On Wed, May 26, 2004 at 06:37:31PM +0200, Jens Schmalzing wrote:
> Hi,
>
> Sven Luther writes:
>
> > So, we only need to trim the discover database to what is in the
> > initrd, or what we want to load at this stage ?
>
> Using discover is pointless if you already know what to load.
Does discov
On Wed, May 26, 2004 at 12:41:54PM -0500, Troy Benjegerdes wrote:
> On Wed, May 26, 2004 at 06:37:31PM +0200, Jens Schmalzing wrote:
> > Hi,
> >
> > Sven Luther writes:
> >
> > > So, we only need to trim the discover database to what is in the
> > > initrd, or what we want to load at this stage ?
Hi,
Sven Luther writes:
> So, we only need to trim the discover database to what is in the
> initrd, or what we want to load at this stage ?
Using discover is pointless if you already know what to load.
> i sincerely doubt that the 4Mo initrd i have generated on my system
> is indeed very speci
Moving this from debian-boot to debian-kernel, as i doubt it has
anything to do with debian-installer, and is more kernel-related.
On Wed, May 26, 2004 at 05:02:30PM +0200, Jens Schmalzing wrote:
> Dear Sven,
>
> > Well, having it automatically called on the subarches that need it.
>
> You shoul
On Wed, May 26, 2004 at 04:46:29PM +0200, Jens Schmalzing wrote:
> Dear Sven,
>
> > > Why not use discover in mkinitrd instead?
>
> > Well, that is what i proposed, didn't i ?
>
> No, you propose to use discover in the initrd.
Ah, ok, i did miss this misinterpretation.
There still is a proble
Dear Sven,
> Well, having it automatically called on the subarches that need it.
You should be able to put it in the postinst_hook in
kernel-img.conf(5). If not, file a bug report, please.
> Yeah, sure, but i think there is a misunderstanding. the duty of
> discover is exactly to find the modul
Dear Sven,
> > Why not use discover in mkinitrd instead?
> Well, that is what i proposed, didn't i ?
No, you propose to use discover in the initrd.
Regards, Jens.
--
J'qbpbe, le m'en fquz pe j'qbpbe!
Le veux aimeb et mqubib panz je pézqbpbe je djuz tqtaj!
On Wed, May 26, 2004 at 12:06:45PM +0200, Gaudenz Steinlin wrote:
> On Wed, May 26, 2004 at 07:47:47AM +0200, Sven Luther wrote:
> > On Tue, May 25, 2004 at 10:22:25PM +0200, Jens Schmalzing wrote:
> > > No. The initrd should only contain the modules needed for bringing up
> > > the root filesyste
On Wed, May 26, 2004 at 07:47:47AM +0200, Sven Luther wrote:
> On Tue, May 25, 2004 at 10:22:25PM +0200, Jens Schmalzing wrote:
> > No. The initrd should only contain the modules needed for bringing up
> > the root filesystem, and maybe using the console. It is mkinitrd's
> > duty to figure out w
On Tue, May 25, 2004 at 10:22:25PM +0200, Jens Schmalzing wrote:
> Hi,
>
> Sven Luther writes:
>
> > Cool. Why is it in the NEW queue though ?
>
> Well, 2.6.6-3 adds tracker packages, 2.6.6-4 adds tracker packages
> *and* removes the g5 flavour.
>
> > And any progress on the mkvmlinuz issue
>
Hi,
Sven Luther writes:
> Cool. Why is it in the NEW queue though ?
Well, 2.6.6-3 adds tracker packages, 2.6.6-4 adds tracker packages
*and* removes the g5 flavour.
> And any progress on the mkvmlinuz issue
What issue?
> On the other side, i still have problem with the initrd, i was able to
>
On Tue, May 25, 2004 at 09:51:03PM +0200, Sven Luther wrote:
> On Tue, May 25, 2004 at 08:10:59PM +0200, Jens Schmalzing wrote:
> > Colin Watson writes:
> > > On reflection, considering my lack of ability to test this stuff, I've
> > > just taken the easy option and arranged for separate g5 2.6 ima
On Tue, May 25, 2004 at 08:10:59PM +0200, Jens Schmalzing wrote:
> Hi,
>
> Colin Watson writes:
>
> > On reflection, considering my lack of ability to test this stuff, I've
> > just taken the easy option and arranged for separate g5 2.6 images to be
> > built. I missed out a build-dependency, but
On Tue, May 25, 2004 at 08:10:59PM +0200, Jens Schmalzing wrote:
> Colin Watson writes:
> > On reflection, considering my lack of ability to test this stuff, I've
> > just taken the easy option and arranged for separate g5 2.6 images to be
> > built. I missed out a build-dependency, but that's fixe
Hi,
Colin Watson writes:
> On reflection, considering my lack of ability to test this stuff, I've
> just taken the easy option and arranged for separate g5 2.6 images to be
> built. I missed out a build-dependency, but that's fixed in svn now.
Excellent timing :)
Two days ago, in revision 2.6.6
On Mon, May 24, 2004 at 05:36:07PM +0200, Sven Luther wrote:
> On Wed, May 05, 2004 at 12:19:47PM +0100, Colin Watson wrote:
> > On Wed, May 05, 2004 at 07:54:52AM +0200, Jens Schmalzing wrote:
> > > Colin Watson writes:
> > > > It's not cheap at all.
> > >
> > > As it stands, it's the price to pa
On Wed, May 05, 2004 at 12:19:47PM +0100, Colin Watson wrote:
> On Wed, May 05, 2004 at 07:54:52AM +0200, Jens Schmalzing wrote:
> > Colin Watson writes:
> > > It's not cheap at all.
> >
> > As it stands, it's the price to pay for supporting the major flagships
> > of the PowerPC architecture out
On Sat, Apr 24, 2004 at 05:02:51PM +1000, Benjamin Herrenschmidt wrote:
> Hi Sven !
>
> The current CDs won't boot a G5 with the -power4 option. The problem
> is that the kernel contains legacy drivers (8250 serial, etc...) that
> are fine for IBM POWER4 hardware, but will lockup a G5. You really
On Mon, May 10, 2004 at 10:38:50AM +0200, Jens Schmalzing wrote:
> > The patches are ugly as hell and thus not in mainline but they work
> > fine
>
> Which patch exactly?
In the SuSE kernel srpms (ftp.suse.com:/pub/people/mantel/kotd/), in
patches.suse most of the suse.ppc32* patches deal with th
Hi,
Colin Watson writes:
> d-i takes your kernel packages and splits them up into a pile of udebs;
Sorry for my misunderstanding.
> However, I do need to know if it's possible to insert modules from
> the -power4 set into a -g5 kernel.
As far as I can tell (from trying out a few non-vital modu
On Mon, May 10, 2004 at 08:46:14AM +0200, Jens Schmalzing wrote:
> Colin Watson writes:
> > In that case, we could save a fair bit of space by shipping both
> > kernels but only one set of modules, and somehow educating d-i about
> > this.
>
> This would add considerable obfuscation at several poi
On Mon, 2004-05-10 at 02:46, Jens Schmalzing wrote:
> Colin Watson writes:
> > In that case, we could save a fair bit of space by shipping both
> > kernels but only one set of modules, and somehow educating d-i about
> > this.
>
> This would add considerable obfuscation at several points. The bu
Hi,
Christoph Hellwig writes:
> Why doesn't it work?
Can't remember, anyway all the patches we tried still led to deadlock.
> The patches are ugly as hell and thus not in mainline but they work
> fine
Which patch exactly?
Regards, Jens.
--
J'qbpbe, le m'en fquz pe j'qbpbe!
Le veux aimeb et
On Wed, May 05, 2004 at 07:54:52AM +0200, Jens Schmalzing wrote:
> We tried this, but our fixes didn't work. And there have been
> discussions among the kernel hackers recently, with no substantial
> outcome.
Why doesn't it work? The patches are ugly as hell and thus not in
mainline but they wor
Hi,
Colin Watson writes:
> Is the -g5 kernel at least compatible with the -power4 modules? (On
> the face of it, it seems that it should be,
> with the exception of the extra serial modules on -g5
Those are gone in 2.6.5-2 ...
> which presumably we don't want to use anyway.
... for precisely
On Wed, 2004-05-05 at 21:19, Colin Watson wrote:
> On Wed, May 05, 2004 at 07:54:52AM +0200, Jens Schmalzing wrote:
> > Colin Watson writes:
> > > It's not cheap at all.
> >
> > As it stands, it's the price to pay for supporting the major flagships
> > of the PowerPC architecture out of the box.
>
On Wed, May 05, 2004 at 07:54:52AM +0200, Jens Schmalzing wrote:
> Colin Watson writes:
> > It's not cheap at all.
>
> As it stands, it's the price to pay for supporting the major flagships
> of the PowerPC architecture out of the box.
Is the -g5 kernel at least compatible with the -power4 module
Hi,
Colin Watson writes:
> ... or patch the serial console driver in some way that doesn't
> break too badly on either ...
We tried this, but our fixes didn't work. And there have been
discussions among the kernel hackers recently, with no substantial
outcome.
> 27MB more space on powerpc CD 1
On Mon, Apr 26, 2004 at 10:22:24AM +0200, Jens Schmalzing wrote:
> Colin Watson writes:
> > Oh, God, *another* flavour?
>
> Sticking with just one flavour for all the power4 machines means we
> either break the Apple G5 or leave the IBM Power4 without serial
> console. Neither is acceptable.
...
Hi,
Colin Watson writes:
> Oh, God, *another* flavour?
Sticking with just one flavour for all the power4 machines means we
either break the Apple G5 or leave the IBM Power4 without serial
console. Neither is acceptable.
And at least we got rid of having one kernel-image package per flavour
*an
On Sat, Apr 24, 2004 at 10:09:20AM +0200, Jens Schmalzing wrote:
> Benjamin Herrenschmidt writes:
> > Which is why you need 2 kernel images on the CD.
>
> So you are suggesting we should create two different kernel-image
> packages, one for the G5, one for the IBM Power4? Sounds reasonable.
Oh,
> Yuck. How about something that doesn't look so PowerPC-specific?
> You're contaminating non-arch drivers with arch-specific code.
>
> if(has_isa_bus)
> return -ENODEV;
>
> Better yet, find a way to flag the driver init functions
> so that they don't even execute until an ISA bus is fou
On Sat, 2004-04-24 at 04:41, Olaf Hering wrote:
> On Sat, Apr 24, Christoph Hellwig wrote:
> > On Sat, Apr 24, 2004 at 05:02:51PM +1000, Benjamin Herrenschmidt wrote:
> > > Hi Sven !
> > >
> > > The current CDs won't boot a G5 with the -power4 option. The problem
> > > is that the kernel contains
On Sat, 2004-04-24 at 18:25, Christoph Hellwig wrote:
> On Sat, Apr 24, 2004 at 05:02:51PM +1000, Benjamin Herrenschmidt wrote:
> > Hi Sven !
> >
> > The current CDs won't boot a G5 with the -power4 option. The problem
> > is that the kernel contains legacy drivers (8250 serial, etc...) that
> > a
On Sat, Apr 24, Christoph Hellwig wrote:
> On Sat, Apr 24, 2004 at 05:02:51PM +1000, Benjamin Herrenschmidt wrote:
> > Hi Sven !
> >
> > The current CDs won't boot a G5 with the -power4 option. The problem
> > is that the kernel contains legacy drivers (8250 serial, etc...) that
> > are fine for
On Sat, Apr 24, Christoph Hellwig wrote:
> On Sat, Apr 24, 2004 at 05:02:51PM +1000, Benjamin Herrenschmidt wrote:
> > Hi Sven !
> >
> > The current CDs won't boot a G5 with the -power4 option. The problem
> > is that the kernel contains legacy drivers (8250 serial, etc...) that
> > are fine for
Hi,
Benjamin Herrenschmidt writes:
> Which is why you need 2 kernel images on the CD.
So you are suggesting we should create two different kernel-image
packages, one for the G5, one for the IBM Power4? Sounds reasonable.
Regards, Jens.
--
J'qbpbe, le m'en fquz pe j'qbpbe!
Le veux aimeb et mq
On Sat, Apr 24, 2004 at 05:02:51PM +1000, Benjamin Herrenschmidt wrote:
> Hi Sven !
>
> The current CDs won't boot a G5 with the -power4 option. The problem
> is that the kernel contains legacy drivers (8250 serial, etc...) that
> are fine for IBM POWER4 hardware, but will lockup a G5. You really
Hi,
Benjamin Herrenschmidt writes:
> The current CDs won't boot a G5 with the -power4 option. The
> problem is that the kernel contains legacy drivers (8250 serial,
> etc...) that are fine for IBM POWER4 hardware, but will lockup a
> G5. You really need different kernel images for now.
What do
On Sat, 2004-04-24 at 17:57, Jens Schmalzing wrote:
> Hi,
>
> Benjamin Herrenschmidt writes:
>
> > The current CDs won't boot a G5 with the -power4 option. The
> > problem is that the kernel contains legacy drivers (8250 serial,
> > etc...) that are fine for IBM POWER4 hardware, but will lockup
Hi Sven !
The current CDs won't boot a G5 with the -power4 option. The problem
is that the kernel contains legacy drivers (8250 serial, etc...) that
are fine for IBM POWER4 hardware, but will lockup a G5. You really need
different kernel images for now.
Ben.
41 matches
Mail list logo
