Re: [Qemu-devel] [PATCH v2] migration: skip sending ram pages released by virtio-balloon driver.

[Paolo Bonzini]

On 28/03/2016 08:59, Michael S. Tsirkin wrote:
>> > +    qemu_mutex_lock_balloon_bitmap();
>> >      for (;;) {
>> >          size_t offset = 0;
>> >          uint32_t pfn;
>> >          elem = virtqueue_pop(vq, sizeof(VirtQueueElement));
>> >          if (!elem) {
>> > +            qemu_mutex_unlock_balloon_bitmap();
>> >              return;
>> >          }
>> >  
>> > @@ -242,6 +245,7 @@ static void virtio_balloon_handle_output(VirtIODevice 
>> > *vdev, VirtQueue *vq)
>> >              addr = section.offset_within_region;
>> >              balloon_page(memory_region_get_ram_ptr(section.mr) + addr,
>> >                           !!(vq == s->dvq));
>> > +            qemu_balloon_bitmap_update(addr, !!(vq == s->dvq));
>> >              memory_region_unref(section.mr);
>> >          }
>> >  
> So the assumption here is that offset_within_region equals
> ram ptr if region is get_system_memory.
> 
> And I'm not sure that's always right.
> 
> Paolo?

Indeed.  It is correct for the main system RAM, but hot-plugged RAM
would also have a zero-based section.offset_within_region.  You need to
add memory_region_get_ram_addr(section.mr), just like the call to
balloon_page adds memory_region_get_ram_ptr(section.mr).

Paolo