Contribute
Register

Hackintosh won't boot after installing nVidia drivers.

Status
Not open for further replies.
Joined
Jan 3, 2019
Messages
7
Motherboard
Tongfang GK5CN6Z Laptop
CPU
i7-8750H
Graphics
UHD 630 + GTX 1060
Mobile Phone
  1. Android
Hi all,

I have macOS High Sierra and it was working very well. I decided to try to enable my dedicated nVidia card (GTX 1060). I unchecked nv_disable=1 and Inject Intel, and checked Inject NvidiaWeb. I also enabled SIP by changing CsrActivateConfig to 0x00. I then restarted, and then downloaded and installed the appropriate web drivers. After restarting, my macOS wouldn't boot any more. I went into Windows and used CMD to mount the EFI and then replace the config.plist with a working backup, but it still won't boot.

I have OsxAptioFix2Drv-free2000.efi and EmuVariableUefi-64.efi installed in my EFI. I also have Lilu and WhateverGreen kexts. I'm not sure what to do at this point.

I'm wondering if the following kext: FakePCIID_Intel_HD_Graphics.kext is interfering with the boot? I tried deleting it in Windows CMD but it just won't let me (even with elevated mode). Please help. :(

Thank you.
 
I also enabled SIP by changing CsrActivateConfig to 0x00
With SIP disabled, when the kernel cache was rebuilt after installing the NVIDIA drivers any unsigned kexts installed to /Library/Extensions will have been excluded.

You can configure Clover on a per-boot basis by pressing 'o' for options at the boot menu e.g. enable the NvidiaWeb function, change graphics injector options, SIP setting etc.

Ensure that you have FakeSMC.kext installed in /EFI/CLOVER/KEXTS/Other, try 'safe boot' with SIP disabled, the kernel cache should get rebuilt with appropriate permissions.
 
With SIP disabled, when the kernel cache was rebuilt after installing the NVIDIA drivers any unsigned kexts installed to /Library/Extensions will have been excluded.

You can configure Clover on a per-boot basis by pressing 'o' for options at the boot menu e.g. enable the NvidiaWeb function, change graphics injector options, SIP setting etc.

Ensure that you have FakeSMC.kext installed in /EFI/CLOVER/KEXTS/Other, try 'safe boot' with SIP disabled, the kernel cache should get rebuilt with appropriate permissions.

Oh thank you very much! Using the Clover Boot Options I removed the nv_disable=1 argument, checked inject nvidia and Web drivers, and that got me into the OS.

However, when checking system information, the system is still recognising the Intel card rather than the nvidia. What can I do to get it recognised, or am I on a fool's errand here?

Also is there a way to uninstall the nvidia Web drivers if I can't get the card to work, so I can avoid boot problems in the future?

Thanks again for helping me get back into my system!

EDIT: just wanted to add that my BIOS doesn't have the option to disable the integrated card so I don't think that route is an option for me.
 
Oh thank you very much! Using the Clover Boot Options I removed the nv_disable=1 argument, checked inject nvidia and Web drivers, and that got me into the OS.
Inject Nvidia should be unchecked.
Also is there a way to uninstall the nvidia Web drivers if I can't get the card to work, so I can avoid boot problems in the future?
In NVIDIA system preferences pane there is an option to uninstall the driver.
just wanted to add that my BIOS doesn't have the option to disable the integrated card so I don't think that route is an option for me.
If you're using a laptop, configuring the discrete card for macOS may not be so straightforward in any case. Maybe use IORegistry Explorer to see what, if any display devices are present and/or drivers loaded on the discrete card and go from there.
 
Inject Nvidia should be unchecked.
In NVIDIA system preferences pane there is an option to uninstall the driver.
If you're using a laptop, configuring the discrete card for macOS may not be so straightforward in any case. Maybe use IORegistry Explorer to see what, if any display devices are present and/or drivers loaded on the discrete card and go from there.

Thanks again for your help. What am I looking for when typing ioreg in the Terminal? I've tried searching for keywords like "Nvidia", "GTX" and "Intel" but got no results.

Thanks.
 
Thanks again for your help. What am I looking for when typing ioreg in the Terminal? I've tried searching for keywords like "Nvidia", "GTX" and "Intel" but got no results.

Thanks.
Look at the IOService tree, you should see your graphics card on one of the PCI ports with or without NVDA-prefixed classes depending on whether NVIDIA drivers are loading. For NVIDIA device, search for 'display'

edit: see also the GUI app IORegistryExplorer, its attached to this post
 
Last edited:
Look at the IOService tree, you should see your graphics card on one of the PCI ports with or without NVDA-prefixed classes depending on whether NVIDIA drivers are loading. For NVIDIA device, search for 'display'

edit: see also the GUI app IORegistryExplorer, its attached to this post

Thanks for the advice. I've just found out that my laptop's graphic card is one of those MUX-less ones.. does that mean I can never get the Nvidia card to work on macOS?

Second question, I know for a fact that my HDMI port is directly connected to the Nvidia card. Does that mean I will never be able to use the HDMI port in macOS?

Thanks again for your patience and help.
 
Thanks for the advice. I've just found out that my laptop's graphic card is one of those MUX-less ones.. does that mean I can never get the Nvidia card to work on macOS?

Second question, I know for a fact that my HDMI port is directly connected to the Nvidia card. Does that mean I will never be able to use the HDMI port in macOS?
I think it is unlikely to work in a straightforward way, whether it will never work in any way i couldn't tell you. I know there is unofficial NVIDIA EGPU support with the web drivers on High Sierra, but on a laptop you may have additional software switching of the device thats not going to happen with the macOS web drivers alone, they are for desktops, the device needs to be enabled at boot.
 
I think it is unlikely to work in a straightforward way, whether it will never work in any way i couldn't tell you. I know there is unofficial NVIDIA EGPU support with the web drivers on High Sierra, but on a laptop you may have additional software switching of the device thats not going to happen with the macOS web drivers alone, they are for desktops, the device needs to be enabled at boot.

Yeah, unfortunately with the way Optimus technology works, it seems very unlikely that I will be able to run the Nvidia card by itself, as it's routed through the Intel card.. I just thought that maybe there is a way to "emulate" that on Mac like there's Bumblebee for Linux. :(
 
Yeah, unfortunately with the way Optimus technology works, it seems very unlikely that I will be able to run the Nvidia card by itself, as it's routed through the Intel card.. I just thought that maybe there is a way to "emulate" that on Mac like there's Bumblebee for Linux. :(
Optimus is known not to work unfortunately, there is plenty of discussion on this forum. If you have Thunderbolt and an enclosure, EGPU is a possibility, recommend supported AMD GPU if you go down that road.
 
Status
Not open for further replies.
Back
Top