Re: [PATCH v2 01/11] hugetlb: set hugetlb page flag before optimizing vmemmap

[Mike Kravetz <mike.kravetz@oracle.com>]

On 09/06/23 01:48, Matthew Wilcox wrote:
> On Tue, Sep 05, 2023 at 02:44:00PM -0700, Mike Kravetz wrote:
> > @@ -456,6 +457,7 @@ int hugetlb_vmemmap_restore(const struct hstate *h, struct page *head)
> >  	unsigned long vmemmap_start = (unsigned long)head, vmemmap_end;
> >  	unsigned long vmemmap_reuse;
> >  
> > +	VM_WARN_ON_ONCE(!PageHuge(head));
> >  	if (!HPageVmemmapOptimized(head))
> >  		return 0;
> >  
> > @@ -550,6 +552,7 @@ void hugetlb_vmemmap_optimize(const struct hstate *h, struct page *head)
> >  	unsigned long vmemmap_start = (unsigned long)head, vmemmap_end;
> >  	unsigned long vmemmap_reuse;
> >  
> > +	VM_WARN_ON_ONCE(!PageHuge(head));
> >  	if (!vmemmap_should_optimize(h, head))
> >  		return;
> 
> Someone who's looking for an easy patch or three should convert both
> of these functions to take a folio instead of a page.  All callers
> pass &folio->page.  Obviously do that work on top of Mike's patch set
> to avoid creating more work for him.

I think Muchun already suggested this.

It would make sense as this series is proposing two new routines taking
a list of folios:
- hugetlb_vmemmap_optimize_folios
- hugetlb_vmemmap_restore_folios

-- 
Mike Kravetz