This repository was archived by the owner on Nov 1, 2021. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 69
/
Copy pathCodeGenFunction.cpp
1612 lines (1359 loc) · 59.2 KB
/
CodeGenFunction.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
//===--- CodeGenFunction.cpp - Emit LLVM Code from ASTs for a Function ----===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This coordinates the per-function state used while generating code.
//
//===----------------------------------------------------------------------===//
#include "CodeGenFunction.h"
#include "CGCUDARuntime.h"
#include "CGCXXABI.h"
#include "CGDebugInfo.h"
#include "CodeGenModule.h"
#include "CodeGenPGO.h"
#include "TargetInfo.h"
#include "clang/AST/ASTContext.h"
#include "clang/AST/Decl.h"
#include "clang/AST/DeclCXX.h"
#include "clang/AST/StmtCXX.h"
#include "clang/Basic/TargetInfo.h"
#include "clang/CodeGen/CGFunctionInfo.h"
#include "clang/Frontend/CodeGenOptions.h"
#include "llvm/IR/DataLayout.h"
#include "llvm/IR/Intrinsics.h"
#include "llvm/IR/MDBuilder.h"
#include "llvm/IR/Operator.h"
using namespace clang;
using namespace CodeGen;
CodeGenFunction::CodeGenFunction(CodeGenModule &cgm, bool suppressNewContext)
: CodeGenTypeCache(cgm), CGM(cgm), Target(cgm.getTarget()),
Builder(cgm.getModule().getContext()), CapturedStmtInfo(0),
SanitizePerformTypeCheck(CGM.getSanOpts().Null |
CGM.getSanOpts().Alignment |
CGM.getSanOpts().ObjectSize |
CGM.getSanOpts().Vptr),
SanOpts(&CGM.getSanOpts()), AutoreleaseResult(false), BlockInfo(0),
BlockPointer(0), LambdaThisCaptureField(0), NormalCleanupDest(0),
NextCleanupDestIndex(1), FirstBlockInfo(0), EHResumeBlock(0),
ExceptionSlot(0), EHSelectorSlot(0), DebugInfo(CGM.getModuleDebugInfo()),
DisableDebugInfo(false), DidCallStackSave(false), IndirectBranch(0),
PGO(cgm), SwitchInsn(0), SwitchWeights(0),
CaseRangeBlock(0), UnreachableBlock(0), NumReturnExprs(0),
NumSimpleReturnExprs(0), CXXABIThisDecl(0), CXXABIThisValue(0),
CXXThisValue(0), CXXDefaultInitExprThis(0),
CXXStructorImplicitParamDecl(0), CXXStructorImplicitParamValue(0),
OutermostConditional(0), CurLexicalScope(0), TerminateLandingPad(0),
TerminateHandler(0), TrapBB(0) {
if (!suppressNewContext)
CGM.getCXXABI().getMangleContext().startNewFunction();
llvm::FastMathFlags FMF;
if (CGM.getLangOpts().FastMath)
FMF.setUnsafeAlgebra();
if (CGM.getLangOpts().FiniteMathOnly) {
FMF.setNoNaNs();
FMF.setNoInfs();
}
Builder.SetFastMathFlags(FMF);
}
CodeGenFunction::~CodeGenFunction() {
assert(LifetimeExtendedCleanupStack.empty() && "failed to emit a cleanup");
// If there are any unclaimed block infos, go ahead and destroy them
// now. This can happen if IR-gen gets clever and skips evaluating
// something.
if (FirstBlockInfo)
destroyBlockInfos(FirstBlockInfo);
}
llvm::Type *CodeGenFunction::ConvertTypeForMem(QualType T) {
return CGM.getTypes().ConvertTypeForMem(T);
}
llvm::Type *CodeGenFunction::ConvertType(QualType T) {
return CGM.getTypes().ConvertType(T);
}
TypeEvaluationKind CodeGenFunction::getEvaluationKind(QualType type) {
type = type.getCanonicalType();
while (true) {
switch (type->getTypeClass()) {
#define TYPE(name, parent)
#define ABSTRACT_TYPE(name, parent)
#define NON_CANONICAL_TYPE(name, parent) case Type::name:
#define DEPENDENT_TYPE(name, parent) case Type::name:
#define NON_CANONICAL_UNLESS_DEPENDENT_TYPE(name, parent) case Type::name:
#include "clang/AST/TypeNodes.def"
llvm_unreachable("non-canonical or dependent type in IR-generation");
case Type::Auto:
llvm_unreachable("undeduced auto type in IR-generation");
// Various scalar types.
case Type::Builtin:
case Type::Pointer:
case Type::BlockPointer:
case Type::LValueReference:
case Type::RValueReference:
case Type::MemberPointer:
case Type::Vector:
case Type::ExtVector:
case Type::FunctionProto:
case Type::FunctionNoProto:
case Type::Enum:
case Type::ObjCObjectPointer:
return TEK_Scalar;
// Complexes.
case Type::Complex:
return TEK_Complex;
// Arrays, records, and Objective-C objects.
case Type::ConstantArray:
case Type::IncompleteArray:
case Type::VariableArray:
case Type::Record:
case Type::ObjCObject:
case Type::ObjCInterface:
return TEK_Aggregate;
// We operate on atomic values according to their underlying type.
