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

Re: [Xen-devel] xsa46-4.2.patch breaks PCI passthrough?

>>> On 08.05.13 at 12:18, Steven Haigh <netwiz@xxxxxxxxx> wrote:
> On 6/05/2013 5:15 PM, Jan Beulich wrote:
>>>>> On 05.05.13 at 12:53, Steven Haigh <netwiz@xxxxxxxxx> wrote:
>>> Debug log is now attached to the bug report. Its a little large to
>>> attach here.
>>> http://xen.crc.id.au/bugs/file_download.php?file_id=13&type=bug 
>>> (XEN) **DBG perms { 16, 1 } = 0
>> I'm surprised by this if the test was done without the xend
>> adjustment, whereas this
>>> (XEN) **DBG perms { 34, 1 } = -22
>>> (XEN)  Domain 2, nr_pirqs 80
>>> (XEN)  dom_pirq_to_irq(34) = 0
>> is expected without a prior physdev_map_pirq() invocation. I'm
>> meanwhile guessing that there might be a second place in xend
>> where under some condition that call is being issued - that could
>> also explain the -EEXIST observed with the xend adjustment in
>> place.
> I'll be the first to admit that this is beyond my knowledge in Xen at 
> this low a level... Is there anything I can do to help debugging progress?

See the other thread ("PCI passthrough problems after legacy
update of xen 4.1") about the same problem. Andreas had put
together a debugging patch that helped narrow it, but it's still
unclear where the conflicting hypercalls originate. Perhaps
continuing the discussion centrally in that other thread (which
has made better progress) would help keep all information


Xen-devel mailing list



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