MPC5200B Startup

[sburck]

I'm filling in for someone who's on vacation on a PowerPC project at a
company which I did a previous PowerPC project years ago - but I haven't
used the PowerPC since, and this member of the family is new to me.

I have modified the old project's startup to work on the 5200B. The
project layout is a three stage one:

Stage 1) BSP initialization, located at physical address 0xFFF00100 on
the flash. This stage is very short, does the board bootstrap, copies
stage two from Flash to SRAM, and jumps to stage 2 in SRAM

Stage 2) Serial application loader. Hangs around on the serial port for
a while, looking for a proprietary protocal. If it comes, stays active for
ever, getting commands to erase and burn different parts of the flash (in
general, reburning stage 3, the application). If it doesn't come, it
copies stage 3 from Flash to SRAM and starts running.

Stage 3) Application code.

Now, under the JTAG I have gotten all three stages working - including the
UART and flash updating. The code is located at physical address in the
flash, and I can update them.

However, when I disconnect the JTAG, and try to boot the board on its own,
nada.

The bootstrap code in stage 1 is pretty much the bootstrap code taken from
the evaluation board, with modifications for different chip selects. I can
see a little noise coming from the UART, but nothing like what I see when
from the JTAG.

Is there anything else I need to know to get this to work?

[Didi]

sburck wrote:
> ...

> I have modified the old project's startup to work on the 5200B. The
> project layout is a three stage one:

> ...

>
> However, when I disconnect the JTAG, and try to boot the board on its own,
> nada.

You might want to pay attention to TRST, I don't remember what
sort of issues I had with it > a year ago but I remember I had to look
into it.
I think the TRST had to be actively driven low and then released, but
this is so vague I would not call it a memory.

You can also try to use the monitor found on my website
for the 5200 (not yet tested on the 5200B - soon to come, though).
If the simplest (no DDR used) version comes out of reset you will
know the issue is other than JTAG. This implies serial port 6 can be
used
by the monitor, however.

Didi

------------------------------------------------------
Dimiter Popoff Transgalactic Instruments

http://www.tgi-sci.com
------------------------------------------------------
http://www.flickr.com/photos/didi_tgi/sets/72157600228621276/

Original message: http://groups.google.com/group/comp.arch.embedded/msg/daac272325ba2037?dmode=source

[sburck]

Problem solved - it turns out that the startup code left to me was loading
MBAR from SFR, before SFR was ever initialized - and as soon as that
happened, the chip selects stopped and everything died. A few hours
setting up a poor man's logic analyser let us count chip selects to the
flash, and see two low bits of the data, with that we figured out both
where the code was running (it was reaching 0xFFF00100) and where it
stopped (loading the MBAR with junk).