Jump to content

Clover General discussion

ErmaC

By ErmaC,
in Clover

 Share
29,929 posts in this topic

Recommended Posts

Matgen84
Matgen84
Posted
Posted
2 minutes ago, Jief_Machak said:

I tried and got the same problem (stop at +++++++++++). What is "unlocked BIOS" ?

 

"Unlocked BIOS" means, I think, Unlock CFG MSR BIOS. I don't try commit r5115 (562e571d3) on my Unlock Z390 Aorus Master, but only on Lock IvyBridge config: stop at +++++++++++ also.

Link to comment
Share on other sites
eSaF
eSaF
Posted
Posted
6 minutes ago, Jief_Machak said:

I have "MSR locked until next reset". What does that lock ?

 

It means exactly what it says - Your MSR is locked until you make a change to the BIOS to unlock it. There is a simple tutorial how to achieve this on Audiogod's page.

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

If you head over to the page the explanation is better than I can explain and more understandable than I can provide. Sir/Madam with a Label 'Coder' under your profile I have the inclination a detail as such, you would be familiar with the subject and what it is and does. Don't get me wrong please I can assure you I am not throwing shades or being disrespectful just curious :)

Edited by eSaF
Added text
  • Like 1
Link to comment
Share on other sites
Matgen84
Matgen84
Posted
Posted
2 minutes ago, Jief_Machak said:

My question wasn't precise enough. What is the MSR that it's locked. What effect does it have ?

 

Note: in config.plist, KernelPm=true, is applied by Clover when MSR 0xE2 is detected as locked. Unlock, KernelPm can be uncheck. Which is useful for aptiomemoryfix or ocquirks.efi

  • Like 2
Link to comment
Share on other sites
yapan4
yapan4
Posted
Posted (edited)

In my case in untouched BIOS are locked 0xE2 and 0x1AA MSR's.

For properly boot in this case used:

KernelPM patch(enabled automatically by Clover)

xcpm_pkg_scope_msrs (c)Pike R. Alpha (Modify lucke TEST) 10.15.4 beta1

patch writes to MSR_MISC_PWR_MGMT

or

KernelPM(enabled automatically by Clover)

xcpm_bootstrap HASwell-E (c) Pike R. Alpha Modify By N.Mano

 

If any one or more are unavailable, disabled or not working it causes a stop at +++++++++++++++

 

I can only guess which one doesn't work in version 5115, I suspect KernelPM???

 

Thank you and sorry my google-translated English

AppleIntelInfo(UnlockedBIOS).rtf

bootlog(unlockedBIOS).txt

config(unlockedBIOS).plist

Edited by yapan4
  • Like 1
Link to comment
Share on other sites
Slice
Slice
Posted
Posted
2 hours ago, Jief_Machak said:

Not sure where that page is...

Looks like it's connected to Kernel Power Management ?

I have to admit I'm a bit lost with all that. HWP, XCPM etc.

I can explain.

MSR 0xE2 according to intel datasheet is for PowerManagement and it's bit 14 (0x8000) has property Read/WriteOnce.

So once BIOS set the bit 0x8000 then the register will be locked until reset and the commands

mov ecx,0xe2

wrmsr 

will cause CPU crash.

This command presents in AppleIntelCPUPowerManagement.kext (10 times) and in kernel:

Two times explicit in procedure _xcpm_idle.

And implicit set in procedure _xcpm_init where wrmsr called three times.

Having locked BIOS we must prevent from writing this register.

May be I made a mistake or not found one more place but I have a BIOS with conditional LOCK CFG and I can check if Clover successfully does the patch with debug output to screen.

  • Like 2
Link to comment
Share on other sites
Slice
Slice
Posted
Posted
35 minutes ago, Jief_Machak said:

Does that mean that I'd have better power management if I'm unlocked ? 

You will have PM with Apple's algorithms but I am not sure if this is the best solution.

 

I know why recent Clover hangs. The problem is in my recent changes to AppleIntelCPUPMPatch. Thinking...

  • Like 1
Link to comment
Share on other sites
dragonmel
dragonmel
Posted
Posted

@Slice an company... 

 

 

looks like I spoke too soon

 

r5115 github release .. looks like a I am getting CMOS coruption / RTC issues

 

happens on reboots...sometimes.. and on wakes occasionally as well.

 

battery is good and the board wasnt powered off... 

