I started with Fedora 19 (KDE) and have installed Xen 4.2 on it. I
followed the install guide, with the exception of the network bridge, as my
current wireless card doesn''t support it (I''m working on
getting a wired
connection).
My (relevant) hardware setup:
AMD FX6300
ASUS m5a99x evo r2.0 (IOMMU & Virtualization enabled)
Radeon 6670 (Dedicated Dom0)
Radeon 7850 (Passthrough to windows)
As the title suggests, I''ve been trying to install Windows 8, and
I''ve run
into the following issue:
- Install completes successfully (Until reboot)
- Install process completes setting up devices
- Install process fails (hangs) after that with no graphical output.
The VM will just hang there, and if I don''t kill it in about 1 minute,
will
bring down my entire system. (Including Dom0)
Here is my configuration:
# Name for the HVM
name = "win8"
# Use HVM
builder = "hvm"
# Create 4 cpus on physical cores 2-5
vcpus = 4
cpus = "2-5"
# 8GB memory
memory = 8192
# Use lvm for windows root, and pass in cdrom for install
disk =
[''phy:/dev/fedora_hypervisor/windows,hda,w'',
''phy:/dev/sr0,hdc:cdrom,r'']
# Uncomment and configure once br0 is created
#vif
# Boot cd rom
boot = "d" #"c"
# Use vnc for display
vnc = 1
# No vnc password
vncpasswd = ""
# Emulate VGA
vga = "stdvga"
# Provide graphics device
nographic = 1
# Emulate usb bus
usb = 1
# Use absolute coordinates for vnc
usbdevice = ''tablet''
# Use graphics card (Radeon 7850) as secondary / optional
gfx_passthrough = 0
# Pass through various PCI devices
# 2:00.0 - Radeon 7850
# 2:00.1 - Radeon 7850 (Sound)
# 0:12.0 - USB Devices (Mouse / Keyboard)
pci = [''02:00.0'',
''02:00.1'', ''00:12.0'']
# Allow windows to see the local time
localtime = 1
# Pass some PCI information to windows
viridian=1
I''ve also tried using irqs=[18], but that did nothing...
Afterwards, I get a notification informing me that IRQ 18 was disabled..
Message from syslogd@Hypervisor at Oct 21 20:28:19 ...
kernel:[ 1312.294827] Disabling IRQ #18
xl dmesg output:
ss = 0xffffffc0
(XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault
address = 0xffffffc0
(XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault
address = 0xffffffc0
(XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault
address = 0xffffffc0
(XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault
address = 0xffffffc0
(XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault
address = 0xffffffc0
(XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault
address = 0xffffffc0
(XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault
address = 0xffffffc0
(XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault
address = 0xffffffc0
(XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault
address = 0xffffffc0
(XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault
address = 0xffffffc0
[truncated output, only first 10 iterations are shown...]
What is the best way to go about solving this problem?
As always just ask for more info,
-John
_______________________________________________
Xen-users mailing list
Xen-users@lists.xen.org
http://lists.xen.org/xen-users
John, I''m not an expert, but the only thing I see in your config that I questioned is the passing of keyboard and mouse to the VM. You would have to have a separate mouse and keyboard and block Dom0 from taking control of them with a boot parameter then pass them to the DomU. Where you are using VNC this doesn''t make a lot of sense to me. VNC will emulate the keyboard and mouse for you that is why the tablet line is in there to VNC tracks correctly. If you need more info on the PCI passthrough the wiki has some great information. Hope this helps Shane On Mon, Oct 21, 2013 at 7:53 PM, John Wargo <wargo.john25@gmail.com> wrote:> I started with Fedora 19 (KDE) and have installed Xen 4.2 on it. I followed > the install guide, with the exception of the network bridge, as my current > wireless card doesn''t support it (I''m working on getting a wired > connection). > > My (relevant) hardware setup: > AMD FX6300 > ASUS m5a99x evo r2.0 (IOMMU & Virtualization enabled) > Radeon 6670 (Dedicated Dom0) > Radeon 7850 (Passthrough to windows) > > As the title suggests, I''ve been trying to install Windows 8, and I''ve run > into the following issue: > > - Install completes successfully (Until reboot) > - Install process completes setting up devices > - Install process fails (hangs) after that with no graphical output. > > The VM will just hang there, and if I don''t kill it in about 1 minute, will > bring down my entire system. (Including Dom0) > > Here is my configuration: > # Name for the HVM > name = "win8" > # Use HVM > builder = "hvm" > # Create 4 cpus on physical cores 2-5 > vcpus = 4 > cpus = "2-5" > # 8GB memory > memory = 8192 > # Use lvm for windows root, and pass in cdrom for install > disk = [''phy:/dev/fedora_hypervisor/windows,hda,w'', > ''phy:/dev/sr0,hdc:cdrom,r''] > # Uncomment and configure once br0 is created > #vif > # Boot cd rom > boot = "d" #"c" > # Use vnc for display > vnc = 1 > # No vnc password > vncpasswd = "" > # Emulate VGA > vga = "stdvga" > # Provide graphics device > nographic = 1 > # Emulate usb bus > usb = 1 > # Use absolute coordinates for vnc > usbdevice = ''tablet'' > # Use graphics card (Radeon 7850) as secondary / optional > gfx_passthrough = 0 > # Pass through various PCI devices > # 2:00.0 - Radeon 7850 > # 2:00.1 - Radeon 7850 (Sound) > # 0:12.0 - USB Devices (Mouse / Keyboard) > pci = [''02:00.0'', ''02:00.1'', ''00:12.0''] > # Allow windows to see the local time > localtime = 1 > # Pass some PCI information to windows > viridian=1 > I''ve also tried using irqs=[18], but that did nothing... > > Afterwards, I get a notification informing me that IRQ 18 was disabled.. > Message from syslogd@Hypervisor at Oct 21 20:28:19 ... > kernel:[ 1312.294827] Disabling IRQ #18 > > xl dmesg output: > ss = 0xffffffc0 > (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault > address = 0xffffffc0 > (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault > address = 0xffffffc0 > (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault > address = 0xffffffc0 > (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault > address = 0xffffffc0 > (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault > address = 0xffffffc0 > (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault > address = 0xffffffc0 > (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault > address = 0xffffffc0 > (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault > address = 0xffffffc0 > (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault > address = 0xffffffc0 > (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault > address = 0xffffffc0 > [truncated output, only first 10 iterations are shown...] > > What is the best way to go about solving this problem? > As always just ask for more info, > -John > > _______________________________________________ > Xen-users mailing list > Xen-users@lists.xen.org > http://lists.xen.org/xen-users-- Shane D. Johnson IT Administrator Rasmussen Equipment
