This commit was generated by cvs2git to track changes on a CVS vendor

branch.

15 files changed, 3356 insertions, 24 deletions

diff --git a/src/lib/libcrypto/bn/asm/ppc.pl b/src/lib/libcrypto/bn/asm/ppc.pl
new file mode 100644
index 0000000000..307c7ccb35
--- /dev/null
+++ b/src/lib/libcrypto/bn/asm/ppc.pl
@@ -0,0 +1,2081 @@
+#!/usr/bin/env perl
+#
+# Implemented as a Perl wrapper as we want to support several different
+# architectures with single file. We pick up the target based on the
+# file name we are asked to generate.
+#
+# It should be noted though that this perl code is nothing like
+# <openssl>/crypto/perlasm/x86*. In this case perl is used pretty much
+# as pre-processor to cover for platform differences in name decoration,
+# linker tables, 32-/64-bit instruction sets...
+#
+# As you might know there're several PowerPC ABI in use. Most notably
+# Linux and AIX use different 32-bit ABIs. Good news are that these ABIs
+# are similar enough to implement leaf(!) functions, which would be ABI
+# neutral. And that's what you find here: ABI neutral leaf functions.
+# In case you wonder what that is...
+#
+#       AIX performance
+#
+#       MEASUREMENTS WITH cc ON a 200 MhZ PowerPC 604e.
+#
+#       The following is the performance of 32-bit compiler
+#       generated code:
+#
+#       OpenSSL 0.9.6c 21 dec 2001
+#       built on: Tue Jun 11 11:06:51 EDT 2002
+#       options:bn(64,32) ...
+#compiler: cc -DTHREADS  -DAIX -DB_ENDIAN -DBN_LLONG -O3
+#                  sign    verify    sign/s verify/s
+#rsa  512 bits   0.0098s   0.0009s    102.0   1170.6
+#rsa 1024 bits   0.0507s   0.0026s     19.7    387.5
+#rsa 2048 bits   0.3036s   0.0085s      3.3    117.1
+#rsa 4096 bits   2.0040s   0.0299s      0.5     33.4
+#dsa  512 bits   0.0087s   0.0106s    114.3     94.5
+#dsa 1024 bits   0.0256s   0.0313s     39.0     32.0    
+#
+#       Same bechmark with this assembler code:
+#
+#rsa  512 bits   0.0056s   0.0005s    178.6   2049.2
+#rsa 1024 bits   0.0283s   0.0015s     35.3    674.1
+#rsa 2048 bits   0.1744s   0.0050s      5.7    201.2
+#rsa 4096 bits   1.1644s   0.0179s      0.9     55.7
+#dsa  512 bits   0.0052s   0.0062s    191.6    162.0
+#dsa 1024 bits   0.0149s   0.0180s     67.0     55.5
+#
+#       Number of operations increases by at almost 75%
+#
+#       Here are performance numbers for 64-bit compiler
+#       generated code:
+#
+#       OpenSSL 0.9.6g [engine] 9 Aug 2002
+#       built on: Fri Apr 18 16:59:20 EDT 2003
+#       options:bn(64,64) ...
+#       compiler: cc -DTHREADS -D_REENTRANT -q64 -DB_ENDIAN -O3
+#                  sign    verify    sign/s verify/s
+#rsa  512 bits   0.0028s   0.0003s    357.1   3844.4
+#rsa 1024 bits   0.0148s   0.0008s     67.5   1239.7
+#rsa 2048 bits   0.0963s   0.0028s     10.4    353.0
+#rsa 4096 bits   0.6538s   0.0102s      1.5     98.1
+#dsa  512 bits   0.0026s   0.0032s    382.5    313.7
+#dsa 1024 bits   0.0081s   0.0099s    122.8    100.6
+#
+#       Same benchmark with this assembler code:
+#
+#rsa  512 bits   0.0020s   0.0002s    510.4   6273.7
+#rsa 1024 bits   0.0088s   0.0005s    114.1   2128.3
+#rsa 2048 bits   0.0540s   0.0016s     18.5    622.5
+#rsa 4096 bits   0.3700s   0.0058s      2.7    171.0
+#dsa  512 bits   0.0016s   0.0020s    610.7    507.1
+#dsa 1024 bits   0.0047s   0.0058s    212.5    173.2
+#       
+#       Again, performance increases by at about 75%
+#
+#       Mac OS X, Apple G5 1.8GHz (Note this is 32 bit code)
+#       OpenSSL 0.9.7c 30 Sep 2003
+#
+#       Original code.
+#
+#rsa  512 bits   0.0011s   0.0001s    906.1  11012.5
+#rsa 1024 bits   0.0060s   0.0003s    166.6   3363.1
+#rsa 2048 bits   0.0370s   0.0010s     27.1    982.4
+#rsa 4096 bits   0.2426s   0.0036s      4.1    280.4
+#dsa  512 bits   0.0010s   0.0012s   1038.1    841.5
+#dsa 1024 bits   0.0030s   0.0037s    329.6    269.7
+#dsa 2048 bits   0.0101s   0.0127s     98.9     78.6
+#
+#       Same benchmark with this assembler code:
+#
+#rsa  512 bits   0.0007s   0.0001s   1416.2  16645.9
+#rsa 1024 bits   0.0036s   0.0002s    274.4   5380.6
+#rsa 2048 bits   0.0222s   0.0006s     45.1   1589.5
+#rsa 4096 bits   0.1469s   0.0022s      6.8    449.6
+#dsa  512 bits   0.0006s   0.0007s   1664.2   1376.2
+#dsa 1024 bits   0.0018s   0.0023s    545.0    442.2
+#dsa 2048 bits   0.0061s   0.0075s    163.5    132.8
+#
+#        Performance increase of ~60%
+#
+#       If you have comments or suggestions to improve code send
+#       me a note at schari@us.ibm.com
+#
+
+$opf = shift;
+
+if ($opf =~ /32\.s/) {
+        $BITS=  32;
+        $BNSZ=  $BITS/8;
+        $ISA=   "\"ppc\"";
+
+        $LD=    "lwz";          # load
+        $LDU=   "lwzu";         # load and update
+        $ST=    "stw";          # store
+        $STU=   "stwu";         # store and update
+        $UMULL= "mullw";        # unsigned multiply low
+        $UMULH= "mulhwu";       # unsigned multiply high
+        $UDIV=  "divwu";        # unsigned divide
+        $UCMPI= "cmplwi";       # unsigned compare with immediate
+        $UCMP=  "cmplw";        # unsigned compare
+        $COUNTZ="cntlzw";       # count leading zeros
+        $SHL=   "slw";          # shift left
+        $SHR=   "srw";          # unsigned shift right
+        $SHRI=  "srwi";         # unsigned shift right by immediate     
+        $SHLI=  "slwi";         # shift left by immediate
+        $CLRU=  "clrlwi";       # clear upper bits
+        $INSR=  "insrwi";       # insert right
+        $ROTL=  "rotlwi";       # rotate left by immediate
+} elsif ($opf =~ /64\.s/) {
+        $BITS=  64;
+        $BNSZ=  $BITS/8;
+        $ISA=   "\"ppc64\"";
+
+        # same as above, but 64-bit mnemonics...
+        $LD=    "ld";           # load
+        $LDU=   "ldu";          # load and update
+        $ST=    "std";          # store
+        $STU=   "stdu";         # store and update
+        $UMULL= "mulld";        # unsigned multiply low
+        $UMULH= "mulhdu";       # unsigned multiply high
+        $UDIV=  "divdu";        # unsigned divide
+        $UCMPI= "cmpldi";       # unsigned compare with immediate
+        $UCMP=  "cmpld";        # unsigned compare
+        $COUNTZ="cntlzd";       # count leading zeros
+        $SHL=   "sld";          # shift left
+        $SHR=   "srd";          # unsigned shift right
+        $SHRI=  "srdi";         # unsigned shift right by immediate     
+        $SHLI=  "sldi";         # shift left by immediate
+        $CLRU=  "clrldi";       # clear upper bits
+        $INSR=  "insrdi";       # insert right 
+        $ROTL=  "rotldi";       # rotate left by immediate
+} else { die "nonsense $opf"; }
+
+( defined shift || open STDOUT,">$opf" ) || die "can't open $opf: $!";
+
+# function entry points from the AIX code
+#
+# There are other, more elegant, ways to handle this. We (IBM) chose
+# this approach as it plays well with scripts we run to 'namespace'
+# OpenSSL .i.e. we add a prefix to all the public symbols so we can
+# co-exist in the same process with other implementations of OpenSSL.
+# 'cleverer' ways of doing these substitutions tend to hide data we
+# need to be obvious.
+#
+my @items = ("bn_sqr_comba4",
+             "bn_sqr_comba8",
+             "bn_mul_comba4",
+             "bn_mul_comba8",
+             "bn_sub_words",
+             "bn_add_words",
+             "bn_div_words",
+             "bn_sqr_words",
+             "bn_mul_words",
+             "bn_mul_add_words");
+
+if    ($opf =~ /linux/) {  do_linux();  }
+elsif ($opf =~ /aix/)   {  do_aix();    }
+elsif ($opf =~ /osx/)   {  do_osx();    }
+else                    {  do_bsd();    }
+
+sub do_linux {
+    $d=&data();
+
+    if ($BITS==64) {
+      foreach $t (@items) {
+        $d =~ s/\.$t:/\
+\t.section\t".opd","aw"\
+\t.align\t3\
+\t.globl\t$t\
+$t:\
+\t.quad\t.$t,.TOC.\@tocbase,0\
+\t.size\t$t,24\
+\t.previous\n\
+\t.type\t.$t,\@function\
+\t.globl\t.$t\
+.$t:/g;
+      }
+    }
+    else {
+      foreach $t (@items) {
+        $d=~s/\.$t/$t/g;
+      }
+    }
+    # hide internal labels to avoid pollution of name table...
+    $d=~s/Lppcasm_/.Lppcasm_/gm;
+    print $d;
+}
+
+sub do_aix {
+    # AIX assembler is smart enough to please the linker without
+    # making us do something special...
+    print &data();
+}
+
+# MacOSX 32 bit
+sub do_osx {
+    $d=&data();
+    # Change the bn symbol prefix from '.' to '_'
+    foreach $t (@items) {
+      $d=~s/\.$t/_$t/g;
+    }
+    # Change .machine to something OS X asm will accept
+    $d=~s/\.machine.*/.text/g;
+    $d=~s/\#/;/g; # change comment from '#' to ';'
+    print $d;
+}
+
+# BSD (Untested)
+sub do_bsd {
+    $d=&data();
+    foreach $t (@items) {
+      $d=~s/\.$t/_$t/g;
+    }
+    print $d;
+}
+
+sub data {
+        local($data)=<<EOF;
+#--------------------------------------------------------------------
+#
+#
+#
+#
+#       File:           ppc32.s
+#
+#       Created by:     Suresh Chari
+#                       IBM Thomas J. Watson Research Library
+#                       Hawthorne, NY
+#
+#
+#       Description:    Optimized assembly routines for OpenSSL crypto
+#                       on the 32 bitPowerPC platform.
+#
+#
+#       Version History
+#
+#       2. Fixed bn_add,bn_sub and bn_div_words, added comments,
+#          cleaned up code. Also made a single version which can
+#          be used for both the AIX and Linux compilers. See NOTE
+#          below.
+#                               12/05/03                Suresh Chari
+#                       (with lots of help from)        Andy Polyakov
+##      
+#       1. Initial version      10/20/02                Suresh Chari
+#
+#
+#       The following file works for the xlc,cc
+#       and gcc compilers.
+#
+#       NOTE:   To get the file to link correctly with the gcc compiler
+#               you have to change the names of the routines and remove
+#               the first .(dot) character. This should automatically
+#               be done in the build process.
+#
+#       Hand optimized assembly code for the following routines
+#       
+#       bn_sqr_comba4
+#       bn_sqr_comba8
+#       bn_mul_comba4
+#       bn_mul_comba8
+#       bn_sub_words
+#       bn_add_words
+#       bn_div_words
+#       bn_sqr_words
+#       bn_mul_words
+#       bn_mul_add_words
+#
+#       NOTE:   It is possible to optimize this code more for
+#       specific PowerPC or Power architectures. On the Northstar
+#       architecture the optimizations in this file do
+#        NOT provide much improvement.
+#
+#       If you have comments or suggestions to improve code send
+#       me a note at schari\@us.ibm.com
+#
+#--------------------------------------------------------------------------
+#
+#       Defines to be used in the assembly code.
+#       
+.set r0,0       # we use it as storage for value of 0
+.set SP,1       # preserved
+.set RTOC,2     # preserved 
+.set r3,3       # 1st argument/return value
+.set r4,4       # 2nd argument/volatile register
+.set r5,5       # 3rd argument/volatile register
+.set r6,6       # ...
+.set r7,7
+.set r8,8
+.set r9,9
+.set r10,10
+.set r11,11
+.set r12,12
+.set r13,13     # not used, nor any other "below" it...
+
+.set BO_IF_NOT,4
+.set BO_IF,12
+.set BO_dCTR_NZERO,16
+.set BO_dCTR_ZERO,18
+.set BO_ALWAYS,20
+.set CR0_LT,0;
+.set CR0_GT,1;
+.set CR0_EQ,2
+.set CR1_FX,4;
+.set CR1_FEX,5;
+.set CR1_VX,6
+.set LR,8
+
+#       Declare function names to be global
+#       NOTE:   For gcc these names MUST be changed to remove
+#               the first . i.e. for example change ".bn_sqr_comba4"
+#               to "bn_sqr_comba4". This should be automatically done
+#               in the build.
+        
+        .globl  .bn_sqr_comba4
+        .globl  .bn_sqr_comba8
+        .globl  .bn_mul_comba4
+        .globl  .bn_mul_comba8
+        .globl  .bn_sub_words
+        .globl  .bn_add_words
+        .globl  .bn_div_words
+        .globl  .bn_sqr_words
+        .globl  .bn_mul_words
+        .globl  .bn_mul_add_words
+        
+# .text section
+        
+        .machine        $ISA
+
+#
+#       NOTE:   The following label name should be changed to
+#               "bn_sqr_comba4" i.e. remove the first dot
+#               for the gcc compiler. This should be automatically
+#               done in the build
+#
+
+.align  4
+.bn_sqr_comba4:
+#
+# Optimized version of bn_sqr_comba4.
+#
+# void bn_sqr_comba4(BN_ULONG *r, BN_ULONG *a)
+# r3 contains r
+# r4 contains a
+#
+# Freely use registers r5,r6,r7,r8,r9,r10,r11 as follows:       
+# 
+# r5,r6 are the two BN_ULONGs being multiplied.
+# r7,r8 are the results of the 32x32 giving 64 bit multiply.
+# r9,r10, r11 are the equivalents of c1,c2, c3.
+# Here's the assembly
+#
+#
+        xor             r0,r0,r0                # set r0 = 0. Used in the addze
+                                                # instructions below
+        
+                                                #sqr_add_c(a,0,c1,c2,c3)
+        $LD             r5,`0*$BNSZ`(r4)                
+        $UMULL          r9,r5,r5                
+        $UMULH          r10,r5,r5               #in first iteration. No need
+                                                #to add since c1=c2=c3=0.
+                                                # Note c3(r11) is NOT set to 0
+                                                # but will be.
+
+        $ST             r9,`0*$BNSZ`(r3)        # r[0]=c1;
+                                                # sqr_add_c2(a,1,0,c2,c3,c1);
+        $LD             r6,`1*$BNSZ`(r4)                
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+                                        
+        addc            r7,r7,r7                # compute (r7,r8)=2*(r7,r8)
+        adde            r8,r8,r8
+        addze           r9,r0                   # catch carry if any.
+                                                # r9= r0(=0) and carry 
+        
+        addc            r10,r7,r10              # now add to temp result.
+        addze           r11,r8                  # r8 added to r11 which is 0 
+        addze           r9,r9
+        
+        $ST             r10,`1*$BNSZ`(r3)       #r[1]=c2; 
+                                                #sqr_add_c(a,1,c3,c1,c2)
+        $UMULL          r7,r6,r6
+        $UMULH          r8,r6,r6
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r0
+                                                #sqr_add_c2(a,2,0,c3,c1,c2)
+        $LD             r6,`2*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        
+        addc            r7,r7,r7
+        adde            r8,r8,r8
+        addze           r10,r10
+        
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r10
+        $ST             r11,`2*$BNSZ`(r3)       #r[2]=c3 
+                                                #sqr_add_c2(a,3,0,c1,c2,c3);
+        $LD             r6,`3*$BNSZ`(r4)                
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        addc            r7,r7,r7
+        adde            r8,r8,r8
+        addze           r11,r0
+        
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r11
+                                                #sqr_add_c2(a,2,1,c1,c2,c3);
+        $LD             r5,`1*$BNSZ`(r4)
+        $LD             r6,`2*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        
+        addc            r7,r7,r7
+        adde            r8,r8,r8
+        addze           r11,r11
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r11
+        $ST             r9,`3*$BNSZ`(r3)        #r[3]=c1
+                                                #sqr_add_c(a,2,c2,c3,c1);
+        $UMULL          r7,r6,r6
+        $UMULH          r8,r6,r6
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r0
+                                                #sqr_add_c2(a,3,1,c2,c3,c1);
+        $LD             r6,`3*$BNSZ`(r4)                
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        addc            r7,r7,r7
+        adde            r8,r8,r8
+        addze           r9,r9
+        
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r9
+        $ST             r10,`4*$BNSZ`(r3)       #r[4]=c2
+                                                #sqr_add_c2(a,3,2,c3,c1,c2);
+        $LD             r5,`2*$BNSZ`(r4)                
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        addc            r7,r7,r7
+        adde            r8,r8,r8
+        addze           r10,r0
+        
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r10
+        $ST             r11,`5*$BNSZ`(r3)       #r[5] = c3
+                                                #sqr_add_c(a,3,c1,c2,c3);
+        $UMULL          r7,r6,r6                
+        $UMULH          r8,r6,r6
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+
+        $ST             r9,`6*$BNSZ`(r3)        #r[6]=c1
+        $ST             r10,`7*$BNSZ`(r3)       #r[7]=c2
+        bclr    BO_ALWAYS,CR0_LT
+        .long   0x00000000
+
+#
+#       NOTE:   The following label name should be changed to
+#               "bn_sqr_comba8" i.e. remove the first dot
+#               for the gcc compiler. This should be automatically
+#               done in the build
+#
+        
+.align  4
+.bn_sqr_comba8:
+#
+# This is an optimized version of the bn_sqr_comba8 routine.
+# Tightly uses the adde instruction
+#
+#
+# void bn_sqr_comba8(BN_ULONG *r, BN_ULONG *a)
+# r3 contains r
+# r4 contains a
+#
+# Freely use registers r5,r6,r7,r8,r9,r10,r11 as follows:       
+# 
+# r5,r6 are the two BN_ULONGs being multiplied.
+# r7,r8 are the results of the 32x32 giving 64 bit multiply.
+# r9,r10, r11 are the equivalents of c1,c2, c3.
+#
+# Possible optimization of loading all 8 longs of a into registers
+# doesnt provide any speedup
+# 
+
+        xor             r0,r0,r0                #set r0 = 0.Used in addze
+                                                #instructions below.
+
+                                                #sqr_add_c(a,0,c1,c2,c3);
+        $LD             r5,`0*$BNSZ`(r4)
+        $UMULL          r9,r5,r5                #1st iteration: no carries.
+        $UMULH          r10,r5,r5
+        $ST             r9,`0*$BNSZ`(r3)        # r[0]=c1;
+                                                #sqr_add_c2(a,1,0,c2,c3,c1);
+        $LD             r6,`1*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6        
+        
+        addc            r10,r7,r10              #add the two register number
+        adde            r11,r8,r0               # (r8,r7) to the three register
+        addze           r9,r0                   # number (r9,r11,r10).NOTE:r0=0
+        
+        addc            r10,r7,r10              #add the two register number
+        adde            r11,r8,r11              # (r8,r7) to the three register
+        addze           r9,r9                   # number (r9,r11,r10).
+        
+        $ST             r10,`1*$BNSZ`(r3)       # r[1]=c2
+                                
+                                                #sqr_add_c(a,1,c3,c1,c2);
+        $UMULL          r7,r6,r6
+        $UMULH          r8,r6,r6
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r0
+                                                #sqr_add_c2(a,2,0,c3,c1,c2);
+        $LD             r6,`2*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r10
+        
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r10
+        
+        $ST             r11,`2*$BNSZ`(r3)       #r[2]=c3
+                                                #sqr_add_c2(a,3,0,c1,c2,c3);
+        $LD             r6,`3*$BNSZ`(r4)        #r6 = a[3]. r5 is already a[0].
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r0
+        
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r11
+                                                #sqr_add_c2(a,2,1,c1,c2,c3);
+        $LD             r5,`1*$BNSZ`(r4)
+        $LD             r6,`2*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r11
+        
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r11
+        
+        $ST             r9,`3*$BNSZ`(r3)        #r[3]=c1;
+                                                #sqr_add_c(a,2,c2,c3,c1);
+        $UMULL          r7,r6,r6
+        $UMULH          r8,r6,r6
+        
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r0
+                                                #sqr_add_c2(a,3,1,c2,c3,c1);
+        $LD             r6,`3*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r9
+        
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r9
+                                                #sqr_add_c2(a,4,0,c2,c3,c1);
+        $LD             r5,`0*$BNSZ`(r4)
+        $LD             r6,`4*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r9
+        
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r9
+        $ST             r10,`4*$BNSZ`(r3)       #r[4]=c2;
+                                                #sqr_add_c2(a,5,0,c3,c1,c2);
+        $LD             r6,`5*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r0
+        
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r10
+                                                #sqr_add_c2(a,4,1,c3,c1,c2);
+        $LD             r5,`1*$BNSZ`(r4)
+        $LD             r6,`4*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r10
+        
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r10
+                                                #sqr_add_c2(a,3,2,c3,c1,c2);
+        $LD             r5,`2*$BNSZ`(r4)
+        $LD             r6,`3*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r10
+        
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r10
+        $ST             r11,`5*$BNSZ`(r3)       #r[5]=c3;
+                                                #sqr_add_c(a,3,c1,c2,c3);
+        $UMULL          r7,r6,r6
+        $UMULH          r8,r6,r6
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r0
+                                                #sqr_add_c2(a,4,2,c1,c2,c3);
+        $LD             r6,`4*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r11
+        
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r11
+                                                #sqr_add_c2(a,5,1,c1,c2,c3);
+        $LD             r5,`1*$BNSZ`(r4)
+        $LD             r6,`5*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r11
+        
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r11
+                                                #sqr_add_c2(a,6,0,c1,c2,c3);
+        $LD             r5,`0*$BNSZ`(r4)
+        $LD             r6,`6*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r11
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r11
+        $ST             r9,`6*$BNSZ`(r3)        #r[6]=c1;
+                                                #sqr_add_c2(a,7,0,c2,c3,c1);
+        $LD             r6,`7*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r0
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r9
+                                                #sqr_add_c2(a,6,1,c2,c3,c1);
+        $LD             r5,`1*$BNSZ`(r4)
+        $LD             r6,`6*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r9
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r9
+                                                #sqr_add_c2(a,5,2,c2,c3,c1);
+        $LD             r5,`2*$BNSZ`(r4)
+        $LD             r6,`5*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r9
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r9
+                                                #sqr_add_c2(a,4,3,c2,c3,c1);
+        $LD             r5,`3*$BNSZ`(r4)
+        $LD             r6,`4*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r9
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r9
+        $ST             r10,`7*$BNSZ`(r3)       #r[7]=c2;
+                                                #sqr_add_c(a,4,c3,c1,c2);
+        $UMULL          r7,r6,r6
+        $UMULH          r8,r6,r6
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r0
+                                                #sqr_add_c2(a,5,3,c3,c1,c2);
+        $LD             r6,`5*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r10
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r10
+                                                #sqr_add_c2(a,6,2,c3,c1,c2);
+        $LD             r5,`2*$BNSZ`(r4)
+        $LD             r6,`6*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r10
+        
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r10
+                                                #sqr_add_c2(a,7,1,c3,c1,c2);
+        $LD             r5,`1*$BNSZ`(r4)
+        $LD             r6,`7*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r10
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r10
+        $ST             r11,`8*$BNSZ`(r3)       #r[8]=c3;
+                                                #sqr_add_c2(a,7,2,c1,c2,c3);
+        $LD             r5,`2*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r0
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r11
+                                                #sqr_add_c2(a,6,3,c1,c2,c3);
+        $LD             r5,`3*$BNSZ`(r4)
+        $LD             r6,`6*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r11
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r11
+                                                #sqr_add_c2(a,5,4,c1,c2,c3);
+        $LD             r5,`4*$BNSZ`(r4)
+        $LD             r6,`5*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r11
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r11
+        $ST             r9,`9*$BNSZ`(r3)        #r[9]=c1;
+                                                #sqr_add_c(a,5,c2,c3,c1);
+        $UMULL          r7,r6,r6
+        $UMULH          r8,r6,r6
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r0
+                                                #sqr_add_c2(a,6,4,c2,c3,c1);
+        $LD             r6,`6*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r9
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r9
+                                                #sqr_add_c2(a,7,3,c2,c3,c1);
+        $LD             r5,`3*$BNSZ`(r4)
+        $LD             r6,`7*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r9
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r9
+        $ST             r10,`10*$BNSZ`(r3)      #r[10]=c2;
+                                                #sqr_add_c2(a,7,4,c3,c1,c2);
+        $LD             r5,`4*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r0
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r10
+                                                #sqr_add_c2(a,6,5,c3,c1,c2);
+        $LD             r5,`5*$BNSZ`(r4)
+        $LD             r6,`6*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r10
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        addze           r10,r10
+        $ST             r11,`11*$BNSZ`(r3)      #r[11]=c3;
+                                                #sqr_add_c(a,6,c1,c2,c3);
+        $UMULL          r7,r6,r6
+        $UMULH          r8,r6,r6
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r0
+                                                #sqr_add_c2(a,7,5,c1,c2,c3)
+        $LD             r6,`7*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r11
+        addc            r9,r7,r9
+        adde            r10,r8,r10
+        addze           r11,r11
+        $ST             r9,`12*$BNSZ`(r3)       #r[12]=c1;
+        
+                                                #sqr_add_c2(a,7,6,c2,c3,c1)
+        $LD             r5,`6*$BNSZ`(r4)
+        $UMULL          r7,r5,r6
+        $UMULH          r8,r5,r6
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r0
+        addc            r10,r7,r10
+        adde            r11,r8,r11
+        addze           r9,r9
+        $ST             r10,`13*$BNSZ`(r3)      #r[13]=c2;
+                                                #sqr_add_c(a,7,c3,c1,c2);
+        $UMULL          r7,r6,r6
+        $UMULH          r8,r6,r6
+        addc            r11,r7,r11
+        adde            r9,r8,r9
+        $ST             r11,`14*$BNSZ`(r3)      #r[14]=c3;
+        $ST             r9, `15*$BNSZ`(r3)      #r[15]=c1;
+
+
+        bclr    BO_ALWAYS,CR0_LT
+
+        .long   0x00000000
+
+#
+#       NOTE:   The following label name should be changed to
+#               "bn_mul_comba4" i.e. remove the first dot
+#               for the gcc compiler. This should be automatically
+#               done in the build
+#
+
+.align  4
+.bn_mul_comba4:
+#
+# This is an optimized version of the bn_mul_comba4 routine.
+#
+# void bn_mul_comba4(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b)
+# r3 contains r
+# r4 contains a
+# r5 contains b
+# r6, r7 are the 2 BN_ULONGs being multiplied.
+# r8, r9 are the results of the 32x32 giving 64 multiply.
+# r10, r11, r12 are the equivalents of c1, c2, and c3.
+#
+        xor     r0,r0,r0                #r0=0. Used in addze below.
+                                        #mul_add_c(a[0],b[0],c1,c2,c3);
+        $LD     r6,`0*$BNSZ`(r4)                
+        $LD     r7,`0*$BNSZ`(r5)                
+        $UMULL  r10,r6,r7               
+        $UMULH  r11,r6,r7               
+        $ST     r10,`0*$BNSZ`(r3)       #r[0]=c1
+                                        #mul_add_c(a[0],b[1],c2,c3,c1);
+        $LD     r7,`1*$BNSZ`(r5)                
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r8,r11
+        adde    r12,r9,r0
+        addze   r10,r0
+                                        #mul_add_c(a[1],b[0],c2,c3,c1);
+        $LD     r6, `1*$BNSZ`(r4)               
+        $LD     r7, `0*$BNSZ`(r5)               
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r8,r11
+        adde    r12,r9,r12
+        addze   r10,r10
+        $ST     r11,`1*$BNSZ`(r3)       #r[1]=c2
+                                        #mul_add_c(a[2],b[0],c3,c1,c2);
+        $LD     r6,`2*$BNSZ`(r4)                
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r8,r12
+        adde    r10,r9,r10
+        addze   r11,r0
+                                        #mul_add_c(a[1],b[1],c3,c1,c2);
+        $LD     r6,`1*$BNSZ`(r4)                
+        $LD     r7,`1*$BNSZ`(r5)                
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r8,r12
+        adde    r10,r9,r10
+        addze   r11,r11
+                                        #mul_add_c(a[0],b[2],c3,c1,c2);
+        $LD     r6,`0*$BNSZ`(r4)                
+        $LD     r7,`2*$BNSZ`(r5)                
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r8,r12
+        adde    r10,r9,r10
+        addze   r11,r11
+        $ST     r12,`2*$BNSZ`(r3)       #r[2]=c3
+                                        #mul_add_c(a[0],b[3],c1,c2,c3);
+        $LD     r7,`3*$BNSZ`(r5)                
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r8,r10
+        adde    r11,r9,r11
+        addze   r12,r0
+                                        #mul_add_c(a[1],b[2],c1,c2,c3);
+        $LD     r6,`1*$BNSZ`(r4)
+        $LD     r7,`2*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r8,r10
+        adde    r11,r9,r11
+        addze   r12,r12
+                                        #mul_add_c(a[2],b[1],c1,c2,c3);
+        $LD     r6,`2*$BNSZ`(r4)
+        $LD     r7,`1*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r8,r10
+        adde    r11,r9,r11
+        addze   r12,r12
+                                        #mul_add_c(a[3],b[0],c1,c2,c3);
+        $LD     r6,`3*$BNSZ`(r4)
+        $LD     r7,`0*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r8,r10
+        adde    r11,r9,r11
+        addze   r12,r12
+        $ST     r10,`3*$BNSZ`(r3)       #r[3]=c1
+                                        #mul_add_c(a[3],b[1],c2,c3,c1);
+        $LD     r7,`1*$BNSZ`(r5)                
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r8,r11
+        adde    r12,r9,r12
+        addze   r10,r0
+                                        #mul_add_c(a[2],b[2],c2,c3,c1);
+        $LD     r6,`2*$BNSZ`(r4)
+        $LD     r7,`2*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r8,r11
+        adde    r12,r9,r12
+        addze   r10,r10
+                                        #mul_add_c(a[1],b[3],c2,c3,c1);
+        $LD     r6,`1*$BNSZ`(r4)
+        $LD     r7,`3*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r8,r11
+        adde    r12,r9,r12
+        addze   r10,r10
+        $ST     r11,`4*$BNSZ`(r3)       #r[4]=c2
+                                        #mul_add_c(a[2],b[3],c3,c1,c2);
+        $LD     r6,`2*$BNSZ`(r4)                
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r8,r12
+        adde    r10,r9,r10
+        addze   r11,r0
+                                        #mul_add_c(a[3],b[2],c3,c1,c2);
+        $LD     r6,`3*$BNSZ`(r4)
+        $LD     r7,`2*$BNSZ`(r4)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r8,r12
+        adde    r10,r9,r10
+        addze   r11,r11
+        $ST     r12,`5*$BNSZ`(r3)       #r[5]=c3
+                                        #mul_add_c(a[3],b[3],c1,c2,c3);
+        $LD     r7,`3*$BNSZ`(r5)                
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r8,r10
+        adde    r11,r9,r11
+
+        $ST     r10,`6*$BNSZ`(r3)       #r[6]=c1
+        $ST     r11,`7*$BNSZ`(r3)       #r[7]=c2
+        bclr    BO_ALWAYS,CR0_LT
+        .long   0x00000000
+
+#
+#       NOTE:   The following label name should be changed to
+#               "bn_mul_comba8" i.e. remove the first dot
+#               for the gcc compiler. This should be automatically
+#               done in the build
+#
+        
+.align  4
+.bn_mul_comba8:
+#
+# Optimized version of the bn_mul_comba8 routine.
+#
+# void bn_mul_comba8(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b)
+# r3 contains r
+# r4 contains a
+# r5 contains b
+# r6, r7 are the 2 BN_ULONGs being multiplied.
+# r8, r9 are the results of the 32x32 giving 64 multiply.
+# r10, r11, r12 are the equivalents of c1, c2, and c3.
+#
+        xor     r0,r0,r0                #r0=0. Used in addze below.
+        
+                                        #mul_add_c(a[0],b[0],c1,c2,c3);
+        $LD     r6,`0*$BNSZ`(r4)        #a[0]
+        $LD     r7,`0*$BNSZ`(r5)        #b[0]
+        $UMULL  r10,r6,r7
+        $UMULH  r11,r6,r7
+        $ST     r10,`0*$BNSZ`(r3)       #r[0]=c1;
+                                        #mul_add_c(a[0],b[1],c2,c3,c1);
+        $LD     r7,`1*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        addze   r12,r9                  # since we didnt set r12 to zero before.
+        addze   r10,r0
+                                        #mul_add_c(a[1],b[0],c2,c3,c1);
+        $LD     r6,`1*$BNSZ`(r4)
+        $LD     r7,`0*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r10
+        $ST     r11,`1*$BNSZ`(r3)       #r[1]=c2;
+                                        #mul_add_c(a[2],b[0],c3,c1,c2);
+        $LD     r6,`2*$BNSZ`(r4)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r0
+                                        #mul_add_c(a[1],b[1],c3,c1,c2);
+        $LD     r6,`1*$BNSZ`(r4)
+        $LD     r7,`1*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r11
+                                        #mul_add_c(a[0],b[2],c3,c1,c2);
+        $LD     r6,`0*$BNSZ`(r4)
+        $LD     r7,`2*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r11
+        $ST     r12,`2*$BNSZ`(r3)       #r[2]=c3;
+                                        #mul_add_c(a[0],b[3],c1,c2,c3);
+        $LD     r7,`3*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r0
+                                        #mul_add_c(a[1],b[2],c1,c2,c3);
+        $LD     r6,`1*$BNSZ`(r4)
+        $LD     r7,`2*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r12
+                
+                                        #mul_add_c(a[2],b[1],c1,c2,c3);
+        $LD     r6,`2*$BNSZ`(r4)
+        $LD     r7,`1*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r12
+                                        #mul_add_c(a[3],b[0],c1,c2,c3);
+        $LD     r6,`3*$BNSZ`(r4)
+        $LD     r7,`0*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r12
+        $ST     r10,`3*$BNSZ`(r3)       #r[3]=c1;
+                                        #mul_add_c(a[4],b[0],c2,c3,c1);
+        $LD     r6,`4*$BNSZ`(r4)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r0
+                                        #mul_add_c(a[3],b[1],c2,c3,c1);
+        $LD     r6,`3*$BNSZ`(r4)
+        $LD     r7,`1*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r10
+                                        #mul_add_c(a[2],b[2],c2,c3,c1);
+        $LD     r6,`2*$BNSZ`(r4)
+        $LD     r7,`2*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r10
+                                        #mul_add_c(a[1],b[3],c2,c3,c1);
+        $LD     r6,`1*$BNSZ`(r4)
+        $LD     r7,`3*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r10
+                                        #mul_add_c(a[0],b[4],c2,c3,c1);
+        $LD     r6,`0*$BNSZ`(r4)
+        $LD     r7,`4*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r10
+        $ST     r11,`4*$BNSZ`(r3)       #r[4]=c2;
+                                        #mul_add_c(a[0],b[5],c3,c1,c2);
+        $LD     r7,`5*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r0
+                                        #mul_add_c(a[1],b[4],c3,c1,c2);
+        $LD     r6,`1*$BNSZ`(r4)                
+        $LD     r7,`4*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r11
+                                        #mul_add_c(a[2],b[3],c3,c1,c2);
+        $LD     r6,`2*$BNSZ`(r4)                
+        $LD     r7,`3*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r11
+                                        #mul_add_c(a[3],b[2],c3,c1,c2);
+        $LD     r6,`3*$BNSZ`(r4)                
+        $LD     r7,`2*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r11
+                                        #mul_add_c(a[4],b[1],c3,c1,c2);
+        $LD     r6,`4*$BNSZ`(r4)                
+        $LD     r7,`1*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r11
+                                        #mul_add_c(a[5],b[0],c3,c1,c2);
+        $LD     r6,`5*$BNSZ`(r4)                
+        $LD     r7,`0*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r11
+        $ST     r12,`5*$BNSZ`(r3)       #r[5]=c3;
+                                        #mul_add_c(a[6],b[0],c1,c2,c3);
+        $LD     r6,`6*$BNSZ`(r4)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r0
+                                        #mul_add_c(a[5],b[1],c1,c2,c3);
+        $LD     r6,`5*$BNSZ`(r4)
+        $LD     r7,`1*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r12
+                                        #mul_add_c(a[4],b[2],c1,c2,c3);
+        $LD     r6,`4*$BNSZ`(r4)
+        $LD     r7,`2*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r12
+                                        #mul_add_c(a[3],b[3],c1,c2,c3);
+        $LD     r6,`3*$BNSZ`(r4)
+        $LD     r7,`3*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r12
+                                        #mul_add_c(a[2],b[4],c1,c2,c3);
+        $LD     r6,`2*$BNSZ`(r4)
+        $LD     r7,`4*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r12
+                                        #mul_add_c(a[1],b[5],c1,c2,c3);
+        $LD     r6,`1*$BNSZ`(r4)
+        $LD     r7,`5*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r12
+                                        #mul_add_c(a[0],b[6],c1,c2,c3);
+        $LD     r6,`0*$BNSZ`(r4)
+        $LD     r7,`6*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r12
+        $ST     r10,`6*$BNSZ`(r3)       #r[6]=c1;
+                                        #mul_add_c(a[0],b[7],c2,c3,c1);
+        $LD     r7,`7*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r0
+                                        #mul_add_c(a[1],b[6],c2,c3,c1);
+        $LD     r6,`1*$BNSZ`(r4)
+        $LD     r7,`6*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r10
+                                        #mul_add_c(a[2],b[5],c2,c3,c1);
+        $LD     r6,`2*$BNSZ`(r4)
+        $LD     r7,`5*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r10
+                                        #mul_add_c(a[3],b[4],c2,c3,c1);
+        $LD     r6,`3*$BNSZ`(r4)
+        $LD     r7,`4*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r10
+                                        #mul_add_c(a[4],b[3],c2,c3,c1);
+        $LD     r6,`4*$BNSZ`(r4)
+        $LD     r7,`3*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r10
+                                        #mul_add_c(a[5],b[2],c2,c3,c1);
+        $LD     r6,`5*$BNSZ`(r4)
+        $LD     r7,`2*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r10
+                                        #mul_add_c(a[6],b[1],c2,c3,c1);
+        $LD     r6,`6*$BNSZ`(r4)
+        $LD     r7,`1*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r10
+                                        #mul_add_c(a[7],b[0],c2,c3,c1);
+        $LD     r6,`7*$BNSZ`(r4)
+        $LD     r7,`0*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r10
+        $ST     r11,`7*$BNSZ`(r3)       #r[7]=c2;
+                                        #mul_add_c(a[7],b[1],c3,c1,c2);
+        $LD     r7,`1*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r0
+                                        #mul_add_c(a[6],b[2],c3,c1,c2);
+        $LD     r6,`6*$BNSZ`(r4)
+        $LD     r7,`2*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r11
+                                        #mul_add_c(a[5],b[3],c3,c1,c2);
+        $LD     r6,`5*$BNSZ`(r4)
+        $LD     r7,`3*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r11
+                                        #mul_add_c(a[4],b[4],c3,c1,c2);
+        $LD     r6,`4*$BNSZ`(r4)
+        $LD     r7,`4*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r11
+                                        #mul_add_c(a[3],b[5],c3,c1,c2);
+        $LD     r6,`3*$BNSZ`(r4)
+        $LD     r7,`5*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r11
+                                        #mul_add_c(a[2],b[6],c3,c1,c2);
+        $LD     r6,`2*$BNSZ`(r4)
+        $LD     r7,`6*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r11
+                                        #mul_add_c(a[1],b[7],c3,c1,c2);
+        $LD     r6,`1*$BNSZ`(r4)
+        $LD     r7,`7*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r11
+        $ST     r12,`8*$BNSZ`(r3)       #r[8]=c3;
+                                        #mul_add_c(a[2],b[7],c1,c2,c3);
+        $LD     r6,`2*$BNSZ`(r4)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r0
+                                        #mul_add_c(a[3],b[6],c1,c2,c3);
+        $LD     r6,`3*$BNSZ`(r4)
+        $LD     r7,`6*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r12
+                                        #mul_add_c(a[4],b[5],c1,c2,c3);
+        $LD     r6,`4*$BNSZ`(r4)
+        $LD     r7,`5*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r12
+                                        #mul_add_c(a[5],b[4],c1,c2,c3);
+        $LD     r6,`5*$BNSZ`(r4)
+        $LD     r7,`4*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r12
+                                        #mul_add_c(a[6],b[3],c1,c2,c3);
+        $LD     r6,`6*$BNSZ`(r4)
+        $LD     r7,`3*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r12
+                                        #mul_add_c(a[7],b[2],c1,c2,c3);
+        $LD     r6,`7*$BNSZ`(r4)
+        $LD     r7,`2*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r12
+        $ST     r10,`9*$BNSZ`(r3)       #r[9]=c1;
+                                        #mul_add_c(a[7],b[3],c2,c3,c1);
+        $LD     r7,`3*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r0
+                                        #mul_add_c(a[6],b[4],c2,c3,c1);
+        $LD     r6,`6*$BNSZ`(r4)
+        $LD     r7,`4*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r10
+                                        #mul_add_c(a[5],b[5],c2,c3,c1);
+        $LD     r6,`5*$BNSZ`(r4)
+        $LD     r7,`5*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r10
+                                        #mul_add_c(a[4],b[6],c2,c3,c1);
+        $LD     r6,`4*$BNSZ`(r4)
+        $LD     r7,`6*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r10
+                                        #mul_add_c(a[3],b[7],c2,c3,c1);
+        $LD     r6,`3*$BNSZ`(r4)
+        $LD     r7,`7*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r10
+        $ST     r11,`10*$BNSZ`(r3)      #r[10]=c2;
+                                        #mul_add_c(a[4],b[7],c3,c1,c2);
+        $LD     r6,`4*$BNSZ`(r4)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r0
+                                        #mul_add_c(a[5],b[6],c3,c1,c2);
+        $LD     r6,`5*$BNSZ`(r4)
+        $LD     r7,`6*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r11
+                                        #mul_add_c(a[6],b[5],c3,c1,c2);
+        $LD     r6,`6*$BNSZ`(r4)
+        $LD     r7,`5*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r11
+                                        #mul_add_c(a[7],b[4],c3,c1,c2);
+        $LD     r6,`7*$BNSZ`(r4)
+        $LD     r7,`4*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        addze   r11,r11
+        $ST     r12,`11*$BNSZ`(r3)      #r[11]=c3;
+                                        #mul_add_c(a[7],b[5],c1,c2,c3);
+        $LD     r7,`5*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r0
+                                        #mul_add_c(a[6],b[6],c1,c2,c3);
+        $LD     r6,`6*$BNSZ`(r4)
+        $LD     r7,`6*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r12
+                                        #mul_add_c(a[5],b[7],c1,c2,c3);
+        $LD     r6,`5*$BNSZ`(r4)
+        $LD     r7,`7*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r10,r10,r8
+        adde    r11,r11,r9
+        addze   r12,r12
+        $ST     r10,`12*$BNSZ`(r3)      #r[12]=c1;
+                                        #mul_add_c(a[6],b[7],c2,c3,c1);
+        $LD     r6,`6*$BNSZ`(r4)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r0
+                                        #mul_add_c(a[7],b[6],c2,c3,c1);
+        $LD     r6,`7*$BNSZ`(r4)
+        $LD     r7,`6*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r11,r11,r8
+        adde    r12,r12,r9
+        addze   r10,r10
+        $ST     r11,`13*$BNSZ`(r3)      #r[13]=c2;
+                                        #mul_add_c(a[7],b[7],c3,c1,c2);
+        $LD     r7,`7*$BNSZ`(r5)
+        $UMULL  r8,r6,r7
+        $UMULH  r9,r6,r7
+        addc    r12,r12,r8
+        adde    r10,r10,r9
+        $ST     r12,`14*$BNSZ`(r3)      #r[14]=c3;
+        $ST     r10,`15*$BNSZ`(r3)      #r[15]=c1;
+        bclr    BO_ALWAYS,CR0_LT
+        .long   0x00000000
+
+#
+#       NOTE:   The following label name should be changed to
+#               "bn_sub_words" i.e. remove the first dot
+#               for the gcc compiler. This should be automatically
+#               done in the build
+#
+#
+.align  4
+.bn_sub_words:
+#
+#       Handcoded version of bn_sub_words
+#
+#BN_ULONG bn_sub_words(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n)
+#
+#       r3 = r
+#       r4 = a
+#       r5 = b
+#       r6 = n
+#
+#       Note:   No loop unrolling done since this is not a performance
+#               critical loop.
+
+        xor     r0,r0,r0        #set r0 = 0
+#
+#       check for r6 = 0 AND set carry bit.
+#
+        subfc.  r7,r0,r6        # If r6 is 0 then result is 0.
+                                # if r6 > 0 then result !=0
+                                # In either case carry bit is set.
+        bc      BO_IF,CR0_EQ,Lppcasm_sub_adios
+        addi    r4,r4,-$BNSZ
+        addi    r3,r3,-$BNSZ
+        addi    r5,r5,-$BNSZ
+        mtctr   r6
+Lppcasm_sub_mainloop:   
+        $LDU    r7,$BNSZ(r4)
+        $LDU    r8,$BNSZ(r5)
+        subfe   r6,r8,r7        # r6 = r7+carry bit + onescomplement(r8)
+                                # if carry = 1 this is r7-r8. Else it
+                                # is r7-r8 -1 as we need.
+        $STU    r6,$BNSZ(r3)
+        bc      BO_dCTR_NZERO,CR0_EQ,Lppcasm_sub_mainloop
+Lppcasm_sub_adios:      
+        subfze  r3,r0           # if carry bit is set then r3 = 0 else -1
+        andi.   r3,r3,1         # keep only last bit.
+        bclr    BO_ALWAYS,CR0_LT
+        .long   0x00000000
+
+
+#
+#       NOTE:   The following label name should be changed to
+#               "bn_add_words" i.e. remove the first dot
+#               for the gcc compiler. This should be automatically
+#               done in the build
+#
+
+.align  4
+.bn_add_words:
+#
+#       Handcoded version of bn_add_words
+#
+#BN_ULONG bn_add_words(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n)
+#
+#       r3 = r
+#       r4 = a
+#       r5 = b
+#       r6 = n
+#
+#       Note:   No loop unrolling done since this is not a performance
+#               critical loop.
+
+        xor     r0,r0,r0
+#
+#       check for r6 = 0. Is this needed?
+#
+        addic.  r6,r6,0         #test r6 and clear carry bit.
+        bc      BO_IF,CR0_EQ,Lppcasm_add_adios
+        addi    r4,r4,-$BNSZ
+        addi    r3,r3,-$BNSZ
+        addi    r5,r5,-$BNSZ
+        mtctr   r6
+Lppcasm_add_mainloop:   
+        $LDU    r7,$BNSZ(r4)
+        $LDU    r8,$BNSZ(r5)
+        adde    r8,r7,r8
+        $STU    r8,$BNSZ(r3)
+        bc      BO_dCTR_NZERO,CR0_EQ,Lppcasm_add_mainloop
+Lppcasm_add_adios:      
+        addze   r3,r0                   #return carry bit.
+        bclr    BO_ALWAYS,CR0_LT
+        .long   0x00000000
+
+#
+#       NOTE:   The following label name should be changed to
+#               "bn_div_words" i.e. remove the first dot
+#               for the gcc compiler. This should be automatically
+#               done in the build
+#
+
+.align  4
+.bn_div_words:
+#
+#       This is a cleaned up version of code generated by
+#       the AIX compiler. The only optimization is to use
+#       the PPC instruction to count leading zeros instead
+#       of call to num_bits_word. Since this was compiled
+#       only at level -O2 we can possibly squeeze it more?
+#       
+#       r3 = h
+#       r4 = l
+#       r5 = d
+        
+        $UCMPI  0,r5,0                  # compare r5 and 0
+        bc      BO_IF_NOT,CR0_EQ,Lppcasm_div1   # proceed if d!=0
+        li      r3,-1                   # d=0 return -1
+        bclr    BO_ALWAYS,CR0_LT        
+Lppcasm_div1:
+        xor     r0,r0,r0                #r0=0
+        $COUNTZ r7,r5                   #r7 = num leading 0s in d.
+        subfic  r8,r7,$BITS             #r8 = BN_num_bits_word(d)
+        cmpi    0,0,r8,$BITS            #
+        bc      BO_IF,CR0_EQ,Lppcasm_div2       #proceed if (r8==$BITS) 
+        li      r9,1                    # r9=1
+        $SHL    r10,r9,r8               # r9<<=r8
+        $UCMP   0,r3,r10                #       
+        bc      BO_IF,CR0_GT,Lppcasm_div2       #or if (h > (1<<r8))
+        $UDIV   r3,r3,r0                #if not assert(0) divide by 0!
+                                        #that's how we signal overflow
+        bclr    BO_ALWAYS,CR0_LT        #return. NEVER REACHED.
+Lppcasm_div2:
+        $UCMP   0,r3,r5                 #h>=d?
+        bc      BO_IF,CR0_LT,Lppcasm_div3       #goto Lppcasm_div3 if not
+        subf    r3,r5,r3                #h-=d ; 
+Lppcasm_div3:                           #r7 = BN_BITS2-i. so r7=i
+        cmpi    0,0,r7,0                # is (i == 0)?
+        bc      BO_IF,CR0_EQ,Lppcasm_div4
+        $SHL    r3,r3,r7                # h = (h<< i)
+        $SHR    r8,r4,r8                # r8 = (l >> BN_BITS2 -i)
+        $SHL    r5,r5,r7                # d<<=i
+        or      r3,r3,r8                # h = (h<<i)|(l>>(BN_BITS2-i))
+        $SHL    r4,r4,r7                # l <<=i
+Lppcasm_div4:
+        $SHRI   r9,r5,`$BITS/2`         # r9 = dh
+                                        # dl will be computed when needed
+                                        # as it saves registers.
+        li      r6,2                    #r6=2
+        mtctr   r6                      #counter will be in count.
+Lppcasm_divouterloop: 
+        $SHRI   r8,r3,`$BITS/2`         #r8 = (h>>BN_BITS4)
+        $SHRI   r11,r4,`$BITS/2`        #r11= (l&BN_MASK2h)>>BN_BITS4
+                                        # compute here for innerloop.
+        $UCMP   0,r8,r9                 # is (h>>BN_BITS4)==dh
+        bc      BO_IF_NOT,CR0_EQ,Lppcasm_div5   # goto Lppcasm_div5 if not
+
+        li      r8,-1
+        $CLRU   r8,r8,`$BITS/2`         #q = BN_MASK2l 
+        b       Lppcasm_div6
+Lppcasm_div5:
+        $UDIV   r8,r3,r9                #q = h/dh
+Lppcasm_div6:
+        $UMULL  r12,r9,r8               #th = q*dh
+        $CLRU   r10,r5,`$BITS/2`        #r10=dl
+        $UMULL  r6,r8,r10               #tl = q*dl
+        
+Lppcasm_divinnerloop:
+        subf    r10,r12,r3              #t = h -th
+        $SHRI   r7,r10,`$BITS/2`        #r7= (t &BN_MASK2H), sort of...
+        addic.  r7,r7,0                 #test if r7 == 0. used below.
+                                        # now want to compute
+                                        # r7 = (t<<BN_BITS4)|((l&BN_MASK2h)>>BN_BITS4)
+                                        # the following 2 instructions do that
+        $SHLI   r7,r10,`$BITS/2`        # r7 = (t<<BN_BITS4)
+        or      r7,r7,r11               # r7|=((l&BN_MASK2h)>>BN_BITS4)
+        $UCMP   1,r6,r7                 # compare (tl <= r7)
+        bc      BO_IF_NOT,CR0_EQ,Lppcasm_divinnerexit
+        bc      BO_IF_NOT,CR1_FEX,Lppcasm_divinnerexit
+        addi    r8,r8,-1                #q--
+        subf    r12,r9,r12              #th -=dh
+        $CLRU   r10,r5,`$BITS/2`        #r10=dl. t is no longer needed in loop.
+        subf    r6,r10,r6               #tl -=dl
+        b       Lppcasm_divinnerloop
+Lppcasm_divinnerexit:
+        $SHRI   r10,r6,`$BITS/2`        #t=(tl>>BN_BITS4)
+        $SHLI   r11,r6,`$BITS/2`        #tl=(tl<<BN_BITS4)&BN_MASK2h;
+        $UCMP   1,r4,r11                # compare l and tl
+        add     r12,r12,r10             # th+=t
+        bc      BO_IF_NOT,CR1_FX,Lppcasm_div7  # if (l>=tl) goto Lppcasm_div7
+        addi    r12,r12,1               # th++
+Lppcasm_div7:
+        subf    r11,r11,r4              #r11=l-tl
+        $UCMP   1,r3,r12                #compare h and th
+        bc      BO_IF_NOT,CR1_FX,Lppcasm_div8   #if (h>=th) goto Lppcasm_div8
+        addi    r8,r8,-1                # q--
+        add     r3,r5,r3                # h+=d
+Lppcasm_div8:
+        subf    r12,r12,r3              #r12 = h-th
+        $SHLI   r4,r11,`$BITS/2`        #l=(l&BN_MASK2l)<<BN_BITS4
+                                        # want to compute
+                                        # h = ((h<<BN_BITS4)|(l>>BN_BITS4))&BN_MASK2
+                                        # the following 2 instructions will do this.
+        $INSR   r11,r12,`$BITS/2`,`$BITS/2`     # r11 is the value we want rotated $BITS/2.
+        $ROTL   r3,r11,`$BITS/2`        # rotate by $BITS/2 and store in r3
+        bc      BO_dCTR_ZERO,CR0_EQ,Lppcasm_div9#if (count==0) break ;
+        $SHLI   r0,r8,`$BITS/2`         #ret =q<<BN_BITS4
+        b       Lppcasm_divouterloop
+Lppcasm_div9:
+        or      r3,r8,r0
+        bclr    BO_ALWAYS,CR0_LT
+        .long   0x00000000
+
+#
+#       NOTE:   The following label name should be changed to
+#               "bn_sqr_words" i.e. remove the first dot
+#               for the gcc compiler. This should be automatically
+#               done in the build
+#
+.align  4
+.bn_sqr_words:
+#
+#       Optimized version of bn_sqr_words
+#
+#       void bn_sqr_words(BN_ULONG *r, BN_ULONG *a, int n)
+#
+#       r3 = r
+#       r4 = a
+#       r5 = n
+#
+#       r6 = a[i].
+#       r7,r8 = product.
+#
+#       No unrolling done here. Not performance critical.
+
+        addic.  r5,r5,0                 #test r5.
+        bc      BO_IF,CR0_EQ,Lppcasm_sqr_adios
+        addi    r4,r4,-$BNSZ
+        addi    r3,r3,-$BNSZ
+        mtctr   r5
+Lppcasm_sqr_mainloop:   
+                                        #sqr(r[0],r[1],a[0]);
+        $LDU    r6,$BNSZ(r4)
+        $UMULL  r7,r6,r6
+        $UMULH  r8,r6,r6
+        $STU    r7,$BNSZ(r3)
+        $STU    r8,$BNSZ(r3)
+        bc      BO_dCTR_NZERO,CR0_EQ,Lppcasm_sqr_mainloop
+Lppcasm_sqr_adios:      
+        bclr    BO_ALWAYS,CR0_LT
+        .long   0x00000000
+
+
+#
+#       NOTE:   The following label name should be changed to
+#               "bn_mul_words" i.e. remove the first dot
+#               for the gcc compiler. This should be automatically
+#               done in the build
+#
+
+.align  4       
+.bn_mul_words:
+#
+# BN_ULONG bn_mul_words(BN_ULONG *rp, BN_ULONG *ap, int num, BN_ULONG w)
+#
+# r3 = rp
+# r4 = ap
+# r5 = num
+# r6 = w
+        xor     r0,r0,r0
+        xor     r12,r12,r12             # used for carry
+        rlwinm. r7,r5,30,2,31           # num >> 2
+        bc      BO_IF,CR0_EQ,Lppcasm_mw_REM
+        mtctr   r7
+Lppcasm_mw_LOOP:        
+                                        #mul(rp[0],ap[0],w,c1);
+        $LD     r8,`0*$BNSZ`(r4)
+        $UMULL  r9,r6,r8
+        $UMULH  r10,r6,r8
+        addc    r9,r9,r12
+        #addze  r10,r10                 #carry is NOT ignored.
+                                        #will be taken care of
+                                        #in second spin below
+                                        #using adde.
+        $ST     r9,`0*$BNSZ`(r3)
+                                        #mul(rp[1],ap[1],w,c1);
+        $LD     r8,`1*$BNSZ`(r4)        
+        $UMULL  r11,r6,r8
+        $UMULH  r12,r6,r8
+        adde    r11,r11,r10
+        #addze  r12,r12
+        $ST     r11,`1*$BNSZ`(r3)
+                                        #mul(rp[2],ap[2],w,c1);
+        $LD     r8,`2*$BNSZ`(r4)
+        $UMULL  r9,r6,r8
+        $UMULH  r10,r6,r8
+        adde    r9,r9,r12
+        #addze  r10,r10
+        $ST     r9,`2*$BNSZ`(r3)
+                                        #mul_add(rp[3],ap[3],w,c1);
+        $LD     r8,`3*$BNSZ`(r4)
+        $UMULL  r11,r6,r8
+        $UMULH  r12,r6,r8
+        adde    r11,r11,r10
+        addze   r12,r12                 #this spin we collect carry into
+                                        #r12
+        $ST     r11,`3*$BNSZ`(r3)
+        
+        addi    r3,r3,`4*$BNSZ`
+        addi    r4,r4,`4*$BNSZ`
+        bc      BO_dCTR_NZERO,CR0_EQ,Lppcasm_mw_LOOP
+
+Lppcasm_mw_REM:
+        andi.   r5,r5,0x3
+        bc      BO_IF,CR0_EQ,Lppcasm_mw_OVER
+                                        #mul(rp[0],ap[0],w,c1);
+        $LD     r8,`0*$BNSZ`(r4)
+        $UMULL  r9,r6,r8
+        $UMULH  r10,r6,r8
+        addc    r9,r9,r12
+        addze   r10,r10
+        $ST     r9,`0*$BNSZ`(r3)
+        addi    r12,r10,0
+        
+        addi    r5,r5,-1
+        cmpli   0,0,r5,0
+        bc      BO_IF,CR0_EQ,Lppcasm_mw_OVER
+
+        
+                                        #mul(rp[1],ap[1],w,c1);
+        $LD     r8,`1*$BNSZ`(r4)        
+        $UMULL  r9,r6,r8
+        $UMULH  r10,r6,r8
+        addc    r9,r9,r12
+        addze   r10,r10
+        $ST     r9,`1*$BNSZ`(r3)
+        addi    r12,r10,0
+        
+        addi    r5,r5,-1
+        cmpli   0,0,r5,0
+        bc      BO_IF,CR0_EQ,Lppcasm_mw_OVER
+        
+                                        #mul_add(rp[2],ap[2],w,c1);
+        $LD     r8,`2*$BNSZ`(r4)
+        $UMULL  r9,r6,r8
+        $UMULH  r10,r6,r8
+        addc    r9,r9,r12
+        addze   r10,r10
+        $ST     r9,`2*$BNSZ`(r3)
+        addi    r12,r10,0
+                
+Lppcasm_mw_OVER:        
+        addi    r3,r12,0
+        bclr    BO_ALWAYS,CR0_LT
+        .long   0x00000000
+
+#
+#       NOTE:   The following label name should be changed to
+#               "bn_mul_add_words" i.e. remove the first dot
+#               for the gcc compiler. This should be automatically
+#               done in the build
+#
+
+.align  4
+.bn_mul_add_words:
+#
+# BN_ULONG bn_mul_add_words(BN_ULONG *rp, BN_ULONG *ap, int num, BN_ULONG w)
+#
+# r3 = rp
+# r4 = ap
+# r5 = num
+# r6 = w
+#
+# empirical evidence suggests that unrolled version performs best!!
+#
+        xor     r0,r0,r0                #r0 = 0
+        xor     r12,r12,r12             #r12 = 0 . used for carry               
+        rlwinm. r7,r5,30,2,31           # num >> 2
+        bc      BO_IF,CR0_EQ,Lppcasm_maw_leftover       # if (num < 4) go LPPCASM_maw_leftover
+        mtctr   r7
+Lppcasm_maw_mainloop:   
+                                        #mul_add(rp[0],ap[0],w,c1);
+        $LD     r8,`0*$BNSZ`(r4)
+        $LD     r11,`0*$BNSZ`(r3)
+        $UMULL  r9,r6,r8
+        $UMULH  r10,r6,r8
+        addc    r9,r9,r12               #r12 is carry.
+        addze   r10,r10
+        addc    r9,r9,r11
+        #addze  r10,r10
+                                        #the above instruction addze
+                                        #is NOT needed. Carry will NOT
+                                        #be ignored. It's not affected
+                                        #by multiply and will be collected
+                                        #in the next spin
+        $ST     r9,`0*$BNSZ`(r3)
+        
+                                        #mul_add(rp[1],ap[1],w,c1);
+        $LD     r8,`1*$BNSZ`(r4)        
+        $LD     r9,`1*$BNSZ`(r3)
+        $UMULL  r11,r6,r8
+        $UMULH  r12,r6,r8
+        adde    r11,r11,r10             #r10 is carry.
+        addze   r12,r12
+        addc    r11,r11,r9
+        #addze  r12,r12
+        $ST     r11,`1*$BNSZ`(r3)
+        
+                                        #mul_add(rp[2],ap[2],w,c1);
+        $LD     r8,`2*$BNSZ`(r4)
+        $UMULL  r9,r6,r8
+        $LD     r11,`2*$BNSZ`(r3)
+        $UMULH  r10,r6,r8
+        adde    r9,r9,r12
+        addze   r10,r10
+        addc    r9,r9,r11
+        #addze  r10,r10
+        $ST     r9,`2*$BNSZ`(r3)
+        
+                                        #mul_add(rp[3],ap[3],w,c1);
+        $LD     r8,`3*$BNSZ`(r4)
+        $UMULL  r11,r6,r8
+        $LD     r9,`3*$BNSZ`(r3)
+        $UMULH  r12,r6,r8
+        adde    r11,r11,r10
+        addze   r12,r12
+        addc    r11,r11,r9
+        addze   r12,r12
+        $ST     r11,`3*$BNSZ`(r3)
+        addi    r3,r3,`4*$BNSZ`
+        addi    r4,r4,`4*$BNSZ`
+        bc      BO_dCTR_NZERO,CR0_EQ,Lppcasm_maw_mainloop
+        
+Lppcasm_maw_leftover:
+        andi.   r5,r5,0x3
+        bc      BO_IF,CR0_EQ,Lppcasm_maw_adios
+        addi    r3,r3,-$BNSZ
+        addi    r4,r4,-$BNSZ
+                                        #mul_add(rp[0],ap[0],w,c1);
+        mtctr   r5
+        $LDU    r8,$BNSZ(r4)
+        $UMULL  r9,r6,r8
+        $UMULH  r10,r6,r8
+        $LDU    r11,$BNSZ(r3)
+        addc    r9,r9,r11
+        addze   r10,r10
+        addc    r9,r9,r12
+        addze   r12,r10
+        $ST     r9,0(r3)
+        
+        bc      BO_dCTR_ZERO,CR0_EQ,Lppcasm_maw_adios
+                                        #mul_add(rp[1],ap[1],w,c1);
+        $LDU    r8,$BNSZ(r4)    
+        $UMULL  r9,r6,r8
+        $UMULH  r10,r6,r8
+        $LDU    r11,$BNSZ(r3)
+        addc    r9,r9,r11
+        addze   r10,r10
+        addc    r9,r9,r12
+        addze   r12,r10
+        $ST     r9,0(r3)
+        
+        bc      BO_dCTR_ZERO,CR0_EQ,Lppcasm_maw_adios
+                                        #mul_add(rp[2],ap[2],w,c1);
+        $LDU    r8,$BNSZ(r4)
+        $UMULL  r9,r6,r8
+        $UMULH  r10,r6,r8
+        $LDU    r11,$BNSZ(r3)
+        addc    r9,r9,r11
+        addze   r10,r10
+        addc    r9,r9,r12
+        addze   r12,r10
+        $ST     r9,0(r3)
+                
+Lppcasm_maw_adios:      
+        addi    r3,r12,0
+        bclr    BO_ALWAYS,CR0_LT
+        .long   0x00000000
+        .align  4
+EOF
+        $data =~ s/\`([^\`]*)\`/eval $1/gem;
+
+        # if some assembler chokes on some simplified mnemonic,
+        # this is the spot to fix it up, e.g.:
+        # GNU as doesn't seem to accept cmplw, 32-bit unsigned compare
+        $data =~ s/^(\s*)cmplw(\s+)([^,]+),(.*)/$1cmpl$2$3,0,$4/gm;
+        # assembler X doesn't accept li, load immediate value
+        #$data =~ s/^(\s*)li(\s+)([^,]+),(.*)/$1addi$2$3,0,$4/gm;
+        return($data);
+}
diff --git a/src/lib/libcrypto/bn/asm/x86_64-gcc.c b/src/lib/libcrypto/bn/asm/x86_64-gcc.c
index 450e8e4322..7378344251 100644
--- a/src/lib/libcrypto/bn/asm/x86_64-gcc.c
+++ b/src/lib/libcrypto/bn/asm/x86_64-gcc.c
@@ -13,20 +13,42 @@
  * A. Well, that's because this code is basically a quick-n-dirty
  *    proof-of-concept hack. As you can see it's implemented with
  *    inline assembler, which means that you're bound to GCC and that
- *    there must be a room for fine-tuning.
+ *    there might be enough room for further improvement.
  *
  * Q. Why inline assembler?
- * A. x86_64 features own ABI I'm not familiar with. Which is why
- *    I decided to let the compiler take care of subroutine
- *    prologue/epilogue as well as register allocation.
+ * A. x86_64 features own ABI which I'm not familiar with. This is
+ *    why I decided to let the compiler take care of subroutine
+ *    prologue/epilogue as well as register allocation. For reference.
+ *    Win64 implements different ABI for AMD64, different from Linux.
  *
  * Q. How much faster does it get?
- * A. Unfortunately people sitting on x86_64 hardware are prohibited
- *    to disclose the performance numbers, so they (SuSE labs to be
- *    specific) wouldn't tell me. However! Very similar coding technique
- *    (reaching out for 128-bit result from 64x64-bit multiplication)
- *    results in >3 times performance improvement on MIPS and I see no
- *    reason why gain on x86_64 would be so much different:-)
+ * A. 'apps/openssl speed rsa dsa' output with no-asm:
+ *
+ *                        sign    verify    sign/s verify/s
+ *      rsa  512 bits   0.0006s   0.0001s   1683.8  18456.2
+ *      rsa 1024 bits   0.0028s   0.0002s    356.0   6407.0
+ *      rsa 2048 bits   0.0172s   0.0005s     58.0   1957.8
+ *      rsa 4096 bits   0.1155s   0.0018s      8.7    555.6
+ *                        sign    verify    sign/s verify/s
+ *      dsa  512 bits   0.0005s   0.0006s   2100.8   1768.3
+ *      dsa 1024 bits   0.0014s   0.0018s    692.3    559.2
+ *      dsa 2048 bits   0.0049s   0.0061s    204.7    165.0
+ *
+ *    'apps/openssl speed rsa dsa' output with this module:
+ *
+ *                        sign    verify    sign/s verify/s
+ *      rsa  512 bits   0.0004s   0.0000s   2767.1  33297.9
+ *      rsa 1024 bits   0.0012s   0.0001s    867.4  14674.7
+ *      rsa 2048 bits   0.0061s   0.0002s    164.0   5270.0
+ *      rsa 4096 bits   0.0384s   0.0006s     26.1   1650.8
+ *                        sign    verify    sign/s verify/s
+ *      dsa  512 bits   0.0002s   0.0003s   4442.2   3786.3
+ *      dsa 1024 bits   0.0005s   0.0007s   1835.1   1497.4
+ *      dsa 2048 bits   0.0016s   0.0020s    620.4    504.6
+ *
+ *    For the reference. IA-32 assembler implementation performs
+ *    very much like 64-bit code compiled with no-asm on the same
+ *    machine.
  */
 
 #define BN_ULONG unsigned long
@@ -151,7 +173,7 @@ BN_ULONG bn_div_words(BN_ULONG h, BN_ULONG l, BN_ULONG d)
 }
 
 BN_ULONG bn_add_words (BN_ULONG *rp, BN_ULONG *ap, BN_ULONG *bp,int n)
-{ BN_ULONG ret,i;
+{ BN_ULONG ret=0,i=0;
 
         if (n <= 0) return 0;
 
@@ -164,7 +186,7 @@ BN_ULONG bn_add_words (BN_ULONG *rp, BN_ULONG *ap, BN_ULONG *bp,int n)
         "       leaq    1(%2),%2        \n"
         "       loop    1b              \n"
         "       sbbq    %0,%0           \n"
-                : "+a"(ret),"+c"(n),"+r"(i)
+                : "=&a"(ret),"+c"(n),"=&r"(i)
                 : "r"(rp),"r"(ap),"r"(bp)
                 : "cc"
         );
@@ -174,7 +196,7 @@ BN_ULONG bn_add_words (BN_ULONG *rp, BN_ULONG *ap, BN_ULONG *bp,int n)
 
 #ifndef SIMICS
 BN_ULONG bn_sub_words (BN_ULONG *rp, BN_ULONG *ap, BN_ULONG *bp,int n)
-{ BN_ULONG ret,i;
+{ BN_ULONG ret=0,i=0;
 
         if (n <= 0) return 0;
 
@@ -187,7 +209,7 @@ BN_ULONG bn_sub_words (BN_ULONG *rp, BN_ULONG *ap, BN_ULONG *bp,int n)
         "       leaq    1(%2),%2        \n"
         "       loop    1b              \n"
         "       sbbq    %0,%0           \n"
-                : "+a"(ret),"+c"(n),"+r"(i)
+                : "=&a"(ret),"+c"(n),"=&r"(i)
                 : "r"(rp),"r"(ap),"r"(bp)
                 : "cc"
         );
@@ -318,7 +340,6 @@ BN_ULONG bn_sub_words(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n)
 
 void bn_mul_comba8(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b)
         {
-        BN_ULONG bl,bh;
         BN_ULONG t1,t2;
         BN_ULONG c1,c2,c3;
 
@@ -423,7 +444,6 @@ void bn_mul_comba8(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b)
 
 void bn_mul_comba4(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b)
         {
-        BN_ULONG bl,bh;
         BN_ULONG t1,t2;
         BN_ULONG c1,c2,c3;
 
@@ -464,7 +484,6 @@ void bn_mul_comba4(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b)
 
 void bn_sqr_comba8(BN_ULONG *r, BN_ULONG *a)
         {
-        BN_ULONG bl,bh;
         BN_ULONG t1,t2;
         BN_ULONG c1,c2,c3;
 
@@ -541,7 +560,6 @@ void bn_sqr_comba8(BN_ULONG *r, BN_ULONG *a)
 
 void bn_sqr_comba4(BN_ULONG *r, BN_ULONG *a)
         {
-        BN_ULONG bl,bh;
         BN_ULONG t1,t2;
         BN_ULONG c1,c2,c3;
 
diff --git a/src/lib/libcrypto/doc/X509_NAME_ENTRY_get_object.pod b/src/lib/libcrypto/doc/X509_NAME_ENTRY_get_object.pod
index d287c18564..11b35f6fd3 100644
--- a/src/lib/libcrypto/doc/X509_NAME_ENTRY_get_object.pod
+++ b/src/lib/libcrypto/doc/X509_NAME_ENTRY_get_object.pod
@@ -13,11 +13,11 @@ ASN1_OBJECT * X509_NAME_ENTRY_get_object(X509_NAME_ENTRY *ne);
 ASN1_STRING * X509_NAME_ENTRY_get_data(X509_NAME_ENTRY *ne);
 
 int X509_NAME_ENTRY_set_object(X509_NAME_ENTRY *ne, ASN1_OBJECT *obj);
-int X509_NAME_ENTRY_set_data(X509_NAME_ENTRY *ne, int type, unsigned char *bytes, int len);
+int X509_NAME_ENTRY_set_data(X509_NAME_ENTRY *ne, int type, const unsigned char *bytes, int len);
 
-X509_NAME_ENTRY *X509_NAME_ENTRY_create_by_txt(X509_NAME_ENTRY **ne, char *field, int type, unsigned char *bytes, int len);
+X509_NAME_ENTRY *X509_NAME_ENTRY_create_by_txt(X509_NAME_ENTRY **ne, const char *field, int type, const unsigned char *bytes, int len);
 X509_NAME_ENTRY *X509_NAME_ENTRY_create_by_NID(X509_NAME_ENTRY **ne, int nid, int type,unsigned char *bytes, int len);
-X509_NAME_ENTRY *X509_NAME_ENTRY_create_by_OBJ(X509_NAME_ENTRY **ne, ASN1_OBJECT *obj, int type,unsigned char *bytes, int len);
+X509_NAME_ENTRY *X509_NAME_ENTRY_create_by_OBJ(X509_NAME_ENTRY **ne, ASN1_OBJECT *obj, int type, const unsigned char *bytes, int len);
 
 =head1 DESCRIPTION
 
diff --git a/src/lib/libcrypto/doc/X509_NAME_add_entry_by_txt.pod b/src/lib/libcrypto/doc/X509_NAME_add_entry_by_txt.pod
index 4472a1c5cf..e2ab4b0d2b 100644
--- a/src/lib/libcrypto/doc/X509_NAME_add_entry_by_txt.pod
+++ b/src/lib/libcrypto/doc/X509_NAME_add_entry_by_txt.pod
@@ -7,10 +7,14 @@ X509_NAME_add_entry, X509_NAME_delete_entry - X509_NAME modification functions
 
 =head1 SYNOPSIS
 
-int X509_NAME_add_entry_by_txt(X509_NAME *name, char *field, int type, unsigned char *bytes, int len, int loc, int set);
+int X509_NAME_add_entry_by_txt(X509_NAME *name, const char *field, int type, const unsigned char *bytes, int len, int loc, int set);
+
 int X509_NAME_add_entry_by_OBJ(X509_NAME *name, ASN1_OBJECT *obj, int type, unsigned char *bytes, int len, int loc, int set);
+
 int X509_NAME_add_entry_by_NID(X509_NAME *name, int nid, int type, unsigned char *bytes, int len, int loc, int set);
+
 int X509_NAME_add_entry(X509_NAME *name,X509_NAME_ENTRY *ne, int loc, int set);
+
 X509_NAME_ENTRY *X509_NAME_delete_entry(X509_NAME *name, int loc);
 
 =head1 DESCRIPTION
diff --git a/src/lib/libcrypto/doc/X509_NAME_print_ex.pod b/src/lib/libcrypto/doc/X509_NAME_print_ex.pod
index 907c04f684..919b908919 100644
--- a/src/lib/libcrypto/doc/X509_NAME_print_ex.pod
+++ b/src/lib/libcrypto/doc/X509_NAME_print_ex.pod
@@ -41,8 +41,8 @@ applications.
 Although there are a large number of possible flags for most purposes
 B<XN_FLAG_ONELINE>, B<XN_FLAG_MULTILINE> or B<XN_FLAG_RFC2253> will suffice.
 As noted on the L<ASN1_STRING_print_ex(3)|ASN1_STRING_print_ex(3)> manual page
-for UTF8 terminals the B<ASN1_STRFLAGS_ESC_MSB> should be unset: so for example
-B<XN_FLAG_ONELINE & ~ASN1_STRFLAGS_ESC_MSB> would be used.
+for UTF8 terminals the B<ASN1_STRFLGS_ESC_MSB> should be unset: so for example
+B<XN_FLAG_ONELINE & ~ASN1_STRFLGS_ESC_MSB> would be used.
 
 The complete set of the flags supported by X509_NAME_print_ex() is listed below.
 
diff --git a/src/lib/libcrypto/evp/e_old.c b/src/lib/libcrypto/evp/e_old.c
new file mode 100644
index 0000000000..92dc498945
--- /dev/null
+++ b/src/lib/libcrypto/evp/e_old.c
@@ -0,0 +1,108 @@
+/* crypto/evp/e_old.c -*- mode:C; c-file-style: "eay" -*- */
+/* Written by Richard Levitte (richard@levitte.org) for the OpenSSL
+ * project 2004.
+ */
+/* ====================================================================
+ * Copyright (c) 2004 The OpenSSL Project.  All rights reserved.
+ *
+ * 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 above 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 acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED 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 OpenSSL PROJECT OR
+ * ITS 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.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <openssl/evp.h>
+
+/* Define some deprecated functions, so older programs
+   don't crash and burn too quickly.  On Windows and VMS,
+   these will never be used, since functions and variables
+   in shared libraries are selected by entry point location,
+   not by name.  */
+
+#ifndef OPENSSL_NO_BF
+#undef EVP_bf_cfb
+const EVP_CIPHER *EVP_bf_cfb(void) { return EVP_bf_cfb64(); }
+#endif
+
+#ifndef OPENSSL_NO_DES
+#undef EVP_des_cfb
+const EVP_CIPHER *EVP_des_cfb(void) { return EVP_des_cfb64(); }
+#undef EVP_des_ede3_cfb
+const EVP_CIPHER *EVP_des_ede3_cfb(void) { return EVP_des_ede3_cfb64(); }
+#undef EVP_des_ede_cfb
+const EVP_CIPHER *EVP_des_ede_cfb(void) { return EVP_des_ede_cfb64(); }
+#endif
+
+#ifndef OPENSSL_NO_IDEA
+#undef EVP_idea_cfb
+const EVP_CIPHER *EVP_idea_cfb(void) { return EVP_idea_cfb64(); }
+#endif
+
+#ifndef OPENSSL_NO_RC2
+#undef EVP_rc2_cfb
+const EVP_CIPHER *EVP_rc2_cfb(void) { return EVP_rc2_cfb64(); }
+#endif
+
+#ifndef OPENSSL_NO_CAST
+#undef EVP_cast5_cfb
+const EVP_CIPHER *EVP_cast5_cfb(void) { return EVP_cast5_cfb64(); }
+#endif
+
+#ifndef OPENSSL_NO_RC5
+#undef EVP_rc5_32_12_16_cfb
+const EVP_CIPHER *EVP_rc5_32_12_16_cfb(void) { return EVP_rc5_32_12_16_cfb64(); }
+#endif
+
+#ifndef OPENSSL_NO_AES
+#undef EVP_aes_128_cfb
+const EVP_CIPHER *EVP_aes_128_cfb(void) { return EVP_aes_128_cfb128(); }
+#undef EVP_aes_192_cfb
+const EVP_CIPHER *EVP_aes_192_cfb(void) { return EVP_aes_192_cfb128(); }
+#undef EVP_aes_256_cfb
+const EVP_CIPHER *EVP_aes_256_cfb(void) { return EVP_aes_256_cfb128(); }
+#endif
diff --git a/src/lib/libcrypto/o_str.c b/src/lib/libcrypto/o_str.c
new file mode 100644
index 0000000000..da8860491d
--- /dev/null
+++ b/src/lib/libcrypto/o_str.c
@@ -0,0 +1,96 @@
+/* crypto/o_str.c -*- mode:C; c-file-style: "eay" -*- */
+/* Written by Richard Levitte (richard@levitte.org) for the OpenSSL
+ * project 2003.
+ */
+/* ====================================================================
+ * Copyright (c) 2003 The OpenSSL Project.  All rights reserved.
+ *
+ * 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 above 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 acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
+ *
+ * 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
+ *    endorse or promote products derived from this software without
+ *    prior written permission. For written permission, please contact
+ *    openssl-core@openssl.org.
+ *
+ * 5. Products derived from this software may not be called "OpenSSL"
+ *    nor may "OpenSSL" appear in their names without prior written
+ *    permission of the OpenSSL Project.
+ *
+ * 6. Redistributions of any form whatsoever must retain the following
+ *    acknowledgment:
+ *    "This product includes software developed by the OpenSSL Project
+ *    for use in the OpenSSL Toolkit (http://www.openssl.org/)"
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
+ * EXPRESSED 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 OpenSSL PROJECT OR
+ * ITS 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.
+ * ====================================================================
+ *
+ * This product includes cryptographic software written by Eric Young
+ * (eay@cryptsoft.com).  This product includes software written by Tim
+ * Hudson (tjh@cryptsoft.com).
+ *
+ */
+
+#include <ctype.h>
+#include <e_os.h>
+#include "o_str.h"
+
+int OPENSSL_strncasecmp(const char *str1, const char *str2, size_t n)
+        {
+#if defined(OPENSSL_IMPLEMENTS_strncasecmp)
+        while (*str1 && *str2 && n)
+                {
+                int res = toupper(*str1) - toupper(*str2);
+                if (res) return res < 0 ? -1 : 1;
+                str1++;
+                str2++;
+                n--;
+                }
+        if (n == 0)
+                return 0;
+        if (*str1)
+                return 1;
+        if (*str2)
+                return -1;
+        return 0;
+#else
+        /* Recursion hazard warning! Whenever strncasecmp is #defined as
+         * OPENSSL_strncasecmp, OPENSSL_IMPLEMENTS_strncasecmp must be
+         * defined as well. */
+        return strncasecmp(str1, str2, n);
+#endif
+        }
+int OPENSSL_strcasecmp(const char *str1, const char *str2)
+        {
+#if defined(OPENSSL_IMPLEMENTS_strncasecmp)
+        return OPENSSL_strncasecmp(str1, str2, (size_t)-1);
+#else
+        return strcasecmp(str1, str2);
+#endif
+        }
+
diff --git a/src/lib/libcrypto/sha/asm/sha1-ia64.pl b/src/lib/libcrypto/sha/asm/sha1-ia64.pl
new file mode 100644
index 0000000000..cb9dfad124
--- /dev/null
+++ b/src/lib/libcrypto/sha/asm/sha1-ia64.pl
@@ -0,0 +1,549 @@
+#!/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.
+# ====================================================================
+#
+# 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. As of this moment
+# performance under little-endian OS such as Linux and Windows will be
+# a bit lower, because data has to be picked in reverse byte-order.
+# It's possible to resolve this issue by implementing third function,
+# sha1_block_asm_data_order_aligned, which would temporarily flip
+# BE field in User Mask register...
+
+$code=<<___;
+.ident  \"sha1-ia64.s, version 1.0\"
+.ident  \"IA-64 ISA artwork by Andy Polyakov <appro\@fy.chalmers.se>\"
+.explicit
+
+___
+
+
+if ($^O eq "hpux") {
+    $ADDP="addp4";
+    for (@ARGV) { $ADDP="add" if (/[\+DD|\-mlp]64/); }
+} else { $ADDP="add"; }
+for (@ARGV) {   $big_endian=1 if (/\-DB_ENDIAN/);
+                $big_endian=0 if (/\-DL_ENDIAN/);   }
+if (!defined($big_endian))
+            {   $big_endian=(unpack('L',pack('N',1))==1);   }
+
+#$human=1;
+if ($human) {   # useful for visual code auditing...
+        ($A,$B,$C,$D,$E,$T)   = ("A","B","C","D","E","T");
+        ($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,$T)   = ("loc0","loc1","loc2","loc3","loc4","loc5");
+        ($h0,$h1,$h2,$h3,$h4) = ("loc6","loc7","loc8","loc9","loc10");
+        ($K_00_19, $K_20_39, $K_40_59, $K_60_79) =
+            (   "r14", "r15", "loc11", "loc12"  );
+        @X= (   "r16", "r17", "r18", "r19", "r20", "r21", "r22", "r23",
+                "r24", "r25", "r26", "r27", "r28", "r29", "r30", "r31"  );
+}
+
+sub BODY_00_15 {
+local   *code=shift;
+local   ($i,$a,$b,$c,$d,$e,$f,$unaligned)=@_;
+
+if ($unaligned) {
+        $code.=<<___;
+{ .mmi; ld1     tmp0=[inp],2                // MSB
+        ld1     tmp1=[tmp3],2           };;
+{ .mmi; ld1     tmp2=[inp],2
+        ld1     $X[$i&0xf]=[tmp3],2         // LSB
+        dep     tmp1=tmp0,tmp1,8,8      };;
+{ .mii; cmp.ne  p16,p0=r0,r0                // no misaligned prefetch
+        dep     $X[$i&0xf]=tmp2,$X[$i&0xf],8,8;;
+        dep     $X[$i&0xf]=tmp1,$X[$i&0xf],16,16        };;
+{ .mmi; nop.m   0
+___
+        }
+elsif ($i<15) {
+        $code.=<<___;
+{ .mmi; ld4     $X[($i+1)&0xf]=[inp],4  // prefetch
+___
+        }
+else    {
+        $code.=<<___;
+{ .mmi; nop.m   0
+___
+        }
+if ($i<15) {
+        $code.=<<___;
+        and     tmp0=$c,$b
+        dep.z   tmp5=$a,5,27            }   // a<<5
+{ .mmi; andcm   tmp1=$d,$b
+        add     tmp4=$e,$K_00_19        };;
+{ .mmi; or      tmp0=tmp0,tmp1              // F_00_19(b,c,d)=(b&c)|(~b&d)
+        add     $f=tmp4,$X[$i&0xf]          // f=xi+e+K_00_19
+        extr.u  tmp1=$a,27,5            };; // a>>27
+{ .mib; add     $f=$f,tmp0                  // f+=F_00_19(b,c,d)
+        shrp    $b=tmp6,tmp6,2          }   // b=ROTATE(b,30)
+{ .mib; or      tmp1=tmp1,tmp5              // ROTATE(a,5)
+        mux2    tmp6=$a,0x44            };; // see b in next iteration
+{ .mii; add     $f=$f,tmp1                  // f+=ROTATE(a,5)
+        mux2    $X[$i&0xf]=$X[$i&0xf],0x44
+        nop.i   0                       };;
+
+___
+        }
+else    {
+        $code.=<<___;
+        and     tmp0=$c,$b
+        dep.z   tmp5=$a,5,27            }   // a<<5 ;;?
+{ .mmi; andcm   tmp1=$d,$b
+        add     tmp4=$e,$K_00_19        };;
+{ .mmi; or      tmp0=tmp0,tmp1              // F_00_19(b,c,d)=(b&c)|(~b&d)
+        add     $f=tmp4,$X[$i&0xf]          // f=xi+e+K_00_19
+        extr.u  tmp1=$a,27,5            }   // a>>27
+{ .mmi; xor     tmp2=$X[($i+0+1)&0xf],$X[($i+2+1)&0xf]  // +1
+        xor     tmp3=$X[($i+8+1)&0xf],$X[($i+13+1)&0xf] // +1
+        nop.i   0                       };;
+{ .mmi; add     $f=$f,tmp0                  // f+=F_00_19(b,c,d)
+        xor     tmp2=tmp2,tmp3              // +1
+        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     $f=$f,tmp1                  // f+=ROTATE(a,5)
+        shrp    $e=tmp2,tmp2,31             // f+1=ROTATE(x[0]^x[2]^x[8]^x[13],1)
+        mux2    $X[$i&0xf]=$X[$i&0xf],0x44  };;
+
+___
+        }
+}
+
+sub BODY_16_19 {
+local   *code=shift;
+local   ($i,$a,$b,$c,$d,$e,$f)=@_;
+
+$code.=<<___;
+{ .mmi; mov     $X[$i&0xf]=$f               // Xupdate
+        and     tmp0=$c,$b
+        dep.z   tmp5=$a,5,27            }   // a<<5
+{ .mmi; andcm   tmp1=$d,$b
+        add     tmp4=$e,$K_00_19        };;
+{ .mmi; or      tmp0=tmp0,tmp1              // F_00_19(b,c,d)=(b&c)|(~b&d)
+        add     $f=$f,tmp4                  // f+=e+K_00_19
+        extr.u  tmp1=$a,27,5            }   // a>>27
+{ .mmi; xor     tmp2=$X[($i+0+1)&0xf],$X[($i+2+1)&0xf]  // +1
+        xor     tmp3=$X[($i+8+1)&0xf],$X[($i+13+1)&0xf] // +1
+        nop.i   0                       };;
+{ .mmi; add     $f=$f,tmp0                  // f+=F_00_19(b,c,d)
+        xor     tmp2=tmp2,tmp3              // +1
+        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     $f=$f,tmp1                  // f+=ROTATE(a,5)
+        shrp    $e=tmp2,tmp2,31             // f+1=ROTATE(x[0]^x[2]^x[8]^x[13],1)
+        nop.i   0                       };;
+
+___
+}
+
+sub BODY_20_39 {
+local   *code=shift;
+local   ($i,$a,$b,$c,$d,$e,$f,$Konst)=@_;
+        $Konst = $K_20_39 if (!defined($Konst));
+
+if ($i<79) {
+$code.=<<___;
+{ .mib; mov     $X[$i&0xf]=$f               // Xupdate
+        dep.z   tmp5=$a,5,27            }   // a<<5
+{ .mib; xor     tmp0=$c,$b
+        add     tmp4=$e,$Konst          };;
+{ .mmi; xor     tmp0=tmp0,$d                // F_20_39(b,c,d)=b^c^d
+        add     $f=$f,tmp4                  // f+=e+K_20_39
+        extr.u  tmp1=$a,27,5            }   // a>>27
+{ .mmi; xor     tmp2=$X[($i+0+1)&0xf],$X[($i+2+1)&0xf]  // +1
+        xor     tmp3=$X[($i+8+1)&0xf],$X[($i+13+1)&0xf] // +1
+        nop.i   0                       };;
+{ .mmi; add     $f=$f,tmp0                  // f+=F_20_39(b,c,d)
+        xor     tmp2=tmp2,tmp3              // +1
+        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     $f=$f,tmp1                  // f+=ROTATE(a,5)
+        shrp    $e=tmp2,tmp2,31             // f+1=ROTATE(x[0]^x[2]^x[8]^x[13],1)
+        nop.i   0                       };;
+
+___
+}
+else {
+$code.=<<___;
+{ .mib; mov     $X[$i&0xf]=$f               // Xupdate
+        dep.z   tmp5=$a,5,27            }   // a<<5
+{ .mib; xor     tmp0=$c,$b
+        add     tmp4=$e,$Konst          };;
+{ .mib; xor     tmp0=tmp0,$d                // F_20_39(b,c,d)=b^c^d
+        extr.u  tmp1=$a,27,5            }   // a>>27
+{ .mib; add     $f=$f,tmp4                  // f+=e+K_20_39
+        add     $h1=$h1,$a              };; // wrap up
+{ .mmi;
+(p16)   ld4.s   $X[0]=[inp],4               // non-faulting prefetch
+        add     $f=$f,tmp0                  // f+=F_20_39(b,c,d)
+        shrp    $b=tmp6,tmp6,2          }   // b=ROTATE(b,30) ;;?
+{ .mmi; or      tmp1=tmp1,tmp5              // ROTATE(a,5)
+        add     $h3=$h3,$c              };; // wrap up
+{ .mib; add     tmp3=1,inp                  // used in unaligned codepath
+        add     $f=$f,tmp1              }   // f+=ROTATE(a,5)
+{ .mib; add     $h2=$h2,$b                  // wrap up
+        add     $h4=$h4,$d              };; // wrap up
+
+___
+}
+}
+
+sub BODY_40_59 {
+local   *code=shift;
+local   ($i,$a,$b,$c,$d,$e,$f)=@_;
+
+$code.=<<___;
+{ .mmi; mov     $X[$i&0xf]=$f               // Xupdate
+        and     tmp0=$c,$b
+        dep.z   tmp5=$a,5,27            }   // a<<5
+{ .mmi; and     tmp1=$d,$b
+        add     tmp4=$e,$K_40_59        };;
+{ .mmi; or      tmp0=tmp0,tmp1              // (b&c)|(b&d)
+        add     $f=$f,tmp4                  // f+=e+K_40_59
+        extr.u  tmp1=$a,27,5            }   // a>>27
+{ .mmi; and     tmp4=$c,$d
+        xor     tmp2=$X[($i+0+1)&0xf],$X[($i+2+1)&0xf]  // +1
+        xor     tmp3=$X[($i+8+1)&0xf],$X[($i+13+1)&0xf] // +1
+        };;
+{ .mmi; or      tmp1=tmp1,tmp5              // ROTATE(a,5)
+        xor     tmp2=tmp2,tmp3              // +1
+        shrp    $b=tmp6,tmp6,2          }   // b=ROTATE(b,30)
+{ .mmi; or      tmp0=tmp0,tmp4              // F_40_59(b,c,d)=(b&c)|(b&d)|(c&d)
+        mux2    tmp6=$a,0x44            };; // see b in next iteration
+{ .mii; add     $f=$f,tmp0                  // f+=F_40_59(b,c,d)
+        shrp    $e=tmp2,tmp2,31;;           // f+1=ROTATE(x[0]^x[2]^x[8]^x[13],1)
+        add     $f=$f,tmp1              };; // f+=ROTATE(a,5)
+
+___
+}
+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_asm_host_order(SHA_CTX *c,const void *p,size_t num);
+.global sha1_block_asm_host_order#
+.proc   sha1_block_asm_host_order#
+.align  32
+sha1_block_asm_host_order:
+        .prologue
+        .fframe 0
+        .save   ar.pfs,r0
+        .save   ar.lc,r3
+{ .mmi; alloc   tmp1=ar.pfs,3,15,0,0
+        $ADDP   tmp0=4,ctx
+        mov     r3=ar.lc                }
+{ .mmi; $ADDP   ctx=0,ctx
+        $ADDP   inp=0,inp
+        mov     r2=pr                   };;
+tmp4=in2;
+tmp5=loc13;
+tmp6=loc14;
+        .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; ld4     $X[0]=[inp],4               // prefetch
+        cmp.ne  p16,p0=r0,in2               // prefecth at loop end
+        mov     ar.lc=in2               };; // brp.loop.imp: too far
+
+.Lhtop:
+{ .mmi; mov     $A=$h0
+        mov     $B=$h1
+        mux2    tmp6=$h1,0x44           }
+{ .mmi; mov     $C=$h2
+        mov     $D=$h3
+        mov     $E=$h4                  };;
+
+___
+
+        &BODY_00_15(\$code, 0,$A,$B,$C,$D,$E,$T);
+        &BODY_00_15(\$code, 1,$T,$A,$B,$C,$D,$E);
+        &BODY_00_15(\$code, 2,$E,$T,$A,$B,$C,$D);
+        &BODY_00_15(\$code, 3,$D,$E,$T,$A,$B,$C);
+        &BODY_00_15(\$code, 4,$C,$D,$E,$T,$A,$B);
+        &BODY_00_15(\$code, 5,$B,$C,$D,$E,$T,$A);
+        &BODY_00_15(\$code, 6,$A,$B,$C,$D,$E,$T);
+        &BODY_00_15(\$code, 7,$T,$A,$B,$C,$D,$E);
+        &BODY_00_15(\$code, 8,$E,$T,$A,$B,$C,$D);
+        &BODY_00_15(\$code, 9,$D,$E,$T,$A,$B,$C);
+        &BODY_00_15(\$code,10,$C,$D,$E,$T,$A,$B);
+        &BODY_00_15(\$code,11,$B,$C,$D,$E,$T,$A);
+        &BODY_00_15(\$code,12,$A,$B,$C,$D,$E,$T);
+        &BODY_00_15(\$code,13,$T,$A,$B,$C,$D,$E);
+        &BODY_00_15(\$code,14,$E,$T,$A,$B,$C,$D);
+        &BODY_00_15(\$code,15,$D,$E,$T,$A,$B,$C);
+
+        &BODY_16_19(\$code,16,$C,$D,$E,$T,$A,$B);
+        &BODY_16_19(\$code,17,$B,$C,$D,$E,$T,$A);
+        &BODY_16_19(\$code,18,$A,$B,$C,$D,$E,$T);
+        &BODY_16_19(\$code,19,$T,$A,$B,$C,$D,$E);
+
+        &BODY_20_39(\$code,20,$E,$T,$A,$B,$C,$D);
+        &BODY_20_39(\$code,21,$D,$E,$T,$A,$B,$C);
+        &BODY_20_39(\$code,22,$C,$D,$E,$T,$A,$B);
+        &BODY_20_39(\$code,23,$B,$C,$D,$E,$T,$A);
+        &BODY_20_39(\$code,24,$A,$B,$C,$D,$E,$T);
+        &BODY_20_39(\$code,25,$T,$A,$B,$C,$D,$E);
+        &BODY_20_39(\$code,26,$E,$T,$A,$B,$C,$D);
+        &BODY_20_39(\$code,27,$D,$E,$T,$A,$B,$C);
+        &BODY_20_39(\$code,28,$C,$D,$E,$T,$A,$B);
+        &BODY_20_39(\$code,29,$B,$C,$D,$E,$T,$A);
+        &BODY_20_39(\$code,30,$A,$B,$C,$D,$E,$T);
+        &BODY_20_39(\$code,31,$T,$A,$B,$C,$D,$E);
+        &BODY_20_39(\$code,32,$E,$T,$A,$B,$C,$D);
+        &BODY_20_39(\$code,33,$D,$E,$T,$A,$B,$C);
+        &BODY_20_39(\$code,34,$C,$D,$E,$T,$A,$B);
+        &BODY_20_39(\$code,35,$B,$C,$D,$E,$T,$A);
+        &BODY_20_39(\$code,36,$A,$B,$C,$D,$E,$T);
+        &BODY_20_39(\$code,37,$T,$A,$B,$C,$D,$E);
+        &BODY_20_39(\$code,38,$E,$T,$A,$B,$C,$D);
+        &BODY_20_39(\$code,39,$D,$E,$T,$A,$B,$C);
+
+        &BODY_40_59(\$code,40,$C,$D,$E,$T,$A,$B);
+        &BODY_40_59(\$code,41,$B,$C,$D,$E,$T,$A);
+        &BODY_40_59(\$code,42,$A,$B,$C,$D,$E,$T);
+        &BODY_40_59(\$code,43,$T,$A,$B,$C,$D,$E);
+        &BODY_40_59(\$code,44,$E,$T,$A,$B,$C,$D);
+        &BODY_40_59(\$code,45,$D,$E,$T,$A,$B,$C);
+        &BODY_40_59(\$code,46,$C,$D,$E,$T,$A,$B);
+        &BODY_40_59(\$code,47,$B,$C,$D,$E,$T,$A);
+        &BODY_40_59(\$code,48,$A,$B,$C,$D,$E,$T);
+        &BODY_40_59(\$code,49,$T,$A,$B,$C,$D,$E);
+        &BODY_40_59(\$code,50,$E,$T,$A,$B,$C,$D);
+        &BODY_40_59(\$code,51,$D,$E,$T,$A,$B,$C);
+        &BODY_40_59(\$code,52,$C,$D,$E,$T,$A,$B);
+        &BODY_40_59(\$code,53,$B,$C,$D,$E,$T,$A);
+        &BODY_40_59(\$code,54,$A,$B,$C,$D,$E,$T);
+        &BODY_40_59(\$code,55,$T,$A,$B,$C,$D,$E);
+        &BODY_40_59(\$code,56,$E,$T,$A,$B,$C,$D);
+        &BODY_40_59(\$code,57,$D,$E,$T,$A,$B,$C);
+        &BODY_40_59(\$code,58,$C,$D,$E,$T,$A,$B);
+        &BODY_40_59(\$code,59,$B,$C,$D,$E,$T,$A);
+
+        &BODY_60_79(\$code,60,$A,$B,$C,$D,$E,$T);
+        &BODY_60_79(\$code,61,$T,$A,$B,$C,$D,$E);
+        &BODY_60_79(\$code,62,$E,$T,$A,$B,$C,$D);
+        &BODY_60_79(\$code,63,$D,$E,$T,$A,$B,$C);
+        &BODY_60_79(\$code,64,$C,$D,$E,$T,$A,$B);
+        &BODY_60_79(\$code,65,$B,$C,$D,$E,$T,$A);
+        &BODY_60_79(\$code,66,$A,$B,$C,$D,$E,$T);
+        &BODY_60_79(\$code,67,$T,$A,$B,$C,$D,$E);
+        &BODY_60_79(\$code,68,$E,$T,$A,$B,$C,$D);
+        &BODY_60_79(\$code,69,$D,$E,$T,$A,$B,$C);
+        &BODY_60_79(\$code,70,$C,$D,$E,$T,$A,$B);
+        &BODY_60_79(\$code,71,$B,$C,$D,$E,$T,$A);
+        &BODY_60_79(\$code,72,$A,$B,$C,$D,$E,$T);
+        &BODY_60_79(\$code,73,$T,$A,$B,$C,$D,$E);
+        &BODY_60_79(\$code,74,$E,$T,$A,$B,$C,$D);
+        &BODY_60_79(\$code,75,$D,$E,$T,$A,$B,$C);
+        &BODY_60_79(\$code,76,$C,$D,$E,$T,$A,$B);
+        &BODY_60_79(\$code,77,$B,$C,$D,$E,$T,$A);
+        &BODY_60_79(\$code,78,$A,$B,$C,$D,$E,$T);
+        &BODY_60_79(\$code,79,$T,$A,$B,$C,$D,$E);
+
+$code.=<<___;
+{ .mmb; add     $h0=$h0,$E
+        nop.m   0
+        br.ctop.dptk.many       .Lhtop  };;
+.Lhend:
+{ .mmi; add     tmp0=4,ctx
+        mov     ar.lc=r3                };;
+{ .mmi; st4     [ctx]=$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_asm_host_order#
+___
+
+
+$code.=<<___;
+// void sha1_block_asm_data_order(SHA_CTX *c,const void *p,size_t num);
+.global sha1_block_asm_data_order#
+.proc   sha1_block_asm_data_order#
+.align  32
+sha1_block_asm_data_order:
+___
+$code.=<<___ if ($big_endian);
+{ .mmi; and     r2=3,inp                                };;
+{ .mib; cmp.eq  p6,p0=r0,r2
+(p6)    br.dptk.many    sha1_block_asm_host_order       };;
+___
+$code.=<<___;
+        .prologue
+        .fframe 0
+        .save   ar.pfs,r0
+        .save   ar.lc,r3
+{ .mmi; alloc   tmp1=ar.pfs,3,15,0,0
+        $ADDP   tmp0=4,ctx
+        mov     r3=ar.lc                }
+{ .mmi; $ADDP   ctx=0,ctx
+        $ADDP   inp=0,inp
+        mov     r2=pr                   };;
+tmp4=in2;
+tmp5=loc13;
+tmp6=loc14;
+        .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                  };;
+
+___
+
+        &BODY_00_15(\$code, 0,$A,$B,$C,$D,$E,$T,1);
+        &BODY_00_15(\$code, 1,$T,$A,$B,$C,$D,$E,1);
+        &BODY_00_15(\$code, 2,$E,$T,$A,$B,$C,$D,1);
+        &BODY_00_15(\$code, 3,$D,$E,$T,$A,$B,$C,1);
+        &BODY_00_15(\$code, 4,$C,$D,$E,$T,$A,$B,1);
+        &BODY_00_15(\$code, 5,$B,$C,$D,$E,$T,$A,1);
+        &BODY_00_15(\$code, 6,$A,$B,$C,$D,$E,$T,1);
+        &BODY_00_15(\$code, 7,$T,$A,$B,$C,$D,$E,1);
+        &BODY_00_15(\$code, 8,$E,$T,$A,$B,$C,$D,1);
+        &BODY_00_15(\$code, 9,$D,$E,$T,$A,$B,$C,1);
+        &BODY_00_15(\$code,10,$C,$D,$E,$T,$A,$B,1);
+        &BODY_00_15(\$code,11,$B,$C,$D,$E,$T,$A,1);
+        &BODY_00_15(\$code,12,$A,$B,$C,$D,$E,$T,1);
+        &BODY_00_15(\$code,13,$T,$A,$B,$C,$D,$E,1);
+        &BODY_00_15(\$code,14,$E,$T,$A,$B,$C,$D,1);
+        &BODY_00_15(\$code,15,$D,$E,$T,$A,$B,$C,1);
+
+        &BODY_16_19(\$code,16,$C,$D,$E,$T,$A,$B);
+        &BODY_16_19(\$code,17,$B,$C,$D,$E,$T,$A);
+        &BODY_16_19(\$code,18,$A,$B,$C,$D,$E,$T);
+        &BODY_16_19(\$code,19,$T,$A,$B,$C,$D,$E);
+
+        &BODY_20_39(\$code,20,$E,$T,$A,$B,$C,$D);
+        &BODY_20_39(\$code,21,$D,$E,$T,$A,$B,$C);
+        &BODY_20_39(\$code,22,$C,$D,$E,$T,$A,$B);
+        &BODY_20_39(\$code,23,$B,$C,$D,$E,$T,$A);
+        &BODY_20_39(\$code,24,$A,$B,$C,$D,$E,$T);
+        &BODY_20_39(\$code,25,$T,$A,$B,$C,$D,$E);
+        &BODY_20_39(\$code,26,$E,$T,$A,$B,$C,$D);
+        &BODY_20_39(\$code,27,$D,$E,$T,$A,$B,$C);
+        &BODY_20_39(\$code,28,$C,$D,$E,$T,$A,$B);
+        &BODY_20_39(\$code,29,$B,$C,$D,$E,$T,$A);
+        &BODY_20_39(\$code,30,$A,$B,$C,$D,$E,$T);
+        &BODY_20_39(\$code,31,$T,$A,$B,$C,$D,$E);
+        &BODY_20_39(\$code,32,$E,$T,$A,$B,$C,$D);
+        &BODY_20_39(\$code,33,$D,$E,$T,$A,$B,$C);
+        &BODY_20_39(\$code,34,$C,$D,$E,$T,$A,$B);
+        &BODY_20_39(\$code,35,$B,$C,$D,$E,$T,$A);
+        &BODY_20_39(\$code,36,$A,$B,$C,$D,$E,$T);
+        &BODY_20_39(\$code,37,$T,$A,$B,$C,$D,$E);
+        &BODY_20_39(\$code,38,$E,$T,$A,$B,$C,$D);
+        &BODY_20_39(\$code,39,$D,$E,$T,$A,$B,$C);
+
+        &BODY_40_59(\$code,40,$C,$D,$E,$T,$A,$B);
+        &BODY_40_59(\$code,41,$B,$C,$D,$E,$T,$A);
+        &BODY_40_59(\$code,42,$A,$B,$C,$D,$E,$T);
+        &BODY_40_59(\$code,43,$T,$A,$B,$C,$D,$E);
+        &BODY_40_59(\$code,44,$E,$T,$A,$B,$C,$D);
+        &BODY_40_59(\$code,45,$D,$E,$T,$A,$B,$C);
+        &BODY_40_59(\$code,46,$C,$D,$E,$T,$A,$B);
+        &BODY_40_59(\$code,47,$B,$C,$D,$E,$T,$A);
+        &BODY_40_59(\$code,48,$A,$B,$C,$D,$E,$T);
+        &BODY_40_59(\$code,49,$T,$A,$B,$C,$D,$E);
+        &BODY_40_59(\$code,50,$E,$T,$A,$B,$C,$D);
+        &BODY_40_59(\$code,51,$D,$E,$T,$A,$B,$C);
+        &BODY_40_59(\$code,52,$C,$D,$E,$T,$A,$B);
+        &BODY_40_59(\$code,53,$B,$C,$D,$E,$T,$A);
+        &BODY_40_59(\$code,54,$A,$B,$C,$D,$E,$T);
+        &BODY_40_59(\$code,55,$T,$A,$B,$C,$D,$E);
+        &BODY_40_59(\$code,56,$E,$T,$A,$B,$C,$D);
+        &BODY_40_59(\$code,57,$D,$E,$T,$A,$B,$C);
+        &BODY_40_59(\$code,58,$C,$D,$E,$T,$A,$B);
+        &BODY_40_59(\$code,59,$B,$C,$D,$E,$T,$A);
+
+        &BODY_60_79(\$code,60,$A,$B,$C,$D,$E,$T);
+        &BODY_60_79(\$code,61,$T,$A,$B,$C,$D,$E);
+        &BODY_60_79(\$code,62,$E,$T,$A,$B,$C,$D);
+        &BODY_60_79(\$code,63,$D,$E,$T,$A,$B,$C);
+        &BODY_60_79(\$code,64,$C,$D,$E,$T,$A,$B);
+        &BODY_60_79(\$code,65,$B,$C,$D,$E,$T,$A);
+        &BODY_60_79(\$code,66,$A,$B,$C,$D,$E,$T);
+        &BODY_60_79(\$code,67,$T,$A,$B,$C,$D,$E);
+        &BODY_60_79(\$code,68,$E,$T,$A,$B,$C,$D);
+        &BODY_60_79(\$code,69,$D,$E,$T,$A,$B,$C);
+        &BODY_60_79(\$code,70,$C,$D,$E,$T,$A,$B);
+        &BODY_60_79(\$code,71,$B,$C,$D,$E,$T,$A);
+        &BODY_60_79(\$code,72,$A,$B,$C,$D,$E,$T);
+        &BODY_60_79(\$code,73,$T,$A,$B,$C,$D,$E);
+        &BODY_60_79(\$code,74,$E,$T,$A,$B,$C,$D);
+        &BODY_60_79(\$code,75,$D,$E,$T,$A,$B,$C);
+        &BODY_60_79(\$code,76,$C,$D,$E,$T,$A,$B);
+        &BODY_60_79(\$code,77,$B,$C,$D,$E,$T,$A);
+        &BODY_60_79(\$code,78,$A,$B,$C,$D,$E,$T);
+        &BODY_60_79(\$code,79,$T,$A,$B,$C,$D,$E);
+
+$code.=<<___;
+{ .mmb; add     $h0=$h0,$E
+        nop.m   0
+        br.ctop.dptk.many       .Ldtop  };;
+.Ldend:
+{ .mmi; add     tmp0=4,ctx
+        mov     ar.lc=r3                };;
+{ .mmi; st4     [ctx]=$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_asm_data_order#
+___
+
+print $code;
diff --git a/src/lib/libcrypto/x509v3/v3_pci.c b/src/lib/libcrypto/x509v3/v3_pci.c
new file mode 100644
index 0000000000..b32d968619
--- /dev/null
+++ b/src/lib/libcrypto/x509v3/v3_pci.c
@@ -0,0 +1,313 @@
+/* v3_pci.c -*- mode:C; c-file-style: "eay" -*- */
+/* Contributed to the OpenSSL Project 2004
+ * by Richard Levitte (richard@levitte.org)
+ */
+/* Copyright (c) 2004 Kungliga Tekniska Högskolan
+ * (Royal Institute of Technology, Stockholm, Sweden).
+ * All rights reserved.
+ *
+ * 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 above 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. Neither the name of the Institute nor the names of its contributors
+ *    may be used to endorse or promote products derived from this software
+ *    without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE INSTITUTE AND CONTRIBUTORS ``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 INSTITUTE 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.
+ */
+
+#include <stdio.h>
+#include "cryptlib.h"
+#include <openssl/conf.h>
+#include <openssl/x509v3.h>
+
+static int i2r_pci(X509V3_EXT_METHOD *method, PROXY_CERT_INFO_EXTENSION *ext,
+        BIO *out, int indent);
+static PROXY_CERT_INFO_EXTENSION *r2i_pci(X509V3_EXT_METHOD *method,
+        X509V3_CTX *ctx, char *str);
+
+X509V3_EXT_METHOD v3_pci =
+        { NID_proxyCertInfo, 0, ASN1_ITEM_ref(PROXY_CERT_INFO_EXTENSION),
+          0,0,0,0,
+          0,0,
+          NULL, NULL,
+          (X509V3_EXT_I2R)i2r_pci,
+          (X509V3_EXT_R2I)r2i_pci,
+          NULL,
+        };
+
+static int i2r_pci(X509V3_EXT_METHOD *method, PROXY_CERT_INFO_EXTENSION *pci,
+        BIO *out, int indent)
+        {
+        BIO_printf(out, "%*sPath Length Constraint: ", indent, "");
+        if (pci->pcPathLengthConstraint)
+          i2a_ASN1_INTEGER(out, pci->pcPathLengthConstraint);
+        else
+          BIO_printf(out, "infinite");
+        BIO_puts(out, "\n");
+        BIO_printf(out, "%*sPolicy Language: ", indent, "");
+        i2a_ASN1_OBJECT(out, pci->proxyPolicy->policyLanguage);
+        BIO_puts(out, "\n");
+        if (pci->proxyPolicy->policy && pci->proxyPolicy->policy->data)
+          BIO_printf(out, "%*sPolicy Text: %s\n", indent, "",
+                     pci->proxyPolicy->policy->data);
+        return 1;
+        }
+
+static int process_pci_value(CONF_VALUE *val,
+        ASN1_OBJECT **language, ASN1_INTEGER **pathlen,
+        ASN1_OCTET_STRING **policy)
+        {
+        int free_policy = 0;
+
+        if (strcmp(val->name, "language") == 0)
+                {
+                if (*language)
+                        {
+                        X509V3err(X509V3_F_R2I_PCI,X509V3_R_POLICY_LANGUAGE_ALREADTY_DEFINED);
+                        X509V3_conf_err(val);
+                        return 0;
+                        }
+                if (!(*language = OBJ_txt2obj(val->value, 0)))
+                        {
+                        X509V3err(X509V3_F_R2I_PCI,X509V3_R_INVALID_OBJECT_IDENTIFIER);
+                        X509V3_conf_err(val);
+                        return 0;
+                        }
+                }
+        else if (strcmp(val->name, "pathlen") == 0)
+                {
+                if (*pathlen)
+                        {
+                        X509V3err(X509V3_F_R2I_PCI,X509V3_R_POLICY_PATH_LENGTH_ALREADTY_DEFINED);
+                        X509V3_conf_err(val);
+                        return 0;
+                        }
+                if (!X509V3_get_value_int(val, pathlen))
+                        {
+                        X509V3err(X509V3_F_R2I_PCI,X509V3_R_POLICY_PATH_LENGTH);
+                        X509V3_conf_err(val);
+                        return 0;
+                        }
+                }
+        else if (strcmp(val->name, "policy") == 0)
+                {
+                unsigned char *tmp_data = NULL;
+                long val_len;
+                if (!*policy)
+                        {
+                        *policy = ASN1_OCTET_STRING_new();
+                        if (!*policy)
+                                {
+                                X509V3err(X509V3_F_R2I_PCI,ERR_R_MALLOC_FAILURE);
+                                X509V3_conf_err(val);
+                                return 0;
+                                }
+                        free_policy = 1;
+                        }
+                if (strncmp(val->value, "hex:", 4) == 0)
+                        {
+                        unsigned char *tmp_data2 =
+                                string_to_hex(val->value + 4, &val_len);
+
+                        if (!tmp_data2) goto err;
+
+                        tmp_data = OPENSSL_realloc((*policy)->data,
+                                (*policy)->length + val_len + 1);
+                        if (tmp_data)
+                                {
+                                (*policy)->data = tmp_data;
+                                memcpy(&(*policy)->data[(*policy)->length],
+                                        tmp_data2, val_len);
+                                (*policy)->length += val_len;
+                                (*policy)->data[(*policy)->length] = '\0';
+                                }
+                        }
+                else if (strncmp(val->value, "file:", 5) == 0)
+                        {
+                        unsigned char buf[2048];
+                        int n;
+                        BIO *b = BIO_new_file(val->value + 5, "r");
+                        if (!b)
+                                {
+                                X509V3err(X509V3_F_R2I_PCI,ERR_R_BIO_LIB);
+                                X509V3_conf_err(val);
+                                goto err;
+                                }
+                        while((n = BIO_read(b, buf, sizeof(buf))) > 0
+                                || (n == 0 && BIO_should_retry(b)))
+                                {
+                                if (!n) continue;
+
+                                tmp_data = OPENSSL_realloc((*policy)->data,
+                                        (*policy)->length + n + 1);
+
+                                if (!tmp_data)
+                                        break;
+
+                                (*policy)->data = tmp_data;
+                                memcpy(&(*policy)->data[(*policy)->length],
+                                        buf, n);
+                                (*policy)->length += n;
+                                (*policy)->data[(*policy)->length] = '\0';
+                                }
+
+                        if (n < 0)
+                                {
+                                X509V3err(X509V3_F_R2I_PCI,ERR_R_BIO_LIB);
+                                X509V3_conf_err(val);
+                                goto err;
+                                }
+                        }
+                else if (strncmp(val->value, "text:", 5) == 0)
+                        {
+                        val_len = strlen(val->value + 5);
+                        tmp_data = OPENSSL_realloc((*policy)->data,
+                                (*policy)->length + val_len + 1);
+                        if (tmp_data)
+                                {
+                                (*policy)->data = tmp_data;
+                                memcpy(&(*policy)->data[(*policy)->length],
+                                        val->value + 5, val_len);
+                                (*policy)->length += val_len;
+                                (*policy)->data[(*policy)->length] = '\0';
+                                }
+                        }
+                else
+                        {
+                        X509V3err(X509V3_F_R2I_PCI,X509V3_R_INCORRECT_POLICY_SYNTAX_TAG);
+                        X509V3_conf_err(val);
+                        goto err;
+                        }
+                if (!tmp_data)
+                        {
+                        X509V3err(X509V3_F_R2I_PCI,ERR_R_MALLOC_FAILURE);
+                        X509V3_conf_err(val);
+                        goto err;
+                        }
+                }
+        return 1;
+err:
+        if (free_policy)
+                {
+                ASN1_OCTET_STRING_free(*policy);
+                *policy = NULL;
+                }
+        return 0;
+        }
+
+static PROXY_CERT_INFO_EXTENSION *r2i_pci(X509V3_EXT_METHOD *method,
+        X509V3_CTX *ctx, char *value)
+        {
+        PROXY_CERT_INFO_EXTENSION *pci = NULL;
+        STACK_OF(CONF_VALUE) *vals;
+        ASN1_OBJECT *language = NULL;
+        ASN1_INTEGER *pathlen = NULL;
+        ASN1_OCTET_STRING *policy = NULL;
+        int i, j;
+
+        vals = X509V3_parse_list(value);
+        for (i = 0; i < sk_CONF_VALUE_num(vals); i++)
+                {
+                CONF_VALUE *cnf = sk_CONF_VALUE_value(vals, i);
+                if (!cnf->name || (*cnf->name != '@' && !cnf->value))
+                        {
+                        X509V3err(X509V3_F_R2I_PCI,X509V3_R_INVALID_PROXY_POLICY_SETTING);
+                        X509V3_conf_err(cnf);
+                        goto err;
+                        }
+                if (*cnf->name == '@')
+                        {
+                        STACK_OF(CONF_VALUE) *sect;
+                        int success_p = 1;
+
+                        sect = X509V3_get_section(ctx, cnf->name + 1);
+                        if (!sect)
+                                {
+                                X509V3err(X509V3_F_R2I_PCI,X509V3_R_INVALID_SECTION);
+                                X509V3_conf_err(cnf);
+                                goto err;
+                                }
+                        for (j = 0; success_p && j < sk_CONF_VALUE_num(sect); j++)
+                                {
+                                success_p =
+                                        process_pci_value(sk_CONF_VALUE_value(sect, j),
+                                                &language, &pathlen, &policy);
+                                }
+                        X509V3_section_free(ctx, sect);
+                        if (!success_p)
+                                goto err;
+                        }
+                else
+                        {
+                        if (!process_pci_value(cnf,
+                                        &language, &pathlen, &policy))
+                                {
+                                X509V3_conf_err(cnf);
+                                goto err;
+                                }
+                        }
+                }
+
+        /* Language is mandatory */
+        if (!language)
+                {
+                X509V3err(X509V3_F_R2I_PCI,X509V3_R_NO_PROXY_CERT_POLICY_LANGUAGE_DEFINED);
+                goto err;
+                }
+        i = OBJ_obj2nid(language);
+        if ((i == NID_Independent || i == NID_id_ppl_inheritAll) && policy)
+                {
+                X509V3err(X509V3_F_R2I_PCI,X509V3_R_POLICY_WHEN_PROXY_LANGUAGE_REQUIRES_NO_POLICY);
+                goto err;
+                }
+
+        pci = PROXY_CERT_INFO_EXTENSION_new();
+        if (!pci)
+                {
+                X509V3err(X509V3_F_R2I_PCI,ERR_R_MALLOC_FAILURE);
+                goto err;
+                }
+        pci->proxyPolicy = PROXY_POLICY_new();
+        if (!pci->proxyPolicy)
+                {
+                X509V3err(X509V3_F_R2I_PCI,ERR_R_MALLOC_FAILURE);
+                goto err;
+                }
+
+        pci->proxyPolicy->policyLanguage = language; language = NULL;
+        pci->proxyPolicy->policy = policy; policy = NULL;
+        pci->pcPathLengthConstraint = pathlen; pathlen = NULL;
+        goto end;
+err:
+        if (language) { ASN1_OBJECT_free(language); language = NULL; }
+        if (pathlen) { ASN1_INTEGER_free(pathlen); pathlen = NULL; }
+        if (policy) { ASN1_OCTET_STRING_free(policy); policy = NULL; }
+        if (pci && pci->proxyPolicy)
+                {
+                PROXY_POLICY_free(pci->proxyPolicy);
+                pci->proxyPolicy = NULL;
+                }
+        if (pci) { PROXY_CERT_INFO_EXTENSION_free(pci); pci = NULL; }
+end:
+        sk_CONF_VALUE_pop_free(vals, X509V3_conf_free);
+        return pci;
+        }
diff --git a/src/lib/libcrypto/x509v3/v3_pcia.c b/src/lib/libcrypto/x509v3/v3_pcia.c
new file mode 100644
index 0000000000..bb362e0e5a
--- /dev/null
+++ b/src/lib/libcrypto/x509v3/v3_pcia.c
@@ -0,0 +1,55 @@
+/* v3_pcia.c -*- mode:C; c-file-style: "eay" -*- */
+/* Contributed to the OpenSSL Project 2004
+ * by Richard Levitte (richard@levitte.org)
+ */
+/* Copyright (c) 2004 Kungliga Tekniska Högskolan
+ * (Royal Institute of Technology, Stockholm, Sweden).
+ * All rights reserved.
+ *
+ * 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 above 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. Neither the name of the Institute nor the names of its contributors
+ *    may be used to endorse or promote products derived from this software
+ *    without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE INSTITUTE AND CONTRIBUTORS ``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 INSTITUTE 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.
+ */
+
+#include <openssl/asn1.h>
+#include <openssl/asn1t.h>
+#include <openssl/x509v3.h>
+
+ASN1_SEQUENCE(PROXY_POLICY) =
+        {
+        ASN1_SIMPLE(PROXY_POLICY,policyLanguage,ASN1_OBJECT),
+        ASN1_OPT(PROXY_POLICY,policy,ASN1_OCTET_STRING)
+} ASN1_SEQUENCE_END(PROXY_POLICY)
+
+IMPLEMENT_ASN1_FUNCTIONS(PROXY_POLICY)
+
+ASN1_SEQUENCE(PROXY_CERT_INFO_EXTENSION) =
+        {
+        ASN1_OPT(PROXY_CERT_INFO_EXTENSION,pcPathLengthConstraint,ASN1_INTEGER),
+        ASN1_SIMPLE(PROXY_CERT_INFO_EXTENSION,proxyPolicy,PROXY_POLICY)
+} ASN1_SEQUENCE_END(PROXY_CERT_INFO_EXTENSION)
+
+IMPLEMENT_ASN1_FUNCTIONS(PROXY_CERT_INFO_EXTENSION)
diff --git a/src/lib/libssl/test/CAss.cnf b/src/lib/libssl/test/CAss.cnf
index b941b7ae15..21da59a73a 100644
--- a/src/lib/libssl/test/CAss.cnf
+++ b/src/lib/libssl/test/CAss.cnf
@@ -23,3 +23,11 @@ organizationName_value		= Dodgy Brothers
 
 commonName                      = Common Name (eg, YOUR name)
 commonName_value                = Dodgy CA
+
+[ v3_ca ]
+subjectKeyIdentifier=hash
+authorityKeyIdentifier=keyid:always,issuer:always
+basicConstraints = CA:true,pathlen:1
+keyUsage = cRLSign, keyCertSign
+issuerAltName=issuer:copy
+
diff --git a/src/lib/libssl/test/P1ss.cnf b/src/lib/libssl/test/P1ss.cnf
new file mode 100644
index 0000000000..876a0d35f8
--- /dev/null
+++ b/src/lib/libssl/test/P1ss.cnf
@@ -0,0 +1,37 @@
+#
+# SSLeay example configuration file.
+# This is mostly being used for generation of certificate requests.
+#
+
+RANDFILE                = ./.rnd
+
+####################################################################
+[ req ]
+default_bits            = 512
+default_keyfile         = keySS.pem
+distinguished_name      = req_distinguished_name
+encrypt_rsa_key         = no
+default_md              = md2
+
+[ req_distinguished_name ]
+countryName                     = Country Name (2 letter code)
+countryName_default             = AU
+countryName_value               = AU
+
+organizationName                = Organization Name (eg, company)
+organizationName_value          = Dodgy Brothers
+
+0.commonName                    = Common Name (eg, YOUR name)
+0.commonName_value              = Brother 1
+
+1.commonName                    = Common Name (eg, YOUR name)
+1.commonName_value              = Brother 2
+
+2.commonName                    = Common Name (eg, YOUR name)
+2.commonName_value              = Proxy 1
+
+[ v3_proxy ]
+basicConstraints=CA:FALSE
+subjectKeyIdentifier=hash
+authorityKeyIdentifier=keyid,issuer:always
+proxyCertInfo=critical,language:id-ppl-anyLanguage,pathlen:1,policy:text:AB
diff --git a/src/lib/libssl/test/P2ss.cnf b/src/lib/libssl/test/P2ss.cnf
new file mode 100644
index 0000000000..373a87e7c2
--- /dev/null
+++ b/src/lib/libssl/test/P2ss.cnf
@@ -0,0 +1,45 @@
+#
+# SSLeay example configuration file.
+# This is mostly being used for generation of certificate requests.
+#
+
+RANDFILE                = ./.rnd
+
+####################################################################
+[ req ]
+default_bits            = 512
+default_keyfile         = keySS.pem
+distinguished_name      = req_distinguished_name
+encrypt_rsa_key         = no
+default_md              = md2
+
+[ req_distinguished_name ]
+countryName                     = Country Name (2 letter code)
+countryName_default             = AU
+countryName_value               = AU
+
+organizationName                = Organization Name (eg, company)
+organizationName_value          = Dodgy Brothers
+
+0.commonName                    = Common Name (eg, YOUR name)
+0.commonName_value              = Brother 1
+
+1.commonName                    = Common Name (eg, YOUR name)
+1.commonName_value              = Brother 2
+
+2.commonName                    = Common Name (eg, YOUR name)
+2.commonName_value              = Proxy 1
+
+3.commonName                    = Common Name (eg, YOUR name)
+3.commonName_value              = Proxy 2
+
+[ v3_proxy ]
+basicConstraints=CA:FALSE
+subjectKeyIdentifier=hash
+authorityKeyIdentifier=keyid,issuer:always
+proxyCertInfo=critical,@proxy_ext
+
+[ proxy_ext ]
+language=id-ppl-anyLanguage
+pathlen=0
+policy=text:BC
diff --git a/src/lib/libssl/test/Uss.cnf b/src/lib/libssl/test/Uss.cnf
index c89692d519..0c0ebb5f67 100644
--- a/src/lib/libssl/test/Uss.cnf
+++ b/src/lib/libssl/test/Uss.cnf
@@ -26,3 +26,11 @@ organizationName_value          = Dodgy Brothers
 
 1.commonName                    = Common Name (eg, YOUR name)
 1.commonName_value              = Brother 2
+
+[ v3_ee ]
+subjectKeyIdentifier=hash
+authorityKeyIdentifier=keyid,issuer:always
+basicConstraints = CA:false
+keyUsage = nonRepudiation, digitalSignature, keyEncipherment
+issuerAltName=issuer:copy
+
diff --git a/src/lib/libssl/test/testsslproxy b/src/lib/libssl/test/testsslproxy
new file mode 100644
index 0000000000..58bbda8ab7
--- /dev/null
+++ b/src/lib/libssl/test/testsslproxy
@@ -0,0 +1,10 @@
+#! /bin/sh
+
+echo 'Testing a lot of proxy conditions.'
+echo 'Some of them may turn out being invalid, which is fine.'
+for auth in A B C BC; do
+    for cond in A B C 'A|B&!C'; do
+        sh ./testssl $1 $2 $3 "-proxy -proxy_auth $auth -proxy_cond $cond"
+        if [ $? = 3 ]; then exit 1; fi
+    done
+done