2 # Copyright 1995-2016 The OpenSSL Project Authors. All Rights Reserved.
4 # Licensed under the Apache License 2.0 (the "License"). You may not use
5 # this file except in compliance with the License. You can obtain a copy
6 # in the file LICENSE in the source distribution or at
7 # https://www.openssl.org/source/license.html
10 # This flag makes the inner loop one cycle longer, but generates
11 # code that runs %30 faster on the pentium pro/II, 44% faster
12 # of PIII, while only %7 slower on the pentium.
13 # By default, this flag is on.
16 $0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
17 push(@INC,"${dir}","${dir}../../perlasm");
21 $output=pop and open STDOUT,">$output";
23 &asm_init($ARGV[0],$ARGV[$#ARGV] eq "386");
38 @F1=("add","xor","sub");
39 @F2=("xor","sub","add");
40 @F3=("sub","add","xor");
42 &CAST_encrypt("CAST_encrypt",1);
43 &CAST_encrypt("CAST_decrypt",0);
44 &cbc("CAST_cbc_encrypt","CAST_encrypt","CAST_decrypt",1,4,5,3,-1,-1);
48 close STDOUT or die "error closing STDOUT";
53 local($win_ex)=<<"EOF";
54 EXTERN _CAST_S_table0:DWORD
55 EXTERN _CAST_S_table1:DWORD
56 EXTERN _CAST_S_table2:DWORD
57 EXTERN _CAST_S_table3:DWORD
59 &main::external_label(
66 &function_begin_B($name,$win_ex);
72 &mov($tmp2,&wparam(0));
77 &comment("Load the 2 words");
78 &mov($L,&DWP(0,$tmp2,"",0));
79 &mov($R,&DWP(4,$tmp2,"",0));
81 &comment('Get short key flag');
82 &mov($tmp3,&DWP(128,$K,"",0));
87 &jnz(&label('cast_dec_skip'));
95 &E_CAST( 0,$S,$L,$R,$K,@F1,$tmp1,$tmp2,$tmp3,$tmp4);
96 &E_CAST( 1,$S,$R,$L,$K,@F2,$tmp1,$tmp2,$tmp3,$tmp4);
97 &E_CAST( 2,$S,$L,$R,$K,@F3,$tmp1,$tmp2,$tmp3,$tmp4);
98 &E_CAST( 3,$S,$R,$L,$K,@F1,$tmp1,$tmp2,$tmp3,$tmp4);
99 &E_CAST( 4,$S,$L,$R,$K,@F2,$tmp1,$tmp2,$tmp3,$tmp4);
100 &E_CAST( 5,$S,$R,$L,$K,@F3,$tmp1,$tmp2,$tmp3,$tmp4);
101 &E_CAST( 6,$S,$L,$R,$K,@F1,$tmp1,$tmp2,$tmp3,$tmp4);
102 &E_CAST( 7,$S,$R,$L,$K,@F2,$tmp1,$tmp2,$tmp3,$tmp4);
103 &E_CAST( 8,$S,$L,$R,$K,@F3,$tmp1,$tmp2,$tmp3,$tmp4);
104 &E_CAST( 9,$S,$R,$L,$K,@F1,$tmp1,$tmp2,$tmp3,$tmp4);
105 &E_CAST(10,$S,$L,$R,$K,@F2,$tmp1,$tmp2,$tmp3,$tmp4);
106 &E_CAST(11,$S,$R,$L,$K,@F3,$tmp1,$tmp2,$tmp3,$tmp4);
107 &comment('test short key flag');
110 &jnz(&label('cast_enc_done'));
111 &E_CAST(12,$S,$L,$R,$K,@F1,$tmp1,$tmp2,$tmp3,$tmp4);
112 &E_CAST(13,$S,$R,$L,$K,@F2,$tmp1,$tmp2,$tmp3,$tmp4);
113 &E_CAST(14,$S,$L,$R,$K,@F3,$tmp1,$tmp2,$tmp3,$tmp4);
114 &E_CAST(15,$S,$R,$L,$K,@F1,$tmp1,$tmp2,$tmp3,$tmp4);
116 &E_CAST(15,$S,$L,$R,$K,@F1,$tmp1,$tmp2,$tmp3,$tmp4);
117 &E_CAST(14,$S,$R,$L,$K,@F3,$tmp1,$tmp2,$tmp3,$tmp4);
118 &E_CAST(13,$S,$L,$R,$K,@F2,$tmp1,$tmp2,$tmp3,$tmp4);
119 &E_CAST(12,$S,$R,$L,$K,@F1,$tmp1,$tmp2,$tmp3,$tmp4);
120 &set_label('cast_dec_skip');
121 &E_CAST(11,$S,$L,$R,$K,@F3,$tmp1,$tmp2,$tmp3,$tmp4);
122 &E_CAST(10,$S,$R,$L,$K,@F2,$tmp1,$tmp2,$tmp3,$tmp4);
123 &E_CAST( 9,$S,$L,$R,$K,@F1,$tmp1,$tmp2,$tmp3,$tmp4);
124 &E_CAST( 8,$S,$R,$L,$K,@F3,$tmp1,$tmp2,$tmp3,$tmp4);
125 &E_CAST( 7,$S,$L,$R,$K,@F2,$tmp1,$tmp2,$tmp3,$tmp4);
126 &E_CAST( 6,$S,$R,$L,$K,@F1,$tmp1,$tmp2,$tmp3,$tmp4);
127 &E_CAST( 5,$S,$L,$R,$K,@F3,$tmp1,$tmp2,$tmp3,$tmp4);
128 &E_CAST( 4,$S,$R,$L,$K,@F2,$tmp1,$tmp2,$tmp3,$tmp4);
129 &E_CAST( 3,$S,$L,$R,$K,@F1,$tmp1,$tmp2,$tmp3,$tmp4);
130 &E_CAST( 2,$S,$R,$L,$K,@F3,$tmp1,$tmp2,$tmp3,$tmp4);
131 &E_CAST( 1,$S,$L,$R,$K,@F2,$tmp1,$tmp2,$tmp3,$tmp4);
132 &E_CAST( 0,$S,$R,$L,$K,@F1,$tmp1,$tmp2,$tmp3,$tmp4);
135 &set_label('cast_enc_done') if $enc;
136 # Why the nop? - Ben 17/1/99
138 &mov($tmp3,&wparam(0));
139 &mov(&DWP(4,$tmp3,"",0),$L);
140 &mov(&DWP(0,$tmp3,"",0),$R);
141 &function_end($name);
145 local($i,$S,$L,$R,$K,$OP1,$OP2,$OP3,$tmp1,$tmp2,$tmp3,$tmp4)=@_;
146 # Ri needs to have 16 pre added.
148 &comment("round $i");
149 &mov( $tmp4, &DWP($i*8,$K,"",1));
151 &mov( $tmp1, &DWP($i*8+4,$K,"",1));
154 &rotl( $tmp4, &LB($tmp1));
160 &movb( &LB($tmp1), &HB($tmp4)); # A
166 &mov( $tmp2, $tmp4); # B
167 &movb( &LB($tmp1), &HB($tmp4)); # A # BAD BAD BAD
173 &movb( &LB($tmp3), &HB($tmp4)); # C # BAD BAD BAD
174 &and( $tmp4, 0xff); # D
176 &mov( $tmp1, &DWP($S1,"",$tmp1,4));
177 &mov( $tmp2, &DWP($S2,"",$tmp2,4));
179 &$OP2( $tmp1, $tmp2);
180 &mov( $tmp2, &DWP($S3,"",$tmp3,4));
182 &$OP3( $tmp1, $tmp2);
183 &mov( $tmp2, &DWP($S4,"",$tmp4,4));
185 &$OP1( $tmp1, $tmp2);