Subject: Re: Autosave of panic backtraces (was Re: Getting into the debugger)
To: Curt Sampson <>
From: Chris G. Demetriou <>
List: tech-kern
Date: 01/16/1999 23:56:22
Curt Sampson <> writes:

> On Sun, 17 Jan 1999, Brian C. Grayson wrote:
> >   Recently I was wondering, how hard would it be to add code to
> > panic() to print out a backtrace, and save it somewhere on disk
> > before dying?
> What's wrong with just running gdb on the kernel and core dump
> produced by savecore(8), and doing a backtrace there?

There's something to be said for having it automatically available...

However, that having been said, I think there's a better way:

	have savecore save the core
	if that's successful, automatically run a script to perform
	    some data extraction from the core file, e.g. stuff dmesg
	    and gdb stack backtrace into a file, plus maybe 'ps'
	    output, and other useful stuff.

I seem to recall that Digital UNIX (formerly DEC OSF/1!!!!!!) does

Chris Demetriou - -
Disclaimer: Not speaking for NetBSD, just expressing my own opinion.