aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
Diffstat
-rw-r--r--lib/CodeGen/CGExpr.cpp5
-rw-r--r--lib/CodeGen/CGExprConstant.cpp8
-rw-r--r--lib/CodeGen/CGObjC.cpp2
-rw-r--r--lib/CodeGen/CGObjCGNU.cpp10
-rw-r--r--lib/CodeGen/CodeGenModule.cpp46
-rw-r--r--lib/CodeGen/CodeGenModule.h29
-rw-r--r--test/CodeGen/c-strings.c34
7 files changed, 102 insertions, 32 deletions
diff --git a/lib/CodeGen/CGExpr.cpp b/lib/CodeGen/CGExpr.cpp
index 6bfd5cb309..e6bc5016e9 100644
--- a/lib/CodeGen/CGExpr.cpp
+++ b/lib/CodeGen/CGExpr.cpp
@@ -488,10 +488,7 @@ LValue CodeGenFunction::EmitUnaryOpLValue(const UnaryOperator *E) {
}
LValue CodeGenFunction::EmitStringLiteralLValue(const StringLiteral *E) {
- llvm::Constant *C =
- CGM.GetAddrOfConstantString(CGM.getStringForStringLiteral(E));
-
- return LValue::MakeAddr(C,0);
+ return LValue::MakeAddr(CGM.GetAddrOfConstantStringFromLiteral(E), 0);
}
LValue CodeGenFunction::EmitPredefinedLValue(const PredefinedExpr *E) {
diff --git a/lib/CodeGen/CGExprConstant.cpp b/lib/CodeGen/CGExprConstant.cpp
index 12620e7101..dfd047c89d 100644
--- a/lib/CodeGen/CGExprConstant.cpp
+++ b/lib/CodeGen/CGExprConstant.cpp
@@ -363,7 +363,7 @@ public:
// Otherwise this must be a string initializing an array in a static
// initializer. Don't emit it as the address of the string, emit the string
// data itself as an inline array.
- return llvm::ConstantArray::get(CGM.getStringForStringLiteral(E), false);
+ return llvm::ConstantArray::get(CGM.GetStringForStringLiteral(E), false);
}
llvm::Constant *VisitDeclRefExpr(DeclRefExpr *E) {
@@ -762,10 +762,8 @@ public:
"Taking the address of a vector component is illegal!");
return llvm::ConstantExpr::getGetElementPtr(Base, &Index, 1);
}
- case Expr::StringLiteralClass: {
- StringLiteral *S = cast<StringLiteral>(E);
- return CGM.GetAddrOfConstantString(CGM.getStringForStringLiteral(S));
- }
+ case Expr::StringLiteralClass:
+ return CGM.GetAddrOfConstantStringFromLiteral(cast<StringLiteral>(E));
case Expr::UnaryOperatorClass: {
UnaryOperator *Exp = cast<UnaryOperator>(E);
switch (Exp->getOpcode()) {
diff --git a/lib/CodeGen/CGObjC.cpp b/lib/CodeGen/CGObjC.cpp
index 7270884572..2e400f4d2a 100644
--- a/lib/CodeGen/CGObjC.cpp
+++ b/lib/CodeGen/CGObjC.cpp
@@ -52,7 +52,7 @@ llvm::Value *CodeGenFunction::EmitObjCMessageExpr(const ObjCMessageExpr *E) {
if (!strcmp(classname, "super")) {
classname = E->getMethodDecl()->getClassInterface()->getName();
}
- llvm::Value *ClassName = CGM.GetAddrOfConstantString(classname);
+ llvm::Value *ClassName = CGM.GetAddrOfConstantCString(classname);
ClassName = Builder.CreateStructGEP(ClassName, 0);
Receiver = Runtime.LookupClass(Builder, ClassName);
} else if (const PredefinedExpr *PDE =
diff --git a/lib/CodeGen/CGObjCGNU.cpp b/lib/CodeGen/CGObjCGNU.cpp
index 61552c4505..c4d236b54d 100644
--- a/lib/CodeGen/CGObjCGNU.cpp
+++ b/lib/CodeGen/CGObjCGNU.cpp
@@ -370,7 +370,7 @@ llvm::Constant *CGObjCGNU::GenerateMethodList(const std::string &ClassName,
std::vector<llvm::Constant*> Elements;
for (unsigned int i = 0, e = MethodTypes.size(); i < e; ++i) {
Elements.clear();
- llvm::Constant *C = CGM.GetAddrOfConstantString(MethodSels[i].getName());
+ llvm::Constant *C = CGM.GetAddrOfConstantCString(MethodSels[i].getName());
Elements.push_back(llvm::ConstantExpr::getGetElementPtr(C, Zeros, 2));
Elements.push_back(
llvm::ConstantExpr::getGetElementPtr(MethodTypes[i], Zeros, 2));
@@ -581,8 +581,8 @@ void CGObjCGNU::GenerateProtocol(const ObjCProtocolDecl *PD) {
std::string TypeStr;
Context.getObjCEncodingForMethodDecl(*iter, TypeStr);
InstanceMethodNames.push_back(
- CGM.GetAddrOfConstantString((*iter)->getSelector().getName()));
- InstanceMethodTypes.push_back(CGM.GetAddrOfConstantString(TypeStr));
+ CGM.GetAddrOfConstantCString((*iter)->getSelector().getName()));
+ InstanceMethodTypes.push_back(CGM.GetAddrOfConstantCString(TypeStr));
}
// Collect information about class methods:
llvm::SmallVector<llvm::Constant*, 16> ClassMethodNames;
@@ -592,8 +592,8 @@ void CGObjCGNU::GenerateProtocol(const ObjCProtocolDecl *PD) {
std::string TypeStr;
Context.getObjCEncodingForMethodDecl((*iter),TypeStr);
ClassMethodNames.push_back(
- CGM.GetAddrOfConstantString((*iter)->getSelector().getName()));
- ClassMethodTypes.push_back(CGM.GetAddrOfConstantString(TypeStr));
+ CGM.GetAddrOfConstantCString((*iter)->getSelector().getName()));
+ ClassMethodTypes.push_back(CGM.GetAddrOfConstantCString(TypeStr));
}
llvm::Constant *ProtocolList = GenerateProtocolList(Protocols);
diff --git a/lib/CodeGen/CodeGenModule.cpp b/lib/CodeGen/CodeGenModule.cpp
index f575300b6d..c6b02f83b4 100644
--- a/lib/CodeGen/CodeGenModule.cpp
+++ b/lib/CodeGen/CodeGenModule.cpp
@@ -261,7 +261,7 @@ void CodeGenModule::EmitObjCCategoryImpl(const ObjCCategoryImplDecl *OCD) {
InstanceMethodSels.push_back((*iter)->getSelector());
std::string TypeStr;
Context.getObjCEncodingForMethodDecl(*iter,TypeStr);
- InstanceMethodTypes.push_back(GetAddrOfConstantString(TypeStr));
+ InstanceMethodTypes.push_back(GetAddrOfConstantCString(TypeStr));
}
// Collect information about class methods
@@ -272,7 +272,7 @@ void CodeGenModule::EmitObjCCategoryImpl(const ObjCCategoryImplDecl *OCD) {
ClassMethodSels.push_back((*iter)->getSelector());
std::string TypeStr;
Context.getObjCEncodingForMethodDecl(*iter,TypeStr);
- ClassMethodTypes.push_back(GetAddrOfConstantString(TypeStr));
+ ClassMethodTypes.push_back(GetAddrOfConstantCString(TypeStr));
}
// Collect the names of referenced protocols
@@ -325,13 +325,13 @@ void CodeGenModule::EmitObjCClassImplementation(
for (ObjCInterfaceDecl::ivar_iterator iter = ClassDecl->ivar_begin(),
endIter = ClassDecl->ivar_end() ; iter != endIter ; iter++) {
// Store the name
- IvarNames.push_back(GetAddrOfConstantString((*iter)->getName()));
+ IvarNames.push_back(GetAddrOfConstantCString((*iter)->getName()));
// Get the type encoding for this ivar
std::string TypeStr;
llvm::SmallVector<const RecordType *, 8> EncodingRecordTypes;
Context.getObjCEncodingForType((*iter)->getType(), TypeStr,
EncodingRecordTypes);
- IvarTypes.push_back(GetAddrOfConstantString(TypeStr));
+ IvarTypes.push_back(GetAddrOfConstantCString(TypeStr));
// Get the offset
int offset =
(int)Layout->getElementOffset(getTypes().getLLVMFieldNo(*iter));
@@ -347,7 +347,7 @@ void CodeGenModule::EmitObjCClassImplementation(
InstanceMethodSels.push_back((*iter)->getSelector());
std::string TypeStr;
Context.getObjCEncodingForMethodDecl((*iter),TypeStr);
- InstanceMethodTypes.push_back(GetAddrOfConstantString(TypeStr));
+ InstanceMethodTypes.push_back(GetAddrOfConstantCString(TypeStr));
}
// Collect information about class methods
@@ -358,7 +358,7 @@ void CodeGenModule::EmitObjCClassImplementation(
ClassMethodSels.push_back((*iter)->getSelector());
std::string TypeStr;
Context.getObjCEncodingForMethodDecl((*iter),TypeStr);
- ClassMethodTypes.push_back(GetAddrOfConstantString(TypeStr));
+ ClassMethodTypes.push_back(GetAddrOfConstantCString(TypeStr));
}
// Collect the names of referenced protocols
llvm::SmallVector<std::string, 16> Protocols;
@@ -888,9 +888,9 @@ GetAddrOfConstantCFString(const std::string &str) {
return GV;
}
-/// getStringForStringLiteral - Return the appropriate bytes for a
+/// GetStringForStringLiteral - Return the appropriate bytes for a
/// string literal, properly padded to match the literal type.
-std::string CodeGenModule::getStringForStringLiteral(const StringLiteral *E) {
+std::string CodeGenModule::GetStringForStringLiteral(const StringLiteral *E) {
assert(!E->isWide() && "FIXME: Wide strings not supported yet!");
const char *StrData = E->getStrData();
unsigned Len = E->getByteLength();
@@ -908,22 +908,37 @@ std::string CodeGenModule::getStringForStringLiteral(const StringLiteral *E) {
return Str;
}
+/// GetAddrOfConstantStringFromLiteral - Return a pointer to a
+/// constant array for the given string literal.
+llvm::Constant *
+CodeGenModule::GetAddrOfConstantStringFromLiteral(const StringLiteral *S) {
+ // FIXME: This can be more efficient.
+ return GetAddrOfConstantString(GetStringForStringLiteral(S));
+}
+
/// GenerateWritableString -- Creates storage for a string literal.
static llvm::Constant *GenerateStringLiteral(const std::string &str,
bool constant,
CodeGenModule &CGM) {
- // Create Constant for this string literal
- llvm::Constant *C = llvm::ConstantArray::get(str);
+ // Create Constant for this string literal. Don't add a '\0'.
+ llvm::Constant *C = llvm::ConstantArray::get(str, false);
// Create a global variable for this string
C = new llvm::GlobalVariable(C->getType(), constant,
llvm::GlobalValue::InternalLinkage,
C, ".str", &CGM.getModule());
+
return C;
}
-/// CodeGenModule::GetAddrOfConstantString -- returns a pointer to the character
-/// array containing the literal. The result is pointer to array type.
+/// GetAddrOfConstantString - Returns a pointer to a character array
+/// containing the literal. This contents are exactly that of the
+/// given string, i.e. it will not be null terminated automatically;
+/// see GetAddrOfConstantCString. Note that whether the result is
+/// actually a pointer to an LLVM constant depends on
+/// Feature.WriteableStrings.
