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

Re: [Xen-users] [Xen-devel] Intel HD Graphics 4600 Xen VGA Passthrough to Windows 7 Ultimate HVM domU



Just to provide another input:
My first try with a IGD4000 passthrough to win7 in a 3GB RAM domU
failed with an old XEN version (about 4.1.3?).
I then step back to use 2GB without issue.
Yesterday, I decided to have another try after moving to XEN 4.3.0.
This time 3GB works for me.
I just read another thread suggesting that this may have something to
do with the PCI address space occupied by the device you passed
through.
The bigger the occupation is, the more likely to fail. I'm not sure if
it has anything to do with the memory allocation to the IGD.

My issue is that the panel constantly flashes rapidly with different
colors once the guest starts.
It lasts until the guest loads the display driver. Quite annoy, but
still usable.

On Thu, Jul 25, 2013 at 10:51 PM, Gordan Bobic <gordan@xxxxxxxxxx> wrote:
> On Thu, 25 Jul 2013 09:41:17 -0500, David Sutton <kantras@xxxxxxxxx> wrote:
>>
>> All,
>>
>> On Thu, Jul 25, 2013 at 4:20 AM, Gordan Bobic  wrote:
>>
>> On Thu, 25 Jul 2013 12:07:25 +0300, Pasi Kärkkäinen  wrote:
>>
>>   On Thu, Jul 25, 2013 at 08:58:58AM +0000, Ren, Yongjie wrote:
>>   > > N.B. I didn't use the latest traditional qemu because of the
>> following bug:
>>  > >
>>
>> http://bugzilla-archived.xenproject.org//bugzilla/show_bug.cgi?id=1865
>> [3]
>>
>>  > >
>>  >
>>  > Hmm.. isn't that bug affecting qemu-traditional in general, not
>> PCI/GPU
>>  > passthru specifically?
>>  > Or is it *only* PCI/GPU passthru that it breaks for you?
>>  >
>>  It's a general device assignment issue.
>>  Any PCI device can't be assigned to a HVM guest with the latest
>> qemu-traditional.
>>
>>  Ok. Is there already a separate thread on xen-devel about the
>>  qemu-traditional PCI assignment issue?
>>  If not, we should create one.
>>
>>  FWIW, I'm using qemu-dm (i.e. traditional) in 4.3 and am successfully
>>  passing USB, audio, GPU PCI devices using it (as long as I stay under
>>  the 2GB domU RAM limit). So whatever the issue is it isn't generic.
>>
>>  That bug report doesn't include the domU config file, and IIRC
>>  I saw a similar error when device_model/device_model_override and
>>  device_model_version don't match or are unavailable in the build
>>  (e.g. building without upstream and specifying traditional or
>>  vice versa). More info needed to verify, I think.
>>
>> And just to add a little more variety, I'm also using traditional
>> (under 4.3) to pass through USB and GPU (Radeon 6770) to a Windows 7
>> DomU configured with 6Gb for memory. Only using a patch to work around
>> a broken BIOS implementation so I have IOMMU support.
>
>
> And you don't suffer from a memory stomp resulting
> in frame buffer corruption shortly followed by a domU
> crash with that much RAM passed to domU?
>
>
> Gordan
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@xxxxxxxxxxxxx
> http://lists.xen.org/xen-devel

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


 


Rackspace

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