-
Notifications
You must be signed in to change notification settings - Fork 392
/
X86BinaryEncoding.cpp
3033 lines (2575 loc) · 120 KB
/
X86BinaryEncoding.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 <algorithm>
#include <stddef.h>
#include <stdint.h>
#include "codegen/BackingStore.hpp"
#include "codegen/CodeGenerator.hpp"
#include "env/FrontEnd.hpp"
#include "codegen/Instruction.hpp"
#include "codegen/Linkage.hpp"
#include "codegen/Linkage_inlines.hpp"
#include "codegen/Machine.hpp"
#include "codegen/MemoryReference.hpp"
#include "codegen/RealRegister.hpp"
#include "codegen/Register.hpp"
#include "codegen/Relocation.hpp"
#include "codegen/Snippet.hpp"
#include "codegen/UnresolvedDataSnippet.hpp"
#include "compile/Compilation.hpp"
#include "compile/ResolvedMethod.hpp"
#include "compile/SymbolReferenceTable.hpp"
#include "control/Options.hpp"
#include "control/Options_inlines.hpp"
#include "env/CompilerEnv.hpp"
#include "env/PersistentInfo.hpp"
#include "env/TRMemory.hpp"
#include "env/jittypes.h"
#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/ResolvedMethodSymbol.hpp"
#include "il/StaticSymbol.hpp"
#include "il/Symbol.hpp"
#include "il/SymbolReference.hpp"
#include "infra/Assert.hpp"
#include "infra/List.hpp"
#include "ras/Debug.hpp"
#include "ras/DebugCounter.hpp"
#include "runtime/CodeCacheManager.hpp"
#include "runtime/Runtime.hpp"
#include "x/codegen/X86Instruction.hpp"
#include "codegen/InstOpCode.hpp"
#include "codegen/StaticRelocation.hpp"
#ifdef J9_PROJECT_SPECIFIC
#include "env/CHTable.hpp"
#include "env/VMJ9.h"
#endif
class TR_OpaqueClassBlock;
class TR_OpaqueMethodBlock;
int32_t memoryBarrierRequired(
TR::InstOpCode &op,
TR::MemoryReference *mr,
TR::CodeGenerator *cg,
bool onlyAskingAboutFences)
{
TR::Compilation *comp = cg->comp();
if (!comp->target().isSMP())
return NoFence;
// LOCKed instructions are serializing instructions and implicitly handle
// any required memory barriers.
//
if (op.needsLockPrefix())
return NoFence;
if (!onlyAskingAboutFences &&
mr->requiresLockPrefix())
{
TR_ASSERT(op.supportsLockPrefix(), "cannot emit a lock prefix with this type of instruction");
return LockPrefix;
}
int32_t barrier = NoFence;
TR::SymbolReference& symRef = mr->getSymbolReference();
TR::Symbol *sym = symRef.getSymbol();
static char *mbou = feGetEnv("TR_MemoryBarriersOnUnresolved");
TR_ASSERT_FATAL(cg->comp()->compileRelocatableCode() || cg->comp()->isOutOfProcessCompilation() || cg->comp()->compilePortableCode() || cg->comp()->target().cpu.requiresLFence() == cg->getX86ProcessorInfo().requiresLFENCE(), "requiresLFence() failed\n");
// Unresolved references are assumed to be volatile until they can be proven otherwise.
// The memory barrier will be removed by PicBuilder when the reference is resolved and
// proven to be non-volatile.
