On 8/16/07, Nadym Salem <[EMAIL PROTECTED]> wrote:
> On Thu, August 16, 2007 15:39, Gilles Chanteperdrix wrote:
>
> >> > > >> trying to use RTnet and having massive synchronisation problems,
> >> a
> >> > > >> problem
> >> > > >> with xenomai latencies came up.
> >> > > >> Our xenomai latency test gives weird values. Attached a
> >> xeno-test
> >> > log.
> >> > > >> Does anyone have an idea ?
> >> > > > Yes, read the FAQ and try the proposed workarounds.
> >> > > Sorry not to mention it. Of course we used all the settings in the
> >> FAQ
> >> > and
> >> > What does your kernel logs say about SMI workaround, is it working ?
> >> It does not say anything about SMI, reading the Torubleshoothing file,
> >> it
> >> sound like there should be a warning if it is not working.
> > But you should get a message "SMI enabled" in the kernel logs. Do you get
> > it ?
>
> Nope...there isn't an entry with SMI at all in all /var/log/*
Aha, this probably means that your chipset is not supporte, could you
run lspci and send us the output ?
>
> >> > > Or am I just too blind to find other workarounds there ?
> >> > There is also the X-server workaround, and the Legacy USB workaround.
> >> No HW Acceleration on.
> >> Legacy USB Option in Bios is missing (like a lot more... :-/ )
> >> > What version of Xenomai are you using ?
> >> Xenomai 2.3.2 with Linux Kernel 2.6.20
> > In case there is a bug in the newly introduced direct access to tsc
> > from posix clock functions, could you re-compile with access to hw tsc
> > disabled and re-run cyclictest ?
>
> I could, just being not able to find the option ;)
It is the option --disable-x86-tsc of the configure script.
--
Gilles Chanteperdrix
_______________________________________________
Xenomai-help mailing list
[email protected]
https://mail.gna.org/listinfo/xenomai-help
