USB Driver Changed Again - PB6

[RehabMan]

@Rehabman and Matt

As I mentioned earlier I've tried all the steps here as well and am having similar issues. However with both the Xosi patch and the DSDT patch on post 208 (without Xosi patch) Matt just completed sleep is working fine until I try to sleep with a usb 3 device plugged into a usb 3 port. If I try to sleep then I have a similar behaviour to what Matt is experiencing. It sleeps fine with usb 2 devices plugged in. Thought this info may help.

Rick

It will be interesting to see the result of the ACPIDebug output.

 

[trboteg]

RehabMan,
Would you mind looking over my IOReg file(s) and seeing if anything stands out to you?

I see my USB 2.0 and USB 3.0 hubs in the Device Tree, however, I seem to have a similar issue as the others on this forum with handoff between XHCI and EHCI, as my USB 2.0 devices do not work in USB 3.0 ports.

The first IOReg file (Brandon's iMac) shows my system with two USB 2.0 devices plugged in and usable (one device utilizing one USB 2.0 port on the internal header, one device utilizing one USB 2.0 port on the back panel), and one USB 3.0 device utilizing one USB 3.0 port on the back panel and usable.

Also, when plugging a USB 2.0 device into a USB 3.0 port with a USB 3.0 device already plugged into a USB 3.0 port, the USB 3.0 device is ejected, both the USB 2.0 device and USB 3.0 device are shown in the Device Tree, yet I cannot access either device, nor are they listed in Disk Utility. The second IOReg file (Brandon's iMac - after) shows this scenario. Removing the devices does not restore original functionality; any device, whether USB 2.0 or USB 3.0, will show in the Device Tree, but is not accessible.

Additionally, the USB layout on my GA-Z68MX-UD2H-B3 Rev. 1.3 is as such:
- 4x USB 2.0 (rear)
- 2x USB 2.0 (header) - one port does not currently work, unsure of reason
- 2x USB 3.0 (rear)

 

[RehabMan]

RehabMan,
Would you mind looking over my IOReg file(s) and seeing if anything stands out to you?

I see my USB 2.0 and USB 3.0 hubs in the Device Tree, however, I seem to have a similar issue as the others on this forum with handoff between XHCI and EHCI, as my USB 2.0 devices do not work in USB 3.0 ports.

The first IOReg file (Brandon's iMac) shows my system with two USB 2.0 devices plugged in and usable (one device utilizing one USB 2.0 port on the internal header, one device utilizing one USB 2.0 port on the back panel), and one USB 3.0 device utilizing one USB 3.0 port on the back panel and usable.

Also, when plugging a USB 2.0 device into a USB 3.0 port with a USB 3.0 device already plugged into a USB 3.0 port, the USB 3.0 device is ejected, both the USB 2.0 device and USB 3.0 device are shown in the Device Tree, yet I cannot access either device, nor are they listed in Disk Utility. The second IOReg file (Brandon's iMac - after) shows this scenario. Removing the devices does not restore original functionality; any device, whether USB 2.0 or USB 3.0, will show in the Device Tree, but is not accessible.

Additionally, the USB layout on my GA-Z68MX-UD2H-B3 Rev. 1.3 is as such:
- 4x USB 2.0 (rear)
- 2x USB 2.0 (header) - one port does not currently work, unsure of reason
- 2x USB 3.0 (rear)

The kind of non-Intel USB3 (I think 1b6f:7023 is Etron) you have... I'm not familiar with...

No (unique) ACPI representation for it, so you might try adding a device for it in DSDT and then creating an injector kext for the ports.

 

[matt77303]

ok,

installed the DSDT, sent the hack to sleep, it decided to wake its self up again straight away.

Last login: Wed Sep 16 13:53:50 on ttys000
Matts-iMac:~ mattgarrad$ grep -y acpidebug /var/log/system.log
Sep 16 13:50:19 localhost kernel[0]: ACPIDebug: Version 0.1.3 starting
Sep 16 13:55:38 Matts-iMac kernel[0]: ACPIDebug: { "^^LPCB.XUSB/XRST", 0x0, 0x1, }
Sep 16 13:55:38 Matts-iMac kernel[0]: ACPIDebug: { "pr2/pr2m/pr3/pr3m", 0x0, 0x3fff, 0x3f, 0x3f, }
Sep 16 13:55:38 Matts-iMac kernel[0]: ACPIDebug: "calling XSEL from XWAK"
Sep 16 13:55:38 Matts-iMac kernel[0]: ACPIDebug: { "pr2/pr2m/pr3/pr3m", 0x3fff, 0x3fff, 0x3f, 0x3f, }
Matts-iMac:~ mattgarrad$

Sep 16 13:55:38 Matts-iMac kernel[0] <Notice>: Wake reason: EHC2

Matt

 

[RehabMan]

ok,

installed the DSDT, sent the hack to sleep, it decided to wake its self up again straight away.

Last login: Wed Sep 16 13:53:50 on ttys000
Matts-iMac:~ mattgarrad$ grep -y acpidebug /var/log/system.log
Sep 16 13:50:19 localhost kernel[0]: ACPIDebug: Version 0.1.3 starting
Sep 16 13:55:38 Matts-iMac kernel[0]: ACPIDebug: { "^^LPCB.XUSB/XRST", 0x0, 0x1, }
Sep 16 13:55:38 Matts-iMac kernel[0]: ACPIDebug: { "pr2/pr2m/pr3/pr3m", 0x0, 0x3fff, 0x3f, 0x3f, }
Sep 16 13:55:38 Matts-iMac kernel[0]: ACPIDebug: "calling XSEL from XWAK"
Sep 16 13:55:38 Matts-iMac kernel[0]: ACPIDebug: { "pr2/pr2m/pr3/pr3m", 0x3fff, 0x3fff, 0x3f, 0x3f, }
Matts-iMac:~ mattgarrad$

Sep 16 13:55:38 Matts-iMac kernel[0] <Notice>: Wake reason: EHC2

Matt

Disable "wake on USB" in BIOS for testing... Or try not having devices plugged into EHC2 that might cause wake.

Note that XWAK is setting PR2 to 0x3fff, disabling XHC USB2 routing to EHC. You are using XHCIMux in this scenario (no ioreg, so can't check myself), right? If so this would be bad... XHCIMux can't stop this write, and it will not be the same as pre-sleep.

XHCIMux can block writes to PR2 by the system (eg. by Apple's USB drivers), but not when it is done via ACPI.

Try the attached DSDT. I want to see what pr2/pr2m/pr3/pr3m was before and after XSEL from _OSC.

 

[matt77303]

Do the same? Set it to sleep etc?

Something I thought was interesting. When I connect to it via screen sharing. I wakes the display up.
Matt

 

[joransrb]

I have GM installed and everything works, almost... I got a strange USB issue (and yes, i have read several places that the usb in el cap has been rewritten, but i cant find anyone else with this issue, i have also been following this thread, and others...) All usb ports works (haven't tried the usb 3 ports as i don't use them), but, there are random "dropouts" of the usb... never had this issue before. The usb drops out for about a second, and comes back, and its at random intervals, maybe 1 - 2 times pr hour...

anyone got a suggestion on where to start investigating this issue...?

 

[RehabMan]

Do the same? Set it to sleep etc?

Something I thought was interesting. When I connect to it via screen sharing. I wakes the display up.
Matt

Yes.

 

[RehabMan]

I have GM installed and everything works, almost... I got a strange USB issue (and yes, i have read several places that the usb in el cap has been rewritten, but i cant find anyone else with this issue, i have also been following this thread, and others...) All usb ports works (haven't tried the usb 3 ports as i don't use them), but, there are random "dropouts" of the usb... never had this issue before. The usb drops out for about a second, and comes back, and its at random intervals, maybe 1 - 2 times pr hour...

anyone got a suggestion on where to start investigating this issue...?

Check ioreg so you know where the devices are attaching (which controller...)

Check system.log for USB related logs.

 

[matt77303]

Yes.

Sep 16 13:50:19 localhost kernel[0]: ACPIDebug: Version 0.1.3 starting
Sep 16 13:55:38 Matts-iMac kernel[0]: ACPIDebug: { "^^LPCB.XUSB/XRST", 0x0, 0x1, }
Sep 16 13:55:38 Matts-iMac kernel[0]: ACPIDebug: { "pr2/pr2m/pr3/pr3m", 0x0, 0x3fff, 0x3f, 0x3f, }
Sep 16 13:55:38 Matts-iMac kernel[0]: ACPIDebug: "calling XSEL from XWAK"
Sep 16 13:55:38 Matts-iMac kernel[0]: ACPIDebug: { "pr2/pr2m/pr3/pr3m", 0x3fff, 0x3fff, 0x3f, 0x3f, }
Sep 16 14:08:44 localhost kernel[0]: ACPIDebug: Version 0.1.3 starting
Sep 16 14:25:34 localhost kernel[0]: ACPIDebug: Version 0.1.3 starting
Sep 16 14:26:39 Matts-Mac-Pro kernel[0]: ACPIDebug: { "^^LPCB.XUSB/XRST", 0x0, 0x1, }
Sep 16 14:26:39 Matts-Mac-Pro kernel[0]: ACPIDebug: { "pr2/pr2m/pr3/pr3m", 0x0, 0x3fff, 0x3f, 0x3f, }
Sep 16 14:26:39 Matts-Mac-Pro kernel[0]: ACPIDebug: "calling XSEL from XWAK"
Sep 16 14:26:39 Matts-Mac-Pro kernel[0]: ACPIDebug: { "pr2/pr2m/pr3/pr3m", 0x3fff, 0x3fff, 0x3f, 0x3f, }
Sep 16 16:16:07 localhost kernel[0]: ACPIDebug: Version 0.1.3 starting
Sep 16 16:18:38 localhost kernel[0]: ACPIDebug: Version 0.1.3 starting
Sep 16 16:18:38 localhost kernel[0]: ACPIDebug: "calling XSEL from _OSC"
Sep 16 16:18:38 localhost kernel[0]: ACPIDebug: { "XSEL: pr2/pr2m/pr3/pr3m", 0x3fff, 0x3fff, 0x3f, 0x3f, }
Sep 16 16:18:38 localhost kernel[0]: ACPIDebug: { "XSEL: pr2/pr2m/pr3/pr3m", 0x3fff, 0x3fff, 0x3f, 0x3f, }
Sep 16 16:19:49 Matts-Mac-Pro kernel[0]: ACPIDebug: { "^^LPCB.XUSB/XRST", 0x0, 0x1, }
Sep 16 16:19:49 Matts-Mac-Pro kernel[0]: ACPIDebug: { "pr2/pr2m/pr3/pr3m", 0x0, 0x3fff, 0x3f, 0x3f, }
Sep 16 16:19:49 Matts-Mac-Pro kernel[0]: ACPIDebug: "calling XSEL from XWAK"
Sep 16 16:19:49 Matts-Mac-Pro kernel[0]: ACPIDebug: { "XSEL: pr2/pr2m/pr3/pr3m", 0x0, 0x3fff, 0x3f, 0x3f, }
Sep 16 16:19:49 Matts-Mac-Pro kernel[0]: ACPIDebug: { "XSEL: pr2/pr2m/pr3/pr3m", 0x3fff, 0x3fff, 0x3f, 0x3f, }
Sep 16 16:19:49 Matts-Mac-Pro kernel[0]: ACPIDebug: { "pr2/pr2m/pr3/pr3m", 0x3fff, 0x3fff, 0x3f, 0x3f, }
Matts-Mac-Pro:~ mattgarrad$

Ignore the MacPro bit, i booted off the wrong HDD that had a different clover config on it.

lol.