Yes, the bug shows at cpu hot-plug patch. What I mean is that the root cause is
some improper handling of MSI/MSI-X on domain-destruction path.
Shan Haitao
-----Original Message-----
From: Keir Fraser [mailto:keir.fraser@xxxxxxxxxxxxx]
Sent: 2008年7月1日 20:44
To: Shan, Haitao; Li, Haicheng; xen-devel@xxxxxxxxxxxxxxxxxxx
Cc: Jiang, Yunhong; Han, Weidong
Subject: Re: [Xen-devel] VMX status report. Xen: #17917 & Xen0: #583 -- no new
issue
Well actually this is not on the domain-destruction path -- it's on the cpu
hotplug path. But anyhow if you have some bugfixes in the pipeline, that's
great!
-- Keir
On 1/7/08 13:37, "Shan, Haitao" <haitao.shan@xxxxxxxxx> wrote:
> Yes. It is caused by improper MSI cleaning up code for domain-destroy code
> path. I have made a patch and basically it can solve the problem. I will test
> it more and send it out tomorrow.
>
> Shan Haitao
>
> -----Original Message-----
> From: Keir Fraser [mailto:keir.fraser@xxxxxxxxxxxxx]
> Sent: 2008年7月1日 17:57
> To: Li, Haicheng; xen-devel@xxxxxxxxxxxxxxxxxxx
> Cc: Jiang, Yunhong; Han, Weidong; Shan, Haitao
> Subject: Re: [Xen-devel] VMX status report. Xen: #17917 & Xen0: #583 -- no new
> issue
>
> On 1/7/08 10:24, "Li, Haicheng" <haicheng.li@xxxxxxxxx> wrote:
>
>> 2. Xen HV crashes while doing dom0 S3.
>> http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=1279
>
> I expect this would crash with host S5 as well (i.e., /sbin/poweroff in
> dom0). The crash is clearly within MSI-related code, so one of the
> developers of those patches should take a look: e.g., Yunhong Jiang, Haitao
> Shan, or Weidong Han. My guess is if one of them can repro this issue then
> the fix will be simple.
>
> -- Keir
>
>
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
|
Home