summaryrefslogtreecommitdiffstats
path: root/crypto
diff options
context:
space:
mode:
authorAndy Polyakov <appro@openssl.org>2004-12-23 21:32:34 +0000
committerAndy Polyakov <appro@openssl.org>2004-12-23 21:32:34 +0000
commit713147109c8df270116fe69b3f1820458ea30583 (patch)
tree0a4a68e0a317a18b00e1bffde64b0843171c380b /crypto
parent76ef6ac956490b4d9f78445599ce52510021858a (diff)
AES x86 assembler implementation.
Diffstat (limited to 'crypto')
-rwxr-xr-xcrypto/aes/asm/aes-586.pl1082
1 files changed, 1082 insertions, 0 deletions
diff --git a/crypto/aes/asm/aes-586.pl b/crypto/aes/asm/aes-586.pl
new file mode 100755
index 0000000000..d728e3e661
--- /dev/null
+++ b/crypto/aes/asm/aes-586.pl
@@ -0,0 +1,1082 @@
+#!/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.
+# ====================================================================
+#
+# You might fail to appreciate this module performance from the first
+# try. If compared to "vanilla" linux-ia32-icc target, i.e. Intel C
+# without -KPIC, performance appears to be virtually identical... But
+# try to configure with shared library support... Aha! Intel compiler
+# "suddenly" lags behind by 30% [on P4]:-) And if compared to
+# position-independent code generated by GNU C, this code performs
+# more than *twice* as fast! Yes, all this buzz about PIC means that
+# [unlike other implementations] this module was explicitly designed
+# to be safe to use even in shared library context...
+#
+# Special note about instruction choice. Do you recall RC4_INT code
+# performing poorly on P4? It might be the time to figure out why.
+# RC4_INT code implies effective address calculations in base+offset*4
+# form. Trouble is that it seems that offset scaling turned to be
+# critical path... At least eliminating scaling resulted in 2.8x RC4
+# performance improvement [as you might recall]. As AES code is hungry
+# for scaling too, I [try to] avoid the latter by favoring off-by-2
+# shifts and masking the result with 0xFF<<2 instead of "boring" 0xFF.
+
+push(@INC,"perlasm","../../perlasm");
+require "x86asm.pl";
+
+&asm_init($ARGV[0],"aes-586.pl",$ARGV[$#ARGV] eq "386");
+
+$small_footprint=1; # $small_footprint=1 code is 5-9% slower, but
+ # 5 times smaller! I default to compact code.
+$s0="eax";
+$s1="ebx";
+$s2="ecx";
+$s3="edx";
+
+sub encstep()
+{ my ($i,$te,@s) = @_;
+ my $tmp,$out;
+
+ if ($i==3) { $out=$s[0]; &mov ("edi",&DWP(12,"esp"));}
+ else { $out="esi"; &mov ($out,$s[0]); }
+ &shr ($out,24-2);
+ &and ($out,0xFF<<2);
+ &mov ($out,&DWP(1024*0,$te,$out));
+
+ if ($i==3) { $tmp=$s[1]; }
+ else { $tmp="edi"; &mov ($tmp,$s[1]); }
+ &shr ($tmp,16-2);
+ &and ($tmp,0xFF<<2);
+ &xor ($out,&DWP(1024*1,$te,$tmp));
+
+ if ($i==3) { $tmp=$s[2]; &mov ($s[1],&DWP(0,"esp")); }
+ else { $tmp="edi"; }
+ &movz ($tmp,&HB($s[2]));
+ &xor ($out,&DWP(1024*2,$te,$tmp,4));
+
+ if ($i==3) { $tmp=$s[3]; &mov ($s[2],&DWP(4,"esp")); }
+ else { $tmp="edi"; &mov ($tmp,$s[3]); }
+ &and ($tmp,0xFF);
+ &xor ($out,&DWP(1024*3,$te,$tmp,4));
+ if ($i<2) { &mov (&DWP(4*$i,"esp"),$out); }
+ if ($i==3) { &mov ($s[3],"esi"); }
+}
+
+sub enclast()
+{ my ($i,$te,@s)=@_;
+ my $tmp,$out;
+
+ if ($i==3) { $out=$s[0]; &mov ("edi",&DWP(12,"esp"));}
+ else { $out="esi"; &mov ($out,$s[0]); }
+ &shr ($out,24-2);
+ &and ($out,0xFF<<2);
+ &mov ($out,&DWP(0,$te,$out));
+ &and ($out,0xff000000);
+
+ if ($i==3) { $tmp=$s[1]; }
+ else { $tmp="edi"; &mov ($tmp,$s[1]); }
+ &shr ($tmp,16-2);
+ &and ($tmp,0xFF<<2);
+ &mov ($tmp,&DWP(0,$te,$tmp));
+ &and ($tmp,0x00ff0000);
+ &xor ($out,$tmp);
+
+ if ($i==3) { $tmp=$s[2]; &mov ($s[1],&DWP(0,"esp")); }
+ else { $tmp="edi"; }
+ &movz ($tmp,&HB($s[2]));
+ &mov ($tmp,&DWP(0,$te,$tmp,4));
+ &and ($tmp,0x0000ff00);
+ &xor ($out,$tmp);
+
+ if ($i==3) { $tmp=$s[3]; &mov ($s[2],&DWP(4,"esp")); }
+ else { $tmp="edi"; &mov ($tmp,$s[3]); }
+ &and ($tmp,0xFF);
+ &mov ($tmp,&DWP(0,$te,$tmp,4));
+ &and ($tmp,0x000000ff);
+ &xor ($out,$tmp);
+ if ($i<2) { &mov (&DWP(4*$i,"esp"),$out); }
+ if ($i==3) { &mov ($s[3],"esi"); }
+}
+
+# void AES_encrypt (const void *inp,void *out,const AES_KEY *key);
+&public_label("AES_Te");
+&function_begin("AES_encrypt");
+ &mov ("esi",&wparam(0)); # load inp
+ &mov ("edi",&wparam(2)); # load key
+
+ &call (&label("pic_point")); # make it PIC!
+&set_label("pic_point");
+ &blindpop("ebp");
+ &lea ("ebp",&DWP(&label("AES_Te")."-".&label("pic_point"),"ebp"));
+
+ # allocate aligned stack frame
+ &mov ("eax","esp");
+ &sub ("esp",20);
+ &and ("esp",-16);
+
+ &mov (&DWP(12,"esp"),"edi"); # save key
+ &mov (&DWP(16,"esp"),"eax"); # save %esp
+
+ &mov ($s0,&DWP(0,"esi")); # load input data
+ &mov ($s1,&DWP(4,"esi"));
+ &mov ($s2,&DWP(8,"esi"));
+ &mov ($s3,&DWP(12,"esi"));
+ #
+ # It's perfectly possible to implement algorithm as
+ # little-endian and get rid of bswaps... It would give
+ # less than 1% performance improvement, so I judge it
+ # doesn't worth the trouble...
