Clover General discussion

[Sherlocks]

Hi,

 

Glad to see you after months, Slice.

 

Well, as for 2, we are not talking about whether HWP is good or not, consider what is KenelPm for - for those who does not have MSR 0xE2 unlocked. It's actually the fact that we have been missing the patch to _xcpm_idle, which should also be a part of 0xE2 patches. That's the point. Just as claimed, attach it to KernelPatchPm.

 

As for 3, my opinion is to rewrite the whole KernelCPU function to simply several lines: Fake as Penryn on older systems (e.g 10.7 and below) and else as Ivy. (Do this via the good new FakeCPUID)

in my case, i don't use _xcpm_idle patch with KernelPM option.

 

also i don't want to force enable HWP option in Clover.

just enable HWP in vectordata enough for my system.

 

나의 LG-F800S 의 Tapatalk에서 보냄

 

 

 

[Sherlocks]

 
I attached ver 138.0.0.0.0 if you want to search with UEFITool. Try plz.
MP51.fd.zip

i have a idea. now i'm at job. i will test it on my system. and i inform you for test. thanks

나의 LG-F800S 의 Tapatalk에서 보냄

[Slice]

To use HWP we need some kexts to set MSR770 after wake. Clover will not work after wake.

[kgp]

2 hours ago, Slice said:

To use HWP we need some kexts to set MSR770 after wake. Clover will not work after wake.

 

The entire iMac Pro Skylake-X/X299 community uses HWP in line with XCPM by enabling Intel Speed Shift Technology in mainboard BIOS with unlocked MSR register. Thus you are saying that some of the wake from sleep issues some people are facing could be due to Clover? Biggest problem are apparently random system wakes, which occasional produce a KP with black screen on wake. The KP should not be TSC related, as we designed TSCResetAdjust.kext, basically a simplification of the respective Voodoo equivalent. Both kexts do not only act at early boot but also during wake from sleep. 

 

Talking about wake issues, I would like to also address another issue... I recently face a 100% reproduceable blackscreen on wake issue and likely related KP with TB devices connected to a GC-Titan Ridge employed in my Skylake-X/X299 system. Is this something that also could be fixed by Clover? Without any TB devices or with USB-C devices connected to the Titan Ridge, sleep/wake works as expected. I was unable to fix the latter issue by means of my TB-SSDT. 

Edited October 16, 2018 by KGP-iMacPro

[Slice]

I mean that Clover set HWP=enable by setting MSR770=1. But after wake I see MSR770=0. Clover can't influence on this. It must be a kext that set MSR770 back to 1 after wake.

Same about TB. Clover can't influence on behaviour after wake.

[kgp]

3 minutes ago, Slice said:

I mean that Clover set HWP=enable by setting MSR770=1. But after wake I see MSR770=0. Clover can't influence on this. It must be a kext that set MSR770 back to 1 after wake.

Same about TB. Clover can't influence on behaviour after wake.

 

Do you think you could come up with a kext that sets MSR770=1 after wake? The latter might remove any HWP related wake issues or KPs.

Edited October 16, 2018 by KGP-iMacPro

[Slice]

I will not test. Simple Intel Enhanced Speedstep is quite enough for me.

[PMheart]

11 hours ago, Sherlocks said:

in my case, i don't use _xcpm_idle patch with KernelPM option.

 

also i don't want to force enable HWP option in Clover.

just enable HWP in vectordata enough for my system.

 

나의 LG-F800S 의 Tapatalk에서 보냄

 

 

 

Hi Sherlocks,

 

Nice to meet you too nowadays.

 

Well, no, _xcpm_idle patch must be a part of KernelPm just as its name shows - XCPM. Plus it will not only affect HWP. No idea why it has been missing for years, but anyway, we are supposed to add it now...

By the way, regarding HWP on iMac Pro, I don't know if FORCING enable it does benefit anything, but I guess it's better to avoid such things on desktop, there must be some negative influences, that's probably also why Apple has not enabled it up till now... (Yes, HWP only for MacBooks, at least for now)

[kgp]

53 minutes ago, PMheart said:

Hi Sherlocks,

 

Nice to meet you too nowadays.

 

Well, no, _xcpm_idle patch must be a part of KernelPm just as its name shows - XCPM. Plus it will not only affect HWP. No idea why it has been missing for years, but anyway, we are supposed to add it now...

By the way, regarding HWP on iMac Pro, I don't know if FORCING enable it does benefit anything, but I guess it's better to avoid such things on desktop, there must be some negative influences, that's probably also why Apple has not enabled it up till now... (Yes, HWP only for MacBooks, at least for now)

 

HWP obviously is also fully enabled on recent Desktops like the iMac Pro and has been proven to work flawless even without additional XCPM support also on our respective hacks. However, as HWP and XCPM also operate on a real iMac Pro in harmony, we successfully use the same approach on our hacks. Now Skylake-X CPUs are not identical with XEONs employed in a real iMacPro, but proven to be fully compatible in any case. Our Skylake-X CPUs and X299 systems absolutely behave as expected, apart from a few minor exceptions described in my former reply. 

 

However, following the recent statement of @Slice, MSR770 seems to become equal to zero on wake on our respective hacks because of yet unknown reason. I don't think that HWP is intended to be disabled after wake from sleep on real iMac Pro systems. This would not comply with any logics, rather seems to be a particular issue only observed on our hacks and indeed would require some quite urgent solution.

 

BTW.. You do not have to force HWP on Skylake-X/X299 with SMBIOS iMacPro1,1. It just works OoB, once enabled in the BIOS. The issue we are talking about is that HWP seems to get disabled on wake, if latest observations reported by Slice are also correct and valid  in case of Skylake-X. Thus what seems necessary is to force MSR770 to be 1 even after wake from sleep. 

 

 

 

 

Edited October 16, 2018 by KGP-iMacPro

[PMheart]

4 minutes ago, KGP-iMacPro said:

 

HWP obviously is also fully enabled on recent Desktops like the iMac Pro and has been proven to work flawless even without additional XCPM support also on our respective hacks. However, as HWP and XCPM also operate on a real iMac Pro in harmony, we successfully use the same approach on our hacks. Now Skylake-X CPUs are not identical with XEONs employed in a real iMacPro, but proven to be fully compatible in any case. Our Skylake-X CPUs and X299 systems absolutely behave as expected, apart from a few minor exceptions described in my former reply. 

  

 However, following the recent statement of @Slice, MSR770 seems to become equal to zero on wake on our respective hacks because of yet unknown reason. I don't think that HWP is intended to be disabled after wake from sleep on real iMac Pro systems. This would not comply with any logics, rather seems to be a particular issue only observed on our hacks and indeed would require some quite urgent solution.

 

 

 

 

Obviously not if you do check /System/Library/Extensions/IOPlatformPluginFamily.kext/Contents/Plugins/X86PlatformPlugin.kext/Contents/Resources/Mac-7BA5B2D9E42DDD94.plist carefully.

Lacking of "hwp" and "epp" power features and so on means that HWP is disabled (0).

 

Well, yes, I do not think that HWP is even supposed to be enabled on any desktop according to Apple's settings like mentioned above.

[kgp]

58 minutes ago, PMheart said:

Obviously not if you do check /System/Library/Extensions/IOPlatformPluginFamily.kext/Contents/Plugins/X86PlatformPlugin.kext/Contents/Resources/Mac-7BA5B2D9E42DDD94.plist carefully.

Lacking of "hwp" and "epp" power features and so on means that HWP is disabled (0).

 

Well, yes, I do not think that HWP is even supposed to be enabled on any desktop according to Apple's settings like mentioned above.

 

And why does a Skylake-X CPU shows proper CPU power management, behaviour and performance with XCPM totally deconfigured or even disabled but Intel SpeedShift technology enabled in BIOS? That's how we used this CPU initially, before enabling any XCPM support in addition. XCPM does basically not change anything on the general Skylake-X behaviour and performance, when used in addition.  

 

I guess Slice already provided the answer..  

 

BTW.. you can even change the HWP performance of Skylake-X with terminal commands or with a respective HWPValue in section CPU of Clover configurator. Carefully tested and approved by myself and respective users. 

Edited October 16, 2018 by KGP-iMacPro

[Slice]

No magic, it is BIOS decision enable or disable HWP after wake. May be someone happy with good BIOS.

[Sherlocks]

On 10/15/2018 at 8:55 AM, Moorre said:

Maybe it can. Click.

 

Can you see at my screenshots $IBIOSI$ click and no EFI Version.

0089 is the same.

But from EFI Version 138.0.0.0.0 a BIOS has more lines click:

 

I attached ver 138.0.0.0.0 if you want to search with UEFITool. Try plz.

MP51.fd.zip

 

hi @Moorre

here is test file.

can you test install HS with 140.0.0.0.0? now test clover automatically use 140.0.0.0.0. just i wonder 140.0.0.0.0 can use or not in older macos.

if can't use it, we consider older macos case.

 

thanks in advance

 

anyone who can install fresh Mojave with test clover file?

i want to clear between 220.200.252.0.0 and 16.16.375.0.0,0 for installation mojave.

try each 220.200.252.0.0 or 16.16.375.0.0,0 in config. then please tell me result. i don't have a time for test 

test.zip

Edited October 16, 2018 by Sherlocks

[Slice]

Hi Sherlocks,

Did you propose to add every macOS build version into Clover?

+              (AsciiStrStr(Entry->BuildVersion, "17G2112") || AsciiStrStr(Entry->BuildVersion, "17G2208") ||

May be there is a check somewhere else?

[Sherlocks]

Hi Sherlocks, Did you propose to add every macOS build version into Clover?

 

+              (AsciiStrStr(Entry->BuildVersion, "17G2112") || AsciiStrStr(Entry->BuildVersion, "17G2208") ||

May be there is a check somewhere else?

 

 

yes right. latest HS 10.13.6 build for MBP15,x.apple divided HS image for MBP15,x. so just added. old image(G2112, factory) and latest image. if you want to left only latest build, remove it

 

나의 LG-F800S 의 Tapatalk에서 보냄

 

 

 

 

 

[Moorre]

17 hours ago, Sherlocks said:

 

hi @Moorre

here is test file.

can you test install HS with 140.0.0.0.0? now test clover automatically use 140.0.0.0.0. just i wonder 140.0.0.0.0 can use or not in older macos.

if can't use it, we consider older macos case.

 

thanks in advance

 

test.zip

Hi! Just booted ok with HD7770 into HS. ROM Version is ok, but Release Date isnt.

 

 

Hardware Overview prints ok: Boot ROM version 140.0.0.0.0 and Mojave installer 10.14.0 seems to be happy:

 

BIOS Information
	Vendor: Apple Inc.
	Version: 140.0.0.0.0
	Release Date: 39/.B/0
	ROM Size: 2048 kB

I have no time to test Mojave installation now, but seems to be ok.

17 hours ago, Sherlocks said:

i want to clear between 220.200.252.0.0 and 16.16.375.0.0,0 for installation mojave.

try each 220.200.252.0.0 or 16.16.375.0.0,0 in config. then please tell me result. i don't have a time for test 

I cant understand what means

Quote

220.200.252.0.0 and 16.16.375.0.0,0

and where it should be located.

[Sherlocks]

and where it should be located.

thank you for testing. i knew releasedate issue. its complicate now. if you not input biosversion, you can see correct releasedate. i'm not perfectly consider releasedate. i will resolve it for all cases.

나의 LG-F800S 의 Tapatalk에서 보냄

[Slice]

3 hours ago, Sherlocks said:

yes right. latest HS 10.13.6 build for MBP15,x.apple divided HS image for MBP15,x. so just added. old image(G2112, factory) and latest image. if you want to left only latest build, remove it

 

나의 LG-F800S 의 Tapatalk에서 보냄

 

 

 

 

 

If you propose the build will not change then let it be.

Otherwise I may propose to check

AsciiStrStr(Entry->BuildVersion, "17G2")

or just check SMBIOS model for MBP15

[LockDown]

Hi

I want to use clover's KextsToPatch's InfoPlistPatch feature.

Lets say i need to change a device id. Which of the following is correct format?

 

do i need to include "<string>" to convert it into base64

<string>xxxxxxxx,xxx</string>

or without

xxxxxxx,xxxx

?

Edited October 18, 2018 by ellaosx

[MICKHAEL]

Both are ok. I prefer shorter

Hi
I want to use clover's KextsToPatch's InfoPlistPatch feature.
Lets say i need to change a device id. Which of the following is correct format?
 
do i need to include "" to convert it into base64

xxxxxxxx,xxx

or without

xxxxxxx,xxxx

?

[LockDown]

18 minutes ago, MICKHAEL said:

Both are ok. I prefer shorter

Thanks for the reply @MICKHAEL

To follow up...

Others are replacing the device id via InfoPlistPatch so kext would load. My plan is the opposite though. Id like to remove a device id inside the plist so the kext wont load.

So...

FIND:  xxxx:1111

REPLACE: xxx:0000

Will work?

 

Edited October 18, 2018 by ellaosx

[MICKHAEL]

I think the need to replace the same amount of symbols.

Thanks for the reply [mention=594753]MICKHAEL[/mention]
To follow up...
Others are replacing the device id via InfoPlistPatch so kext would load. My plan is the opposite though. Id like to remove a device id inside the plist so the kext wont load.
So...
FIND:  xxxx:1111
REPLACE: xxx:0000
Will work?
 

[LockDown]

1 hour ago, MICKHAEL said:

I think the need to replace the same amount of symbols.

Yes. the missing one x is just a typo

[Slice]

6 hours ago, ellaosx said:

Hi

I want to use clover's KextsToPatch's InfoPlistPatch feature.

Lets say i need to change a device id. Which of the following is correct format?

 

do i need to include "<string>" to convert it into base64

<string>xxxxxxxx,xxx</string>

or without

xxxxxxx,xxxx

?

It is better not use Clover for such operation. Use FakeID or DeviceProperties or Legacy.kext or Disabler.kext

[LockDown]

Hi @Slice

Can you show mw how is it done via DeviceProperties. That feature is something new to me.

Thanks