[fab] kernel modules in fedora

Max Spevack mspevack at redhat.com
Thu Aug 31 17:32:04 UTC 2006


As Rex has said previously, this conversation needs to happen out here, 
not just internal to RHAT.  It's ours to answer.

We have our next Board meeting this coming Tuesday.

--Max

---------- Forwarded message ----------
Date: Thu, 31 Aug 2006 13:30:51 -0400 (EDT)
From: Max Spevack <mspevack at redhat.com>
To: memo-list at redhat.com
Subject: kernel modules in fedora (was: 10 open source start-ups to watch)

>> Kernel modules will always lag behind kernel updates, possibly 
>> preventing security updates from landing.  Bug reports will inevitably 
>> go to the kernel component and our (already overworked) kernel folks 
>> will have to suss out what is actually a kernel issue vs a module 
>> issue. The packaging of kernel modules does not mess well _at_ _all_ 
>> with RPM and our updating/build systems.
> 
> Summary: we've got work to do in our release process.
> 
> The kernel is not the only modular software out there. I don't see why 
> it should be special cased, especially when so much useful stuff can 
> only be achieved this way.

I agree.

The kernel is modular.  A lot of cool open source projects (like Asterisk) 
take advantage of that.

For Fedora to be as innovative as it can/should be, we need to be able to 
handle things like this.

One way is by getting things into the upstream kernel, but that's 
obviously not something we can just do by fiat.

Therefore, if we can't get all of the modules upstream, and we are leery 
of supporting them otherwise due to the resources that are required, then 
the solution is either to get more resources internally, or to focus some 
of our community resources on this problem.

Asterisk and CCRMA are two examples of innovative code that should be a 
part of Fedora.  It's up to us to figure out how we can make that happen 
from the engineering perspective, but we *need* to make it happen.

My $.02

--Max




More information about the fedora-advisory-board mailing list