OpenCore General Discussion

[darthsian]

Hi @Download-Fritz, on last commit of 0.5.9 OpenCanopy bootpicker shows "EFI" label instead of "Windows" and wrong hard drive icon (it shows normal icon like for macOS disks) for my Windows partition. I have Windows on separate disk, but EFI part is on disk with OpenCore together (screenshot attached). No problem until last commits of 0.5.9. Did i miss some new change in configuration (can't find anything related to this)? Thanks.

 

EDIT: It started after this commit - OcBootManagementLib: Set 'Windows' name based on booter name

 

My OpenCore bootlog attached.

opencore-2020-05-28-122822.txt.zip

 

Edited May 28, 2020 by darthsian

[apianti]

On 5/26/2020 at 5:43 PM, ajacocks said:

 

I did change it...I had switched because some online guide or other suggested it.

 

It's now iMac17,1.

 

Thanks!

- Alex

 

Oh, no problem. Did you try what I had said previously to that about fixing your ACPI tables?

 

On 5/26/2020 at 9:13 PM, lanzaozijun said:

Lenovo-Air13-IWL

Upgrade to macos catalina 10.15.5 (19F96) hdmi external display black screen!
In macos catalina 10.15.4, boot-args enable igfxonln = 1 igfxagdc = 0 method is invalid.
Is there any other solutions for my friends?

 

 

Sounds like you should ask this in lilu or whatevergreen topics.

 

On 5/26/2020 at 11:06 PM, startergo said:

I wanted to migrate to OC. I just don't know if it even supports Arrandale. I asked questions in that thread. But nobody answered.

 

If you have this working with clover, then just migrate to OC. The boot loader is not giving you support for anything, you just have to enable patches you want done in OC, clover tries to do it automatically only if needed but usually just ends up doing most by default. Also your custom patches need altered slightly but all of this is easily available in the plethora of guides for OC from the AcidAnthera and dortania githubs.

 

On 5/27/2020 at 12:33 AM, SavageAUS said:

Im trying to boot Arch Linux with OC, OC sees my EFI partition for Arch but it just boots to a black screen.

What config do i need to generate a boot log for OC so i can get my PCIRoot address so i can hopefully add an entry in OC.

If this can be done does anyone have a method for this?

 

Use the EFI shell, my dude, to either look at the paths or add a boot entry. Why you not just creating one from in arch? Do you not get all your boot choices from firmware in OC?

 

On 5/27/2020 at 4:32 AM, MacNB said:

 

Support of Arrandale is a macOS issue and nothing to do with the boot loader (Clover, OC, etc).

 

Well, it does somewhat, but in this instance that should be superseded by drivers and Lilu+WEG.

 

On 5/27/2020 at 5:06 AM, kecinzer said:

I have weird minor issue with OpenCore (my files). After I wake up screen and login into macOS, I see for a while booting screen and then normal login screen into macOS. It seems to some buffered thing. This occuring every time. Sleep and wake from sleep solves it, so after sleep it work normally.

I'm attaching video of this issue.
https://drive.google.com/open?id=1_8a8YukSJjiXuySTFPdWWgUdoo2S-PPl
May be it can be solved simply by some option in OpenCore or WEG?

 

This is normal behavior (as point out, even for a mac). It is still booting at that point, lol. FileVault takes place before booting, the disk is encrypted, you have to login to unlock it so it can finish booting. Sometimes when your system goes to sleep it won't be locked because it's a waste of resources to do it too quickly, so quick sleep/wake is not going to trigger it. Deep sleep and hibernation are probably going to trigger it since they need to actually boot again (not exactly but from the viewpoint of the system it actually shutdown and barely anything is still powered, the firmware does a special weird boot and the kernel also does a weird start).

 

On 5/27/2020 at 8:19 AM, MacNB said:

 

I get that on my real MacBook Pro.

It's normal with FileVault.

 

Looks like you turned ON File Vault.

 

Yeah, it's normal.

 

On 5/27/2020 at 9:09 AM, kecinzer said:

Ah, thats weird. I have Fault Vault enabled.

