Jump to content

Can't update Big Sur 11.3 to 11.4


46 posts in this topic

Recommended Posts

Hi guys.

I have a weird problem when trying to upgrade/new-install Big Sur 11.4.

My rig is now running on OC 0.7.0 and  Big Sur 11.3, and everything is working flawlessly.

Yesterday I've tried to update from 11.3 to Big Sur 11.4 but I get panic as soon as the installation reached the Apple logo with the progress bar. It's not a OC problem because I have the exact same behaviour with my OC 0.6.9 EFI when I tested that. I suspected it could be something with my graphic so I changed the CSM-mode in bios from disabled to auto which took me a little bit further showing the progress bar "29 minutes left" on the progress bar before panic. Tested a lot of different Gop-settings under UEFI ---> Output   without success.

Tested with and without WhatEverGreen kext and with different boot arguments too. I prepared a USB installation stick with Big Sur 11.4 and booted on that but get panic as soon as I started the installation. For the moment I'am totally lost so I need some help and guidance.

IMG_0426.jpg

EFI_C422_0_7_0_MacPro7.1-BIOS_3003_RELEASE_FakeSMC.zip

Thanks for replay.

That dosen't make any difference. I've already tried that.

(I use FakeSMC with plugins for working iStat menus)

@obus I quickly scanned your EFI and everything looks good.  In additon to trying @5T33Z0 's recommendation to replace FakeSMC.kext with VirtualSMC.kext (be sure to Reset NVRAM after making the change), you might try changing ACPI>Quirks>ResetLogoStatus from True to False.  I had some strange behavior that I never resolved when this Quirk was True.  After you make the change and reboot, Reset NVRAM before booting macOS.

Edited by tonyx86
4 minutes ago, tonyx86 said:

@obus I quickly scanned your EFI and everything looks good.  In additon to trying @5T33Z0 's recommendation to replace FakeSMC.kext with VirtualSMC.kext (be sure to Reset NVRAM after making the change), you might try changing ACPI>Quirks>ResetLogoStatus from True to False.  I had some strange behavior that I never resolved when this Quirk was True.  After you make the change and reboot, Reset NVRAM before booting macOS.

Reduce your kexts, drivers and kernel flags to a minimum.  Use last known good config.plist.

Posted (edited)

Ok guys I will try what you suggest.

I did a minimal EFI setup (debug version OC 0.7.0) with only Bluetooth and my Samsung 960 Pro NVMe M.2 disk earlier today with no success.

After panic when I was trying install from my USB stick I get this error:

 

Screenshot 2021-06-15 at 22.35.22.png

Edited by obus

@obus Samsung NVMe SSD?  I've had problems with them and don't use Samsung in my hacks any more.  I know many swear by Samsung.   I prefer Western Digital Black.  If all else fails, swap the SSD for something other than Samsung.  See here.  

  • Like 2
4 hours ago, obus said:

Ok guys I will try what you suggest.

I did a minimal EFI setup (debug version OC 0.7.0) with only Bluetooth and my Samsung 960 Pro NVMe M.2 disk earlier today with no success.

After panic when I was trying install from my USB stick I get this error:

 

Screenshot 2021-06-15 at 22.35.22.png

Do you have a booting installation of Big Sur or Catalina?  If so, try installing over top of that if you have at least 40 gigs free.

  • Like 1
1 hour ago, 5T33Z0 said:

Read the entries concerning gIOScreenlockstate: https://dortania.github.io/OpenCore-Install-Guide/troubleshooting/extended/kernel-issues.html#stuck-on-or-near-ioconsoleusers-gioscreenlock-giolockstate-3

 

Which Install Guide did you follow, HEDT > Skylake X?

 

Kernel > Quirks: AppleXcpmXtraMsrs is supposed to be off

I follow HEDT --> Skylake X but this guide is not working without culprit. For example AppleXcpmXtraMsrs has to be true on my rig to work despite the fact my MB has an unlocked MSR0xE2 register. My processor isn't recognised as native by Mac OS.

 

@tonyx86 I'm aware of this problems with Samsung M.2 disk. Tried NVMeFix.kext but always get problems with freeze and panic so I'm not using that. My Samsung 960 Pro NVMe M.2 has worked flawlessly all the time since I bought it 2017. Is there any indications that there is any driver changes in 11.4 regarding NVMe disks? My experience is that problem with graphics could generate funny error like memory and disk error too.

 

@HenryV

I'm not sure I follow you. But if you mean that after trying to upgrade from 11.3 start a new installation on top with my Big Sur USB installation media I've already tried that too.

 

I will now test some other WEG boot fixes and boot arguments. 

Any other suggestions are welcome.

@obus I'm not sure what else to suggest.  Up until a month ago, I had two NVMe SSDs in my rig: a WD Black NVMe SSD for macOS and a Samsung NVMe SSD for Windows.  Both SSDs coexisted in my rig without problems and I would boot macOS or Windows by pressing F9 at boot or by selecting from the OC picker.  I'm not sure what changed (an OC update, a macOS update, a BIOS update), but I began having problems booting macOS.  I removed the Samsung SSD to reduce test variables and my rig started working fine again. I now have two WD SSDs in my rig and it works perfectly.   If all else fails, try a different (non-Samsung) SSD.  

 

Also,  maybe this will help.

 

EDIT:

@obus I noticed that your SSDT-RTC0 patch includes the declaration  External (_SB_.PC00.LPC0.RTC_, DeviceObj), but you don't use this declaration for anything.  Is it possible that your rig's unpatched ACPI has an RTC?  If so, did you declare External RTC with the intent to disable the original RTC (by setting RTC._STA=0)?  See Dortania prebuilt SSDT here.  If you don't figure out the problem, try posting your IORegistryExplorer dump so we can take a look.  Use IORegistryExplorer version 2.1 for the dump.

Edited by tonyx86

Hi @tonyx86

Thank's for your efforts.

The thing is that I can only use one bios (1202) for working thunderbolt with my two Titan Ridge cards, so the only thing that could affect my setup is OC update or Mac OS update.

OC update is not the reason either because I have the same behaviour both with OC 0.6.9 and 0.7.0 so it must be some changes in Big Sur 11.4. My RTC is not a problem. I have problem with broken RTC devices so I need the SSDT-RTC0-Range (even if I named it to just SSDT-RTC0) for booting.

 

 

Spoiler

995754527_Screenshot2021-06-16at16_24_01.thumb.png.902574e0befd6bccb54df14826a7d20f.png

 

@obus I do see a lot of USB issues reported after upgrading to 11.4.  Maybe you should try simplifying your USB devices/map to see if that fixes the problem.  It would be a temporary fix, but might get you looking in the right direction.  Good luck.

Posted (edited)

Hi @yapan4

Just finished a clean install successfully on a old 160 GB intel SSD disk. It seems my Samsung M.2 disk isn't working at all with this new update. I'm not sure yet what the problem is but I will report back after little more testing. It could be disk problem or a memory problem with SMBIOs  Mac Pro7.1, I'm not sure yet. The successful installation was done with iMac Pro1.1 SMBIOS.

Edited by obus
  • Like 2

Glad you solved it. The first screenshot repeatedly points to IONVMeController and I never had any success with Samsung PM981 drives, so my guess would be that your Samsung drive no longer works with 11.4.

You may try with NVMeFix.kext and see if it helps.

 

For what it is worth, I use Toshiba/Kioxia SSDs in my hacks. Consumer-grade XG5 and XG6 work perfectly so far, including with NVMeFix.

Data centre XD5 drives (in M.2 22110 or U.2 format) cause a kernel panic on boot if NVMeFix is present; without NVMeFix they work fine—I suppose that power management by NVMeFix does not play well with Power Loss Protection.

 

 

  • Like 1
13 hours ago, obus said:

I follow HEDT --> Skylake X but this guide is not working without culprit. For example AppleXcpmXtraMsrs has to be true on my rig to work despite the fact my MB has an unlocked MSR0xE2 register. My processor isn't recognised as native by Mac OS.

 

@tonyx86 I'm aware of this problems with Samsung M.2 disk. Tried NVMeFix.kext but always get problems with freeze and panic so I'm not using that. My Samsung 960 Pro NVMe M.2 has worked flawlessly all the time since I bought it 2017. Is there any indications that there is any driver changes in 11.4 regarding NVMe disks? My experience is that problem with graphics could generate funny error like memory and disk error too.

 

@HenryV

I'm not sure I follow you. But if you mean that after trying to upgrade from 11.3 start a new installation on top with my Big Sur USB installation media I've already tried that too.

 

I will now test some other WEG boot fixes and boot arguments. 

Any other suggestions are welcome.

Use SuperDuper to copy your Big Sur install to a new partition and install over the copy.

 

To make sure you don't have a bad installer problem you can download Monterey beta installer and try that.

 

Set your csr-active-config to 67080000 abd that may solve boot and KP problems.

 

  • Like 1

 I have KP and restart all the time now after login screen. This with my installation on my Intel 160 Gb SSD disk. Tested with only one memory and a minimal EFI. No other hardware than my intel disk. Tested different WEG fixes without success. CSM disabled in bios and above 4G enabled. Tested PCIe speed set to 3 gen but nothing seems to stop my KP after login.

Late here in Sweden (00.10) even if its still daylight outside. Will take a nap now and try again tomorrow.      

1 hour ago, HenryV said:

Use SuperDuper to copy your Big Sur install to a new partition and install over the copy.

 

To make sure you don't have a bad installer problem you can download Monterey beta installer and try that.

 

Set your csr-active-config to 67080000 abd that may solve boot and KP problems.

 

If you are using Superduper on Big Sur, make sure it is at least v3.5 beta 1 and up as anything else may not a totally bootable drive.

7 hours ago, obus said:

 I have KP and restart all the time now after login screen.

What kind of kernel panic now?

8 minutes ago, yapan4 said:

What kind of kernel panic now?

Have no idea. Can't see the Apple dump-file because I can't login. My rig is booting mostly fine so OC debug file is saying nothing. It must be something with the new graphics driver Apple introduce in this version. I have no idea how to solve this problem. I've been struggling now for a week.:wallbash: 

@obus

There is no user reports about incompatibility Vega and 11.4 so If you HW works ok in 11.3 and Windows, must work in 11.4 too!

Try complete reset your rig - CMOS clean, reflash BIOS and reset your settings, use Clover instead OC, use 11.5beta or 12 beta instead 11.4...

 

 

Posted (edited)
1 hour ago, yapan4 said:

@obus

There is no user reports about incompatibility Vega and 11.4 so If you HW works ok in 11.3 and Windows, must work in 11.4 too!

Try complete reset your rig - CMOS clean, reflash BIOS and reset your settings, use Clover instead OC, use 11.5beta or 12 beta instead 11.4...

 

 

Thank's @yapan4

I have done this at least ten times. Tested 4 different bios without success. Last crash report is indicating something with HFS so I swapped my HFSPlus driver to to OpenHFSPlus but that dosen't change anything. My rig is booting fine 8 out of ten times but straight after logging in I get KP. Attached is some crash log-files I took out from my 11.4 boot disk, maybe there is one person out there how could draw some conclusions from those, but for me it says nothing. Tested @HenryV suggestion about setting csr-active-config to 67080000 with no joy.

Diagnostic_Reports.zip

Edited by obus
  • Sad 1

@obus Sounds like it's time to revert to your last known working state, thoroughly test to make sure there are no problems and then make incremental changes from there.  Just to recap, potential issues I saw and reported to you were:

  • ACPI>Quirks>ResetLogoStatus: change from True to False
  • SSD: Don't use Samsung
  • RTC patch: your ACPI includes two RTC devices (RTC and RTC0).  Not sure it matters, but Dortania guide's pre-built SSDT disables RTC when injecting RTC0
  • Simplify your USB map (eliminate all but the USB port you require for the installer)

 

I can't speak for everyone else, but my hands are tied without having your original, unpatched ACPI and your IORegistry dump (captured with IORegistryExplorer 2.1). Yeah - it's a lot of work to provide, but think of all the eyes you'll have looking at it to help.

Edited by tonyx86
  • Like 2
Posted (edited)

Hi @tonyx86

 

Here we go!

I will test a little bit further with my original EFI without USBMap.kext (will use one of my four USB 2.0 ports).

Test a new RTC patch. My Samsung disk is not connected.

Back to my original bios 3003. This has AWAC and only RTC0.

 

DSDT.aml

Mac Pro.ioreg

Edited by obus
×
×
  • Create New...