crash on kernel dump test

[manu%netbsd.org@localhost (Emmanuel Dreyfus)]

Hello

In order to track down my NFS-related crashes, I tried to setup a
properly working kernel crash dump on the machine, but my test led to
quite limited success:

While the system was working, I dropped into DDB, then:
db{0}> reboot 0x104

dumping to dev 19,4 offset 479763335
dump fatal page fault in supervisor mode
trap type 6 code 2 eip c05367c4 cs 8 eflags 10246 cr2 cfcefdc0 ilevel 8
kernel: supervisor trap page fault, code=0
Stopped in pid 0.2 (system) at  netbsd:dodumpsys+0x2f4: orb  %al,0(%edx)
db{0}> reboot
syncing disks... 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 giving up
[power cycle required]

Anyone has an idea of what goes wrong?


Moreover, at reboot time, I see that:
Checking for core dump...
savecore: no core dump
(...)
swapctl: setting dump device to /dev/ld0e

That means savecore cannot do its work automatically ar reboot. I assume
I can run it by hand once back in multiuser after a crash, since my dump
device is different from my swap device (the machine is not able to
reboot on its own, it will always require a power cycle after a crash,
so I can afford a manual savecore, unfortunately)



-- 
Emmanuel Dreyfus
http://hcpnet.free.fr/pubz
manu%netbsd.org@localhost