Merge remote-tracking branch 'qatar/master'

* qatar/master: (35 commits) h264_idct_10bit: port x86 assembly to cpuflags. x86inc: clip num_args to 7 on x86-32. x86inc: sync to latest version from x264. fft: rename "z" to "zc" to prevent name collision. wv: return meaningful error codes. wv: return AVERROR_EOF on EOF, not EIO. mp3dec: forward errors for av_get_packet(). mp3dec: remove a pointless local variable. mp3dec: remove commented out cruft. lavfi: bump minor to mark stabilizing the ABI. FATE: add tests for yadif. FATE: add a test for delogo video filter. FATE: add a test for amix audio filter. audiogen: allow specifying random seed as a commandline parameter. vc1dec: Override invalid macroblock quantizer vc1: avoid reading beyond the last line in vc1_draw_sprites() vc1dec: check that coded slice positions and interlacing match. vc1dec: Do not ignore ff_vc1_parse_frame_header_adv return value configure: Move parts that should not be user-selectable to CONFIG_EXTRA lavf: remove commented out cruft in avformat_find_stream_info() ... Conflicts: Makefile configure libavcodec/vc1dec.c libavcodec/x86/h264_deblock.asm libavcodec/x86/h264_deblock_10bit.asm libavcodec/x86/h264dsp_mmx.c libavfilter/version.h libavformat/mp3dec.c libavformat/utils.c libavformat/wv.c libavutil/x86/x86inc.asm Merged-by: Michael Niedermayer <michaelni@gmx.at>
12 years ago · 706bd8ea19
parent 0f8f924847 c83f44dba1
commit 706bd8ea19
40 changed files with 807 additions and 525 deletions
--- a/common.mak
+++ b/common.mak
@ -11,7 +11,7 @@ ifndef V
 Q      = @
 ECHO   = printf "$(1)\t%s\n" $(2)
 BRIEF  = CC CXX AS YASM AR LD HOSTCC STRIP CP
-SILENT = DEPCC DEPAS DEPHOSTCC YASMDEP RM RANLIB
+SILENT = DEPCC DEPAS DEPHOSTCC DEPYASM RM RANLIB
 MSG    = $@
 M      = @$(call ECHO,$(TAG),$@);
 $(foreach VAR,$(BRIEF), \
@ -35,7 +35,7 @@ LDFLAGS    := $(ALLFFLIBS:%=-Llib%) $(LDFLAGS)
 define COMPILE
       $(call $(1)DEP,$(1))
-       $($(1)) $($(1)FLAGS) $($(1)_DEPFLAGS) -c $($(1)_O) $<
+       $($(1)) $($(1)FLAGS) $($(1)_DEPFLAGS) $($(1)_C) $($(1)_O) $<
 endef
 COMPILE_C = $(call COMPILE,CC)
@ -55,7 +55,7 @@ COMPILE_S = $(call COMPILE,AS)
 	$(COMPILE_S)
 %.ho: %.h
-	$(CC) $(CPPFLAGS) $(CFLAGS) -c -o $@ -x c $<
+	$(CC) $(CCFLAGS) -c $(CC_O) -x c $<
 %.ver: %.v
 	$(Q)sed 's/$$MAJOR/$($(basename $(@F))_VERSION_MAJOR)/' $^ > $@
--- a/39
+++ b/39
@ -673,11 +673,15 @@ cc_o(){
    eval printf '%s\\n' $CC_O
 }
 cc_e(){
    eval printf '%s\\n' $CC_E
 }
 check_cc(){
    log check_cc "$@"
    cat > $TMPC
    log_file $TMPC
-    check_cmd $cc $CPPFLAGS $CFLAGS "$@" -c $(cc_o $TMPO) $TMPC
+    check_cmd $cc $CPPFLAGS $CFLAGS "$@" $CC_C $(cc_o $TMPO) $TMPC
 }
 check_cxx(){
@ -691,14 +695,14 @@ check_cpp(){
    log check_cpp "$@"
    cat > $TMPC
    log_file $TMPC
-    check_cmd $cc $CPPFLAGS $CFLAGS "$@" -E -o $TMPO $TMPC
+    check_cmd $cc $CPPFLAGS $CFLAGS "$@" $(cc_e $TMPO) $TMPC
 }
 check_as(){
    log check_as "$@"
    cat > $TMPC
    log_file $TMPC
-    check_cmd $as $CPPFLAGS $ASFLAGS "$@" -c -o $TMPO $TMPC
+    check_cmd $as $CPPFLAGS $ASFLAGS "$@" $AS_C -o $TMPO $TMPC
 }
 check_asm(){
@ -1043,7 +1047,6 @@ PROGRAM_LIST="
 CONFIG_LIST="
    $COMPONENT_LIST
    $PROGRAM_LIST
    ac3dsp
    avcodec
    avdevice
    avfilter
@ -1061,7 +1064,6 @@ CONFIG_LIST="
    fft
    fontconfig
    frei0r
    gcrypt
    gnutls
    gpl
    gray
@ -1107,8 +1109,6 @@ CONFIG_LIST="
    mdct
    memalign_hack
    memory_poisoning
    mpegaudiodsp
    nettle
    network
    nonfree
    openal
@ -1116,11 +1116,9 @@ CONFIG_LIST="
    pic
    postproc
    rdft
    rtpdec
    runtime_cpudetect
    safe_bitstream_reader
    shared
    sinewin
    small
    sram
    static
@ -1341,7 +1339,9 @@ HAVE_LIST="
 # options emitted with CONFIG_ prefix but not available on command line
 CONFIG_EXTRA="
    aandcttables
    ac3dsp
    avutil
    gcrypt
    golomb
    gplv3
    h264chroma
@ -1351,6 +1351,10 @@ CONFIG_EXTRA="
    huffman
    lgplv3
    lpc
    mpegaudiodsp
    nettle
    rtpdec
    sinewin
    vp3dsp
 "
@ -1738,7 +1742,7 @@ v4l_indev_deps="linux_videodev_h"
 v4l2_indev_deps_any="linux_videodev2_h sys_videoio_h"
 vfwcap_indev_deps="capCreateCaptureWindow vfwcap_defines"
 vfwcap_indev_extralibs="-lavicap32"
-x11_grab_device_indev_deps="x11grab"
+x11grab_indev_deps="x11grab"
 # protocols
 bluray_protocol_deps="libbluray"
@ -1968,10 +1972,14 @@ asflags_filter=echo
 cflags_filter=echo
 ldflags_filter=echo
 AS_C='-c'
 AS_O='-o $@'
 CC_C='-c'
 CC_E='-E -o $@'
 CC_O='-o $@'
 CXX_O='-o $@'
 LD_O='-o $@'
 HOSTCC_C='-c'
 HOSTCC_O='-o $@'
 host_cflags='-D_ISOC99_SOURCE -D_XOPEN_SOURCE=600 -O3 -g'
@ -2331,7 +2339,8 @@ probe_cc(){
    pfx=$1
    _cc=$2
-    unset _type _ident _cc_o _flags _cflags _ldflags _depflags _DEPCMD _DEPFLAGS
+    unset _type _ident _cc_c _cc_e _cc_o _flags _cflags _ldflags
    unset _depflags _DEPCMD _DEPFLAGS
    _flags_filter=echo
    if $_cc -v 2>&1 | grep -q '^gcc.*LLVM'; then
@ -2394,6 +2403,7 @@ probe_cc(){
        _ident=$($_cc -version | head -n1 | tr -s ' ')
        _flags='--gcc --abi=eabi -me'
        _cflags='-D__gnuc_va_list=va_list -D__USER_LABEL_PREFIX__='
        _cc_e='-ppl -fe=$@'
        _cc_o='-fe=$@'
        as_default="${cross_prefix}gcc"
        ld_default="${cross_prefix}gcc"
@ -2445,6 +2455,8 @@ probe_cc(){
 }
 set_ccvars(){
    eval ${1}_C=\${_cc_c-\${${1}_C}}
    eval ${1}_E=\${_cc_e-\${${1}_E}}
    eval ${1}_O=\${_cc_o-\${${1}_O}}
    if [ -n "$_depflags" ]; then
@ -3874,7 +3886,7 @@ DEPCCFLAGS=$DEPCCFLAGS \$(CPPFLAGS)
 DEPAS=$as
 DEPASFLAGS=$DEPASFLAGS \$(CPPFLAGS)
 YASM=$yasmexe
-YASMDEP=$yasmexe
+DEPYASM=$yasmexe
 AR=$ar
 RANLIB=$ranlib
 CP=cp -p
@ -3884,7 +3896,9 @@ CPPFLAGS=$CPPFLAGS
 CFLAGS=$CFLAGS
 CXXFLAGS=$CXXFLAGS
 ASFLAGS=$ASFLAGS
 AS_C=$AS_C
 AS_O=$AS_O
 CC_C=$CC_C
 CC_O=$CC_O
 CXX_O=$CXX_O
 LD_O=$LD_O
@ -3919,6 +3933,7 @@ DEPHOSTCCFLAGS=$DEPHOSTCCFLAGS \$(HOSTCCFLAGS)
 HOSTCCDEP=$HOSTCCDEP
 HOSTCCDEP_FLAGS=$HOSTCCDEP_FLAGS
 HOSTCC_DEPFLAGS=$HOSTCC_DEPFLAGS
 HOSTCC_C=$HOSTCC_C
 HOSTCC_O=$HOSTCC_O
 TARGET_EXEC=$target_exec
 TARGET_PATH=$target_path
--- a/libavcodec/vc1dec.c
+++ b/libavcodec/vc1dec.c
@ -1050,9 +1050,10 @@ static void vc1_mc_4mv_chroma4(VC1Context *v)
        if ((edges&8) && s->mb_y == (s->mb_height - 1))        \
            mquant = v->altpq;                                 \
        if (!mquant || mquant > 31) {                          \
-            av_log(v->s.avctx, AV_LOG_ERROR, "invalid mquant %d\n", mquant);   \
+            av_log(v->s.avctx, AV_LOG_ERROR,                   \
-            mquant = 1;                                \
+                   "Overriding invalid mquant %d\n", mquant);  \
-        }                                              \
+            mquant = 1;                                        \
        }                                                      \
    }
 /**
@ -4944,15 +4945,17 @@ static void vc1_draw_sprites(VC1Context *v, SpriteData* sd)
                int      iline  = s->current_picture.f.linesize[plane];
                int      ycoord = yoff[sprite] + yadv[sprite] * row;
                int      yline  = ycoord >> 16;
                int      next_line;
                ysub[sprite] = ycoord & 0xFFFF;
                if (sprite) {
                    iplane = s->last_picture.f.data[plane];
                    iline  = s->last_picture.f.linesize[plane];
                }
                next_line = FFMIN(yline + 1, (v->sprite_height >> !!plane) - 1) * iline;
                if (!(xoff[sprite] & 0xFFFF) && xadv[sprite] == 1 << 16) {
                        src_h[sprite][0] = iplane + (xoff[sprite] >> 16) +  yline      * iline;
                    if (ysub[sprite])
-                        src_h[sprite][1] = iplane + (xoff[sprite] >> 16) + FFMIN(yline + 1, (v->sprite_height>>!!plane)-1) * iline;
+                        src_h[sprite][1] = iplane + (xoff[sprite] >> 16) + next_line;
                } else {
                    if (sr_cache[sprite][0] != yline) {
                        if (sr_cache[sprite][1] == yline) {
@ -4964,7 +4967,9 @@ static void vc1_draw_sprites(VC1Context *v, SpriteData* sd)
                        }
                    }
                    if (ysub[sprite] && sr_cache[sprite][1] != yline + 1) {
-                        v->vc1dsp.sprite_h(v->sr_rows[sprite][1], iplane + FFMIN(yline + 1, (v->sprite_height>>!!plane)-1) * iline, xoff[sprite], xadv[sprite], width);
+                        v->vc1dsp.sprite_h(v->sr_rows[sprite][1],
                                           iplane + next_line, xoff[sprite],
                                           xadv[sprite], width);
                        sr_cache[sprite][1] = yline + 1;
                    }
                    src_h[sprite][0] = v->sr_rows[sprite][0];
@ -5581,8 +5586,10 @@ static int vc1_decode_frame(AVCodecContext *avctx, void *data,
        mb_height = s->mb_height >> v->field_mode;
        for (i = 0; i <= n_slices; i++) {
            if (i > 0 &&  slices[i - 1].mby_start >= mb_height) {
-                if(v->field_mode <= 0) {
+                if (v->field_mode <= 0) {
-                    av_log(v->s.avctx, AV_LOG_ERROR, "invalid end_mb_y %d\n", slices[i - 1].mby_start);
+                    av_log(v->s.avctx, AV_LOG_ERROR, "Slice %d starts beyond "
                           "picture boundary (%d >= %d)\n", i,
                           slices[i - 1].mby_start, mb_height);
                    continue;
                }
                v->second_field = 1;
@ -5597,13 +5604,13 @@ static int vc1_decode_frame(AVCodecContext *avctx, void *data,
                v->pic_header_flag = 0;
                if (v->field_mode && i == n_slices1 + 2) {
                    if (ff_vc1_parse_frame_header_adv(v, &s->gb) < 0) {
-                        av_log(v->s.avctx, AV_LOG_ERROR, "slice header damaged\n");
+                        av_log(v->s.avctx, AV_LOG_ERROR, "Field header damaged\n");
                        continue;
                    }
                } else if (get_bits1(&s->gb)) {
                    v->pic_header_flag = 1;
                    if (ff_vc1_parse_frame_header_adv(v, &s->gb) < 0) {
-                        av_log(v->s.avctx, AV_LOG_ERROR, "slice header damaged\n");
+                        av_log(v->s.avctx, AV_LOG_ERROR, "Slice header damaged\n");
                        continue;
                    }
                }
--- a/libavcodec/x86/dsputil_mmx.c
+++ b/libavcodec/x86/dsputil_mmx.c
@ -2137,10 +2137,10 @@ void ff_ ## OP ## _h264_chroma_mc ## NUM ## _ ## DEPTH ## _ ## OPT      \
                                      (uint8_t *dst, uint8_t *src,      \
                                       int stride, int h, int x, int y);
-CHROMA_MC(put, 2, 10, mmxext)
+CHROMA_MC(put, 2, 10, mmx2)
-CHROMA_MC(avg, 2, 10, mmxext)
+CHROMA_MC(avg, 2, 10, mmx2)
-CHROMA_MC(put, 4, 10, mmxext)
+CHROMA_MC(put, 4, 10, mmx2)
-CHROMA_MC(avg, 4, 10, mmxext)
+CHROMA_MC(avg, 4, 10, mmx2)
 CHROMA_MC(put, 8, 10, sse2)
 CHROMA_MC(avg, 8, 10, sse2)
 CHROMA_MC(put, 8, 10, avx)
@ -2841,10 +2841,10 @@ static void dsputil_init_mmx2(DSPContext *c, AVCodecContext *avctx,
        c->put_h264_chroma_pixels_tab[2] = ff_put_h264_chroma_mc2_mmx2;
    }
    if (bit_depth == 10 && CONFIG_H264CHROMA) {
-        c->put_h264_chroma_pixels_tab[2] = ff_put_h264_chroma_mc2_10_mmxext;
+        c->put_h264_chroma_pixels_tab[2] = ff_put_h264_chroma_mc2_10_mmx2;
-        c->avg_h264_chroma_pixels_tab[2] = ff_avg_h264_chroma_mc2_10_mmxext;
+        c->avg_h264_chroma_pixels_tab[2] = ff_avg_h264_chroma_mc2_10_mmx2;
-        c->put_h264_chroma_pixels_tab[1] = ff_put_h264_chroma_mc4_10_mmxext;
+        c->put_h264_chroma_pixels_tab[1] = ff_put_h264_chroma_mc4_10_mmx2;
-        c->avg_h264_chroma_pixels_tab[1] = ff_avg_h264_chroma_mc4_10_mmxext;
+        c->avg_h264_chroma_pixels_tab[1] = ff_avg_h264_chroma_mc4_10_mmx2;
    }
    c->add_hfyu_median_prediction   = ff_add_hfyu_median_prediction_mmx2;
--- a/libavcodec/x86/fft_mmx.asm
+++ b/libavcodec/x86/fft_mmx.asm
@ -517,23 +517,23 @@ INIT_MMX 3dnow
 FFT48_3DN
-%define Z(x) [zq + o1q*(x&6) + mmsize*(x&1)]
+%define Z(x) [zcq + o1q*(x&6) + mmsize*(x&1)]
-%define Z2(x) [zq + o3q + mmsize*(x&1)]
+%define Z2(x) [zcq + o3q + mmsize*(x&1)]
-%define ZH(x) [zq + o1q*(x&6) + mmsize*(x&1) + mmsize/2]
+%define ZH(x) [zcq + o1q*(x&6) + mmsize*(x&1) + mmsize/2]
-%define Z2H(x) [zq + o3q + mmsize*(x&1) + mmsize/2]
+%define Z2H(x) [zcq + o3q + mmsize*(x&1) + mmsize/2]
 %macro DECL_PASS 2+ ; name, payload
 align 16
 %1:
-DEFINE_ARGS z, w, n, o1, o3
+DEFINE_ARGS zc, w, n, o1, o3
    lea o3q, [nq*3]
    lea o1q, [nq*8]
    shl o3q, 4
 .loop:
    %2
-    add zq, mmsize*2
+    add zcq, mmsize*2
-    add wq, mmsize
+    add  wq, mmsize
-    sub nd, mmsize/8
+    sub  nd, mmsize/8
    jg .loop
    rep ret
 %endmacro
@ -748,7 +748,7 @@ section .text
 ; On x86_32, this function does the register saving and restoring for all of fft.
 ; The others pass args in registers and don't spill anything.
-cglobal fft_dispatch%2, 2,5,8, z, nbits
+cglobal fft_dispatch%2, 2,5,8, zc, nbits
    FFT_DISPATCH fullsuffix, nbits
    RET
 %endmacro ; DECL_FFT
--- a/libavcodec/x86/h264_chromamc_10bit.asm
+++ b/libavcodec/x86/h264_chromamc_10bit.asm
@ -60,10 +60,10 @@ SECTION .text
 ;-----------------------------------------------------------------------------
 ; void put/avg_h264_chroma_mc8(pixel *dst, pixel *src, int stride, int h, int mx, int my)
 ;-----------------------------------------------------------------------------
-%macro CHROMA_MC8 2
+%macro CHROMA_MC8 1
 ; put/avg_h264_chroma_mc8_*(uint8_t *dst /*align 8*/, uint8_t *src /*align 1*/,
 ;                              int stride, int h, int mx, int my)
-cglobal %1_h264_chroma_mc8_10_%2, 6,7,8
+cglobal %1_h264_chroma_mc8_10, 6,7,8
    movsxdifnidn  r2, r2d
    mov          r6d, r5d
    or           r6d, r4d
@ -173,8 +173,8 @@ cglobal %1_h264_chroma_mc8_10_%2, 6,7,8
    add           r0, r2
 %endmacro
-%macro CHROMA_MC4 2
+%macro CHROMA_MC4 1
-cglobal %1_h264_chroma_mc4_10_%2, 6,6,7
+cglobal %1_h264_chroma_mc4_10, 6,6,7
    movsxdifnidn  r2, r2d
    movd          m2, r4m         ; x
    movd          m3, r5m         ; y
@ -203,8 +203,8 @@ cglobal %1_h264_chroma_mc4_10_%2, 6,6,7
 ;-----------------------------------------------------------------------------
 ; void put/avg_h264_chroma_mc2(pixel *dst, pixel *src, int stride, int h, int mx, int my)
 ;-----------------------------------------------------------------------------
-%macro CHROMA_MC2 2
+%macro CHROMA_MC2 1
-cglobal %1_h264_chroma_mc2_10_%2, 6,7
+cglobal %1_h264_chroma_mc2_10, 6,7
    movsxdifnidn  r2, r2d
    mov          r6d, r4d
    shl          r4d, 16
@ -250,24 +250,24 @@ cglobal %1_h264_chroma_mc2_10_%2, 6,7
 %endmacro
 %define CHROMAMC_AVG  NOTHING
-INIT_XMM
+INIT_XMM sse2
-CHROMA_MC8 put, sse2
+CHROMA_MC8 put
 %if HAVE_AVX
-INIT_AVX
+INIT_XMM avx
-CHROMA_MC8 put, avx
+CHROMA_MC8 put
 %endif
-INIT_MMX
+INIT_MMX mmx2
-CHROMA_MC4 put, mmxext
+CHROMA_MC4 put
-CHROMA_MC2 put, mmxext
+CHROMA_MC2 put
 %define CHROMAMC_AVG  AVG
 %define PAVG          pavgw
-INIT_XMM
+INIT_XMM sse2
-CHROMA_MC8 avg, sse2
+CHROMA_MC8 avg
 %if HAVE_AVX
-INIT_AVX
+INIT_XMM avx
-CHROMA_MC8 avg, avx
+CHROMA_MC8 avg
 %endif
-INIT_MMX
+INIT_MMX mmx2
-CHROMA_MC4 avg, mmxext
+CHROMA_MC4 avg
-CHROMA_MC2 avg, mmxext
+CHROMA_MC2 avg
--- a/libavcodec/x86/h264_deblock.asm
+++ b/libavcodec/x86/h264_deblock.asm
@ -282,8 +282,8 @@ cextern pb_A1
 ;-----------------------------------------------------------------------------
 ; void deblock_v_luma( uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0 )
 ;-----------------------------------------------------------------------------
-%macro DEBLOCK_LUMA 1
+%macro DEBLOCK_LUMA 0
-cglobal deblock_v_luma_8_%1, 5,5,10
+cglobal deblock_v_luma_8, 5,5,10
    movd    m8, [r4] ; tc0
    lea     r4, [r1*3]
    dec     r2d        ; alpha-1
@ -327,8 +327,8 @@ cglobal deblock_v_luma_8_%1, 5,5,10
 ;-----------------------------------------------------------------------------
 ; void deblock_h_luma( uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0 )
 ;-----------------------------------------------------------------------------
-INIT_MMX
+INIT_MMX cpuname
-cglobal deblock_h_luma_8_%1, 5,9
+cglobal deblock_h_luma_8, 5,9
    movsxd r7,  r1d
    lea    r8,  [r7+r7*2]
    lea    r6,  [r0-4]
@ -355,7 +355,7 @@ cglobal deblock_h_luma_8_%1, 5,9
 %if WIN64
    mov    [rsp+0x20], r4
 %endif
-    call   deblock_v_luma_8_%1
+    call   deblock_v_luma_8
    ; transpose 16x4 -> original space  (only the middle 4 rows were changed by the filter)
    add    r6, 2
@ -384,26 +384,26 @@ cglobal deblock_h_luma_8_%1, 5,9
    RET
 %endmacro
-INIT_XMM
+INIT_XMM sse2
-DEBLOCK_LUMA sse2
+DEBLOCK_LUMA
 %if HAVE_AVX
-INIT_AVX
+INIT_XMM avx
-DEBLOCK_LUMA avx
+DEBLOCK_LUMA
 %endif
 %else
-%macro DEBLOCK_LUMA 3
+%macro DEBLOCK_LUMA 2
 ;-----------------------------------------------------------------------------
 ; void deblock_v8_luma( uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0 )
 ;-----------------------------------------------------------------------------
-cglobal deblock_%2_luma_8_%1, 5,5
+cglobal deblock_%1_luma_8, 5,5
    lea     r4, [r1*3]
    dec     r2     ; alpha-1
    neg     r4
    dec     r3     ; beta-1
    add     r4, r0 ; pix-3*stride
-    %assign pad 2*%3+12-(stack_offset&15)
+    %assign pad 2*%2+12-(stack_offset&15)
    SUB     esp, pad
    mova    m0, [r4+r1]   ; p1
@ -417,7 +417,7 @@ cglobal deblock_%2_luma_8_%1, 5,5
    movd    m4, [r3] ; tc0
    punpcklbw m4, m4
    punpcklbw m4, m4 ; tc = 4x tc0[3], 4x tc0[2], 4x tc0[1], 4x tc0[0]
-    mova   [esp+%3], m4 ; tc
+    mova   [esp+%2], m4 ; tc
    pcmpgtb m4, m3
    mova    m3, [r4] ; p2
    pand    m4, m7
@ -425,7 +425,7 @@ cglobal deblock_%2_luma_8_%1, 5,5
    DIFF_GT2 m1, m3, m5, m6, m7 ; |p2-p0| > beta-1
    pand    m6, m4
-    pand    m4, [esp+%3] ; tc
+    pand    m4, [esp+%2] ; tc
    psubb   m7, m4, m6
    pand    m6, m4
    LUMA_Q1 m0, m3, [r4], [r4+r1], m6, m4
@ -433,7 +433,7 @@ cglobal deblock_%2_luma_8_%1, 5,5
    mova    m4, [r0+2*r1] ; q2
    DIFF_GT2 m2, m4, m5, m6, m3 ; |q2-q0| > beta-1
    pand    m6, [esp] ; mask
-    mova    m5, [esp+%3] ; tc
+    mova    m5, [esp+%2] ; tc
    psubb   m7, m6
    pand    m5, m6
    mova    m3, [r0+r1]
@ -448,8 +448,8 @@ cglobal deblock_%2_luma_8_%1, 5,5
 ;-----------------------------------------------------------------------------
 ; void deblock_h_luma( uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0 )
 ;-----------------------------------------------------------------------------
-INIT_MMX
+INIT_MMX cpuname
-cglobal deblock_h_luma_8_%1, 0,5
+cglobal deblock_h_luma_8, 0,5
    mov    r0, r0mp
    mov    r3, r1m
    lea    r4, [r3*3]
@ -472,11 +472,11 @@ cglobal deblock_h_luma_8_%1, 0,5
    PUSH   dword r2m
    PUSH   dword 16
    PUSH   dword r0
-    call   deblock_%2_luma_8_%1
+    call   deblock_%1_luma_8
-%ifidn %2, v8
+%ifidn %1, v8
    add    dword [esp   ], 8 ; pix_tmp+0x38
    add    dword [esp+16], 2 ; tc0+2
-    call   deblock_%2_luma_8_%1
+    call   deblock_%1_luma_8
 %endif
    ADD    esp, 20
@ -503,13 +503,13 @@ cglobal deblock_h_luma_8_%1, 0,5
    RET
 %endmacro ; DEBLOCK_LUMA
-INIT_MMX
+INIT_MMX mmx2
-DEBLOCK_LUMA mmxext, v8, 8
+DEBLOCK_LUMA v8, 8
-INIT_XMM
+INIT_XMM sse2
-DEBLOCK_LUMA sse2, v, 16
+DEBLOCK_LUMA v, 16
 %if HAVE_AVX
-INIT_AVX
+INIT_XMM avx
-DEBLOCK_LUMA avx, v, 16
+DEBLOCK_LUMA v, 16
 %endif
 %endif ; ARCH
@ -612,7 +612,7 @@ DEBLOCK_LUMA avx, v, 16
    %define mask1p mask1q
 %endmacro
-%macro DEBLOCK_LUMA_INTRA 2
+%macro DEBLOCK_LUMA_INTRA 1
    %define p1 m0
    %define p0 m1
    %define q0 m2
@ -647,7 +647,7 @@ DEBLOCK_LUMA avx, v, 16
 ;-----------------------------------------------------------------------------
 ; void deblock_v_luma_intra( uint8_t *pix, int stride, int alpha, int beta )
 ;-----------------------------------------------------------------------------
-cglobal deblock_%2_luma_intra_8_%1, 4,6,16
+cglobal deblock_%1_luma_intra_8, 4,6,16
 %if ARCH_X86_64 == 0
    sub     esp, 0x60
 %endif
@ -704,12 +704,12 @@ cglobal deblock_%2_luma_intra_8_%1, 4,6,16
 %endif
    RET
-INIT_MMX
+INIT_MMX cpuname
 %if ARCH_X86_64
 ;-----------------------------------------------------------------------------
 ; void deblock_h_luma_intra( uint8_t *pix, int stride, int alpha, int beta )
 ;-----------------------------------------------------------------------------
-cglobal deblock_h_luma_intra_8_%1, 4,9
+cglobal deblock_h_luma_intra_8, 4,9
    movsxd r7,  r1d
    lea    r8,  [r7*3]
    lea    r6,  [r0-4]
@ -725,7 +725,7 @@ cglobal deblock_h_luma_intra_8_%1, 4,9
    lea    r0,  [pix_tmp+0x40]
    mov    r1,  0x10
-    call   deblock_v_luma_intra_8_%1
+    call   deblock_v_luma_intra_8
    ; transpose 16x6 -> original space (but we can't write only 6 pixels, so really 16x8)
    lea    r5, [r6+r8]
@ -738,7 +738,7 @@ cglobal deblock_h_luma_intra_8_%1, 4,9
    add    rsp, 0x88
    RET
 %else
-cglobal deblock_h_luma_intra_8_%1, 2,4
+cglobal deblock_h_luma_intra_8, 2,4
    lea    r3,  [r1*3]
    sub    r0,  4
    lea    r2,  [r0+r3]
@ -757,10 +757,10 @@ cglobal deblock_h_luma_intra_8_%1, 2,4
    PUSH   dword r2m
    PUSH   dword 16
    PUSH   r0
-    call   deblock_%2_luma_intra_8_%1
+    call   deblock_%1_luma_intra_8
-%ifidn %2, v8
+%ifidn %1, v8
    add    dword [rsp], 8 ; pix_tmp+8
-    call   deblock_%2_luma_intra_8_%1
+    call   deblock_%1_luma_intra_8
 %endif
    ADD    esp, 16
@ -779,18 +779,18 @@ cglobal deblock_h_luma_intra_8_%1, 2,4
 %endif ; ARCH_X86_64
 %endmacro ; DEBLOCK_LUMA_INTRA
-INIT_XMM
+INIT_XMM sse2
-DEBLOCK_LUMA_INTRA sse2, v
+DEBLOCK_LUMA_INTRA v
 %if HAVE_AVX
-INIT_AVX
+INIT_XMM avx
-DEBLOCK_LUMA_INTRA avx , v
+DEBLOCK_LUMA_INTRA v
 %endif
 %if ARCH_X86_64 == 0
-INIT_MMX
+INIT_MMX mmx2
-DEBLOCK_LUMA_INTRA mmxext, v8
+DEBLOCK_LUMA_INTRA v8
 %endif
-INIT_MMX
+INIT_MMX mmx2
 %macro CHROMA_V_START 0
    dec    r2d      ; alpha-1
@ -815,13 +815,13 @@ INIT_MMX
 ;-----------------------------------------------------------------------------
 ; void ff_deblock_v_chroma( uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0 )
 ;-----------------------------------------------------------------------------
-cglobal deblock_v_chroma_8_mmxext, 5,6
+cglobal deblock_v_chroma_8, 5,6
    CHROMA_V_START
    movq  m0, [t5]
    movq  m1, [t5+r1]
    movq  m2, [r0]
    movq  m3, [r0+r1]
-    call ff_chroma_inter_body_mmxext
+    call ff_chroma_inter_body_mmx2
    movq  [t5+r1], m1
    movq  [r0], m2
    RET
@ -829,7 +829,7 @@ cglobal deblock_v_chroma_8_mmxext, 5,6
 ;-----------------------------------------------------------------------------
 ; void ff_deblock_h_chroma( uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0 )
 ;-----------------------------------------------------------------------------
-cglobal deblock_h_chroma_8_mmxext, 5,7
+cglobal deblock_h_chroma_8, 5,7
 %if UNIX64
    %define buf0 [rsp-24]
    %define buf1 [rsp-16]
@ -859,7 +859,7 @@ cglobal deblock_h_chroma_8_mmxext, 5,7
    RET
 ALIGN 16
-ff_chroma_inter_body_mmxext:
+ff_chroma_inter_body_mmx2:
    LOAD_MASK  r2d, r3d
    movd       m6, [r4] ; tc0
    punpcklbw  m6, m6
@ -886,13 +886,13 @@ ff_chroma_inter_body_mmxext:
