[clang] Fix handling of asm string literals (PR #169803)
Abhina Sree via cfe-commits
cfe-commits at lists.llvm.org
Tue Aug 11 06:03:32 PDT 2026
https://github.com/abhina-sree updated https://github.com/llvm/llvm-project/pull/169803
>From e3c12254d60b1448d3051a20f60fdecafa161eb4 Mon Sep 17 00:00:00 2001
From: Abhina Sreeskantharajan <Abhina.Sreeskantharajan at ibm.com>
Date: Fri, 8 May 2026 12:17:22 -0400
Subject: [PATCH 1/6] add ParserConversionAction, do not translate unevaluated
strings
---
clang/include/clang/Parse/Parser.h | 1 +
clang/include/clang/Sema/Sema.h | 4 +++-
clang/lib/Parse/ParseDecl.cpp | 9 +++++++++
clang/lib/Parse/ParseExpr.cpp | 6 +++---
clang/lib/Parse/Parser.cpp | 4 ++++
clang/lib/Sema/SemaExpr.cpp | 8 ++++----
clang/test/CodeGen/systemz-charset-diag.cpp | 8 ++++++++
clang/test/CodeGen/systemz-charset.c | 5 +++++
8 files changed, 37 insertions(+), 8 deletions(-)
diff --git a/clang/include/clang/Parse/Parser.h b/clang/include/clang/Parse/Parser.h
index f0e06473bf615..889df34ab9bfb 100644
--- a/clang/include/clang/Parse/Parser.h
+++ b/clang/include/clang/Parse/Parser.h
@@ -5706,6 +5706,7 @@ class Parser : public CodeCompletionHandler {
bool Finished;
};
ObjCImplParsingDataRAII *CurParsedObjCImpl;
+ ConversionAction ParserConversionAction;
/// StashAwayMethodOrFunctionBodyTokens - Consume the tokens and store them
/// for later parsing.
diff --git a/clang/include/clang/Sema/Sema.h b/clang/include/clang/Sema/Sema.h
index f7d0d493e7081..51e819840c7a4 100644
--- a/clang/include/clang/Sema/Sema.h
+++ b/clang/include/clang/Sema/Sema.h
@@ -55,6 +55,7 @@
#include "clang/Basic/TemplateKinds.h"
#include "clang/Basic/TokenKinds.h"
#include "clang/Basic/TypeTraits.h"
+#include "clang/Lex/TextEncoding.h"
#include "clang/Sema/AnalysisBasedWarnings.h"
#include "clang/Sema/Attr.h"
#include "clang/Sema/CleanupInfo.h"
@@ -7379,7 +7380,8 @@ class Sema final : public SemaBase {
/// from multiple tokens. However, the common case is that StringToks points
/// to one string.
ExprResult ActOnStringLiteral(ArrayRef<Token> StringToks,
- Scope *UDLScope = nullptr);
+ Scope *UDLScope = nullptr,
+ ConversionAction Action = CA_ToLiteralEncoding);
ExprResult ActOnUnevaluatedStringLiteral(ArrayRef<Token> StringToks);
diff --git a/clang/lib/Parse/ParseDecl.cpp b/clang/lib/Parse/ParseDecl.cpp
index 3f41e7c5c6f0d..cbf4b11a719ca 100644
--- a/clang/lib/Parse/ParseDecl.cpp
+++ b/clang/lib/Parse/ParseDecl.cpp
@@ -564,6 +564,9 @@ unsigned Parser::ParseAttributeArgsCommon(
nullptr,
Sema::ExpressionEvaluationContextRecord::EK_AttrArgument);
+ SaveAndRestore<ConversionAction> SavedTranslationState(
+ ParserConversionAction, CA_NoConversion);
+
ExprResult ArgExpr = ParseAssignmentExpression();
if (ArgExpr.isInvalid()) {
SkipUntil(tok::r_paren, StopAtSemi);
@@ -644,6 +647,9 @@ void Parser::ParseGNUAttributeArgs(
ParsedAttr::Kind AttrKind =
ParsedAttr::getParsedKind(AttrName, ScopeName, Form.getSyntax());
+ SaveAndRestore<ConversionAction> SavedTranslationState(ParserConversionAction,
+ CA_NoConversion);
+
if (AttrKind == ParsedAttr::AT_Availability) {
ParseAvailabilityAttribute(*AttrName, AttrNameLoc, Attrs, EndLoc, ScopeName,
ScopeLoc, Form);
@@ -723,6 +729,9 @@ unsigned Parser::ParseClangAttributeArgs(
ParsedAttr::Kind AttrKind =
ParsedAttr::getParsedKind(AttrName, ScopeName, Form.getSyntax());
+ SaveAndRestore<ConversionAction> SavedTranslationState(ParserConversionAction,
+ CA_NoConversion);
+
switch (AttrKind) {
default:
return ParseAttributeArgsCommon(AttrName, AttrNameLoc, Attrs, EndLoc,
diff --git a/clang/lib/Parse/ParseExpr.cpp b/clang/lib/Parse/ParseExpr.cpp
index 451287bbeee3f..bff32eed404f1 100644
--- a/clang/lib/Parse/ParseExpr.cpp
+++ b/clang/lib/Parse/ParseExpr.cpp
@@ -3041,9 +3041,9 @@ ExprResult Parser::ParseStringLiteralExpression(bool AllowUserDefinedLiteral,
}
// Pass the set of string tokens, ready for concatenation, to the actions.
- return Actions.ActOnStringLiteral(StringToks,
- AllowUserDefinedLiteral ? getCurScope()
- : nullptr);
+ return Actions.ActOnStringLiteral(
+ StringToks, AllowUserDefinedLiteral ? getCurScope() : nullptr,
+ ParserConversionAction);
}
ExprResult Parser::ParseGenericSelectionExpression() {
diff --git a/clang/lib/Parse/Parser.cpp b/clang/lib/Parse/Parser.cpp
index 5e1fd4df1a3f0..5a199b842fe8e 100644
--- a/clang/lib/Parse/Parser.cpp
+++ b/clang/lib/Parse/Parser.cpp
@@ -70,6 +70,8 @@ Parser::Parser(Preprocessor &pp, Sema &actions, bool skipFunctionBodies)
NumCachedScopes = 0;
CurParsedObjCImpl = nullptr;
+ ParserConversionAction = CA_ToLiteralEncoding;
+
// Add #pragma handlers. These are removed and destroyed in the
// destructor.
initializePragmaHandlers();
@@ -1551,6 +1553,8 @@ void Parser::ParseKNRParamDeclarations(Declarator &D) {
}
ExprResult Parser::ParseAsmStringLiteral(bool ForAsmLabel) {
+ SaveAndRestore<ConversionAction> SavedTranslationState(ParserConversionAction,
+ CA_NoConversion);
ExprResult AsmString;
if (isTokenStringLiteral()) {
diff --git a/clang/lib/Sema/SemaExpr.cpp b/clang/lib/Sema/SemaExpr.cpp
index 592cb1d588370..ed4fcd3f975dc 100644
--- a/clang/lib/Sema/SemaExpr.cpp
+++ b/clang/lib/Sema/SemaExpr.cpp
@@ -2237,8 +2237,8 @@ Sema::ExpandFunctionLocalPredefinedMacros(ArrayRef<Token> Toks) {
return ExpandedToks;
}
-ExprResult
-Sema::ActOnStringLiteral(ArrayRef<Token> StringToks, Scope *UDLScope) {
+ExprResult Sema::ActOnStringLiteral(ArrayRef<Token> StringToks, Scope *UDLScope,
+ ConversionAction Action) {
assert(!StringToks.empty() && "Must have at least one string!");
// StringToks needs backing storage as it doesn't hold array elements itself
@@ -2246,8 +2246,8 @@ Sema::ActOnStringLiteral(ArrayRef<Token> StringToks, Scope *UDLScope) {
if (getLangOpts().MicrosoftExt)
StringToks = ExpandedToks = ExpandFunctionLocalPredefinedMacros(StringToks);
- StringLiteralParser Literal(
- StringToks, PP, StringLiteralEvalMethod::Evaluated, CA_ToLiteralEncoding);
+ StringLiteralParser Literal(StringToks, PP,
+ StringLiteralEvalMethod::Evaluated, Action);
if (Literal.hadError)
return ExprError();
diff --git a/clang/test/CodeGen/systemz-charset-diag.cpp b/clang/test/CodeGen/systemz-charset-diag.cpp
index 4ed94810150a3..ad08e1f391214 100644
--- a/clang/test/CodeGen/systemz-charset-diag.cpp
+++ b/clang/test/CodeGen/systemz-charset-diag.cpp
@@ -1,3 +1,11 @@
// RUN: %clang_cc1 -triple s390x-none-zos -fexec-charset IBM-1047 %s -std=c++17 -emit-llvm -o - -verify
const char* Computer = "🖥️"; // expected-error-re {{conversion to literal encoding failed: {{.*}}}}
+
+static_assert(false, "Error string"); // expected-error {{static assertion failed: Error string}}
+
+[[deprecated("message")]] void test_deprecated() {return;} // expected-note {{'test_deprecated' has been explicitly marked deprecated here}}
+
+int main() {
