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
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
|
#!/usr/bin/env perl
#
# ====================================================================
# Written by Andy Polyakov <appro@openssl.org> 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/.
# ====================================================================
#
# March 2010
#
# The module implements "4-bit" GCM GHASH function and underlying
# single multiplication operation in GF(2^128). "4-bit" means that it
# uses 256 bytes per-key table [+128 bytes shared table]. Even though
# loops are aggressively modulo-scheduled in respect to references to
# Htbl and Z.hi updates for 8 cycles per byte, measured performance is
# ~12 cycles per processed byte on 21264 CPU. It seems to be a dynamic
# scheduling "glitch," because uprofile(1) indicates uniform sample
# distribution, as if all instruction bundles execute in 1.5 cycles.
# Meaning that it could have been even faster, yet 12 cycles is ~60%
# better than gcc-generated code and ~80% than code generated by vendor
# compiler.
$cnt="v0"; # $0
$t0="t0";
$t1="t1";
$t2="t2";
$Thi0="t3"; # $4
$Tlo0="t4";
$Thi1="t5";
$Tlo1="t6";
$rem="t7"; # $8
#################
$Xi="a0"; # $16, input argument block
$Htbl="a1";
$inp="a2";
$len="a3";
$nlo="a4"; # $20
$nhi="a5";
$Zhi="t8";
$Zlo="t9";
$Xhi="t10"; # $24
$Xlo="t11";
$remp="t12";
$rem_4bit="AT"; # $28
{ my $N;
sub loop() {
$N++;
$code.=<<___;
.align 4
extbl $Xlo,7,$nlo
and $nlo,0xf0,$nhi
sll $nlo,4,$nlo
and $nlo,0xf0,$nlo
addq $nlo,$Htbl,$nlo
ldq $Zlo,8($nlo)
addq $nhi,$Htbl,$nhi
ldq $Zhi,0($nlo)
and $Zlo,0x0f,$remp
sll $Zhi,60,$t0
lda $cnt,6(zero)
extbl $Xlo,6,$nlo
ldq $Tlo1,8($nhi)
s8addq $remp,$rem_4bit,$remp
ldq $Thi1,0($nhi)
srl $Zlo,4,$Zlo
ldq $rem,0($remp)
srl $Zhi,4,$Zhi
xor $t0,$Zlo,$Zlo
and $nlo,0xf0,$nhi
xor $Tlo1,$Zlo,$Zlo
sll $nlo,4,$nlo
xor $Thi1,$Zhi,$Zhi
and $nlo,0xf0,$nlo
addq $nlo,$Htbl,$nlo
ldq $Tlo0,8($nlo)
addq $nhi,$Htbl,$nhi
ldq $Thi0,0($nlo)
.Looplo$N:
and $Zlo,0x0f,$remp
sll $Zhi,60,$t0
subq $cnt,1,$cnt
srl $Zlo,4,$Zlo
ldq $Tlo1,8($nhi)
xor $rem,$Zhi,$Zhi
ldq $Thi1,0($nhi)
s8addq $remp,$rem_4bit,$remp
ldq $rem,0($remp)
srl $Zhi,4,$Zhi
xor $t0,$Zlo,$Zlo
extbl $Xlo,$cnt,$nlo
and $nlo,0xf0,$nhi
xor $Thi0,$Zhi,$Zhi
xor $Tlo0,$Zlo,$Zlo
sll $nlo,4,$nlo
and $Zlo,0x0f,$remp
sll $Zhi,60,$t0
and $nlo,0xf0,$nlo
srl $Zlo,4,$Zlo
s8addq $remp,$rem_4bit,$remp
xor $rem,$Zhi,$Zhi
addq $nlo,$Htbl,$nlo
addq $nhi,$Htbl,$nhi
ldq $rem,0($remp)
srl $Zhi,4,$Zhi
ldq $Tlo0,8($nlo)
xor $t0,$Zlo,$Zlo
xor $Tlo1,$Zlo,$Zlo
xor $Thi1,$Zhi,$Zhi
ldq $Thi0,0($nlo)
bne $cnt,.Looplo$N
and $Zlo,0x0f,$remp
sll $Zhi,60,$t0
lda $cnt,7(zero)
srl $Zlo,4,$Zlo
ldq $Tlo1,8($nhi)
xor $rem,$Zhi,$Zhi
ldq $Thi1,0($nhi)
s8addq $remp,$rem_4bit,$remp
ldq $rem,0($remp)
srl $Zhi,4,$Zhi
xor $t0,$Zlo,$Zlo
extbl $Xhi,$cnt,$nlo
and $nlo,0xf0,$nhi
xor $Thi0,$Zhi,$Zhi
xor $Tlo0,$Zlo,$Zlo
sll $nlo,4,$nlo
and $Zlo,0x0f,$remp
sll $Zhi,60,$t0
and $nlo,0xf0,$nlo
srl $Zlo,4,$Zlo
s8addq $remp,$rem_4bit,$remp
xor $rem,$Zhi,$Zhi
addq $nlo,$Htbl,$nlo
addq $nhi,$Htbl,$nhi
ldq $rem,0($remp)
srl $Zhi,4,$Zhi
ldq $Tlo0,8($nlo)
xor $t0,$Zlo,$Zlo
xor $Tlo1,$Zlo,$Zlo
xor $Thi1,$Zhi,$Zhi
ldq $Thi0,0($nlo)
unop
.Loophi$N:
and $Zlo,0x0f,$remp
sll $Zhi,60,$t0
subq $cnt,1,$cnt
srl $Zlo,4,$Zlo
ldq $Tlo1,8($nhi)
xor $rem,$Zhi,$Zhi
ldq $Thi1,0($nhi)
s8addq $remp,$rem_4bit,$remp
ldq $rem,0($remp)
srl $Zhi,4,$Zhi
xor $t0,$Zlo,$Zlo
extbl $Xhi,$cnt,$nlo
and $nlo,0xf0,$nhi
xor $Thi0,$Zhi,$Zhi
xor $Tlo0,$Zlo,$Zlo
sll $nlo,4,$nlo
and $Zlo,0x0f,$remp
sll $Zhi,60,$t0
and $nlo,0xf0,$nlo
srl $Zlo,4,$Zlo
s8addq $remp,$rem_4bit,$remp
xor $rem,$Zhi,$Zhi
addq $nlo,$Htbl,$nlo
addq $nhi,$Htbl,$nhi
ldq $rem,0($remp)
srl $Zhi,4,$Zhi
ldq $Tlo0,8($nlo)
xor $t0,$Zlo,$Zlo
xor $Tlo1,$Zlo,$Zlo
xor $Thi1,$Zhi,$Zhi
ldq $Thi0,0($nlo)
bne $cnt,.Loophi$N
and $Zlo,0x0f,$remp
sll $Zhi,60,$t0
srl $Zlo,4,$Zlo
ldq $Tlo1,8($nhi)
xor $rem,$Zhi,$Zhi
ldq $Thi1,0($nhi)
s8addq $remp,$rem_4bit,$remp
ldq $rem,0($remp)
srl $Zhi,4,$Zhi
xor $t0,$Zlo,$Zlo
xor $Tlo0,$Zlo,$Zlo
xor $Thi0,$Zhi,$Zhi
and $Zlo,0x0f,$remp
sll $Zhi,60,$t0
srl $Zlo,4,$Zlo
s8addq $remp,$rem_4bit,$remp
xor $rem,$Zhi,$Zhi
ldq $rem,0($remp)
srl $Zhi,4,$Zhi
xor $Tlo1,$Zlo,$Zlo
xor $Thi1,$Zhi,$Zhi
xor $t0,$Zlo,$Zlo
xor $rem,$Zhi,$Zhi
___
}}
$code=<<___;
#ifdef __linux__
#include <asm/regdef.h>
#else
#include <asm.h>
#include <regdef.h>
#endif
.text
.set noat
.set noreorder
.globl gcm_gmult_4bit
.align 4
.ent gcm_gmult_4bit
gcm_gmult_4bit:
.frame sp,0,ra
.prologue 0
ldq $Xlo,8($Xi)
ldq $Xhi,0($Xi)
br $rem_4bit,.Lpic1
.Lpic1: lda $rem_4bit,rem_4bit-.Lpic1($rem_4bit)
___
&loop();
$code.=<<___;
srl $Zlo,24,$t0 # byte swap
srl $Zlo,8,$t1
sll $Zlo,8,$t2
sll $Zlo,24,$Zlo
zapnot $t0,0x11,$t0
zapnot $t1,0x22,$t1
zapnot $Zlo,0x88,$Zlo
or $t0,$t1,$t0
zapnot $t2,0x44,$t2
or $Zlo,$t0,$Zlo
srl $Zhi,24,$t0
srl $Zhi,8,$t1
or $Zlo,$t2,$Zlo
sll $Zhi,8,$t2
sll $Zhi,24,$Zhi
srl $Zlo,32,$Xlo
sll $Zlo,32,$Zlo
zapnot $t0,0x11,$t0
zapnot $t1,0x22,$t1
or $Zlo,$Xlo,$Xlo
zapnot $Zhi,0x88,$Zhi
or $t0,$t1,$t0
zapnot $t2,0x44,$t2
or $Zhi,$t0,$Zhi
or $Zhi,$t2,$Zhi
srl $Zhi,32,$Xhi
sll $Zhi,32,$Zhi
or $Zhi,$Xhi,$Xhi
stq $Xlo,8($Xi)
stq $Xhi,0($Xi)
ret (ra)
.end gcm_gmult_4bit
___
$inhi="s0";
$inlo="s1";
$code.=<<___;
.globl gcm_ghash_4bit
.align 4
.ent gcm_ghash_4bit
gcm_ghash_4bit:
lda sp,-32(sp)
stq ra,0(sp)
stq s0,8(sp)
stq s1,16(sp)
.mask 0x04000600,-32
.frame sp,32,ra
.prologue 0
ldq_u $inhi,0($inp)
ldq_u $Thi0,7($inp)
ldq_u $inlo,8($inp)
ldq_u $Tlo0,15($inp)
ldq $Xhi,0($Xi)
ldq $Xlo,8($Xi)
br $rem_4bit,.Lpic2
.Lpic2: lda $rem_4bit,rem_4bit-.Lpic2($rem_4bit)
.Louter:
extql $inhi,$inp,$inhi
extqh $Thi0,$inp,$Thi0
or $inhi,$Thi0,$inhi
lda $inp,16($inp)
extql $inlo,$inp,$inlo
extqh $Tlo0,$inp,$Tlo0
or $inlo,$Tlo0,$inlo
subq $len,16,$len
xor $Xlo,$inlo,$Xlo
xor $Xhi,$inhi,$Xhi
___
&loop();
$code.=<<___;
srl $Zlo,24,$t0 # byte swap
srl $Zlo,8,$t1
sll $Zlo,8,$t2
sll $Zlo,24,$Zlo
zapnot $t0,0x11,$t0
zapnot $t1,0x22,$t1
zapnot $Zlo,0x88,$Zlo
or $t0,$t1,$t0
zapnot $t2,0x44,$t2
or $Zlo,$t0,$Zlo
srl $Zhi,24,$t0
srl $Zhi,8,$t1
or $Zlo,$t2,$Zlo
sll $Zhi,8,$t2
sll $Zhi,24,$Zhi
srl $Zlo,32,$Xlo
sll $Zlo,32,$Zlo
beq $len,.Ldone
zapnot $t0,0x11,$t0
zapnot $t1,0x22,$t1
or $Zlo,$Xlo,$Xlo
ldq_u $inhi,0($inp)
zapnot $Zhi,0x88,$Zhi
or $t0,$t1,$t0
zapnot $t2,0x44,$t2
ldq_u $Thi0,7($inp)
or $Zhi,$t0,$Zhi
or $Zhi,$t2,$Zhi
ldq_u $inlo,8($inp)
ldq_u $Tlo0,15($inp)
srl $Zhi,32,$Xhi
sll $Zhi,32,$Zhi
or $Zhi,$Xhi,$Xhi
br zero,.Louter
.Ldone:
zapnot $t0,0x11,$t0
zapnot $t1,0x22,$t1
or $Zlo,$Xlo,$Xlo
zapnot $Zhi,0x88,$Zhi
or $t0,$t1,$t0
zapnot $t2,0x44,$t2
or $Zhi,$t0,$Zhi
or $Zhi,$t2,$Zhi
srl $Zhi,32,$Xhi
sll $Zhi,32,$Zhi
or $Zhi,$Xhi,$Xhi
stq $Xlo,8($Xi)
stq $Xhi,0($Xi)
.set noreorder
/*ldq ra,0(sp)*/
ldq s0,8(sp)
ldq s1,16(sp)
lda sp,32(sp)
ret (ra)
.end gcm_ghash_4bit
.align 4
rem_4bit:
.quad 0x0000<<48, 0x1C20<<48, 0x3840<<48, 0x2460<<48
.quad 0x7080<<48, 0x6CA0<<48, 0x48C0<<48, 0x54E0<<48
.quad 0xE100<<48, 0xFD20<<48, 0xD940<<48, 0xC560<<48
.quad 0x9180<<48, 0x8DA0<<48, 0xA9C0<<48, 0xB5E0<<48
.ascii "GHASH for Alpha, CRYPTOGAMS by <appro\@openssl.org>"
.align 4
___
$output=shift and open STDOUT,">$output";
print $code;
close STDOUT;
|
