Re: [RFC PATCH RESEND 21/28] mm: introduce find_and_lock_anon_vma to be used from arch-specific code

[Suren Baghdasaryan]

On Fri, Sep 9, 2022 at 7:38 AM Laurent Dufour <ldufour@xxxxxxxxxxxxx> wrote:
>
> Le 01/09/2022 à 19:35, Suren Baghdasaryan a écrit :
> > Introduce find_and_lock_anon_vma function to lookup and lock an anonymous
> > VMA during page fault handling. When VMA is not found, can't be locked
> > or changes after being locked, the function returns NULL. The lookup is
> > performed under RCU protection to prevent the found VMA from being
> > destroyed before the VMA lock is acquired. VMA lock statistics are
> > updated according to the results.
> >
> > Signed-off-by: Suren Baghdasaryan <surenb@xxxxxxxxxx>
> > ---
> >  include/linux/mm.h |  3 +++
> >  mm/memory.c        | 45 +++++++++++++++++++++++++++++++++++++++++++++
> >  2 files changed, 48 insertions(+)
> >
> > diff --git a/include/linux/mm.h b/include/linux/mm.h
> > index 7c3190eaabd7..a3cbaa7b9119 100644
> > --- a/include/linux/mm.h
> > +++ b/include/linux/mm.h
> > @@ -684,6 +684,9 @@ static inline void vma_assert_no_reader(struct vm_area_struct *vma)
> >                     vma);
> >  }
> >
> > +struct vm_area_struct *find_and_lock_anon_vma(struct mm_struct *mm,
> > +                                           unsigned long address);
> > +
> >  #else /* CONFIG_PER_VMA_LOCK */
> >
> >  static inline void vma_init_lock(struct vm_area_struct *vma) {}
> > diff --git a/mm/memory.c b/mm/memory.c
> > index 29d2f49f922a..bf557f7056de 100644
> > --- a/mm/memory.c
> > +++ b/mm/memory.c
> > @@ -5183,6 +5183,51 @@ vm_fault_t handle_mm_fault(struct vm_area_struct *vma, unsigned long address,
> >  }
> >  EXPORT_SYMBOL_GPL(handle_mm_fault);
> >
> > +#ifdef CONFIG_PER_VMA_LOCK
> > +static inline struct vm_area_struct *find_vma_under_rcu(struct mm_struct *mm,
> > +                                                     unsigned long address)
> > +{
> > +     struct vm_area_struct *vma = __find_vma(mm, address);
> > +
> > +     if (!vma || vma->vm_start > address)
> > +             return NULL;
> > +
> > +     if (!vma_is_anonymous(vma))
> > +             return NULL;
> > +
>
> It looks to me more natural to first check that the VMA is part of the RB
> tree before try read locking it.

I think we want to check that the VMA is still part of the mm _after_
we locked it. Otherwise we might pass the check, then some other
thread does (lock->isolate->unlock) and then we lock the VMA. We would
end up with a VMA that is not part of mm anymore but we assume it is.

>
> > +     if (!vma_read_trylock(vma)) {
> > +             count_vm_vma_lock_event(VMA_LOCK_ABORT);
> > +             return NULL;
> > +     }
> > +
> > +     /* Check if the VMA got isolated after we found it */
> > +     if (RB_EMPTY_NODE(&vma->vm_rb)) {
> > +             vma_read_unlock(vma);
> > +             count_vm_vma_lock_event(VMA_LOCK_MISS);
> > +             return NULL;
> > +     }
> > +
> > +     return vma;
> > +}
> > +
> > +/*
> > + * Lookup and lock and anonymous VMA. Returned VMA is guaranteed to be stable
> > + * and not isolated. If the VMA is not found of is being modified the function
> > + * returns NULL.
> > + */
> > +struct vm_area_struct *find_and_lock_anon_vma(struct mm_struct *mm,
> > +                                           unsigned long address)
> > +{
> > +     struct vm_area_struct *vma;
> > +
> > +     rcu_read_lock();
> > +     vma = find_vma_under_rcu(mm, address);
> > +     rcu_read_unlock();
> > +
> > +     return vma;
> > +}
> > +#endif /* CONFIG_PER_VMA_LOCK */
> > +
> >  #ifndef __PAGETABLE_P4D_FOLDED
> >  /*
> >   * Allocate p4d page table.
>