mirror of
https://github.com/guanzhi/GmSSL.git
synced 2026-07-01 18:23:38 +08:00
quantum init
This commit is contained in:
zhaoxiaomeng
@@ -1,11 +1,4 @@
|
||||
#! /usr/bin/env perl
|
||||
# Copyright 2014-2016 The OpenSSL Project Authors. All Rights Reserved.
|
||||
#
|
||||
# Licensed under the OpenSSL license (the "License"). You may not use
|
||||
# this file except in compliance with the License. You can obtain a copy
|
||||
# in the file LICENSE in the source distribution or at
|
||||
# https://www.openssl.org/source/license.html
|
||||
|
||||
#!/usr/bin/env perl
|
||||
|
||||
##############################################################################
|
||||
# #
|
||||
@@ -67,7 +60,7 @@ $0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
|
||||
( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
|
||||
die "can't locate x86_64-xlate.pl";
|
||||
|
||||
open OUT,"| \"$^X\" \"$xlate\" $flavour \"$output\"";
|
||||
open OUT,"| \"$^X\" $xlate $flavour $output";
|
||||
*STDOUT=*OUT;
|
||||
|
||||
if (`$ENV{CC} -Wa,-v -c -o /dev/null -x assembler /dev/null 2>&1`
|
||||
@@ -1185,19 +1178,18 @@ __ecp_nistz256_sqr_montx:
|
||||
adox $t1, $acc5
|
||||
.byte 0x67,0x67
|
||||
mulx %rdx, $t0, $t4
|
||||
mov $acc0, %rdx
|
||||
mov .Lpoly+8*3(%rip), %rdx
|
||||
adox $t0, $acc6
|
||||
shlx $a_ptr, $acc0, $t0
|
||||
adox $t4, $acc7
|
||||
shrx $a_ptr, $acc0, $t4
|
||||
mov .Lpoly+8*3(%rip), $t1
|
||||
mov %rdx,$t1
|
||||
|
||||
# reduction step 1
|
||||
add $t0, $acc1
|
||||
adc $t4, $acc2
|
||||
|
||||
mulx $t1, $t0, $acc0
|
||||
mov $acc1, %rdx
|
||||
mulx $acc0, $t0, $acc0
|
||||
adc $t0, $acc3
|
||||
shlx $a_ptr, $acc1, $t0
|
||||
adc \$0, $acc0
|
||||
@@ -1207,8 +1199,7 @@ __ecp_nistz256_sqr_montx:
|
||||
add $t0, $acc2
|
||||
adc $t4, $acc3
|
||||
|
||||
mulx $t1, $t0, $acc1
|
||||
mov $acc2, %rdx
|
||||
mulx $acc1, $t0, $acc1
|
||||
adc $t0, $acc0
|
||||
shlx $a_ptr, $acc2, $t0
|
||||
adc \$0, $acc1
|
||||
@@ -1218,8 +1209,7 @@ __ecp_nistz256_sqr_montx:
|
||||
add $t0, $acc3
|
||||
adc $t4, $acc0
|
||||
|
||||
mulx $t1, $t0, $acc2
|
||||
mov $acc3, %rdx
|
||||
mulx $acc2, $t0, $acc2
|
||||
adc $t0, $acc1
|
||||
shlx $a_ptr, $acc3, $t0
|
||||
adc \$0, $acc2
|
||||
@@ -1229,12 +1219,12 @@ __ecp_nistz256_sqr_montx:
|
||||
add $t0, $acc0
|
||||
adc $t4, $acc1
|
||||
|
||||
mulx $t1, $t0, $acc3
|
||||
mulx $acc3, $t0, $acc3
|
||||
adc $t0, $acc2
|
||||
adc \$0, $acc3
|
||||
|
||||
xor $t3, $t3 # cf=0
|
||||
adc $acc0, $acc4 # accumulate upper half
|
||||
xor $t3, $t3
|
||||
add $acc0, $acc4 # accumulate upper half
|
||||
mov .Lpoly+8*1(%rip), $a_ptr
|
||||
adc $acc1, $acc5
|
||||
mov $acc4, $acc0
|
||||
@@ -1243,8 +1233,7 @@ __ecp_nistz256_sqr_montx:
|
||||
mov $acc5, $acc1
|
||||
adc \$0, $t3
|
||||
|
||||
xor %eax, %eax # cf=0
|
||||
sbb \$-1, $acc4 # .Lpoly[0]
|
||||
sub \$-1, $acc4 # .Lpoly[0]
|
||||
mov $acc6, $acc2
|
||||
sbb $a_ptr, $acc5 # .Lpoly[1]
|
||||
sbb \$0, $acc6 # .Lpoly[2]
|
||||
@@ -1378,44 +1367,20 @@ my ($M1,$T2a,$T2b,$TMP2,$M2,$T2a,$T2b,$TMP2)=map("%xmm$_",(8..15));
|
||||
|
||||
$code.=<<___;
|
||||
################################################################################
|
||||
# void ecp_nistz256_scatter_w5(uint64_t *val, uint64_t *in_t, int index);
|
||||
.globl ecp_nistz256_scatter_w5
|
||||
.type ecp_nistz256_scatter_w5,\@abi-omnipotent
|
||||
# void ecp_nistz256_select_w5(uint64_t *val, uint64_t *in_t, int index);
|
||||
.globl ecp_nistz256_select_w5
|
||||
.type ecp_nistz256_select_w5,\@abi-omnipotent
|
||||
.align 32
|
||||
ecp_nistz256_scatter_w5:
|
||||
lea -3($index,$index,2), $index
|
||||
movdqa 0x00($in_t), %xmm0
|
||||
shl \$5, $index
|
||||
movdqa 0x10($in_t), %xmm1
|
||||
movdqa 0x20($in_t), %xmm2
|
||||
movdqa 0x30($in_t), %xmm3
|
||||
movdqa 0x40($in_t), %xmm4
|
||||
movdqa 0x50($in_t), %xmm5
|
||||
movdqa %xmm0, 0x00($val,$index)
|
||||
movdqa %xmm1, 0x10($val,$index)
|
||||
movdqa %xmm2, 0x20($val,$index)
|
||||
movdqa %xmm3, 0x30($val,$index)
|
||||
movdqa %xmm4, 0x40($val,$index)
|
||||
movdqa %xmm5, 0x50($val,$index)
|
||||
|
||||
ret
|
||||
.size ecp_nistz256_scatter_w5,.-ecp_nistz256_scatter_w5
|
||||
|
||||
################################################################################
|
||||
# void ecp_nistz256_gather_w5(uint64_t *val, uint64_t *in_t, int index);
|
||||
.globl ecp_nistz256_gather_w5
|
||||
.type ecp_nistz256_gather_w5,\@abi-omnipotent
|
||||
.align 32
|
||||
ecp_nistz256_gather_w5:
|
||||
ecp_nistz256_select_w5:
|
||||
___
|
||||
$code.=<<___ if ($avx>1);
|
||||
mov OPENSSL_ia32cap_P+8(%rip), %eax
|
||||
test \$`1<<5`, %eax
|
||||
jnz .Lavx2_gather_w5
|
||||
jnz .Lavx2_select_w5
|
||||
___
|
||||
$code.=<<___ if ($win64);
|
||||
lea -0x88(%rsp), %rax
|
||||
.LSEH_begin_ecp_nistz256_gather_w5:
|
||||
.LSEH_begin_ecp_nistz256_select_w5:
|
||||
.byte 0x48,0x8d,0x60,0xe0 #lea -0x20(%rax), %rsp
|
||||
.byte 0x0f,0x29,0x70,0xe0 #movaps %xmm6, -0x20(%rax)
