summaryrefslogtreecommitdiff
path: root/src/lib/libcrypto/bn/asm
diff options
context:
space:
mode:
Diffstat (limited to 'src/lib/libcrypto/bn/asm')
-rw-r--r--src/lib/libcrypto/bn/asm/bn-586.pl593
-rw-r--r--src/lib/libcrypto/bn/asm/co-586.pl286
-rw-r--r--src/lib/libcrypto/bn/asm/ia64.S1605
-rw-r--r--src/lib/libcrypto/bn/asm/pa-risc2.s1618
-rw-r--r--src/lib/libcrypto/bn/asm/pa-risc2W.s1605
-rw-r--r--src/lib/libcrypto/bn/asm/sparcv8.S1458
-rw-r--r--src/lib/libcrypto/bn/asm/sparcv8plus.S1535
-rw-r--r--src/lib/libcrypto/bn/asm/x86.pl28
-rw-r--r--src/lib/libcrypto/bn/asm/x86/add.pl76
-rw-r--r--src/lib/libcrypto/bn/asm/x86/comba.pl277
-rw-r--r--src/lib/libcrypto/bn/asm/x86/div.pl15
-rw-r--r--src/lib/libcrypto/bn/asm/x86/mul.pl77
-rw-r--r--src/lib/libcrypto/bn/asm/x86/mul_add.pl87
-rw-r--r--src/lib/libcrypto/bn/asm/x86/sqr.pl60
-rw-r--r--src/lib/libcrypto/bn/asm/x86/sub.pl76
-rw-r--r--src/lib/libcrypto/bn/asm/x86_64-gcc.c575
16 files changed, 0 insertions, 9971 deletions
diff --git a/src/lib/libcrypto/bn/asm/bn-586.pl b/src/lib/libcrypto/bn/asm/bn-586.pl
deleted file mode 100644
index 9a78f63be1..0000000000
--- a/src/lib/libcrypto/bn/asm/bn-586.pl
+++ /dev/null
@@ -1,593 +0,0 @@
1#!/usr/local/bin/perl
2
3push(@INC,"perlasm","../../perlasm");
4require "x86asm.pl";
5
6&asm_init($ARGV[0],$0);
7
8&bn_mul_add_words("bn_mul_add_words");
9&bn_mul_words("bn_mul_words");
10&bn_sqr_words("bn_sqr_words");
11&bn_div_words("bn_div_words");
12&bn_add_words("bn_add_words");
13&bn_sub_words("bn_sub_words");
14&bn_sub_part_words("bn_sub_part_words") unless $main'openbsd;
15
16&asm_finish();
17
18sub bn_mul_add_words
19 {
20 local($name)=@_;
21
22 &function_begin($name,"");
23
24 &comment("");
25 $Low="eax";
26 $High="edx";
27 $a="ebx";
28 $w="ebp";
29 $r="edi";
30 $c="esi";
31
32 &xor($c,$c); # clear carry
33 &mov($r,&wparam(0)); #
34
35 &mov("ecx",&wparam(2)); #
36 &mov($a,&wparam(1)); #
37
38 &and("ecx",0xfffffff8); # num / 8
39 &mov($w,&wparam(3)); #
40
41 &push("ecx"); # Up the stack for a tmp variable
42
43 &jz(&label("maw_finish"));
44
45 &set_label("maw_loop",0);
46
47 &mov(&swtmp(0),"ecx"); #
48
49 for ($i=0; $i<32; $i+=4)
50 {
51 &comment("Round $i");
52
53 &mov("eax",&DWP($i,$a,"",0)); # *a
54 &mul($w); # *a * w
55 &add("eax",$c); # L(t)+= *r
56 &mov($c,&DWP($i,$r,"",0)); # L(t)+= *r
57 &adc("edx",0); # H(t)+=carry
58 &add("eax",$c); # L(t)+=c
59 &adc("edx",0); # H(t)+=carry
60 &mov(&DWP($i,$r,"",0),"eax"); # *r= L(t);
61 &mov($c,"edx"); # c= H(t);
62 }
63
64 &comment("");
65 &mov("ecx",&swtmp(0)); #
66 &add($a,32);
67 &add($r,32);
68 &sub("ecx",8);
69 &jnz(&label("maw_loop"));
70
71 &set_label("maw_finish",0);
72 &mov("ecx",&wparam(2)); # get num
73 &and("ecx",7);
74 &jnz(&label("maw_finish2")); # helps branch prediction
75 &jmp(&label("maw_end"));
76
77 &set_label("maw_finish2",1);
78 for ($i=0; $i<7; $i++)
79 {
80 &comment("Tail Round $i");
81 &mov("eax",&DWP($i*4,$a,"",0));# *a
82 &mul($w); # *a * w
83 &add("eax",$c); # L(t)+=c
84 &mov($c,&DWP($i*4,$r,"",0)); # L(t)+= *r
85 &adc("edx",0); # H(t)+=carry
86 &add("eax",$c);
87 &adc("edx",0); # H(t)+=carry
88 &dec("ecx") if ($i != 7-1);
89 &mov(&DWP($i*4,$r,"",0),"eax"); # *r= L(t);
90 &mov($c,"edx"); # c= H(t);
91 &jz(&label("maw_end")) if ($i != 7-1);
92 }
93 &set_label("maw_end",0);
94 &mov("eax",$c);
95
96 &pop("ecx"); # clear variable from
97
98 &function_end($name);
99 }
100
101sub bn_mul_words
102 {
103 local($name)=@_;
104
105 &function_begin($name,"");
106
107 &comment("");
108 $Low="eax";
109 $High="edx";
110 $a="ebx";
111 $w="ecx";
112 $r="edi";
113 $c="esi";
114 $num="ebp";
115
116 &xor($c,$c); # clear carry
117 &mov($r,&wparam(0)); #
118 &mov($a,&wparam(1)); #
119 &mov($num,&wparam(2)); #
120 &mov($w,&wparam(3)); #
121
122 &and($num,0xfffffff8); # num / 8
123 &jz(&label("mw_finish"));
124
125 &set_label("mw_loop",0);
126 for ($i=0; $i<32; $i+=4)
127 {
128 &comment("Round $i");
129
130 &mov("eax",&DWP($i,$a,"",0)); # *a
131 &mul($w); # *a * w
132 &add("eax",$c); # L(t)+=c
133 # XXX
134
135 &adc("edx",0); # H(t)+=carry
136 &mov(&DWP($i,$r,"",0),"eax"); # *r= L(t);
137
138 &mov($c,"edx"); # c= H(t);
139 }
140
141 &comment("");
142 &add($a,32);
143 &add($r,32);
144 &sub($num,8);
145 &jz(&label("mw_finish"));
146 &jmp(&label("mw_loop"));
147
148 &set_label("mw_finish",0);
149 &mov($num,&wparam(2)); # get num
150 &and($num,7);
151 &jnz(&label("mw_finish2"));
152 &jmp(&label("mw_end"));
153
154 &set_label("mw_finish2",1);
155 for ($i=0; $i<7; $i++)
156 {
157 &comment("Tail Round $i");
158 &mov("eax",&DWP($i*4,$a,"",0));# *a
159 &mul($w); # *a * w
160 &add("eax",$c); # L(t)+=c
161 # XXX
162 &adc("edx",0); # H(t)+=carry
163 &mov(&DWP($i*4,$r,"",0),"eax");# *r= L(t);
164 &mov($c,"edx"); # c= H(t);
165 &dec($num) if ($i != 7-1);
166 &jz(&label("mw_end")) if ($i != 7-1);
167 }
168 &set_label("mw_end",0);
169 &mov("eax",$c);
170
171 &function_end($name);
172 }
173
174sub bn_sqr_words
175 {
176 local($name)=@_;
177
178 &function_begin($name,"");
179
180 &comment("");
181 $r="esi";
182 $a="edi";
183 $num="ebx";
184
185 &mov($r,&wparam(0)); #
186 &mov($a,&wparam(1)); #
187 &mov($num,&wparam(2)); #
188
189 &and($num,0xfffffff8); # num / 8
190 &jz(&label("sw_finish"));
191
192 &set_label("sw_loop",0);
193 for ($i=0; $i<32; $i+=4)
194 {
195 &comment("Round $i");
196 &mov("eax",&DWP($i,$a,"",0)); # *a
197 # XXX
198 &mul("eax"); # *a * *a
199 &mov(&DWP($i*2,$r,"",0),"eax"); #
200 &mov(&DWP($i*2+4,$r,"",0),"edx");#
201 }
202
203 &comment("");
204 &add($a,32);
205 &add($r,64);
206 &sub($num,8);
207 &jnz(&label("sw_loop"));
208
209 &set_label("sw_finish",0);
210 &mov($num,&wparam(2)); # get num
211 &and($num,7);
212 &jz(&label("sw_end"));
213
214 for ($i=0; $i<7; $i++)
215 {
216 &comment("Tail Round $i");
217 &mov("eax",&DWP($i*4,$a,"",0)); # *a
218 # XXX
219 &mul("eax"); # *a * *a
220 &mov(&DWP($i*8,$r,"",0),"eax"); #
221 &dec($num) if ($i != 7-1);
222 &mov(&DWP($i*8+4,$r,"",0),"edx");
223 &jz(&label("sw_end")) if ($i != 7-1);
224 }
225 &set_label("sw_end",0);
226
227 &function_end($name);
228 }
229
230sub bn_div_words
231 {
232 local($name)=@_;
233
234 &function_begin($name,"");
235 &mov("edx",&wparam(0)); #
236 &mov("eax",&wparam(1)); #
237 &mov("ebx",&wparam(2)); #
238 &div("ebx");
239 &function_end($name);
240 }
241
242sub bn_add_words
243 {
244 local($name)=@_;
245
246 &function_begin($name,"");
247
248 &comment("");
249 $a="esi";
250 $b="edi";
251 $c="eax";
252 $r="ebx";
253 $tmp1="ecx";
254 $tmp2="edx";
255 $num="ebp";
256
257 &mov($r,&wparam(0)); # get r
258 &mov($a,&wparam(1)); # get a
259 &mov($b,&wparam(2)); # get b
260 &mov($num,&wparam(3)); # get num
261 &xor($c,$c); # clear carry
262 &and($num,0xfffffff8); # num / 8
263
264 &jz(&label("aw_finish"));
265
266 &set_label("aw_loop",0);
267 for ($i=0; $i<8; $i++)
268 {
269 &comment("Round $i");
270
271 &mov($tmp1,&DWP($i*4,$a,"",0)); # *a
272 &mov($tmp2,&DWP($i*4,$b,"",0)); # *b
273 &add($tmp1,$c);
274 &mov($c,0);
275 &adc($c,$c);
276 &add($tmp1,$tmp2);
277 &adc($c,0);
278 &mov(&DWP($i*4,$r,"",0),$tmp1); # *r
279 }
280
281 &comment("");
282 &add($a,32);
283 &add($b,32);
284 &add($r,32);
285 &sub($num,8);
286 &jnz(&label("aw_loop"));
287
288 &set_label("aw_finish",0);
289 &mov($num,&wparam(3)); # get num
290 &and($num,7);
291 &jz(&label("aw_end"));
292
293 for ($i=0; $i<7; $i++)
294 {
295 &comment("Tail Round $i");
296 &mov($tmp1,&DWP($i*4,$a,"",0)); # *a
297 &mov($tmp2,&DWP($i*4,$b,"",0));# *b
298 &add($tmp1,$c);
299 &mov($c,0);
300 &adc($c,$c);
301 &add($tmp1,$tmp2);
302 &adc($c,0);
303 &dec($num) if ($i != 6);
304 &mov(&DWP($i*4,$r,"",0),$tmp1); # *r
305 &jz(&label("aw_end")) if ($i != 6);
306 }
307 &set_label("aw_end",0);
308
309# &mov("eax",$c); # $c is "eax"
310
311 &function_end($name);
312 }
313
314sub bn_sub_words
315 {
316 local($name)=@_;
317
318 &function_begin($name,"");
319
320 &comment("");
321 $a="esi";
322 $b="edi";
323 $c="eax";
324 $r="ebx";
325 $tmp1="ecx";
326 $tmp2="edx";
327 $num="ebp";
328
329 &mov($r,&wparam(0)); # get r
330 &mov($a,&wparam(1)); # get a
331 &mov($b,&wparam(2)); # get b
332 &mov($num,&wparam(3)); # get num
333 &xor($c,$c); # clear carry
334 &and($num,0xfffffff8); # num / 8
335
336 &jz(&label("aw_finish"));
337
338 &set_label("aw_loop",0);
339 for ($i=0; $i<8; $i++)
340 {
341 &comment("Round $i");
342
343 &mov($tmp1,&DWP($i*4,$a,"",0)); # *a
344 &mov($tmp2,&DWP($i*4,$b,"",0)); # *b
345 &sub($tmp1,$c);
346 &mov($c,0);
347 &adc($c,$c);
348 &sub($tmp1,$tmp2);
349 &adc($c,0);
350 &mov(&DWP($i*4,$r,"",0),$tmp1); # *r
351 }
352
353 &comment("");
354 &add($a,32);
355 &add($b,32);
356 &add($r,32);
357 &sub($num,8);
358 &jnz(&label("aw_loop"));
359
360 &set_label("aw_finish",0);
361 &mov($num,&wparam(3)); # get num
362 &and($num,7);
363 &jz(&label("aw_end"));
364
365 for ($i=0; $i<7; $i++)
366 {
367 &comment("Tail Round $i");
368 &mov($tmp1,&DWP($i*4,$a,"",0)); # *a
369 &mov($tmp2,&DWP($i*4,$b,"",0));# *b
370 &sub($tmp1,$c);
371 &mov($c,0);
372 &adc($c,$c);
373 &sub($tmp1,$tmp2);
374 &adc($c,0);
375 &dec($num) if ($i != 6);
376 &mov(&DWP($i*4,$r,"",0),$tmp1); # *r
377 &jz(&label("aw_end")) if ($i != 6);
378 }
379 &set_label("aw_end",0);
380
381# &mov("eax",$c); # $c is "eax"
382
383 &function_end($name);
384 }
385
386sub bn_sub_part_words
387 {
388 local($name)=@_;
389
390 &function_begin($name,"");
391
392 &comment("");
393 $a="esi";
394 $b="edi";
395 $c="eax";
396 $r="ebx";
397 $tmp1="ecx";
398 $tmp2="edx";
399 $num="ebp";
400
401 &mov($r,&wparam(0)); # get r
402 &mov($a,&wparam(1)); # get a
403 &mov($b,&wparam(2)); # get b
404 &mov($num,&wparam(3)); # get num
405 &xor($c,$c); # clear carry
406 &and($num,0xfffffff8); # num / 8
407
408 &jz(&label("aw_finish"));
409
410 &set_label("aw_loop",0);
411 for ($i=0; $i<8; $i++)
412 {
413 &comment("Round $i");
414
415 &mov($tmp1,&DWP($i*4,$a,"",0)); # *a
416 &mov($tmp2,&DWP($i*4,$b,"",0)); # *b
417 &sub($tmp1,$c);
418 &mov($c,0);
419 &adc($c,$c);
420 &sub($tmp1,$tmp2);
421 &adc($c,0);
422 &mov(&DWP($i*4,$r,"",0),$tmp1); # *r
423 }
424
425 &comment("");
426 &add($a,32);
427 &add($b,32);
428 &add($r,32);
429 &sub($num,8);
430 &jnz(&label("aw_loop"));
431
432 &set_label("aw_finish",0);
433 &mov($num,&wparam(3)); # get num
434 &and($num,7);
435 &jz(&label("aw_end"));
436
437 for ($i=0; $i<7; $i++)
438 {
439 &comment("Tail Round $i");
440 &mov($tmp1,&DWP(0,$a,"",0)); # *a
441 &mov($tmp2,&DWP(0,$b,"",0));# *b
442 &sub($tmp1,$c);
443 &mov($c,0);
444 &adc($c,$c);
445 &sub($tmp1,$tmp2);
446 &adc($c,0);
447 &mov(&DWP(0,$r,"",0),$tmp1); # *r
448 &add($a, 4);
449 &add($b, 4);
450 &add($r, 4);
451 &dec($num) if ($i != 6);
452 &jz(&label("aw_end")) if ($i != 6);
453 }
454 &set_label("aw_end",0);
455
456 &cmp(&wparam(4),0);
457 &je(&label("pw_end"));
458
459 &mov($num,&wparam(4)); # get dl
460 &cmp($num,0);
461 &je(&label("pw_end"));
462 &jge(&label("pw_pos"));
463
464 &comment("pw_neg");
465 &mov($tmp2,0);
466 &sub($tmp2,$num);
467 &mov($num,$tmp2);
468 &and($num,0xfffffff8); # num / 8
469 &jz(&label("pw_neg_finish"));
470
471 &set_label("pw_neg_loop",0);
472 for ($i=0; $i<8; $i++)
473 {
474 &comment("dl<0 Round $i");
475
476 &mov($tmp1,0);
477 &mov($tmp2,&DWP($i*4,$b,"",0)); # *b
478 &sub($tmp1,$c);
479 &mov($c,0);
480 &adc($c,$c);
481 &sub($tmp1,$tmp2);
482 &adc($c,0);
483 &mov(&DWP($i*4,$r,"",0),$tmp1); # *r
484 }
485
486 &comment("");
487 &add($b,32);
488 &add($r,32);
489 &sub($num,8);
490 &jnz(&label("pw_neg_loop"));
491
492 &set_label("pw_neg_finish",0);
493 &mov($tmp2,&wparam(4)); # get dl
494 &mov($num,0);
495 &sub($num,$tmp2);
496 &and($num,7);
497 &jz(&label("pw_end"));
498
499 for ($i=0; $i<7; $i++)
500 {
501 &comment("dl<0 Tail Round $i");
502 &mov($tmp1,0);
503 &mov($tmp2,&DWP($i*4,$b,"",0));# *b
504 &sub($tmp1,$c);
505 &mov($c,0);
506 &adc($c,$c);
507 &sub($tmp1,$tmp2);
508 &adc($c,0);
509 &dec($num) if ($i != 6);
510 &mov(&DWP($i*4,$r,"",0),$tmp1); # *r
511 &jz(&label("pw_end")) if ($i != 6);
512 }
513
514 &jmp(&label("pw_end"));
515
516 &set_label("pw_pos",0);
517
518 &and($num,0xfffffff8); # num / 8
519 &jz(&label("pw_pos_finish"));
520
521 &set_label("pw_pos_loop",0);
522
523 for ($i=0; $i<8; $i++)
524 {
525 &comment("dl>0 Round $i");
526
527 &mov($tmp1,&DWP($i*4,$a,"",0)); # *a
528 &sub($tmp1,$c);
529 &mov(&DWP($i*4,$r,"",0),$tmp1); # *r
530 &jnc(&label("pw_nc".$i));
531 }
532
533 &comment("");
534 &add($a,32);
535 &add($r,32);
536 &sub($num,8);
537 &jnz(&label("pw_pos_loop"));
538
539 &set_label("pw_pos_finish",0);
540 &mov($num,&wparam(4)); # get dl
541 &and($num,7);
542 &jz(&label("pw_end"));
543
544 for ($i=0; $i<7; $i++)
545 {
546 &comment("dl>0 Tail Round $i");
547 &mov($tmp1,&DWP($i*4,$a,"",0)); # *a
548 &sub($tmp1,$c);
549 &mov(&DWP($i*4,$r,"",0),$tmp1); # *r
550 &jnc(&label("pw_tail_nc".$i));
551 &dec($num) if ($i != 6);
552 &jz(&label("pw_end")) if ($i != 6);
553 }
554 &mov($c,1);
555 &jmp(&label("pw_end"));
556
557 &set_label("pw_nc_loop",0);
558 for ($i=0; $i<8; $i++)
559 {
560 &mov($tmp1,&DWP($i*4,$a,"",0)); # *a
561 &mov(&DWP($i*4,$r,"",0),$tmp1); # *r
562 &set_label("pw_nc".$i,0);
563 }
564
565 &comment("");
566 &add($a,32);
567 &add($r,32);
568 &sub($num,8);
569 &jnz(&label("pw_nc_loop"));
570
571 &mov($num,&wparam(4)); # get dl
572 &and($num,7);
573 &jz(&label("pw_nc_end"));
574
575 for ($i=0; $i<7; $i++)
576 {
577 &mov($tmp1,&DWP($i*4,$a,"",0)); # *a
578 &mov(&DWP($i*4,$r,"",0),$tmp1); # *r
579 &set_label("pw_tail_nc".$i,0);
580 &dec($num) if ($i != 6);
581 &jz(&label("pw_nc_end")) if ($i != 6);
582 }
583
584 &set_label("pw_nc_end",0);
585 &mov($c,0);
586
587 &set_label("pw_end",0);
588
589# &mov("eax",$c); # $c is "eax"
590
591 &function_end($name);
592 }
593
diff --git a/src/lib/libcrypto/bn/asm/co-586.pl b/src/lib/libcrypto/bn/asm/co-586.pl
deleted file mode 100644
index 5d962cb957..0000000000
--- a/src/lib/libcrypto/bn/asm/co-586.pl
+++ /dev/null
@@ -1,286 +0,0 @@
1#!/usr/local/bin/perl
2
3push(@INC,"perlasm","../../perlasm");
4require "x86asm.pl";
5
6&asm_init($ARGV[0],$0);
7
8&bn_mul_comba("bn_mul_comba8",8);
9&bn_mul_comba("bn_mul_comba4",4);
10&bn_sqr_comba("bn_sqr_comba8",8);
11&bn_sqr_comba("bn_sqr_comba4",4);
12
13&asm_finish();
14
15sub mul_add_c
16 {
17 local($a,$ai,$b,$bi,$c0,$c1,$c2,$pos,$i,$na,$nb)=@_;
18
19 # pos == -1 if eax and edx are pre-loaded, 0 to load from next
20 # words, and 1 if load return value
21
22 &comment("mul a[$ai]*b[$bi]");
23
24 # "eax" and "edx" will always be pre-loaded.
25 # &mov("eax",&DWP($ai*4,$a,"",0)) ;
26 # &mov("edx",&DWP($bi*4,$b,"",0));
27
28 &mul("edx");
29 &add($c0,"eax");
30 &mov("eax",&DWP(($na)*4,$a,"",0)) if $pos == 0; # laod next a
31 &mov("eax",&wparam(0)) if $pos > 0; # load r[]
32 ###
33 &adc($c1,"edx");
34 &mov("edx",&DWP(($nb)*4,$b,"",0)) if $pos == 0; # laod next b
35 &mov("edx",&DWP(($nb)*4,$b,"",0)) if $pos == 1; # laod next b
36 ###
37 &adc($c2,0);
38 # is pos > 1, it means it is the last loop
39 &mov(&DWP($i*4,"eax","",0),$c0) if $pos > 0; # save r[];
40 &mov("eax",&DWP(($na)*4,$a,"",0)) if $pos == 1; # laod next a
41 }
42
43sub sqr_add_c
44 {
45 local($r,$a,$ai,$bi,$c0,$c1,$c2,$pos,$i,$na,$nb)=@_;
46
47 # pos == -1 if eax and edx are pre-loaded, 0 to load from next
48 # words, and 1 if load return value
49
50 &comment("sqr a[$ai]*a[$bi]");
51
52 # "eax" and "edx" will always be pre-loaded.
53 # &mov("eax",&DWP($ai*4,$a,"",0)) ;
54 # &mov("edx",&DWP($bi*4,$b,"",0));
55
56 if ($ai == $bi)
57 { &mul("eax");}
58 else
59 { &mul("edx");}
60 &add($c0,"eax");
61 &mov("eax",&DWP(($na)*4,$a,"",0)) if $pos == 0; # load next a
62 ###
63 &adc($c1,"edx");
64 &mov("edx",&DWP(($nb)*4,$a,"",0)) if ($pos == 1) && ($na != $nb);
65 ###
66 &adc($c2,0);
67 # is pos > 1, it means it is the last loop
68 &mov(&DWP($i*4,$r,"",0),$c0) if $pos > 0; # save r[];
69 &mov("eax",&DWP(($na)*4,$a,"",0)) if $pos == 1; # load next b
70 }
71
72sub sqr_add_c2
73 {
74 local($r,$a,$ai,$bi,$c0,$c1,$c2,$pos,$i,$na,$nb)=@_;
75
76 # pos == -1 if eax and edx are pre-loaded, 0 to load from next
77 # words, and 1 if load return value
78
79 &comment("sqr a[$ai]*a[$bi]");
80
81 # "eax" and "edx" will always be pre-loaded.
82 # &mov("eax",&DWP($ai*4,$a,"",0)) ;
83 # &mov("edx",&DWP($bi*4,$a,"",0));
84
85 if ($ai == $bi)
86 { &mul("eax");}
87 else
88 { &mul("edx");}
89 &add("eax","eax");
90 ###
91 &adc("edx","edx");
92 ###
93 &adc($c2,0);
94 &add($c0,"eax");
95 &adc($c1,"edx");
96 &mov("eax",&DWP(($na)*4,$a,"",0)) if $pos == 0; # load next a
97 &mov("eax",&DWP(($na)*4,$a,"",0)) if $pos == 1; # load next b
98 &adc($c2,0);
99 &mov(&DWP($i*4,$r,"",0),$c0) if $pos > 0; # save r[];
100 &mov("edx",&DWP(($nb)*4,$a,"",0)) if ($pos <= 1) && ($na != $nb);
101 ###
102 }
103
104sub bn_mul_comba
105 {
106 local($name,$num)=@_;
107 local($a,$b,$c0,$c1,$c2);
108 local($i,$as,$ae,$bs,$be,$ai,$bi);
109 local($tot,$end);
110
111 &function_begin_B($name,"");
112
113 $c0="ebx";
114 $c1="ecx";
115 $c2="ebp";
116 $a="esi";
117 $b="edi";
118
119 $as=0;
120 $ae=0;
121 $bs=0;
122 $be=0;
123 $tot=$num+$num-1;
124
125 &push("esi");
126 &mov($a,&wparam(1));
127 &push("edi");
128 &mov($b,&wparam(2));
129 &push("ebp");
130 &push("ebx");
131
132 &xor($c0,$c0);
133 &mov("eax",&DWP(0,$a,"",0)); # load the first word
134 &xor($c1,$c1);
135 &mov("edx",&DWP(0,$b,"",0)); # load the first second
136
137 for ($i=0; $i<$tot; $i++)
138 {
139 $ai=$as;
140 $bi=$bs;
141 $end=$be+1;
142
143 &comment("################## Calculate word $i");
144
145 for ($j=$bs; $j<$end; $j++)
146 {
147 &xor($c2,$c2) if ($j == $bs);
148 if (($j+1) == $end)
149 {
150 $v=1;
151 $v=2 if (($i+1) == $tot);
152 }
153 else
154 { $v=0; }
155 if (($j+1) != $end)
156 {
157 $na=($ai-1);
158 $nb=($bi+1);
159 }
160 else
161 {
162 $na=$as+($i < ($num-1));
163 $nb=$bs+($i >= ($num-1));
164 }
165#printf STDERR "[$ai,$bi] -> [$na,$nb]\n";
166 &mul_add_c($a,$ai,$b,$bi,$c0,$c1,$c2,$v,$i,$na,$nb);
167 if ($v)
168 {
169 &comment("saved r[$i]");
170 # &mov("eax",&wparam(0));
171 # &mov(&DWP($i*4,"eax","",0),$c0);
172 ($c0,$c1,$c2)=($c1,$c2,$c0);
173 }
174 $ai--;
175 $bi++;
176 }
177 $as++ if ($i < ($num-1));
178 $ae++ if ($i >= ($num-1));
179
180 $bs++ if ($i >= ($num-1));
181 $be++ if ($i < ($num-1));
182 }
183 &comment("save r[$i]");
184 # &mov("eax",&wparam(0));
185 &mov(&DWP($i*4,"eax","",0),$c0);
186
187 &pop("ebx");
188 &pop("ebp");
189 &pop("edi");
190 &pop("esi");
191 &ret();
192 &function_end_B($name);
193 }
194
195sub bn_sqr_comba
196 {
197 local($name,$num)=@_;
198 local($r,$a,$c0,$c1,$c2)=@_;
199 local($i,$as,$ae,$bs,$be,$ai,$bi);
200 local($b,$tot,$end,$half);
201
202 &function_begin_B($name,"");
203
204 $c0="ebx";
205 $c1="ecx";
206 $c2="ebp";
207 $a="esi";
208 $r="edi";
209
210 &push("esi");
211 &push("edi");
212 &push("ebp");
213 &push("ebx");
214 &mov($r,&wparam(0));
215 &mov($a,&wparam(1));
216 &xor($c0,$c0);
217 &xor($c1,$c1);
218 &mov("eax",&DWP(0,$a,"",0)); # load the first word
219
220 $as=0;
221 $ae=0;
222 $bs=0;
223 $be=0;
224 $tot=$num+$num-1;
225
226 for ($i=0; $i<$tot; $i++)
227 {
228 $ai=$as;
229 $bi=$bs;
230 $end=$be+1;
231
232 &comment("############### Calculate word $i");
233 for ($j=$bs; $j<$end; $j++)
234 {
235 &xor($c2,$c2) if ($j == $bs);
236 if (($ai-1) < ($bi+1))
237 {
238 $v=1;
239 $v=2 if ($i+1) == $tot;
240 }
241 else
242 { $v=0; }
243 if (!$v)
244 {
245 $na=$ai-1;
246 $nb=$bi+1;
247 }
248 else
249 {
250 $na=$as+($i < ($num-1));
251 $nb=$bs+($i >= ($num-1));
252 }
253 if ($ai == $bi)
254 {
255 &sqr_add_c($r,$a,$ai,$bi,
256 $c0,$c1,$c2,$v,$i,$na,$nb);
257 }
258 else
259 {
260 &sqr_add_c2($r,$a,$ai,$bi,
261 $c0,$c1,$c2,$v,$i,$na,$nb);
262 }
263 if ($v)
264 {
265 &comment("saved r[$i]");
266 #&mov(&DWP($i*4,$r,"",0),$c0);
267 ($c0,$c1,$c2)=($c1,$c2,$c0);
268 last;
269 }
270 $ai--;
271 $bi++;
272 }
273 $as++ if ($i < ($num-1));
274 $ae++ if ($i >= ($num-1));
275
276 $bs++ if ($i >= ($num-1));
277 $be++ if ($i < ($num-1));
278 }
279 &mov(&DWP($i*4,$r,"",0),$c0);
280 &pop("ebx");
281 &pop("ebp");
282 &pop("edi");
283 &pop("esi");
284 &ret();
285 &function_end_B($name);
286 }
diff --git a/src/lib/libcrypto/bn/asm/ia64.S b/src/lib/libcrypto/bn/asm/ia64.S
deleted file mode 100644
index 7dfda85566..0000000000
--- a/src/lib/libcrypto/bn/asm/ia64.S
+++ /dev/null
@@ -1,1605 +0,0 @@
1.explicit
2.text
3.ident "ia64.S, Version 2.0"
4.ident "IA-64 ISA artwork by Andy Polyakov <appro@fy.chalmers.se>"
5
6//
7// ====================================================================
8// Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
9// project.
10//
11// Rights for redistribution and usage in source and binary forms are
12// granted according to the OpenSSL license. Warranty of any kind is
13// disclaimed.
14// ====================================================================
15//
16// Version 2.x is Itanium2 re-tune. Few words about how Itanum2 is
17// different from Itanium to this module viewpoint. Most notably, is it
18// "wider" than Itanium? Can you experience loop scalability as
19// discussed in commentary sections? Not really:-( Itanium2 has 6
20// integer ALU ports, i.e. it's 2 ports wider, but it's not enough to
21// spin twice as fast, as I need 8 IALU ports. Amount of floating point
22// ports is the same, i.e. 2, while I need 4. In other words, to this
23// module Itanium2 remains effectively as "wide" as Itanium. Yet it's
24// essentially different in respect to this module, and a re-tune was
25// required. Well, because some intruction latencies has changed. Most
26// noticeably those intensively used:
27//
28// Itanium Itanium2
29// ldf8 9 6 L2 hit
30// ld8 2 1 L1 hit
31// getf 2 5
32// xma[->getf] 7[+1] 4[+0]
33// add[->st8] 1[+1] 1[+0]
34//
35// What does it mean? You might ratiocinate that the original code
36// should run just faster... Because sum of latencies is smaller...
37// Wrong! Note that getf latency increased. This means that if a loop is
38// scheduled for lower latency (and they are), then it will suffer from
39// stall condition and the code will therefore turn anti-scalable, e.g.
40// original bn_mul_words spun at 5*n or 2.5 times slower than expected
41// on Itanium2! What to do? Reschedule loops for Itanium2? But then
42// Itanium would exhibit anti-scalability. So I've chosen to reschedule
43// for worst latency for every instruction aiming for best *all-round*
44// performance.
45
46// Q. How much faster does it get?
47// A. Here is the output from 'openssl speed rsa dsa' for vanilla
48// 0.9.6a compiled with gcc version 2.96 20000731 (Red Hat
49// Linux 7.1 2.96-81):
50//
51// sign verify sign/s verify/s
52// rsa 512 bits 0.0036s 0.0003s 275.3 2999.2
53// rsa 1024 bits 0.0203s 0.0011s 49.3 894.1
54// rsa 2048 bits 0.1331s 0.0040s 7.5 250.9
55// rsa 4096 bits 0.9270s 0.0147s 1.1 68.1
56// sign verify sign/s verify/s
57// dsa 512 bits 0.0035s 0.0043s 288.3 234.8
58// dsa 1024 bits 0.0111s 0.0135s 90.0 74.2
59//
60// And here is similar output but for this assembler
61// implementation:-)
62//
63// sign verify sign/s verify/s
64// rsa 512 bits 0.0021s 0.0001s 549.4 9638.5
65// rsa 1024 bits 0.0055s 0.0002s 183.8 4481.1
66// rsa 2048 bits 0.0244s 0.0006s 41.4 1726.3
67// rsa 4096 bits 0.1295s 0.0018s 7.7 561.5
68// sign verify sign/s verify/s
69// dsa 512 bits 0.0012s 0.0013s 891.9 756.6
70// dsa 1024 bits 0.0023s 0.0028s 440.4 376.2
71//
72// Yes, you may argue that it's not fair comparison as it's
73// possible to craft the C implementation with BN_UMULT_HIGH
74// inline assembler macro. But of course! Here is the output
75// with the macro:
76//
77// sign verify sign/s verify/s
78// rsa 512 bits 0.0020s 0.0002s 495.0 6561.0
79// rsa 1024 bits 0.0086s 0.0004s 116.2 2235.7
80// rsa 2048 bits 0.0519s 0.0015s 19.3 667.3
81// rsa 4096 bits 0.3464s 0.0053s 2.9 187.7
82// sign verify sign/s verify/s
83// dsa 512 bits 0.0016s 0.0020s 613.1 510.5
84// dsa 1024 bits 0.0045s 0.0054s 221.0 183.9
85//
86// My code is still way faster, huh:-) And I believe that even
87// higher performance can be achieved. Note that as keys get
88// longer, performance gain is larger. Why? According to the
89// profiler there is another player in the field, namely
90// BN_from_montgomery consuming larger and larger portion of CPU
91// time as keysize decreases. I therefore consider putting effort
92// to assembler implementation of the following routine:
93//
94// void bn_mul_add_mont (BN_ULONG *rp,BN_ULONG *np,int nl,BN_ULONG n0)
95// {
96// int i,j;
97// BN_ULONG v;
98//
99// for (i=0; i<nl; i++)
100// {
101// v=bn_mul_add_words(rp,np,nl,(rp[0]*n0)&BN_MASK2);
102// nrp++;
103// rp++;
104// if (((nrp[-1]+=v)&BN_MASK2) < v)
105// for (j=0; ((++nrp[j])&BN_MASK2) == 0; j++) ;
106// }
107// }
108//
109// It might as well be beneficial to implement even combaX
110// variants, as it appears as it can literally unleash the
111// performance (see comment section to bn_mul_comba8 below).
112//
113// And finally for your reference the output for 0.9.6a compiled
114// with SGIcc version 0.01.0-12 (keep in mind that for the moment
115// of this writing it's not possible to convince SGIcc to use
116// BN_UMULT_HIGH inline assembler macro, yet the code is fast,
117// i.e. for a compiler generated one:-):
118//
119// sign verify sign/s verify/s
120// rsa 512 bits 0.0022s 0.0002s 452.7 5894.3
121// rsa 1024 bits 0.0097s 0.0005s 102.7 2002.9
122// rsa 2048 bits 0.0578s 0.0017s 17.3 600.2
123// rsa 4096 bits 0.3838s 0.0061s 2.6 164.5
124// sign verify sign/s verify/s
125// dsa 512 bits 0.0018s 0.0022s 547.3 459.6
126// dsa 1024 bits 0.0051s 0.0062s 196.6 161.3
127//
128// Oh! Benchmarks were performed on 733MHz Lion-class Itanium
129// system running Redhat Linux 7.1 (very special thanks to Ray
130// McCaffity of Williams Communications for providing an account).
131//
132// Q. What's the heck with 'rum 1<<5' at the end of every function?
133// A. Well, by clearing the "upper FP registers written" bit of the
134// User Mask I want to excuse the kernel from preserving upper
135// (f32-f128) FP register bank over process context switch, thus
136// minimizing bus bandwidth consumption during the switch (i.e.
137// after PKI opration completes and the program is off doing
138// something else like bulk symmetric encryption). Having said
139// this, I also want to point out that it might be good idea
140// to compile the whole toolkit (as well as majority of the
141// programs for that matter) with -mfixed-range=f32-f127 command
142// line option. No, it doesn't prevent the compiler from writing
143// to upper bank, but at least discourages to do so. If you don't
144// like the idea you have the option to compile the module with
145// -Drum=nop.m in command line.
146//
147
148#if 1
149//
150// bn_[add|sub]_words routines.
151//
152// Loops are spinning in 2*(n+5) ticks on Itanuim (provided that the
153// data reside in L1 cache, i.e. 2 ticks away). It's possible to
154// compress the epilogue and get down to 2*n+6, but at the cost of
155// scalability (the neat feature of this implementation is that it
156// shall automagically spin in n+5 on "wider" IA-64 implementations:-)
157// I consider that the epilogue is short enough as it is to trade tiny
158// performance loss on Itanium for scalability.
159//
160// BN_ULONG bn_add_words(BN_ULONG *rp, BN_ULONG *ap, BN_ULONG *bp,int num)
161//
162.global bn_add_words#
163.proc bn_add_words#
164.align 64
165.skip 32 // makes the loop body aligned at 64-byte boundary
166bn_add_words:
167 .prologue
168 .fframe 0
169 .save ar.pfs,r2
170{ .mii; alloc r2=ar.pfs,4,12,0,16
171 cmp4.le p6,p0=r35,r0 };;
172{ .mfb; mov r8=r0 // return value
173(p6) br.ret.spnt.many b0 };;
174
175 .save ar.lc,r3
176{ .mib; sub r10=r35,r0,1
177 mov r3=ar.lc
178 brp.loop.imp .L_bn_add_words_ctop,.L_bn_add_words_cend-16
179 }
180 .body
181{ .mib;
182#if defined(_HPUX_SOURCE) && defined(_ILP32)
183 addp4 r14=0,r32 // rp
184#else
185 mov r14=r32 // rp
186#endif
187 mov r9=pr };;
188{ .mii;
189#if defined(_HPUX_SOURCE) && defined(_ILP32)
190 addp4 r15=0,r33 // ap
191#else
192 mov r15=r33 // ap
193#endif
194 mov ar.lc=r10
195 mov ar.ec=6 }
196{ .mib;
197#if defined(_HPUX_SOURCE) && defined(_ILP32)
198 addp4 r16=0,r34 // bp
199#else
200 mov r16=r34 // bp
201#endif
202 mov pr.rot=1<<16 };;
203
204.L_bn_add_words_ctop:
205{ .mii; (p16) ld8 r32=[r16],8 // b=*(bp++)
206 (p18) add r39=r37,r34
207 (p19) cmp.ltu.unc p56,p0=r40,r38 }
208{ .mfb; (p0) nop.m 0x0
209 (p0) nop.f 0x0
210 (p0) nop.b 0x0 }
211{ .mii; (p16) ld8 r35=[r15],8 // a=*(ap++)
212 (p58) cmp.eq.or p57,p0=-1,r41 // (p20)
213 (p58) add r41=1,r41 } // (p20)
214{ .mfb; (p21) st8 [r14]=r42,8 // *(rp++)=r
215 (p0) nop.f 0x0
216 br.ctop.sptk .L_bn_add_words_ctop };;
217.L_bn_add_words_cend:
218
219{ .mii;
220(p59) add r8=1,r8 // return value
221 mov pr=r9,0x1ffff
222 mov ar.lc=r3 }
223{ .mbb; nop.b 0x0
224 br.ret.sptk.many b0 };;
225.endp bn_add_words#
226
227//
228// BN_ULONG bn_sub_words(BN_ULONG *rp, BN_ULONG *ap, BN_ULONG *bp,int num)
229//
230.global bn_sub_words#
231.proc bn_sub_words#
232.align 64
233.skip 32 // makes the loop body aligned at 64-byte boundary
234bn_sub_words:
235 .prologue
236 .fframe 0
237 .save ar.pfs,r2
238{ .mii; alloc r2=ar.pfs,4,12,0,16
239 cmp4.le p6,p0=r35,r0 };;
240{ .mfb; mov r8=r0 // return value
241(p6) br.ret.spnt.many b0 };;
242
243 .save ar.lc,r3
244{ .mib; sub r10=r35,r0,1
245 mov r3=ar.lc
246 brp.loop.imp .L_bn_sub_words_ctop,.L_bn_sub_words_cend-16
247 }
248 .body
249{ .mib;
250#if defined(_HPUX_SOURCE) && defined(_ILP32)
251 addp4 r14=0,r32 // rp
252#else
253 mov r14=r32 // rp
254#endif
255 mov r9=pr };;
256{ .mii;
257#if defined(_HPUX_SOURCE) && defined(_ILP32)
258 addp4 r15=0,r33 // ap
259#else
260 mov r15=r33 // ap
261#endif
262 mov ar.lc=r10
263 mov ar.ec=6 }
264{ .mib;
265#if defined(_HPUX_SOURCE) && defined(_ILP32)
266 addp4 r16=0,r34 // bp
267#else
268 mov r16=r34 // bp
269#endif
270 mov pr.rot=1<<16 };;
271
272.L_bn_sub_words_ctop:
273{ .mii; (p16) ld8 r32=[r16],8 // b=*(bp++)
274 (p18) sub r39=r37,r34
275 (p19) cmp.gtu.unc p56,p0=r40,r38 }
276{ .mfb; (p0) nop.m 0x0
277 (p0) nop.f 0x0
278 (p0) nop.b 0x0 }
279{ .mii; (p16) ld8 r35=[r15],8 // a=*(ap++)
280 (p58) cmp.eq.or p57,p0=0,r41 // (p20)
281 (p58) add r41=-1,r41 } // (p20)
282{ .mbb; (p21) st8 [r14]=r42,8 // *(rp++)=r
283 (p0) nop.b 0x0
284 br.ctop.sptk .L_bn_sub_words_ctop };;
285.L_bn_sub_words_cend:
286
287{ .mii;
288(p59) add r8=1,r8 // return value
289 mov pr=r9,0x1ffff
290 mov ar.lc=r3 }
291{ .mbb; nop.b 0x0
292 br.ret.sptk.many b0 };;
293.endp bn_sub_words#
294#endif
295
296#if 0
297#define XMA_TEMPTATION
298#endif
299
300#if 1
301//
302// BN_ULONG bn_mul_words(BN_ULONG *rp, BN_ULONG *ap, int num, BN_ULONG w)
303//
304.global bn_mul_words#
305.proc bn_mul_words#
306.align 64
307.skip 32 // makes the loop body aligned at 64-byte boundary
308bn_mul_words:
309 .prologue
310 .fframe 0
311 .save ar.pfs,r2
312#ifdef XMA_TEMPTATION
313{ .mfi; alloc r2=ar.pfs,4,0,0,0 };;
314#else
315{ .mfi; alloc r2=ar.pfs,4,12,0,16 };;
316#endif
317{ .mib; mov r8=r0 // return value
318 cmp4.le p6,p0=r34,r0
319(p6) br.ret.spnt.many b0 };;
320
321 .save ar.lc,r3
322{ .mii; sub r10=r34,r0,1
323 mov r3=ar.lc
324 mov r9=pr };;
325
326 .body
327{ .mib; setf.sig f8=r35 // w
328 mov pr.rot=0x800001<<16
329 // ------^----- serves as (p50) at first (p27)
330 brp.loop.imp .L_bn_mul_words_ctop,.L_bn_mul_words_cend-16
331 }
332
333#ifndef XMA_TEMPTATION
334
335{ .mii;
336#if defined(_HPUX_SOURCE) && defined(_ILP32)
337 addp4 r14=0,r32 // rp
338 addp4 r15=0,r33 // ap
339#else
340 mov r14=r32 // rp
341 mov r15=r33 // ap
342#endif
343 mov ar.lc=r10 }
344{ .mii; mov r40=0 // serves as r35 at first (p27)
345 mov ar.ec=13 };;
346
347// This loop spins in 2*(n+12) ticks. It's scheduled for data in Itanium
348// L2 cache (i.e. 9 ticks away) as floating point load/store instructions
349// bypass L1 cache and L2 latency is actually best-case scenario for
350// ldf8. The loop is not scalable and shall run in 2*(n+12) even on
351// "wider" IA-64 implementations. It's a trade-off here. n+24 loop
352// would give us ~5% in *overall* performance improvement on "wider"
353// IA-64, but would hurt Itanium for about same because of longer
354// epilogue. As it's a matter of few percents in either case I've
355// chosen to trade the scalability for development time (you can see
356// this very instruction sequence in bn_mul_add_words loop which in
357// turn is scalable).
358.L_bn_mul_words_ctop:
359{ .mfi; (p25) getf.sig r36=f52 // low
360 (p21) xmpy.lu f48=f37,f8
361 (p28) cmp.ltu p54,p50=r41,r39 }
362{ .mfi; (p16) ldf8 f32=[r15],8
363 (p21) xmpy.hu f40=f37,f8
364 (p0) nop.i 0x0 };;
365{ .mii; (p25) getf.sig r32=f44 // high
366 .pred.rel "mutex",p50,p54
367 (p50) add r40=r38,r35 // (p27)
368 (p54) add r40=r38,r35,1 } // (p27)
369{ .mfb; (p28) st8 [r14]=r41,8
370 (p0) nop.f 0x0
371 br.ctop.sptk .L_bn_mul_words_ctop };;
372.L_bn_mul_words_cend:
373
374{ .mii; nop.m 0x0
375.pred.rel "mutex",p51,p55
376(p51) add r8=r36,r0
377(p55) add r8=r36,r0,1 }
378{ .mfb; nop.m 0x0
379 nop.f 0x0
380 nop.b 0x0 }
381
382#else // XMA_TEMPTATION
383
384 setf.sig f37=r0 // serves as carry at (p18) tick
385 mov ar.lc=r10
386 mov ar.ec=5;;
387
388// Most of you examining this code very likely wonder why in the name
389// of Intel the following loop is commented out? Indeed, it looks so
390// neat that you find it hard to believe that it's something wrong
391// with it, right? The catch is that every iteration depends on the
392// result from previous one and the latter isn't available instantly.
393// The loop therefore spins at the latency of xma minus 1, or in other
394// words at 6*(n+4) ticks:-( Compare to the "production" loop above
395// that runs in 2*(n+11) where the low latency problem is worked around
396// by moving the dependency to one-tick latent interger ALU. Note that
397// "distance" between ldf8 and xma is not latency of ldf8, but the
398// *difference* between xma and ldf8 latencies.
399.L_bn_mul_words_ctop:
400{ .mfi; (p16) ldf8 f32=[r33],8
401 (p18) xma.hu f38=f34,f8,f39 }
402{ .mfb; (p20) stf8 [r32]=f37,8
403 (p18) xma.lu f35=f34,f8,f39
404 br.ctop.sptk .L_bn_mul_words_ctop };;
405.L_bn_mul_words_cend:
406
407 getf.sig r8=f41 // the return value
408
409#endif // XMA_TEMPTATION
410
411{ .mii; nop.m 0x0
412 mov pr=r9,0x1ffff
413 mov ar.lc=r3 }
414{ .mfb; rum 1<<5 // clear um.mfh
415 nop.f 0x0
416 br.ret.sptk.many b0 };;
417.endp bn_mul_words#
418#endif
419
420#if 1
421//
422// BN_ULONG bn_mul_add_words(BN_ULONG *rp, BN_ULONG *ap, int num, BN_ULONG w)
423//
424.global bn_mul_add_words#
425.proc bn_mul_add_words#
426.align 64
427//.skip 0 // makes the loop split at 64-byte boundary
428bn_mul_add_words:
429 .prologue
430 .fframe 0
431 .save ar.pfs,r2
432{ .mii; alloc r2=ar.pfs,4,12,0,16
433 cmp4.le p6,p0=r34,r0 };;
434{ .mfb; mov r8=r0 // return value
435(p6) br.ret.spnt.many b0 };;
436
437 .save ar.lc,r3
438{ .mii; sub r10=r34,r0,1
439 mov r3=ar.lc
440 mov r9=pr };;
441
442 .body
443{ .mib; setf.sig f8=r35 // w
444 mov pr.rot=0x800001<<16
445 // ------^----- serves as (p50) at first (p27)
446 brp.loop.imp .L_bn_mul_add_words_ctop,.L_bn_mul_add_words_cend-16
447 }
448{ .mii;
449#if defined(_HPUX_SOURCE) && defined(_ILP32)
450 addp4 r14=0,r32 // rp
451 addp4 r15=0,r33 // ap
452#else
453 mov r14=r32 // rp
454 mov r15=r33 // ap
455#endif
456 mov ar.lc=r10 }
457{ .mii; mov r40=0 // serves as r35 at first (p27)
458#if defined(_HPUX_SOURCE) && defined(_ILP32)
459 addp4 r18=0,r32 // rp copy
460#else
461 mov r18=r32 // rp copy
462#endif
463 mov ar.ec=15 };;
464
465// This loop spins in 3*(n+14) ticks on Itanium and should spin in
466// 2*(n+14) on "wider" IA-64 implementations (to be verified with new
467// µ-architecture manuals as they become available). As usual it's
468// possible to compress the epilogue, down to 10 in this case, at the
469// cost of scalability. Compressed (and therefore non-scalable) loop
470// running at 3*(n+11) would buy you ~10% on Itanium but take ~35%
471// from "wider" IA-64 so let it be scalable! Special attention was
472// paid for having the loop body split at 64-byte boundary. ld8 is
473// scheduled for L1 cache as the data is more than likely there.
474// Indeed, bn_mul_words has put it there a moment ago:-)
475.L_bn_mul_add_words_ctop:
476{ .mfi; (p25) getf.sig r36=f52 // low
477 (p21) xmpy.lu f48=f37,f8
478 (p28) cmp.ltu p54,p50=r41,r39 }
479{ .mfi; (p16) ldf8 f32=[r15],8
480 (p21) xmpy.hu f40=f37,f8
481 (p28) add r45=r45,r41 };;
482{ .mii; (p25) getf.sig r32=f44 // high
483 .pred.rel "mutex",p50,p54
484 (p50) add r40=r38,r35 // (p27)
485 (p54) add r40=r38,r35,1 } // (p27)
486{ .mfb; (p28) cmp.ltu.unc p60,p0=r45,r41
487 (p0) nop.f 0x0
488 (p0) nop.b 0x0 }
489{ .mii; (p27) ld8 r44=[r18],8
490 (p62) cmp.eq.or p61,p0=-1,r46
491 (p62) add r46=1,r46 }
492{ .mfb; (p30) st8 [r14]=r47,8
493 (p0) nop.f 0x0
494 br.ctop.sptk .L_bn_mul_add_words_ctop};;
495.L_bn_mul_add_words_cend:
496
497{ .mii; nop.m 0x0
498.pred.rel "mutex",p53,p57
499(p53) add r8=r38,r0
500(p57) add r8=r38,r0,1 }
501{ .mfb; nop.m 0x0
502 nop.f 0x0
503 nop.b 0x0 };;
504{ .mii;
505(p63) add r8=1,r8
506 mov pr=r9,0x1ffff
507 mov ar.lc=r3 }
508{ .mfb; rum 1<<5 // clear um.mfh
509 nop.f 0x0
510 br.ret.sptk.many b0 };;
511.endp bn_mul_add_words#
512#endif
513
514#if 1
515//
516// void bn_sqr_words(BN_ULONG *rp, BN_ULONG *ap, int num)
517//
518.global bn_sqr_words#
519.proc bn_sqr_words#
520.align 64
521.skip 32 // makes the loop body aligned at 64-byte boundary
522bn_sqr_words:
523 .prologue
524 .fframe 0
525 .save ar.pfs,r2
526{ .mii; alloc r2=ar.pfs,3,0,0,0
527 sxt4 r34=r34 };;
528{ .mii; cmp.le p6,p0=r34,r0
529 mov r8=r0 } // return value
530{ .mfb; nop.f 0x0
531(p6) br.ret.spnt.many b0 };;
532
533 .save ar.lc,r3
534{ .mii; sub r10=r34,r0,1
535 mov r3=ar.lc
536 mov r9=pr };;
537
538 .body
539#if defined(_HPUX_SOURCE) && defined(_ILP32)
540{ .mii; addp4 r32=0,r32
541 addp4 r33=0,r33 };;
542#endif
543{ .mib;
544 mov pr.rot=1<<16
545 brp.loop.imp .L_bn_sqr_words_ctop,.L_bn_sqr_words_cend-16
546 }
547{ .mii; add r34=8,r32
548 mov ar.lc=r10
549 mov ar.ec=18 };;
550
551// 2*(n+17) on Itanium, (n+17) on "wider" IA-64 implementations. It's
552// possible to compress the epilogue (I'm getting tired to write this
553// comment over and over) and get down to 2*n+16 at the cost of
554// scalability. The decision will very likely be reconsidered after the
555// benchmark program is profiled. I.e. if perfomance gain on Itanium
556// will appear larger than loss on "wider" IA-64, then the loop should
557// be explicitely split and the epilogue compressed.
