Qubes 4.0 Freezing When Adding VGA PCI Device to HVM

running QubesOS 4.0 on a Lenovo T430 w/ dGPU 16GB RAM installed.
recently built Heads with dGPU initalization (NVS 5400M on motherboard)
VGA controller appears in dom0 lspci list and in PCI Device list for VM’s.
i have attempted to pass it through to multiple HVM’s with no success-- it freezes the entire OS every time as soon as i click Start qube.
i would ideally like to pass the GPU through to a Fedora 34 HVM running kernel 5.15.11-100 (currently default when ‘none’ is selected), but i could be convinced otherwise. i just want to be able to run programs that i know to be available via rpm repos. If it is easier in Debian 11, or another distro i can build with the Qubes Builder i can make do, but when i tried it failed the same way. i’ve gone through several guides including a Debian HVM guide that suggested Nvidia GPU’s wouldn’t work, only AMD… and no matter what, my machine freezes. in the latest attempt with Fedora 34, I loaded 10000MB RAM straight into the HVM, not included in memory balancing.

i haven’t done this yet-- but i intend to run a test to make sure this isn’t a firmware issue by installing a fresh linux distro on a separate drive to make sure this is a qubes specific issue, but:

am i missing something? do i have to select/build a custom kernel to be compatible with this particular GPU and then use a legacy driver? or something else?

just checking back in to say i did not perform the correct steps initially. i missed the part where i needed to hide the PCI device from dom0 to be able to add it to an appVM.
This guide helped me do what I wanted to do:
Tested in 4.1rc3 with decent results, but doing so removed AEM capabilities in firmware. Holding off on GPU passthrough for now until there is more proper implementation :]