summaryrefslogtreecommitdiff
path: root/gcc/config/xtensa/lib1funcs.asm
blob: 071b9171177c81e7638acc3eb4d1dee075b2386a (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
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
/* Assembly functions for the Xtensa version of libgcc1.
   Copyright (C) 2001, 2002, 2003, 2005, 2006, 2007, 2009
   Free Software Foundation, Inc.
   Contributed by Bob Wilson (bwilson@tensilica.com) at Tensilica.

This file is part of GCC.

GCC is free software; you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free
Software Foundation; either version 3, or (at your option) any later
version.

GCC is distributed in the hope that it will be useful, but WITHOUT ANY
WARRANTY; without even the implied warranty of MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
for more details.

Under Section 7 of GPL version 3, you are granted additional
permissions described in the GCC Runtime Library Exception, version
3.1, as published by the Free Software Foundation.

You should have received a copy of the GNU General Public License and
a copy of the GCC Runtime Library Exception along with this program;
see the files COPYING3 and COPYING.RUNTIME respectively.  If not, see
<http://www.gnu.org/licenses/>.  */

#include "xtensa-config.h"

/* Define macros for the ABS and ADDX* instructions to handle cases
   where they are not included in the Xtensa processor configuration.  */

	.macro	do_abs dst, src, tmp
#if XCHAL_HAVE_ABS
	abs	\dst, \src
#else
	neg	\tmp, \src
	movgez	\tmp, \src, \src
	mov	\dst, \tmp
#endif
	.endm

	.macro	do_addx2 dst, as, at, tmp
#if XCHAL_HAVE_ADDX
	addx2	\dst, \as, \at
#else
	slli	\tmp, \as, 1
	add	\dst, \tmp, \at
#endif
	.endm

	.macro	do_addx4 dst, as, at, tmp
#if XCHAL_HAVE_ADDX
	addx4	\dst, \as, \at
#else
	slli	\tmp, \as, 2
	add	\dst, \tmp, \at
#endif
	.endm

	.macro	do_addx8 dst, as, at, tmp
#if XCHAL_HAVE_ADDX
	addx8	\dst, \as, \at
#else
	slli	\tmp, \as, 3
	add	\dst, \tmp, \at
#endif
	.endm

/* Define macros for leaf function entry and return, supporting either the
   standard register windowed ABI or the non-windowed call0 ABI.  These
   macros do not allocate any extra stack space, so they only work for
   leaf functions that do not need to spill anything to the stack.  */

	.macro leaf_entry reg, size
#if XCHAL_HAVE_WINDOWED && !__XTENSA_CALL0_ABI__
	entry \reg, \size
#else
	/* do nothing */
#endif
	.endm

	.macro leaf_return
#if XCHAL_HAVE_WINDOWED && !__XTENSA_CALL0_ABI__
	retw
#else
	ret
#endif
	.endm


#ifdef L_mulsi3
	.align	4
	.global	__mulsi3
	.type	__mulsi3, @function
__mulsi3:
	leaf_entry sp, 16

#if XCHAL_HAVE_MUL32
	mull	a2, a2, a3

#elif XCHAL_HAVE_MUL16
	or	a4, a2, a3
	srai	a4, a4, 16
	bnez	a4, .LMUL16
	mul16u	a2, a2, a3
	leaf_return
.LMUL16:
	srai	a4, a2, 16
	srai	a5, a3, 16
	mul16u	a7, a4, a3
	mul16u	a6, a5, a2
	mul16u	a4, a2, a3
	add	a7, a7, a6
	slli	a7, a7, 16
	add	a2, a7, a4

#elif XCHAL_HAVE_MAC16
	mul.aa.hl a2, a3
	mula.aa.lh a2, a3
	rsr	a5, ACCLO
	umul.aa.ll a2, a3
	rsr	a4, ACCLO
	slli	a5, a5, 16
	add	a2, a4, a5

#else /* !MUL32 && !MUL16 && !MAC16 */

	/* Multiply one bit at a time, but unroll the loop 4x to better
	   exploit the addx instructions and avoid overhead.
	   Peel the first iteration to save a cycle on init.  */

	/* Avoid negative numbers.  */
	xor	a5, a2, a3	/* Top bit is 1 if one input is negative.  */
	do_abs	a3, a3, a6
	do_abs	a2, a2, a6

	/* Swap so the second argument is smaller.  */
	sub	a7, a2, a3
	mov	a4, a3
	movgez	a4, a2, a7	/* a4 = max (a2, a3) */
	movltz	a3, a2, a7	/* a3 = min (a2, a3) */

	movi	a2, 0
	extui	a6, a3, 0, 1
	movnez	a2, a4, a6

	do_addx2 a7, a4, a2, a7
	extui	a6, a3, 1, 1
	movnez	a2, a7, a6

	do_addx4 a7, a4, a2, a7
	extui	a6, a3, 2, 1
	movnez	a2, a7, a6

	do_addx8 a7, a4, a2, a7
	extui	a6, a3, 3, 1
	movnez	a2, a7, a6

	bgeui	a3, 16, .Lmult_main_loop
	neg	a3, a2
	movltz	a2, a3, a5
	leaf_return

	.align	4
.Lmult_main_loop:
	srli	a3, a3, 4
	slli	a4, a4, 4

	add	a7, a4, a2
	extui	a6, a3, 0, 1
	movnez	a2, a7, a6

	do_addx2 a7, a4, a2, a7
	extui	a6, a3, 1, 1
	movnez	a2, a7, a6

	do_addx4 a7, a4, a2, a7
	extui	a6, a3, 2, 1
	movnez	a2, a7, a6

	do_addx8 a7, a4, a2, a7
	extui	a6, a3, 3, 1
	movnez	a2, a7, a6

	bgeui	a3, 16, .Lmult_main_loop

	neg	a3, a2
	movltz	a2, a3, a5

#endif /* !MUL32 && !MUL16 && !MAC16 */

	leaf_return
	.size	__mulsi3, . - __mulsi3

#endif /* L_mulsi3 */


#ifdef L_umulsidi3

#if !XCHAL_HAVE_MUL16 && !XCHAL_HAVE_MUL32 && !XCHAL_HAVE_MAC16
#define XCHAL_NO_MUL 1
#endif

	.align	4
	.global	__umulsidi3
	.type	__umulsidi3, @function
__umulsidi3:
#if __XTENSA_CALL0_ABI__
	leaf_entry sp, 32
	addi	sp, sp, -32
	s32i	a12, sp, 16
	s32i	a13, sp, 20
	s32i	a14, sp, 24
	s32i	a15, sp, 28
#elif XCHAL_NO_MUL
	/* This is not really a leaf function; allocate enough stack space
	   to allow CALL12s to a helper function.  */
	leaf_entry sp, 48
#else
	leaf_entry sp, 16
#endif

#ifdef __XTENSA_EB__
#define wh a2
#define wl a3
#else
#define wh a3
#define wl a2
#endif /* __XTENSA_EB__ */

	/* This code is taken from the mulsf3 routine in ieee754-sf.S.
	   See more comments there.  */

#if XCHAL_HAVE_MUL32_HIGH
	mull	a6, a2, a3
	muluh	wh, a2, a3
	mov	wl, a6

#else /* ! MUL32_HIGH */

#if __XTENSA_CALL0_ABI__ && XCHAL_NO_MUL
	/* a0 and a8 will be clobbered by calling the multiply function
	   but a8 is not used here and need not be saved.  */
	s32i	a0, sp, 0
#endif

#if XCHAL_HAVE_MUL16 || XCHAL_HAVE_MUL32

#define a2h a4
#define a3h a5

	/* Get the high halves of the inputs into registers.  */
	srli	a2h, a2, 16
	srli	a3h, a3, 16

#define a2l a2
#define a3l a3

#if XCHAL_HAVE_MUL32 && !XCHAL_HAVE_MUL16
	/* Clear the high halves of the inputs.  This does not matter
	   for MUL16 because the high bits are ignored.  */
	extui	a2, a2, 0, 16
	extui	a3, a3, 0, 16
#endif
#endif /* MUL16 || MUL32 */


#if XCHAL_HAVE_MUL16

#define do_mul(dst, xreg, xhalf, yreg, yhalf) \
	mul16u	dst, xreg ## xhalf, yreg ## yhalf

#elif XCHAL_HAVE_MUL32

#define do_mul(dst, xreg, xhalf, yreg, yhalf) \
	mull	dst, xreg ## xhalf, yreg ## yhalf

#elif XCHAL_HAVE_MAC16

/* The preprocessor insists on inserting a space when concatenating after
   a period in the definition of do_mul below.  These macros are a workaround
   using underscores instead of periods when doing the concatenation.  */
#define umul_aa_ll umul.aa.ll
#define umul_aa_lh umul.aa.lh
#define umul_aa_hl umul.aa.hl
#define umul_aa_hh umul.aa.hh

#define do_mul(dst, xreg, xhalf, yreg, yhalf) \
	umul_aa_ ## xhalf ## yhalf	xreg, yreg; \
	rsr	dst, ACCLO

#else /* no multiply hardware */

#define set_arg_l(dst, src) \
	extui	dst, src, 0, 16
#define set_arg_h(dst, src) \
	srli	dst, src, 16

#if __XTENSA_CALL0_ABI__
#define do_mul(dst, xreg, xhalf, yreg, yhalf) \
	set_arg_ ## xhalf (a13, xreg); \
	set_arg_ ## yhalf (a14, yreg); \
	call0	.Lmul_mulsi3; \
	mov	dst, a12
#else
#define do_mul(dst, xreg, xhalf, yreg, yhalf) \
	set_arg_ ## xhalf (a14, xreg); \
	set_arg_ ## yhalf (a15, yreg); \
	call12	.Lmul_mulsi3; \
	mov	dst, a14
#endif /* __XTENSA_CALL0_ABI__ */

#endif /* no multiply hardware */

	/* Add pp1 and pp2 into a6 with carry-out in a9.  */
	do_mul(a6, a2, l, a3, h)	/* pp 1 */
	do_mul(a11, a2, h, a3, l)	/* pp 2 */
	movi	a9, 0
	add	a6, a6, a11
	bgeu	a6, a11, 1f
	addi	a9, a9, 1
1:
	/* Shift the high half of a9/a6 into position in a9.  Note that
	   this value can be safely incremented without any carry-outs.  */
	ssai	16
	src	a9, a9, a6

	/* Compute the low word into a6.  */
	do_mul(a11, a2, l, a3, l)	/* pp 0 */
	sll	a6, a6
	add	a6, a6, a11
	bgeu	a6, a11, 1f
	addi	a9, a9, 1
1:
	/* Compute the high word into wh.  */
	do_mul(wh, a2, h, a3, h)	/* pp 3 */
	add	wh, wh, a9
	mov	wl, a6

#endif /* !MUL32_HIGH */

#if __XTENSA_CALL0_ABI__ && XCHAL_NO_MUL
	/* Restore the original return address.  */
	l32i	a0, sp, 0
#endif
#if __XTENSA_CALL0_ABI__
	l32i	a12, sp, 16
	l32i	a13, sp, 20
	l32i	a14, sp, 24
	l32i	a15, sp, 28
	addi	sp, sp, 32
#endif
	leaf_return

#if XCHAL_NO_MUL

	/* For Xtensa processors with no multiply hardware, this simplified
	   version of _mulsi3 is used for multiplying 16-bit chunks of
	   the floating-point mantissas.  When using CALL0, this function
	   uses a custom ABI: the inputs are passed in a13 and a14, the
	   result is returned in a12, and a8 and a15 are clobbered.  */
	.align	4
.Lmul_mulsi3:
	leaf_entry sp, 16
	.macro mul_mulsi3_body dst, src1, src2, tmp1, tmp2
	movi	\dst, 0
1:	add	\tmp1, \src2, \dst
	extui	\tmp2, \src1, 0, 1
	movnez	\dst, \tmp1, \tmp2

	do_addx2 \tmp1, \src2, \dst, \tmp1
	extui	\tmp2, \src1, 1, 1
	movnez	\dst, \tmp1, \tmp2

	do_addx4 \tmp1, \src2, \dst, \tmp1
	extui	\tmp2, \src1, 2, 1
	movnez	\dst, \tmp1, \tmp2

	do_addx8 \tmp1, \src2, \dst, \tmp1
	extui	\tmp2, \src1, 3, 1
	movnez	\dst, \tmp1, \tmp2

	srli	\src1, \src1, 4
	slli	\src2, \src2, 4
	bnez	\src1, 1b
	.endm
#if __XTENSA_CALL0_ABI__
	mul_mulsi3_body a12, a13, a14, a15, a8
#else
	/* The result will be written into a2, so save that argument in a4.  */
	mov	a4, a2
	mul_mulsi3_body a2, a4, a3, a5, a6
#endif
	leaf_return
#endif /* XCHAL_NO_MUL */

	.size	__umulsidi3, . - __umulsidi3

#endif /* L_umulsidi3 */


/* Define a macro for the NSAU (unsigned normalize shift amount)
   instruction, which computes the number of leading zero bits,
   to handle cases where it is not included in the Xtensa processor
   configuration.  */

	.macro	do_nsau cnt, val, tmp, a
#if XCHAL_HAVE_NSA
	nsau	\cnt, \val
#else
	mov	\a, \val
	movi	\cnt, 0
	extui	\tmp, \a, 16, 16
	bnez	\tmp, 0f
	movi	\cnt, 16
	slli	\a, \a, 16
0:
	extui	\tmp, \a, 24, 8
	bnez	\tmp, 1f
	addi	\cnt, \cnt, 8
	slli	\a, \a, 8
1:
	movi	\tmp, __nsau_data
	extui	\a, \a, 24, 8
	add	\tmp, \tmp, \a
	l8ui	\tmp, \tmp, 0
	add	\cnt, \cnt, \tmp
#endif /* !XCHAL_HAVE_NSA */
	.endm

#ifdef L_clz
	.section .rodata
	.align	4
	.global	__nsau_data
	.type	__nsau_data, @object
__nsau_data:
#if !XCHAL_HAVE_NSA
	.byte	8, 7, 6, 6, 5, 5, 5, 5, 4, 4, 4, 4, 4, 4, 4, 4
	.byte	3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3
	.byte	2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2
	.byte	2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2
	.byte	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1
	.byte	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1
	.byte	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1
	.byte	1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1
	.byte	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
	.byte	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
	.byte	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
	.byte	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
	.byte	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
	.byte	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
	.byte	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
	.byte	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
#endif /* !XCHAL_HAVE_NSA */
	.size	__nsau_data, . - __nsau_data
	.hidden	__nsau_data
#endif /* L_clz */


#ifdef L_clzsi2
	.align	4
	.global	__clzsi2
	.type	__clzsi2, @function
__clzsi2:
	leaf_entry sp, 16
	do_nsau	a2, a2, a3, a4
	leaf_return
	.size	__clzsi2, . - __clzsi2

#endif /* L_clzsi2 */


#ifdef L_ctzsi2
	.align	4
	.global	__ctzsi2
	.type	__ctzsi2, @function
__ctzsi2:
	leaf_entry sp, 16
	neg	a3, a2
	and	a3, a3, a2
	do_nsau	a2, a3, a4, a5
	neg	a2, a2
	addi	a2, a2, 31
	leaf_return
	.size	__ctzsi2, . - __ctzsi2

#endif /* L_ctzsi2 */


#ifdef L_ffssi2
	.align	4
	.global	__ffssi2
	.type	__ffssi2, @function
__ffssi2:
	leaf_entry sp, 16
	neg	a3, a2
	and	a3, a3, a2
	do_nsau	a2, a3, a4, a5
	neg	a2, a2
	addi	a2, a2, 32
	leaf_return
	.size	__ffssi2, . - __ffssi2

#endif /* L_ffssi2 */


#ifdef L_udivsi3
	.align	4
	.global	__udivsi3
	.type	__udivsi3, @function
__udivsi3:
	leaf_entry sp, 16
#if XCHAL_HAVE_DIV32
	quou	a2, a2, a3
#else
	bltui	a3, 2, .Lle_one	/* check if the divisor <= 1 */

	mov	a6, a2		/* keep dividend in a6 */
	do_nsau	a5, a6, a2, a7	/* dividend_shift = nsau (dividend) */
	do_nsau	a4, a3, a2, a7	/* divisor_shift = nsau (divisor) */
	bgeu	a5, a4, .Lspecial

	sub	a4, a4, a5	/* count = divisor_shift - dividend_shift */
	ssl	a4
	sll	a3, a3		/* divisor <<= count */
	movi	a2, 0		/* quotient = 0 */

	/* test-subtract-and-shift loop; one quotient bit on each iteration */
#if XCHAL_HAVE_LOOPS
	loopnez	a4, .Lloopend
#endif /* XCHAL_HAVE_LOOPS */
.Lloop:
	bltu	a6, a3, .Lzerobit
	sub	a6, a6, a3
	addi	a2, a2, 1
.Lzerobit:
	slli	a2, a2, 1
	srli	a3, a3, 1
#if !XCHAL_HAVE_LOOPS
	addi	a4, a4, -1
	bnez	a4, .Lloop
#endif /* !XCHAL_HAVE_LOOPS */
.Lloopend:

	bltu	a6, a3, .Lreturn
	addi	a2, a2, 1	/* increment quotient if dividend >= divisor */
.Lreturn:
	leaf_return

.Lle_one:
	beqz	a3, .Lerror	/* if divisor == 1, return the dividend */
	leaf_return

.Lspecial:
	/* return dividend >= divisor */
	bltu	a6, a3, .Lreturn0
	movi	a2, 1
	leaf_return

.Lerror:
	/* Divide by zero: Use an illegal instruction to force an exception.
	   The subsequent "DIV0" string can be recognized by the exception
	   handler to identify the real cause of the exception.  */
	ill
	.ascii	"DIV0"

.Lreturn0:
	movi	a2, 0
#endif /* XCHAL_HAVE_DIV32 */
	leaf_return
	.size	__udivsi3, . - __udivsi3

#endif /* L_udivsi3 */


#ifdef L_divsi3
	.align	4
	.global	__divsi3
	.type	__divsi3, @function
__divsi3:
	leaf_entry sp, 16
#if XCHAL_HAVE_DIV32
	quos	a2, a2, a3
#else
	xor	a7, a2, a3	/* sign = dividend ^ divisor */
	do_abs	a6, a2, a4	/* udividend = abs (dividend) */
	do_abs	a3, a3, a4	/* udivisor = abs (divisor) */
	bltui	a3, 2, .Lle_one	/* check if udivisor <= 1 */
	do_nsau	a5, a6, a2, a8	/* udividend_shift = nsau (udividend) */
	do_nsau	a4, a3, a2, a8	/* udivisor_shift = nsau (udivisor) */
	bgeu	a5, a4, .Lspecial

	sub	a4, a4, a5	/* count = udivisor_shift - udividend_shift */
	ssl	a4
	sll	a3, a3		/* udivisor <<= count */
	movi	a2, 0		/* quotient = 0 */

	/* test-subtract-and-shift loop; one quotient bit on each iteration */
#if XCHAL_HAVE_LOOPS
	loopnez	a4, .Lloopend
#endif /* XCHAL_HAVE_LOOPS */
.Lloop:
	bltu	a6, a3, .Lzerobit
	sub	a6, a6, a3
	addi	a2, a2, 1
.Lzerobit:
	slli	a2, a2, 1
	srli	a3, a3, 1
#if !XCHAL_HAVE_LOOPS
	addi	a4, a4, -1
	bnez	a4, .Lloop
#endif /* !XCHAL_HAVE_LOOPS */
.Lloopend:

	bltu	a6, a3, .Lreturn
	addi	a2, a2, 1	/* increment if udividend >= udivisor */
.Lreturn:
	neg	a5, a2
	movltz	a2, a5, a7	/* return (sign < 0) ? -quotient : quotient */
	leaf_return

.Lle_one:
	beqz	a3, .Lerror
	neg	a2, a6		/* if udivisor == 1, then return... */
	movgez	a2, a6, a7	/* (sign < 0) ? -udividend : udividend */
	leaf_return

.Lspecial:
	bltu	a6, a3, .Lreturn0 /* if dividend < divisor, return 0 */
	movi	a2, 1
	movi	a4, -1
	movltz	a2, a4, a7	/* else return (sign < 0) ? -1 : 1 */
	leaf_return

.Lerror:
	/* Divide by zero: Use an illegal instruction to force an exception.
	   The subsequent "DIV0" string can be recognized by the exception
	   handler to identify the real cause of the exception.  */
	ill
	.ascii	"DIV0"

.Lreturn0:
	movi	a2, 0
#endif /* XCHAL_HAVE_DIV32 */
	leaf_return
	.size	__divsi3, . - __divsi3

#endif /* L_divsi3 */


#ifdef L_umodsi3
	.align	4
	.global	__umodsi3
	.type	__umodsi3, @function
__umodsi3:
	leaf_entry sp, 16
#if XCHAL_HAVE_DIV32
	remu	a2, a2, a3
#else
	bltui	a3, 2, .Lle_one	/* check if the divisor is <= 1 */

	do_nsau	a5, a2, a6, a7	/* dividend_shift = nsau (dividend) */
	do_nsau	a4, a3, a6, a7	/* divisor_shift = nsau (divisor) */
	bgeu	a5, a4, .Lspecial

	sub	a4, a4, a5	/* count = divisor_shift - dividend_shift */
	ssl	a4
	sll	a3, a3		/* divisor <<= count */

	/* test-subtract-and-shift loop */
#if XCHAL_HAVE_LOOPS
	loopnez	a4, .Lloopend
#endif /* XCHAL_HAVE_LOOPS */
.Lloop:
	bltu	a2, a3, .Lzerobit
	sub	a2, a2, a3
.Lzerobit:
	srli	a3, a3, 1
#if !XCHAL_HAVE_LOOPS
	addi	a4, a4, -1
	bnez	a4, .Lloop
#endif /* !XCHAL_HAVE_LOOPS */
.Lloopend:

.Lspecial:
	bltu	a2, a3, .Lreturn
	sub	a2, a2, a3	/* subtract once more if dividend >= divisor */
.Lreturn:
	leaf_return

.Lle_one:
	bnez	a3, .Lreturn0

	/* Divide by zero: Use an illegal instruction to force an exception.
	   The subsequent "DIV0" string can be recognized by the exception
	   handler to identify the real cause of the exception.  */
	ill
	.ascii	"DIV0"

.Lreturn0:
	movi	a2, 0
#endif /* XCHAL_HAVE_DIV32 */
	leaf_return
	.size	__umodsi3, . - __umodsi3

