sb0 changed the topic of #m-labs to: ARTIQ, Migen, MiSoC, Mixxeo & other M-Labs projects :: fka #milkymist :: Logs http://irclog.whitequark.org/m-labs
<rjo>
sb0: the command can be on a couple of things. i'll solve it with a bit of xdc magic because it gives a nice example of doing other things as well with those user-defined attributes.
<sb0>
rjo, for SPI I'd have used the wishbone CSR generator and the misoc core to create the bus for the RTIO PHY
<sb0>
now there are two different interfaces and code duplication...
<sb0>
or even without the wishbone CSR generator - have your own SPI-core-specific CSR handler that packs the config CSRs into one single address
<sb0>
why did you want to avoid the multiple config addresses that you'd get when using the misoc wishbone csr generator?
<rjo>
sb0: splitting the registers on RTIO would have been too slow and inefficient.
<rjo>
sb0: doing an RTIO event takes orders of magnitude longer than doing a non-RTIO bus access.
<rjo>
i see the convenience of being able to use the autogenerated CSR stuff. but i don't like the speed penalty. an extension to AutoCSR that packs multiple CSRs into one address would be nice.
_whitelogger has joined #m-labs
<rjo>
sb0: by the way, ASYNC_REG never got applied in the vivado AsyncResetSynchronizer.
<rjo>
the wire gets the property, but the actual FFs don't.
rohitksingh has joined #m-labs
rohitksingh has quit [Ping timeout: 255 seconds]
rohitksingh has joined #m-labs
<sb0>
rjo, does it matter for those settings registers?
<rjo>
you tend to switch between devices and different xfer sizes/styles frequently.
<rjo>
ultimately the idea of having profiles here would have been nice as well. then you don't have to manage the config/xfer switching in software at all.