On Tue, Feb 20, 2024 at 7:36 PM Matthew Wilcox <willy@xxxxxxxxxxxxx> wrote: > > On Tue, Feb 20, 2024 at 03:14:22PM -0800, Vishal Moola (Oracle) wrote: > > In order to handle hugetlb faults under the VMA lock, hugetlb can use > > vmf_anon_prepare() to ensure we can safely prepare an anon_vma. Change > > it to be a non-static function so it can be used within hugetlb as well. > > I think the prototype for this should probably live in mm/internal.h? That does make more sense, I'll move it for v2. > > +++ b/include/linux/hugetlb.h > > @@ -272,6 +272,7 @@ void hugetlb_vma_unlock_write(struct vm_area_struct *vma); > > int hugetlb_vma_trylock_write(struct vm_area_struct *vma); > > void hugetlb_vma_assert_locked(struct vm_area_struct *vma); > > void hugetlb_vma_lock_release(struct kref *kref); > > +vm_fault_t vmf_anon_prepare(struct vm_fault *vmf); > > > > int pmd_huge(pmd_t pmd); > > int pud_huge(pud_t pud); > >
