[llvm-branch-commits] [cfe-branch] r318225 - Merging r310905 and r310994:
Tom Stellard via llvm-branch-commits
llvm-branch-commits at lists.llvm.org
Tue Nov 14 15:53:22 PST 2017
Author: tstellar
Date: Tue Nov 14 15:53:22 2017
New Revision: 318225
URL: http://llvm.org/viewvc/llvm-project?rev=318225&view=rev
Log:
Merging r310905 and r310994:
------------------------------------------------------------------------
r310905 | rnk | 2017-08-14 18:17:47 -0700 (Mon, 14 Aug 2017) | 11 lines
Avoid PointerIntPair of constexpr EvalInfo structs
They are stack allocated, so their alignment is not to be trusted.
32-bit MSVC only guarantees 4 byte stack alignment, even though alignof
would tell you otherwise. I tried fixing this with __declspec align, but
that apparently upsets GCC. Hopefully this version will satisfy all
compilers.
See PR32018 for some info about the mingw issues.
Should supercede https://reviews.llvm.org/D34873
------------------------------------------------------------------------
------------------------------------------------------------------------
r310994 | chandlerc | 2017-08-16 00:22:49 -0700 (Wed, 16 Aug 2017) | 6 lines
Fix a UBSan failure where this boolean was copied when uninitialized.
When r310905 moved the pointer and bool out of a PointerIntPair, it made
them end up uninitialized and caused UBSan failures when copying the
uninitialized boolean. However, making the pointer be null should avoid
the reference to the boolean entirely.
------------------------------------------------------------------------
Modified:
cfe/branches/release_50/lib/AST/ExprConstant.cpp
Modified: cfe/branches/release_50/lib/AST/ExprConstant.cpp
URL: http://llvm.org/viewvc/llvm-project/cfe/branches/release_50/lib/AST/ExprConstant.cpp?rev=318225&r1=318224&r2=318225&view=diff
==============================================================================
--- cfe/branches/release_50/lib/AST/ExprConstant.cpp (original)
+++ cfe/branches/release_50/lib/AST/ExprConstant.cpp Tue Nov 14 15:53:22 2017
@@ -537,7 +537,7 @@ namespace {
/// rules. For example, the RHS of (0 && foo()) is not evaluated. We can
/// evaluate the expression regardless of what the RHS is, but C only allows
/// certain things in certain situations.
- struct LLVM_ALIGNAS(/*alignof(uint64_t)*/ 8) EvalInfo {
+ struct EvalInfo {
ASTContext &Ctx;
/// EvalStatus - Contains information about the evaluation.
@@ -977,24 +977,22 @@ namespace {
/// RAII object used to optionally suppress diagnostics and side-effects from
/// a speculative evaluation.
class SpeculativeEvaluationRAII {
- /// Pair of EvalInfo, and a bit that stores whether or not we were
- /// speculatively evaluating when we created this RAII.
- llvm::PointerIntPair<EvalInfo *, 1, bool> InfoAndOldSpecEval;
- Expr::EvalStatus Old;
+ EvalInfo *Info = nullptr;
+ Expr::EvalStatus OldStatus;
+ bool OldIsSpeculativelyEvaluating;
void moveFromAndCancel(SpeculativeEvaluationRAII &&Other) {
- InfoAndOldSpecEval = Other.InfoAndOldSpecEval;
- Old = Other.Old;
- Other.InfoAndOldSpecEval.setPointer(nullptr);
+ Info = Other.Info;
+ OldStatus = Other.OldStatus;
+ Other.Info = nullptr;
}
void maybeRestoreState() {
- EvalInfo *Info = InfoAndOldSpecEval.getPointer();
if (!Info)
return;
- Info->EvalStatus = Old;
- Info->IsSpeculativelyEvaluating = InfoAndOldSpecEval.getInt();
+ Info->EvalStatus = OldStatus;
+ Info->IsSpeculativelyEvaluating = OldIsSpeculativelyEvaluating;
}
public:
@@ -1002,8 +1000,8 @@ namespace {
SpeculativeEvaluationRAII(
EvalInfo &Info, SmallVectorImpl<PartialDiagnosticAt> *NewDiag = nullptr)
- : InfoAndOldSpecEval(&Info, Info.IsSpeculativelyEvaluating),
- Old(Info.EvalStatus) {
+ : Info(&Info), OldStatus(Info.EvalStatus),
+ OldIsSpeculativelyEvaluating(Info.IsSpeculativelyEvaluating) {
Info.EvalStatus.Diag = NewDiag;
Info.IsSpeculativelyEvaluating = true;
}
More information about the llvm-branch-commits
mailing list