Subject: Re: MIPS32 24Kc port
To: Izumi Tsutsui <tsutsui@ceres.dti.ne.jp>
From: Varadarajan Narayanan <1978nv@gmail.com>
List: port-mips
Date: 07/25/2007 19:55:39
Hello Izumi Tsutsui

> > TLB out of universe: ksp 0xc6681fc0 epc 0x801b93e4 vaddr 0xc0001000
>  :
> > I guess, the stack access has caused the issue.
>
> What does "trace" command show?

trace doesn't show much. It just shows the panic function.
mipsX_subr.S changes $sp to 'start', hence the debugger
is not able to backtrack.

Is there a way to figure what the original stack pointer is?

> FYI, I saw the similar errors on kernel stack overflow
> caused by unhandled (or too high rate) interrupts:
> http://mail-index.netbsd.org/port-cobalt/2006/04/23/0000.html

Not sure if it is stack overflow. I tracked the number of interrupts
and it was just 1. The only (enabled) device on the system
operates in a semi-poll-intr mode, hence not a heavy intr load...

Thanks very much for your inputs.

-Varada