1 files changed, 339 insertions, 133 deletions

diff --git a/src/lib/libcrypto/sha/asm/sha1-586.pl b/src/lib/libcrypto/sha/asm/sha1-586.pl
index 0b4dab2bd5..041acc0348 100644
--- a/src/lib/libcrypto/sha/asm/sha1-586.pl
+++ b/src/lib/libcrypto/sha/asm/sha1-586.pl
@@ -1,16 +1,4 @@
-#!/usr/bin/env perl
-
-# ====================================================================
-# [Re]written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
-# project. The module is, however, dual licensed under OpenSSL and
-# CRYPTOGAMS licenses depending on where you obtain it. For further
-# details see http://www.openssl.org/~appro/cryptogams/.
-# ====================================================================
-
-# "[Re]written" was achieved in two major overhauls. In 2004 BODY_*
-# functions were re-implemented to address P4 performance issue [see
-# commentary below], and in 2006 the rest was rewritten in order to
-# gain freedom to liberate licensing terms.
+#!/usr/local/bin/perl
 
 # It was noted that Intel IA-32 C compiler generates code which
 # performs ~30% *faster* on P4 CPU than original *hand-coded*
@@ -21,7 +9,7 @@
 #
 #               compared with original  compared with Intel cc
 #               assembler impl.         generated code
-# Pentium       -16%                    +48%
+# Pentium       -25%                    +37%
 # PIII/AMD      +8%                     +16%
 # P4            +85%(!)                 +45%
 #
@@ -29,115 +17,174 @@
 # improvement on P4 outweights the loss and incorporate this
 # re-tuned code to 0.9.7 and later.
 # ----------------------------------------------------------------
+# Those who for any particular reason absolutely must score on
+# Pentium can replace this module with one from 0.9.6 distribution.
+# This "offer" shall be revoked the moment programming interface to
+# this module is changed, in which case this paragraph should be
+# removed.
+# ----------------------------------------------------------------
 #                                       <appro@fy.chalmers.se>
 
-$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
-push(@INC,"${dir}","${dir}../../perlasm");
+$normal=0;
+
+push(@INC,"perlasm","../../perlasm");
 require "x86asm.pl";
 
 &asm_init($ARGV[0],"sha1-586.pl",$ARGV[$#ARGV] eq "386");
 
 $A="eax";
-$B="ebx";
-$C="ecx";
+$B="ecx";
+$C="ebx";
 $D="edx";
 $E="edi";
 $T="esi";
 $tmp1="ebp";
 
-@V=($A,$B,$C,$D,$E,$T);
+$off=9*4;
+
+@K=(0x5a827999,0x6ed9eba1,0x8f1bbcdc,0xca62c1d6);
+
+&sha1_block_data("sha1_block_asm_data_order");
+
+&asm_finish();
+
+sub Nn
+        {
+        local($p)=@_;
+        local(%n)=($A,$T,$B,$A,$C,$B,$D,$C,$E,$D,$T,$E);
+        return($n{$p});
+        }
+
+sub Np
+        {
+        local($p)=@_;
+        local(%n)=($A,$T,$B,$A,$C,$B,$D,$C,$E,$D,$T,$E);
+        local(%n)=($A,$B,$B,$C,$C,$D,$D,$E,$E,$T,$T,$A);
+        return($n{$p});
+        }
+
+sub Na
+        {
+        local($n)=@_;
+        return( (($n   )&0x0f),
+                (($n+ 2)&0x0f),
+                (($n+ 8)&0x0f),
+                (($n+13)&0x0f),
+                (($n+ 1)&0x0f));
+        }
+
+sub X_expand
+        {
+        local($in)=@_;
+
+        &comment("First, load the words onto the stack in network byte order");
+        for ($i=0; $i<16; $i+=2)
+                {
+                &mov($A,&DWP(($i+0)*4,$in,"",0));# unless $i == 0;
+                 &mov($B,&DWP(($i+1)*4,$in,"",0));
+                &bswap($A);
+                 &bswap($B);
+                &mov(&swtmp($i+0),$A);
+                 &mov(&swtmp($i+1),$B);
+                }
+
+        &comment("We now have the X array on the stack");
+        &comment("starting at sp-4");
+        }
+
+# Rules of engagement
+# F is always trashable at the start, the running total.
+# E becomes the next F so it can be trashed after it has been 'accumulated'
+# F becomes A in the next round.  We don't need to access it much.
+# During the X update part, the result ends up in $X[$n0].
 
 sub BODY_00_15
         {
-        local($n,$a,$b,$c,$d,$e,$f)=@_;
+        local($pos,$K,$X,$n,$a,$b,$c,$d,$e,$f)=@_;
 
         &comment("00_15 $n");
 
-        &mov($f,$c);                    # f to hold F_00_19(b,c,d)
-         if ($n==0)  { &mov($tmp1,$a); }
-         else        { &mov($a,$tmp1); }
+        &mov($tmp1,$a);
+         &mov($f,$c);                   # f to hold F_00_19(b,c,d)
         &rotl($tmp1,5);                 # tmp1=ROTATE(a,5)
          &xor($f,$d);
+        &and($f,$b);
+         &rotr($b,2);                   # b=ROTATE(b,30)
         &add($tmp1,$e);                 # tmp1+=e;
-         &and($f,$b);
-        &mov($e,&swtmp($n%16));         # e becomes volatile and is loaded
-                                        # with xi, also note that e becomes
-                                        # f in next round...
-         &xor($f,$d);                   # f holds F_00_19(b,c,d)
-        &rotr($b,2);                    # b=ROTATE(b,30)
-         &lea($tmp1,&DWP(0x5a827999,$tmp1,$e)); # tmp1+=K_00_19+xi
-
-        if ($n==15) { &add($f,$tmp1); } # f+=tmp1
-        else        { &add($tmp1,$f); } # f becomes a in next round
+         &mov($e,&swtmp($n));           # e becomes volatile and
+                                        # is loaded with xi
+        &xor($f,$d);                    # f holds F_00_19(b,c,d)
+         &lea($tmp1,&DWP($K,$tmp1,$e,1));# tmp1+=K_00_19+xi
+        
+        &add($f,$tmp1);                 # f+=tmp1
         }
 
 sub BODY_16_19
         {
-        local($n,$a,$b,$c,$d,$e,$f)=@_;
+        local($pos,$K,$X,$n,$a,$b,$c,$d,$e,$f)=@_;
+        local($n0,$n1,$n2,$n3,$np)=&Na($n);
 
         &comment("16_19 $n");
 
-        &mov($f,&swtmp($n%16));         # f to hold Xupdate(xi,xa,xb,xc,xd)
+        &mov($f,&swtmp($n1));           # f to hold Xupdate(xi,xa,xb,xc,xd)
          &mov($tmp1,$c);                # tmp1 to hold F_00_19(b,c,d)
-        &xor($f,&swtmp(($n+2)%16));
+        &xor($f,&swtmp($n0));
          &xor($tmp1,$d);
-        &xor($f,&swtmp(($n+8)%16));
+        &xor($f,&swtmp($n2));
          &and($tmp1,$b);                # tmp1 holds F_00_19(b,c,d)
-        &rotr($b,2);                    # b=ROTATE(b,30)
-         &xor($f,&swtmp(($n+13)%16));   # f holds xa^xb^xc^xd
-        &rotl($f,1);                    # f=ROTATE(f,1)
-         &xor($tmp1,$d);                # tmp1=F_00_19(b,c,d)
-        &mov(&swtmp($n%16),$f);         # xi=f
-        &lea($f,&DWP(0x5a827999,$f,$e));# f+=K_00_19+e
+        &xor($f,&swtmp($n3));           # f holds xa^xb^xc^xd
+         &rotr($b,2);                   # b=ROTATE(b,30)
+        &xor($tmp1,$d);                 # tmp1=F_00_19(b,c,d)
+         &rotl($f,1);                   # f=ROATE(f,1)
+        &mov(&swtmp($n0),$f);           # xi=f
+        &lea($f,&DWP($K,$f,$e,1));      # f+=K_00_19+e
          &mov($e,$a);                   # e becomes volatile
-        &rotl($e,5);                    # e=ROTATE(a,5)
-         &add($f,$tmp1);                # f+=F_00_19(b,c,d)
+        &add($f,$tmp1);                 # f+=F_00_19(b,c,d)
+         &rotl($e,5);                   # e=ROTATE(a,5)
         &add($f,$e);                    # f+=ROTATE(a,5)
         }
 
 sub BODY_20_39
         {
-        local($n,$a,$b,$c,$d,$e,$f)=@_;
-        local $K=($n<40)?0x6ed9eba1:0xca62c1d6;
+        local($pos,$K,$X,$n,$a,$b,$c,$d,$e,$f)=@_;
 
