-
Notifications
You must be signed in to change notification settings - Fork 392
/
OMRCodeGenerator.cpp
3137 lines (2714 loc) · 122 KB
/
OMRCodeGenerator.cpp
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
/*******************************************************************************
* Copyright (c) 2000, 2021 IBM Corp. and others
*
* This program and the accompanying materials are made available under
* the terms of the Eclipse Public License 2.0 which accompanies this
* distribution and is available at http://eclipse.org/legal/epl-2.0
* or the Apache License, Version 2.0 which accompanies this distribution
* and is available at https://www.apache.org/licenses/LICENSE-2.0.
*
* This Source Code may also be made available under the following Secondary
* Licenses when the conditions for such availability set forth in the
* Eclipse Public License, v. 2.0 are satisfied: GNU General Public License,
* version 2 with the GNU Classpath Exception [1] and GNU General Public
* License, version 2 with the OpenJDK Assembly Exception [2].
*
* [1] https://www.gnu.org/software/classpath/license.html
* [2] http://openjdk.java.net/legal/assembly-exception.html
*
* SPDX-License-Identifier: EPL-2.0 OR Apache-2.0 OR GPL-2.0 WITH Classpath-exception-2.0 OR LicenseRef-GPL-2.0 WITH Assembly-exception
*******************************************************************************/
#include "codegen/CodeGenerator.hpp"
#include "codegen/CodeGenerator_inlines.hpp"
#include <limits.h>
#include <stdint.h>
#include <string.h>
#include "codegen/BackingStore.hpp"
#include "codegen/ConstantDataSnippet.hpp"
#include "env/FrontEnd.hpp"
#include "codegen/GCStackAtlas.hpp"
#include "codegen/GCStackMap.hpp"
#include "codegen/Instruction.hpp"
#include "codegen/Linkage.hpp"
#include "codegen/Linkage_inlines.hpp"
#include "codegen/LinkageConventionsEnum.hpp"
#include "codegen/LiveRegister.hpp"
#include "codegen/Machine.hpp"
#include "codegen/MemoryReference.hpp"
#include "codegen/RealRegister.hpp"
#include "codegen/RecognizedMethods.hpp"
#include "codegen/Register.hpp"
#include "codegen/RegisterConstants.hpp"
#include "codegen/RegisterIterator.hpp"
#include "codegen/RegisterPressureSimulatorInner.hpp"
#include "codegen/RegisterRematerializationInfo.hpp"
#include "codegen/Snippet.hpp"
#include "codegen/TreeEvaluator.hpp"
#include "codegen/X86Evaluator.hpp"
#ifdef TR_TARGET_64BIT
#include "x/amd64/codegen/AMD64SystemLinkage.hpp"
#else
#include "x/i386/codegen/IA32SystemLinkage.hpp"
#endif
#include "compile/Compilation.hpp"
#include "compile/ResolvedMethod.hpp"
#include "compile/SymbolReferenceTable.hpp"
#include "compile/VirtualGuard.hpp"
#include "control/Options.hpp"
#include "control/Options_inlines.hpp"
#include "control/Recompilation.hpp"
#ifdef J9_PROJECT_SPECIFIC
#include "control/RecompilationInfo.hpp"
#endif
#include "env/CompilerEnv.hpp"
#include "env/IO.hpp"
#include "env/TRMemory.hpp"
#include "env/jittypes.h"
#include "il/Block.hpp"
#include "il/DataTypes.hpp"
#include "il/ILOpCodes.hpp"
#include "il/ILOps.hpp"
#include "il/LabelSymbol.hpp"
#include "il/MethodSymbol.hpp"
#include "il/Node.hpp"
#include "il/Node_inlines.hpp"
#include "il/ParameterSymbol.hpp"
#include "il/ResolvedMethodSymbol.hpp"
#include "il/StaticSymbol.hpp"
#include "il/Symbol.hpp"
#include "il/SymbolReference.hpp"
#include "il/TreeTop.hpp"
#include "il/TreeTop_inlines.hpp"
#include "infra/Assert.hpp"
#include "infra/Bit.hpp"
#include "infra/BitVector.hpp"
#include "infra/Flags.hpp"
#include "infra/IGNode.hpp"
#include "infra/InterferenceGraph.hpp"
#include "infra/List.hpp"
#include "infra/Stack.hpp"
#include "optimizer/RegisterCandidate.hpp"
#include "ras/Debug.hpp"
#include "ras/DebugCounter.hpp"
#include "runtime/CodeCacheManager.hpp"
#include "x/codegen/DataSnippet.hpp"
#include "x/codegen/OutlinedInstructions.hpp"
#include "x/codegen/FPTreeEvaluator.hpp"
#include "x/codegen/X86Instruction.hpp"
#include "codegen/InstOpCode.hpp"
namespace OMR { class RegisterUsage; }
namespace TR { class RegisterDependencyConditions; }
// Hack markers
#define CANT_REMATERIALIZE_ADDRESSES(cg) (cg->comp()->target().is64Bit()) // AMD64 produces a memref with an unassigned addressRegister
TR_X86ProcessorInfo OMR::X86::CodeGenerator::_targetProcessorInfo;
void TR_X86ProcessorInfo::initialize()
{
if (_featureFlags.testAny(TR_X86ProcessorInfoInitialized))
return;
// For now, we only convert the feature bits into a flags32_t, for easier querying.
// To retrieve other information, the VM functions can be called directly.
//
_featureFlags.set(TR::Compiler->target.cpu.getX86ProcessorFeatureFlags());
_featureFlags2.set(TR::Compiler->target.cpu.getX86ProcessorFeatureFlags2());
_featureFlags8.set(TR::Compiler->target.cpu.getX86ProcessorFeatureFlags8());
// Determine the processor vendor.