+ test_deprecated(); // expected-warning {{'test_deprecated' is deprecated: message}}
+}
diff --git a/clang/test/CodeGen/systemz-charset.c b/clang/test/CodeGen/systemz-charset.c
index 766b6a83f00ff..e0d1736bf853f 100644
--- a/clang/test/CodeGen/systemz-charset.c
+++ b/clang/test/CodeGen/systemz-charset.c
@@ -66,3 +66,8 @@ const char* hello_macro = HELLO;
const char* preprocessor_concatenation = HELLO_WORLD;
//CHECK: c"\C8\85\93\93\96@\E6\96\99\93\84Z\00"
//CHECK-UTF8: c"Hello World!\00"
+
+void asm_labeled_fn(void) __asm__("asm_sym");
+void asm_labeled_fn(void) {}
+//CHECK: define{{.*}} void @asm_sym(
+//CHECK-UTF8: define{{.*}} void @asm_sym(
>From b74e511f00d5b6ae75a408cbb4c923089701d050 Mon Sep 17 00:00:00 2001
From: Abhina Sreeskantharajan <Abhina.Sreeskantharajan at ibm.com>
Date: Fri, 3 Jul 2026 14:52:22 -0400
Subject: [PATCH 2/6] remove unnecessary changes
---
clang/lib/Parse/ParseDecl.cpp | 9 ---------
1 file changed, 9 deletions(-)
diff --git a/clang/lib/Parse/ParseDecl.cpp b/clang/lib/Parse/ParseDecl.cpp
index cbf4b11a719ca..3f41e7c5c6f0d 100644
--- a/clang/lib/Parse/ParseDecl.cpp
+++ b/clang/lib/Parse/ParseDecl.cpp
@@ -564,9 +564,6 @@ unsigned Parser::ParseAttributeArgsCommon(
nullptr,
Sema::ExpressionEvaluationContextRecord::EK_AttrArgument);
- SaveAndRestore<ConversionAction> SavedTranslationState(
- ParserConversionAction, CA_NoConversion);
-
ExprResult ArgExpr = ParseAssignmentExpression();
if (ArgExpr.isInvalid()) {
SkipUntil(tok::r_paren, StopAtSemi);
@@ -647,9 +644,6 @@ void Parser::ParseGNUAttributeArgs(
ParsedAttr::Kind AttrKind =
ParsedAttr::getParsedKind(AttrName, ScopeName, Form.getSyntax());
- SaveAndRestore<ConversionAction> SavedTranslationState(ParserConversionAction,
- CA_NoConversion);
-
if (AttrKind == ParsedAttr::AT_Availability) {
ParseAvailabilityAttribute(*AttrName, AttrNameLoc, Attrs, EndLoc, ScopeName,
ScopeLoc, Form);
@@ -729,9 +723,6 @@ unsigned Parser::ParseClangAttributeArgs(
ParsedAttr::Kind AttrKind =
ParsedAttr::getParsedKind(AttrName, ScopeName, Form.getSyntax());
- SaveAndRestore<ConversionAction> SavedTranslationState(ParserConversionAction,
- CA_NoConversion);
-
switch (AttrKind) {
default:
return ParseAttributeArgsCommon(AttrName, AttrNameLoc, Attrs, EndLoc,
>From 15bb83c2a3837db0d40dc520fca8e5ec3a76ceda Mon Sep 17 00:00:00 2001
From: Abhina Sreeskantharajan <Abhina.Sreeskantharajan at ibm.com>
Date: Thu, 9 Jul 2026 13:25:39 -0400
Subject: [PATCH 3/6] remove ParserConversionAction, add paramter to
ParseStringLiteralExpression instead, add conversion to SystemCharset then
back to UTF8 to handle escape sequences in system encoding
---
clang/include/clang/Basic/TargetInfo.h | 3 ++
clang/include/clang/Lex/TextEncoding.h | 11 ++++++--
clang/include/clang/Parse/Parser.h | 6 ++--
clang/lib/Basic/TargetInfo.cpp | 2 ++
clang/lib/Frontend/CompilerInstance.cpp | 4 +--
clang/lib/Lex/TextEncoding.cpp | 37 +++++++++++++++++++++----
