Re: xen-balloon thread using 100% of CPU, regression in 5.4.150

[Jason Andryuk <jandryuk@xxxxxxxxx>]

On Tue, Oct 5, 2021 at 4:05 AM Juergen Gross <jgross@xxxxxxxx> wrote:
>
> On 04.10.21 11:14, Marek Marczykowski-Górecki wrote:
> > On Mon, Oct 04, 2021 at 07:31:40AM +0200, Juergen Gross wrote:
> >> On 03.10.21 06:47, Marek Marczykowski-Górecki wrote:
> >>> Hi,
> >>>
> >>> After updating a PVH domU to 5.4.150, I see xen-balloon thread using
> >>> 100% CPU (one thread).
> >>> This is a domain started with memory=maxmem=716800KiB (via libvirt). Then,
> >>> inside, I see:
> >>>
> >>> # cat /sys/devices/system/xen_memory/xen_memory0/target_kb
> >>> 716924
> >>> # cat /sys/devices/system/xen_memory/xen_memory0/info/current_kb
> >>> 716400
> >>>
> >>> Doing `cat info/current_kb > target_kb` "fixes" the issue. But still,
> >>> something is wrong - on earlier kernel (5.4.143 to be precise), it
> >>> wasn't spinning, with exactly the same values reported in sysfs. It
> >>> shouldn't run in circles if it can't get that much memory it wants. I
> >>> strongly suspect "xen/balloon: use a kernel thread instead a workqueue"
> >>> or related commit being responsible, but I haven't verified it.
> >>
> >> I think you are right. I need to handle the BP_ECANCELED case similar to
> >> BP_EAGAIN in the kernel thread (wait until target size changes again).
> >>
> >> One further question: do you see any kernel message in the guest related
> >> to the looping balloon thread?
> >
> > Nothing, only the usual "xen:balloon: Initialising balloon driver", and
> > nothing related to balloon after that.
>
> Could you try the attached patch, please? I've tested it briefly with
> PV and PVH guests.

I was seeing the CPU spinning in dom0 with xen command line:
dom0_mem=min:420M,max:420M,420M

Your patch eliminated the CPU spinning.

Tested-by: Jason Andryuk <jandryuk@xxxxxxxxx>

Thanks, Juergen

-Jason