Re: OldWorld Mac-specific problems; perhaps drop?
On Thu, Jul 13, 2006 at 03:34:55AM -0400, Daniel Dickinson wrote:
> On Thu, Jul 13, 2006 at 09:01:50AM +0200, Sven Luther wrote:
> > On Wed, Jul 12, 2006 at 10:20:19PM -0400, Daniel Dickinson wrote:
> > > -----BEGIN PGP SIGNED MESSAGE-----
> > > Hash: SHA1
> > >
> > > On Tue, 11 Jul 2006 12:48:57 +0200
> > > Sven Luther <email@example.com> wrote:
> > >
> > > > On Tue, Jul 11, 2006 at 05:12:09AM -0400, Daniel Dickinson wrote:
> > > > > As you know I've been doing a fair bit of playing around with d-i
> > > > > and debian on oldworld powerpc macintosh, although I'm only just
> > > > > now getting a development machine setup. Because of the problems
> > > > > with a lack of dfsg-free means to boot the installer on old world,
> > > > > the fact that the only method available for sarge no longer works
> > > > > (BootX; which can't be used on new world machiens) with 2.6.16
> > > >
> > > > The official method of booting oldworld powermacs is using the miboot
> > > > floppies, and altough miboot is currently non-free, work is under way
> > >
> > > I beg to differ. The manual for sarge didn't mention miboot floppies
> > > at all (because they're non-free and apparently can't even be in
> > Because the powerpc manuals for sarge where mostly sub-par, please help
> > improving the manual so this problem doesn't repeat for sarge.
> > This issue was discussed enough here, and on debian-powerpc, that everyone
> > should be aware of it.
> Do you have any message numbers? Remember I've only started with
> powerpc debian recently, so I'm not one of the 'everyone' that 'should
> be aware of it'. Also, if you're saying that new powerpc users should
> be aware that miBoot is the official way to boot, I think you're
> mistaken. The official manual is the usual way to find things out,
> not reading mailing list archives.
google is your friend.
> > > non-free). CD booting doesn't work on old world (requires non-free
> > > apple boot code), and netboot only works with built-in ethernet cards
> > > (and not on all systems with them). That leaves BootX, which was also
> > > non-free, but didn't require debian directly use it (unlike the miBoot
> > > floppies).
> > And did the manual mention that ?
> Yes, and it gave the url to BootX as well.
So, it could have told about the miboot stuff too, but then the d-i folk
mishandled the sarge miboot floppies badly anyway.
> > > > to free it, and there are other people here who are working on
> > > > bettering the floppies, please participate to that effort or
> > > > something.
> > >
> > > I intend to do what I can, but I don't think we should claim to support
> > > something that we can't, at least not officially.
> > Bah ...
> Well, I agree that intellectual property laws are problematic, but
> that doesn't mean we can ignore them.
> If we can point to the miBoot floppies in the official manual this all
> becomes moot, but I suspect that there are legal liabilities involved
> (perhaps a post to debian-legal is in order?)
What legal liabilities ? I mean, if we can talk about bootx, we can talk about
> > > > There are rumors that you can use also a .coff kernel+ramdisk and use
> > > > that one to boot those boxes, which would be gully free, but nobody
> > > > ever investigated this.
> > >
> > > How would this work? By setting boot-device and boot-file? I'd try
> > > this, but mkvmlinuz complains about the coff format in the initrd.
> > Well, as said, nobody ever tried this, so it may be buggy, i would be happy to
> > debug this with you (but by middle of august, as i go into offlineland for 3
> > weeks on saturday).
> I wouldn't mind working on this if the other possibilities for
> dfsg-free installation haven't come through yet by mid-august, when
> you're available. If you're interested, even if they do. Choice is good.
Ok. Alternatively, work on quick-floppy support with p2mate.
> > > > So what would you gain by dropping support ?
> > >
> > > We wouldn't be claiming to support something that isn't actually
> > > supported, namely installation on old world mac. Also, if the kernel
> > > situation doesn't get resolved, old world mac won't be supported in
> > > etch.
> > What exactly is the kernel problem ? And did you file a bug report about it ?
> Yes, I have filed bug report #375035. The kernel panics claiming that
> it can't find the root device, but the evidence is that it doesn't
> even *try* to use the initramfs before it does this. (no freeing
> memory message, for instance).
ah, that one, none of the kernel team has commented on this one, and it is not
all that clear what you mean in the original report. Can you give us an
un-modified full report with the 2.6.17 kernel. Upto and including the error
message, which was missing in your dmesg output.