558.L_bn_sqr_words_ctop:
559{ .mfi; (p16) ldf8 f32=[r33],8
560 (p25) xmpy.lu f42=f41,f41
561 (p0) nop.i 0x0 }
562{ .mib; (p33) stf8 [r32]=f50,16
563 (p0) nop.i 0x0
564 (p0) nop.b 0x0 }
565{ .mfi; (p0) nop.m 0x0
566 (p25) xmpy.hu f52=f41,f41
567 (p0) nop.i 0x0 }
568{ .mib; (p33) stf8 [r34]=f60,16
569 (p0) nop.i 0x0
570 br.ctop.sptk .L_bn_sqr_words_ctop };;
571.L_bn_sqr_words_cend:
572
573{ .mii; nop.m 0x0
574 mov pr=r9,0x1ffff
575 mov ar.lc=r3 }
576{ .mfb; rum 1<<5 // clear um.mfh
577 nop.f 0x0
578 br.ret.sptk.many b0 };;
579.endp bn_sqr_words#
580#endif
581
582#if 1
583// Apparently we win nothing by implementing special bn_sqr_comba8.
584// Yes, it is possible to reduce the number of multiplications by
585// almost factor of two, but then the amount of additions would
586// increase by factor of two (as we would have to perform those
587// otherwise performed by xma ourselves). Normally we would trade
588// anyway as multiplications are way more expensive, but not this
589// time... Multiplication kernel is fully pipelined and as we drain
590// one 128-bit multiplication result per clock cycle multiplications
591// are effectively as inexpensive as additions. Special implementation
592// might become of interest for "wider" IA-64 implementation as you'll
593// be able to get through the multiplication phase faster (there won't
594// be any stall issues as discussed in the commentary section below and
595// you therefore will be able to employ all 4 FP units)... But these
596// Itanium days it's simply too hard to justify the effort so I just
597// drop down to bn_mul_comba8 code:-)
598//
599// void bn_sqr_comba8(BN_ULONG *r, BN_ULONG *a)
600//
601.global bn_sqr_comba8#
602.proc bn_sqr_comba8#
603.align 64
604bn_sqr_comba8:
605 .prologue
606 .fframe 0
607 .save ar.pfs,r2
608#if defined(_HPUX_SOURCE) && defined(_ILP32)
609{ .mii; alloc r2=ar.pfs,2,1,0,0
610 addp4 r33=0,r33
611 addp4 r32=0,r32 };;
612{ .mii;
613#else
614{ .mii; alloc r2=ar.pfs,2,1,0,0
615#endif
616 mov r34=r33
617 add r14=8,r33 };;
618 .body
619{ .mii; add r17=8,r34
620 add r15=16,r33
621 add r18=16,r34 }
622{ .mfb; add r16=24,r33
623 br .L_cheat_entry_point8 };;
624.endp bn_sqr_comba8#
625#endif
626
627#if 1
628// I've estimated this routine to run in ~120 ticks, but in reality
629// (i.e. according to ar.itc) it takes ~160 ticks. Are those extra
630// cycles consumed for instructions fetch? Or did I misinterpret some
631// clause in Itanium µ-architecture manual? Comments are welcomed and
632// highly appreciated.
633//
634// However! It should be noted that even 160 ticks is darn good result
635// as it's over 10 (yes, ten, spelled as t-e-n) times faster than the
636// C version (compiled with gcc with inline assembler). I really
637// kicked compiler's butt here, didn't I? Yeah! This brings us to the
638// following statement. It's damn shame that this routine isn't called
639// very often nowadays! According to the profiler most CPU time is
640// consumed by bn_mul_add_words called from BN_from_montgomery. In
641// order to estimate what we're missing, I've compared the performance
642// of this routine against "traditional" implementation, i.e. against
643// following routine:
644//
645// void bn_mul_comba8(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b)
646// { r[ 8]=bn_mul_words( &(r[0]),a,8,b[0]);
647// r[ 9]=bn_mul_add_words(&(r[1]),a,8,b[1]);
648// r[10]=bn_mul_add_words(&(r[2]),a,8,b[2]);
649// r[11]=bn_mul_add_words(&(r[3]),a,8,b[3]);
650// r[12]=bn_mul_add_words(&(r[4]),a,8,b[4]);
651// r[13]=bn_mul_add_words(&(r[5]),a,8,b[5]);
652// r[14]=bn_mul_add_words(&(r[6]),a,8,b[6]);
653// r[15]=bn_mul_add_words(&(r[7]),a,8,b[7]);
654// }
655//
656// The one below is over 8 times faster than the one above:-( Even
657// more reasons to "combafy" bn_mul_add_mont...
658//
659// And yes, this routine really made me wish there were an optimizing
660// assembler! It also feels like it deserves a dedication.
661//
662// To my wife for being there and to my kids...
663//
664// void bn_mul_comba8(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b)
665//
666#define carry1 r14
667#define carry2 r15
668#define carry3 r34
669.global bn_mul_comba8#
670.proc bn_mul_comba8#
671.align 64
672bn_mul_comba8:
673 .prologue
674 .fframe 0
675 .save ar.pfs,r2
676#if defined(_HPUX_SOURCE) && defined(_ILP32)
677{ .mii; alloc r2=ar.pfs,3,0,0,0
678 addp4 r33=0,r33
679 addp4 r34=0,r34 };;
680{ .mii; addp4 r32=0,r32
681#else
682{ .mii; alloc r2=ar.pfs,3,0,0,0
683#endif
684 add r14=8,r33
685 add r17=8,r34 }
686 .body
687{ .mii; add r15=16,r33
688 add r18=16,r34
689 add r16=24,r33 }
690.L_cheat_entry_point8:
691{ .mmi; add r19=24,r34
692
693 ldf8 f32=[r33],32 };;
694
695{ .mmi; ldf8 f120=[r34],32
696 ldf8 f121=[r17],32 }
697{ .mmi; ldf8 f122=[r18],32
698 ldf8 f123=[r19],32 };;
699{ .mmi; ldf8 f124=[r34]
700 ldf8 f125=[r17] }
701{ .mmi; ldf8 f126=[r18]
702 ldf8 f127=[r19] }
703
704{ .mmi; ldf8 f33=[r14],32
705 ldf8 f34=[r15],32 }
706{ .mmi; ldf8 f35=[r16],32;;
707 ldf8 f36=[r33] }
708{ .mmi; ldf8 f37=[r14]
709 ldf8 f38=[r15] }
710{ .mfi; ldf8 f39=[r16]
711// -------\ Entering multiplier's heaven /-------
712// ------------\ /------------
713// -----------------\ /-----------------
714// ----------------------\/----------------------
715 xma.hu f41=f32,f120,f0 }
716{ .mfi; xma.lu f40=f32,f120,f0 };; // (*)
717{ .mfi; xma.hu f51=f32,f121,f0 }
718{ .mfi; xma.lu f50=f32,f121,f0 };;
719{ .mfi; xma.hu f61=f32,f122,f0 }
720{ .mfi; xma.lu f60=f32,f122,f0 };;
721{ .mfi; xma.hu f71=f32,f123,f0 }
722{ .mfi; xma.lu f70=f32,f123,f0 };;
723{ .mfi; xma.hu f81=f32,f124,f0 }
724{ .mfi; xma.lu f80=f32,f124,f0 };;
725{ .mfi; xma.hu f91=f32,f125,f0 }
726{ .mfi; xma.lu f90=f32,f125,f0 };;
727{ .mfi; xma.hu f101=f32,f126,f0 }
728{ .mfi; xma.lu f100=f32,f126,f0 };;
729{ .mfi; xma.hu f111=f32,f127,f0 }
730{ .mfi; xma.lu f110=f32,f127,f0 };;//
731// (*) You can argue that splitting at every second bundle would
732// prevent "wider" IA-64 implementations from achieving the peak
733// performance. Well, not really... The catch is that if you
734// intend to keep 4 FP units busy by splitting at every fourth
735// bundle and thus perform these 16 multiplications in 4 ticks,
736// the first bundle *below* would stall because the result from
737// the first xma bundle *above* won't be available for another 3
738// ticks (if not more, being an optimist, I assume that "wider"
739// implementation will have same latency:-). This stall will hold
740// you back and the performance would be as if every second bundle
741// were split *anyway*...
742{ .mfi; getf.sig r16=f40
743 xma.hu f42=f33,f120,f41
744 add r33=8,r32 }
745{ .mfi; xma.lu f41=f33,f120,f41 };;
746{ .mfi; getf.sig r24=f50
747 xma.hu f52=f33,f121,f51 }
748{ .mfi; xma.lu f51=f33,f121,f51 };;
749{ .mfi; st8 [r32]=r16,16
750 xma.hu f62=f33,f122,f61 }
751{ .mfi; xma.lu f61=f33,f122,f61 };;
752{ .mfi; xma.hu f72=f33,f123,f71 }
753{ .mfi; xma.lu f71=f33,f123,f71 };;
754{ .mfi; xma.hu f82=f33,f124,f81 }
755{ .mfi; xma.lu f81=f33,f124,f81 };;
756{ .mfi; xma.hu f92=f33,f125,f91 }
757{ .mfi; xma.lu f91=f33,f125,f91 };;
758{ .mfi; xma.hu f102=f33,f126,f101 }
759{ .mfi; xma.lu f101=f33,f126,f101 };;
760{ .mfi; xma.hu f112=f33,f127,f111 }
761{ .mfi; xma.lu f111=f33,f127,f111 };;//
762//-------------------------------------------------//
763{ .mfi; getf.sig r25=f41
764 xma.hu f43=f34,f120,f42 }
765{ .mfi; xma.lu f42=f34,f120,f42 };;
766{ .mfi; getf.sig r16=f60
767 xma.hu f53=f34,f121,f52 }
768{ .mfi; xma.lu f52=f34,f121,f52 };;
769{ .mfi; getf.sig r17=f51
770 xma.hu f63=f34,f122,f62
771 add r25=r25,r24 }
772{ .mfi; xma.lu f62=f34,f122,f62
773 mov carry1=0 };;
774{ .mfi; cmp.ltu p6,p0=r25,r24
775 xma.hu f73=f34,f123,f72 }
776{ .mfi; xma.lu f72=f34,f123,f72 };;
777{ .mfi; st8 [r33]=r25,16
778 xma.hu f83=f34,f124,f82
779(p6) add carry1=1,carry1 }
780{ .mfi; xma.lu f82=f34,f124,f82 };;
781{ .mfi; xma.hu f93=f34,f125,f92 }
782{ .mfi; xma.lu f92=f34,f125,f92 };;
783{ .mfi; xma.hu f103=f34,f126,f102 }
784{ .mfi; xma.lu f102=f34,f126,f102 };;
785{ .mfi; xma.hu f113=f34,f127,f112 }
786{ .mfi; xma.lu f112=f34,f127,f112 };;//
787//-------------------------------------------------//
788{ .mfi; getf.sig r18=f42
789 xma.hu f44=f35,f120,f43
790 add r17=r17,r16 }
791{ .mfi; xma.lu f43=f35,f120,f43 };;
792{ .mfi; getf.sig r24=f70
793 xma.hu f54=f35,f121,f53 }
794{ .mfi; mov carry2=0
795 xma.lu f53=f35,f121,f53 };;
796{ .mfi; getf.sig r25=f61
797 xma.hu f64=f35,f122,f63
798 cmp.ltu p7,p0=r17,r16 }
799{ .mfi; add r18=r18,r17
800 xma.lu f63=f35,f122,f63 };;
801{ .mfi; getf.sig r26=f52
802 xma.hu f74=f35,f123,f73
803(p7) add carry2=1,carry2 }
804{ .mfi; cmp.ltu p7,p0=r18,r17
805 xma.lu f73=f35,f123,f73
806 add r18=r18,carry1 };;
807{ .mfi;
808 xma.hu f84=f35,f124,f83
809(p7) add carry2=1,carry2 }
810{ .mfi; cmp.ltu p7,p0=r18,carry1
811 xma.lu f83=f35,f124,f83 };;
812{ .mfi; st8 [r32]=r18,16
813 xma.hu f94=f35,f125,f93
814(p7) add carry2=1,carry2 }
815{ .mfi; xma.lu f93=f35,f125,f93 };;
816{ .mfi; xma.hu f104=f35,f126,f103 }
817{ .mfi; xma.lu f103=f35,f126,f103 };;
818{ .mfi; xma.hu f114=f35,f127,f113 }
819{ .mfi; mov carry1=0
820 xma.lu f113=f35,f127,f113
821 add r25=r25,r24 };;//
822//-------------------------------------------------//
823{ .mfi; getf.sig r27=f43
824 xma.hu f45=f36,f120,f44
825 cmp.ltu p6,p0=r25,r24 }
826{ .mfi; xma.lu f44=f36,f120,f44
827 add r26=r26,r25 };;
828{ .mfi; getf.sig r16=f80
829 xma.hu f55=f36,f121,f54
830(p6) add carry1=1,carry1 }
831{ .mfi; xma.lu f54=f36,f121,f54 };;
832{ .mfi; getf.sig r17=f71
833 xma.hu f65=f36,f122,f64
834 cmp.ltu p6,p0=r26,r25 }
835{ .mfi; xma.lu f64=f36,f122,f64
836 add r27=r27,r26 };;
837{ .mfi; getf.sig r18=f62
838 xma.hu f75=f36,f123,f74
839(p6) add carry1=1,carry1 }
840{ .mfi; cmp.ltu p6,p0=r27,r26
841 xma.lu f74=f36,f123,f74
842 add r27=r27,carry2 };;
843{ .mfi; getf.sig r19=f53
844 xma.hu f85=f36,f124,f84
845(p6) add carry1=1,carry1 }
846{ .mfi; xma.lu f84=f36,f124,f84
847 cmp.ltu p6,p0=r27,carry2 };;
848{ .mfi; st8 [r33]=r27,16
849 xma.hu f95=f36,f125,f94
850(p6) add carry1=1,carry1 }
851{ .mfi; xma.lu f94=f36,f125,f94 };;
852{ .mfi; xma.hu f105=f36,f126,f104 }
853{ .mfi; mov carry2=0
854 xma.lu f104=f36,f126,f104
855 add r17=r17,r16 };;
856{ .mfi; xma.hu f115=f36,f127,f114
857 cmp.ltu p7,p0=r17,r16 }
858{ .mfi; xma.lu f114=f36,f127,f114
859 add r18=r18,r17 };;//
860//-------------------------------------------------//
861{ .mfi; getf.sig r20=f44
862 xma.hu f46=f37,f120,f45
863(p7) add carry2=1,carry2 }
864{ .mfi; cmp.ltu p7,p0=r18,r17
865 xma.lu f45=f37,f120,f45
866 add r19=r19,r18 };;
867{ .mfi; getf.sig r24=f90
868 xma.hu f56=f37,f121,f55 }
869{ .mfi; xma.lu f55=f37,f121,f55 };;
870{ .mfi; getf.sig r25=f81
871 xma.hu f66=f37,f122,f65
872(p7) add carry2=1,carry2 }
873{ .mfi; cmp.ltu p7,p0=r19,r18
874 xma.lu f65=f37,f122,f65
875 add r20=r20,r19 };;
876{ .mfi; getf.sig r26=f72
877 xma.hu f76=f37,f123,f75
878(p7) add carry2=1,carry2 }
879{ .mfi; cmp.ltu p7,p0=r20,r19
880 xma.lu f75=f37,f123,f75
881 add r20=r20,carry1 };;
882{ .mfi; getf.sig r27=f63
883 xma.hu f86=f37,f124,f85
884(p7) add carry2=1,carry2 }
885{ .mfi; xma.lu f85=f37,f124,f85
886 cmp.ltu p7,p0=r20,carry1 };;
887{ .mfi; getf.sig r28=f54
888 xma.hu f96=f37,f125,f95
889(p7) add carry2=1,carry2 }
890{ .mfi; st8 [r32]=r20,16
891 xma.lu f95=f37,f125,f95 };;
892{ .mfi; xma.hu f106=f37,f126,f105 }
893{ .mfi; mov carry1=0
894 xma.lu f105=f37,f126,f105
895 add r25=r25,r24 };;
896{ .mfi; xma.hu f116=f37,f127,f115
897 cmp.ltu p6,p0=r25,r24 }
898{ .mfi; xma.lu f115=f37,f127,f115
899 add r26=r26,r25 };;//
900//-------------------------------------------------//
901{ .mfi; getf.sig r29=f45
902 xma.hu f47=f38,f120,f46
903(p6) add carry1=1,carry1 }
904{ .mfi; cmp.ltu p6,p0=r26,r25
905 xma.lu f46=f38,f120,f46
906 add r27=r27,r26 };;
907{ .mfi; getf.sig r16=f100
908 xma.hu f57=f38,f121,f56
909(p6) add carry1=1,carry1 }
910{ .mfi; cmp.ltu p6,p0=r27,r26
911 xma.lu f56=f38,f121,f56
912 add r28=r28,r27 };;
913{ .mfi; getf.sig r17=f91
914 xma.hu f67=f38,f122,f66
915(p6) add carry1=1,carry1 }
916{ .mfi; cmp.ltu p6,p0=r28,r27
917 xma.lu f66=f38,f122,f66
918 add r29=r29,r28 };;
919{ .mfi; getf.sig r18=f82
920 xma.hu f77=f38,f123,f76
921(p6) add carry1=1,carry1 }
922{ .mfi; cmp.ltu p6,p0=r29,r28
923 xma.lu f76=f38,f123,f76
924 add r29=r29,carry2 };;
925{ .mfi; getf.sig r19=f73
926 xma.hu f87=f38,f124,f86
927(p6) add carry1=1,carry1 }
928{ .mfi; xma.lu f86=f38,f124,f86
929 cmp.ltu p6,p0=r29,carry2 };;
930{ .mfi; getf.sig r20=f64
931 xma.hu f97=f38,f125,f96
932(p6) add carry1=1,carry1 }
933{ .mfi; st8 [r33]=r29,16
934 xma.lu f96=f38,f125,f96 };;
935{ .mfi; getf.sig r21=f55
936 xma.hu f107=f38,f126,f106 }
937{ .mfi; mov carry2=0
938 xma.lu f106=f38,f126,f106
939 add r17=r17,r16 };;
940{ .mfi; xma.hu f117=f38,f127,f116
941 cmp.ltu p7,p0=r17,r16 }
942{ .mfi; xma.lu f116=f38,f127,f116
943 add r18=r18,r17 };;//
944//-------------------------------------------------//
945{ .mfi; getf.sig r22=f46
946 xma.hu f48=f39,f120,f47
947(p7) add carry2=1,carry2 }
948{ .mfi; cmp.ltu p7,p0=r18,r17
949 xma.lu f47=f39,f120,f47
950 add r19=r19,r18 };;
951{ .mfi; getf.sig r24=f110
952 xma.hu f58=f39,f121,f57
953(p7) add carry2=1,carry2 }
954{ .mfi; cmp.ltu p7,p0=r19,r18
955 xma.lu f57=f39,f121,f57
956 add r20=r20,r19 };;
957{ .mfi; getf.sig r25=f101
958 xma.hu f68=f39,f122,f67
959(p7) add carry2=1,carry2 }
960{ .mfi; cmp.ltu p7,p0=r20,r19
961 xma.lu f67=f39,f122,f67
962 add r21=r21,r20 };;
963{ .mfi; getf.sig r26=f92
964 xma.hu f78=f39,f123,f77
965(p7) add carry2=1,carry2 }
966{ .mfi; cmp.ltu p7,p0=r21,r20
967 xma.lu f77=f39,f123,f77
968 add r22=r22,r21 };;
969{ .mfi; getf.sig r27=f83
970 xma.hu f88=f39,f124,f87
971(p7) add carry2=1,carry2 }
972{ .mfi; cmp.ltu p7,p0=r22,r21
973 xma.lu f87=f39,f124,f87
974 add r22=r22,carry1 };;
975{ .mfi; getf.sig r28=f74
976 xma.hu f98=f39,f125,f97
977(p7) add carry2=1,carry2 }
978{ .mfi; xma.lu f97=f39,f125,f97
979 cmp.ltu p7,p0=r22,carry1 };;
980{ .mfi; getf.sig r29=f65
981 xma.hu f108=f39,f126,f107
982(p7) add carry2=1,carry2 }
983{ .mfi; st8 [r32]=r22,16
984 xma.lu f107=f39,f126,f107 };;
985{ .mfi; getf.sig r30=f56
986 xma.hu f118=f39,f127,f117 }
987{ .mfi; xma.lu f117=f39,f127,f117 };;//
988//-------------------------------------------------//
989// Leaving muliplier's heaven... Quite a ride, huh?
990
991{ .mii; getf.sig r31=f47
992 add r25=r25,r24
993 mov carry1=0 };;
994{ .mii; getf.sig r16=f111
995 cmp.ltu p6,p0=r25,r24
996 add r26=r26,r25 };;
997{ .mfb; getf.sig r17=f102 }
998{ .mii;
999(p6) add carry1=1,carry1
1000 cmp.ltu p6,p0=r26,r25
1001 add r27=r27,r26 };;
1002{ .mfb; nop.m 0x0 }
1003{ .mii;
1004(p6) add carry1=1,carry1
1005 cmp.ltu p6,p0=r27,r26
1006 add r28=r28,r27 };;
1007{ .mii; getf.sig r18=f93
1008 add r17=r17,r16
1009 mov carry3=0 }
1010{ .mii;
1011(p6) add carry1=1,carry1
1012 cmp.ltu p6,p0=r28,r27
1013 add r29=r29,r28 };;
1014{ .mii; getf.sig r19=f84
1015 cmp.ltu p7,p0=r17,r16 }
1016{ .mii;
1017(p6) add carry1=1,carry1
1018 cmp.ltu p6,p0=r29,r28
1019 add r30=r30,r29 };;
1020{ .mii; getf.sig r20=f75
1021 add r18=r18,r17 }
1022{ .mii;
1023(p6) add carry1=1,carry1
1024 cmp.ltu p6,p0=r30,r29
1025 add r31=r31,r30 };;
1026{ .mfb; getf.sig r21=f66 }
1027{ .mii; (p7) add carry3=1,carry3
1028 cmp.ltu p7,p0=r18,r17
1029 add r19=r19,r18 }
1030{ .mfb; nop.m 0x0 }
1031{ .mii;
1032(p6) add carry1=1,carry1
1033 cmp.ltu p6,p0=r31,r30
1034 add r31=r31,carry2 };;
1035{ .mfb; getf.sig r22=f57 }
1036{ .mii; (p7) add carry3=1,carry3
1037 cmp.ltu p7,p0=r19,r18
1038 add r20=r20,r19 }
1039{ .mfb; nop.m 0x0 }
1040{ .mii;
1041(p6) add carry1=1,carry1
1042 cmp.ltu p6,p0=r31,carry2 };;
1043{ .mfb; getf.sig r23=f48 }
1044{ .mii; (p7) add carry3=1,carry3
1045 cmp.ltu p7,p0=r20,r19
1046 add r21=r21,r20 }
1047{ .mii;
1048(p6) add carry1=1,carry1 }
1049{ .mfb; st8 [r33]=r31,16 };;
1050
1051{ .mfb; getf.sig r24=f112 }
1052{ .mii; (p7) add carry3=1,carry3
1053 cmp.ltu p7,p0=r21,r20
1054 add r22=r22,r21 };;
1055{ .mfb; getf.sig r25=f103 }
1056{ .mii; (p7) add carry3=1,carry3
1057 cmp.ltu p7,p0=r22,r21
1058 add r23=r23,r22 };;
1059{ .mfb; getf.sig r26=f94 }
1060{ .mii; (p7) add carry3=1,carry3
1061 cmp.ltu p7,p0=r23,r22
1062 add r23=r23,carry1 };;
1063{ .mfb; getf.sig r27=f85 }
1064{ .mii; (p7) add carry3=1,carry3
1065 cmp.ltu p7,p8=r23,carry1};;
1066{ .mii; getf.sig r28=f76
1067 add r25=r25,r24
1068 mov carry1=0 }
1069{ .mii; st8 [r32]=r23,16
1070 (p7) add carry2=1,carry3
1071 (p8) add carry2=0,carry3 };;
1072
1073{ .mfb; nop.m 0x0 }
1074{ .mii; getf.sig r29=f67
1075 cmp.ltu p6,p0=r25,r24
1076 add r26=r26,r25 };;
1077{ .mfb; getf.sig r30=f58 }
1078{ .mii;
1079(p6) add carry1=1,carry1
1080 cmp.ltu p6,p0=r26,r25
1081 add r27=r27,r26 };;
1082{ .mfb; getf.sig r16=f113 }
1083{ .mii;
1084(p6) add carry1=1,carry1
1085 cmp.ltu p6,p0=r27,r26
1086 add r28=r28,r27 };;
1087{ .mfb; getf.sig r17=f104 }
1088{ .mii;
1089(p6) add carry1=1,carry1
1090 cmp.ltu p6,p0=r28,r27
1091 add r29=r29,r28 };;
1092{ .mfb; getf.sig r18=f95 }
1093{ .mii;
1094(p6) add carry1=1,carry1
1095 cmp.ltu p6,p0=r29,r28
1096 add r30=r30,r29 };;
1097{ .mii; getf.sig r19=f86
1098 add r17=r17,r16
1099 mov carry3=0 }
1100{ .mii;
1101(p6) add carry1=1,carry1
1102 cmp.ltu p6,p0=r30,r29
1103 add r30=r30,carry2 };;
1104{ .mii; getf.sig r20=f77
1105 cmp.ltu p7,p0=r17,r16
1106 add r18=r18,r17 }
1107{ .mii;
1108(p6) add carry1=1,carry1
1109 cmp.ltu p6,p0=r30,carry2 };;
1110{ .mfb; getf.sig r21=f68 }
1111{ .mii; st8 [r33]=r30,16
1112(p6) add carry1=1,carry1 };;
1113
1114{ .mfb; getf.sig r24=f114 }
1115{ .mii; (p7) add carry3=1,carry3
1116 cmp.ltu p7,p0=r18,r17
1117 add r19=r19,r18 };;
1118{ .mfb; getf.sig r25=f105 }
1119{ .mii; (p7) add carry3=1,carry3
1120 cmp.ltu p7,p0=r19,r18
1121 add r20=r20,r19 };;
1122{ .mfb; getf.sig r26=f96 }
1123{ .mii; (p7) add carry3=1,carry3
1124 cmp.ltu p7,p0=r20,r19
1125 add r21=r21,r20 };;
1126{ .mfb; getf.sig r27=f87 }
1127{ .mii; (p7) add carry3=1,carry3
1128 cmp.ltu p7,p0=r21,r20
1129 add r21=r21,carry1 };;
1130{ .mib; getf.sig r28=f78
1131 add r25=r25,r24 }
1132{ .mib; (p7) add carry3=1,carry3
1133 cmp.ltu p7,p8=r21,carry1};;
1134{ .mii; st8 [r32]=r21,16
1135 (p7) add carry2=1,carry3
1136 (p8) add carry2=0,carry3 }
1137
1138{ .mii; mov carry1=0
1139 cmp.ltu p6,p0=r25,r24
1140 add r26=r26,r25 };;
1141{ .mfb; getf.sig r16=f115 }
1142{ .mii;
1143(p6) add carry1=1,carry1
1144 cmp.ltu p6,p0=r26,r25
1145 add r27=r27,r26 };;
1146{ .mfb; getf.sig r17=f106 }
1147{ .mii;
1148(p6) add carry1=1,carry1
1149 cmp.ltu p6,p0=r27,r26
1150 add r28=r28,r27 };;
1151{ .mfb; getf.sig r18=f97 }
1152{ .mii;
1153(p6) add carry1=1,carry1
1154 cmp.ltu p6,p0=r28,r27
1155 add r28=r28,carry2 };;
1156{ .mib; getf.sig r19=f88
1157 add r17=r17,r16 }
1158{ .mib;
1159(p6) add carry1=1,carry1
1160 cmp.ltu p6,p0=r28,carry2 };;
1161{ .mii; st8 [r33]=r28,16
1162(p6) add carry1=1,carry1 }
1163
1164{ .mii; mov carry2=0
1165 cmp.ltu p7,p0=r17,r16
1166 add r18=r18,r17 };;
1167{ .mfb; getf.sig r24=f116 }
1168{ .mii; (p7) add carry2=1,carry2
1169 cmp.ltu p7,p0=r18,r17
1170 add r19=r19,r18 };;
1171{ .mfb; getf.sig r25=f107 }
1172{ .mii; (p7) add carry2=1,carry2
1173 cmp.ltu p7,p0=r19,r18
1174 add r19=r19,carry1 };;
1175{ .mfb; getf.sig r26=f98 }
1176{ .mii; (p7) add carry2=1,carry2
1177 cmp.ltu p7,p0=r19,carry1};;
1178{ .mii; st8 [r32]=r19,16
1179 (p7) add carry2=1,carry2 }
1180
1181{ .mfb; add r25=r25,r24 };;
1182
1183{ .mfb; getf.sig r16=f117 }
1184{ .mii; mov carry1=0
1185 cmp.ltu p6,p0=r25,r24
1186 add r26=r26,r25 };;
1187{ .mfb; getf.sig r17=f108 }
1188{ .mii;
1189(p6) add carry1=1,carry1
1190 cmp.ltu p6,p0=r26,r25
1191 add r26=r26,carry2 };;
1192{ .mfb; nop.m 0x0 }
1193{ .mii;
1194(p6) add carry1=1,carry1
1195 cmp.ltu p6,p0=r26,carry2 };;
1196{ .mii; st8 [r33]=r26,16
1197(p6) add carry1=1,carry1 }
1198
1199{ .mfb; add r17=r17,r16 };;
1200{ .mfb; getf.sig r24=f118 }
1201{ .mii; mov carry2=0
1202 cmp.ltu p7,p0=r17,r16
1203 add r17=r17,carry1 };;
1204{ .mii; (p7) add carry2=1,carry2
1205 cmp.ltu p7,p0=r17,carry1};;
1206{ .mii; st8 [r32]=r17
1207 (p7) add carry2=1,carry2 };;
1208{ .mfb; add r24=r24,carry2 };;
1209{ .mib; st8 [r33]=r24 }
1210
1211{ .mib; rum 1<<5 // clear um.mfh
1212 br.ret.sptk.many b0 };;
1213.endp bn_mul_comba8#
1214#undef carry3
1215#undef carry2
1216#undef carry1
1217#endif
1218
1219#if 1
1220// It's possible to make it faster (see comment to bn_sqr_comba8), but
1221// I reckon it doesn't worth the effort. Basically because the routine
1222// (actually both of them) practically never called... So I just play
1223// same trick as with bn_sqr_comba8.
1224//
1225// void bn_sqr_comba4(BN_ULONG *r, BN_ULONG *a)
1226//
1227.global bn_sqr_comba4#
1228.proc bn_sqr_comba4#
1229.align 64
1230bn_sqr_comba4:
1231 .prologue
1232 .fframe 0
1233 .save ar.pfs,r2
1234#if defined(_HPUX_SOURCE) && defined(_ILP32)
1235{ .mii; alloc r2=ar.pfs,2,1,0,0
1236 addp4 r32=0,r32
1237 addp4 r33=0,r33 };;
1238{ .mii;
1239#else
1240{ .mii; alloc r2=ar.pfs,2,1,0,0
1241#endif
1242 mov r34=r33
1243 add r14=8,r33 };;
1244 .body
1245{ .mii; add r17=8,r34
1246 add r15=16,r33
1247 add r18=16,r34 }
1248{ .mfb; add r16=24,r33
1249 br .L_cheat_entry_point4 };;
1250.endp bn_sqr_comba4#
1251#endif
1252
1253#if 1
1254// Runs in ~115 cycles and ~4.5 times faster than C. Well, whatever...
1255//
1256// void bn_mul_comba4(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b)
1257//
1258#define carry1 r14
1259#define carry2 r15
1260.global bn_mul_comba4#
1261.proc bn_mul_comba4#
1262.align 64
1263bn_mul_comba4:
1264 .prologue
1265 .fframe 0
1266 .save ar.pfs,r2
1267#if defined(_HPUX_SOURCE) && defined(_ILP32)
1268{ .mii; alloc r2=ar.pfs,3,0,0,0
1269 addp4 r33=0,r33
1270 addp4 r34=0,r34 };;
1271{ .mii; addp4 r32=0,r32
1272#else
1273{ .mii; alloc r2=ar.pfs,3,0,0,0
1274#endif
1275 add r14=8,r33
1276 add r17=8,r34 }
1277 .body
1278{ .mii; add r15=16,r33
1279 add r18=16,r34
1280 add r16=24,r33 };;
1281.L_cheat_entry_point4:
1282{ .mmi; add r19=24,r34
1283
1284 ldf8 f32=[r33] }
1285
1286{ .mmi; ldf8 f120=[r34]
1287 ldf8 f121=[r17] };;
1288{ .mmi; ldf8 f122=[r18]
1289 ldf8 f123=[r19] }
1290
1291{ .mmi; ldf8 f33=[r14]
1292 ldf8 f34=[r15] }
1293{ .mfi; ldf8 f35=[r16]
1294
1295 xma.hu f41=f32,f120,f0 }
1296{ .mfi; xma.lu f40=f32,f120,f0 };;
1297{ .mfi; xma.hu f51=f32,f121,f0 }
1298{ .mfi; xma.lu f50=f32,f121,f0 };;
1299{ .mfi; xma.hu f61=f32,f122,f0 }
1300{ .mfi; xma.lu f60=f32,f122,f0 };;
1301{ .mfi; xma.hu f71=f32,f123,f0 }
1302{ .mfi; xma.lu f70=f32,f123,f0 };;//
1303// Major stall takes place here, and 3 more places below. Result from
1304// first xma is not available for another 3 ticks.
1305{ .mfi; getf.sig r16=f40
1306 xma.hu f42=f33,f120,f41
1307 add r33=8,r32 }
1308{ .mfi; xma.lu f41=f33,f120,f41 };;
1309{ .mfi; getf.sig r24=f50
1310 xma.hu f52=f33,f121,f51 }
1311{ .mfi; xma.lu f51=f33,f121,f51 };;
1312{ .mfi; st8 [r32]=r16,16
1313 xma.hu f62=f33,f122,f61 }
1314{ .mfi; xma.lu f61=f33,f122,f61 };;
1315{ .mfi; xma.hu f72=f33,f123,f71 }
1316{ .mfi; xma.lu f71=f33,f123,f71 };;//
1317//-------------------------------------------------//
1318{ .mfi; getf.sig r25=f41
1319 xma.hu f43=f34,f120,f42 }
1320{ .mfi; xma.lu f42=f34,f120,f42 };;
1321{ .mfi; getf.sig r16=f60
1322 xma.hu f53=f34,f121,f52 }
1323{ .mfi; xma.lu f52=f34,f121,f52 };;
1324{ .mfi; getf.sig r17=f51
1325 xma.hu f63=f34,f122,f62
1326 add r25=r25,r24 }
1327{ .mfi; mov carry1=0
1328 xma.lu f62=f34,f122,f62 };;
1329{ .mfi; st8 [r33]=r25,16
1330 xma.hu f73=f34,f123,f72
1331 cmp.ltu p6,p0=r25,r24 }
1332{ .mfi; xma.lu f72=f34,f123,f72 };;//
1333//-------------------------------------------------//
1334{ .mfi; getf.sig r18=f42
1335 xma.hu f44=f35,f120,f43
1336(p6) add carry1=1,carry1 }
1337{ .mfi; add r17=r17,r16
1338 xma.lu f43=f35,f120,f43
1339 mov carry2=0 };;
1340{ .mfi; getf.sig r24=f70
1341 xma.hu f54=f35,f121,f53
1342 cmp.ltu p7,p0=r17,r16 }
1343{ .mfi; xma.lu f53=f35,f121,f53 };;
1344{ .mfi; getf.sig r25=f61
1345 xma.hu f64=f35,f122,f63
1346 add r18=r18,r17 }
1347{ .mfi; xma.lu f63=f35,f122,f63
1348(p7) add carry2=1,carry2 };;
1349{ .mfi; getf.sig r26=f52
1350 xma.hu f74=f35,f123,f73
1351 cmp.ltu p7,p0=r18,r17 }
1352{ .mfi; xma.lu f73=f35,f123,f73
1353 add r18=r18,carry1 };;
1354//-------------------------------------------------//
1355{ .mii; st8 [r32]=r18,16
1356(p7) add carry2=1,carry2
1357 cmp.ltu p7,p0=r18,carry1 };;
1358
1359{ .mfi; getf.sig r27=f43 // last major stall
1360(p7) add carry2=1,carry2 };;
1361{ .mii; getf.sig r16=f71
1362 add r25=r25,r24
1363 mov carry1=0 };;
1364{ .mii; getf.sig r17=f62
1365 cmp.ltu p6,p0=r25,r24
1366 add r26=r26,r25 };;
1367{ .mii;
1368(p6) add carry1=1,carry1
1369 cmp.ltu p6,p0=r26,r25
1370 add r27=r27,r26 };;
1371{ .mii;
1372(p6) add carry1=1,carry1
1373 cmp.ltu p6,p0=r27,r26
1374 add r27=r27,carry2 };;
1375{ .mii; getf.sig r18=f53
1376(p6) add carry1=1,carry1
1377 cmp.ltu p6,p0=r27,carry2 };;
1378{ .mfi; st8 [r33]=r27,16
1379(p6) add carry1=1,carry1 }
1380
1381{ .mii; getf.sig r19=f44
1382 add r17=r17,r16
1383 mov carry2=0 };;
1384{ .mii; getf.sig r24=f72
1385 cmp.ltu p7,p0=r17,r16
1386 add r18=r18,r17 };;
1387{ .mii; (p7) add carry2=1,carry2
1388 cmp.ltu p7,p0=r18,r17
1389 add r19=r19,r18 };;
1390{ .mii; (p7) add carry2=1,carry2
1391 cmp.ltu p7,p0=r19,r18
1392 add r19=r19,carry1 };;
1393{ .mii; getf.sig r25=f63
1394 (p7) add carry2=1,carry2
1395 cmp.ltu p7,p0=r19,carry1};;
1396{ .mii; st8 [r32]=r19,16
1397 (p7) add carry2=1,carry2 }
1398
1399{ .mii; getf.sig r26=f54
1400 add r25=r25,r24
1401 mov carry1=0 };;
1402{ .mii; getf.sig r16=f73
1403 cmp.ltu p6,p0=r25,r24
1404 add r26=r26,r25 };;
1405{ .mii;
1406(p6) add carry1=1,carry1
1407 cmp.ltu p6,p0=r26,r25
1408 add r26=r26,carry2 };;
1409{ .mii; getf.sig r17=f64
1410(p6) add carry1=1,carry1
1411 cmp.ltu p6,p0=r26,carry2 };;
1412{ .mii; st8 [r33]=r26,16
1413(p6) add carry1=1,carry1 }
1414
1415{ .mii; getf.sig r24=f74
1416 add r17=r17,r16
1417 mov carry2=0 };;
1418{ .mii; cmp.ltu p7,p0=r17,r16
1419 add r17=r17,carry1 };;
1420
1421{ .mii; (p7) add carry2=1,carry2
1422 cmp.ltu p7,p0=r17,carry1};;
1423{ .mii; st8 [r32]=r17,16
1424 (p7) add carry2=1,carry2 };;
1425
1426{ .mii; add r24=r24,carry2 };;
1427{ .mii; st8 [r33]=r24 }
1428
1429{ .mib; rum 1<<5 // clear um.mfh
1430 br.ret.sptk.many b0 };;
1431.endp bn_mul_comba4#
1432#undef carry2
1433#undef carry1
1434#endif
1435
1436#if 1
1437//
1438// BN_ULONG bn_div_words(BN_ULONG h, BN_ULONG l, BN_ULONG d)
1439//
1440// In the nutshell it's a port of my MIPS III/IV implementation.
1441//
1442#define AT r14
1443#define H r16
1444#define HH r20
1445#define L r17
1446#define D r18
1447#define DH r22
1448#define I r21
1449
1450#if 0
1451// Some preprocessors (most notably HP-UX) apper to be allergic to
1452// macros enclosed to parenthesis as these three will be.
1453#define cont p16
1454#define break p0 // p20
1455#define equ p24
1456#else
1457cont=p16
1458break=p0
1459equ=p24
1460#endif
1461
1462.global abort#
1463.global bn_div_words#
1464.proc bn_div_words#
1465.align 64
1466bn_div_words:
1467 .prologue
1468 .fframe 0
1469 .save ar.pfs,r2
1470 .save b0,r3
1471{ .mii; alloc r2=ar.pfs,3,5,0,8
1472 mov r3=b0
1473 mov r10=pr };;
1474{ .mmb; cmp.eq p6,p0=r34,r0
1475 mov r8=-1
1476(p6) br.ret.spnt.many b0 };;
1477
1478 .body
1479{ .mii; mov H=r32 // save h
1480 mov ar.ec=0 // don't rotate at exit
1481 mov pr.rot=0 }
1482{ .mii; mov L=r33 // save l
1483 mov r36=r0 };;
1484
1485.L_divw_shift: // -vv- note signed comparison
1486{ .mfi; (p0) cmp.lt p16,p0=r0,r34 // d
1487 (p0) shladd r33=r34,1,r0 }
1488{ .mfb; (p0) add r35=1,r36
1489 (p0) nop.f 0x0
1490(p16) br.wtop.dpnt .L_divw_shift };;
1491
1492{ .mii; mov D=r34
1493 shr.u DH=r34,32
1494 sub r35=64,r36 };;
1495{ .mii; setf.sig f7=DH
1496 shr.u AT=H,r35
1497 mov I=r36 };;
1498{ .mib; cmp.ne p6,p0=r0,AT
1499 shl H=H,r36
1500(p6) br.call.spnt.clr b0=abort };; // overflow, die...
1501
1502{ .mfi; fcvt.xuf.s1 f7=f7
1503 shr.u AT=L,r35 };;
1504{ .mii; shl L=L,r36
1505 or H=H,AT };;
1506
1507{ .mii; nop.m 0x0
1508 cmp.leu p6,p0=D,H;;
1509(p6) sub H=H,D }
1510
1511{ .mlx; setf.sig f14=D
1512 movl AT=0xffffffff };;
1513///////////////////////////////////////////////////////////
1514{ .mii; setf.sig f6=H
1515 shr.u HH=H,32;;
1516 cmp.eq p6,p7=HH,DH };;
1517{ .mfb;
1518(p6) setf.sig f8=AT
1519(p7) fcvt.xuf.s1 f6=f6
1520(p7) br.call.sptk b6=.L_udiv64_32_b6 };;
1521
1522{ .mfi; getf.sig r33=f8 // q
1523 xmpy.lu f9=f8,f14 }
1524{ .mfi; xmpy.hu f10=f8,f14
1525 shrp H=H,L,32 };;
1526
1527{ .mmi; getf.sig r35=f9 // tl
1528 getf.sig r31=f10 };; // th
1529
1530.L_divw_1st_iter:
1531{ .mii; (p0) add r32=-1,r33
1532 (p0) cmp.eq equ,cont=HH,r31 };;
1533{ .mii; (p0) cmp.ltu p8,p0=r35,D
1534 (p0) sub r34=r35,D
1535 (equ) cmp.leu break,cont=r35,H };;
1536{ .mib; (cont) cmp.leu cont,break=HH,r31
1537 (p8) add r31=-1,r31
1538(cont) br.wtop.spnt .L_divw_1st_iter };;
1539///////////////////////////////////////////////////////////
1540{ .mii; sub H=H,r35
1541 shl r8=r33,32
1542 shl L=L,32 };;
1543///////////////////////////////////////////////////////////
1544{ .mii; setf.sig f6=H
1545 shr.u HH=H,32;;
1546 cmp.eq p6,p7=HH,DH };;
1547{ .mfb;
1548(p6) setf.sig f8=AT
1549(p7) fcvt.xuf.s1 f6=f6
1550(p7) br.call.sptk b6=.L_udiv64_32_b6 };;
1551
1552{ .mfi; getf.sig r33=f8 // q
1553 xmpy.lu f9=f8,f14 }
1554{ .mfi; xmpy.hu f10=f8,f14
1555 shrp H=H,L,32 };;
1556
1557{ .mmi; getf.sig r35=f9 // tl
1558 getf.sig r31=f10 };; // th
1559
1560.L_divw_2nd_iter:
1561{ .mii; (p0) add r32=-1,r33
1562 (p0) cmp.eq equ,cont=HH,r31 };;
1563{ .mii; (p0) cmp.ltu p8,p0=r35,D
1564 (p0) sub r34=r35,D
1565 (equ) cmp.leu break,cont=r35,H };;
1566{ .mib; (cont) cmp.leu cont,break=HH,r31
1567 (p8) add r31=-1,r31
1568(cont) br.wtop.spnt .L_divw_2nd_iter };;
1569///////////////////////////////////////////////////////////
1570{ .mii; sub H=H,r35
1571 or r8=r8,r33
1572 mov ar.pfs=r2 };;
1573{ .mii; shr.u r9=H,I // remainder if anybody wants it
1574 mov pr=r10,0x1ffff }
1575{ .mfb; br.ret.sptk.many b0 };;
1576
1577// Unsigned 64 by 32 (well, by 64 for the moment) bit integer division
1578// procedure.
1579//
1580// inputs: f6 = (double)a, f7 = (double)b
1581// output: f8 = (int)(a/b)
1582// clobbered: f8,f9,f10,f11,pred
1583pred=p15
1584// This procedure is essentially Intel code and therefore is
1585// copyrighted to Intel Corporation (I suppose...). It's sligtly
1586// modified for specific needs.
1587.align 32
1588.skip 16
1589.L_udiv64_32_b6:
1590 frcpa.s1 f8,pred=f6,f7;; // [0] y0 = 1 / b
1591
1592(pred) fnma.s1 f9=f7,f8,f1 // [5] e0 = 1 - b * y0
1593(pred) fmpy.s1 f10=f6,f8;; // [5] q0 = a * y0
1594(pred) fmpy.s1 f11=f9,f9 // [10] e1 = e0 * e0
1595(pred) fma.s1 f10=f9,f10,f10;; // [10] q1 = q0 + e0 * q0
1596(pred) fma.s1 f8=f9,f8,f8 //;; // [15] y1 = y0 + e0 * y0
1597(pred) fma.s1 f9=f11,f10,f10;; // [15] q2 = q1 + e1 * q1
1598(pred) fma.s1 f8=f11,f8,f8 //;; // [20] y2 = y1 + e1 * y1
1599(pred) fnma.s1 f10=f7,f9,f6;; // [20] r2 = a - b * q2
1600(pred) fma.s1 f8=f10,f8,f9;; // [25] q3 = q2 + r2 * y2
1601
1602 fcvt.fxu.trunc.s1 f8=f8 // [30] q = trunc(q3)
1603 br.ret.sptk.many b6;;
1604.endp bn_div_words#
1605#endif
diff --git a/src/lib/libcrypto/bn/asm/pa-risc2.s b/src/lib/libcrypto/bn/asm/pa-risc2.s
deleted file mode 100644
index f3b16290eb..0000000000
--- a/src/lib/libcrypto/bn/asm/pa-risc2.s
+++ /dev/null
@@ -1,1618 +0,0 @@
1;
2; PA-RISC 2.0 implementation of bn_asm code, based on the
3; 64-bit version of the code. This code is effectively the
4; same as the 64-bit version except the register model is
5; slightly different given all values must be 32-bit between
6; function calls. Thus the 64-bit return values are returned
7; in %ret0 and %ret1 vs just %ret0 as is done in 64-bit
8;
9;
10; This code is approximately 2x faster than the C version
11; for RSA/DSA.
12;
13; See http://devresource.hp.com/ for more details on the PA-RISC
14; architecture. Also see the book "PA-RISC 2.0 Architecture"
15; by Gerry Kane for information on the instruction set architecture.
16;
17; Code written by Chris Ruemmler (with some help from the HP C
18; compiler).
19;
20; The code compiles with HP's assembler
21;
22
23 .level 2.0N
24 .space $TEXT$
25 .subspa $CODE$,QUAD=0,ALIGN=8,ACCESS=0x2c,CODE_ONLY
26
27;
28; Global Register definitions used for the routines.
29;
30; Some information about HP's runtime architecture for 32-bits.
31;
32; "Caller save" means the calling function must save the register
33; if it wants the register to be preserved.
34; "Callee save" means if a function uses the register, it must save
35; the value before using it.
36;
37; For the floating point registers
38;
39; "caller save" registers: fr4-fr11, fr22-fr31
40; "callee save" registers: fr12-fr21
41; "special" registers: fr0-fr3 (status and exception registers)
42;
43; For the integer registers
44; value zero : r0
45; "caller save" registers: r1,r19-r26
46; "callee save" registers: r3-r18
47; return register : r2 (rp)
48; return values ; r28,r29 (ret0,ret1)
49; Stack pointer ; r30 (sp)
50; millicode return ptr ; r31 (also a caller save register)
51
52
53;
54; Arguments to the routines
55;
56r_ptr .reg %r26
57a_ptr .reg %r25
58b_ptr .reg %r24
59num .reg %r24
60n .reg %r23
61
62;
63; Note that the "w" argument for bn_mul_add_words and bn_mul_words
64; is passed on the stack at a delta of -56 from the top of stack
65; as the routine is entered.
66;
67
68;
69; Globals used in some routines
70;
71
72top_overflow .reg %r23
73high_mask .reg %r22 ; value 0xffffffff80000000L
74
75
76;------------------------------------------------------------------------------
77;
78; bn_mul_add_words
79;
80;BN_ULONG bn_mul_add_words(BN_ULONG *r_ptr, BN_ULONG *a_ptr,
81; int num, BN_ULONG w)
82;
83; arg0 = r_ptr
84; arg1 = a_ptr
85; arg3 = num
86; -56(sp) = w
87;
88; Local register definitions
89;
90
91fm1 .reg %fr22
92fm .reg %fr23
93ht_temp .reg %fr24
94ht_temp_1 .reg %fr25
95lt_temp .reg %fr26
96lt_temp_1 .reg %fr27
97fm1_1 .reg %fr28
98fm_1 .reg %fr29
99
100fw_h .reg %fr7L
101fw_l .reg %fr7R
102fw .reg %fr7
103
104fht_0 .reg %fr8L
105flt_0 .reg %fr8R
106t_float_0 .reg %fr8
107
108fht_1 .reg %fr9L
109flt_1 .reg %fr9R
110t_float_1 .reg %fr9
111
112tmp_0 .reg %r31
113tmp_1 .reg %r21
114m_0 .reg %r20
115m_1 .reg %r19
116ht_0 .reg %r1
117ht_1 .reg %r3
118lt_0 .reg %r4
119lt_1 .reg %r5
120m1_0 .reg %r6
121m1_1 .reg %r7
122rp_val .reg %r8
123rp_val_1 .reg %r9
124
125bn_mul_add_words
126 .export bn_mul_add_words,entry,NO_RELOCATION,LONG_RETURN
127 .proc
128 .callinfo frame=128
129 .entry
130 .align 64
131
132 STD %r3,0(%sp) ; save r3
133 STD %r4,8(%sp) ; save r4
134 NOP ; Needed to make the loop 16-byte aligned
135 NOP ; needed to make the loop 16-byte aligned
136
137 STD %r5,16(%sp) ; save r5
138 NOP
139 STD %r6,24(%sp) ; save r6
140 STD %r7,32(%sp) ; save r7
141
142 STD %r8,40(%sp) ; save r8
143 STD %r9,48(%sp) ; save r9
144 COPY %r0,%ret1 ; return 0 by default
145 DEPDI,Z 1,31,1,top_overflow ; top_overflow = 1 << 32
146
147 CMPIB,>= 0,num,bn_mul_add_words_exit ; if (num <= 0) then exit
148 LDO 128(%sp),%sp ; bump stack
149
150 ;
151 ; The loop is unrolled twice, so if there is only 1 number
152 ; then go straight to the cleanup code.
