Re: [PATCH 1/9] mm: introduce hw_pte_t for PTE table storage

[Anshuman Khandual <anshuman.khandual@xxxxxxx>]

On Mon, Aug 10, 2026 at 01:20:00PM +0200, David Hildenbrand (Arm) wrote:
> On 8/10/26 12:09, Muhammad Usama Anjum wrote:
> > On 09/08/2026 6:45 pm, Alexander Gordeev wrote:
> >> On Fri, Aug 07, 2026 at 04:24:00PM +0100, Muhammad Usama Anjum wrote:
> >>> Thank you for testing it out on s390.
> >>>
> >>> As __hw_pte_t isn't being used yet in this series, would s390 enablement
> >>> patches add __hw_pte_t to this definition?
> >>
> >> I hope there is a better solution. As I noted m68k, powerpc and sparc
> >> may also be affected, so I would suggest to look into those as well.
> >> I would prefer s390 to use the generic one rather than circumvent a
> >> compile error in a custom way.
> > I've just checked all of these architectures by doing dirty conversion and
> > reached to same conclusion that __hw_pte_t must be defined like:
> >  
> > typedef struct __hw_pte_t { pte_t __pte; } hw_pte_t;
> > 
> > I'lll add __hw_pte_t to this series. (Initially on last email I'd thought
> > that the first user would add __hw_pte_t. But it seems sensible to add it
> > now)
> 
> Yes, do it as part of the introduction. Also a good idea to mention in the 
> patch
> description *why* that is required.

Agreed.

> 
> -- 
> Cheers,
> 
> David