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
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
|
//===-- SparcInstrInfo.td - Target Description for Sparc Target -----------===//
//
// 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 Sparc instructions in TableGen format.
//
//===----------------------------------------------------------------------===//
//===----------------------------------------------------------------------===//
// Instruction format superclass
//===----------------------------------------------------------------------===//
include "SparcInstrFormats.td"
//===----------------------------------------------------------------------===//
// Feature predicates.
//===----------------------------------------------------------------------===//
// True when generating 32-bit code.
def Is32Bit : Predicate<"!Subtarget.is64Bit()">;
// True when generating 64-bit code. This also implies HasV9.
def Is64Bit : Predicate<"Subtarget.is64Bit()">;
// HasV9 - This predicate is true when the target processor supports V9
// instructions. Note that the machine may be running in 32-bit mode.
def HasV9 : Predicate<"Subtarget.isV9()">;
// HasNoV9 - This predicate is true when the target doesn't have V9
// instructions. Use of this is just a hack for the isel not having proper
// costs for V8 instructions that are more expensive than their V9 ones.
def HasNoV9 : Predicate<"!Subtarget.isV9()">;
// HasVIS - This is true when the target processor has VIS extensions.
def HasVIS : Predicate<"Subtarget.isVIS()">;
// HasHardQuad - This is true when the target processor supports quad floating
// point instructions.
def HasHardQuad : Predicate<"Subtarget.hasHardQuad()">;
// UseDeprecatedInsts - This predicate is true when the target processor is a
// V8, or when it is V9 but the V8 deprecated instructions are efficient enough
// to use when appropriate. In either of these cases, the instruction selector
// will pick deprecated instructions.
def UseDeprecatedInsts : Predicate<"Subtarget.useDeprecatedV8Instructions()">;
//===----------------------------------------------------------------------===//
// Instruction Pattern Stuff
//===----------------------------------------------------------------------===//
def simm11 : PatLeaf<(imm), [{ return isInt<11>(N->getSExtValue()); }]>;
def simm13 : PatLeaf<(imm), [{ return isInt<13>(N->getSExtValue()); }]>;
def LO10 : SDNodeXForm<imm, [{
return CurDAG->getTargetConstant((unsigned)N->getZExtValue() & 1023,
MVT::i32);
}]>;
def HI22 : SDNodeXForm<imm, [{
// Transformation function: shift the immediate value down into the low bits.
return CurDAG->getTargetConstant((unsigned)N->getZExtValue() >> 10, MVT::i32);
}]>;
def SETHIimm : PatLeaf<(imm), [{
return isShiftedUInt<22, 10>(N->getZExtValue());
}], HI22>;
// Addressing modes.
def ADDRrr : ComplexPattern<iPTR, 2, "SelectADDRrr", [], []>;
def ADDRri : ComplexPattern<iPTR, 2, "SelectADDRri", [frameindex], []>;
// Address operands
def SparcMEMrrAsmOperand : AsmOperandClass {
let Name = "MEMrr";
let ParserMethod = "parseMEMOperand";
}
def SparcMEMriAsmOperand : AsmOperandClass {
let Name = "MEMri";
let ParserMethod = "parseMEMOperand";
}
def MEMrr : Operand<iPTR> {
let PrintMethod = "printMemOperand";
let MIOperandInfo = (ops ptr_rc, ptr_rc);
let ParserMatchClass = SparcMEMrrAsmOperand;
}
def MEMri : Operand<iPTR> {
let PrintMethod = "printMemOperand";
let MIOperandInfo = (ops ptr_rc, i32imm);
let ParserMatchClass = SparcMEMriAsmOperand;
}
def TLSSym : Operand<iPTR>;
// Branch targets have OtherVT type.
def brtarget : Operand<OtherVT> {
let EncoderMethod = "getBranchTargetOpValue";
}
def calltarget : Operand<i32> {
let EncoderMethod = "getCallTargetOpValue";
}
// Operand for printing out a condition code.
let PrintMethod = "printCCOperand" in
def CCOp : Operand<i32>;
def SDTSPcmpicc :
SDTypeProfile<0, 2, [SDTCisInt<0>, SDTCisSameAs<0, 1>]>;
def SDTSPcmpfcc :
SDTypeProfile<0, 2, [SDTCisFP<0>, SDTCisSameAs<0, 1>]>;
def SDTSPbrcc :
SDTypeProfile<0, 2, [SDTCisVT<0, OtherVT>, SDTCisVT<1, i32>]>;
def SDTSPselectcc :
SDTypeProfile<1, 3, [SDTCisSameAs<0, 1>, SDTCisSameAs<1, 2>, SDTCisVT<3, i32>]>;
def SDTSPFTOI :
SDTypeProfile<1, 1, [SDTCisVT<0, f32>, SDTCisFP<1>]>;
def SDTSPITOF :
SDTypeProfile<1, 1, [SDTCisFP<0>, SDTCisVT<1, f32>]>;
def SDTSPFTOX :
SDTypeProfile<1, 1, [SDTCisVT<0, f64>, SDTCisFP<1>]>;
def SDTSPXTOF :
SDTypeProfile<1, 1, [SDTCisFP<0>, SDTCisVT<1, f64>]>;
def SDTSPtlsadd :
SDTypeProfile<1, 3, [SDTCisInt<0>, SDTCisSameAs<0, 1>, SDTCisPtrTy<2>]>;
def SDTSPtlsld :
SDTypeProfile<1, 2, [SDTCisPtrTy<0>, SDTCisPtrTy<1>]>;
def SPcmpicc : SDNode<"SPISD::CMPICC", SDTSPcmpicc, [SDNPOutGlue]>;
def SPcmpfcc : SDNode<"SPISD::CMPFCC", SDTSPcmpfcc, [SDNPOutGlue]>;
def SPbricc : SDNode<"SPISD::BRICC", SDTSPbrcc, [SDNPHasChain, SDNPInGlue]>;
def SPbrxcc : SDNode<"SPISD::BRXCC", SDTSPbrcc, [SDNPHasChain, SDNPInGlue]>;
def SPbrfcc : SDNode<"SPISD::BRFCC", SDTSPbrcc, [SDNPHasChain, SDNPInGlue]>;
def SPhi : SDNode<"SPISD::Hi", SDTIntUnaryOp>;
def SPlo : SDNode<"SPISD::Lo", SDTIntUnaryOp>;
def SPftoi : SDNode<"SPISD::FTOI", SDTSPFTOI>;
def SPitof : SDNode<"SPISD::ITOF", SDTSPITOF>;
def SPftox : SDNode<"SPISD::FTOX", SDTSPFTOX>;
def SPxtof : SDNode<"SPISD::XTOF", SDTSPXTOF>;
def SPselecticc : SDNode<"SPISD::SELECT_ICC", SDTSPselectcc, [SDNPInGlue]>;
def SPselectxcc : SDNode<"SPISD::SELECT_XCC", SDTSPselectcc, [SDNPInGlue]>;
def SPselectfcc : SDNode<"SPISD::SELECT_FCC", SDTSPselectcc, [SDNPInGlue]>;
// These are target-independent nodes, but have target-specific formats.
def SDT_SPCallSeqStart : SDCallSeqStart<[ SDTCisVT<0, i32> ]>;
def SDT_SPCallSeqEnd : SDCallSeqEnd<[ SDTCisVT<0, i32>,
SDTCisVT<1, i32> ]>;
def callseq_start : SDNode<"ISD::CALLSEQ_START", SDT_SPCallSeqStart,
[SDNPHasChain, SDNPOutGlue]>;
def callseq_end : SDNode<"ISD::CALLSEQ_END", SDT_SPCallSeqEnd,
[SDNPHasChain, SDNPOptInGlue, SDNPOutGlue]>;
def SDT_SPCall : SDTypeProfile<0, -1, [SDTCisVT<0, i32>]>;
def call : SDNode<"SPISD::CALL", SDT_SPCall,
[SDNPHasChain, SDNPOptInGlue, SDNPOutGlue,
SDNPVariadic]>;
def SDT_SPRet : SDTypeProfile<0, 1, [SDTCisVT<0, i32>]>;
def retflag : SDNode<"SPISD::RET_FLAG", SDT_SPRet,
[SDNPHasChain, SDNPOptInGlue, SDNPVariadic]>;
def flushw : SDNode<"SPISD::FLUSHW", SDTNone,
[SDNPHasChain, SDNPSideEffect, SDNPMayStore]>;
def tlsadd : SDNode<"SPISD::TLS_ADD", SDTSPtlsadd>;
def tlsld : SDNode<"SPISD::TLS_LD", SDTSPtlsld>;
def tlscall : SDNode<"SPISD::TLS_CALL", SDT_SPCall,
[SDNPHasChain, SDNPOptInGlue, SDNPOutGlue,
SDNPVariadic]>;
def getPCX : Operand<iPTR> {
let PrintMethod = "printGetPCX";
}
//===----------------------------------------------------------------------===//
// SPARC Flag Conditions
//===----------------------------------------------------------------------===//
// Note that these values must be kept in sync with the CCOp::CondCode enum
// values.
class ICC_VAL<int N> : PatLeaf<(i32 N)>;
def ICC_NE : ICC_VAL< 9>; // Not Equal
def ICC_E : ICC_VAL< 1>; // Equal
def ICC_G : ICC_VAL<10>; // Greater
def ICC_LE : ICC_VAL< 2>; // Less or Equal
def ICC_GE : ICC_VAL<11>; // Greater or Equal
def ICC_L : ICC_VAL< 3>; // Less
def ICC_GU : ICC_VAL<12>; // Greater Unsigned
def ICC_LEU : ICC_VAL< 4>; // Less or Equal Unsigned
def ICC_CC : ICC_VAL<13>; // Carry Clear/Great or Equal Unsigned
def ICC_CS : ICC_VAL< 5>; // Carry Set/Less Unsigned
def ICC_POS : ICC_VAL<14>; // Positive
def ICC_NEG : ICC_VAL< 6>; // Negative
def ICC_VC : ICC_VAL<15>; // Overflow Clear
def ICC_VS : ICC_VAL< 7>; // Overflow Set
class FCC_VAL<int N> : PatLeaf<(i32 N)>;
def FCC_U : FCC_VAL<23>; // Unordered
def FCC_G : FCC_VAL<22>; // Greater
def FCC_UG : FCC_VAL<21>; // Unordered or Greater
def FCC_L : FCC_VAL<20>; // Less
def FCC_UL : FCC_VAL<19>; // Unordered or Less
def FCC_LG : FCC_VAL<18>; // Less or Greater
def FCC_NE : FCC_VAL<17>; // Not Equal
def FCC_E : FCC_VAL<25>; // Equal
def FCC_UE : FCC_VAL<24>; // Unordered or Equal
def FCC_GE : FCC_VAL<25>; // Greater or Equal
def FCC_UGE : FCC_VAL<26>; // Unordered or Greater or Equal
def FCC_LE : FCC_VAL<27>; // Less or Equal
def FCC_ULE : FCC_VAL<28>; // Unordered or Less or Equal
def FCC_O : FCC_VAL<29>; // Ordered
//===----------------------------------------------------------------------===//
// Instruction Class Templates
//===----------------------------------------------------------------------===//
/// F3_12 multiclass - Define a normal F3_1/F3_2 pattern in one shot.
multiclass F3_12<string OpcStr, bits<6> Op3Val, SDNode OpNode,
RegisterClass RC, ValueType Ty, Operand immOp> {
def rr : F3_1<2, Op3Val,
(outs RC:$rd), (ins RC:$rs1, RC:$rs2),
!strconcat(OpcStr, " $rs1, $rs2, $rd"),
[(set Ty:$rd, (OpNode Ty:$rs1, Ty:$rs2))]>;
def ri : F3_2<2, Op3Val,
(outs RC:$rd), (ins RC:$rs1, immOp:$simm13),
!strconcat(OpcStr, " $rs1, $simm13, $rd"),
[(set Ty:$rd, (OpNode Ty:$rs1, (Ty simm13:$simm13)))]>;
}
/// F3_12np multiclass - Define a normal F3_1/F3_2 pattern in one shot, with no
/// pattern.
multiclass F3_12np<string OpcStr, bits<6> Op3Val> {
def rr : F3_1<2, Op3Val,
(outs IntRegs:$rd), (ins IntRegs:$rs1, IntRegs:$rs2),
!strconcat(OpcStr, " $rs1, $rs2, $rd"), []>;
def ri : F3_2<2, Op3Val,
(outs IntRegs:$rd), (ins IntRegs:$rs1, i32imm:$simm13),
!strconcat(OpcStr, " $rs1, $simm13, $rd"), []>;
}
// Load multiclass - Define both Reg+Reg/Reg+Imm patterns in one shot.
multiclass Load<string OpcStr, bits<6> Op3Val, SDPatternOperator OpNode,
RegisterClass RC, ValueType Ty> {
def rr : F3_1<3, Op3Val,
(outs RC:$dst), (ins MEMrr:$addr),
!strconcat(OpcStr, " [$addr], $dst"),
[(set Ty:$dst, (OpNode ADDRrr:$addr))]>;
def ri : F3_2<3, Op3Val,
(outs RC:$dst), (ins MEMri:$addr),
!strconcat(OpcStr, " [$addr], $dst"),
[(set Ty:$dst, (OpNode ADDRri:$addr))]>;
}
// Store multiclass - Define both Reg+Reg/Reg+Imm patterns in one shot.