On 10/22/2013 08:11 AM, Shane Johnson wrote:> John, > I''m not an expert, but the only thing I see in your config that I > questioned is the passing of keyboard and mouse to the VM. You would > have to have a separate mouse and keyboard and block Dom0 from taking > control of them with a boot parameter then pass them to the DomU. > Where you are using VNC this doesn''t make a lot of sense to me. VNC > will emulate the keyboard and mouse for you that is why the tablet > line is in there to VNC tracks correctly. If you need more info on > the PCI passthrough the wiki has some great information. > > Hope this helps > > Shane > > On Mon, Oct 21, 2013 at 7:53 PM, John Wargo <wargo.john25@gmail.com> wrote: >> I started with Fedora 19 (KDE) and have installed Xen 4.2 on it. I followed >> the install guide, with the exception of the network bridge, as my current >> wireless card doesn''t support it (I''m working on getting a wired >> connection). >> >> My (relevant) hardware setup: >> AMD FX6300 >> ASUS m5a99x evo r2.0 (IOMMU & Virtualization enabled) >> Radeon 6670 (Dedicated Dom0) >> Radeon 7850 (Passthrough to windows) >> >> As the title suggests, I''ve been trying to install Windows 8, and I''ve run >> into the following issue: >> >> - Install completes successfully (Until reboot) >> - Install process completes setting up devices >> - Install process fails (hangs) after that with no graphical output. >> >> The VM will just hang there, and if I don''t kill it in about 1 minute, will >> bring down my entire system. (Including Dom0) >> >> Here is my configuration: >> # Name for the HVM >> name = "win8" >> # Use HVM >> builder = "hvm" >> # Create 4 cpus on physical cores 2-5 >> vcpus = 4 >> cpus = "2-5" >> # 8GB memory >> memory = 8192 >> # Use lvm for windows root, and pass in cdrom for install >> disk = [''phy:/dev/fedora_hypervisor/windows,hda,w'', >> ''phy:/dev/sr0,hdc:cdrom,r''] >> # Uncomment and configure once br0 is created >> #vif >> # Boot cd rom >> boot = "d" #"c" >> # Use vnc for display >> vnc = 1 >> # No vnc password >> vncpasswd = "" >> # Emulate VGA >> vga = "stdvga" >> # Provide graphics device >> nographic = 1 >> # Emulate usb bus >> usb = 1 >> # Use absolute coordinates for vnc >> usbdevice = ''tablet'' >> # Use graphics card (Radeon 7850) as secondary / optional >> gfx_passthrough = 0 >> # Pass through various PCI devices >> # 2:00.0 - Radeon 7850 >> # 2:00.1 - Radeon 7850 (Sound) >> # 0:12.0 - USB Devices (Mouse / Keyboard) >> pci = [''02:00.0'', ''02:00.1'', ''00:12.0''] >> # Allow windows to see the local time >> localtime = 1 >> # Pass some PCI information to windows >> viridian=1 >> I''ve also tried using irqs=[18], but that did nothing... >> >> Afterwards, I get a notification informing me that IRQ 18 was disabled.. >> Message from syslogd@Hypervisor at Oct 21 20:28:19 ... >> kernel:[ 1312.294827] Disabling IRQ #18 >> >> xl dmesg output: >> ss = 0xffffffc0 >> (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault >> address = 0xffffffc0 >> (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault >> address = 0xffffffc0 >> (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault >> address = 0xffffffc0 >> (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault >> address = 0xffffffc0 >> (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault >> address = 0xffffffc0 >> (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault >> address = 0xffffffc0 >> (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault >> address = 0xffffffc0 >> (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault >> address = 0xffffffc0 >> (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault >> address = 0xffffffc0 >> (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault >> address = 0xffffffc0 >> [truncated output, only first 10 iterations are shown...] >> >> What is the best way to go about solving this problem? >> As always just ask for more info, >> -John >> >> _______________________________________________ >> Xen-users mailing list >> Xen-users@lists.xen.org >> http://lists.xen.org/xen-users > >It seemed to fix the irq error, however I''m still getting the IO_PAGE_FAULT errors. I''ll go through the debugging page on the wiki, and see what I can''t dig up.. -John
<snip>>>> (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault >>> address = 0xffffffc0 >>> [truncated output, only first 10 iterations are shown...] >>> >>> What is the best way to go about solving this problem? >>> As always just ask for more info, >>> -John >>> >>> _______________________________________________ >>> Xen-users mailing list >>> Xen-users@lists.xen.org >>> http://lists.xen.org/xen-users >> >> >> > It seemed to fix the irq error, however I''m still getting the IO_PAGE_FAULT > errors. > I''ll go through the debugging page on the wiki, and see what I can''t dig > up.. > -JohnJohn, I know when I have need help before, I was asked to include the verbose output of the create command and the log files mentioned in the submitting a bug section. From looking at those logs, those on here who know a heck of a lot more that I do where able to spot what my problem was. If you can''t find it with the steps on the debugging wiki that may be the next best step. Also some details on your hardware and a file or screen print of your xl list command would also probably be beneficial. Good luck Shane -- Shane D. Johnson IT Administrator Rasmussen Equipment
