Help with Hackintosh GPU passtrough RX580

asnando

New Member
Dec 27, 2023
3
0
1
Hello guys,

I followed a couple of guides for creating a VE with mac/windows, and after one week of work I got windows fully working, but my hackintosh vm got a strange behavior when using with GPU passtrough.

Initially the RX580 was not booting at all. Then I restored the original vbios from a RX570 GPU, cuz I saw that it's a common problem for GPUs bought from Aliexpress. Then after that, it started with the following behavior:

If I set the display=default and remove the PCI attached to the GPU, it works fine from the console, but if I set display=none and attach de PCI to GPU it loads the macos, but get stuck in the "ACPI: sleep states S3 S4 S5" for a couple of minutes(3-5 minutes), and after that it shows the password screen and logs normally into the macos with full GPU passtrough.

I tried with macos Sonoma, but same behavior, and now I'm trying again with Ventura.
Also tried to leave everything, except the GPU, unplugged, but no luck.

I am wondering what is causing this video delay, any idea?

ps: i have another ssd fully dedicated to hackintosh(no proxmox) and it works fine within the gpus. i searched this specific problem in hackintosh forums, but no resolution. It looks like some specific proxmox/hackintosh configuration.

Spec:
- Intel i5 10600K
- 16GB RAM
- Primary GPU RX6600M - allocated to windows vm
- Secondary GPU RX580 - allocated to macos vm
- ASUS TUF Gaming B460M-PLUS
- Fenvi BCM

Note: The current BIOS values are displayed on the left side of the picture.

Thanks in advance!
 

Attachments

  • IMG_9128.jpg
    IMG_9128.jpg
    827.7 KB · Views: 11
  • IMG_9127.jpg
    IMG_9127.jpg
    400 KB · Views: 11
I got the same behavior in a try with my RX6600M, and it got stuck on the same step. I have not tried the vendor-reset yet, but Im wondering if with that same scenario from a Navi23 GPU it may be required for the RX580. Do u know if is there any cons?
 
I did the vendor-reset installation and got same behavior. Here is the following dmesg output

root@pve:~# dmesg | grep vfio-pci
[ 5.316024] vfio-pci 0000:04:00.0: vgaarb: deactivate vga console
[ 5.316030] vfio-pci 0000:04:00.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=none:owns=none
[ 88.305659] vfio-pci 0000:04:00.0: AMD_POLARIS10: version 1.1
[ 88.305661] vfio-pci 0000:04:00.0: AMD_POLARIS10: performing pre-reset
[ 88.317573] vfio-pci 0000:04:00.0: AMD_POLARIS10: performing reset
[ 88.317577] vfio-pci 0000:04:00.0: AMD_POLARIS10: CLOCK_CNTL: 0x0, PC: 0x2c8c
[ 88.317578] vfio-pci 0000:04:00.0: AMD_POLARIS10: performing post-reset
[ 88.341432] vfio-pci 0000:04:00.0: AMD_POLARIS10: reset result = 0
[ 90.232681] vfio-pci 0000:04:00.0: enabling device (0400 -> 0403)
[ 90.232807] vfio-pci 0000:04:00.0: AMD_POLARIS10: version 1.1
[ 90.232809] vfio-pci 0000:04:00.0: AMD_POLARIS10: performing pre-reset
[ 90.245042] vfio-pci 0000:04:00.0: AMD_POLARIS10: performing reset
[ 90.245045] vfio-pci 0000:04:00.0: AMD_POLARIS10: CLOCK_CNTL: 0x0, PC: 0x2a44
[ 90.245047] vfio-pci 0000:04:00.0: AMD_POLARIS10: performing post-reset
[ 90.269407] vfio-pci 0000:04:00.0: AMD_POLARIS10: reset result = 0
[ 90.281518] vfio-pci 0000:04:00.1: enabling device (0000 -> 0002)
[ 90.317580] vfio-pci 0000:04:00.0: AMD_POLARIS10: version 1.1
[ 90.317583] vfio-pci 0000:04:00.0: AMD_POLARIS10: performing pre-reset
[ 90.317685] vfio-pci 0000:04:00.0: AMD_POLARIS10: performing reset
[ 90.317688] vfio-pci 0000:04:00.0: AMD_POLARIS10: CLOCK_CNTL: 0x0, PC: 0x2950
[ 90.317689] vfio-pci 0000:04:00.0: AMD_POLARIS10: performing post-reset
[ 90.341562] vfio-pci 0000:04:00.0: AMD_POLARIS10: reset result = 0