         &comment("20_39 $n");
+        local($n0,$n1,$n2,$n3,$np)=&Na($n);
 
-        &mov($tmp1,$b);                 # tmp1 to hold F_20_39(b,c,d)
-         &mov($f,&swtmp($n%16));        # f to hold Xupdate(xi,xa,xb,xc,xd)
-        &rotr($b,2);                    # b=ROTATE(b,30)
-         &xor($f,&swtmp(($n+2)%16));
-        &xor($tmp1,$c);
-         &xor($f,&swtmp(($n+8)%16));
-        &xor($tmp1,$d);                 # tmp1 holds F_20_39(b,c,d)
-         &xor($f,&swtmp(($n+13)%16));   # f holds xa^xb^xc^xd
+        &mov($f,&swtmp($n0));           # f to hold Xupdate(xi,xa,xb,xc,xd)
+         &mov($tmp1,$b);                # tmp1 to hold F_20_39(b,c,d)
+        &xor($f,&swtmp($n1));
+         &rotr($b,2);                   # b=ROTATE(b,30)
+        &xor($f,&swtmp($n2));
+         &xor($tmp1,$c);
+        &xor($f,&swtmp($n3));           # f holds xa^xb^xc^xd
+         &xor($tmp1,$d);                # tmp1 holds F_20_39(b,c,d)
         &rotl($f,1);                    # f=ROTATE(f,1)
-         &add($tmp1,$e);
-        &mov(&swtmp($n%16),$f);         # xi=f
+        &mov(&swtmp($n0),$f);           # xi=f
+        &lea($f,&DWP($K,$f,$e,1));      # f+=K_20_39+e
          &mov($e,$a);                   # e becomes volatile
         &rotl($e,5);                    # e=ROTATE(a,5)
-         &lea($f,&DWP($K,$f,$tmp1));    # f+=K_20_39+e
+         &add($f,$tmp1);                # f+=F_20_39(b,c,d)
         &add($f,$e);                    # f+=ROTATE(a,5)
         }
 
 sub BODY_40_59
         {
-        local($n,$a,$b,$c,$d,$e,$f)=@_;
+        local($pos,$K,$X,$n,$a,$b,$c,$d,$e,$f)=@_;
 
         &comment("40_59 $n");
+        local($n0,$n1,$n2,$n3,$np)=&Na($n);
 
-        &mov($f,&swtmp($n%16));         # f to hold Xupdate(xi,xa,xb,xc,xd)
-         &mov($tmp1,&swtmp(($n+2)%16));
-        &xor($f,$tmp1);
-         &mov($tmp1,&swtmp(($n+8)%16));
-        &xor($f,$tmp1);
-         &mov($tmp1,&swtmp(($n+13)%16));
-        &xor($f,$tmp1);                 # f holds xa^xb^xc^xd
+        &mov($f,&swtmp($n0));           # f to hold Xupdate(xi,xa,xb,xc,xd)
          &mov($tmp1,$b);                # tmp1 to hold F_40_59(b,c,d)
-        &rotl($f,1);                    # f=ROTATE(f,1)
+        &xor($f,&swtmp($n1));
          &or($tmp1,$c);
-        &mov(&swtmp($n%16),$f);         # xi=f
+        &xor($f,&swtmp($n2));
          &and($tmp1,$d);
-        &lea($f,&DWP(0x8f1bbcdc,$f,$e));# f+=K_40_59+e
+        &xor($f,&swtmp($n3));           # f holds xa^xb^xc^xd
+        &rotl($f,1);                    # f=ROTATE(f,1)
+        &mov(&swtmp($n0),$f);           # xi=f
+        &lea($f,&DWP($K,$f,$e,1));      # f+=K_40_59+e
          &mov($e,$b);                   # e becomes volatile and is used
                                         # to calculate F_40_59(b,c,d)
         &rotr($b,2);                    # b=ROTATE(b,30)
@@ -145,75 +192,234 @@ sub BODY_40_59
         &or($tmp1,$e);                  # tmp1 holds F_40_59(b,c,d)             
          &mov($e,$a);
         &rotl($e,5);                    # e=ROTATE(a,5)
-         &add($f,$tmp1);                # f+=tmp1;
-        &add($f,$e);                    # f+=ROTATE(a,5)
+        &add($tmp1,$e);                 # tmp1+=ROTATE(a,5)
+        &add($f,$tmp1);                 # f+=tmp1;
         }
 
-&function_begin("sha1_block_data_order",16);
-        &mov($tmp1,&wparam(0)); # SHA_CTX *c
-        &mov($T,&wparam(1));    # const void *input
-        &mov($A,&wparam(2));    # size_t num
-        &stack_push(16);        # allocate X[16]
-        &shl($A,6);
-        &add($A,$T);
-        &mov(&wparam(2),$A);    # pointer beyond the end of input
-        &mov($E,&DWP(16,$tmp1));# pre-load E
+sub BODY_60_79
+        {
+        &BODY_20_39(@_);
+        }
+
+sub sha1_block_host
+        {
+        local($name, $sclabel)=@_;
 
-        &set_label("loop",16);
+        &function_begin_B($name,"");
 
-        # copy input chunk to X, but reversing byte order!
-        for ($i=0; $i<16; $i+=4)
+        # parameter 1 is the MD5_CTX structure.
+        # A     0
+        # B     4
+        # C     8
+        # D     12
+        # E     16
+
+        &mov("ecx",     &wparam(2));
+         &push("esi");
+        &shl("ecx",6);
+         &mov("esi",    &wparam(1));
+        &push("ebp");
+         &add("ecx","esi");     # offset to leave on
+        &push("ebx");
+         &mov("ebp",    &wparam(0));
+        &push("edi");
+         &mov($D,       &DWP(12,"ebp","",0));
+        &stack_push(18+9);
+         &mov($E,       &DWP(16,"ebp","",0));
+        &mov($C,        &DWP( 8,"ebp","",0));
+         &mov(&swtmp(17),"ecx");
+
+        &comment("First we need to setup the X array");
+
+        for ($i=0; $i<16; $i+=2)
                 {
-                &mov($A,&DWP(4*($i+0),$T));
-                &mov($B,&DWP(4*($i+1),$T));
-                &mov($C,&DWP(4*($i+2),$T));
-                &mov($D,&DWP(4*($i+3),$T));
-                &bswap($A);
-                &bswap($B);
-                &bswap($C);
-                &bswap($D);
+                &mov($A,&DWP(($i+0)*4,"esi","",0));# unless $i == 0;
+                 &mov($B,&DWP(($i+1)*4,"esi","",0));
                 &mov(&swtmp($i+0),$A);
-                &mov(&swtmp($i+1),$B);
-                &mov(&swtmp($i+2),$C);
-                &mov(&swtmp($i+3),$D);
+                 &mov(&swtmp($i+1),$B);
                 }
-        &mov(&wparam(1),$T);    # redundant in 1st spin
-
-        &mov($A,&DWP(0,$tmp1)); # load SHA_CTX
-        &mov($B,&DWP(4,$tmp1));
-        &mov($C,&DWP(8,$tmp1));
-        &mov($D,&DWP(12,$tmp1));
-        # E is pre-loaded
-
-        for($i=0;$i<16;$i++)    { &BODY_00_15($i,@V); unshift(@V,pop(@V)); }
-        for(;$i<20;$i++)        { &BODY_16_19($i,@V); unshift(@V,pop(@V)); }
-        for(;$i<40;$i++)        { &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
-        for(;$i<60;$i++)        { &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
-        for(;$i<80;$i++)        { &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
-
-        (($V[5] eq $D) and ($V[0] eq $E)) or die;       # double-check
-
-        &mov($tmp1,&wparam(0)); # re-load SHA_CTX*
-        &mov($D,&wparam(1));    # D is last "T" and is discarded
-
-        &add($E,&DWP(0,$tmp1)); # E is last "A"...
-        &add($T,&DWP(4,$tmp1));
-        &add($A,&DWP(8,$tmp1));
-        &add($B,&DWP(12,$tmp1));
-        &add($C,&DWP(16,$tmp1));
-
-        &mov(&DWP(0,$tmp1),$E); # update SHA_CTX
-         &add($D,64);           # advance input pointer
-        &mov(&DWP(4,$tmp1),$T);
-         &cmp($D,&wparam(2));   # have we reached the end yet?
-        &mov(&DWP(8,$tmp1),$A);
-         &mov($E,$C);           # C is last "E" which needs to be "pre-loaded"
-        &mov(&DWP(12,$tmp1),$B);
-         &mov($T,$D);           # input pointer
-        &mov(&DWP(16,$tmp1),$C);
-        &jb(&label("loop"));
-
-        &stack_pop(16);
-&function_end("sha1_block_data_order");
+        &jmp($sclabel);
+        &function_end_B($name);
+        }
+
+
+sub sha1_block_data
+        {
+        local($name)=@_;
+
+        &function_begin_B($name,"");
+
+        # parameter 1 is the MD5_CTX structure.
+        # A     0
+        # B     4
+        # C     8
+        # D     12
+        # E     16
+
+        &mov("ecx",     &wparam(2));
+         &push("esi");
+        &shl("ecx",6);
+         &mov("esi",    &wparam(1));
+        &push("ebp");
+         &add("ecx","esi");     # offset to leave on
+        &push("ebx");
+         &mov("ebp",    &wparam(0));
+        &push("edi");
+         &mov($D,       &DWP(12,"ebp","",0));
+        &stack_push(18+9);
+         &mov($E,       &DWP(16,"ebp","",0));
+        &mov($C,        &DWP( 8,"ebp","",0));
+         &mov(&swtmp(17),"ecx");
+
+        &comment("First we need to setup the X array");
+
+        &set_label("start") unless $normal;
+
+        &X_expand("esi");
+         &mov(&wparam(1),"esi");
+
+        &set_label("shortcut", 0, 1);
+        &comment("");
+        &comment("Start processing");
+
+        # odd start
+        &mov($A,        &DWP( 0,"ebp","",0));
+         &mov($B,       &DWP( 4,"ebp","",0));
+        $X="esp";
+        &BODY_00_15(-2,$K[0],$X, 0,$A,$B,$C,$D,$E,$T);
+        &BODY_00_15( 0,$K[0],$X, 1,$T,$A,$B,$C,$D,$E);
+        &BODY_00_15( 0,$K[0],$X, 2,$E,$T,$A,$B,$C,$D);
+        &BODY_00_15( 0,$K[0],$X, 3,$D,$E,$T,$A,$B,$C);
+        &BODY_00_15( 0,$K[0],$X, 4,$C,$D,$E,$T,$A,$B);
+        &BODY_00_15( 0,$K[0],$X, 5,$B,$C,$D,$E,$T,$A);
+        &BODY_00_15( 0,$K[0],$X, 6,$A,$B,$C,$D,$E,$T);
+        &BODY_00_15( 0,$K[0],$X, 7,$T,$A,$B,$C,$D,$E);
+        &BODY_00_15( 0,$K[0],$X, 8,$E,$T,$A,$B,$C,$D);
+        &BODY_00_15( 0,$K[0],$X, 9,$D,$E,$T,$A,$B,$C);
+        &BODY_00_15( 0,$K[0],$X,10,$C,$D,$E,$T,$A,$B);
+        &BODY_00_15( 0,$K[0],$X,11,$B,$C,$D,$E,$T,$A);
+        &BODY_00_15( 0,$K[0],$X,12,$A,$B,$C,$D,$E,$T);
+        &BODY_00_15( 0,$K[0],$X,13,$T,$A,$B,$C,$D,$E);
+        &BODY_00_15( 0,$K[0],$X,14,$E,$T,$A,$B,$C,$D);
+        &BODY_00_15( 1,$K[0],$X,15,$D,$E,$T,$A,$B,$C);
+        &BODY_16_19(-1,$K[0],$X,16,$C,$D,$E,$T,$A,$B);
+        &BODY_16_19( 0,$K[0],$X,17,$B,$C,$D,$E,$T,$A);
+        &BODY_16_19( 0,$K[0],$X,18,$A,$B,$C,$D,$E,$T);
+        &BODY_16_19( 1,$K[0],$X,19,$T,$A,$B,$C,$D,$E);
+
+        &BODY_20_39(-1,$K[1],$X,20,$E,$T,$A,$B,$C,$D);
+        &BODY_20_39( 0,$K[1],$X,21,$D,$E,$T,$A,$B,$C);
+        &BODY_20_39( 0,$K[1],$X,22,$C,$D,$E,$T,$A,$B);
+        &BODY_20_39( 0,$K[1],$X,23,$B,$C,$D,$E,$T,$A);
+        &BODY_20_39( 0,$K[1],$X,24,$A,$B,$C,$D,$E,$T);
+        &BODY_20_39( 0,$K[1],$X,25,$T,$A,$B,$C,$D,$E);
+        &BODY_20_39( 0,$K[1],$X,26,$E,$T,$A,$B,$C,$D);
+        &BODY_20_39( 0,$K[1],$X,27,$D,$E,$T,$A,$B,$C);
+        &BODY_20_39( 0,$K[1],$X,28,$C,$D,$E,$T,$A,$B);
+        &BODY_20_39( 0,$K[1],$X,29,$B,$C,$D,$E,$T,$A);
+        &BODY_20_39( 0,$K[1],$X,30,$A,$B,$C,$D,$E,$T);
+        &BODY_20_39( 0,$K[1],$X,31,$T,$A,$B,$C,$D,$E);
+        &BODY_20_39( 0,$K[1],$X,32,$E,$T,$A,$B,$C,$D);
+        &BODY_20_39( 0,$K[1],$X,33,$D,$E,$T,$A,$B,$C);
+        &BODY_20_39( 0,$K[1],$X,34,$C,$D,$E,$T,$A,$B);
+        &BODY_20_39( 0,$K[1],$X,35,$B,$C,$D,$E,$T,$A);
+        &BODY_20_39( 0,$K[1],$X,36,$A,$B,$C,$D,$E,$T);
+        &BODY_20_39( 0,$K[1],$X,37,$T,$A,$B,$C,$D,$E);
+        &BODY_20_39( 0,$K[1],$X,38,$E,$T,$A,$B,$C,$D);
+        &BODY_20_39( 1,$K[1],$X,39,$D,$E,$T,$A,$B,$C);
+
+        &BODY_40_59(-1,$K[2],$X,40,$C,$D,$E,$T,$A,$B);
+        &BODY_40_59( 0,$K[2],$X,41,$B,$C,$D,$E,$T,$A);
+        &BODY_40_59( 0,$K[2],$X,42,$A,$B,$C,$D,$E,$T);
+        &BODY_40_59( 0,$K[2],$X,43,$T,$A,$B,$C,$D,$E);
+        &BODY_40_59( 0,$K[2],$X,44,$E,$T,$A,$B,$C,$D);
+        &BODY_40_59( 0,$K[2],$X,45,$D,$E,$T,$A,$B,$C);
+        &BODY_40_59( 0,$K[2],$X,46,$C,$D,$E,$T,$A,$B);
+        &BODY_40_59( 0,$K[2],$X,47,$B,$C,$D,$E,$T,$A);
+        &BODY_40_59( 0,$K[2],$X,48,$A,$B,$C,$D,$E,$T);
+        &BODY_40_59( 0,$K[2],$X,49,$T,$A,$B,$C,$D,$E);
+        &BODY_40_59( 0,$K[2],$X,50,$E,$T,$A,$B,$C,$D);
+        &BODY_40_59( 0,$K[2],$X,51,$D,$E,$T,$A,$B,$C);
+        &BODY_40_59( 0,$K[2],$X,52,$C,$D,$E,$T,$A,$B);
+        &BODY_40_59( 0,$K[2],$X,53,$B,$C,$D,$E,$T,$A);
+        &BODY_40_59( 0,$K[2],$X,54,$A,$B,$C,$D,$E,$T);
+        &BODY_40_59( 0,$K[2],$X,55,$T,$A,$B,$C,$D,$E);
+        &BODY_40_59( 0,$K[2],$X,56,$E,$T,$A,$B,$C,$D);
+        &BODY_40_59( 0,$K[2],$X,57,$D,$E,$T,$A,$B,$C);
+        &BODY_40_59( 0,$K[2],$X,58,$C,$D,$E,$T,$A,$B);
+        &BODY_40_59( 1,$K[2],$X,59,$B,$C,$D,$E,$T,$A);
+
+        &BODY_60_79(-1,$K[3],$X,60,$A,$B,$C,$D,$E,$T);
+        &BODY_60_79( 0,$K[3],$X,61,$T,$A,$B,$C,$D,$E);
+        &BODY_60_79( 0,$K[3],$X,62,$E,$T,$A,$B,$C,$D);
+        &BODY_60_79( 0,$K[3],$X,63,$D,$E,$T,$A,$B,$C);
+        &BODY_60_79( 0,$K[3],$X,64,$C,$D,$E,$T,$A,$B);
+        &BODY_60_79( 0,$K[3],$X,65,$B,$C,$D,$E,$T,$A);
+        &BODY_60_79( 0,$K[3],$X,66,$A,$B,$C,$D,$E,$T);
+        &BODY_60_79( 0,$K[3],$X,67,$T,$A,$B,$C,$D,$E);
+        &BODY_60_79( 0,$K[3],$X,68,$E,$T,$A,$B,$C,$D);
+        &BODY_60_79( 0,$K[3],$X,69,$D,$E,$T,$A,$B,$C);
+        &BODY_60_79( 0,$K[3],$X,70,$C,$D,$E,$T,$A,$B);
+        &BODY_60_79( 0,$K[3],$X,71,$B,$C,$D,$E,$T,$A);
+        &BODY_60_79( 0,$K[3],$X,72,$A,$B,$C,$D,$E,$T);
+        &BODY_60_79( 0,$K[3],$X,73,$T,$A,$B,$C,$D,$E);
+        &BODY_60_79( 0,$K[3],$X,74,$E,$T,$A,$B,$C,$D);
+        &BODY_60_79( 0,$K[3],$X,75,$D,$E,$T,$A,$B,$C);
+        &BODY_60_79( 0,$K[3],$X,76,$C,$D,$E,$T,$A,$B);
+        &BODY_60_79( 0,$K[3],$X,77,$B,$C,$D,$E,$T,$A);
+        &BODY_60_79( 0,$K[3],$X,78,$A,$B,$C,$D,$E,$T);
+        &BODY_60_79( 2,$K[3],$X,79,$T,$A,$B,$C,$D,$E);
+
+        &comment("End processing");
+        &comment("");
+        # D is the tmp value
+
+        # E -> A
+        # T -> B
+        # A -> C
+        # B -> D
+        # C -> E
+        # D -> T
+
+        &mov($tmp1,&wparam(0));
+
+         &mov($D,       &DWP(12,$tmp1,"",0));
+        &add($D,$B);
+         &mov($B,       &DWP( 4,$tmp1,"",0));
+        &add($B,$T);
+         &mov($T,       $A);
+        &mov($A,        &DWP( 0,$tmp1,"",0));
+         &mov(&DWP(12,$tmp1,"",0),$D);
+
+        &add($A,$E);
+         &mov($E,       &DWP(16,$tmp1,"",0));
+        &add($E,$C);
+         &mov($C,       &DWP( 8,$tmp1,"",0));
+        &add($C,$T);
+
+         &mov(&DWP( 0,$tmp1,"",0),$A);
+        &mov("esi",&wparam(1));
+         &mov(&DWP( 8,$tmp1,"",0),$C);
+        &add("esi",64);
+         &mov("eax",&swtmp(17));
+        &mov(&DWP(16,$tmp1,"",0),$E);
+         &cmp("esi","eax");
+        &mov(&DWP( 4,$tmp1,"",0),$B);
+         &jb(&label("start"));
+
+        &stack_pop(18+9);
+         &pop("edi");
+        &pop("ebx");
+         &pop("ebp");
+        &pop("esi");
+         &ret();
+
+        # keep a note of shortcut label so it can be used outside
+        # block.
+        my $sclabel = &label("shortcut");
+
+        &function_end_B($name);
+        # Putting this here avoids problems with MASM in debugging mode
+        &sha1_block_host("sha1_block_asm_host_order", $sclabel);
+        }
 
-&asm_finish();