kris/boringssl
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
74944287e1
boringssl/crypto/fipsmodule/sha/asm/sha1-armv4-large.pl

744 lines
19 KiB
Perl
Raw Normal View History

Sync ARM assembly up to 609b0852e4d50251857dbbac3141ba042e35a9ae. This change was made by copying over the files as of that commit and then discarding the parts of the diff which corresponding to our own changes. Change-Id: I28c5d711f7a8cec30749b8174687434129af5209 Reviewed-on: https://boringssl-review.googlesource.com/17111 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:53:40 +01:00
#! /usr/bin/env perl
# Copyright 2007-2016 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
Inital import. Initial fork from f2d678e6e89b6508147086610e985d4e8416e867 (1.0.2 beta). (This change contains substantial changes from the original and effectively starts a new history.)
2014-06-20 20:00:00 +01:00
# ====================================================================
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
Inital import. Initial fork from f2d678e6e89b6508147086610e985d4e8416e867 (1.0.2 beta). (This change contains substantial changes from the original and effectively starts a new history.)
2014-06-20 20:00:00 +01:00
# 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/.
# ====================================================================
# sha1_block procedure for ARMv4.
#
# January 2007.
# Size/performance trade-off
# ====================================================================
# impl size in bytes comp cycles[*] measured performance
# ====================================================================
# thumb 304 3212 4420
# armv4-small 392/+29% 1958/+64% 2250/+96%
# armv4-compact 740/+89% 1552/+26% 1840/+22%
# armv4-large 1420/+92% 1307/+19% 1370/+34%[***]
# full unroll ~5100/+260% ~1260/+4% ~1300/+5%
# ====================================================================
# thumb = same as 'small' but in Thumb instructions[**] and
# with recurring code in two private functions;
# small = detached Xload/update, loops are folded;
# compact = detached Xload/update, 5x unroll;
# large = interleaved Xload/update, 5x unroll;
# full unroll = interleaved Xload/update, full unroll, estimated[!];
#
# [*] Manually counted instructions in "grand" loop body. Measured
# performance is affected by prologue and epilogue overhead,
# i-cache availability, branch penalties, etc.
# [**] While each Thumb instruction is twice smaller, they are not as
# diverse as ARM ones: e.g., there are only two arithmetic
# instructions with 3 arguments, no [fixed] rotate, addressing
# modes are limited. As result it takes more instructions to do
# the same job in Thumb, therefore the code is never twice as
# small and always slower.
# [***] which is also ~35% better than compiler generated code. Dual-
# issue Cortex A8 core was measured to process input block in
# ~990 cycles.
# August 2010.
#
# Rescheduling for dual-issue pipeline resulted in 13% improvement on
# Cortex A8 core and in absolute terms ~870 cycles per input block
# [or 13.6 cycles per byte].
# February 2011.
#
# Profiler-assisted and platform-specific optimization resulted in 10%
# improvement on Cortex A8 core and 12.2 cycles per byte.
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
# September 2013.
#
# Add NEON implementation (see sha1-586.pl for background info). On
# Cortex A8 it was measured to process one byte in 6.7 cycles or >80%
# faster than integer-only code. Because [fully unrolled] NEON code
# is ~2.5x larger and there are some redundant instructions executed
# when processing last block, improvement is not as big for smallest
# blocks, only ~30%. Snapdragon S4 is a tad faster, 6.4 cycles per
ARMv4 assembly pack: add Cortex-A15 performance data. (Imported from upstream's e390ae50e0bc41676994c6fa23f7b65a8afc4d7f) Change-Id: Ifee85b0936c06c42cc7c09f8327d15fec51da48a Reviewed-on: https://boringssl-review.googlesource.com/3832 Reviewed-by: Adam Langley <agl@google.com>
2015-03-09 18:51:51 +00:00
# byte, which is also >80% faster than integer-only code. Cortex-A15
# is even faster spending 5.6 cycles per byte outperforming integer-
# only code by factor of 2.
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
# May 2014.
#
# Add ARMv8 code path performing at 2.35 cpb on Apple A7.
Add assembly support for 32-bit iOS. (Imported from upstream's 313e6ec11fb8a7bda1676ce5804bee8755664141) BUG=338886 Change-Id: Id635e78b9afaad5ca311e3aeed888c9aedeb9637 Reviewed-on: https://boringssl-review.googlesource.com/4490 Reviewed-by: Adam Langley <agl@google.com>
2015-04-21 02:07:38 +01:00
$flavour = shift;
Switch perlasm calling convention. Depending on architecture, perlasm differed on which one or both of: perl foo.pl flavor output.S perl foo.pl flavor > output.S Upstream has now unified on the first form after making a number of changes to their files (the second does not even work for their x86 files anymore). Sync those portions of our perlasm scripts with upstream and update CMakeLists.txt and generate_build_files.py per the new convention. This imports various commits like this one: 184bc45f683c76531d7e065b6553ca9086564576 (this was done by taking a diff, so I don't have the full list) Confirmed that generate_build_files.py sees no change. BUG=14 Change-Id: Id2fb5b8bc2a7369d077221b5df9a6947d41f50d2 Reviewed-on: https://boringssl-review.googlesource.com/8518 Reviewed-by: Adam Langley <agl@google.com>
2016-06-26 18:18:50 +01:00
if ($flavour=~/\w[\w\-]*\.\w+$/) { $output=$flavour; undef $flavour; }
else { while (($output=shift) && ($output!~/\w[\w\-]*\.\w+$/)) {} }
Add assembly support for 32-bit iOS. (Imported from upstream's 313e6ec11fb8a7bda1676ce5804bee8755664141) BUG=338886 Change-Id: Id635e78b9afaad5ca311e3aeed888c9aedeb9637 Reviewed-on: https://boringssl-review.googlesource.com/4490 Reviewed-by: Adam Langley <agl@google.com>
2015-04-21 02:07:38 +01:00
if ($flavour && $flavour ne "void") {
$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
( $xlate="${dir}arm-xlate.pl" and -f $xlate ) or
First part of the FIPS module. Change-Id: Ic3a91ccd2c8cdc364740f256fdb8a7ff66177947 Reviewed-on: https://boringssl-review.googlesource.com/14506 Reviewed-by: Adam Langley <agl@google.com> Commit-Queue: Adam Langley <agl@google.com>
2017-04-04 22:21:43 +01:00
( $xlate="${dir}../../../perlasm/arm-xlate.pl" and -f $xlate) or
Add assembly support for 32-bit iOS. (Imported from upstream's 313e6ec11fb8a7bda1676ce5804bee8755664141) BUG=338886 Change-Id: Id635e78b9afaad5ca311e3aeed888c9aedeb9637 Reviewed-on: https://boringssl-review.googlesource.com/4490 Reviewed-by: Adam Langley <agl@google.com>
2015-04-21 02:07:38 +01:00
die "can't locate arm-xlate.pl";
open STDOUT,"| \"$^X\" $xlate $flavour $output";
} else {
open STDOUT,">$output";
}
Inital import. Initial fork from f2d678e6e89b6508147086610e985d4e8416e867 (1.0.2 beta). (This change contains substantial changes from the original and effectively starts a new history.)
2014-06-20 20:00:00 +01:00
$ctx="r0";
$inp="r1";
$len="r2";
$a="r3";
$b="r4";
$c="r5";
$d="r6";
$e="r7";
$K="r8";
$t0="r9";
$t1="r10";
$t2="r11";
$t3="r12";
$Xi="r14";
@V=($a,$b,$c,$d,$e);
sub Xupdate {
my ($a,$b,$c,$d,$e,$opt1,$opt2)=@_;
$code.=<<___;
ldr $t0,[$Xi,#15*4]
ldr $t1,[$Xi,#13*4]
ldr $t2,[$Xi,#7*4]
add $e,$K,$e,ror#2 @ E+=K_xx_xx
ldr $t3,[$Xi,#2*4]
eor $t0,$t0,$t1
eor $t2,$t2,$t3 @ 1 cycle stall
eor $t1,$c,$d @ F_xx_xx
mov $t0,$t0,ror#31
add $e,$e,$a,ror#27 @ E+=ROR(A,27)
eor $t0,$t0,$t2,ror#31
str $t0,[$Xi,#-4]!
$opt1 @ F_xx_xx
$opt2 @ F_xx_xx
add $e,$e,$t0 @ E+=X[i]
___
}
sub BODY_00_15 {
my ($a,$b,$c,$d,$e)=@_;
$code.=<<___;
#if __ARM_ARCH__<7
ldrb $t1,[$inp,#2]
ldrb $t0,[$inp,#3]
ldrb $t2,[$inp,#1]
add $e,$K,$e,ror#2 @ E+=K_00_19
ldrb $t3,[$inp],#4
orr $t0,$t0,$t1,lsl#8
eor $t1,$c,$d @ F_xx_xx
orr $t0,$t0,$t2,lsl#16
add $e,$e,$a,ror#27 @ E+=ROR(A,27)
orr $t0,$t0,$t3,lsl#24
#else
ldr $t0,[$inp],#4 @ handles unaligned
add $e,$K,$e,ror#2 @ E+=K_00_19
eor $t1,$c,$d @ F_xx_xx
add $e,$e,$a,ror#27 @ E+=ROR(A,27)
#ifdef __ARMEL__
rev $t0,$t0 @ byte swap
#endif
#endif
and $t1,$b,$t1,ror#2
add $e,$e,$t0 @ E+=X[i]
eor $t1,$t1,$d,ror#2 @ F_00_19(B,C,D)
str $t0,[$Xi,#-4]!
add $e,$e,$t1 @ E+=F_00_19(B,C,D)
___
}
sub BODY_16_19 {
my ($a,$b,$c,$d,$e)=@_;
&Xupdate(@_,"and $t1,$b,$t1,ror#2");
$code.=<<___;
eor $t1,$t1,$d,ror#2 @ F_00_19(B,C,D)
add $e,$e,$t1 @ E+=F_00_19(B,C,D)
___
}
sub BODY_20_39 {
my ($a,$b,$c,$d,$e)=@_;
&Xupdate(@_,"eor $t1,$b,$t1,ror#2");
$code.=<<___;
add $e,$e,$t1 @ E+=F_20_39(B,C,D)
___
}
sub BODY_40_59 {
my ($a,$b,$c,$d,$e)=@_;
&Xupdate(@_,"and $t1,$b,$t1,ror#2","and $t2,$c,$d");
$code.=<<___;
add $e,$e,$t1 @ E+=F_40_59(B,C,D)
add $e,$e,$t2,ror#2
___
}
$code=<<___;
Move arm_arch.h and fix up lots of include paths. arm_arch.h is included from ARM asm files, but lives in crypto/, not openssl/include/. Since the asm files are often built from a different location than their position in the source tree, relative include paths are unlikely to work so, rather than having crypto/ be a de-facto, second global include path, this change moves arm_arch.h to include/openssl/. It also removes entries from many include paths because they should be needed as relative includes are always based on the locations of the source file. Change-Id: I638ff43d641ca043a4fc06c0d901b11c6ff73542 Reviewed-on: https://boringssl-review.googlesource.com/5746 Reviewed-by: Adam Langley <agl@google.com>
2015-08-25 02:03:17 +01:00
#include <openssl/arm_arch.h>
Inital import. Initial fork from f2d678e6e89b6508147086610e985d4e8416e867 (1.0.2 beta). (This change contains substantial changes from the original and effectively starts a new history.)
2014-06-20 20:00:00 +01:00
.text
ARMv4 assembly pack: allow Thumb2 even in iOS build, and engage it in most modules. (Imported from upstream's a285992763f3961f69a8d86bf7dfff020a08cef9.) Change-Id: I59df0b567e8e80befe5c399f817d6410ddafc577 Reviewed-on: https://boringssl-review.googlesource.com/17110 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:43:31 +01:00
#if defined(__thumb2__)
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
.syntax unified
.thumb
#else
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
.code 32
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
#endif
Inital import. Initial fork from f2d678e6e89b6508147086610e985d4e8416e867 (1.0.2 beta). (This change contains substantial changes from the original and effectively starts a new history.)
2014-06-20 20:00:00 +01:00
.global sha1_block_data_order
.type sha1_block_data_order,%function
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
.align 5
Inital import. Initial fork from f2d678e6e89b6508147086610e985d4e8416e867 (1.0.2 beta). (This change contains substantial changes from the original and effectively starts a new history.)
2014-06-20 20:00:00 +01:00
sha1_block_data_order:
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
#if __ARM_MAX_ARCH__>=7
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
.Lsha1_block:
adr r3,.Lsha1_block
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
ldr r12,.LOPENSSL_armcap
ldr r12,[r3,r12] @ OPENSSL_armcap_P
Add assembly support for 32-bit iOS. (Imported from upstream's 313e6ec11fb8a7bda1676ce5804bee8755664141) BUG=338886 Change-Id: Id635e78b9afaad5ca311e3aeed888c9aedeb9637 Reviewed-on: https://boringssl-review.googlesource.com/4490 Reviewed-by: Adam Langley <agl@google.com>
2015-04-21 02:07:38 +01:00
#ifdef __APPLE__
ldr r12,[r12]
#endif
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
tst r12,#ARMV8_SHA1
bne .LARMv8
tst r12,#ARMV7_NEON
bne .LNEON
#endif
Inital import. Initial fork from f2d678e6e89b6508147086610e985d4e8416e867 (1.0.2 beta). (This change contains substantial changes from the original and effectively starts a new history.)
2014-06-20 20:00:00 +01:00
stmdb sp!,{r4-r12,lr}
add $len,$inp,$len,lsl#6 @ $len to point at the end of $inp
ldmia $ctx,{$a,$b,$c,$d,$e}
.Lloop:
ldr $K,.LK_00_19
mov $Xi,sp
sub sp,sp,#15*4
mov $c,$c,ror#30
mov $d,$d,ror#30
mov $e,$e,ror#30 @ [6]
.L_00_15:
___
for($i=0;$i<5;$i++) {
&BODY_00_15(@V); unshift(@V,pop(@V));
}
$code.=<<___;
ARMv4 assembly pack: allow Thumb2 even in iOS build, and engage it in most modules. (Imported from upstream's a285992763f3961f69a8d86bf7dfff020a08cef9.) Change-Id: I59df0b567e8e80befe5c399f817d6410ddafc577 Reviewed-on: https://boringssl-review.googlesource.com/17110 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:43:31 +01:00
#if defined(__thumb2__)
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
mov $t3,sp
teq $Xi,$t3
#else
Inital import. Initial fork from f2d678e6e89b6508147086610e985d4e8416e867 (1.0.2 beta). (This change contains substantial changes from the original and effectively starts a new history.)
2014-06-20 20:00:00 +01:00
teq $Xi,sp
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
#endif
Inital import. Initial fork from f2d678e6e89b6508147086610e985d4e8416e867 (1.0.2 beta). (This change contains substantial changes from the original and effectively starts a new history.)
