TheRock/rocm-libraries/.github/workflows/therock-ci-linux.yml

name: TheRock CI Linux

on:
  workflow_call:
    inputs:
      therock_ref:
        type: string
        required: true
      docker_image:
        type: string
        required: true
      rocm_package_version:
        type: string
        required: true
      cmake_options:
        type: string
      projects_to_test:
        type: string
      test_type:
        type: string
      amdgpu_families:
        type: string
      test_runs_on:
        type: string
      build_runs_on:
        type: string

permissions:
  contents: read

jobs:
  therock-build-linux:
    name: Build (${{ inputs.amdgpu_families }})
    runs-on: ${{ inputs.build_runs_on }}
    permissions:
      id-token: write
    container:
      image: ${{ inputs.docker_image }}
    strategy:
      fail-fast: true
    env:
      CACHE_DIR: ${{ github.workspace }}/.container-cache
      CCACHE_CONFIGPATH: ${{ github.workspace }}/.ccache/ccache.conf
      AMDGPU_FAMILIES: ${{ inputs.amdgpu_families }}
      TEATIME_FORCE_INTERACTIVE: 0
    steps:
      - name: "Checking out repository for rocm-libraries"
        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2

      - name: Checkout TheRock repository
        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
        with:
          repository: "ROCm/TheRock"
          path: TheRock
          ref: ${{ inputs.therock_ref }}

      - name: Install python deps
        run: |
          pip install -r TheRock/requirements.txt

      - name: Patch rocm-libraries
        run: |
          # Remove patches here if they cannot be applied cleanly, and they have not been deleted from TheRock repo
          # rm ./TheRock/patches/amd-mainline/rocm-libraries/*.patch
          # git -c user.name="therockbot" -c "user.email=therockbot@amd.com" am --whitespace=nowarn ./TheRock/patches/amd-mainline/rocm-libraries/*.patch

      # safe.directory must be set before Runner health status
      - name: Adjust git config
        run: |
          git config --global --add safe.directory $PWD
          git config fetch.parallel 10

      - name: Setup ccache
        run: |
          ./TheRock/build_tools/setup_ccache.py \
            --config-preset "github-oss-dev" \
            --dir "$(dirname $CCACHE_CONFIGPATH)" \
            --local-path "$CACHE_DIR/ccache"

      - name: Runner health status
        run: |
          ./TheRock/build_tools/health_status.py

      - name: Pull DVC files for rocm-libraries # LOGNAME details here https://github.com/ROCm/rocm-libraries/pull/1617
        run: |
          if command -v dvc &> /dev/null; then
            echo "dvc detected"
          else
            echo "Warning, dvc not detected!"
          fi
          LOGNAME=github-runner dvc pull -v

      - name: Fetch sources
        timeout-minutes: 30
        run: |
          ./TheRock/build_tools/fetch_sources.py --jobs 12 --no-include-rocm-libraries --no-include-ml-frameworks

      - name: Configure Projects
        env:
          amdgpu_families: ${{ env.AMDGPU_FAMILIES }}
          package_version: ${{ inputs.rocm_package_version }}
          extra_cmake_options: "-DTHEROCK_ROCM_LIBRARIES_SOURCE_DIR=../ ${{ inputs.cmake_options }}"
          BUILD_DIR: build
        run: |
          python3 TheRock/build_tools/github_actions/build_configure.py

      - name: Build therock-artifacts and therock-dist
        run: cmake --build TheRock/build --target therock-artifacts therock-dist -- -k 0

      - name: Report
        if: ${{ !cancelled() }}
        run: |
          echo "Full SDK du:"
          echo "------------"
          du -h -d 1 TheRock/build/dist/rocm
          echo "Artifacts:"
          echo "----------"
          du -h -d 1 TheRock/build/artifacts
          echo "CCache Stats:"
          echo "-------------"
          ccache -s -v
          tail -v -n +1 .ccache/compiler_check_cache/* > TheRock/build/logs/ccache_compiler_check_cache.log

      - name: Configure AWS Credentials for non-forked repos
        if: ${{ always() && !github.event.pull_request.head.repo.fork }}
        uses: aws-actions/configure-aws-credentials@99214aa6889fcddfa57764031d71add364327e59 # v6.1.3
        with:
          aws-region: us-east-2
          role-to-assume: arn:aws:iam::692859939525:role/therock-ci-external

      - name: Push artifacts
        run: |
          python TheRock/build_tools/artifact_manager.py push \
            --run-id ${{ github.run_id }} \
            --build-dir TheRock/build \
            --stage all

      - name: Post Build Upload
        if: always()
        run: |
          python3 TheRock/build_tools/github_actions/post_build_upload.py \
            --run-id ${{ github.run_id }} \
            --artifact-group ${{ env.AMDGPU_FAMILIES }} \
            --build-dir TheRock/build \
            --upload

      - name: Notify Teams Channel on Failure
        if: ${{ failure() && !github.event.pull_request.head.repo.fork }}
        env:
          PR_NUMBER: ${{ github.event.pull_request.number }}
          PR_TITLE: ${{ github.event.pull_request.title }}
          MIOPEN_WEBHOOK_URL: ${{ secrets.MIOPEN_CI_WEBHOOK_URL }}
          HIPDNN_WEBHOOK_URL: ${{ secrets.HIPDNN_CI_WEBHOOK_URL }}
          COMPONENT_NAME: ${{ inputs.projects_to_test }}
        run: |
          python3 .github/scripts/notify_teams.py \
            --failure-stage build \
            --log-path TheRock/build/logs \
            --webhook-urls "{\"miopen\":\"${MIOPEN_WEBHOOK_URL}\",\"hipdnn\":\"${HIPDNN_WEBHOOK_URL}\"}" \
            --pr-number "${PR_NUMBER}" \
            --pr-title "${PR_TITLE}" \
            --component-name "${COMPONENT_NAME}"

  therock-test-linux:
    name: Test (${{ inputs.amdgpu_families }})
    needs: [therock-build-linux]
    # TODO(TheRock#3288): Re-enable gfx950-dcgpu runners once we get more capacity
    if: ${{ inputs.amdgpu_families != 'gfx950-dcgpu' }}
    uses: ./.github/workflows/therock-test-packages.yml
    with:
      therock_ref: ${{ inputs.therock_ref }}
      projects_to_test: ${{ inputs.projects_to_test }}
      amdgpu_families: ${{ inputs.amdgpu_families }}
      test_runs_on: ${{ inputs.test_runs_on }}
      platform: "linux"
      test_type: ${{ inputs.test_type }}
    secrets: inherit