|
|
|
@ -684,7 +684,7 @@ static void h264_loop_filter_strength_mmx2( int16_t bS[2][4][4], uint8_t nnz[40] |
|
|
|
|
"movq %%mm6, "#OF"(%1) \n\t" |
|
|
|
|
|
|
|
|
|
#define QPEL_H264(OPNAME, OP, MMX)\ |
|
|
|
|
static void OPNAME ## h264_qpel4_h_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
|
|
|
|
|
static av_noinline void OPNAME ## h264_qpel4_h_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
|
|
|
|
|
int h=4;\
|
|
|
|
|
\
|
|
|
|
|
asm volatile(\
|
|
|
|
@ -724,7 +724,7 @@ static void OPNAME ## h264_qpel4_h_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, i |
|
|
|
|
: "memory"\
|
|
|
|
|
);\
|
|
|
|
|
}\
|
|
|
|
|
static void OPNAME ## h264_qpel4_h_lowpass_l2_ ## MMX(uint8_t *dst, uint8_t *src, uint8_t *src2, int dstStride, int src2Stride){\
|
|
|
|
|
static av_noinline void OPNAME ## h264_qpel4_h_lowpass_l2_ ## MMX(uint8_t *dst, uint8_t *src, uint8_t *src2, int dstStride, int src2Stride){\
|
|
|
|
|
int h=4;\
|
|
|
|
|
asm volatile(\
|
|
|
|
|
"pxor %%mm7, %%mm7 \n\t"\
|
|
|
|
@ -768,7 +768,7 @@ static void OPNAME ## h264_qpel4_h_lowpass_l2_ ## MMX(uint8_t *dst, uint8_t *src |
|
|
|
|
);\
|
|
|
|
|
}while(--h);\
|
|
|
|
|
}\
|
|
|
|
|
static void OPNAME ## h264_qpel4_v_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
|
|
|
|
|
static av_noinline void OPNAME ## h264_qpel4_v_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
|
|
|
|
|
src -= 2*srcStride;\
|
|
|
|
|
asm volatile(\
|
|
|
|
|
"pxor %%mm7, %%mm7 \n\t"\
|
|
|
|
@ -797,7 +797,7 @@ static void OPNAME ## h264_qpel4_v_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, i |
|
|
|
|
: "memory"\
|
|
|
|
|
);\
|
|
|
|
|
}\
|
|
|
|
|
static void OPNAME ## h264_qpel4_hv_lowpass_ ## MMX(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride){\
|
|
|
|
|
static av_noinline void OPNAME ## h264_qpel4_hv_lowpass_ ## MMX(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride){\
|
|
|
|
|
int h=4;\
|
|
|
|
|
int w=3;\
|
|
|
|
|
src -= 2*srcStride+2;\
|
|
|
|
@ -861,7 +861,7 @@ static void OPNAME ## h264_qpel4_hv_lowpass_ ## MMX(uint8_t *dst, int16_t *tmp, |
|
|
|
|
);\
|
|
|
|
|
}\
|
|
|
|
|
\
|
|
|
|
|
static void OPNAME ## h264_qpel8_h_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
|
|
|
|
|
static av_noinline void OPNAME ## h264_qpel8_h_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
|
|
|
|
|
int h=8;\
|
|
|
|
|
asm volatile(\
|
|
|
|
|
"pxor %%mm7, %%mm7 \n\t"\
|
|
|
|
@ -918,7 +918,7 @@ static void OPNAME ## h264_qpel8_h_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, i |
|
|
|
|
);\
|
|
|
|
|
}\
|
|
|
|
|
\
|
|
|
|
|
static void OPNAME ## h264_qpel8_h_lowpass_l2_ ## MMX(uint8_t *dst, uint8_t *src, uint8_t *src2, int dstStride, int src2Stride){\
|
|
|
|
|
static av_noinline void OPNAME ## h264_qpel8_h_lowpass_l2_ ## MMX(uint8_t *dst, uint8_t *src, uint8_t *src2, int dstStride, int src2Stride){\
|
|
|
|
|
int h=8;\
|
|
|
|
|
asm volatile(\
|
|
|
|
|
"pxor %%mm7, %%mm7 \n\t"\
|
|
|
|
@ -981,7 +981,7 @@ static void OPNAME ## h264_qpel8_h_lowpass_l2_ ## MMX(uint8_t *dst, uint8_t *src |
|
|
|
|
}while(--h);\
|
|
|
|
|
}\
|
|
|
|
|
\
|
|
|
|
|
static inline void OPNAME ## h264_qpel8or16_v_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h){\
|
|
|
|
|
static av_noinline void OPNAME ## h264_qpel8or16_v_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h){\
|
|
|
|
|
int w= 2;\
|
|
|
|
|
src -= 2*srcStride;\
|
|
|
|
|
\
|
|
|
|
@ -1036,7 +1036,7 @@ static inline void OPNAME ## h264_qpel8or16_v_lowpass_ ## MMX(uint8_t *dst, uint |
|
|
|
|
dst += 4-h*dstStride;\
|
|
|
|
|
}\
|
|
|
|
|
}\
|
|
|
|
|
static inline void OPNAME ## h264_qpel8or16_hv_lowpass_ ## MMX(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride, int size){\
|
|
|
|
|
static av_noinline void OPNAME ## h264_qpel8or16_hv_lowpass_ ## MMX(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride, int size){\
|
|
|
|
|
int h = size;\
|
|
|
|
|
int w = (size+8)>>2;\
|
|
|
|
|
src -= 2*srcStride+2;\
|
|
|
|
@ -1141,12 +1141,12 @@ static inline void OPNAME ## h264_qpel8or16_hv_lowpass_ ## MMX(uint8_t *dst, int |
|
|
|
|
static void OPNAME ## h264_qpel8_v_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
|
|
|
|
|
OPNAME ## h264_qpel8or16_v_lowpass_ ## MMX(dst , src , dstStride, srcStride, 8);\
|
|
|
|
|
}\
|
|
|
|
|
static void OPNAME ## h264_qpel16_v_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
|
|
|
|
|
static av_noinline void OPNAME ## h264_qpel16_v_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
|
|
|
|
|
OPNAME ## h264_qpel8or16_v_lowpass_ ## MMX(dst , src , dstStride, srcStride, 16);\
|
|
|
|
|
OPNAME ## h264_qpel8or16_v_lowpass_ ## MMX(dst+8, src+8, dstStride, srcStride, 16);\
|
|
|
|
|
}\
|
|
|
|
|
\
|
|
|
|
|
static void OPNAME ## h264_qpel16_h_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
|
|
|
|
|
static av_noinline void OPNAME ## h264_qpel16_h_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
|
|
|
|
|
OPNAME ## h264_qpel8_h_lowpass_ ## MMX(dst , src , dstStride, srcStride);\
|
|
|
|
|
OPNAME ## h264_qpel8_h_lowpass_ ## MMX(dst+8, src+8, dstStride, srcStride);\
|
|
|
|
|
src += 8*srcStride;\
|
|
|
|
@ -1155,7 +1155,7 @@ static void OPNAME ## h264_qpel16_h_lowpass_ ## MMX(uint8_t *dst, uint8_t *src, |
|
|
|
|
OPNAME ## h264_qpel8_h_lowpass_ ## MMX(dst+8, src+8, dstStride, srcStride);\
|
|
|
|
|
}\
|
|
|
|
|
\
|
|
|
|
|
static void OPNAME ## h264_qpel16_h_lowpass_l2_ ## MMX(uint8_t *dst, uint8_t *src, uint8_t *src2, int dstStride, int src2Stride){\
|
|
|
|
|
static av_noinline void OPNAME ## h264_qpel16_h_lowpass_l2_ ## MMX(uint8_t *dst, uint8_t *src, uint8_t *src2, int dstStride, int src2Stride){\
|
|
|
|
|
OPNAME ## h264_qpel8_h_lowpass_l2_ ## MMX(dst , src , src2 , dstStride, src2Stride);\
|
|
|
|
|
OPNAME ## h264_qpel8_h_lowpass_l2_ ## MMX(dst+8, src+8, src2+8, dstStride, src2Stride);\
|
|
|
|
|
src += 8*dstStride;\
|
|
|
|
@ -1173,7 +1173,7 @@ static void OPNAME ## h264_qpel16_hv_lowpass_ ## MMX(uint8_t *dst, int16_t *tmp, |
|
|
|
|
OPNAME ## h264_qpel8or16_hv_lowpass_ ## MMX(dst , tmp , src , dstStride, tmpStride, srcStride, 16);\
|
|
|
|
|
}\
|
|
|
|
|
\
|
|
|
|
|
static void OPNAME ## pixels4_l2_shift5_ ## MMX(uint8_t *dst, int16_t *src16, uint8_t *src8, int dstStride, int src8Stride, int h)\
|
|
|
|
|
static av_noinline void OPNAME ## pixels4_l2_shift5_ ## MMX(uint8_t *dst, int16_t *src16, uint8_t *src8, int dstStride, int src8Stride, int h)\
|
|
|
|
|
{\
|
|
|
|
|
asm volatile(\
|
|
|
|
|
"movq %5, %%mm6 \n\t"\
|
|
|
|
@ -1207,7 +1207,7 @@ static void OPNAME ## pixels4_l2_shift5_ ## MMX(uint8_t *dst, int16_t *src16, ui |
|
|
|
|
:"S"((long)src8Stride), "D"((long)dstStride), "m"(ff_pw_16)\
|
|
|
|
|
:"memory");\
|
|
|
|
|
}\
|
|
|
|
|
static void OPNAME ## pixels8_l2_shift5_ ## MMX(uint8_t *dst, int16_t *src16, uint8_t *src8, int dstStride, int src8Stride, int h)\
|
|
|
|
|
static av_noinline void OPNAME ## pixels8_l2_shift5_ ## MMX(uint8_t *dst, int16_t *src16, uint8_t *src8, int dstStride, int src8Stride, int h)\
|
|
|
|
|
{\
|
|
|
|
|
asm volatile(\
|
|
|
|
|
"movq %0, %%mm6 \n\t"\
|
|
|
|
|