[PATCH v4 08/10] vfio/migration: Fix a check on vbasedev->iommu_dirty_tracking

[Zhenzhong Duan <zhenzhong.duan@intel.com>]

VFIO IOMMU type1 claims to support IOMMU dirty tracking by setting
bcontainer->dirty_pages_supported, but in vfio_migration_realize()
vbasedev->iommu_dirty_tracking is checked, we should pass
bcontainer->dirty_pages_supported to vbasedev->iommu_dirty_tracking
in legacy backend so that the check is accurate.

Fixes: 30b916778517 ("vfio/common: Allow disabling device dirty page tracking")
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
---
 hw/vfio/container-legacy.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/hw/vfio/container-legacy.c b/hw/vfio/container-legacy.c
index dd9c4a6a5a..fa726a2733 100644
--- a/hw/vfio/container-legacy.c
+++ b/hw/vfio/container-legacy.c
@@ -845,6 +845,7 @@ static bool vfio_device_get(VFIOGroup *group, const char *name,
                             VFIODevice *vbasedev, Error **errp)
 {
     g_autofree struct vfio_device_info *info = NULL;
+    VFIOContainer *bcontainer = VFIO_IOMMU(group->container);
     int fd;
 
     fd = vfio_cpr_group_get_device_fd(group->fd, name);
@@ -883,7 +884,8 @@ static bool vfio_device_get(VFIOGroup *group, const char *name,
         }
     }
 
-    vfio_device_prepare(vbasedev, VFIO_IOMMU(group->container), info);
+    vfio_device_prepare(vbasedev, bcontainer, info);
+    vbasedev->iommu_dirty_tracking = bcontainer->dirty_pages_supported;
 
     vbasedev->fd = fd;
     vbasedev->group = group;
-- 
2.47.1