+///
+/// The result has pointer to array type.
llvm::Constant *CodeGenModule::GetAddrOfConstantString(const std::string &str) {
// Don't share any string literals if writable-strings is turned on.
if (Features.WritableStrings)
@@ -940,3 +955,10 @@ llvm::Constant *CodeGenModule::GetAddrOfConstantString(const std::string &str) {
Entry.setValue(C);
return C;
}
+
+/// GetAddrOfConstantCString - Returns a pointer to a character
+/// array containing the literal and a terminating '\-'
+/// character. The result has pointer to array type.
+llvm::Constant *CodeGenModule::GetAddrOfConstantCString(const std::string &str) {
+ return GetAddrOfConstantCString(str + "\0");
+}
diff --git a/lib/CodeGen/CodeGenModule.h b/lib/CodeGen/CodeGenModule.h
index a837b6369c..dd8ac1739f 100644
--- a/lib/CodeGen/CodeGenModule.h
+++ b/lib/CodeGen/CodeGenModule.h
@@ -143,15 +143,34 @@ public:
/// "__builtin_fabsf", return a Function* for "fabsf".
///
llvm::Function *getBuiltinLibFunction(unsigned BuiltinID);
+
+ /// GetStringForStringLiteral - Return the appropriate bytes for a
+ /// string literal, properly padded to match the literal type. If
+ /// only the address of a constant is needed consider using
+ /// GetAddrOfConstantStringLiteral.
+ std::string GetStringForStringLiteral(const StringLiteral *E);
+
llvm::Constant *GetAddrOfConstantCFString(const std::string& str);
- /// getStringForStringLiteral - Return the appropriate bytes for a
- /// string literal, properly padded to match the literal type.
- std::string getStringForStringLiteral(const StringLiteral *E);
+ /// GetAddrOfConstantStringFromLiteral - Return a pointer to a
+ /// constant array for the given string literal.
+ llvm::Constant *GetAddrOfConstantStringFromLiteral(const StringLiteral *S);
- /// GetAddrOfConstantString -- returns a pointer to the character
- /// array containing the literal. The result is pointer to array type.
+ /// GetAddrOfConstantString - Returns a pointer to a character array
+ /// containing the literal. This contents are exactly that of the
+ /// given string, i.e. it will not be null terminated automatically;
+ /// see GetAddrOfConstantCString. Note that whether the result is
+ /// actually a pointer to an LLVM constant depends on
+ /// Feature.WriteableStrings.
+ ///
+ /// The result has pointer to array type.
llvm::Constant *GetAddrOfConstantString(const std::string& str);
+
+ /// GetAddrOfConstantCString - Returns a pointer to a character
+ /// array containing the literal and a terminating '\-'
+ /// character. The result has pointer to array type.
+ llvm::Constant *GetAddrOfConstantCString(const std::string &str);
+
llvm::Function *getMemCpyFn();
llvm::Function *getMemMoveFn();
llvm::Function *getMemSetFn();
diff --git a/test/CodeGen/c-strings.c b/test/CodeGen/c-strings.c
new file mode 100644
index 0000000000..baff4ddfc1
--- /dev/null
+++ b/test/CodeGen/c-strings.c
@@ -0,0 +1,34 @@
+// RUN: clang -emit-llvm -o %t %s &&
+// RUN: grep "hello" %t | count 3 &&
+// RUN: grep 'c"hello\\00"' %t | count 2 &&
+// RUN: grep 'c"hello\\00\\00\\00"' %t | count 1
+
+/* Should be 3 hello string, two global (of different sizes), the rest
+ are shared. */
+
+void f0() {
+ bar("hello");
+}
+
+void f1() {
+ static char *x = "hello";
+ bar(x);
+}
+
+void f2() {
+ static char x[] = "hello";
+ bar(x);
+}
+
+void f3() {
+ static char x[8] = "hello";
+ bar(x);
+}
+
+void f4() {
+ static struct s {
+ char *name;
+ } x = { "hello" };
+ gaz(&x);
+}
+
generated by cgit v1.2.3-18-g5258 (git 2.32.0) at 2025-07-01 18:58:28 +0000