Nested VMware Workstation 17.x Virtualisation Broken

[Tony C]

I'm sorry but this is a bit of a repeating theme... I recently tried to run VMware Workstation 17 inside a Debian 12 VM on Proxmox 8 and despite the VTx flag being recognised and another VM running an older version of VMware player working, this setup didn't use the acceleration offered by VTx.

However the same applies for KVM/QEMU on a Debian 12 desktop. Before I get into the grizzly details should I pursue this issue here or with the KVM/QEMU folks? With the ESXi issue you got involved with upstream. Hence the question.

 

[_gabriel]

is vCPU set to "host"?

 

[Tony C]

Yes both this new Debian 12/VMware Workstation 17 VM and the older working Ubuntu/VMware Player VM have the CPU set to host and the relevant kernel module loaded on the PVE host. The years old VMplayer instance still works in a performant way as expected. Yet the new VMware Workstation 17 instance running on the same PVE host runs unaccelerated. I transferred the Debian 12 VM to an ESXi cluster with CPU passthrough and it worked as expected.

Admittedly I did originally set up the Debian12 VM on KVM/QEMU due to not being able to access my PVE host at the time and then transferred it later (that's why I know it also affects KVM/QEMU on Debian12). So I'll try from scratch directly on the PVE host and then dump out the configs and any log information (there wasn't anything interesting at the time).

 

[_gabriel]

What is the host cpu ?

 

[frankmanzhu]

You need to tweak a bit on the cpu args for version 8. And I am still on version 7. I've tried the 2nd option on Intel and AMD nodes. Both works ok.
Let me know if you can get it working.

https://forum.proxmox.com/threads/f...orkstation-under-proxmox-8.141673/post-689654