src/video/SDL_blit.h
author Sam Lantinga <slouken@libsdl.org>
Fri, 04 Feb 2011 19:50:56 -0800
changeset 5184 d976b67150c5
parent 5163 d72793305335
child 5259 6a65c1fc07af
permissions -rw-r--r--
Restored SDL_BLENDMODE_MOD for MAME
slouken@0
     1
/*
slouken@0
     2
    SDL - Simple DirectMedia Layer
slouken@3697
     3
    Copyright (C) 1997-2010 Sam Lantinga
slouken@0
     4
slouken@0
     5
    This library is free software; you can redistribute it and/or
slouken@1312
     6
    modify it under the terms of the GNU Lesser General Public
slouken@0
     7
    License as published by the Free Software Foundation; either
slouken@1312
     8
    version 2.1 of the License, or (at your option) any later version.
slouken@0
     9
slouken@0
    10
    This library is distributed in the hope that it will be useful,
slouken@0
    11
    but WITHOUT ANY WARRANTY; without even the implied warranty of
slouken@0
    12
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
slouken@1312
    13
    Lesser General Public License for more details.
slouken@0
    14
slouken@1312
    15
    You should have received a copy of the GNU Lesser General Public
slouken@1312
    16
    License along with this library; if not, write to the Free Software
slouken@1312
    17
    Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
slouken@0
    18
slouken@0
    19
    Sam Lantinga
slouken@252
    20
    slouken@libsdl.org
slouken@0
    21
*/
slouken@1402
    22
#include "SDL_config.h"
slouken@0
    23
slouken@0
    24
#ifndef _SDL_blit_h
slouken@0
    25
#define _SDL_blit_h
slouken@0
    26
slouken@4426
    27
#ifdef __MINGW32__
slouken@4426
    28
#include <_mingw.h>
slouken@4426
    29
#endif
slouken@4426
    30
slouken@4426
    31
#if defined(__MINGW32__) && defined(__MINGW64_VERSION_MAJOR)
slouken@4426
    32
#include <intrin.h>
slouken@4426
    33
#else
slouken@2249
    34
#ifdef __MMX__
slouken@2249
    35
#include <mmintrin.h>
slouken@2249
    36
#endif
slouken@2261
    37
#ifdef __3dNOW__
slouken@2261
    38
#include <mm3dnow.h>
slouken@2261
    39
#endif
slouken@2249
    40
#ifdef __SSE__
slouken@2249
    41
#include <xmmintrin.h>
slouken@2249
    42
#endif
slouken@2262
    43
#ifdef __SSE2__
slouken@2262
    44
#include <emmintrin.h>
slouken@2262
    45
#endif
slouken@4426
    46
#endif
slouken@2249
    47
slouken@2250
    48
#include "SDL_cpuinfo.h"
slouken@0
    49
#include "SDL_endian.h"
slouken@5163
    50
#include "SDL_surface.h"
slouken@0
    51
slouken@2262
    52
/* SDL blit copy flags */
slouken@2266
    53
#define SDL_COPY_MODULATE_COLOR     0x00000001
slouken@2266
    54
#define SDL_COPY_MODULATE_ALPHA     0x00000002
slouken@5140
    55
#define SDL_COPY_BLEND              0x00000010
slouken@5140
    56
#define SDL_COPY_ADD                0x00000020
slouken@5184
    57
#define SDL_COPY_MOD                0x00000040
slouken@2266
    58
#define SDL_COPY_COLORKEY           0x00000100
slouken@2266
    59
#define SDL_COPY_NEAREST            0x00000200
slouken@2266
    60
#define SDL_COPY_RLE_DESIRED        0x00001000
slouken@2266
    61
#define SDL_COPY_RLE_COLORKEY       0x00002000
slouken@2266
    62
#define SDL_COPY_RLE_ALPHAKEY       0x00004000
slouken@2853
    63
