On Sun, Jan 08, 2012 at 05:06:08PM +0100, Franck JULLIEN wrote:
> Hi,
> 
> After make my CPU work under Modelsim, I started to take a look at the
> JTAG / debug interface when I realized I needed two JTAG on my board.
> 
> As I'm using an Altera eval board (DE1 and NEEK), I have an integrated
> JTAG interface (hardware built-in into the board).
> So I need to use my USB-Blaster plugged on GPIOs using a special cable
> in order to get OpenOCD running.
> Doesn't sound's good and I don't like it.
> 

The DE boards in my orpsocv2 tree
(http://git.openrisc.net/cgit.cgi/stefan/orpsoc/tree/boards/altera)
has support for using the advanced debug system as an option
(http://opencores.org/project,adv_debug_sys)
That allows you to use the integrated JTAG interface to debug the CPU.
It's working, but I've hade som stability issues with it in my setup
and there is currently no support in OpenOCD for it (you'll need to use
proxy provided with adv_debug_sys).

> I'm thinking of doing some UART to WISHBONE bus IP. This IP would have
> some GPIO (to control cpu reset for example) and a mechanism switching
> between the standard UART and the controlling software. The
> controlling software would be able to download files to RAM (or SPI
> FLASH, CFI,....), read/write/modify some addresses.
> 

Wouldn't that pretty much be achieved with a preloaded u-boot/barebox
and just a normal UART?

> This kind of interface would be slow but I think I worth it. It could
> be an easy way to run test program.
> 
> I know there is already a comparable IP on Opencores. However, it is
> not wishbone compliant and I don't like to reuses someone else's code
> when it's not as a blackbox.
> Moreover, there is no controlling software.
> 
> What do you think about this ?
> 

I'd like to direct valuable developer resources onto getting
advanced debug system support into OpenOCD ;)
I've been (half arsed) experimenting with using the altera vjtag together
with the old debug if (aka mohor) too with some minor success,
that's something that might be worth looking into too.
But in the end I think we want support for adv_debug_sys anyway.

Stefan
_______________________________________________
OpenRISC mailing list
[email protected]
http://lists.openrisc.net/listinfo/openrisc

Reply via email to