Re: [PATCH v2 1/2] mm/rmap: fix and simplify reusing mergeable anon_vma as parent when fork

[Konstantin Khlebnikov]

On 15/01/2020 04.20, Wei Yang wrote:
> On Tue, Jan 14, 2020 at 10:42:52PM +0800, Li Xinhai wrote:
> > On 2020-01-13Âat 19:07ÂKonstantin KhlebnikovÂwrote:
> >
> > > Because I want to keep both heuristics.
> > > This seems most sane way of interaction between them.
> > >
> > > Unfortunately even this patch is slightly broken.
> > > Condition prev->anon_vma->parent == pvma->anon_vma doesn't guarantee that
> > > prev vma has the same set of anon-vmas like current vma.
> > > I.e. anon_vma_clone(vma, prev) might be not enough for keeping connectivity.
> >
> > New patch is required?
>
> My suggestion is separate the fix and new approach instead of mess them into
> one patch.

Yep, it's messy. Maybe it's could be better to revert recent change,
apply second patch from this set and write something new after that.

> > It is necessary to call anon_vma_clone(vma,Âpvma) to link all anon_vma which
> > currently linked by pvma, then link the prev->anon_vma to vma. By this way,
> > connectivity of vma should be maintained, right?
> >
> > > Building such case isn't trivial job but I see nothing that could prevent it.