>Number:         23605
>Category:       port-macppc
>Synopsis:       macppc kernel crashes on postgres db operations
>Confidential:   no
>Severity:       critical
>Priority:       medium
>Responsible:    port-macppc-maintainer
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sun Nov 30 22:22:00 UTC 2003
>Closed-Date:
>Last-Modified:
>Originator:     Martin J. Laubach
>Release:        NetBSD 1.6ZC
>Organization:
>Environment:
System: NetBSD celery.emsi.priv.at 1.6ZC NetBSD 1.6ZC (CELERY) #0: Fri Sep 26 23:03:31 CEST 2003 mjl@celery.emsi.priv.at:/home/users/mjl/netbsd/src/sys/arch/macppc/compile/CELERY macppc
Architecture: powerpc
Machine: macppc
>Description:

  This is a problem that has plagued me for quite some time but I
usually run X so could never get any useful data. This time, without
X, I got a backtrace.

  The problem is rather easily reproduced by running disk i/o intensive
operations in postgres, like creating or dropping a database, vacuuming
or the like.

  Note that an older kernel (1.6D) does work reliably even under stress.

>How-To-Repeat:


% dropdb gruenshop

trap: kernel read DSI trap @ 0x2560812c by 0x22fa48 (DSISR 0x40000000, err=14)
Press a key to panic.
panic: trap
Stopped in pid 1012.1 (postgres) at netbsd:cpu_Debugger+0x10: lwz r0, r1, 0x14
db> bt
0xd59d6bb0: at panic+18c
0xd59d6c70: at trap+bc8
0xd59d6cd0: kernel DSI read trap @ 0x2560812c by pmap_pvo_find_va+48: ssr1=0x1032
	r1=0xd59d6d90 cr=0x4002f032 xer=0 ctr=0x1af54c dsisr=0x40000000
0xd59d6d90: at ADBDevTable+3adefac
0xd59d6db0: at uvm_faul+4e8
0xd59d6ee0: at trap+684
0xd59d6f40: user ISI trap by 0x195cda4: ssr1=0x4000f032
	r1=0xffffde80 cr=44002044 xer=0x20000000 ctr=0


>Fix:
>Release-Note:
>Audit-Trail:
>Unformatted: