[Suzaku:00225] Re: Trouble adding OPB slave

Yasushi SHOJI email@hidden
2005年 1月 14日 (金) 11:18:55 JST


Hi Eike,

At Fri, 14 Jan 2005 11:11:24 +0900,
Eike Suthau wrote:
> 
> Hello (this time in English ;-),
> 
> I have added an instance of the Xilinx opb_ipif (version 3_01_a) to
> my design. I'm using it in posted mode to send data to my own logic (using the 
> decoded WR_CE signal and the data bus only).
> 
> Here is my problem:
> I access a port used by my own logic - the program works and I see the correct 
> behaviour from my design (it's a buffered parallel->serial converter).
> However, when I use "while (1) { ... }" to put the program into an endless 
> loop it quits after a few thousand cycles (this is no constant time).

that's a bug in EDK 6.3i.  see the following url.

to summaries: gcc in the EDK 6.3i generates code with instruction
`cmpu' for loop.  when microblaze gets an interrupt while executing
instruction `cmpu', microblaze dies.

http://lists.atmark-techno.com/pipermail/suzaku/2004-December/000185.html

> What could I be doing wrong or is there a known bug? What is the normal 
> practice for adding slaves to the OPB?

you aren't doing anything wrong, I assume.
--
             yashi



suzaku メーリングリストの案内