* GPU Passthrough without VT-d
@ 2014-03-07 3:03 Dushyant Behl
2014-03-07 7:21 ` Kai Huang
0 siblings, 1 reply; 6+ messages in thread
From: Dushyant Behl @ 2014-03-07 3:03 UTC (permalink / raw)
To: xen-devel
[-- Attachment #1.1: Type: text/plain, Size: 371 bytes --]
Hi all,
I am new to the xen community and I wanted to work on the project "Allowing
guests to boot with a passed-through GPU as the primary display" but the
processor in my system does not supports IO MMU Virtualization (Intel
VT-d), it only supports the Intel Virtualization (Intel VT-x). Would it be
possible for me to work on the project in anyway?
Thanks,
Dushyant
[-- Attachment #1.2: Type: text/html, Size: 465 bytes --]
[-- Attachment #2: Type: text/plain, Size: 126 bytes --]
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: GPU Passthrough without VT-d
2014-03-07 3:03 GPU Passthrough without VT-d Dushyant Behl
@ 2014-03-07 7:21 ` Kai Huang
2014-03-07 17:36 ` Dushyant Behl
0 siblings, 1 reply; 6+ messages in thread
From: Kai Huang @ 2014-03-07 7:21 UTC (permalink / raw)
To: Dushyant Behl, xen-devel
In general you cannot do pci passthrough for HVM without IOMMU (vt-d)
support, as we need VT-d to support guest physical address to machine
address translation for DMA.
However, theoretically, you can do it by hacking Xen to setup identical
(1:1) mappings for guest's memory, but it's not a pretty way and you
should expect to suffer lots of tricky corner case bugs.
Thanks,
-Kai
On 03/07/2014 11:03 AM, Dushyant Behl wrote:
> Hi all,
>
> I am new to the xen community and I wanted to work on the project
> "Allowing guests to boot with a passed-through GPU as the primary
> display" but the processor in my system does not supports IO MMU
> Virtualization (Intel VT-d), it only supports the Intel Virtualization
> (Intel VT-x). Would it be possible for me to work on the project in anyway?
>
> Thanks,
> Dushyant
>
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> http://lists.xen.org/xen-devel
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: GPU Passthrough without VT-d
2014-03-07 7:21 ` Kai Huang
@ 2014-03-07 17:36 ` Dushyant Behl
2014-03-07 19:22 ` Pasi Kärkkäinen
0 siblings, 1 reply; 6+ messages in thread
From: Dushyant Behl @ 2014-03-07 17:36 UTC (permalink / raw)
To: Kai Huang; +Cc: xen-devel
[-- Attachment #1.1: Type: text/plain, Size: 1153 bytes --]
Thanks for the quick reply. I'll try to look into it.
-
Dushyant
On Fri, Mar 7, 2014 at 12:51 PM, Kai Huang <kai.k.huang@linux.intel.com>wrote:
> In general you cannot do pci passthrough for HVM without IOMMU (vt-d)
> support, as we need VT-d to support guest physical address to machine
> address translation for DMA.
>
> However, theoretically, you can do it by hacking Xen to setup identical
> (1:1) mappings for guest's memory, but it's not a pretty way and you should
> expect to suffer lots of tricky corner case bugs.
>
> Thanks,
> -Kai
>
>
> On 03/07/2014 11:03 AM, Dushyant Behl wrote:
>
>> Hi all,
>>
>> I am new to the xen community and I wanted to work on the project
>> "Allowing guests to boot with a passed-through GPU as the primary
>> display" but the processor in my system does not supports IO MMU
>> Virtualization (Intel VT-d), it only supports the Intel Virtualization
>> (Intel VT-x). Would it be possible for me to work on the project in
>> anyway?
>>
>> Thanks,
>> Dushyant
>>
>>
>> _______________________________________________
>> Xen-devel mailing list
>> Xen-devel@lists.xen.org
>> http://lists.xen.org/xen-devel
>>
>>
[-- Attachment #1.2: Type: text/html, Size: 1885 bytes --]
[-- Attachment #2: Type: text/plain, Size: 126 bytes --]
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: GPU Passthrough without VT-d
2014-03-07 17:36 ` Dushyant Behl
@ 2014-03-07 19:22 ` Pasi Kärkkäinen
2014-03-08 5:01 ` Pasi Kärkkäinen
0 siblings, 1 reply; 6+ messages in thread
From: Pasi Kärkkäinen @ 2014-03-07 19:22 UTC (permalink / raw)
To: Dushyant Behl; +Cc: Kai Huang, xen-devel
On Fri, Mar 07, 2014 at 11:06:34PM +0530, Dushyant Behl wrote:
> Thanks for the quick reply. I'll try to look into it.