//
const char *vendor = TR::Compiler->target.cpu.getX86ProcessorVendorId();
if (!strncmp(vendor, "GenuineIntel", 12))
_vendorFlags.set(TR_GenuineIntel);
else if (!strncmp(vendor, "AuthenticAMD", 12))
_vendorFlags.set(TR_AuthenticAMD);
else
_vendorFlags.set(TR_UnknownVendor);
// Finally set this bit so we don't attempt to re-initialize.
//
_featureFlags.set(TR_X86ProcessorInfoInitialized);
// initialize the processor model description
_processorDescription = 0;
// set up the processor family and cache description
uint32_t _processorSignature = TR::Compiler->target.cpu.getX86ProcessorSignature();
if (isGenuineIntel())
{
switch (getCPUFamily(_processorSignature))
{
case 0x05: _processorDescription |= TR_ProcessorIntelPentium; break;
case 0x06:
{
uint32_t extended_model = getCPUModel(_processorSignature) + (getCPUExtendedModel(_processorSignature) << 4);
switch (extended_model)
{
case 0x55:
_processorDescription |= TR_ProcessorIntelSkylake; break;
case 0x4f:
_processorDescription |= TR_ProcessorIntelBroadwell; break;
case 0x3f:
case 0x3c:
_processorDescription |= TR_ProcessorIntelHaswell; break;
case 0x3e:
case 0x3a:
_processorDescription |= TR_ProcessorIntelIvyBridge; break;
case 0x2a:
case 0x2d: // SandyBridge EP
_processorDescription |= TR_ProcessorIntelSandyBridge; break;
case 0x2c: // WestmereEP
case 0x2f: // WestmereEX
_processorDescription |= TR_ProcessorIntelWestmere; break;
case 0x1a: // Nehalem
_processorDescription |= TR_ProcessorIntelNehalem; break;
case 0x17: // Harpertown
case 0x0f: // Woodcrest/Clovertown
_processorDescription |= TR_ProcessorIntelCore2; break;
default: _processorDescription |= TR_ProcessorIntelP6; break;
}
break;
}
case 0x0f: _processorDescription |= TR_ProcessorIntelPentium4; break;
default: _processorDescription |= TR_ProcessorUnknown; break;
}
}
else if (isAuthenticAMD())
{
switch (getCPUFamily(_processorSignature))
{
case 0x05:
if (getCPUModel(_processorSignature) < 0x04)
_processorDescription |= TR_ProcessorAMDK5;
else
_processorDescription |= TR_ProcessorAMDK6;
break;
case 0x06: _processorDescription |= TR_ProcessorAMDAthlonDuron; break;
case 0x0f:
if (getCPUExtendedFamily(_processorSignature) < 6)
_processorDescription |= TR_ProcessorAMDOpteron;
else
_processorDescription |= TR_ProcessorAMDFamily15h;
break;
default: _processorDescription |= TR_ProcessorUnknown; break;
}
}
}
void
OMR::X86::CodeGenerator::initializeX86(TR::Compilation *comp)
{
bool supportsSSE2 = false;
TR_ASSERT_FATAL(comp->compileRelocatableCode() || comp->isOutOfProcessCompilation() || comp->compilePortableCode() || comp->target().cpu.isGenuineIntel() == _targetProcessorInfo.isGenuineIntel(), "isGenuineIntel() failed\n");
TR_ASSERT_FATAL(comp->compileRelocatableCode() || comp->isOutOfProcessCompilation() || comp->compilePortableCode() || comp->target().cpu.isAuthenticAMD() == _targetProcessorInfo.isAuthenticAMD(), "isAuthenticAMD() failed\n");
TR_ASSERT_FATAL(comp->compileRelocatableCode() || comp->isOutOfProcessCompilation() || comp->compilePortableCode() || comp->target().cpu.prefersMultiByteNOP() == _targetProcessorInfo.prefersMultiByteNOP(), "prefersMultiByteNOP() failed\n");
// Pick a padding table
//
if (comp->target().cpu.isGenuineIntel() && comp->target().is32Bit())
_paddingTable = &_old32BitPaddingTable;
else if (comp->target().cpu.isAuthenticAMD())
_paddingTable = &_K8PaddingTable;
else if (comp->target().cpu.prefersMultiByteNOP() && !comp->getOption(TR_DisableZealousCodegenOpts))
_paddingTable = &_intelMultiBytePaddingTable;
else if (comp->target().is32Bit())
_paddingTable = &_old32BitPaddingTable; // Unknown 32-bit target
else
_paddingTable = &_K8PaddingTable; // Unknown 64-bit target
// Determine whether or not x87 or SSE should be used for floating point.
//
#if defined(TR_TARGET_X86)
#if !defined(J9HAMMER)
TR_ASSERT_FATAL(comp->compileRelocatableCode() || comp->isOutOfProcessCompilation() || comp->compilePortableCode() || comp->target().cpu.supportsFeature(OMR_FEATURE_X86_SSE2) == _targetProcessorInfo.supportsSSE2(), "supportsSSE2() failed\n");
if (comp->target().cpu.supportsFeature(OMR_FEATURE_X86_SSE2) && comp->target().cpu.testOSForSSESupport())
supportsSSE2 = true;
#else
// 64-bit targets all support SSE2
supportsSSE2 = true;
#endif // !defined(J9HAMMER)
#endif // defined(TR_TARGET_X86)
TR_ASSERT_FATAL(comp->compileRelocatableCode() || comp->isOutOfProcessCompilation() || comp->compilePortableCode() || comp->target().cpu.supportsFeature(OMR_FEATURE_X86_RTM) == _targetProcessorInfo.supportsTM(), "supportsTM() failed\n");
if (comp->target().cpu.supportsFeature(OMR_FEATURE_X86_RTM) && !comp->getOption(TR_DisableTM))
{
/**
* Due to many verions of Haswell and a small number of Broadwell have defects for TM and then disabled by Intel,
* we will return false for any versions before Broadwell.
*
* TODO: Need to figure out from which mode of Broadwell start supporting TM
*/
TR_ASSERT_FATAL(comp->compileRelocatableCode() || comp->isOutOfProcessCompilation() || comp->compilePortableCode() || comp->target().cpu.is(OMR_PROCESSOR_X86_INTELHASWELL) == _targetProcessorInfo.isIntelHaswell(), "isIntelHaswell() failed\n");
if (!comp->target().cpu.is(OMR_PROCESSOR_X86_INTELHASWELL))
{
if (comp->target().is64Bit())
{
self()->setSupportsTM(); // disable tm on 32bits for now
}
}
}
TR_ASSERT_FATAL(supportsSSE2, "Target processor/OS must support SSE2");
self()->setUseSSEForSinglePrecision();
self()->setUseSSEForDoublePrecision();
self()->setSupportsAutoSIMD();
self()->setSupportsJavaFloatSemantics();
// Choose the best XMM double precision load instruction for the target architecture.
//
TR_ASSERT_FATAL(comp->compileRelocatableCode() || comp->isOutOfProcessCompilation() || comp->compilePortableCode() || comp->target().cpu.isAuthenticAMD() == _targetProcessorInfo.isAuthenticAMD(), "isAuthenticAMD() failed\n");
static char *forceMOVLPD = feGetEnv("TR_forceMOVLPDforDoubleLoads");
if (comp->target().cpu.isAuthenticAMD() || forceMOVLPD)
{
self()->setXMMDoubleLoadOpCode(TR::InstOpCode::MOVLPDRegMem);
}
else
{
self()->setXMMDoubleLoadOpCode(TR::InstOpCode::MOVSDRegMem);
}
self()->setTargetSupportsSoftwarePrefetches();
// Enable software prefetch of the TLH and configure the TLH prefetching
// geometry.
//
TR_ASSERT_FATAL(comp->compileRelocatableCode() || comp->isOutOfProcessCompilation() || comp->compilePortableCode() || comp->target().cpu.is(OMR_PROCESSOR_X86_INTELCORE2) == comp->cg()->getX86ProcessorInfo().isIntelCore2(), "isIntelCore2() failed\n");
TR_ASSERT_FATAL(comp->compileRelocatableCode() || comp->isOutOfProcessCompilation() || comp->compilePortableCode() || comp->target().cpu.is(OMR_PROCESSOR_X86_INTELNEHALEM) == comp->cg()->getX86ProcessorInfo().isIntelNehalem(), "isIntelNehalem() failed\n");
if (((!comp->getOption(TR_DisableTLHPrefetch) && (comp->target().cpu.is(OMR_PROCESSOR_X86_INTELCORE2) || comp->target().cpu.is(OMR_PROCESSOR_X86_INTELNEHALEM))) ||
(comp->getOption(TR_TLHPrefetch) && self()->targetSupportsSoftwarePrefetches())))
{
self()->setEnableTLHPrefetching();
}
self()->setGlobalGPRPartitionLimit(self()->machine()->getGlobalGPRPartitionLimit());
self()->setGlobalFPRPartitionLimit(self()->machine()->getGlobalFPRPartitionLimit());
self()->setLastGlobalGPR(self()->machine()->getLastGlobalGPRRegisterNumber());
self()->setLast8BitGlobalGPR(self()->machine()->getLast8BitGlobalGPRRegisterNumber());
self()->setLastGlobalFPR(self()->machine()->getLastGlobalFPRRegisterNumber());
/*
* GRA does not work with vector registers on 32 bit due to a bug where xmm registers are not being assigned.
* This disables GRA for vector registers on 32 bit.
* This code will be reenabled as part of Issue 2035 which tracks the progress of fixing the GRA bug.
* GRA does not work with vector registers on 64 bit either. So GRA is now being disabled vector registers.
* This code will be reenabled as part of Issue 2280
*/
#if 0
if (comp->target().is64Bit())
{
self()->setFirstGlobalVRF(self()->getFirstGlobalFPR());
self()->setLastGlobalVRF(self()->getLastGlobalFPR());
}
#endif // closes the if 0.
// Initialize Linkage for Code Generator
self()->initializeLinkage();
_linkageProperties = &self()->getLinkage()->getProperties();
_unlatchedRegisterList = (TR::RealRegister**)self()->trMemory()->allocateHeapMemory(sizeof(TR::RealRegister*)*(TR::RealRegister::NumRegisters + 1));
_unlatchedRegisterList[0] = 0; // mark that list is empty
self()->setGlobalRegisterTable(self()->machine()->getGlobalRegisterTable(*_linkageProperties));
self()->machine()->initializeRegisterFile(*_linkageProperties);
self()->getLinkage()->copyLinkageInfoToParameterSymbols();
_switchToInterpreterLabel = NULL;
// Use a virtual frame pointer register. The real frame pointer register to be used
// will be substituted based on _vfpState during size estimation.
//
_frameRegister = self()->machine()->getRealRegister(TR::RealRegister::vfp);
TR::Register *vmThreadRegister = self()->setVMThreadRegister(self()->allocateRegister());
TR::RealRegister::RegNum vmThreadIndex = _linkageProperties->getMethodMetaDataRegister();
if (vmThreadIndex != TR::RealRegister::NoReg)
{
TR::RealRegister *vmThreadReal = self()->machine()->getRealRegister(vmThreadIndex);
vmThreadRegister->setAssignedRegister(vmThreadReal);
vmThreadRegister->setAssociation(vmThreadIndex);
vmThreadReal->setAssignedRegister(vmThreadRegister);
vmThreadReal->setState(TR::RealRegister::Assigned);
}
if (!debug("disableBetterSpillPlacements"))
self()->setEnableBetterSpillPlacements();
self()->setEnableRematerialisation();
self()->setEnableRegisterAssociations();
self()->setEnableRegisterWeights();
self()->setEnableRegisterInterferences();
self()->setEnableSinglePrecisionMethods();
if (!debug("disableRefinedAliasSets"))
self()->setEnableRefinedAliasSets();
if (!comp->getOption(TR_DisableLiveRangeSplitter))
comp->setOption(TR_EnableRangeSplittingGRA);
self()->addSupportedLiveRegisterKind(TR_GPR);
self()->setLiveRegisters(new (self()->trHeapMemory()) TR_LiveRegisters(comp), TR_GPR);
self()->addSupportedLiveRegisterKind(TR_FPR);
self()->setLiveRegisters(new (self()->trHeapMemory()) TR_LiveRegisters(comp), TR_FPR);
self()->addSupportedLiveRegisterKind(TR_VRF);
self()->setLiveRegisters(new (self()->trHeapMemory()) TR_LiveRegisters(comp), TR_VRF);
if (!TR::Compiler->om.canGenerateArraylets())
{
self()->setSupportsArrayCmp();
self()->setSupportsPrimitiveArrayCopy();
if (!comp->getOption(TR_DisableArraySetOpts))
{
self()->setSupportsArraySet();
}
static bool disableX86TRTO = feGetEnv("TR_disableX86TRTO") != NULL;
if (!disableX86TRTO)
{
TR_ASSERT_FATAL(comp->compileRelocatableCode() || comp->isOutOfProcessCompilation() || comp->compilePortableCode() || comp->target().cpu.supportsFeature(OMR_FEATURE_X86_SSE4_1) == self()->getX86ProcessorInfo().supportsSSE4_1(), "supportsSSE4_1() failed\n");
if (comp->target().cpu.supportsFeature(OMR_FEATURE_X86_SSE4_1))
{
self()->setSupportsArrayTranslateTRTO();
}
}
static bool disableX86TROT = feGetEnv("TR_disableX86TROT") != NULL;
if (!disableX86TROT)
{
TR_ASSERT_FATAL(comp->compileRelocatableCode() || comp->isOutOfProcessCompilation() || comp->compilePortableCode() || comp->target().cpu.supportsFeature(OMR_FEATURE_X86_SSE4_1) == self()->getX86ProcessorInfo().supportsSSE4_1(), "supportsSSE4_1() failed\n");
TR_ASSERT_FATAL(comp->compileRelocatableCode() || comp->isOutOfProcessCompilation() || comp->compilePortableCode() || comp->target().cpu.supportsFeature(OMR_FEATURE_X86_SSE2) == self()->getX86ProcessorInfo().supportsSSE2(), "supportsSSE4_1() failed\n");
if (comp->target().cpu.supportsFeature(OMR_FEATURE_X86_SSE4_1))
{
self()->setSupportsArrayTranslateTROT();
}
if (comp->target().cpu.supportsFeature(OMR_FEATURE_X86_SSE2))
{
self()->setSupportsArrayTranslateTROTNoBreak();
}
}
}
self()->setSupportsRecompilation();
self()->setSupportsScaledIndexAddressing();
self()->setSupportsConstantOffsetInAddressing();
self()->setSupportsCompactedLocals();
self()->setSupportsGlRegDeps();
self()->setSupportsEfficientNarrowIntComputation();
self()->setSupportsEfficientNarrowUnsignedIntComputation();
self()->setSupportsVirtualGuardNOPing();
self()->setSupportsDynamicANewArray();
self()->setSupportsSelect();
// TODO (#5642): Re-enable byteswap support on x86 and Power
// self()->setSupportsByteswap();
// allows [i/l]div to decompose to [i/l]mulh in TreeSimplifier
//
static char * enableMulHigh = feGetEnv("TR_X86MulHigh");
if (enableMulHigh)
{
self()->setSupportsLoweringConstIDiv();
if (comp->target().is64Bit())
self()->setSupportsLoweringConstLDiv();
}
self()->setSpillsFPRegistersAcrossCalls(); // TODO:AMD64: Are the preserved XMMRs relevant here?
// Make a conservative estimate of the boundary over which an executable instruction cannot
// be patched.
//
TR_ASSERT_FATAL(comp->compileRelocatableCode() || comp->isOutOfProcessCompilation() || comp->compilePortableCode() || comp->target().cpu.isGenuineIntel() == _targetProcessorInfo.isGenuineIntel(), "isGenuineIntel() failed\n");
TR_ASSERT_FATAL(comp->compileRelocatableCode() || comp->isOutOfProcessCompilation() || comp->compilePortableCode() || comp->target().cpu.isAuthenticAMD() == _targetProcessorInfo.isAuthenticAMD(), "isAuthenticAMD() failed\n");
TR_ASSERT_FATAL(comp->compileRelocatableCode() || comp->isOutOfProcessCompilation() || comp->compilePortableCode() || comp->target().cpu.is(OMR_PROCESSOR_X86_AMDFAMILY15H) == _targetProcessorInfo.isAMD15h(), "isAMD15h() failed\n");
int32_t boundary;
if (comp->target().cpu.isGenuineIntel() || (comp->target().cpu.isAuthenticAMD() && comp->target().cpu.is(OMR_PROCESSOR_X86_AMDFAMILY15H)))
boundary = 32;
else
{
// TR_AuthenticAMD
// TR_UnknownVendor
//
boundary = 8;
}
self()->setInstructionPatchAlignmentBoundary(boundary);
// Smallest code patching boundary that is known to work on all processors we support.
//
self()->setLowestCommonCodePatchingAlignmentBoundary(8);
self()->setPicSlotCount(0);
if (!comp->getOption(TR_DisableRegisterPressureSimulation))
{
for (int32_t i = 0; i < TR_numSpillKinds; i++)
_globalRegisterBitVectors[i].init(self()->getNumberOfGlobalRegisters(), comp->trMemory());
TR::RealRegister::RegNum vmThreadRealRegisterIndex = _linkageProperties->getMethodMetaDataRegister();
for (TR_GlobalRegisterNumber grn=0; grn < self()->getNumberOfGlobalRegisters(); grn++)
{
TR::RealRegister::RegNum reg = (TR::RealRegister::RegNum)self()->getGlobalRegister(grn);
if (grn < self()->getFirstGlobalFPR())
_globalRegisterBitVectors[ TR_gprSpill ].set(grn);
else
_globalRegisterBitVectors[ TR_fprSpill ].set(grn);
if (!self()->getProperties().isPreservedRegister(reg))
_globalRegisterBitVectors[ TR_volatileSpill ].set(grn);
if (self()->getProperties().isIntegerArgumentRegister(reg) || self()->getProperties().isFloatArgumentRegister(reg))
_globalRegisterBitVectors[ TR_linkageSpill ].set(grn);
if (reg == vmThreadRealRegisterIndex)
_globalRegisterBitVectors[ TR_vmThreadSpill ].set(grn);
if (reg == TR::RealRegister::eax)
_globalRegisterBitVectors[ TR_eaxSpill ].set(grn);
if (reg == TR::RealRegister::ecx)
_globalRegisterBitVectors[ TR_ecxSpill ].set(grn);
if (reg == TR::RealRegister::edx)
_globalRegisterBitVectors[ TR_edxSpill ].set(grn);
}
}
if (comp->target().cpu.isI386())
self()->setGenerateMasmListingSyntax();
if (comp->getOption(TR_TraceRA))
{
self()->setGPRegisterIterator(new (self()->trHeapMemory()) TR::RegisterIterator(self()->machine(), TR::RealRegister::FirstGPR, TR::RealRegister::LastAssignableGPR));
self()->setFPRegisterIterator(new (self()->trHeapMemory()) TR::RegisterIterator(self()->machine(), TR::RealRegister::FirstXMMR, TR::RealRegister::LastXMMR));
}
self()->setSupportsProfiledInlining();
}
OMR::X86::CodeGenerator::CodeGenerator(TR::Compilation *comp) :
OMR::CodeGenerator(comp),
_nanoTimeTemp(NULL),
_assignmentDirection(Backward),
_lastCatchAppendInstruction(NULL),
_betterSpillPlacements(NULL),
_dataSnippetList(getTypedAllocator<TR::X86DataSnippet*>(comp->allocator())),
_spilledIntRegisters(getTypedAllocator<TR::Register*>(comp->allocator())),
_liveDiscardableRegisters(getTypedAllocator<TR::Register*>(comp->allocator())),
_dependentDiscardableRegisters(getTypedAllocator<TR::Register*>(comp->allocator())),
_clobberingInstructions(getTypedAllocator<TR::ClobberingInstruction*>(comp->allocator())),
_outlinedInstructionsList(getTypedAllocator<TR_OutlinedInstructions*>(comp->allocator())),
_numReservedIPICTrampolines(0),
_flags(0)
{
}
void
OMR::X86::CodeGenerator::initialize()
{
self()->OMR::CodeGenerator::initialize();
_clobIterator = _clobberingInstructions.begin();
}
TR::Linkage *
OMR::X86::CodeGenerator::createLinkage(TR_LinkageConventions lc)
{
TR::Compilation *comp = self()->comp();
TR::Linkage *linkage = NULL;
switch (lc)
{
case TR_Private:
// HACK HACK HACK "intentional" fall through to system linkage
case TR_Helper:
// Intentional fall through
case TR_System:
if (self()->comp()->target().isLinux() || self()->comp()->target().isOSX() || self()->comp()->target().isBSD())
{
#if defined(TR_TARGET_64BIT)
linkage = new (self()->trHeapMemory()) TR::AMD64ABILinkage(self());
#else
linkage = new (self()->trHeapMemory()) TR::IA32SystemLinkage(self());
#endif
}
else if (self()->comp()->target().isWindows())
{
#if defined(TR_TARGET_64BIT)
linkage = new (self()->trHeapMemory()) TR::AMD64Win64FastCallLinkage(self());
#else
linkage = new (self()->trHeapMemory()) TR::IA32SystemLinkage(self());
#endif
}
else
{
TR_ASSERT(0, "\nTestarossa error: Illegal linkage convention %d\n", lc);
}
break;
default :
TR_ASSERT(0, "\nTestarossa error: Illegal linkage convention %d\n", lc);
}
self()->setLinkage(lc, linkage);
return linkage;
}
void
OMR::X86::CodeGenerator::beginInstructionSelection()
{
TR::Compilation *comp = self()->comp();
_returnTypeInfoInstruction = NULL;
TR::ResolvedMethodSymbol * methodSymbol = comp->getJittedMethodSymbol();
TR::Recompilation * recompilation = comp->getRecompilationInfo();
TR::Node * startNode = comp->getStartTree()->getNode();
if (recompilation && recompilation->generatePrePrologue() != NULL)
{
// Return type info will have been generated by recompilation info
//
if (methodSymbol->getLinkageConvention() == TR_Private)
_returnTypeInfoInstruction = (TR::X86ImmInstruction*)(self()->getAppendInstruction());
if (methodSymbol->getLinkageConvention() == TR_System)
_returnTypeInfoInstruction = (TR::X86ImmInstruction*)(self()->getAppendInstruction());
}
if (methodSymbol->getLinkageConvention() == TR_Private && !_returnTypeInfoInstruction)
{
// linkageInfo word
if (self()->getAppendInstruction())
_returnTypeInfoInstruction = generateImmInstruction(TR::InstOpCode::DDImm4, startNode, 0, self());
else
_returnTypeInfoInstruction = new (self()->trHeapMemory()) TR::X86ImmInstruction((TR::Instruction *)NULL, TR::InstOpCode::DDImm4, 0, self());
}
if (methodSymbol->getLinkageConvention() == TR_System && !_returnTypeInfoInstruction)
{
// linkageInfo word
if (self()->getAppendInstruction())
_returnTypeInfoInstruction = generateImmInstruction(TR::InstOpCode::DDImm4, startNode, 0, self());
else
_returnTypeInfoInstruction = new (self()->trHeapMemory()) TR::X86ImmInstruction((TR::Instruction *)NULL, TR::InstOpCode::DDImm4, 0, self());
}
if (self()->getAppendInstruction())
generateInstruction(TR::InstOpCode::proc, startNode, self());
else
new (self()->trHeapMemory()) TR::Instruction(TR::InstOpCode::proc, (TR::Instruction *)NULL, self());
// Set the default FPCW to single precision mode if we are allowed to.
//
if (self()->enableSinglePrecisionMethods() && comp->getJittedMethodSymbol()->usesSinglePrecisionMode())
{
generateMemInstruction(TR::InstOpCode::LDCWMem, startNode, generateX86MemoryReference(self()->findOrCreate2ByteConstant(startNode, SINGLE_PRECISION_ROUND_TO_NEAREST), self()), self());
}
}
void
OMR::X86::CodeGenerator::endInstructionSelection()
{
TR::Compilation *comp = self()->comp();
if (_returnTypeInfoInstruction != NULL)
{
TR_ReturnInfo returnInfo = comp->getReturnInfo();
// Note: this will get clobbered again in code generation on AMD64
_returnTypeInfoInstruction->setSourceImmediate(returnInfo);
}
// Reset the FPCW in the dummy finally block.
//
if (self()->enableSinglePrecisionMethods() &&
comp->getJittedMethodSymbol()->usesSinglePrecisionMode())
{
TR_ASSERT(self()->getLastCatchAppendInstruction(),
"endInstructionSelection() ==> Could not find the dummy finally block!\n");
generateMemInstruction(self()->getLastCatchAppendInstruction(), TR::InstOpCode::LDCWMem, generateX86MemoryReference(self()->findOrCreate2ByteConstant(self()->getLastCatchAppendInstruction()->getNode(), DOUBLE_PRECISION_ROUND_TO_NEAREST), self()), self());
}
}
int32_t OMR::X86::CodeGenerator::getMaximumNumbersOfAssignableGPRs()
{
return TR::RealRegister::LastAssignableGPR - TR::RealRegister::FirstGPR + 1; // TODO:AMD64: This is including rsp
}
/*
* This method returns TRUE for all the cases we decide NOT to replace the call to CAS
* with inline assembly. The GRA and Evaluator should be consistent about whether to inline CAS natives.
*/
static bool willNotInlineCompareAndSwapNative(TR::Node *node,
int8_t size,
TR::Compilation *comp)
{
#ifdef J9_PROJECT_SPECIFIC
TR::SymbolReference *callSymRef = node->getSymbolReference();
TR::MethodSymbol *methodSymbol = callSymRef->getSymbol()->castToMethodSymbol();
if (TR::Compiler->om.canGenerateArraylets() && !node->isUnsafeGetPutCASCallOnNonArray())
return true;
static char *disableCASInlining = feGetEnv("TR_DisableCASInlining");
if (disableCASInlining /* || comp->useCompressedPointers() */)
return true;
// In Java9 the sun.misc.Unsafe JNI methods have been moved to jdk.internal,
// with a set of wrappers remaining in sun.misc to delegate to the new package.
// We can be called in this function for the wrappers (which we will
// not be converting to assembly), the new jdk.internal JNI methods or the
// Java8 sun.misc JNI methods (both of which we will convert). We can
// differentiate between these cases by testing with isNative() on the method.
if (!methodSymbol->isNative())
return true;
if (size == 4)
{
return false;
}
else if (size == 8 && comp->target().is64Bit())
{
return false;
}
else
{
if (!comp->cg()->getX86ProcessorInfo().supportsCMPXCHG8BInstruction())
return true;
return false;
}
#else
return true;
#endif
}
/** @brief Identify methods which are not transformed into inline assembly.
Some recognized methods are transformed into very simple hardcoded
assembly sequences which don't need register spills as real calls do.
@param node The TR::Node for the method call. NB this function assumes the Node is a call.
@return true if the method will be treated as a normal call, false if the method
will be converted to inline assembly.
*/
bool OMR::X86::CodeGenerator::willBeEvaluatedAsCallByCodeGen(TR::Node *node, TR::Compilation *comp)
{
#ifdef J9_PROJECT_SPECIFIC
TR::SymbolReference *callSymRef = node->getSymbolReference();
TR::MethodSymbol *methodSymbol = callSymRef->getSymbol()->castToMethodSymbol();
switch (methodSymbol->getRecognizedMethod())
{
case TR::sun_misc_Unsafe_compareAndSwapLong_jlObjectJJJ_Z:
return willNotInlineCompareAndSwapNative(node, 8, comp);
case TR::sun_misc_Unsafe_compareAndSwapInt_jlObjectJII_Z:
return willNotInlineCompareAndSwapNative(node, 4, comp);
case TR::sun_misc_Unsafe_compareAndSwapObject_jlObjectJjlObjectjlObject_Z:
return willNotInlineCompareAndSwapNative(node, (comp->target().is64Bit() && !comp->useCompressedPointers()) ? 8 : 4, comp);
default:
break;
}
#endif
return true;
}
int32_t OMR::X86::CodeGenerator::getMaximumNumbersOfAssignableFPRs()
{
return TR::RealRegister::LastAssignableFPR - TR::RealRegister::FirstFPR + 1;
}
// X has a different concept of VR's compared to p/z but since LocalOpts needs this to be hoisted within CG,
// this method is placed here as a dummy until platform specific code is removed from LocalOpt
int32_t OMR::X86::CodeGenerator::getMaximumNumbersOfAssignableVRs()
{
return INT_MAX;
}
void OMR::X86::CodeGenerator::addLiveDiscardableRegister(TR::Register * reg)
{
_liveDiscardableRegisters.push_front(reg);
reg->setIsDiscardable();
}
void OMR::X86::CodeGenerator::removeLiveDiscardableRegister(TR::Register * reg)
{
_liveDiscardableRegisters.remove(reg);
reg->resetIsDiscardable();
}
bool OMR::X86::CodeGenerator::canNullChkBeImplicit(TR::Node * node)
{
return self()->canNullChkBeImplicit(node, true);
}
void OMR::X86::CodeGenerator::clobberLiveDiscardableRegisters(
TR::Instruction *instr,
TR::MemoryReference *mr)
{
TR::Symbol * symbol = mr->getSymbolReference().getSymbol();
if (symbol)
{
TR::ClobberingInstruction * clob = NULL;
TR_IGNode *IGNodeSym = NULL;
if (self()->getLocalsIG())
IGNodeSym = self()->getLocalsIG()->getIGNodeForEntity(symbol);
// If this instruction clobbers the source of any memory discardable register(s),
// those registers must be deactivated (marked as non-discardable) after this point.
// We record which registers are deactivated, so that we can re-activate them when
// we have assigned registers for this instruction.
auto iterator = self()->getLiveDiscardableRegisters().begin();
while (iterator != self()->getLiveDiscardableRegisters().end())
{
TR::Register *registerCursor = *iterator;
if (registerCursor->getRematerializationInfo()->isRematerializableFromMemory())
{
TR::SymbolReference * rmSymRef = registerCursor->getRematerializationInfo()->getSymbolReference();
if ((rmSymRef->getSymbol() == symbol) ||
(rmSymRef->getOffset() == mr->getSymbolReference().getOffset()))
{
if (!clob)
{
clob = new (self()->trHeapMemory()) TR::ClobberingInstruction(instr, self()->trMemory());
self()->addClobberingInstruction(clob);
}
clob->addClobberedRegister(registerCursor);
iterator = self()->getLiveDiscardableRegisters().erase(iterator);
registerCursor->resetIsDiscardable();
if (debug("dumpRemat"))
{
diagnostic("---> Clobbering %s discardable register %s at instruction %p in %s\n",
self()->getDebug()->toString(registerCursor->getRematerializationInfo()),
self()->getDebug()->getName(registerCursor),
instr, self()->comp()->signature());
}
}
// If the symbols are different but will share the same stack slot then a clobber has occurred.
//
else if (IGNodeSym != NULL)
{
TR_IGNode *IGNodeRematSym = self()->getLocalsIG()->getIGNodeForEntity(rmSymRef->getSymbol());
if ((IGNodeRematSym != NULL) &&
(IGNodeSym->getColour() == IGNodeRematSym->getColour()))
{
if (!clob)
{
clob = new (self()->trHeapMemory()) TR::ClobberingInstruction(instr, self()->trMemory());
self()->addClobberingInstruction(clob);
}
clob->addClobberedRegister(registerCursor);
iterator = self()->getLiveDiscardableRegisters().erase(iterator);
registerCursor->resetIsDiscardable();
if (debug("dumpRemat"))
{
diagnostic("---> Clobbering %s discardable register %s at instruction %p because of shared slot in %s\n",
self()->getDebug()->toString(registerCursor->getRematerializationInfo()),
self()->getDebug()->getName(registerCursor),
instr, self()->comp()->signature());
}
}
else
++iterator;
}
else
++iterator;
}
else
++iterator;
}
// If a register-dependent discardable register depends on any of the deactivated
// registers, it (and those that depend on it if any) must also be deactivated.
//
if (clob && self()->supportsIndirectMemoryRematerialization())
{
iterator = clob->getClobberedRegisters().begin();
while (iterator != clob->getClobberedRegisters().end())
{
self()->clobberLiveDependentDiscardableRegisters(clob, *iterator);
++iterator;
}
}
}
}
// During instruction selection, if we know that an instruction clobbers a
// register, we iteratively deactivate all registers that depend on registers
// that have been clobbered/deactivated.
//
void OMR::X86::CodeGenerator::clobberLiveDependentDiscardableRegisters(TR::ClobberingInstruction * clob,
TR::Register * baseReg)
{
TR_Stack<TR::Register *> worklist(self()->trMemory());
worklist.push(baseReg);
while (!worklist.isEmpty())
{
baseReg = worklist.pop();
for (auto iterator = self()->getLiveDiscardableRegisters().begin(); iterator != self()->getLiveDiscardableRegisters().end();)
{
TR::Register * candidate = *iterator;
TR_RematerializationInfo * info = candidate->getRematerializationInfo();
if (info->isIndirect() && info->getBaseRegister() == baseReg)
{
clob->addClobberedRegister(candidate);
iterator = self()->getLiveDiscardableRegisters().erase(iterator);
candidate->resetIsDiscardable();
worklist.push(candidate);
if (debug("dumpRemat"))
{
diagnostic("---> Clobbering %s discardable register %s at instruction %p in %s\n",
self()->getDebug()->toString(info), self()->getDebug()->getName(candidate), clob->getInstruction(),
self()->comp()->signature());
}
}
else
++iterator;
}
}
}
// During register assignment, rematerialisable registers dependent on
// a base register B are re-activated when the spill state of B is
// reversed (i.e. when the register assigner inserts a spill store for it).
// A candidate is only re-activated if it is discardable at the current
// program point.
//
void OMR::X86::CodeGenerator::reactivateDependentDiscardableRegisters(TR::Register * baseReg)
{
TR_Stack<TR::Register *> worklist(self()->trMemory());
worklist.push(baseReg);
if (debug("dumpRemat"))
diagnostic("---> Re-activating rematerializable registers dependent on %s: ",
self()->getDebug()->getName(baseReg));
while (!worklist.isEmpty())
{
baseReg = worklist.pop();
for (auto iterator = self()->getDependentDiscardableRegisters().begin(); iterator != self()->getDependentDiscardableRegisters().end(); ++iterator)
{
if ((*iterator)->isDiscardable() &&
(*iterator)->getRematerializationInfo()->getBaseRegister() == baseReg)
{
if (debug("dumpRemat"))
diagnostic("%s ", self()->getDebug()->getName(*iterator));
(*iterator)->getRematerializationInfo()->setActive();
// If this candidate is currently assigned, then discardable registers
// dependent on it can also be re-activated.
//
if ((*iterator)->getAssignedRegister())
worklist.push(*iterator);
}
}
}
if (debug("dumpRemat"))
diagnostic("\n");
}
// During register assignment, rematerialisable registers dependent on
// a base register B must be deactivated when B is spilled (i.e. when
// the assigner inserts a reload instruction for it).
//
void OMR::X86::CodeGenerator::deactivateDependentDiscardableRegisters(TR::Register * baseReg)
{
TR_Stack<TR::Register *> worklist(self()->trMemory());
worklist.push(baseReg);
if (debug("dumpRemat"))
diagnostic("---> Deactivating rematerialisable registers dependent on %s: ",
self()->getDebug()->getName(baseReg));
while (!worklist.isEmpty())
{
baseReg = worklist.pop();
for (auto iterator = self()->getDependentDiscardableRegisters().begin(); iterator != self()->getDependentDiscardableRegisters().end(); ++iterator)
{
if (baseReg == (*iterator)->getRematerializationInfo()->getBaseRegister())
{
if (debug("dumpRemat"))
diagnostic("%s ", self()->getDebug()->getName(*iterator));
(*iterator)->getRematerializationInfo()->resetActive();
worklist.push(*iterator);
}
}
}
if (debug("dumpRemat"))
diagnostic("\n");
}
#define ALLOWED_TO_REMATERIALIZE(x) \
(getRematerializationOptString() && strstr(getRematerializationOptString(), (x)))
#define CAN_REMATERIALIZE(x) \
(!getRematerializationOptString() || strstr(getRematerializationOptString(), (x)))
static const char *getRematerializationOptString()
{
static char *optString = feGetEnv("TR_REMAT");
return optString;
}
bool OMR::X86::CodeGenerator::supportsConstantRematerialization() { static bool b = CAN_REMATERIALIZE("constant"); return b; }
bool OMR::X86::CodeGenerator::supportsLocalMemoryRematerialization() { static bool b = CAN_REMATERIALIZE("local"); return b; }
bool OMR::X86::CodeGenerator::supportsStaticMemoryRematerialization() { static bool b = CAN_REMATERIALIZE("static"); return !CANT_REMATERIALIZE_ADDRESSES(self()) && b; }
bool OMR::X86::CodeGenerator::supportsXMMRRematerialization() { static bool b = CAN_REMATERIALIZE("xmmr"); return b; }
bool OMR::X86::CodeGenerator::supportsIndirectMemoryRematerialization() { static bool b = ALLOWED_TO_REMATERIALIZE("indirect"); return !CANT_REMATERIALIZE_ADDRESSES(self()) && b;}
bool OMR::X86::CodeGenerator::supportsAddressRematerialization() { static bool b = ALLOWED_TO_REMATERIALIZE("address"); return !CANT_REMATERIALIZE_ADDRESSES(self()) && b; }
#undef ALLOWED_TO_REMATERIALIZE
#undef CAN_REMATERIALIZE
bool
OMR::X86::CodeGenerator::getSupportsOpCodeForAutoSIMD(TR::ILOpCode opcode, TR::DataType dt)
{
/*
* Most of the vector evaluators for opcodes used in AutoSIMD have been implemented.
* The cases that return false are placeholders that should be updated as support for more vector evaluators is added.
*/
// implemented vector opcodes
switch (opcode.getOpCodeValue())
{
case TR::vadd: