Jeremy Fitzhardinge writes:
>
> On 15-Apr-2000 H. Peter Anvin wrote:
> > Pretty much. It still begs the question why you need the virtual
> > filesystem underneath though.
>
> Yep. All you'd need is the 'device present/removed' events from the
> kernel and the daemon could create the nodes.
I considered this a while back. You still want to catch lookup events,
and you also want the kernel to present a list of installed devices to
user-space. A FS is the natural way of doing this.
You also want to have a tighter binding between device nodes and
drivers, which you can't do without devfs.
Also, "catching" device present/removed events is only good after
you've booted user-space. What about the built-in drivers? So you
still need to present the hardware somehow, and a FS is the logical
way. If you present them another way, you *still* have add kernel code
to talk to devfsd (you need more than what autofs gives you). It's
much simpler to have devfs.
I've been over this time and time again. People keep inventing ways of
giving the benefits of devfs without needing devfs, but then there's
so many different things that you can no longer do. And it's offensive
to just dismiss these things, as they are legitimate and provide real
power to people who are trying to solve real problems.
Regards,
Richard....
Permanent: rgooch@atnf.csiro.au
Current: rgooch@ras.ucalgary.ca
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sat Apr 15 2000 - 21:00:27 EST