Repository navigation
Expand file tree
/
Copy pathSDL_blit_N.c
More file actions
3458 lines (3218 loc) · 121 KB
/
Copy pathSDL_blit_N.c
File metadata and controls
3458 lines (3218 loc) · 121 KB
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
/*
Simple DirectMedia Layer
Copyright (C) 1997-2026 Sam Lantinga <slouken@libsdl.org>
This software is provided 'as-is', without any express or implied
warranty. In no event will the authors be held liable for any damages
arising from the use of this software.
Permission is granted to anyone to use this software for any purpose,
including commercial applications, and to alter it and redistribute it
freely, subject to the following restrictions:
1. The origin of this software must not be misrepresented; you must not
claim that you wrote the original software. If you use this software
in a product, an acknowledgment in the product documentation would be
appreciated but is not required.
2. Altered source versions must be plainly marked as such, and must not be
misrepresented as being the original software.
3. This notice may not be removed or altered from any source distribution.
*/
#include "../SDL_internal.h"
#if SDL_HAVE_BLIT_N
#include "SDL_video.h"
#include "SDL_endian.h"
#include "SDL_cpuinfo.h"
#include "SDL_blit.h"
#include "SDL_blit_copy.h"
/* General optimized routines that write char by char */
#define HAVE_FAST_WRITE_INT8 1
/* On some CPU, it's slower than combining and write a word */
#if defined(__MIPS__)
#undef HAVE_FAST_WRITE_INT8
#define HAVE_FAST_WRITE_INT8 0
#endif
/* Functions to blit from N-bit surfaces to other surfaces */
enum blit_features
{
BLIT_FEATURE_NONE = 0,
BLIT_FEATURE_HAS_MMX = 1,
BLIT_FEATURE_HAS_ALTIVEC = 2,
BLIT_FEATURE_ALTIVEC_DONT_USE_PREFETCH = 4,
BLIT_FEATURE_HAS_ARM_SIMD = 8
};
#ifdef SDL_ALTIVEC_BLITTERS
#ifdef HAVE_ALTIVEC_H
#include <altivec.h>
#endif
#ifdef __MACOSX__
#include <sys/sysctl.h>
static size_t GetL3CacheSize(void)
{
const char key[] = "hw.l3cachesize";
u_int64_t result = 0;
size_t typeSize = sizeof(result);
int err = sysctlbyname(key, &result, &typeSize, NULL, 0);
if (0 != err) {
return 0;
}
return result;
}
#else
static size_t GetL3CacheSize(void)
{
/* XXX: Just guess G4 */
return 2097152;
}
#endif /* __MACOSX__ */
#if (defined(__MACOSX__) && (__GNUC__ < 4))
#define VECUINT8_LITERAL(a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p) \
(vector unsigned char) ( a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p )
#define VECUINT16_LITERAL(a,b,c,d,e,f,g,h) \
(vector unsigned short) ( a,b,c,d,e,f,g,h )
#else
#define VECUINT8_LITERAL(a, b, c, d, e, f, g, h, i, j, k, l, m, n, o, p) \
(vector unsigned char) \
{ \
a, b, c, d, e, f, g, h, i, j, k, l, m, n, o, p \
}
#define VECUINT16_LITERAL(a, b, c, d, e, f, g, h) \
(vector unsigned short) \
{ \
a, b, c, d, e, f, g, h \
}
#endif
#define UNALIGNED_PTR(x) (((size_t)x) & 0x0000000F)
#define VSWIZZLE32(a, b, c, d) (vector unsigned char)(0x00 + a, 0x00 + b, 0x00 + c, 0x00 + d, \
0x04 + a, 0x04 + b, 0x04 + c, 0x04 + d, \
0x08 + a, 0x08 + b, 0x08 + c, 0x08 + d, \
0x0C + a, 0x0C + b, 0x0C + c, 0x0C + d)
#define MAKE8888(dstfmt, r, g, b, a) \
(((r << dstfmt->Rshift) & dstfmt->Rmask) | \
((g << dstfmt->Gshift) & dstfmt->Gmask) | \
((b << dstfmt->Bshift) & dstfmt->Bmask) | \
((a << dstfmt->Ashift) & dstfmt->Amask))
10BC8
/*
* Data Stream Touch...Altivec cache prefetching.
*
* Don't use this on a G5...however, the speed boost is very significant
* on a G4.
*/
#define DST_CHAN_SRC 1
#define DST_CHAN_DEST 2
/* macro to set DST control word value... */
#define DST_CTRL(size, count, stride) \
(((size) << 24) | ((count) << 16) | (stride))
#define VEC_ALIGNER(src) ((UNALIGNED_PTR(src)) \
? vec_lvsl(0, src) \
: vec_add(vec_lvsl(8, src), vec_splat_u8(8)))
/* Calculate the permute vector used for 32->32 swizzling */
static vector unsigned char calc_swizzle32(const SDL_PixelFormat *srcfmt, const SDL_PixelFormat *dstfmt)
{
/*
* We have to assume that the bits that aren't used by other
* colors is alpha, and it's one complete byte, since some formats
* leave alpha with a zero mask, but we should still swizzle the bits.
*/
/* ARGB */
static const struct SDL_PixelFormat default_pixel_format = {
0, NULL, 0, 0, { 0, 0 }, 0x00FF0000, 0x0000FF00, 0x000000FF, 0xFF000000, 0, 0, 0, 0, 16, 8, 0, 24, 0, NULL
};
const vector unsigned char plus = VECUINT8_LITERAL(0x00, 0x00, 0x00, 0x00,
0x04, 0x04, 0x04, 0x04,
0x08, 0x08, 0x08, 0x08,
0x0C, 0x0C, 0x0C,
0x0C);
vector unsigned char vswiz;
vector unsigned int srcvec;
Uint32 rmask, gmask, bmask, amask;
if (!srcfmt) {
srcfmt = &default_pixel_format;
}
if (!dstfmt) {
dstfmt = &default_pixel_format;
}
#define RESHIFT(X) (3 - ((X) >> 3))
rmask = RESHIFT(srcfmt->Rshift) << (dstfmt->Rshift);
gmask = RESHIFT(srcfmt->Gshift) << (dstfmt->Gshift);
bmask = RESHIFT(srcfmt->Bshift) << (dstfmt->Bshift);
/* Use zero for alpha if either surface doesn't have alpha */
if (dstfmt->Amask) {
amask =
((srcfmt->Amask) ? RESHIFT(srcfmt->Ashift) : 0x10) << (dstfmt->Ashift);
} else {
amask =
0x10101010 & ((dstfmt->Rmask | dstfmt->Gmask | dstfmt->Bmask) ^
0xFFFFFFFF);
}
#undef RESHIFT
((unsigned int *)(char *)&srcvec)[0] = (rmask | gmask | bmask | amask);
vswiz = vec_add(plus, (vector unsigned char)vec_splat(srcvec, 0));
return (vswiz);
}
#if SDL_BYTEORDER == SDL_LIL_ENDIAN
/* reorder bytes for PowerPC little endian */
static vector unsigned char reorder_ppc64le_vec(vector unsigned char vpermute)
{
/* The result vector of calc_swizzle32 reorder bytes using vec_perm.
The LE transformation for vec_perm has an implicit assumption
that the permutation is being used to reorder vector elements,
not to reorder bytes within those elements.
Unfortunatly the result order is not the expected one for powerpc
little endian when the two first vector parameters of vec_perm are
not of type 'vector char'. This is because the numbering from the
left for BE, and numbering from the right for LE, produces a
different interpretation of what the odd and even lanes are.
Refer to fedora bug 1392465
*/
const vector unsigned char ppc64le_reorder = VECUINT8_LITERAL(
0x01, 0x00, 0x03, 0x02,
0x05, 0x04, 0x07, 0x06,
0x09, 0x08, 0x0B, 0x0A,
0x0D, 0x0C, 0x0F, 0x0E);
vector unsigned char vswiz_ppc64le;
vswiz_ppc64le = vec_perm(vpermute, vpermute, ppc64le_reorder);
return (vswiz_ppc64le);
}
#endif
static void Blit_RGB888_RGB565(SDL_BlitInfo *info);
static void Blit_RGB888_RGB565Altivec(SDL_BlitInfo *info)
{
int height = info->dst_h;
Uint8 *src = (Uint8 *)info->src;
int srcskip = info->src_skip;
Uint8 *dst = (Uint8 *)info->dst;
int dstskip = info->dst_skip;
SDL_PixelFormat *srcfmt = info->src_fmt;
vector unsigned char valpha = vec_splat_u8(0);
vector unsigned char vpermute = calc_swizzle32(srcfmt, NULL);
vector unsigned char vgmerge = VECUINT8_LITERAL(0x00, 0x02, 0x00, 0x06,
0x00, 0x0a, 0x00, 0x0e,
0x00, 0x12, 0x00, 0x16,
0x00, 0x1a, 0x00, 0x1e);
vector unsigned short v1 = vec_splat_u16(1);
vector unsigned short v3 = vec_splat_u16(3);
vector unsigned short v3f =
VECUINT16_LITERAL(0x003f, 0x003f, 0x003f, 0x003f,
0x003f, 0x003f, 0x003f, 0x003f);
vector unsigned short vfc =
VECUINT16_LITERAL(0x00fc, 0x00fc, 0x00fc, 0x00fc,
0x00fc, 0x00fc, 0x00fc, 0x00fc);
vector unsigned short vf800 = (vector unsigned short)vec_splat_u8(-7);
vf800 = vec_sl(vf800, vec_splat_u16(8));
while (height--) {
vector unsigned char valigner;
vector unsigned char voverflow;
vector unsigned char vsrc;
int width = info->dst_w;
int extrawidth;
/* do scalar until we can align... */
#define ONE_PIXEL_BLEND(condition, widthvar) \
while (condition) { \
Uint32 Pixel; \
unsigned sR, sG, sB, sA; \
DISEMBLE_RGBA((Uint8 *)src, 4, srcfmt, Pixel, \
sR, sG, sB, sA); \
*(Uint16 *)(dst) = (((sR << 8) & 0x0000F800) | \
((sG << 3) & 0x000007E0) | \
((sB >> 3) & 0x0000001F)); \
dst += 2; \
src += 4; \
widthvar--; \
}
ONE_PIXEL_BLEND(((UNALIGNED_PTR(dst)) && (width)), width);
/* After all that work, here's the vector part! */
extrawidth = (width % 8); /* trailing unaligned stores */
width -= extrawidth;
vsrc = vec_ld(0, src);
valigner = VEC_ALIGNER(src);
while (width) {
vector unsigned short vpixel, vrpixel, vgpixel, vbpixel;
vector unsigned int vsrc1, vsrc2;
vector unsigned char vdst;
voverflow = vec_ld(15, src);
vsrc = vec_perm(vsrc, voverflow, valigner);
vsrc1 = (vector unsigned int)vec_perm(vsrc, valpha, vpermute);
src += 16;
vsrc = voverflow;
voverflow = vec_ld(15, src);
vsrc = vec_perm(vsrc, voverflow, valigner);
vsrc2 = (vector unsigned int)vec_perm(vsrc, valpha, vpermute);
/* 1555 */
vpixel = (vector unsigned short)vec_packpx(vsrc1, vsrc2);
vgpixel = (vector unsigned short)vec_perm(vsrc1, vsrc2, vgmerge);
vgpixel = vec_and(vgpixel, vfc);
vgpixel = vec_sl(vgpixel, v3);
vrpixel = vec_sl(vpixel, v1);
vrpixel = vec_and(vrpixel, vf800);
vbpixel = vec_and(vpixel, v3f);
vdst =
vec_or((vector unsigned char)vrpixel,
(vector unsigned char)vgpixel);
/* 565 */
vdst = vec_or(vdst, (vector unsigned char)vbpixel);
vec_st(vdst, 0, dst);
width -= 8;
src += 16;
dst += 16;
vsrc = voverflow;
}
SDL_assert(width == 0);
/* do scalar until we can align... */
ONE_PIXEL_BLEND((extrawidth), extrawidth);
#undef ONE_PIXEL_BLEND
src += srcskip; /* move to next row, accounting for pitch. */
dst += dstskip;
}
}
static void Blit_RGB565_32Altivec(SDL_BlitInfo *info)
{
int height = info->dst_h;
Uint8 *src = (Uint8 *)info->src;
int srcskip = info->src_skip;
Uint8 *dst = (Uint8 *)info->dst;
int dstskip = info->dst_skip;
SDL_PixelFormat *srcfmt = info->src_fmt;
SDL_PixelFormat *dstfmt = info->dst_fmt;
unsigned alpha;
vector unsigned char valpha;
vector unsigned char vpermute;
vector unsigned short vf800;
vector unsigned int v8 = vec_splat_u32(8);
vector unsigned int v16 = vec_add(v8, v8);
vector unsigned short v2 = vec_splat_u16(2);
vector unsigned short v3 = vec_splat_u16(3);
/*
0x10 - 0x1f is the alpha
0x00 - 0x0e evens are the red
0x01 - 0x0f odds are zero
*/
vector unsigned char vredalpha1 = VECUINT8_LITERAL(0x10, 0x00, 0x01, 0x01,
0x10, 0x02, 0x01, 0x01,
0x10, 0x04, 0x01, 0x01,
0x10, 0x06, 0x01,
0x01);
vector unsigned char vredalpha2 =
(vector unsigned char)(vec_add((vector unsigned int)vredalpha1, vec_sl(v8, v16)));
/*
0x00 - 0x0f is ARxx ARxx ARxx ARxx
0x11 - 0x0f odds are blue
*/
vector unsigned char vblue1 = VECUINT8_LITERAL(0x00, 0x01, 0x02, 0x11,
0x04, 0x05, 0x06, 0x13,
0x08, 0x09, 0x0a, 0x15,
0x0c, 0x0d, 0x0e, 0x17);
vector unsigned char vblue2 =
(vector unsigned char)(vec_add((vector unsigned int)vblue1, v8));
/*
0x00 - 0x0f is ARxB ARxB ARxB ARxB
0x10 - 0x0e evens are green
*/
vector unsigned char vgreen1 = VECUINT8_LITERAL(0x00, 0x01, 0x10, 0x03,
0x04, 0x05, 0x12, 0x07,
0x08, 0x09, 0x14, 0x0b,
0x0c, 0x0d, 0x16, 0x0f);
vector unsigned char vgreen2 =
(vector unsigned char)(vec_add((vector unsigned int)vgreen1, vec_sl(v8, v8)));
SDL_assert(srcfmt->BytesPerPixel == 2);
SDL_assert(dstfmt->BytesPerPixel == 4);
vf800 = (vector unsigned short)vec_splat_u8(-7);
vf800 = vec_sl(vf800, vec_splat_u16(8));
