Re: unprivileged mounts git tree

[ebiederm@xmission.com (Eric W. Biederman)]

"Serge E. Hallyn" <serue@us.ibm.com> writes:

> Quoting Miklos Szeredi (miklos@szeredi.hu):
>> Here's a git tree of the unprivileged mounts patchset:
>> 
>> git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs.git
> unprivileged-mounts
>> 
>> Could this be added to -mm (and dropped if it's in the way of
>> something) for some testing and added visibility until it's reviewed
>> by Christoph/Al?
>> 
>> I'm not reposting the whole patchset, since it's essentially the same
>> as the last submission, only updated to the latest git.  But if
>> somebody wants it I can post them.
>> 
>> Thanks,
>> Miklos
>> 
>> 
>>  Documentation/filesystems/fuse.txt |   88 ++++++++-
>>  Documentation/filesystems/proc.txt |   40 ++++
>>  fs/filesystems.c                   |   60 ++++++
>>  fs/fuse/inode.c                    |   21 ++
>>  fs/internal.h                      |    3 +-
>> fs/namespace.c | 366 +++++++++++++++++++++++++++---------
>>  fs/pnode.c                         |   22 ++-
>>  fs/pnode.h                         |    2 +
>>  fs/super.c                         |   26 ---
>>  include/linux/fs.h                 |    7 +
>>  include/linux/mount.h              |    4 +
>>  kernel/sysctl.c                    |   16 ++
>>  12 files changed, 527 insertions(+), 128 deletions(-)
>> 
>> Miklos Szeredi (10):
>>       unprivileged mounts: add user mounts to the kernel
>>       unprivileged mounts: allow unprivileged umount
>>       unprivileged mounts: propagate error values from clone_mnt
>>       unprivileged mounts: account user mounts
>>       unprivileged mounts: allow unprivileged bind mounts
>>       unprivileged mounts: allow unprivileged mounts
>>       unprivileged mounts: add sysctl tunable for "safe" property
>>       unprivileged mounts: make fuse safe
>>       unprivileged mounts: propagation: inherit owner from parent
>>       unprivileged mounts: add "no submounts" flag
>
> Hi Miklos,
>
> so on the bright side I pulled this tree today and it compiled and
> passed ltp with no problems.
>
> But then I played around a bit and found I could do the following:
>
> (hmm, i'm trying to remember the exact order :)
>
> as root:
> 	mmount --bind -o user=500 /home/hallyn/etc/ /home/hallyn/etc/
> 	mount --bind /mnt /mnt
> 	mount --make-rshared /mnt
> 	mount --bind /dev /mnt/dev
>
> as hallyn:
> 	mmount --bind /mnt /home/hallyn/etc/mnt
> 	/usr/src/mmount-0.3/mmount --bind mnt/dev mnt/src

You are using relative directory names here which makes it confusing.
I'm assuming you in /home/hallyn/etc ?

>
> Now /mnt/src contained /dev.
>
> Is this what we want?

I don't think so.

I think the simplest answer is to not allow mounting of shared
subtrees controlled by a different user.

Serge I think you are right downgrading the mount from shared to slave
looks like the sane thing to do if the mount owners match.

> Do we want to tell the admin it's his fault for
> not somehow forcing a slave relationship between /mnt and
> /home/hallyn/etc/mnt?  Except I don't think he can do that preemptively,
> it has to be done after hallyn does the mmount.
>
> So does that mean that if non-root user X does:
>
> 	mount a b
>
> where b is user=X but a is not, then if a is shared we should force it
> to be mounted as slave at b?
>
> -serge