[linux-cifs-client] Re: [PATCH] [CIFS] Prevent OOPs when mounting with remote prefixpath

[Steve French]

The DFS mount now works?

On Fri, Feb 13, 2009 at 11:32 AM, Shirish Pargaonkar <
shirishpargaonkar at gmail.com> wrote:

> On Fri, Feb 13, 2009 at 8:36 AM, Jeff Layton <jlayton at redhat.com> wrote:
> > On Thu, 12 Feb 2009 23:18:16 -0600
> > Shirish Pargaonkar <shirishpargaonkar at gmail.com> wrote:
> >
> >> On Tue, Feb 3, 2009 at 4:22 AM, Igor Mammedov <niallain at gmail.com>
> wrote:
> >> > Managed to single out the part of  DFS root support from a big patch
> >> > with a little modification.
> >> > It will report error and fail to mount if perfixpath is on remote
> server.
> >> > As well it will not scare people off with a error
> >> > 'kernel BUG at fs/cifs/cifs_dfs_ref.c:274!'.
> >> >
> >> >
> >> > ---------
> >> > Fixes OOPs with message 'kernel BUG at fs/cifs/cifs_dfs_ref.c:274!'.
> >> > Check if prefixpath in accessible while we are still in cifs_mount
> >> > and fail with reporting a error if we can't access prefixpath
> >> > (i.e. if prefixpath is located on another server)
> >> >
> >> >
> >> > Best regards,
> >> >
> >> > -------------------------
> >> > Igor Mammedov,
> >> > niallain "at" gmail.com
> >> >
> >> >
> >> >
> >> >
> >> >
> >>
> >> We need to bring back NIPQUAD and NIPQUAD_FMT back in cifs_mount
> >> instead of %pI4 i..e.
> >> sprintf(pSesInfo->serverName, NIPQUAD_FMT,
> >>
> NIPQUAD(srvTcp->addr.sockAddr.sin_addr.s_addr));
> >> If we sprintf hex address in serverName instead of dotted-decimal
> >> notation, the Path parameter in Tree Connect
> >> to IPC$ share on the server to get dfs referral fails,
> >> \\hex_address\IPC$ is not a valid UNC path name.
> >>
> >> I have not tried on a IPv6 setup, but I suspect the same applies to
> >> IPv6 formatting.
> >>
> >
> > I'm afraid I don't follow this. %pI4 formats the address in
> > dotted-decimal format, not as a hex address. %pI6 should format it the
> > same way that NIP6_FMT did.
> >
> > --
> > Jeff Layton <jlayton at redhat.com>
> >
>
> Jeff,
>
> You are right, it works. I was using 2.6.29-rc3 cifs code in 2.6.27
> and  I think that is why.
>
> Regards,
>
> Shirish
>



-- 
Thanks,

Steve
-------------- next part --------------
HTML attachment scrubbed and removed