#define SDL_COPY_RLE_MASK           (SDL_COPY_RLE_DESIRED|SDL_COPY_RLE_COLORKEY|SDL_COPY_RLE_ALPHAKEY)
slouken@2262
    64
slouken@2262
    65
/* SDL blit CPU flags */
slouken@2266
    66
#define SDL_CPU_ANY                 0x00000000
slouken@2266
    67
#define SDL_CPU_MMX                 0x00000001
slouken@2266
    68
#define SDL_CPU_3DNOW               0x00000002
slouken@2266
    69
#define SDL_CPU_SSE                 0x00000004
slouken@2266
    70
#define SDL_CPU_SSE2                0x00000008
slouken@2266
    71
#define SDL_CPU_ALTIVEC_PREFETCH    0x00000010
slouken@2266
    72
#define SDL_CPU_ALTIVEC_NOPREFETCH  0x00000020
slouken@2262
    73
slouken@2267
    74
typedef struct
slouken@2267
    75
{
slouken@2262
    76
    Uint8 *src;
slouken@2262
    77
    int src_w, src_h;
slouken@2262
    78
    int src_pitch;
slouken@2267
    79
    int src_skip;
slouken@2262
    80
    Uint8 *dst;
slouken@2262
    81
    int dst_w, dst_h;
slouken@2262
    82
    int dst_pitch;
slouken@2267
    83
    int dst_skip;
slouken@2262
    84
    SDL_PixelFormat *src_fmt;
slouken@2262
    85
    SDL_PixelFormat *dst_fmt;
slouken@1895
    86
    Uint8 *table;
slouken@2262
    87
    int flags;
slouken@2262
    88
    Uint32 colorkey;
slouken@2262
    89
    Uint8 r, g, b, a;
slouken@0
    90
} SDL_BlitInfo;
slouken@0
    91
slouken@2267
    92
typedef void (SDLCALL * SDL_BlitFunc) (SDL_BlitInfo * info);
slouken@2262
    93
slouken@2267
    94
typedef struct
slouken@2267
    95
{
slouken@2262
    96
    Uint32 src_format;
slouken@2262
    97
    Uint32 dst_format;
slouken@2262
    98
    int flags;
slouken@2262
    99
    int cpu;
slouken@2262
   100
    SDL_BlitFunc func;
slouken@2262
   101
} SDL_BlitFuncEntry;
slouken@0
   102
slouken@0
   103
/* Blit mapping definition */
slouken@1895
   104
typedef struct SDL_BlitMap
slouken@1895
   105
{
slouken@1895
   106
    SDL_Surface *dst;
slouken@1895
   107
    int identity;
slouken@2257
   108
    SDL_blit blit;
slouken@2257
   109
    void *data;
slouken@2262
   110
    SDL_BlitInfo info;
slouken@0
   111
slouken@1895
   112
    /* the version count matches the destination; mismatch indicates
slouken@1895
   113
       an invalid mapping */
slouken@1895
   114
    unsigned int format_version;
slouken@0
   115
} SDL_BlitMap;
slouken@0
   116
slouken@0
   117
/* Functions found in SDL_blit.c */
slouken@1895
   118
extern int SDL_CalculateBlit(SDL_Surface * surface);
slouken@0
   119
slouken@2267
   120
/* Functions found in SDL_blit_*.c */
slouken@2267
   121
extern SDL_BlitFunc SDL_CalculateBlit0(SDL_Surface * surface);
slouken@2267
   122
extern SDL_BlitFunc SDL_CalculateBlit1(SDL_Surface * surface);
slouken@2267
   123
extern SDL_BlitFunc SDL_CalculateBlitN(SDL_Surface * surface);
slouken@2267
   124
extern SDL_BlitFunc SDL_CalculateBlitA(SDL_Surface * surface);
slouken@2263
   125
slouken@0
   126
/*
slouken@0
   127
 * Useful macros for blitting routines
slouken@0
   128
 */
slouken@0
   129
slouken@2249
   130
#if defined(__GNUC__)
slouken@2249
   131
#define DECLARE_ALIGNED(t,v,a)  t __attribute__((aligned(a))) v
slouken@2249
   132
#elif defined(_MSC_VER)
slouken@2251
   133
#define DECLARE_ALIGNED(t,v,a)  __declspec(align(a)) t v
slouken@2249
   134
#else
slouken@2249
   135
#define DECLARE_ALIGNED(t,v,a)  t v
slouken@2249
   136
#endif
slouken@2249
   137
slouken@0
   138
#define FORMAT_EQUAL(A, B)						\
slouken@0
   139
    ((A)->BitsPerPixel == (B)->BitsPerPixel				\
slouken@0
   140
     && ((A)->Rmask == (B)->Rmask) && ((A)->Amask == (B)->Amask))
slouken@0
   141
slouken@0
   142
/* Load pixel of the specified format from a buffer and get its R-G-B values */
slouken@0
   143
/* FIXME: rescale values to 0..255 here? */
icculus@1162
   144
#define RGB_FROM_PIXEL(Pixel, fmt, r, g, b)				\
slouken@0
   145
{									\
icculus@1162
   146
	r = (((Pixel&fmt->Rmask)>>fmt->Rshift)<<fmt->Rloss); 		\
icculus@1162
   147
	g = (((Pixel&fmt->Gmask)>>fmt->Gshift)<<fmt->Gloss); 		\
icculus@1162
   148
	b = (((Pixel&fmt->Bmask)>>fmt->Bshift)<<fmt->Bloss); 		\
slouken@0
   149
}
icculus@1162
   150
#define RGB_FROM_RGB565(Pixel, r, g, b)					\
slouken@0
   151
{									\
icculus@1162
   152
	r = (((Pixel&0xF800)>>11)<<3);		 			\
icculus@1162
   153
	g = (((Pixel&0x07E0)>>5)<<2); 					\
icculus@1162
   154
	b = ((Pixel&0x001F)<<3); 					\
slouken@0
   155
}
icculus@1162
   156
#define RGB_FROM_RGB555(Pixel, r, g, b)					\
slouken@0
   157
{									\
icculus@1162
   158
	r = (((Pixel&0x7C00)>>10)<<3);		 			\
icculus@1162
   159
	g = (((Pixel&0x03E0)>>5)<<3); 					\
icculus@1162
   160
	b = ((Pixel&0x001F)<<3); 					\
slouken@0
   161
}
icculus@1162
   162
#define RGB_FROM_RGB888(Pixel, r, g, b)					\
slouken@0
   163
{									\
icculus@1162
   164
	r = ((Pixel&0xFF0000)>>16);		 			\
icculus@1162
   165
	g = ((Pixel&0xFF00)>>8);		 			\
icculus@1162
   166
	b = (Pixel&0xFF);			 			\
slouken@0
   167
}
icculus@1162
   168
#define RETRIEVE_RGB_PIXEL(buf, bpp, Pixel)				   \
slouken@0
   169
do {									   \
slouken@0
   170
	switch (bpp) {							   \
slouken@0
   171
		case 2:							   \
icculus@1162
   172
			Pixel = *((Uint16 *)(buf));			   \
slouken@0
   173
		break;							   \
slouken@0
   174
									   \
slouken@0
   175
		case 3: {						   \
slouken@0
   176
		        Uint8 *B = (Uint8 *)(buf);			   \
slouken@2824
   177
			if (SDL_BYTEORDER == SDL_LIL_ENDIAN) {		   \
icculus@1162
   178
			        Pixel = B[0] + (B[1] << 8) + (B[2] << 16); \
slouken@0
   179
			} else {					   \
icculus@1162
   180
			        Pixel = (B[0] << 16) + (B[1] << 8) + B[2]; \
slouken@0
   181
			}						   \
slouken@0
   182
		}							   \
slouken@0
   183
		break;							   \
slouken@0
   184
									   \
slouken@0
   185
		case 4:							   \
icculus@1162
   186
			Pixel = *((Uint32 *)(buf));			   \
slouken@0
   187
		break;							   \
slouken@0
   188
									   \
slouken@0
   189
		default:						   \
slouken@2824
   190
		        Pixel; /* stop gcc complaints */		   \
slouken@0
   191
		break;							   \
slouken@0
   192
	}								   \
slouken@2824
   193
} while (0)
slouken@0
   194
icculus@1162
   195
#define DISEMBLE_RGB(buf, bpp, fmt, Pixel, r, g, b)			   \
slouken@0
   196
do {									   \
slouken@0
   197
	switch (bpp) {							   \
slouken@0
   198
		case 2:							   \
icculus@1162
   199
			Pixel = *((Uint16 *)(buf));			   \
slouken@2824
   200
			RGB_FROM_PIXEL(Pixel, fmt, r, g, b);		   \
