[PATCH V6 1/2] mm/page_alloc: Make alloc_gigantic_page() available for general use

[Michal Hocko]

On Tue 15-10-19 14:09:56, Michal Hocko wrote:
> On Tue 15-10-19 13:50:02, David Hildenbrand wrote:
> > On 15.10.19 13:47, Michal Hocko wrote:
> > > On Tue 15-10-19 13:42:03, David Hildenbrand wrote:
> > > [...]
> > > > > -static bool pfn_range_valid_gigantic(struct zone *z,
> > > > > -			unsigned long start_pfn, unsigned long nr_pages)
> > > > > -{
> > > > > -	unsigned long i, end_pfn = start_pfn + nr_pages;
> > > > > -	struct page *page;
> > > > > -
> > > > > -	for (i = start_pfn; i < end_pfn; i++) {
> > > > > -		if (!pfn_valid(i))
> > > > > -			return false;
> > > > > -
> > > > > -		page = pfn_to_page(i);
> > > > 
> > > > Am I missing something or should here really be a pfn_to_online_page() here
> > > > instead of a pfn_valid() ?
> > > 
> > > http://lkml.kernel.org/r/20180423000943.GO17484@dhcp22.suse.cz
> > > 
> > 
> > So we managed to add PageReserved(page) but not pfn_to_online_page(). But it
> > is the right thing to do? (or am I missing something?)
> 
> Yeah, pfn_to_online_page is better. But please note that this is an
> optimistic check. The real check has to be done when isolating the
> pageblock because things might change in the meantime.

Except I have missed that we do get zone from the page and other
undefined state. Scratch my above comment.
-- 
Michal Hocko
SUSE Labs