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

[Xen-devel] Re: [xen-unstable test] 2549: regressions - FAIL


  • To: Ian Jackson <Ian.Jackson@xxxxxxxxxxxxx>, Shan Haitao <haitao.shan@xxxxxxxxx>, Han Weidong <weidong.han@xxxxxxxxx>
  • From: Keir Fraser <keir@xxxxxxx>
  • Date: Thu, 04 Nov 2010 16:33:22 +0000
  • Cc: "xen-devel@xxxxxxxxxxxxxxxxxxx" <xen-devel@xxxxxxxxxxxxxxxxxxx>
  • Delivery-date: Thu, 04 Nov 2010 09:34:27 -0700
  • Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=sender:user-agent:date:subject:from:to:cc:message-id:thread-topic :thread-index:in-reply-to:mime-version:content-type :content-transfer-encoding; b=JgiYFkRGgz5XS9Anzs+p7fFE7bSf+Jq3MNkoZ79dCpLFwz8zv2/HFJ5W9vFwwifLGD kQPI7+O5Sn6Ivg6JaL3tkw8b7JrpndqXR6Ms8o65xjq4313848q+gKQcn4d1PnU/VYjp pxXbw6VR/IY8rRCp9P4av8TdMfgE08rXRvyas=
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>
  • Thread-index: Act8Ohakx/+wsZiE00WZEwH6RbdtKQAA+gqR
  • Thread-topic: [xen-unstable test] 2549: regressions - FAIL

On 04/11/2010 16:05, "Keir Fraser" <keir@xxxxxxx> wrote:

>> xc: error: No extended context for VCPU133120 (3 = No such process): Internal
>> error
>> xc: error: Failed to map/save the p2m frame list (3 = No such process):
>> Internal error
>> libxl: error: libxl_dom.c:448:libxl__domain_suspend_common saving domain: No
>> such process
>> xl: fatal error: xl_cmdimpl.c:2610: Unknown error 4294967294:
>> libxl_domain_suspend(&ctx, NULL, domid, fd)
>> 
>> A similar failure occurs with xend and appears in the xend logfile.
> 
> Just as well we didn't check it in to 4.0-testing too, eh?

Obviously the whole patchset was never tested in the common case of xsave
being unsupported or disabled. I have reverted the save/restore patch, and I
have also reverted patch hunks which make XSAVE/AVX visible to guests. The
reason for this is that the code which determines guest CPUID is based on
reading host CPUID -- this will report XSAVE/AVX as available on any CPU
that supports it, even if it has actually been disabled by use_xsave=0 in
the hypervisor (which is the default!). So the guest will try to use the
feature, and at the very least will get its state trashed across context
switches and save/restore (because Xen sees !cpu_has_xsave). Very nice.

 -- Keir



_______________________________________________
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®.