mirror of
https://github.com/luxfi/pqclean.git
synced 2026-11-03 09:57:54 +00:00
Thom WiggersandGitHub
60f4825f49
* Update Classic McEliece This updates classic mceliece to the SUPERCOP-20221025 version. I'm generating this version from the SUPERCOP source code, and apply patches that are listed in the repository. The repository where this codes is coming from can be found at https://github.com/thomwiggers/mceliece-clean. * Fix linting warning in controlbits * MacOS compatibility fixes * Fix more MacOS conflicts * Fix strict prototypes problem
359 lines
8.6 KiB
ArmAsm
359 lines
8.6 KiB
ArmAsm
#include "namespace.h"
|
|
|
|
#define update_asm CRYPTO_NAMESPACE(update_asm)
|
|
#define _update_asm _CRYPTO_NAMESPACE(update_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 s0
|
|
|
|
# qhasm: int64 s1
|
|
|
|
# qhasm: enter update_asm
|
|
.p2align 5
|
|
.global _update_asm
|
|
.global update_asm
|
|
_update_asm:
|
|
update_asm:
|
|
mov % rsp, % r11
|
|
and $31, % r11
|
|
add $0, % r11
|
|
sub % r11, % rsp
|
|
|
|
# qhasm: s1 = input_1
|
|
# asm 1: mov <input_1=int64#2,>s1=int64#2
|
|
# asm 2: mov <input_1=%rsi,>s1=%rsi
|
|
mov % rsi, % rsi
|
|
|
|
# qhasm: s0 = mem64[ input_0 + 0 ]
|
|
# asm 1: movq 0(<input_0=int64#1),>s0=int64#4
|
|
# asm 2: movq 0(<input_0=%rdi),>s0=%rcx
|
|
movq 0( % rdi), % rcx
|
|
|
|
# qhasm: s0 = (s1 s0) >> 1
|
|
# asm 1: shrd $1,<s1=int64#2,<s0=int64#4
|
|
# asm 2: shrd $1,<s1=%rsi,<s0=%rcx
|
|
shrd $1, % rsi, % rcx
|
|
|
|
# qhasm: (uint64) s1 >>= 1
|
|
# asm 1: shr $1,<s1=int64#2
|
|
# asm 2: shr $1,<s1=%rsi
|
|
shr $1, % rsi
|
|
|
|
# qhasm: mem64[ input_0 + 0 ] = s0
|
|
# asm 1: movq <s0=int64#4,0(<input_0=int64#1)
|
|
# asm 2: movq <s0=%rcx,0(<input_0=%rdi)
|
|
movq % rcx, 0( % rdi)
|
|
|
|
# qhasm: input_0 += input_2
|
|
# asm 1: add <input_2=int64#3,<input_0=int64#1
|
|
# asm 2: add <input_2=%rdx,<input_0=%rdi
|
|
add % rdx, % rdi
|
|
|
|
# qhasm: s0 = mem64[ input_0 + 0 ]
|
|
# asm 1: movq 0(<input_0=int64#1),>s0=int64#4
|
|
# asm 2: movq 0(<input_0=%rdi),>s0=%rcx
|
|
movq 0( % rdi), % rcx
|
|
|
|
# qhasm: s0 = (s1 s0) >> 1
|
|
# asm 1: shrd $1,<s1=int64#2,<s0=int64#4
|
|
# asm 2: shrd $1,<s1=%rsi,<s0=%rcx
|
|
shrd $1, % rsi, % rcx
|
|
|
|
# qhasm: (uint64) s1 >>= 1
|
|
# asm 1: shr $1,<s1=int64#2
|
|
# asm 2: shr $1,<s1=%rsi
|
|
shr $1, % rsi
|
|
|
|
# qhasm: mem64[ input_0 + 0 ] = s0
|
|
# asm 1: movq <s0=int64#4,0(<input_0=int64#1)
|
|
# asm 2: movq <s0=%rcx,0(<input_0=%rdi)
|
|
movq % rcx, 0( % rdi)
|
|
|
|
# qhasm: input_0 += input_2
|
