Re: Variable locations in a loaded kernel (bsd.gdb)

[Joshua Lamorie <jpl at xiphos dot ca>]

On Mon, 27 Aug 2001, Eli Zaretskii wrote:
> Does the command "info address" help?

(gdb) info address scpshc_decompress::tp
No symbol "scpshc_decompress::tp" in current context.
(gdb) 

My problem is that I would like to know what address is being used to
refer to the variable.  Or, if it is in the register, which register.  Or
perhaps a stack offset.. though of course, that is context specific.  At
least, I want to see where it is first referenced, then I can trace it
through the assembler code.

Any hints are appreciated, and I'm playing around with all manner of ways
of viewing.  I can't seem to get a kernel to dump, and when I run 'target
kcore /dev/mem' I get...

(gdb) target kcore /dev/mem
can not access 0xc95e5bee, kvm_read: Bad address
can not access 0xc95e5bee, kvm_read: Bad address
can not access 0xe9eeda89, kvm_read: Bad address
can not access 0xe9eeda89, kvm_read: Bad address
cannot read pcb at 0xe9eeda89

I'll have to start firing off messages to the OpenBSD lists to figure out
more info for using 'ddb'.  I'm amazed, I've just loaded a 26Meg kernel!!

... and it runs! :-)

thanks in advance


-- 
Joshua Lamorie                           ph. 514-848-9640
Designer, Space Systems                  fx. 514-848-9644
Xiphos Technologies Inc.                 web http://www.xiphos.ca/