Merge remote-tracking branch 'origin/master'

# Conflicts: # README.md
2026-09-23 22:03:48 +08:00 · 2017-02-14 16:12:29 +08:00
parent d2254170b8
commit 43fed1108d
3503 changed files with 320546 additions and 408546 deletions
--- a/crypto/sha/asm/README
+++ b/crypto/sha/asm/README
@@ -1 +0,0 @@
-C2.pl works
--- a/crypto/sha/asm/sha1-586.pl
+++ b/crypto/sha/asm/sha1-586.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /usr/bin/env perl
+# Copyright 1998-2016 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
+

 # ====================================================================
 # [Re]written by Andy Polyakov <appro@openssl.org> for the OpenSSL
@@ -66,9 +73,9 @@
 # switch to AVX alone improves performance by as little as 4% in
 # comparison to SSSE3 code path. But below result doesn't look like
 # 4% improvement... Trouble is that Sandy Bridge decodes 'ro[rl]' as
-# pair of µ-ops, and it's the additional µ-ops, two per round, that
+# pair of µ-ops, and it's the additional µ-ops, two per round, that
 # make it run slower than Core2 and Westmere. But 'sh[rl]d' is decoded
-# as single µ-op by Sandy Bridge and it's replacing 'ro[rl]' with
+# as single µ-op by Sandy Bridge and it's replacing 'ro[rl]' with
 # equivalent 'sh[rl]d' that is responsible for the impressive 5.1
 # cycles per processed byte. But 'sh[rl]d' is not something that used
 # to be fast, nor does it appear to be fast in upcoming Bulldozer
@@ -113,6 +120,9 @@ $0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
 push(@INC,"${dir}","${dir}../../perlasm");
 require "x86asm.pl";

+$output=pop;
+open STDOUT,">$output";
+
 &asm_init($ARGV[0],"sha1-586.pl",$ARGV[$#ARGV] eq "386");

 $xmm=$ymm=0;
@@ -1474,3 +1484,5 @@ sub Xtail_avx()
 &asciz("SHA1 block transform for x86, CRYPTOGAMS by <appro\@openssl.org>");

 &asm_finish();
+
+close STDOUT;
--- a/crypto/sha/asm/sha1-alpha.pl
+++ b/crypto/sha/asm/sha1-alpha.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /usr/bin/env perl
+# Copyright 2009-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+

 # ====================================================================
 # Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
@@ -317,6 +324,6 @@ $code.=<<___;
 .ascii	"SHA1 block transform for Alpha, CRYPTOGAMS by <appro\@openssl.org>"
 .align	2
 ___
-$output=shift and open STDOUT,">$output";
+$output=pop and open STDOUT,">$output";
 print $code;
 close STDOUT;
--- a/crypto/sha/asm/sha1-armv4-large.pl
+++ b/crypto/sha/asm/sha1-armv4-large.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /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
+

 # ====================================================================
 # Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
@@ -60,14 +67,28 @@
 # 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
-# byte, which is also >80% faster than integer-only code.
+# 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.

 # May 2014.
 #
 # Add ARMv8 code path performing at 2.35 cpb on Apple A7.

-while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
-open STDOUT,">$output";
+$flavour = shift;
+if ($flavour=~/\w[\w\-]*\.\w+$/) { $output=$flavour; undef $flavour; }
+else { while (($output=shift) && ($output!~/\w[\w\-]*\.\w+$/)) {} }
+
+if ($flavour && $flavour ne "void") {
+    $0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+    ( $xlate="${dir}arm-xlate.pl" and -f $xlate ) or
+    ( $xlate="${dir}../../perlasm/arm-xlate.pl" and -f $xlate) or
+    die "can't locate arm-xlate.pl";
+
+    open STDOUT,"| \"$^X\" $xlate $flavour $output";
+} else {
+    open STDOUT,">$output";
+}

 $ctx="r0";
 $inp="r1";
@@ -167,7 +188,12 @@ $code=<<___;
 #include "arm_arch.h"

 .text
+#if defined(__thumb2__)
+.syntax	unified
+.thumb
+#else
 .code	32
+#endif

 .global	sha1_block_data_order
 .type	sha1_block_data_order,%function
@@ -175,9 +201,13 @@ $code=<<___;
 .align	5
 sha1_block_data_order:
 #if __ARM_MAX_ARCH__>=7
-	sub	r3,pc,#8		@ sha1_block_data_order
+.Lsha1_block:
+	adr	r3,.Lsha1_block
 	ldr	r12,.LOPENSSL_armcap
 	ldr	r12,[r3,r12]		@ OPENSSL_armcap_P
+#ifdef	__APPLE__
+	ldr	r12,[r12]
+#endif
 	tst	r12,#ARMV8_SHA1
 	bne	.LARMv8
 	tst	r12,#ARMV7_NEON
@@ -199,7 +229,12 @@ for($i=0;$i<5;$i++) {
 	&BODY_00_15(@V);	unshift(@V,pop(@V));
 }
 $code.=<<___;
+#if defined(__thumb2__)
+	mov	$t3,sp
+	teq	$Xi,$t3
+#else
 	teq	$Xi,sp
+#endif
 	bne	.L_00_15		@ [((11+4)*5+2)*3]
 	sub	sp,sp,#25*4
 ___
@@ -218,7 +253,12 @@ for($i=0;$i<5;$i++) {
 	&BODY_20_39(@V);	unshift(@V,pop(@V));
 }
 $code.=<<___;
+#if defined(__thumb2__)
+	mov	$t3,sp
+	teq	$Xi,$t3
+#else
 	teq	$Xi,sp			@ preserve carry
+#endif
 	bne	.L_20_39_or_60_79	@ [+((12+3)*5+2)*4]
 	bcs	.L_done			@ [+((12+3)*5+2)*4], spare 300 bytes

@@ -230,7 +270,12 @@ for($i=0;$i<5;$i++) {
 	&BODY_40_59(@V);	unshift(@V,pop(@V));
 }
 $code.=<<___;
+#if defined(__thumb2__)
+	mov	$t3,sp
+	teq	$Xi,$t3
+#else
 	teq	$Xi,sp
+#endif
 	bne	.L_40_59		@ [+((12+5)*5+2)*4]

 	ldr	$K,.LK_60_79
@@ -266,7 +311,7 @@ $code.=<<___;
 .LK_60_79:	.word	0xca62c1d6
 #if __ARM_MAX_ARCH__>=7
 .LOPENSSL_armcap:
-.word	OPENSSL_armcap_P-sha1_block_data_order
+.word	OPENSSL_armcap_P-.Lsha1_block
 #endif
 .asciz	"SHA1 block transform for ARMv4/NEON/ARMv8, CRYPTOGAMS by <appro\@openssl.org>"
 .align	5
@@ -441,6 +486,7 @@ sub Xuplast_80 ()

 	&teq		($inp,$len);
 	&sub		($K_XX_XX,$K_XX_XX,16);	# rewind $K_XX_XX
+	&it		("eq");
 	&subeq		($inp,$inp,64);		# reload last block to avoid SEGV
 	&vld1_8		("{@X[-4&7]-@X[-3&7]}","[$inp]!");
 	 eval(shift(@insns));
@@ -491,12 +537,12 @@ sha1_block_data_order_neon:
 	@ dmb				@ errata #451034 on early Cortex A8
 	@ vstmdb	sp!,{d8-d15}	@ ABI specification says so
 	mov	$saved_sp,sp
-	sub	sp,sp,#64		@ alloca
+	sub	$Xfer,sp,#64
 	adr	$K_XX_XX,.LK_00_19
-	bic	sp,sp,#15		@ align for 128-bit stores
+	bic	$Xfer,$Xfer,#15		@ align for 128-bit stores

 	ldmia	$ctx,{$a,$b,$c,$d,$e}	@ load context
-	mov	$Xfer,sp
+	mov	sp,$Xfer		@ alloca

 	vld1.8		{@X[-4&7]-@X[-3&7]},[$inp]!	@ handles unaligned
 	veor		$zero,$zero,$zero
@@ -543,10 +589,13 @@ $code.=<<___;
 	add	$b,$b,$t0
 	add	$c,$c,$t1
 	add	$d,$d,$Xfer
+	it	eq
 	moveq	sp,$saved_sp
 	add	$e,$e,$Ki
+	it	ne
 	ldrne	$Ki,[sp]
 	stmia	$ctx,{$a,$b,$c,$d,$e}
+	itt	ne
 	addne	$Xfer,sp,#3*16
 	bne	.Loop_neon

@@ -567,6 +616,13 @@ my ($W0,$W1,$ABCD_SAVE)=map("q$_",(12..14));

 $code.=<<___;
 #if __ARM_MAX_ARCH__>=7
+
+# if defined(__thumb2__)
+#  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
+
 .type	sha1_block_data_order_armv8,%function
 .align	5
 sha1_block_data_order_armv8:
@@ -660,7 +716,10 @@ ___
 	    # since ARMv7 instructions are always encoded little-endian.
 	    # correct solution is to use .inst directive, but older
 	    # assemblers don't implement it:-(
-	    sprintf ".byte\t0x%02x,0x%02x,0x%02x,0x%02x\t@ %s %s",
+
+	    # 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",
 			$word&0xff,($word>>8)&0xff,
 			($word>>16)&0xff,($word>>24)&0xff,
 			$mnemonic,$arg;
--- a/crypto/sha/asm/sha1-armv8.pl
+++ b/crypto/sha/asm/sha1-armv8.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /usr/bin/env perl
+# Copyright 2014-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+
 #
 # ====================================================================
 # Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
@@ -18,13 +25,22 @@
 # Cortex-A57	2.35			7.88 (+74%)
 # Denver	2.13			3.97 (+0%)(**)
 # X-Gene				8.80 (+200%)
+# Mongoose	2.05			6.50 (+160%)
 #
 # (*)	Software results are presented mostly for reference purposes.
 # (**)	Keep in mind that Denver relies on binary translation, which
 #	optimizes compiler output at run-time.

 $flavour = shift;
-open STDOUT,">".shift;
+$output  = shift;
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}arm-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/arm-xlate.pl" and -f $xlate) or
+die "can't locate arm-xlate.pl";
+
+open OUT,"| \"$^X\" $xlate $flavour $output";
+*STDOUT=*OUT;

 ($ctx,$inp,$num)=("x0","x1","x2");
@Xw=map("w$_",(3..17,19));
@@ -158,11 +174,16 @@ $code.=<<___;

