src/video/SDL_blit.h
author Sam Lantinga <slouken@libsdl.org>
Thu, 03 Feb 2011 02:45:29 -0800
changeset 5163 d72793305335
parent 5140 e743b9c3f6d6
child 5184 d976b67150c5
permissions -rw-r--r--
Making the API simpler, moved the surface drawing functions to the software renderer.
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@2266
    57
#define SDL_COPY_COLORKEY           0x00000100
slouken@2266
    58
#define SDL_COPY_NEAREST            0x00000200
slouken@2266
    59
#define SDL_COPY_RLE_DESIRED        0x00001000
slouken@2266
    60
#define SDL_COPY_RLE_COLORKEY       0x00002000
slouken@2266
    61
#define SDL_COPY_RLE_ALPHAKEY       0x00004000
slouken@2853
    62
#define SDL_COPY_RLE_MASK           (SDL_COPY_RLE_DESIRED|SDL_COPY_RLE_COLORKEY|SDL_COPY_RLE_ALPHAKEY)
slouken@2262
    63
slouken@2262
    64
/* SDL blit CPU flags */
slouken@2266
    65
#define SDL_CPU_ANY                 0x00000000
slouken@2266
    66
#define SDL_CPU_MMX                 0x00000001
slouken@2266
    67
#define SDL_CPU_3DNOW               0x00000002
slouken@2266
    68
#define SDL_CPU_SSE                 0x00000004
slouken@2266
    69
#define SDL_CPU_SSE2                0x00000008
slouken@2266
    70
#define SDL_CPU_ALTIVEC_PREFETCH    0x00000010
slouken@2266
    71
#define SDL_CPU_ALTIVEC_NOPREFETCH  0x00000020
slouken@2262
    72
slouken@2267
    73
typedef struct
slouken@2267
    74
{
slouken@2262
    75
    Uint8 *src;
slouken@2262
    76
    int src_w, src_h;
slouken@2262
    77
    int src_pitch;
slouken@2267
    78
    int src_skip;
slouken@2262
    79
    Uint8 *dst;
slouken@2262
    80
    int dst_w, dst_h;
slouken@2262
    81
    int dst_pitch;
slouken@2267
    82
    int dst_skip;
slouken@2262
    83
    SDL_PixelFormat *src_fmt;
slouken@2262
    84
    SDL_PixelFormat *dst_fmt;
slouken@1895
    85
    Uint8 *table;
slouken@2262
    86
    int flags;
slouken@2262
    87
    Uint32 colorkey;
slouken@2262
    88
    Uint8 r, g, b, a;
slouken@0
    89
} SDL_BlitInfo;
slouken@0
    90
slouken@2267
    91
typedef void (SDLCALL * SDL_BlitFunc) (SDL_BlitInfo * info);
slouken@2262
    92
slouken@2267
    93
typedef struct
slouken@2267
    94
{
slouken@2262
    95
    Uint32 src_format;
slouken@2262
    96
    Uint32 dst_format;
slouken@2262
    97
    int flags;
slouken@2262
    98
    int cpu;
slouken@2262
    99
    SDL_BlitFunc func;
slouken@2262
   100
} SDL_BlitFuncEntry;
slouken@0
   101
slouken@0
   102
/* Blit mapping definition */
slouken@1895
   103
typedef struct SDL_BlitMap
slouken@1895
   104
{
slouken@1895
   105
    SDL_Surface *dst;
slouken@1895
   106
    int identity;
slouken@2257
   107
    SDL_blit blit;
slouken@2257
   108
    void *data;
slouken@2262
   109
    SDL_BlitInfo info;
slouken@0
   110
slouken@1895
   111
    /* the version count matches the destination; mismatch indicates
slouken@1895
   112
       an invalid mapping */
slouken@1895
   113
    unsigned int format_version;
slouken@0
   114
} SDL_BlitMap;
slouken@0
   115
slouken@0
   116
/* Functions found in SDL_blit.c */
slouken@1895
   117
extern int SDL_CalculateBlit(SDL_Surface * surface);
slouken@0
   118
slouken@2267
   119
/* Functions found in SDL_blit_*.c */
slouken@2267
   120
extern SDL_BlitFunc SDL_CalculateBlit0(SDL_Surface * surface);
slouken@2267
   121
extern SDL_BlitFunc SDL_CalculateBlit1(SDL_Surface * surface);
slouken@2267
   122
extern SDL_BlitFunc SDL_CalculateBlitN(SDL_Surface * surface);
slouken@2267
   123
extern SDL_BlitFunc SDL_CalculateBlitA(SDL_Surface * surface);
slouken@2263
   124
slouken@0
   125
/*
slouken@0
   126
 * Useful macros for blitting routines
slouken@0
   127
 */
slouken@0
   128
slouken@2249
   129
#if defined(__GNUC__)
slouken@2249
   130
#define DECLARE_ALIGNED(t,v,a)  t __attribute__((aligned(a))) v
slouken@2249
   131
#elif defined(_MSC_VER)
slouken@2251
   132
#define DECLARE_ALIGNED(t,v,a)  __declspec(align(a)) t v
slouken@2249
   133
#else
slouken@2249
   134
#define DECLARE_ALIGNED(t,v,a)  t v
slouken@2249
   135
#endif
slouken@2249
   136
slouken@0
   137
#define FORMAT_EQUAL(A, B)						\
slouken@0
   138
    ((A)->BitsPerPixel == (B)->BitsPerPixel				\
slouken@0
   139
     && ((A)->Rmask == (B)->Rmask) && ((A)->Amask == (B)->Amask))
slouken@0
   140
slouken@0
   141
/* Load pixel of the specified format from a buffer and get its R-G-B values */
slouken@0
   142
/* FIXME: rescale values to 0..255 here? */
icculus@1162
   143
#define RGB_FROM_PIXEL(Pixel, fmt, r, g, b)				\
slouken@0
   144
{									\
icculus@1162
   145
	r = (((Pixel&fmt->Rmask)>>fmt->Rshift)<<fmt->Rloss); 		\
icculus@1162
   146
	g = (((Pixel&fmt->Gmask)>>fmt->Gshift)<<fmt->Gloss); 		\
icculus@1162
   147
	b = (((Pixel&fmt->Bmask)>>fmt->Bshift)<<fmt->Bloss); 		\
slouken@0
   148
}
icculus@1162
   149
#define RGB_FROM_RGB565(Pixel, r, g, b)					\
slouken@0
   150
{									\
icculus@1162
   151
	r = (((Pixel&0xF800)>>11)<<3);		 			\
icculus@1162
   152
	g = (((Pixel&0x07E0)>>5)<<2); 					\
icculus@1162
   153
	b = ((Pixel&0x001F)<<3); 					\
slouken@0
   154
}
icculus@1162
   155
#define RGB_FROM_RGB555(Pixel, r, g, b)					\
slouken@0
   156
{									\
icculus@1162
   157
	r = (((Pixel&0x7C00)>>10)<<3);		 			\
icculus@1162
   158
	g = (((Pixel&0x03E0)>>5)<<3); 					\
icculus@1162
   159
	b = ((Pixel&0x001F)<<3); 					\
slouken@0
   160
}
icculus@1162
   161
#define RGB_FROM_RGB888(Pixel, r, g, b)					\
slouken@0
   162
{									\
icculus@1162
   163
	r = ((Pixel&0xFF0000)>>16);		 			\
icculus@1162
   164
	g = ((Pixel&0xFF00)>>8);		 			\
icculus@1162
   165
	b = (Pixel&0xFF);			 			\
slouken@0
   166
}
icculus@1162
   167
#define RETRIEVE_RGB_PIXEL(buf, bpp, Pixel)				   \
slouken@0
   168
do {									   \
slouken@0
   169
	switch (bpp) {							   \
slouken@0
   170
		case 2:							   \
icculus@1162
   171
			Pixel = *((Uint16 *)(buf));			   \
slouken@0
   172
		break;							   \
slouken@0
   173
									   \
slouken@0
   174
		case 3: {						   \
slouken@0
   175
		        Uint8 *B = (Uint8 *)(buf);			   \
slouken@2824
   176
			if (SDL_BYTEORDER == SDL_LIL_ENDIAN) {		   \
icculus@1162
   177
			        Pixel = B[0] + (B[1] << 8) + (B[2] << 16); \
slouken@0
   178
			} else {					   \
icculus@1162
   179
			        Pixel = (B[0] << 16) + (B[1] << 8) + B[2]; \
slouken@0
   180
			}						   \
slouken@0
   181
		}							   \
slouken@0
   182
		break;							   \
slouken@0
   183
									   \
slouken@0
   184
		case 4:							   \
icculus@1162
   185
			Pixel = *((Uint32 *)(buf));			   \
slouken@0
   186
		break;							   \
slouken@0
   187
									   \
slouken@0
   188
		default:						   \
slouken@2824
   189
		        Pixel; /* stop gcc complaints */		   \
slouken@0
   190
		break;							   \
slouken@0
   191
	}								   \
slouken@2824
   192
} while (0)
slouken@0
   193
icculus@1162
   194
#define DISEMBLE_RGB(buf, bpp, fmt, Pixel, r, g, b)			   \
slouken@0
   195
do {									   \
slouken@0
   196
	switch (bpp) {							   \
slouken@0
   197
		case 2:							   \
icculus@1162
   198
			Pixel = *((Uint16 *)(buf));			   \
slouken@2824
   199
			RGB_FROM_PIXEL(Pixel, fmt, r, g, b);		   \
slouken@0
   200
		break;							   \
slouken@0
   201
									   \
slouken@2824
   202
		case 3:	{						   \
slouken@2824
   203
                        if (SDL_BYTEORDER == SDL_LIL_ENDIAN) {		   \
slouken@2824
   204
			        r = *((buf)+fmt->Rshift/8);		   \
slouken@2824
   205
				g = *((buf)+fmt->Gshift/8);		   \
slouken@2824
   206
				b = *((buf)+fmt->Bshift/8);		   \
slouken@0
   207
			} else {					   \
slouken@2824
   208
			        r = *((buf)+2-fmt->Rshift/8);		   \
slouken@2824
   209
				g = *((buf)+2-fmt->Gshift/8);		   \
slouken@2824
   210
				b = *((buf)+2-fmt->Bshift/8);		   \
slouken@0
   211
			}						   \
slouken@0
   212
		}							   \
slouken@0
   213
		break;							   \
slouken@0
   214
									   \
slouken@0
   215
		case 4:							   \
icculus@1162
   216
			Pixel = *((Uint32 *)(buf));			   \
slouken@2824
   217
			RGB_FROM_PIXEL(Pixel, fmt, r, g, b);		   \
slouken@0
   218
		break;							   \
slouken@0
   219
									   \
slouken@2824
   220
		default:						   \
slouken@2824
   221
		        Pixel; /* stop gcc complaints */		   \
slouken@0
   222
		break;							   \
slouken@0
   223
	}								   \
slouken@2824
   224
} while (0)
slouken@0
   225
slouken@0
   226
/* Assemble R-G-B values into a specified pixel format and store them */
icculus@1162
   227
#define PIXEL_FROM_RGB(Pixel, fmt, r, g, b)				\
slouken@0
   228
{									\
icculus@1162
   229
	Pixel = ((r>>fmt->Rloss)<<fmt->Rshift)|				\
slouken@0
   230
		((g>>fmt->Gloss)<<fmt->Gshift)|				\
slouken@0
   231
		((b>>fmt->Bloss)<<fmt->Bshift);				\
slouken@0
   232
}
