MP 1,1-5,1 Manually Configured OpenCore on the Mac Pro

[tsialex]

Here is the output: kern.hv_support: 0

It should be "1" ? But I can use virtualization software on that machine and it worked, also you can notice the "VMX" Cpuid flag previously posted.

kern.hv_support: 0 shows that your Mac don't support AppleHypervisor. If I remember correctly, support starts with Macs with with SandyBridge/Westmere CPUs.

This is a problem for people trying to run Docker with old Macs, Nehalem Mac Pros (all of MP4,1 CPUs and some of the original single CPU MP5,1) don't support it too. There is a thread here about it.

 

[jackluke]

kern.hv_support: 0 shows that your Mac don't support AppleHypervisor. If I remember correctly, support starts with Macs with with SandyBridge/Westmere CPUs.

This is a problem for people trying to run Docker with old Macs, Nehalem Mac Pros (all of MP4,1 CPUs and some of the original single CPU MP5,1) don't support it too. There is a thread here about it.

Ok but someone who own a MacPro3,1 could also confirm that they can't spoof a VMM flag with OpenCore ?

MacPro3,1 should use Penryn cpu as MacBook mid 2010 .

Just to be sure that "mobile cpus" don't need a different cpuidmask from OpenCore.

 

[cdf]

Here is the output: kern.hv_support: 0

Just to be sure that "mobile cpus" don't need a different cpuidmask from OpenCore.

I'm not really sure why VMM is not being set in your case. The flag bits are the same for x86 processors. I was just wondering if the flag would not be presented by sysctl if the hardware requirements for the Hypervisor framework are not met...

 

[Fairdinkem]

Why are you blessing apfs.efi when you be should be blessing BOOTx64.efi as per my suggestion? Also why is /usr/standalone/i386/apfs.efi in the EFI folder? What's your mac pro's model?

Edit: Sorry, just noticed that it was an error you were receiving and not a command you were entering. Are you using a cMP 2,1 or 3,1?

I am using a 4,1 which has flashed to a 5,1

 

[Fairdinkem]

Why are you blessing apfs.efi when you be should be blessing BOOTx64.efi as per my suggestion? Also why is /usr/standalone/i386/apfs.efi in the EFI folder? What's your mac pro's model?

Edit: Sorry, just noticed that it was an error you were receiving and not a command you were entering. Are you using a cMP 2,1 or 3,1?

problem solved I got it to work... thanks for your support.

 

[jackluke]

Premise: I have read that doing this from a "Sandy Bridge" spoofing as "Ivy Bridge" (that is the minimum natively Catalina supported architecture with i5 cpu), the OpenCore SMBIOS spoofing worked.

I ask this because I guess spoofing among too different board architecture might lead to a brick, I am on MacBook7,1 mid 2010, but I can assure that with a MacBookPro/Air 2011 (Sandy Bridge) OpenCore SMBIOS spoofing as an iMac 2012 worked. My doubt is about from "Core2Duo Penryn" to "Ivy Bridge".

Hence to those who already tested the OpenCore SMBIOS spoofing (example a MacPro5,1 as a Mojave supported iMac 2012), could it work too if I try to spoof a "C2D Penryn architecture" to an "Ivy Bridge" ?

 

[wash34]

Hello everyone.

First and foremost, I want to acknowledge the efforts you are making to make OpenCore 100% functional.

Despite my multiple readings I can not know the state of development for daily use on a mac pro 4.1 flashed in 5.1 under Catalina?
Follow the tutorial of the first Post is enough for a stable system ? Today I use the DosDude patch, OpenCore changes a lot to use?

And the boot menu, is this functional with a RX Vega 56 with original firmware ?

Can i help you ?

Thank you and good luck for the future!

 

[jackluke]

kern.hv_support: 0 shows that your Mac don't support AppleHypervisor. If I remember correctly, support starts with Macs with with SandyBridge/Westmere CPUs.

This is a problem for people trying to run Docker with old Macs, Nehalem Mac Pros (all of MP4,1 CPUs and some of the original single CPU MP5,1) don't support it too. There is a thread here about it.

and @cdf , you're right, in few words, even if the Core2Duo Penryn does support the "VMX cpu instruction", I've also attempted to unlock the MSR byte, but C2D can't run any hypervisor (example the apple one, ms hyper-v, docker and so on).

This is a summary why (click the blur to reveal):

The Intel Core2Duo cpu lacks two hardware virtualization features:
Extended Page Tables (EPT)
Unrestricted Guest (UG)
those features are available on all Core i-3/5/7 CPUs, but not on older models as Core2Duo.
So from Westmere (2010), Arrandale (2010), Sandy Bridge (2011) and later architecture.

 

[tsialex]

and @cdf , you're right, in few words, even if the Core2Duo Penryn does support the "VMX cpu instruction", I've also attempted to unlock the MSR byte, but C2D can't run any hypervisor (example the apple one, ms hyper-v, docker and so on).

This is a summary why (click the blur to reveal):

The Intel Core2Duo cpu lacks two hardware virtualization features:
Extended Page Tables (EPT)
Unrestricted Guest (UG)
those features are available on all Core i-3/5/7 CPUs, but not on older models as Core2Duo.
So from Westmere (2010), Arrandale (2010), Sandy Bridge (2011) and later architecture.

Nehalem Xeons lack UG too, that’s why Docker don’t work with MP4,1 and original single CPU MP5,1 Xeons and people have to upgrade to Westmere Xeons for it to work.

 

[startergo]

Regarding the HEVC issues you guys report on the real Mac Pros, try to do an ACPI rename of your GPU from whatever it is called by Apple to GFX0. If this is the root of the problem it should enable HEVC acceleration. On a Hackintosh GFX0 is mandatory for this.

According to this:
"Installing WhatEverGreen should take care of renaming all of the required ACPI identities for us so if you have any of the following Clover ACPI Fixes/Renames in your config.plist you should remove or disable them:-

• Change/Rename GFX0 to IGPU
• Change/Rename PEG0 to GFX0
• Change/Rename HECI to IMEI

The above Clover ACPI Fixes/Renames use a global/generic method of renaming the ACPI identities in the ACPI tables that can cause problems for us later on. WhatEverGreen finds the specific ACPI code that requires new ACPI identities using more specific methods and thus does a much better job. "

Renaming ACPI identities + WhateverGreen may be counterproductive

 

[tsialex]

According to this:
"Installing WhatEverGreen should take care of renaming all of the required ACPI identities for us so if you have any of the following Clover ACPI Fixes/Renames in your config.plist you should remove or disable them:-

• Change/Rename GFX0 to IGPU
• Change/Rename PEG0 to GFX0
• Change/Rename HECI to IMEI

The above Clover ACPI Fixes/Renames use a global/generic method of renaming the ACPI identities in the ACPI tables that can cause problems for us later on. WhatEverGreen finds the specific ACPI code that requires new ACPI identities using more specific methods and thus does a much better job. "

Renaming ACPI identities + WhateverGreen may be counterproductive

WTG don’t rename PSX*, you have to do it manually.

* = the number of the PCIe slot your GPU is installed.

 

[startergo]

MacOS App to compile Opencore, common drivers and kexts that are used with OpenCore from source.
It compiles version 0.5.3

Getting-Started-With-OpenCore

A guide For OpenCore

pavo-im.github.io

 

[Ludacrisvp]

WTG don’t rename PSX*, you have to do it manually.

* = the number of the PCIe slot your GPU is installed.

How do we request adding code to WEG to do renames for PSX* to GPU* ?

 

[cdf]

Such requests can probably made here: https://github.com/acidanthera/bugtracker/issues

 

[startergo]

* = the number of the PCIe slot your GPU is installed.

We have renamed PSX1 to GFX0, so it must be
* = the number of the PCIe slot -1 where your GPU is installed.

Right?
For instance using IoJones I can see PSX2@0
So I should patch
rename PSX2 to GFX1?

Just found some relevant info:
"Background:

In OEM Apple ACPI tables, the GPU Devices are named as follows:
iMac13/14/15/17: Discrete GPU = "GFX0" (@0). Integrated GPU = "IGPU" (@2 or 0x00020000)
MacPro3,1/5,1: Discrete GPU#1 = "PSX1" (@0).
MacPro6,1: Discrete GPU#1 = "GFX1" (@0). Discrete GPU#2 = "GFX2" (@0).

The usual school of thought is that each GPU should have a unique name (1st GPU = GFX0, 2nd GPU = GFX1 etc etc etc). However, If you use MacPro6,1 with multiple GPUs, you need to name all GPUs GFX1. Any GPU with a different name will lose video signal after the "Apple Logo" boot progress screen.

 

[Ludacrisvp]

We have renamed PSX1 to GFX0, so it must be
* = the number of the PCIe slot -1 where your GPU is installed.

Right?
For instance using IoJones I can see PXS2@0
So I should patch
rename PSX2 to GFX1?

I believe for AMD hardware accel it can only be used / enabled if the GPU is detected as GFX0 so limiting to first card only is how i'd see it. (or whichever slot you assign as GFX0)

 

[Ludacrisvp]

Such requests can probably made here: https://github.com/acidanthera/bugtracker/issues

Rename MacPro ACPI devices "PXS*" to "GFX*" · Issue #553 · acidanthera/bugtracker

On the Mac Pro (likely 1,1 through 5,1) the PCIe slots show up as PXS* and it would seem that part of the puzzle in getting HEVC hardware accelerated encoding from AMD Polaris or newer is that the ...

github.com

 

[Ludacrisvp]

Rename MacPro ACPI devices "PXS*" to "GFX*" · Issue #553 · acidanthera/bugtracker

On the Mac Pro (likely 1,1 through 5,1) the PCIe slots show up as PXS* and it would seem that part of the puzzle in getting HEVC hardware accelerated encoding from AMD Polaris or newer is that the ...

github.com

it has been closed as irrelevant.
I guess someone with more fame needs to step in as I don't have enough to make it happen on my own.

 

[tsialex]

it has been closed as irrelevant.
I guess someone with more fame needs to step in as I don't have enough to make it happen on my own.

Seems that he didn’t even understood the relevance, maybe we have to make a patch and then submit.

 

[haralds]

I have a 5,1 with EFI flashed Radeon Metal GPU and updated BLE/WiFi. After installation with DosDude Patcher (patches disabled) I can run a vanilla version of Catalina with SIP enabled.
Could anybody explain to me what the advantage (if any) of OpenCore would be in this situation.
I have scanned the thread and scanned the docs and do not see anything.
Does OpenCore do boot display with a non EFI flashed GPU card?

 

[startergo]

I have a 5,1 with EFI flashed Radeon Metal GPU and updated BLE/WiFi. After installation with DosDude Patcher (patches disabled) I can run a vanilla version of Catalina with SIP enabled.
Could anybody explain to me what the advantage (if any) of OpenCore would be in this situation.
I have scanned the thread and scanned the docs and do not see anything.
Does OpenCore do boot display with a non EFI flashed GPU card?

OC allows you to update the Catalina when the VMM flag is placed, dosdude's install does not.

 

[wash34]

OC allows you to update the Catalina when the VMM flag is placed, dosdude's install does not.

Hello, Could you explain me what is the VMM flag ? Is it about virtualization capacity ?

Thanks
[automerge]1573116215[/automerge]

I have a 5,1 with EFI flashed Radeon Metal GPU and updated BLE/WiFi. After installation with DosDude Patcher (patches disabled) I can run a vanilla version of Catalina with SIP enabled.
Could anybody explain to me what the advantage (if any) of OpenCore would be in this situation.
I have scanned the thread and scanned the docs and do not see anything.
Does OpenCore do boot display with a non EFI flashed GPU card?

Hello,

I've the same config with a non flashed GPU (RX 56 Vega, i don't find a way to flash it with EFI bios) and i've kernel panic with AppleIntelCPUPowerManagement.kext and i don't know how to resolv it ...

 

[startergo]

Hello, Could you explain me what is the VMM flag ? Is it about virtualization capacity ?

Catalina thinks it is being installed in a Virtual machine,
I haven't heard of Vega EFi flashing service.

"Currently Mac EFI Flashable:


AMD RX480 Reference Card
AMD RX580 Sapphire Pulse 8GB
AMD RX580 MSI 8GB
AMD RX580 GIGABYTE AORUS 8GB"

But you can buy Vega flashed:

AMD Vega 56 8GB

The #1 place for Mac Pro graphics cards. GPU upgrades from AMD and Nvidia.

www.macvidcards.com

Or ask macvidcards if they can flash one.
[automerge]1573123789[/automerge]
Maybe we can somehow modify OC bootloader to make RX580 appear as an eGPU (for boot screen) as now macmini can do that?

 

[wash34]

Catalina thinks it is being installed in a Virtual machine,
I haven't heard of Vega EFi flashing service.

"Currently Mac EFI Flashable:


AMD RX480 Reference Card
AMD RX580 Sapphire Pulse 8GB
AMD RX580 MSI 8GB
AMD RX580 GIGABYTE AORUS 8GB"

But you can buy Vega flashed:

AMD Vega 56 8GB

The #1 place for Mac Pro graphics cards. GPU upgrades from AMD and Nvidia.

www.macvidcards.com

Or ask macvidcards if they can flash one.
[automerge]1573123789[/automerge]
Maybe we can somehow modify OC bootloader to make RX580 appear as an eGPU (for boot screen) as now macmini can do that?

I've send a mail to Macvidcards to know if they can do something for this card.

Your shared link doesn't give information if the vega is already flashed with an EFI rom.

Do you use OC with a Macpro 4.1 flashed to 5.1? Do you follow first post process to use OC ?

 

[tsialex]

Maybe we can somehow modify OC bootloader to make RX580 appear as an eGPU (for boot screen) as now macmini can do that?

What this will accomplish? Pre-boot configuration support is a BootROM resource and MP5,1 EFI requires UGA. RX 580 is GOP.