Re: [RFC PATCH] mm: move PG_slab flag to page_type

[Hyeonggon Yoo]

On Mon, Sep 19, 2022 at 09:57:08PM +0900, Hyeonggon Yoo wrote:
> For now, only SLAB uses _mapcount field as a number of active objects in
> a slab, and other slab allocators do not use it. As 16 bits are enough
> for that, use remaining 16 bits of _mapcount as page_type even when
> SLAB is used. And then move PG_slab flag to page_type!
> 
> Note that page_type is always placed in upper 16 bits of _mapcount to
> avoid confusing normal _mapcount as page_type. As underflow (actually
> I mean, yeah, overflow) is not a concern anymore, use more lower bits
> except bit zero.
> 
> Add more folio helpers for PAGE_TYPE_OPS() not to break existing
> slab implementations.
> 
> Remove PG_slab check from PAGE_FLAGS_CHECK_AT_FREE. buddy will still
> check if _mapcount is properly set at free.
> 
> Exclude PG_slab from hwpoison and show_page_flags() for now.
> 
> Note that with this patch, page_mapped() and folio_mapped() always return
> false for slab page.
> 
> Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
> Cc: Christoph Lameter <cl@xxxxxxxxx>
> Cc: Pekka Enberg <penberg@xxxxxxxxxx>
> Cc: David Rientjes <rientjes@xxxxxxxxxx>
> Cc: Joonsoo Kim <iamjoonsoo.kim@xxxxxxx>
> Cc: Vlastimil Babka <vbabka@xxxxxxx>
> Cc: Roman Gushchin <roman.gushchin@xxxxxxxxx>
> Cc: Naoya Horiguchi <naoya.horiguchi@xxxxxxx>
> Cc: Miaohe Lin <linmiaohe@xxxxxxxxxx>
> Cc: "Matthew Wilcox (Oracle)" <willy@xxxxxxxxxxxxx>
> Cc: Minchan Kim <minchan@xxxxxxxxxx>
> Cc: Mel Gorman <mgorman@xxxxxxxxxxxxxxxxxxx>
> Cc: Andrea Arcangeli <aarcange@xxxxxxxxxx>
> Cc: Dan Williams <dan.j.williams@xxxxxxxxx>
> Cc: Hugh Dickins <hughd@xxxxxxxxxx>
> Cc: Muchun Song <songmuchun@xxxxxxxxxxxxx>
> Cc: David Hildenbrand <david@xxxxxxxxxx>
> Cc: Andrey Konovalov <andreyknvl@xxxxxxxxx>
> Cc: Marco Elver <elver@xxxxxxxxxx>
> Signed-off-by: Hyeonggon Yoo <42.hyeyoo@xxxxxxxxx>
> ---
> 
> I think this gives us two benefits:
> - it frees a bit in flags field
> - it makes it simpler to distinguish user-mapped pages and
>   not-user-mapped pages.
> 
> Plus I'm writing a bit more of code including:
> 	0) a few cleanup for code that checks
> 	   !PageSlab() && page_mapped() or that does similar thing
> 	1) provide human-readale string of page_type in dump_page
> 	2) add show_page_types() for tracepoint
> 	3) fix hwpoison ...etc.
> 
> Anyway This is an early RFC, I will very appreciate feedbacks!

Looks like I forgot to add RFC in subject :)