multiclass Store<string OpcStr, bits<6> Op3Val, SDPatternOperator OpNode,
RegisterClass RC, ValueType Ty> {
def rr : F3_1<3, Op3Val,
(outs), (ins MEMrr:$addr, RC:$rd),
!strconcat(OpcStr, " $rd, [$addr]"),
[(OpNode Ty:$rd, ADDRrr:$addr)]>;
def ri : F3_2<3, Op3Val,
(outs), (ins MEMri:$addr, RC:$rd),
!strconcat(OpcStr, " $rd, [$addr]"),
[(OpNode Ty:$rd, ADDRri:$addr)]>;
}
//===----------------------------------------------------------------------===//
// Instructions
//===----------------------------------------------------------------------===//
// Pseudo instructions.
class Pseudo<dag outs, dag ins, string asmstr, list<dag> pattern>
: InstSP<outs, ins, asmstr, pattern> {
let isCodeGenOnly = 1;
let isPseudo = 1;
}
// GETPCX for PIC
let Defs = [O7] in {
def GETPCX : Pseudo<(outs getPCX:$getpcseq), (ins), "$getpcseq", [] >;
}
let Defs = [O6], Uses = [O6] in {
def ADJCALLSTACKDOWN : Pseudo<(outs), (ins i32imm:$amt),
"!ADJCALLSTACKDOWN $amt",
[(callseq_start timm:$amt)]>;
def ADJCALLSTACKUP : Pseudo<(outs), (ins i32imm:$amt1, i32imm:$amt2),
"!ADJCALLSTACKUP $amt1",
[(callseq_end timm:$amt1, timm:$amt2)]>;
}
let hasSideEffects = 1, mayStore = 1 in {
let rd = 0, rs1 = 0, rs2 = 0 in
def FLUSHW : F3_1<0b10, 0b101011, (outs), (ins),
"flushw",
[(flushw)]>, Requires<[HasV9]>;
let rd = 0, rs1 = 1, simm13 = 3 in
def TA3 : F3_2<0b10, 0b111010, (outs), (ins),
"ta 3",
[(flushw)]>;
}
let isBarrier = 1, isTerminator = 1, rd = 0b1000, rs1 = 0, simm13 = 5 in
def TA5 : F3_2<0b10, 0b111010, (outs), (ins), "ta 5", [(trap)]>;
let rd = 0 in
def UNIMP : F2_1<0b000, (outs), (ins i32imm:$val),
"unimp $val", []>;
// SELECT_CC_* - Used to implement the SELECT_CC DAG operation. Expanded after
// instruction selection into a branch sequence. This has to handle all
// permutations of selection between i32/f32/f64 on ICC and FCC.
// Expanded after instruction selection.
let Uses = [ICC], usesCustomInserter = 1 in {
def SELECT_CC_Int_ICC
: Pseudo<(outs IntRegs:$dst), (ins IntRegs:$T, IntRegs:$F, i32imm:$Cond),
"; SELECT_CC_Int_ICC PSEUDO!",
[(set i32:$dst, (SPselecticc i32:$T, i32:$F, imm:$Cond))]>;
def SELECT_CC_FP_ICC
: Pseudo<(outs FPRegs:$dst), (ins FPRegs:$T, FPRegs:$F, i32imm:$Cond),
"; SELECT_CC_FP_ICC PSEUDO!",
[(set f32:$dst, (SPselecticc f32:$T, f32:$F, imm:$Cond))]>;
def SELECT_CC_DFP_ICC
: Pseudo<(outs DFPRegs:$dst), (ins DFPRegs:$T, DFPRegs:$F, i32imm:$Cond),
"; SELECT_CC_DFP_ICC PSEUDO!",
[(set f64:$dst, (SPselecticc f64:$T, f64:$F, imm:$Cond))]>;
def SELECT_CC_QFP_ICC
: Pseudo<(outs QFPRegs:$dst), (ins QFPRegs:$T, QFPRegs:$F, i32imm:$Cond),
"; SELECT_CC_QFP_ICC PSEUDO!",
[(set f128:$dst, (SPselecticc f128:$T, f128:$F, imm:$Cond))]>;
}
let usesCustomInserter = 1, Uses = [FCC] in {
def SELECT_CC_Int_FCC
: Pseudo<(outs IntRegs:$dst), (ins IntRegs:$T, IntRegs:$F, i32imm:$Cond),
"; SELECT_CC_Int_FCC PSEUDO!",
[(set i32:$dst, (SPselectfcc i32:$T, i32:$F, imm:$Cond))]>;
def SELECT_CC_FP_FCC
: Pseudo<(outs FPRegs:$dst), (ins FPRegs:$T, FPRegs:$F, i32imm:$Cond),
"; SELECT_CC_FP_FCC PSEUDO!",
[(set f32:$dst, (SPselectfcc f32:$T, f32:$F, imm:$Cond))]>;
def SELECT_CC_DFP_FCC
: Pseudo<(outs DFPRegs:$dst), (ins DFPRegs:$T, DFPRegs:$F, i32imm:$Cond),
"; SELECT_CC_DFP_FCC PSEUDO!",
[(set f64:$dst, (SPselectfcc f64:$T, f64:$F, imm:$Cond))]>;
def SELECT_CC_QFP_FCC
: Pseudo<(outs QFPRegs:$dst), (ins QFPRegs:$T, QFPRegs:$F, i32imm:$Cond),
"; SELECT_CC_QFP_FCC PSEUDO!",
[(set f128:$dst, (SPselectfcc f128:$T, f128:$F, imm:$Cond))]>;
}
// JMPL Instruction.
