Testing with Renode

Renode’s testing capabilities

Renode is very well suited to be a part of an automated tests scenario, e.g. run in the background on a CI server.

Renode is integrated with the Robot Framework testing suite and provides user-friendly scripts for running tests.

It comes with a variety of prepared test scripts, but it also allows you to extend them or add new ones.

Running the robot test script

Running a robot test script in Renode is as simple as executing a single command:

$ renode-test my_test.robot

The above command will:

  • start a Renode instance in the background,

  • enable Renode’s built-in Robot Framework server (providing an interface between Robot Framework and Renode) on port 9999 (the port number can be changed by the user),

  • start the Robot Framework test engine and connect to Renode,

  • run the provided my_test.robot test case,

  • print the progress status on the console,

  • generate the log and the summary after finishing the test.

Below, you can see an example output:

Preparing suites
Started Renode instance on port 9999; pid 2293056
Starting suites
Running tests/platforms/LiteX-VexRiscv.robot
+++++ Starting test 'LiteX-VexRiscv.Timer Test'
+++++ Finished test 'LiteX-VexRiscv.Timer Test' in 46.79 seconds with status OK
+++++ Starting test 'LiteX-VexRiscv.I2C Test'
+++++ Finished test 'LiteX-VexRiscv.I2C Test' in 8.10 seconds with status OK
Cleaning up suites
Closing Renode pid 2293056
Aggregating all robot results
Output:  /home/antmicro/renode/output/tests/robot_output.xml
Log:     /home/antmicro/renode/output/tests/log.html
Report:  /home/antmicro/renode/output/tests/report.html
Tests finished successfully :)

Note

Two entries in the output are associated with a single test case appear - one when the test starts and another when it finishes (useful when running tests in parallel). The second message contains information about the test’s duration and status.

The details of the run can be found in:

  • the robot_output.xml report (suitable for automatic parsing),

  • the log.html and report.html documents (suitable for an interactive inspection).

Creating the test file

Robot Framework uses text files based on a custom syntax to express the test cases. The details of the grammar can be found in the official documentation. In order for the robot files to work with Renode, proper configuration (explained below) is needed.

Here is an example of a simple robot test file that works with Renode:

*** Settings ***
Suite Setup     Setup
Suite Teardown  Teardown
Test Teardown   Test Teardown
Resource        ${RENODEKEYWORDS}

*** Test Cases ***
Should Print Help
    ${x}=  Execute Command     help
           Should Contain      ${x}    Available commands:

The Should Print Help test case executes the help command in Renode’s monitor and verifies the result.

Integration with Renode is achieved by adding entries to the settings section. The RENODEKEYWORDS variable (initialized by the renode-test script) contains the path to the renode-keywords.robot script responsible for setting up the connection with Renode. Other settings configure the suite/test setup and teardowns.

It is recommended to copy the above Settings section to each new robot test file.

Adding new test cases

Each robot test file might contain many test cases. For general instructions on how to define tests cases, please refer to the Robot Framework documentation. In this section we will focus on using the Robot Framework-Renode integration.

The Robot Framework-Renode integration layer provides keywords allowing the user to control and inspect the state of the simulation directly from the Robot Framework test file in a similar manner to the built-in keywords. The basic keywords allow the user to:

  • start the emulation (Start Emulation),

  • clear the emulation (Reset Emulation),

  • execute a command in the Monitor (Execute Command),

  • allocate a file in the Renode temporary folder (Allocate Temporary File),

  • download a file to the Renode temporary folder (Download File),

and more (see the source code for details).

Additionally, Renode provides a set of keywords for interacting with UART devices. They allow:

  • writing a text to UART (Send Key To Uart, Write Char On Uart, Write Line To Uart),

  • waiting for a specific line to appear on UART (Wait For Line On Uart, Wait For Prompt On Uart),

  • waiting for any output on UART (Wait For Next Line On Uart),

  • waiting for the lack of output on UART (Test If Uart Is Idle).

There is also a set of keywords for interacting with network devices that allow:

  • waiting for the next outgoing network packet (Wait For Outgoing Packet),

  • waiting for a specific outgoing network packet (Wait For Outgoing Packet With Bytes At Index).

It is possible to extend the Renode-Robot Framework interface by implementing more keywords in C# if necessary.

For reference on how to use the keywords mentioned in this section, see the robot test files that Renode comes with.

Advanced usage

Test-case dependencies

Normally, test cases in a single .robot file should be independent of each other. In fact, the default test teardown keyword calls Reset Emulation to ensure that the state of one test does not affect the next.

However, there are situations where it may be desirable to split a long-running test scenario into multiple test cases. This allows for better reporting of execution progress and improves overall performance (compared to re-running a common part of the test).

Renode provides custom Robot keywords to annotate situations where a test should continue execution from a state provided by another test:

  • Provides - creates a named snapshot of the simulation state (see State saving and loading),

  • Requires - loads a named snapshot and resumes execution of the test from that state.

An example of usage:

*** Test Cases ***
Boot Linux
    [...]
    Provides               booted-linux

Write to flash
    Requires               booted-linux
    Write Line To Uart     ...
    [...]

Ping another node
    Requires               booted-linux
    Write Line To Uart     ...
    [...]

See Zedboard.robot for a real-life example.

Note

There is also an alternative implementation of the mechanism available, where instead of state snapshots we use keywords recording. In this case when the Requires keyword is used, all the keywords executed before the corresponding Provides are re-executed. To use this mode, pass Reexecution as an additional parameter when using the Provides keyword:

*** Test Cases ***
Boot Linux
    [...]
    Provides               booted-linux        Reexecution

Running many test files with a single command

The example in the previous section showed how to run a single test file (which may still contain many test cases). It is possible to run many test files and aggregate the results into a single report. To do this, you need to pass many test files as arguments to the renode-test command:

renode-test my_tests.robot additional_tests.robot extra_tests.robot

The tests will be executed in the order in which the arguments were provided.

An alternative is to prepare a yaml file with the list of tests to run, e.g:

- my_tests.robot
- additional_tests.robot
- extra_tests.robot

and call renode-test with a special switch:

$ renode-test -t my_tests.yaml

Note

The .yaml notation allows the user to include other .yaml files and to group items that should not be executed in parallel (see the next section).

Running tests in parallel

Test cases from a single file are always executed serially (in the order defined in the file), but it’s possible to run tests from different files in parallel. To do this, run the renode-test command with a special switch:

$ renode-test -j12 my_tests.yaml

This will allow you to run up to 12 Renode instances, each running test cases from a different file. Using the .yaml file allows you to group items that should not be run in parallel (e.g. because they share a resource such as a port number):

- my_tests.robot
- my_group:
    - my_test2.robot
    - my_test3.robot

In the above example, my_test2.robot is executed before my_test3.robot, but in parallel with my_tests.robot.

You can also pass many test files as arguments (i.e. without the .yaml file), but you won’t be able to do the grouping:

$ renode-test -j3 my_tests.robot my_tests2.robot my_tests3.robot

Stopping on error

By default, renode-test will run all the provided test cases. However, it is possible to stop execution on the first error encountered. To do this, run the renode-test script with

renode-test --stop-on-error my_tests.robot

Running multiple instances of renode-test at once

Renode communicates with the Robot Framework executor via a network socket. This means that running two test-renode instances at the same time will result in a network port conflict.

To avoid this, you can explicitly specify the port number to be used for communication between the Robot Framework and Renode:

$ renode-test -P 9997 my_test.robot &
$ renode-test -P 9998 my_test2.robot &

Repeating tests

It is possible to run the specified tests multiple times using

$ renode-test -n 10 my_test.robot

This will re-run all test cases from my_tests.robot 10 times.

Running selected fixtures

It is possible to run only selected test cases from the file using:

$ renode-test -f "*GDB*" my_tests.robot

In the above example, only test cases with GDB in their name will be executed.

Running tests interactively

By default, the renode-test command runs tests in the background and only reports results to the console. However, it is possible to enable printing of log messages to the console in the same way as when running the renode command:

$ renode-test --show-log my_tests.robot

Note

This will cause the test progress messages to be mixed in with the log messages.

You can also show and interact with the monitor and analyzer windows:

$ renode-test --enable-xwt my_tests.robot

Note

Interacting with the running test may affect the results.

Saving state of failed tests

Renode’s testing framework allows the automatic creation of snapshots of failed tests in order to load them later to inspect the state of the simulation and/or run them further. This feature is especially helpful in non-interactive CI environments.

To enable the automatic creation of failed test snapshots, set the RENODE_CI_MODE environment variable before running the renode-test command:

$ RENODE_CI_MODE=YES renode-test my_test.robot

Each time the snapshot is created, it will be given a name corresponding to the failed test and you will see the message in the console informing you about the path to it. All snapshots are stored in the output/tests/snapshots directory.

Note

Enabling the CI mode will also influence the way external resources are handled - the binaries cache will be disabled, so each external file will be downloaded every time it’s referenced.

Inspecting failed tests interactively

With Renode, it is possible to stop the execution of the test suite in order to interactively debug a failed test case using the standard Renode interface (monitor, UART analyzers, etc).

To enable this feature, run the renode-test command with the following switch:

$ renode-test --debug-on-error my_test.robot

This will cause the execution of the test suite to stop on error, display the Renode Monitor and peripheral analyzers, and allow the user to inspect the state of the simulation. Once the interactive session is complete, it’s possible to resume test execution by pressing a button in a command prompt window.

Note

This feature is not currently available in headless environments.


Last update: 2024-04-08