Ok just to close the loop here. Problem persisted with Intel Arc (the only thing I forgot to do was disable REBAR, but I don't think this is a related issue).

Upon installing Radeon RX 6400, pass through worked without any issue. Of note, on the Arc I got two separately numbered PCI devices 0000:0b and 0000:0a where as with the AMD it was 0000.08.01 and .02. I'm not sure if that helps hint at the issue --something to do with grouping and not grouping? This is way above my head but I noticed the difference.

I'm still having issues with pass through into Linux but that is a separate topic. Looks like we have a good discussion of this issue on the other post and so I will stay on that one going forward.

Thank you!

@olivierlambert ok perhaps I didn't include the full address. I've got the card replaced and now managed to corrupt grub (it's a Sunday...) so once I fix that we'll see what this new card has to offer in comparison. As a note I will say the AMD card did immediately pick up the console so I can already see there seems to be a better level of compatibility.

More to come later...!

@TeddyAstie great will take a look this weekend. Is there anything in particular you want us to test / check out?

@andSmv

From LSPCI:

0a:00.0 Audio device: Intel Corporation DG2 Audio Controller
0b:00.0 VGA compatible controller: Intel Corporation DG2 [Arc A310] (rev 05)

With xl pci-assignable-list :

0000:01:00.6
0000:01:00.4
0000:0b:00.0
0000:01:00.0
0000:01:00.3

Here you go...hope this helps!

@olivierlambert I'm going to swap in an AMD card today and will let you know if anything changes!

@TeddyAstie great will take a look this weekend. Is there anything in particular you want us to test / check out?

@olivierlambert yes happy to test as well thank you!

@olivierlambert great thank you. I think there is a Git issue also I found today 744. Not directly related but also driver related where the VGA disable function doesn’t work.

I think if in whatever future driver implementation there can be some more control of how it binds it would be really helpful for these use cases.

Thank you!

Has anyone had success in making PCI passthrough work to a Linux Desktop VM (in particular Hyprland)?

I am getting very mixed results but generally a lot of issues where there seams to be Seat contention for my USB devices (keyboard/mouse) and my GPU vs the Console (Cirrus VGA Driver/Virtual Input Devices).

I've been working through a series of blacklist and priority settings (mostly ChatGPT generated sadly) but was wondering if there is a more standard approach to this that someone knows/has figured out.

I can't imagine it's as hard as I'm making it so hoping one of your experts can help me here as I'm brand new to XCPNG!

Thank you!

@andSmv

From LSPCI:

0a:00.0 Audio device: Intel Corporation DG2 Audio Controller
0b:00.0 VGA compatible controller: Intel Corporation DG2 [Arc A310] (rev 05)

With xl pci-assignable-list :

0000:01:00.6
0000:01:00.4
0000:0b:00.0
0000:01:00.0
0000:01:00.3

Here you go...hope this helps!

Just adding here from my other post---AMD card didn't not reproduce this issue but my Intel Arc card experience matches exactly what others have posted here about the LSPCI vs XE-Assignable / XOA list difference.

Ok just to close the loop here. Problem persisted with Intel Arc (the only thing I forgot to do was disable REBAR, but I don't think this is a related issue).

Upon installing Radeon RX 6400, pass through worked without any issue. Of note, on the Arc I got two separately numbered PCI devices 0000:0b and 0000:0a where as with the AMD it was 0000.08.01 and .02. I'm not sure if that helps hint at the issue --something to do with grouping and not grouping? This is way above my head but I noticed the difference.

I'm still having issues with pass through into Linux but that is a separate topic. Looks like we have a good discussion of this issue on the other post and so I will stay on that one going forward.

Thank you!

@olivierlambert ok perhaps I didn't include the full address. I've got the card replaced and now managed to corrupt grub (it's a Sunday...) so once I fix that we'll see what this new card has to offer in comparison. As a note I will say the AMD card did immediately pick up the console so I can already see there seems to be a better level of compatibility.

More to come later...!

@olivierlambert I'm going to swap in an AMD card today and will let you know if anything changes!

This seems very similar to the problem I am having with my arc Arc310 Not Showing Up in Passthrough.

Perhaps there is a broader issue for these 'less than standard' PCI passthroughs?