CGL and the Linux Foundation

[Joe.MacDonald@xxxxxxxxxxxxx (MacDonald, Joe)]

On Wed, 2007-09-05 at 06:41 -0700, Dan Kohn wrote:
> On May 2, 2007, at 5:20 AM, Joe MacDonald wrote:
> > Do you have specific things the CGL working group did that you can see
> > as clearly separate from the goals of LSB?
> 
> My concern here is that the distros are generally trying to avoid  
> carrying patches of functionality not in the mainline kernel.  So CGL  
> requirements mainly depend on patch adoption by the appropriate  
> kernel subsystem maintainer.

Speaking as someone who works for a distro vendor, specifically as one
of the people who has to look at patches appearing on the various
mailing lists and deciding if they're "right for us" or not, I
completely agree with this statement.

Having said that, many new features that can be quite valuable to end
users (carriers or otherwise) are not accepted on their initial
submission.  We want to strive for a balance between specifying a set of
features already in mainline that constitute Carrier Grade and
identifying features not yet integrated into mainline that we think are
valuable enough to warrant the extra support burden of carrying those
patch sets.

> I think CGL could create both an optional LSB module of functionality  
> expected in any carrier-grade Linux distribution and also a gaps  
> document that lists a set of requirements not currently satisfied by  
> modern Linux distros.

I whole heartedly agree, though we tried to accomplish something like
that with the 4.0 specification.  P1s were requirements that were either
already in the mainline kernel or were available in mature, up-to-date
patch sets in the community (and the distinction between those was
available to anyone with access to the PoC database . . . which is
another discussion entirely).  P2s were requirements that had been
identified as important but not necessarily available on all platforms
or perhaps wasn't as mature as something implementing a P1.  P3s were
items that we had heard were important (several of these came as SCOPE
priority-1 items) but had no suitable implementation either in mainline
or in the community at large.  A good first draft of the gap analysis
you're talking about could come from the list of P3s in the CGL 4.0
specification.

-- 
Joe MacDonald <joe.macdonald@xxxxxxxxxxxxx>
Wind River Systems
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : https://lists.linux-foundation.org/mailman/private/lf_carrier/attachments/20070515/b41aba31/attachment.pgp