RepoMirrors
/
ffmpeg
mirror of https://git.ffmpeg.org/ffmpeg.git
1
0
Fork 0
Code Issues Projects Releases Wiki Activity

Merge commit '08b028c18dc31b6de741861b9555669dcca4d12a' 

* commit '08b028c18dc31b6de741861b9555669dcca4d12a':
  Remove INIT_AVX from x86inc.asm.
  Remove usage of INIT_AVX in h264_intrapred_10bit.asm.
  dv: fix indentation

Conflicts:
	libavformat/dv.c

Merged-by: Michael Niedermayer <michaelni@gmx.at>
This commit is contained in:
Michael Niedermayer 2012-10-30 15:22:13 +01:00
parent e79c3858b3 08b028c18d
commit 6f5f7f705d
3 changed files with 207 additions and 199 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

306
libavcodec/x86/h264_intrapred_10bit.asm
View File

@ -53,8 +53,8 @@ SECTION .text
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred4x4_down_right(pixel *src, const pixel *topright, int stride) ; void pred4x4_down_right(pixel *src, const pixel *topright, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
%macro PRED4x4_DR 1 %macro PRED4x4_DR 0
cglobal pred4x4_down_right_10_%1, 3,3 cglobal pred4x4_down_right_10, 3, 3
sub r0, r2 sub r0, r2
lea r1, [r0+r2*2] lea r1, [r0+r2*2]
movhps m1, [r1-8] movhps m1, [r1-8]
@ -79,21 +79,22 @@ cglobal pred4x4_down_right_10_%1, 3,3
RET RET
%endmacro %endmacro
INIT_XMM INIT_XMM sse2
%define PALIGNR PALIGNR_MMX %define PALIGNR PALIGNR_MMX
PRED4x4_DR sse2 PRED4x4_DR
INIT_XMM ssse3
%define PALIGNR PALIGNR_SSSE3 %define PALIGNR PALIGNR_SSSE3
PRED4x4_DR ssse3 PRED4x4_DR
%if HAVE_AVX_EXTERNAL %if HAVE_AVX_EXTERNAL
INIT_AVX INIT_XMM avx
PRED4x4_DR avx PRED4x4_DR
%endif %endif
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred4x4_vertical_right(pixel *src, const pixel *topright, int stride) ; void pred4x4_vertical_right(pixel *src, const pixel *topright, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
%macro PRED4x4_VR 1 %macro PRED4x4_VR 0
cglobal pred4x4_vertical_right_10_%1, 3,3,6 cglobal pred4x4_vertical_right_10, 3, 3, 6
sub r0, r2 sub r0, r2
lea r1, [r0+r2*2] lea r1, [r0+r2*2]
movq m5, [r0] ; ........t3t2t1t0 movq m5, [r0] ; ........t3t2t1t0
@ -119,21 +120,22 @@ cglobal pred4x4_vertical_right_10_%1, 3,3,6
RET RET
%endmacro %endmacro
INIT_XMM INIT_XMM sse2
%define PALIGNR PALIGNR_MMX %define PALIGNR PALIGNR_MMX
PRED4x4_VR sse2 PRED4x4_VR
INIT_XMM ssse3
%define PALIGNR PALIGNR_SSSE3 %define PALIGNR PALIGNR_SSSE3
PRED4x4_VR ssse3 PRED4x4_VR
%if HAVE_AVX_EXTERNAL %if HAVE_AVX_EXTERNAL
INIT_AVX INIT_XMM avx
PRED4x4_VR avx PRED4x4_VR
%endif %endif
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred4x4_horizontal_down(pixel *src, const pixel *topright, int stride) ; void pred4x4_horizontal_down(pixel *src, const pixel *topright, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
%macro PRED4x4_HD 1 %macro PRED4x4_HD 0
cglobal pred4x4_horizontal_down_10_%1, 3,3 cglobal pred4x4_horizontal_down_10, 3, 3
sub r0, r2 sub r0, r2
lea r1, [r0+r2*2] lea r1, [r0+r2*2]
movq m0, [r0-8] ; lt .. movq m0, [r0-8] ; lt ..
@ -162,14 +164,15 @@ cglobal pred4x4_horizontal_down_10_%1, 3,3
RET RET
%endmacro %endmacro
INIT_XMM INIT_XMM sse2
%define PALIGNR PALIGNR_MMX %define PALIGNR PALIGNR_MMX
PRED4x4_HD sse2 PRED4x4_HD
INIT_XMM ssse3
%define PALIGNR PALIGNR_SSSE3 %define PALIGNR PALIGNR_SSSE3
PRED4x4_HD ssse3 PRED4x4_HD
%if HAVE_AVX_EXTERNAL %if HAVE_AVX_EXTERNAL
INIT_AVX INIT_XMM avx
PRED4x4_HD avx PRED4x4_HD
%endif %endif
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
@ -192,8 +195,8 @@ PRED4x4_HD avx
HADDD %1, %2 HADDD %1, %2
%endmacro %endmacro
INIT_MMX INIT_MMX mmx2
cglobal pred4x4_dc_10_mmxext, 3,3 cglobal pred4x4_dc_10, 3, 3
sub r0, r2 sub r0, r2
lea r1, [r0+r2*2] lea r1, [r0+r2*2]
movq m2, [r0+r2*1-8] movq m2, [r0+r2*1-8]
@ -216,8 +219,8 @@ cglobal pred4x4_dc_10_mmxext, 3,3
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred4x4_down_left(pixel *src, const pixel *topright, int stride) ; void pred4x4_down_left(pixel *src, const pixel *topright, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
%macro PRED4x4_DL 1 %macro PRED4x4_DL 0
cglobal pred4x4_down_left_10_%1, 3,3 cglobal pred4x4_down_left_10, 3, 3
sub r0, r2 sub r0, r2
movq m0, [r0] movq m0, [r0]
movhps m0, [r1] movhps m0, [r1]
@ -236,18 +239,18 @@ cglobal pred4x4_down_left_10_%1, 3,3
RET RET
%endmacro %endmacro
INIT_XMM INIT_XMM sse2
PRED4x4_DL sse2 PRED4x4_DL
%if HAVE_AVX_EXTERNAL %if HAVE_AVX_EXTERNAL
INIT_AVX INIT_XMM avx
PRED4x4_DL avx PRED4x4_DL
%endif %endif
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred4x4_vertical_left(pixel *src, const pixel *topright, int stride) ; void pred4x4_vertical_left(pixel *src, const pixel *topright, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
%macro PRED4x4_VL 1 %macro PRED4x4_VL 0
cglobal pred4x4_vertical_left_10_%1, 3,3 cglobal pred4x4_vertical_left_10, 3, 3
sub r0, r2 sub r0, r2
movu m1, [r0] movu m1, [r0]
movhps m1, [r1] movhps m1, [r1]
@ -265,18 +268,18 @@ cglobal pred4x4_vertical_left_10_%1, 3,3
RET RET
%endmacro %endmacro
INIT_XMM INIT_XMM sse2
PRED4x4_VL sse2 PRED4x4_VL
%if HAVE_AVX_EXTERNAL %if HAVE_AVX_EXTERNAL
INIT_AVX INIT_XMM avx
PRED4x4_VL avx PRED4x4_VL
%endif %endif
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred4x4_horizontal_up(pixel *src, const pixel *topright, int stride) ; void pred4x4_horizontal_up(pixel *src, const pixel *topright, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
INIT_MMX INIT_MMX mmx2
cglobal pred4x4_horizontal_up_10_mmxext, 3,3 cglobal pred4x4_horizontal_up_10, 3, 3
sub r0, r2 sub r0, r2
lea r1, [r0+r2*2] lea r1, [r0+r2*2]
movq m0, [r0+r2*1-8] movq m0, [r0+r2*1-8]
@ -309,8 +312,8 @@ cglobal pred4x4_horizontal_up_10_mmxext, 3,3
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred8x8_vertical(pixel *src, int stride) ; void pred8x8_vertical(pixel *src, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
INIT_XMM INIT_XMM sse2
cglobal pred8x8_vertical_10_sse2, 2,2 cglobal pred8x8_vertical_10, 2, 2
sub r0, r1 sub r0, r1
mova m0, [r0] mova m0, [r0]
%rep 3 %rep 3
@ -325,8 +328,8 @@ cglobal pred8x8_vertical_10_sse2, 2,2
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred8x8_horizontal(pixel *src, int stride) ; void pred8x8_horizontal(pixel *src, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
INIT_XMM INIT_XMM sse2
cglobal pred8x8_horizontal_10_sse2, 2,3 cglobal pred8x8_horizontal_10, 2, 3
mov r2d, 4 mov r2d, 4
.loop: .loop:
movq m0, [r0+r1*0-8] movq m0, [r0+r1*0-8]
@ -355,8 +358,8 @@ cglobal pred8x8_horizontal_10_sse2, 2,3
%endif %endif
%endmacro %endmacro
%macro PRED8x8_DC 2 %macro PRED8x8_DC 1
cglobal pred8x8_dc_10_%1, 2,6 cglobal pred8x8_dc_10, 2, 6
sub r0, r1 sub r0, r1
pxor m4, m4 pxor m4, m4
movq m0, [r0+0] movq m0, [r0+0]
@ -372,7 +375,7 @@ cglobal pred8x8_dc_10_%1, 2,6
