Snow Leopard on Unsupported PowerPC Macs

[barracuda156]

So i’ve now re-downloaded @educovas customised 10A222 .dmg and can clearly see that the file size is marginally different

Same thing here by the way. Re-downloaded, size is not identical. (Quite amazing how is it able to mount?)

IMO, using archive would be better. If an archive is corrupt, it just won’t unpack. With DMG it is weird.

 

[barracuda156]

Then why the hell does it refuse to boot on my PowerMac G5?

Sorry, I did not get you. What exactly refuses to boot? 10a190 (which should boot) or 10a222 (which has not yet worked on G5)?

 

[barracuda156]

So i’ve now re-downloaded @educovas customised 10A222 .dmg and can clearly see that the file size is marginally different, so my first attempt may well have been a failure due to a corruption that occurred via Mega download service. I was unable to verify checksum on either of the disk images within disk utility on 10.5.8 but when looking at the details in the Finder ‘Get Info’ pane, the file size difference is clear.

Anyway, did not help me, second image does not boot either on G5

 

[ChrisCharman]

If that also doesn't work, I'll try to upload the image to a different website.

Then why the hell does it refuse to boot on my PowerMac G5?

Right. There is an issue with this build insofar as it doesn’t seem to work on machines that it should. I now realise this because, after attempting to image the recently downloaded copy to the PowerBook I was using previously, the system decided to hang at the exact same point during boot. It then occurred to me to try and boot it on another machine, and as I have the image on an external firewire drive, I reached for my primary PowerBook, plugged-in the drive and miraculously it booted into the OS X 10.6 build 10A222 welcome screen.

Following this finding my question to both @barracuda156 and @educovas is simply the following; which model of PowerBook G4 can you personally verify the system is bootable on and or created on? The reason I ask is because it now seems to be the case that it has been inadvertently modified to run only on specific hardware.

I have done some fix-ups this ad hoc way myself at times, so I perfectly understand (ask me now what should be replaced to make Wifi working or which components are to be picked from Xcode 10a190 if Xcode 10a222 is used, and I have no idea, and only partial record for the latter), but making builds reproducible makes life easier eventually, and not only for others.

I want to play devils advocate here because I’m sensing some unintentional and unnecessary tension on the thread. I also completely emphatically understand @educovas and @barracuda156 points here, i have on multiple occasions changed and modified so much of the builds on my test systems that it would be impossible for me to recall and recount, having not recorded detailed notes during the investigative modification process.

Having said that, I will echo what I feel is the intended sentiment of @B S Magnet, which i share, which i will frame as the following questions for @educovas:

1) How was the .dmg created, on what system and using which utility?

2) What has been your testing methodology in your journey from non-bootable 10A222 to the disk image that you’ve generously shared?

3) What is your technical background and experience, as you have stated that you were able to create a kext and were aware of and made use of a ‘stubber tool’ unknown to many of us here on this thread?

4) What guided your decisions when it came to which components from which build of OS X were required to be transplanted into 10A222 in order for it to be bootable or has this just been a lengthy trial and error endeavour?

5) What have you learned from creating this bootable 10A222 image that could inform others to continue to improve this build and/or pave the way to booting later builds?

 

[barracuda156]

Right. There is an issue with this build insofar as it doesn’t seem to work on machines that it should. I now realise this because, after attempting to image the recently downloaded copy to the PowerBook I was using previously, the system decided to hang at the exact same point during boot. It then occurred to me to try and boot it on another machine, and as I have the image on an external firewire drive, I reached for my primary PowerBook, plugged-in the drive and miraculously it booted into the OS X 10.6 build 10A222 welcome screen.

Following this finding my question to both @barracuda156 and @educovas is simply the following; which model of PowerBook G4 can you personally verify the system is bootable on and or created on? The reason I ask is because it now seems to be the case that it has been inadvertently modified to run only on specific hardware.

The first part is a bit confusing. Could you say what worked and what did not and on which hardware? (Sorry.)

To the question: I tried on G5 Quad, G5 2.3 DC and PB G4 a1104. The latter worked, both G5 did not (same symptoms, same verbose boot output).
I have two other PowerBooks, but only one at hand, a1010, I can try on it.

 

[educovas]

Right. There is an issue with this build insofar as it doesn’t seem to work on machines that it should. I now realise this because, after attempting to image the recently downloaded copy to the PowerBook I was using previously, the system decided to hang at the exact same point during boot. It then occurred to me to try and boot it on another machine, and as I have the image on an external firewire drive, I reached for my primary PowerBook, plugged-in the drive and miraculously it booted into the OS X 10.6 build 10A222 welcome screen.