|
# asm 1: add <input_2=int64#3,<input_0=int64#1
|
|
# asm 2: add <input_2=%rdx,<input_0=%rdi
|
|
add % rdx, % rdi
|
|
|
|
# qhasm: s0 = mem64[ input_0 + 0 ]
|
|
# asm 1: movq 0(<input_0=int64#1),>s0=int64#4
|
|
# asm 2: movq 0(<input_0=%rdi),>s0=%rcx
|
|
movq 0( % rdi), % rcx
|
|
|
|
# qhasm: s0 = (s1 s0) >> 1
|
|
# asm 1: shrd $1,<s1=int64#2,<s0=int64#4
|
|
# asm 2: shrd $1,<s1=%rsi,<s0=%rcx
|
|
shrd $1, % rsi, % rcx
|
|
|
|
# qhasm: (uint64) s1 >>= 1
|
|
# asm 1: shr $1,<s1=int64#2
|
|
# asm 2: shr $1,<s1=%rsi
|
|
shr $1, % rsi
|
|
|
|
# qhasm: mem64[ input_0 + 0 ] = s0
|
|
# asm 1: movq <s0=int64#4,0(<input_0=int64#1)
|
|
# asm 2: movq <s0=%rcx,0(<input_0=%rdi)
|
|
movq % rcx, 0( % rdi)
|
|
|
|
# qhasm: input_0 += input_2
|
|
# asm 1: add <input_2=int64#3,<input_0=int64#1
|
|
# asm 2: add <input_2=%rdx,<input_0=%rdi
|
|
add % rdx, % rdi
|
|
|
|
# qhasm: s0 = mem64[ input_0 + 0 ]
|
|
# asm 1: movq 0(<input_0=int64#1),>s0=int64#4
|
|
# asm 2: movq 0(<input_0=%rdi),>s0=%rcx
|
|
movq 0( % rdi), % rcx
|
|
|
|
# qhasm: s0 = (s1 s0) >> 1
|
|
# asm 1: shrd $1,<s1=int64#2,<s0=int64#4
|
|
# asm 2: shrd $1,<s1=%rsi,<s0=%rcx
|
|
shrd $1, % rsi, % rcx
|
|
|
|
# qhasm: (uint64) s1 >>= 1
|
|
# asm 1: shr $1,<s1=int64#2
|
|
# asm 2: shr $1,<s1=%rsi
|
|
shr $1, % rsi
|
|
|
|
# qhasm: mem64[ input_0 + 0 ] = s0
|
|
# asm 1: movq <s0=int64#4,0(<input_0=int64#1)
|
|
# asm 2: movq <s0=%rcx,0(<input_0=%rdi)
|
|
movq % rcx, 0( % rdi)
|
|
|
|
# qhasm: input_0 += input_2
|
|
# asm 1: add <input_2=int64#3,<input_0=int64#1
|
|
# asm 2: add <input_2=%rdx,<input_0=%rdi
|
|
add % rdx, % rdi
|
|
|
|
# qhasm: s0 = mem64[ input_0 + 0 ]
|
|
# asm 1: movq 0(<input_0=int64#1),>s0=int64#4
|
|
# asm 2: movq 0(<input_0=%rdi),>s0=%rcx
|
|
movq 0( % rdi), % rcx
|
|
|
|
# qhasm: s0 = (s1 s0) >> 1
|
|
# asm 1: shrd $1,<s1=int64#2,<s0=int64#4
|
|
# asm 2: shrd $1,<s1=%rsi,<s0=%rcx
|
|
shrd $1, % rsi, % rcx
|
|
|
|
# qhasm: (uint64) s1 >>= 1
|
|
# asm 1: shr $1,<s1=int64#2
|
|
# asm 2: shr $1,<s1=%rsi
|
|
shr $1, % rsi
|
|
|
|
# qhasm: mem64[ input_0 + 0 ] = s0
|
|
# asm 1: movq <s0=int64#4,0(<input_0=int64#1)
|
|
# asm 2: movq <s0=%rcx,0(<input_0=%rdi)
|
|
movq % rcx, 0( % rdi)
|
|
|
|
# qhasm: input_0 += input_2
|
|
# asm 1: add <input_2=int64#3,<input_0=int64#1
|
|
# asm 2: add <input_2=%rdx,<input_0=%rdi
|
|
add % rdx, % rdi
|
|
|
|
# qhasm: s0 = mem64[ input_0 + 0 ]
|
|
# asm 1: movq 0(<input_0=int64#1),>s0=int64#4
|
|
# asm 2: movq 0(<input_0=%rdi),>s0=%rcx
|
|
movq 0( % rdi), % rcx
|
|
|
|
# qhasm: s0 = (s1 s0) >> 1
|
|
# asm 1: shrd $1,<s1=int64#2,<s0=int64#4
|
|
# asm 2: shrd $1,<s1=%rsi,<s0=%rcx
|
|
shrd $1, % rsi, % rcx
|
|
|
|
# qhasm: (uint64) s1 >>= 1
|
|
# asm 1: shr $1,<s1=int64#2
|
|
# asm 2: shr $1,<s1=%rsi
|
|
shr $1, % rsi
|
|
|
|
# qhasm: mem64[ input_0 + 0 ] = s0
|
|
# asm 1: movq <s0=int64#4,0(<input_0=int64#1)
|
|
# asm 2: movq <s0=%rcx,0(<input_0=%rdi)
|
|
movq % rcx, 0( % rdi)
|
|
|
|
# qhasm: input_0 += input_2
|
|
# asm 1: add <input_2=int64#3,<input_0=int64#1
|
|
# asm 2: add <input_2=%rdx,<input_0=%rdi
|
|
add % rdx, % rdi
|
|
|
|
# qhasm: s0 = mem64[ input_0 + 0 ]
|
|
# asm 1: movq 0(<input_0=int64#1),>s0=int64#4
|
|
# asm 2: movq 0(<input_0=%rdi),>s0=%rcx
|
|
movq 0( % rdi), % rcx
|
|
|
|
# qhasm: s0 = (s1 s0) >> 1
|
|
# asm 1: shrd $1,<s1=int64#2,<s0=int64#4
|
|
# asm 2: shrd $1,<s1=%rsi,<s0=%rcx
|
|
shrd $1, % rsi, % rcx
|
|
|
|
# qhasm: (uint64) s1 >>= 1
|
|
# asm 1: shr $1,<s1=int64#2
|
|
# asm 2: shr $1,<s1=%rsi
|
|
shr $1, % rsi
|
|
|
|
# qhasm: mem64[ input_0 + 0 ] = s0
|
|
# asm 1: movq <s0=int64#4,0(<input_0=int64#1)
|
|
# asm 2: movq <s0=%rcx,0(<input_0=%rdi)
|
|
movq % rcx, 0( % rdi)
|
|
|
|
# qhasm: input_0 += input_2
|
|
# asm 1: add <input_2=int64#3,<input_0=int64#1
|
|
# asm 2: add <input_2=%rdx,<input_0=%rdi
|
|
add % rdx, % rdi
|
|
|
|
# qhasm: s0 = mem64[ input_0 + 0 ]
|
|
# asm 1: movq 0(<input_0=int64#1),>s0=int64#4
|
|
# asm 2: movq 0(<input_0=%rdi),>s0=%rcx
|
|
movq 0( % rdi), % rcx
|
|
|
|
# qhasm: s0 = (s1 s0) >> 1
|
|
# asm 1: shrd $1,<s1=int64#2,<s0=int64#4
|
|
# asm 2: shrd $1,<s1=%rsi,<s0=%rcx
|
|
shrd $1, % rsi, % rcx
|
|
|
|
# qhasm: (uint64) s1 >>= 1
|
|
# asm 1: shr $1,<s1=int64#2
|
|
# asm 2: shr $1,<s1=%rsi
|
|
shr $1, % rsi
|
|
|
|
# qhasm: mem64[ input_0 + 0 ] = s0
|
|
# asm 1: movq <s0=int64#4,0(<input_0=int64#1)
|
|
# asm 2: movq <s0=%rcx,0(<input_0=%rdi)
|
|
movq % rcx, 0( % rdi)
|
|
|
|
# qhasm: input_0 += input_2
|
|
# asm 1: add <input_2=int64#3,<input_0=int64#1
|
|
# asm 2: add <input_2=%rdx,<input_0=%rdi
|
|
add % rdx, % rdi
|
|
|
|
# qhasm: s0 = mem64[ input_0 + 0 ]
|
|
# asm 1: movq 0(<input_0=int64#1),>s0=int64#4
|
|
# asm 2: movq 0(<input_0=%rdi),>s0=%rcx
|
|
movq 0( % rdi), % rcx
|
|
|
|
# qhasm: s0 = (s1 s0) >> 1
|
|
# asm 1: shrd $1,<s1=int64#2,<s0=int64#4
|
|
# asm 2: shrd $1,<s1=%rsi,<s0=%rcx
|
|
shrd $1, % rsi, % rcx
|
|
|
|
# qhasm: (uint64) s1 >>= 1
|
|
# asm 1: shr $1,<s1=int64#2
|
|
# asm 2: shr $1,<s1=%rsi
|
|
shr $1, % rsi
|
|
|
|
# qhasm: mem64[ input_0 + 0 ] = s0
|
|
# asm 1: movq <s0=int64#4,0(<input_0=int64#1)
|
|
# asm 2: movq <s0=%rcx,0(<input_0=%rdi)
|
|
movq % rcx, 0( % rdi)
|
|
|
|
# qhasm: input_0 += input_2
|
|
# asm 1: add <input_2=int64#3,<input_0=int64#1
|
|
# asm 2: add <input_2=%rdx,<input_0=%rdi
|
|
add % rdx, % rdi
|
|
|
|
# qhasm: s0 = mem64[ input_0 + 0 ]
|
|
# asm 1: movq 0(<input_0=int64#1),>s0=int64#4
|
|
# asm 2: movq 0(<input_0=%rdi),>s0=%rcx
|
|
movq 0( % rdi), % rcx
|
|
|
|
# qhasm: s0 = (s1 s0) >> 1
|
|
# asm 1: shrd $1,<s1=int64#2,<s0=int64#4
|
|
# asm 2: shrd $1,<s1=%rsi,<s0=%rcx
|
|
shrd $1, % rsi, % rcx
|
|
|
|
# qhasm: (uint64) s1 >>= 1
|
|
# asm 1: shr $1,<s1=int64#2
|
|
# asm 2: shr $1,<s1=%rsi
|
|
shr $1, % rsi
|
|
|
|
# qhasm: mem64[ input_0 + 0 ] = s0
|
|
# asm 1: movq <s0=int64#4,0(<input_0=int64#1)
|
|
# asm 2: movq <s0=%rcx,0(<input_0=%rdi)
|
|
movq % rcx, 0( % rdi)
|
|
|
|
# qhasm: input_0 += input_2
|
|
# asm 1: add <input_2=int64#3,<input_0=int64#1
|
|
# asm 2: add <input_2=%rdx,<input_0=%rdi
|
|
add % rdx, % rdi
|
|
|
|
# qhasm: s0 = mem64[ input_0 + 0 ]
|
|
# asm 1: movq 0(<input_0=int64#1),>s0=int64#4
|
|
# asm 2: movq 0(<input_0=%rdi),>s0=%rcx
|
|
movq 0( % rdi), % rcx
|
|
|
|
# qhasm: s0 = (s1 s0) >> 1
|
|
# asm 1: shrd $1,<s1=int64#2,<s0=int64#4
|
|
# asm 2: shrd $1,<s1=%rsi,<s0=%rcx
|
|
shrd $1, % rsi, % rcx
|
|
|
|
# qhasm: (uint64) s1 >>= 1
|
|
# asm 1: shr $1,<s1=int64#2
|
|
# asm 2: shr $1,<s1=%rsi
|
|
shr $1, % rsi
|
|
|
|
# qhasm: mem64[ input_0 + 0 ] = s0
|
|
# asm 1: movq <s0=int64#4,0(<input_0=int64#1)
|
|
# asm 2: movq <s0=%rcx,0(<input_0=%rdi)
|
|
movq % rcx, 0( % rdi)
|
|
|
|
# qhasm: input_0 += input_2
|
|
# asm 1: add <input_2=int64#3,<input_0=int64#1
|
|
# asm 2: add <input_2=%rdx,<input_0=%rdi
|
|
add % rdx, % rdi
|
|
|
|
# qhasm: s0 = mem64[ input_0 + 0 ]
|
|
# asm 1: movq 0(<input_0=int64#1),>s0=int64#4
|
|
# asm 2: movq 0(<input_0=%rdi),>s0=%rcx
|
|
movq 0( % rdi), % rcx
|
|
|
|
# qhasm: s0 = (s1 s0) >> 1
|
|
# asm 1: shrd $1,<s1=int64#2,<s0=int64#4
|
|
# asm 2: shrd $1,<s1=%rsi,<s0=%rcx
|
|
shrd $1, % rsi, % rcx
|
|
|
|
# qhasm: (uint64) s1 >>= 1
|
|
# asm 1: shr $1,<s1=int64#2
|
|
# asm 2: shr $1,<s1=%rsi
|
|
shr $1, % rsi
|
|
|
|
# qhasm: mem64[ input_0 + 0 ] = s0
|
|
# asm 1: movq <s0=int64#4,0(<input_0=int64#1)
|
|
# asm 2: movq <s0=%rcx,0(<input_0=%rdi)
|
|
movq % rcx, 0( % rdi)
|
|
|
|
# qhasm: input_0 += input_2
|
|
# asm 1: add <input_2=int64#3,<input_0=int64#1
|
|
# asm 2: add <input_2=%rdx,<input_0=%rdi
|
|
add % rdx, % rdi
|
|
|
|
# qhasm: return
|
|
add % r11, % rsp
|
|
ret
|