I have also MacBook Pro and I never mentioned it.

It happens only after fresh start or reboot with external displays are connected. I tried gfxrst=4 argument, but result is still same.

 

It's not big deal, but my current setup is almost perfect, so I want to fix this little glitch :).

 

Guessing you have a much smaller/faster/both disk in your macbook pro so it's not taking as long to unlock and you just aren't noticing the final stage of booting there.

 

On 5/27/2020 at 10:34 AM, MacNB said:

 

Why fix something that's not broken ?

 

If it happens on your real Mac, have you tried to ask Apple to fix your real Mac ?

 

I agree, nothing to fix, there's nothing wrong.

 

5 hours ago, LBowgensloth said:

I just would like to ask if bluetooth support in oc is ever going to become a thing or if it is out of scope. (Like bluetooth keyboard support in FileVault.)

The support of BT in the boot environment is probably going to be based entirely on your firmware supporting it and the card itself having a compatible firmware. There are many motherboards that are coming with intel m.2 wifi+bt and you can use BT devices in the firmware. The support for the keyboard in filevault may be a relevant issue depending on why it is not working. Because you don't have the devices working during boot? Or because there needs to be a fix for the drivers? As I recall there needed to be a fix separate for each bus that the devices worked over for filevault so it may actually need a fix for bluetooth as well.

 

5 hours ago, Download-Fritz said:

@LBowgensloth That is most definitely out of scope. However, I don't remember the name of the technology, but some controllers support re-pairing themselves on their own.

 

Betting that he wants an input fix driver for bt keyboard but only exists for usb.... Might be tangentially in scope.

 

4 hours ago, MacNB said:

 

I have an Apple MacBook Pro WiFi Wireless AC BT4.0 CARD (part number 653-0194) BCM943602CS fitted into a PCIe x1 adapter in my desktop.

I also have a Apple Magic 2 keyboard & mouse.

I can access the BIOS via bluetooth before macOS or Windows boots.

 

I can use OPT/Alt, CMD+V, etc OC features before macOS boots.

 

So basically, the BT adapter & Apple peripherals work in LEGACY mode during pre-boot and appears to the system as a USB HID device and once macOS or Windows boots, their drivers take over seamlessly.

 

I have not yet installed FileVault so don't know if the kbd/mouse would work.

 

YMMV 

 

Well.. That answers the first question. Seems that you do indeed want a bt version of the usb keyboard fix. Mouse should work if it works in preboot. You could check and see if the usb driver works but it will cause you to basically brick your install if it doesn't work. You could check the device path of the keyboard device and see what it comes up as and whether or not the usb driver even reconnects with it by using the EFI shell, you can get stuck here though too but with less consequence a hard reboot couldn't solve.

 

2 hours ago, darthsian said:

Hi @Download-Fritz, on last commit of 0.5.9 OpenCanopy bootpicker shows "EFI" label instead of "Windows" and wrong hard drive icon (it shows normal icon like for macOS disks) for my Windows partition. I have Windows on separate disk, but EFI part is on disk with OpenCore together (screenshot attached). No problem until last commits of 0.5.9. Did i miss some new change in configuration (can't find anything related to this)? Thanks.

 

EDIT: It started after this commit - OcBootManagementLib: Set 'Windows' name based on booter name

 

My OpenCore bootlog attached.

opencore-2020-05-28-122822.txt.zip

 

 

There is a problem in that commit. BootEntry->IsGeneric is whether the entry is \EFI\BOOT\... but then the code uses that to check whether windows exists. Windows could also exist at \EFI\Microsoft\Boot\bootmgfw.efi and that is the default entry that windows should actually make for your disk when installed or boots (and detects it needs recovered). Also there is a possibility of an illegal access past the end of the string as the size of EFI_REMOVABLE_MEDIA_FILE_NAME is not checked to be smaller than the offset that is given there to get the address of the string, that could be past the end. That line should occur after checking that the string is larger, later inside the if.

 

EDIT: Also, I'm a little confused, if I have an entry for a generic loader that is not windows but windows is also there, does this not detect every generic loader is windows?

