[help]black screen when uhd630 run with internal screen

[headkaze]

For me Lilu, Whatevergreen and AppleALC load correctly only if installed if L/E. When they are in Clover Kext/Others they don't seem to load (for example Audio is never working in this case and i can't see these kexts in SysInfos>Extensions. I tried to add -lilubetaall as boot arguments for Mojave support, and Clover>Kext Inject=YES without change. So i installed all my kexts in L/E (Except the BRCMPatchRam kexts in S/L/E as expected)

Really you should be placing most of your kext's in EFI/CLOVER/kexts/Other. The only ones I have in L/E are BrcmFirmwareData.kext and BrcmPatchRAM2.kext. Placing kexts in S/L/E or L/E is considered obsolete now. That's not what I say; that's what the authors of Lilu, WhateverGreen and AppleALC say. They also say that now that Clover supports injecting via config.plist/Devices/Properties you're supposed to stop using ACPI and use device properties similarly to Apple.

I don't see the flaky sort of behavior you're seeing either. I've never had a boot up where I don't get the 3 minutes of black screen. That's with SetIntelMaxBacklight disabled and no AppleBacklightInjector.kext. I recommend you download my latest config and try it. It was made for me personally by vit9696. I don't think you're going to find a better config than that!

Also when injecting kexts via Clover you say the kexts aren't getting loaded? You won't see them in System Report...->Extensions. Try running "kextstat | grep -y -E 'Lilu|WhateverGreen|AppleALC'". Also you may want to try rebooting a few times.

At the end of the day though you can take whatever advice you like. For me personally since working with vit9696 on the WhateverGreen framebuffer patching system and witnessing first hand his reverse engineering skills. I can say without a doubt the man knows what he's talking about. So that's why I have so much respect for his advice but you can take it or leave it.

Still no way to have AptioMemoryFix properly working. As before i cannot shutdown correctly.

I will have a talk to vit9696 about it as he's involved in its development.

 

[headkaze]

I am facing this boot to black screen issue as well, though using a white wallpaper I can see the login screen using flashlight, and able to login.

It's certainly starting to sound more like a backlight issue to me now. So if that's the case it could just be related to the bug in Clover. I'll see if I can find out why it hasn't been addressed yet.

 

[headkaze]

For brightness control, see guide

We've all tried to follow your guide and unfortunately it does not fix the problem.

 

[neocoma504]

At the end of the day though you can take whatever advice you like. For me personally since working with vit9696 on the WhateverGreen framebuffer patching system and witnessing first hand his reverse engineering skills. I can say without a doubt the man knows what he's talking about. So that's why I have so much respect for his advice but you can take it or leave it.

Hi!

Don't get me wrong, i totally understand the theory behind this and i'm sure you and vit9696 have far more skills than me in term of software. I was just saying that right now, with my system (Mojave Beta 8), i had more luck and less issues with kext installed in L-E. I tried before to put in Kext/Other and i tried again this morning. Now this is working for audio and indeed i can see the reference with kextstat, didn't now that. But i have some boots issues with "kextd stall" sometimes.

Anyway, i tried again to remove Whatevergreen just to see and so far i had a few nice boots this morning without black screen at all! Maybe there is a mechanism in Whatevergreen to recall backlight, that may create the same kind of conflict with the CFL framebuffer as IntelSetMaxBacklight Clover Option. I don't know, this is far from my skills. Maybe you could ask vit9696 what he thinks about that?

An other thing, about sleep issues : I realised the problem came from USB. When something is connected on a non-internal port, it avoids going to sleep. So i changed the USB3 port of the SD card reader to Internal. And i changed a port where i always have a dongle plugged to Internal as well. Now my laptop goes to sleep as expected, with macOS menu or by closing the lid. And it wakes up correctly after too. The only thing is that this will only work if nothing is plugged on external ports. Which is not a big issue for me. Or maybe i could set all ports to internal but i don't know if this would not create other issue.

 

[headkaze]

But i have some boots issues with "kextd stall" sometimes.

Perhaps it's a BIOS setting you changed? Did you set "Fast Boot" to false by any chance? Changing that setting gave me issues.

Maybe it's some setting in your config.plist causing the issue? Did you give mine a try? You will notice my config.plist is very minimal. No junk so less chance of problems.

Anyway, i tried again to remove Whatevergreen just to see and so far i had a few nice boots this morning without black screen at all! Maybe there is a mechanism in Whatevergreen to recall backlight, that may create the same kind of conflict with the CFL framebuffer as IntelSetMaxBacklight Clover Option. I don't know, this is far from my skills. Maybe you could ask vit9696 what he thinks about that?

I have posted a ticket about the SetIntelMaxBacklight bug in Clover. So as soon as we can get this fixed I'll talk to vit9696 about a possible issue with WhateverGreen. Why I hesitate about it is that it's not consistent which leads me to believe it's an issue with your config. When you don't get a black screen did you make sure WhateverGreen wasn't loaded, did you check you have backlight control, full hardware acceleration etc.

An other thing, about sleep issues : I realised the problem came from USB. When something is connected on a non-internal port, it avoids going to sleep. So i changed the USB3 port of the SD card reader to Internal. And i changed a port where i always have a dongle plugged to Internal as well. Now my laptop goes to sleep as expected, with macOS menu or by closing the lid. And it wakes up correctly after too. The only thing is that this will only work if nothing is plugged on external ports. Which is not a big issue for me. Or maybe i could set all ports to internal but i don't know if this would not create other issue.

I'm just curious are you using SSDT-UIAC.aml or USBPower.kext for USB patching?

 

[RehabMan]

Hi everyone,
I have a laptop Acer Nitro 515-52
i5-8300H
HM370
UHD630 + Nvidia 1050Ti
Samsung 860 Evo 250GB SSD
HighSierra 10.13.6 (17G2208)

I was following this thread and would like to share my experience.

I am facing this boot to black screen issue as well, though using a white wallpaper I can see the login screen using flashlight, and able to login.
Even after logging in, it takes a while, more than 5 minutes. Then the brightness gets back to the last set value.

To enable brightness:
1) I tried Rehabman's guide, using ssdt_pnlf, AppleBacklightInjector.kext and the clover patch.
It works but the max brightness is very low, I know that I need to check the framebuffer for PWMMax value but I am unable to get this value.

2) I removed the AppleBacklightInjector kext and disabled the patch, but kept the SSDT_PNLF in the patched folder, with this I get full brightness due to the AppleBacklight kext.

3) Then I tried by removing the SSDT_PNLF as well and ticked Add_PNLF in clover, this also allows me to use the brightness with full max value.

4) Using the SetIntelBacklight as mentioned by headkaze, I was unable to get brightness control working.

Brightness value persists across reboots, the only problem being 5+ minutes of very low backlight, which seems like black screen.

I tried with mojave beta 4 as well but the black screen problem still persists.

I have attached the whatevergreen igfxdump as well. I am unable to get the PWMmax value from it.

You're disabling SSDT-PNLF.aml:

36:144  0:000  === [ ACPIPatchedAML ] ====================================
36:144  0:000  AutoMerge pass 1
36:144  0:000  Unsorted
36:144  0:000  Disabled: SSDT-PNLF.aml, skip

To discover the correct LMAX, you will need to dump the framebuffer ig-platform data by using WhateverGreen.kext.

Once the LMAX associated with the common ig-platform-id is determined, we can add direct support into SSDT-PNLF.dsl.

 

[headkaze]

To discover the correct LMAX, you will need to dump the framebuffer ig-platform data by using WhateverGreen.kext.

Backlight data is not available in a Coffee Lake dump.

 

[RehabMan]

Backlight data is not available in a Coffee Lake dump.

Bug in WhateverGreen.kext? Or data just not in (runtime generated) ig-platform table?
I have no CFL hardware here, so can't really help discover what is going on...

If there are still enough unknowns with CFL kext, maybe should be spoofing KabyLake or Skylake...

Note: You could get clues regarding PWMMax by looking at ioreg from real Mac running CFL kext, as you can see which brightness profile is used by AppleBacklight.

 

[headkaze]

Bug in WhateverGreen.kext? Or data just not in (runtime generated) ig-platform table?

Not a bug, just the data is not available in the table.

I have no CFL hardware here, so can't really help discover what is going on...

You can use Intel FB-Patcher and the Framebuffer->macOS 10.13.6 (17G2112) menu to view the table extracted from that version. When the full release of Mojave comes out I will be including binary dumps of that version in there too. That way you won't need the hardware to view it.

If there are still enough unknowns with CFL kext, maybe should be spoofing KabyLake or Skylake...

We used to have to unset the CNConnectorAlwaysConnected just to get it to work with KBL which meant no backlight support so we need CFL framebuffer.

Note: You could get clues regarding PWMMax by looking at ioreg from real Mac running CFL kext, as you can see which brightness profile is used by AppleBacklight.

I'm pretty sure 0x56c is the value but either way it's good enough for now. The problem is the 3 to 5 minutes of black screen. Then the screen goes fully bright again. So it's not the max level it's the delay.

 

[RehabMan]

You can use Intel FB-Patcher and the Framebuffer->macOS 10.13.6 (17G2112) menu to view the table extracted from that version. When the full release of Mojave comes out I will be including binary dumps of that version in there too. That way you won't need the hardware to view it.

I'll probably look into it when Mojave is released.
But my understanding was that the table is built at runtime, therefore difficult to dump it without actually executing the kext init code...

I'm pretty sure 0x56c is the value but either way it's good enough for now. The problem is the 3 to 5 minutes of black screen. Then the screen goes fully bright again. So it's not the max level it's the delay.

So far, I have yet to see a CFL setup (PR files) that represent correct brightness implementation assuming 0x56c.