Re: [RFC 19/19] mm, netmem: remove the page pool members in struct page

[Byungchul Park <byungchul@sk.com>]

On Tue, May 13, 2025 at 04:19:03AM +0100, Matthew Wilcox wrote:
> On Tue, May 13, 2025 at 10:42:00AM +0900, Byungchul Park wrote:
> > Just in case, lemme explain what I meant, for *example*:
> 
> I understood what you meant.
> 
> > In here, operating on struct netmem_desc can smash _mapcount and
> > _refcount in struct page unexpectedly, even though sizeof(struct
> > netmem_desc) <= sizeof(struct page).  That's why I think the place holder
> > is necessary until it completely gets separated so as to have its own
> > instance.
> 
> We could tighten up the assert a bit.  eg
> 
> static_assert(sizeof(struct netmem_desc) <= offsetof(struct page, _refcount));

This mitigates what I concern.  I will replace the place holder with
this (but it must never happen to relocate the fields in struct page by
any chance for any reason until the day.  I trust you :).

	Byungchul

> We _can't_ shrink struct page until struct folio is dynamically
> allocated.  The same patch series that dynamically allocates folio will
> do the same for netmem and slab and ptdesc and ...