kernel_task is using 100% CPU when waking from sleep

Hello.

I tried John_Raz method but as several posts here confimr it does not work on late 2013 rmbp. My problem is that when I tried to roll back it didnt get back to normal. My sond is gone and the volume keys does not work.

I'm afraid I did something wrong when trying to roll it back to Mavericks HDA.

Can anyone please supply me with the AppleHDA.kext for a late 2013 rmpb running 10.9.1?

Thank you in andvance.

I had problems too. My fix was to reinstall Mavericks. I think that is your best option.

I also tried it a week or so back and found that the earlier AppleHDA.kext did not work on my late 2013 MacBook Pro. To get it working again with the original kext file, I reversed the instructions to replace the original kext file and then rebooted.

Be sure to go to System preferences > Sound and reselect the sound input and output to get the sound working again. That may be a step you're missing to get your sound back.

I went to local dealers for Apple products and I managed to get myself a few AppleHDA kexts from various 2012 and 2013 MBP 13" and 15" running either Mountain Lion or Mavericks.

The Kext Wizard from John_Raz's file helps a lot. thanks. Just make sure you make a backup to get back. The backup is placed on the desktop. When things go south, just reinstall the from backup and reboot.

To summerize my experiment, kexts from Mountain Lion does not even make sound regardless of the origin. Flash videos do not even load with these ML kexts when I visit youtube.com.

Kexts from Mavericks from older Macbook pros do load and work. But same kernel_task 100% issue ensues when media is streaming and computer is put to sleep and woken up.

I also managed to get the kext from late 13" MBPr running Mavericks 10.9 (not 10.9.1, a rarety in my neighborhood) but it exhibited the same kernel_task issue in the store so I didn't bother to experiment with it.

Here's something interesting about iTunes I noticed. When a song is being played in iTunes and computer is put to sleep, I hear very faint "static" sound in my headphones as if the power on the audio device is going out.

When the MBP is woken up, the song never resumes playing (including iTunes radio).

I never hear that faint sound as the MBP sleeps when something's streaming in any other software/browser. They also always automatically resume playing whatever was playing at the time of sleep.

Whatever trick iTunes is using, I'd like to see that implemented regardless whether it's playing something or not when the computer goes to sleep.

I tried Marc's method as well on a late 2013 rMBP with apple earpods plugged in and it worked. A note to those not savvy with what to do with these command lines.

1. Command + Space to open spotlight

2. Start typing, "Terminal" without quotes and launch Terminal

3. copy and paste "sudo kextunload /System/Library/Extensions/AppleHDA.kext" into Terminal and hit return. If you have a password it will ask you to enter it to make sure you actually want to run it. When you apply it, you won't get a confirmation message, this is normal.

4. Now copy and paste "sudo kextload /System/Library/Extensions/AppleHDA.kext" into Terminal. If you have your apple earpods on you should hear a click and pop and if you start playing music it should come back on.

I have tried John Baz's method as there have been at least a couple of people who said it didn't work for them on a Late 2013 rMBP.

I realize this is a temporary solution, but until Apple fixes this I'm not sure there is a permenant fix for Late 2013 rMBP owners. Could someone make a script to execute this fix from Marc?

Adrius,

It seems that kext unloading/loading only works if program or browser tab that was used for playing/streaming is closed and audio is not actively held by a "process" or some sort.

If audio is still being used/held and if appleHDA kext is forced to unload, it spews out errors ungracefully -- and at least for me, kernel_task still runs high at 100%.

If I need to remember to close my browser tabs and streaming softwares before I put my MBP to sleep, and make note of exactly where I was in the clip or audio that I was playing, and reopen the clip/audio/video/watever again after wakeup and go to that point of the media just seem like so much work.

I left a user feedback to Apple as suggested by caljufi few posts up. I did see that there is a software that can run a script when a computer goes to sleep/wakeup but since I don't want to quit playing whatever I am playing at the time of putting it to sleep, automatic kext unloading/reloading may not be what I'm looking for... but I think it's something to ponder about.

Weird experience today...

I have a windows PC that mounts as a SMB share in finder, and also a USB-drive that is connected to my router. Before I went to sleep last night, I actually ejected those in finder before closing the lid. When I got home from work today I opened the MBPr and expected the kernel to be sky high since my headphones are plugged in too.

Guess what happend... no kernel freakout at all. Maybe that would be the fix for at least me? Eject shared folders in the network before letting the mac fall asleep. 🙂

Has this worked for anyone? Is this safe...? I'm new to Mac, and I dont want to do something thats unsafe. Im on a 2013 MacBook Air 13" base.

Just chipping in here... Found my MacBook Air 2013 with the same symptoms (kernel_task using 100% CPU). I temporarily fixed it by connecting and disconnecting my BlueTooth headphone (a nameless brand which identifies itself as SX-940).

It crops up every now and again.

I had the same problem and it goes aways only if I remove my headphone (w/ mic) from the audio out/in.

Was wondering if upgrading to Maverick will solve that problem I face every week on a MBP running SL!

Looks like the problem exists at least since 2009 (http://reviews.cnet.com/8301-13727_7-10330458-263.html#!)....

I miss the days when Apple was doing QA 😀

Uh, this is so annoying. I already posted 2 bug reports to Apple!

Apart from all the tips with unloading kernel extensions and similar, mine "solution" to get the kernel_process back to normal is simple: I quit all the applications that were playing audio before I put my Mac to sleep. That is usually iTunes, VLC, Spotify or Chrome (YouTube). After several seconds, the kernel_process drops down to normal. (Note: I usually don't have the audio jack plugged in.)

Does anyone know if this problem is fixed in OS X 10.9.2? Is it possible to find bugreport 15046435?

I've seen a few reports on some other forums that at least on the MacBook Air it is fixed in 10.9.2. I'd have to guess the problems with the Pro Retina is fixed also. Hope this helps!

unplug the earphones may help sometimes