Mike Clarke wrote:
I'm getting frequent panics due to spin lock held too long when booting
my recently built 6.3 system with an Athlon 4850e dual core processor
on a Foxconn 6150M2MA motherboard (GeForce 6150 and nForce 430
FreeBSD curlew.lan 6.3-STABLE FreeBSD 6.3-STABLE #2: Sat Jul 12 09:43:21
BST 2008 [EMAIL PROTECTED]:/usr/obj/usr/src/sys/CURLEW i386
I'd welcome any advice on how to understand and fix this.
The panics usually happen immediately after the "Mounting local
filesystem" message, but sometimes just before the end of the boot
The system goes through the motions of producing a crash dump with a
visible count down on the screen followed with...
Automatic reboot in 15 seconds - press a key on the console to abort
cpu_reset: Stopping other CPUs
... then it just hangs until I do a hard reset.
The odd thing is that when I reboot there doesn't appear to be any crash
dump - I see the following messages:
Checking for core dump on /dev/ad4s1b...
savecore: no dumps found
Is the lack of a dump due to some config error of mine, or an effect of
the spin lock problem?
Here's the head of my kernel file ...
nooptions COMPAT_FREEBSD4 # Compatible with FreeBSD4
nooptions COMPAT_FREEBSD5 # Compatible with FreeBSD5
You should add options WITNESS also and this will probably generate
additional debugging and/or a panic before the error occurs.
But I'm still a bit dubious as to whether I've really managed to create
a debugging kernel. I was under the impression that there should be
some *.symbols files in the /boot/kernel directory but there aren't any
(or is this just a rev. 7 thing?) and the kernel directory is only 24
I also noticed when building the kernel there were lines like ...
objcopy --strip-debug kernel.debug kernel
... though I do still have /usr/obj/usr/src/sys/CURLEW/kernel.debug.
should I be doing something with this?
You can just use it post-mortem with kgdb.
email@example.com mailing list
To unsubscribe, send any mail to "[EMAIL PROTECTED]"