Rewrite sha1-586.pl.
authorAndy Polyakov <appro@openssl.org>
Thu, 31 Aug 2006 21:27:30 +0000 (21:27 +0000)
committerAndy Polyakov <appro@openssl.org>
Thu, 31 Aug 2006 21:27:30 +0000 (21:27 +0000)
crypto/sha/asm/sha1-586.pl

index 4f8521f..8a8228f 100644 (file)
@@ -1,4 +1,15 @@
-#!/usr/local/bin/perl
+#!/usr/bin/env perl
+
+# ====================================================================
+# [Re]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.
+# ====================================================================
+
+# "[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.
 
 # It was noted that Intel IA-32 C compiler generates code which
 # performs ~30% *faster* on P4 CPU than original *hand-coded*
 # ----------------------------------------------------------------
 #                                      <appro@fy.chalmers.se>
 
-$normal=0;
-
 push(@INC,"perlasm","../../perlasm");
 require "x86asm.pl";
 
 &asm_init($ARGV[0],"sha1-586.pl",$ARGV[$#ARGV] eq "386");
 
 $A="eax";
-$B="ecx";
-$C="ebx";
+$B="ebx";
+$C="ecx";
 $D="edx";
 $E="edi";
 $T="esi";
 $tmp1="ebp";
 
-$off=9*4;
-
-@K=(0x5a827999,0x6ed9eba1,0x8f1bbcdc,0xca62c1d6);
-
-&sha1_block_data("sha1_block_asm_data_order");
-
-&asm_finish();
-
-sub Nn
-       {
-       local($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].
+@V=($A,$B,$C,$D,$E,$T);
 
 sub BODY_00_15
        {
-       local($pos,$K,$X,$n,$a,$b,$c,$d,$e,$f)=@_;
+       local($n,$a,$b,$c,$d,$e,$f)=@_;
 
        &comment("00_15 $n");
 
@@ -109,37 +62,37 @@ sub BODY_00_15
         else        { &mov($a,$tmp1); }
        &rotl($tmp1,5);                 # tmp1=ROTATE(a,5)
         &xor($f,$d);
-       &and($f,$b);
-        &add($tmp1,$e);                # tmp1+=e;
-       &mov($e,&swtmp($n));            # e becomes volatile and
-                                       # is loaded with xi
+       &add($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($tmp1,&DWP($K,$tmp1,$e,1));# tmp1+=K_00_19+xi
+        &lea($tmp1,&DWP(0x5a827999,$tmp1,$e)); # tmp1+=K_00_19+xi
 
        if ($n==15) { &add($f,$tmp1); } # f+=tmp1
-       else        { &add($tmp1,$f); }
+       else        { &add($tmp1,$f); } # f becomes a in next round
        }
 
 sub BODY_16_19
        {
-       local($pos,$K,$X,$n,$a,$b,$c,$d,$e,$f)=@_;
-       local($n0,$n1,$n2,$n3,$np)=&Na($n);
+       local($n,$a,$b,$c,$d,$e,$f)=@_;
 
        &comment("16_19 $n");
 
-       &mov($f,&swtmp($n1));           # f to hold Xupdate(xi,xa,xb,xc,xd)
+       &mov($f,&swtmp($n%16));         # f to hold Xupdate(xi,xa,xb,xc,xd)
         &mov($tmp1,$c);                # tmp1 to hold F_00_19(b,c,d)
-       &xor($f,&swtmp($n0));
+       &xor($f,&swtmp(($n+2)%16));
         &xor($tmp1,$d);
-       &xor($f,&swtmp($n2));
+       &xor($f,&swtmp(($n+8)%16));
         &and($tmp1,$b);                # tmp1 holds F_00_19(b,c,d)
        &rotr($b,2);                    # b=ROTATE(b,30)
-        &xor($f,&swtmp($n3));          # f holds xa^xb^xc^xd
-       &rotl($f,1);                    # f=ROATE(f,1)
+        &xor($f,&swtmp(($n+13)%16));   # f holds xa^xb^xc^xd
+       &rotl($f,1);                    # f=ROTATE(f,1)
         &xor($tmp1,$d);                # tmp1=F_00_19(b,c,d)
-       &mov(&swtmp($n0),$f);           # xi=f
-       &lea($f,&DWP($K,$f,$e,1));      # f+=K_00_19+e
+       &mov(&swtmp($n%16),$f);         # xi=f
+       &lea($f,&DWP(0x5a827999,$f,$e));# 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)
@@ -148,48 +101,47 @@ sub BODY_16_19
 
 sub BODY_20_39
        {
-       local($pos,$K,$X,$n,$a,$b,$c,$d,$e,$f)=@_;
+       local($n,$a,$b,$c,$d,$e,$f)=@_;
+       local $K=($n<40)?0x6ed9eba1:0xca62c1d6;
 
        &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($f,&swtmp($n0));          # f to hold Xupdate(xi,xa,xb,xc,xd)
+        &mov($f,&swtmp($n%16));        # f to hold Xupdate(xi,xa,xb,xc,xd)
        &rotr($b,2);                    # b=ROTATE(b,30)
-        &xor($f,&swtmp($n1));
+        &xor($f,&swtmp(($n+2)%16));
        &xor($tmp1,$c);
-        &xor($f,&swtmp($n2));
+        &xor($f,&swtmp(($n+8)%16));
        &xor($tmp1,$d);                 # tmp1 holds F_20_39(b,c,d)
-        &xor($f,&swtmp($n3));          # f holds xa^xb^xc^xd
+        &xor($f,&swtmp(($n+13)%16));   # f holds xa^xb^xc^xd
        &rotl($f,1);                    # f=ROTATE(f,1)
         &add($tmp1,$e);
-       &mov(&swtmp($n0),$f);           # xi=f
+       &mov(&swtmp($n%16),$f);         # xi=f
         &mov($e,$a);                   # e becomes volatile
        &rotl($e,5);                    # e=ROTATE(a,5)
-        &lea($f,&DWP($K,$f,$tmp1,1));  # f+=K_20_39+e
+        &lea($f,&DWP($K,$f,$tmp1));    # f+=K_20_39+e
        &add($f,$e);                    # f+=ROTATE(a,5)
        }
 
 sub BODY_40_59
        {
-       local($pos,$K,$X,$n,$a,$b,$c,$d,$e,$f)=@_;
+       local($n,$a,$b,$c,$d,$e,$f)=@_;
 
        &comment("40_59 $n");
-       local($n0,$n1,$n2,$n3,$np)=&Na($n);
 
-       &mov($f,&swtmp($n0));           # f to hold Xupdate(xi,xa,xb,xc,xd)
-        &mov($tmp1,&swtmp($n1));
+       &mov($f,&swtmp($n%16));         # f to hold Xupdate(xi,xa,xb,xc,xd)
+        &mov($tmp1,&swtmp(($n+2)%16));
        &xor($f,$tmp1);
-        &mov($tmp1,&swtmp($n2));
+        &mov($tmp1,&swtmp(($n+8)%16));
        &xor($f,$tmp1);
-        &mov($tmp1,&swtmp($n3));
+        &mov($tmp1,&swtmp(($n+13)%16));
        &xor($f,$tmp1);                 # f holds xa^xb^xc^xd
         &mov($tmp1,$b);                # tmp1 to hold F_40_59(b,c,d)
        &rotl($f,1);                    # f=ROTATE(f,1)
         &or($tmp1,$c);
-       &mov(&swtmp($n0),$f);           # xi=f
+       &mov(&swtmp($n%16),$f);         # xi=f
         &and($tmp1,$d);
-       &lea($f,&DWP($K,$f,$e,1));      # f+=K_40_59+e
+       &lea($f,&DWP(0x8f1bbcdc,$f,$e));# f+=K_40_59+e
         &mov($e,$b);                   # e becomes volatile and is used
                                        # to calculate F_40_59(b,c,d)
        &rotr($b,2);                    # b=ROTATE(b,30)
@@ -201,230 +153,105 @@ sub BODY_40_59
        &add($f,$e);                    # f+=ROTATE(a,5)
        }
 
-sub BODY_60_79
-       {
-       &BODY_20_39(@_);
-       }
+&function_begin("sha1_block_asm_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 sha1_block_host
-       {
-       local($name, $sclabel)=@_;
-
-       &function_begin_B($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($C,        &DWP( 8,"ebp","",0));
-        &mov(&swtmp(17),"ecx");
-
-       &comment("First we need to setup the X array");
-
-       for ($i=0; $i<16; $i+=2)
+       &set_label("loop",16);
+
+       # copy input chunk to X, but reversing byte order!
+       for ($i=0; $i<16; $i+=4)
                {
-               &mov($A,&DWP(($i+0)*4,"esi","",0));# unless $i == 0;
-                &mov($B,&DWP(($i+1)*4,"esi","",0));
+               &mov($A,&DWP(4*($i+0),$T));
+               &mov($B,&DWP(4*($i+1),$T));
+               &mov($C,&DWP(4*($i+2),$T));
+               &mov($D,&DWP(4*($i+3),$T));
+               &bswap($A);
+               &bswap($B);
+               &bswap($C);
+               &bswap($D);
                &mov(&swtmp($i+0),$A);
-                &mov(&swtmp($i+1),$B);
+               &mov(&swtmp($i+1),$B);
+               &mov(&swtmp($i+2),$C);
+               &mov(&swtmp($i+3),$D);
+               }
+       &mov(&wparam(1),$T);    # redundant in 1st spin
+
+       &mov($A,&DWP(0,$tmp1)); # load SHA_CTX
+       &mov($B,&DWP(4,$tmp1));
+       &mov($C,&DWP(8,$tmp1));
+       &mov($D,&DWP(12,$tmp1));
+       # E is pre-loaded
+
+       &set_label("shortcut");
+       # keep a note of shortcut label so it can be used outside the block.
+       $shortcut = &label("shortcut");
+
+       for($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),$C);
+       &jb(&label("loop"));
+
+       &stack_pop(16);
+&function_end("sha1_block_asm_data_order");
+
+&function_begin("sha1_block_asm_host_order",16);
+       &mov($tmp1,&wparam(0)); # SHA_CTX *c
+       &mov($T,&wparam(1));    # const void *input
+       #&mov($A,&wparam(2));   # size_t num, always 1
+       &stack_push(16);        # allocate X[16]
+       &lea($A,&DWP(64,$T));   # this works, because num is always 1 here
+       &mov(&wparam(2),$A);    # pointer beyond the end of input
+
+       # just copy input to X
+       for ($i=0; $i<16; $i+=4)
+               {
+               &mov($A,&DWP(4*($i+0),$T));
+               &mov($B,&DWP(4*($i+1),$T));
+               &mov($C,&DWP(4*($i+2),$T));
+               &mov($D,&DWP(4*($i+3),$T));
+               &mov(&swtmp($i+0),$A);
+               &mov(&swtmp($i+1),$B);
+               &mov(&swtmp($i+2),$C);
+               &mov(&swtmp($i+3),$D);
                }
-       &jmp($sclabel);
-       &function_end_B($name);
-       }
 
+       &mov($A,&DWP(0,$tmp1)); # load SHA_CTX
+       &mov($B,&DWP(4,$tmp1));
+       &mov($C,&DWP(8,$tmp1));
+       &mov($D,&DWP(12,$tmp1));
+       &mov($E,&DWP(16,$tmp1));
 
-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", $sclabel);
-       }
+       &jmp($shortcut);        # this works, because num is always 1
+&function_end_B("sha1_block_asm_host_order");
 
+&asm_finish();