Jump to content

Wake up from sleep and USB issues (10.11.2)


blackswords
 Share

23 posts in this topic

Recommended Posts

Hello, 

 

since I updated from Yosemite to El Capitan I have two main issues that I didn't manage to fix. 

 

The first one is waking up from sleep. The computer enters into sleep mode just fine but when I try to wake it, the screen remains black then after a few seconds the system reboots by itself and I have a message saying that my computer has restarted because of a problem (kernel panic?). I tried different settings via pmset with no luck. When I still was on Yosemite, only hibernation was working but that was okay for me.

 

The second issue I have is with USB ports. Most of the time (but not always), USB ports get disabled when OS X boots. When it happens, I push the reset button and the problem disappears. It's not a huge issue because it just takes me an extra 30s to boot my computer but I would be great if it can be solved. No other USB issues apart this one, all my ports (USB2 & USB3) are working correctly. I think I did the unlocking thing after installing El Capitan but I can't remember the exact procedure. Yesterday I tried the 10.11.3 update but all USB ports stayed disabled no matter what soI just revert to 10.11.2 using CCC.

 

Here are my computer specs:

  • MB : MSI PC Mate Z97
  • CPU : Intel i5-4590k
  • GPU : AMD R9 280X
  • RAM : 8GB
  • SSD : 256GB (OS X boot + W10 boot)
  • HD : 1TB (storage)
  • Bootloader : Clover v3292

I attach a system log including a sleep (yesterday night) and a wake up (this morning) and my Clover config.plist.

 

Thanks for your help 

 

 

 

syslog_and_config.zip

Link to comment
Share on other sites

Okay. So kernel panic (on restart after wake) is caused by this:

29/02/2016 08:45:36,000 kernel[0] Unsynchronized  TSC for cpu 1: 0x00000000793d88cf, delta 0x562e03d6
29/02/2016 08:45:36,000 kernel[0] Unsynchronized  TSC for cpu 2: 0x000000007967ec7a, delta 0x562e03f1
29/02/2016 08:45:36,000 kernel[0] Unsynchronized  TSC for cpu 3: 0x000000007992334b, delta 0x562e03eb

Use VoodooTSCSync.kext to fix this (place it in your clover kext directory).

 

If you have AMD graphics you won't be able to wake your computer without setting IGPU to primary in your BIOS (which means a blind boot). Also you'll have to inject it in Clover with proper framebuffer (most likely 0x04120004).

 

About your USB issues, my advice is to go to 10.11.4, since it has a little bit better USB support.

  • Like 2
Link to comment
Share on other sites

So I tried what you suggested Hervé but it still doesn't work... Maybe I didn't get something wright so here is what I did:

  • Removed NullCPUPowerManagement from /Volumes/EFI/EFI/CLOVER/kexts/10.11
  • Ran ssdtPRGen.sh and copied /Users/Benjamin/Library/ssdtPRGen/ssdt.aml to /Volumes/EFI/EFI/CLOVER/ACPI/patched
  • Verified that KernelPm is checked in CloverConfigurator
  • Disabled hibernation using the commands you gave to me

I also changed my smbios to iMac15,1 since it's closer to my hardware than the one I used before (iMac14,1) with no luck.

 

The bright side is that speedstep is working great now, I can see my CPU going down to 800MHz while the load is low.

 

I attach the syslog concerning the sleep/wake failure as well as my current Clover config.plist, maybe you can spot something else.

 

Many thanks for your help

 

Edit : I just added the VoodooTSCSync kext to /Volumes/EFI/EFI/CLOVER/kexts/ but the "Unsynchronized TSC for cpu x" messages still appears (I set IOCPUNumber to 7 because my CPU has 4 cores, I hope it's the good value)

 

Edit2 : I can't find the VoodooTSCSync kext when running kextstat, is it normal?

syslog_and_config2.zip

Link to comment
Share on other sites

Well, I did what you suggested but the computer still reboots when waking up. The TSC sync errors are still in the syslog...

 

As for the DSDT file, I saw that the one I have is empty so I just removed it from my patched folder and from my config.plist.

 

I can see in the syslog the following line : 01/03/2016 11:39:07,000 kernel[0]: Sleep failure code 0x00000001 0x27006700. Maybe it's some information on what is going wrong?

 

Edit : speedstep doesn't seem to work anymore, the CPU is always clocked at more than 3GHz even with no activity (98% inactivity) 

syslog_and_config3.zip

Link to comment
Share on other sites

Well, I applied your patch to EHC1, EHC2 & XHC (HDEF already have this section and I don't have the two other devices) but the problem persists. 

 

By changing some parameters in my BIOS (USB related mainly) the computer doesn't reboot by itself when waking up but the screen stays black...

Link to comment
Share on other sites

  • 3 weeks later...

I'm having the same problem (I think). I recently changed my motherboard from a GA-Z87MX-D3H to an Asrock Z97M OC Formula, to try to solve a known problem with Gigabyte boards and memory corruption after (long) sleep. 

 

Any time I try to wake the system, it reboots and gives a kernel panic report: launchd exited (signal 8, exit status 0 ), and Thread # crashed. I too am seeing Unsynchronised TSC messages, and VoodooTSCSync didn't solve the problem for longer sleep (with power nap enabled). 

 

I have an i5-4670k @4,4GHz, and it has crossed my mind that this issue is due to that being a Haswell non-refresh CPU in combination with a 9 series board. 

 

My relevant settings in Clover are: KernelPm=TRUE, with other kernel patches disabled. I have FixRegions_10000000 and NewWay_80000000 enabled, but no other patches in Clover because I have a patched DSDT and SSDTs. I also have OsxAptioFix2Drv-64.efi, but I might not need it. 

Link to comment
Share on other sites

After a lot of tests I gave up on sleep, I just spent too much time for what it's worth. But if you somehow find a way to fix it please let me know.

 

I finally managed to install 10.11.3 by reinstalling clover completely and my usb issues seem to be gone now.

Link to comment
Share on other sites

Ah, that's too bad. Thanks for replying. I'll keep working on it for now. I've got VoodooTSCSync back for now, I'll see if all the other ACPI patches and config optimisations I've done will make it do it's job somehow. Last time I tried it, it did reboot after a few scheduled darkwakes. 

Link to comment
Share on other sites

Since this thread was started later, I'll continue here.

 

I did some more testing. I disabled both "Wake for Ethernet network access" and "Enable Power Nap" in the hopes that this would help. Now, without having set a darkwake boot flag, The wake reason when I click my bluetooth trackpad is "GLAN". And the display doesn't come on. Another click is needed. There are still TSC sync issues, but no crashed processes (VoodooTSCSync might be helping here). 

 

I would really like native sleep functionality (Power Nap and WoL/Wake on Demand), so this isn't solved for me. 

Link to comment
Share on other sites

Just a quick update. The logs show there are two different kinds of process crashes after wake. The first to show up are ones similar to these: 

2016-03-24 11:51:19,000 kernel[0]: Process[711] crashed: Image Capture Ex. Too many corpses being created.

Then come the launchd services with messages like: 

2016-03-24 11:51:19,977 com.apple.xpc.launchd[1]: (com.apple.dock.extra[393]) Service exited due to signal: Floating point exception: 8

Does anyone know wether these crashes are even related to TSC?

Link to comment
Share on other sites

While browsing the forums I saw people mentioning the "pmset -g log | grep -i failure" command to debug wake issues and when I run it I get :

2016-03-15 07:41:28 +0100 Failure              Drivers Failure panic during wake due to PEG0():

So it's definitely related to my graphics card (R9 280X). Do you guys have any idea to solve (or at least debug) this error?

 

By the way, I noticed that when my external sound card (Akai EIE Pro) is plugged the speedstep seems disabled (CPU always at maximum frequency) so when I said earlier in this thread that speedstep wasn't working anymore, it may be due to this effect. And since the update to 10.11.3 my USB issues upon boot are gone.

Link to comment
Share on other sites

While browsing the forums I saw people mentioning the "pmset -g log | grep -i failure" command to debug wake issues and when I run it I get :

2016-03-15 07:41:28 +0100 Failure              Drivers Failure panic during wake due to PEG0():

So it's definitely related to my graphics card (R9 280X). Do you guys have any idea to solve (or at least debug) this error?

This is probably related to wake problem with AMD Radeon graphics cards. See this thread. The only solution is to set integrated graphics as primary and enable CSM.

Link to comment
Share on other sites

Yes, the output of "pmset -g log | grep -i failure" for me is: 

2016-03-26 00:07:04 +0100 Failure             	Drivers Failure panic during wake due to HDAU(),SATA(AppleAHCI):           	          
2016-03-26 11:00:06 +0100 Failure             	Drivers Failure panic during wake due to HDAU(),SATA(AppleAHCI):           	          
2016-03-26 13:38:58 +0100 Failure             	Drivers Failure panic during wake due to PEG0(NVDA NVDA NVDA NVDA NVDA NVDA),IGPU():	          
2016-03-26 15:38:52 +0100 Failure             	Drivers Failure panic during wake due to HDAU(),SATA(AppleAHCI):           	          
2016-03-26 17:38:53 +0100 Failure             	Drivers Failure panic during wake due to PEG0(NVDA NVDA NVDA NVDA NVDA NVDA),IGPU():	          
2016-03-27 15:39:17 +0200 Failure             	Drivers Failure panic during wake due to HDAU(),SATA(AppleAHCI):           	          
2016-03-27 17:11:31 +0200 Failure             	Drivers Failure panic during wake due to PEG0(NVDA NVDA NVDA NVDA NVDA NVDA),IGPU():	          
2016-03-27 17:42:20 +0200 Failure             	Drivers Failure panic during wake due to PEG0(NVDA NVDA NVDA NVDA NVDA NVDA),IGPU(),HDEF():	          
2016-03-27 22:01:20 +0200 Failure             	Drivers Failure panic during wake due to PEG0(NVDA NVDA NVDA NVDA NVDA NVDA),IGPU(),HDEF(AppleHDAController):

This is with my GTX 970 set as initial display. But you'll notice SATA also shows up. My theory here is that the TSC sync issue in the processor is also affecting PCIe devices, and these are just a side-effect of the more central problem.

 

I've found threads where the solution was to modify the firmware of their board, or even update the microcode of their processor (for Haswell-E). I'm really hoping it doesn't come to that. 

 

Blackswords, could you let us know wether the settings konrad11901 suggests help in any way? 

Link to comment
Share on other sites

I tried but the computed didn't go to sleep, the image just froze and I was able to unfroze it by doing something with the mouse/keyboard. It may be due to the fact that the integrated GPU wasn't recognized by the system.

 

Anyway, this is not a got solution for me because I need to be able to select the OS I want to boot with inside clover menu and I can't do that during a blind boot.

Link to comment
Share on other sites

I tried but the computed didn't go to sleep, the image just froze and I was able to unfroze it by doing something with the mouse/keyboard. It may be due to the fact that the integrated GPU wasn't recognized by the system.

 

Anyway, this is not a got solution for me because I need to be able to select the OS I want to boot with inside clover menu and I can't do that during a blind boot.

For now it's the only solution. In my case (Radeon R9 270 & i5-4460 & MSI B85-G43 Gaming) I needed to set primary graphics to IGD and Windows 8/8.1 Feature to disabled. Also I set in Clover config Inject->Intel to true (leaving Inject->Ati=true) and ig-platform-id (or something like that) to a value from some guide and it's working.

 

PS: It's not related to restart after wake problem.

Link to comment
Share on other sites

 Share

×
×
  • Create New...