TheRock/rocm-systems/.github/workflows/rocprofiler-sdk-code_coverage.yml

name: rocprofiler-sdk Code Coverage

permissions:
  contents: read

on:
  workflow_dispatch:
  push:
    branches:
      - develop
    paths:
      - 'projects/rocprofiler-sdk/**'
      - '.github/actions/rocprofiler-sdk-util/**'
      - '.github/workflows/rocprofiler-sdk-code_coverage.yml'
      - '.github/actions/rocprofiler-sdk-util/action.yml'
      - '.github/actions/rocprofiler-sdk-util/resolve-ci-config/**'
      - '!**/*.md'
      - '!**/*.rtf'
      - '!**/*.rst'
      - '!**/.markdownlint-ci2.yaml'
      - '!**/.readthedocs.yaml'
      - '!**/.spellcheck.local.yaml'
      - '!**/.wordlist.txt'
      - '!projects/rocprofiler-sdk/CODEOWNERS'
      - '!projects/rocprofiler-sdk/source/docs/**'
  pull_request:
    paths:
      - 'projects/rocprofiler-sdk/**'
      - '.github/actions/rocprofiler-sdk-util/**'
      - '.github/workflows/rocprofiler-sdk-code_coverage.yml'
      - '.github/actions/rocprofiler-sdk-util/action.yml'
      - '.github/actions/rocprofiler-sdk-util/resolve-ci-config/**'
      - '!**/*.md'
      - '!**/*.rtf'
      - '!**/*.rst'
      - '!**/.markdownlint-ci2.yaml'
      - '!**/.readthedocs.yaml'
      - '!**/.spellcheck.local.yaml'
      - '!**/.wordlist.txt'
      - '!projects/rocprofiler-sdk/CODEOWNERS'
      - '!projects/rocprofiler-sdk/source/docs/**'

concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: true

env:
  # TODO(jrmadsen): replace LD_RUNPATH_FLAG, GPU_TARGETS, etc. with internal handling in cmake
  ROCM_PATH: "/opt/rocm"
  GPU_TARGETS: "gfx906 gfx908 gfx90a gfx942 gfx950 gfx1030 gfx1100 gfx1101 gfx1102 gfx1201"
  PATH: "/usr/bin:$PATH"
  SUBMODULE_PATHS: "projects/rocprofiler-sdk projects/aqlprofile projects/rocprofiler-register projects/rocr-runtime projects/clr projects/hip"
  ## No tests should be excluded here except for extreme emergencies; tests should only be disabled in CMake
  ## A task should be assigned directly to fix the issue
  ## Scratch memory tests need to be fixed for ROCm 7.0 release
  navi3_EXCLUDE_TESTS_REGEX: ""
  vega20_EXCLUDE_TESTS_REGEX: ""
  mi200_EXCLUDE_TESTS_REGEX: ""
  mi300_EXCLUDE_TESTS_REGEX: ""
  mi300a_EXCLUDE_TESTS_REGEX: ""
  mi325_EXCLUDE_TESTS_REGEX: ""
  mi3xx_EXCLUDE_TESTS_REGEX: ""
  navi4_EXCLUDE_TESTS_REGEX: ""
  navi3_EXCLUDE_LABEL_REGEX: ""
  vega20_EXCLUDE_LABEL_REGEX: ""
  mi200_EXCLUDE_LABEL_REGEX: ""
  mi300_EXCLUDE_LABEL_REGEX: ""
  mi300a_EXCLUDE_LABEL_REGEX: ""
  mi325_EXCLUDE_LABEL_REGEX: ""
  mi3xx_EXCLUDE_LABEL_REGEX: ""
  navi4_EXCLUDE_LABEL_REGEX: ""
  GLOBAL_CMAKE_OPTIONS: ""
  ENABLE_ROCR_BUILD: "true"
  ENABLE_HIP_CLR_BUILD: "true"
  ENABLE_ROCPROFILER_REGISTER_BUILD: "true"
  ENABLE_AQLPROFILE_BUILD: "true"
  ENABLE_ROCDECODE_BUILD: "true"
  ENABLE_ROCJPEG_BUILD: "true"
  ENABLE_HIPFILE_BUILD: "true"

jobs:
  ci_config:
    if: ${{ github.event_name != 'pull_request' || !github.event.pull_request.head.repo.fork }}
    name: Load CI config
    runs-on: ubuntu-latest
    permissions:
      contents: read
    outputs:
      gfx94x_runner: ${{ steps.runner.outputs.gfx94x_runner }}
      gfx94x_sandbox_runner: ${{ steps.runner.outputs.gfx94x_sandbox_runner }}
    steps:
      - name: Checkout workflow actions
        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
        with:
          sparse-checkout: .github/actions/rocprofiler-sdk-util/resolve-ci-config
          set-safe-directory: true

      - name: Resolve runner labels
        id: runner
        uses: ./.github/actions/rocprofiler-sdk-util/resolve-ci-config

  code-coverage:
    if: ${{ github.event_name != 'pull_request' || !github.event.pull_request.head.repo.fork }}
    needs: ci_config
    name: Code Coverage • ${{ matrix.system.gpu }} • ${{ matrix.system.os }}
    permissions:
      contents: write
      pull-requests: write
    strategy:
      # fail-fast: false
      matrix:
        system:
          - { gpu: 'mi325', os: 'ubuntu-22.04', build-type: 'Release', gpu-target: 'gfx94X' }

    runs-on: ${{ needs.ci_config.outputs.gfx94x_runner }}
    container:
      image: docker.io/rocm/rocprofiler-private:${{ matrix.system.os }}-${{ matrix.system.gpu-target }}-latest
      credentials:
        username: ${{ secrets.ROCPROFILER_AZURE_CI_USER }}
        password: ${{ secrets.ROCPROFILER_AZURE_CI_PASS }}
      env:
        DEBIAN_FRONTEND: noninteractive
      options: --privileged --cap-add=SYS_PTRACE --security-opt seccomp=unconfined

    # define this for containers
    env:
      GIT_DISCOVERY_ACROSS_FILESYSTEM: 1
      GPU_RUNNER: ${{ needs.ci_config.outputs.gfx94x_runner }}
      GCC_COMPILER_VERSION: 11

    steps:
    - name: Clone ROCProfiler SDK & AQLProfile & ROCProfiler Register & ROCR-Runtime
      uses: actions/checkout@v6
      with:
        sparse-checkout: |
          .github/actions/rocprofiler-sdk-util
          projects/rocprofiler-sdk
          projects/aqlprofile
          projects/rocprofiler-register
          projects/rocr-runtime
          projects/clr
          projects/hip
          projects/hipfile
        submodules: false
        set-safe-directory: true

    - name: Compute submodule cache key
      id: submods
      shell: bash
      run: |
        git config --global --add safe.directory '*'
        git submodule status --recursive -- ${SUBMODULE_PATHS} > .git-submodules-status-raw
        awk '{print $1,$2}' .git-submodules-status-raw > .git-submodules-status
        echo "hash=$(sha256sum .git-submodules-status | cut -d' ' -f1)" >> "$GITHUB_OUTPUT"
        # cache scoped submodule working trees plus their .git/modules gitdirs
        awk '{print $2}' .git-submodules-status-raw > .git-submodule-paths
        for p in ${SUBMODULE_PATHS}; do echo ".git/modules/${p}"; done >> .git-submodule-paths
        { echo "paths<<EOF"; cat .git-submodule-paths; echo "EOF"; } >> "$GITHUB_OUTPUT"

    - name: Restore submodule cache
      uses: actions/cache@v6
      with:
        path: |
          ${{ steps.submods.outputs.paths }}
        key: submods-scoped-v1-${{ runner.os }}-${{ steps.submods.outputs.hash }}
        restore-keys: |
          submods-scoped-v1-${{ runner.os }}-

    - name: Init/Update submodules
      run: git submodule update --init --recursive --jobs 16 -- ${SUBMODULE_PATHS}

    - name: Clone ROCDecode
      uses: actions/checkout@v6
      with:
        repository: 'ROCm/rocDecode'
        ref: 'release/rocm-rel-7.0'
        set-safe-directory: true
        path: 'rocDecode'

    - name: Clone ROCJPEG
      uses: actions/checkout@v6
      with:
        repository: 'ROCm/rocJPEG'
        ref: 'release/rocm-rel-7.0'
        set-safe-directory: true
        path: 'rocJPEG'

    - name: Load Existing XML Code Coverage
      if: github.event_name == 'pull_request'
      id: load-coverage
      uses: actions/cache@v6
      with:
        key: ${{ github.event.pull_request.base.sha }}-codecov
        path: .codecov/**

    - name: Copy Existing XML Code Coverage
      if: github.event_name == 'pull_request'
      shell: bash
      run: |
        if [ -d .codecov ]; then cp -r .codecov .codecov.ref; fi

    - name: Configure Env
      shell: bash
      run: |
        echo "${PATH}:/usr/local/bin:${HOME}/.local/bin" >> $GITHUB_PATH
        echo "LD_LIBRARY_PATH=${LD_LIBRARY_PATH}:/usr/local/lib:${HOME}/.local/lib" >> $GITHUB_ENV

    - name: Install requirements
      timeout-minutes: 10
      shell: bash
      working-directory: projects/rocprofiler-sdk
      run: |
        git config --global --add safe.directory '*'
        apt-get update
        apt-get install -y build-essential ccache python3-pip gcovr wkhtmltopdf xvfb xfonts-base xfonts-75dpi xfonts-100dpi xfonts-utils xfonts-encodings libfontconfig libdw-dev libsqlite3-dev libdrm-dev file autoconf pkg-config rpm libzstd-dev
        python3 -m pip install -U --user cmake==3.25.0
        python3 -m pip install -U --user -r requirements.txt

    - name: Sync gcov with compilers
      timeout-minutes: 10
      shell: bash
      run:
        apt-get install -y gcc-${{ env.GCC_COMPILER_VERSION }} g++-${{ env.GCC_COMPILER_VERSION }} &&
        update-alternatives
            --install /usr/bin/gcc  gcc  /usr/bin/gcc-${{ env.GCC_COMPILER_VERSION }} 500
            --slave   /usr/bin/g++  g++  /usr/bin/g++-${{ env.GCC_COMPILER_VERSION }}
            --slave   /usr/bin/gcov gcov /usr/bin/gcov-${{ env.GCC_COMPILER_VERSION }}

    - name: Enable PC Sampling
      if: ${{ contains(matrix.system.gpu, 'mi200') || contains(matrix.system.gpu, 'mi300a') }}
      shell: bash
      working-directory: projects/rocprofiler-sdk
      run: |
          echo 'ROCPROFILER_PC_SAMPLING_BETA_ENABLED=1' >> $GITHUB_ENV

    - name: Build ROCProfiler SDK Dependencies
      uses: ./.github/actions/rocprofiler-sdk-util
      with:
        rocm-path: ${{ env.ROCM_PATH }}
        gpu-target: ${{ matrix.system.gpu-target }}
        compiler-launcher: /usr/bin/ccache
        enable-rocr-build: ${{ env.ENABLE_ROCR_BUILD }}
        enable-hip-clr-build: ${{ env.ENABLE_HIP_CLR_BUILD }}
        enable-rocprofiler-register-build: ${{ env.ENABLE_ROCPROFILER_REGISTER_BUILD }}
        enable-aqlprofile-build: ${{ env.ENABLE_AQLPROFILE_BUILD }}
        enable-rocdecode-build: ${{ env.ENABLE_ROCDECODE_BUILD }}
        enable-rocjpeg-build: ${{ env.ENABLE_ROCJPEG_BUILD }}
        enable-hipfile-build: ${{ env.ENABLE_HIPFILE_BUILD }}
        ccache-key: ccache-${{ matrix.system.os }}-${{ needs.ci_config.outputs.gfx94x_runner }}-${{ matrix.system.gpu }}
        ccache-variant: ccache

