[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#323724: Dual G5 turns off after a while



    "Sven" == Sven Luther <sven.luther@wanadoo.fr> writes:

    Sven> On Wed, Aug 24, 2005 at 12:37:09PM -0700, Shyamal Prasad
    Sven> wrote:

    >>  Actually I'm having the same problem that Paul reported,
    >> except on a PowerMac 7,3 (970FX). The boot fails at the exact
    >> same point. I can try benh's suggestion but it will take some
    >> time (need to install the biarch toolchain and recompile
    >> kernel), and it was more about isolating the place where the
    >> fault happens, right?

    Sven> Yes, please do, the biarch toolchain is in sid, so you just
    Sven> have to install those, and you probably want to do this
    Sven> anyway.

Hi Sven,

I rebuilt a ppc64 kernel with the Debian source package with exactly
the same config as 2.6.12-1-powerpc64 from sid.

I turned on the udbg console for pmac in setup.c as benh suggested. I
basically turned on debug in all the files in arch/ppc64/kernel and
the kernel still seems to halt at the same point with the following
output towards the end (copied by hand)

<quote>
time_init: decrementer frequency 33.333333 Mhz
 -> set_preferred_console()
stdoout is /pci@0,f0000000/NVDA,Parent@10/NVDA,Display-B@1
No preferred console found!
smp_core99_probe
smp_core99_kick_cpu
smp_core99_kick_cpu done
smp_core99_setup_cup 0 done

2.6.12-ppc64
</quote>

(the -ppc64 suffix is what I passed to make-kpkg, probably a bad
choice but....).

I've not had time to determine if this console message from setup.c is
important, and I don't have the knowledge to know on sight (though it
certainly sounds like a problem)! But that seems to be the only
additional debug traces that seem relevant. As far as I can tell the
last line is coming from the pseries set up code where it prints out
UTS_RELEASE.

Any ideas? I can continue on this on the weekend if you can suggest a
way forward (had one too many Friday night beers to keep going right
now ;-)

Finally, do you want to merge this bug with #319986? It does seem like
they are both the same even though the machines are slightly
different.

Cheers!
Shyamal



Reply to: