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
|
//===- ARMInstrVFP.td - VFP support for ARM -------------------------------===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file describes the ARM VFP instruction set.
//
//===----------------------------------------------------------------------===//
def SDT_FTOI :
SDTypeProfile<1, 1, [SDTCisVT<0, f32>, SDTCisFP<1>]>;
def SDT_ITOF :
SDTypeProfile<1, 1, [SDTCisFP<0>, SDTCisVT<1, f32>]>;
def SDT_CMPFP0 :
SDTypeProfile<0, 1, [SDTCisFP<0>]>;
def SDT_VMOVDRR :
SDTypeProfile<1, 2, [SDTCisVT<0, f64>, SDTCisVT<1, i32>,
SDTCisSameAs<1, 2>]>;
def arm_ftoui : SDNode<"ARMISD::FTOUI", SDT_FTOI>;
def arm_ftosi : SDNode<"ARMISD::FTOSI", SDT_FTOI>;
def arm_sitof : SDNode<"ARMISD::SITOF", SDT_ITOF>;
def arm_uitof : SDNode<"ARMISD::UITOF", SDT_ITOF>;
def arm_fmstat : SDNode<"ARMISD::FMSTAT", SDTNone, [SDNPInFlag,SDNPOutFlag]>;
def arm_cmpfp : SDNode<"ARMISD::CMPFP", SDT_ARMCmp, [SDNPOutFlag]>;
def arm_cmpfp0 : SDNode<"ARMISD::CMPFPw0",SDT_CMPFP0, [SDNPOutFlag]>;
def arm_fmdrr : SDNode<"ARMISD::VMOVDRR", SDT_VMOVDRR>;
//===----------------------------------------------------------------------===//
// Operand Definitions.
//
def vfp_f32imm : Operand<f32>,
PatLeaf<(f32 fpimm), [{
return ARM::getVFPf32Imm(N->getValueAPF()) != -1;
}]> {
let PrintMethod = "printVFPf32ImmOperand";
}
def vfp_f64imm : Operand<f64>,
PatLeaf<(f64 fpimm), [{
return ARM::getVFPf64Imm(N->getValueAPF()) != -1;
}]> {
let PrintMethod = "printVFPf64ImmOperand";
}
//===----------------------------------------------------------------------===//
// Load / store Instructions.
//
let canFoldAsLoad = 1, isReMaterializable = 1 in {
def VLDRD : ADI5<0b1101, 0b01, (outs DPR:$dst), (ins addrmode5:$addr),
IIC_fpLoad64, "vldr", ".64\t$dst, $addr",
[(set DPR:$dst, (f64 (load addrmode5:$addr)))]>;
def VLDRS : ASI5<0b1101, 0b01, (outs SPR:$dst), (ins addrmode5:$addr),
IIC_fpLoad32, "vldr", ".32\t$dst, $addr",
[(set SPR:$dst, (load addrmode5:$addr))]>;
} // canFoldAsLoad
def VSTRD : ADI5<0b1101, 0b00, (outs), (ins DPR:$src, addrmode5:$addr),
IIC_fpStore64, "vstr", ".64\t$src, $addr",
[(store (f64 DPR:$src), addrmode5:$addr)]>;
def VSTRS : ASI5<0b1101, 0b00, (outs), (ins SPR:$src, addrmode5:$addr),
IIC_fpStore32, "vstr", ".32\t$src, $addr",
[(store SPR:$src, addrmode5:$addr)]>;
//===----------------------------------------------------------------------===//
// Load / store multiple Instructions.
//
let mayLoad = 1, neverHasSideEffects = 1, hasExtraDefRegAllocReq = 1 in {
def VLDMD : AXDI5<(outs), (ins addrmode5:$addr, pred:$p, reglist:$dsts,
variable_ops), IndexModeNone, IIC_fpLoadm,
"vldm${addr:submode}${p}\t${addr:base}, $dsts", "", []> {
let Inst{20} = 1;
}
def VLDMS : AXSI5<(outs), (ins addrmode5:$addr, pred:$p, reglist:$dsts,
variable_ops), IndexModeNone, IIC_fpLoadm,
"vldm${addr:submode}${p}\t${addr:base}, $dsts", "", []> {
let Inst{20} = 1;
}
def VLDMD_UPD : AXDI5<(outs GPR:$wb), (ins addrmode5:$addr, pred:$p,
reglist:$dsts, variable_ops),
IndexModeUpd, IIC_fpLoadm,
"vldm${addr:submode}${p}\t${addr:base}!, $dsts",
"$addr.base = $wb", []> {
let Inst{20} = 1;
}
def VLDMS_UPD : AXSI5<(outs GPR:$wb), (ins addrmode5:$addr, pred:$p,
reglist:$dsts, variable_ops),
IndexModeUpd, IIC_fpLoadm,
"vldm${addr:submode}${p}\t${addr:base}!, $dsts",
"$addr.base = $wb", []> {
let Inst{20} = 1;
}
} // mayLoad, neverHasSideEffects, hasExtraDefRegAllocReq
let mayStore = 1, neverHasSideEffects = 1, hasExtraSrcRegAllocReq = 1 in {
def VSTMD : AXDI5<(outs), (ins addrmode5:$addr, pred:$p, reglist:$srcs,
variable_ops), IndexModeNone, IIC_fpStorem,
"vstm${addr:submode}${p}\t${addr:base}, $srcs", "", []> {
let Inst{20} = 0;
}
def VSTMS : AXSI5<(outs), (ins addrmode5:$addr, pred:$p, reglist:$srcs,
variable_ops), IndexModeNone, IIC_fpStorem,
"vstm${addr:submode}${p}\t${addr:base}, $srcs", "", []> {
let Inst{20} = 0;
}
def VSTMD_UPD : AXDI5<(outs GPR:$wb), (ins addrmode5:$addr, pred:$p,
reglist:$srcs, variable_ops),
IndexModeUpd, IIC_fpStorem,
"vstm${addr:submode}${p}\t${addr:base}!, $srcs",
"$addr.base = $wb", []> {
let Inst{20} = 0;
}
def VSTMS_UPD : AXSI5<(outs GPR:$wb), (ins addrmode5:$addr, pred:$p,
reglist:$srcs, variable_ops),
IndexModeUpd, IIC_fpStorem,
"vstm${addr:submode}${p}\t${addr:base}!, $srcs",
"$addr.base = $wb", []> {
let Inst{20} = 0;
}
} // mayStore, neverHasSideEffects, hasExtraSrcRegAllocReq
// FLDMX, FSTMX - mixing S/D registers for pre-armv6 cores
//===----------------------------------------------------------------------===//
// FP Binary Operations.
//
def VADDD : ADbI<0b11100, 0b11, 0, 0, (outs DPR:$dst), (ins DPR:$a, DPR:$b),
IIC_fpALU64, "vadd", ".f64\t$dst, $a, $b",
[(set DPR:$dst, (fadd DPR:$a, (f64 DPR:$b)))]>;
def VADDS : ASbIn<0b11100, 0b11, 0, 0, (outs SPR:$dst), (ins SPR:$a, SPR:$b),
IIC_fpALU32, "vadd", ".f32\t$dst, $a, $b",
[(set SPR:$dst, (fadd SPR:$a, SPR:$b))]>;
// These are encoded as unary instructions.
let Defs = [FPSCR] in {
def VCMPED : ADuI<0b11101, 0b11, 0b0100, 0b11, 0, (outs), (ins DPR:$a, DPR:$b),
IIC_fpCMP64, "vcmpe", ".f64\t$a, $b",
[(arm_cmpfp DPR:$a, (f64 DPR:$b))]>;
def VCMPD : ADuI<0b11101, 0b11, 0b0100, 0b01, 0, (outs), (ins DPR:$a, DPR:$b),
IIC_fpCMP64, "vcmp", ".f64\t$a, $b",
[/* For disassembly only; pattern left blank */]>;
def VCMPES : ASuI<0b11101, 0b11, 0b0100, 0b11, 0, (outs), (ins SPR:$a, SPR:$b),
IIC_fpCMP32, "vcmpe", ".f32\t$a, $b",
[(arm_cmpfp SPR:$a, SPR:$b)]>;
def VCMPS : ASuI<0b11101, 0b11, 0b0100, 0b01, 0, (outs), (ins SPR:$a, SPR:$b),
IIC_fpCMP32, "vcmp", ".f32\t$a, $b",
[/* For disassembly only; pattern left blank */]>;
}
def VDIVD : ADbI<0b11101, 0b00, 0, 0, (outs DPR:$dst), (ins DPR:$a, DPR:$b),
IIC_fpDIV64, "vdiv", ".f64\t$dst, $a, $b",
[(set DPR:$dst, (fdiv DPR:$a, (f64 DPR:$b)))]>;
def VDIVS : ASbI<0b11101, 0b00, 0, 0, (outs SPR:$dst), (ins SPR:$a, SPR:$b),
IIC_fpDIV32, "vdiv", ".f32\t$dst, $a, $b",
[(set SPR:$dst, (fdiv SPR:$a, SPR:$b))]>;
def VMULD : ADbI<0b11100, 0b10, 0, 0, (outs DPR:$dst), (ins DPR:$a, DPR:$b),
IIC_fpMUL64, "vmul", ".f64\t$dst, $a, $b",
[(set DPR:$dst, (fmul DPR:$a, (f64 DPR:$b)))]>;
def VMULS : ASbIn<0b11100, 0b10, 0, 0, (outs SPR:$dst), (ins SPR:$a, SPR:$b),
IIC_fpMUL32, "vmul", ".f32\t$dst, $a, $b",
[(set SPR:$dst, (fmul SPR:$a, SPR:$b))]>;
def VNMULD : ADbI<0b11100, 0b10, 1, 0, (outs DPR:$dst), (ins DPR:$a, DPR:$b),
IIC_fpMUL64, "vnmul", ".f64\t$dst, $a, $b",
[(set DPR:$dst, (fneg (fmul DPR:$a, (f64 DPR:$b))))]>;
def VNMULS : ASbI<0b11100, 0b10, 1, 0, (outs SPR:$dst), (ins SPR:$a, SPR:$b),
IIC_fpMUL32, "vnmul", ".f32\t$dst, $a, $b",
[(set SPR:$dst, (fneg (fmul SPR:$a, SPR:$b)))]>;
// Match reassociated forms only if not sign dependent rounding.
def : Pat<(fmul (fneg DPR:$a), (f64 DPR:$b)),
(VNMULD DPR:$a, DPR:$b)>, Requires<[NoHonorSignDependentRounding]>;
def : Pat<(fmul (fneg SPR:$a), SPR:$b),
(VNMULS SPR:$a, SPR:$b)>, Requires<[NoHonorSignDependentRounding]>;
def VSUBD : ADbI<0b11100, 0b11, 1, 0, (outs DPR:$dst), (ins DPR:$a, DPR:$b),
IIC_fpALU64, "vsub", ".f64\t$dst, $a, $b",
[(set DPR:$dst, (fsub DPR:$a, (f64 DPR:$b)))]>;
def VSUBS : ASbIn<0b11100, 0b11, 1, 0, (outs SPR:$dst), (ins SPR:$a, SPR:$b),
IIC_fpALU32, "vsub", ".f32\t$dst, $a, $b",
[(set SPR:$dst, (fsub SPR:$a, SPR:$b))]>;
//===----------------------------------------------------------------------===//
// FP Unary Operations.
//
def VABSD : ADuI<0b11101, 0b11, 0b0000, 0b11, 0, (outs DPR:$dst), (ins DPR:$a),
IIC_fpUNA64, "vabs", ".f64\t$dst, $a",
[(set DPR:$dst, (fabs (f64 DPR:$a)))]>;
def VABSS : ASuIn<0b11101, 0b11, 0b0000, 0b11, 0,(outs SPR:$dst), (ins SPR:$a),
IIC_fpUNA32, "vabs", ".f32\t$dst, $a",
[(set SPR:$dst, (fabs SPR:$a))]>;
let Defs = [FPSCR] in {
def VCMPEZD : ADuI<0b11101, 0b11, 0b0101, 0b11, 0, (outs), (ins DPR:$a),
IIC_fpCMP64, "vcmpe", ".f64\t$a, #0",
[(arm_cmpfp0 (f64 DPR:$a))]>;
def VCMPZD : ADuI<0b11101, 0b11, 0b0101, 0b01, 0, (outs), (ins DPR:$a),
IIC_fpCMP64, "vcmp", ".f64\t$a, #0",
[/* For disassembly only; pattern left blank */]>;
def VCMPEZS : ASuI<0b11101, 0b11, 0b0101, 0b11, 0, (outs), (ins SPR:$a),
IIC_fpCMP32, "vcmpe", ".f32\t$a, #0",
[(arm_cmpfp0 SPR:$a)]>;
def VCMPZS : ASuI<0b11101, 0b11, 0b0101, 0b01, 0, (outs), (ins SPR:$a),
IIC_fpCMP32, "vcmp", ".f32\t$a, #0",
[/* For disassembly only; pattern left blank */]>;
}
def VCVTDS : ASuI<0b11101, 0b11, 0b0111, 0b11, 0, (outs DPR:$dst), (ins SPR:$a),
IIC_fpCVTDS, "vcvt", ".f64.f32\t$dst, $a",
[(set DPR:$dst, (fextend SPR:$a))]>;
// Special case encoding: bits 11-8 is 0b1011.
def VCVTSD : VFPAI<(outs SPR:$dst), (ins DPR:$a), VFPUnaryFrm,
IIC_fpCVTSD, "vcvt", ".f32.f64\t$dst, $a",
[(set SPR:$dst, (fround DPR:$a))]> {
let Inst{27-23} = 0b11101;
let Inst{21-16} = 0b110111;
let Inst{11-8} = 0b1011;
let Inst{7-6} = 0b11;
let Inst{4} = 0;
}
// Between half-precision and single-precision. For disassembly only.
def VCVTBSH: ASuI<0b11101, 0b11, 0b0010, 0b01, 0, (outs SPR:$dst), (ins SPR:$a),
/* FIXME */ IIC_fpCVTSH, "vcvtb", ".f32.f16\t$dst, $a",
[/* For disassembly only; pattern left blank */]>;
def : ARMPat<(f32_to_f16 SPR:$a),
(i32 (COPY_TO_REGCLASS (VCVTBSH SPR:$a), GPR))>;
def VCVTBHS: ASuI<0b11101, 0b11, 0b0011, 0b01, 0, (outs SPR:$dst), (ins SPR:$a),
/* FIXME */ IIC_fpCVTHS, "vcvtb", ".f16.f32\t$dst, $a",
[/* For disassembly only; pattern left blank */]>;
def : ARMPat<(f16_to_f32 GPR:$a),
(VCVTBHS (COPY_TO_REGCLASS GPR:$a, SPR))>;
def VCVTTSH: ASuI<0b11101, 0b11, 0b0010, 0b11, 0, (outs SPR:$dst), (ins SPR:$a),
/* FIXME */ IIC_fpCVTSH, "vcvtt", ".f32.f16\t$dst, $a",
[/* For disassembly only; pattern left blank */]>;
def VCVTTHS: ASuI<0b11101, 0b11, 0b0011, 0b11, 0, (outs SPR:$dst), (ins SPR:$a),
/* FIXME */ IIC_fpCVTHS, "vcvtt", ".f16.f32\t$dst, $a",
[/* For disassembly only; pattern left blank */]>;
let neverHasSideEffects = 1 in {
def VMOVD: ADuI<0b11101, 0b11, 0b0000, 0b01, 0, (outs DPR:$dst), (ins DPR:$a),
IIC_fpUNA64, "vmov", ".f64\t$dst, $a", []>;
def VMOVS: ASuI<0b11101, 0b11, 0b0000, 0b01, 0, (outs SPR:$dst), (ins SPR:$a),
IIC_fpUNA32, "vmov", ".f32\t$dst, $a", []>;
} // neverHasSideEffects
def VNEGD : ADuI<0b11101, 0b11, 0b0001, 0b01, 0, (outs DPR:$dst), (ins DPR:$a),
IIC_fpUNA64, "vneg", ".f64\t$dst, $a",
[(set DPR:$dst, (fneg (f64 DPR:$a)))]>;
def VNEGS : ASuIn<0b11101, 0b11, 0b0001, 0b01, 0,(outs SPR:$dst), (ins SPR:$a),
IIC_fpUNA32, "vneg", ".f32\t$dst, $a",
[(set SPR:$dst, (fneg SPR:$a))]>;
def VSQRTD : ADuI<0b11101, 0b11, 0b0001, 0b11, 0, (outs DPR:$dst), (ins DPR:$a),
IIC_fpSQRT64, "vsqrt", ".f64\t$dst, $a",
[(set DPR:$dst, (fsqrt (f64 DPR:$a)))]>;
def VSQRTS : ASuI<0b11101, 0b11, 0b0001, 0b11, 0, (outs SPR:$dst), (ins SPR:$a),
IIC_fpSQRT32, "vsqrt", ".f32\t$dst, $a",
[(set SPR:$dst, (fsqrt SPR:$a))]>;
//===----------------------------------------------------------------------===//
// FP <-> GPR Copies. Int <-> FP Conversions.
//
def VMOVRS : AVConv2I<0b11100001, 0b1010, (outs GPR:$dst), (ins SPR:$src),
IIC_fpMOVSI, "vmov", "\t$dst, $src",
[(set GPR:$dst, (bitconvert SPR:$src))]>;
def VMOVSR : AVConv4I<0b11100000, 0b1010, (outs SPR:$dst), (ins GPR:$src),
IIC_fpMOVIS, "vmov", "\t$dst, $src",
[(set SPR:$dst, (bitconvert GPR:$src))]>;
let neverHasSideEffects = 1 in {
def VMOVRRD : AVConv3I<0b11000101, 0b1011,
(outs GPR:$wb, GPR:$dst2), (ins DPR:$src),
IIC_fpMOVDI, "vmov", "\t$wb, $dst2, $src",
[/* FIXME: Can't write pattern for multiple result instr*/]> {
let Inst{7-6} = 0b00;
}
def VMOVRRS : AVConv3I<0b11000101, 0b1010,
(outs GPR:$wb, GPR:$dst2), (ins SPR:$src1, SPR:$src2),
IIC_fpMOVDI, "vmov", "\t$wb, $dst2, $src1, $src2",
[/* For disassembly only; pattern left blank */]> {
let Inst{7-6} = 0b00;
}
} // neverHasSideEffects
// FMDHR: GPR -> SPR
// FMDLR: GPR -> SPR
def VMOVDRR : AVConv5I<0b11000100, 0b1011,
(outs DPR:$dst), (ins GPR:$src1, GPR:$src2),
IIC_fpMOVID, "vmov", "\t$dst, $src1, $src2",
[(set DPR:$dst, (arm_fmdrr GPR:$src1, GPR:$src2))]> {
let Inst{7-6} = 0b00;
}
let neverHasSideEffects = 1 in
def VMOVSRR : AVConv5I<0b11000100, 0b1010,
(outs SPR:$dst1, SPR:$dst2), (ins GPR:$src1, GPR:$src2),
IIC_fpMOVID, "vmov", "\t$dst1, $dst2, $src1, $src2",
[/* For disassembly only; pattern left blank */]> {
let Inst{7-6} = 0b00;
}
// FMRDH: SPR -> GPR
// FMRDL: SPR -> GPR
// FMRRS: SPR -> GPR
// FMRX : SPR system reg -> GPR
// FMSRR: GPR -> SPR
// FMXR: GPR -> VFP Sstem reg
// Int to FP:
def VSITOD : AVConv1I<0b11101, 0b11, 0b1000, 0b1011,
(outs DPR:$dst), (ins SPR:$a),
IIC_fpCVTID, "vcvt", ".f64.s32\t$dst, $a",
[(set DPR:$dst, (f64 (arm_sitof SPR:$a)))]> {
let Inst{7} = 1; // s32
}
def VSITOS : AVConv1In<0b11101, 0b11, 0b1000, 0b1010,
(outs SPR:$dst),(ins SPR:$a),
IIC_fpCVTIS, "vcvt", ".f32.s32\t$dst, $a",
[(set SPR:$dst, (arm_sitof SPR:$a))]> {
let Inst{7} = 1; // s32
}
def VUITOD : AVConv1I<0b11101, 0b11, 0b1000, 0b1011,
(outs DPR:$dst), (ins SPR:$a),
IIC_fpCVTID, "vcvt", ".f64.u32\t$dst, $a",
[(set DPR:$dst, (f64 (arm_uitof SPR:$a)))]> {
let Inst{7} = 0; // u32
}
def VUITOS : AVConv1In<0b11101, 0b11, 0b1000, 0b1010,
(outs SPR:$dst), (ins SPR:$a),
IIC_fpCVTIS, "vcvt", ".f32.u32\t$dst, $a",
[(set SPR:$dst, (arm_uitof SPR:$a))]> {
let Inst{7} = 0; // u32
}
// FP to Int:
// Always set Z bit in the instruction, i.e. "round towards zero" variants.
def VTOSIZD : AVConv1I<0b11101, 0b11, 0b1101, 0b1011,
(outs SPR:$dst), (ins DPR:$a),
IIC_fpCVTDI, "vcvt", ".s32.f64\t$dst, $a",
[(set SPR:$dst, (arm_ftosi (f64 DPR:$a)))]> {
let Inst{7} = 1; // Z bit
}
def VTOSIZS : AVConv1In<0b11101, 0b11, 0b1101, 0b1010,
(outs SPR:$dst), (ins SPR:$a),
IIC_fpCVTSI, "vcvt", ".s32.f32\t$dst, $a",
[(set SPR:$dst, (arm_ftosi SPR:$a))]> {
let Inst{7} = 1; // Z bit
}
def VTOUIZD : AVConv1I<0b11101, 0b11, 0b1100, 0b1011,
(outs SPR:$dst), (ins DPR:$a),
IIC_fpCVTDI, "vcvt", ".u32.f64\t$dst, $a",
[(set SPR:$dst, (arm_ftoui (f64 DPR:$a)))]> {
let Inst{7} = 1; // Z bit
}
def VTOUIZS : AVConv1In<0b11101, 0b11, 0b1100, 0b1010,
(outs SPR:$dst), (ins SPR:$a),
IIC_fpCVTSI, "vcvt", ".u32.f32\t$dst, $a",
[(set SPR:$dst, (arm_ftoui SPR:$a))]> {
let Inst{7} = 1; // Z bit
}
// And the Z bit '0' variants, i.e. use the rounding mode specified by FPSCR.
// For disassembly only.
def VTOSIRD : AVConv1I<0b11101, 0b11, 0b1101, 0b1011,
(outs SPR:$dst), (ins DPR:$a),
IIC_fpCVTDI, "vcvtr", ".s32.f64\t$dst, $a",
[/* For disassembly only; pattern left blank */]> {
let Inst{7} = 0; // Z bit
}
def VTOSIRS : AVConv1In<0b11101, 0b11, 0b1101, 0b1010,
(outs SPR:$dst), (ins SPR:$a),
IIC_fpCVTSI, "vcvtr", ".s32.f32\t$dst, $a",
[/* For disassembly only; pattern left blank */]> {
let Inst{7} = 0; // Z bit
}
def VTOUIRD : AVConv1I<0b11101, 0b11, 0b1100, 0b1011,
(outs SPR:$dst), (ins DPR:$a),
IIC_fpCVTDI, "vcvtr", ".u32.f64\t$dst, $a",
[/* For disassembly only; pattern left blank */]> {
let Inst{7} = 0; // Z bit
}
def VTOUIRS : AVConv1In<0b11101, 0b11, 0b1100, 0b1010,
(outs SPR:$dst), (ins SPR:$a),
IIC_fpCVTSI, "vcvtr", ".u32.f32\t$dst, $a",
[/* For disassembly only; pattern left blank */]> {
let Inst{7} = 0; // Z bit
}
// Convert between floating-point and fixed-point
// Data type for fixed-point naming convention:
// S16 (U=0, sx=0) -> SH
// U16 (U=1, sx=0) -> UH
// S32 (U=0, sx=1) -> SL
// U32 (U=1, sx=1) -> UL
let Constraints = "$a = $dst" in {
// FP to Fixed-Point:
def VTOSHS : AVConv1XI<0b11101, 0b11, 0b1110, 0b1010, 0,
(outs SPR:$dst), (ins SPR:$a, i32imm:$fbits),
IIC_fpCVTSI, "vcvt", ".s16.f32\t$dst, $a, $fbits",
[/* For disassembly only; pattern left blank */]>;
def VTOUHS : AVConv1XI<0b11101, 0b11, 0b1111, 0b1010, 0,
(outs SPR:$dst), (ins SPR:$a, i32imm:$fbits),
IIC_fpCVTSI, "vcvt", ".u16.f32\t$dst, $a, $fbits",
[/* For disassembly only; pattern left blank */]>;
def VTOSLS : AVConv1XI<0b11101, 0b11, 0b1110, 0b1010, 1,
(outs SPR:$dst), (ins SPR:$a, i32imm:$fbits),
IIC_fpCVTSI, "vcvt", ".s32.f32\t$dst, $a, $fbits",
[/* For disassembly only; pattern left blank */]>;
def VTOULS : AVConv1XI<0b11101, 0b11, 0b1111, 0b1010, 1,
(outs SPR:$dst), (ins SPR:$a, i32imm:$fbits),
IIC_fpCVTSI, "vcvt", ".u32.f32\t$dst, $a, $fbits",
[/* For disassembly only; pattern left blank */]>;
def VTOSHD : AVConv1XI<0b11101, 0b11, 0b1110, 0b1011, 0,
(outs DPR:$dst), (ins DPR:$a, i32imm:$fbits),
IIC_fpCVTDI, "vcvt", ".s16.f64\t$dst, $a, $fbits",
[/* For disassembly only; pattern left blank */]>;
def VTOUHD : AVConv1XI<0b11101, 0b11, 0b1111, 0b1011, 0,
(outs DPR:$dst), (ins DPR:$a, i32imm:$fbits),
IIC_fpCVTDI, "vcvt", ".u16.f64\t$dst, $a, $fbits",
[/* For disassembly only; pattern left blank */]>;
def VTOSLD : AVConv1XI<0b11101, 0b11, 0b1110, 0b1011, 1,
(outs DPR:$dst), (ins DPR:$a, i32imm:$fbits),
IIC_fpCVTDI, "vcvt", ".s32.f64\t$dst, $a, $fbits",
[/* For disassembly only; pattern left blank */]>;
def VTOULD : AVConv1XI<0b11101, 0b11, 0b1111, 0b1011, 1,
(outs DPR:$dst), (ins DPR:$a, i32imm:$fbits),
IIC_fpCVTDI, "vcvt", ".u32.f64\t$dst, $a, $fbits",
[/* For disassembly only; pattern left blank */]>;
// Fixed-Point to FP:
def VSHTOS : AVConv1XI<0b11101, 0b11, 0b1010, 0b1010, 0,
(outs SPR:$dst), (ins SPR:$a, i32imm:$fbits),
IIC_fpCVTIS, "vcvt", ".f32.s16\t$dst, $a, $fbits",
[/* For disassembly only; pattern left blank */]>;
def VUHTOS : AVConv1XI<0b11101, 0b11, 0b1011, 0b1010, 0,
(outs SPR:$dst), (ins SPR:$a, i32imm:$fbits),
IIC_fpCVTIS, "vcvt", ".f32.u16\t$dst, $a, $fbits",
[/* For disassembly only; pattern left blank */]>;
def VSLTOS : AVConv1XI<0b11101, 0b11, 0b1010, 0b1010, 1,
(outs SPR:$dst), (ins SPR:$a, i32imm:$fbits),
IIC_fpCVTIS, "vcvt", ".f32.s32\t$dst, $a, $fbits",
[/* For disassembly only; pattern left blank */]>;
def VULTOS : AVConv1XI<0b11101, 0b11, 0b1011, 0b1010, 1,
(outs SPR:$dst), (ins SPR:$a, i32imm:$fbits),
IIC_fpCVTIS, "vcvt", ".f32.u32\t$dst, $a, $fbits",
[/* For disassembly only; pattern left blank */]>;
def VSHTOD : AVConv1XI<0b11101, 0b11, 0b1010, 0b1011, 0,
(outs DPR:$dst), (ins DPR:$a, i32imm:$fbits),
IIC_fpCVTID, "vcvt", ".f64.s16\t$dst, $a, $fbits",
[/* For disassembly only; pattern left blank */]>;
def VUHTOD : AVConv1XI<0b11101, 0b11, 0b1011, 0b1011, 0,
(outs DPR:$dst), (ins DPR:$a, i32imm:$fbits),
IIC_fpCVTID, "vcvt", ".f64.u16\t$dst, $a, $fbits",
[/* For disassembly only; pattern left blank */]>;
def VSLTOD : AVConv1XI<0b11101, 0b11, 0b1010, 0b1011, 1,
(outs DPR:$dst), (ins DPR:$a, i32imm:$fbits),
IIC_fpCVTID, "vcvt", ".f64.s32\t$dst, $a, $fbits",
[/* For disassembly only; pattern left blank */]>;
def VULTOD : AVConv1XI<0b11101, 0b11, 0b1011, 0b1011, 1,
(outs DPR:$dst), (ins DPR:$a, i32imm:$fbits),
IIC_fpCVTID, "vcvt", ".f64.u32\t$dst, $a, $fbits",
[/* For disassembly only; pattern left blank */]>;
} // End of 'let Constraints = "$src = $dst" in'
//===----------------------------------------------------------------------===//
// FP FMA Operations.
//
def VMLAD : ADbI_vmlX<0b11100, 0b00, 0, 0,
(outs DPR:$dst), (ins DPR:$dstin, DPR:$a, DPR:$b),
IIC_fpMAC64, "vmla", ".f64\t$dst, $a, $b",
[(set DPR:$dst, (fadd (fmul DPR:$a, DPR:$b),
(f64 DPR:$dstin)))]>,
RegConstraint<"$dstin = $dst">;
def VMLAS : ASbIn<0b11100, 0b00, 0, 0,
(outs SPR:$dst), (ins SPR:$dstin, SPR:$a, SPR:$b),
IIC_fpMAC32, "vmla", ".f32\t$dst, $a, $b",
[(set SPR:$dst, (fadd (fmul SPR:$a, SPR:$b), SPR:$dstin))]>,
RegConstraint<"$dstin = $dst">;
def VNMLSD : ADbI_vmlX<0b11100, 0b01, 0, 0,
(outs DPR:$dst), (ins DPR:$dstin, DPR:$a, DPR:$b),
IIC_fpMAC64, "vnmls", ".f64\t$dst, $a, $b",
[(set DPR:$dst, (fsub (fmul DPR:$a, DPR:$b),
(f64 DPR:$dstin)))]>,
RegConstraint<"$dstin = $dst">;
def VNMLSS : ASbI<0b11100, 0b01, 0, 0,
(outs SPR:$dst), (ins SPR:$dstin, SPR:$a, SPR:$b),
IIC_fpMAC32, "vnmls", ".f32\t$dst, $a, $b",
[(set SPR:$dst, (fsub (fmul SPR:$a, SPR:$b), SPR:$dstin))]>,
RegConstraint<"$dstin = $dst">;
def VMLSD : ADbI_vmlX<0b11100, 0b00, 1, 0,
(outs DPR:$dst), (ins DPR:$dstin, DPR:$a, DPR:$b),
IIC_fpMAC64, "vmls", ".f64\t$dst, $a, $b",
[(set DPR:$dst, (fadd (fneg (fmul DPR:$a, DPR:$b)),
(f64 DPR:$dstin)))]>,
RegConstraint<"$dstin = $dst">;
def VMLSS : ASbIn<0b11100, 0b00, 1, 0,
(outs SPR:$dst), (ins SPR:$dstin, SPR:$a, SPR:$b),
IIC_fpMAC32, "vmls", ".f32\t$dst, $a, $b",
[(set SPR:$dst, (fadd (fneg (fmul SPR:$a, SPR:$b)), SPR:$dstin))]>,
RegConstraint<"$dstin = $dst">;
def : Pat<(fsub DPR:$dstin, (fmul DPR:$a, (f64 DPR:$b))),
(VMLSD DPR:$dstin, DPR:$a, DPR:$b)>, Requires<[DontUseNEONForFP]>;
def : Pat<(fsub SPR:$dstin, (fmul SPR:$a, SPR:$b)),
(VMLSS SPR:$dstin, SPR:$a, SPR:$b)>, Requires<[DontUseNEONForFP]>;
def VNMLAD : ADbI_vmlX<0b11100, 0b01, 1, 0,
(outs DPR:$dst), (ins DPR:$dstin, DPR:$a, DPR:$b),
IIC_fpMAC64, "vnmla", ".f64\t$dst, $a, $b",
[(set DPR:$dst, (fsub (fneg (fmul DPR:$a, DPR:$b)),
(f64 DPR:$dstin)))]>,
RegConstraint<"$dstin = $dst">;
def VNMLAS : ASbI<0b11100, 0b01, 1, 0,
(outs SPR:$dst), (ins SPR:$dstin, SPR:$a, SPR:$b),
IIC_fpMAC32, "vnmla", ".f32\t$dst, $a, $b",
[(set SPR:$dst, (fsub (fneg (fmul SPR:$a, SPR:$b)), SPR:$dstin))]>,
RegConstraint<"$dstin = $dst">;
//===----------------------------------------------------------------------===//
// FP Conditional moves.
//
let neverHasSideEffects = 1 in {
def VMOVDcc : ADuI<0b11101, 0b11, 0b0000, 0b01, 0,
(outs DPR:$dst), (ins DPR:$false, DPR:$true),
IIC_fpUNA64, "vmov", ".f64\t$dst, $true",
[/*(set DPR:$dst, (ARMcmov DPR:$false, DPR:$true, imm:$cc))*/]>,
RegConstraint<"$false = $dst">;
def VMOVScc : ASuI<0b11101, 0b11, 0b0000, 0b01, 0,
(outs SPR:$dst), (ins SPR:$false, SPR:$true),
IIC_fpUNA32, "vmov", ".f32\t$dst, $true",
[/*(set SPR:$dst, (ARMcmov SPR:$false, SPR:$true, imm:$cc))*/]>,
RegConstraint<"$false = $dst">;
def VNEGDcc : ADuI<0b11101, 0b11, 0b0001, 0b01, 0,
(outs DPR:$dst), (ins DPR:$false, DPR:$true),
IIC_fpUNA64, "vneg", ".f64\t$dst, $true",
[/*(set DPR:$dst, (ARMcneg DPR:$false, DPR:$true, imm:$cc))*/]>,
RegConstraint<"$false = $dst">;
def VNEGScc : ASuI<0b11101, 0b11, 0b0001, 0b01, 0,
(outs SPR:$dst), (ins SPR:$false, SPR:$true),
IIC_fpUNA32, "vneg", ".f32\t$dst, $true",
[/*(set SPR:$dst, (ARMcneg SPR:$false, SPR:$true, imm:$cc))*/]>,
RegConstraint<"$false = $dst">;
} // neverHasSideEffects
//===----------------------------------------------------------------------===//
// Misc.
//
// APSR is the application level alias of CPSR. This FPSCR N, Z, C, V flags
// to APSR.
let Defs = [CPSR], Uses = [FPSCR] in
def FMSTAT : VFPAI<(outs), (ins), VFPMiscFrm, IIC_fpSTAT, "vmrs",
"\tapsr_nzcv, fpscr",
[(arm_fmstat)]> {
let Inst{27-20} = 0b11101111;
let Inst{19-16} = 0b0001;
let Inst{15-12} = 0b1111;
let Inst{11-8} = 0b1010;
let Inst{7} = 0;
let Inst{4} = 1;
}
// FPSCR <-> GPR (for disassembly only)
let neverHasSideEffects = 1 in {
let Uses = [FPSCR] in {
def VMRS : VFPAI<(outs GPR:$dst), (ins), VFPMiscFrm, IIC_fpSTAT, "vmrs",
"\t$dst, fpscr",
[/* For disassembly only; pattern left blank */]> {
let Inst{27-20} = 0b11101111;
let Inst{19-16} = 0b0001;
let Inst{11-8} = 0b1010;
let Inst{7} = 0;
let Inst{4} = 1;
}
}
let Defs = [FPSCR] in {
def VMSR : VFPAI<(outs), (ins GPR:$src), VFPMiscFrm, IIC_fpSTAT, "vmsr",
"\tfpscr, $src",
[/* For disassembly only; pattern left blank */]> {
let Inst{27-20} = 0b11101110;
let Inst{19-16} = 0b0001;
let Inst{11-8} = 0b1010;
let Inst{7} = 0;
let Inst{4} = 1;
}
}
} // neverHasSideEffects
// Materialize FP immediates. VFP3 only.
let isReMaterializable = 1 in {
def FCONSTD : VFPAI<(outs DPR:$dst), (ins vfp_f64imm:$imm),
VFPMiscFrm, IIC_fpUNA64,
"vmov", ".f64\t$dst, $imm",
[(set DPR:$dst, vfp_f64imm:$imm)]>, Requires<[HasVFP3]> {
let Inst{27-23} = 0b11101;
let Inst{21-20} = 0b11;
let Inst{11-9} = 0b101;
let Inst{8} = 1;
let Inst{7-4} = 0b0000;
}
def FCONSTS : VFPAI<(outs SPR:$dst), (ins vfp_f32imm:$imm),
VFPMiscFrm, IIC_fpUNA32,
"vmov", ".f32\t$dst, $imm",
[(set SPR:$dst, vfp_f32imm:$imm)]>, Requires<[HasVFP3]> {
let Inst{27-23} = 0b11101;
let Inst{21-20} = 0b11;
let Inst{11-9} = 0b101;
let Inst{8} = 0;
let Inst{7-4} = 0b0000;
}
}
|