sched: Reduce contention in update_cfs_rq_blocked_load()
When running workloads on 2+ socket systems, based on perf profiles, the
update_cfs_rq_blocked_load() function often shows up as taking up a
noticeable % of run time.
Much of the contention is in __update_cfs_rq_tg_load_contrib() when we
update the tg load contribution stats. However, it turns out that in many
cases, they don't need to be updated and "tg_contrib" is 0.
This patch adds a check in __update_cfs_rq_tg_load_contrib() to skip updating
tg load contribution stats when nothing needs to be updated. This reduces the
cacheline contention that would be unnecessary.
Reviewed-by: Ben Segall <firstname.lastname@example.org>
Reviewed-by: Waiman Long <Waiman.Long@hp.com>
Signed-off-by: Jason Low <email@example.com>
Signed-off-by: Peter Zijlstra <firstname.lastname@example.org>
Cc: Paul Turner <email@example.com>
Cc: Yuyang Du <firstname.lastname@example.org>
Cc: Aswin Chandramouleeswaran <email@example.com>
Cc: Chegu Vinod <firstname.lastname@example.org>
Cc: Scott J Norton <email@example.com>
Cc: Tim Chen <firstname.lastname@example.org>
Cc: Linus Torvalds <email@example.com>
Signed-off-by: Ingo Molnar <firstname.lastname@example.org>
1 file changed