Small Issues & Maintenance

-=-=-=-=-=-=-=-=-=-=-=-=-=

-Begin work on Navigation Meshes (simple pathfinding for now, will improve soon)
-More doc on theme overriding
-Upgraded OpenSSL to version without bugs
-Misc bugfixes

37 files changed, 13700 insertions, 0 deletions

diff --git a/drivers/builtin_openssl2/crypto/sha/asm/README b/drivers/builtin_openssl2/crypto/sha/asm/README
new file mode 100644
index 0000000000..b7e755765f
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/README
@@ -0,0 +1 @@
+C2.pl works
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha1-586.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha1-586.pl
new file mode 100644
index 0000000000..1084d227fe
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha1-586.pl
@@ -0,0 +1,1229 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# [Re]written by Andy Polyakov <appro@fy.chalmers.se> 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/.
+# ====================================================================
+
+# "[Re]written" was achieved in two major overhauls. In 2004 BODY_*
+# functions were re-implemented to address P4 performance issue [see
+# commentary below], and in 2006 the rest was rewritten in order to
+# gain freedom to liberate licensing terms.
+
+# January, September 2004.
+#
+# It was noted that Intel IA-32 C compiler generates code which
+# performs ~30% *faster* on P4 CPU than original *hand-coded*
+# SHA1 assembler implementation. To address this problem (and
+# prove that humans are still better than machines:-), the
+# original code was overhauled, which resulted in following
+# performance changes:
+#
+#		compared with original	compared with Intel cc
+#		assembler impl.		generated code
+# Pentium	-16%			+48%
+# PIII/AMD	+8%			+16%
+# P4		+85%(!)			+45%
+#
+# As you can see Pentium came out as looser:-( Yet I reckoned that
+# improvement on P4 outweights the loss and incorporate this
+# re-tuned code to 0.9.7 and later.
+# ----------------------------------------------------------------
+#					<appro@fy.chalmers.se>
+
+# August 2009.
+#
+# George Spelvin has tipped that F_40_59(b,c,d) can be rewritten as
+# '(c&d) + (b&(c^d))', which allows to accumulate partial results
+# and lighten "pressure" on scratch registers. This resulted in
+# >12% performance improvement on contemporary AMD cores (with no
+# degradation on other CPUs:-). Also, the code was revised to maximize
+# "distance" between instructions producing input to 'lea' instruction
+# and the 'lea' instruction itself, which is essential for Intel Atom
+# core and resulted in ~15% improvement.
+
+# October 2010.
+#
+# Add SSSE3, Supplemental[!] SSE3, implementation. The idea behind it
+# is to offload message schedule denoted by Wt in NIST specification,
+# or Xupdate in OpenSSL source, to SIMD unit. The idea is not novel,
+# and in SSE2 context was first explored by Dean Gaudet in 2004, see
+# http://arctic.org/~dean/crypto/sha1.html. Since then several things
+# have changed that made it interesting again:
+#
+# a) XMM units became faster and wider;
+# b) instruction set became more versatile;
+# c) an important observation was made by Max Locktykhin, which made
+#    it possible to reduce amount of instructions required to perform
+#    the operation in question, for further details see
+#    http://software.intel.com/en-us/articles/improving-the-performance-of-the-secure-hash-algorithm-1/.
+
+# April 2011.
+#
+# Add AVX code path, probably most controversial... The thing is that
+# 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
+# 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
+# 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
+# [according to its optimization manual]. Which is why AVX code path
+# is guarded by *both* AVX and synthetic bit denoting Intel CPUs.
+# One can argue that it's unfair to AMD, but without 'sh[rl]d' it
+# makes no sense to keep the AVX code path. If somebody feels that
+# strongly, it's probably more appropriate to discuss possibility of
+# using vector rotate XOP on AMD...
+
+######################################################################
+# Current performance is summarized in following table. Numbers are
+# CPU clock cycles spent to process single byte (less is better).
+#
+#		x86		SSSE3		AVX
+# Pentium	15.7		-
+# PIII		11.5		-
+# P4		10.6		-
+# AMD K8	7.1		-
+# Core2		7.3		6.1/+20%	-
+# Atom		12.5		9.5(*)/+32%	-
+# Westmere	7.3		5.6/+30%	-
+# Sandy Bridge	8.8		6.2/+40%	5.1(**)/+70%
+#
+# (*)	Loop is 1056 instructions long and expected result is ~8.25.
+#	It remains mystery [to me] why ILP is limited to 1.7.
+#
+# (**)	As per above comment, the result is for AVX *plus* sh[rl]d.
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
+require "x86asm.pl";
+
+&asm_init($ARGV[0],"sha1-586.pl",$ARGV[$#ARGV] eq "386");
+
+$xmm=$ymm=0;
+for (@ARGV) { $xmm=1 if (/-DOPENSSL_IA32_SSE2/); }
+
+$ymm=1 if ($xmm &&
+		`$ENV{CC} -Wa,-v -c -o /dev/null -x assembler /dev/null 2>&1`
+			=~ /GNU assembler version ([2-9]\.[0-9]+)/ &&
+		$1>=2.19);	# first version supporting AVX
+
+$ymm=1 if ($xmm && !$ymm && $ARGV[0] eq "win32n" && 
+		`nasm -v 2>&1` =~ /NASM version ([2-9]\.[0-9]+)/ &&
+		$1>=2.03);	# first version supporting AVX
+
+&external_label("OPENSSL_ia32cap_P") if ($xmm);
+
+
+$A="eax";
+$B="ebx";
+$C="ecx";
+$D="edx";
+$E="edi";
+$T="esi";
+$tmp1="ebp";
+
+@V=($A,$B,$C,$D,$E,$T);
+
+$alt=0;	# 1 denotes alternative IALU implementation, which performs
+	# 8% *worse* on P4, same on Westmere and Atom, 2% better on
+	# Sandy Bridge...
+
+sub BODY_00_15
+	{
+	local($n,$a,$b,$c,$d,$e,$f)=@_;
+
+	&comment("00_15 $n");
+
+	&mov($f,$c);			# f to hold F_00_19(b,c,d)
+	 if ($n==0)  { &mov($tmp1,$a); }
+	 else        { &mov($a,$tmp1); }
+	&rotl($tmp1,5);			# tmp1=ROTATE(a,5)
+	 &xor($f,$d);
+	&add($tmp1,$e);			# tmp1+=e;
+	 &mov($e,&swtmp($n%16));	# e becomes volatile and is loaded
+	 				# with xi, also note that e becomes
+					# f in next round...
+	&and($f,$b);
+	&rotr($b,2);			# b=ROTATE(b,30)
+	 &xor($f,$d);			# f holds F_00_19(b,c,d)
+	&lea($tmp1,&DWP(0x5a827999,$tmp1,$e));	# tmp1+=K_00_19+xi
+
+	if ($n==15) { &mov($e,&swtmp(($n+1)%16));# pre-fetch f for next round
+		      &add($f,$tmp1); }	# f+=tmp1
+	else        { &add($tmp1,$f); }	# f becomes a in next round
+	&mov($tmp1,$a)			if ($alt && $n==15);
+	}
+
+sub BODY_16_19
+	{
+	local($n,$a,$b,$c,$d,$e,$f)=@_;
+
+	&comment("16_19 $n");
+
+if ($alt) {
+	&xor($c,$d);
+	 &xor($f,&swtmp(($n+2)%16));	# f to hold Xupdate(xi,xa,xb,xc,xd)
+	&and($tmp1,$c);			# tmp1 to hold F_00_19(b,c,d), b&=c^d
+	 &xor($f,&swtmp(($n+8)%16));
+	&xor($tmp1,$d);			# tmp1=F_00_19(b,c,d)
+	 &xor($f,&swtmp(($n+13)%16));	# f holds xa^xb^xc^xd
+	&rotl($f,1);			# f=ROTATE(f,1)
+	 &add($e,$tmp1);		# e+=F_00_19(b,c,d)
+	&xor($c,$d);			# restore $c
+	 &mov($tmp1,$a);		# b in next round
+	&rotr($b,$n==16?2:7);		# b=ROTATE(b,30)
+	 &mov(&swtmp($n%16),$f);	# xi=f
+	&rotl($a,5);			# ROTATE(a,5)
+	 &lea($f,&DWP(0x5a827999,$f,$e));# f+=F_00_19(b,c,d)+e
+	&mov($e,&swtmp(($n+1)%16));	# pre-fetch f for next round
+	 &add($f,$a);			# f+=ROTATE(a,5)
+} else {
+	&mov($tmp1,$c);			# tmp1 to hold F_00_19(b,c,d)
+	 &xor($f,&swtmp(($n+2)%16));	# f to hold Xupdate(xi,xa,xb,xc,xd)
+	&xor($tmp1,$d);
+	 &xor($f,&swtmp(($n+8)%16));
+	&and($tmp1,$b);
+	 &xor($f,&swtmp(($n+13)%16));	# f holds xa^xb^xc^xd
+	&rotl($f,1);			# f=ROTATE(f,1)
+	 &xor($tmp1,$d);		# tmp1=F_00_19(b,c,d)
+	&add($e,$tmp1);			# e+=F_00_19(b,c,d)
+	 &mov($tmp1,$a);
+	&rotr($b,2);			# b=ROTATE(b,30)
+	 &mov(&swtmp($n%16),$f);	# xi=f
+	&rotl($tmp1,5);			# ROTATE(a,5)
+	 &lea($f,&DWP(0x5a827999,$f,$e));# f+=F_00_19(b,c,d)+e
+	&mov($e,&swtmp(($n+1)%16));	# pre-fetch f for next round
+	 &add($f,$tmp1);		# f+=ROTATE(a,5)
+}
+	}
+
+sub BODY_20_39
+	{
+	local($n,$a,$b,$c,$d,$e,$f)=@_;
+	local $K=($n<40)?0x6ed9eba1:0xca62c1d6;
+
+	&comment("20_39 $n");
+
+if ($alt) {
+	&xor($tmp1,$c);			# tmp1 to hold F_20_39(b,c,d), b^=c
+	 &xor($f,&swtmp(($n+2)%16));	# f to hold Xupdate(xi,xa,xb,xc,xd)
+	&xor($tmp1,$d);			# tmp1 holds F_20_39(b,c,d)
+	 &xor($f,&swtmp(($n+8)%16));
+	&add($e,$tmp1);			# e+=F_20_39(b,c,d)
+	 &xor($f,&swtmp(($n+13)%16));	# f holds xa^xb^xc^xd
+	&rotl($f,1);			# f=ROTATE(f,1)
+	 &mov($tmp1,$a);		# b in next round
+	&rotr($b,7);			# b=ROTATE(b,30)
+	 &mov(&swtmp($n%16),$f)		if($n<77);# xi=f
+	&rotl($a,5);			# ROTATE(a,5)
+	 &xor($b,$c)			if($n==39);# warm up for BODY_40_59
+	&and($tmp1,$b)			if($n==39);
+	 &lea($f,&DWP($K,$f,$e));	# f+=e+K_XX_YY
+	&mov($e,&swtmp(($n+1)%16))	if($n<79);# pre-fetch f for next round
+	 &add($f,$a);			# f+=ROTATE(a,5)
+	&rotr($a,5)			if ($n==79);
+} else {
+	&mov($tmp1,$b);			# tmp1 to hold F_20_39(b,c,d)
+	 &xor($f,&swtmp(($n+2)%16));	# f to hold Xupdate(xi,xa,xb,xc,xd)
+	&xor($tmp1,$c);
+	 &xor($f,&swtmp(($n+8)%16));
+	&xor($tmp1,$d);			# tmp1 holds F_20_39(b,c,d)
+	 &xor($f,&swtmp(($n+13)%16));	# f holds xa^xb^xc^xd
+	&rotl($f,1);			# f=ROTATE(f,1)
+	 &add($e,$tmp1);		# e+=F_20_39(b,c,d)
+	&rotr($b,2);			# b=ROTATE(b,30)
+	 &mov($tmp1,$a);
+	&rotl($tmp1,5);			# ROTATE(a,5)
+	 &mov(&swtmp($n%16),$f) if($n<77);# xi=f
+	&lea($f,&DWP($K,$f,$e));	# f+=e+K_XX_YY
+	 &mov($e,&swtmp(($n+1)%16)) if($n<79);# pre-fetch f for next round
+	&add($f,$tmp1);			# f+=ROTATE(a,5)
+}
+	}
+
+sub BODY_40_59
+	{
+	local($n,$a,$b,$c,$d,$e,$f)=@_;
+
+	&comment("40_59 $n");
+
+if ($alt) {
+	&add($e,$tmp1);			# e+=b&(c^d)
+	 &xor($f,&swtmp(($n+2)%16));	# f to hold Xupdate(xi,xa,xb,xc,xd)
+	&mov($tmp1,$d);
+	 &xor($f,&swtmp(($n+8)%16));
+	&xor($c,$d);			# restore $c
+	 &xor($f,&swtmp(($n+13)%16));	# f holds xa^xb^xc^xd
+	&rotl($f,1);			# f=ROTATE(f,1)
+	 &and($tmp1,$c);
+	&rotr($b,7);			# b=ROTATE(b,30)
+	 &add($e,$tmp1);		# e+=c&d
+	&mov($tmp1,$a);			# b in next round
+	 &mov(&swtmp($n%16),$f);	# xi=f
+	&rotl($a,5);			# ROTATE(a,5)
+	 &xor($b,$c)			if ($n<59);
+	&and($tmp1,$b)			if ($n<59);# tmp1 to hold F_40_59(b,c,d)
+	 &lea($f,&DWP(0x8f1bbcdc,$f,$e));# f+=K_40_59+e+(b&(c^d))
+	&mov($e,&swtmp(($n+1)%16));	# pre-fetch f for next round
+	 &add($f,$a);			# f+=ROTATE(a,5)
+} else {
+	&mov($tmp1,$c);			# tmp1 to hold F_40_59(b,c,d)
+	 &xor($f,&swtmp(($n+2)%16));	# f to hold Xupdate(xi,xa,xb,xc,xd)
+	&xor($tmp1,$d);
+	 &xor($f,&swtmp(($n+8)%16));
+	&and($tmp1,$b);
+	 &xor($f,&swtmp(($n+13)%16));	# f holds xa^xb^xc^xd
+	&rotl($f,1);			# f=ROTATE(f,1)
+	 &add($tmp1,$e);		# b&(c^d)+=e
+	&rotr($b,2);			# b=ROTATE(b,30)
+	 &mov($e,$a);			# e becomes volatile
+	&rotl($e,5);			# ROTATE(a,5)
+	 &mov(&swtmp($n%16),$f);	# xi=f
+	&lea($f,&DWP(0x8f1bbcdc,$f,$tmp1));# f+=K_40_59+e+(b&(c^d))
+	 &mov($tmp1,$c);
+	&add($f,$e);			# f+=ROTATE(a,5)
+	 &and($tmp1,$d);
+	&mov($e,&swtmp(($n+1)%16));	# pre-fetch f for next round
+	 &add($f,$tmp1);		# f+=c&d
+}
+	}
+
+&function_begin("sha1_block_data_order");
+if ($xmm) {
+  &static_label("ssse3_shortcut");
+  &static_label("avx_shortcut")		if ($ymm);
+  &static_label("K_XX_XX");
+
+	&call	(&label("pic_point"));	# make it PIC!
+  &set_label("pic_point");
+	&blindpop($tmp1);
+	&picmeup($T,"OPENSSL_ia32cap_P",$tmp1,&label("pic_point"));
+	&lea	($tmp1,&DWP(&label("K_XX_XX")."-".&label("pic_point"),$tmp1));
+
+	&mov	($A,&DWP(0,$T));
+	&mov	($D,&DWP(4,$T));
+	&test	($D,1<<9);		# check SSSE3 bit
+	&jz	(&label("x86"));
+	&test	($A,1<<24);		# check FXSR bit
+	&jz	(&label("x86"));
+	if ($ymm) {
+		&and	($D,1<<28);		# mask AVX bit
+		&and	($A,1<<30);		# mask "Intel CPU" bit
+		&or	($A,$D);
+		&cmp	($A,1<<28|1<<30);
+		&je	(&label("avx_shortcut"));
+	}
+	&jmp	(&label("ssse3_shortcut"));
+  &set_label("x86",16);
+}
+	&mov($tmp1,&wparam(0));	# SHA_CTX *c
+	&mov($T,&wparam(1));	# const void *input
+	&mov($A,&wparam(2));	# size_t num
+	&stack_push(16+3);	# allocate X[16]
+	&shl($A,6);
+	&add($A,$T);
+	&mov(&wparam(2),$A);	# pointer beyond the end of input
+	&mov($E,&DWP(16,$tmp1));# pre-load E
+	&jmp(&label("loop"));
+
+&set_label("loop",16);
+
+	# copy input chunk to X, but reversing byte order!
+	for ($i=0; $i<16; $i+=4)
+		{
+		&mov($A,&DWP(4*($i+0),$T));
+		&mov($B,&DWP(4*($i+1),$T));
+		&mov($C,&DWP(4*($i+2),$T));
+		&mov($D,&DWP(4*($i+3),$T));
+		&bswap($A);
+		&bswap($B);
+		&bswap($C);
+		&bswap($D);
+		&mov(&swtmp($i+0),$A);
+		&mov(&swtmp($i+1),$B);
+		&mov(&swtmp($i+2),$C);
+		&mov(&swtmp($i+3),$D);
+		}
+	&mov(&wparam(1),$T);	# redundant in 1st spin
+
+	&mov($A,&DWP(0,$tmp1));	# load SHA_CTX
+	&mov($B,&DWP(4,$tmp1));
+	&mov($C,&DWP(8,$tmp1));
+	&mov($D,&DWP(12,$tmp1));
+	# E is pre-loaded
+
+	for($i=0;$i<16;$i++)	{ &BODY_00_15($i,@V); unshift(@V,pop(@V)); }
+	for(;$i<20;$i++)	{ &BODY_16_19($i,@V); unshift(@V,pop(@V)); }
+	for(;$i<40;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+	for(;$i<60;$i++)	{ &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
+	for(;$i<80;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+
+	(($V[5] eq $D) and ($V[0] eq $E)) or die;	# double-check
+
+	&mov($tmp1,&wparam(0));	# re-load SHA_CTX*
+	&mov($D,&wparam(1));	# D is last "T" and is discarded
+
+	&add($E,&DWP(0,$tmp1));	# E is last "A"...
+	&add($T,&DWP(4,$tmp1));
+	&add($A,&DWP(8,$tmp1));
+	&add($B,&DWP(12,$tmp1));
+	&add($C,&DWP(16,$tmp1));
+
+	&mov(&DWP(0,$tmp1),$E);	# update SHA_CTX
+	 &add($D,64);		# advance input pointer
+	&mov(&DWP(4,$tmp1),$T);
+	 &cmp($D,&wparam(2));	# have we reached the end yet?
+	&mov(&DWP(8,$tmp1),$A);
+	 &mov($E,$C);		# C is last "E" which needs to be "pre-loaded"
+	&mov(&DWP(12,$tmp1),$B);
+	 &mov($T,$D);		# input pointer
+	&mov(&DWP(16,$tmp1),$C);
+	&jb(&label("loop"));
+
+	&stack_pop(16+3);
+&function_end("sha1_block_data_order");
+
+if ($xmm) {
+######################################################################
+# The SSSE3 implementation.
+#
+# %xmm[0-7] are used as ring @X[] buffer containing quadruples of last
+# 32 elements of the message schedule or Xupdate outputs. First 4
+# quadruples are simply byte-swapped input, next 4 are calculated
+# according to method originally suggested by Dean Gaudet (modulo
+# being implemented in SSSE3). Once 8 quadruples or 32 elements are
+# collected, it switches to routine proposed by Max Locktyukhin.
+#
+# Calculations inevitably require temporary reqisters, and there are
+# no %xmm registers left to spare. For this reason part of the ring
+# buffer, X[2..4] to be specific, is offloaded to 3 quadriples ring
+# buffer on the stack. Keep in mind that X[2] is alias X[-6], X[3] -
+# X[-5], and X[4] - X[-4]...
+#
+# Another notable optimization is aggressive stack frame compression
+# aiming to minimize amount of 9-byte instructions...
+#
+# Yet another notable optimization is "jumping" $B variable. It means
+# that there is no register permanently allocated for $B value. This
+# allowed to eliminate one instruction from body_20_39...
+#
+my $Xi=4;			# 4xSIMD Xupdate round, start pre-seeded
+my @X=map("xmm$_",(4..7,0..3));	# pre-seeded for $Xi=4
+my @V=($A,$B,$C,$D,$E);
+my $j=0;			# hash round
+my @T=($T,$tmp1);
+my $inp;
+
+my $_rol=sub { &rol(@_) };
+my $_ror=sub { &ror(@_) };
+
+&function_begin("_sha1_block_data_order_ssse3");
+	&call	(&label("pic_point"));	# make it PIC!
+	&set_label("pic_point");
+	&blindpop($tmp1);
+	&lea	($tmp1,&DWP(&label("K_XX_XX")."-".&label("pic_point"),$tmp1));
+&set_label("ssse3_shortcut");
+
+	&movdqa	(@X[3],&QWP(0,$tmp1));		# K_00_19
+	&movdqa	(@X[4],&QWP(16,$tmp1));		# K_20_39
+	&movdqa	(@X[5],&QWP(32,$tmp1));		# K_40_59
+	&movdqa	(@X[6],&QWP(48,$tmp1));		# K_60_79
+	&movdqa	(@X[2],&QWP(64,$tmp1));		# pbswap mask
+
+	&mov	($E,&wparam(0));		# load argument block
+	&mov	($inp=@T[1],&wparam(1));
+	&mov	($D,&wparam(2));
+	&mov	(@T[0],"esp");
+
+	# stack frame layout
+	#
+	# +0	X[0]+K	X[1]+K	X[2]+K	X[3]+K	# XMM->IALU xfer area
+	#	X[4]+K	X[5]+K	X[6]+K	X[7]+K
+	#	X[8]+K	X[9]+K	X[10]+K	X[11]+K
+	#	X[12]+K	X[13]+K	X[14]+K	X[15]+K
+	#
+	# +64	X[0]	X[1]	X[2]	X[3]	# XMM->XMM backtrace area
+	#	X[4]	X[5]	X[6]	X[7]
+	#	X[8]	X[9]	X[10]	X[11]	# even borrowed for K_00_19
+	#
+	# +112	K_20_39	K_20_39	K_20_39	K_20_39	# constants
+	#	K_40_59	K_40_59	K_40_59	K_40_59
+	#	K_60_79	K_60_79	K_60_79	K_60_79
+	#	K_00_19	K_00_19	K_00_19	K_00_19
+	#	pbswap mask
+	#
+	# +192	ctx				# argument block
+	# +196	inp
+	# +200	end
+	# +204	esp
+	&sub	("esp",208);
+	&and	("esp",-64);
+
+	&movdqa	(&QWP(112+0,"esp"),@X[4]);	# copy constants
+	&movdqa	(&QWP(112+16,"esp"),@X[5]);
+	&movdqa	(&QWP(112+32,"esp"),@X[6]);
+	&shl	($D,6);				# len*64
+	&movdqa	(&QWP(112+48,"esp"),@X[3]);
+	&add	($D,$inp);			# end of input
+	&movdqa	(&QWP(112+64,"esp"),@X[2]);
+	&add	($inp,64);
+	&mov	(&DWP(192+0,"esp"),$E);		# save argument block
+	&mov	(&DWP(192+4,"esp"),$inp);
+	&mov	(&DWP(192+8,"esp"),$D);
+	&mov	(&DWP(192+12,"esp"),@T[0]);	# save original %esp
+
+	&mov	($A,&DWP(0,$E));		# load context
+	&mov	($B,&DWP(4,$E));
+	&mov	($C,&DWP(8,$E));
+	&mov	($D,&DWP(12,$E));
+	&mov	($E,&DWP(16,$E));
+	&mov	(@T[0],$B);			# magic seed
+
+	&movdqu	(@X[-4&7],&QWP(-64,$inp));	# load input to %xmm[0-3]
+	&movdqu	(@X[-3&7],&QWP(-48,$inp));
+	&movdqu	(@X[-2&7],&QWP(-32,$inp));
+	&movdqu	(@X[-1&7],&QWP(-16,$inp));
+	&pshufb	(@X[-4&7],@X[2]);		# byte swap
+	&pshufb	(@X[-3&7],@X[2]);
+	&pshufb	(@X[-2&7],@X[2]);
+	&movdqa	(&QWP(112-16,"esp"),@X[3]);	# borrow last backtrace slot
+	&pshufb	(@X[-1&7],@X[2]);
+	&paddd	(@X[-4&7],@X[3]);		# add K_00_19
+	&paddd	(@X[-3&7],@X[3]);
+	&paddd	(@X[-2&7],@X[3]);
+	&movdqa	(&QWP(0,"esp"),@X[-4&7]);	# X[]+K xfer to IALU
+	&psubd	(@X[-4&7],@X[3]);		# restore X[]
+	&movdqa	(&QWP(0+16,"esp"),@X[-3&7]);
+	&psubd	(@X[-3&7],@X[3]);
+	&movdqa	(&QWP(0+32,"esp"),@X[-2&7]);
+	&psubd	(@X[-2&7],@X[3]);
+	&movdqa	(@X[0],@X[-3&7]);
+	&jmp	(&label("loop"));
+
+######################################################################
+# SSE instruction sequence is first broken to groups of indepentent
+# instructions, independent in respect to their inputs and shifter
+# (not all architectures have more than one). Then IALU instructions
+# are "knitted in" between the SSE groups. Distance is maintained for
+# SSE latency of 2 in hope that it fits better upcoming AMD Bulldozer
+# [which allegedly also implements SSSE3]...
+#
+# Temporary registers usage. X[2] is volatile at the entry and at the
+# end is restored from backtrace ring buffer. X[3] is expected to
+# contain current K_XX_XX constant and is used to caclulate X[-1]+K
+# from previous round, it becomes volatile the moment the value is
+# saved to stack for transfer to IALU. X[4] becomes volatile whenever
+# X[-4] is accumulated and offloaded to backtrace ring buffer, at the
+# end it is loaded with next K_XX_XX [which becomes X[3] in next
+# round]...
+#
+sub Xupdate_ssse3_16_31()		# recall that $Xi starts wtih 4
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 40 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&palignr(@X[0],@X[-4&7],8);	# compose "X[-14]" in "X[0]"
+	&movdqa	(@X[2],@X[-1&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &paddd	(@X[3],@X[-1&7]);
+	  &movdqa	(&QWP(64+16*(($Xi-4)%3),"esp"),@X[-4&7]);# save X[] to backtrace buffer
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&psrldq	(@X[2],4);		# "X[-3]", 3 dwords
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&pxor	(@X[0],@X[-4&7]);	# "X[0]"^="X[-16]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pxor	(@X[2],@X[-2&7]);	# "X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pxor	(@X[0],@X[2]);		# "X[0]"^="X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &movdqa	(&QWP(0+16*(($Xi-1)&3),"esp"),@X[3]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&movdqa	(@X[4],@X[0]);
+	&movdqa	(@X[2],@X[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pslldq	(@X[4],12);		# "X[0]"<<96, extract one dword
+	&paddd	(@X[0],@X[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&psrld	(@X[2],31);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&movdqa	(@X[3],@X[4]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&psrld	(@X[4],30);
+	&por	(@X[0],@X[2]);		# "X[0]"<<<=1
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &movdqa	(@X[2],&QWP(64+16*(($Xi-6)%3),"esp")) if ($Xi>5);	# restore X[] from backtrace buffer
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pslld	(@X[3],2);
+	&pxor	(@X[0],@X[4]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &movdqa	(@X[4],&QWP(112-16+16*(($Xi)/5),"esp"));	# K_XX_XX
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pxor	(@X[0],@X[3]);		# "X[0]"^=("X[0]"<<96)<<<2
+	  &movdqa	(@X[1],@X[-2&7])	if ($Xi<7);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	 foreach (@insns) { eval; }	# remaining instructions [if any]
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+}
+
+sub Xupdate_ssse3_32_79()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 to 48 instructions
+  my ($a,$b,$c,$d,$e);
+
+	&movdqa	(@X[2],@X[-1&7])	if ($Xi==8);
+	 eval(shift(@insns));		# body_20_39
+	&pxor	(@X[0],@X[-4&7]);	# "X[0]"="X[-32]"^"X[-16]"
+	&palignr(@X[2],@X[-2&7],8);	# compose "X[-6]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&pxor	(@X[0],@X[-7&7]);	# "X[0]"^="X[-28]"
+	  &movdqa	(&QWP(64+16*(($Xi-4)%3),"esp"),@X[-4&7]);	# save X[] to backtrace buffer
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 if ($Xi%5) {
+	  &movdqa	(@X[4],@X[3]);	# "perpetuate" K_XX_XX...
+	 } else {			# ... or load next one
+	  &movdqa	(@X[4],&QWP(112-16+16*($Xi/5),"esp"));
+	 }
+	  &paddd	(@X[3],@X[-1&7]);
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&pxor	(@X[0],@X[2]);		# "X[0]"^="X[-6]"
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&movdqa	(@X[2],@X[0]);
+	  &movdqa	(&QWP(0+16*(($Xi-1)&3),"esp"),@X[3]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&pslld	(@X[0],2);
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	&psrld	(@X[2],30);
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&por	(@X[0],@X[2]);		# "X[0]"<<<=2
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	  &movdqa	(@X[2],&QWP(64+16*(($Xi-6)%3),"esp")) if($Xi<19);	# restore X[] from backtrace buffer
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	  &movdqa	(@X[3],@X[0])	if ($Xi<19);
+	 eval(shift(@insns));
+
+	 foreach (@insns) { eval; }	# remaining instructions
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+}
+
+sub Xuplast_ssse3_80()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	  &paddd	(@X[3],@X[-1&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &movdqa	(&QWP(0+16*(($Xi-1)&3),"esp"),@X[3]);	# X[]+K xfer IALU
+
+	 foreach (@insns) { eval; }		# remaining instructions
+
+	&mov	($inp=@T[1],&DWP(192+4,"esp"));
+	&cmp	($inp,&DWP(192+8,"esp"));
+	&je	(&label("done"));
+
+	&movdqa	(@X[3],&QWP(112+48,"esp"));	# K_00_19
+	&movdqa	(@X[2],&QWP(112+64,"esp"));	# pbswap mask
+	&movdqu	(@X[-4&7],&QWP(0,$inp));	# load input
+	&movdqu	(@X[-3&7],&QWP(16,$inp));
+	&movdqu	(@X[-2&7],&QWP(32,$inp));
+	&movdqu	(@X[-1&7],&QWP(48,$inp));
+	&add	($inp,64);
+	&pshufb	(@X[-4&7],@X[2]);		# byte swap
+	&mov	(&DWP(192+4,"esp"),$inp);
+	&movdqa	(&QWP(112-16,"esp"),@X[3]);	# borrow last backtrace slot
+
+  $Xi=0;
+}
+
+sub Xloop_ssse3()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&pshufb	(@X[($Xi-3)&7],@X[2]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&paddd	(@X[($Xi-4)&7],@X[3]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&movdqa	(&QWP(0+16*$Xi,"esp"),@X[($Xi-4)&7]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&psubd	(@X[($Xi-4)&7],@X[3]);
+
+	foreach (@insns) { eval; }
+  $Xi++;
+}
+
+sub Xtail_ssse3()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	foreach (@insns) { eval; }
+}
+
+sub body_00_19 () {
+	(
+	'($a,$b,$c,$d,$e)=@V;'.
+	'&add	($e,&DWP(4*($j&15),"esp"));',	# X[]+K xfer
+	'&xor	($c,$d);',
+	'&mov	(@T[1],$a);',	# $b in next round
+	'&$_rol	($a,5);',
+	'&and	(@T[0],$c);',	# ($b&($c^$d))
+	'&xor	($c,$d);',	# restore $c
+	'&xor	(@T[0],$d);',
+	'&add	($e,$a);',
+	'&$_ror	($b,$j?7:2);',	# $b>>>2
+	'&add	($e,@T[0]);'	.'$j++; unshift(@V,pop(@V)); unshift(@T,pop(@T));'
+	);
+}
+
+sub body_20_39 () {
+	(
+	'($a,$b,$c,$d,$e)=@V;'.
+	'&add	($e,&DWP(4*($j++&15),"esp"));',	# X[]+K xfer
+	'&xor	(@T[0],$d);',	# ($b^$d)
+	'&mov	(@T[1],$a);',	# $b in next round
+	'&$_rol	($a,5);',
+	'&xor	(@T[0],$c);',	# ($b^$d^$c)
+	'&add	($e,$a);',
+	'&$_ror	($b,7);',	# $b>>>2
+	'&add	($e,@T[0]);'	.'unshift(@V,pop(@V)); unshift(@T,pop(@T));'
+	);
+}
+
+sub body_40_59 () {
+	(
+	'($a,$b,$c,$d,$e)=@V;'.
+	'&mov	(@T[1],$c);',
+	'&xor	($c,$d);',
+	'&add	($e,&DWP(4*($j++&15),"esp"));',	# X[]+K xfer
+	'&and	(@T[1],$d);',
+	'&and	(@T[0],$c);',	# ($b&($c^$d))
+	'&$_ror	($b,7);',	# $b>>>2
+	'&add	($e,@T[1]);',
+	'&mov	(@T[1],$a);',	# $b in next round
+	'&$_rol	($a,5);',
+	'&add	($e,@T[0]);',
+	'&xor	($c,$d);',	# restore $c
+	'&add	($e,$a);'	.'unshift(@V,pop(@V)); unshift(@T,pop(@T));'
+	);
+}
+
+&set_label("loop",16);
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_32_79(\&body_00_19);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xuplast_ssse3_80(\&body_20_39);	# can jump to "done"
+
+				$saved_j=$j; @saved_V=@V;
+
+	&Xloop_ssse3(\&body_20_39);
+	&Xloop_ssse3(\&body_20_39);
+	&Xloop_ssse3(\&body_20_39);
+
+	&mov	(@T[1],&DWP(192,"esp"));	# update context
+	&add	($A,&DWP(0,@T[1]));
+	&add	(@T[0],&DWP(4,@T[1]));		# $b
+	&add	($C,&DWP(8,@T[1]));
+	&mov	(&DWP(0,@T[1]),$A);
+	&add	($D,&DWP(12,@T[1]));
+	&mov	(&DWP(4,@T[1]),@T[0]);
+	&add	($E,&DWP(16,@T[1]));
+	&mov	(&DWP(8,@T[1]),$C);
+	&mov	($B,@T[0]);
+	&mov	(&DWP(12,@T[1]),$D);
+	&mov	(&DWP(16,@T[1]),$E);
+	&movdqa	(@X[0],@X[-3&7]);
+
+	&jmp	(&label("loop"));
+
+&set_label("done",16);		$j=$saved_j; @V=@saved_V;
+
+	&Xtail_ssse3(\&body_20_39);
+	&Xtail_ssse3(\&body_20_39);
+	&Xtail_ssse3(\&body_20_39);
+
+	&mov	(@T[1],&DWP(192,"esp"));	# update context
+	&add	($A,&DWP(0,@T[1]));
+	&mov	("esp",&DWP(192+12,"esp"));	# restore %esp
+	&add	(@T[0],&DWP(4,@T[1]));		# $b
+	&add	($C,&DWP(8,@T[1]));
+	&mov	(&DWP(0,@T[1]),$A);
+	&add	($D,&DWP(12,@T[1]));
+	&mov	(&DWP(4,@T[1]),@T[0]);
+	&add	($E,&DWP(16,@T[1]));
+	&mov	(&DWP(8,@T[1]),$C);
+	&mov	(&DWP(12,@T[1]),$D);
+	&mov	(&DWP(16,@T[1]),$E);
+
+&function_end("_sha1_block_data_order_ssse3");
+
+if ($ymm) {
+my $Xi=4;			# 4xSIMD Xupdate round, start pre-seeded
+my @X=map("xmm$_",(4..7,0..3));	# pre-seeded for $Xi=4
+my @V=($A,$B,$C,$D,$E);
+my $j=0;			# hash round
+my @T=($T,$tmp1);
+my $inp;
+
+my $_rol=sub { &shld(@_[0],@_) };
+my $_ror=sub { &shrd(@_[0],@_) };
+
+&function_begin("_sha1_block_data_order_avx");
+	&call	(&label("pic_point"));	# make it PIC!
+	&set_label("pic_point");
+	&blindpop($tmp1);
+	&lea	($tmp1,&DWP(&label("K_XX_XX")."-".&label("pic_point"),$tmp1));
+&set_label("avx_shortcut");
+	&vzeroall();
+
+	&vmovdqa(@X[3],&QWP(0,$tmp1));		# K_00_19
+	&vmovdqa(@X[4],&QWP(16,$tmp1));		# K_20_39
+	&vmovdqa(@X[5],&QWP(32,$tmp1));		# K_40_59
+	&vmovdqa(@X[6],&QWP(48,$tmp1));		# K_60_79
+	&vmovdqa(@X[2],&QWP(64,$tmp1));		# pbswap mask
+
+	&mov	($E,&wparam(0));		# load argument block
+	&mov	($inp=@T[1],&wparam(1));
+	&mov	($D,&wparam(2));
+	&mov	(@T[0],"esp");
+
+	# stack frame layout
+	#
+	# +0	X[0]+K	X[1]+K	X[2]+K	X[3]+K	# XMM->IALU xfer area
+	#	X[4]+K	X[5]+K	X[6]+K	X[7]+K
+	#	X[8]+K	X[9]+K	X[10]+K	X[11]+K
+	#	X[12]+K	X[13]+K	X[14]+K	X[15]+K
+	#
+	# +64	X[0]	X[1]	X[2]	X[3]	# XMM->XMM backtrace area
+	#	X[4]	X[5]	X[6]	X[7]
+	#	X[8]	X[9]	X[10]	X[11]	# even borrowed for K_00_19
+	#
+	# +112	K_20_39	K_20_39	K_20_39	K_20_39	# constants
+	#	K_40_59	K_40_59	K_40_59	K_40_59
+	#	K_60_79	K_60_79	K_60_79	K_60_79
+	#	K_00_19	K_00_19	K_00_19	K_00_19
+	#	pbswap mask
+	#
+	# +192	ctx				# argument block
+	# +196	inp
+	# +200	end
+	# +204	esp
+	&sub	("esp",208);
+	&and	("esp",-64);
+
+	&vmovdqa(&QWP(112+0,"esp"),@X[4]);	# copy constants
+	&vmovdqa(&QWP(112+16,"esp"),@X[5]);
+	&vmovdqa(&QWP(112+32,"esp"),@X[6]);
+	&shl	($D,6);				# len*64
+	&vmovdqa(&QWP(112+48,"esp"),@X[3]);
+	&add	($D,$inp);			# end of input
+	&vmovdqa(&QWP(112+64,"esp"),@X[2]);
+	&add	($inp,64);
+	&mov	(&DWP(192+0,"esp"),$E);		# save argument block
+	&mov	(&DWP(192+4,"esp"),$inp);
+	&mov	(&DWP(192+8,"esp"),$D);
+	&mov	(&DWP(192+12,"esp"),@T[0]);	# save original %esp
+
+	&mov	($A,&DWP(0,$E));		# load context
+	&mov	($B,&DWP(4,$E));
+	&mov	($C,&DWP(8,$E));
+	&mov	($D,&DWP(12,$E));
+	&mov	($E,&DWP(16,$E));
+	&mov	(@T[0],$B);			# magic seed
+
+	&vmovdqu(@X[-4&7],&QWP(-64,$inp));	# load input to %xmm[0-3]
+	&vmovdqu(@X[-3&7],&QWP(-48,$inp));
+	&vmovdqu(@X[-2&7],&QWP(-32,$inp));
+	&vmovdqu(@X[-1&7],&QWP(-16,$inp));
+	&vpshufb(@X[-4&7],@X[-4&7],@X[2]);	# byte swap
+	&vpshufb(@X[-3&7],@X[-3&7],@X[2]);
+	&vpshufb(@X[-2&7],@X[-2&7],@X[2]);
+	&vmovdqa(&QWP(112-16,"esp"),@X[3]);	# borrow last backtrace slot
+	&vpshufb(@X[-1&7],@X[-1&7],@X[2]);
+	&vpaddd	(@X[0],@X[-4&7],@X[3]);		# add K_00_19
+	&vpaddd	(@X[1],@X[-3&7],@X[3]);
+	&vpaddd	(@X[2],@X[-2&7],@X[3]);
+	&vmovdqa(&QWP(0,"esp"),@X[0]);		# X[]+K xfer to IALU
+	&vmovdqa(&QWP(0+16,"esp"),@X[1]);
+	&vmovdqa(&QWP(0+32,"esp"),@X[2]);
+	&jmp	(&label("loop"));
+
+sub Xupdate_avx_16_31()		# recall that $Xi starts wtih 4
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 40 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpalignr(@X[0],@X[-3&7],@X[-4&7],8);	# compose "X[-14]" in "X[0]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &vpaddd	(@X[3],@X[3],@X[-1&7]);
+	  &vmovdqa	(&QWP(64+16*(($Xi-4)%3),"esp"),@X[-4&7]);# save X[] to backtrace buffer
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpsrldq(@X[2],@X[-1&7],4);		# "X[-3]", 3 dwords
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpxor	(@X[0],@X[0],@X[-4&7]);		# "X[0]"^="X[-16]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpxor	(@X[2],@X[2],@X[-2&7]);		# "X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vmovdqa	(&QWP(0+16*(($Xi-1)&3),"esp"),@X[3]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpxor	(@X[0],@X[0],@X[2]);		# "X[0]"^="X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpsrld	(@X[2],@X[0],31);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpslldq(@X[4],@X[0],12);		# "X[0]"<<96, extract one dword
+	&vpaddd	(@X[0],@X[0],@X[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpsrld	(@X[3],@X[4],30);
+	&vpor	(@X[0],@X[0],@X[2]);		# "X[0]"<<<=1
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpslld	(@X[4],@X[4],2);
+	  &vmovdqa	(@X[2],&QWP(64+16*(($Xi-6)%3),"esp")) if ($Xi>5);	# restore X[] from backtrace buffer
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpxor	(@X[0],@X[0],@X[3]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpxor	(@X[0],@X[0],@X[4]);		# "X[0]"^=("X[0]"<<96)<<<2
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vmovdqa	(@X[4],&QWP(112-16+16*(($Xi)/5),"esp"));	# K_XX_XX
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	 foreach (@insns) { eval; }	# remaining instructions [if any]
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+}
+
+sub Xupdate_avx_32_79()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 to 48 instructions
+  my ($a,$b,$c,$d,$e);
+
+	&vpalignr(@X[2],@X[-1&7],@X[-2&7],8);	# compose "X[-6]"
+	&vpxor	(@X[0],@X[0],@X[-4&7]);	# "X[0]"="X[-32]"^"X[-16]"
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&vpxor	(@X[0],@X[0],@X[-7&7]);	# "X[0]"^="X[-28]"
+	  &vmovdqa	(&QWP(64+16*(($Xi-4)%3),"esp"),@X[-4&7]);	# save X[] to backtrace buffer
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 if ($Xi%5) {
+	  &vmovdqa	(@X[4],@X[3]);	# "perpetuate" K_XX_XX...
+	 } else {			# ... or load next one
+	  &vmovdqa	(@X[4],&QWP(112-16+16*($Xi/5),"esp"));
+	 }
+	  &vpaddd	(@X[3],@X[3],@X[-1&7]);
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&vpxor	(@X[0],@X[0],@X[2]);		# "X[0]"^="X[-6]"
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&vpsrld	(@X[2],@X[0],30);
+	  &vmovdqa	(&QWP(0+16*(($Xi-1)&3),"esp"),@X[3]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&vpslld	(@X[0],@X[0],2);
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&vpor	(@X[0],@X[0],@X[2]);	# "X[0]"<<<=2
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	  &vmovdqa	(@X[2],&QWP(64+16*(($Xi-6)%3),"esp")) if($Xi<19);	# restore X[] from backtrace buffer
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	 foreach (@insns) { eval; }	# remaining instructions
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+}
+
+sub Xuplast_avx_80()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	  &vpaddd	(@X[3],@X[3],@X[-1&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &vmovdqa	(&QWP(0+16*(($Xi-1)&3),"esp"),@X[3]);	# X[]+K xfer IALU
+
+	 foreach (@insns) { eval; }		# remaining instructions
+
+	&mov	($inp=@T[1],&DWP(192+4,"esp"));
+	&cmp	($inp,&DWP(192+8,"esp"));
+	&je	(&label("done"));
+
+	&vmovdqa(@X[3],&QWP(112+48,"esp"));	# K_00_19
+	&vmovdqa(@X[2],&QWP(112+64,"esp"));	# pbswap mask
+	&vmovdqu(@X[-4&7],&QWP(0,$inp));	# load input
+	&vmovdqu(@X[-3&7],&QWP(16,$inp));
+	&vmovdqu(@X[-2&7],&QWP(32,$inp));
+	&vmovdqu(@X[-1&7],&QWP(48,$inp));
+	&add	($inp,64);
+	&vpshufb(@X[-4&7],@X[-4&7],@X[2]);		# byte swap
+	&mov	(&DWP(192+4,"esp"),$inp);
+	&vmovdqa(&QWP(112-16,"esp"),@X[3]);	# borrow last backtrace slot
+
+  $Xi=0;
+}
+
+sub Xloop_avx()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpshufb	(@X[($Xi-3)&7],@X[($Xi-3)&7],@X[2]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpaddd	(@X[$Xi&7],@X[($Xi-4)&7],@X[3]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vmovdqa	(&QWP(0+16*$Xi,"esp"),@X[$Xi&7]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	foreach (@insns) { eval; }
+  $Xi++;
+}
+
+sub Xtail_avx()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	foreach (@insns) { eval; }
+}
+
+&set_label("loop",16);
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_32_79(\&body_00_19);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xuplast_avx_80(\&body_20_39);	# can jump to "done"
+
+				$saved_j=$j; @saved_V=@V;
+
+	&Xloop_avx(\&body_20_39);
+	&Xloop_avx(\&body_20_39);
+	&Xloop_avx(\&body_20_39);
+
+	&mov	(@T[1],&DWP(192,"esp"));	# update context
+	&add	($A,&DWP(0,@T[1]));
+	&add	(@T[0],&DWP(4,@T[1]));		# $b
+	&add	($C,&DWP(8,@T[1]));
+	&mov	(&DWP(0,@T[1]),$A);
+	&add	($D,&DWP(12,@T[1]));
+	&mov	(&DWP(4,@T[1]),@T[0]);
+	&add	($E,&DWP(16,@T[1]));
+	&mov	(&DWP(8,@T[1]),$C);
+	&mov	($B,@T[0]);
+	&mov	(&DWP(12,@T[1]),$D);
+	&mov	(&DWP(16,@T[1]),$E);
+
+	&jmp	(&label("loop"));
+
+&set_label("done",16);		$j=$saved_j; @V=@saved_V;
+
+	&Xtail_avx(\&body_20_39);
+	&Xtail_avx(\&body_20_39);
+	&Xtail_avx(\&body_20_39);
+
+	&vzeroall();
+
+	&mov	(@T[1],&DWP(192,"esp"));	# update context
+	&add	($A,&DWP(0,@T[1]));
+	&mov	("esp",&DWP(192+12,"esp"));	# restore %esp
+	&add	(@T[0],&DWP(4,@T[1]));		# $b
+	&add	($C,&DWP(8,@T[1]));
+	&mov	(&DWP(0,@T[1]),$A);
+	&add	($D,&DWP(12,@T[1]));
+	&mov	(&DWP(4,@T[1]),@T[0]);
+	&add	($E,&DWP(16,@T[1]));
+	&mov	(&DWP(8,@T[1]),$C);
+	&mov	(&DWP(12,@T[1]),$D);
+	&mov	(&DWP(16,@T[1]),$E);
+&function_end("_sha1_block_data_order_avx");
+}
+&set_label("K_XX_XX",64);
+&data_word(0x5a827999,0x5a827999,0x5a827999,0x5a827999);	# K_00_19
+&data_word(0x6ed9eba1,0x6ed9eba1,0x6ed9eba1,0x6ed9eba1);	# K_20_39
+&data_word(0x8f1bbcdc,0x8f1bbcdc,0x8f1bbcdc,0x8f1bbcdc);	# K_40_59
+&data_word(0xca62c1d6,0xca62c1d6,0xca62c1d6,0xca62c1d6);	# K_60_79
+&data_word(0x00010203,0x04050607,0x08090a0b,0x0c0d0e0f);	# pbswap mask
+}
+&asciz("SHA1 block transform for x86, CRYPTOGAMS by <appro\@openssl.org>");
+
+&asm_finish();
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha1-alpha.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha1-alpha.pl
new file mode 100644
index 0000000000..6c4b9251fd
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha1-alpha.pl
@@ -0,0 +1,322 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> 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 block procedure for Alpha.
+
+# On 21264 performance is 33% better than code generated by vendor
+# compiler, and 75% better than GCC [3.4], and in absolute terms is
+# 8.7 cycles per processed byte. Implementation features vectorized
+# byte swap, but not Xupdate.
+
+@X=(	"\$0",	"\$1",	"\$2",	"\$3",	"\$4",	"\$5",	"\$6",	"\$7",
+	"\$8",	"\$9",	"\$10",	"\$11",	"\$12",	"\$13",	"\$14",	"\$15");
+$ctx="a0";	# $16
+$inp="a1";
+$num="a2";
+$A="a3";
+$B="a4";	# 20
+$C="a5";
+$D="t8";
+$E="t9";	@V=($A,$B,$C,$D,$E);
+$t0="t10";	# 24
+$t1="t11";
+$t2="ra";
+$t3="t12";
+$K="AT";	# 28
+
+sub BODY_00_19 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+$code.=<<___ if ($i==0);
+	ldq_u	@X[0],0+0($inp)
+	ldq_u	@X[1],0+7($inp)
+___
+$code.=<<___ if (!($i&1) && $i<14);
+	ldq_u	@X[$i+2],($i+2)*4+0($inp)
+	ldq_u	@X[$i+3],($i+2)*4+7($inp)
+___
+$code.=<<___ if (!($i&1) && $i<15);
+	extql	@X[$i],$inp,@X[$i]
+	extqh	@X[$i+1],$inp,@X[$i+1]
+
+	or	@X[$i+1],@X[$i],@X[$i]	# pair of 32-bit values are fetched
+
+	srl	@X[$i],24,$t0		# vectorized byte swap
+	srl	@X[$i],8,$t2
+
+	sll	@X[$i],8,$t3
+	sll	@X[$i],24,@X[$i]
+	zapnot	$t0,0x11,$t0
+	zapnot	$t2,0x22,$t2
+
+	zapnot	@X[$i],0x88,@X[$i]
+	or	$t0,$t2,$t0
+	zapnot	$t3,0x44,$t3
+	sll	$a,5,$t1
+
+	or	@X[$i],$t0,@X[$i]
+	addl	$K,$e,$e
+	and	$b,$c,$t2
+	zapnot	$a,0xf,$a
+
+	or	@X[$i],$t3,@X[$i]
+	srl	$a,27,$t0
+	bic	$d,$b,$t3
+	sll	$b,30,$b
+
+	extll	@X[$i],4,@X[$i+1]	# extract upper half
+	or	$t2,$t3,$t2
+	addl	@X[$i],$e,$e
+
+	addl	$t1,$e,$e
+	srl	$b,32,$t3
+	zapnot	@X[$i],0xf,@X[$i]
+
+	addl	$t0,$e,$e
+	addl	$t2,$e,$e
+	or	$t3,$b,$b
+___
+$code.=<<___ if (($i&1) && $i<15);
+	sll	$a,5,$t1
+	addl	$K,$e,$e
+	and	$b,$c,$t2
+	zapnot	$a,0xf,$a
+
+	srl	$a,27,$t0
+	addl	@X[$i%16],$e,$e
+	bic	$d,$b,$t3
+	sll	$b,30,$b
+
+	or	$t2,$t3,$t2
+	addl	$t1,$e,$e
+	srl	$b,32,$t3
+	zapnot	@X[$i],0xf,@X[$i]
+
+	addl	$t0,$e,$e
+	addl	$t2,$e,$e
+	or	$t3,$b,$b
+___
+$code.=<<___ if ($i>=15);	# with forward Xupdate
+	sll	$a,5,$t1
+	addl	$K,$e,$e
+	and	$b,$c,$t2
+	xor	@X[($j+2)%16],@X[$j%16],@X[$j%16]
+
+	zapnot	$a,0xf,$a
+	addl	@X[$i%16],$e,$e
+	bic	$d,$b,$t3
+	xor	@X[($j+8)%16],@X[$j%16],@X[$j%16]
+
+	srl	$a,27,$t0
+	addl	$t1,$e,$e
+	or	$t2,$t3,$t2
+	xor	@X[($j+13)%16],@X[$j%16],@X[$j%16]
+
+	sll	$b,30,$b
+	addl	$t0,$e,$e
+	srl	@X[$j%16],31,$t1
+
+	addl	$t2,$e,$e
+	srl	$b,32,$t3
+	addl	@X[$j%16],@X[$j%16],@X[$j%16]
+
+	or	$t3,$b,$b
+	zapnot	@X[$i%16],0xf,@X[$i%16]
+	or	$t1,@X[$j%16],@X[$j%16]
+___
+}
+
+sub BODY_20_39 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+$code.=<<___ if ($i<79);	# with forward Xupdate
+	sll	$a,5,$t1
+	addl	$K,$e,$e
+	zapnot	$a,0xf,$a
+	xor	@X[($j+2)%16],@X[$j%16],@X[$j%16]
+
+	sll	$b,30,$t3
+	addl	$t1,$e,$e
+	xor	$b,$c,$t2
+	xor	@X[($j+8)%16],@X[$j%16],@X[$j%16]
+
+	srl	$b,2,$b
+	addl	@X[$i%16],$e,$e
+	xor	$d,$t2,$t2
+	xor	@X[($j+13)%16],@X[$j%16],@X[$j%16]
+
+	srl	@X[$j%16],31,$t1
+	addl	$t2,$e,$e
+	srl	$a,27,$t0
+	addl	@X[$j%16],@X[$j%16],@X[$j%16]
+
+	or	$t3,$b,$b
+	addl	$t0,$e,$e
+	or	$t1,@X[$j%16],@X[$j%16]
+___
+$code.=<<___ if ($i<77);
+	zapnot	@X[$i%16],0xf,@X[$i%16]
+___
+$code.=<<___ if ($i==79);	# with context fetch
+	sll	$a,5,$t1
+	addl	$K,$e,$e
+	zapnot	$a,0xf,$a
+	ldl	@X[0],0($ctx)
+
+	sll	$b,30,$t3
+	addl	$t1,$e,$e
+	xor	$b,$c,$t2
+	ldl	@X[1],4($ctx)
+
+	srl	$b,2,$b
+	addl	@X[$i%16],$e,$e
+	xor	$d,$t2,$t2
+	ldl	@X[2],8($ctx)
+
+	srl	$a,27,$t0
+	addl	$t2,$e,$e
+	ldl	@X[3],12($ctx)
+
+	or	$t3,$b,$b
+	addl	$t0,$e,$e
+	ldl	@X[4],16($ctx)
+___
+}
+
+sub BODY_40_59 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+$code.=<<___;	# with forward Xupdate
+	sll	$a,5,$t1
+	addl	$K,$e,$e
+	zapnot	$a,0xf,$a
+	xor	@X[($j+2)%16],@X[$j%16],@X[$j%16]
+
+	srl	$a,27,$t0
+	and	$b,$c,$t2
+	and	$b,$d,$t3
+	xor	@X[($j+8)%16],@X[$j%16],@X[$j%16]
+
+	sll	$b,30,$b
+	addl	$t1,$e,$e
+	xor	@X[($j+13)%16],@X[$j%16],@X[$j%16]
+
+	srl	@X[$j%16],31,$t1
+	addl	$t0,$e,$e
+	or	$t2,$t3,$t2
+	and	$c,$d,$t3
+
+	or	$t2,$t3,$t2
+	srl	$b,32,$t3
+	addl	@X[$i%16],$e,$e
+	addl	@X[$j%16],@X[$j%16],@X[$j%16]
+
+	or	$t3,$b,$b
+	addl	$t2,$e,$e
+	or	$t1,@X[$j%16],@X[$j%16]
+	zapnot	@X[$i%16],0xf,@X[$i%16]
+___
+}
+
+$code=<<___;
+#ifdef __linux__
+#include <asm/regdef.h>
+#else
+#include <asm.h>
+#include <regdef.h>
+#endif
+
+.text
+
+.set	noat
+.set	noreorder
+.globl	sha1_block_data_order
+.align	5
+.ent	sha1_block_data_order
+sha1_block_data_order:
+	lda	sp,-64(sp)
+	stq	ra,0(sp)
+	stq	s0,8(sp)
+	stq	s1,16(sp)
+	stq	s2,24(sp)
+	stq	s3,32(sp)
+	stq	s4,40(sp)
+	stq	s5,48(sp)
+	stq	fp,56(sp)
+	.mask	0x0400fe00,-64
+	.frame	sp,64,ra
+	.prologue 0
+
+	ldl	$A,0($ctx)
+	ldl	$B,4($ctx)
+	sll	$num,6,$num
+	ldl	$C,8($ctx)
+	ldl	$D,12($ctx)
+	ldl	$E,16($ctx)
+	addq	$inp,$num,$num
+
+.Lloop:
+	.set	noreorder
+	ldah	$K,23170(zero)
+	zapnot	$B,0xf,$B
+	lda	$K,31129($K)	# K_00_19
+___
+for ($i=0;$i<20;$i++) { &BODY_00_19($i,@V); unshift(@V,pop(@V)); }
+
+$code.=<<___;
+	ldah	$K,28378(zero)
+	lda	$K,-5215($K)	# K_20_39
+___
+for (;$i<40;$i++) { &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+
+$code.=<<___;
+	ldah	$K,-28900(zero)
+	lda	$K,-17188($K)	# K_40_59
+___
+for (;$i<60;$i++) { &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
+
+$code.=<<___;
+	ldah	$K,-13725(zero)
+	lda	$K,-15914($K)	# K_60_79
+___
+for (;$i<80;$i++) { &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+
+$code.=<<___;
+	addl	@X[0],$A,$A
+	addl	@X[1],$B,$B
+	addl	@X[2],$C,$C
+	addl	@X[3],$D,$D
+	addl	@X[4],$E,$E
+	stl	$A,0($ctx)
+	stl	$B,4($ctx)
+	addq	$inp,64,$inp
+	stl	$C,8($ctx)
+	stl	$D,12($ctx)
+	stl	$E,16($ctx)
+	cmpult	$inp,$num,$t1
+	bne	$t1,.Lloop
+
+	.set	noreorder
+	ldq	ra,0(sp)
+	ldq	s0,8(sp)
+	ldq	s1,16(sp)
+	ldq	s2,24(sp)
+	ldq	s3,32(sp)
+	ldq	s4,40(sp)
+	ldq	s5,48(sp)
+	ldq	fp,56(sp)
+	lda	sp,64(sp)
+	ret	(ra)
+.end	sha1_block_data_order
+.ascii	"SHA1 block transform for Alpha, CRYPTOGAMS by <appro\@openssl.org>"
+.align	2
+___
+$output=shift and open STDOUT,">$output";
+print $code;
+close STDOUT;
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha1-armv4-large.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha1-armv4-large.pl
new file mode 100644
index 0000000000..33da3e0e3c
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha1-armv4-large.pl
@@ -0,0 +1,248 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> 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_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.
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+$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=<<___;
+#include "arm_arch.h"
+
+.text
+
+.global	sha1_block_data_order
+.type	sha1_block_data_order,%function
+
+.align	2
+sha1_block_data_order:
+	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.=<<___;
+	teq	$Xi,sp
+	bne	.L_00_15		@ [((11+4)*5+2)*3]
+	sub	sp,sp,#25*4
+___
+	&BODY_00_15(@V);	unshift(@V,pop(@V));
+	&BODY_16_19(@V);	unshift(@V,pop(@V));
+	&BODY_16_19(@V);	unshift(@V,pop(@V));
+	&BODY_16_19(@V);	unshift(@V,pop(@V));
+	&BODY_16_19(@V);	unshift(@V,pop(@V));
+$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.=<<___;
+	teq	$Xi,sp			@ preserve carry
+	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.=<<___;
+	teq	$Xi,sp
+	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
+.align	2
+.LK_00_19:	.word	0x5a827999
+.LK_20_39:	.word	0x6ed9eba1
+.LK_40_59:	.word	0x8f1bbcdc
+.LK_60_79:	.word	0xca62c1d6
+.size	sha1_block_data_order,.-sha1_block_data_order
+.asciz	"SHA1 block transform for ARMv4, CRYPTOGAMS by <appro\@openssl.org>"
+.align	2
+___
+
+$code =~ s/\bbx\s+lr\b/.word\t0xe12fff1e/gm;	# make it possible to compile with -march=armv4
+print $code;
+close STDOUT; # enforce flush
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha1-ia64.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha1-ia64.pl
new file mode 100644
index 0000000000..02d35d1614
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha1-ia64.pl
@@ -0,0 +1,305 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# Eternal question is what's wrong with compiler generated code? The
+# trick is that it's possible to reduce the number of shifts required
+# to perform rotations by maintaining copy of 32-bit value in upper
+# bits of 64-bit register. Just follow mux2 and shrp instructions...
+# Performance under big-endian OS such as HP-UX is 179MBps*1GHz, which
+# is >50% better than HP C and >2x better than gcc.
+
+$code=<<___;
+.ident  \"sha1-ia64.s, version 1.3\"
+.ident  \"IA-64 ISA artwork by Andy Polyakov <appro\@fy.chalmers.se>\"
+.explicit
+
+___
+
+
+if ($^O eq "hpux") {
+    $ADDP="addp4";
+    for (@ARGV) { $ADDP="add" if (/[\+DD|\-mlp]64/); }
+} else { $ADDP="add"; }
+
+#$human=1;
+if ($human) {	# useful for visual code auditing...
+	($A,$B,$C,$D,$E)   = ("A","B","C","D","E");
+	($h0,$h1,$h2,$h3,$h4) = ("h0","h1","h2","h3","h4");
+	($K_00_19, $K_20_39, $K_40_59, $K_60_79) =
+	    (	"K_00_19","K_20_39","K_40_59","K_60_79"	);
+	@X= (	"X0", "X1", "X2", "X3", "X4", "X5", "X6", "X7",
+		"X8", "X9","X10","X11","X12","X13","X14","X15"	);
+}
+else {
+	($A,$B,$C,$D,$E)   =    ("loc0","loc1","loc2","loc3","loc4");
+	($h0,$h1,$h2,$h3,$h4) = ("loc5","loc6","loc7","loc8","loc9");
+	($K_00_19, $K_20_39, $K_40_59, $K_60_79) =
+	    (	"r14", "r15", "loc10", "loc11"	);
+	@X= (	"r16", "r17", "r18", "r19", "r20", "r21", "r22", "r23",
+		"r24", "r25", "r26", "r27", "r28", "r29", "r30", "r31"	);
+}
+
+sub BODY_00_15 {
+local	*code=shift;
+my	($i,$a,$b,$c,$d,$e)=@_;
+my	$j=$i+1;
+my	$Xn=@X[$j%16];
+
+$code.=<<___ if ($i==0);
+{ .mmi;	ld1	$X[$i]=[inp],2		    // MSB
+	ld1	tmp2=[tmp3],2		};;
+{ .mmi;	ld1	tmp0=[inp],2
+	ld1	tmp4=[tmp3],2		    // LSB
+	dep	$X[$i]=$X[$i],tmp2,8,8	};;
+___
+if ($i<15) {
+	$code.=<<___;
+{ .mmi;	ld1	$Xn=[inp],2		    // forward Xload
+	nop.m	0x0
+	dep	tmp1=tmp0,tmp4,8,8	};;
+{ .mmi;	ld1	tmp2=[tmp3],2		    // forward Xload
+	and	tmp4=$c,$b
+	dep	$X[$i]=$X[$i],tmp1,16,16} //;;
+{ .mmi;	add	$e=$e,$K_00_19		    // e+=K_00_19
+	andcm	tmp1=$d,$b
+	dep.z	tmp5=$a,5,27		};; // a<<5
+{ .mmi;	add	$e=$e,$X[$i]		    // e+=Xload
+	or	tmp4=tmp4,tmp1		    // F_00_19(b,c,d)=(b&c)|(~b&d)
+	extr.u	tmp1=$a,27,5		};; // a>>27
+{ .mmi;	ld1	tmp0=[inp],2		    // forward Xload
+	add	$e=$e,tmp4		    // e+=F_00_19(b,c,d)
+	shrp	$b=tmp6,tmp6,2		}   // b=ROTATE(b,30)
+{ .mmi;	ld1	tmp4=[tmp3],2		    // forward Xload
+	or	tmp5=tmp1,tmp5		    // ROTATE(a,5)
+	mux2	tmp6=$a,0x44		};; // see b in next iteration
+{ .mii;	add	$e=$e,tmp5		    // e+=ROTATE(a,5)
+	dep	$Xn=$Xn,tmp2,8,8	    // forward Xload
+	mux2	$X[$i]=$X[$i],0x44	} //;;
+
+___
+	}
+else	{
+	$code.=<<___;
+{ .mii;	and	tmp3=$c,$b
+	dep	tmp1=tmp0,tmp4,8,8;;
+	dep	$X[$i]=$X[$i],tmp1,16,16} //;;
+{ .mmi;	add	$e=$e,$K_00_19		    // e+=K_00_19
+	andcm	tmp1=$d,$b
+	dep.z	tmp5=$a,5,27		};; // a<<5
+{ .mmi;	add	$e=$e,$X[$i]		    // e+=Xupdate
+	or	tmp4=tmp3,tmp1		    // F_00_19(b,c,d)=(b&c)|(~b&d)
+	extr.u	tmp1=$a,27,5		}   // a>>27
+{ .mmi;	xor	$Xn=$Xn,$X[($j+2)%16]	    // forward Xupdate
+	xor	tmp3=$X[($j+8)%16],$X[($j+13)%16] // forward Xupdate
+	nop.i	0			};;
+{ .mmi;	add	$e=$e,tmp4		    // e+=F_00_19(b,c,d)
+	xor	$Xn=$Xn,tmp3		    // forward Xupdate
+	shrp	$b=tmp6,tmp6,2		}   // b=ROTATE(b,30)
+{ .mmi; or	tmp1=tmp1,tmp5		    // ROTATE(a,5)
+	mux2	tmp6=$a,0x44		};; // see b in next iteration
+{ .mii;	add	$e=$e,tmp1		    // e+=ROTATE(a,5)
+	shrp	$Xn=$Xn,$Xn,31		    // ROTATE(x[0]^x[2]^x[8]^x[13],1)
+	mux2	$X[$i]=$X[$i],0x44	};;
+
+___
+	}
+}
+
+sub BODY_16_19 {
+local	*code=shift;
+my	($i,$a,$b,$c,$d,$e)=@_;
+my	$j=$i+1;
+my	$Xn=@X[$j%16];
+
+$code.=<<___;
+{ .mib;	add	$e=$e,$K_00_19		    // e+=K_00_19
+	dep.z	tmp5=$a,5,27		}   // a<<5
+{ .mib;	andcm	tmp1=$d,$b
+	and	tmp0=$c,$b		};;
+{ .mmi;	add	$e=$e,$X[$i%16]		    // e+=Xupdate
+	or	tmp0=tmp0,tmp1		    // F_00_19(b,c,d)=(b&c)|(~b&d)
+	extr.u	tmp1=$a,27,5		}   // a>>27
+{ .mmi;	xor	$Xn=$Xn,$X[($j+2)%16]	    // forward Xupdate
+	xor	tmp3=$X[($j+8)%16],$X[($j+13)%16]	// forward Xupdate
+	nop.i	0			};;
+{ .mmi;	add	$e=$e,tmp0		    // f+=F_00_19(b,c,d)
+	xor	$Xn=$Xn,tmp3		    // forward Xupdate
+	shrp	$b=tmp6,tmp6,2		}   // b=ROTATE(b,30)
+{ .mmi;	or	tmp1=tmp1,tmp5		    // ROTATE(a,5)
+	mux2	tmp6=$a,0x44		};; // see b in next iteration
+{ .mii;	add	$e=$e,tmp1		    // e+=ROTATE(a,5)
+	shrp	$Xn=$Xn,$Xn,31		    // ROTATE(x[0]^x[2]^x[8]^x[13],1)
+	nop.i	0			};;
+
+___
+}
+
+sub BODY_20_39 {
+local	*code=shift;
+my	($i,$a,$b,$c,$d,$e,$Konst)=@_;
+	$Konst = $K_20_39 if (!defined($Konst));
+my	$j=$i+1;
+my	$Xn=@X[$j%16];
+
+if ($i<79) {
+$code.=<<___;
+{ .mib;	add	$e=$e,$Konst		    // e+=K_XX_XX
+	dep.z	tmp5=$a,5,27		}   // a<<5
+{ .mib;	xor	tmp0=$c,$b
+	xor	$Xn=$Xn,$X[($j+2)%16]	};; // forward Xupdate
+{ .mib;	add	$e=$e,$X[$i%16]		    // e+=Xupdate
+	extr.u	tmp1=$a,27,5		}   // a>>27
+{ .mib;	xor	tmp0=tmp0,$d		    // F_20_39(b,c,d)=b^c^d
+	xor	$Xn=$Xn,$X[($j+8)%16]	};; // forward Xupdate
+{ .mmi;	add	$e=$e,tmp0		    // e+=F_20_39(b,c,d)
+	xor	$Xn=$Xn,$X[($j+13)%16]	    // forward Xupdate
+	shrp	$b=tmp6,tmp6,2		}   // b=ROTATE(b,30)
+{ .mmi;	or	tmp1=tmp1,tmp5		    // ROTATE(a,5)
+	mux2	tmp6=$a,0x44		};; // see b in next iteration
+{ .mii;	add	$e=$e,tmp1		    // e+=ROTATE(a,5)
+	shrp	$Xn=$Xn,$Xn,31		    // ROTATE(x[0]^x[2]^x[8]^x[13],1)
+	nop.i	0			};;
+
+___
+}
+else {
+$code.=<<___;
+{ .mib;	add	$e=$e,$Konst		    // e+=K_60_79
+	dep.z	tmp5=$a,5,27		}   // a<<5
+{ .mib;	xor	tmp0=$c,$b
+	add	$h1=$h1,$a		};; // wrap up
+{ .mib;	add	$e=$e,$X[$i%16]		    // e+=Xupdate
+	extr.u	tmp1=$a,27,5		}   // a>>27
+{ .mib;	xor	tmp0=tmp0,$d		    // F_20_39(b,c,d)=b^c^d
+	add	$h3=$h3,$c		};; // wrap up
+{ .mmi;	add	$e=$e,tmp0		    // e+=F_20_39(b,c,d)
+	or	tmp1=tmp1,tmp5		    // ROTATE(a,5)
+	shrp	$b=tmp6,tmp6,2		};; // b=ROTATE(b,30) ;;?
