Stephen Takacs wrote:
>> cpu0: AMD Sempron(tm) Processor 3000+ ("AuthenticAMD" 686-class, 256KB L2
>> cache) 1.60 GHz
>
> That's interesting. How long have you been running OBSD 4.0 on that
> machine? I have the mobile version of this cpu, and my laptop started
> locking up erratically (also w/o ddb) shortly after upgrading from 3.9
> to 4.0. After about a week of that craziness and no way to troubleshoot
> further (short of running memtest86 and 'make build', neither of which
> revealed any hardware issues), I went back to 3.9, and it's been as
> rock-solid as it used to be.
>
What you are describing is almost certainly the i386-on-amd64 problem.
Solution is to do one of the following (in my order of preference, your
criteria may be different than mine, of course!) :
* run OpenBSD/amd64 (where this problem doesn't exist)
* wait for 4.1 (where it is fixed)
* run -current (where it is fixed)
"fixed" is probably not quite the right word, several people suspect
the PAE support was PROVOKING a real problem elsewhere, but backing
out the PAE support seems to have either quit provoking it or pushed
it down to unseen levels. But then, it's an amd64, if you have no
reason to run i386 code on it, I'd recommend running amd64 on it...
Nick.
