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
|
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office">
<head>
<meta http-equiv="Content-Language" content="en-us" />
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
<title>Welcome to BLTsville</title>
<style type="text/css">
.cmd_line {
background: #000;
color: #fff;
padding: 10px;
}
.inline_code {
font-family: "Courier New", Courier, monospace;
font-size: small;
}
.code_block {
margin-left: 40px;
font-family: "Courier New", Courier, monospace;
font-size: small;
}
.small_code_block_in_table {
font-family: "Courier New", Courier, monospace;
font-size: x-small;
}
.small_code_block {
font-family: "Courier New", Courier, monospace;
font-size: x-small;
margin-left: 40px;
}
.underline_code {
font-family: "Courier New", Courier, monospace;
font-size: small;
text-decoration: underline;
}
.Code_Header {
font-size: xx-large;
font-weight: bold;
text-align: left;
font-family: "Courier New", Courier, monospace;
}
.Code_Header_2 {
font-size: x-large;
font-weight: bold;
text-align: left;
font-family: "Courier New", Courier, monospace;
}
.Code_Header_3 {
font-size: large;
font-weight: bold;
text-align: left;
font-family: "Courier New", Courier, monospace;
}
.imponly {
border: thin solid #666666;
margin-left: 40px;
background-color: #E5E5E5;
font-family: Arial, Helvetica, sans-serif;
color: #333333;
}
.Header1 {
margin: 0px 0 0 0;
font-size: xx-large;
font-weight: bold;
text-align: left;
line-height: normal;
background-color: #E0E0E0;
}
.Header2 {
font-size: xx-large;
font-weight: bold;
margin: 0px;
line-height: 100%;
}
.Header3 {
font-size: x-large;
font-weight: bold;
margin: 0px;
line-height: 100%;
}
.Header4 {
font-size: large;
font-weight: bold;
margin: 0px;
line-height: 100%;
}
.strong_emphasis {
text-decoration: underline;
font-weight: bold;
}
.filename {
font-family: Arial, Helvetica, sans-serif;
font-size: small;
}
.underline {
text-decoration: underline;
}
.grn_left {
text-align: left;
color: #009900;
}
.left_topbord {
text-align: left;
border-top-style: solid;
border-top-width: 1px;
}
.center_topbord {
text-align: center;
border-top-style: solid;
border-top-width: 1px;
}
.blue_left_botbord {
text-align: left;
border-bottom-style: solid;
border-bottom-width: 1px;
color: #0000FF;
}
.blue_center_botbord {
text-align: center;
border-bottom-style: solid;
border-bottom-width: 1px;
color: #0000FF;
}
.red_center {
text-align: center;
color: #FF0000;
}
.red_left {
text-align: left;
color: #FF0000;
}
.grn_center {
text-align: center;
color: #009900;
}
.red_center_topbord {
text-align: center;
border-top-style: solid;
border-top-width: 1px;
color: #FF0000;
}
.blu_center_topbord {
text-align: center;
border-top-style: solid;
border-top-width: 1px;
color: #0000FF;
}
.indent_thick_bord {
border-color: #000000;
border-style: solid;
margin-left: 40px;
}
.thin_bord {
border-style: solid;
border-width: 1px;
}
.thin_bord_dbl_botbord {
border-left-style: solid;
border-left-width: 1px;
border-right-style: solid;
border-right-width: 1px;
border-top-style: solid;
border-top-width: 1px;
border-bottom-style: double;
border-bottom-width: 3px;
}
.dl_link {
float: right;
}
.note {
font-family: Arial, Helvetica, sans-serif;
font-weight: bold;
margin-left: 40px;
}
.small_note {
font-size: xx-small;
line-height: 100%;
}
.ctr {
text-align: center;
}
.glyph_cache {
font-family: "Courier New", Courier, monospace;
font-size: medium;
font-weight: normal;
font-style: normal;
margin-left: 40px;
background-color: #000000;
color: #FFFFFF;
}
.example {
border-style: solid;
border-width: 1px;
margin-left: 40px;
margin-right: 40px;
}
.rt_thick_bord {
border-right-style: solid;
border-right-color: #000000;
}
.indent {
margin-left: 40px;
}
.ctr_thin_bord {
border-style: solid;
border-width: 1px;
text-align: center;
}
.indent_thin_bord {
border-style: solid;
border-width: 1px;
margin-left: 40px;
}
.bold_sans {
font-family: Arial, Helvetica, sans-serif;
font-weight: bold;
}
.nowrap {
border-style: solid;
border-width: 1px;
white-space: nowrap;
font-size: small;
}
</style>
</head>
<body>
<table style="width: 100%; line-height: 100%;">
<tr>
<td style="width: 484px">
<table>
<tr>
<td>
<div style="background-position: center; background-image: url('bvlogo.png'); width: 484px; height: 400px; background-repeat: no-repeat;">
<div style="position: relative; left: 0; top: 0;">
<a href="http://graphics.github.com/ocd">
<img src="ocdtab.png" alt="Now With OCD" style="border-width: 0; position: absolute; top: 0; right: 0;" /></a>
</div>
</div>
</td>
</tr>
<tr>
<td class="ctr"><span class="Header2">Version 2.2</span></td>
</tr>
</table>
</td>
<td>
<p>BLTsville is the open 2-D API designed to provide an abstract interface for both hardware and software 2-D implementations.</p>
<p>BLTs (BLock Transfers) involve the moving around of blocks (rectangles) of pixels. BLTsville is the place
to go for BLTs.</p>
<hr />
<table style="width: 100%">
<tr>
<td>
<div class="dl_link">
<img alt="CC BY-ND" longdesc="Creative Commons Attribution-NoDerivs 3.0 Unported License" src="http://i.creativecommons.org/l/by-nd/3.0/88x31.png" width="88" height="31" /></div>
<p class="Header2">License</p>
</td>
</tr>
<tr>
<td>
<div>
<p class="small_note">The API is designed and maintained by Texas Instruments, Inc., but anyone is free
to use it with no cost or obligation.</p>
<p>This project is licensed under the <a href="http://creativecommons.org/licenses/by-nd/3.0/">Creative
Commons Attribution-NoDerivs 3.0 Unported License</a> (user mode), and the
<a href="http://www.gnu.org/licenses/gpl-2.0.html">GNU General Public License version 2</a> (kernel
mode).</p>
</div>
</td>
</tr>
</table>
<hr />
<table style="width: 100%">
<tr>
<td>
<p class="Header2">Dependencies</p>
</td>
</tr>
<tr>
<td>
<p>This project is dependent on the <a href="http://graphics.github.com/ocd">Open Color format Defintions
(OCD) project</a>.</p>
</td>
</tr>
</table>
<hr />
<table style="width: 100%">
<tr>
<td>
<p class="Header2">Source</p>
</td>
</tr>
<tr>
<td>
<div class="dl_link">
<a href="http://github.com/graphics/bltsville/zipball/master">
<img width="90" alt="download zip" src="http://github.com/images/modules/download/zip.png" /></a>
<a href="http://github.com/graphics/bltsville/tarball/master">
<img width="90" alt="download tar" src="http://github.com/images/modules/download/tar.png" /></a>
</div>
<div>
Get the source code (headers) from GitHub at <a href="http://github.com/graphics/bltsville">github.com/graphics/bltsville</a>,
or download the project in <a href="http://github.com/graphics/bltsville/zipball/master">zip</a> or
<a href="http://github.com/graphics/bltsville/tarball/master">tar</a> format.</div>
<p>You can also clone the project with <a href="http://git-scm.com">Git</a> by running:</p>
<pre><a class="cmd_line">$ git clone git://github.com/graphics/bltsville</a></pre>
</td>
</tr>
<tr>
<td><hr />
<table style="width: 100%">
<tr>
<td class="Header2">Wiki</td>
</tr>
<tr>
<td><a href="https://github.com/graphics/bltsville/wiki">https://github.com/graphics/bltsville/wiki</a></td>
</tr>
</table>
</td>
</tr>
</table>
</td>
</tr>
</table>
<hr />
<p class="Header1">Points of Interest in BLTsville</p>
<table style="width: 100%">
<tr>
<td style="width: 50%" valign="top">
<ul>
<li>Solid fills</li>
<li>Pattern fills</li>
<li>Copies</li>
<li>Color format conversion<ul>
<li>Extensive color format support<ul>
<li>RGB, BGR</li>
<li>RGBA, ARGB, etc.</li>
<li>YCbCr (YUV)<ul>
<li>subsampling</li>
<li>packed</li>
<li>planar</li>
</ul>
</li>
<li>Monochrome</li>
<li>Alpha-only</li>
<li>Look-Up Table (LUT)</li>
</ul>
</li>
<li>Extensible color format</li>
</ul>
</li>
<li>ROP4<ul>
<li>Three inputs</li>
</ul>
</li>
<li>Blends<ul>
<li>Pre-defined Porter-Duff blends</li>
<li>Pre-defined DirectFB support</li>
<li>Extensible blends</li>
</ul>
</li>
<li>Multiple </li>
<li>Filters<ul>
<li>Extensible filters</li>
</ul>
</li>
<li>Independent horizontal and vertical <strong>flipping</strong></li>
<li>Independent <strong>scaling</strong> of all three inputs</li>
<li>Clipping</li>
<li>Independent <strong>rotation</strong> of all three inputs (multiples of 90 degrees)</li>
</ul>
</td>
<td style="width: 50%" valign="top">
<ul>
<li>Choice of <strong>scaling</strong> type<ul>
<li>Quality based choice</li>
<li>Speed based choice</li>
<li>Image type based choice</li>
<li>Specific scale type choice</li>
<li>Extensible scale type</li>
</ul>
</li>
<li>Synchronous operations</li>
<li>Asynchronous operations<ul>
<li>Client notification of BLT completion</li>
</ul>
</li>
<li>Batching<ul>
<li>Combine multiple BLTs into group that can be handled more efficiently by implementations<ul>
<li>Character BLTs</li>
<li>Multi-layer blending</li>
<li>ROP/Blend combination with specified ordering</li>
<li>etc.</li>
</ul>
</li>
<li>Delta BLTs</li>
</ul>
</li>
<li>Dithering<ul>
<li>Quality based choice</li>
<li>Speed based choice</li>
<li>Image type based choice</li>
<li>Specific dither type choice</li>
<li>Extensible dither type</li>
</ul>
</li>
<li>Any implementation support<ul>
<li>CPU</li>
<li>2-D Accelerator</li>
</ul>
</li>
</ul>
</td>
</tr>
</table>
<hr />
<ul>
<li>BLTsville does not dictate capabilities of the implementations<ul>
<li>Operations specified either work or return an error indicating that the operation is not supported</li>
</ul>
</li>
</ul>
<hr />
<p class="Header1">How to Get to BLTsville</p>
<p>BLTsville's API is defined in the BLTsville header files. A client must include <span class="inline_code">bltsville.h</span>
to access the implementations. This header includes the remaining headers (including <span class="inline_code">ocd.h</span>).</p>
<p class="note">NOTE: The <span class="underline_code">bvinternal.h</span><span class="underline"> header is for implementations
only</span> and should not be used by clients.</p>
<p>BLTsville has both user mode and a kernel mode interaces. The kernel mode interface is quite similar to (and compatible
with) the user mode, but due to the minor differences and license issues, there are two different sets of header files.</p>
<hr />
<p class="Header1">History of BLTsville</p>
<br />
<p class="Header4">Versions 1.x</p>
<p>BLTsville was based on a previous closed interface, which had a few implementations and shipped on a few devices.
That interface represented the 1.x versions. A lot was learned from that work, and these lessons were used in the
founding of BLTsville.</p>
<p class="Header4">Version 2.0</p>
<p>This was the initial release of the user mode interface. This version is not compatible with the 1.x versions.
Several minor updates were posted, but the API itself did not change, so no changes to the client or implementation were
required.</p>
<p class="Header4">Version 2.1</p>
<p>This is a minor update to the API, and it adds the kernel mode interface. Some additions to the API have been made.
Details of the changes are below with their compatibility matrices.</p>
<ul>
<li><span class="inline_code"><a href="#bv_cache">bv_cache()</a></span> was added to allow manipulation of the CPU cache.
This is an optional interface meant for hardware implementations.</li>
</ul>
<table class="indent_thin_bord">
<tr>
<td class="thin_bord"> </td>
<td class="ctr_thin_bord"><strong>2.0 Client</strong></td>
<td class="ctr_thin_bord"><strong>2.0 Client</strong><br />
(w/2.1 Headers)</td>
<td class="ctr_thin_bord"><strong>2.1 Client</strong></td>
</tr>
<tr>
<td class="thin_bord"><strong>2.0 Implementation</strong></td>
<td class="ctr_thin_bord">compatible</td>
<td class="ctr_thin_bord">New function and structure definitions have no effect.</td>
<td class="ctr_thin_bord">Client must deal with lack of <span class="inline_code"><a href="#bv_cache">bv_cache()</a></span>.</td>
</tr>
<tr>
<td class="thin_bord"><strong>2.0 Implementation</strong><br />
(w/2.1 Headers)</td>
<td class="ctr_thin_bord">New function and structure definitions have no effect.</td>
<td class="ctr_thin_bord">New function and structure definitions have no effect.</td>
<td class="ctr_thin_bord">Client must deal with lack of <span class="inline_code"><a href="#bv_cache">bv_cache()</a></span>.</td>
</tr>
<tr>
<td class="thin_bord"><strong>2.1 Implementation</strong></td>
<td class="ctr_thin_bord">New function and structures have no effect.</td>
<td class="ctr_thin_bord">New function and structures have no effect.</td>
<td class="ctr_thin_bord">compatible</td>
</tr>
</table>
<ul>
<li><span class="inline_code"><a href="#bvbuffdesc">bvbuffdesc</a></span> was extended with the
<span class="inline_code"><a href="#bvbuffdesc.auxtype">auxtype</a></span> and <span class="inline_code">
<a href="#bvbuffdesc.auxptr">auxptr</a></span> members to allow buffer descriptions beyond a virtual address.
Note that only the kernel mode interface currently includes a standard <span class="inline_code">auxtype</span>, but
user mode interface <span class="inline_code">auxtype</span>s may be added later. Both interfaces provide a mechanism
for individual vendors to add their own <span class="inline_code">auxtype</span>, using the same vendor ID mechanism
as the rest of BLTsville.</li>
</ul>
<table class="indent_thin_bord">
<tr>
<td class="thin_bord"> </td>
<td class="ctr_thin_bord"><strong>2.0 Client</strong></td>
<td class="ctr_thin_bord"><strong>2.0 Client</strong><br />
(w/2.1 Headers)</td>
<td class="ctr_thin_bord"><strong>2.1 Client</strong></td>
</tr>
<tr>
<td class="thin_bord"><strong>2.0 Implementation</strong></td>
<td class="ctr_thin_bord">compatible</td>
<td class="ctr_thin_bord">Client must clear <span class="inline_code"><a href="#bvbuffdesc">bvbuffdesc</a></span>
using <span class="inline_code"><span style="white-space: nowrap">sizeof(<a href="#bvbuffdesc">bvbuffdesc</a>)</span></span>.</td>
<td class="ctr_thin_bord">Implementation must handle <span class="inline_code">
<a href="#bvbuffdesc.structsize">bvbuffdesc.structsize</a> > <span style="white-space: nowrap">sizeof(<a href="#bvbuffdesc">bvbuffdesc</a>)</span></span>.</td>
</tr>
<tr>
<td class="thin_bord"><strong>2.0 Implementation</strong><br />
(w/2.1 Headers)</td>
<td class="ctr_thin_bord">Implementation must handle <span class="inline_code">
<a href="#bvbuffdesc.structsize">bvbuffdesc.structsize</a> < <span style="white-space: nowrap">sizeof(<a href="#bvbuffdesc">bvbuffdesc</a>)</span></span>.</td>
<td class="ctr_thin_bord">Client must clear <span class="inline_code"><a href="#bvbuffdesc">bvbuffdesc</a></span>
using <span class="inline_code"><span style="white-space: nowrap">sizeof(<a href="#bvbuffdesc">bvbuffdesc</a>)</span></span>.</td>
<td class="ctr_thin_bord">Client must deal with implementation that uses <span class="inline_code">
<a href="#bvbuffdesc.virtaddr">bvbuffdesc.virtaddr</a></span> or returns error if <span class="inline_code">
<a href="#bvbuffdesc.virtaddr">bvbuffdesc.virtaddr</a></span> is 0.</td>
</tr>
<tr>
<td class="thin_bord"><strong>2.1 Implementation</strong></td>
<td class="ctr_thin_bord">Implementation must handle <span class="inline_code">
<a href="#bvbuffdesc.structsize">bvbuffdesc.structsize</a> < <span style="white-space: nowrap">sizeof(<a href="#bvbuffdesc">bvbuffdesc</a>)</span></span>.</td>
<td class="ctr_thin_bord">Client must clear <span class="inline_code"><a href="#bvbuffdesc">bvbuffdesc</a></span>
using <span class="inline_code"><span style="white-space: nowrap">sizeof(<a href="#bvbuffdesc">bvbuffdesc</a>)</span></span>.</td>
<td class="ctr_thin_bord">compatible</td>
</tr>
</table>
<ul>
<li>Added documentation of <a href="#NOP">NOP BLT</a> used as synchronization mechanism for <a href="#BVFLAG_ASYNC">
asynchronous BLTs</a>.<ul>
<li>Clients that do not use <a href="#BVFLAG_ASYNC">asynchronous BLTs</a> or the <a href="#NOP">NOP BLT</a> will
not be affected.</li>
<li>Implementations that do not support the NOP BLT will return an error. This will not cause a problem
for clients when using implementations which are actually synchronous. For clients using asynchronous
implementations, an alternate supported but innocuous BLT will be necessary (e.g. copying a pixel to itself).</li>
</ul>
</li>
</ul>
<p class="Header4">Version 2.2</p>
<p>This is a minor update which includes the following:</p>
<ul>
<li>Addition of the <span class="inline_code"><a href="#src2auxdstrect">src2auxdstrect</a></span> and
<span class="inline_code"><a href="#maskauxdstrect">maskauxdstrect</a></span> members to <span class="inline_code">
<a href="#bvbltparams">bvbltparams</a></span> with example.</li>
<li>Addition of <span class="inline_code"><a href="#BVFLAG_SRC2_AUXDSTRECT">BVFLAG_SRC2_AUXDSTRECT</a></span> and <span class="inline_code">
<a href="#BVFLAG_MASK_AUXDSTRECT">BVFLAG_MASK_AUXDSTRECT</a></span> flags.</li>
<li>Added <span class="inline_code"><a href="#BVAT_PHYSADDR">BVAT_PHYSADDR</a></span> to the kernel mode
<span class="inline_code"><a href="#bvbuffdesc.auxtype">bvbuffdesc.auxtype</a></span> enumerations.</li>
<li>Added clarification to the <span class="inline_code"><a href="#bvphysdesc">bvphysdesc</a></span> documentation.</li>
</ul>
<p>Compatibility</p>
<ul>
<li>Clients that do not use the <span class="inline_code"><a href="#BVFLAG_SRC1_AUXDSTRECT">BVFLAG_*_AUXDSTRECT</a></span>
flags will not be affected.</li>
<li>Clients using the new (long) <span class="inline_code"><a href="#bvbltparams">bvbltparams</a></span> will work with
older implementations. If the clients set the <span class="inline_code"><a href="#BVFLAG_SRC2_AUXDSTRECT">
BVFLAG_*_AUXDSTRECT</a></span> flags, the implementations will return <span class="inline_code">BVERR_FLAGS</span>,
indicating the lack of support for this feature.</li>
<li>Implementations supporting the new <span class="inline_code"><a href="#bvbltparams">bvbltparams</a></span> will
accept the older (smaller) version, distinguished by the <span class="inline_code">
<a href="#bvbltparams.structsize">structsize</a></span> member. Clients using the older versions will not set
the <span class="inline_code"><a href="#BVFLAG_SRC2_AUXDSTRECT">BVFLAG_*_AUXDSTRECT</a></span> flags, so the new
members will not be utilized.</li>
<li>Clients using <span class="inline_code"><a href="#BVAT_PHYSADDR">BVAT_PHYSADDR</a></span> will get an error from
implementations that do not support this enumeration. The <span class="inline_code"><a href="#BVAT_PHYSDESC">
BVAT_PHYSDESC</a></span> may be used if supported by the implementation, but care must be taken to ensure the buffer
is defined properly. See <span class="inline_code"><a href="#bvphysdesc">bvphysdesc</a></span> for details.</li>
</ul>
<hr />
<p class="Header1">BLTsville Neighborhoods</p>
<p>Implementations may be software (CPU) or 2-D hardware, and many may coexist. Each implementation will have an individual
entry point, so it can be directly addressed. But there will also be a more general interface for each of these two
types of implementations so that system integrators can choose the most appropriate implementation. In other words,
the system integrator will choose one software and one 2-D hardware implementation to be the "default" used when a client
does not need to choose a specific implementation.</p>
<p class="Header2">User Mode Interface</p>
<p>Clients use the standard names below to access the default implementations. The client then imports the pointers
to the functions. (The specific name decoration and import method will be dictated by the host Operating System (O/S).)
Some examples:</p>
<ul>
<li>CPU: <span class="filename">bltsville_cpu</span><ul>
<li>Linux/Android: <span class="filename">libbltsville_cpu.so</span></li>
<li>Windows: <span class="filename">bltsville_cpu.dll</span></li>
<li>etc.</li>
</ul>
</li>
<li>2-D hardware: <span class="filename">bltsville_hw2d</span><ul>
<li>Linux/Android: <span class="filename">libbltsville_hw2d.so</span></li>
<li>Windows: <span class="filename">bltsville_hw2d.dll</span></li>
<li>etc.</li>
</ul>
</li>
</ul>
<p>Usually these entry points will be symbolic links (either explicit in systems like Linux which support them, or implicit
using a thin wrapper) to the specific implementation. This allows system integrators to connect the client with the
most capable implementation available in the system. For example, <span class="filename">bltsville_hw2d</span> might
be a symbolic link to <span class="filename">bltsville_gc2d</span>.</p>
<p>In addition, there may be more implementations co-existing in a given system. These will have additional unique
names as determined by the vendors. For example:</p>
<ul>
<li>Reference CPU software implementation: <span class="filename">bltsville_refcpu</span></li>
<li>System DMA 2-D hardware implementation: <span class="filename">bltsville_mydma</span></li>
</ul>
<p class="Header3">Initialization</p>
<p>In general, each O/S has the ability to manually load a library. This in turn causes a function in the library
to be called so the library can perform initialization. Unfortunately, not all O/Ss allow this initialization
function to return an error if the initialization fails. Equally unfortunately, it may be necessary for the
initialization to be performed in that function. To accommodate this, BLTsville defers the specific initialization
to the O/S environment.</p>
<p class="Header4">Linux/Android</p>
<p>The client will call <span class="inline_code">dlopen()</span> to open the library. It will then import the
<span class="inline_code">bv_*()</span> functions, and call them as desired. Initialization will occur in
association with one or more of these activities. If the initialization fails, the bv_*() functions will return
the <span class="inline_code">BVERR_RSRC</span> error, indicating that a required resource was not obtained.</p>
<p class="imponly"><strong>Implementations Only<br />
</strong><br />
If the library has designated a function with the <span class="inline_code">__attribute__ ((constructor))</span>, that
function will be called. Linux implementations may use this function to perform initialization (including opening
an interface to an associated kernel module). However, since this function cannot return an error, and thus cannot
fail, if the initialization fails, this must be recorded. Then, when the client calls any of the
<span class="inline_code">bv_*()</span> functions, these should immediately return the <span class="inline_code">
BVERR_RSRC</span> error, indicating that the library was unable to initialize (obtain a necessary resource).<br />
<br />
Linux implementations may also choose to initialize on the first call to a <span class="inline_code">bv_*()</span>
function. Failure is likewise indicated by returning the <span class="inline_code">BVERR_RSRC</span> error.<br />
<br />
<strong>NOTE: Be careful not to repeatedly attempt initialization when a failure is encountered. Some
initializations, and especially initialization failures, can take a long time. This means clients trying to call
</strong><span class="inline_code"><strong>bv_*()</strong></span><strong> functions (presumably before falling back to
alternatives) will be repeatedly penalized if the library can't initialize. Instead, attempt initialization
once, and from them on return <span class="inline_code">BVERR_RSRC</span>.</strong></p>
<p class="Header2">Kernel Mode Interface</p>
<p>For most kernel space BLTsville clients, only a 2-D hardware implementation will be used. However, both types of
implementations are supported. Clients use the standard names below to access the default implementations and obtain
pointers to the functions. (The specific method of obtaining the interface will be dictated by the host Operating
System (O/S).) Some examples:</p>
<ul>
<li>CPU<ul>
<li>Linux/Android <span class="inline_code">bvcpu_entry()</span></li>
<li>etc.</li>
</ul>
</li>
<li>2-D hardware<ul>
<li>Linux/Android <span class="inline_code">bv2d_entry()</span></li>
<li>etc.</li>
</ul>
</li>
</ul>
<p>These entry points may represent the implementations themselves, but more likely they will link the client to the implementations
using more specific names. For example, <span class="inline_code">bv2d_entry()</span> may link the client to
<span class="inline_code">gcbv_entry()</span>.</p>
<p>In addition, there may be more implementations co-existing in the kernel. These will require additional unique
names as determined by the vendors. For example:</p>
<ul>
<li>Reference CPU software implementation: <span class="inline_code">cpurefbv_entry()</span></li>
<li>Vivante GC320 2-D hardware implementation: <span class="inline_code">gcbv_entry()</span></li>
</ul>
<hr />
<p class="Header1">Things To Do In BLTsville</p>
<p>BLTsville's interface consists of three or four functions per implementation, which must be imported by the
client at run time:</p>
<ul>
<li><span class="inline_code"><a href="#bv_map">bv_map()</a></span></li>
<li><span class="inline_code"><a href="#bv_blt">bv_blt()</a></span></li>
<li><span class="inline_code"><a href="#bv_unmap">bv_unmap()</a></span></li>
<li><span class="inline_code"><a href="#bv_cache">bv_cache()</a></span> (optional)</li>
</ul>
<p class="note">NOTE: If the library failed to initialize, these functions will return <span class="inline_code">
BVERR_RSRC</span>, indicating that a required resource was not obtained.</p>
<a name="bv_map" class="Code_Header">bv_map()</a>
<p class="code_block">enum bverror bv_map(<a href="#bvbuffdesc">struct bvbuffdesc* buffdesc</a>);</p>
<p><span class="strong_emphasis">BLTsville does not allocate buffers.</span> Clients must describe a buffer
in BLTsville using the <span class="inline_code"><a href="#bvbuffdesc">bvbuffdesc</a></span> structure so a given implementation
can access the buffer.</p>
<p><span class="inline_code">bv_map()</span> is used to provide the implementation an opportunity to associate hardware
resources with the specified buffer. Most hardware requires this type of mapping, and there is usually appreciable
overhead associated with it. By providing a separate call for this operation, BLTsville allows the client to move
this overhead to the most appropriate time in its execution.</p>
<p>For a given buffer, the client can call the <span class="inline_code">bv_map()</span> function imported from each implementation
to establish the mapping immediately. But this is not required.</p>
<p>As a special bonus, BLTsville clients can call to any implementation's <span class="inline_code">bv_map()</span>.
This is sufficient to indicate that the client can be trusted to make the corresponding call to
<span class="inline_code"><a href="#bv_unmap">bv_unmap()</a></span> upon destruction of the buffer. Then when a client
calls an implementation's <span class="inline_code"><a href="#bv_blt">bv_blt()</a></span>, if the mapping needs to be done,
it's done at that time. But the mapping is maintained, so that the overhead is avoided on subsequent
<span class="inline_code"><a href="#bv_blt">bv_blt()</a></span> calls. This lets implementations use <em>lazy mapping</em>
only as necessary. If an implementation is not called, the mapping is not done.</p>
<p><em>Normally, the lowest overhead </em><span class="inline_code"><em>bv_map()</em></span><em> call will be in the CPU-based
implementation. So most clients will want to make a single, low overhead </em><span class="inline_code"><em>bv_map()</em></span><em>
call to the bltsville_cpu implementation to avoid the mapping/unmapping overhead on each </em><span class="inline_code">
<a href="#bv_blt"><em>bv_blt()</em></a></span><em> call, while avoiding the mapping overhead when possible.</em></p>
<p><em><strong>Calling </strong></em><span class="inline_code"><em><strong>bv_map()</strong></em></span><em><strong> is
actually optional prior to calling </strong></em><span class="inline_code"><a href="#bv_blt"><em><strong>bv_blt()</strong></em></a></span><em><strong>.
However, if it is not called at least once for a given buffer, it must be assumed that </strong></em>
<span class="inline_code"><a href="#bv_unmap"><strong><em>bv_unmap()</em></strong></a></span><em><strong> will not be called.
So the mapping must be done when </strong></em><span class="inline_code"><a href="#bv_blt"><em><strong>bv_blt()</strong></em></a></span><em><strong>
is called, and unmapping done when it is complete. This means the overhead will be incurred for every </strong>
</em><a href="#bv_blt" class="inline_code"><em><strong>bv_blt()</strong></em></a><em><strong> call which uses that buffer.</strong></em></p>
<p class="note">NOTE: Obviously any API cannot add capabilities beyond an implementation's capabilities. So, for example,
if an implementation requires memory to be allocated from a special pool of memory, that responsibility falls upon the client.
The <span class="inline_code">bv_map()</span> function for that implementation will need to check the characteristics of
the memory and return an error if it does not meet the necessary criteria.</p>
<p class="Header4"><a name="bv_map_Function_Sequences">Function Sequences</a></p>
<p>To clarify, here are some function sequences and the operations associated with them:</p>
<table class="indent">
<tr>
<td class="ctr_thin_bord"><strong>Implementation</strong></td>
<td class="ctr_thin_bord"><strong>Function</strong></td>
<td class="ctr_thin_bord"><strong>Operation</strong></td>
</tr>
<tr>
<td class="ctr_thin_bord">A</td>
<td class="thin_bord"><span class="inline_code"><a href="#bv_blt">bv_blt()</a></span></td>
<td class="thin_bord">map A<br />
BLT A<br />
unmap A</td>
</tr>
<tr>
<td class="ctr_thin_bord">A</td>
<td class="thin_bord"><span class="inline_code"><a href="#bv_blt">bv_blt()</a></span></td>
<td class="thin_bord">map A<br />
BLT A<br />
unmap A</td>
</tr>
<tr>
<td class="ctr_thin_bord">B</td>
<td class="thin_bord"><span class="inline_code"><a href="#bv_blt">bv_blt()</a></span></td>
<td class="thin_bord">map B<br />
BLT B<br />
unmap B</td>
</tr>
</table>
<br />
<table class="indent">
<tr>
<td class="ctr_thin_bord"><strong>Implementation</strong></td>
<td class="ctr_thin_bord"><strong>Function</strong></td>
<td class="ctr_thin_bord"><strong>Operation</strong></td>
</tr>
<tr>
<td class="ctr_thin_bord">A</td>
<td class="thin_bord"><span class="inline_code"><a href="#bv_map">bv_map()</a></span></td>
<td class="thin_bord">map A</td>
</tr>
<tr>
<td class="ctr_thin_bord">A</td>
<td class="thin_bord"><span class="inline_code"><a href="#bv_blt">bv_blt()</a></span></td>
<td class="thin_bord">BLT A</td>
</tr>
<tr>
<td class="ctr_thin_bord">A</td>
<td class="thin_bord"><span class="inline_code"><a href="#bv_blt">bv_blt()</a></span></td>
<td class="thin_bord">BLT A</td>
</tr>
<tr>
<td class="ctr_thin_bord">A</td>
<td class="thin_bord"><span class="inline_code"><a href="#bv_unmap">bv_unmap()</a></span></td>
<td class="thin_bord">unmap A</td>
</tr>
</table>
<br />
<table class="indent">
<tr>
<td class="ctr_thin_bord"><strong>Implementation</strong></td>
<td class="ctr_thin_bord"><strong>Function</strong></td>
<td class="ctr_thin_bord"><strong>Operation</strong></td>
</tr>
<tr>
<td class="ctr_thin_bord">A</td>
<td class="thin_bord"><span class="inline_code"><a href="#bv_map">bv_map()</a></span></td>
<td class="thin_bord">map A</td>
</tr>
<tr>
<td class="ctr_thin_bord">B</td>
<td class="thin_bord"><span class="inline_code"><a href="#bv_map">bv_map()</a></span></td>
<td class="thin_bord">map B</td>
</tr>
<tr>
<td class="ctr_thin_bord">A</td>
<td class="thin_bord"><span class="inline_code"><a href="#bv_blt">bv_blt()</a></span></td>
<td class="thin_bord">BLT A</td>
</tr>
<tr>
<td class="ctr_thin_bord">B</td>
<td class="thin_bord"><span class="inline_code"><a href="#bv_blt">bv_blt()</a></span></td>
<td class="thin_bord">BLT B</td>
</tr>
<tr>
<td class="ctr_thin_bord">A</td>
<td class="thin_bord"><span class="inline_code"><a href="#bv_unmap">bv_unmap()</a></span></td>
<td class="thin_bord">unmap A</td>
</tr>
<tr>
<td class="ctr_thin_bord">B</td>
<td class="thin_bord"><span class="inline_code"><a href="#bv_unmap">bv_unmap()</a></span></td>
<td class="thin_bord">unmap B</td>
</tr>
</table>
<br />
<table class="indent">
<tr>
<td class="ctr_thin_bord"><strong>Implementation</strong></td>
<td class="ctr_thin_bord"><strong>Function</strong></td>
<td class="ctr_thin_bord"><strong>Operation</strong></td>
</tr>
<tr>
<td class="ctr_thin_bord">A</td>
<td class="thin_bord"><span class="inline_code"><a href="#bv_map">bv_map()</a></span></td>
<td class="thin_bord">map A</td>
</tr>
<tr>
<td class="ctr_thin_bord">B</td>
<td class="thin_bord"><span class="inline_code"><a href="#bv_blt">bv_blt()</a></span></td>
<td class="thin_bord">map B<br />
BLT B</td>
</tr>
<tr>
<td class="ctr_thin_bord">B</td>
<td class="thin_bord"><span class="inline_code"><a href="#bv_blt">bv_blt()</a></span></td>
<td class="thin_bord">BLT B</td>
</tr>
<tr>
<td class="ctr_thin_bord">A</td>
<td class="thin_bord"><span class="inline_code"><a href="#bv_unmap">bv_unmap()</a></span></td>
<td class="thin_bord">unmap A<br />
unmap B</td>
</tr>
</table>
<br />
<div class="note">
NOTE: Calling <span class="inline_code">bv_map()</span> and <span class="inline_code"><a href="#bv_unmap">
bv_unmap()</a></span> with the same <span class="inline_code"><a href="#bvbuffdesc">bvbuffdesc</a></span> from
different, unsynchronized threads, even (especially) from different implementations, will result in undefined
behavior. This is similar to calling <span class="inline_code">malloc()</span> and <span class="inline_code">
free()</span> using the same buffer pointer in different, unsynchronized threads. While this may work
sometimes and for some implementations and combinations of implementations, BLTsville does not provide any
synchronization mechanism to make this safe. Clients must ensure that these calls are synchronized in cases
where such behavior appears to be necessary.</div>
<br />
<a name="bv_blt" class="Code_Header">bv_blt()</a>
<p class="code_block">enum bverror bv_blt(<a href="#bvbltparams">struct bvbltparams* bltparams</a>);</p>
<p>The main function of BLTsville is <span class="inline_code">bv_blt()</span>. A <span class="inline_code">
<a href="#bvbltparams">bvbltparams</a></span> structure is passed into <span class="inline_code">bv_blt()</span> to trigger
the desired 2-D operation.</p>
<a name="bv_unmap" class="Code_Header">bv_unmap()</a>
<p class="code_block">enum bverror bv_unmap(<a href="#bvbuffdesc">struct bvbuffdesc* buffdesc</a>);</p>
<p><span class="inline_code">bv_unmap()</span> is used to free implementation resources associated with a buffer.
Normally, if <span class="inline_code"><a href="#bv_map">bv_map()</a></span> was called for a given buffer,
<span class="inline_code">bv_unmap()</span> should be called as well.</p>
<p>For convenience, only one <span class="inline_code">bv_unmap()</span> needs to be called for each buffer, regardless
of how many implementations were used, including multiple calls to <span class="inline_code"><a href="#bv_map">bv_map()</a></span>.</p>
<p>Also for convenience, <span class="inline_code">bv_unmap()</span> may be called multiple times on the same buffer.
Note that only the first call will actually free (all) the associated resources. See the
<a href="#bv_map_Function_Sequences">Function Sequences</a> under <span class="inline_code"><a href="#bv_map">bv_map()</a></span>
for more details.</p>
<p class="imponly"><strong>Implementations Only</strong><br />
<br />
Implementations must ensure that unmapping of buffers which are in use by asynchronous BLTs are appropriately delayed to
avoid improper access.</p>
<a name="bv_cache" class="Code_Header">bv_cache()</a>
<p class="code_block">enum bverror bv_cache(<a href="#bvcopprams">struct bvcopparams *copparams</a>);</p>
<p><span class="inline_code">bv_cache()</span> provides manual CPU cache control to maintain cache coherence of surfaces
between the CPU and other hardware. The <a href="#bvbuffdesc">bvcopparams</a> structure provides the information needed
to properly manipulate the CPU cache.</p>
<p>This function is <em>optional</em>. If this function fails to import, it means the implementation does not provide
it, but <span class="inline_code"><a href="#bv_map">bv_map()</a></span>, <span class="inline_code">
<a href="#bv_blt">bv_blt()</a></span>, and <span class="inline_code"><a href="#bv_unmap">bv_unmap()</a></span> may still
be used.</p>
<p><em>In general, this function will be provided with BLTsville implementations which utilize 2-D hardware, even though
it manipulates the CPU cache. This is because most systems require a kernel module to manipulate the cache, and this
is not always practical to include with a user-mode CPU implementation.</em></p>
<p><strong>BEWARE: Manipulation of the CPU cache is tricky. Moreover, different CPUs behave differently, so
cache manipulation that works on one device may fail on another. Also, mismanaged operation of the cache can have
significant impact on overall system performance. And incorrect manipulation of the cache can cause instability or
crashes. Please read and understand all of the discussions below before using this function.</strong></p>
<ol>
<li>To avoid system instability, do not perform cache operations on buffers which would not be accessed by BLTsville.</li>
<li>For maximum performance, combine adjacent rectangles into one <span class="inline_code">bv_cache()</span> call.
For example, when BLTing a line of characters, do not issue a <span class="inline_code">bv_cache()</span> call for each
character. Instead, make one call to bv_cache() which includes all the characters.</li>
<li>When using a hardware BLTsville implementation to read data written into a cached surface by the CPU, use the
<span class="inline_code"><a href="#CPU_TO_DEVICE">BVCACHE_CPU_TO_DEVICE</a></span> operation after the CPU has completed
its operation and before the hardware BLTsville operation is initiated.</li>
<li>When using a hardware BLTsville implementation to write data into a cached surface that will be read by the CPU,
use the <span class="inline_code"><a href="#CPU_FROM_DEVICE">BVCACHE_CPU_FROM_DEVICE</a></span> operation after the
hardware BLTsville operation has completed (note this means after the callback if the BLT is asynchronous) and before
the CPU accesses the surface.</li>
<li>When using a hardware BLTsville implementation to write data into a cached surface that has been written by the
CPU, using the <span class="inline_code"><a href="#CPU_TO_DEVICE">BVCACHE_CPU_TO_DEVICE</a></span> operation after the
CPU has completed its operation and before the hardware BLTsville operation is initiated.<ul>
<li class="bold_sans">NOTE: This cache operation may not be necessary on all hardware, but it is good practice to perform it
anyway. This operation will be necessary for a CPU with a write allocation policy on the cache, but may not
be necessary for CPUs without such a configuration.</li>
<li class="bold_sans"><strong>NOTE WELL: CPU access to a destination buffer is not always initiated by the client. Buffers
recently allocated may be cleared by the CPU on behalf of the client via the allocation call. Failure to perform
this operation may result in image corruption even if no further CPU accesses are performed on the surface!</strong></li>
</ul>
</li>
</ol>
<table class="example">
<tr>
<td>
<p><strong>Example</strong>: On one particular device, a surface was allocated using the standard user mode
<span class="inline_code">malloc()</span>. An image was copied into a portion of this surface using a hardware
implementation of BLTsville. The result was then read by the CPU.</p>
<p>Logically, <span class="inline_code">bv_cache()</span> was used to perform a <span class="inline_code">
<a href="#CPU_FROM_DEVICE">BVCACHE_CPU_FROM_DEVICE</a></span> operation after the hardware-based BLTsville operation
completed, but before the CPU read was performed. However, corruption appeared both inside the image copied,
as well as outside the image!</p>
<p>Both corruptions were caused by not realizing that there was a CPU operation (clear) performed on behalf of the
<span class="inline_code">malloc()</span>, for which the proper cache manipulation was not performed.</p>
<p>The corruption outside the image was due to data in the cache being invalidated before it reached the memory.
As mentioned above, buffers allocated are normally cleared by the system. In this case, since the buffer used
for the surface was configured with a write allocated cache, this meant that not all writes to clear the buffer
were in memory when the <span class="inline_code"><a href="#CPU_FROM_DEVICE">BVCACHE_CPU_FROM_DEVICE</a></span>
operation was performed. As a result, the uncommitted data in the cache was invalidated and lost, and the
previous contents of the memory remained for the CPU to read.</p>
<p>The corruption inside the image was caused by data in the cache being committed to memory after the hardware
BLT completed, but before the <span class="inline_code"><a href="#CPU_FROM_DEVICE">BVCACHE_CPU_FROM_DEVICE</a></span>
operation was executed.</p>
<p>Both corruptions were corrected by performing a <span class="inline_code"><a href="#CPU_TO_DEVICE">BVCACHE_CPU_TO_DEVICE</a></span>
operation on the <span class="underline">destination</span> surface <strong>before</strong> performing the BLT (item
5 above), in addition to the <span class="inline_code"><a href="#CPU_FROM_DEVICE">BVCACHE_CPU_FROM_DEVICE</a></span>
operation performed <strong>after</strong> the BLT (item 3 above).</p>
</td>
</tr>
</table>
<br />
<hr /><a name="bvbltparams" class="Code_Header">bvbltparams</a>
<p><span class="inline_code">bvbltparams</span> is the central structure in BLTsville. This structure holds the details
of the BLT being requested by the client.</p>
<p class="small_code_block">union bvop {<br />
unsigned short <a href="#rop">rop</a>;<br />
enum bvblend <a href="#blend">blend</a>;<br />
struct bvfilter *<a href="#filter">filter</a>;<br />
};<br />
<br />
struct bvinbuff {<br />
<a href="#bvbuffdesc">struct bvbuffdesc</a> *<a href="#src1.desc">desc</a>;<br />
<a href="#bvtileparams">struct bvtileparams</a> *<a href="#src1.tileparams">tileparams</a>;<br />
};<br />
<br />
struct bvbltparams {<br />
unsigned int <a href="#bvbltparams.structsize">structsize</a>;<br />
<br />
char *<a href="#errdesc">errdesc</a>;<br />
<br />
unsigned long <a href="#implementation">implementation</a>;<br />
unsigned long <a href="#flags">flags</a>;<br />
union bvop <a href="#op">op</a>;<br />
<br />
void *<a href="#colorkey">colorkey</a>;<br />
union bvalpha <a href="#globalalpha">globalalpha</a>;<br />
<br />
enum bvscalemode <a href="#scalemode">scalemode</a>;<br />
enum bvdithermode <a href="#dithermode">dithermode</a>;<br />
<br />
<a href="#bvbuffdesc">struct bvbuffdesc</a> *<a href="#dstdesc">dstdesc</a>;<br />
<a href="#bvsurfgeom">struct bvsurfgeom</a> *<a href="#dstgeom">dstgeom</a>;<br />
<a href="#bvrect">struct bvrect</a> <a href="#dstrect">dstrect</a>;<br />
<br />
union bvinbuff <a href="#src1">src1</a>;<br />
<a href="#bvsurfgeom">struct bvsurfgeom</a> *<a href="#src1geom">src1geom</a>;<br />
<a href="#bvrect">struct bvrect</a> <a href="#src1rect">src1rect</a>;<br />
<br />
union bvinbuff <a href="#src2">src2</a>;<br />
<a href="#bvsurfgeom">struct bvsurfgeom</a> *<a href="#src2geom">src2geom</a>;<br />
<a href="#bvrect">struct bvrect</a> <a href="#src2rect">src2rect</a>;<br />
<br />
union bvinbuff <a href="#mask">mask</a>;<br />
<a href="#bvsurfgeom">struct bvsurfgeom</a> *<a href="#maskgeom">maskgeom</a>;<br />
<a href="#bvrect">struct bvrect</a> <a href="#maskrect">maskrect</a>;<br />
<br />
<a href="#bvrect">struct bvrect</a> <a href="#cliprect">cliprect</a>;<br />
<br />
unsigned long <a href="#batchflags">batchflags</a>;<br />
struct bvbatch *<a href="#batch">batch</a>;<br />
<br />
void (*<a href="#callbackfn">callbackfn</a>)(<a href="#bvcallbackerror">struct
bvcallbackerror</a> *err,<br />
unsigned long callbackdata);<br />
unsigned long <a href="#callbackdata">callbackdata</a>;<br />
<br />
<a href="#bvrect">struct bvrect</a> <a href="#src2auxdstrect">src2auxdstrect</a>;<br />
<a href="#bvrect">struct bvrect</a> <a href="#maskauxdstrect">maskauxdstrect</a>;<br />
};</p>
<a name="bvbltparams.structsize" class="Code_Header_2">bvbltparams.structsize</a>
<p><span class="code_block">unsigned long structsize; /* input */</span></p>
<p>This member is used to allow backwards and forwards compatibility between versions of BLTsville. It should be set
to the <span class="inline_code">sizeof()</span> the structure by the client or implementation, whichever allocated the
structure.</p>
<p>BLTsville is designed to be forwards and backwards compatible between client and library versions. But this compatibility
would be eliminated if clients chose to check for a specific version of the BLTsville implementations and fail if the specific
version requested was not in place. So, instead of exporting a version number, BLTsville structures use the
<span class="inline_code">structsize</span> member to indicate the number of bytes in the structure. This is used
to communicate between the client and implementation which portions of the structure exist. This effectively bypasses
the concept of a version and focuses on the specifics of what changes need to be considered to maintain compatibility.</p>
<ol>
<li>When an old client calls into a new implementation, that implementation will realize if the client only provides
a subset of an updated structure. The implementation will handle this and utilize only that information which
has been provided. New features will be disabled, but functionality will be maintained.</li>
<li>When a new client calls into an old implementation, that implementation will ignore the extra members of the structure
and operate in ignorance of them. If these members are necessary for some new functionality, this will be evident
from other fields in the structure, so that the implementation can gracefully fail.</li>
</ol>
<p>If <span class="inline_code">structsize</span> is set to a value that is too small for an implementation, it may return
a <span class="inline_code"><a href="#BVERR_BLTPARAMS_VERS">BVERR_BLTPARAMS_VERS</a></span> error.</p>
<p class="Code_Header_2"><a name="bvbltparams.errdesc">bvbltparams.errdesc</a></p>
<p><span class="code_block">char* errdesc; /* output */</span></p>
<p><span class="inline_code">errdesc</span> is optionally used by implementations to pass a 0-terminated string with additional
debugging information back to clients for debugging purposes. <span class="inline_code">errdesc</span> is not localized
or otherwise meant to provide information that is displayed to users.</p>
<p class="Code_Header_2"><a name="implementation">bvbltparams.implementation</a></p>
<p class="code_block">unsigned long implementation; /* input */</p>
<p>Multiple implementations of BLTsville can be combined under managers which can distribute the BLT requests to the implementations
based on whatever criteria the manager chooses. This might include availability of the operation, performance, loading,
or power state. In such a scenario, the client may need to override or augment the choice made by the manager.
This field allows that control.</p>
<p><strong><em>Note that this feature is extremely complicated, and more detailed documentation needs to be created to allow
creation of managers and smooth integration by a client. There are serious issues that must be understood before any
manager can be put into place, such as CPU cache coherence and multiple implementation operation interdependence.
For now, this field should be set to 0 by clients.</em></strong></p>
<p>If the implementation cannot respond to the <span class="inline_code">implementation</span> flags set, it may return
a <span class="inline_code"><a href="#BVERR_IMPLEMENTATION">BVERR_IMPLEMENTATION</a></span> error.</p>
<p class="Code_Header_2"><a name="flags">bvbltparams.flags</a></p>
<p class="code_block">unsigned long flags; /* input */</p>
<p>The <span class="inline_code">flags</span> member provides the baseline of information to <span class="inline_code">
<a href="#bv_blt">bv_blt()</a></span> about the type of BLT being requested.</p>
<p>To maintain compatibility, unused bits in the flags member should be set to 0.</p>
<p>If the flags set are not supported by the implementation, it may return <span class="inline_code">
<a href="#BVERR_FLAGS">BVERR_FLAGS</a></span>, or a more specific <a href="#bverror">error code</a>.</p>
<p class="Code_Header_3"><a name="BVFLAG_OP_">bvbltparams.flags - BVFLAG_OP_*</a></p>
<p>The <span class="inline_code">op</span> field of the flags member specifies the type of BLT operation to perform.
Currently there are three types of BLT operations defined:</p>
<table class="indent">
<tr>
<td valign="top">1.</td>
<td><span class="inline_code"><strong><a name="BVFLAG_ROP">BVFLAG_ROP</a></strong></span><br />
<p>This flag indicates the operation being performed is a raster operation, and the <span class="inline_code">
<a href="#op">bvbltparams.op</a></span> union is treated as <span class="inline_code"><a href="#rop">rop</a></span>.
Raster OPerations are binary operations performed on the bits of the inputs. See <span class="inline_code">
<a href="#rop">bvbltparams.op.rop</a></span> for details.<br />
<br />
</p>
</td>
</tr>
<tr>
<td valign="top">2.</td>
<td>
<p><span class="inline_code"><strong><a name="BVFLAG_BLEND">BVFLAG_BLEND</a></strong></span><br />
</p>
<p>This flag indicates the operation being performed is a blend, and the <span class="inline_code">
<a href="#op">bvbltparams.op</a></span> union is treated as <span class="inline_code"><a href="#blend">blend</a></span>.
Blending involves mixing multiple layers of pixels using the specified equations. Surrounding pixels are not
involved in blend operations. See <span class="inline_code"><a href="#blend">bvbltparams.op.blend</a></span>
for details.<br />
<br />
</p>
</td>
</tr>
<tr>
<td valign="top">3.</td>
<td><span class="inline_code"><strong><a name="BVFLAG_FILTER">BVFLAG_FILTER</a></strong></span><br />
<br />
This flag indicates the operation being performed is a filter, and the <span class="inline_code"><a href="#op">bvbltparams.op</a></span>
union is treated as <span class="inline_code"><a href="#filter">filter</a></span>. Filtering involves mixing
multiple layers of pixels. Surrounding pixels are involved in filter operations. See
<span class="inline_code"><a href="#filter">bvbltparams.op.filter</a></span> for details.<br />
</td>
</tr>
</table>
<p class="Code_Header_3"><a name="BVFLAG_KEY_SRC">bvbltparams.flags - BVFLAG_KEY_SRC</a>/<a name="BVFLAG_KEY_DST">DST</a></p>
<p>The <span class="inline_code">BVFLAG_KEY_SRC</span> and <span class="inline_code">BVFLAG_KEY_DST</span> enable source
and destination color keying, respectively. When either flag is set, the <span class="inline_code">
<a href="#colorkey">colorkey</a></span> member of <span class="inline_code"><a href="#bvbltparams">bvbltparams</a></span>
is used.</p>
<p><span class="inline_code">BVFLAG_KEY_SRC</span> and <span class="inline_code">BVFLAG_KEY_DST</span> are mutually exclusive.</p>
<p>See <span class="inline_code"><a href="#colorkey">bvbltparams.colorkey</a></span> for details.</p>
<p class="Code_Header_3"><a name="BVFLAG_CLIP">bvbltparams.flags - BVFLAG_CLIP</a></p>
<p>When <span class="inline_code">BVFLAG_CLIP</span> is set, the <span class="inline_code"><a href="#cliprect">cliprect</a></span>
member of <span class="inline_code"><a href="#bvbltparams">bvbltparams</a></span> is used by the implementation as a limiting
rectangle on data written to the destination. See <span class="inline_code"><a href="#cliprect">cliprect</a></span>
for details.</p>
<p class="Code_Header_3"><a name="BVFLAG_SRCMASK">bvbltparams.flags - BVFLAG_SRCMASK</a></p>
<p>Normally, the mask is applied at the destination, after all scaling has been completed (including scaling the mask if
necessary). But some environments require that the mask be applied at the sources, before scaling occurs. The
<span class="inline_code">BVFLAG_SRCMASK</span> flag requests that the implementation use this method if supported.</p>
<p class="Code_Header_3">bvbltparams.flags - BVFLAG_TILE_*</p>
<p>Normally, when a source's size does not match the destination, the source is scaled to fill the destination. But
when the corresponding <span class="inline_code">BVFLAG_TILE_*</span> flag is set, this behavior is modified.</p>
<p>First, the source's size specifies a tile (or pattern, or brush) to be used to fill the destination. This tile
is replicated instead of scaled.</p>
<p>The origin of the source's rectangle is used to locate the tile within a larger surface. </p>
<p>Second, a <span class="inline_code"><a href="#bvbuffdesc">bvbuffdesc</a></span> object is no longer supplied by the client
in the bvbltparams structure. In its place is a <span class="inline_code"><a href="#bvtileparams">bvtileparams</a></span>
object.</p>
<p>Refer to the <span class="inline_code"><a href="#bvtileparams">bvtileparams</a></span> structure definition for details.</p>
<p class="Code_Header_3">bvbltparams.flags - <a name="BVFLAG_HORZ_FLIP">BVFLAG_HORZ</a>/<a name="BVFLAG_VERT_FLIP">VERT_FLIP_*</a></p>
<p>These flags indicate that the corresponding image is flipped horizontally or vertically as it is used by the operation.</p>
<p class="Code_Header_3">bvbltparams.flags - BVFLAG_SCALE/DITHER_RETURN</p>
<p>The scale and dither types can be specified with an implicit type. The implementation will then convert that internally
to an explicit scale or dither type. These flags request that the implementation return the explicit type chosen to
the client in the corresponding <span class="inline_code"><a href="#scalemode">bvbltparams.scalemode</a></span> and
<span class="inline_code"><a href="#dithermode">bvbltparams.dithermode</a></span> members.</p>
<p class="Code_Header_3">bvbltparams.flags - BVFLAG_ASYNC</p>
<p>This flag allows the client to inform the implementation that it can queue the requested BLT and return from
<span class="inline_code"><a href="#bv_blt">bv_blt()</a></span> before it has completed. If this bit is not set, when
the <span class="inline_code"><a href="#bv_blt">bv_blt()</a></span> returns, the operation is complete.</p>
<p>Normally, a client will also utilize the <span class="inline_code"><a href="#callbackfn">bvbltparams.callbackfn</a></span>
and <span class="inline_code"><a href="#callbackdata">bvbltparams.callbackdata</a></span> members to receive a notification
when the BLT has completed.</p>
<p class="note">NOTE: Asynchronous BLTs are performed in the order in which they are submitted within an implementation.
This was done to provide a simple dependency mechanism.
However, synchronization between implementations must be handled by the client, using the callback mechanism.</p>
<p class="note">NOTE: Since asynchronous BLTs are performed in the order in which they are submitted, it follows
that a synchronized BLT after a set of asynchronous BLTs may be used as synchronization as well.</p>
<p class="note"><a name="NOP">NOTE</a>: Certain situations may require manual synchronization without an associated BLT.
Rather than introduce an additional BLTsville function call, the method of handling this will be via a NOP BLT. To
accomplish a NOP BLT, the client should issue a BLT using the <span class="inline_code"><a href="#rop">
bvbltparams.op.rop</a></span> code of <span class="inline_code">0xAAAA</span> (copy destination to destination), and
with the <span class="inline_code">BVFLAG_ASYNC</span> flag <span class="underline">not</span> set. Alternatively, the NOP BLT may set the
<span class="inline_code">BVFLAG_ASYNC</span> and provide a <span class="inline_code"><a href="#callbackfn">
bvbltparams.callbackfn</a></span>. <em>To facilitate implementations, a valid destination surface should be
specified.</em></p>
<p class="imponly"><strong>Implementations Only<br />
<br />
</strong>In general, this BLTsville specification has avoided placing any requirement on implementations for specific
operations. However, in support of this special case, support for these NOP BLTs will need to be an implementation
<span class="underline"><strong>requirement</strong></span>. </p>
<p class="Code_Header_3">bvbltparams.flags - BVFLAG_BATCH_BEGIN/CONTINUE/END</p>
<p>These flags are used to control batching of BLTs for two main reasons:</p>
<ol>
<li>To group small, similar BLTs to consolidate overhead. For example, the BLTs associated with rendering each
character in a word.</li>
<li>To group related BLTs, which may allow an implementation to perform a more efficient, but equivalent set of operations.</li>
</ol>
<p>See <a href="#batching">Batching</a> for details.</p>
<p class="Code_Header_3">bvbltparams.flags - <a name="BVFLAG_SRC2_AUXDSTRECT">BVFLAG_SRC2</a>/<a name="BVFLAG_MASK_AUXDSTRECT">MASK_AUXDSTRECT</a></p>
<p>These flags are used to indicate that the bvbltparams.src2auxdstrect and bvbltparams.maskauxdstrect are to be used.
See these entries below for details. These flags are likely to be ignored except for the special case explained below,
so they should be used only when necessary.</p>
<p class="Code_Header_2"><a name="rop">bvbltparams.op.rop</a></p>
<p class="code_block">unsigned short op; /* input */ </p>
<p>When <span class="inline_code"><a href="#BVFLAG_ROP">BVFLAG_ROP</a></span> is set in the <span class="inline_code">
<a href="#flags">bvbltparams.flags</a></span> member, the <span class="inline_code"><a href="#op">bvbltparams.op</a></span>
union is treated as <span class="inline_code">rop</span>. Raster OPerations are binary operations performed on the
bits of the inputs:</p>
<ul>
<li>ROP1s have one source: the destination. Two bits are sufficient to specify the four possible (2<sup>2</sup>)
ROP1 operations.</li>
<li>ROP2s have two sources: the destination and a source. Four bits are used to specify the sixteen (2<sup>2+2</sup>)
ROP2 operations.</li>
<li>ROP3s have three sources: the destination, a source (source 1), and a pattern (a.k.a. brush), which we call
source 2 in BLTsville. Eight bits are used to specify the 256 (2<sup>2+2+2</sup>) ROP3 operations.</li>
<li>ROP4s have four sources: the destination, two sources, and a mask. Sixteen bits are used to specify
the 65,536 (2<sup>2+2+2+2</sup>) ROP4 operations.</li>
</ul>
<p>BLTsville's <span class="inline_code">rop</span> element is used to specify a ROP4, but anything from ROP1 up to ROP4
can be defined using this member:</p>
<ul>
<li>To specify an 8-bit ROP3 as a 16-bit ROP4, replicate the 8 bits twice: 0x2323.</li>
<li>To specify a 4-bit ROP2 as a 16-bit ROP4, replicate the 4 bits four times: 0x2222.</li>
<li>To specify a 2-bit ROP1 as a 16-bit ROP4, replicate the 2 bits eight times: 0x5555.</li>
</ul>
<p class="note">NOTE:
By far the most common ROP used will be 0xCCCC, which indicates a simple copy from source 1 to the destination.</p>
<p>The table below is the magic decoder ring: </p>
<table class="indent">
<tr>
<td>Mask</td>
<td class="ctr"> 1 </td>
<td class="ctr"> 1 </td>
<td class="ctr"> 1 </td>
<td class="ctr"> 1 </td>
<td class="ctr"> 1 </td>
<td class="ctr"> 1 </td>
<td class="ctr"> 1 </td>
<td class="ctr"> 1 </td>
<td class="ctr"> 0 </td>
<td class="ctr"> 0 </td>
<td class="ctr"> 0 </td>
<td class="ctr"> 0 </td>
<td class="ctr"> 0 </td>
<td class="ctr"> 0 </td>
<td class="ctr"> 0 </td>
<td class="ctr"> 0 </td>
</tr>
<tr>
<td class="red_left">Source 2 </td>
<td class="red_center"> 1 </td>
<td class="red_center"> 1 </td>
<td class="red_center"> 1 </td>
<td class="red_center"> 1 </td>
<td class="red_center"> 0 </td>
<td class="red_center"> 0 </td>
<td class="red_center"> 0 </td>
<td class="red_center"> 0 </td>
<td class="red_center"> 1 </td>
<td class="red_center"> 1 </td>
<td class="red_center"> 1 </td>
<td class="red_center"> 1 </td>
<td class="red_center"> 0 </td>
<td class="red_center"> 0 </td>
<td class="red_center"> 0 </td>
<td class="red_center"> 0 </td>
</tr>
<tr>
<td class="grn_left">Source 1 </td>
<td class="grn_center"> 1 </td>
<td class="grn_center"> 1 </td>
<td class="grn_center"> 0 </td>
<td class="grn_center"> 0 </td>
<td class="grn_center"> 1 </td>
<td class="grn_center"> 1 </td>
<td class="grn_center"> 0 </td>
<td class="grn_center"> 0 </td>
<td class="grn_center"> 1 </td>
<td class="grn_center"> 1 </td>
<td class="grn_center"> 0 </td>
<td class="grn_center"> 0 </td>
<td class="grn_center"> 1 </td>
<td class="grn_center"> 1 </td>
<td class="grn_center"> 0 </td>
<td class="grn_center"> 0 </td>
</tr>
<tr>
<td class="blue_left_botbord">Destination </td>
<td class="blue_center_botbord"> 1 </td>
<td class="blue_center_botbord"> 0 </td>
<td class="blue_center_botbord"> 1 </td>
<td class="blue_center_botbord"> 0 </td>
<td class="blue_center_botbord"> 1 </td>
<td class="blue_center_botbord"> 0 </td>
<td class="blue_center_botbord"> 1 </td>
<td class="blue_center_botbord"> 0 </td>
<td class="blue_center_botbord"> 1 </td>
<td class="blue_center_botbord"> 0 </td>
<td class="blue_center_botbord"> 1 </td>
<td class="blue_center_botbord"> 0 </td>
<td class="blue_center_botbord"> 1 </td>
<td class="blue_center_botbord"> 0 </td>
<td class="blue_center_botbord"> 1 </td>
<td class="blue_center_botbord"> 0 </td>
</tr>
<tr>
<td class="left_topbord">Raster Operation </td>
<td class="center_topbord"> 15 </td>
<td class="center_topbord"> 14 </td>
<td class="center_topbord"> 13 </td>
<td class="center_topbord"> 12 </td>
<td class="center_topbord"> 11 </td>
<td class="center_topbord"> 10 </td>
<td class="center_topbord"> 9 </td>
<td class="center_topbord"> 8 </td>
<td class="center_topbord"> 7 </td>
<td class="center_topbord"> 6 </td>
<td class="center_topbord"> 5 </td>
<td class="center_topbord"> 4 </td>
<td class="center_topbord"> 3 </td>
<td class="center_topbord"> 2 </td>
<td class="center_topbord"> 1 </td>
<td class="center_topbord"> 0 </td>
</tr>
</table>
<br />
For example, to specify an operation that uses the mask to choose between source 1 and destination (source 1 when mask is
1, destination when mask is 0), a client would calculate the bottom line by parsing each column:<br />
<br />
When mask is 1 (the first eight columns), the <span class="inline_code">rop</span> matches the source 1 row. When
mask is 0 (the last eight columns), the <span class="inline_code">rop</span> matches the destination row.<br />
<br />
<table class="indent">
<tr>
<td class="left_topbord">Raster Operation </td>
<td class="red_center_topbord"> 1 </td>
<td class="red_center_topbord"> 1 </td>
<td class="red_center_topbord"> 1 </td>
<td class="red_center_topbord"> 1 </td>
<td class="red_center_topbord"> 0 </td>
<td class="red_center_topbord"> 0 </td>
<td class="red_center_topbord"> 0 </td>
<td class="red_center_topbord"> 0 </td>
<td class="blu_center_topbord"> 1 </td>
<td class="blu_center_topbord"> 0 </td>
<td class="blu_center_topbord"> 1 </td>
<td class="blu_center_topbord"> 0 </td>
<td class="blu_center_topbord"> 1 </td>
<td class="blu_center_topbord"> 0 </td>
<td class="blu_center_topbord"> 1 </td>
<td class="blu_center_topbord"> 0 </td>
</tr>
</table>
<br />
So the <span class="inline_code">rop</span> for this operation would be 0xF0AA.<br />
<br />
Here is a list of some commonly used raster operations that have been given names:<br />
<br />
<table class="indent_thick_bord">
<tr>
<td class="thin_bord_dbl_botbord"><strong>ROP </strong></td>
<td class="thin_bord_dbl_botbord"><strong>Constant</strong></td>
<td class="thin_bord_dbl_botbord"><strong>Description</strong></td>
</tr>
<tr>
<td class="thin_bord">BLACKNESS</td>
<td class="thin_bord">0x0000</td>
<td class="thin_bord">Set all destination bits to black (0). Dest = 0</td>
</tr>
<tr>
<td class="thin_bord">NOTSRCERASE</td>
<td class="thin_bord">0x1111</td>
<td class="thin_bord">Dest = ~Src1 & ~Dest = ~(Src1 | Dest)</td>
</tr>
<tr>
<td class="thin_bord">NOTSRCCOPY</td>
<td class="thin_bord">0x3333</td>
<td class="thin_bord">Dest = ~Src1</td>
</tr>
<tr>
<td class="thin_bord">SRCERASE</td>
<td class="thin_bord">0x4444</td>
<td class="thin_bord">Dest = Src1 & ~Dest</td>
</tr>
<tr>
<td class="thin_bord">DSTINVERT</td>
<td class="thin_bord">0x5555</td>
<td class="thin_bord">Invert (NOT) the destination bits. Dest = ~Dest</td>
</tr>
<tr>
<td class="thin_bord">PATINVERT</td>
<td class="thin_bord">0x5A5A</td>
<td class="thin_bord">XOR with Src2. Dest = Src2 ^ Dest</td>
</tr>
<tr>
<td class="thin_bord">SRCINVERT</td>
<td class="thin_bord">0x6666</td>
<td class="thin_bord">XOR with Src1. Dest = Src1 ^ Dest</td>
</tr>
<tr>
<td class="thin_bord">SRCAND</td>
<td class="thin_bord">0x8888</td>
<td class="thin_bord">Dest = Src1 & Dest</td>
</tr>
<tr>
<td class="thin_bord">NOP</td>
<td class="thin_bord">0xAAAA</td>
<td class="thin_bord">Dest = Dest</td>
</tr>
<tr>
<td class="thin_bord">MERGEPAINT</td>
<td class="thin_bord">0xBBBB</td>
<td class="thin_bord">Dest = ~Src1 | Dest</td>
</tr>
<tr>
<td class="thin_bord">MERGECOPY</td>
<td class="thin_bord">0xC0C0</td>
<td class="thin_bord">Dest = Src1 & Src2</td>
</tr>
<tr>
<td class="thin_bord">SRCCOPY</td>
<td class="thin_bord">0xCCCC</td>
<td class="thin_bord">Dest = Src1</td>
</tr>
<tr>
<td class="thin_bord">SRCPAINT</td>
<td class="thin_bord">0xEEEE</td>
<td class="thin_bord">OR with Src1. Dest = Src1 | Dest</td>
</tr>
<tr>
<td class="thin_bord">PATCOPY</td>
<td class="thin_bord">0xF0F0</td>
<td class="thin_bord">Copy source 2 to destination. Dest = Src2</td>
</tr>
<tr>
<td class="thin_bord">PATPAINT</td>
<td class="thin_bord">0xFBFB</td>
<td class="thin_bord">Dest = ~Src1 | Src2 | Dest</td>
</tr>
<tr>
<td class="thin_bord">WHITENESS</td>
<td class="thin_bord">0xFFFF</td>
<td class="thin_bord">Set all destination bits to white (1). Dest = 1</td>
</tr>
</table>
<br />
<span class="Code_Header_2"><a name="blend">bvbltparams.op.blend</a></span>
<p class="code_block">enum bvblend blend; /* input */</p>
<p>When <span class="inline_code"><a href="#BVFLAG_BLEND">BVFLAG_BLEND</a></span> is set in the
<span class="inline_code"><a href="#flags">bvbltparams.flags</a></span> member, the <span class="inline_code">
<a href="#op">bvbltparams.op</a></span> union is treated as a <span class="inline_code">blend</span>.</p>
<p>To specify the blend, the client fills in <span class="inline_code">blend</span> with one of the
<span class="inline_code"><a href="#bvblend">bvblend</a></span> values.</p>
<p><span class="inline_code"><a href="#bvblend">bvblend</a></span> is an enumeration assembled from sets of fields.
The values specified may be extended beyond those that are explicitly defined using the definitions in the
<span class="filename">bvblend.h</span> header file.</p>
<p>The first 4 bits are the format. Currently two format groups are defined, but others can be added. The remainder
of the bits are used as defined by the individual format:</p>
<table class="indent">
<tr>
<td valign="top">1.</td>
<td><span class="Code_Header_3">BVBLENDDEF_FORMAT_CLASSIC</span><br />
<br />
The <span class="inline_code">BVBLENDDEF_FORMAT_CLASSIC</span> is meant to handle the classic Porter-Duff equations.
It can also handle the DirectFB blending.<br />
<br />
<span class="inline_code">BVBLENDDEF_FORMAT_CLASSIC</span> is based on the following equations:<br />
<div>
<p class="indent">C<sub>d</sub> = K<sub>1</sub>C<sub>1</sub> + K<sub>2</sub>C<sub>2</sub><br />
A<sub>d</sub> = K<sub>3</sub>A<sub>1</sub> + K<sub>4</sub>A<sub>2</sub></p>
</div>
where:<br />
<div>
<p class="indent">C<sub>d</sub>: destination color<br />
C<sub>1</sub>: source 1 color<br />
C<sub>2</sub>: source 2 color<br />
A<sub>d</sub>: destination alpha<br />
A<sub>1</sub>: source 1 alpha<br />
A<sub>2</sub>: source 2 alpha<br />
K<sub>#</sub>: one of the constants defined using the bitfields below</p>
</div>
The 28 bits for <span class="inline_code">BVBLENDDEF_FORMAT_CLASSIC</span> are divided into 5 sections.<br />
<br />
The most significant 4 bits are modifiers, used to include additional alpha values from global or remote sources.<br />
<br />
[27] The most significant bit indicates that a remote alpha is to be included in the blend. The format of this is
defined by <span class="inline_code"><a href="#maskgeom">bvbltparams.maskgeom.format</a></span>.<br />
<br />
[26] The next bit is reserved.<br />
<br />
[25:24] The next 2 bits are used to indicate that a global alpha is to be included, and what its format is:<br />
<div>
<p class="indent">00: no global included<br />
01: global included; bvbltparams.globalalpha.size8 is used (0 -> 255)<br />
10: this value is reserved<br />
11: global included; bvbltparams.flogalalpha.fp is used (0.0 -> 1.0) </p>
</div>
The remaining bits are divided into 4 sections, one to define each of the constants:<br />
<br />
[23:18] - K1<br />
[17:12] - K2<br />
[11:6] - K3<br />
[5:0] - K4<br />
<br />
The format is the same for all 4 constant fields:<br />
<br />
[5:4] The first 2 bits of each field indicates the way in which the other 2 fields are interpreted:<br />
<div>
<p class="indent">00: only As: the other two fields contain only As; there should be only one valid A value
between the two fields<br />
01: minimum: the value of the constant is the minimum of the two fields<br />
10: maximum: the value of the constant is the maximum of the two fields<br />
11: only Cs: the other two fields contain only Cs; there should be only one valid C value between the two fields</p>
</div>
[3:2] The middle 2 bits of each field contain the inverse field:<br />
<div>
<p class="indent">00: 1-C1 ("don't care" for "only As")<br />
01: 1-A1 ("don't care" for "only Cs")<br />
10: 1-C2 ("don't care" for "only As")<br />
11: 1-A2 ("don't care" for "only Cs")</p>
</div>
[1:0] The last 2 bits if each field contain the normal field:<br />
<div>
<p class="indent">00: C1 ("don't care" for "only As")<br />
01: A1 ("don't care" for "only Cs")<br />
10: C2 ("don't care" for "only As")<br />
11: A2 ("don't care" for "only Cs")</p>
</div>
EXCEPTIONS:<br />
<br />
00 00 00 - The value 00 00 00, which normally would indicate "only As" with two "don't care" fields, is interpreted
as a constant of 0.<br />
<br />
11 11 11 - The value 11 11 11, which normally would indicate "only Cs" with two "don't care" fields, is interpreted
as a constant of 1.<br />
<br />
<span class="Header4">Constants</span><br />
<br />
Put together, these can define portions of the blend equations that can be put together in a variety of ways:<br />
<br />
<table class="indent_thick_bord">
<tr>
<td class="rt_thick_bord">
<table>
<tr>
<td class="thin_bord">00 00 00</td>
<td class="thin_bord">undefined -> zero</td>
</tr>
<tr>
<td class="thin_bord">00 00 01</td>
<td class="thin_bord">A1 (preferred)</td>
</tr>
<tr>
<td class="thin_bord">00 00 10</td>
<td class="thin_bord">undefined</td>
</tr>
<tr>
<td class="thin_bord">00 00 11</td>
<td class="thin_bord">A2 (preferred)</td>
</tr>
<tr>
<td class="thin_bord">00 01 00</td>
<td class="thin_bord">1-A1 (preferred)</td>
</tr>
<tr>
<td class="thin_bord">00 01 01</td>
<td class="thin_bord">undefined</td>
</tr>
<tr>
<td class="thin_bord">00 01 10</td>
<td class="thin_bord">1-A1 (use 00 01 00)</td>
</tr>
<tr>
<td class="thin_bord">00 01 11</td>
<td class="thin_bord">undefined</td>
</tr>
<tr>
<td class="thin_bord">00 10 00</td>
<td class="thin_bord">undefined</td>
</tr>
<tr>
<td class="thin_bord">00 10 01</td>
<td class="thin_bord">A1 (use 00 00 01)</td>
</tr>
<tr>
<td class="thin_bord">00 10 10</td>
<td class="thin_bord">undefined</td>
</tr>
<tr>
<td class="thin_bord">00 10 11</td>
<td class="thin_bord">A2 (use 00 00 11)</td>
</tr>
<tr>
<td class="thin_bord">00 11 00</td>
<td class="thin_bord">1-A2 (preferred)</td>
</tr>
<tr>
<td class="thin_bord">00 11 01</td>
<td class="thin_bord">undefined</td>
</tr>
<tr>
<td class="thin_bord">00 11 10</td>
<td class="thin_bord">1-A2 (use 00 11 00)</td>
</tr>
<tr>
<td class="thin_bord">00 11 11</td>
<td class="thin_bord">undefined</td>
</tr>
</table>
</td>
<td class="rt_thick_bord">
<table>
<tr>
<td class="thin_bord">01 00 00</td>
<td class="thin_bord">min(C1,1-C1)</td>
</tr>
<tr>
<td class="thin_bord">01 00 01</td>
<td class="thin_bord">min(A1,1-C1)</td>
</tr>
<tr>
<td class="thin_bord">01 00 10</td>
<td class="thin_bord">min(C2,1-C1)</td>
</tr>
<tr>
<td class="thin_bord">01 00 11</td>
<td class="thin_bord">min(A2,1-C1)</td>
</tr>
<tr>
<td class="thin_bord">01 01 00</td>
<td class="thin_bord">min(C1,1-A1)</td>
</tr>
<tr>
<td class="thin_bord">01 01 01</td>
<td class="thin_bord">min(A1,1-A1)</td>
</tr>
<tr>
<td class="thin_bord">01 01 10</td>
<td class="thin_bord">min(C2,1-A1)</td>
</tr>
<tr>
<td class="thin_bord">01 01 11</td>
<td class="thin_bord">min(A2,1-A1)</td>
</tr>
<tr>
<td class="thin_bord">01 10 00</td>
<td class="thin_bord">min(C1,1-C2)</td>
</tr>
<tr>
<td class="thin_bord">01 10 01</td>
<td class="thin_bord">min(A1,1-C2)</td>
</tr>
<tr>
<td class="thin_bord">01 10 10</td>
<td class="thin_bord">min(C2,1-C2)</td>
</tr>
<tr>
<td class="thin_bord">01 10 11</td>
<td class="thin_bord">min(A2,1-C2)</td>
</tr>
<tr>
<td class="thin_bord">01 11 00</td>
<td class="thin_bord">min(C1,1-A2)</td>
</tr>
<tr>
<td class="thin_bord">01 11 01</td>
<td class="thin_bord">min(A1,1-A2)</td>
</tr>
<tr>
<td class="thin_bord">01 11 10</td>
<td class="thin_bord">min(C2,1-A2)</td>
</tr>
<tr>
<td class="thin_bord">01 11 11</td>
<td class="thin_bord">min(A2,1-A2)</td>
</tr>
</table>
</td>
<td class="rt_thick_bord">
<table>
<tr>
<td class="thin_bord">10 00 00</td>
<td class="thin_bord">max(C1,1-C1)</td>
</tr>
<tr>
<td class="thin_bord">10 00 01</td>
<td class="thin_bord">max(A1,1-C1)</td>
</tr>
<tr>
<td class="thin_bord">10 00 10</td>
<td class="thin_bord">max(C2,1-C1)</td>
</tr>
<tr>
<td class="thin_bord">10 00 11</td>
<td class="thin_bord">max(A2,1-C1)</td>
</tr>
<tr>
<td class="thin_bord">10 01 00</td>
<td class="thin_bord">max(C1,1-A1)</td>
</tr>
<tr>
<td class="thin_bord">10 01 01</td>
<td class="thin_bord">max(A1,1-A1)</td>
</tr>
<tr>
<td class="thin_bord">10 01 10</td>
<td class="thin_bord">max(C2,1-A1)</td>
</tr>
<tr>
<td class="thin_bord">10 01 11</td>
<td class="thin_bord">max(A2,1-A1)</td>
</tr>
<tr>
<td class="thin_bord">10 10 00</td>
<td class="thin_bord">max(C1,1-C2)</td>
</tr>
<tr>
<td class="thin_bord">10 10 01</td>
<td class="thin_bord">max(A1,1-C2)</td>
</tr>
<tr>
<td class="thin_bord">10 10 10</td>
<td class="thin_bord">max(C2,1-C2)</td>
</tr>
<tr>
<td class="thin_bord">10 10 11</td>
<td class="thin_bord">max(A2,1-C2)</td>
</tr>
<tr>
<td class="thin_bord">10 11 00</td>
<td class="thin_bord">max(C1,1-A2)</td>
</tr>
<tr>
<td class="thin_bord">10 11 01</td>
<td class="thin_bord">max(A1,1-A2)</td>
</tr>
<tr>
<td class="thin_bord">10 11 10</td>
<td class="thin_bord">max(C2,1-A2)</td>
</tr>
<tr>
<td class="thin_bord">10 11 11</td>
<td class="thin_bord">max(A2,1-A2)</td>
</tr>
</table>
</td>
<td>
<table>
<tr>
<td class="thin_bord">11 00 00</td>
<td class="thin_bord">undefined</td>
</tr>
<tr>
<td class="thin_bord">11 00 01</td>
<td class="thin_bord">1-C1 (use 11 00 11)</td>
</tr>
<tr>
<td class="thin_bord">11 00 10</td>
<td class="thin_bord">undefined</td>
</tr>
<tr>
<td class="thin_bord">11 00 11</td>
<td class="thin_bord">1-C1 (preferred)</td>
</tr>
<tr>
<td class="thin_bord">11 01 00</td>
<td class="thin_bord">C1 (use 11 11 00)</td>
</tr>
<tr>
<td class="thin_bord">11 01 01</td>
<td class="thin_bord">undefined</td>
</tr>
<tr>
<td class="thin_bord">11 01 10</td>
<td class="thin_bord">C2 (use 11 11 10)</td>
</tr>
<tr>
<td class="thin_bord">11 01 11</td>
<td class="thin_bord">undefined</td>
</tr>
<tr>
<td class="thin_bord">11 10 00</td>
<td class="thin_bord">undefined</td>
</tr>
<tr>
<td class="thin_bord">11 10 01</td>
<td class="thin_bord">1-C2 (use 11 10 11)</td>
</tr>
<tr>
<td class="thin_bord">11 10 10</td>
<td class="thin_bord">undefined</td>
</tr>
<tr>
<td class="thin_bord">11 10 11</td>
<td class="thin_bord">1-C2 (preferred)</td>
</tr>
<tr>
<td class="thin_bord">11 11 00</td>
<td class="thin_bord">C1 (preferred)</td>
</tr>
<tr>
<td class="thin_bord">11 11 01</td>
<td class="thin_bord">undefined</td>
</tr>
<tr>
<td class="thin_bord">11 11 10</td>
<td class="thin_bord">C2 (preferred)</td>
</tr>
<tr>
<td class="thin_bord">11 11 11</td>
<td class="thin_bord">undefined -> one</td>
</tr>
</table>
</td>
</tr>
</table>
<span class="Header4"><br />
DirectFB Example</span><br />
<br />
Putting these together into the proper constants, the blending equations can be built for different APIs.
Here is how DirectFB would be mapped:<br />
<br />
For DirectFB, the
<a href="http://directfb.org/docs/DirectFB_Reference_1_2/IDirectFBSurface_SetSrcBlendFunction.html" class="inline_code">
SetSrcBlendFunction()</a> and
<a href="http://directfb.org/docs/DirectFB_Reference_1_2/IDirectFBSurface_SetDstBlendFunction.html" class="inline_code">
SetDstBlendFunction()</a> can specify 121 combinations of blends (11 x 11). It's impractical to specify these combinations
individually. Instead, the settings indicated by each call should be bitwise OR'd to make the proper single value
used in BLTsville.<br />
<br />
<table class="code_block">
<tr>
<td class="ctr"> </td>
<td colspan="5" class="ctr"><strong>32-bit Binary Value</strong></td>
</tr>
<tr>
<td><strong>
<a href="http://directfb.org/docs/DirectFB_Reference_1_2/IDirectFBSurface_SetSrcBlendFunction.html">SetSrcBlendFunction()</a></strong></td>
<td class="ctr"><strong>[VendorID]</strong></td>
<td class="ctr"><strong> [--K1--] </strong></td>
<td class="ctr"><strong> [--K2--] </strong></td>
<td class="ctr"><strong> [--K3--] </strong></td>
<td class="ctr"><strong> [--K4--] </strong></td>
</tr>
<tr>
<td>DSBF_ZERO</td>
<td class="ctr">0000 0000</td>
<td class="ctr">00 00 00</td>
<td class="ctr">xx xx xx</td>
<td class="ctr">00 00 00</td>
<td class="ctr">xx xx xx</td>
</tr>
<tr>
<td>DSBF_ONE</td>
<td class="ctr">0000 0000</td>
<td class="ctr">11 11 11</td>
<td class="ctr">xx xx xx</td>
<td class="ctr">11 11 11</td>
<td class="ctr">xx xx xx</td>
</tr>
<tr>
<td>DSBF_SRCCOLOR</td>
<td class="ctr">0000 0000</td>
<td class="ctr">11 11 00</td>
<td class="ctr">xx xx xx</td>
<td class="ctr">00 00 01</td>
<td class="ctr">xx xx xx</td>
</tr>
<tr>
<td>DSBF_INVSRCCOLOR</td>
<td class="ctr">0000 0000</td>
<td class="ctr">11 00 11</td>
<td class="ctr">xx xx xx</td>
<td class="ctr">00 01 00</td>
<td class="ctr">xx xx xx</td>
</tr>
<tr>
<td>DSBF_SRCALPHA</td>
<td class="ctr">0000 0000</td>
<td class="ctr">00 00 01</td>
<td class="ctr">xx xx xx</td>
<td class="ctr">00 00 01</td>
<td class="ctr">xx xx xx</td>
</tr>
<tr>
<td>DSBF_INVSRCALPHA</td>
<td class="ctr">0000 0000</td>
<td class="ctr">00 01 00</td>
<td class="ctr">xx xx xx</td>
<td class="ctr">00 01 00</td>
<td class="ctr">xx xx xx</td>
</tr>
<tr>
<td>DSBF_DESTCOLOR</td>
<td class="ctr">0000 0000</td>
<td class="ctr">11 11 10</td>
<td class="ctr">xx xx xx</td>
<td class="ctr">00 00 11</td>
<td class="ctr">xx xx xx</td>
</tr>
<tr>
<td>DSBF_INVDESTCOLOR</td>
<td class="ctr">0000 0000</td>
<td class="ctr">11 10 11</td>
<td class="ctr">xx xx xx</td>
<td class="ctr">00 11 00</td>
<td class="ctr">xx xx xx</td>
</tr>
<tr>
<td>DSBF_DESTALPHA</td>
<td class="ctr">0000 0000</td>
<td class="ctr">00 00 11</td>
<td class="ctr">xx xx xx</td>
<td class="ctr">00 00 11</td>
<td class="ctr">xx xx xx</td>
</tr>
<tr>
<td>DSBF_INVDESTALPHA</td>
<td class="ctr">0000 0000</td>
<td class="ctr">00 11 00</td>
<td class="ctr">xx xx xx</td>
<td class="ctr">00 11 00</td>
<td class="ctr">xx xx xx</td>
</tr>
<tr>
<td>DSBF_SRCALPHASAT</td>
<td class="ctr">0000 0000</td>
<td class="ctr">01 11 01</td>
<td class="ctr">xx xx xx</td>
<td class="ctr">11 11 11</td>
<td class="ctr">xx xx xx</td>
</tr>
</table>
<br />
<table class="code_block">
<tr>
<td class="ctr"> </td>
<td colspan="5" class="ctr"><strong>32-bit Binary Value</strong></td>
</tr>
<tr>
<td><strong>
<a href="http://directfb.org/docs/DirectFB_Reference_1_2/IDirectFBSurface_SetDstBlendFunction.html">SetDstBlendFunction()</a></strong></td>
<td class="ctr"><strong>[VendorID]</strong></td>
<td class="ctr"><strong> [--K1--] </strong></td>
<td class="ctr"><strong> [--K2--] </strong></td>
<td class="ctr"><strong> [--K3--] </strong></td>
<td class="ctr"><strong> [--K4--] </strong></td>
</tr>
<tr>
<td>DSBF_ZERO</td>
<td class="ctr">0000 0000</td>
<td class="ctr">xx xx xx</td>
<td class="ctr">00 00 00</td>
<td class="ctr">xx xx xx</td>
<td class="ctr">00 00 00</td>
</tr>
<tr>
<td>DSBF_ONE</td>
<td class="ctr">0000 0000</td>
<td class="ctr">xx xx xx</td>
<td class="ctr">11 11 11</td>
<td class="ctr">xx xx xx</td>
<td class="ctr">11 11 11</td>
</tr>
<tr>
<td>DSBF_SRCCOLOR</td>
<td class="ctr">0000 0000</td>
<td class="ctr">xx xx xx</td>
<td class="ctr">11 11 00</td>
<td class="ctr">xx xx xx</td>
<td class="ctr">00 00 01</td>
</tr>
<tr>
<td>etc.</td>
<td class="ctr"> </td>
<td class="ctr"> </td>
<td class="ctr"> </td>
<td class="ctr"> </td>
<td> </td>
</tr>
</table>
<br />
<span class="Header4">Porter-Duff</span><br />
<br />
For Porter-Duff blends, the equations can be more specifically defined. For convenience, these are enumerated in
the <span class="inline_code">bvblend.h</span> header. These enumerations utilize only the local alpha in the equations
as indicated. To use global or remote alpha, these enumerations need to be modified. For example, to include the
global alpha in the Porter-Duff <span class="inline_code">BVBLEND_SRC1OVER</span> blend, the blend could be defined
like this:<br />
<br />
<div>
<table class="indent">
<tr>
<td valign="top"><span class="inline_code">params.op.blend =</span></td>
<td><span class="inline_code">BVBLEND_SRC1OVER +<br />
BVBLENDDEF_GLOBAL_UCHAR;</span></td>
</tr>
</table>
</div>
<br />
To include the remote alpha, the blend could be defined like this:<br />
<br />
<div>
<table class="indent">
<tr>
<td valign="top"><span class="inline_code">params.op.blend =</span></td>
<td><span class="inline_code">BVBLEND_SRC1OVER +<br />
BVBLENDDEF_REMOTE;</span></td>
</tr>
</table>
</div>
<br />
And to include both:<br />
<br />
<div>
<table class="indent">
<tr>
<td valign="top"><span class="inline_code">params.op.blend =</span></td>
<td><span class="inline_code">BVBLEND_SRC1OVER +<br />
BVBLENDDEF_GLOBAL_UCHAR +<br />
BVBLENDDEF_REMOTE;</span></td>
</tr>
</table>
</div>
<br />
Note that if the source color formats include local alphas, the local alphas, global alpha, and remote alpha will
be used together.<br />
<br />
Note also that the equations assume the surfaces are premultiplied. So if the surface formats indicate that they
are not premultiplied, the alpha multiplication of each color is done prior to using the surface values in the equations.<br />
<br />
For example, <span class="inline_code">BVBLEND_SRC1OVER</span> specifies the equations:<br />
<table class="indent">
<tr>
<td>C<sub>d</sub> = C<sub>1</sub> + (1 - A<sub>1</sub>)C<sub>2</sub><br />
A<sub>d</sub> = A<sub>1</sub> + (1 - A<sub>1</sub>)A<sub>2</sub> </td>
</tr>
</table>
<br />
If the format of surface 1 is non-premultiplied, the equations are modified to include the multiplication explicitly:<br />
<br />
<table class="indent">
<tr>
<td>C<sub>d</sub> = A<sub>1</sub>C<sub>1</sub> + (1 - A<sub>1</sub>)C<sub>2</sub><br />
A<sub>d</sub> = A<sub>1</sub> + (1 - A<sub>1</sub>)A<sub>2</sub> </td>
</tr>
</table>
<br />
Likewise, if the format of surface 2 is non-premultiplied, the equations are modified for this:<br />
<br />
<table class="indent">
<tr>
<td>
<div>
C<sub>d</sub> = C<sub>1</sub> + (1 - A<sub>1</sub>)A<sub>2</sub>C<sub>2</sub><br />
A<sub>d</sub> = A<sub>1</sub> + (1 - A<sub>1</sub>)A<sub>2</sub> </div>
</td>
</tr>
</table>
<br />
When including global or remote alphas, these values are used to modify the source 1 value values before being used
in the blend equation:<br />
<br />
<table class="indent">
<tr>
<td class="ctr">C<sub>1</sub> = A<sub>g</sub>C<sub>1</sub><br />
A<sub>1</sub> = A<sub>g</sub>A<sub>1</sub></td>
<td style="width: 20%" class="ctr">-or-</td>
<td class="ctr">C<sub>1</sub> = A<sub>r</sub>C<sub>1</sub><br />
A<sub>1</sub> = A<sub>r</sub>A<sub>1</sub></td>
<td class="ctr">-or-</td>
<td class="ctr">C<sub>1</sub> = A<sub>r</sub>A<sub>g</sub>C<sub>1</sub><br />
A<sub>1</sub> = A<sub>r</sub>A<sub>g</sub>A<sub>1</sub></td>
</tr>
</table>
<br />
</td>
</tr>
<tr>
<td valign="top">2.</td>
<td><span class="Code_Header_3"><strong><a name="BVBLENDDEF_FORMAT_ESSENTIAL0">BVBLENDDEF_FORMAT_ESSENTIAL</a></strong></span><br />
<br />
The essential blending equations are based on the blending equations in common image manipulation programs.<pre class="indent"><code>BVBLEND_LIGHTEN max(src1, src2)
BVBLEND_DARKEN min(src1, src2)
BVBLEND_MULTIPLY (src1 * src2) / 255
BVBLEND_AVERAGE (src1 + src2) / 2
BVBLEND_ADD src1 + src2 (saturated)
BVBLEND_SUBTRACT src1 + src2 - 255 (saturated)
BVBLEND_DIFFERENCE abs(src - src2)
BVBLEND_NEGATION 255 - abs(255 - src1 - src2)
BVBLEND_SCREEN 255 - (((255 - src1) * (255 - src2)) / 256)
BVBLEND_EXCLUSION src1 + src2 - ((2 * src1 * src2) / 255)
BVBLEND_OVERLAY (src2 < 128) ? (2 * src1 * src2 / 255) : (255 - 2 * (255 - src1) * (255 - src2) / 255)
BVBLEND_SOFT_LIGHT (src2 < 128) ? (2 * ((src1 >> 1) + 64)) * ((float)src2 / 255) : (255 - (2 * (255 - ((src1 >> 1) + 64)) * (float)(255 - src2) / 255))
BVBLEND_HARD_LIGHT (src1 < 128) ? (2 * src2 * src1 / 255) : (255 - 2 * (255 - src2) * (255 - src1) / 255)
BVBLEND_COLOR_DODGE (src2 == 255) ? src2 : min(255, ((src1 << 8) / (255 - src2))
BVBLEND_COLOR_BURN (src2 == 0) ? src2 : max(0, (255 - ((255 - src1) << 8 ) / src2))))
BVBLEND_LINEAR_DODGE same as BVBLEND_ADD
BVBLEND_LINEAR_BURN same as BVBLEND_SUBTRACT
BVBLEND_LINEAR_LIGHT (src2 < 128) ? LINEAR_BURN(src1,(2 * src2)) : LINEAR_DODGE(src1,(2 * (src2 - 128)))
BVBLEND_VIVID_LIGHT (src2 < 128) ? COLOR_BURN(src1,(2 * src2)) : COLOR_DODGE(src1,(2 * (src2 - 128))))
BVBLEND_PIN_LIGHT (src2 < 128) ? DARKEN(src1,(2 * src2)) : LIGHTEN(src1,(2 * (src2 - 128)))
BVBLEND_HARD_MIX (VIVID_LIGHT(src1, src2) < 128) ? 0 : 255
BVBLEND_REFLECT (src2 == 255) ? src2 : min(255, (src1 * src1 / (255 - src2)))
BVBLEND_GLOW (src1 == 255) ? src1 : min(255, (src2 * src2 / (255 - src1)))
BVBLEND_PHOENIX min(src1, src2) - max(src1, src2) + 255)
BVBLEND_ALPHA alf * src1 + (1 - alf) * src2)</code></pre>
</td>
</tr>
</table>
<a name="filter" class="Code_Header_2">bvbltparams.op.filter</a>
<p class="code_block">struct bvfilter *filter; /* input */</p>
<p>When <span class="inline_code"><a href="#BVFLAG_FILTER">BVFLAG_FILTER</a></span> is set in the
<span class="inline_code"><a href="#flags">bvbltparams.flags</a></span> member, the <span class="inline_code">
<a href="#op">bvbltparams.op</a></span> union is treated as a <span class="inline_code">filter</span>.</p>
<p>To specify the filter, the client fills in <span class="inline_code">filter</span> with one of the
<span class="inline_code">bvfilter</span> values.</p>
<p>These values will be extended as general filter types are requested.</p>
<a name="colorkey" class="Code_Header_2">bvbltparams.colorkey</a>
<p class="code_block">void *colorkey; /* input */</p>
<p>When either <span class="inline_code"><a href="#BVFLAG_KEY_SRC">BVFLAG_KEY_SRC</a></span> or
<span class="inline_code"><a href="#BVFLAG_KEY_DST">BVFLAG_KEY_DST</a></span> is set in the <span class="inline_code">
<a href="#flags">bvbltparams.flags</a></span> member, <span class="inline_code">colorkey</span> points to a single pixel
used as the color key.</p>
<p>The format of this pixel matches the surface being keyed. i.e. <span class="inline_code"><a href="#bvsurfgeom">
src1geom.format</a></span> is the format of the color key if <span class="inline_code">BVFLAG_KEY_SRC</span> is set, or
<span class="inline_code"><a href="#bvsurfgeom">dst.format</a></span> is the format of the color key if
<span class="inline_code">BVFLAG_KEY_DST</span> is set.</p>
<p><em>Subsampled formats do not currently support color keying.</em></p>
<p class="Code_Header_2"><a name="globalalpha">bvbltparams.globalalpha</a></p>
<p class="code_block">union bvalpha globalalpha; /* input */</p>
<p>When <span class="inline_code"><a href="#BVFLAG_BLEND">BVFLAG_BLEND</a></span> is set in the
<span class="inline_code"><a href="#flags">bvbltparams.flags</a></span>, and when the <span class="inline_code">
<a href="#blend">blend</a></span> chosen requires it, <span class="inline_code">globalalpha</span> is used to provide an
alpha blending value for the entire operation. The type is also dependent on the <span class="inline_code">
<a href="#blend">blend</a></span> chosen.</p>
<p>For the <span class="inline_code">BVBLENDDEF_FORMAT_CLASSIC</span> blend types, if the <span class="inline_code">BVBLENDDEF_GLOBAL_MASK</span>
field is not 0, this field is used. Currently <span class="inline_code">BVBLENDDEF_FORMAT_CLASSIC</span> provides
for an 8-bit (unsigned character / byte) format designated by <span class="inline_code">BVBLENDDEF_GLOBAL_UCHAR</span> as
well as a 32-bit floating point format designated by <span class="inline_code">BVBLENDDEF_GLOBAL_FLOAT</span>.</p>
<p class="Code_Header_2"><a name="scalemode">bvbltparams.scalemode</a></p>
<p class="code_block">enum bvscalemode scalemode; /* input/output */</p>
<p>This member allows the client to specify the type of scaling to be used. The enumeration begins with 8 bits indicating
the vendor. The remaining bits are defined by the vendor. <span class="inline_code">BVSCALEDEF_VENDOR_ALL</span>
and <span class="inline_code">BVSCALEDEF_VENDOR_GENERAL</span> are shared by all implementations.</p>
<p><span class="inline_code">BVSCALEDEF_VENDOR_ALL</span> can be used to specify an implicit scale type. This type
is converted to an explicit type by the implementation:</p>
<table class="indent">
<tr>
<td class="inline_code">BVSCALE_FASTEST</td>
<td>The fastest method of scaling available is used. This may include nearest neighbor. The value of
this enumeration is purposely 0, and is the default scale type. No implementation will return an error for
this setting.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_FASTEST_NOT_NEAREST_NEIGHBOR</td>
<td>The fastest method of scaling available that is not nearest neighbor is used. This may include an alternative
point sample technique.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_FASTEST_POINT_SAMPLE</td>
<td>The fastest method of scaling using a point sample technique.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_FASTEST_INTERPOLATED</td>
<td>The fastest method of scaling using an interpolation technique.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_FASTEST_PHOTO</td>
<td>The fastest method of scaling appropriate for photographs is used. This may include nearest neighbor.
No implementation will return an error for this setting.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_FASTEST_DRAWING</td>
<td>The fastest method of scaling appropriate for drawings is used. This may include nearest neighbor.
No implementation will return an error for this setting.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_GOOD</td>
<td>A scaling technique is chosen that may be higher quality than the <span class="inline_code">BVSCALE_FASTEST</span>
choice. This may include nearest neighbor. No implementation will return an error for this setting.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_GOOD_POINT_SAMPLE</td>
<td>A point sample scaling technique is chosen that may be higher quality than the <span class="inline_code">BVSCALE_FASTEST_POINT_SAMPLE</span>
choice. This may include nearest neighbor.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_GOOD_INTERPOLATED</td>
<td>An interpolated scaling technique is chosen that may be higher quality than the <span class="inline_code">BVSCALE_FASTEST_INTERPOLATED</span>
choice.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_GOOD_PHOTO</td>
<td>A scaling technique appropriate for photographs is chosen that may be higher quality than the
<span class="inline_code">BVSCALE_FASTEST_PHOTO</span> choice. This may include nearest neighbor. No
implementation will return an error for this setting.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_GOOD_DRAWING</td>
<td>A scaling technique appropriate for drawings is chosen that may be higher quality than the
<span class="inline_code">BVSCALE_FASTEST_DRAWING</span> choice. This may include nearest neighbor.
No implementation will return an error for this setting.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_BETTER</td>
<td>A scaling technique is chosen that may be higher quality than the <span class="inline_code">BVSCALE_GOOD</span>
choice. This may include nearest neighbor. No implementation will return an error for this setting.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_BETTER_POINT_SAMPLE</td>
<td>A point sample scaling technique is chosen that may be higher quality than the <span class="inline_code">BVSCALE_GOOD_POINT_SAMPLE</span>
choice. This may include nearest neighbor.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_BETTER_INTERPOLATED</td>
<td>An interpolated scaling technique is chosen that may be higher quality than the <span class="inline_code">BVSCALE_GOOD_INTERPOLATED</span>
choice.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_BETTER_PHOTO</td>
<td>A scaling technique appropriate for photographs is chosen that may be higher quality than the
<span class="inline_code">BVSCALE_GOOD_PHOTO</span> choice. This may include nearest neighbor. No implementation
will return an error for this setting.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_BETTER_DRAWING</td>
<td>A scaling technique appropriate for drawings is chosen that may be higher quality than the
<span class="inline_code">BVSCALE_GOOD_DRAWING</span> choice. This may include nearest neighbor. No
implementation will return an error for this setting.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_BEST</td>
<td>The highest quality scaling technique is chosen. This may include nearest neighbor. No implementation
will return an error for this setting.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_BEST_POINT_SAMPLE</td>
<td>The highest quality point sample technique is chosen.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_BEST_INTERPOLATED</td>
<td>The highest quality interpolated scaling technique is chosen.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_BEST_PHOTO</td>
<td>The highest quality scaling technique appropriate for photographs is chosen. This may include nearest
neighbor. No implementation will return an error for this setting.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_BEST_DRAWING</td>
<td>The highest quality scaling technique appropriate for drawings is chosen. This may include nearest neighbor.
No implementation will return an error for this setting.</td>
</tr>
</table>
<br />
<span class="inline_code">BVSCALEDEF_VENDOR_GENERAL</span> can be used to specify one of the shared explicit scale types.
At this point, only a limited number of explicit scale types are defined: <br />
<br />
<table class="indent">
<tr>
<td class="inline_code">BVSCALE_NEAREST_NEIGHBOR</td>
<td>This is a point sample scaling technique where the resampled destination pixel is set to the value of the closest
source pixel.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_BILINEAR</td>
<td>This is an interpolated scaling technique where the resampled destination pixel is set to a value linearly interpolated
in two dimensions from the four closest source pixels.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_BICUBIC</td>
<td>This is an interpolated scaling technique where the resampled destination pixel is set to a value calculated
using cubic interpolation in two dimensions.</td>
</tr>
<tr>
<td class="inline_code">BVSCALE_3x3_TAP</td>
<td> </td>
</tr>
<tr>
<td class="inline_code">BVSCALE_5x5_TAP</td>
<td> </td>
</tr>
<tr>
<td class="inline_code">BVSCALE_7x7_TAP</td>
<td> </td>
</tr>
<tr>
<td class="inline_code">BVSCALE_9x9_TAP</td>
<td> </td>
</tr>
</table>
<p>If the client wants to know the explicit type chosen by a given implementation, it can set <span class="inline_code">
BVFLAG_SCALE_RETURN</span> in the <span class="inline_code"><a href="#flags">bvbltparams.flags</a></span> member, and the
explicit scale type is returned in the <span class="inline_code">scalemode</span> member.</p>
<p class="note">NOTE: Extending the <span class="inline_code">BVSCALEDEF_VENDOR_GENERAL</span> scale types or obtaining
a vendor ID can be accomplished by submitting a patch.</p>
<p class="Code_Header_2"><a name="dithermode">bvbltparams.dithermode</a></p>
<p class="code_block">enum bvdithermode dithermode; /* input/output */</p>
<p>This member allows the client to specify the type of dithering to be used, when the output format has fewer bits of depth
than the internal calculation. The enumeration begins with 8 bits indicating the vendor. The remaining bits
are defined by the vendor. <span class="inline_code">BVDITHERDEF_VENDOR_ALL</span> and <span class="inline_code">BVDITHERDEF_VENDOR_GENERAL</span>
are shared by all implementations.</p>
<p><span class="inline_code">BVDITHERDEF_VENDOR_ALL</span> can be used to specify an implicit dither type. This type
is converted to an explicit type by the implementation:</p>
<table class="indent">
<tr>
<td class="inline_code">BVDITHER_FASTEST</td>
<td>The fastest method of dithering available is used. This may include no dithering (truncation). The
value of this enumeration is purposely 0, and is the default dither type. No implementation will return an
error for this setting.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_FASTEST_ON</td>
<td>The fastest method of dithering available is used. This will not include no dithering.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_FASTEST_RANDOM</td>
<td>The fastest method of dithering using a random technique.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_FASTEST_ORDERED</td>
<td>The fastest method of dithering using an ordered diffusion technique.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_FASTEST_DIFFUSED</td>
<td>The fastest method of dithering using an error diffusion technique.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_FASTEST_PHOTO</td>
<td>The fastest method of dithering appropriate for photographs is used. This may include no dithering.
No implementation will return an error for this setting.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_FASTEST_DRAWING</td>
<td>The fastest method of dithering appropriate for drawings is used. This may include no dithering.
No implementation will return an error for this setting.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_GOOD</td>
<td>A dithering technique is chosen that may be higher quality than the <span class="inline_code">BVDITHER_FASTEST</span>
choice. This may include no dithering. No implementation will return an error for this setting.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_GOOD_ON</td>
<td>Any dithering technique available is used. This will not include no dithering. This may be higher
quality than <span class="inline_code">BVDITHER_FASTEST_ON</span>.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_GOOD_RANDOM</td>
<td>A random dithering technique is chosen that may be higher quality than the <span class="inline_code">BVDITHER_FASTEST_RANDOM</span>
choice.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_GOOD_ORDERED</td>
<td>An ordered dithering technique is chosen that may be higher quality than the <span class="inline_code">BVDITHER_FASTEST_ORDERED</span>
choice.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_GOOD_DIFFUSED</td>
<td>A diffused dithering technique is chosen that may be higher quality than the <span class="inline_code">BVDITHER_FASTEST_DIFFUSED</span>
choice.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_GOOD_PHOTO</td>
<td>A dithering technique appropriate for photographs is chosen that may be higher quality than the
<span class="inline_code">BVDITHER_FASTEST_PHOTO</span> choice. This may include no dithering. No implementation
will return an error for this setting.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_GOOD_DRAWING</td>
<td>A dithering technique appropriate for drawings is chosen that may be higher quality than the
<span class="inline_code">BVDITHER_FASTEST_DRAWING</span> choice. This may include no dithering. No
implementation will return an error for this setting.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_BETTER</td>
<td>A dithering technique is chosen that may be higher quality than the <span class="inline_code">BVDITHER_GOOD</span>
choice. This may include no dithering. No implementation will return an error for this setting.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_BETTER_ON</td>
<td>Any dithering technique available is used. This will not include no dithering. This may be higher
quality than <span class="inline_code">BVDITHER_GOOD_ON</span>.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_BETTER_RANDOM</td>
<td>A random dithering technique is chosen that may be higher quality than the <span class="inline_code">BVDITHER_GOOD_RANDOM</span>
choice.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_BETTER_ORDERED</td>
<td>An ordered dithering technique is chosen that may be higher quality than the <span class="inline_code">BVDITHER_GOOD_ORDERED</span>
choice.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_BETTER_DIFFUSED</td>
<td>A diffused dithering technique is chosen that may be higher quality than the <span class="inline_code">BVDITHER_GOOD_DIFFUSED</span>
choice.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_BETTER_PHOTO</td>
<td>A scaling technique appropriate for photographs is chosen that may be higher quality than the
<span class="inline_code">BVSCALE_GOOD_PHOTO</span> choice. No implementation will return an error for this
setting.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_BETTER_DRAWING</td>
<td>A scaling technique appropriate for drawings is chosen that may be higher quality than the
<span class="inline_code">BVSCALE_GOOD_DRAWING</span> choice. No implementation will return an error for this
setting.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_BEST</td>
<td>The highest quality dithering technique is chosen. This may include no dithering. No implementation
will return an error for this setting.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_BEST_ON</td>
<td>Any dithering technique available is used. This will not include no dithering. This may be higher
quality than <span class="inline_code">BVDITHER_BEST_ON</span>.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_BEST_RANDOM</td>
<td>The highest quality random dithering technique is chosen.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_BEST_ORDERED</td>
<td>The highest quality ordered dithering technique is chosen.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_BEST_DIFFUSED</td>
<td>The highest quality diffused dithering technique is chosen.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_BEST_PHOTO</td>
<td>The highest quality dithering technique appropriate for photographs is chosen. This may include no dithering.
No implementation will return an error for this setting.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_BEST_DRAWING</td>
<td>The highest quality dithering technique appropriate for drawings is chosen. This may include no dithering.
No implementation will return an error for this setting.</td>
</tr>
</table>
<br />
<span class="inline_code">BVDITHERDEF_VENDOR_GENERAL</span> can be used to specify one of the shared explicit dithering
types. At this point, only a limited number of explicit dither types are defined:<br />
<br />
<table class="indent">
<tr>
<td class="inline_code">BVDITHER_NONE</td>
<td>No dithering is performed. Internal pixel component values are truncated to the destination component
bit depth.</td>
</tr>
<tr>
<td class="inline_code">BVDITHER_ORDERED_2x2</td>
<td> </td>
</tr>
<tr>
<td class="inline_code">BVDITHER_ORDERED_4x4</td>
<td> </td>
</tr>
<tr>
<td class="inline_code">BVDITHER_ORDERED_2x2_4x4</td>
<td>2x2 ordered dither is used for components with the lowest bit reduction. 4x4 ordered dither is used for
the components with the highest bit reduction. (E.g. RGB24 to RGB565 will use 2x2 ordered dither for the green
component and 4x4 ordered dither for the red and blue components.)</td>
</tr>
</table>
<p>If the client wants to know the explicit type chosen by a given implementation, it can set <span class="inline_code">
BVFLAG_DITHER_RETURN</span> in the <span class="inline_code"><a href="#flags">bvbltparams.flags</a></span> member, and the
explicit scale type is returned in the <span class="inline_code">dithermode</span> member.</p>
<p class="note">NOTE: Extending the <span class="inline_code">BVDITHERDEF_VENDOR_GENERAL</span> scale types or obtaining
a vendor ID can be accomplished by submitting a patch.</p>
<p class="Code_Header_2"><a name="dstdesc">bvbltparams.dstdesc</a></p>
<p class="code_block"><a href="#bvbuffdesc">struct bvbuffdesc</a> *dstdesc;</p>
<p><span class="inline_code">dstdesc</span> is used to specify the destination buffer. If the buffer has not been
mapped with a call to <span class="inline_code"><a href="#bv_map">bv_map()</a></span>, <span class="inline_code">
<a href="#bv_blt">bv_blt()</a></span> will map the buffer as necessary to perform the BLT and then unmap afterwards.
See <span class="inline_code"><a href="#bvbuffdesc">bvbuffdesc</a></span> for details.</p>
<p class="Code_Header_2"><a name="dstgeom">bvbltparams.dstgeom</a></p>
<p class="code_block"><a href="#bvsurfgeom">struct bvsurfgeom</a> *dstgeom;</p>
<p><span class="inline_code">dstgeom</span> is used to specify the geometry of the surface contained in the destination
buffer. See <span class="inline_code"><a href="#bvsurfgeom">bvsurfgeom</a></span> for details.</p>
<p class="Code_Header_2"><a name="dstrect">bvbltparams.dstrect</a></p>
<p class="code_block"><a href="#bvrect">struct bvrect</a> dstrect;</p>
<p><span class="inline_code">dstrect</span> is used to specify the destination rectangle to receive the BLT. This
rectangle is clipped by <span class="inline_code"><a href="#cliprect">bvbltparams.cliprect</a></span> when
<span class="inline_code"><a href="#BVFLAG_CLIP">BVFLAG_CLIP</a></span> is set in the <span class="inline_code">
<a href="#flags">bvbltparams.flags</a></span> member.</p>
<p class="Code_Header_2">bvbltparams.<a name="src1.desc">src1</a>/<a name="src2.desc">src2</a>/<a name="mask.desc">mask.desc</a></p>
<p class="code_block"><a href="#bvbuffdesc">struct bvbuffdesc</a> *src1.desc;<br />
<a href="#bvbuffdesc">struct bvbuffdesc</a> *src2.desc;<br />
<a href="#bvbuffdesc">struct bvbuffdesc</a> *mask.desc;</p>
<p>These members are used to identify the buffer for the source1, source2, and mask surfaces when the associated
<span class="inline_code">BVFLAG_TILE_*</span> flag is not set. The buffer is the memory in which the surface lies.
See the <span class="inline_code"><a href="#src1geom">bvbltparams.src1/src2/maskgeom</a></span> for the format and layout/geometry
of the surface.</p>
<p class="note">NOTE WELL: Clients should never change the value of a <span class="inline_code">
<a href="#bvbuffdesc">bvbuffdesc</a></span> structure while a buffer is mapped.</p>
<p class="Code_Header_2">bvbltparams.<a name="src1.tileparams">src1</a>/<a name="src2.tileparams">src2</a>/<a name="mask.tileparams">mask.tileparams</a></p>
<p class="code_block"><a href="#bvtileparams">struct bvtileparams</a> *src1.tileparams;<br />
<a href="#bvtileparams">struct bvtileparams</a> *src2.tileparams;<br />
<a href="#bvtileparams">struct bvtileparams</a> *mask.tileparams;</p>
<p>These members are used to identify the buffer for the source1, source2, and mask surfaces when the associated
<span class="inline_code">BVFLAG_TILE_*</span> flag is set. The buffer is the memory in which the surface lies.
This differs from the <span class="inline_code"><a href="#src1.desc">src1/src2/mask.desc</a></span> identity by providing
more information needed for tiling and by not requiring mapping (for hardware implementations that support tiling, the tile
data is usually moved into an on-chip cache).</p>
<p class="Code_Header_2">bvbltparams.<a name="src1geom">src1</a>/<a name="src2geom">src2</a>/<a name="maskgeom">maskgeom</a></p>
<p class="code_block"><a href="#bvsurfgeom">struct bvsurfgeom</a> src1geom;<br />
<a href="#bvsurfgeom">struct bvsurfgeom</a> src2geom;<br />
<a href="#bvsurfgeom">struct bvsurfgeom</a> maskgeom;</p>
<p>These members describe the format and layout/geometry of their respective surfaces. Separating
<span class="inline_code"><a href="#bvsurfgeom">bvsurfgeom</a></span> from the <span class="inline_code">
<a href="#bvbuffdesc">bvbuffdesc</a></span> allows easy use of buffers for multiple geometries without remapping.
See <span class="inline_code"><a href="#bvsurfgeom">bvsurfgeom</a></span> and <span class="inline_code">
<a href="#bvbuffdesc">bvbuffdesc</a></span> for details.</p>
<p class="Code_Header_2">bbvbltparams.src1/src2/maskrect</p>
<p class="code_block"><a href="#bvrect">struct bvrect</a> src1rect;<br />
<a href="#bvrect">struct bvrect</a> src2rect;<br />
<a href="#bvrect">struct bvrect</a> maskrect;</p>
<p>These members specify the rectangle from which data is read for the BLT. These rectangles are clipped by a scaled
version of the <span class="inline_code"><a href="#cliprect">bvbltparams.cliprect</a></span> (scaling is based on
the relationship between them and the <span class="inline_code"><a href="#dstrect">bvbltparams.dstrect</a></span>) when
<span class="inline_code"><a href="#BVFLAG_CLIP">BVFLAG_CLIP</a></span> is set in the <span class="inline_code">
<a href="#flags">bvbltparams.flags</a></span> member.</p>
<table class="example">
<tr>
<td>
<p><strong>Example:</strong></p>
<a href="#src1rect" class="inline_code">src1rect</a> = (0, 0) - (400 x 200)<br />
<a href="#dstrect" class="inline_code">dstrect</a> = (0, 0) - (800 x 600)<br />
<a href="#cliprect" class="inline_code">cliprect</a> = (10, 30) - (300 x 300)<p>The scaling ratio of the
<a href="#dstrect" class="inline_code">dstrect</a> to the <a href="#src1rect" class="inline_code">src1rect</a> is
(800/400, 600/300) or (2, 3). Using this, the effective source 1 clipping rectangle becomes (10/2, 30/3)
- (300/2 x 300/3) or (5, 10) - (150 x 100).</p>
</td>
</tr>
</table>
<p>This approach allows fractional clipping at the source using a method which is simpler to implement than fractional coordinates.</p>
<p class="note">NOTE: In BLTsville, reading outside the source rectangle is forbidden. So scaling algorithms
which require pixels around a particular source pixel must utilize boundary techniques (mirror, repeat, clamp, etc.) at
the edges of the source rectangle. However, if the clipping rectangle, when translated back to the source rectangle,
leaves space between it and the source rectangle, pixels outside the clipped region may be accessed by the implementation.</p>
<p class="Code_Header_2"><a name="cliprect">bvbltparams.cliprect</a></p>
<p class="code_block"><a href="#bvrect">struct bvrect</a> cliprect;</p>
<p><span class="inline_code">cliprect</span> is used to specify a rectangle that limits what region of the destination is
written. This is most useful for scaling operations, where the necessary scaling factor will not allow translation
of the destination rectangle back to the source on an integer pixel boundary.</p>
<p class="note">NOTE: If <span class="inline_code">cliprect</span> exceeds the destination surface, the behavior is
undefined. </p>
<p>For example, if the goal is to show a 640 x 480 video on a 1920 x 1080 screen, the video would be stretched to 1440 x
1080 to maintain the proper aspect ratio. So the relevant rectangles would be:</p>
<table class="indent">
<tr>
<td class="thin_bord"><strong>src1rect</strong></td>
<td class="thin_bord"><strong>dstrect</strong></td>
</tr>
<tr>
<td class="thin_bord">(0, 0) - 640 x 480</td>
<td class="thin_bord">(240, 0) - 1440 x 1080</td>
</tr>
</table>
<p>However, to handle a 640 x 480 pop-up window that appears centered on the screen, in front of the video, the single BLT
may be broken into four smaller BLTs pieced around the popup. These rectangles would need to be:</p>
<table class="indent">
<tr>
<td class="thin_bord"><strong>src1rect</strong></td>
<td class="thin_bord"><strong>dstrect</strong></td>
</tr>
<tr>
<td class="thin_bord">(0, 0) - 640 x 133.333...</td>
<td class="thin_bord">(240, 0) - 1440 x 300</td>
</tr>
<tr>
<td class="thin_bord">(0, 133.333...) - 284.444... x 213.333...</td>
<td class="thin_bord">(240, 300) - 400 x 480</td>
</tr>
<tr>
<td class="thin_bord">(568.888..., 133.333...) - 284.444... x 213.333...</td>
<td class="thin_bord">(1280, 300) - 400 x 480</td>
</tr>
<tr>
<td class="thin_bord">(0, 346.666...) - 640 x 133.333...</td>
<td class="thin_bord">(240, 780) - 1440 x 300</td>
</tr>
</table>
<p>Since this is a scaling factor of 2.25x, translating the required destination rectangles back to the source results in
non-integer coordinates and dimensions, as illustrated above. And adjusting the source rectangles to the nearest integer
values will result in visible discontinuities at the boundaries between the rectangles.</p>
<p>Instead, using the <span class="inline_code">cliprect</span>, this situation can be handled more easily:</p>
<table class="indent">
<tr>
<td class="thin_bord"><strong>src1rect</strong></td>
<td class="thin_bord"><strong>dstrect</strong></td>
<td class="thin_bord"><strong>cliprect</strong></td>
</tr>
<tr>
<td class="thin_bord">(0, 0) - 640 x 480</td>
<td class="thin_bord">(240, 0) - 1440 x 1080</td>
<td class="thin_bord">(240, 0) - 1440 x 300</td>
</tr>
<tr>
<td class="thin_bord">(0, 0) - 640 x 480</td>
<td class="thin_bord">(240, 0) - 1440 x 1080</td>
<td class="thin_bord">(240, 300) - 400 x 480</td>
</tr>
<tr>
<td class="thin_bord">(0, 0) - 640 x 480</td>
<td class="thin_bord">(240, 0) - 1440 x 1080</td>
<td class="thin_bord">(1280, 300) - 400 x 480</td>
</tr>
<tr>
<td class="thin_bord">(0, 0) - 640 x 480</td>
<td class="thin_bord">(240, 0) - 1440 x 1080</td>
<td class="thin_bord">(240, 780) - 1440 x 300</td>
</tr>
</table>
<p class="Code_Header_2"><a name="batchflags">bvbltparams.batchflags</a></p>
<p class="code_block">unsigned long batchflags;</p>
<p><span class="inline_code">batchflags</span> are used by the client as a hint to indicate to the implementation which
parameters are changing between successive BLTs of a batch. The flags may be used when the
<span class="inline_code"><a href="#flags">bvbltparams.flags</a></span> has <span class="inline_code">
<a href="#BVFLAG_BATCH_CONTINUE">BVFLAG_BATCH_CONTINUE</a></span> or <span class="inline_code">
<a href="#BVFLAG_BATCH_END">BVFLAG_BATCH_END</a></span> set.</p>
<table style="" class="indent">
<tr>
<td><span class="inline_code"><a name="BVBATCH_OP">BVBATCH_OP</a></span></td>
<td>indicates that the operation type (<span class="inline_code"><a href="#BVFLAG_ROP">BVFLAG_ROP</a></span>,
<span class="inline_code"><a href="#BVFLAG_BLEND">BVFLAG_BLEND</a></span>, <span class="inline_code">
<a href="#BVFLAG_FILTER">BVFLAG_FILTER</a></span>, etc.) has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_KEY">BVBATCH_KEY</a></span></td>
<td>indicates that the <span class="inline_code"><a href="#colorkey">bvbltparams.colorkey</a></span> or the color
key mode (<span class="inline_code"><a href="#BVFLAG_KEY_SRC">BVFLAG_KEY_SRC</a></span>/<span class="inline_code"><a href="#BVFLAG_KEY_DST">BVFLAG_KEY_DST</a></span>)
has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_MISCFLAGS">BVBATCH_MISCFLAGS</a></span></td>
<td>indicates that <span class="inline_code"><a href="#flags">bvbltparams.flags</a></span> other than the operation,
color key, or clip flag have changes.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_ALPHA">BVBATCH_ALPHA</a></span></td>
<td>indicates that <span class="inline_code"><a href="#globalalpha">bvbltparams.globalalpha</a></span> or global
alpha type has changed.</td>
</tr>
<tr>
<td><a name="BVBATCH_DITHER" class="inline_code">BVBATCH_DITHER</a></td>
<td>indicates that <span class="inline_code"><a href="#dithermode">bvbltparams.dithermode</a></span> has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_SCALE">BVBATCH_SCALE</a></span></td>
<td>indicates that <span class="inline_code"><a href="#scalemode">bvbltparams.scalemode</a></span> has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_DST">BVBATCH_DST</a></span></td>
<td>indicates that the destination surface (<span class="inline_code"><a href="#dstdesc">bvbltparams.dstdesc</a></span>,
<span class="inline_code"><a href="#dstgeom">bvbltparams.dstgeom</a></span> ,or <span class="inline_code">
<a href="#dstrect">bvbltparams.dstrect</a></span>) has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_SRC1">BVBATCH_SRC1</a></span> </td>
<td>indicates that the source 1 surface (<span class="inline_code"><a href="#src1.desc">bvbltparams.src1.desc</a></span>
or <span class="inline_code"><a href="#src1.tileparams">bvbltparams.src1.tileparams</a></span>, or
<span class="inline_code"><a href="#src1geom">bvbltparams.src1geom</a></span>) has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_SRC2">BVBATCH_SRC2</a></span> </td>
<td>indicates that the source 2 surface (<span class="inline_code"><a href="#src2.desc">bvbltparams.src2.desc</a></span>
or <span class="inline_code"><a href="#src2.tileparams">bvbltparams.src2.tileparams</a></span>, or
<span class="inline_code"><a href="#src2geom">bvbltparams.src2geom</a></span>) has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_MASK">BVBATCH_MASK</a></span> </td>
<td>indicates that the mask surface (<span class="inline_code"><a href="#mask.desc">bvbltparams.mask.desc</a></span>
or <span class="inline_code"><a href="#mask.tileparams">bvbltparams.mask.tileparams</a></span>, or
<span class="inline_code"><a href="#maskgeom">bvbltparams.maskgeom</a></span>) has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_DSTRECT_ORIGIN">BVBATCH_DSTRECT_ORIGIN</a></span></td>
<td>indicates that <span class="inline_code"><a href="#dstrect">bvbltparams.dstrect.left</a></span> or
<span class="inline_code"><a href="#dstrect">top</a></span> has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_DSTRECT_SIZE">BVBATCH_DSTRECT_SIZE</a></span></td>
<td>indicates that the <span class="inline_code"><a href="#dstrect">bvbltparams.dstrect.width</a></span> or
<a href="#dstrect">height</a> has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_SRC1RECT_ORIGIN">BVBATCH_SRC1RECT_ORIGIN</a></span></td>
<td>indicates that <span class="inline_code"><a href="#src1rect">bvbltparams.src1rect.left</a></span> or
<span class="inline_code"><a href="#dstrect">top</a></span> has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_SRC1RECT_SIZE">BVBATCH_SRC1RECT_SIZE</a></span></td>
<td>indicates that the <span class="inline_code"><a href="#src1rect">bvbltparams.src1rect.width</a></span> or
<a href="#src1rect">height</a> has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_SRC2RECT_ORIGIN">BVBATCH_SRC2RECT_ORIGIN</a></span></td>
<td>indicates that <span class="inline_code"><a href="#src2rect">bvbltparams.src2rect.left</a></span> or
<span class="inline_code"><a href="#src2rect">top</a></span> has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_SRC2RECT_SIZE">BVBATCH_SRC2RECT_SIZE</a></span></td>
<td>indicates that the <span class="inline_code"><a href="#src2rect">bvbltparams.src2rect.width</a></span> or
<a href="#src2rect">height</a> has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_MASKRECT_ORIGIN">BVBATCH_MASKRECT_ORIGIN</a></span></td>
<td>indicates that <span class="inline_code"><a href="#maskrect">bvbltparams.maskrect.left</a></span> or
<span class="inline_code"><a href="#maskrect">top</a></span> has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_MASKRECT_SIZE">BVBATCH_MASKRECT_SIZE</a></span></td>
<td>indicates that the <span class="inline_code"><a href="#maskrect">bvbltparams.maskrect.width</a></span> or
<a href="#maskrect">height</a> has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_CLIPRECT_ORIGIN">BVBATCH_CLIPRECT_ORIGIN</a></span></td>
<td>indicates that <span class="inline_code"><a href="#cliprect">bvbltparams.cliprect.left</a></span> or
<span class="inline_code"><a href="#cliprect">top</a></span> has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_CLIPRECT_SIZE">BVBATCH_CLIPRECT_SIZE</a></span></td>
<td>indicates that the <span class="inline_code"><a href="#cliprect">bvbltparams.cliprect.width</a></span> or
<a href="#cliprect">height</a> has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_TILE_SRC1">BVBATCH_TILE_SRC1</a></span></td>
<td>indicates that the <span class="inline_code"><a href="#src1.tileparams">bvbltparams.src1.tileparams</a></span>
has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_TILE_SRC2">BVBATCH_TILE_SRC2</a></span></td>
<td>indicates that the <span class="inline_code"><a href="#src2.tileparams">bvbltparams.src2.tileparams</a></span>
has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_TILE_MASK">BVBATCH_TILE_MASK</a></span></td>
<td>indicates that the <span class="inline_code"><a href="#mask.tileparams">bvbltparams.mask.tileparams</a></span>
has changed.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVBATCH_ENDNOP">BVBATCH_ENDNOP</a></span></td>
<td>is a special flag used with <span class="inline_code"><a href="#batch_end">BVFLAG_BATCH_END</a></span>, for
clients that do not have information that a batch is ending until after the last BLT has been issued. When
this flag is set, no BLT is done, but the batch is ended.</td>
</tr>
</table>
<p class="note">NOTE: These flags are hints, and may be used or not by a BLTsville implementation. So if
<span class="inline_code"><a href="#bvbltparams">bvbltparams</a></span> members are changed between BLTs in a batch, but
the <span class="inline_code">bvbltparams.batchflags</span> member is not correctly updated, the resulting behavior on different
implementations will not be consistent.</p>
<p class="Code_Header_2"><a name="batch">bvbltparams.batch</a></p>
<p class="code_block"><a href="#bvbatch">struct bvbatch</a> *batch;</p>
<p>This member is used as a batch handle, so that multiple batches can be under construction at the same time.</p>
<p class="Code_Header_2"><a name="callbackfn">bvbltparams.callbackfn</a></p>
<p class="code_block">void (*callbackfn)(<a href="#bvcallbackerror">struct bvcallbackerror</a> *err, unsigned long
<a href="#callbackdata">callbackdata</a>);</p>
<p>This member is a pointer to a client-supplied function which is called by the implementation when
<span class="inline_code"><a href="#BVFLAG_ASYNC">BVFLAG_ASYNC</a></span> is set and the BLT is complete. If this
member is NULL, no callback is performed. When there is no error, the <span class="inline_code">err</span>
parameter will be set to 0;</p>
<p class="note">NOTE: This function <span class="underline">can be called</span> before the
<span class="inline_code"><a href="#bv_blt">bv_blt()</a></span> call has returned.</p>
<p class="Code_Header_2"><a name="callbackdata">bvbltparams.callbackdata</a></p>
<p class="code_block">unsigned long callbackdata;</p>
<p>This member is used as the parameter passed back by the <span class="inline_code"><a href="#callbackfn">bvbltparams.callbackfn</a></span>.
This can be anything from an identifying index to a pointer used by the client.</p>
<p class="Code_Header_2">bvbltparams.<a name="src2auxdstrect">src2</a>/<a name="maskauxdstrect">maskauxdstrect</a></p>
<p class="code_block">struct bvrect src2auxdstrect;<br />
struct bvrect maskauxdstrect;</p>
<p>These two members are used only when the associated <span class="inline_code"><a href="#BVFLAG_SRC2_AUXDSTRECT">
BVFLAG_SRC2</a></span>/<span class="inline_code"><a href="#BVFLAG_MASK_AUXDSTRECT">MASK_AUXDSTRECT</a></span> flags are
set. They are only necessary (and should only be used) in the case where scaling of the inputs differs and the
entire source images are not being used. <span class="inline_code"><a href="#dstrect">bvbltparams.dstrect</a></span> is always
used to specify the destination of source 1 image. When the associated flags are set, these two members are used
to specify the destination of the source 2 and mask images, instead of <span class="inline_code"><a href="#dstrect">
bvbltparams.dstrect</a></span>.</p>
<p>These flags must be used with the <span class="inline_code"><a href="#BVFLAG_CLIP">BVFLAG_CLIP</a></span> flag.
And if the resulting clipped destination does not include all enabled destination rectangles, the results are undefined.</p>
<div class="example"><strong>Example:</strong> We have two images that we want to merge and view on an 854x480 LCD panel. One
image is a small background image with 16:9 (64x36) aspect ratio that we want to stretch to fill the screen.
The other is a standard definition 720x480 (4:3 aspect ratio) image with transparency we want to blend on top of our
background.<br />
<table align="center">
<tr>
<td>
<table>
<tr>
<td class="ctr"><img alt="" src="concrete-64x36.png" width="64" height="36" /></td>
</tr>
<tr>
<td class="ctr">(shown actual size)</td>
</tr>
</table>
</td>
<td>
<table class="ctr">
<tr>
<td><img alt="" src="clock-720x480_4x3-fauxtrans.jpg" width="360" height="240" /></td>
</tr>
<tr>
<td class="ctr">(shown 1/2x; not adjusted for aspect ratio)</td>
</tr>
</table>
</td>
</tr>
</table>
We want to blend the second image onto the center of the first, scaling both, so that it looks like this:<br />
<table align="center">
<tr>
<td><img alt="" src="blend-854x480.jpg" width="427" height="240" /></td>
</tr>
<tr>
<td class="ctr">(shown 1/2x)</td>
</tr>
</table>
The screen is effectively a 16:9 aspect ratio (we can ignore the fraction of a pixel here), which matches our
background image. So the background image just needs to be scaled from 64x36 to 854x480.<br />
<br />
However, since the second image has a 4:3 aspect ratio, it will not cover the entire background image if we want to
maintain its aspect ratio. Our second image is not as wide as our 16:9 image, which means it's height will match
the screen height, but the width will be smaller. Since the screen is 480 lines (pixels) high, to maintain our 4:3
aspect ratio, our second image will need to be 640 pixels wide (4 * 480 / 3). So it will need to be scaled from
720x480 to 640x480.<br />
<br />
As we mentioned, we would like to center the 640 pixel image on our 854 pixel wide screen. That means the left edge
of the image will be at pixel 107 ( (854 - 640) / 2 ). So the leftmost 107 columns of pixels will just be a copy
of the left portion of the background image. Likewise, the rightmost 107 columns will be a copy of the right
portion of the background image. Only the middle section should be blended.<br />
<table align="center">
<tr>
<td><img alt="" src="blend-854x480-threeblts.jpg" width="427" height="240" /></td>
</tr>
<tr>
<td class="ctr">(shown 1/2x)</td>
</tr>
</table>
The side two BLTs are quite easy with BLTsville, by using the clipping rectangle:<br />
<br />
<table class="indent"><tr><td>
<span class="inline_code">bvbltparams.flags = BVFLAG_ROP | BVFLAG_CLIP;<br />
bvbltparams.op.rop = 0xCCCC;<br />
<br />
bvbltparams.src1.desc = bkgnddesc;<br />
bvbltparams.src1geom = bkgndgeom;<br />
bvbltparams.src1rect.left = 0;<br />
bvbltparams.src1rect.top = 0;<br />
bvbltparams.src1width = 64;<br />
bvbltparams.src1height = 36;<br />
<br />
bvbltparams.dstdesc = screendesc;<br />
bvbltparams.dstgeom = screengeom;<br />
bvbltparams.dstrect.left = 0;<br />
bvbltparams.dstrect.top = 0;<br />
bvbltparams.dstrect.width = 854;<br />
bvbltparams.dstrect.height = 480;<br />
<br />
bvbltparams.cliprect.left = 0;<br />
bvbltparams.cliprect.top = 0;<br />
bvbltparams.cliprect.width = 107;<br />
bvbltparams.cliprect.height = 480;<br />
bv_blt(&bvbltparams);<br />
<br />
bvbltparams.cliprect.left += 640;<br />
bv_blt(&bvbltparams);</span></td></tr></table>
<br />
However, if we try the same approach with the middle BLT, we run into problems:<br />
<br />
<table class="indent"><tr><td>
<span class="inline_code">bvbltparams.flags = BVFLAG_BLEND | BVFLAG_CLIP;<br />
bvbltparams.op.blend = BVBLEND_SRC1OVER;<br />
<br />
bvbltparams.src1.desc = foregnddesc;<br />
bvbltparams.src1geom = foregndgeom;<br />
bvbltparams.src1rect.left = 0;<br />
bvbltparams.src1rect.top = 0;<br />
bvbltparams.src1rect.width = 720;<br />
bvbltparams.src1rect.height = 480;<br />
<br />
bvbltparams.src2.desc = bkgnddesc;<br />
bvbltparams.src2geom = bkgndgeom;<br />
bvbltparams.src2rect.left = 0;<br />
bvbltparams.src2rect.top = 0;<br />
bvbltparams.src2width = 64;<br />
bvbltparams.src2height = 36;<br />
<br />
bvbltparams.cliprect.left = 107;<br />
bvbltparams.cliprect.top = 0;<br />
bvbltparams.cliprect.width = 640;<br />
bvbltparams.cliprect.height = 480;<br />
bv_blt(&bvbltparams);</span></td></tr></table>
<table align="center">
<tr>
<td><img alt="" src="blend-854x480-bad.jpg" width="427" height="240" /></td>
</tr>
<tr>
<td class="ctr">(shown 1/2x)</td>
</tr>
</table>
The result is that the foreground image is stretched horizontally. That's because the scaling factor is
derived from the source (1) rectangle and the destination rectangle, which is the full width of the screen.