+ #
+ &bswap ($s0);
+ &bswap ($s1);
+ &bswap ($s2);
+ &bswap ($s3);
+ &xor ($s0,&DWP(0,"edi"));
+ &xor ($s1,&DWP(4,"edi"));
+ &xor ($s2,&DWP(8,"edi"));
+ &xor ($s3,&DWP(12,"edi"));
+
+ &mov ("esi",&DWP(240,"edi")); # load key->rounds
+
+ if ($small_footprint) {
+ &lea ("esi",&DWP(-2,"esi","esi"));
+ &lea ("esi",&DWP(0,"edi","esi",8));
+ &mov (&DWP(8,"esp"),"esi"); # end of key schedule
+ &align (4);
+ &set_label("loop");
+ &encstep(0,"ebp",$s0,$s1,$s2,$s3);
+ &encstep(1,"ebp",$s1,$s2,$s3,$s0);
+ &encstep(2,"ebp",$s2,$s3,$s0,$s1);
+ &encstep(3,"ebp",$s3,$s0,$s1,$s2);
+ &add ("edi",16); # advance rd_key
+ &xor ($s0,&DWP(0,"edi"));
+ &xor ($s1,&DWP(4,"edi"));
+ &xor ($s2,&DWP(8,"edi"));
+ &xor ($s3,&DWP(12,"edi"));
+ &cmp ("edi",&DWP(8,"esp"));
+ &mov (&DWP(12,"esp"),"edi");
+ &jb (&label("loop"));
+ }
+ else {
+ &cmp ("esi",10);
+ &jle (&label("10rounds"));
+ &cmp ("esi",12);
+ &jle (&label("12rounds"));
+
+ &set_label("14rounds");
+ for ($i=1;$i<3;$i++) {
+ &encstep(0,"ebp",$s0,$s1,$s2,$s3);
+ &encstep(1,"ebp",$s1,$s2,$s3,$s0);
+ &encstep(2,"ebp",$s2,$s3,$s0,$s1);
+ &encstep(3,"ebp",$s3,$s0,$s1,$s2);
+ &xor ($s0,&DWP(16*$i+0,"edi"));
+ &xor ($s1,&DWP(16*$i+4,"edi"));
+ &xor ($s2,&DWP(16*$i+8,"edi"));
+ &xor ($s3,&DWP(16*$i+12,"edi"));
+ }
+ &add ("edi",32);
+ &mov (&DWP(12,"esp"),"edi"); # advance rd_key
+ &set_label("12rounds");
+ for ($i=1;$i<3;$i++) {
+ &encstep(0,"ebp",$s0,$s1,$s2,$s3);
+ &encstep(1,"ebp",$s1,$s2,$s3,$s0);
+ &encstep(2,"ebp",$s2,$s3,$s0,$s1);
+ &encstep(3,"ebp",$s3,$s0,$s1,$s2);
+ &xor ($s0,&DWP(16*$i+0,"edi"));
+ &xor ($s1,&DWP(16*$i+4,"edi"));
+ &xor ($s2,&DWP(16*$i+8,"edi"));
+ &xor ($s3,&DWP(16*$i+12,"edi"));
+ }
+ &add ("edi",32);
+ &mov (&DWP(12,"esp"),"edi"); # advance rd_key
+ &set_label("10rounds");
+ for ($i=1;$i<10;$i++) {
+ &encstep(0,"ebp",$s0,$s1,$s2,$s3);
+ &encstep(1,"ebp",$s1,$s2,$s3,$s0);
+ &encstep(2,"ebp",$s2,$s3,$s0,$s1);
+ &encstep(3,"ebp",$s3,$s0,$s1,$s2);
+ &xor ($s0,&DWP(16*$i+0,"edi"));
+ &xor ($s1,&DWP(16*$i+4,"edi"));
+ &xor ($s2,&DWP(16*$i+8,"edi"));
+ &xor ($s3,&DWP(16*$i+12,"edi"));
+ }
+ }
+
+ &add ("ebp",4*1024); # skip to Te4
+ &enclast(0,"ebp",$s0,$s1,$s2,$s3);
+ &enclast(1,"ebp",$s1,$s2,$s3,$s0);
+ &enclast(2,"ebp",$s2,$s3,$s0,$s1);
+ &enclast(3,"ebp",$s3,$s0,$s1,$s2);
+
+ &mov ("esp",&DWP(16,"esp")); # restore %esp
+ &add ("edi",$small_footprint?16:160);
+ &xor ($s0,&DWP(0,"edi"));
+ &xor ($s1,&DWP(4,"edi"));
+ &xor ($s2,&DWP(8,"edi"));
+ &xor ($s3,&DWP(12,"edi"));
+ &bswap ($s0);
+ &bswap ($s1);
+ &bswap ($s2);
+ &bswap ($s3);
+ &mov ("edi",&wparam(1)); # load out
+ &mov (&DWP(0,"edi"),$s0); # write output data
+ &mov (&DWP(4,"edi"),$s1);
+ &mov (&DWP(8,"edi"),$s2);
+ &mov (&DWP(12,"edi"),$s3);
+
+ &pop ("edi");
+ &pop ("esi");
+ &pop ("ebx");
+ &pop ("ebp");
+ &ret ();
+
+&set_label("AES_Te",64); # Yes! I keep it in the code segment!
+ &data_word(0xc66363a5, 0xf87c7c84, 0xee777799, 0xf67b7b8d);
+ &data_word(0xfff2f20d, 0xd66b6bbd, 0xde6f6fb1, 0x91c5c554);
+ &data_word(0x60303050, 0x02010103, 0xce6767a9, 0x562b2b7d);
+ &data_word(0xe7fefe19, 0xb5d7d762, 0x4dababe6, 0xec76769a);
+ &data_word(0x8fcaca45, 0x1f82829d, 0x89c9c940, 0xfa7d7d87);
+ &data_word(0xeffafa15, 0xb25959eb, 0x8e4747c9, 0xfbf0f00b);
+ &data_word(0x41adadec, 0xb3d4d467, 0x5fa2a2fd, 0x45afafea);
+ &data_word(0x239c9cbf, 0x53a4a4f7, 0xe4727296, 0x9bc0c05b);
+ &data_word(0x75b7b7c2, 0xe1fdfd1c, 0x3d9393ae, 0x4c26266a);
+ &data_word(0x6c36365a, 0x7e3f3f41, 0xf5f7f702, 0x83cccc4f);
+ &data_word(0x6834345c, 0x51a5a5f4, 0xd1e5e534, 0xf9f1f108);
+ &data_word(0xe2717193, 0xabd8d873, 0x62313153, 0x2a15153f);
+ &data_word(0x0804040c, 0x95c7c752, 0x46232365, 0x9dc3c35e);
+ &data_word(0x30181828, 0x379696a1, 0x0a05050f, 0x2f9a9ab5);
+ &data_word(0x0e070709, 0x24121236, 0x1b80809b, 0xdfe2e23d);
+ &data_word(0xcdebeb26, 0x4e272769, 0x7fb2b2cd, 0xea75759f);
+ &data_word(0x1209091b, 0x1d83839e, 0x582c2c74, 0x341a1a2e);
+ &data_word(0x361b1b2d, 0xdc6e6eb2, 0xb45a5aee, 0x5ba0a0fb);
+ &data_word(0xa45252f6, 0x763b3b4d, 0xb7d6d661, 0x7db3b3ce);
+ &data_word(0x5229297b, 0xdde3e33e, 0x5e2f2f71, 0x13848497);
+ &data_word(0xa65353f5, 0xb9d1d168, 0x00000000, 0xc1eded2c);
+ &data_word(0x40202060, 0xe3fcfc1f, 0x79b1b1c8, 0xb65b5bed);
+ &data_word(0xd46a6abe, 0x8dcbcb46, 0x67bebed9, 0x7239394b);
+ &data_word(0x944a4ade, 0x984c4cd4, 0xb05858e8, 0x85cfcf4a);
+ &data_word(0xbbd0d06b, 0xc5efef2a, 0x4faaaae5, 0xedfbfb16);
+ &data_word(0x864343c5, 0x9a4d4dd7, 0x66333355, 0x11858594);
+ &data_word(0x8a4545cf, 0xe9f9f910, 0x04020206, 0xfe7f7f81);
+ &data_word(0xa05050f0, 0x783c3c44, 0x259f9fba, 0x4ba8a8e3);
+ &data_word(0xa25151f3, 0x5da3a3fe, 0x804040c0, 0x058f8f8a);
+ &data_word(0x3f9292ad, 0x219d9dbc, 0x70383848, 0xf1f5f504);
+ &data_word(0x63bcbcdf, 0x77b6b6c1, 0xafdada75, 0x42212163);
+ &data_word(0x20101030, 0xe5ffff1a, 0xfdf3f30e, 0xbfd2d26d);
+ &data_word(0x81cdcd4c, 0x180c0c14, 0x26131335, 0xc3ecec2f);
+ &data_word(0xbe5f5fe1, 0x359797a2, 0x884444cc, 0x2e171739);
+ &data_word(0x93c4c457, 0x55a7a7f2, 0xfc7e7e82, 0x7a3d3d47);
+ &data_word(0xc86464ac, 0xba5d5de7, 0x3219192b, 0xe6737395);
+ &data_word(0xc06060a0, 0x19818198, 0x9e4f4fd1, 0xa3dcdc7f);
+ &data_word(0x44222266, 0x542a2a7e, 0x3b9090ab, 0x0b888883);
+ &data_word(0x8c4646ca, 0xc7eeee29, 0x6bb8b8d3, 0x2814143c);
+ &data_word(0xa7dede79, 0xbc5e5ee2, 0x160b0b1d, 0xaddbdb76);
+ &data_word(0xdbe0e03b, 0x64323256, 0x743a3a4e, 0x140a0a1e);
+ &data_word(0x924949db, 0x0c06060a, 0x4824246c, 0xb85c5ce4);
+ &data_word(0x9fc2c25d, 0xbdd3d36e, 0x43acacef, 0xc46262a6);
+ &data_word(0x399191a8, 0x319595a4, 0xd3e4e437, 0xf279798b);
+ &data_word(0xd5e7e732, 0x8bc8c843, 0x6e373759, 0xda6d6db7);
+ &data_word(0x018d8d8c, 0xb1d5d564, 0x9c4e4ed2, 0x49a9a9e0);
+ &data_word(0xd86c6cb4, 0xac5656fa, 0xf3f4f407, 0xcfeaea25);
+ &data_word(0xca6565af, 0xf47a7a8e, 0x47aeaee9, 0x10080818);
+ &data_word(0x6fbabad5, 0xf0787888, 0x4a25256f, 0x5c2e2e72);
+ &data_word(0x381c1c24, 0x57a6a6f1, 0x73b4b4c7, 0x97c6c651);
+ &data_word(0xcbe8e823, 0xa1dddd7c, 0xe874749c, 0x3e1f1f21);
+ &data_word(0x964b4bdd, 0x61bdbddc, 0x0d8b8b86, 0x0f8a8a85);
+ &data_word(0xe0707090, 0x7c3e3e42, 0x71b5b5c4, 0xcc6666aa);
+ &data_word(0x904848d8, 0x06030305, 0xf7f6f601, 0x1c0e0e12);
+ &data_word(0xc26161a3, 0x6a35355f, 0xae5757f9, 0x69b9b9d0);
+ &data_word(0x17868691, 0x99c1c158, 0x3a1d1d27, 0x279e9eb9);
+ &data_word(0xd9e1e138, 0xebf8f813, 0x2b9898b3, 0x22111133);
+ &data_word(0xd26969bb, 0xa9d9d970, 0x078e8e89, 0x339494a7);
+ &data_word(0x2d9b9bb6, 0x3c1e1e22, 0x15878792, 0xc9e9e920);
+ &data_word(0x87cece49, 0xaa5555ff, 0x50282878, 0xa5dfdf7a);
+ &data_word(0x038c8c8f, 0x59a1a1f8, 0x09898980, 0x1a0d0d17);
+ &data_word(0x65bfbfda, 0xd7e6e631, 0x844242c6, 0xd06868b8);
+ &data_word(0x824141c3, 0x299999b0, 0x5a2d2d77, 0x1e0f0f11);
+ &data_word(0x7bb0b0cb, 0xa85454fc, 0x6dbbbbd6, 0x2c16163a);
+#Te1:
+ &data_word(0xa5c66363, 0x84f87c7c, 0x99ee7777, 0x8df67b7b);
+ &data_word(0x0dfff2f2, 0xbdd66b6b, 0xb1de6f6f, 0x5491c5c5);
+ &data_word(0x50603030, 0x03020101, 0xa9ce6767, 0x7d562b2b);
+ &data_word(0x19e7fefe, 0x62b5d7d7, 0xe64dabab, 0x9aec7676);
+ &data_word(0x458fcaca, 0x9d1f8282, 0x4089c9c9, 0x87fa7d7d);
+ &data_word(0x15effafa, 0xebb25959, 0xc98e4747, 0x0bfbf0f0);
+ &data_word(0xec41adad, 0x67b3d4d4, 0xfd5fa2a2, 0xea45afaf);
+ &data_word(0xbf239c9c, 0xf753a4a4, 0x96e47272, 0x5b9bc0c0);
+ &data_word(0xc275b7b7, 0x1ce1fdfd, 0xae3d9393, 0x6a4c2626);
+ &data_word(0x5a6c3636, 0x417e3f3f, 0x02f5f7f7, 0x4f83cccc);
+ &data_word(0x5c683434, 0xf451a5a5, 0x34d1e5e5, 0x08f9f1f1);
+ &data_word(0x93e27171, 0x73abd8d8, 0x53623131, 0x3f2a1515);
+ &data_word(0x0c080404, 0x5295c7c7, 0x65462323, 0x5e9dc3c3);
+ &data_word(0x28301818, 0xa1379696, 0x0f0a0505, 0xb52f9a9a);
+ &data_word(0x090e0707, 0x36241212, 0x9b1b8080, 0x3ddfe2e2);
+ &data_word(0x26cdebeb, 0x694e2727, 0xcd7fb2b2, 0x9fea7575);
+ &data_word(0x1b120909, 0x9e1d8383, 0x74582c2c, 0x2e341a1a);
+ &data_word(0x2d361b1b, 0xb2dc6e6e, 0xeeb45a5a, 0xfb5ba0a0);
+ &data_word(0xf6a45252, 0x4d763b3b, 0x61b7d6d6, 0xce7db3b3);
+ &data_word(0x7b522929, 0x3edde3e3, 0x715e2f2f, 0x97138484);
+ &data_word(0xf5a65353, 0x68b9d1d1, 0x00000000, 0x2cc1eded);
+ &data_word(0x60402020, 0x1fe3fcfc, 0xc879b1b1, 0xedb65b5b);
+ &data_word(0xbed46a6a, 0x468dcbcb, 0xd967bebe, 0x4b723939);
+ &data_word(0xde944a4a, 0xd4984c4c, 0xe8b05858, 0x4a85cfcf);
+ &data_word(0x6bbbd0d0, 0x2ac5efef, 0xe54faaaa, 0x16edfbfb);
+ &data_word(0xc5864343, 0xd79a4d4d, 0x55663333, 0x94118585);
+ &data_word(0xcf8a4545, 0x10e9f9f9, 0x06040202, 0x81fe7f7f);
+ &data_word(0xf0a05050, 0x44783c3c, 0xba259f9f, 0xe34ba8a8);
+ &data_word(0xf3a25151, 0xfe5da3a3, 0xc0804040, 0x8a058f8f);
+ &data_word(0xad3f9292, 0xbc219d9d, 0x48703838, 0x04f1f5f5);
+ &data_word(0xdf63bcbc, 0xc177b6b6, 0x75afdada, 0x63422121);
+ &data_word(0x30201010, 0x1ae5ffff, 0x0efdf3f3, 0x6dbfd2d2);
+ &data_word(0x4c81cdcd, 0x14180c0c, 0x35261313, 0x2fc3ecec);
+ &data_word(0xe1be5f5f, 0xa2359797, 0xcc884444, 0x392e1717);
+ &data_word(0x5793c4c4, 0xf255a7a7, 0x82fc7e7e, 0x477a3d3d);
+ &data_word(0xacc86464, 0xe7ba5d5d, 0x2b321919, 0x95e67373);
+ &data_word(0xa0c06060, 0x98198181, 0xd19e4f4f, 0x7fa3dcdc);
+ &data_word(0x66442222, 0x7e542a2a, 0xab3b9090, 0x830b8888);
+ &data_word(0xca8c4646, 0x29c7eeee, 0xd36bb8b8, 0x3c281414);
+ &data_word(0x79a7dede, 0xe2bc5e5e, 0x1d160b0b, 0x76addbdb);
+ &data_word(0x3bdbe0e0, 0x56643232, 0x4e743a3a, 0x1e140a0a);
+ &data_word(0xdb924949, 0x0a0c0606, 0x6c482424, 0xe4b85c5c);
+ &data_word(0x5d9fc2c2, 0x6ebdd3d3, 0xef43acac, 0xa6c46262);
+ &data_word(0xa8399191, 0xa4319595, 0x37d3e4e4, 0x8bf27979);
+ &data_word(0x32d5e7e7, 0x438bc8c8, 0x596e3737, 0xb7da6d6d);
+ &data_word(0x8c018d8d, 0x64b1d5d5, 0xd29c4e4e, 0xe049a9a9);
+ &data_word(0xb4d86c6c, 0xfaac5656, 0x07f3f4f4, 0x25cfeaea);
+ &data_word(0xafca6565, 0x8ef47a7a, 0xe947aeae, 0x18100808);
+ &data_word(0xd56fbaba, 0x88f07878, 0x6f4a2525, 0x725c2e2e);
+ &data_word(0x24381c1c, 0xf157a6a6, 0xc773b4b4, 0x5197c6c6);
+ &data_word(0x23cbe8e8, 0x7ca1dddd, 0x9ce87474, 0x213e1f1f);
+ &data_word(0xdd964b4b, 0xdc61bdbd, 0x860d8b8b, 0x850f8a8a);
+ &data_word(0x90e07070, 0x427c3e3e, 0xc471b5b5, 0xaacc6666);
+ &data_word(0xd8904848, 0x05060303, 0x01f7f6f6, 0x121c0e0e);
+ &data_word(0xa3c26161, 0x5f6a3535, 0xf9ae5757, 0xd069b9b9);
+ &data_word(0x91178686, 0x5899c1c1, 0x273a1d1d, 0xb9279e9e);
+ &data_word(0x38d9e1e1, 0x13ebf8f8, 0xb32b9898, 0x33221111);
+ &data_word(0xbbd26969, 0x70a9d9d9, 0x89078e8e, 0xa7339494);
+ &data_word(0xb62d9b9b, 0x223c1e1e, 0x92158787, 0x20c9e9e9);
+ &data_word(0x4987cece, 0xffaa5555, 0x78502828, 0x7aa5dfdf);
+ &data_word(0x8f038c8c, 0xf859a1a1, 0x80098989, 0x171a0d0d);
+ &data_word(0xda65bfbf, 0x31d7e6e6, 0xc6844242, 0xb8d06868);
+ &data_word(0xc3824141, 0xb0299999, 0x775a2d2d, 0x111e0f0f);
+ &data_word(0xcb7bb0b0, 0xfca85454, 0xd66dbbbb, 0x3a2c1616);
+#Te2:
+ &data_word(0x63a5c663, 0x7c84f87c, 0x7799ee77, 0x7b8df67b);
+ &data_word(0xf20dfff2, 0x6bbdd66b, 0x6fb1de6f, 0xc55491c5);
+ &data_word(0x30506030, 0x01030201, 0x67a9ce67, 0x2b7d562b);
+ &data_word(0xfe19e7fe, 0xd762b5d7, 0xabe64dab, 0x769aec76);
+ &data_word(0xca458fca, 0x829d1f82, 0xc94089c9, 0x7d87fa7d);
+ &data_word(0xfa15effa, 0x59ebb259, 0x47c98e47, 0xf00bfbf0);
+ &data_word(0xadec41ad, 0xd467b3d4, 0xa2fd5fa2, 0xafea45af);
+ &data_word(0x9cbf239c, 0xa4f753a4, 0x7296e472, 0xc05b9bc0);
+ &data_word(0xb7c275b7, 0xfd1ce1fd, 0x93ae3d93, 0x266a4c26);
+ &data_word(0x365a6c36, 0x3f417e3f, 0xf702f5f7, 0xcc4f83cc);
+ &data_word(0x345c6834, 0xa5f451a5, 0xe534d1e5, 0xf108f9f1);
+ &data_word(0x7193e271, 0xd873abd8, 0x31536231, 0x153f2a15);
+ &data_word(0x040c0804, 0xc75295c7, 0x23654623, 0xc35e9dc3);
+ &data_word(0x18283018, 0x96a13796, 0x050f0a05, 0x9ab52f9a);
+ &data_word(0x07090e07, 0x12362412, 0x809b1b80, 0xe23ddfe2);
+ &data_word(0xeb26cdeb, 0x27694e27, 0xb2cd7fb2, 0x759fea75);
+ &data_word(0x091b1209, 0x839e1d83, 0x2c74582c, 0x1a2e341a);
+ &data_word(0x1b2d361b, 0x6eb2dc6e, 0x5aeeb45a, 0xa0fb5ba0);
+ &data_word(0x52f6a452, 0x3b4d763b, 0xd661b7d6, 0xb3ce7db3);
+ &data_word(0x297b5229, 0xe33edde3, 0x2f715e2f, 0x84971384);
+ &data_word(0x53f5a653, 0xd168b9d1, 0x00000000, 0xed2cc1ed);
+ &data_word(0x20604020, 0xfc1fe3fc, 0xb1c879b1, 0x5bedb65b);
+ &data_word(0x6abed46a, 0xcb468dcb, 0xbed967be, 0x394b7239);
+ &data_word(0x4ade944a, 0x4cd4984c, 0x58e8b058, 0xcf4a85cf);
+ &data_word(0xd06bbbd0, 0xef2ac5ef, 0xaae54faa, 0xfb16edfb);
+ &data_word(0x43c58643, 0x4dd79a4d, 0x33556633, 0x85941185);
+ &data_word(0x45cf8a45, 0xf910e9f9, 0x02060402, 0x7f81fe7f);
+ &data_word(0x50f0a050, 0x3c44783c, 0x9fba259f, 0xa8e34ba8);
+ &data_word(0x51f3a251, 0xa3fe5da3, 0x40c08040, 0x8f8a058f);
+ &data_word(0x92ad3f92, 0x9dbc219d, 0x38487038, 0xf504f1f5);
+ &data_word(0xbcdf63bc, 0xb6c177b6, 0xda75afda, 0x21634221);
+ &data_word(0x10302010, 0xff1ae5ff, 0xf30efdf3, 0xd26dbfd2);
+ &data_word(0xcd4c81cd, 0x0c14180c, 0x13352613, 0xec2fc3ec);
+ &data_word(0x5fe1be5f, 0x97a23597, 0x44cc8844, 0x17392e17);
+ &data_word(0xc45793c4, 0xa7f255a7, 0x7e82fc7e, 0x3d477a3d);
+ &data_word(0x64acc864, 0x5de7ba5d, 0x192b3219, 0x7395e673);
+ &data_word(0x60a0c060, 0x81981981, 0x4fd19e4f, 0xdc7fa3dc);
+ &data_word(0x22664422, 0x2a7e542a, 0x90ab3b90, 0x88830b88);
+ &data_word(0x46ca8c46, 0xee29c7ee, 0xb8d36bb8, 0x143c2814);