case Type::Atomic:
type = cast<AtomicType>(type)->getValueType();
continue;
}
llvm_unreachable("unknown type kind!");
}
}
void CodeGenFunction::EmitReturnBlock() {
// For cleanliness, we try to avoid emitting the return block for
// simple cases.
llvm::BasicBlock *CurBB = Builder.GetInsertBlock();
if (CurBB) {
assert(!CurBB->getTerminator() && "Unexpected terminated block.");
// We have a valid insert point, reuse it if it is empty or there are no
// explicit jumps to the return block.
if (CurBB->empty() || ReturnBlock.getBlock()->use_empty()) {
ReturnBlock.getBlock()->replaceAllUsesWith(CurBB);
delete ReturnBlock.getBlock();
} else
EmitBlock(ReturnBlock.getBlock());
return;
}
// Otherwise, if the return block is the target of a single direct
// branch then we can just put the code in that block instead. This
// cleans up functions which started with a unified return block.
if (ReturnBlock.getBlock()->hasOneUse()) {
llvm::BranchInst *BI =
dyn_cast<llvm::BranchInst>(*ReturnBlock.getBlock()->use_begin());
if (BI && BI->isUnconditional() &&
BI->getSuccessor(0) == ReturnBlock.getBlock()) {
// Reset insertion point, including debug location, and delete the
// branch. This is really subtle and only works because the next change
// in location will hit the caching in CGDebugInfo::EmitLocation and not
// override this.
Builder.SetCurrentDebugLocation(BI->getDebugLoc());
Builder.SetInsertPoint(BI->getParent());
BI->eraseFromParent();
delete ReturnBlock.getBlock();
return;
}
}
// FIXME: We are at an unreachable point, there is no reason to emit the block
// unless it has uses. However, we still need a place to put the debug
// region.end for now.
EmitBlock(ReturnBlock.getBlock());
}
static void EmitIfUsed(CodeGenFunction &CGF, llvm::BasicBlock *BB) {
if (!BB) return;
if (!BB->use_empty())
return CGF.CurFn->getBasicBlockList().push_back(BB);
delete BB;
}
void CodeGenFunction::FinishFunction(SourceLocation EndLoc) {
assert(BreakContinueStack.empty() &&
"mismatched push/pop in break/continue stack!");
bool OnlySimpleReturnStmts = NumSimpleReturnExprs > 0
&& NumSimpleReturnExprs == NumReturnExprs
&& ReturnBlock.getBlock()->use_empty();
// Usually the return expression is evaluated before the cleanup
// code. If the function contains only a simple return statement,
// such as a constant, the location before the cleanup code becomes
// the last useful breakpoint in the function, because the simple
// return expression will be evaluated after the cleanup code. To be
// safe, set the debug location for cleanup code to the location of
// the return statement. Otherwise the cleanup code should be at the
// end of the function's lexical scope.
//
// If there are multiple branches to the return block, the branch
// instructions will get the location of the return statements and
// all will be fine.
if (CGDebugInfo *DI = getDebugInfo()) {
if (OnlySimpleReturnStmts)
DI->EmitLocation(Builder, LastStopPoint);
else
DI->EmitLocation(Builder, EndLoc);
}
// Pop any cleanups that might have been associated with the
// parameters. Do this in whatever block we're currently in; it's
// important to do this before we enter the return block or return
// edges will be *really* confused.
bool EmitRetDbgLoc = true;
if (EHStack.stable_begin() != PrologueCleanupDepth) {
PopCleanupBlocks(PrologueCleanupDepth);
// Make sure the line table doesn't jump back into the body for
// the ret after it's been at EndLoc.
EmitRetDbgLoc = false;
if (CGDebugInfo *DI = getDebugInfo())
if (OnlySimpleReturnStmts)
DI->EmitLocation(Builder, EndLoc);
}
// Emit function epilog (to return).
EmitReturnBlock();
if (ShouldInstrumentFunction())
EmitFunctionInstrumentation("__cyg_profile_func_exit");
// Emit debug descriptor for function end.
if (CGDebugInfo *DI = getDebugInfo()) {
DI->EmitFunctionEnd(Builder);
}
EmitFunctionEpilog(*CurFnInfo, EmitRetDbgLoc, EndLoc);
EmitEndEHSpec(CurCodeDecl);
assert(EHStack.empty() &&
"did not remove all scopes from cleanup stack!");
// If someone did an indirect goto, emit the indirect goto block at the end of
// the function.
if (IndirectBranch) {
EmitBlock(IndirectBranch->getParent());
Builder.ClearInsertionPoint();
}
// Remove the AllocaInsertPt instruction, which is just a convenience for us.
llvm::Instruction *Ptr = AllocaInsertPt;
AllocaInsertPt = 0;
Ptr->eraseFromParent();
// If someone took the address of a label but never did an indirect goto, we
// made a zero entry PHI node, which is illegal, zap it now.
