2 problems (Nested_Virtualization && pci_hp_register failed )

[Psyc100]

Hello, I have 2 problems with Proxmox (maybe my fault I discover the software I do not understand everything).

For more information about the physical computer is an HP Prodesk whith an Core(TM) i5-10500, 16Gb of ram and the VTX and VTD is enabled in the bios. The Proxmox is freshly downloaded and instaled version 7.1-4 directly with the iso of Proxmox. The virtual machine is an Debian 11 with OVMF Bios, VM Ware display (due to an graphical bug with the standard VGA) on an Q35 Machine type.


First, when I start my virtual machine under Debian 11 I have this error:



Namely that I have linked two pcie devices directly to my VM (an SSD and a network card both work perfectly for the SSD I had to use this setting in the configuration file of my vm: " args: -set device.hostpci1.x-msix-relocation=bar2 " because it generate an error at VM startup) I have of course removed these PCIE links for test but the message is still there. (i have also random try to add "pcid" on cpu but no result )

---

Then my other problem is that my Debian VM will have to host a software that needs sometimes a VM too (with qemu) i have verified my "Nested_Virtualization" on my host machine with :
"root@proxmox:~# cat /sys/module/kvm_intel/parameters/nested"

And i have a " Y " so ok i still tried to :
"options kvm-intel nested=Y" > /etc/modprobe. d/kvm-intel.conf"

Following the tutorial here but same result on my VM when I do :
egrep '(vmx|svm)' --color=always /proc/cpuinfo

I have no result on the VM. I tried to add the option "hv-evmcs" but it tells me that it depends on "hv-vapic" but that doesn't exist and when I add it in the conf file it resets all the cpu configuration

Update : after lot try, i have set the cpu to "host" in Hardware tab and is work !

thank you in advance for your help !

 

[Psyc100]

Hello, every thinks works but i still have the error a boot and more now when i changed the CPU to "host" ...

This is a new capture :


Hardware has no pitchlock
For this one i have juste changed the display config this is ok

no codecs found
if is about media codecs i dont care but if i can dont have this message ? ^^


and i have found nothing to do for the pci_hp_register ...

Thanks you by advance

 

[Lefuneste]

Same issue here on brand new install

 

[Whitterquick]

Same in VMs...

 

[deliam]

Hello,

Disable ACPI Suport and problem solve.


Deliam Silva

 

[Whitterquick]

Hello,

Disable ACPI Suport and problem solve.


Deliam Silva

I thought disabling ACPI is what causes it?

 

[deliam]

Hi,

In my environment, ACPI causes this problem, just turn it off in OPTIONS and the problem goes away.

Deliam Silva

 

[Whitterquick]

Hi,

In my environment, ACPI causes this problem, just turn it off in OPTIONS and the problem goes away.

Deliam Silva

Fantastic! Thank you.

 

[Phydoux]

I am getting this error trying to install AlmaLinux 9. I had the same ACPI issues but turning that off fixed that issue. Now I have this one.

I'm installing it in VirtualBox (having some mouse alignment issues) so I know there's nothing wrong with the ISO in general. I have no idea whata all this means.

 

[markus-r]

Look here:
https://forum.proxmox.com/threads/kernel-panic-when-creating-vms-centos-9-stream-iso.104656/

This seems to be the same for AlmaLinux 9.

 

[LuchS]

Hello, every thinks works but i still have the error a boot and more now when i changed the CPU to "host" ...

This is a new capture :
View attachment 31831

Hardware has no pitchlock
For this one i have juste changed the display config this is ok

no codecs found
if is about media codecs i dont care but if i can dont have this message ? ^^


and i have found nothing to do for the pci_hp_register ...

Thanks you by advance

I have the same problems with my GPU Passthrough on laptop gtx1660ti max-q to Linux guest and i have no problems with my Windows guest. At the Linux guest all works well with Nvidia tesla-driver 418-450, the newest driver has a problem = black screen with thouse text after nvidia driver installation. This does not matter wich distributive of Linux used for guest - Debian/Ubuntu/Fedora/Mint/KDEneon/Manjaro/Garuda/Deepin...
Kernel of guest less than 5 does not support, only well working 5+

This is starts ~ after my Proxmox upgrade to version 7.1+, Version 7.0 worked vell, as i remember.
Now i am using Linux with kernel 5.xx (i tested 5.8/5.10/5.15/5.18 at the Debian 10/11/12tested guest) + Nvidia tesla driver 450. IDK why - but this is works fine... with thouse errors

 

[qingzhuoranguo]

I got the exact same error when installing Ubuntu 20.04, and I think this sounds stupid, but you just wait for like 5 mins, then it will pass to normal boot process, to me it was installation process..

 

[LuchS]

This does not work for me

 

[LuchS]

I have the same problems with my GPU Passthrough on laptop gtx1660ti max-q to Linux guest and i have no problems with my Windows guest. At the Linux guest all works well with Nvidia tesla-driver 418-450, the newest driver has a problem = black screen with thouse text after nvidia driver installation. This does not matter wich distributive of Linux used for guest - Debian/Ubuntu/Fedora/Mint/KDEneon/Manjaro/Garuda/Deepin...
Kernel of guest less than 5 does not support, only well working 5+

This is starts ~ after my Proxmox upgrade to version 7.1+, Version 7.0 worked vell, as i remember.
Now i am using Linux with kernel 5.xx (i tested 5.8/5.10/5.15/5.18 at the Debian 10/11/12tested guest) + Nvidia tesla driver 450. IDK why - but this is works fine... with thouse errors

my problem is over. IDK why, Nvidia told to me upgrade BIOS, but turning off parametr of a virtual machine (delete = do not use x-vga=on/x-vga=1 and use vga: none) allow me to use at linux guests nvidia drivers 510 and higher. But not at the Debian