|
||||
.byte 0x0f,0x29,0x78,0xf0 #movaps %xmm7, -0x10(%rax)
|
||||
@@ -1492,46 +1457,27 @@ $code.=<<___ if ($win64);
|
||||
movaps 0x80(%rsp), %xmm14
|
||||
movaps 0x90(%rsp), %xmm15
|
||||
lea 0xa8(%rsp), %rsp
|
||||
.LSEH_end_ecp_nistz256_gather_w5:
|
||||
.LSEH_end_ecp_nistz256_select_w5:
|
||||
___
|
||||
$code.=<<___;
|
||||
ret
|
||||
.size ecp_nistz256_gather_w5,.-ecp_nistz256_gather_w5
|
||||
.size ecp_nistz256_select_w5,.-ecp_nistz256_select_w5
|
||||
|
||||
################################################################################
|
||||
# void ecp_nistz256_scatter_w7(uint64_t *val, uint64_t *in_t, int index);
|
||||
.globl ecp_nistz256_scatter_w7
|
||||
.type ecp_nistz256_scatter_w7,\@abi-omnipotent
|
||||
# void ecp_nistz256_select_w7(uint64_t *val, uint64_t *in_t, int index);
|
||||
.globl ecp_nistz256_select_w7
|
||||
.type ecp_nistz256_select_w7,\@abi-omnipotent
|
||||
.align 32
|
||||
ecp_nistz256_scatter_w7:
|
||||
movdqu 0x00($in_t), %xmm0
|
||||
shl \$6, $index
|
||||
movdqu 0x10($in_t), %xmm1
|
||||
movdqu 0x20($in_t), %xmm2
|
||||
movdqu 0x30($in_t), %xmm3
|
||||
movdqa %xmm0, 0x00($val,$index)
|
||||
movdqa %xmm1, 0x10($val,$index)
|
||||
movdqa %xmm2, 0x20($val,$index)
|
||||
movdqa %xmm3, 0x30($val,$index)
|
||||
|
||||
ret
|
||||
.size ecp_nistz256_scatter_w7,.-ecp_nistz256_scatter_w7
|
||||
|
||||
################################################################################
|
||||
# void ecp_nistz256_gather_w7(uint64_t *val, uint64_t *in_t, int index);
|
||||
.globl ecp_nistz256_gather_w7
|
||||
.type ecp_nistz256_gather_w7,\@abi-omnipotent
|
||||
.align 32
|
||||
ecp_nistz256_gather_w7:
|
||||
ecp_nistz256_select_w7:
|
||||
___
|
||||
$code.=<<___ if ($avx>1);
|
||||
mov OPENSSL_ia32cap_P+8(%rip), %eax
|
||||
test \$`1<<5`, %eax
|
||||
jnz .Lavx2_gather_w7
|
||||
jnz .Lavx2_select_w7
|
||||
___
|
||||
$code.=<<___ if ($win64);
|
||||
lea -0x88(%rsp), %rax
|
||||
.LSEH_begin_ecp_nistz256_gather_w7:
|
||||
.LSEH_begin_ecp_nistz256_select_w7:
|
||||
.byte 0x48,0x8d,0x60,0xe0 #lea -0x20(%rax), %rsp
|
||||
.byte 0x0f,0x29,0x70,0xe0 #movaps %xmm6, -0x20(%rax)
|
||||
.byte 0x0f,0x29,0x78,0xf0 #movaps %xmm7, -0x10(%rax)
|
||||
@@ -1597,11 +1543,11 @@ $code.=<<___ if ($win64);
|
||||
movaps 0x80(%rsp), %xmm14
|
||||
movaps 0x90(%rsp), %xmm15
|
||||
lea 0xa8(%rsp), %rsp
|
||||
.LSEH_end_ecp_nistz256_gather_w7:
|
||||
.LSEH_end_ecp_nistz256_select_w7:
|
||||