if (dstfmt->Amask && info->a) {
((unsigned char *)&valpha)[0] = alpha = info->a;
valpha = vec_splat(valpha, 0);
} else {
alpha = 0;
valpha = vec_splat_u8(0);
}
vpermute = calc_swizzle32(NULL, dstfmt);
while (height--) {
vector unsigned char valigner;
vector unsigned char voverflow;
vector unsigned char vsrc;
int width = info->dst_w;
int extrawidth;
/* do scalar until we can align... */
#define ONE_PIXEL_BLEND(condition, widthvar) \
while (condition) { \
unsigned sR, sG, sB; \
unsigned short Pixel = *((unsigned short *)src); \
sR = (Pixel >> 8) & 0xf8; \
sG = (Pixel >> 3) & 0xfc; \
sB = (Pixel << 3) & 0xf8; \
ASSEMBLE_RGBA(dst, 4, dstfmt, sR, sG, sB, alpha); \
src += 2; \
dst += 4; \
widthvar--; \
}
ONE_PIXEL_BLEND(((UNALIGNED_PTR(dst)) && (width)), width);
/* After all that work, here's the vector part! */
extrawidth = (width % 8); /* trailing unaligned stores */
width -= extrawidth;
vsrc = vec_ld(0, src);
valigner = VEC_ALIGNER(src);
while (width) {
vector unsigned short vR, vG, vB;
vector unsigned char vdst1, vdst2;
voverflow = vec_ld(15, src);
vsrc = vec_perm(vsrc, voverflow, valigner);
vR = vec_and((vector unsigned short)vsrc, vf800);
vB = vec_sl((vector unsigned short)vsrc, v3);
vG = vec_sl(vB, v2);
vdst1 =
(vector unsigned char)vec_perm((vector unsigned char)vR,
valpha, vredalpha1);
vdst1 = vec_perm(vdst1, (vector unsigned char)vB, vblue1);
vdst1 = vec_perm(vdst1, (vector unsigned char)vG, vgreen1);
vdst1 = vec_perm(vdst1, valpha, vpermute);
vec_st(vdst1, 0, dst);
vdst2 =
(vector unsigned char)vec_perm((vector unsigned char)vR,
valpha, vredalpha2);
vdst2 = vec_perm(vdst2, (vector unsigned char)vB, vblue2);
vdst2 = vec_perm(vdst2, (vector unsigned char)vG, vgreen2);
vdst2 = vec_perm(vdst2, valpha, vpermute);
vec_st(vdst2, 16, dst);
width -= 8;
dst += 32;
src += 16;
vsrc = voverflow;
}
SDL_assert(width == 0);
/* do scalar until we can align... */
ONE_PIXEL_BLEND((extrawidth), extrawidth);
#undef ONE_PIXEL_BLEND
src += srcskip; /* move to next row, accounting for pitch. */
dst += dstskip;
}
}
static void Blit_RGB555_32Altivec(SDL_BlitInfo *info)
{
int height = info->dst_h;
Uint8 *src = (Uint8 *)info->src;
int srcskip = info->src_skip;
Uint8 *dst = (Uint8 *)info->dst;
int dstskip = info->dst_skip;
SDL_PixelFormat *srcfmt = info->src_fmt;
SDL_PixelFormat *dstfmt = info->dst_fmt;
unsigned alpha;
vector unsigned char valpha;
vector unsigned char vpermute;
vector unsigned short vf800;
vector unsigned int v8 = vec_splat_u32(8);
vector unsigned int v16 = vec_add(v8, v8);
vector unsigned short v1 = vec_splat_u16(1);
vector unsigned short v3 = vec_splat_u16(3);
/*
0x10 - 0x1f is the alpha
0x00 - 0x0e evens are the red
0x01 - 0x0f odds are zero
*/
vector unsigned char vredalpha1 = VECUINT8_LITERAL(0x10, 0x00, 0x01, 0x01,
0x10, 0x02, 0x01, 0x01,
0x10, 0x04, 0x01, 0x01,
0x10, 0x06, 0x01,
0x01);
vector unsigned char vredalpha2 =
(vector unsigned char)(vec_add((vector unsigned int)vredalpha1, vec_sl(v8, v16)));
/*
0x00 - 0x0f is ARxx ARxx ARxx ARxx
0x11 - 0x0f odds are blue
*/
vector unsigned char vblue1 = VECUINT8_LITERAL(0x00, 0x01, 0x02, 0x11,
0x04, 0x05, 0x06, 0x13,
0x08, 0x09, 0x0a, 0x15,
0x0c, 0x0d, 0x0e, 0x17);
vector unsigned char vblue2 =
(vector unsigned char)(vec_add((vector unsigned int)vblue1, v8));
/*
0x00 - 0x0f is ARxB ARxB ARxB ARxB
0x10 - 0x0e evens are green
*/
vector unsigned char vgreen1 = VECUINT8_LITERAL(0x00, 0x01, 0x10, 0x03,
