Xen project Mailing List

On Wed, Feb 19, 2014 at 5:40 PM, Gordan Bobic <gordan@xxxxxxxxxx> wrote:

That was a long time ago. IIRC it required applying patches to Xen that ensured vBAR=pBAR and shortly after that was demonstrated to work, Nvidia modified the driver whitelist Quadro cards for virtualized operation. It is purely down to the device ID presented to the driver inside the VM. Modify the card's device ID, and the driver will initialize it even if it detects it is running virtualized.

Somebody mentioned on the list here that they got an unmodified card working by applying a patch that changes the behaviour the driver uses to detect whether it is in a VM, but I have not seen anything about it since.

Cards up to and including 4xx series can be BIOS-modded to change the device ID via the strap. Later cards include two sets of ID registers, one set by the hard-strap before the BIOS initializes, and one subject to being modified by the soft strap. This seems to have been designed specifically for Grid cards. The driver initializes the card based on the hard strap (to decide whether it will initialize it or not). Grid cards can emulate multiple different cards, for example a Quadro or a Tesla, which is effectively done by the soft strap. As far as I can work out, this is done by executing a BIOS payload in the VM to set the soft strap and perform suitable initialization, so a single Grid GPU can present different personalities to different VMs.

For your 285 it doesn't matter - just find a Quadro based on the same model GPU, get it's device ID, find where the device ID strap bits are for that generation of GPUs, re-flash the strap accordingly using nvflash and enjoy VGA passthrough. :)

It is probably only a matter of time before Nvidia do something to prevent this sort of thing in the driver, but we'll worry about it later. Older cards will always work with older drivers (i.e. anything Nvidia do in future drivers doesn't affect older cards since you don't need the new driver), and the current top of the like GTX780Ti is very easily modifiable into Quadro K6000 using a single 1206 33K-ohm resistor. So it would appear that VGA passthrough on the cheap is quite safe until at least the next generation of GPUs comes out in a year or two.

On 02/19/2014 10:29 PM, Kyle Davis wrote:

I'd seen people saying the GTX 285 worked for them with no changes to
the drivers, but I'll see if I can get this to work.

thanks,
-Kyle

On Wed, Feb 19, 2014 at 4:12 PM, Gordan Bobic <gordan@xxxxxxxxxx
<mailto:gordan@xxxxxxxxxx>> wrote:

GeForce card won't work - only Quadro/Tesla/Grid cards are
whitelisted in the driver to make them work. You can modify your
GTX285's device ID by changing the straps in it's BIOS to make it
into a Quadro.

This may be of interest:
http://www.altechnative.net/__2013/11/25/virtualized-gaming-__nvidia-cards-part-3-how-to-__modify-a-fermi-based-geforce-__into-a-quadro-geforce-__gts450gtx470gtx480-to-quadro-__200050006000/

<http://www.altechnative.net/2013/11/25/virtualized-gaming-nvidia-cards-part-3-how-to-modify-a-fermi-based-geforce-into-a-quadro-geforce-gts450gtx470gtx480-to-quadro-200050006000/>

The straps on the 2xx series are a little different (check the
nouveau strap documentation), but the general method will be similar.

Gordan

On 02/19/2014 08:54 PM, Kyle Davis wrote:

Hello everyone,

I'd posted this over at the new user forum on xenproject, and was
directed here. Hopefully I'm not doing something stupid by
copying and
pasting my post from there, but here goes:

I have a core i7-4770 on an ASRock Z87 Extreme 4 motherboard. I
have a
PCI-based USB hub attached to deal with the fact that all the
USB ports
on the motherboard are on the same controller for passthrough
purposes.
I am using Debian Jessie with Xen 4.3 and a kernel based on
3.12.9. The
card I'm trying to passthrough is an NVIDIA GTX 285 (I've found
folks
who say it worked fine for them, including on the boards here).

I created a kernel included the flags suggested by the xen wiki,
switched my toolstack over to xl (was on xm, and I'm unsure
why), and
followed the wiki's guides for setting up xen and VGA passthrough.

I created a Windows 8 (HVM) guest that worked fine, but slowly,
before
beginning the VGA passthrough guide. Changed my windows.cfg file
to the
following:

builder='hvm'
bootloader ="/usr/bin/pygrub"
memory =8192
vcpus=4
name ="windows8"
vif =['bridge=xenbr0']
disk
=['phy:/dev/vg0/windows,hda,w'__,'file:/home/kyle/Desktop/__Windows.iso,hdc:c$

device_model_version ="qemu-xen-traditional"
device_model_override ="/usr/bin/qemu-x86_64"
acpi =1
pci=['01:00.0','04:00.0']
pci_permissive=1
boot="c"
sdl=0
serial='pty'
vnc=1
vnclisten=""
vncpasswd=""

When the config worked, pci_permissive was commented out, as was
the pci
tag. The device override was set to qemu rather than
qemu-x86_64, and
the model version was qemu-xen rather than traditional. I've tried
flipping things just to see if I could make it work to no avail.

