[Spice-devel] Vdagent not working on xen linux hvm DomUs

[Fabio Fantoni]

Il 10/12/2013 15:38, Fabio Fantoni ha scritto:
> Il 06/12/2013 16:39, Wei Liu ha scritto:
>> On Fri, Dec 06, 2013 at 02:55:46PM +0000, Wei Liu wrote:
>>> On Fri, Dec 06, 2013 at 03:49:30PM +0100, Fabio Fantoni wrote:
>>>> Il 06/12/2013 15:40, Wei Liu ha scritto:
>>>>> On Fri, Dec 06, 2013 at 02:51:50PM +0100, Fabio Fantoni wrote:
>>>>>> Il 06/12/2013 14:17, Wei Liu ha scritto:
>>>>>>> On Fri, Dec 06, 2013 at 02:10:24PM +0100, Fabio Fantoni wrote:
>>>>>>> [...]
>>>>>>>> I tried 2 cases with xen_emul_unplug=never:
>>>>>>>> - with xen platform disabled (xen_platform_pci=0 plus konrad 
>>>>>>>> patch)
>>>>>>>> and pci=nomsi
>>>>>>>> - with xen_platform_pci=1 and without pci kernel parameter
>>>>>>>> on both cases qemu crash with same error:
>>>>>>>> xen_ram_addr_from_mapcache, could not find 0x7fdecf624968
>>>>>>>>
>>>>>>>> On your virtio net test you have added only 'model=virtio-net-pci'
>>>>>>>> on vif line of domU's xl cfg or you did other changes?
>>>>>>>>
>>>>>>> No, nothing more. But I'm using Xen's QEMU upstream, not vanilla 
>>>>>>> QEMU.
>>>>>> I'm using xen's upstream qemu (master of
>>>>>> git://xenbits.xen.org/qemu-upstream-unstable.git)
>>>>>>
>>>>> Not only the branch is important but also the changeset.
>>>>>
>>>>> I'm using the hash specified in Config.mk
>>>> qemu of my tests about virtio:
>>>> git log
>>>> commit b97307ecaad98360f41ea36cd9674ef810c4f8cf
>>>> Author: Matthew Daley <mattjd at gmail.com>
>>>> Date:   Thu Oct 10 14:10:48 2013 +0000
>>>>
>>>>      xen_disk: mark ioreq as mapped before unmapping in error case
>>>>
>>> There's lots of changesets between this one and the one I use so doing
>>> bisection is the only way to find out where the regression was
>>> introduced.
>>>
>> And, if you're really going to look into this I suggest you start with
>> virtio related changesets, then Xen mapcache changesets (I suspect
>> there's many changes to this code, but anyway it is worth looking at).
>>
>> Good luck.
>>
>> Wei.
>
> Thanks for your reply.
> Before starting bisection I tried with qemu 1.3.1 from 
> qemu-upstream-4.3-testing.git
> No more crash with virtio net but it needs pci=nomsi to be working, 
> same thing for vdagent, so seems that msi problem is with all virtio 
> devices.
> Then the problems seem 2 different, on your build you have virtio 
> devices working without setting pci=nomsi need to know the differences 
> and find the cause.

Your test with virtio net working without pci=nomsi was on ovmf only or 
you tried also with seabios?
I tested with Ubuntu Saucy and Ubuntu Precise, both with latest 
xen-unstable (based on commit 2f718161bc292bfbdf1aeefd3932b73c0965373d), 
latest commit of qemu-upstream-4.3-testing.git and latest stable seabios 
from debian package 1.7.3-2
On both case pci=nomsi was needed to have virtio net working.

I watch the pdf of virtio spec. of this post:
http://lists.xen.org/archives/html/xen-devel/2013-12/msg01654.html
however, are not able to understand the possible cause of the problem 
encountered with msi on virtio devices with xen.

> About the crash of qemu 1.6.1 with virtio net is confirmed that is a 
> regression, is not critical because is not implement on libxl now but 
> I'll do further research.
>
>

I test with qemu 1.4 and 1.5 and they haven't the regression showing 
xenmap cache error with virtio net.
Watching history seems there aren't commits about xen mapcache between 
1.5 and 1.6, other xen and virtio changes are many, from a quick look I 
could not find commit suspects to be tested.
Someone can suggest me the commits more suspects to be testedplease?

Thanks for any reply.