Merge branch 'main' of https://github.com/llvm/llvm-project into fdate

Author: yiwu0b11

.github/workflows/pr-code-format.yml

@@ -11,14 +11,33 @@ jobs:
       - name: Fetch LLVM sources
         uses: actions/checkout@v4
         with:
-          fetch-depth: 2 # Fetches only the last 2 commits
+          ref: ${{ github.event.pull_request.head.sha }}
+
+      - name: Checkout through merge base
+        uses: rmacklin/fetch-through-merge-base@v0
+        with:
+          base_ref: ${{ github.event.pull_request.base.ref }}
+          head_ref: ${{ github.event.pull_request.head.sha }}
+          deepen_length: 500
 
       - name: Get changed files
         id: changed-files
         uses: tj-actions/changed-files@v39
         with:
           separator: ","
-          fetch_depth: 2000 # Fetches only the last 2000 commits
+          skip_initial_fetch: true
+
+      # We need to make sure that we aren't executing/using any code from the
+      # PR for security reasons as we're using pull_request_target. Checkout
+      # the target branch with the necessary files.
+      - name: Fetch code formatting utils
+        uses: actions/checkout@v4
+        with:
+          sparse-checkout: |
+            llvm/utils/git/requirements_formatting.txt
+            llvm/utils/git/code-format-helper.py
+          sparse-checkout-cone-mode: false
+          path: code-format-tools
 
       - name: "Listed files"
         run: |
@@ -35,10 +54,10 @@ jobs:
         with:
           python-version: '3.11'
           cache: 'pip'
-          cache-dependency-path: 'llvm/utils/git/requirements_formatting.txt'
+          cache-dependency-path: 'code-format-tools/llvm/utils/git/requirements_formatting.txt'
 
       - name: Install python dependencies
-        run: pip install -r llvm/utils/git/requirements_formatting.txt
+        run: pip install -r code-format-tools/llvm/utils/git/requirements_formatting.txt
 
       - name: Run code formatter
         env:
@@ -47,7 +66,7 @@ jobs:
           END_REV: ${{ github.event.pull_request.head.sha }}
           CHANGED_FILES: ${{ steps.changed-files.outputs.all_changed_files }}
         run: |
-          python llvm/utils/git/code-format-helper.py \
+          python ./code-format-tools/llvm/utils/git/code-format-helper.py \
             --token ${{ secrets.GITHUB_TOKEN }} \
             --issue-number $GITHUB_PR_NUMBER \
             --start-rev $START_REV \

bolt/docs/doxygen.cfg.in

@@ -1954,7 +1954,7 @@ PREDEFINED             =
 EXPAND_AS_DEFINED      =
 
 # If the SKIP_FUNCTION_MACROS tag is set to YES then doxygen's preprocessor will
-# remove all refrences to function-like macros that are alone on a line, have an
+# remove all references to function-like macros that are alone on a line, have an
 # all uppercase name, and do not end with a semicolon. Such function macros are
 # typically used for boiler-plate code, and will confuse the parser if not
 # removed.

bolt/include/bolt/Core/BinaryContext.h

@@ -680,6 +680,15 @@ class BinaryContext {
   /// the execution of the binary is completed.
   std::optional<uint64_t> FiniFunctionAddress;
 
+  /// DT_FINI.
+  std::optional<uint64_t> FiniAddress;
+
+  /// DT_FINI_ARRAY. Only used when DT_FINI is not set.
+  std::optional<uint64_t> FiniArrayAddress;
+
+  /// DT_FINI_ARRAYSZ. Only used when DT_FINI is not set.
+  std::optional<uint64_t> FiniArraySize;
+
   /// Page alignment used for code layout.
   uint64_t PageAlign{HugePageSize};
 
@@ -941,7 +950,7 @@ class BinaryContext {
   bool registerFragment(BinaryFunction &TargetFunction,
                         BinaryFunction &Function) const;
 
-  /// Add unterprocedural reference for \p Function to \p Address
+  /// Add interprocedural reference for \p Function to \p Address
   void addInterproceduralReference(BinaryFunction *Function, uint64_t Address) {
     InterproceduralReferences.push_back({Function, Address});
   }

bolt/include/bolt/Core/BinaryFunction.h

@@ -1445,7 +1445,8 @@ class BinaryFunction {
 
   /// Rebuilds BBs layout, ignoring dead BBs. Returns the number of removed
   /// BBs and the removed number of bytes of code.
-  std::pair<unsigned, uint64_t> eraseInvalidBBs();
+  std::pair<unsigned, uint64_t>
+  eraseInvalidBBs(const MCCodeEmitter *Emitter = nullptr);
 
