aboutsummaryrefslogtreecommitdiffstats
path: root/crypto/bn/asm/vis3-mont.pl
diff options
context:
space:
mode:
Diffstat (limited to 'crypto/bn/asm/vis3-mont.pl')
-rwxr-xr-xcrypto/bn/asm/vis3-mont.pl47
1 files changed, 29 insertions, 18 deletions
diff --git a/crypto/bn/asm/vis3-mont.pl b/crypto/bn/asm/vis3-mont.pl
index 263ac02b6f45..04833a0c876d 100755
--- a/crypto/bn/asm/vis3-mont.pl
+++ b/crypto/bn/asm/vis3-mont.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /usr/bin/env perl
+# Copyright 2012-2018 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License"). You may not use
+# this file except in compliance with the License. You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+
# ====================================================================
# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
@@ -9,7 +16,7 @@
# October 2012.
#
-# SPARCv9 VIS3 Montgomery multiplicaion procedure suitable for T3 and
+# SPARCv9 VIS3 Montgomery multiplication procedure suitable for T3 and
# onward. There are three new instructions used here: umulxhi,
# addxc[cc] and initializing store. On T3 RSA private key operations
# are 1.54/1.87/2.11/2.26 times faster for 512/1024/2048/4096-bit key
@@ -18,16 +25,20 @@
# for reference purposes, because T4 has dedicated Montgomery
# multiplication and squaring *instructions* that deliver even more.
-$bits=32;
-for (@ARGV) { $bits=64 if (/\-m64/ || /\-xarch\=v9/); }
-if ($bits==64) { $bias=2047; $frame=192; }
-else { $bias=0; $frame=112; }
+$output = pop;
+open STDOUT,">$output";
+
+$frame = "STACK_FRAME";
+$bias = "STACK_BIAS";
+
+$code.=<<___;
+#include "sparc_arch.h"
-$code.=<<___ if ($bits==64);
+#ifdef __arch64__
.register %g2,#scratch
.register %g3,#scratch
-___
-$code.=<<___;
+#endif
+
.section ".text",#alloc,#execinstr
___
@@ -299,23 +310,23 @@ $code.=<<___;
sub $anp, $num, $anp
sub $rp, $num, $rp
- subc $ovf, %g0, $ovf ! handle upmost overflow bit
- and $tp, $ovf, $ap
- andn $rp, $ovf, $np
- or $np, $ap, $ap ! ap=borrow?tp:rp
+ subccc $ovf, %g0, $ovf ! handle upmost overflow bit
ba .Lcopy
sub $num, 8, $cnt
.align 16
-.Lcopy: ! copy or in-place refresh
- ld [$ap+0], $t2
- ld [$ap+4], $t3
- add $ap, 8, $ap
+.Lcopy: ! conditional copy
+ ld [$tp+0], $t0
+ ld [$tp+4], $t1
+ ld [$rp+0], $t2
+ ld [$rp+4], $t3
stx %g0, [$tp] ! zap
add $tp, 8, $tp
stx %g0, [$anp] ! zap
stx %g0, [$anp+8]
add $anp, 16, $anp
+ movcs %icc, $t0, $t2
+ movcs %icc, $t1, $t3
st $t3, [$rp+0] ! flip order
st $t2, [$rp+4]
add $rp, 8, $rp
@@ -333,7 +344,7 @@ ___
# Purpose of these subroutines is to explicitly encode VIS instructions,
# so that one can compile the module without having to specify VIS
-# extentions on compiler command line, e.g. -xarch=v9 vs. -xarch=v9a.
+# extensions on compiler command line, e.g. -xarch=v9 vs. -xarch=v9a.
# Idea is to reserve for option to produce "universal" binary and let
# programmer detect if current CPU is VIS capable at run-time.
sub unvis3 {