Re: [PATCH v2 3/4] mm/mglru: skip cold PUD subtrees during aging

[Baoquan He <baoquan.he@linux.dev>]

On 11/11/26 at 03:38pm, Kairui Song wrote:
> On Wed, Nov 11, 2026 at 3:19 PM Baoquan He <baoquan.he@linux.dev> wrote:
...snip...
> > > I think we can drop the PUD ptr walk entirely. The filter only
> > > needs a stable identity for the range walk_pud_range() is iterating,
> > > and (mm, PUD-aligned address) gives that on both sides:
> > >
> > >    walk_pud_range(), test and update alike:
> > >       (void *)((unsigned long)args->mm ^ (addr & PUD_MASK))
> > >
> > >    lru_gen_look_around():
> > >       (void *)((unsigned long)vma->vm_mm ^ (pvmw->address & PUD_MASK))
> > >
> > > Will this work? Also I took a look at the hash algorithm of this too:
> >
> > Sorry, missed this mail too, thanks for great suggestion.
> >
> > Yeah, it's a much better way to do pud level filtering. I will
> > switch to it in v3.
> 
> Great, we should have a super fast O(1) key calculation then.
> 
> > And the same idea on PMD has been merged, your suggestion is earlier
> > than it.
> > 65a9bf08a3e4 mm/mglru: don't use stack pointers when walking PUD and PMD tables
> 
> Not sure if I understand this wrongly, but 65a9bf08a3e4 is fixing a
> potential stale stack value, not changing how PMD keys are calculated?

Ah, my bad. It indeed doesn't change how PMD keys are calculated, I
thought about it wrongly on whether pmd key can be calculated the the
similar way as pud. Seems the answer is not worthing.