Hi Jan

Thank you for the reply. I appreciate it.

> The reason for the (perhaps redundant) NOP operation might have been be to 
> collect the status of the preceding DMI operation - op field of the "dmi" 
> JTAG register.

I agree this is used to retrieve the status of the previous DMI operation if 
the NOP function is used in scan_field->out_value. For JTAG driver developer 
like me, we push scan_field->out_value into the JTAG. Writing op field of zero, 
telling the device to not read/write the address indicated, is needed.

The JTAG driver will replace scan_field->in_value with what is read from JTAG. 
Normally we just populate it with zeros, but since it is decided to populate 
with NOP I am presuming it there is a reason for doing so. May be to supply an 
indicator that the field was not written to yet. While it is not important for 
my purpose of “load <image>”, some other callers might be relying on this.

>That is however just a guess, and as you noted, the data is not used anywhere.

Thank you for the confirmation.

> If you'd like to improve the code, you are welcome to post a patch to 
> https://github.com/riscv/riscv-openocd
It will later be propagated here.

Thank you for this information. It saved me another mailing list post to ask 
whether should I to go riscv-openocd or here.

Many thanks and have a good weekend
Cinly



Reply via email to