lokinet/crypto/curve25519/sandy2x/fe51_pack.S

#ifdef IN_SANDY2X

/*
   This file is the result of merging 
   amd64-51/fe25519_pack.c and amd64-51/fe25519_freeze.s.
*/
#include "fe51_namespace.h"
#include "consts_namespace.h"
.p2align 5

#ifdef ASM_HIDE_SYMBOL
ASM_HIDE_SYMBOL fe51_pack
ASM_HIDE_SYMBOL _fe51_pack
#endif
.globl fe51_pack
.globl _fe51_pack
#ifdef __ELF__
.type  fe51_pack, @function
.type _fe51_pack, @function
#endif
fe51_pack:
_fe51_pack:

mov %rsp,%r11
and $31,%r11
add $32,%r11
sub %r11,%rsp
movq %r11,0(%rsp)
movq %r12,8(%rsp)
movq   0(%rsi),%rdx
movq   8(%rsi),%rcx
movq   16(%rsi),%r8
movq   24(%rsi),%r9
movq   32(%rsi),%rsi
movq REDMASK51(%rip),%rax
lea  -18(%rax),%r10
mov  $3,%r11

.p2align 4
._reduceloop:
mov  %rdx,%r12
shr  $51,%r12
and  %rax,%rdx
add  %r12,%rcx
mov  %rcx,%r12
shr  $51,%r12
and  %rax,%rcx
add  %r12,%r8
mov  %r8,%r12
shr  $51,%r12
and  %rax,%r8
add  %r12,%r9
mov  %r9,%r12
shr  $51,%r12
and  %rax,%r9
add  %r12,%rsi
mov  %rsi,%r12
shr  $51,%r12
and  %rax,%rsi
imulq  $19, %r12,%r12
add  %r12,%rdx
sub  $1,%r11
ja ._reduceloop

mov  $1,%r12
cmp  %r10,%rdx
cmovl %r11,%r12
cmp  %rax,%rcx
cmovne %r11,%r12
cmp  %rax,%r8
cmovne %r11,%r12
cmp  %rax,%r9
cmovne %r11,%r12
cmp  %rax,%rsi
cmovne %r11,%r12
neg  %r12
and  %r12,%rax
and  %r12,%r10
sub  %r10,%rdx
sub  %rax,%rcx
sub  %rax,%r8
sub  %rax,%r9
sub  %rax,%rsi
mov  %rdx,%rax
and  $0xFF,%eax
movb   %al,0(%rdi)
mov  %rdx,%rax
shr  $8,%rax
and  $0xFF,%eax
movb   %al,1(%rdi)
mov  %rdx,%rax
shr  $16,%rax
and  $0xFF,%eax
movb   %al,2(%rdi)
mov  %rdx,%rax
shr  $24,%rax
and  $0xFF,%eax
movb   %al,3(%rdi)
mov  %rdx,%rax
shr  $32,%rax
and  $0xFF,%eax
movb   %al,4(%rdi)
mov  %rdx,%rax
shr  $40,%rax
and  $0xFF,%eax
movb   %al,5(%rdi)
mov  %rdx,%rdx
shr  $48,%rdx
mov  %rcx,%rax
shl  $3,%rax
and  $0xF8,%eax
xor  %rdx,%rax
movb   %al,6(%rdi)
mov  %rcx,%rdx
shr  $5,%rdx
and  $0xFF,%edx
movb   %dl,7(%rdi)
mov  %rcx,%rdx
shr  $13,%rdx
and  $0xFF,%edx
movb   %dl,8(%rdi)
mov  %rcx,%rdx
shr  $21,%rdx
and  $0xFF,%edx
movb   %dl,9(%rdi)
mov  %rcx,%rdx
shr  $29,%rdx
and  $0xFF,%edx
movb   %dl,10(%rdi)
mov  %rcx,%rdx
shr  $37,%rdx
and  $0xFF,%edx
movb   %dl,11(%rdi)
mov  %rcx,%rdx
shr  $45,%rdx
mov  %r8,%rcx
shl  $6,%rcx
and  $0xC0,%ecx
xor  %rdx,%rcx
movb   %cl,12(%rdi)
mov  %r8,%rdx
shr  $2,%rdx
and  $0xFF,%edx
movb   %dl,13(%rdi)
mov  %r8,%rdx
shr  $10,%rdx
and  $0xFF,%edx
movb   %dl,14(%rdi)
mov  %r8,%rdx
shr  $18,%rdx
and  $0xFF,%edx
movb   %dl,15(%rdi)
mov  %r8,%rdx
shr  $26,%rdx
and  $0xFF,%edx
movb   %dl,16(%rdi)
mov  %r8,%rdx
shr  $34,%rdx
and  $0xFF,%edx
movb   %dl,17(%rdi)
mov  %r8,%rdx
shr  $42,%rdx
movb   %dl,18(%rdi)
mov  %r8,%rdx
shr  $50,%rdx
mov  %r9,%rcx
shl  $1,%rcx
and  $0xFE,%ecx
xor  %rdx,%rcx
movb   %cl,19(%rdi)
mov  %r9,%rdx
shr  $7,%rdx
and  $0xFF,%edx
movb   %dl,20(%rdi)
mov  %r9,%rdx
shr  $15,%rdx
and  $0xFF,%edx
movb   %dl,21(%rdi)
mov  %r9,%rdx
shr  $23,%rdx
and  $0xFF,%edx
movb   %dl,22(%rdi)
mov  %r9,%rdx
shr  $31,%rdx
and  $0xFF,%edx
movb   %dl,23(%rdi)
mov  %r9,%rdx
shr  $39,%rdx
and  $0xFF,%edx
movb   %dl,24(%rdi)
mov  %r9,%rdx
shr  $47,%rdx
mov  %rsi,%rcx
shl  $4,%rcx
and  $0xF0,%ecx
xor  %rdx,%rcx
movb   %cl,25(%rdi)
mov  %rsi,%rdx
shr  $4,%rdx
and  $0xFF,%edx
movb   %dl,26(%rdi)
mov  %rsi,%rdx
shr  $12,%rdx
and  $0xFF,%edx
movb   %dl,27(%rdi)
mov  %rsi,%rdx
shr  $20,%rdx
and  $0xFF,%edx
movb   %dl,28(%rdi)
mov  %rsi,%rdx
shr  $28,%rdx
and  $0xFF,%edx
movb   %dl,29(%rdi)
mov  %rsi,%rdx
shr  $36,%rdx
and  $0xFF,%edx
movb   %dl,30(%rdi)
mov  %rsi,%rsi
shr  $44,%rsi
movb   %sil,31(%rdi)
movq 0(%rsp),%r11
movq 8(%rsp),%r12
add %r11,%rsp
ret

#endif