diff --git a/libswscale/rgb2rgb.c b/libswscale/rgb2rgb.c index 00e785f0c8..4bba657bd1 100644 --- a/libswscale/rgb2rgb.c +++ b/libswscale/rgb2rgb.c @@ -78,10 +78,6 @@ void (*yuv422ptouyvy)(const uint8_t *ysrc, const uint8_t *usrc, const uint8_t *vsrc, uint8_t *dst, int width, int height, int lumStride, int chromStride, int dstStride); -void (*yuy2toyv12)(const uint8_t *src, uint8_t *ydst, - uint8_t *udst, uint8_t *vdst, - int width, int height, - int lumStride, int chromStride, int srcStride); void (*ff_rgb24toyv12)(const uint8_t *src, uint8_t *ydst, uint8_t *udst, uint8_t *vdst, int width, int height, @@ -95,16 +91,6 @@ void (*interleaveBytes)(const uint8_t *src1, const uint8_t *src2, uint8_t *dst, void (*deinterleaveBytes)(const uint8_t *src, uint8_t *dst1, uint8_t *dst2, int width, int height, int srcStride, int dst1Stride, int dst2Stride); -void (*vu9_to_vu12)(const uint8_t *src1, const uint8_t *src2, - uint8_t *dst1, uint8_t *dst2, - int width, int height, - int srcStride1, int srcStride2, - int dstStride1, int dstStride2); -void (*yvu9_to_yuy2)(const uint8_t *src1, const uint8_t *src2, - const uint8_t *src3, uint8_t *dst, - int width, int height, - int srcStride1, int srcStride2, - int srcStride3, int dstStride); void (*uyvytoyuv420)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src, int width, int height, int lumStride, int chromStride, int srcStride); diff --git a/libswscale/rgb2rgb.h b/libswscale/rgb2rgb.h index 0d0f7c3a28..ed6e08a565 100644 --- a/libswscale/rgb2rgb.h +++ b/libswscale/rgb2rgb.h @@ -108,14 +108,6 @@ extern void (*yuv422ptoyuy2)(const uint8_t *ysrc, const uint8_t *usrc, const uin int width, int height, int lumStride, int chromStride, int dstStride); -/** - * Height should be a multiple of 2 and width should be a multiple of 16. - * (If this is a problem for anyone then tell me, and I will fix it.) - */ -extern void (*yuy2toyv12)(const uint8_t *src, uint8_t *ydst, uint8_t *udst, uint8_t *vdst, - int width, int height, - int lumStride, int chromStride, int srcStride); - /** * Height should be a multiple of 2 and width should be a multiple of 16. * (If this is a problem for anyone then tell me, and I will fix it.) @@ -150,18 +142,6 @@ extern void (*deinterleaveBytes)(const uint8_t *src, uint8_t *dst1, uint8_t *dst int width, int height, int srcStride, int dst1Stride, int dst2Stride); -extern void (*vu9_to_vu12)(const uint8_t *src1, const uint8_t *src2, - uint8_t *dst1, uint8_t *dst2, - int width, int height, - int srcStride1, int srcStride2, - int dstStride1, int dstStride2); - -extern void (*yvu9_to_yuy2)(const uint8_t *src1, const uint8_t *src2, const uint8_t *src3, - uint8_t *dst, - int width, int height, - int srcStride1, int srcStride2, - int srcStride3, int