[v12,00/16] per memcg lru lock

Message

Alex Shi June 11, 2020, 6:16 a.m. UTC

This is a new version which bases on v5.8, only change mm/compaction.c
since mm-compaction-avoid-vm_bug_onpageslab-in-page_mapcount.patch 
removed.

Johannes Weiner has suggested:
"So here is a crazy idea that may be worth exploring:

Right now, pgdat->lru_lock protects both PageLRU *and* the lruvec's
linked list.

Can we make PageLRU atomic and use it to stabilize the lru_lock
instead, and then use the lru_lock only serialize list operations?
..."

With new memcg charge path and this solution, we could isolate
LRU pages to exclusive visit them in compaction, page migration, reclaim,
memcg move_accunt, huge page split etc scenarios while keeping pages' 
memcg stable. Then possible to change per node lru locking to per memcg
lru locking. As to pagevec_lru_move_fn funcs, it would be safe to let
pages remain on lru list, lru lock could guard them for list integrity.

The patchset includes 3 parts:
1, some code cleanup and minimum optimization as a preparation.
2, use TestCleanPageLRU as page isolation's precondition
3, replace per node lru_lock with per memcg per node lru_lock

The 3rd part moves per node lru_lock into lruvec, thus bring a lru_lock for
each of memcg per node. So on a large machine, each of memcg don't
have to suffer from per node pgdat->lru_lock competition. They could go
fast with their self lru_lock

Following Daniel Jordan's suggestion, I have run 208 'dd' with on 104
containers on a 2s * 26cores * HT box with a modefied case:
https://git.kernel.org/pub/scm/linux/kernel/git/wfg/vm-scalability.git/tree/case-lru-file-readtwice

With this patchset, the readtwice performance increased about 80%
in concurrent containers.

Thanks Hugh Dickins and Konstantin Khlebnikov, they both brought this
idea 8 years ago, and others who give comments as well: Daniel Jordan, 
Mel Gorman, Shakeel Butt, Matthew Wilcox etc.

Thanks for Testing support from Intel 0day and Rong Chen, Fengguang Wu,
and Yun Wang. Hugh Dickins also shared his kbuild-swap case. Thanks!


Alex Shi (14):
  mm/vmscan: remove unnecessary lruvec adding
  mm/page_idle: no unlikely double check for idle page counting
  mm/compaction: correct the comments of compact_defer_shift
  mm/compaction: rename compact_deferred as compact_should_defer
  mm/thp: move lru_add_page_tail func to huge_memory.c
  mm/thp: clean up lru_add_page_tail
  mm/thp: narrow lru locking
  mm/memcg: add debug checking in lock_page_memcg
  mm/lru: introduce TestClearPageLRU
  mm/compaction: do page isolation first in compaction
  mm/mlock: reorder isolation sequence during munlock
  mm/lru: replace pgdat lru_lock with lruvec lock
  mm/lru: introduce the relock_page_lruvec function
  mm/pgdat: remove pgdat lru_lock

Hugh Dickins (2):
  mm/vmscan: use relock for move_pages_to_lru
  mm/lru: revise the comments of lru_lock

 Documentation/admin-guide/cgroup-v1/memcg_test.rst |  15 +-
 Documentation/admin-guide/cgroup-v1/memory.rst     |   8 +-
 Documentation/trace/events-kmem.rst                |   2 +-
 Documentation/vm/unevictable-lru.rst               |  22 +--
 include/linux/compaction.h                         |   4 +-
 include/linux/memcontrol.h                         |  92 +++++++++++
 include/linux/mm_types.h                           |   2 +-
 include/linux/mmzone.h                             |   6 +-
 include/linux/page-flags.h                         |   1 +
 include/linux/swap.h                               |   4 +-
 include/trace/events/compaction.h                  |   2 +-
 mm/compaction.c                                    |  96 +++++++-----
 mm/filemap.c                                       |   4 +-
 mm/huge_memory.c                                   |  51 +++++--
 mm/memcontrol.c                                    |  87 ++++++++++-
 mm/mlock.c                                         |  93 ++++++------
 mm/mmzone.c                                        |   1 +
 mm/page_alloc.c                                    |   1 -
 mm/page_idle.c                                     |   8 -
 mm/rmap.c                                          |   2 +-
 mm/swap.c                                          | 112 ++++----------
 mm/swap_state.c                                    |   6 +-
 mm/vmscan.c                                        | 168 +++++++++++----------
 mm/workingset.c                                    |   4 +-
 24 files changed, 481 insertions(+), 310 deletions(-)

Comments

Hugh Dickins June 11, 2020, 10:26 p.m. UTC | #1

On Thu, 11 Jun 2020, Alex Shi wrote:

> This is a new version which bases on v5.8,

No, not even v5.8-rc1 has come out yet.  v12 applied cleanly on
2dca74a40e1e7ff45079d85fc507769383039b9d but I didn't check the build.

> only change mm/compaction.c
> since mm-compaction-avoid-vm_bug_onpageslab-in-page_mapcount.patch 
> removed.
> 
> Johannes Weiner has suggested:
> "So here is a crazy idea that may be worth exploring:
> 
> Right now, pgdat->lru_lock protects both PageLRU *and* the lruvec's
> linked list.
> 
> Can we make PageLRU atomic and use it to stabilize the lru_lock
> instead, and then use the lru_lock only serialize list operations?
> ..."

It was well worth exploring, and may help in a few cases;
Johannes's memcg swap simplifications have helped a lot more;
but crashes under rotate_reclaimable_page() show that this series
still does not give enough protection from mem_cgroup_move_account().

I'll send a couple of fixes to compaction bugs in reply to this:
with those in, compaction appears to be solid.

Hugh

> 
> With new memcg charge path and this solution, we could isolate
> LRU pages to exclusive visit them in compaction, page migration, reclaim,
> memcg move_accunt, huge page split etc scenarios while keeping pages' 
> memcg stable. Then possible to change per node lru locking to per memcg
> lru locking. As to pagevec_lru_move_fn funcs, it would be safe to let
> pages remain on lru list, lru lock could guard them for list integrity.
> 
> The patchset includes 3 parts:
> 1, some code cleanup and minimum optimization as a preparation.
> 2, use TestCleanPageLRU as page isolation's precondition
> 3, replace per node lru_lock with per memcg per node lru_lock
> 
> The 3rd part moves per node lru_lock into lruvec, thus bring a lru_lock for
> each of memcg per node. So on a large machine, each of memcg don't
> have to suffer from per node pgdat->lru_lock competition. They could go
> fast with their self lru_lock
> 
> Following Daniel Jordan's suggestion, I have run 208 'dd' with on 104
> containers on a 2s * 26cores * HT box with a modefied case:
> https://git.kernel.org/pub/scm/linux/kernel/git/wfg/vm-scalability.git/tree/case-lru-file-readtwice
> 
> With this patchset, the readtwice performance increased about 80%
> in concurrent containers.
> 
> Thanks Hugh Dickins and Konstantin Khlebnikov, they both brought this
> idea 8 years ago, and others who give comments as well: Daniel Jordan, 
> Mel Gorman, Shakeel Butt, Matthew Wilcox etc.
> 
> Thanks for Testing support from Intel 0day and Rong Chen, Fengguang Wu,
> and Yun Wang. Hugh Dickins also shared his kbuild-swap case. Thanks!
> 
> 
> Alex Shi (14):
>   mm/vmscan: remove unnecessary lruvec adding
>   mm/page_idle: no unlikely double check for idle page counting
>   mm/compaction: correct the comments of compact_defer_shift
>   mm/compaction: rename compact_deferred as compact_should_defer
>   mm/thp: move lru_add_page_tail func to huge_memory.c
>   mm/thp: clean up lru_add_page_tail
>   mm/thp: narrow lru locking
>   mm/memcg: add debug checking in lock_page_memcg
>   mm/lru: introduce TestClearPageLRU
>   mm/compaction: do page isolation first in compaction
>   mm/mlock: reorder isolation sequence during munlock
>   mm/lru: replace pgdat lru_lock with lruvec lock
>   mm/lru: introduce the relock_page_lruvec function
>   mm/pgdat: remove pgdat lru_lock
> 
> Hugh Dickins (2):
>   mm/vmscan: use relock for move_pages_to_lru
>   mm/lru: revise the comments of lru_lock
> 
>  Documentation/admin-guide/cgroup-v1/memcg_test.rst |  15 +-
>  Documentation/admin-guide/cgroup-v1/memory.rst     |   8 +-
>  Documentation/trace/events-kmem.rst                |   2 +-
>  Documentation/vm/unevictable-lru.rst               |  22 +--
>  include/linux/compaction.h                         |   4 +-
>  include/linux/memcontrol.h                         |  92 +++++++++++
>  include/linux/mm_types.h                           |   2 +-
>  include/linux/mmzone.h                             |   6 +-
>  include/linux/page-flags.h                         |   1 +
>  include/linux/swap.h                               |   4 +-
>  include/trace/events/compaction.h                  |   2 +-
>  mm/compaction.c                                    |  96 +++++++-----
>  mm/filemap.c                                       |   4 +-
>  mm/huge_memory.c                                   |  51 +++++--
>  mm/memcontrol.c                                    |  87 ++++++++++-
>  mm/mlock.c                                         |  93 ++++++------
>  mm/mmzone.c                                        |   1 +
>  mm/page_alloc.c                                    |   1 -
>  mm/page_idle.c                                     |   8 -
>  mm/rmap.c                                          |   2 +-
>  mm/swap.c                                          | 112 ++++----------
>  mm/swap_state.c                                    |   6 +-
>  mm/vmscan.c                                        | 168 +++++++++++----------
>  mm/workingset.c                                    |   4 +-
>  24 files changed, 481 insertions(+), 310 deletions(-)
> 
> -- 
> 1.8.3.1

Alex Shi June 12, 2020, 3:09 a.m. UTC | #2

在 2020/6/12 上午6:26, Hugh Dickins 写道:
>> ..."
> It was well worth exploring, and may help in a few cases;
> Johannes's memcg swap simplifications have helped a lot more;
> but crashes under rotate_reclaimable_page() show that this series
> still does not give enough protection from mem_cgroup_move_account().
> 
> I'll send a couple of fixes to compaction bugs in reply to this:
> with those in, compaction appears to be solid.
> 

Thanks a lot for fixing. I will look into them and try to merge into
the patchset.

Thanks a lot!
Alex