[Libre-soc-dev] Unexpected clock connexions.

Staf Verhaegen (FibraServi) staf at fibraservi.eu
Fri Jun 4 10:44:53 BST 2021


On 3/06/2021 23:04, Luke Kenneth Casson Leighton wrote:
> Staf, attached patch for your review, it is a simple rename:
>
> * sys_clk to sys_clk_0
> * ref_clk returns back to sys_clk

Reason I did not do this is that have to be sure to rename all 
occurrences of sys_clk internal signal otherwise you would end up with 
part still connected directly to sys_clk. With my change it was not 
needed to track down all exisinting of sys_clk so I didn't. I agree that 
number of occurences is low so your patch looks fine.

I did some verification:

* soc-cocotb-sim commit 2957b26
* I just added Verilog model for pll in bypass
* The IDCODE and scan chain test run on changed netlist
* These tests don't actually test anything driven by sys_clk.
   Luke, were you able to do DMI test with cocotb or verilator directly ?

greets,
Staf.

-- 
Chips want to be free.





More information about the Libre-soc-dev mailing list