+{ .mmi;	add	$e=$e,tmp1		    // e+=ROTATE(a,5)
+	add	tmp3=1,inp		    // used in unaligned codepath
+	add	$h4=$h4,$d		};; // wrap up
+
+___
+}
+}
+
+sub BODY_40_59 {
+local	*code=shift;
+my	($i,$a,$b,$c,$d,$e)=@_;
+my	$j=$i+1;
+my	$Xn=@X[$j%16];
+
+$code.=<<___;
+{ .mib;	add	$e=$e,$K_40_59		    // e+=K_40_59
+	dep.z	tmp5=$a,5,27		}   // a<<5
+{ .mib;	and	tmp1=$c,$d
+	xor	tmp0=$c,$d		};;
+{ .mmi;	add	$e=$e,$X[$i%16]		    // e+=Xupdate
+	add	tmp5=tmp5,tmp1		    // a<<5+(c&d)
+	extr.u	tmp1=$a,27,5		}   // a>>27
+{ .mmi;	and	tmp0=tmp0,$b
+	xor	$Xn=$Xn,$X[($j+2)%16]	    // forward Xupdate
+	xor	tmp3=$X[($j+8)%16],$X[($j+13)%16] };;	// forward Xupdate
+{ .mmi;	add	$e=$e,tmp0		    // e+=b&(c^d)
+	add	tmp5=tmp5,tmp1		    // ROTATE(a,5)+(c&d)
+	shrp	$b=tmp6,tmp6,2		}   // b=ROTATE(b,30)
+{ .mmi;	xor	$Xn=$Xn,tmp3
+	mux2	tmp6=$a,0x44		};; // see b in next iteration
+{ .mii;	add	$e=$e,tmp5		    // e+=ROTATE(a,5)+(c&d)
+	shrp	$Xn=$Xn,$Xn,31		    // ROTATE(x[0]^x[2]^x[8]^x[13],1)
+	nop.i	0x0			};;
+
+___
+}
+sub BODY_60_79	{ &BODY_20_39(@_,$K_60_79); }
+
+$code.=<<___;
+.text
+
+tmp0=r8;
+tmp1=r9;
+tmp2=r10;
+tmp3=r11;
+ctx=r32;	// in0
+inp=r33;	// in1
+
+// void sha1_block_data_order(SHA_CTX *c,const void *p,size_t num);
+.global	sha1_block_data_order#
+.proc	sha1_block_data_order#
+.align	32
+sha1_block_data_order:
+	.prologue
+{ .mmi;	alloc	tmp1=ar.pfs,3,14,0,0
+	$ADDP	tmp0=4,ctx
+	.save	ar.lc,r3
+	mov	r3=ar.lc		}
+{ .mmi;	$ADDP	ctx=0,ctx
+	$ADDP	inp=0,inp
+	mov	r2=pr			};;
+tmp4=in2;
+tmp5=loc12;
+tmp6=loc13;
+	.body
+{ .mlx;	ld4	$h0=[ctx],8
+	movl	$K_00_19=0x5a827999	}
+{ .mlx;	ld4	$h1=[tmp0],8
+	movl	$K_20_39=0x6ed9eba1	};;
+{ .mlx;	ld4	$h2=[ctx],8
+	movl	$K_40_59=0x8f1bbcdc	}
+{ .mlx;	ld4	$h3=[tmp0]
+	movl	$K_60_79=0xca62c1d6	};;
+{ .mmi;	ld4	$h4=[ctx],-16
+	add	in2=-1,in2		    // adjust num for ar.lc
+	mov	ar.ec=1			};;
+{ .mmi;	nop.m	0
+	add	tmp3=1,inp
+	mov	ar.lc=in2		};; // brp.loop.imp: too far
+
+.Ldtop:
+{ .mmi;	mov	$A=$h0
+	mov	$B=$h1
+	mux2	tmp6=$h1,0x44		}
+{ .mmi;	mov	$C=$h2
+	mov	$D=$h3
+	mov	$E=$h4			};;
+
+___
+
+{ my $i;
+  my @V=($A,$B,$C,$D,$E);
+
+	for($i=0;$i<16;$i++)	{ &BODY_00_15(\$code,$i,@V); unshift(@V,pop(@V)); }
+	for(;$i<20;$i++)	{ &BODY_16_19(\$code,$i,@V); unshift(@V,pop(@V)); }
+	for(;$i<40;$i++)	{ &BODY_20_39(\$code,$i,@V); unshift(@V,pop(@V)); }
+	for(;$i<60;$i++)	{ &BODY_40_59(\$code,$i,@V); unshift(@V,pop(@V)); }
+	for(;$i<80;$i++)	{ &BODY_60_79(\$code,$i,@V); unshift(@V,pop(@V)); }
+
+	(($V[0] eq $A) and ($V[4] eq $E)) or die;	# double-check
+}
+
+$code.=<<___;
+{ .mmb;	add	$h0=$h0,$A
+	add	$h2=$h2,$C
+	br.ctop.dptk.many	.Ldtop	};;
+.Ldend:
+{ .mmi;	add	tmp0=4,ctx
+	mov	ar.lc=r3		};;
+{ .mmi;	st4	[ctx]=$h0,8
+	st4	[tmp0]=$h1,8		};;
+{ .mmi;	st4	[ctx]=$h2,8
+	st4	[tmp0]=$h3		};;
+{ .mib;	st4	[ctx]=$h4,-16
+	mov	pr=r2,0x1ffff
+	br.ret.sptk.many	b0	};;
+.endp	sha1_block_data_order#
+stringz	"SHA1 block transform for IA64, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+$output=shift and open STDOUT,">$output";
+print $code;
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha1-mips.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha1-mips.pl
new file mode 100644
index 0000000000..f1a702f38f
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha1-mips.pl
@@ -0,0 +1,354 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# SHA1 block procedure for MIPS.
+
+# Performance improvement is 30% on unaligned input. The "secret" is
+# to deploy lwl/lwr pair to load unaligned input. One could have
+# vectorized Xupdate on MIPSIII/IV, but the goal was to code MIPS32-
+# compatible subroutine. There is room for minor optimization on
+# little-endian platforms...
+
+######################################################################
+# There is a number of MIPS ABI in use, O32 and N32/64 are most
+# widely used. Then there is a new contender: NUBI. It appears that if
+# one picks the latter, it's possible to arrange code in ABI neutral
+# manner. Therefore let's stick to NUBI register layout:
+#
+($zero,$at,$t0,$t1,$t2)=map("\$$_",(0..2,24,25));
+($a0,$a1,$a2,$a3,$a4,$a5,$a6,$a7)=map("\$$_",(4..11));
+($s0,$s1,$s2,$s3,$s4,$s5,$s6,$s7,$s8,$s9,$s10,$s11)=map("\$$_",(12..23));
+($gp,$tp,$sp,$fp,$ra)=map("\$$_",(3,28..31));
+#
+# The return value is placed in $a0. Following coding rules facilitate
+# interoperability:
+#
+# - never ever touch $tp, "thread pointer", former $gp;
+# - copy return value to $t0, former $v0 [or to $a0 if you're adapting
+#   old code];
+# - on O32 populate $a4-$a7 with 'lw $aN,4*N($sp)' if necessary;
+#
+# For reference here is register layout for N32/64 MIPS ABIs:
+#
+# ($zero,$at,$v0,$v1)=map("\$$_",(0..3));
+# ($a0,$a1,$a2,$a3,$a4,$a5,$a6,$a7)=map("\$$_",(4..11));
+# ($t0,$t1,$t2,$t3,$t8,$t9)=map("\$$_",(12..15,24,25));
+# ($s0,$s1,$s2,$s3,$s4,$s5,$s6,$s7)=map("\$$_",(16..23));
+# ($gp,$sp,$fp,$ra)=map("\$$_",(28..31));
+#
+$flavour = shift; # supported flavours are o32,n32,64,nubi32,nubi64
+
+if ($flavour =~ /64|n32/i) {
+	$PTR_ADD="dadd";	# incidentally works even on n32
+	$PTR_SUB="dsub";	# incidentally works even on n32
+	$REG_S="sd";
+	$REG_L="ld";
+	$PTR_SLL="dsll";	# incidentally works even on n32
+	$SZREG=8;
+} else {
+	$PTR_ADD="add";
+	$PTR_SUB="sub";
+	$REG_S="sw";
+	$REG_L="lw";
+	$PTR_SLL="sll";
+	$SZREG=4;
+}
+#
+# <appro@openssl.org>
+#
+######################################################################
+
+$big_endian=(`echo MIPSEL | $ENV{CC} -E -P -`=~/MIPSEL/)?1:0;
+
+for (@ARGV) {	$output=$_ if (/^\w[\w\-]*\.\w+$/);   }
+open STDOUT,">$output";
+
+if (!defined($big_endian))
+            {   $big_endian=(unpack('L',pack('N',1))==1);   }
+
+# offsets of the Most and Least Significant Bytes
+$MSB=$big_endian?0:3;
+$LSB=3&~$MSB;
+
+@X=map("\$$_",(8..23));	# a4-a7,s0-s11
+
+$ctx=$a0;
+$inp=$a1;
+$num=$a2;
+$A="\$1";
+$B="\$2";
+$C="\$3";
+$D="\$7";
+$E="\$24";	@V=($A,$B,$C,$D,$E);
+$t0="\$25";
+$t1=$num;	# $num is offloaded to stack
+$t2="\$30";	# fp
+$K="\$31";	# ra
+
+sub BODY_00_14 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+$code.=<<___	if (!$big_endian);
+	srl	$t0,@X[$i],24	# byte swap($i)
+	srl	$t1,@X[$i],8
+	andi	$t2,@X[$i],0xFF00
+	sll	@X[$i],@X[$i],24
+	andi	$t1,0xFF00
+	sll	$t2,$t2,8
+	or	@X[$i],$t0
+	or	$t1,$t2
+	or	@X[$i],$t1
+___
+$code.=<<___;
+	 lwl	@X[$j],$j*4+$MSB($inp)
+	sll	$t0,$a,5	# $i
+	addu	$e,$K
+	 lwr	@X[$j],$j*4+$LSB($inp)
+	srl	$t1,$a,27
+	addu	$e,$t0
+	xor	$t0,$c,$d
+	addu	$e,$t1
+	sll	$t2,$b,30
+	and	$t0,$b
+	srl	$b,$b,2
+	xor	$t0,$d
+	addu	$e,@X[$i]
+	or	$b,$t2
+	addu	$e,$t0
+___
+}
+
+sub BODY_15_19 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+
+$code.=<<___	if (!$big_endian && $i==15);
+	srl	$t0,@X[$i],24	# byte swap($i)
+	srl	$t1,@X[$i],8
+	andi	$t2,@X[$i],0xFF00
+	sll	@X[$i],@X[$i],24
+	andi	$t1,0xFF00
+	sll	$t2,$t2,8
+	or	@X[$i],$t0
+	or	@X[$i],$t1
+	or	@X[$i],$t2
+___
+$code.=<<___;
+	 xor	@X[$j%16],@X[($j+2)%16]
+	sll	$t0,$a,5	# $i
+	addu	$e,$K
+	srl	$t1,$a,27
+	addu	$e,$t0
+	 xor	@X[$j%16],@X[($j+8)%16]
+	xor	$t0,$c,$d
+	addu	$e,$t1
+	 xor	@X[$j%16],@X[($j+13)%16]
+	sll	$t2,$b,30
+	and	$t0,$b
+	 srl	$t1,@X[$j%16],31
+	 addu	@X[$j%16],@X[$j%16]
+	srl	$b,$b,2
+	xor	$t0,$d
+	 or	@X[$j%16],$t1
+	addu	$e,@X[$i%16]
+	or	$b,$t2
+	addu	$e,$t0
+___
+}
+
+sub BODY_20_39 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+$code.=<<___ if ($i<79);
+	 xor	@X[$j%16],@X[($j+2)%16]
+	sll	$t0,$a,5	# $i
+	addu	$e,$K
+	srl	$t1,$a,27
+	addu	$e,$t0
+	 xor	@X[$j%16],@X[($j+8)%16]
+	xor	$t0,$c,$d
+	addu	$e,$t1
+	 xor	@X[$j%16],@X[($j+13)%16]
+	sll	$t2,$b,30
+	xor	$t0,$b
+	 srl	$t1,@X[$j%16],31
+	 addu	@X[$j%16],@X[$j%16]
+	srl	$b,$b,2
+	addu	$e,@X[$i%16]
+	 or	@X[$j%16],$t1
+	or	$b,$t2
+	addu	$e,$t0
+___
+$code.=<<___ if ($i==79);
+	 lw	@X[0],0($ctx)
+	sll	$t0,$a,5	# $i
+	addu	$e,$K
+	 lw	@X[1],4($ctx)
+	srl	$t1,$a,27
+	addu	$e,$t0
+	 lw	@X[2],8($ctx)
+	xor	$t0,$c,$d
+	addu	$e,$t1
+	 lw	@X[3],12($ctx)
+	sll	$t2,$b,30
+	xor	$t0,$b
+	 lw	@X[4],16($ctx)
+	srl	$b,$b,2
+	addu	$e,@X[$i%16]
+	or	$b,$t2
+	addu	$e,$t0
+___
+}
+
+sub BODY_40_59 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+$code.=<<___ if ($i<79);
+	 xor	@X[$j%16],@X[($j+2)%16]
+	sll	$t0,$a,5	# $i
+	addu	$e,$K
+	srl	$t1,$a,27
+	addu	$e,$t0
+	 xor	@X[$j%16],@X[($j+8)%16]
+	and	$t0,$c,$d
+	addu	$e,$t1
+	 xor	@X[$j%16],@X[($j+13)%16]
+	sll	$t2,$b,30
+	addu	$e,$t0
+	 srl	$t1,@X[$j%16],31
+	xor	$t0,$c,$d
+	 addu	@X[$j%16],@X[$j%16]
+	and	$t0,$b
+	srl	$b,$b,2
+	 or	@X[$j%16],$t1
+	addu	$e,@X[$i%16]
+	or	$b,$t2
+	addu	$e,$t0
+___
+}
+
+$FRAMESIZE=16;	# large enough to accomodate NUBI saved registers
+$SAVED_REGS_MASK = ($flavour =~ /nubi/i) ? 0xc0fff008 : 0xc0ff0000;
+
+$code=<<___;
+#ifdef OPENSSL_FIPSCANISTER
+# include <openssl/fipssyms.h>
+#endif
+
+.text
+
+.set	noat
+.set	noreorder
+.align	5
+.globl	sha1_block_data_order
+.ent	sha1_block_data_order
+sha1_block_data_order:
+	.frame	$sp,$FRAMESIZE*$SZREG,$ra
+	.mask	$SAVED_REGS_MASK,-$SZREG
+	.set	noreorder
+	$PTR_SUB $sp,$FRAMESIZE*$SZREG
+	$REG_S	$ra,($FRAMESIZE-1)*$SZREG($sp)
+	$REG_S	$fp,($FRAMESIZE-2)*$SZREG($sp)
+	$REG_S	$s11,($FRAMESIZE-3)*$SZREG($sp)
+	$REG_S	$s10,($FRAMESIZE-4)*$SZREG($sp)
+	$REG_S	$s9,($FRAMESIZE-5)*$SZREG($sp)
+	$REG_S	$s8,($FRAMESIZE-6)*$SZREG($sp)
+	$REG_S	$s7,($FRAMESIZE-7)*$SZREG($sp)
+	$REG_S	$s6,($FRAMESIZE-8)*$SZREG($sp)
+	$REG_S	$s5,($FRAMESIZE-9)*$SZREG($sp)
+	$REG_S	$s4,($FRAMESIZE-10)*$SZREG($sp)
+___
+$code.=<<___ if ($flavour =~ /nubi/i);	# optimize non-nubi prologue
+	$REG_S	$s3,($FRAMESIZE-11)*$SZREG($sp)
+	$REG_S	$s2,($FRAMESIZE-12)*$SZREG($sp)
+	$REG_S	$s1,($FRAMESIZE-13)*$SZREG($sp)
+	$REG_S	$s0,($FRAMESIZE-14)*$SZREG($sp)
+	$REG_S	$gp,($FRAMESIZE-15)*$SZREG($sp)
+___
+$code.=<<___;
+	$PTR_SLL $num,6
+	$PTR_ADD $num,$inp
+	$REG_S	$num,0($sp)
+	lw	$A,0($ctx)
+	lw	$B,4($ctx)
+	lw	$C,8($ctx)
+	lw	$D,12($ctx)
+	b	.Loop
+	lw	$E,16($ctx)
+.align	4
+.Loop:
+	.set	reorder
+	lwl	@X[0],$MSB($inp)
+	lui	$K,0x5a82
+	lwr	@X[0],$LSB($inp)
+	ori	$K,0x7999	# K_00_19
+___
+for ($i=0;$i<15;$i++)	{ &BODY_00_14($i,@V); unshift(@V,pop(@V)); }
+for (;$i<20;$i++)	{ &BODY_15_19($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	lui	$K,0x6ed9
+	ori	$K,0xeba1	# K_20_39
+___
+for (;$i<40;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	lui	$K,0x8f1b
+	ori	$K,0xbcdc	# K_40_59
+___
+for (;$i<60;$i++)	{ &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	lui	$K,0xca62
+	ori	$K,0xc1d6	# K_60_79
+___
+for (;$i<80;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	$PTR_ADD $inp,64
+	$REG_L	$num,0($sp)
+
+	addu	$A,$X[0]
+	addu	$B,$X[1]
+	sw	$A,0($ctx)
+	addu	$C,$X[2]
+	addu	$D,$X[3]
+	sw	$B,4($ctx)
+	addu	$E,$X[4]
+	sw	$C,8($ctx)
+	sw	$D,12($ctx)
+	sw	$E,16($ctx)
+	.set	noreorder
+	bne	$inp,$num,.Loop
+	nop
+
+	.set	noreorder
+	$REG_L	$ra,($FRAMESIZE-1)*$SZREG($sp)
+	$REG_L	$fp,($FRAMESIZE-2)*$SZREG($sp)
+	$REG_L	$s11,($FRAMESIZE-3)*$SZREG($sp)
+	$REG_L	$s10,($FRAMESIZE-4)*$SZREG($sp)
+	$REG_L	$s9,($FRAMESIZE-5)*$SZREG($sp)
+	$REG_L	$s8,($FRAMESIZE-6)*$SZREG($sp)
+	$REG_L	$s7,($FRAMESIZE-7)*$SZREG($sp)
+	$REG_L	$s6,($FRAMESIZE-8)*$SZREG($sp)
+	$REG_L	$s5,($FRAMESIZE-9)*$SZREG($sp)
+	$REG_L	$s4,($FRAMESIZE-10)*$SZREG($sp)
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	$s3,($FRAMESIZE-11)*$SZREG($sp)
+	$REG_L	$s2,($FRAMESIZE-12)*$SZREG($sp)
+	$REG_L	$s1,($FRAMESIZE-13)*$SZREG($sp)
+	$REG_L	$s0,($FRAMESIZE-14)*$SZREG($sp)
+	$REG_L	$gp,($FRAMESIZE-15)*$SZREG($sp)
+___
+$code.=<<___;
+	jr	$ra
+	$PTR_ADD $sp,$FRAMESIZE*$SZREG
+.end	sha1_block_data_order
+.rdata
+.asciiz	"SHA1 for MIPS, CRYPTOGAMS by <appro\@openssl.org>"
+___
+print $code;
+close STDOUT;
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha1-parisc.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha1-parisc.pl
new file mode 100644
index 0000000000..6e5a328a6f
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha1-parisc.pl
@@ -0,0 +1,260 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# SHA1 block procedure for PA-RISC.
+
+# June 2009.
+#
+# On PA-7100LC performance is >30% better than gcc 3.2 generated code
+# for aligned input and >50% better for unaligned. Compared to vendor
+# compiler on PA-8600 it's almost 60% faster in 64-bit build and just
+# few percent faster in 32-bit one (this for aligned input, data for
+# unaligned input is not available).
+#
+# Special thanks to polarhome.com for providing HP-UX account.
+
+$flavour = shift;
+$output = shift;
+open STDOUT,">$output";
+
+if ($flavour =~ /64/) {
+	$LEVEL		="2.0W";
+	$SIZE_T		=8;
+	$FRAME_MARKER	=80;
+	$SAVED_RP	=16;
+	$PUSH		="std";
+	$PUSHMA		="std,ma";
+	$POP		="ldd";
+	$POPMB		="ldd,mb";
+} else {
+	$LEVEL		="1.0";
+	$SIZE_T		=4;
+	$FRAME_MARKER	=48;
+	$SAVED_RP	=20;
+	$PUSH		="stw";
+	$PUSHMA		="stwm";
+	$POP		="ldw";
+	$POPMB		="ldwm";
+}
+
+$FRAME=14*$SIZE_T+$FRAME_MARKER;# 14 saved regs + frame marker
+				#                 [+ argument transfer]
+$ctx="%r26";		# arg0
+$inp="%r25";		# arg1
+$num="%r24";		# arg2
+
+$t0="%r28";
+$t1="%r29";
+$K="%r31";
+
+@X=("%r1", "%r2", "%r3", "%r4", "%r5", "%r6", "%r7", "%r8",
+    "%r9", "%r10","%r11","%r12","%r13","%r14","%r15","%r16",$t0);
+
+@V=($A,$B,$C,$D,$E)=("%r19","%r20","%r21","%r22","%r23");
+
+sub BODY_00_19 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+$code.=<<___ if ($i<15);
+	addl	$K,$e,$e	; $i
+	shd	$a,$a,27,$t1
+	addl	@X[$i],$e,$e
+	and	$c,$b,$t0
+	addl	$t1,$e,$e
+	andcm	$d,$b,$t1
+	shd	$b,$b,2,$b
+	or	$t1,$t0,$t0
+	addl	$t0,$e,$e
+___
+$code.=<<___ if ($i>=15);	# with forward Xupdate
+	addl	$K,$e,$e	; $i
+	shd	$a,$a,27,$t1
+	xor	@X[($j+2)%16],@X[$j%16],@X[$j%16]
+	addl	@X[$i%16],$e,$e
+	and	$c,$b,$t0
+	xor	@X[($j+8)%16],@X[$j%16],@X[$j%16]
+	addl	$t1,$e,$e
+	andcm	$d,$b,$t1
+	shd	$b,$b,2,$b
+	or	$t1,$t0,$t0
+	xor	@X[($j+13)%16],@X[$j%16],@X[$j%16]
+	add	$t0,$e,$e
+	shd	@X[$j%16],@X[$j%16],31,@X[$j%16]
+___
+}
+
+sub BODY_20_39 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+$code.=<<___ if ($i<79);
+	xor	@X[($j+2)%16],@X[$j%16],@X[$j%16]	; $i
+	addl	$K,$e,$e
+	shd	$a,$a,27,$t1
+	xor	@X[($j+8)%16],@X[$j%16],@X[$j%16]
+	addl	@X[$i%16],$e,$e
+	xor	$b,$c,$t0
+	xor	@X[($j+13)%16],@X[$j%16],@X[$j%16]
+	addl	$t1,$e,$e
+	shd	$b,$b,2,$b
+	xor	$d,$t0,$t0
+	shd	@X[$j%16],@X[$j%16],31,@X[$j%16]
+	addl	$t0,$e,$e
+___
+$code.=<<___ if ($i==79);	# with context load
+	ldw	0($ctx),@X[0]	; $i
+	addl	$K,$e,$e
+	shd	$a,$a,27,$t1
+	ldw	4($ctx),@X[1]
+	addl	@X[$i%16],$e,$e
+	xor	$b,$c,$t0
+	ldw	8($ctx),@X[2]
+	addl	$t1,$e,$e
+	shd	$b,$b,2,$b
+	xor	$d,$t0,$t0
+	ldw	12($ctx),@X[3]
+	addl	$t0,$e,$e
+	ldw	16($ctx),@X[4]
+___
+}
+
+sub BODY_40_59 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+$code.=<<___;
+	shd	$a,$a,27,$t1	; $i
+	addl	$K,$e,$e
+	xor	@X[($j+2)%16],@X[$j%16],@X[$j%16]
+	xor	$d,$c,$t0
+	addl	@X[$i%16],$e,$e
+	xor	@X[($j+8)%16],@X[$j%16],@X[$j%16]
+	and	$b,$t0,$t0
+	addl	$t1,$e,$e
+	shd	$b,$b,2,$b
+	xor	@X[($j+13)%16],@X[$j%16],@X[$j%16]
+	addl	$t0,$e,$e
+	and	$d,$c,$t1
+	shd	@X[$j%16],@X[$j%16],31,@X[$j%16]
+	addl	$t1,$e,$e
+___
+}
+
+$code=<<___;
+	.LEVEL	$LEVEL
+	.SPACE	\$TEXT\$
+	.SUBSPA	\$CODE\$,QUAD=0,ALIGN=8,ACCESS=0x2C,CODE_ONLY
+
+	.EXPORT	sha1_block_data_order,ENTRY,ARGW0=GR,ARGW1=GR,ARGW2=GR
+sha1_block_data_order
+	.PROC
+	.CALLINFO	FRAME=`$FRAME-14*$SIZE_T`,NO_CALLS,SAVE_RP,ENTRY_GR=16
+	.ENTRY
+	$PUSH	%r2,-$SAVED_RP(%sp)	; standard prologue
+	$PUSHMA	%r3,$FRAME(%sp)
+	$PUSH	%r4,`-$FRAME+1*$SIZE_T`(%sp)
+	$PUSH	%r5,`-$FRAME+2*$SIZE_T`(%sp)
+	$PUSH	%r6,`-$FRAME+3*$SIZE_T`(%sp)
+	$PUSH	%r7,`-$FRAME+4*$SIZE_T`(%sp)
+	$PUSH	%r8,`-$FRAME+5*$SIZE_T`(%sp)
+	$PUSH	%r9,`-$FRAME+6*$SIZE_T`(%sp)
+	$PUSH	%r10,`-$FRAME+7*$SIZE_T`(%sp)
+	$PUSH	%r11,`-$FRAME+8*$SIZE_T`(%sp)
+	$PUSH	%r12,`-$FRAME+9*$SIZE_T`(%sp)
+	$PUSH	%r13,`-$FRAME+10*$SIZE_T`(%sp)
+	$PUSH	%r14,`-$FRAME+11*$SIZE_T`(%sp)
+	$PUSH	%r15,`-$FRAME+12*$SIZE_T`(%sp)
+	$PUSH	%r16,`-$FRAME+13*$SIZE_T`(%sp)
+
+	ldw	0($ctx),$A
+	ldw	4($ctx),$B
+	ldw	8($ctx),$C
+	ldw	12($ctx),$D
+	ldw	16($ctx),$E
+
+	extru	$inp,31,2,$t0		; t0=inp&3;
+	sh3addl	$t0,%r0,$t0		; t0*=8;
+	subi	32,$t0,$t0		; t0=32-t0;
+	mtctl	$t0,%cr11		; %sar=t0;
+
+L\$oop
+	ldi	3,$t0
+	andcm	$inp,$t0,$t0		; 64-bit neutral
+___
+	for ($i=0;$i<15;$i++) {		# load input block
+	$code.="\tldw	`4*$i`($t0),@X[$i]\n";		}
+$code.=<<___;
+	cmpb,*=	$inp,$t0,L\$aligned
+	ldw	60($t0),@X[15]
+	ldw	64($t0),@X[16]
+___
+	for ($i=0;$i<16;$i++) {		# align input
+	$code.="\tvshd	@X[$i],@X[$i+1],@X[$i]\n";	}
+$code.=<<___;
+L\$aligned
+	ldil	L'0x5a827000,$K		; K_00_19
+	ldo	0x999($K),$K
+___
+for ($i=0;$i<20;$i++)   { &BODY_00_19($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	ldil	L'0x6ed9e000,$K		; K_20_39
+	ldo	0xba1($K),$K
+___
+
+for (;$i<40;$i++)       { &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	ldil	L'0x8f1bb000,$K		; K_40_59
+	ldo	0xcdc($K),$K
+___
+
+for (;$i<60;$i++)       { &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	ldil	L'0xca62c000,$K		; K_60_79
+	ldo	0x1d6($K),$K
+___
+for (;$i<80;$i++)       { &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+
+$code.=<<___;
+	addl	@X[0],$A,$A
+	addl	@X[1],$B,$B
+	addl	@X[2],$C,$C
+	addl	@X[3],$D,$D
+	addl	@X[4],$E,$E
+	stw	$A,0($ctx)
+	stw	$B,4($ctx)
+	stw	$C,8($ctx)
+	stw	$D,12($ctx)
+	stw	$E,16($ctx)
+	addib,*<> -1,$num,L\$oop
+	ldo	64($inp),$inp
+
+	$POP	`-$FRAME-$SAVED_RP`(%sp),%r2	; standard epilogue
+	$POP	`-$FRAME+1*$SIZE_T`(%sp),%r4
+	$POP	`-$FRAME+2*$SIZE_T`(%sp),%r5
+	$POP	`-$FRAME+3*$SIZE_T`(%sp),%r6
+	$POP	`-$FRAME+4*$SIZE_T`(%sp),%r7
+	$POP	`-$FRAME+5*$SIZE_T`(%sp),%r8
+	$POP	`-$FRAME+6*$SIZE_T`(%sp),%r9
+	$POP	`-$FRAME+7*$SIZE_T`(%sp),%r10
+	$POP	`-$FRAME+8*$SIZE_T`(%sp),%r11
+	$POP	`-$FRAME+9*$SIZE_T`(%sp),%r12
+	$POP	`-$FRAME+10*$SIZE_T`(%sp),%r13
+	$POP	`-$FRAME+11*$SIZE_T`(%sp),%r14
+	$POP	`-$FRAME+12*$SIZE_T`(%sp),%r15
+	$POP	`-$FRAME+13*$SIZE_T`(%sp),%r16
+	bv	(%r2)
+	.EXIT
+	$POPMB	-$FRAME(%sp),%r3
+	.PROCEND
+	.STRINGZ "SHA1 block transform for PA-RISC, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+$code =~ s/,\*/,/gm		if ($SIZE_T==4);
+$code =~ s/\bbv\b/bve/gm	if ($SIZE_T==8);
+print $code;
+close STDOUT;
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha1-ppc.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha1-ppc.pl
new file mode 100755
index 0000000000..2140dd2f8d
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha1-ppc.pl
@@ -0,0 +1,326 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# I let hardware handle unaligned input(*), except on page boundaries
+# (see below for details). Otherwise straightforward implementation
+# with X vector in register bank. The module is big-endian [which is
+# not big deal as there're no little-endian targets left around].
+#
+# (*) this means that this module is inappropriate for PPC403? Does
+#     anybody know if pre-POWER3 can sustain unaligned load?
+
+# 			-m64	-m32
+# ----------------------------------
+# PPC970,gcc-4.0.0	+76%	+59%
+# Power6,xlc-7		+68%	+33%
+
+$flavour = shift;
+
+if ($flavour =~ /64/) {
+	$SIZE_T	=8;
+	$LRSAVE	=2*$SIZE_T;
+	$UCMP	="cmpld";
+	$STU	="stdu";
+	$POP	="ld";
+	$PUSH	="std";
+} elsif ($flavour =~ /32/) {
+	$SIZE_T	=4;
+	$LRSAVE	=$SIZE_T;
+	$UCMP	="cmplw";
+	$STU	="stwu";
+	$POP	="lwz";
+	$PUSH	="stw";
+} else { die "nonsense $flavour"; }
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}ppc-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/ppc-xlate.pl" and -f $xlate) or
+die "can't locate ppc-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour ".shift || die "can't call $xlate: $!";
+
+$FRAME=24*$SIZE_T+64;
+$LOCALS=6*$SIZE_T;
+
+$K  ="r0";
+$sp ="r1";
+$toc="r2";
+$ctx="r3";
+$inp="r4";
+$num="r5";
+$t0 ="r15";
+$t1 ="r6";
+
+$A  ="r7";
+$B  ="r8";
+$C  ="r9";
+$D  ="r10";
+$E  ="r11";
+$T  ="r12";
+
+@V=($A,$B,$C,$D,$E,$T);
+@X=("r16","r17","r18","r19","r20","r21","r22","r23",
+    "r24","r25","r26","r27","r28","r29","r30","r31");
+
+sub BODY_00_19 {
+my ($i,$a,$b,$c,$d,$e,$f)=@_;
+my $j=$i+1;
+$code.=<<___ if ($i==0);
+	lwz	@X[$i],`$i*4`($inp)
+___
+$code.=<<___ if ($i<15);
+	lwz	@X[$j],`$j*4`($inp)
+	add	$f,$K,$e
+	rotlwi	$e,$a,5
+	add	$f,$f,@X[$i]
+	and	$t0,$c,$b
+	add	$f,$f,$e
+	andc	$t1,$d,$b
+	rotlwi	$b,$b,30
+	or	$t0,$t0,$t1
+	add	$f,$f,$t0
+___
+$code.=<<___ if ($i>=15);
+	add	$f,$K,$e
+	rotlwi	$e,$a,5
+	xor	@X[$j%16],@X[$j%16],@X[($j+2)%16]
+	add	$f,$f,@X[$i%16]
+	and	$t0,$c,$b
+	xor	@X[$j%16],@X[$j%16],@X[($j+8)%16]
+	add	$f,$f,$e
+	andc	$t1,$d,$b
+	rotlwi	$b,$b,30
+	or	$t0,$t0,$t1
+	xor	@X[$j%16],@X[$j%16],@X[($j+13)%16]
+	add	$f,$f,$t0
+	rotlwi	@X[$j%16],@X[$j%16],1
+___
+}
+
+sub BODY_20_39 {
+my ($i,$a,$b,$c,$d,$e,$f)=@_;
+my $j=$i+1;
+$code.=<<___ if ($i<79);
+	add	$f,$K,$e
+	rotlwi	$e,$a,5
+	xor	@X[$j%16],@X[$j%16],@X[($j+2)%16]
+	add	$f,$f,@X[$i%16]
+	xor	$t0,$b,$c
+	xor	@X[$j%16],@X[$j%16],@X[($j+8)%16]
+	add	$f,$f,$e
+	rotlwi	$b,$b,30
+	xor	$t0,$t0,$d
+	xor	@X[$j%16],@X[$j%16],@X[($j+13)%16]
+	add	$f,$f,$t0
+	rotlwi	@X[$j%16],@X[$j%16],1
+___
+$code.=<<___ if ($i==79);
+	add	$f,$K,$e
+	rotlwi	$e,$a,5
+	lwz	r16,0($ctx)
+	add	$f,$f,@X[$i%16]
+	xor	$t0,$b,$c
+	lwz	r17,4($ctx)
+	add	$f,$f,$e
+	rotlwi	$b,$b,30
+	lwz	r18,8($ctx)
+	xor	$t0,$t0,$d
+	lwz	r19,12($ctx)
+	add	$f,$f,$t0
+	lwz	r20,16($ctx)
+___
+}
+
+sub BODY_40_59 {
+my ($i,$a,$b,$c,$d,$e,$f)=@_;
+my $j=$i+1;
+$code.=<<___;
+	add	$f,$K,$e
+	rotlwi	$e,$a,5
+	xor	@X[$j%16],@X[$j%16],@X[($j+2)%16]
+	add	$f,$f,@X[$i%16]
+	and	$t0,$b,$c
+	xor	@X[$j%16],@X[$j%16],@X[($j+8)%16]
+	add	$f,$f,$e
+	or	$t1,$b,$c
+	rotlwi	$b,$b,30
+	xor	@X[$j%16],@X[$j%16],@X[($j+13)%16]
+	and	$t1,$t1,$d
+	or	$t0,$t0,$t1
+	rotlwi	@X[$j%16],@X[$j%16],1
+	add	$f,$f,$t0
+___
+}
+
+$code=<<___;
+.machine	"any"
+.text
+
+.globl	.sha1_block_data_order
+.align	4
+.sha1_block_data_order:
+	$STU	$sp,-$FRAME($sp)
+	mflr	r0
+	$PUSH	r15,`$FRAME-$SIZE_T*17`($sp)
+	$PUSH	r16,`$FRAME-$SIZE_T*16`($sp)
+	$PUSH	r17,`$FRAME-$SIZE_T*15`($sp)
+	$PUSH	r18,`$FRAME-$SIZE_T*14`($sp)
+	$PUSH	r19,`$FRAME-$SIZE_T*13`($sp)
+	$PUSH	r20,`$FRAME-$SIZE_T*12`($sp)
+	$PUSH	r21,`$FRAME-$SIZE_T*11`($sp)
+	$PUSH	r22,`$FRAME-$SIZE_T*10`($sp)
+	$PUSH	r23,`$FRAME-$SIZE_T*9`($sp)
+	$PUSH	r24,`$FRAME-$SIZE_T*8`($sp)
+	$PUSH	r25,`$FRAME-$SIZE_T*7`($sp)
+	$PUSH	r26,`$FRAME-$SIZE_T*6`($sp)
+	$PUSH	r27,`$FRAME-$SIZE_T*5`($sp)
+	$PUSH	r28,`$FRAME-$SIZE_T*4`($sp)
+	$PUSH	r29,`$FRAME-$SIZE_T*3`($sp)
+	$PUSH	r30,`$FRAME-$SIZE_T*2`($sp)
+	$PUSH	r31,`$FRAME-$SIZE_T*1`($sp)
+	$PUSH	r0,`$FRAME+$LRSAVE`($sp)
+	lwz	$A,0($ctx)
+	lwz	$B,4($ctx)
+	lwz	$C,8($ctx)
+	lwz	$D,12($ctx)
+	lwz	$E,16($ctx)
+	andi.	r0,$inp,3
+	bne	Lunaligned
+Laligned:
+	mtctr	$num
+	bl	Lsha1_block_private
+	b	Ldone
+
+; PowerPC specification allows an implementation to be ill-behaved
+; upon unaligned access which crosses page boundary. "Better safe
+; than sorry" principle makes me treat it specially. But I don't
+; look for particular offending word, but rather for 64-byte input
+; block which crosses the boundary. Once found that block is aligned
+; and hashed separately...
+.align	4
+Lunaligned:
+	subfic	$t1,$inp,4096
+	andi.	$t1,$t1,4095	; distance to closest page boundary
+	srwi.	$t1,$t1,6	; t1/=64
+	beq	Lcross_page
+	$UCMP	$num,$t1
+	ble-	Laligned	; didn't cross the page boundary
+	mtctr	$t1
+	subfc	$num,$t1,$num
+	bl	Lsha1_block_private
+Lcross_page:
+	li	$t1,16
+	mtctr	$t1
+	addi	r20,$sp,$LOCALS	; spot within the frame
+Lmemcpy:
+	lbz	r16,0($inp)
+	lbz	r17,1($inp)
+	lbz	r18,2($inp)
+	lbz	r19,3($inp)
+	addi	$inp,$inp,4
+	stb	r16,0(r20)
+	stb	r17,1(r20)
+	stb	r18,2(r20)
+	stb	r19,3(r20)
+	addi	r20,r20,4
+	bdnz	Lmemcpy
+
+	$PUSH	$inp,`$FRAME-$SIZE_T*18`($sp)
+	li	$t1,1
+	addi	$inp,$sp,$LOCALS
+	mtctr	$t1
+	bl	Lsha1_block_private
+	$POP	$inp,`$FRAME-$SIZE_T*18`($sp)
+	addic.	$num,$num,-1
+	bne-	Lunaligned
+
+Ldone:
+	$POP	r0,`$FRAME+$LRSAVE`($sp)
+	$POP	r15,`$FRAME-$SIZE_T*17`($sp)
+	$POP	r16,`$FRAME-$SIZE_T*16`($sp)
+	$POP	r17,`$FRAME-$SIZE_T*15`($sp)
+	$POP	r18,`$FRAME-$SIZE_T*14`($sp)
+	$POP	r19,`$FRAME-$SIZE_T*13`($sp)
+	$POP	r20,`$FRAME-$SIZE_T*12`($sp)
+	$POP	r21,`$FRAME-$SIZE_T*11`($sp)
+	$POP	r22,`$FRAME-$SIZE_T*10`($sp)
+	$POP	r23,`$FRAME-$SIZE_T*9`($sp)
+	$POP	r24,`$FRAME-$SIZE_T*8`($sp)
+	$POP	r25,`$FRAME-$SIZE_T*7`($sp)
+	$POP	r26,`$FRAME-$SIZE_T*6`($sp)
+	$POP	r27,`$FRAME-$SIZE_T*5`($sp)
+	$POP	r28,`$FRAME-$SIZE_T*4`($sp)
+	$POP	r29,`$FRAME-$SIZE_T*3`($sp)
+	$POP	r30,`$FRAME-$SIZE_T*2`($sp)
+	$POP	r31,`$FRAME-$SIZE_T*1`($sp)
+	mtlr	r0
+	addi	$sp,$sp,$FRAME
+	blr
+	.long	0
+	.byte	0,12,4,1,0x80,18,3,0
+	.long	0
+___
+
+# This is private block function, which uses tailored calling
+# interface, namely upon entry SHA_CTX is pre-loaded to given
+# registers and counter register contains amount of chunks to
+# digest...
+$code.=<<___;
+.align	4
+Lsha1_block_private:
+___
+$code.=<<___;	# load K_00_19
+	lis	$K,0x5a82
+	ori	$K,$K,0x7999
+___
+for($i=0;$i<20;$i++)	{ &BODY_00_19($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;	# load K_20_39
+	lis	$K,0x6ed9
+	ori	$K,$K,0xeba1
+___
+for(;$i<40;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;	# load K_40_59
+	lis	$K,0x8f1b
+	ori	$K,$K,0xbcdc
+___
+for(;$i<60;$i++)	{ &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;	# load K_60_79
+	lis	$K,0xca62
+	ori	$K,$K,0xc1d6
+___
+for(;$i<80;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	add	r16,r16,$E
+	add	r17,r17,$T
+	add	r18,r18,$A
+	add	r19,r19,$B
+	add	r20,r20,$C
+	stw	r16,0($ctx)
+	mr	$A,r16
+	stw	r17,4($ctx)
+	mr	$B,r17
+	stw	r18,8($ctx)
+	mr	$C,r18
+	stw	r19,12($ctx)
+	mr	$D,r19
+	stw	r20,16($ctx)
+	mr	$E,r20
+	addi	$inp,$inp,`16*4`
+	bdnz-	Lsha1_block_private
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+___
+$code.=<<___;
+.asciz	"SHA1 block transform for PPC, CRYPTOGAMS by <appro\@fy.chalmers.se>"
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha1-s390x.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha1-s390x.pl
new file mode 100644
index 0000000000..9193dda45e
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha1-s390x.pl
@@ -0,0 +1,246 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# SHA1 block procedure for s390x.
+
+# April 2007.
+#
+# Performance is >30% better than gcc 3.3 generated code. But the real
+# twist is that SHA1 hardware support is detected and utilized. In
+# which case performance can reach further >4.5x for larger chunks.
+
+# January 2009.
+#
+# Optimize Xupdate for amount of memory references and reschedule
+# instructions to favour dual-issue z10 pipeline. On z10 hardware is
+# "only" ~2.3x faster than software.
+
+# November 2010.
+#
+# Adapt for -m31 build. If kernel supports what's called "highgprs"
+# feature on Linux [see /proc/cpuinfo], it's possible to use 64-bit
+# instructions and achieve "64-bit" performance even in 31-bit legacy
+# application context. The feature is not specific to any particular
+# processor, as long as it's "z-CPU". Latter implies that the code
+# remains z/Architecture specific.
+
+$kimdfunc=1;	# magic function code for kimd instruction
+
+$flavour = shift;
+
+if ($flavour =~ /3[12]/) {
+	$SIZE_T=4;
+	$g="";
+} else {
+	$SIZE_T=8;
+	$g="g";
+}
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+$K_00_39="%r0"; $K=$K_00_39;
+$K_40_79="%r1";
+$ctx="%r2";	$prefetch="%r2";
+$inp="%r3";
+$len="%r4";
+
+$A="%r5";
+$B="%r6";
+$C="%r7";
+$D="%r8";
+$E="%r9";	@V=($A,$B,$C,$D,$E);
+$t0="%r10";
+$t1="%r11";
+@X=("%r12","%r13","%r14");
+$sp="%r15";
+
+$stdframe=16*$SIZE_T+4*8;
+$frame=$stdframe+16*4;
+
+sub Xupdate {
+my $i=shift;
+
+$code.=<<___ if ($i==15);
+	lg	$prefetch,$stdframe($sp)	### Xupdate(16) warm-up
+	lr	$X[0],$X[2]
+___
+return if ($i&1);	# Xupdate is vectorized and executed every 2nd cycle
+$code.=<<___ if ($i<16);
+	lg	$X[0],`$i*4`($inp)	### Xload($i)
+	rllg	$X[1],$X[0],32
+___
+$code.=<<___ if ($i>=16);
+	xgr	$X[0],$prefetch		### Xupdate($i)
+	lg	$prefetch,`$stdframe+4*(($i+2)%16)`($sp)
+	xg	$X[0],`$stdframe+4*(($i+8)%16)`($sp)
+	xgr	$X[0],$prefetch
+	rll	$X[0],$X[0],1
+	rllg	$X[1],$X[0],32
+	rll	$X[1],$X[1],1
+	rllg	$X[0],$X[1],32
+	lr	$X[2],$X[1]		# feedback
+___
+$code.=<<___ if ($i<=70);
+	stg	$X[0],`$stdframe+4*($i%16)`($sp)
+___
+unshift(@X,pop(@X));
+}
+
+sub BODY_00_19 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $xi=$X[1];
+
+	&Xupdate($i);
+$code.=<<___;
+	alr	$e,$K		### $i
+	rll	$t1,$a,5
+	lr	$t0,$d
+	xr	$t0,$c
+	alr	$e,$t1
+	nr	$t0,$b
+	alr	$e,$xi
+	xr	$t0,$d
+	rll	$b,$b,30
+	alr	$e,$t0
+___
+}
+
+sub BODY_20_39 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $xi=$X[1];
+
+	&Xupdate($i);
+$code.=<<___;
+	alr	$e,$K		### $i
+	rll	$t1,$a,5
+	lr	$t0,$b
+	alr	$e,$t1
+	xr	$t0,$c
+	alr	$e,$xi
+	xr	$t0,$d
+	rll	$b,$b,30
+	alr	$e,$t0
+___
+}
+
+sub BODY_40_59 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $xi=$X[1];
+
+	&Xupdate($i);
+$code.=<<___;
+	alr	$e,$K		### $i
+	rll	$t1,$a,5
+	lr	$t0,$b
+	alr	$e,$t1
+	or	$t0,$c
+	lr	$t1,$b
+	nr	$t0,$d
+	nr	$t1,$c
+	alr	$e,$xi
+	or	$t0,$t1
+	rll	$b,$b,30
+	alr	$e,$t0
+___
+}
+
+$code.=<<___;
+.text
+.align	64
+.type	Ktable,\@object
+Ktable: .long	0x5a827999,0x6ed9eba1,0x8f1bbcdc,0xca62c1d6
+	.skip	48	#.long	0,0,0,0,0,0,0,0,0,0,0,0
+.size	Ktable,.-Ktable
+.globl	sha1_block_data_order
+.type	sha1_block_data_order,\@function
+sha1_block_data_order:
+___
+$code.=<<___ if ($kimdfunc);
+	larl	%r1,OPENSSL_s390xcap_P
+	lg	%r0,0(%r1)
+	tmhl	%r0,0x4000	# check for message-security assist
+	jz	.Lsoftware
+	lghi	%r0,0
+	la	%r1,`2*$SIZE_T`($sp)
+	.long	0xb93e0002	# kimd %r0,%r2
+	lg	%r0,`2*$SIZE_T`($sp)
+	tmhh	%r0,`0x8000>>$kimdfunc`
+	jz	.Lsoftware
+	lghi	%r0,$kimdfunc
+	lgr	%r1,$ctx
+	lgr	%r2,$inp
+	sllg	%r3,$len,6
+	.long	0xb93e0002	# kimd %r0,%r2
+	brc	1,.-4		# pay attention to "partial completion"
+	br	%r14
+.align	16
+.Lsoftware:
+___
+$code.=<<___;
+	lghi	%r1,-$frame
+	st${g}	$ctx,`2*$SIZE_T`($sp)
+	stm${g}	%r6,%r15,`6*$SIZE_T`($sp)
+	lgr	%r0,$sp
+	la	$sp,0(%r1,$sp)
+	st${g}	%r0,0($sp)
+
+	larl	$t0,Ktable
+	llgf	$A,0($ctx)
+	llgf	$B,4($ctx)
+	llgf	$C,8($ctx)
+	llgf	$D,12($ctx)
+	llgf	$E,16($ctx)
+
+	lg	$K_00_39,0($t0)
+	lg	$K_40_79,8($t0)
+
+.Lloop:
+	rllg	$K_00_39,$K_00_39,32
+___
+for ($i=0;$i<20;$i++)	{ &BODY_00_19($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	rllg	$K_00_39,$K_00_39,32
+___
+for (;$i<40;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;	$K=$K_40_79;
+	rllg	$K_40_79,$K_40_79,32
+___
+for (;$i<60;$i++)	{ &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	rllg	$K_40_79,$K_40_79,32
+___
+for (;$i<80;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+
+	l${g}	$ctx,`$frame+2*$SIZE_T`($sp)
+	la	$inp,64($inp)
+	al	$A,0($ctx)
+	al	$B,4($ctx)
+	al	$C,8($ctx)
+	al	$D,12($ctx)
+	al	$E,16($ctx)
+	st	$A,0($ctx)
+	st	$B,4($ctx)
+	st	$C,8($ctx)
+	st	$D,12($ctx)
+	st	$E,16($ctx)
+	brct${g} $len,.Lloop
+
+	lm${g}	%r6,%r15,`$frame+6*$SIZE_T`($sp)
+	br	%r14
+.size	sha1_block_data_order,.-sha1_block_data_order
+.string	"SHA1 block transform for s390x, CRYPTOGAMS by <appro\@openssl.org>"
+.comm	OPENSSL_s390xcap_P,16,8
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+
+print $code;
+close STDOUT;
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha1-sparcv9.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha1-sparcv9.pl
new file mode 100644
index 0000000000..5c161cecd6
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha1-sparcv9.pl
@@ -0,0 +1,284 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# Performance improvement is not really impressive on pre-T1 CPU: +8%
+# over Sun C and +25% over gcc [3.3]. While on T1, a.k.a. Niagara, it
+# turned to be 40% faster than 64-bit code generated by Sun C 5.8 and
+# >2x than 64-bit code generated by gcc 3.4. And there is a gimmick.
+# X[16] vector is packed to 8 64-bit registers and as result nothing
+# is spilled on stack. In addition input data is loaded in compact
+# instruction sequence, thus minimizing the window when the code is
+# subject to [inter-thread] cache-thrashing hazard. The goal is to
+# ensure scalability on UltraSPARC T1, or rather to avoid decay when
+# amount of active threads exceeds the number of physical cores.
+
+$bits=32;
+for (@ARGV)	{ $bits=64 if (/\-m64/ || /\-xarch\=v9/); }
+if ($bits==64)	{ $bias=2047; $frame=192; }
+else		{ $bias=0;    $frame=112; }
+
+$output=shift;
+open STDOUT,">$output";
+
+@X=("%o0","%o1","%o2","%o3","%o4","%o5","%g1","%o7");
+$rot1m="%g2";
+$tmp64="%g3";
+$Xi="%g4";
+$A="%l0";
+$B="%l1";
+$C="%l2";
+$D="%l3";
+$E="%l4";
+@V=($A,$B,$C,$D,$E);
+$K_00_19="%l5";
+$K_20_39="%l6";
+$K_40_59="%l7";
+$K_60_79="%g5";
+@K=($K_00_19,$K_20_39,$K_40_59,$K_60_79);
+
+$ctx="%i0";
+$inp="%i1";
+$len="%i2";
+$tmp0="%i3";
+$tmp1="%i4";
+$tmp2="%i5";
+
+sub BODY_00_15 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $xi=($i&1)?@X[($i/2)%8]:$Xi;
+
+$code.=<<___;
+	sll	$a,5,$tmp0		!! $i
+	add	@K[$i/20],$e,$e
+	srl	$a,27,$tmp1
+	add	$tmp0,$e,$e
+	and	$c,$b,$tmp0
+	add	$tmp1,$e,$e
+	sll	$b,30,$tmp2
+	andn	$d,$b,$tmp1
+	srl	$b,2,$b
+	or	$tmp1,$tmp0,$tmp1
+	or	$tmp2,$b,$b
+	add	$xi,$e,$e
+___
+if ($i&1 && $i<15) {
+	$code.=
+	"	srlx	@X[(($i+1)/2)%8],32,$Xi\n";
+}
+$code.=<<___;
+	add	$tmp1,$e,$e
+___
+}
+
+sub Xupdate {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i/2;
+
+if ($i&1) {
+$code.=<<___;
+	sll	$a,5,$tmp0		!! $i
+	add	@K[$i/20],$e,$e
+	srl	$a,27,$tmp1
+___
+} else {
+$code.=<<___;
+	sllx	@X[($j+6)%8],32,$Xi	! Xupdate($i)
+	xor	@X[($j+1)%8],@X[$j%8],@X[$j%8]
+	srlx	@X[($j+7)%8],32,$tmp1
+	xor	@X[($j+4)%8],@X[$j%8],@X[$j%8]
+	sll	$a,5,$tmp0		!! $i
+	or	$tmp1,$Xi,$Xi
+	add	@K[$i/20],$e,$e		!!
+	xor	$Xi,@X[$j%8],@X[$j%8]
+	srlx	@X[$j%8],31,$Xi
+	add	@X[$j%8],@X[$j%8],@X[$j%8]
+	and	$Xi,$rot1m,$Xi
+	andn	@X[$j%8],$rot1m,@X[$j%8]
+	srl	$a,27,$tmp1		!!
+	or	$Xi,@X[$j%8],@X[$j%8]
+___
+}
+}
+
+sub BODY_16_19 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+
+	&Xupdate(@_);
+    if ($i&1) {
+	$xi=@X[($i/2)%8];
+    } else {
+	$xi=$Xi;
+	$code.="\tsrlx	@X[($i/2)%8],32,$xi\n";
+    }
+$code.=<<___;
+	add	$tmp0,$e,$e		!!
+	and	$c,$b,$tmp0
+	add	$tmp1,$e,$e
+	sll	$b,30,$tmp2
+	add	$xi,$e,$e
+	andn	$d,$b,$tmp1
+	srl	$b,2,$b
+	or	$tmp1,$tmp0,$tmp1
+	or	$tmp2,$b,$b
+	add	$tmp1,$e,$e
+___
+}
+
+sub BODY_20_39 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $xi;
+	&Xupdate(@_);
+    if ($i&1) {
+	$xi=@X[($i/2)%8];
+    } else {
+	$xi=$Xi;
+	$code.="\tsrlx	@X[($i/2)%8],32,$xi\n";
+    }
+$code.=<<___;
+	add	$tmp0,$e,$e		!!
+	xor	$c,$b,$tmp0
+	add	$tmp1,$e,$e
+	sll	$b,30,$tmp2
+	xor	$d,$tmp0,$tmp1
+	srl	$b,2,$b
+	add	$tmp1,$e,$e
+	or	$tmp2,$b,$b
+	add	$xi,$e,$e
+___
+}
+
+sub BODY_40_59 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $xi;
+	&Xupdate(@_);
+    if ($i&1) {
+	$xi=@X[($i/2)%8];
+    } else {
+	$xi=$Xi;
+	$code.="\tsrlx	@X[($i/2)%8],32,$xi\n";
+    }
+$code.=<<___;
+	add	$tmp0,$e,$e		!!
+	and	$c,$b,$tmp0
+	add	$tmp1,$e,$e
+	sll	$b,30,$tmp2
+	or	$c,$b,$tmp1
+	srl	$b,2,$b
+	and	$d,$tmp1,$tmp1
+	add	$xi,$e,$e
+	or	$tmp1,$tmp0,$tmp1
+	or	$tmp2,$b,$b
+	add	$tmp1,$e,$e
+___
+}
+
+$code.=<<___ if ($bits==64);
+.register	%g2,#scratch
+.register	%g3,#scratch
+___
+$code.=<<___;
+.section	".text",#alloc,#execinstr
+
+.align	32
+.globl	sha1_block_data_order
+sha1_block_data_order:
+	save	%sp,-$frame,%sp
+	sllx	$len,6,$len
+	add	$inp,$len,$len
+
+	or	%g0,1,$rot1m
+	sllx	$rot1m,32,$rot1m
+	or	$rot1m,1,$rot1m
+
+	ld	[$ctx+0],$A
+	ld	[$ctx+4],$B
+	ld	[$ctx+8],$C
+	ld	[$ctx+12],$D
+	ld	[$ctx+16],$E
+	andn	$inp,7,$tmp0
+
+	sethi	%hi(0x5a827999),$K_00_19
+	or	$K_00_19,%lo(0x5a827999),$K_00_19
+	sethi	%hi(0x6ed9eba1),$K_20_39
+	or	$K_20_39,%lo(0x6ed9eba1),$K_20_39
+	sethi	%hi(0x8f1bbcdc),$K_40_59
+	or	$K_40_59,%lo(0x8f1bbcdc),$K_40_59
+	sethi	%hi(0xca62c1d6),$K_60_79
+	or	$K_60_79,%lo(0xca62c1d6),$K_60_79
+
+.Lloop:
+	ldx	[$tmp0+0],@X[0]
+	ldx	[$tmp0+16],@X[2]
+	ldx	[$tmp0+32],@X[4]
+	ldx	[$tmp0+48],@X[6]
+	and	$inp,7,$tmp1
+	ldx	[$tmp0+8],@X[1]
+	sll	$tmp1,3,$tmp1
+	ldx	[$tmp0+24],@X[3]
+	subcc	%g0,$tmp1,$tmp2	! should be 64-$tmp1, but -$tmp1 works too
+	ldx	[$tmp0+40],@X[5]
+	bz,pt	%icc,.Laligned
+	ldx	[$tmp0+56],@X[7]
+
+	sllx	@X[0],$tmp1,@X[0]
+	ldx	[$tmp0+64],$tmp64
+___
+for($i=0;$i<7;$i++)
+{   $code.=<<___;
+	srlx	@X[$i+1],$tmp2,$Xi
+	sllx	@X[$i+1],$tmp1,@X[$i+1]
+	or	$Xi,@X[$i],@X[$i]
+___
+}
+$code.=<<___;
+	srlx	$tmp64,$tmp2,$tmp64
+	or	$tmp64,@X[7],@X[7]
+.Laligned:
+	srlx	@X[0],32,$Xi
+___
+for ($i=0;$i<16;$i++)	{ &BODY_00_15($i,@V); unshift(@V,pop(@V)); }
+for (;$i<20;$i++)	{ &BODY_16_19($i,@V); unshift(@V,pop(@V)); }
+for (;$i<40;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+for (;$i<60;$i++)	{ &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
+for (;$i<80;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+
+	ld	[$ctx+0],@X[0]
+	ld	[$ctx+4],@X[1]
+	ld	[$ctx+8],@X[2]
+	ld	[$ctx+12],@X[3]
+	add	$inp,64,$inp
+	ld	[$ctx+16],@X[4]
+	cmp	$inp,$len
+
+	add	$A,@X[0],$A
+	st	$A,[$ctx+0]
+	add	$B,@X[1],$B
+	st	$B,[$ctx+4]
+	add	$C,@X[2],$C
+	st	$C,[$ctx+8]
+	add	$D,@X[3],$D
+	st	$D,[$ctx+12]
+	add	$E,@X[4],$E
+	st	$E,[$ctx+16]
+
+	bne	`$bits==64?"%xcc":"%icc"`,.Lloop
+	andn	$inp,7,$tmp0
+
+	ret
+	restore
+.type	sha1_block_data_order,#function
+.size	sha1_block_data_order,(.-sha1_block_data_order)
+.asciz	"SHA1 block transform for SPARCv9, CRYPTOGAMS by <appro\@openssl.org>"
+.align	4
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha1-sparcv9a.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha1-sparcv9a.pl
new file mode 100644
index 0000000000..e65291bbd9
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha1-sparcv9a.pl
@@ -0,0 +1,601 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> 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/.
+# ====================================================================
+
+# January 2009
+#
+# Provided that UltraSPARC VIS instructions are pipe-lined(*) and
+# pairable(*) with IALU ones, offloading of Xupdate to the UltraSPARC
+# Graphic Unit would make it possible to achieve higher instruction-
+# level parallelism, ILP, and thus higher performance. It should be
+# explicitly noted that ILP is the keyword, and it means that this
+# code would be unsuitable for cores like UltraSPARC-Tx. The idea is
+# not really novel, Sun had VIS-powered implementation for a while.
+# Unlike Sun's implementation this one can process multiple unaligned
+# input blocks, and as such works as drop-in replacement for OpenSSL
+# sha1_block_data_order. Performance improvement was measured to be
+# 40% over pure IALU sha1-sparcv9.pl on UltraSPARC-IIi, but 12% on
+# UltraSPARC-III. See below for discussion...
+#
+# The module does not present direct interest for OpenSSL, because
+# it doesn't provide better performance on contemporary SPARCv9 CPUs,
+# UltraSPARC-Tx and SPARC64-V[II] to be specific. Those who feel they
+# absolutely must score on UltraSPARC-I-IV can simply replace
+# crypto/sha/asm/sha1-sparcv9.pl with this module.
+#
+# (*)	"Pipe-lined" means that even if it takes several cycles to
+#	complete, next instruction using same functional unit [but not
+#	depending on the result of the current instruction] can start
+#	execution without having to wait for the unit. "Pairable"
+#	means that two [or more] independent instructions can be
+#	issued at the very same time.
+
+$bits=32;
+for (@ARGV)	{ $bits=64 if (/\-m64/ || /\-xarch\=v9/); }
+if ($bits==64)	{ $bias=2047; $frame=192; }
+else		{ $bias=0;    $frame=112; }
+
+$output=shift;
+open STDOUT,">$output";
+
+$ctx="%i0";
+$inp="%i1";
+$len="%i2";
+$tmp0="%i3";
+$tmp1="%i4";
+$tmp2="%i5";
+$tmp3="%g5";
+
+$base="%g1";
+$align="%g4";
+$Xfer="%o5";
+$nXfer=$tmp3;
+$Xi="%o7";
+
+$A="%l0";
+$B="%l1";
+$C="%l2";
+$D="%l3";
+$E="%l4";
+@V=($A,$B,$C,$D,$E);
+
+$Actx="%o0";
+$Bctx="%o1";
+$Cctx="%o2";
+$Dctx="%o3";
+$Ectx="%o4";
+
+$fmul="%f32";
+$VK_00_19="%f34";
+$VK_20_39="%f36";
+$VK_40_59="%f38";
+$VK_60_79="%f40";
+@VK=($VK_00_19,$VK_20_39,$VK_40_59,$VK_60_79);
+@X=("%f0", "%f1", "%f2", "%f3", "%f4", "%f5", "%f6", "%f7",
+    "%f8", "%f9","%f10","%f11","%f12","%f13","%f14","%f15","%f16");
+
+# This is reference 2x-parallelized VIS-powered Xupdate procedure. It
+# covers even K_NN_MM addition...
+sub Xupdate {
+my ($i)=@_;
+my $K=@VK[($i+16)/20];
+my $j=($i+16)%16;
+
+#	[ provided that GSR.alignaddr_offset is 5, $mul contains
+#	  0x100ULL<<32|0x100 value and K_NN_MM are pre-loaded to
+#	  chosen registers... ]
+$code.=<<___;
+	fxors		@X[($j+13)%16],@X[$j],@X[$j]	!-1/-1/-1:X[0]^=X[13]
+	fxors		@X[($j+14)%16],@X[$j+1],@X[$j+1]! 0/ 0/ 0:X[1]^=X[14]
+	fxor		@X[($j+2)%16],@X[($j+8)%16],%f18! 1/ 1/ 1:Tmp=X[2,3]^X[8,9]
+	fxor		%f18,@X[$j],@X[$j]		! 2/ 4/ 3:X[0,1]^=X[2,3]^X[8,9]
+	faligndata	@X[$j],@X[$j],%f18		! 3/ 7/ 5:Tmp=X[0,1]>>>24
+	fpadd32		@X[$j],@X[$j],@X[$j]		! 4/ 8/ 6:X[0,1]<<=1
+	fmul8ulx16	%f18,$fmul,%f18			! 5/10/ 7:Tmp>>=7, Tmp&=1
+	![fxors		%f15,%f2,%f2]
+	for		%f18,@X[$j],@X[$j]		! 8/14/10:X[0,1]|=Tmp
+	![fxors		%f0,%f3,%f3]			!10/17/12:X[0] dependency
+	fpadd32		$K,@X[$j],%f20
+	std		%f20,[$Xfer+`4*$j`]
+___
+# The numbers delimited with slash are the earliest possible dispatch
+# cycles for given instruction assuming 1 cycle latency for simple VIS
+# instructions, such as on UltraSPARC-I&II, 3 cycles latency, such as
+# on UltraSPARC-III&IV, and 2 cycles latency(*), respectively. Being
+# 2x-parallelized the procedure is "worth" 5, 8.5 or 6 ticks per SHA1
+# round. As [long as] FPU/VIS instructions are perfectly pairable with
+# IALU ones, the round timing is defined by the maximum between VIS
+# and IALU timings. The latter varies from round to round and averages
+# out at 6.25 ticks. This means that USI&II should operate at IALU
+# rate, while USIII&IV - at VIS rate. This explains why performance
+# improvement varies among processors. Well, given that pure IALU
+# sha1-sparcv9.pl module exhibits virtually uniform performance of
+# ~9.3 cycles per SHA1 round. Timings mentioned above are theoretical
+# lower limits. Real-life performance was measured to be 6.6 cycles
+# per SHA1 round on USIIi and 8.3 on USIII. The latter is lower than
+# half-round VIS timing, because there are 16 Xupdate-free rounds,
+# which "push down" average theoretical timing to 8 cycles...
+
+# (*)	SPARC64-V[II] was originally believed to have 2 cycles VIS
+#	latency. Well, it might have, but it doesn't have dedicated
+#	VIS-unit. Instead, VIS instructions are executed by other
+#	functional units, ones used here - by IALU. This doesn't
+#	improve effective ILP...
+}
+
+# The reference Xupdate procedure is then "strained" over *pairs* of
+# BODY_NN_MM and kind of modulo-scheduled in respect to X[n]^=X[n+13]
+# and K_NN_MM addition. It's "running" 15 rounds ahead, which leaves
+# plenty of room to amortize for read-after-write hazard, as well as
+# to fetch and align input for the next spin. The VIS instructions are
+# scheduled for latency of 2 cycles, because there are not enough IALU
+# instructions to schedule for latency of 3, while scheduling for 1
+# would give no gain on USI&II anyway.
+
+sub BODY_00_19 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i&~1;
+my $k=($j+16+2)%16;	# ahead reference
+my $l=($j+16-2)%16;	# behind reference
+my $K=@VK[($j+16-2)/20];
+
+$j=($j+16)%16;
+
+$code.=<<___ if (!($i&1));
+	sll		$a,5,$tmp0			!! $i
+	and		$c,$b,$tmp3
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	 fxors		@X[($j+14)%16],@X[$j+1],@X[$j+1]! 0/ 0/ 0:X[1]^=X[14]
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	 fxor		@X[($j+2)%16],@X[($j+8)%16],%f18! 1/ 1/ 1:Tmp=X[2,3]^X[8,9]
+	sll		$b,30,$tmp2
+	add		$tmp1,$e,$e
+	andn		$d,$b,$tmp1
+	add		$Xi,$e,$e
+	 fxor		%f18,@X[$j],@X[$j]		! 2/ 4/ 3:X[0,1]^=X[2,3]^X[8,9]
+	srl		$b,2,$b
+	or		$tmp1,$tmp3,$tmp1
+	or		$tmp2,$b,$b
+	add		$tmp1,$e,$e
+	 faligndata	@X[$j],@X[$j],%f18		! 3/ 7/ 5:Tmp=X[0,1]>>>24
+___
+$code.=<<___ if ($i&1);
+	sll		$a,5,$tmp0			!! $i
+	and		$c,$b,$tmp3
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	 fpadd32	@X[$j],@X[$j],@X[$j]		! 4/ 8/ 6:X[0,1]<<=1
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	 fmul8ulx16	%f18,$fmul,%f18			! 5/10/ 7:Tmp>>=7, Tmp&=1
+	sll		$b,30,$tmp2
+	add		$tmp1,$e,$e
+	 fpadd32	$K,@X[$l],%f20			!
+	andn		$d,$b,$tmp1
+	add		$Xi,$e,$e
+	 fxors		@X[($k+13)%16],@X[$k],@X[$k]	!-1/-1/-1:X[0]^=X[13]
+	srl		$b,2,$b
+	or		$tmp1,$tmp3,$tmp1
+	 fxor		%f18,@X[$j],@X[$j]		! 8/14/10:X[0,1]|=Tmp
+	or		$tmp2,$b,$b
+	add		$tmp1,$e,$e
+___
+$code.=<<___ if ($i&1 && $i>=2);
+	 std		%f20,[$Xfer+`4*$l`]		!
+___
+}
+
+sub BODY_20_39 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i&~1;
+my $k=($j+16+2)%16;	# ahead reference
+my $l=($j+16-2)%16;	# behind reference
+my $K=@VK[($j+16-2)/20];
+
+$j=($j+16)%16;
+
+$code.=<<___ if (!($i&1) && $i<64);
+	sll		$a,5,$tmp0			!! $i
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	 fxors		@X[($j+14)%16],@X[$j+1],@X[$j+1]! 0/ 0/ 0:X[1]^=X[14]
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	 fxor		@X[($j+2)%16],@X[($j+8)%16],%f18! 1/ 1/ 1:Tmp=X[2,3]^X[8,9]
+	xor		$c,$b,$tmp0
+	add		$tmp1,$e,$e
+	sll		$b,30,$tmp2
+	xor		$d,$tmp0,$tmp1
+	 fxor		%f18,@X[$j],@X[$j]		! 2/ 4/ 3:X[0,1]^=X[2,3]^X[8,9]
+	srl		$b,2,$b
+	add		$tmp1,$e,$e
+	or		$tmp2,$b,$b
+	add		$Xi,$e,$e
+	 faligndata	@X[$j],@X[$j],%f18		! 3/ 7/ 5:Tmp=X[0,1]>>>24
+___
+$code.=<<___ if ($i&1 && $i<64);
+	sll		$a,5,$tmp0			!! $i
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	 fpadd32	@X[$j],@X[$j],@X[$j]		! 4/ 8/ 6:X[0,1]<<=1
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	 fmul8ulx16	%f18,$fmul,%f18			! 5/10/ 7:Tmp>>=7, Tmp&=1
+	xor		$c,$b,$tmp0
+	add		$tmp1,$e,$e
+	 fpadd32	$K,@X[$l],%f20			!
+	sll		$b,30,$tmp2
+	xor		$d,$tmp0,$tmp1
+	 fxors		@X[($k+13)%16],@X[$k],@X[$k]	!-1/-1/-1:X[0]^=X[13]
+	srl		$b,2,$b
+	add		$tmp1,$e,$e
+	 fxor		%f18,@X[$j],@X[$j]		! 8/14/10:X[0,1]|=Tmp
+	or		$tmp2,$b,$b
+	add		$Xi,$e,$e
+	 std		%f20,[$Xfer+`4*$l`]		!
