Manjaro Gpu Passthrough - Host Hard Crashing

Hello again L1T, Long time.

So to get right into it. I have a Fresh install Manjaro system that I have been trying to set up with GPU pass-through to a windows VM. At a specific point the host becomes unstable and hard crashes.
The specific point is adding a sata drive to the VM in the virt-manager “New Virtual Machine Gui”
What I mean by hard crash is, Sudden Black Screens… about 2 seconds go by and the system is posting. Exactly like someone hitting the reset button. This is consistent across two fresh installs of Manjaro on the same system

Specs:
Manjaro w/KDE
Ryzen 2700
Asus Strix x470-f Gaming
Corsair 3000MHz Memory Modules
XFX Vega 64 (host)
EVGA RTX 2060 KO (guest)
WD Black NVME 500GB SSD (host)
Adata 2tb Sata SSD (attempting to pass this to a guest)
WD Blue m.2 1TB SSE(empty)
2TB Seagate spinning rust. (empty)
*added those last two only because they are installed right now. The Seagate is going to be removed and the WD blue will be for manjaro.

So the guide I’ve been following is the Arch Wiki Guide.
https://wiki.archlinux.org/title/PCI_passthrough_via_OVMF#Prerequisites

The only things I’ve done that is pass-through related is verify IOMMU is enabled and the Cards are in separate groups and the guest card is in its own group.
Then I isolated the guest card with vfio-pci.ids in /etc/default/grub (the crashkernel is from me TRYING to configure kdump last night. More on that below.)

GRUB_CMDLINE_LINUX_DEFAULT="crashkernel=256M quiet apparmor=1 security=apparmor udev.log_priority=3 amd_iommu=on iommu=pt vfio-pci.ids=10de:1e89,10de:10f8,10de:1ad8,10de:1ad9"

Last, loading vfio-pci early and verified that they are running vfio-pci driver.

MODULES="vfio_pci vfio vfio_iommu_type1 vfio_virqfd"
HOOKS="base udev autodetect modconf block keyboard keymap consolefont filesystems fsck"
$ lspci -nnk
b:00.0 VGA compatible controller [0300]: NVIDIA Corporation TU104 [GeForce RTX 2060] [10de:1e89] (rev a1)
	Subsystem: eVga.com. Corp. TU104 [GeForce RTX 2060] [3842:2068]
	Kernel driver in use: vfio-pci
	Kernel modules: nvidiafb, nouveau
0b:00.1 Audio device [0403]: NVIDIA Corporation TU104 HD Audio Controller [10de:10f8] (rev a1)
	Subsystem: eVga.com. Corp. TU104 HD Audio Controller [3842:2068]
	Kernel driver in use: vfio-pci
	Kernel modules: snd_hda_intel
0b:00.2 USB controller [0c03]: NVIDIA Corporation TU104 USB 3.1 Host Controller [10de:1ad8] (rev a1)
	Subsystem: eVga.com. Corp. TU104 USB 3.1 Host Controller [3842:2068]
	Kernel driver in use: vfio-pci
0b:00.3 Serial bus controller [0c80]: NVIDIA Corporation TU104 USB Type-C UCSI Controller [10de:1ad9] (rev a1)
	Subsystem: eVga.com. Corp. TU104 USB Type-C UCSI Controller [3842:2068]
	Kernel driver in use: vfio-pci
	Kernel modules: i2c_nvidia_gpu

The point at which the system becomes unstable is when I try to add the Sata drive to the VM. The way I did it in the past on other distros, is creating a storage pool with a source to /dev. Then Selecting the drive, in this case /dev/sda. The bus type is VirtIO.

On Manjaro, after selecting the disk and clicking OK… the system becomes unstable and crashes. Not building the VM or anything, Still configuring it with the GUI. Then after reboot. Will only be up for about 30-60 seconds after login and then hard crash again.

I was able to get systemd logs the manjaro-chroot command from a live environment. (This is new to me in practice). The only thing that sticks out to me, after digging through the log, are these kwin_x11 lines.

Nov 17 17:41:25 Hodor-Linux kwin_x11[1710]: qt.qpa.xcb: QXcbConnection: XCB error: 3 (BadWindow), sequence: 4243, resource id: 33554626, major code: 18 (ChangeProperty), minor code: 0
Nov 17 17:41:25 Hodor-Linux kwin_x11[1710]: qt.qpa.xcb: QXcbConnection: XCB error: 3 (BadWindow), sequence: 4248, resource id: 33554627, major code: 18 (ChangeProperty), minor code: 0
Nov 17 17:41:28 Hodor-Linux kwin_x11[1710]: kwin_core: XCB error: 152 (BadDamage), sequence: 8397, resource id: 12583370, major code: 143 (DAMAGE), minor code: 2 (Destroy)

I’ll be honest here and say I have never diagnosed a problem like this before, but I would love to learn how. As mentioned above, I attempted to set up kdump on the file system from a live usb last night, but I didn’t fully understand how to set it up, And on top of that couldn’t figure out how to do it from a live usb. I also don’t even know if its possible to set up like that.

Hi. It sounds just like my problem: `libvirtd` reboots the system after 1 min. Two reddit threads say it’s because of the /dev allocated. For me, it reboots ~1 min after starting the libvirtd service. By any chance, are you also using the dividebyzer0 tutorial/walthrough/page/whatever?