[Guide] Dell XPS 9560 Mojave VirtualSMC, I2C Trackpad, Clover UEFI Hotpatch

[williamctstaton]

Do you have an i5 7300HQ or i7 7700HQ, might want to fix your profile? Check using latest Intel Power Gadget instead. Try creating a different user and logging in to see if it's software related. Check your battery actual capacity, I'm at about 7500mAh and I can go a good 8 hours of light workload.

Attach debug files if you want someone to look into it further.

Sorry for CPU confusion, I must have mistyped it in. Ive attached my debug information, a screenshot of my power information, and a screenshot of the Intel Power Gadget at idle. As for programs consuming power, nothing is consuming any large amount of power since this install is less than 24hrs old. My batteries condition is 65% of what it used to be according to coconut, however, when I am booted into Arch Linux (with the 1050 manually disabled) I am averaging an hour or more easily with a slightly higher workload (Visual Studio Code, TMUX, Brave Browser, Youtube.)

On a sidenote, thank you very much for this hack and all the effort you have put into it. It has been the better part of these last few days for me!

 

[mazakiss]

Hey @blazinsmokey hope you're great!

Congrats! Usb-c hotpluging works awesome!

Yes, I had to remove both Tb and USB-C SSDT, do more than a couple of NVRAM clearing, then after re-adding them back it worked!

I will report for HDMI 4K @60Hz on Monday when I get back to work (the screen is there).

If anybody has a TB device to test, you are more than welcome!

Cheers!

 

[mazakiss]

Arff....

1) Cold plug (before boot) Hotplug (after) works great!

2) 2nd hotplug and replug is sometimes unpredictable:

- 1st attempt crashed the hack then rebooted automatically on the Dell splash screen.

Here, I think it's because NVRAM hadn't cleared correctly (despite doing many times F11 at clover screen, and also done the remove both SSDT /add KNN / add back + F11 again as you described in post #301.

- After re-clearing NVRAM twice, retried, then worked without crashing. Great!

Also I noticed that if you replug the USB-c/Tb adapter, wether plugged before/after boot, usb devices only reconnect and mount if it wasn't plugged to the adapter prior being replugged to laptop.

Also unplug/replug is very random, while System Info sees the peripheral connected, it is not mounting all the time, sometimes you have a delay before any usb device mounts, other time it doesn't even mount at all. weird.

So we'll see how it goes over time ;-)

Another weird stuff compared to your hack (my config is nearly the same as yours 32Go RAM, no fingerprint sensor, Kabylake 7700HQ) is that in System Configuration shows a different picture (not about our respective adapter specs (see pic attached - marked in red are the USB 3.0 dongle attached to the USB-C/TB Adapter recognized as PCI).
It show 2 PCI devices VS only one on hack

Last, I also tried your config-kb.plist, the boot glitch only disappears if you add lilucpu=8 boot arg to config-kb.plist. ( you added the boot arg the on regular config.plist, not the kabylake one).

The regular config.plist with lilucpu=8 boot arg still displays the glitch (not important, it's not a fashion here lol)

Last, using config-kb.plist is super slow at boot everytime (didn't figured why yet) so I use the regular one now.

All in all thank you for all your hard work @blazinsmokey. Awesome progress!

Cheers.

So only have to test HDMI tomorrow.

Thanks!

 

[InsaneCultist]

Thanks for the update all, I will test with my USB-C dock tomorrow morning, the 2048MB of RAM on the new iGPU device injection is working for me. Will have to investigate the Type C improvements tomorrow and will report back.

EDIT:
Alright, after swapping my TB1 SSDT for the two provided in the 3.2 package, I can safely say that my initial testing confirms hot plug working on my Dell D6000 USB C dock. I started with the dock plugged in from a cold boot and then was able to hot plug repeatedly, showing the same changes in IOReg that are seen in blazinsmokey's posts. I then unplugged, rebooted and then hot plugged once the system was up and the dock was detected again, as expected.

I don't own any Thunderbolt devices, so can't test there. Only down side here is the D6000 doesn't seem to handle display output very well under macOS. HDMI to HDMI only (no adapters allowed) and only one display, but that's an issue for another thread. Thanks for your hard work, blazinsmokey!

EDIT 2:
If I sleep my laptop with the dock plugged in, I have to hotplug it after wake for it to pass power, but data, network, other functionality doesn't work initially after sleep. After unplugging and re-plugging a few times or letting the laptop fully wake up and trying again, then the dock works just fine. This seems infinitely more consistent than the previous implementation. I believe I have a USB C to Ethernet adapter somewhere that I can also test today.

EDIT 3:
I had a kernel panic while my laptop was sleeping and the dock was plugged in, issue with Darwake, but upon further review of the config.plist, I had darwake set to 1 instead of "no" as intended. Will report back if issue persists after this change.

 

[mazakiss]

I tested again today. Hotplug works great.

But replug is what causes me the most issues.

It works very randomly. I had the issues you described with NVRAM.

Did like 10 times the procedure explained in post #301, but still cannot manage to run it stable.

It seems like hot replug works but with a big delay (in minutes) i.e if you unplug usb-c/tb adapter (with USB device directly attached to it in this example), it won't mount back when replugged. But if you wait a couple a minutes then plug it back, it works and usb devices mounts again. Weird.

Twice when I closed the lid and wake from sleep the laptop cold rebooted or had a black screen I had to manually shut the laptop down.

I also did last Bios upgrade 1.14.2 (release date April 25). Didn't solved the issue. Checked bios settings (same setup as described as KNNSpeed). Also updated all my drivers on my windows 10 partition (I'm on dual boot, same drive).

My controller is almost the same as you, same brand, maybe same controller, second on the right here : http://www.wbpinetech.com/index.php?route=product/category&path=20_66

 

[mazakiss]

In the pics attached here, you see what "System Info" is showing once then controller is disconnected from laptop.
Usb device was ejected but is still present on the device while not physically attached to it anymore. Weird.

The second PCI (the controller) disappears which is normal, but the USB 3.1 device still sticks for a while as if it was still detected by the system.

Finally after waiting a few minutes, that USB 3.1 port completely disconnects from system.

After that I can reconnect my controller and plug back all usb devices, it will mount again.

I don't know what creates this delay between each replugs.

 

[mazakiss]

The sleep/ cold reboot issue when lid is closed seems to happen randomly, rarely. Happened twice so far.

 

[InsaneCultist]

The sleep/ cold reboot issue when lid is closed seems to happen randomly, rarely. Happened twice so far.

I am seeing the same as your two posts. I am also getting the cold reboot issue when the lid is closed. Last night, I experienced a KP related to IONVMeFamily - do you by chance also have a PM961 in your XPS? I've seem other reports of instability (despite Sierra running without issue.) blazinsmokey reported a fix HERE using THIS package - is this included in the latest 3.2 zip?

 

[mazakiss]

Last thing, I now it was already an issue, but I also noticed every usb devices connected before sleep won't be properly ejected and won't mount after wake.

It that latter case, while hard disconnected, the USB 3.1 devices are still present in System Info while not mounted.

Cheers.

 

[mazakiss]

I am seeing the same as your two posts. I am also getting the cold reboot issue when the lid is closed. Last night, I experienced a KP related to IONVMeFamily - do you by chance also have a PM961 in your XPS? I've seem other reports of instability (despite Sierra running without issue.) blazinsmokey reported a fix HERE using THIS package - is this included in the latest 3.2 zip?

Hi @InsaneCultist how are you?

No I have the Samsung EVO970, not 4K which now works by default since Mojave (I don't use prior Mac OS versions on this hack).