OpenCore General Discussion

[Henties]

On 9/11/2021 at 8:32 PM, 5T33Z0 said:

@Henties Did you disbale the custom entries to the linux bootloader in the config? I think it is not needed anymore. Otherwise I would create an issue in acidantheras bug tracker. That Linux implementation is pretty new.

The custom entry for Linux was merely disabled, however removing it altogether did not resolve the issue I am experiencing with booting into Linux by bypassing Grub. I experience this problem only with my Haswell and Skylake builds, and not with any of the 2 Comet Lake machines. I agree with your notion that this new booting method is a rather new OC. feature, will thus be a bit patient and wait what refinements developers will apply to the codebase further down the line, if anything. Resolving the issue is certainly not pressing at the moment.

 

@Septendre In the current OC 0.7.3 version, OC_scan_allow_fs_ext is not present in

Misc-->Security referred to in your posting, seems you are on the latest OC 0.7.4 beta code, which if that is the case, confirms the notion that the developers are still fine tuning the code around the new Linux access methodoligy. 

 

Edit:

In the interim, and after finally having read the other postings on this topic, I got a bit wiser and now understand @Septendre's contribution, which confused me initially, it actually referred to the value set for Misc-->ScanPolicy and its relationship to the bit mask arriving at the entities which are actually allowed during the scanning process.

 

Greetings Henties

Edited September 14, 2021 by Henties

[Guest 5T33Z0]

Found a Spreedsheet on a german Hackintosh forum for generating the ScanPolicy old school style. Modified it a bit, so it's easier to change values with breaking the formulae included. This is for Numbers and Excel. Pro: If bits are added/changed you can modify the data on your own. Contra: it's so 1980s…

 

 

Enjoy

ScanPolicy_Spreadsheet.zip

Edited September 14, 2021 by 5T33Z0

[MacNB]

4 hours ago, 5T33Z0 said:

Okay, so to conclude the `ProcessorType" episode:

• Fnd your CPU model from this List  (section "Apple Processor Type Information - Processor Types")
• convert it from hex to decimal.
• Paste it in PlatformInfo > Generic > ProcessorType
• Save, Reboot, check "About this Mac" to see if it is correct
• Be happy!

 

Good summary.

 

I use PlistEdit Pro plist editor (for many many years now) and that allows you to enter/edit any Number in Hex format.

So with that editor, there's no need for any conversion as you can copy and paste the hex value from the source avoiding any errors.

 

 

[Guest 5T33Z0]

@MacNB Nice, I didn't know that.

[MacNB]

28 minutes ago, 5T33Z0 said:

Found a Spreedsheet on a german Hackintosh forum for generating the ScanPolicy old school style. Modified it a bit, so it's easier to change values with breaking the formulae included. This is for Numbers and Excel. Pro: If bits are added/changed you can modify the data on your own. Contra: it's so 1980s…

 

 

Enjoy

ScanPolicy_Spreadsheet.zip 735.98 kB · 1 download

 

Nice find.

 

I use the macOS Calculator app in Programmer mode for Field that has bit setting (like the ScanPolicy).

Basically, open the OC Configuration.pdf and go to the section you are interested in changing.

Open Calculator App

Change View to "Programmer" (CMD+3)

Then click each Bit you wish to set the Calculator will show you the new Hex value.

Copy and Paste that hex value into the plist editor.

If the editor cannot handle hex value, you can convert the hex to decimal in the Calculator App by simply clicking on 10 button before copy-n-paste.

 

 

[deeveedee]

5 hours ago, 5T33Z0 said:

Okay, so to conclude the `ProcessorType" episode:

• Fnd your CPU model from this List  (section "Apple Processor Type Information - Processor Types")
• convert it from hex to decimal.
• Paste it in PlatformInfo > Generic > ProcessorType
• Save, Reboot, check "About this Mac" to see if it is correct
• Be happy!

Nice.  I use XCode, so when I type 0x0704, it's automatically converted to decimal.

[Guest 5T33Z0]

I am wondering if anybody is interested and capable of (I am not) to continue this project: https://github.com/rlerdorf/OCSanity

 

It's the source for the Sanity Checker which has not been updated since 0.6.6 unfortunately: https://opencore.slowgeek.com/

[deeveedee]

@5T33Z0 You've got big cojones for mentioning that here.

[Guest 5T33Z0]

46 minutes ago, tonyx86 said:

@5T33Z0 You've got big cojones for mentioning that here.

 

It was a useful resource for troubleshooting configs. I don't know why this would offend anyone. Because the way it works, it's completely based on the Data provided by Dortanias OpenCore Install Guide. From my understanding it basically takes the data for each CPU family directly from the guide and parses it in the backend for the sanity checks. But I think for each version of OpenCore new masks have to be created manually, so that's why I think the work was too tedious in the end.

Edited September 14, 2021 by 5T33Z0

[Avery B]

ocvalidate is the closest thing right now - that just makes sure that the plist is valid. I don't think there is anything which compares the plist with the values from the guide.

