Stuck with boot of Sonoma on [EB|LOG:EXITBS:START] on AMD Ryzen 7700 and MSI x670e gaming plus wifi

[Andy_rwalker]

Hello everyone,

I apologize if this has been already discussed else where, I did look and tried all possible combinations. Yet I am still stuck.

I understand that [EB|LOG:EXITBS:START] has something to do with kernel patches or quirks however OpenCore boot logs aren't helpful even with dev version of opencore.

Obviously I should be missing or overlooking something but I cannot spot what exactly is that.

Hoping maybe you can guide me what direction to look at or what is wrong in my config.

 

Originally I had Sonoma working using OpenCore with my Intel CPU. Which I upgraded to AMD and modified EFI and config.plist accordingly.

Now trying to boot it with modified configuration.

I identified and whitelisted MMIO regions.

Built USB kext using USBToobox for my MB.

Built SSDT-PLUG-ALT for my MB using SSDTTime.

 

The rest if generic. Based on Dortania, AMD-Vanilla and dozen of other guides.

 

Thank you

opencore-2025-08-24-141453.txt config.plist EFI (2).zip

[strangeron]

config.plist

Try resetting NVRam. I think the kernel/patch is faulty.

[fabiosun]

@Andy_rwalker

your EFI is a disaster

MMIO are all reassigned and this is not good

CPUR and Plug-alt ssdt? only one of these two

 

your motherboard should have this lan:

Realtek® RTL8125BG 2.5Gbps LAN

you use an old kext for intel

 

Initial hang is due to your MMIO all reassigned i suggest to disable all them for now and use these quirk combinations (only to see some steps forward)

 

 

 

which GPU are you using then?

 

[Andy_rwalker]

20 hours ago, strangeron said:

config.plist 88.34 kB · 1 download

Try resetting NVRam. I think the kernel/patch is faulty.

Thank you Strangeron.

Many times I did reset NVRam. Just reseting it itself didn't help. Because underlaying issue was not resolved.

[Andy_rwalker]

20 hours ago, fabiosun said:

@Andy_rwalker

your EFI is a disaster

MMIO are all reassigned and this is not good

CPUR and Plug-alt ssdt? only one of these two

 

your motherboard should have this lan:

Realtek® RTL8125BG 2.5Gbps LAN

you use an old kext for intel

 

Initial hang is due to your MMIO all reassigned i suggest to disable all them for now and use these quirk combinations (only to see some steps forward)

 

 

 

which GPU are you using then?

 

Thank you FabioSun,

That moved it forward indeed.

I deactivated MMIO regions as you suggested and my booter quirks were already set as you suggested.

After that, following NVRam reset, just in case, visually it moved much further. 

Unfortunately I don't see this progress in OpenCore log files, because, I guess control was handed over to MacOS booter and hence not logged by OpenCore.

And I couldn't see it in Apple log files because I didn't manage to boot Apple.

 

But may I ask you before moving forward?

What do you mean by reassigned MMIO?

I thought procedure: is boot dev version of OpenCore, check for MMIO regions which has "skip 0", whitelist them, converting address from hex to dec.

This is what I did.

What I did wrong and how this should be done properly?

 

Now last event which I see on the screen is "Successfully initialized AHCI controller" (see attached screenshot 091) until and turned to support.apple.com/mac/startup abrakadabra (0919).

 

I have RX6600 as a GPU. That's why I have agdpmod=pikera in boot-args. It was working like this before with Inter based built.

 

In regards to SSDT. I suppose it should be both CPUR and Plug-alt. For some reason I had it only CPUR enabled. I will enable both and give it another try. (Also disabled AppleIntel210 kext)

 

I agree, my EFI looks like complete rubbish. This is all result of using various different pre-built EFIs from the Internet. I'll clean it up. I was thinking that it should boot even without wifi or usb or audio support at first.

 

But what is show stopper now?

It is not obvious from the screenshots.

 

Thank you

[fabiosun]

2 hours ago, Andy_rwalker said:

I thought procedure: is boot dev version of OpenCore, check for MMIO regions which has "skip 0", whitelist them, converting address from hex to dec.

This is what I did.

What I did wrong and how this should be done properly?

you have converted and reassigned all of them and system in this way acts as you have in config.plist devirtualizeMMIo disabled

The correct procedure you did is to activate an Opencore debug log txt files

from them you have to reassign only one (to start) MMIO area (reassign means to have it as skip 1 in log)

about MMIO if you have time you can read here to learn a bit:

https://macos86.it/topic/5511-let-talk-aboutdevirtualise-mmio-quirk-and-mmio-whitelist/#comment-125430

 

What i did was to disable MMIO area using DevirtualizeMMIO and DisableVariableWrite quirk (this is not advisable from "gurus" but it works and you can see always progress in your booting part when you are not so skilled to do a correct whitelist for your System

Error i see now could be about wrong USB mapping (screen character corruption and stop sign you see)

I advice to try to map it or change  USB port (if you are booting from an EFI on an USB stick (or disk)

 

2 hours ago, Andy_rwalker said:

n regards to SSDT. I suppose it should be both CPUR and Plug-alt. For some reason I had it only CPUR enabled. I will enable both and give it another try. (Also disabled AppleIntel210 kext)

 

I advice to use SSDTime from CorpNewt github

give it your ACPI folder (calculated with SYSReport option in your config.plist)

then minimal SSDT from it:

SSDT-EC

SSDT-USBX

SSDT-PLUG-ALT

SSDT reset (not mandatory but it could help)

 

For your Lan search in this forum for Lucy lext (i dont remember exactly the name)

 

 

[Andy_rwalker]

5 hours ago, fabiosun said:

you have converted and reassigned all of them and system in this way acts as you have in config.plist devirtualizeMMIo disabled

The correct procedure you did is to activate an Opencore debug log txt files

from them you have to reassign only one (to start) MMIO area (reassign means to have it as skip 1 in log)

about MMIO if you have time you can read here to learn a bit:

https://macos86.it/topic/5511-let-talk-aboutdevirtualise-mmio-quirk-and-mmio-whitelist/#comment-125430

 

What i did was to disable MMIO area using DevirtualizeMMIO and DisableVariableWrite quirk (this is not advisable from "gurus" but it works and you can see always progress in your booting part when you are not so skilled to do a correct whitelist for your System

Error i see now could be about wrong USB mapping (screen character corruption and stop sign you see)

I advice to try to map it or change  USB port (if you are booting from an EFI on an USB stick (or disk)

 

 

I advice to use SSDTime from CorpNewt github

give it your ACPI folder (calculated with SYSReport option in your config.plist)

then minimal SSDT from it:

SSDT-EC

SSDT-USBX

SSDT-PLUG-ALT

SSDT reset (not mandatory but it could help)

 

For your Lan search in this forum for Lucy lext (i dont remember exactly the name)

 

 

Well, I rebuilt SSDTs using SSDTime and mapped them in config.plist

 

Mapped USB using USBToolBox and copied result (UTBMap.kext) as well as USBToolBox.kext to my EFI:

 

All kexts are latest versions.

 

Not 100% sure I did it right, because I do it all under Windows, but I also added my NVME SSD drive, where I have MacOS installed, to Device Properties:

 

Reboot, NVRAM reset, no change.

Still stuck after "Successfully initialized AHCI controller".

 

What else could be wrong?

 

thank you

Appreciate and ideas.

 

[Andy_rwalker]

53 minutes ago, eSaF said:

If I may ask what macOS version are you trying to install?

I am not trying to install new MacOS as such. On this NVMe SSD I have MacOS Sonoma installed. Which I was using with my previous build on Intel CPU and different MB. GPU stays the same.

So I am trying to boot MacOS Sonoma from that disk

[Andy_rwalker]

21 minutes ago, eSaF said:

Ok so as I understand it, the Drive with macOS was used on another build and now you're trying to use it on a different build.

Am I correct with my assumption? If that is the case that is the reason the Boot process stalls or hangs.

 

I am not qualified enough to explain exactly what is happening but I think as the Cache and Boot process was built on another system I feel is the reason for the block.

My assumption of the problem could be completely off or wrong but you should try and build an EFI Folder on the Drive catered for the current system.

 

Hopefully someone more knowledgeable will step forward and completely debunk my theory and offer a more helpful fix.

Whatever happens good luck and here's to hoping for a good outcome.

Your assumption is generally correct. I have an SSD drive with two partitions, one is EFI, another one is where MacOS Sonoma installed. My EFI partition is completely re-built to work properly with new hardware. It has all (as far as I can see) kernel patches, quirks, kexts, new config.plist and so on required to work properly with AMD CPU. Also SSDT files, USB map specific to my new MB. Partition with MacOS installed is not touched. In fact I cannot do much about it because it has Apple file system which is available only under Apple OS.

Edited August 30, 2025 by Andy_rwalker

[fabiosun]

@Andy_rwalker

i think now your EFI is pretty good (post always latest config.plist you used)

MSI x670 and lower had problem with "scramblking devices or loss of some motherboard pcie)

this problem was solved by CaseySJ when he produced some kernel patches for AM5 platform

 

double check your bios

4g and resizebar are ON?

CSM off?

if so i would do better USB mapping fo this MSI (check if you have on there some ASMedia USB and if they are properly supported in OSX)

 

[strangeron]

MSI X670E GAMING PLUS WIFI.zip

Try it with NVRam Reset.

[Andy_rwalker]

1 hour ago, fabiosun said:

@Andy_rwalker

i think now your EFI is pretty good (post always latest config.plist you used)

MSI x670 and lower had problem with "scramblking devices or loss of some motherboard pcie)

this problem was solved by CaseySJ when he produced some kernel patches for AM5 platform

 

double check your bios

4g and resizebar are ON?

CSM off?

if so i would do better USB mapping fo this MSI (check if you have on there some ASMedia USB and if they are properly supported in OSX)

 

Please see attached latest config.plist.

In regards to kernel patches.

Among those I have enabled I have only two from CaseySJ.

 

Both applied successfully:

 

Is that one you are talking about for x670?

 

4g and resizebar are ON and CSM is off.

 

I rebuilt USB mapping. Selected less enabled USB ports (to avoid 15 USB ports limit.). Still the same.

config.plist

[fabiosun]

@Andy_rwalker

try this config based on posted above

do not change nothing else

In boot menu press space and reset Nvram

 

then choose your booting disk

 

PS

I have disabled verbose output (some users have problem with it and 6600 gpu

 

config.plist.zip

Edited August 30, 2025 by fabiosun

[Andy_rwalker]

20 hours ago, strangeron said:

MSI X670E GAMING PLUS WIFI.zip 6.88 MB · 2 downloads

Try it with NVRam Reset.

Thank you Strangeron,

 

I tried your EFI folder.

So you have exactly the same MB. Could be different version of BIOS though. And you have different CPU.

But apparently same number of cores. At least in CPU kernel patch you have proper setting for 8 cores CPU. Like I do. So I didn't have to change it.

 

So, as you said, I simply replaced my entire EFI folder with your one. Nothing changed there. Just did NVRAM reset.

 

Surprisingly I got same issue. Last event "Successfully initialized AHCI controller" but slightly different previous record though.

 

So if this EFI is working on your computer with the same MB and doesn't on mine, then logically what should it be?

BIOS settings??

 

As a next step I am going now revert to my EFI and try config.plist suggested by FabioSun. Will keep you posted.

It's becoming interesting

 

thank you all who's helping

[Andy_rwalker]

15 hours ago, fabiosun said:

@Andy_rwalker

try this config based on posted above

do not change nothing else

In boot menu press space and reset Nvram

 

then choose your booting disk

 

PS

I have disabled verbose output (some users have problem with it and 6600 gpu

 

config.plist.zip 7.15 kB · 2 downloads

 

While using this config.plist (removed -v key in boot-args as far as I can see) boot process stuck at the very beginning. I don't see verbose console. Instead MacOS icon and progress bar down. So it stuck at the very beginning of the progress bar.

 

What I also noticed while playing with EFI folder from @strangeron that changing loaded aml files (SSDT) change the the progress I am able to achieve.

It works for both, my EFI or EFI from strangeron.

If I load no aml files at all I am able to reach up to middle of progress bar (very slowly)

 

The same result I am able to get if I am using my aml files generated for my MB by SSDTTime. Two of them, except of SSDT-PLUG-ALT.aml

 

If I enable again verbose boot it stuck somewhere here:

 

 

So in general, as I said, it is all very interesting but nothing is clear and it looks like fumbling in the dark.

Would be nice if anyone could explain what is happening here and how to fix it.

 

And in regards to BIOS settings, yes.

As far as I can tell, looking at my MSI BIOS 4g and resizebar are ON? and  CSM is off?

 

[fabiosun]

@Andy_rwalker it is from time i cant help an AMD user to have its system up

So we have to find a solution to your problem

Let start from first step to do

BIOS motheboard configuration:

 

1) load bios optimised or default value

2) disable fast boot

3) clear secure boot key (if you have in bios securboot option)

4) other OS as option if available (always in boot section)

5) 4G ON/YEs

6)CSM off/UEFI ON

these are the basics to start in an AMD system 

 

SSDT:

You must use what i said using your ACPI motherboard table/folder to built them with SSDTTime script

 

 

On AM5 MMIO is a very important step to do well, usually user do not understand well that MMIO could be different from also a similar system for different reasons, so we have to calculate them well

this is my MMIO area useful to reassign (now it is disabled because i use below DisableVariableWrite

 

I use a trick for this, and i usually do not advice to do in this way...i do not enable a MMIO whitelist and i use only the quirk useful to skip (for now) them (you can see in my config.plist)

Why this?

because usually i prefer to see some steps forward to have the user happy and not worried about its not working system

 

Kext:

Better start with a minimal kexts list (and this usually are enough to boot from Os Ventura to Tahoe

 

If also in this minimal and safe condition your system hangs always in the same place ( instead i see also other part and it is not relevant if you see a slow apple line progress or a verbose output flowing and hanging )

 

Patches:

i see in some of your config some Nvram declaration of others CPU (7950x)

this CPU has different number of cores than yours and in Nvram it is only a cosmetic parameter to see Cpu name correctly

You can for now skip also restrictevents kext and its parameters (it does the same and other stuff) till you see the screen for inserting your login password

Take the patches from AMD OSX Kernel patches github and use them only changing the first four using 08 instead default 00

I would also disable Visual monotic patches (both) because in my experience no AMD desktop need them but it is your choice 

 

Then others two main parts to check

1)Latest BIOS for different manufacturer as ASUS or Gigabyte need additional ACPI patches, i never seen for now a MSI board needs them (but if you see in your verbose some weird ACPI errors it could be necessary to patch your DSDT

 

2)In the past many AM4 MSI board stopped to work after Monterey OS because they had the same problem we add when amd produced AM5 chipset

CaseySj created a patch for AM5 and also MSI (AM4/Am5) had benefit from this (visit AMD OSX forum for more detailed explanations)

 

Latest advice

some OSX user with RX 6600 gpu had in the past problem to boot in their OS using verbose output, so also this you have to consider if the hang is always in the same place

 

 

 

 

 

Edited August 31, 2025 by fabiosun

[Andy_rwalker]

On 8/31/2025 at 6:44 PM, fabiosun said:

@Andy_rwalker it is from time i cant help an AMD user to have its system up

So we have to find a solution to your problem

Let start from first step to do

BIOS motheboard configuration:

 

1) load bios optimised or default value

2) disable fast boot

3) clear secure boot key (if you have in bios securboot option)

4) other OS as option if available (always in boot section)

5) 4G ON/YEs

6)CSM off/UEFI ON

these are the basics to start in an AMD system 

 

SSDT:

You must use what i said using your ACPI motherboard table/folder to built them with SSDTTime script

 

 

On AM5 MMIO is a very important step to do well, usually user do not understand well that MMIO could be different from also a similar system for different reasons, so we have to calculate them well

this is my MMIO area useful to reassign (now it is disabled because i use below DisableVariableWrite

 

I use a trick for this, and i usually do not advice to do in this way...i do not enable a MMIO whitelist and i use only the quirk useful to skip (for now) them (you can see in my config.plist)

Why this?

because usually i prefer to see some steps forward to have the user happy and not worried about its not working system

 

Kext:

Better start with a minimal kexts list (and this usually are enough to boot from Os Ventura to Tahoe

 

If also in this minimal and safe condition your system hangs always in the same place ( instead i see also other part and it is not relevant if you see a slow apple line progress or a verbose output flowing and hanging )

 

Patches:

i see in some of your config some Nvram declaration of others CPU (7950x)

this CPU has different number of cores than yours and in Nvram it is only a cosmetic parameter to see Cpu name correctly

You can for now skip also restrictevents kext and its parameters (it does the same and other stuff) till you see the screen for inserting your login password

Take the patches from AMD OSX Kernel patches github and use them only changing the first four using 08 instead default 00

I would also disable Visual monotic patches (both) because in my experience no AMD desktop need them but it is your choice 

 

Then others two main parts to check

1)Latest BIOS for different manufacturer as ASUS or Gigabyte need additional ACPI patches, i never seen for now a MSI board needs them (but if you see in your verbose some weird ACPI errors it could be necessary to patch your DSDT

 

2)In the past many AM4 MSI board stopped to work after Monterey OS because they had the same problem we add when amd produced AM5 chipset

CaseySj created a patch for AM5 and also MSI (AM4/Am5) had benefit from this (visit AMD OSX forum for more detailed explanations)

 

Latest advice

some OSX user with RX 6600 gpu had in the past problem to boot in their OS using verbose output, so also this you have to consider if the hang is always in the same place

 

 

 

 

 

Okay, let's start all over again.

 

Loaded default BIOS configuration and configured as per your advice (what I was able to find in my BIOS). Please see below and double check me.

Boot is set to UEFI mode. (UEFI is bold vs CSM in BIOS Mode)

 

TPM is enabled because I need it in Win11.

4G and re-size is enabled:

 

Integrated Video is disabled:

 

Okay, using 3 SSDT flies.

I have a consistent results. With all 3 SSDTs generated on my computer I stuck at "Successfully initialized AHCI controller"

When using only 2 (excluding SSDT-PLUG-ALT) I always stuck at "CoreAnalyticsHub start completed".

This is regardless if I use full set of kexts loaded or minimum, as you advised.

I tried both with and without verbose mode (-v boot arg) as per your recommendation.

In GUI when boot with all 3 SSDTs it stops right at the beginning. if with only 2 SSDTs it slowly goes until roughly middle of the progress bar and stops there. Never boot it.

 

Please see my last config.plist I played with attached.

 

Can you point me out please where have you seen Nvram declaration of others CPU.

I'll fix it. Cannot find myself. But it is possible. Because I took originally this config.plist from one discussion on amd-osx.com

 

About MMIO I did read that article you mentioned. Yet it is not clear to me which regions should be skipped. Article says "On the indications of the Opencore devs we concentrated on the last 4 areas which are the most substantial at the level of pages to possibly devirtualize" It doesn't explain how to chose it. From trying I figured that I can safely select most bottom 3 and system won't halt. If I select either of first two system will halt on [EB|LOG:EXITBS:START]

In any case neither combination bring me close to the boot. I still will stuck at "Successfully initialized AHCI controller" or "CoreAnalyticsHub start completed" depends on SSDT selection.

 

In OpenCore debug logs I see no ACPI errors. (see log attached)

 

And last point.

I tried to find CaseySj patch for AM5. Are you talking about this one? Bug fix for AM5 new firmware motherboards (DSDT-FIX) - General - MacOS86

I tried one DSDT.aml for (MSI X670e Pro Carbon). It killed my computer. 

 

🤷‍♂️

opencore-2025-09-02-083833.txt

config.plist

Edited September 2, 2025 by Andy_rwalker

[fabiosun]

16 hours ago, Andy_rwalker said:

TPM is enabled because I need it in Win11.

And Mostly secure boot could be a first problem (and sorry never used it but there is a section in your config.plist about SecureBootModel

 

16 hours ago, Andy_rwalker said:

About MMIO I did read that article you mentioned

Here i had done a config.plist to skip this initial problem and if you pass the first exit hang error you are fine with your MMIO for now

16 hours ago, Andy_rwalker said:

I still will stuck at "Successfully initialized AHCI controller"

this is the best result you had by now and SSDT_Plug_Alt is mandatory to boot in its or other compatible code

Also SSDT-EC is mandatory

 

I dont know if you disable Secureboot from bios, if this action could create problem to your windows installed OS but i think problem you have now its there and maybe in a not compatible TPM option (i leave it by default and no problem

maybe you can try to search for it and see:

https://www.google.com/search?q=secureboot+in+opencore&rlz=1C5CHFA_enIT1176IT1176&oq=secureboot+in+opencore&gs_lcrp=EgZjaHJvbWUyBggAEEUYOTIKCAEQABgKGBYYHjIHCAIQABjvBTIHCAMQABjvBTIHCAQQABjvBdIBCDUyNzBqMGo3qAIIsAIB8QU2NSNIy-oAx_EFNjUjSMvqAMc&sourceid=chrome&ie=UTF-8

 

maybe @miliuco could help to solve this

 

Edited September 3, 2025 by fabiosun

[Andy_rwalker]

On 9/3/2025 at 7:58 AM, fabiosun said:

And Mostly secure boot could be a first problem (and sorry never used it but there is a section in your config.plist about SecureBootModel

 

Here i had done a config.plist to skip this initial problem and if you pass the first exit hang error you are fine with your MMIO for now

this is the best result you had by now and SSDT_Plug_Alt is mandatory to boot in its or other compatible code

Also SSDT-EC is mandatory

 

I dont know if you disable Secureboot from bios, if this action could create problem to your windows installed OS but i think problem you have now its there and maybe in a not compatible TPM option (i leave it by default and no problem

maybe you can try to search for it and see:

https://www.google.com/search?q=secureboot+in+opencore&rlz=1C5CHFA_enIT1176IT1176&oq=secureboot+in+opencore&gs_lcrp=EgZjaHJvbWUyBggAEEUYOTIKCAEQABgKGBYYHjIHCAIQABjvBTIHCAMQABjvBTIHCAQQABjvBdIBCDUyNzBqMGo3qAIIsAIB8QU2NSNIy-oAx_EFNjUjSMvqAMc&sourceid=chrome&ie=UTF-8

 

maybe @miliuco could help to solve this

 

Just a follow up on my case.

 

I disabled Secure Boot in BIOS. It is not a requirements for Win11. For that I need TPM. Which I left enabled. But Secure Boot is disabled now. It didn't help. It was still suck at "Successfully initialized AHCI controller" if using SSDT_Plug_Alt.

 

Because it was stuck at that phase and it seems to be related to boot drive or controller. (Dortania guide mention USB, when boot is from USB stick or problem with SATA controller)

I have SATA controller in BIOS set to AHCI mode so this shouldn't be a problem.

I tried kext for (SATA unsupported) - didn't help.

So I copied partition with installed Apple system from my M.2 NVME SSD drive to old school normal SATA3 HDD. Now it is loading much further.

I have a feeling it is almost booted. But due to some issues with file system or apple files it reboots at the very end of the load.

 

My current M.2 NVME SSD drive with apple data is Samsung SSD 980. Dortania guide mention to avoid using Samsung PM981.

It is not exactly what I have but close. Also my current M.2 NVME is PCI ver.3 while my MB support ver.4 and even ver.5 So I ordered new M.2 NVME for this. Not the Samsung one.

Yet to figure out was it stuck because of the disk or controller.

Also this is only apple have problem with it. Because my EFI partition is of the same drive and it booted and works fine but apparently couldn't boot MacOS from it.

 

So for now I'll try to boot MacOS from temporary old HDD while waiting for new M.2 NVME SSD drive to come.

[Andy_rwalker]

Managed to get it boot from temporary HDD.

It is bloody slow vs SSD

As suspected it doesn't see my second NVMe drive where MacOS was originally installed.

Now thinking about it. I was using that NVMe SSD on Intel platform. What is different now is controller. So most probably it is not a NVMe SSD the issue but rather controller or missing or incorrect DeviceProperty configuration. In which case new SSD won't resolve this issue. That's pity   But I'll know for sure when new drive arrives.

Good thing now I can boot MacOS.

I can read its logs and see what exactly it doesn't like.

Right?

 

Any further advices?

 

Thank you all

 

 

[Andy_rwalker]

In case anybody is interested. The issue is solved. Case is closed. Mystery is solved. 😁

My last showstopper at "Successfully initialized AHCI controller" was caused by the fact that storage controller with MacOS partition was unavailable under MacOS. I have MacOS installed on M.2 NVME SSD. My MB has 4 suitable M.2 connectors. 2 of which are supported from CPU and 2 supported from MB. My drive was connected to the port supported by MB (whatever it means). It was visible  from BIOS and from Win11 but apparently MacOS didn't see it. Hence boot started okay but in the moment OpenCore handed over boot process to MacOS boot loader it stuck. Because it couldn't see where to boot from. After I reconnected M.2 NVME SSD to the port supported by CPU everything booted properly.