Contribute
Register

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

Malenca said:
I think I learn in the hard way....
Any idea why on radeon Vega 64 there is not much improvement on the tests??
Click to expand...
Because you have a Vega, it's necessary to copy the special version of Radeon Boost driver: RadeonBoost-v1-4+Vega.kext

Copy this into CLOVER/kexts/Other and reboot.
 
iRamon said:
Thank you @CaseySJ
Please find my Thunderbolt SSDT attached below. It's the SSDT from the DROM guide, customized with a unique UID and checksum. I also tried the default SSDT from the guide, and it didn't make any difference.

Again ... no hurry here. With al the incredible work you're doing here for all of us, you are more than entitled to have a wonderful weekend!


I saw your reply to @NCMacGuy, it's crazy how quick you figured it out! But what puzzled me is why the scratch areas and active region would have any differences between boards on a ROM chip (as in read-only)... My guess is that I wouldn't understand the answer to that anyway :lol:

UPDATE: I missed your request to post output of log show --last boot | grep Thunderbolt last night.
A text file with the output is attached below. I noticed that waking my system after sleep created additional entries in the log (NVM not loaded, and No DROM found) ... might reveal a clue of the issue I'm dealing with?
Click to expand...
Hi @CaseySJ ... don't bother looking any further, I doubt you'll find anyhing.

I've been going though an elaborate process of elimination tests, and none have resulted in any change in the two issues since flashing Thunderbolt firmware:
  1. The missing Thunderbolt XHC/USB controller that provides the SSP1 and SSP2 ports.
  2. The odd information in System Report > Thunderbolt.
A list of things I've tested to isolate the issue:
  • macOS installation/configuration
    • Boot from working backup with Catalina 10.15.3.
    • Replace all my kexts, efi drivers and SSDT's with the ones from your Catalina 10.15.4 mini-guide.
    • Fresh Catalina 10.15.4 installation to a spare SSD, following your guide and using kexts, efi drivers and SSDT's exactly as provided.
    • Replacing SSDT-Z390-DESIGNARE-TB3HP-V4 with SSDT-TBOLT3-RP05-PORT7-DESIGNARE-Z390 on both installs (normal and fresh installation on spare).
    • Tried SSDT-TBOLT3-RP05-PORT7-DESIGNARE-Z390 with Device XH3 renamed to XH5.
  • Hardware
    • Disconnect cable from RX590 DP-OUt to Designare DP-IN.
      • Note: I found odd behaviour with hot-plugging or hot-unplugging the DP-cable. My system would freeze completely (mouse, kb, video, audio) and I had to hard reboot.
    • USB: Unplug internal USB headers (F_USB, F_USB30, F_USB31C)
    • PCIE: Remove my AMD RX590 and run with iGPU only.
    • PCIE: Remove Fenvi FV-T919 card.
  • BIOS
    • Thunderbolt: Tried running macOS with different Security Levels (No Security, User Authorization, USB Docking Only).
      • Note: Feels as if switching between TB security levels took longer (approx 5 seconds) since flashing patched TB3 firmware.
    • Thunderbolt: Tried running macOS with GPIO3 Force Power disabled (instead of enabled).
    • Thunderbolt: Tried running macOS with USB Support disabled (instead of enabled).
    • RAM: Change DRAM profile from XMP Profile1 (which runs at 3600 MHz) to fixed 2666 MHz.
    • Physically removed battery from motherboard.
    • Flashed BIOS F8 + CMOS reset + MSR 0x2 Unlock.

The only elimination test left (that I can think of) is installing Windows, but I seriously doubt it will make any difference.

At this point, I feel we can rule out macOS installation and configuration. Also BIOS configuration and BIOS version.
Leaving the cause to be any of the options below:
  • The firmware as present on my chip is corrupted, causing Thunderbolt XHC to malfunction.
    • Least likely, since read checksum after flash was identical.
  • The Winbond chip got damaged (physically or electrically) as result of me attaching the SOIC8 clip and/or flashing.
  • Something else on my motherboard related to the Thunderbolt XHC USB Controller got damaged (physically or electrically).
To eliminate corrupted firmware, I'll try flashing the Winbond again with the original firmware. I'll use @NCMacGuy 's backup.bin he posted here, as my original firmware backups had so many different checksums.

If that does not fix my Thunderbolt XHC issues, I must have fried the chip and/or motherboard :cry: ... and will order a new Designare.
 
Last edited:
CaseySJ said:
Okay, let's try this:

FIND: 085F4849 440C41D0 0501
REPLACE: A3A3A3A3 A3A3A3A3 A3A3


​
EDIT: Another option is to rename PNP0501 to PNP0FFF as follows:
  • Find: 50 4E 50 30 35 30 31
  • Replace: 50 4E 50 30 46 46 46
Click to expand...
Thanks @steve3d for bringing this up!

@CaseySJ
The first worked for me, the second did not. You solved a problem I didn't even know I had. If you have time, how did you come up with this? I get what you're doing in the AML, but how are you converting between the two to get the hex? Just trying to understand.
 
@iRamon,

I also couldn't find anything wrong in the files you posted. Before you do anything further, please try this:
  • Download tbpatch (attached) to the Downloads folder and unzip.
  • In Terminal type this:
Bash: 
cd ~/Downloads
chmod +x tbpatch
xattr -d com.apple.quarantine tbpatch
sudo ./tbpatch list
  • You should see something like this:
Code: 
/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/RP05@1C,4/IOPP/UPSB@0/IOPP/DSB0@0/IOPP/NHI0@0/AppleThunderboltHAL/AppleThunderboltNHIType3/IOThunderboltController/IOThunderboltPort@7/IOThunderboltSwitchType3/IOThunderboltIECSNub/AppleHPMIECS/AppleHPMDevice@0
  Address : 0x00000000
  PID     : 0x2831454341
  UID     : <REDACTED>
  Version : 8
  Build   : a0566aa071c0a0cabf803a9532caee53ddd2ced5_03012018
  Device  : TPS65983 HW0030 FW0001.38.04 ZTBT1

/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEG1@1,1/IOPP/UPSB@0/IOPP/DSB0@0/IOPP/NHI0@0/AppleThunderboltHAL/AppleThunderboltNHIType3/IOThunderboltController/IOThunderboltPort@7/IOThunderboltSwitchType3/IOThunderboltIECSNub/AppleHPMIECS/AppleHPMDevice@0
  Address : 0x00000000
  PID     : 0x2831454341
  UID     : <REDACTED>
  Version : 8
  Build   : 5a8f092c7306c3d9c0764d2f4f132044a8f0cedf_09042017
  Device  : TPS65983 HW0030 FW0001.38.03 ZTBT1
  • Notice the two <REDACTED> items above next to UID.
  • Copy the first UID to clipboard (CMD-C).
  • Then type this, but paste the UID as shown:
Bash: 
sudo ./tbpatch dump -u paste-uid-here -o 0x4000 -s 0x400 -f TB3-Fw-Extract.bin
  • Use HexEdit to examine the resulting file TB3-Fw-Extract.bin and also post it here.
 

Attachments

  • tbpatch.zip
    24.4 KB · Views: 79
CaseySJ said:
There might be a much easier solution, but first:
  • Which GPU are you using on macOS?
  • Which slot is it in?
  • Which slot is the other GPU in?
Click to expand...

WX7100 - macOS - Slot 1

RX580 - Windows - Slot 2

However, I use a secondary display connected to the RX580 in macOS. I could just move it to the WX7100 and disable the secondary gpu?
 
rj510 said:
The APGM section is still useful, so I'm using it, as is, on one machine with an RX580, where I noticed no difference.
Click to expand...

So what are the benefits of this "APGM" section improvement for the RX series?

Thanks a lot!!
 
gandem said:
Thanks @steve3d for bringing this up!

@CaseySJ
The first worked for me, the second did not. You solved a problem I didn't even know I had. If you have time, how did you come up with this? I get what you're doing in the AML, but how are you converting between the two to get the hex? Just trying to understand.
Click to expand...
It's...elementary... :)
www.tonymacx86.com

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

don't use that software on your main config.plist...... It definitely has messed up my older versions, but I keep trying it with each release and have had no issues with 056 yet. Thats a cool first!
www.tonymacx86.com www.tonymacx86.com
 
guiscard said:
I also get a kernel panic without WEG. This is with a Radeon VII.

With just the kext and no SSDT I get much better synthetic scores without any extra fan noise. Tomorrow I have some real work to do and will check actual speeds.
Click to expand...

Really looking forward to seeing these real-world numbers!
 
Current Rig (Shoutout to Casey!)
Z390 DESIGNARE |  i7 9700K |  Vega 56 8GB
Retired Rig (7 years!)
GA-Z77-DS3H |  i7 3770K |  GeForce GTX 760
Hi @steve3d.

Would you be willing to share your EFI- minus serials of course. We have identical hardware and I'd love to try out OC. I have flashed my GC-Titan Ridge to the patched firmware in the repository. Have you tried this? Just curious about your experience, although you don't have TB devices. Wondering if OC would improve my TB3 experience. I'm also running the Radeon Boost kext which has given me significant bench score improvements for my RVII.
 
You must log in or register to reply here.
Copyright © 2010 - 2024 tonymacx86 LLC
Back
Top