The following reply was made to PR kern/33112; it has been noted by GNATS.

From: "Julian Bourne" <julian.bourne@gmail.com>
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: kern/33112
Date: Mon, 26 Feb 2007 21:27:01 -0500

 Its been a long time since I posted that bug, but if I remember correctly,
 some time later I found the problem DIDN'T exhibit itself with the whole disk
 (*[0-9]d) block devices as opposed to the generated disklabel device (*[0-9]a).
 Might be worth testing if nothing else ...
 
 On 2/26/07, Hubert Feyrer <hubert@feyrer.de> wrote:
 > The following reply was made to PR kern/33112; it has been noted by GNATS.
 >
 > From: Hubert Feyrer <hubert@feyrer.de>
 > To: gnats-bugs@netbsd.org
 > Cc:
 > Subject: Re: kern/33112
 > Date: Mon, 26 Feb 2007 22:13:48 +0100 (CET)
 >
 >  More data:
 >
 >  I've tested kernels from the following dates[1]:
 >
 >  20060915                - broken
 >  20060903                - broken (PR time)
 >  20060901                - broken
 >  20060815                - broken
 >  20060801                - broken
 >  20060715                - broken
 >  20060701                - broken
 >  20060615                - broken
 >  20060601                - broken
 >  20060515                - broken
 >  20060501                - broken
 >  20060415                - broken
 >  20060401                - broken
 >  20060315                - broken
 >  20060301                - broken
 >  20060215                - broken
 >  20060201                - broken
 >  20060115                - panic in vndthread->copystrategy
 >  20060101                - panic in vndthread->copystrategy
 >  20051215                - panic in vndthread->copystrategy
 >  20051201                - ok
 >  20051115                - ok
 >  20051101                - ok
 >  20051015                - ok
 >  20051001                - ok
 >  20050901                - ok
 >  20050801                - ok
 >  20050601                - no VND_COMPRESS (introduced in 1.115, this has 1.114)
 >
 >  I'm looking at the chanves between Feb 1 2006 and Dec 1 2005 next.
 >
 >
 >    - Hubert
 >
 >
 >  [1] I've built them all in a cycle, running basically:
 >
 >         for D in <list of dates above>
 >         do
 >                 find . -name obj.i386 | xargs rm -fr
 >                 cvs up -D${D}
 >                 sh build.sh -r tools kernel=GENERIC
 >                 cp sys/arch/i386/compile/obj.i386/GENERIC/netbsd netbsd-${D}
 >         done
 >
 >       actual tests were performed by booting the kernels in qemu, and
 >       performing the commands given in the PR.
 >
 >