Newer
Older
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
;
; PA-RISC 2.0 implementation of bn_asm code, based on the
; 64-bit version of the code. This code is effectively the
; same as the 64-bit version except the register model is
; slightly different given all values must be 32-bit between
; function calls. Thus the 64-bit return values are returned
; in %ret0 and %ret1 vs just %ret0 as is done in 64-bit
;
;
; This code is approximately 2x faster than the C version
; for RSA/DSA.
;
; See http://devresource.hp.com/ for more details on the PA-RISC
; architecture. Also see the book "PA-RISC 2.0 Architecture"
; by Gerry Kane for information on the instruction set architecture.
;
; Code written by Chris Ruemmler (with some help from the HP C
; compiler).
;
; The code compiles with HP's assembler
;
.level 2.0N
.space $TEXT$
.subspa $CODE$,QUAD=0,ALIGN=8,ACCESS=0x2c,CODE_ONLY
;
; Global Register definitions used for the routines.
;
; Some information about HP's runtime architecture for 32-bits.
;
; "Caller save" means the calling function must save the register
; if it wants the register to be preserved.
; "Callee save" means if a function uses the register, it must save
; the value before using it.
;
; For the floating point registers
;
; "caller save" registers: fr4-fr11, fr22-fr31
; "callee save" registers: fr12-fr21
; "special" registers: fr0-fr3 (status and exception registers)
;
; For the integer registers
; value zero : r0
; "caller save" registers: r1,r19-r26
; "callee save" registers: r3-r18
; return register : r2 (rp)
; return values ; r28,r29 (ret0,ret1)
; Stack pointer ; r30 (sp)
; millicode return ptr ; r31 (also a caller save register)
;
; Arguments to the routines
;
r_ptr .reg %r26
a_ptr .reg %r25
b_ptr .reg %r24
num .reg %r24
n .reg %r23
;
; Note that the "w" argument for bn_mul_add_words and bn_mul_words
; is passed on the stack at a delta of -56 from the top of stack
; as the routine is entered.
;
;
; Globals used in some routines
;
top_overflow .reg %r23
high_mask .reg %r22 ; value 0xffffffff80000000L
;------------------------------------------------------------------------------
;
; bn_mul_add_words
;
;BN_ULONG bn_mul_add_words(BN_ULONG *r_ptr, BN_ULONG *a_ptr,
; int num, BN_ULONG w)
;
; arg0 = r_ptr
; arg1 = a_ptr
; arg3 = num
; -56(sp) = w
;
; Local register definitions
;
fm1 .reg %fr22
fm .reg %fr23
ht_temp .reg %fr24
ht_temp_1 .reg %fr25
lt_temp .reg %fr26
lt_temp_1 .reg %fr27
fm1_1 .reg %fr28
fm_1 .reg %fr29
fw_h .reg %fr7L
fw_l .reg %fr7R
fw .reg %fr7
fht_0 .reg %fr8L
flt_0 .reg %fr8R
t_float_0 .reg %fr8
fht_1 .reg %fr9L
flt_1 .reg %fr9R
t_float_1 .reg %fr9
tmp_0 .reg %r31
tmp_1 .reg %r21
m_0 .reg %r20
m_1 .reg %r19
ht_0 .reg %r1
ht_1 .reg %r3
lt_0 .reg %r4
lt_1 .reg %r5
m1_0 .reg %r6
m1_1 .reg %r7
rp_val .reg %r8
rp_val_1 .reg %r9
bn_mul_add_words
.export bn_mul_add_words,entry,NO_RELOCATION,LONG_RETURN
.proc
.callinfo frame=128
.entry
.align 64
STD %r3,0(%sp) ; save r3
STD %r4,8(%sp) ; save r4
NOP ; Needed to make the loop 16-byte aligned
NOP ; needed to make the loop 16-byte aligned
STD %r5,16(%sp) ; save r5
NOP
STD %r6,24(%sp) ; save r6
STD %r7,32(%sp) ; save r7
STD %r8,40(%sp) ; save r8
STD %r9,48(%sp) ; save r9
COPY %r0,%ret1 ; return 0 by default
DEPDI,Z 1,31,1,top_overflow ; top_overflow = 1 << 32
CMPIB,>= 0,num,bn_mul_add_words_exit ; if (num <= 0) then exit
LDO 128(%sp),%sp ; bump stack
;
; The loop is unrolled twice, so if there is only 1 number
; then go straight to the cleanup code.
;
CMPIB,= 1,num,bn_mul_add_words_single_top
FLDD -184(%sp),fw ; (-56-128) load up w into fw (fw_h/fw_l)
;
; This loop is unrolled 2 times (64-byte aligned as well)
;
; PA-RISC 2.0 chips have two fully pipelined multipliers, thus
; two 32-bit mutiplies can be issued per cycle.
;
bn_mul_add_words_unroll2
FLDD 0(a_ptr),t_float_0 ; load up 64-bit value (fr8L) ht(L)/lt(R)
FLDD 8(a_ptr),t_float_1 ; load up 64-bit value (fr8L) ht(L)/lt(R)
LDD 0(r_ptr),rp_val ; rp[0]
LDD 8(r_ptr),rp_val_1 ; rp[1]
XMPYU fht_0,fw_l,fm1 ; m1[0] = fht_0*fw_l
XMPYU fht_1,fw_l,fm1_1 ; m1[1] = fht_1*fw_l
FSTD fm1,-16(%sp) ; -16(sp) = m1[0]
FSTD fm1_1,-48(%sp) ; -48(sp) = m1[1]
XMPYU flt_0,fw_h,fm ; m[0] = flt_0*fw_h
XMPYU flt_1,fw_h,fm_1 ; m[1] = flt_1*fw_h
FSTD fm,-8(%sp) ; -8(sp) = m[0]
FSTD fm_1,-40(%sp) ; -40(sp) = m[1]
XMPYU fht_0,fw_h,ht_temp ; ht_temp = fht_0*fw_h
XMPYU fht_1,fw_h,ht_temp_1 ; ht_temp_1 = fht_1*fw_h
FSTD ht_temp,-24(%sp) ; -24(sp) = ht_temp
FSTD ht_temp_1,-56(%sp) ; -56(sp) = ht_temp_1
XMPYU flt_0,fw_l,lt_temp ; lt_temp = lt*fw_l
XMPYU flt_1,fw_l,lt_temp_1 ; lt_temp = lt*fw_l
FSTD lt_temp,-32(%sp) ; -32(sp) = lt_temp
FSTD lt_temp_1,-64(%sp) ; -64(sp) = lt_temp_1
LDD -8(%sp),m_0 ; m[0]
LDD -40(%sp),m_1 ; m[1]
LDD -16(%sp),m1_0 ; m1[0]
LDD -48(%sp),m1_1 ; m1[1]
LDD -24(%sp),ht_0 ; ht[0]
LDD -56(%sp),ht_1 ; ht[1]
ADD,L m1_0,m_0,tmp_0 ; tmp_0 = m[0] + m1[0];
ADD,L m1_1,m_1,tmp_1 ; tmp_1 = m[1] + m1[1];
LDD -32(%sp),lt_0
LDD -64(%sp),lt_1
CMPCLR,*>>= tmp_0,m1_0, %r0 ; if (m[0] < m1[0])
ADD,L ht_0,top_overflow,ht_0 ; ht[0] += (1<<32)
CMPCLR,*>>= tmp_1,m1_1,%r0 ; if (m[1] < m1[1])
ADD,L ht_1,top_overflow,ht_1 ; ht[1] += (1<<32)
EXTRD,U tmp_0,31,32,m_0 ; m[0]>>32
DEPD,Z tmp_0,31,32,m1_0 ; m1[0] = m[0]<<32
EXTRD,U tmp_1,31,32,m_1 ; m[1]>>32
DEPD,Z tmp_1,31,32,m1_1 ; m1[1] = m[1]<<32
ADD,L ht_0,m_0,ht_0 ; ht[0]+= (m[0]>>32)
ADD,L ht_1,m_1,ht_1 ; ht[1]+= (m[1]>>32)
ADD lt_0,m1_0,lt_0 ; lt[0] = lt[0]+m1[0];
ADD,DC ht_0,%r0,ht_0 ; ht[0]++
ADD lt_1,m1_1,lt_1 ; lt[1] = lt[1]+m1[1];
ADD,DC ht_1,%r0,ht_1 ; ht[1]++
ADD %ret1,lt_0,lt_0 ; lt[0] = lt[0] + c;
ADD,DC ht_0,%r0,ht_0 ; ht[0]++
ADD lt_0,rp_val,lt_0 ; lt[0] = lt[0]+rp[0]
ADD,DC ht_0,%r0,ht_0 ; ht[0]++
LDO -2(num),num ; num = num - 2;
ADD ht_0,lt_1,lt_1 ; lt[1] = lt[1] + ht_0 (c);
ADD,DC ht_1,%r0,ht_1 ; ht[1]++
STD lt_0,0(r_ptr) ; rp[0] = lt[0]
ADD lt_1,rp_val_1,lt_1 ; lt[1] = lt[1]+rp[1]
ADD,DC ht_1,%r0,%ret1 ; ht[1]++
LDO 16(a_ptr),a_ptr ; a_ptr += 2
STD lt_1,8(r_ptr) ; rp[1] = lt[1]
CMPIB,<= 2,num,bn_mul_add_words_unroll2 ; go again if more to do
LDO 16(r_ptr),r_ptr ; r_ptr += 2
CMPIB,=,N 0,num,bn_mul_add_words_exit ; are we done, or cleanup last one
;
; Top of loop aligned on 64-byte boundary
;
bn_mul_add_words_single_top
FLDD 0(a_ptr),t_float_0 ; load up 64-bit value (fr8L) ht(L)/lt(R)
LDD 0(r_ptr),rp_val ; rp[0]
LDO 8(a_ptr),a_ptr ; a_ptr++
XMPYU fht_0,fw_l,fm1 ; m1 = ht*fw_l
FSTD fm1,-16(%sp) ; -16(sp) = m1
XMPYU flt_0,fw_h,fm ; m = lt*fw_h
FSTD fm,-8(%sp) ; -8(sp) = m
XMPYU fht_0,fw_h,ht_temp ; ht_temp = ht*fw_h
FSTD ht_temp,-24(%sp) ; -24(sp) = ht
XMPYU flt_0,fw_l,lt_temp ; lt_temp = lt*fw_l
FSTD lt_temp,-32(%sp) ; -32(sp) = lt
LDD -8(%sp),m_0
LDD -16(%sp),m1_0 ; m1 = temp1
ADD,L m_0,m1_0,tmp_0 ; tmp_0 = m + m1;
LDD -24(%sp),ht_0
LDD -32(%sp),lt_0
CMPCLR,*>>= tmp_0,m1_0,%r0 ; if (m < m1)
ADD,L ht_0,top_overflow,ht_0 ; ht += (1<<32)
EXTRD,U tmp_0,31,32,m_0 ; m>>32
DEPD,Z tmp_0,31,32,m1_0 ; m1 = m<<32
ADD,L ht_0,m_0,ht_0 ; ht+= (m>>32)
ADD lt_0,m1_0,tmp_0 ; tmp_0 = lt+m1;
ADD,DC ht_0,%r0,ht_0 ; ht++
ADD %ret1,tmp_0,lt_0 ; lt = lt + c;
ADD,DC ht_0,%r0,ht_0 ; ht++
ADD lt_0,rp_val,lt_0 ; lt = lt+rp[0]
ADD,DC ht_0,%r0,%ret1 ; ht++
STD lt_0,0(r_ptr) ; rp[0] = lt
bn_mul_add_words_exit
.EXIT
EXTRD,U %ret1,31,32,%ret0 ; for 32-bit, return in ret0/ret1
LDD -80(%sp),%r9 ; restore r9
LDD -88(%sp),%r8 ; restore r8
LDD -96(%sp),%r7 ; restore r7
LDD -104(%sp),%r6 ; restore r6
LDD -112(%sp),%r5 ; restore r5
LDD -120(%sp),%r4 ; restore r4
BVE (%rp)
LDD,MB -128(%sp),%r3 ; restore r3
.PROCEND ;in=23,24,25,26,29;out=28;
;----------------------------------------------------------------------------
;
;BN_ULONG bn_mul_words(BN_ULONG *rp, BN_ULONG *ap, int num, BN_ULONG w)
;
; arg0 = rp
; arg1 = ap
; arg3 = num
; w on stack at -56(sp)
bn_mul_words
.proc
.callinfo frame=128
.entry
.EXPORT bn_mul_words,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
.align 64
STD %r3,0(%sp) ; save r3
STD %r4,8(%sp) ; save r4
NOP
STD %r5,16(%sp) ; save r5
STD %r6,24(%sp) ; save r6
STD %r7,32(%sp) ; save r7
COPY %r0,%ret1 ; return 0 by default
DEPDI,Z 1,31,1,top_overflow ; top_overflow = 1 << 32
CMPIB,>= 0,num,bn_mul_words_exit
LDO 128(%sp),%sp ; bump stack
;
; See if only 1 word to do, thus just do cleanup
;
CMPIB,= 1,num,bn_mul_words_single_top
FLDD -184(%sp),fw ; (-56-128) load up w into fw (fw_h/fw_l)
;
; This loop is unrolled 2 times (64-byte aligned as well)
;
; PA-RISC 2.0 chips have two fully pipelined multipliers, thus
; two 32-bit mutiplies can be issued per cycle.
;
bn_mul_words_unroll2
FLDD 0(a_ptr),t_float_0 ; load up 64-bit value (fr8L) ht(L)/lt(R)
FLDD 8(a_ptr),t_float_1 ; load up 64-bit value (fr8L) ht(L)/lt(R)
XMPYU fht_0,fw_l,fm1 ; m1[0] = fht_0*fw_l
XMPYU fht_1,fw_l,fm1_1 ; m1[1] = ht*fw_l
FSTD fm1,-16(%sp) ; -16(sp) = m1
FSTD fm1_1,-48(%sp) ; -48(sp) = m1
XMPYU flt_0,fw_h,fm ; m = lt*fw_h
XMPYU flt_1,fw_h,fm_1 ; m = lt*fw_h
FSTD fm,-8(%sp) ; -8(sp) = m
FSTD fm_1,-40(%sp) ; -40(sp) = m
XMPYU fht_0,fw_h,ht_temp ; ht_temp = fht_0*fw_h
XMPYU fht_1,fw_h,ht_temp_1 ; ht_temp = ht*fw_h
FSTD ht_temp,-24(%sp) ; -24(sp) = ht
FSTD ht_temp_1,-56(%sp) ; -56(sp) = ht
XMPYU flt_0,fw_l,lt_temp ; lt_temp = lt*fw_l
XMPYU flt_1,fw_l,lt_temp_1 ; lt_temp = lt*fw_l
FSTD lt_temp,-32(%sp) ; -32(sp) = lt
FSTD lt_temp_1,-64(%sp) ; -64(sp) = lt
LDD -8(%sp),m_0
LDD -40(%sp),m_1
LDD -16(%sp),m1_0
LDD -48(%sp),m1_1
LDD -24(%sp),ht_0
LDD -56(%sp),ht_1
ADD,L m1_0,m_0,tmp_0 ; tmp_0 = m + m1;
ADD,L m1_1,m_1,tmp_1 ; tmp_1 = m + m1;
LDD -32(%sp),lt_0
LDD -64(%sp),lt_1
CMPCLR,*>>= tmp_0,m1_0, %r0 ; if (m < m1)
ADD,L ht_0,top_overflow,ht_0 ; ht += (1<<32)
CMPCLR,*>>= tmp_1,m1_1,%r0 ; if (m < m1)
ADD,L ht_1,top_overflow,ht_1 ; ht += (1<<32)
EXTRD,U tmp_0,31,32,m_0 ; m>>32
DEPD,Z tmp_0,31,32,m1_0 ; m1 = m<<32
EXTRD,U tmp_1,31,32,m_1 ; m>>32
DEPD,Z tmp_1,31,32,m1_1 ; m1 = m<<32
ADD,L ht_0,m_0,ht_0 ; ht+= (m>>32)
ADD,L ht_1,m_1,ht_1 ; ht+= (m>>32)
ADD lt_0,m1_0,lt_0 ; lt = lt+m1;
ADD,DC ht_0,%r0,ht_0 ; ht++
ADD lt_1,m1_1,lt_1 ; lt = lt+m1;
ADD,DC ht_1,%r0,ht_1 ; ht++
ADD %ret1,lt_0,lt_0 ; lt = lt + c (ret1);
ADD,DC ht_0,%r0,ht_0 ; ht++
ADD ht_0,lt_1,lt_1 ; lt = lt + c (ht_0)
ADD,DC ht_1,%r0,ht_1 ; ht++
STD lt_0,0(r_ptr) ; rp[0] = lt
STD lt_1,8(r_ptr) ; rp[1] = lt
COPY ht_1,%ret1 ; carry = ht
LDO -2(num),num ; num = num - 2;
LDO 16(a_ptr),a_ptr ; ap += 2
CMPIB,<= 2,num,bn_mul_words_unroll2
LDO 16(r_ptr),r_ptr ; rp++
CMPIB,=,N 0,num,bn_mul_words_exit ; are we done?
;
; Top of loop aligned on 64-byte boundary
;
bn_mul_words_single_top
FLDD 0(a_ptr),t_float_0 ; load up 64-bit value (fr8L) ht(L)/lt(R)
XMPYU fht_0,fw_l,fm1 ; m1 = ht*fw_l
FSTD fm1,-16(%sp) ; -16(sp) = m1
XMPYU flt_0,fw_h,fm ; m = lt*fw_h
FSTD fm,-8(%sp) ; -8(sp) = m
XMPYU fht_0,fw_h,ht_temp ; ht_temp = ht*fw_h
FSTD ht_temp,-24(%sp) ; -24(sp) = ht
XMPYU flt_0,fw_l,lt_temp ; lt_temp = lt*fw_l
FSTD lt_temp,-32(%sp) ; -32(sp) = lt
LDD -8(%sp),m_0
LDD -16(%sp),m1_0
ADD,L m_0,m1_0,tmp_0 ; tmp_0 = m + m1;
LDD -24(%sp),ht_0
LDD -32(%sp),lt_0
CMPCLR,*>>= tmp_0,m1_0,%r0 ; if (m < m1)
ADD,L ht_0,top_overflow,ht_0 ; ht += (1<<32)
EXTRD,U tmp_0,31,32,m_0 ; m>>32
DEPD,Z tmp_0,31,32,m1_0 ; m1 = m<<32
ADD,L ht_0,m_0,ht_0 ; ht+= (m>>32)
ADD lt_0,m1_0,lt_0 ; lt= lt+m1;
ADD,DC ht_0,%r0,ht_0 ; ht++
ADD %ret1,lt_0,lt_0 ; lt = lt + c;
ADD,DC ht_0,%r0,ht_0 ; ht++
COPY ht_0,%ret1 ; copy carry
STD lt_0,0(r_ptr) ; rp[0] = lt
bn_mul_words_exit
.EXIT
EXTRD,U %ret1,31,32,%ret0 ; for 32-bit, return in ret0/ret1
LDD -96(%sp),%r7 ; restore r7
LDD -104(%sp),%r6 ; restore r6
LDD -112(%sp),%r5 ; restore r5
LDD -120(%sp),%r4 ; restore r4
BVE (%rp)
LDD,MB -128(%sp),%r3 ; restore r3
.PROCEND
;----------------------------------------------------------------------------
;
;void bn_sqr_words(BN_ULONG *rp, BN_ULONG *ap, int num)
;
; arg0 = rp
; arg1 = ap
; arg2 = num
;
bn_sqr_words
.proc
.callinfo FRAME=128,ENTRY_GR=%r3,ARGS_SAVED,ORDERING_AWARE
.EXPORT bn_sqr_words,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
.entry
.align 64
STD %r3,0(%sp) ; save r3
STD %r4,8(%sp) ; save r4
NOP
STD %r5,16(%sp) ; save r5
CMPIB,>= 0,num,bn_sqr_words_exit
LDO 128(%sp),%sp ; bump stack
;
; If only 1, the goto straight to cleanup
;
CMPIB,= 1,num,bn_sqr_words_single_top
DEPDI,Z -1,32,33,high_mask ; Create Mask 0xffffffff80000000L
;
; This loop is unrolled 2 times (64-byte aligned as well)
;
bn_sqr_words_unroll2
FLDD 0(a_ptr),t_float_0 ; a[0]
FLDD 8(a_ptr),t_float_1 ; a[1]
XMPYU fht_0,flt_0,fm ; m[0]
XMPYU fht_1,flt_1,fm_1 ; m[1]
FSTD fm,-24(%sp) ; store m[0]
FSTD fm_1,-56(%sp) ; store m[1]
XMPYU flt_0,flt_0,lt_temp ; lt[0]
XMPYU flt_1,flt_1,lt_temp_1 ; lt[1]
FSTD lt_temp,-16(%sp) ; store lt[0]
FSTD lt_temp_1,-48(%sp) ; store lt[1]
XMPYU fht_0,fht_0,ht_temp ; ht[0]
XMPYU fht_1,fht_1,ht_temp_1 ; ht[1]
FSTD ht_temp,-8(%sp) ; store ht[0]
FSTD ht_temp_1,-40(%sp) ; store ht[1]
LDD -24(%sp),m_0
LDD -56(%sp),m_1
AND m_0,high_mask,tmp_0 ; m[0] & Mask
AND m_1,high_mask,tmp_1 ; m[1] & Mask
DEPD,Z m_0,30,31,m_0 ; m[0] << 32+1
DEPD,Z m_1,30,31,m_1 ; m[1] << 32+1
LDD -16(%sp),lt_0
LDD -48(%sp),lt_1
EXTRD,U tmp_0,32,33,tmp_0 ; tmp_0 = m[0]&Mask >> 32-1
EXTRD,U tmp_1,32,33,tmp_1 ; tmp_1 = m[1]&Mask >> 32-1
LDD -8(%sp),ht_0
LDD -40(%sp),ht_1
ADD,L ht_0,tmp_0,ht_0 ; ht[0] += tmp_0
ADD,L ht_1,tmp_1,ht_1 ; ht[1] += tmp_1
ADD lt_0,m_0,lt_0 ; lt = lt+m
ADD,DC ht_0,%r0,ht_0 ; ht[0]++
STD lt_0,0(r_ptr) ; rp[0] = lt[0]
STD ht_0,8(r_ptr) ; rp[1] = ht[1]
ADD lt_1,m_1,lt_1 ; lt = lt+m
ADD,DC ht_1,%r0,ht_1 ; ht[1]++
STD lt_1,16(r_ptr) ; rp[2] = lt[1]
STD ht_1,24(r_ptr) ; rp[3] = ht[1]
LDO -2(num),num ; num = num - 2;
LDO 16(a_ptr),a_ptr ; ap += 2
CMPIB,<= 2,num,bn_sqr_words_unroll2
LDO 32(r_ptr),r_ptr ; rp += 4
CMPIB,=,N 0,num,bn_sqr_words_exit ; are we done?
;
; Top of loop aligned on 64-byte boundary
;
bn_sqr_words_single_top
FLDD 0(a_ptr),t_float_0 ; load up 64-bit value (fr8L) ht(L)/lt(R)
XMPYU fht_0,flt_0,fm ; m
FSTD fm,-24(%sp) ; store m
XMPYU flt_0,flt_0,lt_temp ; lt
FSTD lt_temp,-16(%sp) ; store lt
XMPYU fht_0,fht_0,ht_temp ; ht
FSTD ht_temp,-8(%sp) ; store ht
LDD -24(%sp),m_0 ; load m
AND m_0,high_mask,tmp_0 ; m & Mask
DEPD,Z m_0,30,31,m_0 ; m << 32+1
LDD -16(%sp),lt_0 ; lt
LDD -8(%sp),ht_0 ; ht
EXTRD,U tmp_0,32,33,tmp_0 ; tmp_0 = m&Mask >> 32-1
ADD m_0,lt_0,lt_0 ; lt = lt+m
ADD,L ht_0,tmp_0,ht_0 ; ht += tmp_0
ADD,DC ht_0,%r0,ht_0 ; ht++
STD lt_0,0(r_ptr) ; rp[0] = lt
STD ht_0,8(r_ptr) ; rp[1] = ht
bn_sqr_words_exit
.EXIT
LDD -112(%sp),%r5 ; restore r5
LDD -120(%sp),%r4 ; restore r4
BVE (%rp)
LDD,MB -128(%sp),%r3
.PROCEND ;in=23,24,25,26,29;out=28;
;----------------------------------------------------------------------------
;
;BN_ULONG bn_add_words(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n)
;
; arg0 = rp
; arg1 = ap
; arg2 = bp
; arg3 = n
t .reg %r22
b .reg %r21
l .reg %r20
bn_add_words
.proc
.entry
.callinfo
.EXPORT bn_add_words,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
.align 64
CMPIB,>= 0,n,bn_add_words_exit
COPY %r0,%ret1 ; return 0 by default
;
; If 2 or more numbers do the loop
;
CMPIB,= 1,n,bn_add_words_single_top
NOP
;
; This loop is unrolled 2 times (64-byte aligned as well)
;
bn_add_words_unroll2
LDD 0(a_ptr),t
LDD 0(b_ptr),b
ADD t,%ret1,t ; t = t+c;
ADD,DC %r0,%r0,%ret1 ; set c to carry
ADD t,b,l ; l = t + b[0]
ADD,DC %ret1,%r0,%ret1 ; c+= carry
STD l,0(r_ptr)
LDD 8(a_ptr),t
LDD 8(b_ptr),b
ADD t,%ret1,t ; t = t+c;
ADD,DC %r0,%r0,%ret1 ; set c to carry
ADD t,b,l ; l = t + b[0]
ADD,DC %ret1,%r0,%ret1 ; c+= carry
STD l,8(r_ptr)
LDO -2(n),n
LDO 16(a_ptr),a_ptr
LDO 16(b_ptr),b_ptr
CMPIB,<= 2,n,bn_add_words_unroll2
LDO 16(r_ptr),r_ptr
CMPIB,=,N 0,n,bn_add_words_exit ; are we done?
bn_add_words_single_top
LDD 0(a_ptr),t
LDD 0(b_ptr),b
ADD t,%ret1,t ; t = t+c;
ADD,DC %r0,%r0,%ret1 ; set c to carry (could use CMPCLR??)
ADD t,b,l ; l = t + b[0]
ADD,DC %ret1,%r0,%ret1 ; c+= carry
STD l,0(r_ptr)
bn_add_words_exit
.EXIT
BVE (%rp)
EXTRD,U %ret1,31,32,%ret0 ; for 32-bit, return in ret0/ret1
.PROCEND ;in=23,24,25,26,29;out=28;
;----------------------------------------------------------------------------
;
;BN_ULONG bn_sub_words(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n)
;
; arg0 = rp
; arg1 = ap
; arg2 = bp
; arg3 = n
t1 .reg %r22
t2 .reg %r21
sub_tmp1 .reg %r20
sub_tmp2 .reg %r19
bn_sub_words
.proc
.callinfo
.EXPORT bn_sub_words,ENTRY,PRIV_LEV=3,NO_RELOCATION,LONG_RETURN
.entry
.align 64
CMPIB,>= 0,n,bn_sub_words_exit
COPY %r0,%ret1 ; return 0 by default
;
; If 2 or more numbers do the loop
;
CMPIB,= 1,n,bn_sub_words_single_top
NOP
;
; This loop is unrolled 2 times (64-byte aligned as well)
;
bn_sub_words_unroll2
LDD 0(a_ptr),t1
LDD 0(b_ptr),t2
SUB t1,t2,sub_tmp1 ; t3 = t1-t2;
SUB sub_tmp1,%ret1,sub_tmp1 ; t3 = t3- c;
CMPCLR,*>> t1,t2,sub_tmp2 ; clear if t1 > t2
LDO 1(%r0),sub_tmp2
CMPCLR,*= t1,t2,%r0
COPY sub_tmp2,%ret1
STD sub_tmp1,0(r_ptr)
LDD 8(a_ptr),t1
LDD 8(b_ptr),t2
SUB t1,t2,sub_tmp1 ; t3 = t1-t2;
SUB sub_tmp1,%ret1,sub_tmp1 ; t3 = t3- c;
CMPCLR,*>> t1,t2,sub_tmp2 ; clear if t1 > t2
LDO 1(%r0),sub_tmp2
CMPCLR,*= t1,t2,%r0
COPY sub_tmp2,%ret1
STD sub_tmp1,8(r_ptr)
LDO -2(n),n
LDO 16(a_ptr),a_ptr
LDO 16(b_ptr),b_ptr
CMPIB,<= 2,n,bn_sub_words_unroll2
LDO 16(r_ptr),r_ptr
CMPIB,=,N 0,n,bn_sub_words_exit ; are we done?
bn_sub_words_single_top
LDD 0(a_ptr),t1
LDD 0(b_ptr),t2
SUB t1,t2,sub_tmp1 ; t3 = t1-t2;
SUB sub_tmp1,%ret1,sub_tmp1 ; t3 = t3- c;
CMPCLR,*>> t1,t2,sub_tmp2 ; clear if t1 > t2
LDO 1(%r0),sub_tmp2
CMPCLR,*= t1,t2,%r0
COPY sub_tmp2,%ret1
STD sub_tmp1,0(r_ptr)
bn_sub_words_exit
.EXIT
BVE (%rp)
EXTRD,U %ret1,31,32,%ret0 ; for 32-bit, return in ret0/ret1
.PROCEND ;in=23,24,25,26,29;out=28;
;------------------------------------------------------------------------------
;
; unsigned long bn_div_words(unsigned long h, unsigned long l, unsigned long d)
;
; arg0 = h
; arg1 = l
; arg2 = d
;
; This is mainly just output from the HP C compiler.
;
;------------------------------------------------------------------------------
bn_div_words
.PROC
.EXPORT bn_div_words,ENTRY,PRIV_LEV=3,ARGW0=GR,ARGW1=GR,ARGW2=GR,ARGW3=GR,RTNVAL=GR,LONG_RETURN
.IMPORT BN_num_bits_word,CODE
;--- not PIC .IMPORT __iob,DATA
;--- not PIC .IMPORT fprintf,CODE
.IMPORT abort,CODE
.IMPORT $$div2U,MILLICODE
.CALLINFO CALLER,FRAME=144,ENTRY_GR=%r9,SAVE_RP,ARGS_SAVED,ORDERING_AWARE
.ENTRY
STW %r2,-20(%r30) ;offset 0x8ec
STW,MA %r3,192(%r30) ;offset 0x8f0
STW %r4,-188(%r30) ;offset 0x8f4
DEPD %r5,31,32,%r6 ;offset 0x8f8
STD %r6,-184(%r30) ;offset 0x8fc
DEPD %r7,31,32,%r8 ;offset 0x900
STD %r8,-176(%r30) ;offset 0x904
STW %r9,-168(%r30) ;offset 0x908
LDD -248(%r30),%r3 ;offset 0x90c
COPY %r26,%r4 ;offset 0x910
COPY %r24,%r5 ;offset 0x914
DEPD %r25,31,32,%r4 ;offset 0x918
CMPB,*<> %r3,%r0,$0006000C ;offset 0x91c
DEPD %r23,31,32,%r5 ;offset 0x920
MOVIB,TR -1,%r29,$00060002 ;offset 0x924
EXTRD,U %r29,31,32,%r28 ;offset 0x928
$0006002A
LDO -1(%r29),%r29 ;offset 0x92c
SUB %r23,%r7,%r23 ;offset 0x930
$00060024
SUB %r4,%r31,%r25 ;offset 0x934
AND %r25,%r19,%r26 ;offset 0x938
CMPB,*<>,N %r0,%r26,$00060046 ;offset 0x93c
DEPD,Z %r25,31,32,%r20 ;offset 0x940
OR %r20,%r24,%r21 ;offset 0x944
CMPB,*<<,N %r21,%r23,$0006002A ;offset 0x948
SUB %r31,%r2,%r31 ;offset 0x94c
$00060046
$0006002E
DEPD,Z %r23,31,32,%r25 ;offset 0x950
EXTRD,U %r23,31,32,%r26 ;offset 0x954
AND %r25,%r19,%r24 ;offset 0x958
ADD,L %r31,%r26,%r31 ;offset 0x95c
CMPCLR,*>>= %r5,%r24,%r0 ;offset 0x960
LDO 1(%r31),%r31 ;offset 0x964
$00060032
CMPB,*<<=,N %r31,%r4,$00060036 ;offset 0x968
LDO -1(%r29),%r29 ;offset 0x96c
ADD,L %r4,%r3,%r4 ;offset 0x970
$00060036
ADDIB,=,N -1,%r8,$D0 ;offset 0x974
SUB %r5,%r24,%r28 ;offset 0x978
$0006003A
SUB %r4,%r31,%r24 ;offset 0x97c
SHRPD %r24,%r28,32,%r4 ;offset 0x980
DEPD,Z %r29,31,32,%r9 ;offset 0x984
DEPD,Z %r28,31,32,%r5 ;offset 0x988
$0006001C
EXTRD,U %r4,31,32,%r31 ;offset 0x98c
CMPB,*<>,N %r31,%r2,$00060020 ;offset 0x990
MOVB,TR %r6,%r29,$D1 ;offset 0x994
STD %r29,-152(%r30) ;offset 0x998
$0006000C
EXTRD,U %r3,31,32,%r25 ;offset 0x99c
COPY %r3,%r26 ;offset 0x9a0
EXTRD,U %r3,31,32,%r9 ;offset 0x9a4
EXTRD,U %r4,31,32,%r8 ;offset 0x9a8
.CALL ARGW0=GR,ARGW1=GR,RTNVAL=GR ;in=25,26;out=28;
B,L BN_num_bits_word,%r2 ;offset 0x9ac
EXTRD,U %r5,31,32,%r7 ;offset 0x9b0
LDI 64,%r20 ;offset 0x9b4
DEPD %r7,31,32,%r5 ;offset 0x9b8
DEPD %r8,31,32,%r4 ;offset 0x9bc
DEPD %r9,31,32,%r3 ;offset 0x9c0
CMPB,= %r28,%r20,$00060012 ;offset 0x9c4
COPY %r28,%r24 ;offset 0x9c8
MTSARCM %r24 ;offset 0x9cc
DEPDI,Z -1,%sar,1,%r19 ;offset 0x9d0
CMPB,*>>,N %r4,%r19,$D2 ;offset 0x9d4
$00060012
SUBI 64,%r24,%r31 ;offset 0x9d8
CMPCLR,*<< %r4,%r3,%r0 ;offset 0x9dc
SUB %r4,%r3,%r4 ;offset 0x9e0
$00060016
CMPB,= %r31,%r0,$0006001A ;offset 0x9e4
COPY %r0,%r9 ;offset 0x9e8
MTSARCM %r31 ;offset 0x9ec
DEPD,Z %r3,%sar,64,%r3 ;offset 0x9f0
SUBI 64,%r31,%r26 ;offset 0x9f4
MTSAR %r26 ;offset 0x9f8
SHRPD %r4,%r5,%sar,%r4 ;offset 0x9fc
MTSARCM %r31 ;offset 0xa00
DEPD,Z %r5,%sar,64,%r5 ;offset 0xa04
$0006001A
DEPDI,Z -1,31,32,%r19 ;offset 0xa08
AND %r3,%r19,%r29 ;offset 0xa0c
EXTRD,U %r29,31,32,%r2 ;offset 0xa10
DEPDI,Z -1,63,32,%r6 ;offset 0xa14
MOVIB,TR 2,%r8,$0006001C ;offset 0xa18
EXTRD,U %r3,63,32,%r7 ;offset 0xa1c
$D2
;--- not PIC ADDIL LR'__iob-$global$,%r27,%r1 ;offset 0xa20
;--- not PIC LDIL LR'C$7,%r21 ;offset 0xa24
