Subject: pmap/vm problem on 68040
To: None <port-mac68k@NetBSD.ORG>
From: Erik Bertelsen <erik@sockdev.uni-c.dk>
List: port-mac68k
Date: 03/14/1997 10:49:03
I have a problem with NetBSD/mac68k running -current sources (currently
sup'ed up to 11 March 1997).

Occasionally the system will panic with a Bus error trap, typically under
medium load, e.g. during multiple compilation, but also typically while
running the nightly cron jobs.

The system is networked with a 57600 PPP link connected to a Mac IIcx
also running NetBSD.

I have seen the following stack trace at at least a couple of these
panics. As far as I can tell they might be hardware related or they
may be related to the68040 pmap interface. Any good ideas are welcome.

I have "a feeling" that this problem is new since a few weeks, but as
I have only had the machine for a couple of months, and I have been
reconfiguring it a couple of times (disk+ram), I have really never
had it running for extended periods of time -- so my "feeling" is
not necessarily correct.

_panic("Bus error trap")
_trap
_addrerr(?)
_pmap_testbit(16d0000,10,83fefc,bc10c,16d0000) + e8
_pmap_is_modified(16d0000) + c
_vm_object_page_clean(6c63f00,0,0,0,0) +d4
_vnode_pager_sync(6bb2000, f1, 6bfe500,0,83ffa81) + 96
_sys_sync(6bfc500,83ff88,83ff80)+4a
_syscall(24)+10c
_trap0()+e

(keyed in after a handwritten write down from the console,
 give or take a couple of typos)

regards
Erik Bertelsen