From faaae44a1ef8328a716e8f3a494cbe829691d942 Mon Sep 17 00:00:00 2001 From: diego Date: Sat, 28 Apr 2007 11:44:49 +0000 Subject: cosmetics attack, part I: Remove all tabs and prettyprint/reindent the code. git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@23158 b3059339-0415-0410-9bf9-f77b7e298cf2 --- libswscale/cs_test.c | 6 +- libswscale/rgb2rgb.c | 609 +++-- libswscale/rgb2rgb.h | 69 +- libswscale/rgb2rgb_template.c | 4619 ++++++++++++++++----------------- libswscale/swscale_altivec_template.c | 852 +++--- 5 files changed, 3074 insertions(+), 3081 deletions(-) diff --git a/libswscale/cs_test.c b/libswscale/cs_test.c index 552fafaf6a..876f270411 100644 --- a/libswscale/cs_test.c +++ b/libswscale/cs_test.c @@ -142,7 +142,7 @@ int main(int argc, char **argv) for(i=0; i> 2; - for(i=0; i BGR24 (= B,G,R) */ - dst[3*i + 0] = src[4*i + 1]; - dst[3*i + 1] = src[4*i + 2]; - dst[3*i + 2] = src[4*i + 3]; - #else - dst[3*i + 0] = src[4*i + 2]; - dst[3*i + 1] = src[4*i + 1]; - dst[3*i + 2] = src[4*i + 0]; - #endif - } + long i; + long num_pixels = src_size >> 2; + for (i=0; i BGR24 (= B,G,R) */ + dst[3*i + 0] = src[4*i + 1]; + dst[3*i + 1] = src[4*i + 2]; + dst[3*i + 2] = src[4*i + 3]; + #else + dst[3*i + 0] = src[4*i + 2]; + dst[3*i + 1] = src[4*i + 1]; + dst[3*i + 2] = src[4*i + 0]; + #endif + } } void rgb24tobgr32(const uint8_t *src, uint8_t *dst, long src_size) { - long i; - for(i=0; 3*i BGR32 (= A,R,G,B) */ - dst[4*i + 0] = 0; - dst[4*i + 1] = src[3*i + 0]; - dst[4*i + 2] = src[3*i + 1]; - dst[4*i + 3] = src[3*i + 2]; - #else - dst[4*i + 0] = src[3*i + 2]; - dst[4*i + 1] = src[3*i + 1]; - dst[4*i + 2] = src[3*i + 0]; - dst[4*i + 3] = 0; - #endif - } + long i; + for (i=0; 3*i BGR32 (= A,R,G,B) */ + dst[4*i + 0] = 0; + dst[4*i + 1] = src[3*i + 0]; + dst[4*i + 2] = src[3*i + 1]; + dst[4*i + 3] = src[3*i + 2]; + #else + dst[4*i + 0] = src[3*i + 2]; + dst[4*i + 1] = src[3*i + 1]; + dst[4*i + 2] = src[3*i + 0]; + dst[4*i + 3] = 0; + #endif + } } void rgb16tobgr32(const uint8_t *src, uint8_t *dst, long src_size) { - const uint16_t *end; - uint8_t *d = (uint8_t *)dst; - const uint16_t *s = (uint16_t *)src; - end = s + src_size/2; - while(s < end) - { - register uint16_t bgr; - bgr = *s++; - #ifdef WORDS_BIGENDIAN - *d++ = 0; - *d++ = (bgr&0x1F)<<3; - *d++ = (bgr&0x7E0)>>3; - *d++ = (bgr&0xF800)>>8; - #else - *d++ = (bgr&0xF800)>>8; - *d++ = (bgr&0x7E0)>>3; - *d++ = (bgr&0x1F)<<3; - *d++ = 0; - #endif - } + const uint16_t *end; + uint8_t *d = (uint8_t *)dst; + const uint16_t *s = (uint16_t *)src; + end = s + src_size/2; + while (s < end) + { + register uint16_t bgr; + bgr = *s++; + #ifdef WORDS_BIGENDIAN + *d++ = 0; + *d++ = (bgr&0x1F)<<3; + *d++ = (bgr&0x7E0)>>3; + *d++ = (bgr&0xF800)>>8; + #else + *d++ = (bgr&0xF800)>>8; + *d++ = (bgr&0x7E0)>>3; + *d++ = (bgr&0x1F)<<3; + *d++ = 0; + #endif + } } void rgb16tobgr24(const uint8_t *src, uint8_t *dst, long src_size) { - const uint16_t *end; - uint8_t *d = (uint8_t *)dst; - const uint16_t *s = (const uint16_t *)src; - end = s + src_size/2; - while(s < end) - { - register uint16_t bgr; - bgr = *s++; - *d++ = (bgr&0xF800)>>8; - *d++ = (bgr&0x7E0)>>3; - *d++ = (bgr&0x1F)<<3; - } + const uint16_t *end; + uint8_t *d = (uint8_t *)dst; + const uint16_t *s = (const uint16_t *)src; + end = s + src_size/2; + while (s < end) + { + register uint16_t bgr; + bgr = *s++; + *d++ = (bgr&0xF800)>>8; + *d++ = (bgr&0x7E0)>>3; + *d++ = (bgr&0x1F)<<3; + } } void rgb16tobgr16(const uint8_t *src, uint8_t *dst, long src_size) { - long i; - long num_pixels = src_size >> 1; - - for(i=0; i>5; - b = (rgb&0xF800)>>11; - dst[2*i] = (b&0x1F) | ((g&0x3F)<<5) | ((r&0x1F)<<11); - } + long i; + long num_pixels = src_size >> 1; + + for (i=0; i>5; + b = (rgb&0xF800)>>11; + dst[2*i] = (b&0x1F) | ((g&0x3F)<<5) | ((r&0x1F)<<11); + } } void rgb16tobgr15(const uint8_t *src, uint8_t *dst, long src_size) { - long i; - long num_pixels = src_size >> 1; - - for(i=0; i>5; - b = (rgb&0xF800)>>11; - dst[2*i] = (b&0x1F) | ((g&0x1F)<<5) | ((r&0x1F)<<10); - } + long i; + long num_pixels = src_size >> 1; + + for (i=0; i>5; + b = (rgb&0xF800)>>11; + dst[2*i] = (b&0x1F) | ((g&0x1F)<<5) | ((r&0x1F)<<10); + } } void rgb15tobgr32(const uint8_t *src, uint8_t *dst, long src_size) { - const uint16_t *end; - uint8_t *d = (uint8_t *)dst; - const uint16_t *s = (const uint16_t *)src; - end = s + src_size/2; - while(s < end) - { - register uint16_t bgr; - bgr = *s++; - #ifdef WORDS_BIGENDIAN - *d++ = 0; - *d++ = (bgr&0x1F)<<3; - *d++ = (bgr&0x3E0)>>2; - *d++ = (bgr&0x7C00)>>7; - #else - *d++ = (bgr&0x7C00)>>7; - *d++ = (bgr&0x3E0)>>2; - *d++ = (bgr&0x1F)<<3; - *d++ = 0; - #endif - } + const uint16_t *end; + uint8_t *d = (uint8_t *)dst; + const uint16_t *s = (const uint16_t *)src; + end = s + src_size/2; + while (s < end) + { + register uint16_t bgr; + bgr = *s++; + #ifdef WORDS_BIGENDIAN + *d++ = 0; + *d++ = (bgr&0x1F)<<3; + *d++ = (bgr&0x3E0)>>2; + *d++ = (bgr&0x7C00)>>7; + #else + *d++ = (bgr&0x7C00)>>7; + *d++ = (bgr&0x3E0)>>2; + *d++ = (bgr&0x1F)<<3; + *d++ = 0; + #endif + } } void rgb15tobgr24(const uint8_t *src, uint8_t *dst, long src_size) { - const uint16_t *end; - uint8_t *d = (uint8_t *)dst; - const uint16_t *s = (uint16_t *)src; - end = s + src_size/2; - while(s < end) - { - register uint16_t bgr; - bgr = *s++; - *d++ = (bgr&0x7C00)>>7; - *d++ = (bgr&0x3E0)>>2; - *d++ = (bgr&0x1F)<<3; - } + const uint16_t *end; + uint8_t *d = (uint8_t *)dst; + const uint16_t *s = (uint16_t *)src; + end = s + src_size/2; + while (s < end) + { + register uint16_t bgr; + bgr = *s++; + *d++ = (bgr&0x7C00)>>7; + *d++ = (bgr&0x3E0)>>2; + *d++ = (bgr&0x1F)<<3; + } } void rgb15tobgr16(const uint8_t *src, uint8_t *dst, long src_size) { - long i; - long num_pixels = src_size >> 1; - - for(i=0; i>5; - b = (rgb&0x7C00)>>10; - dst[2*i] = (b&0x1F) | ((g&0x3F)<<5) | ((r&0x1F)<<11); - } + long i; + long num_pixels = src_size >> 1; + + for (i=0; i>5; + b = (rgb&0x7C00)>>10; + dst[2*i] = (b&0x1F) | ((g&0x3F)<<5) | ((r&0x1F)<<11); + } } void rgb15tobgr15(const uint8_t *src, uint8_t *dst, long src_size) { - long i; - long num_pixels = src_size >> 1; - - for(i=0; i>5; - b = (rgb&0x7C00)>>10; - dst[2*i] = (b&0x1F) | ((g&0x1F)<<5) | ((r&0x1F)<<10); - } + long i; + long num_pixels = src_size >> 1; + + for (i=0; i>5; + b = (rgb&0x7C00)>>10; + dst[2*i] = (b&0x1F) | ((g&0x1F)<<5) | ((r&0x1F)<<10); + } } void rgb8tobgr8(const uint8_t *src, uint8_t *dst, long src_size) { - long i; - long num_pixels = src_size; - for(i=0; i>3; - b = (rgb&0xC0)>>6; - dst[i] = ((b<<1)&0x07) | ((g&0x07)<<3) | ((r&0x03)<<6); - } + long i; + long num_pixels = src_size; + for (i=0; i>3; + b = (rgb&0xC0)>>6; + dst[i] = ((b<<1)&0x07) | ((g&0x07)<<3) | ((r&0x03)<<6); + } } diff --git a/libswscale/rgb2rgb.h b/libswscale/rgb2rgb.h index 769811fb8f..5ad3ec71b4 100644 --- a/libswscale/rgb2rgb.h +++ b/libswscale/rgb2rgb.h @@ -1,5 +1,4 @@ /* - * * rgb2rgb.h, Software RGB to RGB convertor * pluralize by Software PAL8 to RGB convertor * Software YUV to YUV convertor @@ -30,18 +29,18 @@ // Note: do not fix the dependence on stdio.h /* A full collection of rgb to rgb(bgr) convertors */ -extern void (*rgb24to32)(const uint8_t *src,uint8_t *dst,long src_size); -extern void (*rgb24to16)(const uint8_t *src,uint8_t *dst,long src_size); -extern void (*rgb24to15)(const uint8_t *src,uint8_t *dst,long src_size); -extern void (*rgb32to24)(const uint8_t *src,uint8_t *dst,long src_size); -extern void (*rgb32to16)(const uint8_t *src,uint8_t *dst,long src_size); -extern void (*rgb32to15)(const uint8_t *src,uint8_t *dst,long src_size); -extern void (*rgb15to16)(const uint8_t *src,uint8_t *dst,long src_size); -extern void (*rgb15to24)(const uint8_t *src,uint8_t *dst,long src_size); -extern void (*rgb15to32)(const uint8_t *src,uint8_t *dst,long src_size); -extern void (*rgb16to15)(const uint8_t *src,uint8_t *dst,long src_size); -extern void (*rgb16to24)(const uint8_t *src,uint8_t *dst,long src_size); -extern void (*rgb16to32)(const uint8_t *src,uint8_t *dst,long src_size); +extern void (*rgb24to32) (const uint8_t *src, uint8_t *dst, long src_size); +extern void (*rgb24to16) (const uint8_t *src, uint8_t *dst, long src_size); +extern void (*rgb24to15) (const uint8_t *src, uint8_t *dst, long src_size); +extern void (*rgb32to24) (const uint8_t *src, uint8_t *dst, long