let isTerminator = 1, hasDelaySlot = 1, isBarrier = 1 in {
def JMPLrr: F3_1<2, 0b111000, (outs IntRegs:$dst), (ins MEMrr:$addr),
"jmpl $addr, $dst", []>;
def JMPLri: F3_2<2, 0b111000, (outs IntRegs:$dst), (ins MEMri:$addr),
"jmpl $addr, $dst", []>;
}
// Section A.3 - Synthetic Instructions, p. 85
// special cases of JMPL:
let isReturn = 1, isTerminator = 1, hasDelaySlot = 1, isBarrier = 1,
isCodeGenOnly = 1 in {
let rd = 0, rs1 = 15 in
def RETL: F3_2<2, 0b111000, (outs), (ins i32imm:$val),
"jmp %o7+$val", [(retflag simm13:$val)]>;
let rd = 0, rs1 = 31 in
def RET: F3_2<2, 0b111000, (outs), (ins i32imm:$val),
"jmp %i7+$val", []>;
}
// Section B.1 - Load Integer Instructions, p. 90
defm LDSB : Load<"ldsb", 0b001001, sextloadi8, IntRegs, i32>;
defm LDSH : Load<"ldsh", 0b001010, sextloadi16, IntRegs, i32>;
defm LDUB : Load<"ldub", 0b000001, zextloadi8, IntRegs, i32>;
defm LDUH : Load<"lduh", 0b000010, zextloadi16, IntRegs, i32>;
defm LD : Load<"ld", 0b000000, load, IntRegs, i32>;
// Section B.2 - Load Floating-point Instructions, p. 92
defm LDF : Load<"ld", 0b100000, load, FPRegs, f32>;
defm LDDF : Load<"ldd", 0b100011, load, DFPRegs, f64>;
defm LDQF : Load<"ldq", 0b100010, load, QFPRegs, f128>,
Requires<[HasV9, HasHardQuad]>;
// Section B.4 - Store Integer Instructions, p. 95
defm STB : Store<"stb", 0b000101, truncstorei8, IntRegs, i32>;
defm STH : Store<"sth", 0b000110, truncstorei16, IntRegs, i32>;
defm ST : Store<"st", 0b000100, store, IntRegs, i32>;
// Section B.5 - Store Floating-point Instructions, p. 97
defm STF : Store<"st", 0b100100, store, FPRegs, f32>;
defm STDF : Store<"std", 0b100111, store, DFPRegs, f64>;
defm STQF : Store<"stq", 0b100110, store, QFPRegs, f128>,
Requires<[HasV9, HasHardQuad]>;
// Section B.9 - SETHI Instruction, p. 104
def SETHIi: F2_1<0b100,
(outs IntRegs:$rd), (ins i32imm:$imm22),
"sethi $imm22, $rd",
[(set i32:$rd, SETHIimm:$imm22)]>;
// Section B.10 - NOP Instruction, p. 105
// (It's a special case of SETHI)
let rd = 0, imm22 = 0 in
def NOP : F2_1<0b100, (outs), (ins), "nop", []>;
// Section B.11 - Logical Instructions, p. 106
defm AND : F3_12<"and", 0b000001, and, IntRegs, i32, i32imm>;
def ANDNrr : F3_1<2, 0b000101,
(outs IntRegs:$rd), (ins IntRegs:$rs1, IntRegs:$rs2),
"andn $rs1, $rs2, $rd",
[(set i32:$rd, (and i32:$rs1, (not i32:$rs2)))]>;
def ANDNri : F3_2<2, 0b000101,
(outs IntRegs:$rd), (ins IntRegs:$rs1, i32imm:$simm13),
"andn $rs1, $simm13, $rd", []>;
defm OR : F3_12<"or", 0b000010, or, IntRegs, i32, i32imm>;
def ORNrr : F3_1<2, 0b000110,
(outs IntRegs:$rd), (ins IntRegs:$rs1, IntRegs:$rs2),
"orn $rs1, $rs2, $rd",
[(set i32:$rd, (or i32:$rs1, (not i32:$rs2)))]>;
def ORNri : F3_2<2, 0b000110,
(outs IntRegs:$rd), (ins IntRegs:$rs1, i32imm:$simm13),
"orn $rs1, $simm13, $rd", []>;
defm XOR : F3_12<"xor", 0b000011, xor, IntRegs, i32, i32imm>;
def XNORrr : F3_1<2, 0b000111,
(outs IntRegs:$rd), (ins IntRegs:$rs1, IntRegs:$rs2),
"xnor $rs1, $rs2, $rd",
[(set i32:$rd, (not (xor i32:$rs1, i32:$rs2)))]>;
def XNORri : F3_2<2, 0b000111,
(outs IntRegs:$rd), (ins IntRegs:$rs1, i32imm:$simm13),
"xnor $rs1, $simm13, $rd", []>;
// Section B.12 - Shift Instructions, p. 107
defm SLL : F3_12<"sll", 0b100101, shl, IntRegs, i32, i32imm>;
defm SRL : F3_12<"srl", 0b100110, srl, IntRegs, i32, i32imm>;
defm SRA : F3_12<"sra", 0b100111, sra, IntRegs, i32, i32imm>;
// Section B.13 - Add Instructions, p. 108
defm ADD : F3_12<"add", 0b000000, add, IntRegs, i32, i32imm>;
// "LEA" forms of add (patterns to make tblgen happy)
let Predicates = [Is32Bit], isCodeGenOnly = 1 in
def LEA_ADDri : F3_2<2, 0b000000,
(outs IntRegs:$dst), (ins MEMri:$addr),
"add ${addr:arith}, $dst",
[(set iPTR:$dst, ADDRri:$addr)]>;
let Defs = [ICC] in
defm ADDCC : F3_12<"addcc", 0b010000, addc, IntRegs, i32, i32imm>;
let Uses = [ICC], Defs = [ICC] in
defm ADDE : F3_12<"addxcc", 0b011000, adde, IntRegs, i32, i32imm>;
// Section B.15 - Subtract Instructions, p. 110
defm SUB : F3_12 <"sub" , 0b000100, sub, IntRegs, i32, i32imm>;
let Uses = [ICC], Defs = [ICC] in
defm SUBE : F3_12 <"subxcc" , 0b011100, sube, IntRegs, i32, i32imm>;
let Defs = [ICC] in
defm SUBCC : F3_12 <"subcc", 0b010100, subc, IntRegs, i32, i32imm>;
let Defs = [ICC], rd = 0 in {
def CMPrr : F3_1<2, 0b010100,
(outs), (ins IntRegs:$rs1, IntRegs:$rs2),
"cmp $rs1, $rs2",
[(SPcmpicc i32:$rs1, i32:$rs2)]>;
def CMPri : F3_2<2, 0b010100,
(outs), (ins IntRegs:$rs1, i32imm:$simm13),
"cmp $rs1, $simm13",
[(SPcmpicc i32:$rs1, (i32 simm13:$simm13))]>;
}
// Section B.18 - Multiply Instructions, p. 113
let Defs = [Y] in {
defm UMUL : F3_12np<"umul", 0b001010>;
defm SMUL : F3_12 <"smul", 0b001011, mul, IntRegs, i32, i32imm>;
}
// Section B.19 - Divide Instructions, p. 115
let Defs = [Y] in {
defm UDIV : F3_12np<"udiv", 0b001110>;
defm SDIV : F3_12np<"sdiv", 0b001111>;
}
// Section B.20 - SAVE and RESTORE, p. 117
defm SAVE : F3_12np<"save" , 0b111100>;
defm RESTORE : F3_12np<"restore", 0b111101>;
// Section B.21 - Branch on Integer Condition Codes Instructions, p. 119
// unconditional branch class.
class BranchAlways<dag ins, string asmstr, list<dag> pattern>
: F2_2<0b010, (outs), ins, asmstr, pattern> {
let isBranch = 1;
let isTerminator = 1;
let hasDelaySlot = 1;
let isBarrier = 1;
}
let cond = 8 in
def BA : BranchAlways<(ins brtarget:$imm22), "ba $imm22", [(br bb:$imm22)]>;
// conditional branch class:
class BranchSP<dag ins, string asmstr, list<dag> pattern>
: F2_2<0b010, (outs), ins, asmstr, pattern> {
let isBranch = 1;
let isTerminator = 1;
let hasDelaySlot = 1;
}
// Indirect branch instructions.
let isTerminator = 1, isBarrier = 1, hasDelaySlot = 1, isBranch =1,
isIndirectBranch = 1, rd = 0, isCodeGenOnly = 1 in {
def BINDrr : F3_1<2, 0b111000,
(outs), (ins MEMrr:$ptr),
"jmp $ptr",
[(brind ADDRrr:$ptr)]>;
def BINDri : F3_2<2, 0b111000,
(outs), (ins MEMri:$ptr),
"jmp $ptr",
[(brind ADDRri:$ptr)]>;
}
let Uses = [ICC] in
def BCOND : BranchSP<(ins brtarget:$imm22, CCOp:$cond),
"b$cond $imm22",
[(SPbricc bb:$imm22, imm:$cond)]>;
// Section B.22 - Branch on Floating-point Condition Codes Instructions, p. 121
// floating-point conditional branch class:
class FPBranchSP<dag ins, string asmstr, list<dag> pattern>
: F2_2<0b110, (outs), ins, asmstr, pattern> {
let isBranch = 1;
let isTerminator = 1;
let hasDelaySlot = 1;
}
let Uses = [FCC] in
def FBCOND : FPBranchSP<(ins brtarget:$imm22, CCOp:$cond),
"fb$cond $imm22",
[(SPbrfcc bb:$imm22, imm:$cond)]>;
// Section B.24 - Call and Link Instruction, p. 125
// This is the only Format 1 instruction
let Uses = [O6],
hasDelaySlot = 1, isCall = 1 in {
def CALL : InstSP<(outs), (ins calltarget:$dst, variable_ops),
"call $dst", []> {
bits<30> disp;
let op = 1;
let Inst{29-0} = disp;
}
// indirect calls: special cases of JMPL.
let isCodeGenOnly = 1, rd = 15 in {
def CALLrr : F3_1<2, 0b111000,
(outs), (ins MEMrr:$ptr, variable_ops),
"call $ptr",
[(call ADDRrr:$ptr)]>;
def CALLri : F3_2<2, 0b111000,
(outs), (ins MEMri:$ptr, variable_ops),
"call $ptr",
[(call ADDRri:$ptr)]>;
}
}
// Section B.28 - Read State Register Instructions
let Uses = [Y], rs1 = 0, rs2 = 0 in
def RDY : F3_1<2, 0b101000,
(outs IntRegs:$dst), (ins),
"rd %y, $dst", []>;
// Section B.29 - Write State Register Instructions
let Defs = [Y], rd = 0 in {
def WRYrr : F3_1<2, 0b110000,
(outs), (ins IntRegs:$b, IntRegs:$c),
"wr $b, $c, %y", []>;
def WRYri : F3_2<2, 0b110000,
(outs), (ins IntRegs:$b, i32imm:$c),
"wr $b, $c, %y", []>;
}
// Convert Integer to Floating-point Instructions, p. 141
def FITOS : F3_3u<2, 0b110100, 0b011000100,
(outs FPRegs:$rd), (ins FPRegs:$rs2),
"fitos $rs2, $rd",
[(set FPRegs:$rd, (SPitof FPRegs:$rs2))]>;
def FITOD : F3_3u<2, 0b110100, 0b011001000,
(outs DFPRegs:$rd), (ins FPRegs:$rs2),
"fitod $rs2, $rd",
[(set DFPRegs:$rd, (SPitof FPRegs:$rs2))]>;
def FITOQ : F3_3u<2, 0b110100, 0b011001100,
(outs QFPRegs:$rd), (ins FPRegs:$rs2),
"fitoq $rs2, $rd",
[(set QFPRegs:$rd, (SPitof FPRegs:$rs2))]>,
Requires<[HasHardQuad]>;
// Convert Floating-point to Integer Instructions, p. 142
def FSTOI : F3_3u<2, 0b110100, 0b011010001,
(outs FPRegs:$rd), (ins FPRegs:$rs2),
"fstoi $rs2, $rd",
[(set FPRegs:$rd, (SPftoi FPRegs:$rs2))]>;
def FDTOI : F3_3u<2, 0b110100, 0b011010010,
(outs FPRegs:$rd), (ins DFPRegs:$rs2),
"fdtoi $rs2, $rd",
[(set FPRegs:$rd, (SPftoi DFPRegs:$rs2))]>;
def FQTOI : F3_3u<2, 0b110100, 0b011010011,
(outs FPRegs:$rd), (ins QFPRegs:$rs2),
"fqtoi $rs2, $rd",
[(set FPRegs:$rd, (SPftoi QFPRegs:$rs2))]>,
Requires<[HasHardQuad]>;
// Convert between Floating-point Formats Instructions, p. 143
def FSTOD : F3_3u<2, 0b110100, 0b011001001,
(outs DFPRegs:$rd), (ins FPRegs:$rs2),
"fstod $rs2, $rd",
[(set f64:$rd, (fextend f32:$rs2))]>;
def FSTOQ : F3_3u<2, 0b110100, 0b011001101,
(outs QFPRegs:$rd), (ins FPRegs:$rs2),
"fstoq $rs2, $rd",
[(set f128:$rd, (fextend f32:$rs2))]>,
Requires<[HasHardQuad]>;
def FDTOS : F3_3u<2, 0b110100, 0b011000110,
(outs FPRegs:$rd), (ins DFPRegs:$rs2),
"fdtos $rs2, $rd",
[(set f32:$rd, (fround f64:$rs2))]>;
def FDTOQ : F3_3u<2, 0b110100, 0b011001110,
(outs QFPRegs:$rd), (ins DFPRegs:$rs2),
"fdtoq $rs2, $rd",
[(set f128:$rd, (fextend f64:$rs2))]>,
Requires<[HasHardQuad]>;
def FQTOS : F3_3u<2, 0b110100, 0b011000111,
(outs FPRegs:$rd), (ins QFPRegs:$rs2),
"fqtos $rs2, $rd",
[(set f32:$rd, (fround f128:$rs2))]>,
Requires<[HasHardQuad]>;
def FQTOD : F3_3u<2, 0b110100, 0b011001011,
(outs DFPRegs:$rd), (ins QFPRegs:$rs2),
"fqtod $rs2, $rd",
[(set f64:$rd, (fround f128:$rs2))]>,
Requires<[HasHardQuad]>;
// Floating-point Move Instructions, p. 144
def FMOVS : F3_3u<2, 0b110100, 0b000000001,
(outs FPRegs:$rd), (ins FPRegs:$rs2),
"fmovs $rs2, $rd", []>;
def FNEGS : F3_3u<2, 0b110100, 0b000000101,
(outs FPRegs:$rd), (ins FPRegs:$rs2),
"fnegs $rs2, $rd",
[(set f32:$rd, (fneg f32:$rs2))]>;
def FABSS : F3_3u<2, 0b110100, 0b000001001,
(outs FPRegs:$rd), (ins FPRegs:$rs2),
"fabss $rs2, $rd",
[(set f32:$rd, (fabs f32:$rs2))]>;
// Floating-point Square Root Instructions, p.145
def FSQRTS : F3_3u<2, 0b110100, 0b000101001,
(outs FPRegs:$rd), (ins FPRegs:$rs2),
"fsqrts $rs2, $rd",
[(set f32:$rd, (fsqrt f32:$rs2))]>;
def FSQRTD : F3_3u<2, 0b110100, 0b000101010,
(outs DFPRegs:$rd), (ins DFPRegs:$rs2),
"fsqrtd $rs2, $rd",
[(set f64:$rd, (fsqrt f64:$rs2))]>;
def FSQRTQ : F3_3u<2, 0b110100, 0b000101011,
(outs QFPRegs:$rd), (ins QFPRegs:$rs2),
"fsqrtq $rs2, $rd",
[(set f128:$rd, (fsqrt f128:$rs2))]>,
Requires<[HasHardQuad]>;
// Floating-point Add and Subtract Instructions, p. 146
def FADDS : F3_3<2, 0b110100, 0b001000001,
(outs FPRegs:$rd), (ins FPRegs:$rs1, FPRegs:$rs2),
"fadds $rs1, $rs2, $rd",
[(set f32:$rd, (fadd f32:$rs1, f32:$rs2))]>;
def FADDD : F3_3<2, 0b110100, 0b001000010,
(outs DFPRegs:$rd), (ins DFPRegs:$rs1, DFPRegs:$rs2),
"faddd $rs1, $rs2, $rd",
[(set f64:$rd, (fadd f64:$rs1, f64:$rs2))]>;
def FADDQ : F3_3<2, 0b110100, 0b001000011,
(outs QFPRegs:$rd), (ins QFPRegs:$rs1, QFPRegs:$rs2),
"faddq $rs1, $rs2, $rd",
[(set f128:$rd, (fadd f128:$rs1, f128:$rs2))]>,
Requires<[HasHardQuad]>;
def FSUBS : F3_3<2, 0b110100, 0b001000101,
(outs FPRegs:$rd), (ins FPRegs:$rs1, FPRegs:$rs2),
"fsubs $rs1, $rs2, $rd",
[(set f32:$rd, (fsub f32:$rs1, f32:$rs2))]>;
def FSUBD : F3_3<2, 0b110100, 0b001000110,
(outs DFPRegs:$rd), (ins DFPRegs:$rs1, DFPRegs:$rs2),
"fsubd $rs1, $rs2, $rd",
[(set f64:$rd, (fsub f64:$rs1, f64:$rs2))]>;
def FSUBQ : F3_3<2, 0b110100, 0b001000111,
(outs QFPRegs:$rd), (ins QFPRegs:$rs1, QFPRegs:$rs2),
"fsubq $rs1, $rs2, $rd",
[(set f128:$rd, (fsub f128:$rs1, f128:$rs2))]>,
Requires<[HasHardQuad]>;
// Floating-point Multiply and Divide Instructions, p. 147
def FMULS : F3_3<2, 0b110100, 0b001001001,
(outs FPRegs:$rd), (ins FPRegs:$rs1, FPRegs:$rs2),
"fmuls $rs1, $rs2, $rd",
[(set f32:$rd, (fmul f32:$rs1, f32:$rs2))]>;
def FMULD : F3_3<2, 0b110100, 0b001001010,
(outs DFPRegs:$rd), (ins DFPRegs:$rs1, DFPRegs:$rs2),
"fmuld $rs1, $rs2, $rd",
[(set f64:$rd, (fmul f64:$rs1, f64:$rs2))]>;
def FMULQ : F3_3<2, 0b110100, 0b001001011,
(outs QFPRegs:$rd), (ins QFPRegs:$rs1, QFPRegs:$rs2),
"fmulq $rs1, $rs2, $rd",
[(set f128:$rd, (fmul f128:$rs1, f128:$rs2))]>,
Requires<[HasHardQuad]>;
def FSMULD : F3_3<2, 0b110100, 0b001101001,
(outs DFPRegs:$rd), (ins FPRegs:$rs1, FPRegs:$rs2),
"fsmuld $rs1, $rs2, $rd",
[(set f64:$rd, (fmul (fextend f32:$rs1),
(fextend f32:$rs2)))]>;
def FDMULQ : F3_3<2, 0b110100, 0b001101110,
(outs QFPRegs:$rd), (ins DFPRegs:$rs1, DFPRegs:$rs2),
"fdmulq $rs1, $rs2, $rd",
[(set f128:$rd, (fmul (fextend f64:$rs1),
(fextend f64:$rs2)))]>,
Requires<[HasHardQuad]>;
def FDIVS : F3_3<2, 0b110100, 0b001001101,
(outs FPRegs:$rd), (ins FPRegs:$rs1, FPRegs:$rs2),
"fdivs $rs1, $rs2, $rd",
[(set f32:$rd, (fdiv f32:$rs1, f32:$rs2))]>;
def FDIVD : F3_3<2, 0b110100, 0b001001110,
(outs DFPRegs:$rd), (ins DFPRegs:$rs1, DFPRegs:$rs2),
"fdivd $rs1, $rs2, $rd",
[(set f64:$rd, (fdiv f64:$rs1, f64:$rs2))]>;
def FDIVQ : F3_3<2, 0b110100, 0b001001111,
(outs QFPRegs:$rd), (ins QFPRegs:$rs1, QFPRegs:$rs2),
"fdivq $rs1, $rs2, $rd",
[(set f128:$rd, (fdiv f128:$rs1, f128:$rs2))]>,
Requires<[HasHardQuad]>;
// Floating-point Compare Instructions, p. 148
// Note: the 2nd template arg is different for these guys.
// Note 2: the result of a FCMP is not available until the 2nd cycle
// after the instr is retired, but there is no interlock in Sparc V8.
// This behavior is modeled with a forced noop after the instruction in
// DelaySlotFiller.
let Defs = [FCC] in {
def FCMPS : F3_3c<2, 0b110101, 0b001010001,
(outs), (ins FPRegs:$rs1, FPRegs:$rs2),
"fcmps $rs1, $rs2",
[(SPcmpfcc f32:$rs1, f32:$rs2)]>;
def FCMPD : F3_3c<2, 0b110101, 0b001010010,
(outs), (ins DFPRegs:$rs1, DFPRegs:$rs2),
"fcmpd $rs1, $rs2",
[(SPcmpfcc f64:$rs1, f64:$rs2)]>;
def FCMPQ : F3_3c<2, 0b110101, 0b001010011,
(outs), (ins QFPRegs:$rs1, QFPRegs:$rs2),
"fcmpq $rs1, $rs2",
[(SPcmpfcc f128:$rs1, f128:$rs2)]>,
Requires<[HasHardQuad]>;
}
//===----------------------------------------------------------------------===//
// Instructions for Thread Local Storage(TLS).
//===----------------------------------------------------------------------===//
let isCodeGenOnly = 1, isAsmParserOnly = 1 in {
def TLS_ADDrr : F3_1<2, 0b000000,
(outs IntRegs:$rd),
(ins IntRegs:$rs1, IntRegs:$rs2, TLSSym:$sym),
"add $rs1, $rs2, $rd, $sym",
[(set i32:$rd,
(tlsadd i32:$rs1, i32:$rs2, tglobaltlsaddr:$sym))]>;
let mayLoad = 1 in
def TLS_LDrr : F3_1<3, 0b000000,
(outs IntRegs:$dst), (ins MEMrr:$addr, TLSSym:$sym),
"ld [$addr], $dst, $sym",
[(set i32:$dst,
(tlsld ADDRrr:$addr, tglobaltlsaddr:$sym))]>;
let Uses = [O6], isCall = 1, hasDelaySlot = 1 in
def TLS_CALL : InstSP<(outs),
(ins calltarget:$disp, TLSSym:$sym, variable_ops),
"call $disp, $sym",
[(tlscall texternalsym:$disp, tglobaltlsaddr:$sym)]> {
bits<30> disp;
let op = 1;
let Inst{29-0} = disp;
}
}
//===----------------------------------------------------------------------===//
// V9 Instructions
//===----------------------------------------------------------------------===//
// V9 Conditional Moves.
let Predicates = [HasV9], Constraints = "$f = $rd" in {
// Move Integer Register on Condition (MOVcc) p. 194 of the V9 manual.
let Uses = [ICC], cc = 0b100 in {
def MOVICCrr
: F4_1<0b101100, (outs IntRegs:$rd),
(ins IntRegs:$rs2, IntRegs:$f, CCOp:$cond),
"mov$cond %icc, $rs2, $rd",
[(set i32:$rd, (SPselecticc i32:$rs2, i32:$f, imm:$cond))]>;
def MOVICCri
: F4_2<0b101100, (outs IntRegs:$rd),
(ins i32imm:$simm11, IntRegs:$f, CCOp:$cond),
"mov$cond %icc, $simm11, $rd",
[(set i32:$rd,
(SPselecticc simm11:$simm11, i32:$f, imm:$cond))]>;
}
let Uses = [FCC], cc = 0b000 in {
def MOVFCCrr
: F4_1<0b101100, (outs IntRegs:$rd),
(ins IntRegs:$rs2, IntRegs:$f, CCOp:$cond),
"mov$cond %fcc0, $rs2, $rd",
[(set i32:$rd, (SPselectfcc i32:$rs2, i32:$f, imm:$cond))]>;
def MOVFCCri
: F4_2<0b101100, (outs IntRegs:$rd),
(ins i32imm:$simm11, IntRegs:$f, CCOp:$cond),
"mov$cond %fcc0, $simm11, $rd",
[(set i32:$rd,
(SPselectfcc simm11:$simm11, i32:$f, imm:$cond))]>;
}
let Uses = [ICC], opf_cc = 0b100 in {
def FMOVS_ICC
: F4_3<0b110101, 0b000001, (outs FPRegs:$rd),
(ins FPRegs:$rs2, FPRegs:$f, CCOp:$cond),
"fmovs$cond %icc, $rs2, $rd",
[(set f32:$rd, (SPselecticc f32:$rs2, f32:$f, imm:$cond))]>;
def FMOVD_ICC
: F4_3<0b110101, 0b000010, (outs DFPRegs:$rd),
(ins DFPRegs:$rs2, DFPRegs:$f, CCOp:$cond),
"fmovd$cond %icc, $rs2, $rd",
[(set f64:$rd, (SPselecticc f64:$rs2, f64:$f, imm:$cond))]>;
def FMOVQ_ICC
: F4_3<0b110101, 0b000011, (outs QFPRegs:$rd),
(ins QFPRegs:$rs2, QFPRegs:$f, CCOp:$cond),
"fmovq$cond %icc, $rs2, $rd",
[(set f128:$rd, (SPselecticc f128:$rs2, f128:$f, imm:$cond))]>,
Requires<[HasHardQuad]>;
}
let Uses = [FCC], opf_cc = 0b000 in {
def FMOVS_FCC
: F4_3<0b110101, 0b000001, (outs FPRegs:$rd),
(ins FPRegs:$rs2, FPRegs:$f, CCOp:$cond),
"fmovs$cond %fcc0, $rs2, $rd",
[(set f32:$rd, (SPselectfcc f32:$rs2, f32:$f, imm:$cond))]>;
def FMOVD_FCC
: F4_3<0b110101, 0b000010, (outs DFPRegs:$rd),
(ins DFPRegs:$rs2, DFPRegs:$f, CCOp:$cond),
"fmovd$cond %fcc0, $rs2, $rd",
[(set f64:$rd, (SPselectfcc f64:$rs2, f64:$f, imm:$cond))]>;
def FMOVQ_FCC
: F4_3<0b110101, 0b000011, (outs QFPRegs:$rd),
(ins QFPRegs:$rs2, QFPRegs:$f, CCOp:$cond),
"fmovq$cond %fcc0, $rs2, $rd",
[(set f128:$rd, (SPselectfcc f128:$rs2, f128:$f, imm:$cond))]>,
Requires<[HasHardQuad]>;
}
}
// Floating-Point Move Instructions, p. 164 of the V9 manual.
let Predicates = [HasV9] in {
def FMOVD : F3_3u<2, 0b110100, 0b000000010,
(outs DFPRegs:$rd), (ins DFPRegs:$rs2),
"fmovd $rs2, $rd", []>;
def FMOVQ : F3_3u<2, 0b110100, 0b000000011,
(outs QFPRegs:$rd), (ins QFPRegs:$rs2),
"fmovq $rs2, $rd", []>,
Requires<[HasHardQuad]>;
def FNEGD : F3_3u<2, 0b110100, 0b000000110,
(outs DFPRegs:$rd), (ins DFPRegs:$rs2),
"fnegd $rs2, $rd",
[(set f64:$rd, (fneg f64:$rs2))]>;
def FNEGQ : F3_3u<2, 0b110100, 0b000000111,
(outs QFPRegs:$rd), (ins QFPRegs:$rs2),
"fnegq $rs2, $rd",
[(set f128:$rd, (fneg f128:$rs2))]>,
Requires<[HasHardQuad]>;
def FABSD : F3_3u<2, 0b110100, 0b000001010,
(outs DFPRegs:$rd), (ins DFPRegs:$rs2),
"fabsd $rs2, $rd",
[(set f64:$rd, (fabs f64:$rs2))]>;
def FABSQ : F3_3u<2, 0b110100, 0b000001011,
(outs QFPRegs:$rd), (ins QFPRegs:$rs2),
"fabsq $rs2, $rd",
[(set f128:$rd, (fabs f128:$rs2))]>,
Requires<[HasHardQuad]>;
}
// POPCrr - This does a ctpop of a 64-bit register. As such, we have to clear
// the top 32-bits before using it. To do this clearing, we use a SRLri X,0.
let rs1 = 0 in
def POPCrr : F3_1<2, 0b101110,
(outs IntRegs:$dst), (ins IntRegs:$src),
"popc $src, $dst", []>, Requires<[HasV9]>;
def : Pat<(ctpop i32:$src),
(POPCrr (SRLri $src, 0))>;
// Atomic swap.
let hasSideEffects =1, rd = 0, rs1 = 0b01111, rs2 = 0 in
def STBAR : F3_1<2, 0b101000, (outs), (ins), "stbar", []>;
let Predicates = [HasV9], hasSideEffects = 1, rd = 0, rs1 = 0b01111 in
def MEMBARi : F3_2<2, 0b101000, (outs), (ins i32imm:$simm13),
"membar $simm13", []>;
let Constraints = "$val = $dst" in {
def SWAPrr : F3_1<3, 0b001111,
(outs IntRegs:$dst), (ins MEMrr:$addr, IntRegs:$val),
"swap [$addr], $dst",
[(set i32:$dst, (atomic_swap_32 ADDRrr:$addr, i32:$val))]>;
def SWAPri : F3_2<3, 0b001111,
(outs IntRegs:$dst), (ins MEMri:$addr, IntRegs:$val),
"swap [$addr], $dst",
[(set i32:$dst, (atomic_swap_32 ADDRri:$addr, i32:$val))]>;
}
let Predicates = [HasV9], Constraints = "$swap = $rd" in
def CASrr: F3_1_asi<3, 0b111100, 0b10000000,
(outs IntRegs:$rd), (ins IntRegs:$rs1, IntRegs:$rs2,
IntRegs:$swap),
"cas [$rs1], $rs2, $rd",
[(set i32:$rd,
(atomic_cmp_swap iPTR:$rs1, i32:$rs2, i32:$swap))]>;
//===----------------------------------------------------------------------===//
// Non-Instruction Patterns
//===----------------------------------------------------------------------===//
// Small immediates.
def : Pat<(i32 simm13:$val),
(ORri (i32 G0), imm:$val)>;
// Arbitrary immediates.
def : Pat<(i32 imm:$val),
(ORri (SETHIi (HI22 imm:$val)), (LO10 imm:$val))>;
// Global addresses, constant pool entries
let Predicates = [Is32Bit] in {
def : Pat<(SPhi tglobaladdr:$in), (SETHIi tglobaladdr:$in)>;
def : Pat<(SPlo tglobaladdr:$in), (ORri (i32 G0), tglobaladdr:$in)>;
def : Pat<(SPhi tconstpool:$in), (SETHIi tconstpool:$in)>;
def : Pat<(SPlo tconstpool:$in), (ORri (i32 G0), tconstpool:$in)>;
// GlobalTLS addresses
def : Pat<(SPhi tglobaltlsaddr:$in), (SETHIi tglobaltlsaddr:$in)>;
def : Pat<(SPlo tglobaltlsaddr:$in), (ORri (i32 G0), tglobaltlsaddr:$in)>;
def : Pat<(add (SPhi tglobaltlsaddr:$in1), (SPlo tglobaltlsaddr:$in2)),
(ADDri (SETHIi tglobaltlsaddr:$in1), (tglobaltlsaddr:$in2))>;
def : Pat<(xor (SPhi tglobaltlsaddr:$in1), (SPlo tglobaltlsaddr:$in2)),
(XORri (SETHIi tglobaltlsaddr:$in1), (tglobaltlsaddr:$in2))>;
// Blockaddress
def : Pat<(SPhi tblockaddress:$in), (SETHIi tblockaddress:$in)>;
def : Pat<(SPlo tblockaddress:$in), (ORri (i32 G0), tblockaddress:$in)>;
// Add reg, lo. This is used when taking the addr of a global/constpool entry.
def : Pat<(add iPTR:$r, (SPlo tglobaladdr:$in)), (ADDri $r, tglobaladdr:$in)>;
def : Pat<(add iPTR:$r, (SPlo tconstpool:$in)), (ADDri $r, tconstpool:$in)>;
def : Pat<(add iPTR:$r, (SPlo tblockaddress:$in)),
(ADDri $r, tblockaddress:$in)>;
}
// Calls:
def : Pat<(call tglobaladdr:$dst),
(CALL tglobaladdr:$dst)>;
def : Pat<(call texternalsym:$dst),
(CALL texternalsym:$dst)>;
// Map integer extload's to zextloads.
def : Pat<(i32 (extloadi1 ADDRrr:$src)), (LDUBrr ADDRrr:$src)>;
def : Pat<(i32 (extloadi1 ADDRri:$src)), (LDUBri ADDRri:$src)>;
def : Pat<(i32 (extloadi8 ADDRrr:$src)), (LDUBrr ADDRrr:$src)>;
def : Pat<(i32 (extloadi8 ADDRri:$src)), (LDUBri ADDRri:$src)>;
def : Pat<(i32 (extloadi16 ADDRrr:$src)), (LDUHrr ADDRrr:$src)>;
def : Pat<(i32 (extloadi16 ADDRri:$src)), (LDUHri ADDRri:$src)>;
// zextload bool -> zextload byte
def : Pat<(i32 (zextloadi1 ADDRrr:$src)), (LDUBrr ADDRrr:$src)>;
def : Pat<(i32 (zextloadi1 ADDRri:$src)), (LDUBri ADDRri:$src)>;
// store 0, addr -> store %g0, addr
def : Pat<(store (i32 0), ADDRrr:$dst), (STrr ADDRrr:$dst, (i32 G0))>;
def : Pat<(store (i32 0), ADDRri:$dst), (STri ADDRri:$dst, (i32 G0))>;
// store bar for all atomic_fence in V8.
let Predicates = [HasNoV9] in
def : Pat<(atomic_fence imm, imm), (STBAR)>;
// atomic_load_32 addr -> load addr
def : Pat<(i32 (atomic_load ADDRrr:$src)), (LDrr ADDRrr:$src)>;
def : Pat<(i32 (atomic_load ADDRri:$src)), (LDri ADDRri:$src)>;
// atomic_store_32 val, addr -> store val, addr
def : Pat<(atomic_store ADDRrr:$dst, i32:$val), (STrr ADDRrr:$dst, $val)>;
def : Pat<(atomic_store ADDRri:$dst, i32:$val), (STri ADDRri:$dst, $val)>;
include "SparcInstr64Bit.td"
include "SparcInstrAliases.td"
|