mm/vmstat.c: erase latency in vmstat_shepherd

Many 100us+ latencies have been deteceted in vmstat_shepherd() on CPX
platform which has 208 logic cpus.  And vmstat_shepherd is queued every
second, which could make the case worse.

Add schedule point in vmstat_shepherd() to erase the latency.

Link: https://lkml.kernel.org/r/20210111035526.1511-1-benbjiang@tencent.com
Signed-off-by: Jiang Biao <benbjiang@tencent.com>
Reported-by: Bin Lai <robinlai@tencent.com>
Cc: Christoph Lameter <cl@linux-foundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

diff --git a/mm/vmstat.c b/mm/vmstat.c
index e60b36f5f0a9ee8c0ba577489f3206869f3a64af..74b2c374b86c8a1728a9a7e118af830796ea9691 100644 (file)
--- a/mm/vmstat.c
+++ b/mm/vmstat.c
@@ -1980,6 +1980,8 @@ static void vmstat_shepherd(struct work_struct *w)
 
                if (!delayed_work_pending(dw) && need_update(cpu))
                        queue_delayed_work_on(cpu, mm_percpu_wq, dw, 0);
+
+               cond_resched();
        }
        put_online_cpus();