dstStride); - extern void (*uyvytoyuv420)(uint8_t *ydst, uint8_t *udst, uint8_t *vdst, const uint8_t *src, int width, int height, int lumStride, int chromStride, int srcStride); diff --git a/libswscale/rgb2rgb_template.c b/libswscale/rgb2rgb_template.c index f71be33ac4..48959b6b5f 100644 --- a/libswscale/rgb2rgb_template.c +++ b/libswscale/rgb2rgb_template.c @@ -528,40 +528,6 @@ static inline void yuv422ptoyuy2_c(const uint8_t *ysrc, const uint8_t *usrc, chromStride, dstStride, 1); } -/** - * Height should be a multiple of 2 and width should be a multiple of 16. - * (If this is a problem for anyone then tell me, and I will fix it.) - */ -static inline void yuy2toyv12_c(const uint8_t *src, uint8_t *ydst, - uint8_t *udst, uint8_t *vdst, - int width, int height, int lumStride, - int chromStride, int srcStride) -{ - int y; - const int chromWidth = width >> 1; - - for (y = 0; y < height; y += 2) { - int i; - for (i = 0; i < chromWidth; i++) { - ydst[2 * i + 0] = src[4 * i + 0]; - udst[i] = src[4 * i + 1]; - ydst[2 * i + 1] = src[4 * i + 2]; - vdst[i] = src[4 * i + 3]; - } - ydst += lumStride; - src += srcStride; - - for (i = 0; i < chromWidth; i++) { - ydst[2 * i + 0] = src[4 * i + 0]; - ydst[2 * i + 1] = src[4 * i + 2]; - } - udst += chromStride; - vdst += chromStride; - ydst += lumStride; - src += srcStride; - } -} - static inline void planar2x_c(const uint8_t *src, uint8_t *dst, int srcWidth, int srcHeight, int srcStride, int dstStride) { @@ -710,59 +676,6 @@ static void deinterleaveBytes_c(const uint8_t *src, uint8_t *dst1, uint8_t *dst2 } } -static inline void vu9_to_vu12_c(const uint8_t *src1, const uint8_t *src2, - uint8_t *dst1, uint8_t *dst2, - int width, int height, - int srcStride1, int srcStride2, - int dstStride1, int dstStride2) -{ - int x, y; - int w = width / 2; - int h = height / 2; - - for (y = 0; y < h; y++) { - const uint8_t *s1 = src1 + srcStride1 * (y >> 1); - uint8_t *d = dst1 + dstStride1 * y; - for (x = 0; x < w; x++) - d[2 * x] = d[2 * x + 1] = s1[x]; - } - for (y = 0; y < h; y++) { - const uint8_t *s2 = src2 + srcStride2 * (y >> 1); - uint8_t *d = dst2 + dstStride2 * y; - for (x = 0; x < w; x++) - d[2 * x] = d[2 * x + 1] = s2[x]; - } -} - -static inline void yvu9_to_yuy2_c(const uint8_t *src1, const uint8_t *src2, - const uint8_t *src3, uint8_t *dst, - int width, int height, - int srcStride1, int srcStride2, - int srcStride3, int dstStride) -{ - int x, y; - int w = width / 2; - int h = height; - - for (y = 0; y < h; y++) { - const uint8_t *yp = src1 + srcStride1 * y; - const uint8_t *up = src2 + srcStride2 * (y >> 2); - const uint8_t *vp = src3 + srcStride3 * (y >> 2); - uint8_t *d = dst + dstStride * y; - for (x = 0; x < w; x++) { - const int x2 = x << 2; - d[8 * x + 0] = yp[x2]; - d[8 * x + 1] = up[x]; - d[8 * x + 2] = yp[x2 + 1]; - d[8 * x + 3] = vp[x]; - d[8 * x + 4] = yp[x2 + 2]; - d[8 * x + 5] = up[x]; - d[8 * x + 6] = yp[x2 + 3]; - d[8 * x + 7] = vp[x]; - } - } -} - static void extract_even_c(const uint8_t *src, uint8_t *dst, int count) { dst += count; @@ -948,13 +861,10 @@ static av_cold void rgb2rgb_init_c(void) yv12touyvy = yv12touyvy_c; yuv422ptoyuy2 = yuv422ptoyuy2_c; yuv422ptouyvy = yuv422ptouyvy_c; - yuy2toyv12 = yuy2toyv12_c; planar2x = planar2x_c; ff_rgb24toyv12 = ff_rgb24toyv12_c; interleaveBytes = interleaveBytes_c; deinterleaveBytes = deinterleaveBytes_c; - vu9_to_vu12 = vu9_to_vu12_c; - yvu9_to_yuy2 = yvu9_to_yuy2_c; uyvytoyuv420 = uyvytoyuv420_c; uyvytoyuv422 = uyvytoyuv422_c; diff --git a/libswscale/x86/rgb2rgb.c b/libswscale/x86/rgb2rgb.c index 48a73bd0b2..f2b434f611 100644 --- a/libswscale/x86/rgb2rgb.c +++ b/libswscale/x86/rgb2rgb.c @@ -1272,107 +1272,6 @@ static inline void yuv422ptoyuy2_mmxext(const uint8_t *ysrc, const uint8_t *usrc yuvPlanartoyuy2_mmxext(ysrc, usrc, vsrc, dst, width, height, lumStride, chromStride, dstStride, 1); } -/** - * Height should be a multiple of 2 and width should be a multiple of 16. - * (If this is a problem for anyone then tell me, and I will fix it.) - */ -static inline void yuy2toyv12_mmxext(const uint8_t *src, uint8_t *ydst, uint8_t *udst, uint8_t *vdst, - int width, int height, - int lumStride, int chromStride, int srcStride) -{ - const x86_reg chromWidth= width>>1; - for (int y = 0; y < height; y += 2) { - __asm__ volatile( - "xor %%"FF_REG_a", %%"FF_REG_a"\n\t" - "pcmpeqw %%mm7, %%mm7 \n\t" - "psrlw $8, %%mm7 \n\t" // FF,00,FF,00... - ".p2align 4 \n\t" - "1: \n\t" - PREFETCH" 64(%0, %%"FF_REG_a", 4) \n\t" - "movq (%0, %%"FF_REG_a", 4), %%mm0 \n\t" // YUYV YUYV(0) - "movq 8(%0, %%"FF_REG_a", 4), %%mm1 \n\t" // YUYV YUYV(4) - "movq %%mm0, %%mm2 \n\t" // YUYV YUYV(0) - "movq %%mm1, %%mm3 \n\t" // YUYV YUYV(4) - "psrlw $8, %%mm0 \n\t" // U0V0 U0V0(0) - "psrlw $8, %%mm1 \n\t" // U0V0 U0V0(4) - "pand %%mm7, %%mm2 \n\t" // Y0Y0 Y0Y0(0) - "pand %%mm7, %%mm3 \n\t" // Y0Y0 Y0Y0(4) - "packuswb %%mm1, %%mm0 \n\t" // UVUV UVUV(0) - "packuswb %%mm3, %%mm2 \n\t" // YYYY YYYY(0) - - MOVNTQ" %%mm2, (%1, %%"FF_REG_a", 2) \n\t" - - "movq 16(%0, %%"FF_REG_a", 4), %%mm1 \n\t" // YUYV YUYV(8) - "movq 24(%0, %%"FF_REG_a", 4), %%mm2 \n\t" // YUYV YUYV(12) - "movq %%mm1, %%mm3 \n\t" // YUYV YUYV(8) - "movq %%mm2, %%mm4 \n\t" // YUYV YUYV(12) - "psrlw $8, %%mm1 \n\t" // U0V0 U0V0(8) - "psrlw $8, %%mm2 \n\t" // U0V0 U0V0(12) - "pand %%mm7, %%mm3 \n\t" // Y0Y0 Y0Y0(8) - "pand %%mm7, %%mm4 \n\t" // Y0Y0 Y0Y0(12) - "packuswb %%mm2, %%mm1 \n\t" // UVUV UVUV(8) - "packuswb %%mm4, %%mm3 \n\t" // YYYY