Jump to content

Clover General discussion

ErmaC

By ErmaC,
in Clover

 Share
29,904 posts in this topic

Recommended Posts

deeveedee
deeveedee
Posted
Posted

@iCanaro Grazie mille!  Thank you for your quick responses!  Just to be clear, I am (was until my disk corrupted) enabling "SSDT->Generate->PluginType" in my config.plist for both r5122 and r5127.  Are you saying that in addition to enabling "Generate->PluginType", I also need to set PluginType->1 in CLOVER config.plist?  If that is correct, are you saying that this requirement for "Plugin Type -> 1" is new for r5127?

Link to comment
Share on other sites
iCanaro
iCanaro
Posted
Posted
12 minutes ago, tonyx86 said:

Are you saying that in addition to enabling "Generate->PluginType", I also need to set PluginType->1 in CLOVER config.plist?  If that is correct, are you saying that this requirement for "Plugin Type -> 1" is new for r5127?

 

yes it's accurate.
But my advice is to use SSDT x PM that you're using with OC and use the same SSDT for both Clover and OC and let go of plugin types in Clover's config 

  • Like 1
  • Thanks 1
Link to comment
Share on other sites
deeveedee
deeveedee
Posted
Posted (edited)
4 minutes ago, iCanaro said:

 

yes it's accurate.
But my advice is to use SSDT x PM that you're using with OC and use the same SSDT for both Clover and OC and let go of plugin types in Clover's config 

Thank you!  I will test soon after I recover.

cc'ing @MacKonsti

 

@iCanaro I would give the same advice about CLOVER's "Fix SBUS."  Disable "Fix SBUS" in CLOVER's config.plist and use an SSDT for my reasons here.

Edited by tonyx86
Link to comment
Share on other sites
MacKonsti
MacKonsti
Posted
Posted (edited)

Thanks for the tagging @tonyx86

 

So wait @iCanaro as it's not clear to me. Sorry to repeat TonyX86:

You say that r5127 and newer, we need on top of this in config.plist 

		<key>SSDT</key>
		<dict>
			<key>Generate</key>
			<dict>
				<key>CStates</key>
				<false/>
				<key>PStates</key>
				<false/>
				<key>PluginType</key>
				<true/>
			</dict>
		</dict>

...we need this from SSDT-PLUG as found on Dortania guides? 

    Method (PMPM, 4, NotSerialized) {
        If ((Arg2 == Zero)) {
           Return (Buffer (One) { 0x03 })
        }

       Return (Package (0x02) {"plugin-type", One})
    }

so that in IORegistry we see the key "plugin-type" under CPU zero?

So what's the purpose of <key>PluginType</key> then in recent versions of Clover? Is it not taken under consideration at all?

@Slice can you enlighten us please mate?

 

With Clover r5123-1 when I open my IORegistryExplorer (SMBIOS is Macmini8,1) I do see plugin-type as 0x01 just fine, without SSDT-PLUG.AML or whatever the Power Management SSDT AML name...

Edited by MacKonsti
Link to comment
Share on other sites
deeveedee
deeveedee
Posted
Posted (edited)

@MacKonsti Lose the SSDT->PluginType in CLOVER config.plist and patch ACPI with your own SSDT (same as you would for OC).

 

I also recommend that you lose CLOVER's "Fix SBUS" and patch SBUS.BUS0 with your own SSDT.  Soon, all CLOVER users will be following @MaLd0n 's ACPI patching guides and we'll be running with patched DSDTs. :)

Edited by tonyx86
Link to comment
Share on other sites
odemolay
odemolay
Posted
Posted (edited)
5 hours ago, chris1111 said:

No I did not use it

Alright with the XhciPortLimit the boot is like a JET :plane: 

Perfect thanks

do you mean you turn this quirks on true ? if yes for what for? if you  have personal usbkext or usb SSDT. why would we need to turn true this quirk ? 

Edited by odemolay
Link to comment
Share on other sites
deeveedee
deeveedee
Posted
Posted (edited)

@iCanaro - I am now running CLOVER r5127 without SSDT->PluginType and with SSDT-PLUG (attached).  You are correct!!!! The Power Saver settings are now working.  It does appear that CLOVER's SSDT-PluginType is not working correctly and needs to be fixed.  @Slice FYI.

 

Spoiler

186629287_ScreenShot2021-01-07at5_28_02PM.png.03d5540d6494f19f7bf87ba5860f6445.png

 

 @MacKonsti - maybe this solves your problem, too.  I have disabled CLOVER r5127's SSDT->PluginType and r5127's "Fix SBUS" and have replaced these CLOVER settings with SSDTs (attached).  All appears to be running perfectly.

 

@MifJpn - this is the solution to the problem.  This IS a CLOVER r5127 bug that needs to be fixed and it is not a configuration problem. Thank you for your help.

 

*** NOTE: The attached SSDT-PLUG.aml is specific to my ACPI (_SB_.PR00).  You will need to make sure that your SSDT-PLUG patch is suitable for your system's ACPI. ***

 

Attached is the SSDT-PLUG.aml (same as what I use for Open Core) that should replace CLOVER r5127's SSDT->PluginType.  Also attached is SSDT-SBUS.aml that should replace CLOVER r5122 and r5127 "Fix SBUS."

 

SSDT-PLUG.zip

 

SSDT-SBUS.zip

Edited by tonyx86
Added note about SSDT-PLUG compatibility with ACPI
  • Like 3
  • Thanks 1
Link to comment
Share on other sites
iCanaro
iCanaro
Posted
Posted

@MacKonsti using translators, my thoughts are not always fully expressed, specified this, we found that the only plugin type flag does not make power management work properly, this in the most recent versions of Clover (5125/5126 etc...) and I reported the thing in this thread. In order for Clover to handle the PM correctly, type 1 plugins must be enabled in addition to the plugin type flag in generate options. That said, since here many like me use OC and Clover at the same time, I think it's more appropriate for the PM to be managed by an identical functional SSDT for both boot loaders

  • Like 1
Link to comment
Share on other sites
deeveedee
deeveedee
Posted
Posted (edited)

@iCanaro The table below represents my testing with r5122 and r5127.  After I replaced CLOVER r5127's Generate->PluginType with SSDT-PLUG.aml, my Geekbench 5 multi-core performance dropped back down to what I was observing with r5122 here.  Notice that my i7-8700 CPU performance is better with CLOVER r5127 and CLOVER's Generate->PluginType (without "Plugin Type = 1").  If @Slice or someone can figure out how to fix Energy Saver settings and retain the increased CPU performance, that would be great.

 

EDIT: I updated this table with an additional column to show that @iCanaro is correct about enabling "Plugin Type = 1" in the CLOVER r5127 config.plist.  I may have overstepped a bit by claiming this to be an r5127 "bug."  Maybe all previous versions of CLOVER should have required "Plugin Type = 1" and this "bug" is now fixed in r5127.  Regardless, this is a change from r5122 to r5127.

 

 

CLOVER R5122 with 

Generate->PluginType

CLOVER R5127 with 
Generate->PluginType

CLOVER R5127 with 
SSDT-PLUG.aml

CLOVER R5127 with 
Generate->PluginType and Plugin Type = 1

Energy Saver Preferences

5 settings available

2 settings available

5 settings available

5 settings available

GeekBench 5 multi-core benchmark

5500+

5600+

5500+

5500+

Lowest CPU multiplier

13

10

13

13

Edited by tonyx86
Added table column with "Plugin Type = 1"
  • Like 1
Link to comment
Share on other sites
deeveedee
deeveedee
Posted
Posted
5 minutes ago, MifJpn said:

After that, for historical reasons, the developer will decide if it can be fixed.

 

I wouldn't want it any other way.  May the developer choose wisely.

Link to comment
Share on other sites
deeveedee
deeveedee
Posted
Posted

Is there a reason that OC 0.6.3 and 0C 0.6.5 are able to properly inject FakeSMC_CPUSensors.kext (6.26-357-gceb835ea.1800) but CLOVER r5127 is not?  Details below...

 
I still use FakeSMC.kext and FakeSMU_CPUSensors.kext with HWMonitor.app ( (all three are version 6.26-357-gceb835ea.1800).  I know - they're old, but I've never had a reason to change and they have served me well all the way through Catalina with CLOVER r5122 and CLOVER r5127 (kexts installed in /Library/Extensions).  This same FakeSMC/Sensor/HWMonitor combo works fine with Catalina and Big Sur and OpenCore 0.6.3 and 0.6.5 (where no 3rd-party kexts are installed in /L/E).

 

To prepare my Catalina 10.15.7 for upgrade to Big Sur with CLOVER r5127, I moved all of my 3rd-party kexts to E/C/k/O, deleted my kexts from /L/E and rebuilt kext cache.  With kexts in E/C/k/O, no 3rd-party kexts in /L/E and CLOVER r5127, HWMonitor.app no longer shows CPU temps/clocks. I'm guessing that CLOVER r5127 is not properly injecting my old FakeSMC_CPUSensors.kext.

 

I'm guessing that upgrading to a newer version of FakeSMC/Sensors/HWMonitor fixes this, but since these old versions have served me so well and still work fine with OC and Big Sur, is there a reason that r5127 doesn't work with these?

 

Thank you.

 

 

Link to comment
Share on other sites
deeveedee
deeveedee
Posted
Posted
6 hours ago, Slice said:

I don't know what is the problem you encountered with plugin-type. Clover has no related changes.

 

Thank you for taking a look at this.  You've done an amazing job with Clover and it has served me very well.  I appreciate everything that you and the Clover Team have done during this amazing journey. It's been a great ride!  Thank you!

  • Like 2
Link to comment
Share on other sites
deeveedee
deeveedee
Posted
Posted
6 hours ago, MifJpn said:

Hello, tonyx86.

I really feel your enthusiasm for making Clover better.

Thank you, but you are giving me way too much credit. ;)

  • Like 2
Link to comment
Share on other sites
deeveedee
deeveedee
Posted
Posted (edited)

This is my last post about this (I promise - some of you may be almost as relieved as I am).  If you are adventurous, experiment with this AT YOUR OWN RISK.  During testing of my HackMini8,1 while going back and forth between r5122 and r5127 and having the "wrong" SSDT>PluginType setting, I experienced a catastrophic SSD crash requiring erase and restore from TimeMachine.  I cannot say with certainty that the crash was caused by the incorrect power management settings.

 

I am now testing with OC 0.6.5 and the modified SSDT-PLUG.aml attached.  The results are similar to using Clover r5127 with SSDT>Generate>PluginType and without SSDT>PluginType=1.  With this modified SSDT-PLUG, my Geekbench 5 multi-core benchmarks are consistently higher AND my low CPU multipliers are 10, 11 and 12 (not limited to 13).  Note that I lose Energy Saver options, so this is a trade-off.  Again, TEST AT YOUR OWN RISK.

 

Note that my rig is running with HPET DISABLED via ACPI patch (HPET._STA=0 just like a real MacMini8,1).  See my reasoning here.  I'm not sure if this affects these results.  I also use RtcMemoryFixup.kext with custom rtcfx_exclude range (with CLOVER, I did not use "Fix RTC").  I also have both DDR4 memory banks populated (I have found that single-bank performance is less).

 

 

CLOVER R5122 (OpenRuntime)
with Generate->PluginType

CLOVER R5127 (OpenRuntime)
with Generate->PluginType

CLOVER R5127 (OpenRuntime)
with SSDT-PLUG.aml (PluginType=1)

CLOVER R5127 (OpenRuntime)
with Generate->PluginType
and PluginType=1

OC 0.6.5 with PluginType=0

(see modified
SSDT-PLUG.aml

attached)

Energy Saver Preferences

5 settings available

2 settings available

5 settings available

5 settings available

2 settings available

GeekBench 5 multi-core benchmark

5500+

5600+

5500+

5500+

5600+

Lowest CPU multiplier

13

10

13

13

10,11,12,13

 

