MP 1,1-5,1 Mac OS 11.3 has broken support for older Mac Pros

[fireheadman]

sub'd for info willing to test when/if others are needed.

 

[RLTechs1]

So a quick note before I go crash out... I was able to shave some time off on my boot times. I am not sure why Martin's 0.6.9 package had he APFS timeout set to 999 so I changed it back to -1 like I have used all along with this machine. It boots a bit faster and out of 25 reboots It went from 25/25 with the error to 18/25 having the error.

I have no clue why, but I figured I did something, and things changed so I'll report it here in the hopes someone else may know why or have the time to look into it.

 

[Syncretic]

May I know if all your test drivers are in APFS? Or just the macOS drive is APFS, and the extra data drives may be in HFS+ or something else?

I wonder if "the number of hard drive" is the variable. Or, "the number of APFS hard drive (or even container)" is the variable to trigger the problem.

All of my SSDs are APFS. The bootable ones were formatted by their respective OS installers (Mojave/Catalina/BS 11.1/BS 11.3). The data drives were formatted by either Mojave or Catalina, I forget which (possibly a mix of the two, definitely not Big Sur). The two spinners are HFS+. All of the drives are "vanilla," single-purpose drives, no creative partitioning (just whatever defaults Disk Utility used). As noted last night, that didn't seem to make a difference, but I think this notion warrants more scrutiny. I will add that to the list of things I'll try to test. (I don't have much time today.)

Another curiosity that I didn't notice last night: aside from the hangs and their immediate successors, all of my boot-ups were "clean," in that I didn't get any "your system was stopped due to an error"-type messages. Likewise, all of my shutdowns were also "clean," in that they were quick (speed felt normal) and error-free. However, this morning I discovered that most or all (I haven't inventoried them) of those "clean" shutdowns generated a "shutdown stall" spindump. At this point, I have no idea whether that's related to the boot-time issues or it's an entirely separate problem; it's yet another thing to investigate.

 

[trifero]

Nobody is disrespecting anything, I wanted to highlight that you are injecting code and components you build yourself from source and offer it to people as a package installer. That's all. I personally believe this is not the right approach. Acidanthera provides a very comprehensive manual for configuring OpenCore. To get the most out of OpenCore, it is important to read the manual and understand its settings, not use a system that does everything for end-user without understanding anything behind the scenes. I believe @vit9696 shares also this philosophy.

Feel free to reach at me with a PM, if you want to talk more about it.

I agree. What a terrible answer. He don´t understand the purpose of this forum.

 

[JohnD]

So a quick note before I go crash out... I was able to shave some time off on my boot times. I am not sure why Martin's 0.6.9 package had he APFS timeout set to 999 so I changed it back to -1 like I have used all along with this machine. It boots a bit faster and out of 25 reboots It went from 25/25 with the error to 18/25 having the error.

The APFS timeout limits the amount of time the SSD has to perform TRIM functions at boot. A -1 value is no limit, and should take longer to boot, not shorter. The value is in microseconds (not seconds, not milliseconds), so Martin's value of 999 is effectively 1 millisecond - I don't think you're making it go any faster than that.

Also, I've always had about a 50% success/fail booting from PCIe storage, so your 18/25 is more in line with normal. Your 25/25 fails is highly unusual. Maybe we should look closer at your system to see what might be causing such a high failure rate, so we can apply inverse logic to help find the root cause.

EDIT: You do have a lot of drives in your signature... are they all powered on at boot? How many are APFS? Copying your signature here for future reference... [cMP 5,1▫️2x Xeon E5620▫️64 GB DDR3 1066 (R) ECC▫️8GB RX580 Pulse OC▫️Sonnet USB 3.2 PCIe▫️1TB WD SSD (OC/macOS)▫️6TB WD HDD (Data)▫️4TB WD HDD (Data)▫️2TB WD HDD (Data)▫️EXT WD 6TB (Data) x2▫️EXT WD 4TB (Backups)▫️EXT WD 6TB (iTunes) x2 JBOD]

 

[Cecco]

With opencore things looked so promising for the cMP after a long dark time when we were at risk to brick our machines with every dot update.
And now this 11.3 update seems to put the nail into the coffin for the cMP. Very sad and very bad timing with the M1X/M2 not yet released.

 

[JohnD]

With opencore things looked so promising for the cMP after a long dark time when we were at risk to brick our machines with every dot update.
And now this 11.3 update seems to put the nail into the coffin for the cMP. Very sad and very bad timing with the M1X/M2 not yet released.

This is helpful to all those trying to solve this issue, how exactly? If you don't have anything useful to contribute, you're just noise in the think tank - please refrain from posting, thanks.

 

[JohnD]

Regarding the 11.3 boot hangs, I have a list of things I want to try (boot-args and others), so I decided that sleep is overrated and I'll get back onto this tonight. I haven't gotten to try a single thing from my list, however, because events took me in another direction. I now offer a few more data points, although I'm not yet certain of their value.

Hello @Syncretic - you inspired me to test your theory. I removed *ALL* drives but my SATA SSD on PCIe running 11.3.1. My record successful reboots to date has been 11x. After removing all drives, I destroyed that record by consistently successfully rebooting 22x before getting a prohibited symbol on the 23rd reboot. Though booting with verbose mode, by the time I looked at the screen, it was a complete unreadable mess. Hope that helps shed more light on the root cause.

UPDATE: After getting that prohibited symbol on the 23rd reboot, I performed many more reboots. I got 7 more successful reboots, then a hang at ethernet initialization. Then 2 more successful reboots, and another hang at ethernet initialization. Then another prohibited symbol, then every reboot after got a prohibited symbol. Seems after the first failure, the condition worsened with every subsequent reboot, to the point of 100% failure.

 

[RLTechs1]

The APFS timeout limits the amount of time the SSD has to perform TRIM functions at boot. A -1 value is no limit, and should take longer to boot, not shorter. The value is in microseconds (not seconds, not milliseconds), so Martin's value of 999 is effectively 1 millisecond - I don't think you're making it go any faster than that.

Also, I've always had about a 50% success/fail booting from PCIe storage, so your 18/25 is more in line with normal. Your 25/25 fails is highly unusual. Maybe we should look closer at your system to see what might be causing such a high failure rate, so we can apply inverse logic to help find the root cause.

EDIT: You do have a lot of drives in your signature... are they all powered on at boot? How many are APFS? Copying your signature here for future reference... [cMP 5,1▫️2x Xeon E5620▫️64 GB DDR3 1066 (R) ECC▫️8GB RX580 Pulse OC▫️Sonnet USB 3.2 PCIe▫️1TB WD SSD (OC/macOS)▫️6TB WD HDD (Data)▫️4TB WD HDD (Data)▫️2TB WD HDD (Data)▫️EXT WD 6TB (Data) x2▫️EXT WD 4TB (Backups)▫️EXT WD 6TB (iTunes) x2 JBOD]

So one the APFS time I do not know why all I know and report is what I'm doing and what I see. That said 1) Glad to know it is microseconds, I don't know where I got it in my head but for some reason I thought it was nanoseconds. 2) I do understand the -1 basically is setting the variable to no limit. My theory on this ws that maybe something does not fire off at all, or within the time limit. Even though there is no KP, something may not have processed or not processed properly. I was hoping by using the -1 it would give that section of the boot process all or none of the time I needed without limiting it if it too say 9999 or 99999. That was the thought process though behind it.

Unless you are counting the drives in the drive sleds as being PCIe storage, I do not have any. May 1TB SSD is in drive bay 1, and formatted APFS. Then there are hard drives (6TB,4TB,2TB) in the other 3 sleds being bay 2,3, & 4, and formatted Mac OS Extended (Journaled) with a GUID Partition Map.