if (IndirectBranch) {
llvm::PHINode *PN = cast<llvm::PHINode>(IndirectBranch->getAddress());
if (PN->getNumIncomingValues() == 0) {
PN->replaceAllUsesWith(llvm::UndefValue::get(PN->getType()));
PN->eraseFromParent();
}
}
EmitIfUsed(*this, EHResumeBlock);
EmitIfUsed(*this, TerminateLandingPad);
EmitIfUsed(*this, TerminateHandler);
EmitIfUsed(*this, UnreachableBlock);
if (CGM.getCodeGenOpts().EmitDeclMetadata)
EmitDeclMetadata();
for (SmallVectorImpl<std::pair<llvm::Instruction *, llvm::Value *> >::iterator
I = DeferredReplacements.begin(),
E = DeferredReplacements.end();
I != E; ++I) {
I->first->replaceAllUsesWith(I->second);
I->first->eraseFromParent();
}
}
/// ShouldInstrumentFunction - Return true if the current function should be
/// instrumented with __cyg_profile_func_* calls
bool CodeGenFunction::ShouldInstrumentFunction() {
if (!CGM.getCodeGenOpts().InstrumentFunctions)
return false;
if (!CurFuncDecl || CurFuncDecl->hasAttr<NoInstrumentFunctionAttr>())
return false;
return true;
}
/// EmitFunctionInstrumentation - Emit LLVM code to call the specified
/// instrumentation function with the current function and the call site, if
/// function instrumentation is enabled.
void CodeGenFunction::EmitFunctionInstrumentation(const char *Fn) {
// void __cyg_profile_func_{enter,exit} (void *this_fn, void *call_site);
llvm::PointerType *PointerTy = Int8PtrTy;
llvm::Type *ProfileFuncArgs[] = { PointerTy, PointerTy };
llvm::FunctionType *FunctionTy =
llvm::FunctionType::get(VoidTy, ProfileFuncArgs, false);
llvm::Constant *F = CGM.CreateRuntimeFunction(FunctionTy, Fn);
llvm::CallInst *CallSite = Builder.CreateCall(
CGM.getIntrinsic(llvm::Intrinsic::returnaddress),
llvm::ConstantInt::get(Int32Ty, 0),
"callsite");
llvm::Value *args[] = {
llvm::ConstantExpr::getBitCast(CurFn, PointerTy),
CallSite
};
EmitNounwindRuntimeCall(F, args);
}
void CodeGenFunction::EmitMCountInstrumentation() {
llvm::FunctionType *FTy = llvm::FunctionType::get(VoidTy, false);
llvm::Constant *MCountFn =
CGM.CreateRuntimeFunction(FTy, getTarget().getMCountName());
EmitNounwindRuntimeCall(MCountFn);
}
// OpenCL v1.2 s5.6.4.6 allows the compiler to store kernel argument
// information in the program executable. The argument information stored
// includes the argument name, its type, the address and access qualifiers used.
static void GenOpenCLArgMetadata(const FunctionDecl *FD, llvm::Function *Fn,
CodeGenModule &CGM,llvm::LLVMContext &Context,
SmallVector <llvm::Value*, 5> &kernelMDArgs,
CGBuilderTy& Builder, ASTContext &ASTCtx) {
// Create MDNodes that represent the kernel arg metadata.
// Each MDNode is a list in the form of "key", N number of values which is
// the same number of values as their are kernel arguments.
// MDNode for the kernel argument address space qualifiers.
SmallVector<llvm::Value*, 8> addressQuals;
addressQuals.push_back(llvm::MDString::get(Context, "kernel_arg_addr_space"));
// MDNode for the kernel argument access qualifiers (images only).
SmallVector<llvm::Value*, 8> accessQuals;
accessQuals.push_back(llvm::MDString::get(Context, "kernel_arg_access_qual"));
// MDNode for the kernel argument type names.
SmallVector<llvm::Value*, 8> argTypeNames;
argTypeNames.push_back(llvm::MDString::get(Context, "kernel_arg_type"));
// MDNode for the kernel argument type qualifiers.
SmallVector<llvm::Value*, 8> argTypeQuals;
argTypeQuals.push_back(llvm::MDString::get(Context, "kernel_arg_type_qual"));
// MDNode for the kernel argument names.
SmallVector<llvm::Value*, 8> argNames;
argNames.push_back(llvm::MDString::get(Context, "kernel_arg_name"));
for (unsigned i = 0, e = FD->getNumParams(); i != e; ++i) {
const ParmVarDecl *parm = FD->getParamDecl(i);
QualType ty = parm->getType();
std::string typeQuals;
if (ty->isPointerType()) {
QualType pointeeTy = ty->getPointeeType();
// Get address qualifier.
addressQuals.push_back(Builder.getInt32(ASTCtx.getTargetAddressSpace(
pointeeTy.getAddressSpace())));
// Get argument type name.
std::string typeName = pointeeTy.getUnqualifiedType().getAsString() + "*";
// Turn "unsigned type" to "utype"
std::string::size_type pos = typeName.find("unsigned");
if (pos != std::string::npos)
typeName.erase(pos+1, 8);
argTypeNames.push_back(llvm::MDString::get(Context, typeName));
// Get argument type qualifiers:
if (ty.isRestrictQualified())
typeQuals = "restrict";
if (pointeeTy.isConstQualified() ||
(pointeeTy.getAddressSpace() == LangAS::opencl_constant))
typeQuals += typeQuals.empty() ? "const" : " const";
if (pointeeTy.isVolatileQualified())
typeQuals += typeQuals.empty() ? "volatile" : " volatile";
} else {
uint32_t AddrSpc = 0;
if (ty->isImageType())
AddrSpc =
CGM.getContext().getTargetAddressSpace(LangAS::opencl_global);
addressQuals.push_back(Builder.getInt32(AddrSpc));
// Get argument type name.
