diff options
Diffstat (limited to 'libavcodec/x86/dsputil_mmx.c')
| -rw-r--r-- | libavcodec/x86/dsputil_mmx.c | 31 |
1 files changed, 30 insertions, 1 deletions
diff --git a/libavcodec/x86/dsputil_mmx.c b/libavcodec/x86/dsputil_mmx.c index 31277d547b..e829bbd8b6 100644 --- a/libavcodec/x86/dsputil_mmx.c +++ b/libavcodec/x86/dsputil_mmx.c @@ -631,6 +631,34 @@ static void add_hfyu_median_prediction_cmov(uint8_t *dst, const uint8_t *top, } #endif +static inline void transpose4x4(uint8_t *dst, uint8_t *src, x86_reg dst_stride, x86_reg src_stride){ + __asm__ volatile( //FIXME could save 1 instruction if done as 8x4 ... + "movd (%1), %%mm0 \n\t" + "add %3, %1 \n\t" + "movd (%1), %%mm1 \n\t" + "movd (%1,%3,1), %%mm2 \n\t" + "movd (%1,%3,2), %%mm3 \n\t" + "punpcklbw %%mm1, %%mm0 \n\t" + "punpcklbw %%mm3, %%mm2 \n\t" + "movq %%mm0, %%mm1 \n\t" + "punpcklwd %%mm2, %%mm0 \n\t" + "punpckhwd %%mm2, %%mm1 \n\t" + "movd %%mm0, (%0) \n\t" + "add %2, %0 \n\t" + "punpckhdq %%mm0, %%mm0 \n\t" + "movd %%mm0, (%0) \n\t" + "movd %%mm1, (%0,%2,1) \n\t" + "punpckhdq %%mm1, %%mm1 \n\t" + "movd %%mm1, (%0,%2,2) \n\t" + + : "+&r" (dst), + "+&r" (src) + : "r" (dst_stride), + "r" (src_stride) + : "memory" + ); +} + #define H263_LOOP_FILTER \ "pxor %%mm7, %%mm7 \n\t" \ "movq %0, %%mm0 \n\t" \ @@ -2902,7 +2930,8 @@ static void dsputil_init_3dnow(DSPContext *c, AVCodecContext *avctx, c->vorbis_inverse_coupling = vorbis_inverse_coupling_3dnow; #if HAVE_7REGS - c->add_hfyu_median_prediction = add_hfyu_median_prediction_cmov; + if (mm_flags & AV_CPU_FLAG_CMOV) + c->add_hfyu_median_prediction = add_hfyu_median_prediction_cmov; #endif } |