Files
kernel/include/linux
Feng Tang 4df910620b mm: memcg: relayout structure mem_cgroup to avoid cache interference
0day reported one -22.7% regression for will-it-scale page_fault2
case [1] on a 4 sockets 144 CPU platform, and bisected to it to be
caused by Waiman's optimization (commit bd0b230fe1) of saving one
'struct page_counter' space for 'struct mem_cgroup'.

Initially we thought it was due to the cache alignment change introduced
by the patch, but further debug shows that it is due to some hot data
members ('vmstats_local', 'vmstats_percpu', 'vmstats') sit in 2 adjacent
cacheline (2N and 2N+1 cacheline), and when adjacent cache line prefetch
is enabled, it triggers an "extended level" of cache false sharing for
2 adjacent cache lines.

So exchange the 2 member blocks, while keeping mostly the original
cache alignment, which can restore and even enhance the performance,
and save 64 bytes of space for 'struct mem_cgroup' (from 2880 to 2816,
with 0day's default RHEL-8.3 kernel config)

[1]. https://lore.kernel.org/lkml/20201102091543.GM31092@shao2-debian/

Fixes: bd0b230fe1 ("mm/memcg: unify swap and memsw page counters")
Reported-by: kernel test robot <rong.a.chen@intel.com>
Signed-off-by: Feng Tang <feng.tang@intel.com>
Acked-by: Waiman Long <longman@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-11-26 09:35:49 -08:00
..
2020-10-28 13:18:56 +01:00
2020-10-02 14:59:25 -07:00
2020-09-21 15:00:40 -07:00
2020-10-02 15:00:49 -07:00
2020-10-16 17:21:51 +02:00
2020-10-27 16:22:43 +01:00
2020-09-30 22:44:26 +02:00
2020-11-19 22:38:29 -05:00
2020-10-13 18:38:32 -07:00
2020-09-23 18:02:49 -07:00
2020-10-07 14:28:39 -04:00
2020-09-24 19:49:36 -07:00
2020-10-28 11:41:15 -06:00
2020-10-18 09:27:10 -07:00
2020-11-06 10:05:18 -08:00
2020-09-26 22:55:05 -04:00
2020-10-18 09:27:10 -07:00
2020-10-05 13:21:49 +02:00