+___
+$code.=<<___ if ($i==64);
+	sll		$a,5,$tmp0			!! $i
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	 fpadd32	$K,@X[$l],%f20
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	xor		$c,$b,$tmp0
+	add		$tmp1,$e,$e
+	sll		$b,30,$tmp2
+	xor		$d,$tmp0,$tmp1
+	 std		%f20,[$Xfer+`4*$l`]
+	srl		$b,2,$b
+	add		$tmp1,$e,$e
+	or		$tmp2,$b,$b
+	add		$Xi,$e,$e
+___
+$code.=<<___ if ($i>64);
+	sll		$a,5,$tmp0			!! $i
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	xor		$c,$b,$tmp0
+	add		$tmp1,$e,$e
+	sll		$b,30,$tmp2
+	xor		$d,$tmp0,$tmp1
+	srl		$b,2,$b
+	add		$tmp1,$e,$e
+	or		$tmp2,$b,$b
+	add		$Xi,$e,$e
+___
+}
+
+sub BODY_40_59 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i&~1;
+my $k=($j+16+2)%16;	# ahead reference
+my $l=($j+16-2)%16;	# behind reference
+my $K=@VK[($j+16-2)/20];
+
+$j=($j+16)%16;
+
+$code.=<<___ if (!($i&1));
+	sll		$a,5,$tmp0			!! $i
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	 fxors		@X[($j+14)%16],@X[$j+1],@X[$j+1]! 0/ 0/ 0:X[1]^=X[14]
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	 fxor		@X[($j+2)%16],@X[($j+8)%16],%f18! 1/ 1/ 1:Tmp=X[2,3]^X[8,9]
+	and		$c,$b,$tmp0
+	add		$tmp1,$e,$e
+	sll		$b,30,$tmp2
+	or		$c,$b,$tmp1
+	 fxor		%f18,@X[$j],@X[$j]		! 2/ 4/ 3:X[0,1]^=X[2,3]^X[8,9]
+	srl		$b,2,$b
+	and		$d,$tmp1,$tmp1
+	add		$Xi,$e,$e
+	or		$tmp1,$tmp0,$tmp1
+	 faligndata	@X[$j],@X[$j],%f18		! 3/ 7/ 5:Tmp=X[0,1]>>>24
+	or		$tmp2,$b,$b
+	add		$tmp1,$e,$e
+	 fpadd32	@X[$j],@X[$j],@X[$j]		! 4/ 8/ 6:X[0,1]<<=1
+___
+$code.=<<___ if ($i&1);
+	sll		$a,5,$tmp0			!! $i
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	 fmul8ulx16	%f18,$fmul,%f18			! 5/10/ 7:Tmp>>=7, Tmp&=1
+	and		$c,$b,$tmp0
+	add		$tmp1,$e,$e
+	 fpadd32	$K,@X[$l],%f20			!
+	sll		$b,30,$tmp2
+	or		$c,$b,$tmp1
+	 fxors		@X[($k+13)%16],@X[$k],@X[$k]	!-1/-1/-1:X[0]^=X[13]
+	srl		$b,2,$b
+	and		$d,$tmp1,$tmp1
+	 fxor		%f18,@X[$j],@X[$j]		! 8/14/10:X[0,1]|=Tmp
+	add		$Xi,$e,$e
+	or		$tmp1,$tmp0,$tmp1
+	or		$tmp2,$b,$b
+	add		$tmp1,$e,$e
+	 std		%f20,[$Xfer+`4*$l`]		!
+___
+}
+
+# If there is more data to process, then we pre-fetch the data for
+# next iteration in last ten rounds...
+sub BODY_70_79 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i&~1;
+my $m=($i%8)*2;
+
+$j=($j+16)%16;
+
+$code.=<<___ if ($i==70);
+	sll		$a,5,$tmp0			!! $i
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	 ldd		[$inp+64],@X[0]
+	xor		$c,$b,$tmp0
+	add		$tmp1,$e,$e
+	sll		$b,30,$tmp2
+	xor		$d,$tmp0,$tmp1
+	srl		$b,2,$b
+	add		$tmp1,$e,$e
+	or		$tmp2,$b,$b
+	add		$Xi,$e,$e
+
+	and		$inp,-64,$nXfer
+	inc		64,$inp
+	and		$nXfer,255,$nXfer
+	alignaddr	%g0,$align,%g0
+	add		$base,$nXfer,$nXfer
+___
+$code.=<<___ if ($i==71);
+	sll		$a,5,$tmp0			!! $i
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	xor		$c,$b,$tmp0
+	add		$tmp1,$e,$e
+	sll		$b,30,$tmp2
+	xor		$d,$tmp0,$tmp1
+	srl		$b,2,$b
+	add		$tmp1,$e,$e
+	or		$tmp2,$b,$b
+	add		$Xi,$e,$e
+___
+$code.=<<___ if ($i>=72);
+	 faligndata	@X[$m],@X[$m+2],@X[$m]
+	sll		$a,5,$tmp0			!! $i
+	ld		[$Xfer+`4*($i%16)`],$Xi
+	srl		$a,27,$tmp1
+	add		$tmp0,$e,$e
+	xor		$c,$b,$tmp0
+	add		$tmp1,$e,$e
+	 fpadd32	$VK_00_19,@X[$m],%f20
+	sll		$b,30,$tmp2
+	xor		$d,$tmp0,$tmp1
+	srl		$b,2,$b
+	add		$tmp1,$e,$e
+	or		$tmp2,$b,$b
+	add		$Xi,$e,$e
+___
+$code.=<<___ if ($i<77);
+	 ldd		[$inp+`8*($i+1-70)`],@X[2*($i+1-70)]
+___
+$code.=<<___ if ($i==77);	# redundant if $inp was aligned
+	 add		$align,63,$tmp0
+	 and		$tmp0,-8,$tmp0
+	 ldd		[$inp+$tmp0],@X[16]
+___
+$code.=<<___ if ($i>=72);
+	 std		%f20,[$nXfer+`4*$m`]
+___
+}
+
+$code.=<<___;
+.section	".text",#alloc,#execinstr
+
+.align	64
+vis_const:
+.long	0x5a827999,0x5a827999	! K_00_19
+.long	0x6ed9eba1,0x6ed9eba1	! K_20_39
+.long	0x8f1bbcdc,0x8f1bbcdc	! K_40_59
+.long	0xca62c1d6,0xca62c1d6	! K_60_79
+.long	0x00000100,0x00000100
+.align	64
+.type	vis_const,#object
+.size	vis_const,(.-vis_const)
+
+.globl	sha1_block_data_order
+sha1_block_data_order:
+	save	%sp,-$frame,%sp
+	add	%fp,$bias-256,$base
+
+1:	call	.+8
+	add	%o7,vis_const-1b,$tmp0
+
+	ldd	[$tmp0+0],$VK_00_19
+	ldd	[$tmp0+8],$VK_20_39
+	ldd	[$tmp0+16],$VK_40_59
+	ldd	[$tmp0+24],$VK_60_79
+	ldd	[$tmp0+32],$fmul
+
+	ld	[$ctx+0],$Actx
+	and	$base,-256,$base
+	ld	[$ctx+4],$Bctx
+	sub	$base,$bias+$frame,%sp
+	ld	[$ctx+8],$Cctx
+	and	$inp,7,$align
+	ld	[$ctx+12],$Dctx
+	and	$inp,-8,$inp
+	ld	[$ctx+16],$Ectx
+
+	! X[16] is maintained in FP register bank
+	alignaddr	%g0,$align,%g0
+	ldd		[$inp+0],@X[0]
+	sub		$inp,-64,$Xfer
+	ldd		[$inp+8],@X[2]
+	and		$Xfer,-64,$Xfer
+	ldd		[$inp+16],@X[4]
+	and		$Xfer,255,$Xfer
+	ldd		[$inp+24],@X[6]
+	add		$base,$Xfer,$Xfer
+	ldd		[$inp+32],@X[8]
+	ldd		[$inp+40],@X[10]
+	ldd		[$inp+48],@X[12]
+	brz,pt		$align,.Laligned
+	ldd		[$inp+56],@X[14]
+
+	ldd		[$inp+64],@X[16]
+	faligndata	@X[0],@X[2],@X[0]
+	faligndata	@X[2],@X[4],@X[2]
+	faligndata	@X[4],@X[6],@X[4]
+	faligndata	@X[6],@X[8],@X[6]
+	faligndata	@X[8],@X[10],@X[8]
+	faligndata	@X[10],@X[12],@X[10]
+	faligndata	@X[12],@X[14],@X[12]
+	faligndata	@X[14],@X[16],@X[14]
+
+.Laligned:
+	mov		5,$tmp0
+	dec		1,$len
+	alignaddr	%g0,$tmp0,%g0
+	fpadd32		$VK_00_19,@X[0],%f16
+	fpadd32		$VK_00_19,@X[2],%f18
+	fpadd32		$VK_00_19,@X[4],%f20
+	fpadd32		$VK_00_19,@X[6],%f22
+	fpadd32		$VK_00_19,@X[8],%f24
+	fpadd32		$VK_00_19,@X[10],%f26
+	fpadd32		$VK_00_19,@X[12],%f28
+	fpadd32		$VK_00_19,@X[14],%f30
+	std		%f16,[$Xfer+0]
+	mov		$Actx,$A
+	std		%f18,[$Xfer+8]
+	mov		$Bctx,$B
+	std		%f20,[$Xfer+16]
+	mov		$Cctx,$C
+	std		%f22,[$Xfer+24]
+	mov		$Dctx,$D
+	std		%f24,[$Xfer+32]
+	mov		$Ectx,$E
+	std		%f26,[$Xfer+40]
+	fxors		@X[13],@X[0],@X[0]
+	std		%f28,[$Xfer+48]
+	ba		.Loop
+	std		%f30,[$Xfer+56]
+.align	32
+.Loop:
+___
+for ($i=0;$i<20;$i++)	{ &BODY_00_19($i,@V); unshift(@V,pop(@V)); }
+for (;$i<40;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+for (;$i<60;$i++)	{ &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
+for (;$i<70;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	tst		$len
+	bz,pn		`$bits==32?"%icc":"%xcc"`,.Ltail
+	nop
+___
+for (;$i<80;$i++)	{ &BODY_70_79($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	add		$A,$Actx,$Actx
+	add		$B,$Bctx,$Bctx
+	add		$C,$Cctx,$Cctx
+	add		$D,$Dctx,$Dctx
+	add		$E,$Ectx,$Ectx
+	mov		5,$tmp0
+	fxors		@X[13],@X[0],@X[0]
+	mov		$Actx,$A
+	mov		$Bctx,$B
+	mov		$Cctx,$C
+	mov		$Dctx,$D
+	mov		$Ectx,$E
+	alignaddr	%g0,$tmp0,%g0	
+	dec		1,$len
+	ba		.Loop
+	mov		$nXfer,$Xfer
+
+.align	32
+.Ltail:
+___
+for($i=70;$i<80;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	add	$A,$Actx,$Actx
+	add	$B,$Bctx,$Bctx
+	add	$C,$Cctx,$Cctx
+	add	$D,$Dctx,$Dctx
+	add	$E,$Ectx,$Ectx
+
+	st	$Actx,[$ctx+0]
+	st	$Bctx,[$ctx+4]
+	st	$Cctx,[$ctx+8]
+	st	$Dctx,[$ctx+12]
+	st	$Ectx,[$ctx+16]
+
+	ret
+	restore
+.type	sha1_block_data_order,#function
+.size	sha1_block_data_order,(.-sha1_block_data_order)
+.asciz	"SHA1 block transform for SPARCv9a, CRYPTOGAMS by <appro\@openssl.org>"
+.align	4
+___
+
+# 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.
+# 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 {
+my ($mnemonic,$rs1,$rs2,$rd)=@_;
+my ($ref,$opf);
+my %visopf = (	"fmul8ulx16"	=> 0x037,
+		"faligndata"	=> 0x048,
+		"fpadd32"	=> 0x052,
+		"fxor"		=> 0x06c,
+		"fxors"		=> 0x06d	);
+
+    $ref = "$mnemonic\t$rs1,$rs2,$rd";
+
+    if ($opf=$visopf{$mnemonic}) {
+	foreach ($rs1,$rs2,$rd) {
+	    return $ref if (!/%f([0-9]{1,2})/);
+	    $_=$1;
+	    if ($1>=32) {
+		return $ref if ($1&1);
+		# re-encode for upper double register addressing
+		$_=($1|$1>>5)&31;
+	    }
+	}
+
+	return	sprintf ".word\t0x%08x !%s",
+			0x81b00000|$rd<<25|$rs1<<14|$opf<<5|$rs2,
+			$ref;
+    } else {
+	return $ref;
+    }
+}
+sub unalignaddr {
+my ($mnemonic,$rs1,$rs2,$rd)=@_;
+my %bias = ( "g" => 0, "o" => 8, "l" => 16, "i" => 24 );
+my $ref="$mnemonic\t$rs1,$rs2,$rd";
+
+    foreach ($rs1,$rs2,$rd) {
+	if (/%([goli])([0-7])/)	{ $_=$bias{$1}+$2; }
+	else			{ return $ref; }
+    }
+    return  sprintf ".word\t0x%08x !%s",
+		    0x81b00300|$rd<<25|$rs1<<14|$rs2,
+		    $ref;
+}
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+$code =~ s/\b(f[^\s]*)\s+(%f[0-9]{1,2}),(%f[0-9]{1,2}),(%f[0-9]{1,2})/
+		&unvis($1,$2,$3,$4)
+	  /gem;
+$code =~ s/\b(alignaddr)\s+(%[goli][0-7]),(%[goli][0-7]),(%[goli][0-7])/
+		&unalignaddr($1,$2,$3,$4)
+	  /gem;
+print $code;
+close STDOUT;
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha1-thumb.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha1-thumb.pl
new file mode 100644
index 0000000000..7c9ea9b029
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha1-thumb.pl
@@ -0,0 +1,259 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# sha1_block for Thumb.
+#
+# January 2007.
+#
+# The code does not present direct interest to OpenSSL, because of low
+# performance. Its purpose is to establish _size_ benchmark. Pretty
+# useless one I must say, because 30% or 88 bytes larger ARMv4 code
+# [avialable on demand] is almost _twice_ as fast. It should also be
+# noted that in-lining of .Lcommon and .Lrotate improves performance
+# by over 40%, while code increases by only 10% or 32 bytes. But once
+# again, the goal was to establish _size_ benchmark, not performance.
+
+$output=shift;
+open STDOUT,">$output";
+
+$inline=0;
+#$cheat_on_binutils=1;
+
+$t0="r0";
+$t1="r1";
+$t2="r2";
+$a="r3";
+$b="r4";
+$c="r5";
+$d="r6";
+$e="r7";
+$K="r8";	# "upper" registers can be used in add/sub and mov insns
+$ctx="r9";
+$inp="r10";
+$len="r11";
+$Xi="r12";
+
+sub common {
+<<___;
+	sub	$t0,#4
+	ldr	$t1,[$t0]
+	add	$e,$K			@ E+=K_xx_xx
+	lsl	$t2,$a,#5
+	add	$t2,$e
+	lsr	$e,$a,#27
+	add	$t2,$e			@ E+=ROR(A,27)
+	add	$t2,$t1			@ E+=X[i]
+___
+}
+sub rotate {
+<<___;
+	mov	$e,$d			@ E=D
+	mov	$d,$c			@ D=C
+	lsl	$c,$b,#30
+	lsr	$b,$b,#2
+	orr	$c,$b			@ C=ROR(B,2)
+	mov	$b,$a			@ B=A
+	add	$a,$t2,$t1		@ A=E+F_xx_xx(B,C,D)
+___
+}
+
+sub BODY_00_19 {
+$code.=$inline?&common():"\tbl	.Lcommon\n";
+$code.=<<___;
+	mov	$t1,$c
+	eor	$t1,$d
+	and	$t1,$b
+	eor	$t1,$d			@ F_00_19(B,C,D)
+___
+$code.=$inline?&rotate():"\tbl	.Lrotate\n";
+}
+
+sub BODY_20_39 {
+$code.=$inline?&common():"\tbl	.Lcommon\n";
+$code.=<<___;
+	mov	$t1,$b
+	eor	$t1,$c
+	eor	$t1,$d			@ F_20_39(B,C,D)
+___
+$code.=$inline?&rotate():"\tbl	.Lrotate\n";
+}
+
+sub BODY_40_59 {
+$code.=$inline?&common():"\tbl	.Lcommon\n";
+$code.=<<___;
+	mov	$t1,$b
+	and	$t1,$c
+	mov	$e,$b
+	orr	$e,$c
+	and	$e,$d
+	orr	$t1,$e			@ F_40_59(B,C,D)
+___
+$code.=$inline?&rotate():"\tbl	.Lrotate\n";
+}
+
+$code=<<___;
+.text
+.code	16
+
+.global	sha1_block_data_order
+.type	sha1_block_data_order,%function
+
+.align	2
+sha1_block_data_order:
+___
+if ($cheat_on_binutils) {
+$code.=<<___;
+.code	32
+	add	r3,pc,#1
+	bx	r3			@ switch to Thumb ISA
+.code	16
+___
+}
+$code.=<<___;
+	push	{r4-r7}
+	mov	r3,r8
+	mov	r4,r9
+	mov	r5,r10
+	mov	r6,r11
+	mov	r7,r12
+	push	{r3-r7,lr}
+	lsl	r2,#6
+	mov	$ctx,r0			@ save context
+	mov	$inp,r1			@ save inp
+	mov	$len,r2			@ save len
+	add	$len,$inp		@ $len to point at inp end
+
+.Lloop:
+	mov	$Xi,sp
+	mov	$t2,sp
+	sub	$t2,#16*4		@ [3]
+.LXload:
+	ldrb	$a,[$t1,#0]		@ $t1 is r1 and holds inp
+	ldrb	$b,[$t1,#1]
+	ldrb	$c,[$t1,#2]
+	ldrb	$d,[$t1,#3]
+	lsl	$a,#24
+	lsl	$b,#16
+	lsl	$c,#8
+	orr	$a,$b
+	orr	$a,$c
+	orr	$a,$d
+	add	$t1,#4
+	push	{$a}
+	cmp	sp,$t2
+	bne	.LXload			@ [+14*16]
+
+	mov	$inp,$t1		@ update $inp
+	sub	$t2,#32*4
+	sub	$t2,#32*4
+	mov	$e,#31			@ [+4]
+.LXupdate:
+	ldr	$a,[sp,#15*4]
+	ldr	$b,[sp,#13*4]
+	ldr	$c,[sp,#7*4]
+	ldr	$d,[sp,#2*4]
+	eor	$a,$b
+	eor	$a,$c
+	eor	$a,$d
+	ror	$a,$e
+	push	{$a}
+	cmp	sp,$t2
+	bne	.LXupdate		@ [+(11+1)*64]
+
+	ldmia	$t0!,{$a,$b,$c,$d,$e}	@ $t0 is r0 and holds ctx
+	mov	$t0,$Xi
+
+	ldr	$t2,.LK_00_19
+	mov	$t1,$t0
+	sub	$t1,#20*4
+	mov	$Xi,$t1
+	mov	$K,$t2			@ [+7+4]
+.L_00_19:
+___
+	&BODY_00_19();
+$code.=<<___;
+	cmp	$Xi,$t0
+	bne	.L_00_19		@ [+(2+9+4+2+8+2)*20]
+
+	ldr	$t2,.LK_20_39
+	mov	$t1,$t0
+	sub	$t1,#20*4
+	mov	$Xi,$t1
+	mov	$K,$t2			@ [+5]
+.L_20_39_or_60_79:
+___
+	&BODY_20_39();
+$code.=<<___;
+	cmp	$Xi,$t0
+	bne	.L_20_39_or_60_79	@ [+(2+9+3+2+8+2)*20*2]
+	cmp	sp,$t0
+	beq	.Ldone			@ [+2]
+
+	ldr	$t2,.LK_40_59
+	mov	$t1,$t0
+	sub	$t1,#20*4
+	mov	$Xi,$t1
+	mov	$K,$t2			@ [+5]
+.L_40_59:
+___
+	&BODY_40_59();
+$code.=<<___;
+	cmp	$Xi,$t0
+	bne	.L_40_59		@ [+(2+9+6+2+8+2)*20]
+
+	ldr	$t2,.LK_60_79
+	mov	$Xi,sp
+	mov	$K,$t2
+	b	.L_20_39_or_60_79	@ [+4]
+.Ldone:
+	mov	$t0,$ctx
+	ldr	$t1,[$t0,#0]
+	ldr	$t2,[$t0,#4]
+	add	$a,$t1
+	ldr	$t1,[$t0,#8]
+	add	$b,$t2
+	ldr	$t2,[$t0,#12]
+	add	$c,$t1
+	ldr	$t1,[$t0,#16]
+	add	$d,$t2
+	add	$e,$t1
+	stmia	$t0!,{$a,$b,$c,$d,$e}	@ [+20]
+
+	add	sp,#80*4		@ deallocate stack frame
+	mov	$t0,$ctx		@ restore ctx
+	mov	$t1,$inp		@ restore inp
+	cmp	$t1,$len
+	beq	.Lexit
+	b	.Lloop			@ [+6] total 3212 cycles
+.Lexit:
+	pop	{r2-r7}
+	mov	r8,r2
+	mov	r9,r3
+	mov	r10,r4
+	mov	r11,r5
+	mov	r12,r6
+	mov	lr,r7
+	pop	{r4-r7}
+	bx	lr
+.align	2
+___
+$code.=".Lcommon:\n".&common()."\tmov	pc,lr\n" if (!$inline);
+$code.=".Lrotate:\n".&rotate()."\tmov	pc,lr\n" if (!$inline);
+$code.=<<___;
+.align	2
+.LK_00_19:	.word	0x5a827999
+.LK_20_39:	.word	0x6ed9eba1
+.LK_40_59:	.word	0x8f1bbcdc
+.LK_60_79:	.word	0xca62c1d6
+.size	sha1_block_data_order,.-sha1_block_data_order
+.asciz	"SHA1 block transform for Thumb, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+print $code;
+close STDOUT; # enforce flush
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha1-x86_64.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha1-x86_64.pl
new file mode 100755
index 0000000000..f15c7ec39b
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha1-x86_64.pl
@@ -0,0 +1,1261 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> 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_block procedure for x86_64.
+#
+# It was brought to my attention that on EM64T compiler-generated code
+# was far behind 32-bit assembler implementation. This is unlike on
+# Opteron where compiler-generated code was only 15% behind 32-bit
+# assembler, which originally made it hard to motivate the effort.
+# There was suggestion to mechanically translate 32-bit code, but I
+# dismissed it, reasoning that x86_64 offers enough register bank
+# capacity to fully utilize SHA-1 parallelism. Therefore this fresh
+# implementation:-) However! While 64-bit code does perform better
+# on Opteron, I failed to beat 32-bit assembler on EM64T core. Well,
+# x86_64 does offer larger *addressable* bank, but out-of-order core
+# reaches for even more registers through dynamic aliasing, and EM64T
+# core must have managed to run-time optimize even 32-bit code just as
+# good as 64-bit one. Performance improvement is summarized in the
+# following table:
+#
+#		gcc 3.4		32-bit asm	cycles/byte
+# Opteron	+45%		+20%		6.8
+# Xeon P4	+65%		+0%		9.9
+# Core2		+60%		+10%		7.0
+
+# August 2009.
+#
+# The code was revised to minimize code size and to maximize
+# "distance" between instructions producing input to 'lea'
+# instruction and the 'lea' instruction itself, which is essential
+# for Intel Atom core.
+
+# October 2010.
+#
+# Add SSSE3, Supplemental[!] SSE3, implementation. The idea behind it
+# is to offload message schedule denoted by Wt in NIST specification,
+# or Xupdate in OpenSSL source, to SIMD unit. See sha1-586.pl module
+# for background and implementation details. The only difference from
+# 32-bit code is that 64-bit code doesn't have to spill @X[] elements
+# to free temporary registers.
+
+# April 2011.
+#
+# Add AVX code path. See sha1-586.pl for further information.
+
+######################################################################
+# Current performance is summarized in following table. Numbers are
+# CPU clock cycles spent to process single byte (less is better).
+#
+#		x86_64		SSSE3		AVX
+# P4		9.8		-
+# Opteron	6.6		-
+# Core2		6.7		6.1/+10%	-
+# Atom		11.0		9.7/+13%	-
+# Westmere	7.1		5.6/+27%	-
+# Sandy Bridge	7.9		6.3/+25%	5.2/+51%
+
+$flavour = shift;
+$output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
+
+$win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
+die "can't locate x86_64-xlate.pl";
+
+$avx=1 if (`$ENV{CC} -Wa,-v -c -o /dev/null -x assembler /dev/null 2>&1`
+		=~ /GNU assembler version ([2-9]\.[0-9]+)/ &&
+	   $1>=2.19);
+$avx=1 if (!$avx && $win64 && ($flavour =~ /nasm/ || $ENV{ASM} =~ /nasm/) &&
+	   `nasm -v 2>&1` =~ /NASM version ([2-9]\.[0-9]+)/ &&
+	   $1>=2.09);
+$avx=1 if (!$avx && $win64 && ($flavour =~ /masm/ || $ENV{ASM} =~ /ml64/) &&
+	   `ml64 2>&1` =~ /Version ([0-9]+)\./ &&
+	   $1>=10);
+
+open OUT,"| \"$^X\" $xlate $flavour $output";
+*STDOUT=*OUT;
+
+$ctx="%rdi";	# 1st arg
+$inp="%rsi";	# 2nd arg
+$num="%rdx";	# 3rd arg
+
+# reassign arguments in order to produce more compact code
+$ctx="%r8";
+$inp="%r9";
+$num="%r10";
+
+$t0="%eax";
+$t1="%ebx";
+$t2="%ecx";
+@xi=("%edx","%ebp");
+$A="%esi";
+$B="%edi";
+$C="%r11d";
+$D="%r12d";
+$E="%r13d";
+
+@V=($A,$B,$C,$D,$E);
+
+sub BODY_00_19 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+$code.=<<___ if ($i==0);
+	mov	`4*$i`($inp),$xi[0]
+	bswap	$xi[0]
+	mov	$xi[0],`4*$i`(%rsp)
+___
+$code.=<<___ if ($i<15);
+	mov	$c,$t0
+	mov	`4*$j`($inp),$xi[1]
+	mov	$a,$t2
+	xor	$d,$t0
+	bswap	$xi[1]
+	rol	\$5,$t2
+	lea	0x5a827999($xi[0],$e),$e
+	and	$b,$t0
+	mov	$xi[1],`4*$j`(%rsp)
+	add	$t2,$e
+	xor	$d,$t0
+	rol	\$30,$b
+	add	$t0,$e
+___
+$code.=<<___ if ($i>=15);
+	mov	`4*($j%16)`(%rsp),$xi[1]
+	mov	$c,$t0
+	mov	$a,$t2
+	xor	`4*(($j+2)%16)`(%rsp),$xi[1]
+	xor	$d,$t0
+	rol	\$5,$t2
+	xor	`4*(($j+8)%16)`(%rsp),$xi[1]
+	and	$b,$t0
+	lea	0x5a827999($xi[0],$e),$e
+	xor	`4*(($j+13)%16)`(%rsp),$xi[1]
+	xor	$d,$t0
+	rol	\$1,$xi[1]
+	add	$t2,$e
+	rol	\$30,$b
+	mov	$xi[1],`4*($j%16)`(%rsp)
+	add	$t0,$e
+___
+unshift(@xi,pop(@xi));
+}
+
+sub BODY_20_39 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+my $K=($i<40)?0x6ed9eba1:0xca62c1d6;
+$code.=<<___ if ($i<79);
+	mov	`4*($j%16)`(%rsp),$xi[1]
+	mov	$c,$t0
+	mov	$a,$t2
+	xor	`4*(($j+2)%16)`(%rsp),$xi[1]
+	xor	$b,$t0
+	rol	\$5,$t2
+	lea	$K($xi[0],$e),$e
+	xor	`4*(($j+8)%16)`(%rsp),$xi[1]
+	xor	$d,$t0
+	add	$t2,$e
+	xor	`4*(($j+13)%16)`(%rsp),$xi[1]
+	rol	\$30,$b
+	add	$t0,$e
+	rol	\$1,$xi[1]
+___
+$code.=<<___ if ($i<76);
+	mov	$xi[1],`4*($j%16)`(%rsp)
+___
+$code.=<<___ if ($i==79);
+	mov	$c,$t0
+	mov	$a,$t2
+	xor	$b,$t0
+	lea	$K($xi[0],$e),$e
+	rol	\$5,$t2
+	xor	$d,$t0
+	add	$t2,$e
+	rol	\$30,$b
+	add	$t0,$e
+___
+unshift(@xi,pop(@xi));
+}
+
+sub BODY_40_59 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $j=$i+1;
+$code.=<<___;
+	mov	`4*($j%16)`(%rsp),$xi[1]
+	mov	$c,$t0
+	mov	$c,$t1
+	xor	`4*(($j+2)%16)`(%rsp),$xi[1]
+	and	$d,$t0
+	mov	$a,$t2
+	xor	`4*(($j+8)%16)`(%rsp),$xi[1]
+	xor	$d,$t1
+	lea	0x8f1bbcdc($xi[0],$e),$e
+	rol	\$5,$t2
+	xor	`4*(($j+13)%16)`(%rsp),$xi[1]
+	add	$t0,$e
+	and	$b,$t1
+	rol	\$1,$xi[1]
+	add	$t1,$e
+	rol	\$30,$b
+	mov	$xi[1],`4*($j%16)`(%rsp)
+	add	$t2,$e
+___
+unshift(@xi,pop(@xi));
+}
+
+$code.=<<___;
+.text
+.extern	OPENSSL_ia32cap_P
+
+.globl	sha1_block_data_order
+.type	sha1_block_data_order,\@function,3
+.align	16
+sha1_block_data_order:
+	mov	OPENSSL_ia32cap_P+0(%rip),%r9d
+	mov	OPENSSL_ia32cap_P+4(%rip),%r8d
+	test	\$`1<<9`,%r8d		# check SSSE3 bit
+	jz	.Lialu
+___
+$code.=<<___ if ($avx);
+	and	\$`1<<28`,%r8d		# mask AVX bit
+	and	\$`1<<30`,%r9d		# mask "Intel CPU" bit
+	or	%r9d,%r8d
+	cmp	\$`1<<28|1<<30`,%r8d
+	je	_avx_shortcut
+___
+$code.=<<___;
+	jmp	_ssse3_shortcut
+
+.align	16
+.Lialu:
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	mov	%rsp,%r11
+	mov	%rdi,$ctx	# reassigned argument
+	sub	\$`8+16*4`,%rsp
+	mov	%rsi,$inp	# reassigned argument
+	and	\$-64,%rsp
+	mov	%rdx,$num	# reassigned argument
+	mov	%r11,`16*4`(%rsp)
+.Lprologue:
+
+	mov	0($ctx),$A
+	mov	4($ctx),$B
+	mov	8($ctx),$C
+	mov	12($ctx),$D
+	mov	16($ctx),$E
+	jmp	.Lloop
+
+.align	16
+.Lloop:
+___
+for($i=0;$i<20;$i++)	{ &BODY_00_19($i,@V); unshift(@V,pop(@V)); }
+for(;$i<40;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+for(;$i<60;$i++)	{ &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
+for(;$i<80;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	add	0($ctx),$A
+	add	4($ctx),$B
+	add	8($ctx),$C
+	add	12($ctx),$D
+	add	16($ctx),$E
+	mov	$A,0($ctx)
+	mov	$B,4($ctx)
+	mov	$C,8($ctx)
+	mov	$D,12($ctx)
+	mov	$E,16($ctx)
+
+	sub	\$1,$num
+	lea	`16*4`($inp),$inp
+	jnz	.Lloop
+
+	mov	`16*4`(%rsp),%rsi
+	mov	(%rsi),%r13
+	mov	8(%rsi),%r12
+	mov	16(%rsi),%rbp
+	mov	24(%rsi),%rbx
+	lea	32(%rsi),%rsp
+.Lepilogue:
+	ret
+.size	sha1_block_data_order,.-sha1_block_data_order
+___
+{{{
+my $Xi=4;
+my @X=map("%xmm$_",(4..7,0..3));
+my @Tx=map("%xmm$_",(8..10));
+my @V=($A,$B,$C,$D,$E)=("%eax","%ebx","%ecx","%edx","%ebp");	# size optimization
+my @T=("%esi","%edi");
+my $j=0;
+my $K_XX_XX="%r11";
+
+my $_rol=sub { &rol(@_) };
+my $_ror=sub { &ror(@_) };
+
+$code.=<<___;
+.type	sha1_block_data_order_ssse3,\@function,3
+.align	16
+sha1_block_data_order_ssse3:
+_ssse3_shortcut:
+	push	%rbx
+	push	%rbp
+	push	%r12
+	lea	`-64-($win64?5*16:0)`(%rsp),%rsp
+___
+$code.=<<___ if ($win64);
+	movaps	%xmm6,64+0(%rsp)
+	movaps	%xmm7,64+16(%rsp)
+	movaps	%xmm8,64+32(%rsp)
+	movaps	%xmm9,64+48(%rsp)
+	movaps	%xmm10,64+64(%rsp)
+.Lprologue_ssse3:
+___
+$code.=<<___;
+	mov	%rdi,$ctx	# reassigned argument
+	mov	%rsi,$inp	# reassigned argument
+	mov	%rdx,$num	# reassigned argument
+
+	shl	\$6,$num
+	add	$inp,$num
+	lea	K_XX_XX(%rip),$K_XX_XX
+
+	mov	0($ctx),$A		# load context
+	mov	4($ctx),$B
+	mov	8($ctx),$C
+	mov	12($ctx),$D
+	mov	$B,@T[0]		# magic seed
+	mov	16($ctx),$E
+
+	movdqa	64($K_XX_XX),@X[2]	# pbswap mask
+	movdqa	0($K_XX_XX),@Tx[1]	# K_00_19
+	movdqu	0($inp),@X[-4&7]	# load input to %xmm[0-3]
+	movdqu	16($inp),@X[-3&7]
+	movdqu	32($inp),@X[-2&7]
+	movdqu	48($inp),@X[-1&7]
+	pshufb	@X[2],@X[-4&7]		# byte swap
+	add	\$64,$inp
+	pshufb	@X[2],@X[-3&7]
+	pshufb	@X[2],@X[-2&7]
+	pshufb	@X[2],@X[-1&7]
+	paddd	@Tx[1],@X[-4&7]		# add K_00_19
+	paddd	@Tx[1],@X[-3&7]
+	paddd	@Tx[1],@X[-2&7]
+	movdqa	@X[-4&7],0(%rsp)	# X[]+K xfer to IALU
+	psubd	@Tx[1],@X[-4&7]		# restore X[]
+	movdqa	@X[-3&7],16(%rsp)
+	psubd	@Tx[1],@X[-3&7]
+	movdqa	@X[-2&7],32(%rsp)
+	psubd	@Tx[1],@X[-2&7]
+	jmp	.Loop_ssse3
+___
+
+sub AUTOLOAD()		# thunk [simplified] 32-bit style perlasm
+{ my $opcode = $AUTOLOAD; $opcode =~ s/.*:://;
+  my $arg = pop;
+    $arg = "\$$arg" if ($arg*1 eq $arg);
+    $code .= "\t$opcode\t".join(',',$arg,reverse @_)."\n";
+}
+
+sub Xupdate_ssse3_16_31()		# recall that $Xi starts wtih 4
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 40 instructions
+  my ($a,$b,$c,$d,$e);
+
+	&movdqa	(@X[0],@X[-3&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&movdqa	(@Tx[0],@X[-1&7]);
+	&palignr(@X[0],@X[-4&7],8);	# compose "X[-14]" in "X[0]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &paddd	(@Tx[1],@X[-1&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&psrldq	(@Tx[0],4);		# "X[-3]", 3 dwords
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&pxor	(@X[0],@X[-4&7]);	# "X[0]"^="X[-16]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pxor	(@Tx[0],@X[-2&7]);	# "X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pxor	(@X[0],@Tx[0]);		# "X[0]"^="X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &movdqa	(eval(16*(($Xi-1)&3))."(%rsp)",@Tx[1]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&movdqa	(@Tx[2],@X[0]);
+	&movdqa	(@Tx[0],@X[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pslldq	(@Tx[2],12);		# "X[0]"<<96, extract one dword
+	&paddd	(@X[0],@X[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&psrld	(@Tx[0],31);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&movdqa	(@Tx[1],@Tx[2]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&psrld	(@Tx[2],30);
+	&por	(@X[0],@Tx[0]);		# "X[0]"<<<=1
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pslld	(@Tx[1],2);
+	&pxor	(@X[0],@Tx[2]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &movdqa	(@Tx[2],eval(16*(($Xi)/5))."($K_XX_XX)");	# K_XX_XX
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&pxor	(@X[0],@Tx[1]);		# "X[0]"^=("X[0]">>96)<<<2
+
+	 foreach (@insns) { eval; }	# remaining instructions [if any]
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+		push(@Tx,shift(@Tx));
+}
+
+sub Xupdate_ssse3_32_79()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 to 48 instructions
+  my ($a,$b,$c,$d,$e);
+
+	&movdqa	(@Tx[0],@X[-1&7])	if ($Xi==8);
+	 eval(shift(@insns));		# body_20_39
+	&pxor	(@X[0],@X[-4&7]);	# "X[0]"="X[-32]"^"X[-16]"
+	&palignr(@Tx[0],@X[-2&7],8);	# compose "X[-6]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&pxor	(@X[0],@X[-7&7]);	# "X[0]"^="X[-28]"
+	 eval(shift(@insns));
+	 eval(shift(@insns))	if (@insns[0] !~ /&ro[rl]/);
+	if ($Xi%5) {
+	  &movdqa	(@Tx[2],@Tx[1]);# "perpetuate" K_XX_XX...
+	} else {			# ... or load next one
+	  &movdqa	(@Tx[2],eval(16*($Xi/5))."($K_XX_XX)");
+	}
+	  &paddd	(@Tx[1],@X[-1&7]);
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&pxor	(@X[0],@Tx[0]);		# "X[0]"^="X[-6]"
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&movdqa	(@Tx[0],@X[0]);
+	  &movdqa	(eval(16*(($Xi-1)&3))."(%rsp)",@Tx[1]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&pslld	(@X[0],2);
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	&psrld	(@Tx[0],30);
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&por	(@X[0],@Tx[0]);		# "X[0]"<<<=2
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	  &movdqa	(@Tx[1],@X[0])	if ($Xi<19);
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+
+	 foreach (@insns) { eval; }	# remaining instructions
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+		push(@Tx,shift(@Tx));
+}
+
+sub Xuplast_ssse3_80()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	  &paddd	(@Tx[1],@X[-1&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &movdqa	(eval(16*(($Xi-1)&3))."(%rsp)",@Tx[1]);	# X[]+K xfer IALU
+
+	 foreach (@insns) { eval; }		# remaining instructions
+
+	&cmp	($inp,$num);
+	&je	(".Ldone_ssse3");
+
+	unshift(@Tx,pop(@Tx));
+
+	&movdqa	(@X[2],"64($K_XX_XX)");		# pbswap mask
+	&movdqa	(@Tx[1],"0($K_XX_XX)");		# K_00_19
+	&movdqu	(@X[-4&7],"0($inp)");		# load input
+	&movdqu	(@X[-3&7],"16($inp)");
+	&movdqu	(@X[-2&7],"32($inp)");
+	&movdqu	(@X[-1&7],"48($inp)");
+	&pshufb	(@X[-4&7],@X[2]);		# byte swap
+	&add	($inp,64);
+
+  $Xi=0;
+}
+
+sub Xloop_ssse3()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&pshufb	(@X[($Xi-3)&7],@X[2]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&paddd	(@X[($Xi-4)&7],@Tx[1]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&movdqa	(eval(16*$Xi)."(%rsp)",@X[($Xi-4)&7]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&psubd	(@X[($Xi-4)&7],@Tx[1]);
+
+	foreach (@insns) { eval; }
+  $Xi++;
+}
+
+sub Xtail_ssse3()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	foreach (@insns) { eval; }
+}
+
+sub body_00_19 () {
+	(
+	'($a,$b,$c,$d,$e)=@V;'.
+	'&add	($e,eval(4*($j&15))."(%rsp)");',	# X[]+K xfer
+	'&xor	($c,$d);',
+	'&mov	(@T[1],$a);',	# $b in next round
+	'&$_rol	($a,5);',
+	'&and	(@T[0],$c);',	# ($b&($c^$d))
+	'&xor	($c,$d);',	# restore $c
+	'&xor	(@T[0],$d);',
+	'&add	($e,$a);',
+	'&$_ror	($b,$j?7:2);',	# $b>>>2
+	'&add	($e,@T[0]);'	.'$j++; unshift(@V,pop(@V)); unshift(@T,pop(@T));'
+	);
+}
+
+sub body_20_39 () {
+	(
+	'($a,$b,$c,$d,$e)=@V;'.
+	'&add	($e,eval(4*($j++&15))."(%rsp)");',	# X[]+K xfer
+	'&xor	(@T[0],$d);',	# ($b^$d)
+	'&mov	(@T[1],$a);',	# $b in next round
+	'&$_rol	($a,5);',
+	'&xor	(@T[0],$c);',	# ($b^$d^$c)
+	'&add	($e,$a);',
+	'&$_ror	($b,7);',	# $b>>>2
+	'&add	($e,@T[0]);'	.'unshift(@V,pop(@V)); unshift(@T,pop(@T));'
+	);
+}
+
+sub body_40_59 () {
+	(
+	'($a,$b,$c,$d,$e)=@V;'.
+	'&mov	(@T[1],$c);',
+	'&xor	($c,$d);',
+	'&add	($e,eval(4*($j++&15))."(%rsp)");',	# X[]+K xfer
+	'&and	(@T[1],$d);',
+	'&and	(@T[0],$c);',	# ($b&($c^$d))
+	'&$_ror	($b,7);',	# $b>>>2
+	'&add	($e,@T[1]);',
+	'&mov	(@T[1],$a);',	# $b in next round
+	'&$_rol	($a,5);',
+	'&add	($e,@T[0]);',
+	'&xor	($c,$d);',	# restore $c
+	'&add	($e,$a);'	.'unshift(@V,pop(@V)); unshift(@T,pop(@T));'
+	);
+}
+$code.=<<___;
+.align	16
+.Loop_ssse3:
+___
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_16_31(\&body_00_19);
+	&Xupdate_ssse3_32_79(\&body_00_19);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_40_59);
+	&Xupdate_ssse3_32_79(\&body_20_39);
+	&Xuplast_ssse3_80(\&body_20_39);	# can jump to "done"
+
+				$saved_j=$j; @saved_V=@V;
+
+	&Xloop_ssse3(\&body_20_39);
+	&Xloop_ssse3(\&body_20_39);
+	&Xloop_ssse3(\&body_20_39);
+
+$code.=<<___;
+	add	0($ctx),$A			# update context
+	add	4($ctx),@T[0]
+	add	8($ctx),$C
+	add	12($ctx),$D
+	mov	$A,0($ctx)
+	add	16($ctx),$E
+	mov	@T[0],4($ctx)
+	mov	@T[0],$B			# magic seed
+	mov	$C,8($ctx)
+	mov	$D,12($ctx)
+	mov	$E,16($ctx)
+	jmp	.Loop_ssse3
+
+.align	16
+.Ldone_ssse3:
+___
+				$j=$saved_j; @V=@saved_V;
+
+	&Xtail_ssse3(\&body_20_39);
+	&Xtail_ssse3(\&body_20_39);
+	&Xtail_ssse3(\&body_20_39);
+
+$code.=<<___;
+	add	0($ctx),$A			# update context
+	add	4($ctx),@T[0]
+	add	8($ctx),$C
+	mov	$A,0($ctx)
+	add	12($ctx),$D
+	mov	@T[0],4($ctx)
+	add	16($ctx),$E
+	mov	$C,8($ctx)
+	mov	$D,12($ctx)
+	mov	$E,16($ctx)
+___
+$code.=<<___ if ($win64);
+	movaps	64+0(%rsp),%xmm6
+	movaps	64+16(%rsp),%xmm7
+	movaps	64+32(%rsp),%xmm8
+	movaps	64+48(%rsp),%xmm9
+	movaps	64+64(%rsp),%xmm10
+___
+$code.=<<___;
+	lea	`64+($win64?5*16:0)`(%rsp),%rsi
+	mov	0(%rsi),%r12
+	mov	8(%rsi),%rbp
+	mov	16(%rsi),%rbx
+	lea	24(%rsi),%rsp
+.Lepilogue_ssse3:
+	ret
+.size	sha1_block_data_order_ssse3,.-sha1_block_data_order_ssse3
+___
+
+if ($avx) {
+my $Xi=4;
+my @X=map("%xmm$_",(4..7,0..3));
+my @Tx=map("%xmm$_",(8..10));
+my @V=($A,$B,$C,$D,$E)=("%eax","%ebx","%ecx","%edx","%ebp");	# size optimization
+my @T=("%esi","%edi");
+my $j=0;
+my $K_XX_XX="%r11";
+
+my $_rol=sub { &shld(@_[0],@_) };
+my $_ror=sub { &shrd(@_[0],@_) };
+
+$code.=<<___;
+.type	sha1_block_data_order_avx,\@function,3
+.align	16
+sha1_block_data_order_avx:
+_avx_shortcut:
+	push	%rbx
+	push	%rbp
+	push	%r12
+	lea	`-64-($win64?5*16:0)`(%rsp),%rsp
+___
+$code.=<<___ if ($win64);
+	movaps	%xmm6,64+0(%rsp)
+	movaps	%xmm7,64+16(%rsp)
+	movaps	%xmm8,64+32(%rsp)
+	movaps	%xmm9,64+48(%rsp)
+	movaps	%xmm10,64+64(%rsp)
+.Lprologue_avx:
+___
+$code.=<<___;
+	mov	%rdi,$ctx	# reassigned argument
+	mov	%rsi,$inp	# reassigned argument
+	mov	%rdx,$num	# reassigned argument
+	vzeroupper
+
+	shl	\$6,$num
+	add	$inp,$num
+	lea	K_XX_XX(%rip),$K_XX_XX
+
+	mov	0($ctx),$A		# load context
+	mov	4($ctx),$B
+	mov	8($ctx),$C
+	mov	12($ctx),$D
+	mov	$B,@T[0]		# magic seed
+	mov	16($ctx),$E
+
+	vmovdqa	64($K_XX_XX),@X[2]	# pbswap mask
+	vmovdqa	0($K_XX_XX),@Tx[1]	# K_00_19
+	vmovdqu	0($inp),@X[-4&7]	# load input to %xmm[0-3]
+	vmovdqu	16($inp),@X[-3&7]
+	vmovdqu	32($inp),@X[-2&7]
+	vmovdqu	48($inp),@X[-1&7]
+	vpshufb	@X[2],@X[-4&7],@X[-4&7]	# byte swap
+	add	\$64,$inp
+	vpshufb	@X[2],@X[-3&7],@X[-3&7]
+	vpshufb	@X[2],@X[-2&7],@X[-2&7]
+	vpshufb	@X[2],@X[-1&7],@X[-1&7]
+	vpaddd	@Tx[1],@X[-4&7],@X[0]	# add K_00_19
+	vpaddd	@Tx[1],@X[-3&7],@X[1]
+	vpaddd	@Tx[1],@X[-2&7],@X[2]
+	vmovdqa	@X[0],0(%rsp)		# X[]+K xfer to IALU
+	vmovdqa	@X[1],16(%rsp)
+	vmovdqa	@X[2],32(%rsp)
+	jmp	.Loop_avx
+___
+
+sub Xupdate_avx_16_31()		# recall that $Xi starts wtih 4
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 40 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpalignr(@X[0],@X[-3&7],@X[-4&7],8);	# compose "X[-14]" in "X[0]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &vpaddd	(@Tx[1],@Tx[1],@X[-1&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpsrldq(@Tx[0],@X[-1&7],4);	# "X[-3]", 3 dwords
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpxor	(@X[0],@X[0],@X[-4&7]);		# "X[0]"^="X[-16]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpxor	(@Tx[0],@Tx[0],@X[-2&7]);	# "X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpxor	(@X[0],@X[0],@Tx[0]);		# "X[0]"^="X[-3]"^"X[-8]"
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vmovdqa	(eval(16*(($Xi-1)&3))."(%rsp)",@Tx[1]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpsrld	(@Tx[0],@X[0],31);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpslldq(@Tx[2],@X[0],12);		# "X[0]"<<96, extract one dword
+	&vpaddd	(@X[0],@X[0],@X[0]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpsrld	(@Tx[1],@Tx[2],30);
+	&vpor	(@X[0],@X[0],@Tx[0]);		# "X[0]"<<<=1
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpslld	(@Tx[2],@Tx[2],2);
+	&vpxor	(@X[0],@X[0],@Tx[1]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	&vpxor	(@X[0],@X[0],@Tx[2]);		# "X[0]"^=("X[0]">>96)<<<2
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	  &vmovdqa	(@Tx[2],eval(16*(($Xi)/5))."($K_XX_XX)");	# K_XX_XX
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+
+	 foreach (@insns) { eval; }	# remaining instructions [if any]
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+		push(@Tx,shift(@Tx));
+}
+
+sub Xupdate_avx_32_79()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 to 48 instructions
+  my ($a,$b,$c,$d,$e);
+
+	&vpalignr(@Tx[0],@X[-1&7],@X[-2&7],8);	# compose "X[-6]"
+	&vpxor	(@X[0],@X[0],@X[-4&7]);		# "X[0]"="X[-32]"^"X[-16]"
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&vpxor	(@X[0],@X[0],@X[-7&7]);		# "X[0]"^="X[-28]"
+	 eval(shift(@insns));
+	 eval(shift(@insns))	if (@insns[0] !~ /&ro[rl]/);
+	if ($Xi%5) {
+	  &vmovdqa	(@Tx[2],@Tx[1]);# "perpetuate" K_XX_XX...
+	} else {			# ... or load next one
+	  &vmovdqa	(@Tx[2],eval(16*($Xi/5))."($K_XX_XX)");
+	}
+	  &vpaddd	(@Tx[1],@Tx[1],@X[-1&7]);
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&vpxor	(@X[0],@X[0],@Tx[0]);		# "X[0]"^="X[-6]"
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+
+	&vpsrld	(@Tx[0],@X[0],30);
+	  &vmovdqa	(eval(16*(($Xi-1)&3))."(%rsp)",@Tx[1]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&vpslld	(@X[0],@X[0],2);
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# ror
+	 eval(shift(@insns));
+
+	&vpor	(@X[0],@X[0],@Tx[0]);		# "X[0]"<<<=2
+	 eval(shift(@insns));		# body_20_39
+	 eval(shift(@insns));
+	  &vmovdqa	(@Tx[1],@X[0])	if ($Xi<19);
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));		# rol
+	 eval(shift(@insns));
+
+	 foreach (@insns) { eval; }	# remaining instructions
+
+  $Xi++;	push(@X,shift(@X));	# "rotate" X[]
+		push(@Tx,shift(@Tx));
+}
+
+sub Xuplast_avx_80()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	  &vpaddd	(@Tx[1],@Tx[1],@X[-1&7]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	  &movdqa	(eval(16*(($Xi-1)&3))."(%rsp)",@Tx[1]);	# X[]+K xfer IALU
+
+	 foreach (@insns) { eval; }		# remaining instructions
+
+	&cmp	($inp,$num);
+	&je	(".Ldone_avx");
+
+	unshift(@Tx,pop(@Tx));
+
+	&vmovdqa(@X[2],"64($K_XX_XX)");		# pbswap mask
+	&vmovdqa(@Tx[1],"0($K_XX_XX)");		# K_00_19
+	&vmovdqu(@X[-4&7],"0($inp)");		# load input
+	&vmovdqu(@X[-3&7],"16($inp)");
+	&vmovdqu(@X[-2&7],"32($inp)");
+	&vmovdqu(@X[-1&7],"48($inp)");
+	&vpshufb(@X[-4&7],@X[-4&7],@X[2]);	# byte swap
+	&add	($inp,64);
+
+  $Xi=0;
+}
+
+sub Xloop_avx()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpshufb(@X[($Xi-3)&7],@X[($Xi-3)&7],@X[2]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vpaddd	(@X[$Xi&7],@X[($Xi-4)&7],@Tx[1]);
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+	&vmovdqa(eval(16*$Xi)."(%rsp)",@X[$Xi&7]);	# X[]+K xfer to IALU
+	 eval(shift(@insns));
+	 eval(shift(@insns));
+
+	foreach (@insns) { eval; }
+  $Xi++;
+}
+
+sub Xtail_avx()
+{ use integer;
+  my $body = shift;
+  my @insns = (&$body,&$body,&$body,&$body);	# 32 instructions
+  my ($a,$b,$c,$d,$e);
+
+	foreach (@insns) { eval; }
+}
+
+$code.=<<___;
+.align	16
+.Loop_avx:
+___
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_16_31(\&body_00_19);
+	&Xupdate_avx_32_79(\&body_00_19);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_40_59);
+	&Xupdate_avx_32_79(\&body_20_39);
+	&Xuplast_avx_80(\&body_20_39);	# can jump to "done"
+
+				$saved_j=$j; @saved_V=@V;
+
+	&Xloop_avx(\&body_20_39);
+	&Xloop_avx(\&body_20_39);
+	&Xloop_avx(\&body_20_39);
+
+$code.=<<___;
+	add	0($ctx),$A			# update context
+	add	4($ctx),@T[0]
+	add	8($ctx),$C
+	add	12($ctx),$D
+	mov	$A,0($ctx)
+	add	16($ctx),$E
+	mov	@T[0],4($ctx)
+	mov	@T[0],$B			# magic seed
+	mov	$C,8($ctx)
+	mov	$D,12($ctx)
+	mov	$E,16($ctx)
+	jmp	.Loop_avx
+
+.align	16
+.Ldone_avx:
+___
+				$j=$saved_j; @V=@saved_V;
+
+	&Xtail_avx(\&body_20_39);
+	&Xtail_avx(\&body_20_39);
+	&Xtail_avx(\&body_20_39);
+
+$code.=<<___;
+	vzeroupper
+
+	add	0($ctx),$A			# update context
+	add	4($ctx),@T[0]
+	add	8($ctx),$C
+	mov	$A,0($ctx)
+	add	12($ctx),$D
+	mov	@T[0],4($ctx)
+	add	16($ctx),$E
+	mov	$C,8($ctx)
+	mov	$D,12($ctx)
+	mov	$E,16($ctx)
+___
+$code.=<<___ if ($win64);
+	movaps	64+0(%rsp),%xmm6
+	movaps	64+16(%rsp),%xmm7
+	movaps	64+32(%rsp),%xmm8
+	movaps	64+48(%rsp),%xmm9
+	movaps	64+64(%rsp),%xmm10
+___
+$code.=<<___;
+	lea	`64+($win64?5*16:0)`(%rsp),%rsi
+	mov	0(%rsi),%r12
+	mov	8(%rsi),%rbp
+	mov	16(%rsi),%rbx
+	lea	24(%rsi),%rsp
+.Lepilogue_avx:
+	ret
+.size	sha1_block_data_order_avx,.-sha1_block_data_order_avx
+___
+}
+$code.=<<___;
+.align	64
+K_XX_XX:
+.long	0x5a827999,0x5a827999,0x5a827999,0x5a827999	# K_00_19
+.long	0x6ed9eba1,0x6ed9eba1,0x6ed9eba1,0x6ed9eba1	# K_20_39
+.long	0x8f1bbcdc,0x8f1bbcdc,0x8f1bbcdc,0x8f1bbcdc	# K_40_59
+.long	0xca62c1d6,0xca62c1d6,0xca62c1d6,0xca62c1d6	# K_60_79
+.long	0x00010203,0x04050607,0x08090a0b,0x0c0d0e0f	# pbswap mask
+___
+}}}
+$code.=<<___;
+.asciz	"SHA1 block transform for x86_64, CRYPTOGAMS by <appro\@openssl.org>"
+.align	64
+___
+
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#		CONTEXT *context,DISPATCHER_CONTEXT *disp)
+if ($win64) {
+$rec="%rcx";
+$frame="%rdx";
+$context="%r8";
+$disp="%r9";
+
+$code.=<<___;
+.extern	__imp_RtlVirtualUnwind
+.type	se_handler,\@abi-omnipotent
+.align	16
+se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	lea	.Lprologue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<.Lprologue
+	jb	.Lcommon_seh_tail
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	lea	.Lepilogue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip>=.Lepilogue
+	jae	.Lcommon_seh_tail
+
+	mov	`16*4`(%rax),%rax	# pull saved stack pointer
+	lea	32(%rax),%rax
+
+	mov	-8(%rax),%rbx
+	mov	-16(%rax),%rbp
+	mov	-24(%rax),%r12
+	mov	-32(%rax),%r13
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore context->R12
+	mov	%r13,224($context)	# restore context->R13
+
+	jmp	.Lcommon_seh_tail
+.size	se_handler,.-se_handler
+
+.type	ssse3_handler,\@abi-omnipotent
+.align	16
+ssse3_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	mov	8($disp),%rsi		# disp->ImageBase
+	mov	56($disp),%r11		# disp->HandlerData
+
+	mov	0(%r11),%r10d		# HandlerData[0]
+	lea	(%rsi,%r10),%r10	# prologue label
+	cmp	%r10,%rbx		# context->Rip<prologue label
+	jb	.Lcommon_seh_tail
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	mov	4(%r11),%r10d		# HandlerData[1]
+	lea	(%rsi,%r10),%r10	# epilogue label
+	cmp	%r10,%rbx		# context->Rip>=epilogue label
+	jae	.Lcommon_seh_tail
+
+	lea	64(%rax),%rsi
+	lea	512($context),%rdi	# &context.Xmm6
+	mov	\$10,%ecx
+	.long	0xa548f3fc		# cld; rep movsq
+	lea	`24+64+5*16`(%rax),%rax	# adjust stack pointer
+
+	mov	-8(%rax),%rbx
+	mov	-16(%rax),%rbp
+	mov	-24(%rax),%r12
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore cotnext->R12
+
+.Lcommon_seh_tail:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+	mov	40($disp),%rdi		# disp->ContextRecord
+	mov	$context,%rsi		# context
+	mov	\$154,%ecx		# sizeof(CONTEXT)
+	.long	0xa548f3fc		# cld; rep movsq
+
+	mov	$disp,%rsi
+	xor	%rcx,%rcx		# arg1, UNW_FLAG_NHANDLER
+	mov	8(%rsi),%rdx		# arg2, disp->ImageBase
+	mov	0(%rsi),%r8		# arg3, disp->ControlPc
+	mov	16(%rsi),%r9		# arg4, disp->FunctionEntry
+	mov	40(%rsi),%r10		# disp->ContextRecord
+	lea	56(%rsi),%r11		# &disp->HandlerData
+	lea	24(%rsi),%r12		# &disp->EstablisherFrame
+	mov	%r10,32(%rsp)		# arg5
+	mov	%r11,40(%rsp)		# arg6
+	mov	%r12,48(%rsp)		# arg7
+	mov	%rcx,56(%rsp)		# arg8, (NULL)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	mov	\$1,%eax		# ExceptionContinueSearch
+	add	\$64,%rsp
+	popfq
+	pop	%r15
+	pop	%r14
+	pop	%r13
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+	pop	%rdi
+	pop	%rsi
+	ret
+.size	ssse3_handler,.-ssse3_handler
+
+.section	.pdata
+.align	4
+	.rva	.LSEH_begin_sha1_block_data_order
+	.rva	.LSEH_end_sha1_block_data_order
+	.rva	.LSEH_info_sha1_block_data_order
+	.rva	.LSEH_begin_sha1_block_data_order_ssse3
+	.rva	.LSEH_end_sha1_block_data_order_ssse3
+	.rva	.LSEH_info_sha1_block_data_order_ssse3
+___
+$code.=<<___ if ($avx);
+	.rva	.LSEH_begin_sha1_block_data_order_avx
+	.rva	.LSEH_end_sha1_block_data_order_avx
+	.rva	.LSEH_info_sha1_block_data_order_avx
+___
+$code.=<<___;
+.section	.xdata
+.align	8
+.LSEH_info_sha1_block_data_order:
+	.byte	9,0,0,0
+	.rva	se_handler
+.LSEH_info_sha1_block_data_order_ssse3:
+	.byte	9,0,0,0
+	.rva	ssse3_handler
+	.rva	.Lprologue_ssse3,.Lepilogue_ssse3	# HandlerData[]
+___
+$code.=<<___ if ($avx);
+.LSEH_info_sha1_block_data_order_avx:
+	.byte	9,0,0,0
+	.rva	ssse3_handler
+	.rva	.Lprologue_avx,.Lepilogue_avx		# HandlerData[]
+___
+}
+
+####################################################################
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha256-586.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha256-586.pl
new file mode 100644
index 0000000000..928ec53123
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha256-586.pl
@@ -0,0 +1,249 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> 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 block transform for x86. September 2007.
+#
+# Performance in clock cycles per processed byte (less is better):
+#
+#		Pentium	PIII	P4	AMD K8	Core2
+# gcc		46	36	41	27	26
+# icc		57	33	38	25	23	
+# x86 asm	40	30	33	20	18
+# x86_64 asm(*)	-	-	21	16	16
+#
+# (*) x86_64 assembler performance is presented for reference
+#     purposes.
+#
+# Performance improvement over compiler generated code varies from
+# 10% to 40% [see above]. Not very impressive on some µ-archs, but
+# it's 5 times smaller and optimizies amount of writes.
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
+require "x86asm.pl";
+
+&asm_init($ARGV[0],"sha512-586.pl",$ARGV[$#ARGV] eq "386");
+
+$A="eax";
+$E="edx";
+$T="ebx";
+$Aoff=&DWP(0,"esp");
+$Boff=&DWP(4,"esp");
+$Coff=&DWP(8,"esp");
+$Doff=&DWP(12,"esp");
+$Eoff=&DWP(16,"esp");
+$Foff=&DWP(20,"esp");
+$Goff=&DWP(24,"esp");
+$Hoff=&DWP(28,"esp");
+$Xoff=&DWP(32,"esp");
+$K256="ebp";
+
+sub BODY_00_15() {
+    my $in_16_63=shift;
+
+	&mov	("ecx",$E);
+	 &add	($T,"edi")			if ($in_16_63);	# T += sigma1(X[-2])
+	&ror	("ecx",25-11);
+	 &mov	("esi",$Foff);
+	&xor	("ecx",$E);
+	&ror	("ecx",11-6);
+	 &mov	(&DWP(4*(8+15),"esp"),$T)	if ($in_16_63);	# save X[0]
+	&xor	("ecx",$E);
+	&ror	("ecx",6);	# Sigma1(e)
+	 &mov	("edi",$Goff);
+	&add	($T,"ecx");	# T += Sigma1(e)
+
+	&xor	("esi","edi");
+	 &mov	($Eoff,$E);	# modulo-scheduled
+	 &mov	("ecx",$A);
+	&and	("esi",$E);
+	 &mov	($E,$Doff);	# e becomes d, which is e in next iteration
+	&xor	("esi","edi");	# Ch(e,f,g)
+	 &mov	("edi",$A);
+	&add	($T,"esi");	# T += Ch(e,f,g)
+
+	&ror	("ecx",22-13);
+	 &add	($T,$Hoff);	# T += h
+	&xor	("ecx",$A);
+	&ror	("ecx",13-2);
+	 &mov	("esi",$Boff);
+	&xor	("ecx",$A);
+	&ror	("ecx",2);	# Sigma0(a)
+	 &add	($E,$T);	# d += T
+	 &mov	("edi",$Coff);
+
+	&add	($T,"ecx");	# T += Sigma0(a)
+	 &mov	($Aoff,$A);	# modulo-scheduled
+
+	&mov	("ecx",$A);
+	 &sub	("esp",4);
+	&or	($A,"esi");	# a becomes h, which is a in next iteration
+	&and	("ecx","esi");
+	&and	($A,"edi");
+	 &mov	("esi",&DWP(0,$K256));
+	&or	($A,"ecx");	# h=Maj(a,b,c)
+
+	&add	($K256,4);
+	&add	($A,$T);	# h += T
+	 &mov	($T,&DWP(4*(8+15+16-1),"esp"))	if ($in_16_63);	# preload T
+	&add	($E,"esi");	# d += K256[i]
+	&add	($A,"esi");	# h += K256[i]
+}
+
+&function_begin("sha256_block_data_order");
+	&mov	("esi",wparam(0));	# ctx
+	&mov	("edi",wparam(1));	# inp
+	&mov	("eax",wparam(2));	# num
+	&mov	("ebx","esp");		# saved sp
+
+	&call	(&label("pic_point"));	# make it PIC!
+&set_label("pic_point");
+	&blindpop($K256);
+	&lea	($K256,&DWP(&label("K256")."-".&label("pic_point"),$K256));
+
+	&sub	("esp",16);
+	&and	("esp",-64);
+
+	&shl	("eax",6);
+	&add	("eax","edi");
+	&mov	(&DWP(0,"esp"),"esi");	# ctx
+	&mov	(&DWP(4,"esp"),"edi");	# inp
+	&mov	(&DWP(8,"esp"),"eax");	# inp+num*128
+	&mov	(&DWP(12,"esp"),"ebx");	# saved sp
+
+&set_label("loop",16);
+    # copy input block to stack reversing byte and dword order
+    for($i=0;$i<4;$i++) {
+	&mov	("eax",&DWP($i*16+0,"edi"));
+	&mov	("ebx",&DWP($i*16+4,"edi"));
+	&mov	("ecx",&DWP($i*16+8,"edi"));
+	&mov	("edx",&DWP($i*16+12,"edi"));
+	&bswap	("eax");
+	&bswap	("ebx");
+	&bswap	("ecx");
+	&bswap	("edx");
+	&push	("eax");
+	&push	("ebx");
+	&push	("ecx");
+	&push	("edx");
+    }
+	&add	("edi",64);
+	&sub	("esp",4*8);		# place for A,B,C,D,E,F,G,H
+	&mov	(&DWP(4*(8+16)+4,"esp"),"edi");
+
+	# copy ctx->h[0-7] to A,B,C,D,E,F,G,H on stack
+	&mov	($A,&DWP(0,"esi"));
+	&mov	("ebx",&DWP(4,"esi"));
+	&mov	("ecx",&DWP(8,"esi"));
+	&mov	("edi",&DWP(12,"esi"));
+	# &mov	($Aoff,$A);
+	&mov	($Boff,"ebx");
+	&mov	($Coff,"ecx");
+	&mov	($Doff,"edi");
+	&mov	($E,&DWP(16,"esi"));	
+	&mov	("ebx",&DWP(20,"esi"));
+	&mov	("ecx",&DWP(24,"esi"));
+	&mov	("edi",&DWP(28,"esi"));
+	# &mov	($Eoff,$E);
+	&mov	($Foff,"ebx");
+	&mov	($Goff,"ecx");
+	&mov	($Hoff,"edi");
+
+&set_label("00_15",16);
+	&mov	($T,&DWP(4*(8+15),"esp"));
+
+	&BODY_00_15();
+
+	&cmp	("esi",0xc19bf174);
+	&jne	(&label("00_15"));
+
+	&mov	($T,&DWP(4*(8+15+16-1),"esp"));	# preloaded in BODY_00_15(1)
+&set_label("16_63",16);
+	&mov	("esi",$T);
+	 &mov	("ecx",&DWP(4*(8+15+16-14),"esp"));
+	&ror	("esi",18-7);
+	 &mov	("edi","ecx");
+	&xor	("esi",$T);
+	&ror	("esi",7);
+	&shr	($T,3);
+
+	&ror	("edi",19-17);
+	 &xor	($T,"esi");			# T = sigma0(X[-15])
+	&xor	("edi","ecx");
+	&ror	("edi",17);
+	&shr	("ecx",10);
+	 &add	($T,&DWP(4*(8+15+16),"esp"));	# T += X[-16]
+	&xor	("edi","ecx");			# sigma1(X[-2])
+
+	 &add	($T,&DWP(4*(8+15+16-9),"esp"));	# T += X[-7]
+	# &add	($T,"edi");			# T += sigma1(X[-2])
+	# &mov	(&DWP(4*(8+15),"esp"),$T);	# save X[0]
+
+	&BODY_00_15(1);
+
+	&cmp	("esi",0xc67178f2);
+	&jne	(&label("16_63"));
+
+	&mov	("esi",&DWP(4*(8+16+64)+0,"esp"));#ctx
+	# &mov	($A,$Aoff);
+	&mov	("ebx",$Boff);
+	&mov	("ecx",$Coff);
+	&mov	("edi",$Doff);
+	&add	($A,&DWP(0,"esi"));
+	&add	("ebx",&DWP(4,"esi"));
+	&add	("ecx",&DWP(8,"esi"));
+	&add	("edi",&DWP(12,"esi"));
+	&mov	(&DWP(0,"esi"),$A);
+	&mov	(&DWP(4,"esi"),"ebx");
+	&mov	(&DWP(8,"esi"),"ecx");
+	&mov	(&DWP(12,"esi"),"edi");
+	# &mov	($E,$Eoff);
+	&mov	("eax",$Foff);
+	&mov	("ebx",$Goff);
+	&mov	("ecx",$Hoff);
+	&mov	("edi",&DWP(4*(8+16+64)+4,"esp"));#inp
+	&add	($E,&DWP(16,"esi"));
+	&add	("eax",&DWP(20,"esi"));
+	&add	("ebx",&DWP(24,"esi"));
+	&add	("ecx",&DWP(28,"esi"));
+	&mov	(&DWP(16,"esi"),$E);
+	&mov	(&DWP(20,"esi"),"eax");
+	&mov	(&DWP(24,"esi"),"ebx");
+	&mov	(&DWP(28,"esi"),"ecx");
+
+	&add	("esp",4*(8+16+64));		# destroy frame
+	&sub	($K256,4*64);			# rewind K
+
+	&cmp	("edi",&DWP(8,"esp"));		# are we done yet?
+	&jb	(&label("loop"));
+
+	&mov	("esp",&DWP(12,"esp"));		# restore sp
+&function_end_A();
+
+&set_label("K256",64);	# Yes! I keep it in the code segment!
+	&data_word(0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5);
+	&data_word(0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5);
+	&data_word(0xd807aa98,0x12835b01,0x243185be,0x550c7dc3);
+	&data_word(0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174);
+	&data_word(0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc);
+	&data_word(0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da);
+	&data_word(0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7);
+	&data_word(0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967);
+	&data_word(0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13);
+	&data_word(0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85);
+	&data_word(0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3);
+	&data_word(0xd192e819,0xd6990624,0xf40e3585,0x106aa070);
+	&data_word(0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5);
+	&data_word(0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3);
+	&data_word(0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208);
+	&data_word(0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2);
+&function_end_B("sha256_block_data_order");
+&asciz("SHA256 block transform for x86, CRYPTOGAMS by <appro\@openssl.org>");
+
+&asm_finish();
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha256-armv4.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha256-armv4.pl
new file mode 100644
index 0000000000..9c84e8d93c
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha256-armv4.pl
@@ -0,0 +1,211 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> 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 block procedure for ARMv4. May 2007.
+
+# Performance is ~2x better than gcc 3.4 generated code and in "abso-
+# lute" terms is ~2250 cycles per 64-byte block or ~35 cycles per
+# byte [on single-issue Xscale PXA250 core].
+
+# July 2010.
+#
+# Rescheduling for dual-issue pipeline resulted in 22% improvement on
+# Cortex A8 core and ~20 cycles per processed byte.
+
+# February 2011.
+#
+# Profiler-assisted and platform-specific optimization resulted in 16%
+# improvement on Cortex A8 core and ~17 cycles per processed byte.
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+$ctx="r0";	$t0="r0";
+$inp="r1";	$t3="r1";
+$len="r2";	$t1="r2";
+$T1="r3";
+$A="r4";
+$B="r5";
+$C="r6";
+$D="r7";
+$E="r8";
+$F="r9";
+$G="r10";
+$H="r11";
+@V=($A,$B,$C,$D,$E,$F,$G,$H);
+$t2="r12";
+$Ktbl="r14";
+
+@Sigma0=( 2,13,22);
+@Sigma1=( 6,11,25);
+@sigma0=( 7,18, 3);
+@sigma1=(17,19,10);
+
+sub BODY_00_15 {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h) = @_;
+
+$code.=<<___ if ($i<16);
+#if __ARM_ARCH__>=7
+	ldr	$T1,[$inp],#4
+#else
+	ldrb	$T1,[$inp,#3]			@ $i
+	ldrb	$t2,[$inp,#2]
+	ldrb	$t1,[$inp,#1]
+	ldrb	$t0,[$inp],#4
+	orr	$T1,$T1,$t2,lsl#8
+	orr	$T1,$T1,$t1,lsl#16
+	orr	$T1,$T1,$t0,lsl#24
+#endif
+___
+$code.=<<___;
+	mov	$t0,$e,ror#$Sigma1[0]
+	ldr	$t2,[$Ktbl],#4			@ *K256++
+	eor	$t0,$t0,$e,ror#$Sigma1[1]
+	eor	$t1,$f,$g
+#if $i>=16
+	add	$T1,$T1,$t3			@ from BODY_16_xx
+#elif __ARM_ARCH__>=7 && defined(__ARMEL__)
+	rev	$T1,$T1
+#endif
+#if $i==15
+	str	$inp,[sp,#17*4]			@ leave room for $t3
+#endif
+	eor	$t0,$t0,$e,ror#$Sigma1[2]	@ Sigma1(e)
+	and	$t1,$t1,$e
+	str	$T1,[sp,#`$i%16`*4]
+	add	$T1,$T1,$t0
+	eor	$t1,$t1,$g			@ Ch(e,f,g)
+	add	$T1,$T1,$h
+	mov	$h,$a,ror#$Sigma0[0]
+	add	$T1,$T1,$t1
+	eor	$h,$h,$a,ror#$Sigma0[1]
+	add	$T1,$T1,$t2
+	eor	$h,$h,$a,ror#$Sigma0[2]		@ Sigma0(a)
+#if $i>=15
+	ldr	$t3,[sp,#`($i+2)%16`*4]		@ from BODY_16_xx
+#endif
+	orr	$t0,$a,$b
+	and	$t1,$a,$b
+	and	$t0,$t0,$c
+	add	$h,$h,$T1
+	orr	$t0,$t0,$t1			@ Maj(a,b,c)
+	add	$d,$d,$T1
+	add	$h,$h,$t0
+___
+}
+
+sub BODY_16_XX {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h) = @_;
+
+$code.=<<___;
+	@ ldr	$t3,[sp,#`($i+1)%16`*4]		@ $i
+	ldr	$t2,[sp,#`($i+14)%16`*4]
+	mov	$t0,$t3,ror#$sigma0[0]
+	ldr	$T1,[sp,#`($i+0)%16`*4]
+	eor	$t0,$t0,$t3,ror#$sigma0[1]
+	ldr	$t1,[sp,#`($i+9)%16`*4]
+	eor	$t0,$t0,$t3,lsr#$sigma0[2]	@ sigma0(X[i+1])
+	mov	$t3,$t2,ror#$sigma1[0]
+	add	$T1,$T1,$t0
+	eor	$t3,$t3,$t2,ror#$sigma1[1]
+	add	$T1,$T1,$t1
+	eor	$t3,$t3,$t2,lsr#$sigma1[2]	@ sigma1(X[i+14])
+	@ add	$T1,$T1,$t3
+___
+	&BODY_00_15(@_);
+}
+
+$code=<<___;
+#include "arm_arch.h"
+
+.text
+.code	32
+
+.type	K256,%object
+.align	5
+K256:
+.word	0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5
+.word	0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5
+.word	0xd807aa98,0x12835b01,0x243185be,0x550c7dc3
+.word	0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174
+.word	0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc
+.word	0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da
+.word	0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7
+.word	0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967
+.word	0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13
+.word	0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85
+.word	0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3
+.word	0xd192e819,0xd6990624,0xf40e3585,0x106aa070
+.word	0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5
+.word	0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3
+.word	0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208
+.word	0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
+.size	K256,.-K256
+
+.global	sha256_block_data_order
+.type	sha256_block_data_order,%function
+sha256_block_data_order:
+	sub	r3,pc,#8		@ sha256_block_data_order
+	add	$len,$inp,$len,lsl#6	@ len to point at the end of inp
+	stmdb	sp!,{$ctx,$inp,$len,r4-r11,lr}
+	ldmia	$ctx,{$A,$B,$C,$D,$E,$F,$G,$H}
+	sub	$Ktbl,r3,#256		@ K256
+	sub	sp,sp,#16*4		@ alloca(X[16])
+.Loop:
+___
+for($i=0;$i<16;$i++)	{ &BODY_00_15($i,@V); unshift(@V,pop(@V)); }
+$code.=".Lrounds_16_xx:\n";
+for (;$i<32;$i++)	{ &BODY_16_XX($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	and	$t2,$t2,#0xff
+	cmp	$t2,#0xf2
+	bne	.Lrounds_16_xx
+
+	ldr	$T1,[sp,#16*4]		@ pull ctx
+	ldr	$t0,[$T1,#0]
+	ldr	$t1,[$T1,#4]
+	ldr	$t2,[$T1,#8]
+	add	$A,$A,$t0
+	ldr	$t0,[$T1,#12]
+	add	$B,$B,$t1
+	ldr	$t1,[$T1,#16]
+	add	$C,$C,$t2
+	ldr	$t2,[$T1,#20]
+	add	$D,$D,$t0
+	ldr	$t0,[$T1,#24]
+	add	$E,$E,$t1
+	ldr	$t1,[$T1,#28]
+	add	$F,$F,$t2
+	ldr	$inp,[sp,#17*4]		@ pull inp
+	ldr	$t2,[sp,#18*4]		@ pull inp+len
+	add	$G,$G,$t0
+	add	$H,$H,$t1
+	stmia	$T1,{$A,$B,$C,$D,$E,$F,$G,$H}
+	cmp	$inp,$t2
+	sub	$Ktbl,$Ktbl,#256	@ rewind Ktbl
+	bne	.Loop
+
+	add	sp,sp,#`16+3`*4	@ destroy frame
+#if __ARM_ARCH__>=5
+	ldmia	sp!,{r4-r11,pc}
+#else
+	ldmia	sp!,{r4-r11,lr}
+	tst	lr,#1
+	moveq	pc,lr			@ be binary compatible with V4, yet
+	bx	lr			@ interoperable with Thumb ISA:-)
+#endif
+.size   sha256_block_data_order,.-sha256_block_data_order
+.asciz  "SHA256 block transform for ARMv4, CRYPTOGAMS by <appro\@openssl.org>"
+.align	2
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+$code =~ s/\bbx\s+lr\b/.word\t0xe12fff1e/gm;	# make it possible to compile with -march=armv4
+print $code;
+close STDOUT; # enforce flush
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha512-586.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha512-586.pl
new file mode 100644
index 0000000000..7eab6a5b88
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha512-586.pl
@@ -0,0 +1,644 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> 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 block transform for x86. September 2007.
+#
+# Performance in clock cycles per processed byte (less is better):
+#
+#		Pentium	PIII	P4	AMD K8	Core2
+# gcc		100	75	116	54	66
+# icc		97	77	95	55	57
+# x86 asm	61	56	82	36	40
+# SSE2 asm	-	-	38	24	20
+# x86_64 asm(*)	-	-	30	10.0	10.5
+#
+# (*) x86_64 assembler performance is presented for reference
+#     purposes.
+#
+# 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%
+# 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
+# though it does not use 128-bit operations. The latter means that
+# SSE2-aware kernel is no longer required to execute the code. Another
+# difference is that new code optimizes amount of writes, but at the
+# cost of increased data cache "footprint" by 1/2KB.
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
+require "x86asm.pl";
+
+&asm_init($ARGV[0],"sha512-586.pl",$ARGV[$#ARGV] eq "386");
+
+$sse2=0;
+for (@ARGV) { $sse2=1 if (/-DOPENSSL_IA32_SSE2/); }
+
+&external_label("OPENSSL_ia32cap_P") if ($sse2);
+
+$Tlo=&DWP(0,"esp");	$Thi=&DWP(4,"esp");
+$Alo=&DWP(8,"esp");	$Ahi=&DWP(8+4,"esp");
+$Blo=&DWP(16,"esp");	$Bhi=&DWP(16+4,"esp");
+$Clo=&DWP(24,"esp");	$Chi=&DWP(24+4,"esp");
+$Dlo=&DWP(32,"esp");	$Dhi=&DWP(32+4,"esp");
+$Elo=&DWP(40,"esp");	$Ehi=&DWP(40+4,"esp");
+$Flo=&DWP(48,"esp");	$Fhi=&DWP(48+4,"esp");
+$Glo=&DWP(56,"esp");	$Ghi=&DWP(56+4,"esp");
+$Hlo=&DWP(64,"esp");	$Hhi=&DWP(64+4,"esp");
+$K512="ebp";
+
+$Asse2=&QWP(0,"esp");
+$Bsse2=&QWP(8,"esp");
+$Csse2=&QWP(16,"esp");
+$Dsse2=&QWP(24,"esp");
+$Esse2=&QWP(32,"esp");
+$Fsse2=&QWP(40,"esp");
+$Gsse2=&QWP(48,"esp");
+$Hsse2=&QWP(56,"esp");
+
+$A="mm0";	# B-D and
+$E="mm4";	# F-H are commonly loaded to respectively mm1-mm3 and
+		# mm5-mm7, but it's done on on-demand basis...
+
+sub BODY_00_15_sse2 {
+    my $prefetch=shift;
+
+	&movq	("mm5",$Fsse2);			# load f
+	&movq	("mm6",$Gsse2);			# load g
+	&movq	("mm7",$Hsse2);			# load h
+
+	&movq	("mm1",$E);			# %mm1 is sliding right
+	&movq	("mm2",$E);			# %mm2 is sliding left
+	&psrlq	("mm1",14);
+	&movq	($Esse2,$E);			# modulo-scheduled save e
+	&psllq	("mm2",23);
+	&movq	("mm3","mm1");			# %mm3 is T1
+	&psrlq	("mm1",4);
+	&pxor	("mm3","mm2");
+	&psllq	("mm2",23);
+	&pxor	("mm3","mm1");
+	&psrlq	("mm1",23);
+	&pxor	("mm3","mm2");
+	&psllq	("mm2",4);
+	&pxor	("mm3","mm1");
+	&paddq	("mm7",QWP(0,$K512));		# h+=K512[i]
+	&pxor	("mm3","mm2");			# T1=Sigma1_512(e)
+
+	&pxor	("mm5","mm6");			# f^=g
+	&movq	("mm1",$Bsse2);			# load b
+	&pand	("mm5",$E);			# f&=e
+	&movq	("mm2",$Csse2);			# load c
+	&pxor	("mm5","mm6");			# f^=g
+	&movq	($E,$Dsse2);			# e = load d
+	&paddq	("mm3","mm5");			# T1+=Ch(e,f,g)
+	&movq	(&QWP(0,"esp"),$A);		# modulo-scheduled save a
+	&paddq	("mm3","mm7");			# T1+=h
+
+	&movq	("mm5",$A);			# %mm5 is sliding right
+	&movq	("mm6",$A);			# %mm6 is sliding left
+	&paddq	("mm3",&QWP(8*9,"esp"));	# T1+=X[0]
+	&psrlq	("mm5",28);
+	&paddq	($E,"mm3");			# e += T1
+	&psllq	("mm6",25);
+	&movq	("mm7","mm5");			# %mm7 is T2
+	&psrlq	("mm5",6);
+	&pxor	("mm7","mm6");
+	&psllq	("mm6",5);
+	&pxor	("mm7","mm5");
+	&psrlq	("mm5",5);
+	&pxor	("mm7","mm6");
+	&psllq	("mm6",6);
+	&pxor	("mm7","mm5");
+	&sub	("esp",8);
+	&pxor	("mm7","mm6");			# T2=Sigma0_512(a)
+
+	&movq	("mm5",$A);			# %mm5=a
+	&por	($A,"mm2");			# a=a|c
+	&movq	("mm6",&QWP(8*(9+16-14),"esp"))	if ($prefetch);
+	&pand	("mm5","mm2");			# %mm5=a&c
+	&pand	($A,"mm1");			# a=(a|c)&b
+	&movq	("mm2",&QWP(8*(9+16-1),"esp"))	if ($prefetch);
+	&por	("mm5",$A);			# %mm5=(a&c)|((a|c)&b)
+	&paddq	("mm7","mm5");			# T2+=Maj(a,b,c)
+	&movq	($A,"mm3");			# a=T1
+
+	&mov	(&LB("edx"),&BP(0,$K512));
+	&paddq	($A,"mm7");			# a+=T2
+	&add	($K512,8);
+}
+
+sub BODY_00_15_x86 {
+	#define Sigma1(x)	(ROTR((x),14) ^ ROTR((x),18)  ^ ROTR((x),41))
+	#	LO		lo>>14^hi<<18 ^ lo>>18^hi<<14 ^ hi>>9^lo<<23
+	#	HI		hi>>14^lo<<18 ^ hi>>18^lo<<14 ^ lo>>9^hi<<23
+	&mov	("ecx",$Elo);
+	&mov	("edx",$Ehi);
+	&mov	("esi","ecx");
+
+	&shr	("ecx",9);	# lo>>9
+	&mov	("edi","edx");
+	&shr	("edx",9);	# hi>>9
+	&mov	("ebx","ecx");
+	&shl	("esi",14);	# lo<<14
+	&mov	("eax","edx");
+	&shl	("edi",14);	# hi<<14
+	&xor	("ebx","esi");
+
+	&shr	("ecx",14-9);	# lo>>14
+	&xor	("eax","edi");
+	&shr	("edx",14-9);	# hi>>14
+	&xor	("eax","ecx");
+	&shl	("esi",18-14);	# lo<<18
+	&xor	("ebx","edx");
+	&shl	("edi",18-14);	# hi<<18
+	&xor	("ebx","esi");
+
+	&shr	("ecx",18-14);	# lo>>18
+	&xor	("eax","edi");
+	&shr	("edx",18-14);	# hi>>18
+	&xor	("eax","ecx");
+	&shl	("esi",23-18);	# lo<<23
+	&xor	("ebx","edx");
+	&shl	("edi",23-18);	# hi<<23
+	&xor	("eax","esi");
+	&xor	("ebx","edi");			# T1 = Sigma1(e)
+
+	&mov	("ecx",$Flo);
+	&mov	("edx",$Fhi);
+	&mov	("esi",$Glo);
+	&mov	("edi",$Ghi);
+	 &add	("eax",$Hlo);
+	 &adc	("ebx",$Hhi);			# T1 += h
+	&xor	("ecx","esi");
+	&xor	("edx","edi");
+	&and	("ecx",$Elo);
+	&and	("edx",$Ehi);
+	 &add	("eax",&DWP(8*(9+15)+0,"esp"));
+	 &adc	("ebx",&DWP(8*(9+15)+4,"esp"));	# T1 += X[0]
+	&xor	("ecx","esi");
+	&xor	("edx","edi");			# Ch(e,f,g) = (f^g)&e)^g
+
+	&mov	("esi",&DWP(0,$K512));
+	&mov	("edi",&DWP(4,$K512));		# K[i]
+	&add	("eax","ecx");
+	&adc	("ebx","edx");			# T1 += Ch(e,f,g)
+	&mov	("ecx",$Dlo);
+	&mov	("edx",$Dhi);
+	&add	("eax","esi");
+	&adc	("ebx","edi");			# T1 += K[i]
+	&mov	($Tlo,"eax");
+	&mov	($Thi,"ebx");			# put T1 away
+	&add	("eax","ecx");
+	&adc	("ebx","edx");			# d += T1
+
+	#define Sigma0(x)	(ROTR((x),28) ^ ROTR((x),34) ^ ROTR((x),39))
+	#	LO		lo>>28^hi<<4  ^ hi>>2^lo<<30 ^ hi>>7^lo<<25
+	#	HI		hi>>28^lo<<4  ^ lo>>2^hi<<30 ^ lo>>7^hi<<25
+	&mov	("ecx",$Alo);
+	&mov	("edx",$Ahi);
+	&mov	($Dlo,"eax");
+	&mov	($Dhi,"ebx");
+	&mov	("esi","ecx");
+
+	&shr	("ecx",2);	# lo>>2
+	&mov	("edi","edx");
+	&shr	("edx",2);	# hi>>2
+	&mov	("ebx","ecx");
+	&shl	("esi",4);	# lo<<4
+	&mov	("eax","edx");
+	&shl	("edi",4);	# hi<<4
+	&xor	("ebx","esi");
+
+	&shr	("ecx",7-2);	# lo>>7
+	&xor	("eax","edi");
+	&shr	("edx",7-2);	# hi>>7
+	&xor	("ebx","ecx");
+	&shl	("esi",25-4);	# lo<<25
+	&xor	("eax","edx");
+	&shl	("edi",25-4);	# hi<<25
+	&xor	("eax","esi");
+
+	&shr	("ecx",28-7);	# lo>>28
+	&xor	("ebx","edi");
+	&shr	("edx",28-7);	# hi>>28
+	&xor	("eax","ecx");
+	&shl	("esi",30-25);	# lo<<30
+	&xor	("ebx","edx");
+	&shl	("edi",30-25);	# hi<<30
+	&xor	("eax","esi");
+	&xor	("ebx","edi");			# Sigma0(a)
+
+	&mov	("ecx",$Alo);
+	&mov	("edx",$Ahi);
+	&mov	("esi",$Blo);
+	&mov	("edi",$Bhi);
+	&add	("eax",$Tlo);
+	&adc	("ebx",$Thi);			# T1 = Sigma0(a)+T1
+	&or	("ecx","esi");
+	&or	("edx","edi");
+	&and	("ecx",$Clo);
+	&and	("edx",$Chi);
+	&and	("esi",$Alo);
+	&and	("edi",$Ahi);
+	&or	("ecx","esi");
+	&or	("edx","edi");			# Maj(a,b,c) = ((a|b)&c)|(a&b)
+
+	&add	("eax","ecx");
+	&adc	("ebx","edx");			# T1 += Maj(a,b,c)
+	&mov	($Tlo,"eax");
+	&mov	($Thi,"ebx");
+
+	&mov	(&LB("edx"),&BP(0,$K512));	# pre-fetch LSB of *K
+	&sub	("esp",8);
+	&lea	($K512,&DWP(8,$K512));		# K++
+}
+
+
+&function_begin("sha512_block_data_order");
+	&mov	("esi",wparam(0));	# ctx
+	&mov	("edi",wparam(1));	# inp
+	&mov	("eax",wparam(2));	# num
+	&mov	("ebx","esp");		# saved sp
+
+	&call	(&label("pic_point"));	# make it PIC!
+&set_label("pic_point");
+	&blindpop($K512);
+	&lea	($K512,&DWP(&label("K512")."-".&label("pic_point"),$K512));
+
+	&sub	("esp",16);
+	&and	("esp",-64);
+
+	&shl	("eax",7);
+	&add	("eax","edi");
+	&mov	(&DWP(0,"esp"),"esi");	# ctx
+	&mov	(&DWP(4,"esp"),"edi");	# inp
+	&mov	(&DWP(8,"esp"),"eax");	# inp+num*128
+	&mov	(&DWP(12,"esp"),"ebx");	# saved sp
+
+if ($sse2) {
+	&picmeup("edx","OPENSSL_ia32cap_P",$K512,&label("K512"));
+	&bt	(&DWP(0,"edx"),26);
+	&jnc	(&label("loop_x86"));
+
+	# load ctx->h[0-7]
+	&movq	($A,&QWP(0,"esi"));
+	&movq	("mm1",&QWP(8,"esi"));
+	&movq	("mm2",&QWP(16,"esi"));
+	&movq	("mm3",&QWP(24,"esi"));
+	&movq	($E,&QWP(32,"esi"));
+	&movq	("mm5",&QWP(40,"esi"));
+	&movq	("mm6",&QWP(48,"esi"));
+	&movq	("mm7",&QWP(56,"esi"));
+	&sub	("esp",8*10);
+
+&set_label("loop_sse2",16);
+	# &movq	($Asse2,$A);
+	&movq	($Bsse2,"mm1");
+	&movq	($Csse2,"mm2");
+	&movq	($Dsse2,"mm3");
+	# &movq	($Esse2,$E);
+	&movq	($Fsse2,"mm5");
+	&movq	($Gsse2,"mm6");
+	&movq	($Hsse2,"mm7");
+
+	&mov	("ecx",&DWP(0,"edi"));
+	&mov	("edx",&DWP(4,"edi"));
+	&add	("edi",8);
+	&bswap	("ecx");
+	&bswap	("edx");
+	&mov	(&DWP(8*9+4,"esp"),"ecx");
+	&mov	(&DWP(8*9+0,"esp"),"edx");
+
+&set_label("00_14_sse2",16);
+	&mov	("eax",&DWP(0,"edi"));
+	&mov	("ebx",&DWP(4,"edi"));
+	&add	("edi",8);
+	&bswap	("eax");
+	&bswap	("ebx");
+	&mov	(&DWP(8*8+4,"esp"),"eax");
+	&mov	(&DWP(8*8+0,"esp"),"ebx");
+
+	&BODY_00_15_sse2();
+
+	&cmp	(&LB("edx"),0x35);
+	&jne	(&label("00_14_sse2"));
+
+	&BODY_00_15_sse2(1);
+
+&set_label("16_79_sse2",16);
+	#&movq	("mm2",&QWP(8*(9+16-1),"esp"));	#prefetched in BODY_00_15 
+	#&movq	("mm6",&QWP(8*(9+16-14),"esp"));
+	&movq	("mm1","mm2");
+
+	&psrlq	("mm2",1);
+	&movq	("mm7","mm6");
+	&psrlq	("mm6",6);
+	&movq	("mm3","mm2");
+
+	&psrlq	("mm2",7-1);
+	&movq	("mm5","mm6");
+	&psrlq	("mm6",19-6);
+	&pxor	("mm3","mm2");
+
+	&psrlq	("mm2",8-7);
+	&pxor	("mm5","mm6");
+	&psrlq	("mm6",61-19);
+	&pxor	("mm3","mm2");
+
+	&movq	("mm2",&QWP(8*(9+16),"esp"));
+
+	&psllq	("mm1",56);
+	&pxor	("mm5","mm6");
+	&psllq	("mm7",3);
+	&pxor	("mm3","mm1");
+
+	&paddq	("mm2",&QWP(8*(9+16-9),"esp"));
+
+	&psllq	("mm1",63-56);
+	&pxor	("mm5","mm7");
+	&psllq	("mm7",45-3);
+	&pxor	("mm3","mm1");
+	&pxor	("mm5","mm7");
+
+	&paddq	("mm3","mm5");
+	&paddq	("mm3","mm2");
+	&movq	(&QWP(8*9,"esp"),"mm3");
+
+	&BODY_00_15_sse2(1);
+
+	&cmp	(&LB("edx"),0x17);
+	&jne	(&label("16_79_sse2"));
+
+	# &movq	($A,$Asse2);
+	&movq	("mm1",$Bsse2);
+	&movq	("mm2",$Csse2);
+	&movq	("mm3",$Dsse2);
+	# &movq	($E,$Esse2);
+	&movq	("mm5",$Fsse2);
+	&movq	("mm6",$Gsse2);
+	&movq	("mm7",$Hsse2);
+
+	&paddq	($A,&QWP(0,"esi"));
+	&paddq	("mm1",&QWP(8,"esi"));
+	&paddq	("mm2",&QWP(16,"esi"));
+	&paddq	("mm3",&QWP(24,"esi"));
+	&paddq	($E,&QWP(32,"esi"));
+	&paddq	("mm5",&QWP(40,"esi"));
+	&paddq	("mm6",&QWP(48,"esi"));
+	&paddq	("mm7",&QWP(56,"esi"));
+
+	&movq	(&QWP(0,"esi"),$A);
+	&movq	(&QWP(8,"esi"),"mm1");
+	&movq	(&QWP(16,"esi"),"mm2");
+	&movq	(&QWP(24,"esi"),"mm3");
+	&movq	(&QWP(32,"esi"),$E);
+	&movq	(&QWP(40,"esi"),"mm5");
+	&movq	(&QWP(48,"esi"),"mm6");
+	&movq	(&QWP(56,"esi"),"mm7");
+
+	&add	("esp",8*80);			# destroy frame
+	&sub	($K512,8*80);			# rewind K
+
+	&cmp	("edi",&DWP(8*10+8,"esp"));	# are we done yet?
+	&jb	(&label("loop_sse2"));
+
+	&emms	();
+	&mov	("esp",&DWP(8*10+12,"esp"));	# restore sp
+&function_end_A();
+}
+&set_label("loop_x86",16);
+    # copy input block to stack reversing byte and qword order
+    for ($i=0;$i<8;$i++) {
+	&mov	("eax",&DWP($i*16+0,"edi"));
+	&mov	("ebx",&DWP($i*16+4,"edi"));
+	&mov	("ecx",&DWP($i*16+8,"edi"));
+	&mov	("edx",&DWP($i*16+12,"edi"));
+	&bswap	("eax");
+	&bswap	("ebx");
+	&bswap	("ecx");
+	&bswap	("edx");
+	&push	("eax");
+	&push	("ebx");
+	&push	("ecx");
+	&push	("edx");
+    }
+	&add	("edi",128);
+	&sub	("esp",9*8);		# place for T,A,B,C,D,E,F,G,H
+	&mov	(&DWP(8*(9+16)+4,"esp"),"edi");
+
+	# copy ctx->h[0-7] to A,B,C,D,E,F,G,H on stack
+	&lea	("edi",&DWP(8,"esp"));
+	&mov	("ecx",16);
+	&data_word(0xA5F3F689);		# rep movsd
+
+&set_label("00_15_x86",16);
+	&BODY_00_15_x86();
+
+	&cmp	(&LB("edx"),0x94);
+	&jne	(&label("00_15_x86"));
+
+&set_label("16_79_x86",16);
+	#define sigma0(x)	(ROTR((x),1)  ^ ROTR((x),8)  ^ ((x)>>7))
+	#	LO		lo>>1^hi<<31  ^ lo>>8^hi<<24 ^ lo>>7^hi<<25
+	#	HI		hi>>1^lo<<31  ^ hi>>8^lo<<24 ^ hi>>7
+	&mov	("ecx",&DWP(8*(9+15+16-1)+0,"esp"));
+	&mov	("edx",&DWP(8*(9+15+16-1)+4,"esp"));
+	&mov	("esi","ecx");
+
+	&shr	("ecx",1);	# lo>>1
+	&mov	("edi","edx");
+	&shr	("edx",1);	# hi>>1
+	&mov	("eax","ecx");
+	&shl	("esi",24);	# lo<<24
+	&mov	("ebx","edx");
+	&shl	("edi",24);	# hi<<24
+	&xor	("ebx","esi");
+
+	&shr	("ecx",7-1);	# lo>>7
+	&xor	("eax","edi");
+	&shr	("edx",7-1);	# hi>>7
+	&xor	("eax","ecx");
+	&shl	("esi",31-24);	# lo<<31
+	&xor	("ebx","edx");
+	&shl	("edi",25-24);	# hi<<25
+	&xor	("ebx","esi");
+
+	&shr	("ecx",8-7);	# lo>>8
+	&xor	("eax","edi");
+	&shr	("edx",8-7);	# hi>>8
+	&xor	("eax","ecx");
+	&shl	("edi",31-25);	# hi<<31
+	&xor	("ebx","edx");
+	&xor	("eax","edi");			# T1 = sigma0(X[-15])
+
+	&mov	(&DWP(0,"esp"),"eax");
+	&mov	(&DWP(4,"esp"),"ebx");		# put T1 away
+
+	#define sigma1(x)	(ROTR((x),19) ^ ROTR((x),61) ^ ((x)>>6))
+	#	LO		lo>>19^hi<<13 ^ hi>>29^lo<<3 ^ lo>>6^hi<<26
+	#	HI		hi>>19^lo<<13 ^ lo>>29^hi<<3 ^ hi>>6
+	&mov	("ecx",&DWP(8*(9+15+16-14)+0,"esp"));
+	&mov	("edx",&DWP(8*(9+15+16-14)+4,"esp"));
+	&mov	("esi","ecx");
+
+	&shr	("ecx",6);	# lo>>6
+	&mov	("edi","edx");
+	&shr	("edx",6);	# hi>>6
+	&mov	("eax","ecx");
+	&shl	("esi",3);	# lo<<3
+	&mov	("ebx","edx");
+	&shl	("edi",3);	# hi<<3
+	&xor	("eax","esi");
+
+	&shr	("ecx",19-6);	# lo>>19
+	&xor	("ebx","edi");
+	&shr	("edx",19-6);	# hi>>19
+	&xor	("eax","ecx");
+	&shl	("esi",13-3);	# lo<<13
+	&xor	("ebx","edx");
+	&shl	("edi",13-3);	# hi<<13
+	&xor	("ebx","esi");
+
+	&shr	("ecx",29-19);	# lo>>29
+	&xor	("eax","edi");
+	&shr	("edx",29-19);	# hi>>29
+	&xor	("ebx","ecx");
+	&shl	("edi",26-13);	# hi<<26
+	&xor	("eax","edx");
+	&xor	("eax","edi");			# sigma1(X[-2])
+
+	&mov	("ecx",&DWP(8*(9+15+16)+0,"esp"));
+	&mov	("edx",&DWP(8*(9+15+16)+4,"esp"));
+	&add	("eax",&DWP(0,"esp"));
+	&adc	("ebx",&DWP(4,"esp"));		# T1 = sigma1(X[-2])+T1
+	&mov	("esi",&DWP(8*(9+15+16-9)+0,"esp"));
+	&mov	("edi",&DWP(8*(9+15+16-9)+4,"esp"));
+	&add	("eax","ecx");
+	&adc	("ebx","edx");			# T1 += X[-16]
+	&add	("eax","esi");
+	&adc	("ebx","edi");			# T1 += X[-7]
+	&mov	(&DWP(8*(9+15)+0,"esp"),"eax");
+	&mov	(&DWP(8*(9+15)+4,"esp"),"ebx");	# save X[0]
+
+	&BODY_00_15_x86();
+
+	&cmp	(&LB("edx"),0x17);
+	&jne	(&label("16_79_x86"));
+
+	&mov	("esi",&DWP(8*(9+16+80)+0,"esp"));# ctx
+	&mov	("edi",&DWP(8*(9+16+80)+4,"esp"));# inp
+    for($i=0;$i<4;$i++) {
+	&mov	("eax",&DWP($i*16+0,"esi"));
+	&mov	("ebx",&DWP($i*16+4,"esi"));
+	&mov	("ecx",&DWP($i*16+8,"esi"));
+	&mov	("edx",&DWP($i*16+12,"esi"));
+	&add	("eax",&DWP(8+($i*16)+0,"esp"));
+	&adc	("ebx",&DWP(8+($i*16)+4,"esp"));
+	&mov	(&DWP($i*16+0,"esi"),"eax");
+	&mov	(&DWP($i*16+4,"esi"),"ebx");
+	&add	("ecx",&DWP(8+($i*16)+8,"esp"));
+	&adc	("edx",&DWP(8+($i*16)+12,"esp"));
+	&mov	(&DWP($i*16+8,"esi"),"ecx");
+	&mov	(&DWP($i*16+12,"esi"),"edx");
+    }
+	&add	("esp",8*(9+16+80));		# destroy frame
+	&sub	($K512,8*80);			# rewind K
+
+	&cmp	("edi",&DWP(8,"esp"));		# are we done yet?
+	&jb	(&label("loop_x86"));
+
+	&mov	("esp",&DWP(12,"esp"));		# restore sp
+&function_end_A();
+
+&set_label("K512",64);	# Yes! I keep it in the code segment!
+	&data_word(0xd728ae22,0x428a2f98);	# u64
+	&data_word(0x23ef65cd,0x71374491);	# u64
+	&data_word(0xec4d3b2f,0xb5c0fbcf);	# u64
+	&data_word(0x8189dbbc,0xe9b5dba5);	# u64
+	&data_word(0xf348b538,0x3956c25b);	# u64
+	&data_word(0xb605d019,0x59f111f1);	# u64
+	&data_word(0xaf194f9b,0x923f82a4);	# u64
+	&data_word(0xda6d8118,0xab1c5ed5);	# u64
+	&data_word(0xa3030242,0xd807aa98);	# u64
+	&data_word(0x45706fbe,0x12835b01);	# u64
+	&data_word(0x4ee4b28c,0x243185be);	# u64
+	&data_word(0xd5ffb4e2,0x550c7dc3);	# u64
+	&data_word(0xf27b896f,0x72be5d74);	# u64
+	&data_word(0x3b1696b1,0x80deb1fe);	# u64
+	&data_word(0x25c71235,0x9bdc06a7);	# u64
+	&data_word(0xcf692694,0xc19bf174);	# u64
+	&data_word(0x9ef14ad2,0xe49b69c1);	# u64
+	&data_word(0x384f25e3,0xefbe4786);	# u64
+	&data_word(0x8b8cd5b5,0x0fc19dc6);	# u64
+	&data_word(0x77ac9c65,0x240ca1cc);	# u64
+	&data_word(0x592b0275,0x2de92c6f);	# u64
+	&data_word(0x6ea6e483,0x4a7484aa);	# u64
+	&data_word(0xbd41fbd4,0x5cb0a9dc);	# u64
+	&data_word(0x831153b5,0x76f988da);	# u64
+	&data_word(0xee66dfab,0x983e5152);	# u64
+	&data_word(0x2db43210,0xa831c66d);	# u64
+	&data_word(0x98fb213f,0xb00327c8);	# u64
+	&data_word(0xbeef0ee4,0xbf597fc7);	# u64
+	&data_word(0x3da88fc2,0xc6e00bf3);	# u64
+	&data_word(0x930aa725,0xd5a79147);	# u64
+	&data_word(0xe003826f,0x06ca6351);	# u64
+	&data_word(0x0a0e6e70,0x14292967);	# u64
+	&data_word(0x46d22ffc,0x27b70a85);	# u64
+	&data_word(0x5c26c926,0x2e1b2138);	# u64
+	&data_word(0x5ac42aed,0x4d2c6dfc);	# u64
+	&data_word(0x9d95b3df,0x53380d13);	# u64
+	&data_word(0x8baf63de,0x650a7354);	# u64
+	&data_word(0x3c77b2a8,0x766a0abb);	# u64
+	&data_word(0x47edaee6,0x81c2c92e);	# u64
+	&data_word(0x1482353b,0x92722c85);	# u64
+	&data_word(0x4cf10364,0xa2bfe8a1);	# u64
+	&data_word(0xbc423001,0xa81a664b);	# u64
+	&data_word(0xd0f89791,0xc24b8b70);	# u64
+	&data_word(0x0654be30,0xc76c51a3);	# u64
+	&data_word(0xd6ef5218,0xd192e819);	# u64
+	&data_word(0x5565a910,0xd6990624);	# u64
+	&data_word(0x5771202a,0xf40e3585);	# u64
+	&data_word(0x32bbd1b8,0x106aa070);	# u64
+	&data_word(0xb8d2d0c8,0x19a4c116);	# u64
+	&data_word(0x5141ab53,0x1e376c08);	# u64
+	&data_word(0xdf8eeb99,0x2748774c);	# u64
+	&data_word(0xe19b48a8,0x34b0bcb5);	# u64
+	&data_word(0xc5c95a63,0x391c0cb3);	# u64
+	&data_word(0xe3418acb,0x4ed8aa4a);	# u64
+	&data_word(0x7763e373,0x5b9cca4f);	# u64
+	&data_word(0xd6b2b8a3,0x682e6ff3);	# u64
+	&data_word(0x5defb2fc,0x748f82ee);	# u64
+	&data_word(0x43172f60,0x78a5636f);	# u64
+	&data_word(0xa1f0ab72,0x84c87814);	# u64
+	&data_word(0x1a6439ec,0x8cc70208);	# u64
+	&data_word(0x23631e28,0x90befffa);	# u64
+	&data_word(0xde82bde9,0xa4506ceb);	# u64
+	&data_word(0xb2c67915,0xbef9a3f7);	# u64
+	&data_word(0xe372532b,0xc67178f2);	# u64
+	&data_word(0xea26619c,0xca273ece);	# u64
+	&data_word(0x21c0c207,0xd186b8c7);	# u64
+	&data_word(0xcde0eb1e,0xeada7dd6);	# u64
+	&data_word(0xee6ed178,0xf57d4f7f);	# u64
+	&data_word(0x72176fba,0x06f067aa);	# u64
+	&data_word(0xa2c898a6,0x0a637dc5);	# u64
+	&data_word(0xbef90dae,0x113f9804);	# u64
+	&data_word(0x131c471b,0x1b710b35);	# u64
+	&data_word(0x23047d84,0x28db77f5);	# u64
+	&data_word(0x40c72493,0x32caab7b);	# u64
+	&data_word(0x15c9bebc,0x3c9ebe0a);	# u64
+	&data_word(0x9c100d4c,0x431d67c4);	# u64
+	&data_word(0xcb3e42b6,0x4cc5d4be);	# u64
+	&data_word(0xfc657e2a,0x597f299c);	# u64
+	&data_word(0x3ad6faec,0x5fcb6fab);	# u64
+	&data_word(0x4a475817,0x6c44198c);	# u64
+&function_end_B("sha512_block_data_order");
+&asciz("SHA512 block transform for x86, CRYPTOGAMS by <appro\@openssl.org>");
+
+&asm_finish();
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha512-armv4.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha512-armv4.pl
new file mode 100644
index 0000000000..7faf37b147
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha512-armv4.pl
@@ -0,0 +1,582 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> 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 block procedure for ARMv4. September 2007.
+
+# This code is ~4.5 (four and a half) times faster than code generated
+# by gcc 3.4 and it spends ~72 clock cycles per byte [on single-issue
+# Xscale PXA250 core].
+#
+# July 2010.
+#
+# Rescheduling for dual-issue pipeline resulted in 6% improvement on
+# Cortex A8 core and ~40 cycles per processed byte.
+
+# February 2011.
+#
+# Profiler-assisted and platform-specific optimization resulted in 7%
+# improvement on Coxtex A8 core and ~38 cycles per byte.
+
+# March 2011.
+#
+# Add NEON implementation. On Cortex A8 it was measured to process
+# one byte in 25.5 cycles or 47% faster than integer-only code.
+
+# Byte order [in]dependence. =========================================
+#
+# Originally caller was expected to maintain specific *dword* order in
+# h[0-7], namely with most significant dword at *lower* address, which
+# was reflected in below two parameters as 0 and 4. Now caller is
+# expected to maintain native byte order for whole 64-bit values.
+$hi="HI";
+$lo="LO";
+# ====================================================================
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+$ctx="r0";	# parameter block
+$inp="r1";
+$len="r2";
+
+$Tlo="r3";
+$Thi="r4";
+$Alo="r5";
+$Ahi="r6";
+$Elo="r7";
+$Ehi="r8";
+$t0="r9";
+$t1="r10";
+$t2="r11";
+$t3="r12";
+############	r13 is stack pointer
+$Ktbl="r14";
+############	r15 is program counter
+
+$Aoff=8*0;
+$Boff=8*1;
+$Coff=8*2;
+$Doff=8*3;
+$Eoff=8*4;
+$Foff=8*5;
+$Goff=8*6;
+$Hoff=8*7;
+$Xoff=8*8;
+
+sub BODY_00_15() {
+my $magic = shift;
+$code.=<<___;
+	@ Sigma1(x)	(ROTR((x),14) ^ ROTR((x),18)  ^ ROTR((x),41))
+	@ LO		lo>>14^hi<<18 ^ lo>>18^hi<<14 ^ hi>>9^lo<<23
+	@ HI		hi>>14^lo<<18 ^ hi>>18^lo<<14 ^ lo>>9^hi<<23
+	mov	$t0,$Elo,lsr#14
+	str	$Tlo,[sp,#$Xoff+0]
+	mov	$t1,$Ehi,lsr#14
+	str	$Thi,[sp,#$Xoff+4]
+	eor	$t0,$t0,$Ehi,lsl#18
+	ldr	$t2,[sp,#$Hoff+0]	@ h.lo
+	eor	$t1,$t1,$Elo,lsl#18
+	ldr	$t3,[sp,#$Hoff+4]	@ h.hi
+	eor	$t0,$t0,$Elo,lsr#18
+	eor	$t1,$t1,$Ehi,lsr#18
+	eor	$t0,$t0,$Ehi,lsl#14
+	eor	$t1,$t1,$Elo,lsl#14
+	eor	$t0,$t0,$Ehi,lsr#9
+	eor	$t1,$t1,$Elo,lsr#9
+	eor	$t0,$t0,$Elo,lsl#23
+	eor	$t1,$t1,$Ehi,lsl#23	@ Sigma1(e)
+	adds	$Tlo,$Tlo,$t0
+	ldr	$t0,[sp,#$Foff+0]	@ f.lo
+	adc	$Thi,$Thi,$t1		@ T += Sigma1(e)
+	ldr	$t1,[sp,#$Foff+4]	@ f.hi
+	adds	$Tlo,$Tlo,$t2
+	ldr	$t2,[sp,#$Goff+0]	@ g.lo
+	adc	$Thi,$Thi,$t3		@ T += h
+	ldr	$t3,[sp,#$Goff+4]	@ g.hi
+
+	eor	$t0,$t0,$t2
+	str	$Elo,[sp,#$Eoff+0]
+	eor	$t1,$t1,$t3
+	str	$Ehi,[sp,#$Eoff+4]
+	and	$t0,$t0,$Elo
+	str	$Alo,[sp,#$Aoff+0]
+	and	$t1,$t1,$Ehi
+	str	$Ahi,[sp,#$Aoff+4]
+	eor	$t0,$t0,$t2
+	ldr	$t2,[$Ktbl,#$lo]	@ K[i].lo
+	eor	$t1,$t1,$t3		@ Ch(e,f,g)
+	ldr	$t3,[$Ktbl,#$hi]	@ K[i].hi
+
+	adds	$Tlo,$Tlo,$t0
+	ldr	$Elo,[sp,#$Doff+0]	@ d.lo
+	adc	$Thi,$Thi,$t1		@ T += Ch(e,f,g)
+	ldr	$Ehi,[sp,#$Doff+4]	@ d.hi
+	adds	$Tlo,$Tlo,$t2
+	and	$t0,$t2,#0xff
+	adc	$Thi,$Thi,$t3		@ T += K[i]
+	adds	$Elo,$Elo,$Tlo
+	ldr	$t2,[sp,#$Boff+0]	@ b.lo
+	adc	$Ehi,$Ehi,$Thi		@ d += T
+	teq	$t0,#$magic
+
+	ldr	$t3,[sp,#$Coff+0]	@ c.lo
+	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
+	@ HI		hi>>28^lo<<4  ^ lo>>2^hi<<30 ^ lo>>7^hi<<25
+	mov	$t0,$Alo,lsr#28
+	mov	$t1,$Ahi,lsr#28
+	eor	$t0,$t0,$Ahi,lsl#4
+	eor	$t1,$t1,$Alo,lsl#4
+	eor	$t0,$t0,$Ahi,lsr#2
+	eor	$t1,$t1,$Alo,lsr#2
+	eor	$t0,$t0,$Alo,lsl#30
+	eor	$t1,$t1,$Ahi,lsl#30
+	eor	$t0,$t0,$Ahi,lsr#7
+	eor	$t1,$t1,$Alo,lsr#7
+	eor	$t0,$t0,$Alo,lsl#25
+	eor	$t1,$t1,$Ahi,lsl#25	@ Sigma0(a)
+	adds	$Tlo,$Tlo,$t0
+	and	$t0,$Alo,$t2
+	adc	$Thi,$Thi,$t1		@ T += Sigma0(a)
+
+	ldr	$t1,[sp,#$Boff+4]	@ b.hi
+	orr	$Alo,$Alo,$t2
+	ldr	$t2,[sp,#$Coff+4]	@ c.hi
+	and	$Alo,$Alo,$t3
+	and	$t3,$Ahi,$t1
+	orr	$Ahi,$Ahi,$t1
+	orr	$Alo,$Alo,$t0		@ Maj(a,b,c).lo
+	and	$Ahi,$Ahi,$t2
+	adds	$Alo,$Alo,$Tlo
+	orr	$Ahi,$Ahi,$t3		@ Maj(a,b,c).hi
+	sub	sp,sp,#8
+	adc	$Ahi,$Ahi,$Thi		@ h += T
+	tst	$Ktbl,#1
+	add	$Ktbl,$Ktbl,#8
+___
+}
+$code=<<___;
+#include "arm_arch.h"
+#ifdef __ARMEL__
+# define LO 0
+# define HI 4
+# define WORD64(hi0,lo0,hi1,lo1)	.word	lo0,hi0, lo1,hi1
+#else
+# define HI 0
+# define LO 4
+# define WORD64(hi0,lo0,hi1,lo1)	.word	hi0,lo0, hi1,lo1
+#endif
+
+.text
+.code	32
+.type	K512,%object
+.align	5
+K512:
+WORD64(0x428a2f98,0xd728ae22, 0x71374491,0x23ef65cd)
+WORD64(0xb5c0fbcf,0xec4d3b2f, 0xe9b5dba5,0x8189dbbc)
+WORD64(0x3956c25b,0xf348b538, 0x59f111f1,0xb605d019)
+WORD64(0x923f82a4,0xaf194f9b, 0xab1c5ed5,0xda6d8118)
+WORD64(0xd807aa98,0xa3030242, 0x12835b01,0x45706fbe)
+WORD64(0x243185be,0x4ee4b28c, 0x550c7dc3,0xd5ffb4e2)
+WORD64(0x72be5d74,0xf27b896f, 0x80deb1fe,0x3b1696b1)
+WORD64(0x9bdc06a7,0x25c71235, 0xc19bf174,0xcf692694)
+WORD64(0xe49b69c1,0x9ef14ad2, 0xefbe4786,0x384f25e3)
+WORD64(0x0fc19dc6,0x8b8cd5b5, 0x240ca1cc,0x77ac9c65)
+WORD64(0x2de92c6f,0x592b0275, 0x4a7484aa,0x6ea6e483)
+WORD64(0x5cb0a9dc,0xbd41fbd4, 0x76f988da,0x831153b5)
+WORD64(0x983e5152,0xee66dfab, 0xa831c66d,0x2db43210)
+WORD64(0xb00327c8,0x98fb213f, 0xbf597fc7,0xbeef0ee4)
+WORD64(0xc6e00bf3,0x3da88fc2, 0xd5a79147,0x930aa725)
+WORD64(0x06ca6351,0xe003826f, 0x14292967,0x0a0e6e70)
+WORD64(0x27b70a85,0x46d22ffc, 0x2e1b2138,0x5c26c926)
+WORD64(0x4d2c6dfc,0x5ac42aed, 0x53380d13,0x9d95b3df)
+WORD64(0x650a7354,0x8baf63de, 0x766a0abb,0x3c77b2a8)
+WORD64(0x81c2c92e,0x47edaee6, 0x92722c85,0x1482353b)
+WORD64(0xa2bfe8a1,0x4cf10364, 0xa81a664b,0xbc423001)
+WORD64(0xc24b8b70,0xd0f89791, 0xc76c51a3,0x0654be30)
+WORD64(0xd192e819,0xd6ef5218, 0xd6990624,0x5565a910)
+WORD64(0xf40e3585,0x5771202a, 0x106aa070,0x32bbd1b8)
+WORD64(0x19a4c116,0xb8d2d0c8, 0x1e376c08,0x5141ab53)
+WORD64(0x2748774c,0xdf8eeb99, 0x34b0bcb5,0xe19b48a8)
+WORD64(0x391c0cb3,0xc5c95a63, 0x4ed8aa4a,0xe3418acb)
+WORD64(0x5b9cca4f,0x7763e373, 0x682e6ff3,0xd6b2b8a3)
+WORD64(0x748f82ee,0x5defb2fc, 0x78a5636f,0x43172f60)
+WORD64(0x84c87814,0xa1f0ab72, 0x8cc70208,0x1a6439ec)
+WORD64(0x90befffa,0x23631e28, 0xa4506ceb,0xde82bde9)
+WORD64(0xbef9a3f7,0xb2c67915, 0xc67178f2,0xe372532b)
+WORD64(0xca273ece,0xea26619c, 0xd186b8c7,0x21c0c207)
+WORD64(0xeada7dd6,0xcde0eb1e, 0xf57d4f7f,0xee6ed178)
+WORD64(0x06f067aa,0x72176fba, 0x0a637dc5,0xa2c898a6)
+WORD64(0x113f9804,0xbef90dae, 0x1b710b35,0x131c471b)
+WORD64(0x28db77f5,0x23047d84, 0x32caab7b,0x40c72493)
+WORD64(0x3c9ebe0a,0x15c9bebc, 0x431d67c4,0x9c100d4c)
+WORD64(0x4cc5d4be,0xcb3e42b6, 0x597f299c,0xfc657e2a)
+WORD64(0x5fcb6fab,0x3ad6faec, 0x6c44198c,0x4a475817)
+.size	K512,.-K512
+.LOPENSSL_armcap:
+.word	OPENSSL_armcap_P-sha512_block_data_order
+.skip	32-4
+
+.global	sha512_block_data_order
+.type	sha512_block_data_order,%function
+sha512_block_data_order:
+	sub	r3,pc,#8		@ sha512_block_data_order
+	add	$len,$inp,$len,lsl#7	@ len to point at the end of inp
+#if __ARM_ARCH__>=7
+	ldr	r12,.LOPENSSL_armcap
+	ldr	r12,[r3,r12]		@ OPENSSL_armcap_P
+	tst	r12,#1
+	bne	.LNEON
+#endif
+	stmdb	sp!,{r4-r12,lr}
+	sub	$Ktbl,r3,#672		@ K512
+	sub	sp,sp,#9*8
+
+	ldr	$Elo,[$ctx,#$Eoff+$lo]
+	ldr	$Ehi,[$ctx,#$Eoff+$hi]
+	ldr	$t0, [$ctx,#$Goff+$lo]
+	ldr	$t1, [$ctx,#$Goff+$hi]
+	ldr	$t2, [$ctx,#$Hoff+$lo]
+	ldr	$t3, [$ctx,#$Hoff+$hi]
+.Loop:
+	str	$t0, [sp,#$Goff+0]
+	str	$t1, [sp,#$Goff+4]
+	str	$t2, [sp,#$Hoff+0]
+	str	$t3, [sp,#$Hoff+4]
+	ldr	$Alo,[$ctx,#$Aoff+$lo]
+	ldr	$Ahi,[$ctx,#$Aoff+$hi]
+	ldr	$Tlo,[$ctx,#$Boff+$lo]
+	ldr	$Thi,[$ctx,#$Boff+$hi]
+	ldr	$t0, [$ctx,#$Coff+$lo]
+	ldr	$t1, [$ctx,#$Coff+$hi]
+	ldr	$t2, [$ctx,#$Doff+$lo]
+	ldr	$t3, [$ctx,#$Doff+$hi]
+	str	$Tlo,[sp,#$Boff+0]
+	str	$Thi,[sp,#$Boff+4]
+	str	$t0, [sp,#$Coff+0]
+	str	$t1, [sp,#$Coff+4]
+	str	$t2, [sp,#$Doff+0]
+	str	$t3, [sp,#$Doff+4]
+	ldr	$Tlo,[$ctx,#$Foff+$lo]
+	ldr	$Thi,[$ctx,#$Foff+$hi]
+	str	$Tlo,[sp,#$Foff+0]
+	str	$Thi,[sp,#$Foff+4]
+
+.L00_15:
+#if __ARM_ARCH__<7
+	ldrb	$Tlo,[$inp,#7]
+	ldrb	$t0, [$inp,#6]
+	ldrb	$t1, [$inp,#5]
+	ldrb	$t2, [$inp,#4]
+	ldrb	$Thi,[$inp,#3]
+	ldrb	$t3, [$inp,#2]
+	orr	$Tlo,$Tlo,$t0,lsl#8
+	ldrb	$t0, [$inp,#1]
+	orr	$Tlo,$Tlo,$t1,lsl#16
+	ldrb	$t1, [$inp],#8
+	orr	$Tlo,$Tlo,$t2,lsl#24
+	orr	$Thi,$Thi,$t3,lsl#8
+	orr	$Thi,$Thi,$t0,lsl#16
+	orr	$Thi,$Thi,$t1,lsl#24
+#else
+	ldr	$Tlo,[$inp,#4]
+	ldr	$Thi,[$inp],#8
+#ifdef __ARMEL__
+	rev	$Tlo,$Tlo
+	rev	$Thi,$Thi
+#endif
+#endif
+___
+	&BODY_00_15(0x94);
+$code.=<<___;
+	tst	$Ktbl,#1
+	beq	.L00_15
+	ldr	$t0,[sp,#`$Xoff+8*(16-1)`+0]
+	ldr	$t1,[sp,#`$Xoff+8*(16-1)`+4]
+	bic	$Ktbl,$Ktbl,#1
+.L16_79:
+	@ sigma0(x)	(ROTR((x),1)  ^ ROTR((x),8)  ^ ((x)>>7))
+	@ LO		lo>>1^hi<<31  ^ lo>>8^hi<<24 ^ lo>>7^hi<<25
+	@ HI		hi>>1^lo<<31  ^ hi>>8^lo<<24 ^ hi>>7
+	mov	$Tlo,$t0,lsr#1
+	ldr	$t2,[sp,#`$Xoff+8*(16-14)`+0]
+	mov	$Thi,$t1,lsr#1
+	ldr	$t3,[sp,#`$Xoff+8*(16-14)`+4]
+	eor	$Tlo,$Tlo,$t1,lsl#31
+	eor	$Thi,$Thi,$t0,lsl#31
+	eor	$Tlo,$Tlo,$t0,lsr#8
+	eor	$Thi,$Thi,$t1,lsr#8
+	eor	$Tlo,$Tlo,$t1,lsl#24
+	eor	$Thi,$Thi,$t0,lsl#24
+	eor	$Tlo,$Tlo,$t0,lsr#7
+	eor	$Thi,$Thi,$t1,lsr#7
+	eor	$Tlo,$Tlo,$t1,lsl#25
+
+	@ sigma1(x)	(ROTR((x),19) ^ ROTR((x),61) ^ ((x)>>6))
+	@ LO		lo>>19^hi<<13 ^ hi>>29^lo<<3 ^ lo>>6^hi<<26
+	@ HI		hi>>19^lo<<13 ^ lo>>29^hi<<3 ^ hi>>6
+	mov	$t0,$t2,lsr#19
+	mov	$t1,$t3,lsr#19
+	eor	$t0,$t0,$t3,lsl#13
+	eor	$t1,$t1,$t2,lsl#13
+	eor	$t0,$t0,$t3,lsr#29
+	eor	$t1,$t1,$t2,lsr#29
+	eor	$t0,$t0,$t2,lsl#3
+	eor	$t1,$t1,$t3,lsl#3
+	eor	$t0,$t0,$t2,lsr#6
+	eor	$t1,$t1,$t3,lsr#6
+	ldr	$t2,[sp,#`$Xoff+8*(16-9)`+0]
+	eor	$t0,$t0,$t3,lsl#26
+
+	ldr	$t3,[sp,#`$Xoff+8*(16-9)`+4]
+	adds	$Tlo,$Tlo,$t0
+	ldr	$t0,[sp,#`$Xoff+8*16`+0]
+	adc	$Thi,$Thi,$t1
+
+	ldr	$t1,[sp,#`$Xoff+8*16`+4]
+	adds	$Tlo,$Tlo,$t2
+	adc	$Thi,$Thi,$t3
+	adds	$Tlo,$Tlo,$t0
+	adc	$Thi,$Thi,$t1
+___
+	&BODY_00_15(0x17);
+$code.=<<___;
+	ldreq	$t0,[sp,#`$Xoff+8*(16-1)`+0]
+	ldreq	$t1,[sp,#`$Xoff+8*(16-1)`+4]
+	beq	.L16_79
+	bic	$Ktbl,$Ktbl,#1
+
+	ldr	$Tlo,[sp,#$Boff+0]
+	ldr	$Thi,[sp,#$Boff+4]
+	ldr	$t0, [$ctx,#$Aoff+$lo]
+	ldr	$t1, [$ctx,#$Aoff+$hi]
+	ldr	$t2, [$ctx,#$Boff+$lo]
+	ldr	$t3, [$ctx,#$Boff+$hi]
+	adds	$t0,$Alo,$t0
+	str	$t0, [$ctx,#$Aoff+$lo]
+	adc	$t1,$Ahi,$t1
+	str	$t1, [$ctx,#$Aoff+$hi]
+	adds	$t2,$Tlo,$t2
+	str	$t2, [$ctx,#$Boff+$lo]
+	adc	$t3,$Thi,$t3
+	str	$t3, [$ctx,#$Boff+$hi]
+
+	ldr	$Alo,[sp,#$Coff+0]
+	ldr	$Ahi,[sp,#$Coff+4]
+	ldr	$Tlo,[sp,#$Doff+0]
+	ldr	$Thi,[sp,#$Doff+4]
+	ldr	$t0, [$ctx,#$Coff+$lo]
+	ldr	$t1, [$ctx,#$Coff+$hi]
+	ldr	$t2, [$ctx,#$Doff+$lo]
+	ldr	$t3, [$ctx,#$Doff+$hi]
+	adds	$t0,$Alo,$t0
+	str	$t0, [$ctx,#$Coff+$lo]
+	adc	$t1,$Ahi,$t1
+	str	$t1, [$ctx,#$Coff+$hi]
+	adds	$t2,$Tlo,$t2
+	str	$t2, [$ctx,#$Doff+$lo]
+	adc	$t3,$Thi,$t3
+	str	$t3, [$ctx,#$Doff+$hi]
+
+	ldr	$Tlo,[sp,#$Foff+0]
+	ldr	$Thi,[sp,#$Foff+4]
+	ldr	$t0, [$ctx,#$Eoff+$lo]
+	ldr	$t1, [$ctx,#$Eoff+$hi]
+	ldr	$t2, [$ctx,#$Foff+$lo]
+	ldr	$t3, [$ctx,#$Foff+$hi]
+	adds	$Elo,$Elo,$t0
+	str	$Elo,[$ctx,#$Eoff+$lo]
+	adc	$Ehi,$Ehi,$t1
+	str	$Ehi,[$ctx,#$Eoff+$hi]
+	adds	$t2,$Tlo,$t2
+	str	$t2, [$ctx,#$Foff+$lo]
+	adc	$t3,$Thi,$t3
+	str	$t3, [$ctx,#$Foff+$hi]
+
+	ldr	$Alo,[sp,#$Goff+0]
+	ldr	$Ahi,[sp,#$Goff+4]
+	ldr	$Tlo,[sp,#$Hoff+0]
+	ldr	$Thi,[sp,#$Hoff+4]
+	ldr	$t0, [$ctx,#$Goff+$lo]
+	ldr	$t1, [$ctx,#$Goff+$hi]
+	ldr	$t2, [$ctx,#$Hoff+$lo]
+	ldr	$t3, [$ctx,#$Hoff+$hi]
+	adds	$t0,$Alo,$t0
+	str	$t0, [$ctx,#$Goff+$lo]
+	adc	$t1,$Ahi,$t1
+	str	$t1, [$ctx,#$Goff+$hi]
+	adds	$t2,$Tlo,$t2
+	str	$t2, [$ctx,#$Hoff+$lo]
+	adc	$t3,$Thi,$t3
+	str	$t3, [$ctx,#$Hoff+$hi]
+
+	add	sp,sp,#640
+	sub	$Ktbl,$Ktbl,#640
+
+	teq	$inp,$len
+	bne	.Loop
+
+	add	sp,sp,#8*9		@ destroy frame
+#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
+___
+
+{
+my @Sigma0=(28,34,39);
+my @Sigma1=(14,18,41);
+my @sigma0=(1, 8, 7);
+my @sigma1=(19,61,6);
+
+my $Ktbl="r3";
+my $cnt="r12";	# volatile register known as ip, intra-procedure-call scratch
+
+my @X=map("d$_",(0..15));
+my @V=($A,$B,$C,$D,$E,$F,$G,$H)=map("d$_",(16..23));
+
+sub NEON_00_15() {
+my $i=shift;
+my ($a,$b,$c,$d,$e,$f,$g,$h)=@_;
+my ($t0,$t1,$t2,$T1,$K,$Ch,$Maj)=map("d$_",(24..31));	# temps
+
+$code.=<<___ if ($i<16 || $i&1);
+	vshr.u64	$t0,$e,#@Sigma1[0]	@ $i
+#if $i<16
+	vld1.64		{@X[$i%16]},[$inp]!	@ handles unaligned
+#endif
+	vshr.u64	$t1,$e,#@Sigma1[1]
+	vshr.u64	$t2,$e,#@Sigma1[2]
+___
+$code.=<<___;
+	vld1.64		{$K},[$Ktbl,:64]!	@ K[i++]
+	vsli.64		$t0,$e,#`64-@Sigma1[0]`
+	vsli.64		$t1,$e,#`64-@Sigma1[1]`
+	vsli.64		$t2,$e,#`64-@Sigma1[2]`
+#if $i<16 && defined(__ARMEL__)
+	vrev64.8	@X[$i],@X[$i]
+#endif
+	vadd.i64	$T1,$K,$h
+	veor		$Ch,$f,$g
+	veor		$t0,$t1
+	vand		$Ch,$e
+	veor		$t0,$t2			@ Sigma1(e)
+	veor		$Ch,$g			@ Ch(e,f,g)
+	vadd.i64	$T1,$t0
+	vshr.u64	$t0,$a,#@Sigma0[0]
+	vadd.i64	$T1,$Ch
+	vshr.u64	$t1,$a,#@Sigma0[1]
+	vshr.u64	$t2,$a,#@Sigma0[2]
+	vsli.64		$t0,$a,#`64-@Sigma0[0]`
+	vsli.64		$t1,$a,#`64-@Sigma0[1]`
+	vsli.64		$t2,$a,#`64-@Sigma0[2]`
+	vadd.i64	$T1,@X[$i%16]
+	vorr		$Maj,$a,$c
+	vand		$Ch,$a,$c
+	veor		$h,$t0,$t1
+	vand		$Maj,$b
+	veor		$h,$t2			@ Sigma0(a)
+	vorr		$Maj,$Ch		@ Maj(a,b,c)
+	vadd.i64	$h,$T1
+	vadd.i64	$d,$T1
+	vadd.i64	$h,$Maj
+___
+}
+
+sub NEON_16_79() {
+my $i=shift;
+
+if ($i&1)	{ &NEON_00_15($i,@_); return; }
+
+# 2x-vectorized, therefore runs every 2nd round
+my @X=map("q$_",(0..7));			# view @X as 128-bit vector
+my ($t0,$t1,$s0,$s1) = map("q$_",(12..15));	# temps
+my ($d0,$d1,$d2) = map("d$_",(24..26));		# temps from NEON_00_15
+my $e=@_[4];					# $e from NEON_00_15
+$i /= 2;
+$code.=<<___;
+	vshr.u64	$t0,@X[($i+7)%8],#@sigma1[0]
+	vshr.u64	$t1,@X[($i+7)%8],#@sigma1[1]
+	vshr.u64	$s1,@X[($i+7)%8],#@sigma1[2]
+	vsli.64		$t0,@X[($i+7)%8],#`64-@sigma1[0]`
+	vext.8		$s0,@X[$i%8],@X[($i+1)%8],#8	@ X[i+1]
+	vsli.64		$t1,@X[($i+7)%8],#`64-@sigma1[1]`
+	veor		$s1,$t0
+	vshr.u64	$t0,$s0,#@sigma0[0]
+	veor		$s1,$t1				@ sigma1(X[i+14])
+	vshr.u64	$t1,$s0,#@sigma0[1]
+	vadd.i64	@X[$i%8],$s1
+	vshr.u64	$s1,$s0,#@sigma0[2]
+	vsli.64		$t0,$s0,#`64-@sigma0[0]`
+	vsli.64		$t1,$s0,#`64-@sigma0[1]`
+	vext.8		$s0,@X[($i+4)%8],@X[($i+5)%8],#8	@ X[i+9]
+	veor		$s1,$t0
+	vshr.u64	$d0,$e,#@Sigma1[0]		@ from NEON_00_15
+	vadd.i64	@X[$i%8],$s0
+	vshr.u64	$d1,$e,#@Sigma1[1]		@ from NEON_00_15
+	veor		$s1,$t1				@ sigma0(X[i+1])
+	vshr.u64	$d2,$e,#@Sigma1[2]		@ from NEON_00_15
+	vadd.i64	@X[$i%8],$s1
+___
+	&NEON_00_15(2*$i,@_);
+}
+
+$code.=<<___;
+#if __ARM_ARCH__>=7
+.fpu	neon
+
+.align	4
+.LNEON:
+	dmb				@ errata #451034 on early Cortex A8
+	vstmdb	sp!,{d8-d15}		@ ABI specification says so
+	sub	$Ktbl,r3,#672		@ K512
+	vldmia	$ctx,{$A-$H}		@ load context
+.Loop_neon:
+___
+for($i=0;$i<16;$i++)	{ &NEON_00_15($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	mov		$cnt,#4
+.L16_79_neon:
+	subs		$cnt,#1
+___
+for(;$i<32;$i++)	{ &NEON_16_79($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	bne		.L16_79_neon
+
+	vldmia		$ctx,{d24-d31}	@ load context to temp
+	vadd.i64	q8,q12		@ vectorized accumulate
+	vadd.i64	q9,q13
+	vadd.i64	q10,q14
+	vadd.i64	q11,q15
+	vstmia		$ctx,{$A-$H}	@ save context
+	teq		$inp,$len
+	sub		$Ktbl,#640	@ rewind K512
+	bne		.Loop_neon
+
+	vldmia	sp!,{d8-d15}		@ epilogue
+	bx	lr
+#endif
+___
+}
+$code.=<<___;
+.size	sha512_block_data_order,.-sha512_block_data_order
+.asciz	"SHA512 block transform for ARMv4/NEON, CRYPTOGAMS by <appro\@openssl.org>"
+.align	2
+.comm	OPENSSL_armcap_P,4,4
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+$code =~ s/\bbx\s+lr\b/.word\t0xe12fff1e/gm;	# make it possible to compile with -march=armv4
+print $code;
+close STDOUT; # enforce flush
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha512-ia64.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha512-ia64.pl
new file mode 100755
index 0000000000..1c6ce56522
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha512-ia64.pl
@@ -0,0 +1,672 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> 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/.
+# ====================================================================
+#
+# SHA256/512_Transform for Itanium.
+#
+# sha512_block runs in 1003 cycles on Itanium 2, which is almost 50%
+# faster than gcc and >60%(!) faster than code generated by HP-UX
+# compiler (yes, HP-UX is generating slower code, because unlike gcc,
+# it failed to deploy "shift right pair," 'shrp' instruction, which
+# substitutes for 64-bit rotate).
+#
+# 924 cycles long sha256_block outperforms gcc by over factor of 2(!)
+# and HP-UX compiler - by >40% (yes, gcc won sha512_block, but lost
+# this one big time). Note that "formally" 924 is about 100 cycles
+# too much. I mean it's 64 32-bit rounds vs. 80 virtually identical
+# 64-bit ones and 1003*64/80 gives 802. Extra cycles, 2 per round,
+# are spent on extra work to provide for 32-bit rotations. 32-bit
+# rotations are still handled by 'shrp' instruction and for this
+# reason lower 32 bits are deposited to upper half of 64-bit register
+# prior 'shrp' issue. And in order to minimize the amount of such
+# operations, X[16] values are *maintained* with copies of lower
+# halves in upper halves, which is why you'll spot such instructions
+# as custom 'mux2', "parallel 32-bit add," 'padd4' and "parallel
+# 32-bit unsigned right shift," 'pshr4.u' instructions here.
+#
+# Rules of engagement.
+#
+# There is only one integer shifter meaning that if I have two rotate,
+# deposit or extract instructions in adjacent bundles, they shall
+# split [at run-time if they have to]. But note that variable and
+# parallel shifts are performed by multi-media ALU and *are* pairable
+# with rotates [and alike]. On the backside MMALU is rather slow: it
+# takes 2 extra cycles before the result of integer operation is
+# available *to* MMALU and 2(*) extra cycles before the result of MM
+# operation is available "back" *to* integer ALU, not to mention that
+# MMALU itself has 2 cycles latency. However! I explicitly scheduled
+# these MM instructions to avoid MM stalls, so that all these extra
+# latencies get "hidden" in instruction-level parallelism.
+#
+# (*) 2 cycles on Itanium 1 and 1 cycle on Itanium 2. But I schedule
+#     for 2 in order to provide for best *overall* performance,
+#     because on Itanium 1 stall on MM result is accompanied by
+#     pipeline flush, which takes 6 cycles:-(
+#
+# Resulting performance numbers for 900MHz Itanium 2 system:
+#
+# The 'numbers' are in 1000s of bytes per second processed.
+# type     16 bytes    64 bytes   256 bytes  1024 bytes  8192 bytes
+# sha1(*)   6210.14k   20376.30k   52447.83k   85870.05k  105478.12k
+# sha256    7476.45k   20572.05k   41538.34k   56062.29k   62093.18k
+# sha512    4996.56k   20026.28k   47597.20k   85278.79k  111501.31k
+#
+# (*) SHA1 numbers are for HP-UX compiler and are presented purely
+#     for reference purposes. I bet it can improved too...
+#
+# To generate code, pass the file name with either 256 or 512 in its
+# name and compiler flags.
+
+$output=shift;
+
+if ($output =~ /512.*\.[s|asm]/) {
+	$SZ=8;
+	$BITS=8*$SZ;
+	$LDW="ld8";
+	$STW="st8";
+	$ADD="add";
+	$SHRU="shr.u";
+	$TABLE="K512";
+	$func="sha512_block_data_order";
+	@Sigma0=(28,34,39);
+	@Sigma1=(14,18,41);
+	@sigma0=(1,  8, 7);
+	@sigma1=(19,61, 6);
+	$rounds=80;
+} elsif ($output =~ /256.*\.[s|asm]/) {
+	$SZ=4;
+	$BITS=8*$SZ;
+	$LDW="ld4";
+	$STW="st4";
+	$ADD="padd4";
+	$SHRU="pshr4.u";
+	$TABLE="K256";
+	$func="sha256_block_data_order";
+	@Sigma0=( 2,13,22);
+	@Sigma1=( 6,11,25);
+	@sigma0=( 7,18, 3);
+	@sigma1=(17,19,10);
+	$rounds=64;
+} else { die "nonsense $output"; }
+
+open STDOUT,">$output" || die "can't open $output: $!";
+
+if ($^O eq "hpux") {
+    $ADDP="addp4";
+    for (@ARGV) { $ADDP="add" if (/[\+DD|\-mlp]64/); }
+} else { $ADDP="add"; }
+for (@ARGV)  {	$big_endian=1 if (/\-DB_ENDIAN/);
+		$big_endian=0 if (/\-DL_ENDIAN/);  }
+if (!defined($big_endian))
+             {	$big_endian=(unpack('L',pack('N',1))==1);  }
+
+$code=<<___;
+.ident  \"$output, version 1.1\"
+.ident  \"IA-64 ISA artwork by Andy Polyakov <appro\@fy.chalmers.se>\"
+.explicit
+.text
+
+pfssave=r2;
+lcsave=r3;
+prsave=r14;
+K=r15;
+A=r16;	B=r17;	C=r18;	D=r19;
+E=r20;	F=r21;	G=r22;	H=r23;
+T1=r24;	T2=r25;
+s0=r26;	s1=r27;	t0=r28;	t1=r29;
+Ktbl=r30;
+ctx=r31;	// 1st arg
+input=r48;	// 2nd arg
+num=r49;	// 3rd arg
+sgm0=r50;	sgm1=r51;	// small constants
+A_=r54;	B_=r55;	C_=r56;	D_=r57;
+E_=r58;	F_=r59;	G_=r60;	H_=r61;
+
+// void $func (SHA_CTX *ctx, const void *in,size_t num[,int host])
+.global	$func#
+.proc	$func#
+.align	32
+$func:
+	.prologue
+	.save	ar.pfs,pfssave
+{ .mmi;	alloc	pfssave=ar.pfs,3,27,0,16
+	$ADDP	ctx=0,r32		// 1st arg
+	.save	ar.lc,lcsave
+	mov	lcsave=ar.lc	}
+{ .mmi;	$ADDP	input=0,r33		// 2nd arg
+	mov	num=r34			// 3rd arg
+	.save	pr,prsave
+	mov	prsave=pr	};;
+
+	.body
+{ .mib;	add	r8=0*$SZ,ctx
+	add	r9=1*$SZ,ctx
+	brp.loop.imp	.L_first16,.L_first16_end-16	}
+{ .mib;	add	r10=2*$SZ,ctx
+	add	r11=3*$SZ,ctx
+	brp.loop.imp	.L_rest,.L_rest_end-16		};;
+
+// load A-H
+.Lpic_point:
+{ .mmi;	$LDW	A_=[r8],4*$SZ
+	$LDW	B_=[r9],4*$SZ
+	mov	Ktbl=ip		}
+{ .mmi;	$LDW	C_=[r10],4*$SZ
+	$LDW	D_=[r11],4*$SZ
+	mov	sgm0=$sigma0[2]	};;
+{ .mmi;	$LDW	E_=[r8]
+	$LDW	F_=[r9]
+	add	Ktbl=($TABLE#-.Lpic_point),Ktbl		}
+{ .mmi;	$LDW	G_=[r10]
+	$LDW	H_=[r11]
+	cmp.ne	p0,p16=0,r0	};;	// used in sha256_block
+___
+$code.=<<___ if ($BITS==64);
+{ .mii;	and	r8=7,input
+	and	input=~7,input;;
+	cmp.eq	p9,p0=1,r8	}
+{ .mmi;	cmp.eq	p10,p0=2,r8
+	cmp.eq	p11,p0=3,r8
+	cmp.eq	p12,p0=4,r8	}
+{ .mmi;	cmp.eq	p13,p0=5,r8
+	cmp.eq	p14,p0=6,r8
+	cmp.eq	p15,p0=7,r8	};;
+___
+$code.=<<___;
+.L_outer:
+.rotr	X[16]
+{ .mmi;	mov	A=A_
+	mov	B=B_
+	mov	ar.lc=14	}
+{ .mmi;	mov	C=C_
+	mov	D=D_
+	mov	E=E_		}
+{ .mmi;	mov	F=F_
+	mov	G=G_
+	mov	ar.ec=2		}
+{ .mmi;	ld1	X[15]=[input],$SZ		// eliminated in 64-bit
+	mov	H=H_
+	mov	sgm1=$sigma1[2]	};;
+
+___
+$t0="t0", $t1="t1", $code.=<<___ if ($BITS==32);
+.align	32
+.L_first16:
+{ .mmi;		add	r9=1-$SZ,input
+		add	r10=2-$SZ,input
+		add	r11=3-$SZ,input	};;
+{ .mmi;		ld1	r9=[r9]
+		ld1	r10=[r10]
+		dep.z	$t1=E,32,32	}
+{ .mmi;		$LDW	K=[Ktbl],$SZ
+		ld1	r11=[r11]
+		zxt4	E=E		};;
+{ .mii;		or	$t1=$t1,E
+		dep	X[15]=X[15],r9,8,8
+		dep	r11=r10,r11,8,8	};;
+{ .mmi;		and	T1=F,E
+		and	T2=A,B
+		dep	X[15]=X[15],r11,16,16	}
+{ .mmi;		andcm	r8=G,E
+		and	r9=A,C
+		mux2	$t0=A,0x44	};;	// copy lower half to upper
+{ .mmi;	(p16)	ld1	X[15-1]=[input],$SZ	// prefetch
+		xor	T1=T1,r8		// T1=((e & f) ^ (~e & g))
+		_rotr	r11=$t1,$Sigma1[0] }	// ROTR(e,14)
+{ .mib;		and	r10=B,C
+		xor	T2=T2,r9	};;
+___
+$t0="A", $t1="E", $code.=<<___ if ($BITS==64);
+// in 64-bit mode I load whole X[16] at once and take care of alignment...
+{ .mmi;	add	r8=1*$SZ,input
+	add	r9=2*$SZ,input
+	add	r10=3*$SZ,input		};;
+{ .mmb;	$LDW	X[15]=[input],4*$SZ
+	$LDW	X[14]=[r8],4*$SZ
+(p9)	br.cond.dpnt.many	.L1byte	};;
+{ .mmb;	$LDW	X[13]=[r9],4*$SZ
+	$LDW	X[12]=[r10],4*$SZ
+(p10)	br.cond.dpnt.many	.L2byte	};;
+{ .mmb;	$LDW	X[11]=[input],4*$SZ
+	$LDW	X[10]=[r8],4*$SZ
+(p11)	br.cond.dpnt.many	.L3byte	};;
+{ .mmb;	$LDW	X[ 9]=[r9],4*$SZ
+	$LDW	X[ 8]=[r10],4*$SZ
+(p12)	br.cond.dpnt.many	.L4byte	};;
+{ .mmb;	$LDW	X[ 7]=[input],4*$SZ
+	$LDW	X[ 6]=[r8],4*$SZ
+(p13)	br.cond.dpnt.many	.L5byte	};;
+{ .mmb;	$LDW	X[ 5]=[r9],4*$SZ
+	$LDW	X[ 4]=[r10],4*$SZ
+(p14)	br.cond.dpnt.many	.L6byte	};;
+{ .mmb;	$LDW	X[ 3]=[input],4*$SZ
+	$LDW	X[ 2]=[r8],4*$SZ
+(p15)	br.cond.dpnt.many	.L7byte	};;
+{ .mmb;	$LDW	X[ 1]=[r9],4*$SZ
+	$LDW	X[ 0]=[r10],4*$SZ
+	br.many	.L_first16		};;
+.L1byte:
+{ .mmi;	$LDW	X[13]=[r9],4*$SZ
+	$LDW	X[12]=[r10],4*$SZ
+	shrp	X[15]=X[15],X[14],56	};;
+{ .mmi;	$LDW	X[11]=[input],4*$SZ
+	$LDW	X[10]=[r8],4*$SZ
+	shrp	X[14]=X[14],X[13],56	}
+{ .mmi;	$LDW	X[ 9]=[r9],4*$SZ
+	$LDW	X[ 8]=[r10],4*$SZ
+	shrp	X[13]=X[13],X[12],56	};;
+{ .mmi;	$LDW	X[ 7]=[input],4*$SZ
+	$LDW	X[ 6]=[r8],4*$SZ
+	shrp	X[12]=X[12],X[11],56	}
+{ .mmi;	$LDW	X[ 5]=[r9],4*$SZ
+	$LDW	X[ 4]=[r10],4*$SZ
+	shrp	X[11]=X[11],X[10],56	};;
+{ .mmi;	$LDW	X[ 3]=[input],4*$SZ
+	$LDW	X[ 2]=[r8],4*$SZ
+	shrp	X[10]=X[10],X[ 9],56	}
+{ .mmi;	$LDW	X[ 1]=[r9],4*$SZ
+	$LDW	X[ 0]=[r10],4*$SZ
+	shrp	X[ 9]=X[ 9],X[ 8],56	};;
+{ .mii;	$LDW	T1=[input]
+	shrp	X[ 8]=X[ 8],X[ 7],56
+	shrp	X[ 7]=X[ 7],X[ 6],56	}
+{ .mii;	shrp	X[ 6]=X[ 6],X[ 5],56
+	shrp	X[ 5]=X[ 5],X[ 4],56	};;
+{ .mii;	shrp	X[ 4]=X[ 4],X[ 3],56
+	shrp	X[ 3]=X[ 3],X[ 2],56	}
+{ .mii;	shrp	X[ 2]=X[ 2],X[ 1],56
+	shrp	X[ 1]=X[ 1],X[ 0],56	}
+{ .mib;	shrp	X[ 0]=X[ 0],T1,56
+	br.many	.L_first16		};;
+.L2byte:
+{ .mmi;	$LDW	X[11]=[input],4*$SZ
+	$LDW	X[10]=[r8],4*$SZ
+	shrp	X[15]=X[15],X[14],48	}
+{ .mmi;	$LDW	X[ 9]=[r9],4*$SZ
+	$LDW	X[ 8]=[r10],4*$SZ
+	shrp	X[14]=X[14],X[13],48	};;
+{ .mmi;	$LDW	X[ 7]=[input],4*$SZ
+	$LDW	X[ 6]=[r8],4*$SZ
+	shrp	X[13]=X[13],X[12],48	}
+{ .mmi;	$LDW	X[ 5]=[r9],4*$SZ
+	$LDW	X[ 4]=[r10],4*$SZ
+	shrp	X[12]=X[12],X[11],48	};;
+{ .mmi;	$LDW	X[ 3]=[input],4*$SZ
+	$LDW	X[ 2]=[r8],4*$SZ
+	shrp	X[11]=X[11],X[10],48	}
+{ .mmi;	$LDW	X[ 1]=[r9],4*$SZ
+	$LDW	X[ 0]=[r10],4*$SZ
+	shrp	X[10]=X[10],X[ 9],48	};;
+{ .mii;	$LDW	T1=[input]
+	shrp	X[ 9]=X[ 9],X[ 8],48
+	shrp	X[ 8]=X[ 8],X[ 7],48	}
+{ .mii;	shrp	X[ 7]=X[ 7],X[ 6],48
+	shrp	X[ 6]=X[ 6],X[ 5],48	};;
+{ .mii;	shrp	X[ 5]=X[ 5],X[ 4],48
+	shrp	X[ 4]=X[ 4],X[ 3],48	}
+{ .mii;	shrp	X[ 3]=X[ 3],X[ 2],48
+	shrp	X[ 2]=X[ 2],X[ 1],48	}
+{ .mii;	shrp	X[ 1]=X[ 1],X[ 0],48
+	shrp	X[ 0]=X[ 0],T1,48	}
+{ .mfb;	br.many	.L_first16		};;
+.L3byte:
+{ .mmi;	$LDW	X[ 9]=[r9],4*$SZ
+	$LDW	X[ 8]=[r10],4*$SZ
+	shrp	X[15]=X[15],X[14],40	};;
+{ .mmi;	$LDW	X[ 7]=[input],4*$SZ
+	$LDW	X[ 6]=[r8],4*$SZ
+	shrp	X[14]=X[14],X[13],40	}
+{ .mmi;	$LDW	X[ 5]=[r9],4*$SZ
+	$LDW	X[ 4]=[r10],4*$SZ
+	shrp	X[13]=X[13],X[12],40	};;
+{ .mmi;	$LDW	X[ 3]=[input],4*$SZ
+	$LDW	X[ 2]=[r8],4*$SZ
+	shrp	X[12]=X[12],X[11],40	}
+{ .mmi;	$LDW	X[ 1]=[r9],4*$SZ
+	$LDW	X[ 0]=[r10],4*$SZ
+	shrp	X[11]=X[11],X[10],40	};;
+{ .mii;	$LDW	T1=[input]
+	shrp	X[10]=X[10],X[ 9],40
+	shrp	X[ 9]=X[ 9],X[ 8],40	}
+{ .mii;	shrp	X[ 8]=X[ 8],X[ 7],40
+	shrp	X[ 7]=X[ 7],X[ 6],40	};;
+{ .mii;	shrp	X[ 6]=X[ 6],X[ 5],40
+	shrp	X[ 5]=X[ 5],X[ 4],40	}
+{ .mii;	shrp	X[ 4]=X[ 4],X[ 3],40
+	shrp	X[ 3]=X[ 3],X[ 2],40	}
+{ .mii;	shrp	X[ 2]=X[ 2],X[ 1],40
+	shrp	X[ 1]=X[ 1],X[ 0],40	}
+{ .mib;	shrp	X[ 0]=X[ 0],T1,40
+	br.many	.L_first16		};;
+.L4byte:
+{ .mmi;	$LDW	X[ 7]=[input],4*$SZ
+	$LDW	X[ 6]=[r8],4*$SZ
+	shrp	X[15]=X[15],X[14],32	}
+{ .mmi;	$LDW	X[ 5]=[r9],4*$SZ
+	$LDW	X[ 4]=[r10],4*$SZ
+	shrp	X[14]=X[14],X[13],32	};;
+{ .mmi;	$LDW	X[ 3]=[input],4*$SZ
+	$LDW	X[ 2]=[r8],4*$SZ
+	shrp	X[13]=X[13],X[12],32	}
+{ .mmi;	$LDW	X[ 1]=[r9],4*$SZ
+	$LDW	X[ 0]=[r10],4*$SZ
+	shrp	X[12]=X[12],X[11],32	};;
+{ .mii;	$LDW	T1=[input]
+	shrp	X[11]=X[11],X[10],32
+	shrp	X[10]=X[10],X[ 9],32	}
+{ .mii;	shrp	X[ 9]=X[ 9],X[ 8],32
+	shrp	X[ 8]=X[ 8],X[ 7],32	};;
+{ .mii;	shrp	X[ 7]=X[ 7],X[ 6],32
+	shrp	X[ 6]=X[ 6],X[ 5],32	}
+{ .mii;	shrp	X[ 5]=X[ 5],X[ 4],32
+	shrp	X[ 4]=X[ 4],X[ 3],32	}
+{ .mii;	shrp	X[ 3]=X[ 3],X[ 2],32
+	shrp	X[ 2]=X[ 2],X[ 1],32	}
+{ .mii;	shrp	X[ 1]=X[ 1],X[ 0],32
+	shrp	X[ 0]=X[ 0],T1,32	}
+{ .mfb;	br.many	.L_first16		};;
+.L5byte:
+{ .mmi;	$LDW	X[ 5]=[r9],4*$SZ
+	$LDW	X[ 4]=[r10],4*$SZ
+	shrp	X[15]=X[15],X[14],24	};;
+{ .mmi;	$LDW	X[ 3]=[input],4*$SZ
+	$LDW	X[ 2]=[r8],4*$SZ
+	shrp	X[14]=X[14],X[13],24	}
+{ .mmi;	$LDW	X[ 1]=[r9],4*$SZ
+	$LDW	X[ 0]=[r10],4*$SZ
+	shrp	X[13]=X[13],X[12],24	};;
+{ .mii;	$LDW	T1=[input]
+	shrp	X[12]=X[12],X[11],24
+	shrp	X[11]=X[11],X[10],24	}
+{ .mii;	shrp	X[10]=X[10],X[ 9],24
+	shrp	X[ 9]=X[ 9],X[ 8],24	};;
+{ .mii;	shrp	X[ 8]=X[ 8],X[ 7],24
+	shrp	X[ 7]=X[ 7],X[ 6],24	}
+{ .mii;	shrp	X[ 6]=X[ 6],X[ 5],24
+	shrp	X[ 5]=X[ 5],X[ 4],24	}
+{ .mii;	shrp	X[ 4]=X[ 4],X[ 3],24
+	shrp	X[ 3]=X[ 3],X[ 2],24	}
+{ .mii;	shrp	X[ 2]=X[ 2],X[ 1],24
+	shrp	X[ 1]=X[ 1],X[ 0],24	}
+{ .mib;	shrp	X[ 0]=X[ 0],T1,24
+	br.many	.L_first16		};;
+.L6byte:
+{ .mmi;	$LDW	X[ 3]=[input],4*$SZ
+	$LDW	X[ 2]=[r8],4*$SZ
+	shrp	X[15]=X[15],X[14],16	}
+{ .mmi;	$LDW	X[ 1]=[r9],4*$SZ
+	$LDW	X[ 0]=[r10],4*$SZ
+	shrp	X[14]=X[14],X[13],16	};;
+{ .mii;	$LDW	T1=[input]
+	shrp	X[13]=X[13],X[12],16
+	shrp	X[12]=X[12],X[11],16	}
+{ .mii;	shrp	X[11]=X[11],X[10],16
+	shrp	X[10]=X[10],X[ 9],16	};;
+{ .mii;	shrp	X[ 9]=X[ 9],X[ 8],16
+	shrp	X[ 8]=X[ 8],X[ 7],16	}
+{ .mii;	shrp	X[ 7]=X[ 7],X[ 6],16
+	shrp	X[ 6]=X[ 6],X[ 5],16	}
+{ .mii;	shrp	X[ 5]=X[ 5],X[ 4],16
+	shrp	X[ 4]=X[ 4],X[ 3],16	}
+{ .mii;	shrp	X[ 3]=X[ 3],X[ 2],16
+	shrp	X[ 2]=X[ 2],X[ 1],16	}
+{ .mii;	shrp	X[ 1]=X[ 1],X[ 0],16
+	shrp	X[ 0]=X[ 0],T1,16	}
+{ .mfb;	br.many	.L_first16		};;
+.L7byte:
+{ .mmi;	$LDW	X[ 1]=[r9],4*$SZ
+	$LDW	X[ 0]=[r10],4*$SZ
+	shrp	X[15]=X[15],X[14],8	};;
+{ .mii;	$LDW	T1=[input]
+	shrp	X[14]=X[14],X[13],8
+	shrp	X[13]=X[13],X[12],8	}
+{ .mii;	shrp	X[12]=X[12],X[11],8
+	shrp	X[11]=X[11],X[10],8	};;
+{ .mii;	shrp	X[10]=X[10],X[ 9],8
+	shrp	X[ 9]=X[ 9],X[ 8],8	}
+{ .mii;	shrp	X[ 8]=X[ 8],X[ 7],8
+	shrp	X[ 7]=X[ 7],X[ 6],8	}
+{ .mii;	shrp	X[ 6]=X[ 6],X[ 5],8
+	shrp	X[ 5]=X[ 5],X[ 4],8	}
+{ .mii;	shrp	X[ 4]=X[ 4],X[ 3],8
+	shrp	X[ 3]=X[ 3],X[ 2],8	}
+{ .mii;	shrp	X[ 2]=X[ 2],X[ 1],8
+	shrp	X[ 1]=X[ 1],X[ 0],8	}
+{ .mib;	shrp	X[ 0]=X[ 0],T1,8
+	br.many	.L_first16		};;
+
+.align	32
+.L_first16:
+{ .mmi;		$LDW	K=[Ktbl],$SZ
+		and	T1=F,E
+		and	T2=A,B		}
+{ .mmi;		//$LDW	X[15]=[input],$SZ	// X[i]=*input++
+		andcm	r8=G,E
+		and	r9=A,C		};;
+{ .mmi;		xor	T1=T1,r8		//T1=((e & f) ^ (~e & g))
+		and	r10=B,C
+		_rotr	r11=$t1,$Sigma1[0] }	// ROTR(e,14)
+{ .mmi;		xor	T2=T2,r9
+		mux1	X[15]=X[15],\@rev };;	// eliminated in big-endian
+___
+$code.=<<___;
+{ .mib;		add	T1=T1,H			// T1=Ch(e,f,g)+h
+		_rotr	r8=$t1,$Sigma1[1] }	// ROTR(e,18)
+{ .mib;		xor	T2=T2,r10		// T2=((a & b) ^ (a & c) ^ (b & c))
+		mov	H=G		};;
+{ .mib;		xor	r11=r8,r11
+		_rotr	r9=$t1,$Sigma1[2] }	// ROTR(e,41)
+{ .mib;		mov	G=F
+		mov	F=E		};;
+{ .mib;		xor	r9=r9,r11		// r9=Sigma1(e)
+		_rotr	r10=$t0,$Sigma0[0] }	// ROTR(a,28)
+{ .mib;		add	T1=T1,K			// T1=Ch(e,f,g)+h+K512[i]
+		mov	E=D		};;
+{ .mib;		add	T1=T1,r9		// T1+=Sigma1(e)
+		_rotr	r11=$t0,$Sigma0[1] }	// ROTR(a,34)
+{ .mib;		mov	D=C
+		mov	C=B		};;
+{ .mib;		add	T1=T1,X[15]		// T1+=X[i]
+		_rotr	r8=$t0,$Sigma0[2] }	// ROTR(a,39)
+{ .mib;		xor	r10=r10,r11
+		mux2	X[15]=X[15],0x44 };;	// eliminated in 64-bit
+{ .mmi;		xor	r10=r8,r10		// r10=Sigma0(a)
+		mov	B=A
+		add	A=T1,T2		};;
+{ .mib;		add	E=E,T1
+		add	A=A,r10			// T2=Maj(a,b,c)+Sigma0(a)
+	br.ctop.sptk	.L_first16	};;
+.L_first16_end:
+
+{ .mii;	mov	ar.lc=$rounds-17
+	mov	ar.ec=1			};;
+
+.align	32
+.L_rest:
+.rotr	X[16]
+{ .mib;		$LDW	K=[Ktbl],$SZ
+		_rotr	r8=X[15-1],$sigma0[0] }	// ROTR(s0,1)
+{ .mib; 	$ADD	X[15]=X[15],X[15-9]	// X[i&0xF]+=X[(i+9)&0xF]
+		$SHRU	s0=X[15-1],sgm0	};;	// s0=X[(i+1)&0xF]>>7
+{ .mib;		and	T1=F,E
+		_rotr	r9=X[15-1],$sigma0[1] }	// ROTR(s0,8)
+{ .mib;		andcm	r10=G,E
+		$SHRU	s1=X[15-14],sgm1 };;	// s1=X[(i+14)&0xF]>>6
+{ .mmi;		xor	T1=T1,r10		// T1=((e & f) ^ (~e & g))
+		xor	r9=r8,r9
+		_rotr	r10=X[15-14],$sigma1[0] };;// ROTR(s1,19)
+{ .mib;		and	T2=A,B		
+		_rotr	r11=X[15-14],$sigma1[1] }// ROTR(s1,61)
+{ .mib;		and	r8=A,C		};;
+___
+$t0="t0", $t1="t1", $code.=<<___ if ($BITS==32);
+// I adhere to mmi; in order to hold Itanium 1 back and avoid 6 cycle
+// pipeline flush in last bundle. Note that even on Itanium2 the
+// latter stalls for one clock cycle...
+{ .mmi;		xor	s0=s0,r9		// s0=sigma0(X[(i+1)&0xF])
+		dep.z	$t1=E,32,32	}
+{ .mmi;		xor	r10=r11,r10
+		zxt4	E=E		};;
+{ .mmi;		or	$t1=$t1,E
+		xor	s1=s1,r10		// s1=sigma1(X[(i+14)&0xF])
+		mux2	$t0=A,0x44	};;	// copy lower half to upper
+{ .mmi;		xor	T2=T2,r8
+		_rotr	r9=$t1,$Sigma1[0] }	// ROTR(e,14)
+{ .mmi;		and	r10=B,C
+		add	T1=T1,H			// T1=Ch(e,f,g)+h
+		$ADD	X[15]=X[15],s0	};;	// X[i&0xF]+=sigma0(X[(i+1)&0xF])
+___
+$t0="A", $t1="E", $code.=<<___ if ($BITS==64);
+{ .mib;		xor	s0=s0,r9		// s0=sigma0(X[(i+1)&0xF])
+		_rotr	r9=$t1,$Sigma1[0] }	// ROTR(e,14)
+{ .mib;		xor	r10=r11,r10
+		xor	T2=T2,r8	};;
+{ .mib;		xor	s1=s1,r10		// s1=sigma1(X[(i+14)&0xF])
+		add	T1=T1,H		}
+{ .mib;		and	r10=B,C
+		$ADD	X[15]=X[15],s0	};;	// X[i&0xF]+=sigma0(X[(i+1)&0xF])
+___
+$code.=<<___;
+{ .mmi;		xor	T2=T2,r10		// T2=((a & b) ^ (a & c) ^ (b & c))
+		mov	H=G
+		_rotr	r8=$t1,$Sigma1[1] };;	// ROTR(e,18)
+{ .mmi;		xor	r11=r8,r9
+		$ADD	X[15]=X[15],s1		// X[i&0xF]+=sigma1(X[(i+14)&0xF])
+		_rotr	r9=$t1,$Sigma1[2] }	// ROTR(e,41)
+{ .mmi;		mov	G=F
+		mov	F=E		};;
+{ .mib;		xor	r9=r9,r11		// r9=Sigma1(e)
+		_rotr	r10=$t0,$Sigma0[0] }	// ROTR(a,28)
+{ .mib;		add	T1=T1,K			// T1=Ch(e,f,g)+h+K512[i]
+		mov	E=D		};;
+{ .mib;		add	T1=T1,r9		// T1+=Sigma1(e)
+		_rotr	r11=$t0,$Sigma0[1] }	// ROTR(a,34)
+{ .mib;		mov	D=C
+		mov	C=B		};;
+{ .mmi;		add	T1=T1,X[15]		// T1+=X[i]
+		xor	r10=r10,r11
+		_rotr	r8=$t0,$Sigma0[2] };;	// ROTR(a,39)
+{ .mmi;		xor	r10=r8,r10		// r10=Sigma0(a)
+		mov	B=A
+		add	A=T1,T2		};;
+{ .mib;		add	E=E,T1
+		add	A=A,r10			// T2=Maj(a,b,c)+Sigma0(a)
+	br.ctop.sptk	.L_rest	};;
+.L_rest_end:
+
+{ .mmi;	add	A_=A_,A
+	add	B_=B_,B
+	add	C_=C_,C			}
+{ .mmi;	add	D_=D_,D
+	add	E_=E_,E
+	cmp.ltu	p16,p0=1,num		};;
+{ .mmi;	add	F_=F_,F
+	add	G_=G_,G
+	add	H_=H_,H			}
+{ .mmb;	add	Ktbl=-$SZ*$rounds,Ktbl
+(p16)	add	num=-1,num
+(p16)	br.dptk.many	.L_outer	};;
+
+{ .mib;	add	r8=0*$SZ,ctx
+	add	r9=1*$SZ,ctx		}
+{ .mib;	add	r10=2*$SZ,ctx
+	add	r11=3*$SZ,ctx		};;
+{ .mmi;	$STW	[r8]=A_,4*$SZ
+	$STW	[r9]=B_,4*$SZ
+	mov	ar.lc=lcsave		}
+{ .mmi;	$STW	[r10]=C_,4*$SZ
+	$STW	[r11]=D_,4*$SZ
+	mov	pr=prsave,0x1ffff	};;
+{ .mmb;	$STW	[r8]=E_
+	$STW	[r9]=F_			}
+{ .mmb;	$STW	[r10]=G_
+	$STW	[r11]=H_
+	br.ret.sptk.many	b0	};;
+.endp	$func#
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+$code =~ s/_rotr(\s+)([^=]+)=([^,]+),([0-9]+)/shrp$1$2=$3,$3,$4/gm;
+if ($BITS==64) {
+    $code =~ s/mux2(\s+)\S+/nop.i$1 0x0/gm;
+    $code =~ s/mux1(\s+)\S+/nop.i$1 0x0/gm	if ($big_endian);
+    $code =~ s/(shrp\s+X\[[^=]+)=([^,]+),([^,]+),([1-9]+)/$1=$3,$2,64-$4/gm
+    						if (!$big_endian);
+    $code =~ s/ld1(\s+)X\[\S+/nop.m$1 0x0/gm;
+}
+
+print $code;
+
+print<<___ if ($BITS==32);
+.align	64
+.type	K256#,\@object
+K256:	data4	0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5
+	data4	0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5
+	data4	0xd807aa98,0x12835b01,0x243185be,0x550c7dc3
+	data4	0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174
+	data4	0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc
+	data4	0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da
+	data4	0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7
+	data4	0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967
+	data4	0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13
+	data4	0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85
+	data4	0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3
+	data4	0xd192e819,0xd6990624,0xf40e3585,0x106aa070
+	data4	0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5
+	data4	0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3
+	data4	0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208
+	data4	0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
+.size	K256#,$SZ*$rounds
+stringz	"SHA256 block transform for IA64, CRYPTOGAMS by <appro\@openssl.org>"
+___
+print<<___ if ($BITS==64);
+.align	64
+.type	K512#,\@object
+K512:	data8	0x428a2f98d728ae22,0x7137449123ef65cd
+	data8	0xb5c0fbcfec4d3b2f,0xe9b5dba58189dbbc
+	data8	0x3956c25bf348b538,0x59f111f1b605d019
+	data8	0x923f82a4af194f9b,0xab1c5ed5da6d8118
+	data8	0xd807aa98a3030242,0x12835b0145706fbe
+	data8	0x243185be4ee4b28c,0x550c7dc3d5ffb4e2
+	data8	0x72be5d74f27b896f,0x80deb1fe3b1696b1
+	data8	0x9bdc06a725c71235,0xc19bf174cf692694
+	data8	0xe49b69c19ef14ad2,0xefbe4786384f25e3
+	data8	0x0fc19dc68b8cd5b5,0x240ca1cc77ac9c65
+	data8	0x2de92c6f592b0275,0x4a7484aa6ea6e483
+	data8	0x5cb0a9dcbd41fbd4,0x76f988da831153b5
+	data8	0x983e5152ee66dfab,0xa831c66d2db43210
+	data8	0xb00327c898fb213f,0xbf597fc7beef0ee4
+	data8	0xc6e00bf33da88fc2,0xd5a79147930aa725
+	data8	0x06ca6351e003826f,0x142929670a0e6e70
+	data8	0x27b70a8546d22ffc,0x2e1b21385c26c926
+	data8	0x4d2c6dfc5ac42aed,0x53380d139d95b3df
+	data8	0x650a73548baf63de,0x766a0abb3c77b2a8
+	data8	0x81c2c92e47edaee6,0x92722c851482353b
+	data8	0xa2bfe8a14cf10364,0xa81a664bbc423001
+	data8	0xc24b8b70d0f89791,0xc76c51a30654be30
+	data8	0xd192e819d6ef5218,0xd69906245565a910
+	data8	0xf40e35855771202a,0x106aa07032bbd1b8
+	data8	0x19a4c116b8d2d0c8,0x1e376c085141ab53
+	data8	0x2748774cdf8eeb99,0x34b0bcb5e19b48a8
+	data8	0x391c0cb3c5c95a63,0x4ed8aa4ae3418acb
+	data8	0x5b9cca4f7763e373,0x682e6ff3d6b2b8a3
+	data8	0x748f82ee5defb2fc,0x78a5636f43172f60
+	data8	0x84c87814a1f0ab72,0x8cc702081a6439ec
+	data8	0x90befffa23631e28,0xa4506cebde82bde9
+	data8	0xbef9a3f7b2c67915,0xc67178f2e372532b
+	data8	0xca273eceea26619c,0xd186b8c721c0c207
+	data8	0xeada7dd6cde0eb1e,0xf57d4f7fee6ed178
+	data8	0x06f067aa72176fba,0x0a637dc5a2c898a6
+	data8	0x113f9804bef90dae,0x1b710b35131c471b
+	data8	0x28db77f523047d84,0x32caab7b40c72493
+	data8	0x3c9ebe0a15c9bebc,0x431d67c49c100d4c
+	data8	0x4cc5d4becb3e42b6,0x597f299cfc657e2a
+	data8	0x5fcb6fab3ad6faec,0x6c44198c4a475817
+.size	K512#,$SZ*$rounds
+stringz	"SHA512 block transform for IA64, CRYPTOGAMS by <appro\@openssl.org>"
+___
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha512-mips.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha512-mips.pl
new file mode 100644
index 0000000000..ffa053bb7d
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha512-mips.pl
@@ -0,0 +1,455 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# SHA2 block procedures for MIPS.
+
+# October 2010.
+#
+# SHA256 performance improvement on MIPS R5000 CPU is ~27% over gcc-
+# generated code in o32 build and ~55% in n32/64 build. SHA512 [which
+# for now can only be compiled for MIPS64 ISA] improvement is modest
+# ~17%, but it comes for free, because it's same instruction sequence.
+# Improvement coefficients are for aligned input.
+
+######################################################################
+# There is a number of MIPS ABI in use, O32 and N32/64 are most
+# widely used. Then there is a new contender: NUBI. It appears that if
+# one picks the latter, it's possible to arrange code in ABI neutral
+# manner. Therefore let's stick to NUBI register layout:
+#
+($zero,$at,$t0,$t1,$t2)=map("\$$_",(0..2,24,25));
+($a0,$a1,$a2,$a3,$a4,$a5,$a6,$a7)=map("\$$_",(4..11));
+($s0,$s1,$s2,$s3,$s4,$s5,$s6,$s7,$s8,$s9,$s10,$s11)=map("\$$_",(12..23));
+($gp,$tp,$sp,$fp,$ra)=map("\$$_",(3,28..31));
+#
+# The return value is placed in $a0. Following coding rules facilitate
+# interoperability:
+#
+# - never ever touch $tp, "thread pointer", former $gp [o32 can be
+#   excluded from the rule, because it's specified volatile];
+# - copy return value to $t0, former $v0 [or to $a0 if you're adapting
+#   old code];
+# - on O32 populate $a4-$a7 with 'lw $aN,4*N($sp)' if necessary;
+#
+# For reference here is register layout for N32/64 MIPS ABIs:
+#
+# ($zero,$at,$v0,$v1)=map("\$$_",(0..3));
+# ($a0,$a1,$a2,$a3,$a4,$a5,$a6,$a7)=map("\$$_",(4..11));
+# ($t0,$t1,$t2,$t3,$t8,$t9)=map("\$$_",(12..15,24,25));
+# ($s0,$s1,$s2,$s3,$s4,$s5,$s6,$s7)=map("\$$_",(16..23));
+# ($gp,$sp,$fp,$ra)=map("\$$_",(28..31));
+#
+$flavour = shift; # supported flavours are o32,n32,64,nubi32,nubi64
+
+if ($flavour =~ /64|n32/i) {
+	$PTR_ADD="dadd";	# incidentally works even on n32
+	$PTR_SUB="dsub";	# incidentally works even on n32
+	$REG_S="sd";
+	$REG_L="ld";
+	$PTR_SLL="dsll";	# incidentally works even on n32
+	$SZREG=8;
+} else {
+	$PTR_ADD="add";
+	$PTR_SUB="sub";
+	$REG_S="sw";
+	$REG_L="lw";
+	$PTR_SLL="sll";
+	$SZREG=4;
+}
+$pf = ($flavour =~ /nubi/i) ? $t0 : $t2;
+#
+# <appro@openssl.org>
+#
+######################################################################
+
+$big_endian=(`echo MIPSEL | $ENV{CC} -E -P -`=~/MIPSEL/)?1:0;
+
+for (@ARGV) {	$output=$_ if (/^\w[\w\-]*\.\w+$/);	}
+open STDOUT,">$output";
+
+if (!defined($big_endian)) { $big_endian=(unpack('L',pack('N',1))==1); }
+
+if ($output =~ /512/) {
+	$label="512";
+	$SZ=8;
+	$LD="ld";		# load from memory
+	$ST="sd";		# store to memory
+	$SLL="dsll";		# shift left logical
+	$SRL="dsrl";		# shift right logical
+	$ADDU="daddu";
+	@Sigma0=(28,34,39);
+	@Sigma1=(14,18,41);
+	@sigma0=( 7, 1, 8);	# right shift first
+	@sigma1=( 6,19,61);	# right shift first
+	$lastK=0x817;
+	$rounds=80;
+} else {
+	$label="256";
+	$SZ=4;
+	$LD="lw";		# load from memory
+	$ST="sw";		# store to memory
+	$SLL="sll";		# shift left logical
+	$SRL="srl";		# shift right logical
+	$ADDU="addu";
+	@Sigma0=( 2,13,22);
+	@Sigma1=( 6,11,25);
+	@sigma0=( 3, 7,18);	# right shift first
+	@sigma1=(10,17,19);	# right shift first
+	$lastK=0x8f2;
+	$rounds=64;
+}
+
+$MSB = $big_endian ? 0 : ($SZ-1);
+$LSB = ($SZ-1)&~$MSB;
+
+@V=($A,$B,$C,$D,$E,$F,$G,$H)=map("\$$_",(1,2,3,7,24,25,30,31));
+@X=map("\$$_",(8..23));
+
+$ctx=$a0;
+$inp=$a1;
+$len=$a2;	$Ktbl=$len;
+
+sub BODY_00_15 {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_;
+my ($T1,$tmp0,$tmp1,$tmp2)=(@X[4],@X[5],@X[6],@X[7]);
+
+$code.=<<___ if ($i<15);
+	${LD}l	@X[1],`($i+1)*$SZ+$MSB`($inp)
+	${LD}r	@X[1],`($i+1)*$SZ+$LSB`($inp)
+___
+$code.=<<___	if (!$big_endian && $i<16 && $SZ==4);
+	srl	$tmp0,@X[0],24		# byte swap($i)
+	srl	$tmp1,@X[0],8
+	andi	$tmp2,@X[0],0xFF00
+	sll	@X[0],@X[0],24
+	andi	$tmp1,0xFF00
+	sll	$tmp2,$tmp2,8
+	or	@X[0],$tmp0
+	or	$tmp1,$tmp2
+	or	@X[0],$tmp1
+___
+$code.=<<___	if (!$big_endian && $i<16 && $SZ==8);
+	ori	$tmp0,$zero,0xFF
+	dsll	$tmp2,$tmp0,32
+	or	$tmp0,$tmp2		# 0x000000FF000000FF
+	and	$tmp1,@X[0],$tmp0	# byte swap($i)
+	dsrl	$tmp2,@X[0],24
+	dsll	$tmp1,24
+	and	$tmp2,$tmp0
+	dsll	$tmp0,8			# 0x0000FF000000FF00
+	or	$tmp1,$tmp2
+	and	$tmp2,@X[0],$tmp0
+	dsrl	@X[0],8
+	dsll	$tmp2,8
+	and	@X[0],$tmp0
+	or	$tmp1,$tmp2
+	or	@X[0],$tmp1
+	dsrl	$tmp1,@X[0],32
+	dsll	@X[0],32
+	or	@X[0],$tmp1
+___
+$code.=<<___;
+	$ADDU	$T1,$X[0],$h			# $i
+	$SRL	$h,$e,@Sigma1[0]
+	xor	$tmp2,$f,$g
+	$SLL	$tmp1,$e,`$SZ*8-@Sigma1[2]`
+	and	$tmp2,$e
+	$SRL	$tmp0,$e,@Sigma1[1]
+	xor	$h,$tmp1
+	$SLL	$tmp1,$e,`$SZ*8-@Sigma1[1]`
+	xor	$h,$tmp0
+	$SRL	$tmp0,$e,@Sigma1[2]
+	xor	$h,$tmp1
+	$SLL	$tmp1,$e,`$SZ*8-@Sigma1[0]`
+	xor	$h,$tmp0
+	xor	$tmp2,$g			# Ch(e,f,g)
+	xor	$tmp0,$tmp1,$h			# Sigma1(e)
+
+	$SRL	$h,$a,@Sigma0[0]
+	$ADDU	$T1,$tmp2
+	$LD	$tmp2,`$i*$SZ`($Ktbl)		# K[$i]
+	$SLL	$tmp1,$a,`$SZ*8-@Sigma0[2]`
+	$ADDU	$T1,$tmp0
+	$SRL	$tmp0,$a,@Sigma0[1]
+	xor	$h,$tmp1
+	$SLL	$tmp1,$a,`$SZ*8-@Sigma0[1]`
+	xor	$h,$tmp0
+	$SRL	$tmp0,$a,@Sigma0[2]
+	xor	$h,$tmp1
+	$SLL	$tmp1,$a,`$SZ*8-@Sigma0[0]`
+	xor	$h,$tmp0
+	$ST	@X[0],`($i%16)*$SZ`($sp)	# offload to ring buffer
+	xor	$h,$tmp1			# Sigma0(a)
+
+	or	$tmp0,$a,$b
+	and	$tmp1,$a,$b
+	and	$tmp0,$c
+	or	$tmp1,$tmp0			# Maj(a,b,c)
+	$ADDU	$T1,$tmp2			# +=K[$i]
+	$ADDU	$h,$tmp1
+
+	$ADDU	$d,$T1
+	$ADDU	$h,$T1
+___
+$code.=<<___ if ($i>=13);
+	$LD	@X[3],`(($i+3)%16)*$SZ`($sp)	# prefetch from ring buffer
+___
+}
+
+sub BODY_16_XX {
+my $i=@_[0];
+my ($tmp0,$tmp1,$tmp2,$tmp3)=(@X[4],@X[5],@X[6],@X[7]);
+
+$code.=<<___;
+	$SRL	$tmp2,@X[1],@sigma0[0]		# Xupdate($i)
+	$ADDU	@X[0],@X[9]			# +=X[i+9]
+	$SLL	$tmp1,@X[1],`$SZ*8-@sigma0[2]`
+	$SRL	$tmp0,@X[1],@sigma0[1]
+	xor	$tmp2,$tmp1
+	$SLL	$tmp1,`@sigma0[2]-@sigma0[1]`
+	xor	$tmp2,$tmp0
+	$SRL	$tmp0,@X[1],@sigma0[2]
+	xor	$tmp2,$tmp1
+
+	$SRL	$tmp3,@X[14],@sigma1[0]
+	xor	$tmp2,$tmp0			# sigma0(X[i+1])
+	$SLL	$tmp1,@X[14],`$SZ*8-@sigma1[2]`
+	$ADDU	@X[0],$tmp2
+	$SRL	$tmp0,@X[14],@sigma1[1]
+	xor	$tmp3,$tmp1
+	$SLL	$tmp1,`@sigma1[2]-@sigma1[1]`
+	xor	$tmp3,$tmp0
+	$SRL	$tmp0,@X[14],@sigma1[2]
+	xor	$tmp3,$tmp1
+
+	xor	$tmp3,$tmp0			# sigma1(X[i+14])
+	$ADDU	@X[0],$tmp3
+___
+	&BODY_00_15(@_);
+}
+
+$FRAMESIZE=16*$SZ+16*$SZREG;
+$SAVED_REGS_MASK = ($flavour =~ /nubi/i) ? 0xc0fff008 : 0xc0ff0000;
+
+$code.=<<___;
+#ifdef OPENSSL_FIPSCANISTER
+# include <openssl/fipssyms.h>
+#endif
+
+.text
+.set	noat
+#if !defined(__vxworks) || defined(__pic__)
+.option	pic2
+#endif
+
+.align	5
+.globl	sha${label}_block_data_order
+.ent	sha${label}_block_data_order
+sha${label}_block_data_order:
+	.frame	$sp,$FRAMESIZE,$ra
+	.mask	$SAVED_REGS_MASK,-$SZREG
+	.set	noreorder
+___
+$code.=<<___ if ($flavour =~ /o32/i);	# o32 PIC-ification
+	.cpload	$pf
+___
+$code.=<<___;
+	$PTR_SUB $sp,$FRAMESIZE
+	$REG_S	$ra,$FRAMESIZE-1*$SZREG($sp)
+	$REG_S	$fp,$FRAMESIZE-2*$SZREG($sp)
+	$REG_S	$s11,$FRAMESIZE-3*$SZREG($sp)
+	$REG_S	$s10,$FRAMESIZE-4*$SZREG($sp)
+	$REG_S	$s9,$FRAMESIZE-5*$SZREG($sp)
+	$REG_S	$s8,$FRAMESIZE-6*$SZREG($sp)
+	$REG_S	$s7,$FRAMESIZE-7*$SZREG($sp)
+	$REG_S	$s6,$FRAMESIZE-8*$SZREG($sp)
+	$REG_S	$s5,$FRAMESIZE-9*$SZREG($sp)
+	$REG_S	$s4,$FRAMESIZE-10*$SZREG($sp)
+___
+$code.=<<___ if ($flavour =~ /nubi/i);	# optimize non-nubi prologue
+	$REG_S	$s3,$FRAMESIZE-11*$SZREG($sp)
+	$REG_S	$s2,$FRAMESIZE-12*$SZREG($sp)
+	$REG_S	$s1,$FRAMESIZE-13*$SZREG($sp)
+	$REG_S	$s0,$FRAMESIZE-14*$SZREG($sp)
+	$REG_S	$gp,$FRAMESIZE-15*$SZREG($sp)
+___
+$code.=<<___;
+	$PTR_SLL @X[15],$len,`log(16*$SZ)/log(2)`
+___
+$code.=<<___ if ($flavour !~ /o32/i);	# non-o32 PIC-ification
+	.cplocal	$Ktbl
+	.cpsetup	$pf,$zero,sha${label}_block_data_order
+___
+$code.=<<___;
+	.set	reorder
+	la	$Ktbl,K${label}		# PIC-ified 'load address'
+
+	$LD	$A,0*$SZ($ctx)		# load context
+	$LD	$B,1*$SZ($ctx)
+	$LD	$C,2*$SZ($ctx)
+	$LD	$D,3*$SZ($ctx)
+	$LD	$E,4*$SZ($ctx)
+	$LD	$F,5*$SZ($ctx)
+	$LD	$G,6*$SZ($ctx)
+	$LD	$H,7*$SZ($ctx)
+
+	$PTR_ADD @X[15],$inp		# pointer to the end of input
+	$REG_S	@X[15],16*$SZ($sp)
+	b	.Loop
+
+.align	5
+.Loop:
+	${LD}l	@X[0],$MSB($inp)
+	${LD}r	@X[0],$LSB($inp)
+___
+for ($i=0;$i<16;$i++)
+{ &BODY_00_15($i,@V); unshift(@V,pop(@V)); push(@X,shift(@X)); }
+$code.=<<___;
+	b	.L16_xx
+.align	4
+.L16_xx:
+___
+for (;$i<32;$i++)
+{ &BODY_16_XX($i,@V); unshift(@V,pop(@V)); push(@X,shift(@X)); }
+$code.=<<___;
+	and	@X[6],0xfff
+	li	@X[7],$lastK
+	.set	noreorder
+	bne	@X[6],@X[7],.L16_xx
+	$PTR_ADD $Ktbl,16*$SZ		# Ktbl+=16
+
+	$REG_L	@X[15],16*$SZ($sp)	# restore pointer to the end of input
+	$LD	@X[0],0*$SZ($ctx)
+	$LD	@X[1],1*$SZ($ctx)
+	$LD	@X[2],2*$SZ($ctx)
+	$PTR_ADD $inp,16*$SZ
+	$LD	@X[3],3*$SZ($ctx)
+	$ADDU	$A,@X[0]
+	$LD	@X[4],4*$SZ($ctx)
+	$ADDU	$B,@X[1]
+	$LD	@X[5],5*$SZ($ctx)
+	$ADDU	$C,@X[2]
+	$LD	@X[6],6*$SZ($ctx)
+	$ADDU	$D,@X[3]
+	$LD	@X[7],7*$SZ($ctx)
+	$ADDU	$E,@X[4]
+	$ST	$A,0*$SZ($ctx)
+	$ADDU	$F,@X[5]
+	$ST	$B,1*$SZ($ctx)
+	$ADDU	$G,@X[6]
+	$ST	$C,2*$SZ($ctx)
+	$ADDU	$H,@X[7]
+	$ST	$D,3*$SZ($ctx)
+	$ST	$E,4*$SZ($ctx)
+	$ST	$F,5*$SZ($ctx)
+	$ST	$G,6*$SZ($ctx)
+	$ST	$H,7*$SZ($ctx)
+
+	bne	$inp,@X[15],.Loop
+	$PTR_SUB $Ktbl,`($rounds-16)*$SZ`	# rewind $Ktbl
+
+	$REG_L	$ra,$FRAMESIZE-1*$SZREG($sp)
+	$REG_L	$fp,$FRAMESIZE-2*$SZREG($sp)
+	$REG_L	$s11,$FRAMESIZE-3*$SZREG($sp)
+	$REG_L	$s10,$FRAMESIZE-4*$SZREG($sp)
+	$REG_L	$s9,$FRAMESIZE-5*$SZREG($sp)
+	$REG_L	$s8,$FRAMESIZE-6*$SZREG($sp)
+	$REG_L	$s7,$FRAMESIZE-7*$SZREG($sp)
+	$REG_L	$s6,$FRAMESIZE-8*$SZREG($sp)
+	$REG_L	$s5,$FRAMESIZE-9*$SZREG($sp)
+	$REG_L	$s4,$FRAMESIZE-10*$SZREG($sp)
+___
+$code.=<<___ if ($flavour =~ /nubi/i);
+	$REG_L	$s3,$FRAMESIZE-11*$SZREG($sp)
+	$REG_L	$s2,$FRAMESIZE-12*$SZREG($sp)
+	$REG_L	$s1,$FRAMESIZE-13*$SZREG($sp)
+	$REG_L	$s0,$FRAMESIZE-14*$SZREG($sp)
+	$REG_L	$gp,$FRAMESIZE-15*$SZREG($sp)
+___
+$code.=<<___;
+	jr	$ra
+	$PTR_ADD $sp,$FRAMESIZE
+.end	sha${label}_block_data_order
+
+.rdata
+.align	5
+K${label}:
+___
+if ($SZ==4) {
+$code.=<<___;
+	.word	0x428a2f98, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5
+	.word	0x3956c25b, 0x59f111f1, 0x923f82a4, 0xab1c5ed5
+	.word	0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3
+	.word	0x72be5d74, 0x80deb1fe, 0x9bdc06a7, 0xc19bf174
+	.word	0xe49b69c1, 0xefbe4786, 0x0fc19dc6, 0x240ca1cc
+	.word	0x2de92c6f, 0x4a7484aa, 0x5cb0a9dc, 0x76f988da
+	.word	0x983e5152, 0xa831c66d, 0xb00327c8, 0xbf597fc7
+	.word	0xc6e00bf3, 0xd5a79147, 0x06ca6351, 0x14292967
+	.word	0x27b70a85, 0x2e1b2138, 0x4d2c6dfc, 0x53380d13
+	.word	0x650a7354, 0x766a0abb, 0x81c2c92e, 0x92722c85
+	.word	0xa2bfe8a1, 0xa81a664b, 0xc24b8b70, 0xc76c51a3
+	.word	0xd192e819, 0xd6990624, 0xf40e3585, 0x106aa070
+	.word	0x19a4c116, 0x1e376c08, 0x2748774c, 0x34b0bcb5
+	.word	0x391c0cb3, 0x4ed8aa4a, 0x5b9cca4f, 0x682e6ff3
+	.word	0x748f82ee, 0x78a5636f, 0x84c87814, 0x8cc70208
+	.word	0x90befffa, 0xa4506ceb, 0xbef9a3f7, 0xc67178f2
+___
+} else {
+$code.=<<___;
+	.dword	0x428a2f98d728ae22, 0x7137449123ef65cd
+	.dword	0xb5c0fbcfec4d3b2f, 0xe9b5dba58189dbbc
+	.dword	0x3956c25bf348b538, 0x59f111f1b605d019
+	.dword	0x923f82a4af194f9b, 0xab1c5ed5da6d8118
+	.dword	0xd807aa98a3030242, 0x12835b0145706fbe
+	.dword	0x243185be4ee4b28c, 0x550c7dc3d5ffb4e2
+	.dword	0x72be5d74f27b896f, 0x80deb1fe3b1696b1
+	.dword	0x9bdc06a725c71235, 0xc19bf174cf692694
+	.dword	0xe49b69c19ef14ad2, 0xefbe4786384f25e3
+	.dword	0x0fc19dc68b8cd5b5, 0x240ca1cc77ac9c65
+	.dword	0x2de92c6f592b0275, 0x4a7484aa6ea6e483
+	.dword	0x5cb0a9dcbd41fbd4, 0x76f988da831153b5
+	.dword	0x983e5152ee66dfab, 0xa831c66d2db43210
+	.dword	0xb00327c898fb213f, 0xbf597fc7beef0ee4
+	.dword	0xc6e00bf33da88fc2, 0xd5a79147930aa725
+	.dword	0x06ca6351e003826f, 0x142929670a0e6e70
+	.dword	0x27b70a8546d22ffc, 0x2e1b21385c26c926
+	.dword	0x4d2c6dfc5ac42aed, 0x53380d139d95b3df
+	.dword	0x650a73548baf63de, 0x766a0abb3c77b2a8
+	.dword	0x81c2c92e47edaee6, 0x92722c851482353b
+	.dword	0xa2bfe8a14cf10364, 0xa81a664bbc423001
+	.dword	0xc24b8b70d0f89791, 0xc76c51a30654be30
+	.dword	0xd192e819d6ef5218, 0xd69906245565a910
+	.dword	0xf40e35855771202a, 0x106aa07032bbd1b8
+	.dword	0x19a4c116b8d2d0c8, 0x1e376c085141ab53
+	.dword	0x2748774cdf8eeb99, 0x34b0bcb5e19b48a8
+	.dword	0x391c0cb3c5c95a63, 0x4ed8aa4ae3418acb
+	.dword	0x5b9cca4f7763e373, 0x682e6ff3d6b2b8a3
+	.dword	0x748f82ee5defb2fc, 0x78a5636f43172f60
+	.dword	0x84c87814a1f0ab72, 0x8cc702081a6439ec
+	.dword	0x90befffa23631e28, 0xa4506cebde82bde9
+	.dword	0xbef9a3f7b2c67915, 0xc67178f2e372532b
+	.dword	0xca273eceea26619c, 0xd186b8c721c0c207
+	.dword	0xeada7dd6cde0eb1e, 0xf57d4f7fee6ed178
+	.dword	0x06f067aa72176fba, 0x0a637dc5a2c898a6
+	.dword	0x113f9804bef90dae, 0x1b710b35131c471b
+	.dword	0x28db77f523047d84, 0x32caab7b40c72493
+	.dword	0x3c9ebe0a15c9bebc, 0x431d67c49c100d4c
+	.dword	0x4cc5d4becb3e42b6, 0x597f299cfc657e2a
+	.dword	0x5fcb6fab3ad6faec, 0x6c44198c4a475817
+___
+}
+$code.=<<___;
+.asciiz	"SHA${label} for MIPS, CRYPTOGAMS by <appro\@openssl.org>"
+.align	5
+
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha512-parisc.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha512-parisc.pl
new file mode 100755
index 0000000000..fc0e15b3c0
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha512-parisc.pl
@@ -0,0 +1,793 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# SHA256/512 block procedure for PA-RISC.
+
+# June 2009.
+#
+# SHA256 performance is >75% better than gcc 3.2 generated code on
+# PA-7100LC. Compared to code generated by vendor compiler this
+# implementation is almost 70% faster in 64-bit build, but delivers
+# virtually same performance in 32-bit build on PA-8600.
+#
+# SHA512 performance is >2.9x better than gcc 3.2 generated code on
+# PA-7100LC, PA-RISC 1.1 processor. Then implementation detects if the
+# code is executed on PA-RISC 2.0 processor and switches to 64-bit
+# code path delivering adequate peformance even in "blended" 32-bit
+# build. Though 64-bit code is not any faster than code generated by
+# vendor compiler on PA-8600...
+#
+# Special thanks to polarhome.com for providing HP-UX account.
+
+$flavour = shift;
+$output = shift;
+open STDOUT,">$output";
+
+if ($flavour =~ /64/) {
+	$LEVEL		="2.0W";
+	$SIZE_T		=8;
+	$FRAME_MARKER	=80;
+	$SAVED_RP	=16;
+	$PUSH		="std";
+	$PUSHMA		="std,ma";
+	$POP		="ldd";
+	$POPMB		="ldd,mb";
+} else {
+	$LEVEL		="1.0";
+	$SIZE_T		=4;
+	$FRAME_MARKER	=48;
+	$SAVED_RP	=20;
+	$PUSH		="stw";
+	$PUSHMA		="stwm";
+	$POP		="ldw";
+	$POPMB		="ldwm";
+}
+
+if ($output =~ /512/) {
+	$func="sha512_block_data_order";
+	$SZ=8;
+	@Sigma0=(28,34,39);
+	@Sigma1=(14,18,41);
+	@sigma0=(1,  8, 7);
+	@sigma1=(19,61, 6);
+	$rounds=80;
+	$LAST10BITS=0x017;
+	$LD="ldd";
+	$LDM="ldd,ma";
+	$ST="std";
+} else {
+	$func="sha256_block_data_order";
+	$SZ=4;
+	@Sigma0=( 2,13,22);
+	@Sigma1=( 6,11,25);
+	@sigma0=( 7,18, 3);
+	@sigma1=(17,19,10);
+	$rounds=64;
+	$LAST10BITS=0x0f2;
+	$LD="ldw";
+	$LDM="ldwm";
+	$ST="stw";
+}
+
+$FRAME=16*$SIZE_T+$FRAME_MARKER;# 16 saved regs + frame marker
+				#                 [+ argument transfer]
+$XOFF=16*$SZ+32;		# local variables
+$FRAME+=$XOFF;
+$XOFF+=$FRAME_MARKER;		# distance between %sp and local variables
+
+$ctx="%r26";	# zapped by $a0
+$inp="%r25";	# zapped by $a1
+$num="%r24";	# zapped by $t0
+
+$a0 ="%r26";
+$a1 ="%r25";
+$t0 ="%r24";
+$t1 ="%r29";
+$Tbl="%r31";
+
+@V=($A,$B,$C,$D,$E,$F,$G,$H)=("%r17","%r18","%r19","%r20","%r21","%r22","%r23","%r28");
+
+@X=("%r1", "%r2", "%r3", "%r4", "%r5", "%r6", "%r7", "%r8",
+    "%r9", "%r10","%r11","%r12","%r13","%r14","%r15","%r16",$inp);
+
+sub ROUND_00_15 {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_;
+$code.=<<___;
+	_ror	$e,$Sigma1[0],$a0
+	and	$f,$e,$t0
+	_ror	$e,$Sigma1[1],$a1
+	addl	$t1,$h,$h
+	andcm	$g,$e,$t1
+	xor	$a1,$a0,$a0
+	_ror	$a1,`$Sigma1[2]-$Sigma1[1]`,$a1
+	or	$t0,$t1,$t1		; Ch(e,f,g)
+	addl	@X[$i%16],$h,$h
+	xor	$a0,$a1,$a1		; Sigma1(e)
+	addl	$t1,$h,$h
+	_ror	$a,$Sigma0[0],$a0
+	addl	$a1,$h,$h
+
+	_ror	$a,$Sigma0[1],$a1
+	and	$a,$b,$t0
+	and	$a,$c,$t1
+	xor	$a1,$a0,$a0
+	_ror	$a1,`$Sigma0[2]-$Sigma0[1]`,$a1
+	xor	$t1,$t0,$t0
+	and	$b,$c,$t1
+	xor	$a0,$a1,$a1		; Sigma0(a)
+	addl	$h,$d,$d
+	xor	$t1,$t0,$t0		; Maj(a,b,c)
+	`"$LDM	$SZ($Tbl),$t1" if ($i<15)`
+	addl	$a1,$h,$h
+	addl	$t0,$h,$h
+
+___
+}
+
+sub ROUND_16_xx {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_;
+$i-=16;
+$code.=<<___;
+	_ror	@X[($i+1)%16],$sigma0[0],$a0
+	_ror	@X[($i+1)%16],$sigma0[1],$a1
+	addl	@X[($i+9)%16],@X[$i],@X[$i]
+	_ror	@X[($i+14)%16],$sigma1[0],$t0
+	_ror	@X[($i+14)%16],$sigma1[1],$t1
+	xor	$a1,$a0,$a0
+	_shr	@X[($i+1)%16],$sigma0[2],$a1
+	xor	$t1,$t0,$t0
+	_shr	@X[($i+14)%16],$sigma1[2],$t1
+	xor	$a1,$a0,$a0		; sigma0(X[(i+1)&0x0f])
+	xor	$t1,$t0,$t0		; sigma1(X[(i+14)&0x0f])
+	$LDM	$SZ($Tbl),$t1
+	addl	$a0,@X[$i],@X[$i]
+	addl	$t0,@X[$i],@X[$i]
+___
+$code.=<<___ if ($i==15);
+	extru	$t1,31,10,$a1
+	comiclr,<> $LAST10BITS,$a1,%r0
+	ldo	1($Tbl),$Tbl		; signal end of $Tbl
+___
+&ROUND_00_15($i+16,$a,$b,$c,$d,$e,$f,$g,$h);
+}
+
+$code=<<___;
+	.LEVEL	$LEVEL
+	.SPACE	\$TEXT\$
+	.SUBSPA	\$CODE\$,QUAD=0,ALIGN=8,ACCESS=0x2C,CODE_ONLY
+
+	.ALIGN	64
+L\$table
+___
+$code.=<<___ if ($SZ==8);
+	.WORD	0x428a2f98,0xd728ae22,0x71374491,0x23ef65cd
+	.WORD	0xb5c0fbcf,0xec4d3b2f,0xe9b5dba5,0x8189dbbc
+	.WORD	0x3956c25b,0xf348b538,0x59f111f1,0xb605d019
+	.WORD	0x923f82a4,0xaf194f9b,0xab1c5ed5,0xda6d8118
+	.WORD	0xd807aa98,0xa3030242,0x12835b01,0x45706fbe
+	.WORD	0x243185be,0x4ee4b28c,0x550c7dc3,0xd5ffb4e2
+	.WORD	0x72be5d74,0xf27b896f,0x80deb1fe,0x3b1696b1
+	.WORD	0x9bdc06a7,0x25c71235,0xc19bf174,0xcf692694
+	.WORD	0xe49b69c1,0x9ef14ad2,0xefbe4786,0x384f25e3
+	.WORD	0x0fc19dc6,0x8b8cd5b5,0x240ca1cc,0x77ac9c65
+	.WORD	0x2de92c6f,0x592b0275,0x4a7484aa,0x6ea6e483
+	.WORD	0x5cb0a9dc,0xbd41fbd4,0x76f988da,0x831153b5
+	.WORD	0x983e5152,0xee66dfab,0xa831c66d,0x2db43210
+	.WORD	0xb00327c8,0x98fb213f,0xbf597fc7,0xbeef0ee4
+	.WORD	0xc6e00bf3,0x3da88fc2,0xd5a79147,0x930aa725
+	.WORD	0x06ca6351,0xe003826f,0x14292967,0x0a0e6e70
+	.WORD	0x27b70a85,0x46d22ffc,0x2e1b2138,0x5c26c926
+	.WORD	0x4d2c6dfc,0x5ac42aed,0x53380d13,0x9d95b3df
+	.WORD	0x650a7354,0x8baf63de,0x766a0abb,0x3c77b2a8
+	.WORD	0x81c2c92e,0x47edaee6,0x92722c85,0x1482353b
+	.WORD	0xa2bfe8a1,0x4cf10364,0xa81a664b,0xbc423001
+	.WORD	0xc24b8b70,0xd0f89791,0xc76c51a3,0x0654be30
+	.WORD	0xd192e819,0xd6ef5218,0xd6990624,0x5565a910
+	.WORD	0xf40e3585,0x5771202a,0x106aa070,0x32bbd1b8
+	.WORD	0x19a4c116,0xb8d2d0c8,0x1e376c08,0x5141ab53
+	.WORD	0x2748774c,0xdf8eeb99,0x34b0bcb5,0xe19b48a8
+	.WORD	0x391c0cb3,0xc5c95a63,0x4ed8aa4a,0xe3418acb
+	.WORD	0x5b9cca4f,0x7763e373,0x682e6ff3,0xd6b2b8a3
+	.WORD	0x748f82ee,0x5defb2fc,0x78a5636f,0x43172f60
+	.WORD	0x84c87814,0xa1f0ab72,0x8cc70208,0x1a6439ec
+	.WORD	0x90befffa,0x23631e28,0xa4506ceb,0xde82bde9
+	.WORD	0xbef9a3f7,0xb2c67915,0xc67178f2,0xe372532b
+	.WORD	0xca273ece,0xea26619c,0xd186b8c7,0x21c0c207
+	.WORD	0xeada7dd6,0xcde0eb1e,0xf57d4f7f,0xee6ed178
+	.WORD	0x06f067aa,0x72176fba,0x0a637dc5,0xa2c898a6
+	.WORD	0x113f9804,0xbef90dae,0x1b710b35,0x131c471b
+	.WORD	0x28db77f5,0x23047d84,0x32caab7b,0x40c72493
+	.WORD	0x3c9ebe0a,0x15c9bebc,0x431d67c4,0x9c100d4c
+	.WORD	0x4cc5d4be,0xcb3e42b6,0x597f299c,0xfc657e2a
+	.WORD	0x5fcb6fab,0x3ad6faec,0x6c44198c,0x4a475817
+___
+$code.=<<___ if ($SZ==4);
+	.WORD	0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5
+	.WORD	0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5
+	.WORD	0xd807aa98,0x12835b01,0x243185be,0x550c7dc3
+	.WORD	0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174
+	.WORD	0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc
+	.WORD	0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da
+	.WORD	0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7
+	.WORD	0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967
+	.WORD	0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13
+	.WORD	0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85
+	.WORD	0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3
+	.WORD	0xd192e819,0xd6990624,0xf40e3585,0x106aa070
+	.WORD	0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5
+	.WORD	0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3
+	.WORD	0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208
+	.WORD	0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
+___
+$code.=<<___;
+
+	.EXPORT	$func,ENTRY,ARGW0=GR,ARGW1=GR,ARGW2=GR
+	.ALIGN	64
+$func
+	.PROC
+	.CALLINFO	FRAME=`$FRAME-16*$SIZE_T`,NO_CALLS,SAVE_RP,ENTRY_GR=18
+	.ENTRY
+	$PUSH	%r2,-$SAVED_RP(%sp)	; standard prologue
+	$PUSHMA	%r3,$FRAME(%sp)
+	$PUSH	%r4,`-$FRAME+1*$SIZE_T`(%sp)
+	$PUSH	%r5,`-$FRAME+2*$SIZE_T`(%sp)
+	$PUSH	%r6,`-$FRAME+3*$SIZE_T`(%sp)
+	$PUSH	%r7,`-$FRAME+4*$SIZE_T`(%sp)
+	$PUSH	%r8,`-$FRAME+5*$SIZE_T`(%sp)
+	$PUSH	%r9,`-$FRAME+6*$SIZE_T`(%sp)
+	$PUSH	%r10,`-$FRAME+7*$SIZE_T`(%sp)
+	$PUSH	%r11,`-$FRAME+8*$SIZE_T`(%sp)
+	$PUSH	%r12,`-$FRAME+9*$SIZE_T`(%sp)
+	$PUSH	%r13,`-$FRAME+10*$SIZE_T`(%sp)
+	$PUSH	%r14,`-$FRAME+11*$SIZE_T`(%sp)
+	$PUSH	%r15,`-$FRAME+12*$SIZE_T`(%sp)
+	$PUSH	%r16,`-$FRAME+13*$SIZE_T`(%sp)
+	$PUSH	%r17,`-$FRAME+14*$SIZE_T`(%sp)
+	$PUSH	%r18,`-$FRAME+15*$SIZE_T`(%sp)
+
+	_shl	$num,`log(16*$SZ)/log(2)`,$num
+	addl	$inp,$num,$num		; $num to point at the end of $inp
+
+	$PUSH	$num,`-$FRAME_MARKER-4*$SIZE_T`(%sp)	; save arguments
+	$PUSH	$inp,`-$FRAME_MARKER-3*$SIZE_T`(%sp)
+	$PUSH	$ctx,`-$FRAME_MARKER-2*$SIZE_T`(%sp)
+
+	blr	%r0,$Tbl
+	ldi	3,$t1
+L\$pic
+	andcm	$Tbl,$t1,$Tbl		; wipe privilege level
+	ldo	L\$table-L\$pic($Tbl),$Tbl
+___
+$code.=<<___ if ($SZ==8 && $SIZE_T==4);
+	ldi	31,$t1
+	mtctl	$t1,%cr11
+	extrd,u,*= $t1,%sar,1,$t1	; executes on PA-RISC 1.0
+	b	L\$parisc1
+	nop
+___
+$code.=<<___;
+	$LD	`0*$SZ`($ctx),$A	; load context
+	$LD	`1*$SZ`($ctx),$B
+	$LD	`2*$SZ`($ctx),$C
+	$LD	`3*$SZ`($ctx),$D
+	$LD	`4*$SZ`($ctx),$E
+	$LD	`5*$SZ`($ctx),$F
+	$LD	`6*$SZ`($ctx),$G
+	$LD	`7*$SZ`($ctx),$H
+
+	extru	$inp,31,`log($SZ)/log(2)`,$t0
+	sh3addl	$t0,%r0,$t0
+	subi	`8*$SZ`,$t0,$t0
+	mtctl	$t0,%cr11		; load %sar with align factor
+
+L\$oop
+	ldi	`$SZ-1`,$t0
+	$LDM	$SZ($Tbl),$t1
+	andcm	$inp,$t0,$t0		; align $inp
+___
+	for ($i=0;$i<15;$i++) {		# load input block
+	$code.="\t$LD	`$SZ*$i`($t0),@X[$i]\n";		}
+$code.=<<___;
+	cmpb,*=	$inp,$t0,L\$aligned
+	$LD	`$SZ*15`($t0),@X[15]
+	$LD	`$SZ*16`($t0),@X[16]
+___
+	for ($i=0;$i<16;$i++) {		# align data
+	$code.="\t_align	@X[$i],@X[$i+1],@X[$i]\n";	}
+$code.=<<___;
+L\$aligned
+	nop	; otherwise /usr/ccs/bin/as is confused by below .WORD
+___
+
+for($i=0;$i<16;$i++)	{ &ROUND_00_15($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+L\$rounds
+	nop	; otherwise /usr/ccs/bin/as is confused by below .WORD
+___
+for(;$i<32;$i++)	{ &ROUND_16_xx($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	bb,>=	$Tbl,31,L\$rounds	; end of $Tbl signalled?
+	nop
+
+	$POP	`-$FRAME_MARKER-2*$SIZE_T`(%sp),$ctx	; restore arguments
+	$POP	`-$FRAME_MARKER-3*$SIZE_T`(%sp),$inp
+	$POP	`-$FRAME_MARKER-4*$SIZE_T`(%sp),$num
+	ldo	`-$rounds*$SZ-1`($Tbl),$Tbl		; rewind $Tbl
+
+	$LD	`0*$SZ`($ctx),@X[0]	; load context
+	$LD	`1*$SZ`($ctx),@X[1]
+	$LD	`2*$SZ`($ctx),@X[2]
+	$LD	`3*$SZ`($ctx),@X[3]
+	$LD	`4*$SZ`($ctx),@X[4]
+	$LD	`5*$SZ`($ctx),@X[5]
+	addl	@X[0],$A,$A
+	$LD	`6*$SZ`($ctx),@X[6]
+	addl	@X[1],$B,$B
+	$LD	`7*$SZ`($ctx),@X[7]
+	ldo	`16*$SZ`($inp),$inp	; advance $inp
+
+	$ST	$A,`0*$SZ`($ctx)	; save context
+	addl	@X[2],$C,$C
+	$ST	$B,`1*$SZ`($ctx)
+	addl	@X[3],$D,$D
+	$ST	$C,`2*$SZ`($ctx)
+	addl	@X[4],$E,$E
+	$ST	$D,`3*$SZ`($ctx)
+	addl	@X[5],$F,$F
+	$ST	$E,`4*$SZ`($ctx)
+	addl	@X[6],$G,$G
+	$ST	$F,`5*$SZ`($ctx)
+	addl	@X[7],$H,$H
+	$ST	$G,`6*$SZ`($ctx)
+	$ST	$H,`7*$SZ`($ctx)
+
+	cmpb,*<>,n $inp,$num,L\$oop
+	$PUSH	$inp,`-$FRAME_MARKER-3*$SIZE_T`(%sp)	; save $inp
+___
+if ($SZ==8 && $SIZE_T==4)	# SHA512 for 32-bit PA-RISC 1.0
+{{
+$code.=<<___;
+	b	L\$done
+	nop
+
+	.ALIGN	64
+L\$parisc1
+___
+
+@V=(  $Ahi,  $Alo,  $Bhi,  $Blo,  $Chi,  $Clo,  $Dhi,  $Dlo,
+      $Ehi,  $Elo,  $Fhi,  $Flo,  $Ghi,  $Glo,  $Hhi,  $Hlo) = 
+   ( "%r1", "%r2", "%r3", "%r4", "%r5", "%r6", "%r7", "%r8",
+     "%r9","%r10","%r11","%r12","%r13","%r14","%r15","%r16");
+$a0 ="%r17";
+$a1 ="%r18";
+$a2 ="%r19";
+$a3 ="%r20";
+$t0 ="%r21";
+$t1 ="%r22";
+$t2 ="%r28";
+$t3 ="%r29";
+$Tbl="%r31";
+
+@X=("%r23","%r24","%r25","%r26");	# zaps $num,$inp,$ctx
+
+sub ROUND_00_15_pa1 {
+my ($i,$ahi,$alo,$bhi,$blo,$chi,$clo,$dhi,$dlo,
+       $ehi,$elo,$fhi,$flo,$ghi,$glo,$hhi,$hlo,$flag)=@_;
+my ($Xhi,$Xlo,$Xnhi,$Xnlo) = @X;
+
+$code.=<<___ if (!$flag);
+	ldw	`-$XOFF+8*(($i+1)%16)`(%sp),$Xnhi
+	ldw	`-$XOFF+8*(($i+1)%16)+4`(%sp),$Xnlo	; load X[i+1]
+___
+$code.=<<___;
+	shd	$ehi,$elo,$Sigma1[0],$t0
+	 add	$Xlo,$hlo,$hlo
+	shd	$elo,$ehi,$Sigma1[0],$t1
+	 addc	$Xhi,$hhi,$hhi		; h += X[i]
+	shd	$ehi,$elo,$Sigma1[1],$t2
+	 ldwm	8($Tbl),$Xhi
+	shd	$elo,$ehi,$Sigma1[1],$t3
+	 ldw	-4($Tbl),$Xlo		; load K[i]
+	xor	$t2,$t0,$t0
+	xor	$t3,$t1,$t1
+	 and	$flo,$elo,$a0
+	 and	$fhi,$ehi,$a1
+	shd	$ehi,$elo,$Sigma1[2],$t2
+	 andcm	$glo,$elo,$a2
+	shd	$elo,$ehi,$Sigma1[2],$t3
+	 andcm	$ghi,$ehi,$a3
+	xor	$t2,$t0,$t0
+	xor	$t3,$t1,$t1		; Sigma1(e)
+	add	$Xlo,$hlo,$hlo
+	 xor	$a2,$a0,$a0
+	addc	$Xhi,$hhi,$hhi		; h += K[i]
+	 xor	$a3,$a1,$a1		; Ch(e,f,g)
+
+	 add	$t0,$hlo,$hlo
+	shd	$ahi,$alo,$Sigma0[0],$t0
+	 addc	$t1,$hhi,$hhi		; h += Sigma1(e)
+	shd	$alo,$ahi,$Sigma0[0],$t1	
+	 add	$a0,$hlo,$hlo
+	shd	$ahi,$alo,$Sigma0[1],$t2
+	 addc	$a1,$hhi,$hhi		; h += Ch(e,f,g)
+	shd	$alo,$ahi,$Sigma0[1],$t3
+
+	xor	$t2,$t0,$t0
+	xor	$t3,$t1,$t1
+	shd	$ahi,$alo,$Sigma0[2],$t2
+	and	$alo,$blo,$a0
+	shd	$alo,$ahi,$Sigma0[2],$t3
+	and	$ahi,$bhi,$a1
+	xor	$t2,$t0,$t0
+	xor	$t3,$t1,$t1		; Sigma0(a)
+
+	and	$alo,$clo,$a2
+	and	$ahi,$chi,$a3
+	xor	$a2,$a0,$a0
+	 add	$hlo,$dlo,$dlo
+	xor	$a3,$a1,$a1
+	 addc	$hhi,$dhi,$dhi		; d += h
+	and	$blo,$clo,$a2
+	 add	$t0,$hlo,$hlo
+	and	$bhi,$chi,$a3
+	 addc	$t1,$hhi,$hhi		; h += Sigma0(a)
+	xor	$a2,$a0,$a0
+	 add	$a0,$hlo,$hlo
+	xor	$a3,$a1,$a1		; Maj(a,b,c)
+	 addc	$a1,$hhi,$hhi		; h += Maj(a,b,c)
+
+___
+$code.=<<___ if ($i==15 && $flag);
+	extru	$Xlo,31,10,$Xlo
+	comiclr,= $LAST10BITS,$Xlo,%r0
+	b	L\$rounds_pa1
+	nop
+___
+push(@X,shift(@X)); push(@X,shift(@X));
+}
+
+sub ROUND_16_xx_pa1 {
+my ($Xhi,$Xlo,$Xnhi,$Xnlo) = @X;
+my ($i)=shift;
+$i-=16;
+$code.=<<___;
+	ldw	`-$XOFF+8*(($i+1)%16)`(%sp),$Xnhi
+	ldw	`-$XOFF+8*(($i+1)%16)+4`(%sp),$Xnlo	; load X[i+1]
+	ldw	`-$XOFF+8*(($i+9)%16)`(%sp),$a1
+	ldw	`-$XOFF+8*(($i+9)%16)+4`(%sp),$a0	; load X[i+9]
+	ldw	`-$XOFF+8*(($i+14)%16)`(%sp),$a3
+	ldw	`-$XOFF+8*(($i+14)%16)+4`(%sp),$a2	; load X[i+14]
+	shd	$Xnhi,$Xnlo,$sigma0[0],$t0
+	shd	$Xnlo,$Xnhi,$sigma0[0],$t1
+	 add	$a0,$Xlo,$Xlo
+	shd	$Xnhi,$Xnlo,$sigma0[1],$t2
+	 addc	$a1,$Xhi,$Xhi
+	shd	$Xnlo,$Xnhi,$sigma0[1],$t3
+	xor	$t2,$t0,$t0
+	shd	$Xnhi,$Xnlo,$sigma0[2],$t2
+	xor	$t3,$t1,$t1
+	extru	$Xnhi,`31-$sigma0[2]`,`32-$sigma0[2]`,$t3
+	xor	$t2,$t0,$t0
+	 shd	$a3,$a2,$sigma1[0],$a0
+	xor	$t3,$t1,$t1		; sigma0(X[i+1)&0x0f])
+	 shd	$a2,$a3,$sigma1[0],$a1
+	add	$t0,$Xlo,$Xlo
+	 shd	$a3,$a2,$sigma1[1],$t2
+	addc	$t1,$Xhi,$Xhi
+	 shd	$a2,$a3,$sigma1[1],$t3
+	xor	$t2,$a0,$a0
+	shd	$a3,$a2,$sigma1[2],$t2
+	xor	$t3,$a1,$a1
+	extru	$a3,`31-$sigma1[2]`,`32-$sigma1[2]`,$t3
+	xor	$t2,$a0,$a0
+	xor	$t3,$a1,$a1		; sigma0(X[i+14)&0x0f])
+	add	$a0,$Xlo,$Xlo
+	addc	$a1,$Xhi,$Xhi
+
+	stw	$Xhi,`-$XOFF+8*($i%16)`(%sp)
+	stw	$Xlo,`-$XOFF+8*($i%16)+4`(%sp)
+___
+&ROUND_00_15_pa1($i,@_,1);
+}
+$code.=<<___;
+	ldw	`0*4`($ctx),$Ahi		; load context
+	ldw	`1*4`($ctx),$Alo
+	ldw	`2*4`($ctx),$Bhi
+	ldw	`3*4`($ctx),$Blo
+	ldw	`4*4`($ctx),$Chi
+	ldw	`5*4`($ctx),$Clo
+	ldw	`6*4`($ctx),$Dhi
+	ldw	`7*4`($ctx),$Dlo
+	ldw	`8*4`($ctx),$Ehi
+	ldw	`9*4`($ctx),$Elo
+	ldw	`10*4`($ctx),$Fhi
+	ldw	`11*4`($ctx),$Flo
+	ldw	`12*4`($ctx),$Ghi
+	ldw	`13*4`($ctx),$Glo
+	ldw	`14*4`($ctx),$Hhi
+	ldw	`15*4`($ctx),$Hlo
+
+	extru	$inp,31,2,$t0
+	sh3addl	$t0,%r0,$t0
+	subi	32,$t0,$t0
+	mtctl	$t0,%cr11		; load %sar with align factor
+
+L\$oop_pa1
+	extru	$inp,31,2,$a3
+	comib,=	0,$a3,L\$aligned_pa1
+	sub	$inp,$a3,$inp
+
+	ldw	`0*4`($inp),$X[0]
+	ldw	`1*4`($inp),$X[1]
+	ldw	`2*4`($inp),$t2
+	ldw	`3*4`($inp),$t3
+	ldw	`4*4`($inp),$a0
+	ldw	`5*4`($inp),$a1
+	ldw	`6*4`($inp),$a2
+	ldw	`7*4`($inp),$a3
+	vshd	$X[0],$X[1],$X[0]
+	vshd	$X[1],$t2,$X[1]
+	stw	$X[0],`-$XOFF+0*4`(%sp)
+	ldw	`8*4`($inp),$t0
+	vshd	$t2,$t3,$t2
+	stw	$X[1],`-$XOFF+1*4`(%sp)
+	ldw	`9*4`($inp),$t1
+	vshd	$t3,$a0,$t3
+___
+{
+my @t=($t2,$t3,$a0,$a1,$a2,$a3,$t0,$t1);
+for ($i=2;$i<=(128/4-8);$i++) {
+$code.=<<___;
+	stw	$t[0],`-$XOFF+$i*4`(%sp)
+	ldw	`(8+$i)*4`($inp),$t[0]
+	vshd	$t[1],$t[2],$t[1]
+___
+push(@t,shift(@t));
+}
+for (;$i<(128/4-1);$i++) {
+$code.=<<___;
+	stw	$t[0],`-$XOFF+$i*4`(%sp)
+	vshd	$t[1],$t[2],$t[1]
+___
+push(@t,shift(@t));
+}
+$code.=<<___;
+	b	L\$collected_pa1
+	stw	$t[0],`-$XOFF+$i*4`(%sp)
+
+___
+}
+$code.=<<___;
+L\$aligned_pa1
+	ldw	`0*4`($inp),$X[0]
+	ldw	`1*4`($inp),$X[1]
+	ldw	`2*4`($inp),$t2
+	ldw	`3*4`($inp),$t3
+	ldw	`4*4`($inp),$a0
+	ldw	`5*4`($inp),$a1
+	ldw	`6*4`($inp),$a2
+	ldw	`7*4`($inp),$a3
+	stw	$X[0],`-$XOFF+0*4`(%sp)
+	ldw	`8*4`($inp),$t0
+	stw	$X[1],`-$XOFF+1*4`(%sp)
+	ldw	`9*4`($inp),$t1
+___
+{
+my @t=($t2,$t3,$a0,$a1,$a2,$a3,$t0,$t1);
+for ($i=2;$i<(128/4-8);$i++) {
+$code.=<<___;
+	stw	$t[0],`-$XOFF+$i*4`(%sp)
+	ldw	`(8+$i)*4`($inp),$t[0]
+___
+push(@t,shift(@t));
+}
+for (;$i<128/4;$i++) {
+$code.=<<___;
+	stw	$t[0],`-$XOFF+$i*4`(%sp)
+___
+push(@t,shift(@t));
+}
+$code.="L\$collected_pa1\n";
+}
+
+for($i=0;$i<16;$i++)	{ &ROUND_00_15_pa1($i,@V); unshift(@V,pop(@V)); unshift(@V,pop(@V)); }
+$code.="L\$rounds_pa1\n";
+for(;$i<32;$i++)	{ &ROUND_16_xx_pa1($i,@V); unshift(@V,pop(@V)); unshift(@V,pop(@V)); }
+
+$code.=<<___;
+	$POP	`-$FRAME_MARKER-2*$SIZE_T`(%sp),$ctx	; restore arguments
+	$POP	`-$FRAME_MARKER-3*$SIZE_T`(%sp),$inp
+	$POP	`-$FRAME_MARKER-4*$SIZE_T`(%sp),$num
+	ldo	`-$rounds*$SZ`($Tbl),$Tbl		; rewind $Tbl
+
+	ldw	`0*4`($ctx),$t1		; update context
+	ldw	`1*4`($ctx),$t0
+	ldw	`2*4`($ctx),$t3
+	ldw	`3*4`($ctx),$t2
+	ldw	`4*4`($ctx),$a1
+	ldw	`5*4`($ctx),$a0
+	ldw	`6*4`($ctx),$a3
+	add	$t0,$Alo,$Alo
+	ldw	`7*4`($ctx),$a2
+	addc	$t1,$Ahi,$Ahi
+	ldw	`8*4`($ctx),$t1
+	add	$t2,$Blo,$Blo
+	ldw	`9*4`($ctx),$t0
+	addc	$t3,$Bhi,$Bhi
+	ldw	`10*4`($ctx),$t3
+	add	$a0,$Clo,$Clo
+	ldw	`11*4`($ctx),$t2
+	addc	$a1,$Chi,$Chi
+	ldw	`12*4`($ctx),$a1
+	add	$a2,$Dlo,$Dlo
+	ldw	`13*4`($ctx),$a0
+	addc	$a3,$Dhi,$Dhi
+	ldw	`14*4`($ctx),$a3
+	add	$t0,$Elo,$Elo
+	ldw	`15*4`($ctx),$a2
+	addc	$t1,$Ehi,$Ehi
+	stw	$Ahi,`0*4`($ctx)
+	add	$t2,$Flo,$Flo
+	stw	$Alo,`1*4`($ctx)
+	addc	$t3,$Fhi,$Fhi
+	stw	$Bhi,`2*4`($ctx)
+	add	$a0,$Glo,$Glo
+	stw	$Blo,`3*4`($ctx)
+	addc	$a1,$Ghi,$Ghi
+	stw	$Chi,`4*4`($ctx)
+	add	$a2,$Hlo,$Hlo
+	stw	$Clo,`5*4`($ctx)
+	addc	$a3,$Hhi,$Hhi
+	stw	$Dhi,`6*4`($ctx)
+	ldo	`16*$SZ`($inp),$inp	; advance $inp
+	stw	$Dlo,`7*4`($ctx)
+	stw	$Ehi,`8*4`($ctx)
+	stw	$Elo,`9*4`($ctx)
+	stw	$Fhi,`10*4`($ctx)
+	stw	$Flo,`11*4`($ctx)
+	stw	$Ghi,`12*4`($ctx)
+	stw	$Glo,`13*4`($ctx)
+	stw	$Hhi,`14*4`($ctx)
+	comb,=	$inp,$num,L\$done
+	stw	$Hlo,`15*4`($ctx)
+	b	L\$oop_pa1
+	$PUSH	$inp,`-$FRAME_MARKER-3*$SIZE_T`(%sp)	; save $inp
+L\$done
+___
+}}
+$code.=<<___;
+	$POP	`-$FRAME-$SAVED_RP`(%sp),%r2		; standard epilogue
+	$POP	`-$FRAME+1*$SIZE_T`(%sp),%r4
+	$POP	`-$FRAME+2*$SIZE_T`(%sp),%r5
+	$POP	`-$FRAME+3*$SIZE_T`(%sp),%r6
+	$POP	`-$FRAME+4*$SIZE_T`(%sp),%r7
+	$POP	`-$FRAME+5*$SIZE_T`(%sp),%r8
+	$POP	`-$FRAME+6*$SIZE_T`(%sp),%r9
+	$POP	`-$FRAME+7*$SIZE_T`(%sp),%r10
+	$POP	`-$FRAME+8*$SIZE_T`(%sp),%r11
+	$POP	`-$FRAME+9*$SIZE_T`(%sp),%r12
+	$POP	`-$FRAME+10*$SIZE_T`(%sp),%r13
+	$POP	`-$FRAME+11*$SIZE_T`(%sp),%r14
+	$POP	`-$FRAME+12*$SIZE_T`(%sp),%r15
+	$POP	`-$FRAME+13*$SIZE_T`(%sp),%r16
+	$POP	`-$FRAME+14*$SIZE_T`(%sp),%r17
+	$POP	`-$FRAME+15*$SIZE_T`(%sp),%r18
+	bv	(%r2)
+	.EXIT
+	$POPMB	-$FRAME(%sp),%r3
+	.PROCEND
+	.STRINGZ "SHA`64*$SZ` block transform for PA-RISC, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+# Explicitly encode PA-RISC 2.0 instructions used in this module, so
+# that it can be compiled with .LEVEL 1.0. It should be noted that I
+# wouldn't have to do this, if GNU assembler understood .ALLOW 2.0
+# directive...
+
+my $ldd = sub {
+  my ($mod,$args) = @_;
+  my $orig = "ldd$mod\t$args";
+
+    if ($args =~ /(\-?[0-9]+)\(%r([0-9]+)\),%r([0-9]+)/) # format 3 suffices
+    {	my $opcode=(0x14<<26)|($2<<21)|($3<<16)|(($1&0x1FF8)<<1)|(($1>>13)&1);
+	$opcode|=(1<<3) if ($mod =~ /^,m/);
+	$opcode|=(1<<2) if ($mod =~ /^,mb/);
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    else { "\t".$orig; }
+};
+
+my $std = sub {
+  my ($mod,$args) = @_;
+  my $orig = "std$mod\t$args";
+
+    if ($args =~ /%r([0-9]+),(\-?[0-9]+)\(%r([0-9]+)\)/) # format 3 suffices
+    {	my $opcode=(0x1c<<26)|($3<<21)|($1<<16)|(($2&0x1FF8)<<1)|(($2>>13)&1);
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    else { "\t".$orig; }
+};
+
+my $extrd = sub {
+  my ($mod,$args) = @_;
+  my $orig = "extrd$mod\t$args";
+
+    # I only have ",u" completer, it's implicitly encoded...
+    if ($args =~ /%r([0-9]+),([0-9]+),([0-9]+),%r([0-9]+)/)	# format 15
+    {	my $opcode=(0x36<<26)|($1<<21)|($4<<16);
+	my $len=32-$3;
+	$opcode |= (($2&0x20)<<6)|(($2&0x1f)<<5);		# encode pos
+	$opcode |= (($len&0x20)<<7)|($len&0x1f);		# encode len
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    elsif ($args =~ /%r([0-9]+),%sar,([0-9]+),%r([0-9]+)/)	# format 12
+    {	my $opcode=(0x34<<26)|($1<<21)|($3<<16)|(2<<11)|(1<<9);
+	my $len=32-$2;
+	$opcode |= (($len&0x20)<<3)|($len&0x1f);		# encode len
+	$opcode |= (1<<13) if ($mod =~ /,\**=/);
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    else { "\t".$orig; }
+};
+
+my $shrpd = sub {
+  my ($mod,$args) = @_;
+  my $orig = "shrpd$mod\t$args";
+
+    if ($args =~ /%r([0-9]+),%r([0-9]+),([0-9]+),%r([0-9]+)/)	# format 14
+    {	my $opcode=(0x34<<26)|($2<<21)|($1<<16)|(1<<10)|$4;
+	my $cpos=63-$3;
+	$opcode |= (($cpos&0x20)<<6)|(($cpos&0x1f)<<5);		# encode sa
+	sprintf "\t.WORD\t0x%08x\t; %s",$opcode,$orig;
+    }
+    elsif ($args =~ /%r([0-9]+),%r([0-9]+),%sar,%r([0-9]+)/)	# format 11
+    {	sprintf "\t.WORD\t0x%08x\t; %s",
+		(0x34<<26)|($2<<21)|($1<<16)|(1<<9)|$3,$orig;
+    }
+    else { "\t".$orig; }
+};
+
+sub assemble {
+  my ($mnemonic,$mod,$args)=@_;
+  my $opcode = eval("\$$mnemonic");
+
+    ref($opcode) eq 'CODE' ? &$opcode($mod,$args) : "\t$mnemonic$mod\t$args";
+}
+
+foreach (split("\n",$code)) {
+	s/\`([^\`]*)\`/eval $1/ge;
+
+	s/shd\s+(%r[0-9]+),(%r[0-9]+),([0-9]+)/
+		$3>31 ? sprintf("shd\t%$2,%$1,%d",$3-32)	# rotation for >=32
+		:       sprintf("shd\t%$1,%$2,%d",$3)/e			or
+	# translate made up instructons: _ror, _shr, _align, _shl
+	s/_ror(\s+)(%r[0-9]+),/
+		($SZ==4 ? "shd" : "shrpd")."$1$2,$2,"/e			or
+
+	s/_shr(\s+%r[0-9]+),([0-9]+),/
+		$SZ==4 ? sprintf("extru%s,%d,%d,",$1,31-$2,32-$2)
+		:        sprintf("extrd,u%s,%d,%d,",$1,63-$2,64-$2)/e	or
+
+	s/_align(\s+%r[0-9]+,%r[0-9]+),/
+		($SZ==4 ? "vshd$1," : "shrpd$1,%sar,")/e		or
+
+	s/_shl(\s+%r[0-9]+),([0-9]+),/
+		$SIZE_T==4 ? sprintf("zdep%s,%d,%d,",$1,31-$2,32-$2)
+		:            sprintf("depd,z%s,%d,%d,",$1,63-$2,64-$2)/e;
+
+	s/^\s+([a-z]+)([\S]*)\s+([\S]*)/&assemble($1,$2,$3)/e if ($SIZE_T==4);
+
+	s/cmpb,\*/comb,/ if ($SIZE_T==4);
+
+	s/\bbv\b/bve/    if ($SIZE_T==8);
+
+	print $_,"\n";
+}
+
+close STDOUT;
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha512-ppc.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha512-ppc.pl
new file mode 100755
index 0000000000..6b44a68e59
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha512-ppc.pl
@@ -0,0 +1,460 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> 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/.
+# ====================================================================
+
+# I let hardware handle unaligned input, except on page boundaries
+# (see below for details). Otherwise straightforward implementation
+# with X vector in register bank. The module is big-endian [which is
+# not big deal as there're no little-endian targets left around].
+
+#			sha256		|	sha512
+# 			-m64	-m32	|	-m64	-m32
+# --------------------------------------+-----------------------
+# PPC970,gcc-4.0.0	+50%	+38%	|	+40%	+410%(*)
+# Power6,xlc-7		+150%	+90%	|	+100%	+430%(*)
+#
+# (*)	64-bit code in 32-bit application context, which actually is
+#	on TODO list. It should be noted that for safe deployment in
+#	32-bit *mutli-threaded* context asyncronous signals should be
+#	blocked upon entry to SHA512 block routine. This is because
+#	32-bit signaling procedure invalidates upper halves of GPRs.
+#	Context switch procedure preserves them, but not signaling:-(
+
+# Second version is true multi-thread safe. Trouble with the original
+# version was that it was using thread local storage pointer register.
+# Well, it scrupulously preserved it, but the problem would arise the
+# moment asynchronous signal was delivered and signal handler would
+# dereference the TLS pointer. While it's never the case in openssl
+# application or test suite, we have to respect this scenario and not
+# use TLS pointer register. Alternative would be to require caller to
+# block signals prior calling this routine. For the record, in 32-bit
+# context R2 serves as TLS pointer, while in 64-bit context - R13.
+
+$flavour=shift;
+$output =shift;
+
+if ($flavour =~ /64/) {
+	$SIZE_T=8;
+	$LRSAVE=2*$SIZE_T;
+	$STU="stdu";
+	$UCMP="cmpld";
+	$SHL="sldi";
+	$POP="ld";
+	$PUSH="std";
+} elsif ($flavour =~ /32/) {
+	$SIZE_T=4;
+	$LRSAVE=$SIZE_T;
+	$STU="stwu";
+	$UCMP="cmplw";
+	$SHL="slwi";
+	$POP="lwz";
+	$PUSH="stw";
+} else { die "nonsense $flavour"; }
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}ppc-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/ppc-xlate.pl" and -f $xlate) or
+die "can't locate ppc-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour $output" || die "can't call $xlate: $!";
+
+if ($output =~ /512/) {
+	$func="sha512_block_data_order";
+	$SZ=8;
+	@Sigma0=(28,34,39);
+	@Sigma1=(14,18,41);
+	@sigma0=(1,  8, 7);
+	@sigma1=(19,61, 6);
+	$rounds=80;
+	$LD="ld";
+	$ST="std";
+	$ROR="rotrdi";
+	$SHR="srdi";
+} else {
+	$func="sha256_block_data_order";
+	$SZ=4;
+	@Sigma0=( 2,13,22);
+	@Sigma1=( 6,11,25);
+	@sigma0=( 7,18, 3);
+	@sigma1=(17,19,10);
+	$rounds=64;
+	$LD="lwz";
+	$ST="stw";
+	$ROR="rotrwi";
+	$SHR="srwi";
+}
+
+$FRAME=32*$SIZE_T+16*$SZ;
+$LOCALS=6*$SIZE_T;
+
+$sp ="r1";
+$toc="r2";
+$ctx="r3";	# zapped by $a0
+$inp="r4";	# zapped by $a1
+$num="r5";	# zapped by $t0
+
+$T  ="r0";
+$a0 ="r3";
+$a1 ="r4";
+$t0 ="r5";
+$t1 ="r6";
+$Tbl="r7";
+
+$A  ="r8";
+$B  ="r9";
+$C  ="r10";
+$D  ="r11";
+$E  ="r12";
+$F  ="r13";	$F="r2" if ($SIZE_T==8);# reassigned to exempt TLS pointer
+$G  ="r14";
+$H  ="r15";
+
+@V=($A,$B,$C,$D,$E,$F,$G,$H);
+@X=("r16","r17","r18","r19","r20","r21","r22","r23",
+    "r24","r25","r26","r27","r28","r29","r30","r31");
+
+$inp="r31";	# reassigned $inp! aliases with @X[15]
+
+sub ROUND_00_15 {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_;
+$code.=<<___;
+	$LD	$T,`$i*$SZ`($Tbl)
+	$ROR	$a0,$e,$Sigma1[0]
+	$ROR	$a1,$e,$Sigma1[1]
+	and	$t0,$f,$e
+	andc	$t1,$g,$e
+	add	$T,$T,$h
+	xor	$a0,$a0,$a1
+	$ROR	$a1,$a1,`$Sigma1[2]-$Sigma1[1]`
+	or	$t0,$t0,$t1		; Ch(e,f,g)
+	add	$T,$T,@X[$i]
+	xor	$a0,$a0,$a1		; Sigma1(e)
+	add	$T,$T,$t0
+	add	$T,$T,$a0
+
+	$ROR	$a0,$a,$Sigma0[0]
+	$ROR	$a1,$a,$Sigma0[1]
+	and	$t0,$a,$b
+	and	$t1,$a,$c
+	xor	$a0,$a0,$a1
+	$ROR	$a1,$a1,`$Sigma0[2]-$Sigma0[1]`
+	xor	$t0,$t0,$t1
+	and	$t1,$b,$c
+	xor	$a0,$a0,$a1		; Sigma0(a)
+	add	$d,$d,$T
+	xor	$t0,$t0,$t1		; Maj(a,b,c)
+	add	$h,$T,$a0
+	add	$h,$h,$t0
+
+___
+}
+
+sub ROUND_16_xx {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_;
+$i-=16;
+$code.=<<___;
+	$ROR	$a0,@X[($i+1)%16],$sigma0[0]
+	$ROR	$a1,@X[($i+1)%16],$sigma0[1]
+	$ROR	$t0,@X[($i+14)%16],$sigma1[0]
+	$ROR	$t1,@X[($i+14)%16],$sigma1[1]
+	xor	$a0,$a0,$a1
+	$SHR	$a1,@X[($i+1)%16],$sigma0[2]
+	xor	$t0,$t0,$t1
+	$SHR	$t1,@X[($i+14)%16],$sigma1[2]
+	add	@X[$i],@X[$i],@X[($i+9)%16]
+	xor	$a0,$a0,$a1		; sigma0(X[(i+1)&0x0f])
+	xor	$t0,$t0,$t1		; sigma1(X[(i+14)&0x0f])
+	add	@X[$i],@X[$i],$a0
+	add	@X[$i],@X[$i],$t0
+___
+&ROUND_00_15($i,$a,$b,$c,$d,$e,$f,$g,$h);
+}
+
+$code=<<___;
+.machine	"any"
+.text
+
+.globl	$func
+.align	6
+$func:
+	$STU	$sp,-$FRAME($sp)
+	mflr	r0
+	$SHL	$num,$num,`log(16*$SZ)/log(2)`
+
+	$PUSH	$ctx,`$FRAME-$SIZE_T*22`($sp)
+
+	$PUSH	$toc,`$FRAME-$SIZE_T*20`($sp)
+	$PUSH	r13,`$FRAME-$SIZE_T*19`($sp)
+	$PUSH	r14,`$FRAME-$SIZE_T*18`($sp)
+	$PUSH	r15,`$FRAME-$SIZE_T*17`($sp)
+	$PUSH	r16,`$FRAME-$SIZE_T*16`($sp)
+	$PUSH	r17,`$FRAME-$SIZE_T*15`($sp)
+	$PUSH	r18,`$FRAME-$SIZE_T*14`($sp)
+	$PUSH	r19,`$FRAME-$SIZE_T*13`($sp)
+	$PUSH	r20,`$FRAME-$SIZE_T*12`($sp)
+	$PUSH	r21,`$FRAME-$SIZE_T*11`($sp)
+	$PUSH	r22,`$FRAME-$SIZE_T*10`($sp)
+	$PUSH	r23,`$FRAME-$SIZE_T*9`($sp)
+	$PUSH	r24,`$FRAME-$SIZE_T*8`($sp)
+	$PUSH	r25,`$FRAME-$SIZE_T*7`($sp)
+	$PUSH	r26,`$FRAME-$SIZE_T*6`($sp)
+	$PUSH	r27,`$FRAME-$SIZE_T*5`($sp)
+	$PUSH	r28,`$FRAME-$SIZE_T*4`($sp)
+	$PUSH	r29,`$FRAME-$SIZE_T*3`($sp)
+	$PUSH	r30,`$FRAME-$SIZE_T*2`($sp)
+	$PUSH	r31,`$FRAME-$SIZE_T*1`($sp)
+	$PUSH	r0,`$FRAME+$LRSAVE`($sp)
+
+	$LD	$A,`0*$SZ`($ctx)
+	mr	$inp,r4				; incarnate $inp
+	$LD	$B,`1*$SZ`($ctx)
+	$LD	$C,`2*$SZ`($ctx)
+	$LD	$D,`3*$SZ`($ctx)
+	$LD	$E,`4*$SZ`($ctx)
+	$LD	$F,`5*$SZ`($ctx)
+	$LD	$G,`6*$SZ`($ctx)
+	$LD	$H,`7*$SZ`($ctx)
+
+	bl	LPICmeup
+LPICedup:
+	andi.	r0,$inp,3
+	bne	Lunaligned
+Laligned:
+	add	$num,$inp,$num
+	$PUSH	$num,`$FRAME-$SIZE_T*24`($sp)	; end pointer
+	$PUSH	$inp,`$FRAME-$SIZE_T*23`($sp)	; inp pointer
+	bl	Lsha2_block_private
+	b	Ldone
+
+; PowerPC specification allows an implementation to be ill-behaved
+; upon unaligned access which crosses page boundary. "Better safe
+; than sorry" principle makes me treat it specially. But I don't
+; look for particular offending word, but rather for the input
+; block which crosses the boundary. Once found that block is aligned
+; and hashed separately...
+.align	4
+Lunaligned:
+	subfic	$t1,$inp,4096
+	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
+	subfc	$num,$t1,$num
+	add	$t1,$inp,$t1
+	$PUSH	$num,`$FRAME-$SIZE_T*25`($sp)	; save real remaining num
+	$PUSH	$t1,`$FRAME-$SIZE_T*24`($sp)	; intermediate end pointer
+	$PUSH	$inp,`$FRAME-$SIZE_T*23`($sp)	; inp pointer
+	bl	Lsha2_block_private
+	; $inp equals to the intermediate end pointer here
+	$POP	$num,`$FRAME-$SIZE_T*25`($sp)	; restore real remaining num
+Lcross_page:
+	li	$t1,`16*$SZ/4`
+	mtctr	$t1
+	addi	r20,$sp,$LOCALS			; aligned spot below the frame
+Lmemcpy:
+	lbz	r16,0($inp)
+	lbz	r17,1($inp)
+	lbz	r18,2($inp)
+	lbz	r19,3($inp)
+	addi	$inp,$inp,4
+	stb	r16,0(r20)
+	stb	r17,1(r20)
+	stb	r18,2(r20)
+	stb	r19,3(r20)
+	addi	r20,r20,4
+	bdnz	Lmemcpy
+
+	$PUSH	$inp,`$FRAME-$SIZE_T*26`($sp)	; save real inp
+	addi	$t1,$sp,`$LOCALS+16*$SZ`	; fictitious end pointer
+	addi	$inp,$sp,$LOCALS		; fictitious inp pointer
+	$PUSH	$num,`$FRAME-$SIZE_T*25`($sp)	; save real num
+	$PUSH	$t1,`$FRAME-$SIZE_T*24`($sp)	; end pointer
+	$PUSH	$inp,`$FRAME-$SIZE_T*23`($sp)	; inp pointer
+	bl	Lsha2_block_private
+	$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
+
+Ldone:
+	$POP	r0,`$FRAME+$LRSAVE`($sp)
+	$POP	$toc,`$FRAME-$SIZE_T*20`($sp)
+	$POP	r13,`$FRAME-$SIZE_T*19`($sp)
+	$POP	r14,`$FRAME-$SIZE_T*18`($sp)
+	$POP	r15,`$FRAME-$SIZE_T*17`($sp)
+	$POP	r16,`$FRAME-$SIZE_T*16`($sp)
+	$POP	r17,`$FRAME-$SIZE_T*15`($sp)
+	$POP	r18,`$FRAME-$SIZE_T*14`($sp)
+	$POP	r19,`$FRAME-$SIZE_T*13`($sp)
+	$POP	r20,`$FRAME-$SIZE_T*12`($sp)
+	$POP	r21,`$FRAME-$SIZE_T*11`($sp)
+	$POP	r22,`$FRAME-$SIZE_T*10`($sp)
+	$POP	r23,`$FRAME-$SIZE_T*9`($sp)
+	$POP	r24,`$FRAME-$SIZE_T*8`($sp)
+	$POP	r25,`$FRAME-$SIZE_T*7`($sp)
+	$POP	r26,`$FRAME-$SIZE_T*6`($sp)
+	$POP	r27,`$FRAME-$SIZE_T*5`($sp)
+	$POP	r28,`$FRAME-$SIZE_T*4`($sp)
+	$POP	r29,`$FRAME-$SIZE_T*3`($sp)
+	$POP	r30,`$FRAME-$SIZE_T*2`($sp)
+	$POP	r31,`$FRAME-$SIZE_T*1`($sp)
+	mtlr	r0
+	addi	$sp,$sp,$FRAME
+	blr
+	.long	0
+	.byte	0,12,4,1,0x80,18,3,0
+	.long	0
+
+.align	4
+Lsha2_block_private:
+___
+for($i=0;$i<16;$i++) {
+$code.=<<___ if ($SZ==4);
+	lwz	@X[$i],`$i*$SZ`($inp)
+___
+# 64-bit loads are split to 2x32-bit ones, as CPU can't handle
+# unaligned 64-bit loads, only 32-bit ones...
+$code.=<<___ if ($SZ==8);
+	lwz	$t0,`$i*$SZ`($inp)
+	lwz	@X[$i],`$i*$SZ+4`($inp)
+	insrdi	@X[$i],$t0,32,0
+___
+	&ROUND_00_15($i,@V);
+	unshift(@V,pop(@V));
+}
+$code.=<<___;
+	li	$T,`$rounds/16-1`
+	mtctr	$T
+.align	4
+Lrounds:
+	addi	$Tbl,$Tbl,`16*$SZ`
+___
+for(;$i<32;$i++) {
+	&ROUND_16_xx($i,@V);
+	unshift(@V,pop(@V));
+}
+$code.=<<___;
+	bdnz-	Lrounds
+
+	$POP	$ctx,`$FRAME-$SIZE_T*22`($sp)
+	$POP	$inp,`$FRAME-$SIZE_T*23`($sp)	; inp pointer
+	$POP	$num,`$FRAME-$SIZE_T*24`($sp)	; end pointer
+	subi	$Tbl,$Tbl,`($rounds-16)*$SZ`	; rewind Tbl
+
+	$LD	r16,`0*$SZ`($ctx)
+	$LD	r17,`1*$SZ`($ctx)
+	$LD	r18,`2*$SZ`($ctx)
+	$LD	r19,`3*$SZ`($ctx)
+	$LD	r20,`4*$SZ`($ctx)
+	$LD	r21,`5*$SZ`($ctx)
+	$LD	r22,`6*$SZ`($ctx)
+	addi	$inp,$inp,`16*$SZ`		; advance inp
+	$LD	r23,`7*$SZ`($ctx)
+	add	$A,$A,r16
+	add	$B,$B,r17
+	$PUSH	$inp,`$FRAME-$SIZE_T*23`($sp)
+	add	$C,$C,r18
+	$ST	$A,`0*$SZ`($ctx)
+	add	$D,$D,r19
+	$ST	$B,`1*$SZ`($ctx)
+	add	$E,$E,r20
+	$ST	$C,`2*$SZ`($ctx)
+	add	$F,$F,r21
+	$ST	$D,`3*$SZ`($ctx)
+	add	$G,$G,r22
+	$ST	$E,`4*$SZ`($ctx)
+	add	$H,$H,r23
+	$ST	$F,`5*$SZ`($ctx)
+	$ST	$G,`6*$SZ`($ctx)
+	$UCMP	$inp,$num
+	$ST	$H,`7*$SZ`($ctx)
+	bne	Lsha2_block_private
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+___
+
+# Ugly hack here, because PPC assembler syntax seem to vary too
+# much from platforms to platform...
+$code.=<<___;
+.align	6
+LPICmeup:
+	mflr	r0
+	bcl	20,31,\$+4
+	mflr	$Tbl	; vvvvvv "distance" between . and 1st data entry
+	addi	$Tbl,$Tbl,`64-8`
+	mtlr	r0
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+	.space	`64-9*4`
+___
+$code.=<<___ if ($SZ==8);
+	.long	0x428a2f98,0xd728ae22,0x71374491,0x23ef65cd
+	.long	0xb5c0fbcf,0xec4d3b2f,0xe9b5dba5,0x8189dbbc
+	.long	0x3956c25b,0xf348b538,0x59f111f1,0xb605d019
+	.long	0x923f82a4,0xaf194f9b,0xab1c5ed5,0xda6d8118
+	.long	0xd807aa98,0xa3030242,0x12835b01,0x45706fbe
+	.long	0x243185be,0x4ee4b28c,0x550c7dc3,0xd5ffb4e2
+	.long	0x72be5d74,0xf27b896f,0x80deb1fe,0x3b1696b1
+	.long	0x9bdc06a7,0x25c71235,0xc19bf174,0xcf692694
+	.long	0xe49b69c1,0x9ef14ad2,0xefbe4786,0x384f25e3
+	.long	0x0fc19dc6,0x8b8cd5b5,0x240ca1cc,0x77ac9c65
+	.long	0x2de92c6f,0x592b0275,0x4a7484aa,0x6ea6e483
+	.long	0x5cb0a9dc,0xbd41fbd4,0x76f988da,0x831153b5
+	.long	0x983e5152,0xee66dfab,0xa831c66d,0x2db43210
+	.long	0xb00327c8,0x98fb213f,0xbf597fc7,0xbeef0ee4
+	.long	0xc6e00bf3,0x3da88fc2,0xd5a79147,0x930aa725
+	.long	0x06ca6351,0xe003826f,0x14292967,0x0a0e6e70
+	.long	0x27b70a85,0x46d22ffc,0x2e1b2138,0x5c26c926
+	.long	0x4d2c6dfc,0x5ac42aed,0x53380d13,0x9d95b3df
+	.long	0x650a7354,0x8baf63de,0x766a0abb,0x3c77b2a8
+	.long	0x81c2c92e,0x47edaee6,0x92722c85,0x1482353b
+	.long	0xa2bfe8a1,0x4cf10364,0xa81a664b,0xbc423001
+	.long	0xc24b8b70,0xd0f89791,0xc76c51a3,0x0654be30
+	.long	0xd192e819,0xd6ef5218,0xd6990624,0x5565a910
+	.long	0xf40e3585,0x5771202a,0x106aa070,0x32bbd1b8
+	.long	0x19a4c116,0xb8d2d0c8,0x1e376c08,0x5141ab53
+	.long	0x2748774c,0xdf8eeb99,0x34b0bcb5,0xe19b48a8
+	.long	0x391c0cb3,0xc5c95a63,0x4ed8aa4a,0xe3418acb
+	.long	0x5b9cca4f,0x7763e373,0x682e6ff3,0xd6b2b8a3
+	.long	0x748f82ee,0x5defb2fc,0x78a5636f,0x43172f60
+	.long	0x84c87814,0xa1f0ab72,0x8cc70208,0x1a6439ec
+	.long	0x90befffa,0x23631e28,0xa4506ceb,0xde82bde9
+	.long	0xbef9a3f7,0xb2c67915,0xc67178f2,0xe372532b
+	.long	0xca273ece,0xea26619c,0xd186b8c7,0x21c0c207
+	.long	0xeada7dd6,0xcde0eb1e,0xf57d4f7f,0xee6ed178
+	.long	0x06f067aa,0x72176fba,0x0a637dc5,0xa2c898a6
+	.long	0x113f9804,0xbef90dae,0x1b710b35,0x131c471b
+	.long	0x28db77f5,0x23047d84,0x32caab7b,0x40c72493
+	.long	0x3c9ebe0a,0x15c9bebc,0x431d67c4,0x9c100d4c
+	.long	0x4cc5d4be,0xcb3e42b6,0x597f299c,0xfc657e2a
+	.long	0x5fcb6fab,0x3ad6faec,0x6c44198c,0x4a475817
+___
+$code.=<<___ if ($SZ==4);
+	.long	0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5
+	.long	0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5
+	.long	0xd807aa98,0x12835b01,0x243185be,0x550c7dc3
+	.long	0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174
+	.long	0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc
+	.long	0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da
+	.long	0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7
+	.long	0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967
+	.long	0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13
+	.long	0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85
+	.long	0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3
+	.long	0xd192e819,0xd6990624,0xf40e3585,0x106aa070
+	.long	0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5
+	.long	0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3
+	.long	0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208
+	.long	0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha512-s390x.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha512-s390x.pl
new file mode 100644
index 0000000000..079a3fc78a
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha512-s390x.pl
@@ -0,0 +1,322 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# SHA256/512 block procedures for s390x.
+
+# April 2007.
+#
+# sha256_block_data_order is reportedly >3 times faster than gcc 3.3
+# generated code (must be a bug in compiler, as improvement is
+# "pathologically" high, in particular in comparison to other SHA
+# modules). But the real twist is that it detects if hardware support
+# for SHA256 is available and in such case utilizes it. Then the
+# performance can reach >6.5x of assembler one for larger chunks.
+#
+# sha512_block_data_order is ~70% faster than gcc 3.3 generated code.
+
+# January 2009.
+#
+# Add support for hardware SHA512 and reschedule instructions to
+# favour dual-issue z10 pipeline. Hardware SHA256/512 is ~4.7x faster
+# than software.
+
+# November 2010.
+#
+# Adapt for -m31 build. If kernel supports what's called "highgprs"
+# feature on Linux [see /proc/cpuinfo], it's possible to use 64-bit
+# instructions and achieve "64-bit" performance even in 31-bit legacy
+# application context. The feature is not specific to any particular
+# processor, as long as it's "z-CPU". Latter implies that the code
+# remains z/Architecture specific. On z900 SHA256 was measured to
+# perform 2.4x and SHA512 - 13x better than code generated by gcc 4.3.
+
+$flavour = shift;
+
+if ($flavour =~ /3[12]/) {
+	$SIZE_T=4;
+	$g="";
+} else {
+	$SIZE_T=8;
+	$g="g";
+}
+
+$t0="%r0";
+$t1="%r1";
+$ctx="%r2";	$t2="%r2";
+$inp="%r3";
+$len="%r4";	# used as index in inner loop
+
+$A="%r5";
+$B="%r6";
+$C="%r7";
+$D="%r8";
+$E="%r9";
+$F="%r10";
+$G="%r11";
+$H="%r12";	@V=($A,$B,$C,$D,$E,$F,$G,$H);
+$tbl="%r13";
+$T1="%r14";
+$sp="%r15";
+
+while (($output=shift) && ($output!~/^\w[\w\-]*\.\w+$/)) {}
+open STDOUT,">$output";
+
+if ($output =~ /512/) {
+	$label="512";
+	$SZ=8;
+	$LD="lg";	# load from memory
+	$ST="stg";	# store to memory
+	$ADD="alg";	# add with memory operand
+	$ROT="rllg";	# rotate left
+	$SHR="srlg";	# logical right shift [see even at the end]
+	@Sigma0=(25,30,36);
+	@Sigma1=(23,46,50);
+	@sigma0=(56,63, 7);
+	@sigma1=( 3,45, 6);
+	$rounds=80;
+	$kimdfunc=3;	# 0 means unknown/unsupported/unimplemented/disabled
+} else {
+	$label="256";
+	$SZ=4;
+	$LD="llgf";	# load from memory
+	$ST="st";	# store to memory
+	$ADD="al";	# add with memory operand
+	$ROT="rll";	# rotate left
+	$SHR="srl";	# logical right shift
+	@Sigma0=(10,19,30);
+	@Sigma1=( 7,21,26);
+	@sigma0=(14,25, 3);
+	@sigma1=(13,15,10);
+	$rounds=64;
+	$kimdfunc=2;	# magic function code for kimd instruction
+}
+$Func="sha${label}_block_data_order";
+$Table="K${label}";
+$stdframe=16*$SIZE_T+4*8;
+$frame=$stdframe+16*$SZ;
+
+sub BODY_00_15 {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h) = @_;
+
+$code.=<<___ if ($i<16);
+	$LD	$T1,`$i*$SZ`($inp)	### $i
+___
+$code.=<<___;
+	$ROT	$t0,$e,$Sigma1[0]
+	$ROT	$t1,$e,$Sigma1[1]
+	 lgr	$t2,$f
+	xgr	$t0,$t1
+	$ROT	$t1,$t1,`$Sigma1[2]-$Sigma1[1]`
+	 xgr	$t2,$g
+	$ST	$T1,`$stdframe+$SZ*($i%16)`($sp)
+	xgr	$t0,$t1			# Sigma1(e)
+	algr	$T1,$h			# T1+=h
+	 ngr	$t2,$e
+	 lgr	$t1,$a
+	algr	$T1,$t0			# T1+=Sigma1(e)
+	$ROT	$h,$a,$Sigma0[0]
+	 xgr	$t2,$g			# Ch(e,f,g)
+	$ADD	$T1,`$i*$SZ`($len,$tbl)	# T1+=K[i]
+	$ROT	$t0,$a,$Sigma0[1]
+	algr	$T1,$t2			# T1+=Ch(e,f,g)
+	 ogr	$t1,$b
+	xgr	$h,$t0
+	 lgr	$t2,$a
+	 ngr	$t1,$c
+	$ROT	$t0,$t0,`$Sigma0[2]-$Sigma0[1]`
+	xgr	$h,$t0			# h=Sigma0(a)
+	 ngr	$t2,$b
+	algr	$h,$T1			# h+=T1
+	 ogr	$t2,$t1			# Maj(a,b,c)
+	algr	$d,$T1			# d+=T1
+	algr	$h,$t2			# h+=Maj(a,b,c)
+___
+}
+
+sub BODY_16_XX {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h) = @_;
+
+$code.=<<___;
+	$LD	$T1,`$stdframe+$SZ*(($i+1)%16)`($sp)	### $i
+	$LD	$t1,`$stdframe+$SZ*(($i+14)%16)`($sp)
+	$ROT	$t0,$T1,$sigma0[0]
+	$SHR	$T1,$sigma0[2]
+	$ROT	$t2,$t0,`$sigma0[1]-$sigma0[0]`
+	xgr	$T1,$t0
+	$ROT	$t0,$t1,$sigma1[0]
+	xgr	$T1,$t2					# sigma0(X[i+1])
+	$SHR	$t1,$sigma1[2]
+	$ADD	$T1,`$stdframe+$SZ*($i%16)`($sp)	# +=X[i]
+	xgr	$t1,$t0
+	$ROT	$t0,$t0,`$sigma1[1]-$sigma1[0]`
+	$ADD	$T1,`$stdframe+$SZ*(($i+9)%16)`($sp)	# +=X[i+9]
+	xgr	$t1,$t0				# sigma1(X[i+14])
+	algr	$T1,$t1				# +=sigma1(X[i+14])
+___
+	&BODY_00_15(@_);
+}
+
+$code.=<<___;
+.text
+.align	64
+.type	$Table,\@object
+$Table:
+___
+$code.=<<___ if ($SZ==4);
+	.long	0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5
+	.long	0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5
+	.long	0xd807aa98,0x12835b01,0x243185be,0x550c7dc3
+	.long	0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174
+	.long	0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc
+	.long	0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da
+	.long	0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7
+	.long	0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967
+	.long	0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13
+	.long	0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85
+	.long	0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3
+	.long	0xd192e819,0xd6990624,0xf40e3585,0x106aa070
+	.long	0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5
+	.long	0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3
+	.long	0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208
+	.long	0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
+___
+$code.=<<___ if ($SZ==8);
+	.quad	0x428a2f98d728ae22,0x7137449123ef65cd
+	.quad	0xb5c0fbcfec4d3b2f,0xe9b5dba58189dbbc
+	.quad	0x3956c25bf348b538,0x59f111f1b605d019
+	.quad	0x923f82a4af194f9b,0xab1c5ed5da6d8118
+	.quad	0xd807aa98a3030242,0x12835b0145706fbe
+	.quad	0x243185be4ee4b28c,0x550c7dc3d5ffb4e2
+	.quad	0x72be5d74f27b896f,0x80deb1fe3b1696b1
+	.quad	0x9bdc06a725c71235,0xc19bf174cf692694
+	.quad	0xe49b69c19ef14ad2,0xefbe4786384f25e3
+	.quad	0x0fc19dc68b8cd5b5,0x240ca1cc77ac9c65
+	.quad	0x2de92c6f592b0275,0x4a7484aa6ea6e483
+	.quad	0x5cb0a9dcbd41fbd4,0x76f988da831153b5
+	.quad	0x983e5152ee66dfab,0xa831c66d2db43210
+	.quad	0xb00327c898fb213f,0xbf597fc7beef0ee4
+	.quad	0xc6e00bf33da88fc2,0xd5a79147930aa725
+	.quad	0x06ca6351e003826f,0x142929670a0e6e70
+	.quad	0x27b70a8546d22ffc,0x2e1b21385c26c926
+	.quad	0x4d2c6dfc5ac42aed,0x53380d139d95b3df
+	.quad	0x650a73548baf63de,0x766a0abb3c77b2a8
+	.quad	0x81c2c92e47edaee6,0x92722c851482353b
+	.quad	0xa2bfe8a14cf10364,0xa81a664bbc423001
+	.quad	0xc24b8b70d0f89791,0xc76c51a30654be30
+	.quad	0xd192e819d6ef5218,0xd69906245565a910
+	.quad	0xf40e35855771202a,0x106aa07032bbd1b8
+	.quad	0x19a4c116b8d2d0c8,0x1e376c085141ab53
+	.quad	0x2748774cdf8eeb99,0x34b0bcb5e19b48a8
+	.quad	0x391c0cb3c5c95a63,0x4ed8aa4ae3418acb
+	.quad	0x5b9cca4f7763e373,0x682e6ff3d6b2b8a3
+	.quad	0x748f82ee5defb2fc,0x78a5636f43172f60
+	.quad	0x84c87814a1f0ab72,0x8cc702081a6439ec
+	.quad	0x90befffa23631e28,0xa4506cebde82bde9
+	.quad	0xbef9a3f7b2c67915,0xc67178f2e372532b
+	.quad	0xca273eceea26619c,0xd186b8c721c0c207
+	.quad	0xeada7dd6cde0eb1e,0xf57d4f7fee6ed178
+	.quad	0x06f067aa72176fba,0x0a637dc5a2c898a6
+	.quad	0x113f9804bef90dae,0x1b710b35131c471b
+	.quad	0x28db77f523047d84,0x32caab7b40c72493
+	.quad	0x3c9ebe0a15c9bebc,0x431d67c49c100d4c
+	.quad	0x4cc5d4becb3e42b6,0x597f299cfc657e2a
+	.quad	0x5fcb6fab3ad6faec,0x6c44198c4a475817
+___
+$code.=<<___;
+.size	$Table,.-$Table
+.globl	$Func
+.type	$Func,\@function
+$Func:
+	sllg	$len,$len,`log(16*$SZ)/log(2)`
+___
+$code.=<<___ if ($kimdfunc);
+	larl	%r1,OPENSSL_s390xcap_P
+	lg	%r0,0(%r1)
+	tmhl	%r0,0x4000	# check for message-security assist
+	jz	.Lsoftware
+	lghi	%r0,0
+	la	%r1,`2*$SIZE_T`($sp)
+	.long	0xb93e0002	# kimd %r0,%r2
+	lg	%r0,`2*$SIZE_T`($sp)
+	tmhh	%r0,`0x8000>>$kimdfunc`
+	jz	.Lsoftware
+	lghi	%r0,$kimdfunc
+	lgr	%r1,$ctx
+	lgr	%r2,$inp
+	lgr	%r3,$len
+	.long	0xb93e0002	# kimd %r0,%r2
+	brc	1,.-4		# pay attention to "partial completion"
+	br	%r14
+.align	16
+.Lsoftware:
+___
+$code.=<<___;
+	lghi	%r1,-$frame
+	la	$len,0($len,$inp)
+	stm${g}	$ctx,%r15,`2*$SIZE_T`($sp)
+	lgr	%r0,$sp
+	la	$sp,0(%r1,$sp)
+	st${g}	%r0,0($sp)
+
+	larl	$tbl,$Table
+	$LD	$A,`0*$SZ`($ctx)
+	$LD	$B,`1*$SZ`($ctx)
+	$LD	$C,`2*$SZ`($ctx)
+	$LD	$D,`3*$SZ`($ctx)
+	$LD	$E,`4*$SZ`($ctx)
+	$LD	$F,`5*$SZ`($ctx)
+	$LD	$G,`6*$SZ`($ctx)
+	$LD	$H,`7*$SZ`($ctx)
+
+.Lloop:
+	lghi	$len,0
+___
+for ($i=0;$i<16;$i++)	{ &BODY_00_15($i,@V); unshift(@V,pop(@V)); }
+$code.=".Lrounds_16_xx:\n";
+for (;$i<32;$i++)	{ &BODY_16_XX($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	aghi	$len,`16*$SZ`
+	lghi	$t0,`($rounds-16)*$SZ`
+	clgr	$len,$t0
+	jne	.Lrounds_16_xx
+
+	l${g}	$ctx,`$frame+2*$SIZE_T`($sp)
+	la	$inp,`16*$SZ`($inp)
+	$ADD	$A,`0*$SZ`($ctx)
+	$ADD	$B,`1*$SZ`($ctx)
+	$ADD	$C,`2*$SZ`($ctx)
+	$ADD	$D,`3*$SZ`($ctx)
+	$ADD	$E,`4*$SZ`($ctx)
+	$ADD	$F,`5*$SZ`($ctx)
+	$ADD	$G,`6*$SZ`($ctx)
+	$ADD	$H,`7*$SZ`($ctx)
+	$ST	$A,`0*$SZ`($ctx)
+	$ST	$B,`1*$SZ`($ctx)
+	$ST	$C,`2*$SZ`($ctx)
+	$ST	$D,`3*$SZ`($ctx)
+	$ST	$E,`4*$SZ`($ctx)
+	$ST	$F,`5*$SZ`($ctx)
+	$ST	$G,`6*$SZ`($ctx)
+	$ST	$H,`7*$SZ`($ctx)
+	cl${g}	$inp,`$frame+4*$SIZE_T`($sp)
+	jne	.Lloop
+
+	lm${g}	%r6,%r15,`$frame+6*$SIZE_T`($sp)	
+	br	%r14
+.size	$Func,.-$Func
+.string	"SHA${label} block transform for s390x, CRYPTOGAMS by <appro\@openssl.org>"
+.comm	OPENSSL_s390xcap_P,16,8
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+# unlike 32-bit shift 64-bit one takes three arguments
+$code =~ s/(srlg\s+)(%r[0-9]+),/$1$2,$2,/gm;
+
+print $code;
+close STDOUT;
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha512-sparcv9.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha512-sparcv9.pl
new file mode 100644
index 0000000000..585740789e
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha512-sparcv9.pl
@@ -0,0 +1,594 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> 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 performance improvement over compiler generated code varies
+# from 40% for Sun C [32-bit build] to 70% for gcc [3.3, 64-bit
+# build]. Just like in SHA1 module I aim to ensure scalability on
+# UltraSPARC T1 by packing X[16] to 8 64-bit registers.
+
+# SHA512 on pre-T1 UltraSPARC.
+#
+# Performance is >75% better than 64-bit code generated by Sun C and
+# over 2x than 32-bit code. X[16] resides on stack, but access to it
+# is scheduled for L2 latency and staged through 32 least significant
+# bits of %l0-%l7. The latter is done to achieve 32-/64-bit ABI
+# duality. Nevetheless it's ~40% faster than SHA256, which is pretty
+# good [optimal coefficient is 50%].
+#
+# SHA512 on UltraSPARC T1.
+#
+# It's not any faster than 64-bit code generated by Sun C 5.8. This is
+# because 64-bit code generator has the advantage of using 64-bit
+# loads(*) to access X[16], which I consciously traded for 32-/64-bit
+# ABI duality [as per above]. But it surpasses 32-bit Sun C generated
+# code by 60%, not to mention that it doesn't suffer from severe decay
+# when running 4 times physical cores threads and that it leaves gcc
+# [3.4] behind by over 4x factor! If compared to SHA256, single thread
+# performance is only 10% better, but overall throughput for maximum
+# amount of threads for given CPU exceeds corresponding one of SHA256
+# by 30% [again, optimal coefficient is 50%].
+#
+# (*)	Unlike pre-T1 UltraSPARC loads on T1 are executed strictly
+#	in-order, i.e. load instruction has to complete prior next
+#	instruction in given thread is executed, even if the latter is
+#	not dependent on load result! This means that on T1 two 32-bit
+#	loads are always slower than one 64-bit load. Once again this
+#	is unlike pre-T1 UltraSPARC, where, if scheduled appropriately,
+#	2x32-bit loads can be as fast as 1x64-bit ones.
+
+$bits=32;
+for (@ARGV)	{ $bits=64 if (/\-m64/ || /\-xarch\=v9/); }
+if ($bits==64)	{ $bias=2047; $frame=192; }
+else		{ $bias=0;    $frame=112; }
+
+$output=shift;
+open STDOUT,">$output";
+
+if ($output =~ /512/) {
+	$label="512";
+	$SZ=8;
+	$LD="ldx";		# load from memory
+	$ST="stx";		# store to memory
+	$SLL="sllx";		# shift left logical
+	$SRL="srlx";		# shift right logical
+	@Sigma0=(28,34,39);
+	@Sigma1=(14,18,41);
+	@sigma0=( 7, 1, 8);	# right shift first
+	@sigma1=( 6,19,61);	# right shift first
+	$lastK=0x817;
+	$rounds=80;
+	$align=4;
+
+	$locals=16*$SZ;		# X[16]
+
+	$A="%o0";
+	$B="%o1";
+	$C="%o2";
+	$D="%o3";
+	$E="%o4";
+	$F="%o5";
+	$G="%g1";
+	$H="%o7";
+	@V=($A,$B,$C,$D,$E,$F,$G,$H);
+} else {
+	$label="256";
+	$SZ=4;
+	$LD="ld";		# load from memory
+	$ST="st";		# store to memory
+	$SLL="sll";		# shift left logical
+	$SRL="srl";		# shift right logical
+	@Sigma0=( 2,13,22);
+	@Sigma1=( 6,11,25);
+	@sigma0=( 3, 7,18);	# right shift first
+	@sigma1=(10,17,19);	# right shift first
+	$lastK=0x8f2;
+	$rounds=64;
+	$align=8;
+
+	$locals=0;		# X[16] is register resident
+	@X=("%o0","%o1","%o2","%o3","%o4","%o5","%g1","%o7");
+	
+	$A="%l0";
+	$B="%l1";
+	$C="%l2";
+	$D="%l3";
+	$E="%l4";
+	$F="%l5";
+	$G="%l6";
+	$H="%l7";
+	@V=($A,$B,$C,$D,$E,$F,$G,$H);
+}
+$T1="%g2";
+$tmp0="%g3";
+$tmp1="%g4";
+$tmp2="%g5";
+
+$ctx="%i0";
+$inp="%i1";
+$len="%i2";
+$Ktbl="%i3";
+$tmp31="%i4";
+$tmp32="%i5";
+
+########### SHA256
+$Xload = sub {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_;
+
+    if ($i==0) {
+$code.=<<___;
+	ldx	[$inp+0],@X[0]
+	ldx	[$inp+16],@X[2]
+	ldx	[$inp+32],@X[4]
+	ldx	[$inp+48],@X[6]
+	ldx	[$inp+8],@X[1]
+	ldx	[$inp+24],@X[3]
+	subcc	%g0,$tmp31,$tmp32 ! should be 64-$tmp31, but -$tmp31 works too
+	ldx	[$inp+40],@X[5]
+	bz,pt	%icc,.Laligned
+	ldx	[$inp+56],@X[7]
+
+	sllx	@X[0],$tmp31,@X[0]
+	ldx	[$inp+64],$T1
+___
+for($j=0;$j<7;$j++)
+{   $code.=<<___;
+	srlx	@X[$j+1],$tmp32,$tmp1
+	sllx	@X[$j+1],$tmp31,@X[$j+1]
+	or	$tmp1,@X[$j],@X[$j]
+___
+}
+$code.=<<___;
+	srlx	$T1,$tmp32,$T1
+	or	$T1,@X[7],@X[7]
+.Laligned:
+___
+    }
+
+    if ($i&1) {
+	$code.="\tadd	@X[$i/2],$h,$T1\n";
+    } else {
+	$code.="\tsrlx	@X[$i/2],32,$T1\n\tadd	$h,$T1,$T1\n";
+    }
+} if ($SZ==4);
+
+########### SHA512
+$Xload = sub {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_;
+my @pair=("%l".eval(($i*2)%8),"%l".eval(($i*2)%8+1),"%l".eval((($i+1)*2)%8));
+
+$code.=<<___ if ($i==0);
+	ld	[$inp+0],%l0
+	ld	[$inp+4],%l1
+	ld	[$inp+8],%l2
+	ld	[$inp+12],%l3
+	ld	[$inp+16],%l4
+	ld	[$inp+20],%l5
+	ld	[$inp+24],%l6
+	ld	[$inp+28],%l7
+___
+$code.=<<___ if ($i<15);
+	sllx	@pair[1],$tmp31,$tmp2	! Xload($i)
+	add	$tmp31,32,$tmp0
+	sllx	@pair[0],$tmp0,$tmp1
+	`"ld	[$inp+".eval(32+0+$i*8)."],@pair[0]"	if ($i<12)`
+	srlx	@pair[2],$tmp32,@pair[1]
+	or	$tmp1,$tmp2,$tmp2
+	or	@pair[1],$tmp2,$tmp2
+	`"ld	[$inp+".eval(32+4+$i*8)."],@pair[1]"	if ($i<12)`
+	add	$h,$tmp2,$T1
+	$ST	$tmp2,[%sp+`$bias+$frame+$i*$SZ`]
+___
+$code.=<<___ if ($i==12);
+	brnz,a	$tmp31,.+8
+	ld	[$inp+128],%l0
+___
+$code.=<<___ if ($i==15);
+	ld	[%sp+`$bias+$frame+(($i+1+1)%16)*$SZ+0`],%l2
+	sllx	@pair[1],$tmp31,$tmp2	! Xload($i)
+	add	$tmp31,32,$tmp0
+	ld	[%sp+`$bias+$frame+(($i+1+1)%16)*$SZ+4`],%l3
+	sllx	@pair[0],$tmp0,$tmp1
+	ld	[%sp+`$bias+$frame+(($i+1+9)%16)*$SZ+0`],%l4
+	srlx	@pair[2],$tmp32,@pair[1]
+	or	$tmp1,$tmp2,$tmp2
+	ld	[%sp+`$bias+$frame+(($i+1+9)%16)*$SZ+4`],%l5
+	or	@pair[1],$tmp2,$tmp2
+	ld	[%sp+`$bias+$frame+(($i+1+14)%16)*$SZ+0`],%l6
+	add	$h,$tmp2,$T1
+	$ST	$tmp2,[%sp+`$bias+$frame+$i*$SZ`]
+	ld	[%sp+`$bias+$frame+(($i+1+14)%16)*$SZ+4`],%l7
+	ld	[%sp+`$bias+$frame+(($i+1+0)%16)*$SZ+0`],%l0
+	ld	[%sp+`$bias+$frame+(($i+1+0)%16)*$SZ+4`],%l1
+___
+} if ($SZ==8);
+
+########### common
+sub BODY_00_15 {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h)=@_;
+
+    if ($i<16) {
+	&$Xload(@_);
+    } else {
+	$code.="\tadd	$h,$T1,$T1\n";
+    }
+
+$code.=<<___;
+	$SRL	$e,@Sigma1[0],$h	!! $i
+	xor	$f,$g,$tmp2
+	$SLL	$e,`$SZ*8-@Sigma1[2]`,$tmp1
+	and	$e,$tmp2,$tmp2
+	$SRL	$e,@Sigma1[1],$tmp0
+	xor	$tmp1,$h,$h
+	$SLL	$e,`$SZ*8-@Sigma1[1]`,$tmp1
+	xor	$tmp0,$h,$h
+	$SRL	$e,@Sigma1[2],$tmp0
+	xor	$tmp1,$h,$h
+	$SLL	$e,`$SZ*8-@Sigma1[0]`,$tmp1
+	xor	$tmp0,$h,$h
+	xor	$g,$tmp2,$tmp2		! Ch(e,f,g)
+	xor	$tmp1,$h,$tmp0		! Sigma1(e)
+
+	$SRL	$a,@Sigma0[0],$h
+	add	$tmp2,$T1,$T1
+	$LD	[$Ktbl+`$i*$SZ`],$tmp2	! K[$i]
+	$SLL	$a,`$SZ*8-@Sigma0[2]`,$tmp1
+	add	$tmp0,$T1,$T1
+	$SRL	$a,@Sigma0[1],$tmp0
+	xor	$tmp1,$h,$h
+	$SLL	$a,`$SZ*8-@Sigma0[1]`,$tmp1
+	xor	$tmp0,$h,$h
+	$SRL	$a,@Sigma0[2],$tmp0
+	xor	$tmp1,$h,$h	
+	$SLL	$a,`$SZ*8-@Sigma0[0]`,$tmp1
+	xor	$tmp0,$h,$h
+	xor	$tmp1,$h,$h		! Sigma0(a)
+
+	or	$a,$b,$tmp0
+	and	$a,$b,$tmp1
+	and	$c,$tmp0,$tmp0
+	or	$tmp0,$tmp1,$tmp1	! Maj(a,b,c)
+	add	$tmp2,$T1,$T1		! +=K[$i]
+	add	$tmp1,$h,$h
+
+	add	$T1,$d,$d
+	add	$T1,$h,$h
+___
+}
+
+########### SHA256
+$BODY_16_XX = sub {
+my $i=@_[0];
+my $xi;
+
+    if ($i&1) {
+	$xi=$tmp32;
+	$code.="\tsrlx	@X[(($i+1)/2)%8],32,$xi\n";
+    } else {
+	$xi=@X[(($i+1)/2)%8];
+    }
+$code.=<<___;
+	srl	$xi,@sigma0[0],$T1		!! Xupdate($i)
+	sll	$xi,`32-@sigma0[2]`,$tmp1
+	srl	$xi,@sigma0[1],$tmp0
+	xor	$tmp1,$T1,$T1
+	sll	$tmp1,`@sigma0[2]-@sigma0[1]`,$tmp1
+	xor	$tmp0,$T1,$T1
+	srl	$xi,@sigma0[2],$tmp0
+	xor	$tmp1,$T1,$T1
+___
+    if ($i&1) {
+	$xi=@X[(($i+14)/2)%8];
+    } else {
+	$xi=$tmp32;
+	$code.="\tsrlx	@X[(($i+14)/2)%8],32,$xi\n";
+    }
+$code.=<<___;
+	srl	$xi,@sigma1[0],$tmp2
+	xor	$tmp0,$T1,$T1			! T1=sigma0(X[i+1])
+	sll	$xi,`32-@sigma1[2]`,$tmp1
+	srl	$xi,@sigma1[1],$tmp0
+	xor	$tmp1,$tmp2,$tmp2
+	sll	$tmp1,`@sigma1[2]-@sigma1[1]`,$tmp1
+	xor	$tmp0,$tmp2,$tmp2
+	srl	$xi,@sigma1[2],$tmp0
+	xor	$tmp1,$tmp2,$tmp2
+___
+    if ($i&1) {
+	$xi=@X[($i/2)%8];
+$code.=<<___;
+	srlx	@X[(($i+9)/2)%8],32,$tmp1	! X[i+9]
+	xor	$tmp0,$tmp2,$tmp2		! sigma1(X[i+14])
+	srl	@X[($i/2)%8],0,$tmp0
+	add	$tmp2,$tmp1,$tmp1
+	add	$xi,$T1,$T1			! +=X[i]
+	xor	$tmp0,@X[($i/2)%8],@X[($i/2)%8]
+	add	$tmp1,$T1,$T1
+
+	srl	$T1,0,$T1
+	or	$T1,@X[($i/2)%8],@X[($i/2)%8]
+___
+    } else {
+	$xi=@X[(($i+9)/2)%8];
+$code.=<<___;
+	srlx	@X[($i/2)%8],32,$tmp1		! X[i]
+	xor	$tmp0,$tmp2,$tmp2		! sigma1(X[i+14])
+	add	$xi,$T1,$T1			! +=X[i+9]
+	add	$tmp2,$tmp1,$tmp1
+	srl	@X[($i/2)%8],0,@X[($i/2)%8]
+	add	$tmp1,$T1,$T1
+
+	sllx	$T1,32,$tmp0
+	or	$tmp0,@X[($i/2)%8],@X[($i/2)%8]
+___
+    }
+    &BODY_00_15(@_);
+} if ($SZ==4);
+
+########### SHA512
+$BODY_16_XX = sub {
+my $i=@_[0];
+my @pair=("%l".eval(($i*2)%8),"%l".eval(($i*2)%8+1));
+
+$code.=<<___;
+	sllx	%l2,32,$tmp0		!! Xupdate($i)
+	or	%l3,$tmp0,$tmp0
+
+	srlx	$tmp0,@sigma0[0],$T1
+	ld	[%sp+`$bias+$frame+(($i+1+1)%16)*$SZ+0`],%l2
+	sllx	$tmp0,`64-@sigma0[2]`,$tmp1
+	ld	[%sp+`$bias+$frame+(($i+1+1)%16)*$SZ+4`],%l3
+	srlx	$tmp0,@sigma0[1],$tmp0
+	xor	$tmp1,$T1,$T1
+	sllx	$tmp1,`@sigma0[2]-@sigma0[1]`,$tmp1
+	xor	$tmp0,$T1,$T1
+	srlx	$tmp0,`@sigma0[2]-@sigma0[1]`,$tmp0
+	xor	$tmp1,$T1,$T1
+	sllx	%l6,32,$tmp2
+	xor	$tmp0,$T1,$T1		! sigma0(X[$i+1])
+	or	%l7,$tmp2,$tmp2
+
+	srlx	$tmp2,@sigma1[0],$tmp1
+	ld	[%sp+`$bias+$frame+(($i+1+14)%16)*$SZ+0`],%l6
+	sllx	$tmp2,`64-@sigma1[2]`,$tmp0
+	ld	[%sp+`$bias+$frame+(($i+1+14)%16)*$SZ+4`],%l7
+	srlx	$tmp2,@sigma1[1],$tmp2
+	xor	$tmp0,$tmp1,$tmp1
+	sllx	$tmp0,`@sigma1[2]-@sigma1[1]`,$tmp0
+	xor	$tmp2,$tmp1,$tmp1
+	srlx	$tmp2,`@sigma1[2]-@sigma1[1]`,$tmp2
+	xor	$tmp0,$tmp1,$tmp1
+	sllx	%l4,32,$tmp0
+	xor	$tmp2,$tmp1,$tmp1	! sigma1(X[$i+14])
+	ld	[%sp+`$bias+$frame+(($i+1+9)%16)*$SZ+0`],%l4
+	or	%l5,$tmp0,$tmp0
+	ld	[%sp+`$bias+$frame+(($i+1+9)%16)*$SZ+4`],%l5
+
+	sllx	%l0,32,$tmp2
+	add	$tmp1,$T1,$T1
+	ld	[%sp+`$bias+$frame+(($i+1+0)%16)*$SZ+0`],%l0
+	or	%l1,$tmp2,$tmp2
+	add	$tmp0,$T1,$T1		! +=X[$i+9]
+	ld	[%sp+`$bias+$frame+(($i+1+0)%16)*$SZ+4`],%l1
+	add	$tmp2,$T1,$T1		! +=X[$i]
+	$ST	$T1,[%sp+`$bias+$frame+($i%16)*$SZ`]
+___
+    &BODY_00_15(@_);
+} if ($SZ==8);
+
+$code.=<<___ if ($bits==64);
+.register	%g2,#scratch
+.register	%g3,#scratch
+___
+$code.=<<___;
+.section	".text",#alloc,#execinstr
+
+.align	64
+K${label}:
+.type	K${label},#object
+___
+if ($SZ==4) {
+$code.=<<___;
+	.long	0x428a2f98, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5
+	.long	0x3956c25b, 0x59f111f1, 0x923f82a4, 0xab1c5ed5
+	.long	0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3
+	.long	0x72be5d74, 0x80deb1fe, 0x9bdc06a7, 0xc19bf174
+	.long	0xe49b69c1, 0xefbe4786, 0x0fc19dc6, 0x240ca1cc
+	.long	0x2de92c6f, 0x4a7484aa, 0x5cb0a9dc, 0x76f988da
+	.long	0x983e5152, 0xa831c66d, 0xb00327c8, 0xbf597fc7
+	.long	0xc6e00bf3, 0xd5a79147, 0x06ca6351, 0x14292967
+	.long	0x27b70a85, 0x2e1b2138, 0x4d2c6dfc, 0x53380d13
+	.long	0x650a7354, 0x766a0abb, 0x81c2c92e, 0x92722c85
+	.long	0xa2bfe8a1, 0xa81a664b, 0xc24b8b70, 0xc76c51a3
+	.long	0xd192e819, 0xd6990624, 0xf40e3585, 0x106aa070
+	.long	0x19a4c116, 0x1e376c08, 0x2748774c, 0x34b0bcb5
+	.long	0x391c0cb3, 0x4ed8aa4a, 0x5b9cca4f, 0x682e6ff3
+	.long	0x748f82ee, 0x78a5636f, 0x84c87814, 0x8cc70208
+	.long	0x90befffa, 0xa4506ceb, 0xbef9a3f7, 0xc67178f2
+___
+} else {
+$code.=<<___;
+	.long	0x428a2f98,0xd728ae22, 0x71374491,0x23ef65cd
+	.long	0xb5c0fbcf,0xec4d3b2f, 0xe9b5dba5,0x8189dbbc
+	.long	0x3956c25b,0xf348b538, 0x59f111f1,0xb605d019
+	.long	0x923f82a4,0xaf194f9b, 0xab1c5ed5,0xda6d8118
+	.long	0xd807aa98,0xa3030242, 0x12835b01,0x45706fbe
+	.long	0x243185be,0x4ee4b28c, 0x550c7dc3,0xd5ffb4e2
+	.long	0x72be5d74,0xf27b896f, 0x80deb1fe,0x3b1696b1
+	.long	0x9bdc06a7,0x25c71235, 0xc19bf174,0xcf692694
+	.long	0xe49b69c1,0x9ef14ad2, 0xefbe4786,0x384f25e3
+	.long	0x0fc19dc6,0x8b8cd5b5, 0x240ca1cc,0x77ac9c65
+	.long	0x2de92c6f,0x592b0275, 0x4a7484aa,0x6ea6e483
+	.long	0x5cb0a9dc,0xbd41fbd4, 0x76f988da,0x831153b5
+	.long	0x983e5152,0xee66dfab, 0xa831c66d,0x2db43210
+	.long	0xb00327c8,0x98fb213f, 0xbf597fc7,0xbeef0ee4
+	.long	0xc6e00bf3,0x3da88fc2, 0xd5a79147,0x930aa725
+	.long	0x06ca6351,0xe003826f, 0x14292967,0x0a0e6e70
+	.long	0x27b70a85,0x46d22ffc, 0x2e1b2138,0x5c26c926
+	.long	0x4d2c6dfc,0x5ac42aed, 0x53380d13,0x9d95b3df
+	.long	0x650a7354,0x8baf63de, 0x766a0abb,0x3c77b2a8
+	.long	0x81c2c92e,0x47edaee6, 0x92722c85,0x1482353b
+	.long	0xa2bfe8a1,0x4cf10364, 0xa81a664b,0xbc423001
+	.long	0xc24b8b70,0xd0f89791, 0xc76c51a3,0x0654be30
+	.long	0xd192e819,0xd6ef5218, 0xd6990624,0x5565a910
+	.long	0xf40e3585,0x5771202a, 0x106aa070,0x32bbd1b8
+	.long	0x19a4c116,0xb8d2d0c8, 0x1e376c08,0x5141ab53
+	.long	0x2748774c,0xdf8eeb99, 0x34b0bcb5,0xe19b48a8
+	.long	0x391c0cb3,0xc5c95a63, 0x4ed8aa4a,0xe3418acb
+	.long	0x5b9cca4f,0x7763e373, 0x682e6ff3,0xd6b2b8a3
+	.long	0x748f82ee,0x5defb2fc, 0x78a5636f,0x43172f60
+	.long	0x84c87814,0xa1f0ab72, 0x8cc70208,0x1a6439ec
+	.long	0x90befffa,0x23631e28, 0xa4506ceb,0xde82bde9
+	.long	0xbef9a3f7,0xb2c67915, 0xc67178f2,0xe372532b
+	.long	0xca273ece,0xea26619c, 0xd186b8c7,0x21c0c207
+	.long	0xeada7dd6,0xcde0eb1e, 0xf57d4f7f,0xee6ed178
+	.long	0x06f067aa,0x72176fba, 0x0a637dc5,0xa2c898a6
+	.long	0x113f9804,0xbef90dae, 0x1b710b35,0x131c471b
+	.long	0x28db77f5,0x23047d84, 0x32caab7b,0x40c72493
+	.long	0x3c9ebe0a,0x15c9bebc, 0x431d67c4,0x9c100d4c
+	.long	0x4cc5d4be,0xcb3e42b6, 0x597f299c,0xfc657e2a
+	.long	0x5fcb6fab,0x3ad6faec, 0x6c44198c,0x4a475817
+___
+}
+$code.=<<___;
+.size	K${label},.-K${label}
+.globl	sha${label}_block_data_order
+sha${label}_block_data_order:
+	save	%sp,`-$frame-$locals`,%sp
+	and	$inp,`$align-1`,$tmp31
+	sllx	$len,`log(16*$SZ)/log(2)`,$len
+	andn	$inp,`$align-1`,$inp
+	sll	$tmp31,3,$tmp31
+	add	$inp,$len,$len
+___
+$code.=<<___ if ($SZ==8); # SHA512
+	mov	32,$tmp32
+	sub	$tmp32,$tmp31,$tmp32
+___
+$code.=<<___;
+.Lpic:	call	.+8
+	add	%o7,K${label}-.Lpic,$Ktbl
+
+	$LD	[$ctx+`0*$SZ`],$A
+	$LD	[$ctx+`1*$SZ`],$B
+	$LD	[$ctx+`2*$SZ`],$C
+	$LD	[$ctx+`3*$SZ`],$D
+	$LD	[$ctx+`4*$SZ`],$E
+	$LD	[$ctx+`5*$SZ`],$F
+	$LD	[$ctx+`6*$SZ`],$G
+	$LD	[$ctx+`7*$SZ`],$H
+
+.Lloop:
+___
+for ($i=0;$i<16;$i++)	{ &BODY_00_15($i,@V); unshift(@V,pop(@V)); }
+$code.=".L16_xx:\n";
+for (;$i<32;$i++)	{ &$BODY_16_XX($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	and	$tmp2,0xfff,$tmp2
+	cmp	$tmp2,$lastK
+	bne	.L16_xx
+	add	$Ktbl,`16*$SZ`,$Ktbl	! Ktbl+=16
+
+___
+$code.=<<___ if ($SZ==4); # SHA256
+	$LD	[$ctx+`0*$SZ`],@X[0]
+	$LD	[$ctx+`1*$SZ`],@X[1]
+	$LD	[$ctx+`2*$SZ`],@X[2]
+	$LD	[$ctx+`3*$SZ`],@X[3]
+	$LD	[$ctx+`4*$SZ`],@X[4]
+	$LD	[$ctx+`5*$SZ`],@X[5]
+	$LD	[$ctx+`6*$SZ`],@X[6]
+	$LD	[$ctx+`7*$SZ`],@X[7]
+
+	add	$A,@X[0],$A
+	$ST	$A,[$ctx+`0*$SZ`]
+	add	$B,@X[1],$B
+	$ST	$B,[$ctx+`1*$SZ`]
+	add	$C,@X[2],$C
+	$ST	$C,[$ctx+`2*$SZ`]
+	add	$D,@X[3],$D
+	$ST	$D,[$ctx+`3*$SZ`]
+	add	$E,@X[4],$E
+	$ST	$E,[$ctx+`4*$SZ`]
+	add	$F,@X[5],$F
+	$ST	$F,[$ctx+`5*$SZ`]
+	add	$G,@X[6],$G
+	$ST	$G,[$ctx+`6*$SZ`]
+	add	$H,@X[7],$H
+	$ST	$H,[$ctx+`7*$SZ`]
+___
+$code.=<<___ if ($SZ==8); # SHA512
+	ld	[$ctx+`0*$SZ+0`],%l0
+	ld	[$ctx+`0*$SZ+4`],%l1
+	ld	[$ctx+`1*$SZ+0`],%l2
+	ld	[$ctx+`1*$SZ+4`],%l3
+	ld	[$ctx+`2*$SZ+0`],%l4
+	ld	[$ctx+`2*$SZ+4`],%l5
+	ld	[$ctx+`3*$SZ+0`],%l6
+
+	sllx	%l0,32,$tmp0
+	ld	[$ctx+`3*$SZ+4`],%l7
+	sllx	%l2,32,$tmp1
+	or	%l1,$tmp0,$tmp0
+	or	%l3,$tmp1,$tmp1
+	add	$tmp0,$A,$A
+	add	$tmp1,$B,$B
+	$ST	$A,[$ctx+`0*$SZ`]
+	sllx	%l4,32,$tmp2
+	$ST	$B,[$ctx+`1*$SZ`]
+	sllx	%l6,32,$T1
+	or	%l5,$tmp2,$tmp2
+	or	%l7,$T1,$T1
+	add	$tmp2,$C,$C
+	$ST	$C,[$ctx+`2*$SZ`]
+	add	$T1,$D,$D
+	$ST	$D,[$ctx+`3*$SZ`]
+
+	ld	[$ctx+`4*$SZ+0`],%l0
+	ld	[$ctx+`4*$SZ+4`],%l1
+	ld	[$ctx+`5*$SZ+0`],%l2
+	ld	[$ctx+`5*$SZ+4`],%l3
+	ld	[$ctx+`6*$SZ+0`],%l4
+	ld	[$ctx+`6*$SZ+4`],%l5
+	ld	[$ctx+`7*$SZ+0`],%l6
+
+	sllx	%l0,32,$tmp0
+	ld	[$ctx+`7*$SZ+4`],%l7
+	sllx	%l2,32,$tmp1
+	or	%l1,$tmp0,$tmp0
+	or	%l3,$tmp1,$tmp1
+	add	$tmp0,$E,$E
+	add	$tmp1,$F,$F
+	$ST	$E,[$ctx+`4*$SZ`]
+	sllx	%l4,32,$tmp2
+	$ST	$F,[$ctx+`5*$SZ`]
+	sllx	%l6,32,$T1
+	or	%l5,$tmp2,$tmp2
+	or	%l7,$T1,$T1
+	add	$tmp2,$G,$G
+	$ST	$G,[$ctx+`6*$SZ`]
+	add	$T1,$H,$H
+	$ST	$H,[$ctx+`7*$SZ`]
+___
+$code.=<<___;
+	add	$inp,`16*$SZ`,$inp		! advance inp
+	cmp	$inp,$len
+	bne	`$bits==64?"%xcc":"%icc"`,.Lloop
+	sub	$Ktbl,`($rounds-16)*$SZ`,$Ktbl	! rewind Ktbl
+
+	ret
+	restore
+.type	sha${label}_block_data_order,#function
+.size	sha${label}_block_data_order,(.-sha${label}_block_data_order)
+.asciz	"SHA${label} block transform for SPARCv9, CRYPTOGAMS by <appro\@openssl.org>"
+.align	4
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/drivers/builtin_openssl2/crypto/sha/asm/sha512-x86_64.pl b/drivers/builtin_openssl2/crypto/sha/asm/sha512-x86_64.pl
new file mode 100755
index 0000000000..8d51678557
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/asm/sha512-x86_64.pl
@@ -0,0 +1,451 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. Rights for redistribution and usage in source and binary
+# forms are granted according to the OpenSSL license.
