FFmpeg4/libavcodec/arm/vp8_armv6.S

249 lines
8.4 KiB
ArmAsm
Raw Normal View History

2023-07-02 12:20:28 +00:00
/*
* Copyright (C) 2010 Mans Rullgard
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "libavutil/arm/asm.S"
.macro rac_get_prob h, bs, buf, cw, pr, t0, t1
adds \bs, \bs, \t0
lsl \cw, \cw, \t0
lsl \t0, \h, \t0
rsb \h, \pr, #256
it cs
ldrhcs \t1, [\buf], #2
smlabb \h, \t0, \pr, \h
T itttt cs
rev16cs \t1, \t1
A orrcs \cw, \cw, \t1, lsl \bs
T lslcs \t1, \t1, \bs
T orrcs \cw, \cw, \t1
subcs \bs, \bs, #16
lsr \h, \h, #8
cmp \cw, \h, lsl #16
itt ge
subge \cw, \cw, \h, lsl #16
subge \h, \t0, \h
.endm
.macro rac_get_128 h, bs, buf, cw, t0, t1
adds \bs, \bs, \t0
lsl \cw, \cw, \t0
lsl \t0, \h, \t0
it cs
ldrhcs \t1, [\buf], #2
mov \h, #128
it cs
rev16cs \t1, \t1
add \h, \h, \t0, lsl #7
A orrcs \cw, \cw, \t1, lsl \bs
T ittt cs
T lslcs \t1, \t1, \bs
T orrcs \cw, \cw, \t1
subcs \bs, \bs, #16
lsr \h, \h, #8
cmp \cw, \h, lsl #16
itt ge
subge \cw, \cw, \h, lsl #16
subge \h, \t0, \h
.endm
function ff_decode_block_coeffs_armv6, export=1
push {r0,r1,r4-r11,lr}
movrelx lr, X(ff_vp56_norm_shift)
ldrd r4, r5, [sp, #44] @ token_prob, qmul
cmp r3, #0
ldr r11, [r5]
ldm r0, {r5-r7} @ high, bits, buf
it ne
pkhtbne r11, r11, r11, asr #16
ldr r8, [r0, #16] @ code_word
0:
ldrb r9, [lr, r5]
add r3, r3, #1
ldrb r0, [r4, #1]
rac_get_prob r5, r6, r7, r8, r0, r9, r10
blt 2f
ldrb r9, [lr, r5]
ldrb r0, [r4, #2]
rac_get_prob r5, r6, r7, r8, r0, r9, r10
ldrb r9, [lr, r5]
bge 3f
add r4, r3, r3, lsl #5
sxth r12, r11
add r4, r4, r2
adds r6, r6, r9
add r4, r4, #11
lsl r8, r8, r9
it cs
ldrhcs r10, [r7], #2
lsl r9, r5, r9
mov r5, #128
it cs
rev16cs r10, r10
add r5, r5, r9, lsl #7
T ittt cs
T lslcs r10, r10, r6
T orrcs r8, r8, r10
A orrcs r8, r8, r10, lsl r6
subcs r6, r6, #16
lsr r5, r5, #8
cmp r8, r5, lsl #16
movrel r10, zigzag_scan-1
itt ge
subge r8, r8, r5, lsl #16
subge r5, r9, r5
ldrb r10, [r10, r3]
it ge
rsbge r12, r12, #0
cmp r3, #16
strh r12, [r1, r10]
bge 6f
5:
ldrb r9, [lr, r5]
ldrb r0, [r4]
rac_get_prob r5, r6, r7, r8, r0, r9, r10
pkhtb r11, r11, r11, asr #16
bge 0b
6:
ldr r0, [sp]
ldr r9, [r0, #12]
cmp r7, r9
it hi
movhi r7, r9
stm r0, {r5-r7} @ high, bits, buf
str r8, [r0, #16] @ code_word
add sp, sp, #8
mov r0, r3
pop {r4-r11,pc}
2:
add r4, r3, r3, lsl #5
cmp r3, #16
add r4, r4, r2
pkhtb r11, r11, r11, asr #16
bne 0b
b 6b
3:
ldrb r0, [r4, #3]
rac_get_prob r5, r6, r7, r8, r0, r9, r10
ldrb r9, [lr, r5]
bge 1f
mov r12, #2
ldrb r0, [r4, #4]
rac_get_prob r5, r6, r7, r8, r0, r9, r10
it ge
addge r12, #1
ldrb r9, [lr, r5]
blt 4f
ldrb r0, [r4, #5]
rac_get_prob r5, r6, r7, r8, r0, r9, r10
it ge
addge r12, #1
ldrb r9, [lr, r5]
b 4f
1:
ldrb r0, [r4, #6]
rac_get_prob r5, r6, r7, r8, r0, r9, r10
ldrb r9, [lr, r5]
bge 3f
ldrb r0, [r4, #7]
rac_get_prob r5, r6, r7, r8, r0, r9, r10
ldrb r9, [lr, r5]
bge 2f
mov r12, #5
mov r0, #159
rac_get_prob r5, r6, r7, r8, r0, r9, r10
it ge
addge r12, r12, #1
ldrb r9, [lr, r5]
b 4f
2:
mov r12, #7
mov r0, #165
rac_get_prob r5, r6, r7, r8, r0, r9, r10
it ge
addge r12, r12, #2
ldrb r9, [lr, r5]
mov r0, #145
rac_get_prob r5, r6, r7, r8, r0, r9, r10
it ge
addge r12, r12, #1
ldrb r9, [lr, r5]
b 4f
3:
ldrb r0, [r4, #8]
rac_get_prob r5, r6, r7, r8, r0, r9, r10
it ge
addge r4, r4, #1
ldrb r9, [lr, r5]
ite ge
movge r12, #2
movlt r12, #0
ldrb r0, [r4, #9]
rac_get_prob r5, r6, r7, r8, r0, r9, r10
mov r9, #8
it ge
addge r12, r12, #1
movrelx r4, X(ff_vp8_dct_cat_prob), r1
lsl r9, r9, r12
ldr r4, [r4, r12, lsl #2]
add r12, r9, #3
mov r1, #0
ldrb r0, [r4], #1
1:
ldrb r9, [lr, r5]
lsl r1, r1, #1
rac_get_prob r5, r6, r7, r8, r0, r9, r10
ldrb r0, [r4], #1
it ge
addge r1, r1, #1
cmp r0, #0
bne 1b
ldrb r9, [lr, r5]
add r12, r12, r1
ldr r1, [sp, #4]
4:
add r4, r3, r3, lsl #5
add r4, r4, r2
add r4, r4, #22
rac_get_128 r5, r6, r7, r8, r9, r10
it ge
rsbge r12, r12, #0
smulbb r12, r12, r11
movrel r9, zigzag_scan-1
ldrb r9, [r9, r3]
cmp r3, #16
strh r12, [r1, r9]
bge 6b
b 5b
endfunc
const zigzag_scan
.byte 0, 2, 8, 16
.byte 10, 4, 6, 12
.byte 18, 24, 26, 20
.byte 14, 22, 28, 30
endconst