 .text

+.extern	OPENSSL_armcap_P
 .globl	sha1_block_data_order
 .type	sha1_block_data_order,%function
 .align	6
 sha1_block_data_order:
+#ifdef	__ILP32__
+	ldrsw	x16,.LOPENSSL_armcap_P
+#else
 	ldr	x16,.LOPENSSL_armcap_P
+#endif
 	adr	x17,.LOPENSSL_armcap_P
 	add	x16,x16,x17
 	ldr	w16,[x16]
@@ -300,7 +321,11 @@ $code.=<<___;
 .long	0x8f1bbcdc,0x8f1bbcdc,0x8f1bbcdc,0x8f1bbcdc	//K_40_59
 .long	0xca62c1d6,0xca62c1d6,0xca62c1d6,0xca62c1d6	//K_60_79
 .LOPENSSL_armcap_P:
+#ifdef	__ILP32__
+.long	OPENSSL_armcap_P-.
+#else
 .quad	OPENSSL_armcap_P-.
+#endif
 .asciz	"SHA1 block transform for ARMv8, CRYPTOGAMS by <appro\@openssl.org>"
 .align	2
 .comm	OPENSSL_armcap_P,4,4
--- a/crypto/sha/asm/sha1-c64xplus.pl
+++ b/crypto/sha/asm/sha1-c64xplus.pl
@@ -0,0 +1,337 @@
+#! /usr/bin/env perl
+# Copyright 2012-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# SHA1 for C64x+.
+#
+# November 2011
+#
+# If compared to compiler-generated code with similar characteristics,
+# i.e. compiled with OPENSSL_SMALL_FOOTPRINT and utilizing SPLOOPs,
+# this implementation is 25% smaller and >2x faster. In absolute terms
+# performance is (quite impressive) ~6.5 cycles per processed byte.
+# Fully unrolled assembler would be ~5x larger and is likely to be
+# ~15% faster. It would be free from references to intermediate ring
+# buffer, but put more pressure on L1P [both because the code would be
+# larger and won't be using SPLOOP buffer]. There are no plans to
+# realize fully unrolled variant though...
+#
+# !!! Note that this module uses AMR, which means that all interrupt
+# service routines are expected to preserve it and for own well-being
+# zero it upon entry.
+
+while (($output=shift) && ($output!~/\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+($CTX,$INP,$NUM) = ("A4","B4","A6");		# arguments
+
+($A,$B,$C,$D,$E, $Arot,$F,$F0,$T,$K) = map("A$_",(16..20, 21..25));
+($X0,$X2,$X8,$X13) = ("A26","B26","A27","B27");
+($TX0,$TX1,$TX2,$TX3) = map("B$_",(28..31));
+($XPA,$XPB) = ("A5","B5");			# X circular buffer
+($Actx,$Bctx,$Cctx,$Dctx,$Ectx) = map("A$_",(3,6..9));	# zaps $NUM
+
+$code=<<___;
+	.text
+
+	.if	.ASSEMBLER_VERSION<7000000
+	.asg	0,__TI_EABI__
+	.endif
+	.if	__TI_EABI__
+	.asg	sha1_block_data_order,_sha1_block_data_order
+	.endif
+
+	.asg	B3,RA
+	.asg	A15,FP
+	.asg	B15,SP
+
+	.if	.BIG_ENDIAN
+	.asg	MV,SWAP2
+	.asg	MV,SWAP4
+	.endif
+
+	.global	_sha1_block_data_order
+_sha1_block_data_order:
+	.asmfunc stack_usage(64)
+	MV	$NUM,A0			; reassign $NUM
+||	MVK	-64,B0
+  [!A0]	BNOP	RA			; if ($NUM==0) return;
+|| [A0]	STW	FP,*SP--[16]		; save frame pointer and alloca(64)
+|| [A0]	MV	SP,FP
+   [A0]	LDW	*${CTX}[0],$A		; load A-E...
+|| [A0]	AND	B0,SP,SP		; align stack at 64 bytes
+   [A0]	LDW	*${CTX}[1],$B
+|| [A0]	SUBAW	SP,2,SP			; reserve two words above buffer
+   [A0]	LDW	*${CTX}[2],$C
+|| [A0]	MVK	0x00404,B0
+   [A0]	LDW	*${CTX}[3],$D
+|| [A0]	MVKH	0x50000,B0		; 0x050404, 64 bytes for $XP[AB]
+   [A0]	LDW	*${CTX}[4],$E
+|| [A0]	MVC	B0,AMR			; setup circular addressing
+	LDNW	*${INP}++,$TX1		; pre-fetch input
+	NOP	1
+
+loop?:
+	MVK	0x00007999,$K
+||	ADDAW	SP,2,$XPA
+||	SUB	A0,1,A0
+||	MVK	13,B0
+	MVKH	0x5a820000,$K		; K_00_19
+||	ADDAW	SP,2,$XPB
+||	MV	$A,$Actx
+||	MV	$B,$Bctx
+;;==================================================
+	SPLOOPD	5			; BODY_00_13
+||	MV	$C,$Cctx
+||	MV	$D,$Dctx
+||	MV	$E,$Ectx
+||	MVC	B0,ILC
+
+	ROTL	$A,5,$Arot
+||	AND	$C,$B,$F
+||	ANDN	$D,$B,$F0
+||	ADD	$K,$E,$T		; T=E+K
+
+	XOR	$F0,$F,$F		; F_00_19(B,C,D)
+||	MV	$D,$E			; E=D
+||	MV	$C,$D			; D=C
+||	SWAP2	$TX1,$TX2
+||	LDNW	*${INP}++,$TX1
+
+	ADD	$F,$T,$T		; T+=F_00_19(B,C,D)
+||	ROTL	$B,30,$C		; C=ROL(B,30)
+||	SWAP4	$TX2,$TX3		; byte swap
+
+	ADD	$Arot,$T,$T		; T+=ROL(A,5)
+||	MV	$A,$B			; B=A
+
+	ADD	$TX3,$T,$A		; A=T+Xi
+||	STW	$TX3,*${XPB}++
+	SPKERNEL
+;;==================================================
+	ROTL	$A,5,$Arot		; BODY_14
+||	AND	$C,$B,$F
+||	ANDN	$D,$B,$F0
+||	ADD	$K,$E,$T		; T=E+K
+
+	XOR	$F0,$F,$F		; F_00_19(B,C,D)
+||	MV	$D,$E			; E=D
+||	MV	$C,$D			; D=C
+||	SWAP2	$TX1,$TX2
+||	LDNW	*${INP}++,$TX1
+
+	ADD	$F,$T,$T		; T+=F_00_19(B,C,D)
+||	ROTL	$B,30,$C		; C=ROL(B,30)
+||	SWAP4	$TX2,$TX2		; byte swap
+||	LDW	*${XPA}++,$X0		; fetches from X ring buffer are
+||	LDW	*${XPB}[4],$X2		; 2 iterations ahead
+
+	ADD	$Arot,$T,$T		; T+=ROL(A,5)
+||	MV	$A,$B			; B=A
+||	LDW	*${XPA}[7],$X8
+||	MV	$TX3,$X13		; ||	LDW	*${XPB}[15],$X13
+||	MV	$TX2,$TX3
+
+	ADD	$TX2,$T,$A		; A=T+Xi
+||	STW	$TX2,*${XPB}++
+;;==================================================
+	ROTL	$A,5,$Arot		; BODY_15
+||	AND	$C,$B,$F
+||	ANDN	$D,$B,$F0
+||	ADD	$K,$E,$T		; T=E+K
+
+	XOR	$F0,$F,$F		; F_00_19(B,C,D)
+||	MV	$D,$E			; E=D
+||	MV	$C,$D			; D=C
+||	SWAP2	$TX1,$TX2
+
+	ADD	$F,$T,$T		; T+=F_00_19(B,C,D)
+||	ROTL	$B,30,$C		; C=ROL(B,30)
+||	SWAP4	$TX2,$TX2		; byte swap
+||	XOR	$X0,$X2,$TX0		; Xupdate XORs are 1 iteration ahead
+||	LDW	*${XPA}++,$X0
+||	LDW	*${XPB}[4],$X2
+
+	ADD	$Arot,$T,$T		; T+=ROL(A,5)
+||	MV	$A,$B			; B=A
+||	XOR	$X8,$X13,$TX1
+||	LDW	*${XPA}[7],$X8
+||	MV	$TX3,$X13		; ||	LDW	*${XPB}[15],$X13
+||	MV	$TX2,$TX3
+
+	ADD	$TX2,$T,$A		; A=T+Xi
+||	STW	$TX2,*${XPB}++
+||	XOR	$TX0,$TX1,$TX1
+||	MVK	3,B0
+;;==================================================
+	SPLOOPD	5			; BODY_16_19
+||	MVC	B0,ILC
+
+	ROTL	$A,5,$Arot
+||	AND	$C,$B,$F
+||	ANDN	$D,$B,$F0
+||	ADD	$K,$E,$T		; T=E+K
+||	ROTL	$TX1,1,$TX2		; Xupdate output
+
+	XOR	$F0,$F,$F		; F_00_19(B,C,D)
+||	MV	$D,$E			; E=D
+||	MV	$C,$D			; D=C
+
+	ADD	$F,$T,$T		; T+=F_00_19(B,C,D)
+||	ROTL	$B,30,$C		; C=ROL(B,30)
+||	XOR	$X0,$X2,$TX0
+||	LDW	*${XPA}++,$X0
+||	LDW	*${XPB}[4],$X2
+
+	ADD	$Arot,$T,$T		; T+=ROL(A,5)
+||	MV	$A,$B			; B=A
+||	XOR	$X8,$X13,$TX1
+||	LDW	*${XPA}[7],$X8
+||	MV	$TX3,$X13		; ||	LDW	*${XPB}[15],$X13
+||	MV	$TX2,$TX3
+
+	ADD	$TX2,$T,$A		; A=T+Xi
+||	STW	$TX2,*${XPB}++
+||	XOR	$TX0,$TX1,$TX1
+	SPKERNEL
+
+	MVK	0xffffeba1,$K
+||	MVK	19,B0
+	MVKH	0x6ed90000,$K		; K_20_39
+___
+sub BODY_20_39 {
+$code.=<<___;
+;;==================================================
+	SPLOOPD	5			; BODY_20_39
+||	MVC	B0,ILC
+
+	ROTL	$A,5,$Arot
+||	XOR	$B,$C,$F
+||	ADD	$K,$E,$T		; T=E+K
+||	ROTL	$TX1,1,$TX2		; Xupdate output
+
+	XOR	$D,$F,$F		; F_20_39(B,C,D)
+||	MV	$D,$E			; E=D
+||	MV	$C,$D			; D=C
+
+	ADD	$F,$T,$T		; T+=F_20_39(B,C,D)
+||	ROTL	$B,30,$C		; C=ROL(B,30)
+||	XOR	$X0,$X2,$TX0
+||	LDW	*${XPA}++,$X0
+||	LDW	*${XPB}[4],$X2
+
+	ADD	$Arot,$T,$T		; T+=ROL(A,5)
+||	MV	$A,$B			; B=A
+||	XOR	$X8,$X13,$TX1
+||	LDW	*${XPA}[7],$X8
+||	MV	$TX3,$X13		; ||	LDW	*${XPB}[15],$X13
+||	MV	$TX2,$TX3
+
+	ADD	$TX2,$T,$A		; A=T+Xi
+||	STW	$TX2,*${XPB}++		; last one is redundant
+||	XOR	$TX0,$TX1,$TX1
+	SPKERNEL
+___
+$code.=<<___ if (!shift);
+	MVK	0xffffbcdc,$K
+	MVKH	0x8f1b0000,$K		; K_40_59
+___
+}	&BODY_20_39();
+$code.=<<___;
+;;==================================================
+	SPLOOPD	5			; BODY_40_59
+||	MVC	B0,ILC
+||	AND	$B,$C,$F
+||	AND	$B,$D,$F0
+
+	ROTL	$A,5,$Arot
+||	XOR	$F0,$F,$F
+||	AND	$C,$D,$F0
+||	ADD	$K,$E,$T		; T=E+K
+||	ROTL	$TX1,1,$TX2		; Xupdate output
+
+	XOR	$F0,$F,$F		; F_40_59(B,C,D)
+||	MV	$D,$E			; E=D
+||	MV	$C,$D			; D=C
+
+	ADD	$F,$T,$T		; T+=F_40_59(B,C,D)
+||	ROTL	$B,30,$C		; C=ROL(B,30)
+||	XOR	$X0,$X2,$TX0
+||	LDW	*${XPA}++,$X0
+||	LDW	*${XPB}[4],$X2
+
+	ADD	$Arot,$T,$T		; T+=ROL(A,5)
+||	MV	$A,$B			; B=A
+||	XOR	$X8,$X13,$TX1
+||	LDW	*${XPA}[7],$X8
+||	MV	$TX3,$X13		; ||	LDW	*${XPB}[15],$X13
+||	MV	$TX2,$TX3
+
+	ADD	$TX2,$T,$A		; A=T+Xi
+||	STW	$TX2,*${XPB}++
+||	XOR	$TX0,$TX1,$TX1
+||	AND	$B,$C,$F
+||	AND	$B,$D,$F0
+	SPKERNEL
+
+	MVK	0xffffc1d6,$K
+||	MVK	18,B0
+	MVKH	0xca620000,$K		; K_60_79
+___
+	&BODY_20_39(-1);		# BODY_60_78
+$code.=<<___;
+;;==================================================
+   [A0]	B	loop?
+||	ROTL	$A,5,$Arot		; BODY_79
+||	XOR	$B,$C,$F
+||	ROTL	$TX1,1,$TX2		; Xupdate output
+
+   [A0]	LDNW	*${INP}++,$TX1		; pre-fetch input
+||	ADD	$K,$E,$T		; T=E+K
+||	XOR	$D,$F,$F		; F_20_39(B,C,D)
+
+	ADD	$F,$T,$T		; T+=F_20_39(B,C,D)
+||	ADD	$Ectx,$D,$E		; E=D,E+=Ectx
+||	ADD	$Dctx,$C,$D		; D=C,D+=Dctx
+||	ROTL	$B,30,$C		; C=ROL(B,30)
+
+	ADD	$Arot,$T,$T		; T+=ROL(A,5)
+||	ADD	$Bctx,$A,$B		; B=A,B+=Bctx
+
+	ADD	$TX2,$T,$A		; A=T+Xi
+
+	ADD	$Actx,$A,$A		; A+=Actx
+||	ADD	$Cctx,$C,$C		; C+=Cctx
+;; end of loop?
+
+	BNOP	RA			; return
+||	MV	FP,SP			; restore stack pointer
+||	LDW	*FP[0],FP		; restore frame pointer
+	STW	$A,*${CTX}[0]		; emit A-E...
+||	MVK	0,B0
+	STW	$B,*${CTX}[1]
+||	MVC	B0,AMR			; clear AMR
+	STW	$C,*${CTX}[2]
+	STW	$D,*${CTX}[3]
+	STW	$E,*${CTX}[4]
+	.endasmfunc
+
+	.sect	.const
+	.cstring "SHA1 block transform for C64x+, CRYPTOGAMS by <appro\@openssl.org>"
+	.align	4
+___
+
+print $code;
+close STDOUT;
--- a/crypto/sha/asm/sha1-ia64.pl
+++ b/crypto/sha/asm/sha1-ia64.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /usr/bin/env perl
+# Copyright 2004-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+
 #
 # ====================================================================
 # Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
@@ -14,6 +21,8 @@
 # Performance under big-endian OS such as HP-UX is 179MBps*1GHz, which
 # is >50% better than HP C and >2x better than gcc.

+$output = pop;
+
 $code=<<___;
 .ident  \"sha1-ia64.s, version 1.3\"
 .ident  \"IA-64 ISA artwork by Andy Polyakov <appro\@fy.chalmers.se>\"
@@ -301,5 +310,5 @@ $code.=<<___;
 stringz	"SHA1 block transform for IA64, CRYPTOGAMS by <appro\@openssl.org>"
 ___

-$output=shift and open STDOUT,">$output";
+open STDOUT,">$output" if $output;
 print $code;
--- a/crypto/sha/asm/sha1-mb-x86_64.pl
+++ b/crypto/sha/asm/sha1-mb-x86_64.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /usr/bin/env perl
+# Copyright 2013-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+

 # ====================================================================
 # Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
@@ -19,6 +26,7 @@
 # Sandy Bridge	(8.16	+5.15=13.3)/n	4.99	5.98		+80%
 # Ivy Bridge	(8.08	+5.14=13.2)/n	4.60	5.54		+68%
 # Haswell(iii)	(8.96	+5.00=14.0)/n	3.57	4.55		+160%
+# Skylake	(8.70	+5.00=13.7)/n	3.64	4.20		+145%
 # Bulldozer	(9.76	+5.76=15.5)/n	5.95	6.37		+64%
 #
 # (i)	multi-block CBC encrypt with 128-bit key;
@@ -58,11 +66,11 @@ if (!$avx && $win64 && ($flavour =~ /masm/ || $ENV{ASM} =~ /ml64/) &&
 	$avx = ($1>=10) + ($1>=11);
 }

-if (!$avx && `$ENV{CC} -v 2>&1` =~ /(^clang version|based on LLVM) ([3-9]\.[0-9]+)/) {
+if (!$avx && `$ENV{CC} -v 2>&1` =~ /((?:^clang|LLVM) version|.*based on LLVM) ([3-9]\.[0-9]+)/) {
 	$avx = ($2>=3.0) + ($2>3.0);
 }

-open OUT,"| \"$^X\" $xlate $flavour $output";
+open OUT,"| \"$^X\" \"$xlate\" $flavour \"$output\"";
 *STDOUT=*OUT;

 # void sha1_multi_block (
@@ -478,7 +486,7 @@ $code.=<<___;
 	jnz	.Loop_grande

 .Ldone:
-	mov	`$REG_SZ*17`(%rsp),%rax		# orignal %rsp
+	mov	`$REG_SZ*17`(%rsp),%rax		# original %rsp
 ___
 $code.=<<___ if ($win64);
 	movaps	-0xb8(%rax),%xmm6
@@ -538,7 +546,7 @@ $code.=<<___;
 	movdqa	K_XX_XX+0x80(%rip),$BSWAP	# byte-n-word swap

