OSE5.1/PPC405 memory layout constraints?

Do you have a question? Post it now! No Registration Necessary

Translate This Thread From English to

Threaded View
Hi,

I was wondering whether there are any PPC405-specific memory layout
constraints.

I am trying to get OSE5.1 (some embedded RTOS) running on a PPC405
embedded system. It went well on a dedicated evaluation board (Memec
VP20, VirtexII-Pro), but "moving" to the target system failed so far.

There is a 32MB SDRAM located at 0xFC000000 - 0xFDFFFFFF.
The OS is configured to use that memory starting at 0xFC900000.
The gap between 0xFC000000 and 0xFC900000 is used for the exception
vector area and system start up logging.

Since the original OS configuration expected the SDRAM at 0 -
0x01FFFFFF, and thus vector base is at 0, I was wondering if 0xFC000000
is still OK? I read that address has to be 64k-aligned.

Is there a default setting for the PPC405 that the exeception area has
to start at 0?

Thanks in advance

    Florian

XPost to comp.sys.powerpc.tech & comp.arch.embedded

Re: OSE5.1/PPC405 memory layout constraints?
Florian,
without knowing OSE5.1 but with the experience of having written
at least one (pretty fat) OS I would speculate that the problem is
not having memory at 0. You can relocate the vectors, sure, but
it could well be that the OS relies on some addresses in the lowest
64k for sort of "direct" access (I do it and if you deprive DPS of the
RAM at 0 just like that all you can expect is a complete mess).
 If you could somehow capture whether it attempts such accesses
this might be revealing.
 Again, this is only speculation based only on related, not direct
experience so please take it with due caution.

Dimiter

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

http://www.tgi-sci.com
------------------------------------------------------


Re: OSE5.1/PPC405 memory layout constraints?
Hi,

I also configured the OS regarding the relocated vectors (who knows
whether there is something hard-coded left).

Nevertheless I will hopefully get a new hardware design in a while, that
  matches the address map of the evaluation board.

   Florian

Re: OSE5.1/PPC405 memory layout constraints?
Florian Boelstler schreef:
Quoted text here. Click to load it

Hai,

Regarding the ppc405: this architecture comes with the Exception Vector
Prefix Register. Your bootstrap, the code that inits the board, should
set this register conform the physical memory.

Ose (5.1) has a config parameter that enables you to tell the kernel
where the exception table ought to be. In your case equal to the EVPR,
but perhaps there is no need to configure it (i.e. the kernel might just
take the EVPR setting).

Regards,

Bartjan

Site Timeline