Edited May 28, 2020 by apianti

[MacNB]

1 hour ago, apianti said:

 

Well.. That answers the first question. Seems that you do indeed want a bt version of the usb keyboard fix. Mouse should work if it works in preboot. You could check and see if the usb driver works but it will cause you to basically brick your install if it doesn't work. You could check the device path of the keyboard device and see what it comes up as and whether or not the usb driver even reconnects with it by using the EFI shell, you can get stuck here though too but with less consequence a hard reboot couldn't solve.

 

 

 

Is that a suggestion for me ??

I don't have a problem with BT.

I was just mentioning that my Apple BT KBD/MSE work fine in BIOS-POST/OC Shell/preboot/etc when used with real Apple BT adapter. 

[kecinzer]

@MacNB @apianti

Thank you a lot for explanation, but why this happens only with external display/s connected? When I use only laptop display, this glitch not occure.

 

And I have a 4 times faster NVMe SSD than on real MB.

[apianti]

1 hour ago, MacNB said:

 

Is that a suggestion for me ??

I don't have a problem with BT.

I was just mentioning that my Apple BT KBD/MSE work fine in BIOS-POST/OC Shell/preboot/etc when used with real Apple BT adapter. 

 

Nah it was suggestion for the other person that the bt does indeed work in boot environment.

 

7 minutes ago, kecinzer said:

@MacNB @apianti

Thank you a lot for explanation, but why this happens only with external display/s connected? When I use only laptop display, this glitch not occure.

 

And I have a 4 times faster NVMe SSD than on real MB.

 

I think you're not understanding the part about it is still booting. It definitely makes sense a graphics change occurs right there as well.

[nmano]

10.15.5 (19F96)

Apple big update XCPM _bootstrap in 10.15.5

I thing C4 ->72

Still testing

 

# This patch for X299(Skylake-X/Cascade Lake-X)

Comment    String  <->xcpm_bootstrap Skylake-X/Cascade Lake-X (c) Pike R. Alpha Modify By N.Mano

Disabled   Boolean <-> No

Find       Data<->   80C37280 FB17777A 0FB6C348 8D

MatchOS    String <-> 10.15.x

Replace    Data<-> 80C36B80 FB17777A 0FB6C348 8D

[SavageAUS]

@apianti in OC I see the EFI partition for Arch but it boots to a black screen. I do not see a Linux entry / vmlinuz etc. I just remembered I can get my path from boot menu but I’m unsure of how to add an entry to OC.

 

 

Sent from my iPhone using Tapatalk

Edited June 1, 2020 by SavageAUS

[MacNB]

4 hours ago, nmano said:

10.15.5 (19F96)

Apple big update XCPM _bootstrap in 10.15.5

I thing C4 ->72

Still testing

 

# This patch for X299(Skylake-X/Cascade Lake-X)

Comment    String  <->xcpm_bootstrap Skylake-X/Cascade Lake-X (c) Pike R. Alpha Modify By N.Mano

Disabled   Boolean <-> No

Find       Data<->   80C37280 FB17777A 0FB6C348 8D

MatchOS    String <-> 10.15.x

Replace    Data<-> 80C36B80 FB17777A 0FB6C348 8D

 

Thanks. But what about Ivy Bridge ? What's the change ?

 

EDIT:

This used to work with 10.15.4 (and apparently 10.5.5. Beta 2).

 

Find: 3B7E2E80 C3C480FB 42

Replace: 007E2E80 C3C680FB 42

 

But with released 10.15.5 now OC cannot find the FIND values in the kernel.

The released kernel has changed 

 

Edited May 29, 2020 by MacNB
Added previous working patch

[hardcorehenry]

14 hours ago, nmano said:

10.15.5 (19F96)

Apple big update XCPM _bootstrap in 10.15.5

I thing C4 ->72

Still testing

 

# This patch for X299(Skylake-X/Cascade Lake-X)

Comment    String  <->xcpm_bootstrap Skylake-X/Cascade Lake-X (c) Pike R. Alpha Modify By N.Mano

Disabled   Boolean <-> No

Find       Data<->   80C37280 FB17777A 0FB6C348 8D

MatchOS    String <-> 10.15.x

Replace    Data<-> 80C36B80 FB17777A 0FB6C348 8D

Shouldn't "Find" look more like this?:

F:8D 43 C4 3C 42 77 22

R:8D 43 XX 3C 42 77 22

XX-calculated value for your CPU

(Only theory, might be worth check it)

Edited May 29, 2020 by hardcorehenry

[apianti]

There are two ways you can discover what the new patch needs to be. First one, disassemble or decompile both kernels, find the original patch, locate the new patch by finding it in the new one. They won't be exactly similar but the code should be close enough that you should be able to determine the change. It also helps if you know the names of functions that are actually being patched. Second, just compare the two kernels in a hex editor, search for the patch in the original and look for the patch around the same area in the newer with a slight change. A note is that maybe the replacement needs modified too, maybe they simply added or removed bits to a default bit mask, which may include something new that is needed, idk.

[apianti]

13 hours ago, SavageAUS said:

[mention]apianti [/mention]in OC I see the EFI partition for Arch but it boots to a black screen. I do not see a Linux entry / vmlinuz etc. I just remembered I can get my path from boot menu but I’m unsure of how to add an entry to OC.

 

I thought that it just added your entries from your firmware to start. Maybe though you need to adjust these:

 

https://github.com/acidanthera/OpenCorePkg/blob/master/Docs/SampleFull.plist#L685

https://github.com/acidanthera/OpenCorePkg/blob/master/Docs/SampleFull.plist#L652

https://github.com/acidanthera/OpenCorePkg/blob/master/Docs/Configuration.pdf

[vit9696]

MacNB, nmano, and the rest. Any adequate the reason to continue to post these patches instead of using builtin quirks for XCPM and virtualising CPU ID? I am kind of tired and confused seeing this from time to time.

[apianti]

1 hour ago, vit9696 said:

MacNB, nmano, and the rest. Any adequate the reason to continue to post these patches instead of using builtin quirks for XCPM and virtualising CPU ID? I am kind of tired and confused seeing this from time to time.

 

Maybe they want to understand how it works and why? Maybe they are experimenting? Maybe those quirks are no longer working? There could be many reasons, I don't understand why you seem to think there is only one way you should ever be able to do anything...

[mhaeuser]

22 minutes ago, apianti said:

I don't understand why you seem to think there is only one way you should ever be able to do anything...

Because when you leave people alone with experimenting you get AptioFix-free20000

[apianti]

5 minutes ago, Download-Fritz said:

Because when you leave people alone with experimenting you get AptioFix-free20000

 

Oh no! That's actually called lack of research and following a s h i t tutorial from probably... tonymac... haha.

 

EDIT: Forgot, you don't believe in experimentation because that's science.... D a m n.

 

EDIT2: Holy s h i t, I can't spell. And the site is getting sooooooooooooooooooooo slow. wtf is going on?

Edited May 30, 2020 by apianti

[SavageAUS]

 
I thought that it just added your entries from your firmware to start. Maybe though you need to adjust these:
 
https://github.com/acidanthera/OpenCorePkg/blob/master/Docs/SampleFull.plist#L685
https://github.com/acidanthera/OpenCorePkg/blob/master/Docs/SampleFull.plist#L652
https://github.com/acidanthera/OpenCorePkg/blob/master/Docs/Configuration.pdf

Yeah it’s possible. I think scanpolicy is set to 0 at the moment. I’m not sure how that all works yet but I can try the scan policy you linked.
Mind you I don’t have a “Linux” entry in my firmware just a uefi os entry.


Sent from my iPhone using Tapatalk

[nmano]

19 hours ago, hardcorehenry said:

Shouldn't "Find" look more like this?:

F:8D 43 C4 3C 42 77 22

R:8D 43 XX 3C 42 77 22

XX-calculated value for your CPU

(Only theory, might be worth check it)

@hardcorehenry I really thanks for your 10.15.5 _bootstrap value I test Its work well.