   /// Get the relative order between two basic blocks in the original
   /// layout.  The result is > 0 if B occurs before A and < 0 if B
@@ -2172,7 +2173,7 @@ class BinaryFunction {
   /// is corrupted. If it is unable to fix it, it returns false.
   bool finalizeCFIState();
 
-  /// Return true if this function needs an address-transaltion table after
+  /// Return true if this function needs an address-translation table after
   /// its code emission.
   bool requiresAddressTranslation() const;
 

bolt/include/bolt/Core/BinarySection.h

@@ -127,7 +127,7 @@ class BinarySection {
     return *ContentsOrErr;
   }
 
-  /// Get the set of relocations refering to data in this section that
+  /// Get the set of relocations referring to data in this section that
   /// has been reordered.  The relocation offsets will be modified to
   /// reflect the new data locations.
   RelocationSetType reorderRelocations(bool Inplace) const;
@@ -375,8 +375,12 @@ class BinarySection {
   /// Add a dynamic relocation at the given /p Offset.
   void addDynamicRelocation(uint64_t Offset, MCSymbol *Symbol, uint64_t Type,
                             uint64_t Addend, uint64_t Value = 0) {
-    assert(Offset < getSize() && "offset not within section bounds");
-    DynamicRelocations.emplace(Relocation{Offset, Symbol, Type, Addend, Value});
+    addDynamicRelocation(Relocation{Offset, Symbol, Type, Addend, Value});
+  }
+
+  void addDynamicRelocation(const Relocation &Reloc) {
+    assert(Reloc.Offset < getSize() && "offset not within section bounds");
+    DynamicRelocations.emplace(Reloc);
   }
 
   /// Add relocation against the original contents of this section.
@@ -410,6 +414,18 @@ class BinarySection {
     return Itr != DynamicRelocations.end() ? &*Itr : nullptr;
   }
 
+  std::optional<Relocation> takeDynamicRelocationAt(uint64_t Offset) {
+    Relocation Key{Offset, 0, 0, 0, 0};
+    auto Itr = DynamicRelocations.find(Key);
+
+    if (Itr == DynamicRelocations.end())
+      return std::nullopt;
+
+    Relocation Reloc = *Itr;
+    DynamicRelocations.erase(Itr);
+    return Reloc;
+  }
+
   uint64_t hash(const BinaryData &BD) const {
     std::map<const BinaryData *, uint64_t> Cache;
     return hash(BD, Cache);

bolt/include/bolt/Core/DebugData.h

@@ -64,7 +64,7 @@ std::optional<AttrInfo> findAttributeInfo(const DWARFDie DIE,
                                           dwarf::Attribute Attr);
 
 // DWARF5 Header in order of encoding.
-// Types represent encodnig sizes.
+// Types represent encoding sizes.
 using UnitLengthType = uint32_t;
 using VersionType = uint16_t;
 using AddressSizeType = uint8_t;
@@ -372,13 +372,13 @@ class DebugAddrWriter {
     }
 
     /// Updates AddressToIndex Map.
-    /// Follows the same symantics as unordered map [].
+    /// Follows the same semantics as unordered map [].
     void updateAddressToIndex(uint64_t Address, uint32_t Index) {
       AddressToIndex[Address] = Index;
     }
 
     /// Updates IndexToAddress Map.
-    /// Follows the same symantics as unordered map [].
+    /// Follows the same semantics as unordered map [].
     void updateIndexToAddrss(uint64_t Address, uint32_t Index) {
       IndexToAddress[Index] = Address;
     }
@@ -474,7 +474,7 @@ class DebugStrWriter {
   }
 
   /// Adds string to .debug_str.
-  /// On first invokation it initializes internal data stractures.
+  /// On first invocation it initializes internal data structures.
   uint32_t addString(StringRef Str);
 
