Re: [gentoo-server] kernel oops

[Kerin Millar]

On Thu, 2004-09-30 at 18:14 +0300, Alex Efros wrote:
> Hi!
>
> On Thu, Sep 30, 2004 at 02:40:17PM +0100, Kerin Millar wrote:
> > Ouch, that's a nasty one. I suspect dabbling in sources variations will
> > not help a great deal because the gentoo-dev-sources are so lightly
> > patched in the first place. If anything, try 2.6.9-rc3. I performed a
> > cusory glance over the ChangeLog for "[NETFILTER]" and, while a few
> > patches have been applied, there was nothing that immediately suggested
> > that it would alleviate your problem.
>
> Is anybody know why all mm-sources masked? Only kernel which never hangs on
> my server was 2.6.5 from mm-sources, but I was forced to upgrade from 2.6.5
> to 2.6.7 some time ago because of some GLSA and I was unable to continue
> using -mm because at that time (and now too) all mm-sources was masked.

It's good that it worked out for you, but most people could testify to
the contrary. The case is basically this:

* Andrew Morton is the 2.6 maintainer
* His branch is a _highly_ experimental playground for code which may or
may not make it mainline later.
* Frequently, massively experimental code makes it own only to be
removed/changed in the next revision

Having said that, Andrew Morton is a superb bug fixer and he often
corrects various bugs ahead of time in his branch. It's an interesting
patchset in which to look for fixes - however I would say that, in the
vast majority of cases, it is madness to consider his branch for
production use. There are no guarantees whatsoever, and there is no
mandate that it should be remotely stable; this is why they are masked.
Not that there's anything to stop you unmasking the ebuilds (or even
grabbing an old one from the CVS Attic and adding the security fixes
yourself).

If you are saying that a modern, mainline instance of 2.6 is not stable
on your system then it would tend to suggest that something important
never made it mainline. You might want to review the ChangeLog for the
instance of -mm that you used to use. You might also want to take this
to the LKML. Also, Greg Kroah-Hartman is an uber-hacker and also happens
to be the gentoo-dev-sources maintainer. Perhaps if you ask him nicely
for advice, he might be able to point you in the right direction in
terms of determining just /why/ a newer kernel would not be stable where
an older, experimental one was.

Regards,

--Kerin Francis Millar