macOS 10.14 Mojave on Unsupported Macs Thread

[xj01]

Now that you reverted to a previous snapshot are you able to boot in single user mode CMD+S ?

Even if I still think that 15 gb are not enough to fit a reinstallation I guess at least 20 gb, anyway you can try re-install, but first from the USB Installer, do a couple of DiskUtility's First Aid on your Mojave Volume.

After the "First Aid", retry to re-install and it should continue.

First aid fails repeatedly and Cmd+S does nothing..

 

[jackluke]

First aid fails repeatedly and Cmd+S does nothing..

If "First Aid" fails after 4-5 attempts, then probably your SSD is becoming faulty.

Anyway try this last attempt to force enter in single user mode at any cost, boot from the USB Installer, open Terminal and type:

nvram boot-args="-no_compat_check -s"

csrutil disable

reboot



At the next reboot at 99% you should boot directly to the single user mode Shell and from there type:

fsck -fy
(this will take a while)

mount -uw /

rm -R /System/Library/Extensions/GeForceTeslaGLDriver.bundle

rm -R /System/Library/UserEventPlugins/com.apple.telemetry.plugin

exit


after a couple of reboot you should land to the desktop, but you have to type "exit" any time since you are now in single user mode and it's an half boot, to return normal mode just from any Terminal type:

nvram boot-args="-no_compat_check"

If in this way you don't fix, then you have to backup and erase, but on a disk that fails more than 5 times a "First Aid" I'd not rely much on it, I'd buy a new one.

This is my humble opinion, can't help you further.

 

[Lars B.]

Hi Guys,

here is my config




can you tell me if the update to Mojave is possible with it ?

do you think it's a good idea ?

Thanks in advance

Yes, it's possible, however I would upgrade to High Sierra instead. High Sierra is still supported by Apple (unlike El Capitan), and almost all programs will still run on it. But because it has been superseded by Mojave, High Sierra will only get security updates that won't break your unsupported installation. With Mojave, you never know if the next version will still work.
Performance with High Sierra will be just as good as with El Capitan, and you won't lose any files and programs if you do an upgrade installation. (You should still have a backup, of course.) You can find dosdude's how-to here: http://dosdude1.com/highsierra/

 

[netsrot39]

This might be a stupid question but is it possible to download a beta installer with 0403979's "macOS Downloader" ? I only manage to download the final version (10.14.4). Also is it possible to download a specific beta version (I wanna dl the 10.14.5b3 since it seems the only viable option right now. Thanks!

 

[Yojx21]

I seem to remember having a similar problem and it was the macOS Mojave.app download was bad. I had thought my USB was bad or had a bad format, but it was the OS D/L.

You where right ! I don't know why but the mojave app downloaded by Dosdude1 patcher was wrong... It was missing something like "basesystem.dmg".

i have installed a fresh copy of High Sierra on one of my MacPro5,1 (the only mac i have that can natively download Mojave), and downloaded Mojave from App Store. Then Mojave Patcher was OK.

I have now another problem... I have setup a patched Mojave on a MacMini3,1, and it reboot radomly. Kernel Parnic report is talking about AppleHDA kext...

Also the Display Port isn't seen as audio output ! :/ Any idea why ?

 

[0403979]

This might be a stupid question but is it possible to download a beta installer with 0403979's "macOS Downloader" ? I only manage to download the final version (10.14.4). Also is it possible to download a specific beta version (I wanna dl the 10.14.5b3 since it seems the only viable option right now. Thanks!

There are no beta installer apps for .x versions. 10.14.4 is the latest installer app so that’s the one you can download. You can download the latest beta update. Right now that’s 10.14.5b3. Tomorrow when 10.14.5b4 might be released then you won’t be able to download 10.14.5b3 anymore since it comes directly from Apple.

 

[xj01]

If "First Aid" fails after 4-5 attempts, then probably your SSD is becoming faulty.

Anyway try this last attempt to force enter in single user mode at any cost, boot from the USB Installer, open Terminal and type:

nvram boot-args="-no_compat_check -s"

csrutil disable

reboot



At the next reboot at 99% you should boot directly to the single user mode Shell and from there type:

fsck -fy
(this will take a while)

mount -uw /

rm -R /System/Library/Extensions/GeForceTeslaGLDriver.bundle

rm -R /System/Library/UserEventPlugins/com.apple.telemetry.plugin

exit


after a couple of reboot you should land to the desktop, but you have to type "exit" any time since you are now in single user mode and it's an half boot, to return normal mode just from any Terminal type:

nvram boot-args="-no_compat_check"

If in this way you don't fix, then you have to backup and erase, but on a disk that fails more than 5 times a "First Aid" I'd not rely much on it, I'd buy a new one.

This is my humble opinion, can't help you further.

This is getting hopeless. I formatted the drive (the above procedure did noy work) and when installing the system, it says “No packages were eligible for install. Contact the software manufacturer for assistance. Quit the installer to restart your computer and try again”. First aid works fine now, I don’t think there’s any hardware problem with the disk. Maybe the installer tool is corrupted?

 

[K two]

You where right ! I don't know why but the mojave app downloaded by Dosdude1 patcher was wrong... It was missing something like "basesystem.dmg".

i have installed a fresh copy of High Sierra on one of my MacPro5,1 (the only mac i have that can natively download Mojave), and downloaded Mojave from App Store. Then Mojave Patcher was OK.

I have now another problem... I have setup a patched Mojave on a MacMini3,1, and it reboot radomly. Kernel Parnic report is talking about AppleHDA kext...

Also the Display Port isn't seen as audio output ! :/ Any idea why ?

Mini3,1 does not provide audio from the MiniDisplayPort. That Mini will KP like infested on .4 but is perfection on .3.

 

[jackluke]

This is getting hopeless. I formatted the drive (the above procedure did noy work) and when installing the system, it says “No packages were eligible for install. Contact the software manufacturer for assistance. Quit the installer to restart your computer and try again”. First aid works fine now, I don’t think there’s any hardware problem with the disk. Maybe the installer tool is corrupted?

Maybe corrupted USB Installer at this point, or packages eligible have a recent compiled data checking, so try from USB Installer Terminal to set the current date/time typing:

date -u 0428190919
reboot

then reboot again to the USB Installer and retry to re-install.

 

[hvds]

If you set the boot argument keepsyms=1, that log will be a lot more helpful.

MBP5,2 with 10.14.5b3 and .4 CoreDisplay, no further replacements (i.e. unchanged from previous post).
Reset the NVRAM and set keepsyms=1 for further running with 9400M. It took almost two days until the first KP happened.
KP happened after I had switched manually to 9600M GT (with system settings, logout/login). gfxCardStatus showed d(iscrete) at that point, was i(nternal) before, both consistent with About this Mac display. Did a shutdown. The KP happened during shutdown (report attached).
After reboot the system still came up in 9400M mode. Tried again to switch to 9600M GT, this time a KP happened during logout/login (attached, involves GeForceTesla this time).
Forced to 9600M GT (by running from a 10.14.3 disk briefly), 10.14.5b3 then came up in 9600M GT mode. But then a KP happended even in this mode (attached, again involving GeForceTesla). System was not usable after that, many KPs, even after resetting NVRAM, SMC, recreating caches - forgotten to try re-apply 1.3.0 patches unfortunately. Left with no idea what had changed to the system.
So did a clean install of 10.14.3 with patcher 1.2.3 over it, to revert from this unstable state of 10.14.5b3. At present running 10.14.3 without problems, even when switching manually between the two GPUs multiple times.
Will run that for a few days to exclude possible GPU or MUX hardware problem. Afterwards upgrade OTA directly from 10.14.3 to 10.14.5b3, post-install patches with 1.3.0, then all replacements according to post #14572.

 

[ASentientBot]

MBP5,2 with 10.14.5b3 and .4 CoreDisplay, no further replacements (i.e. unchanged from previous post).
Reset the NVRAM and set keepsyms=1 for further running with 9400M. It took almost two days until the first KP happened.
KP happened after I had switched manually to 9600M GT (with system settings, logout/login). gfxCardStatus showed d(iscrete) at that point, was i(nternal) before, both consistent with About this Mac display. Did a shutdown. The KP happened during shutdown (report attached).
After reboot the system still came up in 9400M mode. Tried again to switch to 9600M GT, this time a KP happened during logout/login (attached, involves GeForceTesla this time).
Forced to 9600M GT (by running from a 10.14.3 disk briefly), 10.14.5b3 then came up in 9600M GT mode. But then a KP happended even in this mode (attached, again involving GeForceTesla). System was not usable after that, many KPs, even after resetting NVRAM, SMC, recreating caches - forgotten to try re-apply 1.3.0 patches unfortunately. Left with no idea what had changed to the system.
So did a clean install of 10.14.3 with patcher 1.2.3 over it, to revert from this unstable state of 10.14.5b3. At present running 10.14.3 without problems, even when switching manually between the two GPUs multiple times.
Will run that for a few days to exclude possible GPU or MUX hardware problem. Afterwards upgrade OTA directly from 10.14.3 to 10.14.5b3, post-install patches with 1.3.0, then all replacements according to post #14572.

Yikes. The first one looks related to AMFI/codesigning, which is weird. Make sure that SIP is off (csrutil status) but other than that, I'm really not sure.

The second one looks a lot like the much more frequent crash I "fixed" back in July. If it were the only issue, I'd try patching it again, higher up in the call stack than the currently used fix. (I previously deleted the entire clientClose() function that you see at the bottom, and it works for a while, before causing a huge slowdown. But possibly, patching a function like IONVSharedTesla::free() (in between the currently patched nvVirtualAddressSpaceTesla::free() and clientClose() in the call stack) could be a solution, if nothing better were found. This code seems more modification-friendly than most, probably because it's just for freeing resources -- but that's also why my old "fix" caused problems over time. I feel like this could definitely work, but it'd take a lot of testing to be sure that it didn't eventually mess stuff up like last time.)

I don't know about the last one. You could try replacing IONDRVSupport and IOGraphicsFamily (and deleting NDRVShim) with copies from 10.13.6, since that will make them potentially more compatible with the other nVidia kexts from that era. This causes some issues with an AppleGraphicsDevicePolicy kext, but doesn't seem to actually break anything.

I'd like to try to help fix this, but it's hard without having access to the hardware. If you want to try some troubleshooting with the fixes I suggested above, that'd be awesome. And if anyone more knowledgeable (@jackluke @pkouame @parrotgeek1) has any ideas, that'd be awesome too.
[doublepost=1556546744][/doublepost]

This might be a stupid question but is it possible to download a beta installer with 0403979's "macOS Downloader" ? I only manage to download the final version (10.14.4). Also is it possible to download a specific beta version (I wanna dl the 10.14.5b3 since it seems the only viable option right now. Thanks!

Apple only ever provides the latest stable release as a full installer; you have to install 10.14.4 and then update it to 10.14.5 DP3.

It's actually really annoying. Especially on HFS+.

 

[0403979]

Yikes. The first one looks related to AMFI/codesigning, which is weird. Make sure that SIP is off (csrutil status) but other than that, I'm really not sure.

The second one looks a lot like the much more frequent crash I "fixed" back in July. If it were the only issue, I'd try patching it again, higher up in the call stack than the currently used fix. (I previously deleted the entire clientClose() function that you see at the bottom, and it works for a while, before causing a huge slowdown. But possibly, patching a function like IONVSharedTesla::free() (in between the currently patched nvVirtualAddressSpaceTesla::free() and clientClose() in the call stack) could be a solution, if nothing better were found. This code seems more modification-friendly than most, probably because it's just for freeing resources -- but that's also why my old "fix" caused problems over time. I feel like this could definitely work, but it'd take a lot of testing to be sure that it didn't eventually mess stuff up like last time.)

I don't know about the last one. You could try replacing IONDRVSupport and IOGraphicsFamily (and deleting NDRVShim) with copies from 10.13.6, since that will make them potentially more compatible with the other nVidia kexts from that era. This causes some issues with an AppleGraphicsDevicePolicy kext, but doesn't seem to actually break anything.

I'd like to try to help fix this, but it's hard without having access to the hardware. If you want to try some troubleshooting with the fixes I suggested above, that'd be awesome. And if anyone more knowledgeable (@jackluke @pkouame @parrotgeek1) has any ideas, that'd be awesome too.
[doublepost=1556546744][/doublepost]
Apple only ever provides the latest stable release as a full installer; you have to install 10.14.4 and then update it to 10.14.5 DP3.

It's actually really annoying. Especially on HFS+.

What application can be used to modify the kext code and remove the functions? I’d like to inspect the code for fun if I can. Thank you.

 

[jackluke]

Yikes. The first one looks related to AMFI/codesigning, which is weird. Make sure that SIP is off (csrutil status) but other than that, I'm really not sure.

The second one looks a lot like the much more frequent crash I "fixed" back in July. If it were the only issue, I'd try patching it again, higher up in the call stack than the currently used fix. (I previously deleted the entire clientClose() function that you see at the bottom, and it works for a while, before causing a huge slowdown. But possibly, patching a function like IONVSharedTesla::free() (in between the currently patched nvVirtualAddressSpaceTesla::free() and clientClose() in the call stack) could be a solution, if nothing better were found. This code seems more modification-friendly than most, probably because it's just for freeing resources -- but that's also why my old "fix" caused problems over time. I feel like this could definitely work, but it'd take a lot of testing to be sure that it didn't eventually mess stuff up like last time.)

I don't know about the last one. You could try replacing IONDRVSupport and IOGraphicsFamily (and deleting NDRVShim) with copies from 10.13.6, since that will make them potentially more compatible with the other nVidia kexts from that era. This causes some issues with an AppleGraphicsDevicePolicy kext, but doesn't seem to actually break anything.

I'd like to try to help fix this, but it's hard without having access to the hardware. If you want to try some troubleshooting with the fixes I suggested above, that'd be awesome. And if anyone more knowledgeable (@jackluke @pkouame @parrotgeek1) has any ideas, that'd be awesome too.
[doublepost=1556546744][/doublepost]
Apple only ever provides the latest stable release as a full installer; you have to install 10.14.4 and then update it to 10.14.5 DP3.

It's actually really annoying. Especially on HFS+.

I too don't have that dualGPU Nvidia Tesla hardware 2009, and noticed from many posting here that pre-2010 machines need some more adjustments, @hvds could try the packages from #14572 that could make dualGPUs more stable at least in my case dualGPU Nvidia/IntelHD they did, or at this point lets wait for .5 beta 4 to see what happens.

 

[ASentientBot]

What application can be used to modify the kext code and remove the functions? I’d like to inspect the code for fun if I can. Thank you.

https://www.hopperapp.com

 

[pkouame]

MBP5,2 with 10.14.5b3 and .4 CoreDisplay, no further replacements (i.e. unchanged from previous post).
Reset the NVRAM and set keepsyms=1 for further running with 9400M. It took almost two days until the first KP happened.
KP happened after I had switched manually to 9600M GT (with system settings, logout/login). gfxCardStatus showed d(iscrete) at that point, was i(nternal) before, both consistent with About this Mac display. Did a shutdown. The KP happened during shutdown (report attached).
After reboot the system still came up in 9400M mode. Tried again to switch to 9600M GT, this time a KP happened during logout/login (attached, involves GeForceTesla this time).
Forced to 9600M GT (by running from a 10.14.3 disk briefly), 10.14.5b3 then came up in 9600M GT mode. But then a KP happended even in this mode (attached, again involving GeForceTesla). System was not usable after that, many KPs, even after resetting NVRAM, SMC, recreating caches - forgotten to try re-apply 1.3.0 patches unfortunately. Left with no idea what had changed to the system.
So did a clean install of 10.14.3 with patcher 1.2.3 over it, to revert from this unstable state of 10.14.5b3. At present running 10.14.3 without problems, even when switching manually between the two GPUs multiple times.
Will run that for a few days to exclude possible GPU or MUX hardware problem. Afterwards upgrade OTA directly from 10.14.3 to 10.14.5b3, post-install patches with 1.3.0, then all replacements according to post #14572.

While I don't switch GPUs often, my dual-gpu 5,3 (same gpus) has been fairly stable and I usually just leave it on the more powerful 9600M GT.
You started from a clean install - good first step. Can you check the health of your system - fsck and even full system diagnostics if you have them.

 

[pkouame]

MBP5,2 with 10.14.5b3 and .4 CoreDisplay, no further replacements (i.e. unchanged from previous post).
Reset the NVRAM and set keepsyms=1 for further running with 9400M. It took almost two days until the first KP happened.
KP happened after I had switched manually to 9600M GT (with system settings, logout/login). gfxCardStatus showed d(iscrete) at that point, was i(nternal) before, both consistent with About this Mac display. Did a shutdown. The KP happened during shutdown (report attached).
After reboot the system still came up in 9400M mode. Tried again to switch to 9600M GT, this time a KP happened during logout/login (attached, involves GeForceTesla this time).
Forced to 9600M GT (by running from a 10.14.3 disk briefly), 10.14.5b3 then came up in 9600M GT mode. But then a KP happended even in this mode (attached, again involving GeForceTesla). System was not usable after that, many KPs, even after resetting NVRAM, SMC, recreating caches - forgotten to try re-apply 1.3.0 patches unfortunately. Left with no idea what had changed to the system.
So did a clean install of 10.14.3 with patcher 1.2.3 over it, to revert from this unstable state of 10.14.5b3. At present running 10.14.3 without problems, even when switching manually between the two GPUs multiple times.
Will run that for a few days to exclude possible GPU or MUX hardware problem. Afterwards upgrade OTA directly from 10.14.3 to 10.14.5b3, post-install patches with 1.3.0, then all replacements according to post #14572.

I was able to duplicate a KP while switching from 96 to 94: instead of logging out gracefully the machine restarts. Due to my nvram settings my kernel crash log was not as helpful as yours, but I suspect it's a similar crash. Since then, I have switched back and forth a few times with no problem: has to be a latent memory leak type of issue that only shows up after extended use (as you indicate). These types of bugs can be hard to fix without a consistent use case. Also, have you noticed that some gpu switches require multiple logouts: system preferences doesn't register the first logout attempt - you still get the yellow warning. At that point things may have started hosing up - leaving some dangling pointers and memory that can no longer be freed...

 

[0403979]

Does anyone have the 10.14.5 beta 3 update package? I forgot to download it and I want it in my records.

 

[David403]

10.14.5 beta 4 (18F127a) is out and has a new firmware - 144.0.0.0.0. for cMP 5,1 and cMP 4,1->5,1

I will not be updating my firmware. I will be cautious and wait for full release of 10.14.5 final.

 

[jackluke]

Does anyone have the 10.14.5 beta 3 update package? I forgot to download it and I want it in my records.

Even if .5 beta 4 is out right now, I guess there is still time to download the .5 beta 3 using the public beta utility (not the developer one).

 

[TimothyR734]

I bought a MacBook 7,1 and all I have for an old macOS is mountain lion but I want to install High Sierra but I keep getting this for every patcher I want to make and I just downloading 5.7 gigs of nothing

 

[jackluke]

I bought a MacBook 7,1 and all I have for an old macOS is mountain lion but I want to install High Sierra but I keep getting this for every patcher I want to make and I just downloading 5.7 gigs of nothing

On Macbook7,1 you don't need the "patcher" just download HighSierra from here:

https://itunes.apple.com/us/app/macos-high-sierra/id1246284741?mt=12

and let it install officially from there.

Download size should be about 4.8 gb .
[doublepost=1556565227][/doublepost]Just downloaded and installed .5 beta 4 , ll share my experience soon

well well well lets start with first safe good news:

1) .3 OpenGL and .3 GPUSupport aka @dosdude1 patch + .4 CoreDisplay aka @ASentientBot patch still bring full QE/CI on .5 beta 4

2) "Flurry & Co." still work with acceleration

3) .5 beta 4 IOUSB* kext still work OOB

Now I move to the dualGPUs system

4) for dualGPUs to make correctly working MUX switching I confirm the same ".4 set" I used and shared in this post

Using the ".4 frame-kext-set" (included the GPUWrangler required to me for fast GPU MUX switching and power management) doesn't affect in any way the working order of any MacBook with single GPU, infact in both picture I attached are being used those ".4 set" .

In all my unsupported machines .5 beta 4 News.app doesn't crash.

@Larsvonhier .3 Siri patch is still required for Siri working.

For manual patching I just added an HighSierra replacements of AppleHDA.kext and telemetry.plugin (only for C2D).

 

[Larsvonhier]

Great news, @jackluke !
I´m doing round two on my MBA4,1 now after screwing the CoreDisplay.framework copy somehow. Even re-copying the .5 did not help any more - had to re-install 10.14.4 and 10.14.5b4 again.

 

[pkouame]

10.14.5b4 UP on mbp 5,3 dual-gpu (Nvidia 9600+9400) + SSD + APFS ROM patched

Pretty much the same experience as .5b3. Installation was simple:

1. OTA update from preferences while on .5b3
2. latest "patcher" post-install + CoreDisplay (either .3 or .4 CoreDisplay will work)

Enjoy.
[doublepost=1556571898][/doublepost]@hvds - one of the first I tried was switching from 94 to 96 on .5b4 and it was smooth (no KPs - no restarts) Obviously hasn't been stressed much but no glitches.

 

[jackluke]

Great news, @jackluke !
I´m doing round two on my MBA4,1 now after screwing the CoreDisplay.framework copy somehow. Even re-copying the .5 did not help any more - had to re-install 10.14.4 and 10.14.5b4 again.

That's true, infact it seems that the trick "removing the GLDriver" doesn't work more on 5 beta 4, and I guess neither the safe mode, so seems mandatory to replace the "3 acceleration frameworks" even to boot only in framebuffer mode. But next times you have to keep multiple copies of those "3 frameworks", so you can replace them from single user mode.

 

[Larsvonhier]

Success, finally, on MBA4,1 and 10.14.5b4
As described by Jackluke and Pkouame, same "treatment" necessary as before with beta3.

Edit:
MB5,2 also working (with partial trackpad/kbd. patches, night shift, Siri, backlight control)