1
1
mirror of https://github.com/henrydcase/pqc.git synced 2024-11-30 03:11:43 +00:00
pqcrypto/crypto_kem/mceliece460896/sse/syndrome_asm.S
Thom Wiggers ac2c20045c Classic McEliece (#259)
* Add McEliece reference implementations

* Add Vec implementations of McEliece

* Add sse implementations

* Add AVX2 implementations

* Get rid of stuff not supported by Mac ABI

* restrict to two cores

* Ditch .data files

* Remove .hidden from all .S files

* speed up duplicate consistency tests by batching

* make cpuinfo more robust

* Hope to stabilize macos cpuinfo without ccache

* Revert "Hope to stabilize macos cpuinfo without ccache"

This reverts commit 6129c3cabe1abbc8b956bc87e902a698e32bf322.

* Just hardcode what's available at travis

* Fixed-size types in api.h

* namespace all header files in mceliece

* Ditch operations.h

* Get rid of static inline functions

* fixup! Ditch operations.h
2021-03-24 21:02:45 +00:00

961 lines
24 KiB
ArmAsm

# 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: reg128 pp
# qhasm: reg128 ee
# qhasm: reg128 ss
# qhasm: int64 buf_ptr
# qhasm: stack128 buf
# qhasm: enter syndrome_asm
.p2align 5
.global _PQCLEAN_MCELIECE460896_SSE_syndrome_asm
.global PQCLEAN_MCELIECE460896_SSE_syndrome_asm
_PQCLEAN_MCELIECE460896_SSE_syndrome_asm:
PQCLEAN_MCELIECE460896_SSE_syndrome_asm:
mov %rsp,%r11
and $31,%r11
add $32,%r11
sub %r11,%rsp
# qhasm: input_1 += 523740
# asm 1: add $523740,<input_1=int64#2
# asm 2: add $523740,<input_1=%rsi
add $523740,%rsi
# qhasm: buf_ptr = &buf
# asm 1: leaq <buf=stack128#1,>buf_ptr=int64#4
# asm 2: leaq <buf=0(%rsp),>buf_ptr=%rcx
leaq 0(%rsp),%rcx
# qhasm: row = 1248
# asm 1: mov $1248,>row=int64#5
# asm 2: mov $1248,>row=%r8
mov $1248,%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 = mem128[ input_1 + 0 ]
# asm 1: movdqu 0(<input_1=int64#2),>ss=reg128#1
# asm 2: movdqu 0(<input_1=%rsi),>ss=%xmm0
movdqu 0(%rsi),%xmm0
# qhasm: ee = mem128[ input_2 + 156 ]
# asm 1: movdqu 156(<input_2=int64#3),>ee=reg128#2
# asm 2: movdqu 156(<input_2=%rdx),>ee=%xmm1
movdqu 156(%rdx),%xmm1
# qhasm: ss &= ee
# asm 1: pand <ee=reg128#2,<ss=reg128#1
# asm 2: pand <ee=%xmm1,<ss=%xmm0
pand %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 16 ]
# asm 1: movdqu 16(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 16(<input_1=%rsi),>pp=%xmm1
movdqu 16(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 172 ]
# asm 1: movdqu 172(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 172(<input_2=%rdx),>ee=%xmm2
movdqu 172(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 32 ]
# asm 1: movdqu 32(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 32(<input_1=%rsi),>pp=%xmm1
movdqu 32(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 188 ]
# asm 1: movdqu 188(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 188(<input_2=%rdx),>ee=%xmm2
movdqu 188(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 48 ]
# asm 1: movdqu 48(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 48(<input_1=%rsi),>pp=%xmm1
movdqu 48(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 204 ]
# asm 1: movdqu 204(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 204(<input_2=%rdx),>ee=%xmm2
movdqu 204(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 64 ]
# asm 1: movdqu 64(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 64(<input_1=%rsi),>pp=%xmm1
movdqu 64(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 220 ]
# asm 1: movdqu 220(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 220(<input_2=%rdx),>ee=%xmm2
movdqu 220(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 80 ]
# asm 1: movdqu 80(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 80(<input_1=%rsi),>pp=%xmm1
movdqu 80(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 236 ]
# asm 1: movdqu 236(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 236(<input_2=%rdx),>ee=%xmm2
movdqu 236(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 96 ]
# asm 1: movdqu 96(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 96(<input_1=%rsi),>pp=%xmm1
movdqu 96(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 252 ]
# asm 1: movdqu 252(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 252(<input_2=%rdx),>ee=%xmm2
movdqu 252(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 112 ]
# asm 1: movdqu 112(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 112(<input_1=%rsi),>pp=%xmm1
movdqu 112(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 268 ]
# asm 1: movdqu 268(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 268(<input_2=%rdx),>ee=%xmm2
movdqu 268(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 128 ]
# asm 1: movdqu 128(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 128(<input_1=%rsi),>pp=%xmm1
movdqu 128(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 284 ]
# asm 1: movdqu 284(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 284(<input_2=%rdx),>ee=%xmm2
movdqu 284(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 144 ]
# asm 1: movdqu 144(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 144(<input_1=%rsi),>pp=%xmm1
movdqu 144(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 300 ]
# asm 1: movdqu 300(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 300(<input_2=%rdx),>ee=%xmm2
movdqu 300(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 160 ]
# asm 1: movdqu 160(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 160(<input_1=%rsi),>pp=%xmm1
movdqu 160(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 316 ]
# asm 1: movdqu 316(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 316(<input_2=%rdx),>ee=%xmm2
movdqu 316(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 176 ]
# asm 1: movdqu 176(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 176(<input_1=%rsi),>pp=%xmm1
movdqu 176(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 332 ]
# asm 1: movdqu 332(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 332(<input_2=%rdx),>ee=%xmm2
movdqu 332(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 192 ]
# asm 1: movdqu 192(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 192(<input_1=%rsi),>pp=%xmm1
movdqu 192(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 348 ]
# asm 1: movdqu 348(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 348(<input_2=%rdx),>ee=%xmm2
movdqu 348(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 208 ]
# asm 1: movdqu 208(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 208(<input_1=%rsi),>pp=%xmm1
movdqu 208(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 364 ]
# asm 1: movdqu 364(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 364(<input_2=%rdx),>ee=%xmm2
movdqu 364(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 224 ]
# asm 1: movdqu 224(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 224(<input_1=%rsi),>pp=%xmm1
movdqu 224(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 380 ]
# asm 1: movdqu 380(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 380(<input_2=%rdx),>ee=%xmm2
movdqu 380(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 240 ]
# asm 1: movdqu 240(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 240(<input_1=%rsi),>pp=%xmm1
movdqu 240(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 396 ]
# asm 1: movdqu 396(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 396(<input_2=%rdx),>ee=%xmm2
movdqu 396(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 256 ]
# asm 1: movdqu 256(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 256(<input_1=%rsi),>pp=%xmm1
movdqu 256(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 412 ]
# asm 1: movdqu 412(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 412(<input_2=%rdx),>ee=%xmm2
movdqu 412(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 272 ]
# asm 1: movdqu 272(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 272(<input_1=%rsi),>pp=%xmm1
movdqu 272(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 428 ]
# asm 1: movdqu 428(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 428(<input_2=%rdx),>ee=%xmm2
movdqu 428(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 288 ]
# asm 1: movdqu 288(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 288(<input_1=%rsi),>pp=%xmm1
movdqu 288(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 444 ]
# asm 1: movdqu 444(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 444(<input_2=%rdx),>ee=%xmm2
movdqu 444(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 304 ]
# asm 1: movdqu 304(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 304(<input_1=%rsi),>pp=%xmm1
movdqu 304(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 460 ]
# asm 1: movdqu 460(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 460(<input_2=%rdx),>ee=%xmm2
movdqu 460(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 320 ]
# asm 1: movdqu 320(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 320(<input_1=%rsi),>pp=%xmm1
movdqu 320(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 476 ]
# asm 1: movdqu 476(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 476(<input_2=%rdx),>ee=%xmm2
movdqu 476(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 336 ]
# asm 1: movdqu 336(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 336(<input_1=%rsi),>pp=%xmm1
movdqu 336(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 492 ]
# asm 1: movdqu 492(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 492(<input_2=%rdx),>ee=%xmm2
movdqu 492(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 352 ]
# asm 1: movdqu 352(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 352(<input_1=%rsi),>pp=%xmm1
movdqu 352(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 508 ]
# asm 1: movdqu 508(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 508(<input_2=%rdx),>ee=%xmm2
movdqu 508(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 368 ]
# asm 1: movdqu 368(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 368(<input_1=%rsi),>pp=%xmm1
movdqu 368(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 524 ]
# asm 1: movdqu 524(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 524(<input_2=%rdx),>ee=%xmm2
movdqu 524(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 384 ]
# asm 1: movdqu 384(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 384(<input_1=%rsi),>pp=%xmm1
movdqu 384(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 540 ]
# asm 1: movdqu 540(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 540(<input_2=%rdx),>ee=%xmm2
movdqu 540(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: pp = mem128[ input_1 + 400 ]
# asm 1: movdqu 400(<input_1=int64#2),>pp=reg128#2
# asm 2: movdqu 400(<input_1=%rsi),>pp=%xmm1
movdqu 400(%rsi),%xmm1
# qhasm: ee = mem128[ input_2 + 556 ]
# asm 1: movdqu 556(<input_2=int64#3),>ee=reg128#3
# asm 2: movdqu 556(<input_2=%rdx),>ee=%xmm2
movdqu 556(%rdx),%xmm2
# qhasm: pp &= ee
# asm 1: pand <ee=reg128#3,<pp=reg128#2
# asm 2: pand <ee=%xmm2,<pp=%xmm1
pand %xmm2,%xmm1
# qhasm: ss ^= pp
# asm 1: pxor <pp=reg128#2,<ss=reg128#1
# asm 2: pxor <pp=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: buf = ss
# asm 1: movdqa <ss=reg128#1,>buf=stack128#1
# asm 2: movdqa <ss=%xmm0,>buf=0(%rsp)
movdqa %xmm0,0(%rsp)
# qhasm: s = *(uint32 *)(input_1 + 416)
# asm 1: movl 416(<input_1=int64#2),>s=int64#6d
# asm 2: movl 416(<input_1=%rsi),>s=%r9d
movl 416(%rsi),%r9d
# qhasm: e = *(uint32 *)(input_2 + 572)
# asm 1: movl 572(<input_2=int64#3),>e=int64#7d
# asm 2: movl 572(<input_2=%rdx),>e=%eax
movl 572(%rdx),%eax
# qhasm: s &= e
# asm 1: and <e=int64#7,<s=int64#6
# asm 2: and <e=%rax,<s=%r9
and %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: 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 -= 420
# asm 1: sub $420,<input_1=int64#2
# asm 2: sub $420,<input_1=%rsi
sub $420,%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 = mem128[ input_0 + 0 ]
# asm 1: movdqu 0(<input_0=int64#1),>ss=reg128#1
# asm 2: movdqu 0(<input_0=%rdi),>ss=%xmm0
movdqu 0(%rdi),%xmm0
# qhasm: ee = mem128[ input_2 + 0 ]
# asm 1: movdqu 0(<input_2=int64#3),>ee=reg128#2
# asm 2: movdqu 0(<input_2=%rdx),>ee=%xmm1
movdqu 0(%rdx),%xmm1
# qhasm: ss ^= ee
# asm 1: pxor <ee=reg128#2,<ss=reg128#1
# asm 2: pxor <ee=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: mem128[ input_0 + 0 ] = ss
# asm 1: movdqu <ss=reg128#1,0(<input_0=int64#1)
# asm 2: movdqu <ss=%xmm0,0(<input_0=%rdi)
movdqu %xmm0,0(%rdi)
# qhasm: ss = mem128[ input_0 + 16 ]
# asm 1: movdqu 16(<input_0=int64#1),>ss=reg128#1
# asm 2: movdqu 16(<input_0=%rdi),>ss=%xmm0
movdqu 16(%rdi),%xmm0
# qhasm: ee = mem128[ input_2 + 16 ]
# asm 1: movdqu 16(<input_2=int64#3),>ee=reg128#2
# asm 2: movdqu 16(<input_2=%rdx),>ee=%xmm1
movdqu 16(%rdx),%xmm1
# qhasm: ss ^= ee
# asm 1: pxor <ee=reg128#2,<ss=reg128#1
# asm 2: pxor <ee=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: mem128[ input_0 + 16 ] = ss
# asm 1: movdqu <ss=reg128#1,16(<input_0=int64#1)
# asm 2: movdqu <ss=%xmm0,16(<input_0=%rdi)
movdqu %xmm0,16(%rdi)
# qhasm: ss = mem128[ input_0 + 32 ]
# asm 1: movdqu 32(<input_0=int64#1),>ss=reg128#1
# asm 2: movdqu 32(<input_0=%rdi),>ss=%xmm0
movdqu 32(%rdi),%xmm0
# qhasm: ee = mem128[ input_2 + 32 ]
# asm 1: movdqu 32(<input_2=int64#3),>ee=reg128#2
# asm 2: movdqu 32(<input_2=%rdx),>ee=%xmm1
movdqu 32(%rdx),%xmm1
# qhasm: ss ^= ee
# asm 1: pxor <ee=reg128#2,<ss=reg128#1
# asm 2: pxor <ee=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: mem128[ input_0 + 32 ] = ss
# asm 1: movdqu <ss=reg128#1,32(<input_0=int64#1)
# asm 2: movdqu <ss=%xmm0,32(<input_0=%rdi)
movdqu %xmm0,32(%rdi)
# qhasm: ss = mem128[ input_0 + 48 ]
# asm 1: movdqu 48(<input_0=int64#1),>ss=reg128#1
# asm 2: movdqu 48(<input_0=%rdi),>ss=%xmm0
movdqu 48(%rdi),%xmm0
# qhasm: ee = mem128[ input_2 + 48 ]
# asm 1: movdqu 48(<input_2=int64#3),>ee=reg128#2
# asm 2: movdqu 48(<input_2=%rdx),>ee=%xmm1
movdqu 48(%rdx),%xmm1
# qhasm: ss ^= ee
# asm 1: pxor <ee=reg128#2,<ss=reg128#1
# asm 2: pxor <ee=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: mem128[ input_0 + 48 ] = ss
# asm 1: movdqu <ss=reg128#1,48(<input_0=int64#1)
# asm 2: movdqu <ss=%xmm0,48(<input_0=%rdi)
movdqu %xmm0,48(%rdi)
# qhasm: ss = mem128[ input_0 + 64 ]
# asm 1: movdqu 64(<input_0=int64#1),>ss=reg128#1
# asm 2: movdqu 64(<input_0=%rdi),>ss=%xmm0
movdqu 64(%rdi),%xmm0
# qhasm: ee = mem128[ input_2 + 64 ]
# asm 1: movdqu 64(<input_2=int64#3),>ee=reg128#2
# asm 2: movdqu 64(<input_2=%rdx),>ee=%xmm1
movdqu 64(%rdx),%xmm1
# qhasm: ss ^= ee
# asm 1: pxor <ee=reg128#2,<ss=reg128#1
# asm 2: pxor <ee=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: mem128[ input_0 + 64 ] = ss
# asm 1: movdqu <ss=reg128#1,64(<input_0=int64#1)
# asm 2: movdqu <ss=%xmm0,64(<input_0=%rdi)
movdqu %xmm0,64(%rdi)
# qhasm: ss = mem128[ input_0 + 80 ]
# asm 1: movdqu 80(<input_0=int64#1),>ss=reg128#1
# asm 2: movdqu 80(<input_0=%rdi),>ss=%xmm0
movdqu 80(%rdi),%xmm0
# qhasm: ee = mem128[ input_2 + 80 ]
# asm 1: movdqu 80(<input_2=int64#3),>ee=reg128#2
# asm 2: movdqu 80(<input_2=%rdx),>ee=%xmm1
movdqu 80(%rdx),%xmm1
# qhasm: ss ^= ee
# asm 1: pxor <ee=reg128#2,<ss=reg128#1
# asm 2: pxor <ee=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: mem128[ input_0 + 80 ] = ss
# asm 1: movdqu <ss=reg128#1,80(<input_0=int64#1)
# asm 2: movdqu <ss=%xmm0,80(<input_0=%rdi)
movdqu %xmm0,80(%rdi)
# qhasm: ss = mem128[ input_0 + 96 ]
# asm 1: movdqu 96(<input_0=int64#1),>ss=reg128#1
# asm 2: movdqu 96(<input_0=%rdi),>ss=%xmm0
movdqu 96(%rdi),%xmm0
# qhasm: ee = mem128[ input_2 + 96 ]
# asm 1: movdqu 96(<input_2=int64#3),>ee=reg128#2
# asm 2: movdqu 96(<input_2=%rdx),>ee=%xmm1
movdqu 96(%rdx),%xmm1
# qhasm: ss ^= ee
# asm 1: pxor <ee=reg128#2,<ss=reg128#1
# asm 2: pxor <ee=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: mem128[ input_0 + 96 ] = ss
# asm 1: movdqu <ss=reg128#1,96(<input_0=int64#1)
# asm 2: movdqu <ss=%xmm0,96(<input_0=%rdi)
movdqu %xmm0,96(%rdi)
# qhasm: ss = mem128[ input_0 + 112 ]
# asm 1: movdqu 112(<input_0=int64#1),>ss=reg128#1
# asm 2: movdqu 112(<input_0=%rdi),>ss=%xmm0
movdqu 112(%rdi),%xmm0
# qhasm: ee = mem128[ input_2 + 112 ]
# asm 1: movdqu 112(<input_2=int64#3),>ee=reg128#2
# asm 2: movdqu 112(<input_2=%rdx),>ee=%xmm1
movdqu 112(%rdx),%xmm1
# qhasm: ss ^= ee
# asm 1: pxor <ee=reg128#2,<ss=reg128#1
# asm 2: pxor <ee=%xmm1,<ss=%xmm0
pxor %xmm1,%xmm0
# qhasm: mem128[ input_0 + 112 ] = ss
# asm 1: movdqu <ss=reg128#1,112(<input_0=int64#1)
# asm 2: movdqu <ss=%xmm0,112(<input_0=%rdi)
movdqu %xmm0,112(%rdi)
# qhasm: s = mem64[ input_0 + 128 ]
# asm 1: movq 128(<input_0=int64#1),>s=int64#2
# asm 2: movq 128(<input_0=%rdi),>s=%rsi
movq 128(%rdi),%rsi
# qhasm: e = mem64[ input_2 + 128 ]
# asm 1: movq 128(<input_2=int64#3),>e=int64#4
# asm 2: movq 128(<input_2=%rdx),>e=%rcx
movq 128(%rdx),%rcx
# qhasm: s ^= e
# asm 1: xor <e=int64#4,<s=int64#2
# asm 2: xor <e=%rcx,<s=%rsi
xor %rcx,%rsi
# qhasm: mem64[ input_0 + 128 ] = s
# asm 1: movq <s=int64#2,128(<input_0=int64#1)
# asm 2: movq <s=%rsi,128(<input_0=%rdi)
movq %rsi,128(%rdi)
# qhasm: s = mem64[ input_0 + 136 ]
# asm 1: movq 136(<input_0=int64#1),>s=int64#2
# asm 2: movq 136(<input_0=%rdi),>s=%rsi
movq 136(%rdi),%rsi
# qhasm: e = mem64[ input_2 + 136 ]
# asm 1: movq 136(<input_2=int64#3),>e=int64#4
# asm 2: movq 136(<input_2=%rdx),>e=%rcx
movq 136(%rdx),%rcx
# qhasm: s ^= e
# asm 1: xor <e=int64#4,<s=int64#2
# asm 2: xor <e=%rcx,<s=%rsi
xor %rcx,%rsi
# qhasm: mem64[ input_0 + 136 ] = s
# asm 1: movq <s=int64#2,136(<input_0=int64#1)
# asm 2: movq <s=%rsi,136(<input_0=%rdi)
movq %rsi,136(%rdi)
# qhasm: s = mem64[ input_0 + 144 ]
# asm 1: movq 144(<input_0=int64#1),>s=int64#2
# asm 2: movq 144(<input_0=%rdi),>s=%rsi
movq 144(%rdi),%rsi
# qhasm: e = mem64[ input_2 + 144 ]
# asm 1: movq 144(<input_2=int64#3),>e=int64#4
# asm 2: movq 144(<input_2=%rdx),>e=%rcx
movq 144(%rdx),%rcx
# qhasm: s ^= e
# asm 1: xor <e=int64#4,<s=int64#2
# asm 2: xor <e=%rcx,<s=%rsi
xor %rcx,%rsi
# qhasm: mem64[ input_0 + 144 ] = s
# asm 1: movq <s=int64#2,144(<input_0=int64#1)
# asm 2: movq <s=%rsi,144(<input_0=%rdi)
movq %rsi,144(%rdi)
# qhasm: s = *(uint32 *)( input_0 + 152 )
# asm 1: movl 152(<input_0=int64#1),>s=int64#2d
# asm 2: movl 152(<input_0=%rdi),>s=%esi
movl 152(%rdi),%esi
# qhasm: e = *(uint32 *)( input_2 + 152 )
# asm 1: movl 152(<input_2=int64#3),>e=int64#3d
# asm 2: movl 152(<input_2=%rdx),>e=%edx
movl 152(%rdx),%edx
# qhasm: s ^= e
# asm 1: xor <e=int64#3,<s=int64#2
# asm 2: xor <e=%rdx,<s=%rsi
xor %rdx,%rsi
# qhasm: *(uint32 *)( input_0 + 152 ) = s
# asm 1: movl <s=int64#2d,152(<input_0=int64#1)
# asm 2: movl <s=%esi,152(<input_0=%rdi)
movl %esi,152(%rdi)
# qhasm: return
add %r11,%rsp
ret