 .Loop_grande_shaext:
-	mov	$num,`$REG_SZ*17+8`(%rsp)	# orignal $num
+	mov	$num,`$REG_SZ*17+8`(%rsp)	# original $num
 	xor	$num,$num
 ___
 for($i=0;$i<2;$i++) {
@@ -1116,7 +1124,7 @@ $code.=<<___;
 	jnz	.Loop_grande_avx

 .Ldone_avx:
-	mov	`$REG_SZ*17`(%rsp),%rax		# orignal %rsp
+	mov	`$REG_SZ*17`(%rsp),%rax		# original %rsp
 	vzeroupper
 ___
 $code.=<<___ if ($win64);
@@ -1271,7 +1279,7 @@ $code.=<<___;
 	#jnz	.Loop_grande_avx2

 .Ldone_avx2:
-	mov	`$REG_SZ*17`(%rsp),%rax		# orignal %rsp
+	mov	`$REG_SZ*17`(%rsp),%rax		# original %rsp
 	vzeroupper
 ___
 $code.=<<___ if ($win64);
--- a/crypto/sha/asm/sha1-mips.pl
+++ b/crypto/sha/asm/sha1-mips.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /usr/bin/env perl
+# Copyright 2009-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+

 # ====================================================================
 # Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
@@ -70,7 +77,7 @@ if ($flavour =~ /64|n32/i) {

 $big_endian=(`echo MIPSEL | $ENV{CC} -E -`=~/MIPSEL/)?1:0 if ($ENV{CC});

-for (@ARGV) {	$output=$_ if (/^\w[\w\-]*\.\w+$/);   }
+for (@ARGV) {	$output=$_ if (/\w[\w\-]*\.\w+$/);   }
 open STDOUT,">$output";

 if (!defined($big_endian))
@@ -325,8 +332,8 @@ $code.=<<___ if ($i<79);
 ___
 }

-$FRAMESIZE=16;	# large enough to accomodate NUBI saved registers
-$SAVED_REGS_MASK = ($flavour =~ /nubi/i) ? 0xc0fff008 : 0xc0ff0000;
+$FRAMESIZE=16;	# large enough to accommodate NUBI saved registers
+$SAVED_REGS_MASK = ($flavour =~ /nubi/i) ? "0xc0fff008" : "0xc0ff0000";

 $code=<<___;
 #ifdef OPENSSL_FIPSCANISTER
--- a/crypto/sha/asm/sha1-parisc.pl
+++ b/crypto/sha/asm/sha1-parisc.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /usr/bin/env perl
+# Copyright 2009-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+

 # ====================================================================
 # Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
--- a/crypto/sha/asm/sha1-ppc.pl
+++ b/crypto/sha/asm/sha1-ppc.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /usr/bin/env perl
+# Copyright 2006-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+

 # ====================================================================
 # Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
@@ -37,7 +44,7 @@ if ($flavour =~ /64/) {
 	$PUSH	="stw";
 } else { die "nonsense $flavour"; }

-# Define endianess based on flavour
+# Define endianness based on flavour
 # i.e.: linux64le
 $LITTLE_ENDIAN = ($flavour=~/le$/) ? $SIZE_T : 0;

@@ -227,7 +234,7 @@ Lunaligned:
 	srwi.	$t1,$t1,6	; t1/=64
 	beq	Lcross_page
 	$UCMP	$num,$t1
-	ble-	Laligned	; didn't cross the page boundary
+	ble	Laligned	; didn't cross the page boundary
 	mtctr	$t1
 	subfc	$num,$t1,$num
 	bl	Lsha1_block_private
@@ -255,7 +262,7 @@ Lmemcpy:
 	bl	Lsha1_block_private
 	$POP	$inp,`$FRAME-$SIZE_T*18`($sp)
 	addic.	$num,$num,-1
-	bne-	Lunaligned
+	bne	Lunaligned

 Ldone:
 	$POP	r0,`$FRAME+$LRSAVE`($sp)
@@ -329,7 +336,7 @@ $code.=<<___;
 	stw	r20,16($ctx)
 	mr	$E,r20
 	addi	$inp,$inp,`16*4`
-	bdnz-	Lsha1_block_private
+	bdnz	Lsha1_block_private
 	blr
 	.long	0
 	.byte	0,12,0x14,0,0,0,0,0
--- a/crypto/sha/asm/sha1-s390x.pl
+++ b/crypto/sha/asm/sha1-s390x.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /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
+

 # ====================================================================
 # Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
@@ -28,7 +35,8 @@
 # instructions and achieve "64-bit" performance even in 31-bit legacy
 # application context. The feature is not specific to any particular
 # processor, as long as it's "z-CPU". Latter implies that the code
-# remains z/Architecture specific.
+# remains z/Architecture specific. On z990 it was measured to perform
+# 23% better than code generated by gcc 4.3.

 $kimdfunc=1;	# magic function code for kimd instruction

@@ -42,7 +50,7 @@ if ($flavour =~ /3[12]/) {
 	$g="g";
 }

-while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+while (($output=shift) && ($output!~/\w[\w\-]*\.\w+$/)) {}
 open STDOUT,">$output";

 $K_00_39="%r0"; $K=$K_00_39;
@@ -167,10 +175,7 @@ $code.=<<___ if ($kimdfunc);
 	lg	%r0,0(%r1)
 	tmhl	%r0,0x4000	# check for message-security assist
 	jz	.Lsoftware
-	lghi	%r0,0
-	la	%r1,`2*$SIZE_T`($sp)
-	.long	0xb93e0002	# kimd %r0,%r2
-	lg	%r0,`2*$SIZE_T`($sp)
+	lg	%r0,16(%r1)	# check kimd capabilities
 	tmhh	%r0,`0x8000>>$kimdfunc`
 	jz	.Lsoftware
 	lghi	%r0,$kimdfunc
@@ -237,7 +242,7 @@ $code.=<<___;
 	br	%r14
 .size	sha1_block_data_order,.-sha1_block_data_order
 .string	"SHA1 block transform for s390x, CRYPTOGAMS by <appro\@openssl.org>"
-.comm	OPENSSL_s390xcap_P,16,8
+.comm	OPENSSL_s390xcap_P,80,8
 ___

 $code =~ s/\`([^\`]*)\`/eval $1/gem;
--- a/crypto/sha/asm/sha1-sparcv9.pl
+++ b/crypto/sha/asm/sha1-sparcv9.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /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
+

 # ====================================================================
 # Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
@@ -25,7 +32,7 @@
 # single-process result on 8-core processor, or ~9GBps per 2.85GHz
 # socket.

-$output=shift;
+$output=pop;
 open STDOUT,">$output";

@X=("%o0","%o1","%o2","%o3","%o4","%o5","%g1","%o7");
@@ -368,7 +375,7 @@ ___

 # 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 unvis {
--- a/crypto/sha/asm/sha1-sparcv9a.pl
+++ b/crypto/sha/asm/sha1-sparcv9a.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /usr/bin/env perl
+# Copyright 2009-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+

 # ====================================================================
 # Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
@@ -544,7 +551,7 @@ ___

 # 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 unvis {
--- a/crypto/sha/asm/sha1-thumb.pl
+++ b/crypto/sha/asm/sha1-thumb.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /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
+

 # ====================================================================
 # Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
--- a/crypto/sha/asm/sha1-x86_64.pl
+++ b/crypto/sha/asm/sha1-x86_64.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /usr/bin/env perl
+# Copyright 2006-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+
 #
 # ====================================================================
 # Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
@@ -73,13 +80,16 @@
 # Sandy Bridge	7.70		6.10/+26%	4.99/+54%
 # Ivy Bridge	6.06		4.67/+30%	4.60/+32%
 # Haswell	5.45		4.15/+31%	3.57/+53%
+# Skylake	5.18		4.06/+28%	3.54/+46%
 # Bulldozer	9.11		5.95/+53%
 # VIA Nano	9.32		7.15/+30%
 # Atom		10.3		9.17/+12%
 # Silvermont	13.1(*)		9.37/+40%
+# Goldmont	8.13		6.42/+27%	1.70/+380%(**)
 #
 # (*)	obviously suboptimal result, nothing was done about it,
 #	because SSSE3 code is compiled unconditionally;
+# (**)	SHAEXT result

 $flavour = shift;
 $output  = shift;
@@ -107,14 +117,14 @@ if (!$avx && $win64 && ($flavour =~ /masm/ || $ENV{ASM} =~ /ml64/) &&
 	$avx = ($1>=10) + ($1>=11);
 }

-if (!$avx && `$ENV{CC} -v 2>&1` =~ /(^clang version|based on LLVM) ([2-9]\.[0-9]+)/) {
+if (!$avx && `$ENV{CC} -v 2>&1` =~ /((?:^clang|LLVM) version|.*based on LLVM) ([2-9]\.[0-9]+)/) {
 	$avx = ($2>=3.0) + ($2>3.0);
 }

 $shaext=1;	### set to zero if compiling for 1.0.1
 $avx=1		if (!$shaext && $avx);

-open OUT,"| \"$^X\" $xlate $flavour $output";
+open OUT,"| \"$^X\" \"$xlate\" $flavour \"$output\"";
 *STDOUT=*OUT;

 $ctx="%rdi";	# 1st arg
@@ -372,9 +382,9 @@ $code.=<<___;
 .align	16
 .Loop_shaext:
 	dec		$num
-	lea		0x40($inp),%rax		# next input block
+	lea		0x40($inp),%r8		# next input block
 	paddd		@MSG[0],$E
-	cmovne		%rax,$inp
+	cmovne		%r8,$inp
 	movdqa		$ABCD,$ABCD_SAVE	# offload $ABCD
 ___
 for($i=0;$i<20-4;$i+=2) {
--- a/crypto/sha/asm/sha256-586.pl
+++ b/crypto/sha/asm/sha256-586.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /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
+
 #
 # ====================================================================
 # Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
@@ -10,7 +17,7 @@
 # SHA256 block transform for x86. September 2007.
 #
 # Performance improvement over compiler generated code varies from
-# 10% to 40% [see below]. Not very impressive on some µ-archs, but
+# 10% to 40% [see below]. Not very impressive on some µ-archs, but
 # it's 5 times smaller and optimizies amount of writes.
 #
 # May 2012.
@@ -63,6 +70,9 @@ $0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
 push(@INC,"${dir}","${dir}../../perlasm");
 require "x86asm.pl";

+$output=pop;
+open STDOUT,">$output";
+
 &asm_init($ARGV[0],"sha512-586.pl",$ARGV[$#ARGV] eq "386");

 $xmm=$avx=0;
@@ -1279,3 +1289,5 @@ sub bodyx_00_15 () {			# +10%
 &function_end_B("sha256_block_data_order");

 &asm_finish();
+
+close STDOUT;
--- a/crypto/sha/asm/sha256-armv4.pl
+++ b/crypto/sha/asm/sha256-armv4.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /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
+

 # ====================================================================
 # Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
@@ -37,8 +44,20 @@
 #
 # Add ARMv8 code path performing at 2.0 cpb on Apple A7.

-while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
-open STDOUT,">$output";
+$flavour = shift;
+if ($flavour=~/\w[\w\-]*\.\w+$/) { $output=$flavour; undef $flavour; }
+else { while (($output=shift) && ($output!~/\w[\w\-]*\.\w+$/)) {} }
+
+if ($flavour && $flavour ne "void") {
+    $0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+    ( $xlate="${dir}arm-xlate.pl" and -f $xlate ) or
+    ( $xlate="${dir}../../perlasm/arm-xlate.pl" and -f $xlate) or
+    die "can't locate arm-xlate.pl";
+
+    open STDOUT,"| \"$^X\" $xlate $flavour $output";
+} else {
+    open STDOUT,">$output";
+}

 $ctx="r0";	$t0="r0";
 $inp="r1";	$t4="r1";
@@ -73,7 +92,9 @@ $code.=<<___ if ($i<16);
 	eor	$t0,$e,$e,ror#`$Sigma1[1]-$Sigma1[0]`
 	add	$a,$a,$t2			@ h+=Maj(a,b,c) from the past
 	eor	$t0,$t0,$e,ror#`$Sigma1[2]-$Sigma1[0]`	@ Sigma1(e)
+# ifndef __ARMEB__
 	rev	$t1,$t1
+# endif
 #else
 	@ ldrb	$t1,[$inp,#3]			@ $i
 	add	$a,$a,$t2			@ h+=Maj(a,b,c) from the past
@@ -161,15 +182,11 @@ $code=<<___;
 #endif

 .text
-#if __ARM_ARCH__<7
-.code	32
-#else
+#if defined(__thumb2__)
 .syntax unified
-# ifdef __thumb2__
 .thumb
-# else
+#else
 .code   32
-# endif
 #endif

 .type	K256,%object
@@ -195,21 +212,25 @@ K256:
 .word	0				@ terminator
 #if __ARM_MAX_ARCH__>=7 && !defined(__KERNEL__)
 .LOPENSSL_armcap:
-.word	OPENSSL_armcap_P-sha256_block_data_order
+.word	OPENSSL_armcap_P-.Lsha256_block_data_order
 #endif
 .align	5

 .global	sha256_block_data_order
 .type	sha256_block_data_order,%function
 sha256_block_data_order:
-#if __ARM_ARCH__<7
+.Lsha256_block_data_order:
+#if __ARM_ARCH__<7 && !defined(__thumb2__)
 	sub	r3,pc,#8		@ sha256_block_data_order
 #else
-	adr	r3,sha256_block_data_order
+	adr	r3,.Lsha256_block_data_order
 #endif
 #if __ARM_MAX_ARCH__>=7 && !defined(__KERNEL__)
 	ldr	r12,.LOPENSSL_armcap
 	ldr	r12,[r3,r12]		@ OPENSSL_armcap_P
+#ifdef	__APPLE__
+	ldr	r12,[r12]
+#endif
 	tst	r12,#ARMV8_SHA256
 	bne	.LARMv8
 	tst	r12,#ARMV7_NEON
@@ -454,7 +475,8 @@ $code.=<<___;

 .global	sha256_block_data_order_neon
 .type	sha256_block_data_order_neon,%function
-.align	4
+.align	5
+.skip	16
 sha256_block_data_order_neon:
 .LNEON:
 	stmdb	sp!,{r4-r12,lr}
@@ -580,7 +602,7 @@ my $Ktbl="r3";
 $code.=<<___;
 #if __ARM_MAX_ARCH__>=7 && !defined(__KERNEL__)