0x04, 0x05, 0x12, 0x07,
0x08, 0x09, 0x14, 0x0b,
0x0c, 0x0d, 0x16, 0x0f);
vector unsigned char vgreen2 =
(vector unsigned char)(vec_add((vector unsigned int)vgreen1, vec_sl(v8, v8)));
SDL_assert(srcfmt->BytesPerPixel == 2);
SDL_assert(dstfmt->BytesPerPixel == 4);
vf800 = (vector unsigned short)vec_splat_u8(-7);
vf800 = vec_sl(vf800, vec_splat_u16(8));
if (dstfmt->Amask && info->a) {
((unsigned char *)&valpha)[0] = alpha = info->a;
valpha = vec_splat(valpha, 0);
} else {
alpha = 0;
valpha = vec_splat_u8(0);
}
vpermute = calc_swizzle32(NULL, dstfmt);
while (height--) {
vector unsigned char valigner;
vector unsigned char voverflow;
vector unsigned char vsrc;
int width = info->dst_w;
int extrawidth;
/* do scalar until we can align... */
#define ONE_PIXEL_BLEND(condition, widthvar) \
while (condition) { \
unsigned sR, sG, sB; \
unsigned short Pixel = *((unsigned short *)src); \
sR = (Pixel >> 7) & 0xf8; \
sG = (Pixel >> 2) & 0xf8; \
sB = (Pixel << 3) & 0xf8; \
ASSEMBLE_RGBA(dst, 4, dstfmt, sR, sG, sB, alpha); \
src += 2; \
dst += 4; \
widthvar--; \
}
ONE_PIXEL_BLEND(((UNALIGNED_PTR(dst)) && (width)), width);
/* After all that work, here's the vector part! */
extrawidth = (width % 8); /* trailing unaligned stores */
width -= extrawidth;
vsrc = vec_ld(0, src);
valigner = VEC_ALIGNER(src);
while (width) {
vector unsigned short vR, vG, vB;
vector unsigned char vdst1, vdst2;
voverflow = vec_ld(15, src);
vsrc = vec_perm(vsrc, voverflow, valigner);
vR = vec_and(vec_sl((vector unsigned short)vsrc, v1), vf800);
vB = vec_sl((vector unsigned short)vsrc, v3);
vG = vec_sl(vB, v3);
vdst1 =
(vector unsigned char)vec_perm((vector unsigned char)vR,
valpha, vredalpha1);
vdst1 = vec_perm(vdst1, (vector unsigned char)vB, vblue1);
vdst1 = vec_perm(vdst1, (vector unsigned char)vG, vgreen1);
vdst1 = vec_perm(vdst1, valpha, vpermute);
vec_st(vdst1, 0, dst);
vdst2 =
(vector unsigned char)vec_perm((vector unsigned char)vR,
valpha, vredalpha2);
vdst2 = vec_perm(vdst2, (vector unsigned char)vB, vblue2);
vdst2 = vec_perm(vdst2, (vector unsigned char)vG, vgreen2);
vdst2 = vec_perm(vdst2, valpha, vpermute);
vec_st(vdst2, 16, dst);
width -= 8;
dst += 32;
src += 16;
vsrc = voverflow;
}
SDL_assert(width == 0);
/* do scalar until we can align... */
ONE_PIXEL_BLEND((extrawidth), extrawidth);
#undef ONE_PIXEL_BLEND
src += srcskip; /* move to next row, accounting for pitch. */
dst += dstskip;
}
}
static void BlitNtoNKey(SDL_BlitInfo *info);
static void BlitNtoNKeyCopyAlpha(SDL_BlitInfo *info);
static void Blit32to32KeyAltivec(SDL_BlitInfo *info)
{
int height = info->dst_h;
Uint32 *srcp = (Uint32 *)info->src;
int srcskip = info->src_skip / 4;
Uint32 *dstp = (Uint32 *)info->dst;
int dstskip = info->dst_skip / 4;
SDL_PixelFormat *srcfmt = info->src_fmt;
int srcbpp = srcfmt->BytesPerPixel;
SDL_PixelFormat *dstfmt = info->dst_fmt;
int dstbpp = dstfmt->BytesPerPixel;
int copy_alpha = (srcfmt->Amask && dstfmt->Amask);
unsigned alpha = dstfmt->Amask ? info->a : 0;
Uint32 rgbmask = srcfmt->Rmask | srcfmt->Gmask | srcfmt->Bmask;
Uint32 ckey = info->colorkey;
vector unsigned int valpha;
vector unsigned char vpermute;
vector unsigned char vzero;
vector unsigned int vckey;
vector unsigned int vrgbmask;
vpermute = calc_swizzle32(srcfmt, dstfmt);
if (info->dst_w < 16) {
if (copy_alpha) {
BlitNtoNKeyCopyAlpha(info);
} else {
BlitNtoNKey(info);
}
return;
}
vzero = vec_splat_u8(0);
if (alpha) {
((unsigned char *)&valpha)[0] = (unsigned char)alpha;
valpha =
(vector unsigned int)vec_splat((vector unsigned char)valpha, 0);
} else {
valpha = (vector unsigned int)vzero;
}
ckey &= rgbmask;
((unsigned int *)(char *)&vckey)[0] = ckey;
vckey = vec_splat(vckey, 0);
((unsigned int *)(char *)&vrgbmask)[0] = rgbmask;
vrgbmask = vec_splat(vrgbmask, 0);
#if SDL_BYTEORDER == SDL_LIL_ENDIAN
/* reorder bytes for PowerPC little endian */
vpermute = reorder_ppc64le_vec(vpermute);
#endif
while (height--) {
#define ONE_PIXEL_BLEND(condition, widthvar) \
if (copy_alpha) { \
while (condition) { \
Uint32 Pixel; \
unsigned sR, sG, sB, sA; \
DISEMBLE_RGBA((Uint8 *)srcp, srcbpp, srcfmt, Pixel, \
sR, sG, sB, sA); \
if ((Pixel & rgbmask) != ckey) { \
ASSEMBLE_RGBA((Uint8 *)dstp, dstbpp, dstfmt, \
sR, sG, sB, sA); \
} \
dstp = (Uint32 *)(((Uint8 *)dstp) + dstbpp); \
srcp = (Uint32 *)(((Uint8 *)srcp) + srcbpp); \
widthvar--; \
} \
} else { \
while (condition) { \
Uint32 Pixel; \
unsigned sR, sG, sB; \
RETRIEVE_RGB_PIXEL((Uint8 *)srcp, srcbpp, Pixel); \
if (Pixel != ckey) { \
RGB_FROM_PIXEL(Pixel, srcfmt, sR, sG, sB); \
ASSEMBLE_RGBA((Uint8 *)dstp, dstbpp, dstfmt, \
sR, sG, sB, alpha); \
} \
dstp = (Uint32 *)(((Uint8 *)dstp) + dstbpp); \
srcp = (Uint32 *)(((Uint8 *)srcp) + srcbpp); \
widthvar--; \
} \
}
int width = info->dst_w;
ONE_PIXEL_BLEND((UNALIGNED_PTR(dstp)) && (width), width);
SDL_assert(width > 0);
if (width > 0) {
int extrawidth = (width % 4);
vector unsigned char valigner = VEC_ALIGNER(srcp);
vector unsigned int vs = vec_ld(0, srcp);
width -= extrawidth;
SDL_assert(width >= 4);
while (width) {
vector unsigned char vsel;
vector unsigned int vd;
vector unsigned int voverflow = vec_ld(15, srcp);
/* load the source vec */
vs = vec_perm(vs, voverflow, valigner);
/* vsel is set for items that match the key */
vsel = (vector unsigned char)vec_and(vs, vrgbmask);
vsel = (vector unsigned char)vec_cmpeq(vs, vckey);
/* permute the src vec to the dest format */
vs = vec_perm(vs, valpha, vpermute);
/* load the destination vec */
vd = vec_ld(0, dstp);
/* select the source and dest into vs */
vd = (vector unsigned int)vec_sel((vector unsigned char)vs,
(vector unsigned char)vd,
vsel);
vec_st(vd, 0, dstp);
srcp += 4;
width -= 4;
dstp += 4;
vs = voverflow;
}
ONE_PIXEL_BLEND((extrawidth), extrawidth);
#undef ONE_PIXEL_BLEND
srcp += srcskip;
dstp += dstskip;
}
}
}
/* Altivec code to swizzle one 32-bit surface to a different 32-bit format. */
/* Use this on a G5 */
static void ConvertAltivec32to32_noprefetch(SDL_BlitInfo *info)
{
int height = info->dst_h;
Uint32 *src = (Uint32 *)info->src;
int srcskip = info->src_skip / 4;
Uint32 *dst = (Uint32 *)info->dst;
int dstskip = info->dst_skip / 4;
SDL_PixelFormat *srcfmt = info->src_fmt;
SDL_PixelFormat *dstfmt = info->dst_fmt;
vector unsigned int vzero = vec_splat_u32(0);
vector unsigned char vpermute = calc_swizzle32(srcfmt, dstfmt);
if (dstfmt->Amask && !srcfmt->Amask) {
if (info->a) {
vector unsigned char valpha;
((unsigned char *)&valpha)[0] = info->a;
vzero = (vector unsigned int)vec_splat(valpha, 0);
}
}
SDL_assert(srcfmt->BytesPerPixel == 4);
SDL_assert(dstfmt->BytesPerPixel == 4);
#if SDL_BYTEORDER == SDL_LIL_ENDIAN
/* reorder bytes for PowerPC little endian */
vpermute = reorder_ppc64le_vec(vpermute);
#endif
while (height--) {
vector unsigned char valigner;
vector unsigned int vbits;
vector unsigned int voverflow;
