Merge branch 'main' into Uncontrolled-data-used-in-path-expression

Author: MUSTAPHA BARKI

.ci/cache_lit_timing_files.py

@@ -17,6 +17,7 @@
 import glob
 
 from google.cloud import storage
+from google.api_core import exceptions
 
 GCS_PARALLELISM = 100
 
@@ -50,7 +51,14 @@ def _maybe_download_timing_file(blob):
 
 def download_timing_files(storage_client, bucket_name: str):
     bucket = storage_client.bucket(bucket_name)
-    blobs = bucket.list_blobs(prefix="lit_timing")
+    try:
+        blobs = bucket.list_blobs(prefix="lit_timing")
+    except exceptions.ClientError as client_error:
+        print(
+            "::warning file=cache_lit_timing_files.py::Failed to list blobs "
+            "in bucket."
+        )
+        sys.exit(0)
     with multiprocessing.pool.ThreadPool(GCS_PARALLELISM) as thread_pool:
         futures = []
         for timing_file_blob in blobs:
@@ -60,7 +68,13 @@ def download_timing_files(storage_client, bucket_name: str):
                 )
             )
         for future in futures:
-            future.get()
+            future.wait()
+            if not future.successful():
+                print(
+                    "::warning file=cache_lit_timing_files.py::Failed to "
+                    "download lit timing file."
+                )
+                continue
     print("Done downloading")
 
 

.github/new-prs-labeler.yml

@@ -722,7 +722,12 @@ mlgo:
   - llvm/include/llvm/Analysis/IR2Vec.h
   - llvm/lib/Analysis/IR2Vec.cpp
   - llvm/lib/Analysis/models/**
+  - llvm/include/llvm/CodeGen/MIR2Vec.h
+  - llvm/lib/CodeGen/MIR2Vec.cpp
   - llvm/test/Analysis/IR2Vec/**
+  - llvm/test/CodeGen/MIR2Vec/**
+  - llvm/unittests/Analysis/IR2VecTest.cpp
+  - llvm/unittests/CodeGen/MIR2VecTest.cpp
   - llvm/tools/llvm-ir2vec/**
   - llvm/docs/CommandGuide/llvm-ir2vec.rst
 

.github/workflows/build-ci-container-tooling.yml

@@ -0,0 +1,115 @@
+name: Build CI Container
+
+permissions:
+  contents: read
+
+on:
+  push:
+    branches:
+      - main
+    paths:
+      - .github/workflows/build-ci-container-tooling.yml
+      - '.github/workflows/containers/github-action-ci-tooling/**'
+      - llvm/utils/git/requirements_formatting.txt
+      - llvm/utils/git/requirements_linting.txt
+  pull_request:
+    paths:
+      - .github/workflows/build-ci-container-tooling.yml
+      - '.github/workflows/containers/github-action-ci-tooling/**'
+      - llvm/utils/git/requirements_formatting.txt
+      - llvm/utils/git/requirements_linting.txt
+
+jobs:
+  build-ci-container-tooling:
+    if: github.repository_owner == 'llvm'
+    runs-on: ubuntu-24.04
+    steps:
+      - name: Checkout LLVM
+        uses: actions/checkout@08c6903cd8c0fde910a37f88322edcfb5dd907a8 # v5.0.0
+        with:
+          sparse-checkout: |
+            .github/workflows/containers/github-action-ci-tooling/
+            llvm/utils/git/requirements_formatting.txt
+            llvm/utils/git/requirements_linting.txt
+            clang-tools-extra/clang-tidy/tool/clang-tidy-diff.py
+
+      - name: Write Variables
+        id: vars
+        run: |
+          tag=$(git rev-parse --short=12 HEAD)
+          container_name="ghcr.io/$GITHUB_REPOSITORY_OWNER/amd64/ci-ubuntu-24.04"
+          echo "container-name-format=$container_name-code-format" >> $GITHUB_OUTPUT
+          echo "container-name-lint=$container_name-code-lint" >> $GITHUB_OUTPUT
+          echo "container-name-format-tag=$container_name-format:$tag" >> $GITHUB_OUTPUT
+          echo "container-name-lint-tag=$container_name-lint:$tag" >> $GITHUB_OUTPUT
+          echo "container-format-filename=$(echo $container_name-format:$tag  | sed -e 's/\//-/g' -e 's/:/-/g').tar" >> $GITHUB_OUTPUT
+          echo "container-lint-filename=$(echo $container_name-lint:$tag  | sed -e 's/\//-/g' -e 's/:/-/g').tar" >> $GITHUB_OUTPUT
+
+      - name: Build container
+        run: |
+          podman build --target ci-container-code-format \
+          -f .github/workflows/containers/github-action-ci-tooling/Dockerfile \
+          -t ${{ steps.vars.outputs.container-name-format-tag }} .
+          podman build --target ci-container-code-lint \
+          -f .github/workflows/containers/github-action-ci-tooling/Dockerfile \
+          -t ${{ steps.vars.outputs.container-name-lint-tag }} .
+
+      # Save the container so we have it in case the push fails.  This also
+      # allows us to separate the push step into a different job so we can
+      # maintain minimal permissions while building the container.
+      - name: Save container image
+        run: |
+          podman save ${{ steps.vars.outputs.container-name-format-tag }}  >  ${{ steps.vars.outputs.container-format-filename }}
+          podman save ${{ steps.vars.outputs.container-name-lint-tag }}  >  ${{ steps.vars.outputs.container-lint-filename }}
+
+      - name: Upload container image
+        uses: actions/upload-artifact@65c4c4a1ddee5b72f698fdd19549f0f0fb45cf08 # v4.6.0
+        with:
+          name: container-amd64
+          path: "*.tar"
+          retention-days: 14
+
+      - name: Test Container
+        run: |
+          # Use --pull=never to ensure we are testing the just built image.
+          podman run --pull=never --rm -it ${{ steps.vars.outputs.container-name-format-tag }} /usr/bin/bash -x -c 'cd $HOME && clang-format --version | grep version && black --version | grep black'
+          podman run --pull=never --rm -it ${{ steps.vars.outputs.container-name-lint-tag }} /usr/bin/bash -x -c 'cd $HOME && clang-tidy --version | grep version && clang-tidy-diff.py -h | grep usage'
+
+  push-ci-container:
+    if: github.event_name == 'push'
+    needs:
+      - build-ci-container-tooling
+    permissions:
+      packages: write
+    runs-on: ubuntu-24.04
+    env:
+      GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+    steps:
+      - name: Download container
+        uses: actions/download-artifact@634f93cb2916e3fdff6788551b99b062d0335ce0 # v5.0.0
+
+      - name: Push Container
+        run: |
+          function push_container {
+            image_name=$1
+            latest_name=$(echo $image_name | sed 's/:[a-f0-9]\+$/:latest/g')
+            podman tag $image_name $latest_name
+            echo "Pushing $image_name ..."
+            podman push $image_name
+            echo "Pushing $latest_name ..."
+            podman push $latest_name
+          }
+
+          podman login -u ${{ github.actor }} -p $GITHUB_TOKEN ghcr.io
+          for f in $(find . -iname *.tar); do
+            image_name=$(podman load -q -i $f | sed 's/Loaded image: //g')
+            push_container $image_name
+
+            if echo $image_name | grep '/amd64/'; then
+              # For amd64, create an alias with the arch component removed.
+              # This matches the convention used on dockerhub.
+              default_image_name=$(echo $(dirname $(dirname $image_name))/$(basename $image_name))
+              podman tag $image_name $default_image_name
+              push_container $default_image_name
+            fi
+          done

.github/workflows/containers/github-action-ci-tooling/Dockerfile

@@ -0,0 +1,59 @@
+ARG LLVM_VERSION=21.1.0
+
+FROM docker.io/library/ubuntu:24.04 AS llvm-downloader
+ARG LLVM_VERSION
+
+RUN apt-get update && \
+    apt-get install -y wget xz-utils && \
+    wget -O llvm.tar.xz https://github.com/llvm/llvm-project/releases/download/llvmorg-${LLVM_VERSION}/LLVM-${LLVM_VERSION}-Linux-X64.tar.xz && \
+    mkdir -p /llvm-extract && \
+    tar -xvJf llvm.tar.xz -C /llvm-extract \
+        # Only unpack these tools to save space on Github runner.
+        LLVM-${LLVM_VERSION}-Linux-X64/bin/clang-tidy \
+        LLVM-${LLVM_VERSION}-Linux-X64/bin/clang-format && \
+    rm llvm.tar.xz
+
+
+FROM docker.io/library/ubuntu:24.04 AS base
+ENV LLVM_SYSROOT=/opt/llvm
+
+# Need nodejs for some of the GitHub actions.
+# Need git for git-clang-format.
+RUN apt-get update && \
+    DEBIAN_FRONTEND=noninteractive apt-get install -y \
+    git \
+    nodejs \
+    sudo \
+    # These are needed by the premerge pipeline.
+    # Pip is used to install dependent python packages.
+    python3-pip \
+    python-is-python3 && \
+    apt-get clean && \
+    rm -rf /var/lib/apt/lists/*
+
+
+FROM base AS ci-container-code-format
+ARG LLVM_VERSION
+
+COPY --from=llvm-downloader /llvm-extract/LLVM-${LLVM_VERSION}-Linux-X64/bin/clang-format ${LLVM_SYSROOT}/bin/clang-format
+
+ENV PATH=${LLVM_SYSROOT}/bin:${PATH}
+
+# Install dependencies for 'pr-code-format.yml' job
+COPY llvm/utils/git/requirements_formatting.txt requirements_formatting.txt
+RUN pip install -r requirements_formatting.txt --break-system-packages && \
+    rm requirements_formatting.txt
+
+
+FROM base AS ci-container-code-lint
+ARG LLVM_VERSION
+
+COPY --from=llvm-downloader /llvm-extract/LLVM-${LLVM_VERSION}-Linux-X64/bin/clang-tidy ${LLVM_SYSROOT}/bin/
+COPY clang-tools-extra/clang-tidy/tool/clang-tidy-diff.py ${LLVM_SYSROOT}/bin/clang-tidy-diff.py
+
+ENV PATH=${LLVM_SYSROOT}/bin:${PATH}
+
+# Install dependencies for 'pr-code-lint.yml' job
+COPY llvm/utils/git/requirements_linting.txt requirements_linting.txt
+RUN pip install -r requirements_linting.txt --break-system-packages && \
+    rm requirements_linting.txt

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

@@ -236,7 +236,6 @@ jobs:
             **/crash_diagnostics/*
 
   windows:
-    runs-on: windows-2022
     needs: [ stage2 ]
     strategy:
       fail-fast: false
@@ -251,11 +250,12 @@ jobs:
         - { config: mingw-static, mingw: true }
         - { config: mingw-dll-i686, mingw: true }
         - { config: mingw-incomplete-sysroot, mingw: true }
+        - { config: mingw-static, mingw: true, runner: windows-11-arm }
+    runs-on: ${{ matrix.runner != '' && matrix.runner || 'windows-2022' }}
     steps:
       - uses: actions/checkout@08c6903cd8c0fde910a37f88322edcfb5dd907a8 # v5.0.0
       - name: Install dependencies
         run: |
-          choco install -y ninja
           pip install psutil
       - name: Install a current LLVM
         if: ${{ matrix.mingw != true }}
@@ -264,7 +264,7 @@ jobs:
       - name: Install llvm-mingw
         if: ${{ matrix.mingw == true }}
         run: |
-          curl -LO https://github.com/mstorsjo/llvm-mingw/releases/download/20250709/llvm-mingw-20250709-ucrt-x86_64.zip
+          curl -LO https://github.com/mstorsjo/llvm-mingw/releases/download/20250709/llvm-mingw-20250709-ucrt-${{ matrix.runner == 'windows-11-arm' && 'aarch64' || 'x86_64' }}.zip
           powershell Expand-Archive llvm-mingw*.zip -DestinationPath .
           del llvm-mingw*.zip
           mv llvm-mingw* c:\llvm-mingw

bolt/lib/Rewrite/RewriteInstance.cpp

@@ -917,9 +917,6 @@ void RewriteInstance::discoverFileObjects() {
     bool IsData = false;
     uint64_t LastAddr = 0;
     for (const auto &SymInfo : SortedSymbols) {
-      if (LastAddr == SymInfo.Address) // don't repeat markers
-        continue;
-
       MarkerSymType MarkerType = BC->getMarkerType(SymInfo.Symbol);
 
       // Treat ST_Function as code.
@@ -929,8 +926,14 @@ void RewriteInstance::discoverFileObjects() {
         if (IsData) {
           Expected<StringRef> NameOrError = SymInfo.Symbol.getName();
           consumeError(NameOrError.takeError());
-          BC->errs() << "BOLT-WARNING: function symbol " << *NameOrError
-                     << " lacks code marker\n";
+          if (LastAddr == SymInfo.Address) {
+            BC->errs() << "BOLT-WARNING: ignoring data marker conflicting with "
+                          "function symbol "
+                       << *NameOrError << '\n';
+          } else {
+            BC->errs() << "BOLT-WARNING: function symbol " << *NameOrError
+                       << " lacks code marker\n";
+          }
         }
         MarkerType = MarkerSymType::CODE;
       }

bolt/test/AArch64/data-at-0-offset.c

@@ -0,0 +1,23 @@
+## Check that if a data marker is present at the start of a function, the
+## underlying bytes are still treated as code.
+
+# RUN: %clang %cflags %s -o %t.exe
+# RUN: llvm-bolt %t.exe -o %t.bolt --print-cfg 2>&1 | FileCheck %s
+
+# CHECK: BOLT-WARNING: ignoring data marker conflicting with function symbol _start
+
+.text
+.balign 4
+
+## Data marker is emitted because ".long" directive is used instead of ".inst".
+.global _start
+.type _start, %function
+_start:
+  .long 0xcec08000 // sha512su0 v0.2d, v0.2d
+  ret
+.size _start, .-_start
+
+# CHECK-LABEL: Binary Function "_start"
+# CHECK: Entry Point
+# CHECK-NEXT: sha512su0 v0.2d, v0.2d
+

bolt/test/AArch64/function-data-marker.s

@@ -43,6 +43,7 @@
 #include "support/Trace.h"
 #include "clang/AST/Decl.h"
 #include "clang/AST/DeclBase.h"
+#include "clang/AST/DeclTemplate.h"
 #include "clang/Basic/CharInfo.h"
 #include "clang/Basic/LangOptions.h"
 #include "clang/Basic/SourceLocation.h"
@@ -1886,7 +1887,15 @@ class CodeCompleteFlow {
         for (auto &Cand : C.first) {
           if (Cand.SemaResult &&
               Cand.SemaResult->Kind == CodeCompletionResult::RK_Declaration) {
-            auto ID = clangd::getSymbolID(Cand.SemaResult->getDeclaration());
+            const NamedDecl *DeclToLookup = Cand.SemaResult->getDeclaration();
+            // For instantiations of members of class templates, the
+            // documentation will be stored at the member's original
+            // declaration.
+            if (const NamedDecl *Adjusted =
+                    dyn_cast<NamedDecl>(&adjustDeclToTemplate(*DeclToLookup))) {
+              DeclToLookup = Adjusted;
+            }
+            auto ID = clangd::getSymbolID(DeclToLookup);
             if (!ID)
               continue;
             Req.IDs.insert(ID);

clang-tools-extra/clangd/CodeComplete.cpp

@@ -466,7 +466,7 @@ llvm::ArrayRef<ArgStripper::Rule> ArgStripper::rulesFor(llvm::StringRef Arg) {
     } AliasTable[] = {
 #define OPTION(PREFIX, PREFIXED_NAME, ID, KIND, GROUP, ALIAS, ALIASARGS,       \
                FLAGS, VISIBILITY, PARAM, HELPTEXT, HELPTEXTSFORVARIANTS,       \
-               METAVAR, VALUES)                                                \
+               METAVAR, VALUES, SUBCOMMANDIDS_OFFSET)                          \
   {DriverID::OPT_##ID, DriverID::OPT_##ALIAS, ALIASARGS},
 #include "clang/Driver/Options.inc"
 #undef OPTION