153 ;
154 CMPIB,= 1,num,bn_mul_add_words_single_top
155 FLDD -184(%sp),fw ; (-56-128) load up w into fw (fw_h/fw_l)
156
157 ;
158 ; This loop is unrolled 2 times (64-byte aligned as well)
159 ;
160 ; PA-RISC 2.0 chips have two fully pipelined multipliers, thus
161 ; two 32-bit mutiplies can be issued per cycle.
162 ;
163bn_mul_add_words_unroll2
164
165 FLDD 0(a_ptr),t_float_0 ; load up 64-bit value (fr8L) ht(L)/lt(R)
166 FLDD 8(a_ptr),t_float_1 ; load up 64-bit value (fr8L) ht(L)/lt(R)
167 LDD 0(r_ptr),rp_val ; rp[0]
168 LDD 8(r_ptr),rp_val_1 ; rp[1]
169
170 XMPYU fht_0,fw_l,fm1 ; m1[0] = fht_0*fw_l
171 XMPYU fht_1,fw_l,fm1_1 ; m1[1] = fht_1*fw_l
172 FSTD fm1,-16(%sp) ; -16(sp) = m1[0]
173 FSTD fm1_1,-48(%sp) ; -48(sp) = m1[1]
174
175 XMPYU flt_0,fw_h,fm ; m[0] = flt_0*fw_h
176 XMPYU flt_1,fw_h,fm_1 ; m[1] = flt_1*fw_h
177 FSTD fm,-8(%sp) ; -8(sp) = m[0]
178 FSTD fm_1,-40(%sp) ; -40(sp) = m[1]
179
180 XMPYU fht_0,fw_h,ht_temp ; ht_temp = fht_0*fw_h
181 XMPYU fht_1,fw_h,ht_temp_1 ; ht_temp_1 = fht_1*fw_h
182 FSTD ht_temp,-24(%sp) ; -24(sp) = ht_temp
183 FSTD ht_temp_1,-56(%sp) ; -56(sp) = ht_temp_1
184
185 XMPYU flt_0,fw_l,lt_temp ; lt_temp = lt*fw_l
186 XMPYU flt_1,fw_l,lt_temp_1 ; lt_temp = lt*fw_l
187 FSTD lt_temp,-32(%sp) ; -32(sp) = lt_temp
188 FSTD lt_temp_1,-64(%sp) ; -64(sp) = lt_temp_1
189
190 LDD -8(%sp),m_0 ; m[0]
191 LDD -40(%sp),m_1 ; m[1]
192 LDD -16(%sp),m1_0 ; m1[0]
193 LDD -48(%sp),m1_1 ; m1[1]
194
195 LDD -24(%sp),ht_0 ; ht[0]
196 LDD -56(%sp),ht_1 ; ht[1]
197 ADD,L m1_0,m_0,tmp_0 ; tmp_0 = m[0] + m1[0];
198 ADD,L m1_1,m_1,tmp_1 ; tmp_1 = m[1] + m1[1];
199
200 LDD -32(%sp),lt_0
201 LDD -64(%sp),lt_1
202 CMPCLR,*>>= tmp_0,m1_0, %r0 ; if (m[0] < m1[0])
203 ADD,L ht_0,top_overflow,ht_0 ; ht[0] += (1<<32)
204
205 CMPCLR,*>>= tmp_1,m1_1,%r0 ; if (m[1] < m1[1])
206 ADD,L ht_1,top_overflow,ht_1 ; ht[1] += (1<<32)
207 EXTRD,U tmp_0,31,32,m_0 ; m[0]>>32
208 DEPD,Z tmp_0,31,32,m1_0 ; m1[0] = m[0]<<32
209
210 EXTRD,U tmp_1,31,32,m_1 ; m[1]>>32
211 DEPD,Z tmp_1,31,32,m1_1 ; m1[1] = m[1]<<32
212 ADD,L ht_0,m_0,ht_0 ; ht[0]+= (m[0]>>32)
213 ADD,L ht_1,m_1,ht_1 ; ht[1]+= (m[1]>>32)
214
215 ADD lt_0,m1_0,lt_0 ; lt[0] = lt[0]+m1[0];
216 ADD,DC ht_0,%r0,ht_0 ; ht[0]++
217 ADD lt_1,m1_1,lt_1 ; lt[1] = lt[1]+m1[1];
218 ADD,DC ht_1,%r0,ht_1 ; ht[1]++
219
220 ADD %ret1,lt_0,lt_0 ; lt[0] = lt[0] + c;
221 ADD,DC ht_0,%r0,ht_0 ; ht[0]++
222 ADD lt_0,rp_val,lt_0 ; lt[0] = lt[0]+rp[0]
223 ADD,DC ht_0,%r0,ht_0 ; ht[0]++
224
225 LDO -2(num),num ; num = num - 2;
226 ADD ht_0,lt_1,lt_1 ; lt[1] = lt[1] + ht_0 (c);
227 ADD,DC ht_1,%r0,ht_1 ; ht[1]++
228 STD lt_0,0(r_ptr) ; rp[0] = lt[0]
229
230 ADD lt_1,rp_val_1,lt_1 ; lt[1] = lt[1]+rp[1]
231 ADD,DC ht_1,%r0,%ret1 ; ht[1]++
232 LDO 16(a_ptr),a_ptr ; a_ptr += 2
233
234 STD lt_1,8(r_ptr) ; rp[1] = lt[1]
235 CMPIB,<= 2,num,bn_mul_add_words_unroll2 ; go again if more to do
236 LDO 16(r_ptr),r_ptr ; r_ptr += 2
237
238 CMPIB,=,N 0,num,bn_mul_add_words_exit ; are we done, or cleanup last one
239
240 ;
241 ; Top of loop aligned on 64-byte boundary
242 ;
243bn_mul_add_words_single_top
244 FLDD 0(a_ptr),t_float_0 ; load up 64-bit value (fr8L) ht(L)/lt(R)
245 LDD 0(r_ptr),rp_val ; rp[0]
246 LDO 8(a_ptr),a_ptr ; a_ptr++
247 XMPYU fht_0,fw_l,fm1 ; m1 = ht*fw_l
248 FSTD fm1,-16(%sp) ; -16(sp) = m1
249 XMPYU flt_0,fw_h,fm ; m = lt*fw_h
250 FSTD fm,-8(%sp) ; -8(sp) = m
251 XMPYU fht_0,fw_h,ht_temp ; ht_temp = ht*fw_h
252 FSTD ht_temp,-24(%sp) ; -24(sp) = ht
253 XMPYU flt_0,fw_l,lt_temp ; lt_temp = lt*fw_l
254 FSTD lt_temp,-32(%sp) ; -32(sp) = lt
255
256 LDD -8(%sp),m_0
257 LDD -16(%sp),m1_0 ; m1 = temp1
258 ADD,L m_0,m1_0,tmp_0 ; tmp_0 = m + m1;
259 LDD -24(%sp),ht_0
260 LDD -32(%sp),lt_0
261
262 CMPCLR,*>>= tmp_0,m1_0,%r0 ; if (m < m1)
263 ADD,L ht_0,top_overflow,ht_0 ; ht += (1<<32)
264
265 EXTRD,U tmp_0,31,32,m_0 ; m>>32
266 DEPD,Z tmp_0,31,32,m1_0 ; m1 = m<<32
267
268 ADD,L ht_0,m_0,ht_0 ; ht+= (m>>32)
269 ADD lt_0,m1_0,tmp_0 ; tmp_0 = lt+m1;
270 ADD,DC ht_0,%r0,ht_0 ; ht++
271 ADD %ret1,tmp_0,lt_0 ; lt = lt + c;
272 ADD,DC ht_0,%r0,ht_0 ; ht++
273 ADD lt_0,rp_val,lt_0 ; lt = lt+rp[0]
274 ADD,DC ht_0,%r0,%ret1 ; ht++
275 STD lt_0,0(r_ptr) ; rp[0] = lt
276
277bn_mul_add_words_exit
278 .EXIT
279
280 EXTRD,U %ret1,31,32,%ret0 ; for 32-bit, return in ret0/ret1
281 LDD -80(%sp),%r9 ; restore r9
282 LDD -88(%sp),%r8 ; restore r8
283 LDD -96(%sp),%r7 ; restore r7
284 LDD -104(%sp),%r6 ; restore r6
285 LDD -112(%sp),%r5 ; restore r5
286 LDD -120(%sp),%r4 ; restore r4
287 BVE (%rp)
288 LDD,MB -128(%sp),%r3 ; restore r3
289 .PROCEND ;in=23,24,25,26,29;out=28;
290
291;----------------------------------------------------------------------------
292;
293;BN_ULONG bn_mul_words(BN_ULONG *rp, BN_ULONG *ap, int num, BN_ULONG w)
294;
295; arg0 = rp
296; arg1 = ap
297; arg3 = num
298; w on stack at -56(sp)
299
300bn_mul_words
301 .proc
302 .callinfo frame=128
303 .entry
304 .EXPORT bn_mul_words,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
305 .align 64
306
307 STD %r3,0(%sp) ; save r3
308 STD %r4,8(%sp) ; save r4
309 NOP
310 STD %r5,16(%sp) ; save r5
311
312 STD %r6,24(%sp) ; save r6
313 STD %r7,32(%sp) ; save r7
314 COPY %r0,%ret1 ; return 0 by default
315 DEPDI,Z 1,31,1,top_overflow ; top_overflow = 1 << 32
316
317 CMPIB,>= 0,num,bn_mul_words_exit
318 LDO 128(%sp),%sp ; bump stack
319
320 ;
321 ; See if only 1 word to do, thus just do cleanup
322 ;
323 CMPIB,= 1,num,bn_mul_words_single_top
324 FLDD -184(%sp),fw ; (-56-128) load up w into fw (fw_h/fw_l)
325
326 ;
327 ; This loop is unrolled 2 times (64-byte aligned as well)
328 ;
329 ; PA-RISC 2.0 chips have two fully pipelined multipliers, thus
330 ; two 32-bit mutiplies can be issued per cycle.
331 ;
332bn_mul_words_unroll2
333
334 FLDD 0(a_ptr),t_float_0 ; load up 64-bit value (fr8L) ht(L)/lt(R)
335 FLDD 8(a_ptr),t_float_1 ; load up 64-bit value (fr8L) ht(L)/lt(R)
336 XMPYU fht_0,fw_l,fm1 ; m1[0] = fht_0*fw_l
337 XMPYU fht_1,fw_l,fm1_1 ; m1[1] = ht*fw_l
338
339 FSTD fm1,-16(%sp) ; -16(sp) = m1
340 FSTD fm1_1,-48(%sp) ; -48(sp) = m1
341 XMPYU flt_0,fw_h,fm ; m = lt*fw_h
342 XMPYU flt_1,fw_h,fm_1 ; m = lt*fw_h
343
344 FSTD fm,-8(%sp) ; -8(sp) = m
345 FSTD fm_1,-40(%sp) ; -40(sp) = m
346 XMPYU fht_0,fw_h,ht_temp ; ht_temp = fht_0*fw_h
347 XMPYU fht_1,fw_h,ht_temp_1 ; ht_temp = ht*fw_h
348
349 FSTD ht_temp,-24(%sp) ; -24(sp) = ht
350 FSTD ht_temp_1,-56(%sp) ; -56(sp) = ht
351 XMPYU flt_0,fw_l,lt_temp ; lt_temp = lt*fw_l
352 XMPYU flt_1,fw_l,lt_temp_1 ; lt_temp = lt*fw_l
353
354 FSTD lt_temp,-32(%sp) ; -32(sp) = lt
355 FSTD lt_temp_1,-64(%sp) ; -64(sp) = lt
356 LDD -8(%sp),m_0
357 LDD -40(%sp),m_1
358
359 LDD -16(%sp),m1_0
360 LDD -48(%sp),m1_1
361 LDD -24(%sp),ht_0
362 LDD -56(%sp),ht_1
363
364 ADD,L m1_0,m_0,tmp_0 ; tmp_0 = m + m1;
365 ADD,L m1_1,m_1,tmp_1 ; tmp_1 = m + m1;
366 LDD -32(%sp),lt_0
367 LDD -64(%sp),lt_1
368
369 CMPCLR,*>>= tmp_0,m1_0, %r0 ; if (m < m1)
370 ADD,L ht_0,top_overflow,ht_0 ; ht += (1<<32)
371 CMPCLR,*>>= tmp_1,m1_1,%r0 ; if (m < m1)
372 ADD,L ht_1,top_overflow,ht_1 ; ht += (1<<32)
373
374 EXTRD,U tmp_0,31,32,m_0 ; m>>32
375 DEPD,Z tmp_0,31,32,m1_0 ; m1 = m<<32
376 EXTRD,U tmp_1,31,32,m_1 ; m>>32
377 DEPD,Z tmp_1,31,32,m1_1 ; m1 = m<<32
378
379 ADD,L ht_0,m_0,ht_0 ; ht+= (m>>32)
380 ADD,L ht_1,m_1,ht_1 ; ht+= (m>>32)
381 ADD lt_0,m1_0,lt_0 ; lt = lt+m1;
382 ADD,DC ht_0,%r0,ht_0 ; ht++
383
384 ADD lt_1,m1_1,lt_1 ; lt = lt+m1;
385 ADD,DC ht_1,%r0,ht_1 ; ht++
386 ADD %ret1,lt_0,lt_0 ; lt = lt + c (ret1);
387 ADD,DC ht_0,%r0,ht_0 ; ht++
388
389 ADD ht_0,lt_1,lt_1 ; lt = lt + c (ht_0)
390 ADD,DC ht_1,%r0,ht_1 ; ht++
391 STD lt_0,0(r_ptr) ; rp[0] = lt
392 STD lt_1,8(r_ptr) ; rp[1] = lt
393
394 COPY ht_1,%ret1 ; carry = ht
395 LDO -2(num),num ; num = num - 2;
396 LDO 16(a_ptr),a_ptr ; ap += 2
397 CMPIB,<= 2,num,bn_mul_words_unroll2
398 LDO 16(r_ptr),r_ptr ; rp++
399
400 CMPIB,=,N 0,num,bn_mul_words_exit ; are we done?
401
402 ;
403 ; Top of loop aligned on 64-byte boundary
404 ;
405bn_mul_words_single_top
406 FLDD 0(a_ptr),t_float_0 ; load up 64-bit value (fr8L) ht(L)/lt(R)
407
408 XMPYU fht_0,fw_l,fm1 ; m1 = ht*fw_l
409 FSTD fm1,-16(%sp) ; -16(sp) = m1
410 XMPYU flt_0,fw_h,fm ; m = lt*fw_h
411 FSTD fm,-8(%sp) ; -8(sp) = m
412 XMPYU fht_0,fw_h,ht_temp ; ht_temp = ht*fw_h
413 FSTD ht_temp,-24(%sp) ; -24(sp) = ht
414 XMPYU flt_0,fw_l,lt_temp ; lt_temp = lt*fw_l
415 FSTD lt_temp,-32(%sp) ; -32(sp) = lt
416
417 LDD -8(%sp),m_0
418 LDD -16(%sp),m1_0
419 ADD,L m_0,m1_0,tmp_0 ; tmp_0 = m + m1;
420 LDD -24(%sp),ht_0
421 LDD -32(%sp),lt_0
422
423 CMPCLR,*>>= tmp_0,m1_0,%r0 ; if (m < m1)
424 ADD,L ht_0,top_overflow,ht_0 ; ht += (1<<32)
425
426 EXTRD,U tmp_0,31,32,m_0 ; m>>32
427 DEPD,Z tmp_0,31,32,m1_0 ; m1 = m<<32
428
429 ADD,L ht_0,m_0,ht_0 ; ht+= (m>>32)
430 ADD lt_0,m1_0,lt_0 ; lt= lt+m1;
431 ADD,DC ht_0,%r0,ht_0 ; ht++
432
433 ADD %ret1,lt_0,lt_0 ; lt = lt + c;
434 ADD,DC ht_0,%r0,ht_0 ; ht++
435
436 COPY ht_0,%ret1 ; copy carry
437 STD lt_0,0(r_ptr) ; rp[0] = lt
438
439bn_mul_words_exit
440 .EXIT
441 EXTRD,U %ret1,31,32,%ret0 ; for 32-bit, return in ret0/ret1
442 LDD -96(%sp),%r7 ; restore r7
443 LDD -104(%sp),%r6 ; restore r6
444 LDD -112(%sp),%r5 ; restore r5
445 LDD -120(%sp),%r4 ; restore r4
446 BVE (%rp)
447 LDD,MB -128(%sp),%r3 ; restore r3
448 .PROCEND
449
450;----------------------------------------------------------------------------
451;
452;void bn_sqr_words(BN_ULONG *rp, BN_ULONG *ap, int num)
453;
454; arg0 = rp
455; arg1 = ap
456; arg2 = num
457;
458
459bn_sqr_words
460 .proc
461 .callinfo FRAME=128,ENTRY_GR=%r3,ARGS_SAVED,ORDERING_AWARE
462 .EXPORT bn_sqr_words,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
463 .entry
464 .align 64
465
466 STD %r3,0(%sp) ; save r3
467 STD %r4,8(%sp) ; save r4
468 NOP
469 STD %r5,16(%sp) ; save r5
470
471 CMPIB,>= 0,num,bn_sqr_words_exit
472 LDO 128(%sp),%sp ; bump stack
473
474 ;
475 ; If only 1, the goto straight to cleanup
476 ;
477 CMPIB,= 1,num,bn_sqr_words_single_top
478 DEPDI,Z -1,32,33,high_mask ; Create Mask 0xffffffff80000000L
479
480 ;
481 ; This loop is unrolled 2 times (64-byte aligned as well)
482 ;
483
484bn_sqr_words_unroll2
485 FLDD 0(a_ptr),t_float_0 ; a[0]
486 FLDD 8(a_ptr),t_float_1 ; a[1]
487 XMPYU fht_0,flt_0,fm ; m[0]
488 XMPYU fht_1,flt_1,fm_1 ; m[1]
489
490 FSTD fm,-24(%sp) ; store m[0]
491 FSTD fm_1,-56(%sp) ; store m[1]
492 XMPYU flt_0,flt_0,lt_temp ; lt[0]
493 XMPYU flt_1,flt_1,lt_temp_1 ; lt[1]
494
495 FSTD lt_temp,-16(%sp) ; store lt[0]
496 FSTD lt_temp_1,-48(%sp) ; store lt[1]
497 XMPYU fht_0,fht_0,ht_temp ; ht[0]
498 XMPYU fht_1,fht_1,ht_temp_1 ; ht[1]
499
500 FSTD ht_temp,-8(%sp) ; store ht[0]
501 FSTD ht_temp_1,-40(%sp) ; store ht[1]
502 LDD -24(%sp),m_0
503 LDD -56(%sp),m_1
504
505 AND m_0,high_mask,tmp_0 ; m[0] & Mask
506 AND m_1,high_mask,tmp_1 ; m[1] & Mask
507 DEPD,Z m_0,30,31,m_0 ; m[0] << 32+1
508 DEPD,Z m_1,30,31,m_1 ; m[1] << 32+1
509
510 LDD -16(%sp),lt_0
511 LDD -48(%sp),lt_1
512 EXTRD,U tmp_0,32,33,tmp_0 ; tmp_0 = m[0]&Mask >> 32-1
513 EXTRD,U tmp_1,32,33,tmp_1 ; tmp_1 = m[1]&Mask >> 32-1
514
515 LDD -8(%sp),ht_0
516 LDD -40(%sp),ht_1
517 ADD,L ht_0,tmp_0,ht_0 ; ht[0] += tmp_0
518 ADD,L ht_1,tmp_1,ht_1 ; ht[1] += tmp_1
519
520 ADD lt_0,m_0,lt_0 ; lt = lt+m
521 ADD,DC ht_0,%r0,ht_0 ; ht[0]++
522 STD lt_0,0(r_ptr) ; rp[0] = lt[0]
523 STD ht_0,8(r_ptr) ; rp[1] = ht[1]
524
525 ADD lt_1,m_1,lt_1 ; lt = lt+m
526 ADD,DC ht_1,%r0,ht_1 ; ht[1]++
527 STD lt_1,16(r_ptr) ; rp[2] = lt[1]
528 STD ht_1,24(r_ptr) ; rp[3] = ht[1]
529
530 LDO -2(num),num ; num = num - 2;
531 LDO 16(a_ptr),a_ptr ; ap += 2
532 CMPIB,<= 2,num,bn_sqr_words_unroll2
533 LDO 32(r_ptr),r_ptr ; rp += 4
534
535 CMPIB,=,N 0,num,bn_sqr_words_exit ; are we done?
536
537 ;
538 ; Top of loop aligned on 64-byte boundary
539 ;
540bn_sqr_words_single_top
541 FLDD 0(a_ptr),t_float_0 ; load up 64-bit value (fr8L) ht(L)/lt(R)
542
543 XMPYU fht_0,flt_0,fm ; m
544 FSTD fm,-24(%sp) ; store m
545
546 XMPYU flt_0,flt_0,lt_temp ; lt
547 FSTD lt_temp,-16(%sp) ; store lt
548
549 XMPYU fht_0,fht_0,ht_temp ; ht
550 FSTD ht_temp,-8(%sp) ; store ht
551
552 LDD -24(%sp),m_0 ; load m
553 AND m_0,high_mask,tmp_0 ; m & Mask
554 DEPD,Z m_0,30,31,m_0 ; m << 32+1
555 LDD -16(%sp),lt_0 ; lt
556
557 LDD -8(%sp),ht_0 ; ht
558 EXTRD,U tmp_0,32,33,tmp_0 ; tmp_0 = m&Mask >> 32-1
559 ADD m_0,lt_0,lt_0 ; lt = lt+m
560 ADD,L ht_0,tmp_0,ht_0 ; ht += tmp_0
561 ADD,DC ht_0,%r0,ht_0 ; ht++
562
563 STD lt_0,0(r_ptr) ; rp[0] = lt
564 STD ht_0,8(r_ptr) ; rp[1] = ht
565
566bn_sqr_words_exit
567 .EXIT
568 LDD -112(%sp),%r5 ; restore r5
569 LDD -120(%sp),%r4 ; restore r4
570 BVE (%rp)
571 LDD,MB -128(%sp),%r3
572 .PROCEND ;in=23,24,25,26,29;out=28;
573
574
575;----------------------------------------------------------------------------
576;
577;BN_ULONG bn_add_words(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n)
578;
579; arg0 = rp
580; arg1 = ap
581; arg2 = bp
582; arg3 = n
583
584t .reg %r22
585b .reg %r21
586l .reg %r20
587
588bn_add_words
589 .proc
590 .entry
591 .callinfo
592 .EXPORT bn_add_words,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
593 .align 64
594
595 CMPIB,>= 0,n,bn_add_words_exit
596 COPY %r0,%ret1 ; return 0 by default
597
598 ;
599 ; If 2 or more numbers do the loop
600 ;
601 CMPIB,= 1,n,bn_add_words_single_top
602 NOP
603
604 ;
605 ; This loop is unrolled 2 times (64-byte aligned as well)
606 ;
607bn_add_words_unroll2
608 LDD 0(a_ptr),t
609 LDD 0(b_ptr),b
610 ADD t,%ret1,t ; t = t+c;
611 ADD,DC %r0,%r0,%ret1 ; set c to carry
612 ADD t,b,l ; l = t + b[0]
613 ADD,DC %ret1,%r0,%ret1 ; c+= carry
614 STD l,0(r_ptr)
615
616 LDD 8(a_ptr),t
617 LDD 8(b_ptr),b
618 ADD t,%ret1,t ; t = t+c;
619 ADD,DC %r0,%r0,%ret1 ; set c to carry
620 ADD t,b,l ; l = t + b[0]
621 ADD,DC %ret1,%r0,%ret1 ; c+= carry
622 STD l,8(r_ptr)
623
624 LDO -2(n),n
625 LDO 16(a_ptr),a_ptr
626 LDO 16(b_ptr),b_ptr
627
628 CMPIB,<= 2,n,bn_add_words_unroll2
629 LDO 16(r_ptr),r_ptr
630
631 CMPIB,=,N 0,n,bn_add_words_exit ; are we done?
632
633bn_add_words_single_top
634 LDD 0(a_ptr),t
635 LDD 0(b_ptr),b
636
637 ADD t,%ret1,t ; t = t+c;
638 ADD,DC %r0,%r0,%ret1 ; set c to carry (could use CMPCLR??)
639 ADD t,b,l ; l = t + b[0]
640 ADD,DC %ret1,%r0,%ret1 ; c+= carry
641 STD l,0(r_ptr)
642
643bn_add_words_exit
644 .EXIT
645 BVE (%rp)
646 EXTRD,U %ret1,31,32,%ret0 ; for 32-bit, return in ret0/ret1
647 .PROCEND ;in=23,24,25,26,29;out=28;
648
649;----------------------------------------------------------------------------
650;
651;BN_ULONG bn_sub_words(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n)
652;
653; arg0 = rp
654; arg1 = ap
655; arg2 = bp
656; arg3 = n
657
658t1 .reg %r22
659t2 .reg %r21
660sub_tmp1 .reg %r20
661sub_tmp2 .reg %r19
662
663
664bn_sub_words
665 .proc
666 .callinfo
667 .EXPORT bn_sub_words,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
668 .entry
669 .align 64
670
671 CMPIB,>= 0,n,bn_sub_words_exit
672 COPY %r0,%ret1 ; return 0 by default
673
674 ;
675 ; If 2 or more numbers do the loop
676 ;
677 CMPIB,= 1,n,bn_sub_words_single_top
678 NOP
679
680 ;
681 ; This loop is unrolled 2 times (64-byte aligned as well)
682 ;
683bn_sub_words_unroll2
684 LDD 0(a_ptr),t1
685 LDD 0(b_ptr),t2
686 SUB t1,t2,sub_tmp1 ; t3 = t1-t2;
687 SUB sub_tmp1,%ret1,sub_tmp1 ; t3 = t3- c;
688
689 CMPCLR,*>> t1,t2,sub_tmp2 ; clear if t1 > t2
690 LDO 1(%r0),sub_tmp2
691
692 CMPCLR,*= t1,t2,%r0
693 COPY sub_tmp2,%ret1
694 STD sub_tmp1,0(r_ptr)
695
696 LDD 8(a_ptr),t1
697 LDD 8(b_ptr),t2
698 SUB t1,t2,sub_tmp1 ; t3 = t1-t2;
699 SUB sub_tmp1,%ret1,sub_tmp1 ; t3 = t3- c;
700 CMPCLR,*>> t1,t2,sub_tmp2 ; clear if t1 > t2
701 LDO 1(%r0),sub_tmp2
702
703 CMPCLR,*= t1,t2,%r0
704 COPY sub_tmp2,%ret1
705 STD sub_tmp1,8(r_ptr)
706
707 LDO -2(n),n
708 LDO 16(a_ptr),a_ptr
709 LDO 16(b_ptr),b_ptr
710
711 CMPIB,<= 2,n,bn_sub_words_unroll2
712 LDO 16(r_ptr),r_ptr
713
714 CMPIB,=,N 0,n,bn_sub_words_exit ; are we done?
715
716bn_sub_words_single_top
717 LDD 0(a_ptr),t1
718 LDD 0(b_ptr),t2
719 SUB t1,t2,sub_tmp1 ; t3 = t1-t2;
720 SUB sub_tmp1,%ret1,sub_tmp1 ; t3 = t3- c;
721 CMPCLR,*>> t1,t2,sub_tmp2 ; clear if t1 > t2
722 LDO 1(%r0),sub_tmp2
723
724 CMPCLR,*= t1,t2,%r0
725 COPY sub_tmp2,%ret1
726
727 STD sub_tmp1,0(r_ptr)
728
729bn_sub_words_exit
730 .EXIT
731 BVE (%rp)
732 EXTRD,U %ret1,31,32,%ret0 ; for 32-bit, return in ret0/ret1
733 .PROCEND ;in=23,24,25,26,29;out=28;
734
735;------------------------------------------------------------------------------
736;
737; unsigned long bn_div_words(unsigned long h, unsigned long l, unsigned long d)
738;
739; arg0 = h
740; arg1 = l
741; arg2 = d
742;
743; This is mainly just output from the HP C compiler.
744;
745;------------------------------------------------------------------------------
746bn_div_words
747 .PROC
748 .EXPORT bn_div_words,ENTRY,PRIV_LEV=3,ARGW0=GR,ARGW1=GR,ARGW2=GR,ARGW3=GR,RTNVAL=GR,LONG_RETURN
749 .IMPORT BN_num_bits_word,CODE
750 ;--- not PIC .IMPORT __iob,DATA
751 ;--- not PIC .IMPORT fprintf,CODE
752 .IMPORT abort,CODE
753 .IMPORT $$div2U,MILLICODE
754 .CALLINFO CALLER,FRAME=144,ENTRY_GR=%r9,SAVE_RP,ARGS_SAVED,ORDERING_AWARE
755 .ENTRY
756 STW %r2,-20(%r30) ;offset 0x8ec
757 STW,MA %r3,192(%r30) ;offset 0x8f0
758 STW %r4,-188(%r30) ;offset 0x8f4
759 DEPD %r5,31,32,%r6 ;offset 0x8f8
760 STD %r6,-184(%r30) ;offset 0x8fc
761 DEPD %r7,31,32,%r8 ;offset 0x900
762 STD %r8,-176(%r30) ;offset 0x904
763 STW %r9,-168(%r30) ;offset 0x908
764 LDD -248(%r30),%r3 ;offset 0x90c
765 COPY %r26,%r4 ;offset 0x910
766 COPY %r24,%r5 ;offset 0x914
767 DEPD %r25,31,32,%r4 ;offset 0x918
768 CMPB,*<> %r3,%r0,$0006000C ;offset 0x91c
769 DEPD %r23,31,32,%r5 ;offset 0x920
770 MOVIB,TR -1,%r29,$00060002 ;offset 0x924
771 EXTRD,U %r29,31,32,%r28 ;offset 0x928
772$0006002A
773 LDO -1(%r29),%r29 ;offset 0x92c
774 SUB %r23,%r7,%r23 ;offset 0x930
775$00060024
776 SUB %r4,%r31,%r25 ;offset 0x934
777 AND %r25,%r19,%r26 ;offset 0x938
778 CMPB,*<>,N %r0,%r26,$00060046 ;offset 0x93c
779 DEPD,Z %r25,31,32,%r20 ;offset 0x940
780 OR %r20,%r24,%r21 ;offset 0x944
781 CMPB,*<<,N %r21,%r23,$0006002A ;offset 0x948
782 SUB %r31,%r2,%r31 ;offset 0x94c
783$00060046
784$0006002E
785 DEPD,Z %r23,31,32,%r25 ;offset 0x950
786 EXTRD,U %r23,31,32,%r26 ;offset 0x954
787 AND %r25,%r19,%r24 ;offset 0x958
788 ADD,L %r31,%r26,%r31 ;offset 0x95c
789 CMPCLR,*>>= %r5,%r24,%r0 ;offset 0x960
790 LDO 1(%r31),%r31 ;offset 0x964
791$00060032
792 CMPB,*<<=,N %r31,%r4,$00060036 ;offset 0x968
793 LDO -1(%r29),%r29 ;offset 0x96c
794 ADD,L %r4,%r3,%r4 ;offset 0x970
795$00060036
796 ADDIB,=,N -1,%r8,$D0 ;offset 0x974
797 SUB %r5,%r24,%r28 ;offset 0x978
798$0006003A
799 SUB %r4,%r31,%r24 ;offset 0x97c
800 SHRPD %r24,%r28,32,%r4 ;offset 0x980
801 DEPD,Z %r29,31,32,%r9 ;offset 0x984
802 DEPD,Z %r28,31,32,%r5 ;offset 0x988
803$0006001C
804 EXTRD,U %r4,31,32,%r31 ;offset 0x98c
805 CMPB,*<>,N %r31,%r2,$00060020 ;offset 0x990
806 MOVB,TR %r6,%r29,$D1 ;offset 0x994
807 STD %r29,-152(%r30) ;offset 0x998
808$0006000C
809 EXTRD,U %r3,31,32,%r25 ;offset 0x99c
810 COPY %r3,%r26 ;offset 0x9a0
811 EXTRD,U %r3,31,32,%r9 ;offset 0x9a4
812 EXTRD,U %r4,31,32,%r8 ;offset 0x9a8
813 .CALL ARGW0=GR,ARGW1=GR,RTNVAL=GR ;in=25,26;out=28;
814 B,L BN_num_bits_word,%r2 ;offset 0x9ac
815 EXTRD,U %r5,31,32,%r7 ;offset 0x9b0
816 LDI 64,%r20 ;offset 0x9b4
817 DEPD %r7,31,32,%r5 ;offset 0x9b8
818 DEPD %r8,31,32,%r4 ;offset 0x9bc
819 DEPD %r9,31,32,%r3 ;offset 0x9c0
820 CMPB,= %r28,%r20,$00060012 ;offset 0x9c4
821 COPY %r28,%r24 ;offset 0x9c8
822 MTSARCM %r24 ;offset 0x9cc
823 DEPDI,Z -1,%sar,1,%r19 ;offset 0x9d0
824 CMPB,*>>,N %r4,%r19,$D2 ;offset 0x9d4
825$00060012
826 SUBI 64,%r24,%r31 ;offset 0x9d8
827 CMPCLR,*<< %r4,%r3,%r0 ;offset 0x9dc
828 SUB %r4,%r3,%r4 ;offset 0x9e0
829$00060016
830 CMPB,= %r31,%r0,$0006001A ;offset 0x9e4
831 COPY %r0,%r9 ;offset 0x9e8
832 MTSARCM %r31 ;offset 0x9ec
833 DEPD,Z %r3,%sar,64,%r3 ;offset 0x9f0
834 SUBI 64,%r31,%r26 ;offset 0x9f4
835 MTSAR %r26 ;offset 0x9f8
836 SHRPD %r4,%r5,%sar,%r4 ;offset 0x9fc
837 MTSARCM %r31 ;offset 0xa00
838 DEPD,Z %r5,%sar,64,%r5 ;offset 0xa04
839$0006001A
840 DEPDI,Z -1,31,32,%r19 ;offset 0xa08
841 AND %r3,%r19,%r29 ;offset 0xa0c
842 EXTRD,U %r29,31,32,%r2 ;offset 0xa10
843 DEPDI,Z -1,63,32,%r6 ;offset 0xa14
844 MOVIB,TR 2,%r8,$0006001C ;offset 0xa18
845 EXTRD,U %r3,63,32,%r7 ;offset 0xa1c
846$D2
847 ;--- not PIC ADDIL LR'__iob-$global$,%r27,%r1 ;offset 0xa20
848 ;--- not PIC LDIL LR'C$7,%r21 ;offset 0xa24
849 ;--- not PIC LDO RR'__iob-$global$+32(%r1),%r26 ;offset 0xa28
850 ;--- not PIC .CALL ARGW0=GR,ARGW1=GR,ARGW2=GR,RTNVAL=GR ;in=24,25,26;out=28;
851 ;--- not PIC B,L fprintf,%r2 ;offset 0xa2c
852 ;--- not PIC LDO RR'C$7(%r21),%r25 ;offset 0xa30
853 .CALL ;
854 B,L abort,%r2 ;offset 0xa34
855 NOP ;offset 0xa38
856 B $D3 ;offset 0xa3c
857 LDW -212(%r30),%r2 ;offset 0xa40
858$00060020
859 COPY %r4,%r26 ;offset 0xa44
860 EXTRD,U %r4,31,32,%r25 ;offset 0xa48
861 COPY %r2,%r24 ;offset 0xa4c
862 .CALL ;in=23,24,25,26;out=20,21,22,28,29; (MILLICALL)
863 B,L $$div2U,%r31 ;offset 0xa50
864 EXTRD,U %r2,31,32,%r23 ;offset 0xa54
865 DEPD %r28,31,32,%r29 ;offset 0xa58
866$00060022
867 STD %r29,-152(%r30) ;offset 0xa5c
868$D1
869 AND %r5,%r19,%r24 ;offset 0xa60
870 EXTRD,U %r24,31,32,%r24 ;offset 0xa64
871 STW %r2,-160(%r30) ;offset 0xa68
872 STW %r7,-128(%r30) ;offset 0xa6c
873 FLDD -152(%r30),%fr4 ;offset 0xa70
874 FLDD -152(%r30),%fr7 ;offset 0xa74
875 FLDW -160(%r30),%fr8L ;offset 0xa78
876 FLDW -128(%r30),%fr5L ;offset 0xa7c
877 XMPYU %fr8L,%fr7L,%fr10 ;offset 0xa80
878 FSTD %fr10,-136(%r30) ;offset 0xa84
879 XMPYU %fr8L,%fr7R,%fr22 ;offset 0xa88
880 FSTD %fr22,-144(%r30) ;offset 0xa8c
881 XMPYU %fr5L,%fr4L,%fr11 ;offset 0xa90
882 XMPYU %fr5L,%fr4R,%fr23 ;offset 0xa94
883 FSTD %fr11,-112(%r30) ;offset 0xa98
884 FSTD %fr23,-120(%r30) ;offset 0xa9c
885 LDD -136(%r30),%r28 ;offset 0xaa0
886 DEPD,Z %r28,31,32,%r31 ;offset 0xaa4
887 LDD -144(%r30),%r20 ;offset 0xaa8
888 ADD,L %r20,%r31,%r31 ;offset 0xaac
889 LDD -112(%r30),%r22 ;offset 0xab0
890 DEPD,Z %r22,31,32,%r22 ;offset 0xab4
891 LDD -120(%r30),%r21 ;offset 0xab8
892 B $00060024 ;offset 0xabc
893 ADD,L %r21,%r22,%r23 ;offset 0xac0
894$D0
895 OR %r9,%r29,%r29 ;offset 0xac4
896$00060040
897 EXTRD,U %r29,31,32,%r28 ;offset 0xac8
898$00060002
899$L2
900 LDW -212(%r30),%r2 ;offset 0xacc
901$D3
902 LDW -168(%r30),%r9 ;offset 0xad0
903 LDD -176(%r30),%r8 ;offset 0xad4
904 EXTRD,U %r8,31,32,%r7 ;offset 0xad8
905 LDD -184(%r30),%r6 ;offset 0xadc
906 EXTRD,U %r6,31,32,%r5 ;offset 0xae0
907 LDW -188(%r30),%r4 ;offset 0xae4
908 BVE (%r2) ;offset 0xae8
909 .EXIT
910 LDW,MB -192(%r30),%r3 ;offset 0xaec
911 .PROCEND ;in=23,25;out=28,29;fpin=105,107;
912
913
914
915
916;----------------------------------------------------------------------------
917;
918; Registers to hold 64-bit values to manipulate. The "L" part
919; of the register corresponds to the upper 32-bits, while the "R"
920; part corresponds to the lower 32-bits
921;
922; Note, that when using b6 and b7, the code must save these before
923; using them because they are callee save registers
924;
925;
926; Floating point registers to use to save values that
927; are manipulated. These don't collide with ftemp1-6 and
928; are all caller save registers
929;
930a0 .reg %fr22
931a0L .reg %fr22L
932a0R .reg %fr22R
933
934a1 .reg %fr23
935a1L .reg %fr23L
936a1R .reg %fr23R
937
938a2 .reg %fr24
939a2L .reg %fr24L
940a2R .reg %fr24R
941
942a3 .reg %fr25
943a3L .reg %fr25L
944a3R .reg %fr25R
945
946a4 .reg %fr26
947a4L .reg %fr26L
948a4R .reg %fr26R
949
950a5 .reg %fr27
951a5L .reg %fr27L
952a5R .reg %fr27R
953
954a6 .reg %fr28
955a6L .reg %fr28L
956a6R .reg %fr28R
957
958a7 .reg %fr29
959a7L .reg %fr29L
960a7R .reg %fr29R
961
962b0 .reg %fr30
963b0L .reg %fr30L
964b0R .reg %fr30R
965
966b1 .reg %fr31
967b1L .reg %fr31L
968b1R .reg %fr31R
969
970;
971; Temporary floating point variables, these are all caller save
972; registers
973;
974ftemp1 .reg %fr4
975ftemp2 .reg %fr5
976ftemp3 .reg %fr6
977ftemp4 .reg %fr7
978
979;
980; The B set of registers when used.
981;
982
983b2 .reg %fr8
984b2L .reg %fr8L
985b2R .reg %fr8R
986
987b3 .reg %fr9
988b3L .reg %fr9L
989b3R .reg %fr9R
990
991b4 .reg %fr10
992b4L .reg %fr10L
993b4R .reg %fr10R
994
995b5 .reg %fr11
996b5L .reg %fr11L
997b5R .reg %fr11R
998
999b6 .reg %fr12
1000b6L .reg %fr12L
1001b6R .reg %fr12R
1002
1003b7 .reg %fr13
1004b7L .reg %fr13L
1005b7R .reg %fr13R
1006
1007c1 .reg %r21 ; only reg
1008temp1 .reg %r20 ; only reg
1009temp2 .reg %r19 ; only reg
1010temp3 .reg %r31 ; only reg
1011
1012m1 .reg %r28
1013c2 .reg %r23
1014high_one .reg %r1
1015ht .reg %r6
1016lt .reg %r5
1017m .reg %r4
1018c3 .reg %r3
1019
1020SQR_ADD_C .macro A0L,A0R,C1,C2,C3
1021 XMPYU A0L,A0R,ftemp1 ; m
1022 FSTD ftemp1,-24(%sp) ; store m
1023
1024 XMPYU A0R,A0R,ftemp2 ; lt
1025 FSTD ftemp2,-16(%sp) ; store lt
1026
1027 XMPYU A0L,A0L,ftemp3 ; ht
1028 FSTD ftemp3,-8(%sp) ; store ht
1029
1030 LDD -24(%sp),m ; load m
1031 AND m,high_mask,temp2 ; m & Mask
1032 DEPD,Z m,30,31,temp3 ; m << 32+1
1033 LDD -16(%sp),lt ; lt
1034
1035 LDD -8(%sp),ht ; ht
1036 EXTRD,U temp2,32,33,temp1 ; temp1 = m&Mask >> 32-1
1037 ADD temp3,lt,lt ; lt = lt+m
1038 ADD,L ht,temp1,ht ; ht += temp1
1039 ADD,DC ht,%r0,ht ; ht++
1040
1041 ADD C1,lt,C1 ; c1=c1+lt
1042 ADD,DC ht,%r0,ht ; ht++
1043
1044 ADD C2,ht,C2 ; c2=c2+ht
1045 ADD,DC C3,%r0,C3 ; c3++
1046.endm
1047
1048SQR_ADD_C2 .macro A0L,A0R,A1L,A1R,C1,C2,C3
1049 XMPYU A0L,A1R,ftemp1 ; m1 = bl*ht
1050 FSTD ftemp1,-16(%sp) ;
1051 XMPYU A0R,A1L,ftemp2 ; m = bh*lt
1052 FSTD ftemp2,-8(%sp) ;
1053 XMPYU A0R,A1R,ftemp3 ; lt = bl*lt
1054 FSTD ftemp3,-32(%sp)
1055 XMPYU A0L,A1L,ftemp4 ; ht = bh*ht
1056 FSTD ftemp4,-24(%sp) ;
1057
1058 LDD -8(%sp),m ; r21 = m
1059 LDD -16(%sp),m1 ; r19 = m1
1060 ADD,L m,m1,m ; m+m1
1061
1062 DEPD,Z m,31,32,temp3 ; (m+m1<<32)
1063 LDD -24(%sp),ht ; r24 = ht
1064
1065 CMPCLR,*>>= m,m1,%r0 ; if (m < m1)
1066 ADD,L ht,high_one,ht ; ht+=high_one
1067
1068 EXTRD,U m,31,32,temp1 ; m >> 32
1069 LDD -32(%sp),lt ; lt
1070 ADD,L ht,temp1,ht ; ht+= m>>32
1071 ADD lt,temp3,lt ; lt = lt+m1
1072 ADD,DC ht,%r0,ht ; ht++
1073
1074 ADD ht,ht,ht ; ht=ht+ht;
1075 ADD,DC C3,%r0,C3 ; add in carry (c3++)
1076
1077 ADD lt,lt,lt ; lt=lt+lt;
1078 ADD,DC ht,%r0,ht ; add in carry (ht++)
1079
1080 ADD C1,lt,C1 ; c1=c1+lt
1081 ADD,DC,*NUV ht,%r0,ht ; add in carry (ht++)
1082 LDO 1(C3),C3 ; bump c3 if overflow,nullify otherwise
1083
1084 ADD C2,ht,C2 ; c2 = c2 + ht
1085 ADD,DC C3,%r0,C3 ; add in carry (c3++)
1086.endm
1087
1088;
1089;void bn_sqr_comba8(BN_ULONG *r, BN_ULONG *a)
1090; arg0 = r_ptr
1091; arg1 = a_ptr
1092;
1093
1094bn_sqr_comba8
1095 .PROC
1096 .CALLINFO FRAME=128,ENTRY_GR=%r3,ARGS_SAVED,ORDERING_AWARE
1097 .EXPORT bn_sqr_comba8,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
1098 .ENTRY
1099 .align 64
1100
1101 STD %r3,0(%sp) ; save r3
1102 STD %r4,8(%sp) ; save r4
1103 STD %r5,16(%sp) ; save r5
1104 STD %r6,24(%sp) ; save r6
1105
1106 ;
1107 ; Zero out carries
1108 ;
1109 COPY %r0,c1
1110 COPY %r0,c2
1111 COPY %r0,c3
1112
1113 LDO 128(%sp),%sp ; bump stack
1114 DEPDI,Z -1,32,33,high_mask ; Create Mask 0xffffffff80000000L
1115 DEPDI,Z 1,31,1,high_one ; Create Value 1 << 32
1116
1117 ;
1118 ; Load up all of the values we are going to use
1119 ;
1120 FLDD 0(a_ptr),a0
1121 FLDD 8(a_ptr),a1
1122 FLDD 16(a_ptr),a2
1123 FLDD 24(a_ptr),a3
1124 FLDD 32(a_ptr),a4
1125 FLDD 40(a_ptr),a5
1126 FLDD 48(a_ptr),a6
1127 FLDD 56(a_ptr),a7
1128
1129 SQR_ADD_C a0L,a0R,c1,c2,c3
1130 STD c1,0(r_ptr) ; r[0] = c1;
1131 COPY %r0,c1
1132
1133 SQR_ADD_C2 a1L,a1R,a0L,a0R,c2,c3,c1
1134 STD c2,8(r_ptr) ; r[1] = c2;
1135 COPY %r0,c2
1136
1137 SQR_ADD_C a1L,a1R,c3,c1,c2
1138 SQR_ADD_C2 a2L,a2R,a0L,a0R,c3,c1,c2
1139 STD c3,16(r_ptr) ; r[2] = c3;
1140 COPY %r0,c3
1141
1142 SQR_ADD_C2 a3L,a3R,a0L,a0R,c1,c2,c3
1143 SQR_ADD_C2 a2L,a2R,a1L,a1R,c1,c2,c3
1144 STD c1,24(r_ptr) ; r[3] = c1;
1145 COPY %r0,c1
1146
1147 SQR_ADD_C a2L,a2R,c2,c3,c1
1148 SQR_ADD_C2 a3L,a3R,a1L,a1R,c2,c3,c1
1149 SQR_ADD_C2 a4L,a4R,a0L,a0R,c2,c3,c1
1150 STD c2,32(r_ptr) ; r[4] = c2;
1151 COPY %r0,c2
1152
1153 SQR_ADD_C2 a5L,a5R,a0L,a0R,c3,c1,c2
1154 SQR_ADD_C2 a4L,a4R,a1L,a1R,c3,c1,c2
1155 SQR_ADD_C2 a3L,a3R,a2L,a2R,c3,c1,c2
1156 STD c3,40(r_ptr) ; r[5] = c3;
1157 COPY %r0,c3
1158
1159 SQR_ADD_C a3L,a3R,c1,c2,c3
1160 SQR_ADD_C2 a4L,a4R,a2L,a2R,c1,c2,c3
1161 SQR_ADD_C2 a5L,a5R,a1L,a1R,c1,c2,c3
1162 SQR_ADD_C2 a6L,a6R,a0L,a0R,c1,c2,c3
1163 STD c1,48(r_ptr) ; r[6] = c1;
1164 COPY %r0,c1
1165
1166 SQR_ADD_C2 a7L,a7R,a0L,a0R,c2,c3,c1
1167 SQR_ADD_C2 a6L,a6R,a1L,a1R,c2,c3,c1
1168 SQR_ADD_C2 a5L,a5R,a2L,a2R,c2,c3,c1
1169 SQR_ADD_C2 a4L,a4R,a3L,a3R,c2,c3,c1
1170 STD c2,56(r_ptr) ; r[7] = c2;
1171 COPY %r0,c2
1172
1173 SQR_ADD_C a4L,a4R,c3,c1,c2
1174 SQR_ADD_C2 a5L,a5R,a3L,a3R,c3,c1,c2
1175 SQR_ADD_C2 a6L,a6R,a2L,a2R,c3,c1,c2
1176 SQR_ADD_C2 a7L,a7R,a1L,a1R,c3,c1,c2
1177 STD c3,64(r_ptr) ; r[8] = c3;
1178 COPY %r0,c3
1179
1180 SQR_ADD_C2 a7L,a7R,a2L,a2R,c1,c2,c3
1181 SQR_ADD_C2 a6L,a6R,a3L,a3R,c1,c2,c3
1182 SQR_ADD_C2 a5L,a5R,a4L,a4R,c1,c2,c3
1183 STD c1,72(r_ptr) ; r[9] = c1;
1184 COPY %r0,c1
1185
1186 SQR_ADD_C a5L,a5R,c2,c3,c1
1187 SQR_ADD_C2 a6L,a6R,a4L,a4R,c2,c3,c1
1188 SQR_ADD_C2 a7L,a7R,a3L,a3R,c2,c3,c1
1189 STD c2,80(r_ptr) ; r[10] = c2;
1190 COPY %r0,c2
1191
1192 SQR_ADD_C2 a7L,a7R,a4L,a4R,c3,c1,c2
1193 SQR_ADD_C2 a6L,a6R,a5L,a5R,c3,c1,c2
1194 STD c3,88(r_ptr) ; r[11] = c3;
1195 COPY %r0,c3
1196
1197 SQR_ADD_C a6L,a6R,c1,c2,c3
1198 SQR_ADD_C2 a7L,a7R,a5L,a5R,c1,c2,c3
1199 STD c1,96(r_ptr) ; r[12] = c1;
1200 COPY %r0,c1
1201
1202 SQR_ADD_C2 a7L,a7R,a6L,a6R,c2,c3,c1
1203 STD c2,104(r_ptr) ; r[13] = c2;
1204 COPY %r0,c2
1205
1206 SQR_ADD_C a7L,a7R,c3,c1,c2
1207 STD c3, 112(r_ptr) ; r[14] = c3
1208 STD c1, 120(r_ptr) ; r[15] = c1
1209
1210 .EXIT
1211 LDD -104(%sp),%r6 ; restore r6
1212 LDD -112(%sp),%r5 ; restore r5
1213 LDD -120(%sp),%r4 ; restore r4
1214 BVE (%rp)
1215 LDD,MB -128(%sp),%r3
1216
1217 .PROCEND
1218
1219;-----------------------------------------------------------------------------
1220;
1221;void bn_sqr_comba4(BN_ULONG *r, BN_ULONG *a)
1222; arg0 = r_ptr
1223; arg1 = a_ptr
1224;
1225
1226bn_sqr_comba4
1227 .proc
1228 .callinfo FRAME=128,ENTRY_GR=%r3,ARGS_SAVED,ORDERING_AWARE
1229 .EXPORT bn_sqr_comba4,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
1230 .entry
1231 .align 64
1232 STD %r3,0(%sp) ; save r3
1233 STD %r4,8(%sp) ; save r4
1234 STD %r5,16(%sp) ; save r5
1235 STD %r6,24(%sp) ; save r6
1236
1237 ;
1238 ; Zero out carries
1239 ;
1240 COPY %r0,c1
1241 COPY %r0,c2
1242 COPY %r0,c3
1243
1244 LDO 128(%sp),%sp ; bump stack
1245 DEPDI,Z -1,32,33,high_mask ; Create Mask 0xffffffff80000000L
1246 DEPDI,Z 1,31,1,high_one ; Create Value 1 << 32
1247
1248 ;
1249 ; Load up all of the values we are going to use
1250 ;
1251 FLDD 0(a_ptr),a0
1252 FLDD 8(a_ptr),a1
1253 FLDD 16(a_ptr),a2
1254 FLDD 24(a_ptr),a3
1255 FLDD 32(a_ptr),a4
1256 FLDD 40(a_ptr),a5
1257 FLDD 48(a_ptr),a6
1258 FLDD 56(a_ptr),a7
1259
1260 SQR_ADD_C a0L,a0R,c1,c2,c3
1261
1262 STD c1,0(r_ptr) ; r[0] = c1;
1263 COPY %r0,c1
1264
1265 SQR_ADD_C2 a1L,a1R,a0L,a0R,c2,c3,c1
1266
1267 STD c2,8(r_ptr) ; r[1] = c2;
1268 COPY %r0,c2
1269
1270 SQR_ADD_C a1L,a1R,c3,c1,c2
1271 SQR_ADD_C2 a2L,a2R,a0L,a0R,c3,c1,c2
1272
1273 STD c3,16(r_ptr) ; r[2] = c3;
1274 COPY %r0,c3
1275
1276 SQR_ADD_C2 a3L,a3R,a0L,a0R,c1,c2,c3
1277 SQR_ADD_C2 a2L,a2R,a1L,a1R,c1,c2,c3
1278
1279 STD c1,24(r_ptr) ; r[3] = c1;
1280 COPY %r0,c1
1281
1282 SQR_ADD_C a2L,a2R,c2,c3,c1
1283 SQR_ADD_C2 a3L,a3R,a1L,a1R,c2,c3,c1
1284
1285 STD c2,32(r_ptr) ; r[4] = c2;
1286 COPY %r0,c2
1287
1288 SQR_ADD_C2 a3L,a3R,a2L,a2R,c3,c1,c2
1289 STD c3,40(r_ptr) ; r[5] = c3;
1290 COPY %r0,c3
1291
1292 SQR_ADD_C a3L,a3R,c1,c2,c3
1293 STD c1,48(r_ptr) ; r[6] = c1;
1294 STD c2,56(r_ptr) ; r[7] = c2;
1295
1296 .EXIT
1297 LDD -104(%sp),%r6 ; restore r6
1298 LDD -112(%sp),%r5 ; restore r5
1299 LDD -120(%sp),%r4 ; restore r4
1300 BVE (%rp)
1301 LDD,MB -128(%sp),%r3
1302
1303 .PROCEND
1304
1305
1306;---------------------------------------------------------------------------
1307
1308MUL_ADD_C .macro A0L,A0R,B0L,B0R,C1,C2,C3
1309 XMPYU A0L,B0R,ftemp1 ; m1 = bl*ht
1310 FSTD ftemp1,-16(%sp) ;
1311 XMPYU A0R,B0L,ftemp2 ; m = bh*lt
1312 FSTD ftemp2,-8(%sp) ;
1313 XMPYU A0R,B0R,ftemp3 ; lt = bl*lt
1314 FSTD ftemp3,-32(%sp)
1315 XMPYU A0L,B0L,ftemp4 ; ht = bh*ht
1316 FSTD ftemp4,-24(%sp) ;
1317
1318 LDD -8(%sp),m ; r21 = m
1319 LDD -16(%sp),m1 ; r19 = m1
1320 ADD,L m,m1,m ; m+m1
1321
1322 DEPD,Z m,31,32,temp3 ; (m+m1<<32)
1323 LDD -24(%sp),ht ; r24 = ht
1324
1325 CMPCLR,*>>= m,m1,%r0 ; if (m < m1)
1326 ADD,L ht,high_one,ht ; ht+=high_one
1327
1328 EXTRD,U m,31,32,temp1 ; m >> 32
1329 LDD -32(%sp),lt ; lt
1330 ADD,L ht,temp1,ht ; ht+= m>>32
1331 ADD lt,temp3,lt ; lt = lt+m1
1332 ADD,DC ht,%r0,ht ; ht++
1333
1334 ADD C1,lt,C1 ; c1=c1+lt
1335 ADD,DC ht,%r0,ht ; bump c3 if overflow,nullify otherwise
1336
1337 ADD C2,ht,C2 ; c2 = c2 + ht
1338 ADD,DC C3,%r0,C3 ; add in carry (c3++)
1339.endm
1340
1341
1342;
1343;void bn_mul_comba8(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b)
1344; arg0 = r_ptr
1345; arg1 = a_ptr
1346; arg2 = b_ptr
1347;
1348
1349bn_mul_comba8
1350 .proc
1351 .callinfo FRAME=128,ENTRY_GR=%r3,ARGS_SAVED,ORDERING_AWARE
1352 .EXPORT bn_mul_comba8,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
1353 .entry
1354 .align 64
1355
1356 STD %r3,0(%sp) ; save r3
1357 STD %r4,8(%sp) ; save r4
1358 STD %r5,16(%sp) ; save r5
1359 STD %r6,24(%sp) ; save r6
1360 FSTD %fr12,32(%sp) ; save r6
1361 FSTD %fr13,40(%sp) ; save r7
1362
1363 ;
1364 ; Zero out carries
1365 ;
1366 COPY %r0,c1
1367 COPY %r0,c2
1368 COPY %r0,c3
1369
1370 LDO 128(%sp),%sp ; bump stack
1371 DEPDI,Z 1,31,1,high_one ; Create Value 1 << 32
1372
1373 ;
1374 ; Load up all of the values we are going to use
1375 ;
1376 FLDD 0(a_ptr),a0
1377 FLDD 8(a_ptr),a1
1378 FLDD 16(a_ptr),a2
1379 FLDD 24(a_ptr),a3
1380 FLDD 32(a_ptr),a4
1381 FLDD 40(a_ptr),a5
1382 FLDD 48(a_ptr),a6
1383 FLDD 56(a_ptr),a7
1384
1385 FLDD 0(b_ptr),b0
1386 FLDD 8(b_ptr),b1
1387 FLDD 16(b_ptr),b2
1388 FLDD 24(b_ptr),b3
1389 FLDD 32(b_ptr),b4
1390 FLDD 40(b_ptr),b5
1391 FLDD 48(b_ptr),b6
1392 FLDD 56(b_ptr),b7
1393
1394 MUL_ADD_C a0L,a0R,b0L,b0R,c1,c2,c3
1395 STD c1,0(r_ptr)
1396 COPY %r0,c1
1397
1398 MUL_ADD_C a0L,a0R,b1L,b1R,c2,c3,c1
1399 MUL_ADD_C a1L,a1R,b0L,b0R,c2,c3,c1
1400 STD c2,8(r_ptr)
1401 COPY %r0,c2
1402
1403 MUL_ADD_C a2L,a2R,b0L,b0R,c3,c1,c2
1404 MUL_ADD_C a1L,a1R,b1L,b1R,c3,c1,c2
1405 MUL_ADD_C a0L,a0R,b2L,b2R,c3,c1,c2
1406 STD c3,16(r_ptr)
1407 COPY %r0,c3
1408
1409 MUL_ADD_C a0L,a0R,b3L,b3R,c1,c2,c3
1410 MUL_ADD_C a1L,a1R,b2L,b2R,c1,c2,c3
1411 MUL_ADD_C a2L,a2R,b1L,b1R,c1,c2,c3
1412 MUL_ADD_C a3L,a3R,b0L,b0R,c1,c2,c3
1413 STD c1,24(r_ptr)
1414 COPY %r0,c1
1415
1416 MUL_ADD_C a4L,a4R,b0L,b0R,c2,c3,c1
1417 MUL_ADD_C a3L,a3R,b1L,b1R,c2,c3,c1
1418 MUL_ADD_C a2L,a2R,b2L,b2R,c2,c3,c1
1419 MUL_ADD_C a1L,a1R,b3L,b3R,c2,c3,c1
1420 MUL_ADD_C a0L,a0R,b4L,b4R,c2,c3,c1
1421 STD c2,32(r_ptr)
1422 COPY %r0,c2
1423
1424 MUL_ADD_C a0L,a0R,b5L,b5R,c3,c1,c2
1425 MUL_ADD_C a1L,a1R,b4L,b4R,c3,c1,c2
1426 MUL_ADD_C a2L,a2R,b3L,b3R,c3,c1,c2
1427 MUL_ADD_C a3L,a3R,b2L,b2R,c3,c1,c2
1428 MUL_ADD_C a4L,a4R,b1L,b1R,c3,c1,c2
1429 MUL_ADD_C a5L,a5R,b0L,b0R,c3,c1,c2
1430 STD c3,40(r_ptr)
1431 COPY %r0,c3
1432
1433 MUL_ADD_C a6L,a6R,b0L,b0R,c1,c2,c3
1434 MUL_ADD_C a5L,a5R,b1L,b1R,c1,c2,c3
1435 MUL_ADD_C a4L,a4R,b2L,b2R,c1,c2,c3
1436 MUL_ADD_C a3L,a3R,b3L,b3R,c1,c2,c3
1437 MUL_ADD_C a2L,a2R,b4L,b4R,c1,c2,c3
1438 MUL_ADD_C a1L,a1R,b5L,b5R,c1,c2,c3
1439 MUL_ADD_C a0L,a0R,b6L,b6R,c1,c2,c3
1440 STD c1,48(r_ptr)
1441 COPY %r0,c1
1442
1443 MUL_ADD_C a0L,a0R,b7L,b7R,c2,c3,c1
1444 MUL_ADD_C a1L,a1R,b6L,b6R,c2,c3,c1
1445 MUL_ADD_C a2L,a2R,b5L,b5R,c2,c3,c1
1446 MUL_ADD_C a3L,a3R,b4L,b4R,c2,c3,c1
1447 MUL_ADD_C a4L,a4R,b3L,b3R,c2,c3,c1
1448 MUL_ADD_C a5L,a5R,b2L,b2R,c2,c3,c1
1449 MUL_ADD_C a6L,a6R,b1L,b1R,c2,c3,c1
1450 MUL_ADD_C a7L,a7R,b0L,b0R,c2,c3,c1
1451 STD c2,56(r_ptr)
1452 COPY %r0,c2
1453
1454 MUL_ADD_C a7L,a7R,b1L,b1R,c3,c1,c2
1455 MUL_ADD_C a6L,a6R,b2L,b2R,c3,c1,c2
1456 MUL_ADD_C a5L,a5R,b3L,b3R,c3,c1,c2
1457 MUL_ADD_C a4L,a4R,b4L,b4R,c3,c1,c2
1458 MUL_ADD_C a3L,a3R,b5L,b5R,c3,c1,c2
1459 MUL_ADD_C a2L,a2R,b6L,b6R,c3,c1,c2
1460 MUL_ADD_C a1L,a1R,b7L,b7R,c3,c1,c2
1461 STD c3,64(r_ptr)
1462 COPY %r0,c3
1463
1464 MUL_ADD_C a2L,a2R,b7L,b7R,c1,c2,c3
1465 MUL_ADD_C a3L,a3R,b6L,b6R,c1,c2,c3
1466 MUL_ADD_C a4L,a4R,b5L,b5R,c1,c2,c3
1467 MUL_ADD_C a5L,a5R,b4L,b4R,c1,c2,c3
1468 MUL_ADD_C a6L,a6R,b3L,b3R,c1,c2,c3
1469 MUL_ADD_C a7L,a7R,b2L,b2R,c1,c2,c3
1470 STD c1,72(r_ptr)
1471 COPY %r0,c1
1472
1473 MUL_ADD_C a7L,a7R,b3L,b3R,c2,c3,c1
1474 MUL_ADD_C a6L,a6R,b4L,b4R,c2,c3,c1
1475 MUL_ADD_C a5L,a5R,b5L,b5R,c2,c3,c1
1476 MUL_ADD_C a4L,a4R,b6L,b6R,c2,c3,c1
1477 MUL_ADD_C a3L,a3R,b7L,b7R,c2,c3,c1
1478 STD c2,80(r_ptr)
1479 COPY %r0,c2
1480
1481 MUL_ADD_C a4L,a4R,b7L,b7R,c3,c1,c2
1482 MUL_ADD_C a5L,a5R,b6L,b6R,c3,c1,c2
1483 MUL_ADD_C a6L,a6R,b5L,b5R,c3,c1,c2
1484 MUL_ADD_C a7L,a7R,b4L,b4R,c3,c1,c2
1485 STD c3,88(r_ptr)
1486 COPY %r0,c3
1487
1488 MUL_ADD_C a7L,a7R,b5L,b5R,c1,c2,c3
1489 MUL_ADD_C a6L,a6R,b6L,b6R,c1,c2,c3
1490 MUL_ADD_C a5L,a5R,b7L,b7R,c1,c2,c3
1491 STD c1,96(r_ptr)
1492 COPY %r0,c1
1493
1494 MUL_ADD_C a6L,a6R,b7L,b7R,c2,c3,c1
1495 MUL_ADD_C a7L,a7R,b6L,b6R,c2,c3,c1
1496 STD c2,104(r_ptr)
1497 COPY %r0,c2
1498
1499 MUL_ADD_C a7L,a7R,b7L,b7R,c3,c1,c2
1500 STD c3,112(r_ptr)
1501 STD c1,120(r_ptr)
1502
1503 .EXIT
1504 FLDD -88(%sp),%fr13
1505 FLDD -96(%sp),%fr12
1506 LDD -104(%sp),%r6 ; restore r6
1507 LDD -112(%sp),%r5 ; restore r5
1508 LDD -120(%sp),%r4 ; restore r4
1509 BVE (%rp)
1510 LDD,MB -128(%sp),%r3
1511
1512 .PROCEND
1513
1514;-----------------------------------------------------------------------------
1515;
1516;void bn_mul_comba4(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b)
1517; arg0 = r_ptr
1518; arg1 = a_ptr
1519; arg2 = b_ptr
1520;
1521
1522bn_mul_comba4
1523 .proc
1524 .callinfo FRAME=128,ENTRY_GR=%r3,ARGS_SAVED,ORDERING_AWARE
1525 .EXPORT bn_mul_comba4,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
1526 .entry
1527 .align 64
1528
1529 STD %r3,0(%sp) ; save r3
1530 STD %r4,8(%sp) ; save r4
1531 STD %r5,16(%sp) ; save r5
1532 STD %r6,24(%sp) ; save r6
1533 FSTD %fr12,32(%sp) ; save r6
1534 FSTD %fr13,40(%sp) ; save r7
1535
1536 ;
1537 ; Zero out carries
1538 ;
1539 COPY %r0,c1
1540 COPY %r0,c2
1541 COPY %r0,c3
1542
1543 LDO 128(%sp),%sp ; bump stack
1544 DEPDI,Z 1,31,1,high_one ; Create Value 1 << 32
1545
1546 ;
1547 ; Load up all of the values we are going to use
1548 ;
1549 FLDD 0(a_ptr),a0
1550 FLDD 8(a_ptr),a1
1551 FLDD 16(a_ptr),a2
1552 FLDD 24(a_ptr),a3
1553
1554 FLDD 0(b_ptr),b0
1555 FLDD 8(b_ptr),b1
1556 FLDD 16(b_ptr),b2
1557 FLDD 24(b_ptr),b3
1558
1559 MUL_ADD_C a0L,a0R,b0L,b0R,c1,c2,c3
1560 STD c1,0(r_ptr)
1561 COPY %r0,c1
1562
1563 MUL_ADD_C a0L,a0R,b1L,b1R,c2,c3,c1
1564 MUL_ADD_C a1L,a1R,b0L,b0R,c2,c3,c1
1565 STD c2,8(r_ptr)
1566 COPY %r0,c2
1567
1568 MUL_ADD_C a2L,a2R,b0L,b0R,c3,c1,c2
1569 MUL_ADD_C a1L,a1R,b1L,b1R,c3,c1,c2
1570 MUL_ADD_C a0L,a0R,b2L,b2R,c3,c1,c2
1571 STD c3,16(r_ptr)
1572 COPY %r0,c3
1573
1574 MUL_ADD_C a0L,a0R,b3L,b3R,c1,c2,c3
1575 MUL_ADD_C a1L,a1R,b2L,b2R,c1,c2,c3
1576 MUL_ADD_C a2L,a2R,b1L,b1R,c1,c2,c3
1577 MUL_ADD_C a3L,a3R,b0L,b0R,c1,c2,c3
1578 STD c1,24(r_ptr)
1579 COPY %r0,c1
1580
1581 MUL_ADD_C a3L,a3R,b1L,b1R,c2,c3,c1
1582 MUL_ADD_C a2L,a2R,b2L,b2R,c2,c3,c1
1583 MUL_ADD_C a1L,a1R,b3L,b3R,c2,c3,c1
1584 STD c2,32(r_ptr)
1585 COPY %r0,c2
1586
1587 MUL_ADD_C a2L,a2R,b3L,b3R,c3,c1,c2
1588 MUL_ADD_C a3L,a3R,b2L,b2R,c3,c1,c2
1589 STD c3,40(r_ptr)
1590 COPY %r0,c3
1591
1592 MUL_ADD_C a3L,a3R,b3L,b3R,c1,c2,c3
1593 STD c1,48(r_ptr)
1594 STD c2,56(r_ptr)
1595
1596 .EXIT
1597 FLDD -88(%sp),%fr13
1598 FLDD -96(%sp),%fr12
1599 LDD -104(%sp),%r6 ; restore r6
1600 LDD -112(%sp),%r5 ; restore r5
1601 LDD -120(%sp),%r4 ; restore r4
1602 BVE (%rp)
1603 LDD,MB -128(%sp),%r3
1604
1605 .PROCEND
1606
1607
1608;--- not PIC .SPACE $TEXT$
1609;--- not PIC .SUBSPA $CODE$
1610;--- not PIC .SPACE $PRIVATE$,SORT=16
1611;--- not PIC .IMPORT $global$,DATA
1612;--- not PIC .SPACE $TEXT$
1613;--- not PIC .SUBSPA $CODE$
1614;--- not PIC .SUBSPA $LIT$,ACCESS=0x2c
1615;--- not PIC C$7
1616;--- not PIC .ALIGN 8
1617;--- not PIC .STRINGZ "Division would overflow (%d)\n"
1618 .END
diff --git a/src/lib/libcrypto/bn/asm/pa-risc2W.s b/src/lib/libcrypto/bn/asm/pa-risc2W.s
deleted file mode 100644
index a99545754d..0000000000
--- a/src/lib/libcrypto/bn/asm/pa-risc2W.s
+++ /dev/null
@@ -1,1605 +0,0 @@
1;
2; PA-RISC 64-bit implementation of bn_asm code
3;
4; This code is approximately 2x faster than the C version
5; for RSA/DSA.
6;
7; See http://devresource.hp.com/ for more details on the PA-RISC
8; architecture. Also see the book "PA-RISC 2.0 Architecture"
9; by Gerry Kane for information on the instruction set architecture.
10;
11; Code written by Chris Ruemmler (with some help from the HP C
12; compiler).
13;
14; The code compiles with HP's assembler
15;
16
17 .level 2.0W
18 .space $TEXT$
19 .subspa $CODE$,QUAD=0,ALIGN=8,ACCESS=0x2c,CODE_ONLY
20
21;
22; Global Register definitions used for the routines.
23;
24; Some information about HP's runtime architecture for 64-bits.
25;
26; "Caller save" means the calling function must save the register
27; if it wants the register to be preserved.
28; "Callee save" means if a function uses the register, it must save
29; the value before using it.
30;
31; For the floating point registers
32;
33; "caller save" registers: fr4-fr11, fr22-fr31
34; "callee save" registers: fr12-fr21
35; "special" registers: fr0-fr3 (status and exception registers)
36;
37; For the integer registers
38; value zero : r0
39; "caller save" registers: r1,r19-r26
40; "callee save" registers: r3-r18
41; return register : r2 (rp)
42; return values ; r28 (ret0,ret1)
43; Stack pointer ; r30 (sp)
44; global data pointer ; r27 (dp)
45; argument pointer ; r29 (ap)
46; millicode return ptr ; r31 (also a caller save register)
47
48
49;
50; Arguments to the routines
51;
52r_ptr .reg %r26
53a_ptr .reg %r25
54b_ptr .reg %r24
55num .reg %r24
56w .reg %r23
57n .reg %r23
58
59
60;
61; Globals used in some routines
62;
63
64top_overflow .reg %r29
65high_mask .reg %r22 ; value 0xffffffff80000000L
66
67
68;------------------------------------------------------------------------------
69;
70; bn_mul_add_words
71;
72;BN_ULONG bn_mul_add_words(BN_ULONG *r_ptr, BN_ULONG *a_ptr,
73; int num, BN_ULONG w)
74;
75; arg0 = r_ptr
76; arg1 = a_ptr
77; arg2 = num
78; arg3 = w
79;
80; Local register definitions
81;
82
83fm1 .reg %fr22
84fm .reg %fr23
85ht_temp .reg %fr24
86ht_temp_1 .reg %fr25
87lt_temp .reg %fr26
88lt_temp_1 .reg %fr27
89fm1_1 .reg %fr28
90fm_1 .reg %fr29
91
92fw_h .reg %fr7L
93fw_l .reg %fr7R
94fw .reg %fr7
95
96fht_0 .reg %fr8L
97flt_0 .reg %fr8R
98t_float_0 .reg %fr8
99
100fht_1 .reg %fr9L
101flt_1 .reg %fr9R
102t_float_1 .reg %fr9
103
104tmp_0 .reg %r31
105tmp_1 .reg %r21
106m_0 .reg %r20
107m_1 .reg %r19
108ht_0 .reg %r1
109ht_1 .reg %r3
110lt_0 .reg %r4
111lt_1 .reg %r5
112m1_0 .reg %r6
113m1_1 .reg %r7
114rp_val .reg %r8
115rp_val_1 .reg %r9
116
117bn_mul_add_words
118 .export bn_mul_add_words,entry,NO_RELOCATION,LONG_RETURN
119 .proc
120 .callinfo frame=128
121 .entry
122 .align 64
123
124 STD %r3,0(%sp) ; save r3
125 STD %r4,8(%sp) ; save r4
126 NOP ; Needed to make the loop 16-byte aligned
127 NOP ; Needed to make the loop 16-byte aligned
128
129 STD %r5,16(%sp) ; save r5
130 STD %r6,24(%sp) ; save r6
131 STD %r7,32(%sp) ; save r7
132 STD %r8,40(%sp) ; save r8
133
134 STD %r9,48(%sp) ; save r9
135 COPY %r0,%ret0 ; return 0 by default
136 DEPDI,Z 1,31,1,top_overflow ; top_overflow = 1 << 32
137 STD w,56(%sp) ; store w on stack
138
139 CMPIB,>= 0,num,bn_mul_add_words_exit ; if (num <= 0) then exit
140 LDO 128(%sp),%sp ; bump stack
141
142 ;
143 ; The loop is unrolled twice, so if there is only 1 number
144 ; then go straight to the cleanup code.
145 ;
146 CMPIB,= 1,num,bn_mul_add_words_single_top
147 FLDD -72(%sp),fw ; load up w into fp register fw (fw_h/fw_l)
148
149 ;
150 ; This loop is unrolled 2 times (64-byte aligned as well)
151 ;
152 ; PA-RISC 2.0 chips have two fully pipelined multipliers, thus
153 ; two 32-bit mutiplies can be issued per cycle.
154 ;
155bn_mul_add_words_unroll2
156
157 FLDD 0(a_ptr),t_float_0 ; load up 64-bit value (fr8L) ht(L)/lt(R)
158 FLDD 8(a_ptr),t_float_1 ; load up 64-bit value (fr8L) ht(L)/lt(R)
159 LDD 0(r_ptr),rp_val ; rp[0]
160 LDD 8(r_ptr),rp_val_1 ; rp[1]
161
162 XMPYU fht_0,fw_l,fm1 ; m1[0] = fht_0*fw_l
163 XMPYU fht_1,fw_l,fm1_1 ; m1[1] = fht_1*fw_l
164 FSTD fm1,-16(%sp) ; -16(sp) = m1[0]
165 FSTD fm1_1,-48(%sp) ; -48(sp) = m1[1]
166
167 XMPYU flt_0,fw_h,fm ; m[0] = flt_0*fw_h
168 XMPYU flt_1,fw_h,fm_1 ; m[1] = flt_1*fw_h
169 FSTD fm,-8(%sp) ; -8(sp) = m[0]
170 FSTD fm_1,-40(%sp) ; -40(sp) = m[1]
171
172 XMPYU fht_0,fw_h,ht_temp ; ht_temp = fht_0*fw_h
173 XMPYU fht_1,fw_h,ht_temp_1 ; ht_temp_1 = fht_1*fw_h
174 FSTD ht_temp,-24(%sp) ; -24(sp) = ht_temp
175 FSTD ht_temp_1,-56(%sp) ; -56(sp) = ht_temp_1
176
177 XMPYU flt_0,fw_l,lt_temp ; lt_temp = lt*fw_l
178 XMPYU flt_1,fw_l,lt_temp_1 ; lt_temp = lt*fw_l
179 FSTD lt_temp,-32(%sp) ; -32(sp) = lt_temp
180 FSTD lt_temp_1,-64(%sp) ; -64(sp) = lt_temp_1
181
182 LDD -8(%sp),m_0 ; m[0]
183 LDD -40(%sp),m_1 ; m[1]
184 LDD -16(%sp),m1_0 ; m1[0]
185 LDD -48(%sp),m1_1 ; m1[1]
186
187 LDD -24(%sp),ht_0 ; ht[0]
188 LDD -56(%sp),ht_1 ; ht[1]
189 ADD,L m1_0,m_0,tmp_0 ; tmp_0 = m[0] + m1[0];
190 ADD,L m1_1,m_1,tmp_1 ; tmp_1 = m[1] + m1[1];
191
192 LDD -32(%sp),lt_0
193 LDD -64(%sp),lt_1
194 CMPCLR,*>>= tmp_0,m1_0, %r0 ; if (m[0] < m1[0])
195 ADD,L ht_0,top_overflow,ht_0 ; ht[0] += (1<<32)
196
197 CMPCLR,*>>= tmp_1,m1_1,%r0 ; if (m[1] < m1[1])
198 ADD,L ht_1,top_overflow,ht_1 ; ht[1] += (1<<32)
199 EXTRD,U tmp_0,31,32,m_0 ; m[0]>>32
200 DEPD,Z tmp_0,31,32,m1_0 ; m1[0] = m[0]<<32
201
202 EXTRD,U tmp_1,31,32,m_1 ; m[1]>>32
203 DEPD,Z tmp_1,31,32,m1_1 ; m1[1] = m[1]<<32
204 ADD,L ht_0,m_0,ht_0 ; ht[0]+= (m[0]>>32)
205 ADD,L ht_1,m_1,ht_1 ; ht[1]+= (m[1]>>32)
206
207 ADD lt_0,m1_0,lt_0 ; lt[0] = lt[0]+m1[0];
208 ADD,DC ht_0,%r0,ht_0 ; ht[0]++
209 ADD lt_1,m1_1,lt_1 ; lt[1] = lt[1]+m1[1];
210 ADD,DC ht_1,%r0,ht_1 ; ht[1]++
211
212 ADD %ret0,lt_0,lt_0 ; lt[0] = lt[0] + c;
213 ADD,DC ht_0,%r0,ht_0 ; ht[0]++
214 ADD lt_0,rp_val,lt_0 ; lt[0] = lt[0]+rp[0]
215 ADD,DC ht_0,%r0,ht_0 ; ht[0]++
216
217 LDO -2(num),num ; num = num - 2;
218 ADD ht_0,lt_1,lt_1 ; lt[1] = lt[1] + ht_0 (c);
219 ADD,DC ht_1,%r0,ht_1 ; ht[1]++
220 STD lt_0,0(r_ptr) ; rp[0] = lt[0]
221
222 ADD lt_1,rp_val_1,lt_1 ; lt[1] = lt[1]+rp[1]
223 ADD,DC ht_1,%r0,%ret0 ; ht[1]++
224 LDO 16(a_ptr),a_ptr ; a_ptr += 2
225
226 STD lt_1,8(r_ptr) ; rp[1] = lt[1]
227 CMPIB,<= 2,num,bn_mul_add_words_unroll2 ; go again if more to do
228 LDO 16(r_ptr),r_ptr ; r_ptr += 2
229
230 CMPIB,=,N 0,num,bn_mul_add_words_exit ; are we done, or cleanup last one
231
232 ;
233 ; Top of loop aligned on 64-byte boundary
234 ;
235bn_mul_add_words_single_top
236 FLDD 0(a_ptr),t_float_0 ; load up 64-bit value (fr8L) ht(L)/lt(R)
237 LDD 0(r_ptr),rp_val ; rp[0]
238 LDO 8(a_ptr),a_ptr ; a_ptr++
239 XMPYU fht_0,fw_l,fm1 ; m1 = ht*fw_l
240 FSTD fm1,-16(%sp) ; -16(sp) = m1
241 XMPYU flt_0,fw_h,fm ; m = lt*fw_h
242 FSTD fm,-8(%sp) ; -8(sp) = m
243 XMPYU fht_0,fw_h,ht_temp ; ht_temp = ht*fw_h
244 FSTD ht_temp,-24(%sp) ; -24(sp) = ht
245 XMPYU flt_0,fw_l,lt_temp ; lt_temp = lt*fw_l
246 FSTD lt_temp,-32(%sp) ; -32(sp) = lt
247
248 LDD -8(%sp),m_0
249 LDD -16(%sp),m1_0 ; m1 = temp1
250 ADD,L m_0,m1_0,tmp_0 ; tmp_0 = m + m1;
251 LDD -24(%sp),ht_0
252 LDD -32(%sp),lt_0
253
254 CMPCLR,*>>= tmp_0,m1_0,%r0 ; if (m < m1)
255 ADD,L ht_0,top_overflow,ht_0 ; ht += (1<<32)
256
257 EXTRD,U tmp_0,31,32,m_0 ; m>>32
258 DEPD,Z tmp_0,31,32,m1_0 ; m1 = m<<32
259
260 ADD,L ht_0,m_0,ht_0 ; ht+= (m>>32)
261 ADD lt_0,m1_0,tmp_0 ; tmp_0 = lt+m1;
262 ADD,DC ht_0,%r0,ht_0 ; ht++
263 ADD %ret0,tmp_0,lt_0 ; lt = lt + c;
264 ADD,DC ht_0,%r0,ht_0 ; ht++
265 ADD lt_0,rp_val,lt_0 ; lt = lt+rp[0]
266 ADD,DC ht_0,%r0,%ret0 ; ht++
267 STD lt_0,0(r_ptr) ; rp[0] = lt
268
269bn_mul_add_words_exit
270 .EXIT
271 LDD -80(%sp),%r9 ; restore r9
272 LDD -88(%sp),%r8 ; restore r8
273 LDD -96(%sp),%r7 ; restore r7
274 LDD -104(%sp),%r6 ; restore r6
275 LDD -112(%sp),%r5 ; restore r5
276 LDD -120(%sp),%r4 ; restore r4
277 BVE (%rp)
278 LDD,MB -128(%sp),%r3 ; restore r3
279 .PROCEND ;in=23,24,25,26,29;out=28;
280
281;----------------------------------------------------------------------------
282;
283;BN_ULONG bn_mul_words(BN_ULONG *rp, BN_ULONG *ap, int num, BN_ULONG w)
284;
285; arg0 = rp
286; arg1 = ap
287; arg2 = num
288; arg3 = w
289
290bn_mul_words
291 .proc
292 .callinfo frame=128
293 .entry
294 .EXPORT bn_mul_words,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
295 .align 64
296
297 STD %r3,0(%sp) ; save r3
298 STD %r4,8(%sp) ; save r4
299 STD %r5,16(%sp) ; save r5
300 STD %r6,24(%sp) ; save r6
301
302 STD %r7,32(%sp) ; save r7
303 COPY %r0,%ret0 ; return 0 by default
304 DEPDI,Z 1,31,1,top_overflow ; top_overflow = 1 << 32
305 STD w,56(%sp) ; w on stack
306
307 CMPIB,>= 0,num,bn_mul_words_exit
308 LDO 128(%sp),%sp ; bump stack
309
310 ;
311 ; See if only 1 word to do, thus just do cleanup
312 ;
313 CMPIB,= 1,num,bn_mul_words_single_top
314 FLDD -72(%sp),fw ; load up w into fp register fw (fw_h/fw_l)
315
316 ;
317 ; This loop is unrolled 2 times (64-byte aligned as well)
318 ;
319 ; PA-RISC 2.0 chips have two fully pipelined multipliers, thus
320 ; two 32-bit mutiplies can be issued per cycle.
321 ;
322bn_mul_words_unroll2
323
324 FLDD 0(a_ptr),t_float_0 ; load up 64-bit value (fr8L) ht(L)/lt(R)
325 FLDD 8(a_ptr),t_float_1 ; load up 64-bit value (fr8L) ht(L)/lt(R)
326 XMPYU fht_0,fw_l,fm1 ; m1[0] = fht_0*fw_l
327 XMPYU fht_1,fw_l,fm1_1 ; m1[1] = ht*fw_l
328
329 FSTD fm1,-16(%sp) ; -16(sp) = m1
330 FSTD fm1_1,-48(%sp) ; -48(sp) = m1
331 XMPYU flt_0,fw_h,fm ; m = lt*fw_h
332 XMPYU flt_1,fw_h,fm_1 ; m = lt*fw_h
333
334 FSTD fm,-8(%sp) ; -8(sp) = m
335 FSTD fm_1,-40(%sp) ; -40(sp) = m
336 XMPYU fht_0,fw_h,ht_temp ; ht_temp = fht_0*fw_h
337 XMPYU fht_1,fw_h,ht_temp_1 ; ht_temp = ht*fw_h
338
339 FSTD ht_temp,-24(%sp) ; -24(sp) = ht
340 FSTD ht_temp_1,-56(%sp) ; -56(sp) = ht
341 XMPYU flt_0,fw_l,lt_temp ; lt_temp = lt*fw_l
342 XMPYU flt_1,fw_l,lt_temp_1 ; lt_temp = lt*fw_l
343
344 FSTD lt_temp,-32(%sp) ; -32(sp) = lt
345 FSTD lt_temp_1,-64(%sp) ; -64(sp) = lt
346 LDD -8(%sp),m_0
347 LDD -40(%sp),m_1
348
349 LDD -16(%sp),m1_0
350 LDD -48(%sp),m1_1
351 LDD -24(%sp),ht_0
352 LDD -56(%sp),ht_1
353
354 ADD,L m1_0,m_0,tmp_0 ; tmp_0 = m + m1;
355 ADD,L m1_1,m_1,tmp_1 ; tmp_1 = m + m1;
356 LDD -32(%sp),lt_0
357 LDD -64(%sp),lt_1
358
359 CMPCLR,*>>= tmp_0,m1_0, %r0 ; if (m < m1)
360 ADD,L ht_0,top_overflow,ht_0 ; ht += (1<<32)
361 CMPCLR,*>>= tmp_1,m1_1,%r0 ; if (m < m1)
362 ADD,L ht_1,top_overflow,ht_1 ; ht += (1<<32)
363
364 EXTRD,U tmp_0,31,32,m_0 ; m>>32
365 DEPD,Z tmp_0,31,32,m1_0 ; m1 = m<<32
366 EXTRD,U tmp_1,31,32,m_1 ; m>>32
367 DEPD,Z tmp_1,31,32,m1_1 ; m1 = m<<32
368
369 ADD,L ht_0,m_0,ht_0 ; ht+= (m>>32)
370 ADD,L ht_1,m_1,ht_1 ; ht+= (m>>32)
371 ADD lt_0,m1_0,lt_0 ; lt = lt+m1;
372 ADD,DC ht_0,%r0,ht_0 ; ht++
373
374 ADD lt_1,m1_1,lt_1 ; lt = lt+m1;
375 ADD,DC ht_1,%r0,ht_1 ; ht++
376 ADD %ret0,lt_0,lt_0 ; lt = lt + c (ret0);
377 ADD,DC ht_0,%r0,ht_0 ; ht++
378
379 ADD ht_0,lt_1,lt_1 ; lt = lt + c (ht_0)
380 ADD,DC ht_1,%r0,ht_1 ; ht++
381 STD lt_0,0(r_ptr) ; rp[0] = lt
382 STD lt_1,8(r_ptr) ; rp[1] = lt
383
384 COPY ht_1,%ret0 ; carry = ht
385 LDO -2(num),num ; num = num - 2;
386 LDO 16(a_ptr),a_ptr ; ap += 2
387 CMPIB,<= 2,num,bn_mul_words_unroll2
388 LDO 16(r_ptr),r_ptr ; rp++
389
390 CMPIB,=,N 0,num,bn_mul_words_exit ; are we done?
391
392 ;
393 ; Top of loop aligned on 64-byte boundary
394 ;
395bn_mul_words_single_top
396 FLDD 0(a_ptr),t_float_0 ; load up 64-bit value (fr8L) ht(L)/lt(R)
397
398 XMPYU fht_0,fw_l,fm1 ; m1 = ht*fw_l
399 FSTD fm1,-16(%sp) ; -16(sp) = m1
400 XMPYU flt_0,fw_h,fm ; m = lt*fw_h
401 FSTD fm,-8(%sp) ; -8(sp) = m
402 XMPYU fht_0,fw_h,ht_temp ; ht_temp = ht*fw_h
403 FSTD ht_temp,-24(%sp) ; -24(sp) = ht
404 XMPYU flt_0,fw_l,lt_temp ; lt_temp = lt*fw_l
405 FSTD lt_temp,-32(%sp) ; -32(sp) = lt
406
407 LDD -8(%sp),m_0
408 LDD -16(%sp),m1_0
409 ADD,L m_0,m1_0,tmp_0 ; tmp_0 = m + m1;
410 LDD -24(%sp),ht_0
411 LDD -32(%sp),lt_0
412
413 CMPCLR,*>>= tmp_0,m1_0,%r0 ; if (m < m1)
414 ADD,L ht_0,top_overflow,ht_0 ; ht += (1<<32)
415
416 EXTRD,U tmp_0,31,32,m_0 ; m>>32
417 DEPD,Z tmp_0,31,32,m1_0 ; m1 = m<<32
418
419 ADD,L ht_0,m_0,ht_0 ; ht+= (m>>32)
420 ADD lt_0,m1_0,lt_0 ; lt= lt+m1;
421 ADD,DC ht_0,%r0,ht_0 ; ht++
422
423 ADD %ret0,lt_0,lt_0 ; lt = lt + c;
424 ADD,DC ht_0,%r0,ht_0 ; ht++
425
426 COPY ht_0,%ret0 ; copy carry
427 STD lt_0,0(r_ptr) ; rp[0] = lt
428
429bn_mul_words_exit
430 .EXIT
431 LDD -96(%sp),%r7 ; restore r7
432 LDD -104(%sp),%r6 ; restore r6
433 LDD -112(%sp),%r5 ; restore r5
434 LDD -120(%sp),%r4 ; restore r4
435 BVE (%rp)
436 LDD,MB -128(%sp),%r3 ; restore r3
437 .PROCEND ;in=23,24,25,26,29;out=28;
438
439;----------------------------------------------------------------------------
440;
441;void bn_sqr_words(BN_ULONG *rp, BN_ULONG *ap, int num)
442;
443; arg0 = rp
444; arg1 = ap
445; arg2 = num
446;
447
448bn_sqr_words
449 .proc
450 .callinfo FRAME=128,ENTRY_GR=%r3,ARGS_SAVED,ORDERING_AWARE
451 .EXPORT bn_sqr_words,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
452 .entry
453 .align 64
454
455 STD %r3,0(%sp) ; save r3
456 STD %r4,8(%sp) ; save r4
457 NOP
458 STD %r5,16(%sp) ; save r5
459
460 CMPIB,>= 0,num,bn_sqr_words_exit
461 LDO 128(%sp),%sp ; bump stack
462
463 ;
464 ; If only 1, the goto straight to cleanup
465 ;
466 CMPIB,= 1,num,bn_sqr_words_single_top
467 DEPDI,Z -1,32,33,high_mask ; Create Mask 0xffffffff80000000L
468
469 ;
470 ; This loop is unrolled 2 times (64-byte aligned as well)
471 ;
472
473bn_sqr_words_unroll2
474 FLDD 0(a_ptr),t_float_0 ; a[0]
475 FLDD 8(a_ptr),t_float_1 ; a[1]
476 XMPYU fht_0,flt_0,fm ; m[0]
477 XMPYU fht_1,flt_1,fm_1 ; m[1]
478
479 FSTD fm,-24(%sp) ; store m[0]
480 FSTD fm_1,-56(%sp) ; store m[1]
481 XMPYU flt_0,flt_0,lt_temp ; lt[0]
482 XMPYU flt_1,flt_1,lt_temp_1 ; lt[1]
483
484 FSTD lt_temp,-16(%sp) ; store lt[0]
485 FSTD lt_temp_1,-48(%sp) ; store lt[1]
486 XMPYU fht_0,fht_0,ht_temp ; ht[0]
487 XMPYU fht_1,fht_1,ht_temp_1 ; ht[1]
488
489 FSTD ht_temp,-8(%sp) ; store ht[0]
490 FSTD ht_temp_1,-40(%sp) ; store ht[1]
491 LDD -24(%sp),m_0
492 LDD -56(%sp),m_1
493
494 AND m_0,high_mask,tmp_0 ; m[0] & Mask
495 AND m_1,high_mask,tmp_1 ; m[1] & Mask
496 DEPD,Z m_0,30,31,m_0 ; m[0] << 32+1
497 DEPD,Z m_1,30,31,m_1 ; m[1] << 32+1
498
499 LDD -16(%sp),lt_0
500 LDD -48(%sp),lt_1
501 EXTRD,U tmp_0,32,33,tmp_0 ; tmp_0 = m[0]&Mask >> 32-1
502 EXTRD,U tmp_1,32,33,tmp_1 ; tmp_1 = m[1]&Mask >> 32-1
503
504 LDD -8(%sp),ht_0
505 LDD -40(%sp),ht_1
506 ADD,L ht_0,tmp_0,ht_0 ; ht[0] += tmp_0
507 ADD,L ht_1,tmp_1,ht_1 ; ht[1] += tmp_1
508
509 ADD lt_0,m_0,lt_0 ; lt = lt+m
510 ADD,DC ht_0,%r0,ht_0 ; ht[0]++
511 STD lt_0,0(r_ptr) ; rp[0] = lt[0]
512 STD ht_0,8(r_ptr) ; rp[1] = ht[1]
513
514 ADD lt_1,m_1,lt_1 ; lt = lt+m
515 ADD,DC ht_1,%r0,ht_1 ; ht[1]++
516 STD lt_1,16(r_ptr) ; rp[2] = lt[1]
517 STD ht_1,24(r_ptr) ; rp[3] = ht[1]
518
519 LDO -2(num),num ; num = num - 2;
520 LDO 16(a_ptr),a_ptr ; ap += 2
521 CMPIB,<= 2,num,bn_sqr_words_unroll2
522 LDO 32(r_ptr),r_ptr ; rp += 4
523
524 CMPIB,=,N 0,num,bn_sqr_words_exit ; are we done?
525
526 ;
527 ; Top of loop aligned on 64-byte boundary
528 ;
529bn_sqr_words_single_top
530 FLDD 0(a_ptr),t_float_0 ; load up 64-bit value (fr8L) ht(L)/lt(R)
531
532 XMPYU fht_0,flt_0,fm ; m
533 FSTD fm,-24(%sp) ; store m
534
535 XMPYU flt_0,flt_0,lt_temp ; lt
536 FSTD lt_temp,-16(%sp) ; store lt
537
538 XMPYU fht_0,fht_0,ht_temp ; ht
539 FSTD ht_temp,-8(%sp) ; store ht
540
541 LDD -24(%sp),m_0 ; load m
542 AND m_0,high_mask,tmp_0 ; m & Mask
543 DEPD,Z m_0,30,31,m_0 ; m << 32+1
544 LDD -16(%sp),lt_0 ; lt
545
546 LDD -8(%sp),ht_0 ; ht
547 EXTRD,U tmp_0,32,33,tmp_0 ; tmp_0 = m&Mask >> 32-1
548 ADD m_0,lt_0,lt_0 ; lt = lt+m
549 ADD,L ht_0,tmp_0,ht_0 ; ht += tmp_0
550 ADD,DC ht_0,%r0,ht_0 ; ht++
551
552 STD lt_0,0(r_ptr) ; rp[0] = lt
553 STD ht_0,8(r_ptr) ; rp[1] = ht
554
555bn_sqr_words_exit
556 .EXIT
557 LDD -112(%sp),%r5 ; restore r5
558 LDD -120(%sp),%r4 ; restore r4
559 BVE (%rp)
560 LDD,MB -128(%sp),%r3
561 .PROCEND ;in=23,24,25,26,29;out=28;
562
563
564;----------------------------------------------------------------------------
565;
566;BN_ULONG bn_add_words(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n)
567;
568; arg0 = rp
569; arg1 = ap
570; arg2 = bp
571; arg3 = n
572
573t .reg %r22
574b .reg %r21
575l .reg %r20
576
577bn_add_words
578 .proc
579 .entry
580 .callinfo
581 .EXPORT bn_add_words,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
582 .align 64
583
584 CMPIB,>= 0,n,bn_add_words_exit
585 COPY %r0,%ret0 ; return 0 by default
586
587 ;
588 ; If 2 or more numbers do the loop
589 ;
590 CMPIB,= 1,n,bn_add_words_single_top
591 NOP
592
593 ;
594 ; This loop is unrolled 2 times (64-byte aligned as well)
595 ;
596bn_add_words_unroll2
597 LDD 0(a_ptr),t
598 LDD 0(b_ptr),b
599 ADD t,%ret0,t ; t = t+c;
600 ADD,DC %r0,%r0,%ret0 ; set c to carry
601 ADD t,b,l ; l = t + b[0]
602 ADD,DC %ret0,%r0,%ret0 ; c+= carry
603 STD l,0(r_ptr)
604
605 LDD 8(a_ptr),t
606 LDD 8(b_ptr),b
607 ADD t,%ret0,t ; t = t+c;
608 ADD,DC %r0,%r0,%ret0 ; set c to carry
609 ADD t,b,l ; l = t + b[0]
610 ADD,DC %ret0,%r0,%ret0 ; c+= carry
611 STD l,8(r_ptr)
612
613 LDO -2(n),n
614 LDO 16(a_ptr),a_ptr
615 LDO 16(b_ptr),b_ptr
616
617 CMPIB,<= 2,n,bn_add_words_unroll2
618 LDO 16(r_ptr),r_ptr
619
620 CMPIB,=,N 0,n,bn_add_words_exit ; are we done?
621
622bn_add_words_single_top
623 LDD 0(a_ptr),t
624 LDD 0(b_ptr),b
625
626 ADD t,%ret0,t ; t = t+c;
627 ADD,DC %r0,%r0,%ret0 ; set c to carry (could use CMPCLR??)
628 ADD t,b,l ; l = t + b[0]
629 ADD,DC %ret0,%r0,%ret0 ; c+= carry
630 STD l,0(r_ptr)
631
632bn_add_words_exit
633 .EXIT
634 BVE (%rp)
635 NOP
636 .PROCEND ;in=23,24,25,26,29;out=28;
637
638;----------------------------------------------------------------------------
639;
640;BN_ULONG bn_sub_words(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n)
641;
642; arg0 = rp
643; arg1 = ap
644; arg2 = bp
645; arg3 = n
646
647t1 .reg %r22
648t2 .reg %r21
649sub_tmp1 .reg %r20
650sub_tmp2 .reg %r19
651
652
653bn_sub_words
654 .proc
655 .callinfo
656 .EXPORT bn_sub_words,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
657 .entry
658 .align 64
659
660 CMPIB,>= 0,n,bn_sub_words_exit
661 COPY %r0,%ret0 ; return 0 by default
662
663 ;
664 ; If 2 or more numbers do the loop
665 ;
666 CMPIB,= 1,n,bn_sub_words_single_top
667 NOP
668
669 ;
670 ; This loop is unrolled 2 times (64-byte aligned as well)
671 ;
672bn_sub_words_unroll2
673 LDD 0(a_ptr),t1
674 LDD 0(b_ptr),t2
675 SUB t1,t2,sub_tmp1 ; t3 = t1-t2;
676 SUB sub_tmp1,%ret0,sub_tmp1 ; t3 = t3- c;
677
678 CMPCLR,*>> t1,t2,sub_tmp2 ; clear if t1 > t2
679 LDO 1(%r0),sub_tmp2
680
681 CMPCLR,*= t1,t2,%r0
682 COPY sub_tmp2,%ret0
683 STD sub_tmp1,0(r_ptr)
684
685 LDD 8(a_ptr),t1
686 LDD 8(b_ptr),t2
687 SUB t1,t2,sub_tmp1 ; t3 = t1-t2;
688 SUB sub_tmp1,%ret0,sub_tmp1 ; t3 = t3- c;
689 CMPCLR,*>> t1,t2,sub_tmp2 ; clear if t1 > t2
690 LDO 1(%r0),sub_tmp2
691
692 CMPCLR,*= t1,t2,%r0
693 COPY sub_tmp2,%ret0
694 STD sub_tmp1,8(r_ptr)
695
696 LDO -2(n),n
697 LDO 16(a_ptr),a_ptr
698 LDO 16(b_ptr),b_ptr
699
700 CMPIB,<= 2,n,bn_sub_words_unroll2
701 LDO 16(r_ptr),r_ptr
702
703 CMPIB,=,N 0,n,bn_sub_words_exit ; are we done?
704
705bn_sub_words_single_top
706 LDD 0(a_ptr),t1
707 LDD 0(b_ptr),t2
708 SUB t1,t2,sub_tmp1 ; t3 = t1-t2;
709 SUB sub_tmp1,%ret0,sub_tmp1 ; t3 = t3- c;
710 CMPCLR,*>> t1,t2,sub_tmp2 ; clear if t1 > t2
711 LDO 1(%r0),sub_tmp2
712
713 CMPCLR,*= t1,t2,%r0
714 COPY sub_tmp2,%ret0
715
716 STD sub_tmp1,0(r_ptr)
717
718bn_sub_words_exit
719 .EXIT
720 BVE (%rp)
721 NOP
722 .PROCEND ;in=23,24,25,26,29;out=28;
723
724;------------------------------------------------------------------------------
725;
726; unsigned long bn_div_words(unsigned long h, unsigned long l, unsigned long d)
727;
728; arg0 = h
729; arg1 = l
730; arg2 = d
731;
732; This is mainly just modified assembly from the compiler, thus the
733; lack of variable names.
734;
735;------------------------------------------------------------------------------
736bn_div_words
737 .proc
738 .callinfo CALLER,FRAME=272,ENTRY_GR=%r10,SAVE_RP,ARGS_SAVED,ORDERING_AWARE
739 .EXPORT bn_div_words,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
740 .IMPORT BN_num_bits_word,CODE,NO_RELOCATION
741 .IMPORT __iob,DATA
742 .IMPORT fprintf,CODE,NO_RELOCATION
743 .IMPORT abort,CODE,NO_RELOCATION
744 .IMPORT $$div2U,MILLICODE
745 .entry
746 STD %r2,-16(%r30)
747 STD,MA %r3,352(%r30)
748 STD %r4,-344(%r30)
749 STD %r5,-336(%r30)
750 STD %r6,-328(%r30)
751 STD %r7,-320(%r30)
752 STD %r8,-312(%r30)
753 STD %r9,-304(%r30)
754 STD %r10,-296(%r30)
755
756 STD %r27,-288(%r30) ; save gp
757
758 COPY %r24,%r3 ; save d
759 COPY %r26,%r4 ; save h (high 64-bits)
760 LDO -1(%r0),%ret0 ; return -1 by default
761
762 CMPB,*= %r0,%arg2,$D3 ; if (d == 0)
763 COPY %r25,%r5 ; save l (low 64-bits)
764
765 LDO -48(%r30),%r29 ; create ap
766 .CALL ;in=26,29;out=28;
767 B,L BN_num_bits_word,%r2
768 COPY %r3,%r26
769 LDD -288(%r30),%r27 ; restore gp
770 LDI 64,%r21
771
772 CMPB,= %r21,%ret0,$00000012 ;if (i == 64) (forward)
773 COPY %ret0,%r24 ; i
774 MTSARCM %r24
775 DEPDI,Z -1,%sar,1,%r29
776 CMPB,*<<,N %r29,%r4,bn_div_err_case ; if (h > 1<<i) (forward)
777
778$00000012
779 SUBI 64,%r24,%r31 ; i = 64 - i;
780 CMPCLR,*<< %r4,%r3,%r0 ; if (h >= d)
781 SUB %r4,%r3,%r4 ; h -= d
782 CMPB,= %r31,%r0,$0000001A ; if (i)
783 COPY %r0,%r10 ; ret = 0
784 MTSARCM %r31 ; i to shift
785 DEPD,Z %r3,%sar,64,%r3 ; d <<= i;
786 SUBI 64,%r31,%r19 ; 64 - i; redundent
787 MTSAR %r19 ; (64 -i) to shift
788 SHRPD %r4,%r5,%sar,%r4 ; l>> (64-i)
789 MTSARCM %r31 ; i to shift
790 DEPD,Z %r5,%sar,64,%r5 ; l <<= i;
791
792$0000001A
793 DEPDI,Z -1,31,32,%r19
794 EXTRD,U %r3,31,32,%r6 ; dh=(d&0xfff)>>32
795 EXTRD,U %r3,63,32,%r8 ; dl = d&0xffffff
796 LDO 2(%r0),%r9
797 STD %r3,-280(%r30) ; "d" to stack
798
799$0000001C
800 DEPDI,Z -1,63,32,%r29 ;
801 EXTRD,U %r4,31,32,%r31 ; h >> 32
802 CMPB,*=,N %r31,%r6,$D2 ; if ((h>>32) != dh)(forward) div
803 COPY %r4,%r26
804 EXTRD,U %r4,31,32,%r25
805 COPY %r6,%r24
806 .CALL ;in=23,24,25,26;out=20,21,22,28,29; (MILLICALL)
807 B,L $$div2U,%r2
808 EXTRD,U %r6,31,32,%r23
809 DEPD %r28,31,32,%r29
810$D2
811 STD %r29,-272(%r30) ; q
812 AND %r5,%r19,%r24 ; t & 0xffffffff00000000;
813 EXTRD,U %r24,31,32,%r24 ; ???
814 FLDD -272(%r30),%fr7 ; q
815 FLDD -280(%r30),%fr8 ; d
816 XMPYU %fr8L,%fr7L,%fr10
817 FSTD %fr10,-256(%r30)
818 XMPYU %fr8L,%fr7R,%fr22
819 FSTD %fr22,-264(%r30)
820 XMPYU %fr8R,%fr7L,%fr11
821 XMPYU %fr8R,%fr7R,%fr23
822 FSTD %fr11,-232(%r30)
823 FSTD %fr23,-240(%r30)
824 LDD -256(%r30),%r28
825 DEPD,Z %r28,31,32,%r2
826 LDD -264(%r30),%r20
827 ADD,L %r20,%r2,%r31
828 LDD -232(%r30),%r22
829 DEPD,Z %r22,31,32,%r22
830 LDD -240(%r30),%r21
831 B $00000024 ; enter loop
832 ADD,L %r21,%r22,%r23
833
834$0000002A
835 LDO -1(%r29),%r29
836 SUB %r23,%r8,%r23
837$00000024
838 SUB %r4,%r31,%r25
839 AND %r25,%r19,%r26
840 CMPB,*<>,N %r0,%r26,$00000046 ; (forward)
841 DEPD,Z %r25,31,32,%r20
842 OR %r20,%r24,%r21
843 CMPB,*<<,N %r21,%r23,$0000002A ;(backward)
844 SUB %r31,%r6,%r31
845;-------------Break path---------------------
846
847$00000046
848 DEPD,Z %r23,31,32,%r25 ;tl
849 EXTRD,U %r23,31,32,%r26 ;t
850 AND %r25,%r19,%r24 ;tl = (tl<<32)&0xfffffff0000000L
851 ADD,L %r31,%r26,%r31 ;th += t;
852 CMPCLR,*>>= %r5,%r24,%r0 ;if (l<tl)
853 LDO 1(%r31),%r31 ; th++;
854 CMPB,*<<=,N %r31,%r4,$00000036 ;if (n < th) (forward)
855 LDO -1(%r29),%r29 ;q--;
856 ADD,L %r4,%r3,%r4 ;h += d;
857$00000036
858 ADDIB,=,N -1,%r9,$D1 ;if (--count == 0) break (forward)
859 SUB %r5,%r24,%r28 ; l -= tl;
860 SUB %r4,%r31,%r24 ; h -= th;
861 SHRPD %r24,%r28,32,%r4 ; h = ((h<<32)|(l>>32));
862 DEPD,Z %r29,31,32,%r10 ; ret = q<<32
863 b $0000001C
864 DEPD,Z %r28,31,32,%r5 ; l = l << 32
865
866$D1
867 OR %r10,%r29,%r28 ; ret |= q
868$D3
869 LDD -368(%r30),%r2
870$D0
871 LDD -296(%r30),%r10
872 LDD -304(%r30),%r9
873 LDD -312(%r30),%r8
874 LDD -320(%r30),%r7
875 LDD -328(%r30),%r6
876 LDD -336(%r30),%r5
877 LDD -344(%r30),%r4
878 BVE (%r2)
879 .EXIT
880 LDD,MB -352(%r30),%r3
881
882bn_div_err_case
883 MFIA %r6
884 ADDIL L'bn_div_words-bn_div_err_case,%r6,%r1
885 LDO R'bn_div_words-bn_div_err_case(%r1),%r6
886 ADDIL LT'__iob,%r27,%r1
887 LDD RT'__iob(%r1),%r26
888 ADDIL L'C$4-bn_div_words,%r6,%r1
889 LDO R'C$4-bn_div_words(%r1),%r25
890 LDO 64(%r26),%r26
891 .CALL ;in=24,25,26,29;out=28;
892 B,L fprintf,%r2
893 LDO -48(%r30),%r29
894 LDD -288(%r30),%r27
895 .CALL ;in=29;
896 B,L abort,%r2
897 LDO -48(%r30),%r29
898 LDD -288(%r30),%r27
899 B $D0
900 LDD -368(%r30),%r2
901 .PROCEND ;in=24,25,26,29;out=28;
902
903;----------------------------------------------------------------------------
904;
905; Registers to hold 64-bit values to manipulate. The "L" part
906; of the register corresponds to the upper 32-bits, while the "R"
907; part corresponds to the lower 32-bits
908;
909; Note, that when using b6 and b7, the code must save these before
910; using them because they are callee save registers
911;
912;
913; Floating point registers to use to save values that
914; are manipulated. These don't collide with ftemp1-6 and
915; are all caller save registers
916;
917a0 .reg %fr22
918a0L .reg %fr22L
919a0R .reg %fr22R
920
921a1 .reg %fr23
922a1L .reg %fr23L
923a1R .reg %fr23R
924
925a2 .reg %fr24
926a2L .reg %fr24L
927a2R .reg %fr24R
928
929a3 .reg %fr25
930a3L .reg %fr25L
931a3R .reg %fr25R
932
933a4 .reg %fr26
934a4L .reg %fr26L
935a4R .reg %fr26R
936
937a5 .reg %fr27
938a5L .reg %fr27L
939a5R .reg %fr27R
940
941a6 .reg %fr28
942a6L .reg %fr28L
943a6R .reg %fr28R
944
945a7 .reg %fr29
946a7L .reg %fr29L
947a7R .reg %fr29R
948
949b0 .reg %fr30
950b0L .reg %fr30L
951b0R .reg %fr30R
952
953b1 .reg %fr31
954b1L .reg %fr31L
955b1R .reg %fr31R
956
957;
958; Temporary floating point variables, these are all caller save
959; registers
960;
961ftemp1 .reg %fr4
962ftemp2 .reg %fr5
963ftemp3 .reg %fr6
964ftemp4 .reg %fr7
965
966;
967; The B set of registers when used.
968;
969
970b2 .reg %fr8
971b2L .reg %fr8L
972b2R .reg %fr8R
973
974b3 .reg %fr9
975b3L .reg %fr9L
976b3R .reg %fr9R
977
978b4 .reg %fr10
979b4L .reg %fr10L
980b4R .reg %fr10R
981
982b5 .reg %fr11
983b5L .reg %fr11L
984b5R .reg %fr11R
985
986b6 .reg %fr12
987b6L .reg %fr12L
988b6R .reg %fr12R
989
990b7 .reg %fr13
991b7L .reg %fr13L
992b7R .reg %fr13R
993
994c1 .reg %r21 ; only reg
995temp1 .reg %r20 ; only reg
996temp2 .reg %r19 ; only reg
997temp3 .reg %r31 ; only reg
998
999m1 .reg %r28
1000c2 .reg %r23
1001high_one .reg %r1
1002ht .reg %r6
1003lt .reg %r5
1004m .reg %r4
1005c3 .reg %r3
1006
1007SQR_ADD_C .macro A0L,A0R,C1,C2,C3
1008 XMPYU A0L,A0R,ftemp1 ; m
1009 FSTD ftemp1,-24(%sp) ; store m
1010
1011 XMPYU A0R,A0R,ftemp2 ; lt
1012 FSTD ftemp2,-16(%sp) ; store lt
1013
1014 XMPYU A0L,A0L,ftemp3 ; ht
1015 FSTD ftemp3,-8(%sp) ; store ht
1016
1017 LDD -24(%sp),m ; load m
1018 AND m,high_mask,temp2 ; m & Mask
1019 DEPD,Z m,30,31,temp3 ; m << 32+1
1020 LDD -16(%sp),lt ; lt
1021
1022 LDD -8(%sp),ht ; ht
1023 EXTRD,U temp2,32,33,temp1 ; temp1 = m&Mask >> 32-1
1024 ADD temp3,lt,lt ; lt = lt+m
1025 ADD,L ht,temp1,ht ; ht += temp1
1026 ADD,DC ht,%r0,ht ; ht++
1027
1028 ADD C1,lt,C1 ; c1=c1+lt
1029 ADD,DC ht,%r0,ht ; ht++
1030
1031 ADD C2,ht,C2 ; c2=c2+ht
1032 ADD,DC C3,%r0,C3 ; c3++
1033.endm
1034
1035SQR_ADD_C2 .macro A0L,A0R,A1L,A1R,C1,C2,C3
1036 XMPYU A0L,A1R,ftemp1 ; m1 = bl*ht
1037 FSTD ftemp1,-16(%sp) ;
1038 XMPYU A0R,A1L,ftemp2 ; m = bh*lt
1039 FSTD ftemp2,-8(%sp) ;
1040 XMPYU A0R,A1R,ftemp3 ; lt = bl*lt
1041 FSTD ftemp3,-32(%sp)
1042 XMPYU A0L,A1L,ftemp4 ; ht = bh*ht
1043 FSTD ftemp4,-24(%sp) ;
1044
1045 LDD -8(%sp),m ; r21 = m
1046 LDD -16(%sp),m1 ; r19 = m1
1047 ADD,L m,m1,m ; m+m1
1048
1049 DEPD,Z m,31,32,temp3 ; (m+m1<<32)
1050 LDD -24(%sp),ht ; r24 = ht
1051
1052 CMPCLR,*>>= m,m1,%r0 ; if (m < m1)
1053 ADD,L ht,high_one,ht ; ht+=high_one
1054
1055 EXTRD,U m,31,32,temp1 ; m >> 32
1056 LDD -32(%sp),lt ; lt
1057 ADD,L ht,temp1,ht ; ht+= m>>32
1058 ADD lt,temp3,lt ; lt = lt+m1
1059 ADD,DC ht,%r0,ht ; ht++
1060
1061 ADD ht,ht,ht ; ht=ht+ht;
1062 ADD,DC C3,%r0,C3 ; add in carry (c3++)
1063
1064 ADD lt,lt,lt ; lt=lt+lt;
1065 ADD,DC ht,%r0,ht ; add in carry (ht++)
1066
1067 ADD C1,lt,C1 ; c1=c1+lt
1068 ADD,DC,*NUV ht,%r0,ht ; add in carry (ht++)
1069 LDO 1(C3),C3 ; bump c3 if overflow,nullify otherwise
1070
1071 ADD C2,ht,C2 ; c2 = c2 + ht
1072 ADD,DC C3,%r0,C3 ; add in carry (c3++)
1073.endm
1074
1075;
1076;void bn_sqr_comba8(BN_ULONG *r, BN_ULONG *a)
1077; arg0 = r_ptr
1078; arg1 = a_ptr
1079;
1080
1081bn_sqr_comba8
1082 .PROC
1083 .CALLINFO FRAME=128,ENTRY_GR=%r3,ARGS_SAVED,ORDERING_AWARE
1084 .EXPORT bn_sqr_comba8,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
1085 .ENTRY
1086 .align 64
1087
1088 STD %r3,0(%sp) ; save r3
1089 STD %r4,8(%sp) ; save r4
1090 STD %r5,16(%sp) ; save r5
1091 STD %r6,24(%sp) ; save r6
1092
1093 ;
1094 ; Zero out carries
1095 ;
1096 COPY %r0,c1
1097 COPY %r0,c2
1098 COPY %r0,c3
1099
1100 LDO 128(%sp),%sp ; bump stack
1101 DEPDI,Z -1,32,33,high_mask ; Create Mask 0xffffffff80000000L
1102 DEPDI,Z 1,31,1,high_one ; Create Value 1 << 32
1103
1104 ;
1105 ; Load up all of the values we are going to use
1106 ;
1107 FLDD 0(a_ptr),a0
1108 FLDD 8(a_ptr),a1
1109 FLDD 16(a_ptr),a2
1110 FLDD 24(a_ptr),a3
1111 FLDD 32(a_ptr),a4
1112 FLDD 40(a_ptr),a5
1113 FLDD 48(a_ptr),a6
1114 FLDD 56(a_ptr),a7
1115
1116 SQR_ADD_C a0L,a0R,c1,c2,c3
1117 STD c1,0(r_ptr) ; r[0] = c1;
1118 COPY %r0,c1
1119
1120 SQR_ADD_C2 a1L,a1R,a0L,a0R,c2,c3,c1
1121 STD c2,8(r_ptr) ; r[1] = c2;
1122 COPY %r0,c2
1123
1124 SQR_ADD_C a1L,a1R,c3,c1,c2
1125 SQR_ADD_C2 a2L,a2R,a0L,a0R,c3,c1,c2
1126 STD c3,16(r_ptr) ; r[2] = c3;
1127 COPY %r0,c3
1128
1129 SQR_ADD_C2 a3L,a3R,a0L,a0R,c1,c2,c3
1130 SQR_ADD_C2 a2L,a2R,a1L,a1R,c1,c2,c3
1131 STD c1,24(r_ptr) ; r[3] = c1;
1132 COPY %r0,c1
1133
1134 SQR_ADD_C a2L,a2R,c2,c3,c1
1135 SQR_ADD_C2 a3L,a3R,a1L,a1R,c2,c3,c1
1136 SQR_ADD_C2 a4L,a4R,a0L,a0R,c2,c3,c1
1137 STD c2,32(r_ptr) ; r[4] = c2;
1138 COPY %r0,c2
1139
1140 SQR_ADD_C2 a5L,a5R,a0L,a0R,c3,c1,c2
1141 SQR_ADD_C2 a4L,a4R,a1L,a1R,c3,c1,c2
1142 SQR_ADD_C2 a3L,a3R,a2L,a2R,c3,c1,c2
1143 STD c3,40(r_ptr) ; r[5] = c3;
1144 COPY %r0,c3
1145
1146 SQR_ADD_C a3L,a3R,c1,c2,c3
1147 SQR_ADD_C2 a4L,a4R,a2L,a2R,c1,c2,c3
1148 SQR_ADD_C2 a5L,a5R,a1L,a1R,c1,c2,c3
1149 SQR_ADD_C2 a6L,a6R,a0L,a0R,c1,c2,c3
1150 STD c1,48(r_ptr) ; r[6] = c1;
1151 COPY %r0,c1
1152
1153 SQR_ADD_C2 a7L,a7R,a0L,a0R,c2,c3,c1
1154 SQR_ADD_C2 a6L,a6R,a1L,a1R,c2,c3,c1
1155 SQR_ADD_C2 a5L,a5R,a2L,a2R,c2,c3,c1
1156 SQR_ADD_C2 a4L,a4R,a3L,a3R,c2,c3,c1
1157 STD c2,56(r_ptr) ; r[7] = c2;
1158 COPY %r0,c2
1159
1160 SQR_ADD_C a4L,a4R,c3,c1,c2
1161 SQR_ADD_C2 a5L,a5R,a3L,a3R,c3,c1,c2
1162 SQR_ADD_C2 a6L,a6R,a2L,a2R,c3,c1,c2
1163 SQR_ADD_C2 a7L,a7R,a1L,a1R,c3,c1,c2
1164 STD c3,64(r_ptr) ; r[8] = c3;
1165 COPY %r0,c3
1166
1167 SQR_ADD_C2 a7L,a7R,a2L,a2R,c1,c2,c3
1168 SQR_ADD_C2 a6L,a6R,a3L,a3R,c1,c2,c3
1169 SQR_ADD_C2 a5L,a5R,a4L,a4R,c1,c2,c3
1170 STD c1,72(r_ptr) ; r[9] = c1;
1171 COPY %r0,c1
1172
1173 SQR_ADD_C a5L,a5R,c2,c3,c1
1174 SQR_ADD_C2 a6L,a6R,a4L,a4R,c2,c3,c1
1175 SQR_ADD_C2 a7L,a7R,a3L,a3R,c2,c3,c1
1176 STD c2,80(r_ptr) ; r[10] = c2;
1177 COPY %r0,c2
1178
1179 SQR_ADD_C2 a7L,a7R,a4L,a4R,c3,c1,c2
1180 SQR_ADD_C2 a6L,a6R,a5L,a5R,c3,c1,c2
1181 STD c3,88(r_ptr) ; r[11] = c3;
1182 COPY %r0,c3
1183
1184 SQR_ADD_C a6L,a6R,c1,c2,c3
1185 SQR_ADD_C2 a7L,a7R,a5L,a5R,c1,c2,c3
1186 STD c1,96(r_ptr) ; r[12] = c1;
1187 COPY %r0,c1
1188
1189 SQR_ADD_C2 a7L,a7R,a6L,a6R,c2,c3,c1
1190 STD c2,104(r_ptr) ; r[13] = c2;
1191 COPY %r0,c2
1192
1193 SQR_ADD_C a7L,a7R,c3,c1,c2
1194 STD c3, 112(r_ptr) ; r[14] = c3
1195 STD c1, 120(r_ptr) ; r[15] = c1
1196
1197 .EXIT
1198 LDD -104(%sp),%r6 ; restore r6
1199 LDD -112(%sp),%r5 ; restore r5
1200 LDD -120(%sp),%r4 ; restore r4
1201 BVE (%rp)
1202 LDD,MB -128(%sp),%r3
1203
1204 .PROCEND
1205
1206;-----------------------------------------------------------------------------
1207;
1208;void bn_sqr_comba4(BN_ULONG *r, BN_ULONG *a)
1209; arg0 = r_ptr
1210; arg1 = a_ptr
1211;
1212
1213bn_sqr_comba4
1214 .proc
1215 .callinfo FRAME=128,ENTRY_GR=%r3,ARGS_SAVED,ORDERING_AWARE
1216 .EXPORT bn_sqr_comba4,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
1217 .entry
1218 .align 64
1219 STD %r3,0(%sp) ; save r3
1220 STD %r4,8(%sp) ; save r4
1221 STD %r5,16(%sp) ; save r5
1222 STD %r6,24(%sp) ; save r6
1223
1224 ;
1225 ; Zero out carries
1226 ;
1227 COPY %r0,c1
1228 COPY %r0,c2
1229 COPY %r0,c3
1230
1231 LDO 128(%sp),%sp ; bump stack
1232 DEPDI,Z -1,32,33,high_mask ; Create Mask 0xffffffff80000000L
1233 DEPDI,Z 1,31,1,high_one ; Create Value 1 << 32
1234
1235 ;
1236 ; Load up all of the values we are going to use
1237 ;
1238 FLDD 0(a_ptr),a0
1239 FLDD 8(a_ptr),a1
1240 FLDD 16(a_ptr),a2
1241 FLDD 24(a_ptr),a3
1242 FLDD 32(a_ptr),a4
1243 FLDD 40(a_ptr),a5
1244 FLDD 48(a_ptr),a6
1245 FLDD 56(a_ptr),a7
1246
1247 SQR_ADD_C a0L,a0R,c1,c2,c3
1248
1249 STD c1,0(r_ptr) ; r[0] = c1;
1250 COPY %r0,c1
1251
1252 SQR_ADD_C2 a1L,a1R,a0L,a0R,c2,c3,c1
1253
1254 STD c2,8(r_ptr) ; r[1] = c2;
1255 COPY %r0,c2
1256
1257 SQR_ADD_C a1L,a1R,c3,c1,c2
1258 SQR_ADD_C2 a2L,a2R,a0L,a0R,c3,c1,c2
1259
1260 STD c3,16(r_ptr) ; r[2] = c3;
1261 COPY %r0,c3
1262
1263 SQR_ADD_C2 a3L,a3R,a0L,a0R,c1,c2,c3
1264 SQR_ADD_C2 a2L,a2R,a1L,a1R,c1,c2,c3
1265
1266 STD c1,24(r_ptr) ; r[3] = c1;
1267 COPY %r0,c1
1268
1269 SQR_ADD_C a2L,a2R,c2,c3,c1
1270 SQR_ADD_C2 a3L,a3R,a1L,a1R,c2,c3,c1
1271
1272 STD c2,32(r_ptr) ; r[4] = c2;
1273 COPY %r0,c2
1274
1275 SQR_ADD_C2 a3L,a3R,a2L,a2R,c3,c1,c2
1276 STD c3,40(r_ptr) ; r[5] = c3;
1277 COPY %r0,c3
1278
1279 SQR_ADD_C a3L,a3R,c1,c2,c3
1280 STD c1,48(r_ptr) ; r[6] = c1;
1281 STD c2,56(r_ptr) ; r[7] = c2;
1282
1283 .EXIT
1284 LDD -104(%sp),%r6 ; restore r6
1285 LDD -112(%sp),%r5 ; restore r5
1286 LDD -120(%sp),%r4 ; restore r4
1287 BVE (%rp)
1288 LDD,MB -128(%sp),%r3
1289
1290 .PROCEND
1291
1292
1293;---------------------------------------------------------------------------
1294
1295MUL_ADD_C .macro A0L,A0R,B0L,B0R,C1,C2,C3
1296 XMPYU A0L,B0R,ftemp1 ; m1 = bl*ht
1297 FSTD ftemp1,-16(%sp) ;
1298 XMPYU A0R,B0L,ftemp2 ; m = bh*lt
1299 FSTD ftemp2,-8(%sp) ;
1300 XMPYU A0R,B0R,ftemp3 ; lt = bl*lt
1301 FSTD ftemp3,-32(%sp)
1302 XMPYU A0L,B0L,ftemp4 ; ht = bh*ht
1303 FSTD ftemp4,-24(%sp) ;
1304
1305 LDD -8(%sp),m ; r21 = m
1306 LDD -16(%sp),m1 ; r19 = m1
1307 ADD,L m,m1,m ; m+m1
1308
1309 DEPD,Z m,31,32,temp3 ; (m+m1<<32)
1310 LDD -24(%sp),ht ; r24 = ht
1311
1312 CMPCLR,*>>= m,m1,%r0 ; if (m < m1)
1313 ADD,L ht,high_one,ht ; ht+=high_one
1314
1315 EXTRD,U m,31,32,temp1 ; m >> 32
1316 LDD -32(%sp),lt ; lt
1317 ADD,L ht,temp1,ht ; ht+= m>>32
1318 ADD lt,temp3,lt ; lt = lt+m1
1319 ADD,DC ht,%r0,ht ; ht++
1320
1321 ADD C1,lt,C1 ; c1=c1+lt
1322 ADD,DC ht,%r0,ht ; bump c3 if overflow,nullify otherwise
1323
1324 ADD C2,ht,C2 ; c2 = c2 + ht
1325 ADD,DC C3,%r0,C3 ; add in carry (c3++)
1326.endm
1327
1328
1329;
1330;void bn_mul_comba8(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b)
1331; arg0 = r_ptr
1332; arg1 = a_ptr
1333; arg2 = b_ptr
1334;
1335
1336bn_mul_comba8
1337 .proc
1338 .callinfo FRAME=128,ENTRY_GR=%r3,ARGS_SAVED,ORDERING_AWARE
1339 .EXPORT bn_mul_comba8,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
1340 .entry
1341 .align 64
1342
1343 STD %r3,0(%sp) ; save r3
1344 STD %r4,8(%sp) ; save r4
1345 STD %r5,16(%sp) ; save r5
1346 STD %r6,24(%sp) ; save r6
1347 FSTD %fr12,32(%sp) ; save r6
1348 FSTD %fr13,40(%sp) ; save r7
1349
1350 ;
1351 ; Zero out carries
1352 ;
1353 COPY %r0,c1
1354 COPY %r0,c2
1355 COPY %r0,c3
1356
1357 LDO 128(%sp),%sp ; bump stack
1358 DEPDI,Z 1,31,1,high_one ; Create Value 1 << 32
1359
1360 ;
1361 ; Load up all of the values we are going to use
1362 ;
1363 FLDD 0(a_ptr),a0
1364 FLDD 8(a_ptr),a1
1365 FLDD 16(a_ptr),a2
1366 FLDD 24(a_ptr),a3
1367 FLDD 32(a_ptr),a4
1368 FLDD 40(a_ptr),a5
1369 FLDD 48(a_ptr),a6
1370 FLDD 56(a_ptr),a7
1371
1372 FLDD 0(b_ptr),b0
1373 FLDD 8(b_ptr),b1
1374 FLDD 16(b_ptr),b2
1375 FLDD 24(b_ptr),b3
1376 FLDD 32(b_ptr),b4
1377 FLDD 40(b_ptr),b5
1378 FLDD 48(b_ptr),b6
1379 FLDD 56(b_ptr),b7
1380
1381 MUL_ADD_C a0L,a0R,b0L,b0R,c1,c2,c3
1382 STD c1,0(r_ptr)
1383 COPY %r0,c1
1384
1385 MUL_ADD_C a0L,a0R,b1L,b1R,c2,c3,c1
1386 MUL_ADD_C a1L,a1R,b0L,b0R,c2,c3,c1
1387 STD c2,8(r_ptr)
1388 COPY %r0,c2
1389
1390 MUL_ADD_C a2L,a2R,b0L,b0R,c3,c1,c2
1391 MUL_ADD_C a1L,a1R,b1L,b1R,c3,c1,c2
1392 MUL_ADD_C a0L,a0R,b2L,b2R,c3,c1,c2
1393 STD c3,16(r_ptr)
1394 COPY %r0,c3
1395
1396 MUL_ADD_C a0L,a0R,b3L,b3R,c1,c2,c3
1397 MUL_ADD_C a1L,a1R,b2L,b2R,c1,c2,c3
1398 MUL_ADD_C a2L,a2R,b1L,b1R,c1,c2,c3
1399 MUL_ADD_C a3L,a3R,b0L,b0R,c1,c2,c3
1400 STD c1,24(r_ptr)
1401 COPY %r0,c1
1402
1403 MUL_ADD_C a4L,a4R,b0L,b0R,c2,c3,c1
1404 MUL_ADD_C a3L,a3R,b1L,b1R,c2,c3,c1
1405 MUL_ADD_C a2L,a2R,b2L,b2R,c2,c3,c1
1406 MUL_ADD_C a1L,a1R,b3L,b3R,c2,c3,c1
1407 MUL_ADD_C a0L,a0R,b4L,b4R,c2,c3,c1
1408 STD c2,32(r_ptr)
1409 COPY %r0,c2
1410
1411 MUL_ADD_C a0L,a0R,b5L,b5R,c3,c1,c2
1412 MUL_ADD_C a1L,a1R,b4L,b4R,c3,c1,c2
1413 MUL_ADD_C a2L,a2R,b3L,b3R,c3,c1,c2
1414 MUL_ADD_C a3L,a3R,b2L,b2R,c3,c1,c2
1415 MUL_ADD_C a4L,a4R,b1L,b1R,c3,c1,c2
1416 MUL_ADD_C a5L,a5R,b0L,b0R,c3,c1,c2
1417 STD c3,40(r_ptr)
1418 COPY %r0,c3
1419
1420 MUL_ADD_C a6L,a6R,b0L,b0R,c1,c2,c3
1421 MUL_ADD_C a5L,a5R,b1L,b1R,c1,c2,c3
1422 MUL_ADD_C a4L,a4R,b2L,b2R,c1,c2,c3
1423 MUL_ADD_C a3L,a3R,b3L,b3R,c1,c2,c3
1424 MUL_ADD_C a2L,a2R,b4L,b4R,c1,c2,c3
1425 MUL_ADD_C a1L,a1R,b5L,b5R,c1,c2,c3
1426 MUL_ADD_C a0L,a0R,b6L,b6R,c1,c2,c3
1427 STD c1,48(r_ptr)
1428 COPY %r0,c1
1429
1430 MUL_ADD_C a0L,a0R,b7L,b7R,c2,c3,c1
1431 MUL_ADD_C a1L,a1R,b6L,b6R,c2,c3,c1
1432 MUL_ADD_C a2L,a2R,b5L,b5R,c2,c3,c1
1433 MUL_ADD_C a3L,a3R,b4L,b4R,c2,c3,c1
1434 MUL_ADD_C a4L,a4R,b3L,b3R,c2,c3,c1
1435 MUL_ADD_C a5L,a5R,b2L,b2R,c2,c3,c1
1436 MUL_ADD_C a6L,a6R,b1L,b1R,c2,c3,c1
1437 MUL_ADD_C a7L,a7R,b0L,b0R,c2,c3,c1
1438 STD c2,56(r_ptr)
1439 COPY %r0,c2
1440
1441 MUL_ADD_C a7L,a7R,b1L,b1R,c3,c1,c2
1442 MUL_ADD_C a6L,a6R,b2L,b2R,c3,c1,c2
1443 MUL_ADD_C a5L,a5R,b3L,b3R,c3,c1,c2
1444 MUL_ADD_C a4L,a4R,b4L,b4R,c3,c1,c2
1445 MUL_ADD_C a3L,a3R,b5L,b5R,c3,c1,c2
1446 MUL_ADD_C a2L,a2R,b6L,b6R,c3,c1,c2
1447 MUL_ADD_C a1L,a1R,b7L,b7R,c3,c1,c2
1448 STD c3,64(r_ptr)
1449 COPY %r0,c3
1450
1451 MUL_ADD_C a2L,a2R,b7L,b7R,c1,c2,c3
1452 MUL_ADD_C a3L,a3R,b6L,b6R,c1,c2,c3
1453 MUL_ADD_C a4L,a4R,b5L,b5R,c1,c2,c3
1454 MUL_ADD_C a5L,a5R,b4L,b4R,c1,c2,c3
1455 MUL_ADD_C a6L,a6R,b3L,b3R,c1,c2,c3
1456 MUL_ADD_C a7L,a7R,b2L,b2R,c1,c2,c3
1457 STD c1,72(r_ptr)
1458 COPY %r0,c1
1459
1460 MUL_ADD_C a7L,a7R,b3L,b3R,c2,c3,c1
1461 MUL_ADD_C a6L,a6R,b4L,b4R,c2,c3,c1
1462 MUL_ADD_C a5L,a5R,b5L,b5R,c2,c3,c1
1463 MUL_ADD_C a4L,a4R,b6L,b6R,c2,c3,c1
1464 MUL_ADD_C a3L,a3R,b7L,b7R,c2,c3,c1
1465 STD c2,80(r_ptr)
1466 COPY %r0,c2
1467
1468 MUL_ADD_C a4L,a4R,b7L,b7R,c3,c1,c2
1469 MUL_ADD_C a5L,a5R,b6L,b6R,c3,c1,c2
1470 MUL_ADD_C a6L,a6R,b5L,b5R,c3,c1,c2
1471 MUL_ADD_C a7L,a7R,b4L,b4R,c3,c1,c2
1472 STD c3,88(r_ptr)
1473 COPY %r0,c3
1474
1475 MUL_ADD_C a7L,a7R,b5L,b5R,c1,c2,c3
1476 MUL_ADD_C a6L,a6R,b6L,b6R,c1,c2,c3
1477 MUL_ADD_C a5L,a5R,b7L,b7R,c1,c2,c3
1478 STD c1,96(r_ptr)
1479 COPY %r0,c1
1480
1481 MUL_ADD_C a6L,a6R,b7L,b7R,c2,c3,c1
1482 MUL_ADD_C a7L,a7R,b6L,b6R,c2,c3,c1
1483 STD c2,104(r_ptr)
1484 COPY %r0,c2
1485
1486 MUL_ADD_C a7L,a7R,b7L,b7R,c3,c1,c2
1487 STD c3,112(r_ptr)
1488 STD c1,120(r_ptr)
1489
1490 .EXIT
1491 FLDD -88(%sp),%fr13
1492 FLDD -96(%sp),%fr12
1493 LDD -104(%sp),%r6 ; restore r6
1494 LDD -112(%sp),%r5 ; restore r5
1495 LDD -120(%sp),%r4 ; restore r4
1496 BVE (%rp)
1497 LDD,MB -128(%sp),%r3
1498
1499 .PROCEND
1500
1501;-----------------------------------------------------------------------------
1502;
1503;void bn_mul_comba4(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b)
1504; arg0 = r_ptr
1505; arg1 = a_ptr
1506; arg2 = b_ptr
1507;
1508
1509bn_mul_comba4
1510 .proc
1511 .callinfo FRAME=128,ENTRY_GR=%r3,ARGS_SAVED,ORDERING_AWARE
1512 .EXPORT bn_mul_comba4,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
1513 .entry
1514 .align 64
1515
1516 STD %r3,0(%sp) ; save r3
1517 STD %r4,8(%sp) ; save r4
1518 STD %r5,16(%sp) ; save r5
1519 STD %r6,24(%sp) ; save r6
1520 FSTD %fr12,32(%sp) ; save r6
1521 FSTD %fr13,40(%sp) ; save r7
1522
1523 ;
1524 ; Zero out carries
1525 ;
1526 COPY %r0,c1
1527 COPY %r0,c2
1528 COPY %r0,c3
1529
1530 LDO 128(%sp),%sp ; bump stack
1531 DEPDI,Z 1,31,1,high_one ; Create Value 1 << 32
1532
1533 ;
1534 ; Load up all of the values we are going to use
1535 ;
1536 FLDD 0(a_ptr),a0
1537 FLDD 8(a_ptr),a1
1538 FLDD 16(a_ptr),a2
1539 FLDD 24(a_ptr),a3
1540
1541 FLDD 0(b_ptr),b0
1542 FLDD 8(b_ptr),b1
1543 FLDD 16(b_ptr),b2
1544 FLDD 24(b_ptr),b3
1545
1546 MUL_ADD_C a0L,a0R,b0L,b0R,c1,c2,c3
1547 STD c1,0(r_ptr)
1548 COPY %r0,c1
1549
1550 MUL_ADD_C a0L,a0R,b1L,b1R,c2,c3,c1
1551 MUL_ADD_C a1L,a1R,b0L,b0R,c2,c3,c1
1552 STD c2,8(r_ptr)
1553 COPY %r0,c2
1554
1555 MUL_ADD_C a2L,a2R,b0L,b0R,c3,c1,c2
1556 MUL_ADD_C a1L,a1R,b1L,b1R,c3,c1,c2
1557 MUL_ADD_C a0L,a0R,b2L,b2R,c3,c1,c2
1558 STD c3,16(r_ptr)
1559 COPY %r0,c3
1560
1561 MUL_ADD_C a0L,a0R,b3L,b3R,c1,c2,c3
1562 MUL_ADD_C a1L,a1R,b2L,b2R,c1,c2,c3
1563 MUL_ADD_C a2L,a2R,b1L,b1R,c1,c2,c3
1564 MUL_ADD_C a3L,a3R,b0L,b0R,c1,c2,c3
1565 STD c1,24(r_ptr)
1566 COPY %r0,c1
1567
1568 MUL_ADD_C a3L,a3R,b1L,b1R,c2,c3,c1
1569 MUL_ADD_C a2L,a2R,b2L,b2R,c2,c3,c1
1570 MUL_ADD_C a1L,a1R,b3L,b3R,c2,c3,c1
1571 STD c2,32(r_ptr)
1572 COPY %r0,c2
1573
1574 MUL_ADD_C a2L,a2R,b3L,b3R,c3,c1,c2
1575 MUL_ADD_C a3L,a3R,b2L,b2R,c3,c1,c2
1576 STD c3,40(r_ptr)
1577 COPY %r0,c3
1578
1579 MUL_ADD_C a3L,a3R,b3L,b3R,c1,c2,c3
1580 STD c1,48(r_ptr)
1581 STD c2,56(r_ptr)
1582
1583 .EXIT
1584 FLDD -88(%sp),%fr13
1585 FLDD -96(%sp),%fr12
1586 LDD -104(%sp),%r6 ; restore r6
1587 LDD -112(%sp),%r5 ; restore r5
1588 LDD -120(%sp),%r4 ; restore r4
1589 BVE (%rp)
1590 LDD,MB -128(%sp),%r3
1591
1592 .PROCEND
1593
1594
1595 .SPACE $TEXT$
1596 .SUBSPA $CODE$
1597 .SPACE $PRIVATE$,SORT=16
1598 .IMPORT $global$,DATA
1599 .SPACE $TEXT$
1600 .SUBSPA $CODE$
1601 .SUBSPA $LIT$,ACCESS=0x2c
1602C$4
1603 .ALIGN 8
1604 .STRINGZ "Division would overflow (%d)\n"
1605 .END
diff --git a/src/lib/libcrypto/bn/asm/sparcv8.S b/src/lib/libcrypto/bn/asm/sparcv8.S
deleted file mode 100644
index 88c5dc480a..0000000000
--- a/src/lib/libcrypto/bn/asm/sparcv8.S
+++ /dev/null
@@ -1,1458 +0,0 @@
1.ident "sparcv8.s, Version 1.4"
2.ident "SPARC v8 ISA artwork by Andy Polyakov <appro@fy.chalmers.se>"
3
4/*
5 * ====================================================================
6 * Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
7 * project.
8 *
9 * Rights for redistribution and usage in source and binary forms are
10 * granted according to the OpenSSL license. Warranty of any kind is
11 * disclaimed.
12 * ====================================================================
13 */
14
15/*
16 * This is my modest contributon to OpenSSL project (see
17 * http://www.openssl.org/ for more information about it) and is
18 * a drop-in SuperSPARC ISA replacement for crypto/bn/bn_asm.c
19 * module. For updates see http://fy.chalmers.se/~appro/hpe/.
20 *
21 * See bn_asm.sparc.v8plus.S for more details.
22 */
23
24/*
25 * Revision history.
26 *
27 * 1.1 - new loop unrolling model(*);
28 * 1.2 - made gas friendly;
29 * 1.3 - fixed problem with /usr/ccs/lib/cpp;
30 * 1.4 - some retunes;
31 *
32 * (*) see bn_asm.sparc.v8plus.S for details
33 */
34
35.section ".text",#alloc,#execinstr
36.file "bn_asm.sparc.v8.S"
37
38.align 32
39
40.global bn_mul_add_words
41/*
42 * BN_ULONG bn_mul_add_words(rp,ap,num,w)
43 * BN_ULONG *rp,*ap;
44 * int num;
45 * BN_ULONG w;
46 */
47bn_mul_add_words:
48 cmp %o2,0
49 bg,a .L_bn_mul_add_words_proceed
50 ld [%o1],%g2
51 retl
52 clr %o0
53
54.L_bn_mul_add_words_proceed:
55 andcc %o2,-4,%g0
56 bz .L_bn_mul_add_words_tail
57 clr %o5
58
59.L_bn_mul_add_words_loop:
60 ld [%o0],%o4
61 ld [%o1+4],%g3
62 umul %o3,%g2,%g2
63 rd %y,%g1
64 addcc %o4,%o5,%o4
65 addx %g1,0,%g1
66 addcc %o4,%g2,%o4
67 st %o4,[%o0]
68 addx %g1,0,%o5
69
70 ld [%o0+4],%o4
71 ld [%o1+8],%g2
72 umul %o3,%g3,%g3
73 dec 4,%o2
74 rd %y,%g1
75 addcc %o4,%o5,%o4
76 addx %g1,0,%g1
77 addcc %o4,%g3,%o4
78 st %o4,[%o0+4]
79 addx %g1,0,%o5
80
81 ld [%o0+8],%o4
82 ld [%o1+12],%g3
83 umul %o3,%g2,%g2
84 inc 16,%o1
85 rd %y,%g1
86 addcc %o4,%o5,%o4
87 addx %g1,0,%g1
88 addcc %o4,%g2,%o4
89 st %o4,[%o0+8]
90 addx %g1,0,%o5
91
92 ld [%o0+12],%o4
93 umul %o3,%g3,%g3
94 inc 16,%o0
95 rd %y,%g1
96 addcc %o4,%o5,%o4
97 addx %g1,0,%g1
98 addcc %o4,%g3,%o4
99 st %o4,[%o0-4]
100 addx %g1,0,%o5
101 andcc %o2,-4,%g0
102 bnz,a .L_bn_mul_add_words_loop
103 ld [%o1],%g2
104
105 tst %o2
106 bnz,a .L_bn_mul_add_words_tail
107 ld [%o1],%g2
108.L_bn_mul_add_words_return:
109 retl
110 mov %o5,%o0
111 nop
112
113.L_bn_mul_add_words_tail:
114 ld [%o0],%o4
115 umul %o3,%g2,%g2
116 addcc %o4,%o5,%o4
117 rd %y,%g1
118 addx %g1,0,%g1
119 addcc %o4,%g2,%o4
120 addx %g1,0,%o5
121 deccc %o2
122 bz .L_bn_mul_add_words_return
123 st %o4,[%o0]
124
125 ld [%o1+4],%g2
126 ld [%o0+4],%o4
127 umul %o3,%g2,%g2
128 rd %y,%g1
129 addcc %o4,%o5,%o4
130 addx %g1,0,%g1
131 addcc %o4,%g2,%o4
132 addx %g1,0,%o5
133 deccc %o2
134 bz .L_bn_mul_add_words_return
135 st %o4,[%o0+4]
136
137 ld [%o1+8],%g2
138 ld [%o0+8],%o4
139 umul %o3,%g2,%g2
140 rd %y,%g1
141 addcc %o4,%o5,%o4
142 addx %g1,0,%g1
143 addcc %o4,%g2,%o4
144 st %o4,[%o0+8]
145 retl
146 addx %g1,0,%o0
147
148.type bn_mul_add_words,#function
149.size bn_mul_add_words,(.-bn_mul_add_words)
150
151.align 32
152
153.global bn_mul_words
154/*
155 * BN_ULONG bn_mul_words(rp,ap,num,w)
156 * BN_ULONG *rp,*ap;
157 * int num;
158 * BN_ULONG w;
159 */
160bn_mul_words:
161 cmp %o2,0
162 bg,a .L_bn_mul_words_proceeed
163 ld [%o1],%g2
164 retl
165 clr %o0
166
167.L_bn_mul_words_proceeed:
168 andcc %o2,-4,%g0
169 bz .L_bn_mul_words_tail
170 clr %o5
171
172.L_bn_mul_words_loop:
173 ld [%o1+4],%g3
174 umul %o3,%g2,%g2
175 addcc %g2,%o5,%g2
176 rd %y,%g1
177 addx %g1,0,%o5
178 st %g2,[%o0]
179
180 ld [%o1+8],%g2
181 umul %o3,%g3,%g3
182 addcc %g3,%o5,%g3
183 rd %y,%g1
184 dec 4,%o2
185 addx %g1,0,%o5
186 st %g3,[%o0+4]
187
188 ld [%o1+12],%g3
189 umul %o3,%g2,%g2
190 addcc %g2,%o5,%g2
191 rd %y,%g1
192 inc 16,%o1
193 st %g2,[%o0+8]
194 addx %g1,0,%o5
195
196 umul %o3,%g3,%g3
197 addcc %g3,%o5,%g3
198 rd %y,%g1
199 inc 16,%o0
200 addx %g1,0,%o5
201 st %g3,[%o0-4]
202 andcc %o2,-4,%g0
203 nop
204 bnz,a .L_bn_mul_words_loop
205 ld [%o1],%g2
206
207 tst %o2
208 bnz,a .L_bn_mul_words_tail
209 ld [%o1],%g2
210.L_bn_mul_words_return:
211 retl
212 mov %o5,%o0
213 nop
214
215.L_bn_mul_words_tail:
216 umul %o3,%g2,%g2
217 addcc %g2,%o5,%g2
218 rd %y,%g1
219 addx %g1,0,%o5
220 deccc %o2
221 bz .L_bn_mul_words_return
222 st %g2,[%o0]
223 nop
224
225 ld [%o1+4],%g2
226 umul %o3,%g2,%g2
227 addcc %g2,%o5,%g2
228 rd %y,%g1
229 addx %g1,0,%o5
230 deccc %o2
231 bz .L_bn_mul_words_return
232 st %g2,[%o0+4]
233
234 ld [%o1+8],%g2
235 umul %o3,%g2,%g2
236 addcc %g2,%o5,%g2
237 rd %y,%g1
238 st %g2,[%o0+8]
239 retl
240 addx %g1,0,%o0
241
242.type bn_mul_words,#function
243.size bn_mul_words,(.-bn_mul_words)
244
245.align 32
246.global bn_sqr_words
247/*
248 * void bn_sqr_words(r,a,n)
249 * BN_ULONG *r,*a;
250 * int n;
251 */
252bn_sqr_words:
253 cmp %o2,0
254 bg,a .L_bn_sqr_words_proceeed
255 ld [%o1],%g2
256 retl
257 clr %o0
258
259.L_bn_sqr_words_proceeed:
260 andcc %o2,-4,%g0
261 bz .L_bn_sqr_words_tail
262 clr %o5
263
264.L_bn_sqr_words_loop:
265 ld [%o1+4],%g3
266 umul %g2,%g2,%o4
267 st %o4,[%o0]
268 rd %y,%o5
269 st %o5,[%o0+4]
270
271 ld [%o1+8],%g2
272 umul %g3,%g3,%o4
273 dec 4,%o2
274 st %o4,[%o0+8]
275 rd %y,%o5
276 st %o5,[%o0+12]
277 nop
278
279 ld [%o1+12],%g3
280 umul %g2,%g2,%o4
281 st %o4,[%o0+16]
282 rd %y,%o5
283 inc 16,%o1
284 st %o5,[%o0+20]
285
286 umul %g3,%g3,%o4
287 inc 32,%o0
288 st %o4,[%o0-8]
289 rd %y,%o5
290 st %o5,[%o0-4]
291 andcc %o2,-4,%g2
292 bnz,a .L_bn_sqr_words_loop
293 ld [%o1],%g2
294
295 tst %o2
296 nop
297 bnz,a .L_bn_sqr_words_tail
298 ld [%o1],%g2
299.L_bn_sqr_words_return:
300 retl
301 clr %o0
302
303.L_bn_sqr_words_tail:
304 umul %g2,%g2,%o4
305 st %o4,[%o0]
306 deccc %o2
307 rd %y,%o5
308 bz .L_bn_sqr_words_return
309 st %o5,[%o0+4]
310
311 ld [%o1+4],%g2
312 umul %g2,%g2,%o4
313 st %o4,[%o0+8]
314 deccc %o2
315 rd %y,%o5
316 nop
317 bz .L_bn_sqr_words_return
318 st %o5,[%o0+12]
319
320 ld [%o1+8],%g2
321 umul %g2,%g2,%o4
322 st %o4,[%o0+16]
323 rd %y,%o5
324 st %o5,[%o0+20]
325 retl
326 clr %o0
327
328.type bn_sqr_words,#function
329.size bn_sqr_words,(.-bn_sqr_words)
330
331.align 32
332
333.global bn_div_words
334/*
335 * BN_ULONG bn_div_words(h,l,d)
336 * BN_ULONG h,l,d;
337 */
338bn_div_words:
339 wr %o0,%y
340 udiv %o1,%o2,%o0
341 retl
342 nop
343
344.type bn_div_words,#function
345.size bn_div_words,(.-bn_div_words)
346
347.align 32
348
349.global bn_add_words
350/*
351 * BN_ULONG bn_add_words(rp,ap,bp,n)
352 * BN_ULONG *rp,*ap,*bp;
353 * int n;
354 */
355bn_add_words:
356 cmp %o3,0
357 bg,a .L_bn_add_words_proceed
358 ld [%o1],%o4
359 retl
360 clr %o0
361
362.L_bn_add_words_proceed:
363 andcc %o3,-4,%g0
364 bz .L_bn_add_words_tail
365 clr %g1
366 ba .L_bn_add_words_warn_loop
367 addcc %g0,0,%g0 ! clear carry flag
368
369.L_bn_add_words_loop:
370 ld [%o1],%o4
371.L_bn_add_words_warn_loop:
372 ld [%o2],%o5
373 ld [%o1+4],%g3
374 ld [%o2+4],%g4
375 dec 4,%o3
376 addxcc %o5,%o4,%o5
377 st %o5,[%o0]
378
379 ld [%o1+8],%o4
380 ld [%o2+8],%o5
381 inc 16,%o1
382 addxcc %g3,%g4,%g3
383 st %g3,[%o0+4]
384
385 ld [%o1-4],%g3
386 ld [%o2+12],%g4
387 inc 16,%o2
388 addxcc %o5,%o4,%o5
389 st %o5,[%o0+8]
390
391 inc 16,%o0
392 addxcc %g3,%g4,%g3
393 st %g3,[%o0-4]
394 addx %g0,0,%g1
395 andcc %o3,-4,%g0
396 bnz,a .L_bn_add_words_loop
397 addcc %g1,-1,%g0
398
399 tst %o3
400 bnz,a .L_bn_add_words_tail
401 ld [%o1],%o4
402.L_bn_add_words_return:
403 retl
404 mov %g1,%o0
405
406.L_bn_add_words_tail:
407 addcc %g1,-1,%g0
408 ld [%o2],%o5
409 addxcc %o5,%o4,%o5
410 addx %g0,0,%g1
411 deccc %o3
412 bz .L_bn_add_words_return
413 st %o5,[%o0]
414
415 ld [%o1+4],%o4
416 addcc %g1,-1,%g0
417 ld [%o2+4],%o5
418 addxcc %o5,%o4,%o5
419 addx %g0,0,%g1
420 deccc %o3
421 bz .L_bn_add_words_return
422 st %o5,[%o0+4]
423
424 ld [%o1+8],%o4
425 addcc %g1,-1,%g0
426 ld [%o2+8],%o5
427 addxcc %o5,%o4,%o5
428 st %o5,[%o0+8]
429 retl
430 addx %g0,0,%o0
431
432.type bn_add_words,#function
433.size bn_add_words,(.-bn_add_words)
434
435.align 32
436
437.global bn_sub_words
438/*
439 * BN_ULONG bn_sub_words(rp,ap,bp,n)
440 * BN_ULONG *rp,*ap,*bp;
441 * int n;
442 */
443bn_sub_words:
444 cmp %o3,0
445 bg,a .L_bn_sub_words_proceed
446 ld [%o1],%o4
447 retl
448 clr %o0
449
450.L_bn_sub_words_proceed:
451 andcc %o3,-4,%g0
452 bz .L_bn_sub_words_tail
453 clr %g1
454 ba .L_bn_sub_words_warm_loop
455 addcc %g0,0,%g0 ! clear carry flag
456
457.L_bn_sub_words_loop:
458 ld [%o1],%o4
459.L_bn_sub_words_warm_loop:
460 ld [%o2],%o5
461 ld [%o1+4],%g3
462 ld [%o2+4],%g4
463 dec 4,%o3
464 subxcc %o4,%o5,%o5
465 st %o5,[%o0]
466
467 ld [%o1+8],%o4
468 ld [%o2+8],%o5
469 inc 16,%o1
470 subxcc %g3,%g4,%g4
471 st %g4,[%o0+4]
472
473 ld [%o1-4],%g3
474 ld [%o2+12],%g4
475 inc 16,%o2
476 subxcc %o4,%o5,%o5
477 st %o5,[%o0+8]
478
479 inc 16,%o0
480 subxcc %g3,%g4,%g4
481 st %g4,[%o0-4]
482 addx %g0,0,%g1
483 andcc %o3,-4,%g0
484 bnz,a .L_bn_sub_words_loop
485 addcc %g1,-1,%g0
486
487 tst %o3
488 nop
489 bnz,a .L_bn_sub_words_tail
490 ld [%o1],%o4
491.L_bn_sub_words_return:
492 retl
493 mov %g1,%o0
494
495.L_bn_sub_words_tail:
496 addcc %g1,-1,%g0
497 ld [%o2],%o5
498 subxcc %o4,%o5,%o5
499 addx %g0,0,%g1
500 deccc %o3
501 bz .L_bn_sub_words_return
502 st %o5,[%o0]
503 nop
504
505 ld [%o1+4],%o4
506 addcc %g1,-1,%g0
507 ld [%o2+4],%o5
508 subxcc %o4,%o5,%o5
509 addx %g0,0,%g1
510 deccc %o3
511 bz .L_bn_sub_words_return
512 st %o5,[%o0+4]
513
514 ld [%o1+8],%o4
515 addcc %g1,-1,%g0
516 ld [%o2+8],%o5
517 subxcc %o4,%o5,%o5
518 st %o5,[%o0+8]
519 retl
520 addx %g0,0,%o0
521
522.type bn_sub_words,#function
523.size bn_sub_words,(.-bn_sub_words)
524
525#define FRAME_SIZE -96
526
527/*
528 * Here is register usage map for *all* routines below.
529 */
530#define t_1 %o0
531#define t_2 %o1
532#define c_1 %o2
533#define c_2 %o3
534#define c_3 %o4
535
536#define ap(I) [%i1+4*I]
537#define bp(I) [%i2+4*I]
538#define rp(I) [%i0+4*I]
539
540#define a_0 %l0
541#define a_1 %l1
542#define a_2 %l2
543#define a_3 %l3
544#define a_4 %l4
545#define a_5 %l5
546#define a_6 %l6
547#define a_7 %l7
548
549#define b_0 %i3
550#define b_1 %i4
551#define b_2 %i5
552#define b_3 %o5
553#define b_4 %g1
554#define b_5 %g2
555#define b_6 %g3
556#define b_7 %g4
557
558.align 32
559.global bn_mul_comba8
560/*
561 * void bn_mul_comba8(r,a,b)
562 * BN_ULONG *r,*a,*b;
563 */
564bn_mul_comba8:
565 save %sp,FRAME_SIZE,%sp
566 ld ap(0),a_0
567 ld bp(0),b_0
568 umul a_0,b_0,c_1 !=!mul_add_c(a[0],b[0],c1,c2,c3);
569 ld bp(1),b_1
570 rd %y,c_2
571 st c_1,rp(0) !r[0]=c1;
572
573 umul a_0,b_1,t_1 !=!mul_add_c(a[0],b[1],c2,c3,c1);
574 ld ap(1),a_1
575 addcc c_2,t_1,c_2
576 rd %y,t_2
577 addxcc %g0,t_2,c_3 !=
578 addx %g0,%g0,c_1
579 ld ap(2),a_2
580 umul a_1,b_0,t_1 !mul_add_c(a[1],b[0],c2,c3,c1);
581 addcc c_2,t_1,c_2 !=
582 rd %y,t_2
583 addxcc c_3,t_2,c_3
584 st c_2,rp(1) !r[1]=c2;
585 addx c_1,%g0,c_1 !=
586
587 umul a_2,b_0,t_1 !mul_add_c(a[2],b[0],c3,c1,c2);
588 addcc c_3,t_1,c_3
589 rd %y,t_2
590 addxcc c_1,t_2,c_1 !=
591 addx %g0,%g0,c_2
592 ld bp(2),b_2
593 umul a_1,b_1,t_1 !mul_add_c(a[1],b[1],c3,c1,c2);
594 addcc c_3,t_1,c_3 !=
595 rd %y,t_2
596 addxcc c_1,t_2,c_1
597 ld bp(3),b_3
598 addx c_2,%g0,c_2 !=
599 umul a_0,b_2,t_1 !mul_add_c(a[0],b[2],c3,c1,c2);
600 addcc c_3,t_1,c_3
601 rd %y,t_2
602 addxcc c_1,t_2,c_1 !=
603 addx c_2,%g0,c_2
604 st c_3,rp(2) !r[2]=c3;
605
606 umul a_0,b_3,t_1 !mul_add_c(a[0],b[3],c1,c2,c3);
607 addcc c_1,t_1,c_1 !=
608 rd %y,t_2
609 addxcc c_2,t_2,c_2
610 addx %g0,%g0,c_3
611 umul a_1,b_2,t_1 !=!mul_add_c(a[1],b[2],c1,c2,c3);
612 addcc c_1,t_1,c_1
613 rd %y,t_2
614 addxcc c_2,t_2,c_2
615 addx c_3,%g0,c_3 !=
616 ld ap(3),a_3
617 umul a_2,b_1,t_1 !mul_add_c(a[2],b[1],c1,c2,c3);
618 addcc c_1,t_1,c_1
619 rd %y,t_2 !=
620 addxcc c_2,t_2,c_2
621 addx c_3,%g0,c_3
622 ld ap(4),a_4
623 umul a_3,b_0,t_1 !mul_add_c(a[3],b[0],c1,c2,c3);!=
624 addcc c_1,t_1,c_1
625 rd %y,t_2
626 addxcc c_2,t_2,c_2
627 addx c_3,%g0,c_3 !=
628 st c_1,rp(3) !r[3]=c1;
629
630 umul a_4,b_0,t_1 !mul_add_c(a[4],b[0],c2,c3,c1);
631 addcc c_2,t_1,c_2
632 rd %y,t_2 !=
633 addxcc c_3,t_2,c_3
634 addx %g0,%g0,c_1
635 umul a_3,b_1,t_1 !mul_add_c(a[3],b[1],c2,c3,c1);
636 addcc c_2,t_1,c_2 !=
637 rd %y,t_2
638 addxcc c_3,t_2,c_3
639 addx c_1,%g0,c_1
640 umul a_2,b_2,t_1 !=!mul_add_c(a[2],b[2],c2,c3,c1);
641 addcc c_2,t_1,c_2
642 rd %y,t_2
643 addxcc c_3,t_2,c_3
644 addx c_1,%g0,c_1 !=
645 ld bp(4),b_4
646 umul a_1,b_3,t_1 !mul_add_c(a[1],b[3],c2,c3,c1);
647 addcc c_2,t_1,c_2
648 rd %y,t_2 !=
649 addxcc c_3,t_2,c_3
650 addx c_1,%g0,c_1
651 ld bp(5),b_5
652 umul a_0,b_4,t_1 !=!mul_add_c(a[0],b[4],c2,c3,c1);
653 addcc c_2,t_1,c_2
654 rd %y,t_2
655 addxcc c_3,t_2,c_3
656 addx c_1,%g0,c_1 !=
657 st c_2,rp(4) !r[4]=c2;
658
659 umul a_0,b_5,t_1 !mul_add_c(a[0],b[5],c3,c1,c2);
660 addcc c_3,t_1,c_3
661 rd %y,t_2 !=
662 addxcc c_1,t_2,c_1
663 addx %g0,%g0,c_2
664 umul a_1,b_4,t_1 !mul_add_c(a[1],b[4],c3,c1,c2);
665 addcc c_3,t_1,c_3 !=
666 rd %y,t_2
667 addxcc c_1,t_2,c_1
668 addx c_2,%g0,c_2
669 umul a_2,b_3,t_1 !=!mul_add_c(a[2],b[3],c3,c1,c2);
670 addcc c_3,t_1,c_3
671 rd %y,t_2
672 addxcc c_1,t_2,c_1
673 addx c_2,%g0,c_2 !=
674 umul a_3,b_2,t_1 !mul_add_c(a[3],b[2],c3,c1,c2);
675 addcc c_3,t_1,c_3
676 rd %y,t_2
677 addxcc c_1,t_2,c_1 !=
678 addx c_2,%g0,c_2
679 ld ap(5),a_5
680 umul a_4,b_1,t_1 !mul_add_c(a[4],b[1],c3,c1,c2);
681 addcc c_3,t_1,c_3 !=
682 rd %y,t_2
683 addxcc c_1,t_2,c_1
684 ld ap(6),a_6
685 addx c_2,%g0,c_2 !=
686 umul a_5,b_0,t_1 !mul_add_c(a[5],b[0],c3,c1,c2);
687 addcc c_3,t_1,c_3
688 rd %y,t_2
689 addxcc c_1,t_2,c_1 !=
690 addx c_2,%g0,c_2
691 st c_3,rp(5) !r[5]=c3;
692
693 umul a_6,b_0,t_1 !mul_add_c(a[6],b[0],c1,c2,c3);
694 addcc c_1,t_1,c_1 !=
695 rd %y,t_2
696 addxcc c_2,t_2,c_2
697 addx %g0,%g0,c_3
698 umul a_5,b_1,t_1 !=!mul_add_c(a[5],b[1],c1,c2,c3);
699 addcc c_1,t_1,c_1
700 rd %y,t_2
701 addxcc c_2,t_2,c_2
702 addx c_3,%g0,c_3 !=
703 umul a_4,b_2,t_1 !mul_add_c(a[4],b[2],c1,c2,c3);
704 addcc c_1,t_1,c_1
705 rd %y,t_2
706 addxcc c_2,t_2,c_2 !=
707 addx c_3,%g0,c_3
708 umul a_3,b_3,t_1 !mul_add_c(a[3],b[3],c1,c2,c3);
709 addcc c_1,t_1,c_1
710 rd %y,t_2 !=
711 addxcc c_2,t_2,c_2
712 addx c_3,%g0,c_3
713 umul a_2,b_4,t_1 !mul_add_c(a[2],b[4],c1,c2,c3);
714 addcc c_1,t_1,c_1 !=
715 rd %y,t_2
716 addxcc c_2,t_2,c_2
717 ld bp(6),b_6
718 addx c_3,%g0,c_3 !=
719 umul a_1,b_5,t_1 !mul_add_c(a[1],b[5],c1,c2,c3);
720 addcc c_1,t_1,c_1
721 rd %y,t_2
722 addxcc c_2,t_2,c_2 !=
723 addx c_3,%g0,c_3
724 ld bp(7),b_7
725 umul a_0,b_6,t_1 !mul_add_c(a[0],b[6],c1,c2,c3);
726 addcc c_1,t_1,c_1 !=
727 rd %y,t_2
728 addxcc c_2,t_2,c_2
729 st c_1,rp(6) !r[6]=c1;
730 addx c_3,%g0,c_3 !=
731
732 umul a_0,b_7,t_1 !mul_add_c(a[0],b[7],c2,c3,c1);
733 addcc c_2,t_1,c_2
734 rd %y,t_2
735 addxcc c_3,t_2,c_3 !=
736 addx %g0,%g0,c_1
737 umul a_1,b_6,t_1 !mul_add_c(a[1],b[6],c2,c3,c1);
738 addcc c_2,t_1,c_2
739 rd %y,t_2 !=
740 addxcc c_3,t_2,c_3
741 addx c_1,%g0,c_1
742 umul a_2,b_5,t_1 !mul_add_c(a[2],b[5],c2,c3,c1);
743 addcc c_2,t_1,c_2 !=
744 rd %y,t_2
745 addxcc c_3,t_2,c_3
746 addx c_1,%g0,c_1
747 umul a_3,b_4,t_1 !=!mul_add_c(a[3],b[4],c2,c3,c1);
748 addcc c_2,t_1,c_2
749 rd %y,t_2
750 addxcc c_3,t_2,c_3
751 addx c_1,%g0,c_1 !=
752 umul a_4,b_3,t_1 !mul_add_c(a[4],b[3],c2,c3,c1);
753 addcc c_2,t_1,c_2
754 rd %y,t_2
755 addxcc c_3,t_2,c_3 !=
756 addx c_1,%g0,c_1
757 umul a_5,b_2,t_1 !mul_add_c(a[5],b[2],c2,c3,c1);
758 addcc c_2,t_1,c_2
759 rd %y,t_2 !=
760 addxcc c_3,t_2,c_3
761 addx c_1,%g0,c_1
762 ld ap(7),a_7
763 umul a_6,b_1,t_1 !=!mul_add_c(a[6],b[1],c2,c3,c1);
764 addcc c_2,t_1,c_2
765 rd %y,t_2
766 addxcc c_3,t_2,c_3
767 addx c_1,%g0,c_1 !=
768 umul a_7,b_0,t_1 !mul_add_c(a[7],b[0],c2,c3,c1);
769 addcc c_2,t_1,c_2
770 rd %y,t_2
771 addxcc c_3,t_2,c_3 !=
772 addx c_1,%g0,c_1
773 st c_2,rp(7) !r[7]=c2;
774
775 umul a_7,b_1,t_1 !mul_add_c(a[7],b[1],c3,c1,c2);
776 addcc c_3,t_1,c_3 !=
777 rd %y,t_2
778 addxcc c_1,t_2,c_1
779 addx %g0,%g0,c_2
780 umul a_6,b_2,t_1 !=!mul_add_c(a[6],b[2],c3,c1,c2);
781 addcc c_3,t_1,c_3
782 rd %y,t_2
783 addxcc c_1,t_2,c_1
784 addx c_2,%g0,c_2 !=
785 umul a_5,b_3,t_1 !mul_add_c(a[5],b[3],c3,c1,c2);
786 addcc c_3,t_1,c_3
787 rd %y,t_2
788 addxcc c_1,t_2,c_1 !=
789 addx c_2,%g0,c_2
790 umul a_4,b_4,t_1 !mul_add_c(a[4],b[4],c3,c1,c2);
791 addcc c_3,t_1,c_3
792 rd %y,t_2 !=
793 addxcc c_1,t_2,c_1
794 addx c_2,%g0,c_2
795 umul a_3,b_5,t_1 !mul_add_c(a[3],b[5],c3,c1,c2);
796 addcc c_3,t_1,c_3 !=
797 rd %y,t_2
798 addxcc c_1,t_2,c_1
799 addx c_2,%g0,c_2
800 umul a_2,b_6,t_1 !=!mul_add_c(a[2],b[6],c3,c1,c2);
801 addcc c_3,t_1,c_3
802 rd %y,t_2
803 addxcc c_1,t_2,c_1
804 addx c_2,%g0,c_2 !=
805 umul a_1,b_7,t_1 !mul_add_c(a[1],b[7],c3,c1,c2);
806 addcc c_3,t_1,c_3
807 rd %y,t_2
808 addxcc c_1,t_2,c_1 !
809 addx c_2,%g0,c_2
810 st c_3,rp(8) !r[8]=c3;
811
812 umul a_2,b_7,t_1 !mul_add_c(a[2],b[7],c1,c2,c3);
813 addcc c_1,t_1,c_1 !=
814 rd %y,t_2
815 addxcc c_2,t_2,c_2
816 addx %g0,%g0,c_3
817 umul a_3,b_6,t_1 !=!mul_add_c(a[3],b[6],c1,c2,c3);
818 addcc c_1,t_1,c_1
819 rd %y,t_2
820 addxcc c_2,t_2,c_2
821 addx c_3,%g0,c_3 !=
822 umul a_4,b_5,t_1 !mul_add_c(a[4],b[5],c1,c2,c3);
823 addcc c_1,t_1,c_1
824 rd %y,t_2
825 addxcc c_2,t_2,c_2 !=
826 addx c_3,%g0,c_3
827 umul a_5,b_4,t_1 !mul_add_c(a[5],b[4],c1,c2,c3);
828 addcc c_1,t_1,c_1
829 rd %y,t_2 !=
830 addxcc c_2,t_2,c_2
831 addx c_3,%g0,c_3
832 umul a_6,b_3,t_1 !mul_add_c(a[6],b[3],c1,c2,c3);
833 addcc c_1,t_1,c_1 !=
834 rd %y,t_2
835 addxcc c_2,t_2,c_2
836 addx c_3,%g0,c_3
837 umul a_7,b_2,t_1 !=!mul_add_c(a[7],b[2],c1,c2,c3);
838 addcc c_1,t_1,c_1
839 rd %y,t_2
840 addxcc c_2,t_2,c_2
841 addx c_3,%g0,c_3 !=
842 st c_1,rp(9) !r[9]=c1;
843
844 umul a_7,b_3,t_1 !mul_add_c(a[7],b[3],c2,c3,c1);
845 addcc c_2,t_1,c_2
846 rd %y,t_2 !=
847 addxcc c_3,t_2,c_3
848 addx %g0,%g0,c_1
849 umul a_6,b_4,t_1 !mul_add_c(a[6],b[4],c2,c3,c1);
850 addcc c_2,t_1,c_2 !=
851 rd %y,t_2
852 addxcc c_3,t_2,c_3
853 addx c_1,%g0,c_1
854 umul a_5,b_5,t_1 !=!mul_add_c(a[5],b[5],c2,c3,c1);
855 addcc c_2,t_1,c_2
856 rd %y,t_2
857 addxcc c_3,t_2,c_3
858 addx c_1,%g0,c_1 !=
859 umul a_4,b_6,t_1 !mul_add_c(a[4],b[6],c2,c3,c1);
860 addcc c_2,t_1,c_2
861 rd %y,t_2
862 addxcc c_3,t_2,c_3 !=
863 addx c_1,%g0,c_1
864 umul a_3,b_7,t_1 !mul_add_c(a[3],b[7],c2,c3,c1);
865 addcc c_2,t_1,c_2
866 rd %y,t_2 !=
867 addxcc c_3,t_2,c_3
868 addx c_1,%g0,c_1
869 st c_2,rp(10) !r[10]=c2;
870
871 umul a_4,b_7,t_1 !=!mul_add_c(a[4],b[7],c3,c1,c2);
872 addcc c_3,t_1,c_3
873 rd %y,t_2
874 addxcc c_1,t_2,c_1
875 addx %g0,%g0,c_2 !=
876 umul a_5,b_6,t_1 !mul_add_c(a[5],b[6],c3,c1,c2);
877 addcc c_3,t_1,c_3
878 rd %y,t_2
879 addxcc c_1,t_2,c_1 !=
880 addx c_2,%g0,c_2
881 umul a_6,b_5,t_1 !mul_add_c(a[6],b[5],c3,c1,c2);
882 addcc c_3,t_1,c_3
883 rd %y,t_2 !=
884 addxcc c_1,t_2,c_1
885 addx c_2,%g0,c_2
886 umul a_7,b_4,t_1 !mul_add_c(a[7],b[4],c3,c1,c2);
887 addcc c_3,t_1,c_3 !=
888 rd %y,t_2
889 addxcc c_1,t_2,c_1
890 st c_3,rp(11) !r[11]=c3;
891 addx c_2,%g0,c_2 !=
892
893 umul a_7,b_5,t_1 !mul_add_c(a[7],b[5],c1,c2,c3);
894 addcc c_1,t_1,c_1
895 rd %y,t_2
896 addxcc c_2,t_2,c_2 !=
897 addx %g0,%g0,c_3
898 umul a_6,b_6,t_1 !mul_add_c(a[6],b[6],c1,c2,c3);
899 addcc c_1,t_1,c_1
900 rd %y,t_2 !=
901 addxcc c_2,t_2,c_2
902 addx c_3,%g0,c_3
903 umul a_5,b_7,t_1 !mul_add_c(a[5],b[7],c1,c2,c3);
904 addcc c_1,t_1,c_1 !=
905 rd %y,t_2
906 addxcc c_2,t_2,c_2
907 st c_1,rp(12) !r[12]=c1;
908 addx c_3,%g0,c_3 !=
909
910 umul a_6,b_7,t_1 !mul_add_c(a[6],b[7],c2,c3,c1);
911 addcc c_2,t_1,c_2
912 rd %y,t_2
913 addxcc c_3,t_2,c_3 !=
914 addx %g0,%g0,c_1
915 umul a_7,b_6,t_1 !mul_add_c(a[7],b[6],c2,c3,c1);
916 addcc c_2,t_1,c_2
917 rd %y,t_2 !=
918 addxcc c_3,t_2,c_3
919 addx c_1,%g0,c_1
920 st c_2,rp(13) !r[13]=c2;
921
922 umul a_7,b_7,t_1 !=!mul_add_c(a[7],b[7],c3,c1,c2);
923 addcc c_3,t_1,c_3
924 rd %y,t_2
925 addxcc c_1,t_2,c_1
926 nop !=
927 st c_3,rp(14) !r[14]=c3;
928 st c_1,rp(15) !r[15]=c1;
929
930 ret
931 restore %g0,%g0,%o0
932
933.type bn_mul_comba8,#function
934.size bn_mul_comba8,(.-bn_mul_comba8)
935
936.align 32
937
938.global bn_mul_comba4
939/*
940 * void bn_mul_comba4(r,a,b)
941 * BN_ULONG *r,*a,*b;
942 */
943bn_mul_comba4:
944 save %sp,FRAME_SIZE,%sp
945 ld ap(0),a_0
946 ld bp(0),b_0
947 umul a_0,b_0,c_1 !=!mul_add_c(a[0],b[0],c1,c2,c3);
948 ld bp(1),b_1
949 rd %y,c_2
950 st c_1,rp(0) !r[0]=c1;
951
952 umul a_0,b_1,t_1 !=!mul_add_c(a[0],b[1],c2,c3,c1);
953 ld ap(1),a_1
954 addcc c_2,t_1,c_2
955 rd %y,t_2 !=
956 addxcc %g0,t_2,c_3
957 addx %g0,%g0,c_1
958 ld ap(2),a_2
959 umul a_1,b_0,t_1 !=!mul_add_c(a[1],b[0],c2,c3,c1);
960 addcc c_2,t_1,c_2
961 rd %y,t_2
962 addxcc c_3,t_2,c_3
963 addx c_1,%g0,c_1 !=
964 st c_2,rp(1) !r[1]=c2;
965
966 umul a_2,b_0,t_1 !mul_add_c(a[2],b[0],c3,c1,c2);
967 addcc c_3,t_1,c_3
968 rd %y,t_2 !=
969 addxcc c_1,t_2,c_1
970 addx %g0,%g0,c_2
971 ld bp(2),b_2
972 umul a_1,b_1,t_1 !=!mul_add_c(a[1],b[1],c3,c1,c2);
973 addcc c_3,t_1,c_3
974 rd %y,t_2
975 addxcc c_1,t_2,c_1
976 addx c_2,%g0,c_2 !=
977 ld bp(3),b_3
978 umul a_0,b_2,t_1 !mul_add_c(a[0],b[2],c3,c1,c2);
979 addcc c_3,t_1,c_3
980 rd %y,t_2 !=
981 addxcc c_1,t_2,c_1
982 addx c_2,%g0,c_2
983 st c_3,rp(2) !r[2]=c3;
984
985 umul a_0,b_3,t_1 !=!mul_add_c(a[0],b[3],c1,c2,c3);
986 addcc c_1,t_1,c_1
987 rd %y,t_2
988 addxcc c_2,t_2,c_2
989 addx %g0,%g0,c_3 !=
990 umul a_1,b_2,t_1 !mul_add_c(a[1],b[2],c1,c2,c3);
991 addcc c_1,t_1,c_1
992 rd %y,t_2
993 addxcc c_2,t_2,c_2 !=
994 addx c_3,%g0,c_3
995 ld ap(3),a_3
996 umul a_2,b_1,t_1 !mul_add_c(a[2],b[1],c1,c2,c3);
997 addcc c_1,t_1,c_1 !=
998 rd %y,t_2
999 addxcc c_2,t_2,c_2
1000 addx c_3,%g0,c_3
1001 umul a_3,b_0,t_1 !=!mul_add_c(a[3],b[0],c1,c2,c3);
1002 addcc c_1,t_1,c_1
1003 rd %y,t_2
1004 addxcc c_2,t_2,c_2
1005 addx c_3,%g0,c_3 !=
1006 st c_1,rp(3) !r[3]=c1;
1007
1008 umul a_3,b_1,t_1 !mul_add_c(a[3],b[1],c2,c3,c1);
1009 addcc c_2,t_1,c_2
1010 rd %y,t_2 !=
1011 addxcc c_3,t_2,c_3
1012 addx %g0,%g0,c_1
1013 umul a_2,b_2,t_1 !mul_add_c(a[2],b[2],c2,c3,c1);
1014 addcc c_2,t_1,c_2 !=
1015 rd %y,t_2
1016 addxcc c_3,t_2,c_3
1017 addx c_1,%g0,c_1
1018 umul a_1,b_3,t_1 !=!mul_add_c(a[1],b[3],c2,c3,c1);
1019 addcc c_2,t_1,c_2
1020 rd %y,t_2
1021 addxcc c_3,t_2,c_3
1022 addx c_1,%g0,c_1 !=
1023 st c_2,rp(4) !r[4]=c2;
1024
1025 umul a_2,b_3,t_1 !mul_add_c(a[2],b[3],c3,c1,c2);
1026 addcc c_3,t_1,c_3
1027 rd %y,t_2 !=
1028 addxcc c_1,t_2,c_1
1029 addx %g0,%g0,c_2
1030 umul a_3,b_2,t_1 !mul_add_c(a[3],b[2],c3,c1,c2);
1031 addcc c_3,t_1,c_3 !=
1032 rd %y,t_2
1033 addxcc c_1,t_2,c_1
1034 st c_3,rp(5) !r[5]=c3;
1035 addx c_2,%g0,c_2 !=
1036
1037 umul a_3,b_3,t_1 !mul_add_c(a[3],b[3],c1,c2,c3);
1038 addcc c_1,t_1,c_1
1039 rd %y,t_2
1040 addxcc c_2,t_2,c_2 !=
1041 st c_1,rp(6) !r[6]=c1;
1042 st c_2,rp(7) !r[7]=c2;
1043
1044 ret
1045 restore %g0,%g0,%o0
1046
1047.type bn_mul_comba4,#function
1048.size bn_mul_comba4,(.-bn_mul_comba4)
1049
1050.align 32
1051
1052.global bn_sqr_comba8
1053bn_sqr_comba8:
1054 save %sp,FRAME_SIZE,%sp
1055 ld ap(0),a_0
1056 ld ap(1),a_1
1057 umul a_0,a_0,c_1 !=!sqr_add_c(a,0,c1,c2,c3);
1058 rd %y,c_2
1059 st c_1,rp(0) !r[0]=c1;
1060
1061 ld ap(2),a_2
1062 umul a_0,a_1,t_1 !=!sqr_add_c2(a,1,0,c2,c3,c1);
1063 addcc c_2,t_1,c_2
1064 rd %y,t_2
1065 addxcc %g0,t_2,c_3
1066 addx %g0,%g0,c_1 !=
1067 addcc c_2,t_1,c_2
1068 addxcc c_3,t_2,c_3
1069 st c_2,rp(1) !r[1]=c2;
1070 addx c_1,%g0,c_1 !=
1071
1072 umul a_2,a_0,t_1 !sqr_add_c2(a,2,0,c3,c1,c2);
1073 addcc c_3,t_1,c_3
1074 rd %y,t_2
1075 addxcc c_1,t_2,c_1 !=
1076 addx %g0,%g0,c_2
1077 addcc c_3,t_1,c_3
1078 addxcc c_1,t_2,c_1
1079 addx c_2,%g0,c_2 !=
1080 ld ap(3),a_3
1081 umul a_1,a_1,t_1 !sqr_add_c(a,1,c3,c1,c2);
1082 addcc c_3,t_1,c_3
1083 rd %y,t_2 !=
1084 addxcc c_1,t_2,c_1
1085 addx c_2,%g0,c_2
1086 st c_3,rp(2) !r[2]=c3;
1087
1088 umul a_0,a_3,t_1 !=!sqr_add_c2(a,3,0,c1,c2,c3);
1089 addcc c_1,t_1,c_1
1090 rd %y,t_2
1091 addxcc c_2,t_2,c_2
1092 addx %g0,%g0,c_3 !=
1093 addcc c_1,t_1,c_1
1094 addxcc c_2,t_2,c_2
1095 ld ap(4),a_4
1096 addx c_3,%g0,c_3 !=
1097 umul a_1,a_2,t_1 !sqr_add_c2(a,2,1,c1,c2,c3);
1098 addcc c_1,t_1,c_1
1099 rd %y,t_2
1100 addxcc c_2,t_2,c_2 !=
1101 addx c_3,%g0,c_3
1102 addcc c_1,t_1,c_1
1103 addxcc c_2,t_2,c_2
1104 addx c_3,%g0,c_3 !=
1105 st c_1,rp(3) !r[3]=c1;
1106
1107 umul a_4,a_0,t_1 !sqr_add_c2(a,4,0,c2,c3,c1);
1108 addcc c_2,t_1,c_2
1109 rd %y,t_2 !=
1110 addxcc c_3,t_2,c_3
1111 addx %g0,%g0,c_1
1112 addcc c_2,t_1,c_2
1113 addxcc c_3,t_2,c_3 !=
1114 addx c_1,%g0,c_1
1115 umul a_3,a_1,t_1 !sqr_add_c2(a,3,1,c2,c3,c1);
1116 addcc c_2,t_1,c_2
1117 rd %y,t_2 !=
1118 addxcc c_3,t_2,c_3
1119 addx c_1,%g0,c_1
1120 addcc c_2,t_1,c_2
1121 addxcc c_3,t_2,c_3 !=
1122 addx c_1,%g0,c_1
1123 ld ap(5),a_5
1124 umul a_2,a_2,t_1 !sqr_add_c(a,2,c2,c3,c1);
1125 addcc c_2,t_1,c_2 !=
1126 rd %y,t_2
1127 addxcc c_3,t_2,c_3
1128 st c_2,rp(4) !r[4]=c2;
1129 addx c_1,%g0,c_1 !=
1130
1131 umul a_0,a_5,t_1 !sqr_add_c2(a,5,0,c3,c1,c2);
1132 addcc c_3,t_1,c_3
1133 rd %y,t_2
1134 addxcc c_1,t_2,c_1 !=
1135 addx %g0,%g0,c_2
1136 addcc c_3,t_1,c_3
1137 addxcc c_1,t_2,c_1
1138 addx c_2,%g0,c_2 !=
1139 umul a_1,a_4,t_1 !sqr_add_c2(a,4,1,c3,c1,c2);
1140 addcc c_3,t_1,c_3
1141 rd %y,t_2
1142 addxcc c_1,t_2,c_1 !=
1143 addx c_2,%g0,c_2
1144 addcc c_3,t_1,c_3
1145 addxcc c_1,t_2,c_1
1146 addx c_2,%g0,c_2 !=
1147 ld ap(6),a_6
1148 umul a_2,a_3,t_1 !sqr_add_c2(a,3,2,c3,c1,c2);
1149 addcc c_3,t_1,c_3
1150 rd %y,t_2 !=
1151 addxcc c_1,t_2,c_1
1152 addx c_2,%g0,c_2
1153 addcc c_3,t_1,c_3
1154 addxcc c_1,t_2,c_1 !=
1155 addx c_2,%g0,c_2
1156 st c_3,rp(5) !r[5]=c3;
1157
1158 umul a_6,a_0,t_1 !sqr_add_c2(a,6,0,c1,c2,c3);
1159 addcc c_1,t_1,c_1 !=
1160 rd %y,t_2
1161 addxcc c_2,t_2,c_2
1162 addx %g0,%g0,c_3
1163 addcc c_1,t_1,c_1 !=
1164 addxcc c_2,t_2,c_2
1165 addx c_3,%g0,c_3
1166 umul a_5,a_1,t_1 !sqr_add_c2(a,5,1,c1,c2,c3);
1167 addcc c_1,t_1,c_1 !=
1168 rd %y,t_2
1169 addxcc c_2,t_2,c_2
1170 addx c_3,%g0,c_3
1171 addcc c_1,t_1,c_1 !=
1172 addxcc c_2,t_2,c_2
1173 addx c_3,%g0,c_3
1174 umul a_4,a_2,t_1 !sqr_add_c2(a,4,2,c1,c2,c3);
1175 addcc c_1,t_1,c_1 !=
1176 rd %y,t_2
1177 addxcc c_2,t_2,c_2
1178 addx c_3,%g0,c_3
1179 addcc c_1,t_1,c_1 !=
1180 addxcc c_2,t_2,c_2
1181 addx c_3,%g0,c_3
1182 ld ap(7),a_7
1183 umul a_3,a_3,t_1 !=!sqr_add_c(a,3,c1,c2,c3);
1184 addcc c_1,t_1,c_1
1185 rd %y,t_2
1186 addxcc c_2,t_2,c_2
1187 addx c_3,%g0,c_3 !=
1188 st c_1,rp(6) !r[6]=c1;
1189
1190 umul a_0,a_7,t_1 !sqr_add_c2(a,7,0,c2,c3,c1);
1191 addcc c_2,t_1,c_2
1192 rd %y,t_2 !=
1193 addxcc c_3,t_2,c_3
1194 addx %g0,%g0,c_1
1195 addcc c_2,t_1,c_2
1196 addxcc c_3,t_2,c_3 !=
1197 addx c_1,%g0,c_1
1198 umul a_1,a_6,t_1 !sqr_add_c2(a,6,1,c2,c3,c1);
1199 addcc c_2,t_1,c_2
1200 rd %y,t_2 !=
1201 addxcc c_3,t_2,c_3
1202 addx c_1,%g0,c_1
1203 addcc c_2,t_1,c_2
1204 addxcc c_3,t_2,c_3 !=
1205 addx c_1,%g0,c_1
1206 umul a_2,a_5,t_1 !sqr_add_c2(a,5,2,c2,c3,c1);
1207 addcc c_2,t_1,c_2
1208 rd %y,t_2 !=
1209 addxcc c_3,t_2,c_3
1210 addx c_1,%g0,c_1
1211 addcc c_2,t_1,c_2
1212 addxcc c_3,t_2,c_3 !=
1213 addx c_1,%g0,c_1
1214 umul a_3,a_4,t_1 !sqr_add_c2(a,4,3,c2,c3,c1);
1215 addcc c_2,t_1,c_2
1216 rd %y,t_2 !=
1217 addxcc c_3,t_2,c_3
1218 addx c_1,%g0,c_1
1219 addcc c_2,t_1,c_2
1220 addxcc c_3,t_2,c_3 !=
1221 addx c_1,%g0,c_1
1222 st c_2,rp(7) !r[7]=c2;
1223
1224 umul a_7,a_1,t_1 !sqr_add_c2(a,7,1,c3,c1,c2);
1225 addcc c_3,t_1,c_3 !=
1226 rd %y,t_2
1227 addxcc c_1,t_2,c_1
1228 addx %g0,%g0,c_2
1229 addcc c_3,t_1,c_3 !=
1230 addxcc c_1,t_2,c_1
1231 addx c_2,%g0,c_2
1232 umul a_6,a_2,t_1 !sqr_add_c2(a,6,2,c3,c1,c2);
1233 addcc c_3,t_1,c_3 !=
1234 rd %y,t_2
1235 addxcc c_1,t_2,c_1
1236 addx c_2,%g0,c_2
1237 addcc c_3,t_1,c_3 !=
1238 addxcc c_1,t_2,c_1
1239 addx c_2,%g0,c_2
1240 umul a_5,a_3,t_1 !sqr_add_c2(a,5,3,c3,c1,c2);
1241 addcc c_3,t_1,c_3 !=
1242 rd %y,t_2
1243 addxcc c_1,t_2,c_1
1244 addx c_2,%g0,c_2
1245 addcc c_3,t_1,c_3 !=
1246 addxcc c_1,t_2,c_1
1247 addx c_2,%g0,c_2
1248 umul a_4,a_4,t_1 !sqr_add_c(a,4,c3,c1,c2);
1249 addcc c_3,t_1,c_3 !=
1250 rd %y,t_2
1251 addxcc c_1,t_2,c_1
1252 st c_3,rp(8) !r[8]=c3;
1253 addx c_2,%g0,c_2 !=
1254
1255 umul a_2,a_7,t_1 !sqr_add_c2(a,7,2,c1,c2,c3);
1256 addcc c_1,t_1,c_1
1257 rd %y,t_2
1258 addxcc c_2,t_2,c_2 !=
1259 addx %g0,%g0,c_3
1260 addcc c_1,t_1,c_1
1261 addxcc c_2,t_2,c_2
1262 addx c_3,%g0,c_3 !=
1263 umul a_3,a_6,t_1 !sqr_add_c2(a,6,3,c1,c2,c3);
1264 addcc c_1,t_1,c_1
1265 rd %y,t_2
1266 addxcc c_2,t_2,c_2 !=
1267 addx c_3,%g0,c_3
1268 addcc c_1,t_1,c_1
1269 addxcc c_2,t_2,c_2
1270 addx c_3,%g0,c_3 !=
1271 umul a_4,a_5,t_1 !sqr_add_c2(a,5,4,c1,c2,c3);
1272 addcc c_1,t_1,c_1
1273 rd %y,t_2
1274 addxcc c_2,t_2,c_2 !=
1275 addx c_3,%g0,c_3
1276 addcc c_1,t_1,c_1
1277 addxcc c_2,t_2,c_2
1278 addx c_3,%g0,c_3 !=
1279 st c_1,rp(9) !r[9]=c1;
1280
1281 umul a_7,a_3,t_1 !sqr_add_c2(a,7,3,c2,c3,c1);
1282 addcc c_2,t_1,c_2
1283 rd %y,t_2 !=
1284 addxcc c_3,t_2,c_3
1285 addx %g0,%g0,c_1
1286 addcc c_2,t_1,c_2
1287 addxcc c_3,t_2,c_3 !=
1288 addx c_1,%g0,c_1
1289 umul a_6,a_4,t_1 !sqr_add_c2(a,6,4,c2,c3,c1);
1290 addcc c_2,t_1,c_2
1291 rd %y,t_2 !=
1292 addxcc c_3,t_2,c_3
1293 addx c_1,%g0,c_1
1294 addcc c_2,t_1,c_2
1295 addxcc c_3,t_2,c_3 !=
1296 addx c_1,%g0,c_1
1297 umul a_5,a_5,t_1 !sqr_add_c(a,5,c2,c3,c1);
1298 addcc c_2,t_1,c_2
1299 rd %y,t_2 !=
1300 addxcc c_3,t_2,c_3
1301 addx c_1,%g0,c_1
1302 st c_2,rp(10) !r[10]=c2;
1303
1304 umul a_4,a_7,t_1 !=!sqr_add_c2(a,7,4,c3,c1,c2);
1305 addcc c_3,t_1,c_3
1306 rd %y,t_2
1307 addxcc c_1,t_2,c_1
1308 addx %g0,%g0,c_2 !=
1309 addcc c_3,t_1,c_3
1310 addxcc c_1,t_2,c_1
1311 addx c_2,%g0,c_2
1312 umul a_5,a_6,t_1 !=!sqr_add_c2(a,6,5,c3,c1,c2);
1313 addcc c_3,t_1,c_3
1314 rd %y,t_2
1315 addxcc c_1,t_2,c_1
1316 addx c_2,%g0,c_2 !=
1317 addcc c_3,t_1,c_3
1318 addxcc c_1,t_2,c_1
1319 st c_3,rp(11) !r[11]=c3;
1320 addx c_2,%g0,c_2 !=
1321
1322 umul a_7,a_5,t_1 !sqr_add_c2(a,7,5,c1,c2,c3);
1323 addcc c_1,t_1,c_1
1324 rd %y,t_2
1325 addxcc c_2,t_2,c_2 !=
1326 addx %g0,%g0,c_3
1327 addcc c_1,t_1,c_1
1328 addxcc c_2,t_2,c_2
1329 addx c_3,%g0,c_3 !=
1330 umul a_6,a_6,t_1 !sqr_add_c(a,6,c1,c2,c3);
1331 addcc c_1,t_1,c_1
1332 rd %y,t_2
1333 addxcc c_2,t_2,c_2 !=
1334 addx c_3,%g0,c_3
1335 st c_1,rp(12) !r[12]=c1;
1336
1337 umul a_6,a_7,t_1 !sqr_add_c2(a,7,6,c2,c3,c1);
1338 addcc c_2,t_1,c_2 !=
1339 rd %y,t_2
1340 addxcc c_3,t_2,c_3
1341 addx %g0,%g0,c_1
1342 addcc c_2,t_1,c_2 !=
1343 addxcc c_3,t_2,c_3
1344 st c_2,rp(13) !r[13]=c2;
1345 addx c_1,%g0,c_1 !=
1346
1347 umul a_7,a_7,t_1 !sqr_add_c(a,7,c3,c1,c2);
1348 addcc c_3,t_1,c_3
1349 rd %y,t_2
1350 addxcc c_1,t_2,c_1 !=
1351 st c_3,rp(14) !r[14]=c3;
1352 st c_1,rp(15) !r[15]=c1;
1353
1354 ret
1355 restore %g0,%g0,%o0
1356
1357.type bn_sqr_comba8,#function
1358.size bn_sqr_comba8,(.-bn_sqr_comba8)
1359
1360.align 32
1361
1362.global bn_sqr_comba4
1363/*
1364 * void bn_sqr_comba4(r,a)
1365 * BN_ULONG *r,*a;
1366 */
1367bn_sqr_comba4:
1368 save %sp,FRAME_SIZE,%sp
1369 ld ap(0),a_0
1370 umul a_0,a_0,c_1 !sqr_add_c(a,0,c1,c2,c3);
1371 ld ap(1),a_1 !=
1372 rd %y,c_2
1373 st c_1,rp(0) !r[0]=c1;
1374
1375 ld ap(2),a_2
1376 umul a_0,a_1,t_1 !=!sqr_add_c2(a,1,0,c2,c3,c1);
1377 addcc c_2,t_1,c_2
1378 rd %y,t_2
1379 addxcc %g0,t_2,c_3
1380 addx %g0,%g0,c_1 !=
1381 addcc c_2,t_1,c_2
1382 addxcc c_3,t_2,c_3
1383 addx c_1,%g0,c_1 !=
1384 st c_2,rp(1) !r[1]=c2;
1385
1386 umul a_2,a_0,t_1 !sqr_add_c2(a,2,0,c3,c1,c2);
1387 addcc c_3,t_1,c_3
1388 rd %y,t_2 !=
1389 addxcc c_1,t_2,c_1
1390 addx %g0,%g0,c_2
1391 addcc c_3,t_1,c_3
1392 addxcc c_1,t_2,c_1 !=
1393 addx c_2,%g0,c_2
1394 ld ap(3),a_3
1395 umul a_1,a_1,t_1 !sqr_add_c(a,1,c3,c1,c2);
1396 addcc c_3,t_1,c_3 !=
1397 rd %y,t_2
1398 addxcc c_1,t_2,c_1
1399 st c_3,rp(2) !r[2]=c3;
1400 addx c_2,%g0,c_2 !=
1401
1402 umul a_0,a_3,t_1 !sqr_add_c2(a,3,0,c1,c2,c3);
1403 addcc c_1,t_1,c_1
1404 rd %y,t_2
1405 addxcc c_2,t_2,c_2 !=
1406 addx %g0,%g0,c_3
1407 addcc c_1,t_1,c_1
1408 addxcc c_2,t_2,c_2
1409 addx c_3,%g0,c_3 !=
1410 umul a_1,a_2,t_1 !sqr_add_c2(a,2,1,c1,c2,c3);
1411 addcc c_1,t_1,c_1
1412 rd %y,t_2
1413 addxcc c_2,t_2,c_2 !=
1414 addx c_3,%g0,c_3
1415 addcc c_1,t_1,c_1
1416 addxcc c_2,t_2,c_2
1417 addx c_3,%g0,c_3 !=
1418 st c_1,rp(3) !r[3]=c1;
1419
1420 umul a_3,a_1,t_1 !sqr_add_c2(a,3,1,c2,c3,c1);
1421 addcc c_2,t_1,c_2
1422 rd %y,t_2 !=
1423 addxcc c_3,t_2,c_3
1424 addx %g0,%g0,c_1
1425 addcc c_2,t_1,c_2
1426 addxcc c_3,t_2,c_3 !=
1427 addx c_1,%g0,c_1
1428 umul a_2,a_2,t_1 !sqr_add_c(a,2,c2,c3,c1);
1429 addcc c_2,t_1,c_2
1430 rd %y,t_2 !=
1431 addxcc c_3,t_2,c_3
1432 addx c_1,%g0,c_1
1433 st c_2,rp(4) !r[4]=c2;
1434
1435 umul a_2,a_3,t_1 !=!sqr_add_c2(a,3,2,c3,c1,c2);
1436 addcc c_3,t_1,c_3
1437 rd %y,t_2
1438 addxcc c_1,t_2,c_1
1439 addx %g0,%g0,c_2 !=
1440 addcc c_3,t_1,c_3
1441 addxcc c_1,t_2,c_1
1442 st c_3,rp(5) !r[5]=c3;
1443 addx c_2,%g0,c_2 !=
1444
1445 umul a_3,a_3,t_1 !sqr_add_c(a,3,c1,c2,c3);
1446 addcc c_1,t_1,c_1
1447 rd %y,t_2
1448 addxcc c_2,t_2,c_2 !=
1449 st c_1,rp(6) !r[6]=c1;
1450 st c_2,rp(7) !r[7]=c2;
1451
1452 ret
1453 restore %g0,%g0,%o0
1454
1455.type bn_sqr_comba4,#function
1456.size bn_sqr_comba4,(.-bn_sqr_comba4)
1457
1458.align 32
diff --git a/src/lib/libcrypto/bn/asm/sparcv8plus.S b/src/lib/libcrypto/bn/asm/sparcv8plus.S
deleted file mode 100644
index 0074dfdb75..0000000000
--- a/src/lib/libcrypto/bn/asm/sparcv8plus.S
+++ /dev/null
@@ -1,1535 +0,0 @@
1.ident "sparcv8plus.s, Version 1.4"
2.ident "SPARC v9 ISA artwork by Andy Polyakov <appro@fy.chalmers.se>"
3
4/*
5 * ====================================================================
6 * Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
7 * project.
8 *
9 * Rights for redistribution and usage in source and binary forms are
10 * granted according to the OpenSSL license. Warranty of any kind is
11 * disclaimed.
12 * ====================================================================
13 */
14
15/*
16 * This is my modest contributon to OpenSSL project (see
17 * http://www.openssl.org/ for more information about it) and is
18 * a drop-in UltraSPARC ISA replacement for crypto/bn/bn_asm.c
19 * module. For updates see http://fy.chalmers.se/~appro/hpe/.
20 *
21 * Questions-n-answers.
22 *
23 * Q. How to compile?
24 * A. With SC4.x/SC5.x:
25 *
26 * cc -xarch=v8plus -c bn_asm.sparc.v8plus.S -o bn_asm.o
27 *
28 * and with gcc:
29 *
30 * gcc -mcpu=ultrasparc -c bn_asm.sparc.v8plus.S -o bn_asm.o
31 *
32 * or if above fails (it does if you have gas installed):
33 *
34 * gcc -E bn_asm.sparc.v8plus.S | as -xarch=v8plus /dev/fd/0 -o bn_asm.o
35 *
36 * Quick-n-dirty way to fuse the module into the library.
37 * Provided that the library is already configured and built
38 * (in 0.9.2 case with no-asm option):
39 *
40 * # cd crypto/bn
41 * # cp /some/place/bn_asm.sparc.v8plus.S .
42 * # cc -xarch=v8plus -c bn_asm.sparc.v8plus.S -o bn_asm.o
43 * # make
44 * # cd ../..
45 * # make; make test
46 *
47 * Quick-n-dirty way to get rid of it:
48 *
49 * # cd crypto/bn
50 * # touch bn_asm.c
51 * # make
52 * # cd ../..
53 * # make; make test
54 *
55 * Q. V8plus achitecture? What kind of beast is that?
56 * A. Well, it's rather a programming model than an architecture...
57 * It's actually v9-compliant, i.e. *any* UltraSPARC, CPU under
58 * special conditions, namely when kernel doesn't preserve upper
59 * 32 bits of otherwise 64-bit registers during a context switch.
60 *
61 * Q. Why just UltraSPARC? What about SuperSPARC?
62 * A. Original release did target UltraSPARC only. Now SuperSPARC
63 * version is provided along. Both version share bn_*comba[48]
64 * implementations (see comment later in code for explanation).
65 * But what's so special about this UltraSPARC implementation?
66 * Why didn't I let compiler do the job? Trouble is that most of
67 * available compilers (well, SC5.0 is the only exception) don't
68 * attempt to take advantage of UltraSPARC's 64-bitness under
69 * 32-bit kernels even though it's perfectly possible (see next
70 * question).
71 *
72 * Q. 64-bit registers under 32-bit kernels? Didn't you just say it
73 * doesn't work?
74 * A. You can't adress *all* registers as 64-bit wide:-( The catch is
75 * that you actually may rely upon %o0-%o5 and %g1-%g4 being fully
76 * preserved if you're in a leaf function, i.e. such never calling
77 * any other functions. All functions in this module are leaf and
78 * 10 registers is a handful. And as a matter of fact none-"comba"
79 * routines don't require even that much and I could even afford to
80 * not allocate own stack frame for 'em:-)
81 *
82 * Q. What about 64-bit kernels?
83 * A. What about 'em? Just kidding:-) Pure 64-bit version is currently
84 * under evaluation and development...
85 *
86 * Q. What about shared libraries?
87 * A. What about 'em? Kidding again:-) Code does *not* contain any
88 * code position dependencies and it's safe to include it into
89 * shared library as is.
90 *
91 * Q. How much faster does it go?
92 * A. Do you have a good benchmark? In either case below is what I
93 * experience with crypto/bn/expspeed.c test program:
94 *
95 * v8plus module on U10/300MHz against bn_asm.c compiled with:
96 *
97 * cc-5.0 -xarch=v8plus -xO5 -xdepend +7-12%
98 * cc-4.2 -xarch=v8plus -xO5 -xdepend +25-35%
99 * egcs-1.1.2 -mcpu=ultrasparc -O3 +35-45%
100 *
101 * v8 module on SS10/60MHz against bn_asm.c compiled with:
102 *
103 * cc-5.0 -xarch=v8 -xO5 -xdepend +7-10%
104 * cc-4.2 -xarch=v8 -xO5 -xdepend +10%
105 * egcs-1.1.2 -mv8 -O3 +35-45%
106 *
107 * As you can see it's damn hard to beat the new Sun C compiler
108 * and it's in first place GNU C users who will appreciate this
109 * assembler implementation:-)
110 */
111
112/*
113 * Revision history.
114 *
115 * 1.0 - initial release;
116 * 1.1 - new loop unrolling model(*);
117 * - some more fine tuning;
118 * 1.2 - made gas friendly;
119 * - updates to documentation concerning v9;
120 * - new performance comparison matrix;
121 * 1.3 - fixed problem with /usr/ccs/lib/cpp;
122 * 1.4 - native V9 bn_*_comba[48] implementation (15% more efficient)
123 * resulting in slight overall performance kick;
124 * - some retunes;
125 * - support for GNU as added;
126 *
127 * (*) Originally unrolled loop looked like this:
128 * for (;;) {
129 * op(p+0); if (--n==0) break;
130 * op(p+1); if (--n==0) break;
131 * op(p+2); if (--n==0) break;
132 * op(p+3); if (--n==0) break;
133 * p+=4;
134 * }
135 * I unroll according to following:
136 * while (n&~3) {
137 * op(p+0); op(p+1); op(p+2); op(p+3);
138 * p+=4; n=-4;
139 * }
140 * if (n) {
141 * op(p+0); if (--n==0) return;
142 * op(p+2); if (--n==0) return;
143 * op(p+3); return;
144 * }
145 */
146
147/*
148 * GNU assembler can't stand stuw:-(
149 */
150#define stuw st
151
152.section ".text",#alloc,#execinstr
153.file "bn_asm.sparc.v8plus.S"
154
155.align 32
156
157.global bn_mul_add_words
158/*
159 * BN_ULONG bn_mul_add_words(rp,ap,num,w)
160 * BN_ULONG *rp,*ap;
161 * int num;
162 * BN_ULONG w;
163 */
164bn_mul_add_words:
165 brgz,a %o2,.L_bn_mul_add_words_proceed
166 lduw [%o1],%g2
167 retl
168 clr %o0
169
170.L_bn_mul_add_words_proceed:
171 srl %o3,%g0,%o3 ! clruw %o3
172 andcc %o2,-4,%g0
173 bz,pn %icc,.L_bn_mul_add_words_tail
174 clr %o5
175
176.L_bn_mul_add_words_loop: ! wow! 32 aligned!
177 lduw [%o0],%g1
178 lduw [%o1+4],%g3
179 mulx %o3,%g2,%g2
180 add %g1,%o5,%o4
181 nop
182 add %o4,%g2,%o4
183 stuw %o4,[%o0]
184 srlx %o4,32,%o5
185
186 lduw [%o0+4],%g1
187 lduw [%o1+8],%g2
188 mulx %o3,%g3,%g3
189 add %g1,%o5,%o4
190 dec 4,%o2
191 add %o4,%g3,%o4
192 stuw %o4,[%o0+4]
193 srlx %o4,32,%o5
194
195 lduw [%o0+8],%g1
196 lduw [%o1+12],%g3
197 mulx %o3,%g2,%g2
198 add %g1,%o5,%o4
199 inc 16,%o1
200 add %o4,%g2,%o4
201 stuw %o4,[%o0+8]
202 srlx %o4,32,%o5
203
204 lduw [%o0+12],%g1
205 mulx %o3,%g3,%g3
206 add %g1,%o5,%o4
207 inc 16,%o0
208 add %o4,%g3,%o4
209 andcc %o2,-4,%g0
210 stuw %o4,[%o0-4]
211 srlx %o4,32,%o5
212 bnz,a,pt %icc,.L_bn_mul_add_words_loop
213 lduw [%o1],%g2
214
215 brnz,a,pn %o2,.L_bn_mul_add_words_tail
216 lduw [%o1],%g2
217.L_bn_mul_add_words_return:
218 retl
219 mov %o5,%o0
220
221.L_bn_mul_add_words_tail:
222 lduw [%o0],%g1
223 mulx %o3,%g2,%g2
224 add %g1,%o5,%o4
225 dec %o2
226 add %o4,%g2,%o4
227 srlx %o4,32,%o5
228 brz,pt %o2,.L_bn_mul_add_words_return
229 stuw %o4,[%o0]
230
231 lduw [%o1+4],%g2
232 lduw [%o0+4],%g1
233 mulx %o3,%g2,%g2
234 add %g1,%o5,%o4
235 dec %o2
236 add %o4,%g2,%o4
237 srlx %o4,32,%o5
238 brz,pt %o2,.L_bn_mul_add_words_return
239 stuw %o4,[%o0+4]
240
241 lduw [%o1+8],%g2
242 lduw [%o0+8],%g1
243 mulx %o3,%g2,%g2
244 add %g1,%o5,%o4
245 add %o4,%g2,%o4
246 stuw %o4,[%o0+8]
247 retl
248 srlx %o4,32,%o0
249
250.type bn_mul_add_words,#function
251.size bn_mul_add_words,(.-bn_mul_add_words)
252
253.align 32
254
255.global bn_mul_words
256/*
257 * BN_ULONG bn_mul_words(rp,ap,num,w)
258 * BN_ULONG *rp,*ap;
259 * int num;
260 * BN_ULONG w;
261 */
262bn_mul_words:
263 brgz,a %o2,.L_bn_mul_words_proceeed
264 lduw [%o1],%g2
265 retl
266 clr %o0
267
268.L_bn_mul_words_proceeed:
269 srl %o3,%g0,%o3 ! clruw %o3
270 andcc %o2,-4,%g0
271 bz,pn %icc,.L_bn_mul_words_tail
272 clr %o5
273
274.L_bn_mul_words_loop: ! wow! 32 aligned!
275 lduw [%o1+4],%g3
276 mulx %o3,%g2,%g2
277 add %g2,%o5,%o4
278 nop
279 stuw %o4,[%o0]
280 srlx %o4,32,%o5
281
282 lduw [%o1+8],%g2
283 mulx %o3,%g3,%g3
284 add %g3,%o5,%o4
285 dec 4,%o2
286 stuw %o4,[%o0+4]
287 srlx %o4,32,%o5
288
289 lduw [%o1+12],%g3
290 mulx %o3,%g2,%g2
291 add %g2,%o5,%o4
292 inc 16,%o1
293 stuw %o4,[%o0+8]
294 srlx %o4,32,%o5
295
296 mulx %o3,%g3,%g3
297 add %g3,%o5,%o4
298 inc 16,%o0
299 stuw %o4,[%o0-4]
300 srlx %o4,32,%o5
301 andcc %o2,-4,%g0
302 bnz,a,pt %icc,.L_bn_mul_words_loop
303 lduw [%o1],%g2
304 nop
305 nop
306
307 brnz,a,pn %o2,.L_bn_mul_words_tail
308 lduw [%o1],%g2
309.L_bn_mul_words_return:
310 retl
311 mov %o5,%o0
312
313.L_bn_mul_words_tail:
314 mulx %o3,%g2,%g2
315 add %g2,%o5,%o4
316 dec %o2
317 srlx %o4,32,%o5
318 brz,pt %o2,.L_bn_mul_words_return
319 stuw %o4,[%o0]
320
321 lduw [%o1+4],%g2
322 mulx %o3,%g2,%g2
323 add %g2,%o5,%o4
324 dec %o2
325 srlx %o4,32,%o5
326 brz,pt %o2,.L_bn_mul_words_return
327 stuw %o4,[%o0+4]
328
329 lduw [%o1+8],%g2
330 mulx %o3,%g2,%g2
331 add %g2,%o5,%o4
332 stuw %o4,[%o0+8]
333 retl
334 srlx %o4,32,%o0
335
336.type bn_mul_words,#function
337.size bn_mul_words,(.-bn_mul_words)
338
339.align 32
340.global bn_sqr_words
341/*
342 * void bn_sqr_words(r,a,n)
343 * BN_ULONG *r,*a;
344 * int n;
345 */
346bn_sqr_words:
347 brgz,a %o2,.L_bn_sqr_words_proceeed
348 lduw [%o1],%g2
349 retl
350 clr %o0
351
352.L_bn_sqr_words_proceeed:
353 andcc %o2,-4,%g0
354 nop
355 bz,pn %icc,.L_bn_sqr_words_tail
356 nop
357
358.L_bn_sqr_words_loop: ! wow! 32 aligned!
359 lduw [%o1+4],%g3
360 mulx %g2,%g2,%o4
361 stuw %o4,[%o0]
362 srlx %o4,32,%o5
363 stuw %o5,[%o0+4]
364 nop
365
366 lduw [%o1+8],%g2
367 mulx %g3,%g3,%o4
368 dec 4,%o2
369 stuw %o4,[%o0+8]
370 srlx %o4,32,%o5
371 stuw %o5,[%o0+12]
372
373 lduw [%o1+12],%g3
374 mulx %g2,%g2,%o4
375 srlx %o4,32,%o5
376 stuw %o4,[%o0+16]
377 inc 16,%o1
378 stuw %o5,[%o0+20]
379
380 mulx %g3,%g3,%o4
381 inc 32,%o0
382 stuw %o4,[%o0-8]
383 srlx %o4,32,%o5
384 andcc %o2,-4,%g2
385 stuw %o5,[%o0-4]
386 bnz,a,pt %icc,.L_bn_sqr_words_loop
387 lduw [%o1],%g2
388 nop
389
390 brnz,a,pn %o2,.L_bn_sqr_words_tail
391 lduw [%o1],%g2
392.L_bn_sqr_words_return:
393 retl
394 clr %o0
395
396.L_bn_sqr_words_tail:
397 mulx %g2,%g2,%o4
398 dec %o2
399 stuw %o4,[%o0]
400 srlx %o4,32,%o5
401 brz,pt %o2,.L_bn_sqr_words_return
402 stuw %o5,[%o0+4]
403
404 lduw [%o1+4],%g2
405 mulx %g2,%g2,%o4
406 dec %o2
407 stuw %o4,[%o0+8]
408 srlx %o4,32,%o5
409 brz,pt %o2,.L_bn_sqr_words_return
410 stuw %o5,[%o0+12]
411
412 lduw [%o1+8],%g2
413 mulx %g2,%g2,%o4
414 srlx %o4,32,%o5
415 stuw %o4,[%o0+16]
416 stuw %o5,[%o0+20]
417 retl
418 clr %o0
419
420.type bn_sqr_words,#function
421.size bn_sqr_words,(.-bn_sqr_words)
422
423.align 32
424.global bn_div_words
425/*
426 * BN_ULONG bn_div_words(h,l,d)
427 * BN_ULONG h,l,d;
428 */
429bn_div_words:
430 sllx %o0,32,%o0
431 or %o0,%o1,%o0
432 udivx %o0,%o2,%o0
433 retl
434 srl %o0,%g0,%o0 ! clruw %o0
435
436.type bn_div_words,#function
437.size bn_div_words,(.-bn_div_words)
438
439.align 32
440
441.global bn_add_words
442/*
443 * BN_ULONG bn_add_words(rp,ap,bp,n)
444 * BN_ULONG *rp,*ap,*bp;
445 * int n;
446 */
447bn_add_words:
448 brgz,a %o3,.L_bn_add_words_proceed
449 lduw [%o1],%o4
450 retl
451 clr %o0
452
453.L_bn_add_words_proceed:
454 andcc %o3,-4,%g0
455 bz,pn %icc,.L_bn_add_words_tail
456 addcc %g0,0,%g0 ! clear carry flag
457 nop
458
459.L_bn_add_words_loop: ! wow! 32 aligned!
460 dec 4,%o3
461 lduw [%o2],%o5
462 lduw [%o1+4],%g1
463 lduw [%o2+4],%g2
464 lduw [%o1+8],%g3
465 lduw [%o2+8],%g4
466 addccc %o5,%o4,%o5
467 stuw %o5,[%o0]
468
469 lduw [%o1+12],%o4
470 lduw [%o2+12],%o5
471 inc 16,%o1
472 addccc %g1,%g2,%g1
473 stuw %g1,[%o0+4]
474
475 inc 16,%o2
476 addccc %g3,%g4,%g3
477 stuw %g3,[%o0+8]
478
479 inc 16,%o0
480 addccc %o5,%o4,%o5
481 stuw %o5,[%o0-4]
482 and %o3,-4,%g1
483 brnz,a,pt %g1,.L_bn_add_words_loop
484 lduw [%o1],%o4
485
486 brnz,a,pn %o3,.L_bn_add_words_tail
487 lduw [%o1],%o4
488.L_bn_add_words_return:
489 clr %o0
490 retl
491 movcs %icc,1,%o0
492 nop
493
494.L_bn_add_words_tail:
495 lduw [%o2],%o5
496 dec %o3
497 addccc %o5,%o4,%o5
498 brz,pt %o3,.L_bn_add_words_return
499 stuw %o5,[%o0]
500
501 lduw [%o1+4],%o4
502 lduw [%o2+4],%o5
503 dec %o3
504 addccc %o5,%o4,%o5
505 brz,pt %o3,.L_bn_add_words_return
506 stuw %o5,[%o0+4]
507
508 lduw [%o1+8],%o4
509 lduw [%o2+8],%o5
510 addccc %o5,%o4,%o5
511 stuw %o5,[%o0+8]
512 clr %o0
513 retl
514 movcs %icc,1,%o0
515
516.type bn_add_words,#function
517.size bn_add_words,(.-bn_add_words)
518
519.global bn_sub_words
520/*
521 * BN_ULONG bn_sub_words(rp,ap,bp,n)
522 * BN_ULONG *rp,*ap,*bp;
523 * int n;
524 */
525bn_sub_words:
526 brgz,a %o3,.L_bn_sub_words_proceed
527 lduw [%o1],%o4
528 retl
529 clr %o0
530
531.L_bn_sub_words_proceed:
532 andcc %o3,-4,%g0
533 bz,pn %icc,.L_bn_sub_words_tail
534 addcc %g0,0,%g0 ! clear carry flag
535 nop
536
537.L_bn_sub_words_loop: ! wow! 32 aligned!
538 dec 4,%o3
539 lduw [%o2],%o5
540 lduw [%o1+4],%g1
541 lduw [%o2+4],%g2
542 lduw [%o1+8],%g3
543 lduw [%o2+8],%g4
544 subccc %o4,%o5,%o5
545 stuw %o5,[%o0]
546
547 lduw [%o1+12],%o4
548 lduw [%o2+12],%o5
549 inc 16,%o1
550 subccc %g1,%g2,%g2
551 stuw %g2,[%o0+4]
552
553 inc 16,%o2
554 subccc %g3,%g4,%g4
555 stuw %g4,[%o0+8]
556
557 inc 16,%o0
558 subccc %o4,%o5,%o5
559 stuw %o5,[%o0-4]
560 and %o3,-4,%g1
561 brnz,a,pt %g1,.L_bn_sub_words_loop
562 lduw [%o1],%o4
563
564 brnz,a,pn %o3,.L_bn_sub_words_tail
565 lduw [%o1],%o4
566.L_bn_sub_words_return:
567 clr %o0
568 retl
569 movcs %icc,1,%o0
570 nop
571
572.L_bn_sub_words_tail: ! wow! 32 aligned!
573 lduw [%o2],%o5
574 dec %o3
575 subccc %o4,%o5,%o5
576 brz,pt %o3,.L_bn_sub_words_return
577 stuw %o5,[%o0]
578
579 lduw [%o1+4],%o4
580 lduw [%o2+4],%o5
581 dec %o3
582 subccc %o4,%o5,%o5
583 brz,pt %o3,.L_bn_sub_words_return
584 stuw %o5,[%o0+4]
585
586 lduw [%o1+8],%o4
587 lduw [%o2+8],%o5
588 subccc %o4,%o5,%o5
589 stuw %o5,[%o0+8]
590 clr %o0
591 retl
592 movcs %icc,1,%o0
593
594.type bn_sub_words,#function
595.size bn_sub_words,(.-bn_sub_words)
596
597/*
598 * Code below depends on the fact that upper parts of the %l0-%l7
599 * and %i0-%i7 are zeroed by kernel after context switch. In
600 * previous versions this comment stated that "the trouble is that
601 * it's not feasible to implement the mumbo-jumbo in less V9
602 * instructions:-(" which apparently isn't true thanks to
603 * 'bcs,a %xcc,.+8; inc %rd' pair. But the performance improvement
604 * results not from the shorter code, but from elimination of
605 * multicycle none-pairable 'rd %y,%rd' instructions.
606 *
607 * Andy.
608 */
609
610#define FRAME_SIZE -96
611
612/*
613 * Here is register usage map for *all* routines below.
614 */
615#define t_1 %o0
616#define t_2 %o1
617#define c_12 %o2
618#define c_3 %o3
619
620#define ap(I) [%i1+4*I]
621#define bp(I) [%i2+4*I]
622#define rp(I) [%i0+4*I]
623
624#define a_0 %l0
625#define a_1 %l1
626#define a_2 %l2
627#define a_3 %l3
628#define a_4 %l4
629#define a_5 %l5
630#define a_6 %l6
631#define a_7 %l7
632
633#define b_0 %i3
634#define b_1 %i4
635#define b_2 %i5
636#define b_3 %o4
637#define b_4 %o5
638#define b_5 %o7
639#define b_6 %g1
640#define b_7 %g4
641
642.align 32
643.global bn_mul_comba8
644/*
645 * void bn_mul_comba8(r,a,b)
646 * BN_ULONG *r,*a,*b;
647 */
648bn_mul_comba8:
649 save %sp,FRAME_SIZE,%sp
650 mov 1,t_2
651 lduw ap(0),a_0
652 sllx t_2,32,t_2
653 lduw bp(0),b_0 !=
654 lduw bp(1),b_1
655 mulx a_0,b_0,t_1 !mul_add_c(a[0],b[0],c1,c2,c3);
656 srlx t_1,32,c_12
657 stuw t_1,rp(0) !=!r[0]=c1;
658
659 lduw ap(1),a_1
660 mulx a_0,b_1,t_1 !mul_add_c(a[0],b[1],c2,c3,c1);
661 addcc c_12,t_1,c_12
662 clr c_3 !=
663 bcs,a %xcc,.+8
664 add c_3,t_2,c_3
665 lduw ap(2),a_2
666 mulx a_1,b_0,t_1 !=!mul_add_c(a[1],b[0],c2,c3,c1);
667 addcc c_12,t_1,t_1
668 bcs,a %xcc,.+8
669 add c_3,t_2,c_3
670 srlx t_1,32,c_12 !=
671 stuw t_1,rp(1) !r[1]=c2;
672 or c_12,c_3,c_12
673
674 mulx a_2,b_0,t_1 !mul_add_c(a[2],b[0],c3,c1,c2);
675 addcc c_12,t_1,c_12 !=
676 clr c_3
677 bcs,a %xcc,.+8
678 add c_3,t_2,c_3
679 lduw bp(2),b_2 !=
680 mulx a_1,b_1,t_1 !mul_add_c(a[1],b[1],c3,c1,c2);
681 addcc c_12,t_1,c_12
682 bcs,a %xcc,.+8
683 add c_3,t_2,c_3 !=
684 lduw bp(3),b_3
685 mulx a_0,b_2,t_1 !mul_add_c(a[0],b[2],c3,c1,c2);
686 addcc c_12,t_1,t_1
687 bcs,a %xcc,.+8 !=
688 add c_3,t_2,c_3
689 srlx t_1,32,c_12
690 stuw t_1,rp(2) !r[2]=c3;
691 or c_12,c_3,c_12 !=
692
693 mulx a_0,b_3,t_1 !mul_add_c(a[0],b[3],c1,c2,c3);
694 addcc c_12,t_1,c_12
695 clr c_3
696 bcs,a %xcc,.+8 !=
697 add c_3,t_2,c_3
698 mulx a_1,b_2,t_1 !=!mul_add_c(a[1],b[2],c1,c2,c3);
699 addcc c_12,t_1,c_12
700 bcs,a %xcc,.+8 !=
701 add c_3,t_2,c_3
702 lduw ap(3),a_3
703 mulx a_2,b_1,t_1 !mul_add_c(a[2],b[1],c1,c2,c3);
704 addcc c_12,t_1,c_12 !=
705 bcs,a %xcc,.+8
706 add c_3,t_2,c_3
707 lduw ap(4),a_4
708 mulx a_3,b_0,t_1 !=!mul_add_c(a[3],b[0],c1,c2,c3);!=
709 addcc c_12,t_1,t_1
710 bcs,a %xcc,.+8
711 add c_3,t_2,c_3
712 srlx t_1,32,c_12 !=
713 stuw t_1,rp(3) !r[3]=c1;
714 or c_12,c_3,c_12
715
716 mulx a_4,b_0,t_1 !mul_add_c(a[4],b[0],c2,c3,c1);
717 addcc c_12,t_1,c_12 !=
718 clr c_3
719 bcs,a %xcc,.+8
720 add c_3,t_2,c_3
721 mulx a_3,b_1,t_1 !=!mul_add_c(a[3],b[1],c2,c3,c1);
722 addcc c_12,t_1,c_12
723 bcs,a %xcc,.+8
724 add c_3,t_2,c_3
725 mulx a_2,b_2,t_1 !=!mul_add_c(a[2],b[2],c2,c3,c1);
726 addcc c_12,t_1,c_12
727 bcs,a %xcc,.+8
728 add c_3,t_2,c_3
729 lduw bp(4),b_4 !=
730 mulx a_1,b_3,t_1 !mul_add_c(a[1],b[3],c2,c3,c1);
731 addcc c_12,t_1,c_12
732 bcs,a %xcc,.+8
733 add c_3,t_2,c_3 !=
734 lduw bp(5),b_5
735 mulx a_0,b_4,t_1 !mul_add_c(a[0],b[4],c2,c3,c1);
736 addcc c_12,t_1,t_1
737 bcs,a %xcc,.+8 !=
738 add c_3,t_2,c_3
739 srlx t_1,32,c_12
740 stuw t_1,rp(4) !r[4]=c2;
741 or c_12,c_3,c_12 !=
742
743 mulx a_0,b_5,t_1 !mul_add_c(a[0],b[5],c3,c1,c2);
744 addcc c_12,t_1,c_12
745 clr c_3
746 bcs,a %xcc,.+8 !=
747 add c_3,t_2,c_3
748 mulx a_1,b_4,t_1 !mul_add_c(a[1],b[4],c3,c1,c2);
749 addcc c_12,t_1,c_12
750 bcs,a %xcc,.+8 !=
751 add c_3,t_2,c_3
752 mulx a_2,b_3,t_1 !mul_add_c(a[2],b[3],c3,c1,c2);
753 addcc c_12,t_1,c_12
754 bcs,a %xcc,.+8 !=
755 add c_3,t_2,c_3
756 mulx a_3,b_2,t_1 !mul_add_c(a[3],b[2],c3,c1,c2);
757 addcc c_12,t_1,c_12
758 bcs,a %xcc,.+8 !=
759 add c_3,t_2,c_3
760 lduw ap(5),a_5
761 mulx a_4,b_1,t_1 !mul_add_c(a[4],b[1],c3,c1,c2);
762 addcc c_12,t_1,c_12 !=
763 bcs,a %xcc,.+8
764 add c_3,t_2,c_3
765 lduw ap(6),a_6
766 mulx a_5,b_0,t_1 !=!mul_add_c(a[5],b[0],c3,c1,c2);
767 addcc c_12,t_1,t_1
768 bcs,a %xcc,.+8
769 add c_3,t_2,c_3
770 srlx t_1,32,c_12 !=
771 stuw t_1,rp(5) !r[5]=c3;
772 or c_12,c_3,c_12
773
774 mulx a_6,b_0,t_1 !mul_add_c(a[6],b[0],c1,c2,c3);
775 addcc c_12,t_1,c_12 !=
776 clr c_3
777 bcs,a %xcc,.+8
778 add c_3,t_2,c_3
779 mulx a_5,b_1,t_1 !=!mul_add_c(a[5],b[1],c1,c2,c3);
780 addcc c_12,t_1,c_12
781 bcs,a %xcc,.+8
782 add c_3,t_2,c_3
783 mulx a_4,b_2,t_1 !=!mul_add_c(a[4],b[2],c1,c2,c3);
784 addcc c_12,t_1,c_12
785 bcs,a %xcc,.+8
786 add c_3,t_2,c_3
787 mulx a_3,b_3,t_1 !=!mul_add_c(a[3],b[3],c1,c2,c3);
788 addcc c_12,t_1,c_12
789 bcs,a %xcc,.+8
790 add c_3,t_2,c_3
791 mulx a_2,b_4,t_1 !=!mul_add_c(a[2],b[4],c1,c2,c3);
792 addcc c_12,t_1,c_12
793 bcs,a %xcc,.+8
794 add c_3,t_2,c_3
795 lduw bp(6),b_6 !=
796 mulx a_1,b_5,t_1 !mul_add_c(a[1],b[5],c1,c2,c3);
797 addcc c_12,t_1,c_12
798 bcs,a %xcc,.+8
799 add c_3,t_2,c_3 !=
800 lduw bp(7),b_7
801 mulx a_0,b_6,t_1 !mul_add_c(a[0],b[6],c1,c2,c3);
802 addcc c_12,t_1,t_1
803 bcs,a %xcc,.+8 !=
804 add c_3,t_2,c_3
805 srlx t_1,32,c_12
806 stuw t_1,rp(6) !r[6]=c1;
807 or c_12,c_3,c_12 !=
808
809 mulx a_0,b_7,t_1 !mul_add_c(a[0],b[7],c2,c3,c1);
810 addcc c_12,t_1,c_12
811 clr c_3
812 bcs,a %xcc,.+8 !=
813 add c_3,t_2,c_3
814 mulx a_1,b_6,t_1 !mul_add_c(a[1],b[6],c2,c3,c1);
815 addcc c_12,t_1,c_12
816 bcs,a %xcc,.+8 !=
817 add c_3,t_2,c_3
818 mulx a_2,b_5,t_1 !mul_add_c(a[2],b[5],c2,c3,c1);
819 addcc c_12,t_1,c_12
820 bcs,a %xcc,.+8 !=
821 add c_3,t_2,c_3
822 mulx a_3,b_4,t_1 !mul_add_c(a[3],b[4],c2,c3,c1);
823 addcc c_12,t_1,c_12
824 bcs,a %xcc,.+8 !=
825 add c_3,t_2,c_3
826 mulx a_4,b_3,t_1 !mul_add_c(a[4],b[3],c2,c3,c1);
827 addcc c_12,t_1,c_12
828 bcs,a %xcc,.+8 !=
829 add c_3,t_2,c_3
830 mulx a_5,b_2,t_1 !mul_add_c(a[5],b[2],c2,c3,c1);
831 addcc c_12,t_1,c_12
832 bcs,a %xcc,.+8 !=
833 add c_3,t_2,c_3
834 lduw ap(7),a_7
835 mulx a_6,b_1,t_1 !=!mul_add_c(a[6],b[1],c2,c3,c1);
836 addcc c_12,t_1,c_12
837 bcs,a %xcc,.+8
838 add c_3,t_2,c_3
839 mulx a_7,b_0,t_1 !=!mul_add_c(a[7],b[0],c2,c3,c1);
840 addcc c_12,t_1,t_1
841 bcs,a %xcc,.+8
842 add c_3,t_2,c_3
843 srlx t_1,32,c_12 !=
844 stuw t_1,rp(7) !r[7]=c2;
845 or c_12,c_3,c_12
846
847 mulx a_7,b_1,t_1 !=!mul_add_c(a[7],b[1],c3,c1,c2);
848 addcc c_12,t_1,c_12
849 clr c_3
850 bcs,a %xcc,.+8
851 add c_3,t_2,c_3 !=
852 mulx a_6,b_2,t_1 !mul_add_c(a[6],b[2],c3,c1,c2);
853 addcc c_12,t_1,c_12
854 bcs,a %xcc,.+8
855 add c_3,t_2,c_3 !=
856 mulx a_5,b_3,t_1 !mul_add_c(a[5],b[3],c3,c1,c2);
857 addcc c_12,t_1,c_12
858 bcs,a %xcc,.+8
859 add c_3,t_2,c_3 !=
860 mulx a_4,b_4,t_1 !mul_add_c(a[4],b[4],c3,c1,c2);
861 addcc c_12,t_1,c_12
862 bcs,a %xcc,.+8
863 add c_3,t_2,c_3 !=
864 mulx a_3,b_5,t_1 !mul_add_c(a[3],b[5],c3,c1,c2);
865 addcc c_12,t_1,c_12
866 bcs,a %xcc,.+8
867 add c_3,t_2,c_3 !=
868 mulx a_2,b_6,t_1 !mul_add_c(a[2],b[6],c3,c1,c2);
869 addcc c_12,t_1,c_12
870 bcs,a %xcc,.+8
871 add c_3,t_2,c_3 !=
872 mulx a_1,b_7,t_1 !mul_add_c(a[1],b[7],c3,c1,c2);
873 addcc c_12,t_1,t_1
874 bcs,a %xcc,.+8
875 add c_3,t_2,c_3 !=
876 srlx t_1,32,c_12
877 stuw t_1,rp(8) !r[8]=c3;
878 or c_12,c_3,c_12
879
880 mulx a_2,b_7,t_1 !=!mul_add_c(a[2],b[7],c1,c2,c3);
881 addcc c_12,t_1,c_12
882 clr c_3
883 bcs,a %xcc,.+8
884 add c_3,t_2,c_3 !=
885 mulx a_3,b_6,t_1 !mul_add_c(a[3],b[6],c1,c2,c3);
886 addcc c_12,t_1,c_12
887 bcs,a %xcc,.+8 !=
888 add c_3,t_2,c_3
889 mulx a_4,b_5,t_1 !mul_add_c(a[4],b[5],c1,c2,c3);
890 addcc c_12,t_1,c_12
891 bcs,a %xcc,.+8 !=
892 add c_3,t_2,c_3
893 mulx a_5,b_4,t_1 !mul_add_c(a[5],b[4],c1,c2,c3);
894 addcc c_12,t_1,c_12
895 bcs,a %xcc,.+8 !=
896 add c_3,t_2,c_3
897 mulx a_6,b_3,t_1 !mul_add_c(a[6],b[3],c1,c2,c3);
898 addcc c_12,t_1,c_12
899 bcs,a %xcc,.+8 !=
900 add c_3,t_2,c_3
901 mulx a_7,b_2,t_1 !mul_add_c(a[7],b[2],c1,c2,c3);
902 addcc c_12,t_1,t_1
903 bcs,a %xcc,.+8 !=
904 add c_3,t_2,c_3
905 srlx t_1,32,c_12
906 stuw t_1,rp(9) !r[9]=c1;
907 or c_12,c_3,c_12 !=
908
909 mulx a_7,b_3,t_1 !mul_add_c(a[7],b[3],c2,c3,c1);
910 addcc c_12,t_1,c_12
911 clr c_3
912 bcs,a %xcc,.+8 !=
913 add c_3,t_2,c_3
914 mulx a_6,b_4,t_1 !mul_add_c(a[6],b[4],c2,c3,c1);
915 addcc c_12,t_1,c_12
916 bcs,a %xcc,.+8 !=
917 add c_3,t_2,c_3
918 mulx a_5,b_5,t_1 !mul_add_c(a[5],b[5],c2,c3,c1);
919 addcc c_12,t_1,c_12
920 bcs,a %xcc,.+8 !=
921 add c_3,t_2,c_3
922 mulx a_4,b_6,t_1 !mul_add_c(a[4],b[6],c2,c3,c1);
923 addcc c_12,t_1,c_12
924 bcs,a %xcc,.+8 !=
925 add c_3,t_2,c_3
926 mulx a_3,b_7,t_1 !mul_add_c(a[3],b[7],c2,c3,c1);
927 addcc c_12,t_1,t_1
928 bcs,a %xcc,.+8 !=
929 add c_3,t_2,c_3
930 srlx t_1,32,c_12
931 stuw t_1,rp(10) !r[10]=c2;
932 or c_12,c_3,c_12 !=
933
934 mulx a_4,b_7,t_1 !mul_add_c(a[4],b[7],c3,c1,c2);
935 addcc c_12,t_1,c_12
936 clr c_3
937 bcs,a %xcc,.+8 !=
938 add c_3,t_2,c_3
939 mulx a_5,b_6,t_1 !mul_add_c(a[5],b[6],c3,c1,c2);
940 addcc c_12,t_1,c_12
941 bcs,a %xcc,.+8 !=
942 add c_3,t_2,c_3
943 mulx a_6,b_5,t_1 !mul_add_c(a[6],b[5],c3,c1,c2);
944 addcc c_12,t_1,c_12
945 bcs,a %xcc,.+8 !=
946 add c_3,t_2,c_3
947 mulx a_7,b_4,t_1 !mul_add_c(a[7],b[4],c3,c1,c2);
948 addcc c_12,t_1,t_1
949 bcs,a %xcc,.+8 !=
950 add c_3,t_2,c_3
951 srlx t_1,32,c_12
952 stuw t_1,rp(11) !r[11]=c3;
953 or c_12,c_3,c_12 !=
954
955 mulx a_7,b_5,t_1 !mul_add_c(a[7],b[5],c1,c2,c3);
956 addcc c_12,t_1,c_12
957 clr c_3
958 bcs,a %xcc,.+8 !=
959 add c_3,t_2,c_3
960 mulx a_6,b_6,t_1 !mul_add_c(a[6],b[6],c1,c2,c3);
961 addcc c_12,t_1,c_12
962 bcs,a %xcc,.+8 !=
963 add c_3,t_2,c_3
964 mulx a_5,b_7,t_1 !mul_add_c(a[5],b[7],c1,c2,c3);
965 addcc c_12,t_1,t_1
966 bcs,a %xcc,.+8 !=
967 add c_3,t_2,c_3
968 srlx t_1,32,c_12
969 stuw t_1,rp(12) !r[12]=c1;
970 or c_12,c_3,c_12 !=
971
972 mulx a_6,b_7,t_1 !mul_add_c(a[6],b[7],c2,c3,c1);
973 addcc c_12,t_1,c_12
974 clr c_3
975 bcs,a %xcc,.+8 !=
976 add c_3,t_2,c_3
977 mulx a_7,b_6,t_1 !mul_add_c(a[7],b[6],c2,c3,c1);
978 addcc c_12,t_1,t_1
979 bcs,a %xcc,.+8 !=
980 add c_3,t_2,c_3
981 srlx t_1,32,c_12
982 st t_1,rp(13) !r[13]=c2;
983 or c_12,c_3,c_12 !=
984
985 mulx a_7,b_7,t_1 !mul_add_c(a[7],b[7],c3,c1,c2);
986 addcc c_12,t_1,t_1
987 srlx t_1,32,c_12 !=
988 stuw t_1,rp(14) !r[14]=c3;
989 stuw c_12,rp(15) !r[15]=c1;
990
991 ret
992 restore %g0,%g0,%o0 !=
993
994.type bn_mul_comba8,#function
995.size bn_mul_comba8,(.-bn_mul_comba8)
996
997.align 32
998
999.global bn_mul_comba4
1000/*
1001 * void bn_mul_comba4(r,a,b)
1002 * BN_ULONG *r,*a,*b;
1003 */
1004bn_mul_comba4:
1005 save %sp,FRAME_SIZE,%sp
1006 lduw ap(0),a_0
1007 mov 1,t_2
1008 lduw bp(0),b_0
1009 sllx t_2,32,t_2 !=
1010 lduw bp(1),b_1
1011 mulx a_0,b_0,t_1 !mul_add_c(a[0],b[0],c1,c2,c3);
1012 srlx t_1,32,c_12
1013 stuw t_1,rp(0) !=!r[0]=c1;
1014
1015 lduw ap(1),a_1
1016 mulx a_0,b_1,t_1 !mul_add_c(a[0],b[1],c2,c3,c1);
1017 addcc c_12,t_1,c_12
1018 clr c_3 !=
1019 bcs,a %xcc,.+8
1020 add c_3,t_2,c_3
1021 lduw ap(2),a_2
1022 mulx a_1,b_0,t_1 !=!mul_add_c(a[1],b[0],c2,c3,c1);
1023 addcc c_12,t_1,t_1
1024 bcs,a %xcc,.+8
1025 add c_3,t_2,c_3
1026 srlx t_1,32,c_12 !=
1027 stuw t_1,rp(1) !r[1]=c2;
1028 or c_12,c_3,c_12
1029
1030 mulx a_2,b_0,t_1 !mul_add_c(a[2],b[0],c3,c1,c2);
1031 addcc c_12,t_1,c_12 !=
1032 clr c_3
1033 bcs,a %xcc,.+8
1034 add c_3,t_2,c_3
1035 lduw bp(2),b_2 !=
1036 mulx a_1,b_1,t_1 !mul_add_c(a[1],b[1],c3,c1,c2);
1037 addcc c_12,t_1,c_12
1038 bcs,a %xcc,.+8
1039 add c_3,t_2,c_3 !=
1040 lduw bp(3),b_3
1041 mulx a_0,b_2,t_1 !mul_add_c(a[0],b[2],c3,c1,c2);
1042 addcc c_12,t_1,t_1
1043 bcs,a %xcc,.+8 !=
1044 add c_3,t_2,c_3
1045 srlx t_1,32,c_12
1046 stuw t_1,rp(2) !r[2]=c3;
1047 or c_12,c_3,c_12 !=
1048
1049 mulx a_0,b_3,t_1 !mul_add_c(a[0],b[3],c1,c2,c3);
1050 addcc c_12,t_1,c_12
1051 clr c_3
1052 bcs,a %xcc,.+8 !=
1053 add c_3,t_2,c_3
1054 mulx a_1,b_2,t_1 !mul_add_c(a[1],b[2],c1,c2,c3);
1055 addcc c_12,t_1,c_12
1056 bcs,a %xcc,.+8 !=
1057 add c_3,t_2,c_3
1058 lduw ap(3),a_3
1059 mulx a_2,b_1,t_1 !mul_add_c(a[2],b[1],c1,c2,c3);
1060 addcc c_12,t_1,c_12 !=
1061 bcs,a %xcc,.+8
1062 add c_3,t_2,c_3
1063 mulx a_3,b_0,t_1 !mul_add_c(a[3],b[0],c1,c2,c3);!=
1064 addcc c_12,t_1,t_1 !=
1065 bcs,a %xcc,.+8
1066 add c_3,t_2,c_3
1067 srlx t_1,32,c_12
1068 stuw t_1,rp(3) !=!r[3]=c1;
1069 or c_12,c_3,c_12
1070
1071 mulx a_3,b_1,t_1 !mul_add_c(a[3],b[1],c2,c3,c1);
1072 addcc c_12,t_1,c_12
1073 clr c_3 !=
1074 bcs,a %xcc,.+8
1075 add c_3,t_2,c_3
1076 mulx a_2,b_2,t_1 !mul_add_c(a[2],b[2],c2,c3,c1);
1077 addcc c_12,t_1,c_12 !=
1078 bcs,a %xcc,.+8
1079 add c_3,t_2,c_3
1080 mulx a_1,b_3,t_1 !mul_add_c(a[1],b[3],c2,c3,c1);
1081 addcc c_12,t_1,t_1 !=
1082 bcs,a %xcc,.+8
1083 add c_3,t_2,c_3
1084 srlx t_1,32,c_12
1085 stuw t_1,rp(4) !=!r[4]=c2;
1086 or c_12,c_3,c_12
1087
1088 mulx a_2,b_3,t_1 !mul_add_c(a[2],b[3],c3,c1,c2);
1089 addcc c_12,t_1,c_12
1090 clr c_3 !=
1091 bcs,a %xcc,.+8
1092 add c_3,t_2,c_3
1093 mulx a_3,b_2,t_1 !mul_add_c(a[3],b[2],c3,c1,c2);
1094 addcc c_12,t_1,t_1 !=
1095 bcs,a %xcc,.+8
1096 add c_3,t_2,c_3
1097 srlx t_1,32,c_12
1098 stuw t_1,rp(5) !=!r[5]=c3;
1099 or c_12,c_3,c_12
1100
1101 mulx a_3,b_3,t_1 !mul_add_c(a[3],b[3],c1,c2,c3);
1102 addcc c_12,t_1,t_1
1103 srlx t_1,32,c_12 !=
1104 stuw t_1,rp(6) !r[6]=c1;
1105 stuw c_12,rp(7) !r[7]=c2;
1106
1107 ret
1108 restore %g0,%g0,%o0
1109
1110.type bn_mul_comba4,#function
1111.size bn_mul_comba4,(.-bn_mul_comba4)
1112
1113.align 32
1114
1115.global bn_sqr_comba8
1116bn_sqr_comba8:
1117 save %sp,FRAME_SIZE,%sp
1118 mov 1,t_2
1119 lduw ap(0),a_0
1120 sllx t_2,32,t_2
1121 lduw ap(1),a_1
1122 mulx a_0,a_0,t_1 !sqr_add_c(a,0,c1,c2,c3);
1123 srlx t_1,32,c_12
1124 stuw t_1,rp(0) !r[0]=c1;
1125
1126 lduw ap(2),a_2
1127 mulx a_0,a_1,t_1 !=!sqr_add_c2(a,1,0,c2,c3,c1);
1128 addcc c_12,t_1,c_12
1129 clr c_3
1130 bcs,a %xcc,.+8
1131 add c_3,t_2,c_3
1132 addcc c_12,t_1,t_1
1133 bcs,a %xcc,.+8
1134 add c_3,t_2,c_3
1135 srlx t_1,32,c_12
1136 stuw t_1,rp(1) !r[1]=c2;
1137 or c_12,c_3,c_12
1138
1139 mulx a_2,a_0,t_1 !sqr_add_c2(a,2,0,c3,c1,c2);
1140 addcc c_12,t_1,c_12
1141 clr c_3
1142 bcs,a %xcc,.+8
1143 add c_3,t_2,c_3
1144 addcc c_12,t_1,c_12
1145 bcs,a %xcc,.+8
1146 add c_3,t_2,c_3
1147 lduw ap(3),a_3
1148 mulx a_1,a_1,t_1 !sqr_add_c(a,1,c3,c1,c2);
1149 addcc c_12,t_1,t_1
1150 bcs,a %xcc,.+8
1151 add c_3,t_2,c_3
1152 srlx t_1,32,c_12
1153 stuw t_1,rp(2) !r[2]=c3;
1154 or c_12,c_3,c_12
1155
1156 mulx a_0,a_3,t_1 !sqr_add_c2(a,3,0,c1,c2,c3);
1157 addcc c_12,t_1,c_12
1158 clr c_3
1159 bcs,a %xcc,.+8
1160 add c_3,t_2,c_3
1161 addcc c_12,t_1,c_12
1162 bcs,a %xcc,.+8
1163 add c_3,t_2,c_3
1164 lduw ap(4),a_4
1165 mulx a_1,a_2,t_1 !sqr_add_c2(a,2,1,c1,c2,c3);
1166 addcc c_12,t_1,c_12
1167 bcs,a %xcc,.+8
1168 add c_3,t_2,c_3
1169 addcc c_12,t_1,t_1
1170 bcs,a %xcc,.+8
1171 add c_3,t_2,c_3
1172 srlx t_1,32,c_12
1173 st t_1,rp(3) !r[3]=c1;
1174 or c_12,c_3,c_12
1175
1176 mulx a_4,a_0,t_1 !sqr_add_c2(a,4,0,c2,c3,c1);
1177 addcc c_12,t_1,c_12
1178 clr c_3
1179 bcs,a %xcc,.+8
1180 add c_3,t_2,c_3
1181 addcc c_12,t_1,c_12
1182 bcs,a %xcc,.+8
1183 add c_3,t_2,c_3
1184 mulx a_3,a_1,t_1 !sqr_add_c2(a,3,1,c2,c3,c1);
1185 addcc c_12,t_1,c_12
1186 bcs,a %xcc,.+8
1187 add c_3,t_2,c_3
1188 addcc c_12,t_1,c_12
1189 bcs,a %xcc,.+8
1190 add c_3,t_2,c_3
1191 lduw ap(5),a_5
1192 mulx a_2,a_2,t_1 !sqr_add_c(a,2,c2,c3,c1);
1193 addcc c_12,t_1,t_1
1194 bcs,a %xcc,.+8
1195 add c_3,t_2,c_3
1196 srlx t_1,32,c_12
1197 stuw t_1,rp(4) !r[4]=c2;
1198 or c_12,c_3,c_12
1199
1200 mulx a_0,a_5,t_1 !sqr_add_c2(a,5,0,c3,c1,c2);
1201 addcc c_12,t_1,c_12
1202 clr c_3
1203 bcs,a %xcc,.+8
1204 add c_3,t_2,c_3
1205 addcc c_12,t_1,c_12
1206 bcs,a %xcc,.+8
1207 add c_3,t_2,c_3
1208 mulx a_1,a_4,t_1 !sqr_add_c2(a,4,1,c3,c1,c2);
1209 addcc c_12,t_1,c_12
1210 bcs,a %xcc,.+8
1211 add c_3,t_2,c_3
1212 addcc c_12,t_1,c_12
1213 bcs,a %xcc,.+8
1214 add c_3,t_2,c_3
1215 lduw ap(6),a_6
1216 mulx a_2,a_3,t_1 !sqr_add_c2(a,3,2,c3,c1,c2);
1217 addcc c_12,t_1,c_12
1218 bcs,a %xcc,.+8
1219 add c_3,t_2,c_3
1220 addcc c_12,t_1,t_1
1221 bcs,a %xcc,.+8
1222 add c_3,t_2,c_3
1223 srlx t_1,32,c_12
1224 stuw t_1,rp(5) !r[5]=c3;
1225 or c_12,c_3,c_12
1226
1227 mulx a_6,a_0,t_1 !sqr_add_c2(a,6,0,c1,c2,c3);
1228 addcc c_12,t_1,c_12
1229 clr c_3
1230 bcs,a %xcc,.+8
1231 add c_3,t_2,c_3
1232 addcc c_12,t_1,c_12
1233 bcs,a %xcc,.+8
1234 add c_3,t_2,c_3
1235 mulx a_5,a_1,t_1 !sqr_add_c2(a,5,1,c1,c2,c3);
1236 addcc c_12,t_1,c_12
1237 bcs,a %xcc,.+8
1238 add c_3,t_2,c_3
1239 addcc c_12,t_1,c_12
1240 bcs,a %xcc,.+8
1241 add c_3,t_2,c_3
1242 mulx a_4,a_2,t_1 !sqr_add_c2(a,4,2,c1,c2,c3);
1243 addcc c_12,t_1,c_12
1244 bcs,a %xcc,.+8
1245 add c_3,t_2,c_3
1246 addcc c_12,t_1,c_12
1247 bcs,a %xcc,.+8
1248 add c_3,t_2,c_3
1249 lduw ap(7),a_7
1250 mulx a_3,a_3,t_1 !=!sqr_add_c(a,3,c1,c2,c3);
1251 addcc c_12,t_1,t_1
1252 bcs,a %xcc,.+8
1253 add c_3,t_2,c_3
1254 srlx t_1,32,c_12
1255 stuw t_1,rp(6) !r[6]=c1;
1256 or c_12,c_3,c_12
1257
1258 mulx a_0,a_7,t_1 !sqr_add_c2(a,7,0,c2,c3,c1);
1259 addcc c_12,t_1,c_12
1260 clr c_3
1261 bcs,a %xcc,.+8
1262 add c_3,t_2,c_3
1263 addcc c_12,t_1,c_12
1264 bcs,a %xcc,.+8
1265 add c_3,t_2,c_3
1266 mulx a_1,a_6,t_1 !sqr_add_c2(a,6,1,c2,c3,c1);
1267 addcc c_12,t_1,c_12
1268 bcs,a %xcc,.+8
1269 add c_3,t_2,c_3
1270 addcc c_12,t_1,c_12
1271 bcs,a %xcc,.+8
1272 add c_3,t_2,c_3
1273 mulx a_2,a_5,t_1 !sqr_add_c2(a,5,2,c2,c3,c1);
1274 addcc c_12,t_1,c_12
1275 bcs,a %xcc,.+8
1276 add c_3,t_2,c_3
1277 addcc c_12,t_1,c_12
1278 bcs,a %xcc,.+8
1279 add c_3,t_2,c_3
1280 mulx a_3,a_4,t_1 !sqr_add_c2(a,4,3,c2,c3,c1);
1281 addcc c_12,t_1,c_12
1282 bcs,a %xcc,.+8
1283 add c_3,t_2,c_3
1284 addcc c_12,t_1,t_1
1285 bcs,a %xcc,.+8
1286 add c_3,t_2,c_3
1287 srlx t_1,32,c_12
1288 stuw t_1,rp(7) !r[7]=c2;
1289 or c_12,c_3,c_12
1290
1291 mulx a_7,a_1,t_1 !sqr_add_c2(a,7,1,c3,c1,c2);
1292 addcc c_12,t_1,c_12
1293 clr c_3
1294 bcs,a %xcc,.+8
1295 add c_3,t_2,c_3
1296 addcc c_12,t_1,c_12
1297 bcs,a %xcc,.+8
1298 add c_3,t_2,c_3
1299 mulx a_6,a_2,t_1 !sqr_add_c2(a,6,2,c3,c1,c2);
1300 addcc c_12,t_1,c_12
1301 bcs,a %xcc,.+8
1302 add c_3,t_2,c_3
1303 addcc c_12,t_1,c_12
1304 bcs,a %xcc,.+8
1305 add c_3,t_2,c_3
1306 mulx a_5,a_3,t_1 !sqr_add_c2(a,5,3,c3,c1,c2);
1307 addcc c_12,t_1,c_12
1308 bcs,a %xcc,.+8
1309 add c_3,t_2,c_3
1310 addcc c_12,t_1,c_12
1311 bcs,a %xcc,.+8
1312 add c_3,t_2,c_3
1313 mulx a_4,a_4,t_1 !sqr_add_c(a,4,c3,c1,c2);
1314 addcc c_12,t_1,t_1
1315 bcs,a %xcc,.+8
1316 add c_3,t_2,c_3
1317 srlx t_1,32,c_12
1318 stuw t_1,rp(8) !r[8]=c3;
1319 or c_12,c_3,c_12
1320
1321 mulx a_2,a_7,t_1 !sqr_add_c2(a,7,2,c1,c2,c3);
1322 addcc c_12,t_1,c_12
1323 clr c_3
1324 bcs,a %xcc,.+8
1325 add c_3,t_2,c_3
1326 addcc c_12,t_1,c_12
1327 bcs,a %xcc,.+8
1328 add c_3,t_2,c_3
1329 mulx a_3,a_6,t_1 !sqr_add_c2(a,6,3,c1,c2,c3);
1330 addcc c_12,t_1,c_12
1331 bcs,a %xcc,.+8
1332 add c_3,t_2,c_3
1333 addcc c_12,t_1,c_12
1334 bcs,a %xcc,.+8
1335 add c_3,t_2,c_3
1336 mulx a_4,a_5,t_1 !sqr_add_c2(a,5,4,c1,c2,c3);
1337 addcc c_12,t_1,c_12
1338 bcs,a %xcc,.+8
1339 add c_3,t_2,c_3
1340 addcc c_12,t_1,t_1
1341 bcs,a %xcc,.+8
1342 add c_3,t_2,c_3
1343 srlx t_1,32,c_12
1344 stuw t_1,rp(9) !r[9]=c1;
1345 or c_12,c_3,c_12
1346
1347 mulx a_7,a_3,t_1 !sqr_add_c2(a,7,3,c2,c3,c1);
1348 addcc c_12,t_1,c_12
1349 clr c_3
1350 bcs,a %xcc,.+8
1351 add c_3,t_2,c_3
1352 addcc c_12,t_1,c_12
1353 bcs,a %xcc,.+8
1354 add c_3,t_2,c_3
1355 mulx a_6,a_4,t_1 !sqr_add_c2(a,6,4,c2,c3,c1);
1356 addcc c_12,t_1,c_12
1357 bcs,a %xcc,.+8
1358 add c_3,t_2,c_3
1359 addcc c_12,t_1,c_12
1360 bcs,a %xcc,.+8
1361 add c_3,t_2,c_3
1362 mulx a_5,a_5,t_1 !sqr_add_c(a,5,c2,c3,c1);
1363 addcc c_12,t_1,t_1
1364 bcs,a %xcc,.+8
1365 add c_3,t_2,c_3
1366 srlx t_1,32,c_12
1367 stuw t_1,rp(10) !r[10]=c2;
1368 or c_12,c_3,c_12
1369
1370 mulx a_4,a_7,t_1 !sqr_add_c2(a,7,4,c3,c1,c2);
1371 addcc c_12,t_1,c_12
1372 clr c_3
1373 bcs,a %xcc,.+8
1374 add c_3,t_2,c_3
1375 addcc c_12,t_1,c_12
1376 bcs,a %xcc,.+8
1377 add c_3,t_2,c_3
1378 mulx a_5,a_6,t_1 !sqr_add_c2(a,6,5,c3,c1,c2);
1379 addcc c_12,t_1,c_12
1380 bcs,a %xcc,.+8
1381 add c_3,t_2,c_3
1382 addcc c_12,t_1,t_1
1383 bcs,a %xcc,.+8
1384 add c_3,t_2,c_3
1385 srlx t_1,32,c_12
1386 stuw t_1,rp(11) !r[11]=c3;
1387 or c_12,c_3,c_12
1388
1389 mulx a_7,a_5,t_1 !sqr_add_c2(a,7,5,c1,c2,c3);
1390 addcc c_12,t_1,c_12
1391 clr c_3
1392 bcs,a %xcc,.+8
1393 add c_3,t_2,c_3
1394 addcc c_12,t_1,c_12
1395 bcs,a %xcc,.+8
1396 add c_3,t_2,c_3
1397 mulx a_6,a_6,t_1 !sqr_add_c(a,6,c1,c2,c3);
1398 addcc c_12,t_1,t_1
1399 bcs,a %xcc,.+8
1400 add c_3,t_2,c_3
1401 srlx t_1,32,c_12
1402 stuw t_1,rp(12) !r[12]=c1;
1403 or c_12,c_3,c_12
1404
1405 mulx a_6,a_7,t_1 !sqr_add_c2(a,7,6,c2,c3,c1);
1406 addcc c_12,t_1,c_12
1407 clr c_3
1408 bcs,a %xcc,.+8
1409 add c_3,t_2,c_3
1410 addcc c_12,t_1,t_1
1411 bcs,a %xcc,.+8
1412 add c_3,t_2,c_3
1413 srlx t_1,32,c_12
1414 stuw t_1,rp(13) !r[13]=c2;
1415 or c_12,c_3,c_12
1416
1417 mulx a_7,a_7,t_1 !sqr_add_c(a,7,c3,c1,c2);
1418 addcc c_12,t_1,t_1
1419 srlx t_1,32,c_12
1420 stuw t_1,rp(14) !r[14]=c3;
1421 stuw c_12,rp(15) !r[15]=c1;
1422
1423 ret
1424 restore %g0,%g0,%o0
1425
1426.type bn_sqr_comba8,#function
1427.size bn_sqr_comba8,(.-bn_sqr_comba8)
1428
1429.align 32
1430
1431.global bn_sqr_comba4
1432/*
1433 * void bn_sqr_comba4(r,a)
1434 * BN_ULONG *r,*a;
1435 */
1436bn_sqr_comba4:
1437 save %sp,FRAME_SIZE,%sp
1438 mov 1,t_2
1439 lduw ap(0),a_0
1440 sllx t_2,32,t_2
1441 lduw ap(1),a_1
1442 mulx a_0,a_0,t_1 !sqr_add_c(a,0,c1,c2,c3);
1443 srlx t_1,32,c_12
1444 stuw t_1,rp(0) !r[0]=c1;
1445
1446 lduw ap(2),a_2
1447 mulx a_0,a_1,t_1 !sqr_add_c2(a,1,0,c2,c3,c1);
1448 addcc c_12,t_1,c_12
1449 clr c_3
1450 bcs,a %xcc,.+8
1451 add c_3,t_2,c_3
1452 addcc c_12,t_1,t_1
1453 bcs,a %xcc,.+8
1454 add c_3,t_2,c_3
1455 srlx t_1,32,c_12
1456 stuw t_1,rp(1) !r[1]=c2;
1457 or c_12,c_3,c_12
1458
1459 mulx a_2,a_0,t_1 !sqr_add_c2(a,2,0,c3,c1,c2);
1460 addcc c_12,t_1,c_12
1461 clr c_3
1462 bcs,a %xcc,.+8
1463 add c_3,t_2,c_3
1464 addcc c_12,t_1,c_12
1465 bcs,a %xcc,.+8
1466 add c_3,t_2,c_3
1467 lduw ap(3),a_3
1468 mulx a_1,a_1,t_1 !sqr_add_c(a,1,c3,c1,c2);
1469 addcc c_12,t_1,t_1
1470 bcs,a %xcc,.+8
1471 add c_3,t_2,c_3
1472 srlx t_1,32,c_12
1473 stuw t_1,rp(2) !r[2]=c3;
1474 or c_12,c_3,c_12
1475
1476 mulx a_0,a_3,t_1 !sqr_add_c2(a,3,0,c1,c2,c3);
1477 addcc c_12,t_1,c_12
1478 clr c_3
1479 bcs,a %xcc,.+8
1480 add c_3,t_2,c_3
1481 addcc c_12,t_1,c_12
1482 bcs,a %xcc,.+8
1483 add c_3,t_2,c_3
1484 mulx a_1,a_2,t_1 !sqr_add_c2(a,2,1,c1,c2,c3);
1485 addcc c_12,t_1,c_12
1486 bcs,a %xcc,.+8
1487 add c_3,t_2,c_3
1488 addcc c_12,t_1,t_1
1489 bcs,a %xcc,.+8
1490 add c_3,t_2,c_3
1491 srlx t_1,32,c_12
1492 stuw t_1,rp(3) !r[3]=c1;
1493 or c_12,c_3,c_12
1494
1495 mulx a_3,a_1,t_1 !sqr_add_c2(a,3,1,c2,c3,c1);
1496 addcc c_12,t_1,c_12
1497 clr c_3
1498 bcs,a %xcc,.+8
1499 add c_3,t_2,c_3
1500 addcc c_12,t_1,c_12
1501 bcs,a %xcc,.+8
1502 add c_3,t_2,c_3
1503 mulx a_2,a_2,t_1 !sqr_add_c(a,2,c2,c3,c1);
1504 addcc c_12,t_1,t_1
1505 bcs,a %xcc,.+8
1506 add c_3,t_2,c_3
1507 srlx t_1,32,c_12
1508 stuw t_1,rp(4) !r[4]=c2;
1509 or c_12,c_3,c_12
1510
1511 mulx a_2,a_3,t_1 !sqr_add_c2(a,3,2,c3,c1,c2);
1512 addcc c_12,t_1,c_12
1513 clr c_3
1514 bcs,a %xcc,.+8
1515 add c_3,t_2,c_3
1516 addcc c_12,t_1,t_1
1517 bcs,a %xcc,.+8
1518 add c_3,t_2,c_3
1519 srlx t_1,32,c_12
1520 stuw t_1,rp(5) !r[5]=c3;
1521 or c_12,c_3,c_12
1522
1523 mulx a_3,a_3,t_1 !sqr_add_c(a,3,c1,c2,c3);
1524 addcc c_12,t_1,t_1
1525 srlx t_1,32,c_12
1526 stuw t_1,rp(6) !r[6]=c1;
1527 stuw c_12,rp(7) !r[7]=c2;
1528
1529 ret
1530 restore %g0,%g0,%o0
1531
1532.type bn_sqr_comba4,#function
1533.size bn_sqr_comba4,(.-bn_sqr_comba4)
1534
1535.align 32
diff --git a/src/lib/libcrypto/bn/asm/x86.pl b/src/lib/libcrypto/bn/asm/x86.pl
deleted file mode 100644
index 1bc4f1bb27..0000000000
--- a/src/lib/libcrypto/bn/asm/x86.pl
+++ /dev/null
@@ -1,28 +0,0 @@
1#!/usr/local/bin/perl
2
3push(@INC,"perlasm","../../perlasm");
4require "x86asm.pl";
5
6require("x86/mul_add.pl");
7require("x86/mul.pl");
8require("x86/sqr.pl");
9require("x86/div.pl");
10require("x86/add.pl");
11require("x86/sub.pl");
12require("x86/comba.pl");
13
14&asm_init($ARGV[0],$0);
15
16&bn_mul_add_words("bn_mul_add_words");
17&bn_mul_words("bn_mul_words");
18&bn_sqr_words("bn_sqr_words");
19&bn_div_words("bn_div_words");
20&bn_add_words("bn_add_words");
21&bn_sub_words("bn_sub_words");
22&bn_mul_comba("bn_mul_comba8",8);
23&bn_mul_comba("bn_mul_comba4",4);
24&bn_sqr_comba("bn_sqr_comba8",8);
25&bn_sqr_comba("bn_sqr_comba4",4);
26
27&asm_finish();
28
diff --git a/src/lib/libcrypto/bn/asm/x86/add.pl b/src/lib/libcrypto/bn/asm/x86/add.pl
deleted file mode 100644
index 0b5cf583e3..0000000000
--- a/src/lib/libcrypto/bn/asm/x86/add.pl
+++ /dev/null
@@ -1,76 +0,0 @@
1#!/usr/local/bin/perl
2# x86 assember
3
4sub bn_add_words
5 {
6 local($name)=@_;
7
8 &function_begin($name,"");
9
10 &comment("");
11 $a="esi";
12 $b="edi";
13 $c="eax";
14 $r="ebx";
15 $tmp1="ecx";
16 $tmp2="edx";
17 $num="ebp";
18
19 &mov($r,&wparam(0)); # get r
20 &mov($a,&wparam(1)); # get a
21 &mov($b,&wparam(2)); # get b
22 &mov($num,&wparam(3)); # get num
23 &xor($c,$c); # clear carry
24 &and($num,0xfffffff8); # num / 8
25
26 &jz(&label("aw_finish"));
27
28 &set_label("aw_loop",0);
29 for ($i=0; $i<8; $i++)
30 {
31 &comment("Round $i");
32
33 &mov($tmp1,&DWP($i*4,$a,"",0)); # *a
34 &mov($tmp2,&DWP($i*4,$b,"",0)); # *b
35 &add($tmp1,$c);
36 &mov($c,0);
37 &adc($c,$c);
38 &add($tmp1,$tmp2);
39 &adc($c,0);
40 &mov(&DWP($i*4,$r,"",0),$tmp1); # *r
41 }
42
43 &comment("");
44 &add($a,32);
45 &add($b,32);
46 &add($r,32);
47 &sub($num,8);
48 &jnz(&label("aw_loop"));
49
50 &set_label("aw_finish",0);
51 &mov($num,&wparam(3)); # get num
52 &and($num,7);
53 &jz(&label("aw_end"));
54
55 for ($i=0; $i<7; $i++)
56 {
57 &comment("Tail Round $i");
58 &mov($tmp1,&DWP($i*4,$a,"",0)); # *a
59 &mov($tmp2,&DWP($i*4,$b,"",0));# *b
60 &add($tmp1,$c);
61 &mov($c,0);
62 &adc($c,$c);
63 &add($tmp1,$tmp2);
64 &adc($c,0);
65 &dec($num) if ($i != 6);
66 &mov(&DWP($i*4,$r,"",0),$tmp1); # *a
67 &jz(&label("aw_end")) if ($i != 6);
68 }
69 &set_label("aw_end",0);
70
71# &mov("eax",$c); # $c is "eax"
72
73 &function_end($name);
74 }
75
761;
diff --git a/src/lib/libcrypto/bn/asm/x86/comba.pl b/src/lib/libcrypto/bn/asm/x86/comba.pl
deleted file mode 100644
index 2291253629..0000000000
--- a/src/lib/libcrypto/bn/asm/x86/comba.pl
+++ /dev/null
@@ -1,277 +0,0 @@
1#!/usr/local/bin/perl
2# x86 assember
3
4sub mul_add_c
5 {
6 local($a,$ai,$b,$bi,$c0,$c1,$c2,$pos,$i,$na,$nb)=@_;
7
8 # pos == -1 if eax and edx are pre-loaded, 0 to load from next
9 # words, and 1 if load return value
10
11 &comment("mul a[$ai]*b[$bi]");
12
13 # "eax" and "edx" will always be pre-loaded.
