Bug Summary

File:lib/Transforms/Utils/InlineFunction.cpp
Warning:line 563, column 11
Forming reference to null pointer

Annotated Source Code

Press '?' to see keyboard shortcuts

clang -cc1 -triple x86_64-pc-linux-gnu -analyze -disable-free -disable-llvm-verifier -discard-value-names -main-file-name InlineFunction.cpp -analyzer-store=region -analyzer-opt-analyze-nested-blocks -analyzer-checker=core -analyzer-checker=apiModeling -analyzer-checker=unix -analyzer-checker=deadcode -analyzer-checker=cplusplus -analyzer-checker=security.insecureAPI.UncheckedReturn -analyzer-checker=security.insecureAPI.getpw -analyzer-checker=security.insecureAPI.gets -analyzer-checker=security.insecureAPI.mktemp -analyzer-checker=security.insecureAPI.mkstemp -analyzer-checker=security.insecureAPI.vfork -analyzer-checker=nullability.NullPassedToNonnull -analyzer-checker=nullability.NullReturnedFromNonnull -analyzer-output plist -w -mrelocation-model pic -pic-level 2 -mthread-model posix -fmath-errno -masm-verbose -mconstructor-aliases -munwind-tables -fuse-init-array -target-cpu x86-64 -dwarf-column-info -debugger-tuning=gdb -momit-leaf-frame-pointer -ffunction-sections -fdata-sections -resource-dir /usr/lib/llvm-8/lib/clang/8.0.0 -D _DEBUG -D _GNU_SOURCE -D __STDC_CONSTANT_MACROS -D __STDC_FORMAT_MACROS -D __STDC_LIMIT_MACROS -I /build/llvm-toolchain-snapshot-8~svn345461/build-llvm/lib/Transforms/Utils -I /build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils -I /build/llvm-toolchain-snapshot-8~svn345461/build-llvm/include -I /build/llvm-toolchain-snapshot-8~svn345461/include -U NDEBUG -internal-isystem /usr/lib/gcc/x86_64-linux-gnu/6.3.0/../../../../include/c++/6.3.0 -internal-isystem /usr/lib/gcc/x86_64-linux-gnu/6.3.0/../../../../include/x86_64-linux-gnu/c++/6.3.0 -internal-isystem /usr/lib/gcc/x86_64-linux-gnu/6.3.0/../../../../include/x86_64-linux-gnu/c++/6.3.0 -internal-isystem /usr/lib/gcc/x86_64-linux-gnu/6.3.0/../../../../include/c++/6.3.0/backward -internal-isystem /usr/include/clang/8.0.0/include/ -internal-isystem /usr/local/include -internal-isystem /usr/lib/llvm-8/lib/clang/8.0.0/include -internal-externc-isystem /usr/include/x86_64-linux-gnu -internal-externc-isystem /include -internal-externc-isystem /usr/include -O2 -Wno-unused-parameter -Wwrite-strings -Wno-missing-field-initializers -Wno-long-long -Wno-maybe-uninitialized -Wno-comment -std=c++11 -fdeprecated-macro -fdebug-compilation-dir /build/llvm-toolchain-snapshot-8~svn345461/build-llvm/lib/Transforms/Utils -ferror-limit 19 -fmessage-length 0 -fvisibility-inlines-hidden -fobjc-runtime=gcc -fdiagnostics-show-option -vectorize-loops -vectorize-slp -analyzer-output=html -analyzer-config stable-report-filename=true -o /tmp/scan-build-2018-10-27-211344-32123-1 -x c++ /build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp -faddrsig
1//===- InlineFunction.cpp - Code to perform function inlining -------------===//
2//
3// The LLVM Compiler Infrastructure
4//
5// This file is distributed under the University of Illinois Open Source
6// License. See LICENSE.TXT for details.
7//
8//===----------------------------------------------------------------------===//
9//
10// This file implements inlining of a function into a call site, resolving
11// parameters and the return value as appropriate.
12//
13//===----------------------------------------------------------------------===//
14
15#include "llvm/ADT/DenseMap.h"
16#include "llvm/ADT/None.h"
17#include "llvm/ADT/Optional.h"
18#include "llvm/ADT/STLExtras.h"
19#include "llvm/ADT/SetVector.h"
20#include "llvm/ADT/SmallPtrSet.h"
21#include "llvm/ADT/SmallVector.h"
22#include "llvm/ADT/StringExtras.h"
23#include "llvm/ADT/iterator_range.h"
24#include "llvm/Analysis/AliasAnalysis.h"
25#include "llvm/Analysis/AssumptionCache.h"
26#include "llvm/Analysis/BlockFrequencyInfo.h"
27#include "llvm/Analysis/CallGraph.h"
28#include "llvm/Analysis/CaptureTracking.h"
29#include "llvm/Analysis/EHPersonalities.h"
30#include "llvm/Analysis/InstructionSimplify.h"
31#include "llvm/Analysis/ProfileSummaryInfo.h"
32#include "llvm/Transforms/Utils/Local.h"
33#include "llvm/Analysis/ValueTracking.h"
34#include "llvm/IR/Argument.h"
35#include "llvm/IR/BasicBlock.h"
36#include "llvm/IR/CFG.h"
37#include "llvm/IR/CallSite.h"
38#include "llvm/IR/Constant.h"
39#include "llvm/IR/Constants.h"
40#include "llvm/IR/DIBuilder.h"
41#include "llvm/IR/DataLayout.h"
42#include "llvm/IR/DebugInfoMetadata.h"
43#include "llvm/IR/DebugLoc.h"
44#include "llvm/IR/DerivedTypes.h"
45#include "llvm/IR/Dominators.h"
46#include "llvm/IR/Function.h"
47#include "llvm/IR/IRBuilder.h"
48#include "llvm/IR/InstrTypes.h"
49#include "llvm/IR/Instruction.h"
50#include "llvm/IR/Instructions.h"
51#include "llvm/IR/IntrinsicInst.h"
52#include "llvm/IR/Intrinsics.h"
53#include "llvm/IR/LLVMContext.h"
54#include "llvm/IR/MDBuilder.h"
55#include "llvm/IR/Metadata.h"
56#include "llvm/IR/Module.h"
57#include "llvm/IR/Type.h"
58#include "llvm/IR/User.h"
59#include "llvm/IR/Value.h"
60#include "llvm/Support/Casting.h"
61#include "llvm/Support/CommandLine.h"
62#include "llvm/Support/ErrorHandling.h"
63#include "llvm/Transforms/Utils/Cloning.h"
64#include "llvm/Transforms/Utils/ValueMapper.h"
65#include <algorithm>
66#include <cassert>
67#include <cstdint>
68#include <iterator>
69#include <limits>
70#include <string>
71#include <utility>
72#include <vector>
73
74using namespace llvm;
75using ProfileCount = Function::ProfileCount;
76
77static cl::opt<bool>
78EnableNoAliasConversion("enable-noalias-to-md-conversion", cl::init(true),
79 cl::Hidden,
80 cl::desc("Convert noalias attributes to metadata during inlining."));
81
82static cl::opt<bool>
83PreserveAlignmentAssumptions("preserve-alignment-assumptions-during-inlining",
84 cl::init(true), cl::Hidden,
85 cl::desc("Convert align attributes to assumptions during inlining."));
86
87llvm::InlineResult llvm::InlineFunction(CallInst *CI, InlineFunctionInfo &IFI,
88 AAResults *CalleeAAR,
89 bool InsertLifetime) {
90 return InlineFunction(CallSite(CI), IFI, CalleeAAR, InsertLifetime);
91}
92
93llvm::InlineResult llvm::InlineFunction(InvokeInst *II, InlineFunctionInfo &IFI,
94 AAResults *CalleeAAR,
95 bool InsertLifetime) {
96 return InlineFunction(CallSite(II), IFI, CalleeAAR, InsertLifetime);
97}
98
99namespace {
100
101 /// A class for recording information about inlining a landing pad.
102 class LandingPadInliningInfo {
103 /// Destination of the invoke's unwind.
104 BasicBlock *OuterResumeDest;
105
106 /// Destination for the callee's resume.
107 BasicBlock *InnerResumeDest = nullptr;
108
109 /// LandingPadInst associated with the invoke.
110 LandingPadInst *CallerLPad = nullptr;
111
112 /// PHI for EH values from landingpad insts.
113 PHINode *InnerEHValuesPHI = nullptr;
114
115 SmallVector<Value*, 8> UnwindDestPHIValues;
116
117 public:
118 LandingPadInliningInfo(InvokeInst *II)
119 : OuterResumeDest(II->getUnwindDest()) {
120 // If there are PHI nodes in the unwind destination block, we need to keep
121 // track of which values came into them from the invoke before removing
122 // the edge from this block.
123 BasicBlock *InvokeBB = II->getParent();
124 BasicBlock::iterator I = OuterResumeDest->begin();
125 for (; isa<PHINode>(I); ++I) {
126 // Save the value to use for this edge.
127 PHINode *PHI = cast<PHINode>(I);
128 UnwindDestPHIValues.push_back(PHI->getIncomingValueForBlock(InvokeBB));
129 }
130
131 CallerLPad = cast<LandingPadInst>(I);
132 }
133
134 /// The outer unwind destination is the target of
135 /// unwind edges introduced for calls within the inlined function.
136 BasicBlock *getOuterResumeDest() const {
137 return OuterResumeDest;
138 }
139
140 BasicBlock *getInnerResumeDest();
141
142 LandingPadInst *getLandingPadInst() const { return CallerLPad; }
143
144 /// Forward the 'resume' instruction to the caller's landing pad block.
145 /// When the landing pad block has only one predecessor, this is
146 /// a simple branch. When there is more than one predecessor, we need to
147 /// split the landing pad block after the landingpad instruction and jump
148 /// to there.
149 void forwardResume(ResumeInst *RI,
150 SmallPtrSetImpl<LandingPadInst*> &InlinedLPads);
151
152 /// Add incoming-PHI values to the unwind destination block for the given
153 /// basic block, using the values for the original invoke's source block.
154 void addIncomingPHIValuesFor(BasicBlock *BB) const {
155 addIncomingPHIValuesForInto(BB, OuterResumeDest);
156 }
157
158 void addIncomingPHIValuesForInto(BasicBlock *src, BasicBlock *dest) const {
159 BasicBlock::iterator I = dest->begin();
160 for (unsigned i = 0, e = UnwindDestPHIValues.size(); i != e; ++i, ++I) {
161 PHINode *phi = cast<PHINode>(I);
162 phi->addIncoming(UnwindDestPHIValues[i], src);
163 }
164 }
165 };
166
167} // end anonymous namespace
168
169/// Get or create a target for the branch from ResumeInsts.
170BasicBlock *LandingPadInliningInfo::getInnerResumeDest() {
171 if (InnerResumeDest) return InnerResumeDest;
172
173 // Split the landing pad.
174 BasicBlock::iterator SplitPoint = ++CallerLPad->getIterator();
175 InnerResumeDest =
176 OuterResumeDest->splitBasicBlock(SplitPoint,
177 OuterResumeDest->getName() + ".body");
178
179 // The number of incoming edges we expect to the inner landing pad.
180 const unsigned PHICapacity = 2;
181
182 // Create corresponding new PHIs for all the PHIs in the outer landing pad.
183 Instruction *InsertPoint = &InnerResumeDest->front();
184 BasicBlock::iterator I = OuterResumeDest->begin();
185 for (unsigned i = 0, e = UnwindDestPHIValues.size(); i != e; ++i, ++I) {
186 PHINode *OuterPHI = cast<PHINode>(I);
187 PHINode *InnerPHI = PHINode::Create(OuterPHI->getType(), PHICapacity,
188 OuterPHI->getName() + ".lpad-body",
189 InsertPoint);
190 OuterPHI->replaceAllUsesWith(InnerPHI);
191 InnerPHI->addIncoming(OuterPHI, OuterResumeDest);
192 }
193
194 // Create a PHI for the exception values.
195 InnerEHValuesPHI = PHINode::Create(CallerLPad->getType(), PHICapacity,
196 "eh.lpad-body", InsertPoint);
197 CallerLPad->replaceAllUsesWith(InnerEHValuesPHI);
198 InnerEHValuesPHI->addIncoming(CallerLPad, OuterResumeDest);
199
200 // All done.
201 return InnerResumeDest;
202}
203
204/// Forward the 'resume' instruction to the caller's landing pad block.
205/// When the landing pad block has only one predecessor, this is a simple
206/// branch. When there is more than one predecessor, we need to split the
207/// landing pad block after the landingpad instruction and jump to there.
208void LandingPadInliningInfo::forwardResume(
209 ResumeInst *RI, SmallPtrSetImpl<LandingPadInst *> &InlinedLPads) {
210 BasicBlock *Dest = getInnerResumeDest();
211 BasicBlock *Src = RI->getParent();
212
213 BranchInst::Create(Dest, Src);
214
215 // Update the PHIs in the destination. They were inserted in an order which
216 // makes this work.
217 addIncomingPHIValuesForInto(Src, Dest);
218
219 InnerEHValuesPHI->addIncoming(RI->getOperand(0), Src);
220 RI->eraseFromParent();
221}
222
223/// Helper for getUnwindDestToken/getUnwindDestTokenHelper.
224static Value *getParentPad(Value *EHPad) {
225 if (auto *FPI = dyn_cast<FuncletPadInst>(EHPad))
226 return FPI->getParentPad();
227 return cast<CatchSwitchInst>(EHPad)->getParentPad();
228}
229
230using UnwindDestMemoTy = DenseMap<Instruction *, Value *>;
231
232/// Helper for getUnwindDestToken that does the descendant-ward part of
233/// the search.
234static Value *getUnwindDestTokenHelper(Instruction *EHPad,
235 UnwindDestMemoTy &MemoMap) {
236 SmallVector<Instruction *, 8> Worklist(1, EHPad);
237
238 while (!Worklist.empty()) {
239 Instruction *CurrentPad = Worklist.pop_back_val();
240 // We only put pads on the worklist that aren't in the MemoMap. When
241 // we find an unwind dest for a pad we may update its ancestors, but
242 // the queue only ever contains uncles/great-uncles/etc. of CurrentPad,
243 // so they should never get updated while queued on the worklist.
244 assert(!MemoMap.count(CurrentPad))((!MemoMap.count(CurrentPad)) ? static_cast<void> (0) :
__assert_fail ("!MemoMap.count(CurrentPad)", "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 244, __PRETTY_FUNCTION__))
;
245 Value *UnwindDestToken = nullptr;
246 if (auto *CatchSwitch = dyn_cast<CatchSwitchInst>(CurrentPad)) {
247 if (CatchSwitch->hasUnwindDest()) {
248 UnwindDestToken = CatchSwitch->getUnwindDest()->getFirstNonPHI();
249 } else {
250 // Catchswitch doesn't have a 'nounwind' variant, and one might be
251 // annotated as "unwinds to caller" when really it's nounwind (see
252 // e.g. SimplifyCFGOpt::SimplifyUnreachable), so we can't infer the
253 // parent's unwind dest from this. We can check its catchpads'
254 // descendants, since they might include a cleanuppad with an
255 // "unwinds to caller" cleanupret, which can be trusted.
256 for (auto HI = CatchSwitch->handler_begin(),
257 HE = CatchSwitch->handler_end();
258 HI != HE && !UnwindDestToken; ++HI) {
259 BasicBlock *HandlerBlock = *HI;
260 auto *CatchPad = cast<CatchPadInst>(HandlerBlock->getFirstNonPHI());
261 for (User *Child : CatchPad->users()) {
262 // Intentionally ignore invokes here -- since the catchswitch is
263 // marked "unwind to caller", it would be a verifier error if it
264 // contained an invoke which unwinds out of it, so any invoke we'd
265 // encounter must unwind to some child of the catch.
