[GIT PULL] percpu changes for v5.14-rc1

[Dennis Zhou]

Hi Linus,

This pull request contains:
 - percpu chunk depopulation - depopulate backing pages for chunks with
   empty pages when we exceed a global threshold without those pages.
   This lets us reclaim a portion of memory that would previously be
   lost until the full chunk would be freed (possibly never).

 - memcg accounting cleanup - previously separate chunks were managed
   for normal allocations and __GFP_ACCOUNT allocations. These are now
   consolidated which cleans up the code quite a bit.

 - a few misc clean ups for clang warnings

Two notes though. First there is a small conflict in mm/memcontrol.c and
I have an example merge in percpu#for-5.14-merge [1]. Second, I have an
outstanding patch to do some level of batching for percpu depopulation.
Oliver Wang (krobot tests) is reporting a regression only on Knights
Landing cpus [2]. I am working on debugging it and hopefully will have
it figured out soon.

[1] https://git.kernel.org/pub/scm/linux/kernel/git/dennis/percpu.git/log/?h=for-5.14-merge
[2] https://lore.kernel.org/lkml/20210630074500.GB10248@xsang-OptiPlex-9020/

Thanks,
Dennis

The following changes since commit d434405aaab7d0ebc516b68a8fc4100922d7f5ef:

  Linux 5.12-rc7 (2021-04-11 15:16:13 -0700)

are available in the Git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/dennis/percpu.git for-5.14

for you to fetch changes up to e4d777003a43feab2e000749163e531f6c48c385:

  percpu: optimize locking in pcpu_balance_workfn() (2021-06-17 23:05:24 +0000)

----------------------------------------------------------------
Dennis Zhou (2):
      percpu: use pcpu_free_slot instead of pcpu_nr_slots - 1
      percpu: initialize best_upa variable

Roman Gushchin (8):
      percpu: fix a comment about the chunks ordering
      percpu: split __pcpu_balance_workfn()
      percpu: factor out pcpu_check_block_hint()
      percpu: implement partial chunk depopulation
      mm, memcg: mark cgroup_memory_nosocket, nokmem and noswap as __ro_after_init
      mm, memcg: introduce mem_cgroup_kmem_disabled()
      percpu: rework memcg accounting
      percpu: optimize locking in pcpu_balance_workfn()

Wei Yongjun (1):
      percpu: make symbol 'pcpu_free_slot' static

 include/linux/memcontrol.h |   5 +
 mm/memcontrol.c            |  11 +-
 mm/percpu-internal.h       |  56 +------
 mm/percpu-km.c             |  10 +-
 mm/percpu-stats.c          |  44 ++----
 mm/percpu-vm.c             |  35 ++++-
 mm/percpu.c                | 368 +++++++++++++++++++++++++++++++++------------
 7 files changed, 343 insertions(+), 186 deletions(-)