icculus@1162
   233
#define RGB565_FROM_RGB(Pixel, r, g, b)					\
slouken@0
   234
{									\
icculus@1162
   235
	Pixel = ((r>>3)<<11)|((g>>2)<<5)|(b>>3);			\
slouken@0
   236
}
icculus@1162
   237
#define RGB555_FROM_RGB(Pixel, r, g, b)					\
slouken@0
   238
{									\
icculus@1162
   239
	Pixel = ((r>>3)<<10)|((g>>3)<<5)|(b>>3);			\
slouken@0
   240
}
icculus@1162
   241
#define RGB888_FROM_RGB(Pixel, r, g, b)					\
slouken@0
   242
{									\
icculus@1162
   243
	Pixel = (r<<16)|(g<<8)|b;					\
slouken@0
   244
}
slouken@2899
   245
#define ARGB8888_FROM_RGBA(Pixel, r, g, b, a)				\
slouken@2899
   246
{									\
slouken@2899
   247
	Pixel = (a<<24)|(r<<16)|(g<<8)|b;				\
slouken@2899
   248
}
slouken@3054
   249
#define RGBA8888_FROM_RGBA(Pixel, r, g, b, a)				\
slouken@3054
   250
{									\
slouken@3054
   251
	Pixel = (r<<24)|(g<<16)|(b<<8)|a;				\
slouken@3054
   252
}
slouken@3054
   253
#define ABGR8888_FROM_RGBA(Pixel, r, g, b, a)				\
slouken@3054
   254
{									\
slouken@3054
   255
	Pixel = (a<<24)|(b<<16)|(g<<8)|r;				\
slouken@3054
   256
}
slouken@3054
   257
#define BGRA8888_FROM_RGBA(Pixel, r, g, b, a)				\
slouken@3054
   258
{									\
slouken@3054
   259
	Pixel = (b<<24)|(g<<16)|(r<<8)|a;				\
slouken@3054
   260
}
slouken@0
   261
#define ASSEMBLE_RGB(buf, bpp, fmt, r, g, b) 				\
slouken@0
   262
{									\
slouken@0
   263
	switch (bpp) {							\
slouken@0
   264
		case 2: {						\
icculus@1162
   265
			Uint16 Pixel;					\
slouken@0
   266
									\
icculus@1162
   267
			PIXEL_FROM_RGB(Pixel, fmt, r, g, b);		\
icculus@1162
   268
			*((Uint16 *)(buf)) = Pixel;			\
slouken@0
   269
		}							\
slouken@0
   270
		break;							\
slouken@0
   271
									\
slouken@0
   272
		case 3: {						\
slouken@2824
   273
                        if (SDL_BYTEORDER == SDL_LIL_ENDIAN) {		\
slouken@0
   274
			        *((buf)+fmt->Rshift/8) = r;		\
slouken@0
   275
				*((buf)+fmt->Gshift/8) = g;		\
slouken@0
   276
				*((buf)+fmt->Bshift/8) = b;		\
slouken@0
   277
			} else {					\
slouken@0
   278
			        *((buf)+2-fmt->Rshift/8) = r;		\
slouken@0
   279
				*((buf)+2-fmt->Gshift/8) = g;		\
slouken@0
   280
				*((buf)+2-fmt->Bshift/8) = b;		\
slouken@0
   281
			}						\
slouken@0
   282
		}							\
slouken@0
   283
		break;							\
slouken@0
   284
									\
slouken@0
   285
		case 4: {						\
icculus@1162
   286
			Uint32 Pixel;					\
slouken@0
   287
									\
icculus@1162
   288
			PIXEL_FROM_RGB(Pixel, fmt, r, g, b);		\
icculus@1162
   289
			*((Uint32 *)(buf)) = Pixel;			\
slouken@0
   290
		}							\
slouken@0
   291
		break;							\
slouken@0
   292
	}								\
slouken@0
   293
}
slouken@0
   294
#define ASSEMBLE_RGB_AMASK(buf, bpp, fmt, r, g, b, Amask)		\
slouken@0
   295
{									\
slouken@0
   296
	switch (bpp) {							\
slouken@0
   297
		case 2: {						\
slouken@0
   298
			Uint16 *bufp;					\
icculus@1162