266 if (!isa<CleanupPadInst>(Child) && !isa<CatchSwitchInst>(Child))
267 continue;
268
269 Instruction *ChildPad = cast<Instruction>(Child);
270 auto Memo = MemoMap.find(ChildPad);
271 if (Memo == MemoMap.end()) {
272 // Haven't figured out this child pad yet; queue it.
273 Worklist.push_back(ChildPad);
274 continue;
275 }
276 // We've already checked this child, but might have found that
277 // it offers no proof either way.
278 Value *ChildUnwindDestToken = Memo->second;
279 if (!ChildUnwindDestToken)
280 continue;
281 // We already know the child's unwind dest, which can either
282 // be ConstantTokenNone to indicate unwind to caller, or can
283 // be another child of the catchpad. Only the former indicates
284 // the unwind dest of the catchswitch.
285 if (isa<ConstantTokenNone>(ChildUnwindDestToken)) {
286 UnwindDestToken = ChildUnwindDestToken;
287 break;
288 }
289 assert(getParentPad(ChildUnwindDestToken) == CatchPad)((getParentPad(ChildUnwindDestToken) == CatchPad) ? static_cast
<void> (0) : __assert_fail ("getParentPad(ChildUnwindDestToken) == CatchPad"
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 289, __PRETTY_FUNCTION__))
;
290 }
291 }
292 }
293 } else {
294 auto *CleanupPad = cast<CleanupPadInst>(CurrentPad);
295 for (User *U : CleanupPad->users()) {
296 if (auto *CleanupRet = dyn_cast<CleanupReturnInst>(U)) {
297 if (BasicBlock *RetUnwindDest = CleanupRet->getUnwindDest())
298 UnwindDestToken = RetUnwindDest->getFirstNonPHI();
299 else
300 UnwindDestToken = ConstantTokenNone::get(CleanupPad->getContext());
301 break;
302 }
303 Value *ChildUnwindDestToken;
304 if (auto *Invoke = dyn_cast<InvokeInst>(U)) {
305 ChildUnwindDestToken = Invoke->getUnwindDest()->getFirstNonPHI();
306 } else if (isa<CleanupPadInst>(U) || isa<CatchSwitchInst>(U)) {
307 Instruction *ChildPad = cast<Instruction>(U);
308 auto Memo = MemoMap.find(ChildPad);
309 if (Memo == MemoMap.end()) {
310 // Haven't resolved this child yet; queue it and keep searching.
311 Worklist.push_back(ChildPad);
312 continue;
313 }
314 // We've checked this child, but still need to ignore it if it
315 // had no proof either way.
316 ChildUnwindDestToken = Memo->second;
317 if (!ChildUnwindDestToken)
318 continue;
319 } else {
320 // Not a relevant user of the cleanuppad
321 continue;
322 }
323 // In a well-formed program, the child/invoke must either unwind to
324 // an(other) child of the cleanup, or exit the cleanup. In the
325 // first case, continue searching.
326 if (isa<Instruction>(ChildUnwindDestToken) &&
327 getParentPad(ChildUnwindDestToken) == CleanupPad)
328 continue;
329 UnwindDestToken = ChildUnwindDestToken;
330 break;
331 }
332 }
333 // If we haven't found an unwind dest for CurrentPad, we may have queued its
334 // children, so move on to the next in the worklist.
335 if (!UnwindDestToken)
336 continue;
337
338 // Now we know that CurrentPad unwinds to UnwindDestToken. It also exits
339 // any ancestors of CurrentPad up to but not including UnwindDestToken's
340 // parent pad. Record this in the memo map, and check to see if the
341 // original EHPad being queried is one of the ones exited.
342 Value *UnwindParent;
343 if (auto *UnwindPad = dyn_cast<Instruction>(UnwindDestToken))
344 UnwindParent = getParentPad(UnwindPad);
345 else
346 UnwindParent = nullptr;
347 bool ExitedOriginalPad = false;
348 for (Instruction *ExitedPad = CurrentPad;
349 ExitedPad && ExitedPad != UnwindParent;
350 ExitedPad = dyn_cast<Instruction>(getParentPad(ExitedPad))) {
351 // Skip over catchpads since they just follow their catchswitches.
352 if (isa<CatchPadInst>(ExitedPad))
353 continue;
354 MemoMap[ExitedPad] = UnwindDestToken;
355 ExitedOriginalPad |= (ExitedPad == EHPad);
356 }
357
358 if (ExitedOriginalPad)
359 return UnwindDestToken;
360
361 // Continue the search.
362 }
363
364 // No definitive information is contained within this funclet.
365 return nullptr;
366}
367
368/// Given an EH pad, find where it unwinds. If it unwinds to an EH pad,
369/// return that pad instruction. If it unwinds to caller, return
370/// ConstantTokenNone. If it does not have a definitive unwind destination,
371/// return nullptr.
372///
373/// This routine gets invoked for calls in funclets in inlinees when inlining
374/// an invoke. Since many funclets don't have calls inside them, it's queried
375/// on-demand rather than building a map of pads to unwind dests up front.
376/// Determining a funclet's unwind dest may require recursively searching its
377/// descendants, and also ancestors and cousins if the descendants don't provide
378/// an answer. Since most funclets will have their unwind dest immediately
379/// available as the unwind dest of a catchswitch or cleanupret, this routine
380/// searches top-down from the given pad and then up. To avoid worst-case
381/// quadratic run-time given that approach, it uses a memo map to avoid
382/// re-processing funclet trees. The callers that rewrite the IR as they go
383/// take advantage of this, for correctness, by checking/forcing rewritten
384/// pads' entries to match the original callee view.
385static Value *getUnwindDestToken(Instruction *EHPad,
386 UnwindDestMemoTy &MemoMap) {
387 // Catchpads unwind to the same place as their catchswitch;
388 // redirct any queries on catchpads so the code below can
389 // deal with just catchswitches and cleanuppads.
390 if (auto *CPI = dyn_cast<CatchPadInst>(EHPad))
391 EHPad = CPI->getCatchSwitch();
392
393 // Check if we've already determined the unwind dest for this pad.
394 auto Memo = MemoMap.find(EHPad);
395 if (Memo != MemoMap.end())
396 return Memo->second;
397
398 // Search EHPad and, if necessary, its descendants.
399 Value *UnwindDestToken = getUnwindDestTokenHelper(EHPad, MemoMap);
400 assert((UnwindDestToken == nullptr) != (MemoMap.count(EHPad) != 0))(((UnwindDestToken == nullptr) != (MemoMap.count(EHPad) != 0)
) ? static_cast<void> (0) : __assert_fail ("(UnwindDestToken == nullptr) != (MemoMap.count(EHPad) != 0)"
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 400, __PRETTY_FUNCTION__))
;
401 if (UnwindDestToken)
402 return UnwindDestToken;
403
404 // No information is available for this EHPad from itself or any of its
405 // descendants. An unwind all the way out to a pad in the caller would
406 // need also to agree with the unwind dest of the parent funclet, so
407 // search up the chain to try to find a funclet with information. Put
408 // null entries in the memo map to avoid re-processing as we go up.
409 MemoMap[EHPad] = nullptr;
410#ifndef NDEBUG
411 SmallPtrSet<Instruction *, 4> TempMemos;
412 TempMemos.insert(EHPad);
413#endif
414 Instruction *LastUselessPad = EHPad;
415 Value *AncestorToken;
416 for (AncestorToken = getParentPad(EHPad);
417 auto *AncestorPad = dyn_cast<Instruction>(AncestorToken);
418 AncestorToken = getParentPad(AncestorToken)) {
419 // Skip over catchpads since they just follow their catchswitches.
420 if (isa<CatchPadInst>(AncestorPad))
421 continue;
422 // If the MemoMap had an entry mapping AncestorPad to nullptr, since we
423 // haven't yet called getUnwindDestTokenHelper for AncestorPad in this
424 // call to getUnwindDestToken, that would mean that AncestorPad had no
425 // information in itself, its descendants, or its ancestors. If that
426 // were the case, then we should also have recorded the lack of information
427 // for the descendant that we're coming from. So assert that we don't
428 // find a null entry in the MemoMap for AncestorPad.
429 assert(!MemoMap.count(AncestorPad) || MemoMap[AncestorPad])((!MemoMap.count(AncestorPad) || MemoMap[AncestorPad]) ? static_cast
<void> (0) : __assert_fail ("!MemoMap.count(AncestorPad) || MemoMap[AncestorPad]"
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 429, __PRETTY_FUNCTION__))
;
430 auto AncestorMemo = MemoMap.find(AncestorPad);
431 if (AncestorMemo == MemoMap.end()) {
432 UnwindDestToken = getUnwindDestTokenHelper(AncestorPad, MemoMap);
433 } else {
434 UnwindDestToken = AncestorMemo->second;
435 }
436 if (UnwindDestToken)
437 break;
438 LastUselessPad = AncestorPad;
439 MemoMap[LastUselessPad] = nullptr;
440#ifndef NDEBUG
441 TempMemos.insert(LastUselessPad);
442#endif
443 }
444
445 // We know that getUnwindDestTokenHelper was called on LastUselessPad and
446 // returned nullptr (and likewise for EHPad and any of its ancestors up to
447 // LastUselessPad), so LastUselessPad has no information from below. Since
448 // getUnwindDestTokenHelper must investigate all downward paths through
449 // no-information nodes to prove that a node has no information like this,
450 // and since any time it finds information it records it in the MemoMap for
451 // not just the immediately-containing funclet but also any ancestors also
452 // exited, it must be the case that, walking downward from LastUselessPad,
453 // visiting just those nodes which have not been mapped to an unwind dest
454 // by getUnwindDestTokenHelper (the nullptr TempMemos notwithstanding, since
455 // they are just used to keep getUnwindDestTokenHelper from repeating work),
456 // any node visited must have been exhaustively searched with no information
457 // for it found.
458 SmallVector<Instruction *, 8> Worklist(1, LastUselessPad);
459 while (!Worklist.empty()) {
460 Instruction *UselessPad = Worklist.pop_back_val();
461 auto Memo = MemoMap.find(UselessPad);
462 if (Memo != MemoMap.end() && Memo->second) {
463 // Here the name 'UselessPad' is a bit of a misnomer, because we've found
464 // that it is a funclet that does have information about unwinding to
465 // a particular destination; its parent was a useless pad.
466 // Since its parent has no information, the unwind edge must not escape
467 // the parent, and must target a sibling of this pad. This local unwind
468 // gives us no information about EHPad. Leave it and the subtree rooted
469 // at it alone.
470 assert(getParentPad(Memo->second) == getParentPad(UselessPad))((getParentPad(Memo->second) == getParentPad(UselessPad)) ?
static_cast<void> (0) : __assert_fail ("getParentPad(Memo->second) == getParentPad(UselessPad)"
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 470, __PRETTY_FUNCTION__))
;
471 continue;
472 }
473 // We know we don't have information for UselesPad. If it has an entry in
474 // the MemoMap (mapping it to nullptr), it must be one of the TempMemos
475 // added on this invocation of getUnwindDestToken; if a previous invocation
476 // recorded nullptr, it would have had to prove that the ancestors of
477 // UselessPad, which include LastUselessPad, had no information, and that
478 // in turn would have required proving that the descendants of
479 // LastUselesPad, which include EHPad, have no information about
480 // LastUselessPad, which would imply that EHPad was mapped to nullptr in
481 // the MemoMap on that invocation, which isn't the case if we got here.
482 assert(!MemoMap.count(UselessPad) || TempMemos.count(UselessPad))((!MemoMap.count(UselessPad) || TempMemos.count(UselessPad)) ?
static_cast<void> (0) : __assert_fail ("!MemoMap.count(UselessPad) || TempMemos.count(UselessPad)"
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 482, __PRETTY_FUNCTION__))
;
483 // Assert as we enumerate users that 'UselessPad' doesn't have any unwind
484 // information that we'd be contradicting by making a map entry for it
485 // (which is something that getUnwindDestTokenHelper must have proved for
486 // us to get here). Just assert on is direct users here; the checks in
487 // this downward walk at its descendants will verify that they don't have
488 // any unwind edges that exit 'UselessPad' either (i.e. they either have no
489 // unwind edges or unwind to a sibling).
490 MemoMap[UselessPad] = UnwindDestToken;
491 if (auto *CatchSwitch = dyn_cast<CatchSwitchInst>(UselessPad)) {
492 assert(CatchSwitch->getUnwindDest() == nullptr && "Expected useless pad")((CatchSwitch->getUnwindDest() == nullptr && "Expected useless pad"
) ? static_cast<void> (0) : __assert_fail ("CatchSwitch->getUnwindDest() == nullptr && \"Expected useless pad\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 492, __PRETTY_FUNCTION__))
;
493 for (BasicBlock *HandlerBlock : CatchSwitch->handlers()) {
494 auto *CatchPad = HandlerBlock->getFirstNonPHI();
495 for (User *U : CatchPad->users()) {
496 assert((((!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst
>(U)->getUnwindDest()->getFirstNonPHI()) == CatchPad
)) && "Expected useless pad") ? static_cast<void>
(0) : __assert_fail ("(!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst>(U)->getUnwindDest()->getFirstNonPHI()) == CatchPad)) && \"Expected useless pad\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 501, __PRETTY_FUNCTION__))
497 (!isa<InvokeInst>(U) ||(((!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst
>(U)->getUnwindDest()->getFirstNonPHI()) == CatchPad
)) && "Expected useless pad") ? static_cast<void>
(0) : __assert_fail ("(!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst>(U)->getUnwindDest()->getFirstNonPHI()) == CatchPad)) && \"Expected useless pad\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 501, __PRETTY_FUNCTION__))
498 (getParentPad((((!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst
>(U)->getUnwindDest()->getFirstNonPHI()) == CatchPad
)) && "Expected useless pad") ? static_cast<void>
(0) : __assert_fail ("(!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst>(U)->getUnwindDest()->getFirstNonPHI()) == CatchPad)) && \"Expected useless pad\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 501, __PRETTY_FUNCTION__))
499 cast<InvokeInst>(U)->getUnwindDest()->getFirstNonPHI()) ==(((!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst
>(U)->getUnwindDest()->getFirstNonPHI()) == CatchPad
)) && "Expected useless pad") ? static_cast<void>
(0) : __assert_fail ("(!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst>(U)->getUnwindDest()->getFirstNonPHI()) == CatchPad)) && \"Expected useless pad\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 501, __PRETTY_FUNCTION__))
500 CatchPad)) &&(((!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst
>(U)->getUnwindDest()->getFirstNonPHI()) == CatchPad
)) && "Expected useless pad") ? static_cast<void>
(0) : __assert_fail ("(!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst>(U)->getUnwindDest()->getFirstNonPHI()) == CatchPad)) && \"Expected useless pad\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 501, __PRETTY_FUNCTION__))
501 "Expected useless pad")(((!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst
>(U)->getUnwindDest()->getFirstNonPHI()) == CatchPad
)) && "Expected useless pad") ? static_cast<void>
(0) : __assert_fail ("(!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst>(U)->getUnwindDest()->getFirstNonPHI()) == CatchPad)) && \"Expected useless pad\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 501, __PRETTY_FUNCTION__))
;
502 if (isa<CatchSwitchInst>(U) || isa<CleanupPadInst>(U))
503 Worklist.push_back(cast<Instruction>(U));
504 }
505 }
506 } else {
507 assert(isa<CleanupPadInst>(UselessPad))((isa<CleanupPadInst>(UselessPad)) ? static_cast<void
> (0) : __assert_fail ("isa<CleanupPadInst>(UselessPad)"
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 507, __PRETTY_FUNCTION__))
;
508 for (User *U : UselessPad->users()) {
509 assert(!isa<CleanupReturnInst>(U) && "Expected useless pad")((!isa<CleanupReturnInst>(U) && "Expected useless pad"
) ? static_cast<void> (0) : __assert_fail ("!isa<CleanupReturnInst>(U) && \"Expected useless pad\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 509, __PRETTY_FUNCTION__))
;
510 assert((!isa<InvokeInst>(U) ||(((!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst
>(U)->getUnwindDest()->getFirstNonPHI()) == UselessPad
)) && "Expected useless pad") ? static_cast<void>
(0) : __assert_fail ("(!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst>(U)->getUnwindDest()->getFirstNonPHI()) == UselessPad)) && \"Expected useless pad\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 514, __PRETTY_FUNCTION__))
511 (getParentPad((((!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst
>(U)->getUnwindDest()->getFirstNonPHI()) == UselessPad
)) && "Expected useless pad") ? static_cast<void>
(0) : __assert_fail ("(!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst>(U)->getUnwindDest()->getFirstNonPHI()) == UselessPad)) && \"Expected useless pad\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 514, __PRETTY_FUNCTION__))
512 cast<InvokeInst>(U)->getUnwindDest()->getFirstNonPHI()) ==(((!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst
>(U)->getUnwindDest()->getFirstNonPHI()) == UselessPad
)) && "Expected useless pad") ? static_cast<void>
(0) : __assert_fail ("(!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst>(U)->getUnwindDest()->getFirstNonPHI()) == UselessPad)) && \"Expected useless pad\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 514, __PRETTY_FUNCTION__))
513 UselessPad)) &&(((!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst
>(U)->getUnwindDest()->getFirstNonPHI()) == UselessPad
)) && "Expected useless pad") ? static_cast<void>
(0) : __assert_fail ("(!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst>(U)->getUnwindDest()->getFirstNonPHI()) == UselessPad)) && \"Expected useless pad\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 514, __PRETTY_FUNCTION__))
514 "Expected useless pad")(((!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst
>(U)->getUnwindDest()->getFirstNonPHI()) == UselessPad
)) && "Expected useless pad") ? static_cast<void>
(0) : __assert_fail ("(!isa<InvokeInst>(U) || (getParentPad( cast<InvokeInst>(U)->getUnwindDest()->getFirstNonPHI()) == UselessPad)) && \"Expected useless pad\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 514, __PRETTY_FUNCTION__))
;
515 if (isa<CatchSwitchInst>(U) || isa<CleanupPadInst>(U))
516 Worklist.push_back(cast<Instruction>(U));
517 }
518 }
519 }
520
521 return UnwindDestToken;
522}
523
524/// When we inline a basic block into an invoke,
525/// we have to turn all of the calls that can throw into invokes.
526/// This function analyze BB to see if there are any calls, and if so,
527/// it rewrites them to be invokes that jump to InvokeDest and fills in the PHI
528/// nodes in that block with the values specified in InvokeDestPHIValues.
529static BasicBlock *HandleCallsInBlockInlinedThroughInvoke(
530 BasicBlock *BB, BasicBlock *UnwindEdge,
531 UnwindDestMemoTy *FuncletUnwindMap = nullptr) {
532 for (BasicBlock::iterator BBI = BB->begin(), E = BB->end(); BBI != E; ) {
7
Loop condition is true. Entering loop body
533 Instruction *I = &*BBI++;
534
535 // We only need to check for function calls: inlined invoke
536 // instructions require no special handling.
537 CallInst *CI = dyn_cast<CallInst>(I);
538
539 if (!CI || CI->doesNotThrow() || isa<InlineAsm>(CI->getCalledValue()))
8
Taking false branch
540 continue;
541
542 // We do not need to (and in fact, cannot) convert possibly throwing calls
543 // to @llvm.experimental_deoptimize (resp. @llvm.experimental.guard) into
544 // invokes. The caller's "segment" of the deoptimization continuation
545 // attached to the newly inlined @llvm.experimental_deoptimize
546 // (resp. @llvm.experimental.guard) call should contain the exception
547 // handling logic, if any.
548 if (auto *F = CI->getCalledFunction())
9
Taking false branch
549 if (F->getIntrinsicID() == Intrinsic::experimental_deoptimize ||
550 F->getIntrinsicID() == Intrinsic::experimental_guard)
551 continue;
552
553 if (auto FuncletBundle = CI->getOperandBundle(LLVMContext::OB_funclet)) {
10
Assuming the condition is true
11
Taking true branch
554 // This call is nested inside a funclet. If that funclet has an unwind
555 // destination within the inlinee, then unwinding out of this call would
556 // be UB. Rewriting this call to an invoke which targets the inlined
557 // invoke's unwind dest would give the call's parent funclet multiple
558 // unwind destinations, which is something that subsequent EH table
559 // generation can't handle and that the veirifer rejects. So when we
560 // see such a call, leave it as a call.
561 auto *FuncletPad = cast<Instruction>(FuncletBundle->Inputs[0]);
562 Value *UnwindDestToken =
563 getUnwindDestToken(FuncletPad, *FuncletUnwindMap);
12
Forming reference to null pointer
564 if (UnwindDestToken && !isa<ConstantTokenNone>(UnwindDestToken))
565 continue;
566#ifndef NDEBUG
567 Instruction *MemoKey;
568 if (auto *CatchPad = dyn_cast<CatchPadInst>(FuncletPad))
569 MemoKey = CatchPad->getCatchSwitch();
570 else
571 MemoKey = FuncletPad;
572 assert(FuncletUnwindMap->count(MemoKey) &&((FuncletUnwindMap->count(MemoKey) && (*FuncletUnwindMap
)[MemoKey] == UnwindDestToken && "must get memoized to avoid confusing later searches"
) ? static_cast<void> (0) : __assert_fail ("FuncletUnwindMap->count(MemoKey) && (*FuncletUnwindMap)[MemoKey] == UnwindDestToken && \"must get memoized to avoid confusing later searches\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 574, __PRETTY_FUNCTION__))
573 (*FuncletUnwindMap)[MemoKey] == UnwindDestToken &&((FuncletUnwindMap->count(MemoKey) && (*FuncletUnwindMap
)[MemoKey] == UnwindDestToken && "must get memoized to avoid confusing later searches"
) ? static_cast<void> (0) : __assert_fail ("FuncletUnwindMap->count(MemoKey) && (*FuncletUnwindMap)[MemoKey] == UnwindDestToken && \"must get memoized to avoid confusing later searches\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 574, __PRETTY_FUNCTION__))
574 "must get memoized to avoid confusing later searches")((FuncletUnwindMap->count(MemoKey) && (*FuncletUnwindMap
)[MemoKey] == UnwindDestToken && "must get memoized to avoid confusing later searches"
) ? static_cast<void> (0) : __assert_fail ("FuncletUnwindMap->count(MemoKey) && (*FuncletUnwindMap)[MemoKey] == UnwindDestToken && \"must get memoized to avoid confusing later searches\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 574, __PRETTY_FUNCTION__))
;
575#endif // NDEBUG
576 }
577
578 changeToInvokeAndSplitBasicBlock(CI, UnwindEdge);
579 return BB;
580 }
581 return nullptr;
582}
583
584/// If we inlined an invoke site, we need to convert calls
585/// in the body of the inlined function into invokes.
586///
587/// II is the invoke instruction being inlined. FirstNewBlock is the first
588/// block of the inlined code (the last block is the end of the function),
589/// and InlineCodeInfo is information about the code that got inlined.
590static void HandleInlinedLandingPad(InvokeInst *II, BasicBlock *FirstNewBlock,
591 ClonedCodeInfo &InlinedCodeInfo) {
592 BasicBlock *InvokeDest = II->getUnwindDest();
593
594 Function *Caller = FirstNewBlock->getParent();
595
596 // The inlined code is currently at the end of the function, scan from the
597 // start of the inlined code to its end, checking for stuff we need to
598 // rewrite.
599 LandingPadInliningInfo Invoke(II);
600
601 // Get all of the inlined landing pad instructions.
602 SmallPtrSet<LandingPadInst*, 16> InlinedLPads;
603 for (Function::iterator I = FirstNewBlock->getIterator(), E = Caller->end();
1
Loop condition is false. Execution continues on line 610
604 I != E; ++I)
605 if (InvokeInst *II = dyn_cast<InvokeInst>(I->getTerminator()))
606 InlinedLPads.insert(II->getLandingPadInst());
607
608 // Append the clauses from the outer landing pad instruction into the inlined
609 // landing pad instructions.
610 LandingPadInst *OuterLPad = Invoke.getLandingPadInst();
611 for (LandingPadInst *InlinedLPad : InlinedLPads) {
612 unsigned OuterNum = OuterLPad->getNumClauses();
613 InlinedLPad->reserveClauses(OuterNum);
614 for (unsigned OuterIdx = 0; OuterIdx != OuterNum; ++OuterIdx)
615 InlinedLPad->addClause(OuterLPad->getClause(OuterIdx));
616 if (OuterLPad->isCleanup())
617 InlinedLPad->setCleanup(true);
618 }
619
620 for (Function::iterator BB = FirstNewBlock->getIterator(), E = Caller->end();
2
Loop condition is true. Entering loop body
621 BB != E; ++BB) {
622 if (InlinedCodeInfo.ContainsCalls)
3
Assuming the condition is true
4
Taking true branch
623 if (BasicBlock *NewBB = HandleCallsInBlockInlinedThroughInvoke(
5
Passing null pointer value via 3rd parameter 'FuncletUnwindMap'
6
Calling 'HandleCallsInBlockInlinedThroughInvoke'
624 &*BB, Invoke.getOuterResumeDest()))
625 // Update any PHI nodes in the exceptional block to indicate that there
626 // is now a new entry in them.
627 Invoke.addIncomingPHIValuesFor(NewBB);
628
629 // Forward any resumes that are remaining here.
630 if (ResumeInst *RI = dyn_cast<ResumeInst>(BB->getTerminator()))
631 Invoke.forwardResume(RI, InlinedLPads);
632 }
633
634 // Now that everything is happy, we have one final detail. The PHI nodes in
635 // the exception destination block still have entries due to the original
636 // invoke instruction. Eliminate these entries (which might even delete the
637 // PHI node) now.
638 InvokeDest->removePredecessor(II->getParent());
639}
640
641/// If we inlined an invoke site, we need to convert calls
642/// in the body of the inlined function into invokes.
643///
644/// II is the invoke instruction being inlined. FirstNewBlock is the first
645/// block of the inlined code (the last block is the end of the function),
646/// and InlineCodeInfo is information about the code that got inlined.
647static void HandleInlinedEHPad(InvokeInst *II, BasicBlock *FirstNewBlock,
648 ClonedCodeInfo &InlinedCodeInfo) {
649 BasicBlock *UnwindDest = II->getUnwindDest();
650 Function *Caller = FirstNewBlock->getParent();
651
652 assert(UnwindDest->getFirstNonPHI()->isEHPad() && "unexpected BasicBlock!")((UnwindDest->getFirstNonPHI()->isEHPad() && "unexpected BasicBlock!"
) ? static_cast<void> (0) : __assert_fail ("UnwindDest->getFirstNonPHI()->isEHPad() && \"unexpected BasicBlock!\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 652, __PRETTY_FUNCTION__))
;
653
654 // If there are PHI nodes in the unwind destination block, we need to keep
655 // track of which values came into them from the invoke before removing the
656 // edge from this block.
657 SmallVector<Value *, 8> UnwindDestPHIValues;
658 BasicBlock *InvokeBB = II->getParent();
659 for (Instruction &I : *UnwindDest) {
660 // Save the value to use for this edge.
661 PHINode *PHI = dyn_cast<PHINode>(&I);
662 if (!PHI)
663 break;
664 UnwindDestPHIValues.push_back(PHI->getIncomingValueForBlock(InvokeBB));
665 }
666
667 // Add incoming-PHI values to the unwind destination block for the given basic
668 // block, using the values for the original invoke's source block.
669 auto UpdatePHINodes = [&](BasicBlock *Src) {
670 BasicBlock::iterator I = UnwindDest->begin();
671 for (Value *V : UnwindDestPHIValues) {
672 PHINode *PHI = cast<PHINode>(I);
673 PHI->addIncoming(V, Src);
674 ++I;
675 }
676 };
677
678 // This connects all the instructions which 'unwind to caller' to the invoke
679 // destination.
680 UnwindDestMemoTy FuncletUnwindMap;
681 for (Function::iterator BB = FirstNewBlock->getIterator(), E = Caller->end();
682 BB != E; ++BB) {
683 if (auto *CRI = dyn_cast<CleanupReturnInst>(BB->getTerminator())) {
684 if (CRI->unwindsToCaller()) {
685 auto *CleanupPad = CRI->getCleanupPad();
686 CleanupReturnInst::Create(CleanupPad, UnwindDest, CRI);
687 CRI->eraseFromParent();
688 UpdatePHINodes(&*BB);
689 // Finding a cleanupret with an unwind destination would confuse
690 // subsequent calls to getUnwindDestToken, so map the cleanuppad
691 // to short-circuit any such calls and recognize this as an "unwind
692 // to caller" cleanup.
693 assert(!FuncletUnwindMap.count(CleanupPad) ||((!FuncletUnwindMap.count(CleanupPad) || isa<ConstantTokenNone
>(FuncletUnwindMap[CleanupPad])) ? static_cast<void>
(0) : __assert_fail ("!FuncletUnwindMap.count(CleanupPad) || isa<ConstantTokenNone>(FuncletUnwindMap[CleanupPad])"
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 694, __PRETTY_FUNCTION__))
694 isa<ConstantTokenNone>(FuncletUnwindMap[CleanupPad]))((!FuncletUnwindMap.count(CleanupPad) || isa<ConstantTokenNone
>(FuncletUnwindMap[CleanupPad])) ? static_cast<void>
(0) : __assert_fail ("!FuncletUnwindMap.count(CleanupPad) || isa<ConstantTokenNone>(FuncletUnwindMap[CleanupPad])"
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 694, __PRETTY_FUNCTION__))
;
695 FuncletUnwindMap[CleanupPad] =
696 ConstantTokenNone::get(Caller->getContext());
697 }
698 }
699
700 Instruction *I = BB->getFirstNonPHI();
701 if (!I->isEHPad())
702 continue;
703
704 Instruction *Replacement = nullptr;
705 if (auto *CatchSwitch = dyn_cast<CatchSwitchInst>(I)) {
706 if (CatchSwitch->unwindsToCaller()) {
707 Value *UnwindDestToken;
708 if (auto *ParentPad =
709 dyn_cast<Instruction>(CatchSwitch->getParentPad())) {
710 // This catchswitch is nested inside another funclet. If that
711 // funclet has an unwind destination within the inlinee, then
712 // unwinding out of this catchswitch would be UB. Rewriting this
713 // catchswitch to unwind to the inlined invoke's unwind dest would
714 // give the parent funclet multiple unwind destinations, which is
715 // something that subsequent EH table generation can't handle and
716 // that the veirifer rejects. So when we see such a call, leave it
717 // as "unwind to caller".
718 UnwindDestToken = getUnwindDestToken(ParentPad, FuncletUnwindMap);
719 if (UnwindDestToken && !isa<ConstantTokenNone>(UnwindDestToken))
720 continue;
721 } else {
722 // This catchswitch has no parent to inherit constraints from, and
723 // none of its descendants can have an unwind edge that exits it and
724 // targets another funclet in the inlinee. It may or may not have a
725 // descendant that definitively has an unwind to caller. In either
726 // case, we'll have to assume that any unwinds out of it may need to
727 // be routed to the caller, so treat it as though it has a definitive
728 // unwind to caller.
729 UnwindDestToken = ConstantTokenNone::get(Caller->getContext());
730 }
731 auto *NewCatchSwitch = CatchSwitchInst::Create(
732 CatchSwitch->getParentPad(), UnwindDest,
733 CatchSwitch->getNumHandlers(), CatchSwitch->getName(),
734 CatchSwitch);
735 for (BasicBlock *PadBB : CatchSwitch->handlers())
736 NewCatchSwitch->addHandler(PadBB);
737 // Propagate info for the old catchswitch over to the new one in
738 // the unwind map. This also serves to short-circuit any subsequent
739 // checks for the unwind dest of this catchswitch, which would get
740 // confused if they found the outer handler in the callee.
