_cmp(emit, inst);
   case TGSI_OPCODE_COS:
      return emit_sincos(emit, inst);
   case TGSI_OPCODE_DP2A:
      return emit_dp2a(emit, inst);
   case TGSI_OPCODE_DPH:
      return emit_dph(emit, inst);
   case TGSI_OPCODE_DST:
      return emit_dst(emit, inst);
   case TGSI_OPCODE_EX2:
      return emit_ex2(emit, inst);
   case TGSI_OPCODE_EXP:
      return emit_exp(emit, inst);
   case TGSI_OPCODE_IF:
      return emit_if(emit, inst);
   case TGSI_OPCODE_KILL:
      return emit_kill(emit, inst);
   case TGSI_OPCODE_KILL_IF:
      return emit_kill_if(emit, inst);
   case TGSI_OPCODE_LG2:
      return emit_lg2(emit, inst);
   case TGSI_OPCODE_LIT:
      return emit_lit(emit, inst);
   case TGSI_OPCODE_LOG:
      return emit_log(emit, inst);
   case TGSI_OPCODE_LRP:
      return emit_lrp(emit, inst);
   case TGSI_OPCODE_POW:
      return emit_pow(emit, inst);
   case TGSI_OPCODE_RCP:
      return emit_rcp(emit, inst);
   case TGSI_OPCODE_RSQ:
      return emit_rsq(emit, inst);
   case TGSI_OPCODE_SAMPLE:
      return emit_sample(emit, inst);
   case TGSI_OPCODE_SCS:
      return emit_scs(emit, inst);
   case TGSI_OPCODE_SEQ:
      return emit_seq(emit, inst);
   case TGSI_OPCODE_SGE:
      return emit_sge(emit, inst);
   case TGSI_OPCODE_SGT:
      return emit_sgt(emit, inst);
   case TGSI_OPCODE_SIN:
      return emit_sincos(emit, inst);
   case TGSI_OPCODE_SLE:
      return emit_sle(emit, inst);
   case TGSI_OPCODE_SLT:
      return emit_slt(emit, inst);
   case TGSI_OPCODE_SNE:
      return emit_sne(emit, inst);
   case TGSI_OPCODE_SSG:
      return emit_ssg(emit, inst);
   case TGSI_OPCODE_ISSG:
      return emit_issg(emit, inst);
   case TGSI_OPCODE_SUB:
      return emit_sub(emit, inst);
   case TGSI_OPCODE_TEX:
      return emit_tex(emit, inst);
   case TGSI_OPCODE_TXP:
      return emit_txp(emit, inst);
   case TGSI_OPCODE_TXB:
   case TGSI_OPCODE_TXB2:
   case TGSI_OPCODE_TXL:
      return emit_txl_txb(emit, inst);
   case TGSI_OPCODE_TXD:
      return emit_txd(emit, inst);
   case TGSI_OPCODE_TXF:
      return emit_txf(emit, inst);
   case TGSI_OPCODE_TXQ:
      return emit_txq(emit, inst);
   case TGSI_OPCODE_UIF:
      return emit_if(emit, inst);
   case TGSI_OPCODE_XPD:
      return emit_xpd(emit, inst);
   case TGSI_OPCODE_UMUL_HI:
   case TGSI_OPCODE_IMUL_HI:
   case TGSI_OPCODE_UDIV:
   case TGSI_OPCODE_IDIV:
      /* These cases use only the FIRST of two destination registers */
      return emit_simple_1dst(emit, inst, 2, 0);
   case TGSI_OPCODE_UMUL:
   case TGSI_OPCODE_UMOD:
   case TGSI_OPCODE_MOD:
      /* These cases use only the SECOND of two destination registers */
      return emit_simple_1dst(emit, inst, 2, 1);
   case TGSI_OPCODE_END:
      if (!emit_post_helpers(emit))
         return FALSE;
      return emit_simple(emit, inst);

   default:
      debug_printf("Unimplemented tgsi instruction %s\n",
                   tgsi_get_opcode_name(opcode));
      return FALSE;
   }

   return TRUE;
}


/**
 * Emit the extra instructions to adjust the vertex position.
 * There are two possible adjustments:
 * 1. Converting from Gallium to VGPU10 coordinate space by applying the
 *    "prescale" and "pretranslate" values.
 * 2. Undoing the viewport transformation when we use the swtnl/draw path.
 * \param vs_pos_tmp_index  which temporary register contains the vertex pos.
 */
static void
emit_vpos_instructions(struct svga_shader_emitter_v10 *emit,
                       unsigned vs_pos_tmp_index)
{
   struct tgsi_full_src_register tmp_pos_src;
   struct tgsi_full_dst_register pos_dst;

   /* Don't bother to emit any extra vertex instructions if vertex position is
    * not written out
    */
   if (emit->vposition.out_index == INVALID_INDEX)
      return;

   tmp_pos_src = make_src_temp_reg(vs_pos_tmp_index);
   pos_dst = make_dst_output_reg(emit->vposition.out_index);

   /* If non-adjusted vertex position register index
    * is valid, copy the vertex position from the temporary
    * vertex position register before it is modified by the
    * prescale computation.
    */
   if (emit->vposition.so_index != INVALID_INDEX) {
      struct tgsi_full_dst_register pos_so_dst =
         make_dst_output_reg(emit->vposition.so_index);

      /* MOV pos_so, tmp_pos */
      emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &pos_so_dst,
                           &tmp_pos_src, FALSE);
   }

   if (emit->vposition.need_prescale) {
      /* This code adjusts the vertex position to match the VGPU10 convention.
       * If p is the position computed by the shader (usually by applying the
       * modelview and projection matrices), the new position q is computed by:
       *
       * q.x = p.w * trans.x + p.x * scale.x
       * q.y = p.w * trans.y + p.y * scale.y
       * q.z = p.w * trans.z + p.z * scale.z;
       * q.w = p.w * trans.w + p.w;
       */
      struct tgsi_full_src_register tmp_pos_src_w =
         scalar_src(&tmp_pos_src, TGSI_SWIZZLE_W);
      struct tgsi_full_dst_register tmp_pos_dst =
         make_dst_temp_reg(vs_pos_tmp_index);
      struct tgsi_full_dst_register tmp_pos_dst_xyz =
         writemask_dst(&tmp_pos_dst, TGSI_WRITEMASK_XYZ);

      struct tgsi_full_src_register prescale_scale =
         make_src_const_reg(emit->vposition.prescale_scale_index);
      struct tgsi_full_src_register prescale_trans =
         make_src_const_reg(emit->vposition.prescale_trans_index);

      /* MUL tmp_pos.xyz, tmp_pos, prescale.scale */
      emit_instruction_op2(emit, VGPU10_OPCODE_MUL, &tmp_pos_dst_xyz,
                           &tmp_pos_src, &prescale_scale, FALSE);

      /* MAD pos, tmp_pos.wwww, prescale.trans, tmp_pos */
      emit_instruction_op3(emit, VGPU10_OPCODE_MAD, &pos_dst, &tmp_pos_src_w,
                           &prescale_trans, &tmp_pos_src, FALSE);
   }
   else if (emit->key.vs.undo_viewport) {
      /* This code computes the final vertex position from the temporary
       * vertex position by undoing the viewport transformation and the
       * divide-by-W operation (we convert window coords back to clip coords).
       * This is needed when we use the 'draw' module for fallbacks.
       * If p is the temp pos in window coords, then the NDC coord q is:
       *   q.x = (p.x - vp.x_trans) / vp.x_scale * p.w
       *   q.y = (p.y - vp.y_trans) / vp.y_scale * p.w
       *   q.z = p.z * p.w
       *   q.w = p.w
       * CONST[vs_viewport_index] contains:
       *   { 1/vp.x_scale, 1/vp.y_scale, -vp.x_trans, -vp.y_trans }
       */
      struct tgsi_full_dst_register tmp_pos_dst =
         make_dst_temp_reg(vs_pos_tmp_index);
      struct tgsi_full_dst_register tmp_pos_dst_xy =
         writemask_dst(&tmp_pos_dst, TGSI_WRITEMASK_XY);
      struct tgsi_full_src_register tmp_pos_src_wwww =
         scalar_src(&tmp_pos_src, TGSI_SWIZZLE_W);

      struct tgsi_full_dst_register pos_dst_xyz =
         writemask_dst(&pos_dst, TGSI_WRITEMASK_XYZ);
      struct tgsi_full_dst_register pos_dst_w =
         writemask_dst(&pos_dst, TGSI_WRITEMASK_W);

      struct tgsi_full_src_register vp_xyzw =
         make_src_const_reg(emit->vs.viewport_index);
      struct tgsi_full_src_register vp_zwww =
         swizzle_src(&vp_xyzw, TGSI_SWIZZLE_Z, TGSI_SWIZZLE_W,
                     TGSI_SWIZZLE_W, TGSI_SWIZZLE_W);

      /* ADD tmp_pos.xy, tmp_pos.xy, viewport.zwww */
      emit_instruction_op2(emit, VGPU10_OPCODE_ADD, &tmp_pos_dst_xy,
                           &tmp_pos_src, &vp_zwww, FALSE);

      /* MUL tmp_pos.xy, tmp_pos.xyzw, viewport.xyzy */
      emit_instruction_op2(emit, VGPU10_OPCODE_MUL, &tmp_pos_dst_xy,
                           &tmp_pos_src, &vp_xyzw, FALSE);

      /* MUL pos.xyz, tmp_pos.xyz, tmp_pos.www */
      emit_instruction_op2(emit, VGPU10_OPCODE_MUL, &pos_dst_xyz,
                           &tmp_pos_src, &tmp_pos_src_wwww, FALSE);

      /* MOV pos.w, tmp_pos.w */
      emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &pos_dst_w,
                           &tmp_pos_src, FALSE);
   }
   else if (vs_pos_tmp_index != INVALID_INDEX) {
      /* This code is to handle the case where the temporary vertex
       * position register is created when the vertex shader has stream
       * output and prescale is disabled because rasterization is to be
       * discarded.
       */
      struct tgsi_full_dst_register pos_dst =
         make_dst_output_reg(emit->vposition.out_index);