The error I get upon running the config is:

libxl: error: libxl_dm.c:1275:device_model___spawn_outcome:

domain2 devicemodel: spawn failed (rc=-3)
libxl: error:
libxl_create.c:1075:domcreate___devmodel_started: device
model did notstart: -3

libxl: error: libxl_dm.c:1306:libxl____destroy_device_model:

Device Model already exited

When I checked the log under /var/log/xen, I originally had
-gfx_passthru: invalid option. After playing with the config
file, I now
have a hot mess:

usage: qemu-x86_64 [options] program [arguments...]
Linux CPU emulator (compiledfor x86_64 emulation)

Optionsand associated environment variables:

Argument Env-variable Description
-h print this help
-g port QEMU_GDB wait gdb connectionto 'port'
-L path QEMU_LD_PREFIXset the elf interpreter prefixto
'path'
-s size QEMU_STACK_SIZEset the stack sizeto 'size' bytes
-cpu model QEMU_CPU select CPU (-cpu helpfor list)

-E var=value QEMU_SET_ENV sets targets environment
variable (seebelow)
-U var QEMU_UNSET_ENV unsets targets environment
variable (seebelow)
-0 argv0 QEMU_ARGV0 forces target process argv[0]to
be 'argv0'
-r uname QEMU_UNAMEset qemu uname releasestring to 'uname'
-B address QEMU_GUEST_BASEset guest_base addressto 'address'

-R size QEMU_RESERVED_VA reserve 'size' bytesfor guest
virtual address $
-ditem[,...] QEMU_LOG enable loggingof specified items
(use '-d help$
-D logfile QEMU_LOG_FILENAMEwrite logsto 'logfile' (default
stderr)

-p pagesize QEMU_PAGESIZEset the host page sizeto 'pagesize'
-singlestep QEMU_SINGLESTEPrun in singlestep mode
-strace QEMU_STRACElog system calls
-version QEMU_VERSION displayversion informationand
exit

Defaults:
QEMU_LD_PREFIX = /etc/qemu-binfmt/x86_64
QEMU_STACK_SIZE =8388608 byte

You can use -Eand -U optionsor the QEMU_SET_ENVand
QEMU_UNSET_ENV environment variablesto set and unset
environment variablesfor the target process.
Itis possibleto provide several variablesby separating them
by commasin getsubopt(3) style. Additionallyit is possibleto
providethe -Eand -U options multipletimes.
The following lines are equivalent:
-E var1=val2 -E var2=val2 -U LD_PRELOAD -U LD_DEBUG
-E var1=val2,var2=val2 -U LD_PRELOAD,LD_DEBUG
QEMU_SET_ENV=var1=val2,var2=__val2
QEMU_UNSET_ENV=LD_PRELOAD,LD___DEBUG

Notethat if you provide several changesto a single variable
the last change will stayin effect.

Does anyone have some insight on this? I feel like I'm very close to
getting my passthrough working, and that would be really awesome.

Additionally, I tried to go graphically after being unsuccessful
with my
windows.cfg file, and tried the virtual machine manager. When I
start
it, however, I get this:

Unableto connectto libvirt.

Failedto connect socketto '/var/run/libvirt/libvirt-__sock':

No suchfile or directory

Verifythat:
- A Xen host kernel was booted
- The Xen service has been started

and under details :

Unableto connectto libvirt.

Failedto connect socketto '/var/run/libvirt/libvirt-__sock':

No suchfile or directory

Verifythat:
- A Xen host kernel was booted
- The Xen service has been started

Libvirt URIis: xen:///

Traceback (most recent calllast):
File"/usr/share/virt-manager/__virtManager/connection.py",

line1027,in _open_thread
self.vmm = self._try_open()
File"/usr/share/virt-manager/__virtManager/connection.py",
line1009,in _try_open

flags)
File"/usr/lib/python2.7/dist-__packages/libvirt.py",
line105,in openAuth
if retis None:raise libvirtError('__virConnectOpenAuth()

failed')
libvirtError: Failedto connect socketto
'/var/run/libvirt/libvirt-__sock': No suchfile or directory

The virtual machine manager does not, thereby, connect to localhost

I guess I should also mention that Windows 8 didn't always shut down
properly, but it wasn't reporting errors and things like the GPLPV
drivers seemed to install correctly.

Thanks for any insight,
-Kyle

_________________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxx <mailto:Xen-users@xxxxxxxxxxorg>
http://lists.xen.org/xen-users

Re: [Xen-users] Issue with Win8 guest VGA Passthrough