741 FuncletUnwindMap[NewCatchSwitch] = UnwindDestToken;
742 Replacement = NewCatchSwitch;
743 }
744 } else if (!isa<FuncletPadInst>(I)) {
745 llvm_unreachable("unexpected EHPad!")::llvm::llvm_unreachable_internal("unexpected EHPad!", "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 745)
;
746 }
747
748 if (Replacement) {
749 Replacement->takeName(I);
750 I->replaceAllUsesWith(Replacement);
751 I->eraseFromParent();
752 UpdatePHINodes(&*BB);
753 }
754 }
755
756 if (InlinedCodeInfo.ContainsCalls)
757 for (Function::iterator BB = FirstNewBlock->getIterator(),
758 E = Caller->end();
759 BB != E; ++BB)
760 if (BasicBlock *NewBB = HandleCallsInBlockInlinedThroughInvoke(
761 &*BB, UnwindDest, &FuncletUnwindMap))
762 // Update any PHI nodes in the exceptional block to indicate that there
763 // is now a new entry in them.
764 UpdatePHINodes(NewBB);
765
766 // Now that everything is happy, we have one final detail. The PHI nodes in
767 // the exception destination block still have entries due to the original
768 // invoke instruction. Eliminate these entries (which might even delete the
769 // PHI node) now.
770 UnwindDest->removePredecessor(InvokeBB);
771}
772
773/// When inlining a call site that has !llvm.mem.parallel_loop_access metadata,
774/// that metadata should be propagated to all memory-accessing cloned
775/// instructions.
776static void PropagateParallelLoopAccessMetadata(CallSite CS,
777 ValueToValueMapTy &VMap) {
778 MDNode *M =
779 CS.getInstruction()->getMetadata(LLVMContext::MD_mem_parallel_loop_access);
780 if (!M)
781 return;
782
783 for (ValueToValueMapTy::iterator VMI = VMap.begin(), VMIE = VMap.end();
784 VMI != VMIE; ++VMI) {
785 if (!VMI->second)
786 continue;
787
788 Instruction *NI = dyn_cast<Instruction>(VMI->second);
789 if (!NI)
790 continue;
791
792 if (MDNode *PM = NI->getMetadata(LLVMContext::MD_mem_parallel_loop_access)) {
793 M = MDNode::concatenate(PM, M);
794 NI->setMetadata(LLVMContext::MD_mem_parallel_loop_access, M);
795 } else if (NI->mayReadOrWriteMemory()) {
796 NI->setMetadata(LLVMContext::MD_mem_parallel_loop_access, M);
797 }
798 }
799}
800
801/// When inlining a function that contains noalias scope metadata,
802/// this metadata needs to be cloned so that the inlined blocks
803/// have different "unique scopes" at every call site. Were this not done, then
804/// aliasing scopes from a function inlined into a caller multiple times could
805/// not be differentiated (and this would lead to miscompiles because the
806/// non-aliasing property communicated by the metadata could have
807/// call-site-specific control dependencies).
808static void CloneAliasScopeMetadata(CallSite CS, ValueToValueMapTy &VMap) {
809 const Function *CalledFunc = CS.getCalledFunction();
810 SetVector<const MDNode *> MD;
811
812 // Note: We could only clone the metadata if it is already used in the
813 // caller. I'm omitting that check here because it might confuse
814 // inter-procedural alias analysis passes. We can revisit this if it becomes
815 // an efficiency or overhead problem.
816
817 for (const BasicBlock &I : *CalledFunc)
818 for (const Instruction &J : I) {
819 if (const MDNode *M = J.getMetadata(LLVMContext::MD_alias_scope))
820 MD.insert(M);
821 if (const MDNode *M = J.getMetadata(LLVMContext::MD_noalias))
822 MD.insert(M);
823 }
824
825 if (MD.empty())
826 return;
827
828 // Walk the existing metadata, adding the complete (perhaps cyclic) chain to
829 // the set.
830 SmallVector<const Metadata *, 16> Queue(MD.begin(), MD.end());
831 while (!Queue.empty()) {
832 const MDNode *M = cast<MDNode>(Queue.pop_back_val());
833 for (unsigned i = 0, ie = M->getNumOperands(); i != ie; ++i)
834 if (const MDNode *M1 = dyn_cast<MDNode>(M->getOperand(i)))
835 if (MD.insert(M1))
836 Queue.push_back(M1);
837 }
838
839 // Now we have a complete set of all metadata in the chains used to specify
840 // the noalias scopes and the lists of those scopes.
841 SmallVector<TempMDTuple, 16> DummyNodes;
842 DenseMap<const MDNode *, TrackingMDNodeRef> MDMap;
843 for (const MDNode *I : MD) {
844 DummyNodes.push_back(MDTuple::getTemporary(CalledFunc->getContext(), None));
845 MDMap[I].reset(DummyNodes.back().get());
846 }
847
848 // Create new metadata nodes to replace the dummy nodes, replacing old
849 // metadata references with either a dummy node or an already-created new
850 // node.
851 for (const MDNode *I : MD) {
852 SmallVector<Metadata *, 4> NewOps;
853 for (unsigned i = 0, ie = I->getNumOperands(); i != ie; ++i) {
854 const Metadata *V = I->getOperand(i);
855 if (const MDNode *M = dyn_cast<MDNode>(V))
856 NewOps.push_back(MDMap[M]);
857 else
858 NewOps.push_back(const_cast<Metadata *>(V));
859 }
860
861 MDNode *NewM = MDNode::get(CalledFunc->getContext(), NewOps);
862 MDTuple *TempM = cast<MDTuple>(MDMap[I]);
863 assert(TempM->isTemporary() && "Expected temporary node")((TempM->isTemporary() && "Expected temporary node"
) ? static_cast<void> (0) : __assert_fail ("TempM->isTemporary() && \"Expected temporary node\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 863, __PRETTY_FUNCTION__))
;
864
865 TempM->replaceAllUsesWith(NewM);
866 }
867
868 // Now replace the metadata in the new inlined instructions with the
869 // repacements from the map.
870 for (ValueToValueMapTy::iterator VMI = VMap.begin(), VMIE = VMap.end();
871 VMI != VMIE; ++VMI) {
872 if (!VMI->second)
873 continue;
874
875 Instruction *NI = dyn_cast<Instruction>(VMI->second);
876 if (!NI)
877 continue;
878
879 if (MDNode *M = NI->getMetadata(LLVMContext::MD_alias_scope)) {
880 MDNode *NewMD = MDMap[M];
881 // If the call site also had alias scope metadata (a list of scopes to
882 // which instructions inside it might belong), propagate those scopes to
883 // the inlined instructions.
884 if (MDNode *CSM =
885 CS.getInstruction()->getMetadata(LLVMContext::MD_alias_scope))
886 NewMD = MDNode::concatenate(NewMD, CSM);
887 NI->setMetadata(LLVMContext::MD_alias_scope, NewMD);
888 } else if (NI->mayReadOrWriteMemory()) {
889 if (MDNode *M =
890 CS.getInstruction()->getMetadata(LLVMContext::MD_alias_scope))
891 NI->setMetadata(LLVMContext::MD_alias_scope, M);
892 }
893
894 if (MDNode *M = NI->getMetadata(LLVMContext::MD_noalias)) {
895 MDNode *NewMD = MDMap[M];
896 // If the call site also had noalias metadata (a list of scopes with
897 // which instructions inside it don't alias), propagate those scopes to
898 // the inlined instructions.
899 if (MDNode *CSM =
900 CS.getInstruction()->getMetadata(LLVMContext::MD_noalias))
901 NewMD = MDNode::concatenate(NewMD, CSM);
902 NI->setMetadata(LLVMContext::MD_noalias, NewMD);
903 } else if (NI->mayReadOrWriteMemory()) {
904 if (MDNode *M = CS.getInstruction()->getMetadata(LLVMContext::MD_noalias))
905 NI->setMetadata(LLVMContext::MD_noalias, M);
906 }
907 }
908}
909
910/// If the inlined function has noalias arguments,
911/// then add new alias scopes for each noalias argument, tag the mapped noalias
912/// parameters with noalias metadata specifying the new scope, and tag all
913/// non-derived loads, stores and memory intrinsics with the new alias scopes.
914static void AddAliasScopeMetadata(CallSite CS, ValueToValueMapTy &VMap,
915 const DataLayout &DL, AAResults *CalleeAAR) {
916 if (!EnableNoAliasConversion)
917 return;
918
919 const Function *CalledFunc = CS.getCalledFunction();
920 SmallVector<const Argument *, 4> NoAliasArgs;
921
922 for (const Argument &Arg : CalledFunc->args())
923 if (Arg.hasNoAliasAttr() && !Arg.use_empty())
924 NoAliasArgs.push_back(&Arg);
925
926 if (NoAliasArgs.empty())
927 return;
928
929 // To do a good job, if a noalias variable is captured, we need to know if
930 // the capture point dominates the particular use we're considering.
931 DominatorTree DT;
932 DT.recalculate(const_cast<Function&>(*CalledFunc));
933
934 // noalias indicates that pointer values based on the argument do not alias
935 // pointer values which are not based on it. So we add a new "scope" for each
936 // noalias function argument. Accesses using pointers based on that argument
937 // become part of that alias scope, accesses using pointers not based on that
938 // argument are tagged as noalias with that scope.
939
940 DenseMap<const Argument *, MDNode *> NewScopes;
941 MDBuilder MDB(CalledFunc->getContext());
942
943 // Create a new scope domain for this function.
944 MDNode *NewDomain =
945 MDB.createAnonymousAliasScopeDomain(CalledFunc->getName());
946 for (unsigned i = 0, e = NoAliasArgs.size(); i != e; ++i) {
947 const Argument *A = NoAliasArgs[i];
948
949 std::string Name = CalledFunc->getName();
950 if (A->hasName()) {
951 Name += ": %";
952 Name += A->getName();
953 } else {
954 Name += ": argument ";
955 Name += utostr(i);
956 }
957
958 // Note: We always create a new anonymous root here. This is true regardless
959 // of the linkage of the callee because the aliasing "scope" is not just a
960 // property of the callee, but also all control dependencies in the caller.
961 MDNode *NewScope = MDB.createAnonymousAliasScope(NewDomain, Name);
962 NewScopes.insert(std::make_pair(A, NewScope));
963 }
964
965 // Iterate over all new instructions in the map; for all memory-access
966 // instructions, add the alias scope metadata.
967 for (ValueToValueMapTy::iterator VMI = VMap.begin(), VMIE = VMap.end();
968 VMI != VMIE; ++VMI) {
969 if (const Instruction *I = dyn_cast<Instruction>(VMI->first)) {
970 if (!VMI->second)
971 continue;
972
973 Instruction *NI = dyn_cast<Instruction>(VMI->second);
974 if (!NI)
975 continue;
976
977 bool IsArgMemOnlyCall = false, IsFuncCall = false;
978 SmallVector<const Value *, 2> PtrArgs;
979
980 if (const LoadInst *LI = dyn_cast<LoadInst>(I))
981 PtrArgs.push_back(LI->getPointerOperand());
982 else if (const StoreInst *SI = dyn_cast<StoreInst>(I))
983 PtrArgs.push_back(SI->getPointerOperand());
984 else if (const VAArgInst *VAAI = dyn_cast<VAArgInst>(I))
985 PtrArgs.push_back(VAAI->getPointerOperand());
986 else if (const AtomicCmpXchgInst *CXI = dyn_cast<AtomicCmpXchgInst>(I))
987 PtrArgs.push_back(CXI->getPointerOperand());
988 else if (const AtomicRMWInst *RMWI = dyn_cast<AtomicRMWInst>(I))
989 PtrArgs.push_back(RMWI->getPointerOperand());
990 else if (ImmutableCallSite ICS = ImmutableCallSite(I)) {
991 // If we know that the call does not access memory, then we'll still
992 // know that about the inlined clone of this call site, and we don't
993 // need to add metadata.
994 if (ICS.doesNotAccessMemory())
995 continue;
996
997 IsFuncCall = true;
998 if (CalleeAAR) {
999 FunctionModRefBehavior MRB = CalleeAAR->getModRefBehavior(ICS);
1000 if (MRB == FMRB_OnlyAccessesArgumentPointees ||
1001 MRB == FMRB_OnlyReadsArgumentPointees)
1002 IsArgMemOnlyCall = true;
1003 }
1004
1005 for (Value *Arg : ICS.args()) {
1006 // We need to check the underlying objects of all arguments, not just
1007 // the pointer arguments, because we might be passing pointers as
1008 // integers, etc.
1009 // However, if we know that the call only accesses pointer arguments,
1010 // then we only need to check the pointer arguments.
1011 if (IsArgMemOnlyCall && !Arg->getType()->isPointerTy())
1012 continue;
1013
1014 PtrArgs.push_back(Arg);
1015 }
1016 }
1017
1018 // If we found no pointers, then this instruction is not suitable for
1019 // pairing with an instruction to receive aliasing metadata.
1020 // However, if this is a call, this we might just alias with none of the
1021 // noalias arguments.
1022 if (PtrArgs.empty() && !IsFuncCall)
1023 continue;
1024
1025 // It is possible that there is only one underlying object, but you
1026 // need to go through several PHIs to see it, and thus could be
1027 // repeated in the Objects list.
1028 SmallPtrSet<const Value *, 4> ObjSet;
1029 SmallVector<Metadata *, 4> Scopes, NoAliases;
1030
1031 SmallSetVector<const Argument *, 4> NAPtrArgs;
1032 for (const Value *V : PtrArgs) {
1033 SmallVector<Value *, 4> Objects;
1034 GetUnderlyingObjects(const_cast<Value*>(V),
1035 Objects, DL, /* LI = */ nullptr);
1036
1037 for (Value *O : Objects)
1038 ObjSet.insert(O);
1039 }
1040
1041 // Figure out if we're derived from anything that is not a noalias
1042 // argument.
1043 bool CanDeriveViaCapture = false, UsesAliasingPtr = false;
1044 for (const Value *V : ObjSet) {
1045 // Is this value a constant that cannot be derived from any pointer
1046 // value (we need to exclude constant expressions, for example, that
1047 // are formed from arithmetic on global symbols).
1048 bool IsNonPtrConst = isa<ConstantInt>(V) || isa<ConstantFP>(V) ||
1049 isa<ConstantPointerNull>(V) ||
1050 isa<ConstantDataVector>(V) || isa<UndefValue>(V);
1051 if (IsNonPtrConst)
1052 continue;
1053
1054 // If this is anything other than a noalias argument, then we cannot
1055 // completely describe the aliasing properties using alias.scope
1056 // metadata (and, thus, won't add any).
1057 if (const Argument *A = dyn_cast<Argument>(V)) {
1058 if (!A->hasNoAliasAttr())
1059 UsesAliasingPtr = true;
1060 } else {
1061 UsesAliasingPtr = true;
1062 }
1063
1064 // If this is not some identified function-local object (which cannot
1065 // directly alias a noalias argument), or some other argument (which,
1066 // by definition, also cannot alias a noalias argument), then we could
1067 // alias a noalias argument that has been captured).
1068 if (!isa<Argument>(V) &&
1069 !isIdentifiedFunctionLocal(const_cast<Value*>(V)))
1070 CanDeriveViaCapture = true;
1071 }
1072
1073 // A function call can always get captured noalias pointers (via other
1074 // parameters, globals, etc.).
1075 if (IsFuncCall && !IsArgMemOnlyCall)
1076 CanDeriveViaCapture = true;
1077
1078 // First, we want to figure out all of the sets with which we definitely
1079 // don't alias. Iterate over all noalias set, and add those for which:
1080 // 1. The noalias argument is not in the set of objects from which we
1081 // definitely derive.
1082 // 2. The noalias argument has not yet been captured.
1083 // An arbitrary function that might load pointers could see captured
1084 // noalias arguments via other noalias arguments or globals, and so we
1085 // must always check for prior capture.
1086 for (const Argument *A : NoAliasArgs) {
1087 if (!ObjSet.count(A) && (!CanDeriveViaCapture ||
1088 // It might be tempting to skip the
1089 // PointerMayBeCapturedBefore check if
1090 // A->hasNoCaptureAttr() is true, but this is
1091 // incorrect because nocapture only guarantees
1092 // that no copies outlive the function, not
1093 // that the value cannot be locally captured.
1094 !PointerMayBeCapturedBefore(A,
1095 /* ReturnCaptures */ false,
1096 /* StoreCaptures */ false, I, &DT)))
1097 NoAliases.push_back(NewScopes[A]);
1098 }
1099
1100 if (!NoAliases.empty())
1101 NI->setMetadata(LLVMContext::MD_noalias,
1102 MDNode::concatenate(
1103 NI->getMetadata(LLVMContext::MD_noalias),
1104 MDNode::get(CalledFunc->getContext(), NoAliases)));
1105
1106 // Next, we want to figure out all of the sets to which we might belong.
1107 // We might belong to a set if the noalias argument is in the set of
1108 // underlying objects. If there is some non-noalias argument in our list
1109 // of underlying objects, then we cannot add a scope because the fact
1110 // that some access does not alias with any set of our noalias arguments
1111 // cannot itself guarantee that it does not alias with this access
1112 // (because there is some pointer of unknown origin involved and the
1113 // other access might also depend on this pointer). We also cannot add
1114 // scopes to arbitrary functions unless we know they don't access any
1115 // non-parameter pointer-values.
1116 bool CanAddScopes = !UsesAliasingPtr;
1117 if (CanAddScopes && IsFuncCall)
1118 CanAddScopes = IsArgMemOnlyCall;
1119
1120 if (CanAddScopes)
1121 for (const Argument *A : NoAliasArgs) {
1122 if (ObjSet.count(A))
1123 Scopes.push_back(NewScopes[A]);
1124 }
1125
1126 if (!Scopes.empty())
1127 NI->setMetadata(
1128 LLVMContext::MD_alias_scope,
1129 MDNode::concatenate(NI->getMetadata(LLVMContext::MD_alias_scope),
1130 MDNode::get(CalledFunc->getContext(), Scopes)));
1131 }
1132 }
1133}
1134
1135/// If the inlined function has non-byval align arguments, then
1136/// add @llvm.assume-based alignment assumptions to preserve this information.
1137static void AddAlignmentAssumptions(CallSite CS, InlineFunctionInfo &IFI) {
1138 if (!PreserveAlignmentAssumptions || !IFI.GetAssumptionCache)
1139 return;
1140
1141 AssumptionCache *AC = &(*IFI.GetAssumptionCache)(*CS.getCaller());
1142 auto &DL = CS.getCaller()->getParent()->getDataLayout();
1143
1144 // To avoid inserting redundant assumptions, we should check for assumptions
1145 // already in the caller. To do this, we might need a DT of the caller.
1146 DominatorTree DT;
1147 bool DTCalculated = false;
1148
1149 Function *CalledFunc = CS.getCalledFunction();
1150 for (Argument &Arg : CalledFunc->args()) {
1151 unsigned Align = Arg.getType()->isPointerTy() ? Arg.getParamAlignment() : 0;
1152 if (Align && !Arg.hasByValOrInAllocaAttr() && !Arg.hasNUses(0)) {
1153 if (!DTCalculated) {
1154 DT.recalculate(*CS.getCaller());
1155 DTCalculated = true;
1156 }
1157
1158 // If we can already prove the asserted alignment in the context of the
1159 // caller, then don't bother inserting the assumption.
1160 Value *ArgVal = CS.getArgument(Arg.getArgNo());
1161 if (getKnownAlignment(ArgVal, DL, CS.getInstruction(), AC, &DT) >= Align)
1162 continue;
1163
1164 CallInst *NewAsmp = IRBuilder<>(CS.getInstruction())
1165 .CreateAlignmentAssumption(DL, ArgVal, Align);
1166 AC->registerAssumption(NewAsmp);
1167 }
1168 }
1169}
1170
1171/// Once we have cloned code over from a callee into the caller,
1172/// update the specified callgraph to reflect the changes we made.
1173/// Note that it's possible that not all code was copied over, so only
1174/// some edges of the callgraph may remain.
1175static void UpdateCallGraphAfterInlining(CallSite CS,
1176 Function::iterator FirstNewBlock,
1177 ValueToValueMapTy &VMap,
1178 InlineFunctionInfo &IFI) {
1179 CallGraph &CG = *IFI.CG;
1180 const Function *Caller = CS.getCaller();
1181 const Function *Callee = CS.getCalledFunction();
1182 CallGraphNode *CalleeNode = CG[Callee];
1183 CallGraphNode *CallerNode = CG[Caller];
1184
1185 // Since we inlined some uninlined call sites in the callee into the caller,
1186 // add edges from the caller to all of the callees of the callee.
1187 CallGraphNode::iterator I = CalleeNode->begin(), E = CalleeNode->end();
1188
1189 // Consider the case where CalleeNode == CallerNode.
1190 CallGraphNode::CalledFunctionsVector CallCache;
1191 if (CalleeNode == CallerNode) {
1192 CallCache.assign(I, E);
1193 I = CallCache.begin();
1194 E = CallCache.end();
1195 }
1196
1197 for (; I != E; ++I) {
1198 const Value *OrigCall = I->first;
1199
1200 ValueToValueMapTy::iterator VMI = VMap.find(OrigCall);
1201 // Only copy the edge if the call was inlined!
1202 if (VMI == VMap.end() || VMI->second == nullptr)
1203 continue;
1204
1205 // If the call was inlined, but then constant folded, there is no edge to
1206 // add. Check for this case.
1207 Instruction *NewCall = dyn_cast<Instruction>(VMI->second);
1208 if (!NewCall)
1209 continue;
1210
1211 // We do not treat intrinsic calls like real function calls because we
1212 // expect them to become inline code; do not add an edge for an intrinsic.
1213 CallSite CS = CallSite(NewCall);
1214 if (CS && CS.getCalledFunction() && CS.getCalledFunction()->isIntrinsic())
1215 continue;
1216
1217 // Remember that this call site got inlined for the client of
1218 // InlineFunction.
1219 IFI.InlinedCalls.push_back(NewCall);
1220
1221 // It's possible that inlining the callsite will cause it to go from an
1222 // indirect to a direct call by resolving a function pointer. If this
1223 // happens, set the callee of the new call site to a more precise
1224 // destination. This can also happen if the call graph node of the caller
1225 // was just unnecessarily imprecise.
1226 if (!I->second->getFunction())
1227 if (Function *F = CallSite(NewCall).getCalledFunction()) {
1228 // Indirect call site resolved to direct call.
1229 CallerNode->addCalledFunction(CallSite(NewCall), CG[F]);
1230
1231 continue;
1232 }
1233
1234 CallerNode->addCalledFunction(CallSite(NewCall), I->second);
1235 }
1236
1237 // Update the call graph by deleting the edge from Callee to Caller. We must
1238 // do this after the loop above in case Caller and Callee are the same.
1239 CallerNode->removeCallEdgeFor(CS);
1240}
1241
1242static void HandleByValArgumentInit(Value *Dst, Value *Src, Module *M,
1243 BasicBlock *InsertBlock,
1244 InlineFunctionInfo &IFI) {
1245 Type *AggTy = cast<PointerType>(Src->getType())->getElementType();
1246 IRBuilder<> Builder(InsertBlock, InsertBlock->begin());
1247
1248 Value *Size = Builder.getInt64(M->getDataLayout().getTypeStoreSize(AggTy));
1249
1250 // Always generate a memcpy of alignment 1 here because we don't know
1251 // the alignment of the src pointer. Other optimizations can infer
1252 // better alignment.
1253 Builder.CreateMemCpy(Dst, /*DstAlign*/1, Src, /*SrcAlign*/1, Size);
1254}
1255
1256/// When inlining a call site that has a byval argument,
1257/// we have to make the implicit memcpy explicit by adding it.
1258static Value *HandleByValArgument(Value *Arg, Instruction *TheCall,
1259 const Function *CalledFunc,
1260 InlineFunctionInfo &IFI,
1261 unsigned ByValAlignment) {
1262 PointerType *ArgTy = cast<PointerType>(Arg->getType());
1263 Type *AggTy = ArgTy->getElementType();
1264
1265 Function *Caller = TheCall->getFunction();
1266 const DataLayout &DL = Caller->getParent()->getDataLayout();
1267
1268 // If the called function is readonly, then it could not mutate the caller's
1269 // copy of the byval'd memory. In this case, it is safe to elide the copy and
1270 // temporary.
1271 if (CalledFunc->onlyReadsMemory()) {
1272 // If the byval argument has a specified alignment that is greater than the
1273 // passed in pointer, then we either have to round up the input pointer or
1274 // give up on this transformation.
1275 if (ByValAlignment <= 1) // 0 = unspecified, 1 = no particular alignment.
1276 return Arg;
1277
1278 AssumptionCache *AC =
1279 IFI.GetAssumptionCache ? &(*IFI.GetAssumptionCache)(*Caller) : nullptr;
1280
1281 // If the pointer is already known to be sufficiently aligned, or if we can
1282 // round it up to a larger alignment, then we don't need a temporary.
1283 if (getOrEnforceKnownAlignment(Arg, ByValAlignment, DL, TheCall, AC) >=
1284 ByValAlignment)
1285 return Arg;
1286
1287 // Otherwise, we have to make a memcpy to get a safe alignment. This is bad
1288 // for code quality, but rarely happens and is required for correctness.
1289 }
1290
1291 // Create the alloca. If we have DataLayout, use nice alignment.
1292 unsigned Align = DL.getPrefTypeAlignment(AggTy);
1293
1294 // If the byval had an alignment specified, we *must* use at least that
1295 // alignment, as it is required by the byval argument (and uses of the
1296 // pointer inside the callee).
1297 Align = std::max(Align, ByValAlignment);
1298
1299 Value *NewAlloca = new AllocaInst(AggTy, DL.getAllocaAddrSpace(),
1300 nullptr, Align, Arg->getName(),
1301 &*Caller->begin()->begin());
1302 IFI.StaticAllocas.push_back(cast<AllocaInst>(NewAlloca));
1303
1304 // Uses of the argument in the function should use our new alloca
1305 // instead.
1306 return NewAlloca;
1307}
1308
1309// Check whether this Value is used by a lifetime intrinsic.
1310static bool isUsedByLifetimeMarker(Value *V) {
1311 for (User *U : V->users()) {
1312 if (IntrinsicInst *II = dyn_cast<IntrinsicInst>(U)) {
1313 switch (II->getIntrinsicID()) {
1314 default: break;
1315 case Intrinsic::lifetime_start:
1316 case Intrinsic::lifetime_end:
1317 return true;
1318 }
1319 }
1320 }
1321 return false;
1322}
1323
1324// Check whether the given alloca already has
1325// lifetime.start or lifetime.end intrinsics.
1326static bool hasLifetimeMarkers(AllocaInst *AI) {
1327 Type *Ty = AI->getType();
1328 Type *Int8PtrTy = Type::getInt8PtrTy(Ty->getContext(),
1329 Ty->getPointerAddressSpace());
1330 if (Ty == Int8PtrTy)
1331 return isUsedByLifetimeMarker(AI);
1332
1333 // Do a scan to find all the casts to i8*.
1334 for (User *U : AI->users()) {
1335 if (U->getType() != Int8PtrTy) continue;
1336 if (U->stripPointerCasts() != AI) continue;
1337 if (isUsedByLifetimeMarker(U))
1338 return true;
1339 }
1340 return false;
1341}
1342
1343/// Return the result of AI->isStaticAlloca() if AI were moved to the entry
1344/// block. Allocas used in inalloca calls and allocas of dynamic array size
1345/// cannot be static.
1346static bool allocaWouldBeStaticInEntry(const AllocaInst *AI ) {
1347 return isa<Constant>(AI->getArraySize()) && !AI->isUsedWithInAlloca();
1348}
1349
1350/// Update inlined instructions' line numbers to
1351/// to encode location where these instructions are inlined.
1352static void fixupLineNumbers(Function *Fn, Function::iterator FI,
1353 Instruction *TheCall, bool CalleeHasDebugInfo) {
1354 const DebugLoc &TheCallDL = TheCall->getDebugLoc();
1355 if (!TheCallDL)
1356 return;
1357
1358 auto &Ctx = Fn->getContext();
1359 DILocation *InlinedAtNode = TheCallDL;
1360
1361 // Create a unique call site, not to be confused with any other call from the
1362 // same location.
1363 InlinedAtNode = DILocation::getDistinct(
1364 Ctx, InlinedAtNode->getLine(), InlinedAtNode->getColumn(),
1365 InlinedAtNode->getScope(), InlinedAtNode->getInlinedAt());
1366
1367 // Cache the inlined-at nodes as they're built so they are reused, without
1368 // this every instruction's inlined-at chain would become distinct from each
1369 // other.
1370 DenseMap<const MDNode *, MDNode *> IANodes;
1371
1372 for (; FI != Fn->end(); ++FI) {
1373 for (BasicBlock::iterator BI = FI->begin(), BE = FI->end();
1374 BI != BE; ++BI) {
1375 if (DebugLoc DL = BI->getDebugLoc()) {
1376 auto IA = DebugLoc::appendInlinedAt(DL, InlinedAtNode, BI->getContext(),
1377 IANodes);
1378 auto IDL = DebugLoc::get(DL.getLine(), DL.getCol(), DL.getScope(), IA);
1379 BI->setDebugLoc(IDL);
1380 continue;
1381 }
1382
1383 if (CalleeHasDebugInfo)
1384 continue;
1385
1386 // If the inlined instruction has no line number, make it look as if it
1387 // originates from the call location. This is important for
1388 // ((__always_inline__, __nodebug__)) functions which must use caller
1389 // location for all instructions in their function body.
1390
1391 // Don't update static allocas, as they may get moved later.
1392 if (auto *AI = dyn_cast<AllocaInst>(BI))
1393 if (allocaWouldBeStaticInEntry(AI))
1394 continue;
1395
1396 BI->setDebugLoc(TheCallDL);
1397 }
1398 }
1399}
1400
1401/// Update the block frequencies of the caller after a callee has been inlined.
1402///
1403/// Each block cloned into the caller has its block frequency scaled by the
1404/// ratio of CallSiteFreq/CalleeEntryFreq. This ensures that the cloned copy of
1405/// callee's entry block gets the same frequency as the callsite block and the
1406/// relative frequencies of all cloned blocks remain the same after cloning.
1407static void updateCallerBFI(BasicBlock *CallSiteBlock,
1408 const ValueToValueMapTy &VMap,
1409 BlockFrequencyInfo *CallerBFI,
1410 BlockFrequencyInfo *CalleeBFI,
1411 const BasicBlock &CalleeEntryBlock) {
1412 SmallPtrSet<BasicBlock *, 16> ClonedBBs;
1413 for (auto const &Entry : VMap) {
1414 if (!isa<BasicBlock>(Entry.first) || !Entry.second)
1415 continue;
1416 auto *OrigBB = cast<BasicBlock>(Entry.first);
1417 auto *ClonedBB = cast<BasicBlock>(Entry.second);
1418 uint64_t Freq = CalleeBFI->getBlockFreq(OrigBB).getFrequency();
1419 if (!ClonedBBs.insert(ClonedBB).second) {
1420 // Multiple blocks in the callee might get mapped to one cloned block in
1421 // the caller since we prune the callee as we clone it. When that happens,
1422 // we want to use the maximum among the original blocks' frequencies.
1423 uint64_t NewFreq = CallerBFI->getBlockFreq(ClonedBB).getFrequency();
1424 if (NewFreq > Freq)
1425 Freq = NewFreq;
1426 }
1427 CallerBFI->setBlockFreq(ClonedBB, Freq);
1428 }
1429 BasicBlock *EntryClone = cast<BasicBlock>(VMap.lookup(&CalleeEntryBlock));
1430 CallerBFI->setBlockFreqAndScale(
1431 EntryClone, CallerBFI->getBlockFreq(CallSiteBlock).getFrequency(),
1432 ClonedBBs);
1433}
1434
1435/// Update the branch metadata for cloned call instructions.
1436static void updateCallProfile(Function *Callee, const ValueToValueMapTy &VMap,
1437 const ProfileCount &CalleeEntryCount,
1438 const Instruction *TheCall,
1439 ProfileSummaryInfo *PSI,
1440 BlockFrequencyInfo *CallerBFI) {
1441 if (!CalleeEntryCount.hasValue() || CalleeEntryCount.isSynthetic() ||
1442 CalleeEntryCount.getCount() < 1)
1443 return;
1444 auto CallSiteCount = PSI ? PSI->getProfileCount(TheCall, CallerBFI) : None;
1445 uint64_t CallCount =
1446 std::min(CallSiteCount.hasValue() ? CallSiteCount.getValue() : 0,
1447 CalleeEntryCount.getCount());
1448
1449 for (auto const &Entry : VMap)
1450 if (isa<CallInst>(Entry.first))
1451 if (auto *CI = dyn_cast_or_null<CallInst>(Entry.second))
1452 CI->updateProfWeight(CallCount, CalleeEntryCount.getCount());
1453 for (BasicBlock &BB : *Callee)
1454 // No need to update the callsite if it is pruned during inlining.
1455 if (VMap.count(&BB))
1456 for (Instruction &I : BB)
1457 if (CallInst *CI = dyn_cast<CallInst>(&I))
1458 CI->updateProfWeight(CalleeEntryCount.getCount() - CallCount,
1459 CalleeEntryCount.getCount());
1460}
1461
1462/// Update the entry count of callee after inlining.
1463///
1464/// The callsite's block count is subtracted from the callee's function entry
1465/// count.
1466static void updateCalleeCount(BlockFrequencyInfo *CallerBFI, BasicBlock *CallBB,
1467 Instruction *CallInst, Function *Callee,
1468 ProfileSummaryInfo *PSI) {
1469 // If the callee has a original count of N, and the estimated count of
1470 // callsite is M, the new callee count is set to N - M. M is estimated from
1471 // the caller's entry count, its entry block frequency and the block frequency
1472 // of the callsite.
1473 auto CalleeCount = Callee->getEntryCount();
1474 if (!CalleeCount.hasValue() || !PSI)
1475 return;
1476 auto CallCount = PSI->getProfileCount(CallInst, CallerBFI);
1477 if (!CallCount.hasValue())
1478 return;
1479 // Since CallSiteCount is an estimate, it could exceed the original callee
1480 // count and has to be set to 0.
1481 if (CallCount.getValue() > CalleeCount.getCount())
1482 CalleeCount.setCount(0);
1483 else
1484 CalleeCount.setCount(CalleeCount.getCount() - CallCount.getValue());
1485 Callee->setEntryCount(CalleeCount);
1486}
1487
1488/// This function inlines the called function into the basic block of the
1489/// caller. This returns false if it is not possible to inline this call.
1490/// The program is still in a well defined state if this occurs though.
1491///
1492/// Note that this only does one level of inlining. For example, if the
1493/// instruction 'call B' is inlined, and 'B' calls 'C', then the call to 'C' now
1494/// exists in the instruction stream. Similarly this will inline a recursive
1495/// function by one level.
1496llvm::InlineResult llvm::InlineFunction(CallSite CS, InlineFunctionInfo &IFI,
1497 AAResults *CalleeAAR,
1498 bool InsertLifetime,
1499 Function *ForwardVarArgsTo) {
1500 Instruction *TheCall = CS.getInstruction();
1501 assert(TheCall->getParent() && TheCall->getFunction()((TheCall->getParent() && TheCall->getFunction(
) && "Instruction not in function!") ? static_cast<
void> (0) : __assert_fail ("TheCall->getParent() && TheCall->getFunction() && \"Instruction not in function!\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 1502, __PRETTY_FUNCTION__))
1502 && "Instruction not in function!")((TheCall->getParent() && TheCall->getFunction(
) && "Instruction not in function!") ? static_cast<
void> (0) : __assert_fail ("TheCall->getParent() && TheCall->getFunction() && \"Instruction not in function!\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 1502, __PRETTY_FUNCTION__))
;
1503
1504 // If IFI has any state in it, zap it before we fill it in.
1505 IFI.reset();
1506
1507 Function *CalledFunc = CS.getCalledFunction();
1508 if (!CalledFunc || // Can't inline external function or indirect
1509 CalledFunc->isDeclaration()) // call!
1510 return "external or indirect";
1511
1512 // The inliner does not know how to inline through calls with operand bundles
1513 // in general ...
1514 if (CS.hasOperandBundles()) {
1515 for (int i = 0, e = CS.getNumOperandBundles(); i != e; ++i) {
1516 uint32_t Tag = CS.getOperandBundleAt(i).getTagID();
1517 // ... but it knows how to inline through "deopt" operand bundles ...
1518 if (Tag == LLVMContext::OB_deopt)
1519 continue;
1520 // ... and "funclet" operand bundles.
1521 if (Tag == LLVMContext::OB_funclet)
1522 continue;
1523
1524 return "unsupported operand bundle";
1525 }
1526 }
1527
1528 // If the call to the callee cannot throw, set the 'nounwind' flag on any
1529 // calls that we inline.
1530 bool MarkNoUnwind = CS.doesNotThrow();
1531
1532 BasicBlock *OrigBB = TheCall->getParent();
1533 Function *Caller = OrigBB->getParent();
1534
1535 // GC poses two hazards to inlining, which only occur when the callee has GC:
1536 // 1. If the caller has no GC, then the callee's GC must be propagated to the
1537 // caller.
1538 // 2. If the caller has a differing GC, it is invalid to inline.
1539 if (CalledFunc->hasGC()) {
1540 if (!Caller->hasGC())
1541 Caller->setGC(CalledFunc->getGC());
1542 else if (CalledFunc->getGC() != Caller->getGC())
1543 return "incompatible GC";
1544 }
1545
1546 // Get the personality function from the callee if it contains a landing pad.
1547 Constant *CalledPersonality =
1548 CalledFunc->hasPersonalityFn()
1549 ? CalledFunc->getPersonalityFn()->stripPointerCasts()
1550 : nullptr;
1551
1552 // Find the personality function used by the landing pads of the caller. If it
1553 // exists, then check to see that it matches the personality function used in
1554 // the callee.
1555 Constant *CallerPersonality =
1556 Caller->hasPersonalityFn()
1557 ? Caller->getPersonalityFn()->stripPointerCasts()
1558 : nullptr;
1559 if (CalledPersonality) {
1560 if (!CallerPersonality)
1561 Caller->setPersonalityFn(CalledPersonality);
1562 // If the personality functions match, then we can perform the
1563 // inlining. Otherwise, we can't inline.
1564 // TODO: This isn't 100% true. Some personality functions are proper
1565 // supersets of others and can be used in place of the other.
1566 else if (CalledPersonality != CallerPersonality)
1567 return "incompatible personality";
1568 }
1569
1570 // We need to figure out which funclet the callsite was in so that we may
1571 // properly nest the callee.
1572 Instruction *CallSiteEHPad = nullptr;
1573 if (CallerPersonality) {
1574 EHPersonality Personality = classifyEHPersonality(CallerPersonality);
1575 if (isScopedEHPersonality(Personality)) {
1576 Optional<OperandBundleUse> ParentFunclet =
1577 CS.getOperandBundle(LLVMContext::OB_funclet);
1578 if (ParentFunclet)
1579 CallSiteEHPad = cast<FuncletPadInst>(ParentFunclet->Inputs.front());
1580
1581 // OK, the inlining site is legal. What about the target function?
1582
1583 if (CallSiteEHPad) {
1584 if (Personality == EHPersonality::MSVC_CXX) {
1585 // The MSVC personality cannot tolerate catches getting inlined into
1586 // cleanup funclets.
1587 if (isa<CleanupPadInst>(CallSiteEHPad)) {
1588 // Ok, the call site is within a cleanuppad. Let's check the callee
1589 // for catchpads.
1590 for (const BasicBlock &CalledBB : *CalledFunc) {
1591 if (isa<CatchSwitchInst>(CalledBB.getFirstNonPHI()))
1592 return "catch in cleanup funclet";
1593 }
1594 }
1595 } else if (isAsynchronousEHPersonality(Personality)) {
1596 // SEH is even less tolerant, there may not be any sort of exceptional
1597 // funclet in the callee.
1598 for (const BasicBlock &CalledBB : *CalledFunc) {
1599 if (CalledBB.isEHPad())
1600 return "SEH in cleanup funclet";
1601 }
1602 }
1603 }
1604 }
1605 }
1606
1607 // Determine if we are dealing with a call in an EHPad which does not unwind
1608 // to caller.
1609 bool EHPadForCallUnwindsLocally = false;
1610 if (CallSiteEHPad && CS.isCall()) {
1611 UnwindDestMemoTy FuncletUnwindMap;
1612 Value *CallSiteUnwindDestToken =
1613 getUnwindDestToken(CallSiteEHPad, FuncletUnwindMap);
1614
1615 EHPadForCallUnwindsLocally =
1616 CallSiteUnwindDestToken &&
1617 !isa<ConstantTokenNone>(CallSiteUnwindDestToken);
1618 }
1619
1620 // Get an iterator to the last basic block in the function, which will have
1621 // the new function inlined after it.
1622 Function::iterator LastBlock = --Caller->end();
1623
1624 // Make sure to capture all of the return instructions from the cloned
1625 // function.
1626 SmallVector<ReturnInst*, 8> Returns;
1627 ClonedCodeInfo InlinedFunctionInfo;
1628 Function::iterator FirstNewBlock;
1629
1630 { // Scope to destroy VMap after cloning.
1631 ValueToValueMapTy VMap;
1632 // Keep a list of pair (dst, src) to emit byval initializations.
1633 SmallVector<std::pair<Value*, Value*>, 4> ByValInit;
1634
1635 auto &DL = Caller->getParent()->getDataLayout();
1636
1637 // Calculate the vector of arguments to pass into the function cloner, which
1638 // matches up the formal to the actual argument values.
1639 CallSite::arg_iterator AI = CS.arg_begin();
1640 unsigned ArgNo = 0;
1641 for (Function::arg_iterator I = CalledFunc->arg_begin(),
1642 E = CalledFunc->arg_end(); I != E; ++I, ++AI, ++ArgNo) {
1643 Value *ActualArg = *AI;
1644
1645 // When byval arguments actually inlined, we need to make the copy implied
1646 // by them explicit. However, we don't do this if the callee is readonly
1647 // or readnone, because the copy would be unneeded: the callee doesn't
1648 // modify the struct.
1649 if (CS.isByValArgument(ArgNo)) {
1650 ActualArg = HandleByValArgument(ActualArg, TheCall, CalledFunc, IFI,
1651 CalledFunc->getParamAlignment(ArgNo));
1652 if (ActualArg != *AI)
1653 ByValInit.push_back(std::make_pair(ActualArg, (Value*) *AI));
1654 }
1655
1656 VMap[&*I] = ActualArg;
1657 }
1658
1659 // Add alignment assumptions if necessary. We do this before the inlined
1660 // instructions are actually cloned into the caller so that we can easily
1661 // check what will be known at the start of the inlined code.
1662 AddAlignmentAssumptions(CS, IFI);
1663
1664 // We want the inliner to prune the code as it copies. We would LOVE to
1665 // have no dead or constant instructions leftover after inlining occurs
1666 // (which can happen, e.g., because an argument was constant), but we'll be
1667 // happy with whatever the cloner can do.
1668 CloneAndPruneFunctionInto(Caller, CalledFunc, VMap,
1669 /*ModuleLevelChanges=*/false, Returns, ".i",
1670 &InlinedFunctionInfo, TheCall);
1671 // Remember the first block that is newly cloned over.
1672 FirstNewBlock = LastBlock; ++FirstNewBlock;
1673
1674 if (IFI.CallerBFI != nullptr && IFI.CalleeBFI != nullptr)
1675 // Update the BFI of blocks cloned into the caller.
1676 updateCallerBFI(OrigBB, VMap, IFI.CallerBFI, IFI.CalleeBFI,
1677 CalledFunc->front());
1678
1679 updateCallProfile(CalledFunc, VMap, CalledFunc->getEntryCount(), TheCall,
1680 IFI.PSI, IFI.CallerBFI);
1681 // Update the profile count of callee.
1682 updateCalleeCount(IFI.CallerBFI, OrigBB, TheCall, CalledFunc, IFI.PSI);
1683
1684 // Inject byval arguments initialization.
1685 for (std::pair<Value*, Value*> &Init : ByValInit)
1686 HandleByValArgumentInit(Init.first, Init.second, Caller->getParent(),
1687 &*FirstNewBlock, IFI);
1688
1689 Optional<OperandBundleUse> ParentDeopt =
1690 CS.getOperandBundle(LLVMContext::OB_deopt);
1691 if (ParentDeopt) {
1692 SmallVector<OperandBundleDef, 2> OpDefs;
1693
1694 for (auto &VH : InlinedFunctionInfo.OperandBundleCallSites) {
1695 Instruction *I = dyn_cast_or_null<Instruction>(VH);
1696 if (!I) continue; // instruction was DCE'd or RAUW'ed to undef
1697
1698 OpDefs.clear();
1699
1700 CallSite ICS(I);
1701 OpDefs.reserve(ICS.getNumOperandBundles());
1702
1703 for (unsigned i = 0, e = ICS.getNumOperandBundles(); i < e; ++i) {
1704 auto ChildOB = ICS.getOperandBundleAt(i);
1705 if (ChildOB.getTagID() != LLVMContext::OB_deopt) {
1706 // If the inlined call has other operand bundles, let them be
1707 OpDefs.emplace_back(ChildOB);
1708 continue;
1709 }
1710
1711 // It may be useful to separate this logic (of handling operand
1712 // bundles) out to a separate "policy" component if this gets crowded.
1713 // Prepend the parent's deoptimization continuation to the newly
1714 // inlined call's deoptimization continuation.
1715 std::vector<Value *> MergedDeoptArgs;
1716 MergedDeoptArgs.reserve(ParentDeopt->Inputs.size() +
1717 ChildOB.Inputs.size());
1718
1719 MergedDeoptArgs.insert(MergedDeoptArgs.end(),
1720 ParentDeopt->Inputs.begin(),
1721 ParentDeopt->Inputs.end());
1722 MergedDeoptArgs.insert(MergedDeoptArgs.end(), ChildOB.Inputs.begin(),
1723 ChildOB.Inputs.end());
1724
1725 OpDefs.emplace_back("deopt", std::move(MergedDeoptArgs));
1726 }
1727
1728 Instruction *NewI = nullptr;
1729 if (isa<CallInst>(I))
1730 NewI = CallInst::Create(cast<CallInst>(I), OpDefs, I);
1731 else
1732 NewI = InvokeInst::Create(cast<InvokeInst>(I), OpDefs, I);
1733
1734 // Note: the RAUW does the appropriate fixup in VMap, so we need to do
1735 // this even if the call returns void.
1736 I->replaceAllUsesWith(NewI);
1737
1738 VH = nullptr;
1739 I->eraseFromParent();
1740 }
1741 }
1742
1743 // Update the callgraph if requested.
1744 if (IFI.CG)
1745 UpdateCallGraphAfterInlining(CS, FirstNewBlock, VMap, IFI);
1746
1747 // For 'nodebug' functions, the associated DISubprogram is always null.
1748 // Conservatively avoid propagating the callsite debug location to
1749 // instructions inlined from a function whose DISubprogram is not null.