#Kernel->Emulate->Your CPU Original ID

 

#CPUID          Cascade Lake-X 0x050657

 

Cpuid1Data   <->  57060500 00000000 00000000 00000000

Cpuid1Mask  <->   FFFFFFFF 00000000 00000000 00000000

 

#CPUID           Skylake-X 0x050654  /Skylake-H 0x0406E3

 

Cpuid1Data   <->  54060500 00000000 00000000 00000000

Cpuid1Mask  <->   FFFFFFFF 00000000 00000000 00000000

 

#CPUID         0x0306F2   Haswell-E/Haswell 0x0306C3

 

Cpuid1Data   <->  F2060300 00000000 00000000 00000000

Cpuid1Mask  <->   FFFFFFFF 00000000 00000000 00000000 

 

#CPUID         0x0406F1   Broadwell-E/Broadwell 0x0306D4

 

Cpuid1Data   <-> F1060400 00000000 00000000 00000000

Cpuid1Mask  <->  FFFFFFFF 00000000 00000000 00000000

 

#6 This patch for X99(Haswell-E)

Comment    String  <->xcpm_bootstrap HASwell-E (c) hardcorehenry Modify By N.Mano
Disabled   Boolean <-> No
Find       Data<->   8D43C43C 427722
MatchOS    String <-> 10.15.x
Replace    Data<->   8D43C13C 427722

#7 This patch for X299(Skylake-X/Cascade Lake-X)


Comment    String  <->xcpm_bootstrap Skylake-X/Cascade Lake-X (c) hardcorehenry Modify By N.Mano
Disabled   Boolean <-> No
Find       Data<->   8D43C43C 427722
MatchOS    String <-> 10.15.x
Replace    Data<->   8D43BD3C 427722

#8 This patch for X99(Broadwell-E)

Comment    String  <->xcpm_bootstrap Broadwell-E (c) hardcorehenry Modify By N.Mano
Disabled   Boolean <-> No
Find       Data<->   8D43C43C 427722
MatchOS    String <-> 10.15.x
Replace    Data<->   8D43BC3C 427722


#9 This patch for IvyBridge E

Comment    String  <->xcpm_bootstrap IvyBridge E (c) hardcorehenry Modify By N.Mano
Disabled   Boolean <-> No
Find       Data<->   8D43C43C 427722
MatchOS    String <-> 10.15.x
Replace    Data<->   8D43C63C 427722

[hardcorehenry]

For some users like me with crappy CPU(Pentium Haswell), bootstrap patch have to be put manually: Kernel>Patch. Thanks to experiments I managed to boot without this patch and without FAKECPUID, and have "full speedstep".(fakecpufamily patch)

Edited May 30, 2020 by hardcorehenry

[mhaeuser]

@hardcorehenry Did you find out why it does not work? Basically except the few of you nobody ever reported issues. If it's indeed not possible with the current OC patch at all, we might want to update it.

[hardcorehenry]

@Download-Fritz

This is not what I meant, Kernel>Emulate works, from what I understand for Haswell+low-end AVX patch and _xcpm_bootsrap patch have to be put manually, no issue also everything works(checked out of curiosity, bootstrap patch sometimes need to be updated, though).  I haven’t reported such need, because as I said I use fakecpufamily patch since OC 0.1 this patch doesn’t require for my CPU bootstrap or FAKECPUID, I even have set it as headless/connectorless for better compatibility. So no issue at all, everything works, keep up good work.

Thanks!

Edited May 30, 2020 by hardcorehenry

[MacNB]

13 hours ago, vit9696 said:

MacNB, nmano, and the rest. Any adequate the reason to continue to post these patches instead of using builtin quirks for XCPM and virtualising CPU ID? I am kind of tired and confused seeing this from time to time.

 

Simple answer is I do not know HOW to get OC to enable XCPM on my Ivy Bridge i7-3770K.

Yes I have read the configuration.pdf and yes I have read bugtracter#365 it's NOT obvious.

The configuration document does not simply say for example "If you have CPUID 0306A9, then use the following settings to enable XCPM....."

 

I have set:

Cpuid1Data: <A9 06 03 00 00 00 00 00 00 00 00 00 00 00 00 00>
Cpuid1Mask: <FF FF FF FF 00 00 00 00 00 00 00 00 00 00 00 00>

AppleXcpmCfgLock = True
AppleXcpmExtraMsrs = True

But that does not work.

I rather NOT add a patch but it was simpler to use a patch (which worked fine until 10.15.5 arrived).

I just installed latest Clover r5118 and that enables XCPM on 15.15.5 by simply setting KernelPM=True.

 

See attached log & config files and maybe you can spot what I am doing wrong ?

MacNB-config.plist

opencore-2020-05-30-095858.txt

Edited May 30, 2020 by MacNB

[MacNB]

2 hours ago, hardcorehenry said:

For some users like me with crappy CPU(Pentium Haswell), bootstrap patch have to be put manually: Kernel>Patch. Thanks to experiments I managed to boot without this patch and without FAKECPUID, and have "full speedstep".(fakecpufamily patch)

 

...and what's the fakecpufamily patch ?

Thanks.

[pitrysha]

50 minutes ago, MacNB said:

 

Simple answer is I do not know HOW to get OC to enable XCPM on my Ivy Bridge i7-3770K.

Yes I have read the configuration.pdf and yes I have read bugtracter#365 it's NOT obvious.

The configuration document does not simply say for example "If you have CPUID 0306A9, then use the following settings to enable XCPM....."

 

I have set:

Cpuid1Data: <A9 06 03 00 00 00 00 00 00 00 00 00 00 00 00 00>
Cpuid1Mask: <FF FF FF FF 00 00 00 00 00 00 00 00 00 00 00 00>

AppleXcpmCfgLock = True
AppleXcpmExtraMsrs = True

But that does not work.

So it was simpler to use a patch (which worked fine until 10.15.5 arrived).

I just installed latest Clover r5118 and that enables XCPM on 15.15.5 by simply setting KernelPM=True.

 

See attached log & config files and maybe you can spot what I am doing wrong.

MacNB-config.plist

opencore-2020-05-30-095858.txt

Try like this

 

A3060300 00000000 00000000 00000000 

FFFFFFFF 00000000 00000000 00000000 

 

More options

 

A3070400 00000000 00000000 00000000

FFFFFFFF 00000000 00000000 00000000 

 

A3060400 00000000 00000000 00000000

FFFFFFFF 00000000 00000000 00000000 

Edited May 30, 2020 by pitrysha

[MacNB]

35 minutes ago, pitrysha said:

Try like this

 

A3060300 00000000 00000000 00000000 

FFFFFFFF 00000000 00000000 00000000 

 

Thanks but no joy.

Interesting that the OC log now shows:

04:016 00:003 OCAK: AppleCpuPmCfgLock v1 replace count - 10
04:017 00:000 OCAK: Patch v1 success com.apple.driver.AppleIntelCPUPowerManagement
04:017 00:000 OCAK: AppleCpuPmCfgLock v2 replace count - 0

Which seems to imply OC is now patching com.apple.driver.AppleIntelCPUPowerManagement.

In effect, disabling XCPM ?

 

[hardcorehenry]

1 hour ago, MacNB said:

 

...and what's the fakecpufamily patch ?

Thanks.

To boot with Intel Pentium Haswell I normally should emulate IvyBridge CPUID(Kernel>Emulate), because I’m not able to boot with my native CPUID 0x306C3. Instead of faking 0x306C3=>0x306A9 I’m faking cpuhaswellfamily=>cpuivybridgefamily. This patch is no use for you, but since you ask:

Base: 
Comment: Fake Haswell CPUFAMILY
Count: 1
Enabled: YES
Find: bbdc82b2 10
Identifier: kernel
Limit: 0
Mask: ffffffff ff
MaxKernel:
MinKernel:
Replace: bb35e865 1f
ReplaceMask:
Skip: 0

Edited May 30, 2020 by hardcorehenry