[Debian] Re: lenny updates

[dann frazier]

On Mon, Mar 09, 2009 at 08:13:28PM +0100, Ola Lundqvist wrote:
> Hi Dann
> 
> On Sun, Mar 08, 2009 at 11:17:09PM -0600, dann frazier wrote:
> > On Tue, Mar 03, 2009 at 09:44:04PM +0100, Ola Lundqvist wrote:
> > > Hi Dann
> > > 
> > > You asked about the latest status and here it is.
> > > Please tell which ones you want me to fix for the next lenny release of the kernel. I'll prepare
> > > a patch and regression test that version for you.
> > > 
> > > #510787:
> > > Refers to an other bug report that was not openvz specific. Should it be
> > > forwarded to an non-openvz version of the kernel or kept here?
> > 
> > I don't think it really matters - you can reassign to linux-2.6 if you
> > like though.
> 
> Done that now.
> 
> > > In any case I have added latest information to the report and told where
> > > the problem has been forwarded.
> > 
> > Thanks!
> > 
> > > #511165:
> > > Patch exist for 2.6.24 and 2.6.26. Fix is available in
> > > http://git.openvz.org/?p=linux-2.6.26-openvz;a=commit;h=b5e1f74cee5bc2c45bdca53a7218fb8de89215dd
> > > Not sure if this is an ABI breaker.
> > 
> > Seems straightforward, and shouldn't change the ABI. I'll commit it
> > assuming my test build shows that.
> 
> This tells me that there is an easy way to check that. How is that done?
> I assume some files are compared, but I can not find that in the debian directory (without building).

This is just based on looking at the patch - I don't see anything
there that would change an exported symbol. Things to look for
structure definition changes, or changes to exported function calls.
Adding/removing includes can also have non-obvious effects.

My test build did succeed, and this change has been committed.

> > > #500876:
> > > Fix available in:
> > > http://git.openvz.org/?p=linux-2.6.26-openvz;a=commit;h=777e8164ebf8a03e43511983cdec472f8691a8af
> > > Problem is about to be verified. Regression tested without problems seen.
> > 
> > I couldn't reproduce this one (tried dual quad core intel server & a
> > single quad core amd), but user claims this fixed the bug for me and I
> > haven't seen any issues with this patch so its been committed.
> 
> I think you need to have a quad-core amd64 for this. But let us commit it
> as it do not seem to hurt.

I tried on a quad-core/one socket amd64, but maybe it requires
multiple sockets?

> > > #503097:
> > > Reported as http://bugzilla.openvz.org/show_bug.cgi?id=930
> > > Seems to be a duplicate of #500876 above.
> > 
> > Cool. If you think so, it might be good to have Carlos test one of
> > these builds to verify:
> >   http://people.debian.org/~dannf/bugs/500876/
> 
> Ok, I'll ask him at once.

Thanks!

> > (Tomorrow's snapshot builds should also include it)
> > 
> > > #505174:
> > > This is a request to go up to the latest version that includes fixes for
> > > all the ones in this mail that describe that there is a fix available.
> > > Unfortunatly there are ABI breakers...
> > 
> > Its probably a good idea to stick with specific issues/fixes now that
> > its a stable release.
> 
> Maybe so. The openvz development team has proven to provide quite well
> tested kernels. However the safer approach may still be to stick to the
> specific issues.

Yeah, I certainly don't mistrust the openvz team - but all changes
come with a risk of regression, and its easier to regression test
specific changes.

> > > #508773:
> > > Patch available in http://bugzilla.openvz.org/show_bug.cgi?id=1054
> > > Fix in http://git.openvz.org/?p=linux-2.6.24-openvz;a=commit;h=20bd90762d4df4a3c7c247b660c696bdd0a27709
> > > Do not look like an ABI breaker to me.
> > 
> > Yep, definitely shouldn't break the ABI, and seems like a good
> > candidate.
> 
> Good. Please tell if you want me to prepare some patch or check in something.

Patches that apply directly to kernel svn are certainly
welcome. Normally, that should mean adding a changelog entry, a series
file entry, and a new patch file.

This one is pretty trivial, so I'll go ahead and generate a changeset
to commit - but it would be great if you could ensure that the
snapshot builds get regression tested after they become available.

> > > #500145:
> > > Forwarded to http://bugzilla.openvz.org/show_bug.cgi?id=1143
> > > Marked as dupliate of http://bugzilla.openvz.org/show_bug.cgi?id=1067
> > > Not solved yet.
> > 
> > ok
> > 
> > > #501985:
> > > From: maximilian attems
> > > the upstream nfs fixes are abi breakers and thus can't be integrated
> > > at this point they will be for the first point release were abi
> > > breaking will be allowed again.
> > 
> > What is the fix for this - does upstream openvz include it?
> 
> Yes it is found upstream. See the file
> http://download.openvz.org/kernel/branches/2.6.26/current/patches/patch-chekhov.1-combined.gz
> The current patch do not touch any nfs/ files and upstream does. The patch
> now in use was not fully completed when it was incorporated by Maximilian.

I see - so we need to identify which additional changes are needed.
http://git.openvz.org/?p=linux-2.6.26-openvz;a=commitdiff;h=66ec7f7f493fb98e8baa6591e9225086ae640fb8
http://git.openvz.org/?p=linux-2.6.26-openvz;a=commitdiff;h=39bb1ee59237272cd20e1f8696cefbd6a787cfc8

Is this severe enough to fix in a stable release? If we consider this
a regression from etch (since kernel-patch-openvz supplied this), than
maybe so. Is the risk of regression low? Well - these patches would
only get applied on openvz kernels which currently don't support nfs
at all so, assuming these changes are nfs-specific, risk should be
low.

> > > #494445:
> > > There are a number of problems in this area. Fixes are available.
> > > However some of them are ABI breakers.
> > 
> > The nf_conntrack_ipv6 module doesn't appear to be in 2.6.26-13. Maybe
> > it was disabled because of this bug? At this point, turning it
> > on/fixing probably falls into the category of a feature requests that
> > doesn't enable hardware, so wouldn't have a sufficient severity (>=
> > important).
> 
> True.
> 
> > > #500645:
> > > Fix available in http://bugzilla.openvz.org/show_bug.cgi?id=1034
> > > http://git.openvz.org/?p=linux-2.6.26-openvz;a=commit;h=6d18ba377cfa3e86ee830fe6a5fce52b8fd51039
> > > I can not see that this is an ABI breaker, so it should be possibly to
> > > apply this one without problem.
> > 
> > The patch itself certainly looks trivial enough - but the bug is only of
> > severity "normal". If we think this actually deserves a >= important
> > severity, we should bump the severity of the report.
> 
> Yes this one is really important. I'll change the severity now.

ok

-- 
dann frazier