Re: Synchronous board drivers

Alan Cox (alan@lxorguk.ukuu.org.uk)
Sun, 4 Jul 1999 19:11:00 +0100 (BST)


> I agree with Krzysztof that sync board driver should be as simple
> as possible, and syncppp.c (or whatever it will be named) should do
> everything else (registering net device, switching link protocols, etc).

Why ?

What if I want to do things differently. Right now I can re-educate syncppp.
I can for example switch to hardware ppp and use software for X.25

> Then we probably should move from library approach (as in current
> syncppp.c) to a mid-layer approach (register_hdlc_dev(), struct
> hdlc_device_ops).

Almost every "mid layer" in the kernel has proved to be a mistake in the
long term. SCSI is the classic example

> : The "we ought to use pppd" case is quite valid. Its a 'when someone fixes it'
> : task
>
> I haven't looked at the async PPP code yet. What parts of the PPP
> is done by the kernel and what by the pppd?

The ppp kernel code grabs frames of known types (IP etc) the rest are vectored
to the daemon to handle in userspace

-
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/