      /* MOV pos, tmp_pos */
      begin_emit_instruction(emit);
      emit_opcode(emit, VGPU10_OPCODE_MOV, FALSE);
      emit_dst_register(emit, &pos_dst);
      emit_src_register(emit, &tmp_pos_src);
      end_emit_instruction(emit);
   }
}

static void
emit_clipping_instructions(struct svga_shader_emitter_v10 *emit)
{
   if (emit->clip_mode == CLIP_DISTANCE) {
      /* Copy from copy distance temporary to CLIPDIST & the shadow copy */
      emit_clip_distance_instructions(emit);

   } else if (emit->clip_mode == CLIP_VERTEX) {
      /* Convert TGSI CLIPVERTEX to CLIPDIST */
      emit_clip_vertex_instructions(emit);
   }

   /**
    * Emit vertex position and take care of legacy user planes only if
    * there is a valid vertex position register index.
    * This is to take care of the case
    * where the shader doesn't output vertex position. Then in
    * this case, don't bother to emit more vertex instructions.
    */
   if (emit->vposition.out_index == INVALID_INDEX)
      return;

   /**
    * Emit per-vertex clipping instructions for legacy user defined clip planes.
    * NOTE: we must emit the clip distance instructions before the
    * emit_vpos_instructions() call since the later function will change
    * the TEMP[vs_pos_tmp_index] value.
    */
   if (emit->clip_mode == CLIP_LEGACY) {
      /* Emit CLIPDIST for legacy user defined clip planes */
      emit_clip_distance_from_vpos(emit, emit->vposition.tmp_index);
   }
}


/**
 * Emit extra per-vertex instructions.  This includes clip-coordinate
 * space conversion and computing clip distances.  This is called for
 * each GS emit-vertex instruction and at the end of VS translation.
 */
static void
emit_vertex_instructions(struct svga_shader_emitter_v10 *emit)
{
   const unsigned vs_pos_tmp_index = emit->vposition.tmp_index;

   /* Emit clipping instructions based on clipping mode */
   emit_clipping_instructions(emit);

   /**
    * Reset the temporary vertex position register index
    * so that emit_dst_register() will use the real vertex position output
    */
   emit->vposition.tmp_index = INVALID_INDEX;

   /* Emit vertex position instructions */
   emit_vpos_instructions(emit, vs_pos_tmp_index);

   /* Restore original vposition.tmp_index value for the next GS vertex.
    * It doesn't matter for VS.
    */
   emit->vposition.tmp_index = vs_pos_tmp_index;
}

/**
 * Translate the TGSI_OPCODE_EMIT GS instruction.
 */
static boolean
emit_vertex(struct svga_shader_emitter_v10 *emit,
            const struct tgsi_full_instruction *inst)
{
   unsigned ret = TRUE;

   assert(emit->unit == PIPE_SHADER_GEOMETRY);

   emit_vertex_instructions(emit);

   /* We can't use emit_simple() because the TGSI instruction has one
    * operand (vertex stream number) which we must ignore for VGPU10.
    */
   begin_emit_instruction(emit);
   emit_opcode(emit, VGPU10_OPCODE_EMIT, FALSE);
   end_emit_instruction(emit);

   return ret;
}


/**
 * Emit the extra code to convert from VGPU10's boolean front-face
 * register to TGSI's signed front-face register.
 *
 * TODO: Make temporary front-face register a scalar.
 */
static void
emit_frontface_instructions(struct svga_shader_emitter_v10 *emit)
{
   assert(emit->unit == PIPE_SHADER_FRAGMENT);

   if (emit->fs.face_input_index != INVALID_INDEX) {
      /* convert vgpu10 boolean face register to gallium +/-1 value */
      struct tgsi_full_dst_register tmp_dst =
         make_dst_temp_reg(emit->fs.face_tmp_index);
      struct tgsi_full_src_register one =
         make_immediate_reg_float(emit, 1.0f);
      struct tgsi_full_src_register neg_one =
         make_immediate_reg_float(emit, -1.0f);

      /* MOVC face_tmp, IS_FRONT_FACE.x, 1.0, -1.0 */
      begin_emit_instruction(emit);
      emit_opcode(emit, VGPU10_OPCODE_MOVC, FALSE);
      emit_dst_register(emit, &tmp_dst);
      emit_face_register(emit);
      emit_src_register(emit, &one);
      emit_src_register(emit, &neg_one);
      end_emit_instruction(emit);
   }
}


/**
 * Emit the extra code to convert from VGPU10's fragcoord.w value to 1/w.
 */
static void
emit_fragcoord_instructions(struct svga_shader_emitter_v10 *emit)
{
   assert(emit->unit == PIPE_SHADER_FRAGMENT);

   if (emit->fs.fragcoord_input_index != INVALID_INDEX) {
      struct tgsi_full_dst_register tmp_dst =
         make_dst_temp_reg(emit->fs.fragcoord_tmp_index);
      struct tgsi_full_dst_register tmp_dst_xyz =
         writemask_dst(&tmp_dst, TGSI_WRITEMASK_XYZ);
      struct tgsi_full_dst_register tmp_dst_w =
         writemask_dst(&tmp_dst, TGSI_WRITEMASK_W);
      struct tgsi_full_src_register one =
         make_immediate_reg_float(emit, 1.0f);
      struct tgsi_full_src_register fragcoord =
         make_src_reg(TGSI_FILE_INPUT, emit->fs.fragcoord_input_index);

