ghost82

You need to manually install drivers

1050ti is not supported in monterey (max supported os is high sierra with nvidia webdrivers). When you revert back make sure the virtual gpu is on bus 0, slot 1, function 0 and the xml is "mac os compatible".

Ok, can you attach: 1. a ioreg file with the dongle connected to usb-c with the adapter (matisse) 2. a ioreg file with the dongle connected to usb type A port without usb-c adapter (matisse) Before saving the ioreg files better to disconnect all usb devices you can, disconnect hub(s), webcam, usb storages, etc. I'd want to try to inject the usb map kext. I'm not confident that it will work with the hub, nor with the usb plugin (not because of the plugin but because of the emulated controller), but there could be some hope for the other usb type A ports. Do the other type A ports work with usb pendrives? We are ot here, sorry Simon

Sorry, my fault for not understanding, in real I don't know how your headset is made.. Let it be for now emulated controller(s), i.e. passthrourgh device only. Let's consider the passed through usb controller: 1. How many ports does it have and what type of ports it has?I understand at least one is usb-c 2. You are saying that the headset (I understood that you can connect a usb dongle, that connects via wireless/bluetooth to mic and speakers, or a usb cable, is it true?) connects fine both via dongle and usb cable to the usb c port of the matisse controller?Your goal is to connect to another type port or...?

A kext works if it's loaded, doesn't matter if it's injected or loaded from /L/E or S/L/E. When I was talking about the usb map kext I was referring to a kext to map the usb ports of the controller, and manually giving power values. Sorry but maybe I didn't understand, please reply to the following questions: 1. does the dongle works in the mac os vm if connected to a usb port (no hub), by passing through the usb device only (no usb controller passthrough)? 2. are you passing through a usb controller in the vm?if so, does the dongle works if connected to one of these ports (no hub)? Note: by "works", I mean only if it's detected by the os.

I would say...maybe...if the usb 3 controller is compatible and that dongle too with usb 3 (it seems it is since it works on the usb c port), there should be no issue. Maybe there is the need of a usb map kext injected in opencore. Third parties USB hubs don't play well with mac os too, even in official macs..

That's strange, a pegged cpu means the system hung. Send to me with a private message your zipped efi folder and I will update opencore and kexts and check it.

No, unraid runs on x64_x86, M1 is arm

WARNING: before the official release of mac os Ventura, please note that if you have a physical cpu without avx2 instructions (sandy, ivy bridge for example), it is strongly recommended to disable automatic downloads for updates: if you are using a smbios like the imacpro1,1 (offical mac is compatible with ventura) the update should popup, download and install, but your working vm will stop working because cpus without avx2 are not compatible with Ventura. Note that there is no emulation for avx2.

As far as libvirt, everything seems good: I think the issue is with oclp causing issues with a dual gpu setup (vnc and dgpu), or patching something else. Is your SIP value correct in the opencore config.plist (you need to disable authenticated root)? You can try the following, one or the other: 1. Try boot arg: amfi_get_out_of_my_way=1 This in the config.plist of opencore 2. Reinstall mac os monterey with nvidia only (remove the vnc section before booting the installer): you should be able to boot with graphic output on your monitor but without acceleration, then don't use oclp but the geforce kepler patcher, so to be sure to add only the kepler drivers and to not patch anything other: https://github.com/chris1111/Geforce-Kepler-patcher You will need to change the SIP value to be able to patch root volume, the app will tell you what to do.

please post a diagnostics with working conditions and one with non working.

Monterey has no nvidia drivers, that's why it doesn't work. Either you break the volume seal and inject nvidia drivers or you need to switch to a compatible gpu, check dortania guides for this. As far as the errors they are not errors but warnings, you are trying to pass cpu features that your physical cpu doesn't have: they simply are ignored, logging the warnings.

Take into account that if you delete the libvirt.img you will lose the xml for your virtual machine and you will need to recreate the xmls. So, you can delete the libvirt.img, create again the vm by pointing them to where you have their vdisks or real disks.

For vlan this should work: https://forums.unraid.net/topic/57083-vlan-setup-for-vms/?do=findComment&comment=559196

What you have to do is to simply inject files by booting the ovmf bios setup of your current windows virtual machine. You can use only your actual windows machine even for creating the certificate for PK, or use another vm or another pc with linux. If it's not clear what that tutorial describes: 1- Create a private key and certificate to be used for PK 2- Download microsoft files (certificates for KEK and DB) 3- Run a basic virtual machine, with a virtual disk with certificates in it, to boot into uefi shell, ovmf bios and inject the files into the ovmf VARS file (maybe this wasn't clear...:it runs qemu via command line then it manages that basic vm through virt manager to access the gui of the uefi shell and ovmf bios) --> by injecting certificates with a basic vm your ovmf vars file will be "cleaner", because it contains only injected certificates, otherwise vars file contains other info, such as the boot drive and other uefi variables. It isn't needed you can boot your windows virtual machine and inject files from there, check my point (3) What you have to do: 1. Download attachment Microsoft.zip, extract it somewhere, it contains files for KEK and DB: if you don't trust me download files directly from microsoft 2. Generate a private key and a .der certificate: I used a linux virtual machine with openssl installed, run these commands in a terminal: openssl req -newkey rsa:2048 -nodes -keyout PKpriv.key -x509 -days 358000 -out PKtestDER.crt NOTE: the certificate will expire after 980 years!!!(358000 days); for security you may want to decrease this amount of time. Convert the created certificate from .crt to .der: openssl x509 -in PKtestDER.crt -outform der -out PKtestDER.der 3. Create a new vdisk with unraid terminal to copy all the needed files to inject certificates: qemu-img create -f raw /path/to/vdisk.img 2G 4. Format FAT32 that disk: mkfs.vfat -s 16 -F 32 /path/to/vdisk.img 5. Mount vdisk.img somewhere and copy inside: MicCorUEFCA2011_2011-06-27.crt MicWinProPCA2011_2011-10-19.crt MicCorKEKCA2011_2011-06-24.crt PKtestDER.der 6. Unmount vdisk.img, modify your windows vm to include the new vdisk.img, for example add: <disk type='file' device='disk'> <driver name='qemu' type='raw' cache='writeback'/> <source file='/path/to/vdisk.img'/> <target dev='hdc' bus='sata'/> <address type='drive' controller='0' bus='0' target='0' unit='2'/> </disk> 7. run your windows vm, press esc to access ovmf settings 8. inject certificates and enable secure boot: Device Manager / Secure Boot Configuration / Secure Boot Mode: --> change from “Standard Mode” to “Custom Mode” “Custom Secure Boot Options” will show up, click and enter PK Options / Enroll PK / Enroll PK Using File: select the HD corresponding to vdisk.img, select PKtestDer.cer, then “Commit Changes and Exit” KEK Options / Enroll KEK / Enroll KEK Using File: select the HD corresponding to vdisk.img, select MicCorKEKCA2011_2011-06-24.crt, then “Commit Changes and Exit” DB Options / Enroll Signature / Enroll Signature Using File: select the HD corresponding to vdisk.img, select MicWinProPCA2011_2011-10-19.crt, then “Commit Changes and Exit” DB Options / Enroll Signature / Enroll Signature Using File: select the HD corresponding to vdisk.img, select MicCorUEFCA2011_2011-06-27.crt, then “Commit Changes and Exit” Save with F10 and shutdown the vm. Now all the files are injected into the OVMF_VARS.fd file. -------- I'm attaching OVMF-2022-05-TPM-SECBOOT-ENABLED.zip, which contains my compiled version of ovmf 202205 with tpm and secure boot, with secure boot enabled and injected certificates. These files are ready to be used (just change in the xml the paths to ovmf code and ovmf vars), but take into account that PKtestDER.der was generated with a private key which I own. If you want your own PKtestDER.der you need to follow all the above steps. If you open OVMF_VARS_SECBOOT.fd with a hex editor you will notice the certificates injected. NOTE: I compiled this ovmf without smm, so you don't need smm emulation set in libvirt Microsoft.zip OVMF-2022-05-TPM-SECBOOT-ENABLED.zip

He can't as I wrote, the checkmarks are greyed out because probably he has the unraid usb plugged in a usb controller in iommu group 19 (I bet it's the sandisk ultra usb 3.0). Enable acs override, reboot and post the whole diagnostics not only xml.

---> As suggested, you should try to passthrough the device by setting the source bus and port: this will emulate a usb controller to which your bt usb will be attached; since the controller is emulated there is the possibility that it wont work. Moreover if you passthrough the device to the emulated controller it couldn't work if devices are renumbered. You iommu group 19 contains several devices, probably you didn't enable acs override. Enable it and see if iommu group 19 is splitted. As it is now, you can't passthrough 07:00.3 because your unraid usb is probably at 07:00.1, in the same iommu group. Enable acs override, reboot and check if 07:00.3 is splitted, if it is, bind it to vfio, set the vm to use that controller and reboot. Probably because you already used address 05:00.0 for something else, without looking the diagnostics impossible to help.

There's no trace of bluetooth in your post..Why you set the wifi on slot 1? put it on bus 5, slot 0, function 0. As far as bluetooth, it's a usb device, that wifi (or better bluetooth) should have a physical connection to a usb controller, better to passthorugh the whole usb controller to which it is attached, check with lsusb command.

I think this is due to the nvidia gpu running in a vm. Try to change the features block in your xml to: <features> <acpi/> <apic/> <hyperv> <relaxed state="on"/> <vapic state="on"/> <spinlocks state="on" retries="8191"/> <vendor_id state="on" value="1234567890ab"/> </hyperv> <kvm> <hidden state="on"/> </kvm> <vmport state="off"/> <smm state="on"/> </features>

6000 series is not affected. 3060ti should be better but it should cost more than the 6700xt.

That OVMF files are tpm enabled and secure boot compatible (which is enough for installing and running windows 11): this doesn't mean secure boot is enabled, but only capable. To enable secure boot you need microsoft keys injected. In other words ovmf was compiled with these flags: -D SECURE_BOOT_ENABLE -D TPM2_ENABLE or with: -D SECURE_BOOT_ENABLE -D TPM_ENABLE depending on the ovmf version; but microsoft keys injection is up to you. https://projectacrn.github.io/latest/tutorials/waag-secure-boot.html

If you want to use wifi for the server and you have a spare ethernet port, you can use an ethernet --> wifi bridge; for my servers I use: TP-link TL-WR802N (amazon euro 23,99) Vonets VAP11G-500 (amazon euro 36,99) Or any other ethernet --> wifi bridge, setup as client. I prefer the tp-link bridge over the vonets and it's running stable on my servers from a long time.

Just for info, that error meant that there was no hardware attached to that address: this was expected since it referred to the old motherboard.

You did not understand what is unraid. For vms in general, unraid includes qemu and libvirt and adds a gui to manage vms, nothing more; qemu and libvirt are external projects. Switching to something else will not solve your issues, unless the os is using something different than qemu/libvirt and/or other kernel versions. Why not complaining with amd, or nvidia, or with your motherboard manufacturer?If you correctly configured your vm and it still doesn't work the issue may be in your mb bios, or in your gpu firmware. I'm running linux/windows/mac os vms all days from 4 years, everyday, never had an issue with nvidia or amd gpus.

This is very unusual.. Sorry if this may sound stupid, but check that you are not using the digital hdmi audio instead of the analogic one. Do you have your audio jack plugged into the audio mainboard connected to speakers?did you check if you have volume control on your speakers?