Re: [PATCH 11/18] vfio/common: Add device dirty page bitmap sync

[Avihai Horon]

On 28/01/2023 1:37, Alex Williamson wrote:
> External email: Use caution opening links or attachments
>
>
> On Thu, 26 Jan 2023 20:49:41 +0200
> Avihai Horon <avihaih@nvidia.com> wrote:
>
> > From: Joao Martins <joao.m.martins@oracle.com>
> >
> > Add device dirty page bitmap sync functionality. This uses the device
> > DMA logging uAPI to sync dirty page bitmap from the device.
> >
> > Device dirty page bitmap sync is used only if all devices within a
> > container support device dirty page tracking.
> >
> > Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
> > Signed-off-by: Avihai Horon <avihaih@nvidia.com>
> > ---
> >   hw/vfio/common.c | 93 ++++++++++++++++++++++++++++++++++++++++++------
> >   1 file changed, 82 insertions(+), 11 deletions(-)
> >
> > diff --git a/hw/vfio/common.c b/hw/vfio/common.c
> > index 3caa73d6f7..0003f2421d 100644
> > --- a/hw/vfio/common.c
> > +++ b/hw/vfio/common.c
> > @@ -355,6 +355,9 @@ static void vfio_bitmap_dealloc(VFIOBitmap *vbmap)
> >       g_free(vbmap);
> >   }
> >
> > +static int vfio_get_dirty_bitmap(VFIOContainer *container, uint64_t iova,
> > +                                 uint64_t size, ram_addr_t ram_addr);
> > +
> >   bool vfio_mig_active(void)
> >   {
> >       VFIOGroup *group;
> > @@ -582,10 +585,19 @@ static int vfio_dma_unmap(VFIOContainer *container,
> >           .iova = iova,
> >           .size = size,
> >       };
> > +    int ret;
> >
> > -    if (iotlb && container->dirty_pages_supported &&
> > -        vfio_devices_all_running_and_mig_active(container)) {
> > -        return vfio_dma_unmap_bitmap(container, iova, size, iotlb);
> > +    if (iotlb && vfio_devices_all_running_and_mig_active(container)) {
> > +        if (!vfio_devices_all_device_dirty_tracking(container) &&
> > +            container->dirty_pages_supported) {
> > +            return vfio_dma_unmap_bitmap(container, iova, size, iotlb);
> > +        }
> > +
> > +        ret = vfio_get_dirty_bitmap(container, iova, size,
> > +                                    iotlb->translated_addr);
> > +        if (ret) {
> > +            return ret;
> > +        }
> Isn't the ordering backwards here?  Only after the range is unmapped
> can we know that this container can no longer dirty pages within the
> range.

Oops, I thought that it's OK to query the dirty bitmap when we get the 
vIOMMU unmap notification.
I will reverse the order.

Thanks.