My other drives are all connected via USB3 on the Sonnet PCIe USB3 card. They are all formatted Mac OS Extended (Journaled) with a GUID Partition Map, including the 2x 6TB JBOD Raid. The raid I am using is the built in Apple Raid., Yes I know there are better solution, and I have some at my office for critical data, but for iTunes I figure hell if I lose the raid setup or something goes wrong I can re-downladed it all anyways. And yes all my drives are connected at boot, but the externals don't attach until the USB3 card is active (the login screen).

One huge thing I noticed this morning though before eI left for work. Since I put the USB card in PCIe slot 1 and my RX580in slot 2, my RX580 has been running about 20-30% cooler. ???

Something else I do a lot of people don't. On this rig I share it with my daughter, My wife's cMP she shares with our son. Now I know lots of people share, and I highly doubt that is an issue. But the way my setup in my drive sleds works is...

• Drive Bay 1 - WD 1TB SSD - Boot drive, with my OC and macOS and Apps
• Drive Bay 2 - WD 4 TB HDD - For shared data and projects.
• Drive Bay 3 - WD 6TB HDD - This is my HOME drive, my user profile loads this as my data locked to my account.
• Drive Bay 4 - WD 2TB HDD - My daughter's HOME drive, her user profile loads this as locked to her account.
• External - All the external drives are shared with the raid (iTunes) being shared on the network.

Of course, just in case, there is one admin account on the SSD so if one of the home drives fails I can still get into the OS and fix things. I've had this kind of config since OS X on my Power Mac Dual G4 without any issues. For Catalina I had used DosDude patcher as I did not know about OC. Up until that point with Catalina, I've always had a supported OS on a supported machine. Although for a brief amount of time I did have a PC setup with Iaktos (I think that's the one) hack distro. I used that to bridge the gap for a bit between my dual G5 Mac Pro and my 5,1 I have now.

 

[RLTechs1]

By the way JohnD My success/fail rates I listed last night were the number of times I could log in without seeing the "your computer shut down due to an error" dialog during the login process.

 

[Jigga Beef]

Last week i was able to revert back to 11.2.3 and get my Mac Pro 5,1 with NVME up and running again.

I noticed today the computer was being sluggish and that find was not launching. When I looked at disc utlilty i saw the -Update partition that Big Sur adds to my boot drive. Wierd because I do have Automation OS updates turned on, when I reset my computer I can no longer boot into my NVME boot drive. I get the X screen with the URL /Startup.

I updated to OpenCOre .6.9.0, for the past week I use using 6.8.0 if that matters.

My guess is I somehow by mistake jumped to 11.3.1. No biggie I had CCC cloe of my boot drive so I restart the comp, select the CCC copy driver, takes forever to get up and running with start up apps and not being logged to app etc. Than I get a notification that my OS has been updated and is waiting for restart So I scream NO!

I checked again I don't have any auto-updates turned on for the OS. And of course, I Can't boot into now my CCC clone.

I have 1 more CCC clone left (and all important data backed up) but I am worried to try and boot in with that that it will somehow update itself making it no bootable.

Am I crazy?! I have sent the last 4 hours dealing with this...

 

[RLTechs1]

Last week i was able to revert back to 11.2.3 and get my Mac Pro 5,1 with NVME up and running again.

I noticed today the computer was being sluggish and that find was not launching. When I looked at disc utlilty i saw the -Update partition that Big Sur adds to my boot drive. Wierd because I do have Automation OS updates turned on, when I reset my computer I can no longer boot into my NVME boot drive. I get the X screen with the URL /Startup.

I updated to OpenCOre .6.9.0, for the past week I use using 6.8.0 if that matters.

My guess is I somehow by mistake jumped to 11.3.1. No biggie I had CCC cloe of my boot drive so I restart the comp, select the CCC copy driver, takes forever to get up and running with start up apps and not being logged to app etc. Than I get a notification that my OS has been updated and is waiting for restart So I scream NO!

