llvm-project/clang/lib/StaticAnalyzer/Core/MemRegion.cpp

1495 lines
47 KiB
C++
Raw Normal View History

//== MemRegion.cpp - Abstract memory regions for static analysis --*- C++ -*--//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file defines MemRegion and its subclasses. MemRegion defines a
// partially-typed abstraction of memory useful for path-sensitive dataflow
// analyses.
//
//===----------------------------------------------------------------------===//
#include "clang/StaticAnalyzer/Core/PathSensitive/MemRegion.h"
#include "clang/AST/Attr.h"
#include "clang/AST/CharUnits.h"
#include "clang/AST/DeclObjC.h"
#include "clang/AST/RecordLayout.h"
#include "clang/Analysis/AnalysisContext.h"
#include "clang/Analysis/Support/BumpVector.h"
#include "clang/Basic/SourceManager.h"
#include "clang/StaticAnalyzer/Core/PathSensitive/SValBuilder.h"
#include "llvm/Support/raw_ostream.h"
using namespace clang;
using namespace ento;
//===----------------------------------------------------------------------===//
// MemRegion Construction.
//===----------------------------------------------------------------------===//
template<typename RegionTy> struct MemRegionManagerTrait;
template <typename RegionTy, typename A1>
RegionTy* MemRegionManager::getRegion(const A1 a1) {
const typename MemRegionManagerTrait<RegionTy>::SuperRegionTy *superRegion =
MemRegionManagerTrait<RegionTy>::getSuperRegion(*this, a1);
llvm::FoldingSetNodeID ID;
RegionTy::ProfileRegion(ID, a1, superRegion);
void *InsertPos;
RegionTy* R = cast_or_null<RegionTy>(Regions.FindNodeOrInsertPos(ID,
InsertPos));
if (!R) {
R = (RegionTy*) A.Allocate<RegionTy>();
new (R) RegionTy(a1, superRegion);
Regions.InsertNode(R, InsertPos);
}
return R;
}
template <typename RegionTy, typename A1>
RegionTy* MemRegionManager::getSubRegion(const A1 a1,
const MemRegion *superRegion) {
llvm::FoldingSetNodeID ID;
RegionTy::ProfileRegion(ID, a1, superRegion);
void *InsertPos;
RegionTy* R = cast_or_null<RegionTy>(Regions.FindNodeOrInsertPos(ID,
InsertPos));
if (!R) {
R = (RegionTy*) A.Allocate<RegionTy>();
new (R) RegionTy(a1, superRegion);
Regions.InsertNode(R, InsertPos);
}
return R;
}
template <typename RegionTy, typename A1, typename A2>
RegionTy* MemRegionManager::getRegion(const A1 a1, const A2 a2) {
const typename MemRegionManagerTrait<RegionTy>::SuperRegionTy *superRegion =
MemRegionManagerTrait<RegionTy>::getSuperRegion(*this, a1, a2);
llvm::FoldingSetNodeID ID;
RegionTy::ProfileRegion(ID, a1, a2, superRegion);
void *InsertPos;
RegionTy* R = cast_or_null<RegionTy>(Regions.FindNodeOrInsertPos(ID,
InsertPos));
if (!R) {
R = (RegionTy*) A.Allocate<RegionTy>();
new (R) RegionTy(a1, a2, superRegion);
Regions.InsertNode(R, InsertPos);
}
return R;
}
template <typename RegionTy, typename A1, typename A2>
RegionTy* MemRegionManager::getSubRegion(const A1 a1, const A2 a2,
const MemRegion *superRegion) {
llvm::FoldingSetNodeID ID;
RegionTy::ProfileRegion(ID, a1, a2, superRegion);
void *InsertPos;
RegionTy* R = cast_or_null<RegionTy>(Regions.FindNodeOrInsertPos(ID,
InsertPos));
if (!R) {
R = (RegionTy*) A.Allocate<RegionTy>();
new (R) RegionTy(a1, a2, superRegion);
Regions.InsertNode(R, InsertPos);
}
return R;
}
template <typename RegionTy, typename A1, typename A2, typename A3>
RegionTy* MemRegionManager::getSubRegion(const A1 a1, const A2 a2, const A3 a3,
const MemRegion *superRegion) {
llvm::FoldingSetNodeID ID;
RegionTy::ProfileRegion(ID, a1, a2, a3, superRegion);
void *InsertPos;
RegionTy* R = cast_or_null<RegionTy>(Regions.FindNodeOrInsertPos(ID,
InsertPos));
if (!R) {
R = (RegionTy*) A.Allocate<RegionTy>();
new (R) RegionTy(a1, a2, a3, superRegion);
Regions.InsertNode(R, InsertPos);
}
return R;
}
//===----------------------------------------------------------------------===//
// Object destruction.
//===----------------------------------------------------------------------===//
MemRegion::~MemRegion() {}
MemRegionManager::~MemRegionManager() {
// All regions and their data are BumpPtrAllocated. No need to call
// their destructors.
}
//===----------------------------------------------------------------------===//
// Basic methods.
//===----------------------------------------------------------------------===//
bool SubRegion::isSubRegionOf(const MemRegion* R) const {
const MemRegion* r = getSuperRegion();
while (r != 0) {
if (r == R)
return true;
if (const SubRegion* sr = dyn_cast<SubRegion>(r))
r = sr->getSuperRegion();
else
break;
}
return false;
}
MemRegionManager* SubRegion::getMemRegionManager() const {
const SubRegion* r = this;
do {
const MemRegion *superRegion = r->getSuperRegion();
if (const SubRegion *sr = dyn_cast<SubRegion>(superRegion)) {
r = sr;
continue;
}
return superRegion->getMemRegionManager();
} while (1);
}
const StackFrameContext *VarRegion::getStackFrame() const {
const StackSpaceRegion *SSR = dyn_cast<StackSpaceRegion>(getMemorySpace());
return SSR ? SSR->getStackFrame() : NULL;
}
//===----------------------------------------------------------------------===//
// Region extents.
//===----------------------------------------------------------------------===//
DefinedOrUnknownSVal TypedValueRegion::getExtent(SValBuilder &svalBuilder) const {
ASTContext &Ctx = svalBuilder.getContext();
QualType T = getDesugaredValueType(Ctx);
if (isa<VariableArrayType>(T))
return nonloc::SymbolVal(svalBuilder.getSymbolManager().getExtentSymbol(this));
if (T->isIncompleteType())
return UnknownVal();
CharUnits size = Ctx.getTypeSizeInChars(T);
QualType sizeTy = svalBuilder.getArrayIndexType();
return svalBuilder.makeIntVal(size.getQuantity(), sizeTy);
}
DefinedOrUnknownSVal FieldRegion::getExtent(SValBuilder &svalBuilder) const {
// Force callers to deal with bitfields explicitly.
if (getDecl()->isBitField())
return UnknownVal();
DefinedOrUnknownSVal Extent = DeclRegion::getExtent(svalBuilder);
// A zero-length array at the end of a struct often stands for dynamically-
// allocated extra memory.
if (Extent.isZeroConstant()) {
QualType T = getDesugaredValueType(svalBuilder.getContext());
if (isa<ConstantArrayType>(T))
return UnknownVal();
}
return Extent;
}
DefinedOrUnknownSVal AllocaRegion::getExtent(SValBuilder &svalBuilder) const {
return nonloc::SymbolVal(svalBuilder.getSymbolManager().getExtentSymbol(this));
}
DefinedOrUnknownSVal SymbolicRegion::getExtent(SValBuilder &svalBuilder) const {
return nonloc::SymbolVal(svalBuilder.getSymbolManager().getExtentSymbol(this));
}
DefinedOrUnknownSVal StringRegion::getExtent(SValBuilder &svalBuilder) const {
return svalBuilder.makeIntVal(getStringLiteral()->getByteLength()+1,
svalBuilder.getArrayIndexType());
}
ObjCIvarRegion::ObjCIvarRegion(const ObjCIvarDecl *ivd, const MemRegion* sReg)
: DeclRegion(ivd, sReg, ObjCIvarRegionKind) {}
const ObjCIvarDecl *ObjCIvarRegion::getDecl() const {
return cast<ObjCIvarDecl>(D);
}
QualType ObjCIvarRegion::getValueType() const {
return getDecl()->getType();
}
QualType CXXBaseObjectRegion::getValueType() const {
return QualType(getDecl()->getTypeForDecl(), 0);
}
//===----------------------------------------------------------------------===//
// FoldingSet profiling.
//===----------------------------------------------------------------------===//
void MemSpaceRegion::Profile(llvm::FoldingSetNodeID& ID) const {
ID.AddInteger((unsigned)getKind());
}
void StackSpaceRegion::Profile(llvm::FoldingSetNodeID &ID) const {
ID.AddInteger((unsigned)getKind());
ID.AddPointer(getStackFrame());
}
void StaticGlobalSpaceRegion::Profile(llvm::FoldingSetNodeID &ID) const {
ID.AddInteger((unsigned)getKind());
ID.AddPointer(getCodeRegion());
}
void StringRegion::ProfileRegion(llvm::FoldingSetNodeID& ID,
const StringLiteral* Str,
const MemRegion* superRegion) {
ID.AddInteger((unsigned) StringRegionKind);
ID.AddPointer(Str);
ID.AddPointer(superRegion);
}
void ObjCStringRegion::ProfileRegion(llvm::FoldingSetNodeID& ID,
const ObjCStringLiteral* Str,
const MemRegion* superRegion) {
ID.AddInteger((unsigned) ObjCStringRegionKind);
ID.AddPointer(Str);
ID.AddPointer(superRegion);
}
void AllocaRegion::ProfileRegion(llvm::FoldingSetNodeID& ID,
const Expr *Ex, unsigned cnt,
const MemRegion *superRegion) {
ID.AddInteger((unsigned) AllocaRegionKind);
ID.AddPointer(Ex);
ID.AddInteger(cnt);
ID.AddPointer(superRegion);
}
void AllocaRegion::Profile(llvm::FoldingSetNodeID& ID) const {
ProfileRegion(ID, Ex, Cnt, superRegion);
}
void CompoundLiteralRegion::Profile(llvm::FoldingSetNodeID& ID) const {
CompoundLiteralRegion::ProfileRegion(ID, CL, superRegion);
}
void CompoundLiteralRegion::ProfileRegion(llvm::FoldingSetNodeID& ID,
const CompoundLiteralExpr *CL,
const MemRegion* superRegion) {
ID.AddInteger((unsigned) CompoundLiteralRegionKind);
ID.AddPointer(CL);
ID.AddPointer(superRegion);
}
void CXXThisRegion::ProfileRegion(llvm::FoldingSetNodeID &ID,
const PointerType *PT,
const MemRegion *sRegion) {
ID.AddInteger((unsigned) CXXThisRegionKind);
ID.AddPointer(PT);
ID.AddPointer(sRegion);
}
void CXXThisRegion::Profile(llvm::FoldingSetNodeID &ID) const {
CXXThisRegion::ProfileRegion(ID, ThisPointerTy, superRegion);
}
void ObjCIvarRegion::ProfileRegion(llvm::FoldingSetNodeID& ID,
const ObjCIvarDecl *ivd,
const MemRegion* superRegion) {
DeclRegion::ProfileRegion(ID, ivd, superRegion, ObjCIvarRegionKind);
}
void DeclRegion::ProfileRegion(llvm::FoldingSetNodeID& ID, const Decl *D,
const MemRegion* superRegion, Kind k) {
ID.AddInteger((unsigned) k);
ID.AddPointer(D);
ID.AddPointer(superRegion);
}
void DeclRegion::Profile(llvm::FoldingSetNodeID& ID) const {
DeclRegion::ProfileRegion(ID, D, superRegion, getKind());
}
void VarRegion::Profile(llvm::FoldingSetNodeID &ID) const {
VarRegion::ProfileRegion(ID, getDecl(), superRegion);
}
void SymbolicRegion::ProfileRegion(llvm::FoldingSetNodeID& ID, SymbolRef sym,
const MemRegion *sreg) {
ID.AddInteger((unsigned) MemRegion::SymbolicRegionKind);
ID.Add(sym);
ID.AddPointer(sreg);
}
void SymbolicRegion::Profile(llvm::FoldingSetNodeID& ID) const {
SymbolicRegion::ProfileRegion(ID, sym, getSuperRegion());
}
void ElementRegion::ProfileRegion(llvm::FoldingSetNodeID& ID,
QualType ElementType, SVal Idx,
const MemRegion* superRegion) {
ID.AddInteger(MemRegion::ElementRegionKind);
ID.Add(ElementType);
ID.AddPointer(superRegion);
Idx.Profile(ID);
}
void ElementRegion::Profile(llvm::FoldingSetNodeID& ID) const {
ElementRegion::ProfileRegion(ID, ElementType, Index, superRegion);
}
void FunctionTextRegion::ProfileRegion(llvm::FoldingSetNodeID& ID,
const NamedDecl *FD,
const MemRegion*) {
ID.AddInteger(MemRegion::FunctionTextRegionKind);
ID.AddPointer(FD);
}
void FunctionTextRegion::Profile(llvm::FoldingSetNodeID& ID) const {
FunctionTextRegion::ProfileRegion(ID, FD, superRegion);
}
void BlockTextRegion::ProfileRegion(llvm::FoldingSetNodeID& ID,
const BlockDecl *BD, CanQualType,
const AnalysisDeclContext *AC,
const MemRegion*) {
ID.AddInteger(MemRegion::BlockTextRegionKind);
ID.AddPointer(BD);
}
void BlockTextRegion::Profile(llvm::FoldingSetNodeID& ID) const {
BlockTextRegion::ProfileRegion(ID, BD, locTy, AC, superRegion);
}
void BlockDataRegion::ProfileRegion(llvm::FoldingSetNodeID& ID,
const BlockTextRegion *BC,
const LocationContext *LC,
const MemRegion *sReg) {
ID.AddInteger(MemRegion::BlockDataRegionKind);
ID.AddPointer(BC);
ID.AddPointer(LC);
ID.AddPointer(sReg);
}
void BlockDataRegion::Profile(llvm::FoldingSetNodeID& ID) const {
BlockDataRegion::ProfileRegion(ID, BC, LC, getSuperRegion());
}
void CXXTempObjectRegion::ProfileRegion(llvm::FoldingSetNodeID &ID,
Expr const *Ex,
const MemRegion *sReg) {
ID.AddPointer(Ex);
ID.AddPointer(sReg);
}
void CXXTempObjectRegion::Profile(llvm::FoldingSetNodeID &ID) const {
ProfileRegion(ID, Ex, getSuperRegion());
}
void CXXBaseObjectRegion::ProfileRegion(llvm::FoldingSetNodeID &ID,
const CXXRecordDecl *RD,
bool IsVirtual,
const MemRegion *SReg) {
ID.AddPointer(RD);
ID.AddBoolean(IsVirtual);
ID.AddPointer(SReg);
}
void CXXBaseObjectRegion::Profile(llvm::FoldingSetNodeID &ID) const {
ProfileRegion(ID, getDecl(), isVirtual(), superRegion);
}
//===----------------------------------------------------------------------===//
// Region anchors.
//===----------------------------------------------------------------------===//
void GlobalsSpaceRegion::anchor() { }
void HeapSpaceRegion::anchor() { }
void UnknownSpaceRegion::anchor() { }
void StackLocalsSpaceRegion::anchor() { }
void StackArgumentsSpaceRegion::anchor() { }
void TypedRegion::anchor() { }
void TypedValueRegion::anchor() { }
void CodeTextRegion::anchor() { }
void SubRegion::anchor() { }
//===----------------------------------------------------------------------===//
// Region pretty-printing.
//===----------------------------------------------------------------------===//
void MemRegion::dump() const {
dumpToStream(llvm::errs());
}
std::string MemRegion::getString() const {
std::string s;
llvm::raw_string_ostream os(s);
dumpToStream(os);
return os.str();
}
void MemRegion::dumpToStream(raw_ostream &os) const {
os << "<Unknown Region>";
}
void AllocaRegion::dumpToStream(raw_ostream &os) const {
os << "alloca{" << (const void*) Ex << ',' << Cnt << '}';
}
void FunctionTextRegion::dumpToStream(raw_ostream &os) const {
os << "code{" << getDecl()->getDeclName().getAsString() << '}';
}
void BlockTextRegion::dumpToStream(raw_ostream &os) const {
os << "block_code{" << (const void*) this << '}';
}
void BlockDataRegion::dumpToStream(raw_ostream &os) const {
os << "block_data{" << BC << '}';
}
void CompoundLiteralRegion::dumpToStream(raw_ostream &os) const {
// FIXME: More elaborate pretty-printing.
os << "{ " << (const void*) CL << " }";
}
void CXXTempObjectRegion::dumpToStream(raw_ostream &os) const {
os << "temp_object{" << getValueType().getAsString() << ','
<< (const void*) Ex << '}';
2010-11-25 10:07:24 +08:00
}
void CXXBaseObjectRegion::dumpToStream(raw_ostream &os) const {
os << "base{" << superRegion << ',' << getDecl()->getName() << '}';
}
void CXXThisRegion::dumpToStream(raw_ostream &os) const {
os << "this";
}
void ElementRegion::dumpToStream(raw_ostream &os) const {
This is a fairly large patch, which resulted from a cascade of changes made to RegionStore (and related classes) in order to handle some analyzer failures involving casts and manipulation of symbolic memory. The root of the change is in StoreManager::CastRegion(). Instead of using ad hoc heuristics to decide when to layer an ElementRegion on a casted MemRegion, we now always layer an ElementRegion when the cast type is different than the original type of the region. This carries the current cast information associated with a region around without resorting to the error prone recording of "casted types" in GRState. Along with this new policy of layering ElementRegions, I added a new algorithm to strip away existing ElementRegions when they simply represented casts of a base memory object. This algorithm computes the raw "byte offset" that an ElementRegion represents from the base region, and allows the new ElementRegion to be based off that offset. The added benefit is that this naturally handles a series of casts of a MemRegion without building up a set of redundant ElementRegions (thus canonicalizing the region view). Other related changes that cascaded from this one (as tests were failing in RegionStore): - Revamped RegionStoreManager::InvalidateRegion() to completely remove all bindings and default values from a region and all subregions. Now invalidated fields are not bound directly to new symbolic values; instead the base region has a "default" symbol value from which "derived symbols" can be created. The main advantage of this approach is that it allows us to invalidate a region hierarchy and then lazily instantiate new values no matter how deep the hierarchy went (i.e., regardless of the number of field accesses, e.g. x->f->y->z->...). The previous approach did not do this. - Slightly reworked RegionStoreManager::RemoveDeadBindings() to also incorporate live symbols and live regions that do not have direct bindings but also have "default values" used for lazy instantiation. The changes to 'InvalidateRegion' revealed that these were necessary in order to achieve lazy instantiation of values in the region store with those bindings being removed too early. - The changes to InvalidateRegion() and RemoveDeadBindings() revealed a serious bug in 'getSubRegionMap()' where not all region -> subregion relationships involved in actually bindings (explicit and implicit) were being recorded. This has been fixed by using a worklist algorithm to iteratively fill in the region map. - Added special support to RegionStoreManager::Bind()/Retrieve() to handle OSAtomicCompareAndSwap in light of the new 'CastRegion' changes and the layering of ElementRegions. - Fixed a bug in SymbolReaper::isLive() where derived symbols were not being marked live if the symbol they were derived from was also live. This fix was critical for getting lazy instantiation in RegionStore to work. - Tidied up the implementation of ValueManager::getXXXSymbolVal() methods to use SymbolManager::canSymbolicate() to decide whether or not a symbol should be symbolicated. - 'test/Analysis/misc-ps-xfail.m' now passes; that test case has been moved to 'test/Analysis/misc-ps.m'. - Tweaked some pretty-printing of MemRegions, and implemented 'ElementRegion::getRawOffset()' for use with the CastRegion changes. llvm-svn: 77782
2009-08-01 14:17:29 +08:00
os << "element{" << superRegion << ','
<< Index << ',' << getElementType().getAsString() << '}';
}
void FieldRegion::dumpToStream(raw_ostream &os) const {
os << superRegion << "->" << *getDecl();
}
void ObjCIvarRegion::dumpToStream(raw_ostream &os) const {
os << "ivar{" << superRegion << ',' << *getDecl() << '}';
}
void StringRegion::dumpToStream(raw_ostream &os) const {
Str->printPretty(os, 0, PrintingPolicy(getContext().getLangOpts()));
}
void ObjCStringRegion::dumpToStream(raw_ostream &os) const {
Str->printPretty(os, 0, PrintingPolicy(getContext().getLangOpts()));
}
void SymbolicRegion::dumpToStream(raw_ostream &os) const {
os << "SymRegion{" << sym << '}';
}
void VarRegion::dumpToStream(raw_ostream &os) const {
os << *cast<VarDecl>(D);
}
This is a fairly large patch, which resulted from a cascade of changes made to RegionStore (and related classes) in order to handle some analyzer failures involving casts and manipulation of symbolic memory. The root of the change is in StoreManager::CastRegion(). Instead of using ad hoc heuristics to decide when to layer an ElementRegion on a casted MemRegion, we now always layer an ElementRegion when the cast type is different than the original type of the region. This carries the current cast information associated with a region around without resorting to the error prone recording of "casted types" in GRState. Along with this new policy of layering ElementRegions, I added a new algorithm to strip away existing ElementRegions when they simply represented casts of a base memory object. This algorithm computes the raw "byte offset" that an ElementRegion represents from the base region, and allows the new ElementRegion to be based off that offset. The added benefit is that this naturally handles a series of casts of a MemRegion without building up a set of redundant ElementRegions (thus canonicalizing the region view). Other related changes that cascaded from this one (as tests were failing in RegionStore): - Revamped RegionStoreManager::InvalidateRegion() to completely remove all bindings and default values from a region and all subregions. Now invalidated fields are not bound directly to new symbolic values; instead the base region has a "default" symbol value from which "derived symbols" can be created. The main advantage of this approach is that it allows us to invalidate a region hierarchy and then lazily instantiate new values no matter how deep the hierarchy went (i.e., regardless of the number of field accesses, e.g. x->f->y->z->...). The previous approach did not do this. - Slightly reworked RegionStoreManager::RemoveDeadBindings() to also incorporate live symbols and live regions that do not have direct bindings but also have "default values" used for lazy instantiation. The changes to 'InvalidateRegion' revealed that these were necessary in order to achieve lazy instantiation of values in the region store with those bindings being removed too early. - The changes to InvalidateRegion() and RemoveDeadBindings() revealed a serious bug in 'getSubRegionMap()' where not all region -> subregion relationships involved in actually bindings (explicit and implicit) were being recorded. This has been fixed by using a worklist algorithm to iteratively fill in the region map. - Added special support to RegionStoreManager::Bind()/Retrieve() to handle OSAtomicCompareAndSwap in light of the new 'CastRegion' changes and the layering of ElementRegions. - Fixed a bug in SymbolReaper::isLive() where derived symbols were not being marked live if the symbol they were derived from was also live. This fix was critical for getting lazy instantiation in RegionStore to work. - Tidied up the implementation of ValueManager::getXXXSymbolVal() methods to use SymbolManager::canSymbolicate() to decide whether or not a symbol should be symbolicated. - 'test/Analysis/misc-ps-xfail.m' now passes; that test case has been moved to 'test/Analysis/misc-ps.m'. - Tweaked some pretty-printing of MemRegions, and implemented 'ElementRegion::getRawOffset()' for use with the CastRegion changes. llvm-svn: 77782
2009-08-01 14:17:29 +08:00
void RegionRawOffset::dump() const {
dumpToStream(llvm::errs());
}
void RegionRawOffset::dumpToStream(raw_ostream &os) const {
os << "raw_offset{" << getRegion() << ',' << getOffset().getQuantity() << '}';
This is a fairly large patch, which resulted from a cascade of changes made to RegionStore (and related classes) in order to handle some analyzer failures involving casts and manipulation of symbolic memory. The root of the change is in StoreManager::CastRegion(). Instead of using ad hoc heuristics to decide when to layer an ElementRegion on a casted MemRegion, we now always layer an ElementRegion when the cast type is different than the original type of the region. This carries the current cast information associated with a region around without resorting to the error prone recording of "casted types" in GRState. Along with this new policy of layering ElementRegions, I added a new algorithm to strip away existing ElementRegions when they simply represented casts of a base memory object. This algorithm computes the raw "byte offset" that an ElementRegion represents from the base region, and allows the new ElementRegion to be based off that offset. The added benefit is that this naturally handles a series of casts of a MemRegion without building up a set of redundant ElementRegions (thus canonicalizing the region view). Other related changes that cascaded from this one (as tests were failing in RegionStore): - Revamped RegionStoreManager::InvalidateRegion() to completely remove all bindings and default values from a region and all subregions. Now invalidated fields are not bound directly to new symbolic values; instead the base region has a "default" symbol value from which "derived symbols" can be created. The main advantage of this approach is that it allows us to invalidate a region hierarchy and then lazily instantiate new values no matter how deep the hierarchy went (i.e., regardless of the number of field accesses, e.g. x->f->y->z->...). The previous approach did not do this. - Slightly reworked RegionStoreManager::RemoveDeadBindings() to also incorporate live symbols and live regions that do not have direct bindings but also have "default values" used for lazy instantiation. The changes to 'InvalidateRegion' revealed that these were necessary in order to achieve lazy instantiation of values in the region store with those bindings being removed too early. - The changes to InvalidateRegion() and RemoveDeadBindings() revealed a serious bug in 'getSubRegionMap()' where not all region -> subregion relationships involved in actually bindings (explicit and implicit) were being recorded. This has been fixed by using a worklist algorithm to iteratively fill in the region map. - Added special support to RegionStoreManager::Bind()/Retrieve() to handle OSAtomicCompareAndSwap in light of the new 'CastRegion' changes and the layering of ElementRegions. - Fixed a bug in SymbolReaper::isLive() where derived symbols were not being marked live if the symbol they were derived from was also live. This fix was critical for getting lazy instantiation in RegionStore to work. - Tidied up the implementation of ValueManager::getXXXSymbolVal() methods to use SymbolManager::canSymbolicate() to decide whether or not a symbol should be symbolicated. - 'test/Analysis/misc-ps-xfail.m' now passes; that test case has been moved to 'test/Analysis/misc-ps.m'. - Tweaked some pretty-printing of MemRegions, and implemented 'ElementRegion::getRawOffset()' for use with the CastRegion changes. llvm-svn: 77782
2009-08-01 14:17:29 +08:00
}
void StaticGlobalSpaceRegion::dumpToStream(raw_ostream &os) const {
os << "StaticGlobalsMemSpace{" << CR << '}';
}
void GlobalInternalSpaceRegion::dumpToStream(raw_ostream &os) const {
os << "GlobalInternalSpaceRegion";
}
void GlobalSystemSpaceRegion::dumpToStream(raw_ostream &os) const {
os << "GlobalSystemSpaceRegion";
}
void GlobalImmutableSpaceRegion::dumpToStream(raw_ostream &os) const {
os << "GlobalImmutableSpaceRegion";
}
void HeapSpaceRegion::dumpToStream(raw_ostream &os) const {
os << "HeapSpaceRegion";
}
void UnknownSpaceRegion::dumpToStream(raw_ostream &os) const {
os << "UnknownSpaceRegion";
}
void StackArgumentsSpaceRegion::dumpToStream(raw_ostream &os) const {
os << "StackArgumentsSpaceRegion";
}
void StackLocalsSpaceRegion::dumpToStream(raw_ostream &os) const {
os << "StackLocalsSpaceRegion";
}
bool MemRegion::canPrintPretty() const {
return canPrintPrettyAsExpr();
}
bool MemRegion::canPrintPrettyAsExpr() const {
return false;
}
void MemRegion::printPretty(raw_ostream &os) const {
assert(canPrintPretty() && "This region cannot be printed pretty.");
os << "'";
printPrettyAsExpr(os);
os << "'";
return;
}
void MemRegion::printPrettyAsExpr(raw_ostream &os) const {
llvm_unreachable("This region cannot be printed pretty.");
return;
}
bool VarRegion::canPrintPrettyAsExpr() const {
return true;
}
void VarRegion::printPrettyAsExpr(raw_ostream &os) const {
os << getDecl()->getName();
}
bool ObjCIvarRegion::canPrintPrettyAsExpr() const {
return true;
}
void ObjCIvarRegion::printPrettyAsExpr(raw_ostream &os) const {
os << getDecl()->getName();
}
bool FieldRegion::canPrintPretty() const {
return true;
}
bool FieldRegion::canPrintPrettyAsExpr() const {
return superRegion->canPrintPrettyAsExpr();
}
void FieldRegion::printPrettyAsExpr(raw_ostream &os) const {
assert(canPrintPrettyAsExpr());
superRegion->printPrettyAsExpr(os);
os << "." << getDecl()->getName();
}
void FieldRegion::printPretty(raw_ostream &os) const {
if (canPrintPrettyAsExpr()) {
os << "\'";
printPrettyAsExpr(os);
os << "'";
} else {
os << "field " << "\'" << getDecl()->getName() << "'";
}
return;
}
bool CXXBaseObjectRegion::canPrintPrettyAsExpr() const {
return superRegion->canPrintPrettyAsExpr();
}
void CXXBaseObjectRegion::printPrettyAsExpr(raw_ostream &os) const {
superRegion->printPrettyAsExpr(os);
}
//===----------------------------------------------------------------------===//
// MemRegionManager methods.
//===----------------------------------------------------------------------===//
This is a fairly large patch, which resulted from a cascade of changes made to RegionStore (and related classes) in order to handle some analyzer failures involving casts and manipulation of symbolic memory. The root of the change is in StoreManager::CastRegion(). Instead of using ad hoc heuristics to decide when to layer an ElementRegion on a casted MemRegion, we now always layer an ElementRegion when the cast type is different than the original type of the region. This carries the current cast information associated with a region around without resorting to the error prone recording of "casted types" in GRState. Along with this new policy of layering ElementRegions, I added a new algorithm to strip away existing ElementRegions when they simply represented casts of a base memory object. This algorithm computes the raw "byte offset" that an ElementRegion represents from the base region, and allows the new ElementRegion to be based off that offset. The added benefit is that this naturally handles a series of casts of a MemRegion without building up a set of redundant ElementRegions (thus canonicalizing the region view). Other related changes that cascaded from this one (as tests were failing in RegionStore): - Revamped RegionStoreManager::InvalidateRegion() to completely remove all bindings and default values from a region and all subregions. Now invalidated fields are not bound directly to new symbolic values; instead the base region has a "default" symbol value from which "derived symbols" can be created. The main advantage of this approach is that it allows us to invalidate a region hierarchy and then lazily instantiate new values no matter how deep the hierarchy went (i.e., regardless of the number of field accesses, e.g. x->f->y->z->...). The previous approach did not do this. - Slightly reworked RegionStoreManager::RemoveDeadBindings() to also incorporate live symbols and live regions that do not have direct bindings but also have "default values" used for lazy instantiation. The changes to 'InvalidateRegion' revealed that these were necessary in order to achieve lazy instantiation of values in the region store with those bindings being removed too early. - The changes to InvalidateRegion() and RemoveDeadBindings() revealed a serious bug in 'getSubRegionMap()' where not all region -> subregion relationships involved in actually bindings (explicit and implicit) were being recorded. This has been fixed by using a worklist algorithm to iteratively fill in the region map. - Added special support to RegionStoreManager::Bind()/Retrieve() to handle OSAtomicCompareAndSwap in light of the new 'CastRegion' changes and the layering of ElementRegions. - Fixed a bug in SymbolReaper::isLive() where derived symbols were not being marked live if the symbol they were derived from was also live. This fix was critical for getting lazy instantiation in RegionStore to work. - Tidied up the implementation of ValueManager::getXXXSymbolVal() methods to use SymbolManager::canSymbolicate() to decide whether or not a symbol should be symbolicated. - 'test/Analysis/misc-ps-xfail.m' now passes; that test case has been moved to 'test/Analysis/misc-ps.m'. - Tweaked some pretty-printing of MemRegions, and implemented 'ElementRegion::getRawOffset()' for use with the CastRegion changes. llvm-svn: 77782
2009-08-01 14:17:29 +08:00
template <typename REG>
const REG *MemRegionManager::LazyAllocate(REG*& region) {
if (!region) {
region = (REG*) A.Allocate<REG>();
new (region) REG(this);
}
return region;
}
template <typename REG, typename ARG>
const REG *MemRegionManager::LazyAllocate(REG*& region, ARG a) {
if (!region) {
region = (REG*) A.Allocate<REG>();
new (region) REG(this, a);
}
return region;
}
const StackLocalsSpaceRegion*
MemRegionManager::getStackLocalsRegion(const StackFrameContext *STC) {
assert(STC);
StackLocalsSpaceRegion *&R = StackLocalsSpaceRegions[STC];
if (R)
return R;
R = A.Allocate<StackLocalsSpaceRegion>();
new (R) StackLocalsSpaceRegion(this, STC);
return R;
}
const StackArgumentsSpaceRegion *
MemRegionManager::getStackArgumentsRegion(const StackFrameContext *STC) {
assert(STC);
StackArgumentsSpaceRegion *&R = StackArgumentsSpaceRegions[STC];
if (R)
return R;
R = A.Allocate<StackArgumentsSpaceRegion>();
new (R) StackArgumentsSpaceRegion(this, STC);
return R;
}
const GlobalsSpaceRegion
*MemRegionManager::getGlobalsRegion(MemRegion::Kind K,
const CodeTextRegion *CR) {
if (!CR) {
if (K == MemRegion::GlobalSystemSpaceRegionKind)
return LazyAllocate(SystemGlobals);
if (K == MemRegion::GlobalImmutableSpaceRegionKind)
return LazyAllocate(ImmutableGlobals);
assert(K == MemRegion::GlobalInternalSpaceRegionKind);
return LazyAllocate(InternalGlobals);
}
assert(K == MemRegion::StaticGlobalSpaceRegionKind);
StaticGlobalSpaceRegion *&R = StaticsGlobalSpaceRegions[CR];
if (R)
return R;
R = A.Allocate<StaticGlobalSpaceRegion>();
new (R) StaticGlobalSpaceRegion(this, CR);
return R;
}
const HeapSpaceRegion *MemRegionManager::getHeapRegion() {
return LazyAllocate(heap);
}
const MemSpaceRegion *MemRegionManager::getUnknownRegion() {
return LazyAllocate(unknown);
}
const MemSpaceRegion *MemRegionManager::getCodeRegion() {
return LazyAllocate(code);
}
//===----------------------------------------------------------------------===//
// Constructing regions.
//===----------------------------------------------------------------------===//
const StringRegion* MemRegionManager::getStringRegion(const StringLiteral* Str){
return getSubRegion<StringRegion>(Str, getGlobalsRegion());
}
const ObjCStringRegion *
MemRegionManager::getObjCStringRegion(const ObjCStringLiteral* Str){
return getSubRegion<ObjCStringRegion>(Str, getGlobalsRegion());
}
/// Look through a chain of LocationContexts to either find the
/// StackFrameContext that matches a DeclContext, or find a VarRegion
/// for a variable captured by a block.
static llvm::PointerUnion<const StackFrameContext *, const VarRegion *>
getStackOrCaptureRegionForDeclContext(const LocationContext *LC,
const DeclContext *DC,
const VarDecl *VD) {
while (LC) {
if (const StackFrameContext *SFC = dyn_cast<StackFrameContext>(LC)) {
if (cast<DeclContext>(SFC->getDecl()) == DC)
return SFC;
}
if (const BlockInvocationContext *BC =
dyn_cast<BlockInvocationContext>(LC)) {
const BlockDataRegion *BR =
static_cast<const BlockDataRegion*>(BC->getContextData());
// FIXME: This can be made more efficient.
for (BlockDataRegion::referenced_vars_iterator
I = BR->referenced_vars_begin(),
E = BR->referenced_vars_end(); I != E; ++I) {
if (const VarRegion *VR = dyn_cast<VarRegion>(I.getOriginalRegion()))
if (VR->getDecl() == VD)
return cast<VarRegion>(I.getCapturedRegion());
}
}
LC = LC->getParent();
}
return (const StackFrameContext*)0;
}
const VarRegion* MemRegionManager::getVarRegion(const VarDecl *D,
const LocationContext *LC) {
const MemRegion *sReg = 0;
if (D->hasGlobalStorage() && !D->isStaticLocal()) {
// First handle the globals defined in system headers.
if (C.getSourceManager().isInSystemHeader(D->getLocation())) {
// Whitelist the system globals which often DO GET modified, assume the
// rest are immutable.
if (D->getName().find("errno") != StringRef::npos)
sReg = getGlobalsRegion(MemRegion::GlobalSystemSpaceRegionKind);
else
sReg = getGlobalsRegion(MemRegion::GlobalImmutableSpaceRegionKind);
// Treat other globals as GlobalInternal unless they are constants.
} else {
QualType GQT = D->getType();
const Type *GT = GQT.getTypePtrOrNull();
// TODO: We could walk the complex types here and see if everything is
// constified.
if (GT && GQT.isConstQualified() && GT->isArithmeticType())
sReg = getGlobalsRegion(MemRegion::GlobalImmutableSpaceRegionKind);
else
sReg = getGlobalsRegion();
}
// Finally handle static locals.
} else {
// FIXME: Once we implement scope handling, we will need to properly lookup
// 'D' to the proper LocationContext.
const DeclContext *DC = D->getDeclContext();
llvm::PointerUnion<const StackFrameContext *, const VarRegion *> V =
getStackOrCaptureRegionForDeclContext(LC, DC, D);
if (V.is<const VarRegion*>())
return V.get<const VarRegion*>();
const StackFrameContext *STC = V.get<const StackFrameContext*>();
if (!STC)
sReg = getUnknownRegion();
else {
if (D->hasLocalStorage()) {
sReg = isa<ParmVarDecl>(D) || isa<ImplicitParamDecl>(D)
? static_cast<const MemRegion*>(getStackArgumentsRegion(STC))
: static_cast<const MemRegion*>(getStackLocalsRegion(STC));
}
else {
assert(D->isStaticLocal());
const Decl *STCD = STC->getDecl();
if (isa<FunctionDecl>(STCD) || isa<ObjCMethodDecl>(STCD))
sReg = getGlobalsRegion(MemRegion::StaticGlobalSpaceRegionKind,
getFunctionTextRegion(cast<NamedDecl>(STCD)));
else if (const BlockDecl *BD = dyn_cast<BlockDecl>(STCD)) {
// FIXME: The fallback type here is totally bogus -- though it should
// never be queried, it will prevent uniquing with the real
// BlockTextRegion. Ideally we'd fix the AST so that we always had a
// signature.
QualType T;
if (const TypeSourceInfo *TSI = BD->getSignatureAsWritten())
T = TSI->getType();
else
T = getContext().getFunctionNoProtoType(getContext().VoidTy);
const BlockTextRegion *BTR =
getBlockTextRegion(BD, C.getCanonicalType(T),
STC->getAnalysisDeclContext());
sReg = getGlobalsRegion(MemRegion::StaticGlobalSpaceRegionKind,
BTR);
}
else {
sReg = getGlobalsRegion();
}
}
}
}
return getSubRegion<VarRegion>(D, sReg);
}
const VarRegion *MemRegionManager::getVarRegion(const VarDecl *D,
const MemRegion *superR) {
return getSubRegion<VarRegion>(D, superR);
}
const BlockDataRegion *
MemRegionManager::getBlockDataRegion(const BlockTextRegion *BC,
const LocationContext *LC) {
const MemRegion *sReg = 0;
const BlockDecl *BD = BC->getDecl();
if (!BD->hasCaptures()) {
// This handles 'static' blocks.
sReg = getGlobalsRegion(MemRegion::GlobalImmutableSpaceRegionKind);
}
else {
if (LC) {
// FIXME: Once we implement scope handling, we want the parent region
// to be the scope.
const StackFrameContext *STC = LC->getCurrentStackFrame();
assert(STC);
sReg = getStackLocalsRegion(STC);
}
else {
// We allow 'LC' to be NULL for cases where want BlockDataRegions
// without context-sensitivity.
sReg = getUnknownRegion();
}
}
return getSubRegion<BlockDataRegion>(BC, LC, sReg);
}
const CXXTempObjectRegion *
MemRegionManager::getCXXStaticTempObjectRegion(const Expr *Ex) {
return getSubRegion<CXXTempObjectRegion>(
Ex, getGlobalsRegion(MemRegion::GlobalInternalSpaceRegionKind, NULL));
}
const CompoundLiteralRegion*
MemRegionManager::getCompoundLiteralRegion(const CompoundLiteralExpr *CL,
const LocationContext *LC) {
const MemRegion *sReg = 0;
if (CL->isFileScope())
sReg = getGlobalsRegion();
else {
const StackFrameContext *STC = LC->getCurrentStackFrame();
assert(STC);
sReg = getStackLocalsRegion(STC);
}
return getSubRegion<CompoundLiteralRegion>(CL, sReg);
}
const ElementRegion*
MemRegionManager::getElementRegion(QualType elementType, NonLoc Idx,
const MemRegion* superRegion,
ASTContext &Ctx){
QualType T = Ctx.getCanonicalType(elementType).getUnqualifiedType();
llvm::FoldingSetNodeID ID;
ElementRegion::ProfileRegion(ID, T, Idx, superRegion);
void *InsertPos;
MemRegion* data = Regions.FindNodeOrInsertPos(ID, InsertPos);
ElementRegion* R = cast_or_null<ElementRegion>(data);
if (!R) {
R = (ElementRegion*) A.Allocate<ElementRegion>();
new (R) ElementRegion(T, Idx, superRegion);
Regions.InsertNode(R, InsertPos);
}
return R;
}
const FunctionTextRegion *
MemRegionManager::getFunctionTextRegion(const NamedDecl *FD) {
return getSubRegion<FunctionTextRegion>(FD, getCodeRegion());
}
const BlockTextRegion *
MemRegionManager::getBlockTextRegion(const BlockDecl *BD, CanQualType locTy,
AnalysisDeclContext *AC) {
return getSubRegion<BlockTextRegion>(BD, locTy, AC, getCodeRegion());
}
/// getSymbolicRegion - Retrieve or create a "symbolic" memory region.
const SymbolicRegion *MemRegionManager::getSymbolicRegion(SymbolRef sym) {
return getSubRegion<SymbolicRegion>(sym, getUnknownRegion());
}
const SymbolicRegion *MemRegionManager::getSymbolicHeapRegion(SymbolRef Sym) {
return getSubRegion<SymbolicRegion>(Sym, getHeapRegion());
}
const FieldRegion*
MemRegionManager::getFieldRegion(const FieldDecl *d,
const MemRegion* superRegion){
return getSubRegion<FieldRegion>(d, superRegion);
}
const ObjCIvarRegion*
MemRegionManager::getObjCIvarRegion(const ObjCIvarDecl *d,
const MemRegion* superRegion) {
return getSubRegion<ObjCIvarRegion>(d, superRegion);
}
const CXXTempObjectRegion*
MemRegionManager::getCXXTempObjectRegion(Expr const *E,
LocationContext const *LC) {
const StackFrameContext *SFC = LC->getCurrentStackFrame();
assert(SFC);
return getSubRegion<CXXTempObjectRegion>(E, getStackLocalsRegion(SFC));
}
/// Checks whether \p BaseClass is a valid virtual or direct non-virtual base
/// class of the type of \p Super.
static bool isValidBaseClass(const CXXRecordDecl *BaseClass,
const TypedValueRegion *Super,
bool IsVirtual) {
BaseClass = BaseClass->getCanonicalDecl();
const CXXRecordDecl *Class = Super->getValueType()->getAsCXXRecordDecl();
if (!Class)
return true;
if (IsVirtual)
return Class->isVirtuallyDerivedFrom(BaseClass);
for (CXXRecordDecl::base_class_const_iterator I = Class->bases_begin(),
E = Class->bases_end();
I != E; ++I) {
if (I->getType()->getAsCXXRecordDecl()->getCanonicalDecl() == BaseClass)
return true;
}
return false;
}
const CXXBaseObjectRegion *
MemRegionManager::getCXXBaseObjectRegion(const CXXRecordDecl *RD,
const MemRegion *Super,
bool IsVirtual) {
if (isa<TypedValueRegion>(Super)) {
assert(isValidBaseClass(RD, dyn_cast<TypedValueRegion>(Super), IsVirtual));
(void)&isValidBaseClass;
if (IsVirtual) {
// Virtual base regions should not be layered, since the layout rules
// are different.
while (const CXXBaseObjectRegion *Base =
dyn_cast<CXXBaseObjectRegion>(Super)) {
Super = Base->getSuperRegion();
}
assert(Super && !isa<MemSpaceRegion>(Super));
}
}
return getSubRegion<CXXBaseObjectRegion>(RD, IsVirtual, Super);
}
const CXXThisRegion*
MemRegionManager::getCXXThisRegion(QualType thisPointerTy,
const LocationContext *LC) {
const StackFrameContext *STC = LC->getCurrentStackFrame();
assert(STC);
const PointerType *PT = thisPointerTy->getAs<PointerType>();
assert(PT);
return getSubRegion<CXXThisRegion>(PT, getStackArgumentsRegion(STC));
}
const AllocaRegion*
MemRegionManager::getAllocaRegion(const Expr *E, unsigned cnt,
const LocationContext *LC) {
const StackFrameContext *STC = LC->getCurrentStackFrame();
assert(STC);
return getSubRegion<AllocaRegion>(E, cnt, getStackLocalsRegion(STC));
}
const MemSpaceRegion *MemRegion::getMemorySpace() const {
const MemRegion *R = this;
const SubRegion* SR = dyn_cast<SubRegion>(this);
while (SR) {
R = SR->getSuperRegion();
SR = dyn_cast<SubRegion>(R);
}
return dyn_cast<MemSpaceRegion>(R);
}
bool MemRegion::hasStackStorage() const {
return isa<StackSpaceRegion>(getMemorySpace());
}
bool MemRegion::hasStackNonParametersStorage() const {
return isa<StackLocalsSpaceRegion>(getMemorySpace());
}
bool MemRegion::hasStackParametersStorage() const {
return isa<StackArgumentsSpaceRegion>(getMemorySpace());
}
bool MemRegion::hasGlobalsOrParametersStorage() const {
const MemSpaceRegion *MS = getMemorySpace();
return isa<StackArgumentsSpaceRegion>(MS) ||
isa<GlobalsSpaceRegion>(MS);
}
// getBaseRegion strips away all elements and fields, and get the base region
// of them.
const MemRegion *MemRegion::getBaseRegion() const {
const MemRegion *R = this;
while (true) {
switch (R->getKind()) {
case MemRegion::ElementRegionKind:
case MemRegion::FieldRegionKind:
case MemRegion::ObjCIvarRegionKind:
case MemRegion::CXXBaseObjectRegionKind:
R = cast<SubRegion>(R)->getSuperRegion();
continue;
default:
break;
}
break;
}
return R;
}
bool MemRegion::isSubRegionOf(const MemRegion *R) const {
return false;
}
//===----------------------------------------------------------------------===//
// View handling.
//===----------------------------------------------------------------------===//
const MemRegion *MemRegion::StripCasts(bool StripBaseCasts) const {
const MemRegion *R = this;
while (true) {
switch (R->getKind()) {
case ElementRegionKind: {
const ElementRegion *ER = cast<ElementRegion>(R);
if (!ER->getIndex().isZeroConstant())
return R;
R = ER->getSuperRegion();
break;
}
case CXXBaseObjectRegionKind:
if (!StripBaseCasts)
return R;
R = cast<CXXBaseObjectRegion>(R)->getSuperRegion();
break;
default:
return R;
}
}
}
This is a fairly large patch, which resulted from a cascade of changes made to RegionStore (and related classes) in order to handle some analyzer failures involving casts and manipulation of symbolic memory. The root of the change is in StoreManager::CastRegion(). Instead of using ad hoc heuristics to decide when to layer an ElementRegion on a casted MemRegion, we now always layer an ElementRegion when the cast type is different than the original type of the region. This carries the current cast information associated with a region around without resorting to the error prone recording of "casted types" in GRState. Along with this new policy of layering ElementRegions, I added a new algorithm to strip away existing ElementRegions when they simply represented casts of a base memory object. This algorithm computes the raw "byte offset" that an ElementRegion represents from the base region, and allows the new ElementRegion to be based off that offset. The added benefit is that this naturally handles a series of casts of a MemRegion without building up a set of redundant ElementRegions (thus canonicalizing the region view). Other related changes that cascaded from this one (as tests were failing in RegionStore): - Revamped RegionStoreManager::InvalidateRegion() to completely remove all bindings and default values from a region and all subregions. Now invalidated fields are not bound directly to new symbolic values; instead the base region has a "default" symbol value from which "derived symbols" can be created. The main advantage of this approach is that it allows us to invalidate a region hierarchy and then lazily instantiate new values no matter how deep the hierarchy went (i.e., regardless of the number of field accesses, e.g. x->f->y->z->...). The previous approach did not do this. - Slightly reworked RegionStoreManager::RemoveDeadBindings() to also incorporate live symbols and live regions that do not have direct bindings but also have "default values" used for lazy instantiation. The changes to 'InvalidateRegion' revealed that these were necessary in order to achieve lazy instantiation of values in the region store with those bindings being removed too early. - The changes to InvalidateRegion() and RemoveDeadBindings() revealed a serious bug in 'getSubRegionMap()' where not all region -> subregion relationships involved in actually bindings (explicit and implicit) were being recorded. This has been fixed by using a worklist algorithm to iteratively fill in the region map. - Added special support to RegionStoreManager::Bind()/Retrieve() to handle OSAtomicCompareAndSwap in light of the new 'CastRegion' changes and the layering of ElementRegions. - Fixed a bug in SymbolReaper::isLive() where derived symbols were not being marked live if the symbol they were derived from was also live. This fix was critical for getting lazy instantiation in RegionStore to work. - Tidied up the implementation of ValueManager::getXXXSymbolVal() methods to use SymbolManager::canSymbolicate() to decide whether or not a symbol should be symbolicated. - 'test/Analysis/misc-ps-xfail.m' now passes; that test case has been moved to 'test/Analysis/misc-ps.m'. - Tweaked some pretty-printing of MemRegions, and implemented 'ElementRegion::getRawOffset()' for use with the CastRegion changes. llvm-svn: 77782
2009-08-01 14:17:29 +08:00
const SymbolicRegion *MemRegion::getSymbolicBase() const {
const SubRegion *SubR = dyn_cast<SubRegion>(this);
while (SubR) {
if (const SymbolicRegion *SymR = dyn_cast<SymbolicRegion>(SubR))
return SymR;
SubR = dyn_cast<SubRegion>(SubR->getSuperRegion());
}
return 0;
}
This is a fairly large patch, which resulted from a cascade of changes made to RegionStore (and related classes) in order to handle some analyzer failures involving casts and manipulation of symbolic memory. The root of the change is in StoreManager::CastRegion(). Instead of using ad hoc heuristics to decide when to layer an ElementRegion on a casted MemRegion, we now always layer an ElementRegion when the cast type is different than the original type of the region. This carries the current cast information associated with a region around without resorting to the error prone recording of "casted types" in GRState. Along with this new policy of layering ElementRegions, I added a new algorithm to strip away existing ElementRegions when they simply represented casts of a base memory object. This algorithm computes the raw "byte offset" that an ElementRegion represents from the base region, and allows the new ElementRegion to be based off that offset. The added benefit is that this naturally handles a series of casts of a MemRegion without building up a set of redundant ElementRegions (thus canonicalizing the region view). Other related changes that cascaded from this one (as tests were failing in RegionStore): - Revamped RegionStoreManager::InvalidateRegion() to completely remove all bindings and default values from a region and all subregions. Now invalidated fields are not bound directly to new symbolic values; instead the base region has a "default" symbol value from which "derived symbols" can be created. The main advantage of this approach is that it allows us to invalidate a region hierarchy and then lazily instantiate new values no matter how deep the hierarchy went (i.e., regardless of the number of field accesses, e.g. x->f->y->z->...). The previous approach did not do this. - Slightly reworked RegionStoreManager::RemoveDeadBindings() to also incorporate live symbols and live regions that do not have direct bindings but also have "default values" used for lazy instantiation. The changes to 'InvalidateRegion' revealed that these were necessary in order to achieve lazy instantiation of values in the region store with those bindings being removed too early. - The changes to InvalidateRegion() and RemoveDeadBindings() revealed a serious bug in 'getSubRegionMap()' where not all region -> subregion relationships involved in actually bindings (explicit and implicit) were being recorded. This has been fixed by using a worklist algorithm to iteratively fill in the region map. - Added special support to RegionStoreManager::Bind()/Retrieve() to handle OSAtomicCompareAndSwap in light of the new 'CastRegion' changes and the layering of ElementRegions. - Fixed a bug in SymbolReaper::isLive() where derived symbols were not being marked live if the symbol they were derived from was also live. This fix was critical for getting lazy instantiation in RegionStore to work. - Tidied up the implementation of ValueManager::getXXXSymbolVal() methods to use SymbolManager::canSymbolicate() to decide whether or not a symbol should be symbolicated. - 'test/Analysis/misc-ps-xfail.m' now passes; that test case has been moved to 'test/Analysis/misc-ps.m'. - Tweaked some pretty-printing of MemRegions, and implemented 'ElementRegion::getRawOffset()' for use with the CastRegion changes. llvm-svn: 77782
2009-08-01 14:17:29 +08:00
// FIXME: Merge with the implementation of the same method in Store.cpp
static bool IsCompleteType(ASTContext &Ctx, QualType Ty) {
if (const RecordType *RT = Ty->getAs<RecordType>()) {
const RecordDecl *D = RT->getDecl();
if (!D->getDefinition())
This is a fairly large patch, which resulted from a cascade of changes made to RegionStore (and related classes) in order to handle some analyzer failures involving casts and manipulation of symbolic memory. The root of the change is in StoreManager::CastRegion(). Instead of using ad hoc heuristics to decide when to layer an ElementRegion on a casted MemRegion, we now always layer an ElementRegion when the cast type is different than the original type of the region. This carries the current cast information associated with a region around without resorting to the error prone recording of "casted types" in GRState. Along with this new policy of layering ElementRegions, I added a new algorithm to strip away existing ElementRegions when they simply represented casts of a base memory object. This algorithm computes the raw "byte offset" that an ElementRegion represents from the base region, and allows the new ElementRegion to be based off that offset. The added benefit is that this naturally handles a series of casts of a MemRegion without building up a set of redundant ElementRegions (thus canonicalizing the region view). Other related changes that cascaded from this one (as tests were failing in RegionStore): - Revamped RegionStoreManager::InvalidateRegion() to completely remove all bindings and default values from a region and all subregions. Now invalidated fields are not bound directly to new symbolic values; instead the base region has a "default" symbol value from which "derived symbols" can be created. The main advantage of this approach is that it allows us to invalidate a region hierarchy and then lazily instantiate new values no matter how deep the hierarchy went (i.e., regardless of the number of field accesses, e.g. x->f->y->z->...). The previous approach did not do this. - Slightly reworked RegionStoreManager::RemoveDeadBindings() to also incorporate live symbols and live regions that do not have direct bindings but also have "default values" used for lazy instantiation. The changes to 'InvalidateRegion' revealed that these were necessary in order to achieve lazy instantiation of values in the region store with those bindings being removed too early. - The changes to InvalidateRegion() and RemoveDeadBindings() revealed a serious bug in 'getSubRegionMap()' where not all region -> subregion relationships involved in actually bindings (explicit and implicit) were being recorded. This has been fixed by using a worklist algorithm to iteratively fill in the region map. - Added special support to RegionStoreManager::Bind()/Retrieve() to handle OSAtomicCompareAndSwap in light of the new 'CastRegion' changes and the layering of ElementRegions. - Fixed a bug in SymbolReaper::isLive() where derived symbols were not being marked live if the symbol they were derived from was also live. This fix was critical for getting lazy instantiation in RegionStore to work. - Tidied up the implementation of ValueManager::getXXXSymbolVal() methods to use SymbolManager::canSymbolicate() to decide whether or not a symbol should be symbolicated. - 'test/Analysis/misc-ps-xfail.m' now passes; that test case has been moved to 'test/Analysis/misc-ps.m'. - Tweaked some pretty-printing of MemRegions, and implemented 'ElementRegion::getRawOffset()' for use with the CastRegion changes. llvm-svn: 77782
2009-08-01 14:17:29 +08:00
return false;
}
return true;
}
RegionRawOffset ElementRegion::getAsArrayOffset() const {
CharUnits offset = CharUnits::Zero();
This is a fairly large patch, which resulted from a cascade of changes made to RegionStore (and related classes) in order to handle some analyzer failures involving casts and manipulation of symbolic memory. The root of the change is in StoreManager::CastRegion(). Instead of using ad hoc heuristics to decide when to layer an ElementRegion on a casted MemRegion, we now always layer an ElementRegion when the cast type is different than the original type of the region. This carries the current cast information associated with a region around without resorting to the error prone recording of "casted types" in GRState. Along with this new policy of layering ElementRegions, I added a new algorithm to strip away existing ElementRegions when they simply represented casts of a base memory object. This algorithm computes the raw "byte offset" that an ElementRegion represents from the base region, and allows the new ElementRegion to be based off that offset. The added benefit is that this naturally handles a series of casts of a MemRegion without building up a set of redundant ElementRegions (thus canonicalizing the region view). Other related changes that cascaded from this one (as tests were failing in RegionStore): - Revamped RegionStoreManager::InvalidateRegion() to completely remove all bindings and default values from a region and all subregions. Now invalidated fields are not bound directly to new symbolic values; instead the base region has a "default" symbol value from which "derived symbols" can be created. The main advantage of this approach is that it allows us to invalidate a region hierarchy and then lazily instantiate new values no matter how deep the hierarchy went (i.e., regardless of the number of field accesses, e.g. x->f->y->z->...). The previous approach did not do this. - Slightly reworked RegionStoreManager::RemoveDeadBindings() to also incorporate live symbols and live regions that do not have direct bindings but also have "default values" used for lazy instantiation. The changes to 'InvalidateRegion' revealed that these were necessary in order to achieve lazy instantiation of values in the region store with those bindings being removed too early. - The changes to InvalidateRegion() and RemoveDeadBindings() revealed a serious bug in 'getSubRegionMap()' where not all region -> subregion relationships involved in actually bindings (explicit and implicit) were being recorded. This has been fixed by using a worklist algorithm to iteratively fill in the region map. - Added special support to RegionStoreManager::Bind()/Retrieve() to handle OSAtomicCompareAndSwap in light of the new 'CastRegion' changes and the layering of ElementRegions. - Fixed a bug in SymbolReaper::isLive() where derived symbols were not being marked live if the symbol they were derived from was also live. This fix was critical for getting lazy instantiation in RegionStore to work. - Tidied up the implementation of ValueManager::getXXXSymbolVal() methods to use SymbolManager::canSymbolicate() to decide whether or not a symbol should be symbolicated. - 'test/Analysis/misc-ps-xfail.m' now passes; that test case has been moved to 'test/Analysis/misc-ps.m'. - Tweaked some pretty-printing of MemRegions, and implemented 'ElementRegion::getRawOffset()' for use with the CastRegion changes. llvm-svn: 77782
2009-08-01 14:17:29 +08:00
const ElementRegion *ER = this;
const MemRegion *superR = NULL;
ASTContext &C = getContext();
This is a fairly large patch, which resulted from a cascade of changes made to RegionStore (and related classes) in order to handle some analyzer failures involving casts and manipulation of symbolic memory. The root of the change is in StoreManager::CastRegion(). Instead of using ad hoc heuristics to decide when to layer an ElementRegion on a casted MemRegion, we now always layer an ElementRegion when the cast type is different than the original type of the region. This carries the current cast information associated with a region around without resorting to the error prone recording of "casted types" in GRState. Along with this new policy of layering ElementRegions, I added a new algorithm to strip away existing ElementRegions when they simply represented casts of a base memory object. This algorithm computes the raw "byte offset" that an ElementRegion represents from the base region, and allows the new ElementRegion to be based off that offset. The added benefit is that this naturally handles a series of casts of a MemRegion without building up a set of redundant ElementRegions (thus canonicalizing the region view). Other related changes that cascaded from this one (as tests were failing in RegionStore): - Revamped RegionStoreManager::InvalidateRegion() to completely remove all bindings and default values from a region and all subregions. Now invalidated fields are not bound directly to new symbolic values; instead the base region has a "default" symbol value from which "derived symbols" can be created. The main advantage of this approach is that it allows us to invalidate a region hierarchy and then lazily instantiate new values no matter how deep the hierarchy went (i.e., regardless of the number of field accesses, e.g. x->f->y->z->...). The previous approach did not do this. - Slightly reworked RegionStoreManager::RemoveDeadBindings() to also incorporate live symbols and live regions that do not have direct bindings but also have "default values" used for lazy instantiation. The changes to 'InvalidateRegion' revealed that these were necessary in order to achieve lazy instantiation of values in the region store with those bindings being removed too early. - The changes to InvalidateRegion() and RemoveDeadBindings() revealed a serious bug in 'getSubRegionMap()' where not all region -> subregion relationships involved in actually bindings (explicit and implicit) were being recorded. This has been fixed by using a worklist algorithm to iteratively fill in the region map. - Added special support to RegionStoreManager::Bind()/Retrieve() to handle OSAtomicCompareAndSwap in light of the new 'CastRegion' changes and the layering of ElementRegions. - Fixed a bug in SymbolReaper::isLive() where derived symbols were not being marked live if the symbol they were derived from was also live. This fix was critical for getting lazy instantiation in RegionStore to work. - Tidied up the implementation of ValueManager::getXXXSymbolVal() methods to use SymbolManager::canSymbolicate() to decide whether or not a symbol should be symbolicated. - 'test/Analysis/misc-ps-xfail.m' now passes; that test case has been moved to 'test/Analysis/misc-ps.m'. - Tweaked some pretty-printing of MemRegions, and implemented 'ElementRegion::getRawOffset()' for use with the CastRegion changes. llvm-svn: 77782
2009-08-01 14:17:29 +08:00
// FIXME: Handle multi-dimensional arrays.
