Re: [PATCH v3 05/15] mm: introduce and use vma_filebacked_address()

On Mon, Aug 03, 2026 at 11:54:31AM +0200, David Hildenbrand (Arm) wrote:
> On 7/29/26 18:48, Lorenzo Stoakes (ARM) wrote:
> > In cases where we know that the VMA is file-backed, use
> > vma_filebacked_address() rather than vma_address().
> >
> > This lays the foundation for using the anonymous page offset via
> > vma_anon_address().
>
> I was briefly confused why we can't do that immediately.

Because everything has to be switched on at once or it breaks things. Obviously
you could use vma_anon_address() for pure anon but at that point you're only
confusing things.

This all stems from the existing situation where anon folios might be indexed by
EITHER a file offset or an anon offset.

In the end 'laying the foundations' makes the actual 'switch it on' patch (9/15)
nice and (relatively) small :)

>
> >
> > Also add an assert to ensure that the VMA whose address is required is not
> > anonymous.
> Acked-by: David Hildenbrand (Arm) <[email protected]>

Thanks!

>
> --
> Cheers,
>
> David

--
Cheers, Lorenzo