>-----Original Message----- >From: Cédric Le Goater <[email protected]> >Regarding the whole series, > >* [1/5] vfio/iommufd: Add framework code to support getting dirty bitmap >before unmap > > Needs refactoring. iommufd_cdev_unmap_one() seems superfluous >now ?
Yes. > >* [2/5] vfio/iommufd: Query dirty bitmap before DMA unmap > > Looks OK. > > In an extra patch, could we rename 'vfio_dma_unmap_bitmap()' to > 'vfio_legacy_dma_unmap_get_dirty_bitmap()' ? Helps (me) remember > what it does. Sure > >* [3/5] vfio/iommufd: Add IOMMU_HWPT_GET_DIRTY_BITMAP_NO_CLEAR >flag support > > Sorry, I lost track of the discussion between you and Joao regarding > the new '->dirty_tracking_flags' attribute. > > I'd prefer adding a 'backend_flag' parameter, even if it’s currently > only used by IOMMUFD. Since we’re not going to redefine all possible > IOMMU backend flags, we can treat it as an opaque parameter for > the upper container layer and document it as such. > > Is that ok for you ? A bit more churn for you but Joao did provide > some parts. Sure, will cherry-pick Joao's change. > >* [4/5] intel_iommu: Optimize unmap_bitmap during migration > > Ack by Clément may be ? Yes Thanks Zhenzhong > >* [5/5] vfio/migration: Allow live migration with vIOMMU without VFs > using device dirty tracking > > ok. > > >Thanks, > >C.
