studio/cli
2
0
Fork 0
Code Issues Pull Requests Actions Packages Releases 4 Wiki Activity

chore: checkpoint before Python removal

Browse Source
This commit is contained in:
Sienna Meridian Satterwhite
2026-03-26 22:33:59 +00:00
parent 683cec9307
commit e568ddf82a
29972 changed files with 11269302 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1131
vendor/aws-lc-sys/aws-lc/crypto/chacha/asm/chacha-armv4.pl vendored Executable file
View File

File diff suppressed because it is too large Load Diff

1238
vendor/aws-lc-sys/aws-lc/crypto/chacha/asm/chacha-armv8.pl vendored Executable file
View File

File diff suppressed because it is too large Load Diff

756
vendor/aws-lc-sys/aws-lc/crypto/chacha/asm/chacha-x86.pl vendored Executable file
View File

@@ -0,0 +1,756 @@
#! /usr/bin/env perl
# Copyright 2016 The OpenSSL Project Authors. All Rights Reserved.
# SPDX-License-Identifier: Apache-2.0
#
# ====================================================================
# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
# project.
# ====================================================================
#
# January 2015
#
# ChaCha20 for x86.
#
# Performance in cycles per byte out of large buffer.
#
# 1xIALU/gcc 4xSSSE3
# Pentium 17.5/+80%
# PIII 14.2/+60%
# P4 18.6/+84%
# Core2 9.56/+89% 4.83
# Westmere 9.50/+45% 3.35
# Sandy Bridge 10.5/+47% 3.20
# Haswell 8.15/+50% 2.83
# Skylake 7.53/+22% 2.75
# Silvermont 17.4/+36% 8.35
# Goldmont 13.4/+40% 4.36
# Sledgehammer 10.2/+54%
# Bulldozer 13.4/+50% 4.38(*)
#
# (*) Bulldozer actually executes 4xXOP code path that delivers 3.55;
#
# Modified from upstream OpenSSL to remove the XOP code.
# The first two arguments should always be the flavour and output file path.
if ($#ARGV < 1) { die "Not enough arguments provided.
Two arguments are necessary: the flavour and the output file path."; }
$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
push(@INC,"${dir}","${dir}../../perlasm");
require "x86asm.pl";
$output=$ARGV[1];
open STDOUT,">$output";
&asm_init($ARGV[0]);
$xmm=$ymm=1;
$gasver=999; # enable everything
$a="eax";
($b,$b_)=("ebx","ebp");
($c,$c_)=("ecx","esi");
($d,$d_)=("edx","edi");
sub QUARTERROUND {
my ($ai,$bi,$ci,$di,$i)=@_;
my ($an,$bn,$cn,$dn)=map(($_&~3)+(($_+1)&3),($ai,$bi,$ci,$di)); # next
my ($ap,$bp,$cp,$dp)=map(($_&~3)+(($_-1)&3),($ai,$bi,$ci,$di)); # previous
# a b c d
#
# 0 4 8 12 < even round
# 1 5 9 13
# 2 6 10 14
# 3 7 11 15
# 0 5 10 15 < odd round
# 1 6 11 12
# 2 7 8 13
# 3 4 9 14
if ($i==0) {
my $j=4;
($ap,$bp,$cp,$dp)=map(($_&~3)+(($_-$j--)&3),($ap,$bp,$cp,$dp));
} elsif ($i==3) {
my $j=0;
($an,$bn,$cn,$dn)=map(($_&~3)+(($_+$j++)&3),($an,$bn,$cn,$dn));
} elsif ($i==4) {
my $j=4;
($ap,$bp,$cp,$dp)=map(($_&~3)+(($_+$j--)&3),($ap,$bp,$cp,$dp));
} elsif ($i==7) {
my $j=0;
($an,$bn,$cn,$dn)=map(($_&~3)+(($_-$j++)&3),($an,$bn,$cn,$dn));
}
#&add ($a,$b); # see elsewhere
&xor ($d,$a);
&mov (&DWP(4*$cp,"esp"),$c_) if ($ai>0 && $ai<3);
&rol ($d,16);
&mov (&DWP(4*$bp,"esp"),$b_) if ($i!=0);
&add ($c,$d);
&mov ($c_,&DWP(4*$cn,"esp")) if ($ai>0 && $ai<3);
&xor ($b,$c);
&mov ($d_,&DWP(4*$dn,"esp")) if ($di!=$dn);
&rol ($b,12);
&mov ($b_,&DWP(4*$bn,"esp")) if ($i<7);
&mov ($b_,&DWP(128,"esp")) if ($i==7); # loop counter
&add ($a,$b);
&xor ($d,$a);
&mov (&DWP(4*$ai,"esp"),$a);
&rol ($d,8);
&mov ($a,&DWP(4*$an,"esp"));
&add ($c,$d);
&mov (&DWP(4*$di,"esp"),$d) if ($di!=$dn);
&mov ($d_,$d) if ($di==$dn);
&xor ($b,$c);
&add ($a,$b_) if ($i<7); # elsewhere
&rol ($b,7);
($b,$b_)=($b_,$b);
($c,$c_)=($c_,$c);
($d,$d_)=($d_,$d);
}
&static_label("ssse3_data");
&static_label("pic_point");
&function_begin("ChaCha20_ctr32_nohw");
&mov ("esi",&wparam(3)); # key
&mov ("edi",&wparam(4)); # counter and nonce
&stack_push(33);
&mov ("eax",&DWP(4*0,"esi")); # copy key
&mov ("ebx",&DWP(4*1,"esi"));
&mov ("ecx",&DWP(4*2,"esi"));
&mov ("edx",&DWP(4*3,"esi"));
&mov (&DWP(64+4*4,"esp"),"eax");
&mov (&DWP(64+4*5,"esp"),"ebx");
&mov (&DWP(64+4*6,"esp"),"ecx");
&mov (&DWP(64+4*7,"esp"),"edx");
&mov ("eax",&DWP(4*4,"esi"));
&mov ("ebx",&DWP(4*5,"esi"));
&mov ("ecx",&DWP(4*6,"esi"));
&mov ("edx",&DWP(4*7,"esi"));
&mov (&DWP(64+4*8,"esp"),"eax");
&mov (&DWP(64+4*9,"esp"),"ebx");
&mov (&DWP(64+4*10,"esp"),"ecx");
&mov (&DWP(64+4*11,"esp"),"edx");
&mov ("eax",&DWP(4*0,"edi")); # copy counter and nonce
&mov ("ebx",&DWP(4*1,"edi"));
&mov ("ecx",&DWP(4*2,"edi"));
&mov ("edx",&DWP(4*3,"edi"));
&sub ("eax",1);
&mov (&DWP(64+4*12,"esp"),"eax");
&mov (&DWP(64+4*13,"esp"),"ebx");
&mov (&DWP(64+4*14,"esp"),"ecx");
&mov (&DWP(64+4*15,"esp"),"edx");
&jmp (&label("entry"));
&set_label("outer_loop",16);
&mov (&wparam(1),$b); # save input
&mov (&wparam(0),$a); # save output
&mov (&wparam(2),$c); # save len
&set_label("entry");
&mov ($a,0x61707865);
&mov (&DWP(4*1,"esp"),0x3320646e);
&mov (&DWP(4*2,"esp"),0x79622d32);
&mov (&DWP(4*3,"esp"),0x6b206574);
&mov ($b, &DWP(64+4*5,"esp")); # copy key material
&mov ($b_,&DWP(64+4*6,"esp"));
&mov ($c, &DWP(64+4*10,"esp"));
&mov ($c_,&DWP(64+4*11,"esp"));
&mov ($d, &DWP(64+4*13,"esp"));
&mov ($d_,&DWP(64+4*14,"esp"));
&mov (&DWP(4*5,"esp"),$b);
&mov (&DWP(4*6,"esp"),$b_);
&mov (&DWP(4*10,"esp"),$c);
&mov (&DWP(4*11,"esp"),$c_);
&mov (&DWP(4*13,"esp"),$d);
&mov (&DWP(4*14,"esp"),$d_);
&mov ($b, &DWP(64+4*7,"esp"));
&mov ($d_,&DWP(64+4*15,"esp"));
&mov ($d, &DWP(64+4*12,"esp"));
&mov ($b_,&DWP(64+4*4,"esp"));
&mov ($c, &DWP(64+4*8,"esp"));
&mov ($c_,&DWP(64+4*9,"esp"));
&add ($d,1); # counter value
&mov (&DWP(4*7,"esp"),$b);
&mov (&DWP(4*15,"esp"),$d_);
&mov (&DWP(64+4*12,"esp"),$d); # save counter value
&mov ($b,10); # loop counter
&jmp (&label("loop"));
&set_label("loop",16);
&add ($a,$b_); # elsewhere
&mov (&DWP(128,"esp"),$b); # save loop counter
&mov ($b,$b_);
&QUARTERROUND(0, 4, 8, 12, 0);
&QUARTERROUND(1, 5, 9, 13, 1);
&QUARTERROUND(2, 6,10, 14, 2);
&QUARTERROUND(3, 7,11, 15, 3);
&QUARTERROUND(0, 5,10, 15, 4);
&QUARTERROUND(1, 6,11, 12, 5);
&QUARTERROUND(2, 7, 8, 13, 6);
&QUARTERROUND(3, 4, 9, 14, 7);
&dec ($b);
&jnz (&label("loop"));
&mov ($b,&wparam(2)); # load len
&add ($a,0x61707865); # accumulate key material
&add ($b_,&DWP(64+4*4,"esp"));
&add ($c, &DWP(64+4*8,"esp"));
&add ($c_,&DWP(64+4*9,"esp"));
&cmp ($b,64);
&jb (&label("tail"));
&mov ($b,&wparam(1)); # load input pointer
&add ($d, &DWP(64+4*12,"esp"));
&add ($d_,&DWP(64+4*14,"esp"));
&xor ($a, &DWP(4*0,$b)); # xor with input
&xor ($b_,&DWP(4*4,$b));
&mov (&DWP(4*0,"esp"),$a);
&mov ($a,&wparam(0)); # load output pointer
&xor ($c, &DWP(4*8,$b));
&xor ($c_,&DWP(4*9,$b));
&xor ($d, &DWP(4*12,$b));
&xor ($d_,&DWP(4*14,$b));
&mov (&DWP(4*4,$a),$b_); # write output
&mov (&DWP(4*8,$a),$c);
&mov (&DWP(4*9,$a),$c_);
&mov (&DWP(4*12,$a),$d);
&mov (&DWP(4*14,$a),$d_);
&mov ($b_,&DWP(4*1,"esp"));
&mov ($c, &DWP(4*2,"esp"));
&mov ($c_,&DWP(4*3,"esp"));
&mov ($d, &DWP(4*5,"esp"));
&mov ($d_,&DWP(4*6,"esp"));
&add ($b_,0x3320646e); # accumulate key material
&add ($c, 0x79622d32);
&add ($c_,0x6b206574);
&add ($d, &DWP(64+4*5,"esp"));
