nfs kernel server causes: kernel BUG at include/linux/module.h:386!

nfs kernel server causes: kernel BUG at include/linux/module.h:386!

[roma1390]

Hello,

after some nfs server start/stop i hit kernel bug on 2.6.26. There is hope[1] 
that this bug is fixed[2], and fix is merged[3] into 2.6.29. But there is 
missing bugfix for older stable kernels.

Reports for this bug can also be found on distro bug systems[4][5][6][7]. 
Sadly these bugs are still open.

Please backport fix for stable kernels too.

roma1390

1. http://linux-nfs.org/pipermail/nfsv4/2009-March/010018.html
2. 
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=22945e4a1c7454c97f5d8aee1ef526c83fef3223
3. 
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=713404d6082fee34a829a0c6d511e4aec72d7654
4. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=516376
5. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=519348
6. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=532561
7. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=538000

Re: [stable] nfs kernel server causes: kernel BUG at include/linux/module.h:386!

[Greg KH]

On Thu, May 06, 2010 at 08:47:27AM +0300, roma1390 wrote:
> Hello,
>
> after some nfs server start/stop i hit kernel bug on 2.6.26. There is 
> hope[1] that this bug is fixed[2], and fix is merged[3] into 2.6.29. But 
> there is missing bugfix for older stable kernels.

2.6.26 is not supported by the community anymore.

Is this a problem on the latest 2.6.27-stable kernel release?  If so,
which specific patches in Linus's tree do you think resolves this?  Is
it just the 22945e4a1c7454c97f5d8aee1ef526 commit that is needed to be
added there?

thanks,

greg k-h

Re: [stable] nfs kernel server causes: kernel BUG at include/linux/module.h:386!

[roma1390]

Greg KH wrote:
 > On Thu, May 06, 2010 at 08:47:27AM +0300, roma1390 wrote:
 >> Hello,
 >>
 >> after some nfs server start/stop i hit kernel bug on 2.6.26. There is 
hope[1] that this bug is fixed[2], and fix is merged[3] into 2.6.29. But there 
is missing bugfix for older stable kernels.
 >
 > 2.6.26 is not supported by the community anymore.

Sadly, the debian 5.0 (current stable) chose 2.6.26 as distribution kernel. Of 
course they do kernel as part of distribution support by them self.

 > Is this a problem on the latest 2.6.27-stable kernel release?  If so,
 > which specific patches in Linus's tree do you think resolves this?  Is
 > it just the 22945e4a1c7454c97f5d8aee1ef526 commit that is needed to be
 > added there?

Currently there isn't any free resources to do 2.6.27-stable testing. 
Switching from distribution kernel to custom one adds much more work for 
system admin, so this options is the last one none other will work.

If there is any demand to test 22945e4a1c7454c97f5d8aee1ef526, there can be 
such test done. But this issue was found only once, and on first hour after 
nfs server configuration was started.

So the average system uptime from 300 days without nfs droped to 1 hour with nfs.

Currently I didn't know how to reproduce this situation (and didn't tried). 
But as i found this issue is known, is faced by lot of people, and nfs 
maintainer states that 22945e4a1c7454c97f5d8aee1ef526 is cure to this problem.

Re: [stable] nfs kernel server causes: kernel BUG at include/linux/module.h:386!

[Greg KH]

On Fri, May 07, 2010 at 10:10:23AM +0300, roma1390 wrote:
> Greg KH wrote:
> > On Thu, May 06, 2010 at 08:47:27AM +0300, roma1390 wrote:
> >> Hello,
> >>
> >> after some nfs server start/stop i hit kernel bug on 2.6.26.
> There is hope[1] that this bug is fixed[2], and fix is merged[3]
> into 2.6.29. But there is missing bugfix for older stable kernels.
> >
> > 2.6.26 is not supported by the community anymore.
> 
> Sadly, the debian 5.0 (current stable) chose 2.6.26 as distribution
> kernel. Of course they do kernel as part of distribution support by
> them self.
> 
> > Is this a problem on the latest 2.6.27-stable kernel release?  If so,
> > which specific patches in Linus's tree do you think resolves this?  Is
> > it just the 22945e4a1c7454c97f5d8aee1ef526 commit that is needed to be
> > added there?
> 
> Currently there isn't any free resources to do 2.6.27-stable
> testing. Switching from distribution kernel to custom one adds much
> more work for system admin, so this options is the last one none
> other will work.
> 
> If there is any demand to test 22945e4a1c7454c97f5d8aee1ef526, there
> can be such test done. But this issue was found only once, and on
> first hour after nfs server configuration was started.

If you can't test it, if I add that commit id to .27, I don't see how
that would help you out at all, right?

Can you poke the debian developers about this?

thanks,

greg k-h

Re: [stable] nfs kernel server causes: kernel BUG at include/linux/module.h:386!

[roma1390]

Greg KH wrote:
> If you can't test it, if I add that commit id to .27, I don't see how
> that would help you out at all, right?

There isn't a big like to test on production. But rescent that same bug was 
triggered on testing in virtual machine. So theoretically this is only matter 
of human time resource.

There is hope that commit to 2.26.x will be taken by debian, and updated 
kernel will be propagated by debian update system.

> Can you poke the debian developers about this?
Case are reported[1] to debian and waiting for any progress.

1. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=532561#21