>
There's an old patch to do that in the Xen wiki..
-- Pasi
> -
> Dushyant
>
> On Fri, Mar 7, 2014 at 12:51 PM, Kai Huang
> <[1]kai.k.huang@linux.intel.com> wrote:
>
> In general you cannot do pci passthrough for HVM without IOMMU (vt-d)
> support, as we need VT-d to support guest physical address to machine
> address translation for DMA.
>
> However, theoretically, you can do it by hacking Xen to setup identical
> (1:1) mappings for guest's memory, but it's not a pretty way and you
> should expect to suffer lots of tricky corner case bugs.
>
> Thanks,
> -Kai
>
> On 03/07/2014 11:03 AM, Dushyant Behl wrote:
>
> Hi all,
>
> I am new to the xen community and I wanted to work on the project
> "Allowing guests to boot with a passed-through GPU as the primary
> display" but the processor in my system does not supports IO MMU
> Virtualization (Intel VT-d), it only supports the Intel Virtualization
> (Intel VT-x). Would it be possible for me to work on the project in
> anyway?
>
> Thanks,
> Dushyant
>
> _______________________________________________
> Xen-devel mailing list
> [2]Xen-devel@lists.xen.org
> [3]http://lists.xen.org/xen-devel
>
> References
>
> Visible links
> 1. mailto:kai.k.huang@linux.intel.com
> 2. mailto:Xen-devel@lists.xen.org
> 3. http://lists.xen.org/xen-devel
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> http://lists.xen.org/xen-devel
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: GPU Passthrough without VT-d
2014-03-07 19:22 ` Pasi Kärkkäinen
@ 2014-03-08 5:01 ` Pasi Kärkkäinen
2014-03-11 21:08 ` Dushyant Behl
0 siblings, 1 reply; 6+ messages in thread
From: Pasi Kärkkäinen @ 2014-03-08 5:01 UTC (permalink / raw)
To: Dushyant Behl; +Cc: Kai Huang, xen-devel
On Fri, Mar 07, 2014 at 09:22:32PM +0200, Pasi Kärkkäinen wrote:
> On Fri, Mar 07, 2014 at 11:06:34PM +0530, Dushyant Behl wrote:
> > Thanks for the quick reply. I'll try to look into it.
> >
>
> There's an old patch to do that in the Xen wiki..
>
See here:
http://old-list-archives.xenproject.org/archives/html/xen-devel/2009-09/msg00018.html
-- Pasi
>
> > -
> > Dushyant
> >
> > On Fri, Mar 7, 2014 at 12:51 PM, Kai Huang
> > <[1]kai.k.huang@linux.intel.com> wrote:
> >
> > In general you cannot do pci passthrough for HVM without IOMMU (vt-d)
> > support, as we need VT-d to support guest physical address to machine
> > address translation for DMA.
> >
> > However, theoretically, you can do it by hacking Xen to setup identical
> > (1:1) mappings for guest's memory, but it's not a pretty way and you
> > should expect to suffer lots of tricky corner case bugs.
> >
> > Thanks,
> > -Kai
> >
> > On 03/07/2014 11:03 AM, Dushyant Behl wrote:
> >
> > Hi all,
> >
> > I am new to the xen community and I wanted to work on the project
> > "Allowing guests to boot with a passed-through GPU as the primary
> > display" but the processor in my system does not supports IO MMU
> > Virtualization (Intel VT-d), it only supports the Intel Virtualization
> > (Intel VT-x). Would it be possible for me to work on the project in
> > anyway?
