Hi all,

Ive been having this issue for a while, I can not pass though my Nvidia Tesla 1070 server to one of my vms. The Tesla server links to the main system (HP DL380 G9) through PCI cards pictured here:  

 

Here are the groupings:

0b:00.0 PCI bridge [0604]: NVIDIA Corporation NF200 PCIe 2.0 switch for Quadro Plex S4 / Tesla S870 / Tesla S1070 / Tesla S2050 [10de:05be] (rev a3)
0c:00.0 PCI bridge [0604]: NVIDIA Corporation NF200 PCIe 2.0 switch for Quadro Plex S4 / Tesla S870 / Tesla S1070 / Tesla S2050 [10de:05be] (rev a3)
0c:01.0 PCI bridge [0604]: NVIDIA Corporation NF200 PCIe 2.0 switch for Quadro Plex S4 / Tesla S870 / Tesla S1070 / Tesla S2050 [10de:05be] (rev a3)
0c:02.0 PCI bridge [0604]: NVIDIA Corporation NF200 PCIe 2.0 switch for Quadro Plex S4 / Tesla S870 / Tesla S1070 / Tesla S2050 [10de:05be] (rev a3)
0c:03.0 PCI bridge [0604]: NVIDIA Corporation NF200 PCIe 2.0 switch for Quadro Plex S4 / Tesla S870 / Tesla S1070 / Tesla S2050 [10de:05be] (rev a3)
0f:00.0 PCI bridge [0604]: NVIDIA Corporation NF200 PCIe 2.0 switch for Quadro Plex S4 / Tesla S870 / Tesla S1070 / Tesla S2050 [10de:05be] (rev a3)
10:00.0 PCI bridge [0604]: NVIDIA Corporation NF200 PCIe 2.0 switch for Quadro Plex S4 / Tesla S870 / Tesla S1070 / Tesla S2050 [10de:05be] (rev a3)
10:01.0 PCI bridge [0604]: NVIDIA Corporation NF200 PCIe 2.0 switch for Quadro Plex S4 / Tesla S870 / Tesla S1070 / Tesla S2050 [10de:05be] (rev a3)
10:02.0 PCI bridge [0604]: NVIDIA Corporation NF200 PCIe 2.0 switch for Quadro Plex S4 / Tesla S870 / Tesla S1070 / Tesla S2050 [10de:05be] (rev a3)
10:03.0 PCI bridge [0604]: NVIDIA Corporation NF200 PCIe 2.0 switch for Quadro Plex S4 / Tesla S870 / Tesla S1070 / Tesla S2050 [10de:05be] (rev a3)
11:00.0 3D controller [0302]: NVIDIA Corporation GT200GL [Tesla C1060 / M1060] [10de:05e7] (rev a1)
13:00.0 3D controller [0302]: NVIDIA Corporation GT200GL [Tesla C1060 / M1060] [10de:05e7] (rev a1)

 

any help would be much appreciated!

I doubt there are many here with the same configuration so might not get too much answers.

It's not really clear what happens, but from the thread title it doesn't look like you have gotten as far as assigning a card to the VM?

If they don't show up in the list of available Graphics cards, you can either stub the cards so they show up in the other PCI devices list or manually add them to the XML. 

 

Add the ID of the cards to your syslinux.cfg (Found by clicking the flash on the main page of unraid) like below.

 

append vfio-pci.ids=10de:05be,10de:05e7 initrd=/bzroot

You need to add this in the menu entry you use to boot unraid (usually the one labeled default) and reboot.

After the reboot you should have the cards available in the other PCI  devices list in the VM templates. 

further down the line when you try to make a vm, you may also have to allow unsafe interrupt in the syslinux.cfg.

Thanks for the reply! im currently away from the system until Tuesday (Australia time) so ill give it a shot then!