[Libre-soc-dev] daily kan-ban update 18nov2021

[Tobias Platen]

today: found another bug in the PortInterfaceBase, adrok_l does not get
reset in case of the alignment error.

On Thu, 2021-11-18 at 18:28 +0000, lkcl wrote:
> today, i babbled at the EU Bureaucrats and amazingly they listened,
> they are incredibly well trained and practiced at summarising and
> absorbing input and feedback, it was deeply impressive to witness.
> as it started at 8:30 am however i spent the entire afternoon asleep
> :)
> 
> i did a little bit on the core, to try to allow the TestIssuer FSM to
> have
> overlapping instructions: the idea of issuing a DIV followed by an
> ADD
> is working well, the two instructions do in fact create a Dependency
> Hazard and the wrong result is produced (which is, ironically, a good
> sign)
> 
> in core.py, i will have to add a "capture" system of the incoming
> decoded
> instruction, because unfortunately it is necessary to acknowledge
> the "acceptance" of the current instruction before the previous
> pipeline stages (fetch, decode) can be permitted to continue.
> i.e. use an internal ispec() temp copy of the incoming data.
> 
> l.
> 
> _______________________________________________
> Libre-soc-dev mailing list
> Libre-soc-dev at lists.libre-soc.org
> http://lists.libre-soc.org/mailman/listinfo/libre-soc-dev