src_size); +extern void (*rgb32to16) (const uint8_t *src, uint8_t *dst, long src_size); +extern void (*rgb32to15) (const uint8_t *src, uint8_t *dst, long src_size); +extern void (*rgb15to16) (const uint8_t *src, uint8_t *dst, long src_size); +extern void (*rgb15to24) (const uint8_t *src, uint8_t *dst, long src_size); +extern void (*rgb15to32) (const uint8_t *src, uint8_t *dst, long src_size); +extern void (*rgb16to15) (const uint8_t *src, uint8_t *dst, long src_size); +extern void (*rgb16to24) (const uint8_t *src, uint8_t *dst, long src_size); +extern void (*rgb16to32) (const uint8_t *src, uint8_t *dst, long src_size); extern void (*rgb24tobgr24)(const uint8_t *src, uint8_t *dst, long src_size); extern void (*rgb24tobgr16)(const uint8_t *src, uint8_t *dst, long src_size); extern void (*rgb24tobgr15)(const uint8_t *src, uint8_t *dst, long src_size); @@ -59,7 +58,7 @@ extern void rgb15tobgr32(const uint8_t *src, uint8_t *dst, long src_size); extern void rgb15tobgr24(const uint8_t *src, uint8_t *dst, long src_size); extern void rgb15tobgr16(const uint8_t *src, uint8_t *dst, long src_size); extern void rgb15tobgr15(const uint8_t *src, uint8_t *dst, long src_size); -extern void rgb8tobgr8(const uint8_t *src, uint8_t *dst, long src_size); +extern void rgb8tobgr8 (const uint8_t *src, uint8_t *dst, long src_size); extern void palette8torgb32(const uint8_t *src, uint8_t *dst, long num_pixels, const uint8_t *palette); @@ -85,16 +84,16 @@ extern void palette8tobgr15(const uint8_t *src, uint8_t *dst, long num_pixels, c * problem for anyone then tell me, and ill fix it) */ extern void (*yv12toyuy2)(const uint8_t *ysrc, const uint8_t *usrc, const uint8_t *vsrc, uint8_t *dst, - long width, long height, - long lumStride, long chromStride, long dstStride); + long width, long height, + long lumStride, long chromStride, long dstStride); /** * * width should be a multiple of 16 */ extern void (*yuv422ptoyuy2)(const uint8_t *ysrc, const uint8_t *usrc, const uint8_t *vsrc, uint8_t *dst, - long width, long height, - long lumStride, long chromStride, long dstStride); + long width, long height, + long lumStride, long chromStride, long dstStride); /** * @@ -102,8 +101,8 @@ extern void (*yuv422ptoyuy2)(const uint8_t *ysrc, const uint8_t *usrc, const uin * problem for anyone then tell me, and ill fix it) */ extern void (*yuy2toyv12)(const uint8_t *src, uint8_t *ydst, uint8_t *udst, uint8_t *vdst, - long width, long height, - long lumStride, long chromStride, long srcStride); + long width, long height, + long lumStride, long chromStride, long srcStride); /** * @@ -111,8 +110,8 @@ extern void (*yuy2toyv12)(const uint8_t *src, uint8_t *ydst, uint8_t *udst, uint * problem for anyone then tell me, and ill fix it) */ extern void (*yv12touyvy)(const uint8_t *ysrc, const uint8_t *usrc, const uint8_t *vsrc, uint8_t *dst, - long width, long height, - long lumStride, long chromStride, long dstStride); + long width, long height, + long lumStride, long chromStride, long dstStride); /** * @@ -121,26 +120,26 @@ extern void (*yv12touyvy)(const uint8_t *ysrc, const uint8_t *usrc, const uint8_ * chrominance data is only taken from every secound line others are ignored FIXME write HQ version */ extern void (*rgb24toyv12)(const uint8_t *src, uint8_t *ydst, uint8_t *udst, uint8_t *vdst, - long width, long height, - long lumStride, long chromStride, long srcStride); + long width, long height, + long lumStride, long chromStride, long srcStride); extern void (*planar2x)(const uint8_t *src, uint8_t *dst, long width, long height, - long srcStride, long dstStride); + long srcStride, long dstStride); extern void (*interleaveBytes)(uint8_t *src1, uint8_t *src2, uint8_t *dst, - long width, long height, long src1Stride, - long src2Stride, long dstStride); + long width, long height, long src1Stride, + long src2Stride, long dstStride); extern void (*vu9_to_vu12)(const uint8_t *src1, const uint8_t *src2, - uint8_t *dst1, uint8_t *dst2, - long width, long height, - long srcStride1, long srcStride2, - long dstStride1, long dstStride2); + uint8_t *dst1, uint8_t *dst2, + long width, long height, + long srcStride1, long srcStride2, + long