Merge remote-tracking branch 'origin/main' into HEAD

Author: tstellar

.github/workflows/release-binaries.yml

@@ -261,7 +261,6 @@ jobs:
         release_dir=`find ${{ steps.setup-stage.outputs.build-prefix }}/build -iname 'stage2-bins'`
         mv $release_dir/${{ needs.prepare.outputs.release-binary-filename }} .
 
-
     - name: Build Windows
       id: build-windows
       if: runner.os == 'Windows'
@@ -312,8 +311,7 @@ jobs:
       - prepare
       - build-release-package
     if: >-
-      github.event_name != 'pull_request' &&
-      needs.prepare.outputs.upload == 'true'
+      github.event_name != 'pull_request'
     runs-on: ubuntu-24.04
     permissions:
       contents: write # For release uploads
@@ -334,6 +332,6 @@ jobs:
         uses: ./.github/workflows/upload-release-artifact
         with:
           artifact-id: ${{ needs.build-release-package.outputs.artifact-id }}
-          attestation-name: ${{ needs.prepare.outputs.release-binary-filename }}
+          attestation-name: ${{ runner.os }}-${{ runner.arch }}-release-binary-attestation
           digest: ${{ needs.build-release-package.outputs.digest }}
-          upload: true
+          upload: ${{ needs.prepare.outputs.upload }}

.github/workflows/release-sources.yml

@@ -108,7 +108,7 @@ jobs:
       attestations: write
     steps:
       - name: Checkout Release Scripts
-        uses: actions/checkout@08c6903cd8c0fde910a37f88322edcfb5dd907a8 # v5.0.0
+        uses: actions/checkout@8e8c483db84b4bee98b60c0593521ed34d9990e8 # v6.0.1
         with:
           sparse-checkout: |
             .github/workflows/upload-release-artifact

clang-tools-extra/clang-tidy/fuchsia/MultipleInheritanceCheck.cpp

@@ -23,23 +23,19 @@ AST_MATCHER(CXXRecordDecl, hasBases) {
 }
 } // namespace
 
-// Adds a node (by name) to the interface map, if it was not present in the map
+// Adds a node to the interface map, if it was not present in the map
 // previously.
 void MultipleInheritanceCheck::addNodeToInterfaceMap(const CXXRecordDecl *Node,
                                                      bool IsInterface) {
-  assert(Node->getIdentifier());
-  const StringRef Name = Node->getIdentifier()->getName();
-  InterfaceMap.insert(std::make_pair(Name, IsInterface));
+  InterfaceMap.try_emplace(Node, IsInterface);
 }
 
 // Returns "true" if the boolean "isInterface" has been set to the
 // interface status of the current Node. Return "false" if the
 // interface status for the current node is not yet known.
 bool MultipleInheritanceCheck::getInterfaceStatus(const CXXRecordDecl *Node,
                                                   bool &IsInterface) const {
-  assert(Node->getIdentifier());
-  const StringRef Name = Node->getIdentifier()->getName();
-  auto Pair = InterfaceMap.find(Name);
+  auto Pair = InterfaceMap.find(Node);
   if (Pair == InterfaceMap.end())
     return false;
   IsInterface = Pair->second;
@@ -59,9 +55,6 @@ bool MultipleInheritanceCheck::isCurrentClassInterface(
 }
 
 bool MultipleInheritanceCheck::isInterface(const CXXRecordDecl *Node) {
-  if (!Node->getIdentifier())
-    return false;
-
   // Short circuit the lookup if we have analyzed this record before.
   bool PreviousIsInterfaceResult = false;
   if (getInterfaceStatus(Node, PreviousIsInterfaceResult))

clang-tools-extra/clang-tidy/fuchsia/MultipleInheritanceCheck.h

@@ -38,7 +38,7 @@ class MultipleInheritanceCheck : public ClangTidyCheck {
   // Contains the identity of each named CXXRecord as an interface.  This is
   // used to memoize lookup speeds and improve performance from O(N^2) to O(N),
   // where N is the number of classes.
-  llvm::StringMap<bool> InterfaceMap;
+  llvm::DenseMap<const CXXRecordDecl *, bool> InterfaceMap;
 };
 
 } // namespace clang::tidy::fuchsia

clang-tools-extra/docs/ReleaseNotes.rst

@@ -441,6 +441,12 @@ Changes in existing checks
   correctly ignore ``std::array`` and other array-like containers when
   `IgnoreArrays` option is set to `true`.
 
+- Improved :doc:`fuchsia-multiple-inheritance
+  <clang-tidy/checks/fuchsia/multiple-inheritance>`
+  by fixing an issue where the check would only analyze the first class with
+  a given name in the program, missing any subsequent classes with that same
+  name (declared in a different scope).
+
 - Improved :doc:`google-readability-casting
   <clang-tidy/checks/google/readability-casting>` check by adding fix-it
   notes for downcasts and casts to void pointer.

clang-tools-extra/test/clang-tidy/.clang-tidy

@@ -0,0 +1 @@
+Checks: '-*'

clang-tools-extra/test/clang-tidy/checkers/fuchsia/multiple-inheritance.cpp

@@ -148,3 +148,18 @@ void test_no_crash() {
   auto foo = []() {};
   WithTemplBase<decltype(foo)>();
 }
+
+struct S1 {};
+struct S2 {};
+
+struct S3 : S1, S2 {};
+
+namespace N {
+
+struct S1 { int i; };
+struct S2 { int i; };
+
+// CHECK-MESSAGES: [[@LINE+1]]:1: warning: inheriting multiple classes that aren't pure virtual is discouraged [fuchsia-multiple-inheritance]
+struct S3 : S1, S2 {};
+
+} // namespace N

clang-tools-extra/test/clang-tidy/infrastructure/Inputs/param/parameters.txt

@@ -1,2 +1,2 @@
--checks='-*,llvm-namespace-comment'
+-checks=-*,llvm-namespace-comment
 --warnings-as-errors=llvm-namespace-comment

clang/docs/ReleaseNotes.rst

@@ -343,6 +343,7 @@ Modified Compiler Flags
 -----------------------
 - The `-gkey-instructions` compiler flag is now enabled by default when DWARF is emitted for plain C/C++ and optimizations are enabled. (#GH149509)
 - The `-fconstexpr-steps` compiler flag now accepts value `0` to opt out of this limit. (#GH160440)
+- The `-fdevirtualize-speculatively` compiler flag is now supported to enable speculative devirtualization of virtual function calls, it's disabled by default. (#GH159685)
 
 Removed Compiler Flags
 -------------------------

clang/docs/UsersManual.rst

@@ -2352,6 +2352,56 @@ are listed below.
    pure ThinLTO, as all split regular LTO modules are merged and LTO linked
    with regular LTO.
 
+.. option:: -fdevirtualize-speculatively
+
+   Enable speculative devirtualization optimization where a virtual call
+   can be transformed into a direct call under the assumption that its
+   object is of a particular type. A runtime check is inserted to validate
+   the assumption before making the direct call, and if the check fails,
+   the original virtual call is made instead. This optimization can enable
+   more inlining opportunities and better optimization of the direct call.
+   This is different from whole program devirtualization optimization
+   that rely on global analysis and hidden visibility of the objects to prove
+   that the object is always of a particular type at a virtual call site.
+   This optimization doesn't require global analysis or hidden visibility.
+   This optimization doesn't devirtualize all virtual calls, but only
+   when there's a single implementation of the virtual function in the module.
+   There could be a single implementation of the virtual function
+   either because the function is not overridden in any derived class,
+   or because all objects are instances of the same class/type.
+
+   Ex of IR before the optimization:
+
+   .. code-block:: llvm
+
+     %vtable = load ptr, ptr %BV, align 8, !tbaa !6
+     %0 = tail call i1 @llvm.public.type.test(ptr %vtable, metadata !"_ZTS4Base")
+     tail call void @llvm.assume(i1 %0)
+     %0 = load ptr, ptr %vtable, align 8
+     tail call void %0(ptr noundef nonnull align 8 dereferenceable(8) %BV)
+     ret void
+
+   IR after the optimization:
+
+   .. code-block:: llvm
+
+     %vtable = load ptr, ptr %BV, align 8, !tbaa !12
+     %0 = load ptr, ptr %vtable, align 8
+     %1 = icmp eq ptr %0, @_ZN4Base17virtual_function1Ev
+     br i1 %1, label %if.true.direct_targ, label %if.false.orig_indirect, !prof !15
+     if.true.direct_targ:                              ; preds = %entry
+       tail call void @_ZN4Base17virtual_function1Ev(ptr noundef nonnull align 8 dereferenceable(8) %BV)
+       br label %if.end.icp
+     if.false.orig_indirect:                           ; preds = %entry
+       tail call void %0(ptr noundef nonnull align 8 dereferenceable(8) %BV)
+       br label %if.end.icp
+     if.end.icp:                                       ; preds = %if.false.orig_indirect, %if.true.direct_targ
+       ret void
+
+   This feature is temporarily ignored at the LLVM side when LTO is enabled.
+   TODO: Update the comment when the LLVM side supports this feature for LTO.
+   This feature is turned off by default.
+
 .. option:: -f[no-]unique-source-file-names
 
    When enabled, allows the compiler to assume that each object file
@@ -5216,6 +5266,8 @@ Execute ``clang-cl /?`` to see a list of supported options:
       -fstandalone-debug      Emit full debug info for all types used by the program
       -fstrict-aliasing	      Enable optimizations based on strict aliasing rules
       -fsyntax-only           Run the preprocessor, parser and semantic analysis stages
+      -fdevirtualize-speculatively
+                              Enables speculative devirtualization optimization.
       -fwhole-program-vtables Enables whole-program vtable optimization. Requires -flto
       -gcodeview-ghash        Emit type record hashes in a .debug$H section
       -gcodeview              Generate CodeView debug information