paddw m1, m3 paddw m1, m3
punpcklwd m0, m1 punpcklwd m0, m1
%endif %endif
%2 m2, m0, 00001110b %1 m2, m0, 00001110b
paddw m0, m2 paddw m0, m2
lea r5, [r1*3] lea r5, [r1*3]
@ -397,8 +400,8 @@ cglobal pred8x8_dc_10_%1, 2,6
punpcklwd m2, m3 punpcklwd m2, m3
punpckldq m0, m2 ; s0, s1, s2, s3 punpckldq m0, m2 ; s0, s1, s2, s3
%2 m3, m0, 11110110b ; s2, s1, s3, s3 %1 m3, m0, 11110110b ; s2, s1, s3, s3
%2 m0, m0, 01110100b ; s0, s1, s3, s1 %1 m0, m0, 01110100b ; s0, s1, s3, s1
paddw m0, m3 paddw m0, m3
psrlw m0, 2 psrlw m0, 2
pavgw m0, m4 ; s0+s2, s1, s3, s1+s3 pavgw m0, m4 ; s0+s2, s1, s3, s1+s3
@ -424,16 +427,16 @@ cglobal pred8x8_dc_10_%1, 2,6
RET RET
%endmacro %endmacro
INIT_MMX INIT_MMX mmx2
PRED8x8_DC mmxext, pshufw PRED8x8_DC pshufw
INIT_XMM INIT_XMM sse2
PRED8x8_DC sse2 , pshuflw PRED8x8_DC pshuflw
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred8x8_top_dc(pixel *src, int stride) ; void pred8x8_top_dc(pixel *src, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
INIT_XMM INIT_XMM sse2
cglobal pred8x8_top_dc_10_sse2, 2,4 cglobal pred8x8_top_dc_10, 2, 4
sub r0, r1 sub r0, r1
mova m0, [r0] mova m0, [r0]
pshuflw m1, m0, 0x4e pshuflw m1, m0, 0x4e
@ -459,8 +462,8 @@ cglobal pred8x8_top_dc_10_sse2, 2,4
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred8x8_plane(pixel *src, int stride) ; void pred8x8_plane(pixel *src, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
INIT_XMM INIT_XMM sse2
cglobal pred8x8_plane_10_sse2, 2,7,7 cglobal pred8x8_plane_10, 2, 7, 7
sub r0, r1 sub r0, r1
lea r2, [r1*3] lea r2, [r1*3]
lea r3, [r0+r1*4] lea r3, [r0+r1*4]
@ -522,8 +525,8 @@ cglobal pred8x8_plane_10_sse2, 2,7,7
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred8x8l_128_dc(pixel *src, int has_topleft, int has_topright, int stride) ; void pred8x8l_128_dc(pixel *src, int has_topleft, int has_topright, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
%macro PRED8x8L_128_DC 1 %macro PRED8x8L_128_DC 0
cglobal pred8x8l_128_dc_10_%1, 4,4 cglobal pred8x8l_128_dc_10, 4, 4
mova m0, [pw_512] ; (1<<(BIT_DEPTH-1)) mova m0, [pw_512] ; (1<<(BIT_DEPTH-1))
lea r1, [r3*3] lea r1, [r3*3]
lea r2, [r0+r3*4] lea r2, [r0+r3*4]
@ -538,16 +541,16 @@ cglobal pred8x8l_128_dc_10_%1, 4,4
RET RET
%endmacro %endmacro
INIT_MMX INIT_MMX mmx2
PRED8x8L_128_DC mmxext PRED8x8L_128_DC
INIT_XMM INIT_XMM sse2
PRED8x8L_128_DC sse2 PRED8x8L_128_DC
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred8x8l_top_dc(pixel *src, int has_topleft, int has_topright, int stride) ; void pred8x8l_top_dc(pixel *src, int has_topleft, int has_topright, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
%macro PRED8x8L_TOP_DC 1 %macro PRED8x8L_TOP_DC 0
cglobal pred8x8l_top_dc_10_%1, 4,4,6 cglobal pred8x8l_top_dc_10, 4, 4, 6
sub r0, r3 sub r0, r3
mova m0, [r0] mova m0, [r0]
shr r1d, 14 shr r1d, 14
@ -575,19 +578,19 @@ cglobal pred8x8l_top_dc_10_%1, 4,4,6
RET RET
%endmacro %endmacro
INIT_XMM INIT_XMM sse2
PRED8x8L_TOP_DC sse2 PRED8x8L_TOP_DC
%if HAVE_AVX_EXTERNAL %if HAVE_AVX_EXTERNAL
INIT_AVX INIT_XMM avx
PRED8x8L_TOP_DC avx PRED8x8L_TOP_DC
%endif %endif
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
;void pred8x8l_dc(pixel *src, int has_topleft, int has_topright, int stride) ;void pred8x8l_dc(pixel *src, int has_topleft, int has_topright, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
;TODO: see if scalar is faster ;TODO: see if scalar is faster
%macro PRED8x8L_DC 1 %macro PRED8x8L_DC 0
cglobal pred8x8l_dc_10_%1, 4,6,6 cglobal pred8x8l_dc_10, 4, 6, 6
sub r0, r3 sub r0, r3
lea r4, [r0+r3*4] lea r4, [r0+r3*4]
lea r5, [r3*3] lea r5, [r3*3]
@ -634,18 +637,18 @@ cglobal pred8x8l_dc_10_%1, 4,6,6
RET RET
%endmacro %endmacro
INIT_XMM INIT_XMM sse2
