03:43 <cr1901> I don't think this is great, but I really _really_ wanted to get this out of my system and try it... I made a wrapper over boneless' assembler to automatically handle making an FSM: https://github.com/cr1901/rc-adc/blob/fsm-dsl/src/rc_adc/daq/driver.py#L145-L214

10:15 <Chips4MakersakaS> <whitequark[cis]> "and it could be prototyped..." <- I had to subclass the XilinxPlatform class to be sure the PS clock pin gets the right annotations in the output files for Vivado to be able to do it's thing.

12:53 <whitequark[cis]> yep, I'd expect something like that would be needed

12:53 <whitequark[cis]> also, I think resource assignment gets kind of weird with PS

14:35 <Chips4MakersakaS> My idea was that the class would define resources like the memory interface etc. and you could `request` them as other resources defined in amaranth-soc in your code. The class would then take care of instantiating the PS. I only have defined the PS clk pin for now and nothing else yet...

14:35 <Chips4MakersakaS> And I will be focusing on other things for the moment.

14:35 <Chips4MakersakaS> s/soc/boards/

14:37 <whitequark[cis]> yeah that would be my idea too, for when it's integrated upstream

14:37 <whitequark[cis]> unfortunately i haven't really used SoC FPGAs much

14:44 <galibert[m]> You need a de10-nano?

14:51 <whitequark[cis]> i need more free time when i'm not in pain

14:51 <whitequark[cis]> i have enough boards

15:15 <Chips4MakersakaS> <eigenform[m]> "on the topic of zynq parts..." <- eigenform: Do you plan to use Amaranth with Zynq in near future ?

15:29 <tpw_rules> i still need to upstream my de10 nano stuff

15:30 <tpw_rules> (and migrate to streams) but i think it's not a bad pattern and would at least like to bring it to the discussion table. there's some weird junk with internal logic i remember not being sure how to solve nicely

15:30 <tpw_rules> but it works very well

16:06 <galibert[m]> Catherine: that's... harder to provide, sadly

18:42 <eigenform[m]> not for anything practical, but i do have one of those digilent zybo boards

18:43 <eigenform[m]> i'll probably try to play around with it sometime soon