Re: OldWorld ROM Macintoshes

Hi Cedar,

On 8/25/25 7:09 PM, Cedar Maxwell wrote:
> Adrian:
> 
> Thanks for your reply.  According to that article in my previous email,
> those instructions should still work with kernel 6.1 or earlier.  Are
> there any offline installer ISOs at
> https://cdimage.debian.org/cdimage/ports/ which contain kernel <=6.1?
> It appears they're all 'net installs, which will download the latest
> kernel from the repository.  Correct me if I'm wrong.
> 
> P.S. Does anyone know why this article would claim kernel >6.1 won't
> work?


There were issues with kernels around 5.10 not booting correctly on 
PowerBook 3400c and Wallstreet systems (please refer to the link that 
Adrian sent). Kernel 6.1 and later work fine on Wallstreet; I'm running 
a custom 6.16.2 kernel on my Wallstreet now.

I don't recommend using Debian's default powerpc kernels. Instead, I 
suggest creating custom kernels depending on your specific hardware as 
determined by looking at the System Profiler output in Mac OS. I try to 
avoid using modules whenever possible. I use the same custom kernel on 
my Wallstreet, Lombard and Pismo systems. On Lombard and Pismo, I boot 
using yaboot; on Wallstreet, I boot using BootX.


> 
> On Sat, 2025-08-23 at 12:55 +0200, John Paul Adrian Glaubitz wrote:
>> Hello,
>>
>> On Fri, 2025-08-22 at 20:12 -0500, Cedar Maxwell wrote:
>>> Can anyone provide any guidance on how to install Debian onto an
>>> OldWorld ROM Macintosh?
>>>
>>> I followed this gentleman's guide to the tee (before it was changed
>>> to
>>> read "OUTDATED") on my PDQ, but as soon as I selected "Linux" from
>>> the
>>> BootX prompt, I received a black screen, and was unable to progress
>>> further.
>>>
>>> https://distrosprojects.blogspot.com/2023/09/modern-debian-on-1998-powerbook-g3.html
>>>
>>> I have struggled to find solid evidence to back the partitioning
>>> choices, and nothing to explain my actual issue.
>>
>> Last time I tried a recent kernel on an OldWorld PowerMac, the kernel
>> crashed
>> quite early and I didn't get to see any usable boot messages. There
>> was a
>> discussion on the topic on this list, see [1].
>>
>> While I have an OldWorld PowerBook somewhere in my storage, I don't
>> really have
>> the time to look into this issue now so that someone else interested
>> needs to
>> do that.
>>
>> It's not unusual for the kernel not to work on hardware on which it
>> hasn't been
>> tested on for a long time.
>>
>> Adrian
>>
>>> [1] https://lists.debian.org/debian-powerpc/2021/07/msg00050.html
> 


Wallstreet systems can only boot from BootX in Mac OS. I believe it's 
also possible to use an install ISO by copying the kernel and initrd 
from the ISO to your local disk and then use BootX to boot that kernel 
and initrd. Neither GRUB nor yaboot will work for booting, since the 
Wallstreet is an OldWorld system. (Note that GRUB also fails on the 
Lombard, even though it is NewWorld, though yaboot works fine there).

If you have another, perhaps faster, G3 system that has SCSI (such as a 
Lombard or a Beige G3 Desktop), I suggest installing there to an 
external SCSI drive, then boot the Wallstreet from that SCSI drive and 
optionally copy the installation (using dump/restore or equivalent) to 
your Wallstreet disk.

To run Debian SID on a Wallstreet, you should have at least 6 GB of 
available disk space and 128 MiB memory (512 MiB, the maximum memory for 
Wallstreet, is of course better). I use pdisk to partition disks from 
Mac OS, or mac-fdisk (or parted) to partition disks from Debian SID or 
Gentoo.

There seems to be an issue right now getting X-Windows to work on the 
Wallstreet with the current Debian SID. My Wallstreet has ATY 3D RAGE LT 
PRO graphics, but wdm and startx tell me that no screens are found. My 
Pismo works; it has ATY Rage128 graphics, so I suspect an upstream issue 
with some version of xorg-xserver. To confirm this, I first need to go 
back to a version of Debian SID that worked, then boot kernel 6.16.2. If 
that works, then I'll see if I can figure out which version of 
xorg-xserver may have caused a regression.

-Stan