Skip to content

vmm_tests: mark linux direct pcie tests as unstable#3548

Merged
smalis-msft merged 1 commit into
microsoft:mainfrom
tjones60:pcie_unstable
May 21, 2026
Merged

vmm_tests: mark linux direct pcie tests as unstable#3548
smalis-msft merged 1 commit into
microsoft:mainfrom
tjones60:pcie_unstable

Conversation

@tjones60
Copy link
Copy Markdown
Contributor

Our PCIe tests have been failing with Linux direct, so mark them as unstable for now. This issue will hopefully be fixed once the new test kernel is onboarded: #3530

@tjones60 tjones60 requested a review from a team as a code owner May 21, 2026 20:25
Copilot AI review requested due to automatic review settings May 21, 2026 20:25
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

Marks the Linux-direct x64 PCIe integration tests as unstable so failures no longer block PRs, while keeping the same test coverage for UEFI-based guests.

Changes:

  • Switched all #[openvmm_test(linux_direct_x64, ...)] usages in the PCIe test module to unstable_linux_direct_x64.
  • Added a file-level comment explaining why Linux-direct PCIe tests are currently unstable and that this should be reverted after the kernel update.

Comment on lines +4 to +6
// All of the PCIe tests on Linux direct have been marked as unstable due to
// what is presumed to be a bug in the guest kernel currently being used.
// TODO: remove the unstable designations once the kernel is updated.
@github-actions
Copy link
Copy Markdown

@smalis-msft smalis-msft merged commit 108baea into microsoft:main May 21, 2026
67 checks passed
@jackschefer-msft
Copy link
Copy Markdown
Contributor

A guest MANA/GDMA issue would only affect the pcie_devices test (that's the only one that adds MANA emulators), are we seeing failures elsewhere?

@smalis-msft
Copy link
Copy Markdown
Contributor

The test failures we're seeing are kernel panics, but they seem to come only from these tests. I think once we update the test kernel we'll reenable these and reevaluate. I've had copilot run on a bunch of logs of these failures and it keeps blaming the kernel, and 6.1 is reallllly old.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants