Contribute
Register

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

Let's keep Verbose enabled for the time being (position of the progress bar is not very useful).

Is your display monitor connected to the Vega 64 or to HDMI port on motherboard?


Vega 64, with motherboard settings accordingly
 
if I boot from the iGPU with V2 plist, I no longer get stuck at the same point. I now get right past that to the machine check of the CPU. let me try with V1 plist real quick.

Edit - no dice, will try with your new EFI below. Thanks!
 

Attachments

  • IMG_3506.JPG
    IMG_3506.JPG
    2.9 MB · Views: 258
Last edited:
I think it's a good idea to stick to iGPU for now. Attached is a modified EFI that should enable HDMI port on the motherboard. It uses Lilu and WhateverGreen in addition to USBInjectAll and XHCI-Unsupported.

Your existing EFI folder (the entire folder) should be backed up and fully replaced with the attached version.

Standard Disclaimer:
Because this is a different motherboard from the Designare, I cannot test the changes I'm making, so the changes are all "theoretical".
 

Attachments

  • EFI-frickeregg.zip
    4.4 MB · Views: 247
@CaseySJ

Thank you , I used your method for UHD 630, and is working.

and I may have found a solution for the glitches, just switch to iMac18,1. (found the solution online).
 
@CaseySJ , OK, so I think we are getting somewhere. Now, I get hung at "AppleEthernetAquantiaAqtion: :start :321 : delay for fw upload.

My motherboard has an onboard 10GBe Nic, an aquantia nic that is supposed to be the same one in the new macs. This is one difference between our Mobos. I saw someone referencing a south tree kext that could be used to try to fix this, so I may try to do some reading on that. I wouldn't know how to include it though.



Late Edit - Actually, loading it again, it looks like it loaded the appropriate firmware "AEthernetAquantiaAqctionFirmware 1.0.35" and went on, but got stuck just after @kec.corecrypto 1.0 - this is right before it previously told me some info about the CPU.

Edit V2 - now it is hanging on "virtual IOReturn IONVMeController: :CreateSubmissionQueue(unit16_t, uint8_t)::2777:SQ index=1 entrysize=64". Thumbnail updated to reflect this.
 

Attachments

  • IMG_3509.JPG
    IMG_3509.JPG
    3.7 MB · Views: 271
Last edited:
@CaseySJ , OK, so I think we are getting somewhere. Now, I get hung at "AppleEthernetAquantiaAqtion: :start :321 : delay for fw upload.

My motherboard has an onboard 10GBe Nic, an aquantia nic that is supposed to be the same one in the new macs. This is one difference between our Mobos. I saw someone referencing a south tree kext that could be used to try to fix this, so I may try to do some reading on that. I wouldn't know how to include it though.
Interesting...I would have expected it to simply timeout and proceed. Anyway, let me find the SmallTree kext and upload it with instructions. One moment...
 
Plz Post that youtube link !!!

Sure, here's what I've found. Actually, that one with Clover and BIOS video was not a Z170 but Z87X.

Asus Z370-A + ThunderboltEX 3:

Gigabyte GA-Z170MX:

Gigabyte Z87X (this one is actually allows BIOS/Clover):

Asus Sabertooth Z87 + ThunderboltEX 2:

Anyway, this still remains a complete mystery for me, why Apple Thunderbolt Display doesn't work with Titan Ridge onboard controller of Z390 Designare motherboard. JHL7540 in its current implementation seems to be not able to work in pure Thunderbolt switch mode which is absolutely mandatory for Apple Thunderbolt Display. Be it Windows, Ubuntu or macOS. I've tried everything and still no luck. None of the OS is able to see the monitor.

Gigabyte's engineers definitely broke something that worked fine before. BIOS and onboard TB firmware need significant work now to properly support Apple Thunderbolt monitor.

Accordingly to official Thunderbolt 3 Technical Brief:
https://thunderbolttechnology.net/sites/default/files/Thunderbolt3_TechBrief_FINAL.pdf

There are three different operating modes for Thunderbolt silicon. USB, DisplayPort and Thunderbolt. Depending on what device is being connected, the chip switches its operating mode. So JHL7540 should switch when it gets the TB device, but it doesn't.

It becomes clear for me, that for some reason Titan Ridge on Z390 Designare can operate only in USB or DisplayPort mode. And completely ignores pure Thunderbolt mode when working with Thunderbolt monitors (pure Thunderbolt, not just DisplayPort over USB-C capable).

It also completely explains the fact, that people are currently only able to connect USB devices (various drives and TB storage solutions, external audio interfaces) as well as displays that require only DisplayPort portion (they are typically offering fallback via DP over USB-C instead of being a pure Thunderbolt capable).

In conversation with Intel engineers I've been told that there is also a power delivery controller which is a specific companion to TB chip. It helps to properly select and switch mode and adjust the power settings so the device can connect. And it is all currently up to motherboard vendors how to properly implement these things both in hardware (actual PCB layout) and software (BIOS and firmwares).

My guess is this issue may be due to faulty or very early BIOS or Thunderbolt or Power Delivery controller firmware code for this motherboard model. I already sent an eSupport ticket to Gigabyte where I described this problem. Hopefully, they'll fix something in the next BIOS updates.

On our side of things, there's also a lot to be done to get a proper Thunderbolt support in macOS. Accordingly to @kgp researches it is still not possible to achieve a complete, fully initialized Thunderbolt Local Node implementation on Hackintosh. When it's done right, it is a completely expanded and filled with devices (all switches and TB ports with properly initialized and assigned DPIn and PCIDown Adapters) sub-tree inside NHIType3 Controller device in IOReg.

Unfortunately, in our Hackintoshes, we're only getting empty IOThunderboltLocalNode and IOThunderboltPort in the end of that tree. Of course, we have implemented DSB1 and DSB4 as well as DSB2 with mapped Titan Ridge XHC controller ports, but that's still not enough.

I do believe that with proper and complete Thunderbolt SSDT patch or even better – with a proper Thunderbolt kext (just like we use for Ethernet or Bluetooth or USB or anything else) or even with a EFI driver, it will be possible to fully initialize the Titan Ridge on Hackintosh.

Now we're only half way there. But we'll make it happen one day!
 
@frickergg MacOS should natively support several Aquantia 10GBe cards, but yours may not be one of them. The first step would be to grab a BIOS image of the Peripherals and Chipset pages. The device ID of this item should be in one of those sections of BIOS...
 
@Racke @xanderevo

Sorry for bothering, but I seems that I cannot get the IGPU into headless mode, because FCP still crashes during startup.

My system:
  • i5 9600k
  • RX 580 (PCIe slot 2)

BIOS:
  • Version 1.4
  • Peripherals - Initial Display Output = PCIe 2 Slot
  • Chipset - Internal Graphics = Enabled

config.plist
  • Devices/IntelGFX = 0x0
  • Graphics/FB Name = Orinoco
  • Graphics/ig-platform-id = 0x
  • SMBIOS: iMac18,3 / iMac18,2 / iMacPro1,1

The FCP crash report is the same as described in another thread.

Any ideas?
I have bios 6a, pci1,imac18,
Please look at my first post and compare my config.plist for the correct framebuffers, device ids.

The rest of my efi is like Casey described in his guide
 
@CaseySJ

Thank you , I used your method for UHD 630, and is working.

and I may have found a solution for the glitches, just switch to iMac18,1. (found the solution online).
Glad to hear it works on the Gigabyte Z390 AORUS PRO as well, and thanks for the tip regarding iMac18,1!

Will try it later this afternoon.
 
Back
Top