[RFC] Moving Intrepid to 2.6.27 kernel

[Daniel J Blueman]

On Sun, Aug 24, 2008 at 1:40 PM, Matt Zimmerman <mdz at ubuntu.com> wrote:
> On Wed, Aug 20, 2008 at 04:02:47PM -0400, Ben Collins wrote:
> The amount of testing it has received is trivial compared to what 2.6.26 has
> seen.  It's been tested on our certified systems and by many thousands of
> community members on their own hardware.  The fact that it's been smoke
> tested by a few people doesn't give me much confidence that it's free of
> serious regressions.

As you'll know, the level of upstream testing is pretty wide; I've
been running 2.6.27-rc kernels on Intrepid since -rc1 and have been
struggling to find any issues so far with my testsuites, hardware and
maximal debug enabled, to log the usual bug reports I do early in
kernel cycles; this was in contrast to 2.6.26 which had a rougher
landing (but is now mature).

I'd also say there are/will be a number of non-trivial but non-fatal
(eg performance/functionality) bugs which are fixed in 2.6.27, but
won't be backported to 2.6.26.x.

>> Yesterday, we discussed the possibility of moving to 2.6.27 before
>> feature freeze on #ubuntu-kernel during out weekly IRC meeting. In
>> attendance were the kernel team, Steve Langasek, Chuck Short and Soren
>> Hansen. Soren and Chuck gave particularly good reasons for moving to
>> 2.6.27, which included xen64 and updated KVM.

A good example is, when the GSPCA and UVC video cam drivers were
merged for 2.6.27, a number of (non-fatal) bugs were quickly fixed
which will unlikely be backported to 2.6.26 - this is a key usability
area for Ubuntu desktop, so contributes to the justification.

Sticking with 2.6.26, there are key ALSA fixes we'd need to backport
to 2.6.26 to get recording working for ALC883 on certain motherboards,
introducing extra work and risk. There are nVidia driver patches which
work for 2.6.27 too.

Daniel
-- 
Daniel J Blueman