[Gluster-devel] RFC: FUSE kernel features to be adopted by GlusterFS

[Niels de Vos]

On Thu, Nov 09, 2017 at 06:01:10PM -0500, Jeff Darcy wrote:
> > So - nothing inherent to libfuse and nothing that would be relevant as
> > of today.
> > 
> > But then, let me put it like this: what reason could we have to *not* go
> > with xglfs in 4.0? It's true that the deliverables are present in libfuse
> > and
> > libgfapi and it's just a thin glue. As such it seems to be almost devoid
> > of design concerns, it just bridges the two interfaces in a
> > straightforward manner. Superficially it seems to be a superior approach
> > - what snag holds us back to embrace it wholeheartedly?
> 
> What about https://review.gluster.org/#/c/3341/ and its antecedents? 
> Libfuse used to be unable to deal with SELinux's behavior of trying to
> issue a getxattr from within the mount call.  Have either libfuse or
> SELinux fixed that?  There might be other local changes that we'd need
> to verify in similar fashion.

The FUSE kernel module reverted patches that made it support SELinux.
Currently no FUSE filesystem can support SELinux :-(

There has been some work done through
https://bugzilla.redhat.com/1272868 , but we need to spend more time on
it to get the kernel work functional. IIRC it needs some changes in the
SELinux part, VFS-hooks and FUSE. Once that is in place, we can see how
it works together with the selinux-policy, our fuse implementation and
libfuse.

Niels