Ryzen 7000 Testing

I found something interesting, can you try adding the acpi=0x10000 boot arg and booting vanilla Monterey?
 
  • Like
Reactions: CaseySJ
I found something interesting, can you try adding the acpi=0x10000 boot arg and booting vanilla Monterey?
Just tried this -- unfortunately, VanillaMon goes into a boot loop very early in the process.

I should add that a PCIe card (Fenvi) is installed, which I believe is what we want.
 
Just tried this -- unfortunately, VanillaMon goes into a boot loop very early in the process.

I should add that a PCIe card (Fenvi) is installed, which I believe is what we want.
Where exactly does it panic? NVMe initialization?
 
Where exactly does it panic? NVMe initialization?
NVMe has been removed. Only Thunderbolt and Fenvi are installed. Have stepped out for lunch. Be back shortly.
 
  • Like
Reactions: ExtremeXT
Where exactly does it panic? NVMe initialization?

NVMe has been removed. Only Thunderbolt and Fenvi are installed. Have stepped out for lunch. Be back shortly.
Some Updates:
  • Thunderbolt disabled
  • NVMe SSD removed (since early this morning)
  • Fenvi FV-T919 installed
  • boot-arg includes acpi=0x10000
Results:
  • Vanilla Monterey boots up fully -- this is being posted from that boot
  • Fenvi WiFi appears in IOReg, but no driver attaches, as shown:
Screen Shot 2022-10-15 at 1.35.00 PM.png
 
  • Like
Reactions: ExtremeXT
Some Updates:
  • Thunderbolt disabled
  • NVMe SSD removed (since early this morning)
  • Fenvi FV-T919 installed
  • boot-arg includes acpi=0x10000
Results:
  • Vanilla Monterey boots up fully -- this is being posted from that boot
  • Fenvi WiFi appears in IOReg, but no driver attaches, as shown:
View attachment 7885
I think the only chance we have at properly fixing this issue is serial debugging, would you be able to get one of those set-up?
 
I think the only chance we have at properly fixing this issue is serial debugging, would you be able to get one of those set-up?
I read through the Dortania article, which seems to require the presence of an RS-232 serial port or header. Alas, this motherboard does not have RS-232.

An update to previous post:
  • Having the Fenvi by itself allows VanillaMon to boot with or without acpi=0x10000
  • Enabling on-board Thunderbolt, however, causes a boot loop in the late stage regardless of acpi=0x10000
 
Last edited:
which seems to require the presence of an RS-232 serial port or header. Alas, this motherboard does not have RS-232.
Also having it on board or via an USB RS232 could be a no go, because it could need of a UEFI driver
In the past I did some tests about it but I have had no success (I have used PCIEx RS 232 add-on and an USB One
OC devs said in the past I should have need of a UEFi driver for it
Maybe today things are different
 
Also having it on board or via an USB RS232 could be a no go, because it could need of a UEFI driver
In the past I did some tests about it but I have had no success (I have used PCIEx RS 232 add-on and an USB One
OC devs said in the past I should have need of a UEFi driver for it
Maybe today things are different
I already have a USB-to-RS232 cable that uses a FDTI chipset. Just verified that CoolTerm in macOS works with it. I’ll need a second such cable, preferably with null modem configuration— or a separate null modem adapter. These are cheap so I can set up the experiment soon.
 
Last edited:
Cross my fingers for you, but it seems the same steps I did and no useful data because I miss an Efi drivers in problematic rig
Reading data from another pc with macOS or windows didn’t produce a serial debug
 
Cross my fingers for you, but it seems the same steps I did and no useful data because I miss an Efi drivers in problematic rig
Reading data from another pc with macOS or windows didn’t produce a serial debug
Have you tried OpenCore's Serial feature? This seems to imply that OpenCore includes pre-boot drivers...
Screenshot 2022-10-16 at 4.14.01 AM.png
 
there was not this feature in OC 065 I used in that time
 
I already have a USB-to-RS232 cable that uses a FTDI chipset. Just verified that CoolTerm in macOS works with it. I’ll need a second such cable, preferably with null modem configuration— or a separate null modem adapter. These are cheap so I can set up the experiment soon.
Not sure if USB Serial will work, but we'll see...
 
  • Like
Reactions: CaseySJ
Not sure if USB Serial will work, but we'll see...

these for me didn't work
 
  • Like
Reactions: CaseySJ

these for me didn't work
I have just ordered this one, which has a CP2102 chipset.


Screenshot 2022-10-16 at 8.03.21 AM.png
 
Last edited:
I have a new idea, what if we redirect the kexts like IONVMeFamily to use another IOPCIFamily version which we'd inject with OC? It can be changed in the Info.plist of the kexts.
@CaseySJ Can you dump Monterey's IONVMeFamily?
 
  • Like
Reactions: CaseySJ
I have a new idea, what if we redirect the kexts like IONVMeFamily to use another IOPCIFamily version which we'd inject with OC? It can be changed in the Info.plist of the kexts.
@CaseySJ Can you dump Monterey's IONVMeFamily?
Sure -- kext attached (Monterey 12.6)
 

Attachments

@CaseySJ Please try this EFI, it's the latest one you sent but with:
  • Patched IOPCIFamily from Big Sur named BigSurIOPCIFamily.kext
  • Patched IONVMeFamily from Monterey named PatchedIONVMeFamily.kext
  • Patched AppleIntelI210Ethernet from Monterey named PatchedAppleIntelI210Ethernet.kext

The BundleIdentifier of the kexts was changed and the IOProbeScore was set higher than the original kexts to they attach to the devices instead of the vanilla ones.
Try with disabled Thunderbolt since that kext was not patched, if everything goes right we should have I225-V LAN and NVMe working without modifying the system partition.
 

Attachments

  • Like
Reactions: CaseySJ
If you get class errors, try these, you might also need to block IONVMeFamily with a kernel block (in the EFI but disabled).

You could also record the boot process.
 

Attachments

  • Like
Reactions: CaseySJ
@CaseySJ Please try this EFI, it's the latest one you sent but with:
  • Patched IOPCIFamily from Big Sur named BigSurIOPCIFamily.kext
  • Patched IONVMeFamily from Monterey named PatchedIONVMeFamily.kext
  • Patched AppleIntelI210Ethernet from Monterey named PatchedAppleIntelI210Ethernet.kext

The BundleIdentifier of the kexts was changed and the IOProbeScore was set higher than the original kexts to they attach to the devices instead of the vanilla ones.
Try with disabled Thunderbolt since that kext was not patched, if everything goes right we should have I225-V LAN and NVMe working without modifying the system partition.
It seems all of the patched kexts were skipped because the standard versions of those kexts have the same class names as those in the patched versions...

IMG_7598.PNG
 
  • Like
Reactions: ExtremeXT
  AdBlock Detected
Sure, ad-blocking software does a great job at blocking ads, but it also blocks some useful and important features of our website. For the best possible site experience please take a moment to disable your AdBlocker.