Weird as* sound problem

[Remington L.]

Well disabling ACPI is an option but im waiting until 5.1, as it stands with
5.0 I go into kernel panic on disable. Unfortunetly my skill level is not up
to par with yours so I'm afraid it would be hard to pass on information. I
have tried ogle though on my machine, I am getting the same problem.

BUT I have done alittle testing. As you know I have to move my mouse
constantly in order to hear sound. I found another the thing I can do that
does it.  If I ping flood(ping -i 0 -s 4 xxx.xxx.xxx.xxx) another machine my
sound works, but only on another machine, not if I do it to myself or
loopback.

The question now is what do moving my mouse and ping flooding have in common

-----Original Message-----
From: Iain Templeton [mailto:iain.templeton at cisra.canon.com.au] 
Sent: Sunday, May 18, 2003 5:12 PM
To: Remington L.
Cc: freebsd-mobile at FreeBSD.ORG


Remington L wrote:

[ Replying to freebsd-mobile only - not sure which is the best list,
  current perhaps? ]

--Original Message--:
>FreeBSD 5.0-RELEASE-p7
>
>I have a Sony VAIO GRX570 with a YAMAHA AC-XG Audio. My sound works on in
>XMMS playing mp3s, here is the catch, the sound dies the second I stop
>moving the mouse. How the helldoes that work and how do I fix it
>
> [dmesg trimmed]

Hmm. I have the same problem with a Vaio PCG-R505TFP. Looking at your
dmesg it looks like it is basically the same motherboard chipset
(ICH3-based)
but with different processor (mine's a PIII-M 1200).

Anyway, I did a little bit of snooping and found that pcm0's interrupt
doesn't
seem to be getting as far as the interrupt handler (via judicious use of
printf()).

The symptom I have is is that xmms doesn't work at all, and ogle works for
about
one samples worth of data (probably because pcm0 and the DVD-ROM on firewire
share irq 9).

Have you tried booting the machine with ACPI disabled? All mine does is
panic
in what might be the PNPBIOS code. I expect that if/when you use Windows on
the
the machine the sound works fine?

I can also supply fairly verbose boot messages (with lotsa ACPI debug) and
extracted DSDT files if necessary (although not until tomorrow).

(Personally I blame interrupt routing, but I have no evidence at all to
back that up).

Iain