slouken@0
   201
		break;							   \
slouken@0
   202
									   \
slouken@2824
   203
		case 3:	{						   \
slouken@2824
   204
                        if (SDL_BYTEORDER == SDL_LIL_ENDIAN) {		   \
slouken@2824
   205
			        r = *((buf)+fmt->Rshift/8);		   \
slouken@2824
   206
				g = *((buf)+fmt->Gshift/8);		   \
slouken@2824
   207
				b = *((buf)+fmt->Bshift/8);		   \
slouken@0
   208
			} else {					   \
slouken@2824
   209
			        r = *((buf)+2-fmt->Rshift/8);		   \
slouken@2824
   210
				g = *((buf)+2-fmt->Gshift/8);		   \
slouken@2824
   211
				b = *((buf)+2-fmt->Bshift/8);		   \
slouken@0
   212
			}						   \
slouken@0
   213
		}							   \
slouken@0
   214
		break;							   \
slouken@0
   215
									   \
slouken@0
   216
		case 4:							   \
icculus@1162
   217
			Pixel = *((Uint32 *)(buf));			   \
slouken@2824
   218
			RGB_FROM_PIXEL(Pixel, fmt, r, g, b);		   \
slouken@0
   219
		break;							   \
slouken@0
   220
									   \
slouken@2824
   221
		default:						   \
slouken@2824
   222
		        Pixel; /* stop gcc complaints */		   \
slouken@0
   223
		break;							   \
slouken@0
   224
	}								   \
slouken@2824
   225
} while (0)
slouken@0
   226
slouken@0
   227
/* Assemble R-G-B values into a specified pixel format and store them */
icculus@1162
   228
#define PIXEL_FROM_RGB(Pixel, fmt, r, g, b)				\
slouken@0
   229
{									\
icculus@1162
   230
	Pixel = ((r>>fmt->Rloss)<<fmt->Rshift)|				\
slouken@0
   231
		((g>>fmt->Gloss)<<fmt->Gshift)|				\
slouken@0
   232
		((b>>fmt->Bloss)<<fmt->Bshift);				\
slouken@0
   233
}
icculus@1162
   234
#define RGB565_FROM_RGB(Pixel, r, g, b)					\
slouken@0
   235
{									\
icculus@1162
   236
	Pixel = ((r>>3)<<11)|((g>>2)<<5)|(b>>3);			\
slouken@0
   237
}
icculus@1162
   238
#define RGB555_FROM_RGB(Pixel, r, g, b)					\
slouken@0
   239
{									\
icculus@1162
   240
	Pixel = ((r>>3)<<10)|((g>>3)<<5)|(b>>3);			\
slouken@0
   241
}
icculus@1162
   242
#define RGB888_FROM_RGB(Pixel, r, g, b)					\
slouken@0
   243
{									\
icculus@1162
   244
	Pixel = (r<<16)|(g<<8)|b;					\
slouken@0
   245
}
slouken@2899
   246
#define ARGB8888_FROM_RGBA(Pixel, r, g, b, a)				\
slouken@2899
   247
{									\
slouken@2899
   248
	Pixel = (a<<24)|(r<<16)|(g<<8)|b;				\
slouken@2899
   249
}
slouken@3054
   250
#define RGBA8888_FROM_RGBA(Pixel, r, g, b, a)				\
slouken@3054
   251
{									\
slouken@3054
   252
	Pixel = (r<<24)|(g<<16)|(b<<8)|a;				\
slouken@3054
   253
}
slouken@3054
   254
#define ABGR8888_FROM_RGBA(Pixel, r, g, b, a)				\
slouken@3054
   255
{									\
slouken@3054
   256
	Pixel = (a<<24)|(b<<16)|(g<<8)|r;				\
slouken@3054
   257
}
slouken@3054
   258
#define BGRA8888_FROM_RGBA(Pixel, r, g, b, a)				\
slouken@3054
   259
{									\
slouken@3054
   260
	Pixel = (b<<24)|(g<<16)|(r<<8)|a;				\
slouken@3054
   261
}
slouken@0
   262
#define ASSEMBLE_RGB(buf, bpp, fmt, r, g, b) 				\
slouken@0
   263
{									\
slouken@0
   264
	switch (bpp) {							\
slouken@0
   265
		case 2: {						\
icculus@1162
   266
			Uint16 Pixel;					\
slouken@0
   267
									\
icculus@1162
   268
			PIXEL_FROM_RGB(Pixel, fmt, r, g, b);		\
icculus@1162
   269
			*((Uint16 *)(buf)) = Pixel;			\