YYYY(8) - - MOVNTQ" %%mm3, 8(%1, %%"FF_REG_a", 2) \n\t" - - "movq %%mm0, %%mm2 \n\t" // UVUV UVUV(0) - "movq %%mm1, %%mm3 \n\t" // UVUV UVUV(8) - "psrlw $8, %%mm0 \n\t" // V0V0 V0V0(0) - "psrlw $8, %%mm1 \n\t" // V0V0 V0V0(8) - "pand %%mm7, %%mm2 \n\t" // U0U0 U0U0(0) - "pand %%mm7, %%mm3 \n\t" // U0U0 U0U0(8) - "packuswb %%mm1, %%mm0 \n\t" // VVVV VVVV(0) - "packuswb %%mm3, %%mm2 \n\t" // UUUU UUUU(0) - - MOVNTQ" %%mm0, (%3, %%"FF_REG_a") \n\t" - MOVNTQ" %%mm2, (%2, %%"FF_REG_a") \n\t" - - "add $8, %%"FF_REG_a" \n\t" - "cmp %4, %%"FF_REG_a" \n\t" - " jb 1b \n\t" - ::"r"(src), "r"(ydst), "r"(udst), "r"(vdst), "g" (chromWidth) - : "memory", "%"FF_REG_a - ); - - ydst += lumStride; - src += srcStride; - - __asm__ volatile( - "xor %%"FF_REG_a", %%"FF_REG_a"\n\t" - ".p2align 4 \n\t" - "1: \n\t" - PREFETCH" 64(%0, %%"FF_REG_a", 4) \n\t" - "movq (%0, %%"FF_REG_a", 4), %%mm0 \n\t" // YUYV YUYV(0) - "movq 8(%0, %%"FF_REG_a", 4), %%mm1 \n\t" // YUYV YUYV(4) - "movq 16(%0, %%"FF_REG_a", 4), %%mm2 \n\t" // YUYV YUYV(8) - "movq 24(%0, %%"FF_REG_a", 4), %%mm3 \n\t" // YUYV YUYV(12) - "pand %%mm7, %%mm0 \n\t" // Y0Y0 Y0Y0(0) - "pand %%mm7, %%mm1 \n\t" // Y0Y0 Y0Y0(4) - "pand %%mm7, %%mm2 \n\t" // Y0Y0 Y0Y0(8) - "pand %%mm7, %%mm3 \n\t" // Y0Y0 Y0Y0(12) - "packuswb %%mm1, %%mm0 \n\t" // YYYY YYYY(0) - "packuswb %%mm3, %%mm2 \n\t" // YYYY YYYY(8) - - MOVNTQ" %%mm0, (%1, %%"FF_REG_a", 2) \n\t" - MOVNTQ" %%mm2, 8(%1, %%"FF_REG_a", 2) \n\t" - - "add $8, %%"FF_REG_a"\n\t" - "cmp %4, %%"FF_REG_a"\n\t" - " jb 1b \n\t" - - ::"r"(src), "r"(ydst), "r"(udst), "r"(vdst), "g" (chromWidth) - : "memory", "%"FF_REG_a - ); - udst += chromStride; - vdst += chromStride; - ydst += lumStride; - src += srcStride; - } - __asm__ volatile(EMMS" \n\t" - SFENCE" \n\t" - :::"memory"); -} - static inline void planar2x_mmxext(const uint8_t *src, uint8_t *dst, int srcWidth, int srcHeight, int srcStride, int dstStride) { dst[0]= src[0]; @@ -1677,182 +1576,6 @@ static inline void rgb24toyv12_mmxext(const uint8_t *src, uint8_t *ydst, uint8_t } #endif /* HAVE_7REGS */ -static inline void vu9_to_vu12_mmxext(const uint8_t *src1, const uint8_t *src2, - uint8_t *dst1, uint8_t *dst2, - int width, int height, - int srcStride1, int srcStride2, - int dstStride1, int dstStride2) -{ - int w,h; - w=width/2; h=height/2; - __asm__ volatile( - PREFETCH" %0 \n\t" - PREFETCH" %1 \n\t" - ::"m"(*(src1+srcStride1)),"m"(*(src2+srcStride2)):"memory"); - for (x86_reg y = 0; y < h; y++) { - const uint8_t* s1=src1+srcStride1*(y>>1); - uint8_t* d=dst1+dstStride1*y; - x86_reg x = 0; - for (;x>1); - uint8_t* d=dst2+dstStride2*y; - x86_reg x = 0; - for (;x>2); - const uint8_t* vp=src3+srcStride3*(y>>2); - uint8_t* d=dst+dstStride*y; - x86_reg x = 0; - for (;x