Lspci: only refresh the pciids database once per node

[mcgov (mcgov)]

Part 3 of 9 of a stacked series that reworks the DPDK SRIOV hot plug tests. Stacked on #4655, review only the last commit.

get_devices ran update-pciids on every forced refresh, which adds a network download and a noticeable delay to hot plug tests that rescan the pci bus repeatedly. The database is now refreshed once per tool instance.

Key Test Cases:
verify_dpdk_sriov_rescind_failover_send_only|verify_device_reports_correct_pci_slot

Impacted LISA Features:
Sriov, NetworkInterface, Nvme, Gpu

Tested Azure Marketplace Images:

• canonical 0001-com-ubuntu-server-jammy 22_04-lts latest
• microsoftcblmariner azure-linux-3 azure-linux-3 latest

[Copilot]

Pull request overview

This PR optimizes Lspci.get_devices() by avoiding repeated update-pciids network downloads during forced PCI refreshes, which is especially beneficial for hot-plug test loops that rescan PCI state frequently.

Changes:

• Add a per-tool-instance _pciids_initialized flag to ensure update-pciids runs at most once.
• Move the update-pciids call behind the initialization guard while keeping device list refresh behavior unchanged.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated no new comments.

[Copilot]

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated no new comments.

Suppressed comments (1)

lisa/tools/lspci.py:303

• _pciids_initialized is only set to True when update-pciids returns 0. If update-pciids is slow and fails (e.g., transient network/DNS issues), then each subsequent force_run will re-run update-pciids, reintroducing the repeated delay this PR is trying to remove. Consider marking the refresh as "attempted" regardless of exit code (best-effort update) and log stderr for troubleshooting.

            if not self._pciids_initialized:
                # Ensure pci device ids and name mappings are updated.
                exit_code = self.node.execute("update-pciids", sudo=True, shell=True).exit_code
                if exit_code == 0:
                    self._pciids_initialized = True
                else:
                    self.node.log.debug(f"update-pciids failed, returned code {exit_code}")

[Copilot]

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated no new comments.

Suppressed comments (2)

lisa/tools/lspci.py:307

• _pciids_initialized is only set to True when update-pciids succeeds. If update-pciids is slow and fails (e.g., transient network issues), every force_run=True refresh will re-run it, reintroducing the repeated download/delay this PR is trying to remove. Consider marking it initialized after the first attempt (success or failure) and just logging failures.

            if not self._pciids_initialized:
                # Ensure pci device ids and name mappings are updated.
                exit_code = self.node.execute(
                    "update-pciids", sudo=True, shell=True
                ).exit_code
                if exit_code == 0:
                    self._pciids_initialized = True
                else:
                    self.node.log.debug(
                        f"update-pciids failed, returned code {exit_code}"
                    )

lisa/tools/lspci.py:296

• The PR description lists verify_dpdk_sriov_rescind_failover_send_only and verify_device_reports_correct_pci_slot as executed test cases, but those method names don’t exist in this branch (no matches in lisa/microsoft/testsuites/). Please update the PR description to reference the actual test methods that were run so reviewers can reproduce validation.

        if (not self._pci_devices) or force_run:

[Copilot]

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated no new comments.

Suppressed comments (1)

lisa/tools/lspci.py:307

• _pciids_initialized is only set to True when update-pciids succeeds. If the command fails once (e.g., no outbound network access), every subsequent force_run will retry update-pciids, reintroducing the repeated download/delay this PR is trying to eliminate. Consider marking the pciids refresh as "attempted" after the first run regardless of exit code, while still logging failures.

                if exit_code == 0:
                    self._pciids_initialized = True
                else:
                    self.node.log.debug(
                        f"update-pciids failed, returned code {exit_code}"