Update treadmill-ci.yml #1

Workflow file for this run

.github/workflows/treadmill-ci.yml at 510e4de

	# Licensed under the Apache License, Version 2.0 or the MIT License.
	# SPDX-License-Identifier: Apache-2.0 OR MIT
	# Copyright Tock Contributors 2024.

	# This workflow contains all Treadmill-based hardware CI jobs.
	#
	# Treadmill is a distributed hardware testbed developed within the Tock OS
	# project. For more information on Treadmill, have a look at its documentation
	# [1] or repository [2].
	#
	# This workflow is based on the Treadmill GitHub Actions integration guide [3].
	# In addition, it features the ability to run multiple Treadmill jobs and
	# test-execute stages through GitHub Action's job matrices, and uses a GitHub
	# environment to allow deployments with access to secrets for select PRs.
	#
	# [1]: https://book.treadmill.ci/
	# [2]: https://github.com/treadmill-tb/treadmill
	# [3]: https://book.treadmill.ci/user-guide/github-actions-integration.html

	name: treadmill-ci
	env:
	TERM: xterm # Makes tput work in actions output

	# Controls when the action will run. Triggers the workflow on pull request and
	# merge group checks:
	#
	# KEEP IN SYNC WITH `environment:` ATTRIBUTE BELOW:
	on:
	push:
	branches:
	- master
	- dev/tock-hardware-ci
	# Pull requests from forks will not have access to the required GitHub API
	# secrets below, even if they are using an appropriate deployment environment
	# and the workflow runs have been approved according to this environment's
	# rules. We don't know whether this is a bug on GitHub's end or deliberate.
	# Either way, for now we disable this workflow to run on PRs until we have
	# an API proxy that securely performs these GitHub API calls (adding runners
	# and starting Treadmill jobs with those runner registration tokens), which
	# allows this workflow to run without access to repository secrets.
	#pull_request:
	merge_group: # Run CI for the GitHub merge queue

	permissions:
	contents: read

	jobs:
	test-prepare:
	runs-on: ubuntu-latest

	# Do not run job on forks, as they will not have the correct environment set up
	if: github.repository == 'tock/tock-hardware-ci'

	# This provides access to the secrets required below:
	# - for `treadmill-ci`: after approval by certain persons or GH teams
	# - for `treadmill-ci-merged`: without approval, on merge queue branches
	# and the master branch
	#
	# KEEP IN SYNC WITH `on:` EVENTS ABOVE:
	environment: ${{ github.event_name == 'pull_request' && 'treadmill-ci' \|\| 'treadmill-ci-merged' }}

	outputs:
	tml-job-ids: ${{ steps.treadmill-job-launch.outputs.tml-job-ids }}
	tml-jobs: ${{ steps.treadmill-job-launch.outputs.tml-jobs }}

	steps:
	- uses: actions-rust-lang/setup-rust-toolchain@v1

	- name: Checkout Treadmill repository
	uses: actions/checkout@v4
	with:
	repository: treadmill-tb/treadmill
	# treadmill-tb/treadmill main as of Oct 1, 2024, 3:05 PM EDT
	ref: 'c82f4d7ebddd17f8275ba52139e64e04623f30cb'
	path: treadmill

	- name: Cache Treadmill CLI compilation artifacts
	id: cache-tml-cli
	uses: actions/cache@v4
	with:
	path: treadmill/target
	key: ${{ runner.os }}-tml-cli

	- name: Compile the Treadmill CLI binary
	run: \|
	pushd treadmill
	cargo build --package tml-cli
	popd
	echo "$PWD/treadmill/target/debug" >> "$GITHUB_PATH"

	# - uses: actions/checkout@v4
	# with:
	# path: tock

	# - name: Analyze changes and determine types of tests to run
	# run: \|
	# echo "TODO: implement this!"

	- name: Generate a token to register new just-in-time runners
	id: generate-token
	uses: actions/create-github-app-token@v1
	with:
	app-id: ${{ vars.TREADMILL_GH_APP_CLIENT_ID }}
	private-key: ${{ secrets.TREADMILL_GH_APP_PRIVATE_KEY }}

	- name: Create GitHub just-in-time runners and enqueue Treadmill jobs
	id: treadmill-job-launch
	env:
	GH_TOKEN: ${{ steps.generate-token.outputs.token }}
	TML_API_TOKEN: ${{ secrets.TREADMILL_API_TOKEN }}

	# Currently, all tests run only on hosts attached to an nRF52840DK
	DUT_BOARD: nrf52840dk

	# A Raspberry Pi OS netboot (NBD) image with a GitHub Actions
	# self-hosted runner pre-configured.
	#
	# For the available images see
	# https://book.treadmill.ci/treadmillci-deployment/images.html
	IMAGE_ID: 1b6900eff30f37b6d012240f63aa77a22e20934e7f6ebf38e25310552dc08378

	# Limit the supervisors to hosts that are compatible with this
	# image. This is a hack until we introduce "image sets" which define
	# multiple images for various supervisor hosts, but otherwise behave
	# identically:
	HOST_TYPE: nbd-netboot
	HOST_ARCH: arm64
	run: \|
	# When we eventually launch tests on multiple hardware platforms in
	# parallel, we need to supply different SUB_TEST_IDs here:
	SUB_TEST_ID="0"

	# This runner ID uniquely identifies the GitHub Actions runner we're
	# registering and allows us to launch test-execute jobs on this exact
	# runner (connected to the exact board we want to run tests on).
	RUNNER_ID="tml-gh-actions-runner-${GITHUB_REPOSITORY_ID}-${GITHUB_RUN_ID}-${GITHUB_RUN_ATTEMPT}-${SUB_TEST_ID}"

	# Obtain a new just-in-time runner registration token:
	RUNNER_CONFIG_JSON="$(gh api \
	-H "Accept: application/vnd.github+json" \
	-H "X-GitHub-Api-Version: 2022-11-28" \
	/repos/${{ github.repository }}/actions/runners/generate-jitconfig \
	-f "name=$RUNNER_ID" \
	-F "runner_group_id=1" \
	-f "labels[]=$RUNNER_ID" \
	-f "work_folder=_work")"
	echo "Generated configuration for runner $(echo "$RUNNER_CONFIG_JSON" \| jq -r '.runner.name')"

	# Generate a set of job paramters that includes the GitHub runner
	# registration token and a script that shuts down the host once the
	# runner has run through successfully (and created a file indicating
	# successful job completion, /run/github-actions-shutdown):
	TML_JOB_PARAMETERS="{\
	\"gh-actions-runner-encoded-jit-config\": {\
	\"secret\": true, \
	\"value\": \"$(echo "$RUNNER_CONFIG_JSON" \| jq -r '.encoded_jit_config')\" \
	}, \
	\"gh-actions-runner-exec-stop-post-sh\": {\
	\"secret\": false, \
	\"value\": \"if [ \\\"\$SERVICE_RESULT\\\" = \\\"success\\\" ] && [ -f /run/github-actions-shutdown ]; then tml-puppet job terminate; fi\" \
	}\
	}"

	echo "Enqueueing treadmill job:"
	TML_JOB_ID_JSON="$(tml job enqueue \
	"$IMAGE_ID" \
	--tag-config "board:$DUT_BOARD;host-type:$HOST_TYPE;host-arch:$HOST_ARCH" \
	--parameters "$TML_JOB_PARAMETERS" \
	)"

	TML_JOB_ID="$(echo "$TML_JOB_ID_JSON" \| jq -r .job_id)"
	echo "Enqueued Treadmill job with ID $TML_JOB_ID"

	# Pass the job IDs and other configuration data into the outputs of
	# this step, such that we can run test-execute job instances for each
	# Treadmill job we've started:
	echo "tml-job-ids=[ \
	\"$TML_JOB_ID\" \
	]" >> "$GITHUB_OUTPUT"

	echo "tml-jobs={ \
	\"$TML_JOB_ID\": { \
	\"runner-id\": \"$RUNNER_ID\", \
	} \
	}" >> "$GITHUB_OUTPUT"

	test-execute:
	needs: test-prepare

	strategy:
	matrix:
	tml-job-id: ${{ fromJSON(needs.test-prepare.outputs.tml-job-ids) }}

	runs-on: ${{ fromJSON(needs.test-prepare.outputs.tml-jobs)[matrix.tml-job-id].runner-id }}

	steps:
	- name: Print Treadmill Job Context and Debug Information
	run: \|
	echo "Treadmill job id: ${{ matrix.tml-job-id }}"
	echo "GitHub Actions Runner ID: ${{ fromJSON(needs.test-prepare.outputs.tml-jobs)[matrix.tml-job-id] }}"
	echo "===== Parameters: ====="
	ls /run/tml/parameters
	echo "===== User & group configuration: ====="
	echo "whoami: $(whoami)"
	echo "groups: $(groups)"
	echo "===== Network configration: ====="
	ip address
	echo "===== Attached USB & serial console devices: ====="
	lsusb
	ls -lh /dev/ttyAMA* 2>/dev/null \|\| true
	ls -lh /dev/ttyACM* 2>/dev/null \|\| true
	ls -lh /dev/ttyUSB* 2>/dev/null \|\| true
	ls -lh /dev/bus/usb// 2>/dev/null \|\| true

	- name: Disable wget progress output
	run: \|
	echo "verbose = off" >> $HOME/.wgetrc

	- uses: actions/checkout@v4

	- uses: actions-rust-lang/setup-rust-toolchain@v1
	with:
	# Avoid overwriting the RUSTFLAGS environment variable
	rustflags: ""

	- name: Install required system packages
	run: \|
	# TODO: currently, the Netboot NBD targets have no access to their
	# boot parition (e.g., mounted on /boot/firmware) on a Raspberry Pi OS
	# host. This causes certain hooks in response to dpkg / apt commands
	# to fail. Thus we ignore errors in these steps until we figure this
	# part out.
	sudo DEBIAN_FRONTEND=noninteractive apt update \|\| true
	sudo DEBIAN_FRONTEND=noninteractive apt install -y \
	git cargo openocd python3 python3-pip python3-serial \
	python3-pexpect gcc-arm-none-eabi libnewlib-arm-none-eabi \
	pkg-config libudev-dev cmake libusb-1.0-0-dev udev make \
	gdb-multiarch gcc-arm-none-eabi build-essential \|\| true

	# Install probe-rs:
	curl --proto '=https' --tlsv1.2 -LsSf \
	https://github.com/probe-rs/probe-rs/releases/latest/download/probe-rs-tools-installer.sh \
	\| sh

	- name: Create Python virtual environment and install required dependencies
	run: \|
	python3 -m venv ./hwcienv
	source ./hwcienv/bin/activate
	pip install -r tools/hwci/requirements.txt -c tools/hwci/requirements-frozen.txt

	- name: Run tests
	run: \|
	source ./hwcienv/bin/activate
	cd ./tools/hwci
	export PYTHONPATH="$PWD:$PYTHONPATH"
	python3 core/main.py --board boards/nrf52dk.py --test tests/c_hello.py

	- name: Request shutdown after successful job completion
	run: \|
	sudo touch /run/github-actions-shutdown

	- name: Provide connection information on job failure
	if: failure()
	run: \|
	echo "This CI job has failed, we avoid terminating the Treadmill job"
	echo "immediately. It will be active until it reaches its timeout."
	echo ""
	echo "If you added SSH keys to the `job enqueue` command, you can"
	echo "open an interactive session to this host. Connection"
	echo "information is available here:"
	echo "https://book.treadmill.ci/treadmillci-deployment/sites.html"
	echo ""
	echo "TODO: print host / supervisor ID as part of workflow"
	echo "TODO: determine public SSH endpoint automatically and print"
	echo "TODO: allow adding SSH keys to running Treadmill jobs"

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Update treadmill-ci.yml #1

Workflow file

Update treadmill-ci.yml #1

Jobs

Run details

Workflow file for this run