Skip to content

Latest commit

 

History

History
275 lines (213 loc) · 14.6 KB

cli.rst

File metadata and controls

275 lines (213 loc) · 14.6 KB

Command Line Interface

A :class:`VUnit <vunit.ui.VUnit>` object can be created from command line arguments by using the :meth:`from_argv <vunit.ui.VUnit.from_argv>` method effectively creating a custom command line tool for running tests in the user project. Source files and libraries are added to the project by using methods on the VUnit object. The configuration is followed by a call to the :meth:`main <vunit.ui.VUnit.main>` method which will execute the function specified by the command line arguments and exit the script. The added source files are automatically scanned for test cases.

Usage

.. argparse::
   :ref: vunit.vunit_cli._parser_for_documentation
   :prog: run.py

Example Session

The :vunit_example:`VHDL User Guide Example <vhdl/user_guide/>` can be run to produce the following output:

> python run.py -l
lib.tb_example.all
lib.tb_example_many.test_pass
lib.tb_example_many.test_fail
Listed 3 tests
> python run.py -v lib.tb_example*
Running test: lib.tb_example.all
Running test: lib.tb_example_many.test_pass
Running test: lib.tb_example_many.test_fail
Running 3 tests

running lib.tb_example.all
Hello World!
pass( P=1 S=0 F=0 T=3) lib.tb_example.all (0.1 seconds)

running lib.tb_example.test_pass
This will pass
pass (P=2 S=0 F=0 T=3) lib.tb_example_many.test_pass (0.1 seconds)

running lib.tb_example.test_fail
Error: It fails
fail (P=2 S=0 F=1 T=3) lib.tb_example_many.test_fail (0.1 seconds)

==== Summary =========================================
pass lib.tb_example.all            (0.1 seconds)
pass lib.tb_example_many.test_pass (0.1 seconds)
fail lib.tb_example_many.test_fail (0.1 seconds)
======================================================
pass 2 of 3
fail 1 of 3
======================================================
Total time was 0.3 seconds
Elapsed time was 0.3 seconds
======================================================
Some failed!
> python run.py -v lib.tb_example.all
Running test: lib.tb_example.all
Running 1 tests

Starting lib.tb_example.all
Hello world!
pass (P=1 S=0 F=0 T=1) lib.tb_example.all (0.1 seconds)

==== Summary ==========================
pass lib.tb_example.all (0.9 seconds)
=======================================
pass 1 of 1
=======================================
Total time was 0.9 seconds
Elapsed time was 1.2 seconds
=======================================
All passed!

Opening a Test Case in Simulator GUI

Sometimes the textual error messages and logs are not enough to pinpoint the error and a test case needs to be opened in the GUI for visual debugging using single stepping, breakpoints and wave form viewing. VUnit makes it easy to open a test case in the GUI by having a -g/--gui command line flag:

> python run.py --gui my_test_case &

This launches a simulator GUI window with the top level for the selected test case loaded and ready to run. Depending on the simulator a help text is printed were a few TCL functions are pre-defined:

# vunit_help
#   - Prints this help
# vunit_load [vsim_extra_args]
#   - Load design with correct generics for the test
#   - Optional first argument are passed as extra flags to vsim
# vunit_user_init
#   - Re-runs the user defined init file
# vunit_run
#   - Run test, must do vunit_load first
# vunit_compile
#   - Recompiles the source files
# vunit_restart
#   - Recompiles the source files
#   - and re-runs the simulation if the compile was successful

The test bench has already been loaded with the vunit_load command. Breakpoints can now be set and signals added to the log or to the waveform viewer manually by the user. The test case is then run using the vunit_run command. Recompilation can be performed without closing the GUI by running vunit_compile. It is also possible to perform run.py with the --compile flag in a separate terminal.

Test Output Paths

VUnit creates a separate output directory for each test to provide isolation. The test output paths are located under OUTPUT_PATH/test_output/. The test names have been washed of any unsuitable characters and a hash has been added as a suffix to ensure uniqueness.

On Windows the paths can be shortened to avoid path length limitations. This behavior can be controlled by setting the relevant :ref:`environment variables <test_output_envs>`.

To get the exact test name to test output path mapping the file OUTPUT_PATH/test_output/test_name_to_path_mapping.txt can be used. Each line contains a test output path followed by a space seperator and then a test name.

Note

When using the run_all_in_same_sim pragma all tests within the test bench share the same output folder named after the test bench.

Simulator Selection

VUnit automatically detects which simulators are available on the PATH environment variable and by default selects the first one found. For people who have multiple simulators installed the VUNIT_SIMULATOR environment variable can be set to one of activehdl, rivierapro, ghdl or modelsim to explicitly specify which simulator to use.

In addition to VUnit scanning the PATH the simulator executable path can be explicitly configured by setting a VUNIT_<SIMULATOR_NAME>_PATH environment variable.

VUNIT_GHDL_PATH=/opt/ghdl/bin

Simulator Specific Environment Variables

  • VUNIT_MODELSIM_INI By default VUnit copies the modelsim.ini file from the tool install folder as a starting point. Setting this environment variable selects another modelsim.ini file as the starting point allowing the user to customize it.

Test Output Path Length Environment Variables

  • VUNIT_SHORT_TEST_OUTPUT_PATHS Unfortunately file system paths are still practically limited to 260 characters on Windows. VUnit tries to limit the length of the test output paths on Windows to avoid this limitation but still includes as much of the test name name as possible leaving a margin of 100 characters. VUnit however cannot forsee user specifc test output file lengths and this environment variable can be set to minimize output path lengths on Windows. On other operating systems this limitation is not relevant.
  • VUNIT_TEST_OUTPUT_PATH_MARGIN Can be used to change the test output path margin on Windows. By default the test output path is shortened to allow a 100 character margin.

Continuous Integration (CI) Environment

Because VUnit features the functionality needed to realize continuous and automated testing of HDL code, it is a very valuable resource in continuous integration environments. Once a project run.py has been setup, tests can be run in a headless environment with standardized Xunit style output to a file; which allows dynamic interpretation of results avoiding custom (and error-prone) parsing of the logs.

 python run.py --xunit-xml test_output.xml

After tests have finished running, the test_output.xml file can be parsed using standard xUnit test parsers such as Jenkins xUnit Plugin.

Furthermore, VUnit can be easily executed in many different platforms (either operating systems or architectures), because it is written in Python, which is an interpreted language. However, besides the sources and VUnit, a HDL compiler/simulator is required in order to run the tests. Due to performance, all the HDL simulators are written in compiled languages, which makes the releases platform specific. I.e., each simulator needs to be specifically compiled for a given architecture and operating system. This might represent a burden for the adoption of continuous integration in hardware development teams, as it falls into the category of dev ops.

Nevertheless, thanks to the striking research about portable development environment solutions in the last decade, there are a bunch of alternatives to ease the path. The 'classic' approach is to use virtual machines with tools such as VirtualBox, QEMU or VMware. This is still an option, but for most uses cases sharing complete system images is overkill. Here, containerization or operating-system-level virtualization comes into the game. Without going into technical details, containers are a kind of lightweight virtual machines, and the most known product that uses such a technology is Docker. Indeed, products such as Vagrant are meant to simplify the usage of virtual machines and/or containers by providing a common (black) box approach. In the end, there are enough open/non-open and free/non-free solutions for each user/company to choose the one that best fits their needs. From the hardware designer point-of-view, we 'just' need a box (no matter the exact underlying technology) that includes VUnit and a simulator.

Fortunately, contributors of project GHDL provide ready-to-use docker images at hub.docker.com/u/ghdl/dashboard. Some of these include not only GHDL but also VUnit:

  • ghdl/ext:vunit: Debian Stretch image with GHDL built from the latest commit of the master branch, and the latest release of VUnit installed through pip.
  • ghdl/ext:vunit-master: Debian Stretch with GHDL built from the latest commit of the master branch, and the latest commit of VUnit from the master branch.

As a result, the burden for the adoption of continuous integration for VUnit users is reduced to using docker; which is available in GNU/Linux, FreeBSD, Windows and macOS, and is supported in most cloud services (Travis CI, AWS, Codefresh, etc.) or CI frameworks (Jenkins, Drone, GitLab Runner, etc.).

For example, script :vunit_file:`examples/vhdl/docker_runall.sh <examples/vhdl/docker_runall.sh>` shows how to run all the VHDL examples in any x86 platform:

docker run --rm -t \
  -v /$(pwd)://work \
  -w //work \
  ghdl/ext:vunit-master sh -c ' \
    VUNIT_SIMULATOR=ghdl; \
    for f in $(find ./ -name 'run.py'); do python3 $f; done \
  '

where:

  • run: create and start a container.
  • --rm: automatically remove the container when it exits.
  • -t: allocate a pseudo-TTY, to get the stdout of the container forwarded.
  • -v: bind mount a volume, to share a folder between the host and the container. In this example the current path in the host is used ($(pwd)), and it is bind to /work inside the container. Note that both paths must be absolute.
  • -w: sets the working directory inside the container, i.e. where the commands we provide as arguments are executed.
  • ghdl/ext:vunit-master: the image we want to create a container from.
  • sh -c: the command that is executed as soon as the container is created.

Note that:

  • The arguments to sh -c are the same commands that you would execute locally, shall all the dependencies be installed in the host:

    VUNIT_SIMULATOR=ghdl
for f in $(find ./ -name 'run.py'); do python3 $f; done

  • The leading slashes in /$(pwd) and //work are only required for the paths to be properly handled in MINGW shells, and are ignored in other shells. See docker/for-win#1509.

Final comments:

  • All the (automated) flow to generate ghdl docker images is open source and public, in order to let any user learn and extend it. You can easily replicate it to build you own images with other development dependencies you use.
    • There are ready-to-use images available with additional tools on top of GHDL and VUnit. For example, ghdl/ext:vunit-gtkwave includes GTKWave.
  • Although the licenses of most commercial simulators do not allow to share ready-to-use docker images, it is straightforward to mimic the process.
  • Both GHDL and VUnit are free software. Docker is almost fully open source, but this depends on the host platform. See Is Docker still free and open source?.

Further info: