Subject: Re: Ensoniq Audio PCI
To: Manuel Bouyer <>
From: Perry E. Metzger <>
List: current-users
Date: 07/29/2001 22:07:14
I know this is a month old thread, but...

I know a bunch about the problem in question. It impacts only some
versions of the card. It is caused by lost interrupts. I own a card,
for example, that is a victim of the problem, but others do not. This
leads people to post "hey, it works for me", even though some revs of
the card most certainly do not work properly.

I tried very hard to figure out why interrupts were being lost a long
time ago and failed.


Manuel Bouyer <> writes:

> On Sat, Jun 30, 2001 at 03:51:48PM -0300, Paulo Alexandre Pinto Pires wrote:
> > Hello, All.
> > 
> > For a few months now, I have been experiencing the following problem
> > with my Ensoniq Audio PCI card: when playing sounds, it works fine
> > for the first few times, but at some point the card seems to "hang",
> > blocking on further write()s.  When this happens, I have to kill the
> > program that is sending data, but it takes about half a minute for the
> > program to die.  If I keep trying, by starting and killing the appli-
> > cation until it succeeds, it will eventually open the device and play
> > the sound successfully, but the problem eventually happens again.
> I have the same problem when listening live audio from realplayer.
> I don't know if this is realplayer or the driver's fault.
> This seems to happen under heavy disk I/O only. Maybe it's related to
> interrupt latency, maybe it's realplayer that can't write fast enouth
> to the device and loose sync.
> --
> Manuel Bouyer <>
> --

Perry E. Metzger
NetBSD Development, Support & CDs.