      /* save the input index */
      unsigned fragcoord_input_index = emit->fs.fragcoord_input_index;
      /* set to invalid to prevent substitution in emit_src_register() */
      emit->fs.fragcoord_input_index = INVALID_INDEX;

      /* MOV fragcoord_tmp.xyz, fragcoord.xyz */
      begin_emit_instruction(emit);
      emit_opcode(emit, VGPU10_OPCODE_MOV, FALSE);
      emit_dst_register(emit, &tmp_dst_xyz);
      emit_src_register(emit, &fragcoord);
      end_emit_instruction(emit);

      /* DIV fragcoord_tmp.w, 1.0, fragcoord.w */
      begin_emit_instruction(emit);
      emit_opcode(emit, VGPU10_OPCODE_DIV, FALSE);
      emit_dst_register(emit, &tmp_dst_w);
      emit_src_register(emit, &one);
      emit_src_register(emit, &fragcoord);
      end_emit_instruction(emit);

      /* restore saved value */
      emit->fs.fragcoord_input_index = fragcoord_input_index;
   }
}


/**
 * Emit extra instructions to adjust VS inputs/attributes.  This can
 * mean casting a vertex attribute from int to float or setting the
 * W component to 1, or both.
 */
static void
emit_vertex_attrib_instructions(struct svga_shader_emitter_v10 *emit)
{
   const unsigned save_w_1_mask = emit->key.vs.adjust_attrib_w_1;
   const unsigned save_itof_mask = emit->key.vs.adjust_attrib_itof;
   const unsigned save_utof_mask = emit->key.vs.adjust_attrib_utof;
   const unsigned save_is_bgra_mask = emit->key.vs.attrib_is_bgra;
   const unsigned save_puint_to_snorm_mask = emit->key.vs.attrib_puint_to_snorm;
   const unsigned save_puint_to_uscaled_mask = emit->key.vs.attrib_puint_to_uscaled;
   const unsigned save_puint_to_sscaled_mask = emit->key.vs.attrib_puint_to_sscaled;

   unsigned adjust_mask = (save_w_1_mask |
                           save_itof_mask |
                           save_utof_mask |
                           save_is_bgra_mask |
                           save_puint_to_snorm_mask |
                           save_puint_to_uscaled_mask |
                           save_puint_to_sscaled_mask);

   assert(emit->unit == PIPE_SHADER_VERTEX);

   if (adjust_mask) {
      struct tgsi_full_src_register one =
         make_immediate_reg_float(emit, 1.0f);

      struct tgsi_full_src_register one_int =
         make_immediate_reg_int(emit, 1);

      /* We need to turn off these bitmasks while emitting the
       * instructions below, then restore them afterward.
       */
      emit->key.vs.adjust_attrib_w_1 = 0;
      emit->key.vs.adjust_attrib_itof = 0;
      emit->key.vs.adjust_attrib_utof = 0;
      emit->key.vs.attrib_is_bgra = 0;
      emit->key.vs.attrib_puint_to_snorm = 0;
      emit->key.vs.attrib_puint_to_uscaled = 0;
      emit->key.vs.attrib_puint_to_sscaled = 0;

      while (adjust_mask) {
         unsigned index = u_bit_scan(&adjust_mask);

         /* skip the instruction if this vertex attribute is not being used */
         if (emit->info.input_usage_mask[index] == 0)
            continue;

         unsigned tmp = emit->vs.adjusted_input[index];
         struct tgsi_full_src_register input_src =
            make_src_reg(TGSI_FILE_INPUT, index);

         struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);
         struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
         struct tgsi_full_dst_register tmp_dst_w =
            writemask_dst(&tmp_dst, TGSI_WRITEMASK_W);

         /* ITOF/UTOF/MOV tmp, input[index] */
         if (save_itof_mask & (1 << index)) {
            emit_instruction_op1(emit, VGPU10_OPCODE_ITOF,
                                 &tmp_dst, &input_src, FALSE);
         }
         else if (save_utof_mask & (1 << index)) {
            emit_instruction_op1(emit, VGPU10_OPCODE_UTOF,
                                 &tmp_dst, &input_src, FALSE);
         }
         else if (save_puint_to_snorm_mask & (1 << index)) {
            emit_puint_to_snorm(emit, &tmp_dst, &input_src);
         }
         else if (save_puint_to_uscaled_mask & (1 << index)) {
            emit_puint_to_uscaled(emit, &tmp_dst, &input_src);
         }
         else if (save_puint_to_sscaled_mask & (1 << index)) {
            emit_puint_to_sscaled(emit, &tmp_dst, &input_src);
         }
         else {
            assert((save_w_1_mask | save_is_bgra_mask) & (1 << index));
            emit_instruction_op1(emit, VGPU10_OPCODE_MOV,
                                 &tmp_dst, &input_src, FALSE);
         }

         if (save_is_bgra_mask & (1 << index)) {
            emit_swap_r_b(emit, &tmp_dst, &tmp_src);
         }

         if (save_w_1_mask & (1 << index)) {
            /* MOV tmp.w, 1.0 */
            if (emit->key.vs.attrib_is_pure_int & (1 << index)) {
               emit_instruction_op1(emit, VGPU10_OPCODE_MOV,
                                    &tmp_dst_w, &one_int, FALSE);
            }
            else {
               emit_instruction_op1(emit, VGPU10_OPCODE_MOV,
                                    &tmp_dst_w, &one, FALSE);
            }
         }
      }

      emit->key.vs.adjust_attrib_w_1 = save_w_1_mask;
      emit->key.vs.adjust_attrib_itof = save_itof_mask;
      emit->key.vs.adjust_attrib_utof = save_utof_mask;
      emit->key.vs.attrib_is_bgra = save_is_bgra_mask;
      emit->key.vs.attrib_puint_to_snorm = save_puint_to_snorm_mask;
      emit->key.vs.attrib_puint_to_uscaled = save_puint_to_uscaled_mask;
      emit->key.vs.attrib_puint_to_sscaled = save_puint_to_sscaled_mask;
   }
}


