Oh my goodness! I've removed rombar=0 and voila! (Strangely even with other 
configurations it's failed before but I must have been doing something else 
wrong. It's not worth questioning it now :P.)

Cheers Alex, you're a lifesaver.

-Matthew

Sent from [ProtonMail](https://protonmail.com), Swiss-based encrypted email.



-------- Original Message --------
Subject: Re: [vfio-users] [help] Pci passthrough - I've probably done something 
really obvious wrong
Local Time: March 19, 2016 3:06 PM
UTC Time: March 19, 2016 3:06 PM
From: [email protected]
To: [email protected]
CC: [email protected]




On Sat, Mar 19, 2016 at 7:38 AM, Matthew Wolffsohn <[email protected]> 
wrote:

Hi Vfio-users,

I'm trying to passthrough a single GPU (originally I was trying to passthrough 
two, but I have removed the other from the system to reduce the chance it could 
be causing a problem) to a standard Windows 8 VM. Unfortunately upon starting 
the VM I get no output to either of my displays and the qemu output is spammed 
with dma errors such as below:

dma: command 78 not supported
dma: command 78 not supported
dma: command 80 not supported
dma: command 80 not supported
dma: command c8 not supported
dma: command c8 not supported
dma: command 80 not supported
dma: command 80 not supported
dma: command c8 not supported
dma: command c8 not supported
dma: command f8 not supported
dma: command c not supported
dma: command 84 not supported
...snip...

These are not coming from vfio, they're generated by the i8257 dma controller 
in QEMU.


Let me run through some of the other specifics of my system. I'm currently 
assigning the GPU to pci-stub using the grub commandline:
[root@UnforseenCircumstances ~]# cat /proc/cmdline
BOOT_IMAGE=/boot/vmlinuz-linux-vfio 
root=UUID=8aa1247d-5b3c-4282-b666-d4000ff5a517 rw quiet splash 
resume=UUID=17129c2f-d322-4662-8bf5-d1ca010968e6 i915.enable_hd_vgaarb=1 
intel_iommu=on pci-stub.ids=10de:1200,10de:0e0c

I am also attempting to use an Intel iGPU has the host's display, for that 
reason I'm trying to use the vga arbitration patch. Strangely however, I still 
seem to have hardware accelleration on my iGPU - Shouldn't that be impossible 
due to the nature of the patch?
After booting I use the following command to rebind the gpu to vfio-pci:
[root@UnforseenCircumstances ~]# cat bindcmd.txt
sudo /usr/bin/vfio-bind 0000:01:00.0 0000:01:00.1

I cannot use OVMF as my GPU is a Geforce 560 Ti, it has not UEFI ROM support so 
SeaBios is my only option.
[root@UnforseenCircumstances ~]# lspci -nnk | grep NVIDIA
01:00.0 VGA compatible controller [0300]: NVIDIA Corporation GF114 [GeForce GTX 
560 Ti] [10de:1200] (rev a1)
01:00.1 Audio device [0403]: NVIDIA Corporation GF114 HDMI Audio Controller 
[10de:0e0c] (rev a1)

Fermi cards don't have a great success record with device assignment for some 
reason.


And finally, as my GPU ROM won't play ball with me I'm using a downloaded ROM 
from TechpowerUp (For the same model!)
[root@UnforseenCircumstances ~]# cat startvm.sh
#!/bin/bash

qemu-system-x86_64 \
-enable-kvm -m 1024 -cpu host,kvm=off \
-smp 4,sockets=1,cores=4,threads=1 \
-device vfio-pci,host=01:00.0,x-vga=on,rombar=0,romfile=/root/gpurom.rom \

rombar=0 is clearly wrong, but I don't know if it's the source of your problems.



-device vfio-pci,host=01:00.1 \
-vga none \
-device virtio-scsi-pci,id=scsi \
-drive file=/home/mwolffsohn/kvm/win8.iso,id=isocd,format=raw,if=none -device 
scsi-cd,drive=isocd \
-drive file=/root/win.img,id=disk,format=qcow2,if=none,cache=writeback -device 
scsi-hd,drive=disk \
-drive file=/home/mwolffsohn/kvm/virtio-win.iso,id=virtiocd,if=none,format=raw 
-device ide-cd,bus=ide.1,drive=virtiocd

Hopefully I'm doing something really obvious wrong because this has been 
confusing me for ages!
Many thanks in advance,

-Matthew Wolffsohn

Sent from [ProtonMail](https://protonmail.com), Swiss-based encrypted email.

_______________________________________________
vfio-users mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/vfio-users
_______________________________________________
vfio-users mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/vfio-users

Reply via email to