+ &data_word(0xde79a7de, 0x5ee2bc5e, 0x0b1d160b, 0xdb76addb);
+ &data_word(0xe03bdbe0, 0x32566432, 0x3a4e743a, 0x0a1e140a);
+ &data_word(0x49db9249, 0x060a0c06, 0x246c4824, 0x5ce4b85c);
+ &data_word(0xc25d9fc2, 0xd36ebdd3, 0xacef43ac, 0x62a6c462);
+ &data_word(0x91a83991, 0x95a43195, 0xe437d3e4, 0x798bf279);
+ &data_word(0xe732d5e7, 0xc8438bc8, 0x37596e37, 0x6db7da6d);
+ &data_word(0x8d8c018d, 0xd564b1d5, 0x4ed29c4e, 0xa9e049a9);
+ &data_word(0x6cb4d86c, 0x56faac56, 0xf407f3f4, 0xea25cfea);
+ &data_word(0x65afca65, 0x7a8ef47a, 0xaee947ae, 0x08181008);
+ &data_word(0xbad56fba, 0x7888f078, 0x256f4a25, 0x2e725c2e);
+ &data_word(0x1c24381c, 0xa6f157a6, 0xb4c773b4, 0xc65197c6);
+ &data_word(0xe823cbe8, 0xdd7ca1dd, 0x749ce874, 0x1f213e1f);
+ &data_word(0x4bdd964b, 0xbddc61bd, 0x8b860d8b, 0x8a850f8a);
+ &data_word(0x7090e070, 0x3e427c3e, 0xb5c471b5, 0x66aacc66);
+ &data_word(0x48d89048, 0x03050603, 0xf601f7f6, 0x0e121c0e);
+ &data_word(0x61a3c261, 0x355f6a35, 0x57f9ae57, 0xb9d069b9);
+ &data_word(0x86911786, 0xc15899c1, 0x1d273a1d, 0x9eb9279e);
+ &data_word(0xe138d9e1, 0xf813ebf8, 0x98b32b98, 0x11332211);
+ &data_word(0x69bbd269, 0xd970a9d9, 0x8e89078e, 0x94a73394);
+ &data_word(0x9bb62d9b, 0x1e223c1e, 0x87921587, 0xe920c9e9);
+ &data_word(0xce4987ce, 0x55ffaa55, 0x28785028, 0xdf7aa5df);
+ &data_word(0x8c8f038c, 0xa1f859a1, 0x89800989, 0x0d171a0d);
+ &data_word(0xbfda65bf, 0xe631d7e6, 0x42c68442, 0x68b8d068);
+ &data_word(0x41c38241, 0x99b02999, 0x2d775a2d, 0x0f111e0f);
+ &data_word(0xb0cb7bb0, 0x54fca854, 0xbbd66dbb, 0x163a2c16);
+#Te3:
+ &data_word(0x6363a5c6, 0x7c7c84f8, 0x777799ee, 0x7b7b8df6);
+ &data_word(0xf2f20dff, 0x6b6bbdd6, 0x6f6fb1de, 0xc5c55491);
+ &data_word(0x30305060, 0x01010302, 0x6767a9ce, 0x2b2b7d56);
+ &data_word(0xfefe19e7, 0xd7d762b5, 0xababe64d, 0x76769aec);
+ &data_word(0xcaca458f, 0x82829d1f, 0xc9c94089, 0x7d7d87fa);
+ &data_word(0xfafa15ef, 0x5959ebb2, 0x4747c98e, 0xf0f00bfb);
+ &data_word(0xadadec41, 0xd4d467b3, 0xa2a2fd5f, 0xafafea45);
+ &data_word(0x9c9cbf23, 0xa4a4f753, 0x727296e4, 0xc0c05b9b);
+ &data_word(0xb7b7c275, 0xfdfd1ce1, 0x9393ae3d, 0x26266a4c);
+ &data_word(0x36365a6c, 0x3f3f417e, 0xf7f702f5, 0xcccc4f83);
+ &data_word(0x34345c68, 0xa5a5f451, 0xe5e534d1, 0xf1f108f9);
+ &data_word(0x717193e2, 0xd8d873ab, 0x31315362, 0x15153f2a);
+ &data_word(0x04040c08, 0xc7c75295, 0x23236546, 0xc3c35e9d);
+ &data_word(0x18182830, 0x9696a137, 0x05050f0a, 0x9a9ab52f);
+ &data_word(0x0707090e, 0x12123624, 0x80809b1b, 0xe2e23ddf);
+ &data_word(0xebeb26cd, 0x2727694e, 0xb2b2cd7f, 0x75759fea);
+ &data_word(0x09091b12, 0x83839e1d, 0x2c2c7458, 0x1a1a2e34);
+ &data_word(0x1b1b2d36, 0x6e6eb2dc, 0x5a5aeeb4, 0xa0a0fb5b);
+ &data_word(0x5252f6a4, 0x3b3b4d76, 0xd6d661b7, 0xb3b3ce7d);
+ &data_word(0x29297b52, 0xe3e33edd, 0x2f2f715e, 0x84849713);
+ &data_word(0x5353f5a6, 0xd1d168b9, 0x00000000, 0xeded2cc1);
+ &data_word(0x20206040, 0xfcfc1fe3, 0xb1b1c879, 0x5b5bedb6);
+ &data_word(0x6a6abed4, 0xcbcb468d, 0xbebed967, 0x39394b72);
+ &data_word(0x4a4ade94, 0x4c4cd498, 0x5858e8b0, 0xcfcf4a85);
+ &data_word(0xd0d06bbb, 0xefef2ac5, 0xaaaae54f, 0xfbfb16ed);
+ &data_word(0x4343c586, 0x4d4dd79a, 0x33335566, 0x85859411);
+ &data_word(0x4545cf8a, 0xf9f910e9, 0x02020604, 0x7f7f81fe);
+ &data_word(0x5050f0a0, 0x3c3c4478, 0x9f9fba25, 0xa8a8e34b);
+ &data_word(0x5151f3a2, 0xa3a3fe5d, 0x4040c080, 0x8f8f8a05);
+ &data_word(0x9292ad3f, 0x9d9dbc21, 0x38384870, 0xf5f504f1);
+ &data_word(0xbcbcdf63, 0xb6b6c177, 0xdada75af, 0x21216342);
+ &data_word(0x10103020, 0xffff1ae5, 0xf3f30efd, 0xd2d26dbf);
+ &data_word(0xcdcd4c81, 0x0c0c1418, 0x13133526, 0xecec2fc3);
+ &data_word(0x5f5fe1be, 0x9797a235, 0x4444cc88, 0x1717392e);
+ &data_word(0xc4c45793, 0xa7a7f255, 0x7e7e82fc, 0x3d3d477a);
+ &data_word(0x6464acc8, 0x5d5de7ba, 0x19192b32, 0x737395e6);
+ &data_word(0x6060a0c0, 0x81819819, 0x4f4fd19e, 0xdcdc7fa3);
+ &data_word(0x22226644, 0x2a2a7e54, 0x9090ab3b, 0x8888830b);
+ &data_word(0x4646ca8c, 0xeeee29c7, 0xb8b8d36b, 0x14143c28);
+ &data_word(0xdede79a7, 0x5e5ee2bc, 0x0b0b1d16, 0xdbdb76ad);
+ &data_word(0xe0e03bdb, 0x32325664, 0x3a3a4e74, 0x0a0a1e14);
+ &data_word(0x4949db92, 0x06060a0c, 0x24246c48, 0x5c5ce4b8);
+ &data_word(0xc2c25d9f, 0xd3d36ebd, 0xacacef43, 0x6262a6c4);
+ &data_word(0x9191a839, 0x9595a431, 0xe4e437d3, 0x79798bf2);
+ &data_word(0xe7e732d5, 0xc8c8438b, 0x3737596e, 0x6d6db7da);
+ &data_word(0x8d8d8c01, 0xd5d564b1, 0x4e4ed29c, 0xa9a9e049);
+ &data_word(0x6c6cb4d8, 0x5656faac, 0xf4f407f3, 0xeaea25cf);
+ &data_word(0x6565afca, 0x7a7a8ef4, 0xaeaee947, 0x08081810);
+ &data_word(0xbabad56f, 0x787888f0, 0x25256f4a, 0x2e2e725c);
+ &data_word(0x1c1c2438, 0xa6a6f157, 0xb4b4c773, 0xc6c65197);
+ &data_word(0xe8e823cb, 0xdddd7ca1, 0x74749ce8, 0x1f1f213e);
+ &data_word(0x4b4bdd96, 0xbdbddc61, 0x8b8b860d, 0x8a8a850f);
+ &data_word(0x707090e0, 0x3e3e427c, 0xb5b5c471, 0x6666aacc);
+ &data_word(0x4848d890, 0x03030506, 0xf6f601f7, 0x0e0e121c);
+ &data_word(0x6161a3c2, 0x35355f6a, 0x5757f9ae, 0xb9b9d069);
+ &data_word(0x86869117, 0xc1c15899, 0x1d1d273a, 0x9e9eb927);
+ &data_word(0xe1e138d9, 0xf8f813eb, 0x9898b32b, 0x11113322);
+ &data_word(0x6969bbd2, 0xd9d970a9, 0x8e8e8907, 0x9494a733);
+ &data_word(0x9b9bb62d, 0x1e1e223c, 0x87879215, 0xe9e920c9);
+ &data_word(0xcece4987, 0x5555ffaa, 0x28287850, 0xdfdf7aa5);
+ &data_word(0x8c8c8f03, 0xa1a1f859, 0x89898009, 0x0d0d171a);
+ &data_word(0xbfbfda65, 0xe6e631d7, 0x4242c684, 0x6868b8d0);
+ &data_word(0x4141c382, 0x9999b029, 0x2d2d775a, 0x0f0f111e);
+ &data_word(0xb0b0cb7b, 0x5454fca8, 0xbbbbd66d, 0x16163a2c);
+#Te4:
+ &data_word(0x63636363, 0x7c7c7c7c, 0x77777777, 0x7b7b7b7b);
+ &data_word(0xf2f2f2f2, 0x6b6b6b6b, 0x6f6f6f6f, 0xc5c5c5c5);
+ &data_word(0x30303030, 0x01010101, 0x67676767, 0x2b2b2b2b);
+ &data_word(0xfefefefe, 0xd7d7d7d7, 0xabababab, 0x76767676);
+ &data_word(0xcacacaca, 0x82828282, 0xc9c9c9c9, 0x7d7d7d7d);
+ &data_word(0xfafafafa, 0x59595959, 0x47474747, 0xf0f0f0f0);
+ &data_word(0xadadadad, 0xd4d4d4d4, 0xa2a2a2a2, 0xafafafaf);
+ &data_word(0x9c9c9c9c, 0xa4a4a4a4, 0x72727272, 0xc0c0c0c0);
+ &data_word(0xb7b7b7b7, 0xfdfdfdfd, 0x93939393, 0x26262626);
+ &data_word(0x36363636, 0x3f3f3f3f, 0xf7f7f7f7, 0xcccccccc);
+ &data_word(0x34343434, 0xa5a5a5a5, 0xe5e5e5e5, 0xf1f1f1f1);
+ &data_word(0x71717171, 0xd8d8d8d8, 0x31313131, 0x15151515);
+ &data_word(0x04040404, 0xc7c7c7c7, 0x23232323, 0xc3c3c3c3);
+ &data_word(0x18181818, 0x96969696, 0x05050505, 0x9a9a9a9a);
+ &data_word(0x07070707, 0x12121212, 0x80808080, 0xe2e2e2e2);
+ &data_word(0xebebebeb, 0x27272727, 0xb2b2b2b2, 0x75757575);
+ &data_word(0x09090909, 0x83838383, 0x2c2c2c2c, 0x1a1a1a1a);
+ &data_word(0x1b1b1b1b, 0x6e6e6e6e, 0x5a5a5a5a, 0xa0a0a0a0);
+ &data_word(0x52525252, 0x3b3b3b3b, 0xd6d6d6d6, 0xb3b3b3b3);
+ &data_word(0x29292929, 0xe3e3e3e3, 0x2f2f2f2f, 0x84848484);
+ &data_word(0x53535353, 0xd1d1d1d1, 0x00000000, 0xedededed);
+ &data_word(0x20202020, 0xfcfcfcfc, 0xb1b1b1b1, 0x5b5b5b5b);
+ &data_word(0x6a6a6a6a, 0xcbcbcbcb, 0xbebebebe, 0x39393939);
+ &data_word(0x4a4a4a4a, 0x4c4c4c4c, 0x58585858, 0xcfcfcfcf);
+ &data_word(0xd0d0d0d0, 0xefefefef, 0xaaaaaaaa, 0xfbfbfbfb);
+ &data_word(0x43434343, 0x4d4d4d4d, 0x33333333, 0x85858585);
+ &data_word(0x45454545, 0xf9f9f9f9, 0x02020202, 0x7f7f7f7f);
+ &data_word(0x50505050, 0x3c3c3c3c, 0x9f9f9f9f, 0xa8a8a8a8);
+ &data_word(0x51515151, 0xa3a3a3a3, 0x40404040, 0x8f8f8f8f);
+ &data_word(0x92929292, 0x9d9d9d9d, 0x38383838, 0xf5f5f5f5);
+ &data_word(0xbcbcbcbc, 0xb6b6b6b6, 0xdadadada, 0x21212121);
+ &data_word(0x10101010, 0xffffffff, 0xf3f3f3f3, 0xd2d2d2d2);
+ &data_word(0xcdcdcdcd, 0x0c0c0c0c, 0x13131313, 0xecececec);
+ &data_word(0x5f5f5f5f, 0x97979797, 0x44444444, 0x17171717);
+ &data_word(0xc4c4c4c4, 0xa7a7a7a7, 0x7e7e7e7e, 0x3d3d3d3d);
+ &data_word(0x64646464, 0x5d5d5d5d, 0x19191919, 0x73737373);
+ &data_word(0x60606060, 0x81818181, 0x4f4f4f4f, 0xdcdcdcdc);
+ &data_word(0x22222222, 0x2a2a2a2a, 0x90909090, 0x88888888);
+ &data_word(0x46464646, 0xeeeeeeee, 0xb8b8b8b8, 0x14141414);
+ &data_word(0xdededede, 0x5e5e5e5e, 0x0b0b0b0b, 0xdbdbdbdb);
+ &data_word(0xe0e0e0e0, 0x32323232, 0x3a3a3a3a, 0x0a0a0a0a);
+ &data_word(0x49494949, 0x06060606, 0x24242424, 0x5c5c5c5c);
+ &data_word(0xc2c2c2c2, 0xd3d3d3d3, 0xacacacac, 0x62626262);
+ &data_word(0x91919191, 0x95959595, 0xe4e4e4e4, 0x79797979);
+ &data_word(0xe7e7e7e7, 0xc8c8c8c8, 0x37373737, 0x6d6d6d6d);
+ &data_word(0x8d8d8d8d, 0xd5d5d5d5, 0x4e4e4e4e, 0xa9a9a9a9);
+ &data_word(0x6c6c6c6c, 0x56565656, 0xf4f4f4f4, 0xeaeaeaea);
+ &data_word(0x65656565, 0x7a7a7a7a, 0xaeaeaeae, 0x08080808);
+ &data_word(0xbabababa, 0x78787878, 0x25252525, 0x2e2e2e2e);
+ &data_word(0x1c1c1c1c, 0xa6a6a6a6, 0xb4b4b4b4, 0xc6c6c6c6);
+ &data_word(0xe8e8e8e8, 0xdddddddd, 0x74747474, 0x1f1f1f1f);
+ &data_word(0x4b4b4b4b, 0xbdbdbdbd, 0x8b8b8b8b, 0x8a8a8a8a);
+ &data_word(0x70707070, 0x3e3e3e3e, 0xb5b5b5b5, 0x66666666);
+ &data_word(0x48484848, 0x03030303, 0xf6f6f6f6, 0x0e0e0e0e);
+ &data_word(0x61616161, 0x35353535, 0x57575757, 0xb9b9b9b9);
+ &data_word(0x86868686, 0xc1c1c1c1, 0x1d1d1d1d, 0x9e9e9e9e);
+ &data_word(0xe1e1e1e1, 0xf8f8f8f8, 0x98989898, 0x11111111);
+ &data_word(0x69696969, 0xd9d9d9d9, 0x8e8e8e8e, 0x94949494);
+ &data_word(0x9b9b9b9b, 0x1e1e1e1e, 0x87878787, 0xe9e9e9e9);
+ &data_word(0xcececece, 0x55555555, 0x28282828, 0xdfdfdfdf);
+ &data_word(0x8c8c8c8c, 0xa1a1a1a1, 0x89898989, 0x0d0d0d0d);
+ &data_word(0xbfbfbfbf, 0xe6e6e6e6, 0x42424242, 0x68686868);
+ &data_word(0x41414141, 0x99999999, 0x2d2d2d2d, 0x0f0f0f0f);
+ &data_word(0xb0b0b0b0, 0x54545454, 0xbbbbbbbb, 0x16161616);
+&function_end_B("AES_encrypt");
+
+sub decstep()
+{ my ($i,$td,@s) = @_;
+ my $tmp,$out;
+
+ if ($i==3) { $out=$s[0]; &mov ("edi",&DWP(12,"esp"));}
+ else { $out="esi"; &mov ($out,$s[0]); }
+ &shr ($out,24-2);
+ &and ($out,0xFF<<2);
+ &mov ($out,&DWP(1024*0,$td,$out));
+
+ if ($i==3) { $tmp=$s[1]; }
+ else { $tmp="edi"; &mov ($tmp,$s[1]); }
+ &shr ($tmp,16-2);
+ &and ($tmp,0xFF<<2);
+ &xor ($out,&DWP(1024*1,$td,$tmp));
+
+ if ($i==3) { $tmp=$s[2]; &mov ($s[1],"esi"); }
+ else { $tmp="edi"; }
+ &movz ($tmp,&HB($s[2]));
+ &xor ($out,&DWP(1024*2,$td,$tmp,4));
+
+ if ($i==3) { $tmp=$s[3]; &mov ($s[2],&DWP(4,"esp")); }
+ else { $tmp="edi"; &mov ($tmp,$s[3]); }
+ &and ($tmp,0xFF);
+ &xor ($out,&DWP(1024*3,$td,$tmp,4));
+ if ($i<2) { &mov (&DWP(4*$i,"esp"),$out); }
+ if ($i==3) { &mov ($s[3],&DWP(0,"esp")); }
+}
+
+sub declast()
+{ my ($i,$td,@s)=@_;
+ my $tmp,$out;
+
+ if ($i==3) { $out=$s[0]; &mov ("edi",&DWP(12,"esp"));}
+ else { $out="esi"; &mov ($out,$s[0]); }
+ &shr ($out,24-2);
+ &and ($out,0xFF<<2);
+ &mov ($out,&DWP(0,$td,$out));
+ &and ($out,0xff000000);
+
+ if ($i==3) { $tmp=$s[1]; }
+ else { $tmp="edi"; &mov ($tmp,$s[1]); }
+ &shr ($tmp,16-2);
+ &and ($tmp,0xFF<<2);
+ &mov ($tmp,&DWP(0,$td,$tmp));
+ &and ($tmp,0x00ff0000);
+ &xor ($out,$tmp);
+
+ if ($i==3) { $tmp=$s[2]; &mov ($s[1],"esi"); }
+ else { $tmp="edi"; }
+ &movz ($tmp,&HB($s[2]));
+ &mov ($tmp,&DWP(0,$td,$tmp,4));
+ &and ($tmp,0x0000ff00);
+ &xor ($out,$tmp);
+
+ if ($i==3) { $tmp=$s[3]; &mov ($s[2],&DWP(4,"esp")); }
+ else { $tmp="edi"; &mov ($tmp,$s[3]); }
+ &and ($tmp,0xFF);
+ &mov ($tmp,&DWP(0,$td,$tmp,4));
+ &and ($tmp,0x000000ff);
+ &xor ($out,$tmp);
+ if ($i<2) { &mov (&DWP(4*$i,"esp"),$out); }
+ if ($i==3) { &mov ($s[3],&DWP(0,"esp")); }
+}
+
+# void AES_decrypt (const void *inp,void *out,const AES_KEY *key);
+&public_label("AES_Td");
+&function_begin("AES_decrypt");
+ &mov ("esi",&wparam(0)); # load inp
+ &mov ("edi",&wparam(2)); # load key
+
+ &call (&label("pic_point")); # make it PIC!
+&set_label("pic_point");
+ &blindpop("ebp");
+ &lea ("ebp",&DWP(&label("AES_Td")."-".&label("pic_point"),"ebp"));
+
+ # allocate aligned stack frame
+ &mov ("eax","esp");
+ &sub ("esp",20);
+ &and ("esp",-16);
+
+ &mov (&DWP(12,"esp"),"edi"); # save key
+ &mov (&DWP(16,"esp"),"eax"); # save %esp
+
+ &mov ($s0,&DWP(0,"esi")); # load input data
+ &mov ($s1,&DWP(4,"esi"));
+ &mov ($s2,&DWP(8,"esi"));
+ &mov ($s3,&DWP(12,"esi"));
+ &bswap ($s0);
+ &bswap ($s1);
+ &bswap ($s2);
+ &bswap ($s3);
+ &xor ($s0,&DWP(0,"edi"));
+ &xor ($s1,&DWP(4,"edi"));
+ &xor ($s2,&DWP(8,"edi"));
+ &xor ($s3,&DWP(12,"edi"));
+
+ &mov ("esi",&DWP(240,"edi")); # load key->rounds
+
+ if ($small_footprint) {
+ &lea ("esi",&DWP(-2,"esi","esi"));
+ &lea ("esi",&DWP(0,"edi","esi",8));
+ &mov (&DWP(8,"esp"),"esi"); # end of key schedule
+ &align (4);
+ &set_label("loop");
+ &decstep(0,"ebp",$s0,$s3,$s2,$s1);
+ &decstep(1,"ebp",$s1,$s0,$s3,$s2);
+ &decstep(2,"ebp",$s2,$s1,$s0,$s3);
+ &decstep(3,"ebp",$s3,$s2,$s1,$s0);
+ &add ("edi",16); # advance rd_key
+ &xor ($s0,&DWP(0,"edi"));
+ &xor ($s1,&DWP(4,"edi"));
+ &xor ($s2,&DWP(8,"edi"));
+ &xor ($s3,&DWP(12,"edi"));
+ &cmp ("edi",&DWP(8,"esp"));
+ &mov (&DWP(12,"esp"),"edi");
+ &jb (&label("loop"));
+ }
+ else {
+ &cmp ("esi",10);
+ &jle (&label("10rounds"));
+ &cmp ("esi",12);
+ &jle (&label("12rounds"));
+
+ &set_label("14rounds");
+ for ($i=1;$i<3;$i++) {
+ &decstep(0,"ebp",$s0,$s3,$s2,$s1);
+ &decstep(1,"ebp",$s1,$s0,$s3,$s2);
+ &decstep(2,"ebp",$s2,$s1,$s0,$s3);
+ &decstep(3,"ebp",$s3,$s2,$s1,$s0);
+ &xor ($s0,&DWP(16*$i+0,"edi"));
+ &xor ($s1,&DWP(16*$i+4,"edi"));
+ &xor ($s2,&DWP(16*$i+8,"edi"));
+ &xor ($s3,&DWP(16*$i+12,"edi"));
+ }
+ &add ("edi",32);
+ &mov (&DWP(12,"esp"),"edi"); # advance rd_key
+ &set_label("12rounds");
+ for ($i=1;$i<3;$i++) {
+ &decstep(0,"ebp",$s0,$s3,$s2,$s1);
+ &decstep(1,"ebp",$s1,$s0,$s3,$s2);
+ &decstep(2,"ebp",$s2,$s1,$s0,$s3);
+ &decstep(3,"ebp",$s3,$s2,$s1,$s0);
+ &xor ($s0,&DWP(16*$i+0,"edi"));
+ &xor ($s1,&DWP(16*$i+4,"edi"));
+ &xor ($s2,&DWP(16*$i+8,"edi"));
+ &xor ($s3,&DWP(16*$i+12,"edi"));
+ }
+ &add ("edi",32);
+ &mov (&DWP(12,"esp"),"edi"); # advance rd_key
+ &set_label("10rounds");
+ for ($i=1;$i<10;$i++) {
+ &decstep(0,"ebp",$s0,$s3,$s2,$s1);
+ &decstep(1,"ebp",$s1,$s0,$s3,$s2);
+ &decstep(2,"ebp",$s2,$s1,$s0,$s3);
+ &decstep(3,"ebp",$s3,$s2,$s1,$s0);
+ &xor ($s0,&DWP(16*$i+0,"edi"));
+ &xor ($s1,&DWP(16*$i+4,"edi"));
+ &xor ($s2,&DWP(16*$i+8,"edi"));
+ &xor ($s3,&DWP(16*$i+12,"edi"));
+ }
+ }
+
+ &add ("ebp",4*1024); # skip to Te4
+ &declast(0,"ebp",$s0,$s3,$s2,$s1);
+ &declast(1,"ebp",$s1,$s0,$s3,$s2);
+ &declast(2,"ebp",$s2,$s1,$s0,$s3);
+ &declast(3,"ebp",$s3,$s2,$s1,$s0);
+
+ &mov ("esp",&DWP(16,"esp")); # restore %esp
+ &add ("edi",$small_footprint?16:160);
+ &xor ($s0,&DWP(0,"edi"));
+ &xor ($s1,&DWP(4,"edi"));
+ &xor ($s2,&DWP(8,"edi"));
+ &xor ($s3,&DWP(12,"edi"));
+ &bswap ($s0);
+ &bswap ($s1);
+ &bswap ($s2);
+ &bswap ($s3);
+ &mov ("edi",&wparam(1)); # load out
+ &mov (&DWP(0,"edi"),$s0); # write output data
+ &mov (&DWP(4,"edi"),$s1);
+ &mov (&DWP(8,"edi"),$s2);
+ &mov (&DWP(12,"edi"),$s3);
+
+ &pop ("edi");
+ &pop ("esi");
+ &pop ("ebx");
+ &pop ("ebp");
+ &ret ();
+
+&set_label("AES_Td",64); # Yes! I keep it in the code segment!
+ &data_word(0x51f4a750, 0x7e416553, 0x1a17a4c3, 0x3a275e96);
+ &data_word(0x3bab6bcb, 0x1f9d45f1, 0xacfa58ab, 0x4be30393);
+ &data_word(0x2030fa55, 0xad766df6, 0x88cc7691, 0xf5024c25);
+ &data_word(0x4fe5d7fc, 0xc52acbd7, 0x26354480, 0xb562a38f);
+ &data_word(0xdeb15a49, 0x25ba1b67, 0x45ea0e98, 0x5dfec0e1);
+ &data_word(0xc32f7502, 0x814cf012, 0x8d4697a3, 0x6bd3f9c6);
+ &data_word(0x038f5fe7, 0x15929c95, 0xbf6d7aeb, 0x955259da);
+ &data_word(0xd4be832d, 0x587421d3, 0x49e06929, 0x8ec9c844);
+ &data_word(0x75c2896a, 0xf48e7978, 0x99583e6b, 0x27b971dd);
+ &data_word(0xbee14fb6, 0xf088ad17, 0xc920ac66, 0x7dce3ab4);
+ &data_word(0x63df4a18, 0xe51a3182, 0x97513360, 0x62537f45);
+ &data_word(0xb16477e0, 0xbb6bae84, 0xfe81a01c, 0xf9082b94);
+ &data_word(0x70486858, 0x8f45fd19, 0x94de6c87, 0x527bf8b7);
+ &data_word(0xab73d323, 0x724b02e2, 0xe31f8f57, 0x6655ab2a);
+ &data_word(0xb2eb2807, 0x2fb5c203, 0x86c57b9a, 0xd33708a5);
+ &data_word(0x302887f2, 0x23bfa5b2, 0x02036aba, 0xed16825c);
+ &data_word(0x8acf1c2b, 0xa779b492, 0xf307f2f0, 0x4e69e2a1);
+ &data_word(0x65daf4cd, 0x0605bed5, 0xd134621f, 0xc4a6fe8a);
+ &data_word(0x342e539d, 0xa2f355a0, 0x058ae132, 0xa4f6eb75);
+ &data_word(0x0b83ec39, 0x4060efaa, 0x5e719f06, 0xbd6e1051);
+ &data_word(0x3e218af9, 0x96dd063d, 0xdd3e05ae, 0x4de6bd46);
+ &data_word(0x91548db5, 0x71c45d05, 0x0406d46f, 0x605015ff);
+ &data_word(0x1998fb24, 0xd6bde997, 0x894043cc, 0x67d99e77);
+ &data_word(0xb0e842bd, 0x07898b88, 0xe7195b38, 0x79c8eedb);
+ &data_word(0xa17c0a47, 0x7c420fe9, 0xf8841ec9, 0x00000000);
+ &data_word(0x09808683, 0x322bed48, 0x1e1170ac, 0x6c5a724e);
+ &data_word(0xfd0efffb, 0x0f853856, 0x3daed51e, 0x362d3927);
+ &data_word(0x0a0fd964, 0x685ca621, 0x9b5b54d1, 0x24362e3a);
+ &data_word(0x0c0a67b1, 0x9357e70f, 0xb4ee96d2, 0x1b9b919e);
+ &data_word(0x80c0c54f, 0x61dc20a2, 0x5a774b69, 0x1c121a16);
+ &data_word(0xe293ba0a, 0xc0a02ae5, 0x3c22e043, 0x121b171d);
+ &data_word(0x0e090d0b, 0xf28bc7ad, 0x2db6a8b9, 0x141ea9c8);
+ &data_word(0x57f11985, 0xaf75074c, 0xee99ddbb, 0xa37f60fd);
+ &data_word(0xf701269f, 0x5c72f5bc, 0x44663bc5, 0x5bfb7e34);
+ &data_word(0x8b432976, 0xcb23c6dc, 0xb6edfc68, 0xb8e4f163);
+ &data_word(0xd731dcca, 0x42638510, 0x13972240, 0x84c61120);
+ &data_word(0x854a247d, 0xd2bb3df8, 0xaef93211, 0xc729a16d);
+ &data_word(0x1d9e2f4b, 0xdcb230f3, 0x0d8652ec, 0x77c1e3d0);
+ &data_word(0x2bb3166c, 0xa970b999, 0x119448fa, 0x47e96422);
+ &data_word(0xa8fc8cc4, 0xa0f03f1a, 0x567d2cd8, 0x223390ef);
+ &data_word(0x87494ec7, 0xd938d1c1, 0x8ccaa2fe, 0x98d40b36);
+ &data_word(0xa6f581cf, 0xa57ade28, 0xdab78e26, 0x3fadbfa4);
+ &data_word(0x2c3a9de4, 0x5078920d, 0x6a5fcc9b, 0x547e4662);
+ &data_word(0xf68d13c2, 0x90d8b8e8, 0x2e39f75e, 0x82c3aff5);
+ &data_word(0x9f5d80be, 0x69d0937c, 0x6fd52da9, 0xcf2512b3);
+ &data_word(0xc8ac993b, 0x10187da7, 0xe89c636e, 0xdb3bbb7b);
+ &data_word(0xcd267809, 0x6e5918f4, 0xec9ab701, 0x834f9aa8);