Re: [PATCH v3 3/5] mm: Add RCU-based VMA lookup helper that waits for writers

On Mon, Aug 3, 2026 at 4:28=E2=80=AFAM Lorenzo Stoakes (ARM) <[email protected]=
g> wrote:
>
> On Sun, Aug 02, 2026 at 02:54:57PM -0700, Suren Baghdasaryan wrote:
> > From: Dave Hansen <[email protected]>
> >
> > =3D=3D Background =3D=3D
> >
> > There are basically two parallel ways to look up a VMA: the
> > traditional way, which is protected by mmap_read_lock, and the RCU-base=
d
> > per-VMA lock way which is based on RCU and refcounts.
> >
> > =3D=3D Problem =3D=3D
> >
> > The mmap_lock one is more straightforward to use but it has a big
> > disadvantage in that it can not be mixed with page faults since those
> > can take mmap_lock for read, which can deadlock when mixed with nested
> > page faults and parallel writers.
> > For example:
> >
> >       mmap_read_lock(mm);
> >       // Another thread does mmap_write_lock().
> >       // New mmap_lock readers are blocked.
> >       vma =3D vma_lookup(mm, address);
> >       // This deadlocks on mmap_read_lock() if it faults:
> >       copy_from_user(address);
> >       mmap_read_unlock(mm);
> >
> > The per-VMA lock can be mixed with faults, but they can fail and need t=
o
> > be able to fall back to the traditional way.
> >
> > =3D=3D Solution =3D=3D
> >
> > Add a variant of the RCU-based lookup that waits for writers. This is
> > basically the same as the existing RCU-based lookup, but on a failure t=
o
> > lock it temporarily takes mmap_lock for read and waits for writers
> > to finish before locking the VMA, dropping the mmap_lock and returning
> > the locked VMA. This has some advantages:
> >
> >  1. Callers do not need to have a fallback path for when they
> >     collide with writers.
> >  2. It can be used in contexts where page faults can happen because
> >     it can take the mmap_lock for read but never *holds* it.
> >  3. Its fast path does not require taking mmap_lock for read.
> >
> > Basically, when applied correctly, this approach results in faster
> > *and* simpler code.
> >
>
> Would be nice to have a:
>
> Suggested-by: Lorenzo Stoakes (ARM) <[email protected]>

Ack.

>
> Here given https://lore.kernel.org/linux-mm/af4Zx0gJIWbdDeY2@lucifer/ :)
>
> > Signed-off-by: Dave Hansen <[email protected]>
> > Signed-off-by: Suren Baghdasaryan <[email protected]>
> > Cc: Suren Baghdasaryan <[email protected]>
> > Cc: Andrew Morton <[email protected]>
> > Cc: "Liam R. Howlett" <[email protected]>
> > Cc: Lorenzo Stoakes <[email protected]>
> > Cc: Vlastimil Babka <[email protected]>
> > Cc: Shakeel Butt <[email protected]>
> > Cc: [email protected]
> > Cc: Greg Kroah-Hartman <[email protected]>
> > Cc: Arve Hj=C3=B8nnev=C3=A5g <[email protected]>
> > Cc: Todd Kjos <[email protected]>
> > Cc: Christian Brauner <[email protected]>
> > Cc: Carlos Llamas <[email protected]>
> > Cc: Alice Ryhl <[email protected]>
> > Cc: "David S. Miller" <[email protected]>
> > Cc: David Ahern <[email protected]>
> > Cc: [email protected]
> > ---
> >  include/linux/mmap_lock.h | 15 +++++++++++----
> >  mm/mmap_lock.c            | 29 +++++++++++++++++++++++++++++
> >  mm/userfaultfd.c          |  6 ++++--
> >  3 files changed, 44 insertions(+), 6 deletions(-)
> >
> > diff --git a/include/linux/mmap_lock.h b/include/linux/mmap_lock.h
> > index eb32b482434e..fdd8f5cf5722 100644
> > --- a/include/linux/mmap_lock.h
> > +++ b/include/linux/mmap_lock.h
> > @@ -228,10 +228,12 @@ static inline void vma_refcount_put(struct vm_are=
a_struct *vma)
> >  }
> >
> >  /*
> > - * Use only while holding mmap read lock which guarantees that locking=
 will not
> > - * fail (nobody can concurrently write-lock the vma). vma_start_read()=
 should
> > + * Use only while holding mmap read lock which guarantees that vma loc=
k is not
> > + * contended (nobody can concurrently write-lock the vma). vma_start_r=
ead() should
> >   * not be used in such cases because it might fail due to mm_lock_seq =
overflow.
> >   * This functionality is used to obtain vma read lock and drop the mma=
p read lock.
> > + * VMA can't be detached while we are holding mmap lock, therefore in =
practice this
> > + * function can fail only when there are so many readers that vm_refcn=
t overflows.
> >   */
> >  static inline bool vma_start_read_locked_nested(struct vm_area_struct =
*vma, int subclass)
> >  {
> > @@ -247,16 +249,21 @@ static inline bool vma_start_read_locked_nested(s=
truct vm_area_struct *vma, int
> >  }
> >
> >  /*
> > - * Use only while holding mmap read lock which guarantees that locking=
 will not
> > - * fail (nobody can concurrently write-lock the vma). vma_start_read()=
 should
> > + * Use only while holding mmap read lock which guarantees that vma loc=
k is not
> > + * contended (nobody can concurrently write-lock the vma). vma_start_r=
ead() should
> >   * not be used in such cases because it might fail due to mm_lock_seq =
overflow.
> >   * This functionality is used to obtain vma read lock and drop the mma=
p read lock.
> > + * VMA can't be detached while we are holding mmap lock, therefore in =
practice this
> > + * function can fail only when there are so many readers that vm_refcn=
t overflows.
> >   */
> >  static inline bool vma_start_read_locked(struct vm_area_struct *vma)
> >  {
> >       return vma_start_read_locked_nested(vma, 0);
> >  }
> >
> > +struct vm_area_struct *vma_start_read_unlocked(struct mm_struct *mm,
> > +                                            unsigned long address);
> > +
> >  static inline void vma_end_read(struct vm_area_struct *vma)
> >  {
> >       vma_refcount_put(vma);
> > diff --git a/mm/mmap_lock.c b/mm/mmap_lock.c
> > index e20d01e8d38f..6ff05e68e61b 100644
> > --- a/mm/mmap_lock.c
> > +++ b/mm/mmap_lock.c
> > @@ -338,6 +338,35 @@ struct vm_area_struct *lock_vma_under_rcu(struct m=
m_struct *mm,
> >       return NULL;
> >  }
> >
> > +/*
>
> Why not a kdoc comment?

Indeed. Will change.

>
> > + * Find the VMA covering 'address' and lock it for reading. Waits for =
writers to
> > + * finish if the VMA is being modified. Returns NULL if there is no VM=
A covering
> > + * 'address'.
> > + *
> > + * Use only in code paths where no mmap_lock and no VMA lock is held.
>
> Well, a VMA read lock can be held which would make this a noop essentiall=
y.
>
> If a VMA write lock is held you're also ok as the mmap read lock will pre=
clude
> an mmap write lock, meaning vma_end_write_all() will have been called and=
 the
> write lock released.
>
> So I think you can just drop this line?

Maybe instead of this I should say: Use when mmap_lock is not held,
otherwise use vma_start_read_locked() ? I think that was the original
reason for this comment.

>
> > + *
> > + * The fast path does not take mmap_lock.
> > + */
> > +struct vm_area_struct *vma_start_read_unlocked(struct mm_struct *mm,
> > +                                            unsigned long address)
> > +{
> > +     struct vm_area_struct *vma;
> > +
> > +     /* Fast path: return stable VMA covering 'address': */
> > +     vma =3D lock_vma_under_rcu(mm, address);
> > +     if (vma)
> > +             return vma;
> > +
> > +     /* Slow path: preclude VMA writers by temporarily getting mmap re=
ad lock. */
> > +     mmap_read_lock(mm);
> > +     vma =3D vma_lookup(mm, address);
> > +     if (vma && !vma_start_read_locked(vma))
>
> This maybe warrants an unlikely() given it can only happen if refcount
> overflows? Also worth having a comment to that effect here?

vma_start_read_locked()
  vma_start_read_locked()
    vma_start_read_locked_nested()
      unlikely(!__refcount_inc_not_zero_limited_acquire())

already contains "unlikely" clause. Do we need to add it in all its users?
I can modify the comment for vma_start_read_locked() stating that
refcount overflow is very unlikely. Would that work or do you want
callers to have that too?

>
> > +             vma =3D NULL;
> > +     mmap_read_unlock(mm);
> > +
> > +     return vma;
> > +}
> > +
> >  static struct vm_area_struct *lock_next_vma_under_mmap_lock(struct mm_=
struct *mm,
> >                                                           struct vma_it=
erator *vmi,
> >                                                           unsigned long=
 from_addr)
> > diff --git a/mm/userfaultfd.c b/mm/userfaultfd.c
> > index edd90892f8cc..c3a0c38a3dc3 100644
> > --- a/mm/userfaultfd.c
> > +++ b/mm/userfaultfd.c
> > @@ -129,8 +129,10 @@ struct vm_area_struct *find_vma_and_prepare_anon(s=
truct mm_struct *mm,
> >   *
> >   * Should be called without holding mmap_lock.
> >   *
> > - * Return: A locked vma containing @address, -ENOENT if no vma is foun=
d, or
> > - * -ENOMEM if anon_vma couldn't be allocated.
> > + * Return: A locked vma containing @address, -ENOENT if no vma is foun=
d,
> > + * -ENOMEM if anon_vma couldn't be allocated, or -EAGAIN if vma refcou=
nt
> > + * overflow happened due to high number of readers and the caller shou=
ld
> > + * retry later.
>
> I'm guessing you're fixing this up as part of the patch? But it feels a b=
it
> random, I mean fine but you should mention this change in the commit mess=
age +
> explain why.

I'm fixing that because Vlastimil asked about these inconsistencies in
his previous review :)
I can package these fixes as a separate patch or add a comment in the
changelog, smth like "While at it, fix the comments for related
functions". Would that work?

>
> >   */
> >  static struct vm_area_struct *uffd_lock_vma(struct mm_struct *mm,
> >                                      unsigned long address)
> > --
> > 2.55.0.508.g3f0d502094-goog
> >
>
> --
> Cheers, Lorenzo