Re: Oops in 2.1.104p1 + Trace

Keith Owens (kaos@ocs.com.au)
Sun, 31 May 1998 10:50:16 +1000


On Fri, 29 May 1998 01:16:23 -0400 (EDT),
Jordan Mendelson <jordy@wserv.com> wrote:
>It took me about an hour to write down this Oops message, and I'm sure I
>still didn't get it right. You'd think someone would put an option in the
>kernel to automatically attempt to do a call trace in something I can
>actually read.

ftp://ftp.ocs.com.au/pub/patch-ikd-2.1.103.gz. Integrated Kernel
Debugging patch including SysRQ-loGs to dump the logs to all consoles,
including serial consoles.

The problem with dumping logs automatically is where do you dump it?
If your file system code is impacted, you cannot dump to disk.
Interrupt lockup is also a problem after an oops. Serial consoles are
polling only and tend to work after all else fails. Of course you need
to compile the kernel with a serial console and run a wire to another
machine to capture the log.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu