linux-mips-fnet
[Top] [All Lists]

Re: DECStation zs driver continued...

To: linux-mips@fnet.fr
Subject: Re: DECStation zs driver continued...
From: Harald Koerfgen <Harald.Koerfgen@home.ivm.de>
Date: Wed, 30 Jun 1999 17:56:23 +0200 (MEST)
In-reply-to: <3779C1F8.A000B45D@niisi.msk.ru>
Organization: none
Reply-to: "Harald Koerfgen" <Harald.Koerfgen@home.ivm.de>
Sender: harry@franz.no.dom
Hi,

On 30-Jun-99 Gleb O. Raiko wrote:
> I'd like to note that I debugged DECstation zs at the point where it's
> stable enough to carry terminal, keyboard, and mouse. The only thing I
> have in TODO list is wait_until_sent. I think I'll ready to send a patch
> during next week.

Excellent! Do you see a chance to unify your Baget driver and the
DECstation driver?

As a side note, the B ports on the DECstations can do DMA...
 
> BTW, we've got at least 5 lance drivers (PC, Sparc, Atari, DECstation,
> and Baget).

Not to forget the HP 300 Series.

Yes, i noticed that too. I was thinking about using 7990.c and writing a
DECstation backend similar to hplance.c. I didn't do this because the
layouts of the shared memory (yes, there are three different ones for three
lance implementations!) are so weird that it's hardly possible to write a
driver that covers all three DECstation interfaces. That's why the support
for PMAD-AA cards (TURBOchannel ethernet interfaces) is broken.

The NetBSD people are working with shadow copies of the init block and the
ring descriptors. They're manipulating the copies and copy them to the
shared memory afterwards with specialzed copy routines. Although this is
an elegant solution from a code point of view I was trying to avoid this
additional copy.

Nevertheless, I haven't given up yet and I still have some ideas for a
unified lance driver.
---
Regards,
Harald

<Prev in Thread] Current Thread [Next in Thread>