Subject: Re: panic in -current + sommerfeld_i386mp_1
To: None <>
From: Sean Doran <>
List: tech-smp
Date: 05/11/2001 20:20:49
Bill Sommerfeld <> writes:

> Someone who should know better committed untested code to the mp
> branch.  I've backed out the change from the branch until I have time
> to figure out the correct fix.

Well I don't mind pull-ups which lead to a build
completing but resulting in a kernel which crashes & burns.
In fact, there are days when I prefer that. :-)

OTOH, maybe the practice of adding some degree of warning
like "untested, but {might build,does build}" to the
commit message could be useful -- who knows?  That's for
you guys to figure out amongst yourselves, but everyone
who tracks the latest stuff in CVS should be prepared for
some degree of risk of everything from simple build
failure to outright disaster.

> Using -current as of ~may 8th should do the trick with the MP branch
> should do the trick. (cvs update -D)

Yup - NetBSD 1.5V (SCREAM) #0: Tue May 8 02:48:06 CEST 2001 
was very happy, and thanks to those of you who keep
integrating trunk changes into this branch.  Recent
features have made "scream" feel alot faster, so I've
been having fun building from -current + this branch
every day or two.