This repository has been archived by the owner on Feb 11, 2021. It is now read-only.
/
SDL_blit.h
542 lines (504 loc) · 14.9 KB
1
2
/*
SDL - Simple DirectMedia Layer
3
Copyright (C) 1997-2011 Sam Lantinga
4
5
This library is free software; you can redistribute it and/or
6
modify it under the terms of the GNU Lesser General Public
7
License as published by the Free Software Foundation; either
8
version 2.1 of the License, or (at your option) any later version.
9
10
11
12
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13
Lesser General Public License for more details.
14
15
16
17
You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
18
19
Sam Lantinga
20
slouken@libsdl.org
21
*/
22
#include "SDL_config.h"
23
24
25
26
#ifndef _SDL_blit_h
#define _SDL_blit_h
27
28
29
30
31
32
33
#ifdef __MINGW32__
#include <_mingw.h>
#endif
#if defined(__MINGW32__) && defined(__MINGW64_VERSION_MAJOR)
#include <intrin.h>
#else
34
35
#ifdef __MMX__
#include <mmintrin.h>
36
#endif
37
38
39
#ifdef __SSE__
#include <xmmintrin.h>
#endif
40
41
42
#ifdef __SSE2__
#include <emmintrin.h>
#endif
43
#endif
44
45
#include "SDL_cpuinfo.h"
46
#include "SDL_endian.h"
47
#include "SDL_surface.h"
48
49
/* SDL blit copy flags */
50
51
#define SDL_COPY_MODULATE_COLOR 0x00000001
#define SDL_COPY_MODULATE_ALPHA 0x00000002
52
53
#define SDL_COPY_BLEND 0x00000010
#define SDL_COPY_ADD 0x00000020
54
#define SDL_COPY_MOD 0x00000040
55
56
57
58
59
#define SDL_COPY_COLORKEY 0x00000100
#define SDL_COPY_NEAREST 0x00000200
#define SDL_COPY_RLE_DESIRED 0x00001000
#define SDL_COPY_RLE_COLORKEY 0x00002000
#define SDL_COPY_RLE_ALPHAKEY 0x00004000
60
#define SDL_COPY_RLE_MASK (SDL_COPY_RLE_DESIRED|SDL_COPY_RLE_COLORKEY|SDL_COPY_RLE_ALPHAKEY)
61
62
/* SDL blit CPU flags */
63
64
65
66
#define SDL_CPU_ANY 0x00000000
#define SDL_CPU_MMX 0x00000001
#define SDL_CPU_SSE 0x00000004
#define SDL_CPU_SSE2 0x00000008
67
68
69
typedef struct
{
70
71
72
Uint8 *src;
int src_w, src_h;
int src_pitch;
73
int src_skip;
74
75
76
Uint8 *dst;
int dst_w, dst_h;
int dst_pitch;
77
int dst_skip;
78
79
SDL_PixelFormat *src_fmt;
SDL_PixelFormat *dst_fmt;
80
Uint8 *table;
81
82
83
int flags;
Uint32 colorkey;
Uint8 r, g, b, a;
84
85
} SDL_BlitInfo;
86
typedef void (SDLCALL * SDL_BlitFunc) (SDL_BlitInfo * info);
87
88
89
typedef struct
{
90
91
92
93
94
95
Uint32 src_format;
Uint32 dst_format;
int flags;
int cpu;
SDL_BlitFunc func;
} SDL_BlitFuncEntry;
96
97
/* Blit mapping definition */
98
99
100
101
typedef struct SDL_BlitMap
{
SDL_Surface *dst;
int identity;
102
103
SDL_blit blit;
void *data;
104
SDL_BlitInfo info;
105
106
107
108
/* the version count matches the destination; mismatch indicates
an invalid mapping */
unsigned int format_version;
109
110
111
} SDL_BlitMap;
/* Functions found in SDL_blit.c */
112
extern int SDL_CalculateBlit(SDL_Surface * surface);
113
114
115
116
117
118
/* Functions found in SDL_blit_*.c */
extern SDL_BlitFunc SDL_CalculateBlit0(SDL_Surface * surface);
extern SDL_BlitFunc SDL_CalculateBlit1(SDL_Surface * surface);
extern SDL_BlitFunc SDL_CalculateBlitN(SDL_Surface * surface);
extern SDL_BlitFunc SDL_CalculateBlitA(SDL_Surface * surface);
119
120
121
122
123
/*
* Useful macros for blitting routines
*/
124
125
126
#if defined(__GNUC__)
#define DECLARE_ALIGNED(t,v,a) t __attribute__((aligned(a))) v
#elif defined(_MSC_VER)
127
#define DECLARE_ALIGNED(t,v,a) __declspec(align(a)) t v
128
129
130
131
#else
#define DECLARE_ALIGNED(t,v,a) t v
#endif
132
133
134
135
136
137
#define FORMAT_EQUAL(A, B) \
((A)->BitsPerPixel == (B)->BitsPerPixel \
&& ((A)->Rmask == (B)->Rmask) && ((A)->Amask == (B)->Amask))
/* Load pixel of the specified format from a buffer and get its R-G-B values */
/* FIXME: rescale values to 0..255 here? */
138
#define RGB_FROM_PIXEL(Pixel, fmt, r, g, b) \
139
{ \
140
141
142
r = (((Pixel&fmt->Rmask)>>fmt->Rshift)<<fmt->Rloss); \
g = (((Pixel&fmt->Gmask)>>fmt->Gshift)<<fmt->Gloss); \
b = (((Pixel&fmt->Bmask)>>fmt->Bshift)<<fmt->Bloss); \
143
}
144
#define RGB_FROM_RGB565(Pixel, r, g, b) \
145
{ \
146
147
148
r = (((Pixel&0xF800)>>11)<<3); \
g = (((Pixel&0x07E0)>>5)<<2); \
b = ((Pixel&0x001F)<<3); \
149
}
150
#define RGB_FROM_RGB555(Pixel, r, g, b) \
151
{ \
152
153
154
r = (((Pixel&0x7C00)>>10)<<3); \
g = (((Pixel&0x03E0)>>5)<<3); \
b = ((Pixel&0x001F)<<3); \
155
}
156
#define RGB_FROM_RGB888(Pixel, r, g, b) \
157
{ \
158
159
160
r = ((Pixel&0xFF0000)>>16); \
g = ((Pixel&0xFF00)>>8); \
b = (Pixel&0xFF); \
161
}
162
#define RETRIEVE_RGB_PIXEL(buf, bpp, Pixel) \
163
164
165
do { \
switch (bpp) { \
case 2: \
166
Pixel = *((Uint16 *)(buf)); \
167
168
169
170
break; \
\
case 3: { \
Uint8 *B = (Uint8 *)(buf); \
171
if (SDL_BYTEORDER == SDL_LIL_ENDIAN) { \
172
Pixel = B[0] + (B[1] << 8) + (B[2] << 16); \
173
} else { \
174
Pixel = (B[0] << 16) + (B[1] << 8) + B[2]; \
175
176
177
178
179
} \
} \
break; \
\
case 4: \
180
Pixel = *((Uint32 *)(buf)); \
181
182
183
break; \
\
default: \
184
Pixel; /* stop gcc complaints */ \
185
186
break; \
} \
187
} while (0)
188
189
#define DISEMBLE_RGB(buf, bpp, fmt, Pixel, r, g, b) \
190
191
192
do { \
switch (bpp) { \
case 2: \
193
Pixel = *((Uint16 *)(buf)); \
194
RGB_FROM_PIXEL(Pixel, fmt, r, g, b); \
195
196
break; \
\
197
198
199
200
201
case 3: { \
if (SDL_BYTEORDER == SDL_LIL_ENDIAN) { \
r = *((buf)+fmt->Rshift/8); \
g = *((buf)+fmt->Gshift/8); \
b = *((buf)+fmt->Bshift/8); \
202
} else { \
203
204
205
r = *((buf)+2-fmt->Rshift/8); \
g = *((buf)+2-fmt->Gshift/8); \
b = *((buf)+2-fmt->Bshift/8); \
206
207
208
209
210
} \
} \
break; \
\
case 4: \
211
Pixel = *((Uint32 *)(buf)); \
212
RGB_FROM_PIXEL(Pixel, fmt, r, g, b); \
213
214
break; \
\
215
216
default: \
Pixel; /* stop gcc complaints */ \
