Contribute
Register

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

Status
Not open for further replies.
You could also try and disable CPUFriend with boot arguments "-cpufoff"
Keep in mind then the laptop won't idle down to 0.8Ghz and will do as what a MBP 14,1 would do which IIRC is 1.3GHz.
Mix and match the options above and let us know.

CPUFriend seems to be causing it. I disabled it and my laptop seems to idle around 11watts after sleep now. The problem is that my external mouse stops working and opening the lid doesn't wake the computer, I need to push the power button to get it to wake up.

That being said, I am still using the other CLOVER folder from this thread. Nothing against what you have done, but I get better battery life (it idles the CPU down to 800Mhz) witch translates to about 9.5w idle usage. I'm not sure what's going on to get it down to 800Mhz, but it works as shown in Intel Power Gadget.
 
CPUFriend seems to be causing it. I disabled it and my laptop seems to idle around 11watts after sleep now. The problem is that my external mouse stops working and opening the lid doesn't wake the computer, I need to push the power button to get it to wake up.

That being said, I am still using the other CLOVER folder from this thread. Nothing against what you have done, but I get better battery life (it idles the CPU down to 800Mhz) witch translates to about 9.5w idle usage. I'm not sure what's going on to get it down to 800Mhz, but it works as shown in Intel Power Gadget.

With my files it does idles down to 800Mhz on Intel Power Gadget. During my testing, rarely does it actually go down that low and stay there as macOS is pretty busy. I'm sure someone using my files can confirm it. As far as you using another guides files and asking why your having all these USB and lid wake issues you really gotta take it to that thread. I'm not here to offer troubleshooting and support from another set of files and guide.
 
With my files it does idles down to 800Mhz on Intel Power Gadget. During my testing, rarely does it actually go down that low and stay there as macOS is pretty busy. I'm sure someone using my files can confirm it. As far as you using another guides files and asking why your having all these USB and lid wake issues you really gotta take it to that thread. I'm not here to offer troubleshooting and support from another set of files and guide.

Sorry, I didn't mean to make it seem like I was looking for support on other guides. I was saying that if I remove CPUFriend or disable it from your config, then my 9650 becomes unstable and the computer doesn't wake from sleep correctly and I have trouble with the trackpad. It does resolve the high battery usage issue though. Sorry if I wasn't clear on my explanation.

As of right now, the issue I have is:

1. Wake from sleep results in higher (double wattage) battery usage.
 
Last edited:
Do any of you have issues with Combo Jack? It works like a charm until I put my dell into sleep. When it wakes up I do not have any audio (neither speaker or headphones), when I try to reconnect the headphones popup appears with a lag (usually 10-30 secs) and selecting any options does not have any effect. Is it a know issue? How can I help to debug it?
 
Do any of you have issues with Combo Jack? It works like a charm until I put my dell into sleep. When it wakes up I do not have any audio (neither speaker or headphones), when I try to reconnect the headphones popup appears with a lag (usually 10-30 secs) and selecting any options does not have any effect. Is it a know issue? How can I help to debug it?

Definitely do not have this problem. Just tested after sleep and speakers work as well as the jack.
 
@jmayniac
And to jmayniac, I think you helped find this issue with higher power usage after wake but all your other issues, lol. I bet it's from switching files back and forth, not clearing nvram and rebuilding kext caches and permissions. Sometimes you need to do both and completely shut down in between.

@nikoel Below will definitely be of interest to you too
 
EDIT!! Ok so without deep idle sleep works, lid open wakes but now after sleep it's idling at 16-17Watts. :banghead:

Clover 3.4 is released due to some findings about a botched CPUDataProvider.
  • New CPUFriendDataProvider.kexts, options two options available
    • Default Power Save
  • Removed SSDT-DeepIdle
    • Fixed sleep issues
    • Fixed opening lid didn't wake screen, had to press power button
  • TB/USB-C options are now in a separate folder, you decide which one to use, none is included by default
  • Latest Clover, Kexts, and AptioMemoryFix R27 as of today
If your just here for the files, it's attached to the first post.....

coconutBattery, all it really does it pull info from about System Info and pretties it up. I also have the coolest battery on earth apparently. While I'm not sure I would trust anything other then power discharge, it's consistent enough for me to realize there is a problem with power after sleep. I have been using my laptop docked for quite some time now so I didn't realize you poor souls running on battery are getting less than you should. Sorry!

I set the status bar to show me milliamp discharge and it's exactly the same as what's already provided to us using System Report Power section and it only updates when the system updates it. So you can refresh the System Info and it'll stay the same regardless of what's happening and once that changes coconutBattery changes as well.

coconutbattery.png

Problem was evident after testing. 9-12W during idle but after sleep it stayed pretty stuck in the 20s even after like 25 minutes.

So I went back and investigated the CPUFriendDataProvider. Learned a decent amount too. I generated two new ones using Power Saving method and a Balanced setting. Included default will be Power Saving.
These technically should only be used for i7 processors but give it a shot and report back to me with your findings.

Included are DataProviders with two frequency vectors and EPP adjustments. You have the option to select what you prefer. Both have been altered to change the SpeedShift’s EPP value and both have LFM(lowest frequency mode) set to 800Mhz. The standard is 1.3GHz for MacBook Pro 14,1. Both versions will also resume HWP after sleep.

Default installed is the Power Saving as there’s no ill effects after testing and I ran Geekbench on both with negligible differences in score. Max CPU speeds on all cores can still be reached.

Balanced will technically scale the performance on the higher end using higher clock speeds than Power savings. This I believe is the default for MBP. I did notice the cpu in Intel Power Gadget did not idle as close to 800Mhz as often as the default.
—————————————————————
If you care about the details… continue reading.

Balanced method is EPP 0x20 in decimal it’s 32. So think of that as EPP 32.

Power Saving is EPP 0xC0 which is EPP 192.

In short, EPP(energy performance preference) is the value at which Intel SpeedShift (HWP) will consumes/saves power. I said in short!

The range of values are supposed to be between 0-255. 00 being HIGH performance, max CPU speeds all the time. 255 the opposite.

You can read more about EPP and a starting point on what to change in our board id’s plist in SLE/IOPlatformPluginFamily.kext and merge it with the resource script that comes with CPUFriend.


Heres what it looks like now at idle before and after sleep.
Idle.png

And here is the reading right when it wakes from sleep.
Screen Shot 2019-07-03 at 1.36.50 AM.png
Like I said before I been using docked and hardly put it to sleep so I started retesting with the new CPUFriendProvider and the damn laptop would not sleep no matter what. Only Deep Idle. After a few testings I found SSDT-DeepIdle to be the culprit. I researched it up and KNN grabbed these patches from Pike and there's really not much info out there on it. We have fully working HWP so this bit of DeepIdle is not necessary(?). It's like the middle ground between idle and sleep. Display blacks out but the machine is technically still on. So figure it's better to have sleep then be stuck in the middle.

Would like you guys to report to see if you notice anything different without it. It is still included in a separate folder in Misc. Maybe some of you can test sleep as well.
 
Last edited:
For anybody that wants to try, i have OpenCore 0.0.3 working for Dell XPS 9560 to boot MacOS. I can't get BrcmpatchRAM2 to load for bluetooth to work. If anyone can look into that and help with that, it'll be a good thing. Any modifications or fixes is welcomed also.
 

Attachments

  • EFI.zip
    9.1 MB · Views: 94
EDIT!! Ok so without deep idle sleep works, lid open wakes but now after sleep it's idling at 16-17Watts. :banghead:

Clover 3.4 is released due to some findings about a botched CPUDataProvider.
  • New CPUFriendDataProvider.kexts, options two options available
    • Default Power Save
  • Removed SSDT-DeepIdle
    • Fixed sleep issues
    • Fixed opening lid didn't wake screen, had to press power button
  • TB/USB-C options are now in a separate folder, you decide which one to use, none is included by default
  • Latest Clover, Kexts, and AptioMemoryFix R27 as of today
If your just here for the files, it's attached to the first post.....

coconutBattery, all it really does it pull info from about System Info and pretties it up. I also have the coolest battery on earth apparently. While I'm not sure I would trust anything other then power discharge, it's consistent enough for me to realize there is a problem with power after sleep. I have been using my laptop docked for quite some time now so I didn't realize you poor souls running on battery are getting less than you should. Sorry!

I set the status bar to show me milliamp discharge and it's exactly the same as what's already provided to us using System Report Power section and it only updates when the system updates it. So you can refresh the System Info and it'll stay the same regardless of what's happening and once that changes coconutBattery changes as well.

View attachment 414689

Problem was evident after testing. 9-12W during idle but after sleep it stayed pretty stuck in the 20s even after like 25 minutes.

So I went back and investigated the CPUFriendDataProvider. Learned a decent amount too. I generated two new ones using Power Saving method and a Balanced setting. Included default will be Power Saving.
These technically should only be used for i7 processors but give it a shot and report back to me with your findings.

Included are DataProviders with two frequency vectors and EPP adjustments. You have the option to select what you prefer. Both have been altered to change the SpeedShift’s EPP value and both have LFM(lowest frequency mode) set to 800Mhz. The standard is 1.3GHz for MacBook Pro 14,1. Both versions will also resume HWP after sleep.

Default installed is the Power Saving as there’s no ill effects after testing and I ran Geekbench on both with negligible differences in score. Max CPU speeds on all cores can still be reached.

Balanced will technically scale the performance on the higher end using higher clock speeds than Power savings. This I believe is the default for MBP. I did notice the cpu in Intel Power Gadget did not idle as close to 800Mhz as often as the default.
—————————————————————
If you care about the details… continue reading.

Balanced method is EPP 0x20 in decimal it’s 32. So think of that as EPP 32.

Power Saving is EPP 0xC0 which is EPP 192.

In short, EPP(energy performance preference) is the value at which Intel SpeedShift (HWP) will consumes/saves power. I said in short!

The range of values are supposed to be between 0-255. 00 being HIGH performance, max CPU speeds all the time. 255 the opposite.

You can read more about EPP and a starting point on what to change in our board id’s plist in SLE/IOPlatformPluginFamily.kext and merge it with the resource script that comes with CPUFriend.


Heres what it looks like now at idle before and after sleep.
View attachment 414692

And here is the reading right when it wakes from sleep.
View attachment 414693
Like I said before I been using docked and hardly put it to sleep so I started retesting with the new CPUFriendProvider and the damn laptop would not sleep no matter what. Only Deep Idle. After a few testings I found SSDT-DeepIdle to be the culprit. I researched it up and KNN grabbed these patches from Pike and there's really not much info out there on it. We have fully working HWP so this bit of DeepIdle is not necessary(?). It's like the middle ground between idle and sleep. Display blacks out but the machine is technically still on. So figure it's better to have sleep then be stuck in the middle.

Would like you guys to report to see if you notice anything different without it. It is still included in a separate folder in Misc. Maybe some of you can test sleep as well.

Ugh... Confused. Maybe I didn't read this correctly. This is what i can gather. Just had a look at your first post but there were no additions. So... you've noticed the increased power consumption. Fixed it. Uploaded the solution. But then encountered an other problem and the consumption is back at 17-18 so you used strike-through and left the post as is. Correct?
 
Blazinsmokey
Instruction in OP about VoltageShift is incomplete
./voltageshift offset -125 -90 -125 vil only work after cold/reboot.But after sleep/wake these values return to -0.
You can check with this
./voltageshift info
To solve it you should use -125 -90 -125 0 0 0 1
the last digit "1" will scheduled check the settings every 1 min. and amend if needed. (after wake from sleep)
More details here:
 
Last edited:
Status
Not open for further replies.
Back
Top