Intel i5-8400 using UHD 630 Graphics

[HackinMax]

@CaseySJ @jaymonkey

HDMI Seems VERY happy at Index 2 0x01 |18 HDMI 0x000003C7

Apparently, DP likes Index 3 0x06 |18 DP 0x000003C7

Based on your suggestions, I have modified the config using Hackintool's patch creator (REALLY nifty tool BTW. Thanks @headkaze!!! )... one-by-one tracking down the DP that works in coexistence with HDMI. I have been using the data from our previous versions to pin down some good data... This is what I have come up with:

Look familiar? I know you've seen it in your head. It ain't pretty on startup. But, it works!

Yes! It is not perfect. But, it IS the closest we have come so far to a perfect pindown on this board. It still 'feels' incomplete somehow. But, I don't know anything more than you've taught me so far.

Startup is still very shakey and flashing green blocks. This occurs until I complete login (sometimes I still have to login blind) and, make at least one action with the pointer. I usually just do a right click (two finger tap). Once I do this, the flashing stops. And, so far has not reappeared.
I can get picture on both monitors. And, both report @60Hz. They both reported 60Hz for a few reboot cycles. But, without ANY other changes whatsoever, DP is BACK to 30Hz and, won't come back! Frustrating.

This is the extent of my abilities at this point. Where do we go from here?

Having gotten this far, I am very leery of dashing forward until I have more information.


Thanks for all of the help thus far!

 

[CaseySJ]

@HackinMax This does seem much better. There is still something strange in the IOReg dump:

• Your av-signal-type is 0x10 for both HDMI and DP ports. This is the signal type for DisplayPort.
• IOReg confirms that 2 displays are active and they are on Ports 6 and 7 (Index 2 and 3).

On my system, I just checked the following:

• AMD RX 580 feeding HDMI to 4K monitor: av-signal-type = 0x08 and connector type = 0x0008.
• Intel UHD 630 feeding HDMI to 1080p monitor: av-signal-type = 0x08 and connector type = 0x0008.
• Intel UHD 630 feeding DP to 4K monitor: av-signal-type = 0x10 and connector type = 0x0004.

What happens if you use the same BusIDs, but change HDMI to DP at Index 2?

 

[HackinMax]

@HackinMax This does seem much better. There is still something strange in the IOReg dump:

• Your av-signal-type is 0x10 for both HDMI and DP ports. This is the signal type for DisplayPort.
• IOReg confirms that 2 displays are active and they are on Ports 6 and 7 (Index 2 and 3).

On my system, I just checked the following:

• AMD RX 580 feeding HDMI to 4K monitor: av-signal-type = 0x08 and connector type = 0x0008.
• Intel UHD 630 feeding HDMI to 1080p monitor: av-signal-type = 0x08 and connector type = 0x0008.
• Intel UHD 630 feeding DP to 4K monitor: av-signal-type = 0x10 and connector type = 0x0004.

What happens if you use the same BusIDs, but change HDMI to DP at Index 2?

@CaseySJ

On boot, HDMI monitor returns to flashing wildly until login is complete. And, DP has a green horizontal line which moves from bottom to top and then repeats the trip several times. DP remains at 30Hz. But, nothing else has visually changed.

I kinda wish we had someone with the i5 MacMini8,1 who could run IOReg and just tell us all the answers.

Anyway, I will leave things as they are and let it simmer overnight. I enjoyed what was for me, a major success today with getting all of my USB ports patched properly. So, there's that.

If I boot with only DP connected. It is a beautiful, smooth process. If I did not absolutely need both ports, I would just leave it this way and move on.

If I boot with only HDMI it is a mess. glitchy all over until login is complete and, some screen action is performed.

 

[HackinMax]

Update/Summary:

Still using the config profiled in post #91 above.

- Boot goes black after IGPU initialization. (HDMI & DP)
- Blind login brings stable video to HDMI and DP after long delay.
- DP runs only at 30Hz, unless I switch it to HD 1080p rez.

I know these are just the remaining 'symptoms' of a port/signal issue.
No idea how to proceed troubleshooting on video ports.

Updated to WeG and Lilu latest (beta) per @toleda due to audio layout testing. When audio finally lit up, video flashing on HDMI completely stopped has been MUCH decreased. Coincidence? Maybe. But, I'll take it!

At this point, these three video problems are the only outstanding issue with this build.

Any path forward?

 

[HackinMax]

I went about cleaning up unneeded stuff today. I checked each-and-every kext file and driver to make sure there was nothing left over from previous, unsuccessful experiments.

I only found a few. But, one of them is a complete mystery. I had a RTL8111 kext in my L/E/
for the life of me, I cannot remember putting it there. (This board has only intel GBe. So, Realtek was never a....) Anyway... I removed it and a couple of other things I don't need.

It was merely to produce a best-version of the current EFI to move to my Unibeast Stick.

Wasn't expecting anything. But, the system booted normally - no black screen, no flashing. But, there is about a 15 second delay before it lights up.

I returned the DP monitor to 4K rez. And, all heck broke loose. Visual disturbances abounded. And. Then. Stopped.
Checked SysProf and DP had claimed 4K@60Hz!!

Screen shot from SystemProfiler attached.
IOReg attached

Just in case this is for real.

EDIT: Three reboots now. And, it seems to be working. Can an orphan driver really cause so much trouble?
We will see as I continue testing.

@CaseySJ ,Would you mind taking a look at the dump and, see if that signal type discrepancy has been resolved? OR anything else we need to troubleshoot?

Thanks SO much. I could not have done this without your help. If it stays working like this, then there is nothing on this system not working at this time.
Another dozen or so reboots will tell the tale. But, I am hoping to be able to call this build a [SUCCESS].

EDIT: Several cold boots and sixteen warm restarts later. I still have to login blind. But, the video performance after that is still hanging in there.

EDIT: After MANY different 'stress' tests (changing rez/rebooting mostly) I broke it. DP went back to 30Hz and, it took a while to figure out a way to get it back. Somehow, I 'tricked' the system into setting DP @60Hz. But, it doesn't survive reboot anymore.
Long story short... It's still broken.

 

[CaseySJ]

Sure, will take a look shortly...

 

[CaseySJ]

The av-signal-type is still 0x10 for both HDMI and DP, but I think this might actually be okay because you have HDMI 2.0 driven by the custom controller. Because both ports support 4K60, it might make sense that av-signal-type is the same for both. So for now I would not worry about that...

I understand that the current Framebuffer configuration is as follows:

Index 1 (whatever it has now)
Index 2, BusID 0x01, Type HDMI
Index 3, BusID 0x06, Type DP
Index -1, BusID 0x00, Type DUMMY

I would suggest trying out a final set of 2 options:

Index 1, BusID 0x00, Type DUMMY
Index 2, BusID 0x01, Type HDMI
Index 3, BusID 0x06, Type DP
Index -1, BusID 0x00, Type DUMMY

and

Index -1, BusID 0x00, Type DUMMY
Index 2, BusID 0x01, Type HDMI
Index 3, BusID 0x06, Type DP
Index -1, BusID 0x00, Type DUMMY

This can also be rewritten as follows:

Index 2, BusID 0x01, Type HDMI
Index 3, BusID 0x06, Type DP
Index -1, BusID 0x00, Type DUMMY
Index -1, BusID 0x00, Type DUMMY

In Hackintool, you could also try setting FB Port Limit to 2 in the Advanced page of the Patch tab:


Because you're driving 4K displays through the iGPU, you should also include the -cdfon option in Boot arguments. This option is described in Jaymonkey's guide (use browser's Find or Search function to look for "cdfon"):

https://www.tonymacx86.com/threads/an-idiots-guide-to-lilu-and-its-plug-ins.260063/

 

[HackinMax]

The av-signal-type is still 0x10 for both HDMI and DP, but I think this might actually be okay because you have HDMI 2.0 driven by the custom controller. Because both ports support 4K60, it might make sense that av-signal-type is the same for both. So for now I would not worry about that...

I understand that the current Framebuffer configuration is as follows:

Index 1 (whatever it has now)
Index 2, BusID 0x01, Type HDMI
Index 3, BusID 0x06, Type DP
Index -1, BusID 0x00, Type DUMMY

I would suggest trying out a final set of 2 options:

Index 1, BusID 0x00, Type DUMMY
Index 2, BusID 0x01, Type HDMI
Index 3, BusID 0x06, Type DP
Index -1, BusID 0x00, Type DUMMY

and

Index -1, BusID 0x00, Type DUMMY
Index 2, BusID 0x01, Type HDMI
Index 3, BusID 0x06, Type DP
Index -1, BusID 0x00, Type DUMMY

This can also be rewritten as follows:

Index 2, BusID 0x01, Type HDMI
Index 3, BusID 0x06, Type DP
Index -1, BusID 0x00, Type DUMMY
Index -1, BusID 0x00, Type DUMMY

In Hackintool, you could also try setting FB Port Limit to 2 in the Advanced page of the Patch tab:

View attachment 381197
Because you're driving 4K displays through the iGPU, you should also include the -cdfon option in Boot arguments. This option is described in Jaymonkey's guide (use browser's Find or Search function to look for "cdfon"):

https://www.tonymacx86.com/threads/an-idiots-guide-to-lilu-and-its-plug-ins.260063/

Thank you for the follow-up!

I will get to this as soon as I get back home tonight.

 

[HackinMax]

@CaseySJ ;

Honestly, these attempts made things go back to the bad-ole days of....well, just bad.

In every case, the system simply ignored the values in the patch code. Hackintool reported completely different data than what was created in the patch code.
Neither case recognizes DP. And, hotplugging DP causes reboot.

The FB ports = 2 did not seem to improve anything.

Both of us have done everything humanly possible at this point. Unless there is a super-human out there, I think it may be time to accept defeat.

But, if anyone lurking out there has the answers, it would be a GREAT time to share.

 

[CaseySJ]

@CaseySJ ;

Honestly, these attempts made things go back to the bad-ole days of....well, just bad.

In every case, the system simply ignored the values in the patch code. Hackintool reported completely different data than what was created in the patch code.
Neither case recognizes DP. And, hotplugging DP causes reboot.

The FB ports = 2 did not seem to improve anything.

Both of us have done everything humanly possible at this point. Unless there is a super-human out there, I think it may be time to accept defeat.

But, if anyone lurking out there has the answers, it would be a GREAT time to share.

Agreed. Best to stay with the settings that (tolerably) work until a more reasoned approach can be found to solving this problem as opposed to a continuous stream of trials-and-errors. At least we attempted every reasonable solution that has worked elsewhere.