[SOLVED] TASK ERROR: IOMMU not present adding two port nic to VM

digitalbots

Member
Dec 7, 2018
23
0
21
43
So i am trying to load pfSense on a VM and do a PCI passthrough for the nic.

I am on 7.x

i followed these steps to enable Dells version of IOMMU
https://www.dell.com/support/kbdoc/...g-sr-iov-technology-on-dell-poweredge-servers

then followed
https://www.youtube.com/watch?v=GoZaMgEgrHw&t=378s

these steps.

rebooted and now all I am getting is DMAR: [DMA read] errors on the screen even though i put in quiet in the grub file?

Also i have 2 raid controllers and now it only sees one of the 2 even though the bios shows both of them up and running.

what did i do wrong? did i miss a setting?

(attached is the output of dmesg | grep -e DMAR -e IOMMU)
 

Attachments

Last edited:
I am having the same issue. I read among the lines of my Dell 3880 Desktop that trusted execution needs to be activated to activate VT-d
 
I am having the same issue. I read among the lines of my Dell 3880 Desktop that trusted execution needs to be activated to activate VT-d
Im not out of the woods yet BUT i fixed most of my issues
(https://pve.proxmox.com/wiki/Pci_passthrough#Enable_the_IOMMU)

as per that doc i needed to add iommu=pt to the grub entry and did this "echo "options vfio_iommu_type1 allow_unsafe_interrupts=1" > /etc/modprobe.d/iommu_unsafe_interrupts.conf"

rebooted the errors are GONE! BUTTTT when i do a PCI passthrough my whole box locks up. still working on it
 
im about at my wits end here guys. Here are some outputs.


root@pve7:~# dmesg | grep -i vfio
[ 10.533481] VFIO - User Level meta-driver version: 0.3

root@pve7:~# find /sys/kernel/iommu_groups/ -type l | grep 01:00
/sys/kernel/iommu_groups/16/devices/0000:01:00.0
/sys/kernel/iommu_groups/16/devices/0000:01:00.1

So it shows the card. I add it to my VM via the pci passthrough turn the VM on and BOOM i lose connect via ssh and https to proxmox and I have to hardwire to it to reboot because all VMS go offline when turn on the VM with the pci passthrough.

also these are the settings for my VM

/etc/pve/qemu-server/105.conf
boot: order=scsi0
cores: 2
hostpci0: 0000:01:00,rombar=1
memory: 5048
name: pfsense
numa: 0
ostype: l26
scsi0: local-lvm:vm-105-disk-0,size=32G
scsihw: virtio-scsi-pci
smbios1: uuid=57eccb20-0773-4443-b0fa-e6f216565a54
sockets: 2
vmgenid: f6849b76-f0e6-4039-978e-35ef837abe5a
 
Why do you feel you need to do pci passthrough? would it not be simpler to use the vm bridge method?
 
OK so i took the card out and found that it has the SAME NAME AS MY ONBOARD LAN. Because I have a poweredge I have 2 onboard Lans already. So it came down to the fact I picked the wrong one.

I would recommend anyone that tries to do this write down your PCI device ID's for your current NIC's so that you dont do what i did!
 
OK so i took the card out and found that it has the SAME NAME AS MY ONBOARD LAN. Because I have a poweredge I have 2 onboard Lans already. So it came down to the fact I picked the wrong one.

I would recommend anyone that tries to do this write down your PCI device ID's for your current NIC's so that you dont do what i did!
How do you find your PCI device IDs for the current NIC