src/video/SDL_blit.h
author Sam Lantinga <slouken@libsdl.org>
Tue, 22 Feb 2011 21:44:36 -0800
changeset 5389 24903690f48a
parent 5288 d4381f3b0d1e
child 5423 b69fa50e80d7
permissions -rw-r--r--
Re-added the 3DNow! and AltiVec instruction support.
slouken@0
     1
/*
slouken@0
     2
    SDL - Simple DirectMedia Layer
slouken@5262
     3
    Copyright (C) 1997-2011 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@2250
    27
#include "SDL_cpuinfo.h"
slouken@0
    28
#include "SDL_endian.h"
slouken@5163
    29
#include "SDL_surface.h"
slouken@0
    30
slouken@2262
    31
/* SDL blit copy flags */
slouken@2266
    32
#define SDL_COPY_MODULATE_COLOR     0x00000001
slouken@2266
    33
#define SDL_COPY_MODULATE_ALPHA     0x00000002
slouken@5140
    34
#define SDL_COPY_BLEND              0x00000010
slouken@5140
    35
#define SDL_COPY_ADD                0x00000020
slouken@5184
    36
#define SDL_COPY_MOD                0x00000040
slouken@2266
    37
#define SDL_COPY_COLORKEY           0x00000100
slouken@2266
    38
#define SDL_COPY_NEAREST            0x00000200
slouken@2266
    39
#define SDL_COPY_RLE_DESIRED        0x00001000
slouken@2266
    40
#define SDL_COPY_RLE_COLORKEY       0x00002000
slouken@2266
    41
#define SDL_COPY_RLE_ALPHAKEY       0x00004000
slouken@2853
    42
#define SDL_COPY_RLE_MASK           (SDL_COPY_RLE_DESIRED|SDL_COPY_RLE_COLORKEY|SDL_COPY_RLE_ALPHAKEY)
slouken@2262
    43
slouken@2262
    44
/* SDL blit CPU flags */
slouken@2266
    45
#define SDL_CPU_ANY                 0x00000000
slouken@2266
    46
#define SDL_CPU_MMX                 0x00000001
slouken@5389
    47
#define SDL_CPU_3DNOW               0x00000002
slouken@2266
    48
#define SDL_CPU_SSE                 0x00000004
slouken@2266
    49
#define SDL_CPU_SSE2                0x00000008
slouken@5389
    50
#define SDL_CPU_ALTIVEC_PREFETCH    0x00000010
slouken@5389
    51
#define SDL_CPU_ALTIVEC_NOPREFETCH  0x00000020
slouken@2262
    52
slouken@2267
    53
typedef struct
slouken@2267
    54
{
slouken@2262
    55
    Uint8 *src;
slouken@2262
    56
    int src_w, src_h;
slouken@2262
    57
    int src_pitch;
slouken@2267
    58
    int src_skip;
slouken@2262
    59
    Uint8 *dst;
slouken@2262
    60
    int dst_w, dst_h;
slouken@2262
    61
    int dst_pitch;
slouken@2267
    62
    int dst_skip;
slouken@2262
    63
    SDL_PixelFormat *src_fmt;
slouken@2262
    64
    SDL_PixelFormat *dst_fmt;
slouken@1895
    65
    Uint8 *table;
slouken@2262
    66
    int flags;
slouken@2262
    67
    Uint32 colorkey;
slouken@2262
    68
    Uint8 r, g, b, a;
slouken@0
    69
} SDL_BlitInfo;
slouken@0
    70
slouken@2267
    71
typedef void (SDLCALL * SDL_BlitFunc) (SDL_BlitInfo * info);
slouken@2262
    72
slouken@2267
    73
typedef struct
slouken@2267
    74
{
slouken@2262
    75
    Uint32 src_format;
slouken@2262
    76
    Uint32 dst_format;
slouken@2262
    77
    int flags;
slouken@2262
    78
    int cpu;
slouken@2262
    79
    SDL_BlitFunc func;
slouken@2262
    80
} SDL_BlitFuncEntry;
slouken@0
    81
slouken@0
    82
/* Blit mapping definition */
slouken@1895
    83
typedef struct SDL_BlitMap
slouken@1895
    84
{
slouken@1895
    85
    SDL_Surface *dst;
slouken@1895
    86
    int identity;
slouken@2257
    87
    SDL_blit blit;
slouken@2257
    88
    void *data;
slouken@2262
    89
    SDL_BlitInfo info;
slouken@0
    90
slouken@1895
    91
    /* the version count matches the destination; mismatch indicates
slouken@1895
    92
       an invalid mapping */
slouken@5288
    93
    Uint32 palette_version;
slouken@0
    94
} SDL_BlitMap;
slouken@0
    95
slouken@0
    96
/* Functions found in SDL_blit.c */
slouken@1895
    97
extern int SDL_CalculateBlit(SDL_Surface * surface);
slouken@0
    98
slouken@2267
    99
/* Functions found in SDL_blit_*.c */
slouken@2267
   100
extern SDL_BlitFunc SDL_CalculateBlit0(SDL_Surface * surface);
slouken@2267
   101
extern SDL_BlitFunc SDL_CalculateBlit1(SDL_Surface * surface);
slouken@2267
   102
extern SDL_BlitFunc SDL_CalculateBlitN(SDL_Surface * surface);
slouken@2267
   103
extern SDL_BlitFunc SDL_CalculateBlitA(SDL_Surface * surface);
slouken@2263
   104
slouken@0
   105
/*
slouken@0
   106
 * Useful macros for blitting routines
slouken@0
   107
 */
slouken@0
   108
slouken@2249
   109
#if defined(__GNUC__)
slouken@2249
   110
#define DECLARE_ALIGNED(t,v,a)  t __attribute__((aligned(a))) v
slouken@2249
   111
#elif defined(_MSC_VER)
slouken@2251
   112
#define DECLARE_ALIGNED(t,v,a)  __declspec(align(a)) t v
slouken@2249
   113
#else
slouken@2249
   114
#define DECLARE_ALIGNED(t,v,a)  t v
slouken@2249
   115
#endif
slouken@2249
   116
slouken@0
   117
/* Load pixel of the specified format from a buffer and get its R-G-B values */
slouken@0
   118
/* FIXME: rescale values to 0..255 here? */
icculus@1162
   119
#define RGB_FROM_PIXEL(Pixel, fmt, r, g, b)				\
slouken@0
   120
{									\
icculus@1162
   121
	r = (((Pixel&fmt->Rmask)>>fmt->Rshift)<<fmt->Rloss); 		\
icculus@1162
   122
	g = (((Pixel&fmt->Gmask)>>fmt->Gshift)<<fmt->Gloss); 		\
icculus@1162
   123
	b = (((Pixel&fmt->Bmask)>>fmt->Bshift)<<fmt->Bloss); 		\
slouken@0
   124
}
icculus@1162
   125
#define RGB_FROM_RGB565(Pixel, r, g, b)					\
slouken@0
   126
{									\
icculus@1162
   127
	r = (((Pixel&0xF800)>>11)<<3);		 			\
icculus@1162
   128
	g = (((Pixel&0x07E0)>>5)<<2); 					\
icculus@1162
   129
	b = ((Pixel&0x001F)<<3); 					\
slouken@0
   130
}
icculus@1162
   131
#define RGB_FROM_RGB555(Pixel, r, g, b)					\
slouken@0
   132
{									\
icculus@1162
   133
	r = (((Pixel&0x7C00)>>10)<<3);		 			\
icculus@1162
   134
	g = (((Pixel&0x03E0)>>5)<<3); 					\
icculus@1162
   135
	b = ((Pixel&0x001F)<<3); 					\
slouken@0
   136
}
icculus@1162
   137
#define RGB_FROM_RGB888(Pixel, r, g, b)					\
slouken@0
   138
{									\
icculus@1162
   139
	r = ((Pixel&0xFF0000)>>16);		 			\
icculus@1162
   140
