Re: [PATCH 3/3] mm/mglru: add tracepoint for inc_max_seq()

[Ridong Chen]

On 9/11/2026 10:15 PM, Steven Rostedt wrote:
> On Fri, 11 Sep 2026 15:28:48 +0800
> Ridong Chen <ridong.chen@xxxxxxxxx> wrote:
>
> > +static void trace_inc_max_seq(struct lruvec *lruvec)
> > +{
> > +	int type, gen;
> > +	unsigned long nr[ANON_AND_FILE][MAX_NR_GENS];
> > +	struct lru_gen_folio *lrugen = &lruvec->lrugen;
> > +
> > +	if (!trace_mm_mglru_inc_max_seq_enabled())
> > +		return;
> > +
> > +	for (type = 0; type < ANON_AND_FILE; type++)
> > +		for (gen = 0; gen < MAX_NR_GENS; gen++)
> > +			nr[type][gen] = lru_gen_seq_nr_pages(lrugen, gen, type);
> > +
> > +	trace_mm_mglru_inc_max_seq(mem_cgroup_id(lruvec_memcg(lruvec)),
> > +				   lrugen->max_seq,
> > +				   lrugen->min_seq[LRU_GEN_ANON],
> > +				   lrugen->min_seq[LRU_GEN_FILE],
> > +				   nr[LRU_GEN_ANON], nr[LRU_GEN_FILE]);
> > +}
>
> I would always look at trying to move as much logic into the
> TP_fast_assign() and not have it be where the tracepoint is called.
>
> -- Steve

Thanks. Will update.

--
Best regards
Ridong