Fixing 'Error: start failed: QEMU exited with code 1'

[jriley]

Alright so there's plenty of other posts about this issue but I was hoping to get some insight into my particular situation, more context here, where I had to force intel_iommu on. I can now add the PCI passthrough device and try to start the VM, but it throws 'Error: start failed: QEMU exited with code 1'. On the proxmox box I get
genirq: Flags mismatch irq 16. 00000000 (vfio-intx(0000:06:04.0)) vs 00000080 (ehci_hcd:usb1).

Further inspection shows the ehci_hcd:usb1 is the kernel module for the integrated Intel USB chipset. The NIC is trying to use the chipsets interrupt handler and they're preventing the VM from starting.

So far, my only fix is to rmmod the kernel module for ehci_hcd and ehci_pci but then it disables all USB functionality on the physical machine. This isn't a huge deal, I can always just SSH into it, but then the modules are reloaded on reboot.

Any suggestions?

 

[needle]

Remarkably, I have exactly the same problem.....

 

[jriley]

Remarkably, I have exactly the same problem.....

you can just unload the module with rmmod. You can find out the module name and any other modules that depends on it by using lsmod

keep in mind whatever the module is for, if you unload it, it obviously removes functionality for that device. In my cause, none of the USB ports on the device work


I actually set up a job in root’s crontab file https://unix.stackexchange.com/ques...n-rmmod-not-executing-at-reboot/734902#734902

Apparently the module can be blacklisted. I added the module names to /etc/modprobe.d/pve-blacklist.conf which didn’t work, because someone else on stackexchange said “if you have an initramfs you'd need to update that too”. Haven’t tried it out yet though