On Sun, 3 Sep 2006, Valeriy E. Ushakov wrote:
> On Sun, Sep 03, 2006 at 16:45:27 +0100, Stephen Borrill wrote:
>> I'm running an NFS 3.1_RC1 userland with a GENERIC 4.99.1 kernel on my
>> Jornada 690. I've 128MB of NFS swap configured and I've been compiling up
>> (some of) pkgsrc. My network card is a Linksys WPC11 v3 (wi0). At
>> seemingly random times (sometimes after 24 hours of uptime, but sometimes
>> only an hour), I've seen the machine apparently lock up (logged in over
>> ssh), but the screen's been blank when I've got to the machine itself and
>> I've hit the reset button. Today however, I caught it at a panic:
>>
>> fatal address error (load) in kernel mode
>>  spc 0 ssr 400001f0
>> panic: general_exception
>> kloader: kernel file name: /netbsd
>> panic: tlb_exception: no copyin/out fault handler (page not found)
>> expevt=c0 va=c0077000 ssr=400001f1 spc=8c0766d6 lwp=0x8cf2ba4c onfault=0x0
>>
>> Would I have more joy with a kernel that wasn't -current? Am I just
>> hitting a lack of memory?
>
> Hmm, EXPEVT=C0 is protection violation on store.  Can you do in DDB:
>
> db> bt                  # backtrace
> db> x/i 8c0766d6        # spc
> db> show reg

From GENERIC:

#options        DDB             # in-kernel debugger
#options        DDB_ONPANIC=1   # see also sysctl(8): `ddb.onpanic'

So, no, not at the moment :-/

I'll compile up a kernel with DDB and try again.

Thanks,

-- 
Stehen