-rw-r--r-- 13940 libmceliece-20230612/crypto_kem/348864/avx/syndrome_asm.S raw
// 20221231 djb: port hidden to macos; tnx thom wiggers
// 20221230 djb: add linker line

// linker define syndrome_asm

#include "crypto_asm_hidden.h"

#define syndrome_asm CRYPTO_SHARED_NAMESPACE(syndrome_asm)
#define _syndrome_asm _CRYPTO_SHARED_NAMESPACE(syndrome_asm)

# qhasm: int64 input_0

# qhasm: int64 input_1

# qhasm: int64 input_2

# qhasm: int64 input_3

# qhasm: int64 input_4

# qhasm: int64 input_5

# qhasm: stack64 input_6

# qhasm: stack64 input_7

# qhasm: int64 caller_r11

# qhasm: int64 caller_r12

# qhasm: int64 caller_r13

# qhasm: int64 caller_r14

# qhasm: int64 caller_r15

# qhasm: int64 caller_rbx

# qhasm: int64 caller_rbp

# qhasm: int64 b64

# qhasm: int64 synd

# qhasm: int64 addr

# qhasm: int64 c

# qhasm: int64 c_all

# qhasm: int64 row

# qhasm: int64 p

# qhasm: int64 e

# qhasm: int64 s

# qhasm: reg256 pp

# qhasm: reg256 ee

# qhasm: reg256 ss

# qhasm: int64 buf_ptr

# qhasm: stack256 buf

# qhasm: enter syndrome_asm
.p2align 5
ASM_HIDDEN _syndrome_asm
ASM_HIDDEN syndrome_asm
.global _syndrome_asm
.global syndrome_asm
_syndrome_asm:
syndrome_asm:
mov %rsp,%r11
and $31,%r11
add $32,%r11
sub %r11,%rsp

# qhasm: input_1 += 260780
# asm 1: add  $260780,<input_1=int64#2
# asm 2: add  $260780,<input_1=%rsi
add  $260780,%rsi

# qhasm: buf_ptr = &buf
# asm 1: leaq <buf=stack256#1,>buf_ptr=int64#4
# asm 2: leaq <buf=0(%rsp),>buf_ptr=%rcx
leaq 0(%rsp),%rcx

# qhasm: row = 768
# asm 1: mov  $768,>row=int64#5
# asm 2: mov  $768,>row=%r8
mov  $768,%r8

# qhasm: loop:
._loop:

# qhasm: row -= 1
# asm 1: sub  $1,<row=int64#5
# asm 2: sub  $1,<row=%r8
sub  $1,%r8

# qhasm: ss = mem256[ input_1 + 0 ]
# asm 1: vmovupd   0(<input_1=int64#2),>ss=reg256#1
# asm 2: vmovupd   0(<input_1=%rsi),>ss=%ymm0
vmovupd   0(%rsi),%ymm0

# qhasm: ee = mem256[ input_2 + 96 ]
# asm 1: vmovupd   96(<input_2=int64#3),>ee=reg256#2
# asm 2: vmovupd   96(<input_2=%rdx),>ee=%ymm1
vmovupd   96(%rdx),%ymm1

# qhasm: ss &= ee
# asm 1: vpand <ee=reg256#2,<ss=reg256#1,<ss=reg256#1
# asm 2: vpand <ee=%ymm1,<ss=%ymm0,<ss=%ymm0
vpand %ymm1,%ymm0,%ymm0

# qhasm: pp = mem256[ input_1 + 32 ]
# asm 1: vmovupd   32(<input_1=int64#2),>pp=reg256#2
# asm 2: vmovupd   32(<input_1=%rsi),>pp=%ymm1
vmovupd   32(%rsi),%ymm1

# qhasm: ee = mem256[ input_2 + 128 ]
# asm 1: vmovupd   128(<input_2=int64#3),>ee=reg256#3
# asm 2: vmovupd   128(<input_2=%rdx),>ee=%ymm2
vmovupd   128(%rdx),%ymm2

# qhasm: pp &= ee
# asm 1: vpand <ee=reg256#3,<pp=reg256#2,<pp=reg256#2
# asm 2: vpand <ee=%ymm2,<pp=%ymm1,<pp=%ymm1
vpand %ymm2,%ymm1,%ymm1

# qhasm: ss ^= pp
# asm 1: vpxor <pp=reg256#2,<ss=reg256#1,<ss=reg256#1
# asm 2: vpxor <pp=%ymm1,<ss=%ymm0,<ss=%ymm0
vpxor %ymm1,%ymm0,%ymm0

# qhasm: pp = mem256[ input_1 + 64 ]
# asm 1: vmovupd   64(<input_1=int64#2),>pp=reg256#2
# asm 2: vmovupd   64(<input_1=%rsi),>pp=%ymm1
vmovupd   64(%rsi),%ymm1

# qhasm: ee = mem256[ input_2 + 160 ]
# asm 1: vmovupd   160(<input_2=int64#3),>ee=reg256#3
# asm 2: vmovupd   160(<input_2=%rdx),>ee=%ymm2
vmovupd   160(%rdx),%ymm2

# qhasm: pp &= ee
# asm 1: vpand <ee=reg256#3,<pp=reg256#2,<pp=reg256#2
# asm 2: vpand <ee=%ymm2,<pp=%ymm1,<pp=%ymm1
vpand %ymm2,%ymm1,%ymm1

# qhasm: ss ^= pp
# asm 1: vpxor <pp=reg256#2,<ss=reg256#1,<ss=reg256#1
# asm 2: vpxor <pp=%ymm1,<ss=%ymm0,<ss=%ymm0
vpxor %ymm1,%ymm0,%ymm0

# qhasm: pp = mem256[ input_1 + 96 ]
# asm 1: vmovupd   96(<input_1=int64#2),>pp=reg256#2
# asm 2: vmovupd   96(<input_1=%rsi),>pp=%ymm1
vmovupd   96(%rsi),%ymm1

# qhasm: ee = mem256[ input_2 + 192 ]
# asm 1: vmovupd   192(<input_2=int64#3),>ee=reg256#3
# asm 2: vmovupd   192(<input_2=%rdx),>ee=%ymm2
vmovupd   192(%rdx),%ymm2

# qhasm: pp &= ee
# asm 1: vpand <ee=reg256#3,<pp=reg256#2,<pp=reg256#2
# asm 2: vpand <ee=%ymm2,<pp=%ymm1,<pp=%ymm1
vpand %ymm2,%ymm1,%ymm1

# qhasm: ss ^= pp
# asm 1: vpxor <pp=reg256#2,<ss=reg256#1,<ss=reg256#1
# asm 2: vpxor <pp=%ymm1,<ss=%ymm0,<ss=%ymm0
vpxor %ymm1,%ymm0,%ymm0

# qhasm: pp = mem256[ input_1 + 128 ]
# asm 1: vmovupd   128(<input_1=int64#2),>pp=reg256#2
# asm 2: vmovupd   128(<input_1=%rsi),>pp=%ymm1
vmovupd   128(%rsi),%ymm1

# qhasm: ee = mem256[ input_2 + 224 ]
# asm 1: vmovupd   224(<input_2=int64#3),>ee=reg256#3
# asm 2: vmovupd   224(<input_2=%rdx),>ee=%ymm2
vmovupd   224(%rdx),%ymm2

# qhasm: pp &= ee
# asm 1: vpand <ee=reg256#3,<pp=reg256#2,<pp=reg256#2
# asm 2: vpand <ee=%ymm2,<pp=%ymm1,<pp=%ymm1
vpand %ymm2,%ymm1,%ymm1

# qhasm: ss ^= pp
# asm 1: vpxor <pp=reg256#2,<ss=reg256#1,<ss=reg256#1
# asm 2: vpxor <pp=%ymm1,<ss=%ymm0,<ss=%ymm0
vpxor %ymm1,%ymm0,%ymm0

# qhasm: pp = mem256[ input_1 + 160 ]
# asm 1: vmovupd   160(<input_1=int64#2),>pp=reg256#2
# asm 2: vmovupd   160(<input_1=%rsi),>pp=%ymm1
vmovupd   160(%rsi),%ymm1

# qhasm: ee = mem256[ input_2 + 256 ]
# asm 1: vmovupd   256(<input_2=int64#3),>ee=reg256#3
# asm 2: vmovupd   256(<input_2=%rdx),>ee=%ymm2
vmovupd   256(%rdx),%ymm2

# qhasm: pp &= ee
# asm 1: vpand <ee=reg256#3,<pp=reg256#2,<pp=reg256#2
# asm 2: vpand <ee=%ymm2,<pp=%ymm1,<pp=%ymm1
vpand %ymm2,%ymm1,%ymm1

# qhasm: ss ^= pp
# asm 1: vpxor <pp=reg256#2,<ss=reg256#1,<ss=reg256#1
# asm 2: vpxor <pp=%ymm1,<ss=%ymm0,<ss=%ymm0
vpxor %ymm1,%ymm0,%ymm0

# qhasm: pp = mem256[ input_1 + 192 ]
# asm 1: vmovupd   192(<input_1=int64#2),>pp=reg256#2
# asm 2: vmovupd   192(<input_1=%rsi),>pp=%ymm1
vmovupd   192(%rsi),%ymm1

# qhasm: ee = mem256[ input_2 + 288 ]
# asm 1: vmovupd   288(<input_2=int64#3),>ee=reg256#3
# asm 2: vmovupd   288(<input_2=%rdx),>ee=%ymm2
vmovupd   288(%rdx),%ymm2

# qhasm: pp &= ee
# asm 1: vpand <ee=reg256#3,<pp=reg256#2,<pp=reg256#2
# asm 2: vpand <ee=%ymm2,<pp=%ymm1,<pp=%ymm1
vpand %ymm2,%ymm1,%ymm1

# qhasm: ss ^= pp
# asm 1: vpxor <pp=reg256#2,<ss=reg256#1,<ss=reg256#1
# asm 2: vpxor <pp=%ymm1,<ss=%ymm0,<ss=%ymm0
vpxor %ymm1,%ymm0,%ymm0

# qhasm: pp = mem256[ input_1 + 224 ]
# asm 1: vmovupd   224(<input_1=int64#2),>pp=reg256#2
# asm 2: vmovupd   224(<input_1=%rsi),>pp=%ymm1
vmovupd   224(%rsi),%ymm1

# qhasm: ee = mem256[ input_2 + 320 ]
# asm 1: vmovupd   320(<input_2=int64#3),>ee=reg256#3
# asm 2: vmovupd   320(<input_2=%rdx),>ee=%ymm2
vmovupd   320(%rdx),%ymm2

# qhasm: pp &= ee
# asm 1: vpand <ee=reg256#3,<pp=reg256#2,<pp=reg256#2
# asm 2: vpand <ee=%ymm2,<pp=%ymm1,<pp=%ymm1
vpand %ymm2,%ymm1,%ymm1

# qhasm: ss ^= pp
# asm 1: vpxor <pp=reg256#2,<ss=reg256#1,<ss=reg256#1
# asm 2: vpxor <pp=%ymm1,<ss=%ymm0,<ss=%ymm0
vpxor %ymm1,%ymm0,%ymm0

# qhasm: pp = mem256[ input_1 + 256 ]
# asm 1: vmovupd   256(<input_1=int64#2),>pp=reg256#2
# asm 2: vmovupd   256(<input_1=%rsi),>pp=%ymm1
vmovupd   256(%rsi),%ymm1

# qhasm: ee = mem256[ input_2 + 352 ]
# asm 1: vmovupd   352(<input_2=int64#3),>ee=reg256#3
# asm 2: vmovupd   352(<input_2=%rdx),>ee=%ymm2
vmovupd   352(%rdx),%ymm2

# qhasm: pp &= ee
# asm 1: vpand <ee=reg256#3,<pp=reg256#2,<pp=reg256#2
# asm 2: vpand <ee=%ymm2,<pp=%ymm1,<pp=%ymm1
vpand %ymm2,%ymm1,%ymm1

# qhasm: ss ^= pp
# asm 1: vpxor <pp=reg256#2,<ss=reg256#1,<ss=reg256#1
# asm 2: vpxor <pp=%ymm1,<ss=%ymm0,<ss=%ymm0
vpxor %ymm1,%ymm0,%ymm0

# qhasm: pp = mem256[ input_1 + 288 ]
# asm 1: vmovupd   288(<input_1=int64#2),>pp=reg256#2
# asm 2: vmovupd   288(<input_1=%rsi),>pp=%ymm1
vmovupd   288(%rsi),%ymm1

# qhasm: ee = mem256[ input_2 + 384 ]
# asm 1: vmovupd   384(<input_2=int64#3),>ee=reg256#3
# asm 2: vmovupd   384(<input_2=%rdx),>ee=%ymm2
vmovupd   384(%rdx),%ymm2

# qhasm: pp &= ee
# asm 1: vpand <ee=reg256#3,<pp=reg256#2,<pp=reg256#2
# asm 2: vpand <ee=%ymm2,<pp=%ymm1,<pp=%ymm1
vpand %ymm2,%ymm1,%ymm1

# qhasm: ss ^= pp
# asm 1: vpxor <pp=reg256#2,<ss=reg256#1,<ss=reg256#1
# asm 2: vpxor <pp=%ymm1,<ss=%ymm0,<ss=%ymm0
vpxor %ymm1,%ymm0,%ymm0

# qhasm: buf = ss
# asm 1: vmovapd <ss=reg256#1,>buf=stack256#1
# asm 2: vmovapd <ss=%ymm0,>buf=0(%rsp)
vmovapd %ymm0,0(%rsp)

# qhasm: s = *(uint64 *)(input_1 + 320)
# asm 1: movq   320(<input_1=int64#2),>s=int64#6
# asm 2: movq   320(<input_1=%rsi),>s=%r9
movq   320(%rsi),%r9

# qhasm: e = *(uint64 *)(input_2 + 416)
# asm 1: movq   416(<input_2=int64#3),>e=int64#7
# asm 2: movq   416(<input_2=%rdx),>e=%rax
movq   416(%rdx),%rax

# qhasm: s &= e
# asm 1: and  <e=int64#7,<s=int64#6
# asm 2: and  <e=%rax,<s=%r9
and  %rax,%r9

# qhasm: p = *(uint64 *)(input_1 + 328)
# asm 1: movq   328(<input_1=int64#2),>p=int64#7
# asm 2: movq   328(<input_1=%rsi),>p=%rax
movq   328(%rsi),%rax

# qhasm: e = *(uint64 *)(input_2 + 424)
# asm 1: movq   424(<input_2=int64#3),>e=int64#8
# asm 2: movq   424(<input_2=%rdx),>e=%r10
movq   424(%rdx),%r10

# qhasm: p &= e
# asm 1: and  <e=int64#8,<p=int64#7
# asm 2: and  <e=%r10,<p=%rax
and  %r10,%rax

# qhasm: s ^= p
# asm 1: xor  <p=int64#7,<s=int64#6
# asm 2: xor  <p=%rax,<s=%r9
xor  %rax,%r9

# qhasm: p = *(uint32 *)(input_1 + 336)
# asm 1: movl   336(<input_1=int64#2),>p=int64#7d
# asm 2: movl   336(<input_1=%rsi),>p=%eax
movl   336(%rsi),%eax

# qhasm: e = *(uint32 *)(input_2 + 432)
# asm 1: movl   432(<input_2=int64#3),>e=int64#8d
# asm 2: movl   432(<input_2=%rdx),>e=%r10d
movl   432(%rdx),%r10d

# qhasm: p &= e
# asm 1: and  <e=int64#8,<p=int64#7
# asm 2: and  <e=%r10,<p=%rax
and  %r10,%rax

# qhasm: s ^= p
# asm 1: xor  <p=int64#7,<s=int64#6
# asm 2: xor  <p=%rax,<s=%r9
xor  %rax,%r9

# qhasm: c_all = count(s)
# asm 1: popcnt <s=int64#6, >c_all=int64#6
# asm 2: popcnt <s=%r9, >c_all=%r9
popcnt %r9, %r9

# qhasm: b64 = mem64[ buf_ptr + 0 ]
# asm 1: movq   0(<buf_ptr=int64#4),>b64=int64#7
# asm 2: movq   0(<buf_ptr=%rcx),>b64=%rax
movq   0(%rcx),%rax

# qhasm: c = count(b64)
# asm 1: popcnt <b64=int64#7, >c=int64#7
# asm 2: popcnt <b64=%rax, >c=%rax
popcnt %rax, %rax

# qhasm: c_all ^= c
# asm 1: xor  <c=int64#7,<c_all=int64#6
# asm 2: xor  <c=%rax,<c_all=%r9
xor  %rax,%r9

# qhasm: b64 = mem64[ buf_ptr + 8 ]
# asm 1: movq   8(<buf_ptr=int64#4),>b64=int64#7
# asm 2: movq   8(<buf_ptr=%rcx),>b64=%rax
movq   8(%rcx),%rax

# qhasm: c = count(b64)
# asm 1: popcnt <b64=int64#7, >c=int64#7
# asm 2: popcnt <b64=%rax, >c=%rax
popcnt %rax, %rax

# qhasm: c_all ^= c
# asm 1: xor  <c=int64#7,<c_all=int64#6
# asm 2: xor  <c=%rax,<c_all=%r9
xor  %rax,%r9

# qhasm: b64 = mem64[ buf_ptr + 16 ]
# asm 1: movq   16(<buf_ptr=int64#4),>b64=int64#7
# asm 2: movq   16(<buf_ptr=%rcx),>b64=%rax
movq   16(%rcx),%rax

# qhasm: c = count(b64)
# asm 1: popcnt <b64=int64#7, >c=int64#7
# asm 2: popcnt <b64=%rax, >c=%rax
popcnt %rax, %rax

# qhasm: c_all ^= c
# asm 1: xor  <c=int64#7,<c_all=int64#6
# asm 2: xor  <c=%rax,<c_all=%r9
xor  %rax,%r9

# qhasm: b64 = mem64[ buf_ptr + 24 ]
# asm 1: movq   24(<buf_ptr=int64#4),>b64=int64#7
# asm 2: movq   24(<buf_ptr=%rcx),>b64=%rax
movq   24(%rcx),%rax

# qhasm: c = count(b64)
# asm 1: popcnt <b64=int64#7, >c=int64#7
# asm 2: popcnt <b64=%rax, >c=%rax
popcnt %rax, %rax

# qhasm: c_all ^= c
# asm 1: xor  <c=int64#7,<c_all=int64#6
# asm 2: xor  <c=%rax,<c_all=%r9
xor  %rax,%r9

# qhasm: addr = row
# asm 1: mov  <row=int64#5,>addr=int64#7
# asm 2: mov  <row=%r8,>addr=%rax
mov  %r8,%rax

# qhasm: (uint64) addr >>= 3
# asm 1: shr  $3,<addr=int64#7
# asm 2: shr  $3,<addr=%rax
shr  $3,%rax

# qhasm: addr += input_0
# asm 1: add  <input_0=int64#1,<addr=int64#7
# asm 2: add  <input_0=%rdi,<addr=%rax
add  %rdi,%rax

# qhasm: synd = *(uint8  *) (addr + 0)
# asm 1: movzbq 0(<addr=int64#7),>synd=int64#8
# asm 2: movzbq 0(<addr=%rax),>synd=%r10
movzbq 0(%rax),%r10

# qhasm: synd <<= 1
# asm 1: shl  $1,<synd=int64#8
# asm 2: shl  $1,<synd=%r10
shl  $1,%r10

# qhasm: (uint32) c_all &= 1
# asm 1: and  $1,<c_all=int64#6d
# asm 2: and  $1,<c_all=%r9d
and  $1,%r9d

# qhasm: synd |= c_all
# asm 1: or   <c_all=int64#6,<synd=int64#8
# asm 2: or   <c_all=%r9,<synd=%r10
or   %r9,%r10

# qhasm: *(uint8  *) (addr + 0) = synd
# asm 1: movb   <synd=int64#8b,0(<addr=int64#7)
# asm 2: movb   <synd=%r10b,0(<addr=%rax)
movb   %r10b,0(%rax)

# qhasm: input_1 -= 340
# asm 1: sub  $340,<input_1=int64#2
# asm 2: sub  $340,<input_1=%rsi
sub  $340,%rsi

# qhasm: =? row-0
# asm 1: cmp  $0,<row=int64#5
# asm 2: cmp  $0,<row=%r8
cmp  $0,%r8
# comment:fp stack unchanged by jump

# qhasm: goto loop if !=
jne ._loop

# qhasm: ss = mem256[ input_0 + 0 ]
# asm 1: vmovupd   0(<input_0=int64#1),>ss=reg256#1
# asm 2: vmovupd   0(<input_0=%rdi),>ss=%ymm0
vmovupd   0(%rdi),%ymm0

# qhasm: ee = mem256[ input_2 + 0 ]
# asm 1: vmovupd   0(<input_2=int64#3),>ee=reg256#2
# asm 2: vmovupd   0(<input_2=%rdx),>ee=%ymm1
vmovupd   0(%rdx),%ymm1

# qhasm: ss ^= ee
# asm 1: vpxor <ee=reg256#2,<ss=reg256#1,<ss=reg256#1
# asm 2: vpxor <ee=%ymm1,<ss=%ymm0,<ss=%ymm0
vpxor %ymm1,%ymm0,%ymm0

# qhasm: mem256[ input_0 + 0 ] = ss
# asm 1: vmovupd   <ss=reg256#1,0(<input_0=int64#1)
# asm 2: vmovupd   <ss=%ymm0,0(<input_0=%rdi)
vmovupd   %ymm0,0(%rdi)

# qhasm: ss = mem256[ input_0 + 32 ]
# asm 1: vmovupd   32(<input_0=int64#1),>ss=reg256#1
# asm 2: vmovupd   32(<input_0=%rdi),>ss=%ymm0
vmovupd   32(%rdi),%ymm0

# qhasm: ee = mem256[ input_2 + 32 ]
# asm 1: vmovupd   32(<input_2=int64#3),>ee=reg256#2
# asm 2: vmovupd   32(<input_2=%rdx),>ee=%ymm1
vmovupd   32(%rdx),%ymm1

# qhasm: ss ^= ee
# asm 1: vpxor <ee=reg256#2,<ss=reg256#1,<ss=reg256#1
# asm 2: vpxor <ee=%ymm1,<ss=%ymm0,<ss=%ymm0
vpxor %ymm1,%ymm0,%ymm0

# qhasm: mem256[ input_0 + 32 ] = ss
# asm 1: vmovupd   <ss=reg256#1,32(<input_0=int64#1)
# asm 2: vmovupd   <ss=%ymm0,32(<input_0=%rdi)
vmovupd   %ymm0,32(%rdi)

# qhasm: ss = mem256[ input_0 + 64 ]
# asm 1: vmovupd   64(<input_0=int64#1),>ss=reg256#1
# asm 2: vmovupd   64(<input_0=%rdi),>ss=%ymm0
vmovupd   64(%rdi),%ymm0

# qhasm: ee = mem256[ input_2 + 64 ]
# asm 1: vmovupd   64(<input_2=int64#3),>ee=reg256#2
# asm 2: vmovupd   64(<input_2=%rdx),>ee=%ymm1
vmovupd   64(%rdx),%ymm1

# qhasm: ss ^= ee
# asm 1: vpxor <ee=reg256#2,<ss=reg256#1,<ss=reg256#1
# asm 2: vpxor <ee=%ymm1,<ss=%ymm0,<ss=%ymm0
vpxor %ymm1,%ymm0,%ymm0

# qhasm: mem256[ input_0 + 64 ] = ss
# asm 1: vmovupd   <ss=reg256#1,64(<input_0=int64#1)
# asm 2: vmovupd   <ss=%ymm0,64(<input_0=%rdi)
vmovupd   %ymm0,64(%rdi)

# qhasm: return
add %r11,%rsp
ret