On Thu, Jul 9, 2026 at 10:32 AM Brian Cain <[email protected]>
wrote:

> On Thu, Jul 9, 2026 at 2:51 AM Stefan Hajnoczi <[email protected]> wrote:
> >
> > On Wed, Jul 1, 2026 at 10:21 PM Brian Cain <[email protected]>
> wrote:
> > > On 6/30/2026 11:01 AM, Stefan Hajnoczi wrote:
> > > > On Sun, Jun 28, 2026 at 9:46 PM Brian Cain <[email protected]>
> wrote:
> > > >> I'm working on upstreaming support for hexagon system emulation and
> it'd be useful to add the h2 vmm to QEMU's roms/ in order to simplify
> booting OSs like Linux.
> > > >>
> > > >> To that end, I think the initial step in the process is to mirror
> the repo in the qemu-project?
> https://github.com/qualcomm/hexagon-hypervisor/ is where the code is
> hosted now.  And once that is done I suppose I would add QEMU patches to
> add the build process and build output into roms/.  Can one of you do that
> mirror step or suggest how I can do it?
> > > > Hi Brian,
> > > > I have created a mirror of the stable branch here:
> > > > https://gitlab.com/qemu-project/hexagon-hypervisor
> > > >
> > > > Please take a look and let me know if you need any changes.
> > >
> > >
> > > Excellent, thanks!
> > >
> > >
> > > I have a feeling that the 'stable' branch of this repo was a 'good
> > > intention' that does not have a process by which it gets updated.  I've
> > > asked the maintainers if they plan to keep this branch active or prune
> > > it.  And I'll ask if they can generate a tag/release that we could use
> > > as a basis.
> >
> > Hi Brian,
> > Would you like me to change the mirroring configuration to follow
> > "master" instead of "stable"?
>
> Yes, please do.
>
> > > What's the process by which we update our fork?  It has an automated
> > > mirror process driven by gitlab?  What if we need changes that
> > > can't/won't be upstreamed - can we make those locally?  I don't yet
> have
> > > any plans to make local changes, just want to understand how the
> process
> > > works.
> >
> > GitLab automatically mirrors the upstream repo.
> >
> > It's possible to change the repo configuration to act as a fork where
> > you can push downstream-only patches, but at the moment it acts purely
> > as a mirror. There is only 1 repo I'm aware of where it was necessary
> > to do this. Usually it's possible to get patches upstream. Please let
> > me know in case you need downstream-only patches and we can set it up
> > (this should be a last resort).
>
> Agreed - I expect we'll be able to upstream the changes required.
>

Long term forks are an enormous pain. At most one should use the fork as a
staging
area in advance of upstreaming...

Warner

Reply via email to