2012-01-27 18:21:55 +03:00
|
|
|
;******************************************************************************
|
|
|
|
;* x86 optimizations for PNG decoding
|
|
|
|
;*
|
|
|
|
;* Copyright (c) 2008 Loren Merritt <lorenm@u.washington.edu>
|
2012-01-27 18:28:28 +03:00
|
|
|
;* Copyright (c) 2012 Ronald S. Bultje <rsbultje@gmail.com>
|
2012-01-27 18:21:55 +03:00
|
|
|
;*
|
2013-08-16 00:12:51 +03:00
|
|
|
;* This file is part of FFmpeg.
|
2012-01-27 18:21:55 +03:00
|
|
|
;*
|
2013-08-16 00:12:51 +03:00
|
|
|
;* FFmpeg is free software; you can redistribute it and/or
|
2012-01-27 18:21:55 +03:00
|
|
|
;* modify it under the terms of the GNU Lesser General Public
|
|
|
|
;* License as published by the Free Software Foundation; either
|
|
|
|
;* version 2.1 of the License, or (at your option) any later version.
|
|
|
|
;*
|
2013-08-16 00:12:51 +03:00
|
|
|
;* FFmpeg is distributed in the hope that it will be useful,
|
2012-01-27 18:21:55 +03:00
|
|
|
;* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
;* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
|
|
;* Lesser General Public License for more details.
|
|
|
|
;*
|
|
|
|
;* You should have received a copy of the GNU Lesser General Public
|
2013-08-16 00:12:51 +03:00
|
|
|
;* License along with FFmpeg; if not, write to the Free Software
|
2012-01-27 18:21:55 +03:00
|
|
|
;* 51, Inc., Foundation Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
|
|
|
;******************************************************************************
|
|
|
|
|
2012-07-15 12:48:21 +03:00
|
|
|
%include "libavutil/x86/x86util.asm"
|
2012-01-27 18:21:55 +03:00
|
|
|
|
|
|
|
SECTION_RODATA
|
|
|
|
|
|
|
|
cextern pw_255
|
|
|
|
|
2015-05-27 21:38:14 +02:00
|
|
|
SECTION .text
|
2012-01-27 18:21:55 +03:00
|
|
|
|
|
|
|
; %1 = nr. of xmm registers used
|
|
|
|
%macro ADD_BYTES_FN 1
|
|
|
|
cglobal add_bytes_l2, 4, 6, %1, dst, src1, src2, wa, w, i
|
|
|
|
%if ARCH_X86_64
|
|
|
|
movsxd waq, wad
|
|
|
|
%endif
|
|
|
|
xor iq, iq
|
|
|
|
|
|
|
|
; vector loop
|
|
|
|
mov wq, waq
|
|
|
|
and waq, ~(mmsize*2-1)
|
|
|
|
jmp .end_v
|
|
|
|
.loop_v:
|
2014-12-02 15:31:49 +02:00
|
|
|
movu m0, [src2q+iq]
|
|
|
|
movu m1, [src2q+iq+mmsize]
|
|
|
|
paddb m0, [src1q+iq]
|
|
|
|
paddb m1, [src1q+iq+mmsize]
|
|
|
|
movu [dstq+iq ], m0
|
|
|
|
movu [dstq+iq+mmsize], m1
|
2012-01-27 18:21:55 +03:00
|
|
|
add iq, mmsize*2
|
|
|
|
.end_v:
|
|
|
|
cmp iq, waq
|
|
|
|
jl .loop_v
|
|
|
|
|
2012-01-27 18:23:05 +03:00
|
|
|
%if mmsize == 16
|
|
|
|
; vector loop
|
2012-02-07 23:51:21 +03:00
|
|
|
mov waq, wq
|
2012-01-27 18:23:05 +03:00
|
|
|
and waq, ~7
|
|
|
|
jmp .end_l
|
|
|
|
.loop_l:
|
|
|
|
movq mm0, [src1q+iq]
|
|
|
|
paddb mm0, [src2q+iq]
|
|
|
|
movq [dstq+iq ], mm0
|
|
|
|
add iq, 8
|
|
|
|
.end_l:
|
|
|
|
cmp iq, waq
|
|
|
|
jl .loop_l
|
|
|
|
%endif
|
|
|
|
|
2012-01-27 18:21:55 +03:00
|
|
|
; scalar loop for leftover
|
|
|
|
jmp .end_s
|
|
|
|
.loop_s:
|
|
|
|
mov wab, [src1q+iq]
|
|
|
|
add wab, [src2q+iq]
|
|
|
|
mov [dstq+iq], wab
|
|
|
|
inc iq
|
|
|
|
.end_s:
|
|
|
|
cmp iq, wq
|
|
|
|
jl .loop_s
|
|
|
|
REP_RET
|
|
|
|
%endmacro
|
|
|
|
|
2012-01-27 18:23:05 +03:00
|
|
|
%if ARCH_X86_32
|
2012-01-27 18:21:55 +03:00
|
|
|
INIT_MMX mmx
|
|
|
|
ADD_BYTES_FN 0
|
2012-01-27 18:23:05 +03:00
|
|
|
%endif
|
|
|
|
|
|
|
|
INIT_XMM sse2
|
|
|
|
ADD_BYTES_FN 2
|
2012-01-27 18:21:55 +03:00
|
|
|
|
|
|
|
%macro ADD_PAETH_PRED_FN 1
|
|
|
|
cglobal add_png_paeth_prediction, 5, 7, %1, dst, src, top, w, bpp, end, cntr
|
|
|
|
%if ARCH_X86_64
|
|
|
|
movsxd bppq, bppd
|
|
|
|
movsxd wq, wd
|
|
|
|
%endif
|
|
|
|
lea endq, [dstq+wq-(mmsize/2-1)]
|
|
|
|
sub topq, dstq
|
|
|
|
sub srcq, dstq
|
|
|
|
sub dstq, bppq
|
|
|
|
pxor m7, m7
|
2012-01-27 18:28:28 +03:00
|
|
|
|
|
|
|
PUSH dstq
|
|
|
|
lea cntrq, [bppq-1]
|
|
|
|
shr cntrq, 2 + mmsize/16
|
|
|
|
.bpp_loop:
|
|
|
|
lea dstq, [dstq+cntrq*(mmsize/2)]
|
2012-01-27 18:21:55 +03:00
|
|
|
movh m0, [dstq]
|
|
|
|
movh m1, [topq+dstq]
|
|
|
|
punpcklbw m0, m7
|
|
|
|
punpcklbw m1, m7
|
|
|
|
add dstq, bppq
|
|
|
|
.loop:
|
|
|
|
mova m2, m1
|
|
|
|
movh m1, [topq+dstq]
|
|
|
|
mova m3, m2
|
|
|
|
punpcklbw m1, m7
|
|
|
|
mova m4, m2
|
|
|
|
psubw m3, m1
|
|
|
|
psubw m4, m0
|
|
|
|
mova m5, m3
|
|
|
|
paddw m5, m4
|
|
|
|
%if cpuflag(ssse3)
|
|
|
|
pabsw m3, m3
|
|
|
|
pabsw m4, m4
|
|
|
|
pabsw m5, m5
|
|
|
|
%else ; !cpuflag(ssse3)
|
|
|
|
psubw m7, m5
|
|
|
|
pmaxsw m5, m7
|
|
|
|
pxor m6, m6
|
|
|
|
pxor m7, m7
|
|
|
|
psubw m6, m3
|
|
|
|
psubw m7, m4
|
|
|
|
pmaxsw m3, m6
|
|
|
|
pmaxsw m4, m7
|
|
|
|
pxor m7, m7
|
|
|
|
%endif ; cpuflag(ssse3)
|
|
|
|
mova m6, m4
|
|
|
|
pminsw m6, m5
|
|
|
|
pcmpgtw m3, m6
|
|
|
|
pcmpgtw m4, m5
|
|
|
|
mova m6, m4
|
|
|
|
pand m4, m3
|
|
|
|
pandn m6, m3
|
|
|
|
pandn m3, m0
|
|
|
|
movh m0, [srcq+dstq]
|
|
|
|
pand m6, m1
|
|
|
|
pand m2, m4
|
|
|
|
punpcklbw m0, m7
|
|
|
|
paddw m0, m6
|
|
|
|
paddw m3, m2
|
|
|
|
paddw m0, m3
|
|
|
|
pand m0, [pw_255]
|
|
|
|
mova m3, m0
|
|
|
|
packuswb m3, m3
|
|
|
|
movh [dstq], m3
|
|
|
|
add dstq, bppq
|
|
|
|
cmp dstq, endq
|
2014-12-07 13:28:37 +02:00
|
|
|
jl .loop
|
2012-01-27 18:28:28 +03:00
|
|
|
|
|
|
|
mov dstq, [rsp]
|
|
|
|
dec cntrq
|
|
|
|
jge .bpp_loop
|
|
|
|
POP dstq
|
|
|
|
RET
|
2012-01-27 18:21:55 +03:00
|
|
|
%endmacro
|
|
|
|
|
2012-07-10 01:04:18 +03:00
|
|
|
INIT_MMX mmxext
|
2012-01-27 18:21:55 +03:00
|
|
|
ADD_PAETH_PRED_FN 0
|
|
|
|
|
|
|
|
INIT_MMX ssse3
|
|
|
|
ADD_PAETH_PRED_FN 0
|