Re: [ECOS] Help Setting up Abatron bdiGDB to Debug RedBoot

[Todd Smith <ecos at footmandog dot com>]

Herbold, Steve wrote:

> I have been banging my head trying to get the Abatron bdiGDB JTAG debugger
> to help me step through the RedBoot initialization code.  I have tried
> following the recommendations from Abatron to get their VTAB entries set up
> properly, but almost always end up with "TARGET: communication with debug
> handler failed" after stepping past some points.
>
>  
...

> I think some of the problems are that in the initialization memory is being
> swapped around (SDRAM to low memory), and MMU and caches being turned on and
> flushed. I have tried to set breakpoints past these points in the hope that
> funny behavior is just artifact.
I have used the Abatron to debug Redboot, but it was on an MPC8250, and 
it was long enough ago that I dont remember the details. But this may be 
of some help anyway. The gist of getting it to work for me was that I 
had to do two things:

1) Change the code word in the Abatron config to be "dumb" mode. The 
code word on an MPC8250 is part of the boot sequence for the processor. 
It gets loaded by the Abatron as part of the first fetches the processor 
makes. The code word tells the processor, among other things, where RAM 
is located. Anyway, I wanted to tell the processor to run out of RAM.

2) Build REDBOOT as RAM only. I don't believe that redboot for the 
MPC8xxx variants actually enabled the MMU, so there was nothing else to 
configure.

Hope that helps even though my memory is foggy and you have a different 
processor. Maybe it will trigger something in your head.

--Todd


--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss