CI: use aws codebuild for job dist-arm-linux

Author: marcoieni

.github/workflows/ci.yml

@@ -91,6 +91,17 @@ jobs:
         # Check the `calculate_matrix` job to see how is the matrix defined.
         include: ${{ fromJSON(needs.calculate_matrix.outputs.jobs) }}
     steps:
+      - name: Install cargo in AWS CodeBuild
+        if: matrix.codebuild
+        run: |
+          # Check if cargo is installed
+          if ! command -v cargo &> /dev/null; then
+            echo "Cargo not found, installing Rust..."
+            curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y --profile=minimal
+            # Make cargo available in PATH
+            echo "$HOME/.cargo/bin" >> $GITHUB_PATH
+          fi
+
       - name: disable git crlf conversion
         run: git config --global core.autocrlf false
 
@@ -168,6 +179,8 @@ jobs:
         run: src/ci/scripts/install-ninja.sh
 
       - name: enable ipv6 on Docker
+        # Don't run on codebuild because systemctl is not available
+        if: ${{ !matrix.codebuild }}
         run: src/ci/scripts/enable-docker-ipv6.sh
 
       # Disable automatic line ending conversion (again). On Windows, when we're

src/ci/citool/src/jobs.rs

@@ -3,12 +3,15 @@ mod tests;
 
 use std::collections::BTreeMap;
 
+use anyhow::Context as _;
 use serde_yaml::Value;
 
 use crate::GitHubContext;
+use crate::utils::load_env_var;
 
 /// Representation of a job loaded from the `src/ci/github-actions/jobs.yml` file.
 #[derive(serde::Deserialize, Debug, Clone)]
+#[serde(deny_unknown_fields)]
 pub struct Job {
     /// Name of the job, e.g. mingw-check
     pub name: String,
@@ -26,6 +29,8 @@ pub struct Job {
     pub free_disk: Option<bool>,
     /// Documentation link to a resource that could help people debug this CI job.
     pub doc_url: Option<String>,
+    /// Whether the job is executed on AWS CodeBuild.
+    pub codebuild: Option<bool>,
 }
 
 impl Job {
@@ -80,7 +85,7 @@ impl JobDatabase {
 }
 
 pub fn load_job_db(db: &str) -> anyhow::Result<JobDatabase> {
-    let mut db: Value = serde_yaml::from_str(&db)?;
+    let mut db: Value = serde_yaml::from_str(db)?;
 
     // We need to expand merge keys (<<), because serde_yaml can't deal with them
     // `apply_merge` only applies the merge once, so do it a few times to unwrap nested merges.
@@ -107,6 +112,29 @@ struct GithubActionsJob {
     free_disk: Option<bool>,
     #[serde(skip_serializing_if = "Option::is_none")]
     doc_url: Option<String>,
+    #[serde(skip_serializing_if = "Option::is_none")]
+    codebuild: Option<bool>,
+}
+
+/// Replace GitHub context variables with environment variables in job configs.
+/// Used for codebuild jobs like
+/// `codebuild-ubuntu-22-8c-$github.run_id-$github.run_attempt`
+fn substitute_github_vars(jobs: Vec<Job>) -> anyhow::Result<Vec<Job>> {
+    let run_id = load_env_var("GITHUB_RUN_ID")?;
+    let run_attempt = load_env_var("GITHUB_RUN_ATTEMPT")?;
+
+    let jobs = jobs
+        .into_iter()
+        .map(|mut job| {
+            job.os = job
+                .os
+                .replace("$github.run_id", &run_id)
+                .replace("$github.run_attempt", &run_attempt);
+            job
+        })
+        .collect();
+
+    Ok(jobs)
 }
 
 /// Skip CI jobs that are not supposed to be executed on the given `channel`.
@@ -177,6 +205,8 @@ fn calculate_jobs(
         }
         RunType::AutoJob => (db.auto_jobs.clone(), "auto", &db.envs.auto_env),
     };
+    let jobs = substitute_github_vars(jobs.clone())
+        .context("Failed to substitute GitHub context variables in jobs")?;
     let jobs = skip_jobs(jobs, channel);
     let jobs = jobs
         .into_iter()
@@ -207,6 +237,7 @@ fn calculate_jobs(
                 continue_on_error: job.continue_on_error,
                 free_disk: job.free_disk,
                 doc_url: job.doc_url,
+                codebuild: job.codebuild,
             }
         })
         .collect();