while (ER) {
superR = ER->getSuperRegion();
This is a fairly large patch, which resulted from a cascade of changes made to RegionStore (and related classes) in order to handle some analyzer failures involving casts and manipulation of symbolic memory. The root of the change is in StoreManager::CastRegion(). Instead of using ad hoc heuristics to decide when to layer an ElementRegion on a casted MemRegion, we now always layer an ElementRegion when the cast type is different than the original type of the region. This carries the current cast information associated with a region around without resorting to the error prone recording of "casted types" in GRState. Along with this new policy of layering ElementRegions, I added a new algorithm to strip away existing ElementRegions when they simply represented casts of a base memory object. This algorithm computes the raw "byte offset" that an ElementRegion represents from the base region, and allows the new ElementRegion to be based off that offset. The added benefit is that this naturally handles a series of casts of a MemRegion without building up a set of redundant ElementRegions (thus canonicalizing the region view). Other related changes that cascaded from this one (as tests were failing in RegionStore): - Revamped RegionStoreManager::InvalidateRegion() to completely remove all bindings and default values from a region and all subregions. Now invalidated fields are not bound directly to new symbolic values; instead the base region has a "default" symbol value from which "derived symbols" can be created. The main advantage of this approach is that it allows us to invalidate a region hierarchy and then lazily instantiate new values no matter how deep the hierarchy went (i.e., regardless of the number of field accesses, e.g. x->f->y->z->...). The previous approach did not do this. - Slightly reworked RegionStoreManager::RemoveDeadBindings() to also incorporate live symbols and live regions that do not have direct bindings but also have "default values" used for lazy instantiation. The changes to 'InvalidateRegion' revealed that these were necessary in order to achieve lazy instantiation of values in the region store with those bindings being removed too early. - The changes to InvalidateRegion() and RemoveDeadBindings() revealed a serious bug in 'getSubRegionMap()' where not all region -> subregion relationships involved in actually bindings (explicit and implicit) were being recorded. This has been fixed by using a worklist algorithm to iteratively fill in the region map. - Added special support to RegionStoreManager::Bind()/Retrieve() to handle OSAtomicCompareAndSwap in light of the new 'CastRegion' changes and the layering of ElementRegions. - Fixed a bug in SymbolReaper::isLive() where derived symbols were not being marked live if the symbol they were derived from was also live. This fix was critical for getting lazy instantiation in RegionStore to work. - Tidied up the implementation of ValueManager::getXXXSymbolVal() methods to use SymbolManager::canSymbolicate() to decide whether or not a symbol should be symbolicated. - 'test/Analysis/misc-ps-xfail.m' now passes; that test case has been moved to 'test/Analysis/misc-ps.m'. - Tweaked some pretty-printing of MemRegions, and implemented 'ElementRegion::getRawOffset()' for use with the CastRegion changes. llvm-svn: 77782
2009-08-01 14:17:29 +08:00
// FIXME: generalize to symbolic offsets.
SVal index = ER->getIndex();
if (Optional<nonloc::ConcreteInt> CI = index.getAs<nonloc::ConcreteInt>()) {
This is a fairly large patch, which resulted from a cascade of changes made to RegionStore (and related classes) in order to handle some analyzer failures involving casts and manipulation of symbolic memory. The root of the change is in StoreManager::CastRegion(). Instead of using ad hoc heuristics to decide when to layer an ElementRegion on a casted MemRegion, we now always layer an ElementRegion when the cast type is different than the original type of the region. This carries the current cast information associated with a region around without resorting to the error prone recording of "casted types" in GRState. Along with this new policy of layering ElementRegions, I added a new algorithm to strip away existing ElementRegions when they simply represented casts of a base memory object. This algorithm computes the raw "byte offset" that an ElementRegion represents from the base region, and allows the new ElementRegion to be based off that offset. The added benefit is that this naturally handles a series of casts of a MemRegion without building up a set of redundant ElementRegions (thus canonicalizing the region view). Other related changes that cascaded from this one (as tests were failing in RegionStore): - Revamped RegionStoreManager::InvalidateRegion() to completely remove all bindings and default values from a region and all subregions. Now invalidated fields are not bound directly to new symbolic values; instead the base region has a "default" symbol value from which "derived symbols" can be created. The main advantage of this approach is that it allows us to invalidate a region hierarchy and then lazily instantiate new values no matter how deep the hierarchy went (i.e., regardless of the number of field accesses, e.g. x->f->y->z->...). The previous approach did not do this. - Slightly reworked RegionStoreManager::RemoveDeadBindings() to also incorporate live symbols and live regions that do not have direct bindings but also have "default values" used for lazy instantiation. The changes to 'InvalidateRegion' revealed that these were necessary in order to achieve lazy instantiation of values in the region store with those bindings being removed too early. - The changes to InvalidateRegion() and RemoveDeadBindings() revealed a serious bug in 'getSubRegionMap()' where not all region -> subregion relationships involved in actually bindings (explicit and implicit) were being recorded. This has been fixed by using a worklist algorithm to iteratively fill in the region map. - Added special support to RegionStoreManager::Bind()/Retrieve() to handle OSAtomicCompareAndSwap in light of the new 'CastRegion' changes and the layering of ElementRegions. - Fixed a bug in SymbolReaper::isLive() where derived symbols were not being marked live if the symbol they were derived from was also live. This fix was critical for getting lazy instantiation in RegionStore to work. - Tidied up the implementation of ValueManager::getXXXSymbolVal() methods to use SymbolManager::canSymbolicate() to decide whether or not a symbol should be symbolicated. - 'test/Analysis/misc-ps-xfail.m' now passes; that test case has been moved to 'test/Analysis/misc-ps.m'. - Tweaked some pretty-printing of MemRegions, and implemented 'ElementRegion::getRawOffset()' for use with the CastRegion changes. llvm-svn: 77782
2009-08-01 14:17:29 +08:00
// Update the offset.
int64_t i = CI->getValue().getSExtValue();
This is a fairly large patch, which resulted from a cascade of changes made to RegionStore (and related classes) in order to handle some analyzer failures involving casts and manipulation of symbolic memory. The root of the change is in StoreManager::CastRegion(). Instead of using ad hoc heuristics to decide when to layer an ElementRegion on a casted MemRegion, we now always layer an ElementRegion when the cast type is different than the original type of the region. This carries the current cast information associated with a region around without resorting to the error prone recording of "casted types" in GRState. Along with this new policy of layering ElementRegions, I added a new algorithm to strip away existing ElementRegions when they simply represented casts of a base memory object. This algorithm computes the raw "byte offset" that an ElementRegion represents from the base region, and allows the new ElementRegion to be based off that offset. The added benefit is that this naturally handles a series of casts of a MemRegion without building up a set of redundant ElementRegions (thus canonicalizing the region view). Other related changes that cascaded from this one (as tests were failing in RegionStore): - Revamped RegionStoreManager::InvalidateRegion() to completely remove all bindings and default values from a region and all subregions. Now invalidated fields are not bound directly to new symbolic values; instead the base region has a "default" symbol value from which "derived symbols" can be created. The main advantage of this approach is that it allows us to invalidate a region hierarchy and then lazily instantiate new values no matter how deep the hierarchy went (i.e., regardless of the number of field accesses, e.g. x->f->y->z->...). The previous approach did not do this. - Slightly reworked RegionStoreManager::RemoveDeadBindings() to also incorporate live symbols and live regions that do not have direct bindings but also have "default values" used for lazy instantiation. The changes to 'InvalidateRegion' revealed that these were necessary in order to achieve lazy instantiation of values in the region store with those bindings being removed too early. - The changes to InvalidateRegion() and RemoveDeadBindings() revealed a serious bug in 'getSubRegionMap()' where not all region -> subregion relationships involved in actually bindings (explicit and implicit) were being recorded. This has been fixed by using a worklist algorithm to iteratively fill in the region map. - Added special support to RegionStoreManager::Bind()/Retrieve() to handle OSAtomicCompareAndSwap in light of the new 'CastRegion' changes and the layering of ElementRegions. - Fixed a bug in SymbolReaper::isLive() where derived symbols were not being marked live if the symbol they were derived from was also live. This fix was critical for getting lazy instantiation in RegionStore to work. - Tidied up the implementation of ValueManager::getXXXSymbolVal() methods to use SymbolManager::canSymbolicate() to decide whether or not a symbol should be symbolicated. - 'test/Analysis/misc-ps-xfail.m' now passes; that test case has been moved to 'test/Analysis/misc-ps.m'. - Tweaked some pretty-printing of MemRegions, and implemented 'ElementRegion::getRawOffset()' for use with the CastRegion changes. llvm-svn: 77782
2009-08-01 14:17:29 +08:00
if (i != 0) {
QualType elemType = ER->getElementType();
This is a fairly large patch, which resulted from a cascade of changes made to RegionStore (and related classes) in order to handle some analyzer failures involving casts and manipulation of symbolic memory. The root of the change is in StoreManager::CastRegion(). Instead of using ad hoc heuristics to decide when to layer an ElementRegion on a casted MemRegion, we now always layer an ElementRegion when the cast type is different than the original type of the region. This carries the current cast information associated with a region around without resorting to the error prone recording of "casted types" in GRState. Along with this new policy of layering ElementRegions, I added a new algorithm to strip away existing ElementRegions when they simply represented casts of a base memory object. This algorithm computes the raw "byte offset" that an ElementRegion represents from the base region, and allows the new ElementRegion to be based off that offset. The added benefit is that this naturally handles a series of casts of a MemRegion without building up a set of redundant ElementRegions (thus canonicalizing the region view). Other related changes that cascaded from this one (as tests were failing in RegionStore): - Revamped RegionStoreManager::InvalidateRegion() to completely remove all bindings and default values from a region and all subregions. Now invalidated fields are not bound directly to new symbolic values; instead the base region has a "default" symbol value from which "derived symbols" can be created. The main advantage of this approach is that it allows us to invalidate a region hierarchy and then lazily instantiate new values no matter how deep the hierarchy went (i.e., regardless of the number of field accesses, e.g. x->f->y->z->...). The previous approach did not do this. - Slightly reworked RegionStoreManager::RemoveDeadBindings() to also incorporate live symbols and live regions that do not have direct bindings but also have "default values" used for lazy instantiation. The changes to 'InvalidateRegion' revealed that these were necessary in order to achieve lazy instantiation of values in the region store with those bindings being removed too early. - The changes to InvalidateRegion() and RemoveDeadBindings() revealed a serious bug in 'getSubRegionMap()' where not all region -> subregion relationships involved in actually bindings (explicit and implicit) were being recorded. This has been fixed by using a worklist algorithm to iteratively fill in the region map. - Added special support to RegionStoreManager::Bind()/Retrieve() to handle OSAtomicCompareAndSwap in light of the new 'CastRegion' changes and the layering of ElementRegions. - Fixed a bug in SymbolReaper::isLive() where derived symbols were not being marked live if the symbol they were derived from was also live. This fix was critical for getting lazy instantiation in RegionStore to work. - Tidied up the implementation of ValueManager::getXXXSymbolVal() methods to use SymbolManager::canSymbolicate() to decide whether or not a symbol should be symbolicated. - 'test/Analysis/misc-ps-xfail.m' now passes; that test case has been moved to 'test/Analysis/misc-ps.m'. - Tweaked some pretty-printing of MemRegions, and implemented 'ElementRegion::getRawOffset()' for use with the CastRegion changes. llvm-svn: 77782
2009-08-01 14:17:29 +08:00
// If we are pointing to an incomplete type, go no further.
if (!IsCompleteType(C, elemType)) {
superR = ER;
break;
}
CharUnits size = C.getTypeSizeInChars(elemType);
This is a fairly large patch, which resulted from a cascade of changes made to RegionStore (and related classes) in order to handle some analyzer failures involving casts and manipulation of symbolic memory. The root of the change is in StoreManager::CastRegion(). Instead of using ad hoc heuristics to decide when to layer an ElementRegion on a casted MemRegion, we now always layer an ElementRegion when the cast type is different than the original type of the region. This carries the current cast information associated with a region around without resorting to the error prone recording of "casted types" in GRState. Along with this new policy of layering ElementRegions, I added a new algorithm to strip away existing ElementRegions when they simply represented casts of a base memory object. This algorithm computes the raw "byte offset" that an ElementRegion represents from the base region, and allows the new ElementRegion to be based off that offset. The added benefit is that this naturally handles a series of casts of a MemRegion without building up a set of redundant ElementRegions (thus canonicalizing the region view). Other related changes that cascaded from this one (as tests were failing in RegionStore): - Revamped RegionStoreManager::InvalidateRegion() to completely remove all bindings and default values from a region and all subregions. Now invalidated fields are not bound directly to new symbolic values; instead the base region has a "default" symbol value from which "derived symbols" can be created. The main advantage of this approach is that it allows us to invalidate a region hierarchy and then lazily instantiate new values no matter how deep the hierarchy went (i.e., regardless of the number of field accesses, e.g. x->f->y->z->...). The previous approach did not do this. - Slightly reworked RegionStoreManager::RemoveDeadBindings() to also incorporate live symbols and live regions that do not have direct bindings but also have "default values" used for lazy instantiation. The changes to 'InvalidateRegion' revealed that these were necessary in order to achieve lazy instantiation of values in the region store with those bindings being removed too early. - The changes to InvalidateRegion() and RemoveDeadBindings() revealed a serious bug in 'getSubRegionMap()' where not all region -> subregion relationships involved in actually bindings (explicit and implicit) were being recorded. This has been fixed by using a worklist algorithm to iteratively fill in the region map. - Added special support to RegionStoreManager::Bind()/Retrieve() to handle OSAtomicCompareAndSwap in light of the new 'CastRegion' changes and the layering of ElementRegions. - Fixed a bug in SymbolReaper::isLive() where derived symbols were not being marked live if the symbol they were derived from was also live. This fix was critical for getting lazy instantiation in RegionStore to work. - Tidied up the implementation of ValueManager::getXXXSymbolVal() methods to use SymbolManager::canSymbolicate() to decide whether or not a symbol should be symbolicated. - 'test/Analysis/misc-ps-xfail.m' now passes; that test case has been moved to 'test/Analysis/misc-ps.m'. - Tweaked some pretty-printing of MemRegions, and implemented 'ElementRegion::getRawOffset()' for use with the CastRegion changes. llvm-svn: 77782
2009-08-01 14:17:29 +08:00
offset += (i * size);
}
// Go to the next ElementRegion (if any).
ER = dyn_cast<ElementRegion>(superR);
continue;
}
This is a fairly large patch, which resulted from a cascade of changes made to RegionStore (and related classes) in order to handle some analyzer failures involving casts and manipulation of symbolic memory. The root of the change is in StoreManager::CastRegion(). Instead of using ad hoc heuristics to decide when to layer an ElementRegion on a casted MemRegion, we now always layer an ElementRegion when the cast type is different than the original type of the region. This carries the current cast information associated with a region around without resorting to the error prone recording of "casted types" in GRState. Along with this new policy of layering ElementRegions, I added a new algorithm to strip away existing ElementRegions when they simply represented casts of a base memory object. This algorithm computes the raw "byte offset" that an ElementRegion represents from the base region, and allows the new ElementRegion to be based off that offset. The added benefit is that this naturally handles a series of casts of a MemRegion without building up a set of redundant ElementRegions (thus canonicalizing the region view). Other related changes that cascaded from this one (as tests were failing in RegionStore): - Revamped RegionStoreManager::InvalidateRegion() to completely remove all bindings and default values from a region and all subregions. Now invalidated fields are not bound directly to new symbolic values; instead the base region has a "default" symbol value from which "derived symbols" can be created. The main advantage of this approach is that it allows us to invalidate a region hierarchy and then lazily instantiate new values no matter how deep the hierarchy went (i.e., regardless of the number of field accesses, e.g. x->f->y->z->...). The previous approach did not do this. - Slightly reworked RegionStoreManager::RemoveDeadBindings() to also incorporate live symbols and live regions that do not have direct bindings but also have "default values" used for lazy instantiation. The changes to 'InvalidateRegion' revealed that these were necessary in order to achieve lazy instantiation of values in the region store with those bindings being removed too early. - The changes to InvalidateRegion() and RemoveDeadBindings() revealed a serious bug in 'getSubRegionMap()' where not all region -> subregion relationships involved in actually bindings (explicit and implicit) were being recorded. This has been fixed by using a worklist algorithm to iteratively fill in the region map. - Added special support to RegionStoreManager::Bind()/Retrieve() to handle OSAtomicCompareAndSwap in light of the new 'CastRegion' changes and the layering of ElementRegions. - Fixed a bug in SymbolReaper::isLive() where derived symbols were not being marked live if the symbol they were derived from was also live. This fix was critical for getting lazy instantiation in RegionStore to work. - Tidied up the implementation of ValueManager::getXXXSymbolVal() methods to use SymbolManager::canSymbolicate() to decide whether or not a symbol should be symbolicated. - 'test/Analysis/misc-ps-xfail.m' now passes; that test case has been moved to 'test/Analysis/misc-ps.m'. - Tweaked some pretty-printing of MemRegions, and implemented 'ElementRegion::getRawOffset()' for use with the CastRegion changes. llvm-svn: 77782
2009-08-01 14:17:29 +08:00
return NULL;
}
This is a fairly large patch, which resulted from a cascade of changes made to RegionStore (and related classes) in order to handle some analyzer failures involving casts and manipulation of symbolic memory. The root of the change is in StoreManager::CastRegion(). Instead of using ad hoc heuristics to decide when to layer an ElementRegion on a casted MemRegion, we now always layer an ElementRegion when the cast type is different than the original type of the region. This carries the current cast information associated with a region around without resorting to the error prone recording of "casted types" in GRState. Along with this new policy of layering ElementRegions, I added a new algorithm to strip away existing ElementRegions when they simply represented casts of a base memory object. This algorithm computes the raw "byte offset" that an ElementRegion represents from the base region, and allows the new ElementRegion to be based off that offset. The added benefit is that this naturally handles a series of casts of a MemRegion without building up a set of redundant ElementRegions (thus canonicalizing the region view). Other related changes that cascaded from this one (as tests were failing in RegionStore): - Revamped RegionStoreManager::InvalidateRegion() to completely remove all bindings and default values from a region and all subregions. Now invalidated fields are not bound directly to new symbolic values; instead the base region has a "default" symbol value from which "derived symbols" can be created. The main advantage of this approach is that it allows us to invalidate a region hierarchy and then lazily instantiate new values no matter how deep the hierarchy went (i.e., regardless of the number of field accesses, e.g. x->f->y->z->...). The previous approach did not do this. - Slightly reworked RegionStoreManager::RemoveDeadBindings() to also incorporate live symbols and live regions that do not have direct bindings but also have "default values" used for lazy instantiation. The changes to 'InvalidateRegion' revealed that these were necessary in order to achieve lazy instantiation of values in the region store with those bindings being removed too early. - The changes to InvalidateRegion() and RemoveDeadBindings() revealed a serious bug in 'getSubRegionMap()' where not all region -> subregion relationships involved in actually bindings (explicit and implicit) were being recorded. This has been fixed by using a worklist algorithm to iteratively fill in the region map. - Added special support to RegionStoreManager::Bind()/Retrieve() to handle OSAtomicCompareAndSwap in light of the new 'CastRegion' changes and the layering of ElementRegions. - Fixed a bug in SymbolReaper::isLive() where derived symbols were not being marked live if the symbol they were derived from was also live. This fix was critical for getting lazy instantiation in RegionStore to work. - Tidied up the implementation of ValueManager::getXXXSymbolVal() methods to use SymbolManager::canSymbolicate() to decide whether or not a symbol should be symbolicated. - 'test/Analysis/misc-ps-xfail.m' now passes; that test case has been moved to 'test/Analysis/misc-ps.m'. - Tweaked some pretty-printing of MemRegions, and implemented 'ElementRegion::getRawOffset()' for use with the CastRegion changes. llvm-svn: 77782
2009-08-01 14:17:29 +08:00
assert(superR && "super region cannot be NULL");
return RegionRawOffset(superR, offset);
This is a fairly large patch, which resulted from a cascade of changes made to RegionStore (and related classes) in order to handle some analyzer failures involving casts and manipulation of symbolic memory. The root of the change is in StoreManager::CastRegion(). Instead of using ad hoc heuristics to decide when to layer an ElementRegion on a casted MemRegion, we now always layer an ElementRegion when the cast type is different than the original type of the region. This carries the current cast information associated with a region around without resorting to the error prone recording of "casted types" in GRState. Along with this new policy of layering ElementRegions, I added a new algorithm to strip away existing ElementRegions when they simply represented casts of a base memory object. This algorithm computes the raw "byte offset" that an ElementRegion represents from the base region, and allows the new ElementRegion to be based off that offset. The added benefit is that this naturally handles a series of casts of a MemRegion without building up a set of redundant ElementRegions (thus canonicalizing the region view). Other related changes that cascaded from this one (as tests were failing in RegionStore): - Revamped RegionStoreManager::InvalidateRegion() to completely remove all bindings and default values from a region and all subregions. Now invalidated fields are not bound directly to new symbolic values; instead the base region has a "default" symbol value from which "derived symbols" can be created. The main advantage of this approach is that it allows us to invalidate a region hierarchy and then lazily instantiate new values no matter how deep the hierarchy went (i.e., regardless of the number of field accesses, e.g. x->f->y->z->...). The previous approach did not do this. - Slightly reworked RegionStoreManager::RemoveDeadBindings() to also incorporate live symbols and live regions that do not have direct bindings but also have "default values" used for lazy instantiation. The changes to 'InvalidateRegion' revealed that these were necessary in order to achieve lazy instantiation of values in the region store with those bindings being removed too early. - The changes to InvalidateRegion() and RemoveDeadBindings() revealed a serious bug in 'getSubRegionMap()' where not all region -> subregion relationships involved in actually bindings (explicit and implicit) were being recorded. This has been fixed by using a worklist algorithm to iteratively fill in the region map. - Added special support to RegionStoreManager::Bind()/Retrieve() to handle OSAtomicCompareAndSwap in light of the new 'CastRegion' changes and the layering of ElementRegions. - Fixed a bug in SymbolReaper::isLive() where derived symbols were not being marked live if the symbol they were derived from was also live. This fix was critical for getting lazy instantiation in RegionStore to work. - Tidied up the implementation of ValueManager::getXXXSymbolVal() methods to use SymbolManager::canSymbolicate() to decide whether or not a symbol should be symbolicated. - 'test/Analysis/misc-ps-xfail.m' now passes; that test case has been moved to 'test/Analysis/misc-ps.m'. - Tweaked some pretty-printing of MemRegions, and implemented 'ElementRegion::getRawOffset()' for use with the CastRegion changes. llvm-svn: 77782
2009-08-01 14:17:29 +08:00
}
/// Returns true if \p Base is an immediate base class of \p Child
static bool isImmediateBase(const CXXRecordDecl *Child,
const CXXRecordDecl *Base) {
// Note that we do NOT canonicalize the base class here, because
// ASTRecordLayout doesn't either. If that leads us down the wrong path,
// so be it; at least we won't crash.
for (CXXRecordDecl::base_class_const_iterator I = Child->bases_begin(),
E = Child->bases_end();
I != E; ++I) {
if (I->getType()->getAsCXXRecordDecl() == Base)
return true;
}
return false;
}
RegionOffset MemRegion::getAsOffset() const {
const MemRegion *R = this;
const MemRegion *SymbolicOffsetBase = 0;
int64_t Offset = 0;
while (1) {
switch (R->getKind()) {
case GenericMemSpaceRegionKind:
case StackLocalsSpaceRegionKind:
case StackArgumentsSpaceRegionKind:
case HeapSpaceRegionKind:
case UnknownSpaceRegionKind:
case StaticGlobalSpaceRegionKind:
case GlobalInternalSpaceRegionKind:
case GlobalSystemSpaceRegionKind:
case GlobalImmutableSpaceRegionKind:
// Stores can bind directly to a region space to set a default value.
assert(Offset == 0 && !SymbolicOffsetBase);
goto Finish;
case FunctionTextRegionKind:
case BlockTextRegionKind:
case BlockDataRegionKind:
// These will never have bindings, but may end up having values requested
// if the user does some strange casting.
if (Offset != 0)
SymbolicOffsetBase = R;
goto Finish;
case SymbolicRegionKind:
case AllocaRegionKind:
case CompoundLiteralRegionKind:
case CXXThisRegionKind:
case StringRegionKind:
case ObjCStringRegionKind:
case VarRegionKind:
case CXXTempObjectRegionKind:
// Usual base regions.
goto Finish;
case ObjCIvarRegionKind:
// This is a little strange, but it's a compromise between
// ObjCIvarRegions having unknown compile-time offsets (when using the
// non-fragile runtime) and yet still being distinct, non-overlapping
// regions. Thus we treat them as "like" base regions for the purposes
// of computing offsets.
goto Finish;
case CXXBaseObjectRegionKind: {
const CXXBaseObjectRegion *BOR = cast<CXXBaseObjectRegion>(R);
R = BOR->getSuperRegion();
QualType Ty;
bool RootIsSymbolic = false;
if (const TypedValueRegion *TVR = dyn_cast<TypedValueRegion>(R)) {
Ty = TVR->getDesugaredValueType(getContext());
} else if (const SymbolicRegion *SR = dyn_cast<SymbolicRegion>(R)) {
// If our base region is symbolic, we don't know what type it really is.
// Pretend the type of the symbol is the true dynamic type.
// (This will at least be self-consistent for the life of the symbol.)
Ty = SR->getSymbol()->getType()->getPointeeType();
RootIsSymbolic = true;
}
const CXXRecordDecl *Child = Ty->getAsCXXRecordDecl();
if (!Child) {
// We cannot compute the offset of the base class.
SymbolicOffsetBase = R;
}
if (RootIsSymbolic) {
// Base layers on symbolic regions may not be type-correct.
// Double-check the inheritance here, and revert to a symbolic offset
// if it's invalid (e.g. due to a reinterpret_cast).
if (BOR->isVirtual()) {
if (!Child->isVirtuallyDerivedFrom(BOR->getDecl()))
SymbolicOffsetBase = R;
} else {
if (!isImmediateBase(Child, BOR->getDecl()))
SymbolicOffsetBase = R;
}
}
// Don't bother calculating precise offsets if we already have a
// symbolic offset somewhere in the chain.
if (SymbolicOffsetBase)
continue;
CharUnits BaseOffset;
const ASTRecordLayout &Layout = getContext().getASTRecordLayout(Child);
if (BOR->isVirtual())
BaseOffset = Layout.getVBaseClassOffset(BOR->getDecl());
else
BaseOffset = Layout.getBaseClassOffset(BOR->getDecl());
// The base offset is in chars, not in bits.
Offset += BaseOffset.getQuantity() * getContext().getCharWidth();
break;
}
case ElementRegionKind: {
const ElementRegion *ER = cast<ElementRegion>(R);
R = ER->getSuperRegion();
QualType EleTy = ER->getValueType();
if (!IsCompleteType(getContext(), EleTy)) {
// We cannot compute the offset of the base class.
SymbolicOffsetBase = R;
continue;
}
SVal Index = ER->getIndex();
if (Optional<nonloc::ConcreteInt> CI =
Index.getAs<nonloc::ConcreteInt>()) {
// Don't bother calculating precise offsets if we already have a
// symbolic offset somewhere in the chain.
if (SymbolicOffsetBase)
continue;
int64_t i = CI->getValue().getSExtValue();
// This type size is in bits.
Offset += i * getContext().getTypeSize(EleTy);
} else {
// We cannot compute offset for non-concrete index.
SymbolicOffsetBase = R;
}
break;
}
case FieldRegionKind: {
const FieldRegion *FR = cast<FieldRegion>(R);
R = FR->getSuperRegion();
const RecordDecl *RD = FR->getDecl()->getParent();
if (RD->isUnion() || !RD->isCompleteDefinition()) {
// We cannot compute offset for incomplete type.
// For unions, we could treat everything as offset 0, but we'd rather
// treat each field as a symbolic offset so they aren't stored on top
// of each other, since we depend on things in typed regions actually
// matching their types.
SymbolicOffsetBase = R;
}
// Don't bother calculating precise offsets if we already have a
// symbolic offset somewhere in the chain.
if (SymbolicOffsetBase)
continue;
// Get the field number.
unsigned idx = 0;
for (RecordDecl::field_iterator FI = RD->field_begin(),
FE = RD->field_end(); FI != FE; ++FI, ++idx)
if (FR->getDecl() == *FI)
break;
const ASTRecordLayout &Layout = getContext().getASTRecordLayout(RD);
// This is offset in bits.
Offset += Layout.getFieldOffset(idx);
break;
}
}
}
Finish:
if (SymbolicOffsetBase)
return RegionOffset(SymbolicOffsetBase, RegionOffset::Symbolic);
return RegionOffset(R, Offset);
}
//===----------------------------------------------------------------------===//
// BlockDataRegion
//===----------------------------------------------------------------------===//
std::pair<const VarRegion *, const VarRegion *>
BlockDataRegion::getCaptureRegions(const VarDecl *VD) {
MemRegionManager &MemMgr = *getMemRegionManager();
const VarRegion *VR = 0;
const VarRegion *OriginalVR = 0;
if (!VD->getAttr<BlocksAttr>() && VD->hasLocalStorage()) {
VR = MemMgr.getVarRegion(VD, this);
OriginalVR = MemMgr.getVarRegion(VD, LC);
}
else {
if (LC) {
VR = MemMgr.getVarRegion(VD, LC);
OriginalVR = VR;
}
else {
VR = MemMgr.getVarRegion(VD, MemMgr.getUnknownRegion());
OriginalVR = MemMgr.getVarRegion(VD, LC);
}
}
return std::make_pair(VR, OriginalVR);
}
void BlockDataRegion::LazyInitializeReferencedVars() {
if (ReferencedVars)
return;
AnalysisDeclContext *AC = getCodeRegion()->getAnalysisDeclContext();
AnalysisDeclContext::referenced_decls_iterator I, E;
llvm::tie(I, E) = AC->getReferencedBlockVars(BC->getDecl());
if (I == E) {
ReferencedVars = (void*) 0x1;
return;
}
MemRegionManager &MemMgr = *getMemRegionManager();
llvm::BumpPtrAllocator &A = MemMgr.getAllocator();
BumpVectorContext BC(A);
typedef BumpVector<const MemRegion*> VarVec;
VarVec *BV = (VarVec*) A.Allocate<VarVec>();
new (BV) VarVec(BC, E - I);
VarVec *BVOriginal = (VarVec*) A.Allocate<VarVec>();
new (BVOriginal) VarVec(BC, E - I);
for ( ; I != E; ++I) {
const VarRegion *VR = 0;
const VarRegion *OriginalVR = 0;
llvm::tie(VR, OriginalVR) = getCaptureRegions(*I);
assert(VR);
assert(OriginalVR);
BV->push_back(VR, BC);
BVOriginal->push_back(OriginalVR, BC);
}
ReferencedVars = BV;
OriginalVars = BVOriginal;
}
BlockDataRegion::referenced_vars_iterator
BlockDataRegion::referenced_vars_begin() const {
const_cast<BlockDataRegion*>(this)->LazyInitializeReferencedVars();
BumpVector<const MemRegion*> *Vec =
static_cast<BumpVector<const MemRegion*>*>(ReferencedVars);
if (Vec == (void*) 0x1)
return BlockDataRegion::referenced_vars_iterator(0, 0);
BumpVector<const MemRegion*> *VecOriginal =
static_cast<BumpVector<const MemRegion*>*>(OriginalVars);
return BlockDataRegion::referenced_vars_iterator(Vec->begin(),
VecOriginal->begin());
}
BlockDataRegion::referenced_vars_iterator
BlockDataRegion::referenced_vars_end() const {
const_cast<BlockDataRegion*>(this)->LazyInitializeReferencedVars();
BumpVector<const MemRegion*> *Vec =
static_cast<BumpVector<const MemRegion*>*>(ReferencedVars);
if (Vec == (void*) 0x1)
return BlockDataRegion::referenced_vars_iterator(0, 0);
BumpVector<const MemRegion*> *VecOriginal =
static_cast<BumpVector<const MemRegion*>*>(OriginalVars);
return BlockDataRegion::referenced_vars_iterator(Vec->end(),
VecOriginal->end());
}
const VarRegion *BlockDataRegion::getOriginalRegion(const VarRegion *R) const {
for (referenced_vars_iterator I = referenced_vars_begin(),
E = referenced_vars_end();
I != E; ++I) {
if (I.getCapturedRegion() == R)
return I.getOriginalRegion();
}
return 0;
}
//===----------------------------------------------------------------------===//
// RegionAndSymbolInvalidationTraits
//===----------------------------------------------------------------------===//
void RegionAndSymbolInvalidationTraits::setTrait(SymbolRef Sym,
InvalidationKinds IK) {
SymTraitsMap[Sym] |= IK;
}
void RegionAndSymbolInvalidationTraits::setTrait(const MemRegion *MR,
InvalidationKinds IK) {
assert(MR);
if (const SymbolicRegion *SR = dyn_cast<SymbolicRegion>(MR))
setTrait(SR->getSymbol(), IK);
else
MRTraitsMap[MR] |= IK;
}
bool RegionAndSymbolInvalidationTraits::hasTrait(SymbolRef Sym,
InvalidationKinds IK) {
const_symbol_iterator I = SymTraitsMap.find(Sym);
if (I != SymTraitsMap.end())
return I->second & IK;
return false;
}
bool RegionAndSymbolInvalidationTraits::hasTrait(const MemRegion *MR,
InvalidationKinds IK) {
if (!MR)
return false;
if (const SymbolicRegion *SR = dyn_cast<SymbolicRegion>(MR))
return hasTrait(SR->getSymbol(), IK);
const_region_iterator I = MRTraitsMap.find(MR);
if (I != MRTraitsMap.end())
return I->second & IK;
return false;
}