&add ($d_,&DWP(64+4*6,"esp"));
&xor ($b_,&DWP(4*1,$b));
&xor ($c, &DWP(4*2,$b));
&xor ($c_,&DWP(4*3,$b));
&xor ($d, &DWP(4*5,$b));
&xor ($d_,&DWP(4*6,$b));
&mov (&DWP(4*1,$a),$b_);
&mov (&DWP(4*2,$a),$c);
&mov (&DWP(4*3,$a),$c_);
&mov (&DWP(4*5,$a),$d);
&mov (&DWP(4*6,$a),$d_);
&mov ($b_,&DWP(4*7,"esp"));
&mov ($c, &DWP(4*10,"esp"));
&mov ($c_,&DWP(4*11,"esp"));
&mov ($d, &DWP(4*13,"esp"));
&mov ($d_,&DWP(4*15,"esp"));
&add ($b_,&DWP(64+4*7,"esp"));
&add ($c, &DWP(64+4*10,"esp"));
&add ($c_,&DWP(64+4*11,"esp"));
&add ($d, &DWP(64+4*13,"esp"));
&add ($d_,&DWP(64+4*15,"esp"));
&xor ($b_,&DWP(4*7,$b));
&xor ($c, &DWP(4*10,$b));
&xor ($c_,&DWP(4*11,$b));
&xor ($d, &DWP(4*13,$b));
&xor ($d_,&DWP(4*15,$b));
&lea ($b,&DWP(4*16,$b));
&mov (&DWP(4*7,$a),$b_);
&mov ($b_,&DWP(4*0,"esp"));
&mov (&DWP(4*10,$a),$c);
&mov ($c,&wparam(2)); # len
&mov (&DWP(4*11,$a),$c_);
&mov (&DWP(4*13,$a),$d);
&mov (&DWP(4*15,$a),$d_);
&mov (&DWP(4*0,$a),$b_);
&lea ($a,&DWP(4*16,$a));
&sub ($c,64);
&jnz (&label("outer_loop"));
&jmp (&label("done"));
&set_label("tail");
&add ($d, &DWP(64+4*12,"esp"));
&add ($d_,&DWP(64+4*14,"esp"));
&mov (&DWP(4*0,"esp"),$a);
&mov (&DWP(4*4,"esp"),$b_);
&mov (&DWP(4*8,"esp"),$c);
&mov (&DWP(4*9,"esp"),$c_);
&mov (&DWP(4*12,"esp"),$d);
&mov (&DWP(4*14,"esp"),$d_);
&mov ($b_,&DWP(4*1,"esp"));
&mov ($c, &DWP(4*2,"esp"));
&mov ($c_,&DWP(4*3,"esp"));
&mov ($d, &DWP(4*5,"esp"));
&mov ($d_,&DWP(4*6,"esp"));
&add ($b_,0x3320646e); # accumulate key material
&add ($c, 0x79622d32);
&add ($c_,0x6b206574);
&add ($d, &DWP(64+4*5,"esp"));
&add ($d_,&DWP(64+4*6,"esp"));
&mov (&DWP(4*1,"esp"),$b_);
&mov (&DWP(4*2,"esp"),$c);
&mov (&DWP(4*3,"esp"),$c_);
&mov (&DWP(4*5,"esp"),$d);
&mov (&DWP(4*6,"esp"),$d_);
&mov ($b_,&DWP(4*7,"esp"));
&mov ($c, &DWP(4*10,"esp"));
&mov ($c_,&DWP(4*11,"esp"));
&mov ($d, &DWP(4*13,"esp"));
&mov ($d_,&DWP(4*15,"esp"));
&add ($b_,&DWP(64+4*7,"esp"));
&add ($c, &DWP(64+4*10,"esp"));
&add ($c_,&DWP(64+4*11,"esp"));
&add ($d, &DWP(64+4*13,"esp"));
&add ($d_,&DWP(64+4*15,"esp"));
&mov (&DWP(4*7,"esp"),$b_);
&mov ($b_,&wparam(1)); # load input
&mov (&DWP(4*10,"esp"),$c);
&mov ($c,&wparam(0)); # load output
&mov (&DWP(4*11,"esp"),$c_);
&xor ($c_,$c_);
&mov (&DWP(4*13,"esp"),$d);
&mov (&DWP(4*15,"esp"),$d_);
&xor ("eax","eax");
&xor ("edx","edx");
&set_label("tail_loop");
&movb ("al",&BP(0,$c_,$b_));
&movb ("dl",&BP(0,"esp",$c_));
&lea ($c_,&DWP(1,$c_));
&xor ("al","dl");
&mov (&BP(-1,$c,$c_),"al");
&dec ($b);
&jnz (&label("tail_loop"));
&set_label("done");
&stack_pop(33);
&function_end("ChaCha20_ctr32_nohw");
if ($xmm) {
my ($xa,$xa_,$xb,$xb_,$xc,$xc_,$xd,$xd_)=map("xmm$_",(0..7));
my ($out,$inp,$len)=("edi","esi","ecx");
sub QUARTERROUND_SSSE3 {
my ($ai,$bi,$ci,$di,$i)=@_;
my ($an,$bn,$cn,$dn)=map(($_&~3)+(($_+1)&3),($ai,$bi,$ci,$di)); # next
my ($ap,$bp,$cp,$dp)=map(($_&~3)+(($_-1)&3),($ai,$bi,$ci,$di)); # previous
# a b c d
#
# 0 4 8 12 < even round
# 1 5 9 13
# 2 6 10 14
# 3 7 11 15
# 0 5 10 15 < odd round
# 1 6 11 12
# 2 7 8 13
# 3 4 9 14
if ($i==0) {
my $j=4;
($ap,$bp,$cp,$dp)=map(($_&~3)+(($_-$j--)&3),($ap,$bp,$cp,$dp));
} elsif ($i==3) {
my $j=0;
($an,$bn,$cn,$dn)=map(($_&~3)+(($_+$j++)&3),($an,$bn,$cn,$dn));
} elsif ($i==4) {
my $j=4;
($ap,$bp,$cp,$dp)=map(($_&~3)+(($_+$j--)&3),($ap,$bp,$cp,$dp));
} elsif ($i==7) {
my $j=0;
($an,$bn,$cn,$dn)=map(($_&~3)+(($_-$j++)&3),($an,$bn,$cn,$dn));
}
#&paddd ($xa,$xb); # see elsewhere
#&pxor ($xd,$xa); # see elsewhere
&movdqa(&QWP(16*$cp-128,"ebx"),$xc_) if ($ai>0 && $ai<3);
&pshufb ($xd,&QWP(0,"eax")); # rot16
&movdqa(&QWP(16*$bp-128,"ebx"),$xb_) if ($i!=0);
&paddd ($xc,$xd);
&movdqa($xc_,&QWP(16*$cn-128,"ebx")) if ($ai>0 && $ai<3);
&pxor ($xb,$xc);
&movdqa($xb_,&QWP(16*$bn-128,"ebx")) if ($i<7);
&movdqa ($xa_,$xb); # borrow as temporary
&pslld ($xb,12);
&psrld ($xa_,20);
&por ($xb,$xa_);
&movdqa($xa_,&QWP(16*$an-128,"ebx"));
&paddd ($xa,$xb);
&movdqa($xd_,&QWP(16*$dn-128,"ebx")) if ($di!=$dn);
&pxor ($xd,$xa);
&movdqa (&QWP(16*$ai-128,"ebx"),$xa);
&pshufb ($xd,&QWP(16,"eax")); # rot8
&paddd ($xc,$xd);
&movdqa (&QWP(16*$di-128,"ebx"),$xd) if ($di!=$dn);
&movdqa ($xd_,$xd) if ($di==$dn);
&pxor ($xb,$xc);
&paddd ($xa_,$xb_) if ($i<7); # elsewhere
&movdqa ($xa,$xb); # borrow as temporary
&pslld ($xb,7);
&psrld ($xa,25);
&pxor ($xd_,$xa_) if ($i<7); # elsewhere
&por ($xb,$xa);
($xa,$xa_)=($xa_,$xa);
($xb,$xb_)=($xb_,$xb);
($xc,$xc_)=($xc_,$xc);
($xd,$xd_)=($xd_,$xd);
}
&function_begin("ChaCha20_ctr32_ssse3");
&call (&label("pic_point"));
&set_label("pic_point");
&blindpop("eax");
&mov ($out,&wparam(0));
&mov ($inp,&wparam(1));
&mov ($len,&wparam(2));
&mov ("edx",&wparam(3)); # key
&mov ("ebx",&wparam(4)); # counter and nonce
&mov ("ebp","esp");
&stack_push (131);
&and ("esp",-64);
&mov (&DWP(512,"esp"),"ebp");
&lea ("eax",&DWP(&label("ssse3_data")."-".
&label("pic_point"),"eax"));
&movdqu ("xmm3",&QWP(0,"ebx")); # counter and nonce
if (defined($gasver) && $gasver>=2.17) { # even though we encode
# pshufb manually, we
# handle only register
# operands, while this
# segment uses memory
# operand...
&cmp ($len,64*4);
&jb (&label("1x"));
&mov (&DWP(512+4,"esp"),"edx"); # offload pointers
&mov (&DWP(512+8,"esp"),"ebx");
&sub ($len,64*4); # bias len
&lea ("ebp",&DWP(256+128,"esp")); # size optimization
&movdqu ("xmm7",&QWP(0,"edx")); # key
&pshufd ("xmm0","xmm3",0x00);
&pshufd ("xmm1","xmm3",0x55);
&pshufd ("xmm2","xmm3",0xaa);
&pshufd ("xmm3","xmm3",0xff);
&paddd ("xmm0",&QWP(16*3,"eax")); # fix counters
&pshufd ("xmm4","xmm7",0x00);
&pshufd ("xmm5","xmm7",0x55);
&psubd ("xmm0",&QWP(16*4,"eax"));
&pshufd ("xmm6","xmm7",0xaa);
&pshufd ("xmm7","xmm7",0xff);
&movdqa (&QWP(16*12-128,"ebp"),"xmm0");
&movdqa (&QWP(16*13-128,"ebp"),"xmm1");
&movdqa (&QWP(16*14-128,"ebp"),"xmm2");
&movdqa (&QWP(16*15-128,"ebp"),"xmm3");
&movdqu ("xmm3",&QWP(16,"edx")); # key
&movdqa (&QWP(16*4-128,"ebp"),"xmm4");
&movdqa (&QWP(16*5-128,"ebp"),"xmm5");
&movdqa (&QWP(16*6-128,"ebp"),"xmm6");
&movdqa (&QWP(16*7-128,"ebp"),"xmm7");
&movdqa ("xmm7",&QWP(16*2,"eax")); # sigma
&lea ("ebx",&DWP(128,"esp")); # size optimization
&pshufd ("xmm0","xmm3",0x00);
&pshufd ("xmm1","xmm3",0x55);
&pshufd ("xmm2","xmm3",0xaa);
&pshufd ("xmm3","xmm3",0xff);
&pshufd ("xmm4","xmm7",0x00);
&pshufd ("xmm5","xmm7",0x55);
&pshufd ("xmm6","xmm7",0xaa);
&pshufd ("xmm7","xmm7",0xff);
&movdqa (&QWP(16*8-128,"ebp"),"xmm0");
&movdqa (&QWP(16*9-128,"ebp"),"xmm1");
&movdqa (&QWP(16*10-128,"ebp"),"xmm2");
&movdqa (&QWP(16*11-128,"ebp"),"xmm3");
&movdqa (&QWP(16*0-128,"ebp"),"xmm4");
&movdqa (&QWP(16*1-128,"ebp"),"xmm5");
&movdqa (&QWP(16*2-128,"ebp"),"xmm6");
&movdqa (&QWP(16*3-128,"ebp"),"xmm7");
&lea ($inp,&DWP(128,$inp)); # size optimization
&lea ($out,&DWP(128,$out)); # size optimization
&jmp (&label("outer_loop"));
&set_label("outer_loop",16);
#&movdqa ("xmm0",&QWP(16*0-128,"ebp")); # copy key material
&movdqa ("xmm1",&QWP(16*1-128,"ebp"));
&movdqa ("xmm2",&QWP(16*2-128,"ebp"));
&movdqa ("xmm3",&QWP(16*3-128,"ebp"));
#&movdqa ("xmm4",&QWP(16*4-128,"ebp"));
&movdqa ("xmm5",&QWP(16*5-128,"ebp"));
&movdqa ("xmm6",&QWP(16*6-128,"ebp"));
&movdqa ("xmm7",&QWP(16*7-128,"ebp"));
#&movdqa (&QWP(16*0-128,"ebx"),"xmm0");
&movdqa (&QWP(16*1-128,"ebx"),"xmm1");
&movdqa (&QWP(16*2-128,"ebx"),"xmm2");
&movdqa (&QWP(16*3-128,"ebx"),"xmm3");
#&movdqa (&QWP(16*4-128,"ebx"),"xmm4");
&movdqa (&QWP(16*5-128,"ebx"),"xmm5");
&movdqa (&QWP(16*6-128,"ebx"),"xmm6");
&movdqa (&QWP(16*7-128,"ebx"),"xmm7");
#&movdqa ("xmm0",&QWP(16*8-128,"ebp"));
#&movdqa ("xmm1",&QWP(16*9-128,"ebp"));
&movdqa ("xmm2",&QWP(16*10-128,"ebp"));
&movdqa ("xmm3",&QWP(16*11-128,"ebp"));
&movdqa ("xmm4",&QWP(16*12-128,"ebp"));
&movdqa ("xmm5",&QWP(16*13-128,"ebp"));
&movdqa ("xmm6",&QWP(16*14-128,"ebp"));
&movdqa ("xmm7",&QWP(16*15-128,"ebp"));
&paddd ("xmm4",&QWP(16*4,"eax")); # counter value
#&movdqa (&QWP(16*8-128,"ebx"),"xmm0");
#&movdqa (&QWP(16*9-128,"ebx"),"xmm1");
&movdqa (&QWP(16*10-128,"ebx"),"xmm2");
&movdqa (&QWP(16*11-128,"ebx"),"xmm3");
&movdqa (&QWP(16*12-128,"ebx"),"xmm4");
&movdqa (&QWP(16*13-128,"ebx"),"xmm5");
&movdqa (&QWP(16*14-128,"ebx"),"xmm6");
&movdqa (&QWP(16*15-128,"ebx"),"xmm7");
&movdqa (&QWP(16*12-128,"ebp"),"xmm4"); # save counter value
&movdqa ($xa, &QWP(16*0-128,"ebp"));
&movdqa ($xd, "xmm4");
&movdqa ($xb_,&QWP(16*4-128,"ebp"));
&movdqa ($xc, &QWP(16*8-128,"ebp"));
&movdqa ($xc_,&QWP(16*9-128,"ebp"));
&mov ("edx",10); # loop counter
&nop ();
&set_label("loop",16);
&paddd ($xa,$xb_); # elsewhere
&movdqa ($xb,$xb_);
&pxor ($xd,$xa); # elsewhere
&QUARTERROUND_SSSE3(0, 4, 8, 12, 0);
&QUARTERROUND_SSSE3(1, 5, 9, 13, 1);
&QUARTERROUND_SSSE3(2, 6,10, 14, 2);
&QUARTERROUND_SSSE3(3, 7,11, 15, 3);
&QUARTERROUND_SSSE3(0, 5,10, 15, 4);
&QUARTERROUND_SSSE3(1, 6,11, 12, 5);
&QUARTERROUND_SSSE3(2, 7, 8, 13, 6);
&QUARTERROUND_SSSE3(3, 4, 9, 14, 7);
&dec ("edx");
&jnz (&label("loop"));
&movdqa (&QWP(16*4-128,"ebx"),$xb_);
&movdqa (&QWP(16*8-128,"ebx"),$xc);
&movdqa (&QWP(16*9-128,"ebx"),$xc_);
&movdqa (&QWP(16*12-128,"ebx"),$xd);
&movdqa (&QWP(16*14-128,"ebx"),$xd_);
my ($xa0,$xa1,$xa2,$xa3,$xt0,$xt1,$xt2,$xt3)=map("xmm$_",(0..7));
#&movdqa ($xa0,&QWP(16*0-128,"ebx")); # it's there
&movdqa ($xa1,&QWP(16*1-128,"ebx"));
&movdqa ($xa2,&QWP(16*2-128,"ebx"));
&movdqa ($xa3,&QWP(16*3-128,"ebx"));
for($i=0;$i<256;$i+=64) {
&paddd ($xa0,&QWP($i+16*0-128,"ebp")); # accumulate key material
&paddd ($xa1,&QWP($i+16*1-128,"ebp"));
&paddd ($xa2,&QWP($i+16*2-128,"ebp"));
&paddd ($xa3,&QWP($i+16*3-128,"ebp"));
&movdqa ($xt2,$xa0); # "de-interlace" data
&punpckldq ($xa0,$xa1);
&movdqa ($xt3,$xa2);
&punpckldq ($xa2,$xa3);
&punpckhdq ($xt2,$xa1);
&punpckhdq ($xt3,$xa3);
&movdqa ($xa1,$xa0);
&punpcklqdq ($xa0,$xa2); # "a0"
&movdqa ($xa3,$xt2);
&punpcklqdq ($xt2,$xt3); # "a2"
&punpckhqdq ($xa1,$xa2); # "a1"
&punpckhqdq ($xa3,$xt3); # "a3"
#($xa2,$xt2)=($xt2,$xa2);
&movdqu ($xt0,&QWP(64*0-128,$inp)); # load input
&movdqu ($xt1,&QWP(64*1-128,$inp));
&movdqu ($xa2,&QWP(64*2-128,$inp));
&movdqu ($xt3,&QWP(64*3-128,$inp));
&lea ($inp,&QWP($i<192?16:(64*4-16*3),$inp));
&pxor ($xt0,$xa0);
&movdqa ($xa0,&QWP($i+16*4-128,"ebx")) if ($i<192);
&pxor ($xt1,$xa1);
&movdqa ($xa1,&QWP($i+16*5-128,"ebx")) if ($i<192);
&pxor ($xt2,$xa2);
&movdqa ($xa2,&QWP($i+16*6-128,"ebx")) if ($i<192);
&pxor ($xt3,$xa3);
&movdqa ($xa3,&QWP($i+16*7-128,"ebx")) if ($i<192);
&movdqu (&QWP(64*0-128,$out),$xt0); # store output
&movdqu (&QWP(64*1-128,$out),$xt1);
&movdqu (&QWP(64*2-128,$out),$xt2);
&movdqu (&QWP(64*3-128,$out),$xt3);
&lea ($out,&QWP($i<192?16:(64*4-16*3),$out));
}
&sub ($len,64*4);
&jnc (&label("outer_loop"));
&add ($len,64*4);
&jz (&label("done"));
&mov ("ebx",&DWP(512+8,"esp")); # restore pointers
&lea ($inp,&DWP(-128,$inp));
&mov ("edx",&DWP(512+4,"esp"));
&lea ($out,&DWP(-128,$out));
&movd ("xmm2",&DWP(16*12-128,"ebp")); # counter value
&movdqu ("xmm3",&QWP(0,"ebx"));
&paddd ("xmm2",&QWP(16*6,"eax")); # +four
&pand ("xmm3",&QWP(16*7,"eax"));
&por ("xmm3","xmm2"); # counter value
}
{
my ($a,$b,$c,$d,$t,$t1,$rot16,$rot24)=map("xmm$_",(0..7));
sub SSSE3ROUND { # critical path is 20 "SIMD ticks" per round
&paddd ($a,$b);
&pxor ($d,$a);
&pshufb ($d,$rot16);
&paddd ($c,$d);
&pxor ($b,$c);
&movdqa ($t,$b);
&psrld ($b,20);
&pslld ($t,12);
&por ($b,$t);
&paddd ($a,$b);
&pxor ($d,$a);
&pshufb ($d,$rot24);
&paddd ($c,$d);
&pxor ($b,$c);
&movdqa ($t,$b);
&psrld ($b,25);
&pslld ($t,7);
&por ($b,$t);
}
&set_label("1x");
&movdqa ($a,&QWP(16*2,"eax")); # sigma
&movdqu ($b,&QWP(0,"edx"));
&movdqu ($c,&QWP(16,"edx"));
#&movdqu ($d,&QWP(0,"ebx")); # already loaded
&movdqa ($rot16,&QWP(0,"eax"));
&movdqa ($rot24,&QWP(16,"eax"));
&mov (&DWP(16*3,"esp"),"ebp");
&movdqa (&QWP(16*0,"esp"),$a);
&movdqa (&QWP(16*1,"esp"),$b);
&movdqa (&QWP(16*2,"esp"),$c);
&movdqa (&QWP(16*3,"esp"),$d);
&mov ("edx",10);
&jmp (&label("loop1x"));
&set_label("outer1x",16);
&movdqa ($d,&QWP(16*5,"eax")); # one
&movdqa ($a,&QWP(16*0,"esp"));
&movdqa ($b,&QWP(16*1,"esp"));
&movdqa ($c,&QWP(16*2,"esp"));
&paddd ($d,&QWP(16*3,"esp"));
&mov ("edx",10);
&movdqa (&QWP(16*3,"esp"),$d);
&jmp (&label("loop1x"));
&set_label("loop1x",16);
&SSSE3ROUND();
&pshufd ($c,$c,0b01001110);
&pshufd ($b,$b,0b00111001);
&pshufd ($d,$d,0b10010011);
&nop ();
&SSSE3ROUND();
&pshufd ($c,$c,0b01001110);
&pshufd ($b,$b,0b10010011);
&pshufd ($d,$d,0b00111001);
&dec ("edx");
&jnz (&label("loop1x"));
&paddd ($a,&QWP(16*0,"esp"));
&paddd ($b,&QWP(16*1,"esp"));
&paddd ($c,&QWP(16*2,"esp"));
&paddd ($d,&QWP(16*3,"esp"));
&cmp ($len,64);
&jb (&label("tail"));
&movdqu ($t,&QWP(16*0,$inp));
&movdqu ($t1,&QWP(16*1,$inp));
&pxor ($a,$t); # xor with input
&movdqu ($t,&QWP(16*2,$inp));
&pxor ($b,$t1);
&movdqu ($t1,&QWP(16*3,$inp));
&pxor ($c,$t);
&pxor ($d,$t1);
&lea ($inp,&DWP(16*4,$inp)); # inp+=64
&movdqu (&QWP(16*0,$out),$a); # write output
&movdqu (&QWP(16*1,$out),$b);
&movdqu (&QWP(16*2,$out),$c);
&movdqu (&QWP(16*3,$out),$d);
&lea ($out,&DWP(16*4,$out)); # inp+=64
&sub ($len,64);
&jnz (&label("outer1x"));
&jmp (&label("done"));
&set_label("tail");
&movdqa (&QWP(16*0,"esp"),$a);
&movdqa (&QWP(16*1,"esp"),$b);
&movdqa (&QWP(16*2,"esp"),$c);
&movdqa (&QWP(16*3,"esp"),$d);
&xor ("eax","eax");
&xor ("edx","edx");
&xor ("ebp","ebp");
&set_label("tail_loop");
&movb ("al",&BP(0,"esp","ebp"));
&movb ("dl",&BP(0,$inp,"ebp"));
&lea ("ebp",&DWP(1,"ebp"));
&xor ("al","dl");
&movb (&BP(-1,$out,"ebp"),"al");
&dec ($len);
&jnz (&label("tail_loop"));
}
&set_label("done");
&mov ("esp",&DWP(512,"esp"));
&function_end("ChaCha20_ctr32_ssse3");
&align (64);
&set_label("ssse3_data");
&data_byte(0x2,0x3,0x0,0x1, 0x6,0x7,0x4,0x5, 0xa,0xb,0x8,0x9, 0xe,0xf,0xc,0xd);
&data_byte(0x3,0x0,0x1,0x2, 0x7,0x4,0x5,0x6, 0xb,0x8,0x9,0xa, 0xf,0xc,0xd,0xe);
&data_word(0x61707865,0x3320646e,0x79622d32,0x6b206574);
&data_word(0,1,2,3);
&data_word(4,4,4,4);
&data_word(1,0,0,0);
&data_word(4,0,0,0);
&data_word(0,-1,-1,-1);
&align (64);
}
&asciz ("ChaCha20 for x86, CRYPTOGAMS by <appro\@openssl.org>");
&asm_finish();
close STDOUT or die "error closing STDOUT: $!";

