Contribute
Register

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

congrat casey. but your DTPG.aml source from KGP's code not CJOE. CJOE has been stolen that file. btw should be change author name of SSDT-DTPG.aml from CJOE to KGP.

how do you think?
Sure, no problem. I'll rename the author and upload a new ZIP.

Update: SSDT-DTPG.aml has been replaced by KGP's version from his Github Repository. I tested and verified this version.
 
Last edited:
This build was for video editing on Adobe Premiere pro and DaVinci Resolve.

FYI

Thunderbolt
The Thunderbolt works great on my build. I'm using the Startech Thunderbolt 3 to Thunderbolt adapter to connect to external TB drive and also to a broadcast monitor through the Blackmagic Ultrastudio mini monitor. Everything works great and is hot plug-able. From what I gathered, the type of adapter and the length of thunderbolt cable does make a difference to the connection.

Wifi/Bluetooth
I use the Fenvi wifi/bluetooth adapter and installed according to Casey's guide. Works great!

Display
I connect 2 dell monitors through the HDMI and display port on the RX580.

This build was very smooth, thanks much to Casey comprehensive guide. The only caveat is my Noctua U14s CPU cooler was a little too tight for the RX-580 to be on the first PCIe slot, I have to install it on the second slot instead.
 
All:

Because the General Framebuffer Patching Guide is applicable to many other systems, the best place for it is in:

Post Installation --> Graphics

The moderators have moved and pinned the post to that forum.
 
This build was for video editing on Adobe Premiere pro and DaVinci Resolve.

FYI

Thunderbolt
the thunderbolt works great on my build. I'm using the Startech thunderbolt 3 to thunderbolt adapter to connect to external TB drive and also to a broadcast monitor through the Blackmagic Ultrastudio mini monitor. Everything works great and is hot pluggable. From what I gathered, the type of adapter and the length of thunderbolt cable does make a difference to the connection.

Wifi/Bluetooth
I use the Fenvi wifi/bluetooth adapter and installed according to Casey's guide. Works great!

Display
I connect 2 dell monitors through the HDMI and display port on the RX580.

This build was very smooth, thanks much to Casey comprehensive guide. The only caveat is my Noctua U14s CPU cooler was a little too tight for the RX-580 to be on the first PCIe slot, I have to install it on the second slot instead.
Thank you for the positive comments! :)

Which specific Thunderbolt 3 external drive enclosure are you using? I would like to add make/model to the build guide, Thunderbolt 3 Experiences.
 
The sample IOReg in the screenshot is actually the bare minimum -- and it shows no detected devices. Only the controller itself appears in the tree. Because Thunderbolt 2 and Thunderbolt 3 devices by Universal Audio seem to work fine (several of them have been tested successfully), I would tend to agree with your assessment about compatibility with Thunderbolt 1.

As @AlexD pointed out earlier today, a real 2018 Mac with Thunderbolt 3 (e.g., MacBook Pro 2018) has both the older Alpine Ridge controller and the newer Titan Ridge controller. It's very curious why Apple chose to keep the old controller, and we're wondering whether it has something to do with maintaining compatibility with old Thunderbolt 1 and 2 devices. If that's true, then our Titan Ridge based system might simply not work with some of those older devices. BUT we don't know this for sure.

So my EFI folder as it is should work for "normal" Thunderbolt 3,2 and 1 devices? Should the ioregistry only show the bare minimum or should there be more?

Is the problem simply the adapter, that fails to convert the Thunderbolt 1 signal correctly, or are more complex Thunderbolt 1 devices simply not recognizable in titan ridge environments?

Maybe I'm misunderstanding what a Thunderbolt controller is, but isn't this a physical controller mounted physically on the motherboard? So apple has both a Titan Rdge and an Alpine Ridge controller on the same motherboard and those work on both ports all the time? If true, apple really outdid themselves on this one...

I will try a Startech Thunderbolt 3 to Thunderbolt 2 adapter tomorrow and see what happens. Startech claims to be Mac as well as Windows compatible, but I'm skeptical now that the official apple one is not working.

The Desginare board does have a Thunderbolt header... would a Thunderbolt 1 PCIe card be the solution?
 
Curious about the BIOS. Here's what my Peripherals page looks like. And by double-clicking the top Ethernet device, we get the second image that contains the PCI Device ID.
View attachment 377380 View attachment 377381

In addition to Device ID and Vendor ID, we also need to determine the Sub Device ID and Sub Vendor ID. Here's what these terms mean:
  • Let's say Aquantia makes the 10GbE Ethernet chip. These chips can be used in Aquantia's own products or third parties can buy them for their own products. In both cases, the Device ID and Vendor ID will refer to the Aquantia chip.
  • But if company "SmallTree" for example uses this chip in its own product line, then they will be the "sub vendor" of the chip. And the Sub Vendor ID and Sub Device ID will be assigned by SmallTree.
  • So if we plan to use SmallTree's driver, the problem is that the driver will look specifically for SmallTree's own Sub Vendor ID and Sub Device ID. Sometimes only the Sub Vendor ID will be different (i.e. the Sub Device ID might be the same as the original Device ID).
  • If this happens, we can spoof the Sub Vendor ID and, if needed, the Sub Device ID to match SmallTree's IDs. And then we can happily use SmallTree's driver.
  • This is what was done to enable the Intel i211 Ethernet port on the Designare Z390. For this motherboard, the Sub Vendor ID belonged to Intel, but the SmallTree driver (that supports the i211) was looking for its own Sub Vendor ID.
So the next task is to see if you can determine the Sub Vendor ID and Sub Device ID for this. Hopefully you can find this information from Windows.

I've made a few other changes to your config.plist (added ACPI patches for USBInjectAll). This version may or may not make a difference, but it's worth trying (file needs to be renamed of course). Also, try moving your USB flash disk to one of the two black USB 2.0 ports on the rear IO panel if not there already.

@CaseySJ - So, I think I have found what may be the issue, but I am not sure it is helpful. I have tried your V3 config and I am having the same issue.

First off, attached is my BIOS images, it only details the Intel NIC and does not speak of the Aquantia. Secondly, taking more images of the panic, I see a few interesting things:

"Invalidation IPI Timeout, unresponsive CPU bitmap 0x1101"
"Create IOReporters ..... No service found"
"configure interests ...... failed to configure interests"

These all seem like bad things to me, is this ringing any bells?

Also, attached is my current EFI - am I missing any kexts that I am being dumb and not including?

Thank you again for all the help.
 

Attachments

  • IMG_3519.jpg
    IMG_3519.jpg
    5.7 MB · Views: 84
  • IMG_3520.jpg
    IMG_3520.jpg
    5.1 MB · Views: 99
  • IMG_3523.jpg
    IMG_3523.jpg
    7.1 MB · Views: 94
  • IMG_3525.jpg
    IMG_3525.jpg
    7.2 MB · Views: 90
  • IMG_3526.jpg
    IMG_3526.jpg
    7.2 MB · Views: 84
  • IMG_3527.jpg
    IMG_3527.jpg
    7.8 MB · Views: 88
  • EFI - 2019-01-09.zip
    4.4 MB · Views: 106
So my EFI folder as it is should work for "normal" Thunderbolt 3,2 and 1 devices? Should the ioregistry only show the bare minimum or should there be more?
As soon as a valid TB3 device is attached, a sub-node in the tree will expand. Here the sub-tree belongs to my OWC 14-port TB3 dock.

Screen Shot 2019-01-09 at 8.47.55 PM.png

Is the problem simply the adapter, that fails to convert the thunderbolt 1 signal correctly, or are more complex thunderbolt 1 devices simply not recognizable in titan ridge environments?
This is the question we're wondering ourselves. Why would Apple, a company known for cutting unnecessary components and headphone jacks, include both the Alpine Ridge and Titan Ridge controllers? We think it may have to do with *some* TB1 and TB2 devices, but we don't know for sure.

Maybe I'm misunderstanding what a thunderbolt controller is, but isn't this a physical controller mounted physically on the motherboard? So apple has both a titan ridge and an alpine ridge controller on the same motherboard and those work on both ports all the time? If true, apple really outdid themselves on this one...
Yes they work on all ports at all times. The parent node is controlled by Alpine Ridge, but its first child node is a Titan Ridge controller. Apple has certainly created a hybrid.

I will try a startech thunderbolt 3 to thunderbolt 2 adapter tomorrow and see what happens. Startech claims to be mac as well as windows comptatible, but I'm skeptical now that the official apple one is not working.

The desginare board does have a thudnerbolt header... would a thunderbolt 1 pcie card be the solution?
Are you sure it has a Thunderbolt header on the motherboard? ;)
 
Lastly, I tried it with Clover 4843 just to make sure that this wasn't just an issue of clover being broken, and I had the same issues.

"expected 16 acts but received 8 after 11068200 loops in ....." immediately precedes the panic.
 

Attachments

  • EFI with Clover 4843 as test.zip
    4.5 MB · Views: 103
@CaseySJ - So, I think I have found what may be the issue, but I am not sure it is helpful. I have tried your V3 config and I am having the same issue.

First off, attached is my bios images, it only details the intel nic and does not speak of the aquantia. Secondly, taking more images of the panic, I see a few interesting things:

"Invalidation IPI Timeout, unresponsive CPU bitmap 0x1101"
"Create IOReporters ..... No service found"
"configure interests ...... failed to configure interests"

These all seem like bad things to me, is this ringing any bells?

Also, attached is my current EFI - am I missing any kexts that I am being dumb and not including?

Thank you again for all the help.
I see what you mean about the missing Aquantia device. But it may not be the cause for the kernel panic. Even in normal boot sequences there are a number of error messages encountered along the way, but they aren't fatal. So issues like "No service found" or "Failed to configure interests" may be benign.

I would suggest capturing a video of the verbose boot sequence with your smartphone and repeating the boot sequence 3 times to see if the crash occurs at the same place every time. If so, one can scrub the video and capture the last 3 or 4 frames.
 
Back
Top