14 # &mov("eax",&DWP($ai*4,$a,"",0)) ;
15 # &mov("edx",&DWP($bi*4,$b,"",0));
16
17 &mul("edx");
18 &add($c0,"eax");
19 &mov("eax",&DWP(($na)*4,$a,"",0)) if $pos == 0; # laod next a
20 &mov("eax",&wparam(0)) if $pos > 0; # load r[]
21 ###
22 &adc($c1,"edx");
23 &mov("edx",&DWP(($nb)*4,$b,"",0)) if $pos == 0; # laod next b
24 &mov("edx",&DWP(($nb)*4,$b,"",0)) if $pos == 1; # laod next b
25 ###
26 &adc($c2,0);
27 # is pos > 1, it means it is the last loop
28 &mov(&DWP($i*4,"eax","",0),$c0) if $pos > 0; # save r[];
29 &mov("eax",&DWP(($na)*4,$a,"",0)) if $pos == 1; # laod next a
30 }
31
32sub sqr_add_c
33 {
34 local($r,$a,$ai,$bi,$c0,$c1,$c2,$pos,$i,$na,$nb)=@_;
35
36 # pos == -1 if eax and edx are pre-loaded, 0 to load from next
37 # words, and 1 if load return value
38
39 &comment("sqr a[$ai]*a[$bi]");
40
41 # "eax" and "edx" will always be pre-loaded.
42 # &mov("eax",&DWP($ai*4,$a,"",0)) ;
43 # &mov("edx",&DWP($bi*4,$b,"",0));
44
45 if ($ai == $bi)
46 { &mul("eax");}
47 else
48 { &mul("edx");}
49 &add($c0,"eax");
50 &mov("eax",&DWP(($na)*4,$a,"",0)) if $pos == 0; # load next a
51 ###
52 &adc($c1,"edx");
53 &mov("edx",&DWP(($nb)*4,$a,"",0)) if ($pos == 1) && ($na != $nb);
54 ###
55 &adc($c2,0);
56 # is pos > 1, it means it is the last loop
57 &mov(&DWP($i*4,$r,"",0),$c0) if $pos > 0; # save r[];
58 &mov("eax",&DWP(($na)*4,$a,"",0)) if $pos == 1; # load next b
59 }
60
61sub sqr_add_c2
62 {
63 local($r,$a,$ai,$bi,$c0,$c1,$c2,$pos,$i,$na,$nb)=@_;
64
65 # pos == -1 if eax and edx are pre-loaded, 0 to load from next
66 # words, and 1 if load return value
67
68 &comment("sqr a[$ai]*a[$bi]");
69
70 # "eax" and "edx" will always be pre-loaded.
71 # &mov("eax",&DWP($ai*4,$a,"",0)) ;
72 # &mov("edx",&DWP($bi*4,$a,"",0));
73
74 if ($ai == $bi)
75 { &mul("eax");}
76 else
77 { &mul("edx");}
78 &add("eax","eax");
79 ###
80 &adc("edx","edx");
81 ###
82 &adc($c2,0);
83 &add($c0,"eax");
84 &adc($c1,"edx");
85 &mov("eax",&DWP(($na)*4,$a,"",0)) if $pos == 0; # load next a
86 &mov("eax",&DWP(($na)*4,$a,"",0)) if $pos == 1; # load next b
87 &adc($c2,0);
88 &mov(&DWP($i*4,$r,"",0),$c0) if $pos > 0; # save r[];
89 &mov("edx",&DWP(($nb)*4,$a,"",0)) if ($pos <= 1) && ($na != $nb);
90 ###
91 }
92
93sub bn_mul_comba
94 {
95 local($name,$num)=@_;
96 local($a,$b,$c0,$c1,$c2);
97 local($i,$as,$ae,$bs,$be,$ai,$bi);
98 local($tot,$end);
99
100 &function_begin_B($name,"");
101
102 $c0="ebx";
103 $c1="ecx";
104 $c2="ebp";
105 $a="esi";
106 $b="edi";
107
108 $as=0;
109 $ae=0;
110 $bs=0;
111 $be=0;
112 $tot=$num+$num-1;
113
114 &push("esi");
115 &mov($a,&wparam(1));
116 &push("edi");
117 &mov($b,&wparam(2));
118 &push("ebp");
119 &push("ebx");
120
121 &xor($c0,$c0);
122 &mov("eax",&DWP(0,$a,"",0)); # load the first word
123 &xor($c1,$c1);
124 &mov("edx",&DWP(0,$b,"",0)); # load the first second
125
126 for ($i=0; $i<$tot; $i++)
127 {
128 $ai=$as;
129 $bi=$bs;
130 $end=$be+1;
131
132 &comment("################## Calculate word $i");
133
134 for ($j=$bs; $j<$end; $j++)
135 {
136 &xor($c2,$c2) if ($j == $bs);
137 if (($j+1) == $end)
138 {
139 $v=1;
140 $v=2 if (($i+1) == $tot);
141 }
142 else
143 { $v=0; }
144 if (($j+1) != $end)
145 {
146 $na=($ai-1);
147 $nb=($bi+1);
148 }
149 else
150 {
151 $na=$as+($i < ($num-1));
152 $nb=$bs+($i >= ($num-1));
153 }
154#printf STDERR "[$ai,$bi] -> [$na,$nb]\n";
155 &mul_add_c($a,$ai,$b,$bi,$c0,$c1,$c2,$v,$i,$na,$nb);
156 if ($v)
157 {
158 &comment("saved r[$i]");
159 # &mov("eax",&wparam(0));
160 # &mov(&DWP($i*4,"eax","",0),$c0);
161 ($c0,$c1,$c2)=($c1,$c2,$c0);
162 }
163 $ai--;
164 $bi++;
165 }
166 $as++ if ($i < ($num-1));
167 $ae++ if ($i >= ($num-1));
168
169 $bs++ if ($i >= ($num-1));
170 $be++ if ($i < ($num-1));
171 }
172 &comment("save r[$i]");
173 # &mov("eax",&wparam(0));
174 &mov(&DWP($i*4,"eax","",0),$c0);
175
176 &pop("ebx");
177 &pop("ebp");
178 &pop("edi");
179 &pop("esi");
180 &ret();
181 &function_end_B($name);
182 }
183
184sub bn_sqr_comba
185 {
186 local($name,$num)=@_;
187 local($r,$a,$c0,$c1,$c2)=@_;
188 local($i,$as,$ae,$bs,$be,$ai,$bi);
189 local($b,$tot,$end,$half);
190
191 &function_begin_B($name,"");
192
193 $c0="ebx";
194 $c1="ecx";
195 $c2="ebp";
196 $a="esi";
197 $r="edi";
198
199 &push("esi");
200 &push("edi");
201 &push("ebp");
202 &push("ebx");
203 &mov($r,&wparam(0));
204 &mov($a,&wparam(1));
205 &xor($c0,$c0);
206 &xor($c1,$c1);
207 &mov("eax",&DWP(0,$a,"",0)); # load the first word
208
209 $as=0;
210 $ae=0;
211 $bs=0;
212 $be=0;
213 $tot=$num+$num-1;
214
215 for ($i=0; $i<$tot; $i++)
216 {
217 $ai=$as;
218 $bi=$bs;
219 $end=$be+1;
220
221 &comment("############### Calculate word $i");
222 for ($j=$bs; $j<$end; $j++)
223 {
224 &xor($c2,$c2) if ($j == $bs);
225 if (($ai-1) < ($bi+1))
226 {
227 $v=1;
228 $v=2 if ($i+1) == $tot;
229 }
230 else
231 { $v=0; }
232 if (!$v)
233 {
234 $na=$ai-1;
235 $nb=$bi+1;
236 }
237 else
238 {
239 $na=$as+($i < ($num-1));
240 $nb=$bs+($i >= ($num-1));
241 }
242 if ($ai == $bi)
243 {
244 &sqr_add_c($r,$a,$ai,$bi,
245 $c0,$c1,$c2,$v,$i,$na,$nb);
246 }
247 else
248 {
249 &sqr_add_c2($r,$a,$ai,$bi,
250 $c0,$c1,$c2,$v,$i,$na,$nb);
251 }
252 if ($v)
253 {
254 &comment("saved r[$i]");
255 #&mov(&DWP($i*4,$r,"",0),$c0);
256 ($c0,$c1,$c2)=($c1,$c2,$c0);
257 last;
258 }
259 $ai--;
260 $bi++;
261 }
262 $as++ if ($i < ($num-1));
263 $ae++ if ($i >= ($num-1));
264
265 $bs++ if ($i >= ($num-1));
266 $be++ if ($i < ($num-1));
267 }
268 &mov(&DWP($i*4,$r,"",0),$c0);
269 &pop("ebx");
270 &pop("ebp");
271 &pop("edi");
272 &pop("esi");
273 &ret();
274 &function_end_B($name);
275 }
276
2771;
diff --git a/src/lib/libcrypto/bn/asm/x86/div.pl b/src/lib/libcrypto/bn/asm/x86/div.pl
deleted file mode 100644
index 0e90152caa..0000000000
--- a/src/lib/libcrypto/bn/asm/x86/div.pl
+++ /dev/null
@@ -1,15 +0,0 @@
1#!/usr/local/bin/perl
2# x86 assember
3
4sub bn_div_words
5 {
6 local($name)=@_;
7
8 &function_begin($name,"");
9 &mov("edx",&wparam(0)); #
10 &mov("eax",&wparam(1)); #
11 &mov("ebx",&wparam(2)); #
12 &div("ebx");
13 &function_end($name);
14 }
151;
diff --git a/src/lib/libcrypto/bn/asm/x86/mul.pl b/src/lib/libcrypto/bn/asm/x86/mul.pl
deleted file mode 100644
index 674cb9b055..0000000000
--- a/src/lib/libcrypto/bn/asm/x86/mul.pl
+++ /dev/null
@@ -1,77 +0,0 @@
1#!/usr/local/bin/perl
2# x86 assember
3
4sub bn_mul_words
5 {
6 local($name)=@_;
7
8 &function_begin($name,"");
9
10 &comment("");
11 $Low="eax";
12 $High="edx";
13 $a="ebx";
14 $w="ecx";
15 $r="edi";
16 $c="esi";
17 $num="ebp";
18
19 &xor($c,$c); # clear carry
20 &mov($r,&wparam(0)); #
21 &mov($a,&wparam(1)); #
22 &mov($num,&wparam(2)); #
23 &mov($w,&wparam(3)); #
24
25 &and($num,0xfffffff8); # num / 8
26 &jz(&label("mw_finish"));
27
28 &set_label("mw_loop",0);
29 for ($i=0; $i<32; $i+=4)
30 {
31 &comment("Round $i");
32
33 &mov("eax",&DWP($i,$a,"",0)); # *a
34 &mul($w); # *a * w
35 &add("eax",$c); # L(t)+=c
36 # XXX
37
38 &adc("edx",0); # H(t)+=carry
39 &mov(&DWP($i,$r,"",0),"eax"); # *r= L(t);
40
41 &mov($c,"edx"); # c= H(t);
42 }
43
44 &comment("");
45 &add($a,32);
46 &add($r,32);
47 &sub($num,8);
48 &jz(&label("mw_finish"));
49 &jmp(&label("mw_loop"));
50
51 &set_label("mw_finish",0);
52 &mov($num,&wparam(2)); # get num
53 &and($num,7);
54 &jnz(&label("mw_finish2"));
55 &jmp(&label("mw_end"));
56
57 &set_label("mw_finish2",1);
58 for ($i=0; $i<7; $i++)
59 {
60 &comment("Tail Round $i");
61 &mov("eax",&DWP($i*4,$a,"",0));# *a
62 &mul($w); # *a * w
63 &add("eax",$c); # L(t)+=c
64 # XXX
65 &adc("edx",0); # H(t)+=carry
66 &mov(&DWP($i*4,$r,"",0),"eax");# *r= L(t);
67 &mov($c,"edx"); # c= H(t);
68 &dec($num) if ($i != 7-1);
69 &jz(&label("mw_end")) if ($i != 7-1);
70 }
71 &set_label("mw_end",0);
72 &mov("eax",$c);
73
74 &function_end($name);
75 }
76
771;
diff --git a/src/lib/libcrypto/bn/asm/x86/mul_add.pl b/src/lib/libcrypto/bn/asm/x86/mul_add.pl
deleted file mode 100644
index 61830d3a90..0000000000
--- a/src/lib/libcrypto/bn/asm/x86/mul_add.pl
+++ /dev/null
@@ -1,87 +0,0 @@
1#!/usr/local/bin/perl
2# x86 assember
3
4sub bn_mul_add_words
5 {
6 local($name)=@_;
7
8 &function_begin($name,"");
9
10 &comment("");
11 $Low="eax";
12 $High="edx";
13 $a="ebx";
14 $w="ebp";
15 $r="edi";
16 $c="esi";
17
18 &xor($c,$c); # clear carry
19 &mov($r,&wparam(0)); #
20
21 &mov("ecx",&wparam(2)); #
22 &mov($a,&wparam(1)); #
23
24 &and("ecx",0xfffffff8); # num / 8
25 &mov($w,&wparam(3)); #
26
27 &push("ecx"); # Up the stack for a tmp variable
28
29 &jz(&label("maw_finish"));
30
31 &set_label("maw_loop",0);
32
33 &mov(&swtmp(0),"ecx"); #
34
35 for ($i=0; $i<32; $i+=4)
36 {
37 &comment("Round $i");
38
39 &mov("eax",&DWP($i,$a,"",0)); # *a
40 &mul($w); # *a * w
41 &add("eax",$c); # L(t)+= *r
42 &mov($c,&DWP($i,$r,"",0)); # L(t)+= *r
43 &adc("edx",0); # H(t)+=carry
44 &add("eax",$c); # L(t)+=c
45 &adc("edx",0); # H(t)+=carry
46 &mov(&DWP($i,$r,"",0),"eax"); # *r= L(t);
47 &mov($c,"edx"); # c= H(t);
48 }
49
50 &comment("");
51 &mov("ecx",&swtmp(0)); #
52 &add($a,32);
53 &add($r,32);
54 &sub("ecx",8);
55 &jnz(&label("maw_loop"));
56
57 &set_label("maw_finish",0);
58 &mov("ecx",&wparam(2)); # get num
59 &and("ecx",7);
60 &jnz(&label("maw_finish2")); # helps branch prediction
61 &jmp(&label("maw_end"));
62
63 &set_label("maw_finish2",1);
64 for ($i=0; $i<7; $i++)
65 {
66 &comment("Tail Round $i");
67 &mov("eax",&DWP($i*4,$a,"",0));# *a
68 &mul($w); # *a * w
69 &add("eax",$c); # L(t)+=c
70 &mov($c,&DWP($i*4,$r,"",0)); # L(t)+= *r
71 &adc("edx",0); # H(t)+=carry
72 &add("eax",$c);
73 &adc("edx",0); # H(t)+=carry
74 &dec("ecx") if ($i != 7-1);
75 &mov(&DWP($i*4,$r,"",0),"eax"); # *r= L(t);
76 &mov($c,"edx"); # c= H(t);
77 &jz(&label("maw_end")) if ($i != 7-1);
78 }
79 &set_label("maw_end",0);
80 &mov("eax",$c);
81
82 &pop("ecx"); # clear variable from
83
84 &function_end($name);
85 }
86
871;
diff --git a/src/lib/libcrypto/bn/asm/x86/sqr.pl b/src/lib/libcrypto/bn/asm/x86/sqr.pl
deleted file mode 100644
index 1f90993cf6..0000000000
--- a/src/lib/libcrypto/bn/asm/x86/sqr.pl
+++ /dev/null
@@ -1,60 +0,0 @@
1#!/usr/local/bin/perl
2# x86 assember
3
4sub bn_sqr_words
5 {
6 local($name)=@_;
7
8 &function_begin($name,"");
9
10 &comment("");
11 $r="esi";
12 $a="edi";
13 $num="ebx";
14
15 &mov($r,&wparam(0)); #
16 &mov($a,&wparam(1)); #
17 &mov($num,&wparam(2)); #
18
19 &and($num,0xfffffff8); # num / 8
20 &jz(&label("sw_finish"));
21
22 &set_label("sw_loop",0);
23 for ($i=0; $i<32; $i+=4)
24 {
25 &comment("Round $i");
26 &mov("eax",&DWP($i,$a,"",0)); # *a
27 # XXX
28 &mul("eax"); # *a * *a
29 &mov(&DWP($i*2,$r,"",0),"eax"); #
30 &mov(&DWP($i*2+4,$r,"",0),"edx");#
31 }
32
33 &comment("");
34 &add($a,32);
35 &add($r,64);
36 &sub($num,8);
37 &jnz(&label("sw_loop"));
38
39 &set_label("sw_finish",0);
40 &mov($num,&wparam(2)); # get num
41 &and($num,7);
42 &jz(&label("sw_end"));
43
44 for ($i=0; $i<7; $i++)
45 {
46 &comment("Tail Round $i");
47 &mov("eax",&DWP($i*4,$a,"",0)); # *a
48 # XXX
49 &mul("eax"); # *a * *a
50 &mov(&DWP($i*8,$r,"",0),"eax"); #
51 &dec($num) if ($i != 7-1);
52 &mov(&DWP($i*8+4,$r,"",0),"edx");
53 &jz(&label("sw_end")) if ($i != 7-1);
54 }
55 &set_label("sw_end",0);
56
57 &function_end($name);
58 }
59
601;
diff --git a/src/lib/libcrypto/bn/asm/x86/sub.pl b/src/lib/libcrypto/bn/asm/x86/sub.pl
deleted file mode 100644
index 837b0e1b07..0000000000
--- a/src/lib/libcrypto/bn/asm/x86/sub.pl
+++ /dev/null
@@ -1,76 +0,0 @@
1#!/usr/local/bin/perl
2# x86 assember
3
4sub bn_sub_words
5 {
6 local($name)=@_;
7
8 &function_begin($name,"");
9
10 &comment("");
11 $a="esi";
12 $b="edi";
13 $c="eax";
14 $r="ebx";
15 $tmp1="ecx";
16 $tmp2="edx";
17 $num="ebp";
18
19 &mov($r,&wparam(0)); # get r
20 &mov($a,&wparam(1)); # get a
21 &mov($b,&wparam(2)); # get b
22 &mov($num,&wparam(3)); # get num
23 &xor($c,$c); # clear carry
24 &and($num,0xfffffff8); # num / 8
25
26 &jz(&label("aw_finish"));
27
28 &set_label("aw_loop",0);
29 for ($i=0; $i<8; $i++)
30 {
31 &comment("Round $i");
32
33 &mov($tmp1,&DWP($i*4,$a,"",0)); # *a
34 &mov($tmp2,&DWP($i*4,$b,"",0)); # *b
35 &sub($tmp1,$c);
36 &mov($c,0);
37 &adc($c,$c);
38 &sub($tmp1,$tmp2);
39 &adc($c,0);
40 &mov(&DWP($i*4,$r,"",0),$tmp1); # *r
41 }
42
43 &comment("");
44 &add($a,32);
45 &add($b,32);
46 &add($r,32);
47 &sub($num,8);
48 &jnz(&label("aw_loop"));
49
50 &set_label("aw_finish",0);
51 &mov($num,&wparam(3)); # get num
52 &and($num,7);
53 &jz(&label("aw_end"));
54
55 for ($i=0; $i<7; $i++)
56 {
57 &comment("Tail Round $i");
58 &mov($tmp1,&DWP($i*4,$a,"",0)); # *a
59 &mov($tmp2,&DWP($i*4,$b,"",0));# *b
60 &sub($tmp1,$c);
61 &mov($c,0);
62 &adc($c,$c);
63 &sub($tmp1,$tmp2);
64 &adc($c,0);
65 &dec($num) if ($i != 6);
66 &mov(&DWP($i*4,$r,"",0),$tmp1); # *a
67 &jz(&label("aw_end")) if ($i != 6);
68 }
69 &set_label("aw_end",0);
70
71# &mov("eax",$c); # $c is "eax"
72
73 &function_end($name);
74 }
75
761;
diff --git a/src/lib/libcrypto/bn/asm/x86_64-gcc.c b/src/lib/libcrypto/bn/asm/x86_64-gcc.c
deleted file mode 100644
index b97b394661..0000000000
--- a/src/lib/libcrypto/bn/asm/x86_64-gcc.c
+++ /dev/null
@@ -1,575 +0,0 @@
1/*
2 * x86_64 BIGNUM accelerator version 0.1, December 2002.
3 *
4 * Implemented by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
5 * project.
6 *
7 * Rights for redistribution and usage in source and binary forms are
8 * granted according to the OpenSSL license. Warranty of any kind is
9 * disclaimed.
10 *
11 * Q. Version 0.1? It doesn't sound like Andy, he used to assign real
12 * versions, like 1.0...
13 * A. Well, that's because this code is basically a quick-n-dirty
14 * proof-of-concept hack. As you can see it's implemented with
15 * inline assembler, which means that you're bound to GCC and that
16 * there must be a room for fine-tuning.
17 *
18 * Q. Why inline assembler?
19 * A. x86_64 features own ABI I'm not familiar with. Which is why
20 * I decided to let the compiler take care of subroutine
21 * prologue/epilogue as well as register allocation.
22 *
23 * Q. How much faster does it get?
24 * A. Unfortunately people sitting on x86_64 hardware are prohibited
25 * to disclose the performance numbers, so they (SuSE labs to be
26 * specific) wouldn't tell me. However! Very similar coding technique
27 * (reaching out for 128-bit result from 64x64-bit multiplication)
28 * results in >3 times performance improvement on MIPS and I see no
29 * reason why gain on x86_64 would be so much different:-)
30 */
31
32#define BN_ULONG unsigned long
33
34/*
35 * "m"(a), "+m"(r) is the way to favor DirectPath µ-code;
36 * "g"(0) let the compiler to decide where does it
37 * want to keep the value of zero;
38 */
39#define mul_add(r,a,word,carry) do { \
40 register BN_ULONG high,low; \
41 asm ("mulq %3" \
42 : "=a"(low),"=d"(high) \
43 : "a"(word),"m"(a) \
44 : "cc"); \
45 asm ("addq %2,%0; adcq %3,%1" \
46 : "+r"(carry),"+d"(high)\
47 : "a"(low),"g"(0) \
48 : "cc"); \
49 asm ("addq %2,%0; adcq %3,%1" \
50 : "+m"(r),"+d"(high) \
51 : "r"(carry),"g"(0) \
52 : "cc"); \
53 carry=high; \
54 } while (0)
55
56#define mul(r,a,word,carry) do { \
57 register BN_ULONG high,low; \
58 asm ("mulq %3" \
59 : "=a"(low),"=d"(high) \
60 : "a"(word),"g"(a) \
61 : "cc"); \
62 asm ("addq %2,%0; adcq %3,%1" \
63 : "+r"(carry),"+d"(high)\
64 : "a"(low),"g"(0) \
65 : "cc"); \
66 (r)=carry, carry=high; \
67 } while (0)
68
69#define sqr(r0,r1,a) \
70 asm ("mulq %2" \
71 : "=a"(r0),"=d"(r1) \
72 : "a"(a) \
73 : "cc");
74
75BN_ULONG bn_mul_add_words(BN_ULONG *rp, BN_ULONG *ap, int num, BN_ULONG w)
76 {
77 BN_ULONG c1=0;
78
79 if (num <= 0) return(c1);
80
81 while (num&~3)
82 {
83 mul_add(rp[0],ap[0],w,c1);
84 mul_add(rp[1],ap[1],w,c1);
85 mul_add(rp[2],ap[2],w,c1);
86 mul_add(rp[3],ap[3],w,c1);
87 ap+=4; rp+=4; num-=4;
88 }
89 if (num)
90 {
91 mul_add(rp[0],ap[0],w,c1); if (--num==0) return c1;
92 mul_add(rp[1],ap[1],w,c1); if (--num==0) return c1;
93 mul_add(rp[2],ap[2],w,c1); return c1;
94 }
95
96 return(c1);
97 }
98
99BN_ULONG bn_mul_words(BN_ULONG *rp, BN_ULONG *ap, int num, BN_ULONG w)
100 {
101 BN_ULONG c1=0;
102
103 if (num <= 0) return(c1);
104
105 while (num&~3)
106 {
107 mul(rp[0],ap[0],w,c1);
108 mul(rp[1],ap[1],w,c1);
109 mul(rp[2],ap[2],w,c1);
110 mul(rp[3],ap[3],w,c1);
111 ap+=4; rp+=4; num-=4;
112 }
113 if (num)
114 {
115 mul(rp[0],ap[0],w,c1); if (--num == 0) return c1;
116 mul(rp[1],ap[1],w,c1); if (--num == 0) return c1;
117 mul(rp[2],ap[2],w,c1);
118 }
119 return(c1);
120 }
121
122void bn_sqr_words(BN_ULONG *r, BN_ULONG *a, int n)
123 {
124 if (n <= 0) return;
125
126 while (n&~3)
127 {
128 sqr(r[0],r[1],a[0]);
129 sqr(r[2],r[3],a[1]);
130 sqr(r[4],r[5],a[2]);
131 sqr(r[6],r[7],a[3]);
132 a+=4; r+=8; n-=4;
133 }
134 if (n)
135 {
136 sqr(r[0],r[1],a[0]); if (--n == 0) return;
137 sqr(r[2],r[3],a[1]); if (--n == 0) return;
138 sqr(r[4],r[5],a[2]);
139 }
140 }
141
142BN_ULONG bn_div_words(BN_ULONG h, BN_ULONG l, BN_ULONG d)
143{ BN_ULONG ret,waste;
144
145 asm ("divq %3"
146 : "=a"(ret),"=d"(waste)
147 : "a"(l),"d"(h),"g"(d)
148 : "cc");
149
150 return ret;
151}
152
153BN_ULONG bn_add_words (BN_ULONG *rp, BN_ULONG *ap, BN_ULONG *bp,int n)
154{ BN_ULONG ret,i;
155
156 if (n <= 0) return 0;
157
158 asm (
159 " subq %2,%2 \n"
160 ".align 16 \n"
161 "1: movq (%4,%2,8),%0 \n"
162 " adcq (%5,%2,8),%0 \n"
163 " movq %0,(%3,%2,8) \n"
164 " leaq 1(%2),%2 \n"
165 " loop 1b \n"
166 " sbbq %0,%0 \n"
167 : "+a"(ret),"+c"(n),"+r"(i)
168 : "r"(rp),"r"(ap),"r"(bp)
169 : "cc"
170 );
171
172 return ret&1;
173}
174
175#ifndef SIMICS
176BN_ULONG bn_sub_words (BN_ULONG *rp, BN_ULONG *ap, BN_ULONG *bp,int n)
177{ BN_ULONG ret,i;
178
179 if (n <= 0) return 0;
180
181 asm (
182 " subq %2,%2 \n"
183 ".align 16 \n"
184 "1: movq (%4,%2,8),%0 \n"
185 " sbbq (%5,%2,8),%0 \n"
186 " movq %0,(%3,%2,8) \n"
187 " leaq 1(%2),%2 \n"
188 " loop 1b \n"
189 " sbbq %0,%0 \n"
190 : "+a"(ret),"+c"(n),"+r"(i)
191 : "r"(rp),"r"(ap),"r"(bp)
192 : "cc"
193 );
194
195 return ret&1;
196}
197#else
198/* Simics 1.4<7 has buggy sbbq:-( */
199#define BN_MASK2 0xffffffffffffffffL
200BN_ULONG bn_sub_words(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n)
201 {
202 BN_ULONG t1,t2;
203 int c=0;
204
205 if (n <= 0) return((BN_ULONG)0);
206
207 for (;;)
208 {
209 t1=a[0]; t2=b[0];
210 r[0]=(t1-t2-c)&BN_MASK2;
211 if (t1 != t2) c=(t1 < t2);
212 if (--n <= 0) break;
213
214 t1=a[1]; t2=b[1];
215 r[1]=(t1-t2-c)&BN_MASK2;
216 if (t1 != t2) c=(t1 < t2);
217 if (--n <= 0) break;
218
219 t1=a[2]; t2=b[2];
220 r[2]=(t1-t2-c)&BN_MASK2;
221 if (t1 != t2) c=(t1 < t2);
222 if (--n <= 0) break;
223
224 t1=a[3]; t2=b[3];
225 r[3]=(t1-t2-c)&BN_MASK2;
226 if (t1 != t2) c=(t1 < t2);
227 if (--n <= 0) break;
228
229 a+=4;
230 b+=4;
231 r+=4;
232 }
233 return(c);
234 }
235#endif
236
237/* mul_add_c(a,b,c0,c1,c2) -- c+=a*b for three word number c=(c2,c1,c0) */
238/* mul_add_c2(a,b,c0,c1,c2) -- c+=2*a*b for three word number c=(c2,c1,c0) */
239/* sqr_add_c(a,i,c0,c1,c2) -- c+=a[i]^2 for three word number c=(c2,c1,c0) */
240/* sqr_add_c2(a,i,c0,c1,c2) -- c+=2*a[i]*a[j] for three word number c=(c2,c1,c0) */
241
242#if 0
243/* original macros are kept for reference purposes */
244#define mul_add_c(a,b,c0,c1,c2) { \
245 BN_ULONG ta=(a),tb=(b); \
246 t1 = ta * tb; \
247 t2 = BN_UMULT_HIGH(ta,tb); \
248 c0 += t1; t2 += (c0<t1)?1:0; \
249 c1 += t2; c2 += (c1<t2)?1:0; \
250 }
251
252#define mul_add_c2(a,b,c0,c1,c2) { \
253 BN_ULONG ta=(a),tb=(b),t0; \
254 t1 = BN_UMULT_HIGH(ta,tb); \
255 t0 = ta * tb; \
256 t2 = t1+t1; c2 += (t2<t1)?1:0; \
257 t1 = t0+t0; t2 += (t1<t0)?1:0; \
258 c0 += t1; t2 += (c0<t1)?1:0; \
259 c1 += t2; c2 += (c1<t2)?1:0; \
260 }
261#else
262#define mul_add_c(a,b,c0,c1,c2) do { \
263 asm ("mulq %3" \
264 : "=a"(t1),"=d"(t2) \
265 : "a"(a),"m"(b) \
266 : "cc"); \
267 asm ("addq %2,%0; adcq %3,%1" \
268 : "+r"(c0),"+d"(t2) \
269 : "a"(t1),"g"(0) \
270 : "cc"); \
271 asm ("addq %2,%0; adcq %3,%1" \
272 : "+r"(c1),"+r"(c2) \
273 : "d"(t2),"g"(0) \
274 : "cc"); \
275 } while (0)
276
277#define sqr_add_c(a,i,c0,c1,c2) do { \
278 asm ("mulq %2" \
279 : "=a"(t1),"=d"(t2) \
280 : "a"(a[i]) \
281 : "cc"); \
282 asm ("addq %2,%0; adcq %3,%1" \
283 : "+r"(c0),"+d"(t2) \
284 : "a"(t1),"g"(0) \
285 : "cc"); \
286 asm ("addq %2,%0; adcq %3,%1" \
287 : "+r"(c1),"+r"(c2) \
288 : "d"(t2),"g"(0) \
289 : "cc"); \
290 } while (0)
291
292#define mul_add_c2(a,b,c0,c1,c2) do { \
293 asm ("mulq %3" \
294 : "=a"(t1),"=d"(t2) \
295 : "a"(a),"m"(b) \
296 : "cc"); \
297 asm ("addq %0,%0; adcq %2,%1" \
298 : "+d"(t2),"+r"(c2) \
299 : "g"(0) \
300 : "cc"); \
301 asm ("addq %0,%0; adcq %2,%1" \
302 : "+a"(t1),"+d"(t2) \
303 : "g"(0) \
304 : "cc"); \
305 asm ("addq %2,%0; adcq %3,%1" \
306 : "+r"(c0),"+d"(t2) \
307 : "a"(t1),"g"(0) \
308 : "cc"); \
309 asm ("addq %2,%0; adcq %3,%1" \
310 : "+r"(c1),"+r"(c2) \
311 : "d"(t2),"g"(0) \
312 : "cc"); \
313 } while (0)
314#endif
315
316#define sqr_add_c2(a,i,j,c0,c1,c2) \
317 mul_add_c2((a)[i],(a)[j],c0,c1,c2)
318
319void bn_mul_comba8(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b)
320 {
321 BN_ULONG bl,bh;
322 BN_ULONG t1,t2;
323 BN_ULONG c1,c2,c3;
324
325 c1=0;
326 c2=0;
327 c3=0;
328 mul_add_c(a[0],b[0],c1,c2,c3);
329 r[0]=c1;
330 c1=0;
331 mul_add_c(a[0],b[1],c2,c3,c1);
332 mul_add_c(a[1],b[0],c2,c3,c1);
333 r[1]=c2;
334 c2=0;
335 mul_add_c(a[2],b[0],c3,c1,c2);
336 mul_add_c(a[1],b[1],c3,c1,c2);
337 mul_add_c(a[0],b[2],c3,c1,c2);
338 r[2]=c3;
339 c3=0;
340 mul_add_c(a[0],b[3],c1,c2,c3);
341 mul_add_c(a[1],b[2],c1,c2,c3);
342 mul_add_c(a[2],b[1],c1,c2,c3);
343 mul_add_c(a[3],b[0],c1,c2,c3);
344 r[3]=c1;
345 c1=0;
346 mul_add_c(a[4],b[0],c2,c3,c1);
347 mul_add_c(a[3],b[1],c2,c3,c1);
348 mul_add_c(a[2],b[2],c2,c3,c1);
349 mul_add_c(a[1],b[3],c2,c3,c1);
350 mul_add_c(a[0],b[4],c2,c3,c1);
351 r[4]=c2;
352 c2=0;
353 mul_add_c(a[0],b[5],c3,c1,c2);
354 mul_add_c(a[1],b[4],c3,c1,c2);
355 mul_add_c(a[2],b[3],c3,c1,c2);
356 mul_add_c(a[3],b[2],c3,c1,c2);
357 mul_add_c(a[4],b[1],c3,c1,c2);
358 mul_add_c(a[5],b[0],c3,c1,c2);
359 r[5]=c3;
360 c3=0;
361 mul_add_c(a[6],b[0],c1,c2,c3);
362 mul_add_c(a[5],b[1],c1,c2,c3);
363 mul_add_c(a[4],b[2],c1,c2,c3);
364 mul_add_c(a[3],b[3],c1,c2,c3);
365 mul_add_c(a[2],b[4],c1,c2,c3);
366 mul_add_c(a[1],b[5],c1,c2,c3);
367 mul_add_c(a[0],b[6],c1,c2,c3);
368 r[6]=c1;
369 c1=0;
370 mul_add_c(a[0],b[7],c2,c3,c1);
371 mul_add_c(a[1],b[6],c2,c3,c1);
372 mul_add_c(a[2],b[5],c2,c3,c1);
373 mul_add_c(a[3],b[4],c2,c3,c1);
374 mul_add_c(a[4],b[3],c2,c3,c1);
375 mul_add_c(a[5],b[2],c2,c3,c1);
376 mul_add_c(a[6],b[1],c2,c3,c1);
377 mul_add_c(a[7],b[0],c2,c3,c1);
378 r[7]=c2;
379 c2=0;
380 mul_add_c(a[7],b[1],c3,c1,c2);
381 mul_add_c(a[6],b[2],c3,c1,c2);
382 mul_add_c(a[5],b[3],c3,c1,c2);
383 mul_add_c(a[4],b[4],c3,c1,c2);
384 mul_add_c(a[3],b[5],c3,c1,c2);
385 mul_add_c(a[2],b[6],c3,c1,c2);
386 mul_add_c(a[1],b[7],c3,c1,c2);
387 r[8]=c3;
388 c3=0;
389 mul_add_c(a[2],b[7],c1,c2,c3);
390 mul_add_c(a[3],b[6],c1,c2,c3);
391 mul_add_c(a[4],b[5],c1,c2,c3);
392 mul_add_c(a[5],b[4],c1,c2,c3);
393 mul_add_c(a[6],b[3],c1,c2,c3);
394 mul_add_c(a[7],b[2],c1,c2,c3);
395 r[9]=c1;
396 c1=0;
397 mul_add_c(a[7],b[3],c2,c3,c1);
398 mul_add_c(a[6],b[4],c2,c3,c1);
399 mul_add_c(a[5],b[5],c2,c3,c1);
400 mul_add_c(a[4],b[6],c2,c3,c1);
401 mul_add_c(a[3],b[7],c2,c3,c1);
402 r[10]=c2;
403 c2=0;
404 mul_add_c(a[4],b[7],c3,c1,c2);
405 mul_add_c(a[5],b[6],c3,c1,c2);
406 mul_add_c(a[6],b[5],c3,c1,c2);
407 mul_add_c(a[7],b[4],c3,c1,c2);
408 r[11]=c3;
409 c3=0;
410 mul_add_c(a[7],b[5],c1,c2,c3);
411 mul_add_c(a[6],b[6],c1,c2,c3);
412 mul_add_c(a[5],b[7],c1,c2,c3);
413 r[12]=c1;
414 c1=0;
415 mul_add_c(a[6],b[7],c2,c3,c1);
416 mul_add_c(a[7],b[6],c2,c3,c1);
417 r[13]=c2;
418 c2=0;
419 mul_add_c(a[7],b[7],c3,c1,c2);
420 r[14]=c3;
421 r[15]=c1;
422 }
423
424void bn_mul_comba4(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b)
425 {
426 BN_ULONG bl,bh;
427 BN_ULONG t1,t2;
428 BN_ULONG c1,c2,c3;
429
430 c1=0;
431 c2=0;
432 c3=0;
433 mul_add_c(a[0],b[0],c1,c2,c3);
434 r[0]=c1;
435 c1=0;
436 mul_add_c(a[0],b[1],c2,c3,c1);
437 mul_add_c(a[1],b[0],c2,c3,c1);
438 r[1]=c2;
439 c2=0;
440 mul_add_c(a[2],b[0],c3,c1,c2);
441 mul_add_c(a[1],b[1],c3,c1,c2);
442 mul_add_c(a[0],b[2],c3,c1,c2);
443 r[2]=c3;
444 c3=0;
445 mul_add_c(a[0],b[3],c1,c2,c3);
446 mul_add_c(a[1],b[2],c1,c2,c3);
447 mul_add_c(a[2],b[1],c1,c2,c3);
448 mul_add_c(a[3],b[0],c1,c2,c3);
449 r[3]=c1;
450 c1=0;
451 mul_add_c(a[3],b[1],c2,c3,c1);
452 mul_add_c(a[2],b[2],c2,c3,c1);
453 mul_add_c(a[1],b[3],c2,c3,c1);
454 r[4]=c2;
455 c2=0;
456 mul_add_c(a[2],b[3],c3,c1,c2);
457 mul_add_c(a[3],b[2],c3,c1,c2);
458 r[5]=c3;
459 c3=0;
460 mul_add_c(a[3],b[3],c1,c2,c3);
461 r[6]=c1;
462 r[7]=c2;
463 }
464
465void bn_sqr_comba8(BN_ULONG *r, BN_ULONG *a)
466 {
467 BN_ULONG bl,bh;
468 BN_ULONG t1,t2;
469 BN_ULONG c1,c2,c3;
470
471 c1=0;
472 c2=0;
473 c3=0;
474 sqr_add_c(a,0,c1,c2,c3);
475 r[0]=c1;
476 c1=0;
477 sqr_add_c2(a,1,0,c2,c3,c1);
478 r[1]=c2;
479 c2=0;
480 sqr_add_c(a,1,c3,c1,c2);
481 sqr_add_c2(a,2,0,c3,c1,c2);
482 r[2]=c3;
483 c3=0;
484 sqr_add_c2(a,3,0,c1,c2,c3);
485 sqr_add_c2(a,2,1,c1,c2,c3);
486 r[3]=c1;
487 c1=0;
488 sqr_add_c(a,2,c2,c3,c1);
489 sqr_add_c2(a,3,1,c2,c3,c1);
490 sqr_add_c2(a,4,0,c2,c3,c1);
491 r[4]=c2;
492 c2=0;
493 sqr_add_c2(a,5,0,c3,c1,c2);
494 sqr_add_c2(a,4,1,c3,c1,c2);
495 sqr_add_c2(a,3,2,c3,c1,c2);
496 r[5]=c3;
497 c3=0;
498 sqr_add_c(a,3,c1,c2,c3);
499 sqr_add_c2(a,4,2,c1,c2,c3);
500 sqr_add_c2(a,5,1,c1,c2,c3);
501 sqr_add_c2(a,6,0,c1,c2,c3);
502 r[6]=c1;
503 c1=0;
504 sqr_add_c2(a,7,0,c2,c3,c1);
505 sqr_add_c2(a,6,1,c2,c3,c1);
506 sqr_add_c2(a,5,2,c2,c3,c1);
507 sqr_add_c2(a,4,3,c2,c3,c1);
508 r[7]=c2;
509 c2=0;
510 sqr_add_c(a,4,c3,c1,c2);
511 sqr_add_c2(a,5,3,c3,c1,c2);
512 sqr_add_c2(a,6,2,c3,c1,c2);
513 sqr_add_c2(a,7,1,c3,c1,c2);
514 r[8]=c3;
515 c3=0;
516 sqr_add_c2(a,7,2,c1,c2,c3);
517 sqr_add_c2(a,6,3,c1,c2,c3);
518 sqr_add_c2(a,5,4,c1,c2,c3);
519 r[9]=c1;
520 c1=0;
521 sqr_add_c(a,5,c2,c3,c1);
522 sqr_add_c2(a,6,4,c2,c3,c1);
523 sqr_add_c2(a,7,3,c2,c3,c1);
524 r[10]=c2;
525 c2=0;
526 sqr_add_c2(a,7,4,c3,c1,c2);
527 sqr_add_c2(a,6,5,c3,c1,c2);
528 r[11]=c3;
529 c3=0;
530 sqr_add_c(a,6,c1,c2,c3);
531 sqr_add_c2(a,7,5,c1,c2,c3);
532 r[12]=c1;
533 c1=0;
534 sqr_add_c2(a,7,6,c2,c3,c1);
535 r[13]=c2;
536 c2=0;
537 sqr_add_c(a,7,c3,c1,c2);
538 r[14]=c3;
539 r[15]=c1;
540 }
541
542void bn_sqr_comba4(BN_ULONG *r, BN_ULONG *a)
543 {
544 BN_ULONG bl,bh;
545 BN_ULONG t1,t2;
546 BN_ULONG c1,c2,c3;
547
548 c1=0;
549 c2=0;
550 c3=0;
551 sqr_add_c(a,0,c1,c2,c3);
552 r[0]=c1;
553 c1=0;
554 sqr_add_c2(a,1,0,c2,c3,c1);
555 r[1]=c2;
556 c2=0;
557 sqr_add_c(a,1,c3,c1,c2);
558 sqr_add_c2(a,2,0,c3,c1,c2);
559 r[2]=c3;
560 c3=0;
561 sqr_add_c2(a,3,0,c1,c2,c3);
562 sqr_add_c2(a,2,1,c1,c2,c3);
563 r[3]=c1;
564 c1=0;
565 sqr_add_c(a,2,c2,c3,c1);
566 sqr_add_c2(a,3,1,c2,c3,c1);
567 r[4]=c2;
568 c2=0;
569 sqr_add_c2(a,3,2,c3,c1,c2);
570 r[5]=c3;
571 c3=0;
572 sqr_add_c(a,3,c1,c2,c3);
573 r[6]=c1;
574 r[7]=c2;
575 }
generated by cgit v1.2.3-55-g6feb (git 2.39.1) at 2025-03-19 21:44:59 +0000