On TrueNAS SCALE, when performing GPU passthrough with high-memory cards QEMU options are required
Description
Problem/Justification
None
Impact
None
Activity
Show:
Bug Clerk September 27, 2022 at 4:42 PM
Michelle Johnson September 7, 2022 at 12:27 PM
Thank you for your report, !
This issue is now in the queue for review. An Engineering representative will update with further details or questions in the near future.
Complete
Pinned fields
Click on the next to a field label to start pinning.
Details
Details
Assignee
Muhammad RehanReporter
Matthew JohnsonLabels
Impact
Medium
Time remaining
0m
Components
Fix versions
Affects versions
Priority
LowKatalon Platform
Linked Test Cases, Katalon Defect Results, Katalon Studio Test Results
Katalon Platform
Linked Test Cases, Katalon Defect Results, Katalon Studio Test Results
Created September 2, 2022 at 5:09 PM
Updated October 6, 2022 at 3:14 PM
Resolved October 6, 2022 at 12:03 PM
When attempting to perform GPU Passthrough to a UEFI VM with a high-memory card such as a Tesla M40 or RTX 3090 on a system that supports resizable BAR additional QEMU options are required. Without these options the VM will not start.
The domain XML needs to be edited as such
domain block needs the schema added as below
<domain type='kvm' xmlns:qemu='http://libvirt.org/schemas/domain/qemu/1.0'>then the following configuration items are added at the end before the final </domain>
<qemu:commandline> <qemu:arg value='-fw_cfg'/> <qemu:arg value='opt/ovmf/X-PciMmio64Mb,string=65536'/> </qemu:commandline>Alternatively, the following can be passed on the QEMU command line -fw_cfg name=opt/ovmf/X-PciMmio64Mb,string=65536