	g = ((Pixel&0xFF00)>>8);		 			\
icculus@1162
   141
	b = (Pixel&0xFF);			 			\
slouken@0
   142
}
icculus@1162
   143
#define RETRIEVE_RGB_PIXEL(buf, bpp, Pixel)				   \
slouken@0
   144
do {									   \
slouken@0
   145
	switch (bpp) {							   \
slouken@0
   146
		case 2:							   \
icculus@1162
   147
			Pixel = *((Uint16 *)(buf));			   \
slouken@0
   148
		break;							   \
slouken@0
   149
									   \
slouken@0
   150
		case 3: {						   \
slouken@0
   151
		        Uint8 *B = (Uint8 *)(buf);			   \
slouken@2824
   152
			if (SDL_BYTEORDER == SDL_LIL_ENDIAN) {		   \
icculus@1162
   153
			        Pixel = B[0] + (B[1] << 8) + (B[2] << 16); \
slouken@0
   154
			} else {					   \
icculus@1162
   155
			        Pixel = (B[0] << 16) + (B[1] << 8) + B[2]; \
slouken@0
   156
			}						   \
slouken@0
   157
		}							   \
slouken@0
   158
		break;							   \
slouken@0
   159
									   \
slouken@0
   160
		case 4:							   \
icculus@1162
   161
			Pixel = *((Uint32 *)(buf));			   \
slouken@0
   162
		break;							   \
slouken@0
   163
									   \
slouken@0
   164
		default:						   \
slouken@2824
   165
		        Pixel; /* stop gcc complaints */		   \
slouken@0
   166
		break;							   \
slouken@0
   167
	}								   \
slouken@2824
   168
} while (0)
slouken@0
   169
icculus@1162
   170
#define DISEMBLE_RGB(buf, bpp, fmt, Pixel, r, g, b)			   \
slouken@0
   171
do {									   \
slouken@0
   172
	switch (bpp) {							   \
slouken@0
   173
		case 2:							   \
icculus@1162
   174
			Pixel = *((Uint16 *)(buf));			   \
slouken@2824
   175
			RGB_FROM_PIXEL(Pixel, fmt, r, g, b);		   \
slouken@0
   176
		break;							   \
slouken@0
   177
									   \
slouken@2824
   178
		case 3:	{						   \
slouken@2824
   179
                        if (SDL_BYTEORDER == SDL_LIL_ENDIAN) {		   \
slouken@2824
   180
			        r = *((buf)+fmt->Rshift/8);		   \
slouken@2824
   181
				g = *((buf)+fmt->Gshift/8);		   \
slouken@2824
   182
				b = *((buf)+fmt->Bshift/8);		   \
slouken@0
   183
			} else {					   \
slouken@2824
   184
			        r = *((buf)+2-fmt->Rshift/8);		   \
slouken@2824
   185
				g = *((buf)+2-fmt->Gshift/8);		   \
slouken@2824
   186
				b = *((buf)+2-fmt->Bshift/8);		   \
slouken@0
   187
			}						   \
slouken@0
   188
		}							   \
slouken@0
   189
		break;							   \
slouken@0
   190
									   \
slouken@0
   191
		case 4:							   \
icculus@1162
   192
			Pixel = *((Uint32 *)(buf));			   \
slouken@2824
   193
			RGB_FROM_PIXEL(Pixel, fmt, r, g, b);		   \
slouken@0
   194
		break;							   \
slouken@0
   195
									   \
slouken@2824
   196
		default:						   \
slouken@2824
   197
		        Pixel; /* stop gcc complaints */		   \
slouken@0
   198
		break;							   \
slouken@0
   199
	}								   \
slouken@2824
   200
} while (0)
slouken@0
   201
slouken@0
   202
/* Assemble R-G-B values into a specified pixel format and store them */
icculus@1162
   203
#define PIXEL_FROM_RGB(Pixel, fmt, r, g, b)				\
slouken@0
   204
{									\
icculus@1162
   205
	Pixel = ((r>>fmt->Rloss)<<fmt->Rshift)|				\
slouken@0
   206
		((g>>fmt->Gloss)<<fmt->Gshift)|				\
slouken@0
   207
		((b>>fmt->Bloss)<<fmt->Bshift);				\
slouken@0
   208
}
icculus@1162
   209
#define RGB565_FROM_RGB(Pixel, r, g, b)					\
slouken@0
   210
{									\
icculus@1162
   211
	Pixel = ((r>>3)<<11)|((g>>2)<<5)|(b>>3);			\
slouken@0
   212
}
icculus@1162
   213
#define RGB555_FROM_RGB(Pixel, r, g, b)					\
slouken@0
   214
{									\
icculus@1162
   215
	Pixel = ((r>>3)<<10)|((g>>3)<<5)|(b>>3);			\
slouken@0
   216
}
icculus@1162
   217
#define RGB888_FROM_RGB(Pixel, r, g, b)					\
slouken@0
   218
{									\
icculus@1162
   219
	Pixel = (r<<16)|(g<<8)|b;					\
slouken@0
   220
}
slouken@2899
   221
#define ARGB8888_FROM_RGBA(Pixel, r, g, b, a)				\
slouken@2899
   222
{									\
slouken@2899
   223
	Pixel = (a<<24)|(r<<16)|(g<<8)|b;				\
slouken@2899
   224
}
slouken@3054
   225
#define RGBA8888_FROM_RGBA(Pixel, r, g, b, a)				\
slouken@3054
   226
{									\
slouken@3054
   227
	Pixel = (r<<24)|(g<<16)|(b<<8)|a;				\
slouken@3054
   228
}
slouken@3054
   229
#define ABGR8888_FROM_RGBA(Pixel, r, g, b, a)				\
slouken@3054
   230
{									\
slouken@3054
   231
	Pixel = (a<<24)|(b<<16)|(g<<8)|r;				\
slouken@3054
   232
}
slouken@3054
   233
#define BGRA8888_FROM_RGBA(Pixel, r, g, b, a)				\
slouken@3054
   234
{									\
slouken@3054
   235
	Pixel = (b<<24)|(g<<16)|(r<<8)|a;				\
slouken@3054
   236
}
slouken@0
   237
#define ASSEMBLE_RGB(buf, bpp, fmt, r, g, b) 				\
slouken@0
   238
{									\
slouken@0
   239
	switch (bpp) {							\
slouken@0
   240
		case 2: {						\
icculus@1162
   241
			Uint16 Pixel;					\
slouken@0
   242
									\
icculus@1162
   243
			PIXEL_FROM_RGB(Pixel, fmt, r, g, b);		\
icculus@1162
   244
			*((Uint16 *)(buf)) = Pixel;			\
slouken@0
   245
		}							\
slouken@0
   246
		break;							\
slouken@0
   247
									\
slouken@0
   248
		case 3: {						\
slouken@2824
   249
                        if (SDL_BYTEORDER == SDL_LIL_ENDIAN) {		\
slouken@0
   250
			        *((buf)+fmt->Rshift/8) = r;		\
slouken@0
   251
				*((buf)+fmt->Gshift/8) = g;		\
slouken@0
   252
				*((buf)+fmt->Bshift/8) = b;		\
slouken@0
   253
			} else {					\
slouken@0
   254
			        *((buf)+2-fmt->Rshift/8) = r;		\
slouken@0
   255
				*((buf)+2-fmt->Gshift/8) = g;		\
slouken@0
   256
				*((buf)+2-fmt->Bshift/8) = b;		\
slouken@0
   257
			}						\
slouken@0
   258
		}							\
slouken@0
   259
		break;							\
slouken@0
   260
									\
slouken@0
   261
		case 4: {						\
icculus@1162
   262
			Uint32 Pixel;					\
slouken@0
   263
									\
icculus@1162
   264
			PIXEL_FROM_RGB(Pixel, fmt, r, g, b);		\
icculus@1162
   265
			*((Uint32 *)(buf)) = Pixel;			\
slouken@0
   266
		}							\
slouken@0
   267
		break;							\
slouken@0
   268
	}								\
slouken@0
   269
}
slouken@0
   270
#define ASSEMBLE_RGB_AMASK(buf, bpp, fmt, r, g, b, Amask)		\
slouken@0
   271
{									\
slouken@0
   272
	switch (bpp) {							\
slouken@0
   273
		case 2: {						\
slouken@0
   274
			Uint16 *bufp;					\
icculus@1162
   275
			Uint16 Pixel;					\
slouken@0
   276
									\
slouken@0
   277
			bufp = (Uint16 *)buf;				\
icculus@1162
   278
			PIXEL_FROM_RGB(Pixel, fmt, r, g, b);		\
icculus@1162
   279
			*bufp = Pixel | (*bufp & Amask);		\
slouken@0
   280
		}							\
slouken@0
   281
		break;							\
slouken@0
   282
									\
slouken@0
   283
		case 3: {						\
slouken@2824
   284
                        if (SDL_BYTEORDER == SDL_LIL_ENDIAN) {		\
slouken@0
   285
			        *((buf)+fmt->Rshift/8) = r;		\
slouken@0
   286
				*((buf)+fmt->Gshift/8) = g;		\
slouken@0
   287
				*((buf)+fmt->Bshift/8) = b;		\
slouken@0
   288
			} else {					\
slouken@0
   289
			        *((buf)+2-fmt->Rshift/8) = r;		\
slouken@0
   290
				*((buf)+2-fmt->Gshift/8) = g;		\
slouken@0
   291
				*((buf)+2-fmt->Bshift/8) = b;		\
slouken@0
   292
			}						\
slouken@0
   293
		}							\
slouken@0
   294
		break;							\
slouken@0
   295
									\
slouken@0
   296
		case 4: {						\
slouken@0
   297
			Uint32 *bufp;					\
icculus@1162
   298
			Uint32 Pixel;					\
slouken@0
   299
									\
slouken@0
   300
			bufp = (Uint32 *)buf;				\
icculus@1162
   301
			PIXEL_FROM_RGB(Pixel, fmt, r, g, b);		\
icculus@1162
   302
			*bufp = Pixel | (*bufp & Amask);		\
slouken@0
   303
		}							\
slouken@0
   304
		break;							\
slouken@0
   305
	}								\
slouken@0
   306
}
slouken@0
   307
slouken@0
   308
/* FIXME: Should we rescale alpha into 0..255 here? */
icculus@1162
   309
#define RGBA_FROM_PIXEL(Pixel, fmt, r, g, b, a)				\
slouken@0
   310
{									\
icculus@1162
   311
	r = ((Pixel&fmt->Rmask)>>fmt->Rshift)<<fmt->Rloss; 		\
icculus@1162
   312
	g = ((Pixel&fmt->Gmask)>>fmt->Gshift)<<fmt->Gloss; 		\
icculus@1162
   313
	b = ((Pixel&fmt->Bmask)>>fmt->Bshift)<<fmt->Bloss; 		\
icculus@1162
   314
	a = ((Pixel&fmt->Amask)>>fmt->Ashift)<<fmt->Aloss;	 	\
slouken@0
   315
}
icculus@1162
   316
#define RGBA_FROM_8888(Pixel, fmt, r, g, b, a)	\
slouken@0
   317
{						\
icculus@1162
   318
	r = (Pixel&fmt->Rmask)>>fmt->Rshift;	\
icculus@1162
   319
	g = (Pixel&fmt->Gmask)>>fmt->Gshift;	\
icculus@1162
   320
	b = (Pixel&fmt->Bmask)>>fmt->Bshift;	\
icculus@1162
   321
	a = (Pixel&fmt->Amask)>>fmt->Ashift;	\
slouken@0
   322
}
icculus@1162
   323
#define RGBA_FROM_RGBA8888(Pixel, r, g, b, a)				\
slouken@0
   324
{									\
icculus@1162
   325
	r = (Pixel>>24);						\
icculus@1162
   326
	g = ((Pixel>>16)&0xFF);						\
icculus@1162
   327
	b = ((Pixel>>8)&0xFF);						\
icculus@1162
   328
	a = (Pixel&0xFF);						\
slouken@0
   329
}
icculus@1162
   330
#define RGBA_FROM_ARGB8888(Pixel, r, g, b, a)				\
slouken@0
   331
{									\
icculus@1162
   332
	r = ((Pixel>>16)&0xFF);						\
icculus@1162
   333
	g = ((Pixel>>8)&0xFF);						\
icculus@1162