Uint32 bits;
Uint8 r, g, b, a;
int width = info->dst_w;
int extrawidth;
/* do scalar until we can align... */
while ((UNALIGNED_PTR(dst)) && (width)) {
bits = *(src++);
RGBA_FROM_8888(bits, srcfmt, r, g, b, a);
if (!srcfmt->Amask)
a = info->a;
*(dst++) = MAKE8888(dstfmt, r, g, b, a);
width--;
}
/* After all that work, here's the vector part! */
extrawidth = (width % 4);
width -= extrawidth;
valigner = VEC_ALIGNER(src);
vbits = vec_ld(0, src);
while (width) {
voverflow = vec_ld(15, src);
src += 4;
width -= 4;
vbits = vec_perm(vbits, voverflow, valigner); /* src is ready. */
vbits = vec_perm(vbits, vzero, vpermute); /* swizzle it. */
vec_st(vbits, 0, dst); /* store it back out. */
dst += 4;
vbits = voverflow;
}
SDL_assert(width == 0);
/* cover pixels at the end of the row that didn't fit in 16 bytes. */
while (extrawidth) {
bits = *(src++); /* max 7 pixels, don't bother with prefetch. */
RGBA_FROM_8888(bits, srcfmt, r, g, b, a);
if (!srcfmt->Amask)
a = info->a;
*(dst++) = MAKE8888(dstfmt, r, g, b, a);
extrawidth--;
}
src += srcskip;
dst += dstskip;
}
}
/* Altivec code to swizzle one 32-bit surface to a different 32-bit format. */
/* Use this on a G4 */
static void ConvertAltivec32to32_prefetch(SDL_BlitInfo *info)
{
const int scalar_dst_lead = sizeof(Uint32) * 4;
const int vector_dst_lead = sizeof(Uint32) * 16;
int height = info->dst_h;
Uint32 *src = (Uint32 *)info->src;
int srcskip = info->src_skip / 4;
Uint32 *dst = (Uint32 *)info->dst;
int dstskip = info->dst_skip / 4;
SDL_PixelFormat *srcfmt = info->src_fmt;
SDL_PixelFormat *dstfmt = info->dst_fmt;
vector unsigned int vzero = vec_splat_u32(0);
vector unsigned char vpermute = calc_swizzle32(srcfmt, dstfmt);
if (dstfmt->Amask && !srcfmt->Amask) {
if (info->a) {
vector unsigned char valpha;
((unsigned char *)&valpha)[0] = info->a;
vzero = (vector unsigned int)vec_splat(valpha, 0);
}
}
SDL_assert(srcfmt->BytesPerPixel == 4);
SDL_assert(dstfmt->BytesPerPixel == 4);
#if SDL_BYTEORDER == SDL_LIL_ENDIAN
/* reorder bytes for PowerPC little endian */
vpermute = reorder_ppc64le_vec(vpermute);
#endif
while (height--) {
vector unsigned char valigner;
vector unsigned int vbits;
vector unsigned int voverflow;
Uint32 bits;
Uint8 r, g, b, a;
int width = info->dst_w;
int extrawidth;
/* do scalar until we can align... */
while ((UNALIGNED_PTR(dst)) && (width)) {
vec_dstt(src + scalar_dst_lead, DST_CTRL(2, 32, 1024),
DST_CHAN_SRC);
vec_dstst(dst + scalar_dst_lead, DST_CTRL(2, 32, 1024),
DST_CHAN_DEST);
bits = *(src++);
RGBA_FROM_8888(bits, srcfmt, r, g, b, a);
if (!srcfmt->Amask)
a = info->a;
*(dst++) = MAKE8888(dstfmt, r, g, b, a);
width--;
}
/* After all that work, here's the vector part! */
extrawidth = (width % 4);
width -= extrawidth;
valigner = VEC_ALIGNER(src);
vbits = vec_ld(0, src);
while (width) {
vec_dstt(src + vector_dst_lead, DST_CTRL(2, 32, 1024),
DST_CHAN_SRC);
vec_dstst(dst + vector_dst_lead, DST_CTRL(2, 32, 1024),
DST_CHAN_DEST);
voverflow = vec_ld(15, src);
src += 4;
width -= 4;
vbits = vec_perm(vbits, voverflow, valigner); /* src is ready. */
vbits = vec_perm(vbits, vzero, vpermute); /* swizzle it. */
vec_st(vbits, 0, dst); /* store it back out. */
dst += 4;
vbits = voverflow;
}
SDL_assert(width == 0);
/* cover pixels at the end of the row that didn't fit in 16 bytes. */
while (extrawidth) {
bits = *(src++); /* max 7 pixels, don't bother with prefetch. */
RGBA_FROM_8888(bits, srcfmt, r, g, b, a);
if (!srcfmt->Amask)
a = info->a;
*(dst++) = MAKE8888(dstfmt, r, g, b, a);
extrawidth--;
}
src += srcskip;
dst += dstskip;
}
vec_dss(DST_CHAN_SRC);
vec_dss(DST_CHAN_DEST);
}
static enum blit_features GetBlitFeatures(void)
{
static enum blit_features features = -1;
if (features == (enum blit_features) - 1) {
/* Provide an override for testing .. */
char *override = SDL_getenv("SDL_ALTIVEC_BLIT_FEATURES");
if (override) {
unsigned int features_as_uint = 0;
SDL_sscanf(override, "%u", &features_as_uint);
features = (enum blit_features)features_as_uint;
} else {
99A1
features = (0
/* Feature 1 is has-MMX */
| ((SDL_HasMMX()) ? BLIT_FEATURE_HAS_MMX : 0)
/* Feature 2 is has-AltiVec */
| ((SDL_HasAltiVec()) ? BLIT_FEATURE_HAS_ALTIVEC : 0)
/* Feature 4 is dont-use-prefetch */
/* !!!! FIXME: Check for G5 or later, not the cache size! Always prefetch on a G4. */
| ((GetL3CacheSize() == 0) ? BLIT_FEATURE_ALTIVEC_DONT_USE_PREFETCH : 0));
}
}
return features;
}
#ifdef __MWERKS__
#pragma altivec_model off
#endif
#else
/* Feature 1 is has-MMX */
#define GetBlitFeatures() ((SDL_HasMMX() ? BLIT_FEATURE_HAS_MMX : 0) | (SDL_HasARMSIMD() ? BLIT_FEATURE_HAS_ARM_SIMD : 0))
#endif
#ifdef SDL_ARM_SIMD_BLITTERS
void Blit_BGR888_RGB888ARMSIMDAsm(int32_t w, int32_t h, uint32_t *dst, int32_t dst_stride, uint32_t *src, int32_t src_stride);
static void Blit_BGR888_RGB888ARMSIMD(SDL_BlitInfo *info)
{
int32_t width = info->dst_w;
int32_t height = info->dst_h;
uint32_t *dstp = (uint32_t *)info->dst;
int32_t dststride = width + (info->dst_skip >> 2);
uint32_t *srcp = (uint32_t *)info->src;
int32_t srcstride = width + (info->src_skip >> 2);
Blit_BGR888_RGB888ARMSIMDAsm(width, height, dstp, dststride, srcp, srcstride);
}
void Blit_RGB444_RGB888ARMSIMDAsm(int32_t w, int32_t h, uint32_t *dst, int32_t dst_stride, uint16_t *src, int32_t src_stride);
static void Blit_RGB444_RGB888ARMSIMD(SDL_BlitInfo *info)
{
int32_t width = info->dst_w;
int32_t height = info->dst_h;
uint32_t *dstp = (uint32_t *)info->dst;
int32_t dststride = width + (info->dst_skip >> 2);
uint16_t *srcp = (uint16_t *)info->src;
int32_t srcstride = width + (info->src_skip >> 1);
Blit_RGB444_RGB888ARMSIMDAsm(width, height, dstp, dststride, srcp, srcstride);
}
#endif
/* This is now endian dependent */
#if SDL_BYTEORDER == SDL_LIL_ENDIAN
#define HI 1
#define LO 0
#else /* SDL_BYTEORDER == SDL_BIG_ENDIAN */
#define HI 0
#define LO 1
#endif
/* Special optimized blit for RGB 8-8-8 --> RGB 3-3-2 */
#define RGB888_RGB332(dst, src) \
{ \
dst = (Uint8)((((src)&0x00E00000) >> 16) | \
(((src)&0x0000E000) >> 11) | \
(((src)&0x000000C0) >> 6)); \
}
static void Blit_RGB888_index8(SDL_BlitInfo *info)
{
#ifndef USE_DUFFS_LOOP
int c;
#endif
int width, height;
Uint32 *src;
const Uint8 *map;
Uint8 *dst;
int srcskip, dstskip;
/* Set up some basic variables */
width = info->dst_w;
height = info->dst_h;
src = (Uint32 *)info->src;
srcskip = info->src_skip / 4;
dst = info->dst;
dstskip = info->dst_skip;
map = info->table;
if (!map) {
while (height--) {
#ifdef USE_DUFFS_LOOP
/* *INDENT-OFF* */ /* clang-format off */
DUFFS_LOOP(
RGB888_RGB332(*dst++, *src);
, width);
/* *INDENT-ON* */ /* clang-format on */
#else
for (c = width / 4; c; --c) {
/* Pack RGB into 8bit pixel */
++src;
RGB888_RGB332(*dst++, *src);
++src;
RGB888_RGB332(*dst++, *src);
++src;
RGB888_RGB332(*dst++, *src);
++src;
}
switch (width & 3) {
case 3: