Contribute
Register

Thunderbolt 3 Modified Firmware Custom SSDT Discussion

Wow, very interesting- and actually quite disappointing- since this would seem to have implications for using the cards on systems that don't have thunderbolt BIOS options, and probably also for using multiple cards in a system. i know i had seen at least one or two reports of people turning off thunderbolt support in BIOS and saying the cards still functioned using the 3-5 jumper, flashed firmware and TDROM SSDT.

Now i wonder if this could be firmware related. it seems feasible for instance that maybe we would need a firmware based on a real apple firmware for this to work since BIOS may need some kind of hooks for the ICM mode or whatever it was called that intel/windows uses for device management over thunderbolt. Granted that so far i dont think any of the apple based modified firmwares have worked as well as these recent (barely modified) official firmwares have, it might still be worth testing to see if flashing multiple cards with the semi-working apple based firmwares will allow them all to work together in macOS without a BIOS component (thunderbolt disabled in bios, or on system with no thunderbolt settings available in bios, or with multiple cards in one system).

g\

We know it works without the header because it works in real Mac 5.1 system. Others says it works when you turn it off in the bios, that is not the case for me. It works somewhat in my z170a system but it bugs out the grahipx card and some other stuff. Also I do not think the titan worked only the alpine.
 
@qthegaijin please keep us in the loop about your z170x results. I have a couple of clients with those motherboards and it is also the onboard controller equivalent of whats on the GC-AR Rev1 cards (DHL6540 DEV ID 1577).

@mango1122 following your posts in other threads, looks very good RE the onboard for Asrock Z390 itx using the patched NVM20. I have two of those boards so i might start there. Did you have to do a grub BIOS patch to get GPIO Force PWR? If so can you link to your post on the subject? i have a hazy memory of seeing it at some point.

Thanks,
g\
Details in here - https://www.tonymacx86.com/threads/...tb3-igpu-mojave-sff-build.277418/post-2092659
 
Update on my Alpine Ridge tests, sadly not good:
 
Some of this info is well documented but some of it is new so i am going to post it all together for completeness. It turns out all the programmable chips on my current motherboards are not accessible without removing them from their cases, so i will have to be doing that over the next few days. But i went through and surveyed the thunderbolt status on each of the boards i am currently using- pre any flashing. Also please note i own a 2019 mac pro 7,1 so if there is ever any info you need pulled from it just let me know- its a thunderbolt monster.

Mac Pro 7,1 - Titan JHL7540 15EB - NVM49.4
Z390-Desig - Titan JHL7540 15EB - NVM33

ASrock Z390 PG itx/ac - Alpine JHL6240 15BF - NVM14

X99 Designare EX - Alpine DHL6540 1577 - NVM21 (after official update)
X299 Deisgnare EX - Alpine JHL6540 15D2 - NVM26*
*This is interesting because this is the first time we are seeing NVM26 for this controller model. Unfortunately this board has a backplate covering the chip so on top of removing the motherboard from the case it will be even more disassembly to access the chip.

GC-AR Rev1 - Alpine DHL6540 1577 - NVM21 (after official update)
GC-AR Rev2 - Alpine JHL6540 15D2 - NVM20
GC-TR - Titan JHL7540 15EB - NVM43 (after official update)

I'll update the list above as needed.

g\
 
X299 Deisgnare EX - Alpine JHL6540 15D2 - NVM26*
*This is interesting because this is the first time we are seeing NVM26 for this controller model. Unfortunately this board has a backplate covering the chip so on top of removing the motherboard from the case it will be even more disassembly to access the chip.

Awesome, I use this board at my workplace and would love to know more about it as well. I saw you post in the other thread about the Tbt locations and how they could play a part in native functionality. Hopefully that leads to good places!
 
Mac Pro 7,1 - Titan JHL7540 15EB - NVM49.4

Any chance you pulled the firmware or willing to pull the firmware? Seems you are kind of versed with this. I tried to find the firmware on my Mac mini but I guess I am blind because I could not locate it.
 
i don't remember v# but ya i should have the original emails, i will dig it out and post it here. actually i think i did post it on this forum back in the day but thats probably going to be hard to find so i will dig it up and post it here.

g\
@qthegaijin

oops, almost completely forgot. attached is the "custom" BIOS which finally enabled thunderbolt AIC support on the z170x-designare.

g\
 

Attachments

  • Z170XDesignareGA.zip
    5.7 MB · Views: 81
Any chance you pulled the firmware or willing to pull the firmware? Seems you are kind of versed with this. I tried to find the firmware on my Mac mini but I guess I am blind because I could not locate it.
yep still on the list. hopefully the linux live boots work on the mac pro. And i guess i might as well list out all the other real macs i have just for posterity, nthough it does look like you can get all this info by investigating the firmware updates included with macOS updaters.
g\
 
@qthegaijin

oops, almost completely forgot. attached is the "custom" BIOS which finally enabled thunderbolt AIC support on the z170x-designare.

g\
Thank you! You rock.
Just tested this BIOS out. Weirdly it still doesn't want to show the alpine ridge AIC (even tried with my buddies GCAR AIC that wasnt flashed). Thats with stock firmware. Starting to think it may be this mobo but I am not sure. The Titan Ridge seems to get better results overall. Looking forward to your results ripping the MP7,1 firmware.
 
@CaseySJ could you provide any insight as to where to poke around for fixing thunderbolt devices not reconnecting on sleep inside our current SSDT?

I've been experimenting with loading various Mac thunderbolt SSDT's from their Darwin Dumps as well. With a few patches I'm able to get them to load but there is no Hotplug functionality. In fact, any SSDT that calls on GPE will throw a panic in the background. I did notice however that my Apollo Twin (which is daisy chained to the x16) shows up as DSB4 underneath DSB1 while the x16 still is loaded as a pci bridge. Neat! That is when using native SSDT's.

Lastly, the EX3 seems to be the only card that carries an empty HPM device. It loads no UUID. I wonder if these are related.
 
Back
Top