#endif /* L_umodsi3 */


#ifdef L_modsi3
	.align	4
	.global	__modsi3
	.type	__modsi3, @function
__modsi3:
	leaf_entry sp, 16
#if XCHAL_HAVE_DIV32
	rems	a2, a2, a3
#else
	mov	a7, a2		/* save original (signed) dividend */
	do_abs	a2, a2, a4	/* udividend = abs (dividend) */
	do_abs	a3, a3, a4	/* udivisor = abs (divisor) */
	bltui	a3, 2, .Lle_one	/* check if udivisor <= 1 */
	do_nsau	a5, a2, a6, a8	/* udividend_shift = nsau (udividend) */
	do_nsau	a4, a3, a6, a8	/* udivisor_shift = nsau (udivisor) */
	bgeu	a5, a4, .Lspecial

	sub	a4, a4, a5	/* count = udivisor_shift - udividend_shift */
	ssl	a4
	sll	a3, a3		/* udivisor <<= count */

	/* test-subtract-and-shift loop */
#if XCHAL_HAVE_LOOPS
	loopnez	a4, .Lloopend
#endif /* XCHAL_HAVE_LOOPS */
.Lloop:
	bltu	a2, a3, .Lzerobit
	sub	a2, a2, a3
.Lzerobit:
	srli	a3, a3, 1
#if !XCHAL_HAVE_LOOPS
	addi	a4, a4, -1
	bnez	a4, .Lloop
#endif /* !XCHAL_HAVE_LOOPS */
.Lloopend:

.Lspecial:
	bltu	a2, a3, .Lreturn
	sub	a2, a2, a3	/* subtract again if udividend >= udivisor */
.Lreturn:
	bgez	a7, .Lpositive
	neg	a2, a2		/* if (dividend < 0), return -udividend */
.Lpositive:
	leaf_return

.Lle_one:
	bnez	a3, .Lreturn0

	/* Divide by zero: Use an illegal instruction to force an exception.
	   The subsequent "DIV0" string can be recognized by the exception
	   handler to identify the real cause of the exception.  */
	ill
	.ascii	"DIV0"

.Lreturn0:
	movi	a2, 0
#endif /* XCHAL_HAVE_DIV32 */
	leaf_return
	.size	__modsi3, . - __modsi3

#endif /* L_modsi3 */


#ifdef __XTENSA_EB__
#define uh a2
#define ul a3
#else
#define uh a3
#define ul a2
#endif /* __XTENSA_EB__ */


#ifdef L_ashldi3
	.align	4
	.global	__ashldi3
	.type	__ashldi3, @function
__ashldi3:
	leaf_entry sp, 16
	ssl	a4
	bgei	a4, 32, .Llow_only
	src	uh, uh, ul
	sll	ul, ul
	leaf_return

.Llow_only:
	sll	uh, ul
	movi	ul, 0
	leaf_return
	.size	__ashldi3, . - __ashldi3

#endif /* L_ashldi3 */


#ifdef L_ashrdi3
	.align	4
	.global	__ashrdi3
	.type	__ashrdi3, @function
__ashrdi3:
	leaf_entry sp, 16
	ssr	a4
	bgei	a4, 32, .Lhigh_only
	src	ul, uh, ul
	sra	uh, uh
	leaf_return

.Lhigh_only:
	sra	ul, uh
	srai	uh, uh, 31
	leaf_return
	.size	__ashrdi3, . - __ashrdi3

#endif /* L_ashrdi3 */


#ifdef L_lshrdi3
	.align	4
	.global	__lshrdi3
	.type	__lshrdi3, @function
__lshrdi3:
	leaf_entry sp, 16
	ssr	a4
	bgei	a4, 32, .Lhigh_only1
	src	ul, uh, ul
	srl	uh, uh
	leaf_return

.Lhigh_only1:
	srl	ul, uh
	movi	uh, 0
	leaf_return
	.size	__lshrdi3, . - __lshrdi3

#endif /* L_lshrdi3 */


#include "ieee754-df.S"
#include "ieee754-sf.S"