//
if ((symRef.isUnresolved() && mbou) ||
(sym && sym->isVolatile() && !mr->ignoreVolatile()))
{
if (op.sourceIsMemRef())
{
if (op.modifiesSource())
{
// use lock or as default fence unless specified to use mfence
if (!comp->getOption(TR_X86UseMFENCE))
barrier |= LockOR;
else
barrier |= kMemoryFence;
}
else if (cg->comp()->target().cpu.requiresLFence())
barrier |= kLoadFence;
}
else
{
if (op.modifiesTarget())
{
// use lock or as default fence unless specified to use mfence
if (!comp->getOption(TR_X86UseMFENCE))
barrier |= LockOR;
else
barrier |= kMemoryFence;
}
else if (op.usesTarget() && cg->comp()->target().cpu.requiresLFence())
barrier |= kLoadFence;
}
}
static char *disableExplicitFences = feGetEnv("TR_DisableExplicitFences");
if (barrier)
{
TR_ASSERT_FATAL(cg->comp()->compileRelocatableCode() || cg->comp()->isOutOfProcessCompilation() || cg->comp()->compilePortableCode() || cg->comp()->target().cpu.supportsLFence() == cg->getX86ProcessorInfo().supportsLFence(), "supportsLFence() failed\n");
TR_ASSERT_FATAL(cg->comp()->compileRelocatableCode() || cg->comp()->isOutOfProcessCompilation() || cg->comp()->compilePortableCode() || cg->comp()->target().cpu.supportsMFence() == cg->getX86ProcessorInfo().supportsMFence(), "supportsMFence() failed\n");
if ((!cg->comp()->target().cpu.supportsLFence() ||
!cg->comp()->target().cpu.supportsMFence()) || disableExplicitFences)
{
if (op.supportsLockPrefix())
barrier |= LockPrefix;
else
barrier |= LockOR;
}
}
return barrier;
}
int32_t estimateMemoryBarrierBinaryLength(int32_t barrier, TR::CodeGenerator *cg)
{
int32_t length = 0;
if (!cg->comp()->getOption(TR_DisableNewX86VolatileSupport))
return 0;
if (barrier & LockOR)
length = 5;
else if ((barrier & kLoadFence) && cg->comp()->target().cpu.requiresLFence())
length = TR::InstOpCode(TR::InstOpCode::LFENCE).length();
else if ((barrier & kMemoryFence) == kMemoryFence)
length = TR::InstOpCode(TR::InstOpCode::MFENCE).length();
else if (barrier & kStoreFence)
length = TR::InstOpCode(TR::InstOpCode::SFENCE).length();
return length;
}
uint8_t getMemoryBarrierBinaryLengthLowerBound(int32_t barrier, TR::CodeGenerator *cg)
{
return estimateMemoryBarrierBinaryLength(barrier, cg);
}
// -----------------------------------------------------------------------------
// OMR::X86::Instruction:: member functions
bool OMR::X86::Instruction::needsRepPrefix()
{
return self()->getOpCode().needsRepPrefix() != 0;
}
bool OMR::X86::Instruction::needsLockPrefix()
{
return self()->getOpCode().needsLockPrefix() != 0;
}
uint8_t* OMR::X86::Instruction::generateBinaryEncoding()
{
// Normally the loop only executes once,
// it is needed to avoid recursive calls when generateOperand() requests to regenerate the binary code by returning NULL;
for(;;)
{
uint8_t *instructionStart = self()->cg()->getBinaryBufferCursor();
uint8_t *cursor = instructionStart;
self()->setBinaryEncoding(instructionStart);
if (self()->needsRepPrefix())
{
*cursor++ = 0xf3;
}
if (self()->needsLockPrefix())
{
*cursor++ = 0xf0;
}
cursor = self()->generateRepeatedRexPrefix(cursor);
cursor = self()->getOpCode().binary(cursor, self()->rexBits());
cursor = self()->generateOperand(cursor);
// cursor is normal not NULL, and hence we can finish binary encoding and exit the loop
// cursor is NULL when generateOperand() requests to regenerate the binary code, which may happen during encoding of memref with unresolved symbols on 64-bit
if (cursor)
{
if (!self()->getSource2ndRegister())
{
self()->getOpCode().finalize(instructionStart);
}
self()->setBinaryLength(static_cast<int8_t>(cursor - instructionStart));
self()->cg()->addAccumulatedInstructionLengthError(self()->getEstimatedBinaryLength() - self()->getBinaryLength());
return cursor;
}
}
}
int32_t OMR::X86::Instruction::estimateBinaryLength(int32_t currentEstimate)
{
self()->setEstimatedBinaryLength(self()->getOpCode().length(self()->rexBits()) + (self()->needsRepPrefix() ? 1 : 0));
return currentEstimate + self()->getEstimatedBinaryLength();
}
// -----------------------------------------------------------------------------
// TR::X86PaddingInstruction:: member functions
uint8_t *TR::X86PaddingInstruction::generateBinaryEncoding()
{
uint8_t *instructionStart = cg()->getBinaryBufferCursor();
uint8_t *cursor = instructionStart;
cursor = cg()->generatePadding(cursor, _length, this, _properties);
setBinaryLength(static_cast<int8_t>(cursor - instructionStart));
setBinaryEncoding(instructionStart);
return cursor;
}
int32_t TR::X86PaddingInstruction::estimateBinaryLength(int32_t currentEstimate)
{
setEstimatedBinaryLength(_length);
return currentEstimate + getEstimatedBinaryLength();
}
TR::Snippet *TR::X86PaddingSnippetInstruction::getSnippetForGC()
{
return _unresolvedSnippet;
}
// -----------------------------------------------------------------------------
// TR::X86BoundaryAvoidanceInstruction:: member functions
OMR::X86::EnlargementResult
TR::X86BoundaryAvoidanceInstruction::enlarge(
int32_t requestedEnlargementSize,
int32_t maxEnlargementSize,
bool allowPartialEnlargement)
{
static char *disableMinPadding = feGetEnv("TR_DisableBoundaryAvoidanceMinPadding");
if (disableMinPadding)
return OMR::X86::EnlargementResult(0, 0);
if ((maxEnlargementSize < requestedEnlargementSize && !allowPartialEnlargement) || requestedEnlargementSize < 1)
return OMR::X86::EnlargementResult(0, 0);
int32_t enlargementSize = std::min(requestedEnlargementSize, maxEnlargementSize);
_minPaddingLength += enlargementSize;
setEstimatedBinaryLength(getEstimatedBinaryLength() + enlargementSize);
return OMR::X86::EnlargementResult(enlargementSize, enlargementSize);
}
int32_t TR::X86BoundaryAvoidanceInstruction::estimateBinaryLength(int32_t currentEstimate)
{
uint8_t totalLength = _minPaddingLength;
// Conceivably, each atomic region could add length-1 bytes of padding
// TODO:AMD64: This is usually very conservative; can we do better?
//
for (const TR_AtomicRegion *cur=_atomicRegions; cur->getLength() > 0; cur++)
{
totalLength += cur->getLength()-1;
}
totalLength = std::max(totalLength, _boundarySpacing);
totalLength = std::min(totalLength, _maxPadding);
totalLength += getSizeOfProtectiveNop();
setEstimatedBinaryLength(totalLength);
return currentEstimate + getEstimatedBinaryLength();
}
uint8_t *TR::X86BoundaryAvoidanceInstruction::generateBinaryEncoding()
{
uint8_t *instructionStart = cg()->getBinaryBufferCursor();
int32_t modulo = ((uintptr_t)(instructionStart + _minPaddingLength)) % _boundarySpacing;
int32_t padLength = 0;
const TR_AtomicRegion *cur = _atomicRegions;
while (cur->getLength() > 0)
{
TR_ASSERT(cur->getLength() <= _boundarySpacing, "Atomic regions of %d bytes can't fit between %d-byte boundaries", cur->getLength(), _boundarySpacing);
int32_t start = cur->getStart() + padLength + modulo;
int32_t end = start + cur->getLength()-1;
if (start / _boundarySpacing == end / _boundarySpacing)
{
// This atomic region can be safely patched
//
cur++;
}
else
{
// Atomic region crosses a boundary; adjust padLength and start over
//
int32_t newPadLength = betterPadLength(padLength, cur, start);
if (newPadLength > _maxPadding)
{
// Can't accomodate this region, so skip it
}
else
{
padLength = newPadLength;
}
// Restart checking atomic regions
//
cur = _atomicRegions;
}
}
setBinaryLength(_minPaddingLength + padLength + getSizeOfProtectiveNop());
setBinaryEncoding(instructionStart);
cg()->addAccumulatedInstructionLengthError(getEstimatedBinaryLength() - getBinaryLength());
uint8_t * afterPadding = cg()->generatePadding(instructionStart, _minPaddingLength + padLength, this);
return cg()->generatePadding(afterPadding, getSizeOfProtectiveNop(), this);
}
// -----------------------------------------------------------------------------
// TR::X86PatchableCodeAlignment
int32_t
TR::X86PatchableCodeAlignmentInstruction::betterPadLength(
int32_t oldPadLength,
const TR_AtomicRegion *unaccommodatedRegion,
int32_t unaccommodatedRegionStart)
{
int32_t result = TR::X86BoundaryAvoidanceInstruction::betterPadLength(oldPadLength, unaccommodatedRegion, unaccommodatedRegionStart);
TR_ASSERT(result < getBoundarySpacing(), "Patchable code has no viable padding that will let it be updated atomically");
return result;
}
// -----------------------------------------------------------------------------
// TR::X86AlignmentInstruction:: member functions
OMR::X86::EnlargementResult
TR::X86AlignmentInstruction::enlarge(
int32_t requestedEnlargementSize,
int32_t maxEnlargementSize,
bool allowPartialEnlargement)
{
static char *disableMinPadding = feGetEnv("TR_DisableAlignmentMinPadding");
if (disableMinPadding)
return OMR::X86::EnlargementResult(0, 0);
if ((maxEnlargementSize < requestedEnlargementSize && !allowPartialEnlargement) || requestedEnlargementSize < 1)
return OMR::X86::EnlargementResult(0, 0);
int32_t enlargementSize = std::min(requestedEnlargementSize, maxEnlargementSize);
_minPaddingLength += enlargementSize;
setEstimatedBinaryLength(getEstimatedBinaryLength() + enlargementSize);
return OMR::X86::EnlargementResult(enlargementSize, enlargementSize);
}
int32_t TR::X86AlignmentInstruction::estimateBinaryLength(int32_t currentEstimate)
{
// Worst case padding is _boundary-1 regardless of margin
//
setEstimatedBinaryLength(_boundary - 1 + _minPaddingLength);
return currentEstimate + getEstimatedBinaryLength();
}
uint8_t *TR::X86AlignmentInstruction::generateBinaryEncoding()
{
uint8_t *instructionStart = cg()->getBinaryBufferCursor();
uint8_t *cursor = instructionStart;
intptr_t paddingLength = cg()->alignment(cursor + _minPaddingLength, _boundary, _margin);
cursor = cg()->generatePadding(cursor, _minPaddingLength + paddingLength, this);
setBinaryLength(static_cast<int8_t>(cursor - instructionStart));
cg()->addAccumulatedInstructionLengthError(getEstimatedBinaryLength() - getBinaryLength());
setBinaryEncoding(instructionStart);
return cursor;
}
// -----------------------------------------------------------------------------
// TR::X86LabelInstruction:: member functions
void
TR::X86LabelInstruction::addMetaDataForCodeAddress(uint8_t *cursor)
{
if (!getOpCode().hasRelativeBranchDisplacement() &&
!(getOpCodeValue() == TR::InstOpCode::label))
{
if (getReloType() == TR_AbsoluteMethodAddress)
{
cg()->addProjectSpecializedRelocation(cursor, NULL, NULL, TR_AbsoluteMethodAddress, __FILE__, __LINE__, getNode());
}
}
}
uint8_t *TR::X86LabelInstruction::generateBinaryEncoding()
{
uint8_t *instructionStart = cg()->getBinaryBufferCursor();
TR::LabelSymbol *label = getLabelSymbol();
uint8_t *cursor = instructionStart;
TR::Compilation *comp = cg()->comp();
uint8_t *immediateCursor;
if (getOpCode().hasRelativeBranchDisplacement())
{
int32_t distance;
if (label == NULL)
{
// Dummy instruction - stays a long branch and gets a zero distance
//
cursor = getOpCode().binary(instructionStart, self()->rexBits());
immediateCursor = cursor;
*(uint32_t *)cursor = 0;
cursor += 4;
}
else
{
if (label->getCodeLocation() != NULL)
{
// Actual distance if target address is known exactly.
//
distance = static_cast<int32_t>(label->getCodeLocation() - (cursor + IA32LengthOfShortBranch));
}
else
{
// Conservative estimate of distance if target address is not known exactly.
// (e.g., a forward relative branch)
//
distance = static_cast<int32_t>(cg()->getBinaryBufferStart()
// +4 == Temporary and possibly incomplete fix for WebSphere problem,
// the buffer start is -4 from the start of the method
//
+ 4
+ label->getEstimatedCodeLocation()
- (cursor + IA32LengthOfShortBranch +
cg()->getAccumulatedInstructionLengthError()));
}
TR_ASSERT(getOpCodeValue() != TR::InstOpCode::XBEGIN4 || !_permitShortening, "TR::InstOpCode::XBEGIN4 cannot be shortened and can only be used with a label instruction that cannot shorten - use generateLongLabel!\n");
if (distance >= -128 && distance <= 127 &&
getOpCode().isBranchOp() && _permitShortening)
{
// Convert long branch to short branch.
//
if (!getOpCode().isShortBranchOp())
{
getOpCode().convertLongBranchToShort();
}
cursor = getOpCode().binary(instructionStart, self()->rexBits());
immediateCursor = cursor;
if (label->getCodeLocation() != NULL)
{
*(int8_t *)cursor = (int8_t)distance;
}
else
{
cg()->addRelocation(new (cg()->trHeapMemory()) TR::LabelRelative8BitRelocation(cursor, label));
*cursor = (uint8_t)(-(intptr_t)(cursor+1));
}
cursor += 1;
}
else
{
if (getOpCode().isShortBranchOp())
{
// If a short branch was explicitly requested but the displacement
// is too large then abort the compile. This should NEVER happen
// in practice because short form branch instructions should
// only be used in controlled circumstances (e.g., PICs) where
// distances are guaranteed.
