On Tue, Oct 19, 2010 at 9:03 AM, Xiaofan Chen <[email protected]> wrote:
> On Tue, Oct 19, 2010 at 2:38 PM, Peter Stuge <[email protected]> wrote:
>>> For the OpenOCD, we know there are some optimizations do be done on
>>> the software to accelerate the USB JTAG cable.
>>
>> The bottleneck is stupid adapters.

I don't know that "stupid" is derogatory when we talk about
hardware. KISS often does it...

> People can actually get decent result from the so-called stupid
> adapters. For examples, Urjtag project is now mainly using libusb-1.0
> and libftdi-1.0 instead of the old libusb-0.1 and libftdi-0.1x to get
> much better result because of the async API support. Maybe
> as the first step, OpenOCD can be modified to make use of the
> async API provided by libusb-1.0 and libftdi-1.0.

OpenOCD will create an entire read/write sequence before sending
*anything* to the USB adapter. This would have been a bad thing if
PC's were not infinitely fast.

One interesting modification to OpenOCD might be to add some profiling
metrics on inter-jtag_execute_queue() times.

If there are large inter-jtag_execute_queue() times then I'd expect an
async interface to help.

OpenOCD already has a very nice mechanism to create long read/write
chains, and it might have a lot of the potential benefit of an async API
rolled into that optimization already.

-- 
Øyvind Harboe
US toll free 1-866-980-3434 / International +47 51 63 25 00
http://www.zylin.com/zy1000.html
ARM7 ARM9 ARM11 XScale Cortex
JTAG debugger and flash programmer
_______________________________________________
Openocd-development mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/openocd-development

Reply via email to