-# ifdef __thumb2__
+# if defined(__thumb2__)
 #  define INST(a,b,c,d)	.byte	c,d|0xc,a,b
 # else
 #  define INST(a,b,c,d)	.byte	a,b,c,d
@@ -591,14 +613,11 @@ $code.=<<___;
 sha256_block_data_order_armv8:
 .LARMv8:
 	vld1.32	{$ABCD,$EFGH},[$ctx]
-# ifdef __thumb2__
-	adr	$Ktbl,.LARMv8
-	sub	$Ktbl,$Ktbl,#.LARMv8-K256
-# else
-	adrl	$Ktbl,K256
-# endif
+	sub	$Ktbl,$Ktbl,#256+32
 	add	$len,$inp,$len,lsl#6	@ len to point at the end of inp
+	b	.Loop_v8

+.align	4
 .Loop_v8:
 	vld1.8		{@MSG[0]-@MSG[1]},[$inp]!
 	vld1.8		{@MSG[2]-@MSG[3]},[$inp]!
--- a/crypto/sha/asm/sha256-c64xplus.pl
+++ b/crypto/sha/asm/sha256-c64xplus.pl
@@ -0,0 +1,320 @@
+#! /usr/bin/env perl
+# Copyright 2012-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# SHA256 for C64x+.
+#
+# January 2012
+#
+# Performance is just below 10 cycles per processed byte, which is
+# almost 40% faster than compiler-generated code. Unroll is unlikely
+# to give more than ~8% improvement...
+#
+# !!! Note that this module uses AMR, which means that all interrupt
+# service routines are expected to preserve it and for own well-being
+# zero it upon entry.
+
+while (($output=shift) && ($output!~/\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+($CTXA,$INP,$NUM) = ("A4","B4","A6");            # arguments
+ $K256="A3";
+
+($A,$Actx,$B,$Bctx,$C,$Cctx,$D,$Dctx,$T2,$S0,$s1,$t0a,$t1a,$t2a,$X9,$X14)
+	=map("A$_",(16..31));
+($E,$Ectx,$F,$Fctx,$G,$Gctx,$H,$Hctx,$T1,$S1,$s0,$t0e,$t1e,$t2e,$X1,$X15)
+	=map("B$_",(16..31));
+
+($Xia,$Xib)=("A5","B5");			# circular/ring buffer
+ $CTXB=$t2e;
+
+($Xn,$X0,$K)=("B7","B8","B9");
+($Maj,$Ch)=($T2,"B6");
+
+$code.=<<___;
+	.text
+
+	.if	.ASSEMBLER_VERSION<7000000
+	.asg	0,__TI_EABI__
+	.endif
+	.if	__TI_EABI__
+	.nocmp
+	.asg	sha256_block_data_order,_sha256_block_data_order
+	.endif
+
+	.asg	B3,RA
+	.asg	A15,FP
+	.asg	B15,SP
+
+	.if	.BIG_ENDIAN
+	.asg	SWAP2,MV
+	.asg	SWAP4,MV
+	.endif
+
+	.global	_sha256_block_data_order
+_sha256_block_data_order:
+__sha256_block:
+	.asmfunc stack_usage(64)
+	MV	$NUM,A0				; reassign $NUM
+||	MVK	-64,B0
+  [!A0]	BNOP	RA				; if ($NUM==0) return;
+|| [A0]	STW	FP,*SP--[16]			; save frame pointer and alloca(64)
+|| [A0]	MV	SP,FP
+   [A0]	ADDKPC	__sha256_block,B2
+|| [A0]	AND	B0,SP,SP			; align stack at 64 bytes
+	.if	__TI_EABI__
+   [A0]	MVK	0x00404,B1
+|| [A0]	MVKL	\$PCR_OFFSET(K256,__sha256_block),$K256
+   [A0]	MVKH	0x50000,B1
+|| [A0]	MVKH	\$PCR_OFFSET(K256,__sha256_block),$K256
+	.else
+   [A0]	MVK	0x00404,B1
+|| [A0]	MVKL	(K256-__sha256_block),$K256
+   [A0]	MVKH	0x50000,B1
+|| [A0]	MVKH	(K256-__sha256_block),$K256
+	.endif
+   [A0]	MVC	B1,AMR				; setup circular addressing
+|| [A0]	MV	SP,$Xia
+   [A0]	MV	SP,$Xib
+|| [A0]	ADD	B2,$K256,$K256
+|| [A0]	MV	$CTXA,$CTXB
+|| [A0]	SUBAW	SP,2,SP				; reserve two words above buffer
+	LDW	*${CTXA}[0],$A			; load ctx
+||	LDW	*${CTXB}[4],$E
+	LDW	*${CTXA}[1],$B
+||	LDW	*${CTXB}[5],$F
+	LDW	*${CTXA}[2],$C
+||	LDW	*${CTXB}[6],$G
+	LDW	*${CTXA}[3],$D
+||	LDW	*${CTXB}[7],$H
+
+	LDNW	*$INP++,$Xn			; pre-fetch input
+	LDW	*$K256++,$K			; pre-fetch K256[0]
+	MVK	14,B0				; loop counters
+	MVK	47,B1
+||	ADDAW	$Xia,9,$Xia
+outerloop?:
+	SUB	A0,1,A0
+||	MV	$A,$Actx
+||	MV	$E,$Ectx
+||	MVD	$B,$Bctx
+||	MVD	$F,$Fctx
+	MV	$C,$Cctx
+||	MV	$G,$Gctx
+||	MVD	$D,$Dctx
+||	MVD	$H,$Hctx
+||	SWAP4	$Xn,$X0
+
+	SPLOOPD	8				; BODY_00_14
+||	MVC	B0,ILC
+||	SWAP2	$X0,$X0
+
+	LDNW	*$INP++,$Xn
+||	ROTL	$A,30,$S0
+||	OR	$A,$B,$Maj
+||	AND	$A,$B,$t2a
+||	ROTL	$E,26,$S1
+||	AND	$F,$E,$Ch
+||	ANDN	$G,$E,$t2e
+	ROTL	$A,19,$t0a
+||	AND	$C,$Maj,$Maj
+||	ROTL	$E,21,$t0e
+||	XOR	$t2e,$Ch,$Ch			; Ch(e,f,g) = (e&f)^(~e&g)
+	ROTL	$A,10,$t1a
+||	OR	$t2a,$Maj,$Maj			; Maj(a,b,c) = ((a|b)&c)|(a&b)
+||	ROTL	$E,7,$t1e
+||	ADD	$K,$H,$T1			; T1 = h + K256[i]
+	ADD	$X0,$T1,$T1			; T1 += X[i];
+||	STW	$X0,*$Xib++
+||	XOR	$t0a,$S0,$S0
+||	XOR	$t0e,$S1,$S1
+	XOR	$t1a,$S0,$S0			; Sigma0(a)
+||	XOR	$t1e,$S1,$S1			; Sigma1(e)
+||	LDW	*$K256++,$K			; pre-fetch K256[i+1]
+||	ADD	$Ch,$T1,$T1			; T1 += Ch(e,f,g)
+	ADD	$S1,$T1,$T1			; T1 += Sigma1(e)
+||	ADD	$S0,$Maj,$T2			; T2 = Sigma0(a) + Maj(a,b,c)
+||	ROTL	$G,0,$H				; h = g
+||	MV	$F,$G				; g = f
+||	MV	$X0,$X14
+||	SWAP4	$Xn,$X0
+	SWAP2	$X0,$X0
+||	MV	$E,$F				; f = e
+||	ADD	$D,$T1,$E			; e = d + T1
+||	MV	$C,$D				; d = c
+	MV	$B,$C				; c = b
+||	MV	$A,$B				; b = a
+||	ADD	$T1,$T2,$A			; a = T1 + T2
+	SPKERNEL
+
+	ROTL	$A,30,$S0			; BODY_15
+||	OR	$A,$B,$Maj
+||	AND	$A,$B,$t2a
+||	ROTL	$E,26,$S1
+||	AND	$F,$E,$Ch
+||	ANDN	$G,$E,$t2e
+||	LDW	*${Xib}[1],$Xn			; modulo-scheduled
+	ROTL	$A,19,$t0a
+||	AND	$C,$Maj,$Maj
+||	ROTL	$E,21,$t0e
+||	XOR	$t2e,$Ch,$Ch			; Ch(e,f,g) = (e&f)^(~e&g)
+||	LDW	*${Xib}[2],$X1			; modulo-scheduled
+	ROTL	$A,10,$t1a
+||	OR	$t2a,$Maj,$Maj			; Maj(a,b,c) = ((a|b)&c)|(a&b)
+||	ROTL	$E,7,$t1e
+||	ADD	$K,$H,$T1			; T1 = h + K256[i]
+	ADD	$X0,$T1,$T1			; T1 += X[i];
+||	STW	$X0,*$Xib++
+||	XOR	$t0a,$S0,$S0
+||	XOR	$t0e,$S1,$S1
+	XOR	$t1a,$S0,$S0			; Sigma0(a)
+||	XOR	$t1e,$S1,$S1			; Sigma1(e)
+||	LDW	*$K256++,$K			; pre-fetch K256[i+1]
+||	ADD	$Ch,$T1,$T1			; T1 += Ch(e,f,g)
+	ADD	$S1,$T1,$T1			; T1 += Sigma1(e)
+||	ADD	$S0,$Maj,$T2			; T2 = Sigma0(a) + Maj(a,b,c)
+||	ROTL	$G,0,$H				; h = g
+||	MV	$F,$G				; g = f
+||	MV	$X0,$X15
+	MV	$E,$F				; f = e
+||	ADD	$D,$T1,$E			; e = d + T1
+||	MV	$C,$D				; d = c
+||	MV	$Xn,$X0				; modulo-scheduled
+||	LDW	*$Xia,$X9			; modulo-scheduled
+||	ROTL	$X1,25,$t0e			; modulo-scheduled
+||	ROTL	$X14,15,$t0a			; modulo-scheduled
+	SHRU	$X1,3,$s0			; modulo-scheduled
+||	SHRU	$X14,10,$s1			; modulo-scheduled
+||	ROTL	$B,0,$C				; c = b
+||	MV	$A,$B				; b = a
+||	ADD	$T1,$T2,$A			; a = T1 + T2
+
+	SPLOOPD	10				; BODY_16_63
+||	MVC	B1,ILC
+||	ROTL	$X1,14,$t1e			; modulo-scheduled
+||	ROTL	$X14,13,$t1a			; modulo-scheduled
+
+	XOR	$t0e,$s0,$s0
+||	XOR	$t0a,$s1,$s1
+||	MV	$X15,$X14
+||	MV	$X1,$Xn
+	XOR	$t1e,$s0,$s0			; sigma0(X[i+1])
+||	XOR	$t1a,$s1,$s1			; sigma1(X[i+14])
+||	LDW	*${Xib}[2],$X1			; module-scheduled
+	ROTL	$A,30,$S0
+||	OR	$A,$B,$Maj
+||	AND	$A,$B,$t2a
+||	ROTL	$E,26,$S1
+||	AND	$F,$E,$Ch
+||	ANDN	$G,$E,$t2e
+||	ADD	$X9,$X0,$X0			; X[i] += X[i+9]
+	ROTL	$A,19,$t0a
+||	AND	$C,$Maj,$Maj
+||	ROTL	$E,21,$t0e
+||	XOR	$t2e,$Ch,$Ch			; Ch(e,f,g) = (e&f)^(~e&g)
+||	ADD	$s0,$X0,$X0			; X[i] += sigma1(X[i+1])
+	ROTL	$A,10,$t1a
+||	OR	$t2a,$Maj,$Maj			; Maj(a,b,c) = ((a|b)&c)|(a&b)
+||	ROTL	$E,7,$t1e
+||	ADD	$H,$K,$T1			; T1 = h + K256[i]
+||	ADD	$s1,$X0,$X0			; X[i] += sigma1(X[i+14])
+	XOR	$t0a,$S0,$S0
+||	XOR	$t0e,$S1,$S1
+||	ADD	$X0,$T1,$T1			; T1 += X[i]
+||	STW	$X0,*$Xib++
+	XOR	$t1a,$S0,$S0			; Sigma0(a)
+||	XOR	$t1e,$S1,$S1			; Sigma1(e)
+||	ADD	$Ch,$T1,$T1			; T1 += Ch(e,f,g)
+||	MV	$X0,$X15
+||	ROTL	$G,0,$H				; h = g
+||	LDW	*$K256++,$K			; pre-fetch K256[i+1]
+	ADD	$S1,$T1,$T1			; T1 += Sigma1(e)
+||	ADD	$S0,$Maj,$T2			; T2 = Sigma0(a) + Maj(a,b,c)
+||	MV	$F,$G				; g = f
+||	MV	$Xn,$X0				; modulo-scheduled
+||	LDW	*++$Xia,$X9			; modulo-scheduled
+||	ROTL	$X1,25,$t0e			; module-scheduled
+||	ROTL	$X14,15,$t0a			; modulo-scheduled
+	ROTL	$X1,14,$t1e			; modulo-scheduled
+||	ROTL	$X14,13,$t1a			; modulo-scheduled
+||	MV	$E,$F				; f = e
+||	ADD	$D,$T1,$E			; e = d + T1
+||	MV	$C,$D				; d = c
+||	MV	$B,$C				; c = b
+	MV	$A,$B				; b = a
+||	ADD	$T1,$T2,$A			; a = T1 + T2
+||	SHRU	$X1,3,$s0			; modulo-scheduled
+||	SHRU	$X14,10,$s1			; modulo-scheduled
+	SPKERNEL
+
+   [A0]	B	outerloop?
+|| [A0]	LDNW	*$INP++,$Xn			; pre-fetch input
+|| [A0]	ADDK	-260,$K256			; rewind K256
+||	ADD	$Actx,$A,$A			; accumulate ctx
+||	ADD	$Ectx,$E,$E
+||	ADD	$Bctx,$B,$B
+	ADD	$Fctx,$F,$F
+||	ADD	$Cctx,$C,$C
+||	ADD	$Gctx,$G,$G
+||	ADD	$Dctx,$D,$D
+||	ADD	$Hctx,$H,$H
+|| [A0]	LDW	*$K256++,$K			; pre-fetch K256[0]
+
+  [!A0]	BNOP	RA
+||[!A0]	MV	$CTXA,$CTXB
+  [!A0]	MV	FP,SP				; restore stack pointer
+||[!A0]	LDW	*FP[0],FP			; restore frame pointer
+  [!A0]	STW	$A,*${CTXA}[0]  		; save ctx
+||[!A0]	STW	$E,*${CTXB}[4]
+||[!A0]	MVK	0,B0
+  [!A0]	STW	$B,*${CTXA}[1]
+||[!A0]	STW	$F,*${CTXB}[5]
+||[!A0]	MVC	B0,AMR				; clear AMR
+	STW	$C,*${CTXA}[2]
+||	STW	$G,*${CTXB}[6]
+	STW	$D,*${CTXA}[3]
+||	STW	$H,*${CTXB}[7]
+	.endasmfunc
+
+	.if	__TI_EABI__
+	.sect	".text:sha_asm.const"
+	.else
+	.sect	".const:sha_asm"
+	.endif
+	.align	128
+K256:
+	.uword	0x428a2f98, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5
+	.uword	0x3956c25b, 0x59f111f1, 0x923f82a4, 0xab1c5ed5
+	.uword	0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3
+	.uword	0x72be5d74, 0x80deb1fe, 0x9bdc06a7, 0xc19bf174
+	.uword	0xe49b69c1, 0xefbe4786, 0x0fc19dc6, 0x240ca1cc
+	.uword	0x2de92c6f, 0x4a7484aa, 0x5cb0a9dc, 0x76f988da
+	.uword	0x983e5152, 0xa831c66d, 0xb00327c8, 0xbf597fc7
+	.uword	0xc6e00bf3, 0xd5a79147, 0x06ca6351, 0x14292967
+	.uword	0x27b70a85, 0x2e1b2138, 0x4d2c6dfc, 0x53380d13
+	.uword	0x650a7354, 0x766a0abb, 0x81c2c92e, 0x92722c85
+	.uword	0xa2bfe8a1, 0xa81a664b, 0xc24b8b70, 0xc76c51a3
+	.uword	0xd192e819, 0xd6990624, 0xf40e3585, 0x106aa070
+	.uword	0x19a4c116, 0x1e376c08, 0x2748774c, 0x34b0bcb5
+	.uword	0x391c0cb3, 0x4ed8aa4a, 0x5b9cca4f, 0x682e6ff3
+	.uword	0x748f82ee, 0x78a5636f, 0x84c87814, 0x8cc70208
+	.uword	0x90befffa, 0xa4506ceb, 0xbef9a3f7, 0xc67178f2
+	.cstring "SHA256 block transform for C64x+, CRYPTOGAMS by <appro\@openssl.org>"
+	.align	4
+
+___
+
+print $code;
+close STDOUT;
--- a/crypto/sha/asm/sha256-mb-x86_64.pl
+++ b/crypto/sha/asm/sha256-mb-x86_64.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /usr/bin/env perl
+# Copyright 2013-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+

 # ====================================================================
 # Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
@@ -19,6 +26,7 @@
 # Sandy Bridge	(20.5	+5.15=25.7)/n	11.6	13.0		+103%
 # Ivy Bridge	(20.4	+5.14=25.5)/n	10.3	11.6		+82%
 # Haswell(iii)	(21.0	+5.00=26.0)/n	7.80	8.79		+170%
+# Skylake	(18.9	+5.00=23.9)/n	7.70	8.17		+170%
 # Bulldozer	(21.6	+5.76=27.4)/n	13.6	13.7		+100%
 #
 # (i)	multi-block CBC encrypt with 128-bit key;
@@ -59,11 +67,11 @@ if (!$avx && $win64 && ($flavour =~ /masm/ || $ENV{ASM} =~ /ml64/) &&
 	$avx = ($1>=10) + ($1>=11);
 }

-if (!$avx && `$ENV{CC} -v 2>&1` =~ /(^clang version|based on LLVM) ([3-9]\.[0-9]+)/) {
+if (!$avx && `$ENV{CC} -v 2>&1` =~ /((?:^clang|LLVM) version|.*based on LLVM) ([3-9]\.[0-9]+)/) {
 	$avx = ($2>=3.0) + ($2>3.0);
 }

-open OUT,"| \"$^X\" $xlate $flavour $output";
+open OUT,"| \"$^X\" \"$xlate\" $flavour \"$output\"";
 *STDOUT=*OUT;

 # void sha256_multi_block (
@@ -382,7 +390,7 @@ $code.=<<___;
 	jnz	.Loop_grande

 .Ldone:
-	mov	`$REG_SZ*17`(%rsp),%rax		# orignal %rsp
+	mov	`$REG_SZ*17`(%rsp),%rax		# original %rsp
 ___
 $code.=<<___ if ($win64);
 	movaps	-0xb8(%rax),%xmm6
@@ -442,7 +450,7 @@ $code.=<<___;
 	lea	K256_shaext+0x80(%rip),$Tbl

 .Loop_grande_shaext:
-	mov	$num,`$REG_SZ*17+8`(%rsp)	# orignal $num
+	mov	$num,`$REG_SZ*17+8`(%rsp)	# original $num
 	xor	$num,$num
 ___
 for($i=0;$i<2;$i++) {
@@ -1061,7 +1069,7 @@ $code.=<<___;
 	jnz	.Loop_grande_avx

 .Ldone_avx:
-	mov	`$REG_SZ*17`(%rsp),%rax		# orignal %rsp
+	mov	`$REG_SZ*17`(%rsp),%rax		# original %rsp
 	vzeroupper
 ___
 $code.=<<___ if ($win64);
@@ -1237,7 +1245,7 @@ $code.=<<___;
 	#jnz	.Loop_grande_avx2

 .Ldone_avx2:
-	mov	`$REG_SZ*17`(%rsp),%rax		# orignal %rsp
+	mov	`$REG_SZ*17`(%rsp),%rax		# original %rsp
 	vzeroupper
 ___
 $code.=<<___ if ($win64);
--- a/crypto/sha/asm/sha512-586.pl
+++ b/crypto/sha/asm/sha512-586.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /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
+
 #
 # ====================================================================
 # Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
@@ -29,6 +36,7 @@
 # VIA Nano	91	-	52	33	14.7
 # Atom		126	-	68	48(***)	14.7
 # Silvermont	97	-	58	42(***)	17.5
+# Goldmont	80	-	48	19.5	12.0
 #
 # (*)	whichever best applicable.
 # (**)	x86_64 assembler performance is presented for reference
@@ -37,7 +45,7 @@
 #
 # IALU code-path is optimized for elder Pentiums. On vanilla Pentium
 # performance improvement over compiler generated code reaches ~60%,
-# while on PIII - ~35%. On newer µ-archs improvement varies from 15%
+# while on PIII - ~35%. On newer µ-archs improvement varies from 15%
 # to 50%, but it's less important as they are expected to execute SSE2
 # code-path, which is commonly ~2-3x faster [than compiler generated
 # code]. SSE2 code-path is as fast as original sha512-sse2.pl, even
@@ -50,6 +58,9 @@ $0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
 push(@INC,"${dir}","${dir}../../perlasm");
 require "x86asm.pl";

+$output=pop;
+open STDOUT,">$output";
+
 &asm_init($ARGV[0],"sha512-586.pl",$ARGV[$#ARGV] eq "386");

 $sse2=0;
@@ -909,3 +920,5 @@ sub BODY_00_15_ssse3 {		# "phase-less" copy of BODY_00_15_sse2
 &asciz("SHA512 block transform for x86, CRYPTOGAMS by <appro\@openssl.org>");

 &asm_finish();
+
+close STDOUT;
--- a/crypto/sha/asm/sha512-armv4.pl
+++ b/crypto/sha/asm/sha512-armv4.pl
@@ -1,10 +1,19 @@
-#!/usr/bin/env perl
+#! /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
+

 # ====================================================================
 # Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
 # project. The module is, however, dual licensed under OpenSSL and
 # CRYPTOGAMS licenses depending on where you obtain it. For further
 # details see http://www.openssl.org/~appro/cryptogams/.
+#
+# Permission to use under GPL terms is granted.
 # ====================================================================

 # SHA512 block procedure for ARMv4. September 2007.
@@ -34,16 +43,9 @@
 # terms it's 22.6 cycles per byte, which is disappointing result.
 # Technical writers asserted that 3-way S4 pipeline can sustain
 # multiple NEON instructions per cycle, but dual NEON issue could
-# not be observed, and for NEON-only sequences IPC(*) was found to
-# be limited by 1:-( 0.33 and 0.66 were measured for sequences with
-# ILPs(*) of 1 and 2 respectively. This in turn means that you can
-# even find yourself striving, as I did here, for achieving IPC
-# adequate to one delivered by Cortex A8 [for reference, it's
-# 0.5 for ILP of 1, and 1 for higher ILPs].
-#
-# (*) ILP, instruction-level parallelism, how many instructions
-#     *can* execute at the same time. IPC, instructions per cycle,
-#     indicates how many instructions actually execute.
+# not be observed, see http://www.openssl.org/~appro/Snapdragon-S4.html
+# for further details. On side note Cortex-A15 processes one byte in
+# 16 cycles.

 # Byte order [in]dependence. =========================================
 #
@@ -55,8 +57,20 @@ $hi="HI";
 $lo="LO";
 # ====================================================================

-while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
-open STDOUT,">$output";
+$flavour = shift;
+if ($flavour=~/\w[\w\-]*\.\w+$/) { $output=$flavour; undef $flavour; }
+else { while (($output=shift) && ($output!~/\w[\w\-]*\.\w+$/)) {} }
+
+if ($flavour && $flavour ne "void") {
+    $0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+    ( $xlate="${dir}arm-xlate.pl" and -f $xlate ) or
+    ( $xlate="${dir}../../perlasm/arm-xlate.pl" and -f $xlate) or
+    die "can't locate arm-xlate.pl";
+
+    open STDOUT,"| \"$^X\" $xlate $flavour $output";
+} else {
+    open STDOUT,">$output";
+}

 $ctx="r0";	# parameter block
 $inp="r1";
@@ -143,6 +157,9 @@ $code.=<<___;
 	teq	$t0,#$magic

 	ldr	$t3,[sp,#$Coff+0]	@ c.lo
+#if __ARM_ARCH__>=7
+	it	eq			@ Thumb2 thing, sanity check in ARM
+#endif
 	orreq	$Ktbl,$Ktbl,#1
 	@ Sigma0(x)	(ROTR((x),28) ^ ROTR((x),34) ^ ROTR((x),39))
 	@ LO		lo>>28^hi<<4  ^ hi>>2^lo<<30 ^ hi>>7^lo<<25
@@ -180,7 +197,17 @@ $code.=<<___;
 ___
 }
 $code=<<___;
-#include "arm_arch.h"
+#ifndef __KERNEL__
+# include "arm_arch.h"
+# define VFP_ABI_PUSH	vstmdb	sp!,{d8-d15}
+# define VFP_ABI_POP	vldmia	sp!,{d8-d15}
+#else
+# define __ARM_ARCH__ __LINUX_ARM_ARCH__
+# define __ARM_MAX_ARCH__ 7
+# define VFP_ABI_PUSH
+# define VFP_ABI_POP
+#endif
+
 #ifdef __ARMEL__
 # define LO 0
 # define HI 4
@@ -192,7 +219,14 @@ $code=<<___;
 #endif

 .text
+#if defined(__thumb2__)
+.syntax unified
+.thumb
+# define adrl adr
+#else
 .code	32
+#endif
+
 .type	K512,%object
 .align	5
 K512:
@@ -237,9 +271,9 @@ WORD64(0x3c9ebe0a,0x15c9bebc, 0x431d67c4,0x9c100d4c)
 WORD64(0x4cc5d4be,0xcb3e42b6, 0x597f299c,0xfc657e2a)
 WORD64(0x5fcb6fab,0x3ad6faec, 0x6c44198c,0x4a475817)
 .size	K512,.-K512
-#if __ARM_MAX_ARCH__>=7
+#if __ARM_MAX_ARCH__>=7 && !defined(__KERNEL__)
 .LOPENSSL_armcap:
-.word	OPENSSL_armcap_P-sha512_block_data_order
+.word	OPENSSL_armcap_P-.Lsha512_block_data_order
 .skip	32-4
 #else
 .skip	32
@@ -248,14 +282,22 @@ WORD64(0x5fcb6fab,0x3ad6faec, 0x6c44198c,0x4a475817)
 .global	sha512_block_data_order
 .type	sha512_block_data_order,%function
 sha512_block_data_order:
+.Lsha512_block_data_order:
+#if __ARM_ARCH__<7 && !defined(__thumb2__)
 	sub	r3,pc,#8		@ sha512_block_data_order
-	add	$len,$inp,$len,lsl#7	@ len to point at the end of inp
-#if __ARM_MAX_ARCH__>=7
+#else
+	adr	r3,.Lsha512_block_data_order
+#endif
+#if __ARM_MAX_ARCH__>=7 && !defined(__KERNEL__)
 	ldr	r12,.LOPENSSL_armcap
 	ldr	r12,[r3,r12]		@ OPENSSL_armcap_P
-	tst	r12,#1
+#ifdef	__APPLE__
+	ldr	r12,[r12]
+#endif
+	tst	r12,#ARMV7_NEON
 	bne	.LNEON
 #endif
+	add	$len,$inp,$len,lsl#7	@ len to point at the end of inp
 	stmdb	sp!,{r4-r12,lr}
 	sub	$Ktbl,r3,#672		@ K512
 	sub	sp,sp,#9*8
@@ -369,6 +411,9 @@ $code.=<<___;
 ___
 	&BODY_00_15(0x17);
 $code.=<<___;
+#if __ARM_ARCH__>=7
+	ittt	eq			@ Thumb2 thing, sanity check in ARM
+#endif
 	ldreq	$t0,[sp,#`$Xoff+8*(16-1)`+0]
 	ldreq	$t1,[sp,#`$Xoff+8*(16-1)`+4]
 	beq	.L16_79
@@ -453,6 +498,7 @@ $code.=<<___;
 	moveq	pc,lr			@ be binary compatible with V4, yet
 	bx	lr			@ interoperable with Thumb ISA:-)
 #endif
+.size	sha512_block_data_order,.-sha512_block_data_order
 ___

 {
@@ -559,11 +605,15 @@ $code.=<<___;
 .arch	armv7-a
 .fpu	neon

+.global	sha512_block_data_order_neon
+.type	sha512_block_data_order_neon,%function
 .align	4
+sha512_block_data_order_neon:
 .LNEON:
 	dmb				@ errata #451034 on early Cortex A8
-	vstmdb	sp!,{d8-d15}		@ ABI specification says so
-	sub	$Ktbl,r3,#672		@ K512
+	add	$len,$inp,$len,lsl#7	@ len to point at the end of inp
+	adr	$Ktbl,K512
+	VFP_ABI_PUSH
 	vldmia	$ctx,{$A-$H}		@ load context
 .Loop_neon:
 ___
@@ -588,16 +638,16 @@ $code.=<<___;
 	sub		$Ktbl,#640	@ rewind K512
 	bne		.Loop_neon

-	vldmia	sp!,{d8-d15}		@ epilogue
+	VFP_ABI_POP
 	ret				@ bx lr
+.size	sha512_block_data_order_neon,.-sha512_block_data_order_neon
 #endif
 ___
 }
 $code.=<<___;
-.size	sha512_block_data_order,.-sha512_block_data_order
 .asciz	"SHA512 block transform for ARMv4/NEON, CRYPTOGAMS by <appro\@openssl.org>"
 .align	2
-#if __ARM_MAX_ARCH__>=7
+#if __ARM_MAX_ARCH__>=7 && !defined(__KERNEL__)
 .comm	OPENSSL_armcap_P,4,4
 #endif
 ___
@@ -605,5 +655,14 @@ ___
 $code =~ s/\`([^\`]*)\`/eval $1/gem;
 $code =~ s/\bbx\s+lr\b/.word\t0xe12fff1e/gm;	# make it possible to compile with -march=armv4
 $code =~ s/\bret\b/bx	lr/gm;
+
+open SELF,$0;
+while(<SELF>) {
+	next if (/^#!/);
+	last if (!s/^#/@/ and !/^$/);
+	print;
+}
+close SELF;
+
 print $code;
 close STDOUT; # enforce flush
--- a/crypto/sha/asm/sha512-armv8.pl
+++ b/crypto/sha/asm/sha512-armv8.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /usr/bin/env perl
+# Copyright 2014-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+
 #
 # ====================================================================
 # Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
@@ -18,6 +25,7 @@
 # Cortex-A57	2.31		11.6 (+86%)	7.51 (+260%(***))
 # Denver	2.01		10.5 (+26%)	6.70 (+8%)
 # X-Gene			20.0 (+100%)	12.8 (+300%(***))
+# Mongoose	2.36		13.0 (+50%)	8.36 (+33%)
 # 
 # (*)	Software SHA256 results are of lesser relevance, presented
 #	mostly for informational purposes.
@@ -31,7 +39,14 @@

 $flavour=shift;
 $output=shift;
-open STDOUT,">$output";
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}arm-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/arm-xlate.pl" and -f $xlate) or
+die "can't locate arm-xlate.pl";
+
+open OUT,"| \"$^X\" $xlate $flavour $output";
+*STDOUT=*OUT;

 if ($output =~ /512/) {
 	$BITS=512;
@@ -155,13 +170,18 @@ $code.=<<___;

 .text

+.extern	OPENSSL_armcap_P
 .globl	$func
 .type	$func,%function
 .align	6
 $func:
 ___
 $code.=<<___	if ($SZ==4);
+#ifdef	__ILP32__
+	ldrsw	x16,.LOPENSSL_armcap_P
+#else
 	ldr	x16,.LOPENSSL_armcap_P
+#endif
 	adr	x17,.LOPENSSL_armcap_P
 	add	x16,x16,x17
 	ldr	w16,[x16]
@@ -184,7 +204,7 @@ $code.=<<___;
 	ldp	$E,$F,[$ctx,#4*$SZ]
 	add	$num,$inp,$num,lsl#`log(16*$SZ)/log(2)`	// end of input
 	ldp	$G,$H,[$ctx,#6*$SZ]
-	adr	$Ktbl,K$BITS
+	adr	$Ktbl,.LK$BITS
 	stp	$ctx,$num,[x29,#96]

 .Loop:
@@ -234,8 +254,8 @@ $code.=<<___;
 .size	$func,.-$func

 .align	6
-.type	K$BITS,%object
-K$BITS:
+.type	.LK$BITS,%object
+.LK$BITS:
 ___
 $code.=<<___ if ($SZ==8);
 	.quad	0x428a2f98d728ae22,0x7137449123ef65cd
@@ -300,10 +320,14 @@ $code.=<<___ if ($SZ==4);
 	.long	0	//terminator
 ___
 $code.=<<___;
-.size	K$BITS,.-K$BITS
+.size	.LK$BITS,.-.LK$BITS
 .align	3
 .LOPENSSL_armcap_P:
+#ifdef	__ILP32__
+	.long	OPENSSL_armcap_P-.
+#else
 	.quad	OPENSSL_armcap_P-.
+#endif
 .asciz	"SHA$BITS block transform for ARMv8, CRYPTOGAMS by <appro\@openssl.org>"
 .align	2
 ___
@@ -325,7 +349,7 @@ sha256_block_armv8:
 	add		x29,sp,#0

 	ld1.32		{$ABCD,$EFGH},[$ctx]
-	adr		$Ktbl,K256
+	adr		$Ktbl,.LK256

 .Loop_hw:
 	ld1		{@MSG[0]-@MSG[3]},[$inp],#64
--- a/crypto/sha/asm/sha512-c64xplus.pl
+++ b/crypto/sha/asm/sha512-c64xplus.pl
@@ -0,0 +1,438 @@
+#! /usr/bin/env perl
+# Copyright 2012-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# SHA512 for C64x+.
+#
+# January 2012
+#
+# Performance is 19 cycles per processed byte. Compared to block
+# transform function from sha512.c compiled with cl6x with -mv6400+
+# -o2 -DOPENSSL_SMALL_FOOTPRINT it's almost 7x faster and 2x smaller.
+# Loop unroll won't make it, this implementation, any faster, because
+# it's effectively dominated by SHRU||SHL pairs and you can't schedule
+# more of them.
+#
+# !!! Note that this module uses AMR, which means that all interrupt
+# service routines are expected to preserve it and for own well-being
+# zero it upon entry.
+
+while (($output=shift) && ($output!~/\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+($CTXA,$INP,$NUM) = ("A4","B4","A6");            # arguments
+ $K512="A3";
+
+($Ahi,$Actxhi,$Bhi,$Bctxhi,$Chi,$Cctxhi,$Dhi,$Dctxhi,
+ $Ehi,$Ectxhi,$Fhi,$Fctxhi,$Ghi,$Gctxhi,$Hhi,$Hctxhi)=map("A$_",(16..31));
+($Alo,$Actxlo,$Blo,$Bctxlo,$Clo,$Cctxlo,$Dlo,$Dctxlo,
+ $Elo,$Ectxlo,$Flo,$Fctxlo,$Glo,$Gctxlo,$Hlo,$Hctxlo)=map("B$_",(16..31));
+
+($S1hi,$CHhi,$S0hi,$t0hi)=map("A$_",(10..13));
+($S1lo,$CHlo,$S0lo,$t0lo)=map("B$_",(10..13));
+($T1hi,         $T2hi)=         ("A6","A7");
+($T1lo,$T1carry,$T2lo,$T2carry)=("B6","B7","B8","B9");
+($Khi,$Klo)=("A9","A8");
+($MAJhi,$MAJlo)=($T2hi,$T2lo);
+($t1hi,$t1lo)=($Khi,"B2");
+ $CTXB=$t1lo;
+
+($Xihi,$Xilo)=("A5","B5");			# circular/ring buffer
+
+$code.=<<___;
+	.text
+
+	.if	.ASSEMBLER_VERSION<7000000
+	.asg	0,__TI_EABI__
+	.endif
+	.if	__TI_EABI__
+	.nocmp
+	.asg	sha512_block_data_order,_sha512_block_data_order
+	.endif
+
+	.asg	B3,RA
+	.asg	A15,FP
+	.asg	B15,SP
+
+	.if	.BIG_ENDIAN
+	.asg	$Khi,KHI
+	.asg	$Klo,KLO
+	.else
+	.asg	$Khi,KLO
+	.asg	$Klo,KHI
+	.endif
+
+	.global	_sha512_block_data_order
+_sha512_block_data_order:
+__sha512_block:
+	.asmfunc stack_usage(40+128)
+	MV	$NUM,A0				; reassign $NUM
+||	MVK	-128,B0
+  [!A0]	BNOP	RA				; if ($NUM==0) return;
+|| [A0]	STW	FP,*SP--(40)			; save frame pointer
+|| [A0]	MV	SP,FP
+   [A0]	STDW	B13:B12,*SP[4]
+|| [A0]	MVK	0x00404,B1
+   [A0]	STDW	B11:B10,*SP[3]
+|| [A0]	STDW	A13:A12,*FP[-3]
+|| [A0]	MVKH	0x60000,B1
+   [A0]	STDW	A11:A10,*SP[1]
+|| [A0]	MVC	B1,AMR				; setup circular addressing
+|| [A0]	ADD	B0,SP,SP			; alloca(128)
+	.if	__TI_EABI__
+   [A0]	AND	B0,SP,SP			; align stack at 128 bytes
+|| [A0]	ADDKPC	__sha512_block,B1
+|| [A0]	MVKL	\$PCR_OFFSET(K512,__sha512_block),$K512
+   [A0]	MVKH	\$PCR_OFFSET(K512,__sha512_block),$K512
+|| [A0]	SUBAW	SP,2,SP				; reserve two words above buffer
+	.else
+   [A0]	AND	B0,SP,SP			; align stack at 128 bytes
+|| [A0]	ADDKPC	__sha512_block,B1
+|| [A0]	MVKL	(K512-__sha512_block),$K512
+   [A0]	MVKH	(K512-__sha512_block),$K512
+|| [A0]	SUBAW	SP,2,SP				; reserve two words above buffer
+	.endif
+	ADDAW	SP,3,$Xilo
+	ADDAW	SP,2,$Xihi
+
+||	MV	$CTXA,$CTXB
+	LDW	*${CTXA}[0^.LITTLE_ENDIAN],$Ahi	; load ctx
+||	LDW	*${CTXB}[1^.LITTLE_ENDIAN],$Alo
+||	ADD	B1,$K512,$K512
+	LDW	*${CTXA}[2^.LITTLE_ENDIAN],$Bhi
+||	LDW	*${CTXB}[3^.LITTLE_ENDIAN],$Blo
+	LDW	*${CTXA}[4^.LITTLE_ENDIAN],$Chi
+||	LDW	*${CTXB}[5^.LITTLE_ENDIAN],$Clo
+	LDW	*${CTXA}[6^.LITTLE_ENDIAN],$Dhi
+||	LDW	*${CTXB}[7^.LITTLE_ENDIAN],$Dlo
+	LDW	*${CTXA}[8^.LITTLE_ENDIAN],$Ehi
+||	LDW	*${CTXB}[9^.LITTLE_ENDIAN],$Elo
+	LDW	*${CTXA}[10^.LITTLE_ENDIAN],$Fhi
+||	LDW	*${CTXB}[11^.LITTLE_ENDIAN],$Flo
+	LDW	*${CTXA}[12^.LITTLE_ENDIAN],$Ghi
+||	LDW	*${CTXB}[13^.LITTLE_ENDIAN],$Glo
+	LDW	*${CTXA}[14^.LITTLE_ENDIAN],$Hhi
+||	LDW	*${CTXB}[15^.LITTLE_ENDIAN],$Hlo
+
+	LDNDW	*$INP++,B11:B10			; pre-fetch input
+	LDDW	*$K512++,$Khi:$Klo		; pre-fetch K512[0]
+outerloop?:
+	MVK	15,B0				; loop counters
+||	MVK	64,B1
+||	SUB	A0,1,A0
+	MV	$Ahi,$Actxhi
+||	MV	$Alo,$Actxlo
+||	MV	$Bhi,$Bctxhi
+||	MV	$Blo,$Bctxlo
+||	MV	$Chi,$Cctxhi
+||	MV	$Clo,$Cctxlo
+||	MVD	$Dhi,$Dctxhi
+||	MVD	$Dlo,$Dctxlo
+	MV	$Ehi,$Ectxhi
+||	MV	$Elo,$Ectxlo
+||	MV	$Fhi,$Fctxhi
+||	MV	$Flo,$Fctxlo
+||	MV	$Ghi,$Gctxhi
+||	MV	$Glo,$Gctxlo
+||	MVD	$Hhi,$Hctxhi
+||	MVD	$Hlo,$Hctxlo
+loop0_15?:
+	.if	.BIG_ENDIAN
+	MV	B11,$T1hi
+||	MV	B10,$T1lo
+	.else
+	SWAP4	B10,$T1hi
+||	SWAP4	B11,$T1lo
+	SWAP2	$T1hi,$T1hi
+||	SWAP2	$T1lo,$T1lo
+	.endif
+loop16_79?:
+	STW	$T1hi,*$Xihi++[2]
+||	STW	$T1lo,*$Xilo++[2]			; X[i] = T1
+||	ADD	$Hhi,$T1hi,$T1hi
+||	ADDU	$Hlo,$T1lo,$T1carry:$T1lo		; T1 += h
+||	SHRU	$Ehi,14,$S1hi
+||	SHL	$Ehi,32-14,$S1lo
+	XOR	$Fhi,$Ghi,$CHhi
+||	XOR	$Flo,$Glo,$CHlo
+||	ADD	KHI,$T1hi,$T1hi
+||	ADDU	KLO,$T1carry:$T1lo,$T1carry:$T1lo	; T1 += K512[i]
+||	SHRU	$Elo,14,$t0lo
+||	SHL	$Elo,32-14,$t0hi
+	XOR	$t0hi,$S1hi,$S1hi
+||	XOR	$t0lo,$S1lo,$S1lo
+||	AND	$Ehi,$CHhi,$CHhi
+||	AND	$Elo,$CHlo,$CHlo
+||	ROTL	$Ghi,0,$Hhi
+||	ROTL	$Glo,0,$Hlo				; h = g
+||	SHRU	$Ehi,18,$t0hi
+||	SHL	$Ehi,32-18,$t0lo
+	XOR	$t0hi,$S1hi,$S1hi
+||	XOR	$t0lo,$S1lo,$S1lo
+||	XOR	$Ghi,$CHhi,$CHhi
+||	XOR	$Glo,$CHlo,$CHlo			; Ch(e,f,g) = ((f^g)&e)^g
+||	ROTL	$Fhi,0,$Ghi
+||	ROTL	$Flo,0,$Glo				; g = f
+||	SHRU	$Elo,18,$t0lo
+||	SHL	$Elo,32-18,$t0hi
+	XOR	$t0hi,$S1hi,$S1hi
+||	XOR	$t0lo,$S1lo,$S1lo
+||	OR	$Ahi,$Bhi,$MAJhi
+||	OR	$Alo,$Blo,$MAJlo
+||	ROTL	$Ehi,0,$Fhi
+||	ROTL	$Elo,0,$Flo				; f = e
+||	SHRU	$Ehi,41-32,$t0lo
+||	SHL	$Ehi,64-41,$t0hi
+	XOR	$t0hi,$S1hi,$S1hi
+||	XOR	$t0lo,$S1lo,$S1lo
+||	AND	$Chi,$MAJhi,$MAJhi
+||	AND	$Clo,$MAJlo,$MAJlo
+||	ROTL	$Dhi,0,$Ehi
+||	ROTL	$Dlo,0,$Elo				; e = d
+||	SHRU	$Elo,41-32,$t0hi
+||	SHL	$Elo,64-41,$t0lo
+	XOR	$t0hi,$S1hi,$S1hi
+||	XOR	$t0lo,$S1lo,$S1lo			; Sigma1(e)
+||	AND	$Ahi,$Bhi,$t1hi
+||	AND	$Alo,$Blo,$t1lo
+||	ROTL	$Chi,0,$Dhi
+||	ROTL	$Clo,0,$Dlo				; d = c
+||	SHRU	$Ahi,28,$S0hi
+||	SHL	$Ahi,32-28,$S0lo
+	OR	$t1hi,$MAJhi,$MAJhi
+||	OR	$t1lo,$MAJlo,$MAJlo			; Maj(a,b,c) = ((a|b)&c)|(a&b)
+||	ADD	$CHhi,$T1hi,$T1hi
+||	ADDU	$CHlo,$T1carry:$T1lo,$T1carry:$T1lo	; T1 += Ch(e,f,g)
+||	ROTL	$Bhi,0,$Chi
+||	ROTL	$Blo,0,$Clo				; c = b
+||	SHRU	$Alo,28,$t0lo
+||	SHL	$Alo,32-28,$t0hi
+	XOR	$t0hi,$S0hi,$S0hi
+||	XOR	$t0lo,$S0lo,$S0lo
+||	ADD	$S1hi,$T1hi,$T1hi
+||	ADDU	$S1lo,$T1carry:$T1lo,$T1carry:$T1lo	; T1 += Sigma1(e)
+||	ROTL	$Ahi,0,$Bhi
+||	ROTL	$Alo,0,$Blo				; b = a
+||	SHRU	$Ahi,34-32,$t0lo
+||	SHL	$Ahi,64-34,$t0hi
+	XOR	$t0hi,$S0hi,$S0hi
+||	XOR	$t0lo,$S0lo,$S0lo
+||	ADD	$MAJhi,$T1hi,$T2hi
+||	ADDU	$MAJlo,$T1carry:$T1lo,$T2carry:$T2lo	; T2 = T1+Maj(a,b,c)
+||	SHRU	$Alo,34-32,$t0hi
+||	SHL	$Alo,64-34,$t0lo
+	XOR	$t0hi,$S0hi,$S0hi
+||	XOR	$t0lo,$S0lo,$S0lo
+||	ADD	$Ehi,$T1hi,$T1hi
+||	ADDU	$Elo,$T1carry:$T1lo,$T1carry:$T1lo	; T1 += e
+|| [B0]	BNOP	loop0_15?
+||	SHRU	$Ahi,39-32,$t0lo
+||	SHL	$Ahi,64-39,$t0hi
+	XOR	$t0hi,$S0hi,$S0hi
+||	XOR	$t0lo,$S0lo,$S0lo
+|| [B0]	LDNDW	*$INP++,B11:B10				; pre-fetch input
+||[!B1]	BNOP	break?
+||	SHRU	$Alo,39-32,$t0hi
+||	SHL	$Alo,64-39,$t0lo
+	XOR	$t0hi,$S0hi,$S0hi
+||	XOR	$t0lo,$S0lo,$S0lo			; Sigma0(a)
+||	ADD	$T1carry,$T1hi,$Ehi
+||	MV	$T1lo,$Elo				; e = T1
+||[!B0]	LDW	*${Xihi}[28],$T1hi
+||[!B0]	LDW	*${Xilo}[28],$T1lo			; X[i+14]
+	ADD	$S0hi,$T2hi,$T2hi
+||	ADDU	$S0lo,$T2carry:$T2lo,$T2carry:$T2lo	; T2 += Sigma0(a)
+|| [B1]	LDDW	*$K512++,$Khi:$Klo			; pre-fetch K512[i]
+	NOP						; avoid cross-path stall
+	ADD	$T2carry,$T2hi,$Ahi
+||	MV	$T2lo,$Alo				; a = T2
+|| [B0]	SUB	B0,1,B0
+;;===== branch to loop00_15? is taken here
+	NOP
+;;===== branch to break? is taken here
+	LDW	*${Xihi}[2],$T2hi
+||	LDW	*${Xilo}[2],$T2lo			; X[i+1]
+||	SHRU	$T1hi,19,$S1hi
+||	SHL	$T1hi,32-19,$S1lo
+	SHRU	$T1lo,19,$t0lo
+||	SHL	$T1lo,32-19,$t0hi
+	XOR	$t0hi,$S1hi,$S1hi
+||	XOR	$t0lo,$S1lo,$S1lo
+||	SHRU	$T1hi,61-32,$t0lo
+||	SHL	$T1hi,64-61,$t0hi
+	XOR	$t0hi,$S1hi,$S1hi
+||	XOR	$t0lo,$S1lo,$S1lo
+||	SHRU	$T1lo,61-32,$t0hi
+||	SHL	$T1lo,64-61,$t0lo
+	XOR	$t0hi,$S1hi,$S1hi
+||	XOR	$t0lo,$S1lo,$S1lo
+||	SHRU	$T1hi,6,$t0hi
+||	SHL	$T1hi,32-6,$t0lo
+	XOR	$t0hi,$S1hi,$S1hi
+||	XOR	$t0lo,$S1lo,$S1lo
+||	SHRU	$T1lo,6,$t0lo
+||	LDW	*${Xihi}[18],$T1hi
+||	LDW	*${Xilo}[18],$T1lo			; X[i+9]
+	XOR	$t0lo,$S1lo,$S1lo			; sigma1(Xi[i+14])
+
+||	LDW	*${Xihi}[0],$CHhi
+||	LDW	*${Xilo}[0],$CHlo			; X[i]
+||	SHRU	$T2hi,1,$S0hi
+||	SHL	$T2hi,32-1,$S0lo
+	SHRU	$T2lo,1,$t0lo
+||	SHL	$T2lo,32-1,$t0hi
+	XOR	$t0hi,$S0hi,$S0hi
+||	XOR	$t0lo,$S0lo,$S0lo
+||	SHRU	$T2hi,8,$t0hi
+||	SHL	$T2hi,32-8,$t0lo
+	XOR	$t0hi,$S0hi,$S0hi
+||	XOR	$t0lo,$S0lo,$S0lo
+||	SHRU	$T2lo,8,$t0lo
+||	SHL	$T2lo,32-8,$t0hi
+	XOR	$t0hi,$S0hi,$S0hi
+||	XOR	$t0lo,$S0lo,$S0lo
+||	ADD	$S1hi,$T1hi,$T1hi
+||	ADDU	$S1lo,$T1lo,$T1carry:$T1lo		; T1 = X[i+9]+sigma1()
+|| [B1]	BNOP	loop16_79?
+||	SHRU	$T2hi,7,$t0hi
+||	SHL	$T2hi,32-7,$t0lo
+	XOR	$t0hi,$S0hi,$S0hi
+||	XOR	$t0lo,$S0lo,$S0lo
+||	ADD	$CHhi,$T1hi,$T1hi
+||	ADDU	$CHlo,$T1carry:$T1lo,$T1carry:$T1lo	; T1 += X[i]
+||	SHRU	$T2lo,7,$t0lo
+	XOR	$t0lo,$S0lo,$S0lo			; sigma0(Xi[i+1]
+
+	ADD	$S0hi,$T1hi,$T1hi
+||	ADDU	$S0lo,$T1carry:$T1lo,$T1carry:$T1lo	; T1 += sigma0()
+|| [B1]	SUB	B1,1,B1
+	NOP						; avoid cross-path stall
+	ADD	$T1carry,$T1hi,$T1hi
+;;===== branch to loop16_79? is taken here
+
+break?:
+	ADD	$Ahi,$Actxhi,$Ahi		; accumulate ctx
+||	ADDU	$Alo,$Actxlo,$Actxlo:$Alo
+|| [A0]	LDNDW	*$INP++,B11:B10			; pre-fetch input
+|| [A0]	ADDK	-640,$K512			; rewind pointer to K512
+	ADD	$Bhi,$Bctxhi,$Bhi
+||	ADDU	$Blo,$Bctxlo,$Bctxlo:$Blo
+|| [A0]	LDDW	*$K512++,$Khi:$Klo		; pre-fetch K512[0]
+	ADD	$Chi,$Cctxhi,$Chi
+||	ADDU	$Clo,$Cctxlo,$Cctxlo:$Clo
+||	ADD	$Actxlo,$Ahi,$Ahi
+||[!A0]	MV	$CTXA,$CTXB
+	ADD	$Dhi,$Dctxhi,$Dhi
+||	ADDU	$Dlo,$Dctxlo,$Dctxlo:$Dlo
+||	ADD	$Bctxlo,$Bhi,$Bhi
+||[!A0]	STW	$Ahi,*${CTXA}[0^.LITTLE_ENDIAN]	; save ctx
+||[!A0]	STW	$Alo,*${CTXB}[1^.LITTLE_ENDIAN]
+	ADD	$Ehi,$Ectxhi,$Ehi
+||	ADDU	$Elo,$Ectxlo,$Ectxlo:$Elo
+||	ADD	$Cctxlo,$Chi,$Chi
+|| [A0]	BNOP	outerloop?
+||[!A0]	STW	$Bhi,*${CTXA}[2^.LITTLE_ENDIAN]
+||[!A0]	STW	$Blo,*${CTXB}[3^.LITTLE_ENDIAN]
+	ADD	$Fhi,$Fctxhi,$Fhi
+||	ADDU	$Flo,$Fctxlo,$Fctxlo:$Flo
+||	ADD	$Dctxlo,$Dhi,$Dhi
+||[!A0]	STW	$Chi,*${CTXA}[4^.LITTLE_ENDIAN]
+||[!A0]	STW	$Clo,*${CTXB}[5^.LITTLE_ENDIAN]
+	ADD	$Ghi,$Gctxhi,$Ghi
+||	ADDU	$Glo,$Gctxlo,$Gctxlo:$Glo
+||	ADD	$Ectxlo,$Ehi,$Ehi
+||[!A0]	STW	$Dhi,*${CTXA}[6^.LITTLE_ENDIAN]
+||[!A0]	STW	$Dlo,*${CTXB}[7^.LITTLE_ENDIAN]
+	ADD	$Hhi,$Hctxhi,$Hhi
+||	ADDU	$Hlo,$Hctxlo,$Hctxlo:$Hlo
+||	ADD	$Fctxlo,$Fhi,$Fhi
+||[!A0]	STW	$Ehi,*${CTXA}[8^.LITTLE_ENDIAN]
+||[!A0]	STW	$Elo,*${CTXB}[9^.LITTLE_ENDIAN]
+	ADD	$Gctxlo,$Ghi,$Ghi
+||[!A0]	STW	$Fhi,*${CTXA}[10^.LITTLE_ENDIAN]
+||[!A0]	STW	$Flo,*${CTXB}[11^.LITTLE_ENDIAN]
+	ADD	$Hctxlo,$Hhi,$Hhi
+||[!A0]	STW	$Ghi,*${CTXA}[12^.LITTLE_ENDIAN]
+||[!A0]	STW	$Glo,*${CTXB}[13^.LITTLE_ENDIAN]
+;;===== branch to outerloop? is taken here
+
+	STW	$Hhi,*${CTXA}[14^.LITTLE_ENDIAN]
+||	STW	$Hlo,*${CTXB}[15^.LITTLE_ENDIAN]
+||	MVK	-40,B0
+	ADD	FP,B0,SP			; destroy circular buffer
+||	LDDW	*FP[-4],A11:A10
+	LDDW	*SP[2],A13:A12
+||	LDDW	*FP[-2],B11:B10
+	LDDW	*SP[4],B13:B12
+||	BNOP	RA
+	LDW	*++SP(40),FP			; restore frame pointer
+	MVK	0,B0
+	MVC	B0,AMR				; clear AMR
+	NOP	2				; wait till FP is committed
+	.endasmfunc
+
+	.if	__TI_EABI__
+	.sect	".text:sha_asm.const"
+	.else
+	.sect	".const:sha_asm"
+	.endif
+	.align	128
+K512:
+	.uword	0x428a2f98,0xd728ae22, 0x71374491,0x23ef65cd
+	.uword	0xb5c0fbcf,0xec4d3b2f, 0xe9b5dba5,0x8189dbbc
+	.uword	0x3956c25b,0xf348b538, 0x59f111f1,0xb605d019
+	.uword	0x923f82a4,0xaf194f9b, 0xab1c5ed5,0xda6d8118
+	.uword	0xd807aa98,0xa3030242, 0x12835b01,0x45706fbe
+	.uword	0x243185be,0x4ee4b28c, 0x550c7dc3,0xd5ffb4e2
+	.uword	0x72be5d74,0xf27b896f, 0x80deb1fe,0x3b1696b1
+	.uword	0x9bdc06a7,0x25c71235, 0xc19bf174,0xcf692694
+	.uword	0xe49b69c1,0x9ef14ad2, 0xefbe4786,0x384f25e3
+	.uword	0x0fc19dc6,0x8b8cd5b5, 0x240ca1cc,0x77ac9c65
+	.uword	0x2de92c6f,0x592b0275, 0x4a7484aa,0x6ea6e483
+	.uword	0x5cb0a9dc,0xbd41fbd4, 0x76f988da,0x831153b5
+	.uword	0x983e5152,0xee66dfab, 0xa831c66d,0x2db43210
+	.uword	0xb00327c8,0x98fb213f, 0xbf597fc7,0xbeef0ee4
+	.uword	0xc6e00bf3,0x3da88fc2, 0xd5a79147,0x930aa725
+	.uword	0x06ca6351,0xe003826f, 0x14292967,0x0a0e6e70
+	.uword	0x27b70a85,0x46d22ffc, 0x2e1b2138,0x5c26c926
+	.uword	0x4d2c6dfc,0x5ac42aed, 0x53380d13,0x9d95b3df
+	.uword	0x650a7354,0x8baf63de, 0x766a0abb,0x3c77b2a8
+	.uword	0x81c2c92e,0x47edaee6, 0x92722c85,0x1482353b
+	.uword	0xa2bfe8a1,0x4cf10364, 0xa81a664b,0xbc423001
+	.uword	0xc24b8b70,0xd0f89791, 0xc76c51a3,0x0654be30
+	.uword	0xd192e819,0xd6ef5218, 0xd6990624,0x5565a910
+	.uword	0xf40e3585,0x5771202a, 0x106aa070,0x32bbd1b8
+	.uword	0x19a4c116,0xb8d2d0c8, 0x1e376c08,0x5141ab53
+	.uword	0x2748774c,0xdf8eeb99, 0x34b0bcb5,0xe19b48a8
+	.uword	0x391c0cb3,0xc5c95a63, 0x4ed8aa4a,0xe3418acb
+	.uword	0x5b9cca4f,0x7763e373, 0x682e6ff3,0xd6b2b8a3
+	.uword	0x748f82ee,0x5defb2fc, 0x78a5636f,0x43172f60
+	.uword	0x84c87814,0xa1f0ab72, 0x8cc70208,0x1a6439ec
+	.uword	0x90befffa,0x23631e28, 0xa4506ceb,0xde82bde9
+	.uword	0xbef9a3f7,0xb2c67915, 0xc67178f2,0xe372532b
+	.uword	0xca273ece,0xea26619c, 0xd186b8c7,0x21c0c207
+	.uword	0xeada7dd6,0xcde0eb1e, 0xf57d4f7f,0xee6ed178
+	.uword	0x06f067aa,0x72176fba, 0x0a637dc5,0xa2c898a6
+	.uword	0x113f9804,0xbef90dae, 0x1b710b35,0x131c471b
+	.uword	0x28db77f5,0x23047d84, 0x32caab7b,0x40c72493
+	.uword	0x3c9ebe0a,0x15c9bebc, 0x431d67c4,0x9c100d4c
+	.uword	0x4cc5d4be,0xcb3e42b6, 0x597f299c,0xfc657e2a
+	.uword	0x5fcb6fab,0x3ad6faec, 0x6c44198c,0x4a475817
+	.cstring "SHA512 block transform for C64x+, CRYPTOGAMS by <appro\@openssl.org>"
+	.align	4
+___
+
+print $code;
+close STDOUT;
--- a/crypto/sha/asm/sha512-ia64.pl
+++ b/crypto/sha/asm/sha512-ia64.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /usr/bin/env perl
+# Copyright 2004-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+
 #
 # ====================================================================
 # Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
@@ -68,7 +75,7 @@
 # To generate code, pass the file name with either 256 or 512 in its
 # name and compiler flags.

-$output=shift;
+$output=pop;

 if ($output =~ /512.*\.[s|asm]/) {
 	$SZ=8;
--- a/crypto/sha/asm/sha512-mips.pl
+++ b/crypto/sha/asm/sha512-mips.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /usr/bin/env perl
+# Copyright 2010-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+

 # ====================================================================
 # Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
@@ -52,6 +59,7 @@
 $flavour = shift || "o32"; # supported flavours are o32,n32,64,nubi32,nubi64

 if ($flavour =~ /64|n32/i) {
+	$PTR_LA="dla";
 	$PTR_ADD="dadd";	# incidentally works even on n32
 	$PTR_SUB="dsub";	# incidentally works even on n32
 	$REG_S="sd";
@@ -59,6 +67,7 @@ if ($flavour =~ /64|n32/i) {
 	$PTR_SLL="dsll";	# incidentally works even on n32
 	$SZREG=8;
 } else {
+	$PTR_LA="la";
 	$PTR_ADD="add";
 	$PTR_SUB="sub";
 	$REG_S="sw";
@@ -74,7 +83,7 @@ $pf = ($flavour =~ /nubi/i) ? $t0 : $t2;

 $big_endian=(`echo MIPSEL | $ENV{CC} -E -`=~/MIPSEL/)?1:0 if ($ENV{CC});

-for (@ARGV) {	$output=$_ if (/^\w[\w\-]*\.\w+$/);	}
+for (@ARGV) {	$output=$_ if (/\w[\w\-]*\.\w+$/);	}
 open STDOUT,">$output";

 if (!defined($big_endian)) { $big_endian=(unpack('L',pack('N',1))==1); }
@@ -286,7 +295,7 @@ ___
 }

 $FRAMESIZE=16*$SZ+16*$SZREG;
-$SAVED_REGS_MASK = ($flavour =~ /nubi/i) ? 0xc0fff008 : 0xc0ff0000;
+$SAVED_REGS_MASK = ($flavour =~ /nubi/i) ? "0xc0fff008" : "0xc0ff0000";

 $code.=<<___;
 #ifdef OPENSSL_FIPSCANISTER
@@ -343,7 +352,7 @@ $code.=<<___ if ($flavour !~ /o32/i);	# non-o32 PIC-ification
 ___
 $code.=<<___;
 	.set	reorder
-	la	$Ktbl,K${label}		# PIC-ified 'load address'
+	$PTR_LA	$Ktbl,K${label}		# PIC-ified 'load address'

 	$LD	$A,0*$SZ($ctx)		# load context
 	$LD	$B,1*$SZ($ctx)
--- a/crypto/sha/asm/sha512-parisc.pl
+++ b/crypto/sha/asm/sha512-parisc.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /usr/bin/env perl
+# Copyright 2009-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+

 # ====================================================================
 # Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
@@ -19,7 +26,7 @@
 # SHA512 performance is >2.9x better than gcc 3.2 generated code on
 # PA-7100LC, PA-RISC 1.1 processor. Then implementation detects if the
 # code is executed on PA-RISC 2.0 processor and switches to 64-bit
-# code path delivering adequate peformance even in "blended" 32-bit
+# code path delivering adequate performance even in "blended" 32-bit
 # build. Though 64-bit code is not any faster than code generated by
 # vendor compiler on PA-8600...
 #
--- a/crypto/sha/asm/sha512-ppc.pl
+++ b/crypto/sha/asm/sha512-ppc.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /usr/bin/env perl
+# Copyright 2006-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+

 # ====================================================================
 # Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
@@ -259,7 +266,7 @@ Lunaligned:
 	andi.	$t1,$t1,`4096-16*$SZ`	; distance to closest page boundary
 	beq	Lcross_page
 	$UCMP	$num,$t1
-	ble-	Laligned		; didn't cross the page boundary
+	ble	Laligned		; didn't cross the page boundary
 	subfc	$num,$t1,$num
 	add	$t1,$inp,$t1
 	$PUSH	$num,`$FRAME-$SIZE_T*25`($sp)	; save real remaining num
@@ -317,7 +324,7 @@ $code.=<<___;
 	$POP	$inp,`$FRAME-$SIZE_T*26`($sp)	; restore real inp
 	$POP	$num,`$FRAME-$SIZE_T*25`($sp)	; restore real num
 	addic.	$num,$num,`-16*$SZ`		; num--
-	bne-	Lunaligned
+	bne	Lunaligned

 Ldone:
 	$POP	r0,`$FRAME+$LRSAVE`($sp)
@@ -396,7 +403,7 @@ for(;$i<32;$i++) {
 	unshift(@V,pop(@V));
 }
 $code.=<<___;
-	bdnz-	Lrounds
+	bdnz	Lrounds

 	$POP	$ctx,`$FRAME-$SIZE_T*22`($sp)
 	$POP	$inp,`$FRAME-$SIZE_T*23`($sp)	; inp pointer
@@ -644,7 +651,7 @@ for(;$i<32;$i++) {
 	($a0,$a1,$a2,$a3) = ($a2,$a3,$a0,$a1);
 }
 $code.=<<___;
-	bdnz-	Lrounds
+	bdnz	Lrounds

 	$POP	$ctx,`$FRAME-$SIZE_T*22`($sp)
 	$POP	$inp,`$FRAME-$SIZE_T*23`($sp)	; inp pointer
--- a/crypto/sha/asm/sha512-s390x.pl
+++ b/crypto/sha/asm/sha512-s390x.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /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
+

 # ====================================================================
 # Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
@@ -33,7 +40,7 @@
 # instructions and achieve "64-bit" performance even in 31-bit legacy
 # application context. The feature is not specific to any particular
 # processor, as long as it's "z-CPU". Latter implies that the code
-# remains z/Architecture specific. On z900 SHA256 was measured to
+# remains z/Architecture specific. On z990 SHA256 was measured to
 # perform 2.4x and SHA512 - 13x better than code generated by gcc 4.3.

 $flavour = shift;
@@ -64,7 +71,7 @@ $tbl="%r13";
 $T1="%r14";
 $sp="%r15";

-while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+while (($output=shift) && ($output!~/\w[\w\-]*\.\w+$/)) {}
 open STDOUT,">$output";

 if ($output =~ /512/) {
@@ -240,10 +247,7 @@ $code.=<<___ if ($kimdfunc);
 	lg	%r0,0(%r1)
 	tmhl	%r0,0x4000	# check for message-security assist
 	jz	.Lsoftware
-	lghi	%r0,0
-	la	%r1,`2*$SIZE_T`($sp)
-	.long	0xb93e0002	# kimd %r0,%r2
-	lg	%r0,`2*$SIZE_T`($sp)
+	lg	%r0,16(%r1)	# check kimd capabilities
 	tmhh	%r0,`0x8000>>$kimdfunc`
 	jz	.Lsoftware
 	lghi	%r0,$kimdfunc
@@ -311,7 +315,7 @@ $code.=<<___;
 	br	%r14
 .size	$Func,.-$Func
 .string	"SHA${label} block transform for s390x, CRYPTOGAMS by <appro\@openssl.org>"
-.comm	OPENSSL_s390xcap_P,16,8
+.comm	OPENSSL_s390xcap_P,80,8
 ___

 $code =~ s/\`([^\`]*)\`/eval $1/gem;
--- a/crypto/sha/asm/sha512-sparcv9.pl
+++ b/crypto/sha/asm/sha512-sparcv9.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /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
+

 # ====================================================================
 # Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
@@ -49,7 +56,7 @@
 # saturates at 11.5x single-process result on 8-core processor, or
 # ~11/16GBps per 2.85GHz socket.

-$output=shift;
+$output=pop;
 open STDOUT,">$output";

 if ($output =~ /512/) {
@@ -791,7 +798,7 @@ ___

 # 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 unvis {
--- a/crypto/sha/asm/sha512-x86_64.pl
+++ b/crypto/sha/asm/sha512-x86_64.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /usr/bin/env perl
+# Copyright 2005-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+
 #
 # ====================================================================
 # Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
@@ -34,7 +41,7 @@
 # level parallelism, on a given CPU implementation in this case.
 #
 # Special note on Intel EM64T. While Opteron CPU exhibits perfect
-# perfromance ratio of 1.5 between 64- and 32-bit flavors [see above],
+# performance ratio of 1.5 between 64- and 32-bit flavors [see above],
 # [currently available] EM64T CPUs apparently are far from it. On the
 # contrary, 64-bit version, sha512_block, is ~30% *slower* than 32-bit
 # sha256_block:-( This is presumably because 64-bit shifts/rotates
@@ -86,12 +93,14 @@
 # Sandy Bridge	17.4	14.2(+23%)  11.6(+50%(**))  11.2    8.10(+38%(**))
 # Ivy Bridge	12.6	10.5(+20%)  10.3(+22%)	    8.17    7.22(+13%)
 # Haswell	12.2	9.28(+31%)  7.80(+56%)	    7.66    5.40(+42%)
+# Skylake	11.4	9.03(+26%)  7.70(+48%)      7.25    5.20(+40%)
 # Bulldozer	21.1	13.6(+54%)  13.6(+54%(***)) 13.5    8.58(+57%)
 # VIA Nano	23.0	16.5(+39%)  -		    14.7    -
 # Atom		23.0	18.9(+22%)  -		    14.7    -
 # Silvermont	27.4	20.6(+33%)  -               17.5    -
+# Goldmont	18.9	14.3(+32%)  4.16(+350%)     12.0    -
 #
-# (*)	whichever best applicable;
+# (*)	whichever best applicable, including SHAEXT;
 # (**)	switch from ror to shrd stands for fair share of improvement;
 # (***)	execution time is fully determined by remaining integer-only
 #	part, body_00_15; reducing the amount of SIMD instructions
@@ -124,14 +133,14 @@ if (!$avx && $win64 && ($flavour =~ /masm/ || $ENV{ASM} =~ /ml64/) &&
 	$avx = ($1>=10) + ($1>=11);
 }

-if (!$avx && `$ENV{CC} -v 2>&1` =~ /(^clang version|based on LLVM) ([3-9]\.[0-9]+)/) {
+if (!$avx && `$ENV{CC} -v 2>&1` =~ /((?:^clang|LLVM) version|.*based on LLVM) ([3-9]\.[0-9]+)/) {
 	$avx = ($2>=3.0) + ($2>3.0);
 }

 $shaext=1;	### set to zero if compiling for 1.0.1
 $avx=1		if (!$shaext && $avx);

-open OUT,"| \"$^X\" $xlate $flavour $output";
+open OUT,"| \"$^X\" \"$xlate\" $flavour \"$output\"";
 *STDOUT=*OUT;

 if ($output =~ /512/) {
--- a/crypto/sha/asm/sha512p8-ppc.pl
+++ b/crypto/sha/asm/sha512p8-ppc.pl
@@ -1,4 +1,11 @@
-#!/usr/bin/env perl
+#! /usr/bin/env perl
+# Copyright 2014-2016 The OpenSSL Project Authors. All Rights Reserved.
+#
+# Licensed under the OpenSSL license (the "License").  You may not use
+# this file except in compliance with the License.  You can obtain a copy
+# in the file LICENSE in the source distribution or at
+# https://www.openssl.org/source/license.html
+

 # ====================================================================
 # Written by Andy Polyakov <appro@openssl.org> for the OpenSSL