Hi, What about isolate a core in todays multicore cpu, to do the RT work, using a custom support layer avoiding using linux itself, RTAI or Xenomai?
Even ARMs are going multicore today. I presume that you can have the NON-rt running in standard Linux user land, and then communicate through shared memory with the isolated core. I don't know much about the architecture, but I presume that the isolated core can probably generate its own timer interrupts. I presume that the kernel probably can ignore a hardware card like a parallel port and lend its responsibility/ownership to the core devoted to RT. I'm not an expert but it looks like it can not be really difficult to make HAL's RT part to tun isolated on that core. Most setups are using two threads, and lots just one. A custom support layer without RTAI or XENOMAI can be developed. Also I presume that there is some Open Source POXIX like RTOS, that can be easier o use than to develop. Such a philosophy, if possible, will make LinuxCNC pretty independent of linux and even deployable in other contexes. Would standard Linux in such a theoretical setup still ruin the RT performance of the standalone core running LinuxCNC's real time part?. Cheers, Javier ------------------------------------------------------------------------------ Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ _______________________________________________ Emc-developers mailing list Emc-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-developers
