aboutsummaryrefslogtreecommitdiffstats
path: root/crypto/aes/asm/aest4-sparcv9.pl
blob: 536f23b47c70bd3f428cec8841ba65fa1d21695f (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
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
#!/usr/bin/env perl

# ====================================================================
# Written by David S. Miller <davem@devemloft.net> and Andy Polyakov
# <appro@openssl.org>. The module is licensed under 2-clause BSD
# license. October 2012. All rights reserved.
# ====================================================================

######################################################################
# AES for SPARC T4.
#
# AES round instructions complete in 3 cycles and can be issued every
# cycle. It means that round calculations should take 4*rounds cycles,
# because any given round instruction depends on result of *both*
# previous instructions:
#
#	|0 |1 |2 |3 |4
#	|01|01|01|
#	   |23|23|23|
#	            |01|01|...
#	               |23|...
#
# Provided that fxor [with IV] takes 3 cycles to complete, critical
# path length for CBC encrypt would be 3+4*rounds, or in other words
# it should process one byte in at least (3+4*rounds)/16 cycles. This
# estimate doesn't account for "collateral" instructions, such as
# fetching input from memory, xor-ing it with zero-round key and
# storing the result. Yet, *measured* performance [for data aligned
# at 64-bit boundary!] deviates from this equation by less than 0.5%:
#
#		128-bit key	192-		256-
# CBC encrypt	2.70/2.90(*)	3.20/3.40	3.70/3.90
#			 (*) numbers after slash are for
#			     misaligned data.
#
# Out-of-order execution logic managed to fully overlap "collateral"
# instructions with those on critical path. Amazing!
#
# As with Intel AES-NI, question is if it's possible to improve
# performance of parallelizeable modes by interleaving round
# instructions. Provided round instruction latency and throughput
# optimal interleave factor is 2. But can we expect 2x performance
# improvement? Well, as round instructions can be issued one per
# cycle, they don't saturate the 2-way issue pipeline and therefore
# there is room for "collateral" calculations... Yet, 2x speed-up
# over CBC encrypt remains unattaintable:
#
#		128-bit key	192-		256-
# CBC decrypt	1.64/2.11	1.89/2.37	2.23/2.61
# CTR		1.64/2.08(*)	1.89/2.33	2.23/2.61
#			 (*) numbers after slash are for
#			     misaligned data.
#
# Estimates based on amount of instructions under assumption that
# round instructions are not pairable with any other instruction
# suggest that latter is the actual case and pipeline runs
# underutilized. It should be noted that T4 out-of-order execution
# logic is so capable that performance gain from 2x interleave is
# not even impressive, ~7-13% over non-interleaved code, largest
# for 256-bit keys.

# To anchor to something else, software implementation processes
# one byte in 29 cycles with 128-bit key on same processor. Intel
# Sandy Bridge encrypts byte in 5.07 cycles in CBC mode and decrypts
# in 0.93, naturally with AES-NI.

$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
push(@INC,"${dir}","${dir}../../perlasm");
require "sparcv9_modes.pl";

&asm_init(@ARGV);

$::evp=1;	# if $evp is set to 0, script generates module with
# AES_[en|de]crypt, AES_set_[en|de]crypt_key and AES_cbc_encrypt entry
# points. These however are not fully compatible with openssl/aes.h,
# because they expect AES_KEY to be aligned at 64-bit boundary. When
# used through EVP, alignment is arranged at EVP layer. Second thing
# that is arranged by EVP is at least 32-bit alignment of IV.

######################################################################
# single-round subroutines
#
{
my ($inp,$out,$key,$rounds,$tmp,$mask)=map("%o$_",(0..5));

$code.=<<___ if ($::abibits==64);
.register	%g2,#scratch
.register	%g3,#scratch

___
$code.=<<___;
.text

.globl	aes_t4_encrypt
.align	32
aes_t4_encrypt:
	andcc		$inp, 7, %g1		! is input aligned?
	andn		$inp, 7, $inp

	ldx		[$key + 0], %g4
	ldx		[$key + 8], %g5

	ldx		[$inp + 0], %o4
	bz,pt		%icc, 1f
	ldx		[$inp + 8], %o5
	ldx		[$inp + 16], $inp
	sll		%g1, 3, %g1
	sub		%g0, %g1, %o3
	sllx		%o4, %g1, %o4
	sllx		%o5, %g1, %g1
	srlx		%o5, %o3, %o5
	srlx		$inp, %o3, %o3
	or		%o5, %o4, %o4
	or		%o3, %g1, %o5
1:
	ld		[$key + 240], $rounds
	ldd		[$key + 16], %f12
	ldd		[$key + 24], %f14
	xor		%g4, %o4, %o4
	xor		%g5, %o5, %o5
	movxtod		%o4, %f0
	movxtod		%o5, %f2
	srl		$rounds, 1, $rounds
	ldd		[$key + 32], %f16
	sub		$rounds, 1, $rounds
	ldd		[$key + 40], %f18
	add		$key, 48, $key