   299
			Uint16 Pixel;					\
slouken@0
   300
									\
slouken@0
   301
			bufp = (Uint16 *)buf;				\
icculus@1162
   302
			PIXEL_FROM_RGB(Pixel, fmt, r, g, b);		\
icculus@1162
   303
			*bufp = Pixel | (*bufp & Amask);		\
slouken@0
   304
		}							\
slouken@0
   305
		break;							\
slouken@0
   306
									\
slouken@0
   307
		case 3: {						\
slouken@2824
   308
                        if (SDL_BYTEORDER == SDL_LIL_ENDIAN) {		\
slouken@0
   309
			        *((buf)+fmt->Rshift/8) = r;		\
slouken@0
   310
				*((buf)+fmt->Gshift/8) = g;		\
slouken@0
   311
				*((buf)+fmt->Bshift/8) = b;		\
slouken@0
   312
			} else {					\
slouken@0
   313
			        *((buf)+2-fmt->Rshift/8) = r;		\
slouken@0
   314
				*((buf)+2-fmt->Gshift/8) = g;		\
slouken@0
   315
				*((buf)+2-fmt->Bshift/8) = b;		\
slouken@0
   316
			}						\
slouken@0
   317
		}							\
slouken@0
   318
		break;							\
slouken@0
   319
									\
slouken@0
   320
		case 4: {						\
slouken@0
   321
			Uint32 *bufp;					\
icculus@1162
   322
			Uint32 Pixel;					\
slouken@0
   323
									\
slouken@0
   324
			bufp = (Uint32 *)buf;				\
icculus@1162
   325
			PIXEL_FROM_RGB(Pixel, fmt, r, g, b);		\
icculus@1162
   326
			*bufp = Pixel | (*bufp & Amask);		\
slouken@0
   327
		}							\
slouken@0
   328
		break;							\
slouken@0
   329
	}								\
slouken@0
   330
}
slouken@0
   331
slouken@0
   332
/* FIXME: Should we rescale alpha into 0..255 here? */
icculus@1162
   333
#define RGBA_FROM_PIXEL(Pixel, fmt, r, g, b, a)				\
slouken@0
   334
{									\
icculus@1162
   335
	r = ((Pixel&fmt->Rmask)>>fmt->Rshift)<<fmt->Rloss; 		\
icculus@1162
   336
	g = ((Pixel&fmt->Gmask)>>fmt->Gshift)<<fmt->Gloss; 		\
icculus@1162
   337
	b = ((Pixel&fmt->Bmask)>>fmt->Bshift)<<fmt->Bloss; 		\
icculus@1162
   338
	a = ((Pixel&fmt->Amask)>>fmt->Ashift)<<fmt->Aloss;	 	\
slouken@0
   339
}
icculus@1162
   340
#define RGBA_FROM_8888(Pixel, fmt, r, g, b, a)	\
slouken@0
   341
{						\
icculus@1162
   342
	r = (Pixel&fmt->Rmask)>>fmt->Rshift;	\
icculus@1162
   343
	g = (Pixel&fmt->Gmask)>>fmt->Gshift;	\
icculus@1162
   344
	b = (Pixel&fmt->Bmask)>>fmt->Bshift;	\
icculus@1162
   345
	a = (Pixel&fmt->Amask)>>fmt->Ashift;	\
slouken@0
   346
}
icculus@1162
   347
#define RGBA_FROM_RGBA8888(Pixel, r, g, b, a)				\
slouken@0
   348
{									\
icculus@1162
   349
	r = (Pixel>>24);						\
icculus@1162
   350
	g = ((Pixel>>16)&0xFF);						\
icculus@1162
   351
	b = ((Pixel>>8)&0xFF);						\
icculus@1162
   352
	a = (Pixel&0xFF);						\
slouken@0
   353
}
icculus@1162
   354
#define RGBA_FROM_ARGB8888(Pixel, r, g, b, a)				\
slouken@0
   355
{									\
icculus@1162
   356
	r = ((Pixel>>16)&0xFF);						\
icculus@1162
   357
	g = ((Pixel>>8)&0xFF);						\
icculus@1162
   358
	b = (Pixel&0xFF);						\
icculus@1162
   359
	a = (Pixel>>24);						\
slouken@0
   360
}
icculus@1162
   361
#define RGBA_FROM_ABGR8888(Pixel, r, g, b, a)				\
