Conversation
tjones60
added
the
release-ci-required
There was a problem hiding this comment.
Pull request overview
This PR increases parallelism for vmm_tests by reducing nextest thread reservations per test (reflecting that many guest VPs are idle), and standardizes processor-topology configuration by introducing reusable “heavy” presets.
Changes:
- Adjust
.config/nextest.tomlthreads-requiredoverrides to allow morevmm_teststo run concurrently. - Add
ProcessorTopology::heavy()/ProcessorTopology::very_heavy()helpers and update multiple tests to use them. - Clarify VP-count selection logic in
memstattests based on(isolation_type, machine_arch).
Reviewed changes
Copilot reviewed 6 out of 6 changed files in this pull request and generated 1 comment.
Show a summary per file
| File | Description |
|---|---|
vmm_tests/vmm_tests/tests/tests/x86_64/openhcl_uefi.rs |
Switch heavy/very-heavy tests to use the new ProcessorTopology presets. |
vmm_tests/vmm_tests/tests/tests/multiarch/vmbus_relay.rs |
Use ProcessorTopology::heavy() for the heavy vmbus relay test. |
vmm_tests/vmm_tests/tests/tests/multiarch/memstat.rs |
Make large-VM VP-count selection explicit by isolation type and architecture. |
vmm_tests/vmm_tests/tests/tests/multiarch.rs |
Use ProcessorTopology::heavy() for the heavy boot test. |
petri/src/vm/mod.rs |
Introduce reusable ProcessorTopology::heavy() and ::very_heavy() constructors. |
.config/nextest.toml |
Reduce per-test thread reservations for vmm_tests (including heavy/very-heavy filters) to increase CI concurrency. |
| 64 | ||
| } | ||
| } | ||
| TestVPCount::LargeVPCount => match (isolation_type, machine_arch) { |
There was a problem hiding this comment.
@ Daniel Paulius is this change ok? This will presumably affect the baselines here.
There was a problem hiding this comment.
This should be fine. Logically it should set up the same test configurations given the isolation type and architecture. I will check test results and update baselines as necessary though.
There was a problem hiding this comment.
Yeah the goal was for it to be the same logic, just a little easier to read.
Allow tests to run with less threads reserved, since many of the threads are often idle. This allows more tests to run in parallel, improving CI times.
Also standardizes and clarifies some processor configuration.