1750 fixupLineNumbers(Caller, FirstNewBlock, TheCall,
1751 CalledFunc->getSubprogram() != nullptr);
1752
1753 // Clone existing noalias metadata if necessary.
1754 CloneAliasScopeMetadata(CS, VMap);
1755
1756 // Add noalias metadata if necessary.
1757 AddAliasScopeMetadata(CS, VMap, DL, CalleeAAR);
1758
1759 // Propagate llvm.mem.parallel_loop_access if necessary.
1760 PropagateParallelLoopAccessMetadata(CS, VMap);
1761
1762 // Register any cloned assumptions.
1763 if (IFI.GetAssumptionCache)
1764 for (BasicBlock &NewBlock :
1765 make_range(FirstNewBlock->getIterator(), Caller->end()))
1766 for (Instruction &I : NewBlock) {
1767 if (auto *II = dyn_cast<IntrinsicInst>(&I))
1768 if (II->getIntrinsicID() == Intrinsic::assume)
1769 (*IFI.GetAssumptionCache)(*Caller).registerAssumption(II);
1770 }
1771 }
1772
1773 // If there are any alloca instructions in the block that used to be the entry
1774 // block for the callee, move them to the entry block of the caller. First
1775 // calculate which instruction they should be inserted before. We insert the
1776 // instructions at the end of the current alloca list.
1777 {
1778 BasicBlock::iterator InsertPoint = Caller->begin()->begin();
1779 for (BasicBlock::iterator I = FirstNewBlock->begin(),
1780 E = FirstNewBlock->end(); I != E; ) {
1781 AllocaInst *AI = dyn_cast<AllocaInst>(I++);
1782 if (!AI) continue;
1783
1784 // If the alloca is now dead, remove it. This often occurs due to code
1785 // specialization.
1786 if (AI->use_empty()) {
1787 AI->eraseFromParent();
1788 continue;
1789 }
1790
1791 if (!allocaWouldBeStaticInEntry(AI))
1792 continue;
1793
1794 // Keep track of the static allocas that we inline into the caller.
1795 IFI.StaticAllocas.push_back(AI);
1796
1797 // Scan for the block of allocas that we can move over, and move them
1798 // all at once.
1799 while (isa<AllocaInst>(I) &&
1800 allocaWouldBeStaticInEntry(cast<AllocaInst>(I))) {
1801 IFI.StaticAllocas.push_back(cast<AllocaInst>(I));
1802 ++I;
1803 }
1804
1805 // Transfer all of the allocas over in a block. Using splice means
1806 // that the instructions aren't removed from the symbol table, then
1807 // reinserted.
1808 Caller->getEntryBlock().getInstList().splice(
1809 InsertPoint, FirstNewBlock->getInstList(), AI->getIterator(), I);
1810 }
1811 // Move any dbg.declares describing the allocas into the entry basic block.
1812 DIBuilder DIB(*Caller->getParent());
1813 for (auto &AI : IFI.StaticAllocas)
1814 replaceDbgDeclareForAlloca(AI, AI, DIB, DIExpression::NoDeref, 0,
1815 DIExpression::NoDeref);
1816 }
1817
1818 SmallVector<Value*,4> VarArgsToForward;
1819 SmallVector<AttributeSet, 4> VarArgsAttrs;
1820 for (unsigned i = CalledFunc->getFunctionType()->getNumParams();
1821 i < CS.getNumArgOperands(); i++) {
1822 VarArgsToForward.push_back(CS.getArgOperand(i));
1823 VarArgsAttrs.push_back(CS.getAttributes().getParamAttributes(i));
1824 }
1825
1826 bool InlinedMustTailCalls = false, InlinedDeoptimizeCalls = false;
1827 if (InlinedFunctionInfo.ContainsCalls) {
1828 CallInst::TailCallKind CallSiteTailKind = CallInst::TCK_None;
1829 if (CallInst *CI = dyn_cast<CallInst>(TheCall))
1830 CallSiteTailKind = CI->getTailCallKind();
1831
1832 // For inlining purposes, the "notail" marker is the same as no marker.
1833 if (CallSiteTailKind == CallInst::TCK_NoTail)
1834 CallSiteTailKind = CallInst::TCK_None;
1835
1836 for (Function::iterator BB = FirstNewBlock, E = Caller->end(); BB != E;
1837 ++BB) {
1838 for (auto II = BB->begin(); II != BB->end();) {
1839 Instruction &I = *II++;
1840 CallInst *CI = dyn_cast<CallInst>(&I);
1841 if (!CI)
1842 continue;
1843
1844 // Forward varargs from inlined call site to calls to the
1845 // ForwardVarArgsTo function, if requested, and to musttail calls.
1846 if (!VarArgsToForward.empty() &&
1847 ((ForwardVarArgsTo &&
1848 CI->getCalledFunction() == ForwardVarArgsTo) ||
1849 CI->isMustTailCall())) {
1850 // Collect attributes for non-vararg parameters.
1851 AttributeList Attrs = CI->getAttributes();
1852 SmallVector<AttributeSet, 8> ArgAttrs;
1853 if (!Attrs.isEmpty() || !VarArgsAttrs.empty()) {
1854 for (unsigned ArgNo = 0;
1855 ArgNo < CI->getFunctionType()->getNumParams(); ++ArgNo)
1856 ArgAttrs.push_back(Attrs.getParamAttributes(ArgNo));
1857 }
1858
1859 // Add VarArg attributes.
1860 ArgAttrs.append(VarArgsAttrs.begin(), VarArgsAttrs.end());
1861 Attrs = AttributeList::get(CI->getContext(), Attrs.getFnAttributes(),
1862 Attrs.getRetAttributes(), ArgAttrs);
1863 // Add VarArgs to existing parameters.
1864 SmallVector<Value *, 6> Params(CI->arg_operands());
1865 Params.append(VarArgsToForward.begin(), VarArgsToForward.end());
1866 CallInst *NewCI =
1867 CallInst::Create(CI->getCalledFunction() ? CI->getCalledFunction()
1868 : CI->getCalledValue(),
1869 Params, "", CI);
1870 NewCI->setDebugLoc(CI->getDebugLoc());
1871 NewCI->setAttributes(Attrs);
1872 NewCI->setCallingConv(CI->getCallingConv());
1873 CI->replaceAllUsesWith(NewCI);
1874 CI->eraseFromParent();
1875 CI = NewCI;
1876 }
1877
1878 if (Function *F = CI->getCalledFunction())
1879 InlinedDeoptimizeCalls |=
1880 F->getIntrinsicID() == Intrinsic::experimental_deoptimize;
1881
1882 // We need to reduce the strength of any inlined tail calls. For
1883 // musttail, we have to avoid introducing potential unbounded stack
1884 // growth. For example, if functions 'f' and 'g' are mutually recursive
1885 // with musttail, we can inline 'g' into 'f' so long as we preserve
1886 // musttail on the cloned call to 'f'. If either the inlined call site
1887 // or the cloned call site is *not* musttail, the program already has
1888 // one frame of stack growth, so it's safe to remove musttail. Here is
1889 // a table of example transformations:
1890 //
1891 // f -> musttail g -> musttail f ==> f -> musttail f
1892 // f -> musttail g -> tail f ==> f -> tail f
1893 // f -> g -> musttail f ==> f -> f
1894 // f -> g -> tail f ==> f -> f
1895 //
1896 // Inlined notail calls should remain notail calls.
1897 CallInst::TailCallKind ChildTCK = CI->getTailCallKind();
1898 if (ChildTCK != CallInst::TCK_NoTail)
1899 ChildTCK = std::min(CallSiteTailKind, ChildTCK);
1900 CI->setTailCallKind(ChildTCK);
1901 InlinedMustTailCalls |= CI->isMustTailCall();
1902
1903 // Calls inlined through a 'nounwind' call site should be marked
1904 // 'nounwind'.
1905 if (MarkNoUnwind)
1906 CI->setDoesNotThrow();
1907 }
1908 }
1909 }
1910
1911 // Leave lifetime markers for the static alloca's, scoping them to the
1912 // function we just inlined.
1913 if (InsertLifetime && !IFI.StaticAllocas.empty()) {
1914 IRBuilder<> builder(&FirstNewBlock->front());
1915 for (unsigned ai = 0, ae = IFI.StaticAllocas.size(); ai != ae; ++ai) {
1916 AllocaInst *AI = IFI.StaticAllocas[ai];
1917 // Don't mark swifterror allocas. They can't have bitcast uses.
1918 if (AI->isSwiftError())
1919 continue;
1920
1921 // If the alloca is already scoped to something smaller than the whole
1922 // function then there's no need to add redundant, less accurate markers.
1923 if (hasLifetimeMarkers(AI))
1924 continue;
1925
1926 // Try to determine the size of the allocation.
1927 ConstantInt *AllocaSize = nullptr;
1928 if (ConstantInt *AIArraySize =
1929 dyn_cast<ConstantInt>(AI->getArraySize())) {
1930 auto &DL = Caller->getParent()->getDataLayout();
1931 Type *AllocaType = AI->getAllocatedType();
1932 uint64_t AllocaTypeSize = DL.getTypeAllocSize(AllocaType);
1933 uint64_t AllocaArraySize = AIArraySize->getLimitedValue();
1934
1935 // Don't add markers for zero-sized allocas.
1936 if (AllocaArraySize == 0)
1937 continue;
1938
1939 // Check that array size doesn't saturate uint64_t and doesn't
1940 // overflow when it's multiplied by type size.
1941 if (AllocaArraySize != std::numeric_limits<uint64_t>::max() &&
1942 std::numeric_limits<uint64_t>::max() / AllocaArraySize >=
1943 AllocaTypeSize) {
1944 AllocaSize = ConstantInt::get(Type::getInt64Ty(AI->getContext()),
1945 AllocaArraySize * AllocaTypeSize);
1946 }
1947 }
1948
1949 builder.CreateLifetimeStart(AI, AllocaSize);
1950 for (ReturnInst *RI : Returns) {
1951 // Don't insert llvm.lifetime.end calls between a musttail or deoptimize
1952 // call and a return. The return kills all local allocas.
1953 if (InlinedMustTailCalls &&
1954 RI->getParent()->getTerminatingMustTailCall())
1955 continue;
1956 if (InlinedDeoptimizeCalls &&
1957 RI->getParent()->getTerminatingDeoptimizeCall())
1958 continue;
1959 IRBuilder<>(RI).CreateLifetimeEnd(AI, AllocaSize);
1960 }
1961 }
1962 }
1963
1964 // If the inlined code contained dynamic alloca instructions, wrap the inlined
1965 // code with llvm.stacksave/llvm.stackrestore intrinsics.
1966 if (InlinedFunctionInfo.ContainsDynamicAllocas) {
1967 Module *M = Caller->getParent();
1968 // Get the two intrinsics we care about.
1969 Function *StackSave = Intrinsic::getDeclaration(M, Intrinsic::stacksave);
1970 Function *StackRestore=Intrinsic::getDeclaration(M,Intrinsic::stackrestore);
1971
1972 // Insert the llvm.stacksave.
1973 CallInst *SavedPtr = IRBuilder<>(&*FirstNewBlock, FirstNewBlock->begin())
1974 .CreateCall(StackSave, {}, "savedstack");
1975
1976 // Insert a call to llvm.stackrestore before any return instructions in the
1977 // inlined function.
1978 for (ReturnInst *RI : Returns) {
1979 // Don't insert llvm.stackrestore calls between a musttail or deoptimize
1980 // call and a return. The return will restore the stack pointer.
1981 if (InlinedMustTailCalls && RI->getParent()->getTerminatingMustTailCall())
1982 continue;
1983 if (InlinedDeoptimizeCalls && RI->getParent()->getTerminatingDeoptimizeCall())
1984 continue;
1985 IRBuilder<>(RI).CreateCall(StackRestore, SavedPtr);
1986 }
1987 }
1988
1989 // If we are inlining for an invoke instruction, we must make sure to rewrite
1990 // any call instructions into invoke instructions. This is sensitive to which
1991 // funclet pads were top-level in the inlinee, so must be done before
1992 // rewriting the "parent pad" links.
1993 if (auto *II = dyn_cast<InvokeInst>(TheCall)) {
1994 BasicBlock *UnwindDest = II->getUnwindDest();
1995 Instruction *FirstNonPHI = UnwindDest->getFirstNonPHI();
1996 if (isa<LandingPadInst>(FirstNonPHI)) {
1997 HandleInlinedLandingPad(II, &*FirstNewBlock, InlinedFunctionInfo);
1998 } else {
1999 HandleInlinedEHPad(II, &*FirstNewBlock, InlinedFunctionInfo);
2000 }
2001 }
2002
2003 // Update the lexical scopes of the new funclets and callsites.
2004 // Anything that had 'none' as its parent is now nested inside the callsite's
2005 // EHPad.
2006
2007 if (CallSiteEHPad) {
2008 for (Function::iterator BB = FirstNewBlock->getIterator(),
2009 E = Caller->end();
2010 BB != E; ++BB) {
2011 // Add bundle operands to any top-level call sites.
2012 SmallVector<OperandBundleDef, 1> OpBundles;
2013 for (BasicBlock::iterator BBI = BB->begin(), E = BB->end(); BBI != E;) {
2014 Instruction *I = &*BBI++;
2015 CallSite CS(I);
2016 if (!CS)
2017 continue;
2018
2019 // Skip call sites which are nounwind intrinsics.
2020 auto *CalledFn =
2021 dyn_cast<Function>(CS.getCalledValue()->stripPointerCasts());
2022 if (CalledFn && CalledFn->isIntrinsic() && CS.doesNotThrow())
2023 continue;
2024
2025 // Skip call sites which already have a "funclet" bundle.
2026 if (CS.getOperandBundle(LLVMContext::OB_funclet))
2027 continue;
2028
2029 CS.getOperandBundlesAsDefs(OpBundles);
2030 OpBundles.emplace_back("funclet", CallSiteEHPad);
2031
2032 Instruction *NewInst;
2033 if (CS.isCall())
2034 NewInst = CallInst::Create(cast<CallInst>(I), OpBundles, I);
2035 else
2036 NewInst = InvokeInst::Create(cast<InvokeInst>(I), OpBundles, I);
2037 NewInst->takeName(I);
2038 I->replaceAllUsesWith(NewInst);
2039 I->eraseFromParent();
2040
2041 OpBundles.clear();
2042 }
2043
2044 // It is problematic if the inlinee has a cleanupret which unwinds to
2045 // caller and we inline it into a call site which doesn't unwind but into
2046 // an EH pad that does. Such an edge must be dynamically unreachable.
2047 // As such, we replace the cleanupret with unreachable.
2048 if (auto *CleanupRet = dyn_cast<CleanupReturnInst>(BB->getTerminator()))
2049 if (CleanupRet->unwindsToCaller() && EHPadForCallUnwindsLocally)
2050 changeToUnreachable(CleanupRet, /*UseLLVMTrap=*/false);
2051
2052 Instruction *I = BB->getFirstNonPHI();
2053 if (!I->isEHPad())
2054 continue;
2055
2056 if (auto *CatchSwitch = dyn_cast<CatchSwitchInst>(I)) {
2057 if (isa<ConstantTokenNone>(CatchSwitch->getParentPad()))
2058 CatchSwitch->setParentPad(CallSiteEHPad);
2059 } else {
2060 auto *FPI = cast<FuncletPadInst>(I);
2061 if (isa<ConstantTokenNone>(FPI->getParentPad()))
2062 FPI->setParentPad(CallSiteEHPad);
2063 }
2064 }
2065 }
2066
2067 if (InlinedDeoptimizeCalls) {
2068 // We need to at least remove the deoptimizing returns from the Return set,
2069 // so that the control flow from those returns does not get merged into the
2070 // caller (but terminate it instead). If the caller's return type does not
2071 // match the callee's return type, we also need to change the return type of
2072 // the intrinsic.
2073 if (Caller->getReturnType() == TheCall->getType()) {
2074 auto NewEnd = llvm::remove_if(Returns, [](ReturnInst *RI) {
2075 return RI->getParent()->getTerminatingDeoptimizeCall() != nullptr;
2076 });
2077 Returns.erase(NewEnd, Returns.end());
2078 } else {
2079 SmallVector<ReturnInst *, 8> NormalReturns;
2080 Function *NewDeoptIntrinsic = Intrinsic::getDeclaration(
2081 Caller->getParent(), Intrinsic::experimental_deoptimize,
2082 {Caller->getReturnType()});
2083
2084 for (ReturnInst *RI : Returns) {
2085 CallInst *DeoptCall = RI->getParent()->getTerminatingDeoptimizeCall();
2086 if (!DeoptCall) {
2087 NormalReturns.push_back(RI);
2088 continue;
2089 }
2090
2091 // The calling convention on the deoptimize call itself may be bogus,
2092 // since the code we're inlining may have undefined behavior (and may
2093 // never actually execute at runtime); but all
2094 // @llvm.experimental.deoptimize declarations have to have the same
2095 // calling convention in a well-formed module.
2096 auto CallingConv = DeoptCall->getCalledFunction()->getCallingConv();
2097 NewDeoptIntrinsic->setCallingConv(CallingConv);
2098 auto *CurBB = RI->getParent();
2099 RI->eraseFromParent();
2100
2101 SmallVector<Value *, 4> CallArgs(DeoptCall->arg_begin(),
2102 DeoptCall->arg_end());
2103
2104 SmallVector<OperandBundleDef, 1> OpBundles;
2105 DeoptCall->getOperandBundlesAsDefs(OpBundles);
2106 DeoptCall->eraseFromParent();
2107 assert(!OpBundles.empty() &&((!OpBundles.empty() && "Expected at least the deopt operand bundle"
) ? static_cast<void> (0) : __assert_fail ("!OpBundles.empty() && \"Expected at least the deopt operand bundle\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 2108, __PRETTY_FUNCTION__))
2108 "Expected at least the deopt operand bundle")((!OpBundles.empty() && "Expected at least the deopt operand bundle"
) ? static_cast<void> (0) : __assert_fail ("!OpBundles.empty() && \"Expected at least the deopt operand bundle\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 2108, __PRETTY_FUNCTION__))
;
2109
2110 IRBuilder<> Builder(CurBB);
2111 CallInst *NewDeoptCall =
2112 Builder.CreateCall(NewDeoptIntrinsic, CallArgs, OpBundles);
2113 NewDeoptCall->setCallingConv(CallingConv);
2114 if (NewDeoptCall->getType()->isVoidTy())
2115 Builder.CreateRetVoid();
2116 else
2117 Builder.CreateRet(NewDeoptCall);
2118 }
2119
2120 // Leave behind the normal returns so we can merge control flow.
2121 std::swap(Returns, NormalReturns);
2122 }
2123 }
2124
2125 // Handle any inlined musttail call sites. In order for a new call site to be
2126 // musttail, the source of the clone and the inlined call site must have been
2127 // musttail. Therefore it's safe to return without merging control into the
2128 // phi below.
2129 if (InlinedMustTailCalls) {
2130 // Check if we need to bitcast the result of any musttail calls.
2131 Type *NewRetTy = Caller->getReturnType();
2132 bool NeedBitCast = !TheCall->use_empty() && TheCall->getType() != NewRetTy;
2133
2134 // Handle the returns preceded by musttail calls separately.
2135 SmallVector<ReturnInst *, 8> NormalReturns;
2136 for (ReturnInst *RI : Returns) {
2137 CallInst *ReturnedMustTail =
2138 RI->getParent()->getTerminatingMustTailCall();
2139 if (!ReturnedMustTail) {
2140 NormalReturns.push_back(RI);
2141 continue;
2142 }
2143 if (!NeedBitCast)
2144 continue;
2145
2146 // Delete the old return and any preceding bitcast.
2147 BasicBlock *CurBB = RI->getParent();
2148 auto *OldCast = dyn_cast_or_null<BitCastInst>(RI->getReturnValue());
2149 RI->eraseFromParent();
2150 if (OldCast)
2151 OldCast->eraseFromParent();
2152
2153 // Insert a new bitcast and return with the right type.
2154 IRBuilder<> Builder(CurBB);
2155 Builder.CreateRet(Builder.CreateBitCast(ReturnedMustTail, NewRetTy));
2156 }
2157
2158 // Leave behind the normal returns so we can merge control flow.
2159 std::swap(Returns, NormalReturns);
2160 }
2161
2162 // Now that all of the transforms on the inlined code have taken place but
2163 // before we splice the inlined code into the CFG and lose track of which
2164 // blocks were actually inlined, collect the call sites. We only do this if
2165 // call graph updates weren't requested, as those provide value handle based
2166 // tracking of inlined call sites instead.
2167 if (InlinedFunctionInfo.ContainsCalls && !IFI.CG) {
2168 // Otherwise just collect the raw call sites that were inlined.
2169 for (BasicBlock &NewBB :
2170 make_range(FirstNewBlock->getIterator(), Caller->end()))
2171 for (Instruction &I : NewBB)
2172 if (auto CS = CallSite(&I))
2173 IFI.InlinedCallSites.push_back(CS);
2174 }
2175
2176 // If we cloned in _exactly one_ basic block, and if that block ends in a
2177 // return instruction, we splice the body of the inlined callee directly into
2178 // the calling basic block.
2179 if (Returns.size() == 1 && std::distance(FirstNewBlock, Caller->end()) == 1) {
2180 // Move all of the instructions right before the call.
2181 OrigBB->getInstList().splice(TheCall->getIterator(),
2182 FirstNewBlock->getInstList(),
2183 FirstNewBlock->begin(), FirstNewBlock->end());
2184 // Remove the cloned basic block.
2185 Caller->getBasicBlockList().pop_back();
2186
2187 // If the call site was an invoke instruction, add a branch to the normal
2188 // destination.
2189 if (InvokeInst *II = dyn_cast<InvokeInst>(TheCall)) {
2190 BranchInst *NewBr = BranchInst::Create(II->getNormalDest(), TheCall);
2191 NewBr->setDebugLoc(Returns[0]->getDebugLoc());
2192 }
2193
2194 // If the return instruction returned a value, replace uses of the call with
2195 // uses of the returned value.
2196 if (!TheCall->use_empty()) {
2197 ReturnInst *R = Returns[0];
2198 if (TheCall == R->getReturnValue())
2199 TheCall->replaceAllUsesWith(UndefValue::get(TheCall->getType()));
2200 else
2201 TheCall->replaceAllUsesWith(R->getReturnValue());
2202 }
2203 // Since we are now done with the Call/Invoke, we can delete it.
2204 TheCall->eraseFromParent();
2205
2206 // Since we are now done with the return instruction, delete it also.
2207 Returns[0]->eraseFromParent();
2208
2209 // We are now done with the inlining.
2210 return true;
2211 }
2212
2213 // Otherwise, we have the normal case, of more than one block to inline or
2214 // multiple return sites.
2215
2216 // We want to clone the entire callee function into the hole between the
2217 // "starter" and "ender" blocks. How we accomplish this depends on whether
2218 // this is an invoke instruction or a call instruction.
2219 BasicBlock *AfterCallBB;
2220 BranchInst *CreatedBranchToNormalDest = nullptr;
2221 if (InvokeInst *II = dyn_cast<InvokeInst>(TheCall)) {
2222
2223 // Add an unconditional branch to make this look like the CallInst case...
2224 CreatedBranchToNormalDest = BranchInst::Create(II->getNormalDest(), TheCall);
2225
2226 // Split the basic block. This guarantees that no PHI nodes will have to be
2227 // updated due to new incoming edges, and make the invoke case more
2228 // symmetric to the call case.
2229 AfterCallBB =
2230 OrigBB->splitBasicBlock(CreatedBranchToNormalDest->getIterator(),
2231 CalledFunc->getName() + ".exit");
2232
2233 } else { // It's a call
2234 // If this is a call instruction, we need to split the basic block that
2235 // the call lives in.
2236 //
2237 AfterCallBB = OrigBB->splitBasicBlock(TheCall->getIterator(),
2238 CalledFunc->getName() + ".exit");
2239 }
2240
2241 if (IFI.CallerBFI) {
2242 // Copy original BB's block frequency to AfterCallBB
2243 IFI.CallerBFI->setBlockFreq(
2244 AfterCallBB, IFI.CallerBFI->getBlockFreq(OrigBB).getFrequency());
2245 }
2246
2247 // Change the branch that used to go to AfterCallBB to branch to the first
2248 // basic block of the inlined function.
2249 //
2250 Instruction *Br = OrigBB->getTerminator();
2251 assert(Br && Br->getOpcode() == Instruction::Br &&((Br && Br->getOpcode() == Instruction::Br &&
"splitBasicBlock broken!") ? static_cast<void> (0) : __assert_fail
("Br && Br->getOpcode() == Instruction::Br && \"splitBasicBlock broken!\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 2252, __PRETTY_FUNCTION__))
2252 "splitBasicBlock broken!")((Br && Br->getOpcode() == Instruction::Br &&
"splitBasicBlock broken!") ? static_cast<void> (0) : __assert_fail
("Br && Br->getOpcode() == Instruction::Br && \"splitBasicBlock broken!\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 2252, __PRETTY_FUNCTION__))
;
2253 Br->setOperand(0, &*FirstNewBlock);
2254
2255 // Now that the function is correct, make it a little bit nicer. In
2256 // particular, move the basic blocks inserted from the end of the function
2257 // into the space made by splitting the source basic block.
2258 Caller->getBasicBlockList().splice(AfterCallBB->getIterator(),
2259 Caller->getBasicBlockList(), FirstNewBlock,
2260 Caller->end());
2261
2262 // Handle all of the return instructions that we just cloned in, and eliminate
2263 // any users of the original call/invoke instruction.
2264 Type *RTy = CalledFunc->getReturnType();
2265
2266 PHINode *PHI = nullptr;
2267 if (Returns.size() > 1) {
2268 // The PHI node should go at the front of the new basic block to merge all
2269 // possible incoming values.
2270 if (!TheCall->use_empty()) {
2271 PHI = PHINode::Create(RTy, Returns.size(), TheCall->getName(),
2272 &AfterCallBB->front());
2273 // Anything that used the result of the function call should now use the
2274 // PHI node as their operand.
2275 TheCall->replaceAllUsesWith(PHI);
2276 }
2277
2278 // Loop over all of the return instructions adding entries to the PHI node
2279 // as appropriate.
2280 if (PHI) {
2281 for (unsigned i = 0, e = Returns.size(); i != e; ++i) {
2282 ReturnInst *RI = Returns[i];
2283 assert(RI->getReturnValue()->getType() == PHI->getType() &&((RI->getReturnValue()->getType() == PHI->getType() &&
"Ret value not consistent in function!") ? static_cast<void
> (0) : __assert_fail ("RI->getReturnValue()->getType() == PHI->getType() && \"Ret value not consistent in function!\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 2284, __PRETTY_FUNCTION__))
2284 "Ret value not consistent in function!")((RI->getReturnValue()->getType() == PHI->getType() &&
"Ret value not consistent in function!") ? static_cast<void
> (0) : __assert_fail ("RI->getReturnValue()->getType() == PHI->getType() && \"Ret value not consistent in function!\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 2284, __PRETTY_FUNCTION__))
;
2285 PHI->addIncoming(RI->getReturnValue(), RI->getParent());
2286 }
2287 }
2288
2289 // Add a branch to the merge points and remove return instructions.
2290 DebugLoc Loc;
2291 for (unsigned i = 0, e = Returns.size(); i != e; ++i) {
2292 ReturnInst *RI = Returns[i];
2293 BranchInst* BI = BranchInst::Create(AfterCallBB, RI);
2294 Loc = RI->getDebugLoc();
2295 BI->setDebugLoc(Loc);
2296 RI->eraseFromParent();
2297 }
2298 // We need to set the debug location to *somewhere* inside the
2299 // inlined function. The line number may be nonsensical, but the
2300 // instruction will at least be associated with the right
2301 // function.
2302 if (CreatedBranchToNormalDest)
2303 CreatedBranchToNormalDest->setDebugLoc(Loc);
2304 } else if (!Returns.empty()) {
2305 // Otherwise, if there is exactly one return value, just replace anything
2306 // using the return value of the call with the computed value.
2307 if (!TheCall->use_empty()) {
2308 if (TheCall == Returns[0]->getReturnValue())
2309 TheCall->replaceAllUsesWith(UndefValue::get(TheCall->getType()));
2310 else
2311 TheCall->replaceAllUsesWith(Returns[0]->getReturnValue());
2312 }
2313
2314 // Update PHI nodes that use the ReturnBB to use the AfterCallBB.
2315 BasicBlock *ReturnBB = Returns[0]->getParent();
2316 ReturnBB->replaceAllUsesWith(AfterCallBB);
2317
2318 // Splice the code from the return block into the block that it will return
2319 // to, which contains the code that was after the call.
2320 AfterCallBB->getInstList().splice(AfterCallBB->begin(),
2321 ReturnBB->getInstList());
2322
2323 if (CreatedBranchToNormalDest)
2324 CreatedBranchToNormalDest->setDebugLoc(Returns[0]->getDebugLoc());
2325
2326 // Delete the return instruction now and empty ReturnBB now.
2327 Returns[0]->eraseFromParent();
2328 ReturnBB->eraseFromParent();
2329 } else if (!TheCall->use_empty()) {
2330 // No returns, but something is using the return value of the call. Just
2331 // nuke the result.
2332 TheCall->replaceAllUsesWith(UndefValue::get(TheCall->getType()));
2333 }
2334
2335 // Since we are now done with the Call/Invoke, we can delete it.
2336 TheCall->eraseFromParent();
2337
2338 // If we inlined any musttail calls and the original return is now
2339 // unreachable, delete it. It can only contain a bitcast and ret.
2340 if (InlinedMustTailCalls && pred_begin(AfterCallBB) == pred_end(AfterCallBB))
2341 AfterCallBB->eraseFromParent();
2342
2343 // We should always be able to fold the entry block of the function into the
2344 // single predecessor of the block...
2345 assert(cast<BranchInst>(Br)->isUnconditional() && "splitBasicBlock broken!")((cast<BranchInst>(Br)->isUnconditional() &&
"splitBasicBlock broken!") ? static_cast<void> (0) : __assert_fail
("cast<BranchInst>(Br)->isUnconditional() && \"splitBasicBlock broken!\""
, "/build/llvm-toolchain-snapshot-8~svn345461/lib/Transforms/Utils/InlineFunction.cpp"
, 2345, __PRETTY_FUNCTION__))
;
2346 BasicBlock *CalleeEntry = cast<BranchInst>(Br)->getSuccessor(0);
2347
2348 // Splice the code entry block into calling block, right before the
2349 // unconditional branch.
2350 CalleeEntry->replaceAllUsesWith(OrigBB); // Update PHI nodes
2351 OrigBB->getInstList().splice(Br->getIterator(), CalleeEntry->getInstList());
2352
2353 // Remove the unconditional branch.
2354 OrigBB->getInstList().erase(Br);
2355
2356 // Now we can remove the CalleeEntry block, which is now empty.
2357 Caller->getBasicBlockList().erase(CalleeEntry);
2358
2359 // If we inserted a phi node, check to see if it has a single value (e.g. all
2360 // the entries are the same or undef). If so, remove the PHI so it doesn't
2361 // block other optimizations.
2362 if (PHI) {
2363 AssumptionCache *AC =
2364 IFI.GetAssumptionCache ? &(*IFI.GetAssumptionCache)(*Caller) : nullptr;
2365 auto &DL = Caller->getParent()->getDataLayout();
2366 if (Value *V = SimplifyInstruction(PHI, {DL, nullptr, nullptr, AC})) {
2367 PHI->replaceAllUsesWith(V);
2368 PHI->eraseFromParent();
2369 }
2370 }
2371
2372 return true;
2373}