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

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

src/ci/docker/run.sh

@@ -288,7 +288,7 @@ args="$args --privileged"
 # `LOCAL_USER_ID` (recognized in `src/ci/run.sh`) to ensure that files are all
 # read/written as the same user as the bare-metal user.
 if [ -f /.dockerenv ]; then
-  docker create -v /checkout --name checkout alpine:3.4 /bin/true
+  docker create -v /checkout --name checkout ghcr.io/rust-lang/alpine:3.4 /bin/true
   docker cp . checkout:/checkout
   args="$args --volumes-from checkout"
 else

src/ci/github-actions/jobs.yml

@@ -56,6 +56,15 @@ runners:
   - &job-aarch64-linux-8c
     os: ubuntu-24.04-arm64-8core-32gb
     <<: *base-job
+
+  # Codebuild runners are provisioned in
+  # https://github.com/rust-lang/simpleinfra/blob/b7ddd5e6bec8a93ec30510cdddec02c5666fefe9/terragrunt/accounts/ci-prod/ci-runners/terragrunt.hcl#L2
+  - &job-linux-36c-codebuild
+    free_disk: true
+    codebuild: true
+    os: codebuild-ubuntu-22-36c-$github.run_id-$github.run_attempt
+    <<: *base-job
+
 envs:
   env-x86_64-apple-tests: &env-x86_64-apple-tests
     SCRIPT: ./x.py check compiletest --set build.compiletest-use-stage0-libtest=true && ./x.py --stage 2 test --skip tests/ui --skip tests/rustdoc -- --exact
@@ -153,7 +162,7 @@ auto:
     <<: *job-linux-4c
 
   - name: dist-arm-linux
-    <<: *job-linux-8c
+    <<: *job-linux-36c-codebuild
 
   - name: dist-armhf-linux
     <<: *job-linux-4c

src/ci/scripts/free-disk-space.sh

@@ -14,6 +14,17 @@ isX86() {
     fi
 }
 
+# Check if we're on a GitHub hosted runner.
+# In aws codebuild, the variable RUNNER_ENVIRONMENT is "self-hosted".
+isGitHubRunner() {
+    # `:-` means "use the value of RUNNER_ENVIRONMENT if it exists, otherwise use an empty string".
+    if [[ "${RUNNER_ENVIRONMENT:-}" == "github-hosted" ]]; then
+        return 0
+    else
+        return 1
+    fi
+}
+
 # print a line of the specified character
 printSeparationLine() {
     for ((i = 0; i < 80; i++)); do
@@ -32,7 +43,7 @@ getAvailableSpace() {
 # make Kb human readable (assume the input is Kb)
 # REF: https://unix.stackexchange.com/a/44087/60849
 formatByteCount() {
-    numfmt --to=iec-i --suffix=B --padding=7 "$1"'000'
+    numfmt --to=iec-i --suffix=B --padding=7 "${1}000"
 }
 
 # macro to output saved space
@@ -45,6 +56,11 @@ printSavedSpace() {
     after=$(getAvailableSpace)
     local saved=$((after - before))
 
+    if [ "$saved" -lt 0 ]; then
+        echo "::warning::Saved space is negative: $saved. Using '0' as saved space."
+        saved=0
+    fi
+
     echo ""
     printSeparationLine "*"
     if [ -n "${title}" ]; then
@@ -118,10 +134,14 @@ removeUnusedFilesAndDirs() {
         # Azure
         "/opt/az"
         "/usr/share/az_"*
+    )
 
+    if [ -n "${AGENT_TOOLSDIRECTORY:-}" ]; then
         # Environment variable set by GitHub Actions
-        "$AGENT_TOOLSDIRECTORY"
-    )
+        to_remove+=(
+            "${AGENT_TOOLSDIRECTORY}"
+        )
+    fi
 
     for element in "${to_remove[@]}"; do
         if [ ! -e "$element" ]; then
@@ -155,20 +175,25 @@ cleanPackages() {
         '^dotnet-.*'
         '^llvm-.*'
         '^mongodb-.*'
-        'azure-cli'
         'firefox'
         'libgl1-mesa-dri'
         'mono-devel'
         'php.*'
     )
 
-    if isX86; then
+    if isGitHubRunner; then
         packages+=(
-            'google-chrome-stable'
-            'google-cloud-cli'
-            'google-cloud-sdk'
-            'powershell'
+            azure-cli
         )
+
+        if isX86; then
+            packages+=(
+                'google-chrome-stable'
+                'google-cloud-cli'
+                'google-cloud-sdk'
+                'powershell'
+            )
+        fi
     fi
 
     sudo apt-get -qq remove -y --fix-missing "${packages[@]}"