>> UVM is unstable, but not *that* unstable.  From the time it was
>> committed to the tree it was good enough to complete 'make build', at
>> least with the new pmap code.

>Not true.   On my 64-meg Pentium, I have never gotten through a make
>build without a spontaneous reboot or some kind of weird error (e.g.,
>rm: command not found).


hmm.   that doesn't sound very fun.   yours is the first report i've
seen of that sort of problem.   i would be interested in knowing how
common it is.  


one thing you can try is switching off PMAP_NEW if you've got it on
to see if that makes any difference.


for reference, my primary development system is a P6-200 with 32MB RAM
and a 2940UW-based SCSI bus.

chuck