Subject: Re: Simple fix for 1.5.x/i386, trouble elsewhere (was: an driver trouble)
To: None <port-alpha@netbsd.org>
From: Juergen Weiss <weiss@Uni-Mainz.DE>
List: port-alpha
Date: 03/03/2001 20:04:39
Feico Dillema writes:
 > On Fri, Mar 02, 2001 at 12:44:44AM +0100, Feico Dillema wrote:
 > > On Thu, Mar 01, 2001 at 12:49:03PM +0100, Feico Dillema wrote:
 > > Finally, I'm trying to get the an card to work on 1.5.1/alpha for
 > > my Multia. The an-driver itself seems ok sofar, but PCMCIA interrupts
 > > still do not work properly; changing IST_EDGE into IST_LEVEL in the
 > > pcic-ic code helped. 
 > Ok, sorry abuot this one. Just read the comments in -current source
 > that this is plain wrong and doesn't fix anything. Should have spotted
 > that one earlier...

This change was in my December fixes for the Multia. It was
supposed to be used with the Multia only. Certainly ISA
interrupts should be edge triggered, but unfortunately the
Multia does not know this :-(.  The two com devices and the
pcmcia controller work with level triggered interrupts, but they
seem not to work with edge triggered interrupts. In NetBSD 1.4
and older NetBSD versions the com devices were always probed
with `using edge triggered on level triggered' warnings, because
the srm firmware initialized the interrupt logic that way (level
triggered). The changes in NetBSD 1.5 (using the com devices
edge trigged) just did not work on the Multia -- I got spurious
interrupts and the devices did not work. I do not know if there
is something special with the hardware or something strange with
the old palcode or maybe there is a bug in other parts of the
lca (21066) code.

Jürgen Weiß

-- 
Juergen Weiss	      | Universitaet Mainz, Zentrum f"ur Datenverarbeitung,
weiss@uni-mainz.de    | 55099 Mainz, Tel: 06131/39-26361, FAX: 06131/39-26407