Gigabyte Z490 Vision D (Thunderbolt 3) + i5-10400 + AMD RX 580

[emistral]

@

If the PCI path for the 6800XT is correctly specified, and Class Code (and other device properties) are set to specific values (such as FF FF FF FF for Class Code) then theoretically the device should be disabled. If IOReg shows that Device Properties that we want to inject are not in fact being injected, then:

• We'll need to double check all of the steps we took to modify those properties.
• macOS could be overriding the device properties.

How is macOS treating the 6800XT? Does it attach the default frame buffer to it (IONDRVFrameBuffer)? It would be helpful to see the IOReg file.

@CaseySJ Here is a screenshot from IOReg. It does attach the default framebuffer. Name and IOName device properties are injected correctly s per either th config.plist or the SSDT. But the class-code is not as you can see in the below screenshot.

 

[joe12south]

Hey there,

I have exactly the same problem. No idea how to solve it, and it doesn't really bother me that much, so I've just left it. Not much help, I'm afraid.

Although if you do find a solution, I'd love to know what it is...

Cheers,
V

After some more searching, it appears it might be a Radeon Vega thing that can only be solved with a GPU firmware update.

 

[CaseySJ]

@

@CaseySJ Here is a screenshot from IOReg. It does attach the default framebuffer. Name and IOName device properties are injected correctly s per either th config.plist or the SSDT. But the class-code is not as you can see in the below screenshot.
View attachment 513100

That is indeed strange. If you post the SSDT you're using to modify the device properties, I can take a closer look. Just to confirm: This is on the Z490 Vision D with AMD 6800 XT in the top long x16 slot?

 

[CaseySJ]

Unfortunately 11.3 Beta 4 is not stable for me. What I've observed so far:

• I believe changes were made to the NVMe driver (IONVMeFamily.kext) to address SSD caching issues with M1 Macs.
• Unfortunately, the new version has been crashing my system repeatedly. Kernel panics for the past 15 of 15 crashes since updating to Beta 4 have all occurred in IONVMeFamily.kext.
• Restoring back to Beta 3 fixes this problem completely.
• I have reported the last 5 of these crashes to Apple via the crash dump reporter.

Note 1: The problem may depend on make/model of NVMe SSD. My system has both Silicon Power and Sabrent Rocket (PCIe 3.0) NVMe SSDs.

Note 2: The problem may not occur on some motherboards. For example, my Gigabyte B550 Vision D (AMD Ryzen) with Silicon Power NVMe SSD and Sabrent Rocket (PCIe 4.0 version) remains rock solid on Beta 4.

** Update Big Sur 11.3 Beta 5 **

MacOS 11.3 Beta 4 was very troublesome as described in the post quoted above, so I had reverted back to Beta 3. Because Beta 5 was just released to the public, I decided to update from Beta 3 directly to Beta 5, but ran into the same problems as with Beta 4.

But there is good news...

It seems the culprit was my flashed GC-Titan Ridge. Although the Z490 Vision D has an on-board Titan Ridge controller, I am also using a flashed GC-Titan Ridge for a total of 4 Thunderbolt ports. During the macOS install process, however, a fatal error occurs while attempting to activate the XHC (USB) controller on this card.

Physically removing the card allowed the update to proceed flawlessly, and the system is stable.

Update:

• Installing the flashed GC-Titan Ridge after the update to 11.3 Beta 5 causes the system to crash after a couple of minutes. Same as with 11.3 Beta 4, but this does not happen in Beta 3 and earlier.
• So for now the flashed GC-Titan Ridge will remain disconnected.

 

[emistral]

That is indeed strange. If you post the SSDT you're using to modify the device properties, I can take a closer look. Just to confirm: This is on the Z490 Vision D with AMD 6800 XT in the top long x16 slot?

@CaseySJ Yes and no. Yes because this on on a Z490 Vision D with a 6800 XT. No because the 6800 XT is in the second slot. If I put it in the first slot and the 5600 XT in the second slot, I have only a couple of millimeters of clearance between the 2 cards. I am concerned that this will impact the airflow of the 6800 XT.
Attached is the SSDT.

 

[CaseySJ]

@CaseySJ Yes and no. Yes because this on on a Z490 Vision D with a 6800 XT. No because the 6800 XT is in the second slot. If I put it in the first slot and the 5600 XT in the second slot, I have only a couple of millimeters of clearance between the 2 cards. I am concerned that this will impact the airflow of the 6800 XT.
Attached is the SSDT.

I have uploaded new versions of the GPU disabling SSDTs to Post 1 of the Designare Z390 thread. Attaching them here as well. Please try the new SSDT-Designare-Z390-Disable-Slot-2-GPU.aml.

 

[ckrueger99]

The ASUS ThunderboltEX 4 AIC is now available on the ASUS website. I haven't seen any proper reviews of it yet. I'm also unclear what motherboard header is required and therefore which motherboards are compatible OOTB. (Is the compatibility list the same as for the tbEX 3 AIC?) Here is the Compatible Model List. It has a 14-1 pin TB header, which is the same form factor as the tbEX 3.

 

[zirkaiva]

I have uploaded new versions of the GPU disabling SSDTs to Post 1 of the Designare Z390 thread. Attaching them here as well. Please try the new SSDT-Designare-Z390-Disable-Slot-2-GPU.aml.

So How can I modify the SSDT if my ACPI path is:
PCIROOT(0)#PCI(0700)#PCI(0000)#PCI(0400)#PCI(0000)
ACPI(_SB_)#ACPI(PCI0)#ACPI(IOU1)#PCI(0000)#PCI(0400)#PCI(0000)
It is an expansion board connected to slot 2.

 

[imnotjack]

Hi again! I just came to a bottle neck and I can't seem to get my head around it as to what is actually happening. So, I use this new hackintosh for colour grading and video post. I have a Promise Pegasus RAID that works great! The thunderbolt is just super fast and usually makes the workflow great. But, as of right now, I don't have the necessary space to render a giant project inside the same RAID, so I plugged a USB 3.1 Adata Hard Drive to render a Master, which will be around 2-3Tb in size. But, whenever the computer is using the RAID, it seems that the USB starts working really slow, as a USB 2, but this only happens when I'm using the RAID, does this makes sense? I thought that this board had separate controllers and a extensive amount of USB... I've checked many times the BIOS config and is just as the post says, is it a OS thing?

 

[CaseySJ]

So How can I modify the SSDT if my ACPI path is:
PCIROOT(0)#PCI(0700)#PCI(0000)#PCI(0400)#PCI(0000)
ACPI(_SB_)#ACPI(PCI0)#ACPI(IOU1)#PCI(0000)#PCI(0400)#PCI(0000)
It is an expansion board connected to slot 2.

We will need to see:

• Make/model of the expansion board
• IOReg file
• System DSDT