Jonas Horberg wrote:
>> The attached patch add the khz and speed_div functions to
>> the parport interface driver.
>>
>> This enables the jtag_khz command for parport interfaces.
>>
>> The patch also adds a command named parport_toggling_time
>> that tells the parport driver how long (in nanoseconds) it
>> takes for the hardware to toggle RTCK. The khz and speed_div
>> functions needs this data. The default value is set to
>> 1000 ns, perhaps it exist a better typical value. My PC
>> have a 1.37 us toggle time.
>>     
>
> Comments?
>
> Configuration scripts with jtag_khz or jtag_rclk fails if these
> commands are unsupported.
> jtag_speed works just as before.
>   
It seems we do need this, however, the implementation seems wrong.

First, I do not see how RTCK comes into play - AFAIK, the parport
interface does not support RTCK, so only TCK frequency should be
relevant for jtag_khz.

Second, using arbitrary timing values that are measured on one system
seems doubly wrong. If we need a specified clock, we need to make sure
that the code really produces that clock speed (at most), so the could
should auto-adapt to the port timing on the machine it runs on (note
that we also do not want overly pessimistic timings - clocking slower
than requested degrades performance).

cu
Michael

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

Reply via email to