Jason Thorpe wrote:
> 
>  > Hmm, just built a new -current kernel, and under it was
>  > rebuilding the bootblocks, when I hit a trap in
>  > pmap_enter_ptpage.  It looked like a dereference of a null
>  > pointer.  Unfortunately, I had NFS filesystems mounted,
>  > so I had to hard-reset to get it to reboot.
>  >
>  > Useful numbers :
>  >
>  > fault address 0xc
>  > fault pc      pmap_enter_ptpage+0x776
>  >
>  > (should I be filing this as a PR? port-m68k?)
> 
> Please file a bug report for port-hp300.  Also, if you built the kernel
> with full debugging symbols (i.e. "makeoptions DEBUG="-g""), I may
> have you do a few things... (did you get a crash dump?)

How do you force a dump from within DDB?  I couldn't get a
copy of the trap message from the message buffer because I
had to hard reset to get it to reboot, so I just noted the
FA and PC.
 
> Hmm... Taking a quick look, pmap_enter_ptpage+0x776 is in pmap_bootstrap().
> I'm not sure how that could have happened... I assume you got that info
> from DDB?  A backtrace and the _exact_ info printed by the trap handler
> would be good...

No, I looked it up using "nm netbsd | sort | less"  8)  Obviously a
more detailed report would be better (why I didn't file a PR 
straight off).

Can you run gdb-remote at DDB on the hp300?

Another point; I've noticed that NetBSD seems to have the same
problem that FreeBSD has wrt. text corruption in sticky text;
if an executable's memory image is corrupted for some reason
(VM error, memory error, pagein error etc.) and the executable
is killed, the stickiness of its text remains, so that if you
try to run it again, it will die of the same cause.

I've talked this over with John Dyson in the past, and I realise
it's not something simple to deal with, but I observed it 
last night, possibly in conjunction with another disk dying, and
thought it was worth mentioning.
 
> Jason R. Thorpe                                       thorpej@nas.nasa.gov

--
Mike Smith  *BSD hack  Unix hardware collector
The question "why are the fundamental laws of nature mathematical"
invites the trivial response "because we define as fundamental those
laws which are mathematical".  Paul Davies, _The_Mind_of_God_