quantum init

This commit is contained in:
zhaoxiaomeng
2018-01-04 13:38:57 +08:00
committed by Simon
parent d11f845fde
commit 53af3b51ae
2361 changed files with 387455 additions and 144458 deletions

View File

@@ -1,11 +1,4 @@
#! /usr/bin/env perl
# Copyright 1995-2016 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/local/bin/perl
# Normal is the
# md5_block_x86(MD5_CTX *c, ULONG *X);
@@ -18,9 +11,6 @@ $0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
push(@INC,"${dir}","${dir}../../perlasm");
require "x86asm.pl";
$output=pop;
open STDOUT,">$output";
&asm_init($ARGV[0],$0);
$A="eax";
@@ -43,8 +33,6 @@ $X="esi";
&md5_block("md5_block_asm_data_order");
&asm_finish();
close STDOUT;
sub Np
{
local($p)=@_;
@@ -68,14 +56,14 @@ sub R0
&lea($a,&DWP($t,$a,$tmp2,1));
&xor($tmp1,$d); # F function - part 4
&mov($tmp2,&DWP($xo[$ki+1]*4,$K,"",0)) if ($pos != 2);
&add($a,$tmp1);
&mov($tmp1,&Np($c)) if $pos < 1; # next tmp1 for R0
&mov($tmp1,&Np($c)) if $pos == 1; # next tmp1 for R1
&rotl($a,$s);
&mov($tmp1,&Np($c)) if $pos < 1; # next tmp1 for R0
&mov($tmp1,&Np($c)) if $pos == 1; # next tmp1 for R1
&mov($tmp2,&DWP($xo[$ki+1]*4,$K,"",0)) if ($pos != 2);
&add($a,$b);
}
@@ -86,12 +74,13 @@ sub R1
&comment("R1 $ki");
&xor($tmp1,$b); # G function - part 2
&and($tmp1,$d); # G function - part 3
&lea($a,&DWP($t,$a,$tmp2,1));
&xor($tmp1,$c); # G function - part 4
&xor($tmp1,$b); # G function - part 2
&and($tmp1,$d); # G function - part 3
&mov($tmp2,&DWP($xo[$ki+1]*4,$K,"",0)) if ($pos != 2);
&xor($tmp1,$c); # G function - part 4
&add($a,$tmp1);
&mov($tmp1,&Np($c)) if $pos < 1; # G function - part 1
@@ -119,10 +108,10 @@ if (($n & 1) == 0)
&lea($a,&DWP($t,$a,$tmp2,1));
&add($a,$tmp1);
&mov($tmp2,&DWP($xo[$ki+1]*4,$K,"",0));
&rotl($a,$s);
&mov($tmp2,&DWP($xo[$ki+1]*4,$K,"",0));
&mov($tmp1,&Np($c));
}
else
@@ -131,11 +120,11 @@ else
# make sure to do 'D' first, not 'B', else we clash with
# the last add from the previous round.
&lea($a,&DWP($t,$a,$tmp2,1));
&add($b,$c); # MOVED FORWARD
&xor($tmp1,$d); # H function - part 2
&lea($a,&DWP($t,$a,$tmp2,1));
&xor($tmp1,$b); # H function - part 3
&mov($tmp2,&DWP($xo[$ki+1]*4,$K,"",0)) if ($pos != 2);

View File

@@ -1,13 +1,3 @@
/*
*
* Copyright 2005-2016 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
*/
/* Copyright (c) 2005 Hewlett-Packard Development Company, L.P.
Permission is hereby granted, free of charge, to any person obtaining

View File

@@ -1,11 +1,4 @@
#! /usr/bin/env perl
# Copyright 2012-2016 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
# ====================================================================
# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
@@ -24,7 +17,7 @@
# single-process result on 8-core processor, or ~11GBps per 2.85GHz
# socket.
$output=pop;
$output=shift;
open STDOUT,">$output";
use integer;
@@ -378,7 +371,7 @@ ___
# Purpose of these subroutines is to explicitly encode VIS instructions,
# so that one can compile the module without having to specify VIS
# extensions on compiler command line, e.g. -xarch=v9 vs. -xarch=v9a.
# extentions on compiler command line, e.g. -xarch=v9 vs. -xarch=v9a.
# Idea is to reserve for option to produce "universal" binary and let
# programmer detect if current CPU is VIS capable at run-time.
sub unvis {

View File

@@ -1,13 +1,11 @@
#! /usr/bin/env perl
# Author: Marc Bevand <bevand_m (at) epita.fr>
# Copyright 2005-2016 The OpenSSL Project Authors. All Rights Reserved.
#!/usr/bin/perl -w
#
# 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
# MD5 optimized for AMD64.
#
# Author: Marc Bevand <bevand_m (at) epita.fr>
# Licence: I hereby disclaim the copyright on this code and place it
# in the public domain.
#
use strict;
@@ -27,8 +25,8 @@ sub round1_step
xor $y, %r11d /* y ^ ... */
lea $T_i($dst,%r10d),$dst /* Const + dst + ... */
and $x, %r11d /* x & ... */
mov $k_next*4(%rsi),%r10d /* (NEXT STEP) X[$k_next] */
xor $z, %r11d /* z ^ ... */
mov $k_next*4(%rsi),%r10d /* (NEXT STEP) X[$k_next] */
add %r11d, $dst /* dst += ... */
rol \$$s, $dst /* dst <<< s */
mov $y, %r11d /* (NEXT STEP) z' = $y */
@@ -45,12 +43,13 @@ EOF
sub round2_step
{
my ($pos, $dst, $x, $y, $z, $k_next, $T_i, $s) = @_;
$code .= " mov 1*4(%rsi), %r10d /* (NEXT STEP) X[1] */\n" if ($pos == -1);
$code .= " mov %edx, %r11d /* (NEXT STEP) z' = %edx */\n" if ($pos == -1);
$code .= " mov %edx, %r12d /* (NEXT STEP) z' = %edx */\n" if ($pos == -1);
$code .= <<EOF;
not %r11d /* not z */
and $x, %r12d /* x & z */
lea $T_i($dst,%r10d),$dst /* Const + dst + ... */
and $x, %r12d /* x & z */
and $y, %r11d /* y & (not z) */
mov $k_next*4(%rsi),%r10d /* (NEXT STEP) X[$k_next] */
or %r11d, %r12d /* (y & (not z)) | (x & z) */
@@ -67,31 +66,21 @@ EOF
# %r10d = X[k_next]
# %r11d = y' (copy of y for the next step)
# Each round3_step() takes about 4.2 clocks (8 instructions, 1.9 IPC)
{ my $round3_alter=0;
sub round3_step
{
my ($pos, $dst, $x, $y, $z, $k_next, $T_i, $s) = @_;
$code .= " mov 5*4(%rsi), %r10d /* (NEXT STEP) X[5] */\n" if ($pos == -1);
$code .= " mov %ecx, %r11d /* (NEXT STEP) y' = %ecx */\n" if ($pos == -1);
$code .= <<EOF;
lea $T_i($dst,%r10d),$dst /* Const + dst + ... */
xor $z, %r11d /* z ^ ... */
mov $k_next*4(%rsi),%r10d /* (NEXT STEP) X[$k_next] */
xor $z, %r11d /* z ^ ... */
xor $x, %r11d /* x ^ ... */
add %r11d, $dst /* dst += ... */
EOF
$code .= <<EOF if ($round3_alter);
rol \$$s, $dst /* dst <<< s */
mov $x, %r11d /* (NEXT STEP) y' = $x */
EOF
$code .= <<EOF if (!$round3_alter);
mov $x, %r11d /* (NEXT STEP) y' = $x */
rol \$$s, $dst /* dst <<< s */
EOF
$code .= <<EOF;
add $x, $dst /* dst += x */
EOF
$round3_alter^=1;
}
}
# round4_step() does:
@@ -102,15 +91,16 @@ EOF
sub round4_step
{
my ($pos, $dst, $x, $y, $z, $k_next, $T_i, $s) = @_;
$code .= " mov 0*4(%rsi), %r10d /* (NEXT STEP) X[0] */\n" if ($pos == -1);
$code .= " mov \$0xffffffff, %r11d\n" if ($pos == -1);
$code .= " xor %edx, %r11d /* (NEXT STEP) not z' = not %edx*/\n"
if ($pos == -1);
$code .= <<EOF;
lea $T_i($dst,%r10d),$dst /* Const + dst + ... */
or $x, %r11d /* x | ... */
mov $k_next*4(%rsi),%r10d /* (NEXT STEP) X[$k_next] */
xor $y, %r11d /* y ^ ... */
add %r11d, $dst /* dst += ... */
mov $k_next*4(%rsi),%r10d /* (NEXT STEP) X[$k_next] */
mov \$0xffffffff, %r11d
rol \$$s, $dst /* dst <<< s */
xor $y, %r11d /* (NEXT STEP) not z' = not $y */
@@ -130,7 +120,7 @@ $0 =~ m/(.*[\/\\])[^\/\\]+$/; my $dir=$1; my $xlate;
( $xlate="${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;
$code .= <<EOF;
@@ -189,7 +179,7 @@ round1_step( 0,'%ebx','%ecx','%edx','%eax','12','0x895cd7be','22');
round1_step( 0,'%eax','%ebx','%ecx','%edx','13','0x6b901122', '7');
round1_step( 0,'%edx','%eax','%ebx','%ecx','14','0xfd987193','12');
round1_step( 0,'%ecx','%edx','%eax','%ebx','15','0xa679438e','17');
round1_step( 1,'%ebx','%ecx','%edx','%eax', '1','0x49b40821','22');
round1_step( 1,'%ebx','%ecx','%edx','%eax', '0','0x49b40821','22');
round2_step(-1,'%eax','%ebx','%ecx','%edx', '6','0xf61e2562', '5');
round2_step( 0,'%edx','%eax','%ebx','%ecx','11','0xc040b340', '9');
@@ -206,7 +196,7 @@ round2_step( 0,'%ebx','%ecx','%edx','%eax','13','0x455a14ed','20');
round2_step( 0,'%eax','%ebx','%ecx','%edx', '2','0xa9e3e905', '5');
round2_step( 0,'%edx','%eax','%ebx','%ecx', '7','0xfcefa3f8', '9');
round2_step( 0,'%ecx','%edx','%eax','%ebx','12','0x676f02d9','14');
round2_step( 1,'%ebx','%ecx','%edx','%eax', '5','0x8d2a4c8a','20');
round2_step( 1,'%ebx','%ecx','%edx','%eax', '0','0x8d2a4c8a','20');
round3_step(-1,'%eax','%ebx','%ecx','%edx', '8','0xfffa3942', '4');
round3_step( 0,'%edx','%eax','%ebx','%ecx','11','0x8771f681','11');