Re: [PATCH] mm: hugetlb: independent PMD page table shared count

[David Hildenbrand]

On 16.12.24 19:33, jane.chu@xxxxxxxxxx wrote:
> On 12/14/2024 2:44 AM, Liu Shixin wrote:
>
> > The folio refcount may be increased unexpectly through try_get_folio() by
> > caller such as split_huge_pages. In huge_pmd_unshare(), we use refcount to
> > check whether a pmd page table is shared. The check is incorrect if the
> > refcount is increased by the above caller, and this can cause the page
> > table leaked:
>
> hugetlb and THP don't overlap, right?  how does split_huge_pages() end
> up messing up huge_pmd_share() ?
>
> Am I missing something?

If first grabs a reference to then check if it's a THP. So we can end up 
grabbing anything temporarily.


In general, we'll have to be smarter about what we try grabbing, but 
handling races for now is tricky.

--
Cheers,

David / dhildenb