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

[Steve French]

Trying it just now, the patch causes an extra QueryPathInfo on every mount,
and presumably it can't be cached information since it doesn't use the
normal Unix vs. Non-Unix Query Path Info check and because we special cause
the lookup of the root inode.

On Sat, Feb 7, 2009 at 8:34 PM, Steve French <smfrench at gmail.com> wrote:

> Have to get rid of the "sparse" warning (CHECK   fs/cifs/connect.c
> fs/cifs/connect.c:2184:1: warning: symbol 'is_path_accessible' was not
> declared. Should it be static?)
>
> and also verify that this does not cause an extra QueryPathInfo on every
> mount in the normal path.
>
>
> On Wed, Feb 4, 2009 at 7:31 AM, Jeff Layton <jlayton at redhat.com> wrote:
>
>> On Wed, 04 Feb 2009 16:25:46 +0300
>> Igor Mammedov <niallain at gmail.com> wrote:
>>
>> > Patch with Jeff's suggestions is attached.
>> >
>> > Jeff Layton wrote:
>> > > On Wed, 04 Feb 2009 13:35:16 +0300
>> > ...
>> > >
>> > > Maybe something like:
>> > >
>> > >     cERROR(1, ("Path %s not accessable: %d", full_path, rc));
>> > >
>> > > ...of course, you'd need to not kfree full_path until after this is
>> > > printed but that's not too tough.
>> > >
>> > > Won't this error also fire when trying to mount with an inaccessable
>> > > prefixpath, even if DFS is involved? If so, then mentioning DFS in
>> > > the error is probably going to be confusing for users/admins.
>> > >
>> > > Actually, there's another problem too. If build_path_to_root ends up
>> > > returning NULL, then we'll just skip this check. Shouldn't we return
>> > > -ENOMEM or something at that point? It's probably better to fail the
>> > > mount than to leave the client subject to a later oops...
>> > >
>> >
>>
>> Looks good to me:
>>
>> Acked-by: Jeff Layton <jlayton at redhat.com>
>>
>
>
>
> --
> Thanks,
>
> Steve
>



-- 
Thanks,

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