On 10/22/2013 02:51 PM, Shane Johnson wrote:> <snip> >>>> (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault >>>> address = 0xffffffc0 >>>> [truncated output, only first 10 iterations are shown...] >>>> >>>> What is the best way to go about solving this problem? >>>> As always just ask for more info, >>>> -John >>>> >>>> _______________________________________________ >>>> Xen-users mailing list >>>> Xen-users@lists.xen.org >>>> http://lists.xen.org/xen-users >>> >>> >> It seemed to fix the irq error, however I''m still getting the IO_PAGE_FAULT >> errors. >> I''ll go through the debugging page on the wiki, and see what I can''t dig >> up.. >> -John > John, > I know when I have need help before, I was asked to include the > verbose output of the create command and the log files mentioned in > the submitting a bug section. From looking at those logs, those on > here who know a heck of a lot more that I do where able to spot what > my problem was. If you can''t find it with the steps on the debugging > wiki that may be the next best step. Also some details on your > hardware and a file or screen print of your xl list command would also > probably be beneficial. > > Good luck > Shane >After some debugging, here is what I''ve been able to come up with, if there''s an issue with attachments, just let me know how to properly post the logs. Included in the zip : create.xen - The result of "xl create -d ./win8.cfg" dmesg-boot.xen - The result of "xl dmesg" immediately after boot dmesg-win8.xen - The result of "xl dmesg" after failing to install windows 8 grub - My grub configuration file (/etc/default/grub) info.xen - The result of "xl info" win8.cfg - The configuration file I''m using for windows 8 -John _______________________________________________ Xen-users mailing list Xen-users@lists.xen.org http://lists.xen.org/xen-users
On Tue, Oct 22, 2013 at 4:30 PM, John Wargo <wargo.john25@gmail.com> wrote:> On 10/22/2013 02:51 PM, Shane Johnson wrote: > > <snip> > > (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault > address = 0xffffffc0 > [truncated output, only first 10 iterations are shown...] > > What is the best way to go about solving this problem? > As always just ask for more info, > -John > > _______________________________________________ > Xen-users mailing list > Xen-users@lists.xen.org > http://lists.xen.org/xen-users > > > It seemed to fix the irq error, however I''m still getting the IO_PAGE_FAULT > errors. > I''ll go through the debugging page on the wiki, and see what I can''t dig > up.. > -John > > John, > I know when I have need help before, I was asked to include the > verbose output of the create command and the log files mentioned in > the submitting a bug section. From looking at those logs, those on > here who know a heck of a lot more that I do where able to spot what > my problem was. If you can''t find it with the steps on the debugging > wiki that may be the next best step. Also some details on your > hardware and a file or screen print of your xl list command would also > probably be beneficial. > > Good luck > Shane > > > After some debugging, here is what I''ve been able to come up with, if > there''s an issue with attachments, just let me know how to properly post the > logs. > > Included in the zip : > create.xen - The result of "xl create -d ./win8.cfg" > dmesg-boot.xen - The result of "xl dmesg" immediately after boot > dmesg-win8.xen - The result of "xl dmesg" after failing to install windows 8 > grub - My grub configuration file (/etc/default/grub) > info.xen - The result of "xl info" > win8.cfg - The configuration file I''m using for windows 8 > > -JohnSorry John, Memory isn''t as good as it used to be. I went back and read the entire conversation again and I think something clicked. (Or maybe just broke). Try disabling viridian. I just remembered when I loaded Win8.1 just to test it out to see if it was bad as everyone said, I ran into something similar. If I remember correctly I completely disabled viridian or at least until I had 8.1 up and running. Like I said I''m not an expert on this. Looking over your logs I don''t see anything that stands out. Hopefully something here helps. If not hopefully one of the guru''s will step in. Just out of curiosity, Is there is reason for disabling the IOMMU remapping in your grub file? -- Shane D. Johnson IT Administrator Rasmussen Equipment
On Tue, Oct 22, 2013 at 8:37 PM, Shane Johnson <sdj@rasmussenequipment.com>wrote:> On Tue, Oct 22, 2013 at 4:30 PM, John Wargo <wargo.john25@gmail.com> > wrote: > > On 10/22/2013 02:51 PM, Shane Johnson wrote: > > > > <snip> > > > > (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault > > address = 0xffffffc0 > > [truncated output, only first 10 iterations are shown...] > > > > What is the best way to go about solving this problem? > > As always just ask for more info, > > -John > > > > _______________________________________________ > > Xen-users mailing list > > Xen-users@lists.xen.org > > http://lists.xen.org/xen-users > > > > > > It seemed to fix the irq error, however I''m still getting the > IO_PAGE_FAULT > > errors. > > I''ll go through the debugging page on the wiki, and see what I can''t dig > > up.. > > -John > > > > John, > > I know when I have need help before, I was asked to include the > > verbose output of the create command and the log files mentioned in > > the submitting a bug section. From looking at those logs, those on > > here who know a heck of a lot more that I do where able to spot what > > my problem was. If you can''t find it with the steps on the debugging > > wiki that may be the next best step. Also some details on your > > hardware and a file or screen print of your xl list command would also > > probably be beneficial. > > > > Good luck > > Shane > > > > > > After some debugging, here is what I''ve been able to come up with, if > > there''s an issue with attachments, just let me know how to properly post > the > > logs. > > > > Included in the zip : > > create.xen - The result of "xl create -d ./win8.cfg" > > dmesg-boot.xen - The result of "xl dmesg" immediately after boot > > dmesg-win8.xen - The result of "xl dmesg" after failing to install > windows 8 > > grub - My grub configuration file (/etc/default/grub) > > info.xen - The result of "xl info" > > win8.cfg - The configuration file I''m using for windows 8 > > > > -John > > > Sorry John, > Memory isn''t as good as it used to be. I went back and read the > entire conversation again and I think something clicked. (Or maybe > just broke). Try disabling viridian. I just remembered when I loaded > Win8.1 just to test it out to see if it was bad as everyone said, I > ran into something similar. If I remember correctly I completely > disabled viridian or at least until I had 8.1 up and running. > Like I said I''m not an expert on this. > Looking over your logs I don''t see anything that stands out. > Hopefully something here helps. If not hopefully one of the guru''s > will step in. > Just out of curiosity, Is there is reason for disabling the IOMMU > remapping in your grub file? > > -- > Shane D. Johnson > IT Administrator > Rasmussen Equipment > > _______________________________________________ > Xen-users mailing list > Xen-users@lists.xen.org > http://lists.xen.org/xen-users >Shane, When I disable viridian, I can complete the "getting devices ready" phase, however after that the vnc window goes black. When I look at the VCPUs though they are definitively processing something, so I think that it may be a graphics issue caused by one of my other attempts at fixing it, but I''ll have to look into it tomorrow. Unfortunately I am still getting the IO page fault from AMD Vi, when I look at ''xl dmesg''. To answer your question, I set iommu=no-intremap (I assume that this is what you''re talking about), because every other option I''ve tried will result in an error disabling IOMMU on startup. (I have one of those FX processors with the buggy tables..) Thanks for your help, -John _______________________________________________ Xen-users mailing list Xen-users@lists.xen.org http://lists.xen.org/xen-users
On 10/22/2013 08:37 PM, Shane Johnson wrote:> On Tue, Oct 22, 2013 at 4:30 PM, John Wargo <wargo.john25@gmail.com> wrote: >> On 10/22/2013 02:51 PM, Shane Johnson wrote: >> >> <snip> >> >> (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault >> address = 0xffffffc0 >> [truncated output, only first 10 iterations are shown...] >> >> What is the best way to go about solving this problem? >> As always just ask for more info, >> -John >> >> _______________________________________________ >> Xen-users mailing list >> Xen-users@lists.xen.org >> http://lists.xen.org/xen-users >> >> >> It seemed to fix the irq error, however I''m still getting the IO_PAGE_FAULT >> errors. >> I''ll go through the debugging page on the wiki, and see what I can''t dig >> up.. >> -John >> >> John, >> I know when I have need help before, I was asked to include the >> verbose output of the create command and the log files mentioned in >> the submitting a bug section. From looking at those logs, those on >> here who know a heck of a lot more that I do where able to spot what >> my problem was. If you can''t find it with the steps on the debugging >> wiki that may be the next best step. Also some details on your >> hardware and a file or screen print of your xl list command would also >> probably be beneficial. >> >> Good luck >> Shane >> >> >> After some debugging, here is what I''ve been able to come up with, if >> there''s an issue with attachments, just let me know how to properly post the >> logs. >> >> Included in the zip : >> create.xen - The result of "xl create -d ./win8.cfg" >> dmesg-boot.xen - The result of "xl dmesg" immediately after boot >> dmesg-win8.xen - The result of "xl dmesg" after failing to install windows 8 >> grub - My grub configuration file (/etc/default/grub) >> info.xen - The result of "xl info" >> win8.cfg - The configuration file I''m using for windows 8 >> >> -John > > Sorry John, > Memory isn''t as good as it used to be. I went back and read the > entire conversation again and I think something clicked. (Or maybe > just broke). Try disabling viridian. I just remembered when I loaded > Win8.1 just to test it out to see if it was bad as everyone said, I > ran into something similar. If I remember correctly I completely > disabled viridian or at least until I had 8.1 up and running. > Like I said I''m not an expert on this. > Looking over your logs I don''t see anything that stands out. > Hopefully something here helps. If not hopefully one of the guru''s > will step in. > Just out of curiosity, Is there is reason for disabling the IOMMU > remapping in your grub file? >Hmm, I disabled the VGA passthrough, and it all seems to work fine, I''m going to try passing my GPU through now that it''s installed, but I doubt that it''ll work. I''m using an XFX Radeon 7850 for passthrough, it''s reported to work, so I think that it''s the issues that I''ve been having with my AMD FX-6300 processor and IOMMU.. If anybody knows of how to fix these issues, that would be great.. -John
John, Some other things I have thought of while reviewing your problem. I don''t remember where I read it but I remember reading that someone couldn''t get passthrough working during install and only by waiting till Windows 7 I believe it was was fully installed then they installed the driver and then enabled passthrough. Another thing you might want to try is reducing the amount of RAM you are throwing at the DomU to below 4G I think it is. I know there are some posts on the forums about people having issues with this and if I remember right there is a solution for running more memory but I don''t know exactly what it is. Shane On Wed, Oct 23, 2013 at 4:53 AM, John Wargo <wargo.john25@gmail.com> wrote:> On 10/22/2013 08:37 PM, Shane Johnson wrote: >> >> On Tue, Oct 22, 2013 at 4:30 PM, John Wargo <wargo.john25@gmail.com> >> wrote: >>> >>> On 10/22/2013 02:51 PM, Shane Johnson wrote: >>> >>> <snip> >>> >>> (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault >>> address = 0xffffffc0 >>> [truncated output, only first 10 iterations are shown...] >>> >>> What is the best way to go about solving this problem? >>> As always just ask for more info, >>> -John >>> >>> _______________________________________________ >>> Xen-users mailing list >>> Xen-users@lists.xen.org >>> http://lists.xen.org/xen-users >>> >>> >>> It seemed to fix the irq error, however I''m still getting the >>> IO_PAGE_FAULT >>> errors. >>> I''ll go through the debugging page on the wiki, and see what I can''t dig >>> up.. >>> -John >>> >>> John, >>> I know when I have need help before, I was asked to include the >>> verbose output of the create command and the log files mentioned in >>> the submitting a bug section. From looking at those logs, those on >>> here who know a heck of a lot more that I do where able to spot what >>> my problem was. If you can''t find it with the steps on the debugging >>> wiki that may be the next best step. Also some details on your >>> hardware and a file or screen print of your xl list command would also >>> probably be beneficial. >>> >>> Good luck >>> Shane >>> >>> >>> After some debugging, here is what I''ve been able to come up with, if >>> there''s an issue with attachments, just let me know how to properly post >>> the >>> logs. >>> >>> Included in the zip : >>> create.xen - The result of "xl create -d ./win8.cfg" >>> dmesg-boot.xen - The result of "xl dmesg" immediately after boot >>> dmesg-win8.xen - The result of "xl dmesg" after failing to install >>> windows 8 >>> grub - My grub configuration file (/etc/default/grub) >>> info.xen - The result of "xl info" >>> win8.cfg - The configuration file I''m using for windows 8 >>> >>> -John >> >> >> Sorry John, >> Memory isn''t as good as it used to be. I went back and read the >> entire conversation again and I think something clicked. (Or maybe >> just broke). Try disabling viridian. I just remembered when I loaded >> Win8.1 just to test it out to see if it was bad as everyone said, I >> ran into something similar. If I remember correctly I completely >> disabled viridian or at least until I had 8.1 up and running. >> Like I said I''m not an expert on this. >> Looking over your logs I don''t see anything that stands out. >> Hopefully something here helps. If not hopefully one of the guru''s >> will step in. >> Just out of curiosity, Is there is reason for disabling the IOMMU >> remapping in your grub file? >> > Hmm, I disabled the VGA passthrough, and it all seems to work fine, I''m > going to try passing my GPU > through now that it''s installed, but I doubt that it''ll work. > > I''m using an XFX Radeon 7850 for passthrough, it''s reported to work, so I > think that it''s the issues > that I''ve been having with my AMD FX-6300 processor and IOMMU.. > > If anybody knows of how to fix these issues, that would be great.. > -John-- Shane D. Johnson IT Administrator Rasmussen Equipment
On 10/23/2013 08:15 AM, Shane Johnson wrote:> John, > Some other things I have thought of while reviewing your problem. I > don''t remember where I read it but I remember reading that someone > couldn''t get passthrough working during install and only by waiting > till Windows 7 I believe it was was fully installed then they > installed the driver and then enabled passthrough. Another thing you > might want to try is reducing the amount of RAM you are throwing at > the DomU to below 4G I think it is. I know there are some posts on > the forums about people having issues with this and if I remember > right there is a solution for running more memory but I don''t know > exactly what it is. > > Shane > > On Wed, Oct 23, 2013 at 4:53 AM, John Wargo <wargo.john25@gmail.com> wrote: >> On 10/22/2013 08:37 PM, Shane Johnson wrote: >>> On Tue, Oct 22, 2013 at 4:30 PM, John Wargo <wargo.john25@gmail.com> >>> wrote: >>>> On 10/22/2013 02:51 PM, Shane Johnson wrote: >>>> >>>> <snip> >>>> >>>> (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault >>>> address = 0xffffffc0 >>>> [truncated output, only first 10 iterations are shown...] >>>> >>>> What is the best way to go about solving this problem? >>>> As always just ask for more info, >>>> -John >>>> >>>> _______________________________________________ >>>> Xen-users mailing list >>>> Xen-users@lists.xen.org >>>> http://lists.xen.org/xen-users >>>> >>>> >>>> It seemed to fix the irq error, however I''m still getting the >>>> IO_PAGE_FAULT >>>> errors. >>>> I''ll go through the debugging page on the wiki, and see what I can''t dig >>>> up.. >>>> -John >>>> >>>> John, >>>> I know when I have need help before, I was asked to include the >>>> verbose output of the create command and the log files mentioned in >>>> the submitting a bug section. From looking at those logs, those on >>>> here who know a heck of a lot more that I do where able to spot what >>>> my problem was. If you can''t find it with the steps on the debugging >>>> wiki that may be the next best step. Also some details on your >>>> hardware and a file or screen print of your xl list command would also >>>> probably be beneficial. >>>> >>>> Good luck >>>> Shane >>>> >>>> >>>> After some debugging, here is what I''ve been able to come up with, if >>>> there''s an issue with attachments, just let me know how to properly post >>>> the >>>> logs. >>>> >>>> Included in the zip : >>>> create.xen - The result of "xl create -d ./win8.cfg" >>>> dmesg-boot.xen - The result of "xl dmesg" immediately after boot >>>> dmesg-win8.xen - The result of "xl dmesg" after failing to install >>>> windows 8 >>>> grub - My grub configuration file (/etc/default/grub) >>>> info.xen - The result of "xl info" >>>> win8.cfg - The configuration file I''m using for windows 8 >>>> >>>> -John >>> >>> Sorry John, >>> Memory isn''t as good as it used to be. I went back and read the >>> entire conversation again and I think something clicked. (Or maybe >>> just broke). Try disabling viridian. I just remembered when I loaded >>> Win8.1 just to test it out to see if it was bad as everyone said, I >>> ran into something similar. If I remember correctly I completely >>> disabled viridian or at least until I had 8.1 up and running. >>> Like I said I''m not an expert on this. >>> Looking over your logs I don''t see anything that stands out. >>> Hopefully something here helps. If not hopefully one of the guru''s >>> will step in. >>> Just out of curiosity, Is there is reason for disabling the IOMMU >>> remapping in your grub file? >>> >> Hmm, I disabled the VGA passthrough, and it all seems to work fine, I''m >> going to try passing my GPU >> through now that it''s installed, but I doubt that it''ll work. >> >> I''m using an XFX Radeon 7850 for passthrough, it''s reported to work, so I >> think that it''s the issues >> that I''ve been having with my AMD FX-6300 processor and IOMMU.. >> >> If anybody knows of how to fix these issues, that would be great.. >> -John > >I was browsing the wiki earlier today, and I found this : http://old-list-archives.xenproject.org/archives/html/xen-devel/2010-10/txtfLpL6CdMGC.txt It''s a source-code patch for xen. How do I apply it? Also, I''m thinking about building Xen-4.3.0-RC6 from source, is there a reason why I shouldn''t try to see if a fix has been implemented there? Thanks, -John _______________________________________________ Xen-users mailing list Xen-users@lists.xen.org http://lists.xen.org/xen-users
John, Only thing I would hesitate over is it isn''t a release version and may still have bugs in it. As for that patch, it''s over 3 years old. I would imagine it''s in the version you''re running. I would see what version it was written for and what you''re running and maybe a little research to see if the patch was accepted. I think your best bet would be either stable or, if you don''t mind helping the developers debug it, the 4.3.0-RC6 you mentioned. Shane On Wed, Oct 23, 2013 at 1:55 PM, John Wargo <wargo.john25@gmail.com> wrote:> On 10/23/2013 08:15 AM, Shane Johnson wrote: > > John, > Some other things I have thought of while reviewing your problem. I > don''t remember where I read it but I remember reading that someone > couldn''t get passthrough working during install and only by waiting > till Windows 7 I believe it was was fully installed then they > installed the driver and then enabled passthrough. Another thing you > might want to try is reducing the amount of RAM you are throwing at > the DomU to below 4G I think it is. I know there are some posts on > the forums about people having issues with this and if I remember > right there is a solution for running more memory but I don''t know > exactly what it is. > > Shane > > On Wed, Oct 23, 2013 at 4:53 AM, John Wargo <wargo.john25@gmail.com> wrote: > > On 10/22/2013 08:37 PM, Shane Johnson wrote: > > On Tue, Oct 22, 2013 at 4:30 PM, John Wargo <wargo.john25@gmail.com> > wrote: > > On 10/22/2013 02:51 PM, Shane Johnson wrote: > > <snip> > > (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault > address = 0xffffffc0 > [truncated output, only first 10 iterations are shown...] > > What is the best way to go about solving this problem? > As always just ask for more info, > -John > > _______________________________________________ > Xen-users mailing list > Xen-users@lists.xen.org > http://lists.xen.org/xen-users > > > It seemed to fix the irq error, however I''m still getting the > IO_PAGE_FAULT > errors. > I''ll go through the debugging page on the wiki, and see what I can''t dig > up.. > -John > > John, > I know when I have need help before, I was asked to include the > verbose output of the create command and the log files mentioned in > the submitting a bug section. From looking at those logs, those on > here who know a heck of a lot more that I do where able to spot what > my problem was. If you can''t find it with the steps on the debugging > wiki that may be the next best step. Also some details on your > hardware and a file or screen print of your xl list command would also > probably be beneficial. > > Good luck > Shane > > > After some debugging, here is what I''ve been able to come up with, if > there''s an issue with attachments, just let me know how to properly post > the > logs. > > Included in the zip : > create.xen - The result of "xl create -d ./win8.cfg" > dmesg-boot.xen - The result of "xl dmesg" immediately after boot > dmesg-win8.xen - The result of "xl dmesg" after failing to install > windows 8 > grub - My grub configuration file (/etc/default/grub) > info.xen - The result of "xl info" > win8.cfg - The configuration file I''m using for windows 8 > > -John > > Sorry John, > Memory isn''t as good as it used to be. I went back and read the > entire conversation again and I think something clicked. (Or maybe > just broke). Try disabling viridian. I just remembered when I loaded > Win8.1 just to test it out to see if it was bad as everyone said, I > ran into something similar. If I remember correctly I completely > disabled viridian or at least until I had 8.1 up and running. > Like I said I''m not an expert on this. > Looking over your logs I don''t see anything that stands out. > Hopefully something here helps. If not hopefully one of the guru''s > will step in. > Just out of curiosity, Is there is reason for disabling the IOMMU > remapping in your grub file? > > Hmm, I disabled the VGA passthrough, and it all seems to work fine, I''m > going to try passing my GPU > through now that it''s installed, but I doubt that it''ll work. > > I''m using an XFX Radeon 7850 for passthrough, it''s reported to work, so I > think that it''s the issues > that I''ve been having with my AMD FX-6300 processor and IOMMU.. > > If anybody knows of how to fix these issues, that would be great.. > -John > > > I was browsing the wiki earlier today, and I found this : > http://old-list-archives.xenproject.org/archives/html/xen-devel/2010-10/txtfLpL6CdMGC.txt > It''s a source-code patch for xen. How do I apply it? > Also, I''m thinking about building Xen-4.3.0-RC6 from source, is there a > reason why I shouldn''t try to see if a fix has been implemented there? > > Thanks, > -John-- Shane D. Johnson IT Administrator Rasmussen Equipment
On 10/23/2013 03:10 PM, Shane Johnson wrote:> John, > Only thing I would hesitate over is it isn''t a release version and may > still have bugs in it. As for that patch, it''s over 3 years old. I > would imagine it''s in the version you''re running. I would see what > version it was written for and what you''re running and maybe a little > research to see if the patch was accepted. I think your best bet > would be either stable or, if you don''t mind helping the developers > debug it, the 4.3.0-RC6 you mentioned. > > Shane > > > On Wed, Oct 23, 2013 at 1:55 PM, John Wargo <wargo.john25@gmail.com> wrote: >> On 10/23/2013 08:15 AM, Shane Johnson wrote: >> >> John, >> Some other things I have thought of while reviewing your problem. I >> don''t remember where I read it but I remember reading that someone >> couldn''t get passthrough working during install and only by waiting >> till Windows 7 I believe it was was fully installed then they >> installed the driver and then enabled passthrough. Another thing you >> might want to try is reducing the amount of RAM you are throwing at >> the DomU to below 4G I think it is. I know there are some posts on >> the forums about people having issues with this and if I remember >> right there is a solution for running more memory but I don''t know >> exactly what it is. >> >> Shane >> >> On Wed, Oct 23, 2013 at 4:53 AM, John Wargo <wargo.john25@gmail.com> wrote: >> >> On 10/22/2013 08:37 PM, Shane Johnson wrote: >> >> On Tue, Oct 22, 2013 at 4:30 PM, John Wargo <wargo.john25@gmail.com> >> wrote: >> >> On 10/22/2013 02:51 PM, Shane Johnson wrote: >> >> <snip> >> >> (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault >> address = 0xffffffc0 >> [truncated output, only first 10 iterations are shown...] >> >> What is the best way to go about solving this problem? >> As always just ask for more info, >> -John >> >> _______________________________________________ >> Xen-users mailing list >> Xen-users@lists.xen.org >> http://lists.xen.org/xen-users >> >> >> It seemed to fix the irq error, however I''m still getting the >> IO_PAGE_FAULT >> errors. >> I''ll go through the debugging page on the wiki, and see what I can''t dig >> up.. >> -John >> >> John, >> I know when I have need help before, I was asked to include the >> verbose output of the create command and the log files mentioned in >> the submitting a bug section. From looking at those logs, those on >> here who know a heck of a lot more that I do where able to spot what >> my problem was. If you can''t find it with the steps on the debugging >> wiki that may be the next best step. Also some details on your >> hardware and a file or screen print of your xl list command would also >> probably be beneficial. >> >> Good luck >> Shane >> >> >> After some debugging, here is what I''ve been able to come up with, if >> there''s an issue with attachments, just let me know how to properly post >> the >> logs. >> >> Included in the zip : >> create.xen - The result of "xl create -d ./win8.cfg" >> dmesg-boot.xen - The result of "xl dmesg" immediately after boot >> dmesg-win8.xen - The result of "xl dmesg" after failing to install >> windows 8 >> grub - My grub configuration file (/etc/default/grub) >> info.xen - The result of "xl info" >> win8.cfg - The configuration file I''m using for windows 8 >> >> -John >> >> Sorry John, >> Memory isn''t as good as it used to be. I went back and read the >> entire conversation again and I think something clicked. (Or maybe >> just broke). Try disabling viridian. I just remembered when I loaded >> Win8.1 just to test it out to see if it was bad as everyone said, I >> ran into something similar. If I remember correctly I completely >> disabled viridian or at least until I had 8.1 up and running. >> Like I said I''m not an expert on this. >> Looking over your logs I don''t see anything that stands out. >> Hopefully something here helps. If not hopefully one of the guru''s >> will step in. >> Just out of curiosity, Is there is reason for disabling the IOMMU >> remapping in your grub file? >> >> Hmm, I disabled the VGA passthrough, and it all seems to work fine, I''m >> going to try passing my GPU >> through now that it''s installed, but I doubt that it''ll work. >> >> I''m using an XFX Radeon 7850 for passthrough, it''s reported to work, so I >> think that it''s the issues >> that I''ve been having with my AMD FX-6300 processor and IOMMU.. >> >> If anybody knows of how to fix these issues, that would be great.. >> -John >> >> >> I was browsing the wiki earlier today, and I found this : >> http://old-list-archives.xenproject.org/archives/html/xen-devel/2010-10/txtfLpL6CdMGC.txt >> It''s a source-code patch for xen. How do I apply it? >> Also, I''m thinking about building Xen-4.3.0-RC6 from source, is there a >> reason why I shouldn''t try to see if a fix has been implemented there? >> >> Thanks, >> -John > >Shane, Thanks for the advice, I just realized that on the wiki it stated that it needed to install AMD''s official drivers before secondary passthrough would work, so I''m working on that right now. If that doesn''t work then I''ll reduce the memory. If neither work, then I''ll probably give RC6 a go, and notify the dev''s if I''m still having issues. Thanks, -John
On 10/23/2013 03:10 PM, Shane Johnson wrote:> John, > Only thing I would hesitate over is it isn''t a release version and may > still have bugs in it. As for that patch, it''s over 3 years old. I > would imagine it''s in the version you''re running. I would see what > version it was written for and what you''re running and maybe a little > research to see if the patch was accepted. I think your best bet > would be either stable or, if you don''t mind helping the developers > debug it, the 4.3.0-RC6 you mentioned. > > Shane > > > On Wed, Oct 23, 2013 at 1:55 PM, John Wargo <wargo.john25@gmail.com> wrote: >> On 10/23/2013 08:15 AM, Shane Johnson wrote: >> >> John, >> Some other things I have thought of while reviewing your problem. I >> don''t remember where I read it but I remember reading that someone >> couldn''t get passthrough working during install and only by waiting >> till Windows 7 I believe it was was fully installed then they >> installed the driver and then enabled passthrough. Another thing you >> might want to try is reducing the amount of RAM you are throwing at >> the DomU to below 4G I think it is. I know there are some posts on >> the forums about people having issues with this and if I remember >> right there is a solution for running more memory but I don''t know >> exactly what it is. >> >> Shane >> >> On Wed, Oct 23, 2013 at 4:53 AM, John Wargo <wargo.john25@gmail.com> wrote: >> >> On 10/22/2013 08:37 PM, Shane Johnson wrote: >> >> On Tue, Oct 22, 2013 at 4:30 PM, John Wargo <wargo.john25@gmail.com> >> wrote: >> >> On 10/22/2013 02:51 PM, Shane Johnson wrote: >> >> <snip> >> >> (XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault >> address = 0xffffffc0 >> [truncated output, only first 10 iterations are shown...] >> >> What is the best way to go about solving this problem? >> As always just ask for more info, >> -John >> >> _______________________________________________ >> Xen-users mailing list >> Xen-users@lists.xen.org >> http://lists.xen.org/xen-users >> >> >> It seemed to fix the irq error, however I''m still getting the >> IO_PAGE_FAULT >> errors. >> I''ll go through the debugging page on the wiki, and see what I can''t dig >> up.. >> -John >> >> John, >> I know when I have need help before, I was asked to include the >> verbose output of the create command and the log files mentioned in >> the submitting a bug section. From looking at those logs, those on >> here who know a heck of a lot more that I do where able to spot what >> my problem was. If you can''t find it with the steps on the debugging >> wiki that may be the next best step. Also some details on your >> hardware and a file or screen print of your xl list command would also >> probably be beneficial. >> >> Good luck >> Shane >> >> >> After some debugging, here is what I''ve been able to come up with, if >> there''s an issue with attachments, just let me know how to properly post >> the >> logs. >> >> Included in the zip : >> create.xen - The result of "xl create -d ./win8.cfg" >> dmesg-boot.xen - The result of "xl dmesg" immediately after boot >> dmesg-win8.xen - The result of "xl dmesg" after failing to install >> windows 8 >> grub - My grub configuration file (/etc/default/grub) >> info.xen - The result of "xl info" >> win8.cfg - The configuration file I''m using for windows 8 >> >> -John >> >> Sorry John, >> Memory isn''t as good as it used to be. I went back and read the >> entire conversation again and I think something clicked. (Or maybe >> just broke). Try disabling viridian. I just remembered when I loaded >> Win8.1 just to test it out to see if it was bad as everyone said, I >> ran into something similar. If I remember correctly I completely >> disabled viridian or at least until I had 8.1 up and running. >> Like I said I''m not an expert on this. >> Looking over your logs I don''t see anything that stands out. >> Hopefully something here helps. If not hopefully one of the guru''s >> will step in. >> Just out of curiosity, Is there is reason for disabling the IOMMU >> remapping in your grub file? >> >> Hmm, I disabled the VGA passthrough, and it all seems to work fine, I''m >> going to try passing my GPU >> through now that it''s installed, but I doubt that it''ll work. >> >> I''m using an XFX Radeon 7850 for passthrough, it''s reported to work, so I >> think that it''s the issues >> that I''ve been having with my AMD FX-6300 processor and IOMMU.. >> >> If anybody knows of how to fix these issues, that would be great.. >> -John >> >> >> I was browsing the wiki earlier today, and I found this : >> http://old-list-archives.xenproject.org/archives/html/xen-devel/2010-10/txtfLpL6CdMGC.txt >> It''s a source-code patch for xen. How do I apply it? >> Also, I''m thinking about building Xen-4.3.0-RC6 from source, is there a >> reason why I shouldn''t try to see if a fix has been implemented there? >> >> Thanks, >> -John > >After installing the xen drivers and the catalyst drivers, I rebooted windows. Upon restarting, it went back to the "Configuring Devices n%" screen, where it proceeded to crash once again. This time though, the crash didn''t bring down my entire system. The entire dmesg output is filled with "(XEN) AMD-Vi: IO_PAGE_FAULT: domain = 0, device id = 0x0090, fault address = 0xffffffc0" (or at least until the terminal buffer). The full file is attached. Any thoughts? -John _______________________________________________ Xen-users mailing list Xen-users@lists.xen.org http://lists.xen.org/xen-users
John, On Wed, Oct 23, 2013 at 2:55 PM, John Wargo <wargo.john25@gmail.com> wrote:> I was browsing the wiki earlier today, and I found this : > http://old-list-archives.xenproject.org/archives/html/xen-devel/2010-10/txtfLpL6CdMGC.txt > It''s a source-code patch for xen. How do I apply it? > Also, I''m thinking about building Xen-4.3.0-RC6 from source, is there a > reason why I shouldn''t try to see if a fix has been implemented there? > > Thanks, > -John > > That particular patch is still used today by some people, when workingwith passing an AMD card through to a guest domU. Theres a copy of it included with the ArchLinux xen aur packages, cleaned up so it will apply cleanly to xen v4.3.0. Its not been accepted into the source code at this time. To install it, typically you''d grab a copy of the xen source code, download the patch file, then use the ''patch'' command to update the source code with the patches changes. Passthrough of AMD graphics cards can be a little problematic at times; often people have great access the first time the domain is started, but restarts of the domain cause the card to lock up and often require the host system to be rebooted to reset the graphics card. I have a system where I''ve been passing through a 7750 to a Windows 7 guest, and it experiences issues if I have to restart the guest for any reason (until I reboot the host). It appears partly to be related to driver issues and getting the card to reset properly. I''ve also been using a modified Nvidia GTX 670 (reports back as a Grid K2, but have Physx working) and its been much more stable (the Grid K2 drivers are better designed for a virtualized environment) Regards, David _______________________________________________ Xen-users mailing list Xen-users@lists.xen.org http://lists.xen.org/xen-users
On 10/23/2013 07:53 PM, David Sutton wrote:> John, > > On Wed, Oct 23, 2013 at 2:55 PM, John Wargo <wargo.john25@gmail.com > <mailto:wargo.john25@gmail.com>> wrote: > > I was browsing the wiki earlier today, and I found this : > http://old-list-archives.xenproject.org/archives/html/xen-devel/2010-10/txtfLpL6CdMGC.txt > It''s a source-code patch for xen. How do I apply it? > Also, I''m thinking about building Xen-4.3.0-RC6 from source, is > there a reason why I shouldn''t try to see if a fix has been > implemented there? > > Thanks, > -John > > That particular patch is still used today by some people, when working > with passing an AMD card through to a guest domU. Theres a copy of it > included with the ArchLinux xen aur packages, cleaned up so it will > apply cleanly to xen v4.3.0. Its not been accepted into the source > code at this time. To install it, typically you''d grab a copy of the > xen source code, download the patch file, then use the ''patch'' command > to update the source code with the patches changes. > > Passthrough of AMD graphics cards can be a little problematic at > times; often people have great access the first time the domain is > started, but restarts of the domain cause the card to lock up and > often require the host system to be rebooted to reset the graphics > card. I have a system where I''ve been passing through a 7750 to a > Windows 7 guest, and it experiences issues if I have to restart the > guest for any reason (until I reboot the host). It appears partly to > be related to driver issues and getting the card to reset properly. > I''ve also been using a modified Nvidia GTX 670 (reports back as a Grid > K2, but have Physx working) and its been much more stable (the Grid K2 > drivers are better designed for a virtualized environment) > > Regards, > > David >So far no luck, as of right now I''m compiling the source code for the 4.3 RC, and I''ll be contacting the developers if the problem persists. Thank you both for your time, -John _______________________________________________ Xen-users mailing list Xen-users@lists.xen.org http://lists.xen.org/xen-users