2014-06-20 20:00:00 +01:00
bne .L_00_15 @ [((11+4)*5+2)*3]
sub sp,sp,#25*4
___
&BODY_00_15(@V); unshift(@V,pop(@V));
&BODY_16_19(@V); unshift(@V,pop(@V));
&BODY_16_19(@V); unshift(@V,pop(@V));
&BODY_16_19(@V); unshift(@V,pop(@V));
&BODY_16_19(@V); unshift(@V,pop(@V));
$code.=<<___;
ldr $K,.LK_20_39 @ [+15+16*4]
cmn sp,#0 @ [+3], clear carry to denote 20_39
.L_20_39_or_60_79:
___
for($i=0;$i<5;$i++) {
&BODY_20_39(@V); unshift(@V,pop(@V));
}
$code.=<<___;
ARMv4 assembly pack: allow Thumb2 even in iOS build, and engage it in most modules. (Imported from upstream's a285992763f3961f69a8d86bf7dfff020a08cef9.) Change-Id: I59df0b567e8e80befe5c399f817d6410ddafc577 Reviewed-on: https://boringssl-review.googlesource.com/17110 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:43:31 +01:00
#if defined(__thumb2__)
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
mov $t3,sp
teq $Xi,$t3
#else
Inital import. Initial fork from f2d678e6e89b6508147086610e985d4e8416e867 (1.0.2 beta). (This change contains substantial changes from the original and effectively starts a new history.)
2014-06-20 20:00:00 +01:00
teq $Xi,sp @ preserve carry
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
#endif
Inital import. Initial fork from f2d678e6e89b6508147086610e985d4e8416e867 (1.0.2 beta). (This change contains substantial changes from the original and effectively starts a new history.)
2014-06-20 20:00:00 +01:00
bne .L_20_39_or_60_79 @ [+((12+3)*5+2)*4]
bcs .L_done @ [+((12+3)*5+2)*4], spare 300 bytes
ldr $K,.LK_40_59
sub sp,sp,#20*4 @ [+2]
.L_40_59:
___
for($i=0;$i<5;$i++) {
&BODY_40_59(@V); unshift(@V,pop(@V));
}
$code.=<<___;
ARMv4 assembly pack: allow Thumb2 even in iOS build, and engage it in most modules. (Imported from upstream's a285992763f3961f69a8d86bf7dfff020a08cef9.) Change-Id: I59df0b567e8e80befe5c399f817d6410ddafc577 Reviewed-on: https://boringssl-review.googlesource.com/17110 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:43:31 +01:00
#if defined(__thumb2__)
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
mov $t3,sp
teq $Xi,$t3
#else
Inital import. Initial fork from f2d678e6e89b6508147086610e985d4e8416e867 (1.0.2 beta). (This change contains substantial changes from the original and effectively starts a new history.)
2014-06-20 20:00:00 +01:00
teq $Xi,sp
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
#endif
Inital import. Initial fork from f2d678e6e89b6508147086610e985d4e8416e867 (1.0.2 beta). (This change contains substantial changes from the original and effectively starts a new history.)
2014-06-20 20:00:00 +01:00
bne .L_40_59 @ [+((12+5)*5+2)*4]
ldr $K,.LK_60_79
sub sp,sp,#20*4
cmp sp,#0 @ set carry to denote 60_79
b .L_20_39_or_60_79 @ [+4], spare 300 bytes
.L_done:
add sp,sp,#80*4 @ "deallocate" stack frame
ldmia $ctx,{$K,$t0,$t1,$t2,$t3}
add $a,$K,$a
add $b,$t0,$b
add $c,$t1,$c,ror#2
add $d,$t2,$d,ror#2
add $e,$t3,$e,ror#2
stmia $ctx,{$a,$b,$c,$d,$e}
teq $inp,$len
bne .Lloop @ [+18], total 1307
#if __ARM_ARCH__>=5
ldmia sp!,{r4-r12,pc}
#else
ldmia sp!,{r4-r12,lr}
tst lr,#1
moveq pc,lr @ be binary compatible with V4, yet
bx lr @ interoperable with Thumb ISA:-)
#endif
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
.size sha1_block_data_order,.-sha1_block_data_order
.align 5
Inital import. Initial fork from f2d678e6e89b6508147086610e985d4e8416e867 (1.0.2 beta). (This change contains substantial changes from the original and effectively starts a new history.)
2014-06-20 20:00:00 +01:00
.LK_00_19: .word 0x5a827999
.LK_20_39: .word 0x6ed9eba1
.LK_40_59: .word 0x8f1bbcdc
.LK_60_79: .word 0xca62c1d6
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
#if __ARM_MAX_ARCH__>=7
.LOPENSSL_armcap:
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
.word OPENSSL_armcap_P-.Lsha1_block
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
#endif
.asciz "SHA1 block transform for ARMv4/NEON/ARMv8, CRYPTOGAMS by <appro\@openssl.org>"
.align 5
___
#####################################################################
# NEON stuff
#
{{{
my @V=($a,$b,$c,$d,$e);
my ($K_XX_XX,$Ki,$t0,$t1,$Xfer,$saved_sp)=map("r$_",(8..12,14));
my $Xi=4;
my @X=map("q$_",(8..11,0..3));
my @Tx=("q12","q13");
my ($K,$zero)=("q14","q15");
my $j=0;
sub AUTOLOAD() # thunk [simplified] x86-style perlasm
{ my $opcode = $AUTOLOAD; $opcode =~ s/.*:://; $opcode =~ s/_/\./;
my $arg = pop;
$arg = "#$arg" if ($arg*1 eq $arg);
$code .= "\t$opcode\t".join(',',@_,$arg)."\n";
}
sub body_00_19 () {
(
'($a,$b,$c,$d,$e)=@V;'. # '$code.="@ $j\n";'.
'&bic ($t0,$d,$b)',
'&add ($e,$e,$Ki)', # e+=X[i]+K
'&and ($t1,$c,$b)',
'&ldr ($Ki,sprintf "[sp,#%d]",4*(($j+1)&15))',
'&add ($e,$e,$a,"ror#27")', # e+=ROR(A,27)
'&eor ($t1,$t1,$t0)', # F_00_19
'&mov ($b,$b,"ror#2")', # b=ROR(b,2)
'&add ($e,$e,$t1);'. # e+=F_00_19
'$j++; unshift(@V,pop(@V));'
)
}
sub body_20_39 () {
(
'($a,$b,$c,$d,$e)=@V;'. # '$code.="@ $j\n";'.
'&eor ($t0,$b,$d)',
'&add ($e,$e,$Ki)', # e+=X[i]+K
'&ldr ($Ki,sprintf "[sp,#%d]",4*(($j+1)&15)) if ($j<79)',
'&eor ($t1,$t0,$c)', # F_20_39
'&add ($e,$e,$a,"ror#27")', # e+=ROR(A,27)
'&mov ($b,$b,"ror#2")', # b=ROR(b,2)
'&add ($e,$e,$t1);'. # e+=F_20_39
'$j++; unshift(@V,pop(@V));'
)
}
sub body_40_59 () {
(
'($a,$b,$c,$d,$e)=@V;'. # '$code.="@ $j\n";'.
'&add ($e,$e,$Ki)', # e+=X[i]+K
'&and ($t0,$c,$d)',
'&ldr ($Ki,sprintf "[sp,#%d]",4*(($j+1)&15))',
'&add ($e,$e,$a,"ror#27")', # e+=ROR(A,27)
'&eor ($t1,$c,$d)',
'&add ($e,$e,$t0)',
'&and ($t1,$t1,$b)',
'&mov ($b,$b,"ror#2")', # b=ROR(b,2)
'&add ($e,$e,$t1);'. # e+=F_40_59
'$j++; unshift(@V,pop(@V));'
)
}
sub Xupdate_16_31 ()
{ use integer;
my $body = shift;
my @insns = (&$body,&$body,&$body,&$body);
my ($a,$b,$c,$d,$e);
&vext_8 (@X[0],@X[-4&7],@X[-3&7],8); # compose "X[-14]" in "X[0]"
eval(shift(@insns));
eval(shift(@insns));
eval(shift(@insns));
&vadd_i32 (@Tx[1],@X[-1&7],$K);
eval(shift(@insns));
&vld1_32 ("{$K\[]}","[$K_XX_XX,:32]!") if ($Xi%5==0);
eval(shift(@insns));
&vext_8 (@Tx[0],@X[-1&7],$zero,4); # "X[-3]", 3 words
eval(shift(@insns));
eval(shift(@insns));
eval(shift(@insns));
&veor (@X[0],@X[0],@X[-4&7]); # "X[0]"^="X[-16]"
eval(shift(@insns));
eval(shift(@insns));
&veor (@Tx[0],@Tx[0],@X[-2&7]); # "X[-3]"^"X[-8]"
eval(shift(@insns));
eval(shift(@insns));
&veor (@Tx[0],@Tx[0],@X[0]); # "X[0]"^="X[-3]"^"X[-8]
eval(shift(@insns));
eval(shift(@insns));
&vst1_32 ("{@Tx[1]}","[$Xfer,:128]!"); # X[]+K xfer
&sub ($Xfer,$Xfer,64) if ($Xi%4==0);
eval(shift(@insns));
eval(shift(@insns));
&vext_8 (@Tx[1],$zero,@Tx[0],4); # "X[0]"<<96, extract one dword
eval(shift(@insns));
eval(shift(@insns));
&vadd_i32 (@X[0],@Tx[0],@Tx[0]);
eval(shift(@insns));
eval(shift(@insns));
&vsri_32 (@X[0],@Tx[0],31); # "X[0]"<<<=1
eval(shift(@insns));
eval(shift(@insns));
eval(shift(@insns));
&vshr_u32 (@Tx[0],@Tx[1],30);
eval(shift(@insns));
eval(shift(@insns));
&vshl_u32 (@Tx[1],@Tx[1],2);
eval(shift(@insns));
eval(shift(@insns));
&veor (@X[0],@X[0],@Tx[0]);
eval(shift(@insns));
eval(shift(@insns));
&veor (@X[0],@X[0],@Tx[1]); # "X[0]"^=("X[0]">>96)<<<2
foreach (@insns) { eval; } # remaining instructions [if any]
$Xi++; push(@X,shift(@X)); # "rotate" X[]
}
sub Xupdate_32_79 ()
{ use integer;
my $body = shift;
my @insns = (&$body,&$body,&$body,&$body);
my ($a,$b,$c,$d,$e);
&vext_8 (@Tx[0],@X[-2&7],@X[-1&7],8); # compose "X[-6]"
eval(shift(@insns));
eval(shift(@insns));
eval(shift(@insns));
&veor (@X[0],@X[0],@X[-4&7]); # "X[0]"="X[-32]"^"X[-16]"
eval(shift(@insns));
eval(shift(@insns));
&veor (@X[0],@X[0],@X[-7&7]); # "X[0]"^="X[-28]"
eval(shift(@insns));
eval(shift(@insns));
&vadd_i32 (@Tx[1],@X[-1&7],$K);
eval(shift(@insns));
&vld1_32 ("{$K\[]}","[$K_XX_XX,:32]!") if ($Xi%5==0);
eval(shift(@insns));
&veor (@Tx[0],@Tx[0],@X[0]); # "X[-6]"^="X[0]"
eval(shift(@insns));
eval(shift(@insns));
&vshr_u32 (@X[0],@Tx[0],30);
eval(shift(@insns));
eval(shift(@insns));
&vst1_32 ("{@Tx[1]}","[$Xfer,:128]!"); # X[]+K xfer
&sub ($Xfer,$Xfer,64) if ($Xi%4==0);
eval(shift(@insns));
eval(shift(@insns));
&vsli_32 (@X[0],@Tx[0],2); # "X[0]"="X[-6]"<<<2
foreach (@insns) { eval; } # remaining instructions [if any]
$Xi++; push(@X,shift(@X)); # "rotate" X[]
}
sub Xuplast_80 ()
{ use integer;
my $body = shift;
my @insns = (&$body,&$body,&$body,&$body);
my ($a,$b,$c,$d,$e);
&vadd_i32 (@Tx[1],@X[-1&7],$K);
eval(shift(@insns));
eval(shift(@insns));
&vst1_32 ("{@Tx[1]}","[$Xfer,:128]!");
&sub ($Xfer,$Xfer,64);
&teq ($inp,$len);
&sub ($K_XX_XX,$K_XX_XX,16); # rewind $K_XX_XX
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
&it ("eq");
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
&subeq ($inp,$inp,64); # reload last block to avoid SEGV
&vld1_8 ("{@X[-4&7]-@X[-3&7]}","[$inp]!");
eval(shift(@insns));
eval(shift(@insns));
&vld1_8 ("{@X[-2&7]-@X[-1&7]}","[$inp]!");
eval(shift(@insns));
eval(shift(@insns));
&vld1_32 ("{$K\[]}","[$K_XX_XX,:32]!"); # load K_00_19
eval(shift(@insns));
eval(shift(@insns));
&vrev32_8 (@X[-4&7],@X[-4&7]);
foreach (@insns) { eval; } # remaining instructions
$Xi=0;
}
sub Xloop()
{ use integer;
my $body = shift;
my @insns = (&$body,&$body,&$body,&$body);
my ($a,$b,$c,$d,$e);
&vrev32_8 (@X[($Xi-3)&7],@X[($Xi-3)&7]);
eval(shift(@insns));
eval(shift(@insns));
&vadd_i32 (@X[$Xi&7],@X[($Xi-4)&7],$K);
eval(shift(@insns));
eval(shift(@insns));
&vst1_32 ("{@X[$Xi&7]}","[$Xfer,:128]!");# X[]+K xfer to IALU
foreach (@insns) { eval; }
$Xi++;
}
$code.=<<___;
#if __ARM_MAX_ARCH__>=7
.arch armv7-a
.fpu neon
.type sha1_block_data_order_neon,%function
.align 4
sha1_block_data_order_neon:
.LNEON:
stmdb sp!,{r4-r12,lr}
add $len,$inp,$len,lsl#6 @ $len to point at the end of $inp
@ dmb @ errata #451034 on early Cortex A8
@ vstmdb sp!,{d8-d15} @ ABI specification says so
mov $saved_sp,sp
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
sub $Xfer,sp,#64
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
adr $K_XX_XX,.LK_00_19
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
bic $Xfer,$Xfer,#15 @ align for 128-bit stores
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
ldmia $ctx,{$a,$b,$c,$d,$e} @ load context
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
mov sp,$Xfer @ alloca
Inital import. Initial fork from f2d678e6e89b6508147086610e985d4e8416e867 (1.0.2 beta). (This change contains substantial changes from the original and effectively starts a new history.)
2014-06-20 20:00:00 +01:00
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
vld1.8 {@X[-4&7]-@X[-3&7]},[$inp]! @ handles unaligned
veor $zero,$zero,$zero
vld1.8 {@X[-2&7]-@X[-1&7]},[$inp]!
vld1.32 {${K}\[]},[$K_XX_XX,:32]! @ load K_00_19
vrev32.8 @X[-4&7],@X[-4&7] @ yes, even on
vrev32.8 @X[-3&7],@X[-3&7] @ big-endian...
vrev32.8 @X[-2&7],@X[-2&7]
vadd.i32 @X[0],@X[-4&7],$K
vrev32.8 @X[-1&7],@X[-1&7]
vadd.i32 @X[1],@X[-3&7],$K
vst1.32 {@X[0]},[$Xfer,:128]!
vadd.i32 @X[2],@X[-2&7],$K
vst1.32 {@X[1]},[$Xfer,:128]!
vst1.32 {@X[2]},[$Xfer,:128]!
ldr $Ki,[sp] @ big RAW stall
.Loop_neon:
___
&Xupdate_16_31(\&body_00_19);
&Xupdate_16_31(\&body_00_19);
&Xupdate_16_31(\&body_00_19);
&Xupdate_16_31(\&body_00_19);
&Xupdate_32_79(\&body_00_19);
&Xupdate_32_79(\&body_20_39);
&Xupdate_32_79(\&body_20_39);
&Xupdate_32_79(\&body_20_39);
&Xupdate_32_79(\&body_20_39);
&Xupdate_32_79(\&body_20_39);
&Xupdate_32_79(\&body_40_59);
&Xupdate_32_79(\&body_40_59);
&Xupdate_32_79(\&body_40_59);
&Xupdate_32_79(\&body_40_59);
&Xupdate_32_79(\&body_40_59);
&Xupdate_32_79(\&body_20_39);
&Xuplast_80(\&body_20_39);
&Xloop(\&body_20_39);
&Xloop(\&body_20_39);
&Xloop(\&body_20_39);
$code.=<<___;
ldmia $ctx,{$Ki,$t0,$t1,$Xfer} @ accumulate context
add $a,$a,$Ki
ldr $Ki,[$ctx,#16]
add $b,$b,$t0
add $c,$c,$t1
add $d,$d,$Xfer
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
it eq
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
moveq sp,$saved_sp
add $e,$e,$Ki
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
it ne
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
ldrne $Ki,[sp]
stmia $ctx,{$a,$b,$c,$d,$e}
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
itt ne
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
addne $Xfer,sp,#3*16
bne .Loop_neon
@ vldmia sp!,{d8-d15}
ldmia sp!,{r4-r12,pc}
.size sha1_block_data_order_neon,.-sha1_block_data_order_neon
Inital import. Initial fork from f2d678e6e89b6508147086610e985d4e8416e867 (1.0.2 beta). (This change contains substantial changes from the original and effectively starts a new history.)
2014-06-20 20:00:00 +01:00
#endif
___
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
}}}
#####################################################################
# ARMv8 stuff
#
{{{
my ($ABCD,$E,$E0,$E1)=map("q$_",(0..3));
my @MSG=map("q$_",(4..7));
my @Kxx=map("q$_",(8..11));
my ($W0,$W1,$ABCD_SAVE)=map("q$_",(12..14));
$code.=<<___;
#if __ARM_MAX_ARCH__>=7
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
ARMv4 assembly pack: allow Thumb2 even in iOS build, and engage it in most modules. (Imported from upstream's a285992763f3961f69a8d86bf7dfff020a08cef9.) Change-Id: I59df0b567e8e80befe5c399f817d6410ddafc577 Reviewed-on: https://boringssl-review.googlesource.com/17110 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:43:31 +01:00
# if defined(__thumb2__)
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
# define INST(a,b,c,d) .byte c,d|0xf,a,b
# else
# define INST(a,b,c,d) .byte a,b,c,d|0x10
# endif
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
.type sha1_block_data_order_armv8,%function
.align 5
sha1_block_data_order_armv8:
.LARMv8:
vstmdb sp!,{d8-d15} @ ABI specification says so
veor $E,$E,$E
adr r3,.LK_00_19
vld1.32 {$ABCD},[$ctx]!
vld1.32 {$E\[0]},[$ctx]
sub $ctx,$ctx,#16
vld1.32 {@Kxx[0]\[]},[r3,:32]!
vld1.32 {@Kxx[1]\[]},[r3,:32]!
vld1.32 {@Kxx[2]\[]},[r3,:32]!
vld1.32 {@Kxx[3]\[]},[r3,:32]
.Loop_v8:
vld1.8 {@MSG[0]-@MSG[1]},[$inp]!
vld1.8 {@MSG[2]-@MSG[3]},[$inp]!
vrev32.8 @MSG[0],@MSG[0]
vrev32.8 @MSG[1],@MSG[1]
vadd.i32 $W0,@Kxx[0],@MSG[0]
vrev32.8 @MSG[2],@MSG[2]
vmov $ABCD_SAVE,$ABCD @ offload
subs $len,$len,#1
vadd.i32 $W1,@Kxx[0],@MSG[1]
vrev32.8 @MSG[3],@MSG[3]
sha1h $E1,$ABCD @ 0
sha1c $ABCD,$E,$W0
vadd.i32 $W0,@Kxx[$j],@MSG[2]
sha1su0 @MSG[0],@MSG[1],@MSG[2]
___
for ($j=0,$i=1;$i<20-3;$i++) {
my $f=("c","p","m","p")[$i/5];
$code.=<<___;
sha1h $E0,$ABCD @ $i
sha1$f $ABCD,$E1,$W1
vadd.i32 $W1,@Kxx[$j],@MSG[3]
sha1su1 @MSG[0],@MSG[3]
___
$code.=<<___ if ($i<20-4);
sha1su0 @MSG[1],@MSG[2],@MSG[3]
___
($E0,$E1)=($E1,$E0); ($W0,$W1)=($W1,$W0);
push(@MSG,shift(@MSG)); $j++ if ((($i+3)%5)==0);
}
$code.=<<___;
sha1h $E0,$ABCD @ $i
sha1p $ABCD,$E1,$W1
vadd.i32 $W1,@Kxx[$j],@MSG[3]
sha1h $E1,$ABCD @ 18
sha1p $ABCD,$E0,$W0
sha1h $E0,$ABCD @ 19
sha1p $ABCD,$E1,$W1
vadd.i32 $E,$E,$E0
vadd.i32 $ABCD,$ABCD,$ABCD_SAVE
bne .Loop_v8
vst1.32 {$ABCD},[$ctx]!
