[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] VGA passthrough is not working



Hi!

Thanks for your answer!

1. Version Xen:
(XEN) Xen version 4.4.1-pre (root@xxxxxxxxxxxxx) (gcc (Debian 4.8.3-2) 4.8.3) debug=n Tue Jun 10 08:25:03 CDT 2014
2. Kernel:
Linux hv2 3.14-1-amd64 #1 SMP Debian 3.14.4-1 (2014-05-13) x86_64 GNU/Linux


Variant using :

A). Config file my VM:
device_model_version =Â "qemu-xen-traditional"
device_model_override = "/usr/local/lib/xen/bin/qemu-dm"
A.1) Trouble in VM(use conf file VM 3):
ÂÂÂ VM is not starting!
register_real_device: Real physical device 03:00.0 registered successfuly!
IRQ type = INTx
pt_pci_write_config: [00:05:0] Warning: Guest attempt to set address to unused Base Address Register. [Offset:14h][Length:4]
pt_pci_write_config: Internal error: Invalid write emulation return value[-1]. I/O emulator exit.

OR


B) Config my VM:
device_model_version =Â "qemu-xen"
B.1)
My VM is starting! %) But It have a problem...
Please, look at attach file.

qemu-dm file is have size 5.1Mb. Are you sure attaching file in here?

nvidia card into VM:Â Region 1: Memory at 100000000 (64-bit, prefetchable) [size=4G]
nvidia card into DOM0: Region 1: Memory at 5800000000 (64-bit, prefetchable) [size=16G]

Why? Where 12G was lost?



I try use memory for VM == 1G... there's no difference at all

10.06.2014 16:58, Konrad Rzeszutek Wilk ÐÐÑÐÑ:
On Tue, Jun 10, 2014 at 01:10:01PM +0400, Maxim Gordeev wrote:
Sorry, Xen don`t map memory BAR1 when I use "qemu-dm" or
"qemu-xen-traditional"...
Could you attach your qemu-dm files, the lspci -vvv from dom0 and
from the guest.

Does it work if you use less memory? Say mem=2G?

10.06.2014 10:56, Maxim Gordeev ÐÐÑÐÑ:
Hi!

Thanks for your answer!

I tried starting my VM with device_model_version = "qemu-xen.
VM is working, but I don`t see NVIDIA in list pci devices...

This is output message of command "xl create:"
Hi!

Thanks for your answer!

I tried starting my VM with device_model_version = "qemu-xen.
VM is working, but I don`t see NVIDIA in list pci devices...

This is output message of command xl create:
dmesg:
[81234.316270] xen_pciback: vpci: 0000:03:00.0: assign to virtual slot 0
It is All.

Maybe I will try use this parameter in Xen 4.3?

09.06.2014 16:48, Gordan Bobic ÐÐÑÐÑ:
Yes, I think so. I have a sneaky suspicion that qemu cannot handle
BARs over a certain size. Have you tried with
device_model_version = "qemu-xen"
instead of qemu-xen-traditional ?

On 06/09/2014 01:42 PM, Maxim Gordeev wrote:
Do you understand me? :)

I can give you logs!
09.06.2014 16:40, Maxim Gordeev ÐÐÑÐÑ:
no, no.
With 4.3 is bad work! With 4.4 is do not starting.
I see into lspci: BAR1 is not possible assignment. With 4.3 can not
maping BAR1.
Thanks!

09.06.2014 16:29, Gordan Bobic ÐÐÑÐÑ:
On 06/09/2014 01:21 PM, Maxim Gordeev wrote:
16Gb this is BAR1.
This is example my VM config:
/builder='hvm'//
//memory = 16000//
//name = "xen-hvm-cheap1"//
//vcpus=8//
//acpi=0//
//disk = [ //
//// 'file:/opt/VM/cheap1_glinux-1.img,hda,w',//
// 'file:/opt/VM/cheap1_glinux-2.img,hdb,w'//
// ]//
// xen_platform_pci = 1//
////
////
////
//device_model_version = 'qemu-xen-traditional'//
//boot="c"//
//vnc=1//
//vnclisten="0.0.0.0"//
//vncdisplay=1//
//vncunused=1//
//vncpasswd=''//
//stdvga=0//
//serial='pty'//
//monitor=1//
//gfx_passthru=0//
//pci=['03:00.0' ] /
Xen 4.3 start this VM, but Xen 4.4 is not starting VM.
So you are saying that this works fine with 4.3 but breaks with 4.4?

Gordan

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel

Attachment: dmesg.log
Description: Text Data

Attachment: lspci_nvidia_DOM0.txt
Description: Text document

Attachment: lspci_nvidia_intoVM.txt
Description: Text document

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.