 ;-----------------------------------------------------------------------------
 ; void ff_deblock_v_chroma_intra( uint8_t *pix, int stride, int alpha, int beta )
 ;-----------------------------------------------------------------------------
-cglobal deblock_v_chroma_intra_8_mmxext, 4,5
+cglobal deblock_v_chroma_intra_8, 4,5
    CHROMA_V_START
    movq  m0, [t5]
    movq  m1, [t5+r1]
    movq  m2, [r0]
    movq  m3, [r0+r1]
-    call ff_chroma_intra_body_mmxext
+    call ff_chroma_intra_body_mmx2
    movq  [t5+r1], m1
    movq  [r0], m2
    RET
@ -900,15 +900,15 @@ cglobal deblock_v_chroma_intra_8_mmxext, 4,5
 ;-----------------------------------------------------------------------------
 ; void ff_deblock_h_chroma_intra( uint8_t *pix, int stride, int alpha, int beta )
 ;-----------------------------------------------------------------------------
-cglobal deblock_h_chroma_intra_8_mmxext, 4,6
+cglobal deblock_h_chroma_intra_8, 4,6
    CHROMA_H_START
    TRANSPOSE4x8_LOAD  bw, wd, dq, PASS8ROWS(t5, r0, r1, t6)
-    call ff_chroma_intra_body_mmxext
+    call ff_chroma_intra_body_mmx2
    TRANSPOSE8x4B_STORE PASS8ROWS(t5, r0, r1, t6)
    RET
 ALIGN 16
-ff_chroma_intra_body_mmxext:
+ff_chroma_intra_body_mmx2:
    LOAD_MASK r2d, r3d
    movq   m5, m1
    movq   m6, m2
--- a/libavcodec/x86/h264_deblock_10bit.asm
+++ b/libavcodec/x86/h264_deblock_10bit.asm
@ -151,11 +151,11 @@ cextern pw_4
 %endif
 %endmacro
-%macro DEBLOCK_LUMA 1
+%macro DEBLOCK_LUMA 0
 ;-----------------------------------------------------------------------------
 ; void deblock_v_luma( uint16_t *pix, int stride, int alpha, int beta, int8_t *tc0 )
 ;-----------------------------------------------------------------------------
-cglobal deblock_v_luma_10_%1, 5,5,8*(mmsize/16)
+cglobal deblock_v_luma_10, 5,5,8*(mmsize/16)
    %assign pad 5*mmsize+12-(stack_offset&15)
    %define tcm [rsp]
    %define ms1 [rsp+mmsize]
@ -210,7 +210,7 @@ cglobal deblock_v_luma_10_%1, 5,5,8*(mmsize/16)
    ADD         rsp, pad
    RET
-cglobal deblock_h_luma_10_%1, 5,6,8*(mmsize/16)
+cglobal deblock_h_luma_10, 5,6,8*(mmsize/16)
    %assign pad 7*mmsize+12-(stack_offset&15)
    %define tcm [rsp]
    %define ms1 [rsp+mmsize]
@ -301,7 +301,6 @@ cglobal deblock_h_luma_10_%1, 5,6,8*(mmsize/16)
    RET
 %endmacro
 INIT_XMM
 %if ARCH_X86_64
 ; in:  m0=p1, m1=p0, m2=q0, m3=q1, m8=p2, m9=q2
 ;      m12=alpha, m13=beta
@ -339,8 +338,8 @@ INIT_XMM
    SWAP         3, 9
 %endmacro
-%macro DEBLOCK_LUMA_64 1
+%macro DEBLOCK_LUMA_64 0
-cglobal deblock_v_luma_10_%1, 5,5,15
+cglobal deblock_v_luma_10, 5,5,15
    %define p2 m8
    %define p1 m0
    %define p0 m1
@ -377,7 +376,7 @@ cglobal deblock_v_luma_10_%1, 5,5,15
    jg .loop
    REP_RET
-cglobal deblock_h_luma_10_%1, 5,7,15
+cglobal deblock_h_luma_10, 5,7,15
    shl        r2d, 2
    shl        r3d, 2
    LOAD_AB    m12, m13, r2d, r3d
@ -417,11 +416,11 @@ cglobal deblock_h_luma_10_%1, 5,7,15
    REP_RET
 %endmacro
-INIT_XMM
+INIT_XMM sse2
-DEBLOCK_LUMA_64 sse2
+DEBLOCK_LUMA_64
 %if HAVE_AVX
-INIT_AVX
+INIT_XMM avx
-DEBLOCK_LUMA_64 avx
+DEBLOCK_LUMA_64
 %endif
 %endif
@ -604,8 +603,8 @@ DEBLOCK_LUMA_64 avx
 ;-----------------------------------------------------------------------------
 ; void deblock_v_luma_intra( uint16_t *pix, int stride, int alpha, int beta )
 ;-----------------------------------------------------------------------------
-%macro DEBLOCK_LUMA_INTRA_64 1
+%macro DEBLOCK_LUMA_INTRA_64 0
-cglobal deblock_v_luma_intra_10_%1, 4,7,16
+cglobal deblock_v_luma_intra_10, 4,7,16
    %define t0 m1
    %define t1 m2
    %define t2 m4
@ -655,7 +654,7 @@ cglobal deblock_v_luma_intra_10_%1, 4,7,16
 ;-----------------------------------------------------------------------------
 ; void deblock_h_luma_intra( uint16_t *pix, int stride, int alpha, int beta )
 ;-----------------------------------------------------------------------------
-cglobal deblock_h_luma_intra_10_%1, 4,7,16
+cglobal deblock_h_luma_intra_10, 4,7,16
    %define t0 m15
    %define t1 m14
    %define t2 m2
@ -714,20 +713,20 @@ cglobal deblock_h_luma_intra_10_%1, 4,7,16
    RET
 %endmacro
-INIT_XMM
+INIT_XMM sse2
-DEBLOCK_LUMA_INTRA_64 sse2
+DEBLOCK_LUMA_INTRA_64
 %if HAVE_AVX
-INIT_AVX
+INIT_XMM avx
-DEBLOCK_LUMA_INTRA_64 avx
+DEBLOCK_LUMA_INTRA_64
 %endif
 %endif
-%macro DEBLOCK_LUMA_INTRA 1
+%macro DEBLOCK_LUMA_INTRA 0
 ;-----------------------------------------------------------------------------
 ; void deblock_v_luma_intra( uint16_t *pix, int stride, int alpha, int beta )
 ;-----------------------------------------------------------------------------
-cglobal deblock_v_luma_intra_10_%1, 4,7,8*(mmsize/16)
+cglobal deblock_v_luma_intra_10, 4,7,8*(mmsize/16)
    LUMA_INTRA_INIT 3
    lea     r4, [r1*4]
    lea     r5, [r1*3]
@ -755,7 +754,7 @@ cglobal deblock_v_luma_intra_10_%1, 4,7,8*(mmsize/16)
 ;-----------------------------------------------------------------------------
 ; void deblock_h_luma_intra( uint16_t *pix, int stride, int alpha, int beta )
 ;-----------------------------------------------------------------------------
-cglobal deblock_h_luma_intra_10_%1, 4,7,8*(mmsize/16)
+cglobal deblock_h_luma_intra_10, 4,7,8*(mmsize/16)
    LUMA_INTRA_INIT 8
 %if mmsize == 8
    lea     r4, [r1*3]
@ -797,16 +796,16 @@ cglobal deblock_h_luma_intra_10_%1, 4,7,8*(mmsize/16)
 %endmacro
 %if ARCH_X86_64 == 0
-INIT_MMX
+INIT_MMX mmx2
-DEBLOCK_LUMA mmxext
+DEBLOCK_LUMA
-DEBLOCK_LUMA_INTRA mmxext
+DEBLOCK_LUMA_INTRA
-INIT_XMM
+INIT_XMM sse2
-DEBLOCK_LUMA sse2
+DEBLOCK_LUMA
-DEBLOCK_LUMA_INTRA sse2
+DEBLOCK_LUMA_INTRA
 %if HAVE_AVX
-INIT_AVX
+INIT_XMM avx
-DEBLOCK_LUMA avx
+DEBLOCK_LUMA
-DEBLOCK_LUMA_INTRA avx
+DEBLOCK_LUMA_INTRA
 %endif
 %endif
@ -849,11 +848,11 @@ DEBLOCK_LUMA_INTRA avx
    psraw       %1, 6
 %endmacro
-%macro DEBLOCK_CHROMA 1
+%macro DEBLOCK_CHROMA 0
 ;-----------------------------------------------------------------------------
 ; void deblock_v_chroma( uint16_t *pix, int stride, int alpha, int beta, int8_t *tc0 )
 ;-----------------------------------------------------------------------------
-cglobal deblock_v_chroma_10_%1, 5,7-(mmsize/16),8*(mmsize/16)
+cglobal deblock_v_chroma_10, 5,7-(mmsize/16),8*(mmsize/16)
    mov         r5, r0
    sub         r0, r1
    sub         r0, r1
@ -887,7 +886,7 @@ cglobal deblock_v_chroma_10_%1, 5,7-(mmsize/16),8*(mmsize/16)
 ;-----------------------------------------------------------------------------
 ; void deblock_v_chroma_intra( uint16_t *pix, int stride, int alpha, int beta )
 ;-----------------------------------------------------------------------------
-cglobal deblock_v_chroma_intra_10_%1, 4,6-(mmsize/16),8*(mmsize/16)
+cglobal deblock_v_chroma_intra_10, 4,6-(mmsize/16),8*(mmsize/16)
    mov         r4, r0
    sub         r0, r1
    sub         r0, r1
@ -914,12 +913,12 @@ cglobal deblock_v_chroma_intra_10_%1, 4,6-(mmsize/16),8*(mmsize/16)
 %endmacro
 %if ARCH_X86_64 == 0
-INIT_MMX
+INIT_MMX mmx2
-DEBLOCK_CHROMA mmxext
+DEBLOCK_CHROMA
 %endif
-INIT_XMM
+INIT_XMM sse2
-DEBLOCK_CHROMA sse2
+DEBLOCK_CHROMA
 %if HAVE_AVX
-INIT_AVX
+INIT_XMM avx
-DEBLOCK_CHROMA avx
+DEBLOCK_CHROMA
 %endif
--- a/libavcodec/x86/h264_idct_10bit.asm
+++ b/libavcodec/x86/h264_idct_10bit.asm
@ -72,25 +72,25 @@ SECTION .text
    STORE_DIFFx2 m2, m3, m4, m5, %1, %3
 %endmacro
-%macro IDCT_ADD_10 1
+%macro IDCT_ADD_10 0
-cglobal h264_idct_add_10_%1, 3,3
+cglobal h264_idct_add_10, 3,3
    IDCT4_ADD_10 r0, r1, r2
    RET
 %endmacro
-INIT_XMM
+INIT_XMM sse2
-IDCT_ADD_10 sse2
+IDCT_ADD_10
 %if HAVE_AVX
-INIT_AVX
+INIT_XMM avx
-IDCT_ADD_10 avx
+IDCT_ADD_10
 %endif
 ;-----------------------------------------------------------------------------
 ; h264_idct_add16(pixel *dst, const int *block_offset, dctcoef *block, int stride, const uint8_t nnzc[6*8])
 ;-----------------------------------------------------------------------------
 ;;;;;;; NO FATE SAMPLES TRIGGER THIS
-%macro ADD4x4IDCT 1
+%macro ADD4x4IDCT 0
-add4x4_idct_%1:
+add4x4_idct %+ SUFFIX:
    add   r5, r0
    mova  m0, [r2+ 0]
    mova  m1, [r2+16]
@ -107,52 +107,52 @@ add4x4_idct_%1:
    ret
 %endmacro
-INIT_XMM
+INIT_XMM sse2
 ALIGN 16
-ADD4x4IDCT sse2
+ADD4x4IDCT
 %if HAVE_AVX
-INIT_AVX
+INIT_XMM avx
 ALIGN 16
-ADD4x4IDCT avx
+ADD4x4IDCT
 %endif
-%macro ADD16_OP 3
+%macro ADD16_OP 2
-    cmp          byte [r4+%3], 0
+    cmp          byte [r4+%2], 0
-    jz .skipblock%2
+    jz .skipblock%1
-    mov         r5d, [r1+%2*4]
+    mov         r5d, [r1+%1*4]
-    call add4x4_idct_%1
+    call add4x4_idct %+ SUFFIX
-.skipblock%2:
+.skipblock%1:
-%if %2<15
+%if %1<15
    add          r2, 64
 %endif
 %endmacro
-%macro IDCT_ADD16_10 1
+%macro IDCT_ADD16_10 0
-cglobal h264_idct_add16_10_%1, 5,6
+cglobal h264_idct_add16_10, 5,6
-    ADD16_OP %1, 0, 4+1*8
+    ADD16_OP 0, 4+1*8
-    ADD16_OP %1, 1, 5+1*8
+    ADD16_OP 1, 5+1*8
-    ADD16_OP %1, 2, 4+2*8
+    ADD16_OP 2, 4+2*8
-    ADD16_OP %1, 3, 5+2*8
+    ADD16_OP 3, 5+2*8
-    ADD16_OP %1, 4, 6+1*8
+    ADD16_OP 4, 6+1*8
-    ADD16_OP %1, 5, 7+1*8
+    ADD16_OP 5, 7+1*8
-    ADD16_OP %1, 6, 6+2*8
+    ADD16_OP 6, 6+2*8
-    ADD16_OP %1, 7, 7+2*8
+    ADD16_OP 7, 7+2*8
-    ADD16_OP %1, 8, 4+3*8
+    ADD16_OP 8, 4+3*8
-    ADD16_OP %1, 9, 5+3*8
+    ADD16_OP 9, 5+3*8
-    ADD16_OP %1, 10, 4+4*8
+    ADD16_OP 10, 4+4*8
-    ADD16_OP %1, 11, 5+4*8
+    ADD16_OP 11, 5+4*8
-    ADD16_OP %1, 12, 6+3*8
+    ADD16_OP 12, 6+3*8
-    ADD16_OP %1, 13, 7+3*8
+    ADD16_OP 13, 7+3*8
-    ADD16_OP %1, 14, 6+4*8
+    ADD16_OP 14, 6+4*8
-    ADD16_OP %1, 15, 7+4*8
+    ADD16_OP 15, 7+4*8
    REP_RET
 %endmacro
-INIT_XMM
+INIT_XMM sse2
-IDCT_ADD16_10 sse2
+IDCT_ADD16_10
 %if HAVE_AVX
-INIT_AVX
+INIT_XMM avx
-IDCT_ADD16_10 avx
+IDCT_ADD16_10
 %endif
 ;-----------------------------------------------------------------------------
@ -185,8 +185,8 @@ IDCT_ADD16_10 avx
    mova [%1+%3  ], m4
 %endmacro
-INIT_MMX
+INIT_MMX mmx2
-cglobal h264_idct_dc_add_10_mmx2,3,3
+cglobal h264_idct_dc_add_10,3,3
    movd      m0, [r1]
    paddd     m0, [pd_32]
    psrad     m0, 6
@ -199,8 +199,8 @@ cglobal h264_idct_dc_add_10_mmx2,3,3
 ;-----------------------------------------------------------------------------
 ; void h264_idct8_dc_add(pixel *dst, dctcoef *block, int stride)
 ;-----------------------------------------------------------------------------
-%macro IDCT8_DC_ADD 1
+%macro IDCT8_DC_ADD 0
-cglobal h264_idct8_dc_add_10_%1,3,3,7
+cglobal h264_idct8_dc_add_10,3,3,7
    mov      r1d, [r1]
    add       r1, 32
    sar       r1, 6
@ -214,45 +214,45 @@ cglobal h264_idct8_dc_add_10_%1,3,3,7
    RET
 %endmacro
-INIT_XMM
+INIT_XMM sse2
-IDCT8_DC_ADD sse2
+IDCT8_DC_ADD
 %if HAVE_AVX
-INIT_AVX
+INIT_XMM avx
-IDCT8_DC_ADD avx
+IDCT8_DC_ADD
 %endif
 ;-----------------------------------------------------------------------------
 ; h264_idct_add16intra(pixel *dst, const int *block_offset, dctcoef *block, int stride, const uint8_t nnzc[6*8])
 ;-----------------------------------------------------------------------------
-%macro AC 2
+%macro AC 1
-.ac%2
+.ac%1
-    mov  r5d, [r1+(%2+0)*4]
+    mov  r5d, [r1+(%1+0)*4]
-    call add4x4_idct_%1
+    call add4x4_idct %+ SUFFIX
-    mov  r5d, [r1+(%2+1)*4]
+    mov  r5d, [r1+(%1+1)*4]
    add  r2, 64
-    call add4x4_idct_%1
+    call add4x4_idct %+ SUFFIX
    add  r2, 64
-    jmp .skipadd%2
+    jmp .skipadd%1
 %endmacro
 %assign last_block 16
-%macro ADD16_OP_INTRA 3
+%macro ADD16_OP_INTRA 2
-    cmp      word [r4+%3], 0
+    cmp      word [r4+%2], 0
-    jnz .ac%2
+    jnz .ac%1
    mov      r5d, [r2+ 0]
    or       r5d, [r2+64]
-    jz .skipblock%2
+    jz .skipblock%1
-    mov      r5d, [r1+(%2+0)*4]
+    mov      r5d, [r1+(%1+0)*4]
-    call idct_dc_add_%1
+    call idct_dc_add %+ SUFFIX
-.skipblock%2:
+.skipblock%1:
-%if %2<last_block-2
+%if %1<last_block-2
    add       r2, 128
 %endif
-.skipadd%2:
+.skipadd%1:
 %endmacro
-%macro IDCT_ADD16INTRA_10 1
+%macro IDCT_ADD16INTRA_10 0
-idct_dc_add_%1:
+idct_dc_add %+ SUFFIX:
    add       r5, r0
    movq      m0, [r2+ 0]
    movhps    m0, [r2+64]
@ -265,46 +265,46 @@ idct_dc_add_%1:
    IDCT_DC_ADD_OP_10 r5, r3, r6
    ret
-cglobal h264_idct_add16intra_10_%1,5,7,8
+cglobal h264_idct_add16intra_10,5,7,8
-    ADD16_OP_INTRA %1, 0, 4+1*8
+    ADD16_OP_INTRA 0, 4+1*8
-    ADD16_OP_INTRA %1, 2, 4+2*8
+    ADD16_OP_INTRA 2, 4+2*8
-    ADD16_OP_INTRA %1, 4, 6+1*8
+    ADD16_OP_INTRA 4, 6+1*8
-    ADD16_OP_INTRA %1, 6, 6+2*8
+    ADD16_OP_INTRA 6, 6+2*8
-    ADD16_OP_INTRA %1, 8, 4+3*8
+    ADD16_OP_INTRA 8, 4+3*8
-    ADD16_OP_INTRA %1, 10, 4+4*8
+    ADD16_OP_INTRA 10, 4+4*8
-    ADD16_OP_INTRA %1, 12, 6+3*8
+    ADD16_OP_INTRA 12, 6+3*8
-    ADD16_OP_INTRA %1, 14, 6+4*8
+    ADD16_OP_INTRA 14, 6+4*8
    REP_RET
-    AC %1, 8
+    AC 8
-    AC %1, 10
+    AC 10
-    AC %1, 12
+    AC 12
-    AC %1, 14
+    AC 14
-    AC %1, 0
+    AC 0
-    AC %1, 2
+    AC 2
-    AC %1, 4
+    AC 4
-    AC %1, 6
+    AC 6
 %endmacro
-INIT_XMM
+INIT_XMM sse2
-IDCT_ADD16INTRA_10 sse2
+IDCT_ADD16INTRA_10
 %if HAVE_AVX
-INIT_AVX
+INIT_XMM avx
-IDCT_ADD16INTRA_10 avx
+IDCT_ADD16INTRA_10
 %endif
 %assign last_block 36
 ;-----------------------------------------------------------------------------
 ; h264_idct_add8(pixel **dst, const int *block_offset, dctcoef *block, int stride, const uint8_t nnzc[6*8])
 ;-----------------------------------------------------------------------------
-%macro IDCT_ADD8 1
+%macro IDCT_ADD8 0
-cglobal h264_idct_add8_10_%1,5,8,7
+cglobal h264_idct_add8_10,5,8,7
 %if ARCH_X86_64
    mov      r7, r0
 %endif
    add      r2, 1024
    mov      r0, [r0]
-    ADD16_OP_INTRA %1, 16, 4+ 6*8
+    ADD16_OP_INTRA 16, 4+ 6*8
-    ADD16_OP_INTRA %1, 18, 4+ 7*8
+    ADD16_OP_INTRA 18, 4+ 7*8
    add      r2, 1024-128*2
 %if ARCH_X86_64
    mov      r0, [r7+gprsize]
@ -312,21 +312,21 @@ cglobal h264_idct_add8_10_%1,5,8,7
    mov      r0, r0m
    mov      r0, [r0+gprsize]
 %endif
-    ADD16_OP_INTRA %1, 32, 4+11*8
+    ADD16_OP_INTRA 32, 4+11*8
-    ADD16_OP_INTRA %1, 34, 4+12*8
+    ADD16_OP_INTRA 34, 4+12*8
    REP_RET
-    AC %1, 16
+    AC 16
-    AC %1, 18
+    AC 18
-    AC %1, 32
+    AC 32
-    AC %1, 34
+    AC 34
 %endmacro ; IDCT_ADD8
-INIT_XMM
+INIT_XMM sse2
-IDCT_ADD8 sse2
+IDCT_ADD8
 %if HAVE_AVX
-INIT_AVX
+INIT_XMM avx
-IDCT_ADD8 avx
+IDCT_ADD8
 %endif
 ;-----------------------------------------------------------------------------
@ -432,19 +432,19 @@ IDCT_ADD8 avx
    STORE_DIFFx2 m0, m1, m6, m7, %1, %3
 %endmacro
-%macro IDCT8_ADD 1
+%macro IDCT8_ADD 0
-cglobal h264_idct8_add_10_%1, 3,4,16
+cglobal h264_idct8_add_10, 3,4,16
 %if UNIX64 == 0
    %assign pad 16-gprsize-(stack_offset&15)
    sub  rsp, pad
-    call h264_idct8_add1_10_%1
+    call h264_idct8_add1_10 %+ SUFFIX
    add  rsp, pad
    RET
 %endif
 ALIGN 16
 ; TODO: does not need to use stack
-h264_idct8_add1_10_%1:
+h264_idct8_add1_10 %+ SUFFIX:
 %assign pad 256+16-gprsize
    sub          rsp, pad
    add   dword [r1], 32
@ -499,31 +499,31 @@ h264_idct8_add1_10_%1:
    ret
 %endmacro
-INIT_XMM
+INIT_XMM sse2
-IDCT8_ADD sse2
+IDCT8_ADD
 %if HAVE_AVX
-INIT_AVX
+INIT_XMM avx
-IDCT8_ADD avx
+IDCT8_ADD
 %endif
 ;-----------------------------------------------------------------------------
 ; h264_idct8_add4(pixel **dst, const int *block_offset, dctcoef *block, int stride, const uint8_t nnzc[6*8])
 ;-----------------------------------------------------------------------------
 ;;;;;;; NO FATE SAMPLES TRIGGER THIS
-%macro IDCT8_ADD4_OP 3
+%macro IDCT8_ADD4_OP 2
-    cmp       byte [r4+%3], 0
+    cmp       byte [r4+%2], 0
-    jz .skipblock%2
+    jz .skipblock%1
-    mov      r0d, [r6+%2*4]
+    mov      r0d, [r6+%1*4]
    add       r0, r5
-    call h264_idct8_add1_10_%1
+    call h264_idct8_add1_10 %+ SUFFIX
-.skipblock%2:
+.skipblock%1:
-%if %2<12
+%if %1<12
    add       r1, 256
 %endif
 %endmacro
-%macro IDCT8_ADD4 1
+%macro IDCT8_ADD4 0
-cglobal h264_idct8_add4_10_%1, 0,7,16
+cglobal h264_idct8_add4_10, 0,7,16
    %assign pad 16-gprsize-(stack_offset&15)
    SUB      rsp, pad
    mov       r5, r0mp
@ -531,17 +531,17 @@ cglobal h264_idct8_add4_10_%1, 0,7,16
    mov       r1, r2mp
    mov      r2d, r3m
    movifnidn r4, r4mp
-    IDCT8_ADD4_OP %1,  0, 4+1*8
+    IDCT8_ADD4_OP  0, 4+1*8
-    IDCT8_ADD4_OP %1,  4, 6+1*8
+    IDCT8_ADD4_OP  4, 6+1*8
-    IDCT8_ADD4_OP %1,  8, 4+3*8
+    IDCT8_ADD4_OP  8, 4+3*8
-    IDCT8_ADD4_OP %1, 12, 6+3*8
+    IDCT8_ADD4_OP 12, 6+3*8
    ADD       rsp, pad
    RET
 %endmacro ; IDCT8_ADD4
-INIT_XMM
+INIT_XMM sse2
-IDCT8_ADD4 sse2
+IDCT8_ADD4
 %if HAVE_AVX
-INIT_AVX
+INIT_XMM avx
-IDCT8_ADD4 avx
+IDCT8_ADD4
 %endif
--- a/libavcodec/x86/h264dsp_mmx.c
+++ b/libavcodec/x86/h264dsp_mmx.c
@ -249,12 +249,12 @@ void ff_deblock_ ## DIR ## _ ## TYPE ## _ ## DEPTH ## _ ## OPT (uint8_t *pix, in
                                                                int alpha, int beta);
 #define LF_FUNCS(type, depth)\
-LF_FUNC (h,  chroma,       depth, mmxext)\
+LF_FUNC (h,  chroma,       depth, mmx2)\
-LF_IFUNC(h,  chroma_intra, depth, mmxext)\
+LF_IFUNC(h,  chroma_intra, depth, mmx2)\
-LF_FUNC (v,  chroma,       depth, mmxext)\
+LF_FUNC (v,  chroma,       depth, mmx2)\
-LF_IFUNC(v,  chroma_intra, depth, mmxext)\
+LF_IFUNC(v,  chroma_intra, depth, mmx2)\
-LF_FUNC (h,  luma,         depth, mmxext)\
+LF_FUNC (h,  luma,         depth, mmx2)\
-LF_IFUNC(h,  luma_intra,   depth, mmxext)\
+LF_IFUNC(h,  luma_intra,   depth, mmx2)\
 LF_FUNC (h,  luma,         depth, sse2)\
 LF_IFUNC(h,  luma_intra,   depth, sse2)\
 LF_FUNC (v,  luma,         depth, sse2)\
@ -276,24 +276,24 @@ LF_FUNCS( uint8_t,  8)
 LF_FUNCS(uint16_t, 10)
 #if ARCH_X86_32 && HAVE_YASM
-LF_FUNC (v8, luma,             8, mmxext)
+LF_FUNC (v8, luma,             8, mmx2)
-static void ff_deblock_v_luma_8_mmxext(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0)
+static void ff_deblock_v_luma_8_mmx2(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0)
 {
    if((tc0[0] & tc0[1]) >= 0)
-        ff_deblock_v8_luma_8_mmxext(pix+0, stride, alpha, beta, tc0);
+        ff_deblock_v8_luma_8_mmx2(pix+0, stride, alpha, beta, tc0);
    if((tc0[2] & tc0[3]) >= 0)
-        ff_deblock_v8_luma_8_mmxext(pix+8, stride, alpha, beta, tc0+2);
+        ff_deblock_v8_luma_8_mmx2(pix+8, stride, alpha, beta, tc0+2);
 }
-LF_IFUNC(v8, luma_intra,        8, mmxext)
+LF_IFUNC(v8, luma_intra,        8, mmx2)
-static void ff_deblock_v_luma_intra_8_mmxext(uint8_t *pix, int stride, int alpha, int beta)
+static void ff_deblock_v_luma_intra_8_mmx2(uint8_t *pix, int stride, int alpha, int beta)
 {
-    ff_deblock_v8_luma_intra_8_mmxext(pix+0, stride, alpha, beta);
+    ff_deblock_v8_luma_intra_8_mmx2(pix+0, stride, alpha, beta);
-    ff_deblock_v8_luma_intra_8_mmxext(pix+8, stride, alpha, beta);
+    ff_deblock_v8_luma_intra_8_mmx2(pix+8, stride, alpha, beta);
 }
 #endif /* ARCH_X86_32 */
-LF_FUNC (v,  luma,            10, mmxext)
+LF_FUNC (v,  luma,            10, mmx2)
-LF_IFUNC(v,  luma_intra,      10, mmxext)
+LF_IFUNC(v,  luma_intra,      10, mmx2)
 /***********************************/
 /* weighted prediction */
@ -373,17 +373,17 @@ void ff_h264dsp_init_x86(H264DSPContext *c, const int bit_depth, const int chrom
                c->h264_idct_add8  = ff_h264_idct_add8_8_mmx2;
            c->h264_idct_add16intra= ff_h264_idct_add16intra_8_mmx2;
-            c->h264_v_loop_filter_chroma= ff_deblock_v_chroma_8_mmxext;
+            c->h264_v_loop_filter_chroma= ff_deblock_v_chroma_8_mmx2;
-            c->h264_v_loop_filter_chroma_intra= ff_deblock_v_chroma_intra_8_mmxext;
+            c->h264_v_loop_filter_chroma_intra= ff_deblock_v_chroma_intra_8_mmx2;
            if (chroma_format_idc == 1) {
-                c->h264_h_loop_filter_chroma= ff_deblock_h_chroma_8_mmxext;
+                c->h264_h_loop_filter_chroma= ff_deblock_h_chroma_8_mmx2;
-                c->h264_h_loop_filter_chroma_intra= ff_deblock_h_chroma_intra_8_mmxext;
+                c->h264_h_loop_filter_chroma_intra= ff_deblock_h_chroma_intra_8_mmx2;
            }
 #if ARCH_X86_32
-            c->h264_v_loop_filter_luma= ff_deblock_v_luma_8_mmxext;
+            c->h264_v_loop_filter_luma= ff_deblock_v_luma_8_mmx2;
-            c->h264_h_loop_filter_luma= ff_deblock_h_luma_8_mmxext;
+            c->h264_h_loop_filter_luma= ff_deblock_h_luma_8_mmx2;
-            c->h264_v_loop_filter_luma_intra = ff_deblock_v_luma_intra_8_mmxext;
+            c->h264_v_loop_filter_luma_intra = ff_deblock_v_luma_intra_8_mmx2;
-            c->h264_h_loop_filter_luma_intra = ff_deblock_h_luma_intra_8_mmxext;
+            c->h264_h_loop_filter_luma_intra = ff_deblock_h_luma_intra_8_mmx2;
 #endif
            c->weight_h264_pixels_tab[0]= ff_h264_weight_16_mmx2;
            c->weight_h264_pixels_tab[1]= ff_h264_weight_8_mmx2;
@ -436,12 +436,12 @@ void ff_h264dsp_init_x86(H264DSPContext *c, const int bit_depth, const int chrom
    if (mm_flags & AV_CPU_FLAG_MMX) {
        if (mm_flags & AV_CPU_FLAG_MMX2) {
 #if ARCH_X86_32
-            c->h264_v_loop_filter_chroma= ff_deblock_v_chroma_10_mmxext;
+            c->h264_v_loop_filter_chroma= ff_deblock_v_chroma_10_mmx2;
-            c->h264_v_loop_filter_chroma_intra= ff_deblock_v_chroma_intra_10_mmxext;
+            c->h264_v_loop_filter_chroma_intra= ff_deblock_v_chroma_intra_10_mmx2;
-            c->h264_v_loop_filter_luma= ff_deblock_v_luma_10_mmxext;
+            c->h264_v_loop_filter_luma= ff_deblock_v_luma_10_mmx2;
-            c->h264_h_loop_filter_luma= ff_deblock_h_luma_10_mmxext;
+            c->h264_h_loop_filter_luma= ff_deblock_h_luma_10_mmx2;
-            c->h264_v_loop_filter_luma_intra = ff_deblock_v_luma_intra_10_mmxext;
+            c->h264_v_loop_filter_luma_intra = ff_deblock_v_luma_intra_10_mmx2;
-            c->h264_h_loop_filter_luma_intra = ff_deblock_h_luma_intra_10_mmxext;
+            c->h264_h_loop_filter_luma_intra = ff_deblock_h_luma_intra_10_mmx2;
 #endif
            c->h264_idct_dc_add= ff_h264_idct_dc_add_10_mmx2;
            if (mm_flags&AV_CPU_FLAG_SSE2) {
--- a/libavcodec/x86/rv34dsp.asm
+++ b/libavcodec/x86/rv34dsp.asm
@ -46,7 +46,7 @@ SECTION .text
 %endmacro
 %macro rv34_idct 1
-cglobal rv34_idct_%1_mmx2, 1, 2, 0
+cglobal rv34_idct_%1, 1, 2, 0
    movsx   r1, word [r0]
    IDCT_DC r1
    movd    m0, r1d
@ -58,14 +58,15 @@ cglobal rv34_idct_%1_mmx2, 1, 2, 0
    REP_RET
 %endmacro
-INIT_MMX
+INIT_MMX mmx2
 %define IDCT_DC IDCT_DC_ROUND
 rv34_idct dc
 %define IDCT_DC IDCT_DC_NOROUND
 rv34_idct dc_noround
 ; ff_rv34_idct_dc_add_mmx(uint8_t *dst, int stride, int dc);
-cglobal rv34_idct_dc_add_mmx, 3, 3
+INIT_MMX mmx
 cglobal rv34_idct_dc_add, 3, 3
    ; calculate DC
    IDCT_DC_ROUND r2
    pxor       m1, m1
@ -167,8 +168,8 @@ cglobal rv34_idct_add, 3,3,0, d, s, b
    ret
 ; ff_rv34_idct_dc_add_sse4(uint8_t *dst, int stride, int dc);
-INIT_XMM
+INIT_XMM sse4
-cglobal rv34_idct_dc_add_sse4, 3, 3, 6
+cglobal rv34_idct_dc_add, 3, 3, 6
    ; load data
    IDCT_DC_ROUND r2
    pxor       m1, m1
--- a/libavcodec/x86/vp3dsp.asm
+++ b/libavcodec/x86/vp3dsp.asm
@ -102,8 +102,8 @@ SECTION .text
    mov  [r0+r3  -1], r2w
 %endmacro
-INIT_MMX
+INIT_MMX mmx2
-cglobal vp3_v_loop_filter_mmx2, 3, 4
+cglobal vp3_v_loop_filter, 3, 4
 %if ARCH_X86_64
    movsxd        r1, r1d
 %endif
@ -120,7 +120,7 @@ cglobal vp3_v_loop_filter_mmx2, 3, 4
    movq     [r0   ], m3
    RET
-cglobal vp3_h_loop_filter_mmx2, 3, 4
+cglobal vp3_h_loop_filter, 3, 4
 %if ARCH_X86_64
    movsxd        r1, r1d
 %endif
@ -354,38 +354,6 @@ cglobal vp3_h_loop_filter_mmx2, 3, 4
    movq        I(2), m2
 %endmacro
 %macro VP3_IDCT_mmx 1
    ; eax = quantized input
    ; ebx = dequantizer matrix
    ; ecx = IDCT constants
    ;  M(I) = ecx + MaskOffset(0) + I * 8
    ;  C(I) = ecx + CosineOffset(32) + (I-1) * 8
    ; edx = output
    ; r0..r7 = mm0..mm7
 %define OC_8 [pw_8]
 %define C(x) [vp3_idct_data+16*(x-1)]
    ; at this point, function has completed dequantization + dezigzag +
    ; partial transposition; now do the idct itself
 %define I(x) [%1+16* x     ]
 %define J(x) [%1+16*(x-4)+8]
    RowIDCT
    Transpose
 %define I(x) [%1+16* x   +64]
 %define J(x) [%1+16*(x-4)+72]
    RowIDCT
    Transpose
 %define I(x) [%1+16*x]
 %define J(x) [%1+16*x]
    ColumnIDCT
 %define I(x) [%1+16*x+8]
 %define J(x) [%1+16*x+8]
    ColumnIDCT
 %endmacro
 %macro VP3_1D_IDCT_SSE2 0
    movdqa        m2, I(3)      ; xmm2 = i3
    movdqa        m6, C(3)      ; xmm6 = c3
@ -501,7 +469,8 @@ cglobal vp3_h_loop_filter_mmx2, 3, 4
    movdqa      O(7), m%8
 %endmacro
-%macro VP3_IDCT_sse2 1
+%macro VP3_IDCT 1
 %if mmsize == 16
 %define I(x) [%1+16*x]
 %define O(x) [%1+16*x]
 %define C(x) [vp3_idct_data+16*(x-1)]
@ -519,11 +488,42 @@ cglobal vp3_h_loop_filter_mmx2, 3, 4
 %define ADD(x)   paddsw x, [pw_8]
        VP3_1D_IDCT_SSE2
        PUT_BLOCK 0, 1, 2, 3, 4, 5, 6, 7
 %else ; mmsize == 8
    ; eax = quantized input
    ; ebx = dequantizer matrix
    ; ecx = IDCT constants
    ;  M(I) = ecx + MaskOffset(0) + I * 8
    ;  C(I) = ecx + CosineOffset(32) + (I-1) * 8
    ; edx = output
    ; r0..r7 = mm0..mm7
 %define OC_8 [pw_8]
 %define C(x) [vp3_idct_data+16*(x-1)]
    ; at this point, function has completed dequantization + dezigzag +
    ; partial transposition; now do the idct itself
 %define I(x) [%1+16* x     ]
 %define J(x) [%1+16*(x-4)+8]
    RowIDCT
    Transpose
 %define I(x) [%1+16* x   +64]
 %define J(x) [%1+16*(x-4)+72]
    RowIDCT
    Transpose
 %define I(x) [%1+16*x]
 %define J(x) [%1+16*x]
    ColumnIDCT
 %define I(x) [%1+16*x+8]
 %define J(x) [%1+16*x+8]
    ColumnIDCT
 %endif ; mmsize == 16/8
 %endmacro
-%macro vp3_idct_funcs 1
+%macro vp3_idct_funcs 0
-cglobal vp3_idct_put_%1, 3, 4, 9
+cglobal vp3_idct_put, 3, 4, 9
-    VP3_IDCT_%1   r2
+    VP3_IDCT      r2
    movsxdifnidn  r1, r1d
    mova          m4, [pb_80]
@ -565,8 +565,8 @@ cglobal vp3_idct_put_%1, 3, 4, 9
 %endrep
    RET
-cglobal vp3_idct_add_%1, 3, 4, 9
+cglobal vp3_idct_add, 3, 4, 9
-    VP3_IDCT_%1   r2
+    VP3_IDCT      r2
    mov           r3, 4
    pxor          m4, m4
@ -607,10 +607,13 @@ cglobal vp3_idct_add_%1, 3, 4, 9
    RET
 %endmacro
-INIT_MMX
+%if ARCH_X86_32
-vp3_idct_funcs mmx
+INIT_MMX mmx
-INIT_XMM
+vp3_idct_funcs
-vp3_idct_funcs sse2
+%endif
 INIT_XMM sse2
 vp3_idct_funcs
 %macro DC_ADD 0
    movq          m2, [r0     ]
@ -631,8 +634,8 @@ vp3_idct_funcs sse2
    movq   [r0+r3  ], m5
 %endmacro
-INIT_MMX
+INIT_MMX mmx2
-cglobal vp3_idct_dc_add_mmx2, 3, 4
+cglobal vp3_idct_dc_add, 3, 4
 %if ARCH_X86_64
    movsxd        r1, r1d
 %endif
--- a/libavcodec/x86/vp3dsp_init.c
+++ b/libavcodec/x86/vp3dsp_init.c
@ -41,11 +41,13 @@ av_cold void ff_vp3dsp_init_x86(VP3DSPContext *c, int flags)
 #if HAVE_YASM
    int cpuflags = av_get_cpu_flags();
 #if ARCH_X86_32
    if (HAVE_MMX && cpuflags & AV_CPU_FLAG_MMX) {
        c->idct_put  = ff_vp3_idct_put_mmx;
        c->idct_add  = ff_vp3_idct_add_mmx;
        c->idct_perm = FF_PARTTRANS_IDCT_PERM;
    }
 #endif
    if (HAVE_MMX2 && cpuflags & AV_CPU_FLAG_MMX2) {
        c->idct_dc_add = ff_vp3_idct_dc_add_mmx2;
--- a/libavcodec/x86/vp56dsp.asm
+++ b/libavcodec/x86/vp56dsp.asm
@ -27,7 +27,8 @@ cextern pw_64
 SECTION .text
-%macro DIAG4_MMX 6
+%macro DIAG4 6
 %if mmsize == 8
    movq          m0, [%1+%2]
    movq          m1, [%1+%3]
    movq          m3, m0
@ -64,9 +65,7 @@ SECTION .text
    psraw         m3, 7
    packuswb      m0, m3
    movq        [%6], m0
-%endmacro
+%else ; mmsize == 16
 %macro DIAG4_SSE2 6
    movq          m0, [%1+%2]
    movq          m1, [%1+%3]
    punpcklbw     m0, m7
@ -86,9 +85,11 @@ SECTION .text
    psraw         m0, 7
    packuswb      m0, m0
    movq        [%6], m0
 %endif ; mmsize == 8/16
 %endmacro
-%macro SPLAT4REGS_MMX 0
+%macro SPLAT4REGS 0
 %if mmsize == 8
    movq         m5, m3
    punpcklwd    m3, m3
    movq         m4, m3
@ -102,9 +103,7 @@ SECTION .text
    movq [rsp+8*12], m4
    movq [rsp+8*13], m5
    movq [rsp+8*14], m2
-%endmacro
+%else ; mmsize == 16
 %macro SPLAT4REGS_SSE2 0
    pshuflw      m4, m3, 0x0
    pshuflw      m5, m3, 0x55
    pshuflw      m6, m3, 0xAA
@ -113,15 +112,16 @@ SECTION .text
    punpcklqdq   m5, m5
    punpcklqdq   m6, m6
    punpcklqdq   m3, m3
 %endif ; mmsize == 8/16
 %endmacro
-%macro vp6_filter_diag4 2
+%macro vp6_filter_diag4 0
 ; void ff_vp6_filter_diag4_<opt>(uint8_t *dst, uint8_t *src, int stride,
 ;                                const int16_t h_weight[4], const int16_t v_weights[4])
-cglobal vp6_filter_diag4_%1, 5, 7, %2
+cglobal vp6_filter_diag4, 5, 7, 8
    mov          r5, rsp         ; backup stack pointer
    and         rsp, ~(mmsize-1) ; align stack
-%ifidn %1, sse2
+%if mmsize == 16
    sub         rsp, 8*11
 %else
    sub         rsp, 8*15
@ -162,12 +162,10 @@ cglobal vp6_filter_diag4_%1, 5, 7, %2
    RET
 %endmacro
-INIT_MMX
+%if ARCH_X86_32
-%define DIAG4      DIAG4_MMX
+INIT_MMX mmx
-%define SPLAT4REGS SPLAT4REGS_MMX
+vp6_filter_diag4
-vp6_filter_diag4 mmx,  0
+%endif
-INIT_XMM
+INIT_XMM sse2
-%define DIAG4      DIAG4_SSE2
+vp6_filter_diag4
 %define SPLAT4REGS SPLAT4REGS_SSE2
 vp6_filter_diag4 sse2, 8
--- a/libavcodec/x86/vp56dsp_init.c
+++ b/libavcodec/x86/vp56dsp_init.c
@ -36,9 +36,11 @@ av_cold void ff_vp56dsp_init_x86(VP56DSPContext* c, enum CodecID codec)
    int mm_flags = av_get_cpu_flags();
    if (CONFIG_VP6_DECODER && codec == CODEC_ID_VP6) {
 #if ARCH_X86_32
        if (mm_flags & AV_CPU_FLAG_MMX) {
            c->vp6_filter_diag4 = ff_vp6_filter_diag4_mmx;
        }
 #endif
        if (mm_flags & AV_CPU_FLAG_SSE2) {
            c->vp6_filter_diag4 = ff_vp6_filter_diag4_sse2;
--- a/libavdevice/Makefile
+++ b/libavdevice/Makefile
@ -35,7 +35,7 @@ OBJS-$(CONFIG_SNDIO_OUTDEV)              += sndio_common.o sndio_enc.o
 OBJS-$(CONFIG_V4L2_INDEV)                += v4l2.o timefilter.o
 OBJS-$(CONFIG_V4L_INDEV)                 += v4l.o
 OBJS-$(CONFIG_VFWCAP_INDEV)              += vfwcap.o
-OBJS-$(CONFIG_X11_GRAB_DEVICE_INDEV)     += x11grab.o
+OBJS-$(CONFIG_X11GRAB_INDEV)             += x11grab.o
 # external libraries
 OBJS-$(CONFIG_LIBCDIO_INDEV)             += libcdio.o
--- a/libavdevice/alldevices.c
+++ b/libavdevice/alldevices.c
@ -55,7 +55,7 @@ void avdevice_register_all(void)
    REGISTER_INDEV    (V4L2, v4l2);
 //    REGISTER_INDEV    (V4L, v4l
    REGISTER_INDEV    (VFWCAP, vfwcap);
-    REGISTER_INDEV    (X11_GRAB_DEVICE, x11_grab_device);
+    REGISTER_INDEV    (X11GRAB, x11grab);
    /* external libraries */
    REGISTER_INDEV    (LIBCDIO, libcdio);
--- a/libavdevice/avdevice.h
+++ b/libavdevice/avdevice.h
@ -36,7 +36,7 @@
 * (de)muxers in libavdevice are of the AVFMT_NOFILE type (they use their own
 * I/O functions). The filename passed to avformat_open_input() often does not
 * refer to an actually existing file, but has some special device-specific
- * meaning - e.g. for the x11grab device it is the display name.
+ * meaning - e.g. for x11grab it is the display name.
 *
 * To use libavdevice, simply call avdevice_register_all() to register all
 * compiled muxers and demuxers. They all use standard libavformat API.
--- a/libavdevice/fbdev.c
+++ b/libavdevice/fbdev.c
@ -39,6 +39,7 @@
 #include "libavutil/log.h"
 #include "libavutil/mem.h"
 #include "libavutil/opt.h"
 #include "libavutil/time.h"
 #include "libavutil/parseutils.h"
 #include "libavutil/pixdesc.h"
 #include "avdevice.h"
--- a/libavdevice/jack_audio.c
+++ b/libavdevice/jack_audio.c
@ -27,6 +27,7 @@
 #include "libavutil/log.h"
 #include "libavutil/fifo.h"
 #include "libavutil/opt.h"
 #include "libavutil/time.h"
 #include "libavcodec/avcodec.h"
 #include "libavformat/avformat.h"
 #include "libavformat/internal.h"
--- a/libavdevice/oss_audio.c
+++ b/libavdevice/oss_audio.c
@ -36,6 +36,7 @@
 #include "libavutil/log.h"
 #include "libavutil/opt.h"
 #include "libavutil/time.h"
 #include "libavcodec/avcodec.h"
 #include "avdevice.h"
 #include "libavformat/internal.h"
--- a/libavdevice/x11grab.c
+++ b/libavdevice/x11grab.c
@ -41,6 +41,7 @@
 #include "libavutil/log.h"
 #include "libavutil/opt.h"
 #include "libavutil/parseutils.h"
 #include "libavutil/time.h"
 #include <time.h>
 #include <X11/X.h>
 #include <X11/Xlib.h>
@ -56,8 +57,7 @@
 /**
 * X11 Device Demuxer context
 */
-struct x11_grab
+struct x11grab {
 {
    const AVClass *class;    /**< Class for private options. */
    int frame_size;          /**< Size in bytes of a grabbed frame */
    AVRational time_base;    /**< Time base */
@ -84,10 +84,10 @@ struct x11_grab
 /**
 * Draw grabbing region window
 *
- * @param s x11_grab context
+ * @param s x11grab context
 */
 static void
-x11grab_draw_region_win(struct x11_grab *s)
+x11grab_draw_region_win(struct x11grab *s)
 {
    Display *dpy = s->dpy;
    int screen;
@ -109,10 +109,10 @@ x11grab_draw_region_win(struct x11_grab *s)
 /**
 * Initialize grabbing region window
 *
- * @param s x11_grab context
+ * @param s x11grab context
 */
 static void
-x11grab_region_win_init(struct x11_grab *s)
+x11grab_region_win_init(struct x11grab *s)
 {
    Display *dpy = s->dpy;
    int screen;
@ -154,7 +154,7 @@ x11grab_region_win_init(struct x11_grab *s)
 static int
 x11grab_read_header(AVFormatContext *s1)
 {
-    struct x11_grab *x11grab = s1->priv_data;
+    struct x11grab *x11grab = s1->priv_data;
    Display *dpy;
    AVStream *st = NULL;
    enum PixelFormat input_pixfmt;
@ -330,7 +330,7 @@ out:
 *          coordinates
 */
 static void
-paint_mouse_pointer(XImage *image, struct x11_grab *s)
+paint_mouse_pointer(XImage *image, struct x11grab *s)
 {
    int x_off = s->x_off;
    int y_off = s->y_off;
@ -444,7 +444,7 @@ xget_zpixmap(Display *dpy, Drawable d, XImage *image, int x, int y)
 static int
 x11grab_read_packet(AVFormatContext *s1, AVPacket *pkt)
 {
-    struct x11_grab *s = s1->priv_data;
+    struct x11grab *s = s1->priv_data;
    Display *dpy = s->dpy;
    XImage *image = s->image;
    int x_off = s->x_off;
@ -554,7 +554,7 @@ x11grab_read_packet(AVFormatContext *s1, AVPacket *pkt)
 static int
 x11grab_read_close(AVFormatContext *s1)
 {
-    struct x11_grab *x11grab = s1->priv_data;
+    struct x11grab *x11grab = s1->priv_data;
    /* Detach cleanly from shared mem */
    if (x11grab->use_shm) {
@ -578,7 +578,7 @@ x11grab_read_close(AVFormatContext *s1)
    return 0;
 }
-#define OFFSET(x) offsetof(struct x11_grab, x)
+#define OFFSET(x) offsetof(struct x11grab, x)
 #define DEC AV_OPT_FLAG_DECODING_PARAM
 static const AVOption options[] = {
    { "video_size", "A string describing frame size, such as 640x480 or hd720.", OFFSET(width), AV_OPT_TYPE_IMAGE_SIZE, {.str = "vga"}, 0, 0, DEC },
@ -599,10 +599,10 @@ static const AVClass x11_class = {
 };
 /** x11 grabber device demuxer declaration */
-AVInputFormat ff_x11_grab_device_demuxer = {
+AVInputFormat ff_x11grab_demuxer = {
    .name           = "x11grab",
    .long_name      = NULL_IF_CONFIG_SMALL("X11grab"),
-    .priv_data_size = sizeof(struct x11_grab),
+    .priv_data_size = sizeof(struct x11grab),
    .read_header    = x11grab_read_header,
    .read_packet    = x11grab_read_packet,
    .read_close     = x11grab_read_close,
--- a/libavfilter/version.h
+++ b/libavfilter/version.h
@ -29,7 +29,7 @@
 #include "libavutil/avutil.h"
 #define LIBAVFILTER_VERSION_MAJOR  3
-#define LIBAVFILTER_VERSION_MINOR  4
+#define LIBAVFILTER_VERSION_MINOR  5
 #define LIBAVFILTER_VERSION_MICRO 100
 #define LIBAVFILTER_VERSION_INT AV_VERSION_INT(LIBAVFILTER_VERSION_MAJOR, \
--- a/libavformat/hls.c
+++ b/libavformat/hls.c
@ -420,8 +420,6 @@ reload:
    ret = ffurl_read(v->input, buf, buf_size);
    if (ret > 0)
        return ret;
    if (ret < 0 && ret != AVERROR_EOF)
        return ret;
    ffurl_close(v->input);
    v->input = NULL;
    v->cur_seq_no++;
--- a/libavformat/mp3dec.c
+++ b/libavformat/mp3dec.c
@ -196,7 +196,6 @@ static int mp3_read_packet(AVFormatContext *s, AVPacket *pkt)
    MP3Context *mp3 = s->priv_data;
    int ret, size;
    int64_t pos;
    //    AVStream *st = s->streams[0];
    size= MP3_PACKET_SIZE;
    pos = avio_tell(s->pb);
@ -204,15 +203,15 @@ static int mp3_read_packet(AVFormatContext *s, AVPacket *pkt)
        size= FFMIN(size, mp3->filesize - pos);
    ret= av_get_packet(s->pb, pkt, size);
    pkt->flags &= ~AV_PKT_FLAG_CORRUPT;
    pkt->stream_index = 0;
    if (ret <= 0) {
        if(ret<0)
            return ret;
        return AVERROR_EOF;
    }
    pkt->flags &= ~AV_PKT_FLAG_CORRUPT;
    pkt->stream_index = 0;
    if (ret >= ID3v1_TAG_SIZE &&
        memcmp(&pkt->data[ret - ID3v1_TAG_SIZE], "TAG", 3) == 0)
        ret -= ID3v1_TAG_SIZE;
--- a/libavformat/rtpdec.c
+++ b/libavformat/rtpdec.c
@ -21,6 +21,7 @@
 #include "libavutil/mathematics.h"
 #include "libavutil/avstring.h"
 #include "libavutil/time.h"
 #include "libavcodec/get_bits.h"
 #include "avformat.h"
 #include "mpegts.h"
--- a/libavformat/rtsp.c
+++ b/libavformat/rtsp.c
@ -27,6 +27,7 @@
 #include "libavutil/random_seed.h"
 #include "libavutil/dict.h"
 #include "libavutil/opt.h"
 #include "libavutil/time.h"
 #include "avformat.h"
 #include "avio_internal.h"
--- a/libavformat/rtspdec.c
+++ b/libavformat/rtspdec.c
@ -23,6 +23,7 @@
 #include "libavutil/intreadwrite.h"
 #include "libavutil/mathematics.h"
 #include "libavutil/random_seed.h"
 #include "libavutil/time.h"
 #include "avformat.h"
 #include "internal.h"
--- a/libavformat/rtspenc.c
+++ b/libavformat/rtspenc.c
@ -31,6 +31,7 @@
 #include "avio_internal.h"
 #include "libavutil/intreadwrite.h"
 #include "libavutil/avstring.h"
 #include "libavutil/time.h"
 #include "url.h"
 #define SDP_MAX_SIZE 16384
--- a/libavformat/sapenc.c
+++ b/libavformat/sapenc.c
@ -24,6 +24,7 @@
 #include "libavutil/random_seed.h"
 #include "libavutil/avstring.h"
 #include "libavutil/intreadwrite.h"
 #include "libavutil/time.h"
 #include "internal.h"
 #include "network.h"
 #include "os_support.h"
--- a/libavformat/tls.c
+++ b/libavformat/tls.c
@ -266,7 +266,7 @@ static int tls_read(URLContext *h, uint8_t *buf, int size)
        if (ret > 0)
            return ret;
        if (ret == 0)
-            return AVERROR(EIO);
+            return AVERROR_EOF;
        if ((ret = do_tls_poll(h, ret)) < 0)
            return ret;
    }
@ -281,7 +281,7 @@ static int tls_write(URLContext *h, const uint8_t *buf, int size)
        if (ret > 0)
            return ret;
        if (ret == 0)
-            return AVERROR(EIO);
+            return AVERROR_EOF;
        if ((ret = do_tls_poll(h, ret)) < 0)
            return ret;
    }
--- a/libavformat/utils.c
+++ b/libavformat/utils.c
@ -37,6 +37,7 @@
 #include "libavutil/avstring.h"
 #include "libavutil/mathematics.h"
 #include "libavutil/parseutils.h"
 #include "libavutil/time.h"
 #include "libavutil/timestamp.h"
 #include "riff.h"
 #include "audiointerleave.h"
@ -2765,9 +2766,7 @@ int avformat_find_stream_info(AVFormatContext *ic, AVDictionary **options)
            if (tb_unreliable(st->codec) && st->info->duration_count > 15 && st->info->duration_gcd > FFMAX(1, st->time_base.den/(500LL*st->time_base.num)) && !st->r_frame_rate.num)
                av_reduce(&st->r_frame_rate.num, &st->r_frame_rate.den, st->time_base.den, st->time_base.num * st->info->duration_gcd, INT_MAX);
            if (st->info->duration_count && !st->r_frame_rate.num
-               && tb_unreliable(st->codec) /*&&
+                && tb_unreliable(st->codec)) {
               //FIXME we should not special-case MPEG-2, but this needs testing with non-MPEG-2 ...
               st->time_base.num*duration_sum[i]/st->info->duration_count*101LL > st->time_base.den*/){
                int num = 0;
                double best_error= 0.01;
--- a/libavformat/wv.c
+++ b/libavformat/wv.c
@ -90,17 +90,17 @@ static int wv_read_block_header(AVFormatContext *ctx, AVIOContext *pb, int appen
    if(!append){
        tag = avio_rl32(pb);
        if (tag != MKTAG('w', 'v', 'p', 'k'))
-            return -1;
+            return AVERROR_INVALIDDATA;
        size = avio_rl32(pb);
        if(size < 24 || size > WV_BLOCK_LIMIT){
            av_log(ctx, AV_LOG_ERROR, "Incorrect block size %i\n", size);
-            return -1;
+            return AVERROR_INVALIDDATA;
        }
        wc->blksize = size;
        ver = avio_rl16(pb);
        if(ver < 0x402 || ver > 0x410){
            av_log(ctx, AV_LOG_ERROR, "Unsupported version %03X\n", ver);
-            return -1;
+            return AVERROR_PATCHWELCOME;
        }
        avio_r8(pb); // track no
        avio_r8(pb); // track sub index
@ -128,7 +128,7 @@ static int wv_read_block_header(AVFormatContext *ctx, AVIOContext *pb, int appen
        int64_t block_end = avio_tell(pb) + wc->blksize - 24;
        if(!pb->seekable){
            av_log(ctx, AV_LOG_ERROR, "Cannot determine additional parameters\n");
-            return -1;
+            return AVERROR_INVALIDDATA;
        }
        while(avio_tell(pb) < block_end){
            int id, size;
@ -141,7 +141,7 @@ static int wv_read_block_header(AVFormatContext *ctx, AVIOContext *pb, int appen
            case 0xD:
                if(size <= 1){
                    av_log(ctx, AV_LOG_ERROR, "Insufficient channel information\n");
-                    return -1;
+                    return AVERROR_INVALIDDATA;
                }
                chan = avio_r8(pb);
                switch(size - 2){
@ -164,7 +164,7 @@ static int wv_read_block_header(AVFormatContext *ctx, AVIOContext *pb, int appen
                    break;
                default:
                    av_log(ctx, AV_LOG_ERROR, "Invalid channel info size %d\n", size);
-                    return -1;
+                    return AVERROR_INVALIDDATA;
                }
                break;
            case 0x27:
@ -178,7 +178,7 @@ static int wv_read_block_header(AVFormatContext *ctx, AVIOContext *pb, int appen
        }
        if(rate == -1){
            av_log(ctx, AV_LOG_ERROR, "Cannot determine custom sampling rate\n");
-            return -1;
+            return AVERROR_INVALIDDATA;
        }
        avio_seek(pb, block_end - wc->blksize + 24, SEEK_SET);
    }
@ -189,15 +189,15 @@ static int wv_read_block_header(AVFormatContext *ctx, AVIOContext *pb, int appen
    if(wc->flags && bpp != wc->bpp){
        av_log(ctx, AV_LOG_ERROR, "Bits per sample differ, this block: %i, header block: %i\n", bpp, wc->bpp);
-        return -1;
+        return AVERROR_INVALIDDATA;
    }
    if(wc->flags && !wc->multichannel && chan != wc->chan){
        av_log(ctx, AV_LOG_ERROR, "Channels differ, this block: %i, header block: %i\n", chan, wc->chan);
-        return -1;
+        return AVERROR_INVALIDDATA;
    }
    if(wc->flags && rate != -1 && rate != wc->rate){
        av_log(ctx, AV_LOG_ERROR, "Sampling rate differ, this block: %i, header block: %i\n", rate, wc->rate);
-        return -1;
+        return AVERROR_INVALIDDATA;
    }
    wc->blksize = size - 24;
    return 0;
@ -208,11 +208,12 @@ static int wv_read_header(AVFormatContext *s)
    AVIOContext *pb = s->pb;
    WVContext *wc = s->priv_data;
    AVStream *st;
    int ret;
    wc->block_parsed = 0;
    for(;;){
-        if(wv_read_block_header(s, pb, 0) < 0)
+        if ((ret = wv_read_block_header(s, pb, 0)) < 0)
-            return -1;
+            return ret;
        if(!AV_RN32(wc->extra))
            avio_skip(pb, wc->blksize - 24);
        else
@ -222,7 +223,7 @@ static int wv_read_header(AVFormatContext *s)
    /* now we are ready: build format streams */
    st = avformat_new_stream(s, NULL);
    if (!st)
-        return -1;
+        return AVERROR(ENOMEM);
    st->codec->codec_type = AVMEDIA_TYPE_AUDIO;
    st->codec->codec_id = CODEC_ID_WAVPACK;
    st->codec->channels = wc->chan;
@ -254,10 +255,10 @@ static int wv_read_packet(AVFormatContext *s,
    uint32_t block_samples;
    if (url_feof(s->pb))
-        return AVERROR(EIO);
+        return AVERROR_EOF;
    if(wc->block_parsed){
-        if(wv_read_block_header(s, s->pb, 0) < 0)
+        if ((ret = wv_read_block_header(s, s->pb, 0)) < 0)
-            return -1;
+            return ret;
    }
    pos = wc->pos;
@ -275,7 +276,7 @@ static int wv_read_packet(AVFormatContext *s,
    while(!(wc->flags & WV_END_BLOCK)){
        if(avio_rl32(s->pb) != MKTAG('w', 'v', 'p', 'k')){
            av_free_packet(pkt);
-            return -1;
+            return AVERROR_INVALIDDATA;
        }
        if((ret = av_append_packet(s->pb, pkt, 4)) < 0){
            av_free_packet(pkt);
@ -285,14 +286,14 @@ static int wv_read_packet(AVFormatContext *s,
        if(size < 24 || size > WV_BLOCK_LIMIT){
            av_free_packet(pkt);
            av_log(s, AV_LOG_ERROR, "Incorrect block size %d\n", size);
-            return -1;
+            return AVERROR_INVALIDDATA;
        }
        wc->blksize = size;
        ver = avio_rl16(s->pb);
        if(ver < 0x402 || ver > 0x410){
            av_free_packet(pkt);
            av_log(s, AV_LOG_ERROR, "Unsupported version %03X\n", ver);
-            return -1;
+            return AVERROR_PATCHWELCOME;
        }
        avio_r8(s->pb); // track no
        avio_r8(s->pb); // track sub index
@ -304,9 +305,9 @@ static int wv_read_packet(AVFormatContext *s,
        }
        memcpy(wc->extra, pkt->data + pkt->size - WV_EXTRA_SIZE, WV_EXTRA_SIZE);
-        if(wv_read_block_header(s, s->pb, 1) < 0){
+        if ((ret = wv_read_block_header(s, s->pb, 1)) < 0){
            av_free_packet(pkt);
-            return -1;
+            return ret;
        }
        ret = av_append_packet(s->pb, pkt, wc->blksize);
        if(ret < 0){
@ -345,14 +346,14 @@ static int wv_read_seek(AVFormatContext *s, int stream_index, int64_t timestamp,
    }
    /* if timestamp is out of bounds, return error */
    if(timestamp < 0 || timestamp >= s->duration)
-        return -1;
+        return AVERROR(EINVAL);
    pos = avio_tell(s->pb);
    do{
        ret = av_read_frame(s, pkt);
        if (ret < 0){
            avio_seek(s->pb, pos, SEEK_SET);
-            return -1;
+            return ret;
        }
        pts = pkt->pts;
        av_free_packet(pkt);
--- a/libavutil/x86/x86inc.asm
+++ b/libavutil/x86/x86inc.asm
@ -36,8 +36,8 @@
 %define program_name ff
 %define UNIX64 0
 %define WIN64  0
 %define UNIX64 0
 %if ARCH_X86_64
    %ifidn __OUTPUT_FORMAT__,win32
        %define WIN64  1
@ -54,11 +54,6 @@
    %define mangle(x) x
 %endif
 ; FIXME: All of the 64bit asm functions that take a stride as an argument
 ; via register, assume that the high dword of that register is filled with 0.
 ; This is true in practice (since we never do any 64bit arithmetic on strides,
 ; and x264's strides are all positive), but is not guaranteed by the ABI.
 ; Name of the .rodata section.
