common_cells treewide: Replace non-resettable FFs

Following https://github.com/pulp-platform/snitch_cluster/pull/154, there are still some non-resettable FFs in the Snitch cluster, that can be traced back to the isochronous_spill_register IP.

This PR is a placeholder for addressing this issue. The current solution has to be revised to ensure safety against metastability conditions.

Nov 05 '24 15:11 colluca

@colluca any updates on this?

Feb 04 '25 04:02 niwis

Hi @niwis, haven't had the time to look into this. I'm also not an expert in CDCs, reset synchronization, etc. so if anyone could provide input on this it would be helpful, as properly handling the resets of the two domains is not trivial.

Feb 04 '25 09:02 colluca

@niwis After double checking this, it does look indeed safe to me. The FIFO registers which were already clocked by the source clock should also simply be reset by the source reset.

Jul 16 '25 08:07 colluca

This looks fine to me as the STA engine should properly time all paths anyway. We don't change or lose the described timing behavior so its also not an interface change to me.

Sep 18 '25 16:09 phsauter