On Fri, 2007-03-30 at 09:45 +0200, Eric Noulard wrote: > 2007/3/29, RAKOTOSALAMA, Nirilanto <[EMAIL PROTECTED]>: > > Hello everybody ! > > > > I installed xenomai 2.3.0 on a FC6 linux distribution with vanilla kernel > > 2.6.19.7 following the > > http://www.captain.at/xenomai-2.1-installation.php installation guide. > > It worked well until I install a new xenomai version (2.3.1) and linux > > kernel (2.6.20.4) > > with the same kernel configuration than the first install. > > In your .config you have > CONFIG_HPET_TIMER=y > > Can you provider boot log in order to see if your box effectively > have HPET discovered at boot time? > > On my box with Xenomai 2.3.0 + linux 2.6.19 this setup make > xeno-test to hard lock, > see: > https://mail.gna.org/public/xenomai-help/2007-02/msg00085.html > > I don't know if HPET_TIMER is an issue in Xenomai 2.3.1 > anymore since my FAQ comment on this issues seems to > have been removed > see > (http://www.xenomai.org/index.php?title=FAQs&diff=1674&oldid=1642) > > Nevertheless SVN history says the HPET issue is not confirmed... >
It's generally not, because there is no issue with HPET over x86_64, even if there _might_ be some issue over x86, I don't know yet. Maybe this issue happens when HPET is combined with other options; looking at the HPET support code does not immediately ring a bell about a potential conflict with the I-pipe / Xenomai. I need to test this configuration first and reproduce the issue unambiguously, but don't have any HPET-enabled x86 hw at hand right now. Until that is done, it is preferable not to document this issue as a known limitation, since it's either a wild goose or a high priority must fix anyway. -- Philippe. _______________________________________________ Xenomai-help mailing list [email protected] https://mail.gna.org/listinfo/xenomai-help
