Re: config(5) break down

["Michael L. Hitch" <mhitch%lightning.msu.montana.edu@localhost>]

On Tue, 27 Apr 2010, Michael L. Hitch wrote:

> It kind of looks like possibly a problem in bt_insbusy() called from 
> vmem_xalloc() when memory gets tight.  I haven't had much time to try
> to debug this yet (the RaidFrame parity really hurts here).  I also
> tried a NetBSD-4.0_STABLE kernel, with the same results.  I noticed there 
> were a number of changes in subr_vmem.c between my 4.0_BETA and 4.0_STABLE.

  An update....

  I copied this system to a 5000/260 with 180M of memory, and it was able
to boot up to multiuser with the NetBSD-4.0_STABLE.  A 5.0.1 kernel, and 
an older NetBSD-5.0_STABLE kernel both hang when clamd is starting.  and
the only thing I've been able to determine so far is that it's in the trap 
handling when I halt it.  I'd guess it's looping with memory exceptions at 
a high IPL.  If I disable the clamd startup, then 5.0_STABLE boots fine
and seems to run OK.

--
Michael L. Hitch                        mhitch%montana.edu@localhost
Computer Consultant
Information Technology Center
Montana State University        Bozeman, MT     USA