clang/lib/Parse/ParseExpr.cpp | 6 ++--
clang/lib/Parse/Parser.cpp | 8 ++----
clang/lib/Sema/SemaStmtAsm.cpp | 8 ++++++
clang/test/CodeGen/systemz-charset.c | 6 ++--
10 files changed, 67 insertions(+), 24 deletions(-)
diff --git a/clang/include/clang/Basic/TargetInfo.h b/clang/include/clang/Basic/TargetInfo.h
index ad735d7b0b0e5..d1188e2ca5434 100644
--- a/clang/include/clang/Basic/TargetInfo.h
+++ b/clang/include/clang/Basic/TargetInfo.h
@@ -38,6 +38,7 @@
#include "llvm/IR/DerivedTypes.h"
#include "llvm/Support/DataTypes.h"
#include "llvm/Support/Error.h"
+#include "llvm/Support/TextEncoding.h"
#include "llvm/Support/VersionTuple.h"
#include "llvm/TargetParser/Triple.h"
#include <cassert>
@@ -323,6 +324,8 @@ class TargetInfo : public TransferrableTargetInfo,
virtual ~TargetInfo();
+ std::unique_ptr<llvm::TextEncodingConverter> TargetStrConverter;
+
/// Retrieve the target options.
TargetOptions &getTargetOpts() const {
assert(TargetOpts && "Missing target options");
diff --git a/clang/include/clang/Lex/TextEncoding.h b/clang/include/clang/Lex/TextEncoding.h
index 55663cdc072da..02cb16cef9f1b 100644
--- a/clang/include/clang/Lex/TextEncoding.h
+++ b/clang/include/clang/Lex/TextEncoding.h
@@ -10,6 +10,7 @@
#define LLVM_CLANG_LEX_TEXTENCODING_H
#include "clang/Basic/LangOptions.h"
+#include "clang/Basic/TargetInfo.h"
#include "llvm/ADT/StringRef.h"
namespace llvm {
@@ -17,16 +18,22 @@ class TextEncodingConverter;
} // namespace llvm
namespace clang {
-enum ConversionAction { CA_NoConversion, CA_ToLiteralEncoding };
+enum ConversionAction {
+ CA_NoConversion,
+ CA_ToSystemEncoding,
+ CA_ToLiteralEncoding
+};
class TextEncoding {
llvm::StringRef LiteralEncoding;
std::unique_ptr<llvm::TextEncodingConverter> ToLiteralEncodingConverter;
+ std::unique_ptr<llvm::TextEncodingConverter> ToSystemEncodingConverter;
public:
llvm::TextEncodingConverter *getConverter(ConversionAction Action) const;
static std::error_code
- setConvertersFromOptions(TextEncoding &TE, const clang::LangOptions &Opts);
+ setConvertersFromOptions(TextEncoding &TE, const clang::LangOptions &Opts,
+ clang::TargetInfo &TInfo);
llvm::StringRef getLiteralEncoding() { return LiteralEncoding; }
};
diff --git a/clang/include/clang/Parse/Parser.h b/clang/include/clang/Parse/Parser.h
index 889df34ab9bfb..019316296f232 100644
--- a/clang/include/clang/Parse/Parser.h
+++ b/clang/include/clang/Parse/Parser.h
@@ -3866,8 +3866,9 @@ class Parser : public CodeCompletionHandler {
/// associated with the class's definition.
void PopParsingClass(Sema::ParsingClassState);
- ExprResult ParseStringLiteralExpression(bool AllowUserDefinedLiteral,
- bool Unevaluated);
+ ExprResult
+ ParseStringLiteralExpression(bool AllowUserDefinedLiteral, bool Unevaluated,
+ ConversionAction CA = CA_ToLiteralEncoding);
/// This routine is called when the '@' is seen and consumed.