2762
vendor/aws-lc-sys/aws-lc/crypto/chacha/asm/chacha-x86_64.pl vendored Executable file
View File

File diff suppressed because it is too large Load Diff

236
vendor/aws-lc-sys/aws-lc/crypto/chacha/chacha.c vendored Normal file
View File

@@ -0,0 +1,236 @@
// Copyright (c) 2014, Google Inc.
// SPDX-License-Identifier: ISC
// Adapted from the public domain, estream code by D. Bernstein.
#include <openssl/chacha.h>
#include <assert.h>
#include <string.h>
#include "../internal.h"
#include "internal.h"
// sigma contains the ChaCha constants, which happen to be an ASCII string.
// "expand 32-byte k"
static const uint32_t sigma_words[4] = {
0x61707865, // 'e' 0x65, 'x' 0x78, 'p' 0x70, 'a' 0x61,
0x3320646e, // 'n' 0x6E, 'd' 0x64, ' ' 0x20, '3' 0x33,
0x79622d32, // '2' 0x32, '-' 0x2D, 'b' 0x62, 'y' 0x79,
0x6b206574 // 't' 0x74, 'e' 0x65, ' ' 0x20, 'k' 0x6B
};
// QUARTERROUND updates a, b, c, d with a ChaCha "quarter" round.
#define QUARTERROUND(a, b, c, d) \
x[a] += x[b]; \
x[d] = CRYPTO_rotl_u32(x[d] ^ x[a], 16); \
x[c] += x[d]; \
x[b] = CRYPTO_rotl_u32(x[b] ^ x[c], 12); \
x[a] += x[b]; \
x[d] = CRYPTO_rotl_u32(x[d] ^ x[a], 8); \
x[c] += x[d]; \
x[b] = CRYPTO_rotl_u32(x[b] ^ x[c], 7);
void CRYPTO_hchacha20(uint8_t out[32], const uint8_t key[32],
const uint8_t nonce[16]) {
uint32_t x[16];
OPENSSL_memcpy(x, sigma_words, sizeof(sigma_words));
#ifdef OPENSSL_BIG_ENDIAN
for(size_t i = 4; i < 12; i++) {
x[i] = CRYPTO_load_u32_le(key + (i-4) * sizeof(uint32_t));
}
for(size_t i = 12; i < 16; i++) {
x[i] = CRYPTO_load_u32_le(nonce + (i-12) * sizeof(uint32_t));
}
#else
OPENSSL_memcpy(&x[4], key, 32);
OPENSSL_memcpy(&x[12], nonce, 16);
#endif
for (size_t i = 0; i < 20; i += 2) {
QUARTERROUND(0, 4, 8, 12)
QUARTERROUND(1, 5, 9, 13)
QUARTERROUND(2, 6, 10, 14)
QUARTERROUND(3, 7, 11, 15)
QUARTERROUND(0, 5, 10, 15)
QUARTERROUND(1, 6, 11, 12)
QUARTERROUND(2, 7, 8, 13)
QUARTERROUND(3, 4, 9, 14)
}
#ifdef OPENSSL_BIG_ENDIAN
for(size_t i = 0; i < 4; i++) {
CRYPTO_store_u32_le(out + i * sizeof(uint32_t), x[i]);
}
for(size_t i = 12; i < 16; i++) {
CRYPTO_store_u32_le(out + (i-8) * sizeof(uint32_t), x[i]);
}
#else
OPENSSL_memcpy(out, &x[0], sizeof(uint32_t) * 4);
OPENSSL_memcpy(&out[16], &x[12], sizeof(uint32_t) * 4);
#endif
}
#if defined(CHACHA20_ASM_NOHW)
static void ChaCha20_ctr32(uint8_t *out, const uint8_t *in, size_t in_len,
const uint32_t key[8], const uint32_t counter[4]) {
#if defined(CHACHA20_ASM_NEON)
if (ChaCha20_ctr32_neon_capable(in_len)) {
ChaCha20_ctr32_neon(out, in, in_len, key, counter);
return;
}
#endif
#if defined(CHACHA20_ASM_AVX2) && !defined(MY_ASSEMBLER_IS_TOO_OLD_FOR_512AVX)
if (ChaCha20_ctr32_avx2_capable(in_len)) {
ChaCha20_ctr32_avx2(out, in, in_len, key, counter);
return;
}
#endif
#if defined(CHACHA20_ASM_SSSE3_4X)
if (ChaCha20_ctr32_ssse3_4x_capable(in_len)) {
ChaCha20_ctr32_ssse3_4x(out, in, in_len, key, counter);
return;
}
#endif
#if defined(CHACHA20_ASM_SSSE3)
if (ChaCha20_ctr32_ssse3_capable(in_len)) {
ChaCha20_ctr32_ssse3(out, in, in_len, key, counter);
return;
}
#endif
if (in_len > 0) {
ChaCha20_ctr32_nohw(out, in, in_len, key, counter);
}
}
#endif
#if defined(CHACHA20_ASM_NOHW)
void CRYPTO_chacha_20(uint8_t *out, const uint8_t *in, size_t in_len,
const uint8_t key[32], const uint8_t nonce[12],
uint32_t counter) {
assert(!buffers_alias(out, in_len, in, in_len) || in == out);
uint32_t counter_nonce[4];
counter_nonce[0] = counter;
counter_nonce[1] = CRYPTO_load_u32_le(nonce + 0);
counter_nonce[2] = CRYPTO_load_u32_le(nonce + 4);
counter_nonce[3] = CRYPTO_load_u32_le(nonce + 8);
const uint32_t *key_ptr = (const uint32_t *)key;
#if !defined(OPENSSL_X86) && !defined(OPENSSL_X86_64)
// The assembly expects the key to be four-byte aligned.
uint32_t key_u32[8];
if ((((uintptr_t)key) & 3) != 0) {
key_u32[0] = CRYPTO_load_u32_le(key + 0);
key_u32[1] = CRYPTO_load_u32_le(key + 4);
key_u32[2] = CRYPTO_load_u32_le(key + 8);
key_u32[3] = CRYPTO_load_u32_le(key + 12);
key_u32[4] = CRYPTO_load_u32_le(key + 16);
key_u32[5] = CRYPTO_load_u32_le(key + 20);
key_u32[6] = CRYPTO_load_u32_le(key + 24);
key_u32[7] = CRYPTO_load_u32_le(key + 28);
key_ptr = key_u32;
}
#endif
while (in_len > 0) {
// The assembly functions do not have defined overflow behavior. While
// overflow is almost always a bug in the caller, we prefer our functions to
// behave the same across platforms, so divide into multiple calls to avoid
// this case.
uint64_t todo = 64 * ((UINT64_C(1) << 32) - counter_nonce[0]);
if (todo > in_len) {
todo = in_len;
}
ChaCha20_ctr32(out, in, (size_t)todo, key_ptr, counter_nonce);
in += todo;
out += todo;
in_len -= todo;
// We're either done and will next break out of the loop, or we stopped at
// the wraparound point and the counter should continue at zero.
counter_nonce[0] = 0;
}
}
#else
// chacha_core performs 20 rounds of ChaCha on the input words in
// |input| and writes the 64 output bytes to |output|.
static void chacha_core(uint8_t output[64], const uint32_t input[16]) {
uint32_t x[16];
int i;
OPENSSL_memcpy(x, input, sizeof(uint32_t) * 16);
for (i = 20; i > 0; i -= 2) {
QUARTERROUND(0, 4, 8, 12)
QUARTERROUND(1, 5, 9, 13)
QUARTERROUND(2, 6, 10, 14)
QUARTERROUND(3, 7, 11, 15)
QUARTERROUND(0, 5, 10, 15)
QUARTERROUND(1, 6, 11, 12)
QUARTERROUND(2, 7, 8, 13)
QUARTERROUND(3, 4, 9, 14)
}
for (i = 0; i < 16; ++i) {
x[i] += input[i];
}
for (i = 0; i < 16; ++i) {
CRYPTO_store_u32_le(output + 4 * i, x[i]);
}
}
void CRYPTO_chacha_20(uint8_t *out, const uint8_t *in, size_t in_len,
const uint8_t key[32], const uint8_t nonce[12],
uint32_t counter) {
assert(!buffers_alias(out, in_len, in, in_len) || in == out);
uint32_t input[16];
uint8_t buf[64];
size_t todo, i;
input[0] = sigma_words[0];
input[1] = sigma_words[1];
input[2] = sigma_words[2];
input[3] = sigma_words[3];
input[4] = CRYPTO_load_u32_le(key + 0);
input[5] = CRYPTO_load_u32_le(key + 4);
input[6] = CRYPTO_load_u32_le(key + 8);
input[7] = CRYPTO_load_u32_le(key + 12);
input[8] = CRYPTO_load_u32_le(key + 16);
input[9] = CRYPTO_load_u32_le(key + 20);
input[10] = CRYPTO_load_u32_le(key + 24);
input[11] = CRYPTO_load_u32_le(key + 28);
input[12] = counter;
input[13] = CRYPTO_load_u32_le(nonce + 0);
input[14] = CRYPTO_load_u32_le(nonce + 4);
input[15] = CRYPTO_load_u32_le(nonce + 8);
while (in_len > 0) {
todo = sizeof(buf);
if (in_len < todo) {
todo = in_len;
}
chacha_core(buf, input);
for (i = 0; i < todo; i++) {
out[i] = in[i] ^ buf[i];
}
out += todo;
in += todo;
in_len -= todo;
input[12]++;
}
}
#endif

390
vendor/aws-lc-sys/aws-lc/crypto/chacha/chacha_test.cc vendored Normal file
View File

@@ -0,0 +1,390 @@
// Copyright (c) 2016, Google Inc.
// SPDX-License-Identifier: ISC
#include <stdio.h>
#include <stdint.h>
#include <string.h>
#include <memory>
#include <gtest/gtest.h>
#include <openssl/crypto.h>
#include <openssl/chacha.h>
#include "internal.h"
#include "../internal.h"
#include "../test/abi_test.h"
#include "../test/test_util.h"
static const uint8_t kKey[32] = {
0x98, 0xbe, 0xf1, 0x46, 0x9b, 0xe7, 0x26, 0x98, 0x37, 0xa4, 0x5b,
0xfb, 0xc9, 0x2a, 0x5a, 0x6a, 0xc7, 0x62, 0x50, 0x7c, 0xf9, 0x64,
0x43, 0xbf, 0x33, 0xb9, 0x6b, 0x1b, 0xd4, 0xc6, 0xf8, 0xf6,
};
static const uint8_t kNonce[12] = {
0x44, 0xe7, 0x92, 0xd6, 0x33, 0x35, 0xab, 0xb1, 0x58, 0x2e, 0x92, 0x53,
};
static uint32_t kCounter = 42;
static const uint8_t kInput[] = {
0x58, 0x28, 0xd5, 0x30, 0x36, 0x2c, 0x60, 0x55, 0x29, 0xf8, 0xe1, 0x8c,
0xae, 0x15, 0x15, 0x26, 0xf2, 0x3a, 0x73, 0xa0, 0xf3, 0x12, 0xa3, 0x88,
0x5f, 0x2b, 0x74, 0x23, 0x3d, 0xc9, 0x05, 0x23, 0xc6, 0x54, 0x49, 0x1e,
0x44, 0x88, 0x14, 0xd9, 0xda, 0x37, 0x15, 0xdc, 0xb7, 0xe4, 0x23, 0xb3,
0x9d, 0x7e, 0x16, 0x68, 0x35, 0xfc, 0x02, 0x6d, 0xcc, 0x8a, 0xe5, 0xdd,
0x5f, 0xe4, 0xd2, 0x56, 0x6f, 0x12, 0x9c, 0x9c, 0x7d, 0x6a, 0x38, 0x48,
0xbd, 0xdf, 0xd9, 0xac, 0x1b, 0xa2, 0x4d, 0xc5, 0x43, 0x04, 0x3c, 0xd7,
0x99, 0xe1, 0xa7, 0x13, 0x9c, 0x51, 0xc2, 0x6d, 0xf9, 0xcf, 0x07, 0x3b,
0xe4, 0xbf, 0x93, 0xa3, 0xa9, 0xb4, 0xc5, 0xf0, 0x1a, 0xe4, 0x8d, 0x5f,
0xc6, 0xc4, 0x7c, 0x69, 0x7a, 0xde, 0x1a, 0xc1, 0xc9, 0xcf, 0xc2, 0x4e,
0x7a, 0x25, 0x2c, 0x32, 0xe9, 0x17, 0xba, 0x68, 0xf1, 0x37, 0x5d, 0x62,
0x84, 0x46, 0xf5, 0x80, 0x7f, 0x1a, 0x71, 0xf7, 0xbe, 0x72, 0x4b, 0xb8,
0x1c, 0xfe, 0x3e, 0xbd, 0xae, 0x0d, 0x73, 0x0d, 0x87, 0x4a, 0x31, 0xc3,
0x3d, 0x46, 0x6f, 0xb3, 0xd7, 0x6b, 0xe3, 0xb8, 0x70, 0x17, 0x8e, 0x7a,
0x6a, 0x0e, 0xbf, 0xa8, 0xbc, 0x2b, 0xdb, 0xfa, 0x4f, 0xb6, 0x26, 0x20,
0xee, 0x63, 0xf0, 0x6d, 0x26, 0xac, 0x6a, 0x18, 0x37, 0x6e, 0x59, 0x81,
0xd1, 0x60, 0xe6, 0x40, 0xd5, 0x6d, 0x68, 0xba, 0x8b, 0x65, 0x4a, 0xf9,
0xf1, 0xae, 0x56, 0x24, 0x8f, 0xe3, 0x8e, 0xe7, 0x7e, 0x6f, 0xcf, 0x92,
0xdf, 0xa9, 0x75, 0x3a, 0xd6, 0x2e, 0x1c, 0xaf, 0xf2, 0xd6, 0x8b, 0x39,
0xad, 0xd2, 0x5d, 0xfb, 0xd7, 0xdf, 0x05, 0x57, 0x0d, 0xf7, 0xf6, 0x8f,
0x2d, 0x14, 0xb0, 0x4e, 0x1a, 0x3c, 0x77, 0x04, 0xcd, 0x3c, 0x5c, 0x58,
0x52, 0x10, 0x6f, 0xcf, 0x5c, 0x03, 0xc8, 0x5f, 0x85, 0x2b, 0x05, 0x82,
0x60, 0xda, 0xcc, 0xcd, 0xd6, 0x88, 0xbf, 0xc0, 0x10, 0xb3, 0x6f, 0x54,
0x54, 0x42, 0xbc, 0x4b, 0x77, 0x21, 0x4d, 0xee, 0x87, 0x45, 0x06, 0x4c,
0x60, 0x38, 0xd2, 0x7e, 0x1d, 0x30, 0x6c, 0x55, 0xf0, 0x38, 0x80, 0x1c,
0xde, 0x3d, 0xea, 0x68, 0x3e, 0xf6, 0x3e, 0x59, 0xcf, 0x0d, 0x08, 0xae,
0x8c, 0x02, 0x0b, 0xc1, 0x72, 0x6a, 0xb4, 0x6d, 0xf3, 0xf7, 0xb3, 0xef,
0x3a, 0xb1, 0x06, 0xf2, 0xf4, 0xd6, 0x69, 0x7b, 0x3e, 0xa2, 0x16, 0x31,
0x31, 0x79, 0xb6, 0x33, 0xa9, 0xca, 0x8a, 0xa8, 0xbe, 0xf3, 0xe9, 0x38,
0x28, 0xd1, 0xe1, 0x3b, 0x4e, 0x2e, 0x47, 0x35, 0xa4, 0x61, 0x14, 0x1e,
0x42, 0x2c, 0x49, 0x55, 0xea, 0xe3, 0xb3, 0xce, 0x39, 0xd3, 0xb3, 0xef,
0x4a, 0x4d, 0x78, 0x49, 0xbd, 0xf6, 0x7c, 0x0a, 0x2c, 0xd3, 0x26, 0xcb,
0xd9, 0x6a, 0xad, 0x63, 0x93, 0xa7, 0x29, 0x92, 0xdc, 0x1f, 0xaf, 0x61,
0x82, 0x80, 0x74, 0xb2, 0x9c, 0x4a, 0x86, 0x73, 0x50, 0xd8, 0xd1, 0xff,
0xee, 0x1a, 0xe2, 0xdd, 0xa2, 0x61, 0xbd, 0x10, 0xc3, 0x5f, 0x67, 0x9f,
0x29, 0xe4, 0xd3, 0x70, 0xe5, 0x67, 0x3a, 0xd2, 0x20, 0x00, 0xcc, 0x25,
0x15, 0x96, 0x54, 0x45, 0x85, 0xed, 0x82, 0x88, 0x3b, 0x9f, 0x3b, 0xc3,
0x04, 0xd4, 0x23, 0xb1, 0x0d, 0xdc, 0xc8, 0x26, 0x9d, 0x28, 0xb3, 0x25,
0x4d, 0x52, 0xe5, 0x33, 0xf3, 0xed, 0x2c, 0xb8, 0x1a, 0xcf, 0xc3, 0x52,
0xb4, 0x2f, 0xc7, 0x79, 0x96, 0x14, 0x7d, 0x72, 0x27, 0x72, 0x85, 0xea,
0x6d, 0x41, 0xa0, 0x22, 0x13, 0x6d, 0x06, 0x83, 0xa4, 0xdd, 0x0f, 0x69,
0xd2, 0x01, 0xcd, 0xc6, 0xb8, 0x64, 0x5c, 0x2c, 0x79, 0xd1, 0xc7, 0xd3,
0x31, 0xdb, 0x2c, 0xff, 0xda, 0xd0, 0x69, 0x31, 0xad, 0x83, 0x5f, 0xed,
0x6a, 0x97, 0xe4, 0x00, 0x43, 0xb0, 0x2e, 0x97, 0xae, 0x00, 0x5f, 0x5c,
0xb9, 0xe8, 0x39, 0x80, 0x10, 0xca, 0x0c, 0xfa, 0xf0, 0xb5, 0xcd, 0xaa,
0x27, 0x11, 0x60, 0xd9, 0x21, 0x86, 0x93, 0x91, 0x9f, 0x2d, 0x1a, 0x8e,
0xde, 0x0b, 0xb5, 0xcb, 0x05, 0x24, 0x30, 0x45, 0x4d, 0x11, 0x75, 0xfd,
0xe5, 0xa0, 0xa9, 0x4e, 0x3a, 0x8c, 0x3b, 0x52, 0x5a, 0x37, 0x18, 0x05,
0x4a, 0x7a, 0x09, 0x6a, 0xe6, 0xd5, 0xa9, 0xa6, 0x71, 0x47, 0x4c, 0x50,
0xe1, 0x3e, 0x8a, 0x21, 0x2b, 0x4f, 0x0e, 0xe3, 0xcb, 0x72, 0xc5, 0x28,
0x3e, 0x5a, 0x33, 0xec, 0x48, 0x92, 0x2e, 0xa1, 0x24, 0x57, 0x09, 0x0f,
0x01, 0x85, 0x3b, 0x34, 0x39, 0x7e, 0xc7, 0x90, 0x62, 0xe2, 0xdc, 0x5d,
0x0a, 0x2c, 0x51, 0x26, 0x95, 0x3a, 0x95, 0x92, 0xa5, 0x39, 0x8f, 0x0c,
0x83, 0x0b, 0x9d, 0x38, 0xab, 0x98, 0x2a, 0xc4, 0x01, 0xc4, 0x0d, 0x77,
0x13, 0xcb, 0xca, 0xf1, 0x28, 0x31, 0x52, 0x75, 0x27, 0x2c, 0xf0, 0x04,
0x86, 0xc8, 0xf3, 0x3d, 0xf2, 0x9d, 0x8f, 0x55, 0x52, 0x40, 0x3f, 0xaa,
0x22, 0x7f, 0xe7, 0x69, 0x3b, 0xee, 0x44, 0x09, 0xde, 0xff, 0xb0, 0x69,
0x3a, 0xae, 0x74, 0xe9, 0x9d, 0x33, 0xae, 0x8b, 0x6d, 0x60, 0x04, 0xff,
0x53, 0x3f, 0x88, 0xe9, 0x63, 0x9b, 0xb1, 0x6d, 0x2c, 0x22, 0x15, 0x5a,
0x15, 0xd9, 0xe5, 0xcb, 0x03, 0x78, 0x3c, 0xca, 0x59, 0x8c, 0xc8, 0xc2,
0x86, 0xff, 0xd2, 0x79, 0xd6, 0xc6, 0xec, 0x5b, 0xbb, 0xa0, 0xae, 0x01,
0x20, 0x09, 0x2e, 0x38, 0x5d, 0xda, 0x5d, 0xe0, 0x59, 0x4e, 0xe5, 0x8b,
0x84, 0x8f, 0xb6, 0xe0, 0x56, 0x9f, 0x21, 0xa1, 0xcf, 0xb2, 0x0f, 0x2c,
0x93, 0xf8, 0xcf, 0x37, 0xc1, 0x9f, 0x32, 0x98, 0x21, 0x65, 0x52, 0x66,
0x6e, 0xd3, 0x71, 0x98, 0x55, 0xb9, 0x46, 0x9f, 0x1a, 0x35, 0xc4, 0x47,
0x69, 0x62, 0x70, 0x4b, 0x77, 0x9e, 0xe4, 0x21, 0xe6, 0x32, 0x5a, 0x26,
0x05, 0xba, 0x57, 0x53, 0xd7, 0x9b, 0x55, 0x3c, 0xbb, 0x53, 0x79, 0x60,
0x9c, 0xc8, 0x4d, 0xf7, 0xf5, 0x1d, 0x54, 0x02, 0x91, 0x68, 0x0e, 0xaa,
0xca, 0x5a, 0x78, 0x0c, 0x28, 0x9a, 0xc3, 0xac, 0x49, 0xc0, 0xf4, 0x85,
0xee, 0x59, 0x76, 0x7e, 0x28, 0x4e, 0xf1, 0x5c, 0x63, 0xf7, 0xce, 0x0e,
0x2c, 0x21, 0xa0, 0x58, 0xe9, 0x01, 0xfd, 0xeb, 0xd1, 0xaf, 0xe6, 0xef,
0x93, 0xb3, 0x95, 0x51, 0x60, 0xa2, 0x74, 0x40, 0x15, 0xe5, 0xf4, 0x0a,
0xca, 0x6d, 0x9a, 0x37, 0x42, 0x4d, 0x5a, 0x58, 0x49, 0x0f, 0xe9, 0x02,
0xfc, 0x77, 0xd8, 0x59, 0xde, 0xdd, 0xad, 0x4b, 0x99, 0x2e, 0x64, 0x73,
0xad, 0x42, 0x2f, 0xf3, 0x2c, 0x0d, 0x49, 0xe4, 0x2e, 0x6c, 0xa4, 0x73,
0x75, 0x18, 0x14, 0x85, 0xbb, 0x64, 0xb4, 0xa1, 0xb0, 0x6e, 0x01, 0xc0,
0xcf, 0x17, 0x9c, 0xc5, 0x28, 0xc3, 0x2d, 0x6c, 0x17, 0x2a, 0x3d, 0x06,
0x5c, 0xf3, 0xb4, 0x49, 0x75, 0xad, 0x17, 0x69, 0xd4, 0xca, 0x65, 0xae,
0x44, 0x71, 0xa5, 0xf6, 0x0d, 0x0f, 0x8e, 0x37, 0xc7, 0x43, 0xce, 0x6b,
0x08, 0xe9, 0xd1, 0x34, 0x48, 0x8f, 0xc9, 0xfc, 0xf3, 0x5d, 0x2d, 0xec,
0x62, 0xd3, 0xf0, 0xb3, 0xfe, 0x2e, 0x40, 0x55, 0x76, 0x54, 0xc7, 0xb4,
0x61, 0x16, 0xcc, 0x7c, 0x1c, 0x19, 0x24, 0xe6, 0x4d, 0xd4, 0xc3, 0x77,
0x67, 0x1f, 0x3c, 0x74, 0x79, 0xa1, 0xf8, 0x85, 0x88, 0x1d, 0x6f, 0xa4,
0x7e, 0x2c, 0x21, 0x9f, 0x49, 0xf5, 0xaa, 0x4e, 0xf3, 0x4a, 0xfa, 0x9d,
0xbe, 0xf6, 0xce, 0xda, 0xb5, 0xab, 0x39, 0xbd, 0x16, 0x41, 0xa9, 0x4a,
0xac, 0x09, 0x01, 0xca,
};
static const uint8_t kOutput[] = {
0x54, 0x30, 0x6a, 0x13, 0xda, 0x59, 0x6b, 0x6d, 0x59, 0x49, 0xc8, 0xc5,
0xab, 0x26, 0xd4, 0x8a, 0xad, 0xc0, 0x3d, 0xaf, 0x14, 0xb9, 0x15, 0xb8,
0xca, 0xdf, 0x17, 0xa7, 0x03, 0xd3, 0xc5, 0x06, 0x01, 0xef, 0x21, 0xdd,
0xa3, 0x0b, 0x9e, 0x48, 0xb8, 0x5e, 0x0b, 0x87, 0x9f, 0x95, 0x23, 0x68,
0x85, 0x69, 0xd2, 0x5d, 0xaf, 0x57, 0xe9, 0x27, 0x11, 0x3d, 0x49, 0xfa,
0xf1, 0x08, 0xcc, 0x15, 0xec, 0x1d, 0x19, 0x16, 0x12, 0x9b, 0xc8, 0x66,
0x1f, 0xfa, 0x2c, 0x93, 0xf4, 0x99, 0x11, 0x27, 0x31, 0x0e, 0xd8, 0x46,
0x47, 0x40, 0x11, 0x70, 0x01, 0xca, 0xe8, 0x5b, 0xc5, 0x91, 0xc8, 0x3a,
0xdc, 0xaa, 0xf3, 0x4b, 0x80, 0xe5, 0xbc, 0x03, 0xd0, 0x89, 0x72, 0xbc,
0xce, 0x2a, 0x76, 0x0c, 0xf5, 0xda, 0x4c, 0x10, 0x06, 0x35, 0x41, 0xb1,
0xe6, 0xb4, 0xaa, 0x7a, 0xef, 0xf0, 0x62, 0x4a, 0xc5, 0x9f, 0x2c, 0xaf,
0xb8, 0x2f, 0xd9, 0xd1, 0x01, 0x7a, 0x36, 0x2f, 0x3e, 0x83, 0xa5, 0xeb,
0x81, 0x70, 0xa0, 0x57, 0x17, 0x46, 0xea, 0x9e, 0xcb, 0x0e, 0x74, 0xd3,
0x44, 0x57, 0x1d, 0x40, 0x06, 0xf8, 0xb7, 0xcb, 0x5f, 0xf4, 0x79, 0xbd,
0x11, 0x19, 0xd6, 0xee, 0xf8, 0xb0, 0xaa, 0xdd, 0x00, 0x62, 0xad, 0x3b,
0x88, 0x9a, 0x88, 0x5b, 0x1b, 0x07, 0xc9, 0xae, 0x9e, 0xa6, 0x94, 0xe5,
0x55, 0xdb, 0x45, 0x23, 0xb9, 0x2c, 0xcd, 0x29, 0xd3, 0x54, 0xc3, 0x88,
0x1e, 0x5f, 0x52, 0xf2, 0x09, 0x00, 0x26, 0x26, 0x1a, 0xed, 0xf5, 0xc2,
0xa9, 0x7d, 0xf9, 0x21, 0x5a, 0xaf, 0x6d, 0xab, 0x8e, 0x16, 0x84, 0x96,
0xb5, 0x4f, 0xcf, 0x1e, 0xa3, 0xaf, 0x08, 0x9f, 0x79, 0x86, 0xc3, 0xbe,
0x0c, 0x70, 0xcb, 0x8f, 0xf3, 0xc5, 0xf8, 0xe8, 0x4b, 0x21, 0x7d, 0x18,
0xa9, 0xed, 0x8b, 0xfb, 0x6b, 0x5a, 0x6f, 0x26, 0x0b, 0x56, 0x04, 0x7c,
0xfe, 0x0e, 0x1e, 0xc1, 0x3f, 0x82, 0xc5, 0x73, 0xbd, 0x53, 0x0c, 0xf0,
0xe2, 0xc9, 0xf3, 0x3d, 0x1b, 0x6d, 0xba, 0x70, 0xc1, 0x6d, 0xb6, 0x00,
0x28, 0xe1, 0xc4, 0x78, 0x62, 0x04, 0xda, 0x23, 0x86, 0xc3, 0xda, 0x74,
0x3d, 0x7c, 0xd6, 0x76, 0x29, 0xb2, 0x27, 0x2e, 0xb2, 0x35, 0x42, 0x60,
0x82, 0xcf, 0x30, 0x2c, 0x59, 0xe4, 0xe3, 0xd0, 0x74, 0x1f, 0x58, 0xe8,
0xda, 0x47, 0x45, 0x73, 0x1c, 0x05, 0x93, 0xae, 0x75, 0xbe, 0x1f, 0x81,
0xd8, 0xb7, 0xb3, 0xff, 0xfc, 0x8b, 0x52, 0x9e, 0xed, 0x8b, 0x37, 0x9f,
0xe0, 0xb8, 0xa2, 0x66, 0xe1, 0x6a, 0xc5, 0x1f, 0x1d, 0xf0, 0xde, 0x3f,
0x3d, 0xb0, 0x28, 0xf3, 0xaa, 0x4e, 0x4d, 0x31, 0xb0, 0x26, 0x79, 0x2b,
0x08, 0x0f, 0xe9, 0x2f, 0x79, 0xb3, 0xc8, 0xdd, 0xa7, 0x89, 0xa8, 0xa8,
0x1d, 0x59, 0x0e, 0x4f, 0x1e, 0x93, 0x1f, 0x70, 0x7f, 0x4e, 0x7e, 0xfe,
0xb8, 0xca, 0x63, 0xe0, 0xa6, 0x05, 0xcc, 0xd7, 0xde, 0x2a, 0x49, 0x31,
0x78, 0x5c, 0x5f, 0x44, 0xb2, 0x9b, 0x91, 0x99, 0x14, 0x29, 0x63, 0x09,
0x12, 0xdd, 0x02, 0xd9, 0x7b, 0xe9, 0xf5, 0x12, 0x07, 0xd0, 0xe7, 0xe6,
0xe8, 0xdd, 0xda, 0xa4, 0x73, 0xc4, 0x8e, 0xbd, 0x7b, 0xb7, 0xbb, 0xcb,
0x83, 0x2f, 0x43, 0xf6, 0x1c, 0x50, 0xae, 0x9b, 0x2e, 0x52, 0x80, 0x18,
0x85, 0xa8, 0x23, 0x52, 0x7a, 0x6a, 0xf7, 0x42, 0x36, 0xca, 0x91, 0x5a,
0x3d, 0x2a, 0xa0, 0x35, 0x7d, 0x70, 0xfc, 0x4c, 0x18, 0x7c, 0x57, 0x72,
0xcf, 0x9b, 0x29, 0xd6, 0xd0, 0xb4, 0xd7, 0xe6, 0x89, 0x70, 0x69, 0x22,
0x5e, 0x45, 0x09, 0x4d, 0x49, 0x87, 0x84, 0x5f, 0x8a, 0x5f, 0xe4, 0x15,
0xd3, 0xe3, 0x72, 0xaf, 0xb2, 0x30, 0x9c, 0xc1, 0xff, 0x8e, 0x6d, 0x2a,
0x76, 0x9e, 0x08, 0x03, 0x7e, 0xe0, 0xc3, 0xc2, 0x97, 0x06, 0x6b, 0x33,
0x2b, 0x08, 0xe3, 0xd5, 0x0b, 0xd8, 0x32, 0x67, 0x61, 0x10, 0xed, 0x6b,
0xed, 0x50, 0xef, 0xd7, 0x1c, 0x1b, 0xe0, 0x6d, 0xa1, 0x64, 0x19, 0x34,
0x2f, 0xe4, 0xe8, 0x54, 0xbf, 0x84, 0x0e, 0xdf, 0x0e, 0x8b, 0xd8, 0xdd,
0x77, 0x96, 0xb8, 0x54, 0xab, 0xf2, 0x95, 0x59, 0x0d, 0x0d, 0x0a, 0x15,
0x6e, 0x01, 0xf2, 0x24, 0xab, 0xa0, 0xd8, 0xdf, 0x38, 0xea, 0x97, 0x58,
0x76, 0x88, 0xbe, 0xaf, 0x45, 0xe3, 0x56, 0x4f, 0x68, 0xe8, 0x4b, 0xe7,
0x2b, 0x22, 0x18, 0x96, 0x82, 0x89, 0x25, 0x34, 0xd1, 0xdd, 0x08, 0xea,
0x7e, 0x21, 0xef, 0x57, 0x55, 0x43, 0xf7, 0xfa, 0xca, 0x1c, 0xde, 0x99,
0x2e, 0x8b, 0xd8, 0xc3, 0xcf, 0x89, 0x4d, 0xfc, 0x3b, 0x7d, 0x4a, 0xc9,
0x99, 0xc4, 0x31, 0xb6, 0x7a, 0xae, 0xf8, 0x49, 0xb2, 0x46, 0xc1, 0x60,
0x05, 0x75, 0xf3, 0x3d, 0xf2, 0xc9, 0x84, 0xa4, 0xb9, 0x8a, 0x87, 0x2a,
0x87, 0x5c, 0x0a, 0xbc, 0x51, 0x7d, 0x9a, 0xf5, 0xc9, 0x24, 0x2d, 0x5e,
0xe6, 0xc6, 0xe3, 0xcd, 0x7e, 0xe4, 0xaf, 0x8a, 0x6c, 0x00, 0x04, 0xc8,
0xd7, 0xa5, 0xad, 0xfa, 0xb2, 0x08, 0x4a, 0x26, 0x9b, 0x7c, 0xd0, 0xc6,
0x13, 0xb1, 0xb9, 0x65, 0x3f, 0x70, 0x30, 0xf9, 0x98, 0x9d, 0x87, 0x99,
0x57, 0x71, 0x3e, 0xb1, 0xc3, 0x24, 0xf0, 0xa6, 0xa2, 0x60, 0x9d, 0x66,
0xd2, 0x5f, 0xae, 0xe3, 0x94, 0x87, 0xea, 0xd1, 0xea, 0x0d, 0x2a, 0x77,
0xef, 0x31, 0xcc, 0xeb, 0xf9, 0x0c, 0xdc, 0x9c, 0x12, 0x80, 0xbb, 0xb0,
0x8e, 0xab, 0x9a, 0x04, 0xcd, 0x4b, 0x95, 0x4f, 0x7a, 0x0b, 0x53, 0x7c,
0x16, 0xcc, 0x0e, 0xb1, 0x73, 0x10, 0xdd, 0xaa, 0x76, 0x94, 0x90, 0xd9,
0x8b, 0x66, 0x41, 0x31, 0xed, 0x8c, 0x7d, 0x74, 0xc4, 0x33, 0xfa, 0xc3,
0x43, 0x8d, 0x10, 0xbc, 0x84, 0x4d, 0x0e, 0x95, 0x32, 0xdf, 0x17, 0x43,
0x6d, 0xd2, 0x5e, 0x12, 0xb9, 0xed, 0x33, 0xd9, 0x97, 0x6f, 0x4a, 0xcd,
0xc3, 0xcd, 0x81, 0x34, 0xbe, 0x7e, 0xa2, 0xd0, 0xa7, 0x91, 0x5d, 0x90,
0xf6, 0x5e, 0x4a, 0x25, 0x0f, 0xcc, 0x24, 0xeb, 0xe1, 0xe4, 0x62, 0x6c,
0x8f, 0x45, 0x36, 0x97, 0x5d, 0xda, 0x20, 0x2b, 0x86, 0x00, 0x8c, 0x94,
0xa9, 0x6a, 0x69, 0xb2, 0xe9, 0xbb, 0x82, 0x8e, 0x41, 0x95, 0xb4, 0xb7,
0xf1, 0x55, 0x52, 0x30, 0x39, 0x48, 0xb3, 0x25, 0x82, 0xa9, 0x10, 0x27,
0x89, 0xb5, 0xe5, 0x1f, 0xab, 0x72, 0x3c, 0x70, 0x08, 0xce, 0xe6, 0x61,
0xbf, 0x19, 0xc8, 0x90, 0x2b, 0x29, 0x30, 0x3e, 0xb8, 0x4c, 0x33, 0xf0,
0xf0, 0x15, 0x2e, 0xb7, 0x25, 0xca, 0x99, 0x4b, 0x6f, 0x4b, 0x41, 0x50,
0xee, 0x56, 0x99, 0xcf, 0x2b, 0xa4, 0xc4, 0x7c, 0x5c, 0xa6, 0xd4, 0x67,
0x04, 0x5c, 0x5d, 0x5f, 0x26, 0x9e, 0x0f, 0xe2, 0x58, 0x68, 0x4c, 0x30,
0xcd, 0xef, 0x46, 0xdb, 0x37, 0x6f, 0xbb, 0xc4, 0x80, 0xca, 0x8a, 0x54,
0x5d, 0x71, 0x9d, 0x0c, 0xe8, 0xb8, 0x2c, 0x10, 0x90, 0x44, 0xa4, 0x88,
0x3f, 0xbc, 0x15, 0x3c, 0xd2, 0xca, 0x0e, 0xc3, 0xe4, 0x6e, 0xef, 0xb0,
0xcb, 0xfd, 0x61, 0x7c, 0x27, 0xf2, 0x25, 0xea, 0x71, 0x6d, 0xf7, 0x49,
0x9c, 0x81, 0x27, 0xf0, 0x61, 0x33, 0xcf, 0x55, 0x68, 0xd3, 0x73, 0xa4,
0xed, 0x35, 0x65, 0x2a, 0xf2, 0x3e, 0xcf, 0x90, 0x98, 0x54, 0x6d, 0x95,
0x6a, 0x0c, 0x9c, 0x24, 0x0e, 0xb4, 0xb7, 0x9b, 0x8d, 0x6e, 0x1c, 0xbc,
0xeb, 0x17, 0x10, 0x86, 0xda, 0x91, 0x6d, 0x89, 0x4c, 0xeb, 0xf5, 0x50,
0x8f, 0x40, 0xcf, 0x4a,
};
static uint32_t kOverflowCounter = 0xffffffff;
static const uint8_t kOverflowOutput[] = {
0x37, 0x64, 0x38, 0xcb, 0x25, 0x69, 0x2c, 0xf5, 0x88, 0x8a, 0xfe, 0x6d,
0x3b, 0x10, 0x07, 0x3c, 0x77, 0xac, 0xcd, 0x1c, 0x0c, 0xa7, 0x17, 0x31,
0x1d, 0xc3, 0x81, 0xd1, 0xa5, 0x20, 0x55, 0xea, 0xd3, 0x00, 0xc9, 0x84,
0xde, 0xe2, 0xe5, 0x5e, 0x7b, 0x28, 0x28, 0x59, 0x73, 0x3a, 0x8e, 0x57,
0x62, 0x18, 0x50, 0x55, 0x97, 0xca, 0x50, 0x3e, 0x8a, 0x84, 0x61, 0x28,
0x4c, 0x22, 0x93, 0x50, 0x48, 0x7e, 0x65, 0x78, 0x06, 0x5a, 0xcd, 0x2b,
0x11, 0xf7, 0x10, 0xfd, 0x6f, 0x41, 0x92, 0x82, 0x7c, 0x3a, 0x71, 0x07,
0x67, 0xd0, 0x7e, 0xb7, 0xdf, 0xdc, 0xfc, 0xee, 0xe6, 0x55, 0xdd, 0x6f,
0x79, 0x23, 0xf3, 0xae, 0xb1, 0x21, 0x96, 0xbe, 0xea, 0x0e, 0x1b, 0x58,
0x0b, 0x3f, 0x63, 0x51, 0xd4, 0xce, 0x98, 0xfe, 0x1a, 0xc7, 0xa7, 0x43,
0x7f, 0x0c, 0xe8, 0x62, 0xcf, 0x78, 0x3f, 0x4e, 0x31, 0xbf, 0x2b, 0x76,
0x91, 0xcd, 0x19, 0x80, 0x0d, 0x7f, 0x11, 0x8b, 0x76, 0xef, 0x43, 0x3c,
0x4f, 0x61, 0x86, 0xc5, 0x64, 0xa8, 0xc2, 0x73, 0xc2, 0x64, 0x39, 0xa0,
0x8b, 0xe6, 0x7f, 0xf6, 0x26, 0xd4, 0x47, 0x4f, 0xe4, 0x46, 0xe2, 0xf5,
0x9e, 0xe6, 0xc7, 0x76, 0x6c, 0xa9, 0x0f, 0x1d, 0x1b, 0x22, 0xa5, 0x62,
0x0a, 0x88, 0x3e, 0x8c, 0xf0, 0xbc, 0x4c, 0x11, 0x3f, 0x0d, 0xf7, 0x85,
0x67, 0x0b, 0x4c, 0xa3, 0x3f, 0xa8, 0xf1, 0x2a, 0x65, 0x2e, 0x00, 0x03,
0xc9, 0x49, 0x91, 0x48, 0xb7, 0xc8, 0x29, 0x28, 0x2f, 0x46, 0x8e, 0x8b,
0xd6, 0x73, 0x19, 0x06, 0x3e, 0x6f, 0x92, 0xc8, 0x3d, 0x3f, 0x4d, 0x68,
0xbc, 0x02, 0xc0, 0x8f, 0x71, 0x46, 0x0d, 0x28, 0x63, 0xfe, 0xad, 0x14,
0x81, 0x04, 0xb7, 0x23, 0xfd, 0x21, 0x0a, 0xf0, 0x6f, 0xcd, 0x47, 0x0b,
0x0e, 0x93, 0xa3, 0xa8, 0x44, 0x15, 0xd6, 0xae, 0x06, 0x44, 0x6b, 0xbc,
0xff, 0x8a, 0x56, 0x60, 0x3c, 0x38, 0xd6, 0xed, 0x03, 0x2d, 0x79, 0x2a,
0xe9, 0x15, 0xef, 0xfc, 0x92, 0x1f, 0x83, 0xa4, 0x60, 0x8f, 0xc9, 0x29,
0xb2, 0xb4, 0x9e, 0x3f, 0xa9, 0xe8, 0xfb, 0xa2, 0x62, 0x20, 0x2e, 0xc9,
0x43, 0xb2, 0xd1, 0x36, 0x85, 0x1e, 0xa4, 0xb3, 0x4f, 0x8c, 0x9e, 0x81,
0x75, 0x68, 0xbc, 0xf1, 0x52, 0xd5, 0x03, 0x22, 0xcf, 0xdf, 0x64, 0xb0,
0x28, 0xd2, 0x45, 0x18, 0x38, 0x8c, 0xd0, 0xf6, 0x30, 0x3c, 0x04, 0xd9,
0x8d, 0xb6, 0xb2, 0x57, 0x2a, 0xee, 0x28, 0xeb, 0x5f, 0x1a, 0x10, 0x6e,
0x88, 0x79, 0x08, 0x23, 0x19, 0x84, 0xf8, 0x80, 0x1a, 0x7d, 0x6f, 0x8b,
0xc1, 0x8e, 0x5f, 0x5f, 0x54, 0x14, 0x2a, 0xdc, 0x41, 0x5d, 0xeb, 0x00,
0xf2, 0x50, 0xae, 0xd3, 0x55, 0x32, 0xf6, 0xd9, 0x34, 0xf4, 0xb2, 0xf2,
0xf5, 0x90, 0x05, 0x8a, 0x9c, 0xc7, 0x94, 0x5d, 0x2d, 0x5a, 0x0f, 0xdd,
0x03, 0xde, 0xbe, 0x18, 0xb3, 0xe3, 0x07, 0x6b, 0x57, 0xfa, 0x1b, 0x7b,
0x75, 0xcb, 0xc2, 0x4d, 0xf7, 0x88, 0xfe, 0xf9, 0xc0, 0x6c, 0xdb, 0x5f,
0xf6, 0x48, 0x00, 0x4a, 0x5d, 0x75, 0xfa, 0x6b, 0x45, 0x43, 0xc4, 0x7f,
0x97, 0x31, 0x22, 0xb4, 0x9c, 0xa3, 0xee, 0x2f, 0x27, 0xa9, 0x9f, 0x0e,
0xdc, 0x40, 0x67, 0x17, 0x2e, 0xcb, 0xfd, 0x9e, 0xe7, 0xb2, 0x85, 0xcd,
0x49, 0x24, 0xc8, 0x8a, 0x59, 0x6b, 0x1f, 0xec, 0x72, 0x89, 0xf8, 0x30,
0xdf, 0x82, 0x61, 0x3b, 0x8b, 0xc9, 0x80, 0xe4, 0x27, 0x0d, 0xfe, 0x42,
0x27, 0x6c, 0xaf, 0x62, 0x3e, 0x2f, 0x1d, 0x38, 0xb6, 0x88, 0x8f, 0x71,
0x5a, 0x54, 0x6c, 0x68, 0x57, 0x40, 0x49, 0x7a, 0xb2, 0xe8, 0xb6, 0x97,
0xab, 0xd6, 0x3c, 0x35, 0xf3, 0x95, 0x12, 0xde, 0xa2, 0x39, 0x54, 0x52,
0x8c, 0x38, 0x2a, 0x2b, 0xe7, 0x21, 0x38, 0x63, 0xb0, 0xd6, 0xad, 0x94,
0x44, 0xaf, 0x49, 0x5d, 0xfc, 0x49, 0x6b, 0x30, 0xdf, 0xe9, 0x19, 0x1e,
0xed, 0x98, 0x0d, 0x4a, 0x3d, 0x56, 0x5e, 0x74, 0xad, 0x13, 0x8b, 0x68,
0x45, 0x08, 0xbe, 0x0e, 0x6c, 0xb4, 0x62, 0x93, 0x27, 0x8b, 0x4f, 0xab,
0x3e, 0xba, 0xe1, 0xe5, 0xff, 0xa8, 0x5d, 0x33, 0x32, 0xff, 0x34, 0xf9,
0x8d, 0x67, 0x24, 0x4a, 0xbb, 0x2c, 0x60, 0xb5, 0x88, 0x96, 0x1b, 0xcc,
0x53, 0xfb, 0x2e, 0x05, 0x1d, 0x8b, 0xc2, 0xa0, 0xde, 0x21, 0x41, 0x5e,
0x11, 0x1b, 0x96, 0xd9, 0xa6, 0xae, 0xbd, 0xf0, 0x91, 0xad, 0x69, 0x2b,
0xd2, 0x3f, 0xe4, 0x3d, 0x16, 0x69, 0xa6, 0xb2, 0x9c, 0xbe, 0x59, 0x7b,
0x87, 0x79, 0xf5, 0xc2, 0x5a, 0xcc, 0xdf, 0xfe, 0x7f, 0xf9, 0xa6, 0x52,
0xde, 0x5f, 0x46, 0x91, 0x21, 0x2c, 0x2c, 0x49, 0x25, 0x00, 0xd5, 0xe4,
0x81, 0x6b, 0x85, 0xad, 0x98, 0xaf, 0x06, 0x4a, 0x83, 0xb2, 0xe3, 0x42,
0x39, 0x31, 0x50, 0xe1, 0x2d, 0x22, 0xe6, 0x07, 0x24, 0x65, 0x29, 0x3f,
0x4c, 0xbd, 0x14, 0x8d, 0xfa, 0x31, 0xfa, 0xa4, 0xb5, 0x99, 0x04, 0xa2,
0xa5, 0xcc, 0x3b, 0x12, 0xb1, 0xaa, 0x6a, 0x17, 0x78, 0x8b, 0xb3, 0xe4,
0x3c, 0x4c, 0xc5, 0xaa, 0x79, 0x12, 0x17, 0xe0, 0x22, 0x4d, 0xf4, 0xa9,
0xd5, 0xd0, 0xed, 0xf8, 0xfe, 0x0a, 0x45, 0x80, 0x9f, 0x3b, 0x74, 0xa0,
0xb1, 0xda, 0x18, 0xfa, 0xc2, 0x7d, 0xf6, 0x18, 0x2e, 0xa9, 0x2b, 0x7e,
0x69, 0x06, 0x43, 0x2d, 0x62, 0x09, 0x42, 0x10, 0x9f, 0x83, 0xad, 0xd9,
0xdd, 0xcd, 0xcb, 0x1b, 0x33, 0x32, 0x3e, 0x1f, 0xf6, 0xac, 0x3b, 0xa3,
0x29, 0xd7, 0xc0, 0x88, 0xf9, 0xb7, 0x4c, 0xcd, 0x0a, 0x1f, 0xb8, 0x0f,
0xe6, 0xf7, 0xd7, 0x4d, 0x5f, 0x06, 0x12, 0x8a, 0x12, 0xa6, 0x2d, 0xbe,
0x5c, 0x57, 0xf8, 0x7f, 0x54, 0x3f, 0x90, 0x83, 0x2c, 0x0a, 0xc5, 0x3d,
0x03, 0x78, 0x8a, 0x68, 0xf0, 0xbd, 0xa5, 0x3e, 0xe7, 0x07, 0xab, 0xc8,
0x58, 0x2f, 0x5c, 0xfd, 0xb5, 0x39, 0xe3, 0xc6, 0x1c, 0x27, 0xf9, 0x0b,
0xc7, 0x4c, 0xcc, 0x67, 0x62, 0xe6, 0x79, 0xe8, 0xc1, 0x0a, 0x86, 0x8a,
0xb2, 0x32, 0x7b, 0x90, 0x36, 0x50, 0x92, 0x1f, 0x3e, 0x68, 0x39, 0x1c,
0x4d, 0x5d, 0xf8, 0x2b, 0xe8, 0x7d, 0xe2, 0x34, 0x61, 0x9e, 0xc3, 0x77,
0xb9, 0x4c, 0x34, 0x08, 0xda, 0x31, 0xc9, 0x1d, 0xbd, 0x3b, 0x7b, 0xf1,
0x14, 0xba, 0x3a, 0x34, 0x13, 0xaa, 0x5e, 0xa8, 0x36, 0xf6, 0xfe, 0xed,
0x5b, 0xef, 0xaf, 0x24, 0x42, 0xba, 0xfc, 0xc9, 0x30, 0x84, 0xec, 0x49,
0x14, 0xab, 0x58, 0x71, 0xfe, 0x4b, 0x6d, 0x7b, 0x9f, 0xbb, 0x3c, 0x83,
0xdf, 0x3a, 0xfb, 0x54, 0xff, 0x36, 0xaa, 0x6c, 0x47, 0x94, 0xc0, 0xde,
0x89, 0x2e, 0xac, 0x68, 0xee, 0xe8, 0xf4, 0xae, 0xa3, 0xe0, 0x91, 0x55,
0x0b, 0x0c, 0xd7, 0xf4, 0x33, 0xb5, 0xf9, 0xf2, 0x9e, 0xda, 0x78, 0xe5,
0x75, 0xec, 0xdb, 0xf6, 0xed, 0x27, 0x9f, 0x44, 0x19, 0x9f, 0xb7, 0xf0,
0xac, 0x1b, 0x3a, 0xf5, 0x77, 0xc7, 0x76, 0x1e, 0x3f, 0x78, 0x12, 0x48,
0x1d, 0xb8, 0xe0, 0x30, 0x29, 0x9a, 0x8c, 0x8f, 0x21, 0x44, 0x9c, 0x89,
0xec, 0x8e, 0xd0, 0x81, 0xf5, 0x6a, 0xd0, 0xac, 0x5e, 0xf0, 0x0f, 0x88,
0x86, 0x31, 0x2e, 0x15, 0x1e, 0x0d, 0x2d, 0xeb, 0x56, 0x30, 0x27, 0x02,
0x93, 0xf4, 0x07, 0x07, 0xba, 0xf7, 0xbd, 0xe8, 0x27, 0x4f, 0xc6, 0xd9,
0x57, 0x10, 0x3b, 0xf0, 0xff, 0x2f, 0x2d, 0x6b, 0xd0, 0x17, 0xb3, 0x49,
0xeb, 0xc2, 0x49, 0xdb,
};
static_assert(sizeof(kInput) == sizeof(kOutput),
"Input and output lengths don't match.");
static_assert(sizeof(kInput) == sizeof(kOverflowOutput),
"Input and output lengths don't match.");
TEST(ChaChaTest, TestVector) {
// Run the test with the test vector at all lengths.
for (size_t len = 0; len <= sizeof(kInput); len++) {
SCOPED_TRACE(len);
std::unique_ptr<uint8_t[]> buf(new uint8_t[len]);
CRYPTO_chacha_20(buf.get(), kInput, len, kKey, kNonce, kCounter);
EXPECT_EQ(Bytes(kOutput, len), Bytes(buf.get(), len));
// Test the in-place version.
OPENSSL_memcpy(buf.get(), kInput, len);
CRYPTO_chacha_20(buf.get(), buf.get(), len, kKey, kNonce, kCounter);
EXPECT_EQ(Bytes(kOutput, len), Bytes(buf.get(), len));
}
}
TEST(ChaChaTest, CounterOverflow) {
// Run the test with the test vector at all lengths.
for (size_t len = 0; len <= sizeof(kInput); len++) {
SCOPED_TRACE(len);
std::unique_ptr<uint8_t[]> buf(new uint8_t[len]);
CRYPTO_chacha_20(buf.get(), kInput, len, kKey, kNonce, kOverflowCounter);
EXPECT_EQ(Bytes(kOverflowOutput, len), Bytes(buf.get(), len));
// Test the in-place version.
OPENSSL_memcpy(buf.get(), kInput, len);
CRYPTO_chacha_20(buf.get(), buf.get(), len, kKey, kNonce, kOverflowCounter);
EXPECT_EQ(Bytes(kOverflowOutput, len), Bytes(buf.get(), len));
}
}
#if defined(SUPPORTS_ABI_TEST)
static void check_abi(uint8_t *out, const uint8_t *in, size_t in_len,
const uint32_t key[8], const uint32_t counter[4]) {
#if defined(CHACHA20_ASM_NEON)
if (ChaCha20_ctr32_neon_capable(in_len)) {
CHECK_ABI(ChaCha20_ctr32_neon, out, in, in_len, key, counter);
}
#endif
#if defined(CHACHA20_ASM_AVX2) && !defined(MY_ASSEMBLER_IS_TOO_OLD_FOR_512AVX)
if (ChaCha20_ctr32_avx2_capable(in_len)) {
CHECK_ABI(ChaCha20_ctr32_avx2, out, in, in_len, key, counter);
}
#endif
#if defined(CHACHA20_ASM_SSSE3_4X)
if (ChaCha20_ctr32_ssse3_4x_capable(in_len)) {
CHECK_ABI(ChaCha20_ctr32_ssse3_4x, out, in, in_len, key, counter);
}
#endif
#if defined(CHACHA20_ASM_SSSE3)
if (ChaCha20_ctr32_ssse3_capable(in_len)) {
CHECK_ABI(ChaCha20_ctr32_ssse3, out, in, in_len, key, counter);
}
#endif
#if defined(CHACHA20_ASM_NOHW)
if (in_len > 0) {
CHECK_ABI(ChaCha20_ctr32_nohw, out, in, in_len, key, counter);
}
#endif
}
TEST(ChaChaTest, ABI) {
uint32_t key[8];
OPENSSL_memcpy(key, kKey, sizeof(key));
static const uint32_t kCounterNonce[4] = {0};
std::unique_ptr<uint8_t[]> buf(new uint8_t[sizeof(kInput)]);
for (size_t len = 0; len <= 32; len++) {
SCOPED_TRACE(len);
check_abi(buf.get(), kInput, len, key, kCounterNonce);
}
for (size_t len : {32 * 2, 32 * 4, 32 * 8, 32 * 16, 32 * 24}) {
SCOPED_TRACE(len);
check_abi(buf.get(), kInput, len, key, kCounterNonce);
// Cover the partial block paths.
check_abi(buf.get(), kInput, len + 15, key, kCounterNonce);
}
}
#endif // SUPPORTS_ABI_TEST

