aboutsummaryrefslogtreecommitdiff
path: root/openssl/crypto/sha/asm/sha1-parisc.pl
blob: 6e5a328a6f1f377387863fabdb09e382eee5f9fe (plain)
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
#!/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 PA-RISC.

# June 2009.
#
# On PA-7100LC performance is >30% better than gcc 3.2 generated code
# for aligned input and >50% better for unaligned. Compared to vendor
# compiler on PA-8600 it's almost 60% faster in 64-bit build and just
# few percent faster in 32-bit one (this for aligned input, data for
# unaligned input is not available).
#
# Special thanks to polarhome.com for providing HP-UX account.

$flavour = shift;
$output = shift;
open STDOUT,">$output";

if ($flavour =~ /64/) {
	$LEVEL		="2.0W";
	$SIZE_T		=8;
	$FRAME_MARKER	=80;
	$SAVED_RP	=16;
	$PUSH		="std";
	$PUSHMA		="std,ma";
	$POP		="ldd";
	$POPMB		="ldd,mb";
} else {
	$LEVEL		="1.0";
	$SIZE_T		=4;
	$FRAME_MARKER	=48;
	$SAVED_RP	=20;
	$PUSH		="stw";
	$PUSHMA		="stwm";
	$POP		="ldw";
	$POPMB		="ldwm";
}

$FRAME=14*$SIZE_T+$FRAME_MARKER;# 14 saved regs + frame marker
				#                 [+ argument transfer]
$ctx="%r26";		# arg0
$inp="%r25";		# arg1
$num="%r24";		# arg2

$t0="%r28";
$t1="%r29";
$K="%r31";

@X=("%r1", "%r2", "%r3", "%r4", "%r5", "%r6", "%r7", "%r8",
    "%r9", "%r10","%r11","%r12","%r13","%r14","%r15","%r16",$t0);

@V=($A,$B,$C,$D,$E)=("%r19","%r20","%r21","%r22","%r23");

sub BODY_00_19 {
my ($i,$a,$b,$c,$d,$e)=@_;
my $j=$i+1;
$code.=<<___ if ($i<15);
	addl	$K,$e,$e	; $i
	shd	$a,$a,27,$t1
	addl	@X[$i],$e,$e
	and	$c,$b,$t0
	addl	$t1,$e,$e
	andcm	$d,$b,$t1
	shd	$b,$b,2,$b
	or	$t1,$t0,$t0
	addl	$t0,$e,$e
___
$code.=<<___ if ($i>=15);	# with forward Xupdate
	addl	$K,$e,$e	; $i
	shd	$a,$a,27,$t1
	xor	@X[($j+2)%16],@X[$j%16],@X[$j%16]
	addl	@X[$i%16],$e,$e
	and	$c,$b,$t0
	xor	@X[($j+8)%16],@X[$j%16],@X[$j%16]
	addl	$t1,$e,$e
	andcm	$d,$b,$t1
	shd	$b,$b,2,$b
	or	$t1,$t0,$t0
	xor	@X[($j+13)%16],@X[$j%16],@X[$j%16]
	add	$t0,$e,$e
	shd	@X[$j%16],@X[$j%16],31,@X[$j%16]
___
}

sub BODY_20_39 {
my ($i,$a,$b,$c,$d,$e)=@_;
my $j=$i+1;
$code.=<<___ if ($i<79);
	xor	@X[($j+2)%16],@X[$j%16],@X[$j%16]	; $i
	addl	$K,$e,$e
	shd	$a,$a,27,$t1
	xor	@X[($j+8)%16],@X[$j%16],@X[$j%16]
	addl	@X[$i%16],$e,$e
	xor	$b,$c,$t0
	xor	@X[($j+13)%16],@X[$j%16],@X[$j%16]
	addl	$t1,$e,$e
	shd	$b,$b,2,$b
	xor	$d,$t0,$t0
	shd	@X[$j%16],@X[$j%16],31,@X[$j%16]
	addl	$t0,$e,$e
___
$code.=<<___ if ($i==79);	# with context load
	ldw	0($ctx),@X[0]	; $i
	addl	$K,$e,$e
	shd	$a,$a,27,$t1
	ldw	4($ctx),@X[1]
	addl	@X[$i%16],$e,$e
	xor	$b,$c,$t0
	ldw	8($ctx),@X[2]
	addl	$t1,$e,$e
	shd	$b,$b,2,$b
	xor	$d,$t0,$t0
	ldw	12($ctx),@X[3]
	addl	$t0,$e,$e
	ldw	16($ctx),@X[4]
___
}

sub BODY_40_59 {
my ($i,$a,$b,$c,$d,$e)=@_;
my $j=$i+1;
$code.=<<___;
	shd	$a,$a,27,$t1	; $i
	addl	$K,$e,$e
	xor	@X[($j+2)%16],@X[$j%16],@X[$j%16]
	xor	$d,$c,$t0
	addl	@X[$i%16],$e,$e
	xor	@X[($j+8)%16],@X[$j%16],@X[$j%16]
	and	$b,$t0,$t0
	addl	$t1,$e,$e
	shd	$b,$b,2,$b
	xor	@X[($j+13)%16],@X[$j%16],@X[$j%16]
	addl	$t0,$e,$e
	and	$d,$c,$t1
	shd	@X[$j%16],@X[$j%16],31,@X[$j%16]
	addl	$t1,$e,$e
___
}

$code=<<___;
	.LEVEL	$LEVEL
	.SPACE	\$TEXT\$
	.SUBSPA	\$CODE\$,QUAD=0,ALIGN=8,ACCESS=0x2C,CODE_ONLY

	.EXPORT	sha1_block_data_order,ENTRY,ARGW0=GR,ARGW1=GR,ARGW2=GR
sha1_block_data_order
	.PROC
	.CALLINFO	FRAME=`$FRAME-14*$SIZE_T`,NO_CALLS,SAVE_RP,ENTRY_GR=16
	.ENTRY
	$PUSH	%r2,-$SAVED_RP(%sp)	; standard prologue
	$PUSHMA	%r3,$FRAME(%sp)
	$PUSH	%r4,`-$FRAME+1*$SIZE_T`(%sp)
	$PUSH	%r5,`-$FRAME+2*$SIZE_T`(%sp)
	$PUSH	%r6,`-$FRAME+3*$SIZE_T`(%sp)
	$PUSH	%r7,`-$FRAME+4*$SIZE_T`(%sp)
	$PUSH	%r8,`-$FRAME+5*$SIZE_T`(%sp)
	$PUSH	%r9,`-$FRAME+6*$SIZE_T`(%sp)
	$PUSH	%r10,`-$FRAME+7*$SIZE_T`(%sp)
	$PUSH	%r11,`-$FRAME+8*$SIZE_T`(%sp)
	$PUSH	%r12,`-$FRAME+9*$SIZE_T`(%sp)
	$PUSH	%r13,`-$FRAME+10*$SIZE_T`(%sp)
	$PUSH	%r14,`-$FRAME+11*$SIZE_T`(%sp)
	$PUSH	%r15,`-$FRAME+12*$SIZE_T`(%sp)
	$PUSH	%r16,`-$FRAME+13*$SIZE_T`(%sp)

	ldw	0($ctx),$A
	ldw	4($ctx),$B
	ldw	8($ctx),$C
	ldw	12($ctx),$D
	ldw	16($ctx),$E

	extru	$inp,31,2,$t0		; t0=inp&3;
	sh3addl	$t0,%r0,$t0		; t0*=8;
	subi	32,$t0,$t0		; t0=32-t0;
	mtctl	$t0,%cr11		; %sar=t0;

L\$oop
	ldi	3,$t0
	andcm	$inp,$t0,$t0		; 64-bit neutral
___
	for ($i=0;$i<15;$i++) {		# load input block
	$code.="\tldw	`4*$i`($t0),@X[$i]\n";		}
$code.=<<___;
	cmpb,*=	$inp,$t0,L\$aligned
	ldw	60($t0),@X[15]
	ldw	64($t0),@X[16]
___
	for ($i=0;$i<16;$i++) {		# align input
	$code.="\tvshd	@X[$i],@X[$i+1],@X[$i]\n";	}
$code.=<<___;
L\$aligned
	ldil	L'0x5a827000,$K		; K_00_19
	ldo	0x999($K),$K
___
for ($i=0;$i<20;$i++)   { &BODY_00_19($i,@V); unshift(@V,pop(@V)); }
$code.=<<___;
	ldil	L'0x6ed9e000,$K		; K_20_39
	ldo	0xba1($K),$K
___

for (;$i<40;$i++)       { &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
$code.=<<___;
	ldil	L'0x8f1bb000,$K		; K_40_59
	ldo	0xcdc($K),$K
___

for (;$i<60;$i++)       { &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
$code.=<<___;
	ldil	L'0xca62c000,$K		; K_60_79
	ldo	0x1d6($K),$K
___
for (;$i<80;$i++)       { &BODY_20_39($i,@V); unshift(@V,pop(@V)); }

$code.=<<___;
	addl	@X[0],$A,$A
	addl	@X[1],$B,$B
	addl	@X[2],$C,$C
	addl	@X[3],$D,$D
	addl	@X[4],$E,$E
	stw	$A,0($ctx)
	stw	$B,4($ctx)
	stw	$C,8($ctx)
	stw	$D,12($ctx)
	stw	$E,16($ctx)
	addib,*<> -1,$num,L\$oop
	ldo	64($inp),$inp

	$POP	`-$FRAME-$SAVED_RP`(%sp),%r2	; standard epilogue
	$POP	`-$FRAME+1*$SIZE_T`(%sp),%r4
	$POP	`-$FRAME+2*$SIZE_T`(%sp),%r5
	$POP	`-$FRAME+3*$SIZE_T`(%sp),%r6
	$POP	`-$FRAME+4*$SIZE_T`(%sp),%r7
	$POP	`-$FRAME+5*$SIZE_T`(%sp),%r8
	$POP	`-$FRAME+6*$SIZE_T`(%sp),%r9
	$POP	`-$FRAME+7*$SIZE_T`(%sp),%r10
	$POP	`-$FRAME+8*$SIZE_T`(%sp),%r11
	$POP	`-$FRAME+9*$SIZE_T`(%sp),%r12
	$POP	`-$FRAME+10*$SIZE_T`(%sp),%r13
	$POP	`-$FRAME+11*$SIZE_T`(%sp),%r14
	$POP	`-$FRAME+12*$SIZE_T`(%sp),%r15
	$POP	`-$FRAME+13*$SIZE_T`(%sp),%r16
	bv	(%r2)
	.EXIT
	$POPMB	-$FRAME(%sp),%r3
	.PROCEND
	.STRINGZ "SHA1 block transform for PA-RISC, CRYPTOGAMS by <appro\@openssl.org>"
___

$code =~ s/\`([^\`]*)\`/eval $1/gem;
$code =~ s/,\*/,/gm		if ($SIZE_T==4);
$code =~ s/\bbv\b/bve/gm	if ($SIZE_T==8);
print $code;
close STDOUT;