Xen devel - Jan 2005 - [PATCH] USB virtualisation (experimental)

Hi all,

A number of people have expressed interest in the USB virtualisation code.  
I''m making a patch against 2.0-testing available for those who wish to
have a
play with it.

Description:
USB virtualisation allows otherwise unprivileged domains to be given control 
of a USB port on the host.  The unprivileged domain sees a virtual USB host 
controller with only the ports it has been granted access to.  For instance, 
you might dedicate a USB webcam to one domain, a USB pendrive to another 
domain and a USB CD-burner to another.

This patch includes the kernel modifications (backend and frontend drivers) to 
2.4.28 and modifications to the userspace toolset.  This is sufficient to set 
up USB virtualisation for 2.4.28 on xen-2.0-testing.

There are still a few sticky bugs that are really annoying me at the moment 
but it is possible to mount USB flash drives, capture realtime video from a 
webcam, etc. in a domU by virtualising through dom0.

Disclaimers:
* Don''t use it on production systems
* Don''t trust it with data you care about
* Some USB device drivers may need tweaking - if you have problems with a 
device, let me know
* YMMV but it works for me ;-)

Installation:
1) download the patch from
http://www.cambridge.intel-research.net/~mwilli2/xen-2.0-testing-usb.patch
2) apply to a fresh xen-2.0-testing.bk using
patch -p1 < xen-2.0-testing-usb.patch
(say yes to all the SCCS checkouts)
3) set up some appropriate kernel configs in 
install/boot/config-2.4.28-xen{0,U}.  You should enable 
CONFIG_XEN_USB_FRONTEND, CONFIG_USB in the configuration for any devices you 
want to run.
Don''t (!) build drivers for devices you want to virtualise into
dom0''s kernel,
or dom0 will grab the device before the backend gets a chance to export it.
4) build & install the updated kernel & tools on your test machine. 
Note
again you _have_ to use the 2.4 kernels, 2.6 isn''t supported (yet).
5) for frontend domains, add a
usb = [ ''port_number'', ''port_number'', ... ]
to the config file to give a domain control of those physical USB ports.  You 
can specify ports on a hub by giving the path through the device tree e.g. 
''1/2'' = second port on the hub plugged into the first port on
the host.

Any questions, comments, problems, success reports - let us know!

Cheers,
Mark


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It''s fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

Mark Williamson wrote:
> This patch includes the kernel modifications (backend and frontend drivers)
to
> 2.4.28 and modifications to the userspace toolset.  This is sufficient to
set
> up USB virtualisation for 2.4.28 on xen-2.0-testing.
> 
> There are still a few sticky bugs that are really annoying me at the moment
> but it is possible to mount USB flash drives, capture realtime video from a
> webcam, etc. in a domU by virtualising through dom0.
> Any questions, comments, problems, success reports - let us know!
Nice work, Mark! Will be checking this out, but need
a 2.6 version. Will you be working on the 2.6
port in the near future? If not, perhaps we could
help..

thanks,
Nivedita



-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It''s fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

> Nice work, Mark!
Cheers!  Don''t get too excited until it works for you too though ;-)
> Will be checking this out, but need 
> a 2.6 version. Will you be working on the 2.6
> port in the near future?
It''ll probably be a while - once I''ve chased down a few more
issues and merged
the patch into BitKeeper, I''m keen to go work on something else for a
bit ;-)

I suspect porting the backend will be reasonably straightforward as I
don''t
think the USB client API has changed much (at all?).

The frontend will probably require a little of work - the host controller 
interface has changed a bit.  A quick scan of the code doesn''t suggest 
anything dramatic, however.
> If not, perhaps we could 
> help..
That would be very welcome :-)  If you have the time to try it out I''d
love to
see a patch that I could merge.

Another thing I''d like to see someday is a BSD frontend.  My frontend
is GPL
so it''s probably easiest to rewrite a BSD frontend from scratch.  If
anyone
in the community is interested I could provide assistance.

