abc/GmSSL
1
0
Fork 0
mirror of https://github.com/guanzhi/GmSSL.git synced 2026-08-12 08:53:39 +08:00
Code Issues Packages Projects Releases Wiki Activity

Merge remote-tracking branch 'origin/master'

Browse Source 
# Conflicts:
#	README.md
This commit is contained in:
Zhi Guan
2017-02-14 16:12:29 +08:00
parent d2254170b8
commit 43fed1108d
3503 changed files with 320546 additions and 408546 deletions
Download Patch File Download Diff File Expand all files Collapse all files

60
crypto/bn/asm/x86-mont.pl
View File

@@ -1,4 +1,11 @@
#!/usr/bin/env perl
#! /usr/bin/env perl
# 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
# ====================================================================
# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
@@ -30,6 +37,9 @@ $0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
push(@INC,"${dir}","${dir}../../perlasm");
require "x86asm.pl";
$output = pop;
open STDOUT,">$output";
&asm_init($ARGV[0],$0);
$sse2=0;
@@ -63,33 +73,57 @@ $frame=32; # size of above frame rounded up to 16n
&lea ("esi",&wparam(0)); # put aside pointer to argument block
&lea ("edx",&wparam(1)); # load ap
&mov ("ebp","esp"); # saved stack pointer!
&add ("edi",2); # extra two words on top of tp
&neg ("edi");
&lea ("esp",&DWP(-$frame,"esp","edi",4)); # alloca($frame+4*(num+2))
&lea ("ebp",&DWP(-$frame,"esp","edi",4)); # future alloca($frame+4*(num+2))
&neg ("edi");
# minimize cache contention by arraning 2K window between stack
# pointer and ap argument [np is also position sensitive vector,
# but it's assumed to be near ap, as it's allocated at ~same
# time].
&mov ("eax","esp");
&mov ("eax","ebp");
&sub ("eax","edx");
&and ("eax",2047);
&sub ("esp","eax"); # this aligns sp and ap modulo 2048
&sub ("ebp","eax"); # this aligns sp and ap modulo 2048
&xor ("edx","esp");
&xor ("edx","ebp");
&and ("edx",2048);
&xor ("edx",2048);
&sub ("esp","edx"); # this splits them apart modulo 4096
&sub ("ebp","edx"); # this splits them apart modulo 4096
&and ("esp",-64); # align to cache line
&and ("ebp",-64); # align to cache line
# An OS-agnostic version of __chkstk.
#
# Some OSes (Windows) insist on stack being "wired" to
# physical memory in strictly sequential manner, i.e. if stack
# allocation spans two pages, then reference to farmost one can
# be punishable by SEGV. But page walking can do good even on
# other OSes, because it guarantees that villain thread hits
# the guard page before it can make damage to innocent one...
&mov ("eax","esp");
&sub ("eax","ebp");
&and ("eax",-4096);
&mov ("edx","esp"); # saved stack pointer!
&lea ("esp",&DWP(0,"ebp","eax"));
&mov ("eax",&DWP(0,"esp"));
&cmp ("esp","ebp");
&ja (&label("page_walk"));
&jmp (&label("page_walk_done"));
&set_label("page_walk",16);
&lea ("esp",&DWP(-4096,"esp"));
&mov ("eax",&DWP(0,"esp"));
&cmp ("esp","ebp");
&ja (&label("page_walk"));
&set_label("page_walk_done");
################################# load argument block...
&mov ("eax",&DWP(0*4,"esi"));# BN_ULONG *rp
&mov ("ebx",&DWP(1*4,"esi"));# const BN_ULONG *ap
&mov ("ecx",&DWP(2*4,"esi"));# const BN_ULONG *bp
&mov ("edx",&DWP(3*4,"esi"));# const BN_ULONG *np
&mov ("ebp",&DWP(3*4,"esi"));# const BN_ULONG *np
&mov ("esi",&DWP(4*4,"esi"));# const BN_ULONG *n0
#&mov ("edi",&DWP(5*4,"esi"));# int num
@@ -97,11 +131,11 @@ $frame=32; # size of above frame rounded up to 16n
&mov ($_rp,"eax"); # ... save a copy of argument block
&mov ($_ap,"ebx");
&mov ($_bp,"ecx");
&mov ($_np,"edx");
&mov ($_np,"ebp");
&mov ($_n0,"esi");
&lea ($num,&DWP(-3,"edi")); # num=num-1 to assist modulo-scheduling
#&mov ($_num,$num); # redundant as $num is not reused
&mov ($_sp,"ebp"); # saved stack pointer!
&mov ($_sp,"edx"); # saved stack pointer!
if($sse2) {
$acc0="mm0"; # mmx register bank layout
@@ -267,7 +301,7 @@ if (0) {
&xor ("eax","eax"); # signal "not fast enough [yet]"
&jmp (&label("just_leave"));
# While the below code provides competitive performance for
# all key lengthes on modern Intel cores, it's still more
# all key lengths on modern Intel cores, it's still more
# than 10% slower for 4096-bit key elsewhere:-( "Competitive"
# means compared to the original integer-only assembler.
# 512-bit RSA sign is better by ~40%, but that's about all
@@ -591,3 +625,5 @@ $sbit=$num;
&asciz("Montgomery Multiplication for x86, CRYPTOGAMS by <appro\@openssl.org>");
&asm_finish();
close STDOUT;