 

I had this patched in my DSDT and not using clover hot patching for the RTC in many years

 

this is an older board.. and if I recall had a length of 0x04 just like my 2008 macbook.. I think I check allowrtc8 

I have also always used darkwake=no on this board.. it has always seemed to work best allowing self sleep and wake usually wihtout issues

 

 

should I check fixRTC and should I leave allow8 on ?   running mojave with dosdude on my macbook and the RTC DSDT definition is the same on the macbook as this x58 .. and everything is working good there... strange

 

anything else casue this?

IMG_3919.jpg.0262bba86181df7c6e68d98d7af291dc.jpg

 

@eSaF

 

I noticed you were also runnnig an x58 board... 

 

I was wondering how stable its been for you with mojave and catalina if you are running those.  I am still trying to work out some issues with my setup movoing from a very stable elcap and 660 card to mojave with vega has been giving me issues...

 

if you have the time to look over my DSDT or could zip my your config and acpi stuff I can compare notes!! 

 

thanks

 

r5115 prebuilt release is actually working pretty good on a VERY OLD DX58so... I have had issues transitioning this board that was rock solid running clover 43xx and ElCap with a 660 graphics card.. more reliable than my MacBook but moving to a Vega and clover 5000+ has been.. well interesting

 

It has brought native NVRAM to this intel half baked EUFI that this board has.. but it broke the ability for the board to boot without manually hitting F10 every boot to get to the bios boot picker... clover added a Mac OS X header and my UEFI drive showed up there but if left to boot on its own would go black screen with a solid cursor and not boot into clover at all...  

The other day in the clover menu I decided to hit apply clover boot option to all volumes and boom.. now a whole bunch of boot.efi entries show up above the MAC OS X header and the board does auto boot them.. it unfortunately as mentioned does not respect the picked drive but will default to preboot.. so I removed all boot options, used clover to hide all but my macOS volume and recovery..  rebooted, added clover boot options to all which added boot.efi lines only for macOS and recovery.. and defualts to the first drive in line which is macOS.. so one issue solved

 

I too have noticed that i now get a cmos reset on reboot and sometimes even on start after full shutdown. I have a heavily patched DSDT and don't use clovers auto fix..  do 10.14.6 2020-002 change the RTC requirements and need a new patch?  

 

I have also lately decided to go split personality on the SMBIOS section.  I was running straight MacPro5,1 as the board and the W3680 are pretty damn close to a late 2012 Mac Pro ...  but now I put in JUST the board-id of the iMac-pro instead of injecting it as a shiki variable... I now have FULL 264 AND 265 encode/decode with the vega and at least DRM with prime video using just pikera and skikigva=80 args along with dart=0 and darkwake=no

 

is there a way like with open core to change a board-it without it showing up in iMessage debug?  like they are doing with real macpro5,1 and open core..?  doesn't really mater.. apple did kick my iMessage but signing out, rebooting and signing back in have SO FAR worked so even though  I am using macpro5,1 def, serial, firmware etc.. and the iMacpro1,1 board id... the rig works pretty damn good.

 

I can still only boot with one monitor connected at DP1 .. if I try and boot the HDMI monitor or have both on at boot, or boot the one and turn on the other after getting to the desktop, black screen and console logging GPU.restart .. 

so I have to boot single, sleep, wake and that wake is sometimes hit/miss.. but if it wakes.. the second monitor can now be on and subsequent sleep/wakes usually are reliable and both monitors can remain on.. I have yet to figure that out

 

I know that open core is the future and that a lot of development is focused there but I just don't think the x58 will work very well with it .. and with no way to use my modded DSDT I would have to figure out all the find and replaces and that is just way too far past my ability or paygraide.. 

 

keep up to good work boys... 

 

Link to comment
Share on other sites
eSaF
eSaF
Posted
Posted
46 minutes ago, dragonmel said:

@Slice an company... 

 

 

looks like I spoke too soon

 

r5115 github release .. looks like a I am getting CMOS coruption / RTC issues

 

happens on reboots...sometimes.. and on wakes occasionally as well.

 

battery is good and the board wasnt powered off... 

 

I had this patched in my DSDT and not using clover hot patching for the RTC in many years

 

this is an older board.. and if I recall had a length of 0x04 just like my 2008 macbook.. I think I check allowrtc8 

I have also always used darkwake=no on this board.. it has always seemed to work best allowing self sleep and wake usually wihtout issues

 

 

should I check fixRTC and should I leave allow8 on ?   running mojave with dosdude on my macbook and the RTC DSDT definition is the same on the macbook as this x58 .. and everything is working good there... strange

 

anything else casue this?

IMG_3919.jpg.0262bba86181df7c6e68d98d7af291dc.jpg

 

@eSaF

 

I noticed you were also runnnig an x58 board... 

