Ryzen 7000 Testing

[fabiosun]

@PoMpIs intersting this:


it seems also Real MacPro 2019 have this wrong detection for fulfilled slot

 

[hvv_]

IOUserEthernet

@CaseySJ
I did it but unfortunately the same symptoms.


I guess the only solution is to bypass AppleVTD or emulate it?

 

[ExtremeXT]

Back online!

 

[CaseySJ]

So I finally booted from a USB stick using the RELEASE-Mode EFI, on my X570 Creator/5950X build and I don't know if I did it right b/c at the bootloader I simply chose my working MacOS install and everything seemed to be working normally. I listened to several minutes of Spotify and it sounded perfect.

that said, the only thing I changed was OpenCore.efi which sounds like it doesn't change all that much.

like.. I don't need new Kerenal Patches and Kexts to go along with the TSC/FSB fix?

Really?

next I will copy the OpenCore.efi to the EFI partition of my actual working MacOS drive to test. but i'm too scared to do that just now.

I think I just have to Install Ventura on my 7950X system and hope for the best.

Yes, the audio fix is a change to OpenCore.efi only. Plug and play -- no config.plist changes needed.

 

[CaseySJ]

Interim Update #18:

• We see the first rays of hope for Aquantia 10GbE in Monterey 12.6 on AMD
• I purchased an inexpensive AQC-107 card that arrived late last night, and installed it this morning in my AM5 system
• Have a couple of appointments shortly so I'll explain details later today
• Meanwhile, here's something to whet your appetite!

 

[CaseySJ]

More Information About Aquantia 10GbE:

• In the previous post I was using an Aquantia AQC-107
• In the screenshot below we have an Aquantia AQC-113 (Gigabyte GC-AQC113C)
• This just demonstrates that both 107 and 113 models function properly; we would expect them to, but empirical proof is preferred over theory
• A simple SSDT is necessary to change the GC-AQC113C's device-id to 0x94c0

 

[DarkSilentSC]

So it confirms that overwriting those big sur kexts over will make it work? Is it not possible to “inject” big sur kext via Kernel/Add in OC? Just curious on how OC adding kext works.

 

[CaseySJ]

Details of Current Aquantia Status:

The steps to enable Aquantia 10GbE in Monterey are not ideal at the moment, but at least they provide two important pieces of information:

• They establish proof-of-concept, which means we now know that Aquantia 10GbE cards will work on AMD platform in latest version of Monterey despite lack of AppleVTD
• They isolate the problem to two areas:
  • IONetworkingFamily
  • DriverKit

We observed previously that the key to enabling Intel i225-V in Monterey and Ventura is to use boot argument e1000=0, which disables the AppleEthernetE1000 DriverKit extension (dext). So the question was, is there a way to disable the AppleEthernetIXGBE dext as well?

Looking at the disassembly of AppleEthernetE1000, we can see where e1000 boot argument is parsed and acted upon. So it was a straightforward task to find the equivalent in AppleEthernetIXGBE. The two screenshots are shown below.

This is AppleEthernetE1000: (this is Big Sur's version, hence "dk.e1000")

This is AppleEthernetIXGBE:

• From this we find a boot argument called ixgbe and we see that it is treated the same way as e1000
• Hence, adding boot-argument ixgbe=0 disables AppleEthernetIXGBE dext
  • This boot argument by itself is insufficient; it will cause system to freeze or render a self-assigned IP when cable is connected
• And we use the previously described surgical procedure to import IONetworkingFamily from Monterey 12.0.x into Monterey 12.6.x
• The task before us is to see whether we can find a way to patch AppleEthernetAquantiaAqtion so we can completely avoid the surgical procedure
  • I tried using OpenCore to inject IONetworkingFamily and AppleEthernetAquantiaAqtion, but it did not work
  • Would be great if others can join the search because that would avoid having to research a patch
• We also have to verify that just replacing AppleEthernetAquantiaAqtion (and not the entire IONetworkingFamily) will also work
  • This is what I plan to do shortly

 

[CaseySJ]

Brief Update:

• Just confirmed that it is not necessary to replace the entire IONetworkingFamily.kext
• Instead, we can just replace the single plug-in AppleEthernetAquantiaAqtion.kext
• This also means we may (or may not) be able to create a patch; if we needed to replace multiple kexts, a patch would be much harder to create

Important Notes (for testers only):

• On every cold boot, macOS will upload Aquantia firmware to the card from AppleEthernetAquantiaAqtionFirmware.kext, but on a warm boot, firmware is not uploaded
  • This will cause the next test to fail unless we do a cold boot between tests as described in next bullet
• When we want to test different combinations of IONetworkingFamily, we should perform a cold boot each time before testing
  • A cold boot means (a) shutdown the computer, (b) flip power switch on PSU to OFF for about 10 seconds, then (c) power back up
• This is also the reason why booting to Windows and then booting to macOS causes the Aquantia to fail
  • After booting to windows we can perform a cold boot to macOS in order to solve the problem

 

[Galve2000]

** Enabling Intel I225-V in Monterey **

To enable I225-V in Monterey, we just need to add boot argument e1000=0

​

You guys are geniuses!




this is being done on my X570/5950X running Ventura so I apologize for cross posting on this 7000 testing thread and the necesary steps for Monterey may be different so YMMV,

BUT..

I used an SSDT found in this thread and added AppleIntel210Ethernet.kext to my OC as well as the boot-arg e1000=0 and suddenly my long dormant i225-V sprang to life.

honestly I had been chugging along just fine with my Fenvi T-919 Broadcom WiFI.. but ethernet feels much more robust

I will say, it is a bit jarring that my ethernet connection is populated as en2 -- back in the chimera days if you did not have ethernet as en0, you did not have a TCP/IP connection.

but i'm trying not to be too OCD about cosmetic stuff like that.

you will note that in the above screenshot, wifi is disabled for testing purposes, and yet I am web browsing away on ethernet like it's the only game in town.

thank you all so much.

next up, we conquer Aquantia 10GB bc obviously, 2.,5 GB will just not do in 2022!

one final caveat.. this EFI is on a USB stick along with the updated RELEASE OpenCore.efi for Audio issues in Rizen 7000 CPUs. I have not yet transfered the OC to my NVNE, but I am feeling emboldened!

 

[ExtremeXT]

You guys are geniuses!


View attachment 8765

this is being done on my X570/5950X running Ventura so I apologize for cross posting on this 7000 testing thread and the necesary steps for Monterey may be different so YMMV,

BUT..

I used an SSDT found in this thread and added AppleIntel210Ethernet.kext to my OC as well as the boot-arg e1000=0 and suddenly my long dormant i225-V sprang to life.

honestly I had been chugging along just fine with my Fenvi T-919 Broadcom WiFI.. but ethernet feels much more robust

I will say, it is a bit jarring that my ethernet connection is populated as en2 -- back in the chimera days if you did not have ethernet as en0, you did not have a TCP/IP connection.

but i'm trying not to be too OCD about cosmetic stuff like that.

you will note that in the above screenshot, wifi is disabled for testing purposes, and yet I am web browsing away on ethernet like it's the only game in town.

thank you all so much.

next up, we conquer Aquantia 10GB bc obviously, 2.,5 GB will just not do in 2022!

one final caveat.. this EFI is on a USB stick along with the updated RELEASE OpenCore.efi for Audio issues in Rizen 7000 CPUs. I have not yet transfered the OC to my NVNE, but I am feeling emboldened!

I would use the SSDT uploaded in Casey's post, the base one from OC-Little has a lot of unnecessary properties.

 

[ExtremeXT]

Details of Current Aquantia Status:

The steps to enable Aquantia 10GbE in Monterey are not ideal at the moment, but at least they provide two important pieces of information:

• They establish proof-of-concept, which means we now know that Aquantia 10GbE cards will work on AMD platform in latest version of Monterey despite lack of AppleVTD
• They isolate the problem to two areas:
  • IONetworkingFamily
  • DriverKit

We observed previously that the key to enabling Intel i225-V in Monterey and Ventura is to use boot argument e1000=0, which disables the AppleEthernetE1000 DriverKit extension (dext). So the question was, is there a way to disable the AppleEthernetIXGBE dext as well?

Looking at the disassembly of AppleEthernetE1000, we can see where e1000 boot argument is parsed and acted upon. So it was a straightforward task to find the equivalent in AppleEthernetIXGBE. The two screenshots are shown below.

This is AppleEthernetE1000: (this is Big Sur's version, hence "dk.e1000")
View attachment 8760
This is AppleEthernetIXGBE:
View attachment 8759

• From this we find a boot argument called ixgbe and we see that it is treated the same way as e1000
• Hence, adding boot-argument ixgbe=0 disables AppleEthernetIXGBE dext
  • This boot argument by itself is insufficient; it will cause system to freeze or render a self-assigned IP when cable is connected
• And we use the previously described surgical procedure to import IONetworkingFamily from Monterey 12.0.x into Monterey 12.6.x
• The task before us is to see whether we can find a way to patch AppleEthernetAquantiaAqtionso we can completely avoid the surgical procedure
  • I tried using OpenCore to inject IONetworkingFamily and AppleEthernetAquantiaAqtion, but it did not work
  • Would be great if others can join the search because that would avoid having to research a patch
• We also have to verify that just replacing AppleEthernetAquantiaAqtion (and not the entire IONetworkingFamily) will also work
  • This is what I plan to do shortly

Could you upload the working and non working version of the Aquantia kext? IIRC 12.2.3 worked but 12.3 broke it, so having those 2 versions would limit the amount of unrelated changes when comparing the Assembly of the 2 versions.

 

[CaseySJ]

Could you upload the working and non working version of the Aquantia kext? IIRC 12.2.3 worked but 12.3 broke it, so having those 2 versions would limit the amount of unrelated changes when comparing the Assembly of the 2 versions.

The working version from 12.0.x was uploaded a couple of days ago:

Ryzen 7000 Testing

Great post but it should probably be in a Thread of its own, not on page 69. So it is better seen and the others with Aquantia Ethernet issues can also join in testing.

forum.amd-osx.com

Non-working version can be copied out of any recent version of Monterey.

 

[hvv_]

Could you upload the working and non working version of the Aquantia kext? IIRC 12.2.3 worked but 12.3 broke it, so having those 2 versions would limit the amount of unrelated changes when comparing the Assembly of the 2 versions.

No, Monterey 12.2.3 don't work only until 12.2.1

 

[ExtremeXT]

No, Monterey 12.2.3 don't work only until 12.2.1

Okay, is it possible to get the kext from both 12.2.1 and 12.2.3?

 

[CaseySJ]

I found a copy of Monterey Installer 17.2.02, which might be macOS 12.2 public beta 2. In any case, the Aquantia driver contained therein works in Monterey 12.6.x so I've added that to the Surgical Procedure post.

I started by examining and comparing the interrupt handler functions, which contain a call to medium_changed. There are a handful of these functions:

• medium_changed
• medium_enable
• medium_disable

These functions may be related to the detection of a cable (i.e. a medium for transferring the signal) and the subsequent attempt to establish a connection, but I'm not sure. I plan to spend more time on this later today.

 

[Galve2000]

I would use the SSDT uploaded in Casey's post, the base one from OC-Little has a lot of unnecessary properties.

I think I have both copies of the SSDT saved somewhere but IIRC I used the one from the thread that I linked to.

I will say, it's tough crowd in here.. everyone's a critic!

A month ago in the discord I was telling people how I had to use a special SSDT to get my PowerColor RDU 6900XT to work, and I was chided for using a "suboptimal" SSDT.

sometimes I wish we would notice that "It just works" and call it a day.

/vent

 

[fabiosun]

have any of you tried to reissue the 4 dimm slots with 4 32 gb modules to get 128 Gb?
in case you did with what brand and what frequencies?
I read very conflicting opinions about the robustness of the system with all slots full

Thank you

 

[ExtremeXT]

I think I have both copies of the SSDT saved somewhere but IIRC I used the one from the thread that I linked to.

I will say, it's tough crowd in here.. everyone's a critic!

A month ago in the discord I was telling people how I had to use a special SSDT to get my PowerColor RDU 6900XT to work, and I was chided for using a "suboptimal" SSDT.

sometimes I wish we would notice that "It just works" and call it a day.

/vent

The "if it works, it works" mentality is bad regarding to Hackintoshing. The SSDT you're using injects properties that your I225-V already has, which makes it useless, the SSDT I modifed for Casey only changes what's necessary.

 

[leesurone]

** Some Updates **

• I decided to assemble the system yesterday and started with Big Sur 11.6. No issues getting the O/S to install
  • AMD Ryzen 7 7700X 8-core 16-thread CPU
  • Asus Crosshair X670E Gene mATX board
  • G.SKill DDR5 memory (not EXPO)
  • BeQuiet! Pure Loop 2 FX 240mm liquid AIO cooler
  • PowerColor Radeon RX 5700XT GPU
• BIOS 0705 (latest)
• Big Sur is installed on SATA SSD mounted to an internal SATA port; will install NVMe SSD later
• iGPU disabled via SSDT
• Just completed the USB port map:
  • Aura LED controller enabled (see post below for liquidctl effects)
  • Realtek ALC4080 USB audio enabled
  • Intel Bluetooth enabled
  • Intel WiFi enabled
• Thunderbolt SSDT modified for this board, but no luck yet connecting Thunderbolt components; haven't spent much time on this yet
• Components that work:
  • Intel i225-V Ethernet port works out of the box
  • Intel WiFi and Intel Bluetooth both work properly with OpenIntelWireless drivers
    • Not using BCM94360NG or Fenvi FV-T919 at this time
  • USB 2, USB 3 Type A, USB 3 Type C ports all working

Issues / To-Do:

• Audio is occasionally choppy regardless of (a) Realtek 4080, (b) HDMI/DP output, (c) external USB DAC/Amp
• Disabling NB Azalia didn't seem to help, but will try again
• Each reboot causes a BIOS alert, requiring me to enter BIOS and Save/Exit
  • Fortunately all default BIOS parameters are compatible, but still need to figure this out
• Will work on sleep/wake soon

View attachment 7765View attachment 7766

Hi Casey, I decided to join the AM5 fray and parts are on the way. I decided to go with the same Asus board you are using in the end. For me it was always either the Asus Rog Crosshair X670E Gene or the Rog Strix X670E-I wifi as I prefer smaller form factor PCs. The Gene will fit perfectly into one of the small footprint Silverstone Alta G1M case my X570 system is currently in. Would you mind sharing the current state of your EFI folder in the Ventura flavor? Thanks in advance