On Fri, 2 Jul 2010, Andy Whitcroft wrote:
> With all things for releases which are already shipped, it is a trade
> off between the size/invasivness off the patches against the possible
> benefits. A quick look at the upstream patches you point to seems to
> say that its a pretty big patch set? It would depend to some extent how
> much of that is additions and how much modifications.
Yes it is, majority of which is additions. It also only touches the
sunrpc code. I haven't checked if .35 has changes that fix issues with the
Oh and DES still works, of course.
> You note that this stuff is going to be in 2.6.35, which would mean that
> the lts-backports-maverick kernel would presumably work for you? Have
> you considered that as an option?
I have, but I wonder what issues there might be running a server kernel on
desktops (or is there going to be a desktop flavor too)? I guess some of
the choices can be selected on runtime though. Another issue is that we'd
need to wait for the release first, or run prerelease code. We are in
production next month.
A DKMS package of the sunrpc modules might be another option, but I'd be
surprised that if there weren't other people interested in seeing this
code in the official kernel
DKMS is something that I'd like to test
anyway, to see how easy it would be to backport fixes to the NFSv4
client.. or is it even possible.
Systems Specialist, Aalto IT
kernel-team mailing list