I've been trying to passthrough an Elgato HD60 Pro to a Windows 10 Pro VM, but when the VM starts it will hang on POST or will continuously show the spinner and reboot after a while.
I attempted to boot the VM into safe mode, but got the same result.
There aren't any errors displayed in ProxMox, and nothing outside of the VM is affected.
Forcing the VM to stop or restart will post the following error on the host:
Running lspci after this will display the following:
A similar issue has been reported in a thread for unRAID, but the posted solutions of adding 'disable_idle_d3=1' to the vfio entry for the capture card hasn't helped.
The other posted fix of patching the kernel with an update to support Yuan-Yuan capture devices, and enable 'quirk_no_pm_reset' and 'quirk_no_bus_reset', is several years old at this point and I don't know if it would even be supported on newer kernel versions.
I currently have two PCIe devices passed into the VM (excluding the capture card); a GTX 760, and a USB 3 PCIe card.
I've tried swapping the positions of the capture & USB cards without any luck and have tried removing the USB card from the MV config to see if there was some sort of interference.
I'm running a dual Xeon E5-2670 setup on an ASRock Rack ep2c602-4l/d16, the system has full IOMMU separation.
I attempted to boot the VM into safe mode, but got the same result.
There aren't any errors displayed in ProxMox, and nothing outside of the VM is affected.
Forcing the VM to stop or restart will post the following error on the host:
Code:
vfio-pci 0000:02:00:0: timed out waiting for pending transaction; performing level reset anywayRunning lspci after this will display the following:
Code:
02:00.0 0000: 12ab:0380 (rev ff) (prog-if ff)
!!! Unknown header type 7f
Kernel driver in use: vfio-pciA similar issue has been reported in a thread for unRAID, but the posted solutions of adding 'disable_idle_d3=1' to the vfio entry for the capture card hasn't helped.
The other posted fix of patching the kernel with an update to support Yuan-Yuan capture devices, and enable 'quirk_no_pm_reset' and 'quirk_no_bus_reset', is several years old at this point and I don't know if it would even be supported on newer kernel versions.
I currently have two PCIe devices passed into the VM (excluding the capture card); a GTX 760, and a USB 3 PCIe card.
I've tried swapping the positions of the capture & USB cards without any luck and have tried removing the USB card from the MV config to see if there was some sort of interference.
I'm running a dual Xeon E5-2670 setup on an ASRock Rack ep2c602-4l/d16, the system has full IOMMU separation.