Since we were also scaling the background, we set the destination rectangle to cover the screen, as we did in the
previous two BLTs.<br />
<br />
The edges of our foreground image are also cropped, since we were only modifying the middle of the screen.<br />
<br />
What if we change the destination rectangle?<br />
<br />
<table class="indent"><tr><td>
<span class="inline_code">bvbltparams.dstrect.left = 107;<br />
bvbltparams.dstrect.top = 0;<br />
bvbltparams.dstrect.width = 640;<br />
bvbltparams.dstrect.height = 480;<br />
<br />
bv_blt(&bvbltparams);</span></td></tr></table>
<table align="center">
<tr>
<td><img alt="" src="blend-854x480-bad2.jpg" width="427" height="240" /></td>
</tr>
<tr>
<td class="ctr">(shown 1/2x)</td>
</tr>
</table>
Here we get the proper scaling of the foreground image, but the background image is scaled improperly.<br />
<br />
What if we adjust the source rectangles?
For our purposes, we want all of the foreground image, but we only need the middle of the background image. So
we can manually specify the middle of the background image by modifying the source 2 rectangle:<br />
<br />
<table class="indent"><tr><td>
<span class="inline_code">bvbltparams.src2rect.left = 107 * 64 / 854;<br />
bvbltparams.src2rect.width = 640 * 64 / 854;</span></td></tr></table><br />
Nice, but what are those values?<br />
<br />
<table class="indent"><tr><td>
107 * 1280 / 854 = 8.0187...<br />
640 * 1280 / 854 = 47.9625...<br />
</td></tr></table>
<br />
In BLTsville, all rectangle parameters are expressed in integers (this also allows BLTsville to be used in the
kernels where floating point variables are not allowed). The clipping rectangle then handles introducing the
necessary source pixel subdivision (by translating the clipping rectangle back to the source rectangle in the
implementation). So what happens if we actually do use these values as integers?<br />
<br />
<table class="indent"><tr><td>
<span class="inline_code">bvbltparams.src2rect.left = 8;<br />
bvbltparams.src2rect.top = 0;<br />
bvbltparams.src2rect.width = 47;<br />
bvbltparams.src2height = 36;<br />
<br />
bv_blt(&bvbltparams);</span></td></tr></table>
<br />
And this is what we get:<br />
<table align="center">
<tr>
<td><img alt="" src="blend-854x480-roundingerror.jpg" width="427" height="240" /></td>
</tr>
<tr>
<td class="ctr">(shown 1/2x)</td>
</tr>
</table>
Closer, but not quite. Rounding the values above to integers still results in visible errors at the boundaries
between the middle and the side BLTs (the one on the right is a bit more visible at this reduced size, but if you
view the full image, you'll see the left one as well), because the left edge and scaling (and right edge as a
result) don't match the alignment and scaling done for the BLTs on the side. <p class="note">NOTE: This artifact is not always obvious in still images.
The images here were chosen to make the artifacts obvious in this documentation. But even if the static images
appear correct, movement of the images (e.g. moving the foreground image across the background image) or changes in
the blending (e.g. fading the foreground image out and finally removing it), will show these less obvious
discrepancies.</p>This is actually what the
clipping rectangle is for. It's meant to allow us to always specify the source and destination rectangles the
same, but move the clipping window around on the destination to get just the pixels we want. That way the
scaling and alignment area always the same. Unfortunately, for this special case, we really need a way to
specify different scaling factors for the different inputs. The src2auxdstrect (and maskauxdstrect, when
needed) have been added to provide this capability.<br />
<br />
Here is how this set of BLTs can be done:<br />
<br />
<table class="indent"><tr><td>
<span class="inline_code">bvbltparams.flags = BVFLAG_ROP | BVFLAG_CLIP;<br />
bvbltparams.op.rop = 0xCCCC;<br />
<br />
bvbltparams.src1.desc = bkgnddesc;<br />
bvbltparams.src1geom = bkgndgeom;<br />
bvbltparams.src1rect.left = 0;<br />
bvbltparams.src1rect.top = 0;<br />
bvbltparams.src1width = 64;<br />
bvbltparams.src1height = 36;<br />
<br />
bvbltparams.dstdesc = screendesc;<br />
bvbltparams.dstgeom = screengeom;<br />
bvbltparams.dstrect.left = 0;<br />
bvbltparams.dstrect.top = 0;<br />
bvbltparams.dstrect.width = 854;<br />
bvbltparams.dstrect.height = 480;<br />
<br />
bvbltparams.cliprect.left = 0;<br />
bvbltparams.cliprect.top = 0;<br />
bvbltparams.cliprect.width = 107;<br />
bvbltparams.cliprect.height = 480;<br />
bv_blt(&bvbltparams);<br />
<br />
bvbltparams.cliprect.left += 640;<br />
bv_blt(&bvbltparams);<br />
<br />
bvbltparams.flags = BVFLAG_BLEND | BVFLAG_CLIP | BVFLAG_SRC2_AUXDSTRECT;<br />
bvbltparams.op.blend = BVBLEND_SRC1OVER;<br />
<br />
bvbltparams.src1.desc = foregnddesc;<br />
bvbltparams.src1geom = foregndgeom;<br />
bvbltparams.src1rect.left = 0;<br />
bvbltparams.src1rect.top = 0;<br />
bvbltparams.src1rect.width = 720;<br />
bvbltparams.src1rect.height = 480;<br />
<br />
bvbltparams.dstrect.left = 107;<br />
bvbltparams.dstrect.top = 0;<br />
bvbltparams.dstrect.width = 640;<br />
bvbltparams.dstrect.height = 480;<br />
<br />
bvbltparams.src2.desc = bkgnddesc;<br />
bvbltparams.src2geom = bkgndgeom;<br />
bvbltparams.src2rect.left = 0;<br />
bvbltparams.src2rect.top = 0;<br />
bvbltparams.src2width = 64;<br />
bvbltparams.src2height = 36;<br />
<br />
bvbltparams.src2auxdstrect.left = 0;<br />
bvbltparams.src2auxdstrect.top = 0;<br />
bvbltparams.src2auxdstrect.width = 854;<br />
bvbltparams.src2auxdstrect.height = 480;<br />
<br />
bvbltparams.cliprect.left = 107;<br />
bvbltparams.cliprect.top = 0;<br />
bvbltparams.cliprect.width = 640;<br />
bvbltparams.cliprect.height = 480;<br />
bv_blt(&bvbltparams);</span></td></tr></table>
<br />
Using this approach, we get the desired output:<br />
<table align="center">
<tr>
<td><img alt="" src="blend-854x480.jpg" width="427" height="240" /></td>
</tr>
<tr>
<td class="ctr">(shown 1/2x)</td>
</tr>
</table>
It may also be clear that in that last BLT, the clip rectangle isn't really necessary. This is good, because
it frees up the clipping rectangle to be used to further subdivide the image if necessary (e.g. if partially
occluded).<br />
</div>
<br />
<hr />
<p class="Code_Header"><a name="bvrect">bvrect</a></p>
<p class="small_code_block">struct bvrect {<br />
int <a href="#bvrect.left">left</a>;<br />
int <a href="#bvrect.top">top</a>;<br />
unsigned int <a href="#bvrect.width">width</a>;<br />
unsigned int <a href="#bvrect.height">height</a>;<br />
};</p>
<p class="Code_Header_2"><a name="bvrect.left">bvrect.left</a></p>
<p class="code_block">int left;</p>
<p>This member indicates the left edge of the rectangle, measured in pixels from the left edge of the surface. Note
that this value <span class="underline">can</span> be negative, indicating that the rectangle begins before the left edge
of the surface. However, this is only allowed when a rectangle is clipped to the surface. If, after clipping,
the left edge of the rectangle is still negative, this is an error.</p>
<p class="Code_Header_2"><a name="bvrect.top">bvrect.top</a></p>
<p class="code_block">int top;</p>
<p>This member indicates the top edge of the rectangle, measured in lines of
<a href="#bfbuffdesc.virtstride" class="inline_code">bvbuffdesc.virtstride</a> bytes from the top edge of the surface.
Note that this value <span class="underline">can</span> be negative, indicating that the rectangle begins before the top
edge of the surface. However, this is only allowed when a rectangle is clipped to the surface. If, after clipping,
the top edge of the rectangle is still negative, this is an error.</p>
<p class="Code_Header_2"><a name="bvrect.width">bvrect.width</a></p>
<p class="code_block">unsigned int width;</p>
<p>This member indicates the width of the rectangle, measured in pixels. Note that this value
<span class="underline">cannot</span> be negative. (Horizontal flipping is indicated using the
<span class="inline_code"><a href="#BVFLAG_HORZ_FLIP">BVFLAG_HORZ_FLIP_*</a></span> flags.) The value of this member
may exceed the width of the associated surface. However, this is only allowed when a rectangle is clipped to the surface.
If, after clipping, the right edge of the rectangle still exceeds the width of the surface, this is an error.</p>
<p class="Code_Header_2"><a name="bvrect.height">bvrect.height</a></p>
<p class="code_block">unsigned int height;</p>
<p>This member indicates the height of the rectangle, measured in lines of <span class="inline_code">
<a href="#bvbuffdesc.virtstride">bvbuffdesc.virtstride</a></span> bytes. Note that this value
<span class="underline">cannot</span> be negative. (Vertical flipping is indicated using the
<span class="inline_code"><a href="#BVFLAG_VERT_FLIP">BVFLAG_VERT_FLIP_*</a></span> flags.) The value of this member
may exceed the width of the associated surface. However, this is only allowed when a rectangle is clipped to the surface.
If, after clipping, the right edge of the rectangle still exceeds the height of the surface, this is an error.</p>
<hr />
<a name="bvcopparams" class="Code_Header">bvcopparams</a>
<p><span class="inline_code">bvcopparams</span> is used to define the cache operation to be performed by
<span class="inline_code"><a href="#bv_cache">bv_cache()</a></span>.</p>
<p class="small_code_block">struct bvcopparams {<br />
unsigned int <a href="#bvcopparams.structsize">structsize</a>;<br />
<a href="#bvbuffdesc">struct bvbuffdesc</a> *<a href="#bvcopparams.desc">desc</a>;<br />
<a href="#bvsurfgeom">struct bvsurfgeom</a> *<a href="#bvcopparams.geom">geom</a>;<br />
<a href="#bvrect">struct bvrect</a> *<a href="#bvcopparams.rect">rect</a>;<br />
enum bvcacheop <a href="#bvcopparams.cacheop">cacheop</a>;<br />
};</p>
<a name="bvcopparams.structsize" class="Code_Header_2">bvcopparams.structsize</a>
<p><span class="code_block">unsigned long structsize; /* input */</span></p>
<p>This member is used for compatibility between BLTsville versions. (See <span class="inline_code">
<a href="#bvbltparams.structsize">bvbltparams.structsize</a></span> for an explanation.) </p>
<p class="Code_Header_2"><a name="bvcopparams.desc">bvcopparams.desc</a></p>
<p class="code_block"><a href="#bvbuffdesc">struct bvbuffdesc</a> *desc;</p>
<p>This member points to the <span class="inline_code"><a href="#bvbuffdesc">bvbuffdesc</a></span> of the surface for which
the cache is being manipulated. This buffer should have been mapped with a call to <span class="inline_code">
<a href="#bv_map">bv_map()</a></span>.</p>
<p class="note">NOTE: Implementations may choose to dynamically map the surface as with <span class="inline_code">
<a href="#bv_blt">bv_blt()</a></span>, however in many systems, this will not function properly due to dynamic paging which
can occur when a surface is not locked.</p>
<p><span class="Code_Header_2"><a name="bvcopparams.geom">bvcopparams.geom</a></span></p>
<p class="code_block"><a href="#bvsurfgeom">struct bvsurfgeom</a> *geom;</p>
<p>This member points to the <span class="inline_code"><a href="#bvsurfgeom">bvsurfgeom</a></span> of the surface for which
the cache is being manipulated.</p>
<p><span class="Code_Header_2"><a name="bvcopparams.rect">bvcopparams.rect</a></span></p>
<p class="code_block"><a href="#bvrect">struct bvrect</a> *rect;</p>
<p>This member points to the <span class="inline_code"><a href="bvrect">bvrect</a></span> describing the rectangle of the
surface which is being manipulated.</p>
<p><span class="Code_Header_2"><a name="bvcopparams.cacheop">bvcopparams.cacheop</a></span></p>
<p class="code_block">enum bvcacheop cacheop;</p>
<p>This member specifies the cache operation to be performed. It is an enumeration from the following list:</p>
<table style="" class="indent">
<tr>
<td><span class="inline_code"><a name="BVCACHE_BIDIRECTIONAL">BVCACHE_BIDIRECTIONAL</a></span></td>
<td>(This usually performs a cache flush operation.)</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVCACHE_CPU_TO_DEVICE">BVCACHE_CPU_TO_DEVICE</a></span></td>
<td>Performs the appropriate cache operation to ensure data can be transferred correctly when it was written with
the CPU, but will be read by the 2-D device. (This is usually a cache clean operation.)</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVCACHE_CPU_FROM_DEVICE">BVCACHE_CPU_FROM_DEVICE</a></span></td>
<td>Performs the appropriate cache operation to ensure data can be transferred correctly when it was written by
the 2-D device, but will be read by the CPU. (This is usually a cache invalidate operation.)</td>
</tr>
</table>
<br />
<hr />
<p class="Code_Header"><a name="bvbuffdesc">bvbuffdesc</a></p>
<p>This structure is used in conjunction with a <span class="inline_code"><a href="#bvsurfgeom">bvsurfgeom</a></span> structure
to specify the characteristics of a graphic surface. This structure specifies the memory buffer itself.</p>
<p class="small_code_block">struct bvbuffdesc {<br />
unsigned int <a href="#bvbuffdesc.structsize">structsize</a>;<br />
void *<a href="#bvbuffdesc.virtaddr">virtaddr</a>;<br />
unsigned long <a href="#bvbuffdesc.length">length</a>;<br />
<a href="#bvbuffmap">struct bvbuffmap</a> *<a href="#bvbuffdesc.map">map</a>;<br />
enum bvauxtype <a href="#bvbuffdesc.auxtype">auxtype</a>;<br />
void *<a href="#bvbuffdesc.auxptr">auxptr</a>;<br />
};</p>
<p class="Code_Header_2"><a name="bvbuffdesc.structsize">bvbuffdesc.structsize</a></p>
<p class="code_block">unsigned int structsize;</p>
<p>This member is used for compatibility between BLTsville versions. (See <span class="inline_code">
<a href="#bvbltparams.structsize">bvbltparams.structsize</a></span> for an explanation.) </p>
<p class="Code_Header_2"><a name="bvbuffdesc.virtaddr">bvbuffdesc.virtaddr</a></p>
<p class="code_block">void *virtaddr;</p>
<p>This member is used to indicate the CPU virtual address of the start of the buffer. This value must be provided
unless the <span class="inline_code"><a href="#bvbuffdesc.auxtype">auxtype</a></span>/<span class="inline_code"><a href="#bvbuffdesc.auxptr">auxptr</a></span>
members below are used. At that time, this member is optional, and the <span class="inline_code">
<a href="#auxptr">auxptr</a></span> usually has higher priority than this member.</p>
<p class="imponly"><strong>Implementations Only</strong><br />
<br />
Note that this is always the beginning of the buffer. This means that if the <span class="inline_code">
<a href="#bvsurfgeom.virtaddr">bvsurfgeom.virtstride</a></span> is negative, or the <a href="#bvsurfgeom.orientation">bvsurfgeom.orientation</a>
does not normalize to 0º (i.e. <span class="inline_code">orientation % 360 != 0</span>), implementations may need
to use a modified version of <span class="inline_code">virtaddr</span> internally to operate correctly.</p>
<p class="Code_Header_2"><a name="bvbuffdesc.length">bvbuffdesc.length</a></p>
<p class="code_block">unsigned long length;</p>
<p>This member specifies the length of the buffer in bytes.</p>
<p class="note">NOTE: When used with a <span class="inline_code"><a href="#bvsurfgeom">bvsurfgeom</a></span> structure,
<span class="inline_code">length</span> should be greater than or equal to <span class="inline_code">
<a href="#bvsurfgeom.height">bvsurfgeom.height</a> * <a href="#bvsurfgeom.virtstride">bvsurfgeom.virtstride</a></span>.</p>
<p class="Code_Header_2"><a name="bvbuffdesc.map">bvbuffdesc.map</a></p>
<p class="code_block">struct bvbuffmap *map;</p>
<p>This member is used by the implementations and should <span class="underline"><strong>NEVER</strong></span> be manipulated
by the client. When the <span class="inline_code">bvbuffdesc</span> structure is created, this member should be set
to 0, indicating that no implementations have mapped the buffer. After a buffer has been mapped using a call to
<span class="inline_code"><a href="#bv_map">bv_map()</a></span>, this member should be left as-is by clients. (The
implementation will set this back to 0 before returning from <span class="inline_code"><a href="#bv_unmap">bv_unmap()</a></span>.)</p>
<p class="imponly"><strong>Implementations Only</strong><br />
<br />
This member points to a linked list of <span class="inline_code"><a href="#bvbuffmap">bvbuffmap</a></span> structures associated
with the buffer. Each <span class="inline_code"><a href="#bvbuffmap">bvbuffmap</a></span> is added to the list as
the buffer is mapped by a given implementation. This may be done with an explicit call to
<span class="inline_code"><a href="#bv_map">bv_map()</a></span>, or implicitly with a call to <span class="inline_code">
<a href="#bv_blt">bv_blt()</a></span>, after a call to <span class="inline_code"><a href="#bv_map">bv_map()</a></span> from
a different implementation.<br />
<br />
Implementations should not assume that the first entry in the list is their <span class="inline_code">
<a href="#bvbuffmap">bvbuffmap</a></span>. Instead, implementations should compare the <span class="inline_code">
<a href="#bv_unmap">bv_unmap()</a></span> pointer in the structure to their own function address.</p>
<p class="Code_Header_2"><a name="bvbuffdesc.auxtype">bvbuffdesc.auxtype</a></p>
<p class="code_block">enum bvauxtype auxtype;</p>
<p>This member is used to identify the type of additional information about the buffer provided by
<span class="inline_code"><a href="#bvbuffdesc.auxptr">auxptr</a></span>. Currently no values are defined for the
user mode interface, so it should be initialized to 0 or <span class="inline_code">BVAT_NONE</span>. See the
<a href="#Kernel_Mode_Interface">Kernel Mode Interface</a> for details on the values defined for the kernel mode interface.</p>
<p class="Code_Header_2"><a name="bvbuffdesc.auxptr">bvbuffdesc.auxptr</a></p>
<p class="code_block">void *auxptr;</p>
<p>This member is used to point to additional information about the buffer. The type of this pointer is determined
by the <span class="inline_code"><a href="#auxtype">auxtype</a></span> value. Currently there are no types defined
for the user mode interface, so this member is ignored. See the <a href="#Kernel_Mode_Interface">Kernel Mode Interface</a>
for details on the types defined for the kernel mode interface. </p>
<hr /><br />
<table style="" class="imponly">
<tr>
<td>
<p><strong>Implementations Only</strong></p>
<p class="Code_Header"><a name="bvbuffmap">bvbuffmap</a></p>
<p>This structure is used from the bvbuffdesc.map member to allow implementations to associate their own data with
a buffer.</p>
<p class="small_code_block"><span class="small_code_block_in_table">struct bvbuffmap {<br />
unsigned int <a href="#bvbuffmap.structsize">structsize</a>;<br />
BVFN_UNMAP <a href="#bvbuffmap.bv_unmap">bv_unmap</a>;<br />
unsigned long <a href="#bvbuffmap.handle">handle</a>;<br />
struct bvbuffmap *<a href="#bvbuffmap.nextmap">nextmap</a>;<br />
};</span></p>
<p class="Code_Header_2"><a name="bvbuffmap.structsize">bvbuffmap.structsize</a></p>
<p class="code_block">unsigned int structsize;</p>
<p>This member is used for compatibility between BLTsville versions. (See <span class="inline_code">
<a href="#bvbltparams.structsize">bvbltparams.structsize</a></span> for an explanation.) </p>
<p class="Code_Header_2"><a name="bvbuffmap.bv_unmap">bvbuffmap.bv_unmap</a></p>
<p class="code_block">BVFN_UNMAP bv_unmap;</p>
<p>This member holds the pointer to the <span class="inline_code"><a href="#bv_unmap">bv_unmap()</a></span> function
of the implementation associated with the <span class="inline_code">bvbuffmap</span> structure. It serves
to allow implementations to identify their <span class="inline_code">bvbuffmap</span> structure in the linked list,
as well as to allow implementations to call each other's <span class="inline_code"><a href="#bv_unmap">bv_unmap()</a></span>
calls from their own.</p>
<p class="Code_Header_2"><a name="bvbuffmap.handle">bvbuffmap.handle</a></p>
<p class="code_block">unsigned long handle;</p>
<p>This member is used to hold an implementation-specific piece of data.</p>
<p class="Code_Header_2"><a name="bvbuffmap.nextmap">bvbuffmap.nextmap</a></p>
<p class="code_block">struct bvbuffmap *nextmap;</p>
<p>This member holds a pointer to the next bvbuffmap structure in the linked list. If this member is 0, there
are no more entries in the list.<br />
<br />
<span class="note">NOTE: The Linux/Android Kernel Mode Interface differs slightly from this structure.
Refer to the <a href="#Kernel_Mode_Interface">Kernel Mode Interface</a> section for details.</span></p>
</td>
</tr>
</table>
<br />
<hr />
<p class="Code_Header"><a name="bvsurfgeom">bvsurfgeom</a></p>
<p>This structure is used in conjunction with a <span class="inline_code"><a href="#bvbuffdesc">bvbuffdesc</a></span> structure
to specify the characteristics of a graphic surface. This structure specifies the surface geometric characteristics.</p>
<p class="note">NOTE: This structure was separated from <span class="inline_code"><a href="#bvbuffdesc">bvbuffdesc</a></span>
to afford much flexibility to the client. Using the same <span class="inline_code"><a href="#bvbuffdesc">bvbuffdesc</a></span>
structure with different <span class="inline_code">bvsurfgeom</span> structures or using the same
<span class="inline_code">bvsurfgeom</span> structure with different <span class="inline_code"><a href="#bvbuffdesc">bvbuffdesc</a></span>
structures may be of benefit. See the <a href="#bvsurfgeom_examples">examples</a> at the bottom of this section.</p>
<p class="small_code_block">struct bvcopparams {<br />
unsigned int <a href="#bvsurfgeom.structsize">structsize</a>;<br />
<a href="http://graphics.github.com/ocd/">enum ocdformat</a> format;<br />
unsigned int width;<br />
unsigned int height;<br />
int orientation;<br />
long virtstride;<br />
<a href="http://graphics.github.com/ocd/">enum ocdformat</a> paletteformat;<br />
void *palette;<br />
};</p>
<p class="Code_Header_2"><a name="bvsurfgeom.structsize">bvsurfgeom.structsize</a></p>
<p class="inline_code">unsigned int structsize;</p>
<p>This member is used for compatibility between BLTsville versions. (See <span class="inline_code">
<a href="#bvbltparams.structsize">bvbltparams.structsize</a></span> for an explanation.) </p>
<p class="Code_Header_2"><a name="bvsurfgeom.format">bvsurfgeom.format</a></p>
<p class="code_block"><a href="http://graphics.github.com/ocd/">enum ocdformat</a> format;</p>
<p>This member specifies the format of the surface using the <a href="http://graphics.github.com/ocd">Open Color format
Definitions (OCD)</a>.</p>
<p class="Code_Header_2"><a name="bvsurfgeom.width">bvsurfgeom.width</a></p>
<p class="code_block">unsigned int width;</p>
<p>This member specifies the width of the surface in pixels. This size does not have to be equivalent to the
<span class="inline_code"><a href="#bvsurfgeom.virtstride">virtstride</a></span> size.</p>
<p class="imponly"><strong>Implementations Only</strong><br />
<br />
Implementations should never assume that <span class="inline_code">width</span> is equivalent to
<span class="inline_code"><a href="#bvsurfgeom.virtstride">virtstride</a></span>.</p>
<p class="Code_Header_2"><a name="bvsurfgeom.height">bvsurfgeom.height</a></p>
<p class="code_block">unsigned int height;</p>
<p>This member specifies the height of the surface in lines of <span class="inline_code">
<a href="#bvsurfgeom.virtstride">virtstride</a></span> width.</p>
<p class="Code_Header_2"><a name="bvsurfgeom.orientation">bvsurfgeom.orientation</a></p>
<p class="code_block">int orientation;</p>
<p>This member specifies the orientation or angle of the surface in degrees. Since BLTsville is designed only to specify
orthogonal rectangles, this value must be a multiple of 90º. This value <span class="underline">may</span> be negative.
<em>(Extending BLTsville to handle non-orthogonal rectangles may be considered if there is sufficient interest.)</em></p>
<p class="imponly"><strong>Implementations Only</strong><br />
<br />
Implementations should normalize orientation angles. For example, a client that sets the orientation to -450º should
behave as if the value of 270º were specified. </p>
<p class="Code_Header_2"><a name="bvsurfgeom.virtstride">bvsurfgeom.virtstride</a></p>
<p class="code_block">long virtstride;</p>
<p>This member specifies the horizontal stride of the surface in bytes for an unrotated surface. The stride represents
the number of bytes needed to move from one pixel to the pixel immediately below it. This value
<span class="underline">may</span> be negative.</p>
<p class="note">NOTE: This means the <span class="inline_code">orientation</span> does not affect the
<span class="inline_code">virtstride</span>. However, rotating a surface usually results in a different configuration
(i.e. <span class="inline_code">width</span>), which <span class="underline">will</span> affect the
<span class="inline_code">virtstride</span>. For example, a 320 x 240 x 32 bpp 0º surface might have a
<span class="inline_code">virtstride</span> of 1280 bytes (320 pixels/line * 32 bits/pixel / 8 bits/byte). When the
orientation is set to 180º, the <span class="inline_code">virtstride</span> would be the same. But when the orientation
is set to 90º (or 270º), the <span class="inline_code">virtstride</span> would most likely need to be set to 960 bytes (240
pixels/line * 32 bits/pixel / 8 bits/byte). </p>
<p class="imponly"><strong>Implementations Only</strong><br />
<br />
Implementations that do not support a negative <span class="inline_code">virtstride</span> must compensate using whatever
mechanism is appropriate for the implementation. For example, using a vertical flipping/mirroring setting.</p>
<p class="note">NOTE: The <span class="inline_code">virtstride</span> name must be maintained for backwards compatibility.
However, no situation should arise where the client would need to provide two different strides for the virtual and physical
views of a surface (there are situations where a physical stride will need to be available within the implementation, but
the client will not be the one to supply it), so <em>physstride</em> will most likely never be needed. However, when
a client provides a physical description of the buffer (see the <a href="#Kernel_Mode_Interface">Kernel Mode Interface</a>
section below), the <span class="inline_code">virtstride</span> entry should be used to provide the physical stride.</p>
<p class="Code_Header_2"><a name="bvsurfgeom.paletteformat">bvsurfgeom.paletteformat</a></p>
<p class="code_block"><a href="http://graphics.github.com/ocd/">enum ocdformat</a> paletteformat;</p>
<p>This member specifies the format of the palette supplied via the <span class="inline_code">
<a href="#bvsurfgeom.palette">palette</a></span> member for palettized formats using the
<a href="http://graphics.github.com/ocd">Open Color format Definitions (OCD)</a>.</p>
<p class="Code_Header_2"><a name="bvsurfgeom.palette">bvsurfgeom.palette</a></p>
<p class="code_block">void *palette;</p>
<p>This member points to a palette used for palettized formats. The format of the palette is specified by the
<span class="inline_code"><a href="#bvsurfgeom.palette">paletteformat</a></span> member. Palettes are packed based
on their container size:</p>
<table class="indent">
<tr>
<td class="ctr_thin_bord"><strong>Palette Format</strong></td>
<td class="ctr_thin_bord"><strong>Palette Layout (byte address)</strong></td>
<td class="ctr_thin_bord"><strong>Palette Layout (little endian)</strong></td>
</tr>
<tr class="small_code_block_in_table">
<td class="thin_bord">OCDFMT_xRGB12</td>
<td class="ctr_thin_bord">n/a</td>
<td class="thin_bord">
<table style="width: 100%">
<tr>
<td class="thin_bord">*(((unsigned short *)palette) + 0)</td>
<td class="thin_bord">0xFrgb</td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned short *)palette) + 1)</td>
<td class="thin_bord">0xFrgb</td>
</tr>
<tr>
<td class="thin_bord">...</td>
<td class="thin_bord">...</td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned short *)palette) + n - 1)</td>
<td class="thin_bord">0xFrgb</td>
</tr>
</table>
</td>
</tr>
<tr class="small_code_block_in_table">
<td class="thin_bord">OCDFMT_RGB24</td>
<td class="thin_bord">
<table style="width: 100%">
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + 0)</td>
<td class="thin_bord">red0</td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + 1)</td>
<td class="thin_bord">green0</td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + 2)</td>
<td class="thin_bord">blue0</td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + 3)</td>
<td class="thin_bord">red1</td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + 4)</td>
<td class="thin_bord">green1</td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + 5)</td>
<td class="thin_bord">blue1</td>
</tr>
<tr>
<td class="thin_bord">...</td>
<td class="thin_bord"> </td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + (3 * n) - 3)</td>
<td class="thin_bord">redNm1</td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + (3 * n) - 2)</td>
<td class="thin_bord">greenNm1</td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + (3 * n) - 1)</td>
<td class="thin_bord">blueNm1</td>
</tr>
</table>
</td>
<td class="ctr_thin_bord">n/a</td>
</tr>
<tr class="small_code_block_in_table">
<td class="thin_bord">OCDFMT_RGBx24</td>
<td class="thin_bord">
<table style="width: 100%">
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + 0)</td>
<td class="thin_bord">red0</td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + 1)</td>
<td class="thin_bord">green0</td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + 2)</td>
<td class="thin_bord">blue0</td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + 3)</td>
<td class="thin_bord">0xFF</td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + 4)</td>
<td class="thin_bord">red1</td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + 5)</td>
<td class="thin_bord">green1</td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + 6)</td>
<td class="thin_bord">blue1</td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + 7)</td>
<td class="thin_bord">0xFF</td>
</tr>
<tr>
<td class="thin_bord">...</td>
<td class="thin_bord"> </td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + (4 * n) - 4)</td>
<td class="thin_bord">redNm1</td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + (4 * n) - 3)</td>
<td class="thin_bord">greenNm1</td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + (4 * n) - 2)</td>
<td class="thin_bord">blueNm1</td>
</tr>
<tr>
<td class="thin_bord">*(((unsigned char *)palette) + (4 * n) - 1)</td>
<td class="thin_bord">0xFF</td>
</tr>
</table>
</td>
<td class="thin_bord">
<table style="width: 100%">
<tr>
<td class="thin_bord">*(((unsigned long *)palette) + 0)</td>
<td class="thin_bord">0xFFbbggrr<br />
</td>
</tr>
<tr class="thin_bord">
<td class="thin_bord"> </td>
<td class="thin_bord"> </td>
</tr>
<tr class="thin_bord">
<td class="thin_bord"> </td>
<td class="thin_bord"> </td>
</tr>
<tr class="thin_bord">
<td class="thin_bord"> </td>
<td class="thin_bord"> </td>
</tr>
<tr class="thin_bord">
<td class="thin_bord">*(((unsigned long *)palette) + 1)<br />
</td>
<td class="thin_bord">0xFFbbggrr<br />
</td>
</tr>
<tr class="thin_bord">
<td class="thin_bord"> </td>
<td class="thin_bord"> </td>
</tr>
<tr class="thin_bord">
<td class="thin_bord"> </td>
<td class="thin_bord"> </td>
</tr>
<tr class="thin_bord">
<td class="thin_bord"> </td>
<td class="thin_bord"> </td>
</tr>
<tr class="thin_bord">
<td class="thin_bord">...</td>
<td class="thin_bord"> </td>
</tr>
<tr class="thin_bord">
<td class="thin_bord">*(((unsigned long *)palette) + n - 1)|<br />
</td>
<td class="thin_bord">0xFFbbggrr<br />
</td>
</tr>
<tr class="thin_bord">
<td class="thin_bord"> </td>
<td class="thin_bord"> </td>
</tr>
<tr class="thin_bord">
<td class="thin_bord"> </td>
<td class="thin_bord"> </td>
</tr>
<tr class="thin_bord">
<td class="thin_bord"> </td>
<td class="thin_bord"> </td>
</tr>
</table>
</td>
</tr>
</table>
<p class="note">NOTE: Use of subsampled formats for <span class="inline_code">paletteformat</span> is currently undefined.</p>
<p class="Header4"><a name="bvsurfgeom_examples">Examples</a></p>
<p>Mixing and matching <span class="inline_code"><a href="#bvbuffdesc">bvbuffdesc</a></span> and
<span class="inline_code">bvsurfgeom</span> structures provides maximum flexibility for a client.</p>
<table style="width: 100%" class="example">
<tr>
<td><strong>Example:</strong> Using two different <span class="inline_code">bvsurfgeom</span> structures with
the same <span class="inline_code"><a href="#bvbuffdesc">bvbuffdesc</a></span> structure allows in-place format
conversion:</td>
</tr>
<tr>
<td>
<p class="indent"><span class="small_code_block_in_table">...<br />
// Convert premultiplied image to non-premultiplied in place<br />
struct
bvbltparams parms;<br />
...<br />
struct
bvbuffdesc buff;<br />
...<br />
struct
bvsurfgeom srcgeom, dstgeom;<br />
...<br />
srcgeom.format = OCDFMT_RGBA24;<br />
dstgeom.format = OCDFMT_nRGBA24;<br />
...<br />
parms.src1.desc = &buff;<br />
parms.src1geom = &srcgeom;<br />
parms.dstdesc = &buff;<br />
parms.dstgeom = &dstgeom;<br />
...<br />
bv_blt(&parms);<br />
... </span></p>
</td>
</tr>
</table>
<br />
<table style="width: 100%" class="example">
<tr>
<td><strong>Example:</strong> Using three different <span class="inline_code"><a href="#bvbuffdesc">bvbuffdesc</a></span>
structures with the same <span class="inline_code">bvsurfgeom</span> structure reduces code and copy errors:</td>
</tr>
<tr>
<td>
<p class="indent"><span class="small_code_block_in_table">...<br />
// Blend two images of the same size<br />
struct
bvbltparams parms;<br />
...<br />
struct
bvbuffdesc src1buff, src2buff, dstbuff;<br />
...<br />
struct
bvsurfgeom geom;<br />
...<br />
parms.src1.desc = &src1buff;<br />
parms.src1geom = &geom;<br />
parms.src2.desc = &src2buff;<br />
parms.src2geom = &geom;<br />
parms.dstdesc = &dstbuff;<br />
parms.dstgeom = &dstgeom;<br />
...<br />
bv_blt(&parms);<br />
... </span></p>
</td>
</tr>
</table>
<br />
<hr />
<p class="Code_Header"><a name="bvtileparams">bvtileparams</a></p>
<p>This structure is used to define the parameters necessary to use a small image as a tile or block that will be repeated
when used as a source. This structure is used in conjunction with the associated <span class="inline_code">
<a href="#bvsurfgeom">bvsurfgeom</a></span> and the associated <span class="inline_code"><a href="#bvrect">bvrect</a></span>
to determine the operation that is performed.</p>
<p class="small_code_block">struct bvcopparams {<br />
unsigned int <a href="#bvtileparams.structsize">structsize</a>;<br />
unsigned long <a href="#bvtileparams.flags">flags</a>;<br />
void *<a href="#bvtileparams.virtaddr">virtaddr</a>;<br />
int <a href="#bvtileparams.dstleft">dstleft</a>;<br />
int <a href="#bvtileparams.dsttop">dsttop</a>;<br />
unsigned int <a href="#bvtileparams.srcwidth">srcwidth</a>;<br />
unsigned int <a href="#bvtileparams.srcheight">srcheight</a>;<br />
};</p>
<p class="Code_Header_2"><a name="bvtileparams.structsize">bvtileparams.structsize</a></p>
<p class="code_block">unsigned int structsize;</p>
<p>This member is used for compatibility between BLTsville versions. (See <span class="inline_code">
<a href="#bvbltparams.structsize">bvbltparams.structsize</a></span> for an explanation.) </p>
<p class="Code_Header_2"><a name="bvtileparams.flags">bvtileparams.flags</a></p>
<p class="code_block">unsigned long flags;</p>
<p>This member specifies some additional information for the tiling operation. It can be composed as the binary OR
of one selection for each edge (left, top, right, and bottom) from the following flags:</p>
<table style="" class="indent">
<tr>
<td><span class="inline_code"><a name="BVTILE_LEFT_REPEAT">BVTILE_LEFT_REPEAT</a></span></td>
<td>indicates that the tile is repeated to the left of the destination alignment location.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVTILE_TOP_REPEAT">BVTILE_TOP_REPEAT</a></span></td>
<td>indicates that the tile is repeated above the destination alignment location.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVTILE_RIGHT_REPEAT">BVTILE_RIGHT_REPEAT</a></span></td>
<td>indicates that the tile is repeated to the right of the destination alignment location.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVTILE_BOTTOM_REPEAT">BVTILE_BOTTOM_REPEAT</a></span></td>
<td>indicates that the tile is repeated below the destination alignment location.</td>
</tr>
<tr>
<td><a name="BVTILE_LEFT_MIRROR" class="inline_code">BVTILE_LEFT_MIRROR</a></td>
<td>indicates that the tile is mirrored to the left of the destination alignment location.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVTILE_TOP_MIRROR">BVTILE_TOP_MIRROR</a></span></td>
<td>indicates that the tile is mirrored above the destination alignment location.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVTILE_RIGHT_MIRROR">BVTILE_RIGHT_MIRROR</a></span></td>
<td>indicates that the tile is mirrored to the right of the destination alignment location.</td>
</tr>
<tr>
<td><span class="inline_code"><a name="BVTILE_BOTTOM_MIRROR">BVTILE_BOTTOM_MIRROR</a></span></td>
<td>indicates that the tile is mirrored below the destination alignment location.</td>
</tr>
</table>
<p class="Code_Header_2"><a name="bvtileparams.virtaddr">bvtileparams.virtaddr</a></p>
<p class="code_block">void *virtaddr;</p>
<p>This member is used to indicate the CPU virtual address of the start of the buffer.</p>
<p class="imponly"><strong>Implementations Only</strong><br />
<br />
Note that this is always the beginning of the buffer. This means that if the <span class="inline_code">
<a href="#bvsurfgeom.virtaddr">bvsurfgeom.virtstride</a></span> is negative, or the <a href="#bvsurfgeom.orientation">bvsurfgeom.orientation</a>
does not normalize to 0º (i.e. <span class="inline_code">orientation % 360 != 0</span>), implementations may need
to use a modified version of <span class="inline_code">virtaddr</span> internally to operate correctly.</p>
<p class="Code_Header_2"><a name="bvtileparams.dstleft">bvtileparams.dstleft</a></p>
<p class="code_block">int dstleft;</p>
<p>This member is used to designate the left edge of the location of the tile in the destination for alignment purposes
(alignment location). Note that the <span class="inline_code"><a href="#bvrect">bvrect</a></span> of the destination
specifies the region which is filled by the tile.</p>
<p class="Code_Header_2"><a name="bvtileparams.dsttop">bvtileparams.dsttop</a></p>
<p class="code_block">int dsttop;</p>
<p>This member is used to designate the top edge of the location of the tile in the destination for alignment purposes (alignment
location). Note that the <span class="inline_code"><a href="#bvrect">bvrect</a></span> of the destination specifies
the region which is filled by the tile.</p>
<p class="Code_Header_2"><a name="bvtileparams.srcwidth">bvtileparams.srcwidth</a></p>
<p class="code_block">unsigned int srcwidth;</p>
<p>This member is used to designate the width of the source for purposes of scaling. The relationship between this
field and the <span class="inline_code"><a href="#bvrect.width">bvrect.width</a></span> of the associated source surface
determines the horizontal scaling factor.</p>
<p class="Code_Header_2"><a name="bvtileparams.srcheight">bvtileparams.srcheight</a></p>
<p class="code_block">unsigned int srcheight;</p>
<p>This member is used to designate the height of the source for purposes of scaling. The relationship between this
field and the <span class="inline_code"><a href="#bvrect.height">bvrect.height</a></span> of the associated source surface
determines the vertical scaling factor.</p>
<hr />
<p class="Code_Header"><a name="bvcallbackerror">bvcallbackerror</a></p>
<p>This structure is used to provide error information to the client of a BLT that failed within an asynchronous operation.
The errors will be limited to those that occur within the implementation.</p>
<p class="note">NOTE: Parameter errors should never be returned in this structure. These should have been returned
to the client before the BLT was ever initiated.</p>
<p class="small_code_block">struct bvcallbackerror {<br />
unsigned int <a href="#bvcallbackerror.structsize">structsize</a>;<br />
<a href="#bverror">enum bverror</a> <a href="#bvcallbackerror.error">error</a>;<br />
char *<a href="#bvcallbackerror.errdesc">errdesc</a>;<br />
};</p>
<p class="Code_Header_2"><a name="bvcallbackerror.structsize">bvcallbackerror.structsize</a></p>
<p class="code_block">unsigned int structsize;</p>
<p>This member is used for compatibility between BLTsville versions. (See <span class="inline_code">
<a href="#bvbltparams.structsize">bvbltparams.structsize</a></span> for an explanation.) </p>
<p class="Code_Header_2"><a name="bvcallbackerror.error">bvcallbackerror.error</a></p>
<p class="code_block"><a href="#bverror">enum bverror</a> error;</p>
<p>This member is used to indicate the error encountered. In general, these will be error like these:</p>
<table class="indent">
<tr>
<td class="inline_code">BVERR_OP_FAILED</td>
<td>The operation failed for unspecified reasons. The destination buffer was not modified.</td>
</tr>
<tr>
<td class="inline_code">BVERR_OP_INCOMPLETE</td>
<td>The operation only partially completed. The destination buffer is in an undefined state.</td>
</tr>
<tr>
<td class="inline_code">BVERR_MEMORY_ERROR</td>
<td>The operation resulted in a memory error, most likely due to an attempt to access invalid memory. The
destination buffer is in an undefined state.</td>
</tr>
</table>
<p class="Code_Header_2"><a name="bvcallbackerror.errdesc">bvcallbackerror.errdesc</a></p>
<p class="code_block">char *errdesc;</p>
<p><span class="inline_code">errdesc</span> is optionally used by implementations to pass a 0-terminated string with additional
debugging information back to clients for debugging purposes. <span class="inline_code">errdesc</span> is not localized
or otherwise meant to provide information that is displayed to users.</p>
<hr />
<p class="Header1">Batching<a name="batching"></a></p>
<p>Batching is the single most powerful feature in BLTsville. It is used for two major purposes:</p>
<ol>
<li>To group similar BLTs which use most of the same parameters so that they can be handled more efficiently by the
implementation.</li>
<li>To group BLTs that should go together so that implementations can use special features that go beyond what seems
to be expressed by the BLTsville API.</li>
</ol>
<p class="note">NOTE: It is important to realize that BLTs batched together may be done <span class="underline">in
any order</span>, and in fact may not even be done in the way specified. This includes the BLTs being done as they
are submitted, or no operations performed until the batch submission is completed with
<a href="#BVFLAG_BATCH_END" class="inline_code">BVFLAG_BATCH_END</a>. This means the client must not rely on intermediate
results within a batch.</p>
<p class="note">NOTE: Because BLTs can be performed in a variety of ways, callbacks for individual BLTs would have
no consistent meaning. So, when batching is mixed with <span class="inline_code"><a href="#BVFLAG_ASYNC">BVFLAG_ASYNC</a></span>,
only the callback for the last BLT occurs.</p>
<p class="note">NOTE: Since implementations can perform batched BLTs in a variety of ways, even synchronous batched
BLTs can be effectively asynchronous. Therefore, only the last BLT determines the synchronicity of the entire batch.
i.e. the <span class="inline_code"><a href="#BVFLAG_ASYNC">BVFLAG_ASYNC</a></span> flag is only heeded when combined with
<span class="inline_code"><a href="#BVFLAG_BATCH_END">BVFLAG_BATCH_END</a>.</span></p>
<p class="note">NOTE: Failure during the performance of a batch (different from an error on submission--indicated by the
contents of the <a href="#bvcallbackerror" class="inline_code">bvcallbackerror</a> structure) will result in an unknown
state for all destination buffers. Do not assume that a given implementation's state in this case represents the state
which will be encountered for a different implementation.</p>
<p class="note">NOTE: Because of the indeterminate nature of the execution of a batch of BLTs, a "batch abort" would not
result in a known state either. As stated above, a given implementation may have already performed earlier BLTs in
a batch as the batch is submitted. So errors encountered during the submission of a batch must be handled by the client,
and then the batch must be terminated normally using <a href="#BVFLAG_BATCH_END" class="inline_code">BVFLAG_BATCH_END</a>.</p>
<p class="Header2">Batches For Grouping Similar BLTs</p>
<p>Often, groups of similar BLTs are performed, with changes to only a few parameters. Some implementations have the
ability to re-use previous settings, coupled with these changes, to perform new BLTs.</p>
<p>One good example of this in in rendering text, similar to that you are reading now. In most systems, a glyph cache
is maintained to hold the characters of a given font, rasterized with the specific characteristics desired (e.g. bold, italics,
etc.). Each font in the glyph cache is normally created using a font rasterization engine from a vector-based font,
such as FreeType. This technology allows fonts to be described in terms of curves and lines instead of pixels, which
means they can be created as needed, in any size desirable.</p>
<table style="" class="glyph_cache">
<tr>
<td class="glyph_cache"> !"#$%&'()*+'-./0123456789:;<=>?<br />
@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_<br />
`abcdefghijklmnopqrstuvwxyz{|}~</td>
</tr>
</table>
<p>Then, when a character needs to be rendered, it is copied from the pre-rendered glyph cache. This is much more
efficient than performing the font rasterization from the vector description each time a character is used.</p>
<p>With some hardware implementations, the setup to trigger the copy of these characters from the glyph cache to the target
surface can be quite significant, when compared to the number of pixels actually affected. For example, each character
might consist of something on the order of 10 x 14, or about 140 pixels. Programming a typical hardware BLTer
may require tens of commands for each character.</p>
<p>But note that each of these BLTs differs by only a few parameters. Specifically, once the source and destination
surfaces have been specified, and the operation described, only the source and destination rectangles change between BLTs.