> >
> > Thanks,
> > Dushyant
> >
> > _______________________________________________
> > Xen-devel mailing list
> > [2]Xen-devel@lists.xen.org
> > [3]http://lists.xen.org/xen-devel
> >
> > References
> >
> > Visible links
> > 1. mailto:kai.k.huang@linux.intel.com
> > 2. mailto:Xen-devel@lists.xen.org
> > 3. http://lists.xen.org/xen-devel
>
> > _______________________________________________
> > Xen-devel mailing list
> > Xen-devel@lists.xen.org
> > http://lists.xen.org/xen-devel
>
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> http://lists.xen.org/xen-devel
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: GPU Passthrough without VT-d
2014-03-08 5:01 ` Pasi Kärkkäinen
@ 2014-03-11 21:08 ` Dushyant Behl
0 siblings, 0 replies; 6+ messages in thread
From: Dushyant Behl @ 2014-03-11 21:08 UTC (permalink / raw)
To: Pasi Kärkkäinen; +Cc: xen-devel
[-- Attachment #1.1: Type: text/plain, Size: 2758 bytes --]
Sorry I was outside so couldn't reply. I found the patch earlier but as
this would be my first project so i was a little conscious about the corner
case bugs.
I was also planning to buy some new hardware so i though i could give it a
go then, Nevertheless i'll try to use the patch in the meantime.
Thanks,
Dushyant
On Sat, Mar 8, 2014 at 10:31 AM, Pasi Kärkkäinen <pasik@iki.fi> wrote:
> On Fri, Mar 07, 2014 at 09:22:32PM +0200, Pasi Kärkkäinen wrote:
> > On Fri, Mar 07, 2014 at 11:06:34PM +0530, Dushyant Behl wrote:
> > > Thanks for the quick reply. I'll try to look into it.
> > >
> >
> > There's an old patch to do that in the Xen wiki..
> >
>
> See here:
>
> http://old-list-archives.xenproject.org/archives/html/xen-devel/2009-09/msg00018.html
>
> -- Pasi
>
> >
> > > -
> > > Dushyant
> > >
> > > On Fri, Mar 7, 2014 at 12:51 PM, Kai Huang
> > > <[1]kai.k.huang@linux.intel.com> wrote:
> > >
> > > In general you cannot do pci passthrough for HVM without IOMMU
> (vt-d)
> > > support, as we need VT-d to support guest physical address to
> machine
> > > address translation for DMA.
> > >
> > > However, theoretically, you can do it by hacking Xen to setup
> identical
> > > (1:1) mappings for guest's memory, but it's not a pretty way and
> you
> > > should expect to suffer lots of tricky corner case bugs.
> > >
> > > Thanks,
> > > -Kai
> > >
> > > On 03/07/2014 11:03 AM, Dushyant Behl wrote:
> > >
> > > Hi all,
> > >
> > > I am new to the xen community and I wanted to work on the
> project
> > > "Allowing guests to boot with a passed-through GPU as the
> primary
> > > display" but the processor in my system does not supports IO MMU
> > > Virtualization (Intel VT-d), it only supports the Intel
> Virtualization
> > > (Intel VT-x). Would it be possible for me to work on the
> project in
> > > anyway?
> > >
> > > Thanks,
> > > Dushyant
> > >
> > > _______________________________________________
> > > Xen-devel mailing list
> > > [2]Xen-devel@lists.xen.org
> > > [3]http://lists.xen.org/xen-devel
> > >
> > > References
> > >
> > > Visible links
> > > 1. mailto:kai.k.huang@linux.intel.com
> > > 2. mailto:Xen-devel@lists.xen.org
> > > 3. http://lists.xen.org/xen-devel
> >
> > > _______________________________________________
> > > Xen-devel mailing list
> > > Xen-devel@lists.xen.org
> > > http://lists.xen.org/xen-devel
> >
> >
> > _______________________________________________
> > Xen-devel mailing list
> > Xen-devel@lists.xen.org
> > http://lists.xen.org/xen-devel
>
[-- Attachment #1.2: Type: text/html, Size: 4319 bytes --]
[-- Attachment #2: Type: text/plain, Size: 126 bytes --]
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2014-03-11 21:08 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2014-03-07 3:03 GPU Passthrough without VT-d Dushyant Behl
2014-03-07 7:21 ` Kai Huang
2014-03-07 17:36 ` Dushyant Behl
2014-03-07 19:22 ` Pasi Kärkkäinen
2014-03-08 5:01 ` Pasi Kärkkäinen
2014-03-11 21:08 ` Dushyant Behl
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).