Re: [PATCH v4 1/5] mm: Make per-VMA locks available universally

[Lorenzo Stoakes (ARM)]

On Mon, Aug 10, 2026 at 09:52:11AM +0100, Lorenzo Stoakes (ARM) wrote:
> On Sat, Aug 08, 2026 at 02:12:50AM +0100, Matthew Wilcox wrote:
> > On Thu, Aug 06, 2026 at 01:05:44PM -0700, Suren Baghdasaryan wrote:
> > > +++ b/kernel/bpf/stackmap.c
> > > @@ -272,13 +272,8 @@ struct stack_map_vma_lock {
> > >  /*
> > >   * Acquire a stable read-side reference on the VMA covering @ip.
> > >   *
> > > - * With CONFIG_PER_VMA_LOCK=y this returns a VMA with its per-VMA read
> > > - * lock held and mmap_lock dropped, so the caller may sleep.
> > > - *
> > > - * With CONFIG_PER_VMA_LOCK=n it returns a VMA with mmap_lock still
> > > - * held; the caller must snapshot any fields it needs and pin vm_file
> > > - * with get_file() before stack_map_unlock_vma() drops mmap_lock, as
> > > - * the VMA may be split, merged, or freed after that.
> > > + * This returns a VMA with its per-VMA read lock held and mmap_lock
> > > + * dropped, so the caller may sleep.
> >
> > I don't know if BPF is compatible with !MMU or not, but the comment
> > is inconsistent with the code.  How about:
>
> <requisite nommu rant>
>
> I do think there are components that simply don't think to depend on CONFIG_MMU
> even though they do.
>
> In fact more than think - have run into exactly that before.
>
> It's another thing that speaks to nommu being a legacy barnacle that bashes us
> on the head fairly regularly for little to no gain (and nobody is testing it for
> tip kernel AFAICT).
>
> </requisite nommu rant>
>
> >
> >  * On NOMMU configurations, returns with the mmap_lock held.  If the MMU
> >  * is enabled, the per-VMA lock will be held instead.  The lock
> >  * should be released with stack_map_unlock_vma() which will release the
> >  * appropriate lock.  Once the lock is released, the VMA may be freed.
>
> I mean I suppose it's accurate but I don't love the idea of essentially implying
> nommu+bpf is a thing and also treating it as so important that it must be called
> out here.
>
> I'd rather it be inaccurate for nommu as are most comments in mm and mm-adjacent
> components, it's kinda implied in general. Those who care can look at the code.

Actually scratch that I'm wrong, firstly Suren's series explicitly does some
nommu-specific logic in the bpf code and secondly BPF already has nommu-specific
stuff in it.

It seems nommu bpf is (kinda) a supported path.

>
> >
> > >   * Returns NULL on failure, in which case no lock is held.
> > >   */

--
Cheers, Lorenzo