Merge branch 'main' into issue-97715/lsv-insert-casts

Author: gandhi56

bolt/docs/BinaryAnalysis.md

@@ -180,12 +180,6 @@ The following are current known cases of false negatives:
    [prototype branch](
    https://github.com/llvm/llvm-project/compare/main...kbeyls:llvm-project:bolt-gadget-scanner-prototype).
 
-BOLT cannot currently handle functions with `cfi_negate_ra_state` correctly,
-i.e. any binaries built with `-mbranch-protection=pac-ret`. The scanner is meant
-to be used on specifically such binaries, so this is a major limitation! Work is
-going on in PR [#120064](https://github.com/llvm/llvm-project/pull/120064) to
-fix this.
-
 ## How to add your own binary analysis
 
 _TODO: this section needs to be written. Ideally, we should have a simple

bolt/lib/Core/BinaryFunction.cpp

@@ -65,6 +65,8 @@ extern cl::opt<bool> StrictMode;
 extern cl::opt<bool> UpdateDebugSections;
 extern cl::opt<unsigned> Verbosity;
 
+extern bool BinaryAnalysisMode;
+extern bool HeatmapMode;
 extern bool processAllFunctions();
 
 static cl::opt<bool> CheckEncoding(
@@ -2760,13 +2762,19 @@ struct CFISnapshot {
     }
     case MCCFIInstruction::OpAdjustCfaOffset:
     case MCCFIInstruction::OpWindowSave:
-    case MCCFIInstruction::OpNegateRAState:
     case MCCFIInstruction::OpNegateRAStateWithPC:
     case MCCFIInstruction::OpLLVMDefAspaceCfa:
     case MCCFIInstruction::OpLabel:
     case MCCFIInstruction::OpValOffset:
       llvm_unreachable("unsupported CFI opcode");
       break;
+    case MCCFIInstruction::OpNegateRAState:
+      if (!(opts::BinaryAnalysisMode || opts::HeatmapMode)) {
+        llvm_unreachable("BOLT-ERROR: binaries using pac-ret hardening (e.g. "
+                         "as produced by '-mbranch-protection=pac-ret') are "
+                         "currently not supported by BOLT.");
+      }
+      break;
     case MCCFIInstruction::OpRememberState:
     case MCCFIInstruction::OpRestoreState:
     case MCCFIInstruction::OpGnuArgsSize:
@@ -2900,13 +2908,19 @@ struct CFISnapshotDiff : public CFISnapshot {
       return CFAReg == Instr.getRegister() && CFAOffset == Instr.getOffset();
     case MCCFIInstruction::OpAdjustCfaOffset:
     case MCCFIInstruction::OpWindowSave:
-    case MCCFIInstruction::OpNegateRAState:
     case MCCFIInstruction::OpNegateRAStateWithPC:
     case MCCFIInstruction::OpLLVMDefAspaceCfa:
     case MCCFIInstruction::OpLabel:
     case MCCFIInstruction::OpValOffset:
       llvm_unreachable("unsupported CFI opcode");
       return false;
+    case MCCFIInstruction::OpNegateRAState:
+      if (!(opts::BinaryAnalysisMode || opts::HeatmapMode)) {
+        llvm_unreachable("BOLT-ERROR: binaries using pac-ret hardening (e.g. "
+                         "as produced by '-mbranch-protection=pac-ret') are "
+                         "currently not supported by BOLT.");
+      }
+      break;
     case MCCFIInstruction::OpRememberState:
     case MCCFIInstruction::OpRestoreState:
     case MCCFIInstruction::OpGnuArgsSize:
@@ -3051,13 +3065,19 @@ BinaryFunction::unwindCFIState(int32_t FromState, int32_t ToState,
       break;
     case MCCFIInstruction::OpAdjustCfaOffset:
     case MCCFIInstruction::OpWindowSave:
-    case MCCFIInstruction::OpNegateRAState:
     case MCCFIInstruction::OpNegateRAStateWithPC:
     case MCCFIInstruction::OpLLVMDefAspaceCfa:
     case MCCFIInstruction::OpLabel:
     case MCCFIInstruction::OpValOffset:
       llvm_unreachable("unsupported CFI opcode");
       break;
+    case MCCFIInstruction::OpNegateRAState:
+      if (!(opts::BinaryAnalysisMode || opts::HeatmapMode)) {
+        llvm_unreachable("BOLT-ERROR: binaries using pac-ret hardening (e.g. "
+                         "as produced by '-mbranch-protection=pac-ret') are "
+                         "currently not supported by BOLT.");
+      }
+      break;
     case MCCFIInstruction::OpGnuArgsSize:
       // do not affect CFI state
       break;

bolt/lib/Rewrite/RewriteInstance.cpp

@@ -2697,20 +2697,19 @@ void RewriteInstance::handleRelocation(const SectionRef &RelocatedSection,
                BD->nameStartsWith("_ZTCN"))) { // construction vtable
       BinaryFunction *BF = BC->getBinaryFunctionContainingAddress(
           SymbolAddress, /*CheckPastEnd*/ false, /*UseMaxSize*/ true);
-      if (!BF || BF->getAddress() != SymbolAddress) {
-        BC->errs()
-            << "BOLT-ERROR: the virtual function table entry at offset 0x"
-            << Twine::utohexstr(Rel.getOffset());
-        if (BF)
-          BC->errs() << " points to the middle of a function @ 0x"
-                     << Twine::utohexstr(BF->getAddress()) << "\n";
-        else
-          BC->errs() << " does not point to any function\n";
-        exit(1);
+      if (BF) {
+        if (BF->getAddress() != SymbolAddress) {
+          BC->errs()
+              << "BOLT-ERROR: the virtual function table entry at offset 0x"
+              << Twine::utohexstr(Rel.getOffset())
+              << " points to the middle of a function @ 0x"
+              << Twine::utohexstr(BF->getAddress()) << "\n";
+          exit(1);
+        }
+        BC->addRelocation(Rel.getOffset(), BF->getSymbol(), RType, Addend,
+                          ExtractedValue);
+        return;
       }
-      BC->addRelocation(Rel.getOffset(), BF->getSymbol(), RType, Addend,
-                        ExtractedValue);
-      return;
     }
   }
 

bolt/test/AArch64/skip-non-vfuncptr-reloc-in-relative-vtable.s

@@ -0,0 +1,35 @@
+// Test the fix that BOLT should skip special handling of any non-virtual
+// function pointer relocations in relative vtable.
+
+// RUN: llvm-mc -filetype=obj -triple aarch64-unknown-gnu %s -o %t.o
+// RUN: %clang %cxxflags -fuse-ld=lld %t.o -o %t.so -Wl,-q
+// RUN: llvm-bolt %t.so -o %t.bolted.so
+
+	.text
+	.p2align	2
+	.type	foo,@function
+foo:
+	.cfi_startproc
+	adrp	x8, _ZTV3gooE
+	add	x8, x8, :lo12:_ZTV3gooE
+	ldr	x0, [x8]
+	ret
+.Lfunc_end0:
+	.size	foo, .Lfunc_end0-foo
+	.cfi_endproc
+
+	.type	_fake_rtti_data,@object
+	.section	.rodata.cst16._fake_rtti_data,"aMG",@progbits,16,_fake_rtti_data,comdat
+	.p2align	3, 0x0
+_fake_rtti_data:
+	.ascii	"_FAKE_RTTI_DATA_"
+	.size	_fake_rtti_data, 16
+
+	.type	_ZTV3gooE,@object
+	.section	.rodata,"a",@progbits
+	.p2align	2, 0x0
+_ZTV3gooE:
+	.word	0
+	.word	_fake_rtti_data-_ZTV3gooE-8
+	.word	foo@PLT-_ZTV3gooE-8
+	.size	_ZTV3gooE, 12

bolt/test/non-empty-debug-line.test

@@ -4,7 +4,7 @@
 REQUIRES: system-linux
 
 RUN: %clang %cflags %S/Inputs/hello.c -g -o %t
-RUN: llvm-bolt %t -o %t1 --update-debug-sections --funcs=_start
+RUN: llvm-bolt %t -o %t1 --update-debug-sections --funcs=_start --keep-nops
 RUN: llvm-readobj -S %t > %t2
 RUN: llvm-readobj -S %t1 >> %t2
 RUN: FileCheck %s --input-file %t2

clang-tools-extra/clang-doc/BitcodeReader.cpp

@@ -79,7 +79,8 @@ static llvm::Error decodeRecord(const Record &R, std::optional<Location> &Field,
   if (R[0] > INT_MAX)
     return llvm::createStringError(llvm::inconvertibleErrorCode(),
                                    "integer too large to parse");
-  Field.emplace(static_cast<int>(R[0]), Blob, static_cast<bool>(R[1]));
+  Field.emplace(static_cast<int>(R[0]), static_cast<int>(R[1]), Blob,
+                static_cast<bool>(R[2]));
   return llvm::Error::success();
 }
 
@@ -130,7 +131,8 @@ static llvm::Error decodeRecord(const Record &R,
   if (R[0] > INT_MAX)
     return llvm::createStringError(llvm::inconvertibleErrorCode(),
                                    "integer too large to parse");
-  Field.emplace_back(static_cast<int>(R[0]), Blob, static_cast<bool>(R[1]));
+  Field.emplace_back(static_cast<int>(R[0]), static_cast<int>(R[1]), Blob,
+                     static_cast<bool>(R[2]));
   return llvm::Error::success();
 }
 

clang-tools-extra/clang-doc/BitcodeWriter.cpp

@@ -78,13 +78,16 @@ static void genLocationAbbrev(std::shared_ptr<llvm::BitCodeAbbrev> &Abbrev) {
       {// 0. Fixed-size integer (line number)
        llvm::BitCodeAbbrevOp(llvm::BitCodeAbbrevOp::Fixed,
                              BitCodeConstants::LineNumberSize),
-       // 1. Boolean (IsFileInRootDir)
+       // 1. Fixed-size integer (start line number)
+       llvm::BitCodeAbbrevOp(llvm::BitCodeAbbrevOp::Fixed,
+                             BitCodeConstants::LineNumberSize),
+       // 2. Boolean (IsFileInRootDir)
        llvm::BitCodeAbbrevOp(llvm::BitCodeAbbrevOp::Fixed,
                              BitCodeConstants::BoolSize),
-       // 2. Fixed-size integer (length of the following string (filename))
+       // 3. Fixed-size integer (length of the following string (filename))
        llvm::BitCodeAbbrevOp(llvm::BitCodeAbbrevOp::Fixed,
                              BitCodeConstants::StringLengthSize),
-       // 3. The string blob
+       // 4. The string blob
        llvm::BitCodeAbbrevOp(llvm::BitCodeAbbrevOp::Blob)});
 }
 
@@ -357,7 +360,8 @@ void ClangDocBitcodeWriter::emitRecord(const Location &Loc, RecordId ID) {
   if (!prepRecordData(ID, true))
     return;
   // FIXME: Assert that the line number is of the appropriate size.
-  Record.push_back(Loc.LineNumber);
+  Record.push_back(Loc.StartLineNumber);
+  Record.push_back(Loc.EndLineNumber);
   assert(Loc.Filename.size() < (1U << BitCodeConstants::StringLengthSize));
   Record.push_back(Loc.IsFileInRootDir);
   Record.push_back(Loc.Filename.size());

clang-tools-extra/clang-doc/HTMLGenerator.cpp

@@ -455,7 +455,7 @@ static std::unique_ptr<TagNode> writeSourceFileRef(const ClangDocContext &CDCtx,
 
   if (!L.IsFileInRootDir && !CDCtx.RepositoryUrl)
     return std::make_unique<TagNode>(
-        HTMLTag::TAG_P, "Defined at line " + std::to_string(L.LineNumber) +
+        HTMLTag::TAG_P, "Defined at line " + std::to_string(L.StartLineNumber) +
                             " of file " + L.Filename);
 
   SmallString<128> FileURL(CDCtx.RepositoryUrl.value_or(""));
@@ -472,13 +472,14 @@ static std::unique_ptr<TagNode> writeSourceFileRef(const ClangDocContext &CDCtx,
           llvm::sys::path::Style::windows));
   auto Node = std::make_unique<TagNode>(HTMLTag::TAG_P);
   Node->Children.emplace_back(std::make_unique<TextNode>("Defined at line "));
-  auto LocNumberNode =
-      std::make_unique<TagNode>(HTMLTag::TAG_A, std::to_string(L.LineNumber));
+  auto LocNumberNode = std::make_unique<TagNode>(
+      HTMLTag::TAG_A, std::to_string(L.StartLineNumber));
   // The links to a specific line in the source code use the github /
   // googlesource notation so it won't work for all hosting pages.
   LocNumberNode->Attributes.emplace_back(
-      "href", formatv("{0}#{1}{2}", FileURL,
-                      CDCtx.RepositoryLinePrefix.value_or(""), L.LineNumber));
+      "href",
+      formatv("{0}#{1}{2}", FileURL, CDCtx.RepositoryLinePrefix.value_or(""),
+              L.StartLineNumber));
   Node->Children.emplace_back(std::move(LocNumberNode));
   Node->Children.emplace_back(std::make_unique<TextNode>(" of file "));
   auto LocFileNode = std::make_unique<TagNode>(

clang-tools-extra/clang-doc/MDGenerator.cpp

@@ -56,12 +56,12 @@ static void writeSourceFileRef(const ClangDocContext &CDCtx, const Location &L,
                                raw_ostream &OS) {
 
   if (!CDCtx.RepositoryUrl) {
-    OS << "*Defined at " << L.Filename << "#" << std::to_string(L.LineNumber)
-       << "*";
+    OS << "*Defined at " << L.Filename << "#"
+       << std::to_string(L.StartLineNumber) << "*";
   } else {
 
     OS << formatv("*Defined at [#{0}{1}{2}](#{0}{1}{3})*",
-                  CDCtx.RepositoryLinePrefix.value_or(""), L.LineNumber,
+                  CDCtx.RepositoryLinePrefix.value_or(""), L.StartLineNumber,
                   L.Filename, *CDCtx.RepositoryUrl);
   }
   OS << "\n\n";

clang-tools-extra/clang-doc/Mapper.cpp

@@ -28,6 +28,17 @@ template <typename T> static bool isTypedefAnonRecord(const T *D) {
   return false;
 }
 
+Location MapASTVisitor::getDeclLocation(const NamedDecl *D) const {
+  bool IsFileInRootDir;
+  llvm::SmallString<128> File =
+      getFile(D, D->getASTContext(), CDCtx.SourceRoot, IsFileInRootDir);
+  SourceManager &SM = D->getASTContext().getSourceManager();
+  int Start = SM.getPresumedLoc(D->getBeginLoc()).getLine();
+  int End = SM.getPresumedLoc(D->getEndLoc()).getLine();
+
+  return Location(Start, End, File, IsFileInRootDir);
+}
+
 void MapASTVisitor::HandleTranslationUnit(ASTContext &Context) {
   TraverseDecl(Context.getTranslationUnitDecl());
 }
@@ -59,9 +70,9 @@ bool MapASTVisitor::mapDecl(const T *D, bool IsDefinition) {
   bool IsFileInRootDir;
   llvm::SmallString<128> File =
       getFile(D, D->getASTContext(), CDCtx.SourceRoot, IsFileInRootDir);
-  auto [Child, Parent] = serialize::emitInfo(
-      D, getComment(D, D->getASTContext()), getLine(D, D->getASTContext()),
-      File, IsFileInRootDir, CDCtx.PublicOnly);
+  auto [Child, Parent] =
+      serialize::emitInfo(D, getComment(D, D->getASTContext()),
+                          getDeclLocation(D), CDCtx.PublicOnly);
 
   // A null in place of a valid Info indicates that the serializer is skipping
   // this decl for some reason (e.g. we're only reporting public decls).

clang-tools-extra/clang-doc/Mapper.h

@@ -46,6 +46,9 @@ class MapASTVisitor : public clang::RecursiveASTVisitor<MapASTVisitor>,
   template <typename T> bool mapDecl(const T *D, bool IsDefinition);
 
   int getLine(const NamedDecl *D, const ASTContext &Context) const;
+
+  Location getDeclLocation(const NamedDecl *D) const;
+
   llvm::SmallString<128> getFile(const NamedDecl *D, const ASTContext &Context,
                                  StringRef RootDir,
                                  bool &IsFileInRootDir) const;

clang-tools-extra/clang-doc/Representation.h

@@ -241,31 +241,29 @@ struct MemberTypeInfo : public FieldTypeInfo {
 };
 
 struct Location {
-  Location(int LineNumber = 0, StringRef Filename = StringRef(),
-           bool IsFileInRootDir = false)
-      : LineNumber(LineNumber), Filename(Filename),
-        IsFileInRootDir(IsFileInRootDir) {}
+  Location(int StartLineNumber = 0, int EndLineNumber = 0,
+           StringRef Filename = StringRef(), bool IsFileInRootDir = false)
+      : StartLineNumber(StartLineNumber), EndLineNumber(EndLineNumber),
+        Filename(Filename), IsFileInRootDir(IsFileInRootDir) {}
 
   bool operator==(const Location &Other) const {
-    return std::tie(LineNumber, Filename) ==
-           std::tie(Other.LineNumber, Other.Filename);
+    return std::tie(StartLineNumber, EndLineNumber, Filename) ==
+           std::tie(Other.StartLineNumber, Other.EndLineNumber, Other.Filename);
   }
 
-  bool operator!=(const Location &Other) const {
-    return std::tie(LineNumber, Filename) !=
-           std::tie(Other.LineNumber, Other.Filename);
-  }
+  bool operator!=(const Location &Other) const { return !(*this == Other); }
 
   // This operator is used to sort a vector of Locations.
   // No specific order (attributes more important than others) is required. Any
   // sort is enough, the order is only needed to call std::unique after sorting
   // the vector.
   bool operator<(const Location &Other) const {
-    return std::tie(LineNumber, Filename) <
-           std::tie(Other.LineNumber, Other.Filename);
+    return std::tie(StartLineNumber, EndLineNumber, Filename) <
+           std::tie(Other.StartLineNumber, Other.EndLineNumber, Other.Filename);
   }
 
-  int LineNumber = 0;           // Line number of this Location.
+  int StartLineNumber = 0; // Line number of this Location.
+  int EndLineNumber = 0;
   SmallString<32> Filename;     // File for this Location.
   bool IsFileInRootDir = false; // Indicates if file is inside root directory
 };