Re: kern/42113 (null pointer dereference in audio_calcwater if using hdaudio)

["Jared D. McNeill" <jmcneill%invisible.ca@localhost>]

The following reply was made to PR kern/42113; it has been noted by GNATS.

From: "Jared D. McNeill" <jmcneill%invisible.ca@localhost>
To: "List Mail User" <track%Plectere.com@localhost>,
        <gnats-bugs%NetBSD.org@localhost>
Cc: <jmcneill%NetBSD.org@localhost>
Subject: Re: kern/42113 (null pointer dereference in audio_calcwater if using 
hdaudio)
Date: Sun, 27 Sep 2009 07:16:43 -0400

 That message should only be printed if HDAUDIO_AFG_DEBUG is defined. The 
 reason the code runs endlessly is that it is polling for the presence of 
 headphones so it can automatically mute/unmute the built-in speakers as 
 appropriate.
 
 --------------------------------------------------
 From: "List Mail User" <track%Plectere.com@localhost>
 Sent: Sunday, September 27, 2009 7:08 AM
 To: <gnats-bugs%NetBSD.org@localhost>
 Cc: <jmcneill%NetBSD.org@localhost>
 Subject: Re: kern/42113 (null pointer dereference in audio_calcwater if 
 using hdaudio)
 
 >>...
 >>From: "Jared D. McNeill" <jmcneill%invisible.ca@localhost>
 >>...
 >>Cc: <track%Plectere.com@localhost>
 >>Subject: Re: kern/42113 (null pointer dereference in audio_calcwater if 
 >>using hdaudio)
 >>Date: Sat, 26 Sep 2009 21:55:33 -0400
 >>
 >> You're right! Your audio_calcwater panic was caused by a different 
 >> problem
 >> which should have been resolved by the following commit:
 >>
 >> http://mail-index.netbsd.org/source-changes/2009/09/25/msg001185.html
 >>
 >> Cheers,
 >> Jared
 >>...
 >
 > OK, now with up-to-date cvs (i.e. ~3AM PST today), the panic
 > is gone.  However, now there is an endless stream of
 > "hdafg0: NID 00 out of range (10-28)"
 > kernel messages.  These occur at ~2-4/second and render the machine
 > unusable (you can't see more than an occasional blue "flash" in sysinst).
 >
 > I guess the original PR can now be properly closed - the panic
 > is gone;  Time for a new PR though :-(  (though commenting out the 
 > offending
 > message is an "easy" work-around, it is also not a satisfactory "fix" for
 > having the kernel running an "endless" loop of useless code).
 >