   334
	b = (Pixel&0xFF);						\
icculus@1162
   335
	a = (Pixel>>24);						\
slouken@0
   336
}
icculus@1162
   337
#define RGBA_FROM_ABGR8888(Pixel, r, g, b, a)				\
slouken@0
   338
{									\
icculus@1162
   339
	r = (Pixel&0xFF);						\
icculus@1162
   340
	g = ((Pixel>>8)&0xFF);						\
icculus@1162
   341
	b = ((Pixel>>16)&0xFF);						\
icculus@1162
   342
	a = (Pixel>>24);						\
slouken@0
   343
}
slouken@3054
   344
#define RGBA_FROM_BGRA8888(Pixel, r, g, b, a)				\
slouken@3054
   345
{									\
slouken@3054
   346
	r = ((Pixel>>8)&0xFF);						\
slouken@3054
   347
	g = ((Pixel>>16)&0xFF);						\
slouken@3054
   348
	b = (Pixel>>24);						\
slouken@3054
   349
	a = (Pixel&0xFF);						\
slouken@3054
   350
}
icculus@1162
   351
#define DISEMBLE_RGBA(buf, bpp, fmt, Pixel, r, g, b, a)			   \
slouken@0
   352
do {									   \
slouken@0
   353
	switch (bpp) {							   \
slouken@0
   354
		case 2:							   \
icculus@1162
   355
			Pixel = *((Uint16 *)(buf));			   \
slouken@2824
   356
			RGBA_FROM_PIXEL(Pixel, fmt, r, g, b, a);	   \
slouken@0
   357
		break;							   \
slouken@0
   358
									   \
slouken@2824
   359
		case 3:	{						   \
slouken@2824
   360
                        if (SDL_BYTEORDER == SDL_LIL_ENDIAN) {		   \
slouken@2824
   361
			        r = *((buf)+fmt->Rshift/8);		   \
slouken@2824
   362
				g = *((buf)+fmt->Gshift/8);		   \
slouken@2824
   363
				b = *((buf)+fmt->Bshift/8);		   \
slouken@0
   364
			} else {					   \
slouken@2824
   365
			        r = *((buf)+2-fmt->Rshift/8);		   \
slouken@2824
   366
				g = *((buf)+2-fmt->Gshift/8);		   \
slouken@2824
   367
				b = *((buf)+2-fmt->Bshift/8);		   \
slouken@0
   368
			}						   \
slouken@2824
   369
			a = 0xFF;					   \
slouken@0
   370
		}							   \
slouken@0
   371
		break;							   \
slouken@0
   372
									   \
slouken@0
   373
		case 4:							   \
icculus@1162
   374
			Pixel = *((Uint32 *)(buf));			   \
slouken@2824
   375
			RGBA_FROM_PIXEL(Pixel, fmt, r, g, b, a);	   \
slouken@0
   376
		break;							   \
slouken@0
   377
									   \
slouken@0
   378
		default:						   \
slouken@2824
   379
		        Pixel; /* stop gcc complaints */		   \
slouken@0
   380
		break;							   \
slouken@0
   381
	}								   \
slouken@2824
   382
} while (0)
slouken@0
   383
slouken@0
   384
/* FIXME: this isn't correct, especially for Alpha (maximum != 255) */
icculus@1162
   385
#define PIXEL_FROM_RGBA(Pixel, fmt, r, g, b, a)				\
slouken@0
   386
{									\
icculus@1162
   387
	Pixel = ((r>>fmt->Rloss)<<fmt->Rshift)|				\
slouken@0
   388
		((g>>fmt->Gloss)<<fmt->Gshift)|				\
slouken@0
   389
		((b>>fmt->Bloss)<<fmt->Bshift)|				\
slouken@535
   390
		((a>>fmt->Aloss)<<fmt->Ashift);				\
slouken@0
   391
}
slouken@0
   392
#define ASSEMBLE_RGBA(buf, bpp, fmt, r, g, b, a)			\
slouken@0
   393
{									\
slouken@0
   394
	switch (bpp) {							\
slouken@0
   395
		case 2: {						\
icculus@1162
   396
			Uint16 Pixel;					\
