-
-
Notifications
You must be signed in to change notification settings - Fork 62
Qubes update cannot handle hvms with attached pci devices #10617
Copy link
Copy link
Closed
Copy link
Labels
C: updatesThis issue pertains to updating Qubes OS.This issue pertains to updating Qubes OS.P: defaultPriority: default. Default priority for new issues, to be replaced given sufficient information.Priority: default. Default priority for new issues, to be replaced given sufficient information.affects-4.2This issue affects Qubes OS 4.2.This issue affects Qubes OS 4.2.affects-4.3This issue affects Qubes OS 4.3.This issue affects Qubes OS 4.3.devel-host-cur-testdiagnosedTechnical diagnosis of this issue has been performed.Technical diagnosis of this issue has been performed.good first issueThis is a recommended issue for first-time contributors.This is a recommended issue for first-time contributors.pr submittedA pull request has been submitted for this issue.A pull request has been submitted for this issue.
Description
Activity
Metadata
Metadata
Assignees
Labels
C: updatesThis issue pertains to updating Qubes OS.This issue pertains to updating Qubes OS.P: defaultPriority: default. Default priority for new issues, to be replaced given sufficient information.Priority: default. Default priority for new issues, to be replaced given sufficient information.affects-4.2This issue affects Qubes OS 4.2.This issue affects Qubes OS 4.2.affects-4.3This issue affects Qubes OS 4.3.This issue affects Qubes OS 4.3.devel-host-cur-testdiagnosedTechnical diagnosis of this issue has been performed.Technical diagnosis of this issue has been performed.good first issueThis is a recommended issue for first-time contributors.This is a recommended issue for first-time contributors.pr submittedA pull request has been submitted for this issue.A pull request has been submitted for this issue.
How to file a helpful issue
Qubes OS release
4.2.4, 4.3
Brief summary
Qubes update GUI (with concurrency set to 1) and
qubes-vm-update --max-concurrency=1are unable to handle multiple hvms that operate at least one same pcie device. Next qube in the queue is started without waiting for previous one to shut down completely, which causes pcie device conflicts.Steps to reproduce
Expected behavior
If currently updated vm has at least one same pcie device attached to it as the next vm in queue, wait for complete shutdown before starting the next vm.
Actual behavior
Updating mechanism attempts to boot the second hvm immediately after initiating shutdown of the first hvm, which results in a pcie device conflict.
Additional information
It would be nice to see this fixed for higher concurrency settings as well :)