I checked again I don't have any auto-updates turned on for the OS. And of course, I Can't boot into now my CCC clone.

I have 1 more CCC clone left (and all important data backed up) but I am worried to try and boot in with that that it will somehow update itself making it no bootable.

Am I crazy?! I have sent the last 4 hours dealing with this...

No, you are not crazy. BS 11.3 and 11.3.1 have created some issues which I'm sure with everyone on it we'll all get figured out. I am wondering though what you mean by 1 CCC left? I use CCC to keep a bootable backup and I have never seen a limitation on it. Can you explain further?

 

[xlw108]

thank you all for your tireless work! I tried all suggestions for a week to get my cMP 5.1 OC 0.6.9 to start after 11.3 and 11.3.1 upgrade. It always ended with a black screen. Then i stumbled over another post: „I removed AVVideoCard.kext and AVFrameBuffer.kext and the problem disappeared instantly“ Up and running again . Maybe not the full solution but hopefully a useful hint

 

[Jigga Beef]

No, you are not crazy. BS 11.3 and 11.3.1 have created some issues which I'm sure with everyone on it we'll all get figured out. I am wondering though what you mean by 1 CCC left? I use CCC to keep a bootable backup and I have never seen a limitation on it. Can you explain further?

Sorry I Wrote this half alseep at 1AM.

Everything with CCC was fine, the bootable backup booted up just fine. The issue was I have about a dozen apps that launch at StartUp so the start-up was real slow, Dropbox wanted me to log back in etc. So I that meantime the Mac OS downloaded itself the update (with no notification) then told me ti was complete about 10 mins after boot up.

So once that was done I noticed the finder was frozen ./ wouldn't launch, a computer reset showed me I went from 11.2.3 to 11.3.1.

I am going to try booting to my other bootable backup but with no internet connection so nothing can be automatically downloaded and I can just restore my NVME from a Time Machine backup.

I am just worried that even with auto-updates turned off, they will still somehow be installed like they were on my past 2 drives and make the computer unuseable

 

[RLTechs1]

Sorry I Wrote this half alseep at 1AM.

Everything with CCC was fine, the bootable backup booted up just fine. The issue was I have about a dozen apps that launch at StartUp so the start-up was real slow, Dropbox wanted me to log back in etc. So I that meantime the Mac OS downloaded itself the update (with no notification) then told me ti was complete about 10 mins after boot up.

So once that was done I noticed the finder was frozen ./ wouldn't launch, a computer reset showed me I went from 11.2.3 to 11.3.1.

I am going to try booting to my other bootable backup but with no internet connection so nothing can be automatically downloaded and I can just restore my NVME from a Time Machine backup.

I am just worried that even with auto-updates turned off, they will still somehow be installed like they were on my past 2 drives and make the computer unuseable

If you make sure your LAN is not connected, you should be fine. Make sure both the RJ45 jacks are disconnected as well as the WiFi. I made that mistake once. I had forgot about the WiFi and kept having issues. Anyways once you boot up with no LAN at all Go to 'System Preferences> Software Updates' and make sure the 'Automatically keep my Mac up to date' checkbox is Unchecked.

If it is checked and you uncheck it you will probably get this dialog...

Notice the small not bold print. "System data files and security updates will still be installed." This is what is causing you to auto download and install. Click then 'Turn Off" button, and your Software Update Prerences should look like this...

Now here comes the catch... Most people think 'Ok no checkmark so we are now good to go." - Nope! Apple started a "slipstream" process in 11.2.3 to make sure 'critical' files are installed for security reasons. If you click on the 'Advanced' button you will see this...

Notice that even though you told it no updates and it looks turned off from the main dialog, certain actions are still enabled. I have noticed as well that if you uncheck all of them eventually the very bottom one rechecks itself (no clue where this is controlled). So here is how mine is setup...

For some reason (so far) by having only the one setting checked it will check for updates, but even for security updates it just keeps nagging every once in a while in the Notification Center. So far, it has not auto-installed anything I did not click 'Install' on.

