Xen Guru group | XCP-ng and XO forum

TeddyAstie

@benapetr said in New project - XenAdminQt - a cross-platform GNU/Linux, macOS, Windows native thick client:

@Pilow I know them a little bit, I will have a look, but I am now working on another new cool thing! It's called xen_exporter: https://github.com/benapetr/xen_exporter

It's a prometheus exporter that hooks directly to xen kernel via xenctrl library from dom0 and extract all low-level metrics from the host, allowing very detailed graphs with very low granularity with stuff I always missed in both XenOrchestra and XenAdmin:

...

We have a similar project : https://github.com/xcp-ng/xcp-metrics, but unfortunately it's not used as of today (though it could get revived as Rust for Xen matures, i.e easier to build).
There is also Xen Orchestra OpenMetrics support but it's not on XCP-ng itself.

TeddyAstie

@DustyArmstrong

EDIT: It looks like I may just have a fake BIOS? The settings to enable all the relevant components (IOMMU, DMAr support etc) don't actually seem to do anything, they might just be for show - dmesg | grep -i iommu returns nothing, dmesg | grep -i -e dmar -e vfio -e pciback only shows pciback info, and cat /proc/cmdline contains nothing about IOMMU. Oddly, XO is still reporting that IOMMU is enabled:

dmesg in the Dom0 will not report the information you're looking for.
To know if PCI Passthrough is supported (e.g IOMMU enabled), you should check xl info | grep virt_caps and look for hvm_directio. You can also look for IOMMU-related stuff in xl dmesg.
As you managed to passthrough the device (even if not working in the guest), I don't see a issue there.

[ 4.655776] amdgpu 0000:00:08.0: amdgpu: Unable to locate a BIOS ROM
[ 4.655797] amdgpu 0000:00:08.0: amdgpu: Fatal error during GPU init
[ 4.655812] amdgpu 0000:00:08.0: amdgpu: amdgpu: finishing device.
[ 4.656681] amdgpu 0000:00:08.0: probe with driver amdgpu failed with error -22

Is there a trick to this, has anyone had success with this kind of AMD GPU? On my old hosts, enabling pass through was enough for it to just kind of work (Intel HD 530). The host machine outputs to a display normally when the card is in-use by the host. I am of the understanding the ROM is just part of the motherboard/GPU, there is some suggestion it can be dumped from the host-side, but I'm unsure on this.

Looks like the GPU ROMBAR is missing in the guest, while it's ok for many devices, many others will fail to work without it (like this GPU).
To me, there's something missing on the PCI Passthrough logic, I just brought the topic internally to see what we can do.

dinhngtu

@rniesen Do you mean occasional hangs? Yes, Server 2025 is known to hang if these options are not present.

dinhngtu

@rniesen Your VM is missing Viridian extensions (Hyper-V-compatible VM enlightenments implemented by Xen). This is typically due to the wrong VM template being chosen at the time of creation; the built-in Windows template should be used for all Windows VMs. There's an issue with Server 2025 that causes it to not run well without these features enabled.

Please power off your VM, run this command to add them back, then reboot:

xe vm-param-set uuid=<vm-uuid> platform:viridian=true platform:viridian_time_ref_count=true platform:viridian_reference_tsc=true platform:viridian_apic_assist=true platform:viridian_crash_ctl=true platform:viridian_stimer=true

dinhngtu

@rniesen

Hi,

Could you check the VM's Viridian settings from the host:

xe vm-param-get uuid=... param-name=platform

Followed by checking the Windows BCDEDIT settings:

bcdedit /enum all

TeddyAstie

@jbamford
Hello, can you give the output of xl info in Dom0 ? If you see hvm_directio in front of virt_caps, then PCI Passthrough is supported (i.e you have VT-d).

You can also check xl dmesg for vt-d specific messages (you can also give the output of it).

Xen Guru

Posts

Member List