;--- not PIC LDO RR'__iob-$global$+32(%r1),%r26 ;offset 0xa28
;--- not PIC .CALL ARGW0=GR,ARGW1=GR,ARGW2=GR,RTNVAL=GR ;in=24,25,26;out=28;
;--- not PIC B,L fprintf,%r2 ;offset 0xa2c
;--- not PIC LDO RR'C$7(%r21),%r25 ;offset 0xa30
.CALL ;
B,L abort,%r2 ;offset 0xa34
NOP ;offset 0xa38
B $D3 ;offset 0xa3c
LDW -212(%r30),%r2 ;offset 0xa40
$00060020
COPY %r4,%r26 ;offset 0xa44
EXTRD,U %r4,31,32,%r25 ;offset 0xa48
COPY %r2,%r24 ;offset 0xa4c
.CALL ;in=23,24,25,26;out=20,21,22,28,29; (MILLICALL)
B,L $$div2U,%r31 ;offset 0xa50
EXTRD,U %r2,31,32,%r23 ;offset 0xa54
DEPD %r28,31,32,%r29 ;offset 0xa58
$00060022
STD %r29,-152(%r30) ;offset 0xa5c
$D1
AND %r5,%r19,%r24 ;offset 0xa60
EXTRD,U %r24,31,32,%r24 ;offset 0xa64
STW %r2,-160(%r30) ;offset 0xa68
STW %r7,-128(%r30) ;offset 0xa6c
FLDD -152(%r30),%fr4 ;offset 0xa70
FLDD -152(%r30),%fr7 ;offset 0xa74
FLDW -160(%r30),%fr8L ;offset 0xa78
FLDW -128(%r30),%fr5L ;offset 0xa7c
XMPYU %fr8L,%fr7L,%fr10 ;offset 0xa80
FSTD %fr10,-136(%r30) ;offset 0xa84
XMPYU %fr8L,%fr7R,%fr22 ;offset 0xa88
FSTD %fr22,-144(%r30) ;offset 0xa8c
XMPYU %fr5L,%fr4L,%fr11 ;offset 0xa90
XMPYU %fr5L,%fr4R,%fr23 ;offset 0xa94
FSTD %fr11,-112(%r30) ;offset 0xa98
FSTD %fr23,-120(%r30) ;offset 0xa9c
LDD -136(%r30),%r28 ;offset 0xaa0
DEPD,Z %r28,31,32,%r31 ;offset 0xaa4
LDD -144(%r30),%r20 ;offset 0xaa8
ADD,L %r20,%r31,%r31 ;offset 0xaac
LDD -112(%r30),%r22 ;offset 0xab0
DEPD,Z %r22,31,32,%r22 ;offset 0xab4
LDD -120(%r30),%r21 ;offset 0xab8
B $00060024 ;offset 0xabc
ADD,L %r21,%r22,%r23 ;offset 0xac0
$D0
OR %r9,%r29,%r29 ;offset 0xac4
$00060040
EXTRD,U %r29,31,32,%r28 ;offset 0xac8
$00060002
$L2
LDW -212(%r30),%r2 ;offset 0xacc
$D3
LDW -168(%r30),%r9 ;offset 0xad0
LDD -176(%r30),%r8 ;offset 0xad4
EXTRD,U %r8,31,32,%r7 ;offset 0xad8
LDD -184(%r30),%r6 ;offset 0xadc
EXTRD,U %r6,31,32,%r5 ;offset 0xae0
LDW -188(%r30),%r4 ;offset 0xae4
BVE (%r2) ;offset 0xae8
.EXIT
LDW,MB -192(%r30),%r3 ;offset 0xaec
.PROCEND ;in=23,25;out=28,29;fpin=105,107;
;----------------------------------------------------------------------------
;
; Registers to hold 64-bit values to manipulate. The "L" part
; of the register corresponds to the upper 32-bits, while the "R"
; part corresponds to the lower 32-bits
;
; Note, that when using b6 and b7, the code must save these before
; using them because they are callee save registers
;
;
; Floating point registers to use to save values that
; are manipulated. These don't collide with ftemp1-6 and
; are all caller save registers
;
a0 .reg %fr22
a0L .reg %fr22L
a0R .reg %fr22R
a1 .reg %fr23
a1L .reg %fr23L
a1R .reg %fr23R
a2 .reg %fr24
a2L .reg %fr24L
a2R .reg %fr24R
a3 .reg %fr25
a3L .reg %fr25L
a3R .reg %fr25R
a4 .reg %fr26
a4L .reg %fr26L
a4R .reg %fr26R
a5 .reg %fr27
a5L .reg %fr27L
a5R .reg %fr27R
a6 .reg %fr28
a6L .reg %fr28L
a6R .reg %fr28R
a7 .reg %fr29
a7L .reg %fr29L
a7R .reg %fr29R
b0 .reg %fr30
b0L .reg %fr30L
b0R .reg %fr30R
b1 .reg %fr31
b1L .reg %fr31L
b1R .reg %fr31R
;
; Temporary floating point variables, these are all caller save
; registers
;
ftemp1 .reg %fr4
ftemp2 .reg %fr5
ftemp3 .reg %fr6
ftemp4 .reg %fr7
;
; The B set of registers when used.
;
b2 .reg %fr8
b2L .reg %fr8L
b2R .reg %fr8R
b3 .reg %fr9
b3L .reg %fr9L
b3R .reg %fr9R
b4 .reg %fr10
b4L .reg %fr10L
b4R .reg %fr10R
b5 .reg %fr11
b5L .reg %fr11L
b5R .reg %fr11R
b6 .reg %fr12
b6L .reg %fr12L