/// Current token is an Identifier and is not a 'try'. This
@@ -5706,7 +5707,6 @@ class Parser : public CodeCompletionHandler {
bool Finished;
};
ObjCImplParsingDataRAII *CurParsedObjCImpl;
- ConversionAction ParserConversionAction;
/// StashAwayMethodOrFunctionBodyTokens - Consume the tokens and store them
/// for later parsing.
diff --git a/clang/lib/Basic/TargetInfo.cpp b/clang/lib/Basic/TargetInfo.cpp
index 9a25384347073..e58a0d73c429a 100644
--- a/clang/lib/Basic/TargetInfo.cpp
+++ b/clang/lib/Basic/TargetInfo.cpp
@@ -194,6 +194,8 @@ TargetInfo::TargetInfo(const llvm::Triple &T) : Triple(T) {
MaxOpenCLWorkGroupSize = 1024;
MaxBitIntWidth.reset();
+
+ TargetStrConverter = nullptr;
}
// Out of line virtual dtor for TargetInfo.
diff --git a/clang/lib/Frontend/CompilerInstance.cpp b/clang/lib/Frontend/CompilerInstance.cpp
index 941c9e40f0c1d..a6bc432efdf8b 100644
--- a/clang/lib/Frontend/CompilerInstance.cpp
+++ b/clang/lib/Frontend/CompilerInstance.cpp
@@ -561,8 +561,8 @@ void CompilerInstance::createPreprocessor(TranslationUnitKind TUKind) {
if (GetDependencyDirectives)
PP->setDependencyDirectivesGetter(*GetDependencyDirectives);
- if (auto EC = TextEncoding::setConvertersFromOptions(PP->getTextEncoding(),
- getLangOpts()))
+ if (auto EC = TextEncoding::setConvertersFromOptions(
+ PP->getTextEncoding(), getLangOpts(), getTarget()))
PP->getDiagnostics().Report(clang::diag::err_fe_text_encoding_config)
<< PP->getTextEncoding().getLiteralEncoding();
}
diff --git a/clang/lib/Lex/TextEncoding.cpp b/clang/lib/Lex/TextEncoding.cpp
index eb4fde8928a68..effd05185faf7 100644
--- a/clang/lib/Lex/TextEncoding.cpp
+++ b/clang/lib/Lex/TextEncoding.cpp
@@ -17,6 +17,8 @@ TextEncoding::getConverter(ConversionAction Action) const {
switch (Action) {
case CA_ToLiteralEncoding:
return ToLiteralEncodingConverter.get();
+ case CA_ToSystemEncoding:
+ return ToSystemEncodingConverter.get();
default:
return nullptr;
}
@@ -24,23 +26,46 @@ TextEncoding::getConverter(ConversionAction Action) const {
std::error_code
TextEncoding::setConvertersFromOptions(TextEncoding &TE,
- const clang::LangOptions &Opts) {
+ const clang::LangOptions &Opts,
+ clang::TargetInfo &TInfo) {
using namespace llvm;
const char *UTF8 = "UTF-8";
TE.LiteralEncoding =
Opts.LiteralEncoding.empty() ? UTF8 : Opts.LiteralEncoding.c_str();
- // Create converter between internal and literal encoding specified
- // in fexec-charset option.
- if (TE.LiteralEncoding == UTF8)
+ if (TE.LiteralEncoding != UTF8) {
+ ErrorOr<TextEncodingConverter> ErrorOrLiteralConverter =
+ llvm::TextEncodingConverter::create(UTF8, TE.LiteralEncoding);
+ if (ErrorOrLiteralConverter)
+ TE.ToLiteralEncodingConverter = std::make_unique<TextEncodingConverter>(
+ std::move(*ErrorOrLiteralConverter));
+ else
+ return ErrorOrLiteralConverter.getError();
+ }
+
+ if (TInfo.getDefaultOrdinaryLiteralEncoding() == UTF8)
return std::error_code();
+
+ // Create converter between internal and default ordinary encoding for the
+ // target
ErrorOr<TextEncodingConverter> ErrorOrConverter =
- llvm::TextEncodingConverter::create(UTF8, TE.LiteralEncoding);
+ llvm::TextEncodingConverter::create(
+ UTF8, TInfo.getDefaultOrdinaryLiteralEncoding());
+ if (ErrorOrConverter)
+ TE.ToSystemEncodingConverter =
+ std::make_unique<TextEncodingConverter>(std::move(*ErrorOrConverter));
+ else
+ return ErrorOrConverter.getError();
+
+ ErrorOrConverter = llvm::TextEncodingConverter::create(
+ TInfo.getDefaultOrdinaryLiteralEncoding(), UTF8);
+
if (ErrorOrConverter)
- TE.ToLiteralEncodingConverter =
+ TInfo.TargetStrConverter =
std::make_unique<TextEncodingConverter>(std::move(*ErrorOrConverter));
else
return ErrorOrConverter.getError();
+
return std::error_code();
}
diff --git a/clang/lib/Parse/ParseExpr.cpp b/clang/lib/Parse/ParseExpr.cpp
index bff32eed404f1..642de927f750f 100644
--- a/clang/lib/Parse/ParseExpr.cpp
+++ b/clang/lib/Parse/ParseExpr.cpp
@@ -3021,7 +3021,8 @@ ExprResult Parser::ParseUnevaluatedStringLiteralExpression() {
}
ExprResult Parser::ParseStringLiteralExpression(bool AllowUserDefinedLiteral,
- bool Unevaluated) {
+ bool Unevaluated,
+ ConversionAction CA) {
assert(tokenIsLikeStringLiteral(Tok, getLangOpts()) &&
"Not a string-literal-like token!");
@@ -3042,8 +3043,7 @@ ExprResult Parser::ParseStringLiteralExpression(bool AllowUserDefinedLiteral,
// Pass the set of string tokens, ready for concatenation, to the actions.
return Actions.ActOnStringLiteral(
- StringToks, AllowUserDefinedLiteral ? getCurScope() : nullptr,
- ParserConversionAction);
+ StringToks, AllowUserDefinedLiteral ? getCurScope() : nullptr, CA);
}
ExprResult Parser::ParseGenericSelectionExpression() {
diff --git a/clang/lib/Parse/Parser.cpp b/clang/lib/Parse/Parser.cpp
index 5a199b842fe8e..419db5963077c 100644
--- a/clang/lib/Parse/Parser.cpp
+++ b/clang/lib/Parse/Parser.cpp
@@ -70,8 +70,6 @@ Parser::Parser(Preprocessor &pp, Sema &actions, bool skipFunctionBodies)
NumCachedScopes = 0;
CurParsedObjCImpl = nullptr;
- ParserConversionAction = CA_ToLiteralEncoding;
-
// Add #pragma handlers. These are removed and destroyed in the
// destructor.
initializePragmaHandlers();
@@ -1553,12 +1551,12 @@ void Parser::ParseKNRParamDeclarations(Declarator &D) {
}
ExprResult Parser::ParseAsmStringLiteral(bool ForAsmLabel) {
- SaveAndRestore<ConversionAction> SavedTranslationState(ParserConversionAction,
- CA_NoConversion);
ExprResult AsmString;
if (isTokenStringLiteral()) {
- AsmString = ParseStringLiteralExpression();
+ AsmString = ParseStringLiteralExpression(/*AllowUserDefinedLiteral=*/false,
+ /*Unevaluated=*/false,
+ CA_ToSystemEncoding);
if (AsmString.isInvalid())
return AsmString;
diff --git a/clang/lib/Sema/SemaStmtAsm.cpp b/clang/lib/Sema/SemaStmtAsm.cpp
index 6bf12d9cd98da..f1aab455b4411 100644
--- a/clang/lib/Sema/SemaStmtAsm.cpp
+++ b/clang/lib/Sema/SemaStmtAsm.cpp
