Subject: Re: Call for review: The New Block/Character Device Switch
To: None <greywolf@starwolf.com>
From: MAEKAWA Masahide <maekawa@math.sci.kobe-u.ac.jp>
List: tech-kern
Date: 06/28/2001 17:22:41
Greywolf <greywolf@starwolf.com> wrote:
># >Okay, fine. Now how does this bind to special devices so that the
># >dynamic assignment maps to something usable in /dev? Or do you plan
># >to also provide a freebsd style devfs at the same time?
># 
># Their implementation needs uid/gid hard-codings.
># I don't think it is a Good Thing. At least I will take
># an other way to support a dynamic devnode configuration.
># 
># "Can we accept hard-codings?"
># It's the only essential issue for this, I think.
>
>You only need hard-codings until the devfs populates and you can mount
>the root filesystem at which point an /etc/dev.conf or some such can
>be taken into account at some point during the boot process.
>
>I don't think it's nearly so important as to which devices have what
>permissions until you begin the transition to multi-user mode.

I would say "hard-codings of uid/gid should not be in kernel".

--- MAEKAWA Masahide