slouken@0
   270
		}							\
slouken@0
   271
		break;							\
slouken@0
   272
									\
slouken@0
   273
		case 3: {						\
slouken@2824
   274
                        if (SDL_BYTEORDER == SDL_LIL_ENDIAN) {		\
slouken@0
   275
			        *((buf)+fmt->Rshift/8) = r;		\
slouken@0
   276
				*((buf)+fmt->Gshift/8) = g;		\
slouken@0
   277
				*((buf)+fmt->Bshift/8) = b;		\
slouken@0
   278
			} else {					\
slouken@0
   279
			        *((buf)+2-fmt->Rshift/8) = r;		\
slouken@0
   280
				*((buf)+2-fmt->Gshift/8) = g;		\
slouken@0
   281
				*((buf)+2-fmt->Bshift/8) = b;		\
slouken@0
   282
			}						\
slouken@0
   283
		}							\
slouken@0
   284
		break;							\
slouken@0
   285
									\
slouken@0
   286
		case 4: {						\
icculus@1162
   287
			Uint32 Pixel;					\
slouken@0
   288
									\
icculus@1162
   289
			PIXEL_FROM_RGB(Pixel, fmt, r, g, b);		\
icculus@1162
   290
			*((Uint32 *)(buf)) = Pixel;			\
slouken@0
   291
		}							\
slouken@0
   292
		break;							\
slouken@0
   293
	}								\
slouken@0
   294
}
slouken@0
   295
#define ASSEMBLE_RGB_AMASK(buf, bpp, fmt, r, g, b, Amask)		\
slouken@0
   296
{									\
slouken@0
   297
	switch (bpp) {							\
slouken@0
   298
		case 2: {						\
slouken@0
   299
			Uint16 *bufp;					\
icculus@1162
   300
			Uint16 Pixel;					\
slouken@0
   301
									\
slouken@0
   302
			bufp = (Uint16 *)buf;				\
icculus@1162
   303
			PIXEL_FROM_RGB(Pixel, fmt, r, g, b);		\
icculus@1162
   304
			*bufp = Pixel | (*bufp & Amask);		\
slouken@0
   305
		}							\
slouken@0
   306
		break;							\
slouken@0
   307
									\
slouken@0
   308
		case 3: {						\
slouken@2824
   309
                        if (SDL_BYTEORDER == SDL_LIL_ENDIAN) {		\
slouken@0
   310
			        *((buf)+fmt->Rshift/8) = r;		\
slouken@0
   311
				*((buf)+fmt->Gshift/8) = g;		\
slouken@0
   312
				*((buf)+fmt->Bshift/8) = b;		\
slouken@0
   313
			} else {					\
slouken@0
   314
			        *((buf)+2-fmt->Rshift/8) = r;		\
slouken@0
   315
				*((buf)+2-fmt->Gshift/8) = g;		\
slouken@0
   316
				*((buf)+2-fmt->Bshift/8) = b;		\
slouken@0
   317
			}						\
slouken@0
   318
		}							\
slouken@0
   319
		break;							\
slouken@0
   320
									\
slouken@0
   321
		case 4: {						\
slouken@0
   322
			Uint32 *bufp;					\
icculus@1162
   323
			Uint32 Pixel;					\
slouken@0
   324
									\
slouken@0
   325
			bufp = (Uint32 *)buf;				\
icculus@1162
   326
			PIXEL_FROM_RGB(Pixel, fmt, r, g, b);		\
icculus@1162
   327
			*bufp = Pixel | (*bufp & Amask);		\
slouken@0
   328
		}							\
slouken@0
   329
		break;							\
slouken@0
   330
	}								\
slouken@0
   331
}
slouken@0
   332
slouken@0
   333
/* FIXME: Should we rescale alpha into 0..255 here? */
icculus@1162
   334
#define RGBA_FROM_PIXEL(Pixel, fmt, r, g, b, a)				\
slouken@0
   335
{									\
icculus@1162
   336
	r = ((Pixel&fmt->Rmask)>>fmt->Rshift)<<fmt->Rloss; 		\
icculus@1162
   337
	g = ((Pixel&fmt->Gmask)>>fmt->Gshift)<<fmt->Gloss; 		\
icculus@1162
   338
	b = ((Pixel&fmt->Bmask)>>fmt->Bshift)<<fmt->Bloss; 		\
icculus@1162
   339
	a = ((Pixel&fmt->Amask)>>fmt->Ashift)<<fmt->Aloss;	 	\
slouken@0
   340
}
icculus@1162
   341
#define RGBA_FROM_8888(Pixel, fmt, r, g, b, a)	\
slouken@0
   342
{						\
icculus@1162
   343
	r = (Pixel&fmt->Rmask)>>fmt->Rshift;	\
icculus@1162
   344
	g = (Pixel&fmt->Gmask)>>fmt->Gshift;	\
icculus@1162
   345
	b = (Pixel&fmt->Bmask)>>fmt->Bshift;	\
icculus@1162
   346
	a = (Pixel&fmt->Amask)>>fmt->Ashift;	\
slouken@0
   347
}
icculus@1162
   348
#define RGBA_FROM_RGBA8888(Pixel, r, g, b, a)				\
slouken@0
   349
{									\
icculus@1162
   350
	r = (Pixel>>24);						\
icculus@1162
   351
	g = ((Pixel>>16)&0xFF);						\
icculus@1162
   352
	b = ((Pixel>>8)&0xFF);						\
icculus@1162
   353
	a = (Pixel&0xFF);						\
slouken@0
   354
}
icculus@1162
   355
#define RGBA_FROM_ARGB8888(Pixel, r, g, b, a)				\
slouken@0
   356
{									\
icculus@1162
   357
	r = ((Pixel>>16)&0xFF);						\
icculus@1162
   358
	g = ((Pixel>>8)&0xFF);						\
icculus@1162
   359
	b = (Pixel&0xFF);						\
icculus@1162
   360
	a = (Pixel>>24);						\
slouken@0
   361
}
icculus@1162
   362
#define RGBA_FROM_ABGR8888(Pixel, r, g, b, a)				\
slouken@0
   363
{									\
icculus@1162
   364
	r = (Pixel&0xFF);						\
icculus@1162
   365
	g = ((Pixel>>8)&0xFF);						\
icculus@1162
   366
	b = ((Pixel>>16)&0xFF);						\
icculus@1162
   367
	a = (Pixel>>24);						\
slouken@0
   368
}
slouken@3054
   369
#define RGBA_FROM_BGRA8888(Pixel, r, g, b, a)				\
slouken@3054
   370
{									\
slouken@3054
   371
	r = ((Pixel>>8)&0xFF);						\
slouken@3054
   372
	g = ((Pixel>>16)&0xFF);						\
slouken@3054
   373
	b = (Pixel>>24);						\
slouken@3054
   374
	a = (Pixel&0xFF);						\
slouken@3054
   375
}
icculus@1162
   376
#define DISEMBLE_RGBA(buf, bpp, fmt, Pixel, r, g, b, a)			   \
slouken@0
   377
do {									   \
slouken@0
   378
	switch (bpp) {							   \
slouken@0
   379
		case 2:							   \
icculus@1162
   380
			Pixel = *((Uint16 *)(buf));			   \
slouken@2824
   381
			RGBA_FROM_PIXEL(Pixel, fmt, r, g, b, a);	   \
slouken@0
   382
		break;							   \
slouken@0
   383
									   \
slouken@2824
   384
		case 3:	{						   \
slouken@2824
   385
                        if (SDL_BYTEORDER == SDL_LIL_ENDIAN) {		   \
slouken@2824
   386
			        r = *((buf)+fmt->Rshift/8);		   \
slouken@2824
   387
				g = *((buf)+fmt->Gshift/8);		   \
slouken@2824
   388
				b = *((buf)+fmt->Bshift/8);		   \
slouken@0
   389
			} else {					   \
slouken@2824
   390
			        r = *((buf)+2-fmt->Rshift/8);		   \
slouken@2824
   391
				g = *((buf)+2-fmt->Gshift/8);		   \
slouken@2824
   392
				b = *((buf)+2-fmt->Bshift/8);		   \
slouken@0
   393
			}						   \
slouken@2824
   394
			a = 0xFF;					   \
slouken@0
   395
		}							   \
slouken@0
   396
		break;							   \
slouken@0
   397
									   \
slouken@0
   398
		case 4:							   \
icculus@1162
   399
			Pixel = *((Uint32 *)(buf));			   \
slouken@2824
   400
			RGBA_FROM_PIXEL(Pixel, fmt, r, g, b, a);	   \
slouken@0
   401
		break;							   \
slouken@0
   402
									   \
slouken@0
   403
		default:						   \
slouken@2824
   404
		        Pixel; /* stop gcc complaints */		   \
slouken@0
   405
		break;							   \
slouken@0
   406
	}								   \
slouken@2824
   407
} while (0)
slouken@0
   408
slouken@0
   409
/* FIXME: this isn't correct, especially for Alpha (maximum != 255) */
icculus@1162
   410
#define PIXEL_FROM_RGBA(Pixel, fmt, r, g, b, a)				\
slouken@0
   411
{									\
icculus@1162
   412
	Pixel = ((r>>fmt->Rloss)<<fmt->Rshift)|				\
slouken@0
   413
		((g>>fmt->Gloss)<<fmt->Gshift)|				\
slouken@0
   414
		((b>>fmt->Bloss)<<fmt->Bshift)|				\
slouken@535
   415
		((a>>fmt->Aloss)<<fmt->Ashift);				\
slouken@0
   416
}
slouken@0
   417
#define ASSEMBLE_RGBA(buf, bpp, fmt, r, g, b, a)			\
slouken@0
   418
{									\
slouken@0
   419
	switch (bpp) {							\
slouken@0
   420
		case 2: {						\
icculus@1162
   421
			Uint16 Pixel;					\
slouken@0
   422
									\
icculus@1162
   423
			PIXEL_FROM_RGBA(Pixel, fmt, r, g, b, a);	\
icculus@1162
   424
			*((Uint16 *)(buf)) = Pixel;			\
slouken@0
   425
		}							\
slouken@0
   426
		break;							\
slouken@0
   427
									\
slouken@2824
   428
		case 3: {						\
slouken@2824
   429
                        if (SDL_BYTEORDER == SDL_LIL_ENDIAN) {		\
slouken@0
   430
			        *((buf)+fmt->Rshift/8) = r;		\
slouken@0
   431
				*((buf)+fmt->Gshift/8) = g;		\
slouken@0
   432
				*((buf)+fmt->Bshift/8) = b;		\
slouken@0
   433
			} else {					\
slouken@0
   434
			        *((buf)+2-fmt->Rshift/8) = r;		\
slouken@0
   435
				*((buf)+2-fmt->Gshift/8) = g;		\
slouken@0
   436
				*((buf)+2-fmt->Bshift/8) = b;		\
slouken@0
   437
			}						\
slouken@0
   438
		}							\
slouken@0
   439
		break;							\
slouken@0
   440
									\
slouken@0
   441
		case 4: {						\
icculus@1162
   442
			Uint32 Pixel;					\
slouken@0
   443
									\
icculus@1162
   444
			PIXEL_FROM_RGBA(Pixel, fmt, r, g, b, a);	\
icculus@1162
   445
			*((Uint32 *)(buf)) = Pixel;			\
slouken@0
   446
		}							\
slouken@0
   447
		break;							\
slouken@0
   448
	}								\
slouken@0
   449
}
slouken@0
   450
icculus@1162
   451
/* Blend the RGB values of two Pixels based on a source alpha value */
slouken@0
   452
#define ALPHA_BLEND(sR, sG, sB, A, dR, dG, dB)	\
slouken@0
   453
do {						\
icculus@3631
   454
	dR = ((((int)(sR-dR)*(int)A)/255)+dR);	\
icculus@3631
   455
	dG = ((((int)(sG-dG)*(int)A)/255)+dG);	\
icculus@3631
   456
	dB = ((((int)(sB-dB)*(int)A)/255)+dB);	\
icculus@1047
   457
} while(0)
icculus@1047
   458
icculus@1047
   459
slouken@0
   460
/* This is a very useful loop for optimizing blitters */
slouken@553
   461
#if defined(_MSC_VER) && (_MSC_VER == 1300)
slouken@553
   462
/* There's a bug in the Visual C++ 7 optimizer when compiling this code */
slouken@553
   463
#else
slouken@0
   464
#define USE_DUFFS_LOOP
slouken@553
   465
#endif
slouken@0
   466
#ifdef USE_DUFFS_LOOP
slouken@0
   467
slouken@0
   468
/* 8-times unrolled loop */
slouken@0
   469
#define DUFFS_LOOP8(pixel_copy_increment, width)			\
slouken@0
   470
{ int n = (width+7)/8;							\
slouken@91
   471
	switch (width & 7) {						\
slouken@0
   472
	case 0: do {	pixel_copy_increment;				\
slouken@0
   473
	case 7:		pixel_copy_increment;				\
slouken@0
   474
	case 6:		pixel_copy_increment;				\
slouken@0
   475
	case 5:		pixel_copy_increment;				\
slouken@0
   476
	case 4:		pixel_copy_increment;				\
slouken@0
   477
	case 3:		pixel_copy_increment;				\
slouken@0
   478
	case 2:		pixel_copy_increment;				\
slouken@0
   479
	case 1:		pixel_copy_increment;				\
slouken@0
   480
		} while ( --n > 0 );					\
slouken@0
   481
	}								\
slouken@0
   482
}
slouken@0
   483
slouken@0
   484
/* 4-times unrolled loop */
slouken@0
   485
#define DUFFS_LOOP4(pixel_copy_increment, width)			\
slouken@0
   486
{ int n = (width+3)/4;							\
slouken@91
   487
	switch (width & 3) {						\
slouken@0
   488
	case 0: do {	pixel_copy_increment;				\
slouken@0
   489
	case 3:		pixel_copy_increment;				\
slouken@0
   490
	case 2:		pixel_copy_increment;				\
slouken@0
   491
	case 1:		pixel_copy_increment;				\
slouken@3035
   492
		} while (--n > 0);					\
slouken@689
   493
	}								\
slouken@689
   494
}
slouken@689
   495
slouken@0
   496
/* Use the 8-times version of the loop by default */
slouken@0
   497
#define DUFFS_LOOP(pixel_copy_increment, width)				\
slouken@0
   498
	DUFFS_LOOP8(pixel_copy_increment, width)
slouken@0
   499
slouken@3035
   500
/* Special version of Duff's device for even more optimization */
slouken@3035
   501
#define DUFFS_LOOP_124(pixel_copy_increment1,				\
slouken@3035
   502
                       pixel_copy_increment2,				\
slouken@3035
   503
                       pixel_copy_increment4, width)			\
slouken@3035
   504
{ int n = width;							\
slouken@3035
   505
	if (n & 1) {							\
slouken@3035
   506
		pixel_copy_increment1; n -= 1;				\
slouken@3035
   507
	}								\
slouken@3035
   508
	if (n & 2) {							\
slouken@3035
   509
		pixel_copy_increment2; n -= 2;				\
slouken@3035
   510
	}								\
slouken@3035
   511
	if (n) {							\
slouken@3035
   512
		n = (n+7)/ 8;						\
slouken@3035
   513
		switch (n & 4) {					\
slouken@3035
   514
		case 0: do {	pixel_copy_increment4;			\
slouken@3035
   515
		case 4:		pixel_copy_increment4;			\
slouken@3035
   516
			} while (--n > 0);				\
slouken@3035
   517
		}							\
slouken@3035
   518
	}								\
slouken@689
   519
}
slouken@689
   520
slouken@3035
   521
#else
slouken@689
   522
slouken@689
   523
/* Don't use Duff's device to unroll loops */
slouken@0
   524
#define DUFFS_LOOP(pixel_copy_increment, width)				\
slouken@0
   525
{ int n;								\
slouken@0
   526
	for ( n=width; n > 0; --n ) {					\
slouken@0
   527
		pixel_copy_increment;					\
slouken@0
   528
	}								\
slouken@0
   529
}
slouken@0
   530
#define DUFFS_LOOP8(pixel_copy_increment, width)			\
slouken@0
   531
	DUFFS_LOOP(pixel_copy_increment, width)
slouken@0
   532
#define DUFFS_LOOP4(pixel_copy_increment, width)			\
slouken@0
   533
	DUFFS_LOOP(pixel_copy_increment, width)
slouken@3035
   534
#define DUFFS_LOOP_124(pixel_copy_increment1,				\
slouken@3035
   535
                       pixel_copy_increment2,				\
slouken@3035
   536
                       pixel_copy_increment4, width)			\
slouken@3035
   537
	DUFFS_LOOP(pixel_copy_increment1, width)
slouken@0
   538
slouken@0
   539
#endif /* USE_DUFFS_LOOP */
slouken@0
   540
slouken@0
   541
/* Prevent Visual C++ 6.0 from printing out stupid warnings */
slouken@0
   542
#if defined(_MSC_VER) && (_MSC_VER >= 600)
slouken@0
   543
#pragma warning(disable: 4550)
slouken@0
   544
#endif
slouken@0
   545
slouken@0
   546
#endif /* _SDL_blit_h */
slouken@2898
   547
slouken@1895
   548
/* vi: set ts=4 sw=4 expandtab: */