92
vendor/aws-lc-sys/aws-lc/crypto/chacha/internal.h vendored Normal file
View File

@@ -0,0 +1,92 @@
// Copyright (c) 2018, Google Inc.
// SPDX-License-Identifier: ISC
#ifndef OPENSSL_HEADER_CHACHA_INTERNAL
#define OPENSSL_HEADER_CHACHA_INTERNAL
#include <openssl/base.h>
#include "../fipsmodule/cpucap/internal.h"
#include "../internal.h"
#if defined(__cplusplus)
extern "C" {
#endif
// CRYPTO_hchacha20 computes the HChaCha20 function, which should only be used
// as part of XChaCha20.
void CRYPTO_hchacha20(uint8_t out[32], const uint8_t key[32],
const uint8_t nonce[16]);
#if !defined(OPENSSL_NO_ASM) && defined(OPENSSL_X86)
#define CHACHA20_ASM_NOHW
#define CHACHA20_ASM_SSSE3
OPENSSL_INLINE int ChaCha20_ctr32_ssse3_capable(size_t len) {
// Unlike the x86_64 version, the x86 SSSE3 routine runs for all non-zero
// lengths.
return len > 0 && CRYPTO_is_SSSE3_capable() && CRYPTO_is_FXSR_capable();
}
void ChaCha20_ctr32_ssse3(uint8_t *out, const uint8_t *in, size_t in_len,
const uint32_t key[8], const uint32_t counter[4]);
#elif !defined(OPENSSL_NO_ASM) && \
(defined(OPENSSL_ARM) || defined(OPENSSL_AARCH64))
#define CHACHA20_ASM_NOHW
#define CHACHA20_ASM_NEON
OPENSSL_INLINE int ChaCha20_ctr32_neon_capable(size_t len) {
return len >= 192 && CRYPTO_is_NEON_capable();
}
void ChaCha20_ctr32_neon(uint8_t *out, const uint8_t *in, size_t in_len,
const uint32_t key[8], const uint32_t counter[4]);
#elif !defined(OPENSSL_NO_ASM) && defined(OPENSSL_X86_64)
#define CHACHA20_ASM_NOHW
#define CHACHA20_ASM_AVX2
OPENSSL_INLINE int ChaCha20_ctr32_avx2_capable(size_t len) {
return len > 128 && CRYPTO_is_AVX2_capable();
}
void ChaCha20_ctr32_avx2(uint8_t *out, const uint8_t *in, size_t in_len,
const uint32_t key[8], const uint32_t counter[4]);
#define CHACHA20_ASM_SSSE3_4X
OPENSSL_INLINE int ChaCha20_ctr32_ssse3_4x_capable(size_t len) {
int capable = len > 128 && CRYPTO_is_SSSE3_capable();
int faster = len > 192 || !CRYPTO_cpu_perf_is_like_silvermont();
return capable && faster;
}
void ChaCha20_ctr32_ssse3_4x(uint8_t *out, const uint8_t *in, size_t in_len,
const uint32_t key[8], const uint32_t counter[4]);
#define CHACHA20_ASM_SSSE3
OPENSSL_INLINE int ChaCha20_ctr32_ssse3_capable(size_t len) {
return len > 128 && CRYPTO_is_SSSE3_capable();
}
void ChaCha20_ctr32_ssse3(uint8_t *out, const uint8_t *in, size_t in_len,
const uint32_t key[8], const uint32_t counter[4]);
#endif
#if defined(CHACHA20_ASM_NOHW)
// ChaCha20_ctr32_nohw encrypts |in_len| bytes from |in| and writes the result
// to |out|. If |in| and |out| alias, they must be equal. |in_len| may not be
// zero.
//
// |counter[0]| is the initial 32-bit block counter, and the remainder is the
// 96-bit nonce. If the counter overflows, the output is undefined. The function
// will produce output, but the output may vary by machine and may not be
// self-consistent. (On some architectures, the assembly implements a mix of
// 64-bit and 32-bit counters.)
void ChaCha20_ctr32_nohw(uint8_t *out, const uint8_t *in, size_t in_len,
const uint32_t key[8], const uint32_t counter[4]);
#endif
#if defined(__cplusplus)
} // extern C
#endif
#endif // OPENSSL_HEADER_CHACHA_INTERNAL