slouken@0
   397
									\
icculus@1162
   398
			PIXEL_FROM_RGBA(Pixel, fmt, r, g, b, a);	\
icculus@1162
   399
			*((Uint16 *)(buf)) = Pixel;			\
slouken@0
   400
		}							\
slouken@0
   401
		break;							\
slouken@0
   402
									\
slouken@2824
   403
		case 3: {						\
slouken@2824
   404
                        if (SDL_BYTEORDER == SDL_LIL_ENDIAN) {		\
slouken@0
   405
			        *((buf)+fmt->Rshift/8) = r;		\
slouken@0
   406
				*((buf)+fmt->Gshift/8) = g;		\
slouken@0
   407
				*((buf)+fmt->Bshift/8) = b;		\
slouken@0
   408
			} else {					\
slouken@0
   409
			        *((buf)+2-fmt->Rshift/8) = r;		\
slouken@0
   410
				*((buf)+2-fmt->Gshift/8) = g;		\
slouken@0
   411
				*((buf)+2-fmt->Bshift/8) = b;		\
slouken@0
   412
			}						\
slouken@0
   413
		}							\
slouken@0
   414
		break;							\
slouken@0
   415
									\
slouken@0
   416
		case 4: {						\
icculus@1162
   417
			Uint32 Pixel;					\
slouken@0
   418
									\
icculus@1162
   419
			PIXEL_FROM_RGBA(Pixel, fmt, r, g, b, a);	\
icculus@1162
   420
			*((Uint32 *)(buf)) = Pixel;			\
slouken@0
   421
		}							\
slouken@0
   422
		break;							\
slouken@0
   423
	}								\
slouken@0
   424
}
slouken@0
   425
icculus@1162
   426
/* Blend the RGB values of two Pixels based on a source alpha value */
slouken@0
   427
#define ALPHA_BLEND(sR, sG, sB, A, dR, dG, dB)	\
slouken@0
   428
do {						\
icculus@3631
   429
	dR = ((((int)(sR-dR)*(int)A)/255)+dR);	\
icculus@3631
   430
	dG = ((((int)(sG-dG)*(int)A)/255)+dG);	\
icculus@3631
   431
	dB = ((((int)(sB-dB)*(int)A)/255)+dB);	\
icculus@1047
   432
} while(0)
icculus@1047
   433
icculus@1047
   434
slouken@0
   435
/* This is a very useful loop for optimizing blitters */
slouken@553
   436
#if defined(_MSC_VER) && (_MSC_VER == 1300)
slouken@553
   437
/* There's a bug in the Visual C++ 7 optimizer when compiling this code */
slouken@553
   438
#else
slouken@0
   439
#define USE_DUFFS_LOOP
slouken@553
   440
#endif
slouken@0
   441
#ifdef USE_DUFFS_LOOP
slouken@0
   442
slouken@0
   443
/* 8-times unrolled loop */
slouken@0
   444
#define DUFFS_LOOP8(pixel_copy_increment, width)			\
slouken@0
   445
{ int n = (width+7)/8;							\
slouken@91
   446
	switch (width & 7) {						\
slouken@0
   447
	case 0: do {	pixel_copy_increment;				\
slouken@0
   448
	case 7:		pixel_copy_increment;				\
slouken@0
   449
	case 6:		pixel_copy_increment;				\
slouken@0
   450
	case 5:		pixel_copy_increment;				\
slouken@0
   451
	case 4:		pixel_copy_increment;				\
slouken@0
   452
	case 3:		pixel_copy_increment;				\
slouken@0
   453
	case 2:		pixel_copy_increment;				\
slouken@0
   454
	case 1:		pixel_copy_increment;				\
slouken@0
   455
		} while ( --n > 0 );					\
slouken@0
   456
	}								\
slouken@0
   457
}
slouken@0
   458
slouken@0
   459
/* 4-times unrolled loop */
slouken@0
   460
#define DUFFS_LOOP4(pixel_copy_increment, width)			\
slouken@0
   461
{ int n = (width+3)/4;							\
