;*****************************************************************************;* mc-a2.asm: h264 encoder library;*****************************************************************************;* Copyright (C) 2005-2008 x264 project;*;* Authors: Loren Merritt <lorenm@u.washington.edu>;*          Jason Garrett-Glaser <darkshikari@gmail.com>;*          Holger Lubitz <hal@duncan.ol.sub.de>;*          Mathieu Monnier <manao@melix.net>;*;* This program is free software; you can redistribute it and/or modify;* it under the terms of the GNU General Public License as published by;* the Free Software Foundation; either version 2 of the License, or;* (at your option) any later version.;*;* This program is distributed in the hope that it will be useful,;* but WITHOUT ANY WARRANTY; without even the implied warranty of;* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the;* GNU General Public License for more details.;*;* You should have received a copy of the GNU General Public License;* along with this program; if not, write to the Free Software;* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02111, USA.;*****************************************************************************%include "x86inc.asm"%include "x86util.asm"SECTION_RODATAfilt_mul20: times 16 db 20filt_mul51: times 8 db 1, -5pw_1:  times 8 dw 1pw_16: times 8 dw 16pw_32: times 8 dw 32SECTION .text%macro LOAD_ADD 4    movh       %4, %3    movh       %1, %2    punpcklbw  %4, m0    punpcklbw  %1, m0    paddw      %1, %4%endmacro%macro LOAD_ADD_2 6    mova       %5, %3    mova       %1, %4    mova       %6, %5    mova       %2, %1    punpcklbw  %5, m0    punpcklbw  %1, m0    punpckhbw  %6, m0    punpckhbw  %2, m0    paddw      %1, %5    paddw      %2, %6%endmacro%macro FILT_V2 0    psubw  m1, m2  ; a-b    psubw  m4, m5    psubw  m2, m3  ; b-c    psubw  m5, m6    psllw  m2, 2    psllw  m5, 2    psubw  m1, m2  ; a-5*b+4*c    psubw  m4, m5    psllw  m3, 4    psllw  m6, 4    paddw  m1, m3  ; a-5*b+20*c    paddw  m4, m6%endmacro%macro FILT_H 3    psubw  %1, %2  ; a-b    psraw  %1, 2   ; (a-b)/4    psubw  %1, %2  ; (a-b)/4-b    paddw  %1, %3  ; (a-b)/4-b+c    psraw  %1, 2   ; ((a-b)/4-b+c)/4    paddw  %1, %3  ; ((a-b)/4-b+c)/4+c = (a-5*b+20*c)/16%endmacro%macro FILT_H2 6    psubw  %1, %2    psubw  %4, %5    psraw  %1, 2    psraw  %4, 2    psubw  %1, %2    psubw  %4, %5    paddw  %1, %3    paddw  %4, %6    psraw  %1, 2    psraw  %4, 2    paddw  %1, %3    paddw  %4, %6%endmacro%macro FILT_PACK 3    paddw     %1, m7    paddw     %2, m7    psraw     %1, %3    psraw     %2, %3    packuswb  %1, %2%endmacroINIT_MMX%macro HPEL_V 1-2 0;-----------------------------------------------------------------------------; void x264_hpel_filter_v_mmxext( uint8_t *dst, uint8_t *src, int16_t *buf, int stride, int width );;-----------------------------------------------------------------------------cglobal x264_hpel_filter_v_%1, 5,6,%2%ifdef WIN64    movsxd   r4, r4d%endif    lea r5, [r1+r3]    sub r1, r3    sub r1, r3    add r0, r4    lea r2, [r2+r4*2]    neg r4%ifnidn %1, ssse3    pxor m0, m0%else    mova m0, [filt_mul51 GLOBAL]%endif.loop:%ifidn %1, ssse3    mova m1, [r1]    mova m4, [r1+r3]    mova m2, [r5+r3*2]    mova m5, [r5+r3]    mova m3, [r1+r3*2]    mova m6, [r5]    SBUTTERFLY bw, 1, 4, 7    SBUTTERFLY bw, 2, 5, 7    SBUTTERFLY bw, 3, 6, 7    pmaddubsw m1, m0    pmaddubsw m4, m0    pmaddubsw m2, m0    pmaddubsw m5, m0    pmaddubsw m3, [filt_mul20 GLOBAL]    pmaddubsw m6, [filt_mul20 GLOBAL]    paddw  m1, m2    paddw  m4, m5    paddw  m1, m3    paddw  m4, m6%else    LOAD_ADD_2 m1, m4, [r1     ], [r5+r3*2], m6, m7            ; a0 / a1    LOAD_ADD_2 m2, m5, [r1+r3  ], [r5+r3  ], m6, m7            ; b0 / b1    LOAD_ADD   m3,     [r1+r3*2], [r5     ], m7                ; c0    LOAD_ADD   m6,     [r1+r3*2+mmsize/2], [r5+mmsize/2], m7 ; c1    FILT_V2%endif    mova      m7, [pw_16 GLOBAL]    mova      [r2+r4*2], m1    mova      [r2+r4*2+mmsize], m4    paddw     m1, m7    paddw     m4, m7    psraw     m1, 5    psraw     m4, 5    packuswb  m1, m4    mova     [r0+r4], m1    add r1, mmsize    add r5, mmsize    add r4, mmsize    jl .loop    REP_RET%endmacroHPEL_V mmxext;-----------------------------------------------------------------------------; void x264_hpel_filter_c_mmxext( uint8_t *dst, int16_t *buf, int width );;-----------------------------------------------------------------------------cglobal x264_hpel_filter_c_mmxext, 3,3    add r0, r2    lea r1, [r1+r2*2]    neg r2    %define src r1+r2*2    movq m7, [pw_32 GLOBAL].loop:    movq   m1, [src-4]    movq   m2, [src-2]    movq   m3, [src  ]    movq   m4, [src+4]    movq   m5, [src+6]    paddw  m3, [src+2]  ; c0    paddw  m2, m4       ; b0    paddw  m1, m5       ; a0    movq   m6, [src+8]    paddw  m4, [src+14] ; a1    paddw  m5, [src+12] ; b1    paddw  m6, [src+10] ; c1    FILT_H2 m1, m2, m3, m4, m5, m6    FILT_PACK m1, m4, 6    movntq [r0+r2], m1    add r2, 8    jl .loop    REP_RET;-----------------------------------------------------------------------------; void x264_hpel_filter_h_mmxext( uint8_t *dst, uint8_t *src, int width );;-----------------------------------------------------------------------------cglobal x264_hpel_filter_h_mmxext, 3,3    add r0, r2    add r1, r2    neg r2    %define src r1+r2    pxor m0, m0.loop:    movd       m1, [src-2]    movd       m2, [src-1]    movd       m3, [src  ]    movd       m6, [src+1]    movd       m4, [src+2]    movd       m5, [src+3]    punpcklbw  m1, m0    punpcklbw  m2, m0    punpcklbw  m3, m0    punpcklbw  m6, m0    punpcklbw  m4, m0    punpcklbw  m5, m0    paddw      m3, m6 ; c0    paddw      m2, m4 ; b0    paddw      m1, m5 ; a0    movd       m7, [src+7]    movd       m6, [src+6]    punpcklbw  m7, m0    punpcklbw  m6, m0    paddw      m4, m7 ; c1    paddw      m5, m6 ; b1    movd       m7, [src+5]    movd       m6, [src+4]    punpcklbw  m7, m0    punpcklbw  m6, m0    paddw      m6, m7 ; a1    movq       m7, [pw_1 GLOBAL]    FILT_H2 m1, m2, m3, m4, m5, m6    FILT_PACK m1, m4, 1    movntq     [r0+r2], m1    add r2, 8    jl .loop    REP_RETINIT_XMM%macro HPEL_C 1;-----------------------------------------------------------------------------; void x264_hpel_filter_c_sse2( uint8_t *dst, int16_t *buf, int width );;-----------------------------------------------------------------------------cglobal x264_hpel_filter_c_%1, 3,3,9    add r0, r2    lea r1, [r1+r2*2]    neg r2    %define src r1+r2*2%ifidn %1, ssse3    mova    m7, [pw_32 GLOBAL]    %define tpw_32 m7%elifdef ARCH_X86_64    mova    m8, [pw_32 GLOBAL]    %define tpw_32 m8%else    %define tpw_32 [pw_32 GLOBAL]%endif.loop:%ifidn %1,sse2_misalign    movu    m0, [src-4]    movu    m1, [src-2]    mova    m2, [src]    paddw   m0, [src+6]    paddw   m1, [src+4]    paddw   m2, [src+2]%else    mova    m6, [src-16]    mova    m2, [src]    mova    m3, [src+16]    mova    m0, m2    mova    m1, m2    mova    m4, m3    mova    m5, m3    PALIGNR m3, m2, 2, m7    PALIGNR m4, m2, 4, m7    PALIGNR m5, m2, 6, m7    PALIGNR m0, m6, 12, m7    PALIGNR m1, m6, 14, m7    paddw   m2, m3    paddw   m1, m4    paddw   m0, m5%endif    FILT_H  m0, m1, m2    paddw   m0, tpw_32    psraw   m0, 6    packuswb m0, m0    movq [r0+r2], m0    add r2, 8    jl .loop    REP_RET%endmacro;-----------------------------------------------------------------------------; void x264_hpel_filter_h_sse2( uint8_t *dst, uint8_t *src, int width );;-----------------------------------------------------------------------------cglobal x264_hpel_filter_h_sse2, 3,3,8    add r0, r2    add r1, r2    neg r2    %define src r1+r2    pxor m0, m0.loop:    movh       m1, [src-2]    movh       m2, [src-1]    movh       m3, [src  ]    movh       m4, [src+1]    movh       m5, [src+2]    movh       m6, [src+3]    punpcklbw  m1, m0    punpcklbw  m2, m0    punpcklbw  m3, m0    punpcklbw  m4, m0    punpcklbw  m5, m0    punpcklbw  m6, m0    paddw      m3, m4 ; c0    paddw      m2, m5 ; b0    paddw      m1, m6 ; a0    movh       m4, [src+6]    movh       m5, [src+7]    movh       m6, [src+10]    movh       m7, [src+11]    punpcklbw  m4, m0    punpcklbw  m5, m0    punpcklbw  m6, m0    punpcklbw  m7, m0    paddw      m5, m6 ; b1    paddw      m4, m7 ; a1    movh       m6, [src+8]    movh       m7, [src+9]    punpcklbw  m6, m0    punpcklbw  m7, m0    paddw      m6, m7 ; c1    mova       m7, [pw_1 GLOBAL] ; FIXME xmm8    FILT_H2 m1, m2, m3, m4, m5, m6    FILT_PACK m1, m4, 1    movntdq    [r0+r2], m1    add r2, 16    jl .loop    REP_RET%ifndef ARCH_X86_64;-----------------------------------------------------------------------------; void x264_hpel_filter_h_ssse3( uint8_t *dst, uint8_t *src, int width );;-----------------------------------------------------------------------------cglobal x264_hpel_filter_h_ssse3, 3,3    add r0, r2    add r1, r2    neg r2    %define src r1+r2    pxor m0, m0    movh m1, [src-8]    punpcklbw m1, m0         ; 00 -1 00 -2 00 -3 00 -4 00 -5 00 -6 00 -7 00 -8    movh m2, [src]    punpcklbw m2, m0    mova       m7, [pw_1 GLOBAL].loop:    movh       m3, [src+8]    punpcklbw  m3, m0    mova       m4, m2    palignr    m2, m1, 14    mova       m5, m3    palignr    m3, m4, 4    paddw      m3, m2    mova       m2, m4    palignr    m4, m1, 12    mova       m1, m5    palignr    m5, m2, 6    paddw      m5, m4    mova       m4, m1    palignr    m1, m2, 2    paddw      m1, m2    FILT_H     m5, m3, m1    movh       m1, [src+16]    punpcklbw  m1, m0    mova       m3, m4    palignr    m4, m2, 14    mova       m6, m1    palignr    m1, m3, 4    paddw      m1, m4    mova       m4, m3    palignr    m3, m2, 12    mova       m2, m6    palignr    m6, m4, 6    paddw      m6, m3    mova       m3, m2    palignr    m2, m4, 2    paddw      m2, m4    FILT_H m6, m1, m2    FILT_PACK m5, m6, 1    movdqa    [r0+r2], m5    add r2, 16    mova      m2, m3    mova      m1, m4    jl .loop    REP_RET%endif%define PALIGNR PALIGNR_MMX%ifndef ARCH_X86_64HPEL_C sse2%endifHPEL_V sse2, 8HPEL_C sse2_misalign%define PALIGNR PALIGNR_SSSE3HPEL_C ssse3HPEL_V ssse3%ifdef ARCH_X86_64%macro DO_FILT_V 6%ifidn %6, ssse3    mova m1, [r3]    mova m2, [r3+r2]    mova %3, [r3+r2*2]    mova m3, [r1]    mova %4, [r1+r2]    mova m0, [r1+r2*2]    mova %2, [filt_mul51 GLOBAL]    mova m4, m1    punpcklbw m1, m2    punpckhbw m4, m2    mova m2, m0    punpcklbw m0, %4    punpckhbw m2, %4    mova %1, m3    punpcklbw m3, %3    punpckhbw %1, %3    mova %3, m3    mova %4, %1    pmaddubsw m1, %2    pmaddubsw m4, %2    pmaddubsw m0, %2    pmaddubsw m2, %2    pmaddubsw m3, [filt_mul20 GLOBAL]    pmaddubsw %1, [filt_mul20 GLOBAL]    psrlw     %3, 8    psrlw     %4, 8    paddw m1, m0    paddw m4, m2    paddw m1, m3    paddw m4, %1%else    LOAD_ADD_2 m1, m4, [r3     ], [r1+r2*2], m2, m5            ; a0 / a1    LOAD_ADD_2 m2, m5, [r3+r2  ], [r1+r2  ], m3, m6            ; b0 / b1    LOAD_ADD_2 m3, m6, [r3+r2*2], [r1     ], %3, %4            ; c0 / c1    FILT_V2%endif    mova      %1, m1    mova      %2, m4    paddw     m1, m15    paddw     m4, m15    add       r3, 16    add       r1, 16    psraw     m1, 5    psraw     m4, 5    packuswb  m1, m4    movntps  [r11+r4+%5], m1%endmacro%macro DO_FILT_H 4    mova      m1, %2    PALIGNR   m1, %1, 12, m4    mova      m2, %2    PALIGNR   m2, %1, 14, m4    mova      %1, %3    PALIGNR   %3, %2, 6, m4    mova      m3, %1    PALIGNR   m3, %2, 4, m4    mova      m4, %1    paddw     %3, m1    PALIGNR   m4, %2, 2, m1    paddw     m3, m2    paddw     m4, %2    FILT_H    %3, m3, m4    paddw     %3, m15    psraw     %3, %4%endmacro%macro DO_FILT_CC 4    DO_FILT_H %1, %2, %3, 6    DO_FILT_H %2, %1, %4, 6    packuswb  %3, %4    movntps   [r5+r4], %3%endmacro%macro DO_FILT_HH 4    DO_FILT_H %1, %2, %3, 1    DO_FILT_H %2, %1, %4, 1    packuswb  %3, %4    movntps   [r0+r4], %3%endmacro%macro DO_FILT_H2 6    DO_FILT_H %1, %2, %3, 6    psrlw    m15, 5    DO_FILT_H %4, %5, %6, 1    packuswb  %6, %3%endmacro