Hello, I made various measurements with a ls1021a-twr and a rt-preemption-patch kernel. I was able to capture the latency with cyclictest which can be seen in picture 1. In addtion I captured the periodic jitter of a realtime-application which outputs a square wave which can be seen in picture 2. This Applikation was initialised with a priority of 99 as well as the arguments of cyclictest. In short I could measure these times with a system without load: Latency with the command: cyclictest -l 3600000 -m -n -p 99 -i 1000 -> 47 us Periodic jitter with prio 99, mlockall, clock_nanosleep and monotic_clock -> 17.04 us How is this possible? I know there is the possibility that a oscilloscope can have blind times but not with that amount of cycles. (21600000 with a cycle time of 1ms) Picture 1: http://imgur.com/a/Kaaje Picture 2: http://imgur.com/a/PNxbo I would be grateful for an answer! Best regards J. Weinert -- To unsubscribe from this list: send the line "unsubscribe linux-rt-users" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