Cheers,
Mark


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It''s fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

Mark, how did you manage to include in the patch also main.bak, main.c2? ;-)

I am longing to see 2.6 USB virtualization soon. 

Thank you,
AQ


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It''s fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

> Mark, how did you manage to include in the patch also main.bak, main.c2?
> ;-)
Oops, sorry :-)  They were just hanging around in the src tree and got plopped 
out when I did the diff.  Ignore them.

Let me know how you get on if you try this out.

Cheers,
Mark


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It''s fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

All,

Here is an updated USB virtualisation patch.  Things are rather more well 
behaved now.  I just dumped 8 meg of video out of a USB webcam and onto a USB 
pendrive (with both of these peripherals running through the USB virt 
system), no problems at all.

I''m intending to merge this into -unstable fairly soon but in the
meantime,
the patches are against 2.0-testing.

Enjoy,
Mark

On Monday 10 January 2005 17:03, Mark Williamson wrote:
> Hi all,
>
> A number of people have expressed interest in the USB virtualisation code.
> I''m making a patch against 2.0-testing available for those who
wish to have
> a play with it.
>
> Description:
> USB virtualisation allows otherwise unprivileged domains to be given
> control of a USB port on the host.  The unprivileged domain sees a virtual
> USB host controller with only the ports it has been granted access to.  For
> instance, you might dedicate a USB webcam to one domain, a USB pendrive to
> another domain and a USB CD-burner to another.
>
> This patch includes the kernel modifications (backend and frontend drivers)
> to 2.4.28 and modifications to the userspace toolset.  This is sufficient
> to set up USB virtualisation for 2.4.28 on xen-2.0-testing.
>
> There are still a few sticky bugs that are really annoying me at the moment
> but it is possible to mount USB flash drives, capture realtime video from a
> webcam, etc. in a domU by virtualising through dom0.
>
> Disclaimers:
> * Don''t use it on production systems
> * Don''t trust it with data you care about
> * Some USB device drivers may need tweaking - if you have problems with a
> device, let me know
> * YMMV but it works for me ;-)
>
> Installation:
> 1) download the patch from
> http://www.cambridge.intel-research.net/~mwilli2/xen-2.0-testing-usb.patch
> 2) apply to a fresh xen-2.0-testing.bk using
> patch -p1 < xen-2.0-testing-usb.patch
> (say yes to all the SCCS checkouts)
> 3) set up some appropriate kernel configs in
> install/boot/config-2.4.28-xen{0,U}.  You should enable
> CONFIG_XEN_USB_FRONTEND, CONFIG_USB in the configuration for any devices
> you want to run.
> Don''t (!) build drivers for devices you want to virtualise into
dom0''s
> kernel, or dom0 will grab the device before the backend gets a chance to
> export it. 4) build & install the updated kernel & tools on your
test
> machine.  Note again you _have_ to use the 2.4 kernels, 2.6 isn''t
supported
> (yet). 5) for frontend domains, add a
> usb = [ ''port_number'', ''port_number'',
... ]
> to the config file to give a domain control of those physical USB ports. 
> You can specify ports on a hub by giving the path through the device tree
> e.g. ''1/2'' = second port on the hub plugged into the
first port on the
> host.
>
> Any questions, comments, problems, success reports - let us know!
>
> Cheers,
> Mark
>
>
> -------------------------------------------------------
> The SF.Net email is sponsored by: Beat the post-holiday blues
> Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
> It''s fun and FREE -- well,
almost....http://www.thinkgeek.com/sfshirt
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/xen-devel

All,

I have updated the USB virtualisation patch at:
http://www.cambridge.intel-research.net/~mwilli2/xen-2.0-testing-usb.patch

Description:
* Give guest operating systems under Xen access to USB devices plugged into 
the host.  Adds a backend driver (for domain 0), a frontend driver (for 
guests) and extra support code in the tools.

Features:
* USB mass storage devices are now fully working (new!)
* USB isochronous transfers (larger buffers now supported, provided they are 
machine-contiguous)

Still to do:
* proper support for hot plugging of the physical devices
* stronger checking for iso transfers in the backend
* a little more work needed to make it work for all USB devices
* iron out a few lingering backend bugs

I''m intending to get this all done as soon as possible and then push
into the
-unstable tree.

Cheers,
Mark

On Monday 10 January 2005 17:03, Mark Williamson wrote:
> Hi all,
>
> A number of people have expressed interest in the USB virtualisation code.
> I''m making a patch against 2.0-testing available for those who
wish to have
> a play with it.
>
> Description:
> USB virtualisation allows otherwise unprivileged domains to be given
> control of a USB port on the host.  The unprivileged domain sees a virtual
> USB host controller with only the ports it has been granted access to.  For
> instance, you might dedicate a USB webcam to one domain, a USB pendrive to
> another domain and a USB CD-burner to another.
>
> This patch includes the kernel modifications (backend and frontend drivers)
> to 2.4.28 and modifications to the userspace toolset.  This is sufficient
> to set up USB virtualisation for 2.4.28 on xen-2.0-testing.
>
> There are still a few sticky bugs that are really annoying me at the moment
> but it is possible to mount USB flash drives, capture realtime video from a
> webcam, etc. in a domU by virtualising through dom0.
>
> Disclaimers:
> * Don''t use it on production systems
> * Don''t trust it with data you care about
> * Some USB device drivers may need tweaking - if you have problems with a
> device, let me know
> * YMMV but it works for me ;-)
>
> Installation:
> 1) download the patch from
> http://www.cambridge.intel-research.net/~mwilli2/xen-2.0-testing-usb.patch
> 2) apply to a fresh xen-2.0-testing.bk using
> patch -p1 < xen-2.0-testing-usb.patch
> (say yes to all the SCCS checkouts)
> 3) set up some appropriate kernel configs in
> install/boot/config-2.4.28-xen{0,U}.  You should enable
> CONFIG_XEN_USB_FRONTEND, CONFIG_USB in the configuration for any devices
> you want to run.
> Don''t (!) build drivers for devices you want to virtualise into
dom0''s
> kernel, or dom0 will grab the device before the backend gets a chance to
> export it. 4) build & install the updated kernel & tools on your
test
> machine.  Note again you _have_ to use the 2.4 kernels, 2.6 isn''t
supported
> (yet). 5) for frontend domains, add a
> usb = [ ''port_number'', ''port_number'',
... ]
> to the config file to give a domain control of those physical USB ports. 
> You can specify ports on a hub by giving the path through the device tree
> e.g. ''1/2'' = second port on the hub plugged into the
first port on the
> host.
>
> Any questions, comments, problems, success reports - let us know!
>
> Cheers,
> Mark
>
>
> -------------------------------------------------------
> The SF.Net email is sponsored by: Beat the post-holiday blues
> Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
> It''s fun and FREE -- well,
almost....http://www.thinkgeek.com/sfshirt
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/xen-devel

-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It''s fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

Mark Williamson wrote:
> * USB mass storage devices are now fully working (new!)
Woo, nice! I have an external MP3 player that is an USB mass storage
hard drive - if the drive happens to spin down when plugging the cable
in, it manages to freeze the USB subsystem in the kernel. Now I can
just put it in a separate virtual machine and reboot it :-)

-- Naked




-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It''s fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

> Woo, nice! I have an external MP3 player that is an USB mass storage
> hard drive - if the drive happens to spin down when plugging the cable
Hmm, that''s a nasty problem - I wonder what causes it.
> in, it manages to freeze the USB subsystem in the kernel. Now I can
> just put it in a separate virtual machine and reboot it :-)
If it''s freezing the USB subsystem (and nothing else), it implies a
problem
with the USB core and / or with the host controller driver.  If it''s
the
latter, you''d have to run the driver for your (hardware) host
controller in
the domain so that could be restarted too - the appropriate tool for that 
would be the existing PCI driver domains stuff.

What kind of host controller do you have?  (or if you don''t know, do
you have
USB 2.0?).  Do all other USB devices freeze as well when you plug in the MP3 
player?  And do you get any interesting output from dmesg regarding the 
error?

Cheers,
Mark


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It''s fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

On Fri, 14 Jan 2005, Mark Williamson wrote:
> All,
>
> I have updated the USB virtualisation patch at:
> http://www.cambridge.intel-research.net/~mwilli2/xen-2.0-testing-usb.patch
>
> Description:
> * Give guest operating systems under Xen access to USB devices plugged into
> the host.  Adds a backend driver (for domain 0), a frontend driver (for
> guests) and extra support code in the tools.
>
> Features:
> * USB mass storage devices are now fully working (new!)
> * USB isochronous transfers (larger buffers now supported, provided they
are
> machine-contiguous)
>
> Still to do:
> * proper support for hot plugging of the physical devices
> * stronger checking for iso transfers in the backend
> * a little more work needed to make it work for all USB devices
> * iron out a few lingering backend bugs
>
> I''m intending to get this all done as soon as possible and then
push into the
> -unstable tree.
Could it be possible to emulate a USB device in dom0 userspace?


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It''s fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

> Could it be possible to emulate a USB device in dom0 userspace?
Yes, that''s doable in principle.  You could, for instance, connect the 
guest''s USB frontend to a userspace process that implemented:
a) the Xen USBif protocol (as implemented by the current backend driver)
b) software emulation of one or more actual USB devices per domain

This should actually be simpler to achieve than implementing virtual USB 
devices for a whole-machine emulator like Bochs,  QEmu, VMWare, etc. because 
the paravirtualised interface is higher-level.

Cheers,
Mark


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It''s fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

Mark Williamson wrote:
>> Woo, nice! I have an external MP3 player that is an USB mass
>> storage hard drive - if the drive happens to spin down when
>> plugging the cable
>
> Hmm, that''s a nasty problem - I wonder what causes it.
I''d say it is just a trivial bug in the kernel USB stuff - expecting
some reply to always arrive, when it doesn''t if the drive is unplugged
at the unfortunate moment (or the drive dies, as in this case).
>> in, it manages to freeze the USB subsystem in the kernel. Now I can
>> just put it in a separate virtual machine and reboot it :-)
>
> If it''s freezing the USB subsystem (and nothing else), it implies
a
> problem with the USB core and / or with the host controller driver.
> If it''s the latter, you''d have to run the driver for your
(hardware)
> host controller in the domain so that could be restarted too - the
> appropriate tool for that would be the existing PCI driver domains
> stuff.
>
> What kind of host controller do you have?  (or if you don''t know,
do
> you have USB 2.0?).  Do all other USB devices freeze as well when
> you plug in the MP3 player?  And do you get any interesting output
> from dmesg regarding the error?
Dmesg of boot:

ohci_hcd: 2004 Feb 02 USB 1.1 ''Open'' Host Controller (OHCI)
Driver (PCI)
ACPI: PCI interrupt 0000:00:02.0[A] -> GSI 22 (level, high) -> IRQ 177
ohci_hcd 0000:00:02.0: nVidia Corporation nForce2 USB Controller
PCI: Setting latency timer of device 0000:00:02.0 to 64
ohci_hcd 0000:00:02.0: irq 177, pci mem f896e000
ohci_hcd 0000:00:02.0: new USB bus registered, assigned bus number 1
hub 1-0:1.0: USB hub found
hub 1-0:1.0: 3 ports detected
ACPI: PCI interrupt 0000:00:02.1[B] -> GSI 21 (level, high) -> IRQ 185
ohci_hcd 0000:00:02.1: nVidia Corporation nForce2 USB Controller (#2)
PCI: Setting latency timer of device 0000:00:02.1 to 64
ohci_hcd 0000:00:02.1: irq 185, pci mem f8970000
ohci_hcd 0000:00:02.1: new USB bus registered, assigned bus number 2
hub 2-0:1.0: USB hub found
hub 2-0:1.0: 3 ports detected
ACPI: PCI interrupt 0000:00:02.2[C] -> GSI 20 (level, high) -> IRQ 193
ehci_hcd 0000:00:02.2: nVidia Corporation nForce2 USB Controller
PCI: Setting latency timer of device 0000:00:02.2 to 64
ehci_hcd 0000:00:02.2: irq 193, pci mem f8972000
ehci_hcd 0000:00:02.2: new USB bus registered, assigned bus number 3
PCI: cache line size of 64 is not supported by device 0000:00:02.2
ehci_hcd 0000:00:02.2: USB 2.0 enabled, EHCI 1.00, driver 2004-May-10
hub 3-0:1.0: USB hub found
hub 3-0:1.0: 6 ports detected
ieee1394: Initialized config rom entry `ip1394''
ohci1394: $Rev: 1223 $ Ben Collins <bcollins@debian.org>
ACPI: PCI interrupt 0000:00:0d.0[A] -> GSI 22 (level, high) -> IRQ 177
PCI: Setting latency timer of device 0000:00:0d.0 to 64
ohci1394: fw-host0: OHCI-1394 1.1 (PCI): IRQ=[177]  MMIO=[e2084000-e20847ff] 
Max Packet=[2048]
ohci_hcd 0000:00:02.0: wakeup

Dmesg of successful plugging in of the drive:

usb 3-5: new high speed USB device using address 3
scsi0 : SCSI emulation for USB Mass Storage devices
  Vendor: FUJITSU   Model: MHR2040AT         Rev: 40BA
  Type:   Direct-Access                      ANSI SCSI revision: 02
SCSI device sda: 78140160 512-byte hdwr sectors (40008 MB)
sda: assuming drive cache: write through
 /dev/scsi/host0/bus0/target0/lun0: p1
Attached scsi disk sda at scsi0, channel 0, id 0, lun 0
USB Mass Storage device found at 3

Dmesg of an unsuccessful plugging in of the drive:

usb 3-5: new high speed USB device using address 5
scsi2 : SCSI emulation for USB Mass Storage devices

After that, nothing, no replugging or anything else helps. Only reboot.

-- Naked



-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It''s fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

Nuutti Kotivuori wrote:
> After that, nothing, no replugging or anything else helps. Only reboot.
Neither removing usb modules?

-jkt

-- 
cd /local/pub && more beer > /dev/mouth



-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It''s fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

Jan Kundrát wrote:
> Nuutti Kotivuori wrote:
>> After that, nothing, no replugging or anything else helps. Only
>> reboot.
>
> Neither removing usb modules?
Usb is compiled into the kernel at the moment. I remember that
removing usb-storage didn''t help - removing all the usb modules
might''ve helped, or it might''ve been that I couldn''t
remove them
because they were in use.

-- Naked


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It''s fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel

> I''d say it is just a trivial bug in the kernel USB stuff -
expecting
> some reply to always arrive, when it doesn''t if the drive is
unplugged
> at the unfortunate moment (or the drive dies, as in this case).
I was just wondering if it was in the storage driver, the low level host 
controller driver or maybe the USB core code.  If all USB devices stop 
working, it suggests one of the latter two, otherwise it suggests the former 
and shouldn''t even stop other USB storage devices from working.
> ohci_hcd: 2004 Feb 02 USB 1.1 ''Open'' Host Controller
(OHCI) Driver (PCI)
I didn''t think OHCI was very common on PCs but I guess
> Dmesg of an unsuccessful plugging in of the drive:
>
> usb 3-5: new high speed USB device using address 5
> scsi2 : SCSI emulation for USB Mass Storage devices
Yuk!

The USB storage code has a certain amount of statefulness, so if it''s
hanging
for that drive once, it may stop it working for the rest of the uptime.  If 
the bug is in the USB storage code, you''ll probably find that other
storage
devices still work, otherwise it may be somewhere else.

Cheers,
Mark


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It''s fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/xen-devel