___
|
||||
$code.=<<___;
|
||||
ret
|
||||
.size ecp_nistz256_gather_w7,.-ecp_nistz256_gather_w7
|
||||
.size ecp_nistz256_select_w7,.-ecp_nistz256_select_w7
|
||||
___
|
||||
}
|
||||
if ($avx>1) {
|
||||
@@ -1612,16 +1558,16 @@ my ($M1,$T1a,$T1b,$T1c,$TMP1)=map("%ymm$_",(10..14));
|
||||
|
||||
$code.=<<___;
|
||||
################################################################################
|
||||
# void ecp_nistz256_avx2_gather_w5(uint64_t *val, uint64_t *in_t, int index);
|
||||
.type ecp_nistz256_avx2_gather_w5,\@abi-omnipotent
|
||||
# void ecp_nistz256_avx2_select_w5(uint64_t *val, uint64_t *in_t, int index);
|
||||
.type ecp_nistz256_avx2_select_w5,\@abi-omnipotent
|
||||
.align 32
|
||||
ecp_nistz256_avx2_gather_w5:
|
||||
.Lavx2_gather_w5:
|
||||
ecp_nistz256_avx2_select_w5:
|
||||
.Lavx2_select_w5:
|
||||
vzeroupper
|
||||
___
|
||||
$code.=<<___ if ($win64);
|
||||
lea -0x88(%rsp), %rax
|
||||
.LSEH_begin_ecp_nistz256_avx2_gather_w5:
|
||||
.LSEH_begin_ecp_nistz256_avx2_select_w5:
|
||||
.byte 0x48,0x8d,0x60,0xe0 #lea -0x20(%rax), %rsp
|
||||
.byte 0xc5,0xf8,0x29,0x70,0xe0 #vmovaps %xmm6, -0x20(%rax)
|
||||
.byte 0xc5,0xf8,0x29,0x78,0xf0 #vmovaps %xmm7, -0x10(%rax)
|
||||
@@ -1699,11 +1645,11 @@ $code.=<<___ if ($win64);
|
||||
movaps 0x80(%rsp), %xmm14
|
||||
movaps 0x90(%rsp), %xmm15
|
||||
lea 0xa8(%rsp), %rsp
|
||||
.LSEH_end_ecp_nistz256_avx2_gather_w5:
|
||||
.LSEH_end_ecp_nistz256_avx2_select_w5:
|
||||
___
|
||||
$code.=<<___;
|
||||
ret
|
||||
.size ecp_nistz256_avx2_gather_w5,.-ecp_nistz256_avx2_gather_w5
|
||||
.size ecp_nistz256_avx2_select_w5,.-ecp_nistz256_avx2_select_w5
|
||||
___
|
||||
}
|
||||
if ($avx>1) {
|
||||
@@ -1716,17 +1662,17 @@ my ($M2,$T2a,$T2b,$TMP2)=map("%ymm$_",(12..15));
|
||||
$code.=<<___;
|
||||
|
||||
################################################################################
|
||||
# void ecp_nistz256_avx2_gather_w7(uint64_t *val, uint64_t *in_t, int index);
|
||||
.globl ecp_nistz256_avx2_gather_w7
|
||||
.type ecp_nistz256_avx2_gather_w7,\@abi-omnipotent
|
||||
# void ecp_nistz256_avx2_select_w7(uint64_t *val, uint64_t *in_t, int index);
|
||||
.globl ecp_nistz256_avx2_select_w7
|
||||
.type ecp_nistz256_avx2_select_w7,\@abi-omnipotent
|
||||
.align 32
|
||||
ecp_nistz256_avx2_gather_w7:
|
||||
.Lavx2_gather_w7:
|
||||
ecp_nistz256_avx2_select_w7:
|
||||
.Lavx2_select_w7:
|
||||
vzeroupper
|
||||
___
|
||||
$code.=<<___ if ($win64);
|
||||