To alleviate much of this overhead, most implementations will allow the configuration of a previous BLT to be used again,
with only those parameters which change provided for the subsequent BLTs.</p>
<p>BLTsville provides access to this capability via the batch mechanism.</p>
<p>For rendering a word using a monospaced font like this, the client might construct the batch like this:</p>
<p class="small_code_block">struct bvbuffdesc screendesc = {sizeof(struct bvbuffdesc}, 0};<br />
struct bvsurfgeom screengeom = {sizeof(struct bvsurfgeom), 0};<br />
struct bvbuffdesc glyphcachedesc = {sizeof(struct bvbuffdesc), 0};<br />
struct bvsurfgeom glyphcachegeom = {sizeof(struct bvsurfgeom), 0};<br />
struct bvtileparams solidcolortileparams = {sizeof(struct bvtileparams), 0};<br />
struct bvbuffgeom solidcolorgeom = {sizeof(struct bvsurfgeom), 0};<br />
<br />
struct bvbltparams bltparams = {sizeof(struct bvbltparams), 0};<br />
<br />
int charsperline = 32;<br />
int fontwidth = 10;<br />
int fontheight = 14;<br />
int i = 0;<br />
<br />
screendesc.virtaddr = screenaddr;<br />
screendesc.length = screenstride * screenheight;<br />
screengeom.format = OCDFMT_RGB24;<br />
screengeom.width = screenwidth;<br />
screengeom.height = screenheight;<br />
screengeom.virtstride = screenstride;<br />
<br />
glyphcachedesc.virtaddr = glyphcacheaddr;<br />
glyphcachedesc.length = glyphcachestride * glyphcacheheight;<br />
glyphcachegeom.format = OCDFMT_ALPHA8;<br />
glyphcachegeom.width = glyphcachewidth;<br />
glyphcachegeom.height = glyphcacheheight;<br />
glyphcachegeom.virtstride = glyphstride;<br />
<br />
solidcolortileparams.virtaddr = &solidcolor;<br />
solidcolortileparams.srcwidth = 1;<br />
solidcolortileparams.srcheight = 1;<br />
solidcolorgeom.format = OCDFMT_RGB24;<br />
<br />
bltparams.flags = BVFLAG_BLEND | BVFLAG_SRC1_TILED | BVFLAG_BATCH_BEGIN;<br />
bltparams.op.blend = BVBLEND_SRCOVER + BVBLENDDEF_REMOTE;<br />
bltparams.dstdesc = &screendesc;<br />
bltparams.dstgeom = &screengeom;<br />
bltparams.src1.tileparams = &solidcolortileparams;<br />
bltparams.src1geom = &solidcolorgeom;<br />
bltparams.src2.desc = &screendesc;<br />
bltparams.src2geom = &screengeom;<br />
bltparams.mask.desc = &glyphcachedesc;<br />
bltparams.maskgeom = &glyphcachegeom;<br />
<br />
bltparams.dstrect.left = bltparams.src2rect.left = screenrect.left;<br />
bltparams.dstrect.top = bltparams.src2rect.top = screenrect.top;<br />
<br />
bltparams.maskrect.width = bltparams.dstrect.width = bltparams.src2rect.width = fontwidth;<br />
bltparams.maskrect.height = bltparams.dstrect.height = bltparams.src2rect.height = fontheight;<br />
<br />
bltparams.maskrect.left = ((text[i] - ' ') % charsperline) * fontwidth;<br />
bltparams.maskrect.top = ((text[i] - ' ') / charsperline) * fontheight;<br />
<br />
bv_blt(&bltparams);<br />
<br />
i++;<br />
if(i < textlen)<br />
{<br />
bltparams.flags = (bltparams.flags & ~BVFLAG_BATCH_MASK) | BVFLAG_BATCH_CONTINUE;<br />
bltparams.batchflags = BVBATCH_DSTRECT_ORIGIN | BVBATCH_SRC2RECT_ORIGIN | BVBATCH_MASKRECT_ORIGIN;<br />
<br />
do<br />
{<br />
bltparams.dstrect.left += fontwidth;<br />
bltparams.src2rect.left = bltparams.dstrect.left;<br />
<br />
bltparams.maskrect.left = ((text[i] - ' ') % charsperline) * fontwidth;<br />
bltparams.maskrect.top = ((text[i] - ' ') / charsperline) * fontheight;<br />
<br />
bv_blt(&bltparams);<br />
<br />
i++;<br />
}while(i < textlen);<br />
}<br />
<br />
bltparams.flags = (bltparams.flags & ~BVFLAG_BATCH_MASK) | BVFLAG_BATCH_END;<br />
bltparams.batchflags = BVBATCH_ENDNOP;<br />
<br />
bv_blt(&bltparams);</p>
<p class="note">NOTE: bvbltparams.batchflags is just a hit. Not all implementations support deltas in
batching, so clients must not change the values of members of <span class="inline_code"><a href="#bvbltparams">
bvbltparams</a></span> (or structures it
references) between BLTs. These values may be used.</p>
<p class="Header2">Batches For Special Feature BLTs</p>
<p>Enabling special features of some implementations is a special challenge. But BLTsville is up the task.</p>
<p>For example, perhaps an implementation is capable of blending four layers at the same time. But BLTsville only allows
blending to be specified using two layers at a time. How can this be accomplished?</p>
<p>The most prevalent blending reference used is the <a href="http://dx.doi.org/10.1145/800031.808606">Porter-Duff
whitepaper</a>, which specifies blending of two sources (A and B). So any N-source blend (N > 2) would require the blends to be
specified as a grouping of N - 1 two-source blends in order to utilize the
Porter-Duff equations. That's how such a blend is specified in BLTsville:</p>
<p class="small_code_block">bltparams.dstrect.width = bltparams.src1rect.width = bltparams.src2rect.width = dstgeom.width;<br />
bltparams.dstrect.height = bltparams.src1rect.height = bltparams.src2rect.height = dstgeom.height; <br />
<br />
bltparams.flags = BVFLAG_BLEND | BVFLAG_BATCH_BEGIN;<br />
bltparams.op.blend = BVBLEND_SRCOVER;<br />
bltparams.dstdesc = &dstdesc;<br />
bltparams.dstgeom = &dstgeom;<br />
bltparams.src1.desc = &src1desc;<br />
bltparams.src1geom = &src1geom;<br />
bltparams.src2.desc = &src2desc;<br />
bltparams.src2geom = &src2geom;<br />
<br />
bv_blt(&bltparams);<br />
<br />
bltparams.src1.desc = &src3desc;<br />
bltparams.src1geom = &src3geom;<br />
bltparams.dstdesc = &dstdesc;<br />
bltparams.dstgeom = &dstgeom;<br />
<br />
bltparams.flags = (bltparams.flags & ~BVFLAG_BATCH_MASK) | BVFLAG_BATCH_CONTINUE;<br />
bltparams.batch = BVBATCH_SRC1 | BVBATCH_SRC2;<br />
<br />
bv_blt(&bltparams);<br />
<br />
bltparams.src1.desc = &src4desc;<br />
bltparams.src1geom = &src4geom;<br />
<br />
bltparams.flags = (bltparams.flags & ~BVFLAG_BATCH_MASK) | BVFLAG_BATCH_END;<br />
bltparams.batch = BVBATCH_SRC1;<br />
<br />
bv_blt(&bltparams);</p>
<p>The driver for an implementation that can perform this pair of operations as one BLT would be tasked with recognizing
that the batch contained BLTs which can be combined.</p>
<p>The fantastic thing about this approach is that an implementation without the ability to blend N sources in one pass would perform
the blends separately, but the result would be identical. Moreover, implementations with the ability to combine
different numbers of operations would likewise produce the same results, even they they used a different number of
internal steps. Here's an example:</p>
<table align="center">
<tr>
<td>
<table>
<tr>
<td class="ctr_thin_bord"><strong>Number of<br />
Layers to<br />
Blend</strong></td>
<td class="ctr_thin_bord"><strong>BLTsville<br />
Operations</strong></td>
<td class="ctr_thin_bord"><strong>Implementation<br />
Capable of<br />
Blending One<br />
Source with a<br />
Destination</strong><br />
(2 inputs)</td>
<td class="ctr_thin_bord"><strong>Implementation<br />
Capable of<br />
Blending Two<br />
Sources to a<br />
Destination</strong><br />
(2 inputs)</td>
<td class="ctr_thin_bord"><strong>Implementation<br />
Capable of<br />
Blending Four<br />
Sources to a<br />
Destination</strong><br />
(4 inputs)</td>
<td class="ctr_thin_bord"><strong>Implementation<br />
Capable of<br />
Blending Eight<br />
Sources with<br />
a Destination</strong><br />
(5 inputs)</td>
</tr>
<tr>
<td class="ctr_thin_bord">2</td>
<td class="nowrap">A over B => O</td>
<td class="nowrap">B => O<br />
A over O => O</td>
<td class="nowrap">A over B => O</td>
<td class="nowrap">A over B => O</td>
<td class="nowrap">A over B => O</td>
</tr>
<tr>
<td class="ctr_thin_bord">3</td>
<td class="nowrap">B over C => O<br />
A over O => O</td>
<td class="nowrap">C => O<br />
B over O => O<br />
A over O => O</td>
<td class="nowrap">B over C => O<br />
A over O => O</td>
<td class="nowrap">A over B over C => O</td>
<td class="nowrap">A over B over C => O</td>
</tr>
<tr>
<td class="ctr_thin_bord">4</td>
<td class="nowrap">C over D => O<br />
B over O => O<br />
A over O => O</td>
<td class="nowrap">D => O<br />
C over O => O<br />
B over O => O<br />
A over O => O</td>
<td class="nowrap"> C over D => O<br />
B over O => O<br />
A over O => O</td>
<td class="nowrap"> A over B over C over D => O</td>
<td class="nowrap"> A over B over C over D => O</td>
</tr>
<tr>
<td class="ctr_thin_bord">5</td>
<td class="nowrap">D over E => O<br />
C over O => O<br />
B over O => O<br />
A over O => O</td>
<td class="nowrap">E => O<br />
D over O => O<br />
C over O => O<br />
B over O => O<br />
A over O => O</td>
<td class="nowrap">D over E => O<br />
C over O => O<br />
B over O => O<br />
A over O => O</td>
<td class="nowrap">D over E => O<br />
A over B over C over O => O</td>
<td class="nowrap">E => O<br />
A over B over C over D over O => O</td>
</tr>
<tr>
<td class="ctr_thin_bord">6</td>
<td class="nowrap">E over F => O<br />
D over O => O<br />
C over O => O<br />
B over O => O<br />
A over O => O</td>
<td class="nowrap">F => O<br />
E over O => O<br />
D over O => O<br />
C over O => O<br />
B over O => O<br />
A over O => O</td>
<td class="nowrap">E over F => O<br />
D over O => O<br />
C over O => O<br />
B over O => O<br />
A over O => O</td>
<td class="nowrap">D over E over F => O<br />
A over B over C over O => O</td>
<td class="nowrap">E over F => O<br />
A over B over C over D over O => O</td>
</tr>
<tr>
<td class="ctr_thin_bord">7</td>
<td class="nowrap">F over G => O<br />
E over O => O<br />
D over O => O<br />
C over O => O<br />
B over O => O<br />
A over O => O</td>
<td class="nowrap">G => O<br />
F over O => O<br />
E over O => O<br />
D over O => O<br />
C over O => O<br />
B over O => O<br />
A over O => O</td>
<td class="nowrap">F over G => O<br />
E over O => O<br />
D over O => O<br />
C over O => O<br />
B over O => O<br />
A over O => O</td>
<td class="nowrap">D over E over F over G => O<br />
A over B over C over O => O</td>
<td class="nowrap">E over F over G => O<br />
A over B over C over D over O => O</td>
</tr>
<tr>
<td class="ctr_thin_bord">8</td>
<td class="nowrap">G over H => O<br />
F over O => O<br />
E over O => O<br />
D over O => O<br />
C over O => O<br />
B over O => O<br />
A over O => O</td>
<td class="nowrap">H => O<br />
G over O => O<br />
F over O => O<br />
E over O => O<br />
D over O => O<br />
C over O => O<br />
B over O => O<br />
A over O => O</td>
<td class="nowrap">G over H => O<br />
F over O => O<br />
E over O => O<br />
D over O => O<br />
C over O => O<br />
B over O => O<br />
A over O => O</td>
<td class="nowrap">G over H => O<br />
D over E over F over O => O<br />
A over B over C over O => O</td>
<td class="nowrap">E over F over G over H => O<br />
A over B over C over D over O => O</td>
</tr>
<tr>
<td class="ctr_thin_bord">9</td>
<td class="nowrap">H over I => O<br />
G over O => O<br />
F over O => O<br />
E over O => O<br />
D over O => O<br />
C over O => O<br />
B over O => O<br />
A over O => O</td>
<td class="nowrap">I => O<br />
H over O => O<br />
G over O => O<br />
F over O => O<br />
E over O => O<br />
D over O => O<br />
C over O => O<br />
B over O => O<br />
A over O => O</td>
<td class="nowrap">H over I => O<br />
G over O => O<br />
F over O => O<br />
E over O => O<br />
D over O => O<br />
C over O => O<br />
B over O => O<br />
A over O => O</td>
<td class="nowrap">G over H over I => O<br />
D over E over F over O => O<br />
A over B over C over O => O</td>
<td class="nowrap">I => O<br />
E over F over G over H over O => O<br />
A over B over C over D over O => O</td>
</tr>
</table>
</td>
</tr>
<tr>
<td class="ctr">Comparison of batched BLTsville calls with internal operations, based on implementation capabilities.
</td>
</tr>
</table>
<p class="note">NOTE: As mentioned above a batch of BLTs may be serviced in any number of ways. In this example, the
destination buffer may be used for intermediate results, so it is important that this buffer not be used during the batch--i.e.
as a displayed buffer.</p>
<hr />
<p class="Header1"><a name="start">Where to Start</a></p>
<p><em>(Note that error checking is omitted in all the examples below for clarity.)</em> </p>
<p>1. Clients begin by opening one or more BLTsville implementations dynamically. The specific method of doing
this is dependent on the operating system. For example, Linux might do this like this:</p>
<p class="small_code_block">struct bltsvillelib<br />
{<br />
char* name;<br />
void* handle;<br />
BVFN_MAP bv_map;<br />
BVFN_BLT bv_blt;<br />
BVFN_UNMAP bv_unmap;<br />
}; <br />
<br />
struct bltsville bvlib[] =<br />
{<br />
{ "libbltsville_cpu.so", 0 },<br />
{ "libbltsville_2d.so", 0 }<br />
};<br />
const int NUMBVLIBS = sizeof(bvlib) / sizeof(struct bltsvillelib);<br />
<br />
for(int i = 0; i < NUMLIBS; i++)<br />
{<br />
bvlib[i].handle = dlopen(bvlib[i].name, RTLD_LOCAL | RTLD_LAZY);<br />
bvlib[i].bv_map = (BVFN_MAP)dlsym(bvlib[i].handle, "bv_map");<br />
bvlib[i].bv_blt = (BVFN_BLT)dlsym(bvlib[i].handle, "bv_blt");<br />
bvlib[i].bv_unmap = (BVFN_BLT)dlsym(bvlib[i].handle, "bv_unmap");<br />
}<br />
</p>
<p>2. Clients then need to create a <span class="inline_code"><a href="#bvbuffdesc">bvbuffdesc</a></span> object for
each buffer to be accessed in BLTsville:</p>
<table class="indent">
<tr>
<td valign="top">
<p class="small_code_block_in_table">struct bvbuffdesc buff =<br />
{sizeof(struct bvbuffdesc), 0};<br />
<br />
buff.virtaddr = buffptr;<br />
buff.length = bufflength;</p>
</td>
<td class="ctr"> or </td>
<td valign="top">
<p class="inline_code"><span class="small_code_block_in_table">struct bvbuffdesc buff;<br />
<br />
memset(&buff, 0, sizeof(buff));<br />
buff.structsize = sizeof(buff);<br />
buff.virtaddr = buffptr;<br />
buff.length = bufflength;</span></p>
</td>
</tr>
</table>
<p class="strong_emphasis">Note that the client must ensure that the map element and any additional members in
<span class="inline_code"><a href="#bvbuffdesc">bvbuffdesc</a></span> are initialized to 0.</p>
<p>3. Next the buffer can be mapped to give the hardware implementations a chance to associate any necessary resources
with the buffer:</p>
<table class="indent">
<tr>
<td valign="top">
<p class="small_code_block_in_table">/* do nothing */ </p>
</td>
<td class="ctr"> or </td>
<td valign="top">
<p class="small_code_block_in_table">bvlib[0].bv_map(&buff); </p>
</td>
<td class="ctr"> or </td>
<td valign="top">
<p class="small_code_block_in_table">for(int i = 0; i < NUMLIBS; i++)<br />
{<br />
if(bvlib[i].bv_map)<br />
bvlib[i].bv_map(&buff);<br />
}</p>
</td>
</tr>
</table>
<br />
<table style="width: 100%">
<tr>
<td valign="top">a. </td>
<td>This step is actually optional, as indicated above. However, if the client does not explicitly call
<span class="inline_code"><a href="#bv_map">bv_map()</a></span>, the mapping must be done by the implementation
to associate the necessary resources with the buffer. So this mapping must be done later, when
<span class="inline_code"><a href="#bv_blt">bv_blt()</a></span> is called. Additionally, since the client
did not call <span class="inline_code"><a href="#bv_map">bv_map()</a></span>, it is unlikely that the client will
call <span class="inline_code"><a href="#bv_unmap">bv_unmap()</a></span> to allow the implementation to free the
resources associated with the buffer. So the implementation will internally unmap the resources after completing
the BLT. This means that the mapping and unmapping overhead will be encountered on every call to
<span class="inline_code"><a href="#bv_blt">bv_blt()</a></span>.<br />
<em><br />
In general, the CPU implementations have (almost) no overhead associated with mapping and unmapping. So opting
not to make the <span class="inline_code"><a href="#bv_map">bv_map()</a></span> call for CPU implementations is
likely to have negligible difference in <span class="inline_code"><a href="#bv_blt">bv_blt()</a></span> performance.<br />
</em></td>
</tr>
<tr>
<td valign="top">b. </td>
<td>Calling <span class="inline_code"><a href="#bv_map">bv_map()</a></span> once for each buffer is enough to tell
the implementations that the client can be trusted to call <span class="inline_code"><a href="#bv_unmap">bv_unmap()</a></span>
when work with the buffer is complete, as indicated above. It does not matter which implementation's
<span class="inline_code"><a href="#bv_map">bv_map()</a></span> is called. However, that implementation is
the only one which will perform the mapping immediately. All other implementations will perform a <em>lazy
mapping</em> only when their <span class="inline_code"><a href="#bv_blt">bv_blt()</a></span> call is invoked.<br />
<br />
This allows the client to avoid the overhead of mapping and unmapping the buffers on each
<span class="inline_code"><a href="#bv_blt">bv_blt()</a></span> call. It also avoids the associated mapping
and unmapping overhead if a given implementation is never used.<br />
<br />
<em>As mentioned above, the CPU implementations have (almost) no overhead associated with mapping and unmapping,
so they are a good choice to use for the call to <span class="inline_code"><a href="#bv_map">bv_map()</a></span>.<br />
</em></td>
</tr>
<tr>
<td valign="top">c. </td>
<td>If the client wants direct control over the mapping and unmapping overhead, it can call the
<span class="inline_code"><a href="#bv_map">bv_map()</a></span> function of each implementation, as indicated above.
Each implementation will perform the mapping at that time, so that the overhead will not appear on subsequent calls
to <span class="inline_code"><a href="#bv_blt">bv_blt()</a></span>. </td>
</tr>
</table>
<p>4. Next the client must create <span class="inline_code"><a href="#bvsurfgeom">bvsurfgeom</a></span> objects for
each way in which a buffer will be accessed. Often, there is only one way in which a buffer is accessed, so there
will be the same number of buffers, <span class="inline_code"><a href="#bvbuffdesc">bvbuffdesc</a></span>, and
<span class="inline_code"><a href="#bvsurfgeom">bvsurfgeom</a></span> objects. If that's the case, it may be convenient
for the client to combine them into a parent structure. It may even be possible to share a single bvbuffgeom structure
among buffers. Or there will be times when it is necessary to treat a buffer in different ways for different BLTs.
Having these two structures separated allows all of these combinations.</p>
<table class="indent">
<tr>
<td valign="top">
<p class="small_code_block_in_table">struct bvsurfgeom geom =<br />
{sizeof(struct bvsurfgeom), 0};<br />
<br />
geom.format = OCDFMT_RGB24;<br />
geom.width = width;<br />
geom.height = height;<br />
geom.virtstride = stride;</p>
</td>
<td class="ctr"> or </td>
<td valign="top">
<p class="inline_code"><span class="small_code_block_in_table">struct bvsurfgeom geom;
<br />
memset(&geom, 0, sizeof(geom));<br />
geom.structsize = sizeof(geom);<br />
geom.width = width;<br />
geom.height = height;<br />
geom.virtstride = stride;</span></p>
</td>
</tr>
</table>
<p class="strong_emphasis">Note that the client must ensure that any additional members in <span class="inline_code">
<a href="#bvsurfgeom">bvsurfgeom</a></span> are initialized to 0 for future compatibility.</p>
<p>5. Now the client is ready to fill in a bvbltparams structure to specify the type of BLT requested. Here
is an example of a simple copy from the lower right corner of a surface to the upper left:</p>
<p class="small_code_block">struct bvbltparams bltparams = {sizeof(struct bvbltparams), 0};<br />
<br />
bltparams.flags = BVFLAG_ROP;<br />
bltparams.op.rop = 0xCCCC; /* SRCCOPY */<br />
bltparams.dstdesc = &buff;<br />
bltparams.dstgeom = &geom;<br />
bltparams.dstrect.left = 0;<br />
bltparams.dstrect.top = 0;<br />
bltparams.dstwidth = width / 2;<br />
bltparams.dstheight = height / 2;<br />
bltparams.src1.desc = &buff;<br />
bltparams.src1geom = &geom;<br />
bltparams.src1rect.left = width / 2;<br />
bltparams.src1rect.top = height / 2;<br />
bltparams.src1rect.width = width / 2;<br />
bltparams.src1rect.height = height / 2;</p>
<p>6. And next the client can trigger the BLT by calling <span class="inline_code"><a href="#bv_blt">bv_blt()</a></span>:</p>
<p class="small_code_block">bv_blt(&bltparams); </p>
<p><em>If the client cannot complete the requested BLT, it returns a </em><span class="inline_code"><a href="#bverror">
<em>bverror</em></a></span><em> indicating the issue. </em></p>
<p>7. Finally, the client should clean up:</p>
<p class="small_code_block">bv_unmap(&buff); </p>
<hr />
<p class="Header1"><a name="Kernel_Mode_Interface">Kernel Mode Interface</a></p>
<p>The kernel mode interface differs only slightly from the user mode interface. Currently there are two differences
in the general kernel interface, and one in the Linux/Android interface:</p>
<p class="Code_Header_2">bvbuffdesc.auxtype/auxptr</p>
<p><span class="inline_code"><a href="#bvbuffdesc.auxtype">bvbuffdesc.auxtype</a></span> is an <span class="inline_code">enum</span>,
indicating the type of the
<span class="inline_code"><a href="#bvbuffdesc.auxptr">bvbuffdesc.auxptr</a></span>. The enumeration values and
the associated types are:</p>
<table class="indent_thick_bord">
<tr>
<td class="thin_bord_dbl_botbord"><span class="inline_code"><a href="#bvbuffdesc.auxtype">bvbuffdesc.auxtype</a></span></td>
<td class="thin_bord_dbl_botbord">
<span class="inline_code"><a href="#bvbuffdesc.auxptr">bvbuffdesc.auxptr</a></span> type</td>
<td class="thin_bord_dbl_botbord">Notes</td>
</tr>
<tr>
<td class="thin_bord"><span class="inline_code"><a name="BVAT_PHYSDESC">BVAT_PHYSDESC</a></span></td>
<td class="thin_bord">
<a href="#bvphysdesc" class="inline_code">bvphysdesc</a></td>
<td class="thin_bord">Used to specify the physical pages of a physically discontiguous buffer constructed using
a single page size. This may be used with physically contiguous buffers as well, but
<span class="inline_code"><a href="#BVAT_PHYSADDR">BVAT_PHYSADDR</a></span> is preferred.</td>
</tr>
<tr>
<td class="thin_bord"><span class="inline_code"><a name="BVAT_PHYSADDR">BVAT_PHYSADDR</a></span></td>
<td class="thin_bord">physical address</td>
<td class="thin_bord">Used to specify the starting physical address of a physically contiguous buffer.</td>
</tr>
</table>
<p>The methods of describing the buffer using physical addresses is not exposed in user mode for security reasons.</p>
<hr />
<p class="Code_Header"><a name="bvphysdesc">bvphysdesc</a></p>
<p class="small_code_block">struct bvphysdesc {<br />
unsigned int <a href="#bvphysdesc.structsize">structsize</a>;<br />
unsigned long <a href="#bvphysdesc.pagesize">pagesize</a>;<br />
unsigned long *<a href="#bvphysdesc.pagearray">pagearray</a>;<br />
unsigned int <a href="#bvphysdesc.pagecount">pagecount</a>;<br />
unsigned long <a href="#bvphysdesc.pageoffset">pageoffset</a>;<br />
};</p>
<p class="Code_Header_2"><a name="bvphysdesc.structsize">bvphysdesc.structsize</a></p>
<p class="code_block">unsigned int structsize;</p>
<p>This member is used for compatibility between BLTsville versions. (See <span class="inline_code">
<a href="#bvbltparams.structsize">bvbltparams.structsize</a></span> for an explanation.) </p>
<p class="Code_Header_2"><a name="bvphysdesc.pagesize">bvphysdesc.pagesize</a></p>
<p class="code_block">unsigned long pagesize;</p>
<p>This member indicates the size of the physical pages containing the buffer. <span class="inline_code">BVAT_PHYSDESC</span>/<span class="inline_code">bvphysdesc
</span>does not support buffers which reside in pages that are not all the same size. <span class="inline_code">
bvphysdesc.pagesize</span> is used to indicate the length of the pages in the <span class="inline_code">
<a href="#bvphysdesc.pagearray">bvphysdesc.pagearray</a></span> as well as the expected alignment of those pages. If this value is 0, the default
page size of the system is assumed.</p>
<p class="note">NOTE: When used with physically contiguous buffers, this member should be set to the length of the
buffer, which is the same as the value in <span class="inline_code"><a href="#bvbuffdesc.length">bvbuffdesc.length</a></span>.</p>
<p class="Code_Header_2"><a name="bvphysdesc.pagearray">bvphysdesc.pagearray</a></p>
<p class="code_block">unsigned long *pagearray;</p>
<p>This member is an array of <span class="inline_code">unsigned long</span>s holding the physical addresses of the pages
holding the buffer. The array contains <span class="inline_code"><a href="#bvphysdesc.pagecount">pagecount</a></span>
entries. The specific format of the physical addresses is O/S dependent. However, <span class="inline_code">
BVAT_PHYSDESC</span>/<span class="inline_code">bvphysdesc</span> only supports 32-bit physical addresses.</p>
<p>Addresses in this array must be aligned on <span class="inline_code"><a href="#bvphysdesc.pagesize">
bvphysdesc.pagesize</a></span> boundaries. Use the <span class="inline_code"><a href="#bvphysdesc.pageoffset">
bvphysdesc.pageoffset</a></span> member to indicate the offset from the start of the first page to the beginning of the
buffer.</p>
<p class="note">NOTE: When used with physically contiguous buffers, the first (only) address in this array should
be aligned on the system default page boundary, and the <span class="inline_code"><a href="#bvphysdesc.pageoffset">
bvphysdesc.pageoffset</a></span> member should be used to indicate the offset from that address to the beginning of the
buffer.</p>
<p class="Code_Header_2"><a name="bvphysdesc.pagecount">bvphysdesc.pagecount</a></p>
<p class="code_block">unsigned int pagecount;</p>
<p>This member indicates the number of pages in the array pointed to by <span class="inline_code">
<a href="#bvphysdesc.pagearray">bvphysdesc.pagearray</a></span>.</p>
<p class="note">NOTE: When used with physically contiguous buffers, this member should be set to 1.</p>
<p class="Code_Header_2"><a name="bvphysdesc.pageoffset">bvphysdesc.pageoffset</a></p>
<p class="code_block">unsigned long pageoffset;</p>
<p>This member indicates the number of bytes from the start of the first page (<span class="inline_code">*pagearray</span>)
to the start of the buffer. The value must be less than <span class="inline_code"><a href="#bvphysdesc.pagesize">
bvphysdesc.pagesize</a></span>.</p>
<p class="imponly"><strong>Implementations Only</strong><br />
<br />
Implementations should not ignore this member.</p>
<hr />
<p class="Header2">bventry</p>
<p>Kernel mode entry cannot be the same as the user mode. The specific method of accessing the kernel interface is
O/S specific. However, the following interface is currently defined for the specified O/Ss:</p>
<table class="example">
<tr>
<td>
<p class="Header4">Linux/Android</p>
<p class="Code_Header"><a name="bventry">bventry</a></p>
<p>This structure is used to obtain the pointers to the implementation's BLTsville calls. The client can call
the default <span class="inline_code">bv2d_entry()</span> function to obtain the pointers to the implementation
chosen by the system integrators, or it can call a specific function to get the pointers for a specific implementation
(e.g. <span class="inline_code">gcbv_entry()</span>).</p>
<p class="small_code_block">struct bventry {<br />
unsigned int <a href="#bventry.structsize">structsize</a>;<br />
<a href="#bv_map">BVFN_MAP</a> <a href="#bventry.bv_map">bv_map</a>;<br />
<a href="#bv_unmap">BVFN_UNMAP</a> <a href="#bventry.bv_unmap">bv_unmap</a>;<br />
<a href="#bv_blt">BVFN_BLT</a> <a href="#bventry.bv_blt">bv_blt</a>;<br />
<a href="#bv_cache">BVFN_CACHE</a> <a href="#bventry.bv_cache">bv_cache</a>;<br />
};</p>
<p class="Code_Header_2"><a name="bventry.structsize">bventry.structsize</a></p>
<p class="code_block">unsigned int structsize;</p>
<p>This member is used for compatibility between BLTsville versions. (See <span class="inline_code">
<a href="#bvbltparams.structsize">bvbltparams.structsize</a></span> for an explanation.) </p>
<p class="Code_Header_2"><a name="bventry.bv_map">bventry.bv_map</a>/<a name="bventry.bv_unmap">bv_unmap</a>/<a name="bventry.bv_blt">bv_blt</a>/<a name="bventry.bv_cache">bv_cache</a></p>
<p class="code_block">BVFN_MAP bv_map;<br />
BVFN_UNMAP bv_unmap;<br />
BVFN_BLT bv_blt;<br />
BVFN_CACHE bv_cache;</p>
<p>These members hold pointers to the functions for the specific implementation queried with a call to
<span class="inline_code">*_entry()</span>.</p>
<p class="note">NOTE: <span class="inline_code"><a href="#bv_cache">bv_cache()</a></span> is optional, so
this pointer may be set to 0.</p>
</td>
</tr>
</table>
<br />
<hr /><br />
<p class="Header2">Linux/Android Deviation</p>
<p>Although the linked list used in the <span class="inline_code"><a href="#bvbuffmap">bvbuffmap</a></span> structure is
not complicated, there may be a requirement to use the standard Linux/Android kernel linked list in that environment.
To facilitate this, the <span class="inline_code"><a href="#bvbuffmap.map">bvbuffmap.map</a></span> entry is replaced by
the following entry for Linux/Android <span class="underline">kernel mode only</span>:</p>
<p class="Code_Header"><a name="bvbuffmap.node" class="Code_Header_2">bvbuffmap.node</a></p>
<p class="code_block">struct list_head node;</p>
<p>This member is used to reference the containing linked list for the <span class="inline_code"><a href="#bvbuffmap">bvbuffmap</a></span>
structures associated with the buffer.</p>
</body>
</html>
|
