Re: 2.4 & 2.6 kernels, should sarge be 2.6 only at least for powerpc ?

[Sven Luther <sven.luther@wanadoo.fr>]

On Sat, Jul 03, 2004 at 07:06:30PM +0200, Christoph Hellwig wrote:
> > So, i am seriously considering dropping all 2.4 powerpc kernels, and
> > going with 2.6 only, and would like to get feedback both from
> > debian-kernel as well as debian-powerpc, feedback i didn't get in the
> > past.
> 
> I'd be in favour of that.

Yeah, in the meantime, i need to release a 2.4.26 kernel as there are
people clamoring for it, since the 2.4.25 has some security holes whose
fixes where not backported, and 2.6 is not considered stable by some
users yet (and if you want to fight it, you are welcome to engage in a
flamewar with Branden, i certainly won't :).

The patch from 2.4.26 -benh tree apply cleanly but doesn't build
(probably because of two conflicting libata patches, but maybe because
some other bits and pieces), and altough i worked at it some, it is
still failing with :

fs/fs.o(.text+0x119cc4): dans la fonction « syncd »:
: undefined reference to `sigmask_lock'
fs/fs.o(.text+0x119cdc): dans la fonction « syncd »:
: undefined reference to `__recalc_sigpending'
fs/fs.o(.text+0x119ce0): dans la fonction « syncd »:
: undefined reference to `sigmask_unlock'

Could you (or someone else) have a look at the stuff ? The patch against
kernel-source-2.4.26 is at : 

  http://people.debian.org/~luther/benh.diff.gz

And the config file used at :

  http://people.debian.org/~luther/config-powerpc

I don't feel like fixing stuff in that fs part, and may cause something
to break horribly if i go fixing stuff.

> > Ah, and i am seriously considering dropping support for apus from the
> > kernels (and thus debian-installer). I believe that they are only a
> > handfull of apus users left, and those are happily running self built
> > 2.2 kernels. Furthermore, i have some evidence that not only where the
> > debian apus kernels never tried on apus, but also that there is big
> > chance they don't even work. I don't have apus hardware anymore, so ...
> 
> If you want my 2cent:  If we don't have someone who actively tests a
> subarchitecture and provides feedback consider a (sub-) architecture
> dead.  

Yeah, well, after this consideration, everything but new-world pmac
would be dead, except for me, and i am speaking of prep (and we have a
prep debian machine) and chrp-rs6k, who don't have any active testers,
not speaking about oldworld pmac who is only half tested.

Friendly,

Sven Luther