You are not logged in.
Hey all,
I'm in the process of putting the finishing touches on my virtualized windows 10 installation. The windows instance is actually installed on a physical disk and can be booted bare metal also. Despite my efforts following the wiki and other examples I've found, I can't seem to get rid of error 43 on my passed-through nvidia 2080. I'm using remotefx and freerdp to display my vm on my laptop screen.
I added the lines to mask the hypervisor as is outlined in the wiki troubleshooting section and can be seen in my config xml below.
libvirt xml
http://ix.io/1Vq6
grub args
GRUB_CMDLINE_LINUX_DEFAULT=... acpi_osi= video.use_native_backlight=1 nvidia-drm.modeset=1 intel_iommu=on,igfx_off iommu=pt kvm.ignore_msrs=1
* could the nvidia-drm enabling be interfering with the passthrough? I was looking at messing around with the newer nvidia driver to get switching working on my laptop but abandoned that.
IOMMU Group 1:
00:01.0 PCI bridge [0604]: Intel Corporation Xeon E3-1200 v5/E3-1500 v5/6th Gen Core Processor PCIe Controller (x16) [8086:1901] (rev 07)
01:00.0 VGA compatible controller [0300]: NVIDIA Corporation TU104M [GeForce RTX 2080 Mobile] [10de:1e90] (rev a1)
01:00.1 Audio device [0403]: NVIDIA Corporation Device [10de:10f8] (rev a1)
01:00.2 USB controller [0c03]: NVIDIA Corporation Device [10de:1ad8] (rev a1)
01:00.3 Serial bus controller [0c80]: NVIDIA Corporation Device [10de:1ad9] (rev a1)In virt-manager I added host pci device 01:00.0
<hostdev mode="subsystem" type="pci" managed="yes">
<source>
<address domain="0" bus="1" slot="0" function="0"/>
</source>
</hostdev>lspci info showing vfio-pci loaded appropriately.
01:00.0 VGA compatible controller [0300]: NVIDIA Corporation TU104M [GeForce RTX 2080 Mobile] [10de:1e90] (rev a1)
Subsystem: Gigabyte Technology Co., Ltd TU104M [GeForce RTX 2080 Mobile] [1458:75a7]
Kernel driver in use: vfio-pci
Kernel modules: nouveau, nvidia_drm, nvidia
01:00.1 Audio device [0403]: NVIDIA Corporation Device [10de:10f8] (rev a1)
Subsystem: Gigabyte Technology Co., Ltd Device [1458:75a7]
Kernel driver in use: vfio-pci
Kernel modules: snd_hda_intel
01:00.2 USB controller [0c03]: NVIDIA Corporation Device [10de:1ad8] (rev a1)
Subsystem: Gigabyte Technology Co., Ltd Device [1458:75a7]
Kernel driver in use: vfio-pci
Kernel modules: xhci_pci
01:00.3 Serial bus controller [0c80]: NVIDIA Corporation Device [10de:1ad9] (rev a1)
Subsystem: Gigabyte Technology Co., Ltd Device [1458:75a7]
Kernel driver in use: vfio-pci
Kernel modules: i2c_nvidia_gpuI've also blacklisted nvidia and nouveau.
I'm at a loss of where to go next and could really use some input.
Thanks for reading!
Kwertii
**Edited to fix typo in grub cmdline arguments
Last edited by kwertii (2019-09-17 12:40:26)
Offline
Edit: nope, looks like you already have this covered.
Don't know if you've seen this..
https://wiki.archlinux.org/index.php/PC … F/Examples
I saw where you added lines to mask the hypervisor and I know this is esxi but I thought it was interesting..
https://petr.io/en/blog/2017/12/26/runn … ware-esxi/
Is this right ?......GRUB_CMDLINE_LINUX_DEFAULT=... acpi_osi= video.use_native_backlight=1 nvidia-drm.modeset=1 ***intel_iommu=one***,igfx_off iommu=pt kvm.ignore_msrs=1
Last edited by Zod (2019-09-17 02:48:02)
Offline
Whoops, that was a typo on the grub cmdline!
I'm going through more examples now, though I skimmed through a bunch earlier.
In terms of the esxi article, I had read about masking the kernel at the cpu level, but then it apparently changes how the clock is interacted with by the system, I can give it a shot though.
One more thing to note, when I print out the vbios, there doesn't appear to be any entry for uefi. Could that be a part of my issue?
Offline
Maybe this is some windows register bug/error (link, link). Have you tried googling "nvidia error 43" or "nvidia error 43 fixer"? You may try to download and run that "fixer" (nvidia-error43-fixer_112.zip) under win10 vm (before that you may backup vm or do the snapshot).
Last edited by xerxes_ (2019-09-18 17:37:22)
Offline
I have googled those and played around with the system quite a bit. I currently have a post in on the forums for gopupd to resolve some issues when patching my vbios to add uefi (it's not an mxm card, so I don't know if this is actually okay to do).
In relation to those links, they are specifically intended for egpus it seems, though I did give them a read. I also tested the script to no avail.
Still exploring possibilities, but if this uefi rom isn't the right answer, I think I may be out of leads ![]()
**EDIT**
Might be worth noting that after running the error43 fixer script, the gpu reappears as functional to the system, then half a second later it refreshes with the same error 43. Seems like it still may be hypervisor detection related, but I could be crazy.
Last edited by kwertii (2019-09-18 18:16:16)
Offline