Re: [PULL 13/31] accel/nvmm: Replace @dirty field by generic CPUState::vcpu_dirty field

["Alex Bennée" <alex.bennee@linaro.org>]

Markus Armbruster <armbru@redhat.com> writes:

> Reinoud Zandijk <reinoud@NetBSD.org> writes:
>
>> On Wed, Jan 14, 2026 at 01:17:20PM +0100, Paolo Bonzini wrote:
>>> On 1/14/26 12:28, Reinoud Zandijk wrote:
>
> [...]
>
>>> We're not asking you to keep CI up-to-date (which Thomas is doing now,
>>> despite having no specific need that I know of to support NetBSD), just to
>>> *report* failure to build from source and tell us "hey, that's how we fixed
>>> it".  Otherwise we have the false impression that no one even cares about
>>> new QEMU on NetBSD.
>>
>> As pkgsrc normally tracks releases and has a directory with local patches on
>> top of that, build errors only show up when we bump the version. Version 10.2
>> is now in pkgsrc though I am running 10.1.3 so there is surely interest in
>> keeping it running. I could try to commit some more trivial patches in the
>> repo here to reduce the diffs some more.
>
> Fixing the build long after it broke is less than ideal.  What would it
> take to get NetBSD covered in CI?

We have the ability to run tests via the vm-build:

  make vm-build-netbsd

however to take full advantage that needs a runner with a reasonable
amount of memory and access to KVM so it can run reasonably fast. Does
anyone want to host a dedicated x86_64 custom runner?


-- 
Alex Bennée
Virtualisation Tech Lead @ Linaro