Following this finding my question to both @barracuda156 and @educovas is simply the following; which model of PowerBook G4 can you personally verify the system is bootable on and or created on? The reason I ask is because it now seems to be the case that it has been inadvertently modified to run only on specific hardware.




I want to play devils advocate here because I’m sensing some unintentional and unnecessary tension on the thread. I also completely emphatically understand @educovas and @barracuda156 points here, i have on multiple occasions changed and modified so much of the builds on my test systems that it would be impossible for me to recall and recount, having not recorded detailed notes during the investigative modification process.

Having said that, I will echo what I feel is the intended sentiment of @B S Magnet, which i share, which i will frame as the following questions for @educovas:

1) How was the .dmg created, on what system and using which utility?

2) What has been your testing methodology in your journey from non-bootable 10A222 to the disk image that you’ve generously shared?

3) What is your technical background and experience, as you have stated that you were able to create a kext and were aware of and made use of a ‘stubber tool’ unknown to many of us here on this thread?

4) What guided your decisions when it came to which components from which build of OS X were required to be transplanted into 10A222 in order for it to be bootable or has this just been a lengthy trial and error endeavour?

5) What have you learned from creating this bootable 10A222 image that could inform others to continue to improve this build and/or pave the way to booting later builds?

1. Disk utility, imaged the 10A222 partition using the compressed image format. 10.5.8 running on the same PowerBook5,5.

2. First step was to use the kernel from 10A222 on 10A190 to check if it would be possible to boot. Doing that I figured out the missing symbols in the kernel that were preventing the IONDVRSupport kext from loading. After creating a kext to stub those symbols, 10A190 successfully booted with the kernel from 10A222 and I could try to boot 10A222 So I've installed it on a Intel Mac, imaged the install and transferred to my PowerBook. Once it was in there, I had to figure out what was removed since 10A190 so I ran a script to use 'file' to get every binary with PPC code, then diff to figure out exactly what was removed. Once I had that list, the next step was to downgrade all of them to 10A190 and try to boot it. It got a bit further but something was still crashing. With the help of @ASentientBot, we figured out that the notifyd from 10A190 wasn't working and we've compiled the 10.6 GM version from AOSP. After that it finally booted and I could fix what still wasn't working reading crash logs and guessing what could fix something.

3. I'm the person who created most of the root patches in OpenCore Legacy patcher for the past couple years.

4. All my work is usually based on trial and error and usually the only clue I have is what I see in crash logs. For the QE/CI patch, I've used what I learned from the non-Metal patches in OpenCore Legacy Patcher.

5. The first step to boot any builds after 10A190 is to restore the binaries that were removed. Usually when the booting process is stuck right after launchd started is something in /usr. During my tests I noticed a bug where fat binaries from 10A190 and 10A222 might be incompatible with older and newer builds and it's required to use lips to keep only the ppc code. I might be forgetting something but I can't remember anything else right now.

 

[barracuda156]

The same image which failed to boot on the Quad now, boots on another PowerBook (a1010). I literally boot from one of the Quad’s drives over a FW now:

 

[ChrisCharman]

1. Disk utility, imaged the 10A222 partition using the compressed image format. 10.5.8 running on the same PowerBook5,5.

2. First step was to use the kernel from 10A222 on 10A190 to check if it would be possible to boot. Doing that I figured out the missing symbols in the kernel that were preventing the IONDVRSupport kext from loading. After creating a kext to stub those symbols, 10A190 successfully booted with the kernel from 10A222 and I could try to boot 10A222 So I've installed it on a Intel Mac, imaged the install and transferred to my PowerBook. Once it was in there, I had to figure out what was removed since 10A190 so I ran a script to use 'file' to get every binary with PPC code, then diff to figure out exactly what was removed. Once I had that list, the next step was to downgrade all of them to 10A190 and try to boot it. It got a bit further but something was still crashing. With the help of @ASentientBot, we figured out that the notifyd from 10A190 wasn't working and we've compiled the 10.6 GM version from AOSP. After that it finally booted and I could fix what still wasn't working reading crash logs and guessing what could fix something.

3. I'm the person who created most of the root patches in OpenCore Legacy patcher for the past couple years.

4. All my work is usually based on trial and error and usually the only clue I have is what I see in crash logs. For the QE/CI patch, I've used what I learned from the non-Metal patches in OpenCore Legacy Patcher.

5. The first step to boot any builds after 10A190 is to restore the binaries that were removed. Usually when the booting process is stuck right after launchd started is something in /usr. During my tests I noticed a bug where fat binaries from 10A190 and 10A222 might be incompatible with older and newer builds and it's required to use lips to keep only the ppc code. I might be forgetting something but I can't remember anything else right now.

This is perfect! Thank you @educovas

You’ve basically been doing exactly what I was trying to do when i was tinkering with 10A222 a while back but you clearly have a lot more knowledge and experience than any of us working on this by ourselves over the last few years. I use OpenCore Legacy patcher on some of my other machines and it is a well loved and greatly appreciated project so thank you for that as well!

 

[ChrisCharman]

The first part is a bit confusing. Could you say what worked and what did not and on which hardware? (Sorry.)

To the question: I tried on G5 Quad, G5 2.3 DC and PB G4 a1104. The latter worked, both G5 did not (same symptoms, same verbose boot output).
I have two other PowerBooks, but only one at hand, a1010, I can try on it.

To clarify, @educovas modified .dmg does not work on my a1046 PowerBook G4 even after downloading a second time. That same second download DOES boot successfully on my a1106 PowerBook G4.

 

[educovas]

To clarify, @educovas modified .dmg does not work on my a1046 PowerBook G4 even after downloading a second time. That same second download DOES boot successfully on my a1106 PowerBook G4.

So, the image does work. The photo on #2,304 is where it get's stuck? If so, I see something that could be a problem: InterfaceNamer: timed out waiting for IOkit to quiesce. I don't have that when booting here.

 

[barracuda156]

So, the image does work. The photo on #2,304 is where it get's stuck? If so, I see something that could be a problem: InterfaceNamer: timed out waiting for IOkit to quiesce. I don't have that when booting here.

For me on G5 it is always here:

 

[educovas]

For me on G5 it is always here:

Looking at where it's stuck, I'd guess launchd is failing to launch what it should. that's very difficult to figure out why it's failing since there are no errors. The only thing I can suggest it to try to replace some bins from /usr or /sbin with the ones from 10A190.

 

[barracuda156]

Looking at where it's stuck, I'd guess launchd is failing to launch what it should. that's very difficult to figure out why it's failing since there are no errors. The only thing I can suggest it to try to replace some bins from /usr or /sbin with the ones from 10A190.

Do you know roughly what is supposed to be used from there? Just to narrow down scope of choices. (I have no idea how OS boot works, it is not something I ever touched.)

 

[educovas]

Do you know roughly what is supposed to be used from there? Just to narrow down scope of choices. (I have no idea how OS boot works, it is not something I ever touched.)

I have no idea what could be wrong, unfortunately. Today I think I found something that was causing some instability that you could try but I'm not sure if that would help you.

Remove the 2 AppleFSCompression, BootCache and autofs kexts and replace them with the kexts from a 10A222 installer.

 

[barracuda156]

I have no idea what could be wrong, unfortunately. Today I think I found something that was causing some instability that you could try but I'm not sure if that would help you.

Remove the 2 AppleFSCompression, BootCache and autofs kexts and replace them with the kexts from a 10A222 installer.

Thanks, I will try.

Meanwhile, it boot on MacMini G4 normally:

 

[B S Magnet]

Having said that, I will echo what I feel is the intended sentiment of @B S Magnet, which i share, which i will frame as the following questions for @educovas:

1) How was the .dmg created, on what system and using which utility?

2) What has been your testing methodology in your journey from non-bootable 10A222 to the disk image that you’ve generously shared?

3) What is your technical background and experience, as you have stated that you were able to create a kext and were aware of and made use of a ‘stubber tool’ unknown to many of us here on this thread?

4) What guided your decisions when it came to which components from which build of OS X were required to be transplanted into 10A222 in order for it to be bootable or has this just been a lengthy trial and error endeavour?

5) What have you learned from creating this bootable 10A222 image that could inform others to continue to improve this build and/or pave the way to booting later builds?

Yah, I guess, but I’m not that stringent.

For example, I’d be less preoccupied by #3 than I would one’s initiative and eagerness to dive into the work. #2 is particularly important from my vantage, as is #4.

I’m sorry for coming across as terse, if not stern and admonishing. I wasn’t striving for that at all.

Also, as much as I want to get back into this project, I’m not in a place right now to be able to do that. So from the sidelines, I’m mostly trying to keep folks aware of how we’re chronicling our work.

 

[educovas]

Thanks, I will try.

Meanwhile, it boot on MacMini G4 normally:

Does QE/CI work?

 

[ChrisCharman]

So, the image does work. The photo on #2,304 is where it get's stuck? If so, I see something that could be a problem: InterfaceNamer: timed out waiting for IOkit to quiesce. I don't have that when booting here.

