Plan for tomorrows (20070726) FESCO meeting

David Woodhouse dwmw2 at infradead.org
Thu Jul 26 16:43:33 UTC 2007


On Thu, 2007-07-26 at 11:59 -0400, Jesse Keating wrote:
> On Thu, 26 Jul 2007 17:48:57 +0200
> Hans de Goede <j.w.r.degoede at hhs.nl> wrote:
> 
> > I must say I like this approach, it avoids the whole problem of
> > having to rebuild kmods all the time and of wether to delay kernel
> > security updates until all kmods are fixetd etc. I do think however
> > that this might cause some pain for Dave Jones, whose job already is
> > hard. Maybe we should ask him what he thinks about this?
> 
> Well, if the module doesn't build fine with an update, most likely
> what's going to happen is the module gets disabled, which could result
> in even more frequent kernel updates just for getting a single module
> to build again.

I assume you're thinking of updates, not rawhide. In rawhide, Dave or
Chuck send you a mail saying "your $FOO patch broke; I turned it off".
You fix it, and it appears again in rawhide in a day or two.

In updates, such things are actually much less likely to break, except
when we rebase to a newer kernel -- and if you've kept your driver
working in rawhide then it should work fine when we rebase the release
to the newer kernel anyway. Even if not, we tend to be quite
conservative about releasing new kernels anyway -- they end up in
updates-testing for some time, and that gives you time to get it
working. The urgent security fixes are usually relatively small and
unlikely to break drivers.

It's not something that scales hugely, but we don't _want_ it to. We
should be sticking close to upstream, and not shipping drivers which
aren't getting merged.

-- 
dwmw2




More information about the fedora-devel-list mailing list