dstStride1, long dstStride2); extern void (*yvu9_to_yuy2)(const uint8_t *src1, const uint8_t *src2, const uint8_t *src3, - uint8_t *dst, - long width, long height, - long srcStride1, long srcStride2, - long srcStride3, long dstStride); + uint8_t *dst, + long width, long height, + long srcStride1, long srcStride2, + long srcStride3, long dstStride); void sws_rgb2rgb_init(int flags); diff --git a/libswscale/rgb2rgb_template.c b/libswscale/rgb2rgb_template.c index c089e39fa7..6dab0aa539 100644 --- a/libswscale/rgb2rgb_template.c +++ b/libswscale/rgb2rgb_template.c @@ -1,5 +1,4 @@ /* - * * rgb2rgb.c, Software RGB to RGB convertor * pluralize by Software PAL8 to RGB convertor * Software YUV to YUV convertor @@ -53,11 +52,11 @@ #ifdef HAVE_3DNOW #define PREFETCH "prefetch" #define PREFETCHW "prefetchw" -#define PAVGB "pavgusb" +#define PAVGB "pavgusb" #elif defined ( HAVE_MMX2 ) #define PREFETCH "prefetchnta" #define PREFETCHW "prefetcht0" -#define PAVGB "pavgb" +#define PAVGB "pavgb" #else #ifdef __APPLE__ #define PREFETCH "#" @@ -85,148 +84,148 @@ static inline void RENAME(rgb24to32)(const uint8_t *src,uint8_t *dst,long src_size) { - uint8_t *dest = dst; - const uint8_t *s = src; - const uint8_t *end; -#ifdef HAVE_MMX - const uint8_t *mm_end; -#endif - end = s + src_size; -#ifdef HAVE_MMX - __asm __volatile(PREFETCH" %0"::"m"(*s):"memory"); - mm_end = end - 23; - __asm __volatile("movq %0, %%mm7"::"m"(mask32):"memory"); - while(s < mm_end) - { - __asm __volatile( - PREFETCH" 32%1\n\t" - "movd %1, %%mm0\n\t" - "punpckldq 3%1, %%mm0\n\t" - "movd 6%1, %%mm1\n\t" - "punpckldq 9%1, %%mm1\n\t" - "movd 12%1, %%mm2\n\t" - "punpckldq 15%1, %%mm2\n\t" - "movd 18%1, %%mm3\n\t" - "punpckldq 21%1, %%mm3\n\t" - "pand %%mm7, %%mm0\n\t" - "pand %%mm7, %%mm1\n\t" - "pand %%mm7, %%mm2\n\t" - "pand %%mm7, %%mm3\n\t" - MOVNTQ" %%mm0, %0\n\t" - MOVNTQ" %%mm1, 8%0\n\t" - MOVNTQ" %%mm2, 16%0\n\t" - MOVNTQ" %%mm3, 24%0" - :"=m"(*dest) - :"m"(*s) - :"memory"); - dest += 32; - s += 24; - } - __asm __volatile(SFENCE:::"memory"); - __asm __volatile(EMMS:::"memory"); -#endif - while(s < end) - { -#ifdef WORDS_BIGENDIAN - /* RGB24 (= R,G,B) -> RGB32 (= A,B,G,R) */ - *dest++ = 0; - *dest++ = s[2]; - *dest++ = s[1]; - *dest++ = s[0]; - s+=3; -#else - *dest++ = *s++; - *dest++ = *s++; - *dest++ = *s++; - *dest++ = 0; -#endif - } + uint8_t *dest = dst; + const uint8_t *s = src; + const uint8_t *end; + #ifdef HAVE_MMX + const uint8_t *mm_end; + #endif + end = s + src_size; + #ifdef HAVE_MMX + __asm __volatile(PREFETCH" %0"::"m"(*s):"memory"); + mm_end = end - 23; + __asm __volatile("movq %0, %%mm7"::"m"(mask32):"memory"); + while (s < mm_end) + { + __asm __volatile( + PREFETCH" 32%1 \n\t" + "movd %1, %%mm0 \n\t" + "punpckldq 3%1, %%mm0 \n\t" + "movd 6%1, %%mm1 \n\t" + "punpckldq 9%1, %%mm1 \n\t" + "movd 12%1, %%mm2 \n\t" + "punpckldq 15%1, %%mm2 \n\t" + "movd 18%1, %%mm3 \n\t" + "punpckldq 21%1, %%mm3 \n\t" + "pand %%mm7, %%mm0 \n\t" + "pand %%mm7, %%mm1 \n\t" + "pand %%mm7, %%mm2 \n\t" + "pand %%mm7, %%mm3 \n\t" + MOVNTQ" %%mm0, %0 \n\t" + MOVNTQ" %%mm1, 8%0 \n\t" + MOVNTQ" %%mm2, 16%0 \n\t" + MOVNTQ" %%mm3, 24%0" + :"=m"(*dest) + :"m"(*s) + :"memory"); + dest += 32; + s += 24; + } + __asm __volatile(SFENCE:::"memory"); + __asm __volatile(EMMS:::"memory"); + #endif + while (s < end) + { + #ifdef WORDS_BIGENDIAN + /* RGB24 (= R,G,B) -> RGB32 (= A,B,G,R) */ + *dest++ = 0; + *dest++ = s[2]; + *dest++ = s[1]; + *dest++ = s[0]; + s+=3; + #else + *dest++ = *s++; + *dest++ = *s++; + *dest++ = *s++; + *dest++ = 0; + #endif + } } static inline void RENAME(rgb32to24)(const uint8_t *src,uint8_t *dst,long src_size) { - uint8_t *dest = dst; - const uint8_t *s = src; - const uint8_t *end; + uint8_t *dest = dst; + const uint8_t *s = src; + const uint8_t *end; #ifdef HAVE_MMX - const uint8_t *mm_end; + const uint8_t *mm_end; #endif - end = s + src_size; + end = s + src_size; #ifdef HAVE_MMX - __asm __volatile(PREFETCH" %0"::"m"(*s):"memory"); - mm_end = end - 31; - while(s < mm_end) - { - __asm __volatile( - PREFETCH" 32%1\n\t" - "movq %1, %%mm0\n\t" - "movq 8%1, %%mm1\n\t" - "movq 16%1, %%mm4\n\t" - "movq 24%1, %%mm5\n\t" - "movq %%mm0, %%mm2\n\t" - "movq %%mm1, %%mm3\n\t" - "movq %%mm4, %%mm6\n\t" - "movq %%mm5, %%mm7\n\t" - "psrlq $8, %%mm2\n\t" - "psrlq $8, %%mm3\n\t" - "psrlq $8, %%mm6\n\t" - "psrlq $8, %%mm7\n\t" - "pand %2, %%mm0\n\t" - "pand %2, %%mm1\n\t" - "pand %2, %%mm4\n\t" - "pand %2, %%mm5\n\t" - "pand %3, %%mm2\n\t" - "pand %3, %%mm3\n\t" - "pand %3, %%mm6\n\t" - "pand %3, %%mm7\n\t" - "por %%mm2, %%mm0\n\t" - "por %%mm3, %%mm1\n\t" - "por %%mm6, %%mm4\n\t" - "por %%mm7, %%mm5\n\t" - - "movq %%mm1, %%mm2\n\t" - "movq %%mm4, %%mm3\n\t" - "psllq $48, %%mm2\n\t" - "psllq $32, %%mm3\n\t" - "pand %4, %%mm2\n\t" - "pand %5, %%mm3\n\t" - "por %%mm2, %%mm0\n\t" - "psrlq $16, %%mm1\n\t" - "psrlq $32, %%mm4\n\t" - "psllq $16, %%mm5\n\t" - "por %%mm3, %%mm1\n\t" - "pand %6, %%mm5\n\t" - "por %%mm5, %%mm4\n\t" - - MOVNTQ" %%mm0, %0\n\t" - MOVNTQ" %%mm1, 8%0\n\t" - MOVNTQ" %%mm4, 16%0" - :"=m"(*dest) - :"m"(*s),"m"(mask24l), - "m"(mask24h),"m"(mask24hh),"m"(mask24hhh),"m"(mask24hhhh) - :"memory"); - dest += 24; - s += 32; - } - __asm __volatile(SFENCE:::"memory"); - __asm __volatile(EMMS:::"memory"); -#endif - while(s < end) - { + __asm __volatile(PREFETCH" %0"::"m"(*s):"memory"); + mm_end = end - 31; + while (s < mm_end) + { + __asm __volatile( + PREFETCH" 32%1 \n\t" + "movq %1, %%mm0 \n\t" + "movq 8%1, %%mm1 \n\t" + "movq 16%1, %%mm4 \n\t" + "movq 24%1, %%mm5 \n\t" + "movq %%mm0, %%mm2 \n\t" + "movq %%mm1, %%mm3 \n\t" + "movq %%mm4, %%mm6 \n\t" + "movq %%mm5, %%mm7 \n\t" + "psrlq $8, %%mm2 \n\t" + "psrlq $8, %%mm3 \n\t" + "psrlq $8, %%mm6 \n\t" + "psrlq $8, %%mm7 \n\t" + "pand %2, %%mm0 \n\t" + "pand %2, %%mm1 \n\t" + "pand %2, %%mm4 \n\t" + "pand %2, %%mm5 \n\t" + "pand %3, %%mm2 \n\t" + "pand %3, %%mm3 \n\t" + "pand %3, %%mm6 \n\t" + "pand %3, %%mm7 \n\t" + "por %%mm2, %%mm0 \n\t" + "por %%mm3, %%mm1 \n\t" + "por %%mm6, %%mm4 \n\t" + "por %%mm7, %%mm5 \n\t" + + "movq %%mm1, %%mm2 \n\t" + "movq %%mm4, %%mm3 \n\t" + "psllq $48, %%mm2 \n\t" + "psllq $32, %%mm3 \n\t" + "pand %4, %%mm2 \n\t" + "pand %5, %%mm3 \n\t" + "por %%mm2, %%mm0 \n\t" + "psrlq $16, %%mm1 \n\t" + "psrlq $32, %%mm4 \n\t" + "psllq $16, %%mm5 \n\t" + "por %%mm3, %%mm1 \n\t" + "pand %6, %%mm5 \n\t" + "por %%mm5, %%mm4 \n\t" + + MOVNTQ" %%mm0, %0 \n\t" + MOVNTQ" %%mm1, 8%0 \n\t" + MOVNTQ" %%mm4, 16%0" + :"=m"(*dest) + :"m"(*s),"m"(mask24l), + "m"(mask24h),"m"(mask24hh),"m"(mask24hhh),"m"(mask24hhhh) + :"memory"); + dest += 24; + s += 32; + } + __asm __volatile(SFENCE:::"memory"); + __asm __volatile(EMMS:::"memory"); +#endif + while (s < end) + { #ifdef WORDS_BIGENDIAN - /* RGB32 (= A,B,G,R) -> RGB24 (= R,G,B) */ - s++; - dest[2] = *s++; - dest[1] = *s++; - dest[0] = *s++; - dest += 3; + /* RGB32 (= A,B,G,R) -> RGB24 (= R,G,B) */ + s++; + dest[2] = *s++; + dest[1] = *s++; + dest[0] = *s++; + dest += 3; #else - *dest++ = *s++; - *dest++ = *s++; - *dest++ = *s++; - s++; + *dest++ = *s++; + *dest++ = *s++; + *dest++ = *s++; + s++; #endif - } + } } /* @@ -237,677 +236,677 @@ static inline void RENAME(rgb32to24)(const uint8_t *src,uint8_t *dst,long src_si */ static inline void RENAME(rgb15to16)(const uint8_t *src,uint8_t *dst,long src_size) { - register const uint8_t* s=src; - register uint8_t* d=dst; - register const uint8_t *end; - const uint8_t *mm_end; - end = s + src_size; + register const uint8_t* s=src; + register uint8_t* d=dst; + register const uint8_t *end; + const uint8_t *mm_end; + end = s + src_size; #ifdef HAVE_MMX - __asm __volatile(PREFETCH" %0"::"m"(*s)); - __asm __volatile("movq %0, %%mm4"::"m"(mask15s)); - mm_end = end - 15; - while(s>1)&0x7FE07FE0) | (x&0x001F001F); - s+=4; - d+=4; + register uint32_t x= *((uint32_t *)s); + *((uint32_t *)d) = ((x>>1)&0x7FE07FE0) | (x&0x001F001F); + s+=4; + d+=4; } - if(s < end) + if (s < end) { - register uint16_t x= *((uint16_t *)s); - *((uint16_t *)d) = ((x>>1)&0x7FE0) | (x&0x001F); - s+=2; - d+=2; + register uint16_t x= *((uint16_t *)s); + *((uint16_t *)d) = ((x>>1)&0x7FE0) | (x&0x001F); + s+=2; + d+=2; } } static inline void RENAME(rgb32to16)(const uint8_t *src, uint8_t *dst, long src_size) { - const uint8_t *s = src; - const uint8_t *end; + const uint8_t *s = src; + const uint8_t *end; #ifdef HAVE_MMX - const uint8_t *mm_end; + const uint8_t *mm_end; #endif - uint16_t *d = (uint16_t *)dst; - end = s + src_size; + uint16_t *d = (uint16_t *)dst; + end = s + src_size; #ifdef HAVE_MMX - mm_end = end - 15; + mm_end = end - 15; #if 1 //is faster only if multiplies are reasonable fast (FIXME figure out on which cpus this is faster, on Athlon its slightly faster) - asm volatile( - "movq %3, %%mm5 \n\t" - "movq %4, %%mm6 \n\t" - "movq %5, %%mm7 \n\t" - "jmp 2f \n\t" - ASMALIGN(4) - "1: \n\t" - PREFETCH" 32(%1) \n\t" - "movd (%1), %%mm0 \n\t" - "movd 4(%1), %%mm3 \n\t" - "punpckldq 8(%1), %%mm0 \n\t" - "punpckldq 12(%1), %%mm3 \n\t" - "movq %%mm0, %%mm1 \n\t" - "movq %%mm3, %%mm4 \n\t" - "pand %%mm6, %%mm0 \n\t" - "pand %%mm6, %%mm3 \n\t" - "pmaddwd %%mm7, %%mm0 \n\t" - "pmaddwd %%mm7, %%mm3 \n\t" - "pand %%mm5, %%mm1 \n\t" - "pand %%mm5, %%mm4 \n\t" - "por %%mm1, %%mm0 \n\t" - "por %%mm4, %%mm3 \n\t" - "psrld $5, %%mm0 \n\t" - "pslld $11, %%mm3 \n\t" - "por %%mm3, %%mm0 \n\t" - MOVNTQ" %%mm0, (%0) \n\t" - "add $16, %1 \n\t" - "add $8, %0 \n\t" - "2: \n\t" - "cmp %2, %1 \n\t" - " jb 1b \n\t" - : "+r" (d), "+r"(s) - : "r" (mm_end), "m" (mask3216g), "m" (mask3216br), "m" (mul3216) - ); + asm volatile( + "movq %3, %%mm5 \n\t" + "movq %4, %%mm6 \n\t" + "movq %5, %%mm7 \n\t" + "jmp 2f \n\t" + ASMALIGN(4) + "1: \n\t" + PREFETCH" 32(%1) \n\t" + "movd (%1), %%mm0 \n\t" + "movd 4(%1), %%mm3 \n\t" + "punpckldq 8(%1), %%mm0 \n\t" + "punpckldq 12(%1), %%mm3 \n\t" + "movq %%mm0, %%mm1 \n\t" + "movq %%mm3, %%mm4 \n\t" + "pand %%mm6, %%mm0 \n\t" + "pand %%mm6, %%mm3 \n\t" + "pmaddwd %%mm7, %%mm0 \n\t" + "pmaddwd %%mm7, %%mm3 \n\t" + "pand %%mm5, %%mm1 \n\t" + "pand %%mm5, %%mm4 \n\t" + "por %%mm1, %%mm0 \n\t" + "por %%mm4, %%mm3 \n\t" + "psrld $5, %%mm0 \n\t" + "pslld $11, %%mm3 \n\t" + "por %%mm3, %%mm0 \n\t" + MOVNTQ" %%mm0, (%0) \n\t" + "add $16, %1 \n\t" + "add $8, %0 \n\t" + "2: \n\t" + "cmp %2, %1 \n\t" + " jb 1b \n\t" + : "+r" (d), "+r"(s) + : "r" (mm_end), "m" (mask3216g), "m" (mask3216br), "m" (mul3216) + ); #else - __asm __volatile(PREFETCH" %0"::"m"(*src):"memory"); - __asm __volatile( - "movq %0, %%mm7\n\t" - "movq %1, %%mm6\n\t" - ::"m"(red_16mask),"m"(green_16mask)); - while(s < mm_end) - { - __asm __volatile( - PREFETCH" 32%1\n\t" - "movd %1, %%mm0\n\t" - "movd 4%1, %%mm3\n\t" - "punpckldq 8%1, %%mm0\n\t" - "punpckldq 12%1, %%mm3\n\t" - "movq %%mm0, %%mm1\n\t" - "movq %%mm0, %%mm2\n\t" - "movq %%mm3, %%mm4\n\t" - "movq %%mm3, %%mm5\n\t" - "psrlq $3, %%mm0\n\t" - "psrlq $3, %%mm3\n\t" - "pand %2, %%mm0\n\t" - "pand %2, %%mm3\n\t" - "psrlq $5, %%mm1\n\t" - "psrlq $5, %%mm4\n\t" - "pand %%mm6, %%mm1\n\t" - "pand %%mm6, %%mm4\n\t" - "psrlq $8, %%mm2\n\t" - "psrlq $8, %%mm5\n\t" - "pand %%mm7, %%mm2\n\t" - "pand %%mm7, %%mm5\n\t" - "por %%mm1, %%mm0\n\t" - "por %%mm4, %%mm3\n\t" - "por %%mm2, %%mm0\n\t" - "por %%mm5, %%mm3\n\t" - "psllq $16, %%mm3\n\t" - "por %%mm3, %%mm0\n\t" - MOVNTQ" %%mm0, %0\n\t" - :"=m"(*d):"m"(*s),"m"(blue_16mask):"memory"); - d += 4; - s += 16; - } -#endif - __asm __volatile(SFENCE:::"memory"); - __asm __volatile(EMMS:::"memory"); -#endif - while(s < end) - { - register int rgb = *(uint32_t*)s; s += 4; - *d++ = ((rgb&0xFF)>>3) + ((rgb&0xFC00)>>5) + ((rgb&0xF80000)>>8); - } + __asm __volatile(PREFETCH" %0"::"m"(*src):"memory"); + __asm __volatile( + "movq %0, %%mm7 \n\t" + "movq %1, %%mm6 \n\t" + ::"m"(red_16mask),"m"(green_16mask)); + while (s < mm_end) + { + __asm __volatile( + PREFETCH" 32%1 \n\t" + "movd %1, %%mm0 \n\t" + "movd 4%1, %%mm3 \n\t" + "punpckldq 8%1, %%mm0 \n\t" + "punpckldq 12%1, %%mm3 \n\t" + "movq %%mm0, %%mm1 \n\t" + "movq %%mm0, %%mm2 \n\t" + "movq %%mm3, %%mm4 \n\t" + "movq %%mm3, %%mm5 \n\t" + "psrlq $3, %%mm0 \n\t" + "psrlq $3, %%mm3 \n\t" + "pand %2, %%mm0 \n\t" + "pand %2, %%mm3 \n\t" + "psrlq $5, %%mm1 \n\t" + "psrlq $5, %%mm4 \n\t" + "pand %%mm6, %%mm1 \n\t" + "pand %%mm6, %%mm4 \n\t" + "psrlq $8, %%mm2 \n\t" + "psrlq $8, %%mm5 \n\t" + "pand %%mm7, %%mm2 \n\t" + "pand %%mm7, %%mm5 \n\t" + "por %%mm1, %%mm0 \n\t" + "por %%mm4, %%mm3 \n\t" + "por %%mm2, %%mm0 \n\t" + "por %%mm5, %%mm3 \n\t" + "psllq $16, %%mm3 \n\t" + "por %%mm3, %%mm0 \n\t" + MOVNTQ" %%mm0, %0 \n\t" + :"=m"(*d):"m"(*s),"m"(blue_16mask):"memory"); + d += 4; + s += 16; + } +#endif + __asm __volatile(SFENCE:::"memory"); + __asm __volatile(EMMS:::"memory"); +#endif + while (s < end) + { + register int rgb = *(uint32_t*)s; s += 4; + *d++ = ((rgb&0xFF)>>3) + ((rgb&0xFC00)>>5) + ((rgb&0xF80000)>>8); + } } static inline void RENAME(rgb32tobgr16)(const uint8_t *src, uint8_t *dst, long src_size) { - const uint8_t *s = src; - const uint8_t *end; + const uint8_t *s = src; + const uint8_t *end; #ifdef HAVE_MMX - const uint8_t *mm_end; + const uint8_t *mm_end; #endif - uint16_t *d = (uint16_t *)dst; - end = s + src_size; + uint16_t *d = (uint16_t *)dst; + end = s + src_size; #ifdef HAVE_MMX - __asm __volatile(PREFETCH" %0"::"m"(*src):"memory"); - __asm __volatile( - "movq %0, %%mm7\n\t" - "movq %1, %%mm6\n\t" - ::"m"(red_16mask),"m"(green_16mask)); - mm_end = end - 15; - while(s < mm_end) - { - __asm __volatile( - PREFETCH" 32%1\n\t" - "movd %1, %%mm0\n\t" - "movd 4%1, %%mm3\n\t" - "punpckldq 8%1, %%mm0\n\t" - "punpckldq 12%1, %%mm3\n\t" - "movq %%mm0, %%mm1\n\t" - "movq %%mm0, %%mm2\n\t" - "movq %%mm3, %%mm4\n\t" - "movq %%mm3, %%mm5\n\t" - "psllq $8, %%mm0\n\t" - "psllq $8, %%mm3\n\t" - "pand %%mm7, %%mm0\n\t" - "pand %%mm7, %%mm3\n\t" - "psrlq $5, %%mm1\n\t" - "psrlq $5, %%mm4\n\t" - "pand %%mm6, %%mm1\n\t" - "pand %%mm6, %%mm4\n\t" - "psrlq $19, %%mm2\n\t" - "psrlq $19, %%mm5\n\t" - "pand %2, %%mm2\n\t" - "pand %2, %%mm5\n\t" - "por %%mm1, %%mm0\n\t" - "por %%mm4, %%mm3\n\t" - "por %%mm2, %%mm0\n\t" - "por %%mm5, %%mm3\n\t" - "psllq $16, %%mm3\n\t" - "por %%mm3, %%mm0\n\t" - MOVNTQ" %%mm0, %0\n\t" - :"=m"(*d):"m"(*s),"m"(blue_16mask):"memory"); - d += 4; - s += 16; - } - __asm __volatile(SFENCE:::"memory"); - __asm __volatile(EMMS:::"memory"); -#endif - while(s < end) - { - register int rgb = *(uint32_t*)s; s += 4; - *d++ = ((rgb&0xF8)<<8) + ((rgb&0xFC00)>>5) + ((rgb&0xF80000)>>19); - } + __asm __volatile(PREFETCH" %0"::"m"(*src):"memory"); + __asm __volatile( + "movq %0, %%mm7 \n\t" + "movq %1, %%mm6 \n\t" + ::"m"(red_16mask),"m"(green_16mask)); + mm_end = end - 15; + while (s < mm_end) + { + __asm __volatile( + PREFETCH" 32%1 \n\t" + "movd %1, %%mm0 \n\t" + "movd 4%1, %%mm3 \n\t" + "punpckldq 8%1, %%mm0 \n\t" + "punpckldq 12%1, %%mm3 \n\t" + "movq %%mm0, %%mm1 \n\t" + "movq %%mm0, %%mm2 \n\t" + "movq %%mm3, %%mm4 \n\t" + "movq %%mm3, %%mm5 \n\t" + "psllq $8, %%mm0 \n\t" + "psllq $8, %%mm3 \n\t" + "pand %%mm7, %%mm0 \n\t" + "pand %%mm7, %%mm3 \n\t" + "psrlq $5, %%mm1 \n\t" + "psrlq $5, %%mm4 \n\t" + "pand %%mm6, %%mm1 \n\t" + "pand %%mm6, %%mm4 \n\t" + "psrlq $19, %%mm2 \n\t" + "psrlq $19, %%mm5 \n\t" + "pand %2, %%mm2 \n\t" + "pand %2, %%mm5 \n\t" + "por %%mm1, %%mm0 \n\t" + "por %%mm4, %%mm3 \n\t" + "por %%mm2, %%mm0 \n\t" + "por %%mm5, %%mm3 \n\t" + "psllq $16, %%mm3 \n\t" + "por %%mm3, %%mm0 \n\t" + MOVNTQ" %%mm0, %0 \n\t" + :"=m"(*d):"m"(*s),"m"(blue_16mask):"memory"); + d += 4; + s += 16; + } + __asm __volatile(SFENCE:::"memory"); + __asm __volatile(EMMS:::"memory"); +#endif + while (s < end) + { + register int rgb = *(uint32_t*)s; s += 4; + *d++ = ((rgb&0xF8)<<8) + ((rgb&0xFC00)>>5) + ((rgb&0xF80000)>>19); + } } static inline void RENAME(rgb32to15)(const uint8_t *src, uint8_t *dst, long src_size) { - const uint8_t *s = src; - const uint8_t *end; + const uint8_t *s = src; + const uint8_t *end; #ifdef HAVE_MMX - const uint8_t *mm_end; + const uint8_t *mm_end; #endif - uint16_t *d = (uint16_t *)dst; - end = s + src_size; + uint16_t *d = (uint16_t *)dst; + end = s + src_size; #ifdef HAVE_MMX - mm_end = end - 15; + mm_end = end - 15; #if 1 //is faster only if multiplies are reasonable fast (FIXME figure out on which cpus this is faster, on Athlon its slightly faster) - asm volatile( - "movq %3, %%mm5 \n\t" - "movq %4, %%mm6 \n\t" - "movq %5, %%mm7 \n\t" - "jmp 2f \n\t" - ASMALIGN(4) - "1: \n\t" - PREFETCH" 32(%1) \n\t" - "movd (%1), %%mm0 \n\t" - "movd 4(%1), %%mm3 \n\t" - "punpckldq 8(%1), %%mm0 \n\t" - "punpckldq 12(%1), %%mm3 \n\t" - "movq %%mm0, %%mm1 \n\t" - "movq %%mm3, %%mm4 \n\t" - "pand %%mm6, %%mm0 \n\t" - "pand %%mm6, %%mm3 \n\t" - "pmaddwd %%mm7, %%mm0 \n\t" - "pmaddwd %%mm7, %%mm3 \n\t" - "pand %%mm5, %%mm1 \n\t" - "pand %%mm5, %%mm4 \n\t" - "por %%mm1, %%mm0 \n\t" - "por %%mm4, %%mm3 \n\t" - "psrld $6, %%mm0 \n\t" - "pslld $10, %%mm3 \n\t" - "por %%mm3, %%mm0 \n\t" - MOVNTQ" %%mm0, (%0) \n\t" - "add $16, %1 \n\t" - "add $8, %0 \n\t" - "2: \n\t" - "cmp %2, %1 \n\t" - " jb 1b \n\t" - : "+r" (d), "+r"(s) - : "r" (mm_end), "m" (mask3215g), "m" (mask3216br), "m" (mul3215) - ); + asm volatile( + "movq %3, %%mm5 \n\t" + "movq %4, %%mm6 \n\t" + "movq %5, %%mm7 \n\t" + "jmp 2f \n\t" + ASMALIGN(4) + "1: \n\t" + PREFETCH" 32(%1) \n\t" + "movd (%1), %%mm0 \n\t" + "movd 4(%1), %%mm3 \n\t" + "punpckldq 8(%1), %%mm0 \n\t" + "punpckldq 12(%1), %%mm3 \n\t" + "movq %%mm0, %%mm1 \n\t" + "movq %%mm3, %%mm4 \n\t" + "pand %%mm6, %%mm0 \n\t" + "pand %%mm6, %%mm3 \n\t" + "pmaddwd %%mm7, %%mm0 \n\t" + "pmaddwd %%mm7, %%mm3 \n\t" + "pand %%mm5, %%mm1 \n\t" + "pand %%mm5, %%mm4 \n\t" + "por %%mm1, %%mm0 \n\t" + "por %%mm4, %%mm3 \n\t" + "psrld $6, %%mm0 \n\t" + "pslld $10, %%mm3 \n\t" + "por %%mm3, %%mm0 \n\t" + MOVNTQ" %%mm0, (%0) \n\t" + "add $16, %1 \n\t" + "add $8, %0 \n\t" + "2: \n\t" + "cmp %2, %1 \n\t" + " jb 1b \n\t" + : "+r" (d), "+r"(s) + : "r" (mm_end), "m" (mask3215g), "m" (mask3216br), "m" (mul3215) + ); #else - __asm __volatile(PREFETCH" %0"::"m"(*src):"memory"); - __asm __volatile( - "movq %0, %%mm7\n\t" - "movq %1, %%mm6\n\t" - ::"m"(red_15mask),"m"(green_15mask)); - while(s < mm_end) - { - __asm __volatile( - PREFETCH" 32%1\n\t" - "movd %1, %%mm0\n\t" - "movd 4%1, %%mm3\n\t" - "punpckldq 8%1, %%mm0\n\t" - "punpckldq 12%1, %%mm3\n\t" - "movq %%mm0, %%mm1\n\t" - "movq %%mm0, %%mm2\n\t" - "movq %%mm3, %%mm4\n\t" - "movq %%mm3, %%mm5\n\t" - "psrlq $3, %%mm0\n\t" - "psrlq $3, %%mm3\n\t" - "pand %2, %%mm0\n\t" - "pand %2, %%mm3\n\t" - "psrlq $6, %%mm1\n\t" - "psrlq $6, %%mm4\n\t" - "pand %%mm6, %%mm1\n\t" - "pand %%mm6, %%mm4\n\t" - "psrlq $9, %%mm2\n\t" - "psrlq $9, %%mm5\n\t" - "pand %%mm7, %%mm2\n\t" - "pand %%mm7, %%mm5\n\t" - "por %%mm1, %%mm0\n\t" - "por %%mm4, %%mm3\n\t" - "por %%mm2, %%mm0\n\t" - "por %%mm5, %%mm3\n\t" - "psllq $16, %%mm3\n\t" - "por %%mm3, %%mm0\n\t" - MOVNTQ" %%mm0, %0\n\t" - :"=m"(*d):"m"(*s),"m"(blue_15mask):"memory"); - d += 4; - s += 16; - } -#endif - __asm __volatile(SFENCE:::"memory"); - __asm __volatile(EMMS:::"memory"); -#endif - while(s < end) - { - register int rgb = *(uint32_t*)s; s += 4; - *d++ = ((rgb&0xFF)>>3) + ((rgb&0xF800)>>6) + ((rgb&0xF80000)>>9); - } + __asm __volatile(PREFETCH" %0"::"m"(*src):"memory"); + __asm __volatile( + "movq %0, %%mm7 \n\t" + "movq %1, %%mm6 \n\t" + ::"m"(red_15mask),"m"(green_15mask)); + while (s < mm_end) + { + __asm __volatile( + PREFETCH" 32%1 \n\t" + "movd %1, %%mm0 \n\t" + "movd 4%1, %%mm3 \n\t" + "punpckldq 8%1, %%mm0 \n\t" + "punpckldq 12%1, %%mm3 \n\t" + "movq %%mm0, %%mm1 \n\t" + "movq %%mm0, %%mm2 \n\t" + "movq %%mm3, %%mm4 \n\t" + "movq %%mm3, %%mm5 \n\t" + "psrlq $3, %%mm0 \n\t" + "psrlq $3, %%mm3 \n\t" + "pand %2, %%mm0 \n\t" + "pand %2, %%mm3 \n\t" + "psrlq $6, %%mm1 \n\t" + "psrlq $6, %%mm4 \n\t" + "pand %%mm6, %%mm1 \n\t" + "pand %%mm6, %%mm4 \n\t" + "psrlq $9, %%mm2 \n\t" + "psrlq $9, %%mm5 \n\t" + "pand %%mm7, %%mm2 \n\t" + "pand %%mm7, %%mm5 \n\t" + "por %%mm1, %%mm0 \n\t" + "por %%mm4, %%mm3 \n\t" + "por %%mm2, %%mm0 \n\t" + "por %%mm5, %%mm3 \n\t" + "psllq $16, %%mm3 \n\t" + "por %%mm3, %%mm0 \n\t" + MOVNTQ" %%mm0, %0 \n\t" + :"=m"(*d):"m"(*s),"m"(blue_15mask):"memory"); + d += 4; + s += 16; + } +#endif + __asm __volatile(SFENCE:::"memory"); + __asm __volatile(EMMS:::"memory"); +#endif + while (s < end) + { + register int rgb = *(uint32_t*)s; s += 4; + *d++ = ((rgb&0xFF)>>3) + ((rgb&0xF800)>>6) + ((rgb&0xF80000)>>9); + } } static inline void RENAME(rgb32tobgr15)(const uint8_t *src, uint8_t *dst, long src_size) { - const uint8_t *s = src; - const uint8_t *end; + const uint8_t *s = src; + const uint8_t *end; #ifdef HAVE_MMX - const uint8_t *mm_end; + const uint8_t *mm_end; #endif - uint16_t *d = (uint16_t *)dst; - end = s + src_size; + uint16_t *d = (uint16_t *)dst; + end = s + src_size; #ifdef HAVE_MMX - __asm __volatile(