On 14.09.21 15:47, Philippe Mathieu-Daudé wrote:
On 9/14/21 3:17 PM, Dr. David Alan Gilbert (git) wrote:
From: "Dr. David Alan Gilbert" <[email protected]>
The subsection name for page-poison was typo'd as:
vitio-balloon-device/page-poison
Note the missing 'r' in virtio.
When we have a machine type that enables page poison, and the guest
enables it (which needs a new kernel), things fail rather unpredictably.
IIUC since v5.1 guests have 'page-poison'=true but once migrated
they become 'page-poison'=unset=false?
We only migrate the subsection if the guest supports the feature (->
newer guest kernel).
If we migrate, it's a broken migration stream. If we don't migrate,
everything is fine.
--
Thanks,
David / dhildenb
