Re: [PATCH v5 2/4] fuse: Support fuse filesystems outside of init_user_ns

[Andy Lutomirski <luto@amacapital.net>]

On Fri, Nov 21, 2014 at 10:14 AM, Eric W. Biederman
<ebiederm@xmission.com> wrote:
> - Tweak the file capability code to look at s_user_ns and treat it
>   properly.
>

> - Tweak the security checks to allow setting file capabilities and
>   other security xattrs if we have the appropriate capabilities in
>   s_user_ns.
>

Thinking about this some more, what do you mean by tweaking the file
capability code to look at s_user_ns and treat it properly?

I think that the semantics should be that cap_inode_setxattr should
check ns_capable wrt s_user_ns, but that the fscap *consumer* should
check the mount as in my may_suid patch (and maybe also check
s_user_ns).  There is legacy code that starts a FUSE server as global
root, mounts the thing in a mount namespace belonging to an
unprivileged user ns, and (I think) hands the /dev/fuse fd to that
unprivileged code.

Without the mount ns check, that FUSE server can take over the system.
With the mount ns check, it's safe.

--Andy

>
> When those bits are done we can tweak the fuse patches to also set
> s_user_ns.
>
> As for MNT_NO_SUID if fuse wants to enforce that in some way.  I don't
> particularly care, but I don't think that makes sense as a vfs property.
>
> Eric



-- 
Andy Lutomirski
AMA Capital Management, LLC