lea -0x88(%rsp), %rax
|
||||
.LSEH_begin_ecp_nistz256_avx2_gather_w7:
|
||||
.LSEH_begin_ecp_nistz256_avx2_select_w7:
|
||||
.byte 0x48,0x8d,0x60,0xe0 #lea -0x20(%rax), %rsp
|
||||
.byte 0xc5,0xf8,0x29,0x70,0xe0 #vmovaps %xmm6, -0x20(%rax)
|
||||
.byte 0xc5,0xf8,0x29,0x78,0xf0 #vmovaps %xmm7, -0x10(%rax)
|
||||
@@ -1819,21 +1765,21 @@ $code.=<<___ if ($win64);
|
||||
movaps 0x80(%rsp), %xmm14
|
||||
movaps 0x90(%rsp), %xmm15
|
||||
lea 0xa8(%rsp), %rsp
|
||||
.LSEH_end_ecp_nistz256_avx2_gather_w7:
|
||||
.LSEH_end_ecp_nistz256_avx2_select_w7:
|
||||
___
|
||||
$code.=<<___;
|
||||
ret
|
||||
.size ecp_nistz256_avx2_gather_w7,.-ecp_nistz256_avx2_gather_w7
|
||||
.size ecp_nistz256_avx2_select_w7,.-ecp_nistz256_avx2_select_w7
|
||||
___
|
||||
} else {
|
||||
$code.=<<___;
|
||||
.globl ecp_nistz256_avx2_gather_w7
|
||||
.type ecp_nistz256_avx2_gather_w7,\@function,3
|
||||
.globl ecp_nistz256_avx2_select_w7
|
||||
.type ecp_nistz256_avx2_select_w7,\@function,3
|
||||
.align 32
|
||||
ecp_nistz256_avx2_gather_w7:
|
||||
ecp_nistz256_avx2_select_w7:
|
||||
.byte 0x0f,0x0b # ud2
|
||||
ret
|
||||
.size ecp_nistz256_avx2_gather_w7,.-ecp_nistz256_avx2_gather_w7
|
||||
.size ecp_nistz256_avx2_select_w7,.-ecp_nistz256_avx2_select_w7
|
||||
___
|
||||
}
|
||||
{{{
|
||||
@@ -3052,36 +2998,6 @@ ___
|
||||
}
|
||||
}}}
|
||||
|
||||
########################################################################
|
||||
# Convert ecp_nistz256_table.c to layout expected by ecp_nistz_gather_w7
|
||||
#
|
||||
open TABLE,"<ecp_nistz256_table.c" or
|
||||
open TABLE,"<${dir}../ecp_nistz256_table.c" or
|
||||
die "failed to open ecp_nistz256_table.c:",$!;
|
||||
|
||||
use integer;
|
||||
|
||||
foreach(<TABLE>) {
|
||||
s/TOBN\(\s*(0x[0-9a-f]+),\s*(0x[0-9a-f]+)\s*\)/push @arr,hex($2),hex($1)/geo;
|
||||
}
|
||||
close TABLE;
|
||||
|
||||
die "insane number of elements" if ($#arr != 64*16*37-1);
|
||||
|
||||
print <<___;
|
||||
.text
|
||||
.globl ecp_nistz256_precomputed
|
||||
.type ecp_nistz256_precomputed,\@object
|
||||
.align 4096
|
||||
ecp_nistz256_precomputed:
|
||||
___
|
||||
while (@line=splice(@arr,0,16)) {
|
||||
print ".long\t",join(',',map { sprintf "0x%08x",$_} @line),"\n";
|
||||
}
|
||||
print <<___;
|
||||
.size ecp_nistz256_precomputed,.-ecp_nistz256_precomputed
|
||||
___
|
||||
|
||||
$code =~ s/\`([^\`]*)\`/eval $1/gem;
|
||||
print $code;
|
||||
close STDOUT;
|
||||
|
||||
Reference in New Issue
Block a user