If the kernel attempts to execute some code in an invalid area of memory
(in my case it was at 0x20303020) then the Oops reporting code starts
and correctly displays the EIP, registers and stack. It then tries to
display the code at that location and causes its own memory access
violation, which results in another Oops dump referring to the Oops
code itself!
I don't know how the memory-related parts of the kernel work so I'm not
in a position to fix this, but at least I can bring the bug to your
attention.
Nick.
-- Kralizec Dialup Internet System Data: +61-2-837-1183, 837-1868 Zeta Microcomputer Software Fax: +61-2-837-3753 Voice: 837-1397 P.O. Box 177, Riverstone NSW 2765 http://www.kralizec.net.au/