[Guest 5T33Z0]

I've added a new method to the OpenCore Update Guide using OCAuxiliaryTools (OCAT). No more headaches and manual fiddling when trying to bring an outdated config.plist up to par with the newest feature-set.

 

[Guest 5T33Z0]

Some decision tree kinda thingy quick start guide to troubleshooting boot issues... work in progress...

[ghost8282]

On 9/14/2021 at 9:07 AM, ghost8282 said:

I'm still unable to update the os without a forced shutdown..

Happened right now with the ota delta update to big sur 11.6.

 

Steps:

1. Downloaded the delta ota update from system software update

2. After finishing the download the system asks to reboot <-- It doesn't reboot, monitor shutdown, virsh is reporting my vm is still in execution <-- Waited for 15 min, I had to force the shutdown (virsh destroy BigSur) and manually start again the vm (virsh start BigSur)

3. Update continues with no problem, reboot (OK)

4. Update continues with no problem, reboot (OK)

 

In general I don't have any issue with shutdown/reboot from within the os.

Any idea on what's going on? @vit9696 do you know what can I check to see what is preventing the os to reboot?

 

It seems a bug in libvirt or qemu (I suspect libvirt).

Not working: qemu 6.1, libvirt 7.6.0/7.7.0

working (same vm configuration, same oc): qemu 5.1, libvirt 6.5.0

 

I opened an issue in libvirt gitlab, hopefully someone will be able to sort it out.

 

So nothing to do with oc, sorry.

[Guest 5T33Z0]

Does somebody have some more insight into what RestrictEvents.kext does exactly? The description on the github page is really vague, yet on the other hand it's used a lot... which I find a bit confusing.

[miliuco]

2 hours ago, 5T33Z0 said:

Does somebody have some more insight into what RestrictEvents.kext does exactly? The description on the github page is really vague, yet on the other hand it's used a lot... which I find a bit confusing.

I find the github description clear:
 

"Lilu Kernel extension for blocking unwanted processes causing compatibility issues on different hardware and unlocking the support for certain features restricted to other hardware. The list of blocks currently includes:

• /System/Library/CoreServices/ExpansionSlotNotification
• /System/Library/CoreServices/MemorySlotNotification

The list of patches currently includes:

• Disabled MacBookAir model memory replacement UI (comes in pair with SystemMemoryStatus = Upgradablequirk).
• Disabled MacPro7,1 PCI Expansion view and RAM view.
• CPU brand string patch for non-Intel CPUs."

Yes, it’ used a lot by those who have MacPro7.1 SMBIOS (like me) or AMD cpus o MacBooAir of soldered ram in SMBIOS. 

[Guest 5T33Z0]

"Unlocking the support for certain features (which?) restricted to other hardware (which hardware?)". To me this sounds as if it does more than what's listed next. It just sounds vague.

 

Otherwise he could have just said, RestrictEvents does the following and then drop the list.

Edited September 20, 2021 by 5T33Z0

[Guest 5T33Z0]

Some features of OpenCore are so brilliant, you only notice it once you need them. Like the MinKernel and MaxKernel Feature in the Kernel Section.

 

[FredWst]

Hi,

 

It makes a long time I didn't post here.

But, i got a new toy   Lenovo Thinkpad T15 Gen 1, all is working except mic sound (need to investigate or help). (fingerprint too but will never work)

 

At the moment Monterey but also works under Big Sur. Major difference is that all PCI device show this error "family specific matching fails"

 

@vit9696

But something strange and unusual I got : see screenshot below.

MacOs is booting without Unlocked CFG and no kernel patch !

Patch CFG Lock (0x3E) impossible with know tools (OcTool, Grub tool or Ru refi)  because EFI's protect. (I thing patched bios is needed and hard tool to flash), but will it make any differences ? I don't think all working fine with about 10 hours of battery life normal using. (no heavy task)

Another thing strange is CPU frequency setting ... 

 

 

 

Fred

Edited September 20, 2021 by FredWst

[deeveedee]

@FredWst If I understood you correctly, you're wondering why your rig boots without AppleCpuPmConfigLock=True and/or AppleXcpmCfgLock=True even though your MSR 0xE2 is locked and you are unable to unlock it in BIOS.  I experience the same behavior with my rig.  Further, when I was using CLOVER, I was unable to boot without CLOVER's KernelPM fix enabled, so the behavior is different with OC than with CLOVER.  Others who have used my posted OC EFIs have indicated that if they don't set AppleXcpmCfgLock=True, their rig freezes under heavy CPU load even though it boots fine.  

If you're looking for a suggestion, I would recommend that you set AppleXcpmCfgLock=True if you can't unlock MSR 0xE2.

[ghost8282]

On 9/18/2021 at 6:26 PM, ghost8282 said:

It seems a bug in libvirt or qemu (I suspect libvirt).

It's not qemu, nor libvirt....it's OVMF...older versions (1 year ago?) work as expected..it's the second time in one year they break something

Going there to see if they can fix it.

