Re: [PATCH v3] mm: mglru: fix stale batch updates after memcg reparenting

[Qi Zheng]

Hi Johannes,

On 6/26/26 2:41 AM, Johannes Weiner wrote:
> On Thu, Jun 25, 2026 at 11:15:54PM +0800, Qi Zheng wrote:
> > From: Qi Zheng <zhengqi.arch@xxxxxxxxxxxxx>
> >
> > The mglru page table walker batches per-generation size deltas in
> > walk->nr_pages while walking page tables without holding the lruvec lock.
> > The reset_batch_size() later folds those deltas into walk->lruvec under
> > the lruvec lock.
> >
> > The page table walker can run concurrently with the memcg reparenting path
> > as follows:
> >
> > CPU0                           CPU1
> > ====                           ====
> >
> > walk_mm
> > --> walk_page_range
> >      --> update_batch_size
> >          --> walk->nr_pages += delta
> >
> >                                mem_cgroup_css_offline
> >                                --> memcg_reparent_objcgs
> >                                    --> lock lruvec
> >                                        lru_gen_reparent_memcg
> >                                        --> reparent child folios to parent
> >                                        unlock lruvec
> >
> >      lock lruvec
> >      reset_batch_size
> >      --> child lrugen->nr_pages += delta
> >
> > This will trigger the following warning in lru_gen_exit_memcg():
> >
> > 	VM_WARN_ON_ONCE(memchr_inv(lruvec->lrugen.nr_pages, 0,
> > 				   sizeof(lruvec->lrugen.nr_pages)));
> >
> > And the user-visible impact of underestimated nr_pages in MGLRU was
> > premature OOMs because MGLRU does not try to reclaim memory when nr_pages
> > reaches zero, but there are still more pages.
> >
> > To fix it, make reset_batch_size() check CSS_DYING under RCU before
> > flushing the pending batch. A non-dying memcg keeps the original lruvec
> > stable against RCU-delayed offlining; a dying memcg redirects the deltas
> > to the first non-dying ancestor.
> >
> > Reported-by: Peiyang He <peiyang_he@xxxxxxxxxxxxxxxx>
> > Closes: https://lore.kernel.org/all/5A9E929D82717101+12fcf643-efb8-4b9a-a53a-1e28cc894f0b@xxxxxxxxxxxxxxxx
> > Fixes: f304652609ea ("mm: vmscan: prepare for reparenting MGLRU folios")
> > Cc: <stable@xxxxxxxxxxxxxxx>
> > Signed-off-by: Qi Zheng <zhengqi.arch@xxxxxxxxxxxxx>
> > ---
> > Changes in v3:
> >   - re-implement lock_batch_lruvec() by checking CSS_DYING under the RCU lock
> >     (suggested by Harry)
> >   - update the commit message (suggested by Harry)
> >   - temporarily drop the previous Reviewed-by tags
> >     (since the sync method has changed)
> >   - rebase onto the next-20260624
> >
> > Changes in v2:
> >   - update the commit message (pointed by Barry)
> >   - collect Reviewed-by
> >
> >   mm/vmscan.c | 45 ++++++++++++++++++++++++++++++++++++++-------
> >   1 file changed, 38 insertions(+), 7 deletions(-)
> >
> > diff --git a/mm/vmscan.c b/mm/vmscan.c
> > index 35c3bb15ae96..1ec8c23c72b9 100644
> > --- a/mm/vmscan.c
> > +++ b/mm/vmscan.c
> > @@ -3262,10 +3262,44 @@ static void update_batch_size(struct lru_gen_mm_walk *walk, struct folio *folio,
> >   	walk->nr_pages[new_gen][type][zone] += delta;
> >   }
> >   
> > +#ifdef CONFIG_MEMCG
> > +static struct lruvec *lock_batch_lruvec(struct lruvec *lruvec)
> > +{
> > +	struct pglist_data *pgdat = lruvec_pgdat(lruvec);
> > +	struct mem_cgroup *memcg = lruvec_memcg(lruvec);
> > +
> > +	rcu_read_lock();
>
> Where is this unlocked?

The lruvec_unlock_irq() in reset_batch_size() will handle the unlocking.

> > +	/*
> > +	 * The memcg can be NULL when the memory controller is disabled.
> > +	 * Otherwise, the caller keeps the memcg owning @lruvec alive.
> > +	 */
> > +	if (!memcg || !css_is_dying(&memcg->css))
> > +		goto lock;
> > +
> > +	do {
> > +		memcg = parent_mem_cgroup(memcg);
> > +	} while (memcg && css_is_dying(&memcg->css));
> > +	lruvec = mem_cgroup_lruvec(memcg, pgdat);
>
> 	while (unlikely(memcg && css_is_dying(&memcg->css))) {
> 		memcg = parent_mem_cgroup(memcg);
> 		lruvec = mem_cgroup_lruvec(memcg, pgdat);

There is no need to acquire the lruvec before finding the first
non-dying memcg.

Thanks,
Qi

> 	}