 %macro SECTION_RODATA 0-1 16
    ; Kludge: Something on OS X fails to align .rodata even given an align
@ -152,34 +147,38 @@ CPU amdnop
 ; registers:
 ; rN and rNq are the native-size register holding function argument N
 ; rNd, rNw, rNb are dword, word, and byte size
 ; rNh is the high 8 bits of the word size
 ; rNm is the original location of arg N (a register or on the stack), dword
 ; rNmp is native size
-%macro DECLARE_REG 5-6
+%macro DECLARE_REG 2-3
    %define r%1q %2
-    %define r%1d %3
+    %define r%1d %2d
-    %define r%1w %4
+    %define r%1w %2w
-    %define r%1b %5
+    %define r%1b %2b
-    %if %0 == 5
+    %define r%1h %2h
-        %define r%1m  %3
+    %if %0 == 2
        %define r%1m  %2d
        %define r%1mp %2
    %elif ARCH_X86_64 ; memory
-        %define r%1m [rsp + stack_offset + %6]
+        %define r%1m [rsp + stack_offset + %3]
        %define r%1mp qword r %+ %1 %+ m
    %else
-        %define r%1m [esp + stack_offset + %6]
+        %define r%1m [esp + stack_offset + %3]
        %define r%1mp dword r %+ %1 %+ m
    %endif
    %define r%1  %2
 %endmacro
-%macro DECLARE_REG_SIZE 2
+%macro DECLARE_REG_SIZE 3
    %define r%1q r%1
    %define e%1q r%1
    %define r%1d e%1
    %define e%1d e%1
    %define r%1w %1
    %define e%1w %1
    %define r%1h %3
    %define e%1h %3
    %define r%1b %2
    %define e%1b %2
 %if ARCH_X86_64 == 0
@ -187,13 +186,13 @@ CPU amdnop
 %endif
 %endmacro
-DECLARE_REG_SIZE ax, al
+DECLARE_REG_SIZE ax, al, ah
-DECLARE_REG_SIZE bx, bl
+DECLARE_REG_SIZE bx, bl, bh
-DECLARE_REG_SIZE cx, cl
+DECLARE_REG_SIZE cx, cl, ch
-DECLARE_REG_SIZE dx, dl
+DECLARE_REG_SIZE dx, dl, dh
-DECLARE_REG_SIZE si, sil
+DECLARE_REG_SIZE si, sil, null
-DECLARE_REG_SIZE di, dil
+DECLARE_REG_SIZE di, dil, null
-DECLARE_REG_SIZE bp, bpl
+DECLARE_REG_SIZE bp, bpl, null
 ; t# defines for when per-arch register allocation is more complex than just function arguments
@ -211,6 +210,7 @@ DECLARE_REG_SIZE bp, bpl
        %define t%1q t%1 %+ q
        %define t%1d t%1 %+ d
        %define t%1w t%1 %+ w
        %define t%1h t%1 %+ h
        %define t%1b t%1 %+ b
        %rotate 1
    %endrep
@ -300,6 +300,7 @@ DECLARE_REG_TMP_SIZE 0,1,2,3,4,5,6,7,8,9,10,11,12,13,14
            CAT_UNDEF arg_name %+ %%i, q
            CAT_UNDEF arg_name %+ %%i, d
            CAT_UNDEF arg_name %+ %%i, w
            CAT_UNDEF arg_name %+ %%i, h
            CAT_UNDEF arg_name %+ %%i, b
            CAT_UNDEF arg_name %+ %%i, m
            CAT_UNDEF arg_name %+ %%i, mp
@ -315,6 +316,7 @@ DECLARE_REG_TMP_SIZE 0,1,2,3,4,5,6,7,8,9,10,11,12,13,14
        %xdefine %1q r %+ %%i %+ q
        %xdefine %1d r %+ %%i %+ d
        %xdefine %1w r %+ %%i %+ w
        %xdefine %1h r %+ %%i %+ h
        %xdefine %1b r %+ %%i %+ b
        %xdefine %1m r %+ %%i %+ m
        %xdefine %1mp r %+ %%i %+ mp
@ -328,21 +330,21 @@ DECLARE_REG_TMP_SIZE 0,1,2,3,4,5,6,7,8,9,10,11,12,13,14
 %if WIN64 ; Windows x64 ;=================================================
-DECLARE_REG 0,  rcx, ecx,  cx,   cl
+DECLARE_REG 0,  rcx
-DECLARE_REG 1,  rdx, edx,  dx,   dl
+DECLARE_REG 1,  rdx
-DECLARE_REG 2,  R8,  R8D,  R8W,  R8B
+DECLARE_REG 2,  R8
-DECLARE_REG 3,  R9,  R9D,  R9W,  R9B
+DECLARE_REG 3,  R9
-DECLARE_REG 4,  R10, R10D, R10W, R10B, 40
+DECLARE_REG 4,  R10, 40
-DECLARE_REG 5,  R11, R11D, R11W, R11B, 48
+DECLARE_REG 5,  R11, 48
-DECLARE_REG 6,  rax, eax,  ax,   al,   56
+DECLARE_REG 6,  rax, 56
-DECLARE_REG 7,  rdi, edi,  di,   dil,  64
+DECLARE_REG 7,  rdi, 64
-DECLARE_REG 8,  rsi, esi,  si,   sil,  72
+DECLARE_REG 8,  rsi, 72
-DECLARE_REG 9,  rbx, ebx,  bx,   bl,   80
+DECLARE_REG 9,  rbx, 80
-DECLARE_REG 10, rbp, ebp,  bp,   bpl,  88
+DECLARE_REG 10, rbp, 88
-DECLARE_REG 11, R12, R12D, R12W, R12B, 96
+DECLARE_REG 11, R12, 96
-DECLARE_REG 12, R13, R13D, R13W, R13B, 104
+DECLARE_REG 12, R13, 104
-DECLARE_REG 13, R14, R14D, R14W, R14B, 112
+DECLARE_REG 13, R14, 112
-DECLARE_REG 14, R15, R15D, R15W, R15B, 120
+DECLARE_REG 14, R15, 120
 %macro PROLOGUE 2-4+ 0 ; #args, #regs, #xmm_regs, arg_names...
    %assign num_args %1
@ -389,6 +391,8 @@ DECLARE_REG 14, R15, R15D, R15W, R15B, 120
    %assign xmm_regs_used 0
 %endmacro
 %define has_epilogue regs_used > 7 || xmm_regs_used > 6 || mmsize == 32
 %macro RET 0
    WIN64_RESTORE_XMM_INTERNAL rsp
    POP_IF_USED 14, 13, 12, 11, 10, 9, 8, 7
@ -398,31 +402,23 @@ DECLARE_REG 14, R15, R15D, R15W, R15B, 120
    ret
 %endmacro
 %macro REP_RET 0
    %if regs_used > 7 || xmm_regs_used > 6 || mmsize == 32
        RET
    %else
        rep ret
    %endif
 %endmacro
 %elif ARCH_X86_64 ; *nix x64 ;=============================================
-DECLARE_REG 0,  rdi, edi,  di,   dil
+DECLARE_REG 0,  rdi
-DECLARE_REG 1,  rsi, esi,  si,   sil
+DECLARE_REG 1,  rsi
-DECLARE_REG 2,  rdx, edx,  dx,   dl
+DECLARE_REG 2,  rdx
-DECLARE_REG 3,  rcx, ecx,  cx,   cl
+DECLARE_REG 3,  rcx
-DECLARE_REG 4,  R8,  R8D,  R8W,  R8B
+DECLARE_REG 4,  R8
-DECLARE_REG 5,  R9,  R9D,  R9W,  R9B
+DECLARE_REG 5,  R9
-DECLARE_REG 6,  rax, eax,  ax,   al,   8
+DECLARE_REG 6,  rax, 8
-DECLARE_REG 7,  R10, R10D, R10W, R10B, 16
+DECLARE_REG 7,  R10, 16
-DECLARE_REG 8,  R11, R11D, R11W, R11B, 24
+DECLARE_REG 8,  R11, 24
-DECLARE_REG 9,  rbx, ebx,  bx,   bl,   32
+DECLARE_REG 9,  rbx, 32
-DECLARE_REG 10, rbp, ebp,  bp,   bpl,  40
+DECLARE_REG 10, rbp, 40
-DECLARE_REG 11, R12, R12D, R12W, R12B, 48
+DECLARE_REG 11, R12, 48
-DECLARE_REG 12, R13, R13D, R13W, R13B, 56
+DECLARE_REG 12, R13, 56
-DECLARE_REG 13, R14, R14D, R14W, R14B, 64
+DECLARE_REG 13, R14, 64
-DECLARE_REG 14, R15, R15D, R15W, R15B, 72
+DECLARE_REG 14, R15, 72
 %macro PROLOGUE 2-4+ ; #args, #regs, #xmm_regs, arg_names...
    %assign num_args %1
@ -434,6 +430,8 @@ DECLARE_REG 14, R15, R15D, R15W, R15B, 72
    DEFINE_ARGS %4
 %endmacro
 %define has_epilogue regs_used > 9 || mmsize == 32
 %macro RET 0
    POP_IF_USED 14, 13, 12, 11, 10, 9
 %if mmsize == 32
@ -442,23 +440,15 @@ DECLARE_REG 14, R15, R15D, R15W, R15B, 72
    ret
 %endmacro
 %macro REP_RET 0
    %if regs_used > 9 || mmsize == 32
        RET
    %else
        rep ret
    %endif
 %endmacro
 %else ; X86_32 ;==============================================================
-DECLARE_REG 0, eax, eax, ax, al,   4
+DECLARE_REG 0, eax, 4
-DECLARE_REG 1, ecx, ecx, cx, cl,   8
+DECLARE_REG 1, ecx, 8
-DECLARE_REG 2, edx, edx, dx, dl,   12
+DECLARE_REG 2, edx, 12
-DECLARE_REG 3, ebx, ebx, bx, bl,   16
+DECLARE_REG 3, ebx, 16
-DECLARE_REG 4, esi, esi, si, null, 20
+DECLARE_REG 4, esi, 20
-DECLARE_REG 5, edi, edi, di, null, 24
+DECLARE_REG 5, edi, 24
-DECLARE_REG 6, ebp, ebp, bp, null, 28
+DECLARE_REG 6, ebp, 28
 %define rsp esp
 %macro DECLARE_ARG 1-*
@ -474,6 +464,9 @@ DECLARE_ARG 7, 8, 9, 10, 11, 12, 13, 14
 %macro PROLOGUE 2-4+ ; #args, #regs, #xmm_regs, arg_names...
    %assign num_args %1
    %assign regs_used %2
    %if num_args > 7
        %assign num_args 7
    %endif
    %if regs_used > 7
        %assign regs_used 7
    %endif
@ -483,6 +476,8 @@ DECLARE_ARG 7, 8, 9, 10, 11, 12, 13, 14
    DEFINE_ARGS %4
 %endmacro
 %define has_epilogue regs_used > 3 || mmsize == 32
 %macro RET 0
    POP_IF_USED 6, 5, 4, 3
 %if mmsize == 32
@ -491,14 +486,6 @@ DECLARE_ARG 7, 8, 9, 10, 11, 12, 13, 14
    ret
 %endmacro
 %macro REP_RET 0
    %if regs_used > 3 || mmsize == 32
        RET
    %else
        rep ret
    %endif
 %endmacro
 %endif ;======================================================================
 %if WIN64 == 0
@ -508,6 +495,23 @@ DECLARE_ARG 7, 8, 9, 10, 11, 12, 13, 14
 %endmacro
 %endif
 %macro REP_RET 0
    %if has_epilogue
        RET
    %else
        rep ret
    %endif
 %endmacro
 %macro TAIL_CALL 2 ; callee, is_nonadjacent
    %if has_epilogue
        call %1
        RET
    %elif %2
        jmp %1
    %endif
 %endmacro
 ;=============================================================================
 ; arch-independent part
 ;=============================================================================
@ -597,6 +601,8 @@ SECTION .note.GNU-stack noalloc noexec nowrite progbits
 %assign cpuflags_avx      (1<<11)| cpuflags_sse42
 %assign cpuflags_xop      (1<<12)| cpuflags_avx
 %assign cpuflags_fma4     (1<<13)| cpuflags_avx
 %assign cpuflags_avx2     (1<<14)| cpuflags_avx
 %assign cpuflags_fma3     (1<<15)| cpuflags_avx
 %assign cpuflags_cache32  (1<<16)
 %assign cpuflags_cache64  (1<<17)
@ -605,6 +611,9 @@ SECTION .note.GNU-stack noalloc noexec nowrite progbits
 %assign cpuflags_misalign (1<<20)
 %assign cpuflags_aligned  (1<<21) ; not a cpu feature, but a function variant
 %assign cpuflags_atom     (1<<22)
 %assign cpuflags_bmi1     (1<<23)
 %assign cpuflags_bmi2     (1<<24)|cpuflags_bmi1
 %assign cpuflags_tbm      (1<<25)|cpuflags_bmi1
 %define    cpuflag(x) ((cpuflags & (cpuflags_ %+ x)) == (cpuflags_ %+ x))
 %define notcpuflag(x) ((cpuflags & (cpuflags_ %+ x)) != (cpuflags_ %+ x))
@ -875,25 +884,38 @@ INIT_XMM
 %endrep
 %undef i
 %macro CHECK_AVX_INSTR_EMU 3-*
    %xdefine %%opcode %1
    %xdefine %%dst %2
    %rep %0-2
        %ifidn %%dst, %3
            %error non-avx emulation of ``%%opcode'' is not supported
        %endif
        %rotate 1
    %endrep
 %endmacro
 ;%1 == instruction
 ;%2 == 1 if float, 0 if int
 ;%3 == 1 if 4-operand (xmm, xmm, xmm, imm), 0 if 2- or 3-operand (xmm, xmm, xmm)
 ;%4 == number of operands given
 ;%5+: operands
 %macro RUN_AVX_INSTR 6-7+
-    %ifid %5
+    %ifid %6
-        %define %%size sizeof%5
+        %define %%sizeofreg sizeof%6
    %elifid %5
        %define %%sizeofreg sizeof%5
    %else
-        %define %%size mmsize
+        %define %%sizeofreg mmsize
    %endif
-    %if %%size==32
+    %if %%sizeofreg==32
-        %if %0 >= 7
+        %if %4>=3
            v%1 %5, %6, %7
        %else
            v%1 %5, %6
        %endif
    %else
-        %if %%size==8
+        %if %%sizeofreg==8
            %define %%regmov movq
        %elif %2
            %define %%regmov movaps
@ -903,16 +925,17 @@ INIT_XMM
        %if %4>=3+%3
            %ifnidn %5, %6
-                %if avx_enabled && sizeof%5==16
+                %if avx_enabled && %%sizeofreg==16
                    v%1 %5, %6, %7
                %else
                    CHECK_AVX_INSTR_EMU {%1 %5, %6, %7}, %5, %7
                    %%regmov %5, %6
                    %1 %5, %7
                %endif
            %else
                %1 %5, %7
            %endif
-        %elif %3
+        %elif %4>=3
            %1 %5, %6, %7
        %else
            %1 %5, %6
@ -943,7 +966,7 @@ INIT_XMM
 ;%1 == instruction
 ;%2 == 1 if float, 0 if int
-;%3 == 1 if 4-operand (xmm, xmm, xmm, imm), 0 if 3-operand (xmm, xmm, xmm)
+;%3 == 1 if 4-operand (xmm, xmm, xmm, imm), 0 if 2- or 3-operand (xmm, xmm, xmm)
 ;%4 == 1 if symmetric (i.e. doesn't matter which src arg is which), 0 if not
 %macro AVX_INSTR 4
    %macro %1 2-9 fnord, fnord, fnord, %1, %2, %3, %4
@ -1008,6 +1031,9 @@ AVX_INSTR mulsd, 1, 0, 1
 AVX_INSTR mulss, 1, 0, 1
 AVX_INSTR orpd, 1, 0, 1
 AVX_INSTR orps, 1, 0, 1
 AVX_INSTR pabsb, 0, 0, 0
 AVX_INSTR pabsw, 0, 0, 0
 AVX_INSTR pabsd, 0, 0, 0
 AVX_INSTR packsswb, 0, 0, 0
 AVX_INSTR packssdw, 0, 0, 0
 AVX_INSTR packuswb, 0, 0, 0
@ -1059,6 +1085,7 @@ AVX_INSTR pminsd, 0, 0, 1
 AVX_INSTR pminub, 0, 0, 1
 AVX_INSTR pminuw, 0, 0, 1
 AVX_INSTR pminud, 0, 0, 1
 AVX_INSTR pmovmskb, 0, 0, 0
 AVX_INSTR pmulhuw, 0, 0, 1
 AVX_INSTR pmulhrsw, 0, 0, 1
 AVX_INSTR pmulhw, 0, 0, 1
@ -1069,6 +1096,9 @@ AVX_INSTR pmuldq, 0, 0, 1
 AVX_INSTR por, 0, 0, 1
 AVX_INSTR psadbw, 0, 0, 1
 AVX_INSTR pshufb, 0, 0, 0
 AVX_INSTR pshufd, 0, 1, 0
 AVX_INSTR pshufhw, 0, 1, 0
 AVX_INSTR pshuflw, 0, 1, 0
 AVX_INSTR psignb, 0, 0, 0
 AVX_INSTR psignw, 0, 0, 0
 AVX_INSTR psignd, 0, 0, 0
@ -1090,6 +1120,7 @@ AVX_INSTR psubsb, 0, 0, 0
 AVX_INSTR psubsw, 0, 0, 0
 AVX_INSTR psubusb, 0, 0, 0
 AVX_INSTR psubusw, 0, 0, 0
 AVX_INSTR ptest, 0, 0, 0
 AVX_INSTR punpckhbw, 0, 0, 0
 AVX_INSTR punpckhwd, 0, 0, 0
 AVX_INSTR punpckhdq, 0, 0, 0
@ -1154,3 +1185,7 @@ FMA_INSTR  fmaddps,   mulps, addps
 FMA_INSTR  pmacsdd,  pmulld, paddd
 FMA_INSTR  pmacsww,  pmullw, paddw
 FMA_INSTR pmadcswd, pmaddwd, paddd
 ; tzcnt is equivalent to "rep bsf" and is backwards-compatible with bsf.
 ; This lets us use tzcnt without bumping the yasm version requirement yet.
 %define tzcnt rep bsf
--- a/library.mak
+++ b/library.mak
@ -17,7 +17,7 @@ $(SUBDIR)%-test.o: $(SUBDIR)%.c
 	$(COMPILE_C)
 $(SUBDIR)x86/%.o: $(SUBDIR)x86/%.asm
-	$(YASMDEP) $(YASMFLAGS) -I $(<D)/ -M -o $@ $< > $(@:.o=.d)
+	$(DEPYASM) $(YASMFLAGS) -I $(<D)/ -M -o $@ $< > $(@:.o=.d)
 	$(YASM) $(YASMFLAGS) -I $(<D)/ -o $@ $<
 $(OBJS) $(OBJS:.o=.s) $(SUBDIR)%.ho $(TESTOBJS): CPPFLAGS += -DHAVE_AV_CONFIG_H
--- a/tests/audiogen.c
+++ b/tests/audiogen.c
@ -144,8 +144,8 @@ int main(int argc, char **argv)
    int nb_channels = 2;
    char *ext;
-    if (argc < 2 || argc > 4) {
+    if (argc < 2 || argc > 5) {
-        printf("usage: %s file [<sample rate> [<channels>]]\n"
+        printf("usage: %s file [<sample rate> [<channels>] [<random seed>]]\n"
               "generate a test raw 16 bit audio stream\n"
               "If the file extension is .wav a WAVE header will be added.\n"
               "default: 44100 Hz stereo\n", argv[0]);
@ -168,6 +168,9 @@ int main(int argc, char **argv)
        }
    }
    if (argc > 4)
        seed = atoi(argv[4]);
    outfile = fopen(argv[1], "wb");
    if (!outfile) {
        perror(argv[1]);
--- a/tests/fate/filter.mak
+++ b/tests/fate/filter.mak
@ -1,3 +1,25 @@
 FATE_AMIX += fate-filter-amix-simple
 fate-filter-amix-simple: CMD = ffmpeg -filter_complex amix -i $(SRC) -ss 3 -i $(SRC1) -f f32le -
 fate-filter-amix-simple: REF = $(SAMPLES)/filter/amix_simple.pcm
 FATE_AMIX += fate-filter-amix-first
 fate-filter-amix-first: CMD = ffmpeg -filter_complex amix=duration=first -ss 4 -i $(SRC) -i $(SRC1) -f f32le -
 fate-filter-amix-first: REF = $(SAMPLES)/filter/amix_first.pcm
 FATE_AMIX += fate-filter-amix-transition
 fate-filter-amix-transition: tests/data/asynth-44100-2-3.wav
 fate-filter-amix-transition: SRC2 = $(TARGET_PATH)/tests/data/asynth-44100-2-3.wav
 fate-filter-amix-transition: CMD = ffmpeg -filter_complex amix=inputs=3:dropout_transition=0.5 -i $(SRC) -ss 2 -i $(SRC1) -ss 4 -i $(SRC2) -f f32le -
 fate-filter-amix-transition: REF = $(SAMPLES)/filter/amix_transition.pcm
 $(FATE_AMIX): tests/data/asynth-44100-2.wav tests/data/asynth-44100-2-2.wav
 $(FATE_AMIX): SRC  = $(TARGET_PATH)/tests/data/asynth-44100-2.wav
 $(FATE_AMIX): SRC1 = $(TARGET_PATH)/tests/data/asynth-44100-2-2.wav
 $(FATE_AMIX): CMP  = oneoff
 FATE_FILTER += $(FATE_AMIX)
 FATE_SAMPLES_AVCONV += $(FATE_AMIX)
 FATE_ASYNCTS += fate-filter-asyncts
 fate-filter-asyncts: SRC = $(SAMPLES)/nellymoser/nellymoser-discont.flv
 fate-filter-asyncts: CMD = pcm -i $(SRC) -af aresample=min_comp=0.001:min_hard_comp=0.1
@ -7,4 +29,18 @@ fate-filter-asyncts: REF = $(SAMPLES)/nellymoser/nellymoser-discont.pcm
 FATE_FILTER += $(FATE_ASYNCTS)
 FATE_SAMPLES_AVCONV += $(FATE_ASYNCTS)
 fate-filter-delogo: CMD = framecrc -i $(SAMPLES)/real/rv30.rm -vf delogo=show=0:x=290:y=25:w=26:h=16 -an
 FATE_FILTER += fate-filter-delogo
 FATE_SAMPLES_AVCONV += fate-filter-delogo
 FATE_YADIF += fate-filter-yadif-mode0
 fate-filter-yadif-mode0: CMD = framecrc -i $(SAMPLES)/mpeg2/mpeg2_field_encoding.ts -vf yadif=0
 FATE_YADIF += fate-filter-yadif-mode1
 fate-filter-yadif-mode1: CMD = framecrc -i $(SAMPLES)/mpeg2/mpeg2_field_encoding.ts -vf yadif=1
 FATE_FILTER += $(FATE_YADIF)
 FATE_SAMPLES_AVCONV += $(FATE_YADIF)
 fate-filter: $(FATE_FILTER)
--- a/tests/ref/fate/filter-delogo
+++ b/tests/ref/fate/filter-delogo
@ -0,0 +1,110 @@
 #tb 0: 32768/982057
 0,          0,          0,        1,   126720, 0x689de87e
 0,          1,          1,        1,   126720, 0x3db9e91c
 0,          2,          2,        1,   126720, 0x3db9e91c
 0,          3,          3,        1,   126720, 0x3db9e91c
 0,          4,          4,        1,   126720, 0xfa6ae95e
 0,          5,          5,        1,   126720, 0x5bcbf0e6
 0,          6,          6,        1,   126720, 0x94a0f126
 0,          7,          7,        1,   126720, 0x0250f106
 0,          8,          8,        1,   126720, 0xcf6ab4bc
 0,          9,          9,        1,   126720, 0x44aeb57c
 0,         10,         10,        1,   126720, 0x33b0b5bc
 0,         11,         11,        1,   126720, 0xc4bab591
 0,         12,         12,        1,   126720, 0xa492b5ec
 0,         13,         13,        1,   126720, 0x1459b85c
 0,         14,         14,        1,   126720, 0x806fb8dc
 0,         15,         15,        1,   126720, 0xd241b871
 0,         16,         16,        1,   126720, 0x698eb5cc
 0,         17,         17,        1,   126720, 0x4719aa98
 0,         18,         18,        1,   126720, 0x9ca1962c
 0,         19,         19,        1,   126720, 0x18cda460
 0,         20,         20,        1,   126720, 0xc230b716
 0,         21,         21,        1,   126720, 0x8451a4e2
 0,         22,         22,        1,   126720, 0x59e9a7ea
 0,         23,         23,        1,   126720, 0xc77ca73d
 0,         24,         24,        1,   126720, 0x725fb976
 0,         25,         25,        1,   126720, 0xb30da3b3
 0,         26,         26,        1,   126720, 0x7af2ea86
 0,         27,         27,        1,   126720, 0x40d4b4eb
 0,         28,         28,        1,   126720, 0x49d00307
 0,         29,         29,        1,   126720, 0x44c8848e
 0,         30,         30,        1,   126720, 0xc6990101
 0,         31,         31,        1,   126720, 0x2e01b963
 0,         32,         32,        1,   126720, 0xd0e903f0
 0,         33,         33,        1,   126720, 0x3457d592
 0,         34,         34,        1,   126720, 0x4f1ddb3c
 0,         35,         35,        1,   126720, 0x3980ace5
 0,         36,         36,        1,   126720, 0xb1e37954
 0,         37,         37,        1,   126720, 0x619fc554
 0,         38,         38,        1,   126720, 0x945fb39e
 0,         39,         39,        1,   126720, 0xb1d5e0ce
 0,         40,         40,        1,   126720, 0xf26e1dcc
 0,         41,         41,        1,   126720, 0x04d5783e
 0,         42,         42,        1,   126720, 0xbaa0479e
 0,         43,         43,        1,   126720, 0x20d88b01
 0,         44,         44,        1,   126720, 0x59d99901
 0,         45,         45,        1,   126720, 0x1c6e09f6
 0,         46,         46,        1,   126720, 0xeec50fc5
 0,         47,         47,        1,   126720, 0xb3a92827
 0,         48,         48,        1,   126720, 0xf62dd2b6
 0,         49,         49,        1,   126720, 0x75b1e619
 0,         50,         50,        1,   126720, 0x6bbce2c0
 0,         51,         51,        1,   126720, 0xd93e023c
 0,         52,         52,        1,   126720, 0xbbe8e7c2
 0,         53,         53,        1,   126720, 0x2272ec17
 0,         54,         54,        1,   126720, 0xf5e4ee6e
 0,         55,         55,        1,   126720, 0x751d2607
 0,         56,         56,        1,   126720, 0x44c499c9
 0,         57,         57,        1,   126720, 0xddccd842
 0,         58,         58,        1,   126720, 0x508dd214
 0,         59,         59,        1,   126720, 0x8eb10272
 0,         60,         60,        1,   126720, 0x7224b1c6
 0,         61,         61,        1,   126720, 0x50ff456c
 0,         62,         62,        1,   126720, 0xa81e2731
 0,         63,         63,        1,   126720, 0x7e50456d
 0,         64,         64,        1,   126720, 0x44802978
 0,         65,         65,        1,   126720, 0x86e88743
 0,         66,         66,        1,   126720, 0x0b1087d6
 0,         67,         67,        1,   126720, 0xb0227d21
 0,         68,         68,        1,   126720, 0x29d10bd2
 0,         69,         69,        1,   126720, 0x04b43afa
 0,         70,         70,        1,   126720, 0xb48e9698
 0,         71,         71,        1,   126720, 0x75d760fb
 0,         72,         72,        1,   126720, 0xa2ab1fdb
 0,         73,         73,        1,   126720, 0xec30a5ee
 0,         74,         74,        1,   126720, 0xbdab7c8c
 0,         75,         75,        1,   126720, 0xac5c3f2c
 0,         76,         76,        1,   126720, 0xce6350be
 0,         77,         77,        1,   126720, 0xb109657a
 0,         78,         78,        1,   126720, 0x723865a4
 0,         79,         79,        1,   126720, 0xa9869124
 0,         80,         80,        1,   126720, 0xc41af558
 0,         81,         81,        1,   126720, 0xcbe6a402
 0,         82,         82,        1,   126720, 0xb6735ecb
 0,         83,         83,        1,   126720, 0xba3059f2
 0,         84,         84,        1,   126720, 0xe7d63b8d
 0,         85,         85,        1,   126720, 0x8f115906
 0,         86,         86,        1,   126720, 0xaf6a8dcb
 0,         87,         87,        1,   126720, 0xb73e846e
 0,         88,         88,        1,   126720, 0xedd6380f
 0,         89,         89,        1,   126720, 0xd9026acf
 0,         90,         90,        1,   126720, 0xa03a650b
 0,         91,         91,        1,   126720, 0x262765bc
 0,         92,         92,        1,   126720, 0xaaa9ded1
 0,         93,         93,        1,   126720, 0xe4f42665
 0,         94,         94,        1,   126720, 0x78daf760
 0,         95,         95,        1,   126720, 0x3b0c6ef8
 0,         96,         96,        1,   126720, 0xb745df80
 0,         97,         97,        1,   126720, 0x08e57b90
 0,         98,         98,        1,   126720, 0x6f883ab0
 0,         99,         99,        1,   126720, 0x934b4dd5
 0,        100,        100,        1,   126720, 0x762f108f
 0,        101,        101,        1,   126720, 0x91ee0f2b
 0,        102,        102,        1,   126720, 0x9af6e5e8
 0,        103,        103,        1,   126720, 0xdcd95e0a
 0,        104,        104,        1,   126720, 0x22c33a6e
 0,        105,        105,        1,   126720, 0x21c1b7f4
 0,        106,        106,        1,   126720, 0x0a66a1ed
 0,        107,        107,        1,   126720, 0x53fea81b
 0,        108,        108,        1,   126720, 0x597f5567
--- a/tests/ref/fate/filter-yadif-mode0
+++ b/tests/ref/fate/filter-yadif-mode0
@ -0,0 +1,32 @@
 #tb 0: 1/25
 0,          9,          9,        1,   622080, 0x1511cae9
 0,         10,         10,        1,   622080, 0x6e77e746
 0,         11,         11,        1,   622080, 0x89aac777
 0,         12,         12,        1,   622080, 0x7e0a9335
 0,         13,         13,        1,   622080, 0x5f34759b
 0,         14,         14,        1,   622080, 0xfac498a6
 0,         15,         15,        1,   622080, 0xe60e7a9e
 0,         16,         16,        1,   622080, 0x44875bbd
 0,         17,         17,        1,   622080, 0xfa761aab
 0,         18,         18,        1,   622080, 0x59be119c
 0,         19,         19,        1,   622080, 0x21316b36
 0,         20,         20,        1,   622080, 0x929fde5b
 0,         21,         21,        1,   622080, 0xfca8990c
 0,         22,         22,        1,   622080, 0x1ec87d02
 0,         23,         23,        1,   622080, 0x5768eea0
 0,         24,         24,        1,   622080, 0x1a0894ab
 0,         25,         25,        1,   622080, 0xb4e61323
 0,         26,         26,        1,   622080, 0xb773341a
 0,         27,         27,        1,   622080, 0x8a914cf7
 0,         28,         28,        1,   622080, 0xf1cfbc7d
 0,         29,         29,        1,   622080, 0xebaeb317
 0,         30,         30,        1,   622080, 0xbae9adf4
 0,         31,         31,        1,   622080, 0x593544fd
 0,         32,         32,        1,   622080, 0x2cd8ec0b
 0,         33,         33,        1,   622080, 0x8032d9d4
 0,         34,         34,        1,   622080, 0x5c67ace7
 0,         35,         35,        1,   622080, 0x95714528
 0,         36,         36,        1,   622080, 0xa11cbed2
 0,         37,         37,        1,   622080, 0x7389f8f1
 0,         38,         38,        1,   622080, 0xa694f3f2
 0,         39,         39,        1,   622080, 0xac3a3d09
--- a/tests/ref/fate/filter-yadif-mode1
+++ b/tests/ref/fate/filter-yadif-mode1
@ -0,0 +1,34 @@
 #tb 0: 1/25
 0,          9,          9,        1,   622080, 0x1511cae9
 0,         10,         10,        1,   622080, 0xb88ca855
 0,         11,         11,        1,   622080, 0x6e77e746
 0,         12,         12,        1,   622080, 0x5da19198
 0,         13,         13,        1,   622080, 0xee31c8a8
 0,         14,         14,        1,   622080, 0xcbb7aac5
 0,         15,         15,        1,   622080, 0x19972f1a
 0,         16,         16,        1,   622080, 0xac7d34b9
 0,         17,         17,        1,   622080, 0x4adfe592
 0,         18,         18,        1,   622080, 0x5d738330
 0,         19,         19,        1,   622080, 0xb60b4447
 0,         20,         20,        1,   622080, 0x1e11acf4
 0,         21,         21,        1,   622080, 0x5ed635d0
 0,         22,         22,        1,   622080, 0x939857af
 0,         23,         23,        1,   622080, 0x530b28fd
 0,         24,         24,        1,   622080, 0x3bc0d5d3
 0,         25,         25,        1,   622080, 0x77e0fe99
 0,         26,         26,        1,   622080, 0xd2151c1e
 0,         27,         27,        1,   622080, 0xe021a815
 0,         28,         28,        1,   622080, 0xceae4f12
 0,         29,         29,        1,   622080, 0x4c2f3330
 0,         30,         30,        1,   622080, 0xf534c392
 0,         31,         31,        1,   622080, 0x88f01c11
 0,         32,         32,        1,   622080, 0x654d5df2
 0,         33,         33,        1,   622080, 0x89ef6f8a
 0,         34,         34,        1,   622080, 0x78a7b5f1
 0,         35,         35,        1,   622080, 0x8152d67f
 0,         36,         36,        1,   622080, 0x6590ff5f
 0,         37,         37,        1,   622080, 0x51d2be96
 0,         38,         38,        1,   622080, 0x483f65f7
 0,         39,         39,        1,   622080, 0x7a69143d
 0,         40,         40,        1,   622080, 0xeccc58ff
 0,         41,         41,        1,   622080, 0xc4d2c370