[Libre-soc-dev] Unexpected clock connexions.

[Jean-Paul Chaput]

On Sat, 2021-06-05 at 13:17 +0100, Luke Kenneth Casson Leighton wrote:
> i've added the dummy (fake) PLL to experiments10_verilog using nsxlib with
> the (fake) PLL cell
> https://git.libre-soc.org/?p=soclayout.git;a=commitdiff;h=a876c5dd61a53230de74b2ca2900e0e09490da95
> 
> also worked out how to fix the nsxlib error "jtag_tck_from_pad" is not a
> clock tree error
> https://git.libre-soc.org/?p=soclayout.git;a=commitdiff;h=c24e13b1078fa2d358d6b7f91cfd96905c100769
> 
> i am however experimenting tracking down the right H-Tree clock for use in
> ls180Conf, it's probably
> 
>     ls180Conf.useHTree('por_clk)
> 
> not
> 
>     ls180Conf.useHTree('core.por_clk')

  Nono. That's correct. It is really 'core.por_clk'.

  Explanation:
    When building a complete chip *from* a core, say "ls180",
    the core2chip procedure creates two levels of additional
    hierarchy :

      1. corona, which instantiate the core model, and it's
         instance name will be 'core'.

      2. The chip level, which instantiate *only* the corona
         and all the I/O pads.

    However, the P&R is done not at ls180 model level, but
    at corona model level. So, as "por_clk" is an *internal*
    signal of ls180 (comes out from the PLL) and have no
    "representative" at corona level, we must reach it through
    the path : 'core.por_clk' (path rooted in corona model,
    one instance 'core', which is the ls180).
 

> also, is it likely that "coresync_rst" also has to be added?

  Depends. It is in the "middle range". Many sinks, but maybe
  not that much to need an H-Tree. If it needs a low skew, maybe.

-- 

      .-.     J e a n - P a u l   C h a p u t  /  Administrateur Systeme
      /v\     Jean-Paul.Chaput at lip6.fr
    /(___)\   work: (33) 01.44.27.53.99              
     ^^ ^^    cell:      06.66.25.35.55   home: 09.65.29.83.38

    U P M C   Universite Pierre & Marie Curie
    L I P 6   Laboratoire d'Informatique de Paris VI
    S o C     System On Chip