aboutsummaryrefslogtreecommitdiffstats
path: root/crypto/sha/asm/sha1-mips.pl
blob: 197bc6b50e95f53a3e12114d5a81dde6e9055509 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
#!/usr/bin/env perl

# ====================================================================
# 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/.
# ====================================================================

# SHA1 block procedure for MIPS.

# Performance improvement is 30% on unaligned input. The "secret" is
# to deploy lwl/lwr pair to load unaligned input. One could have
# vectorized Xupdate on MIPSIII/IV, but the goal was to code MIPS32-
# compatible subroutine. There is room for minor optimization on
# little-endian platforms...

######################################################################
# There is a number of MIPS ABI in use, O32 and N32/64 are most
# widely used. Then there is a new contender: NUBI. It appears that if
# one picks the latter, it's possible to arrange code in ABI neutral
# manner. Therefore let's stick to NUBI register layout:
#
($zero,$at,$t0,$t1,$t2)=map("\$$_",(0..2,24,25));
($a0,$a1,$a2,$a3,$a4,$a5,$a6,$a7)=map("\$$_",(4..11));
($s0,$s1,$s2,$s3,$s4,$s5,$s6,$s7,$s8,$s9,$s10,$s11)=map("\$$_",(12..23));
($gp,$tp,$sp,$fp,$ra)=map("\$$_",(3,28..31));
#
# The return value is placed in $a0. Following coding rules facilitate
# interoperability:
#
# - never ever touch $tp, "thread pointer", former $gp;
# - copy return value to $t0, former $v0 [or to $a0 if you're adapting
#   old code];
# - on O32 populate $a4-$a7 with 'lw $aN,4*N($sp)' if necessary;
#
# For reference here is register layout for N32/64 MIPS ABIs:
#
# ($zero,$at,$v0,$v1)=map("\$$_",(0..3));
# ($a0,$a1,$a2,$a3,$a4,$a5,$a6,$a7)=map("\$$_",(4..11));
# ($t0,$t1,$t2,$t3,$t8,$t9)=map("\$$_",(12..15,24,25));
# ($s0,$s1,$s2,$s3,$s4,$s5,$s6,$s7)=map("\$$_",(16..23));
# ($gp,$sp,$fp,$ra)=map("\$$_",(28..31));
#
$flavour = shift; # supported flavours are o32,n32,64,nubi32,nubi64

if ($flavour =~ /64|n32/i) {
	$PTR_ADD="dadd";	# incidentally works even on n32
	$PTR_SUB="dsub";	# incidentally works even on n32
	$REG_S="sd";
	$REG_L="ld";
	$PTR_SLL="dsll";	# incidentally works even on n32
	$SZREG=8;
} else {
	$PTR_ADD="add";
	$PTR_SUB="sub";
	$REG_S="sw";
	$REG_L="lw";
	$PTR_SLL="sll";
	$SZREG=4;
}
#
# <appro@openssl.org>
#
######################################################################

$big_endian=(`echo MIPSEL | $ENV{CC} -E -`=~/MIPSEL/)?1:0 if ($ENV{CC});

for (@ARGV) {	$output=$_ if (/^\w[\w\-]*\.\w+$/);   }
open STDOUT,">$output";

if (!defined($big_endian))
            {   $big_endian=(unpack('L',pack('N',1))==1);   }

# offsets of the Most and Least Significant Bytes
$MSB=$big_endian?0:3;
$LSB=3&~$MSB;

@X=map("\$$_",(8..23));	# a4-a7,s0-s11

$ctx=$a0;
$inp=$a1;
$num=$a2;
$A="\$1";
$B="\$2";
$C="\$3";
$D="\$7";
$E="\$24";	@V=($A,$B,$C,$D,$E);
$t0="\$25";
$t1=$num;	# $num is offloaded to stack
$t2="\$30";	# fp
$K="\$31";	# ra

sub BODY_00_14 {
my ($i,$a,$b,$c,$d,$e)=@_;
my $j=$i+1;
$code.=<<___	if (!$big_endian);
	srl	$t0,@X[$i],24	# byte swap($i)
	srl	$t1,@X[$i],8
	andi	$t2,@X[$i],0xFF00
	sll	@X[$i],@X[$i],24
	andi	$t1,0xFF00
	sll	$t2,$t2,8
	or	@X[$i],$t0
	or	$t1,$t2
	or	@X[$i],$t1
___
$code.=<<___;
	 lwl	@X[$j],$j*4+$MSB($inp)
	sll	$t0,$a,5	# $i
	addu	$e,$K
	 lwr	@X[$j],$j*4+$LSB($inp)
	srl	$t1,$a,27
	addu	$e,$t0
	xor	$t0,$c,$d
	addu	$e,$t1
	sll	$t2,$b,30
	and	$t0,$b
	srl	$b,$b,2
	xor	$t0,$d
	addu	$e,@X[$i]
	or	$b,$t2
	addu	$e,$t0
___
}

sub BODY_15_19 {
my ($i,$a,$b,$c,$d,$e)=@_;
my $j=$i+1;

$code.=<<___	if (!$big_endian && $i==15);
	srl	$t0,@X[$i],24	# byte swap($i)
	srl	$t1,@X[$i],8
	andi	$t2,@X[$i],0xFF00
	sll	@X[$i],@X[$i],24
	andi	$t1,0xFF00
	sll	$t2,$t2,8
	or	@X[$i],$t0
	or	@X[$i],$t1
	or	@X[$i],$t2
___
$code.=<<___;
	 xor	@X[$j%16],@X[($j+2)%16]
	sll	$t0,$a,5	# $i
	addu	$e,$K
	srl	$t1,$a,27
	addu	$e,$t0
	 xor	@X[$j%16],@X[($j+8)%16]
	xor	$t0,$c,$d
	addu	$e,$t1
	 xor	@X[$j%16],@X[($j+13)%16]
	sll	$t2,$b,30
	and	$t0,$b
	 srl	$t1,@X[$j%16],31
	 addu	@X[$j%16],@X[$j%16]
	srl	$b,$b,2
	xor	$t0,$d
	 or	@X[$j%16],$t1
	addu	$e,@X[$i%16]
	or	$b,$t2
	addu	$e,$t0
___
}

sub BODY_20_39 {
my ($i,$a,$b,$c,$d,$e)=@_;
my $j=$i+1;
$code.=<<___ if ($i<79);
	 xor	@X[$j%16],@X[($j+2)%16]
	sll	$t0,$a,5	# $i
	addu	$e,$K
	srl	$t1,$a,27
	addu	$e,$t0
	 xor	@X[$j%16],@X[($j+8)%16]
	xor	$t0,$c,$d
	addu	$e,$t1
	 xor	@X[$j%16],@X[($j+13)%16]
	sll	$t2,$b,30
	xor	$t0,$b
	 srl	$t1,@X[$j%16],31
	 addu	@X[$j%16],@X[$j%16]
	srl	$b,$b,2
	addu	$e,@X[$i%16]
	 or	@X[$j%16],$t1
	or	$b,$t2
	addu	$e,$t0
___
$code.=<<___ if ($i==79);
	 lw	@X[0],0($ctx)
	sll	$t0,$a,5	# $i
	addu	$e,$K
	 lw	@X[1],4($ctx)
	srl	$t1,$a,27
	addu	$e,$t0
	 lw	@X[2],8($ctx)
	xor	$t0,$c,$d
	addu	$e,$t1
	 lw	@X[3],12($ctx)
	sll	$t2,$b,30
	xor	$t0,$b
	 lw	@X[4],16($ctx)
	srl	$b,$b,2
	addu	$e,@X[$i%16]
	or	$b,$t2
	addu	$e,$t0
___
}

sub BODY_40_59 {
my ($i,$a,$b,$c,$d,$e)=@_;
my $j=$i+1;
$code.=<<___ if ($i<79);
	 xor	@X[$j%16],@X[($j+2)%16]
	sll	$t0,$a,5	# $i
	addu	$e,$K
	srl	$t1,$a,27
	addu	$e,$t0
	 xor	@X[$j%16],@X[($j+8)%16]
	and	$t0,$c,$d
	addu	$e,$t1
	 xor	@X[$j%16],@X[($j+13)%16]
	sll	$t2,$b,30
	addu	$e,$t0
	 srl	$t1,@X[$j%16],31
	xor	$t0,$c,$d
	 addu	@X[$j%16],@X[$j%16]
	and	$t0,$b
	srl	$b,$b,2
	 or	@X[$j%16],$t1
	addu	$e,@X[$i%16]
	or	$b,$t2
	addu	$e,$t0
___
}

$FRAMESIZE=16;	# large enough to accomodate NUBI saved registers
$SAVED_REGS_MASK = ($flavour =~ /nubi/i) ? 0xc0fff008 : 0xc0ff0000;

$code=<<___;
#ifdef OPENSSL_FIPSCANISTER
# include <openssl/fipssyms.h>
#endif

.text

.set	noat
.set	noreorder
.align	5
.globl	sha1_block_data_order
.ent	sha1_block_data_order
sha1_block_data_order:
	.frame	$sp,$FRAMESIZE*$SZREG,$ra
	.mask	$SAVED_REGS_MASK,-$SZREG
	.set	noreorder
	$PTR_SUB $sp,$FRAMESIZE*$SZREG
	$REG_S	$ra,($FRAMESIZE-1)*$SZREG($sp)
	$REG_S	$fp,($FRAMESIZE-2)*$SZREG($sp)
	$REG_S	$s11,($FRAMESIZE-3)*$SZREG($sp)
	$REG_S	$s10,($FRAMESIZE-4)*$SZREG($sp)
	$REG_S	$s9,($FRAMESIZE-5)*$SZREG($sp)
	$REG_S	$s8,($FRAMESIZE-6)*$SZREG($sp)
	$REG_S	$s7,($FRAMESIZE-7)*$SZREG($sp)
	$REG_S	$s6,($FRAMESIZE-8)*$SZREG($sp)
	$REG_S	$s5,($FRAMESIZE-9)*$SZREG($sp)
	$REG_S	$s4,($FRAMESIZE-10)*$SZREG($sp)
___
$code.=<<___ if ($flavour =~ /nubi/i);	# optimize non-nubi prologue
	$REG_S	$s3,($FRAMESIZE-11)*$SZREG($sp)
	$REG_S	$s2,($FRAMESIZE-12)*$SZREG($sp)
	$REG_S	$s1,($FRAMESIZE-13)*$SZREG($sp)
	$REG_S	$s0,($FRAMESIZE-14)*$SZREG($sp)
	$REG_S	$gp,($FRAMESIZE-15)*$SZREG($sp)
___
$code.=<<___;
	$PTR_SLL $num,6
	$PTR_ADD $num,$inp
	$REG_S	$num,0($sp)
	lw	$A,0($ctx)
	lw	$B,4($ctx)
	lw	$C,8($ctx)
	lw	$D,12($ctx)
	b	.Loop
	lw	$E,16($ctx)
.align	4
.Loop:
	.set	reorder
	lwl	@X[0],$MSB($inp)
	lui	$K,0x5a82
	lwr	@X[0],$LSB($inp)
	ori	$K,0x7999	# K_00_19
___
for ($i=0;$i<15;$i++)	{ &BODY_00_14($i,@V); unshift(@V,pop(@V)); }
for (;$i<20;$i++)	{ &BODY_15_19($i,@V); unshift(@V,pop(@V)); }
$code.=<<___;
	lui	$K,0x6ed9
	ori	$K,0xeba1	# K_20_39
___
for (;$i<40;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
$code.=<<___;
	lui	$K,0x8f1b
	ori	$K,0xbcdc	# K_40_59
___
for (;$i<60;$i++)	{ &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
$code.=<<___;
	lui	$K,0xca62
	ori	$K,0xc1d6	# K_60_79
___
for (;$i<80;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
$code.=<<___;
	$PTR_ADD $inp,64
	$REG_L	$num,0($sp)

	addu	$A,$X[0]
	addu	$B,$X[1]
	sw	$A,0($ctx)
	addu	$C,$X[2]
	addu	$D,$X[3]
	sw	$B,4($ctx)
	addu	$E,$X[4]
	sw	$C,8($ctx)
	sw	$D,12($ctx)
	sw	$E,16($ctx)
	.set	noreorder
	bne	$inp,$num,.Loop
	nop

	.set	noreorder
	$REG_L	$ra,($FRAMESIZE-1)*$SZREG($sp)
	$REG_L	$fp,($FRAMESIZE-2)*$SZREG($sp)
	$REG_L	$s11,($FRAMESIZE-3)*$SZREG($sp)
	$REG_L	$s10,($FRAMESIZE-4)*$SZREG($sp)
	$REG_L	$s9,($FRAMESIZE-5)*$SZREG($sp)
	$REG_L	$s8,($FRAMESIZE-6)*$SZREG($sp)
	$REG_L	$s7,($FRAMESIZE-7)*$SZREG($sp)
	$REG_L	$s6,($FRAMESIZE-8)*$SZREG($sp)
	$REG_L	$s5,($FRAMESIZE-9)*$SZREG($sp)
	$REG_L	$s4,($FRAMESIZE-10)*$SZREG($sp)
___
$code.=<<___ if ($flavour =~ /nubi/i);
	$REG_L	$s3,($FRAMESIZE-11)*$SZREG($sp)
	$REG_L	$s2,($FRAMESIZE-12)*$SZREG($sp)
	$REG_L	$s1,($FRAMESIZE-13)*$SZREG($sp)
	$REG_L	$s0,($FRAMESIZE-14)*$SZREG($sp)
	$REG_L	$gp,($FRAMESIZE-15)*$SZREG($sp)
___
$code.=<<___;
	jr	$ra
	$PTR_ADD $sp,$FRAMESIZE*$SZREG
.end	sha1_block_data_order
.rdata
.asciiz	"SHA1 for MIPS, CRYPTOGAMS by <appro\@openssl.org>"
___
print $code;
close STDOUT;