On Sat, 2009-04-04 at 06:30 +0200, Thomas Gleixner wrote:
quoted text
> On Fri, 3 Apr 2009, Fernando Lopez-Lezcano wrote:
> > On Fri, 2009-04-03 at 19:21 +0200, Thomas Gleixner wrote:
> > > We are pleased to announce the next update to our new preempt-rt
> > > series.
> > >   
> > > 	- rebased to 2.6.29.1
> > > 
> > > 	- powerpc64 tlb rework fix by Peter
> > 
> > Thanks for cc'ing me...
> > 
> > Just a heads up. I'm still having shutdown/suspend problems with this
> > release as before (I had not reported the suspend problem, I think -
> > machine hangs and has to be power cycled). Booting with "noreplace-smp"
> > fixes both. Is this something I could disable through configuration
> > options?
> 
> No. I have this shutdown/suspend thing on my todo list. 

Ah, ok. No problem. If it were not for this feature I'd probably release
a test version for Planet CCRMA. 

quoted text
> Can you send me your .config please ? 

Attached. 

quoted text
> This happens with the stock -rt kernel or do
> you have other patches on top of it ?

Other patches on top of it. I'm building rpm packages merging the latest
Fedora .src.rpm's with rt (as I have been doing for a long long time). I
have to have as close a kernel as possible to the Fedora kernels
(drivers + configs, within reason - some of the stuff does not patch
cleanly of course[*]). The kernel I'm testing is merged with
2.6.29.1-15 / fc10. 

Sigh, I know, I know, I have to try to test with the vanilla tree. I
don't anticipate having time to do that soon (concert tomorrow, trip
starting next week). 

quoted text
> > I was also seeing a hard hang when running a java program (with the
> > Fedora 10 open source java) - the swingosc gui component of
> > SuperCollider 3 (a computer music synthesis program). Impossible to know
> > what had failed, machine completely catatonic and nothing left on the
> > logs. 
> > 
> > I managed to start the same program _once_ (have not tested again) with
> > 2.6.29.1/rt4 + some patches & config from fc10' koji 2.6.29.1-15. The
> > mouse seemed to pause a couple of times and then I killed the program.
> > Restarting it again was successful. Sorry to not have more hard info...
> 
> Too bad that I can't find my crystal ball :)

You __have__ one?? Wow. Anyway, it seems that the java problem is gone
or something else killed it (another software upgrade perhaps?). Which
is good. 

Thanks again, looking forward to the next release!
-- Fernando

[*] over the years I have tried different approaches and this is the one
that has given me kernels that work and are most compatible with the
rest of Fedora.