Re: [PATCH v2 1/5] mm,hugetlb: Change mechanism to detect a COW on private mapping

[Oscar Salvador]

On Mon, Jun 23, 2025 at 04:09:51PM +0200, David Hildenbrand wrote:
> 
> > -	/*
> > -	 * hugetlb_wp() requires page locks of pte_page(vmf.orig_pte) and
> > -	 * pagecache_folio, so here we need take the former one
> > -	 * when folio != pagecache_folio or !pagecache_folio.
> > -	 */
> > +	/* hugetlb_wp() requires page locks of pte_page(vmf.orig_pte) */
> >   	folio = page_folio(pte_page(vmf.orig_pte));
> > -	if (folio != pagecache_folio)
> > -		if (!folio_trylock(folio)) {
> > -			need_wait_lock = 1;
> > -			goto out_ptl;
> > -		}
> > -
> > +	folio_lock(folio);
> >   	folio_get(folio);
> 
> Just realized that this won't work for this patch here, as we are holding
> the PTL.
> 
> In patch #2 you do the right thing.

Yap, missed that.
I might have to do the lock-unlock-dance here, and then in patch#2 move
it to hugetlb_wp.
Sounds reasonable?

 

-- 
Oscar Salvador
SUSE Labs