Subject: Re: INITIALLY_LEVEL_TRIGGERED() lossage (DANGER WILL ROBINSON!)
To: Charles M. Hannum <firstname.lastname@example.org>
From: Juergen Weiss <weiss@Uni-Mainz.DE>
Date: 12/19/2000 11:33:57
Charles M. Hannum writes:
> So, anyone who was around when I tried to boot NetBSD on a PWS500
> earlier this year would have recalled that this is the code that broke
> it -- it caused the machine to panic at boot time. Coupled with the
> fact that resetting the state at shutdown time fitted the original
> reason for the code, this is why I *explicitly removed it*.
> If we're going to work around Multia `brain damage' (and the stray
> interrupts are not really braindamage, except inasmuchas the kernel
> shouldn't be bitching about them), could we *PLEASE* do it in a way
> that's not going to screw other people over?
> If someone would like to provide a *working* Alpha Multia, I could be
> convinced to figure out the real origin of the stray interrupts and
> fix it.
Sorry, that the fixes for the multia from 1.4 broke the PWS500
port. The serial line interrupts (3 and 4) have to be level
triggered on the multia. If they are edge triggered, you get
stray interrupts and the serial lines are not usable any more.
Any ideas why this happens? Level triggered is active low, edge
triggers on high to low or low to high?
Juergen Weiss | Universitaet Mainz, Zentrum f"ur Datenverarbeitung,
email@example.com | 55099 Mainz, Tel: 06131/39-26361, FAX: 06131/39-26407