On Tue, 04.01.11 01:14, Michael Biebl (mbi...@gmail.com) wrote: > > Hmm, I normally would be happy to do that, but I am a bit concerned > > about the gid=tty parameter. Before we pass that off to the kernel we'd > > have to look that up and pass a numeric gid on. But doing NSS queries > > from PID 1 is a little bit dangerous. Hence I'd actually prefer if > > people apply this in /etc/fstab. > > > > Ideas? Suggestions? > > The mountdevsubfs init script does indeed use a numeric value (= 5) > for gid as that is guaranteed to be tty on Debian. > I just checked a Ubuntu 10.10, Fedora 14 and openSUSE 11.4 system and > they all use tty:5. > > If this is an assumption though that can be safely made accros distros, I > dunno.
Hmm, we could lookup the ID in a configure check at build time which fails if the "tty" group doesn't exist. I guess that wouldn't be too bad. Might have weird effects on cross-compilation possibly -- but if we provide --with-tty-gid= as configure parameter this should be fine. Anybody has a good reason why we shouldn't solve this with a compile-time NSS lookup? Lennart -- Lennart Poettering - Red Hat, Inc. _______________________________________________ systemd-devel mailing list systemd-devel@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/systemd-devel