 

I was wondering how stable its been for you with mojave and catalina if you are running those.  I am still trying to work out some issues with my setup movoing from a very stable elcap and 660 card to mojave with vega has been giving me issues...

 

if you have the time to look over my DSDT or could zip my your config and acpi stuff I can compare notes!! 

 

thanks

 

r5115 prebuilt release is actually working pretty good on a VERY OLD DX58so... I have had issues transitioning this board that was rock solid running clover 43xx and ElCap with a 660 graphics card.. more reliable than my MacBook but moving to a Vega and clover 5000+ has been.. well interesting

 

It has brought native NVRAM to this intel half baked EUFI that this board has.. but it broke the ability for the board to boot without manually hitting F10 every boot to get to the bios boot picker... clover added a Mac OS X header and my UEFI drive showed up there but if left to boot on its own would go black screen with a solid cursor and not boot into clover at all...  

The other day in the clover menu I decided to hit apply clover boot option to all volumes and boom.. now a whole bunch of boot.efi entries show up above the MAC OS X header and the board does auto boot them.. it unfortunately as mentioned does not respect the picked drive but will default to preboot.. so I removed all boot options, used clover to hide all but my macOS volume and recovery..  rebooted, added clover boot options to all which added boot.efi lines only for macOS and recovery.. and defualts to the first drive in line which is macOS.. so one issue solved

 

I too have noticed that i now get a cmos reset on reboot and sometimes even on start after full shutdown. I have a heavily patched DSDT and don't use clovers auto fix..  do 10.14.6 2020-002 change the RTC requirements and need a new patch?  

 

I have also lately decided to go split personality on the SMBIOS section.  I was running straight MacPro5,1 as the board and the W3680 are pretty damn close to a late 2012 Mac Pro ...  but now I put in JUST the board-id of the iMac-pro instead of injecting it as a shiki variable... I now have FULL 264 AND 265 encode/decode with the vega and at least DRM with prime video using just pikera and skikigva=80 args along with dart=0 and darkwake=no

 

is there a way like with open core to change a board-it without it showing up in iMessage debug?  like they are doing with real macpro5,1 and open core..?  doesn't really mater.. apple did kick my iMessage but signing out, rebooting and signing back in have SO FAR worked so even though  I am using macpro5,1 def, serial, firmware etc.. and the iMacpro1,1 board id... the rig works pretty damn good.

 

I can still only boot with one monitor connected at DP1 .. if I try and boot the HDMI monitor or have both on at boot, or boot the one and turn on the other after getting to the desktop, black screen and console logging GPU.restart .. 

so I have to boot single, sleep, wake and that wake is sometimes hit/miss.. but if it wakes.. the second monitor can now be on and subsequent sleep/wakes usually are reliable and both monitors can remain on.. I have yet to figure that out

 

I know that open core is the future and that a lot of development is focused there but I just don't think the x58 will work very well with it .. and with no way to use my modded DSDT I would have to figure out all the find and replaces and that is just way too far past my ability or paygraide.. 

 

keep up to good work boys... 

 

 

 

Sorry man I really need to update my profile to truly reflect my every day rig - the X58 was retired long ago with the last OS X it was running which was High/Sierra. I no longer have that board and to this day I still think compared to my z390, it was my best for installing OS X even if it was just a legacy board. It was rock solid dual booting windows and OS X no problem, sorry I couldn't help in anyway.

Link to comment
Share on other sites
dragonmel
dragonmel
Posted
Posted
3 hours ago, eSaF said:

 

 

Sorry man I really need to update my profile to truly reflect my every day rig - the X58 was retired long ago with the last OS X it was running which was High/Sierra. I no longer have that board and to this day I still think compared to my z390, it was my best for installing OS X even if it was just a legacy board. It was rock solid dual booting windows and OS X no problem, sorry I couldn't help in anyway.

 

 

hehe... quitter.. going to something easier... hehe.. kidding

 

yeah.. until recently with this VEGA I had a pretty good DSDT and was running elcap using clover pretty much for smbios and that's it... 

 

I had elcap working with UEFI booting but it didn't have native nvram with clover 43xx so was using clover for that as well

 

sometime after clover 4900 I noticed that nvram was working ... so I am booting with UEFI and native nvram.. but honestly, with intels first gen UEFI environment as half baked as it was... I can't rule out that this might be a bit of the problem and going back to emulated nvram might be necessary. but .. time will tell...

 

  • Like 1
Link to comment
Share on other sites
 Share
Go to topic listing
×
×
  • Create New...