Re: VFS, NFS security bug? Should CAP_MKNOD and CAP_LINUX_IMMUTABLE be added to CAP_FS_MASK?

["J. Bruce Fields" <bfields@fieldses.org>]

On Thu, Mar 12, 2009 at 11:03:00AM -0500, Serge E. Hallyn wrote:
> Quoting J. Bruce Fields (bfields@fieldses.org):
> > On Wed, Mar 11, 2009 at 03:53:34PM +0300, Igor Zhbanov wrote:
> > > Hello!
> > > 
> > > It seems that CAP_MKNOD and CAP_LINUX_IMMUTABLE were forgotten to be
> 
> (Still looking into this, but meanwhile...)
> 
> > > added to CAP_FS_MASK_B0 in linux-2.6.x and to CAP_FS_MASK in
> > > linux-2.4.x. Both capabilities affects file system and can be
> > > considered file system capabilities.
> > 
> > Sounds right to me--I'd expect rootsquash to guarantee that new device
> > nodes can't be created from the network.  Cc'ing random people from the
> > git log for include/linux/capability.h in hopes they can help.
> > 
> > --b.
> > 
> > (Also: my copy of mknod(2) claims "Linux... does not have the CAP_MKNOD
> > capability".  I assume the manpage is out of date?)
> 
> No, the whole paragraph is:
> 
> EPERM  mode  requested creation of something other than a regular file, FIFO
> (named pipe), or Unix domain socket, and the caller is not privileged
> (Linux: does not have the CAP_MKNOD capability);
> 
> So it's saying that 'caller is not privileged', in linux, can be
> interpreted to mean 'the caller does not have CAP_MKNOD'.

Ah!  Foiled by punctuation.  Apologies!--b.