[PATCH v2 0/8] per lruvec lru_lock for memcg
From: Alex Shi
Date: Tue Nov 12 2019 - 09:06:44 EST
Hi all,
This patchset move lru_lock into lruvec, give a lru_lock for each of
lruvec, thus bring a lru_lock for each of memcg per node.
According to Daniel Jordan's suggestion, I run 64 'dd' with on 32
containers on my 2s* 8 core * HT box with the modefied case:
https://git.kernel.org/pub/scm/linux/kernel/git/wfg/vm-scalability.git/tree/case-lru-file-readtwice
With this change above lru_lock censitive testing improved 17% with multiple
containers scenario. And no performance lose w/o mem_cgroup.
Thanks Hugh Dickins and Konstantin Khlebnikov, they both bring the same idea
7 years ago. I don't know why they didn't go further, but according to my
testing, and google internal usage. This feathre is clearly benefit
multi-container user.
So I like to introduce it here.
v2: bypass a performance regression bug and fix some function issues
---
Documentation/admin-guide/cgroup-v1/memcg_test.rst | 15 +++------------
Documentation/admin-guide/cgroup-v1/memory.rst | 6 +++---
Documentation/trace/events-kmem.rst | 2 +-
Documentation/vm/unevictable-lru.rst | 22 ++++++++--------------
include/linux/memcontrol.h | 67 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
include/linux/mm_types.h | 2 +-
include/linux/mmzone.h | 7 +++++--
mm/compaction.c | 62 ++++++++++++++++++++++++++++++++++++++++++--------------------
mm/filemap.c | 4 ++--
mm/huge_memory.c | 16 ++++++----------
mm/memcontrol.c | 64 +++++++++++++++++++++++++++++++++++++++++++++++++++-------------
mm/mlock.c | 27 ++++++++++++++-------------
mm/mmzone.c | 1 +
mm/page_alloc.c | 1 -
mm/page_idle.c | 5 +++--
mm/rmap.c | 2 +-
mm/swap.c | 77 +++++++++++++++++++++++++++++++----------------------------------------------
mm/vmscan.c | 74 ++++++++++++++++++++++++++++++++++++++------------------------------------
18 files changed, 277 insertions(+), 177 deletions(-)
[PATCH v2 1/8] mm/lru: add per lruvec lock for memcg
[PATCH v2 2/8] mm/lruvec: add irqsave flags into lruvec struct
[PATCH v2 3/8] mm/lru: replace pgdat lru_lock with lruvec lock
[PATCH v2 4/8] mm/lru: only change the lru_lock iff page's lruvec is
[PATCH v2 5/8] mm/pgdat: remove pgdat lru_lock
[PATCH v2 6/8] mm/lru: remove rcu_read_lock to fix performance
[PATCH v2 7/8] mm/lru: likely enhancement
[PATCH v2 8/8] mm/lru: revise the comments of lru_lock