Re: [PATCH v2] mm/userfaultfd: fix hugetlb fault mutex hash calculation

[Jianhui Zhou <jianhuizzzzz@gmail.com>]

On Sun, Mar 08, 2026, Hugh Dickins wrote:
> I have not thought it through, nor checked (someone else please do so
> before this might reach stable trees); but I believe it's very likely
> that that Fixes attribution to a 4.11 commit is wrong - more likely 6.7's
> a08c7193e4f1 ("mm/filemap: remove hugetlb special casing in filemap.c").

You are right. Before a08c7193e4f1, linear_page_index() called
linear_hugepage_index() for hugetlb VMAs, which returned the index in
huge page units. The bug was introduced when a08c7193e4f1 removed that
special casing but missed updating the caller in mm/userfaultfd.c.

I will fix the Fixes tag in v3. Thanks!

Hugh Dickins <hughd@google.com> 于2026年3月9日周一 10:09写道：
>
> On Sat, 7 Mar 2026, Jianhui Zhou wrote:
>
> > In mfill_atomic_hugetlb(), linear_page_index() is used to calculate the
> > page index for hugetlb_fault_mutex_hash(). However, linear_page_index()
> > returns the index in PAGE_SIZE units, while hugetlb_fault_mutex_hash()
> > expects the index in huge page units (as calculated by
> > vma_hugecache_offset()). This mismatch means that different addresses
> > within the same huge page can produce different hash values, leading to
> > the use of different mutexes for the same huge page. This can cause
> > races between faulting threads, which can corrupt the reservation map
> > and trigger the BUG_ON in resv_map_release().
> >
> > Fix this by replacing linear_page_index() with vma_hugecache_offset()
> > and applying huge_page_mask() to align the address properly. To make
> > vma_hugecache_offset() available outside of mm/hugetlb.c, move it to
> > include/linux/hugetlb.h as a static inline function.
> >
> > Fixes: 60d4d2d2b40e ("userfaultfd: hugetlbfs: add __mcopy_atomic_hugetlb for huge page UFFDIO_COPY")
>
> I have not thought it through, nor checked (someone else please do so
> before this might reach stable trees); but I believe it's very likely
> that that Fixes attribution to a 4.11 commit is wrong - more likely 6.7's
> a08c7193e4f1 ("mm/filemap: remove hugetlb special casing in filemap.c").
>
> Hugh