Once you have that done, just as a precaution before hooking back ups to your LAN, I would reboot, and then verify the changes persisted through the reboot and did not reset. If they did not you should be good to go on going back online and this should prevent your system from doing any auto-updates, including App Store apps, without you knowing about it.

Also, make sure once you do a macOS update or security patch, to go back in and set this up again, as it seems to revert to auto update after each patch no matter how small.

 

[Jigga Beef]

If you make sure your LAN is not connected, you should be fine. Make sure both the RJ45 jacks are disconnected as well as the WiFi. I made that mistake once. I had forgot about the WiFi and kept having issues. Anyways once you boot up with no LAN at all Go to 'System Preferences> Software Updates' and make sure the 'Automatically keep my Mac up to date' checkbox is Unchecked.

View attachment 1771880

If it is checked and you uncheck it you will probably get this dialog...

View attachment 1771881

Notice the small not bold print. "System data files and security updates will still be installed." This is what is causing you to auto download and install. Click then 'Turn Off" button, and your Software Update Prerences should look like this...

View attachment 1771882

Now here comes the catch... Most people think 'Ok no checkmark so we are now good to go." - Nope! Apple started a "slipstream" process in 11.2.3 to make sure 'critical' files are installed for security reasons. If you click on the 'Advanced' button you will see this...

View attachment 1771885

Notice that even though you told it no updates and it looks turned off from the main dialog, certain actions are still enabled. I have noticed as well that if you uncheck all of them eventually the very bottom one rechecks itself (no clue where this is controlled). So here is how mine is setup...

View attachment 1771894

For some reason (so far) by having only the one setting checked it will check for updates, but even for security updates it just keeps nagging every once in a while in the Notification Center. So far, it has not auto-installed anything I did not click 'Install' on.

Once you have that done, just as a precaution before hooking back ups to your LAN, I would reboot, and then verify the changes persisted through the reboot and did not reset. If they did not you should be good to go on going back online and this should prevent your system from doing any auto-updates, including App Store apps, without you knowing about it.

Also, make sure once you do a macOS update or security patch, to go back in and set this up again, as it seems to revert to auto update after each patch no matter how small.

Ugh I have checked so far on all 3 of my Boot drives with Big Sur those have all been unchecked. Boot up takes 15 to 20 mins for all of the user LaunchAgents. I still can't get into Finder on any of those 3 versions of Big Sur, it hangs the entire time, force quit and relaunch does not help. Luckily I have path Finder to navigate the folders but I can't even go into Users and turn off the apps to launch at system start. Even with LAN disconnected i can't get finder up and running. I made a Fresh Big Sur 111.2.3 Install on a flash drive right now, hopefully, I can boot into that with no finder issue than re-start in Recovery mode and try and restore from Time Machine from a few days ago before I was having issues.

Now I am trying to get the Big Sur Recovery Mode to boot, I can boot "fine" into a Big Sur Boot drive from the Boot Picker but if I hold down CMD+R the Mojove syle Recovery Mode pops up and won't allow me to store from Time Machine drive (I assume because its Big Sur)

 

[Jigga Beef]

If you make sure your LAN is not connected, you should be fine. Make sure both the RJ45 jacks are disconnected as well as the WiFi. I made that mistake once. I had forgot about the WiFi and kept having issues. Anyways once you boot up with no LAN at all Go to 'System Preferences> Software Updates' and make sure the 'Automatically keep my Mac up to date' checkbox is Unchecked.

View attachment 1771880

If it is checked and you uncheck it you will probably get this dialog...

View attachment 1771881

Notice the small not bold print. "System data files and security updates will still be installed." This is what is causing you to auto download and install. Click then 'Turn Off" button, and your Software Update Prerences should look like this...

View attachment 1771882

Now here comes the catch... Most people think 'Ok no checkmark so we are now good to go." - Nope! Apple started a "slipstream" process in 11.2.3 to make sure 'critical' files are installed for security reasons. If you click on the 'Advanced' button you will see this...

View attachment 1771885

Notice that even though you told it no updates and it looks turned off from the main dialog, certain actions are still enabled. I have noticed as well that if you uncheck all of them eventually the very bottom one rechecks itself (no clue where this is controlled). So here is how mine is setup...

View attachment 1771894

For some reason (so far) by having only the one setting checked it will check for updates, but even for security updates it just keeps nagging every once in a while in the Notification Center. So far, it has not auto-installed anything I did not click 'Install' on.

Once you have that done, just as a precaution before hooking back ups to your LAN, I would reboot, and then verify the changes persisted through the reboot and did not reset. If they did not you should be good to go on going back online and this should prevent your system from doing any auto-updates, including App Store apps, without you knowing about it.

Also, make sure once you do a macOS update or security patch, to go back in and set this up again, as it seems to revert to auto update after each patch no matter how small.

So after all of this...The issue had nothing to do with those Auto-Downloads. As you said they weren't installed since I didn't hit installed. The Finder issue I was having seems to be tied to iCloud. Once I Turned off iCloud one of my BootDrives functions properly again....So doesn't even seem to be OpenCOre Related. Smh thanks for the help though!

 

[RLTechs1]

So after all of this...The issue had nothing to do with those Auto-Downloads. As you said they weren't installed since I didn't hit installed. The Finder issue I was having seems to be tied to iCloud. Once I Turned off iCloud one of my BootDrives functions properly again....So doesn't even seem to be OpenCOre Related. Smh thanks for the help though!

Interesting I will add that to my notes and play with a new user not logged into iCloud to see if there is any difference.

 

[KvR]

Yesterday, by accident, my boot issues vanished. I’ve now booted into 11.3.1 11x times in a row and counting.

Unfortunately I don’t know what was the deciding change. I have three installations on two separate disks I can boot from, Mojave, 11.2.3 and 11.3.1. And a three data disks formatted in APFS.

I was messing around with a graphics card I still needed to flash and I replaced my PCIe WiFI card with an original small mini PCIe card. I previously already had taken out the original card. The graphics card I’ve taken out again. So effectively the only hardware change is the WiFI card.
While I was at it I also upgraded the OCLP. This is something I do regularly, and already did this week, so I don’t now which build I was on. Definitely the latest version.

 

[startergo]

I added these debug arguments with debug versions of Lilu and other Lilu child kexts:

-liludbgall msgbuf=1048576 -nvmefaspm

Then issued:

sudo dmesg | grep -i "Lilu" > ~/Desktop/Log_"$(date '+%Y-%m-%d_%H-%M-%S')".log

I found this in the debug log:

[   11.507578]: AppleNVMe Assert failed: 0 == (status) Exit file: /System/Volumes/Data/SWE/macOS/BuildRoots/577555a5c2/Library/Caches/com.apple.xbs/Sources/IONVMeFamily/IONVMeFamily-557.120.1/Common/IONVMeController.cpp Lilu     iokit: @ (DBG) getOSData vendor-id was not found
[   61.545687]: Lilu     nvram: @ (DBG) read oem-product is missing
[   61.646676]: NVMeFix    quirks: @ (DBG) Failed to find LiluVendorGuid:oem-product
[   61.747672]: Lilu     nvram: @ (DBG) read oem-vendor is missing
[   61.848656]: NVMeFix    quirks: @ (DBG) Failed to find LiluVendorGuid:oem-vendor
[   61.949657]: Lilu     nvram: @ (DBG) read oem-board is missing
[   62.050649]: NVMeFix    quirks: @ (DBG) Failed to find LiluVendorGuid:oem-board

There is a lot of vendor-id was not found. I wonder if this is a dealbreaker @Syncretic ?

 

[eVasilis]

I added these debug arguments with debug versions of Lilu and other Lilu child kexts:

-liludbgall msgbuf=1048576 -nvmefaspm liludelay=100

Then issued:

sudo dmesg | grep -i "Lilu" > ~/Desktop/Log_"$(date '+%Y-%m-%d_%H-%M-%S')".log

I found this in the debug log:

[   11.507578]: AppleNVMe Assert failed: 0 == (status) Exit file: /System/Volumes/Data/SWE/macOS/BuildRoots/577555a5c2/Library/Caches/com.apple.xbs/Sources/IONVMeFamily/IONVMeFamily-557.120.1/Common/IONVMeController.cpp Lilu     iokit: @ (DBG) getOSData vendor-id was not found
[   61.545687]: Lilu     nvram: @ (DBG) read oem-product is missing
[   61.646676]: NVMeFix    quirks: @ (DBG) Failed to find LiluVendorGuid:oem-product
[   61.747672]: Lilu     nvram: @ (DBG) read oem-vendor is missing
[   61.848656]: NVMeFix    quirks: @ (DBG) Failed to find LiluVendorGuid:oem-vendor
[   61.949657]: Lilu     nvram: @ (DBG) read oem-board is missing
[   62.050649]: NVMeFix    quirks: @ (DBG) Failed to find LiluVendorGuid:oem-board

There is a lot of vendor-id was not found. I wonder if this is a dealbreaker @Syncretic ?

removing lilu would mean loss of acceleration but have you tried it?

 

[Syncretic]

Does anyone have a quick and/or straightforward way to replace system kexts on Big Sur 11.3? I've never tinkered with the sealed volume stuff, so I don't know how big a can of worms that is.

I've spent the past two days digging through the MacOS kernel via both the two-Mac debugger and the source code/disassemblies. I've found several things that influence the boot hangs, but no actual fixes yet. I've tried dozens of undocumented boot-args, with countless reboots. I'm exhausted.

My two main areas of focus at the moment are /S/L/E/AppleSMC.kext and /S/L/E/IOAHCIFamily.kext/Contents/PlugIns/IOAHCIBlockStorage.kext. Those are the two I'd like to try replacing with the ones from 11.2.x, separately and/or together (and possibly including the parent /S/L/E/IOAHCIFamily.kext/ itself).

The "hang" we're seeing is MacOS still in single-thread mode, with the thread in an idle loop, waiting for something to happen - and nothing ever does. The players at that point (aside from the kernel itself) are AppleSMC, IOAHCIBlockStorage, and IONVMeFamily (and sometimes IOUSB*/AppleUSB*). I've tried altering timeouts, debug flags, and other flags via their respective boot-args, but while I can improve the situation somewhat, I don't yet have a real solution (or a complete understanding of what's going wrong).

A surprising amount of stuff has changed between dot-revs. It looks like they either changed compilers or changed the settings, because some functions that were inlined are now standalone, while others that were standalone are inlined; there are also some functional changes to the kernel startup code that I'm still puzzling through.

If you can point me toward a simplified guide to replacing kexts in 11.3, I'd appreciate it. If you're inclined to try replacing those kexts yourself, go for it - aside from whatever headaches the sealed volume creates, I don't think the risk is very high.

 

[khronokernel]

Does anyone have a quick and/or straightforward way to replace system kexts on Big Sur 11.3? I've never tinkered with the sealed volume stuff, so I don't know how big a can of worms that is.

With OpenCore Legacy Patcher, this is how we mount the root volume for adding legacy graphics acceleration (note SIP and SBM need to be disabled, SecureBootModel is only for OpenCore users):

# Find your System volume
diskutil list

# From the below list, we can see our System volume is disk5s5
/dev/disk5 (synthesized):
   #:                       TYPE NAME                    SIZE       IDENTIFIER
   0:      APFS Container Scheme -                      +255.7 GB   disk5
                                 Physical Store disk4s2
   1:                APFS Volume ⁨Big Sur HD - Data⁩       122.5 GB   disk5s1
   2:                APFS Volume ⁨Preboot⁩                 309.4 MB   disk5s2
   3:                APFS Volume ⁨Recovery⁩                887.8 MB   disk5s3
   4:                APFS Volume ⁨VM⁩                      1.1 MB     disk5s4
   5:                APFS Volume ⁨Big Sur HD⁩              16.2 GB    disk5s5
   6:              APFS Snapshot ⁨com.apple.os.update-...⁩ 16.2 GB    disk5s5s

# Mount the drive(ie. disk5s5)
sudo mount -o nobrowse -t apfs  /dev/disk5s5 /System/Volumes/Update/mnt1

# Remove old kext
sudo rm -R /System/Volumes/Update/mnt1/System/Library/Extensions/Example.kext

# Copy over your kext
sudo cp -R ~/Desktop/Example.kext /System/Volumes/Update/mnt1/System/Library/Extensions/

# Fix permissions
sudo chmod -Rf 755 /System/Volumes/Update/mnt1/System/Library/Extensions/Example.kext
sudo chown -Rf root:wheel /System/Volumes/Update/mnt1/System/Library/Extensions/Example.kext

# Rebuild kernel cache
sudo kmutil install --volume-root /System/Volumes/Update/mnt1/ --update-all

# Create new bootable snapshot
sudo bless --folder /System/Volumes/Update/mnt1/System/Library/CoreServices --bootefi --create-snapshot

Note that we simple use /System/Volumes/Update/mnt1/ as that's what Apple's updater uses as well. Theoretically it can be mounted anywhere empty. Hope this clarifies how to edit the root volume on Big Sur! For SIP, I usually just flip everything but AppleInternal so would be csr-active-config | data | EF0F0000

 

[startergo]

AppleInternal so would be csr-active-config | data | EF0F0000

Mine is

nvram csr-active-config
csr-active-config    %7f%08%00%00
csrutil status
System Integrity Protection status: disabled.
csrutil authenticated-root status
Authenticated Root status: disabled

This is the result from:

csrutil disable
csrutil authenticated-root disable

Isn't that enough?

 

[khronokernel]

Isn't that enough?

Oh should be, EF0F0000 is just me being lazy and not interested in calculating the correct mask. For reference:

0x87F:

- CSR_ALLOW_UNTRUSTED_KEXTS                True
- CSR_ALLOW_UNRESTRICTED_FS                True
- CSR_ALLOW_TASK_FOR_PID                   True
- CSR_ALLOW_KERNEL_DEBUGGER                True
- CSR_ALLOW_APPLE_INTERNAL                 True
- CSR_ALLOW_UNRESTRICTED_DTRACE            True
- CSR_ALLOW_UNRESTRICTED_NVRAM             True
- CSR_ALLOW_DEVICE_CONFIGURATION           False
- CSR_ALLOW_ANY_RECOVERY_OS                False
- CSR_ALLOW_UNAPPROVED_KEXTS               False
- CSR_ALLOW_EXECUTABLE_POLICY_OVERRIDE     False
- CSR_ALLOW_UNAUTHENTICATED_ROOT           True

0xFEF:

- CSR_ALLOW_UNTRUSTED_KEXTS                True
- CSR_ALLOW_UNRESTRICTED_FS                True
- CSR_ALLOW_TASK_FOR_PID                   True
- CSR_ALLOW_KERNEL_DEBUGGER                True
- CSR_ALLOW_APPLE_INTERNAL                 False
- CSR_ALLOW_UNRESTRICTED_DTRACE            True
- CSR_ALLOW_UNRESTRICTED_NVRAM             True
- CSR_ALLOW_DEVICE_CONFIGURATION           True
- CSR_ALLOW_ANY_RECOVERY_OS                True
- CSR_ALLOW_UNAPPROVED_KEXTS               True
- CSR_ALLOW_EXECUTABLE_POLICY_OVERRIDE     True
- CSR_ALLOW_UNAUTHENTICATED_ROOT           True

As you can see, no relevant args are changed between the 2 SIP values, either will work