Booting Imac G5

[J. Mayer]

Hi,

On Mon, 2004-11-15 at 10:26, Segher Boessenkool wrote:
> >> 
> >> - Have to unplug/replug the USB keyboard
> >>   after kernel boot to make it work with kernel
> >>   2.6.10-rc1. No such problem with 2.6.9.
> >
> > There have been various USB related issue in 2.6.10-rc*, have you tried
> > the latest bk ?
> 
> Do you have an url for that?  USB is misbehaving on my 7,2 as well.

I just tested 2.6.10-rc2 and USB related problems I saw are now fixed.

> >> - Lot's of segfaults occuring when multiple
> >>   concurent processes are running, especially
> >>   during compilations. Maybe the RAM is bad
> >>   (I'm trying to port memtest86 to check this)
> >>   or the CPU state is not completely saved /
> >>   restored when rescheduling: it seems not
> >>   to occur when compiling without doing
> >>   anything else concurently.
> >>
> >> Do you have any idea about the last point ? Could this be Altivec
> >> context save / restore problems ?
> >
> > I very much doubt it has anything to do with CPU context
> > saving/restoring... Could be lots of different things, difficult to say
> > at this point. Thermal problem ? Clock chip setup problem ? Bad
> > RAMs ? ...
> 
> Doesn't sound like a hardware problem, as it only occurs if he is
> running multiple processes...
> 
> Or maybe it is...  try running without X?

I think I found the problem.
While testing my RTC driver, I had some problems which seemed to be
cache related.
I made a try:
in arch/ppc64/kernel/misc.S, in flush_dcache_range and
flush_dcache_phys_range, I added a 
dcbf 0,r6 just after the dcbst 0,r6.
This fixed the problems I had in the RTC driver and I can notice that I
got no more segfault using this patch (I even can now update my Gentoo
under X11 using Gnome !).
I guess I could remove the dcbst, as dcbz does the same plus L1 cache
invalidation.
I'm not completly sure of what the problem is, but I think it can be
either the dbcst is not appropriate, as it does not invalidate the L1
cache in 970FX, or HID registers are not well programmed.

I'll send soon a new patch which will make the RTC accesses, reboot and
halt available on SMU based machines (just need a few more tests and
cleanups).

Regards.

-- 
J. Mayer <l_indien at magic.fr>
Never organized