[SUCCESS] Gigabyte Designare Z390 (Thunderbolt 3) + i7-9700K + AMD RX 580

[CaseySJ]

@CaseySJ

8086-3EC2 is not our XHC2 device id !!

I have the following information about this ID :
8th Gen Core Processor Host Bridge/DRAM Registers

| | +-o MCHC@0 <class IOPCIDevice, id 0x1000001e3, registered, matched, active, busy 0 (0 ms), retain 8>
| | | {
| | | "name" = <"pci8086,3ec2">
| | | "compatible" = <"pci8086,7270","pci8086,3ec2","pciclass,060000","MCHC">
| | | "subsystem-vendor-id" = <86800000>
| | | "IOName" = "pci8086,3ec2"
| | | "reg" = <0000000000000000000000000000000000000000>
| | | "device-id" = <c23e0000>
| | | "IOPowerManagement" = {"CapabilityFlags"=258,"CurrentPowerState"=2,"ChildProxyPowerState"=2,"MaxPowerState"=3}
| | | "built-in" = <00>
| | | "IOPCIResourced" = Yes
| | | "revision-id" = <07000000>
| | | "vendor-id" = <86800000>
| | | "pcidebug" = "0:0:0"
| | | "class-code" = <00000600>
| | | "acpi-path" = "IOACPIPlane:/_SB/PCI0@0/MCHC@0"
| | | "subsystem-id" = <70720000>
| | | "acpi-device" = "IOACPIPlatformDevice is not serializable"

Extracted from iMac19,1 IOReg, with Cannonlake PCH !

That's MCHC! Does this mean we're using the wrong register/GPIO for XHC2? In other words, some problem with this?

 

[Elias64Fr]

That's MCHC! Does this mean we're using the wrong register/GPIO for XHC2? In other words, some problem with this?

View attachment 447327

No Problem ! Just reading VID/DID when component is not available or visible ! We might always have 0xFFFFFFF that mean component is not ready/connected via tunneling ...

 

[docfiru]

From my recent experience, you are not dumb. I had to reset the MSR 4 times before it would stay set. Each time I carefully followed the instructions and each time it reported the variable as 0x00. On boot back to Windows, with consequent adjustments to BIOS it somehow reset itself to 0x01. Finally replaced EFI folder with a working one and solved the problem but was curious as to how a variable that cannot be set by normal methods could revert.

However it now works and I have just updated to Catalina 10.15.3. Curiously 4 boots with each progress bar looking as though it had stopped. Finally a bar with time that actually counted down correctly and then black screen as the monitor decided to go to sleep claiming no input. A keypress or two and voilà...

I've found that if you change things in the BIOS and save them, the variable will get re-set to 0x01 again. I've had to unlock CSR after every BIOS change. If I don't touch the BIOS, it stays unlocked across all reboots etc.

 

[AlexD]

@gandem

I guess, I found the true reason for my system failing to boot. And it is not removed EmuVariableUefi.
But very messy, fragmented memmap that can drastically change on a sudden occasion.

I was finally able to succesfully boot with iGPU enabled, but only with Pre-Allocated DVMT lowered from 64Mb to 32Mb and Total from 256Mb to 128Mb.

There were multiple "error allocating pages" messages during many unsuccesful boot attempts and each time system needed around 0x11000-0x12000 pages.

But when I looked at memmap output I was surprised that there are no such region with that number of pages available.

Instead, there are many regions with much smaller number of available pages. And sometimes larger region appear which allows me to boot system.

Btw, there's always one large region with 70000+ pages available, starting at 100000000, but calculating slide value for it gives us enormous 2048, which doesn't seem to be right value to pass as a boot arg.

Configurations tested:
[OK] - system boots with iGPU disabled
[OK] - system boots with iGPU enabled and lowered iGPU memory settings (32/128 instead of default 64/256)
[OK] - system boots with iGPU enabled and default iGPU memory settings (64 / 256)

Final Conclusion:
Native NVRAM works on Z390 Aorus Pro.
FwRuntimeServices + OCQuirks + plist combo required for a proper memory mapping.

@CaseySJ I should excuse for posting this in Z390 Designare thread. Just want to explore and implement your Native NVRAM guide so many more users with other Z390 boards can apply it on their systems.

 

[CaseySJ]

I've found that if you change things in the BIOS and save them, the variable will get re-set to 0x01 again. I've had to unlock CSR after every BIOS change. If I don't touch the BIOS, it stays unlocked across all reboots etc.

Ever since I unlocked MSR 0xE2 on my main system months and months and months ago it has never been relocked automatically except after upgrading BIOS from F6 to F8. But with CMOS Reset, it did not lock itself.

 

[CaseySJ]

No Problem ! Just reading VID/DID when component is not available or visible ! We might always have 0xFFFFFFF that mean component is not ready/connected via tunneling ...

On your system I believe you also get XHC2 VID/DID = 0x0000FFFF (or maybe 0xFFFFFFFF), but your system still works correctly. So maybe this is not a big issue at this time...?

EDIT: Just saw an update of your earlier post:

"Finally after analyzing last modification ! This not a great idea .. just a mistake ! We always have 0xFFFFFFFF like on previous results .. this last modif just allow reading VID/DID in this false case !!!"

 

[CaseySJ]

Hi @CaseySJ,

Thanks for your Catalina installation guide. I followed your guide with most of the recommended hardware. Everything works fine except Airplay and Sidecar. I can Airplay to my homepod but fail to Airplay to my Apple TV. When I try to Airplay to Apple TV or Sidecar to my iPad Pro (2018 version), they just go to black screen. Any idea how to solve it? Is it related to the nvram.plist in the EFI folder (I haven’t copied the nvram.plist to the EFI folder)?

Keith

Hello @keithwongmk,

Let's address the Sidecar problem first:

• Do you have Location Services enabled in System Preferences --> Security & Privacy --> Privacy --> Enable Location Services?
  
• Is iGPU enabled properly? This can be checked by running IORegistryExplorer then scrolling the device tree on left side (do not use the 'search' box) until you get to IGPU@2. If the first child node below that is AppleIntelFramebuffer... then it's good.
• Of course, Bluetooth and WiFi must be enabled and working properly.

 

[rankrotten]

Hi everyone! For some strange reason, native NVRAM isn't working for me on Z390 Aorus Pro and Mojave 10.14.6 using those mini-guide @CaseySJ kindly published a few pages ago.

I was surprised a little bit, when @rankrotten instantly reported success here with NVRAM working on his Z390 Aorus Pro.

So, MSR 0xE2 was succesfully unlocked, SSDT-NVRAM.aml added, Clover RC scripts removed, KernelPM and AppleIntelCPUPM unchecked. System still boots normally.

But when I finally remove EmuVariableUefi from CLOVER/drivers/UEFI I instantly get panic on boot and system never loads. There's something about Kernel trap at 0x... type 14=page fault,... Fault CR2.

Maybe some of you guys have any ideas, why removing EmuVariableUefi may cause kernel panic.

Try my EFI if you are still having issues, it is working for native NVRAM on Z390 Aorus Pro. Have generated new serials so it is ready to go.
iGPU set to default settings of 64/256MB but only managed to load it in memmap by updating to latest BIOS (F12d) and performing a CMOS clear by shorting the two pins on MB.

 

[CaseySJ]

Hello, ive followed the instructions listed with one exception, unibeast was not behaving and only copying 5kb/s so i created the installation media through terminal instead, anyways i have a 9700k, designare, samsung 970, got the system booted, im just running intigrated graphics until amd releases the rdna2 card that supposedly smokes the 2080ti,

I like that plan!

sleep isnt working and hardware acceleration is not working, ive tried several methods including the hacktool guide outlined here and most result in a black screen half way through the boot,

Please post your config.plist from EFI partition of the macOS SSD. You may remove serial numbers from SMBIOS section before posting the file.

also as an aside is it possible to get a module to pop into the slot for the wifi card on the designare, i dont really care about bluetooth but it would be sweet to not have another pci card

Alas, nope! That slot is reserved for Intel CNVi modules only...unless someone can hack it.

 

[CaseySJ]

Even After Following every step got the error while booting directly from Catalina SSD Drive. Share the Screenshot for the same.

Please Help me out

Hello @luvlastcall

Can you please edit your post quoted above and delete the Catalina Mini-Guide?



You said that you enabled verbose mode, but your screenshot here and in your follow-up post do not show the verbose log. Are you sure you activated verbose mode correctly?

At Clover Boot Menu you can press spacebar then use arrow keys to move down to "Verbose" and press spacebar again to toggle it ON. Then press ESC to return to Clover Boot Menu and try booting again.

It seems you're using the old Catalina Mini-Guide. The new one is referenced in Post #1 or just click here.

I am getting no signal screen even after doing the fresh install and nothing is happening just getting no signal even after using -v argument while booting seems to be like still an issue with rx580 one. But yes I can boot with bootable usb and even if I choose to boot from catalina ssd I can get the screen. But if boot directly from ssd getting no signal screen every time. Sharing the screen shot for the same

Does this mean:

• If you boot with USB install disk, then choose Boot macOS from Catalina (internal SSD) then everyone is OKAY?
• But if you boot directly from internal SSD, then you get black screen (no signal)?

UPDATE: The quoted Catalina Mini-Guide has been truncated by site administrators.