Ahh see i thought i had missed something ... here is the qemu log attached.
I will try the changeset and 2.6.18 kernel as suggested.
Tuesday, December 1, 2009, 10:11:14 AM, you wrote:
> Can you post qemu log (/var/log/xen/qemu-dm-xxx.log)?
> Maybe need to rebase the patch on latest xen with pv-ops, I will look at it
> in several days. I you want to have a quick try, I suggest you to use xen
> changeset 20145, 2.6.18 dom0, and qemu commit: a05958b6e. It worked stably
> for me with IGD pass-through.
> -----Original Message-----
> From: Sander Eikelenboom [mailto:linux@xxxxxxxxxxxxxx]
> Sent: Tuesday, December 01, 2009 2:24 AM
> To: Han, Weidong
> Subject: Re: [Xen-devel] Partial succes VGA passthrough Ati Radeon HD4850 on
> Asus P5Q-EM DO (Q45 chipset)
> I previously tried passing through both functions of the IGD, but from what i
> saw from Tom Rotenberg, i now tried only passing through 00:02.0.
> It now doesn't continue to spew error messages, but the guest domain didn't
> completely start either. After the errors stopped i was able to switch to the
> hypervisor and get a full dump.
> The IGD is at this moment the primary (boot) display adapter (and the only
> display adapter present), could this be a problem and should a boot from a
> pci adapter as primary, or is that not required ?
> - serial console output + hypervisor dump is attached
> - "xm info" with the changeset is attached
> - QEMU_TAG ?= f72b6e0ffc3bb84d4442c5a7493bffbdce2a4468 from Config.mk
> - lspci -vvvknn output is attached
> - .cfg of HVM is attached
> I hope i have given you all the right info this time.
> Monday, November 30, 2009, 2:05:59 AM, you wrote:
>> Pls provide xen changeset, dom0 and qemu commit, and qemu log
>> -----Original Message-----
>> From: Sander Eikelenboom [mailto:linux@xxxxxxxxxxxxxx]
>> Sent: Sunday, November 29, 2009 6:55 PM
>> To: Han, Weidong
>> Cc: xen-devel@xxxxxxxxxxxxxxxxxxx
>> Subject: Re: [Xen-devel] Partial succes VGA passthrough Ati Radeon HD4850 on
>> Asus P5Q-EM DO (Q45 chipset)
>> Hmm less succes with the IGD :(, it keeps (infinite) giving IOMMU faults, in
>> a rate which locks up the box.
>> I have experienced this before, but didn't have serial console.
>> But since i do now i have something to report :-)
Description: Binary data
Xen-devel mailing list