Re: [Xen-devel] [PATCH v3 0/3] map grant refs at pfn = mfn

[Wei Liu <wei.liu2@xxxxxxxxxx>]

On Wed, Aug 06, 2014 at 06:50:04PM +0100, Thomas Leonard wrote:
> On 6 August 2014 16:59, Wei Liu <wei.liu2@xxxxxxxxxx> wrote:
> > Thomas, can you give this patch a quick try? I only compile test it
> > because I don't have handy testing environment at the moment.
> >
> > Apply it to Dom0 kernel.
> >
> > ---8<---
> > From b32aa53ba8b79f0006780549953510930513a3ac Mon Sep 17 00:00:00 2001
> > From: Wei Liu <wei.liu2@xxxxxxxxxx>
> > Date: Wed, 6 Aug 2014 16:50:06 +0100
> > Subject: [PATCH] xen-netback: don't stop kthreads until all in-flight 
> > packets
> >  are processed
> >
> > Reference count the number of packets in host stack, so that we don't
> > stop the deallocation thread too early. If not, we can end up with
> > xenvif_free permanently waiting for deallocation thread to unmap grefs.
> >
> > Signed-off-by: Wei Liu <wei.liu2@xxxxxxxxxx>
> 
> That stops the errors being printed, but I still have a domain I can't 
> destroy:
> 
> $ xl list
> Name                                        ID   Mem VCPUs      State   
> Time(s)
> Domain-0                                     0   512     2     r-----      
> 35.6
> (null)                                       1     0     1     --ps-d       
> 0.1
> 
> I do see this after a bit:
> 
> [  240.062339] INFO: task xenwatch:18 blocked for more than 120 seconds.
> [  240.062404]       Not tainted 3.16.0-rc7-00004-g9d89db9 #1
> [  240.062426] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs"
> disables this message.
> 
> The load average is 1.
> 

If you do "ps aux | grep vif" do you see some kthreads? If so, what is
the status of these threads? What's the status of "xenwatch" thread?  It
looks like that xenbus thread hangs waiting to be woken up, which means
there's yet another exit point.

This problem involes at least four execution entities (softirq +
kthreads) so to be honest it's not straightforward to figure out what
goes wrong. I will take another stab as soon as possible.

Wei.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel