Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Status
The first post of this thread is a WikiPost and can be edited by anyone with the appropiate permissions. Your edits will be public.
I did upgrade directly from High Sierra, do you think it's worth wiping the partition and trying again with a clean install? (Leaving my Bootcamp partition untouched)

If you don't have important data on it you could try, but in that case there is a little risk that Bootcamp partition could misalign, so if Bootcamp was created from HighSierra, then you should know to make it bootable again with hybrid MBR (if you have important datas on Bootcamp I mean).
 
  • Like
Reactions: Jas-Singh685
Apart "diskutil apfs updatepreboot /" I don't know another way how to set the right boot.efi to the correct APFS Volume, the fact you had two folders inside your /Volumes/Recovery means that something got wrong in past, but I think a next update could correct that. Meanwhile nvram parameters win over everything, so not a big issue.
[doublepost=1552841207][/doublepost]

Simply from Mojave Terminal: sudo nvram boot-args="-no_compat_check"
[doublepost=1552841452][/doublepost]

The fact you can reach the GUI Mode in safe-mode means that some graphics kext to gain OpenGL or CoreBrightness or maybe other kext are incorrect. If you are on 10.14.4 you should apply from the safe mode the dosdude1 accelpatch, if you are on 10.14.3 or lower I think you should re-apply the dosdude1's Post Install Video patch, and from there force rebuild cache.



Okay I gonna give it a try. I am doing at the moment a Time-Machine-Backup to be sure via the safemode. It takes still round about 3 hours it says. But just to be sure.

I got version 10.14.3. The video resolution is laggy which was once time ago also like this. I was forced to kextload the AMD kext. But after that Backup I gonna try to run the Post Mac installer. And then I gonna give you a new info.

Since I put into the nvram boot-args="-no_compat_check -v -x", do I need to take care of something? to undo it or anything else?

What if the mac post instaler does not fix this issue? since I tried that serveral times..but it didnt work.
The display does not shows any annomalie at the moment.
 
Okay I gonna give it a try. I am doing at the moment a Time-Machine-Backup to be sure via the safemode. It takes still round about 3 hours it says. But just to be sure.

I got version 10.14.3. The video resolution is laggy which was once time ago also like this. I was forced to kextload the AMD kext. But after that Backup I gonna try to run the Post Mac installer. And then I gonna give you a new info.

Since I put into the nvram boot-args="-no_compat_check -v -x", do I need to take care of something? to undo it or anything else?

What if the mac post instaler does not fix this issue? since I tried that serveral times..but it didnt work.
The display does not shows any annomalie at the moment.

No, nothing else to care, just drop -v (verbose) and -x (safe mode) from nvram when you want attempt a normal boot.

Anyway you could simply re-install over your current Mojave Volume, even a previous version (down-grade) I did and it worked from 10.14.4 to 10.14.2 for example, you won't lose your datas.
 
If you don't have important data on it you could try, but in that case there is a little risk that Bootcamp partition could misalign, so if Bootcamp was created from HighSierra, then you should know to make it bootable again with hybrid MBR (if you have important datas on Bootcamp I mean).

Yes I've had to do this before when I dual booted Mojave and HS then I lost bootcamp and recovered it with hybrid MBR. I'll wipe my drive and try again. Thanks!
 
  • Like
Reactions: jackluke
No, nothing else to care, just drop -v (verbose) and -x (safe mode) from nvram when you want attempt a normal boot.

Anyway you could simply re-install over your current Mojave Volume, even a previous version (down-grade) I did and it worked from 10.14.4 to 10.14.2 for example, you won't lose your datas.

What I did my friend was just reinstalling the Mojave version a third time and tried the post installer, and as I already mentioned. It just got always stucked. only when it moved to some point 75% or close to 90% loading, was the safe mode thingy. and also the safe mode allowed me to login twice. so now I try to fix it.

how can I downgrade the version?
What I did while doing a time machine backup was installing the disable gpu from dosdude1.
Just to be sure, that it works... .
what would you do in the first place?
 
What I did my friend was just reinstalling the Mojave version a third time and tried the post installer, and as I already mentioned. It just got always stucked. only when it moved to some point 75% or close to 90% loading, was the safe mode thingy. and also the safe mode allowed me to login twice. so now I try to fix it.

how can I downgrade the version?
What I did while doing a time machine backup was installing the disable gpu from dosdude1.
Just to be sure, that it works... .
what would you do in the first place?

I don't have those dual GPUs machines so I don't know the exactly behavior but since you can boot to the safe mode then try boot from there again and from Mojave Terminal:
sudo rm -R /System/Library/Extensions/AMDRadeonX3000.kext
that should exclude the OpenGL acceleration from the patched ATI/AMDs, so you should be able to boot in "normal" mode.
 
Apart "diskutil apfs updatepreboot /" I don't know another way how to set the right boot.efi to the correct APFS Volume, the fact you had two folders inside your /Volumes/Recovery means that something got wrong in past, but I think a next update could correct that. Meanwhile nvram parameters win over everything, so not a big issue.
[doublepost=1552841207][/doublepost]

Simply from Mojave Terminal: sudo nvram boot-args="-no_compat_check"
[doublepost=1552841452][/doublepost]

The fact you can reach the GUI Mode in safe-mode means that some graphics kext to gain OpenGL or CoreBrightness or maybe other kext are incorrect. If you are on 10.14.4 you should apply from the safe mode the dosdude1 accelpatch, if you are on 10.14.3 or lower I think you should re-apply the dosdude1's Post Install Video patch, and from there force rebuild cache.

Thanks for your immediate reply.
 
  • Like
Reactions: jackluke
Wiped my drive clean (even bootcamp), removed all partitions, created a fresh new APFS container and reinstalled Mojave. I now only have one folder in the Recovery volume. After adding -no_compat_check to boot.plst, still no luck :( Again it's finding the boot.efi from S/L/CoreServices in Mojave, not in the recovery partition. Again did the update preboot commands too but still no go.

No loss, I feel better having done a clean install instead of an upgrade, will just use sudo nvram boot-args="-no_compat_check" for now. Really appreciate all the help @jackluke
 
Last edited:
  • Like
Reactions: jackluke
Wiped my drive clean (even bootcamp), removed all partitions, created a fresh new APFS container and reinstalled Mojave. I now only have one folder in the Recovery volume. After adding -no_compat_check to boot.plst, still no luck :( Again it's finding the boot.efi from S/L/CoreServices in Mojave, not in the recovery partition. Again did the update preboot commands too but still no go.

No loss, I feel better having done a clean install instead of an upgrade, will just use sudo nvram boot-args="-no_compat_check" for now. Really appreciate all the help @jackluke

Really weird is the fact that system won't let you complete the routine "diskutil apfs updatepreboot".
But have the APFS UUID changed after your new installation ? They must change all (the containers and disks).
Have you Erased the whole disk selecting from DiskUtility the "Vendor name" and not only the main partition?

I could think a kind of incompatibility with the internal drive you're using, don't forget that from Mojave launch APFS is better using with an SSD while not so much with a spinning disk where HFS+ is still preferred.

Anyway we should observe that more forward go (example 10.14.4 and later) less compatibility will encounter.

I still think that if when 10.14.4 final will be released and you do the OTA update from your current Mojave, this will auto-fix this weird issue.

Otherwise my last advise is to find somewhere or exploring some Mojave Installer this package or similar named: macOSUpd10.14.3.RecoveryHDUpdate.pkg
Trying to "update" adjusting the Recovery "boot.efi" partition with it.

Perhaps findable here (open it with TextEdit):
https://swscan.apple.com/content/catalogs/others/index-10.14seed.merged-1.sucatalog

here is exactly: http://swcdn.apple.com/content/down...OSUpd10.14.3Supplemental.RecoveryHDUpdate.pkg
 
Last edited:
Wiped my drive clean (even bootcamp), removed all partitions, created a fresh new APFS container and reinstalled Mojave. I now only have one folder in the Recovery volume. After adding -no_compat_check to boot.plst, still no luck :( Again it's finding the boot.efi from S/L/CoreServices in Mojave, not in the recovery partition. Again did the update preboot commands too but still no go.

No loss, I feel better having done a clean install instead of an upgrade, will just use sudo nvram boot-args="-no_compat_check" for now. Really appreciate all the help @jackluke

Would it be at all possible that your changes to the boot.plst were not saved since it is locked?
Probably a very dumb suggestion but I remember having to work around that myself.
 
Would it be at all possible that your changes to the boot.plst were not saved since it is locked?
Probably a very dumb suggestion but I remember having to work around that myself.

You're right, in-fact from his previous pictures the APFS Recovery "com.apple.boot.plist" window appears "locked", while shouldn't.

And of course using nvram boot-args win even over a locked file.
 
  • Like
Reactions: Jas-Singh685
@jackluke The UUID did indeed change! I did erase the drive by clicking on the outermost volume which the name of the drive, SanDisk 960GB Solid State. I will try the installer, I’m really curious to find out why it’s acting like this!

@avz Yes I did notice that! I don’t know why the file is locked, I wasn’t sure if it should be? But I copied the file to my desktop, modified it and then copied it back so the changes were saved. If I tried to modify it in the recovery location then the changes wouldn’t save.

As an experiment, I locked my Mac using iCloud. It reboot into recovery just fine!! Asking for the password and everything. Then I rebooted normally, but CMD+R still doesn’t work. Here’s a picture of when I used the iCloud lock. It seemed to use the recovery volume boot.efi just fine. Notice the changed UUID from the reinstall of Mojave.
 

Attachments

  • 04B6958A-81BA-4C35-99DE-0CC8EDD5AA07.jpeg
    04B6958A-81BA-4C35-99DE-0CC8EDD5AA07.jpeg
    3 MB · Views: 252
  • Like
Reactions: jackluke
@jackluke The UUID did indeed change! I did erase the drive by clicking on the outermost volume which the name of the drive, SanDisk 960GB Solid State. I will try the installer, I’m really curious to find out why it’s acting like this!

@avz Yes I did notice that! I don’t know why the file is locked, I wasn’t sure if it should be? But I copied the file to my desktop, modified it and then copied it back so the changes were saved. If I tried to modify it in the recovery location then the changes wouldn’t save.

As an experiment, I locked my Mac using iCloud. It reboot into recovery just fine!! Asking for the password and everything. Then I rebooted normally, but CMD+R still doesn’t work. Here’s a picture of when I used the iCloud lock. It seemed to use the recovery volume boot.efi just fine. Notice the changed UUID from the reinstall of Mojave.

It is correct, but the only way I know to adjust the APFS Recovery "boot.efi" associating with "CMD+R" is "diskutil apfs updatepreboot", because "bless" command with APFS seems almost ineffective, I've checked that on almost all my machines I do have always the nvram setting with -no_compat_check, it's just an habit but it makes things work.

Before trying to install manually the Recovery package, do this stupid test, select from Mojave's System Preferences / Startup Disk and pick your Mojave Volume and restart, then check once again CMD+R.

Please note that inside your new APFS Recovery path the com.apple.boot.list must contain -no_compat_check, because the BaseSystem.dmg that Recovery will mount in RAM does have inside other PlatformSupport.plist and com.apple.Boot.plist unsupported, that why must be imposed. While on your machine I guess is not necessary to replace the immutablekernel since it's a Sandy Bridge.
 
Any chance you can find a better quality image?
Dark theme? Safari dev. preview setting? Or how do you view macrumors website in dark theme? Screenshot, pls...
Thanks!

Edit:
I´ve tested the latest patcher of 03-17-2019 (but had to replace the prelinked kernel again with my one for the MB 4,1).
It works fine then on these test machines, including correct and fast trackpad behavior:
MacBook Air 4,1
MacBook 4,1
MacBook 5,2
MacBook Pro 5,2
 
Last edited:
I don't have those dual GPUs machines so I don't know the exactly behavior but since you can boot to the safe mode then try boot from there again and from Mojave Terminal:
sudo rm -R /System/Library/Extensions/AMDRadeonX3000.kext
that should exclude the OpenGL acceleration from the patched ATI/AMDs, so you should be able to boot in "normal" mode.
Worked fine. Just did it as you said it and it was fine. Wasn’t even forced to remove the .kext files again or even the only one. Just in Safemode the terminal command fixed it. While it didn’t get fixed with cmd+s and then the command.
I appreciate your help a lot guys!
You make the world a better place with your help! Hope that I can contribute sometimes also something :)
 
  • Like
Reactions: olad and jackluke
To try force entering into "safe mode" try this from single user moder (CMD+S) typing exacly as is:

nvram boot-args="-no_compat_check -x -v"
reboot


to exit from safe mode hold after power on chime CMD+ALT+P+R or re-enter CMD+S typing:
nvram boot-args="-no_compat_check"
reboot

I've seen that all over and it didn't click. "Thank you ever so much!" I got in!! to the log in window and after entering my PW, it's sitting there. I hope it's doing something in the . . . just got a dark screen . . . and a reboot back to the 2/3rds bar and "About 8 minutes remaining". So it's finishing the 14.5 install. We'll have to see where I land after this. It's up in it's full beauty in 14.4 and not the 14.5 that the installed said it was installing. The updater says 10.14.4 is up to date. Great a step further.

But during its very verbose initial start up, I swear that I saw in a couple of lines of text that flashed by it call me some nasty names.

In Safe Boot now on to the rest and hopefully can get DD1s patch installed. Hm? Now to install the graphics patch . . . . Now stuck in getting it to change the dir. I've done this a billions times and I'm drawing a blank . . . I'm in, it's running and for those that watch this fly by, I noticed it says it can't over write the files. But it looks like it does. Because everything is up and running nicely after the reboot, the Patcher popped up and is Updating itself and patching other things. It took a couple reboots to finish it all and we're Golden now!!

"Perfection!!" Got everything installed. AND it boots a lot faster that it has before. That's nice. Very Nice.

Thanks to Dosdude1 and especially jackluke for all the materials and guidance to feed those Geek Genes of mine that just can't leave well enough alone. Thanks guys!!
 
Last edited:
  • Like
Reactions: olad and jackluke
I don't have those dual GPUs machines so I don't know the exactly behavior but since you can boot to the safe mode then try boot from there again and from Mojave Terminal:
sudo rm -R /System/Library/Extensions/AMDRadeonX3000.kext
that should exclude the OpenGL acceleration from the patched ATI/AMDs, so you should be able to boot in "normal" mode.
By the way. I checked instantly after it booted normal if there are some .kext 3000 files in the system extension folder. There were nothing. I also thought first of all that they gonna be added again but they didn’t.
What I would like to ask and to know also. Do i need to update via the App Store or via the patcher from dosdude1?

Best regard
 
Worked fine. Just did it as you said it and it was fine. Wasn’t even forced to remove the .kext files again or even the only one. Just in Safemode the terminal command fixed it. While it didn’t get fixed with cmd+s and then the command.
I appreciate your help a lot guys!
You make the world a better place with your help! Hope that I can contribute sometimes also something :)

It didn't worked from CMD+S cause you had to put first "mount -uw /" and since it's a shell bash you don't need "sudo" command from there.
[doublepost=1552858969][/doublepost]
By the way. I checked instantly after it booted normal if there are some .kext 3000 files in the system extension folder. There were nothing. I also thought first of all that they gonna be added again but they didn’t.
What I would like to ask and to know also. Do i need to update via the App Store or via the patcher from dosdude1?

Best regard

You can update from both ways, but I'd say waiting for the final release of an update through dosdude1's Patcher is safer for unsupported machines.
 
  • Like
Reactions: FarmerBob and jklas
Dark theme? Safari dev. preview setting? Or how do you view macrumors website in dark theme? Screenshot, pls...
Thanks!

Edit:
I´ve tested the latest patcher of 03-17-2019 (but had to replace the prelinked kernel again with my one for the MB 4,1).
It works fine then on these test machines, including correct and fast trackpad behavior:
MacBook Air 4,1
MacBook 4,1
MacBook 5,2
MacBook Pro 5,2
It’s a chrome extension called dark reader. I have included the trackpad patch for the 5,2 but not in the installer.
 
  • Like
Reactions: jackluke
It’s a chrome extension called dark reader. I have included the trackpad patch for the 5,2 but not in the installer.
Oh, I see. But what about the prelinked kernel - I think it would be safe to include it for MB4,1 and "one size fits all" for all newer machines as well. I tested it on various platforms - all worked!

Attached is a pic with alpha channel. Should work now on all backgrounds...
 

Attachments

  • MB4,1_5,1.png
    MB4,1_5,1.png
    14.9 KB · Views: 215
  • Like
Reactions: jackluke
Oh, I see. But what about the prelinked kernel - I think it would be safe to include it for MB4,1 and "one size fits all" for all newer machines as well. I tested it on various platforms - all worked!

Attached is a pic with alpha channel. Should work now on all backgrounds...

Your prelinkedkernel for MB4,1 (18,4 MB) , even with many refused kext, does allow booting Mojave Installer 10.14.3 from a MBP 2012 (Metal supported with XHCI USB 3.0 host controller). Keyboard, trackpad, wifi, DiskUtility, Terminal and probably a clean Installation work correctly, so I'd say it's enough safe for almost all machines.
 
  • Like
Reactions: Larsvonhier
It didn't worked from CMD+S cause you had to put first "mount -uw /" and since it's a shell bash you don't need "sudo" command from there.
[doublepost=1552858969][/doublepost]

You can update from both ways, but I'd say waiting for the final release of an update through dosdude1's Patcher is safer for unsupported machines.

Okay, that makes sense. Since you told me that trick everything seems to work kinda fine. Only thing that does not work like it did before is the Brithness mode with F1 and F2 as key. I used the programm "Brightness Slider" since people recommended it here from that Forum thread here : https://forums.macrumors.com/thread...ntel-integrated-gpu-efi-variable-fix.2037591/

So any possibility to "fix" this? If not thats a thing that I can easy care less but its nice to have. :D

Anyway, thanks a lot again to you guys for your help and patience with all the people that keep asking including me! :)
 
  • Like
Reactions: jackluke
Okay, that makes sense. Since you told me that trick everything seems to work kinda fine. Only thing that does not work like it did before is the Brithness mode with F1 and F2 as key. I used the programm "Brightness Slider" since people recommended it here from that Forum thread here : https://forums.macrumors.com/thread...ntel-integrated-gpu-efi-variable-fix.2037591/

So any possibility to "fix" this? If not thats a thing that I can easy care less but its nice to have. :D

Anyway, thanks a lot again to you guys for your help and patience with all the people that keep asking including me! :)

Unluckily I don't know how to fix native Brightness on dual gpu mac, the trick I knew was when Mojave becomes un-bootable or trying safe mode or avoiding the OpenGL driver (that is an earlier ASentientBot workaround).
 
It’s a chrome extension called dark reader. I have included the trackpad patch for the 5,2 but not in the installer.

I LOVE DARK READER!!! My brain would be more mushier without it. I have it on all my browsers. Glad people are catching on that dark pages is good. Bright white is not.
 
Last edited:
I LOVE DARK READER!!! My brain would be more mushier without it. I have it on all my browsers. Glad people are catching on that dark pages is good. Bright white is not.
I have used Mojave since August and dark reader since probably a bit before that. I’ve been using Chrome Canary and for the past few months they’ve had a dark mode. It’s now in stable Chrome too. I jailbroke my iPad a few weeks ago and I’ve been using a dark mode on there too. It looks amazing. I use a dark mode on sublime text and iterm for my bash scripting development. Dark mode on github too btw.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.