std::string typeName = ty.getUnqualifiedType().getAsString();
// Turn "unsigned type" to "utype"
std::string::size_type pos = typeName.find("unsigned");
if (pos != std::string::npos)
typeName.erase(pos+1, 8);
argTypeNames.push_back(llvm::MDString::get(Context, typeName));
// Get argument type qualifiers:
if (ty.isConstQualified())
typeQuals = "const";
if (ty.isVolatileQualified())
typeQuals += typeQuals.empty() ? "volatile" : " volatile";
}
argTypeQuals.push_back(llvm::MDString::get(Context, typeQuals));
// Get image access qualifier:
if (ty->isImageType()) {
const OpenCLImageAccessAttr *A = parm->getAttr<OpenCLImageAccessAttr>();
if (A && A->isWriteOnly())
accessQuals.push_back(llvm::MDString::get(Context, "write_only"));
else
accessQuals.push_back(llvm::MDString::get(Context, "read_only"));
// FIXME: what about read_write?
} else
accessQuals.push_back(llvm::MDString::get(Context, "none"));
// Get argument name.
argNames.push_back(llvm::MDString::get(Context, parm->getName()));
}
kernelMDArgs.push_back(llvm::MDNode::get(Context, addressQuals));
kernelMDArgs.push_back(llvm::MDNode::get(Context, accessQuals));
kernelMDArgs.push_back(llvm::MDNode::get(Context, argTypeNames));
kernelMDArgs.push_back(llvm::MDNode::get(Context, argTypeQuals));
kernelMDArgs.push_back(llvm::MDNode::get(Context, argNames));
}
void CodeGenFunction::EmitOpenCLKernelMetadata(const FunctionDecl *FD,
llvm::Function *Fn)
{
if (!FD->hasAttr<OpenCLKernelAttr>())
return;
llvm::LLVMContext &Context = getLLVMContext();
SmallVector <llvm::Value*, 5> kernelMDArgs;
kernelMDArgs.push_back(Fn);
if (CGM.getCodeGenOpts().EmitOpenCLArgMetadata)
GenOpenCLArgMetadata(FD, Fn, CGM, Context, kernelMDArgs,
Builder, getContext());
if (const VecTypeHintAttr *A = FD->getAttr<VecTypeHintAttr>()) {
QualType hintQTy = A->getTypeHint();
const ExtVectorType *hintEltQTy = hintQTy->getAs<ExtVectorType>();
bool isSignedInteger =
hintQTy->isSignedIntegerType() ||
(hintEltQTy && hintEltQTy->getElementType()->isSignedIntegerType());
llvm::Value *attrMDArgs[] = {
llvm::MDString::get(Context, "vec_type_hint"),
llvm::UndefValue::get(CGM.getTypes().ConvertType(A->getTypeHint())),
llvm::ConstantInt::get(
llvm::IntegerType::get(Context, 32),
llvm::APInt(32, (uint64_t)(isSignedInteger ? 1 : 0)))
};
kernelMDArgs.push_back(llvm::MDNode::get(Context, attrMDArgs));
}
if (const WorkGroupSizeHintAttr *A = FD->getAttr<WorkGroupSizeHintAttr>()) {
llvm::Value *attrMDArgs[] = {
llvm::MDString::get(Context, "work_group_size_hint"),
Builder.getInt32(A->getXDim()),
Builder.getInt32(A->getYDim()),
Builder.getInt32(A->getZDim())
};
kernelMDArgs.push_back(llvm::MDNode::get(Context, attrMDArgs));
}
if (const ReqdWorkGroupSizeAttr *A = FD->getAttr<ReqdWorkGroupSizeAttr>()) {
llvm::Value *attrMDArgs[] = {
llvm::MDString::get(Context, "reqd_work_group_size"),
Builder.getInt32(A->getXDim()),
Builder.getInt32(A->getYDim()),
Builder.getInt32(A->getZDim())
};
kernelMDArgs.push_back(llvm::MDNode::get(Context, attrMDArgs));
}
llvm::MDNode *kernelMDNode = llvm::MDNode::get(Context, kernelMDArgs);
llvm::NamedMDNode *OpenCLKernelMetadata =
CGM.getModule().getOrInsertNamedMetadata("opencl.kernels");
OpenCLKernelMetadata->addOperand(kernelMDNode);
}
void CodeGenFunction::StartFunction(GlobalDecl GD,
QualType RetTy,
llvm::Function *Fn,
const CGFunctionInfo &FnInfo,
const FunctionArgList &Args,
SourceLocation StartLoc) {
const Decl *D = GD.getDecl();
DidCallStackSave = false;
CurCodeDecl = D;
CurFuncDecl = (D ? D->getNonClosureContext() : 0);
FnRetTy = RetTy;
CurFn = Fn;
CurFnInfo = &FnInfo;
assert(CurFn->isDeclaration() && "Function already has body?");
if (CGM.getSanitizerBlacklist().isIn(*Fn)) {
SanOpts = &SanitizerOptions::Disabled;
SanitizePerformTypeCheck = false;
}
// Pass inline keyword to optimizer if it appears explicitly on any
// declaration. Also, in the case of -fno-inline attach NoInline
// attribute to all function that are not marked AlwaysInline.