vst1.32 {$E\[0]},[$ctx]
vldmia sp!,{d8-d15}
ret @ bx lr
.size sha1_block_data_order_armv8,.-sha1_block_data_order_armv8
#endif
___
}}}
$code.=<<___;
#if __ARM_MAX_ARCH__>=7
.comm OPENSSL_armcap_P,4,4
Revert assembly changes in "Hide CPU capability symbols in C." This partially reverts commit 38636aba74400272afbb7155bbdfef86be6ed715. Some build on Android seems to break now. I'm not really sure what the situation is, but if the weird common symbols are still there (can we remove them?), they probably ought to have the right flags. Change-Id: Ief589d763d16b995ac6be536505acf7596a87b30 Reviewed-on: https://boringssl-review.googlesource.com/22404 Commit-Queue: David Benjamin <davidben@google.com> Commit-Queue: Steven Valdez <svaldez@google.com> Reviewed-by: Steven Valdez <svaldez@google.com> CQ-Verified: CQ bot account: commit-bot@chromium.org <commit-bot@chromium.org>
2017-10-30 20:13:04 +00:00
.hidden OPENSSL_armcap_P
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
#endif
___
{ my %opcode = (
"sha1c" => 0xf2000c40, "sha1p" => 0xf2100c40,
"sha1m" => 0xf2200c40, "sha1su0" => 0xf2300c40,
"sha1h" => 0xf3b902c0, "sha1su1" => 0xf3ba0380 );
sub unsha1 {
my ($mnemonic,$arg)=@_;
if ($arg =~ m/q([0-9]+)(?:,\s*q([0-9]+))?,\s*q([0-9]+)/o) {
my $word = $opcode{$mnemonic}|(($1&7)<<13)|(($1&8)<<19)
|(($2&7)<<17)|(($2&8)<<4)
|(($3&7)<<1) |(($3&8)<<2);
# since ARMv7 instructions are always encoded little-endian.
# correct solution is to use .inst directive, but older
# assemblers don't implement it:-(
ARMv4 assembly pack: implement support for Thumb2. As some of ARM processors, more specifically Cortex-Mx series, are Thumb2-only, we need to support Thumb2-only builds even in assembly. (Imported from upstream's 11208dcfb9105e8afa37233185decefd45e89e17.) Change-Id: I7cb48ce6a842cf3cfdf553f6e6e6227d52d525c0 Reviewed-on: https://boringssl-review.googlesource.com/17108 Reviewed-by: Adam Langley <agl@google.com>
2017-06-12 23:31:15 +01:00
# this fix-up provides Thumb encoding in conjunction with INST
$word &= ~0x10000000 if (($word & 0x0f000000) == 0x02000000);
sprintf "INST(0x%02x,0x%02x,0x%02x,0x%02x)\t@ %s %s",
Update SHA-1/256/512 asm for ARM from upstream. This change syncs these asm files with upstream's 1.0.2 branch. The important change is that they contain ARMv8 code that allows 32-bit ARM code to use the hardware support in ARMv8 when running on such a chip. Change-Id: Id37cb1ff0cbc98a8e328612df7cf60340ca96064 Reviewed-on: https://boringssl-review.googlesource.com/2921 Reviewed-by: David Benjamin <davidben@chromium.org> Reviewed-by: Adam Langley <agl@google.com>
2015-01-16 00:06:23 +00:00
$word&0xff,($word>>8)&0xff,
($word>>16)&0xff,($word>>24)&0xff,
$mnemonic,$arg;
}
}
}
foreach (split($/,$code)) {
s/{q([0-9]+)\[\]}/sprintf "{d%d[],d%d[]}",2*$1,2*$1+1/eo or
s/{q([0-9]+)\[0\]}/sprintf "{d%d[0]}",2*$1/eo;
s/\b(sha1\w+)\s+(q.*)/unsha1($1,$2)/geo;
s/\bret\b/bx lr/o or
s/\bbx\s+lr\b/.word\t0xe12fff1e/o; # make it possible to compile with -march=armv4
print $_,$/;
}
Inital import. Initial fork from f2d678e6e89b6508147086610e985d4e8416e867 (1.0.2 beta). (This change contains substantial changes from the original and effectively starts a new history.)
2014-06-20 20:00:00 +01:00
close STDOUT; # enforce flush
Reference in New Issue Copy Permalink