On Tue, Oct 14, 2008 at 12:21 PM, Duane Ellis <[EMAIL PROTECTED]> wrote:
> Øyvind Harboe wrote:
>>
>> A target event wouldn't work terribly well here, as OpenOCD is founded
>> on the false assumption that the JTAG chain never changes throughout
>> OpenOCDs
>> lifespan.
>>
>
> Yes, but - I think - the JRC only allows insertions and deletions in a
> specific order. Not total and random re-arrangements.
>
> If that is true, then as far as the TDI/TDO chain goes - perhaps we can view
> the problem like this:
>
> Traditionally - a tap is INSERTED and thus put in bypass mode. And the
> TDI/TDO chain is +1 bit longer.
>
> What I think OpenOCD needs is another "state" - ie: "Not Inserted", ie:
> TDI/TDO is +0 or +HUGE longer.
>
> That would - I think - keep the "tap order" the same and simplify some
> things in OpenOCD.

This does sound like a solution that is no more complex than it needs
to be to handle the problem.

So a TAP controller can then be in one of three states:

- connected and not in bypass
- connected and in bypass
- not connected



-- 
Øyvind Harboe
http://www.zylin.com/zy1000.html
ARM7 ARM9 XScale Cortex
JTAG debugger and flash programmer

Free eCos workshop in Oslo October 21!
http://www.zylin.com/workshop.html
_______________________________________________
Openocd-development mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/openocd-development

Reply via email to