[FW: [Fedora-xen] Plans for paravirt_ops kernel-xen]

[Daniel P. Berrange]

Forwarding the message below from Mark McLoughlin for wider exposure...

To make it crystal clear - we are almost certainly *NOT* going to have
Xen Dom0 host support in Fedora 9 GA. People requiring a Xen Dom0 host
should expect to remain on Fedora 8. We *WILL* be providing Xen DomU
guests running Fedora 9, with a kernel that is at last synced to bare
metal versions in a supportable manner.

If Dom0 support in F10 rawhide is sufficiently stable, we may add Dom0
to F9 post-GA, but don't count on it.

The ancient "upstream" Xen community kernels are never coming back.
Celebrate :-)

Dan.

----- Forwarded message from Mark McLoughlin <markmc at redhat.com> -----

> From: Mark McLoughlin <markmc at redhat.com>
> To: fedora-xen at redhat.com
> Date: Thu, 06 Mar 2008 15:29:17 +0000
> Subject: [Fedora-xen] Plans for paravirt_ops kernel-xen
> 
> Hi,
>         Fedora's Xen hackers have been working hard towards switching
> our kernel-xen package from a forward-ported Xensource kernel tree to a
> state-of-the-art upstream, paravirt_ops based, kernel in Fedora 9 as
> described here:
> 
>   http://fedoraproject.org/wiki/Features/XenPvops
>   http://www.redhat.com/archives/fedora-xen/2007-November/msg00106.html
> 
>         Some great progress has been made, and tomorrow's rawhide will
> have a kernel-xen update with:
> 
>   + A very recent 2.6.25-rc4 base
> 
>   + Xen paravirt_ops DomU from upstream
> 
>   + x86_64 Xen paravirt_ops DomU support
> 
>   + Paravirt framebuffer
> 
>         However, although the Dom0 paravirt_ops work is well advanced at
> this point, we still don't have backend drivers or x86_64 Dom0 working.
> 
>         With the feature freeze looming next week, we have make the
> difficult decision to focus the Fedora 9 efforts on DomU and postpone
> the inclusion of paravirt_ops Dom0 support.
> 
>         The alternative course of action was to keep shipping the
> 2.6.21.7 based kernel-xen in Fedora 9, but we have ruled this out as a
> supportable option. This kernel is almost a year old now and we cannot
> expect Fedora hackers to keep the distribution working on such an old
> kernel. Examples of the kind of issues we see cropping up are:
> 
>   1) Broken installs due to old squashfs:
> 
>        https://bugzilla.redhat.com/431109
> 
>   2) Broken SELinux due to old SELinux:
> 
>        https://bugzilla.redhat.com/436173
> 
>   3) Broken networking due to old netlink:
> 
>        https://bugzilla.redhat.com/431179
> 
>         We feel that making significant investment across the
> distribution to keep this old kernel working for the sake of Dom0
> support would be wasting effort on a dead codebase.
> 
>         Work will continue apace on the Dom0 paravirt_ops effort for
> Fedora 10 and we hope to introduce the first build to rawhide soon after
> Fedora 9 been branched. This first build should include backend drivers
> and x86_64 support. If all goes well with the Dom0 support in Fedora 10
> rawhide, we may well pull it into Fedora 9 as a post-GA update.
> 
>         So, in summary:
> 
>   1) Try out the F9 rawhide/beta paravirt_ops kernel-xen in your DomUs
> 
>   2) Keep your Dom0 on Fedora 8 for now
> 
>   3) If you want to help out with Dom0 paravirt_ops testing, then be
>      ready to jump onto Fedora 10 rawhide
> 
> Thanks,
> Mark.
> 
> --
> Fedora-xen mailing list
> Fedora-xen at redhat.com
> https://www.redhat.com/mailman/listinfo/fedora-xen
> 
----- End forwarded message -----

-- 
|=- Red Hat, Engineering, Emerging Technologies, Boston.  +1 978 392 2496 -=|
|=-           Perl modules: http://search.cpan.org/~danberr/              -=|
|=-               Projects: http://freshmeat.net/~danielpb/               -=|
|=-  GnuPG: 7D3B9505   F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505  -=|