Re: [PATCH v2 8/9] mm: introduce and use vm_normal_page_pud()

[Lorenzo Stoakes <lorenzo.stoakes@oracle.com>]

On Fri, Jul 18, 2025 at 01:06:30PM +0200, David Hildenbrand wrote:
> On 18.07.25 12:47, Lorenzo Stoakes wrote:
> > On Thu, Jul 17, 2025 at 10:14:33PM +0200, David Hildenbrand wrote:
> > > On 17.07.25 22:03, Lorenzo Stoakes wrote:
> > > > On Thu, Jul 17, 2025 at 01:52:11PM +0200, David Hildenbrand wrote:
> > > > > Let's introduce vm_normal_page_pud(), which ends up being fairly simple
> > > > > because of our new common helpers and there not being a PUD-sized zero
> > > > > folio.
> > > > >
> > > > > Use vm_normal_page_pud() in folio_walk_start() to resolve a TODO,
> > > > > structuring the code like the other (pmd/pte) cases. Defer
> > > > > introducing vm_normal_folio_pud() until really used.
> > > >
> > > > I mean fine :P but does anybody really use this?
> > >
> > > This is a unified PFN walker (!hugetlb + hugetlb), so you can easily run
> > > into hugetlb PUDs, DAX PUDs and huge pfnmap (vfio) PUDs :)
> >
> > Ahhh ok. I hate hugetlb so very very much.
> >
> > Oscar is doing the Lord's work improving things but the trauma is still
> > there... :P
> >
> > Also yeah DAX ahem.
> >
> > I'm not familiar with huge pfnmap PUDs, could you give me a hint on this? :>)
>
> vmf_insert_pfn_pmd(), called from  drivers/vfio/pci/vfio_pci_core.c
>
> Essentially, we create huge PUDs when mapping device BARs to user space.

Ah makes sense. Thanks!