Re: [PATCH v10 03/12] vfio/migration: Allow migration without VFIO IOMMU dirty tracking support

[Alex Williamson]

On Wed, 15 Feb 2023 16:38:10 -0400
Jason Gunthorpe <jgg@nvidia.com> wrote:

> On Wed, Feb 15, 2023 at 01:14:35PM -0700, Alex Williamson wrote:
> 
> > We'll need to consider whether we want to keep "dumb" dirty tracking,
> > or even any form of dirty tracking in the type1 uAPI, under an
> > experimental opt-in.  Thanks,  
> 
> I was expecting we'd delete the kernel code for type 1 dirty tracking
> once the v2 parts are merged to qemu since we don't and won't have any
> kernel implementation of it..
> 
> The big point of this to allow qmeu to continue on with a future
> kernel that no longer reports it supports this.

Right, in the part-1 series adding v2 support, we have no other dirty
tracking, so it serves a purpose until we have device-level dirty
tracking in the part-2 series.  After that, we can certainly at least
remove the type1 dirty tracking since there never were, and likely never
will be, any in-kernel implementations.  I could go either way if we
want to keep the in-QEMU dirty-everything tracking around as an
experimental option to aid migration driver development/debug (ie.
maybe enabling it preempts device dirty tracking as a comparison for
debug).  Thanks,

Alex