nfs root problem in linux-next

nfs root problem in linux-next

[Mark Salter]

As of Friday's linux-next I've been unable to successfully boot using an
nfs root with a c6x board. I bisected it to this patch:

  commit cc3b296fd9a01c9e654bcb9373af70dd6e32fb24
  Author: Trond Myklebust <Trond.Myklebust-HgOvQuBEEgTQT0dZR+AlfA@public.gmane.org>
  Date:   Thu Jun 21 14:53:10 2012 -0400

      SUNRPC: Clean up xdr_read_pages and xdr_enter_page

I placed "good-boot" and "bad-boot" packet capture files at:

  http://people.redhat.com/~msalter/linux-next/

The only difference in the kernels used is the above patch. The
bad-boot mounted the nfs root but was unable to find the init
file which caused a panic. The nfsroot= parameter uses v3,tcp
flags and the server side is running rhel6.

I'm not at all familiar with the nfs code, but can help with
testing or providing any addtional info needed.

--Mark


--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: nfs root problem in linux-next

[Myklebust, Trond]

On Mon, 2012-06-25 at 16:26 -0400, Mark Salter wrote:
> As of Friday's linux-next I've been unable to successfully boot using an
> nfs root with a c6x board. I bisected it to this patch:
> 
>   commit cc3b296fd9a01c9e654bcb9373af70dd6e32fb24
>   Author: Trond Myklebust <Trond.Myklebust@netapp.com>
>   Date:   Thu Jun 21 14:53:10 2012 -0400
> 
>       SUNRPC: Clean up xdr_read_pages and xdr_enter_page
> 
> I placed "good-boot" and "bad-boot" packet capture files at:
> 
>   http://people.redhat.com/~msalter/linux-next/
> 
> The only difference in the kernels used is the above patch. The
> bad-boot mounted the nfs root but was unable to find the init
> file which caused a panic. The nfsroot= parameter uses v3,tcp
> flags and the server side is running rhel6.
> 
> I'm not at all familiar with the nfs code, but can help with
> testing or providing any addtional info needed.

Hi Mark,

So you are saying that reverting that patch alone suffices to fix the
problem that you are seeing?

Firstly, a big "Thank You!!!" for testing the linux-next branch. I very
much appreciate that you report the bugs before they get merged
upstream.

I'll have a look to see if I can find out what is the problem. The only
significant change that patch should be introducing would be the use of
xdr_stream_pos().

-- 
Trond Myklebust
Linux NFS client maintainer

NetApp
Trond.Myklebust@netapp.com
www.netapp.com

Re: nfs root problem in linux-next

[Mark Salter]

On Mon, 2012-06-25 at 23:14 +0000, Myklebust, Trond wrote:
> On Mon, 2012-06-25 at 16:26 -0400, Mark Salter wrote:
> > As of Friday's linux-next I've been unable to successfully boot using an
> > nfs root with a c6x board. I bisected it to this patch:
> > 
> >   commit cc3b296fd9a01c9e654bcb9373af70dd6e32fb24
> >   Author: Trond Myklebust <Trond.Myklebust-HgOvQuBEEgTQT0dZR+AlfA@public.gmane.org>
> >   Date:   Thu Jun 21 14:53:10 2012 -0400
> > 
> >       SUNRPC: Clean up xdr_read_pages and xdr_enter_page
> > 
> > I placed "good-boot" and "bad-boot" packet capture files at:
> > 
> >   http://people.redhat.com/~msalter/linux-next/
> > 
> > The only difference in the kernels used is the above patch. The
> > bad-boot mounted the nfs root but was unable to find the init
> > file which caused a panic. The nfsroot= parameter uses v3,tcp
> > flags and the server side is running rhel6.
> > 
> > I'm not at all familiar with the nfs code, but can help with
> > testing or providing any addtional info needed.
> 
> Hi Mark,
> 
> So you are saying that reverting that patch alone suffices to fix the
> problem that you are seeing?

Yes. If I check out a tree at commit cc3b296fd9a, the resulting kernel
does not boot (it panics that it cannot find init). If I check out a
tree at  cc3b296fd9a~1, it boots the nfs root successfully.

...
> I'll have a look to see if I can find out what is the problem. The only
> significant change that patch should be introducing would be the use of
> xdr_stream_pos().

Okay. I am happy to help with any testing if needed to help characterize
the problem better.

--Mark




--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: nfs root problem in linux-next

[Myklebust, Trond]

On Mon, 2012-06-25 at 23:17 -0400, Mark Salter wrote:
> On Mon, 2012-06-25 at 23:14 +0000, Myklebust, Trond wrote:
> > On Mon, 2012-06-25 at 16:26 -0400, Mark Salter wrote:
> > > As of Friday's linux-next I've been unable to successfully boot using an
> > > nfs root with a c6x board. I bisected it to this patch:
> > > 
> > >   commit cc3b296fd9a01c9e654bcb9373af70dd6e32fb24
> > >   Author: Trond Myklebust <Trond.Myklebust@netapp.com>
> > >   Date:   Thu Jun 21 14:53:10 2012 -0400
> > > 
> > >       SUNRPC: Clean up xdr_read_pages and xdr_enter_page
> > > 
> > > I placed "good-boot" and "bad-boot" packet capture files at:
> > > 
> > >   http://people.redhat.com/~msalter/linux-next/
> > > 
> > > The only difference in the kernels used is the above patch. The
> > > bad-boot mounted the nfs root but was unable to find the init
> > > file which caused a panic. The nfsroot= parameter uses v3,tcp
> > > flags and the server side is running rhel6.
> > > 
> > > I'm not at all familiar with the nfs code, but can help with
> > > testing or providing any addtional info needed.
> > 
> > Hi Mark,
> > 
> > So you are saying that reverting that patch alone suffices to fix the
> > problem that you are seeing?
> 
> Yes. If I check out a tree at commit cc3b296fd9a, the resulting kernel
> does not boot (it panics that it cannot find init). If I check out a
> tree at  cc3b296fd9a~1, it boots the nfs root successfully.
> 
> ...
> > I'll have a look to see if I can find out what is the problem. The only
> > significant change that patch should be introducing would be the use of
> > xdr_stream_pos().
> 
> Okay. I am happy to help with any testing if needed to help characterize
> the problem better.

Hi Mark and Fengguang,

I believe that I've found the bug that was being triggered in the above
patch. It was a mistake in the earlier patch "SUNRPC: Don't decode
beyond the end of the RPC reply message", which was failing to correctly
initialise the xdr->nwords field.

I've now fixed it and updated the linux-next branch.

Cheers
  Trond
-- 
Trond Myklebust
Linux NFS client maintainer

NetApp
Trond.Myklebust@netapp.com
www.netapp.com

Re: nfs root problem in linux-next

[Mark Salter]

On Fri, 2012-06-29 at 13:55 +0000, Myklebust, Trond wrote:
> Hi Mark and Fengguang,
> 
> I believe that I've found the bug that was being triggered in the
> above
> patch. It was a mistake in the earlier patch "SUNRPC: Don't decode
> beyond the end of the RPC reply message", which was failing to
> correctly
> initialise the xdr->nwords field.
> 
> I've now fixed it and updated the linux-next branch.

Yes, I can confirm that today's linux-next is booting correctly
for me on an nfsroot.

Thanks!

Re: nfs root problem in linux-next

[Fengguang Wu]

On Fri, Jun 29, 2012 at 10:48:48AM -0400, Mark Salter wrote:
> On Fri, 2012-06-29 at 13:55 +0000, Myklebust, Trond wrote:
> > Hi Mark and Fengguang,
> > 
> > I believe that I've found the bug that was being triggered in the
> > above
> > patch. It was a mistake in the earlier patch "SUNRPC: Don't decode
> > beyond the end of the RPC reply message", which was failing to
> > correctly
> > initialise the xdr->nwords field.
> > 
> > I've now fixed it and updated the linux-next branch.
> 
> Yes, I can confirm that today's linux-next is booting correctly
> for me on an nfsroot.
> 
> Thanks!
 
I can confirm it, too. Thanks!