Auto merge of rust-lang#133912 - MarcoIeni:test-codebuild-runner, r=<…

…try>

[experiment] test code build runner

try-job: x86_64-fuchsia
try-job: x86_64-gnu
try-job: dist-arm-linux
try-job: dist-powerpc64le-linux
try-job: dist-x86_64-linux
try-job: dist-x86_64-linux-alt

.github/workflows/ci.yml

@@ -112,6 +112,26 @@ jobs:
         run: src/ci/scripts/free-disk-space.sh
         if: matrix.free_disk
 
+      - name: free up more disk space
+        if: matrix.free_disk
+        run: |
+          sudo apt purge -y --autoremove \
+            '^java-*'                    \
+            'groff'                      \
+            'groff-base'                 \
+            '^libllvm.*'                 \
+            '^llvm.*'                    \
+            'gcc'                        \
+            'gcc-11'                     \
+            'libicu-dev'                 \
+            '^vim.*'                     \
+            'python3-breezy'             \
+          && sudo rm -rf /var/lib/apt/lists/*
+          df -h /
+
+      - name: Show installed packages ordered by size
+        run: dpkg-query -W --showformat='${Installed-Size} ${Package}\n' | sort -nr
+
       # Rust Log Analyzer can't currently detect the PR number of a GitHub
       # Actions build on its own, so a hint in the log message is needed to
       # point it in the right direction.
@@ -174,6 +194,8 @@ jobs:
 
       - name: enable ipv6 on Docker
         run: src/ci/scripts/enable-docker-ipv6.sh
+        # Don't run on codebuild because systemctl is not available
+        if: ${{ !contains(matrix.os, 'codebuild-ubuntu') }}
 
       # Disable automatic line ending conversion (again). On Windows, when we're
       # installing dependencies, something switches the git configuration directory or

src/bootstrap/bootstrap.py

@@ -1002,10 +1002,14 @@ def build_bootstrap(self):
         env = os.environ.copy()
         if "GITHUB_ACTIONS" in env:
             print("::group::Building bootstrap")
+            print("Current home directory:", os.path.expanduser("~"))
+            print("Current working directory:", os.getcwd())
+            print("Current HOME:", env["HOME"])
         else:
             eprint("Building bootstrap")
 
         args = self.build_bootstrap_cmd(env)
+        print("Running", args, "in", self.rust_root)
         # Run this from the source directory so cargo finds .cargo/config
         run(args, env=env, verbose=self.verbose, cwd=self.rust_root)
 
@@ -1015,6 +1019,7 @@ def build_bootstrap(self):
     def build_bootstrap_cmd(self, env):
         """For tests."""
         build_dir = os.path.join(self.build_dir, "bootstrap")
+        print("Building bootstrap in", build_dir)
         if self.clean and os.path.exists(build_dir):
             shutil.rmtree(build_dir)
         # `CARGO_BUILD_TARGET` breaks bootstrap build.

src/ci/docker/host-x86_64/dist-arm-linux/Dockerfile

@@ -1,4 +1,4 @@
-FROM ubuntu:22.04
+FROM ghcr.io/marcoieni/ubuntu:22.04
 
 COPY scripts/cross-apt-packages.sh /scripts/
 RUN sh /scripts/cross-apt-packages.sh

src/ci/docker/host-x86_64/dist-powerpc64le-linux/Dockerfile

@@ -1,4 +1,4 @@
-FROM ubuntu:22.04
+FROM ghcr.io/marcoieni/ubuntu:22.04
 
 COPY scripts/cross-apt-packages.sh /scripts/
 RUN sh /scripts/cross-apt-packages.sh

src/ci/docker/host-x86_64/dist-x86_64-linux/Dockerfile

@@ -2,7 +2,7 @@
 # CentOS 7 has headers for kernel 3.10, but that's fine as long as we don't
 # actually use newer APIs in rustc or std without a fallback. It's more
 # important that we match glibc for ELF symbol versioning.
-FROM centos:7
+FROM ghcr.io/marcoieni/centos:7
 
 WORKDIR /build
 

src/ci/docker/host-x86_64/x86_64-fuchsia/Dockerfile

@@ -2,7 +2,7 @@
 # Fuchsia as an integration test of the toolchain. See the build-fuchsia.sh
 # script in this directory for more details.
 
