Has someone tried PREEMPT_RT patch for Linux on i.MX6?
If yes, did you try measuring RT-latency via cyclictest -- if so, can you please share your observations?
What problem I am observing: if I use '-M' command line argument to cyclictest (that causes screen refresh only when latency is increasing), cyclictest doesn't report any latency, it just sits silent.
If I run it without '-M' option to cyclictest, I am observing latencies in order of 100s of microseconds, when there's no load at all on the CPU.
Thanks,
Rajan