Memory corrupted after longer sleep

[brousseau6933]

Report:

Gone to Yosemite 10.10.3 with an SMBIOS of iMac 15,1 (i7 4790k). Ran again the ssdtPRGen.sh script along with the Toleda lpc ssdt.

24 hours without a KP but it is too soon to declare the bug resolved. Time will tell.

Side effect:

With iMac 14,1 SMBIOS, the CPU ran cooler and lowered at 0,8 gHz. Now, it won't go lower than 1,64 gHz and stays a lot in the 3,8 -4,0 gHz after waking up. I think that it is related to the fact that OS X expect a real 4790k (4 gHz, 4,4 gHz Turbo) and behave accordingly.

 

[oddjobav8r]

After several days, I finally had a kp during deep sleep with 10.10.3. I guess I will shut it down at night until we get this figured out.

 

[brousseau6933]

After several days, I finally had a kp during deep sleep with 10.10.3. I guess I will shut it down at night until we get this figured out.

Same for me yesterday after three days without KP.

This one involved IOUSBMassStorage.kext I think.

How many of us have external USB storage devices? Maybe there's a cause lying there!

 

[derMatze]

I don't have any external devices...
But KPs on about every second or third wake up.

 

[Mariolillo]

I haven't experienced any more KP in deep sleep, 10.10.3. I switched my keyboard from a USB3.0 port to a USB2.0 port. I only have a webcam connected in one USB3.0 port, all others are empty. I didn't reverted the wastez23 when I did it so I will revert it now and test. If a KP happens after the change then it must be a combination of changing the value wastez23 suggested and which devices you have connected to which USB ports.

 

[Mariolillo]

I haven't experienced any more KP in deep sleep, 10.10.3. I switched my keyboard from a USB3.0 port to a USB2.0 port. I only have a webcam connected in one USB3.0 port, all others are empty. I didn't reverted the wastez23 when I did it so I will revert it now and test. If a KP happens after the change then it must be a combination of changing the value wastez23 suggested and which devices you have connected to which USB ports.

After trying again, I can confirm that removing IOPlatformSystemSleepPolicy allows the PC to keep sleeping without crashing while doing darkwakes.

 

[0utrunner]

After trying again, I can confirm that removing IOPlatformSystemSleepPolicy allows the PC to keep sleeping without crashing while doing darkwakes.

what does this mean - is that a fix? if so - how does one implement it

thanks

 

[Mariolillo]

what does this mean - is that a fix? if so - how does one implement it

thanks

It is a "fix", as in, we don't know what those values are used for in the system. All we know so far is that removing them solves this deep sleep problem.

To implement it, you need the board-id you are currently using (determined by the model id you are using in your SMBIOS).

To check your board-id:

ioreg -l | grep "board-id" | awk -F\" '{print $4}'

Next, go to the path :

/System/Library/Extensions/IOPlatformPluginFamily.kext/Contents/PlugIns/X86PlatformPlugin.kext/Contents/Resources

and find the plist that matches your hardware. Make a backup copy of it and then edit the plist file, removing the IOPlatformSystemSleepPolicy section. Rebuild your kernel cache and test.

 

[brousseau6933]

The trick of IOSystemPolicy didn't worked for me (nor any other tricks BTW!).

I gave up and set sleep to never and KP's are gone.

It looks like even some Windows users are plagued with Blue Screen Of Death related to Haswell c-sates scheme.

 

[g3power]

I can confirm that swapping the memory modules for Kingston ones fixes this issue for me. Haven’t had a panic in weeks. Neither on 10.10.2 nor 10.10.3. It was expensive, but worth it IMHO.

Using both the Kingston Kit and the Crucial Kit (for a total of 32 GB) exhibits the problem again.

Thanks pista7!