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

RE: [Xen-devel] GPLPV Drivers 0.9.11-pre4 hang during boot - WindowsXP Pro SP2


  • To: "Sandro Sigala" <sandro@xxxxxxxxx>
  • From: "James Harper" <james.harper@xxxxxxxxxxxxxxxx>
  • Date: Fri, 4 Jul 2008 22:39:12 +1000
  • Cc: xen-devel@xxxxxxxxxxxxxxxxxxx
  • Delivery-date: Fri, 04 Jul 2008 05:39:40 -0700
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>
  • Thread-index: Acjdy34PxHulzdfvTQ++T4sW11TCQAAA9wkw
  • Thread-topic: [Xen-devel] GPLPV Drivers 0.9.11-pre4 hang during boot - WindowsXP Pro SP2

> James Harper wrote:
> > I have just recompiled and uploaded the correct version. Sorry for
> > giving you the run around. Hopefully we'll get to the bottom of this
> > yet!
> 
> Happens! :-)
> 
> Attached there are the new logs with /GPLPV and without.

Do you have any timestamps to go with your logs? Don't worry about
resubmitting these ones but on any future ones it would be useful.

It looks like requests being put on the ring are not being actioned...
maybe we are corrupting something somewhere.

Looking at this bit of the logs:

"
XenVbd --- HwScsiStartIo (Still figuring out ring)
"

That 'Still figuring out ring' bit means that we got a request before we
had detected whether we were talking to a 32 or 64 bit ring, so we told
windows to give us the request back again later when we were ready. I
haven't seen that before in my logs, so something about your system
means that windows is giving you a request at an earlier point than my
test systems...

Then we see this:

"
XenVbd     ring_detect_state = 0, operation = ff, id = 0, status = 0
XenVbd     ring_detect_state = 1, operation = 0, id = ff, status = 0
"

Which means that the first response we got back had operation ff which
is what we expect, but the second one had operaton = 0 and id = ff which
means that the ring is out of alignment and we must be using the
opposite ring size (64 bits in this case).

Originally I thought that maybe Windows wasn't re-issuing the original
request, but obviously it is:

"
XenVbd --> HwScsiStartIo PathId = 0, TargetId = 0, Lun = 0
XenVbd     SRB_FUNCTION_EXECUTE_SCSI
XenVbd     Command = READ/WRITE
XenVbd --> XenVbd_PutSrbOnRing
XenVbd <-- XenVbd_PutSrbOnRing
"

But then it is never getting a response so it resets the bus:

"
XenVbd --> HwScsiResetBus
XenVbd     IRQL = 22
XenVbd <-- HwScsiResetBus
"

I wonder if it is a problem under XP... I'll see what I can do about
setting up a new XP test environment (my old one broke :)

Otherwise, I'm out of ideas for the moment.

James


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel


 


Rackspace

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