Re: [PATCH v1 04/26] mm: vmscan: refactor move_folios_to_lru()

[Harry Yoo <harry.yoo@oracle.com>]

On Fri, Nov 07, 2025 at 02:41:13PM +0800, Qi Zheng wrote:
> Hi Harry,
> 
> On 11/7/25 1:11 PM, Harry Yoo wrote:
> > On Tue, Oct 28, 2025 at 09:58:17PM +0800, Qi Zheng wrote:
> > > From: Muchun Song <songmuchun@bytedance.com>
> > > 
> > > In a subsequent patch, we'll reparent the LRU folios. The folios that are
> > > moved to the appropriate LRU list can undergo reparenting during the
> > > move_folios_to_lru() process. Hence, it's incorrect for the caller to hold
> > > a lruvec lock. Instead, we should utilize the more general interface of
> > > folio_lruvec_relock_irq() to obtain the correct lruvec lock.
> > > 
> > > This patch involves only code refactoring and doesn't introduce any
> > > functional changes.
> > > 
> > > Signed-off-by: Muchun Song <songmuchun@bytedance.com>
> > > Acked-by: Johannes Weiner <hannes@cmpxchg.org>
> > > Signed-off-by: Qi Zheng <zhengqi.arch@bytedance.com>
> > > ---
> > >   mm/vmscan.c | 46 +++++++++++++++++++++++-----------------------
> > >   1 file changed, 23 insertions(+), 23 deletions(-)
> > > 
> > > diff --git a/mm/vmscan.c b/mm/vmscan.c
> > > index 3a1044ce30f1e..660cd40cfddd4 100644
> > > --- a/mm/vmscan.c
> > > +++ b/mm/vmscan.c
> > > @@ -2016,9 +2016,9 @@ static unsigned long shrink_inactive_list(unsigned long nr_to_scan,
> > >   	nr_reclaimed = shrink_folio_list(&folio_list, pgdat, sc, &stat, false,
> > >   					 lruvec_memcg(lruvec));
> > > -	spin_lock_irq(&lruvec->lru_lock);
> > > -	move_folios_to_lru(lruvec, &folio_list);
> > > +	move_folios_to_lru(&folio_list);
> > > +	spin_lock_irq(&lruvec->lru_lock);
> > >   	__mod_lruvec_state(lruvec, PGDEMOTE_KSWAPD + reclaimer_offset(sc),
> > >   					stat.nr_demoted);
> > 
> > Maybe I'm missing something or just confused for now, but let me ask...
> > 
> > How do we make sure the lruvec (and the mem_cgroup containing the
> > lruvec) did not disappear (due to offlining) after move_folios_to_lru()?
> 
> We obtained lruvec through the following method:
> 
> memcg = mem_cgroup_iter(target_memcg, NULL, partial);
> do {
>     struct lruvec *lruvec = mem_cgroup_lruvec(memcg, pgdat);
> 
>     shrink_lruvec(lruvec, sc);
>     --> shrink_inactive_list
> } while ((memcg = mem_cgroup_iter(target_memcg, memcg, partial)));
> 
> The mem_cgroup_iter() will hold the refcount of this memcg, so IIUC,
> the memcg will not disappear at this time.

Ah, right!

It can be offlined, but won't be released due to the refcount.

Thanks for the explanation.

> > >   	__mod_node_page_state(pgdat, NR_ISOLATED_ANON + file, -nr_taken);
> > > @@ -2166,11 +2166,10 @@ static void shrink_active_list(unsigned long nr_to_scan,
> > >   	/*
> > >   	 * Move folios back to the lru list.
> > >   	 */
> > > -	spin_lock_irq(&lruvec->lru_lock);
> > > -
> > > -	nr_activate = move_folios_to_lru(lruvec, &l_active);
> > > -	nr_deactivate = move_folios_to_lru(lruvec, &l_inactive);
> > > +	nr_activate = move_folios_to_lru(&l_active);
> > > +	nr_deactivate = move_folios_to_lru(&l_inactive);
> > > +	spin_lock_irq(&lruvec->lru_lock);
> > >   	__count_vm_events(PGDEACTIVATE, nr_deactivate);
> > >   	count_memcg_events(lruvec_memcg(lruvec), PGDEACTIVATE, nr_deactivate);
> > > @@ -4735,14 +4734,15 @@ static int evict_folios(unsigned long nr_to_scan, struct lruvec *lruvec,
> > >   			set_mask_bits(&folio->flags.f, LRU_REFS_FLAGS, BIT(PG_active));
> > >   	}
> > > -	spin_lock_irq(&lruvec->lru_lock);
> > > -
> > > -	move_folios_to_lru(lruvec, &list);
> > > +	move_folios_to_lru(&list);
> > > +	local_irq_disable();
> > >   	walk = current->reclaim_state->mm_walk;
> > >   	if (walk && walk->batched) {
> > >   		walk->lruvec = lruvec;
> > > +		spin_lock(&lruvec->lru_lock);
> > >   		reset_batch_size(walk);
> > > +		spin_unlock(&lruvec->lru_lock);
> > >   	}
> > 
> > Cc'ing RT folks as they may not want to disable IRQ on PREEMPT_RT.
> > 
> > IIRC there has been some effort in MM to reduce the scope of
> > IRQ-disabled section in MM when PREEMPT_RT config was added to the
> > mainline. spin_lock_irq() doesn't disable IRQ on PREEMPT_RT.
> 
> Thanks for this information.
> 
> > 
> > Also, this will break RT according to Documentation/locking/locktypes.rst:
> > > The changes in spinlock_t and rwlock_t semantics on PREEMPT_RT kernels
> > > have a few implications. For example, on a non-PREEMPT_RT kernel
> > > the following code sequence works as expected:
> > > 
> > > local_irq_disable();
> > > spin_lock(&lock);
> > > 
> > > and is fully equivalent to:
> > > 
> > > spin_lock_irq(&lock);
> > > Same applies to rwlock_t and the _irqsave() suffix variants.
> > > 
> > > On PREEMPT_RT kernel this code sequence breaks because RT-mutex requires
> > > a fully preemptible context. Instead, use spin_lock_irq() or
> > > spin_lock_irqsave() and their unlock counterparts.
> > > 
> > > In cases where the interrupt disabling and locking must remain separate,
> > > PREEMPT_RT offers a local_lock mechanism. Acquiring the local_lock pins
> > > the task to a CPU, allowing things like per-CPU interrupt disabled locks
> > > to be acquired. However, this approach should be used only where absolutely
> > > necessary.
> 
> But how do we determine if it's necessary?

Although it's mentioned in the locking documentation, I'm afraid that
local_lock is not the right interface to use here. Preemption will be
disabled anyway (on both PREEMPT_RT and !PREEMPT_RT) when the stats are
updated (in __mod_node_page_state()).

Here we just want to disable IRQ only on !PREEMPT_RT (to update
the stats safely).

Maybe we'll need some ifdeffery? I don't see a better way around...

-- 
Cheers,
Harry / Hyeonggon