Jump to content

AMD High Sierra Kernel Release and Testing


1,119 posts in this topic

Recommended Posts

Yes it was update from 10.13.5. So I did fresh install from your installer from amd-osx forum and updated following your video tutorial. and everything is great.

Thank you so very much!

Would be awesome if you could make installer for 10.13.6 too.

Link to comment
Share on other sites

Hi Shaneee, tried the combo updater and your video tutorial (thanks!) on my backup drive. 

Unfortunately the kextcache command gives me 2 errors:

-----------

Can't use AppleBusPowerController.kext - not linked.
Prelink failed for com.apple.iokit.IOUSBHostFamily; omitting from prelinked kernel.
Can't use IOUSBHostFamily.kext - not linked.
Prelink failed for com.apple.driver.KernelRelayHost; aborting prelink.
Failed to generate prelinked kernel.
Child process /usr/sbin/kextcache[1299] exited with status 71.

-----------

 

System is working for some minutes, but then reboots randomly.

Link to comment
Share on other sites

25 minutes ago, AntarcticP said:

Hi Shaneee, tried the combo updater and your video tutorial (thanks!) on my backup drive. 

Unfortunately the kextcache command gives me 2 errors:

-----------

Can't use AppleBusPowerController.kext - not linked.
Prelink failed for com.apple.iokit.IOUSBHostFamily; omitting from prelinked kernel.
Can't use IOUSBHostFamily.kext - not linked.
Prelink failed for com.apple.driver.KernelRelayHost; aborting prelink.
Failed to generate prelinked kernel.
Child process /usr/sbin/kextcache[1299] exited with status 71.

-----------

 

System is working for some minutes, but then reboots randomly.

 

Did you use the System kext and updated kernel? If so try adding these kexts, repair permissions and run kextcache again,

 

 

Kexts.zip

  • Like 1
Link to comment
Share on other sites

Updated from 10.13.3 to 10.13.6 following Shaneee‘s tutorial, worked very well. Thanks to all of you involved in building the new kernel! No issues with building prelinkedkernel after rolling back to Shaneee‘s IONetworkingFamily.kext. I do not use any modified System.kext. System is running on latest Clover with VirtualSMC. I also did not use Shaneee‘s rp file but repaired permissions manually (didn‘t find sources for rp and don‘t like unknown binarys working on my system files).

Edited by SPO-Fan
  • Like 1
Link to comment
Share on other sites

Now tried with the 3 kext you provided. getting this:

------------

Can't use AppleBusPowerController.kext - not linked.

Prelink failed for com.apple.iokit.IOUSBHostFamily; omitting from prelinked kernel.

Can't use IOUSBHostFamily.kext - not linked.

Prelink failed for com.apple.driver.KernelRelayHost; aborting prelink.

Failed to generate prelinked kernel.

Child process /usr/sbin/kextcache[681] exited with status 71.

--------------

System does boot, typing this from 10.13.6, as it did yesterday. Will inform you if it crashes again.

Link to comment
Share on other sites

On 10/3/2018 at 9:20 PM, AlGrey said:

here is kernel for 10.13.6. 

Backup your IONetworkingFamily.kext from 10.13.3 and restore it once you updated to 10.13.6 otherwise prelinkedkernel won't build.

kernel_10.13.6_minimal+opemu.zip

Im noticing jagged video playback with this kernel. Using VLC player. Attached is sysctrl output text. 

sysctrl -a | grep hw.txt

Link to comment
Share on other sites

7 hours ago, Shaneee said:

Just to update everyone. We are currently trying to work out why Nvidia performance is worse when used with an AMD CPU. We have a few willing devs helping out now.

 

Something I’ve noticed is that AMD cards in the 7xxx series run benchmarks the same as they would on an intel system but higher end RX cards have a 10-15 % drop versus on an intel system.    

Cinebench is a bit odd my RX 460 scores the same as my RX 480 and Vega 64.   My R7 265 is about equal with those cards in the 80s.   With Intel they break 100.

 

Geekbench on the other hand the R7 265 is about equal to the RX 460 but the RX 480 and Vega are about 10-15% lower than intel systems.   Could it be a PCIE lane width issue?   It seems like the high end cards are running on 8 lanes instead of 16 but they list as 16 lanes in system report.   

Nvidia cards suffer a severe loss in comparison but maybe there is something that is holding both cards back.

  • Like 1
Link to comment
Share on other sites

System running now stable for over an hour on 10.13.6 on my AMD FX 8350. Seems that it had nothing to do with the kexts in my previous post.

I have updated Clover to the latest version (4700) and deleted a busratio=20 flag that I was using in 10.13.3.

That did the trick.

 

thanks a lot Shaneee for this kernel!!!

  • Like 2
Link to comment
Share on other sites

On 10/25/2018 at 3:57 AM, Gigamaxx said:

Just curious, has anyone tried iMac Pro definition?   I get a constant kernel panic with this.   I have tested a few times with my RX Vega to see if there is a performance benefit but no luck.   

