- Joined
- Oct 24, 2013
- Messages
- 1,253
- Motherboard
- Asus ROG Z690 Formula
- CPU
- i9-13900K
- Graphics
- RX 6900 XT
- Mac
- Mobile Phone
Why does the new SSDT work to enable Thunderbolt bus and local node, bypassing the need to patch the NVM ?
Thunderbolt 4, modified firmware, Full Thunderbolt Bus tree
- Thread starter Elias64Fr
- Start date
CaseySJ
Moderator
- Joined
- Nov 11, 2018
- Messages
- 22,212
- Motherboard
- Asus ProArt Z690-Creator
- CPU
- i7-12700K
- Graphics
- RX 6800 XT
- Mac
- Classic Mac
- Mobile Phone
This is my speculative answer:
- As we know, there are write-once registers such as CFG-Lock
- They can only be written to once, which then prevent further changes to these registers (and to the resources they control) until system is rebooted
- It is possible that there's a write-once register that controls the ability to put a Thunderbolt Controller into either (a) ICM [Internal Connection Manager] or (b) Host Connection Manager mode
- If so, Asus Thunderbolt drivers may be unique in not locking that register, which then allows macOS to toggle Host Connection Manager (i.e. Thunderbolt Local Node)
In our early experiments, we actually succeeded in enabling Host connection manager (i.e. Thunderbolt Local Node) even on Gigabyte Z390 Designare using just SSDT.
Remember this post?
[SUCCESS] Gigabyte Designare Z390 (Thunderbolt 3) + i7-9700K + AMD RX 580
Do you have any other DAWs installed? Are those DAWs exhibiting the same symptoms? The reason I ask is because I was experiencing something similar and it ended up being a bug specific to the latest Cubase update which was causing a memory leak and abnormal CPU consumption of certain VST3...
www.tonymacx86.com
But we were never able to connect and disconnect devices reliably, or even recover them from sleep. And sometimes after 20 seconds, "something" would flip that register back to ICM, and Thunderbolt Local Node would disappear. But something in Asus' EFI firmware allows macOS to change whatever parameter it needs to change such that Local Node both activates and runs reliably.
Last edited:
- Joined
- Nov 18, 2011
- Messages
- 193
- Motherboard
- Gigabyte Z170X Gaming G1-F22j-OC 1.0.0
- CPU
- i7-6700K
- Graphics
- HD 530/Vega 64
- Mac
- Mobile Phone
Do you think is possible to achieve this by changing NVRAM variables with modGRUBShell.efi or RU.efi ?This is my speculative answer:
Linux source code and Apple's own Thunderbolt SSDT from iMac19,1 shows that a single bit in a Thunderbolt register switches between ICM and Host management modes.
- As we know, there are write-once registers such as CFG-Lock
- They can only be written to once, which then prevent further changes to these registers (and to the resources they control) until system is rebooted
- It is possible that there's a write-once register that controls the ability to put a Thunderbolt Controller into either (a) ICM [Internal Connection Manager] or (b) Host Connection Manager mode
- If so, Asus Thunderbolt drivers may be unique in not locking that register, which then allows macOS to toggle Host Connection Manager (i.e. Thunderbolt Local Node)
In our early experiments, we actually succeeded in enabling Host connection manager (i.e. Thunderbolt Local Node) even on Gigabyte Z390 Designare using just SSDT.
Remember this post?
[SUCCESS] Gigabyte Designare Z390 (Thunderbolt 3) + i7-9700K + AMD RX 580
Do you have any other DAWs installed? Are those DAWs exhibiting the same symptoms? The reason I ask is because I was experiencing something similar and it ended up being a bug specific to the latest Cubase update which was causing a memory leak and abnormal CPU consumption of certain VST3...
But we were never able to connect and disconnect devices reliably, or even recover them from sleep. And sometimes after 20 seconds, "something" would flip that register back to ICM, and Thunderbolt Local Node would disappear. But something in Asus' EFI firmware allows macOS to change whatever parameter it needs to change such that Local Node both activates and runs reliably.
- Joined
- Nov 18, 2011
- Messages
- 193
- Motherboard
- Gigabyte Z170X Gaming G1-F22j-OC 1.0.0
- CPU
- i7-6700K
- Graphics
- HD 530/Vega 64
- Mac
- Mobile Phone
Maybe we need to find-out if there is an auto option instead of setting manually the Thunderbolt controller to ICM or HCM...
- Joined
- Oct 24, 2013
- Messages
- 1,253
- Motherboard
- Asus ROG Z690 Formula
- CPU
- i9-13900K
- Graphics
- RX 6900 XT
- Mac
- Mobile Phone
Interesting. Doesn’t usb4 and maple ridge run in software connection manager mode (unlike titan ridge)? Also with memory protection from the IOMMU ?This is my speculative answer:
Linux source code and Apple's own Thunderbolt SSDT from iMac19,1 shows that a single bit in a Thunderbolt register switches between ICM and Host management modes.
- As we know, there are write-once registers such as CFG-Lock
- They can only be written to once, which then prevent further changes to these registers (and to the resources they control) until system is rebooted
- It is possible that there's a write-once register that controls the ability to put a Thunderbolt Controller into either (a) ICM [Internal Connection Manager] or (b) Host Connection Manager mode
- If so, Asus Thunderbolt drivers may be unique in not locking that register, which then allows macOS to toggle Host Connection Manager (i.e. Thunderbolt Local Node)
In our early experiments, we actually succeeded in enabling Host connection manager (i.e. Thunderbolt Local Node) even on Gigabyte Z390 Designare using just SSDT.
Remember this post?
[SUCCESS] Gigabyte Designare Z390 (Thunderbolt 3) + i7-9700K + AMD RX 580
Do you have any other DAWs installed? Are those DAWs exhibiting the same symptoms? The reason I ask is because I was experiencing something similar and it ended up being a bug specific to the latest Cubase update which was causing a memory leak and abnormal CPU consumption of certain VST3...
But we were never able to connect and disconnect devices reliably, or even recover them from sleep. And sometimes after 20 seconds, "something" would flip that register back to ICM, and Thunderbolt Local Node would disappear. But something in Asus' EFI firmware allows macOS to change whatever parameter it needs to change such that Local Node both activates and runs reliably.
CaseySJ
Moderator
- Joined
- Nov 11, 2018
- Messages
- 22,212
- Motherboard
- Asus ProArt Z690-Creator
- CPU
- i7-12700K
- Graphics
- RX 6800 XT
- Mac
- Classic Mac
- Mobile Phone
Because macOS is unable to activate Local Node on Gigabyte motherboards even though the controller and firmware (NVM) are the same, it suggests that Gigabyte’s BIOS is not putting the Maple Ridge controller into the proper state.
When we examined Gigabyte Z690 BIOS almost two years ago we found a SSDT containing ACPI methods for supposedly switching between ICM and HCM. We tried to invoke those methods, but we weren’t successful. We also tried to replace the SSDT with a modified version, but that also failed. But maybe that experiment is worth redoing — because now we know that TBTFlags is the key (along with class-code and spoofing NHI device-id).
Last edited:
- Joined
- Nov 18, 2011
- Messages
- 193
- Motherboard
- Gigabyte Z170X Gaming G1-F22j-OC 1.0.0
- CPU
- i7-6700K
- Graphics
- HD 530/Vega 64
- Mac
- Mobile Phone
@Casey have you check-out Osy's repos at Github? There he talks about different aspects of Thunderbolt controller including ACPI related methods. I think he is very close...
- Joined
- Feb 16, 2012
- Messages
- 490
- Motherboard
- Asus Z790-P WiFi
- CPU
- i9-14900KF
- Graphics
- RX 6600
- Mac
- Classic Mac
- Mobile Phone
Welcome back to the thread about thunderbolt 4. I found an old Gigabyte z590i board that has been sitting on the shelf for 3 years and I still have a chip tray soldered in it.
www.tonymacx86.com
I also have a GC-MapleRidge and ordered an ASUS ThunderboltEX 4,
and an ASUS USB4 PCIe Gen4 Card.
Write if you need any from me.
Z490 & Z590 - Will Z590 ever have macOS Support ?
I am thinking of using a socket for Winbond 25Q80 flash memory. In order not to desolder it, it would have to apply an additional 3.3V to pin 8. But once my hand twitched and I burned the controller. Therefore, I would rather not risk anymore. Unless someone has an idea for a programmer that has...
www.tonymacx86.com
I also have a GC-MapleRidge and ordered an ASUS ThunderboltEX 4,
and an ASUS USB4 PCIe Gen4 Card.
Write if you need any from me.
Last edited:
CaseySJ
Moderator
- Joined
- Nov 11, 2018
- Messages
- 22,212
- Motherboard
- Asus ProArt Z690-Creator
- CPU
- i7-12700K
- Graphics
- RX 6800 XT
- Mac
- Classic Mac
- Mobile Phone
@ori69,Welcome back to the thread about thunderbolt 4. I found an old Gigabyte z590i board that has been sitting on the shelf for 3 years and I still have a chip tray soldered in it.
Z490 & Z590 - Will Z590 ever have macOS Support ?
I am thinking of using a socket for Winbond 25Q80 flash memory. In order not to desolder it, it would have to apply an additional 3.3V to pin 8. But once my hand twitched and I burned the controller. Therefore, I would rather not risk anymore. Unless someone has an idea for a programmer that has...
I also have a GC-MapleRidge and ordered an ASUS ThunderboltEX 4,
and an ASUS USB4 PCIe Gen4 Card.
Write if you need any from me.
Some immediate questions/suggestions:
- Have you tried installing GC-Maple Ridge into your Asus Z790-P WiFi with (a) original NVM firmware and (b) modified NVM firmware?
- If so, for original NVM firmware, does Thunderbolt Local Node appear using just SSDT and cold start?
- If modified NVM firmware has been flashed, does sleep work when Thunderbolt Local Node is enabled?
- Regarding the Gigabyte Z590i board, is it assembled and ready to boot? Is it running macOS Sonoma?
- Have you tried flashing the firmware?
- If so, does sleep work?
- Does hot plug work?