hqjenny-chipyard/.circleci

Chipyard CI

Website: https://circleci.com/gh/ucb-bar/project-template

CircleCI Brief Explanation

CircleCI is controlled by the config.yml script. It consists of a workflow which has a series of jobs within it that do particular tasks. All jobs in the workflow must pass for the CI run to be successful.

At the bottom of the config.yml there is a workflows: section that specifies the order in which the jobs of the workflow should run. For example:

- prepare-rocketchip:
    requires:
        - install-riscv-toolchain
        - install-verilator

This specifies that the prepare-rocketchip job needs the install-riscv-toolchain and install-verilator steps to run before it can run.

All jobs in the CI workflow are specified at the top of config.yml They specify a docker image to use (in this case a riscv-boom image since that is already available and works nicely) and an environment. Finally, in the steps: section, the steps are run sequentially and state persists throughout a job. So when you run something like checkout the next step has the checked out code. Caching in the job is done by giving a file to cache on. restore_cache: loads the cache into the environment if the key matches while save_cache: writes to the cache with the key IF IT IS NOT PRESENT. Note, if the cache is already present for that key, the write to it is ignored. Here the key is built from a string where the checksum portion converts the file given into a hash.

.circleci directory

This directory contains all the collateral for the Chipyard CI to work. The following is included:

build-toolchains.sh # build either riscv-tools or esp-tools
build-verilator.sh  # build verilator
create-hash.sh      # create hashes of riscv-tools/esp-tools so circleci caching can work
do-rtl-build.sh     # use verilator to build a sim executable
config.yml          # main circleci config script to enumerate jobs/workflows

How things are setup for Chipyard

The steps for CI to run are as follows. 1st, build the toolchains in parallel (note: esp-tools is currently not used in the run). The docker image sets up the PATH and RISCV variable so that riscv-tools is the default (currently the env.sh script that is created at tool build is unused). 2nd, install verilator using the *.mk to cache unique versions of verilator (mainly for if verilator is bumped). 3rd, create the simulator binary. This requires the riscv-tools for fesvr and verilator to be able to build the binary. This stores all collateral for the tests (srcs, generated-srcs, sim binary, etc) to run "out of the gate" in the next job (make needs everything or else it will run again). 4th, finally run the tests that were wanted.