macOS 10.15 Catalina on Unsupported Macs

[K two]

Yet another strong motivation to upgrade and employ an old unsupported Mac patched by @dosdude1 or 0403979.

AFP reports - "Some factories across China are using forced Uighur labour under a state-sponsored labour transfer scheme that is tainting the global supply chain."

The brands, included "Apple, BMW, Gap, Huawei, Nike, Samsung, Sony and Volkswagen."

 

[K two]

Yet another strong motivation to upgrade and employ an old unsupported Mac patched by @dosdude1 or 0403979.

AFP reports - "Some factories across China are using forced Uighur labour under a state-sponsored labour transfer scheme that is tainting the global supply chain."

The brands, included "Apple, BMW, Gap, Huawei, Nike, Samsung, Sony and Volkswagen."

Sorry for the double post! 😬

 

[haralds]

The only I noticed running 10.15.4 Dev beta 3 compared to running 10.15.3 is that 10.15.4 dev beta 3 uses more memory and my CPU runs about 20 degrees hot on 10.15.3 it would run at 95 degrees F and on 10.14.4 dev beta 3 it runs at 114 degrees F

I "ran out of Application memory" with a 48GB RAM 5,1 due to a memory leaking Safari on DP3. Looking to upgrade to DP4 when it drops despite that being a bit of a PITA. Meanwhile killing Safari fixes it.

 

[Pluda]

I "ran out of Application memory" with a 48GB RAM 5,1 due to a memory leaking Safari on DP3. Looking to upgrade to DP4 when it drops despite that being a bit of a PITA. Meanwhile killing Safari fixes it.

I did not ran out of memory, but... safari is turning on my fans and drainig my battery. I’m developing a wordpress theme and on the process need to use the developer tolls. On firefox or chrome it works like a charm, in my ipad or on my mobile android phone also ok, in safari (desktop) it is a pain... it was on 10.15.0, 10.15.1 and now on 10.15.3

 

[ASentientBot]

Just crossed my mind: has anyone tried booting Cat on Penryn with a stock telemetry plugin and MouSSE? Purely academic of course, but this has been a mystery for almost two years.

@jackluke 👆

 

[jackluke]

Just crossed my mind: has anyone tried booting Cat on Penryn with a stock telemetry plugin and MouSSE? Purely academic of course, but this has been a mystery for almost two years.

@jackluke 👆

Tried, on C2D using the stock 10.15.4 telemetry plugin and MouSSE in kextcache but still the same kernel panic before loginUI.

kextstat | grep -v com.apple

Index Refs Address            Size       Wired      Name (Version) UUID <Linked Against>
   40    0 0xffffff7f82921000 0x5000     0x5000     AAA.LoadEarly.MouSSE (0.38)
7DDA536B-57F7-33B2-A5DC-7ABB923F9C59 <8 6 5 3 2 1>


sysctl -a | grep machdep.cpu.features

machdep.cpu.features: FPU VME DE PSE TSC MSR PAE MCE CX8 APIC SEP MTRR
PGE MCA CMOV PAT PSE36 CLFSH DS ACPI MMX FXSR SSE SSE2 SS HTT TM PBE
SSE3 DTES64 MON DSCPL VMX SMX EST TM2 SSSE3 CX16 TPR PDCM SSE4.1 XSAVE

With MouSSE kext from Catalina verbose getting many errors "duplicate job" and the same kp, removing telemetry anyway the kext is loaded but no SSE4.2 is added to Core2Duo features, but I'd exclude that telemetry since Mojave panicked for SSE4.2 instruction reason.

As you know using the HighSierra telemetry will work on Catalina.

Another my non-clever test that I did in Mojave too, you can even delete all the UserEventPlugins except the com.apple.remoted.plugin and Catalina still boot and work correctly.
Not sure for what is needed but it's the only required plugin.

 

[K two]

Apparently, supported Mac owners are experiencing constant crashing via the "BAD MAGIC" bug in Catalina. The thread is 22 pages long, here - https://discussions.apple.com/thread/250826263?page=22. Any "unsupported" Macs seeing this? Three different unsupported Mac models on 10.15.3 NO PROBLEMS AT ALL. YMMV 😎

 

[Larsvonhier]

Apparently, supported Mac owners are experiencing constant crashing via the "BAD MAGIC" bug in Catalina. The thread is 22 pages long, here - https://discussions.apple.com/thread/250826263?page=22. Any "unsupported" Macs seeing this? Three different unsupported Mac models on 10.15.3 NO PROBLEMS AT ALL. YMMV 😎

Sounds more like a hardware related quality problem to me (some users reported that it stopped after Apple changed the mainboard and touchID electronics under warranty). Then, on the other hand, crashes that seem to be triggered every 30min. or so could be some really "bad magic" due to intel system management mode (SMM) not handled correctly. That´s what folks found when they modified their Thinkpad EFI to exclude that spyware...

 

[K two]

Sounds more like a hardware related quality problem to me (some users reported that it stopped after Apple changed the mainboard and touchID electronics under warranty). Then, on the other hand, crashes that seem to be triggered every 30min. or so could be some really "bad magic" due to intel system management mode (SMM) not handled correctly. That´s what folks found when they modified their Thinkpad EFI to exclude that spyware...

Suspect the most recent silent TCC patch from Apple as the trigger???

 

[TimothyR734]

Tried, on C2D using the stock 10.15.4 telemetry plugin and MouSSE in kextcache but still the same kernel panic before loginUI.

kextstat | grep -v com.apple

Index Refs Address            Size       Wired      Name (Version) UUID <Linked Against>
   40    0 0xffffff7f82921000 0x5000     0x5000     AAA.LoadEarly.MouSSE (0.38)
7DDA536B-57F7-33B2-A5DC-7ABB923F9C59 <8 6 5 3 2 1>


sysctl -a | grep machdep.cpu.features

machdep.cpu.features: FPU VME DE PSE TSC MSR PAE MCE CX8 APIC SEP MTRR
PGE MCA CMOV PAT PSE36 CLFSH DS ACPI MMX FXSR SSE SSE2 SS HTT TM PBE
SSE3 DTES64 MON DSCPL VMX SMX EST TM2 SSSE3 CX16 TPR PDCM SSE4.1 XSAVE

With MouSSE kext from Catalina verbose getting many errors "duplicate job" and the same kp, removing telemetry anyway the kext is loaded but no SSE4.2 is added to Core2Duo features, but I'd exclude that telemetry since Mojave panicked for SSE4.2 instruction reason.

As you know using the HighSierra telemetry will work on Catalina.

Another my non-clever test that I did in Mojave too, you can even delete all the UserEventPlugins except the com.apple.remoted.plugin and Catalina still boot and work correctly.
Not sure for what is needed but it's the only required plugin.

[automerge]1583258966[/automerge]
I have installed the AAAMouSSE kext using kext utility since in was released do the macOS Catalina Patcher disable Telemetry plugin

 

[Larsvonhier]

...beta4 out now.
downloading the 2.95GB...

edit: works on MBA 11,7" intel HD3000

 

[jackluke]

View attachment 897192
[automerge]1583258966[/automerge]
I have installed the AAAMouSSE kext using kext utility since in was released do the macOS Catalina Patcher disable Telemetry plugin

Premise: the telemetry plugin is almost unuseful.

From Mojave only the i5-i7 core cpus can use the stock telemetry plugin, of course for Core2Duo machines the macOS Catalina Patcher "disables" the telemetry plugin, the c2d kernel panic using the stock one is not due to AAAMouSSE, but this kext currently it's more useful for those who use AMD Metal GPU.

 

[TimothyR734]

macOS Catalina 10.15.4 Dev Beta 4 is out

 

[jackluke]

...beta4 out now.
downloading the 2.95GB...

Just updated, known OTA methods worked, essentially it's only a matter of keeping AMFI disabled.

edit:
ASentientBot just released a new method to use 10.15.4 with AMFI enabled: https://forums.macrumors.com/threads/macos-10-15-catalina-on-unsupported-macs.2183772/post-28257563

 

[deeveedee]

Did Apple do something funky with the .3 update where they only now support HDMI and DisplayPort audio?

I see that you're still patiently looking for audio solutions for your MP5,1. I'm running Catalina 10.15.3 on my HackPro5,1 without any problems. I was originally running with legacy NVidia graphics / DosDude and switched to a Sapphire Pulse RX580 (which required an in-place install of 10.15.3, removal of the Legacy NVidia kexts and removal of the patch updater prefpane). The upgrade was painless (I performed the upgrade from 10.14.6 and from 10.15.2 while I was testing).

Is your MP5,1 a dual-CPU model? If so, could that explain the audio issue?

 

[ASentientBot]

Just came across a new way of disabling Library Validation that doesn't seem to be known or mentioned online anywhere!

sudo defaults write /Library/Preferences/com.apple.security.libraryvalidation.plist DisableLibraryValidation -bool true

Confirmed that this allows booting to the desktop on a system with my wrappers and without DisableLibraryValidation.kext or amfi_get_out_of_my_way=1. I don't think it offers any benefits (compared to the boot argument anyways, the kext has known issues) currently, but we should keep it in mind for the future. And it's always fun discovering new things!

@dosdude1 @0403979 may find this interesting.

Edit: Searching that command actually brings up one result, so I'm not the first to discover it after all (though I did independently). Figured I should correct that so I don't get accused of stealing! But it's still a useful and little-known piece of info.

I also discovered some other keys, such as Entitlements and AllowUnsafeDynamicLinking in /Library/Preferences/com.apple.security.coderequirements.plist. I have yet to test them.

 

[jackluke]

Just came across a new way of disabling Library Validation that doesn't seem to be known or mentioned online anywhere!

sudo defaults write /Library/Preferences/com.apple.security.libraryvalidation.plist DisableLibraryValidation -bool true

Confirmed that this allows booting to the desktop on a system with my wrappers and without DisableLibraryValidation.kext or amfi_get_out_of_my_way=1. I don't think it offers any benefits (compared to the boot argument anyways, the kext has known issues) currently, but we should keep it in mind for the future. And it's always fun discovering new things!

@dosdude1 @0403979 may find this interesting.

Just tried your new library validation script, to test properly I have re-enabled AMFI both in nvram and Preboot, and while from 10.15.4 beta 4 verbose I still get some warnings about your wrappers signatures, but with AMFI enabled booting with Catalina GUI and video acceleration worked!

Once again amazing work @ASentientBot !

 

[chris1111]

Done Beta 4

and yes its work
Confirm by me
AppleScript
do shell script "defaults write /Library/Preferences/com.apple.security.libraryvalidation.plist DisableLibraryValidation -bool true" with administrator privileges

 

[TimothyR734]

No luck with the script on my iMac 9,1 can't get to my desktop and if I run the amfi get out of my way in the terminal on 10.15.3 Final I won't let me into the desktop

 

[chris1111]

No luck with the script on my iMac 9,1 can't get to my desktop and if I run the amfi get out of my way in the terminal on 10.15.3 Final I won't let me into the desktop

I am using this Applescript on 10.15.4 B4 and its work
do shell script "defaults write /Library/Preferences/com.apple.security.libraryvalidation.plist DisableLibraryValidation -bool true" with administrator privileges

 

[jackluke]

No luck with the script on my iMac 9,1 can't get to my desktop and if I run the amfi get out of my way in the terminal on 10.15.3 Final I won't let me into the desktop

I am using this Applescript on 10.15.4 B4 and its work
do shell script "defaults write /Library/Preferences/com.apple.security.libraryvalidation.plist DisableLibraryValidation -bool true" with administrator privileges

Yes, the new ASentientBot DisableLibraryValidation script worked correctly with AMFI enabled, but some my remarks:

- ASentientBot with his script makes a new plist (non-xml non-plain-text) in binary format, so in case someone complete an OTA update, at the first run of his script in Catalina you should have AMFI disabled in nvram at least, because from single user mode AFAIK you can't write any new plist to a com.apple domain host . (But instead you could use a pre-copied com.apple.libraryvalidation.plist file and copy it in /Library/Preferences/ from single user mode and should work anyway)

- Probably the "non-mac users" on 10.15.4 don't needed to use the nvram "AMFI disable argument" because they had previously installed the DisableLibraryValidation.kext , but as also @ASentientBot and @Larsvonhier noticed, that kext breaks some core apps, so it's not advisable to use, then better to remove it and use the ASentientBot new method and/or "AMFI disable boot-args/kernel flag".

Hence my advise, together with ASentientBot valid and working libraryvalidation script, I'd continue to use too the nvram AMFI disabling way.

For @TimothyR734 (and maybe others who use the OTA updates), skipping an annoying 10.15.3 reinstalling, to disable AMFI on-the-fly simply boot with CMD+S (single user mode) targeting the 10.15.4 volume (or any other OS X volume), and then type (exactly as is with quotes and underscores):

mount -uw /
nvram boot-args="-no_compat_check amfi_get_out_of_my_way=1"
reboot

 

[Larsvonhier]

Need help "debugging":
I ran into the situation (10.15.4b4) that stocks does not work anymore.
After minimal patched installation, it worked. Then either deleting language sets (with monolingual) or applying @fabioroberto ´s CatalinaUI patch got me this outcome when starting stocks (everything else seems to work):

Termination Reason: DYLD, [0x4] Symbol missing
Application Specific Information:
dyld: launch, loading dependent libraries
Dyld Error Message:
Symbol not found: _OBJC_CLASS_$__NSDatePickerCalendarOverlay
Referenced from: /System/Library/PrivateFrameworks/UIKitMacHelper.framework/Versions/A/UIKitMacHelper
Expected in: /System/Library/Frameworks/AppKit.framework/AppKit
in /System/Library/PrivateFrameworks/UIKitMacHelper.framework/Versions/A/UIKitMacHelper

Can some of you guys try to confirm which mod does harm the stocks app?
(I do not want to re-install the current beta now just for that). Thanks!

 

[jackluke]

Need help "debugging":
I ran into the situation (10.15.4b4) that stocks does not work anymore.
After minimal patched installation, it worked. Then either deleting language sets (with monolingual) or applying @fabioroberto ´s CatalinaUI patch got me this outcome when starting stocks (everything else seems to work):

Termination Reason: DYLD, [0x4] Symbol missing
Application Specific Information:
dyld: launch, loading dependent libraries
Dyld Error Message:
Symbol not found: _OBJC_CLASS_$__NSDatePickerCalendarOverlay
Referenced from: /System/Library/PrivateFrameworks/UIKitMacHelper.framework/Versions/A/UIKitMacHelper
Expected in: /System/Library/Frameworks/AppKit.framework/AppKit
in /System/Library/PrivateFrameworks/UIKitMacHelper.framework/Versions/A/UIKitMacHelper

Can some of you guys try to confirm which mod does harm the stocks app?
(I do not want to re-install the current beta now just for that). Thanks!

Stock stocks.app is working correctly on my 4b4 Ota updated (amfi disabled), I have checked my unpatched 4b4 Appkit exec is compiled by apple on "2 March 2020 at 8:30" (as any other 4b4 exec), to avoid a re-install I'd try to replace this attached in yours /S/L/Frameworks/AppKit.framework/Versions/C/
try this unpatched 4b4 attached, maybe it adds that symbol and fixes that dyld error.

Check also if your UIKitMacHelper has a right date modified of "2 march 2020 at 8:30".

 

[Larsvonhier]

The binary of the AppKit framework was the culprit. Now the question is: Who modified it? Monolingual is to be ruled out - it does not replace those things, just deletes localization. Then only the CatalinaUI patch is a suspect. But interestingly: After restoring the AppKit framework binary, the UI patch for the correct corners still works (quick checked spotlight overlay and context dock popups on light backgrounds).
...?
Thanks, @jackluke, for the quick provision of the framework(s)!

 

[0403979]

Yes, the new ASentientBot DisableLibraryValidation script worked correctly with AMFI enabled, but some my remarks:

- ASentientBot with his script makes a new plist (non-xml non-plain-text) in binary format, so in case someone complete an OTA update, at the first run of his script in Catalina you should have AMFI disabled in nvram at least, because from single user mode AFAIK you can't write any new plist to a com.apple domain host . (But instead you could use a pre-copied com.apple.libraryvalidation.plist file and copy it in /Library/Preferences/ from single user mode and should work anyway)

- Probably the "non-mac users" on 10.15.4 don't needed to use the nvram "AMFI disable argument" because they had previously installed the DisableLibraryValidation.kext , but as also @ASentientBot and @Larsvonhier noticed, that kext breaks some core apps, so it's not advisable to use, then better to remove it and use the ASentientBot new method and/or "AMFI disable boot-args/kernel flag".

Hence my advise, together with ASentientBot valid and working libraryvalidation script, I'd continue to use too the nvram AMFI disabling way.

For @TimothyR734 (and maybe others who use the OTA updates), skipping an annoying 10.15.3 reinstalling, to disable AMFI on-the-fly simply boot with CMD+S (single user mode) targeting the 10.15.4 volume (or any other OS X volume), and then type (exactly as is with quotes and underscores):

mount -uw /
nvram boot-args="-no_compat_check amfi_get_out_of_my_way=1"
reboot

Untested but I assume this launchdaemon would work.