1
0
Fork 0
mirror of https://github.com/ossrs/srs.git synced 2025-03-09 15:49:59 +00:00

Upgrade openssl from 1.1.0e to 1.1.1b, with source code. 4.0.78

This commit is contained in:
winlin 2021-03-01 20:47:57 +08:00
parent 8f1c992379
commit 96dbd7bced
1476 changed files with 616554 additions and 4 deletions

View file

@ -0,0 +1,318 @@
#! /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
# Normal is the
# md5_block_x86(MD5_CTX *c, ULONG *X);
# version, non-normal is the
# md5_block_x86(MD5_CTX *c, ULONG *X,int blocks);
$normal=0;
$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
push(@INC,"${dir}","${dir}../../perlasm");
require "x86asm.pl";
$output=pop;
open STDOUT,">$output";
&asm_init($ARGV[0]);
$A="eax";
$B="ebx";
$C="ecx";
$D="edx";
$tmp1="edi";
$tmp2="ebp";
$X="esi";
# What we need to load into $tmp for the next round
%Ltmp1=("R0",&Np($C), "R1",&Np($C), "R2",&Np($C), "R3",&Np($D));
@xo=(
0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, # R0
1, 6, 11, 0, 5, 10, 15, 4, 9, 14, 3, 8, 13, 2, 7, 12, # R1
5, 8, 11, 14, 1, 4, 7, 10, 13, 0, 3, 6, 9, 12, 15, 2, # R2
0, 7, 14, 5, 12, 3, 10, 1, 8, 15, 6, 13, 4, 11, 2, 9, # R3
);
&md5_block("md5_block_asm_data_order");
&asm_finish();
close STDOUT;
sub Np
{
local($p)=@_;
local(%n)=($A,$D,$B,$A,$C,$B,$D,$C);
return($n{$p});
}
sub R0
{
local($pos,$a,$b,$c,$d,$K,$ki,$s,$t)=@_;
&mov($tmp1,$C) if $pos < 0;
&mov($tmp2,&DWP($xo[$ki]*4,$K,"",0)) if $pos < 0; # very first one
# body proper
&comment("R0 $ki");
&xor($tmp1,$d); # F function - part 2
&and($tmp1,$b); # F function - part 3
&lea($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);
&rotl($a,$s);
&mov($tmp1,&Np($c)) if $pos < 1; # next tmp1 for R0
&mov($tmp1,&Np($c)) if $pos == 1; # next tmp1 for R1
&add($a,$b);
}
sub R1
{
local($pos,$a,$b,$c,$d,$K,$ki,$s,$t)=@_;
&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,$c); # G function - part 4
&mov($tmp2,&DWP($xo[$ki+1]*4,$K,"",0)) if ($pos != 2);
&add($a,$tmp1);
&mov($tmp1,&Np($c)) if $pos < 1; # G function - part 1
&mov($tmp1,&Np($c)) if $pos == 1; # G function - part 1
&rotl($a,$s);
&add($a,$b);
}
sub R2
{
local($n,$pos,$a,$b,$c,$d,$K,$ki,$s,$t)=@_;
# This one is different, only 3 logical operations
if (($n & 1) == 0)
{
&comment("R2 $ki");
# make sure to do 'D' first, not 'B', else we clash with
# the last add from the previous round.
&xor($tmp1,$d); # H function - part 2
&xor($tmp1,$b); # H function - part 3
&lea($a,&DWP($t,$a,$tmp2,1));
&add($a,$tmp1);
&mov($tmp2,&DWP($xo[$ki+1]*4,$K,"",0));
&rotl($a,$s);
&mov($tmp1,&Np($c));
}
else
{
&comment("R2 $ki");
# make sure to do 'D' first, not 'B', else we clash with
# the last add from the previous round.
&add($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);
&add($a,$tmp1);
&mov($tmp1,&Np($c)) if $pos < 1; # H function - part 1
&mov($tmp1,-1) if $pos == 1; # I function - part 1
&rotl($a,$s);
&add($a,$b);
}
}
sub R3
{
local($pos,$a,$b,$c,$d,$K,$ki,$s,$t)=@_;
&comment("R3 $ki");
# &not($tmp1)
&xor($tmp1,$d) if $pos < 0; # I function - part 2
&or($tmp1,$b); # I function - part 3
&lea($a,&DWP($t,$a,$tmp2,1));
&xor($tmp1,$c); # I function - part 4
&mov($tmp2,&DWP($xo[$ki+1]*4,$K,"",0)) if $pos != 2; # load X/k value
&mov($tmp2,&wparam(0)) if $pos == 2;
&add($a,$tmp1);
&mov($tmp1,-1) if $pos < 1; # H function - part 1
&add($K,64) if $pos >=1 && !$normal;
&rotl($a,$s);
&xor($tmp1,&Np($d)) if $pos <= 0; # I function - part = first time
&mov($tmp1,&DWP( 0,$tmp2,"",0)) if $pos > 0;
&add($a,$b);
}
sub md5_block
{
local($name)=@_;
&function_begin_B($name,"",3);
# parameter 1 is the MD5_CTX structure.
# A 0
# B 4
# C 8
# D 12
&push("esi");
&push("edi");
&mov($tmp1, &wparam(0)); # edi
&mov($X, &wparam(1)); # esi
&mov($C, &wparam(2));
&push("ebp");
&shl($C, 6);
&push("ebx");
&add($C, $X); # offset we end at
&sub($C, 64);
&mov($A, &DWP( 0,$tmp1,"",0));
&push($C); # Put on the TOS
&mov($B, &DWP( 4,$tmp1,"",0));
&mov($C, &DWP( 8,$tmp1,"",0));
&mov($D, &DWP(12,$tmp1,"",0));
&set_label("start") unless $normal;
&comment("");
&comment("R0 section");
&R0(-2,$A,$B,$C,$D,$X, 0, 7,0xd76aa478);
&R0( 0,$D,$A,$B,$C,$X, 1,12,0xe8c7b756);
&R0( 0,$C,$D,$A,$B,$X, 2,17,0x242070db);
&R0( 0,$B,$C,$D,$A,$X, 3,22,0xc1bdceee);
&R0( 0,$A,$B,$C,$D,$X, 4, 7,0xf57c0faf);
&R0( 0,$D,$A,$B,$C,$X, 5,12,0x4787c62a);
&R0( 0,$C,$D,$A,$B,$X, 6,17,0xa8304613);
&R0( 0,$B,$C,$D,$A,$X, 7,22,0xfd469501);
&R0( 0,$A,$B,$C,$D,$X, 8, 7,0x698098d8);
&R0( 0,$D,$A,$B,$C,$X, 9,12,0x8b44f7af);
&R0( 0,$C,$D,$A,$B,$X,10,17,0xffff5bb1);
&R0( 0,$B,$C,$D,$A,$X,11,22,0x895cd7be);
&R0( 0,$A,$B,$C,$D,$X,12, 7,0x6b901122);
&R0( 0,$D,$A,$B,$C,$X,13,12,0xfd987193);
&R0( 0,$C,$D,$A,$B,$X,14,17,0xa679438e);
&R0( 1,$B,$C,$D,$A,$X,15,22,0x49b40821);
&comment("");
&comment("R1 section");
&R1(-1,$A,$B,$C,$D,$X,16, 5,0xf61e2562);
&R1( 0,$D,$A,$B,$C,$X,17, 9,0xc040b340);
&R1( 0,$C,$D,$A,$B,$X,18,14,0x265e5a51);
&R1( 0,$B,$C,$D,$A,$X,19,20,0xe9b6c7aa);
&R1( 0,$A,$B,$C,$D,$X,20, 5,0xd62f105d);
&R1( 0,$D,$A,$B,$C,$X,21, 9,0x02441453);
&R1( 0,$C,$D,$A,$B,$X,22,14,0xd8a1e681);
&R1( 0,$B,$C,$D,$A,$X,23,20,0xe7d3fbc8);
&R1( 0,$A,$B,$C,$D,$X,24, 5,0x21e1cde6);
&R1( 0,$D,$A,$B,$C,$X,25, 9,0xc33707d6);
&R1( 0,$C,$D,$A,$B,$X,26,14,0xf4d50d87);
&R1( 0,$B,$C,$D,$A,$X,27,20,0x455a14ed);
&R1( 0,$A,$B,$C,$D,$X,28, 5,0xa9e3e905);
&R1( 0,$D,$A,$B,$C,$X,29, 9,0xfcefa3f8);
&R1( 0,$C,$D,$A,$B,$X,30,14,0x676f02d9);
&R1( 1,$B,$C,$D,$A,$X,31,20,0x8d2a4c8a);
&comment("");
&comment("R2 section");
&R2( 0,-1,$A,$B,$C,$D,$X,32, 4,0xfffa3942);
&R2( 1, 0,$D,$A,$B,$C,$X,33,11,0x8771f681);
&R2( 2, 0,$C,$D,$A,$B,$X,34,16,0x6d9d6122);
&R2( 3, 0,$B,$C,$D,$A,$X,35,23,0xfde5380c);
&R2( 4, 0,$A,$B,$C,$D,$X,36, 4,0xa4beea44);
&R2( 5, 0,$D,$A,$B,$C,$X,37,11,0x4bdecfa9);
&R2( 6, 0,$C,$D,$A,$B,$X,38,16,0xf6bb4b60);
&R2( 7, 0,$B,$C,$D,$A,$X,39,23,0xbebfbc70);
&R2( 8, 0,$A,$B,$C,$D,$X,40, 4,0x289b7ec6);
&R2( 9, 0,$D,$A,$B,$C,$X,41,11,0xeaa127fa);
&R2(10, 0,$C,$D,$A,$B,$X,42,16,0xd4ef3085);
&R2(11, 0,$B,$C,$D,$A,$X,43,23,0x04881d05);
&R2(12, 0,$A,$B,$C,$D,$X,44, 4,0xd9d4d039);
&R2(13, 0,$D,$A,$B,$C,$X,45,11,0xe6db99e5);
&R2(14, 0,$C,$D,$A,$B,$X,46,16,0x1fa27cf8);
&R2(15, 1,$B,$C,$D,$A,$X,47,23,0xc4ac5665);
&comment("");
&comment("R3 section");
&R3(-1,$A,$B,$C,$D,$X,48, 6,0xf4292244);
&R3( 0,$D,$A,$B,$C,$X,49,10,0x432aff97);
&R3( 0,$C,$D,$A,$B,$X,50,15,0xab9423a7);
&R3( 0,$B,$C,$D,$A,$X,51,21,0xfc93a039);
&R3( 0,$A,$B,$C,$D,$X,52, 6,0x655b59c3);
&R3( 0,$D,$A,$B,$C,$X,53,10,0x8f0ccc92);
&R3( 0,$C,$D,$A,$B,$X,54,15,0xffeff47d);
&R3( 0,$B,$C,$D,$A,$X,55,21,0x85845dd1);
&R3( 0,$A,$B,$C,$D,$X,56, 6,0x6fa87e4f);
&R3( 0,$D,$A,$B,$C,$X,57,10,0xfe2ce6e0);
&R3( 0,$C,$D,$A,$B,$X,58,15,0xa3014314);
&R3( 0,$B,$C,$D,$A,$X,59,21,0x4e0811a1);
&R3( 0,$A,$B,$C,$D,$X,60, 6,0xf7537e82);
&R3( 0,$D,$A,$B,$C,$X,61,10,0xbd3af235);
&R3( 0,$C,$D,$A,$B,$X,62,15,0x2ad7d2bb);
&R3( 2,$B,$C,$D,$A,$X,63,21,0xeb86d391);
# &mov($tmp2,&wparam(0)); # done in the last R3
# &mov($tmp1, &DWP( 0,$tmp2,"",0)); # done is the last R3
&add($A,$tmp1);
&mov($tmp1, &DWP( 4,$tmp2,"",0));
&add($B,$tmp1);
&mov($tmp1, &DWP( 8,$tmp2,"",0));
&add($C,$tmp1);
&mov($tmp1, &DWP(12,$tmp2,"",0));
&add($D,$tmp1);
&mov(&DWP( 0,$tmp2,"",0),$A);
&mov(&DWP( 4,$tmp2,"",0),$B);
&mov($tmp1,&swtmp(0)) unless $normal;
&mov(&DWP( 8,$tmp2,"",0),$C);
&mov(&DWP(12,$tmp2,"",0),$D);
&cmp($tmp1,$X) unless $normal; # check count
&jae(&label("start")) unless $normal;
&pop("eax"); # pop the temp variable off the stack
&pop("ebx");
&pop("ebp");
&pop("edi");
&pop("esi");
&ret();
&function_end_B($name);
}

View file

@ -0,0 +1,437 @@
#! /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
# ====================================================================
# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
# project. The module is, however, dual licensed under OpenSSL and
# CRYPTOGAMS licenses depending on where you obtain it. For further
# details see http://www.openssl.org/~appro/cryptogams/.
#
# Hardware SPARC T4 support by David S. Miller.
# ====================================================================
# MD5 for SPARCv9, 6.9 cycles per byte on UltraSPARC, >40% faster than
# code generated by Sun C 5.2.
# SPARC T4 MD5 hardware achieves 3.20 cycles per byte, which is 2.1x
# faster than software. Multi-process benchmark saturates at 12x
# single-process result on 8-core processor, or ~11GBps per 2.85GHz
# socket.
$output=pop;
open STDOUT,">$output";
use integer;
($ctx,$inp,$len)=("%i0","%i1","%i2"); # input arguments
# 64-bit values
@X=("%o0","%o1","%o2","%o3","%o4","%o5","%o7","%g1","%g2");
$tx="%g3";
($AB,$CD)=("%g4","%g5");
# 32-bit values
@V=($A,$B,$C,$D)=map("%l$_",(0..3));
($t1,$t2,$t3,$saved_asi)=map("%l$_",(4..7));
($shr,$shl1,$shl2)=("%i3","%i4","%i5");
my @K=( 0xd76aa478,0xe8c7b756,0x242070db,0xc1bdceee,
0xf57c0faf,0x4787c62a,0xa8304613,0xfd469501,
0x698098d8,0x8b44f7af,0xffff5bb1,0x895cd7be,
0x6b901122,0xfd987193,0xa679438e,0x49b40821,
0xf61e2562,0xc040b340,0x265e5a51,0xe9b6c7aa,
0xd62f105d,0x02441453,0xd8a1e681,0xe7d3fbc8,
0x21e1cde6,0xc33707d6,0xf4d50d87,0x455a14ed,
0xa9e3e905,0xfcefa3f8,0x676f02d9,0x8d2a4c8a,
0xfffa3942,0x8771f681,0x6d9d6122,0xfde5380c,
0xa4beea44,0x4bdecfa9,0xf6bb4b60,0xbebfbc70,
0x289b7ec6,0xeaa127fa,0xd4ef3085,0x04881d05,
0xd9d4d039,0xe6db99e5,0x1fa27cf8,0xc4ac5665,
0xf4292244,0x432aff97,0xab9423a7,0xfc93a039,
0x655b59c3,0x8f0ccc92,0xffeff47d,0x85845dd1,
0x6fa87e4f,0xfe2ce6e0,0xa3014314,0x4e0811a1,
0xf7537e82,0xbd3af235,0x2ad7d2bb,0xeb86d391, 0 );
sub R0 {
my ($i,$a,$b,$c,$d) = @_;
my $rot = (7,12,17,22)[$i%4];
my $j = ($i+1)/2;
if ($i&1) {
$code.=<<___;
srlx @X[$j],$shr,@X[$j] ! align X[`$i+1`]
and $b,$t1,$t1 ! round $i
sllx @X[$j+1],$shl1,$tx
add $t2,$a,$a
sllx $tx,$shl2,$tx
xor $d,$t1,$t1
or $tx,@X[$j],@X[$j]
sethi %hi(@K[$i+1]),$t2
add $t1,$a,$a
or $t2,%lo(@K[$i+1]),$t2
sll $a,$rot,$t3
add @X[$j],$t2,$t2 ! X[`$i+1`]+K[`$i+1`]
srl $a,32-$rot,$a
add $b,$t3,$t3
xor $b,$c,$t1
add $t3,$a,$a
___
} else {
$code.=<<___;
srlx @X[$j],32,$tx ! extract X[`2*$j+1`]
and $b,$t1,$t1 ! round $i
add $t2,$a,$a
xor $d,$t1,$t1
sethi %hi(@K[$i+1]),$t2
add $t1,$a,$a
or $t2,%lo(@K[$i+1]),$t2
sll $a,$rot,$t3
add $tx,$t2,$t2 ! X[`2*$j+1`]+K[`$i+1`]
srl $a,32-$rot,$a
add $b,$t3,$t3
xor $b,$c,$t1
add $t3,$a,$a
___
}
}
sub R0_1 {
my ($i,$a,$b,$c,$d) = @_;
my $rot = (7,12,17,22)[$i%4];
$code.=<<___;
srlx @X[0],32,$tx ! extract X[1]
and $b,$t1,$t1 ! round $i
add $t2,$a,$a
xor $d,$t1,$t1
sethi %hi(@K[$i+1]),$t2
add $t1,$a,$a
or $t2,%lo(@K[$i+1]),$t2
sll $a,$rot,$t3
add $tx,$t2,$t2 ! X[1]+K[`$i+1`]
srl $a,32-$rot,$a
add $b,$t3,$t3
andn $b,$c,$t1
add $t3,$a,$a
___
}
sub R1 {
my ($i,$a,$b,$c,$d) = @_;
my $rot = (5,9,14,20)[$i%4];
my $j = $i<31 ? (1+5*($i+1))%16 : (5+3*($i+1))%16;
my $xi = @X[$j/2];
$code.=<<___ if ($j&1 && ($xi=$tx));
srlx @X[$j/2],32,$xi ! extract X[$j]
___
$code.=<<___;
and $b,$d,$t3 ! round $i
add $t2,$a,$a
or $t3,$t1,$t1
sethi %hi(@K[$i+1]),$t2
add $t1,$a,$a
or $t2,%lo(@K[$i+1]),$t2
sll $a,$rot,$t3
add $xi,$t2,$t2 ! X[$j]+K[`$i+1`]
srl $a,32-$rot,$a
add $b,$t3,$t3
`$i<31?"andn":"xor"` $b,$c,$t1
add $t3,$a,$a
___
}
sub R2 {
my ($i,$a,$b,$c,$d) = @_;
my $rot = (4,11,16,23)[$i%4];
my $j = $i<47 ? (5+3*($i+1))%16 : (0+7*($i+1))%16;
my $xi = @X[$j/2];
$code.=<<___ if ($j&1 && ($xi=$tx));
srlx @X[$j/2],32,$xi ! extract X[$j]
___
$code.=<<___;
add $t2,$a,$a ! round $i
xor $b,$t1,$t1
sethi %hi(@K[$i+1]),$t2
add $t1,$a,$a
or $t2,%lo(@K[$i+1]),$t2
sll $a,$rot,$t3
add $xi,$t2,$t2 ! X[$j]+K[`$i+1`]
srl $a,32-$rot,$a
add $b,$t3,$t3
xor $b,$c,$t1
add $t3,$a,$a
___
}
sub R3 {
my ($i,$a,$b,$c,$d) = @_;
my $rot = (6,10,15,21)[$i%4];
my $j = (0+7*($i+1))%16;
my $xi = @X[$j/2];
$code.=<<___;
add $t2,$a,$a ! round $i
___
$code.=<<___ if ($j&1 && ($xi=$tx));
srlx @X[$j/2],32,$xi ! extract X[$j]
___
$code.=<<___;
orn $b,$d,$t1
sethi %hi(@K[$i+1]),$t2
xor $c,$t1,$t1
or $t2,%lo(@K[$i+1]),$t2
add $t1,$a,$a
sll $a,$rot,$t3
add $xi,$t2,$t2 ! X[$j]+K[`$i+1`]
srl $a,32-$rot,$a
add $b,$t3,$t3
add $t3,$a,$a
___
}
$code.=<<___;
#include "sparc_arch.h"
#ifdef __arch64__
.register %g2,#scratch
.register %g3,#scratch
#endif
.section ".text",#alloc,#execinstr
#ifdef __PIC__
SPARC_PIC_THUNK(%g1)
#endif
.globl md5_block_asm_data_order
.align 32
md5_block_asm_data_order:
SPARC_LOAD_ADDRESS_LEAF(OPENSSL_sparcv9cap_P,%g1,%g5)
ld [%g1+4],%g1 ! OPENSSL_sparcv9cap_P[1]
andcc %g1, CFR_MD5, %g0
be .Lsoftware
nop
mov 4, %g1
andcc %o1, 0x7, %g0
lda [%o0 + %g0]0x88, %f0 ! load context
lda [%o0 + %g1]0x88, %f1
add %o0, 8, %o0
lda [%o0 + %g0]0x88, %f2
lda [%o0 + %g1]0x88, %f3
bne,pn %icc, .Lhwunaligned
sub %o0, 8, %o0
.Lhw_loop:
ldd [%o1 + 0x00], %f8
ldd [%o1 + 0x08], %f10
ldd [%o1 + 0x10], %f12
ldd [%o1 + 0x18], %f14
ldd [%o1 + 0x20], %f16
ldd [%o1 + 0x28], %f18
ldd [%o1 + 0x30], %f20
subcc %o2, 1, %o2 ! done yet?
ldd [%o1 + 0x38], %f22
add %o1, 0x40, %o1
prefetch [%o1 + 63], 20
.word 0x81b02800 ! MD5
bne,pt SIZE_T_CC, .Lhw_loop
nop
.Lhwfinish:
sta %f0, [%o0 + %g0]0x88 ! store context
sta %f1, [%o0 + %g1]0x88
add %o0, 8, %o0
sta %f2, [%o0 + %g0]0x88
sta %f3, [%o0 + %g1]0x88
retl
nop
.align 8
.Lhwunaligned:
alignaddr %o1, %g0, %o1
ldd [%o1 + 0x00], %f10
.Lhwunaligned_loop:
ldd [%o1 + 0x08], %f12
ldd [%o1 + 0x10], %f14
ldd [%o1 + 0x18], %f16
ldd [%o1 + 0x20], %f18
ldd [%o1 + 0x28], %f20
ldd [%o1 + 0x30], %f22
ldd [%o1 + 0x38], %f24
subcc %o2, 1, %o2 ! done yet?
ldd [%o1 + 0x40], %f26
add %o1, 0x40, %o1
prefetch [%o1 + 63], 20
faligndata %f10, %f12, %f8
faligndata %f12, %f14, %f10
faligndata %f14, %f16, %f12
faligndata %f16, %f18, %f14
faligndata %f18, %f20, %f16
faligndata %f20, %f22, %f18
faligndata %f22, %f24, %f20
faligndata %f24, %f26, %f22
.word 0x81b02800 ! MD5
bne,pt SIZE_T_CC, .Lhwunaligned_loop
for %f26, %f26, %f10 ! %f10=%f26
ba .Lhwfinish
nop
.align 16
.Lsoftware:
save %sp,-STACK_FRAME,%sp
rd %asi,$saved_asi
wr %g0,0x88,%asi ! ASI_PRIMARY_LITTLE
and $inp,7,$shr
andn $inp,7,$inp
sll $shr,3,$shr ! *=8
mov 56,$shl2
ld [$ctx+0],$A
sub $shl2,$shr,$shl2
ld [$ctx+4],$B
and $shl2,32,$shl1
add $shl2,8,$shl2
ld [$ctx+8],$C
sub $shl2,$shl1,$shl2 ! shr+shl1+shl2==64
ld [$ctx+12],$D
nop
.Loop:
cmp $shr,0 ! was inp aligned?
ldxa [$inp+0]%asi,@X[0] ! load little-endian input
ldxa [$inp+8]%asi,@X[1]
ldxa [$inp+16]%asi,@X[2]
ldxa [$inp+24]%asi,@X[3]
ldxa [$inp+32]%asi,@X[4]
sllx $A,32,$AB ! pack A,B
ldxa [$inp+40]%asi,@X[5]
sllx $C,32,$CD ! pack C,D
ldxa [$inp+48]%asi,@X[6]
or $B,$AB,$AB
ldxa [$inp+56]%asi,@X[7]
or $D,$CD,$CD
bnz,a,pn %icc,.+8
ldxa [$inp+64]%asi,@X[8]
srlx @X[0],$shr,@X[0] ! align X[0]
sllx @X[1],$shl1,$tx
sethi %hi(@K[0]),$t2
sllx $tx,$shl2,$tx
or $t2,%lo(@K[0]),$t2
or $tx,@X[0],@X[0]
xor $C,$D,$t1
add @X[0],$t2,$t2 ! X[0]+K[0]
___
for ($i=0;$i<15;$i++) { &R0($i,@V); unshift(@V,pop(@V)); }
for (;$i<16;$i++) { &R0_1($i,@V); unshift(@V,pop(@V)); }
for (;$i<32;$i++) { &R1($i,@V); unshift(@V,pop(@V)); }
for (;$i<48;$i++) { &R2($i,@V); unshift(@V,pop(@V)); }
for (;$i<64;$i++) { &R3($i,@V); unshift(@V,pop(@V)); }
$code.=<<___;
srlx $AB,32,$t1 ! unpack A,B,C,D and accumulate
add $inp,64,$inp ! advance inp
srlx $CD,32,$t2
add $t1,$A,$A
subcc $len,1,$len ! done yet?
add $AB,$B,$B
add $t2,$C,$C
add $CD,$D,$D
srl $B,0,$B ! clruw $B
bne SIZE_T_CC,.Loop
srl $D,0,$D ! clruw $D
st $A,[$ctx+0] ! write out ctx
st $B,[$ctx+4]
st $C,[$ctx+8]
st $D,[$ctx+12]
wr %g0,$saved_asi,%asi
ret
restore
.type md5_block_asm_data_order,#function
.size md5_block_asm_data_order,(.-md5_block_asm_data_order)
.asciz "MD5 block transform for SPARCv9, CRYPTOGAMS by <appro\@openssl.org>"
.align 4
___
# Purpose of these subroutines is to explicitly encode VIS instructions,
# so that one can compile the module without having to specify VIS
# extensions on compiler command line, e.g. -xarch=v9 vs. -xarch=v9a.
# Idea is to reserve for option to produce "universal" binary and let
# programmer detect if current CPU is VIS capable at run-time.
sub unvis {
my ($mnemonic,$rs1,$rs2,$rd)=@_;
my $ref,$opf;
my %visopf = ( "faligndata" => 0x048,
"for" => 0x07c );
$ref = "$mnemonic\t$rs1,$rs2,$rd";
if ($opf=$visopf{$mnemonic}) {
foreach ($rs1,$rs2,$rd) {
return $ref if (!/%f([0-9]{1,2})/);
$_=$1;
if ($1>=32) {
return $ref if ($1&1);
# re-encode for upper double register addressing
$_=($1|$1>>5)&31;
}
}
return sprintf ".word\t0x%08x !%s",
0x81b00000|$rd<<25|$rs1<<14|$opf<<5|$rs2,
$ref;
} else {
return $ref;
}
}
sub unalignaddr {
my ($mnemonic,$rs1,$rs2,$rd)=@_;
my %bias = ( "g" => 0, "o" => 8, "l" => 16, "i" => 24 );
my $ref="$mnemonic\t$rs1,$rs2,$rd";
foreach ($rs1,$rs2,$rd) {
if (/%([goli])([0-7])/) { $_=$bias{$1}+$2; }
else { return $ref; }
}
return sprintf ".word\t0x%08x !%s",
0x81b00300|$rd<<25|$rs1<<14|$rs2,
$ref;
}
foreach (split("\n",$code)) {
s/\`([^\`]*)\`/eval $1/ge;
s/\b(f[^\s]*)\s+(%f[0-9]{1,2}),\s*(%f[0-9]{1,2}),\s*(%f[0-9]{1,2})/
&unvis($1,$2,$3,$4)
/ge;
s/\b(alignaddr)\s+(%[goli][0-7]),\s*(%[goli][0-7]),\s*(%[goli][0-7])/
&unalignaddr($1,$2,$3,$4)
/ge;
print $_,"\n";
}
close STDOUT;

View file

@ -0,0 +1,393 @@
#! /usr/bin/env perl
# Author: Marc Bevand <bevand_m (at) epita.fr>
# 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
# MD5 optimized for AMD64.
use strict;
my $code;
# round1_step() does:
# dst = x + ((dst + F(x,y,z) + X[k] + T_i) <<< s)
# %r10d = X[k_next]
# %r11d = z' (copy of z for the next step)
# Each round1_step() takes about 5.3 clocks (9 instructions, 1.7 IPC)
sub round1_step
{
my ($pos, $dst, $x, $y, $z, $k_next, $T_i, $s) = @_;
$code .= " mov 0*4(%rsi), %r10d /* (NEXT STEP) X[0] */\n" if ($pos == -1);
$code .= " mov %edx, %r11d /* (NEXT STEP) z' = %edx */\n" if ($pos == -1);
$code .= <<EOF;
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 ^ ... */
add %r11d, $dst /* dst += ... */
rol \$$s, $dst /* dst <<< s */
mov $y, %r11d /* (NEXT STEP) z' = $y */
add $x, $dst /* dst += x */
EOF
}
# round2_step() does:
# dst = x + ((dst + G(x,y,z) + X[k] + T_i) <<< s)
# %r10d = X[k_next]
# %r11d = z' (copy of z for the next step)
# %r12d = z' (copy of z for the next step)
# Each round2_step() takes about 5.4 clocks (11 instructions, 2.0 IPC)
sub round2_step
{
my ($pos, $dst, $x, $y, $z, $k_next, $T_i, $s) = @_;
$code .= " mov %edx, %r11d /* (NEXT STEP) z' = %edx */\n" if ($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 $y, %r11d /* y & (not z) */
mov $k_next*4(%rsi),%r10d /* (NEXT STEP) X[$k_next] */
or %r11d, %r12d /* (y & (not z)) | (x & z) */
mov $y, %r11d /* (NEXT STEP) z' = $y */
add %r12d, $dst /* dst += ... */
mov $y, %r12d /* (NEXT STEP) z' = $y */
rol \$$s, $dst /* dst <<< s */
add $x, $dst /* dst += x */
EOF
}
# round3_step() does:
# dst = x + ((dst + H(x,y,z) + X[k] + T_i) <<< s)
# %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) = @_;
$code .= " mov %ecx, %r11d /* (NEXT STEP) y' = %ecx */\n" if ($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 $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
$round3_alter^=1;
}
}
# round4_step() does:
# dst = x + ((dst + I(x,y,z) + X[k] + T_i) <<< s)
# %r10d = X[k_next]
# %r11d = not z' (copy of not z for the next step)
# Each round4_step() takes about 5.2 clocks (9 instructions, 1.7 IPC)
sub round4_step
{
my ($pos, $dst, $x, $y, $z, $k_next, $T_i, $s) = @_;
$code .= " mov \$0xffffffff, %r11d\n" if ($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, $dst /* dst += ... */
mov \$0xffffffff, %r11d
rol \$$s, $dst /* dst <<< s */
xor $y, %r11d /* (NEXT STEP) not z' = not $y */
add $x, $dst /* dst += x */
EOF
}
no warnings qw(uninitialized);
my $flavour = shift;
my $output = shift;
if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
my $win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
$0 =~ m/(.*[\/\\])[^\/\\]+$/; my $dir=$1; my $xlate;
( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
die "can't locate x86_64-xlate.pl";
open OUT,"| \"$^X\" \"$xlate\" $flavour \"$output\"";
*STDOUT=*OUT;
$code .= <<EOF;
.text
.align 16
.globl md5_block_asm_data_order
.type md5_block_asm_data_order,\@function,3
md5_block_asm_data_order:
.cfi_startproc
push %rbp
.cfi_push %rbp
push %rbx
.cfi_push %rbx
push %r12
.cfi_push %r12
push %r14
.cfi_push %r14
push %r15
.cfi_push %r15
.Lprologue:
# rdi = arg #1 (ctx, MD5_CTX pointer)
# rsi = arg #2 (ptr, data pointer)
# rdx = arg #3 (nbr, number of 16-word blocks to process)
mov %rdi, %rbp # rbp = ctx
shl \$6, %rdx # rdx = nbr in bytes
lea (%rsi,%rdx), %rdi # rdi = end
mov 0*4(%rbp), %eax # eax = ctx->A
mov 1*4(%rbp), %ebx # ebx = ctx->B
mov 2*4(%rbp), %ecx # ecx = ctx->C
mov 3*4(%rbp), %edx # edx = ctx->D
# end is 'rdi'
# ptr is 'rsi'
# A is 'eax'
# B is 'ebx'
# C is 'ecx'
# D is 'edx'
cmp %rdi, %rsi # cmp end with ptr
je .Lend # jmp if ptr == end
# BEGIN of loop over 16-word blocks
.Lloop: # save old values of A, B, C, D
mov %eax, %r8d
mov %ebx, %r9d
mov %ecx, %r14d
mov %edx, %r15d
EOF
round1_step(-1,'%eax','%ebx','%ecx','%edx', '1','0xd76aa478', '7');
round1_step( 0,'%edx','%eax','%ebx','%ecx', '2','0xe8c7b756','12');
round1_step( 0,'%ecx','%edx','%eax','%ebx', '3','0x242070db','17');
round1_step( 0,'%ebx','%ecx','%edx','%eax', '4','0xc1bdceee','22');
round1_step( 0,'%eax','%ebx','%ecx','%edx', '5','0xf57c0faf', '7');
round1_step( 0,'%edx','%eax','%ebx','%ecx', '6','0x4787c62a','12');
round1_step( 0,'%ecx','%edx','%eax','%ebx', '7','0xa8304613','17');
round1_step( 0,'%ebx','%ecx','%edx','%eax', '8','0xfd469501','22');
round1_step( 0,'%eax','%ebx','%ecx','%edx', '9','0x698098d8', '7');
round1_step( 0,'%edx','%eax','%ebx','%ecx','10','0x8b44f7af','12');
round1_step( 0,'%ecx','%edx','%eax','%ebx','11','0xffff5bb1','17');
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');
round2_step(-1,'%eax','%ebx','%ecx','%edx', '6','0xf61e2562', '5');
round2_step( 0,'%edx','%eax','%ebx','%ecx','11','0xc040b340', '9');
round2_step( 0,'%ecx','%edx','%eax','%ebx', '0','0x265e5a51','14');
round2_step( 0,'%ebx','%ecx','%edx','%eax', '5','0xe9b6c7aa','20');
round2_step( 0,'%eax','%ebx','%ecx','%edx','10','0xd62f105d', '5');
round2_step( 0,'%edx','%eax','%ebx','%ecx','15', '0x2441453', '9');
round2_step( 0,'%ecx','%edx','%eax','%ebx', '4','0xd8a1e681','14');
round2_step( 0,'%ebx','%ecx','%edx','%eax', '9','0xe7d3fbc8','20');
round2_step( 0,'%eax','%ebx','%ecx','%edx','14','0x21e1cde6', '5');
round2_step( 0,'%edx','%eax','%ebx','%ecx', '3','0xc33707d6', '9');
round2_step( 0,'%ecx','%edx','%eax','%ebx', '8','0xf4d50d87','14');
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');
round3_step(-1,'%eax','%ebx','%ecx','%edx', '8','0xfffa3942', '4');
round3_step( 0,'%edx','%eax','%ebx','%ecx','11','0x8771f681','11');
round3_step( 0,'%ecx','%edx','%eax','%ebx','14','0x6d9d6122','16');
round3_step( 0,'%ebx','%ecx','%edx','%eax', '1','0xfde5380c','23');
round3_step( 0,'%eax','%ebx','%ecx','%edx', '4','0xa4beea44', '4');
round3_step( 0,'%edx','%eax','%ebx','%ecx', '7','0x4bdecfa9','11');
round3_step( 0,'%ecx','%edx','%eax','%ebx','10','0xf6bb4b60','16');
round3_step( 0,'%ebx','%ecx','%edx','%eax','13','0xbebfbc70','23');
round3_step( 0,'%eax','%ebx','%ecx','%edx', '0','0x289b7ec6', '4');
round3_step( 0,'%edx','%eax','%ebx','%ecx', '3','0xeaa127fa','11');
round3_step( 0,'%ecx','%edx','%eax','%ebx', '6','0xd4ef3085','16');
round3_step( 0,'%ebx','%ecx','%edx','%eax', '9', '0x4881d05','23');
round3_step( 0,'%eax','%ebx','%ecx','%edx','12','0xd9d4d039', '4');
round3_step( 0,'%edx','%eax','%ebx','%ecx','15','0xe6db99e5','11');
round3_step( 0,'%ecx','%edx','%eax','%ebx', '2','0x1fa27cf8','16');
round3_step( 1,'%ebx','%ecx','%edx','%eax', '0','0xc4ac5665','23');
round4_step(-1,'%eax','%ebx','%ecx','%edx', '7','0xf4292244', '6');
round4_step( 0,'%edx','%eax','%ebx','%ecx','14','0x432aff97','10');
round4_step( 0,'%ecx','%edx','%eax','%ebx', '5','0xab9423a7','15');
round4_step( 0,'%ebx','%ecx','%edx','%eax','12','0xfc93a039','21');
round4_step( 0,'%eax','%ebx','%ecx','%edx', '3','0x655b59c3', '6');
round4_step( 0,'%edx','%eax','%ebx','%ecx','10','0x8f0ccc92','10');
round4_step( 0,'%ecx','%edx','%eax','%ebx', '1','0xffeff47d','15');
round4_step( 0,'%ebx','%ecx','%edx','%eax', '8','0x85845dd1','21');
round4_step( 0,'%eax','%ebx','%ecx','%edx','15','0x6fa87e4f', '6');
round4_step( 0,'%edx','%eax','%ebx','%ecx', '6','0xfe2ce6e0','10');
round4_step( 0,'%ecx','%edx','%eax','%ebx','13','0xa3014314','15');
round4_step( 0,'%ebx','%ecx','%edx','%eax', '4','0x4e0811a1','21');
round4_step( 0,'%eax','%ebx','%ecx','%edx','11','0xf7537e82', '6');
round4_step( 0,'%edx','%eax','%ebx','%ecx', '2','0xbd3af235','10');
round4_step( 0,'%ecx','%edx','%eax','%ebx', '9','0x2ad7d2bb','15');
round4_step( 1,'%ebx','%ecx','%edx','%eax', '0','0xeb86d391','21');
$code .= <<EOF;
# add old values of A, B, C, D
add %r8d, %eax
add %r9d, %ebx
add %r14d, %ecx
add %r15d, %edx
# loop control
add \$64, %rsi # ptr += 64
cmp %rdi, %rsi # cmp end with ptr
jb .Lloop # jmp if ptr < end
# END of loop over 16-word blocks
.Lend:
mov %eax, 0*4(%rbp) # ctx->A = A
mov %ebx, 1*4(%rbp) # ctx->B = B
mov %ecx, 2*4(%rbp) # ctx->C = C
mov %edx, 3*4(%rbp) # ctx->D = D
mov (%rsp),%r15
.cfi_restore %r15
mov 8(%rsp),%r14
.cfi_restore %r14
mov 16(%rsp),%r12
.cfi_restore %r12
mov 24(%rsp),%rbx
.cfi_restore %rbx
mov 32(%rsp),%rbp
.cfi_restore %rbp
add \$40,%rsp
.cfi_adjust_cfa_offset -40
.Lepilogue:
ret
.cfi_endproc
.size md5_block_asm_data_order,.-md5_block_asm_data_order
EOF
# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
# CONTEXT *context,DISPATCHER_CONTEXT *disp)
if ($win64) {
my $rec="%rcx";
my $frame="%rdx";
my $context="%r8";
my $disp="%r9";
$code.=<<___;
.extern __imp_RtlVirtualUnwind
.type se_handler,\@abi-omnipotent
.align 16
se_handler:
push %rsi
push %rdi
push %rbx
push %rbp
push %r12
push %r13
push %r14
push %r15
pushfq
sub \$64,%rsp
mov 120($context),%rax # pull context->Rax
mov 248($context),%rbx # pull context->Rip
lea .Lprologue(%rip),%r10
cmp %r10,%rbx # context->Rip<.Lprologue
jb .Lin_prologue
mov 152($context),%rax # pull context->Rsp
lea .Lepilogue(%rip),%r10
cmp %r10,%rbx # context->Rip>=.Lepilogue
jae .Lin_prologue
lea 40(%rax),%rax
mov -8(%rax),%rbp
mov -16(%rax),%rbx
mov -24(%rax),%r12
mov -32(%rax),%r14
mov -40(%rax),%r15
mov %rbx,144($context) # restore context->Rbx
mov %rbp,160($context) # restore context->Rbp
mov %r12,216($context) # restore context->R12
mov %r14,232($context) # restore context->R14
mov %r15,240($context) # restore context->R15
.Lin_prologue:
mov 8(%rax),%rdi
mov 16(%rax),%rsi
mov %rax,152($context) # restore context->Rsp
mov %rsi,168($context) # restore context->Rsi
mov %rdi,176($context) # restore context->Rdi
mov 40($disp),%rdi # disp->ContextRecord
mov $context,%rsi # context
mov \$154,%ecx # sizeof(CONTEXT)
.long 0xa548f3fc # cld; rep movsq
mov $disp,%rsi
xor %rcx,%rcx # arg1, UNW_FLAG_NHANDLER
mov 8(%rsi),%rdx # arg2, disp->ImageBase
mov 0(%rsi),%r8 # arg3, disp->ControlPc
mov 16(%rsi),%r9 # arg4, disp->FunctionEntry
mov 40(%rsi),%r10 # disp->ContextRecord
lea 56(%rsi),%r11 # &disp->HandlerData
lea 24(%rsi),%r12 # &disp->EstablisherFrame
mov %r10,32(%rsp) # arg5
mov %r11,40(%rsp) # arg6
mov %r12,48(%rsp) # arg7
mov %rcx,56(%rsp) # arg8, (NULL)
call *__imp_RtlVirtualUnwind(%rip)
mov \$1,%eax # ExceptionContinueSearch
add \$64,%rsp
popfq
pop %r15
pop %r14
pop %r13
pop %r12
pop %rbp
pop %rbx
pop %rdi
pop %rsi
ret
.size se_handler,.-se_handler
.section .pdata
.align 4
.rva .LSEH_begin_md5_block_asm_data_order
.rva .LSEH_end_md5_block_asm_data_order
.rva .LSEH_info_md5_block_asm_data_order
.section .xdata
.align 8
.LSEH_info_md5_block_asm_data_order:
.byte 9,0,0,0
.rva se_handler
___
}
print $code;
close STDOUT;

View file

@ -0,0 +1,11 @@
LIBS=../../libcrypto
SOURCE[../../libcrypto]=\
md5_dgst.c md5_one.c {- $target{md5_asm_src} -}
GENERATE[md5-586.s]=asm/md5-586.pl \
$(PERLASM_SCHEME) $(LIB_CFLAGS) $(LIB_CPPFLAGS)
GENERATE[md5-x86_64.s]=asm/md5-x86_64.pl $(PERLASM_SCHEME)
GENERATE[md5-sparcv9.S]=asm/md5-sparcv9.pl $(PERLASM_SCHEME)
INCLUDE[md5-sparcv9.o]=..

View file

@ -0,0 +1,164 @@
/*
* 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
*/
#include <stdio.h>
#include "md5_locl.h"
#include <openssl/opensslv.h>
/*
* Implemented from RFC1321 The MD5 Message-Digest Algorithm
*/
#define INIT_DATA_A (unsigned long)0x67452301L
#define INIT_DATA_B (unsigned long)0xefcdab89L
#define INIT_DATA_C (unsigned long)0x98badcfeL
#define INIT_DATA_D (unsigned long)0x10325476L
int MD5_Init(MD5_CTX *c)
{
memset(c, 0, sizeof(*c));
c->A = INIT_DATA_A;
c->B = INIT_DATA_B;
c->C = INIT_DATA_C;
c->D = INIT_DATA_D;
return 1;
}
#ifndef md5_block_data_order
# ifdef X
# undef X
# endif
void md5_block_data_order(MD5_CTX *c, const void *data_, size_t num)
{
const unsigned char *data = data_;
register unsigned MD32_REG_T A, B, C, D, l;
# ifndef MD32_XARRAY
/* See comment in crypto/sha/sha_locl.h for details. */
unsigned MD32_REG_T XX0, XX1, XX2, XX3, XX4, XX5, XX6, XX7,
XX8, XX9, XX10, XX11, XX12, XX13, XX14, XX15;
# define X(i) XX##i
# else
MD5_LONG XX[MD5_LBLOCK];
# define X(i) XX[i]
# endif
A = c->A;
B = c->B;
C = c->C;
D = c->D;
for (; num--;) {
(void)HOST_c2l(data, l);
X(0) = l;
(void)HOST_c2l(data, l);
X(1) = l;
/* Round 0 */
R0(A, B, C, D, X(0), 7, 0xd76aa478L);
(void)HOST_c2l(data, l);
X(2) = l;
R0(D, A, B, C, X(1), 12, 0xe8c7b756L);
(void)HOST_c2l(data, l);
X(3) = l;
R0(C, D, A, B, X(2), 17, 0x242070dbL);
(void)HOST_c2l(data, l);
X(4) = l;
R0(B, C, D, A, X(3), 22, 0xc1bdceeeL);
(void)HOST_c2l(data, l);
X(5) = l;
R0(A, B, C, D, X(4), 7, 0xf57c0fafL);
(void)HOST_c2l(data, l);
X(6) = l;
R0(D, A, B, C, X(5), 12, 0x4787c62aL);
(void)HOST_c2l(data, l);
X(7) = l;
R0(C, D, A, B, X(6), 17, 0xa8304613L);
(void)HOST_c2l(data, l);
X(8) = l;
R0(B, C, D, A, X(7), 22, 0xfd469501L);
(void)HOST_c2l(data, l);
X(9) = l;
R0(A, B, C, D, X(8), 7, 0x698098d8L);
(void)HOST_c2l(data, l);
X(10) = l;
R0(D, A, B, C, X(9), 12, 0x8b44f7afL);
(void)HOST_c2l(data, l);
X(11) = l;
R0(C, D, A, B, X(10), 17, 0xffff5bb1L);
(void)HOST_c2l(data, l);
X(12) = l;
R0(B, C, D, A, X(11), 22, 0x895cd7beL);
(void)HOST_c2l(data, l);
X(13) = l;
R0(A, B, C, D, X(12), 7, 0x6b901122L);
(void)HOST_c2l(data, l);
X(14) = l;
R0(D, A, B, C, X(13), 12, 0xfd987193L);
(void)HOST_c2l(data, l);
X(15) = l;
R0(C, D, A, B, X(14), 17, 0xa679438eL);
R0(B, C, D, A, X(15), 22, 0x49b40821L);
/* Round 1 */
R1(A, B, C, D, X(1), 5, 0xf61e2562L);
R1(D, A, B, C, X(6), 9, 0xc040b340L);
R1(C, D, A, B, X(11), 14, 0x265e5a51L);
R1(B, C, D, A, X(0), 20, 0xe9b6c7aaL);
R1(A, B, C, D, X(5), 5, 0xd62f105dL);
R1(D, A, B, C, X(10), 9, 0x02441453L);
R1(C, D, A, B, X(15), 14, 0xd8a1e681L);
R1(B, C, D, A, X(4), 20, 0xe7d3fbc8L);
R1(A, B, C, D, X(9), 5, 0x21e1cde6L);
R1(D, A, B, C, X(14), 9, 0xc33707d6L);
R1(C, D, A, B, X(3), 14, 0xf4d50d87L);
R1(B, C, D, A, X(8), 20, 0x455a14edL);
R1(A, B, C, D, X(13), 5, 0xa9e3e905L);
R1(D, A, B, C, X(2), 9, 0xfcefa3f8L);
R1(C, D, A, B, X(7), 14, 0x676f02d9L);
R1(B, C, D, A, X(12), 20, 0x8d2a4c8aL);
/* Round 2 */
R2(A, B, C, D, X(5), 4, 0xfffa3942L);
R2(D, A, B, C, X(8), 11, 0x8771f681L);
R2(C, D, A, B, X(11), 16, 0x6d9d6122L);
R2(B, C, D, A, X(14), 23, 0xfde5380cL);
R2(A, B, C, D, X(1), 4, 0xa4beea44L);
R2(D, A, B, C, X(4), 11, 0x4bdecfa9L);
R2(C, D, A, B, X(7), 16, 0xf6bb4b60L);
R2(B, C, D, A, X(10), 23, 0xbebfbc70L);
R2(A, B, C, D, X(13), 4, 0x289b7ec6L);
R2(D, A, B, C, X(0), 11, 0xeaa127faL);
R2(C, D, A, B, X(3), 16, 0xd4ef3085L);
R2(B, C, D, A, X(6), 23, 0x04881d05L);
R2(A, B, C, D, X(9), 4, 0xd9d4d039L);
R2(D, A, B, C, X(12), 11, 0xe6db99e5L);
R2(C, D, A, B, X(15), 16, 0x1fa27cf8L);
R2(B, C, D, A, X(2), 23, 0xc4ac5665L);
/* Round 3 */
R3(A, B, C, D, X(0), 6, 0xf4292244L);
R3(D, A, B, C, X(7), 10, 0x432aff97L);
R3(C, D, A, B, X(14), 15, 0xab9423a7L);
R3(B, C, D, A, X(5), 21, 0xfc93a039L);
R3(A, B, C, D, X(12), 6, 0x655b59c3L);
R3(D, A, B, C, X(3), 10, 0x8f0ccc92L);
R3(C, D, A, B, X(10), 15, 0xffeff47dL);
R3(B, C, D, A, X(1), 21, 0x85845dd1L);
R3(A, B, C, D, X(8), 6, 0x6fa87e4fL);
R3(D, A, B, C, X(15), 10, 0xfe2ce6e0L);
R3(C, D, A, B, X(6), 15, 0xa3014314L);
R3(B, C, D, A, X(13), 21, 0x4e0811a1L);
R3(A, B, C, D, X(4), 6, 0xf7537e82L);
R3(D, A, B, C, X(11), 10, 0xbd3af235L);
R3(C, D, A, B, X(2), 15, 0x2ad7d2bbL);
R3(B, C, D, A, X(9), 21, 0xeb86d391L);
A = c->A += A;
B = c->B += B;
C = c->C += C;
D = c->D += D;
}
}
#endif

View file

@ -0,0 +1,80 @@
/*
* 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
*/
#include <stdlib.h>
#include <string.h>
#include <openssl/e_os2.h>
#include <openssl/md5.h>
#ifdef MD5_ASM
# if defined(__i386) || defined(__i386__) || defined(_M_IX86) || \
defined(__x86_64) || defined(__x86_64__) || defined(_M_AMD64) || defined(_M_X64)
# define md5_block_data_order md5_block_asm_data_order
# elif defined(__ia64) || defined(__ia64__) || defined(_M_IA64)
# define md5_block_data_order md5_block_asm_data_order
# elif defined(__sparc) || defined(__sparc__)
# define md5_block_data_order md5_block_asm_data_order
# endif
#endif
void md5_block_data_order(MD5_CTX *c, const void *p, size_t num);
#define DATA_ORDER_IS_LITTLE_ENDIAN
#define HASH_LONG MD5_LONG
#define HASH_CTX MD5_CTX
#define HASH_CBLOCK MD5_CBLOCK
#define HASH_UPDATE MD5_Update
#define HASH_TRANSFORM MD5_Transform
#define HASH_FINAL MD5_Final
#define HASH_MAKE_STRING(c,s) do { \
unsigned long ll; \
ll=(c)->A; (void)HOST_l2c(ll,(s)); \
ll=(c)->B; (void)HOST_l2c(ll,(s)); \
ll=(c)->C; (void)HOST_l2c(ll,(s)); \
ll=(c)->D; (void)HOST_l2c(ll,(s)); \
} while (0)
#define HASH_BLOCK_DATA_ORDER md5_block_data_order
#include "internal/md32_common.h"
/*-
#define F(x,y,z) (((x) & (y)) | ((~(x)) & (z)))
#define G(x,y,z) (((x) & (z)) | ((y) & (~(z))))
*/
/*
* As pointed out by Wei Dai, the above can be simplified to the code
* below. Wei attributes these optimizations to Peter Gutmann's
* SHS code, and he attributes it to Rich Schroeppel.
*/
#define F(b,c,d) ((((c) ^ (d)) & (b)) ^ (d))
#define G(b,c,d) ((((b) ^ (c)) & (d)) ^ (c))
#define H(b,c,d) ((b) ^ (c) ^ (d))
#define I(b,c,d) (((~(d)) | (b)) ^ (c))
#define R0(a,b,c,d,k,s,t) { \
a+=((k)+(t)+F((b),(c),(d))); \
a=ROTATE(a,s); \
a+=b; };\
#define R1(a,b,c,d,k,s,t) { \
a+=((k)+(t)+G((b),(c),(d))); \
a=ROTATE(a,s); \
a+=b; };
#define R2(a,b,c,d,k,s,t) { \
a+=((k)+(t)+H((b),(c),(d))); \
a=ROTATE(a,s); \
a+=b; };
#define R3(a,b,c,d,k,s,t) { \
a+=((k)+(t)+I((b),(c),(d))); \
a=ROTATE(a,s); \
a+=b; };

View file

@ -0,0 +1,47 @@
/*
* 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
*/
#include <stdio.h>
#include <string.h>
#include <openssl/md5.h>
#include <openssl/crypto.h>
#ifdef CHARSET_EBCDIC
# include <openssl/ebcdic.h>
#endif
unsigned char *MD5(const unsigned char *d, size_t n, unsigned char *md)
{
MD5_CTX c;
static unsigned char m[MD5_DIGEST_LENGTH];
if (md == NULL)
md = m;
if (!MD5_Init(&c))
return NULL;
#ifndef CHARSET_EBCDIC
MD5_Update(&c, d, n);
#else
{
char temp[1024];
unsigned long chunk;
while (n > 0) {
chunk = (n > sizeof(temp)) ? sizeof(temp) : n;
ebcdic2ascii(temp, d, chunk);
MD5_Update(&c, temp, chunk);
n -= chunk;
d += chunk;
}
}
#endif
MD5_Final(md, &c);
OPENSSL_cleanse(&c, sizeof(c)); /* security consideration */
return md;
}