#include "libavutil/arm/asm.S"
function ff_decode_hf_neon, export=1
push {r4-r5,lr}
add r2, r2, r3
ldr r3, [sp, #12]
ldrd r4, r5, [sp, #16]
add r3, r3, r4, lsl #3
add r1, r1, r4, lsl #2
add r0, r0, r4, lsl #5
1: ldr_post lr, r1, #4
add r4, r4, #1
add lr, r2, lr, lsl #5
cmp r4, r5
vld1.32 {d7}, [r3]!
vld1.8 {d0}, [lr,:64]
vcvt.f32.s32 d7, d7, #4
vmovl.s8 q1, d0
vmovl.s16 q0, d2
vmovl.s16 q1, d3
vcvt.f32.s32 q0, q0
vcvt.f32.s32 q1, q1
vmul.f32 q0, q0, d7[0]
vmul.f32 q1, q1, d7[0]
vst1.32 {q0-q1}, [r0,:128]!
bne 1b
pop {r4-r5,pc}
endfunc
function ff_dca_lfe_fir0_neon, export=1
push {r4-r6,lr}
mov r3, #32 @ decifactor
mov r6, #256/32
b dca_lfe_fir
endfunc
function ff_dca_lfe_fir1_neon, export=1
push {r4-r6,lr}
mov r3, #64 @ decifactor
mov r6, #256/64
dca_lfe_fir:
add r4, r0, r3, lsl #2 @ out2
add r5, r2, #256*4-16 @ cf1
sub r1, r1, #12
mov lr, #-16
1:
vmov.f32 q2, #0.0 @ v0
vmov.f32 q3, #0.0 @ v1
mov r12, r6
2:
vld1.32 {q8}, [r2,:128]! @ cf0
vld1.32 {q9}, [r5,:128], lr @ cf1
vld1.32 {q1}, [r1], lr @ in
subs r12, r12, #4
vrev64.32 q10, q8
vmla.f32 q3, q1, q9
vmla.f32 d4, d2, d21
vmla.f32 d5, d3, d20
bne 2b
add r1, r1, r6, lsl #2
subs r3, r3, #1
vadd.f32 d4, d4, d5
vadd.f32 d6, d6, d7
vpadd.f32 d5, d4, d6
vst1.32 {d5[0]}, [r0,:32]!
vst1.32 {d5[1]}, [r4,:32]!
bne 1b
pop {r4-r6,pc}
endfunc