DKMS into Fedora Extras

[Fernando Lopez-Lezcano]

On Mon, 2005-02-21 at 11:05, seth vidal wrote:
> On Mon, 2005-02-21 at 19:57 +0100, Thorsten Leemhuis wrote:
> > Am Montag, den 21.02.2005, 20:47 +0200 schrieb Ville Skyttä:
> > > Let's say we have kernel-V-R installed, and kernel-module-foo built
> > > for
> > > that installed.  Now, we get an update for the module package only;
> > > the
> > > kernel does not change.  Obviously, the already installed module
> > > package
> > > and the new one conflict.  What happens when one does "yum update"?
> > 
> > It fails:
> > 
> > https://devel.linux.duke.edu/bugzilla/show_bug.cgi?id=394
> 
> I'm not sure what the other modes should be.
> 
> we might be able to do something like:
> if two kernel modules have the same name but different versions then
> it's an update.
> 
> that would require:
>  - kernel-version-in-module-package-name
>  - provides: kernel-module in the header
>  - consistent use.
>
> if that's do-able I should be able to write that code w/o it being too
> excruciating.

Has anyone had problems with architecture matching between external
kernel module packages and the main kernel package?

I did (in Planet CCRMA), albeit a long time ago, and maybe as a result
of old versions of apt not doing the right thing. 

But it was possible to install a kernel package of, let's say,
architecture i586, on top of a i686 kernel. Obviously that does not
work. There are no dependencies at the rpm level to keep that from
happening. To prevent that from happening again I have both a provides
(in the kernel package) and matching requires (in the kernel-module-*
packages) so that it is impossible, from the point of view of rpm, to
mismatch architectures.

Could someone please add this, if deemed appropriate, to the list of
requirements when packaging kernel modules?

-- Fernando