.Lenc:
	aes_eround01	%f12, %f0, %f2, %f4
	aes_eround23	%f14, %f0, %f2, %f2
	ldd		[$key + 0], %f12
	ldd		[$key + 8], %f14
	sub		$rounds,1,$rounds
	aes_eround01	%f16, %f4, %f2, %f0
	aes_eround23	%f18, %f4, %f2, %f2
	ldd		[$key + 16], %f16
	ldd		[$key + 24], %f18
	brnz,pt		$rounds, .Lenc
	add		$key, 32, $key

	andcc		$out, 7, $tmp		! is output aligned?
	aes_eround01	%f12, %f0, %f2, %f4
	aes_eround23	%f14, %f0, %f2, %f2
	aes_eround01_l	%f16, %f4, %f2, %f0
	aes_eround23_l	%f18, %f4, %f2, %f2

	bnz,pn		%icc, 2f
	nop

	std		%f0, [$out + 0]
	retl
	std		%f2, [$out + 8]

2:	alignaddrl	$out, %g0, $out
	mov		0xff, $mask
	srl		$mask, $tmp, $mask

	faligndata	%f0, %f0, %f4
	faligndata	%f0, %f2, %f6
	faligndata	%f2, %f2, %f8

	stda		%f4, [$out + $mask]0xc0	! partial store
	std		%f6, [$out + 8]
	add		$out, 16, $out
	orn		%g0, $mask, $mask
	retl
	stda		%f8, [$out + $mask]0xc0	! partial store
.type	aes_t4_encrypt,#function
.size	aes_t4_encrypt,.-aes_t4_encrypt

.globl	aes_t4_decrypt
.align	32
aes_t4_decrypt:
	andcc		$inp, 7, %g1		! is input aligned?
	andn		$inp, 7, $inp

	ldx		[$key + 0], %g4
	ldx		[$key + 8], %g5

	ldx		[$inp + 0], %o4
	bz,pt		%icc, 1f
	ldx		[$inp + 8], %o5
	ldx		[$inp + 16], $inp
	sll		%g1, 3, %g1
	sub		%g0, %g1, %o3
	sllx		%o4, %g1, %o4
	sllx		%o5, %g1, %g1
	srlx		%o5, %o3, %o5
	srlx		$inp, %o3, %o3
	or		%o5, %o4, %o4
	or		%o3, %g1, %o5
1:
	ld		[$key + 240], $rounds
	ldd		[$key + 16], %f12
	ldd		[$key + 24], %f14
	xor		%g4, %o4, %o4
	xor		%g5, %o5, %o5
	movxtod		%o4, %f0
	movxtod		%o5, %f2
	srl		$rounds, 1, $rounds
	ldd		[$key + 32], %f16
	sub		$rounds, 1, $rounds
	ldd		[$key + 40], %f18
	add		$key, 48, $key

.Ldec:
	aes_dround01	%f12, %f0, %f2, %f4
	aes_dround23	%f14, %f0, %f2, %f2
	ldd		[$key + 0], %f12
	ldd		[$key + 8], %f14
	sub		$rounds,1,$rounds
	aes_dround01	%f16, %f4, %f2, %f0
	aes_dround23	%f18, %f4, %f2, %f2
	ldd		[$key + 16], %f16
	ldd		[$key + 24], %f18
	brnz,pt		$rounds, .Ldec
	add		$key, 32, $key

	andcc		$out, 7, $tmp		! is output aligned?
	aes_dround01	%f12, %f0, %f2, %f4
	aes_dround23	%f14, %f0, %f2, %f2
	aes_dround01_l	%f16, %f4, %f2, %f0
	aes_dround23_l	%f18, %f4, %f2, %f2

	bnz,pn		%icc, 2f
	nop

	std		%f0, [$out + 0]
	retl
	std		%f2, [$out + 8]

2:	alignaddrl	$out, %g0, $out
	mov		0xff, $mask
	srl		$mask, $tmp, $mask

	faligndata	%f0, %f0, %f4
	faligndata	%f0, %f2, %f6
	faligndata	%f2, %f2, %f8

	stda		%f4, [$out + $mask]0xc0	! partial store
	std		%f6, [$out + 8]
	add		$out, 16, $out
	orn		%g0, $mask, $mask
	retl
	stda		%f8, [$out + $mask]0xc0	! partial store
.type	aes_t4_decrypt,#function
.size	aes_t4_decrypt,.-aes_t4_decrypt
___
}

######################################################################
# key setup subroutines
#
{
my ($inp,$bits,$out,$tmp)=map("%o$_",(0..5));
$code.=<<___;
.globl	aes_t4_set_encrypt_key
.align	32
aes_t4_set_encrypt_key:
.Lset_encrypt_key:
	and		$inp, 7, $tmp
	alignaddr	$inp, %g0, $inp
	cmp		$bits, 192
	ldd		[$inp + 0], %f0
	bl,pt		%icc,.L128
	ldd		[$inp + 8], %f2

	be,pt		%icc,.L192
	ldd		[$inp + 16], %f4
	brz,pt		$tmp, .L256aligned
	ldd		[$inp + 24], %f6

	ldd		[$inp + 32], %f8
	faligndata	%f0, %f2, %f0
	faligndata	%f2, %f4, %f2
	faligndata	%f4, %f6, %f4
	faligndata	%f6, %f8, %f6
.L256aligned:
___
for ($i=0; $i<6; $i++) {
    $code.=<<___;
	std		%f0, [$out + `32*$i+0`]
	aes_kexpand1	%f0, %f6, $i, %f0
	std		%f2, [$out + `32*$i+8`]
	aes_kexpand2	%f2, %f0, %f2
	std		%f4, [$out + `32*$i+16`]
	aes_kexpand0	%f4, %f2, %f4
	std		%f6, [$out + `32*$i+24`]
	aes_kexpand2	%f6, %f4, %f6
___
}
$code.=<<___;
	std		%f0, [$out + `32*$i+0`]
	aes_kexpand1	%f0, %f6, $i, %f0
	std		%f2, [$out + `32*$i+8`]
	aes_kexpand2	%f2, %f0, %f2
	std		%f4, [$out + `32*$i+16`]
	std		%f6, [$out + `32*$i+24`]
	std		%f0, [$out + `32*$i+32`]
	std		%f2, [$out + `32*$i+40`]

	mov		14, $tmp
	st		$tmp, [$out + 240]
	retl
	xor		%o0, %o0, %o0

.align	16
.L192:
	brz,pt		$tmp, .L192aligned
	nop

	ldd		[$inp + 24], %f6
	faligndata	%f0, %f2, %f0
	faligndata	%f2, %f4, %f2
	faligndata	%f4, %f6, %f4
.L192aligned:
___
for ($i=0; $i<7; $i++) {
    $code.=<<___;
	std		%f0, [$out + `24*$i+0`]
	aes_kexpand1	%f0, %f4, $i, %f0
	std		%f2, [$out + `24*$i+8`]
	aes_kexpand2	%f2, %f0, %f2
	std		%f4, [$out + `24*$i+16`]
	aes_kexpand2	%f4, %f2, %f4
___
}
$code.=<<___;
	std		%f0, [$out + `24*$i+0`]
	aes_kexpand1	%f0, %f4, $i, %f0
	std		%f2, [$out + `24*$i+8`]
	aes_kexpand2	%f2, %f0, %f2
	std		%f4, [$out + `24*$i+16`]
	std		%f0, [$out + `24*$i+24`]
	std		%f2, [$out + `24*$i+32`]

	mov		12, $tmp
	st		$tmp, [$out + 240]
	retl
	xor		%o0, %o0, %o0

.align	16
.L128:
	brz,pt		$tmp, .L128aligned
	nop

	ldd		[$inp + 16], %f4
	faligndata	%f0, %f2, %f0
	faligndata	%f2, %f4, %f2
.L128aligned:
___
for ($i=0; $i<10; $i++) {
    $code.=<<___;
	std		%f0, [$out + `16*$i+0`]
	aes_kexpand1	%f0, %f2, $i, %f0
	std		%f2, [$out + `16*$i+8`]
	aes_kexpand2	%f2, %f0, %f2
___
}
$code.=<<___;
	std		%f0, [$out + `16*$i+0`]
	std		%f2, [$out + `16*$i+8`]

	mov		10, $tmp
	st		$tmp, [$out + 240]
	retl
	xor		%o0, %o0, %o0
.type	aes_t4_set_encrypt_key,#function
.size	aes_t4_set_encrypt_key,.-aes_t4_set_encrypt_key

.globl	aes_t4_set_decrypt_key
.align	32
aes_t4_set_decrypt_key:
	mov		%o7, %o5
	call		.Lset_encrypt_key
	nop

	mov		%o5, %o7
	sll		$tmp, 4, $inp		! $tmp is number of rounds
	add		$tmp, 2, $tmp
	add		$out, $inp, $inp	! $inp=$out+16*rounds
	srl		$tmp, 2, $tmp		! $tmp=(rounds+2)/4

.Lkey_flip:
	ldd		[$out + 0],  %f0
	ldd		[$out + 8],  %f2
	ldd		[$out + 16], %f4
	ldd		[$out + 24], %f6
	ldd		[$inp + 0],  %f8
	ldd		[$inp + 8],  %f10
	ldd		[$inp - 16], %f12
	ldd		[$inp - 8],  %f14
	sub		$tmp, 1, $tmp
	std		%f0, [$inp + 0]
	std		%f2, [$inp + 8]
	std		%f4, [$inp - 16]
	std		%f6, [$inp - 8]
	std		%f8, [$out + 0]
	std		%f10, [$out + 8]
	std		%f12, [$out + 16]
	std		%f14, [$out + 24]
	add		$out, 32, $out
	brnz		$tmp, .Lkey_flip
	sub		$inp, 32, $inp

	retl
	xor		%o0, %o0, %o0
.type	aes_t4_set_decrypt_key,#function
.size	aes_t4_set_decrypt_key,.-aes_t4_set_decrypt_key
___
}

{{{
my ($inp,$out,$len,$key,$ivec,$enc)=map("%i$_",(0..5));
my ($ileft,$iright,$ooff,$omask,$ivoff)=map("%l$_",(1..7));

$code.=<<___;
.align	32
_aes128_encrypt_1x:
___
for ($i=0; $i<4; $i++) {
    $code.=<<___;
	aes_eround01	%f`16+8*$i+0`, %f0, %f2, %f4
	aes_eround23	%f`16+8*$i+2`, %f0, %f2, %f2
	aes_eround01	%f`16+8*$i+4`, %f4, %f2, %f0
	aes_eround23	%f`16+8*$i+6`, %f4, %f2, %f2
___
}
$code.=<<___;
	aes_eround01	%f48, %f0, %f2, %f4
	aes_eround23	%f50, %f0, %f2, %f2
	aes_eround01_l	%f52, %f4, %f2, %f0
	retl
	aes_eround23_l	%f54, %f4, %f2, %f2
.type	_aes128_encrypt_1x,#function
.size	_aes128_encrypt_1x,.-_aes128_encrypt_1x

.align	32
_aes128_encrypt_2x:
___
for ($i=0; $i<4; $i++) {
    $code.=<<___;
	aes_eround01	%f`16+8*$i+0`, %f0, %f2, %f8
	aes_eround23	%f`16+8*$i+2`, %f0, %f2, %f2
	aes_eround01	%f`16+8*$i+0`, %f4, %f6, %f10
	aes_eround23	%f`16+8*$i+2`, %f4, %f6, %f6
	aes_eround01	%f`16+8*$i+4`, %f8, %f2, %f0
	aes_eround23	%f`16+8*$i+6`, %f8, %f2, %f2
	aes_eround01	%f`16+8*$i+4`, %f10, %f6, %f4
	aes_eround23	%f`16+8*$i+6`, %f10, %f6, %f6
___
}
$code.=<<___;
	aes_eround01	%f48, %f0, %f2, %f8
	aes_eround23	%f50, %f0, %f2, %f2
	aes_eround01	%f48, %f4, %f6, %f10
	aes_eround23	%f50, %f4, %f6, %f6
	aes_eround01_l	%f52, %f8, %f2, %f0
	aes_eround23_l	%f54, %f8, %f2, %f2
	aes_eround01_l	%f52, %f10, %f6, %f4
	retl
	aes_eround23_l	%f54, %f10, %f6, %f6
.type	_aes128_encrypt_2x,#function
.size	_aes128_encrypt_2x,.-_aes128_encrypt_2x

.align	32
_aes128_loadkey:
	ldx		[$key + 0], %g4
	ldx		[$key + 8], %g5
___
for ($i=2; $i<22;$i++) {			# load key schedule
    $code.=<<___;
	ldd		[$key + `8*$i`], %f`12+2*$i`
___
}
$code.=<<___;
	retl
	nop
.type	_aes128_loadkey,#function
.size	_aes128_loadkey,.-_aes128_loadkey
_aes128_load_enckey=_aes128_loadkey
_aes128_load_deckey=_aes128_loadkey

___

&alg_cbc_encrypt_implement("aes",128);
if ($::evp) {
    &alg_ctr32_implement("aes",128);
    &alg_xts_implement("aes",128,"en");
    &alg_xts_implement("aes",128,"de");
}
&alg_cbc_decrypt_implement("aes",128);

$code.=<<___;
.align	32
_aes128_decrypt_1x:
___
for ($i=0; $i<4; $i++) {
    $code.=<<___;
	aes_dround01	%f`16+8*$i+0`, %f0, %f2, %f4
	aes_dround23	%f`16+8*$i+2`, %f0, %f2, %f2
	aes_dround01	%f`16+8*$i+4`, %f4, %f2, %f0
	aes_dround23	%f`16+8*$i+6`, %f4, %f2, %f2
___
}
$code.=<<___;
	aes_dround01	%f48, %f0, %f2, %f4
	aes_dround23	%f50, %f0, %f2, %f2
	aes_dround01_l	%f52, %f4, %f2, %f0
	retl
	aes_dround23_l	%f54, %f4, %f2, %f2
.type	_aes128_decrypt_1x,#function
.size	_aes128_decrypt_1x,.-_aes128_decrypt_1x

.align	32
_aes128_decrypt_2x:
___
for ($i=0; $i<4; $i++) {
    $code.=<<___;
	aes_dround01	%f`16+8*$i+0`, %f0, %f2, %f8
	aes_dround23	%f`16+8*$i+2`, %f0, %f2, %f2
	aes_dround01	%f`16+8*$i+0`, %f4, %f6, %f10
	aes_dround23	%f`16+8*$i+2`, %f4, %f6, %f6
	aes_dround01	%f`16+8*$i+4`, %f8, %f2, %f0
	aes_dround23	%f`16+8*$i+6`, %f8, %f2, %f2
	aes_dround01	%f`16+8*$i+4`, %f10, %f6, %f4
	aes_dround23	%f`16+8*$i+6`, %f10, %f6, %f6
___
}
$code.=<<___;
	aes_dround01	%f48, %f0, %f2, %f8
	aes_dround23	%f50, %f0, %f2, %f2
	aes_dround01	%f48, %f4, %f6, %f10
	aes_dround23	%f50, %f4, %f6, %f6
	aes_dround01_l	%f52, %f8, %f2, %f0
	aes_dround23_l	%f54, %f8, %f2, %f2
	aes_dround01_l	%f52, %f10, %f6, %f4
	retl
	aes_dround23_l	%f54, %f10, %f6, %f6
.type	_aes128_decrypt_2x,#function
.size	_aes128_decrypt_2x,.-_aes128_decrypt_2x
___

$code.=<<___;
.align	32
_aes192_encrypt_1x:
___
for ($i=0; $i<5; $i++) {
    $code.=<<___;
	aes_eround01	%f`16+8*$i+0`, %f0, %f2, %f4
	aes_eround23	%f`16+8*$i+2`, %f0, %f2, %f2
	aes_eround01	%f`16+8*$i+4`, %f4, %f2, %f0
	aes_eround23	%f`16+8*$i+6`, %f4, %f2, %f2
___
}
$code.=<<___;
	aes_eround01	%f56, %f0, %f2, %f4
	aes_eround23	%f58, %f0, %f2, %f2
	aes_eround01_l	%f60, %f4, %f2, %f0
	retl
	aes_eround23_l	%f62, %f4, %f2, %f2
.type	_aes192_encrypt_1x,#function
.size	_aes192_encrypt_1x,.-_aes192_encrypt_1x

.align	32
_aes192_encrypt_2x:
___
for ($i=0; $i<5; $i++) {
    $code.=<<___;
	aes_eround01	%f`16+8*$i+0`, %f0, %f2, %f8
	aes_eround23	%f`16+8*$i+2`, %f0, %f2, %f2
	aes_eround01	%f`16+8*$i+0`, %f4, %f6, %f10
	aes_eround23	%f`16+8*$i+2`, %f4, %f6, %f6
	aes_eround01	%f`16+8*$i+4`, %f8, %f2, %f0
	aes_eround23	%f`16+8*$i+6`, %f8, %f2, %f2
	aes_eround01	%f`16+8*$i+4`, %f10, %f6, %f4
	aes_eround23	%f`16+8*$i+6`, %f10, %f6, %f6
___
}
$code.=<<___;
	aes_eround01	%f56, %f0, %f2, %f8
	aes_eround23	%f58, %f0, %f2, %f2
	aes_eround01	%f56, %f4, %f6, %f10
	aes_eround23	%f58, %f4, %f6, %f6
	aes_eround01_l	%f60, %f8, %f2, %f0
	aes_eround23_l	%f62, %f8, %f2, %f2
	aes_eround01_l	%f60, %f10, %f6, %f4
	retl
	aes_eround23_l	%f62, %f10, %f6, %f6
.type	_aes192_encrypt_2x,#function
.size	_aes192_encrypt_2x,.-_aes192_encrypt_2x

.align	32
_aes256_encrypt_1x:
	aes_eround01	%f16, %f0, %f2, %f4
	aes_eround23	%f18, %f0, %f2, %f2
	ldd		[$key + 208], %f16
	ldd		[$key + 216], %f18
	aes_eround01	%f20, %f4, %f2, %f0
	aes_eround23	%f22, %f4, %f2, %f2
	ldd		[$key + 224], %f20
	ldd		[$key + 232], %f22
___
for ($i=1; $i<6; $i++) {
    $code.=<<___;
	aes_eround01	%f`16+8*$i+0`, %f0, %f2, %f4
	aes_eround23	%f`16+8*$i+2`, %f0, %f2, %f2
	aes_eround01	%f`16+8*$i+4`, %f4, %f2, %f0
	aes_eround23	%f`16+8*$i+6`, %f4, %f2, %f2
___
}
$code.=<<___;
	aes_eround01	%f16, %f0, %f2, %f4
	aes_eround23	%f18, %f0, %f2, %f2
	ldd		[$key + 16], %f16
	ldd		[$key + 24], %f18
	aes_eround01_l	%f20, %f4, %f2, %f0
	aes_eround23_l	%f22, %f4, %f2, %f2
	ldd		[$key + 32], %f20
	retl
	ldd		[$key + 40], %f22
.type	_aes256_encrypt_1x,#function
.size	_aes256_encrypt_1x,.-_aes256_encrypt_1x

.align	32
_aes256_encrypt_2x:
	aes_eround01	%f16, %f0, %f2, %f8
	aes_eround23	%f18, %f0, %f2, %f2
	aes_eround01	%f16, %f4, %f6, %f10
	aes_eround23	%f18, %f4, %f6, %f6
	ldd		[$key + 208], %f16
	ldd		[$key + 216], %f18
	aes_eround01	%f20, %f8, %f2, %f0
	aes_eround23	%f22, %f8, %f2, %f2
	aes_eround01	%f20, %f10, %f6, %f4
	aes_eround23	%f22, %f10, %f6, %f6
	ldd		[$key + 224], %f20
	ldd		[$key + 232], %f22
___
for ($i=1; $i<6; $i++) {
    $code.=<<___;
	aes_eround01	%f`16+8*$i+0`, %f0, %f2, %f8
	aes_eround23	%f`16+8*$i+2`, %f0, %f2, %f2
	aes_eround01	%f`16+8*$i+0`, %f4, %f6, %f10
	aes_eround23	%f`16+8*$i+2`, %f4, %f6, %f6
	aes_eround01	%f`16+8*$i+4`, %f8, %f2, %f0
	aes_eround23	%f`16+8*$i+6`, %f8, %f2, %f2
	aes_eround01	%f`16+8*$i+4`, %f10, %f6, %f4
	aes_eround23	%f`16+8*$i+6`, %f10, %f6, %f6
___
}
$code.=<<___;
	aes_eround01	%f16, %f0, %f2, %f8
	aes_eround23	%f18, %f0, %f2, %f2
	aes_eround01	%f16, %f4, %f6, %f10
	aes_eround23	%f18, %f4, %f6, %f6
	ldd		[$key + 16], %f16
	ldd		[$key + 24], %f18
	aes_eround01_l	%f20, %f8, %f2, %f0
	aes_eround23_l	%f22, %f8, %f2, %f2
	aes_eround01_l	%f20, %f10, %f6, %f4
	aes_eround23_l	%f22, %f10, %f6, %f6
	ldd		[$key + 32], %f20
	retl
	ldd		[$key + 40], %f22
.type	_aes256_encrypt_2x,#function
.size	_aes256_encrypt_2x,.-_aes256_encrypt_2x

.align	32
_aes192_loadkey:
	ldx		[$key + 0], %g4
	ldx		[$key + 8], %g5
___
for ($i=2; $i<26;$i++) {			# load key schedule
    $code.=<<___;
	ldd		[$key + `8*$i`], %f`12+2*$i`
___
}
$code.=<<___;
	retl
	nop
.type	_aes192_loadkey,#function
.size	_aes192_loadkey,.-_aes192_loadkey
_aes256_loadkey=_aes192_loadkey
_aes192_load_enckey=_aes192_loadkey
_aes192_load_deckey=_aes192_loadkey
_aes256_load_enckey=_aes192_loadkey
_aes256_load_deckey=_aes192_loadkey
___

&alg_cbc_encrypt_implement("aes",256);
&alg_cbc_encrypt_implement("aes",192);
if ($::evp) {
    &alg_ctr32_implement("aes",256);
    &alg_xts_implement("aes",256,"en");
    &alg_xts_implement("aes",256,"de");
    &alg_ctr32_implement("aes",192);
}
&alg_cbc_decrypt_implement("aes",192);
&alg_cbc_decrypt_implement("aes",256);

$code.=<<___;
.align	32
_aes256_decrypt_1x:
	aes_dround01	%f16, %f0, %f2, %f4
	aes_dround23	%f18, %f0, %f2, %f2
	ldd		[$key + 208], %f16
	ldd		[$key + 216], %f18
	aes_dround01	%f20, %f4, %f2, %f0
	aes_dround23	%f22, %f4, %f2, %f2
	ldd		[$key + 224], %f20
	ldd		[$key + 232], %f22
___
for ($i=1; $i<6; $i++) {
    $code.=<<___;
	aes_dround01	%f`16+8*$i+0`, %f0, %f2, %f4
	aes_dround23	%f`16+8*$i+2`, %f0, %f2, %f2
	aes_dround01	%f`16+8*$i+4`, %f4, %f2, %f0
	aes_dround23	%f`16+8*$i+6`, %f4, %f2, %f2
___
}
$code.=<<___;
	aes_dround01	%f16, %f0, %f2, %f4
	aes_dround23	%f18, %f0, %f2, %f2
	ldd		[$key + 16], %f16
	ldd		[$key + 24], %f18
	aes_dround01_l	%f20, %f4, %f2, %f0
	aes_dround23_l	%f22, %f4, %f2, %f2
	ldd		[$key + 32], %f20
	retl
	ldd		[$key + 40], %f22
.type	_aes256_decrypt_1x,#function
.size	_aes256_decrypt_1x,.-_aes256_decrypt_1x

.align	32
_aes256_decrypt_2x:
	aes_dround01	%f16, %f0, %f2, %f8
	aes_dround23	%f18, %f0, %f2, %f2
	aes_dround01	%f16, %f4, %f6, %f10
	aes_dround23	%f18, %f4, %f6, %f6
	ldd		[$key + 208], %f16
	ldd		[$key + 216], %f18
	aes_dround01	%f20, %f8, %f2, %f0
	aes_dround23	%f22, %f8, %f2, %f2
	aes_dround01	%f20, %f10, %f6, %f4
	aes_dround23	%f22, %f10, %f6, %f6
	ldd		[$key + 224], %f20
	ldd		[$key + 232], %f22
___
for ($i=1; $i<6; $i++) {
    $code.=<<___;
	aes_dround01	%f`16+8*$i+0`, %f0, %f2, %f8
	aes_dround23	%f`16+8*$i+2`, %f0, %f2, %f2
	aes_dround01	%f`16+8*$i+0`, %f4, %f6, %f10
	aes_dround23	%f`16+8*$i+2`, %f4, %f6, %f6
	aes_dround01	%f`16+8*$i+4`, %f8, %f2, %f0
	aes_dround23	%f`16+8*$i+6`, %f8, %f2, %f2
	aes_dround01	%f`16+8*$i+4`, %f10, %f6, %f4
	aes_dround23	%f`16+8*$i+6`, %f10, %f6, %f6
___
}
$code.=<<___;
	aes_dround01	%f16, %f0, %f2, %f8
	aes_dround23	%f18, %f0, %f2, %f2
	aes_dround01	%f16, %f4, %f6, %f10
	aes_dround23	%f18, %f4, %f6, %f6
	ldd		[$key + 16], %f16
	ldd		[$key + 24], %f18
	aes_dround01_l	%f20, %f8, %f2, %f0
	aes_dround23_l	%f22, %f8, %f2, %f2
	aes_dround01_l	%f20, %f10, %f6, %f4
	aes_dround23_l	%f22, %f10, %f6, %f6
	ldd		[$key + 32], %f20
	retl
	ldd		[$key + 40], %f22
.type	_aes256_decrypt_2x,#function
.size	_aes256_decrypt_2x,.-_aes256_decrypt_2x

.align	32
_aes192_decrypt_1x:
___
for ($i=0; $i<5; $i++) {
    $code.=<<___;
	aes_dround01	%f`16+8*$i+0`, %f0, %f2, %f4
	aes_dround23	%f`16+8*$i+2`, %f0, %f2, %f2
	aes_dround01	%f`16+8*$i+4`, %f4, %f2, %f0
	aes_dround23	%f`16+8*$i+6`, %f4, %f2, %f2
___
}
$code.=<<___;
	aes_dround01	%f56, %f0, %f2, %f4
	aes_dround23	%f58, %f0, %f2, %f2
	aes_dround01_l	%f60, %f4, %f2, %f0
	retl
	aes_dround23_l	%f62, %f4, %f2, %f2
.type	_aes192_decrypt_1x,#function
.size	_aes192_decrypt_1x,.-_aes192_decrypt_1x

.align	32
_aes192_decrypt_2x:
___
for ($i=0; $i<5; $i++) {
    $code.=<<___;
	aes_dround01	%f`16+8*$i+0`, %f0, %f2, %f8
	aes_dround23	%f`16+8*$i+2`, %f0, %f2, %f2
	aes_dround01	%f`16+8*$i+0`, %f4, %f6, %f10
	aes_dround23	%f`16+8*$i+2`, %f4, %f6, %f6
	aes_dround01	%f`16+8*$i+4`, %f8, %f2, %f0
	aes_dround23	%f`16+8*$i+6`, %f8, %f2, %f2
	aes_dround01	%f`16+8*$i+4`, %f10, %f6, %f4
	aes_dround23	%f`16+8*$i+6`, %f10, %f6, %f6
___
}
$code.=<<___;
	aes_dround01	%f56, %f0, %f2, %f8
	aes_dround23	%f58, %f0, %f2, %f2
	aes_dround01	%f56, %f4, %f6, %f10
	aes_dround23	%f58, %f4, %f6, %f6
	aes_dround01_l	%f60, %f8, %f2, %f0
	aes_dround23_l	%f62, %f8, %f2, %f2
	aes_dround01_l	%f60, %f10, %f6, %f4
	retl
	aes_dround23_l	%f62, %f10, %f6, %f6
.type	_aes192_decrypt_2x,#function
.size	_aes192_decrypt_2x,.-_aes192_decrypt_2x
___
}}}

if (!$::evp) {
$code.=<<___;
.global	AES_encrypt
AES_encrypt=aes_t4_encrypt
.global	AES_decrypt
AES_decrypt=aes_t4_decrypt
.global	AES_set_encrypt_key
.align	32
AES_set_encrypt_key:
	andcc		%o2, 7, %g0		! check alignment
	bnz,a,pn	%icc, 1f
	mov		-1, %o0
	brz,a,pn	%o0, 1f
	mov		-1, %o0
	brz,a,pn	%o2, 1f
	mov		-1, %o0
	andncc		%o1, 0x1c0, %g0
	bnz,a,pn	%icc, 1f
	mov		-2, %o0
	cmp		%o1, 128
	bl,a,pn		%icc, 1f
	mov		-2, %o0
	b		aes_t4_set_encrypt_key
	nop
1:	retl
	nop
.type	AES_set_encrypt_key,#function
.size	AES_set_encrypt_key,.-AES_set_encrypt_key

.global	AES_set_decrypt_key
.align	32
AES_set_decrypt_key:
	andcc		%o2, 7, %g0		! check alignment
	bnz,a,pn	%icc, 1f
	mov		-1, %o0
	brz,a,pn	%o0, 1f
	mov		-1, %o0
	brz,a,pn	%o2, 1f
	mov		-1, %o0
	andncc		%o1, 0x1c0, %g0
	bnz,a,pn	%icc, 1f
	mov		-2, %o0
	cmp		%o1, 128
	bl,a,pn		%icc, 1f
	mov		-2, %o0
	b		aes_t4_set_decrypt_key
	nop
1:	retl
	nop
.type	AES_set_decrypt_key,#function
.size	AES_set_decrypt_key,.-AES_set_decrypt_key
___

my ($inp,$out,$len,$key,$ivec,$enc)=map("%o$_",(0..5));

$code.=<<___;
.globl	AES_cbc_encrypt
.align	32
AES_cbc_encrypt:
	ld		[$key + 240], %g1
	nop
	brz		$enc, .Lcbc_decrypt
	cmp		%g1, 12

	bl,pt		%icc, aes128_t4_cbc_encrypt
	nop
	be,pn		%icc, aes192_t4_cbc_encrypt
	nop
	ba		aes256_t4_cbc_encrypt
	nop

.Lcbc_decrypt:
	bl,pt		%icc, aes128_t4_cbc_decrypt
	nop
	be,pn		%icc, aes192_t4_cbc_decrypt
	nop
	ba		aes256_t4_cbc_decrypt
	nop
.type	AES_cbc_encrypt,#function
.size	AES_cbc_encrypt,.-AES_cbc_encrypt
___
}
$code.=<<___;
.asciz	"AES for SPARC T4, David S. Miller, Andy Polyakov"
.align	4
___

&emit_assembler();

close STDOUT;