Re: OldWorld ROM Macintoshes

I doubt that this is the same issue. The Debian kernel referenced in the 
links was Linux version 4.9.0-2-powerpc.

The Wallstreet has worked on and off since then, with kernel regressions 
introduced as recently as a few years ago regarding KUEP and KUAP.

If you have the .config file that they used to compile their custom 
kernel, we can try compiling a recent kernel. Note that v6.1 works; the 
possible kernel regression occurred between v6.2-rc1 (good) and v6.2-rc8 
(bad).

The possible kernel regression we may be seeing now is that the 
"video=ofonly" fallback stopped working on the Wallstreet. But we should 
first determine whether the xorg mach64 driver can be made to work (why 
doesn't mach64 show up in /usr/lib/xorg/modules/drivers?)

And the complete freeze may be a different problem than X11 not working 
but still getting a text login.


On 9/11/25 10:26 AM, Cedar Maxwell wrote:
> Looks like some folks at Arch fixed this with a custom kernel:
> 
> https://bbs.archlinux32.org/viewtopic.php?id=2776
> 
> Adrian, Riccardo and others discussed this exact same issue at length 9
> years ago, so looks like this isn't new.:
> 
> https://linux.debian.ports.powerpc.narkive.com/vQmkxrXj/xorg-fails-on-ati-after-update-mmio-aperture
> 
> But this "regression" is just about the Xorg server failing to launch
> with Mach64, it has nothing to do with the complete system hang we are
> seeing on kernel 6.2+
> 
> On Thu, 2025-09-11 at 09:32 -0600, Stan Johnson wrote:
>> Hi Cedar,
>>
>> On 9/10/25 9:49 AM, Cedar Maxwell wrote:
>>> On Mon, 2025-09-08 at 21:15 -0600, Stan Johnson wrote:
>>>> ...
>>>>>>
>>>>>> My BootX configuration is as follows (working X11 in 6.1.0 in
>>>>>> Gentoo):
>>>>>>         Kernel: vmlinux-6.1.0-pmac (custom kernel, no modules)
>>>>>>         Boot Device: /dev/sda13 (Gentoo partition)
>>>>>>         More kernel arguments:
>>>>>> video=atyfb:vmode:14,cmode:32,mclk:71
>>>>>>         No video driver: checked
>>>>>>         Options:
>>>>>>           Force SCSI ON: checked
>>>>>>           Force video settings: checked
>>>>>>           Use specified RAM disk: not checked
>>>>>
>>>>> ...
>>>
>>> In your Xorg.log it should say which "video=" option you are using.
>>> ...
>>
>> On a Wallstreet running kernel 6.1 in Gentoo and using twm:
>>
>> $ fgrep video xorg.0.log
>> [    70.783] Kernel command line: root=/dev/sda13
>> video=atyfb:vmode:14,cmode:32,mclk:71 video=ofonly
>> [    72.023] (II) FBDEV(0): hardware: OFfb ATY,RageLT (video memory:
>> 3072kB)
>>
>> $ cat /proc/cmdline
>> root=/dev/sda13 video=atyfb:vmode:14,cmode:32,mclk:71 video=ofonly
>>
>> Like you noticed, it appears that "video=ofonly" is also being passed
>> since I checked "No video driver".
>>
>> See Xorg.0.log_no_video_driver_checked.xz, attached. X11 works.
>>
>> If I uncheck "No video driver", video=ofonly is no longer being
>> passed
>> to the kernel, and X11 does not work (text login prompt only):
>>
>> $ fgrep video Xorg.0.log
>> [    66.565] Kernel command line: root=/dev/sda13
>> video=atyfb:vmode:14,cmode:32,mclk:71
>>
>> $ cat /proc/cmdline
>> root=/dev/sda13 video=atyfb:vmode:14,cmode:32,mclk:71
>>
>> See Xorg.0.log_no_video_driver_unchecked.xz, attached. X11 doesn't
>> work.
>>
>> I think the atyfb driver is the same as Mach64; I see this in
>> Xorg.0.log
>> when video=ofonly isn't passed to the kernel:
>>
>> $ grep -i Mach64 Xorg.0.log
>> [    67.315] (II) LoadModule: "mach64"
>> [    67.323] (WW) Warning, couldn't open module mach64
>> [    67.323] (EE) Failed to load module "mach64" (module does not
>> exist, 0)
>>
>> So we may need to investigate why xorg module mach64 doesn't exist in
>> /usr/lib/xorg/modules/drivers.
>>
>>
>>   > ...
>