Hi Philippe, thanks for your answer, I ran autotune but the output didn't change, sometimes the assertion just fails as before. Do you think that it could lead to any problem in the standard Xenomai usage? Or is it something that I may safely ignore? I didn't notice any strange behaviour so far, but I'm wondering if this could be an indicator of something going wrong in the system I'm testing.
Just in case it can be of any use, you can find attached the config of my kernel. Thanks again for your kind support, Regards, Andrea 2016-04-22 11:06 GMT+02:00 Philippe Gerum <[email protected]>: > On 04/22/2016 10:51 AM, Andrea wrote: > >> Hello, >> >> I'm currently using Xenomai 3.0.2, under kernel 4.1.8 with i-pipe patch. >> I just noticed that I'm experiencing an error on some run of xeno-test. >> >> The reported error is: >> *posix-clock.c:420, assertion failed: diff < 1000000000* >> */usr/xenomai/bin/smokey: test posix_clock failed: Invalid argument* >> >> The specific function where the assert is issued in posix-clock.c is >> "clock_decrease_after_periodic_timer_first_tick(void)" >> >> On multiple executions it is not reproducible, I am wondering if anybody >> could hint me to the source of the problem on my system. >> >> > This test is highly time-dependent. Try running the autotuner > (/usr/xenomai/bin/autotune) for calibrating the core timer with respect to > system latencies prior to running the tests (aka "timer gravity" settings). > > > -- > Philippe. > -------------- next part -------------- A non-text attachment was scrubbed... Name: config-4.1.18.gz Type: application/x-gzip Size: 44905 bytes Desc: not available URL: <http://xenomai.org/pipermail/xenomai/attachments/20160422/fd6f2167/attachment.bin> _______________________________________________ Xenomai mailing list [email protected] https://xenomai.org/mailman/listinfo/xenomai
