[SRU][jammy:linux][PATCH] mm/memcontrol.c: remove the redundant updating of stats_flush_threshold

[Philip Cox]

From: Jiebin Sun <jiebin.sun at intel.com>

BugLink: https://bugs.launchpad.net/bugs/2052827

Remove the redundant updating of stats_flush_threshold.  If the global var
stats_flush_threshold has exceeded the trigger value for
__mem_cgroup_flush_stats, further increment is unnecessary.

Apply the patch and test the pts/hackbench-1.0.0 Count:4 (160 threads).

Score gain: 1.95x
Reduce CPU cycles in __mod_memcg_lruvec_state (44.88% -> 0.12%)

CPU: ICX 8380 x 2 sockets
Core number: 40 x 2 physical cores
Benchmark: pts/hackbench-1.0.0 Count:4 (160 threads)

Link: https://lkml.kernel.org/r/20220722164949.47760-1-jiebin.sun@intel.com
Signed-off-by: Jiebin Sun <jiebin.sun at intel.com>
Acked-by: Shakeel Butt <shakeelb at google.com>
Reviewed-by: Roman Gushchin <roman.gushchin at linux.dev>
Reviewed-by: Tim Chen <tim.c.chen at linux.intel.com>
Acked-by: Muchun Song <songmuchun at bytedance.com>
Cc: Johannes Weiner <hannes at cmpxchg.org>
Cc: Michal Hocko <mhocko at kernel.org>
Cc: "Huang, Ying" <ying.huang at intel.com>
Cc: Amadeusz Sawiski <amadeuszx.slawinski at linux.intel.com>
Signed-off-by: Andrew Morton <akpm at linux-foundation.org>
(cherry picked from commit 873f64b791a2b43c246e78b7d9fdd64ce909685b)
Signed-off-by: Philip Cox <philip.cox at canonical.com>
---
 mm/memcontrol.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 6f969ba0d688..f922613ddf8c 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -662,7 +662,14 @@ static inline void memcg_rstat_updated(struct mem_cgroup *memcg, int val)
 
 	x = __this_cpu_add_return(stats_updates, abs(val));
 	if (x > MEMCG_CHARGE_BATCH) {
-		atomic_add(x / MEMCG_CHARGE_BATCH, &stats_flush_threshold);
+		/*
+		 * If stats_flush_threshold exceeds the threshold
+		 * (>num_online_cpus()), cgroup stats update will be triggered
+		 * in __mem_cgroup_flush_stats(). Increasing this var further
+		 * is redundant and simply adds overhead in atomic update.
+		 */
+		if (atomic_read(&stats_flush_threshold) <= num_online_cpus())
+			atomic_add(x / MEMCG_CHARGE_BATCH, &stats_flush_threshold);
 		__this_cpu_write(stats_updates, 0);
 	}
 }
-- 
2.34.1