Contribute
Register
I thought the best results were on ASUS Prime Deluxe (hot swap fully functioning)
I was referring to the person who carried out the tests. I explained myself badly :)
What I think is that it is a power management problem, as you suggested some time ago
 
Does this mean I could buy an ASUS extension card and get hot swap working, or is the ASUS card only compatible with ASUS mainboards?
 
Does this mean I could buy an ASUS extension card and get hot swap working, or is the ASUS card only compatible with ASUS mainboards?

I'm not completely certain what configurations will work or not, It just seems that the TB3 configs that are working have TB3 cards connected to CPU PCIe lanes. Gigabyte MBs don't seem to support this option. (But nothing is preventing anyone from trying)

ALSO: like @LeleTuratti pointed out, this isn't a hardware limitation (as Windows can address the cards just fine) It might be how OSX addresses TB3 configs through Platform Controller Hubs....
That said... Their kabylake machines... (Imac 18,3) If their video cards are running at 16x then the TB3 MUST be implemented on the PCH. If they're running at 8x, then it's possible the TB3 controllers are connected directly to PCU lanes.

Imac 18,3 Radoen 570 is 16x which means the TB3 ports must be through the PCH/DMI link..
So, it must be still possible to implement.
 
Last edited:
I cleared my config and tried again! To this point I've only had my Focusrite Clarett 4Pre to try with, but I managed to find a FW800 disk and a TB adapter and try with that, and I actually managed to wake from sleep and use it!

The strange thing with my Clarett 4Pre is that when I start my Hack without it connected my Focusrite Control software says "No Hardware Connected" or such. But when I wake from sleep with it connected (also connected prior to first boot), the TB indicator lights up on the device and the software still reads my interface, but it doesn't connect or interact with the software somehow and therefore no sound! It's strange but seeing the FW800 disk work after sleep, still gave some hope for the future!
 
Last edited:
No Gigabyte X299 board allow to have thunderbolt on CPU lanes, they all will be on PCH lanes.

The following SSDT made TB3 (with USB-C) hotplug work perfectly on a GB x299 Gaming 7 Pro.
You may have to adjust it for your need.
To do so, with IOJones (or IORegistry Explorer) search for NHI, look at the ACPI path.
In the Gaming 7, it's \_SB.PCI0.RP21

adjust accordingly

Seems that it doesn't work with Designare however.

I've used this SSDT for my Z170x-UD5-TH without much success (with some adaptions for my specific case). Hot swap doesn't work at all, and when I compare ioreg TB activity of my genuine MBP compared to my Hack with the SSDT, it looks completely different. Much more happening on the MBP.
I wonder if this SSDT is meant for boards with 2 combined TB/USB3.1 ports, I observe TB devices connecting on pci-bridges and not to TB itself.

Amazing that it works at all, TB on hacks.
 
What *I* don't understand about the PCI addresses, is that natively PCI0.RP05 in the Z170X-UD5 TH shows up at 7:0:0 (PXSX at 8:0:0 and TB Controller at 9:0:0) But as soon as we go to the SSDT those addresses start at 65:0:0. I have no understanding to ACPI/ASL but I find that interesting. also the 'reg' field gets corrupted with the SSDT and I'm not sure how to fix that...
 
Anyone have an update on ASRock Z270 mobo's?

I'm trying to get my Focusrite Clarett to reconnect after Sleep since it's a real bummer it disconnects upon Sleep. The Thunderbolt light turns off after Sleep, and turns on when exiting Sleep, and my audio device remains the Focusrite, but it doesn't play any audio. It remains "active" in IOReg whether connected or not.

A restart/cold start always reinitializes the audio.

Please find attached screenshots of my IOReg and the SSDT I'm using (based on version "B" in Post 1 of this thread). Thanks all
 

Attachments

  • Screen Shot 2018-04-14 at 6.21.15 PM.png
    Screen Shot 2018-04-14 at 6.21.15 PM.png
    302 KB · Views: 224
  • Screen Shot 2018-04-14 at 6.21.07 PM.png
    Screen Shot 2018-04-14 at 6.21.07 PM.png
    321 KB · Views: 228
  • SSDT-TB3-PD.aml
    682 bytes · Views: 159
I have tried all three variants attached in Post 1 ("A," "B," "C") as well as the three variants attached to this post. None of them make any difference to device usability: USB 3.1 devices still work before Sleep and after (though they are illegally ejected upon Sleep), and Thunderbolt 3 devices work before Sleep but not after (I can only test with my Thunderbolt 2 Focusrite Clarett audio interface).

My Clarett still remains active and unaffected in IOReg before and after Sleep; same outcome if I boot with it plugged in then unplug it -- it remains active and unaffected in IOReg.

I will also attach STOCK_SYS_OUTPUT.dsl which is my overall system's DSDT

Anyone have any idea how to edit these entries to enable my Clarett to re-engage after Sleep??
 

Attachments

  • SSDT-TB3-PD.aml
    610 bytes · Views: 103
  • SSDT-C-TB3.aml
    2.2 KB · Views: 148
  • SSDT-TB3-PD2.aml
    522 bytes · Views: 158
  • STOCK_SYS_OUTPUT.dsl
    1.1 MB · Views: 84
Last edited:
I thought maybe disabling or returning 1 or 0 for _PS0 and _PS3 would make a difference -- it does not. Attached are my most recent SSDT's as based on those from an iMac 18,3.

HELPER works on the RP05 node only, calling the methods used on an iMac 18,3 and calling TBXX in place of PXSX

PD3 works on the RP05.PXSX node, calling all the Thunderbolt devices TB00, TB01, and TB02 instead of my stock pci-bridge0, pci-bridge1, and pci-bridge2

Still the same result. My audio interface does not work post-Sleep, but works perfectly on coldboot
I am at a loss
 

Attachments

  • SSDT-TB3-PD3.aml
    1.3 KB · Views: 133
  • SSDT-TB3-PD3.dsl
    12.2 KB · Views: 137
  • SSDT-TB3-HELPER.aml
    188 bytes · Views: 154
  • SSDT-TB3-HELPER.dsl
    1.2 KB · Views: 143
Back
Top