slouken@91
   462
	switch (width & 3) {						\
slouken@0
   463
	case 0: do {	pixel_copy_increment;				\
slouken@0
   464
	case 3:		pixel_copy_increment;				\
slouken@0
   465
	case 2:		pixel_copy_increment;				\
slouken@0
   466
	case 1:		pixel_copy_increment;				\
slouken@3035
   467
		} while (--n > 0);					\
slouken@689
   468
	}								\
slouken@689
   469
}
slouken@689
   470
slouken@0
   471
/* Use the 8-times version of the loop by default */
slouken@0
   472
#define DUFFS_LOOP(pixel_copy_increment, width)				\
slouken@0
   473
	DUFFS_LOOP8(pixel_copy_increment, width)
slouken@0
   474
slouken@3035
   475
/* Special version of Duff's device for even more optimization */
slouken@3035
   476
#define DUFFS_LOOP_124(pixel_copy_increment1,				\
slouken@3035
   477
                       pixel_copy_increment2,				\
slouken@3035
   478
                       pixel_copy_increment4, width)			\
slouken@3035
   479
{ int n = width;							\
slouken@3035
   480
	if (n & 1) {							\
slouken@3035
   481
		pixel_copy_increment1; n -= 1;				\
slouken@3035
   482
	}								\
slouken@3035
   483
	if (n & 2) {							\
slouken@3035
   484
		pixel_copy_increment2; n -= 2;				\
slouken@3035
   485
	}								\
slouken@3035
   486
	if (n) {							\
slouken@3035
   487
		n = (n+7)/ 8;						\
slouken@3035
   488
		switch (n & 4) {					\
slouken@3035
   489
		case 0: do {	pixel_copy_increment4;			\
slouken@3035
   490
		case 4:		pixel_copy_increment4;			\
slouken@3035
   491
			} while (--n > 0);				\
slouken@3035
   492
		}							\
slouken@3035
   493
	}								\
slouken@689
   494
}
slouken@689
   495
slouken@3035
   496
#else
slouken@689
   497
slouken@689
   498
/* Don't use Duff's device to unroll loops */
slouken@0
   499
#define DUFFS_LOOP(pixel_copy_increment, width)				\
slouken@0
   500
{ int n;								\
slouken@0
   501
	for ( n=width; n > 0; --n ) {					\
slouken@0
   502
		pixel_copy_increment;					\
slouken@0
   503
	}								\
slouken@0
   504
}
slouken@0
   505
#define DUFFS_LOOP8(pixel_copy_increment, width)			\
slouken@0
   506
	DUFFS_LOOP(pixel_copy_increment, width)
slouken@0
   507
#define DUFFS_LOOP4(pixel_copy_increment, width)			\
slouken@0
   508
	DUFFS_LOOP(pixel_copy_increment, width)
slouken@3035
   509
#define DUFFS_LOOP_124(pixel_copy_increment1,				\
slouken@3035
   510
                       pixel_copy_increment2,				\
slouken@3035
   511
                       pixel_copy_increment4, width)			\
slouken@3035
   512
	DUFFS_LOOP(pixel_copy_increment1, width)
slouken@0
   513
slouken@0
   514
#endif /* USE_DUFFS_LOOP */
slouken@0
   515
slouken@0
   516
/* Prevent Visual C++ 6.0 from printing out stupid warnings */
slouken@0
   517
#if defined(_MSC_VER) && (_MSC_VER >= 600)
slouken@0
   518
#pragma warning(disable: 4550)
slouken@0
   519
#endif
slouken@0
   520
slouken@0
   521
#endif /* _SDL_blit_h */
slouken@2898
   522
slouken@1895
   523
/* vi: set ts=4 sw=4 expandtab: */