Contribute
Register

How to build your own iMac Pro [Successful Build/Extended Guide]

Status
Not open for further replies.
Your PWM fans are controlled by the bios. Normally, osx has no access to control them. On real macs, the os controls the fans via the SMC. Try removing the motherboard sensors kext, as that kext gives control access via fakesmc. Without that, there should be no way whatsoever for osx to control your fans.

That's what I did.

Current uptime: 1d 13h, the maximum I've ever had (never passed 24h uptime before). I removed ACPISensors.kext two days ago and it seems to have resolved my random freezes/KP... But I also switched from APFS to HFS+ at the same time + some DSDT changes... So I'm not sure about what appears to have resolved my freezes either ACPISensors.kext removal or the other changes, that's what I need others who've been having the same KP tracedump to help me figure out. For now I'm still stuck trying to achieve at least 2 days of uptime to really confirm that my ramdom freezes issue is gone before testing sleep/wake with these modifications... I don't want to restart or try sleep to avoid reset my uptime counter. :(

Waiting for @AnaktuvGod to provide me with the list of kexts he has in Clover so we can cross-check that.

In the meantime if someone wants to try removing ACPISensors.kext from their kexts folder and check if E3/KP sleep/wake issue is gone with that modification that'd help! :)
 
Last edited:
Yeah, I don't think @kgp's fans are related to the sleep problem, sorry. I don't use any sensors and I still have e3 sleep problem, but no KP.

Ok thanks for confirming this point. So maybe E3 is different from KP. Waiting for @AnaktuvGod to provide his list of kexts :(.

Such a time consuming issue to debunk.
 
Yeah, I don't think @kgp's fans are related to the sleep problem, sorry. I don't use any sensors and I still have e3 sleep problem, but no KP.

Using xcpm_core_scope_msrs © Pike R you should neither face Error E3 nor KP.
 
That patch disables a big part of xcpm_init, where the kernel writes a couple of MSRs, based on the detected processor. One of these writes lock our system, thus the E3 error. Interestingly, this code also runs on boot, and we CAN boot without this patch. Anyway, we don't know which MSR is causing the issue. Pike had a kernel debugger that stopped working a couple of os versions ago, but he is apparently not fixing it, so I have no idea how to debug this further.

Your PWM fans are controlled by the bios. Normally, osx has no access to control them. On real macs, the os controls the fans via the SMC. Try removing the motherboard sensors kext, as that kext gives control access via fakesmc. Without that, there should be no way whatsoever for osx to control your fans.

Hope this helps.

I currently don't use any sensor kexts..

Screen Shot 2018-01-30 at 12.14.08.png


and still I have the problem with my PWR fans after wake from sleep. Sleep/Wake else works flawless and my system is rock solid!
 
How do you use the links? You should copy all curl commands and drop them into a terminal. Is this what you are doing?
Thanks. With terminal it works. Before I copied the urls into the browser. Donwloads every file but the plist. But in terminal it worked :)
 
Ok thanks for confirming this point. So maybe E3 is different from KP. Waiting for @AnaktuvGod to provide his list of kexts :(.

Such a time consuming issue to debunk.

@Thireus I'm not using any additional kext outside of what @kgp provided in the EFI folder. Came home to a panic with the same acpi error.

*Note. I noticed as well once disabled the xcpm_core_scope_msrs © Pike R patch, it prevented my system from sleeping. the display would turn off but not the system. I can tell you now that since enabling, the computer actually responds to sleep again and wakes no problem. Runs fine like when I first was successful with my initial install.

Only thing I changed I believe yesterday was a new lilu kext I seen was released recently which I don't believe has anything to do with this issue. Just though I let you know of all kext changes outside of what @kgp provided
 
@Thireus I'm not using any additional kext outside of what @kgp provided in the EFI folder. Came home to a panic with the same acpi error.

Ok, so this is not the issue. Do you mind making a backup of the EFI folder you know results in KP? I also have a backup of the previous one that was causing KPs, but there are many changes I've made since then. If it happens that I've resolved the issue (time will tell, no KP so far) we will have to crosscheck our EFI folders.
 
@Thireus I'm not using any additional kext outside of what @kgp provided in the EFI folder. Came home to a panic with the same acpi error.

*Note. I noticed as well once disabled the xcpm_core_scope_msrs © Pike R patch, it prevented my system from sleeping. the display would turn off but not the system. I can tell you now that since enabling, the computer actually responds to sleep again and wakes no problem. Runs fine like when I first was successful with my initial install.

Only thing I changed I believe yesterday was a new lilu kext I seen was released recently which I don't believe has anything to do with this issue. Just though I let you know of all kext changes outside of what @kgp provided

@AnaktuvGod, seems that you are addressing now a different issue and I think I already have the solution.

You are running an Nvidia GPU, right?

In all my previous EFI-Folder distributions, I checked the "nv_spanmodepolicy=1" boot flag in Clover Configurator. The latest release has this boot flag unchecked, as I was made ridiculous by somebody to use this boot flag meant for 5K Displays. :)

However, as expected from my side, it seems absolutely mandatory to check this boot flag in order to avoid a black screen with Nvidia GPUs on Wake.

If your system is accessible via remote access and the mainboard display shows 30, the issue is definitely your GPU!

Can you please try to check "nv_spanmodepolicy=1" in Section "Boot" of Clover Configurator and see if this solves your new issue.

I am convinced that will work.

I just checked it myself.

Using boot flag "nv_spanmodepolicy=1", everything works as expected.

Not using boot flag "nv_spanmodepolicy=1", black screen on Wake with Nvidia..
 
Last edited:
Ok, so this is not the issue. Do you mind making a backup of the EFI folder you know results in KP? I also have a backup of the previous one that was causing KPs, but there are many changes I've made since then. If it happens that I've resolved the issue (time will tell, no KP so far) we will have to crosscheck our EFI folders.

I literally use the same folder @kgp provides us with. All I’ve done was open the config.plist and enable/disable kernel PM and Xcpm core scope msrs. But if you really want the one I’m using alright give me a minute. I’m actually in the bed and have to work 2 more 12 hour shifts this week. I’m responding from my phone. But give me a moment
 
I literally use the same folder @kgp provides us with. All I’ve done was open the config.plist and enable/disable kernel PM and Xcpm core scope msrs. But if you really want the one I’m using alright give me a minute. I’m actually in the bed and have to work 2 more 12 hour shifts this week. I’m responding from my phone. But give me a moment

See post #4644.... >>I told you to wait a moment ;):lol:
 
Status
Not open for further replies.
Back
Top