iGPU HD530 passthrough issue on PVE 9

Z

Zippi

Member
Nov 17, 2022
41
3
13
Hi everybody,
I just updated my HP 800 G2 mini to ProxMox 9, and the HD530's iGPU passthrough no longer works.
I checked all the PCI passthrough configurations and everything is fine, but Windows 11 still gives me error 43.
Is this a known issue with ProxMox 9, or am I missing something?
I apologize if this has already been addressed, but I couldn't find anything.
Thanks
 
Does not work with PVE9. It was reported to work with an external module, but I could not get it to work.
It works with PVE8.

dakralex

Post in thread 'updated to PVE 9.0 then GPU passthrough stopped working on Windows'

dakralex said:
I haven't been too close to the Intel IGD quirks in QEMU before, but looking at a diff between v9.2.0 (the latest pve8 version), v10.0.2 (the current pve9 version) and the current upstream master, then there seems that v10.0.2 introduced a bug that was fixed by master with [1], where certain VGA memory regions aren't exposed to the guest anymore and the Windows driver seems to depend on those. I'll post more if this fixes the problem.
Click to expand...
Unfortunately, that was a false flag as the patch that introduced the bug is not in v10.0.2, but is in the upstream's...
  • Like
 
  • Like
Reactions: Zippi
uzumo said:
Does not work with PVE9. It was reported to work with an external module, but I could not get it to work.
It works with PVE8.

dakralex

Post in thread 'updated to PVE 9.0 then GPU passthrough stopped working on Windows'

dakralex said:
I haven't been too close to the Intel IGD quirks in QEMU before, but looking at a diff between v9.2.0 (the latest pve8 version), v10.0.2 (the current pve9 version) and the current upstream master, then there seems that v10.0.2 introduced a bug that was fixed by master with [1], where certain VGA memory regions aren't exposed to the guest anymore and the Windows driver seems to depend on those. I'll post more if this fixes the problem.
Click to expand...
Unfortunately, that was a false flag as the patch that introduced the bug is not in v10.0.2, but is in the upstream's...
  • Like
Click to expand...
Yes, it worked fine for me with PVE8 too.
So, can we hope for a fix in some future ProxMox update or not?
Thanks for the quick reply!
 
uzumo said:
Only Prokmox knows how this will turn out, so no one else can say.

Let's wait for progress on the thread under discussion.
Click to expand...
OK, I just found the other thread I hadn't seen before.

Let's wait hopefully...

Thanks again for the quick reply, hopefully we'll have some news soon!
 
@LongQT-sea
Hi,
I tried the procedure you suggested and... IT WORKS!
I'll explain in detail what I did and the hardware I have so I can perhaps help someone else:

1 - ProxMox 9.1.2
2 - Hardware: HP EliteDesk 800 G2 Mini
3 - CPU SkyLake Intel core i7 6700 and 6700T ( i've 3 nodes mix )

I initially tried the file: SKL_CML_GOPv9_igd.rom ( for Skylake to Comet Lake (6/7/8/9/10th gen) ) but unfortunately it didn't work, the VM with Windows 11 started but I had 100% memory and 25% CPU but it never finished loading the VM completely until the end.
I then tried Universal_noGOP_igd.rom and IT WORKS!!!!!
The only doubt I had was about point 2 of the guide, that is, the choice between LEGACY MODE and UPT MODE, I chose the first and I did this:
qm set [VMID] --machine pc \
--vga none \
--bios ovmf \
--hostpci0 0000:00:02.0,legacy-igd=1,romfile=igd.rom

But is this a choice I have to make? In other words, how do I want to set up the VM?

By the way, in my Win11 VM I had already set BIOS OVMF (UEFI) but as a machine I still have i440fx and not Q35
The fact is, everything is working fine now, and finally the infamous error 43 is GONE in the Win11 VM!
FANTASTIC!

YOU CHANGED MY DAY!

Do you think it's okay now, or should I try modifying something to use the first ROM that didn't work, because that might be the best one to use... or should I just leave it as is?

Thanks again, and if you have any questions, I'm here to help!
 
You must log in or register to reply here.
Top Bottom