 

Yeah. I had to modify the AppleGraphicsDevicePolicy.kext to stop no signal being displayed on my monitor but no KP. Can you share an image of the panic?

Link to comment
Share on other sites

Hello,

 

I have updated to 10.13.6 using the disk clone method from Shaneee. However since the update (not sure if prior to the update it was the case, but almost certainly not). I have noticed that the usb audio devices skips every so often at interval of anywhere between 30-40 seconds. I have done a little bit of digging around and I am now almost certain that this is a kernel related issue. I am using a ryzen 1800x, I have the bus ratio correctly set to 36, I am not using voodoo, I have cleaned up my other/kexts to the bare minimum and don't have any extra kexts in S/L/E.

 

I then started looking at the logs in the Console app and I have noticed the exact same error every time the audio skips.

 

default    14:17:59.176707 -0400    kernel    IOAudioEngineUserClient[<private>] - AUDIO OFFLINE. online=1. state=0. loopCount=0, lastLoopTime=0
error    14:17:59.176721 -0400    coreaudiod    HALS_IOA1Engine.cpp:365:EndWriting:  HALS_IOA1Engine::EndWriting: got an error from the kernel trap, Error: 0xE00002D7
default    14:17:59.176757 -0400    com.apple.WebKit.WebContent    HALC_ProxyIOContext.cpp:1068:IOWorkLoop:  HALC_ProxyIOContext::IOWorkLoop: skipping cycle due to overload
default    14:17:59.179454 -0400    kernel    IOAudioEngineUserClient[<private>] - AUDIO OFFLINE. online=1. state=0. loopCount=0, lastLoopTime=0   

 

and it repeats from there. So it starts with the kernel reporting AUDIO OFFLINE, then coreaudiod crashing and then the next line is whatever application that was playing the audio reporting HALC_ProxyIOContext::IOWorkLoop: skipping cycle due to overload.

 

Also this only seems to occur with usb audio devices. I have tested bluetooth headphones and also sound from the line out port and those don't seem to have the same problem.

 

Link to comment
Share on other sites

6 hours ago, fk2106 said:

Hello,

 

I have updated to 10.13.6 using the disk clone method from Shaneee. However since the update (not sure if prior to the update it was the case, but almost certainly not). I have noticed that the usb audio devices skips every so often at interval of anywhere between 30-40 seconds. I have done a little bit of digging around and I am now almost certain that this is a kernel related issue. I am using a ryzen 1800x, I have the bus ratio correctly set to 36, I am not using voodoo, I have cleaned up my other/kexts to the bare minimum and don't have any extra kexts in S/L/E.

 

I then started looking at the logs in the Console app and I have noticed the exact same error every time the audio skips.

 

default    14:17:59.176707 -0400    kernel    IOAudioEngineUserClient[<private>] - AUDIO OFFLINE. online=1. state=0. loopCount=0, lastLoopTime=0
error    14:17:59.176721 -0400    coreaudiod    HALS_IOA1Engine.cpp:365:EndWriting:  HALS_IOA1Engine::EndWriting: got an error from the kernel trap, Error: 0xE00002D7
default    14:17:59.176757 -0400    com.apple.WebKit.WebContent    HALC_ProxyIOContext.cpp:1068:IOWorkLoop:  HALC_ProxyIOContext::IOWorkLoop: skipping cycle due to overload
default    14:17:59.179454 -0400    kernel    IOAudioEngineUserClient[<private>] - AUDIO OFFLINE. online=1. state=0. loopCount=0, lastLoopTime=0   

 

and it repeats from there. So it starts with the kernel reporting AUDIO OFFLINE, then coreaudiod crashing and then the next line is whatever application that was playing the audio reporting HALC_ProxyIOContext::IOWorkLoop: skipping cycle due to overload.

 

Also this only seems to occur with usb audio devices. I have tested bluetooth headphones and also sound from the line out port and those don't seem to have the same problem.

 

 

The new kernels have fixed the need to set busratio=, have you tested without?

Link to comment
Share on other sites

I can remove the busratio argument and see what happens. And yes I had noticed that busratio was affecting it, because when I was overclocking the bus ratio wasn't matching but all the audio sources would still work the same.

 

Thank you

Link to comment
Share on other sites

Hello I have updated to 10.13.6 using the disk clone method from Shaneee. The system boots up fine and works for a few minutes but it often freezes and reboots. I am currently on 10.13.3. That is working fine.

I am using the latest clover without any busratio flag. 

What could be the issue? I also tryed to add the kexts you included to AntarcticP but no luck.

 

EDIT:
And I am also unable to restart or shutdown my PC without manually pushing the Power Button on my PC. I attached the code that I get when I try to restart/shutdown(I dont know where I can find this code but if you tell me I can send you that)

 

My system specs:

Amd Ryzen 5 1600 (stock no OC)
GIGABYTE GA-A320M-S2H

12 GB Ram

AMD R9 280x 3 GB

IMG_3386.jpg

Edited by erik96joe
Link to comment
Share on other sites

 Share

×
×
  • Create New...