+# ====================================================================
+#
+# sha256/512_block procedure for x86_64.
+#
+# 40% improvement over compiler-generated code on Opteron. On EM64T
+# sha256 was observed to run >80% faster and sha512 - >40%. No magical
+# tricks, just straight implementation... I really wonder why gcc
+# [being armed with inline assembler] fails to generate as fast code.
+# The only thing which is cool about this module is that it's very
+# same instruction sequence used for both SHA-256 and SHA-512. In
+# former case the instructions operate on 32-bit operands, while in
+# latter - on 64-bit ones. All I had to do is to get one flavor right,
+# the other one passed the test right away:-)
+#
+# sha256_block runs in ~1005 cycles on Opteron, which gives you
+# asymptotic performance of 64*1000/1005=63.7MBps times CPU clock
+# frequency in GHz. sha512_block runs in ~1275 cycles, which results
+# in 128*1000/1275=100MBps per GHz. Is there room for improvement?
+# Well, if you compare it to IA-64 implementation, which maintains
+# X[16] in register bank[!], tends to 4 instructions per CPU clock
+# cycle and runs in 1003 cycles, 1275 is very good result for 3-way
+# issue Opteron pipeline and X[16] maintained in memory. So that *if*
+# there is a way to improve it, *then* the only way would be to try to
+# offload X[16] updates to SSE unit, but that would require "deeper"
+# loop unroll, which in turn would naturally cause size blow-up, not
+# to mention increased complexity! And once again, only *if* it's
+# actually possible to noticeably improve overall ILP, instruction
+# 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],
+# [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
+# apparently are not atomic instructions, but implemented in microcode.
+
+$flavour = shift;
+$output  = shift;
+if ($flavour =~ /\./) { $output = $flavour; undef $flavour; }
+
+$win64=0; $win64=1 if ($flavour =~ /[nm]asm|mingw64/ || $output =~ /\.asm$/);
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
+die "can't locate x86_64-xlate.pl";
+
+open OUT,"| \"$^X\" $xlate $flavour $output";
+*STDOUT=*OUT;
+
+if ($output =~ /512/) {
+	$func="sha512_block_data_order";
+	$TABLE="K512";
+	$SZ=8;
+	@ROT=($A,$B,$C,$D,$E,$F,$G,$H)=("%rax","%rbx","%rcx","%rdx",
+					"%r8", "%r9", "%r10","%r11");
+	($T1,$a0,$a1,$a2)=("%r12","%r13","%r14","%r15");
+	@Sigma0=(28,34,39);
+	@Sigma1=(14,18,41);
+	@sigma0=(1,  8, 7);
+	@sigma1=(19,61, 6);
+	$rounds=80;
+} else {
+	$func="sha256_block_data_order";
+	$TABLE="K256";
+	$SZ=4;
+	@ROT=($A,$B,$C,$D,$E,$F,$G,$H)=("%eax","%ebx","%ecx","%edx",
+					"%r8d","%r9d","%r10d","%r11d");
+	($T1,$a0,$a1,$a2)=("%r12d","%r13d","%r14d","%r15d");
+	@Sigma0=( 2,13,22);
+	@Sigma1=( 6,11,25);
+	@sigma0=( 7,18, 3);
+	@sigma1=(17,19,10);
+	$rounds=64;
+}
+
+$ctx="%rdi";	# 1st arg
+$round="%rdi";	# zaps $ctx
+$inp="%rsi";	# 2nd arg
+$Tbl="%rbp";
+
+$_ctx="16*$SZ+0*8(%rsp)";
+$_inp="16*$SZ+1*8(%rsp)";
+$_end="16*$SZ+2*8(%rsp)";
+$_rsp="16*$SZ+3*8(%rsp)";
+$framesz="16*$SZ+4*8";
+
+
+sub ROUND_00_15()
+{ my ($i,$a,$b,$c,$d,$e,$f,$g,$h) = @_;
+
+$code.=<<___;
+	ror	\$`$Sigma1[2]-$Sigma1[1]`,$a0
+	mov	$f,$a2
+	mov	$T1,`$SZ*($i&0xf)`(%rsp)
+
+	ror	\$`$Sigma0[2]-$Sigma0[1]`,$a1
+	xor	$e,$a0
+	xor	$g,$a2			# f^g
+
+	ror	\$`$Sigma1[1]-$Sigma1[0]`,$a0
+	add	$h,$T1			# T1+=h
+	xor	$a,$a1
+
+	add	($Tbl,$round,$SZ),$T1	# T1+=K[round]
+	and	$e,$a2			# (f^g)&e
+	mov	$b,$h
+
+	ror	\$`$Sigma0[1]-$Sigma0[0]`,$a1
+	xor	$e,$a0
+	xor	$g,$a2			# Ch(e,f,g)=((f^g)&e)^g
+
+	xor	$c,$h			# b^c
+	xor	$a,$a1
+	add	$a2,$T1			# T1+=Ch(e,f,g)
+	mov	$b,$a2
+
+	ror	\$$Sigma1[0],$a0	# Sigma1(e)
+	and	$a,$h			# h=(b^c)&a
+	and	$c,$a2			# b&c
+
+	ror	\$$Sigma0[0],$a1	# Sigma0(a)
+	add	$a0,$T1			# T1+=Sigma1(e)
+	add	$a2,$h			# h+=b&c (completes +=Maj(a,b,c)
+
+	add	$T1,$d			# d+=T1
+	add	$T1,$h			# h+=T1
+	lea	1($round),$round	# round++
+	add	$a1,$h			# h+=Sigma0(a)
+
+___
+}
+
+sub ROUND_16_XX()
+{ my ($i,$a,$b,$c,$d,$e,$f,$g,$h) = @_;
+
+$code.=<<___;
+	mov	`$SZ*(($i+1)&0xf)`(%rsp),$a0
+	mov	`$SZ*(($i+14)&0xf)`(%rsp),$a1
+	mov	$a0,$T1
+	mov	$a1,$a2
+
+	ror	\$`$sigma0[1]-$sigma0[0]`,$T1
+	xor	$a0,$T1
+	shr	\$$sigma0[2],$a0
+
+	ror	\$$sigma0[0],$T1
+	xor	$T1,$a0			# sigma0(X[(i+1)&0xf])
+	mov	`$SZ*(($i+9)&0xf)`(%rsp),$T1
+
+	ror	\$`$sigma1[1]-$sigma1[0]`,$a2
+	xor	$a1,$a2
+	shr	\$$sigma1[2],$a1
+
+	ror	\$$sigma1[0],$a2
+	add	$a0,$T1
+	xor	$a2,$a1			# sigma1(X[(i+14)&0xf])
+
+	add	`$SZ*($i&0xf)`(%rsp),$T1
+	mov	$e,$a0
+	add	$a1,$T1
+	mov	$a,$a1
+___
+	&ROUND_00_15(@_);
+}
+
+$code=<<___;
+.text
+
+.globl	$func
+.type	$func,\@function,4
+.align	16
+$func:
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	mov	%rsp,%r11		# copy %rsp
+	shl	\$4,%rdx		# num*16
+	sub	\$$framesz,%rsp
+	lea	($inp,%rdx,$SZ),%rdx	# inp+num*16*$SZ
+	and	\$-64,%rsp		# align stack frame
+	mov	$ctx,$_ctx		# save ctx, 1st arg
+	mov	$inp,$_inp		# save inp, 2nd arh
+	mov	%rdx,$_end		# save end pointer, "3rd" arg
+	mov	%r11,$_rsp		# save copy of %rsp
+.Lprologue:
+
+	lea	$TABLE(%rip),$Tbl
+
+	mov	$SZ*0($ctx),$A
+	mov	$SZ*1($ctx),$B
+	mov	$SZ*2($ctx),$C
+	mov	$SZ*3($ctx),$D
+	mov	$SZ*4($ctx),$E
+	mov	$SZ*5($ctx),$F
+	mov	$SZ*6($ctx),$G
+	mov	$SZ*7($ctx),$H
+	jmp	.Lloop
+
+.align	16
+.Lloop:
+	xor	$round,$round
+___
+	for($i=0;$i<16;$i++) {
+		$code.="	mov	$SZ*$i($inp),$T1\n";
+		$code.="	mov	@ROT[4],$a0\n";
+		$code.="	mov	@ROT[0],$a1\n";
+		$code.="	bswap	$T1\n";
+		&ROUND_00_15($i,@ROT);
+		unshift(@ROT,pop(@ROT));
+	}
+$code.=<<___;
+	jmp	.Lrounds_16_xx
+.align	16
+.Lrounds_16_xx:
+___
+	for(;$i<32;$i++) {
+		&ROUND_16_XX($i,@ROT);
+		unshift(@ROT,pop(@ROT));
+	}
+
+$code.=<<___;
+	cmp	\$$rounds,$round
+	jb	.Lrounds_16_xx
+
+	mov	$_ctx,$ctx
+	lea	16*$SZ($inp),$inp
+
+	add	$SZ*0($ctx),$A
+	add	$SZ*1($ctx),$B
+	add	$SZ*2($ctx),$C
+	add	$SZ*3($ctx),$D
+	add	$SZ*4($ctx),$E
+	add	$SZ*5($ctx),$F
+	add	$SZ*6($ctx),$G
+	add	$SZ*7($ctx),$H
+
+	cmp	$_end,$inp
+
+	mov	$A,$SZ*0($ctx)
+	mov	$B,$SZ*1($ctx)
+	mov	$C,$SZ*2($ctx)
+	mov	$D,$SZ*3($ctx)
+	mov	$E,$SZ*4($ctx)
+	mov	$F,$SZ*5($ctx)
+	mov	$G,$SZ*6($ctx)
+	mov	$H,$SZ*7($ctx)
+	jb	.Lloop
+
+	mov	$_rsp,%rsi
+	mov	(%rsi),%r15
+	mov	8(%rsi),%r14
+	mov	16(%rsi),%r13
+	mov	24(%rsi),%r12
+	mov	32(%rsi),%rbp
+	mov	40(%rsi),%rbx
+	lea	48(%rsi),%rsp
+.Lepilogue:
+	ret
+.size	$func,.-$func
+___
+
+if ($SZ==4) {
+$code.=<<___;
+.align	64
+.type	$TABLE,\@object
+$TABLE:
+	.long	0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5
+	.long	0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5
+	.long	0xd807aa98,0x12835b01,0x243185be,0x550c7dc3
+	.long	0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174
+	.long	0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc
+	.long	0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da
+	.long	0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7
+	.long	0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967
+	.long	0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13
+	.long	0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85
+	.long	0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3
+	.long	0xd192e819,0xd6990624,0xf40e3585,0x106aa070
+	.long	0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5
+	.long	0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3
+	.long	0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208
+	.long	0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
+___
+} else {
+$code.=<<___;
+.align	64
+.type	$TABLE,\@object
+$TABLE:
+	.quad	0x428a2f98d728ae22,0x7137449123ef65cd
+	.quad	0xb5c0fbcfec4d3b2f,0xe9b5dba58189dbbc
+	.quad	0x3956c25bf348b538,0x59f111f1b605d019
+	.quad	0x923f82a4af194f9b,0xab1c5ed5da6d8118
+	.quad	0xd807aa98a3030242,0x12835b0145706fbe
+	.quad	0x243185be4ee4b28c,0x550c7dc3d5ffb4e2
+	.quad	0x72be5d74f27b896f,0x80deb1fe3b1696b1
+	.quad	0x9bdc06a725c71235,0xc19bf174cf692694
+	.quad	0xe49b69c19ef14ad2,0xefbe4786384f25e3
+	.quad	0x0fc19dc68b8cd5b5,0x240ca1cc77ac9c65
+	.quad	0x2de92c6f592b0275,0x4a7484aa6ea6e483
+	.quad	0x5cb0a9dcbd41fbd4,0x76f988da831153b5
+	.quad	0x983e5152ee66dfab,0xa831c66d2db43210
+	.quad	0xb00327c898fb213f,0xbf597fc7beef0ee4
+	.quad	0xc6e00bf33da88fc2,0xd5a79147930aa725
+	.quad	0x06ca6351e003826f,0x142929670a0e6e70
+	.quad	0x27b70a8546d22ffc,0x2e1b21385c26c926
+	.quad	0x4d2c6dfc5ac42aed,0x53380d139d95b3df
+	.quad	0x650a73548baf63de,0x766a0abb3c77b2a8
+	.quad	0x81c2c92e47edaee6,0x92722c851482353b
+	.quad	0xa2bfe8a14cf10364,0xa81a664bbc423001
+	.quad	0xc24b8b70d0f89791,0xc76c51a30654be30
+	.quad	0xd192e819d6ef5218,0xd69906245565a910
+	.quad	0xf40e35855771202a,0x106aa07032bbd1b8
+	.quad	0x19a4c116b8d2d0c8,0x1e376c085141ab53
+	.quad	0x2748774cdf8eeb99,0x34b0bcb5e19b48a8
+	.quad	0x391c0cb3c5c95a63,0x4ed8aa4ae3418acb
+	.quad	0x5b9cca4f7763e373,0x682e6ff3d6b2b8a3
+	.quad	0x748f82ee5defb2fc,0x78a5636f43172f60
+	.quad	0x84c87814a1f0ab72,0x8cc702081a6439ec
+	.quad	0x90befffa23631e28,0xa4506cebde82bde9
+	.quad	0xbef9a3f7b2c67915,0xc67178f2e372532b
+	.quad	0xca273eceea26619c,0xd186b8c721c0c207
+	.quad	0xeada7dd6cde0eb1e,0xf57d4f7fee6ed178
+	.quad	0x06f067aa72176fba,0x0a637dc5a2c898a6
+	.quad	0x113f9804bef90dae,0x1b710b35131c471b
+	.quad	0x28db77f523047d84,0x32caab7b40c72493
+	.quad	0x3c9ebe0a15c9bebc,0x431d67c49c100d4c
+	.quad	0x4cc5d4becb3e42b6,0x597f299cfc657e2a
+	.quad	0x5fcb6fab3ad6faec,0x6c44198c4a475817
+___
+}
+
+# EXCEPTION_DISPOSITION handler (EXCEPTION_RECORD *rec,ULONG64 frame,
+#		CONTEXT *context,DISPATCHER_CONTEXT *disp)
+if ($win64) {
+$rec="%rcx";
+$frame="%rdx";
+$context="%r8";
+$disp="%r9";
+
+$code.=<<___;
+.extern	__imp_RtlVirtualUnwind
+.type	se_handler,\@abi-omnipotent
+.align	16
+se_handler:
+	push	%rsi
+	push	%rdi
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	pushfq
+	sub	\$64,%rsp
+
+	mov	120($context),%rax	# pull context->Rax
+	mov	248($context),%rbx	# pull context->Rip
+
+	lea	.Lprologue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip<.Lprologue
+	jb	.Lin_prologue
+
+	mov	152($context),%rax	# pull context->Rsp
+
+	lea	.Lepilogue(%rip),%r10
+	cmp	%r10,%rbx		# context->Rip>=.Lepilogue
+	jae	.Lin_prologue
+
+	mov	16*$SZ+3*8(%rax),%rax	# pull $_rsp
+	lea	48(%rax),%rax
+
+	mov	-8(%rax),%rbx
+	mov	-16(%rax),%rbp
+	mov	-24(%rax),%r12
+	mov	-32(%rax),%r13
+	mov	-40(%rax),%r14
+	mov	-48(%rax),%r15
+	mov	%rbx,144($context)	# restore context->Rbx
+	mov	%rbp,160($context)	# restore context->Rbp
+	mov	%r12,216($context)	# restore context->R12
+	mov	%r13,224($context)	# restore context->R13
+	mov	%r14,232($context)	# restore context->R14
+	mov	%r15,240($context)	# restore context->R15
+
+.Lin_prologue:
+	mov	8(%rax),%rdi
+	mov	16(%rax),%rsi
+	mov	%rax,152($context)	# restore context->Rsp
+	mov	%rsi,168($context)	# restore context->Rsi
+	mov	%rdi,176($context)	# restore context->Rdi
+
+	mov	40($disp),%rdi		# disp->ContextRecord
+	mov	$context,%rsi		# context
+	mov	\$154,%ecx		# sizeof(CONTEXT)
+	.long	0xa548f3fc		# cld; rep movsq
+
+	mov	$disp,%rsi
+	xor	%rcx,%rcx		# arg1, UNW_FLAG_NHANDLER
+	mov	8(%rsi),%rdx		# arg2, disp->ImageBase
+	mov	0(%rsi),%r8		# arg3, disp->ControlPc
+	mov	16(%rsi),%r9		# arg4, disp->FunctionEntry
+	mov	40(%rsi),%r10		# disp->ContextRecord
+	lea	56(%rsi),%r11		# &disp->HandlerData
+	lea	24(%rsi),%r12		# &disp->EstablisherFrame
+	mov	%r10,32(%rsp)		# arg5
+	mov	%r11,40(%rsp)		# arg6
+	mov	%r12,48(%rsp)		# arg7
+	mov	%rcx,56(%rsp)		# arg8, (NULL)
+	call	*__imp_RtlVirtualUnwind(%rip)
+
+	mov	\$1,%eax		# ExceptionContinueSearch
+	add	\$64,%rsp
+	popfq
+	pop	%r15
+	pop	%r14
+	pop	%r13
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+	pop	%rdi
+	pop	%rsi
+	ret
+.size	se_handler,.-se_handler
+
+.section	.pdata
+.align	4
+	.rva	.LSEH_begin_$func
+	.rva	.LSEH_end_$func
+	.rva	.LSEH_info_$func
+
+.section	.xdata
+.align	8
+.LSEH_info_$func:
+	.byte	9,0,0,0
+	.rva	se_handler
+___
+}
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;
diff --git a/drivers/builtin_openssl2/crypto/sha/sha.c b/drivers/builtin_openssl2/crypto/sha/sha.c
new file mode 100644
index 0000000000..42126551d1
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/sha.c
@@ -0,0 +1,124 @@
+/* crypto/sha/sha.c */
+/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
+ * All rights reserved.
+ *
+ * This package is an SSL implementation written
+ * by Eric Young (eay@cryptsoft.com).
+ * The implementation was written so as to conform with Netscapes SSL.
+ * 
+ * This library is free for commercial and non-commercial use as long as
+ * the following conditions are aheared to.  The following conditions
+ * apply to all code found in this distribution, be it the RC4, RSA,
+ * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
+ * included with this distribution is covered by the same copyright terms
+ * except that the holder is Tim Hudson (tjh@cryptsoft.com).
+ * 
+ * Copyright remains Eric Young's, and as such any Copyright notices in
+ * the code are not to be removed.
+ * If this package is used in a product, Eric Young should be given attribution
+ * as the author of the parts of the library used.
+ * This can be in the form of a textual message at program startup or
+ * in documentation (online or textual) provided with the package.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. All advertising materials mentioning features or use of this software
+ *    must display the following acknowledgement:
+ *    "This product includes cryptographic software written by
+ *     Eric Young (eay@cryptsoft.com)"
+ *    The word 'cryptographic' can be left out if the rouines from the library
+ *    being used are not cryptographic related :-).
+ * 4. If you include any Windows specific code (or a derivative thereof) from 
+ *    the apps directory (application code) you must include an acknowledgement:
+ *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
+ * 
+ * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ * 
+ * The licence and distribution terms for any publically available version or
+ * derivative of this code cannot be changed.  i.e. this code cannot simply be
+ * copied and put under another distribution licence
+ * [including the GNU Public Licence.]
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <openssl/sha.h>
+
+#define BUFSIZE	1024*16
+
+void do_fp(FILE *f);
+void pt(unsigned char *md);
+int read(int, void *, unsigned int);
+int main(int argc, char **argv)
+	{
+	int i,err=0;
+	FILE *IN;
+
+	if (argc == 1)
+		{
+		do_fp(stdin);
+		}
+	else
+		{
+		for (i=1; i<argc; i++)
+			{
+			IN=fopen(argv[i],"r");
+			if (IN == NULL)
+				{
+				perror(argv[i]);
+				err++;
+				continue;
+				}
+			printf("SHA(%s)= ",argv[i]);
+			do_fp(IN);
+			fclose(IN);
+			}
+		}
+	exit(err);
+	}
+
+void do_fp(FILE *f)
+	{
+	SHA_CTX c;
+	unsigned char md[SHA_DIGEST_LENGTH];
+	int fd;
+	int i;
+	unsigned char buf[BUFSIZE];
+
+	fd=fileno(f);
+	SHA_Init(&c);
+	for (;;)
+		{
+		i=read(fd,buf,BUFSIZE);
+		if (i <= 0) break;
+		SHA_Update(&c,buf,(unsigned long)i);
+		}
+	SHA_Final(&(md[0]),&c);
+	pt(md);
+	}
+
+void pt(unsigned char *md)
+	{
+	int i;
+
+	for (i=0; i<SHA_DIGEST_LENGTH; i++)
+		printf("%02x",md[i]);
+	printf("\n");
+	}
+
diff --git a/drivers/builtin_openssl2/crypto/sha/sha1.c b/drivers/builtin_openssl2/crypto/sha/sha1.c
new file mode 100644
index 0000000000..d350c88ee4
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/sha1.c
@@ -0,0 +1,127 @@
+/* crypto/sha/sha1.c */
+/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
+ * All rights reserved.
+ *
+ * This package is an SSL implementation written
+ * by Eric Young (eay@cryptsoft.com).
+ * The implementation was written so as to conform with Netscapes SSL.
+ * 
+ * This library is free for commercial and non-commercial use as long as
+ * the following conditions are aheared to.  The following conditions
+ * apply to all code found in this distribution, be it the RC4, RSA,
+ * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
+ * included with this distribution is covered by the same copyright terms
+ * except that the holder is Tim Hudson (tjh@cryptsoft.com).
+ * 
+ * Copyright remains Eric Young's, and as such any Copyright notices in
+ * the code are not to be removed.
+ * If this package is used in a product, Eric Young should be given attribution
+ * as the author of the parts of the library used.
+ * This can be in the form of a textual message at program startup or
+ * in documentation (online or textual) provided with the package.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. All advertising materials mentioning features or use of this software
+ *    must display the following acknowledgement:
+ *    "This product includes cryptographic software written by
+ *     Eric Young (eay@cryptsoft.com)"
+ *    The word 'cryptographic' can be left out if the rouines from the library
+ *    being used are not cryptographic related :-).
+ * 4. If you include any Windows specific code (or a derivative thereof) from 
+ *    the apps directory (application code) you must include an acknowledgement:
+ *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
+ * 
+ * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ * 
+ * The licence and distribution terms for any publically available version or
+ * derivative of this code cannot be changed.  i.e. this code cannot simply be
+ * copied and put under another distribution licence
+ * [including the GNU Public Licence.]
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <openssl/sha.h>
+
+#define BUFSIZE	1024*16
+
+void do_fp(FILE *f);
+void pt(unsigned char *md);
+#ifndef _OSD_POSIX
+int read(int, void *, unsigned int);
+#endif
+
+int main(int argc, char **argv)
+	{
+	int i,err=0;
+	FILE *IN;
+
+	if (argc == 1)
+		{
+		do_fp(stdin);
+		}
+	else
+		{
+		for (i=1; i<argc; i++)
+			{
+			IN=fopen(argv[i],"r");
+			if (IN == NULL)
+				{
+				perror(argv[i]);
+				err++;
+				continue;
+				}
+			printf("SHA1(%s)= ",argv[i]);
+			do_fp(IN);
+			fclose(IN);
+			}
+		}
+	exit(err);
+	}
+
+void do_fp(FILE *f)
+	{
+	SHA_CTX c;
+	unsigned char md[SHA_DIGEST_LENGTH];
+	int fd;
+	int i;
+	unsigned char buf[BUFSIZE];
+
+	fd=fileno(f);
+	SHA1_Init(&c);
+	for (;;)
+		{
+		i=read(fd,buf,BUFSIZE);
+		if (i <= 0) break;
+		SHA1_Update(&c,buf,(unsigned long)i);
+		}
+	SHA1_Final(&(md[0]),&c);
+	pt(md);
+	}
+
+void pt(unsigned char *md)
+	{
+	int i;
+
+	for (i=0; i<SHA_DIGEST_LENGTH; i++)
+		printf("%02x",md[i]);
+	printf("\n");
+	}
+
diff --git a/drivers/builtin_openssl2/crypto/sha/sha1_one.c b/drivers/builtin_openssl2/crypto/sha/sha1_one.c
new file mode 100644
index 0000000000..c56ec94020
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/sha1_one.c
@@ -0,0 +1,78 @@
+/* crypto/sha/sha1_one.c */
+/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
+ * All rights reserved.
+ *
+ * This package is an SSL implementation written
+ * by Eric Young (eay@cryptsoft.com).
+ * The implementation was written so as to conform with Netscapes SSL.
+ * 
+ * This library is free for commercial and non-commercial use as long as
+ * the following conditions are aheared to.  The following conditions
+ * apply to all code found in this distribution, be it the RC4, RSA,
+ * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
+ * included with this distribution is covered by the same copyright terms
+ * except that the holder is Tim Hudson (tjh@cryptsoft.com).
+ * 
+ * Copyright remains Eric Young's, and as such any Copyright notices in
+ * the code are not to be removed.
+ * If this package is used in a product, Eric Young should be given attribution
+ * as the author of the parts of the library used.
+ * This can be in the form of a textual message at program startup or
+ * in documentation (online or textual) provided with the package.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. All advertising materials mentioning features or use of this software
+ *    must display the following acknowledgement:
+ *    "This product includes cryptographic software written by
+ *     Eric Young (eay@cryptsoft.com)"
+ *    The word 'cryptographic' can be left out if the rouines from the library
+ *    being used are not cryptographic related :-).
+ * 4. If you include any Windows specific code (or a derivative thereof) from 
+ *    the apps directory (application code) you must include an acknowledgement:
+ *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
+ * 
+ * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ * 
+ * The licence and distribution terms for any publically available version or
+ * derivative of this code cannot be changed.  i.e. this code cannot simply be
+ * copied and put under another distribution licence
+ * [including the GNU Public Licence.]
+ */
+
+#include <stdio.h>
+#include <string.h>
+#include <openssl/crypto.h>
+#include <openssl/sha.h>
+
+#ifndef OPENSSL_NO_SHA1
+unsigned char *SHA1(const unsigned char *d, size_t n, unsigned char *md)
+	{
+	SHA_CTX c;
+	static unsigned char m[SHA_DIGEST_LENGTH];
+
+	if (md == NULL) md=m;
+	if (!SHA1_Init(&c))
+		return NULL;
+	SHA1_Update(&c,d,n);
+	SHA1_Final(md,&c);
+	OPENSSL_cleanse(&c,sizeof(c));
+	return(md);
+	}
+#endif
diff --git a/drivers/builtin_openssl2/crypto/sha/sha1dgst.c b/drivers/builtin_openssl2/crypto/sha/sha1dgst.c
new file mode 100644
index 0000000000..a98690225f
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/sha1dgst.c
@@ -0,0 +1,75 @@
+/* crypto/sha/sha1dgst.c */
+/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
+ * All rights reserved.
+ *
+ * This package is an SSL implementation written
+ * by Eric Young (eay@cryptsoft.com).
+ * The implementation was written so as to conform with Netscapes SSL.
+ * 
+ * This library is free for commercial and non-commercial use as long as
+ * the following conditions are aheared to.  The following conditions
+ * apply to all code found in this distribution, be it the RC4, RSA,
+ * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
+ * included with this distribution is covered by the same copyright terms
+ * except that the holder is Tim Hudson (tjh@cryptsoft.com).
+ * 
+ * Copyright remains Eric Young's, and as such any Copyright notices in
+ * the code are not to be removed.
+ * If this package is used in a product, Eric Young should be given attribution
+ * as the author of the parts of the library used.
+ * This can be in the form of a textual message at program startup or
+ * in documentation (online or textual) provided with the package.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. All advertising materials mentioning features or use of this software
+ *    must display the following acknowledgement:
+ *    "This product includes cryptographic software written by
+ *     Eric Young (eay@cryptsoft.com)"
+ *    The word 'cryptographic' can be left out if the rouines from the library
+ *    being used are not cryptographic related :-).
+ * 4. If you include any Windows specific code (or a derivative thereof) from 
+ *    the apps directory (application code) you must include an acknowledgement:
+ *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
+ * 
+ * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ * 
+ * The licence and distribution terms for any publically available version or
+ * derivative of this code cannot be changed.  i.e. this code cannot simply be
+ * copied and put under another distribution licence
+ * [including the GNU Public Licence.]
+ */
+
+#include <openssl/crypto.h>
+#include <openssl/opensslconf.h>
+#if !defined(OPENSSL_NO_SHA1) && !defined(OPENSSL_NO_SHA)
+
+#undef  SHA_0
+#define SHA_1
+
+#include <openssl/opensslv.h>
+
+const char SHA1_version[]="SHA1" OPENSSL_VERSION_PTEXT;
+
+/* The implementation is in ../md32_common.h */
+
+#include "sha_locl.h"
+
+#endif
+
diff --git a/drivers/builtin_openssl2/crypto/sha/sha1test.c b/drivers/builtin_openssl2/crypto/sha/sha1test.c
new file mode 100644
index 0000000000..6feb3964c7
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/sha1test.c
@@ -0,0 +1,178 @@
+/* crypto/sha/sha1test.c */
+/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
+ * All rights reserved.
+ *
+ * This package is an SSL implementation written
+ * by Eric Young (eay@cryptsoft.com).
+ * The implementation was written so as to conform with Netscapes SSL.
+ * 
+ * This library is free for commercial and non-commercial use as long as
+ * the following conditions are aheared to.  The following conditions
+ * apply to all code found in this distribution, be it the RC4, RSA,
+ * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
+ * included with this distribution is covered by the same copyright terms
+ * except that the holder is Tim Hudson (tjh@cryptsoft.com).
+ * 
+ * Copyright remains Eric Young's, and as such any Copyright notices in
+ * the code are not to be removed.
+ * If this package is used in a product, Eric Young should be given attribution
+ * as the author of the parts of the library used.
+ * This can be in the form of a textual message at program startup or
+ * in documentation (online or textual) provided with the package.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. All advertising materials mentioning features or use of this software
+ *    must display the following acknowledgement:
+ *    "This product includes cryptographic software written by
+ *     Eric Young (eay@cryptsoft.com)"
+ *    The word 'cryptographic' can be left out if the rouines from the library
+ *    being used are not cryptographic related :-).
+ * 4. If you include any Windows specific code (or a derivative thereof) from 
+ *    the apps directory (application code) you must include an acknowledgement:
+ *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
+ * 
+ * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ * 
+ * The licence and distribution terms for any publically available version or
+ * derivative of this code cannot be changed.  i.e. this code cannot simply be
+ * copied and put under another distribution licence
+ * [including the GNU Public Licence.]
+ */
+
+#include <stdio.h>
+#include <string.h>
+#include <stdlib.h>
+
+#include "../e_os.h"
+
+#ifdef OPENSSL_NO_SHA
+int main(int argc, char *argv[])
+{
+    printf("No SHA support\n");
+    return(0);
+}
+#else
+#include <openssl/evp.h>
+#include <openssl/sha.h>
+
+#ifdef CHARSET_EBCDIC
+#include <openssl/ebcdic.h>
+#endif
+
+#undef SHA_0 /* FIPS 180 */
+#define  SHA_1 /* FIPS 180-1 */
+
+static char *test[]={
+	"abc",
+	"abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq",
+	NULL,
+	};
+
+#ifdef SHA_0
+static char *ret[]={
+	"0164b8a914cd2a5e74c4f7ff082c4d97f1edf880",
+	"d2516ee1acfa5baf33dfc1c471e438449ef134c8",
+	};
+static char *bigret=
+	"3232affa48628a26653b5aaa44541fd90d690603";
+#endif
+#ifdef SHA_1
+static char *ret[]={
+	"a9993e364706816aba3e25717850c26c9cd0d89d",
+	"84983e441c3bd26ebaae4aa1f95129e5e54670f1",
+	};
+static char *bigret=
+	"34aa973cd4c4daa4f61eeb2bdbad27316534016f";
+#endif
+
+static char *pt(unsigned char *md);
+int main(int argc, char *argv[])
+	{
+	int i,err=0;
+	char **P,**R;
+	static unsigned char buf[1000];
+	char *p,*r;
+	EVP_MD_CTX c;
+	unsigned char md[SHA_DIGEST_LENGTH];
+
+#ifdef CHARSET_EBCDIC
+	ebcdic2ascii(test[0], test[0], strlen(test[0]));
+	ebcdic2ascii(test[1], test[1], strlen(test[1]));
+#endif
+
+	EVP_MD_CTX_init(&c);
+	P=test;
+	R=ret;
+	i=1;
+	while (*P != NULL)
+		{
+		EVP_Digest(*P,strlen((char *)*P),md,NULL,EVP_sha1(), NULL);
+		p=pt(md);
+		if (strcmp(p,(char *)*R) != 0)
+			{
+			printf("error calculating SHA1 on '%s'\n",*P);
+			printf("got %s instead of %s\n",p,*R);
+			err++;
+			}
+		else
+			printf("test %d ok\n",i);
+		i++;
+		R++;
+		P++;
+		}
+
+	memset(buf,'a',1000);
+#ifdef CHARSET_EBCDIC
+	ebcdic2ascii(buf, buf, 1000);
+#endif /*CHARSET_EBCDIC*/
+	EVP_DigestInit_ex(&c,EVP_sha1(), NULL);
+	for (i=0; i<1000; i++)
+		EVP_DigestUpdate(&c,buf,1000);
+	EVP_DigestFinal_ex(&c,md,NULL);
+	p=pt(md);
+
+	r=bigret;
+	if (strcmp(p,r) != 0)
+		{
+		printf("error calculating SHA1 on 'a' * 1000\n");
+		printf("got %s instead of %s\n",p,r);
+		err++;
+		}
+	else
+		printf("test 3 ok\n");
+
+#ifdef OPENSSL_SYS_NETWARE
+    if (err) printf("ERROR: %d\n", err);
+#endif
+	EXIT(err);
+	EVP_MD_CTX_cleanup(&c);
+	return(0);
+	}
+
+static char *pt(unsigned char *md)
+	{
+	int i;
+	static char buf[80];
+
+	for (i=0; i<SHA_DIGEST_LENGTH; i++)
+		sprintf(&(buf[i*2]),"%02x",md[i]);
+	return(buf);
+	}
+#endif
diff --git a/drivers/builtin_openssl2/crypto/sha/sha256.c b/drivers/builtin_openssl2/crypto/sha/sha256.c
new file mode 100644
index 0000000000..4eae074849
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/sha256.c
@@ -0,0 +1,282 @@
+/* crypto/sha/sha256.c */
+/* ====================================================================
+ * Copyright (c) 2004 The OpenSSL Project.  All rights reserved
+ * according to the OpenSSL license [found in ../../LICENSE].
+ * ====================================================================
+ */
+#include <openssl/opensslconf.h>
+#if !defined(OPENSSL_NO_SHA) && !defined(OPENSSL_NO_SHA256)
+
+#include <stdlib.h>
+#include <string.h>
+
+#include <openssl/crypto.h>
+#include <openssl/sha.h>
+#include <openssl/opensslv.h>
+
+const char SHA256_version[]="SHA-256" OPENSSL_VERSION_PTEXT;
+
+fips_md_init_ctx(SHA224, SHA256)
+	{
+	memset (c,0,sizeof(*c));
+	c->h[0]=0xc1059ed8UL;	c->h[1]=0x367cd507UL;
+	c->h[2]=0x3070dd17UL;	c->h[3]=0xf70e5939UL;
+	c->h[4]=0xffc00b31UL;	c->h[5]=0x68581511UL;
+	c->h[6]=0x64f98fa7UL;	c->h[7]=0xbefa4fa4UL;
+	c->md_len=SHA224_DIGEST_LENGTH;
+	return 1;
+	}
+
+fips_md_init(SHA256)
+	{
+	memset (c,0,sizeof(*c));
+	c->h[0]=0x6a09e667UL;	c->h[1]=0xbb67ae85UL;
+	c->h[2]=0x3c6ef372UL;	c->h[3]=0xa54ff53aUL;
+	c->h[4]=0x510e527fUL;	c->h[5]=0x9b05688cUL;
+	c->h[6]=0x1f83d9abUL;	c->h[7]=0x5be0cd19UL;
+	c->md_len=SHA256_DIGEST_LENGTH;
+	return 1;
+	}
+
+unsigned char *SHA224(const unsigned char *d, size_t n, unsigned char *md)
+	{
+	SHA256_CTX c;
+	static unsigned char m[SHA224_DIGEST_LENGTH];
+
+	if (md == NULL) md=m;
+	SHA224_Init(&c);
+	SHA256_Update(&c,d,n);
+	SHA256_Final(md,&c);
+	OPENSSL_cleanse(&c,sizeof(c));
+	return(md);
+	}
+
+unsigned char *SHA256(const unsigned char *d, size_t n, unsigned char *md)
+	{
+	SHA256_CTX c;
+	static unsigned char m[SHA256_DIGEST_LENGTH];
+
+	if (md == NULL) md=m;
+	SHA256_Init(&c);
+	SHA256_Update(&c,d,n);
+	SHA256_Final(md,&c);
+	OPENSSL_cleanse(&c,sizeof(c));
+	return(md);
+	}
+
+int SHA224_Update(SHA256_CTX *c, const void *data, size_t len)
+{   return SHA256_Update (c,data,len);   }
+int SHA224_Final (unsigned char *md, SHA256_CTX *c)
+{   return SHA256_Final (md,c);   }
+
+#define	DATA_ORDER_IS_BIG_ENDIAN
+
+#define	HASH_LONG		SHA_LONG
+#define	HASH_CTX		SHA256_CTX
+#define	HASH_CBLOCK		SHA_CBLOCK
+/*
+ * Note that FIPS180-2 discusses "Truncation of the Hash Function Output."
+ * default: case below covers for it. It's not clear however if it's
+ * permitted to truncate to amount of bytes not divisible by 4. I bet not,
+ * but if it is, then default: case shall be extended. For reference.
+ * Idea behind separate cases for pre-defined lenghts is to let the
+ * compiler decide if it's appropriate to unroll small loops.
+ */
+#define	HASH_MAKE_STRING(c,s)	do {	\
+	unsigned long ll;		\
+	unsigned int  nn;		\
+	switch ((c)->md_len)		\
+	{   case SHA224_DIGEST_LENGTH:	\
+		for (nn=0;nn<SHA224_DIGEST_LENGTH/4;nn++)	\
+		{   ll=(c)->h[nn]; (void)HOST_l2c(ll,(s));   }	\
+		break;			\
+	    case SHA256_DIGEST_LENGTH:	\
+		for (nn=0;nn<SHA256_DIGEST_LENGTH/4;nn++)	\
+		{   ll=(c)->h[nn]; (void)HOST_l2c(ll,(s));   }	\
+		break;			\
+	    default:			\
+		if ((c)->md_len > SHA256_DIGEST_LENGTH)	\
+		    return 0;				\
+		for (nn=0;nn<(c)->md_len/4;nn++)		\
+		{   ll=(c)->h[nn]; (void)HOST_l2c(ll,(s));   }	\
+		break;			\
+	}				\
+	} while (0)
+
+#define	HASH_UPDATE		SHA256_Update
+#define	HASH_TRANSFORM		SHA256_Transform
+#define	HASH_FINAL		SHA256_Final
+#define	HASH_BLOCK_DATA_ORDER	sha256_block_data_order
+#ifndef SHA256_ASM
+static
+#endif
+void sha256_block_data_order (SHA256_CTX *ctx, const void *in, size_t num);
+
+#include "md32_common.h"
+
+#ifndef SHA256_ASM
+static const SHA_LONG K256[64] = {
+	0x428a2f98UL,0x71374491UL,0xb5c0fbcfUL,0xe9b5dba5UL,
+	0x3956c25bUL,0x59f111f1UL,0x923f82a4UL,0xab1c5ed5UL,
+	0xd807aa98UL,0x12835b01UL,0x243185beUL,0x550c7dc3UL,
+	0x72be5d74UL,0x80deb1feUL,0x9bdc06a7UL,0xc19bf174UL,
+	0xe49b69c1UL,0xefbe4786UL,0x0fc19dc6UL,0x240ca1ccUL,
+	0x2de92c6fUL,0x4a7484aaUL,0x5cb0a9dcUL,0x76f988daUL,
+	0x983e5152UL,0xa831c66dUL,0xb00327c8UL,0xbf597fc7UL,
+	0xc6e00bf3UL,0xd5a79147UL,0x06ca6351UL,0x14292967UL,
+	0x27b70a85UL,0x2e1b2138UL,0x4d2c6dfcUL,0x53380d13UL,
+	0x650a7354UL,0x766a0abbUL,0x81c2c92eUL,0x92722c85UL,
+	0xa2bfe8a1UL,0xa81a664bUL,0xc24b8b70UL,0xc76c51a3UL,
+	0xd192e819UL,0xd6990624UL,0xf40e3585UL,0x106aa070UL,
+	0x19a4c116UL,0x1e376c08UL,0x2748774cUL,0x34b0bcb5UL,
+	0x391c0cb3UL,0x4ed8aa4aUL,0x5b9cca4fUL,0x682e6ff3UL,
+	0x748f82eeUL,0x78a5636fUL,0x84c87814UL,0x8cc70208UL,
+	0x90befffaUL,0xa4506cebUL,0xbef9a3f7UL,0xc67178f2UL };
+
+/*
+ * FIPS specification refers to right rotations, while our ROTATE macro
+ * is left one. This is why you might notice that rotation coefficients
+ * differ from those observed in FIPS document by 32-N...
+ */
+#define Sigma0(x)	(ROTATE((x),30) ^ ROTATE((x),19) ^ ROTATE((x),10))
+#define Sigma1(x)	(ROTATE((x),26) ^ ROTATE((x),21) ^ ROTATE((x),7))
+#define sigma0(x)	(ROTATE((x),25) ^ ROTATE((x),14) ^ ((x)>>3))
+#define sigma1(x)	(ROTATE((x),15) ^ ROTATE((x),13) ^ ((x)>>10))
+
+#define Ch(x,y,z)	(((x) & (y)) ^ ((~(x)) & (z)))
+#define Maj(x,y,z)	(((x) & (y)) ^ ((x) & (z)) ^ ((y) & (z)))
+
+#ifdef OPENSSL_SMALL_FOOTPRINT
+
+static void sha256_block_data_order (SHA256_CTX *ctx, const void *in, size_t num)
+	{
+	unsigned MD32_REG_T a,b,c,d,e,f,g,h,s0,s1,T1,T2;
+	SHA_LONG	X[16],l;
+	int i;
+	const unsigned char *data=in;
+
+			while (num--) {
+
+	a = ctx->h[0];	b = ctx->h[1];	c = ctx->h[2];	d = ctx->h[3];
+	e = ctx->h[4];	f = ctx->h[5];	g = ctx->h[6];	h = ctx->h[7];
+
+	for (i=0;i<16;i++)
+		{
+		HOST_c2l(data,l); T1 = X[i] = l;
+		T1 += h + Sigma1(e) + Ch(e,f,g) + K256[i];
+		T2 = Sigma0(a) + Maj(a,b,c);
+		h = g;	g = f;	f = e;	e = d + T1;
+		d = c;	c = b;	b = a;	a = T1 + T2;
+		}
+
+	for (;i<64;i++)
+		{
+		s0 = X[(i+1)&0x0f];	s0 = sigma0(s0);
+		s1 = X[(i+14)&0x0f];	s1 = sigma1(s1);
+
+		T1 = X[i&0xf] += s0 + s1 + X[(i+9)&0xf];
+		T1 += h + Sigma1(e) + Ch(e,f,g) + K256[i];
+		T2 = Sigma0(a) + Maj(a,b,c);
+		h = g;	g = f;	f = e;	e = d + T1;
+		d = c;	c = b;	b = a;	a = T1 + T2;
+		}
+
+	ctx->h[0] += a;	ctx->h[1] += b;	ctx->h[2] += c;	ctx->h[3] += d;
+	ctx->h[4] += e;	ctx->h[5] += f;	ctx->h[6] += g;	ctx->h[7] += h;
+
+			}
+}
+
+#else
+
+#define	ROUND_00_15(i,a,b,c,d,e,f,g,h)		do {	\
+	T1 += h + Sigma1(e) + Ch(e,f,g) + K256[i];	\
+	h = Sigma0(a) + Maj(a,b,c);			\
+	d += T1;	h += T1;		} while (0)
+
+#define	ROUND_16_63(i,a,b,c,d,e,f,g,h,X)	do {	\
+	s0 = X[(i+1)&0x0f];	s0 = sigma0(s0);	\
+	s1 = X[(i+14)&0x0f];	s1 = sigma1(s1);	\
+	T1 = X[(i)&0x0f] += s0 + s1 + X[(i+9)&0x0f];	\
+	ROUND_00_15(i,a,b,c,d,e,f,g,h);		} while (0)
+
+static void sha256_block_data_order (SHA256_CTX *ctx, const void *in, size_t num)
+	{
+	unsigned MD32_REG_T a,b,c,d,e,f,g,h,s0,s1,T1;
+	SHA_LONG	X[16];
+	int i;
+	const unsigned char *data=in;
+	const union { long one; char little; } is_endian = {1};
+
+			while (num--) {
+
+	a = ctx->h[0];	b = ctx->h[1];	c = ctx->h[2];	d = ctx->h[3];
+	e = ctx->h[4];	f = ctx->h[5];	g = ctx->h[6];	h = ctx->h[7];
+
+	if (!is_endian.little && sizeof(SHA_LONG)==4 && ((size_t)in%4)==0)
+		{
+		const SHA_LONG *W=(const SHA_LONG *)data;
+
+		T1 = X[0] = W[0];	ROUND_00_15(0,a,b,c,d,e,f,g,h);
+		T1 = X[1] = W[1];	ROUND_00_15(1,h,a,b,c,d,e,f,g);
+		T1 = X[2] = W[2];	ROUND_00_15(2,g,h,a,b,c,d,e,f);
+		T1 = X[3] = W[3];	ROUND_00_15(3,f,g,h,a,b,c,d,e);
+		T1 = X[4] = W[4];	ROUND_00_15(4,e,f,g,h,a,b,c,d);
+		T1 = X[5] = W[5];	ROUND_00_15(5,d,e,f,g,h,a,b,c);
+		T1 = X[6] = W[6];	ROUND_00_15(6,c,d,e,f,g,h,a,b);
+		T1 = X[7] = W[7];	ROUND_00_15(7,b,c,d,e,f,g,h,a);
+		T1 = X[8] = W[8];	ROUND_00_15(8,a,b,c,d,e,f,g,h);
+		T1 = X[9] = W[9];	ROUND_00_15(9,h,a,b,c,d,e,f,g);
+		T1 = X[10] = W[10];	ROUND_00_15(10,g,h,a,b,c,d,e,f);
+		T1 = X[11] = W[11];	ROUND_00_15(11,f,g,h,a,b,c,d,e);
+		T1 = X[12] = W[12];	ROUND_00_15(12,e,f,g,h,a,b,c,d);
+		T1 = X[13] = W[13];	ROUND_00_15(13,d,e,f,g,h,a,b,c);
+		T1 = X[14] = W[14];	ROUND_00_15(14,c,d,e,f,g,h,a,b);
+		T1 = X[15] = W[15];	ROUND_00_15(15,b,c,d,e,f,g,h,a);
+
+		data += SHA256_CBLOCK;
+		}
+	else
+		{
+		SHA_LONG l;
+
+		HOST_c2l(data,l); T1 = X[0] = l;  ROUND_00_15(0,a,b,c,d,e,f,g,h);
+		HOST_c2l(data,l); T1 = X[1] = l;  ROUND_00_15(1,h,a,b,c,d,e,f,g);
+		HOST_c2l(data,l); T1 = X[2] = l;  ROUND_00_15(2,g,h,a,b,c,d,e,f);
+		HOST_c2l(data,l); T1 = X[3] = l;  ROUND_00_15(3,f,g,h,a,b,c,d,e);
+		HOST_c2l(data,l); T1 = X[4] = l;  ROUND_00_15(4,e,f,g,h,a,b,c,d);
+		HOST_c2l(data,l); T1 = X[5] = l;  ROUND_00_15(5,d,e,f,g,h,a,b,c);
+		HOST_c2l(data,l); T1 = X[6] = l;  ROUND_00_15(6,c,d,e,f,g,h,a,b);
+		HOST_c2l(data,l); T1 = X[7] = l;  ROUND_00_15(7,b,c,d,e,f,g,h,a);
+		HOST_c2l(data,l); T1 = X[8] = l;  ROUND_00_15(8,a,b,c,d,e,f,g,h);
+		HOST_c2l(data,l); T1 = X[9] = l;  ROUND_00_15(9,h,a,b,c,d,e,f,g);
+		HOST_c2l(data,l); T1 = X[10] = l; ROUND_00_15(10,g,h,a,b,c,d,e,f);
+		HOST_c2l(data,l); T1 = X[11] = l; ROUND_00_15(11,f,g,h,a,b,c,d,e);
+		HOST_c2l(data,l); T1 = X[12] = l; ROUND_00_15(12,e,f,g,h,a,b,c,d);
+		HOST_c2l(data,l); T1 = X[13] = l; ROUND_00_15(13,d,e,f,g,h,a,b,c);
+		HOST_c2l(data,l); T1 = X[14] = l; ROUND_00_15(14,c,d,e,f,g,h,a,b);
+		HOST_c2l(data,l); T1 = X[15] = l; ROUND_00_15(15,b,c,d,e,f,g,h,a);
+		}
+
+	for (i=16;i<64;i+=8)
+		{
+		ROUND_16_63(i+0,a,b,c,d,e,f,g,h,X);
+		ROUND_16_63(i+1,h,a,b,c,d,e,f,g,X);
+		ROUND_16_63(i+2,g,h,a,b,c,d,e,f,X);
+		ROUND_16_63(i+3,f,g,h,a,b,c,d,e,X);
+		ROUND_16_63(i+4,e,f,g,h,a,b,c,d,X);
+		ROUND_16_63(i+5,d,e,f,g,h,a,b,c,X);
+		ROUND_16_63(i+6,c,d,e,f,g,h,a,b,X);
+		ROUND_16_63(i+7,b,c,d,e,f,g,h,a,X);
+		}
+
+	ctx->h[0] += a;	ctx->h[1] += b;	ctx->h[2] += c;	ctx->h[3] += d;
+	ctx->h[4] += e;	ctx->h[5] += f;	ctx->h[6] += g;	ctx->h[7] += h;
+
+			}
+	}
+
+#endif
+#endif /* SHA256_ASM */
+
+#endif /* OPENSSL_NO_SHA256 */
diff --git a/drivers/builtin_openssl2/crypto/sha/sha256t.c b/drivers/builtin_openssl2/crypto/sha/sha256t.c
new file mode 100644
index 0000000000..6b4a3bd001
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/sha256t.c
@@ -0,0 +1,147 @@
+/* crypto/sha/sha256t.c */
+/* ====================================================================
+ * Copyright (c) 2004 The OpenSSL Project.  All rights reserved.
+ * ====================================================================
+ */
+#include <stdio.h>
+#include <string.h>
+#include <stdlib.h>
+
+#include <openssl/sha.h>
+#include <openssl/evp.h>
+
+#if defined(OPENSSL_NO_SHA) || defined(OPENSSL_NO_SHA256)
+int main(int argc, char *argv[])
+{
+    printf("No SHA256 support\n");
+    return(0);
+}
+#else
+
+unsigned char app_b1[SHA256_DIGEST_LENGTH] = {
+	0xba,0x78,0x16,0xbf,0x8f,0x01,0xcf,0xea,
+	0x41,0x41,0x40,0xde,0x5d,0xae,0x22,0x23,
+	0xb0,0x03,0x61,0xa3,0x96,0x17,0x7a,0x9c,
+	0xb4,0x10,0xff,0x61,0xf2,0x00,0x15,0xad	};
+
+unsigned char app_b2[SHA256_DIGEST_LENGTH] = {
+	0x24,0x8d,0x6a,0x61,0xd2,0x06,0x38,0xb8,
+	0xe5,0xc0,0x26,0x93,0x0c,0x3e,0x60,0x39,
+	0xa3,0x3c,0xe4,0x59,0x64,0xff,0x21,0x67,
+	0xf6,0xec,0xed,0xd4,0x19,0xdb,0x06,0xc1	};
+
+unsigned char app_b3[SHA256_DIGEST_LENGTH] = {
+	0xcd,0xc7,0x6e,0x5c,0x99,0x14,0xfb,0x92,
+	0x81,0xa1,0xc7,0xe2,0x84,0xd7,0x3e,0x67,
+	0xf1,0x80,0x9a,0x48,0xa4,0x97,0x20,0x0e,
+	0x04,0x6d,0x39,0xcc,0xc7,0x11,0x2c,0xd0	};
+
+unsigned char addenum_1[SHA224_DIGEST_LENGTH] = {
+	0x23,0x09,0x7d,0x22,0x34,0x05,0xd8,0x22,
+	0x86,0x42,0xa4,0x77,0xbd,0xa2,0x55,0xb3,
+	0x2a,0xad,0xbc,0xe4,0xbd,0xa0,0xb3,0xf7,
+	0xe3,0x6c,0x9d,0xa7 };
+
+unsigned char addenum_2[SHA224_DIGEST_LENGTH] = {
+	0x75,0x38,0x8b,0x16,0x51,0x27,0x76,0xcc,
+	0x5d,0xba,0x5d,0xa1,0xfd,0x89,0x01,0x50,
+	0xb0,0xc6,0x45,0x5c,0xb4,0xf5,0x8b,0x19,
+	0x52,0x52,0x25,0x25 };
+
+unsigned char addenum_3[SHA224_DIGEST_LENGTH] = {
+	0x20,0x79,0x46,0x55,0x98,0x0c,0x91,0xd8,
+	0xbb,0xb4,0xc1,0xea,0x97,0x61,0x8a,0x4b,
+	0xf0,0x3f,0x42,0x58,0x19,0x48,0xb2,0xee,
+	0x4e,0xe7,0xad,0x67 };
+
+int main (int argc,char **argv)
+{ unsigned char md[SHA256_DIGEST_LENGTH];
+  int		i;
+  EVP_MD_CTX	evp;
+
+    fprintf(stdout,"Testing SHA-256 ");
+
+    EVP_Digest ("abc",3,md,NULL,EVP_sha256(),NULL);
+    if (memcmp(md,app_b1,sizeof(app_b1)))
+    {	fflush(stdout);
+	fprintf(stderr,"\nTEST 1 of 3 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    EVP_Digest ("abcdbcde""cdefdefg""efghfghi""ghijhijk"
+		"ijkljklm""klmnlmno""mnopnopq",56,md,NULL,EVP_sha256(),NULL);
+    if (memcmp(md,app_b2,sizeof(app_b2)))
+    {	fflush(stdout);
+	fprintf(stderr,"\nTEST 2 of 3 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    EVP_MD_CTX_init (&evp);
+    EVP_DigestInit_ex (&evp,EVP_sha256(),NULL);
+    for (i=0;i<1000000;i+=160)
+	EVP_DigestUpdate (&evp,	"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa",
+				(1000000-i)<160?1000000-i:160);
+    EVP_DigestFinal_ex (&evp,md,NULL);
+    EVP_MD_CTX_cleanup (&evp);
+
+    if (memcmp(md,app_b3,sizeof(app_b3)))
+    {	fflush(stdout);
+	fprintf(stderr,"\nTEST 3 of 3 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    fprintf(stdout," passed.\n"); fflush(stdout);
+
+    fprintf(stdout,"Testing SHA-224 ");
+
+    EVP_Digest ("abc",3,md,NULL,EVP_sha224(),NULL);
+    if (memcmp(md,addenum_1,sizeof(addenum_1)))
+    {	fflush(stdout);
+	fprintf(stderr,"\nTEST 1 of 3 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    EVP_Digest ("abcdbcde""cdefdefg""efghfghi""ghijhijk"
+		"ijkljklm""klmnlmno""mnopnopq",56,md,NULL,EVP_sha224(),NULL);
+    if (memcmp(md,addenum_2,sizeof(addenum_2)))
+    {	fflush(stdout);
+	fprintf(stderr,"\nTEST 2 of 3 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    EVP_MD_CTX_init (&evp);
+    EVP_DigestInit_ex (&evp,EVP_sha224(),NULL);
+    for (i=0;i<1000000;i+=64)
+	EVP_DigestUpdate (&evp,	"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa",
+				(1000000-i)<64?1000000-i:64);
+    EVP_DigestFinal_ex (&evp,md,NULL);
+    EVP_MD_CTX_cleanup (&evp);
+
+    if (memcmp(md,addenum_3,sizeof(addenum_3)))
+    {	fflush(stdout);
+	fprintf(stderr,"\nTEST 3 of 3 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    fprintf(stdout," passed.\n"); fflush(stdout);
+
+  return 0;
+}
+#endif
diff --git a/drivers/builtin_openssl2/crypto/sha/sha512.c b/drivers/builtin_openssl2/crypto/sha/sha512.c
new file mode 100644
index 0000000000..50c229ddeb
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/sha512.c
@@ -0,0 +1,604 @@
+/* crypto/sha/sha512.c */
+/* ====================================================================
+ * Copyright (c) 2004 The OpenSSL Project.  All rights reserved
+ * according to the OpenSSL license [found in ../../LICENSE].
+ * ====================================================================
+ */
+#include <openssl/opensslconf.h>
+#if !defined(OPENSSL_NO_SHA) && !defined(OPENSSL_NO_SHA512)
+/*
+ * IMPLEMENTATION NOTES.
+ *
+ * As you might have noticed 32-bit hash algorithms:
+ *
+ * - permit SHA_LONG to be wider than 32-bit (case on CRAY);
+ * - optimized versions implement two transform functions: one operating
+ *   on [aligned] data in host byte order and one - on data in input
+ *   stream byte order;
+ * - share common byte-order neutral collector and padding function
+ *   implementations, ../md32_common.h;
+ *
+ * Neither of the above applies to this SHA-512 implementations. Reasons
+ * [in reverse order] are:
+ *
+ * - it's the only 64-bit hash algorithm for the moment of this writing,
+ *   there is no need for common collector/padding implementation [yet];
+ * - by supporting only one transform function [which operates on
+ *   *aligned* data in input stream byte order, big-endian in this case]
+ *   we minimize burden of maintenance in two ways: a) collector/padding
+ *   function is simpler; b) only one transform function to stare at;
+ * - SHA_LONG64 is required to be exactly 64-bit in order to be able to
+ *   apply a number of optimizations to mitigate potential performance
+ *   penalties caused by previous design decision;
+ *
+ * Caveat lector.
+ *
+ * Implementation relies on the fact that "long long" is 64-bit on
+ * both 32- and 64-bit platforms. If some compiler vendor comes up
+ * with 128-bit long long, adjustment to sha.h would be required.
+ * As this implementation relies on 64-bit integer type, it's totally
+ * inappropriate for platforms which don't support it, most notably
+ * 16-bit platforms.
+ *					<appro@fy.chalmers.se>
+ */
+#include <stdlib.h>
+#include <string.h>
+
+#include <openssl/crypto.h>
+#include <openssl/sha.h>
+#include <openssl/opensslv.h>
+
+#include "cryptlib.h"
+
+const char SHA512_version[]="SHA-512" OPENSSL_VERSION_PTEXT;
+
+#if defined(__i386) || defined(__i386__) || defined(_M_IX86) || \
+    defined(__x86_64) || defined(_M_AMD64) || defined(_M_X64) || \
+    defined(__s390__) || defined(__s390x__) || \
+    defined(SHA512_ASM)
+#define SHA512_BLOCK_CAN_MANAGE_UNALIGNED_DATA
+#endif
+
+fips_md_init_ctx(SHA384, SHA512)
+	{
+	c->h[0]=U64(0xcbbb9d5dc1059ed8);
+	c->h[1]=U64(0x629a292a367cd507);
+	c->h[2]=U64(0x9159015a3070dd17);
+	c->h[3]=U64(0x152fecd8f70e5939);
+	c->h[4]=U64(0x67332667ffc00b31);
+	c->h[5]=U64(0x8eb44a8768581511);
+	c->h[6]=U64(0xdb0c2e0d64f98fa7);
+	c->h[7]=U64(0x47b5481dbefa4fa4);
+
+        c->Nl=0;        c->Nh=0;
+        c->num=0;       c->md_len=SHA384_DIGEST_LENGTH;
+        return 1;
+	}
+
+fips_md_init(SHA512)
+	{
+	c->h[0]=U64(0x6a09e667f3bcc908);
+	c->h[1]=U64(0xbb67ae8584caa73b);
+	c->h[2]=U64(0x3c6ef372fe94f82b);
+	c->h[3]=U64(0xa54ff53a5f1d36f1);
+	c->h[4]=U64(0x510e527fade682d1);
+	c->h[5]=U64(0x9b05688c2b3e6c1f);
+	c->h[6]=U64(0x1f83d9abfb41bd6b);
+	c->h[7]=U64(0x5be0cd19137e2179);
+
+        c->Nl=0;        c->Nh=0;
+        c->num=0;       c->md_len=SHA512_DIGEST_LENGTH;
+        return 1;
+	}
+
+#ifndef SHA512_ASM
+static
+#endif
+void sha512_block_data_order (SHA512_CTX *ctx, const void *in, size_t num);
+
+int SHA512_Final (unsigned char *md, SHA512_CTX *c)
+	{
+	unsigned char *p=(unsigned char *)c->u.p;
+	size_t n=c->num;
+
+	p[n]=0x80;	/* There always is a room for one */
+	n++;
+	if (n > (sizeof(c->u)-16))
+		memset (p+n,0,sizeof(c->u)-n), n=0,
+		sha512_block_data_order (c,p,1);
+
+	memset (p+n,0,sizeof(c->u)-16-n);
+#ifdef	B_ENDIAN
+	c->u.d[SHA_LBLOCK-2] = c->Nh;
+	c->u.d[SHA_LBLOCK-1] = c->Nl;
+#else
+	p[sizeof(c->u)-1]  = (unsigned char)(c->Nl);
+	p[sizeof(c->u)-2]  = (unsigned char)(c->Nl>>8);
+	p[sizeof(c->u)-3]  = (unsigned char)(c->Nl>>16);
+	p[sizeof(c->u)-4]  = (unsigned char)(c->Nl>>24);
+	p[sizeof(c->u)-5]  = (unsigned char)(c->Nl>>32);
+	p[sizeof(c->u)-6]  = (unsigned char)(c->Nl>>40);
+	p[sizeof(c->u)-7]  = (unsigned char)(c->Nl>>48);
+	p[sizeof(c->u)-8]  = (unsigned char)(c->Nl>>56);
+	p[sizeof(c->u)-9]  = (unsigned char)(c->Nh);
+	p[sizeof(c->u)-10] = (unsigned char)(c->Nh>>8);
+	p[sizeof(c->u)-11] = (unsigned char)(c->Nh>>16);
+	p[sizeof(c->u)-12] = (unsigned char)(c->Nh>>24);
+	p[sizeof(c->u)-13] = (unsigned char)(c->Nh>>32);
+	p[sizeof(c->u)-14] = (unsigned char)(c->Nh>>40);
+	p[sizeof(c->u)-15] = (unsigned char)(c->Nh>>48);
+	p[sizeof(c->u)-16] = (unsigned char)(c->Nh>>56);
+#endif
+
+	sha512_block_data_order (c,p,1);
+
+	if (md==0) return 0;
+
+	switch (c->md_len)
+		{
+		/* Let compiler decide if it's appropriate to unroll... */
+		case SHA384_DIGEST_LENGTH:
+			for (n=0;n<SHA384_DIGEST_LENGTH/8;n++)
+				{
+				SHA_LONG64 t = c->h[n];
+
+				*(md++)	= (unsigned char)(t>>56);
+				*(md++)	= (unsigned char)(t>>48);
+				*(md++)	= (unsigned char)(t>>40);
+				*(md++)	= (unsigned char)(t>>32);
+				*(md++)	= (unsigned char)(t>>24);
+				*(md++)	= (unsigned char)(t>>16);
+				*(md++)	= (unsigned char)(t>>8);
+				*(md++)	= (unsigned char)(t);
+				}
+			break;
+		case SHA512_DIGEST_LENGTH:
+			for (n=0;n<SHA512_DIGEST_LENGTH/8;n++)
+				{
+				SHA_LONG64 t = c->h[n];
+
+				*(md++)	= (unsigned char)(t>>56);
+				*(md++)	= (unsigned char)(t>>48);
+				*(md++)	= (unsigned char)(t>>40);
+				*(md++)	= (unsigned char)(t>>32);
+				*(md++)	= (unsigned char)(t>>24);
+				*(md++)	= (unsigned char)(t>>16);
+				*(md++)	= (unsigned char)(t>>8);
+				*(md++)	= (unsigned char)(t);
+				}
+			break;
+		/* ... as well as make sure md_len is not abused. */
+		default:	return 0;
+		}
+
+	return 1;
+	}
+
+int SHA384_Final (unsigned char *md,SHA512_CTX *c)
+{   return SHA512_Final (md,c);   }
+
+int SHA512_Update (SHA512_CTX *c, const void *_data, size_t len)
+	{
+	SHA_LONG64	l;
+	unsigned char  *p=c->u.p;
+	const unsigned char *data=(const unsigned char *)_data;
+
+	if (len==0) return  1;
+
+	l = (c->Nl+(((SHA_LONG64)len)<<3))&U64(0xffffffffffffffff);
+	if (l < c->Nl)		c->Nh++;
+	if (sizeof(len)>=8)	c->Nh+=(((SHA_LONG64)len)>>61);
+	c->Nl=l;
+
+	if (c->num != 0)
+		{
+		size_t n = sizeof(c->u) - c->num;
+
+		if (len < n)
+			{
+			memcpy (p+c->num,data,len), c->num += (unsigned int)len;
+			return 1;
+			}
+		else	{
+			memcpy (p+c->num,data,n), c->num = 0;
+			len-=n, data+=n;
+			sha512_block_data_order (c,p,1);
+			}
+		}
+
+	if (len >= sizeof(c->u))
+		{
+#ifndef SHA512_BLOCK_CAN_MANAGE_UNALIGNED_DATA
+		if ((size_t)data%sizeof(c->u.d[0]) != 0)
+			while (len >= sizeof(c->u))
+				memcpy (p,data,sizeof(c->u)),
+				sha512_block_data_order (c,p,1),
+				len  -= sizeof(c->u),
+				data += sizeof(c->u);
+		else
+#endif
+			sha512_block_data_order (c,data,len/sizeof(c->u)),
+			data += len,
+			len  %= sizeof(c->u),
+			data -= len;
+		}
+
+	if (len != 0)	memcpy (p,data,len), c->num = (int)len;
+
+	return 1;
+	}
+
+int SHA384_Update (SHA512_CTX *c, const void *data, size_t len)
+{   return SHA512_Update (c,data,len);   }
+
+void SHA512_Transform (SHA512_CTX *c, const unsigned char *data)
+	{
+#ifndef SHA512_BLOCK_CAN_MANAGE_UNALIGNED_DATA
+	if ((size_t)data%sizeof(c->u.d[0]) != 0)
+		memcpy(c->u.p,data,sizeof(c->u.p)),
+		data = c->u.p;
+#endif
+	sha512_block_data_order (c,data,1);
+	}
+
+unsigned char *SHA384(const unsigned char *d, size_t n, unsigned char *md)
+	{
+	SHA512_CTX c;
+	static unsigned char m[SHA384_DIGEST_LENGTH];
+
+	if (md == NULL) md=m;
+	SHA384_Init(&c);
+	SHA512_Update(&c,d,n);
+	SHA512_Final(md,&c);
+	OPENSSL_cleanse(&c,sizeof(c));
+	return(md);
+	}
+
+unsigned char *SHA512(const unsigned char *d, size_t n, unsigned char *md)
+	{
+	SHA512_CTX c;
+	static unsigned char m[SHA512_DIGEST_LENGTH];
+
+	if (md == NULL) md=m;
+	SHA512_Init(&c);
+	SHA512_Update(&c,d,n);
+	SHA512_Final(md,&c);
+	OPENSSL_cleanse(&c,sizeof(c));
+	return(md);
+	}
+
+#ifndef SHA512_ASM
+static const SHA_LONG64 K512[80] = {
+        U64(0x428a2f98d728ae22),U64(0x7137449123ef65cd),
+        U64(0xb5c0fbcfec4d3b2f),U64(0xe9b5dba58189dbbc),
+        U64(0x3956c25bf348b538),U64(0x59f111f1b605d019),
+        U64(0x923f82a4af194f9b),U64(0xab1c5ed5da6d8118),
+        U64(0xd807aa98a3030242),U64(0x12835b0145706fbe),
+        U64(0x243185be4ee4b28c),U64(0x550c7dc3d5ffb4e2),
+        U64(0x72be5d74f27b896f),U64(0x80deb1fe3b1696b1),
+        U64(0x9bdc06a725c71235),U64(0xc19bf174cf692694),
+        U64(0xe49b69c19ef14ad2),U64(0xefbe4786384f25e3),
+        U64(0x0fc19dc68b8cd5b5),U64(0x240ca1cc77ac9c65),
+        U64(0x2de92c6f592b0275),U64(0x4a7484aa6ea6e483),
+        U64(0x5cb0a9dcbd41fbd4),U64(0x76f988da831153b5),
+        U64(0x983e5152ee66dfab),U64(0xa831c66d2db43210),
+        U64(0xb00327c898fb213f),U64(0xbf597fc7beef0ee4),
+        U64(0xc6e00bf33da88fc2),U64(0xd5a79147930aa725),
+        U64(0x06ca6351e003826f),U64(0x142929670a0e6e70),
+        U64(0x27b70a8546d22ffc),U64(0x2e1b21385c26c926),
+        U64(0x4d2c6dfc5ac42aed),U64(0x53380d139d95b3df),
+        U64(0x650a73548baf63de),U64(0x766a0abb3c77b2a8),
+        U64(0x81c2c92e47edaee6),U64(0x92722c851482353b),
+        U64(0xa2bfe8a14cf10364),U64(0xa81a664bbc423001),
+        U64(0xc24b8b70d0f89791),U64(0xc76c51a30654be30),
+        U64(0xd192e819d6ef5218),U64(0xd69906245565a910),
+        U64(0xf40e35855771202a),U64(0x106aa07032bbd1b8),
+        U64(0x19a4c116b8d2d0c8),U64(0x1e376c085141ab53),
+        U64(0x2748774cdf8eeb99),U64(0x34b0bcb5e19b48a8),
+        U64(0x391c0cb3c5c95a63),U64(0x4ed8aa4ae3418acb),
+        U64(0x5b9cca4f7763e373),U64(0x682e6ff3d6b2b8a3),
+        U64(0x748f82ee5defb2fc),U64(0x78a5636f43172f60),
+        U64(0x84c87814a1f0ab72),U64(0x8cc702081a6439ec),
+        U64(0x90befffa23631e28),U64(0xa4506cebde82bde9),
+        U64(0xbef9a3f7b2c67915),U64(0xc67178f2e372532b),
+        U64(0xca273eceea26619c),U64(0xd186b8c721c0c207),
+        U64(0xeada7dd6cde0eb1e),U64(0xf57d4f7fee6ed178),
+        U64(0x06f067aa72176fba),U64(0x0a637dc5a2c898a6),
+        U64(0x113f9804bef90dae),U64(0x1b710b35131c471b),
+        U64(0x28db77f523047d84),U64(0x32caab7b40c72493),
+        U64(0x3c9ebe0a15c9bebc),U64(0x431d67c49c100d4c),
+        U64(0x4cc5d4becb3e42b6),U64(0x597f299cfc657e2a),
+        U64(0x5fcb6fab3ad6faec),U64(0x6c44198c4a475817) };
+
+#ifndef PEDANTIC
+# if defined(__GNUC__) && __GNUC__>=2 && !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM)
+#  if defined(__x86_64) || defined(__x86_64__)
+#   define ROTR(a,n)	({ SHA_LONG64 ret;		\
+				asm ("rorq %1,%0"	\
+				: "=r"(ret)		\
+				: "J"(n),"0"(a)		\
+				: "cc"); ret;		})
+#   if !defined(B_ENDIAN)
+#    define PULL64(x) ({ SHA_LONG64 ret=*((const SHA_LONG64 *)(&(x)));	\
+				asm ("bswapq	%0"		\
+				: "=r"(ret)			\
+				: "0"(ret)); ret;		})
+#   endif
+#  elif (defined(__i386) || defined(__i386__)) && !defined(B_ENDIAN)
+#   if defined(I386_ONLY)
+#    define PULL64(x) ({ const unsigned int *p=(const unsigned int *)(&(x));\
+			 unsigned int hi=p[0],lo=p[1];		\
+				asm("xchgb %%ah,%%al;xchgb %%dh,%%dl;"\
+				    "roll $16,%%eax; roll $16,%%edx; "\
+				    "xchgb %%ah,%%al;xchgb %%dh,%%dl;" \
+				: "=a"(lo),"=d"(hi)		\
+				: "0"(lo),"1"(hi) : "cc");	\
+				((SHA_LONG64)hi)<<32|lo;	})
+#   else
+#    define PULL64(x) ({ const unsigned int *p=(const unsigned int *)(&(x));\
+			 unsigned int hi=p[0],lo=p[1];		\
+				asm ("bswapl %0; bswapl %1;"	\
+				: "=r"(lo),"=r"(hi)		\
+				: "0"(lo),"1"(hi));		\
+				((SHA_LONG64)hi)<<32|lo;	})
+#   endif
+#  elif (defined(_ARCH_PPC) && defined(__64BIT__)) || defined(_ARCH_PPC64)
+#   define ROTR(a,n)	({ SHA_LONG64 ret;		\
+				asm ("rotrdi %0,%1,%2"	\
+				: "=r"(ret)		\
+				: "r"(a),"K"(n)); ret;	})
+#  endif
+# elif defined(_MSC_VER)
+#  if defined(_WIN64)	/* applies to both IA-64 and AMD64 */
+#   pragma intrinsic(_rotr64)
+#   define ROTR(a,n)	_rotr64((a),n)
+#  endif
+#  if defined(_M_IX86) && !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM)
+#   if defined(I386_ONLY)
+    static SHA_LONG64 __fastcall __pull64be(const void *x)
+    {	_asm	mov	edx, [ecx + 0]
+	_asm	mov	eax, [ecx + 4]
+	_asm	xchg	dh,dl
+	_asm	xchg	ah,al
+	_asm	rol	edx,16
+	_asm	rol	eax,16
+	_asm	xchg	dh,dl
+	_asm	xchg	ah,al
+    }
+#   else
+    static SHA_LONG64 __fastcall __pull64be(const void *x)
+    {	_asm	mov	edx, [ecx + 0]
+	_asm	mov	eax, [ecx + 4]
+	_asm	bswap	edx
+	_asm	bswap	eax
+    }
+#   endif
+#   define PULL64(x) __pull64be(&(x))
+#   if _MSC_VER<=1200
+#    pragma inline_depth(0)
+#   endif
+#  endif
+# endif
+#endif
+
+#ifndef PULL64
+#define B(x,j)    (((SHA_LONG64)(*(((const unsigned char *)(&x))+j)))<<((7-j)*8))
+#define PULL64(x) (B(x,0)|B(x,1)|B(x,2)|B(x,3)|B(x,4)|B(x,5)|B(x,6)|B(x,7))
+#endif
+
+#ifndef ROTR
+#define ROTR(x,s)	(((x)>>s) | (x)<<(64-s))
+#endif
+
+#define Sigma0(x)	(ROTR((x),28) ^ ROTR((x),34) ^ ROTR((x),39))
+#define Sigma1(x)	(ROTR((x),14) ^ ROTR((x),18) ^ ROTR((x),41))
+#define sigma0(x)	(ROTR((x),1)  ^ ROTR((x),8)  ^ ((x)>>7))
+#define sigma1(x)	(ROTR((x),19) ^ ROTR((x),61) ^ ((x)>>6))
+
+#define Ch(x,y,z)	(((x) & (y)) ^ ((~(x)) & (z)))
+#define Maj(x,y,z)	(((x) & (y)) ^ ((x) & (z)) ^ ((y) & (z)))
+
+
+#if defined(__i386) || defined(__i386__) || defined(_M_IX86)
+/*
+ * This code should give better results on 32-bit CPU with less than
+ * ~24 registers, both size and performance wise...
+ */
+static void sha512_block_data_order (SHA512_CTX *ctx, const void *in, size_t num)
+	{
+	const SHA_LONG64 *W=in;
+	SHA_LONG64	A,E,T;
+	SHA_LONG64	X[9+80],*F;
+	int i;
+
+			while (num--) {
+
+	F    = X+80;
+	A    = ctx->h[0];	F[1] = ctx->h[1];
+	F[2] = ctx->h[2];	F[3] = ctx->h[3];
+	E    = ctx->h[4];	F[5] = ctx->h[5];
+	F[6] = ctx->h[6];	F[7] = ctx->h[7];
+
+	for (i=0;i<16;i++,F--)
+		{
+#ifdef B_ENDIAN
+		T = W[i];
+#else
+		T = PULL64(W[i]);
+#endif
+		F[0] = A;
+		F[4] = E;
+		F[8] = T;
+		T   += F[7] + Sigma1(E) + Ch(E,F[5],F[6]) + K512[i];
+		E    = F[3] + T;
+		A    = T + Sigma0(A) + Maj(A,F[1],F[2]);
+		}
+
+	for (;i<80;i++,F--)
+		{
+		T    = sigma0(F[8+16-1]);
+		T   += sigma1(F[8+16-14]);
+		T   += F[8+16] + F[8+16-9];
+
+		F[0] = A;
+		F[4] = E;
+		F[8] = T;
+		T   += F[7] + Sigma1(E) + Ch(E,F[5],F[6]) + K512[i];
+		E    = F[3] + T;
+		A    = T + Sigma0(A) + Maj(A,F[1],F[2]);
+		}
+
+	ctx->h[0] += A;		ctx->h[1] += F[1];
+	ctx->h[2] += F[2];	ctx->h[3] += F[3];
+	ctx->h[4] += E;		ctx->h[5] += F[5];
+	ctx->h[6] += F[6];	ctx->h[7] += F[7];
+
+			W+=SHA_LBLOCK;
+			}
+	}
+
+#elif defined(OPENSSL_SMALL_FOOTPRINT)
+
+static void sha512_block_data_order (SHA512_CTX *ctx, const void *in, size_t num)
+	{
+	const SHA_LONG64 *W=in;
+	SHA_LONG64	a,b,c,d,e,f,g,h,s0,s1,T1,T2;
+	SHA_LONG64	X[16];
+	int i;
+
+			while (num--) {
+
+	a = ctx->h[0];	b = ctx->h[1];	c = ctx->h[2];	d = ctx->h[3];
+	e = ctx->h[4];	f = ctx->h[5];	g = ctx->h[6];	h = ctx->h[7];
+
+	for (i=0;i<16;i++)
+		{
+#ifdef B_ENDIAN
+		T1 = X[i] = W[i];
+#else
+		T1 = X[i] = PULL64(W[i]);
+#endif
+		T1 += h + Sigma1(e) + Ch(e,f,g) + K512[i];
+		T2 = Sigma0(a) + Maj(a,b,c);
+		h = g;	g = f;	f = e;	e = d + T1;
+		d = c;	c = b;	b = a;	a = T1 + T2;
+		}
+
+	for (;i<80;i++)
+		{
+		s0 = X[(i+1)&0x0f];	s0 = sigma0(s0);
+		s1 = X[(i+14)&0x0f];	s1 = sigma1(s1);
+
+		T1 = X[i&0xf] += s0 + s1 + X[(i+9)&0xf];
+		T1 += h + Sigma1(e) + Ch(e,f,g) + K512[i];
+		T2 = Sigma0(a) + Maj(a,b,c);
+		h = g;	g = f;	f = e;	e = d + T1;
+		d = c;	c = b;	b = a;	a = T1 + T2;
+		}
+
+	ctx->h[0] += a;	ctx->h[1] += b;	ctx->h[2] += c;	ctx->h[3] += d;
+	ctx->h[4] += e;	ctx->h[5] += f;	ctx->h[6] += g;	ctx->h[7] += h;
+
+			W+=SHA_LBLOCK;
+			}
+	}
+
+#else
+
+#define	ROUND_00_15(i,a,b,c,d,e,f,g,h)		do {	\
+	T1 += h + Sigma1(e) + Ch(e,f,g) + K512[i];	\
+	h = Sigma0(a) + Maj(a,b,c);			\
+	d += T1;	h += T1;		} while (0)
+
+#define	ROUND_16_80(i,j,a,b,c,d,e,f,g,h,X)	do {	\
+	s0 = X[(j+1)&0x0f];	s0 = sigma0(s0);	\
+	s1 = X[(j+14)&0x0f];	s1 = sigma1(s1);	\
+	T1 = X[(j)&0x0f] += s0 + s1 + X[(j+9)&0x0f];	\
+	ROUND_00_15(i+j,a,b,c,d,e,f,g,h);		} while (0)
+
+static void sha512_block_data_order (SHA512_CTX *ctx, const void *in, size_t num)
+	{
+	const SHA_LONG64 *W=in;
+	SHA_LONG64	a,b,c,d,e,f,g,h,s0,s1,T1;
+	SHA_LONG64	X[16];
+	int i;
+
+			while (num--) {
+
+	a = ctx->h[0];	b = ctx->h[1];	c = ctx->h[2];	d = ctx->h[3];
+	e = ctx->h[4];	f = ctx->h[5];	g = ctx->h[6];	h = ctx->h[7];
+
+#ifdef B_ENDIAN
+	T1 = X[0] = W[0];	ROUND_00_15(0,a,b,c,d,e,f,g,h);
+	T1 = X[1] = W[1];	ROUND_00_15(1,h,a,b,c,d,e,f,g);
+	T1 = X[2] = W[2];	ROUND_00_15(2,g,h,a,b,c,d,e,f);
+	T1 = X[3] = W[3];	ROUND_00_15(3,f,g,h,a,b,c,d,e);
+	T1 = X[4] = W[4];	ROUND_00_15(4,e,f,g,h,a,b,c,d);
+	T1 = X[5] = W[5];	ROUND_00_15(5,d,e,f,g,h,a,b,c);
+	T1 = X[6] = W[6];	ROUND_00_15(6,c,d,e,f,g,h,a,b);
+	T1 = X[7] = W[7];	ROUND_00_15(7,b,c,d,e,f,g,h,a);
+	T1 = X[8] = W[8];	ROUND_00_15(8,a,b,c,d,e,f,g,h);
+	T1 = X[9] = W[9];	ROUND_00_15(9,h,a,b,c,d,e,f,g);
+	T1 = X[10] = W[10];	ROUND_00_15(10,g,h,a,b,c,d,e,f);
+	T1 = X[11] = W[11];	ROUND_00_15(11,f,g,h,a,b,c,d,e);
+	T1 = X[12] = W[12];	ROUND_00_15(12,e,f,g,h,a,b,c,d);
+	T1 = X[13] = W[13];	ROUND_00_15(13,d,e,f,g,h,a,b,c);
+	T1 = X[14] = W[14];	ROUND_00_15(14,c,d,e,f,g,h,a,b);
+	T1 = X[15] = W[15];	ROUND_00_15(15,b,c,d,e,f,g,h,a);
+#else
+	T1 = X[0]  = PULL64(W[0]);	ROUND_00_15(0,a,b,c,d,e,f,g,h);
+	T1 = X[1]  = PULL64(W[1]);	ROUND_00_15(1,h,a,b,c,d,e,f,g);
+	T1 = X[2]  = PULL64(W[2]);	ROUND_00_15(2,g,h,a,b,c,d,e,f);
+	T1 = X[3]  = PULL64(W[3]);	ROUND_00_15(3,f,g,h,a,b,c,d,e);
+	T1 = X[4]  = PULL64(W[4]);	ROUND_00_15(4,e,f,g,h,a,b,c,d);
+	T1 = X[5]  = PULL64(W[5]);	ROUND_00_15(5,d,e,f,g,h,a,b,c);
+	T1 = X[6]  = PULL64(W[6]);	ROUND_00_15(6,c,d,e,f,g,h,a,b);
+	T1 = X[7]  = PULL64(W[7]);	ROUND_00_15(7,b,c,d,e,f,g,h,a);
+	T1 = X[8]  = PULL64(W[8]);	ROUND_00_15(8,a,b,c,d,e,f,g,h);
+	T1 = X[9]  = PULL64(W[9]);	ROUND_00_15(9,h,a,b,c,d,e,f,g);
+	T1 = X[10] = PULL64(W[10]);	ROUND_00_15(10,g,h,a,b,c,d,e,f);
+	T1 = X[11] = PULL64(W[11]);	ROUND_00_15(11,f,g,h,a,b,c,d,e);
+	T1 = X[12] = PULL64(W[12]);	ROUND_00_15(12,e,f,g,h,a,b,c,d);
+	T1 = X[13] = PULL64(W[13]);	ROUND_00_15(13,d,e,f,g,h,a,b,c);
+	T1 = X[14] = PULL64(W[14]);	ROUND_00_15(14,c,d,e,f,g,h,a,b);
+	T1 = X[15] = PULL64(W[15]);	ROUND_00_15(15,b,c,d,e,f,g,h,a);
+#endif
+
+	for (i=16;i<80;i+=16)
+		{
+		ROUND_16_80(i, 0,a,b,c,d,e,f,g,h,X);
+		ROUND_16_80(i, 1,h,a,b,c,d,e,f,g,X);
+		ROUND_16_80(i, 2,g,h,a,b,c,d,e,f,X);
+		ROUND_16_80(i, 3,f,g,h,a,b,c,d,e,X);
+		ROUND_16_80(i, 4,e,f,g,h,a,b,c,d,X);
+		ROUND_16_80(i, 5,d,e,f,g,h,a,b,c,X);
+		ROUND_16_80(i, 6,c,d,e,f,g,h,a,b,X);
+		ROUND_16_80(i, 7,b,c,d,e,f,g,h,a,X);
+		ROUND_16_80(i, 8,a,b,c,d,e,f,g,h,X);
+		ROUND_16_80(i, 9,h,a,b,c,d,e,f,g,X);
+		ROUND_16_80(i,10,g,h,a,b,c,d,e,f,X);
+		ROUND_16_80(i,11,f,g,h,a,b,c,d,e,X);
+		ROUND_16_80(i,12,e,f,g,h,a,b,c,d,X);
+		ROUND_16_80(i,13,d,e,f,g,h,a,b,c,X);
+		ROUND_16_80(i,14,c,d,e,f,g,h,a,b,X);
+		ROUND_16_80(i,15,b,c,d,e,f,g,h,a,X);
+		}
+
+	ctx->h[0] += a;	ctx->h[1] += b;	ctx->h[2] += c;	ctx->h[3] += d;
+	ctx->h[4] += e;	ctx->h[5] += f;	ctx->h[6] += g;	ctx->h[7] += h;
+
+			W+=SHA_LBLOCK;
+			}
+	}
+
+#endif
+
+#endif /* SHA512_ASM */
+
+#else /* !OPENSSL_NO_SHA512 */
+
+#if defined(PEDANTIC) || defined(__DECC) || defined(OPENSSL_SYS_MACOSX)
+static void *dummy=&dummy;
+#endif
+
+#endif /* !OPENSSL_NO_SHA512 */
diff --git a/drivers/builtin_openssl2/crypto/sha/sha512t.c b/drivers/builtin_openssl2/crypto/sha/sha512t.c
new file mode 100644
index 0000000000..210041d435
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/sha512t.c
@@ -0,0 +1,184 @@
+/* crypto/sha/sha512t.c */
+/* ====================================================================
+ * Copyright (c) 2004 The OpenSSL Project.  All rights reserved.
+ * ====================================================================
+ */
+#include <stdio.h>
+#include <string.h>
+#include <stdlib.h>
+
+#include <openssl/sha.h>
+#include <openssl/evp.h>
+#include <openssl/crypto.h>
+
+#if defined(OPENSSL_NO_SHA) || defined(OPENSSL_NO_SHA512)
+int main(int argc, char *argv[])
+{
+    printf("No SHA512 support\n");
+    return(0);
+}
+#else
+
+unsigned char app_c1[SHA512_DIGEST_LENGTH] = {
+	0xdd,0xaf,0x35,0xa1,0x93,0x61,0x7a,0xba,
+	0xcc,0x41,0x73,0x49,0xae,0x20,0x41,0x31,
+	0x12,0xe6,0xfa,0x4e,0x89,0xa9,0x7e,0xa2,
+	0x0a,0x9e,0xee,0xe6,0x4b,0x55,0xd3,0x9a,
+	0x21,0x92,0x99,0x2a,0x27,0x4f,0xc1,0xa8,
+	0x36,0xba,0x3c,0x23,0xa3,0xfe,0xeb,0xbd,
+	0x45,0x4d,0x44,0x23,0x64,0x3c,0xe8,0x0e,
+	0x2a,0x9a,0xc9,0x4f,0xa5,0x4c,0xa4,0x9f };
+
+unsigned char app_c2[SHA512_DIGEST_LENGTH] = {
+	0x8e,0x95,0x9b,0x75,0xda,0xe3,0x13,0xda,
+	0x8c,0xf4,0xf7,0x28,0x14,0xfc,0x14,0x3f,
+	0x8f,0x77,0x79,0xc6,0xeb,0x9f,0x7f,0xa1,
+	0x72,0x99,0xae,0xad,0xb6,0x88,0x90,0x18,
+	0x50,0x1d,0x28,0x9e,0x49,0x00,0xf7,0xe4,
+	0x33,0x1b,0x99,0xde,0xc4,0xb5,0x43,0x3a,
+	0xc7,0xd3,0x29,0xee,0xb6,0xdd,0x26,0x54,
+	0x5e,0x96,0xe5,0x5b,0x87,0x4b,0xe9,0x09 };
+
+unsigned char app_c3[SHA512_DIGEST_LENGTH] = {
+	0xe7,0x18,0x48,0x3d,0x0c,0xe7,0x69,0x64,
+	0x4e,0x2e,0x42,0xc7,0xbc,0x15,0xb4,0x63,
+	0x8e,0x1f,0x98,0xb1,0x3b,0x20,0x44,0x28,
+	0x56,0x32,0xa8,0x03,0xaf,0xa9,0x73,0xeb,
+	0xde,0x0f,0xf2,0x44,0x87,0x7e,0xa6,0x0a,
+	0x4c,0xb0,0x43,0x2c,0xe5,0x77,0xc3,0x1b,
+	0xeb,0x00,0x9c,0x5c,0x2c,0x49,0xaa,0x2e,
+	0x4e,0xad,0xb2,0x17,0xad,0x8c,0xc0,0x9b };
+
+unsigned char app_d1[SHA384_DIGEST_LENGTH] = {
+	0xcb,0x00,0x75,0x3f,0x45,0xa3,0x5e,0x8b,
+	0xb5,0xa0,0x3d,0x69,0x9a,0xc6,0x50,0x07,
+	0x27,0x2c,0x32,0xab,0x0e,0xde,0xd1,0x63,
+	0x1a,0x8b,0x60,0x5a,0x43,0xff,0x5b,0xed,
+	0x80,0x86,0x07,0x2b,0xa1,0xe7,0xcc,0x23,
+	0x58,0xba,0xec,0xa1,0x34,0xc8,0x25,0xa7 };
+
+unsigned char app_d2[SHA384_DIGEST_LENGTH] = {
+	0x09,0x33,0x0c,0x33,0xf7,0x11,0x47,0xe8,
+	0x3d,0x19,0x2f,0xc7,0x82,0xcd,0x1b,0x47,
+	0x53,0x11,0x1b,0x17,0x3b,0x3b,0x05,0xd2,
+	0x2f,0xa0,0x80,0x86,0xe3,0xb0,0xf7,0x12,
+	0xfc,0xc7,0xc7,0x1a,0x55,0x7e,0x2d,0xb9,
+	0x66,0xc3,0xe9,0xfa,0x91,0x74,0x60,0x39 };
+
+unsigned char app_d3[SHA384_DIGEST_LENGTH] = {
+	0x9d,0x0e,0x18,0x09,0x71,0x64,0x74,0xcb,
+	0x08,0x6e,0x83,0x4e,0x31,0x0a,0x4a,0x1c,
+	0xed,0x14,0x9e,0x9c,0x00,0xf2,0x48,0x52,
+	0x79,0x72,0xce,0xc5,0x70,0x4c,0x2a,0x5b,
+	0x07,0xb8,0xb3,0xdc,0x38,0xec,0xc4,0xeb,
+	0xae,0x97,0xdd,0xd8,0x7f,0x3d,0x89,0x85 };
+
+int main (int argc,char **argv)
+{ unsigned char md[SHA512_DIGEST_LENGTH];
+  int		i;
+  EVP_MD_CTX	evp;
+
+#ifdef OPENSSL_IA32_SSE2
+    /* Alternative to this is to call OpenSSL_add_all_algorithms...
+     * The below code is retained exclusively for debugging purposes. */
+    { char      *env;
+
+	if ((env=getenv("OPENSSL_ia32cap")))
+	    OPENSSL_ia32cap = strtoul (env,NULL,0);
+    }
+#endif
+
+    fprintf(stdout,"Testing SHA-512 ");
+
+    EVP_Digest ("abc",3,md,NULL,EVP_sha512(),NULL);
+    if (memcmp(md,app_c1,sizeof(app_c1)))
+    {	fflush(stdout);
+	fprintf(stderr,"\nTEST 1 of 3 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    EVP_Digest ("abcdefgh""bcdefghi""cdefghij""defghijk"
+		"efghijkl""fghijklm""ghijklmn""hijklmno"
+		"ijklmnop""jklmnopq""klmnopqr""lmnopqrs"
+		"mnopqrst""nopqrstu",112,md,NULL,EVP_sha512(),NULL);
+    if (memcmp(md,app_c2,sizeof(app_c2)))
+    {	fflush(stdout);
+	fprintf(stderr,"\nTEST 2 of 3 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    EVP_MD_CTX_init (&evp);
+    EVP_DigestInit_ex (&evp,EVP_sha512(),NULL);
+    for (i=0;i<1000000;i+=288)
+	EVP_DigestUpdate (&evp,	"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa",
+				(1000000-i)<288?1000000-i:288);
+    EVP_DigestFinal_ex (&evp,md,NULL);
+    EVP_MD_CTX_cleanup (&evp);
+
+    if (memcmp(md,app_c3,sizeof(app_c3)))
+    {	fflush(stdout);
+	fprintf(stderr,"\nTEST 3 of 3 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    fprintf(stdout," passed.\n"); fflush(stdout);
+
+    fprintf(stdout,"Testing SHA-384 ");
+
+    EVP_Digest ("abc",3,md,NULL,EVP_sha384(),NULL);
+    if (memcmp(md,app_d1,sizeof(app_d1)))
+    {	fflush(stdout);
+	fprintf(stderr,"\nTEST 1 of 3 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    EVP_Digest ("abcdefgh""bcdefghi""cdefghij""defghijk"
+		"efghijkl""fghijklm""ghijklmn""hijklmno"
+		"ijklmnop""jklmnopq""klmnopqr""lmnopqrs"
+		"mnopqrst""nopqrstu",112,md,NULL,EVP_sha384(),NULL);
+    if (memcmp(md,app_d2,sizeof(app_d2)))
+    {	fflush(stdout);
+	fprintf(stderr,"\nTEST 2 of 3 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    EVP_MD_CTX_init (&evp);
+    EVP_DigestInit_ex (&evp,EVP_sha384(),NULL);
+    for (i=0;i<1000000;i+=64)
+	EVP_DigestUpdate (&evp,	"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa"
+				"aaaaaaaa""aaaaaaaa""aaaaaaaa""aaaaaaaa",
+				(1000000-i)<64?1000000-i:64);
+    EVP_DigestFinal_ex (&evp,md,NULL);
+    EVP_MD_CTX_cleanup (&evp);
+
+    if (memcmp(md,app_d3,sizeof(app_d3)))
+    {	fflush(stdout);
+	fprintf(stderr,"\nTEST 3 of 3 failed.\n");
+	return 1;
+    }
+    else
+	fprintf(stdout,"."); fflush(stdout);
+
+    fprintf(stdout," passed.\n"); fflush(stdout);
+
+  return 0;
+}
+#endif
diff --git a/drivers/builtin_openssl2/crypto/sha/sha_dgst.c b/drivers/builtin_openssl2/crypto/sha/sha_dgst.c
new file mode 100644
index 0000000000..fb63b17ff2
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/sha_dgst.c
@@ -0,0 +1,75 @@
+/* crypto/sha/sha1dgst.c */
+/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
+ * All rights reserved.
+ *
+ * This package is an SSL implementation written
+ * by Eric Young (eay@cryptsoft.com).
+ * The implementation was written so as to conform with Netscapes SSL.
+ * 
+ * This library is free for commercial and non-commercial use as long as
+ * the following conditions are aheared to.  The following conditions
+ * apply to all code found in this distribution, be it the RC4, RSA,
+ * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
+ * included with this distribution is covered by the same copyright terms
+ * except that the holder is Tim Hudson (tjh@cryptsoft.com).
+ * 
+ * Copyright remains Eric Young's, and as such any Copyright notices in
+ * the code are not to be removed.
+ * If this package is used in a product, Eric Young should be given attribution
+ * as the author of the parts of the library used.
+ * This can be in the form of a textual message at program startup or
+ * in documentation (online or textual) provided with the package.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. All advertising materials mentioning features or use of this software
+ *    must display the following acknowledgement:
+ *    "This product includes cryptographic software written by
+ *     Eric Young (eay@cryptsoft.com)"
+ *    The word 'cryptographic' can be left out if the rouines from the library
+ *    being used are not cryptographic related :-).
+ * 4. If you include any Windows specific code (or a derivative thereof) from 
+ *    the apps directory (application code) you must include an acknowledgement:
+ *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
+ * 
+ * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ * 
+ * The licence and distribution terms for any publically available version or
+ * derivative of this code cannot be changed.  i.e. this code cannot simply be
+ * copied and put under another distribution licence
+ * [including the GNU Public Licence.]
+ */
+
+#include <openssl/crypto.h>
+#include <openssl/opensslconf.h>
+#if !defined(OPENSSL_NO_SHA0) && !defined(OPENSSL_NO_SHA)
+
+#undef  SHA_1
+#define SHA_0
+
+#include <openssl/opensslv.h>
+
+const char SHA_version[]="SHA" OPENSSL_VERSION_PTEXT;
+
+/* The implementation is in ../md32_common.h */
+
+#include "sha_locl.h"
+
+#endif
+
diff --git a/drivers/builtin_openssl2/crypto/sha/sha_locl.h b/drivers/builtin_openssl2/crypto/sha/sha_locl.h
new file mode 100644
index 0000000000..d673255f78
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/sha_locl.h
@@ -0,0 +1,441 @@
+/* crypto/sha/sha_locl.h */
+/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
+ * All rights reserved.
+ *
+ * This package is an SSL implementation written
+ * by Eric Young (eay@cryptsoft.com).
+ * The implementation was written so as to conform with Netscapes SSL.
+ * 
+ * This library is free for commercial and non-commercial use as long as
+ * the following conditions are aheared to.  The following conditions
+ * apply to all code found in this distribution, be it the RC4, RSA,
+ * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
+ * included with this distribution is covered by the same copyright terms
+ * except that the holder is Tim Hudson (tjh@cryptsoft.com).
+ * 
+ * Copyright remains Eric Young's, and as such any Copyright notices in
+ * the code are not to be removed.
+ * If this package is used in a product, Eric Young should be given attribution
+ * as the author of the parts of the library used.
+ * This can be in the form of a textual message at program startup or
+ * in documentation (online or textual) provided with the package.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. All advertising materials mentioning features or use of this software
+ *    must display the following acknowledgement:
+ *    "This product includes cryptographic software written by
+ *     Eric Young (eay@cryptsoft.com)"
+ *    The word 'cryptographic' can be left out if the rouines from the library
+ *    being used are not cryptographic related :-).
+ * 4. If you include any Windows specific code (or a derivative thereof) from 
+ *    the apps directory (application code) you must include an acknowledgement:
+ *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
+ * 
+ * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ * 
+ * The licence and distribution terms for any publically available version or
+ * derivative of this code cannot be changed.  i.e. this code cannot simply be
+ * copied and put under another distribution licence
+ * [including the GNU Public Licence.]
+ */
+
+#include <stdlib.h>
+#include <string.h>
+
+#include <openssl/opensslconf.h>
+#include <openssl/sha.h>
+
+#define DATA_ORDER_IS_BIG_ENDIAN
+
+#define HASH_LONG               SHA_LONG
+#define HASH_CTX                SHA_CTX
+#define HASH_CBLOCK             SHA_CBLOCK
+#define HASH_MAKE_STRING(c,s)   do {	\
+	unsigned long ll;		\
+	ll=(c)->h0; (void)HOST_l2c(ll,(s));	\
+	ll=(c)->h1; (void)HOST_l2c(ll,(s));	\
+	ll=(c)->h2; (void)HOST_l2c(ll,(s));	\
+	ll=(c)->h3; (void)HOST_l2c(ll,(s));	\
+	ll=(c)->h4; (void)HOST_l2c(ll,(s));	\
+	} while (0)
+
+#if defined(SHA_0)
+
+# define HASH_UPDATE             	SHA_Update
+# define HASH_TRANSFORM          	SHA_Transform
+# define HASH_FINAL              	SHA_Final
+# define HASH_INIT			SHA_Init
+# define HASH_BLOCK_DATA_ORDER   	sha_block_data_order
+# define Xupdate(a,ix,ia,ib,ic,id)	(ix=(a)=(ia^ib^ic^id))
+
+static void sha_block_data_order (SHA_CTX *c, const void *p,size_t num);
+
+#elif defined(SHA_1)
+
+# define HASH_UPDATE             	SHA1_Update
+# define HASH_TRANSFORM          	SHA1_Transform
+# define HASH_FINAL              	SHA1_Final
+# define HASH_INIT			SHA1_Init
+# define HASH_BLOCK_DATA_ORDER   	sha1_block_data_order
+# if defined(__MWERKS__) && defined(__MC68K__)
+   /* Metrowerks for Motorola fails otherwise:-( <appro@fy.chalmers.se> */
+#  define Xupdate(a,ix,ia,ib,ic,id)	do { (a)=(ia^ib^ic^id);		\
+					     ix=(a)=ROTATE((a),1);	\
+					} while (0)
+# else
+#  define Xupdate(a,ix,ia,ib,ic,id)	( (a)=(ia^ib^ic^id),	\
+					  ix=(a)=ROTATE((a),1)	\
+					)
+# endif
+
+#ifndef SHA1_ASM
+static
+#endif
+void sha1_block_data_order (SHA_CTX *c, const void *p,size_t num);
+
+#else
+# error "Either SHA_0 or SHA_1 must be defined."
+#endif
+
+#include "md32_common.h"
+
+#define INIT_DATA_h0 0x67452301UL
+#define INIT_DATA_h1 0xefcdab89UL
+#define INIT_DATA_h2 0x98badcfeUL
+#define INIT_DATA_h3 0x10325476UL
+#define INIT_DATA_h4 0xc3d2e1f0UL
+
+#ifdef SHA_0
+fips_md_init(SHA)
+#else
+fips_md_init_ctx(SHA1, SHA)
+#endif
+	{
+	memset (c,0,sizeof(*c));
+	c->h0=INIT_DATA_h0;
+	c->h1=INIT_DATA_h1;
+	c->h2=INIT_DATA_h2;
+	c->h3=INIT_DATA_h3;
+	c->h4=INIT_DATA_h4;
+	return 1;
+	}
+
+#define K_00_19	0x5a827999UL
+#define K_20_39 0x6ed9eba1UL
+#define K_40_59 0x8f1bbcdcUL
+#define K_60_79 0xca62c1d6UL
+
+/* As  pointed out by Wei Dai <weidai@eskimo.com>, F() below can be
+ * simplified to the code in F_00_19.  Wei attributes these optimisations
+ * to Peter Gutmann's SHS code, and he attributes it to Rich Schroeppel.
+ * #define F(x,y,z) (((x) & (y))  |  ((~(x)) & (z)))
+ * I've just become aware of another tweak to be made, again from Wei Dai,
+ * in F_40_59, (x&a)|(y&a) -> (x|y)&a
+ */
+#define	F_00_19(b,c,d)	((((c) ^ (d)) & (b)) ^ (d)) 
+#define	F_20_39(b,c,d)	((b) ^ (c) ^ (d))
+#define F_40_59(b,c,d)	(((b) & (c)) | (((b)|(c)) & (d))) 
+#define	F_60_79(b,c,d)	F_20_39(b,c,d)
+
+#ifndef OPENSSL_SMALL_FOOTPRINT
+
+#define BODY_00_15(i,a,b,c,d,e,f,xi) \
+	(f)=xi+(e)+K_00_19+ROTATE((a),5)+F_00_19((b),(c),(d)); \
+	(b)=ROTATE((b),30);
+
+#define BODY_16_19(i,a,b,c,d,e,f,xi,xa,xb,xc,xd) \
+	Xupdate(f,xi,xa,xb,xc,xd); \
+	(f)+=(e)+K_00_19+ROTATE((a),5)+F_00_19((b),(c),(d)); \
+	(b)=ROTATE((b),30);
+
+#define BODY_20_31(i,a,b,c,d,e,f,xi,xa,xb,xc,xd) \
+	Xupdate(f,xi,xa,xb,xc,xd); \
+	(f)+=(e)+K_20_39+ROTATE((a),5)+F_20_39((b),(c),(d)); \
+	(b)=ROTATE((b),30);
+
+#define BODY_32_39(i,a,b,c,d,e,f,xa,xb,xc,xd) \
+	Xupdate(f,xa,xa,xb,xc,xd); \
+	(f)+=(e)+K_20_39+ROTATE((a),5)+F_20_39((b),(c),(d)); \
+	(b)=ROTATE((b),30);
+
+#define BODY_40_59(i,a,b,c,d,e,f,xa,xb,xc,xd) \
+	Xupdate(f,xa,xa,xb,xc,xd); \
+	(f)+=(e)+K_40_59+ROTATE((a),5)+F_40_59((b),(c),(d)); \
+	(b)=ROTATE((b),30);
+
+#define BODY_60_79(i,a,b,c,d,e,f,xa,xb,xc,xd) \
+	Xupdate(f,xa,xa,xb,xc,xd); \
+	(f)=xa+(e)+K_60_79+ROTATE((a),5)+F_60_79((b),(c),(d)); \
+	(b)=ROTATE((b),30);
+
+#ifdef X
+#undef X
+#endif
+#ifndef MD32_XARRAY
+  /*
+   * Originally X was an array. As it's automatic it's natural
+   * to expect RISC compiler to accomodate at least part of it in
+   * the register bank, isn't it? Unfortunately not all compilers
+   * "find" this expectation reasonable:-( On order to make such
+   * compilers generate better code I replace X[] with a bunch of
+   * X0, X1, etc. See the function body below...
+   *					<appro@fy.chalmers.se>
+   */
+# define X(i)	XX##i
+#else
+  /*
+   * However! Some compilers (most notably HP C) get overwhelmed by
+   * that many local variables so that we have to have the way to
+   * fall down to the original behavior.
+   */
+# define X(i)	XX[i]
+#endif
+
+#if !defined(SHA_1) || !defined(SHA1_ASM)
+static void HASH_BLOCK_DATA_ORDER (SHA_CTX *c, const void *p, size_t num)
+	{
+	const unsigned char *data=p;
+	register unsigned MD32_REG_T A,B,C,D,E,T,l;
+#ifndef MD32_XARRAY
+	unsigned MD32_REG_T	XX0, XX1, XX2, XX3, XX4, XX5, XX6, XX7,
+				XX8, XX9,XX10,XX11,XX12,XX13,XX14,XX15;
+#else
+	SHA_LONG	XX[16];
+#endif
+
+	A=c->h0;
+	B=c->h1;
+	C=c->h2;
+	D=c->h3;
+	E=c->h4;
+
+	for (;;)
+			{
+	const union { long one; char little; } is_endian = {1};
+
+	if (!is_endian.little && sizeof(SHA_LONG)==4 && ((size_t)p%4)==0)
+		{
+		const SHA_LONG *W=(const SHA_LONG *)data;
+
+		X( 0) = W[0];				X( 1) = W[ 1];
+		BODY_00_15( 0,A,B,C,D,E,T,X( 0));	X( 2) = W[ 2];
+		BODY_00_15( 1,T,A,B,C,D,E,X( 1));	X( 3) = W[ 3];
+		BODY_00_15( 2,E,T,A,B,C,D,X( 2));	X( 4) = W[ 4];
+		BODY_00_15( 3,D,E,T,A,B,C,X( 3));	X( 5) = W[ 5];
+		BODY_00_15( 4,C,D,E,T,A,B,X( 4));	X( 6) = W[ 6];
+		BODY_00_15( 5,B,C,D,E,T,A,X( 5));	X( 7) = W[ 7];
+		BODY_00_15( 6,A,B,C,D,E,T,X( 6));	X( 8) = W[ 8];
+		BODY_00_15( 7,T,A,B,C,D,E,X( 7));	X( 9) = W[ 9];
+		BODY_00_15( 8,E,T,A,B,C,D,X( 8));	X(10) = W[10];
+		BODY_00_15( 9,D,E,T,A,B,C,X( 9));	X(11) = W[11];
+		BODY_00_15(10,C,D,E,T,A,B,X(10));	X(12) = W[12];
+		BODY_00_15(11,B,C,D,E,T,A,X(11));	X(13) = W[13];
+		BODY_00_15(12,A,B,C,D,E,T,X(12));	X(14) = W[14];
+		BODY_00_15(13,T,A,B,C,D,E,X(13));	X(15) = W[15];
+		BODY_00_15(14,E,T,A,B,C,D,X(14));
+		BODY_00_15(15,D,E,T,A,B,C,X(15));
+
+		data += SHA_CBLOCK;
+		}
+	else
+		{
+		(void)HOST_c2l(data,l); X( 0)=l;	(void)HOST_c2l(data,l); X( 1)=l;
+		BODY_00_15( 0,A,B,C,D,E,T,X( 0));	(void)HOST_c2l(data,l); X( 2)=l;
+		BODY_00_15( 1,T,A,B,C,D,E,X( 1));	(void)HOST_c2l(data,l); X( 3)=l;
+		BODY_00_15( 2,E,T,A,B,C,D,X( 2));	(void)HOST_c2l(data,l); X( 4)=l;
+		BODY_00_15( 3,D,E,T,A,B,C,X( 3));	(void)HOST_c2l(data,l); X( 5)=l;
+		BODY_00_15( 4,C,D,E,T,A,B,X( 4));	(void)HOST_c2l(data,l); X( 6)=l;
+		BODY_00_15( 5,B,C,D,E,T,A,X( 5));	(void)HOST_c2l(data,l); X( 7)=l;
+		BODY_00_15( 6,A,B,C,D,E,T,X( 6));	(void)HOST_c2l(data,l); X( 8)=l;
+		BODY_00_15( 7,T,A,B,C,D,E,X( 7));	(void)HOST_c2l(data,l); X( 9)=l;
+		BODY_00_15( 8,E,T,A,B,C,D,X( 8));	(void)HOST_c2l(data,l); X(10)=l;
+		BODY_00_15( 9,D,E,T,A,B,C,X( 9));	(void)HOST_c2l(data,l); X(11)=l;
+		BODY_00_15(10,C,D,E,T,A,B,X(10));	(void)HOST_c2l(data,l); X(12)=l;
+		BODY_00_15(11,B,C,D,E,T,A,X(11));	(void)HOST_c2l(data,l); X(13)=l;
+		BODY_00_15(12,A,B,C,D,E,T,X(12));	(void)HOST_c2l(data,l); X(14)=l;
+		BODY_00_15(13,T,A,B,C,D,E,X(13));	(void)HOST_c2l(data,l); X(15)=l;
+		BODY_00_15(14,E,T,A,B,C,D,X(14));
+		BODY_00_15(15,D,E,T,A,B,C,X(15));
+		}
+
+	BODY_16_19(16,C,D,E,T,A,B,X( 0),X( 0),X( 2),X( 8),X(13));
+	BODY_16_19(17,B,C,D,E,T,A,X( 1),X( 1),X( 3),X( 9),X(14));
+	BODY_16_19(18,A,B,C,D,E,T,X( 2),X( 2),X( 4),X(10),X(15));
+	BODY_16_19(19,T,A,B,C,D,E,X( 3),X( 3),X( 5),X(11),X( 0));
+
+	BODY_20_31(20,E,T,A,B,C,D,X( 4),X( 4),X( 6),X(12),X( 1));
+	BODY_20_31(21,D,E,T,A,B,C,X( 5),X( 5),X( 7),X(13),X( 2));
+	BODY_20_31(22,C,D,E,T,A,B,X( 6),X( 6),X( 8),X(14),X( 3));
+	BODY_20_31(23,B,C,D,E,T,A,X( 7),X( 7),X( 9),X(15),X( 4));
+	BODY_20_31(24,A,B,C,D,E,T,X( 8),X( 8),X(10),X( 0),X( 5));
+	BODY_20_31(25,T,A,B,C,D,E,X( 9),X( 9),X(11),X( 1),X( 6));
+	BODY_20_31(26,E,T,A,B,C,D,X(10),X(10),X(12),X( 2),X( 7));
+	BODY_20_31(27,D,E,T,A,B,C,X(11),X(11),X(13),X( 3),X( 8));
+	BODY_20_31(28,C,D,E,T,A,B,X(12),X(12),X(14),X( 4),X( 9));
+	BODY_20_31(29,B,C,D,E,T,A,X(13),X(13),X(15),X( 5),X(10));
+	BODY_20_31(30,A,B,C,D,E,T,X(14),X(14),X( 0),X( 6),X(11));
+	BODY_20_31(31,T,A,B,C,D,E,X(15),X(15),X( 1),X( 7),X(12));
+
+	BODY_32_39(32,E,T,A,B,C,D,X( 0),X( 2),X( 8),X(13));
+	BODY_32_39(33,D,E,T,A,B,C,X( 1),X( 3),X( 9),X(14));
+	BODY_32_39(34,C,D,E,T,A,B,X( 2),X( 4),X(10),X(15));
+	BODY_32_39(35,B,C,D,E,T,A,X( 3),X( 5),X(11),X( 0));
+	BODY_32_39(36,A,B,C,D,E,T,X( 4),X( 6),X(12),X( 1));
+	BODY_32_39(37,T,A,B,C,D,E,X( 5),X( 7),X(13),X( 2));
+	BODY_32_39(38,E,T,A,B,C,D,X( 6),X( 8),X(14),X( 3));
+	BODY_32_39(39,D,E,T,A,B,C,X( 7),X( 9),X(15),X( 4));
+
+	BODY_40_59(40,C,D,E,T,A,B,X( 8),X(10),X( 0),X( 5));
+	BODY_40_59(41,B,C,D,E,T,A,X( 9),X(11),X( 1),X( 6));
+	BODY_40_59(42,A,B,C,D,E,T,X(10),X(12),X( 2),X( 7));
+	BODY_40_59(43,T,A,B,C,D,E,X(11),X(13),X( 3),X( 8));
+	BODY_40_59(44,E,T,A,B,C,D,X(12),X(14),X( 4),X( 9));
+	BODY_40_59(45,D,E,T,A,B,C,X(13),X(15),X( 5),X(10));
+	BODY_40_59(46,C,D,E,T,A,B,X(14),X( 0),X( 6),X(11));
+	BODY_40_59(47,B,C,D,E,T,A,X(15),X( 1),X( 7),X(12));
+	BODY_40_59(48,A,B,C,D,E,T,X( 0),X( 2),X( 8),X(13));
+	BODY_40_59(49,T,A,B,C,D,E,X( 1),X( 3),X( 9),X(14));
+	BODY_40_59(50,E,T,A,B,C,D,X( 2),X( 4),X(10),X(15));
+	BODY_40_59(51,D,E,T,A,B,C,X( 3),X( 5),X(11),X( 0));
+	BODY_40_59(52,C,D,E,T,A,B,X( 4),X( 6),X(12),X( 1));
+	BODY_40_59(53,B,C,D,E,T,A,X( 5),X( 7),X(13),X( 2));
+	BODY_40_59(54,A,B,C,D,E,T,X( 6),X( 8),X(14),X( 3));
+	BODY_40_59(55,T,A,B,C,D,E,X( 7),X( 9),X(15),X( 4));
+	BODY_40_59(56,E,T,A,B,C,D,X( 8),X(10),X( 0),X( 5));
+	BODY_40_59(57,D,E,T,A,B,C,X( 9),X(11),X( 1),X( 6));
+	BODY_40_59(58,C,D,E,T,A,B,X(10),X(12),X( 2),X( 7));
+	BODY_40_59(59,B,C,D,E,T,A,X(11),X(13),X( 3),X( 8));
+
+	BODY_60_79(60,A,B,C,D,E,T,X(12),X(14),X( 4),X( 9));
+	BODY_60_79(61,T,A,B,C,D,E,X(13),X(15),X( 5),X(10));
+	BODY_60_79(62,E,T,A,B,C,D,X(14),X( 0),X( 6),X(11));
+	BODY_60_79(63,D,E,T,A,B,C,X(15),X( 1),X( 7),X(12));
+	BODY_60_79(64,C,D,E,T,A,B,X( 0),X( 2),X( 8),X(13));
+	BODY_60_79(65,B,C,D,E,T,A,X( 1),X( 3),X( 9),X(14));
+	BODY_60_79(66,A,B,C,D,E,T,X( 2),X( 4),X(10),X(15));
+	BODY_60_79(67,T,A,B,C,D,E,X( 3),X( 5),X(11),X( 0));
+	BODY_60_79(68,E,T,A,B,C,D,X( 4),X( 6),X(12),X( 1));
+	BODY_60_79(69,D,E,T,A,B,C,X( 5),X( 7),X(13),X( 2));
+	BODY_60_79(70,C,D,E,T,A,B,X( 6),X( 8),X(14),X( 3));
+	BODY_60_79(71,B,C,D,E,T,A,X( 7),X( 9),X(15),X( 4));
+	BODY_60_79(72,A,B,C,D,E,T,X( 8),X(10),X( 0),X( 5));
+	BODY_60_79(73,T,A,B,C,D,E,X( 9),X(11),X( 1),X( 6));
+	BODY_60_79(74,E,T,A,B,C,D,X(10),X(12),X( 2),X( 7));
+	BODY_60_79(75,D,E,T,A,B,C,X(11),X(13),X( 3),X( 8));
+	BODY_60_79(76,C,D,E,T,A,B,X(12),X(14),X( 4),X( 9));
+	BODY_60_79(77,B,C,D,E,T,A,X(13),X(15),X( 5),X(10));
+	BODY_60_79(78,A,B,C,D,E,T,X(14),X( 0),X( 6),X(11));
+	BODY_60_79(79,T,A,B,C,D,E,X(15),X( 1),X( 7),X(12));
+	
+	c->h0=(c->h0+E)&0xffffffffL; 
+	c->h1=(c->h1+T)&0xffffffffL;
+	c->h2=(c->h2+A)&0xffffffffL;
+	c->h3=(c->h3+B)&0xffffffffL;
+	c->h4=(c->h4+C)&0xffffffffL;
+
+	if (--num == 0) break;
+
+	A=c->h0;
+	B=c->h1;
+	C=c->h2;
+	D=c->h3;
+	E=c->h4;
+
+			}
+	}
+#endif
+
+#else	/* OPENSSL_SMALL_FOOTPRINT */
+
+#define BODY_00_15(xi)		 do {	\
+	T=E+K_00_19+F_00_19(B,C,D);	\
+	E=D, D=C, C=ROTATE(B,30), B=A;	\
+	A=ROTATE(A,5)+T+xi;	    } while(0)
+
+#define BODY_16_19(xa,xb,xc,xd)	 do {	\
+	Xupdate(T,xa,xa,xb,xc,xd);	\
+	T+=E+K_00_19+F_00_19(B,C,D);	\
+	E=D, D=C, C=ROTATE(B,30), B=A;	\
+	A=ROTATE(A,5)+T;	    } while(0)
+
+#define BODY_20_39(xa,xb,xc,xd)	 do {	\
+	Xupdate(T,xa,xa,xb,xc,xd);	\
+	T+=E+K_20_39+F_20_39(B,C,D);	\
+	E=D, D=C, C=ROTATE(B,30), B=A;	\
+	A=ROTATE(A,5)+T;	    } while(0)
+
+#define BODY_40_59(xa,xb,xc,xd)	 do {	\
+	Xupdate(T,xa,xa,xb,xc,xd);	\
+	T+=E+K_40_59+F_40_59(B,C,D);	\
+	E=D, D=C, C=ROTATE(B,30), B=A;	\
+	A=ROTATE(A,5)+T;	    } while(0)
+
+#define BODY_60_79(xa,xb,xc,xd)	 do {	\
+	Xupdate(T,xa,xa,xb,xc,xd);	\
+	T=E+K_60_79+F_60_79(B,C,D);	\
+	E=D, D=C, C=ROTATE(B,30), B=A;	\
+	A=ROTATE(A,5)+T+xa;	    } while(0)
+
+#if !defined(SHA_1) || !defined(SHA1_ASM)
+static void HASH_BLOCK_DATA_ORDER (SHA_CTX *c, const void *p, size_t num)
+	{
+	const unsigned char *data=p;
+	register unsigned MD32_REG_T A,B,C,D,E,T,l;
+	int i;
+	SHA_LONG	X[16];
+
+	A=c->h0;
+	B=c->h1;
+	C=c->h2;
+	D=c->h3;
+	E=c->h4;
+
+	for (;;)
+		{
+	for (i=0;i<16;i++)
+	{ HOST_c2l(data,l); X[i]=l; BODY_00_15(X[i]); }
+	for (i=0;i<4;i++)
+	{ BODY_16_19(X[i],       X[i+2],      X[i+8],     X[(i+13)&15]); }
+	for (;i<24;i++)
+	{ BODY_20_39(X[i&15],    X[(i+2)&15], X[(i+8)&15],X[(i+13)&15]); }
+	for (i=0;i<20;i++)
+	{ BODY_40_59(X[(i+8)&15],X[(i+10)&15],X[i&15],    X[(i+5)&15]);  }
+	for (i=4;i<24;i++)
+	{ BODY_60_79(X[(i+8)&15],X[(i+10)&15],X[i&15],    X[(i+5)&15]);  }
+
+	c->h0=(c->h0+A)&0xffffffffL; 
+	c->h1=(c->h1+B)&0xffffffffL;
+	c->h2=(c->h2+C)&0xffffffffL;
+	c->h3=(c->h3+D)&0xffffffffL;
+	c->h4=(c->h4+E)&0xffffffffL;
+
+	if (--num == 0) break;
+
+	A=c->h0;
+	B=c->h1;
+	C=c->h2;
+	D=c->h3;
+	E=c->h4;
+
+		}
+	}
+#endif
+
+#endif
diff --git a/drivers/builtin_openssl2/crypto/sha/sha_one.c b/drivers/builtin_openssl2/crypto/sha/sha_one.c
new file mode 100644
index 0000000000..3bae623ce8
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/sha_one.c
@@ -0,0 +1,78 @@
+/* crypto/sha/sha_one.c */
+/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
+ * All rights reserved.
+ *
+ * This package is an SSL implementation written
+ * by Eric Young (eay@cryptsoft.com).
+ * The implementation was written so as to conform with Netscapes SSL.
+ * 
+ * This library is free for commercial and non-commercial use as long as
+ * the following conditions are aheared to.  The following conditions
+ * apply to all code found in this distribution, be it the RC4, RSA,
+ * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
+ * included with this distribution is covered by the same copyright terms
+ * except that the holder is Tim Hudson (tjh@cryptsoft.com).
+ * 
+ * Copyright remains Eric Young's, and as such any Copyright notices in
+ * the code are not to be removed.
+ * If this package is used in a product, Eric Young should be given attribution
+ * as the author of the parts of the library used.
+ * This can be in the form of a textual message at program startup or
+ * in documentation (online or textual) provided with the package.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. All advertising materials mentioning features or use of this software
+ *    must display the following acknowledgement:
+ *    "This product includes cryptographic software written by
+ *     Eric Young (eay@cryptsoft.com)"
+ *    The word 'cryptographic' can be left out if the rouines from the library
+ *    being used are not cryptographic related :-).
+ * 4. If you include any Windows specific code (or a derivative thereof) from 
+ *    the apps directory (application code) you must include an acknowledgement:
+ *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
+ * 
+ * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ * 
+ * The licence and distribution terms for any publically available version or
+ * derivative of this code cannot be changed.  i.e. this code cannot simply be
+ * copied and put under another distribution licence
+ * [including the GNU Public Licence.]
+ */
+
+#include <stdio.h>
+#include <string.h>
+#include <openssl/sha.h>
+#include <openssl/crypto.h>
+
+#ifndef OPENSSL_NO_SHA0
+unsigned char *SHA(const unsigned char *d, size_t n, unsigned char *md)
+	{
+	SHA_CTX c;
+	static unsigned char m[SHA_DIGEST_LENGTH];
+
+	if (md == NULL) md=m;
+	if (!SHA_Init(&c))
+		return NULL;
+	SHA_Update(&c,d,n);
+	SHA_Final(md,&c);
+	OPENSSL_cleanse(&c,sizeof(c));
+	return(md);
+	}
+#endif
diff --git a/drivers/builtin_openssl2/crypto/sha/shatest.c b/drivers/builtin_openssl2/crypto/sha/shatest.c
new file mode 100644
index 0000000000..27614646d1
--- /dev/null
+++ b/drivers/builtin_openssl2/crypto/sha/shatest.c
@@ -0,0 +1,178 @@
+/* crypto/sha/shatest.c */
+/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
+ * All rights reserved.
+ *
+ * This package is an SSL implementation written
+ * by Eric Young (eay@cryptsoft.com).
+ * The implementation was written so as to conform with Netscapes SSL.
+ * 
+ * This library is free for commercial and non-commercial use as long as
+ * the following conditions are aheared to.  The following conditions
+ * apply to all code found in this distribution, be it the RC4, RSA,
+ * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
+ * included with this distribution is covered by the same copyright terms
+ * except that the holder is Tim Hudson (tjh@cryptsoft.com).
+ * 
+ * Copyright remains Eric Young's, and as such any Copyright notices in
+ * the code are not to be removed.
+ * If this package is used in a product, Eric Young should be given attribution
+ * as the author of the parts of the library used.
+ * This can be in the form of a textual message at program startup or
+ * in documentation (online or textual) provided with the package.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. All advertising materials mentioning features or use of this software
+ *    must display the following acknowledgement:
+ *    "This product includes cryptographic software written by
+ *     Eric Young (eay@cryptsoft.com)"
+ *    The word 'cryptographic' can be left out if the rouines from the library
+ *    being used are not cryptographic related :-).
+ * 4. If you include any Windows specific code (or a derivative thereof) from 
+ *    the apps directory (application code) you must include an acknowledgement:
+ *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
+ * 
+ * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ * 
+ * The licence and distribution terms for any publically available version or
+ * derivative of this code cannot be changed.  i.e. this code cannot simply be
+ * copied and put under another distribution licence
+ * [including the GNU Public Licence.]
+ */
+
+#include <stdio.h>
+#include <string.h>
+#include <stdlib.h>
+
+#include "../e_os.h"
+
+#if defined(OPENSSL_NO_SHA) || defined(OPENSSL_NO_SHA0)
+int main(int argc, char *argv[])
+{
+    printf("No SHA0 support\n");
+    return(0);
+}
+#else
+#include <openssl/evp.h>
+#include <openssl/sha.h>
+
+#ifdef CHARSET_EBCDIC
+#include <openssl/ebcdic.h>
+#endif
+
+#define SHA_0 /* FIPS 180 */
+#undef  SHA_1 /* FIPS 180-1 */
+
+static char *test[]={
+	"abc",
+	"abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq",
+	NULL,
+	};
+
+#ifdef SHA_0
+static char *ret[]={
+	"0164b8a914cd2a5e74c4f7ff082c4d97f1edf880",
+	"d2516ee1acfa5baf33dfc1c471e438449ef134c8",
+	};
+static char *bigret=
+	"3232affa48628a26653b5aaa44541fd90d690603";
+#endif
+#ifdef SHA_1
+static char *ret[]={
+	"a9993e364706816aba3e25717850c26c9cd0d89d",
+	"84983e441c3bd26ebaae4aa1f95129e5e54670f1",
+	};
+static char *bigret=
+	"34aa973cd4c4daa4f61eeb2bdbad27316534016f";
+#endif
+
+static char *pt(unsigned char *md);
+int main(int argc, char *argv[])
+	{
+	int i,err=0;
+	char **P,**R;
+	static unsigned char buf[1000];
+	char *p,*r;
+	EVP_MD_CTX c;
+	unsigned char md[SHA_DIGEST_LENGTH];
+
+#ifdef CHARSET_EBCDIC
+	ebcdic2ascii(test[0], test[0], strlen(test[0]));
+	ebcdic2ascii(test[1], test[1], strlen(test[1]));
+#endif
+
+	EVP_MD_CTX_init(&c);
+	P=test;
+	R=ret;
+	i=1;
+	while (*P != NULL)
+		{
+		EVP_Digest(*P,strlen(*P),md,NULL,EVP_sha(), NULL);
+		p=pt(md);
+		if (strcmp(p,*R) != 0)
+			{
+			printf("error calculating SHA on '%s'\n",*P);
+			printf("got %s instead of %s\n",p,*R);
+			err++;
+			}
+		else
+			printf("test %d ok\n",i);
+		i++;
+		R++;
+		P++;
+		}
+
+	memset(buf,'a',1000);
+#ifdef CHARSET_EBCDIC
+	ebcdic2ascii(buf, buf, 1000);
+#endif /*CHARSET_EBCDIC*/
+	EVP_DigestInit_ex(&c,EVP_sha(), NULL);
+	for (i=0; i<1000; i++)
+		EVP_DigestUpdate(&c,buf,1000);
+	EVP_DigestFinal_ex(&c,md,NULL);
+	p=pt(md);
+
+	r=bigret;
+	if (strcmp(p,r) != 0)
+		{
+		printf("error calculating SHA on '%s'\n",p);
+		printf("got %s instead of %s\n",p,r);
+		err++;
+		}
+	else
+		printf("test 3 ok\n");
+
+#ifdef OPENSSL_SYS_NETWARE
+    if (err) printf("ERROR: %d\n", err);
+#endif
+	EVP_MD_CTX_cleanup(&c);
+	EXIT(err);
+	return(0);
+	}
+
+static char *pt(unsigned char *md)
+	{
+	int i;
+	static char buf[80];
+
+	for (i=0; i<SHA_DIGEST_LENGTH; i++)
+		sprintf(&(buf[i*2]),"%02x",md[i]);
+	return(buf);
+	}
+#endif