slouken@0
   362
{									\
icculus@1162
   363
	r = (Pixel&0xFF);						\
icculus@1162
   364
	g = ((Pixel>>8)&0xFF);						\
icculus@1162
   365
	b = ((Pixel>>16)&0xFF);						\
icculus@1162
   366
	a = (Pixel>>24);						\
slouken@0
   367
}
slouken@3054
   368
#define RGBA_FROM_BGRA8888(Pixel, r, g, b, a)				\
slouken@3054
   369
{									\
slouken@3054
   370
	r = ((Pixel>>8)&0xFF);						\
slouken@3054
   371
	g = ((Pixel>>16)&0xFF);						\
slouken@3054
   372
	b = (Pixel>>24);						\
slouken@3054
   373
	a = (Pixel&0xFF);						\
slouken@3054
   374
}
icculus@1162
   375
#define DISEMBLE_RGBA(buf, bpp, fmt, Pixel, r, g, b, a)			   \
slouken@0
   376
do {									   \
slouken@0
   377
	switch (bpp) {							   \
slouken@0
   378
		case 2:							   \
icculus@1162
   379
			Pixel = *((Uint16 *)(buf));			   \
slouken@2824
   380
			RGBA_FROM_PIXEL(Pixel, fmt, r, g, b, a);	   \
slouken@0
   381
		break;							   \
slouken@0
   382
									   \
slouken@2824
   383
		case 3:	{						   \
slouken@2824
   384
                        if (SDL_BYTEORDER == SDL_LIL_ENDIAN) {		   \
slouken@2824
   385
			        r = *((buf)+fmt->Rshift/8);		   \
slouken@2824
   386
				g = *((buf)+fmt->Gshift/8);		   \
slouken@2824
   387
				b = *((buf)+fmt->Bshift/8);		   \
slouken@0
   388
			} else {					   \
slouken@2824
   389
			        r = *((buf)+2-fmt->Rshift/8);		   \
slouken@2824
   390
				g = *((buf)+2-fmt->Gshift/8);		   \
slouken@2824
   391
				b = *((buf)+2-fmt->Bshift/8);		   \
slouken@0
   392
			}						   \
slouken@2824
   393
			a = 0xFF;					   \
slouken@0
   394
		}							   \
slouken@0
   395
		break;							   \
slouken@0
   396
									   \
slouken@0
   397
		case 4:							   \
icculus@1162
   398
			Pixel = *((Uint32 *)(buf));			   \
slouken@2824
   399
			RGBA_FROM_PIXEL(Pixel, fmt, r, g, b, a);	   \
slouken@0
   400
		break;							   \
slouken@0
   401
									   \
slouken@0
   402
		default:						   \
slouken@2824
   403
		        Pixel; /* stop gcc complaints */		   \
slouken@0
   404
		break;							   \
slouken@0
   405
	}								   \
slouken@2824
   406
} while (0)
slouken@0
   407
slouken@0
   408
/* FIXME: this isn't correct, especially for Alpha (maximum != 255) */
icculus@1162
   409
#define PIXEL_FROM_RGBA(Pixel, fmt, r, g, b, a)				\
slouken@0
   410
{									\
icculus@1162
   411
	Pixel = ((r>>fmt->Rloss)<<fmt->Rshift)|				\
slouken@0
   412
		((g>>fmt->Gloss)<<fmt->Gshift)|				\
slouken@0
   413
		((b>>fmt->Bloss)<<fmt->Bshift)|				\
slouken@535
   414
		((a>>fmt->Aloss)<<fmt->Ashift);				\
slouken@0
   415
}
slouken@0
   416
#define ASSEMBLE_RGBA(buf, bpp, fmt, r, g, b, a)			\
slouken@0
   417
{									\
slouken@0
   418
	switch (bpp) {							\
slouken@0
   419
		case 2: {						\
icculus@1162
   420
			Uint16 Pixel;					\
slouken@0
   421
									\
icculus@1162
   422
			PIXEL_FROM_RGBA(Pixel, fmt, r, g, b, a);	\
icculus@1162
   423
			*((Uint16 *)(buf)) = Pixel;			\
slouken@0
   424
		}							\
slouken@0
   425
		break;							\
slouken@0
   426
									\
slouken@2824
   427
		case 3: {						\
slouken@2824
   428
                        if (SDL_BYTEORDER == SDL_LIL_ENDIAN) {		\
slouken@0
   429
			        *((buf)+fmt->Rshift/8) = r;		\
slouken@0
   430
				*((buf)+fmt->Gshift/8) = g;		\
slouken@0
   431
				*((buf)+fmt->Bshift/8) = b;		\
slouken@0
   432
			} else {					\
slouken@0
   433
			        *((buf)+2-fmt->Rshift/8) = r;		\
slouken@0
   434
				*((buf)+2-fmt->Gshift/8) = g;		\
slouken@0
   435
				*((buf)+2-fmt->Bshift/8) = b;		\
slouken@0
   436
			}						\
slouken@0
   437
		}							\
slouken@0
   438
		break;							\
slouken@0
   439
									\
slouken@0
   440
		case 4: {						\
icculus@1162
   441
			Uint32 Pixel;					\
slouken@0
   442
									\
icculus@1162
   443
			PIXEL_FROM_RGBA(Pixel, fmt, r, g, b, a);	\
icculus@1162
   444
			*((Uint32 *)(buf)) = Pixel;			\
slouken@0
   445
		}							\
slouken@0
   446
		break;							\
slouken@0
   447
	}								\
slouken@0
   448
}
slouken@0
   449
icculus@1162
   450
/* Blend the RGB values of two Pixels based on a source alpha value */
slouken@0
   451
#define ALPHA_BLEND(sR, sG, sB, A, dR, dG, dB)	\
slouken@0
   452
do {						\
icculus@3631
   453
	dR = ((((int)(sR-dR)*(int)A)/255)+dR);	\
icculus@3631
   454
	dG = ((((int)(sG-dG)*(int)A)/255)+dG);	\
icculus@3631
   455
	dB = ((((int)(sB-dB)*(int)A)/255)+dB);	\
icculus@1047
   456
} while(0)
icculus@1047
   457
icculus@1047
   458
slouken@0
   459
/* This is a very useful loop for optimizing blitters */
slouken@553
   460
#if defined(_MSC_VER) && (_MSC_VER == 1300)
slouken@553
   461
/* There's a bug in the Visual C++ 7 optimizer when compiling this code */
slouken@553
   462
#else
slouken@0
   463
#define USE_DUFFS_LOOP
slouken@553
   464
#endif
slouken@0
   465
#ifdef USE_DUFFS_LOOP
slouken@0
   466
slouken@0
   467
/* 8-times unrolled loop */
slouken@0
   468
#define DUFFS_LOOP8(pixel_copy_increment, width)			\
slouken@0
   469
{ int n = (width+7)/8;							\
slouken@91
   470
	switch (width & 7) {						\
slouken@0
   471
	case 0: do {	pixel_copy_increment;				\
slouken@0
   472
	case 7:		pixel_copy_increment;				\
slouken@0
   473
	case 6:		pixel_copy_increment;				\
slouken@0
   474
	case 5:		pixel_copy_increment;				\
slouken@0
   475
	case 4:		pixel_copy_increment;				\
slouken@0
   476
	case 3:		pixel_copy_increment;				\
slouken@0
   477
	case 2:		pixel_copy_increment;				\
slouken@0
   478
	case 1:		pixel_copy_increment;				\
slouken@0
   479
		} while ( --n > 0 );					\
slouken@0
   480
	}								\
slouken@0
   481
}
slouken@0
   482
slouken@0
   483
/* 4-times unrolled loop */
slouken@0
   484
#define DUFFS_LOOP4(pixel_copy_increment, width)			\
slouken@0
   485
{ int n = (width+3)/4;							\
slouken@91
   486
	switch (width & 3) {						\
slouken@0
   487
	case 0: do {	pixel_copy_increment;				\
slouken@0
   488
	case 3:		pixel_copy_increment;				\
slouken@0
   489
	case 2:		pixel_copy_increment;				\
slouken@0
   490
	case 1:		pixel_copy_increment;				\
slouken@3035
   491
		} while (--n > 0);					\
slouken@689
   492
	}								\
slouken@689
   493
}
slouken@689
   494
slouken@0
   495
/* Use the 8-times version of the loop by default */
slouken@0
   496
#define DUFFS_LOOP(pixel_copy_increment, width)				\
slouken@0
   497
	DUFFS_LOOP8(pixel_copy_increment, width)
slouken@0
   498
slouken@3035
   499
/* Special version of Duff's device for even more optimization */
slouken@3035
   500
#define DUFFS_LOOP_124(pixel_copy_increment1,				\
slouken@3035
   501
                       pixel_copy_increment2,				\
slouken@3035
   502
                       pixel_copy_increment4, width)			\
slouken@3035
   503
{ int n = width;							\
slouken@3035
   504
	if (n & 1) {							\
slouken@3035
   505
		pixel_copy_increment1; n -= 1;				\
slouken@3035
   506
	}								\
slouken@3035
   507
	if (n & 2) {							\
slouken@3035
   508
		pixel_copy_increment2; n -= 2;				\
slouken@3035
   509
	}								\
slouken@3035
   510
	if (n) {							\
slouken@3035
   511
		n = (n+7)/ 8;						\
slouken@3035
   512
		switch (n & 4) {					\
slouken@3035
   513
		case 0: do {	pixel_copy_increment4;			\
slouken@3035
   514
		case 4:		pixel_copy_increment4;			\
slouken@3035
   515
			} while (--n > 0);				\
slouken@3035
   516
		}							\
slouken@3035
   517
	}								\
slouken@689
   518
}
slouken@689
   519
slouken@3035
   520
#else
slouken@689
   521
slouken@689
   522
/* Don't use Duff's device to unroll loops */
slouken@0
   523
#define DUFFS_LOOP(pixel_copy_increment, width)				\
slouken@0
   524
{ int n;								\
slouken@0
   525
	for ( n=width; n > 0; --n ) {					\
slouken@0
   526
		pixel_copy_increment;					\
slouken@0
   527
	}								\
slouken@0
   528
}
slouken@0
   529
#define DUFFS_LOOP8(pixel_copy_increment, width)			\
slouken@0
   530
	DUFFS_LOOP(pixel_copy_increment, width)
slouken@0
   531
#define DUFFS_LOOP4(pixel_copy_increment, width)			\
slouken@0
   532
	DUFFS_LOOP(pixel_copy_increment, width)
slouken@3035
   533
#define DUFFS_LOOP_124(pixel_copy_increment1,				\
slouken@3035
   534
                       pixel_copy_increment2,				\
slouken@3035
   535
                       pixel_copy_increment4, width)			\
slouken@3035
   536
	DUFFS_LOOP(pixel_copy_increment1, width)
slouken@0
   537
slouken@0
   538
#endif /* USE_DUFFS_LOOP */
slouken@0
   539
slouken@0
   540
/* Prevent Visual C++ 6.0 from printing out stupid warnings */
slouken@0
   541
#if defined(_MSC_VER) && (_MSC_VER >= 600)
slouken@0
   542
#pragma warning(disable: 4550)
slouken@0
   543
#endif
slouken@0
   544
slouken@0
   545
#endif /* _SDL_blit_h */
slouken@2898
   546
slouken@1895
   547
/* vi: set ts=4 sw=4 expandtab: */