//
TR_ASSERT(0, "short form branch displacement too large: instr=%p, distance=%d\n", this, distance);
comp->failCompilation<TR::CompilationException>("short form branch displacement too large");
}
cursor = getOpCode().binary(instructionStart, self()->rexBits());
immediateCursor = cursor;
if (label->getCodeLocation() != NULL)
{
*(int32_t *)cursor = distance + IA32LengthOfShortBranch - getOpCode().length(self()->rexBits()) - 4;
}
else
{
cg()->addRelocation(new (cg()->trHeapMemory()) TR::LabelRelative32BitRelocation(cursor, label));
*(uint32_t *)cursor = (uint32_t)(-(intptr_t)(cursor+4));
}
cursor += 4;
}
}
}
else if (getOpCodeValue() == TR::InstOpCode::label)
{
label->setCodeLocation(instructionStart);
immediateCursor = cursor;
}
else // assume absolute code address referencing instruction like push label
{
cursor = getOpCode().binary(instructionStart, self()->rexBits());
immediateCursor = cursor;
cg()->addRelocation(new (cg()->trHeapMemory()) TR::LabelAbsoluteRelocation(cursor, label));
*(uint32_t *)cursor = 0;
cursor += 4;
}
addMetaDataForCodeAddress(immediateCursor);
setBinaryLength(static_cast<int8_t>(cursor - instructionStart));
cg()->addAccumulatedInstructionLengthError(getEstimatedBinaryLength() - getBinaryLength());
setBinaryEncoding(instructionStart);
return cursor;
}
uint8_t TR::X86LabelInstruction::getBinaryLengthLowerBound()
{
if (getOpCodeValue() == TR::InstOpCode::label)
return 0;
if (getOpCodeValue() == TR::InstOpCode::vgnop)
{
return 0;
}
if (getOpCode().isBranchOp())
return getOpCode().length(self()->rexBits()) + (_permitShortening ? 0 : 4);
return getOpCode().length(self()->rexBits()) + 4; // assume absolute code reference
}
OMR::X86::EnlargementResult
TR::X86LabelInstruction::enlarge(
int32_t requestedEnlargementSize,
int32_t maxEnlargementSize,
bool allowPartialEnlargement)
{
static char *disableJMPExpansion = feGetEnv("TR_DisableJMPExpansion");
if (disableJMPExpansion)
return OMR::X86::EnlargementResult(0, 0);
// enlargement for label instructions operates on the following principle:
// in the X86 codegen a branch instruction has a size eg TR::InstOpCode::JMP4 and that size is the
// maximum requested size - we never make a TR::InstOpCode::JMP1 a TR::InstOpCode::JMP4 for example - we trust
// people choosing a lower maximum size know what they are doing
// the binary encoding will, usually try to use the smallest instruction it can
// and so if a TR::InstOpCode::JMP4 can be encoded as a TR::InstOpCode::JMP1 it will be
// this function prevents this instruction shrinking by calling prohibitShortening()
// this increases the size lowerbound of the instruction, thus "enlarging" the
// instruction and, hopefully, making it a suitable candidate for patching
if (!getOpCode().isBranchOp())
return OMR::X86::EnlargementResult(0, 0);
if (!getOpCode().hasIntImmediate() || !_permitShortening || getOpCodeValue() == TR::InstOpCode::vgnop)
return OMR::X86::EnlargementResult(0, 0);
if ((maxEnlargementSize < requestedEnlargementSize && !allowPartialEnlargement)
|| (requestedEnlargementSize > 4 && !allowPartialEnlargement)
|| maxEnlargementSize < 4)
return OMR::X86::EnlargementResult(0, 0);
prohibitShortening();
setEstimatedBinaryLength(getEstimatedBinaryLength() + 4);
return OMR::X86::EnlargementResult(4, 4);
}
int32_t TR::X86LabelInstruction::estimateBinaryLength(int32_t currentEstimate)
{
TR::Compilation *comp = cg()->comp();
if (getOpCode().isBranchOp())
{
uint32_t immediateLength = 1;
if (getOpCode().hasIntImmediate())
{
immediateLength = 4;
if (getLabelSymbol() && getLabelSymbol()->getEstimatedCodeLocation())
{
int32_t distance = getLabelSymbol()->getEstimatedCodeLocation() - (currentEstimate + IA32LengthOfShortBranch);
if (distance >= -128 && distance < 0 && _permitShortening)
{
immediateLength = 0; // really 1, but for conditional branches (all excep TR::InstOpCode::JMP4) the opcode entry will be 1 too big for short branch
// because the short branch op is 1 byte, but the long branch op for conditionals is 2
if (getOpCodeValue() == TR::InstOpCode::JMP4)
{
immediateLength = 1;
}
}
}
}
setEstimatedBinaryLength(getOpCode().length(self()->rexBits()) + immediateLength);
}
else if (getOpCodeValue() == TR::InstOpCode::label)
{
getLabelSymbol()->setEstimatedCodeLocation(currentEstimate);
}
else // assume absolute code reference
{
setEstimatedBinaryLength(getOpCode().length(self()->rexBits()) + 4); // full offset for absolute address reference
}
return currentEstimate + getEstimatedBinaryLength();
}
// -----------------------------------------------------------------------------
// TR::X86FenceInstruction:: member functions
void
TR::X86FenceInstruction::addMetaDataForCodeAddress(uint8_t *cursor)
{
cg()->addProjectSpecializedRelocation( cursor, 0, NULL, TR_AbsoluteMethodAddress,
__FILE__, __LINE__, _fenceNode);
}
uint8_t *TR::X86FenceInstruction::generateBinaryEncoding()
{
uint8_t *instructionStart = cg()->getBinaryBufferCursor();
if (_fenceNode->getRelocationType() == TR_AbsoluteAddress)
{
for (auto i = 0U; i < _fenceNode->getNumRelocations(); ++i)
{
*(uint8_t **)(_fenceNode->getRelocationDestination(i)) = instructionStart;
}
}
else if (_fenceNode->getRelocationType() == TR_ExternalAbsoluteAddress)
{
for (auto i = 0U; i < _fenceNode->getNumRelocations(); ++i)
{
*(uint8_t **)(_fenceNode->getRelocationDestination(i)) = instructionStart;
addMetaDataForCodeAddress( (uint8_t *)_fenceNode->getRelocationDestination(i) );
}
}
else if (_fenceNode->getRelocationType() == TR_EntryRelative32Bit)
{
for (auto i = 0U; i < _fenceNode->getNumRelocations(); ++i)
{
*(uint32_t *)(_fenceNode->getRelocationDestination(i)) = cg()->getCodeLength();
}
}
else // entryrelative16bit
{
for (auto i = 0U; i < _fenceNode->getNumRelocations(); ++i)
{
*(uint16_t *)(_fenceNode->getRelocationDestination(i)) = (uint16_t)cg()->getCodeLength();
}
}
setBinaryEncoding(instructionStart);
return instructionStart;
}
#ifdef J9_PROJECT_SPECIFIC
// -----------------------------------------------------------------------------
// TR::X86VirtualGuardNOPInstruction
int32_t TR::X86VirtualGuardNOPInstruction::estimateBinaryLength(int32_t currentEstimate)
{
setEstimatedBinaryLength(self()->cg()->comp()->target().is64Bit() ? 5 : 6); //TODO:AMD64: What if patched instruction needs a Rex? Is 6 enough?
return currentEstimate + getEstimatedBinaryLength();
}
uint8_t *TR::X86VirtualGuardNOPInstruction::generateBinaryEncoding()
{
TR::Compilation *comp = cg()->comp();
uint8_t *instructionStart = cg()->getBinaryBufferCursor();
uint8_t *cursor = instructionStart;
TR::LabelSymbol *label = getLabelSymbol();
int32_t offset;
uint8_t *patchCursor = cursor;
TR::Instruction *guardForPatching = cg()->getVirtualGuardForPatching(this);
static char *disableOSRNOPs = feGetEnv("TR_disableOSRGuardNOPs");
if (disableOSRNOPs && getNode()->isOSRGuard())
{
setBinaryLength(0);
setBinaryEncoding(instructionStart);
cg()->addAccumulatedInstructionLengthError(getEstimatedBinaryLength());
return cursor;
}
// a previous guard is patching to the same destination and we can recycle the patch point
// so setup the patching location to use this previous guard and generate no instructions
// ourselves
if (guardForPatching != this)
{
_site->setLocation(guardForPatching->getBinaryEncoding());
setBinaryLength(0);
setBinaryEncoding(cursor);
if (label->getCodeLocation() == NULL)
{
cg()->addRelocation(new (cg()->trHeapMemory()) TR::LabelAbsoluteRelocation
((uint8_t *) (&_site->getDestination()), label));
}
else
{
_site->setDestination(label->getCodeLocation());
}
cg()->addAccumulatedInstructionLengthError(getEstimatedBinaryLength() - getBinaryLength());
return cursor;
}
_site->setLocation(patchCursor);
if (label->getCodeLocation() == NULL)
{
// Conservative offset estimate
//
offset = cg()->getBinaryBufferStart() +
label->getEstimatedCodeLocation() -
(patchCursor + IA32LengthOfShortBranch +
cg()->getAccumulatedInstructionLengthError());
// Can't call _site->setDestination because we don't know the destination
// yet, so use a relocation instead.
//
cg()->addRelocation(new (cg()->trHeapMemory()) TR::LabelAbsoluteRelocation
((uint8_t *) (&_site->getDestination()), label));
}
else
{
offset = label->getCodeLocation() - (patchCursor + IA32LengthOfShortBranch);
_site->setDestination(label->getCodeLocation());
}
// guards that do not require atomic patching have a more relaxed sizing constraing since they are only patched while all threads are stopped
bool requiresAtomicPatching = this->requiresAtomicPatching();
uint8_t next = requiresAtomicPatching ?
cg()->sizeOfInstructionToBePatched(this) :
cg()->sizeOfInstructionToBePatchedHCRGuard(this);
//If the current instruction is already a patchable instruction, we cannot use it as a vgnop guard site.
//We will need to insert nops in this case.
//
TR::Instruction *instToBePatched = cg()->getInstructionToBePatched(this);
int32_t shortJumpWidth = 2;
int32_t longJumpWidth = cg()->comp()->target().is64Bit() ? 5 : 6;
_nopSize = 0;
if (offset >= -128 && offset <= 127)
{
if (next < shortJumpWidth || !instToBePatched)
_nopSize = shortJumpWidth;
}
else
{
if (next < longJumpWidth || !instToBePatched)
_nopSize = longJumpWidth;
}
if (cg()->nopsAlsoProcessedByRelocations())
{
_nopSize = longJumpWidth;
}
// in the case of an HCR guard merged with a Profiled Guard, we know the instruction to patch
// is the Profield Guard conditional jump. Because the HCR Guard jumps to the same destination
// we are guaranteed that the patch site will be big enough for the required jump and so we
// force the patch size to longJumpWidth to suppress NOP emission
else if (getNode()->isProfiledGuard())
{
_nopSize = 0;
}
// HCR guards are special because they occur in stop the world so we don't require atomic
// nop sequences - all other guards require a single instruction to prevent possible inconsistent
// state seen by the processor while decoding when patching occurs
if (instToBePatched && _nopSize > 0 && !cg()->nopsAlsoProcessedByRelocations())
{
TR_ASSERT(_nopSize > next, "Emitting nops when the instruction is big enough is always wrong!");
OMR::X86::EnlargementResult enlargement = instToBePatched->enlarge(_nopSize - next, getEstimatedBinaryLength(), !requiresAtomicPatching);
if (enlargement.getPatchGrowth() + next >= _nopSize)
{
_nopSize = 0;
}
else if (!requiresAtomicPatching)
{
_nopSize -= next + enlargement.getPatchGrowth();
if (_nopSize < 0)
_nopSize = 0;
}
else
{
TR_ASSERT(enlargement.getPatchGrowth() == 0, "Instructions should only partially enlarge when dealing with an HCR guard - any other partial enlargement is a waste of code cache!");
}
// donate some of our estimated size to the newly englarged instruction
// the instruction to be patched will have already taken the extra size when computing enlarge
if (enlargement.getEncodingGrowth() > 0)
setEstimatedBinaryLength(getEstimatedBinaryLength() - enlargement.getEncodingGrowth());
}
cursor = cg()->generatePadding(cursor, _nopSize, this,
requiresAtomicPatching ? TR_AtomicNoOpPadding : TR_NoOpPadding);
setBinaryLength(cursor - instructionStart);
setBinaryEncoding(instructionStart);
cg()->addAccumulatedInstructionLengthError(getEstimatedBinaryLength() - getBinaryLength());
return cursor;
}
#endif
// -----------------------------------------------------------------------------
// TR::X86ImmInstruction:: member functions
void
TR::X86ImmInstruction::addMetaDataForCodeAddress(uint8_t *cursor)
{
if (getOpCode().hasIntImmediate())
{
if (needsAOTRelocation())
{
cg()->addExternalRelocation(new (cg()->trHeapMemory()) TR::ExternalRelocation(cursor, 0, TR_BodyInfoAddress, cg()),
__FILE__,
__LINE__,
getNode());
}
if (getReloKind() != -1) // TODO: need to change Body info one to use this
{
TR::SymbolType symbolKind = TR::SymbolType::typeClass;
switch (getReloKind())
{
case TR_StaticRamMethodConst:
case TR_VirtualRamMethodConst:
case TR_SpecialRamMethodConst:
cg()->addExternalRelocation(new (cg()->trHeapMemory()) TR::ExternalRelocation(cursor, (uint8_t *)getNode()->getSymbolReference(),
(uint8_t *)(intptr_t)getNode()->getInlinedSiteIndex() , (TR_ExternalRelocationTargetKind) getReloKind(), cg()),
__FILE__, __LINE__, getNode());
break;
case TR_MethodPointer:
if (getNode() && getNode()->getInlinedSiteIndex() == -1 &&
(void *)(uintptr_t) getSourceImmediate() == cg()->comp()->getCurrentMethod()->resolvedMethodAddress())
setReloKind(TR_RamMethod);
// intentional fall-through
case TR_RamMethod:
symbolKind = TR::SymbolType::typeMethod;
// intentional fall-through
case TR_ClassPointer:
if (cg()->comp()->getOption(TR_UseSymbolValidationManager))
{
cg()->addExternalRelocation(new (cg()->trHeapMemory()) TR::ExternalRelocation(cursor,
(uint8_t *)(uintptr_t)getSourceImmediate(),
(uint8_t *)symbolKind,
TR_SymbolFromManager,
cg()),
__FILE__, __LINE__,
getNode());
}
else
{
cg()->addExternalRelocation(new (cg()->trHeapMemory()) TR::ExternalRelocation(cursor,
(uint8_t*)getNode(),
(TR_ExternalRelocationTargetKind) _reloKind,
cg()),
__FILE__,
__LINE__,
getNode());
}
break;
default:
cg()->addExternalRelocation(new (cg()->trHeapMemory()) TR::ExternalRelocation(cursor, 0, (TR_ExternalRelocationTargetKind)getReloKind(), cg()),
__FILE__,
__LINE__,
getNode());
}
}
if (std::find(cg()->comp()->getStaticHCRPICSites()->begin(), cg()->comp()->getStaticHCRPICSites()->end(), this) != cg()->comp()->getStaticHCRPICSites()->end())
{
cg()->jitAdd32BitPicToPatchOnClassRedefinition(((void *)(uintptr_t) getSourceImmediateAsAddress()), (void *) cursor);
}
}
}
uint8_t* TR::X86ImmInstruction::generateOperand(uint8_t* cursor)
{
uint8_t *immediateCursor = cursor;
if (getOpCode().hasIntImmediate())
{
*(int32_t *)cursor = (int32_t)getSourceImmediate();
if (getOpCode().isCallImmOp())
{
*(int32_t *)cursor -= (int32_t)(intptr_t)(cursor + 4);
}
cursor += 4;
}
else if (getOpCode().hasByteImmediate() || getOpCode().hasSignExtendImmediate())
{
*(int8_t *)cursor = (int8_t)getSourceImmediate();
cursor += 1;
}
else
{
*(int16_t *)cursor = (int16_t)getSourceImmediate();
cursor += 2;
}
addMetaDataForCodeAddress(immediateCursor);
return cursor;
}
uint8_t TR::X86ImmInstruction::getBinaryLengthLowerBound()
{
uint8_t len = getOpCode().length(self()->rexBits());
if (getOpCode().hasIntImmediate())
len += 4;
else if (getOpCode().hasByteImmediate() || getOpCode().hasSignExtendImmediate())
len += 1;
else if (getOpCode().hasShortImmediate())
len += 2;
return len;
}
int32_t TR::X86ImmInstruction::estimateBinaryLength(int32_t currentEstimate)
{
uint32_t immediateLength = 1;
if (getOpCode().hasIntImmediate())
{
immediateLength = 4;
}
else if (getOpCode().hasShortImmediate())
{
immediateLength = 2;
}
setEstimatedBinaryLength(getOpCode().length(self()->rexBits()) + immediateLength);
return currentEstimate + getEstimatedBinaryLength();
}
// -----------------------------------------------------------------------------
// TR::X86ImmSnippetInstruction:: member functions
void
TR::X86ImmSnippetInstruction::addMetaDataForCodeAddress(uint8_t *cursor)
{
if (getOpCode().hasIntImmediate())
{
if (std::find(cg()->comp()->getStaticHCRPICSites()->begin(), cg()->comp()->getStaticHCRPICSites()->end(), this) != cg()->comp()->getStaticHCRPICSites()->end())
{
cg()->jitAdd32BitPicToPatchOnClassRedefinition(((void *)(uintptr_t) getSourceImmediateAsAddress()), (void *) cursor);