From alsa-user-owner@alsa.jcu.cz  Sun Jul 12 18:44:59 1998
Received: from pump2.york.ac.uk (pump2.york.ac.uk [144.32.128.12])
	by marvin.jcu.cz (8.8.8/8.8.8) with ESMTP id SAA27838
	for <alsa-user@alsa.jcu.cz>; Sun, 12 Jul 1998 18:44:13 +0200
Received: from tower.york.ac.uk (tower.york.ac.uk [144.32.128.6])
	by pump2.york.ac.uk (8.8.7/8.8.7) with SMTP id RAA27154
	for <alsa-user@alsa.jcu.cz>; Sun, 12 Jul 1998 17:44:13 +0100 (BST)
Date: Sun, 12 Jul 1998 17:44:13 +0100 (BST)
From: Stephen Oxley <so101@york.ac.uk>
To: alsa-user@alsa.jcu.cz
Subject: Re: SB16 ViBRA16X PnP partial solution -addendum-
Message-ID: <Pine.SGI.3.95L.980712172433.22231B-100000@tower.york.ac.uk>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Reply-To: alsa-user@alsa.jcu.cz
Sender: alsa-user-owner@alsa.jcu.cz
Precedence: list

Okay perhaps I should have tested my solution thourougly before posting
it..... 

It turned out that with the configuration I specified my computer would
completely hang after anywhere between 10 mins and 2 hours of sound play. 
I mean a complete *freeze* up and nothing in the system logs.....very
similar to what would happen with both the oss's. I tried changing from
2.0.33 to 2.0.32 a solution which I had seen in a news group some time
back, but this didn't cure the problem. 

Maybe I'm being a bit premature in posting this revised solution but after
20 hours or so of music play and no hangs I'm almost confident it's not
going to happen (touch wood). 

It seems that only the first dma on the vibra16x can emulate 16 bit wide
sound (dma 0)  , and for some reason or other when channel 0 is used for
this on the pc it is liable to hang. So on the pc dma channel 1 or 3
should be used instead (I'm using 1). ie for dma 3 on the pc: 

/etc/isapnp.conf:
(DMA 0 (CHANNEL 3))
(DMA 1 (CHANNEL 0))

/etc/conf.modules
options snd-sb16 snd_port=0x220 snd_mpu_port=0x330 snd_irq=7 snd_dma8=9 
snd_dma16=3

Does anyone know the source of these problems?

I am still using kernel 2.0.32 and should really have tested it with
2.0.33 but wanted to send this addendum out as soon as possible. 

Stephen Oxley.