217
218
break; \
} \
219
} while (0)
220
221
/* Assemble R-G-B values into a specified pixel format and store them */
222
#define PIXEL_FROM_RGB(Pixel, fmt, r, g, b) \
223
{ \
224
Pixel = ((r>>fmt->Rloss)<<fmt->Rshift)| \
225
226
227
((g>>fmt->Gloss)<<fmt->Gshift)| \
((b>>fmt->Bloss)<<fmt->Bshift); \
}
228
#define RGB565_FROM_RGB(Pixel, r, g, b) \
229
{ \
230
Pixel = ((r>>3)<<11)|((g>>2)<<5)|(b>>3); \
231
}
232
#define RGB555_FROM_RGB(Pixel, r, g, b) \
233
{ \
234
Pixel = ((r>>3)<<10)|((g>>3)<<5)|(b>>3); \
235
}
236
#define RGB888_FROM_RGB(Pixel, r, g, b) \
237
{ \
238
Pixel = (r<<16)|(g<<8)|b; \
239
}
240
241
242
243
#define ARGB8888_FROM_RGBA(Pixel, r, g, b, a) \
{ \
Pixel = (a<<24)|(r<<16)|(g<<8)|b; \
}
244
245
246
247
248
249
250
251
252
253
254
255
#define RGBA8888_FROM_RGBA(Pixel, r, g, b, a) \
{ \
Pixel = (r<<24)|(g<<16)|(b<<8)|a; \
}
#define ABGR8888_FROM_RGBA(Pixel, r, g, b, a) \
{ \
Pixel = (a<<24)|(b<<16)|(g<<8)|r; \
}
#define BGRA8888_FROM_RGBA(Pixel, r, g, b, a) \
{ \
Pixel = (b<<24)|(g<<16)|(r<<8)|a; \
}
256
257
258
259
#define ASSEMBLE_RGB(buf, bpp, fmt, r, g, b) \
{ \
switch (bpp) { \
case 2: { \
260
Uint16 Pixel; \
261
\
262
263
PIXEL_FROM_RGB(Pixel, fmt, r, g, b); \
*((Uint16 *)(buf)) = Pixel; \
264
265
266
267
} \
break; \
\
case 3: { \
268
if (SDL_BYTEORDER == SDL_LIL_ENDIAN) { \
269
270
271
272
273
274
275
276
277
278
279
280
*((buf)+fmt->Rshift/8) = r; \
*((buf)+fmt->Gshift/8) = g; \
*((buf)+fmt->Bshift/8) = b; \
} else { \
*((buf)+2-fmt->Rshift/8) = r; \
*((buf)+2-fmt->Gshift/8) = g; \
*((buf)+2-fmt->Bshift/8) = b; \
} \
} \
break; \
\
case 4: { \
281
Uint32 Pixel; \
282
\
283
284
PIXEL_FROM_RGB(Pixel, fmt, r, g, b); \
*((Uint32 *)(buf)) = Pixel; \
285
286
287
288
289
290
291
292
293
} \
break; \
} \
}
#define ASSEMBLE_RGB_AMASK(buf, bpp, fmt, r, g, b, Amask) \
{ \
switch (bpp) { \
case 2: { \
Uint16 *bufp; \
294
Uint16 Pixel; \
295
296
\
bufp = (Uint16 *)buf; \
297
298
PIXEL_FROM_RGB(Pixel, fmt, r, g, b); \
*bufp = Pixel | (*bufp & Amask); \
299
300
301
302
} \
break; \
\
case 3: { \
303
if (SDL_BYTEORDER == SDL_LIL_ENDIAN) { \
304
305
306
307
308
309
310
311
312
313
314
315
316
*((buf)+fmt->Rshift/8) = r; \
*((buf)+fmt->Gshift/8) = g; \
*((buf)+fmt->Bshift/8) = b; \
} else { \
*((buf)+2-fmt->Rshift/8) = r; \
*((buf)+2-fmt->Gshift/8) = g; \
*((buf)+2-fmt->Bshift/8) = b; \
} \
} \
break; \
\
case 4: { \
Uint32 *bufp; \
317
Uint32 Pixel; \
318
319
\
bufp = (Uint32 *)buf; \
320
321
PIXEL_FROM_RGB(Pixel, fmt, r, g, b); \
*bufp = Pixel | (*bufp & Amask); \
322
323
324
325
326
327
} \
break; \
} \
}
/* FIXME: Should we rescale alpha into 0..255 here? */
328
#define RGBA_FROM_PIXEL(Pixel, fmt, r, g, b, a) \
329
{ \
330
331
332
333
r = ((Pixel&fmt->Rmask)>>fmt->Rshift)<<fmt->Rloss; \
g = ((Pixel&fmt->Gmask)>>fmt->Gshift)<<fmt->Gloss; \
b = ((Pixel&fmt->Bmask)>>fmt->Bshift)<<fmt->Bloss; \
a = ((Pixel&fmt->Amask)>>fmt->Ashift)<<fmt->Aloss; \
334
}
335
#define RGBA_FROM_8888(Pixel, fmt, r, g, b, a) \
336
{ \
337
338
339
340
r = (Pixel&fmt->Rmask)>>fmt->Rshift; \
g = (Pixel&fmt->Gmask)>>fmt->Gshift; \
b = (Pixel&fmt->Bmask)>>fmt->Bshift; \
a = (Pixel&fmt->Amask)>>fmt->Ashift; \
341
}
342
#define RGBA_FROM_RGBA8888(Pixel, r, g, b, a) \
343
{ \
344
345
346
347
r = (Pixel>>24); \
g = ((Pixel>>16)&0xFF); \
b = ((Pixel>>8)&0xFF); \
a = (Pixel&0xFF); \
348
}
349
#define RGBA_FROM_ARGB8888(Pixel, r, g, b, a) \
350
{ \
351
352
353
354
r = ((Pixel>>16)&0xFF); \
g = ((Pixel>>8)&0xFF); \
b = (Pixel&0xFF); \
a = (Pixel>>24); \
355
}
356
#define RGBA_FROM_ABGR8888(Pixel, r, g, b, a) \
357
{ \
358
359
360
361
r = (Pixel&0xFF); \
g = ((Pixel>>8)&0xFF); \
b = ((Pixel>>16)&0xFF); \
a = (Pixel>>24); \
362
}
363
364
365
366
367
368
369
#define RGBA_FROM_BGRA8888(Pixel, r, g, b, a) \
{ \
r = ((Pixel>>8)&0xFF); \
g = ((Pixel>>16)&0xFF); \
b = (Pixel>>24); \
a = (Pixel&0xFF); \
}
370
#define DISEMBLE_RGBA(buf, bpp, fmt, Pixel, r, g, b, a) \
371
372
373
do { \
switch (bpp) { \
case 2: \
374
Pixel = *((Uint16 *)(buf)); \
375
RGBA_FROM_PIXEL(Pixel, fmt, r, g, b, a); \
376
377
break; \
\
378
379
380
381
382
case 3: { \
if (SDL_BYTEORDER == SDL_LIL_ENDIAN) { \
r = *((buf)+fmt->Rshift/8); \
g = *((buf)+fmt->Gshift/8); \
b = *((buf)+fmt->Bshift/8); \
383
} else { \
384
385
386
r = *((buf)+2-fmt->Rshift/8); \
g = *((buf)+2-fmt->Gshift/8); \
b = *((buf)+2-fmt->Bshift/8); \
387
} \
388
a = 0xFF; \
389
390
391
392
} \
break; \
\
case 4: \
393
Pixel = *((Uint32 *)(buf)); \
394
RGBA_FROM_PIXEL(Pixel, fmt, r, g, b, a); \
395
396
397
break; \
\
default: \
398
Pixel; /* stop gcc complaints */ \
399
400
break; \
} \
401
} while (0)
402
403
/* FIXME: this isn't correct, especially for Alpha (maximum != 255) */
404
#define PIXEL_FROM_RGBA(Pixel, fmt, r, g, b, a) \
405
{ \
406
Pixel = ((r>>fmt->Rloss)<<fmt->Rshift)| \
407
408
((g>>fmt->Gloss)<<fmt->Gshift)| \
((b>>fmt->Bloss)<<fmt->Bshift)| \
409
((a>>fmt->Aloss)<<fmt->Ashift); \
410
411
412
413
414
}
#define ASSEMBLE_RGBA(buf, bpp, fmt, r, g, b, a) \
{ \
switch (bpp) { \
case 2: { \
415
Uint16 Pixel; \
416
\
417
418
PIXEL_FROM_RGBA(Pixel, fmt, r, g, b, a); \
*((Uint16 *)(buf)) = Pixel; \
419
420
421
} \
break; \
\
422
423
case 3: { \
if (SDL_BYTEORDER == SDL_LIL_ENDIAN) { \
424
425
426
427
428
429
430
431
432
433
434
435
*((buf)+fmt->Rshift/8) = r; \
*((buf)+fmt->Gshift/8) = g; \
*((buf)+fmt->Bshift/8) = b; \
} else { \
*((buf)+2-fmt->Rshift/8) = r; \
*((buf)+2-fmt->Gshift/8) = g; \
*((buf)+2-fmt->Bshift/8) = b; \
} \
} \
break; \
\
case 4: { \
436
Uint32 Pixel; \
437
\
438
439
PIXEL_FROM_RGBA(Pixel, fmt, r, g, b, a); \
*((Uint32 *)(buf)) = Pixel; \
440
441
442
443
444
} \
break; \
} \
}
445
/* Blend the RGB values of two Pixels based on a source alpha value */
446
447
#define ALPHA_BLEND(sR, sG, sB, A, dR, dG, dB) \
do { \
448
449
450
dR = ((((int)(sR-dR)*(int)A)/255)+dR); \
dG = ((((int)(sG-dG)*(int)A)/255)+dG); \
dB = ((((int)(sB-dB)*(int)A)/255)+dB); \
451
452
453
} while(0)
454
/* This is a very useful loop for optimizing blitters */
455
456
457
#if defined(_MSC_VER) && (_MSC_VER == 1300)
/* There's a bug in the Visual C++ 7 optimizer when compiling this code */
#else
458
#define USE_DUFFS_LOOP
459
#endif
460
461
462
463
464
#ifdef USE_DUFFS_LOOP
/* 8-times unrolled loop */
#define DUFFS_LOOP8(pixel_copy_increment, width) \
{ int n = (width+7)/8; \
465
switch (width & 7) { \
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
case 0: do { pixel_copy_increment; \
case 7: pixel_copy_increment; \
case 6: pixel_copy_increment; \
case 5: pixel_copy_increment; \
case 4: pixel_copy_increment; \
case 3: pixel_copy_increment; \
case 2: pixel_copy_increment; \
case 1: pixel_copy_increment; \
} while ( --n > 0 ); \
} \
}
/* 4-times unrolled loop */
#define DUFFS_LOOP4(pixel_copy_increment, width) \
{ int n = (width+3)/4; \
481
switch (width & 3) { \
482
483
484
485
case 0: do { pixel_copy_increment; \
case 3: pixel_copy_increment; \
case 2: pixel_copy_increment; \
case 1: pixel_copy_increment; \
486
} while (--n > 0); \
487
488
489
} \
}
490
491
492
493
/* Use the 8-times version of the loop by default */
#define DUFFS_LOOP(pixel_copy_increment, width) \
DUFFS_LOOP8(pixel_copy_increment, width)
494
495
496
497
498
499
500
501
502
503
/* Special version of Duff's device for even more optimization */
#define DUFFS_LOOP_124(pixel_copy_increment1, \
pixel_copy_increment2, \
pixel_copy_increment4, width) \
{ int n = width; \
if (n & 1) { \
pixel_copy_increment1; n -= 1; \
} \
if (n & 2) { \
pixel_copy_increment2; n -= 2; \
504
} \
505
506
507
508
509
510
511
if (n) { \
n = (n+7)/ 8; \
switch (n & 4) { \
case 0: do { pixel_copy_increment4; \
case 4: pixel_copy_increment4; \
} while (--n > 0); \
} \
512
513
514
} \
}
515
516
#else
517
518
519
520
521
522
523
524
525
526
527
/* Don't use Duff's device to unroll loops */
#define DUFFS_LOOP(pixel_copy_increment, width) \
{ int n; \
for ( n=width; n > 0; --n ) { \
pixel_copy_increment; \
} \
}
#define DUFFS_LOOP8(pixel_copy_increment, width) \
DUFFS_LOOP(pixel_copy_increment, width)
#define DUFFS_LOOP4(pixel_copy_increment, width) \
DUFFS_LOOP(pixel_copy_increment, width)
528
529
530
531
#define DUFFS_LOOP_124(pixel_copy_increment1, \
pixel_copy_increment2, \
pixel_copy_increment4, width) \
DUFFS_LOOP(pixel_copy_increment1, width)
532
533
534
535
536
537
538
539
540
#endif /* USE_DUFFS_LOOP */
/* Prevent Visual C++ 6.0 from printing out stupid warnings */
#if defined(_MSC_VER) && (_MSC_VER >= 600)
#pragma warning(disable: 4550)
#endif
#endif /* _SDL_blit_h */
541
542
/* vi: set ts=4 sw=4 expandtab: */