   /// Returns False if no strings were added to .debug_str.
@@ -485,7 +485,7 @@ class DebugStrWriter {
   std::mutex WriterMutex;
   /// Initializes Buffer and Stream.
   void initialize();
-  /// Creats internal data stractures.
+  /// Creates internal data structures.
   void create();
   std::unique_ptr<DebugStrBufferVector> StrBuffer;
   std::unique_ptr<raw_svector_ostream> StrStream;
@@ -585,7 +585,7 @@ class DebugLoclistWriter : public DebugLocWriter {
   void finalize(DIEBuilder &DIEBldr, DIE &Die) override;
 
   /// Returns CU ID.
-  /// For Skelton CU it is a CU Offset.
+  /// For Skeleton CU it is a CU Offset.
   /// For DWO CU it is a DWO ID.
   uint64_t getCUID() const {
     return CU.isDWOUnit() ? *CU.getDWOId() : CU.getOffset();

bolt/include/bolt/Core/DynoStats.h

@@ -121,7 +121,7 @@ class DynoStats {
   /// Maps instruction opcodes to:
   /// 1. Accumulated executed instruction counts.
   /// 2. a multimap that records highest execution counts, function names,
-  /// and BB offsets where intructions of these opcodes occur.
+  /// and BB offsets where instructions of these opcodes occur.
   using MaxOpcodeHistogramTy =
       std::multimap<uint64_t, std::pair<StringRef, uint32_t>>;
   using OpcodeHistogramTy =

bolt/include/bolt/Core/Exceptions.h

@@ -77,7 +77,7 @@ class EHFrameParser {
   ///   void PatcherCallback(uint64_t Value, uint64_t Offset, uint64_t Type);
   ///
   /// where Value is a value of the reference, Offset - is an offset into the
-  /// frame data at which the reference occured, and Type is a DWARF encoding
+  /// frame data at which the reference occurred, and Type is a DWARF encoding
   /// type of the reference.
   static Error parse(DWARFDataExtractor Data, uint64_t EHFrameAddress,
                      PatcherCallbackTy PatcherCallback);

bolt/include/bolt/Core/MCPlus.h

@@ -32,11 +32,16 @@ namespace MCPlus {
 /// pad and the uint64_t represents the action.
 using MCLandingPad = std::pair<const MCSymbol *, uint64_t>;
 
-/// An extension to MCInst is provided via an extra operand of type MCInst with
-/// ANNOTATION_LABEL opcode (i.e. we are tying an annotation instruction to an
-/// existing one). The annotation instruction contains a list of Immediate
-/// operands. Each operand either contains a value, or is a pointer to
-/// an instance of class MCAnnotation.
+/// An extension to MCInst is provided via extra operands, i.e. operands that
+/// are not used in the instruction assembly. Any kind of metadata can be
+/// attached to MCInst with this "annotation" extension using MCPlusBuilder
+/// interface.
+//
+/// The first extra operand must be of type kInst with an empty (nullptr)
+/// value. The kInst operand type is unused on most non-VLIW architectures.
+/// We use it to mark the beginning of annotations operands. The rest of the
+/// operands are of Immediate type with annotation info encoded into the value
+/// of the immediate.
 ///
 /// There are 2 distinct groups of annotations. The first group is a first-class
 /// annotation that affects semantics of the instruction, such as an
@@ -55,7 +60,7 @@ using MCLandingPad = std::pair<const MCSymbol *, uint64_t>;
 /// of their corresponding operand.
 ///
 /// Annotations in the second group could be addressed either by name, or by
-/// by and index which could be queried by providing a name.
+/// by index which could be queried by providing the name.
 class MCAnnotation {
 public:
   enum Kind {
@@ -106,10 +111,11 @@ template <typename ValueType> class MCSimpleAnnotation : public MCAnnotation {
 /// Return a number of operands in \Inst excluding operands representing
 /// annotations.
 inline unsigned getNumPrimeOperands(const MCInst &Inst) {
-  if (Inst.getNumOperands() > 0 && std::prev(Inst.end())->isInst()) {
-    assert(std::prev(Inst.end())->getInst()->getOpcode() ==
-           TargetOpcode::ANNOTATION_LABEL);
-    return Inst.getNumOperands() - 1;
+  for (signed I = Inst.getNumOperands() - 1; I >= 0; --I) {
+    if (Inst.getOperand(I).isInst())
+      return I;
+    if (!Inst.getOperand(I).isImm())
+      return Inst.getNumOperands();
   }
   return Inst.getNumOperands();
 }