/**
 * Some common values like 0.0, 1.0, 0.5, etc. are frequently needed
 * to implement some instructions.  We pre-allocate those values here
 * in the immediate constant buffer.
 */
static void
alloc_common_immediates(struct svga_shader_emitter_v10 *emit)
{
   unsigned n = 0;

   emit->common_immediate_pos[n++] =
      alloc_immediate_float4(emit, 0.0f, 1.0f, 0.5f, -1.0f);

   emit->common_immediate_pos[n++] =
      alloc_immediate_float4(emit, 128.0f, -128.0f, 2.0f, 3.0f);

   emit->common_immediate_pos[n++] =
      alloc_immediate_int4(emit, 0, 1, 0, -1);

   if (emit->key.vs.attrib_puint_to_snorm) {
      emit->common_immediate_pos[n++] =
         alloc_immediate_float4(emit, -2.0f, -2.0f, -2.0f, -1.66666f);
   }

   if (emit->key.vs.attrib_puint_to_uscaled) {
      emit->common_immediate_pos[n++] =
         alloc_immediate_float4(emit, 1023.0f, 3.0f, 0.0f, 0.0f);
   }

   if (emit->key.vs.attrib_puint_to_sscaled) {
      emit->common_immediate_pos[n++] =
         alloc_immediate_int4(emit, 22, 12, 2, 0);

      emit->common_immediate_pos[n++] =
         alloc_immediate_int4(emit, 22, 30, 0, 0);
   }

   assert(n <= Elements(emit->common_immediate_pos));
   emit->num_common_immediates = n;
}


/**
 * Emit any extra/helper declarations/code that we might need between
 * the declaration section and code section.
 */
static boolean
emit_pre_helpers(struct svga_shader_emitter_v10 *emit)
{
   /* Properties */
   if (emit->unit == PIPE_SHADER_GEOMETRY)
      emit_property_instructions(emit);

   /* Declare inputs */
   if (!emit_input_declarations(emit))
      return FALSE;

   /* Declare outputs */
   if (!emit_output_declarations(emit))
      return FALSE;

   /* Declare temporary registers */
   emit_temporaries_declaration(emit);

   /* Declare constant registers */
   emit_constant_declaration(emit);

   /* Declare samplers and resources */
   emit_sampler_declarations(emit);
   emit_resource_declarations(emit);

   /* Declare clip distance output registers */
   if (emit->unit == PIPE_SHADER_VERTEX ||
       emit->unit == PIPE_SHADER_GEOMETRY) {
      emit_clip_distance_declarations(emit);
   }

   alloc_common_immediates(emit);

   if (emit->unit == PIPE_SHADER_FRAGMENT &&
       emit->key.fs.alpha_func != SVGA3D_CMP_ALWAYS) {
      float alpha = emit->key.fs.alpha_ref;
      emit->fs.alpha_ref_index =
         alloc_immediate_float4(emit, alpha, alpha, alpha, alpha);
   }

   /* Now, emit the constant block containing all the immediates
    * declared by shader, as well as the extra ones seen above.
    */
   emit_vgpu10_immediates_block(emit);

   if (emit->unit == PIPE_SHADER_FRAGMENT) {
      emit_frontface_instructions(emit);
      emit_fragcoord_instructions(emit);
   }
   else if (emit->unit == PIPE_SHADER_VERTEX) {
      emit_vertex_attrib_instructions(emit);
   }

   return TRUE;
}


/**
 * Emit alpha test code.  This compares TEMP[fs_color_tmp_index].w
 * against the alpha reference value and discards the fragment if the
 * comparison fails.
 */
static void
emit_alpha_test_instructions(struct svga_shader_emitter_v10 *emit,
                             unsigned fs_color_tmp_index)
{
   /* compare output color's alpha to alpha ref and kill */
   unsigned tmp = get_temp_index(emit);
   struct tgsi_full_src_register tmp_src = make_src_temp_reg(tmp);
   struct tgsi_full_src_register tmp_src_x =
      scalar_src(&tmp_src, TGSI_SWIZZLE_X);
   struct tgsi_full_dst_register tmp_dst = make_dst_temp_reg(tmp);
   struct tgsi_full_src_register color_src =
      make_src_temp_reg(fs_color_tmp_index);
   struct tgsi_full_src_register color_src_w =
      scalar_src(&color_src, TGSI_SWIZZLE_W);
   struct tgsi_full_src_register ref_src =
      make_src_immediate_reg(emit->fs.alpha_ref_index);
   struct tgsi_full_dst_register color_dst =
      make_dst_output_reg(emit->fs.color_out_index[0]);

   assert(emit->unit == PIPE_SHADER_FRAGMENT);

   /* dst = src0 'alpha_func' src1 */
   emit_comparison(emit, emit->key.fs.alpha_func, &tmp_dst,
                   &color_src_w, &ref_src);

   /* DISCARD if dst.x == 0 */
   begin_emit_instruction(emit);
   emit_discard_opcode(emit, FALSE);  /* discard if src0.x is zero */
   emit_src_register(emit, &tmp_src_x);
   end_emit_instruction(emit);

   /* If we don't need to broadcast the color below or set fragments to
    * white, emit final color here.
    */
   if (emit->key.fs.write_color0_to_n_cbufs <= 1 &&
       !emit->key.fs.white_fragments) {
      /* MOV output.color, tempcolor */
      emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &color_dst,
                           &color_src, FALSE);     /* XXX saturate? */
   }

   free_temp_indexes(emit);
}


/**
 * When we need to emit white for all fragments (for emulating XOR logicop
 * mode), this function copies white into the temporary color output register.
 */
static void
emit_set_color_white(struct svga_shader_emitter_v10 *emit,
                     unsigned fs_color_tmp_index)
{
   struct tgsi_full_dst_register color_dst =
      make_dst_temp_reg(fs_color_tmp_index);
   struct tgsi_full_src_register white =
      make_immediate_reg_float(emit, 1.0f);

   emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &color_dst, &white, FALSE);
}


/**
 * Emit instructions for writing a single color output to multiple
 * color buffers.
 * This is used when the TGSI_PROPERTY_FS_COLOR0_WRITES_ALL_CBUFS (or
 * when key.fs.white_fragments is true).
 * property is set and the number of render targets is greater than one.
 * \param fs_color_tmp_index  index of the temp register that holds the
 *                            color to broadcast.
 */
static void
emit_broadcast_color_instructions(struct svga_shader_emitter_v10 *emit,
                                 unsigned fs_color_tmp_index)
{
   const unsigned n = emit->key.fs.write_color0_to_n_cbufs;
   unsigned i;
   struct tgsi_full_src_register color_src =
      make_src_temp_reg(fs_color_tmp_index);

   assert(emit->unit == PIPE_SHADER_FRAGMENT);

   for (i = 0; i < n; i++) {
      unsigned output_reg = emit->fs.color_out_index[i];
      struct tgsi_full_dst_register color_dst =
         make_dst_output_reg(output_reg);

      /* Fill in this semantic here since we'll use it later in
       * emit_dst_register().
       */
      emit->info.output_semantic_name[output_reg] = TGSI_SEMANTIC_COLOR;

      /* MOV output.color[i], tempcolor */
      emit_instruction_op1(emit, VGPU10_OPCODE_MOV, &color_dst,
                           &color_src, FALSE);     /* XXX saturate? */
   }
}


/**
 * Emit extra helper code after the original shader code, but before the
 * last END/RET instruction.
 * For vertex shaders this means emitting the extra code to apply the
 * prescale scale/translation.
 */
static boolean
emit_post_helpers(struct svga_shader_emitter_v10 *emit)
{
   if (emit->unit == PIPE_SHADER_VERTEX) {
      emit_vertex_instructions(emit);
   }
   else if (emit->unit == PIPE_SHADER_FRAGMENT) {
      const unsigned fs_color_tmp_index = emit->fs.color_tmp_index;

      /* We no longer want emit_dst_register() to substitute the
       * temporary fragment color register for the real color output.
       */
      emit->fs.color_tmp_index = INVALID_INDEX;

      if (emit->key.fs.alpha_func != SVGA3D_CMP_ALWAYS) {
         emit_alpha_test_instructions(emit, fs_color_tmp_index);
      }
      if (emit->key.fs.white_fragments) {
         emit_set_color_white(emit, fs_color_tmp_index);
      }
      if (emit->key.fs.write_color0_to_n_cbufs > 1 ||
          emit->key.fs.white_fragments) {
         emit_broadcast_color_instructions(emit, fs_color_tmp_index);
      }
   }

   return TRUE;
}


/**
 * Translate the TGSI tokens into VGPU10 tokens.
 */
static boolean
emit_vgpu10_instructions(struct svga_shader_emitter_v10 *emit,
                         const struct tgsi_token *tokens)
{
   struct tgsi_parse_context parse;
   boolean ret = TRUE;
   boolean pre_helpers_emitted = FALSE;
   unsigned inst_number = 0;

   tgsi_parse_init(&parse, tokens);

   while (!tgsi_parse_end_of_tokens(&parse)) {
      tgsi_parse_token(&parse);

      switch (parse.FullToken.Token.Type) {
      case TGSI_TOKEN_TYPE_IMMEDIATE:
         ret = emit_vgpu10_immediate(emit, &parse.FullToken.FullImmediate);
         if (!ret)
            goto done;
         break;

      case TGSI_TOKEN_TYPE_DECLARATION:
         ret = emit_vgpu10_declaration(emit, &parse.FullToken.FullDeclaration);
         if (!ret)
            goto done;
         break;

      case TGSI_TOKEN_TYPE_INSTRUCTION:
         if (!pre_helpers_emitted) {
            ret = emit_pre_helpers(emit);
            if (!ret)
               goto done;
            pre_helpers_emitted = TRUE;
         }
         ret = emit_vgpu10_instruction(emit, inst_number++,
                                       &parse.FullToken.FullInstruction);
         if (!ret)
            goto done;
         break;

      case TGSI_TOKEN_TYPE_PROPERTY:
         ret = emit_vgpu10_property(emit, &parse.FullToken.FullProperty);
         if (!ret)
            goto done;
         break;

      default:
         break;
      }
   }

done:
   tgsi_parse_free(&parse);
   return ret;
}


/**
 * Emit the first VGPU10 shader tokens.
 */
static boolean
emit_vgpu10_header(struct svga_shader_emitter_v10 *emit)
{
   VGPU10ProgramToken ptoken;

   /* First token: VGPU10ProgramToken  (version info, program type (VS,GS,PS)) */
   ptoken.majorVersion = 4;
   ptoken.minorVersion = 0;
   ptoken.programType = translate_shader_type(emit->unit);
   if (!emit_dword(emit, ptoken.value))
      return FALSE;

   /* Second token: total length of shader, in tokens.  We can't fill this
    * in until we're all done.  Emit zero for now.
    */
   return emit_dword(emit, 0);
}


static boolean
emit_vgpu10_tail(struct svga_shader_emitter_v10 *emit)
{
   VGPU10ProgramToken *tokens;

   /* Replace the second token with total shader length */
   tokens = (VGPU10ProgramToken *) emit->buf;
   tokens[1].value = emit_get_num_tokens(emit);

   return TRUE;
}


/**
 * Modify the FS to read the BCOLORs and use the FACE register
 * to choose between the front/back colors.
 */
static const struct tgsi_token *
transform_fs_twoside(const struct tgsi_token *tokens)
{
   if (0) {
      debug_printf("Before tgsi_add_two_side ------------------\n");
      tgsi_dump(tokens,0);
   }
   tokens = tgsi_add_two_side(tokens);
   if (0) {
      debug_printf("After tgsi_add_two_side ------------------\n");
      tgsi_dump(tokens, 0);
   }
   return tokens;
}


/**
 * Modify the FS to do polygon stipple.
 */
static const struct tgsi_token *
transform_fs_pstipple(struct svga_shader_emitter_v10 *emit,
                      const struct tgsi_token *tokens)
{
   const struct tgsi_token *new_tokens;
   unsigned unit;

   if (0) {
      debug_printf("Before pstipple ------------------\n");
      tgsi_dump(tokens,0);
   }

   new_tokens = util_pstipple_create_fragment_shader(tokens, &unit, 0,
                                                     TGSI_FILE_INPUT);

   emit->fs.pstipple_sampler_unit = unit;

   /* Setup texture state for stipple */
   emit->key.tex[unit].texture_target = PIPE_TEXTURE_2D;
   emit->key.tex[unit].swizzle_r = TGSI_SWIZZLE_X;
   emit->key.tex[unit].swizzle_g = TGSI_SWIZZLE_Y;
   emit->key.tex[unit].swizzle_b = TGSI_SWIZZLE_Z;
   emit->key.tex[unit].swizzle_a = TGSI_SWIZZLE_W;

   if (0) {
      debug_printf("After pstipple ------------------\n");
      tgsi_dump(new_tokens, 0);
   }

   return new_tokens;
}

/**
 * Modify the FS to support anti-aliasing point.
 */
static const struct tgsi_token *
transform_fs_aapoint(const struct tgsi_token *tokens,
                     int aa_coord_index)
{
   if (0) {
      debug_printf("Before tgsi_add_aa_point ------------------\n");
      tgsi_dump(tokens,0);
   }
   tokens = tgsi_add_aa_point(tokens, aa_coord_index);
   if (0) {
      debug_printf("After tgsi_add_aa_point ------------------\n");
      tgsi_dump(tokens, 0);
   }
   return tokens;
}

/**
 * This is the main entrypoint for the TGSI -> VPGU10 translator.
 */
struct svga_shader_variant *
svga_tgsi_vgpu10_translate(struct svga_context *svga,
                           const struct svga_shader *shader,
                           const struct svga_compile_key *key,
                           unsigned unit)
{
   struct svga_shader_variant *variant = NULL;
   struct svga_shader_emitter_v10 *emit;
   const struct tgsi_token *tokens = shader->tokens;
   struct svga_vertex_shader *vs = svga->curr.vs;
   struct svga_geometry_shader *gs = svga->curr.gs;

   assert(unit == PIPE_SHADER_VERTEX ||
          unit == PIPE_SHADER_GEOMETRY ||
          unit == PIPE_SHADER_FRAGMENT);

   /* These two flags cannot be used together */
   assert(key->vs.need_prescale + key->vs.undo_viewport <= 1);

   /*
    * Setup the code emitter
    */
   emit = alloc_emitter();
   if (!emit)
      return NULL;

   emit->unit = unit;
   emit->key = *key;

   emit->vposition.need_prescale = (emit->key.vs.need_prescale ||
                                   emit->key.gs.need_prescale);
   emit->vposition.tmp_index = INVALID_INDEX;
   emit->vposition.so_index = INVALID_INDEX;
   emit->vposition.out_index = INVALID_INDEX;

   emit->fs.color_tmp_index = INVALID_INDEX;
   emit->fs.face_input_index = INVALID_INDEX;
   emit->fs.fragcoord_input_index = INVALID_INDEX;

   emit->gs.prim_id_index = INVALID_INDEX;

   emit->clip_dist_out_index = INVALID_INDEX;
   emit->clip_dist_tmp_index = INVALID_INDEX;
   emit->clip_dist_so_index = INVALID_INDEX;
   emit->clip_vertex_out_index = INVALID_INDEX;

   if (emit->key.fs.alpha_func == SVGA3D_CMP_INVALID) {
      emit->key.fs.alpha_func = SVGA3D_CMP_ALWAYS;
   }

   if (unit == PIPE_SHADER_FRAGMENT) {
      if (key->fs.light_twoside) {
         tokens = transform_fs_twoside(tokens);
      }
      if (key->fs.pstipple) {
         const struct tgsi_token *new_tokens =
            transform_fs_pstipple(emit, tokens);
         if (tokens != shader->tokens) {
            /* free the two-sided shader tokens */
            tgsi_free_tokens(tokens);
         }
         tokens = new_tokens;
      }
      if (key->fs.aa_point) {
         tokens = transform_fs_aapoint(tokens, key->fs.aa_point_coord_index);
      }
   }

   if (SVGA_DEBUG & DEBUG_TGSI) {
      debug_printf("#####################################\n");
      debug_printf("### TGSI Shader %u\n", shader->id);
      tgsi_dump(tokens, 0);
   }

   /**
    * Rescan the header if the token string is different from the one
    * included in the shader; otherwise, the header info is already up-to-date
    */
   if (tokens != shader->tokens) {
      tgsi_scan_shader(tokens, &emit->info);
   } else {
      emit->info = shader->info;
   }

   emit->num_outputs = emit->info.num_outputs;

   if (unit == PIPE_SHADER_FRAGMENT) {
      /* Compute FS input remapping to match the output from VS/GS */
      if (gs) {
         svga_link_shaders(&gs->base.info, &emit->info, &emit->linkage);
      } else {
         assert(vs);
         svga_link_shaders(&vs->base.info, &emit->info, &emit->linkage);
      }
   } else if (unit == PIPE_SHADER_GEOMETRY) {
      assert(vs);
      svga_link_shaders(&vs->base.info, &emit->info, &emit->linkage);
   }

   determine_clipping_mode(emit);

   if (unit == PIPE_SHADER_GEOMETRY || unit == PIPE_SHADER_VERTEX) {
      if (shader->stream_output != NULL || emit->clip_mode == CLIP_DISTANCE) {
         /* if there is stream output declarations associated
          * with this shader or the shader writes to ClipDistance
          * then reserve extra registers for the non-adjusted vertex position
          * and the ClipDistance shadow copy
          */
         emit->vposition.so_index = emit->num_outputs++;

         if (emit->clip_mode == CLIP_DISTANCE) {
            emit->clip_dist_so_index = emit->num_outputs++;
            if (emit->info.num_written_clipdistance > 4)
               emit->num_outputs++;
         }
      }
   }

   /*
    * Do actual shader translation.
    */
   if (!emit_vgpu10_header(emit)) {
      debug_printf("svga: emit VGPU10 header failed\n");
      goto cleanup;
   }

   if (!emit_vgpu10_instructions(emit, tokens)) {
      debug_printf("svga: emit VGPU10 instructions failed\n");
      goto cleanup;
   }

   if (!emit_vgpu10_tail(emit)) {
      debug_printf("svga: emit VGPU10 tail failed\n");
      goto cleanup;
   }

   if (emit->register_overflow) {
      goto cleanup;
   }

   /*
    * Create, initialize the 'variant' object.
    */
   variant = svga_new_shader_variant(svga);
   if (!variant)
      goto cleanup;

   variant->shader = shader;
   variant->nr_tokens = emit_get_num_tokens(emit);
   variant->tokens = (const unsigned *)emit->buf;
   emit->buf = NULL;  /* buffer is no longer owed by emitter context */
   memcpy(&variant->key, key, sizeof(*key));
   variant->id = UTIL_BITMASK_INVALID_INDEX;

   /* The extra constant starting offset starts with the number of
    * shader constants declared in the shader.
    */
   variant->extra_const_start = emit->num_shader_consts[0];
   if (key->gs.wide_point) {
      /**
       * The extra constant added in the transformed shader
       * for inverse viewport scale is to be supplied by the driver.
       * So the extra constant starting offset needs to be reduced by 1.
       */
      assert(variant->extra_const_start > 0);
      variant->extra_const_start--;
   }

   variant->pstipple_sampler_unit = emit->fs.pstipple_sampler_unit;

   /* If there was exactly one write to a fragment shader output register
    * and it came from a constant buffer, we know all fragments will have
    * the same color (except for blending).
    */
   variant->constant_color_output =
      emit->constant_color_output && emit->num_output_writes == 1;

   /** keep track in the variant if flat interpolation is used
    *  for any of the varyings.
    */
   variant->uses_flat_interp = emit->uses_flat_interp;

   if (tokens != shader->tokens) {
      tgsi_free_tokens(tokens);
   }

cleanup:
   free_emitter(emit);

   return variant;
}