Øyvind Harboe wrote:
>> I wonder if we can "genericize" jtag_add_runtest(xruntest, xendir) to
>> not be so tied to TAP_IDLE.
>>     
>
> If this should be added to OpenOCD, then the above looks like a good
> approach. The bitbang.c driver is very simple and should be a good
> point place to put a reference implementation.
>
> There is the issue with backwards compatibility. This is only required
> for FPGA programming, today, so it should be possible to implement
> only runtest but to delete that fn once the more general version
> has been implemented in the driver.
>
> The need for this behaviour smacks of perversion of the JTAG protocol,
> except that I've heard that the JTAG protocol is designed to be able
> to support a continiously running clock I've heard...
>   

OK, I can use a small hand crafted SVF file to generate the stream I 
need for the XSVF player to call the functions needing to be tested.  
Bitbang is where I'll start and use the new dummy state tracker to 
verify the looping and state transitions.


My cable is the OLIMEX Usb (ARM-USB-OCD), so by Tuesday I'll be testing 
that driver too.  I get more boards then.

After thinking about it, we will probably have to implement a new driver 
handler in the function code switch(), then later come back and make 
jtag_add_runtest() be a macro that calls the new api after all the 
drivers support it.  Or do the global search and replace then.  Until 
then they will both have to exist for backwards compatibility.

Using existing jtag_add_runtest() as a starting point, I think it should 
be pretty straight forward. 

I cannot foresee the thickness of the forest right now, brain is tired.

Dick


_______________________________________________
Openocd-development mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/openocd-development

Reply via email to