On 11/17/2011 1:36 PM, Jon Elson wrote: > 张 恒 wrote: >> Hello everyone, >> I try to port EMC2 to Mini2440, a platform based on S3C2440 ARM9 >> processor. >> I want to ask if anyone have ported EMC2 to platforms based on arm >> processors. > We tried this by getting a Beagle Board and sending it to Torsten > Koschorrek > in Germany. He is the ARM maintainer for RTAI. That was 18 months ago, > and he still has not produces a port for the Beagle. There is a Source > Forge > project on this > http://sourceforge.net/projects/rtai-cortex/ > > Lately, it seems there has been work on adding the rt-preempt kernel > to thise supported by our rtapi interface, and there is an rt-preempt > kernel > for the Beagle, so that might be close. > > One nasty thing we found out on the Beagle is that the OMAP CPU has > multiplexed GPIO hardware, so the GPIO pins can only be updated every > 240 ns (either read or write). Apparently accessing the GPIO also > incurs a 240 ns wait state to the processor. This was SO WELL hidden in > the TI OMAP manual that I still can't find any place where this > bottleneck > is fully described.
Well, I certainly never heard of this until now, and I spent a fair amount of time reading the docs when you were first talking about the Beagle Board. Now I don't feel so bad that some time ago I got tired of waiting on results from Torsten and decided to build the BB I bought into a pedestrian, non-real time project. I fear this kind of hidden gotcha is going to show up in every hot, new SoC. Regards, Kent ------------------------------------------------------------------------------ All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-novd2d _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
