merge main into amd-staging

reverts: breaks rocBLAS build
d57230c [AMDGPU][MC] Disallow op_sel in some VOP3P dot instructions (llvm#100485)

Change-Id: I19e8a59504e111ac0eeca51f30a0fbe36c6b57d5

Author: animeshk-amd

.github/workflows/libcxx-build-and-test.yaml

@@ -37,7 +37,7 @@ jobs:
   stage1:
     if: github.repository_owner == 'llvm'
     runs-on: libcxx-self-hosted-linux
-    container: ghcr.io/llvm/libcxx-linux-builder:0fd6f684b9c84c32d6cbfd9742402e788b2879f1
+    container: ghcr.io/llvm/libcxx-linux-builder:d8a0709b1090350a7fe3604d8ab78c7d62f10698
     continue-on-error: false
     strategy:
       fail-fast: false
@@ -74,7 +74,7 @@ jobs:
   stage2:
     if: github.repository_owner == 'llvm'
     runs-on: libcxx-self-hosted-linux
-    container: ghcr.io/llvm/libcxx-linux-builder:0fd6f684b9c84c32d6cbfd9742402e788b2879f1
+    container: ghcr.io/llvm/libcxx-linux-builder:d8a0709b1090350a7fe3604d8ab78c7d62f10698
     needs: [ stage1 ]
     continue-on-error: false
     strategy:
@@ -162,7 +162,7 @@ jobs:
         - config: 'generic-msan'
           machine: libcxx-self-hosted-linux
     runs-on: ${{ matrix.machine }}
-    container: ghcr.io/llvm/libcxx-linux-builder:0fd6f684b9c84c32d6cbfd9742402e788b2879f1
+    container: ghcr.io/llvm/libcxx-linux-builder:d8a0709b1090350a7fe3604d8ab78c7d62f10698
     steps:
       - uses: actions/checkout@v4
       - name: ${{ matrix.config }}

bolt/lib/Rewrite/LinuxKernelRewriter.cpp

@@ -123,6 +123,30 @@ inline raw_ostream &operator<<(raw_ostream &OS, const ORCState &E) {
 
 namespace {
 
+/// Extension to DataExtractor that supports reading addresses stored in
+/// PC-relative format.
+class AddressExtractor : public DataExtractor {
+  uint64_t DataAddress;
+
+public:
+  AddressExtractor(StringRef Data, uint64_t DataAddress, bool IsLittleEndian,
+                   uint8_t AddressSize)
+      : DataExtractor(Data, IsLittleEndian, AddressSize),
+        DataAddress(DataAddress) {}
+
+  /// Extract 32-bit PC-relative address/pointer.
+  uint64_t getPCRelAddress32(Cursor &C) {
+    const uint64_t Base = DataAddress + C.tell();
+    return Base + (int32_t)getU32(C);
+  }
+
+  /// Extract 64-bit PC-relative address/pointer.
+  uint64_t getPCRelAddress64(Cursor &C) {
+    const uint64_t Base = DataAddress + C.tell();
+    return Base + (int64_t)getU64(C);
+  }
+};
+
 class LinuxKernelRewriter final : public MetadataRewriter {
   /// Information required for updating metadata referencing an instruction.
   struct InstructionFixup {
@@ -423,13 +447,13 @@ Error LinuxKernelRewriter::processSMPLocks() {
     return createStringError(errc::executable_format_error,
                              "bad size of .smp_locks section");
 
-  DataExtractor DE = DataExtractor(SMPLocksSection->getContents(),
-                                   BC.AsmInfo->isLittleEndian(),
-                                   BC.AsmInfo->getCodePointerSize());
-  DataExtractor::Cursor Cursor(0);
+  AddressExtractor AE(SMPLocksSection->getContents(), SectionAddress,
+                      BC.AsmInfo->isLittleEndian(),
+                      BC.AsmInfo->getCodePointerSize());
+  AddressExtractor::Cursor Cursor(0);
   while (Cursor && Cursor.tell() < SectionSize) {
     const uint64_t Offset = Cursor.tell();
-    const uint64_t IP = SectionAddress + Offset + (int32_t)DE.getU32(Cursor);
+    const uint64_t IP = AE.getPCRelAddress32(Cursor);
 
     // Consume the status of the cursor.
     if (!Cursor)
@@ -499,20 +523,17 @@ Error LinuxKernelRewriter::readORCTables() {
     return createStringError(errc::executable_format_error,
                              "ORC entries number mismatch detected");
 
-  const uint64_t IPSectionAddress = ORCUnwindIPSection->getAddress();
-  DataExtractor OrcDE = DataExtractor(ORCUnwindSection->getContents(),
-                                      BC.AsmInfo->isLittleEndian(),
-                                      BC.AsmInfo->getCodePointerSize());
-  DataExtractor IPDE = DataExtractor(ORCUnwindIPSection->getContents(),
-                                     BC.AsmInfo->isLittleEndian(),
-                                     BC.AsmInfo->getCodePointerSize());
+  DataExtractor OrcDE(ORCUnwindSection->getContents(),
+                      BC.AsmInfo->isLittleEndian(),
+                      BC.AsmInfo->getCodePointerSize());
+  AddressExtractor IPAE(
+      ORCUnwindIPSection->getContents(), ORCUnwindIPSection->getAddress(),
+      BC.AsmInfo->isLittleEndian(), BC.AsmInfo->getCodePointerSize());
   DataExtractor::Cursor ORCCursor(0);
   DataExtractor::Cursor IPCursor(0);
   uint64_t PrevIP = 0;
   for (uint32_t Index = 0; Index < NumORCEntries; ++Index) {
-    const uint64_t IP =
-        IPSectionAddress + IPCursor.tell() + (int32_t)IPDE.getU32(IPCursor);
-
+    const uint64_t IP = IPAE.getPCRelAddress32(IPCursor);
     // Consume the status of the cursor.
     if (!IPCursor)
       return createStringError(errc::executable_format_error,
@@ -856,15 +877,13 @@ Error LinuxKernelRewriter::validateORCTables() {
   if (!ORCUnwindIPSection)
     return Error::success();
 
-  const uint64_t IPSectionAddress = ORCUnwindIPSection->getAddress();
-  DataExtractor IPDE = DataExtractor(ORCUnwindIPSection->getOutputContents(),
-                                     BC.AsmInfo->isLittleEndian(),
-                                     BC.AsmInfo->getCodePointerSize());
-  DataExtractor::Cursor IPCursor(0);
+  AddressExtractor IPAE(
+      ORCUnwindIPSection->getOutputContents(), ORCUnwindIPSection->getAddress(),
+      BC.AsmInfo->isLittleEndian(), BC.AsmInfo->getCodePointerSize());
+  AddressExtractor::Cursor IPCursor(0);
   uint64_t PrevIP = 0;
   for (uint32_t Index = 0; Index < NumORCEntries; ++Index) {
-    const uint64_t IP =
-        IPSectionAddress + IPCursor.tell() + (int32_t)IPDE.getU32(IPCursor);
+    const uint64_t IP = IPAE.getPCRelAddress32(IPCursor);
     if (!IPCursor)
       return createStringError(errc::executable_format_error,
                                "out of bounds while reading ORC IP table: %s",
@@ -916,16 +935,14 @@ Error LinuxKernelRewriter::readStaticCalls() {
                              "static call table size error");
 
   const uint64_t SectionAddress = StaticCallSection->getAddress();
-  DataExtractor DE(StaticCallSection->getContents(),
-                   BC.AsmInfo->isLittleEndian(),
-                   BC.AsmInfo->getCodePointerSize());
-  DataExtractor::Cursor Cursor(StaticCallTableAddress - SectionAddress);
+  AddressExtractor AE(StaticCallSection->getContents(), SectionAddress,
+                      BC.AsmInfo->isLittleEndian(),
+                      BC.AsmInfo->getCodePointerSize());
+  AddressExtractor::Cursor Cursor(StaticCallTableAddress - SectionAddress);
   uint32_t EntryID = 0;
   while (Cursor && Cursor.tell() < Stop->getAddress() - SectionAddress) {
-    const uint64_t CallAddress =
-        SectionAddress + Cursor.tell() + (int32_t)DE.getU32(Cursor);
-    const uint64_t KeyAddress =
-        SectionAddress + Cursor.tell() + (int32_t)DE.getU32(Cursor);
+    const uint64_t CallAddress = AE.getPCRelAddress32(Cursor);
+    const uint64_t KeyAddress = AE.getPCRelAddress32(Cursor);
 
     // Consume the status of the cursor.
     if (!Cursor)
@@ -1027,18 +1044,15 @@ Error LinuxKernelRewriter::readExceptionTable() {
     return createStringError(errc::executable_format_error,
                              "exception table size error");
 
-  const uint64_t SectionAddress = ExceptionsSection->getAddress();
-  DataExtractor DE(ExceptionsSection->getContents(),
-                   BC.AsmInfo->isLittleEndian(),
-                   BC.AsmInfo->getCodePointerSize());
-  DataExtractor::Cursor Cursor(0);
+  AddressExtractor AE(
+      ExceptionsSection->getContents(), ExceptionsSection->getAddress(),
+      BC.AsmInfo->isLittleEndian(), BC.AsmInfo->getCodePointerSize());
+  AddressExtractor::Cursor Cursor(0);
   uint32_t EntryID = 0;
   while (Cursor && Cursor.tell() < ExceptionsSection->getSize()) {
-    const uint64_t InstAddress =
-        SectionAddress + Cursor.tell() + (int32_t)DE.getU32(Cursor);
-    const uint64_t FixupAddress =
-        SectionAddress + Cursor.tell() + (int32_t)DE.getU32(Cursor);
-    const uint64_t Data = DE.getU32(Cursor);
+    const uint64_t InstAddress = AE.getPCRelAddress32(Cursor);
+    const uint64_t FixupAddress = AE.getPCRelAddress32(Cursor);
+    const uint64_t Data = AE.getU32(Cursor);
 
     // Consume the status of the cursor.
     if (!Cursor)
@@ -1134,9 +1148,9 @@ Error LinuxKernelRewriter::readParaInstructions() {
   if (!ParavirtualPatchSection)
     return Error::success();
 
-  DataExtractor DE = DataExtractor(ParavirtualPatchSection->getContents(),
-                                   BC.AsmInfo->isLittleEndian(),
-                                   BC.AsmInfo->getCodePointerSize());
+  DataExtractor DE(ParavirtualPatchSection->getContents(),
+                   BC.AsmInfo->isLittleEndian(),
+                   BC.AsmInfo->getCodePointerSize());
   uint32_t EntryID = 0;
   DataExtractor::Cursor Cursor(0);
   while (Cursor && !DE.eof(Cursor)) {
@@ -1235,15 +1249,14 @@ Error LinuxKernelRewriter::readBugTable() {
     return createStringError(errc::executable_format_error,
                              "bug table size error");
 
-  const uint64_t SectionAddress = BugTableSection->getAddress();
-  DataExtractor DE(BugTableSection->getContents(), BC.AsmInfo->isLittleEndian(),
-                   BC.AsmInfo->getCodePointerSize());
-  DataExtractor::Cursor Cursor(0);
+  AddressExtractor AE(
+      BugTableSection->getContents(), BugTableSection->getAddress(),
+      BC.AsmInfo->isLittleEndian(), BC.AsmInfo->getCodePointerSize());
+  AddressExtractor::Cursor Cursor(0);
   uint32_t EntryID = 0;
   while (Cursor && Cursor.tell() < BugTableSection->getSize()) {
     const uint64_t Pos = Cursor.tell();
-    const uint64_t InstAddress =
-        SectionAddress + Pos + (int32_t)DE.getU32(Cursor);
+    const uint64_t InstAddress = AE.getPCRelAddress32(Cursor);
     Cursor.seek(Pos + BUG_TABLE_ENTRY_SIZE);
 
     if (!Cursor)
@@ -1402,23 +1415,20 @@ Error LinuxKernelRewriter::readAltInstructions() {
 Error LinuxKernelRewriter::tryReadAltInstructions(uint32_t AltInstFeatureSize,
                                                   bool AltInstHasPadLen,
                                                   bool ParseOnly) {
-  const uint64_t Address = AltInstrSection->getAddress();
-  DataExtractor DE = DataExtractor(AltInstrSection->getContents(),
-                                   BC.AsmInfo->isLittleEndian(),
-                                   BC.AsmInfo->getCodePointerSize());
+  AddressExtractor AE(
+      AltInstrSection->getContents(), AltInstrSection->getAddress(),
+      BC.AsmInfo->isLittleEndian(), BC.AsmInfo->getCodePointerSize());
+  AddressExtractor::Cursor Cursor(0);
   uint64_t EntryID = 0;
-  DataExtractor::Cursor Cursor(0);
-  while (Cursor && !DE.eof(Cursor)) {
-    const uint64_t OrgInstAddress =
-        Address + Cursor.tell() + (int32_t)DE.getU32(Cursor);
-    const uint64_t AltInstAddress =
-        Address + Cursor.tell() + (int32_t)DE.getU32(Cursor);
-    const uint64_t Feature = DE.getUnsigned(Cursor, AltInstFeatureSize);
-    const uint8_t OrgSize = DE.getU8(Cursor);
-    const uint8_t AltSize = DE.getU8(Cursor);
+  while (Cursor && !AE.eof(Cursor)) {
+    const uint64_t OrgInstAddress = AE.getPCRelAddress32(Cursor);
+    const uint64_t AltInstAddress = AE.getPCRelAddress32(Cursor);
+    const uint64_t Feature = AE.getUnsigned(Cursor, AltInstFeatureSize);
+    const uint8_t OrgSize = AE.getU8(Cursor);
+    const uint8_t AltSize = AE.getU8(Cursor);
 
     // Older kernels may have the padlen field.
-    const uint8_t PadLen = AltInstHasPadLen ? DE.getU8(Cursor) : 0;
+    const uint8_t PadLen = AltInstHasPadLen ? AE.getU8(Cursor) : 0;
 
     if (!Cursor)
       return createStringError(
@@ -1537,19 +1547,17 @@ Error LinuxKernelRewriter::readPCIFixupTable() {
     return createStringError(errc::executable_format_error,
                              "PCI fixup table size error");
 
-  const uint64_t Address = PCIFixupSection->getAddress();
-  DataExtractor DE = DataExtractor(PCIFixupSection->getContents(),
-                                   BC.AsmInfo->isLittleEndian(),
-                                   BC.AsmInfo->getCodePointerSize());
+  AddressExtractor AE(
+      PCIFixupSection->getContents(), PCIFixupSection->getAddress(),
+      BC.AsmInfo->isLittleEndian(), BC.AsmInfo->getCodePointerSize());
+  AddressExtractor::Cursor Cursor(0);
   uint64_t EntryID = 0;
-  DataExtractor::Cursor Cursor(0);
-  while (Cursor && !DE.eof(Cursor)) {
-    const uint16_t Vendor = DE.getU16(Cursor);
-    const uint16_t Device = DE.getU16(Cursor);
-    const uint32_t Class = DE.getU32(Cursor);
-    const uint32_t ClassShift = DE.getU32(Cursor);
-    const uint64_t HookAddress =
-        Address + Cursor.tell() + (int32_t)DE.getU32(Cursor);
+  while (Cursor && !AE.eof(Cursor)) {
+    const uint16_t Vendor = AE.getU16(Cursor);
+    const uint16_t Device = AE.getU16(Cursor);
+    const uint32_t Class = AE.getU32(Cursor);
+    const uint32_t ClassShift = AE.getU32(Cursor);
+    const uint64_t HookAddress = AE.getPCRelAddress32(Cursor);
 
     if (!Cursor)
       return createStringError(errc::executable_format_error,
@@ -1654,18 +1662,15 @@ Error LinuxKernelRewriter::readStaticKeysJumpTable() {
                              "static keys jump table size error");
 
   const uint64_t SectionAddress = StaticKeysJumpSection->getAddress();
-  DataExtractor DE(StaticKeysJumpSection->getContents(),
-                   BC.AsmInfo->isLittleEndian(),
-                   BC.AsmInfo->getCodePointerSize());
-  DataExtractor::Cursor Cursor(StaticKeysJumpTableAddress - SectionAddress);
+  AddressExtractor AE(StaticKeysJumpSection->getContents(), SectionAddress,
+                      BC.AsmInfo->isLittleEndian(),
+                      BC.AsmInfo->getCodePointerSize());
+  AddressExtractor::Cursor Cursor(StaticKeysJumpTableAddress - SectionAddress);
   uint32_t EntryID = 0;
   while (Cursor && Cursor.tell() < Stop->getAddress() - SectionAddress) {
-    const uint64_t JumpAddress =
-        SectionAddress + Cursor.tell() + (int32_t)DE.getU32(Cursor);
-    const uint64_t TargetAddress =
-        SectionAddress + Cursor.tell() + (int32_t)DE.getU32(Cursor);
-    const uint64_t KeyAddress =
-        SectionAddress + Cursor.tell() + (int64_t)DE.getU64(Cursor);
+    const uint64_t JumpAddress = AE.getPCRelAddress32(Cursor);
+    const uint64_t TargetAddress = AE.getPCRelAddress32(Cursor);
+    const uint64_t KeyAddress = AE.getPCRelAddress64(Cursor);
 
     // Consume the status of the cursor.
     if (!Cursor)
@@ -1859,21 +1864,18 @@ Error LinuxKernelRewriter::updateStaticKeysJumpTablePostEmit() {
     return Error::success();
 
   const uint64_t SectionAddress = StaticKeysJumpSection->getAddress();
-  DataExtractor DE(StaticKeysJumpSection->getOutputContents(),
-                   BC.AsmInfo->isLittleEndian(),
-                   BC.AsmInfo->getCodePointerSize());
-  DataExtractor::Cursor Cursor(StaticKeysJumpTableAddress - SectionAddress);
+  AddressExtractor AE(StaticKeysJumpSection->getOutputContents(),
+                      SectionAddress, BC.AsmInfo->isLittleEndian(),
+                      BC.AsmInfo->getCodePointerSize());
+  AddressExtractor::Cursor Cursor(StaticKeysJumpTableAddress - SectionAddress);
   const BinaryData *Stop = BC.getBinaryDataByName("__stop___jump_table");
   uint32_t EntryID = 0;
   uint64_t NumShort = 0;
   uint64_t NumLong = 0;
   while (Cursor && Cursor.tell() < Stop->getAddress() - SectionAddress) {
-    const uint64_t JumpAddress =
-        SectionAddress + Cursor.tell() + (int32_t)DE.getU32(Cursor);
-    const uint64_t TargetAddress =
-        SectionAddress + Cursor.tell() + (int32_t)DE.getU32(Cursor);
-    const uint64_t KeyAddress =
-        SectionAddress + Cursor.tell() + (int64_t)DE.getU64(Cursor);
+    const uint64_t JumpAddress = AE.getPCRelAddress32(Cursor);
+    const uint64_t TargetAddress = AE.getPCRelAddress32(Cursor);
+    const uint64_t KeyAddress = AE.getPCRelAddress64(Cursor);
 
     // Consume the status of the cursor.
     if (!Cursor)

clang-tools-extra/clang-tidy/readability/ContainerSizeEmptyCheck.h

@@ -18,10 +18,11 @@ namespace clang::tidy::readability {
 /// a call to `empty()`.
 ///
 /// The emptiness of a container should be checked using the `empty()` method
-/// instead of the `size()` method. It shows clearer intent to use `empty()`.
-/// Furthermore some containers may implement the `empty()` method but not
-/// implement the `size()` method. Using `empty()` whenever possible makes it
-/// easier to switch to another container in the future.
+/// instead of the `size()`/`length()` method. It shows clearer intent to use
+/// `empty()`. Furthermore some containers may implement the `empty()` method
+/// but not implement the `size()` or `length()` method. Using `empty()`
+/// whenever possible makes it easier to switch to another container in the
+/// future.
 class ContainerSizeEmptyCheck : public ClangTidyCheck {
 public:
   ContainerSizeEmptyCheck(StringRef Name, ClangTidyContext *Context);

clang-tools-extra/test/clang-tidy/checkers/cppcoreguidelines/pro-type-vararg.cpp

@@ -51,8 +51,8 @@ void my_printf(const char* format, ...) {
 
 int my_vprintf(const char* format, va_list arg ); // OK to declare function taking va_list
 
-void ignoredBuiltinsTest() {
-  (void)__builtin_assume_aligned(0, 8);
+void ignoredBuiltinsTest(void *ptr) {
+  (void)__builtin_assume_aligned(ptr, 8);
   (void)__builtin_constant_p(0);
   (void)__builtin_fpclassify(0, 0, 0, 0, 0, 0.f);
   (void)__builtin_isinf_sign(0.f);

clang/docs/ReleaseNotes.rst

@@ -453,6 +453,10 @@ New Compiler Flags
 - The ``-Warray-compare-cxx26`` warning has been added to warn about array comparison
   starting from C++26, this warning is enabled as an error by default.
 
+- '-fsanitize-merge' (default) and '-fno-sanitize-merge' have been added for
+  fine-grained control of which UBSan checks are allowed to be merged by the
+  backend (for example, -fno-sanitize-merge=bool,enum).
+
 Deprecated Compiler Flags
 -------------------------
 
@@ -511,6 +515,8 @@ Removed Compiler Flags
   derivatives) is now removed, since it's no longer possible to suppress the
   diagnostic (see above). Users can expect an `unknown warning` diagnostic if
   it's still in use.
+- The experimental flag '-ubsan-unique-traps' has been removed. It is
+  superseded by '-fno-sanitize-merge'.
 
 Attribute Changes in Clang
 --------------------------
@@ -882,6 +888,7 @@ Bug Fixes to C++ Support
 - Clang no longer rejects deleting a pointer of incomplete enumeration type. (#GH99278)
 - Fixed recognition of ``std::initializer_list`` when it's surrounded with ``extern "C++"`` and exported
   out of a module (which is the case e.g. in MSVC's implementation of ``std`` module). (#GH118218)
+- Fixed a pack expansion issue in checking unexpanded parameter sizes. (#GH17042)
 
 Bug Fixes to AST Handling
 ^^^^^^^^^^^^^^^^^^^^^^^^^
@@ -1233,6 +1240,8 @@ Sanitizers
   ``-fsanitize=type`` flag. This sanitizer detects violations of C/C++ type-based
   aliasing rules.
 
+- Implemented ``-f[no-]sanitize-trap=local-bounds``, and ``-f[no-]sanitize-recover=local-bounds``.
+
 Python Binding Changes
 ----------------------
 - Fixed an issue that led to crashes when calling ``Type.get_exception_specification_kind``.