public inbox for linux-s390@vger.kernel.org
 help / color / mirror / Atom feed
* [PATCH 00/25] mm: Page fault accounting cleanups
@ 2020-06-15 22:15 Peter Xu
  2020-06-15 22:23 ` [PATCH 19/25] mm/s390: Use mm_fault_accounting() Peter Xu
  2020-06-16 18:55 ` [PATCH 00/25] mm: Page fault accounting cleanups Linus Torvalds
  0 siblings, 2 replies; 13+ messages in thread
From: Peter Xu @ 2020-06-15 22:15 UTC (permalink / raw)
  To: linux-kernel
  Cc: Gerald Schaefer, Andrew Morton, peterx, Linus Torvalds,
	Andrea Arcangeli

Gerald Schaefer reported an issue a week ago regarding to incorrect page fault
accountings for retried page fault after commit 4064b9827063 ("mm: allow
VM_FAULT_RETRY for multiple times"):

  https://lore.kernel.org/lkml/20200610174811.44b94525@thinkpad/

When I was looking at the issue, I actually found some other trivial issues too
related to the page fault accounting, namely:

  - Incorrect accountings

    - The major issue that was reported by Gerald, that we did multiple
      accounting when the page fault retried while we should only do it once
      (for either perf event accounting or per-task accounting).

    - In many archs, major fault is only accounted when the 1st page fault is a
      major fault, or the last page fault is a major fault.  Ideally we should
      account the page fault as a major fault as long as any of the page fault
      retries is a major fault and keep the same behavior across archs.

  - Missing accountings of perf events (PERF_COUNT_SW_PAGE_FAULTS[_[MAJ|MIN]]).

  - Doing accounting with mmap_sem: not a big deal, but logically we can move
    it after releasing the mmap_sem because accounting does not need it.

This series tries to address all of them by introducing mm_fault_accounting()
first, so that we move all the page fault accounting into the common code base,
then call it properly from arch pf handlers just like handle_mm_fault().  There
are some special cases, e.g., the um arch does not have "regs" in the fault
handler, so it's still using the manual statistics.

For each of the patch that fixes a specific arch, I'm CCing the maintainers and
the arch list if there is.  Besides, I only lightly tested this series on x86.

Please have a look, thanks.

Peter Xu (25):
  mm/um: Fix extra accounting for page fault retries
  mm: Introduce mm_fault_accounting()
  mm/alpha: Use mm_fault_accounting()
  mm/arc: Use mm_fault_accounting()
  mm/arm: Use mm_fault_accounting()
  mm/arm64: Use mm_fault_accounting()
  mm/csky: Use mm_fault_accounting()
  mm/hexagon: Use mm_fault_accounting()
  mm/ia64: Use mm_fault_accounting()
  mm/m68k: Use mm_fault_accounting()
  mm/microblaze: Use mm_fault_accounting()
  mm/mips: Use mm_fault_accounting()
  mm/nds32: Use mm_fault_accounting()
  mm/nios2: Use mm_fault_accounting()
  mm/openrisc: Use mm_fault_accounting()
  mm/parisc: Use mm_fault_accounting()
  mm/powerpc: Use mm_fault_accounting()
  mm/riscv: Use mm_fault_accounting()
  mm/s390: Use mm_fault_accounting()
  mm/sh: Use mm_fault_accounting()
  mm/sparc32: Use mm_fault_accounting()
  mm/sparc64: Use mm_fault_accounting()
  mm/unicore32: Use mm_fault_accounting()
  mm/x86: Use mm_fault_accounting()
  mm/xtensa: Use mm_fault_accounting()

 arch/alpha/mm/fault.c      |  9 ++++-----
 arch/arc/mm/fault.c        | 15 +++------------
 arch/arm/mm/fault.c        | 21 ++++-----------------
 arch/arm64/mm/fault.c      | 17 ++---------------
 arch/csky/mm/fault.c       | 13 +------------
 arch/hexagon/mm/vm_fault.c |  9 +++------
 arch/ia64/mm/fault.c       |  8 +++-----
 arch/m68k/mm/fault.c       | 13 +++----------
 arch/microblaze/mm/fault.c |  8 +++-----
 arch/mips/mm/fault.c       | 14 +++-----------
 arch/nds32/mm/fault.c      | 19 +++----------------
 arch/nios2/mm/fault.c      | 13 +++----------
 arch/openrisc/mm/fault.c   |  8 +++-----
 arch/parisc/mm/fault.c     |  8 +++-----
 arch/powerpc/mm/fault.c    | 13 ++++---------
 arch/riscv/mm/fault.c      | 21 +++------------------
 arch/s390/mm/fault.c       | 21 +++++----------------
 arch/sh/mm/fault.c         | 15 +++------------
 arch/sparc/mm/fault_32.c   | 16 +++-------------
 arch/sparc/mm/fault_64.c   | 16 ++++------------
 arch/um/kernel/trap.c      | 13 +++++++------
 arch/unicore32/mm/fault.c  | 15 +++++++--------
 arch/x86/mm/fault.c        | 10 +---------
 arch/xtensa/mm/fault.c     | 14 +++-----------
 include/linux/mm.h         |  2 ++
 mm/memory.c                | 18 ++++++++++++++++++
 26 files changed, 101 insertions(+), 248 deletions(-)

-- 
2.26.2

^ permalink raw reply	[flat|nested] 13+ messages in thread
end of thread, other threads:[~2020-06-17 16:44 UTC | newest]

Thread overview: 13+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2020-06-15 22:15 [PATCH 00/25] mm: Page fault accounting cleanups Peter Xu
2020-06-15 22:23 ` [PATCH 19/25] mm/s390: Use mm_fault_accounting() Peter Xu
2020-06-16 15:59   ` Alexander Gordeev
2020-06-16 16:35     ` Peter Xu
2020-06-17  6:19       ` Christian Borntraeger
2020-06-17 16:06         ` Peter Xu
2020-06-17 16:14           ` Christian Borntraeger
2020-06-17 16:44             ` Peter Xu
2020-06-16 18:55 ` [PATCH 00/25] mm: Page fault accounting cleanups Linus Torvalds
2020-06-16 21:03   ` Peter Xu
2020-06-17  0:55   ` Michael Ellerman
2020-06-17  8:04     ` Will Deacon
2020-06-17 16:10       ` Peter Xu

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox