Re: [Pkg-mol-devel] Bug#445359: [RFH] Mac-on-Linux for Debian lenny
Hi, Gaudenz.
Sorry for being MIA, but I'm working on many things now (including
working on lame and doing things for my own research---if I continue
this way, I'll be a Mr. Mediocrity Researcher on my own area).
On Jul 16 2008, Gaudenz Steinlin wrote:
> Just a short update on this bug if somebody has more time to work on
> this. I will be only partly available until the begining of august.
Will you be able to attent DebConf 8? I have a pressing need of *not*
going, for some reasons beyond my control. :-(
> I will be able to respond to mail, but most likely not be able to work
> on the package.
>
> * The current version of mol in testing and unstable works with the
> patches from SuSE. So no update to mol or the driver packages is
> needed to fix the RC bug on mol-source.
Nice.
> * The fix needed to be able to build mol kernel modules with the current
> Debian kernel was rejected by the kernel team. They only accept
> pachtes also accepted upstream and upstream does not like the patch.
> The longterm way forward is to integrate the mol modules into the
> upstream kernel.
One thing that I have not checked is the following: are the patches from
SuSE that change the kernel only the symbol export or are they more
"intrusive"?
I feel that getting mol integrated into the upstream kernel would be a
huge task (and may be faced with some resistance). On the other hand, it
would be exposed for more people with knowledge of PowerPC arches (like
the people from Australia and the Samba Team---among others).
> * So the only short term option I see is to update mol-source based on
> the SuSE patches and to advise user to build their own kernel with the
> handl_mm_fault patch if they want to run mol.
This is a pity. :-(
I am, OTOH, "on the way out" with powerpc. I can't have my rt2500usb
adaptor working on x86-64 (amd64) with the new kernel (2.6.26), let
alone working on PowerPC. The rt2x00 maintainer doesn't seem to have
much idea of what is happening after seeing a register dump of a working
and a non-working situation (the kernel has regressed).
This is, for me, the only way to communicate using my iBook and using a
wireless connection. It just seems too hard to get things working on a
second-class citizen architecture. As much as I love my iBook G3, I
can't even play some movies on it and this is, for me, a dealbreaker
(together with the lack of wireless connection). :-(
I can't work on hotels while I am attending conferences where the only
way to use the net on the room is via wireless, which forces me to use
MacOS X, which, again, has been orphaned by Apple. :-(
> > The SuSE patches can be found here:
> > http://bb.zhang.free.fr/suse-mol-patch-for-2.6.25/
> > I used all pachtes except suse-ppc32-mol-handle-mm-fault plus the
> > attached mol-mm.patch.
(...)
> > diff --git a/arch/powerpc/kernel/ppc_ksyms.c b/arch/powerpc/kernel/ppc_ksyms.c
> > index c6b1aa3..455fb38 100644
> > --- a/arch/powerpc/kernel/ppc_ksyms.c
> > +++ b/arch/powerpc/kernel/ppc_ksyms.c
> > @@ -189,6 +189,7 @@ EXPORT_SYMBOL(set_context);
> > extern long mol_trampoline;
> > EXPORT_SYMBOL(mol_trampoline); /* For MOL */
> > EXPORT_SYMBOL(flush_hash_pages); /* For MOL */
> > +EXPORT_SYMBOL(handle_mm_fault); /* For MOL */
> > #ifdef CONFIG_SMP
> > extern int mmu_hash_lock;
> > EXPORT_SYMBOL(mmu_hash_lock); /* For MOL */
Really a pity that this can't be included in Debian kernels... It's just
a symbol export. Why not remove the other "for MOL" symbols then
(assuming that they are *just* for mol)? I'll try to submit this patch
upstream, but it seems that some ego's are getting in the way of having
things fixed.
Sorry for the bleak picture, but I'm a bit sad with the situation of
PPC.
Regards, Rogério Brito.
--
Rogério Brito : rbrito@{mackenzie,ime.usp}.br : GPG key 1024D/7C2CAEB8
http://www.ime.usp.br/~rbrito : http://meusite.mackenzie.com.br/rbrito
Projects: algorithms.berlios.de : lame.sf.net : vrms.alioth.debian.org
Reply to: