Re: [RFC V2 10/18] famfs_fuse: Basic fuse kernel ABI enablement for famfs

["Darrick J. Wong" <djwong@kernel.org>]

On Thu, Jul 03, 2025 at 05:45:48PM -0500, John Groves wrote:
> On 25/07/03 01:50PM, John Groves wrote:
> > * FUSE_DAX_FMAP flag in INIT request/reply
> > 
> > * fuse_conn->famfs_iomap (enable famfs-mapped files) to denote a
> >   famfs-enabled connection
> > 
> > Signed-off-by: John Groves <john@groves.net>
> > ---
> >  fs/fuse/fuse_i.h          |  3 +++
> >  fs/fuse/inode.c           | 14 ++++++++++++++
> >  include/uapi/linux/fuse.h |  4 ++++
> >  3 files changed, 21 insertions(+)
> > 
> > diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h
> > index 9d87ac48d724..a592c1002861 100644
> > --- a/fs/fuse/fuse_i.h
> > +++ b/fs/fuse/fuse_i.h
> > @@ -873,6 +873,9 @@ struct fuse_conn {
> >  	/* Use io_uring for communication */
> >  	unsigned int io_uring;
> >  
> > +	/* dev_dax_iomap support for famfs */
> > +	unsigned int famfs_iomap:1;
> > +
> >  	/** Maximum stack depth for passthrough backing files */
> >  	int max_stack_depth;
> >  
> > diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c
> > index 29147657a99f..e48e11c3f9f3 100644
> > --- a/fs/fuse/inode.c
> > +++ b/fs/fuse/inode.c
> > @@ -1392,6 +1392,18 @@ static void process_init_reply(struct fuse_mount *fm, struct fuse_args *args,
> >  			}
> >  			if (flags & FUSE_OVER_IO_URING && fuse_uring_enabled())
> >  				fc->io_uring = 1;
> > +			if (IS_ENABLED(CONFIG_FUSE_FAMFS_DAX) &&
> > +			    flags & FUSE_DAX_FMAP) {
> > +				/* XXX: Should also check that fuse server
> > +				 * has CAP_SYS_RAWIO and/or CAP_SYS_ADMIN,
> > +				 * since it is directing the kernel to access
> > +				 * dax memory directly - but this function
> > +				 * appears not to be called in fuse server
> > +				 * process context (b/c even if it drops
> > +				 * those capabilities, they are held here).
> > +				 */
> > +				fc->famfs_iomap = 1;
> 
> I think there should be a check here that the fuse server is 
> capable(CAP_SYS_RAWIO) (or maybe CAP_SYS_ADMIN), but this function doesn't 
> run in fuse server context. A famfs fuse server is providing fmaps, which 
> map files to devdax memory, which should not be an unprivileged operation.

I thought process_init_reply /does/ run in the fuse server's context.
It calls process_init_limits, which checks for capable(CAP_SYS_ADMIN)...

--D

> 1) Does fs/fuse already store the capabilities of the fuse server?
> 2) If not, where do you suggest I do that, and where do you suggest I store
> that info? The only dead-obvious place (to me) that fs/fuse runs in server
> context is in fuse_dev_open(), but it doesn't store anything...
> 
> @Miklos, I'd appreciate your advice here.
> 
> Thanks!
> John
> 
>