> On Tue, Mar 03, 2009 at 05:19:18PM +0100, Christian Gieseler wrote: > > As timing source i use the same sources that are given in your user > > and mini examples. > > The minimal (kernel) example uses a kernel timer, that is usually not a > good timing source at all. The userspace example uses the setitimer() > function to generate SIG_ALRM signals, which has the same behaviour. > These two examples use on-board methods go create cyclic-tasks to get > something running without great efford. They're not meant to create > jitter-free cycles. > > Do you have the possibility to try RTAI or an RT-preempt kernel? Hello Florian,
thank you for your explanations about the timing sources. I wasn´t sure if there are any more precise alternatives to the used clock sources exept from the ons offered by a patched kernel. Just now I am already working on setting up an RTAI patched kernel. I see that this would help with the jitter values. But i don´t understand how this should change something with the timed out datagrams in idle mode. And since the master module is not an RTAI module this problem probably wouldn´t change. Or do i get something wrong and by setting the rtai directory when configuring the master modules changes and gets another clocksoure if the rtai modules are loaded? Best regards Christian -- Psssst! Schon vom neuen GMX MultiMessenger gehört? Der kann`s mit allen: http://www.gmx.net/de/go/multimessenger01 _______________________________________________ etherlab-users mailing list [email protected] http://lists.etherlab.org/mailman/listinfo/etherlab-users
