I'm having the same problem.  Did you ever find a solution?

>Hello,
>I have a problem host booting my SHARC 21161 Multiprocessor system. I
get
>System hangs due to a HBG# hang each time I download the loader file
>generated with the Visual DSP environment. The same code just runs fine
out
>of an EEPROM (I ensure to use a different bootloader for those two
tasks).
>
>The ADSP 21161 Hardware Reference states that you can get those hangs
>writing to a non existent IOP register. (HR ed3: p. 7-55) So far I used
>0x04 wich is directly assoiated with the EPB0.
>
>As I said, I use a multiprocessor system. I wonder wheater I can also
use
>the muliprocessor memory space to boot my processors.
>Is writing to 0x10 0004 equal with writing to 0x00 0004 using CS001# in
both
>cases? May a HBG# hang result from such a write? 
>
>What other reasons could cause those hangs? It is really difficult to
debug
>as the hangs are not deterministic.
>
>Thank you for any hints.
>
>-- 
>Viele Gr��e
>Philipp Drewes
>
>Reply by mail sticking "philipp.drewes" in front of "tu-harburg.de".
>
>




		
This message was sent using the Comp.DSP web interface on
www.DSPRelated.com

Hello,
I have a problem host booting my SHARC 21161 Multiprocessor system. I get
System hangs due to a HBG# hang each time I download the loader file
generated with the Visual DSP environment. The same code just runs fine out
of an EEPROM (I ensure to use a different bootloader for those two tasks).

The ADSP 21161 Hardware Reference states that you can get those hangs
writing to a non existent IOP register. (HR ed3: p. 7-55) So far I used
0x04 wich is directly assoiated with the EPB0.

As I said, I use a multiprocessor system. I wonder wheater I can also use
the muliprocessor memory space to boot my processors.
Is writing to 0x10 0004 equal with writing to 0x00 0004 using CS001# in both
cases? May a HBG# hang result from such a write? 

What other reasons could cause those hangs? It is really difficult to debug
as the hangs are not deterministic.

Thank you for any hints.

-- 
Viele Gr��e
Philipp Drewes

Reply by mail sticking "philipp.drewes" in front of "tu-harburg.de".