From alsa-user-owner@alsa.jcu.cz  Fri Jun 26 17:41:37 1998
Received: from onyx.ftech.net (onyx.ftech.net [195.200.12.17])
	by marvin.jcu.cz (8.8.8/8.8.8) with ESMTP id RAA08478
	for <alsa-user@jcu.cz>; Fri, 26 Jun 1998 17:41:13 +0200
Received: from mailgate (ruby.ftech.net [195.200.12.8])
	by onyx.ftech.net (8.8.5/8.8.5) with SMTP id QAA14198
	for <alsa-user@jcu.cz>; Fri, 26 Jun 1998 16:41:09 +0100
Message-ID: <3593BF84.145B2E46@bigfoot.com>
Date: Fri, 26 Jun 1998 16:34:28 +0100
From: Reinier Mostert <reinier.mostert@bigfoot.com>
MIME-Version: 1.0
To: alsa-user@jcu.cz
Subject: Re: SB 16 hangs my PPro when playing sound
References: <19980623142843.1307.qmail@hotmail.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Reply-To: alsa-user@alsa.jcu.cz
Sender: alsa-user-owner@alsa.jcu.cz
Precedence: list

Hi,

I tried removing the 16bit dma setting from my conf.modules, but alsa still
detected it after restarting snd driver.  I also checked /proc/dma, it gives:
1: SB 16
4: cascade
5: SB 16

Will it show if there's a DMA clash?  What specific tests can I do?

(I will try the new version of the driver tonight)

Reinier

Reinier Mostert wrote:
> 
> Hi,
> 
> I'll try you suggestions below tonight.  Some notes:
> 1) It hangs using SMP and non-SMP kernels.  I got the second CPU only a
> couple of weeks ago, and it used to hang before that.
> 
> 2) I never thought of DMA conflicts - I'll definitely check on that.  I
> am using a 16 bit DMA, and just selected the one suggested.  BTW, I
> checked last night, and the IRQ is actually 5 (and not 7, as I said).
> Still, I set the SB 16 IRQ the same as the SB 1.5.
> 
> 3) No other instabilities.  And it is just too pat, the machine will be
> on for a week, I start playing audio, and after a random time it freezes
> solid.  It's never hung while I'm not playing sound.
> 
> My RAM amount is set to mem=79.x (x < 5) MB - I had it on 80 MB, and had
> some problems.  I played around, and settled on this amount.  No
> problems after that.
> 
> So, I'll fiddle with DMA and let you know.
> 
> >1.  It's relatively easy to compile a non-SMP kernel and test this
> setup with
> >    SMP disabled and running only on one processor.  I know JK says it
> should
> >    be SMP-safe, but it won't hurt to double check.  If you still have
> problems
> >    with SMP disabled, it will rule that out as a possibility and we
> can
> >    stop worrying about it.
> >
> >    I've got a dual PPro system I can use (SuperMicro MB), but it
> currently
> >    has no sound card.  I could put a GUS in it, but I don't have an
> SB16
> >    I can use.  If I get around to testing it, I'll let you know if I
> have
> >    any problems.
> >
> >2.  One difference between an SB and an SB 16 is that the SB 16 is a 16
> bit
> >    card (duh) and therefore can use a 16 bit DMA.  Are you using a 16
> bit
> >    DMA?  Are you sure that DMA is otherwise unused?  I think a DMA
> conflict
> >    is a more likely problem than an IO base or IRQ conflict.
> >
> >    Can you run applications in 8-bit mode, using the 8-bit DMA used
> with
> >    your original card?  If the problems go away, you have a problem
> with
> >    your 16-bit DMA.
> >
> >3.  Are you experiencing any other instability at any other time?  I
> notice
> >    that you have >64 MB of RAM, which linux cannot automatically
> detect.
> >    You need a special lilo appendage to get that to work.  If that is
> not
> >    done correctly, if the ammount of memory actually installed is not
> less
> >    than or equal to the amount of usable memory in the system (taking
> into
> >    account anything used by the bios) you can have random system
> crashes
> >    when linux tries to use memory that isn't there.  (See the
> BootPromt
> >    HOWTO for more info on the mem= appendage.)
> >
> >    Try ommitting the mem= line from lilo if you've got one. (This
> should
> >    set it to use at most 64 MB of RAM.)  If the problem goes away, you
> >    have a RAM problem.  You should play with the amount of memory
> listed
> >    on the mem= line until you have as much memory as possible without
> >    having any crashes.  There's probably some better way to do this.
> >
> >That's all I can think of off the top of my head.  It shouldn't be hard
> to
> >try these things and see what happens.
> >
> >Eric
> >------
> >To unsubscribe from <alsa-user@alsa.jcu.cz> mailing list send message
> >'unsubscribe' in the body of message to
> <alsa-user-request@alsa.jcu.cz>.
> >
> >
> 
> ______________________________________________________
> Get Your Private, Free Email at http://www.hotmail.com
> ------
> To unsubscribe from <alsa-user@alsa.jcu.cz> mailing list send message
> 'unsubscribe' in the body of message to <alsa-user-request@alsa.jcu.cz>.