@@ -245,6 +245,14 @@ ExprResult Sema::ActOnGCCAsmStmtString(Expr *Expr, bool ForAsmLabel) {
Diag(Expr->getBeginLoc(), diag::err_asm_operand_empty_string)
<< SL->getSourceRange();
}
+ if (Context.getTargetInfo().TargetStrConverter) {
+ SmallString<16> ConvertedAsm;
+ Context.getTargetInfo().TargetStrConverter->convert(SL->getString(),
+ ConvertedAsm);
+ return StringLiteral::Create(Context, ConvertedAsm,
+ StringLiteralKind::Ordinary,
+ /*Pascal*/ false, {}, SL->getBeginLoc());
+ }
return SL;
}
if (DiagnoseUnexpandedParameterPack(Expr))
diff --git a/clang/test/CodeGen/systemz-charset.c b/clang/test/CodeGen/systemz-charset.c
index e0d1736bf853f..f3bb1055ddf5e 100644
--- a/clang/test/CodeGen/systemz-charset.c
+++ b/clang/test/CodeGen/systemz-charset.c
@@ -67,7 +67,7 @@ const char* preprocessor_concatenation = HELLO_WORLD;
//CHECK: c"\C8\85\93\93\96@\E6\96\99\93\84Z\00"
//CHECK-UTF8: c"Hello World!\00"
-void asm_labeled_fn(void) __asm__("asm_sym");
+void asm_labeled_fn(void) __asm__("\174asm_sym");
void asm_labeled_fn(void) {}
-//CHECK: define{{.*}} void @asm_sym(
-//CHECK-UTF8: define{{.*}} void @asm_sym(
+//CHECK: define{{.*}}void @"@asm_sym"(
+//CHECK-UTF8: define{{.*}}void @"@asm_sym"(
>From 593afe501fb24e039d952f6138806948f3c4dbb6 Mon Sep 17 00:00:00 2001
From: Abhina Sreeskantharajan <Abhina.Sreeskantharajan at ibm.com>
Date: Fri, 10 Jul 2026 15:07:07 -0400
Subject: [PATCH 4/6] address comments, attempt to fix CI
---
clang/include/clang/Basic/TargetInfo.h | 2 +-
clang/lib/Basic/TargetInfo.cpp | 2 +-
clang/lib/Lex/TextEncoding.cpp | 2 +-
clang/lib/Sema/SemaStmtAsm.cpp | 10 ++++++----
clang/test/CodeGen/systemz-charset.c | 4 ++++
5 files changed, 13 insertions(+), 7 deletions(-)
diff --git a/clang/include/clang/Basic/TargetInfo.h b/clang/include/clang/Basic/TargetInfo.h
index d1188e2ca5434..fb0e699234379 100644
--- a/clang/include/clang/Basic/TargetInfo.h
+++ b/clang/include/clang/Basic/TargetInfo.h
@@ -324,7 +324,7 @@ class TargetInfo : public TransferrableTargetInfo,
virtual ~TargetInfo();
- std::unique_ptr<llvm::TextEncodingConverter> TargetStrConverter;
+ std::unique_ptr<llvm::TextEncodingConverter> FromSystemEncodingConverter;
/// Retrieve the target options.
TargetOptions &getTargetOpts() const {
diff --git a/clang/lib/Basic/TargetInfo.cpp b/clang/lib/Basic/TargetInfo.cpp
index e58a0d73c429a..c318f2cfc40b3 100644
--- a/clang/lib/Basic/TargetInfo.cpp
+++ b/clang/lib/Basic/TargetInfo.cpp
@@ -195,7 +195,7 @@ TargetInfo::TargetInfo(const llvm::Triple &T) : Triple(T) {
MaxBitIntWidth.reset();
- TargetStrConverter = nullptr;
+ FromSystemEncodingConverter = nullptr;
}
// Out of line virtual dtor for TargetInfo.
diff --git a/clang/lib/Lex/TextEncoding.cpp b/clang/lib/Lex/TextEncoding.cpp
index effd05185faf7..fd59f3df59bdf 100644
--- a/clang/lib/Lex/TextEncoding.cpp
+++ b/clang/lib/Lex/TextEncoding.cpp
@@ -62,7 +62,7 @@ TextEncoding::setConvertersFromOptions(TextEncoding &TE,
TInfo.getDefaultOrdinaryLiteralEncoding(), UTF8);
if (ErrorOrConverter)
- TInfo.TargetStrConverter =
+ TInfo.FromSystemEncodingConverter =
std::make_unique<TextEncodingConverter>(std::move(*ErrorOrConverter));
else
return ErrorOrConverter.getError();
diff --git a/clang/lib/Sema/SemaStmtAsm.cpp b/clang/lib/Sema/SemaStmtAsm.cpp
index f1aab455b4411..8b61228db3533 100644
--- a/clang/lib/Sema/SemaStmtAsm.cpp
+++ b/clang/lib/Sema/SemaStmtAsm.cpp
@@ -245,13 +245,15 @@ ExprResult Sema::ActOnGCCAsmStmtString(Expr *Expr, bool ForAsmLabel) {
Diag(Expr->getBeginLoc(), diag::err_asm_operand_empty_string)
<< SL->getSourceRange();
}
- if (Context.getTargetInfo().TargetStrConverter) {
+ if (Context.getTargetInfo().FromSystemEncodingConverter) {
SmallString<16> ConvertedAsm;
- Context.getTargetInfo().TargetStrConverter->convert(SL->getString(),
- ConvertedAsm);
+ Context.getTargetInfo().FromSystemEncodingConverter->convert(
+ SL->getString(), ConvertedAsm);
+ QualType StrTy = Context.getStringLiteralArrayType(Context.CharTy,
+ ConvertedAsm.size());
return StringLiteral::Create(Context, ConvertedAsm,
StringLiteralKind::Ordinary,
- /*Pascal*/ false, {}, SL->getBeginLoc());
+ /*Pascal*/ false, StrTy, SL->getBeginLoc());
}
return SL;
}
diff --git a/clang/test/CodeGen/systemz-charset.c b/clang/test/CodeGen/systemz-charset.c
index f3bb1055ddf5e..4ccc9934e4bde 100644
--- a/clang/test/CodeGen/systemz-charset.c
+++ b/clang/test/CodeGen/systemz-charset.c
@@ -71,3 +71,7 @@ void asm_labeled_fn(void) __asm__("\174asm_sym");
void asm_labeled_fn(void) {}
//CHECK: define{{.*}}void @"@asm_sym"(
//CHECK-UTF8: define{{.*}}void @"@asm_sym"(
+
+void asm_stmt_fn(void) { __asm__("\225\226\227nop"); }
+//CHECK: call void asm{{.*}}"nopnop"
+//CHECK-UTF8: call void asm{{.*}}"nopnop"
>From 9e48d18b741184e649859daa640f8ec49e70d59f Mon Sep 17 00:00:00 2001
From: Abhina Sreeskantharajan <Abhina.Sreeskantharajan at ibm.com>
Date: Fri, 17 Jul 2026 14:31:50 -0400
Subject: [PATCH 5/6] check for UTF-8 to avoid conversion codepath if no
conversion is needed
---
clang/lib/Sema/SemaStmtAsm.cpp | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/clang/lib/Sema/SemaStmtAsm.cpp b/clang/lib/Sema/SemaStmtAsm.cpp
index 8b61228db3533..e0c3f7c7fd84f 100644
--- a/clang/lib/Sema/SemaStmtAsm.cpp
+++ b/clang/lib/Sema/SemaStmtAsm.cpp
@@ -245,7 +245,7 @@ ExprResult Sema::ActOnGCCAsmStmtString(Expr *Expr, bool ForAsmLabel) {
Diag(Expr->getBeginLoc(), diag::err_asm_operand_empty_string)
<< SL->getSourceRange();
}
- if (Context.getTargetInfo().FromSystemEncodingConverter) {
+ if (Context.getTargetInfo().FromSystemEncodingConverter != nullptr) {
SmallString<16> ConvertedAsm;
Context.getTargetInfo().FromSystemEncodingConverter->convert(
SL->getString(), ConvertedAsm);
>From 0f9bf91671d7109e32650413c7f500d47e937c99 Mon Sep 17 00:00:00 2001
From: Abhina Sreeskantharajan <Abhina.Sreeskantharajan at ibm.com>
Date: Tue, 11 Aug 2026 09:02:17 -0400
Subject: [PATCH 6/6] Remove unrelated testcases
---
clang/test/CodeGen/systemz-charset-diag.cpp | 8 --------
1 file changed, 8 deletions(-)
diff --git a/clang/test/CodeGen/systemz-charset-diag.cpp b/clang/test/CodeGen/systemz-charset-diag.cpp
index ad08e1f391214..4ed94810150a3 100644
--- a/clang/test/CodeGen/systemz-charset-diag.cpp
+++ b/clang/test/CodeGen/systemz-charset-diag.cpp
@@ -1,11 +1,3 @@
// RUN: %clang_cc1 -triple s390x-none-zos -fexec-charset IBM-1047 %s -std=c++17 -emit-llvm -o - -verify
const char* Computer = "🖥️"; // expected-error-re {{conversion to literal encoding failed: {{.*}}}}
-
-static_assert(false, "Error string"); // expected-error {{static assertion failed: Error string}}
-
-[[deprecated("message")]] void test_deprecated() {return;} // expected-note {{'test_deprecated' has been explicitly marked deprecated here}}
-
-int main() {
- test_deprecated(); // expected-warning {{'test_deprecated' is deprecated: message}}
-}
More information about the cfe-commits
mailing list