    - name: Configure, Build, and Test (Total Code Coverage)
      timeout-minutes: 30
      shell: bash
      working-directory: projects/rocprofiler-sdk
      run:
        python3 ./source/scripts/run-ci.py -B build
          --name ${{ github.repository }}-${{ github.ref_name }}-${{ matrix.system.os }}-${{ matrix.system.gpu }}-codecov
          --build-jobs 16
          --site ${{ needs.ci_config.outputs.gfx94x_runner }}
          --gpu-targets ${{ env.GPU_TARGETS }}
          --coverage all
          --run-attempt ${{ github.run_attempt }}
          --
          -DCMAKE_BUILD_TYPE=${{ matrix.system.build-type }}
          -DCMAKE_PREFIX_PATH='${{ env.ROCM_PATH }};${{ env.ROCM_PATH }}/llvm'
          -DPython3_EXECUTABLE=$(which python3)
          ${{ env.GLOBAL_CMAKE_OPTIONS }}
          --
          -LE "${${{ matrix.system.gpu }}_EXCLUDE_LABEL_REGEX}"
          -E "${${{ matrix.system.gpu }}_EXCLUDE_TESTS_REGEX}"

    - name: Configure, Build, and Test (Tests Code Coverage)
      timeout-minutes: 30
      shell: bash
      working-directory: projects/rocprofiler-sdk
      run:
        find build -type f | egrep '\.gcda$' | xargs rm &&
        python3 ./source/scripts/run-ci.py -B build
          --name ${{ github.repository }}-${{ github.ref_name }}-${{ matrix.system.os }}-${{ matrix.system.gpu }}-codecov-tests
          --build-jobs 16
          --site ${{ needs.ci_config.outputs.gfx94x_runner }}
          --gpu-targets ${{ env.GPU_TARGETS }}
          --coverage tests
          --run-attempt ${{ github.run_attempt }}
          --
          -DCMAKE_BUILD_TYPE=${{ matrix.system.build-type }}
          -DCMAKE_PREFIX_PATH='${{ env.ROCM_PATH }};${{ env.ROCM_PATH }}/llvm'
          -DPython3_EXECUTABLE=$(which python3)
          ${{ env.GLOBAL_CMAKE_OPTIONS }}
          --
          -LE "${${{ matrix.system.gpu }}_EXCLUDE_LABEL_REGEX}"
          -E "${${{ matrix.system.gpu }}_EXCLUDE_TESTS_REGEX}"

    - name: Configure, Build, and Test (Samples Code Coverage)
      timeout-minutes: 30
      working-directory: projects/rocprofiler-sdk
      shell: bash
      run:
        find build -type f | egrep '\.gcda$' | xargs rm &&
        python3 ./source/scripts/run-ci.py -B build
          --name ${{ github.repository }}-${{ github.ref_name }}-${{ matrix.system.os }}-${{ matrix.system.gpu }}-codecov-samples
          --build-jobs 16
          --site ${{ needs.ci_config.outputs.gfx94x_runner }}
          --gpu-targets ${{ env.GPU_TARGETS }}
          --coverage samples
          --run-attempt ${{ github.run_attempt }}
          --
          -DCMAKE_BUILD_TYPE=${{ matrix.system.build-type }}
          -DCMAKE_PREFIX_PATH='${{ env.ROCM_PATH }};${{ env.ROCM_PATH }}/llvm'
          -DPython3_EXECUTABLE=$(which python3)
          ${{ env.GLOBAL_CMAKE_OPTIONS }}
          --
          -LE "${${{ matrix.system.gpu }}_EXCLUDE_LABEL_REGEX}"
          -E "${${{ matrix.system.gpu }}_EXCLUDE_TESTS_REGEX}"

    - name: Save XML Code Coverage
      id: save-coverage
      uses: actions/cache/save@v4
      with:
        key: ${{ github.sha }}-codecov
        path: |
          projects/rocprofiler-sdk/.codecov/*.xml

    - id: generatereport
      name: Generate Code Coverage Comment
      if: github.event_name == 'pull_request'
      timeout-minutes: 10
      working-directory: projects/rocprofiler-sdk
      shell: bash
      run: |
        echo "PWD: ${PWD}"
        ls -la

        for i in "all" "tests" "samples"; do
            wkhtmltoimage --enable-local-file-access --quality 70 .codecov/${i}.html .codecov/${i}.png
        done
        ls -la .codecov
        which -a git
        git --version

        ./source/scripts/upload-image-to-github.py --bot --token ${{ secrets.GITHUB_TOKEN }} --files .codecov/{all,tests,samples}.png --output-dir .codecov --name pr-${{ github.event.pull_request.number }}

        echo -e "\n${PWD}:"
        ls -la .

        echo -e "\n.codecov:"
        ls -la .codecov

        get-base-md-contents() { cat .codecov/${1}.png.md; }
        get-full-md-contents() { cat .codecov/${1}.png.md .codecov/${1}.md; }
        cat << EOF > .codecov/report.md
        # Code Coverage Report

        ## Tests Only
        $(get-base-md-contents tests)

        ## Samples Only
        $(get-base-md-contents samples)

        ## Tests + Samples
        $(get-base-md-contents all)

        <!-- code-coverage-comment-identifier -->
        EOF

        echo 'CODECOVERAGE_REPORT<<EOF' > $GITHUB_OUTPUT
        cat .codecov/report.md >> $GITHUB_OUTPUT
        echo 'EOF' >> $GITHUB_OUTPUT

    - name: Write Code Coverage Comment
      if: github.event_name == 'pull_request'
      timeout-minutes: 10
      uses: actions/github-script@v6
      env:
        COMMENT_BODY: |
          <details>
          <summary><h2>Code Coverage Report</h2></summary>

          <!-- Keep the empty line above to render markdown properly -->
          ${{ steps.generatereport.outputs.CODECOVERAGE_REPORT }}
          </details>
      with:
        github-token: ${{ secrets.GITHUB_TOKEN }}
        script: |
          const commentIdentifier = '<!-- code-coverage-comment-identifier -->'; // Used to identify codecov comment
            const commentBody = process.env.COMMENT_BODY;

            // Fetch existing comments
            const { data: comments } = await github.rest.issues.listComments({
              owner: context.repo.owner,
              repo: context.repo.repo,
              issue_number: context.payload.pull_request.number,
            });

            // Check for an existing comment
            const existingComment = comments.find(comment => comment.body.includes(commentIdentifier));

            if (existingComment) {
              // Update the existing comment
              await github.rest.issues.updateComment({
                owner: context.repo.owner,
                repo: context.repo.repo,
                comment_id: existingComment.id,
                body: commentBody,
              });
              core.info(`Updated comment with ID: ${existingComment.id}`);
            } else {
              // Create a new comment
              await github.rest.issues.createComment({
                owner: context.repo.owner,
                repo: context.repo.repo,
                issue_number: context.payload.pull_request.number,
                body: commentBody,
              });
              core.info('Created a new comment.');
            }

    - name: Archive Code Coverage Data
      if: ${{ github.event_name == 'workflow_dispatch' }}
      uses: actions/upload-artifact@v7
      with:
        name: code-coverage-details
        path: |
          ${{github.workspace}}/projects/rocprofiler-sdk/.codecov/*

    - name: Verify Test Labels
      timeout-minutes: 5
      shell: bash
      working-directory: projects/rocprofiler-sdk
      run: |
        pushd build
        #
        # if following fails, there is a test that does not have
        # a label identifying it as sample or test (unit or integration).
        # Recommended labels are:
        #   - samples
        #   - unittests
        #   - integration-tests
        #
        ctest -N -LE 'samples|tests' -E "${${{ matrix.system.gpu }}_EXCLUDE_TESTS_REGEX}" -O ctest.mislabeled.log
        grep 'Total Tests: 0' ctest.mislabeled.log
        #
        # if following fails, then there is overlap between the labels.
        # A test cannot both be a sample and (unit/integration) test.
        #
        ctest -N -O ctest.all.log
        ctest -N -O ctest.samples.log -L samples
        ctest -N -O ctest.tests.log -L tests
        NUM_ALL=$(grep 'Total Tests:' ctest.all.log | awk '{print $NF}')
        NUM_SAMPLE=$(grep 'Total Tests:' ctest.samples.log | awk '{print $NF}')
        NUM_TEST=$(grep 'Total Tests:' ctest.tests.log | awk '{print $NF}')
        NUM_SUM=$((${NUM_SAMPLE} + ${NUM_TEST}))
        echo "Total tests: ${NUM_ALL}"
        echo "Total labeled tests: ${NUM_SUM}"
        if [ ${NUM_ALL} != ${NUM_SUM} ]; then
            echo "Test label overlap"
            exit 1
        fi
        popd