if (const FunctionDecl *FD = dyn_cast_or_null<FunctionDecl>(D)) {
if (!CGM.getCodeGenOpts().NoInline) {
for (FunctionDecl::redecl_iterator RI = FD->redecls_begin(),
RE = FD->redecls_end(); RI != RE; ++RI)
if (RI->isInlineSpecified()) {
Fn->addFnAttr(llvm::Attribute::InlineHint);
break;
}
} else if (!FD->hasAttr<AlwaysInlineAttr>())
Fn->addFnAttr(llvm::Attribute::NoInline);
}
if (getLangOpts().OpenCL) {
// Add metadata for a kernel function.
if (const FunctionDecl *FD = dyn_cast_or_null<FunctionDecl>(D))
EmitOpenCLKernelMetadata(FD, Fn);
}
// If we are checking function types, emit a function type signature as
// prefix data.
if (getLangOpts().CPlusPlus && SanOpts->Function) {
if (const FunctionDecl *FD = dyn_cast_or_null<FunctionDecl>(D)) {
if (llvm::Constant *PrefixSig =
CGM.getTargetCodeGenInfo().getUBSanFunctionSignature(CGM)) {
llvm::Constant *FTRTTIConst =
CGM.GetAddrOfRTTIDescriptor(FD->getType(), /*ForEH=*/true);
llvm::Constant *PrefixStructElems[] = { PrefixSig, FTRTTIConst };
llvm::Constant *PrefixStructConst =
llvm::ConstantStruct::getAnon(PrefixStructElems, /*Packed=*/true);
Fn->setPrefixData(PrefixStructConst);
}
}
}
llvm::BasicBlock *EntryBB = createBasicBlock("entry", CurFn);
// Create a marker to make it easy to insert allocas into the entryblock
// later. Don't create this with the builder, because we don't want it
// folded.
llvm::Value *Undef = llvm::UndefValue::get(Int32Ty);
AllocaInsertPt = new llvm::BitCastInst(Undef, Int32Ty, "", EntryBB);
if (Builder.isNamePreserving())
AllocaInsertPt->setName("allocapt");
ReturnBlock = getJumpDestInCurrentScope("return");
Builder.SetInsertPoint(EntryBB);
// Emit subprogram debug descriptor.
if (CGDebugInfo *DI = getDebugInfo()) {
SmallVector<QualType, 16> ArgTypes;
for (FunctionArgList::const_iterator i = Args.begin(), e = Args.end();
i != e; ++i) {
ArgTypes.push_back((*i)->getType());
}
QualType FnType =
getContext().getFunctionType(RetTy, ArgTypes,
FunctionProtoType::ExtProtoInfo());
DI->setLocation(StartLoc);
DI->EmitFunctionStart(GD, FnType, CurFn, Builder);
}
if (ShouldInstrumentFunction())
EmitFunctionInstrumentation("__cyg_profile_func_enter");
if (CGM.getCodeGenOpts().InstrumentForProfiling)
EmitMCountInstrumentation();
PGO.assignRegionCounters(GD);
if (CGM.getPGOData() && D) {
// Turn on InlineHint attribute for hot functions.
if (CGM.getPGOData()->isHotFunction(CGM.getMangledName(GD)))
Fn->addFnAttr(llvm::Attribute::InlineHint);
// Turn on Cold attribute for cold functions.
else if (CGM.getPGOData()->isColdFunction(CGM.getMangledName(GD)))
Fn->addFnAttr(llvm::Attribute::Cold);
}
if (RetTy->isVoidType()) {
// Void type; nothing to return.
ReturnValue = 0;
} else if (CurFnInfo->getReturnInfo().getKind() == ABIArgInfo::Indirect &&
!hasScalarEvaluationKind(CurFnInfo->getReturnType())) {
// Indirect aggregate return; emit returned value directly into sret slot.
// This reduces code size, and affects correctness in C++.
ReturnValue = CurFn->arg_begin();
} else if (CurFnInfo->getReturnInfo().getKind() == ABIArgInfo::InAlloca &&
!hasScalarEvaluationKind(CurFnInfo->getReturnType())) {
// Load the sret pointer from the argument struct and return into that.
unsigned Idx = CurFnInfo->getReturnInfo().getInAllocaFieldIndex();
llvm::Function::arg_iterator EI = CurFn->arg_end();
--EI;
llvm::Value *Addr = Builder.CreateStructGEP(EI, Idx);
ReturnValue = Builder.CreateLoad(Addr, "agg.result");
} else {
ReturnValue = CreateIRTemp(RetTy, "retval");
// Tell the epilog emitter to autorelease the result. We do this
// now so that various specialized functions can suppress it
// during their IR-generation.
if (getLangOpts().ObjCAutoRefCount &&
!CurFnInfo->isReturnsRetained() &&
RetTy->isObjCRetainableType())
AutoreleaseResult = true;
}
EmitStartEHSpec(CurCodeDecl);
PrologueCleanupDepth = EHStack.stable_begin();
EmitFunctionProlog(*CurFnInfo, CurFn, Args);
if (D && isa<CXXMethodDecl>(D) && cast<CXXMethodDecl>(D)->isInstance()) {
CGM.getCXXABI().EmitInstanceFunctionProlog(*this);
const CXXMethodDecl *MD = cast<CXXMethodDecl>(D);
if (MD->getParent()->isLambda() &&
MD->getOverloadedOperator() == OO_Call) {
// We're in a lambda; figure out the captures.
MD->getParent()->getCaptureFields(LambdaCaptureFields,
LambdaThisCaptureField);
if (LambdaThisCaptureField) {
// If this lambda captures this, load it.
LValue ThisLValue = EmitLValueForLambdaField(LambdaThisCaptureField);
CXXThisValue = EmitLoadOfLValue(ThisLValue,
SourceLocation()).getScalarVal();
}
} else {
// Not in a lambda; just use 'this' from the method.
// FIXME: Should we generate a new load for each use of 'this'? The
// fast register allocator would be happier...
CXXThisValue = CXXABIThisValue;
}
}
// If any of the arguments have a variably modified type, make sure to
// emit the type size.
for (FunctionArgList::const_iterator i = Args.begin(), e = Args.end();
i != e; ++i) {
const VarDecl *VD = *i;
// Dig out the type as written from ParmVarDecls; it's unclear whether
// the standard (C99 6.9.1p10) requires this, but we're following the
// precedent set by gcc.
QualType Ty;
if (const ParmVarDecl *PVD = dyn_cast<ParmVarDecl>(VD))
Ty = PVD->getOriginalType();
else
Ty = VD->getType();
if (Ty->isVariablyModifiedType())
EmitVariablyModifiedType(Ty);
}
// Emit a location at the end of the prologue.
if (CGDebugInfo *DI = getDebugInfo())
DI->EmitLocation(Builder, StartLoc);
}
void CodeGenFunction::EmitFunctionBody(FunctionArgList &Args,
const Stmt *Body) {
RegionCounter Cnt = getPGORegionCounter(Body);
Cnt.beginRegion(Builder);
if (const CompoundStmt *S = dyn_cast<CompoundStmt>(Body))
EmitCompoundStmtWithoutScope(*S);
else
EmitStmt(Body);
}
/// When instrumenting to collect profile data, the counts for some blocks
/// such as switch cases need to not include the fall-through counts, so
/// emit a branch around the instrumentation code. When not instrumenting,
/// this just calls EmitBlock().
void CodeGenFunction::EmitBlockWithFallThrough(llvm::BasicBlock *BB,
RegionCounter &Cnt) {
llvm::BasicBlock *SkipCountBB = 0;
if (HaveInsertPoint() && CGM.getCodeGenOpts().ProfileInstrGenerate) {
// When instrumenting for profiling, the fallthrough to certain
// statements needs to skip over the instrumentation code so that we
// get an accurate count.
SkipCountBB = createBasicBlock("skipcount");
EmitBranch(SkipCountBB);
}
EmitBlock(BB);
Cnt.beginRegion(Builder, /*AddIncomingFallThrough=*/true);
if (SkipCountBB)
EmitBlock(SkipCountBB);
}
/// Tries to mark the given function nounwind based on the
/// non-existence of any throwing calls within it. We believe this is
/// lightweight enough to do at -O0.
static void TryMarkNoThrow(llvm::Function *F) {
// LLVM treats 'nounwind' on a function as part of the type, so we
// can't do this on functions that can be overwritten.
if (F->mayBeOverridden()) return;
for (llvm::Function::iterator FI = F->begin(), FE = F->end(); FI != FE; ++FI)
for (llvm::BasicBlock::iterator
BI = FI->begin(), BE = FI->end(); BI != BE; ++BI)
if (llvm::CallInst *Call = dyn_cast<llvm::CallInst>(&*BI)) {
if (!Call->doesNotThrow())
return;
} else if (isa<llvm::ResumeInst>(&*BI)) {
return;
}
F->setDoesNotThrow();
}
static void EmitSizedDeallocationFunction(CodeGenFunction &CGF,
const FunctionDecl *UnsizedDealloc) {
// This is a weak discardable definition of the sized deallocation function.
CGF.CurFn->setLinkage(llvm::Function::LinkOnceAnyLinkage);
// Call the unsized deallocation function and forward the first argument
// unchanged.
llvm::Constant *Unsized = CGF.CGM.GetAddrOfFunction(UnsizedDealloc);
CGF.Builder.CreateCall(Unsized, &*CGF.CurFn->arg_begin());
}
void CodeGenFunction::GenerateCode(GlobalDecl GD, llvm::Function *Fn,
const CGFunctionInfo &FnInfo) {
const FunctionDecl *FD = cast<FunctionDecl>(GD.getDecl());
// Check if we should generate debug info for this function.
if (FD->hasAttr<NoDebugAttr>())
DebugInfo = NULL; // disable debug info indefinitely for this function
FunctionArgList Args;
QualType ResTy = FD->getReturnType();
CurGD = GD;
const CXXMethodDecl *MD = dyn_cast<CXXMethodDecl>(FD);
if (MD && MD->isInstance()) {
if (CGM.getCXXABI().HasThisReturn(GD))
ResTy = MD->getThisType(getContext());
CGM.getCXXABI().buildThisParam(*this, Args);
}
for (unsigned i = 0, e = FD->getNumParams(); i != e; ++i)
Args.push_back(FD->getParamDecl(i));
if (MD && (isa<CXXConstructorDecl>(MD) || isa<CXXDestructorDecl>(MD)))
CGM.getCXXABI().addImplicitStructorParams(*this, ResTy, Args);
SourceRange BodyRange;
if (Stmt *Body = FD->getBody()) BodyRange = Body->getSourceRange();
CurEHLocation = BodyRange.getEnd();
// Emit the standard function prologue.
StartFunction(GD, ResTy, Fn, FnInfo, Args, BodyRange.getBegin());
// Generate the body of the function.
if (isa<CXXDestructorDecl>(FD))
EmitDestructorBody(Args);
else if (isa<CXXConstructorDecl>(FD))
EmitConstructorBody(Args);
else if (getLangOpts().CUDA &&
!CGM.getCodeGenOpts().CUDAIsDevice &&
FD->hasAttr<CUDAGlobalAttr>())
CGM.getCUDARuntime().EmitDeviceStubBody(*this, Args);
else if (isa<CXXConversionDecl>(FD) &&
cast<CXXConversionDecl>(FD)->isLambdaToBlockPointerConversion()) {
// The lambda conversion to block pointer is special; the semantics can't be
// expressed in the AST, so IRGen needs to special-case it.
EmitLambdaToBlockPointerBody(Args);
} else if (isa<CXXMethodDecl>(FD) &&
cast<CXXMethodDecl>(FD)->isLambdaStaticInvoker()) {
// The lambda static invoker function is special, because it forwards or
// clones the body of the function call operator (but is actually static).
EmitLambdaStaticInvokeFunction(cast<CXXMethodDecl>(FD));
} else if (FD->isDefaulted() && isa<CXXMethodDecl>(FD) &&
(cast<CXXMethodDecl>(FD)->isCopyAssignmentOperator() ||
cast<CXXMethodDecl>(FD)->isMoveAssignmentOperator())) {
// Implicit copy-assignment gets the same special treatment as implicit
// copy-constructors.
emitImplicitAssignmentOperatorBody(Args);
} else if (Stmt *Body = FD->getBody()) {
EmitFunctionBody(Args, Body);
} else if (FunctionDecl *UnsizedDealloc =
FD->getCorrespondingUnsizedGlobalDeallocationFunction()) {
// Global sized deallocation functions get an implicit weak definition if
// they don't have an explicit definition.
EmitSizedDeallocationFunction(*this, UnsizedDealloc);
} else
llvm_unreachable("no definition for emitted function");
// C++11 [stmt.return]p2:
// Flowing off the end of a function [...] results in undefined behavior in
// a value-returning function.
// C11 6.9.1p12:
// If the '}' that terminates a function is reached, and the value of the
// function call is used by the caller, the behavior is undefined.
if (getLangOpts().CPlusPlus && !FD->hasImplicitReturnZero() &&
!FD->getReturnType()->isVoidType() && Builder.GetInsertBlock()) {
if (SanOpts->Return)
EmitCheck(Builder.getFalse(), "missing_return",
EmitCheckSourceLocation(FD->getLocation()),
ArrayRef<llvm::Value *>(), CRK_Unrecoverable);
else if (CGM.getCodeGenOpts().OptimizationLevel == 0)
Builder.CreateCall(CGM.getIntrinsic(llvm::Intrinsic::trap));
Builder.CreateUnreachable();
Builder.ClearInsertionPoint();
}
// Emit the standard function epilogue.
FinishFunction(BodyRange.getEnd());
// If we haven't marked the function nothrow through other means, do
// a quick pass now to see if we can.
if (!CurFn->doesNotThrow())
TryMarkNoThrow(CurFn);
PGO.emitWriteoutFunction(CurGD);
PGO.destroyRegionCounters();
}
/// ContainsLabel - Return true if the statement contains a label in it. If
/// this statement is not executed normally, it not containing a label means
/// that we can just remove the code.
bool CodeGenFunction::ContainsLabel(const Stmt *S, bool IgnoreCaseStmts) {
// Null statement, not a label!
if (S == 0) return false;
// If this is a label, we have to emit the code, consider something like:
// if (0) { ... foo: bar(); } goto foo;
//
// TODO: If anyone cared, we could track __label__'s, since we know that you
// can't jump to one from outside their declared region.
if (isa<LabelStmt>(S))
return true;
// If this is a case/default statement, and we haven't seen a switch, we have
// to emit the code.
if (isa<SwitchCase>(S) && !IgnoreCaseStmts)
return true;
// If this is a switch statement, we want to ignore cases below it.
if (isa<SwitchStmt>(S))
IgnoreCaseStmts = true;
// Scan subexpressions for verboten labels.
for (Stmt::const_child_range I = S->children(); I; ++I)
if (ContainsLabel(*I, IgnoreCaseStmts))
return true;
return false;
}
/// containsBreak - Return true if the statement contains a break out of it.
/// If the statement (recursively) contains a switch or loop with a break
/// inside of it, this is fine.
bool CodeGenFunction::containsBreak(const Stmt *S) {
// Null statement, not a label!
if (S == 0) return false;
// If this is a switch or loop that defines its own break scope, then we can
// include it and anything inside of it.
if (isa<SwitchStmt>(S) || isa<WhileStmt>(S) || isa<DoStmt>(S) ||
isa<ForStmt>(S))
return false;
if (isa<BreakStmt>(S))
return true;
// Scan subexpressions for verboten breaks.
for (Stmt::const_child_range I = S->children(); I; ++I)
if (containsBreak(*I))
return true;
return false;
}
/// ConstantFoldsToSimpleInteger - If the specified expression does not fold
/// to a constant, or if it does but contains a label, return false. If it
/// constant folds return true and set the boolean result in Result.
bool CodeGenFunction::ConstantFoldsToSimpleInteger(const Expr *Cond,
bool &ResultBool) {
llvm::APSInt ResultInt;
if (!ConstantFoldsToSimpleInteger(Cond, ResultInt))
return false;
ResultBool = ResultInt.getBoolValue();
return true;
}
/// ConstantFoldsToSimpleInteger - If the specified expression does not fold
/// to a constant, or if it does but contains a label, return false. If it
/// constant folds return true and set the folded value.
bool CodeGenFunction::
ConstantFoldsToSimpleInteger(const Expr *Cond, llvm::APSInt &ResultInt) {
// FIXME: Rename and handle conversion of other evaluatable things
// to bool.
llvm::APSInt Int;
if (!Cond->EvaluateAsInt(Int, getContext()))
return false; // Not foldable, not integer or not fully evaluatable.
if (CodeGenFunction::ContainsLabel(Cond))
return false; // Contains a label.
ResultInt = Int;
return true;
}
/// EmitBranchOnBoolExpr - Emit a branch on a boolean condition (e.g. for an if
/// statement) to the specified blocks. Based on the condition, this might try
/// to simplify the codegen of the conditional based on the branch.
///
void CodeGenFunction::EmitBranchOnBoolExpr(const Expr *Cond,
llvm::BasicBlock *TrueBlock,
llvm::BasicBlock *FalseBlock,
uint64_t TrueCount) {
Cond = Cond->IgnoreParens();
if (const BinaryOperator *CondBOp = dyn_cast<BinaryOperator>(Cond)) {
// Handle X && Y in a condition.
if (CondBOp->getOpcode() == BO_LAnd) {
RegionCounter Cnt = getPGORegionCounter(CondBOp);
// If we have "1 && X", simplify the code. "0 && X" would have constant
// folded if the case was simple enough.
bool ConstantBool = false;
if (ConstantFoldsToSimpleInteger(CondBOp->getLHS(), ConstantBool) &&
ConstantBool) {
// br(1 && X) -> br(X).
Cnt.beginRegion(Builder);
return EmitBranchOnBoolExpr(CondBOp->getRHS(), TrueBlock, FalseBlock,
TrueCount);
}
// If we have "X && 1", simplify the code to use an uncond branch.
// "X && 0" would have been constant folded to 0.
if (ConstantFoldsToSimpleInteger(CondBOp->getRHS(), ConstantBool) &&
ConstantBool) {
// br(X && 1) -> br(X).
return EmitBranchOnBoolExpr(CondBOp->getLHS(), TrueBlock, FalseBlock,
TrueCount);
}
// Emit the LHS as a conditional. If the LHS conditional is false, we
// want to jump to the FalseBlock.
llvm::BasicBlock *LHSTrue = createBasicBlock("land.lhs.true");
// The counter tells us how often we evaluate RHS, and all of TrueCount
// can be propagated to that branch.
uint64_t RHSCount = Cnt.getCount();
ConditionalEvaluation eval(*this);
EmitBranchOnBoolExpr(CondBOp->getLHS(), LHSTrue, FalseBlock, RHSCount);
EmitBlock(LHSTrue);
// Any temporaries created here are conditional.
Cnt.beginRegion(Builder);
eval.begin(*this);
EmitBranchOnBoolExpr(CondBOp->getRHS(), TrueBlock, FalseBlock, TrueCount);
eval.end(*this);
return;
}
if (CondBOp->getOpcode() == BO_LOr) {
RegionCounter Cnt = getPGORegionCounter(CondBOp);
// If we have "0 || X", simplify the code. "1 || X" would have constant
// folded if the case was simple enough.
bool ConstantBool = false;
if (ConstantFoldsToSimpleInteger(CondBOp->getLHS(), ConstantBool) &&
!ConstantBool) {
// br(0 || X) -> br(X).
Cnt.beginRegion(Builder);
return EmitBranchOnBoolExpr(CondBOp->getRHS(), TrueBlock, FalseBlock,
TrueCount);
}
// If we have "X || 0", simplify the code to use an uncond branch.
// "X || 1" would have been constant folded to 1.
if (ConstantFoldsToSimpleInteger(CondBOp->getRHS(), ConstantBool) &&