Hi Ben,
Ben Warren wrote:
Yes. Everything's polled. No interrupt funny business.
Thanks, that's quite comforting to know.
I worked a bit with a Coldfire (5328 I think) a couple of years ago and remember that its Ethernet controller was very similar to the SCC/FCC/TSEC controllers that you find on 68360 and PowerQUICC chips. I have no personal experience here, but I would wager that it's also pretty similar to the FEC on the older MCF52x2 Coldfires. Bottom line - this should be a cut & paste job. Look in cpu/bcf52x2/fec.c, cpu/mpc8xx/fec.c, ...
That's what I would have thought, but I've heard this is not true, and I've seen some small differences myself. I would like to know which ColdFire variant it is closest to, or that is is so different that I should not bother and start from scratch.
I also grabbed Freescale's dBug sources last night, which has working drivers for the mcf532x. Does anyone know if those are interrupt-driven? (Yeah, I know, look at the source! Just, if someone already knows...)
There's no point putting your BSP in the main tree until it's complete and tested. That said, you're always welcome to post code for review. Just clearly label it as such and be patient.
Well, yes, there is a point - it would allow others to use it, or even better, work with me to improve it :-P . But of course, your are right that it shouldn't be in the main tree until it is ready. So if it is the operating convention here, I'm happy to comply.
So what do I do? Create a patch and post it?
Is that what all those "[PATCH]' messages are? I had thought those were enhancements/fixes for existing code.
Not big questions, and I would think I could make it clear enough on my own, but I don't want to make any troubles here.
regards, Ben