diff options
Diffstat (limited to 'libavcodec')
30 files changed, 232 insertions, 159 deletions
diff --git a/libavcodec/ac3dsp.c b/libavcodec/ac3dsp.c index 4e1e4bd..49866eb 100644 --- a/libavcodec/ac3dsp.c +++ b/libavcodec/ac3dsp.c @@ -256,6 +256,6 @@ av_cold void ff_ac3dsp_init(AC3DSPContext *c, int bit_exact) if (ARCH_ARM) ff_ac3dsp_init_arm(c, bit_exact); - if (HAVE_MMX) + if (ARCH_X86) ff_ac3dsp_init_x86(c, bit_exact); } diff --git a/libavcodec/cavsdsp.c b/libavcodec/cavsdsp.c index 7eb8eca..a9b136e 100644 --- a/libavcodec/cavsdsp.c +++ b/libavcodec/cavsdsp.c @@ -546,5 +546,6 @@ av_cold void ff_cavsdsp_init(CAVSDSPContext* c, AVCodecContext *avctx) { c->cavs_idct8_add = cavs_idct8_add_c; c->idct_perm = FF_NO_IDCT_PERM; - if (HAVE_MMX) ff_cavsdsp_init_mmx(c, avctx); + if (ARCH_X86) + ff_cavsdsp_init_x86(c, avctx); } diff --git a/libavcodec/cavsdsp.h b/libavcodec/cavsdsp.h index b281133..f6e3e18 100644 --- a/libavcodec/cavsdsp.h +++ b/libavcodec/cavsdsp.h @@ -37,6 +37,6 @@ typedef struct CAVSDSPContext { } CAVSDSPContext; void ff_cavsdsp_init(CAVSDSPContext* c, AVCodecContext *avctx); -void ff_cavsdsp_init_mmx(CAVSDSPContext* c, AVCodecContext *avctx); +void ff_cavsdsp_init_x86(CAVSDSPContext* c, AVCodecContext *avctx); #endif /* AVCODEC_CAVSDSP_H */ diff --git a/libavcodec/dct.c b/libavcodec/dct.c index 2782bd9..e2ac0a8 100644 --- a/libavcodec/dct.c +++ b/libavcodec/dct.c @@ -209,8 +209,8 @@ av_cold int ff_dct_init(DCTContext *s, int nbits, enum DCTTransformType inverse) } s->dct32 = ff_dct32_float; - if (HAVE_MMX) - ff_dct_init_mmx(s); + if (ARCH_X86) + ff_dct_init_x86(s); return 0; } diff --git a/libavcodec/dct.h b/libavcodec/dct.h index bb17d75..8995f10 100644 --- a/libavcodec/dct.h +++ b/libavcodec/dct.h @@ -47,6 +47,6 @@ struct DCTContext { int ff_dct_init(DCTContext *s, int nbits, enum DCTTransformType type); void ff_dct_end (DCTContext *s); -void ff_dct_init_mmx(DCTContext *s); +void ff_dct_init_x86(DCTContext *s); #endif /* AVCODEC_DCT_H */ diff --git a/libavcodec/fft.c b/libavcodec/fft.c index e5bdcbd..00c434a 100644 --- a/libavcodec/fft.c +++ b/libavcodec/fft.c @@ -160,7 +160,7 @@ av_cold int ff_fft_init(FFTContext *s, int nbits, int inverse) #if CONFIG_FFT_FLOAT if (ARCH_ARM) ff_fft_init_arm(s); if (HAVE_ALTIVEC) ff_fft_init_altivec(s); - if (HAVE_MMX) ff_fft_init_mmx(s); + if (ARCH_X86) ff_fft_init_x86(s); if (CONFIG_MDCT) s->mdct_calcw = s->mdct_calc; if (HAVE_MIPSFPU) ff_fft_init_mips(s); #else diff --git a/libavcodec/fft.h b/libavcodec/fft.h index 15e5a12..9d92b2c 100644 --- a/libavcodec/fft.h +++ b/libavcodec/fft.h @@ -135,7 +135,7 @@ int ff_fft_init(FFTContext *s, int nbits, int inverse); #if CONFIG_FFT_FLOAT void ff_fft_init_altivec(FFTContext *s); -void ff_fft_init_mmx(FFTContext *s); +void ff_fft_init_x86(FFTContext *s); void ff_fft_init_arm(FFTContext *s); void ff_fft_init_mips(FFTContext *s); #else diff --git a/libavcodec/fmtconvert.c b/libavcodec/fmtconvert.c index 372f2a3..79e9645 100644 --- a/libavcodec/fmtconvert.c +++ b/libavcodec/fmtconvert.c @@ -85,7 +85,7 @@ av_cold void ff_fmt_convert_init(FmtConvertContext *c, AVCodecContext *avctx) if (ARCH_ARM) ff_fmt_convert_init_arm(c, avctx); if (HAVE_ALTIVEC) ff_fmt_convert_init_altivec(c, avctx); - if (HAVE_MMX) ff_fmt_convert_init_x86(c, avctx); + if (ARCH_X86) ff_fmt_convert_init_x86(c, avctx); if (HAVE_MIPSFPU) ff_fmt_convert_init_mips(c); } diff --git a/libavcodec/h264dsp.c b/libavcodec/h264dsp.c index ce6545b..69d0536 100644 --- a/libavcodec/h264dsp.c +++ b/libavcodec/h264dsp.c @@ -130,5 +130,5 @@ void ff_h264dsp_init(H264DSPContext *c, const int bit_depth, const int chroma_fo if (ARCH_ARM) ff_h264dsp_init_arm(c, bit_depth, chroma_format_idc); if (HAVE_ALTIVEC) ff_h264dsp_init_ppc(c, bit_depth, chroma_format_idc); - if (HAVE_MMX) ff_h264dsp_init_x86(c, bit_depth, chroma_format_idc); + if (ARCH_X86) ff_h264dsp_init_x86(c, bit_depth, chroma_format_idc); } diff --git a/libavcodec/h264pred.c b/libavcodec/h264pred.c index 2d8b9a4..5619efd 100644 --- a/libavcodec/h264pred.c +++ b/libavcodec/h264pred.c @@ -549,5 +549,5 @@ void ff_h264_pred_init(H264PredContext *h, int codec_id, const int bit_depth, co } if (ARCH_ARM) ff_h264_pred_init_arm(h, codec_id, bit_depth, chroma_format_idc); - if (HAVE_MMX) ff_h264_pred_init_x86(h, codec_id, bit_depth, chroma_format_idc); + if (ARCH_X86) ff_h264_pred_init_x86(h, codec_id, bit_depth, chroma_format_idc); } diff --git a/libavcodec/lpc.c b/libavcodec/lpc.c index ac8984a..c8d24c0 100644 --- a/libavcodec/lpc.c +++ b/libavcodec/lpc.c @@ -262,7 +262,7 @@ av_cold int ff_lpc_init(LPCContext *s, int blocksize, int max_order, s->lpc_apply_welch_window = lpc_apply_welch_window_c; s->lpc_compute_autocorr = lpc_compute_autocorr_c; - if (HAVE_MMX) + if (ARCH_X86) ff_lpc_init_x86(s); return 0; diff --git a/libavcodec/mpegaudiodsp.c b/libavcodec/mpegaudiodsp.c index 82a3652..aadc747 100644 --- a/libavcodec/mpegaudiodsp.c +++ b/libavcodec/mpegaudiodsp.c @@ -41,7 +41,7 @@ void ff_mpadsp_init(MPADSPContext *s) s->imdct36_blocks_fixed = ff_imdct36_blocks_fixed; if (ARCH_ARM) ff_mpadsp_init_arm(s); - if (HAVE_MMX) ff_mpadsp_init_mmx(s); + if (ARCH_X86) ff_mpadsp_init_x86(s); if (HAVE_ALTIVEC) ff_mpadsp_init_altivec(s); if (HAVE_MIPSFPU) ff_mpadsp_init_mipsfpu(s); if (HAVE_MIPSDSPR1) ff_mpadsp_init_mipsdspr1(s); diff --git a/libavcodec/mpegaudiodsp.h b/libavcodec/mpegaudiodsp.h index aae8ac0..623d2df 100644 --- a/libavcodec/mpegaudiodsp.h +++ b/libavcodec/mpegaudiodsp.h @@ -56,7 +56,7 @@ void ff_mpa_synth_filter_float(MPADSPContext *s, float *sb_samples); void ff_mpadsp_init_arm(MPADSPContext *s); -void ff_mpadsp_init_mmx(MPADSPContext *s); +void ff_mpadsp_init_x86(MPADSPContext *s); void ff_mpadsp_init_altivec(MPADSPContext *s); void ff_mpadsp_init_mipsfpu(MPADSPContext *s); void ff_mpadsp_init_mipsdspr1(MPADSPContext *s); diff --git a/libavcodec/pngdsp.c b/libavcodec/pngdsp.c index 75ec996..1ee8b57 100644 --- a/libavcodec/pngdsp.c +++ b/libavcodec/pngdsp.c @@ -44,5 +44,5 @@ void ff_pngdsp_init(PNGDSPContext *dsp) dsp->add_bytes_l2 = add_bytes_l2_c; dsp->add_paeth_prediction = ff_add_png_paeth_prediction; - if (HAVE_MMX) ff_pngdsp_init_x86(dsp); + if (ARCH_X86) ff_pngdsp_init_x86(dsp); } diff --git a/libavcodec/proresdsp.c b/libavcodec/proresdsp.c index 867e403..9e063b0 100644 --- a/libavcodec/proresdsp.c +++ b/libavcodec/proresdsp.c @@ -75,7 +75,7 @@ void ff_proresdsp_init(ProresDSPContext *dsp, AVCodecContext *avctx) dsp->idct_put = prores_idct_put_c; dsp->idct_permutation_type = FF_NO_IDCT_PERM; - if (HAVE_MMX) ff_proresdsp_x86_init(dsp, avctx); + if (ARCH_X86) ff_proresdsp_x86_init(dsp, avctx); ff_init_scantable_permutation(dsp->idct_permutation, dsp->idct_permutation_type); diff --git a/libavcodec/rv34dsp.c b/libavcodec/rv34dsp.c index ac14ed9..25e8c3d 100644 --- a/libavcodec/rv34dsp.c +++ b/libavcodec/rv34dsp.c @@ -137,6 +137,6 @@ av_cold void ff_rv34dsp_init(RV34DSPContext *c, DSPContext* dsp) { if (HAVE_NEON) ff_rv34dsp_init_neon(c, dsp); - if (HAVE_MMX) + if (ARCH_X86) ff_rv34dsp_init_x86(c, dsp); } diff --git a/libavcodec/rv40dsp.c b/libavcodec/rv40dsp.c index 66f5326..2799f1a 100644 --- a/libavcodec/rv40dsp.c +++ b/libavcodec/rv40dsp.c @@ -604,7 +604,7 @@ av_cold void ff_rv40dsp_init(RV34DSPContext *c, DSPContext* dsp) { c->rv40_loop_filter_strength[0] = rv40_h_loop_filter_strength; c->rv40_loop_filter_strength[1] = rv40_v_loop_filter_strength; - if (HAVE_MMX) + if (ARCH_X86) ff_rv40dsp_init_x86(c, dsp); if (HAVE_NEON) ff_rv40dsp_init_neon(c, dsp); diff --git a/libavcodec/sbrdsp.c b/libavcodec/sbrdsp.c index 8c88fb3..781ec83 100644 --- a/libavcodec/sbrdsp.c +++ b/libavcodec/sbrdsp.c @@ -243,6 +243,6 @@ av_cold void ff_sbrdsp_init(SBRDSPContext *s) if (ARCH_ARM) ff_sbrdsp_init_arm(s); - if (HAVE_MMX) + if (ARCH_X86) ff_sbrdsp_init_x86(s); } diff --git a/libavcodec/vc1dsp.c b/libavcodec/vc1dsp.c index c2b7eff..a874509 100644 --- a/libavcodec/vc1dsp.c +++ b/libavcodec/vc1dsp.c @@ -850,6 +850,6 @@ av_cold void ff_vc1dsp_init(VC1DSPContext* dsp) { if (HAVE_ALTIVEC) ff_vc1dsp_init_altivec(dsp); - if (HAVE_MMX) - ff_vc1dsp_init_mmx(dsp); + if (ARCH_X86) + ff_vc1dsp_init_x86(dsp); } diff --git a/libavcodec/vc1dsp.h b/libavcodec/vc1dsp.h index d96853a..8775818 100644 --- a/libavcodec/vc1dsp.h +++ b/libavcodec/vc1dsp.h @@ -74,6 +74,6 @@ typedef struct VC1DSPContext { void ff_vc1dsp_init(VC1DSPContext* c); void ff_vc1dsp_init_altivec(VC1DSPContext* c); -void ff_vc1dsp_init_mmx(VC1DSPContext* dsp); +void ff_vc1dsp_init_x86(VC1DSPContext* dsp); #endif /* AVCODEC_VC1DSP_H */ diff --git a/libavcodec/vp56dsp.c b/libavcodec/vp56dsp.c index 339c3d2..a72c48e 100644 --- a/libavcodec/vp56dsp.c +++ b/libavcodec/vp56dsp.c @@ -90,5 +90,5 @@ void ff_vp56dsp_init(VP56DSPContext *s, enum AVCodecID codec) } if (ARCH_ARM) ff_vp56dsp_init_arm(s, codec); - if (HAVE_MMX) ff_vp56dsp_init_x86(s, codec); + if (ARCH_X86) ff_vp56dsp_init_x86(s, codec); } diff --git a/libavcodec/vp8dsp.c b/libavcodec/vp8dsp.c index bacd10e..a53643c 100644 --- a/libavcodec/vp8dsp.c +++ b/libavcodec/vp8dsp.c @@ -521,7 +521,7 @@ av_cold void ff_vp8dsp_init(VP8DSPContext *dsp) VP8_BILINEAR_MC_FUNC(1, 8); VP8_BILINEAR_MC_FUNC(2, 4); - if (HAVE_MMX) + if (ARCH_X86) ff_vp8dsp_init_x86(dsp); if (HAVE_ALTIVEC) ff_vp8dsp_init_altivec(dsp); diff --git a/libavcodec/x86/Makefile b/libavcodec/x86/Makefile index 39d2727..4ddd06d 100644 --- a/libavcodec/x86/Makefile +++ b/libavcodec/x86/Makefile @@ -1,42 +1,43 @@ +OBJS += x86/fmtconvert_init.o + +OBJS-$(CONFIG_AAC_DECODER) += x86/sbrdsp_init.o +OBJS-$(CONFIG_AC3DSP) += x86/ac3dsp_init.o +OBJS-$(CONFIG_CAVS_DECODER) += x86/cavsdsp.o OBJS-$(CONFIG_DNXHD_ENCODER) += x86/dnxhdenc.o +OBJS-$(CONFIG_FFT) += x86/fft_init.o +OBJS-$(CONFIG_GPL) += x86/idct_mmx.o +OBJS-$(CONFIG_H264DSP) += x86/h264dsp_init.o +OBJS-$(CONFIG_H264PRED) += x86/h264_intrapred_init.o +OBJS-$(CONFIG_LPC) += x86/lpc.o OBJS-$(CONFIG_MLP_DECODER) += x86/mlpdsp.o +OBJS-$(CONFIG_MPEGAUDIODSP) += x86/mpegaudiodec.o OBJS-$(CONFIG_MPEGVIDEO) += x86/mpegvideo.o OBJS-$(CONFIG_MPEGVIDEOENC) += x86/mpegvideoenc.o +OBJS-$(CONFIG_PNG_DECODER) += x86/pngdsp_init.o +OBJS-$(CONFIG_PRORES_LGPL_DECODER) += x86/proresdsp_init.o +OBJS-$(CONFIG_RV30_DECODER) += x86/rv34dsp_init.o +OBJS-$(CONFIG_RV40_DECODER) += x86/rv34dsp_init.o \ + x86/rv40dsp_init.o +OBJS-$(CONFIG_V210_DECODER) += x86/v210-init.o OBJS-$(CONFIG_TRUEHD_DECODER) += x86/mlpdsp.o +OBJS-$(CONFIG_VC1_DECODER) += x86/vc1dsp_init.o OBJS-$(CONFIG_VP3DSP) += x86/vp3dsp_init.o +OBJS-$(CONFIG_VP5_DECODER) += x86/vp56dsp_init.o +OBJS-$(CONFIG_VP6_DECODER) += x86/vp56dsp_init.o +OBJS-$(CONFIG_VP8_DECODER) += x86/vp8dsp_init.o OBJS-$(CONFIG_XMM_CLOBBER_TEST) += x86/w64xmmtest.o MMX-OBJS += x86/dsputil_mmx.o \ x86/fdct.o \ - x86/fmtconvert_init.o \ x86/idct_mmx_xvid.o \ x86/idct_sse2_xvid.o \ x86/simple_idct.o \ -MMX-OBJS-$(CONFIG_AAC_DECODER) += x86/sbrdsp_init.o -MMX-OBJS-$(CONFIG_AC3DSP) += x86/ac3dsp_init.o -MMX-OBJS-$(CONFIG_CAVS_DECODER) += x86/cavsdsp.o MMX-OBJS-$(CONFIG_DWT) += x86/snowdsp.o \ x86/dwt.o MMX-OBJS-$(CONFIG_ENCODERS) += x86/dsputilenc_mmx.o \ x86/motion_est.o -MMX-OBJS-$(CONFIG_FFT) += x86/fft_init.o -MMX-OBJS-$(CONFIG_GPL) += x86/idct_mmx.o -MMX-OBJS-$(CONFIG_H264DSP) += x86/h264dsp_init.o -MMX-OBJS-$(CONFIG_H264PRED) += x86/h264_intrapred_init.o -MMX-OBJS-$(CONFIG_LPC) += x86/lpc.o -MMX-OBJS-$(CONFIG_MPEGAUDIODSP) += x86/mpegaudiodec.o -MMX-OBJS-$(CONFIG_PNG_DECODER) += x86/pngdsp_init.o -MMX-OBJS-$(CONFIG_PRORES_DECODER) += x86/proresdsp_init.o -MMX-OBJS-$(CONFIG_PRORES_LGPL_DECODER) += x86/proresdsp_init.o -MMX-OBJS-$(CONFIG_RV30_DECODER) += x86/rv34dsp_init.o -MMX-OBJS-$(CONFIG_RV40_DECODER) += x86/rv34dsp_init.o \ - x86/rv40dsp_init.o -MMX-OBJS-$(CONFIG_V210_DECODER) += x86/v210-init.o MMX-OBJS-$(CONFIG_VC1_DECODER) += x86/vc1dsp_mmx.o -MMX-OBJS-$(CONFIG_VP5_DECODER) += x86/vp56dsp_init.o -MMX-OBJS-$(CONFIG_VP6_DECODER) += x86/vp56dsp_init.o -MMX-OBJS-$(CONFIG_VP8_DECODER) += x86/vp8dsp_init.o YASM-OBJS-$(CONFIG_AAC_DECODER) += x86/sbrdsp.o YASM-OBJS-$(CONFIG_AC3DSP) += x86/ac3dsp.o diff --git a/libavcodec/x86/cavsdsp.c b/libavcodec/x86/cavsdsp.c index 03a8b87..9643b81 100644 --- a/libavcodec/x86/cavsdsp.c +++ b/libavcodec/x86/cavsdsp.c @@ -25,11 +25,13 @@ #include "libavutil/common.h" #include "libavutil/cpu.h" #include "libavutil/x86/asm.h" +#include "libavutil/x86/cpu.h" #include "libavcodec/dsputil.h" #include "libavcodec/cavsdsp.h" #include "dsputil_mmx.h" +#include "config.h" -#if HAVE_INLINE_ASM +#if (HAVE_MMXEXT_INLINE || HAVE_AMD3DNOW_INLINE) /* in/out: mma=mma+mmb, mmb=mmb-mma */ #define SUMSUB_BA( a, b ) \ @@ -433,15 +435,12 @@ static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc03_ ## MMX(uint8_t *dst, ui "pavgb " #temp ", " #a " \n\t"\ "mov" #size " " #a ", " #b " \n\t" -QPEL_CAVS(put_, PUT_OP, 3dnow) -QPEL_CAVS(avg_, AVG_3DNOW_OP, 3dnow) +#endif /* (HAVE_MMXEXT_INLINE || HAVE_AMD3DNOW_INLINE) */ + +#if HAVE_MMXEXT_INLINE QPEL_CAVS(put_, PUT_OP, mmx2) QPEL_CAVS(avg_, AVG_MMX2_OP, mmx2) -CAVS_MC(put_, 8, 3dnow) -CAVS_MC(put_, 16,3dnow) -CAVS_MC(avg_, 8, 3dnow) -CAVS_MC(avg_, 16,3dnow) CAVS_MC(put_, 8, mmx2) CAVS_MC(put_, 16,mmx2) CAVS_MC(avg_, 8, mmx2) @@ -463,6 +462,16 @@ static void ff_cavsdsp_init_mmx2(CAVSDSPContext* c, AVCodecContext *avctx) { c->cavs_idct8_add = cavs_idct8_add_mmx; c->idct_perm = FF_TRANSPOSE_IDCT_PERM; } +#endif /* HAVE_MMXEXT_INLINE */ + +#if HAVE_AMD3DNOW_INLINE +QPEL_CAVS(put_, PUT_OP, 3dnow) +QPEL_CAVS(avg_, AVG_3DNOW_OP, 3dnow) + +CAVS_MC(put_, 8, 3dnow) +CAVS_MC(put_, 16,3dnow) +CAVS_MC(avg_, 8, 3dnow) +CAVS_MC(avg_, 16,3dnow) static void ff_cavsdsp_init_3dnow(CAVSDSPContext* c, AVCodecContext *avctx) { #define dspfunc(PFX, IDX, NUM) \ @@ -480,15 +489,16 @@ static void ff_cavsdsp_init_3dnow(CAVSDSPContext* c, AVCodecContext *avctx) { c->cavs_idct8_add = cavs_idct8_add_mmx; c->idct_perm = FF_TRANSPOSE_IDCT_PERM; } +#endif /* HAVE_AMD3DNOW_INLINE */ -#endif /* HAVE_INLINE_ASM */ - -void ff_cavsdsp_init_mmx(CAVSDSPContext *c, AVCodecContext *avctx) +av_cold void ff_cavsdsp_init_x86(CAVSDSPContext *c, AVCodecContext *avctx) { int mm_flags = av_get_cpu_flags(); -#if HAVE_INLINE_ASM +#if HAVE_MMXEXT_INLINE if (mm_flags & AV_CPU_FLAG_MMXEXT) ff_cavsdsp_init_mmx2(c, avctx); +#endif /* HAVE_MMXEXT_INLINE */ +#if HAVE_AMD3DNOW_INLINE if (mm_flags & AV_CPU_FLAG_3DNOW) ff_cavsdsp_init_3dnow(c, avctx); -#endif /* HAVE_INLINE_ASM */ +#endif /* HAVE_AMD3DNOW_INLINE */ } diff --git a/libavcodec/x86/fft_init.c b/libavcodec/x86/fft_init.c index d7a31d4..a9d9579 100644 --- a/libavcodec/x86/fft_init.c +++ b/libavcodec/x86/fft_init.c @@ -22,7 +22,7 @@ #include "libavcodec/dct.h" #include "fft.h" -av_cold void ff_fft_init_mmx(FFTContext *s) +av_cold void ff_fft_init_x86(FFTContext *s) { int has_vectors = av_get_cpu_flags(); #if ARCH_X86_32 @@ -56,7 +56,7 @@ av_cold void ff_fft_init_mmx(FFTContext *s) } #if CONFIG_DCT -av_cold void ff_dct_init_mmx(DCTContext *s) +av_cold void ff_dct_init_x86(DCTContext *s) { int has_vectors = av_get_cpu_flags(); if (EXTERNAL_SSE(has_vectors)) diff --git a/libavcodec/x86/mpegaudiodec.c b/libavcodec/x86/mpegaudiodec.c index a8b55ae..44cbf9a 100644 --- a/libavcodec/x86/mpegaudiodec.c +++ b/libavcodec/x86/mpegaudiodec.c @@ -232,7 +232,7 @@ DECL_IMDCT_BLOCKS(avx,avx) #endif #endif /* HAVE_YASM */ -void ff_mpadsp_init_mmx(MPADSPContext *s) +void ff_mpadsp_init_x86(MPADSPContext *s) { int mm_flags = av_get_cpu_flags(); diff --git a/libavcodec/x86/rv40dsp_init.c b/libavcodec/x86/rv40dsp_init.c index 966e3f2..902ae19 100644 --- a/libavcodec/x86/rv40dsp_init.c +++ b/libavcodec/x86/rv40dsp_init.c @@ -195,12 +195,12 @@ void ff_rv40dsp_init_x86(RV34DSPContext *c, DSPContext *dsp) if (EXTERNAL_MMX(mm_flags)) { c->put_chroma_pixels_tab[0] = ff_put_rv40_chroma_mc8_mmx; c->put_chroma_pixels_tab[1] = ff_put_rv40_chroma_mc4_mmx; -#if HAVE_INLINE_ASM +#if HAVE_MMX_INLINE c->put_pixels_tab[0][15] = ff_put_rv40_qpel16_mc33_mmx; c->put_pixels_tab[1][15] = ff_put_rv40_qpel8_mc33_mmx; c->avg_pixels_tab[0][15] = ff_avg_rv40_qpel16_mc33_mmx; c->avg_pixels_tab[1][15] = ff_avg_rv40_qpel8_mc33_mmx; -#endif /* HAVE_INLINE_ASM */ +#endif /* HAVE_MMX_INLINE */ #if ARCH_X86_32 QPEL_MC_SET(put_, _mmx) #endif diff --git a/libavcodec/x86/vc1dsp.h b/libavcodec/x86/vc1dsp.h new file mode 100644 index 0000000..fdd4de1 --- /dev/null +++ b/libavcodec/x86/vc1dsp.h @@ -0,0 +1,29 @@ +/* + * VC-1 and WMV3 decoder - X86 DSP init functions + * + * This file is part of FFmpeg. + * + * FFmpeg is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * FFmpeg is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with FFmpeg; if not, write to the Free Software + * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA + */ + +#ifndef AVCODEC_X86_VC1DSP_H +#define AVCODEC_X86_VC1DSP_H + +#include "libavcodec/vc1dsp.h" + +void ff_vc1dsp_init_mmx(VC1DSPContext *dsp); +void ff_vc1dsp_init_mmxext(VC1DSPContext *dsp); + +#endif /* AVCODEC_X86_VC1DSP_H */ diff --git a/libavcodec/x86/vc1dsp_init.c b/libavcodec/x86/vc1dsp_init.c new file mode 100644 index 0000000..d2548fc --- /dev/null +++ b/libavcodec/x86/vc1dsp_init.c @@ -0,0 +1,123 @@ +/* + * VC-1 and WMV3 - DSP functions MMX-optimized + * Copyright (c) 2007 Christophe GISQUET <christophe.gisquet@free.fr> + * + * Permission is hereby granted, free of charge, to any person + * obtaining a copy of this software and associated documentation + * files (the "Software"), to deal in the Software without + * restriction, including without limitation the rights to use, + * copy, modify, merge, publish, distribute, sublicense, and/or sell + * copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following + * conditions: + * + * The above copyright notice and this permission notice shall be + * included in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, + * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES + * OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND + * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT + * HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, + * WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + */ + +#include "libavutil/cpu.h" +#include "libavutil/x86/cpu.h" +#include "libavcodec/vc1dsp.h" +#include "vc1dsp.h" +#include "config.h" + +#define LOOP_FILTER(EXT) \ +void ff_vc1_v_loop_filter4_ ## EXT(uint8_t *src, int stride, int pq); \ +void ff_vc1_h_loop_filter4_ ## EXT(uint8_t *src, int stride, int pq); \ +void ff_vc1_v_loop_filter8_ ## EXT(uint8_t *src, int stride, int pq); \ +void ff_vc1_h_loop_filter8_ ## EXT(uint8_t *src, int stride, int pq); \ +\ +static void vc1_v_loop_filter16_ ## EXT(uint8_t *src, int stride, int pq) \ +{ \ + ff_vc1_v_loop_filter8_ ## EXT(src, stride, pq); \ + ff_vc1_v_loop_filter8_ ## EXT(src+8, stride, pq); \ +} \ +\ +static void vc1_h_loop_filter16_ ## EXT(uint8_t *src, int stride, int pq) \ +{ \ + ff_vc1_h_loop_filter8_ ## EXT(src, stride, pq); \ + ff_vc1_h_loop_filter8_ ## EXT(src+8*stride, stride, pq); \ +} + +#if HAVE_YASM +LOOP_FILTER(mmx2) +LOOP_FILTER(sse2) +LOOP_FILTER(ssse3) + +void ff_vc1_h_loop_filter8_sse4(uint8_t *src, int stride, int pq); + +static void vc1_h_loop_filter16_sse4(uint8_t *src, int stride, int pq) +{ + ff_vc1_h_loop_filter8_sse4(src, stride, pq); + ff_vc1_h_loop_filter8_sse4(src+8*stride, stride, pq); +} +#endif /* HAVE_YASM */ + +void ff_put_vc1_chroma_mc8_mmx_nornd (uint8_t *dst, uint8_t *src, + int stride, int h, int x, int y); +void ff_avg_vc1_chroma_mc8_mmx2_nornd (uint8_t *dst, uint8_t *src, + int stride, int h, int x, int y); +void ff_avg_vc1_chroma_mc8_3dnow_nornd(uint8_t *dst, uint8_t *src, + int stride, int h, int x, int y); +void ff_put_vc1_chroma_mc8_ssse3_nornd(uint8_t *dst, uint8_t *src, + int stride, int h, int x, int y); +void ff_avg_vc1_chroma_mc8_ssse3_nornd(uint8_t *dst, uint8_t *src, + int stride, int h, int x, int y); + + +av_cold void ff_vc1dsp_init_x86(VC1DSPContext *dsp) +{ + int mm_flags = av_get_cpu_flags(); + + if (INLINE_MMX(mm_flags)) + ff_vc1dsp_init_mmx(dsp); + + if (INLINE_MMXEXT(mm_flags)) + ff_vc1dsp_init_mmxext(dsp); + +#define ASSIGN_LF(EXT) \ + dsp->vc1_v_loop_filter4 = ff_vc1_v_loop_filter4_ ## EXT; \ + dsp->vc1_h_loop_filter4 = ff_vc1_h_loop_filter4_ ## EXT; \ + dsp->vc1_v_loop_filter8 = ff_vc1_v_loop_filter8_ ## EXT; \ + dsp->vc1_h_loop_filter8 = ff_vc1_h_loop_filter8_ ## EXT; \ + dsp->vc1_v_loop_filter16 = vc1_v_loop_filter16_ ## EXT; \ + dsp->vc1_h_loop_filter16 = vc1_h_loop_filter16_ ## EXT + +#if HAVE_YASM + if (mm_flags & AV_CPU_FLAG_MMX) { + dsp->put_no_rnd_vc1_chroma_pixels_tab[0]= ff_put_vc1_chroma_mc8_mmx_nornd; + } + + if (mm_flags & AV_CPU_FLAG_MMXEXT) { + ASSIGN_LF(mmx2); + dsp->avg_no_rnd_vc1_chroma_pixels_tab[0]= ff_avg_vc1_chroma_mc8_mmx2_nornd; + } else if (mm_flags & AV_CPU_FLAG_3DNOW) { + dsp->avg_no_rnd_vc1_chroma_pixels_tab[0]= ff_avg_vc1_chroma_mc8_3dnow_nornd; + } + + if (mm_flags & AV_CPU_FLAG_SSE2) { + dsp->vc1_v_loop_filter8 = ff_vc1_v_loop_filter8_sse2; + dsp->vc1_h_loop_filter8 = ff_vc1_h_loop_filter8_sse2; + dsp->vc1_v_loop_filter16 = vc1_v_loop_filter16_sse2; + dsp->vc1_h_loop_filter16 = vc1_h_loop_filter16_sse2; + } + if (mm_flags & AV_CPU_FLAG_SSSE3) { + ASSIGN_LF(ssse3); + dsp->put_no_rnd_vc1_chroma_pixels_tab[0]= ff_put_vc1_chroma_mc8_ssse3_nornd; + dsp->avg_no_rnd_vc1_chroma_pixels_tab[0]= ff_avg_vc1_chroma_mc8_ssse3_nornd; + } + if (mm_flags & AV_CPU_FLAG_SSE4) { + dsp->vc1_h_loop_filter8 = ff_vc1_h_loop_filter8_sse4; + dsp->vc1_h_loop_filter16 = vc1_h_loop_filter16_sse4; + } +#endif /* HAVE_YASM */ +} diff --git a/libavcodec/x86/vc1dsp_mmx.c b/libavcodec/x86/vc1dsp_mmx.c index 66251d0..adf89b2 100644 --- a/libavcodec/x86/vc1dsp_mmx.c +++ b/libavcodec/x86/vc1dsp_mmx.c @@ -27,9 +27,11 @@ #include "libavutil/cpu.h" #include "libavutil/mem.h" #include "libavutil/x86/asm.h" +#include "libavutil/x86/cpu.h" #include "libavcodec/dsputil.h" #include "dsputil_mmx.h" #include "libavcodec/vc1dsp.h" +#include "vc1dsp.h" #if HAVE_INLINE_ASM @@ -685,57 +687,8 @@ static void vc1_inv_trans_8x8_dc_mmx2(uint8_t *dest, int linesize, DCTELEM *bloc ); } -#endif /* HAVE_INLINE_ASM */ - -#define LOOP_FILTER(EXT) \ -void ff_vc1_v_loop_filter4_ ## EXT(uint8_t *src, int stride, int pq); \ -void ff_vc1_h_loop_filter4_ ## EXT(uint8_t *src, int stride, int pq); \ -void ff_vc1_v_loop_filter8_ ## EXT(uint8_t *src, int stride, int pq); \ -void ff_vc1_h_loop_filter8_ ## EXT(uint8_t *src, int stride, int pq); \ -\ -static void vc1_v_loop_filter16_ ## EXT(uint8_t *src, int stride, int pq) \ -{ \ - ff_vc1_v_loop_filter8_ ## EXT(src, stride, pq); \ - ff_vc1_v_loop_filter8_ ## EXT(src+8, stride, pq); \ -} \ -\ -static void vc1_h_loop_filter16_ ## EXT(uint8_t *src, int stride, int pq) \ -{ \ - ff_vc1_h_loop_filter8_ ## EXT(src, stride, pq); \ - ff_vc1_h_loop_filter8_ ## EXT(src+8*stride, stride, pq); \ -} - -#if HAVE_YASM -LOOP_FILTER(mmx2) -LOOP_FILTER(sse2) -LOOP_FILTER(ssse3) - -void ff_vc1_h_loop_filter8_sse4(uint8_t *src, int stride, int pq); - -static void vc1_h_loop_filter16_sse4(uint8_t *src, int stride, int pq) -{ - ff_vc1_h_loop_filter8_sse4(src, stride, pq); - ff_vc1_h_loop_filter8_sse4(src+8*stride, stride, pq); -} -#endif /* HAVE_YASM */ - -void ff_put_vc1_chroma_mc8_mmx_nornd (uint8_t *dst, uint8_t *src, - int stride, int h, int x, int y); -void ff_avg_vc1_chroma_mc8_mmx2_nornd (uint8_t *dst, uint8_t *src, - int stride, int h, int x, int y); -void ff_avg_vc1_chroma_mc8_3dnow_nornd(uint8_t *dst, uint8_t *src, - int stride, int h, int x, int y); -void ff_put_vc1_chroma_mc8_ssse3_nornd(uint8_t *dst, uint8_t *src, - int stride, int h, int x, int y); -void ff_avg_vc1_chroma_mc8_ssse3_nornd(uint8_t *dst, uint8_t *src, - int stride, int h, int x, int y); - -void ff_vc1dsp_init_mmx(VC1DSPContext *dsp) +av_cold void ff_vc1dsp_init_mmx(VC1DSPContext *dsp) { - int mm_flags = av_get_cpu_flags(); - -#if HAVE_INLINE_ASM - if (mm_flags & AV_CPU_FLAG_MMX) { dsp->put_vc1_mspel_pixels_tab[ 0] = ff_put_vc1_mspel_mc00_mmx; dsp->put_vc1_mspel_pixels_tab[ 4] = put_vc1_mspel_mc01_mmx; dsp->put_vc1_mspel_pixels_tab[ 8] = put_vc1_mspel_mc02_mmx; @@ -755,12 +708,10 @@ void ff_vc1dsp_init_mmx(VC1DSPContext *dsp) dsp->put_vc1_mspel_pixels_tab[ 7] = put_vc1_mspel_mc31_mmx; dsp->put_vc1_mspel_pixels_tab[11] = put_vc1_mspel_mc32_mmx; dsp->put_vc1_mspel_pixels_tab[15] = put_vc1_mspel_mc33_mmx; +} - if (HAVE_YASM) - dsp->put_no_rnd_vc1_chroma_pixels_tab[0]= ff_put_vc1_chroma_mc8_mmx_nornd; - } - - if (mm_flags & AV_CPU_FLAG_MMXEXT) { +av_cold void ff_vc1dsp_init_mmxext(VC1DSPContext *dsp) +{ dsp->avg_vc1_mspel_pixels_tab[ 0] = ff_avg_vc1_mspel_mc00_mmx2; dsp->avg_vc1_mspel_pixels_tab[ 4] = avg_vc1_mspel_mc01_mmx2; dsp->avg_vc1_mspel_pixels_tab[ 8] = avg_vc1_mspel_mc02_mmx2; @@ -785,47 +736,5 @@ void ff_vc1dsp_init_mmx(VC1DSPContext *dsp) dsp->vc1_inv_trans_4x8_dc = vc1_inv_trans_4x8_dc_mmx2; dsp->vc1_inv_trans_8x4_dc = vc1_inv_trans_8x4_dc_mmx2; dsp->vc1_inv_trans_4x4_dc = vc1_inv_trans_4x4_dc_mmx2; - - if (HAVE_YASM) - dsp->avg_no_rnd_vc1_chroma_pixels_tab[0]= ff_avg_vc1_chroma_mc8_mmx2_nornd; - } else if (HAVE_YASM && mm_flags & AV_CPU_FLAG_3DNOW) { - dsp->avg_no_rnd_vc1_chroma_pixels_tab[0]= ff_avg_vc1_chroma_mc8_3dnow_nornd; - } - - if (HAVE_YASM && mm_flags & AV_CPU_FLAG_SSSE3) { - dsp->put_no_rnd_vc1_chroma_pixels_tab[0]= ff_put_vc1_chroma_mc8_ssse3_nornd; - dsp->avg_no_rnd_vc1_chroma_pixels_tab[0]= ff_avg_vc1_chroma_mc8_ssse3_nornd; - } -#endif /* HAVE_INLINE_ASM */ - -#define ASSIGN_LF(EXT) \ - dsp->vc1_v_loop_filter4 = ff_vc1_v_loop_filter4_ ## EXT; \ - dsp->vc1_h_loop_filter4 = ff_vc1_h_loop_filter4_ ## EXT; \ - dsp->vc1_v_loop_filter8 = ff_vc1_v_loop_filter8_ ## EXT; \ - dsp->vc1_h_loop_filter8 = ff_vc1_h_loop_filter8_ ## EXT; \ - dsp->vc1_v_loop_filter16 = vc1_v_loop_filter16_ ## EXT; \ - dsp->vc1_h_loop_filter16 = vc1_h_loop_filter16_ ## EXT - -#if HAVE_YASM - if (mm_flags & AV_CPU_FLAG_MMX) { - } - - if (mm_flags & AV_CPU_FLAG_MMXEXT) { - ASSIGN_LF(mmx2); - } - - if (mm_flags & AV_CPU_FLAG_SSE2) { - dsp->vc1_v_loop_filter8 = ff_vc1_v_loop_filter8_sse2; - dsp->vc1_h_loop_filter8 = ff_vc1_h_loop_filter8_sse2; - dsp->vc1_v_loop_filter16 = vc1_v_loop_filter16_sse2; - dsp->vc1_h_loop_filter16 = vc1_h_loop_filter16_sse2; - } - if (mm_flags & AV_CPU_FLAG_SSSE3) { - ASSIGN_LF(ssse3); - } - if (mm_flags & AV_CPU_FLAG_SSE4) { - dsp->vc1_h_loop_filter8 = ff_vc1_h_loop_filter8_sse4; - dsp->vc1_h_loop_filter16 = vc1_h_loop_filter16_sse4; - } -#endif /* HAVE_YASM */ } +#endif /* HAVE_INLINE_ASM */ |
