Re: locking against myself in sched_nextlwp()

[Mindaugas Rasiukevicius <rmind%NetBSD.org@localhost>]

Michael Lorenz <macallan%netbsd.org@localhost> wrote:
> Hello,
> 
> a kernel built from yesterday's sources panics reproducably in  
> sched_nextlwp() - the lock is a spinlock, the machine is a dual G4,  
> the scheduler is M2. Before I dig deeper - does that ring a bell  
> anywhere?
> It usually happens a few minutes into build.sh -j4 distribution.
> 
> have fun
> Michael

Email did not get into the GNATS, but anyway, I think it is because of
current LWP migration mechanism:

http://mail-index.netbsd.org/netbsd-bugs/2008/05/20/msg003504.html

-- 
Best regards,
Mindaugas
www.NetBSD.org