Le mer. 6 avr. 2022 à 16:14, Mouse <
mouse%rodents-montreal.org@localhost> a écrit :
> Oh, I don't. If someone _does_ have an SBus interface for a pi, I want
> to know about it!
When it comes to the feasibility of such an interface, it depends what you expect (as far as I know, none exists yet).
For a random RPi, using GPIO, I very much doubt it would be possible to match the timings on all pins needed to support a 16 MHz SBus, let alone a faster one.
On the other hand, if you mean "sort of RPi-like", then anything with a PCIe lane might be doable, including a RPi4 Compute Module.
Many "reasonably" priced FPGA (less than $100, some less than $50) have the ability to interface with PCIe, and there's at least one open-source "PCIe endpoint" (to implement a PCIe device) available, in Litex (the SoC infrastructure I use). My SBusFPGA has a working SBus side already, the only missing part is the SBus controller itself, and that's well documented by Sun and IEEE-1496 (i was able to implement the master & slave part of SBus with no major issue). So it should be possible to create a PCIe board with a FPGA to interface between a PCIe host and a SBus device - there should not be any major technical hurdle. I suspect the hardest part would be software support, either to use the SBus device from the host itself, or to be able to expose it to a QEMU/sparc virtual machine in a way that would enable the emulated OS to use it.
Probably only 3-4 years ago it would have been a crazy project for a hobbyist, but now with the expansion of FOSH it's very doable for someone willing to invest the time and effort in it (and money for prototypes, but probably no more than, say, $1000).
Cordially,
--
Romain Dolbeau