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

[CaseySJ]

UPDATE: I'm honestly not sure how, but after booting into the Install MacOS Catalina USB and running Disk Utility, it appears my main drive in the hackintosh was pretty much wiped. I had about ~100GB on there and now it's virtually empty. Proceeding with a fresh install again. Luckily all my important things were backed up elsewhere. YIKES.

-----

I just had a monumental crash that I'm worried I won't be able to recover from without some help.

Things have been going along swimmingly with my Hack for the last couple of weeks, and then today I ran Carbon Copy Cloner to refresh my backup from a week ago. I may very well have screwed myself by doing this, but since my install (as of this afternoon) seemed like the perfect starting point for a restore should I have any issues going forward, I decided to create a fresh backup and ignore the one I made a week ago. I know...sounds like famous last words to me, too. But anyway I re-formatted my backup drive and ran the task--still saved from last week--again anew. I had the EFIclone script and everything in place as before, and as the full backup was taking awhile I left the room and had dinner.

When I came back, the display had gone to sleep but the system itself was still up (I have "prevent computer from sleeping automatically when display is off" enabled). In the upper right hand corner I had a notification that the backup was complete, and a prompt to log back in. But when I typed in my password, it was rejected. I've had the same password on my computer for a long time so I'm sure it wasn't wrong. I typed it again, and it kept rejecting the password. I tried it again bypassing the numeric keypad altogether, and still no dice. I then selected "switch user." I don't know why exactly, but I guess I wanted to see if my other user account would let me in.

Upon doing that...the screen went blank and I had only a rainbow wheel of death cursor. After about ten minutes I restarted the computer. From that point on to now the Hack simply will not boot. It goes directly to the BIOS no matter what volume I try, including my original Install Catalina USB stick. To be clear, since getting everything up and running I've had zero serious issues like this. The hack has basically been running like any other Mac.

- I've checked to see if the motherboard relocked itself, but the value of 0x5C1 was still set to 0x0
- I booted from the Install macOS Catalina USB stick and got to the Clover menu, but the actual MacOS installer doesn't get any farther than my main macOS drive

This is the error I'm getting when I try to boot either from my internal m.2 or from the USB installer:

View attachment 468753

I've seen others post this error but it seems to have a few different possible causes. I tried resetting the NVRAM in the Clover boot loader from the USB using F11, but that also did not help.

So I booted up my old Mini (which is where I'm typing this from) and checked out the contents of my bootable backup. For some reason...everything on the backup is empty. I've got folders for everything, but no files. My Photos library is there, but that is essentially a folder as well (or a package in any case), and its size is 0 KB. My only thought is that while CCC was running something odd might have happened to my EFI partition, which the EFIclone script opens to do its thing.

I suppose it's possible I hit revert instead of clone...but I distinctly remember clicking the clone button, and watching the progress as it copied more and more files TO the USB drive.

At this point the only thing I can think to try is a fresh re-install...but I can't even get the USB installer to boot. I'm at a loss! But will keep plugging away trying to at least get SOMETHING to boot...if anyone here has any ideas let me know.

Ouch... let's first do a CMOS Reset by shorting the two designated pins with a flathead screwdriver (please see Designare User Manual for proper instructions). This will revert all settings to factory defaults so configure BIOS parameters again, starting with Load Optimized Defaults. Then unlock MSR 0xE2 once more.

 

[CaseySJ]

Thanks, in the debug log (Thanks @MuffinCrumbs) I noticed it didn't load, removed it from config and added it back to the config.plist and now it works..‍♂

Now I'm trying to get the iGPU active for accelerated rendering while using my Radeon to drive the screen. (plaform-id set to 0300923E but no success yet)

On 9th Gen CPUs we should use Platform ID 0300983E.

 

[CaseySJ]

Regarding Flashing the TB.

I've just managed to install Windows 10 via Opencore 0.5.8 on a separate NVMe Dual Boot, ( thanks @CaseySJ )
Trying to do some streaming with OBS and Mac 10.15.4 version crashes (developer bug), so decided to try Windows..

Now.. because I've flashed my Thunderbolt will it work or be seen by the Windows side?

My UAD Apollo 8 with TB2/3 Adapter which now runs great on the Mac side is not recognised on the Windows side.
So I thought it's obvious because I've flashed the TB it won't be recognised..
However, yesterday for a short period it showed up in Windows 10.. Can't replicate now though.

Any ideas?? TIA

When on-board Thunderbolt firmware is flashed, then on a cold start (i.e. with PSU turned off or power strip turned off) we must boot into macOS first and then warm-boot into Windows.

 

[WhosMac]

Please check the 4 tiny LEDs on the bottom right of the motherboard. They should turn on and quickly turn off, but if one or more of them remains on then it indicates a problem with that component.

I tried to see if it would but none stay on, it does it’s boot sequence to check components like the z390 normally does from a cold boot after CMOS and flashes them one by one till it gets to boot. But none stay on once it has cycled thru them

Edit: It power cycles 3 times after CMOS, which in the past is normal while it does its checks
Cpu light flash then reset
Cpu and dram Flash then reset
Cpu, Ram, and vga flash then reset
And then it cycles thru the 4 and makes it took boot led and makes it to bios

 

[faithie999]

I've done lots of testing/experimenting with this the last couple of days. I think we are getting closer but not there yet.

First, thanks to @CaseySJ for translating the DROM property from OC to Clover.

Netting out lots of cold boot/warm boot/CMOS reset and BIOS reset:

My goal is for TB display and display peripherals (mic, display audio out, camera) to work after booting. Sometimes, the display and peripherals work; sometimes, the display works but not the peripherals; sometimes the peripherals work, but not the display. (When the display doesn't light up after the boot is complete, I turn on a secondary display that I have connected via DisplayPort from my RX 590, and I can test the camera, mic, and display sound.)

Sometimes, when I get an "incomplete" boot (display but no peripherals or peripherals but no display), in order to have the TB display light up at the beginning of the next boot, I need to either do a cold boot (unplug power plug from PSU) or sometimes I need to do a CMOS reset.

One thing I've just discovered is that, when I have a successful boot and when the monitor is connected to TB output 1, the TB tree in IOReg is very different from the tree when the monitor is connected to TB output 2. In the first instance, the tree shows up under DSB1, and in the second instance it shows up under DSB4. Both IOReg outputs are attached.

After an incomplete boot, the TB tree shows up with the display attached to Port5. even though the tree shows the camera and display audio connected, those devices don't work. Output attached.

I've tried connecting a USB-c to USB 3 dongle to the unused TB port, but the USB stick I have plugged into the dongle isn't recognized. This isn't a problem because I don't need more USB ports.

Interestingly, prior to booting, when I attach a TB1 hard disk dock to the TB connection on the back of the monitor, BIOS recognizes it, and I am able to boot from a HD loaded in the dock. Not something I plan to do but interesting nonetheless.

You said to look for a checksum error in Hackintool-->logs by searching for DROM. no error shown. Output attached.

Finally, you said to run gfxutil to see if NHI0 was properly configured. It shows as
PCI0.RP21.UPSB.DSB0.NHI0 = PciRoot(0x0)/Pci(0x1b,0x4)/Pci(0x0,0x0)/Pci(0x0,0x0)/Pci(0x0,0x0)
I don't know if that means I need to change the path in Devices--property from Pci(0x1c,0x4) to Pci(0x1b,0x4). But ,when I do so, I never get a boot with the peripherals working. To get back to Pci(0x1c,0x4) in DROM properties, I usually need to do a CMOS reset. Output attached.

I also tried a wake from sleep, which doesn't work when I have flashed a modded firmware. That is something I really want to have working. There are two types of wakes that I have experienced--one is the display wakes up with one keyboard press but peripherals don't work, and the other is the system is frozen when I press a key.


Sorry for the long message. But I wanted to supply as much info as possible.

UPDATE:

subsequent to the above posting I realized I had never methodically tested the TB monitor connected to TB port 2 on the AR card. I did so and my machine has successfully completed a couple of sleep/wake cycles (display and peripherals working after wake), which never happened when attached to TB port 1. as I mentioned earlier, the IOReg TB tree is very different port 1 vs port 2.

I have attached IOReg for port 2 connection.

also, IIRC when I was using Elias64Fr's patched firmware and the DROM SSDT from the guide, IOReg showed the device attached to ThunderboltPort7. what is the significance of Port 7 vs Port 5?
thanks!

 

[CaseySJ]

I tried to see if it would but none stay on, it does it’s boot sequence to check components like the z390 normally does from a cold boot after CMOS and flashes them one by one till it gets to boot. But none stay on once it has cycled thru them

Edit: It power cycles 3 times after CMOS, which in the past is normal while it does its checks
Cpu light flash then reset
Cpu and dram Flash then reset
Cpu, Ram, and vga flash then reset
And then it cycles thru the 4 and makes it took boot led and makes it to bios

Try reseating all PCIe devices and reconnecting all cables.

If that doesn't help, try removing all PCIe devices one by one. We can even more GPU and set Platform ID to 0x3E9B0007 and connect monitor to on-board HDMI or Thunderbolt port.

 

[faithie999]

Good to know!

If anyone wants to test:

1) With TB Device on, fully enter sleep and then wake. The device shouldn't reconnect.
2) Reboot. Device shouldn't connect as, in my case, it needs a shutdown and unplug.
3) Check IOResources.

I'm wondering if this only applies TB via card. In other words, if you attempt sleep and break your TB and it's simply fixed by a reboot then it's likely you won't get the same missing USB power info.

Anyways, the hope is that the next time someone attempts to tackle the sleep issue is that this is a useful clue.

Happy Mothers days where it applies!

that wake behavior had been my experience with my flashed AR card, so I was ready to give up on the AIC approach.

then I tried @dgsga's SSDT-only approach for Alpine Ridge (see post 19,995). after help from @dgsga and @CaseySJ, and a few fits and starts, I have that approach working (see post 20,687)--but for some reason, successful wake only works when the Apple TB display is attached to TB port 2 on my AR card. when attached to port 1, I get the behavior you describe.

 

[michelb76]

On 9th Gen CPUs we should use Platform ID 0300983E.
View attachment 468785

That didn't do much. I followed the opencore guide to set this up, I have different entries, is there something else I can check? I connect my display to the Vega through hdmi atm, thta works fine. I just would like to use the IGPU to accelerate video rendering.

 

[CaseySJ]

That didn't do much. I followed the opencore guide to set this up, I have different entries, is there something else I can check? I connect my display to the Vega through hdmi atm, thta works fine. I just would like to use the IGPU to accelerate video rendering.

Why do you say that iGPU acceleration is not working? In other words, what symptom or system report are you looking at? Have you looked at IORegistryExplorer --> iGPU@2 to see it AppleIntelFramebufferController appears as a child device as shown below? If this appears, it means iGPU is accelerated.

 

[CaseySJ]

I was thinking this wasn't working at all for me, but apparently with this, I can still wake via USB devices plugged into my LG UltraFine 5K. Since I plug the Magic Keyboard into it, it still wakes in one key. Since I plug my iPhone into it, that wakes as well (which is what I was hoping wouldn't happen. I like to plug it in after putting the PC to sleep and before I go to sleep myself and when the phone gets power it wakes the PC.
...

When you connect USB devices to the Thunderbolt-based LG UltraFine 5K, those USB devices are most likely managed by the Titan Ridge USB controller instead of the Intel Chipset USB controller.

The Designare Z390 has two USB controllers:

• XHC -- Intel chipset-based controller
• XHC2 (on RP05) -- Thunderbolt USB controller

The procedure to disable wake-from-USB affects XHC instead of XHC2.