[Xenomai-help] XenoPPC + Comedi

2005-10-15 Thread smannori
Hi: with the help of Gilles now I have a working PowerBook with Xenomai. USB-DUX works with Comedi-compiled-for-PPC BUT if I try to mix realtime Xenomai with NON-realtime Comedi the application start, works for a while the stop, exactly as the x86 version. Any suggestions ? Simone

[Xenomai-help] XenoPPC : Latency test results

2005-10-15 Thread smannori
Hi: if I run the latency test and I contunue to use the PPC as usual, after a while, I noticed some overruns (see report below). Question: it is a serious problems or a normal behavior ? Simone Il mattino ha l'oro in bocca

Re: [Xenomai-help] XenoPPC : Latency test results

2005-10-15 Thread Gilles Chanteperdrix
[EMAIL PROTECTED] wrote: Hi: if I run the latency test and I contunue to use the PPC as usual, after a while, I noticed some overruns (see report below). Question: it is a serious problems or a normal behavior ? 150 us is high, you likely have a hardware problem. Do you also have

[Xenomai-help] XenoPPC :- 2.6.13.4-ipipe :)

2005-10-15 Thread smannori
Hi: - Apple PowerBook 1.67 - Mandriva Limited 2005 (Mandrake 10.2) for PPC - Linux 2.6.13.4 - Adeos patch x ppc (32bit - G4) (www.adeos.org) (Thanks Philippe :) ) - Xenomai: configured with NO FPU SUPPORT (Thanks Gilles :) ) Everything works perfect (for the internal test). CPU pushed steady

Re: [Xenomai-help] XenoPPC :- 2.6.13.4-ipipe :)

2005-10-15 Thread Philippe Gerum
[EMAIL PROTECTED] wrote: Hi: - Apple PowerBook 1.67 - Mandriva Limited 2005 (Mandrake 10.2) for PPC - Linux 2.6.13.4 - Adeos patch x ppc (32bit - G4) (www.adeos.org) (Thanks Philippe :) ) - Xenomai: configured with NO FPU SUPPORT (Thanks Gilles :) ) Everything works perfect (for the internal

Re: [Xenomai-help] XenoPPC :- 2.6.13.4-ipipe :)

2005-10-15 Thread Gilles Chanteperdrix
[EMAIL PROTECTED] wrote: Max latency = 44 us Overrun = zero nil nada zip. You should not conclude to rapidly. Running the dohell script for ten hours would make me a bit more confident about your results. I think that the save/restore functions of xenomai must be aware of the presence

[Xenomai-help] XenoPPC :

2005-10-15 Thread smannori
Hi: - Apple PowerBook 1.67 - Mandriva Limited 2005 (Mandrake 10.2) for PPC - Linux 2.6.13.4 - Adeos patch x ppc (32bit - G4) (www.adeos.org) (Thanks Philippe :) ) - Xenomai: configured with NO FPU SUPPORT (Thanks Gilles :) ) CPU pushed steady to 99%: - glxgear (NO DRI, software emululation

Re: [Xenomai-help] XenoPPC :

2005-10-15 Thread Gilles Chanteperdrix
[EMAIL PROTECTED] wrote: If you have any suggestion regarding the testing... I'm ready :) Yes: could you try compiling a vanilla (non adeos patched) kernel without Altivec support, and tell us if it works : 1- without X running 2- with X running If this works: try compiling an Adeos patched