Yeah i noticed that as well, seems like it could be an IOKit driver issue. I will replace the kernel with the debug kernel next and see if any further output is displayed and review the crashreporter log files.

 

[MacPro2006VBox]

1. Disk utility, imaged the 10A222 partition using the compressed image format. 10.5.8 running on the same PowerBook5,5.

2. First step was to use the kernel from 10A222 on 10A190 to check if it would be possible to boot. Doing that I figured out the missing symbols in the kernel that were preventing the IONDVRSupport kext from loading. After creating a kext to stub those symbols, 10A190 successfully booted with the kernel from 10A222 and I could try to boot 10A222 So I've installed it on a Intel Mac, imaged the install and transferred to my PowerBook. Once it was in there, I had to figure out what was removed since 10A190 so I ran a script to use 'file' to get every binary with PPC code, then diff to figure out exactly what was removed. Once I had that list, the next step was to downgrade all of them to 10A190 and try to boot it. It got a bit further but something was still crashing. With the help of @ASentientBot, we figured out that the notifyd from 10A190 wasn't working and we've compiled the 10.6 GM version from AOSP. After that it finally booted and I could fix what still wasn't working reading crash logs and guessing what could fix something.

3. I'm the person who created most of the root patches in OpenCore Legacy patcher for the past couple years.

4. All my work is usually based on trial and error and usually the only clue I have is what I see in crash logs. For the QE/CI patch, I've used what I learned from the non-Metal patches in OpenCore Legacy Patcher.

5. The first step to boot any builds after 10A190 is to restore the binaries that were removed. Usually when the booting process is stuck right after launchd started is something in /usr. During my tests I noticed a bug where fat binaries from 10A190 and 10A222 might be incompatible with older and newer builds and it's required to use lips to keep only the ppc code. I might be forgetting something but I can't remember anything else right now.

You must’ve not stubbed out calls for the G5 or something.

 

[educovas]

I've imaged a 10A190 install with the 10.0.0d3 kernel, the same kexts and some other patches I used on 10A222 in case someone wants to see if the kernel/kexts or one of those patches are the reason 10A222 won't boot on G5 or the other G4 machines that could't boot it.

Image: https://mega.nz/file/tBZQXAyJ#XQOu1pb5cvH5BCAZWQ3baJCxSvATXcoagzMU0rig8bo

Replaced files: https://mega.nz/file/JBQEVYxZ#Q5WyiYtBV8dhLGUngzxCiFYW390k29lH5ouO5fAiyMo

 

[MacPro2006VBox]

I've imaged a 10A190 install with the 10.0.0d3 kernel, the same kexts and some other patches I used on 10A222 in case someone wants to see if the kernel/kexts or one of those patches are the reason 10A222 won't boot on G5 or the other G4 machines that could't boot it.

Image: https://mega.nz/file/tBZQXAyJ#XQOu1pb5cvH5BCAZWQ3baJCxSvATXcoagzMU0rig8bo

Replaced files: https://mega.nz/file/JBQEVYxZ#Q5WyiYtBV8dhLGUngzxCiFYW390k29lH5ouO5fAiyMo

I bet it’s the kernel. Why else would it KP on my G5? Get the kernel from 10A190 working, it seems to be the last one to work on G5, or better yet GET A G5!

 

[Jazzzny]

I bet it’s the kernel. Why else would it KP on my G5? Get the kernel from 10A190 working, it seems to be the last one to work on G5, or better yet GET A G5!

Why should anyone listen to you? You clearly don't seem to know much about what you're saying. Instead of making pointless (and incorrect) assumptions, you could instead fix it yourself or keep it to yourself.

 

[MacPro2006VBox]

Why should anyone listen to you? You clearly don't seem to know much about what you're saying. Instead of making pointless (and incorrect) assumptions, you could instead fix it yourself or keep it to yourself.

I just restored the patched 10a190 with the D3 kernel found in 10a222 they had shared and it also KP’d on my PMG5 2.0 DP.

 

[educovas]

I bet it’s the kernel. Why else would it KP on my G5? Get the kernel from 10A190 working, it seems to be the last one to work on G5, or better yet GET A G5!

I would get a G5 if I had money for that.

 

[barracuda156]

I would get a G5 if I had money for that.

Unless you look only for Quad or 2.7 DP, everything else is regularly available at very low prices, and in the US perhaps often for free. (But shipping may involve insane costs, so it depends.)

Or, given that you are developing for the platform, maybe ask here in a dedicated thread or/and on Reddit if someone in your geographical vicinity may donate a PowerMac?