-FROM ubuntu:22.04
+FROM ghcr.io/marcoieni/ubuntu:22.04
 
 ARG DEBIAN_FRONTEND=noninteractive
 RUN apt-get update && apt-get install -y --no-install-recommends \
@@ -75,5 +75,7 @@ ENV RUST_CONFIGURE_ARGS \
   --set target.x86_64-unknown-fuchsia.linker=/usr/local/bin/ld.lld
 
 ENV SCRIPT \
+    echo "x install" && df -h / && \
     python3 ../x.py install --target $TARGETS compiler/rustc library/std clippy && \
+    echo "build fuchsia" && df -h / && \
     bash ../src/ci/docker/host-x86_64/x86_64-fuchsia/build-fuchsia.sh

src/ci/docker/host-x86_64/x86_64-fuchsia/build-fuchsia.sh

@@ -43,6 +43,7 @@ jiri=.jiri_root/bin/jiri
 set -x
 
 if [ -z "$KEEP_CHECKOUT" ]; then
+    echo "downloading jiri" && df -h /
     # This script will:
     # - create a directory named "fuchsia" if it does not exist
     # - download "jiri" to "fuchsia/.jiri_root/bin"
@@ -52,11 +53,13 @@ if [ -z "$KEEP_CHECKOUT" ]; then
 
     cd $checkout
 
+    echo "running jiri init" && df -h /
     $jiri init \
         -partial=true \
         -analytics-opt=false \
         .
 
+    echo "running jiri import" && df -h /
     $jiri import \
         -name=integration \
         -revision=$INTEGRATION_SHA \
@@ -65,24 +68,41 @@ if [ -z "$KEEP_CHECKOUT" ]; then
         "https://fuchsia.googlesource.com/integration"
 
     if [ -d ".git" ]; then
+        echo "Wipe out any local changes" && df -h /
         # Wipe out any local changes if we're reusing a checkout.
         git checkout --force JIRI_HEAD
     fi
 
+    echo "running jiri update" && df -h /
     $jiri update -autoupdate=false
 
     echo integration commit = $(git -C integration rev-parse HEAD)
 
     for git_ref in "${PICK_REFS[@]}"; do
+        echo "running git fetch" && df -h /
         git fetch https://fuchsia.googlesource.com/fuchsia $git_ref
+        echo "running git cherry-pick" && df -h /
         git cherry-pick --no-commit FETCH_HEAD
     done
 else
-    echo Reusing existing Fuchsia checkout
+    echo Reusing existing Fuchsia checkout && df -h /
     cd $checkout
 fi
 
 # Run the script inside the Fuchsia checkout responsible for building Fuchsia.
 # You can change arguments to the build by setting KEEP_CHECKOUT=1 above and
 # modifying them in build_fuchsia_from_rust_ci.sh.
+echo "running build_fuchsia_from_rust_ci.sh" && df -h /
+
+# Start background monitoring of disk space
+(
+    while true; do
+        df -h /
+        sleep 20
+    done
+) &
+MONITOR_PID=$!
+
 bash scripts/rust/build_fuchsia_from_rust_ci.sh
+
+kill $MONITOR_PID

src/ci/docker/host-x86_64/x86_64-gnu/Dockerfile

@@ -1,4 +1,4 @@
-FROM ubuntu:22.04
+FROM ghcr.io/marcoieni/ubuntu:22.04
 
 ARG DEBIAN_FRONTEND=noninteractive
 RUN apt-get update && apt-get install -y --no-install-recommends \

src/ci/docker/run.sh

@@ -39,6 +39,7 @@ root_dir="`dirname $src_dir`"
 source "$ci_dir/shared.sh"
 
 if isCI; then
+    echo "CI detected"
     objdir=$root_dir/obj
 else
     objdir=$root_dir/obj/$image
@@ -53,6 +54,7 @@ fi
 CACHE_DOMAIN="${CACHE_DOMAIN:-ci-caches.rust-lang.org}"
 
 if [ -f "$docker_dir/$image/Dockerfile" ]; then
+    echo "Dockerfile found for $image"
     hash_key=/tmp/.docker-hash-key.txt
     rm -f "${hash_key}"
     echo $image >> $hash_key
@@ -153,6 +155,7 @@ if [ -f "$docker_dir/$image/Dockerfile" ]; then
           --output=type=docker
     # On auto/try builds, we can also write to the cache.
     else
+        echo "Logging into the Docker registry"
         # Log into the Docker registry, so that we can read/write cache and the final image
         echo ${DOCKER_TOKEN} | docker login ${REGISTRY} \
             --username ${REGISTRY_USERNAME} \
@@ -163,6 +166,7 @@ if [ -f "$docker_dir/$image/Dockerfile" ]; then
         docker buildx create --use --driver docker-container \
           --driver-opt image=${GHCR_BUILDKIT_IMAGE}
 
+        echo "Building Docker image with cache"
         # Build the image using registry caching backend
         retry docker \
           buildx \
@@ -171,11 +175,13 @@ if [ -f "$docker_dir/$image/Dockerfile" ]; then
           --cache-to type=registry,ref=${CACHE_IMAGE_TAG},compression=zstd \
           --output=type=docker
 
+        echo "Docker image built"
         # Print images for debugging purposes
         docker images
 
         # Tag the built image and push it to the registry
         docker tag rust-ci "${IMAGE_TAG}"
+        echo "Pushing Docker image to the registry"
         docker push "${IMAGE_TAG}"
 
         # Record the container registry tag/url for reuse, e.g. by rustup.rs builds
@@ -227,6 +233,7 @@ else
     exit 1
 fi
 
+echo "Creating directories"
 mkdir -p $HOME/.cargo
 mkdir -p $objdir/tmp
 mkdir -p $objdir/cores
@@ -282,24 +289,32 @@ args="$args --privileged"
 # `LOCAL_USER_ID` (recognized in `src/ci/run.sh`) to ensure that files are all
 # read/written as the same user as the bare-metal user.
 if [ -f /.dockerenv ]; then
-  docker create -v /checkout --name checkout alpine:3.4 /bin/true
+  echo "Dockerenv detected. We are in docker-in-docker scenario."
+  docker create -v /checkout --name checkout ghcr.io/marcoieni/alpine:3.4 /bin/true
   docker cp . checkout:/checkout
   args="$args --volumes-from checkout"
 else
   args="$args --volume $root_dir:/checkout$SRC_MOUNT_OPTION"
   args="$args --volume $objdir:/checkout/obj"
   args="$args --volume $HOME/.cargo:/cargo"
   args="$args --volume /tmp/toolstate:/tmp/toolstate"
+fi
 
-  id=$(id -u)
-  if [[ "$id" != 0 && "$(docker version)" =~ Podman ]]; then
-    # Rootless podman creates a separate user namespace, where an inner
-    # LOCAL_USER_ID will map to a different subuid range on the host.
-    # The "keep-id" mode maps the current UID directly into the container.
-    args="$args --env NO_CHANGE_USER=1 --userns=keep-id"
-  else
-    args="$args --env LOCAL_USER_ID=$id"
-  fi
+id=$(id -u)
+if [[ "$id" != 0 && "$(docker version)" =~ Podman ]]; then
+  # Rootless podman creates a separate user namespace, where an inner
+  # LOCAL_USER_ID will map to a different subuid range on the host.
+  # The "keep-id" mode maps the current UID directly into the container.
+  echo "Running in rootless podman"
+  args="$args --env NO_CHANGE_USER=1 --userns=keep-id"
+elif [[ "$id" != 0 ]]; then
+  echo "Running in docker as non-root"
+  args="$args --env LOCAL_USER_ID=$id"
+else
+  echo "Running in docker as root. Using id 1001."
+  # If we're running as root, we don't want to run the container as root,
+  # so we set id `1001` instead of `0`.
+  args="$args --env LOCAL_USER_ID=1001"
 fi
 
 if [ "$dev" = "1" ]

src/ci/github-actions/ci.py

@@ -195,6 +195,19 @@ def format_run_type(run_type: WorkflowRunType) -> str:
         raise AssertionError()
 
 
+# Add new function before main:
+def substitute_github_vars(jobs: list) -> list:
+    """Replace GitHub context variables with environment variables in job configs."""
+    for job in jobs:
+        if "os" in job:
+            job["os"] = (
+                job["os"]
+                .replace("${{ github.run_id }}", os.environ["GITHUB_RUN_ID"])
+                .replace("${{ github.run_attempt }}", os.environ["GITHUB_RUN_ATTEMPT"])
+            )
+    return jobs
+
+
 def get_job_image(job: Job) -> str:
     """
     By default, the Docker image of a job is based on its name.
@@ -265,6 +278,7 @@ def calculate_job_matrix(job_data: Dict[str, Any]):
     if run_type is not None:
         jobs = calculate_jobs(run_type, job_data)
     jobs = skip_jobs(jobs, channel)
+    jobs = substitute_github_vars(jobs)
 
     if not jobs:
         raise Exception("Scheduled job list is empty, this is an error")

src/ci/github-actions/jobs.yml

@@ -52,6 +52,16 @@ runners:
     free_disk: true
     os: ubuntu-22.04-arm
 
+  - &job-linux-8c-codebuild
+    free_disk: true
+    os: codebuild-ubuntu-22-8c-${{ github.run_id }}-${{ github.run_attempt }}
+    <<: *base-job
+
+  - &job-linux-36c-codebuild
+    free_disk: true
+    os: codebuild-ubuntu-22-36c-${{ github.run_id }}-${{ github.run_attempt }}
+    <<: *base-job
+
   - &job-aarch64-linux-8c
     os: ubuntu-22.04-arm64-8core-32gb
 envs:
@@ -152,7 +162,7 @@ auto:
     <<: *job-linux-4c
 
   - name: dist-arm-linux
-    <<: *job-linux-8c
+    <<: *job-linux-8c-codebuild
 
   - name: dist-armhf-linux
     <<: *job-linux-4c
@@ -182,7 +192,7 @@ auto:
     <<: *job-linux-4c
 
   - name: dist-powerpc64le-linux
-    <<: *job-linux-4c-largedisk
+    <<: *job-linux-36c-codebuild
 
   - name: dist-riscv64-linux
     <<: *job-linux-4c
@@ -205,13 +215,13 @@ auto:
   - name: dist-x86_64-linux
     env:
       CODEGEN_BACKENDS: llvm,cranelift
-    <<: *job-linux-16c
+    <<: *job-linux-36c-codebuild
 
   - name: dist-x86_64-linux-alt
     env:
       IMAGE: dist-x86_64-linux
       CODEGEN_BACKENDS: llvm,cranelift
-    <<: *job-linux-16c
+    <<: *job-linux-36c-codebuild
 
   - name: dist-x86_64-musl
     env:
@@ -264,15 +274,15 @@ auto:
     # nightly features to compile, and this job would fail if
     # executed on beta and stable.
     only_on_channel: nightly
-    <<: *job-linux-8c
+    <<: *job-linux-36c-codebuild
 
   # Tests integration with Rust for Linux.
   # Builds stage 1 compiler and tries to compile a few RfL examples with it.
   - name: x86_64-rust-for-linux
     <<: *job-linux-4c
 
   - name: x86_64-gnu
-    <<: *job-linux-4c
+    <<: *job-linux-8c-codebuild
 
   # This job ensures commits landing on nightly still pass the full
   # test suite on the stable channel. There are some UI tests that
@@ -295,10 +305,10 @@ auto:
   - name: x86_64-gnu-debug
     # This seems to be needed because a full stage 2 build + run-make tests
     # overwhelms the storage capacity of the standard 4c runner.
-    <<: *job-linux-4c-largedisk
+    <<: *job-linux-8c-codebuild
 
   - name: x86_64-gnu-distcheck
-    <<: *job-linux-8c
+    <<: *job-linux-8c-codebuild
 
   # The x86_64-gnu-llvm-19 job is split into multiple jobs to run tests in parallel.
   # x86_64-gnu-llvm-19-1 skips tests that run in x86_64-gnu-llvm-19-{2,3}.