Geekbench 5 Benchmarks with modified SSDT-PLUG

Spoiler

169289447_ScreenShot2021-01-08at1_18_55PM.thumb.png.ce5c60d610b24d6d309b024e6e1d966a.png

 

 

When running OC 0.6.5, OC correctly injects my FakeSMC_CPUSensors.kext (6.26-357-gceb835ea.1800).  CLOVER r5127 does not properly inject this kext (from E/C/k/O), so I am unable to run HWMonitor.app to show CPU.temps/clocks with r5127 unless I install kexts in /Library/Extensions.  OC does not have this problem.

 

HWMonitor.app showing 10x CPU multiplier

Spoiler

1769407279_ScreenShot2021-01-08at1_29_47PM.png.154a7fe76c543316829a9e9724899ca2.png

 

*** The attached SSDT-PLUG patch is specific to my ACPI (_SB_.PR00).  If you use this patch, make sure it is compatible with your ACPI ***

 

SSDT-PLUG.zip

Edited by tonyx86
Added OpenRuntime and PluginType notes to table
  • Like 3
Link to comment
Share on other sites
Slice
Slice
Posted
Posted
2 hours ago, Hervé said:

It's supposed to fix the briefly garbled Apple logo at startup, something that used to be called the "2nd stage boot glitch" in the Yosemite years and later. It causes something like this on screen for like 1/10th of a second:

Boot_glitch.png

 

Initially, it was addressed by a binary patch of IOGraphicsFamily kext. Since then, WEG has taken care of this though you may still see a very minor glitch, shorter than what most people ever experienced before the fix.

 

Name of the patch is arguable, some people see 4 to 5 logos, others see 12 or 13! :lol:

https://www.insanelymac.com/forum/topic/300431-glitches-on-boot/

 

This is the Apple's bug presents on a real MacBook. So it is not hackintosh problem and should not be cured.

  • Like 1
Link to comment
Share on other sites
deeveedee
deeveedee
Posted
Posted (edited)

"This is the Apple's bug presents on a real MacBook. So it is not hackintosh problem and should not be cured."

 

@Hervé - It's actually very simple - the problem is not exclusive to hacks and doesn't necessitate any further Clover fixes/revisions.  I speak 'Slice' - let me know if you need further translation.

Edited by tonyx86
Link to comment
Share on other sites
Slice
Slice
Posted
Posted
1 hour ago, Hervé said:

You lost me there @Slice.

 

You mentioned the glitch yourself several times in the past and referred to it with your "8 apples" denomination...

https://www.insanelymac.com/forum/topic/327824-dell-latitude-e6430-full-solution/

https://www.insanelymac.com/forum/topic/284656-clover-general-discussion/?do=findComment&comment=2724151

 

Yes, I was fighting with the glitch sometimes ago. But now I think it doesn't matter.

Link to comment
Share on other sites
Slice
Slice
Posted
Posted
11 hours ago, 5T33Z0 said:

 

Well, the question was, what is "Apple Eight" Kernel patch for. And that's what is for… so, I guess it can appear on hackintoshes although I never had it ;)

Just forget it.

  • Like 1
Link to comment
Share on other sites
Matgen84
Matgen84
Posted
Posted (edited)
1 hour ago, iCanaro said:

@Matgen84 if you ally your EFI, without serials, I can see to modify it according to my ideas and see if it then works to you

 

@iCanaro My EFI Folder attached (latest r5128 commit). The r5127 EFI folder works well, can boot BS 11.2 Beta 1
 

EDIT: Maybe, I've got a problem to update correctly opencorepkg submodule with Buildme option (Update and Update reset). What script do you use ?

 

Spoiler

1929738210_Capturedcran2021-01-1413_05_44.png.839234ed05bdcdf6dd9a1689dad0a4f9.png

 

EFI.zip

Edited by Matgen84
Link to comment
Share on other sites
 Share
Go to topic listing
×
×
  • Create New...