[Pavo]

On 9/18/2021 at 12:26 PM, ghost8282 said:

emu

It seems a bug in libvirt or qemu (I suspect libvirt).

Not working: qemu 6.1, libvirt 7.6.0/7.7.0

working (same vm configuration, same oc): qemu 5.1, libvirt 6.5.0

 

I opened an issue in libvirt gitlab, hopefully someone will be able to sort it out.

 

So nothing to do with oc, sorry.

Never had any installation/updating issues using KVM+Qemu even OTA updates have worked fine in all versions of Monterey. Currently on Libvirt 7.3.0 and Qemu 6.0.0 on Unraid 6.10.0-rc1.

[FredWst]

On 9/20/2021 at 2:26 PM, tonyx86 said:

@FredWst If I understood you correctly, you're wondering why your rig boots without AppleCpuPmConfigLock=True and/or AppleXcpmCfgLock=True even though your MSR 0xE2 is locked and you are unable to unlock it in BIOS.  I experience the same behavior with my rig.  Further, when I was using CLOVER, I was unable to boot without CLOVER's KernelPM fix enabled, so the behavior is different with OC than with CLOVER.  Others who have used my posted OC EFIs have indicated that if they don't set AppleXcpmCfgLock=True, their rig freezes under heavy CPU load even though it boots fine.  

If you're looking for a suggestion, I would recommend that you set AppleXcpmCfgLock=True if you can't unlock MSR 0xE2.

 

@tonyx86

I've no issue during heavy load or after many hour of use, even wake up after many days.  (without patch)

I read your post before posting, I activate xclm patch ... but not sure it is correctly working. wait and see if it's better or not.

 

14:567 00:033 OCAK: Read kernel version 21.1.0 (210100)
14:579 00:012 OCAK: Replacing _xcpm_core_scope_msrs data 226 76
14:586 00:006 OCAK: Replacing _xcpm_core_scope_msrs data 226 406416
14:592 00:006 OCAK: Replacing _xcpm_core_scope_msrs data 226 16384
14:599 00:006 OCAK: Not matching _xcpm_core_scope_msrs data 46 422876
14:610 00:011 OCAK: 64-bit XcpmCfgLockRel replace count - 2

 

Fred

[ghost8282]

On 9/20/2021 at 4:24 PM, Pavo said:

Never had any installation/updating issues using KVM+Qemu even OTA updates have worked fine in all versions of Monterey. Currently on Libvirt 7.3.0 and Qemu 6.0.0 on Unraid 6.10.0-rc1.

It's ovmf not qemu, nor libvirt.

Latest working version is 202011 (stable). I suggest to not update OVMF_CODE and VARS.

Unraid uses old versions of ovmf.

Thanks to this I was able to track the culprit to ovmf.

[Septendre]

On 9/12/2021 at 7:55 PM, antuneddu said:

I think your tool is out of date, masks are missing in the ScanPolicy ( OC_SCAN_ALLOW_FS_LINUX_DATA and  OC_SCAN_ALLOW_FS_XBOOTLDR )

However I tried again for further confirmation
With ScanPolicy = 2703107 and also with ScanPolicy = 0   ( of course with Linux drivers enabled and without Entries in Misc    you don t need to add nothing in Misc > Entries if use the driver ext4_x64 and openlinuxboot )   

 

 all systems are detected and started fine , but Windows and Linux are placed in pole position

 

With ScanPolicy 2687747  without any driver , but with the Entries for Windows and Linux declared in Misc

                

 

At boot it is like this, which is what I prefer ... to have osx in pole position

 

 

 

 

 

 

I finally got openlinuxboot working on my rig, after a lot of fiddling, reading and testing. 

With Misc:Security:ScanPolicy set to 16129, UEFI:APFS:EnableJumpstart= true, and UEFI:Drivers:OpenLinuxBoot.efi + Flags=0x0000018F, everything works, and linux boot real quick (without using grub) !

So I moved my USB test EFI to the computer boot disk, and got rid of Clover (after long years of faithful service). 

The only annoying thing left is since Ubuntu doesn't manage the external screen resolution of 2560 x 1440  (it stops at 1080p) when I come back to OSX I have to power off, wait and then power on the external screen so that Catalina have the framebuffer activated with the 2560 x 1440 resolution. Otherwise both screens flick on and off trying to adjust, the internal one showing the desktop (on and off), and the external remaining black. I wonder if there is a NVRAM UUID addressing that ... more reading to do. When I come back from windows 10 everything is fine, probably since Windows manage the screen resolution that I use in OSX.

 

By the way Antuneddu where is your icon set from? They look real nice.

Edited September 21, 2021 by Septendre

[Anto65]

Well, welcome aboard ... you're okay with Linux too 👍
The icons come from myself, every now and then it takes my fancy and I create some sets 🙂  (Backgrounds are resized to your 2560x1440 resolution)

 

 

https://drive.google.com/file/d/1yiuk4G2fnbGIAngP0LhWjhqNPYAXGVeA/view?usp=sharing