Re: Domctl series as a fix to PV_SHIM_EXCLUSIVE, was: [PATCH] xen/x86: move domctl.o out of PV_SHIM_EXCLUSIVE

[Oleksii Kurochko <oleksii.kurochko@xxxxxxxxx>]

Hello Stefano,

On 9/3/25 12:16 AM, Stefano Stabellini wrote:

> Hi Oleksii,
>
> We previously discussed the PV_SHIM_EXCLUSIVE build issue on Matrix
> and agreed on resolving it after the feature freeze as a fix. This
> conversation took place before the feature freeze was rescheduled to
> September 5. I am documenting this on xen-devel both for record-keeping
> and because Matrix is currently offline. I am proceeding under the
> assumption that the conclusions from that discussion still apply.
>     
Thank you for documenting this on xen-devel.
> I would like to request an additional clarification. While there is no
> consensus among the maintainers on the preferred short-term compromise,
> there is agreement that reviewing and committing the domctl series [1]
> in time would be the best outcome. (Together with unifying CONFIG_SYSCTL
> and CONFIG_DOMCTL into a single CONFIG_MGMT_HYPERCALLS for simplicify.)
>
> Are you still OK with us going ahead with reviewing and committing the
> domctl series as a fix to the PV_SHIM_EXCLUSIVE issue, potentially going
> past the feature freeze by a week?
>     
Regarding your request for clarification on the domctl series [1], I agree
that reviewing and committing this series as a fix for the PV_SHIM_EXCLUSIV
issue is the preferred short-term solution.
Given that the series primarily consists of refactoring changes with minimal
functional impact, I am comfortable with making an exception to the feature
freeze to allow its inclusion, potentially extending up to one week past
September 5 (only for this patch series).

But if it will/could take more time then I think that we should consider
an option just to revert some patches, IIUC it should fix the mentioned issue
too.

~ Oleksii
> Stefano
>
> [1] https://marc.info/?l=xen-devel&m=175421442423500
>