Contribute
Register

10.8.4: no USB on wake up after sleep (Z77x-UD5H)

Status
Not open for further replies.
Joined
Apr 7, 2013
Messages
30
Motherboard
GB Z77X-UD5H
CPU
i7 3770S
Graphics
ATI Radeon RX570
Mac
  1. MacBook Pro
Classic Mac
  1. Apple
Mobile Phone
  1. iOS
tl;dr: Z77X-UD5H mobo - can't get USB to work reliably after sleep on OSX 10.8.4, unless I trick GenericUSBXHCI into leaving one USB device to AppleUSBEHCI to manage, and work with all others in -gux_nosleep mode (reset on wake up). Experiencing everything from devices not working post-sleep, all the way to KPs. Any help appreciated. Details below.


---

hi guys

anyone else had this? I updated a fully working 10.8.3 to 10.8.4 via combo updater. Rebooted, checked the machine works, slept it, then woke it up, some of USB devices refuse to come back until I disconnect/reconnect or am able to trigger re-enumeration of sorts.

Reinstalled FakeSMC and plugins, as well as the USB 3 extension, but it does not seem to help.
The various boot flags listed for GenericUSBXHCI do not make any difference.
I reinstalled the older version from Multibeast 5.3.1 but that didn't make any difference either.
Also I completely removed this extension and rebuilt the caches. I remember OSX waking up fine without this extension, on this machine, when it was 10.8.3. 10.8.4 seems to have broken this.

:/ Anyone any ideas?

Edit, from the logs, with GenericUSBXHCI, regardless of boot flags
Wake reason = ?
RestoreControllerStateFromSleep: xHC Restore Error

From the logs without GenericUSBXHCI
Aug 14 11:15:19 macx kernel[0]: Wake reason: ?
Aug 14 11:15:19 macx kernel[0]: RTC: PowerByCalendarDate setting ignored
Aug 14 11:15:19 --- last message repeated 1 time ---
Aug 14 11:15:19 macx kernel[0]: Previous Sleep Cause: 5
Aug 14 11:15:19 macx kernel[0]: The USB device 4-Port USB 3.0 Hub (Port 3 of Hub at 0x14800000) may have caused a wake by issuing a remote wakeup (3)
Aug 14 11:15:19 macx kernel[0]: The USB device 4-Port USB 3.0 Hub (Port 4 of Hub at 0x14800000) may have caused a wake by issuing a remote wakeup (3)
Aug 14 11:15:19 macx kernel[0]: [0xffffff801f6ae600](0)/(5) Device not responding


The second one is the behaviour I observed before (maybe without the "Device not responding" message, which I cannot track down)

Any help is appreciated.

Gigabyte Z77x-UD5H, i7-3770S, 16GB RAM, GeForce GTX 650 (vanilla), BIOS as per recommended install settings. Everything USB (EHCI/XHCI) in the BIOS is set to Auto.


Edit:
I'm seeing this with a D-link USB Hub, to which I have connected an Apple keyboard (itself a hub), and a Wacom tablet. On wake-up the tablet is dead, so is the keyboard. If I unplug the tablet and plug it back in, nothing changes. But if I unplug the keyboard, the tablet wakes up immediately. The hub is powered so I don't expect this to be power-related.
I'll do some more testing now..


Edit 2:
It does seem related to the fact that the Apple USB keyboard was plugged into the hub. This way it could also be power-related although in a roundabout way (OSX 10.8.3 had no issues with this but the USB was "jittery" for the first 30 seconds after wakeup).

I reinstalled GenericUSBXHCI, then moved the keyboard to another USB port and now the tablet wakes up correctly. System.log still reports "RestoreControllerStateFromSleep: xHC Restore Error" though. That brings us back to the kernel boot flags which previously didn't help.. Will report back.

Edit 3:
Various port combinations yield different results in terms of what works at boot and what works after sleep (usually different). Incredibly frustrating. I'll keep looking.

Edit 4 & 5:
Have plugged everything into the blue USB ports at the back of machine.

I can reliably wake up the keyboard and tablet only under very specific circumstances - I have to have something plugged into the top blue usb port at the back, and then the hub and keyboard plugged into the bottom blue usb ports at the back. Then xhcdump slots shows 9 working slots and working endpoints at each slot

It looks like 10.8.4 introduces a problem with high-powered devices after sleep. I'll report back if I find out anything else.

Edit 6:
Looks like I've reached a point where I can repeatedly make the system unstable after waking up:

Ports at back:
Top left blue USB slot: iLok (port presents itself as USB Hi-Speed Bus/USB 2.0)
Bottom left USB slot: Apple USB keyboard (port listed as USB 3.0)
Bottom right USB slot: USB Hub (port listed as USB 3.0, my hub listed as USB 2.0 hub)

Triggering an error: open System Profiler to USB tab, sleep, wake up, and press Cmd+R to refresh devices.

so consistently the first/top port is assigned to the USB 2.0 class of ports, however I have nothing plugged into any of the 2.0 ports. I wonder if the XHCI kext inadvertently shifts a USB 3 port device into the wrong category..

Outcome:
I'm seeing devices jump categories all the time upon plugging-unplugging, when booting with -gux_defer_usb2. I think there are some conflicts going on between this kext, apple's kext, and the ehci controller (auto) on motherboard. I don't think I can take this research past this point. Any help is appreciated.

Conclusion:
I've reached a compromise by having the system boot with -gux_nosleep and making sure there's something in the first blue usb port at the back. That device registers as USB Hi-speed one (which I can live with). The keyboard and hub register as USB 3.0 Bus devices (and get managed by usbxhci). I think the system, as it stands, is broken. I'll wait for a fix, in the meantime I can't even summarise a workaround :/

Meanwhile:
xhcdump does not show any changes in information before and after sleep. so far I've had best results when that kext manages everything, but that has also made the system unreliable. View attachment xhcdump_working.txt
 

Attachments

  • xhcdump_pre_sleep.txt
    5.8 KB · Views: 291
tl;dr: Z77X-UD5H mobo - can't get USB to work reliably after sleep on OSX 10.8.4, unless I trick GenericUSBXHCI into leaving one USB device to AppleUSBEHCI to manage, and work with all others in -gux_nosleep mode (reset on wake up). Experiencing everything from devices not working post-sleep, all the way to KPs. Any help appreciated. Details below.


---

hi guys

anyone else had this? I updated a fully working 10.8.3 to 10.8.4 via combo updater. Rebooted, checked the machine works, slept it, then woke it up, some of USB devices refuse to come back until I disconnect/reconnect or am able to trigger re-enumeration of sorts.

Reinstalled FakeSMC and plugins, as well as the USB 3 extension, but it does not seem to help.
The various boot flags listed for GenericUSBXHCI do not make any difference.
I reinstalled the older version from Multibeast 5.3.1 but that didn't make any difference either.
Also I completely removed this extension and rebuilt the caches. I remember OSX waking up fine without this extension, on this machine, when it was 10.8.3. 10.8.4 seems to have broken this.

:/ Anyone any ideas?

Edit, from the logs, with GenericUSBXHCI, regardless of boot flags


From the logs without GenericUSBXHCI



The second one is the behaviour I observed before (maybe without the "Device not responding" message, which I cannot track down)

Any help is appreciated.

Gigabyte Z77x-UD5H, i7-3770S, 16GB RAM, GeForce GTX 650 (vanilla), BIOS as per recommended install settings. Everything USB (EHCI/XHCI) in the BIOS is set to Auto.


Edit:
I'm seeing this with a D-link USB Hub, to which I have connected an Apple keyboard (itself a hub), and a Wacom tablet. On wake-up the tablet is dead, so is the keyboard. If I unplug the tablet and plug it back in, nothing changes. But if I unplug the keyboard, the tablet wakes up immediately. The hub is powered so I don't expect this to be power-related.
I'll do some more testing now..


Edit 2:
It does seem related to the fact that the Apple USB keyboard was plugged into the hub. This way it could also be power-related although in a roundabout way (OSX 10.8.3 had no issues with this but the USB was "jittery" for the first 30 seconds after wakeup).

I reinstalled GenericUSBXHCI, then moved the keyboard to another USB port and now the tablet wakes up correctly. System.log still reports "RestoreControllerStateFromSleep: xHC Restore Error" though. That brings us back to the kernel boot flags which previously didn't help.. Will report back.

Edit 3:
Various port combinations yield different results in terms of what works at boot and what works after sleep (usually different). Incredibly frustrating. I'll keep looking.

Edit 4 & 5:
Have plugged everything into the blue USB ports at the back of machine.

I can reliably wake up the keyboard and tablet only under very specific circumstances - I have to have something plugged into the top blue usb port at the back, and then the hub and keyboard plugged into the bottom blue usb ports at the back. Then xhcdump slots shows 9 working slots and working endpoints at each slot

It looks like 10.8.4 introduces a problem with high-powered devices after sleep. I'll report back if I find out anything else.

Edit 6:
Looks like I've reached a point where I can repeatedly make the system unstable after waking up:

Ports at back:
Top left blue USB slot: iLok (port presents itself as USB Hi-Speed Bus/USB 2.0)
Bottom left USB slot: Apple USB keyboard (port listed as USB 3.0)
Bottom right USB slot: USB Hub (port listed as USB 3.0, my hub listed as USB 2.0 hub)

Triggering an error: open System Profiler to USB tab, sleep, wake up, and press Cmd+R to refresh devices.

so consistently the first/top port is assigned to the USB 2.0 class of ports, however I have nothing plugged into any of the 2.0 ports. I wonder if the XHCI kext inadvertently shifts a USB 3 port device into the wrong category..

Outcome:
I'm seeing devices jump categories all the time upon plugging-unplugging, when booting with -gux_defer_usb2. I think there are some conflicts going on between this kext, apple's kext, and the ehci controller (auto) on motherboard. I don't think I can take this research past this point. Any help is appreciated.

Conclusion:
I've reached a compromise by having the system boot with -gux_nosleep and making sure there's something in the first blue usb port at the back. That device registers as USB Hi-speed one (which I can live with). The keyboard and hub register as USB 3.0 Bus devices (and get managed by usbxhci). I think the system, as it stands, is broken. I'll wait for a fix, in the meantime I can't even summarise a workaround :/

Meanwhile:
xhcdump does not show any changes in information before and after sleep. so far I've had best results when that kext manages everything, but that has also made the system unreliable. View attachment 64310

have you tried disabling cpu pll overvoltage in the bios? fixed all my USB wake from sleep problems.
 
have you tried disabling cpu pll overvoltage in the bios? fixed all my USB wake from sleep problems.

wow, what's the reasoning?
 
Status
Not open for further replies.
Back
Top