[ECOS] eCos PowerPC Startup

[ECOS] eCos PowerPC Startup

[amassa]

Since I have been having so much trouble getting ecos running on my MBX
board I have stepped as far back to the beginning as possible.  I have been
stepping through the startup code (in vectors.s) on my MBX board because it
seems to be hanging each time.  I am using the EPPC-bug firmware to step
through each assembly line.

When I compile the code I see that the srec file is generated with an offset
of 0x10000 (if no exception vectors are copied) or 0x10100 (when exception
vectors are present).

I had to manually define CYGSEM_HAL_POWERPC_COPY_VECTORS in vectors.s
because it would not copy the vectors when I have it configured that way in
my ecos configuration.

I checked and found the correct assembly code present at the specified
address.

I ran the code in two different situations, the first was without any
exception vectors copied.  I stepped through the code and the board hung as
soon as I tried to branch from vectors.s in the _start routine to mbx.s in
the hal_hardware_init routine.

The next time I ran I didn't get very far at all.  The IP was at 0x10100
which is the reset vector.  I step through the vector which basically just
jumps to _start at address 0x12000.  As soon as I tried to step to the
branch instructions (bclr), the board hangs.

Any insight into what might be going on would be greatly appreciated.

Thanks,
Anthony

[ECOS] Re: eCos PowerPC Startup

[Jesper Skov]

>>>>> "amassa@cts" == amassa@cts com <amassa@cts.com> writes:

amassa@cts> I had to manually define CYGSEM_HAL_POWERPC_COPY_VECTORS
amassa@cts> in vectors.s because it would not copy the vectors when I
amassa@cts> have it configured that way in my ecos configuration.

Then you probably _didn't_ change CYGSEM_HAL_USE_ROM_MONITOR - when
that is set, eCos relies on the ROM monitor to provide such things as
exception vectors.

Jesper

[ECOS] Re: eCos PowerPC Startup

[amassa]

I'm pretty sure I had that set properly.  However, I rebuilt and have an
image which runs through the code that copies the vectors.

Do you have any other insight with the other parts to my question - about
the assembly problems?

Thanks.

[ECOS] Re: eCos PowerPC Startup

[Jesper Skov]

>>>>> "amassa@cts" == amassa@cts com <amassa@cts.com> writes:

amassa@cts> I'm pretty sure I had that set properly.  However, I
amassa@cts> rebuilt and have an image which runs through the code that
amassa@cts> copies the vectors.

amassa@cts> Do you have any other insight with the other parts to my
amassa@cts> question - about the assembly problems?

No.

FWIW I've been updating stubs on a MBX board today - 50MHz version,
and there were no problems whatsoever.

Jesper