On 05/06/2011 07:12 PM, Konrad Rzeszutek Wilk wrote:
On Fri, May 06, 2011 at 06:16:26PM +0200, Igor Mammedov wrote:
On 05/05/2011 11:28 PM, Konrad Rzeszutek Wilk wrote:
On Wed, May 04, 2011 at 05:13:20PM -0400, Konrad Rzeszutek Wilk wrote:
On Wed, May 04, 2011 at 11:03:04PM +0200, Igor Mammedov wrote:
On 05/04/2011 10:19 PM, Konrad Rzeszutek Wilk wrote:
On Wed, May 04, 2011 at 09:58:01PM +0200, Igor Mammedov wrote:
On restore 32bit PV guest crashes on x64_86 host, while x64_86 PV guest
works as expected.
I tried this with 2.6.39-rc6 and could not reproduce this. Does it work
for you with 2.6.39-rc6? Or is it _only_ failing with 2.6.38?
It's failing with 2.6.39-rc6 as well. And 2.6.38 with suggested
patches as well.
But since I do like fast builds, I've used fast PC with 2 NUMA nodes
at least and
Hmm, the guest is a single CPU one, so perhaps you are hitting a hypervisor
issue. What happens if you turn NUMA off on the hypervisor line?
Looks like removing numa=on from hypervisor init line helped, guest
stays alive after restore.
it turns out that the bug is not reproducible on PC with only one
cpu (HPZ800
1 socket config). So you can reproduce it if you try a PC with at
least 2 cpus
nodes.
Ok, I've an old IBM x366 that has four sockets - that should do it -
but will have to wait till next week when I am back from traveling.
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
|