>Date: Thu, 17 Dec 1998 13:53:55 -0800
>From: Tom Spindler <dogcow@isi.net>
>
>I've found that that my AWE64 misbehaves in rather strange ways under
>netbsd 1.3.2 - it only accepts the first write to /dev/audio and subsequent
>writes get queued but never played. A further interesting wrinkle is that
>when the program doing the writing is ^C'ed, the program continues to block
>for about 10-30 seconds, plays one final blip of sound, and then exits.
>
>Has anybody else encountered this problem?
>
>I've attached a lttle program that sends some output to the speaker, as
>well as the results on my machine.
>
>(It's a ASUS P5A, K6-2 @350mhz machine, if that helps.)

>/* This is mostly stolen from the xmame source */

Been there, wrote that :-)

>	a_info.blocksize = 0;

I'm guessing that this is a big part of your problem.  Why did you set
the blocksize to 0?

>	a_info.mode = AUMODE_PLAY | AUMODE_PLAY_ALL;

You probably don't want AUMODE_PLAY_ALL.  The sound system in xmame is
really screwy, and I found it necessary to use that flag to retain
synchronization with the graphics.

I probably can't be of much more help...I don't even have a sound card
in any of my NetBSD/i386 boxes at the moment.  But I do know that the
xmame driver did work at one time, with exactly the card you're using
(AWE64).  It probably wasn't under 1.3.2, but some version of -current
from a year or so ago.

Cheers,
entropy

--
entropy -- it's not just a good idea, it's the second law.