On Wed, Oct 15, 2014 at 07:58:59AM -0700, Andy Lutomirski wrote: > On 10/14/2014 07:25 AM, Seth Forshee wrote: > > Unprivileged users are normally restricted from mounting with the > > allow_other option by system policy, but this could be bypassed > > for a mount done with user namespace root permissions. In such > > cases allow_other should not allow users outside the user > > namespace to access the mount as doing so would give the > > unprivileged user the ability to manipulate processes it would > > otherwise be unable to manipulate. > > What threat is this intended to protect against? I think that, if this > is needed, tasks outside the userns or its descendents should be > blocked, even if the user ids match. That is, I think you should check > the namespace, not the uid and gid. allow_other is an existing option in fuse to protect against DoS attacks against more privileged users by making file operations block indefinitely. So this change makes it work the same way inside a user namespace but only to users mapped into the namespace. Checking the namespace does seem to make more sense, so I'll make that change. Thanks, Seth -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
