Re: [RFC v3 2/4] mm: move PG_slab flag to page_type

[David Hildenbrand]

On 08.02.23 10:44, Mike Rapoport wrote:
> On Fri, Feb 03, 2023 at 05:04:01PM +0100, David Hildenbrand wrote:
> > On 30.01.23 06:11, Matthew Wilcox wrote:
> > > On Mon, Jan 30, 2023 at 01:34:59PM +0900, Hyeonggon Yoo wrote:
> > > > > Seems like quite some changes to page_type to accomodate SLAB, which is
> > > > > hopefully going away soon(TM). Could we perhaps avoid that?
> > > >
> > > > If it could be done with less changes, I'll try to avoid that.
> > >
> > > Let me outline the idea I had for removing PG_slab:
> > >
> > > Observe that PG_reserved and PG_slab are mutually exclusive.  Also,
> >
> > I recall that there are SetPageReserved() calls on pages allocated via slab.
>
> I did a quick scan, and it seems that all allocated reserved pages come
> from alloc_pages() or vmalloc().
>   

Thanks for checking, good that my memory was wrong :)

> BTW, looking at the current usage of SetPageReserved, it seems that some
> are bogus/historical and some would justify a different page type if we are
> to have 15 bits for PG_kernel.

I remember raising in the past that something like PG_hole might be helpful.

I also recall that most SetPageReserved usage in drivers is to make 
ioremap happy. PG_ioremappable or smth like that would be better.

So yes, that would even help to cleanup the existing PG_reserved usage.

--
Thanks,

David / dhildenb