On Wed, Mar 15, 2017 at 1:07 PM, Sébastien Bourdeauducq via ARTIQ
wrote:
> Hi,
>
> RTIOCollision is a bit tedious to implement with DRTIO, since the master
> does not know if a given channel should do replace or collision. The
> satellite would need to report this information for each of its channels
> (and this also needs to be passed from the gateware scripts to the satellite
> manager firmware), then the master should program it into its gateware. Do
> we strongly need it, or can we simply have the replace behavior at all
> times? According to this mailing list discussion, the replace behavior is
> actually wanted:
> https://ssl.serverraum.org/lists-archive/artiq/2016-November/001052.html
Even on the same (data) address, many channels are not idempotent (the
DDS, SAWG, any SPI devices). More generally any channel where state is
kept can be non-idempotent. I don't see how we can track that at the
master.
Why not do blind submission and do the replacement at the satellite
side plus asynchronous error reporting like RTIOBusy?
--
Robert Jördens.
___
ARTIQ mailing list
https://ssl.serverraum.org/lists/listinfo/artiq