PRED8x8L_DC sse2 PRED8x8L_DC
%if HAVE_AVX_EXTERNAL %if HAVE_AVX_EXTERNAL
INIT_AVX INIT_XMM avx
PRED8x8L_DC avx PRED8x8L_DC
%endif %endif
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred8x8l_vertical(pixel *src, int has_topleft, int has_topright, int stride) ; void pred8x8l_vertical(pixel *src, int has_topleft, int has_topright, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
%macro PRED8x8L_VERTICAL 1 %macro PRED8x8L_VERTICAL 0
cglobal pred8x8l_vertical_10_%1, 4,4,6 cglobal pred8x8l_vertical_10, 4, 4, 6
sub r0, r3 sub r0, r3
mova m0, [r0] mova m0, [r0]
shr r1d, 14 shr r1d, 14
@ -669,18 +672,18 @@ cglobal pred8x8l_vertical_10_%1, 4,4,6
RET RET
%endmacro %endmacro
INIT_XMM INIT_XMM sse2
PRED8x8L_VERTICAL sse2 PRED8x8L_VERTICAL
%if HAVE_AVX_EXTERNAL %if HAVE_AVX_EXTERNAL
INIT_AVX INIT_XMM avx
PRED8x8L_VERTICAL avx PRED8x8L_VERTICAL
%endif %endif
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred8x8l_horizontal(uint8_t *src, int has_topleft, int has_topright, int stride) ; void pred8x8l_horizontal(uint8_t *src, int has_topleft, int has_topright, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
%macro PRED8x8L_HORIZONTAL 1 %macro PRED8x8L_HORIZONTAL 0
cglobal pred8x8l_horizontal_10_%1, 4,4,5 cglobal pred8x8l_horizontal_10, 4, 4, 5
mova m0, [r0-16] mova m0, [r0-16]
shr r1d, 14 shr r1d, 14
dec r1 dec r1
@ -723,21 +726,22 @@ cglobal pred8x8l_horizontal_10_%1, 4,4,5
RET RET
%endmacro %endmacro
INIT_XMM INIT_XMM sse2
%define PALIGNR PALIGNR_MMX %define PALIGNR PALIGNR_MMX
PRED8x8L_HORIZONTAL sse2 PRED8x8L_HORIZONTAL
INIT_XMM ssse3
%define PALIGNR PALIGNR_SSSE3 %define PALIGNR PALIGNR_SSSE3
PRED8x8L_HORIZONTAL ssse3 PRED8x8L_HORIZONTAL
%if HAVE_AVX_EXTERNAL %if HAVE_AVX_EXTERNAL
INIT_AVX INIT_XMM avx
PRED8x8L_HORIZONTAL avx PRED8x8L_HORIZONTAL
%endif %endif
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
;void pred8x8l_down_left(pixel *src, int has_topleft, int has_topright, int stride) ;void pred8x8l_down_left(pixel *src, int has_topleft, int has_topright, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
%macro PRED8x8L_DOWN_LEFT 1 %macro PRED8x8L_DOWN_LEFT 0
cglobal pred8x8l_down_left_10_%1, 4,4,7 cglobal pred8x8l_down_left_10, 4, 4, 7
sub r0, r3 sub r0, r3
mova m3, [r0] mova m3, [r0]
shr r1d, 14 shr r1d, 14
@ -792,23 +796,24 @@ cglobal pred8x8l_down_left_10_%1, 4,4,7
jmp .do_topright jmp .do_topright
%endmacro %endmacro
INIT_XMM INIT_XMM sse2
%define PALIGNR PALIGNR_MMX %define PALIGNR PALIGNR_MMX
PRED8x8L_DOWN_LEFT sse2 PRED8x8L_DOWN_LEFT
INIT_XMM ssse3
%define PALIGNR PALIGNR_SSSE3 %define PALIGNR PALIGNR_SSSE3
PRED8x8L_DOWN_LEFT ssse3 PRED8x8L_DOWN_LEFT
%if HAVE_AVX_EXTERNAL %if HAVE_AVX_EXTERNAL
INIT_AVX INIT_XMM avx
PRED8x8L_DOWN_LEFT avx PRED8x8L_DOWN_LEFT
%endif %endif
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
;void pred8x8l_down_right(pixel *src, int has_topleft, int has_topright, int stride) ;void pred8x8l_down_right(pixel *src, int has_topleft, int has_topright, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
%macro PRED8x8L_DOWN_RIGHT 1 %macro PRED8x8L_DOWN_RIGHT 0
; standard forbids this when has_topleft is false ; standard forbids this when has_topleft is false
; no need to check ; no need to check
cglobal pred8x8l_down_right_10_%1, 4,5,8 cglobal pred8x8l_down_right_10, 4, 5, 8
sub r0, r3 sub r0, r3
lea r4, [r0+r3*4] lea r4, [r0+r3*4]
lea r1, [r3*3] lea r1, [r3*3]
@ -867,22 +872,23 @@ cglobal pred8x8l_down_right_10_%1, 4,5,8
RET RET
%endmacro %endmacro
INIT_XMM INIT_XMM sse2
%define PALIGNR PALIGNR_MMX %define PALIGNR PALIGNR_MMX
PRED8x8L_DOWN_RIGHT sse2 PRED8x8L_DOWN_RIGHT
INIT_XMM ssse3
%define PALIGNR PALIGNR_SSSE3 %define PALIGNR PALIGNR_SSSE3
PRED8x8L_DOWN_RIGHT ssse3 PRED8x8L_DOWN_RIGHT
%if HAVE_AVX_EXTERNAL %if HAVE_AVX_EXTERNAL
INIT_AVX INIT_XMM avx
PRED8x8L_DOWN_RIGHT avx PRED8x8L_DOWN_RIGHT
%endif %endif
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred8x8l_vertical_right(pixel *src, int has_topleft, int has_topright, int stride) ; void pred8x8l_vertical_right(pixel *src, int has_topleft, int has_topright, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
%macro PRED8x8L_VERTICAL_RIGHT 1 %macro PRED8x8L_VERTICAL_RIGHT 0
; likewise with 8x8l_down_right ; likewise with 8x8l_down_right
cglobal pred8x8l_vertical_right_10_%1, 4,5,7 cglobal pred8x8l_vertical_right_10, 4, 5, 7
sub r0, r3 sub r0, r3
lea r4, [r0+r3*4] lea r4, [r0+r3*4]
lea r1, [r3*3] lea r1, [r3*3]
@ -938,21 +944,22 @@ cglobal pred8x8l_vertical_right_10_%1, 4,5,7
RET RET
%endmacro %endmacro
INIT_XMM INIT_XMM sse2
%define PALIGNR PALIGNR_MMX %define PALIGNR PALIGNR_MMX
PRED8x8L_VERTICAL_RIGHT sse2 PRED8x8L_VERTICAL_RIGHT
INIT_XMM ssse3
%define PALIGNR PALIGNR_SSSE3 %define PALIGNR PALIGNR_SSSE3
PRED8x8L_VERTICAL_RIGHT ssse3 PRED8x8L_VERTICAL_RIGHT
%if HAVE_AVX_EXTERNAL %if HAVE_AVX_EXTERNAL
INIT_AVX INIT_XMM avx
PRED8x8L_VERTICAL_RIGHT avx PRED8x8L_VERTICAL_RIGHT
%endif %endif
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred8x8l_horizontal_up(pixel *src, int has_topleft, int has_topright, int stride) ; void pred8x8l_horizontal_up(pixel *src, int has_topleft, int has_topright, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
%macro PRED8x8L_HORIZONTAL_UP 1 %macro PRED8x8L_HORIZONTAL_UP 0
cglobal pred8x8l_horizontal_up_10_%1, 4,4,6 cglobal pred8x8l_horizontal_up_10, 4, 4, 6
mova m0, [r0+r3*0-16] mova m0, [r0+r3*0-16]
punpckhwd m0, [r0+r3*1-16] punpckhwd m0, [r0+r3*1-16]
shr r1d, 14 shr r1d, 14
@ -1000,14 +1007,15 @@ cglobal pred8x8l_horizontal_up_10_%1, 4,4,6
RET RET
%endmacro %endmacro
INIT_XMM INIT_XMM sse2
%define PALIGNR PALIGNR_MMX %define PALIGNR PALIGNR_MMX
PRED8x8L_HORIZONTAL_UP sse2 PRED8x8L_HORIZONTAL_UP
INIT_XMM ssse3
%define PALIGNR PALIGNR_SSSE3 %define PALIGNR PALIGNR_SSSE3
PRED8x8L_HORIZONTAL_UP ssse3 PRED8x8L_HORIZONTAL_UP
%if HAVE_AVX_EXTERNAL %if HAVE_AVX_EXTERNAL
INIT_AVX INIT_XMM avx
PRED8x8L_HORIZONTAL_UP avx PRED8x8L_HORIZONTAL_UP
%endif %endif
@ -1023,8 +1031,8 @@ PRED8x8L_HORIZONTAL_UP avx
%endif %endif
%endmacro %endmacro
%macro PRED16x16_VERTICAL 1 %macro PRED16x16_VERTICAL 0
cglobal pred16x16_vertical_10_%1, 2,3 cglobal pred16x16_vertical_10, 2, 3
sub r0, r1 sub r0, r1
mov r2d, 8 mov r2d, 8
mova m0, [r0+ 0] mova m0, [r0+ 0]
@ -1042,16 +1050,16 @@ cglobal pred16x16_vertical_10_%1, 2,3
REP_RET REP_RET
%endmacro %endmacro
INIT_MMX INIT_MMX mmx2
PRED16x16_VERTICAL mmxext PRED16x16_VERTICAL
INIT_XMM INIT_XMM sse2
PRED16x16_VERTICAL sse2 PRED16x16_VERTICAL
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred16x16_horizontal(pixel *src, int stride) ; void pred16x16_horizontal(pixel *src, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
%macro PRED16x16_HORIZONTAL 1 %macro PRED16x16_HORIZONTAL 0
cglobal pred16x16_horizontal_10_%1, 2,3 cglobal pred16x16_horizontal_10, 2, 3
mov r2d, 8 mov r2d, 8
.vloop: .vloop:
movd m0, [r0+r1*0-4] movd m0, [r0+r1*0-4]
@ -1066,16 +1074,16 @@ cglobal pred16x16_horizontal_10_%1, 2,3
REP_RET REP_RET
%endmacro %endmacro
INIT_MMX INIT_MMX mmx2
PRED16x16_HORIZONTAL mmxext PRED16x16_HORIZONTAL
INIT_XMM INIT_XMM sse2
PRED16x16_HORIZONTAL sse2 PRED16x16_HORIZONTAL
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred16x16_dc(pixel *src, int stride) ; void pred16x16_dc(pixel *src, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
%macro PRED16x16_DC 1 %macro PRED16x16_DC 0
cglobal pred16x16_dc_10_%1, 2,6 cglobal pred16x16_dc_10, 2, 6
mov r5, r0 mov r5, r0
sub r0, r1 sub r0, r1
mova m0, [r0+0] mova m0, [r0+0]
@ -1112,16 +1120,16 @@ cglobal pred16x16_dc_10_%1, 2,6
REP_RET REP_RET
%endmacro %endmacro
INIT_MMX INIT_MMX mmx2
PRED16x16_DC mmxext PRED16x16_DC
INIT_XMM INIT_XMM sse2
PRED16x16_DC sse2 PRED16x16_DC
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred16x16_top_dc(pixel *src, int stride) ; void pred16x16_top_dc(pixel *src, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
%macro PRED16x16_TOP_DC 1 %macro PRED16x16_TOP_DC 0
cglobal pred16x16_top_dc_10_%1, 2,3 cglobal pred16x16_top_dc_10, 2, 3
sub r0, r1 sub r0, r1
mova m0, [r0+0] mova m0, [r0+0]
paddw m0, [r0+mmsize] paddw m0, [r0+mmsize]
@ -1144,16 +1152,16 @@ cglobal pred16x16_top_dc_10_%1, 2,3
REP_RET REP_RET
%endmacro %endmacro
INIT_MMX INIT_MMX mmx2
PRED16x16_TOP_DC mmxext PRED16x16_TOP_DC
INIT_XMM INIT_XMM sse2
PRED16x16_TOP_DC sse2 PRED16x16_TOP_DC
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred16x16_left_dc(pixel *src, int stride) ; void pred16x16_left_dc(pixel *src, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
%macro PRED16x16_LEFT_DC 1 %macro PRED16x16_LEFT_DC 0
cglobal pred16x16_left_dc_10_%1, 2,6 cglobal pred16x16_left_dc_10, 2, 6
mov r5, r0 mov r5, r0
sub r0, 2 sub r0, 2
@ -1181,16 +1189,16 @@ cglobal pred16x16_left_dc_10_%1, 2,6
REP_RET REP_RET
%endmacro %endmacro
INIT_MMX INIT_MMX mmx2
PRED16x16_LEFT_DC mmxext PRED16x16_LEFT_DC
INIT_XMM INIT_XMM sse2
PRED16x16_LEFT_DC sse2 PRED16x16_LEFT_DC
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
; void pred16x16_128_dc(pixel *src, int stride) ; void pred16x16_128_dc(pixel *src, int stride)
;----------------------------------------------------------------------------- ;-----------------------------------------------------------------------------
%macro PRED16x16_128_DC 1 %macro PRED16x16_128_DC 0
cglobal pred16x16_128_dc_10_%1, 2,3 cglobal pred16x16_128_dc_10, 2,3
mova m0, [pw_512] mova m0, [pw_512]
mov r2d, 8 mov r2d, 8
.loop: .loop:
@ -1202,7 +1210,7 @@ cglobal pred16x16_128_dc_10_%1, 2,3
REP_RET REP_RET
%endmacro %endmacro
INIT_MMX INIT_MMX mmx2
PRED16x16_128_DC mmxext PRED16x16_128_DC
INIT_XMM INIT_XMM sse2
PRED16x16_128_DC sse2 PRED16x16_128_DC

40
libavcodec/x86/h264_intrapred_init.c
View File

@ -25,7 +25,7 @@
#define PRED4x4(TYPE, DEPTH, OPT) \ #define PRED4x4(TYPE, DEPTH, OPT) \
void ff_pred4x4_ ## TYPE ## _ ## DEPTH ## _ ## OPT (uint8_t *src, const uint8_t *topright, int stride); void ff_pred4x4_ ## TYPE ## _ ## DEPTH ## _ ## OPT (uint8_t *src, const uint8_t *topright, int stride);
PRED4x4(dc, 10, mmxext) PRED4x4(dc, 10, mmx2)
PRED4x4(down_left, 10, sse2) PRED4x4(down_left, 10, sse2)
PRED4x4(down_left, 10, avx) PRED4x4(down_left, 10, avx)
PRED4x4(down_right, 10, sse2) PRED4x4(down_right, 10, sse2)
@ -36,7 +36,7 @@ PRED4x4(vertical_left, 10, avx)
PRED4x4(vertical_right, 10, sse2) PRED4x4(vertical_right, 10, sse2)
PRED4x4(vertical_right, 10, ssse3) PRED4x4(vertical_right, 10, ssse3)
PRED4x4(vertical_right, 10, avx) PRED4x4(vertical_right, 10, avx)
PRED4x4(horizontal_up, 10, mmxext) PRED4x4(horizontal_up, 10, mmx2)
PRED4x4(horizontal_down, 10, sse2) PRED4x4(horizontal_down, 10, sse2)
PRED4x4(horizontal_down, 10, ssse3) PRED4x4(horizontal_down, 10, ssse3)
PRED4x4(horizontal_down, 10, avx) PRED4x4(horizontal_down, 10, avx)
@ -44,7 +44,7 @@ PRED4x4(horizontal_down, 10, avx)
#define PRED8x8(TYPE, DEPTH, OPT) \ #define PRED8x8(TYPE, DEPTH, OPT) \
void ff_pred8x8_ ## TYPE ## _ ## DEPTH ## _ ## OPT (uint8_t *src, int stride); void ff_pred8x8_ ## TYPE ## _ ## DEPTH ## _ ## OPT (uint8_t *src, int stride);
PRED8x8(dc, 10, mmxext) PRED8x8(dc, 10, mmx2)
PRED8x8(dc, 10, sse2) PRED8x8(dc, 10, sse2)
PRED8x8(top_dc, 10, sse2) PRED8x8(top_dc, 10, sse2)
PRED8x8(plane, 10, sse2) PRED8x8(plane, 10, sse2)
@ -56,7 +56,7 @@ void ff_pred8x8l_ ## TYPE ## _ ## DEPTH ## _ ## OPT (uint8_t *src, int has_tople
PRED8x8L(dc, 10, sse2) PRED8x8L(dc, 10, sse2)
PRED8x8L(dc, 10, avx) PRED8x8L(dc, 10, avx)
PRED8x8L(128_dc, 10, mmxext) PRED8x8L(128_dc, 10, mmx2)
PRED8x8L(128_dc, 10, sse2) PRED8x8L(128_dc, 10, sse2)
PRED8x8L(top_dc, 10, sse2) PRED8x8L(top_dc, 10, sse2)
PRED8x8L(top_dc, 10, avx) PRED8x8L(top_dc, 10, avx)
@ -81,17 +81,17 @@ PRED8x8L(horizontal_up, 10, avx)
#define PRED16x16(TYPE, DEPTH, OPT)\ #define PRED16x16(TYPE, DEPTH, OPT)\
void ff_pred16x16_ ## TYPE ## _ ## DEPTH ## _ ## OPT (uint8_t *src, int stride); void ff_pred16x16_ ## TYPE ## _ ## DEPTH ## _ ## OPT (uint8_t *src, int stride);
PRED16x16(dc, 10, mmxext) PRED16x16(dc, 10, mmx2)
PRED16x16(dc, 10, sse2) PRED16x16(dc, 10, sse2)
PRED16x16(top_dc, 10, mmxext) PRED16x16(top_dc, 10, mmx2)
PRED16x16(top_dc, 10, sse2) PRED16x16(top_dc, 10, sse2)
PRED16x16(128_dc, 10, mmxext) PRED16x16(128_dc, 10, mmx2)
PRED16x16(128_dc, 10, sse2) PRED16x16(128_dc, 10, sse2)
PRED16x16(left_dc, 10, mmxext) PRED16x16(left_dc, 10, mmx2)
PRED16x16(left_dc, 10, sse2) PRED16x16(left_dc, 10, sse2)
PRED16x16(vertical, 10, mmxext) PRED16x16(vertical, 10, mmx2)
PRED16x16(vertical, 10, sse2) PRED16x16(vertical, 10, sse2)
PRED16x16(horizontal, 10, mmxext) PRED16x16(horizontal, 10, mmx2)
PRED16x16(horizontal, 10, sse2) PRED16x16(horizontal, 10, sse2)
void ff_pred16x16_vertical_mmx (uint8_t *src, int stride); void ff_pred16x16_vertical_mmx (uint8_t *src, int stride);
@ -309,20 +309,20 @@ void ff_h264_pred_init_x86(H264PredContext *h, int codec_id, const int bit_depth
} }
} else if (bit_depth == 10) { } else if (bit_depth == 10) {
if (EXTERNAL_MMXEXT(mm_flags)) { if (EXTERNAL_MMXEXT(mm_flags)) {
h->pred4x4[DC_PRED ] = ff_pred4x4_dc_10_mmxext; h->pred4x4[DC_PRED ] = ff_pred4x4_dc_10_mmx2;
h->pred4x4[HOR_UP_PRED ] = ff_pred4x4_horizontal_up_10_mmxext; h->pred4x4[HOR_UP_PRED ] = ff_pred4x4_horizontal_up_10_mmx2;
if (chroma_format_idc == 1) if (chroma_format_idc == 1)
h->pred8x8[DC_PRED8x8 ] = ff_pred8x8_dc_10_mmxext; h->pred8x8[DC_PRED8x8 ] = ff_pred8x8_dc_10_mmx2;
h->pred8x8l[DC_128_PRED ] = ff_pred8x8l_128_dc_10_mmxext; h->pred8x8l[DC_128_PRED ] = ff_pred8x8l_128_dc_10_mmx2;
h->pred16x16[DC_PRED8x8 ] = ff_pred16x16_dc_10_mmxext; h->pred16x16[DC_PRED8x8 ] = ff_pred16x16_dc_10_mmx2;
h->pred16x16[TOP_DC_PRED8x8 ] = ff_pred16x16_top_dc_10_mmxext; h->pred16x16[TOP_DC_PRED8x8 ] = ff_pred16x16_top_dc_10_mmx2;
h->pred16x16[DC_128_PRED8x8 ] = ff_pred16x16_128_dc_10_mmxext; h->pred16x16[DC_128_PRED8x8 ] = ff_pred16x16_128_dc_10_mmx2;
h->pred16x16[LEFT_DC_PRED8x8 ] = ff_pred16x16_left_dc_10_mmxext; h->pred16x16[LEFT_DC_PRED8x8 ] = ff_pred16x16_left_dc_10_mmx2;
h->pred16x16[VERT_PRED8x8 ] = ff_pred16x16_vertical_10_mmxext; h->pred16x16[VERT_PRED8x8 ] = ff_pred16x16_vertical_10_mmx2;
h->pred16x16[HOR_PRED8x8 ] = ff_pred16x16_horizontal_10_mmxext; h->pred16x16[HOR_PRED8x8 ] = ff_pred16x16_horizontal_10_mmx2;
} }
if (EXTERNAL_SSE2(mm_flags)) { if (EXTERNAL_SSE2(mm_flags)) {
h->pred4x4[DIAG_DOWN_LEFT_PRED ] = ff_pred4x4_down_left_10_sse2; h->pred4x4[DIAG_DOWN_LEFT_PRED ] = ff_pred4x4_down_left_10_sse2;

60
libavformat/dv.c
View File

@ -236,24 +236,24 @@ static int dv_extract_audio_info(DVDemuxContext* c, uint8_t* frame)
/* Dynamic handling of the audio streams in DV */ /* Dynamic handling of the audio streams in DV */
for (i = 0; i < ach; i++) { for (i = 0; i < ach; i++) {
if (!c->ast[i]) { if (!c->ast[i]) {
c->ast[i] = avformat_new_stream(c->fctx, NULL); c->ast[i] = avformat_new_stream(c->fctx, NULL);
if (!c->ast[i]) if (!c->ast[i])
break; break;
avpriv_set_pts_info(c->ast[i], 64, 1, 30000); avpriv_set_pts_info(c->ast[i], 64, 1, 30000);
c->ast[i]->codec->codec_type = AVMEDIA_TYPE_AUDIO; c->ast[i]->codec->codec_type = AVMEDIA_TYPE_AUDIO;
c->ast[i]->codec->codec_id = AV_CODEC_ID_PCM_S16LE; c->ast[i]->codec->codec_id = AV_CODEC_ID_PCM_S16LE;
av_init_packet(&c->audio_pkt[i]); av_init_packet(&c->audio_pkt[i]);
c->audio_pkt[i].size = 0; c->audio_pkt[i].size = 0;
c->audio_pkt[i].data = c->audio_buf[i]; c->audio_pkt[i].data = c->audio_buf[i];
c->audio_pkt[i].stream_index = c->ast[i]->index; c->audio_pkt[i].stream_index = c->ast[i]->index;
c->audio_pkt[i].flags |= AV_PKT_FLAG_KEY; c->audio_pkt[i].flags |= AV_PKT_FLAG_KEY;
} }
c->ast[i]->codec->sample_rate = dv_audio_frequency[freq]; c->ast[i]->codec->sample_rate = dv_audio_frequency[freq];
c->ast[i]->codec->channels = 2; c->ast[i]->codec->channels = 2;
c->ast[i]->codec->bit_rate = 2 * dv_audio_frequency[freq] * 16; c->ast[i]->codec->bit_rate = 2 * dv_audio_frequency[freq] * 16;
c->ast[i]->start_time = 0; c->ast[i]->start_time = 0;
} }
c->ach = i; c->ach = i;
@ -271,14 +271,14 @@ static int dv_extract_video_info(DVDemuxContext *c, uint8_t* frame)
avctx = c->vst->codec; avctx = c->vst->codec;
avpriv_set_pts_info(c->vst, 64, c->sys->time_base.num, avpriv_set_pts_info(c->vst, 64, c->sys->time_base.num,
c->sys->time_base.den); c->sys->time_base.den);
avctx->time_base= c->sys->time_base; avctx->time_base= c->sys->time_base;
/* finding out SAR is a little bit messy */ /* finding out SAR is a little bit messy */
vsc_pack = dv_extract_pack(frame, dv_video_control); vsc_pack = dv_extract_pack(frame, dv_video_control);
apt = frame[4] & 0x07; apt = frame[4] & 0x07;
is16_9 = (vsc_pack && ((vsc_pack[2] & 0x07) == 0x02 || is16_9 = (vsc_pack && ((vsc_pack[2] & 0x07) == 0x02 ||
(!apt && (vsc_pack[2] & 0x07) == 0x07))); (!apt && (vsc_pack[2] & 0x07) == 0x07)));
c->vst->sample_aspect_ratio = c->sys->sar[is16_9]; c->vst->sample_aspect_ratio = c->sys->sar[is16_9];
avctx->bit_rate = av_rescale_q(c->sys->frame_size, (AVRational){8,1}, avctx->bit_rate = av_rescale_q(c->sys->frame_size, (AVRational){8,1},
c->sys->time_base); c->sys->time_base);
@ -336,12 +336,12 @@ int avpriv_dv_get_packet(DVDemuxContext *c, AVPacket *pkt)
int i; int i;
for (i = 0; i < c->ach; i++) { for (i = 0; i < c->ach; i++) {
if (c->ast[i] && c->audio_pkt[i].size) { if (c->ast[i] && c->audio_pkt[i].size) {
*pkt = c->audio_pkt[i]; *pkt = c->audio_pkt[i];
c->audio_pkt[i].size = 0; c->audio_pkt[i].size = 0;
size = pkt->size; size = pkt->size;
break; break;
} }
} }
return size; return size;
@ -356,17 +356,17 @@ int avpriv_dv_produce_packet(DVDemuxContext *c, AVPacket *pkt,
if (buf_size < DV_PROFILE_BYTES || if (buf_size < DV_PROFILE_BYTES ||
!(c->sys = avpriv_dv_frame_profile(c->sys, buf, buf_size)) || !(c->sys = avpriv_dv_frame_profile(c->sys, buf, buf_size)) ||
buf_size < c->sys->frame_size) { buf_size < c->sys->frame_size) {
return -1; /* Broken frame, or not enough data */ return -1; /* Broken frame, or not enough data */
} }
/* Queueing audio packet */ /* Queueing audio packet */
/* FIXME: in case of no audio/bad audio we have to do something */ /* FIXME: in case of no audio/bad audio we have to do something */
size = dv_extract_audio_info(c, buf); size = dv_extract_audio_info(c, buf);
for (i = 0; i < c->ach; i++) { for (i = 0; i < c->ach; i++) {
c->audio_pkt[i].pos = pos; c->audio_pkt[i].pos = pos;
c->audio_pkt[i].size = size; c->audio_pkt[i].size = size;
c->audio_pkt[i].pts = c->abytes * 30000*8 / c->ast[i]->codec->bit_rate; c->audio_pkt[i].pts = c->abytes * 30000 * 8 / c->ast[i]->codec->bit_rate;
ppcm[i] = c->audio_buf[i]; ppcm[i] = c->audio_buf[i];
} }
if (c->ach) if (c->ach)
dv_extract_audio(buf, ppcm, c->sys); dv_extract_audio(buf, ppcm, c->sys);