Re: [PATCH v3] migration: clear the memory region dirty bitmap when skipping free pages

[David Hildenbrand]

On 22.07.21 16:51, Peter Xu wrote:
> On Thu, Jul 22, 2021 at 09:57:13AM +0000, Wang, Wei W wrote:
> > On Thursday, July 22, 2021 5:48 PM, David Hildenbrand wrote:
> > > On 22.07.21 10:30, Wei Wang wrote:
> > > > When skipping free pages to send, their corresponding dirty bits in
> > > > the memory region dirty bitmap need to be cleared. Otherwise the
> > > > skipped pages will be sent in the next round after the migration
> > > > thread syncs dirty bits from the memory region dirty bitmap.
> > > >
> > > > Cc: David Hildenbrand <david@redhat.com>
> > > > Cc: Peter Xu <peterx@redhat.com>
> > > > Cc: Michael S. Tsirkin <mst@redhat.com>
> > > > Reported-by: David Hildenbrand <david@redhat.com>
> > > > Signed-off-by: Wei Wang <wei.w.wang@intel.com>
> > > > ---
> > > >    migration/ram.c | 74
> > > +++++++++++++++++++++++++++++++++++++------------
> > > >    1 file changed, 56 insertions(+), 18 deletions(-)
> > >
> > > LGTM, thanks
> > >
> > > Reviewed-by: David Hildenbrand <david@redhat.com>
> >
> > Thanks. Please remember to have a regression test together with Peterx's that 
> > patch when you get a chance.
>
> I can continue to try that; but it's slightly low priority to me so it could be
> a bit delayed.  If either of you could help that would be great, as I still
> don't know last time why I didn't use free-page-hint right (note: I definitely
> checked lsmod for sure; so it's there).  So I'll need to figure that out first.

Just for the records:

Add the following to your QEMU cmdline to make it work:

    -object iothread,id=t0 \
    -device virtio-balloon-pci,id=balloon,iothread=t0,free-page-hint=on

Most upstream distributions (e.g., Fedora) should have virtio-balloon 
build as module and there is no action inside the guest required to get 
it running. Just wait until Linux in the guest booted up.

--
Thanks,

David / dhildenb