[GIT PULL] Additional MM updates for 6.16-rc1

[GIT PULL] Additional MM updates for 6.16-rc1

[Andrew Morton]

Linus, please merge this second batch of MM updates for the 6.16-rcX cycle.

I'm seeing a single small conflict in mm/memcontrol.c.  Stephen's
resolution is at
https://lore.kernel.org/all/20250529111223.3987a514@canb.auug.org.au/T/#u

Thanks.


The following changes since commit c544a952ba61b1a025455098033c17e0573ab085:

  mm: pcp: increase pcp->free_count threshold to trigger free_high (2025-05-27 19:38:27 -0700)

are available in the Git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm tags/mm-stable-2025-06-01-14-06

for you to fetch changes up to 0b43b8bc8ef88bb45b018b2d4853d38bfc5ce2a7:

  mm/khugepaged: clean up refcount check using folio_expected_ref_count() (2025-05-31 22:46:16 -0700)

----------------------------------------------------------------
- The 2 patch series "zram: support algorithm-specific parameters" from
  Sergey Senozhatsky adds infrastructure for passing algorithm-specific
  parameters into zram.  A single parameter `winbits' is implemented at
  this time.

- The 5 patch series "memcg: nmi-safe kmem charging" from Shakeel Butt
  makes memcg charging nmi-safe, which is required by BFP, which can
  operate in NMI context.

- The 5 patch series "Some random fixes and cleanup to shmem" from
  Kemeng Shi implements small fixes and cleanups in the shmem code.

- The 2 patch series "Skip mm selftests instead when kernel features are
  not present" from Zi Yan fixes some issues in the MM selftest code.

- The 2 patch series "mm/damon: build-enable essential DAMON components
  by default" from SeongJae Park reworks DAMON Kconfig to make it easier
  to enable CONFIG_DAMON.

- The 2 patch series "sched/numa: add statistics of numa balance task
  migration" from Libo Chen adds more info into sysfs and procfs files to
  improve visibility into the NUMA balancer's task migration activity.

- The 4 patch series "selftests/mm: cow and gup_longterm cleanups" from
  Mark Brown provides various updates to some of the MM selftests to make
  them play better with the overall containing framework.

----------------------------------------------------------------
Akinobu Mita (1):
      mm/damon/core: avoid destroyed target reference from DAMOS quota

Alice Ryhl (2):
      mm: rust: make CONFIG_MMU ifdefs more narrow
      kcov: rust: add flags for KCOV with Rust

Chen Yu (1):
      sched/numa: add statistics of numa balance task

Dan Carpenter (1):
      tools/testing: check correct variable in open_procmap()

David Hildenbrand (1):
      selftests/mm: two fixes for the pfnmap test

Enze Li (1):
      selftests/damon/_damon_sysfs: skip testcases if CONFIG_DAMON_SYSFS is disabled

Jann Horn (2):
      mmu_notifiers: remove leftover stub macros
      mm/gup: update comment explaining why gup_fast() disables IRQs

Juan Yescas (1):
      mm: add CONFIG_PAGE_BLOCK_ORDER to select page block order

Kemeng Shi (5):
      mm: shmem: avoid unpaired folio_unlock() in shmem_swapin_folio()
      mm: shmem: add missing shmem_unacct_size() in __shmem_file_setup()
      mm/shmem: fix potential dead loop in shmem_unuse()
      mm: shmem: only remove inode from swaplist when it's swapped page count is 0
      mm/shmem: remove unneeded xa_is_value() check in shmem_unuse_swap_entries()

Libo Chen (1):
      sched/numa: fix task swap by skipping kernel threads

Lorenzo Stoakes (1):
      tools/testing/vma: add missing function stub

Mark Brown (7):
      selftests/mm: deduplicate test logging in test_mlock_lock()
      selftests/mm: deduplicate default page size test results in thuge-gen
      selftests/mm: deduplicate test names in madv_populate
      selftests/mm: use standard ksft_finished() in cow and gup_longterm
      selftests/mm: add helper for logging test start and results
      selftests/mm: report unique test names for each cow test
      selftests/mm: fix test result reporting in gup_longterm

Matthew Wilcox (Oracle) (4):
      m68k: remove use of page->index
      mm: rename page->index to page->__folio_index
      ntfs3: use folios more in ntfs_compress_write()
      iov: remove copy_page_from_iter_atomic()

Roman Gushchin (1):
      mmu_gather: move tlb flush for VM_PFNMAP/VM_MIXEDMAP vmas into free_pgtables()

SeongJae Park (2):
      mm/damon/Kconfig: set DAMON_{VADDR,PADDR,SYSFS} default to DAMON
      mm/damon/Kconfig: enable CONFIG_DAMON by default

Sergey Senozhatsky (2):
      zram: rename ZCOMP_PARAM_NO_LEVEL
      zram: support deflate-specific params

Shakeel Butt (5):
      memcg: disable kmem charging in nmi for unsupported arch
      memcg: nmi safe memcg stats for specific archs
      memcg: add nmi-safe update for MEMCG_KMEM
      memcg: nmi-safe slab stats updates
      memcg: make memcg_rstat_updated nmi safe

Shivank Garg (2):
      mm/khugepaged: fix race with folio split/free using temporary reference
      mm/khugepaged: clean up refcount check using folio_expected_ref_count()

Wenjie Xu (1):
      hugetlb: show nr_huge_pages in report_hugepages()

Zi Yan (2):
      selftests/mm: skip guard_regions.uffd tests when uffd is not present
      selftests/mm: skip hugevm test if kernel config file is not present

 Documentation/admin-guide/cgroup-v2.rst           |   6 +
 arch/arm/mm/flush.c                               |   4 +-
 arch/m68k/mm/motorola.c                           |   3 +-
 drivers/block/zram/backend_deflate.c              |  12 +-
 drivers/block/zram/backend_lz4.c                  |   2 +-
 drivers/block/zram/backend_lz4hc.c                |   2 +-
 drivers/block/zram/backend_zstd.c                 |   2 +-
 drivers/block/zram/zcomp.h                        |   9 +-
 drivers/block/zram/zram_drv.c                     |  21 +-
 fs/ntfs3/file.c                                   |  31 +-
 include/asm-generic/tlb.h                         |  46 ++-
 include/linux/memcontrol.h                        |  10 +
 include/linux/mm.h                                |   6 +-
 include/linux/mm_types.h                          |   6 +-
 include/linux/mmu_notifier.h                      |   3 -
 include/linux/mmzone.h                            |  16 +
 include/linux/pageblock-flags.h                   |   8 +-
 include/linux/sched.h                             |   4 +
 include/linux/uio.h                               |  10 +-
 include/linux/vm_event_item.h                     |   2 +
 init/Kconfig                                      |  14 +
 kernel/futex/core.c                               |   2 +-
 kernel/sched/core.c                               |   9 +-
 kernel/sched/debug.c                              |   4 +
 kernel/sched/fair.c                               |   3 +-
 lib/iov_iter.c                                    |  29 +-
 mm/Kconfig                                        |  34 +++
 mm/damon/Kconfig                                  |   4 +
 mm/damon/core.c                                   |   8 +
 mm/filemap.c                                      |   4 +-
 mm/gup.c                                          |   2 +-
 mm/hugetlb.c                                      |   2 +-
 mm/khugepaged.c                                   |  35 ++-
 mm/memcontrol.c                                   | 127 +++++++-
 mm/memory.c                                       |   6 +-
 mm/mm_init.c                                      |   2 +-
 mm/mmu_gather.c                                   |   1 +
 mm/page-writeback.c                               |   6 +-
 mm/shmem.c                                        |  23 +-
 mm/truncate.c                                     |   2 +-
 mm/vmstat.c                                       |   2 +
 mm/zpdesc.h                                       |   4 +-
 rust/Makefile                                     |   1 +
 rust/kernel/mm.rs                                 |  56 +---
 rust/kernel/mm/mmput_async.rs                     |  68 +++++
 scripts/Makefile.kcov                             |   6 +
 scripts/Makefile.lib                              |   3 +
 tools/testing/selftests/damon/_damon_sysfs.py     |   4 +
 tools/testing/selftests/mm/cow.c                  | 340 ++++++++++++++--------
 tools/testing/selftests/mm/guard-regions.c        |  17 +-
 tools/testing/selftests/mm/gup_longterm.c         | 158 ++++++----
 tools/testing/selftests/mm/madv_populate.c        |  18 +-
 tools/testing/selftests/mm/mlock2-tests.c         |   2 +-
 tools/testing/selftests/mm/pfnmap.c               |  61 +++-
 tools/testing/selftests/mm/thuge-gen.c            |   4 +-
 tools/testing/selftests/mm/va_high_addr_switch.sh |  26 +-
 tools/testing/selftests/mm/vm_util.c              |   2 +-
 tools/testing/selftests/mm/vm_util.h              |  20 ++
 tools/testing/vma/vma_internal.h                  |   5 +
 59 files changed, 909 insertions(+), 408 deletions(-)
 create mode 100644 rust/kernel/mm/mmput_async.rs

Re: [GIT PULL] Additional MM updates for 6.16-rc1

[Matthew Wilcox]

On Sun, Jun 01, 2025 at 03:12:22PM -0700, Andrew Morton wrote:
> Linus, please merge this second batch of MM updates for the 6.16-rcX cycle.

A word of warning for Linus since he wouldn't've been cc'd on the
earlier email.  This branch is based on v6.15-rc6 but contains a
commit which depends on 97dfbbd135cb which was merged during the
current merge window.  You might want to do a rebase or ask Andrew to
do it in order to prevent a bisection hazard.

0day-ci found the problem here:
https://lore.kernel.org/linux-mm/202506022027.IYQzZghL-lkp@intel.com/

Re: [GIT PULL] Additional MM updates for 6.16-rc1

[Andrew Morton]

On Mon, 2 Jun 2025 16:05:38 +0100 Matthew Wilcox <willy@infradead.org> wrote:

> On Sun, Jun 01, 2025 at 03:12:22PM -0700, Andrew Morton wrote:
> > Linus, please merge this second batch of MM updates for the 6.16-rcX cycle.
> 
> A word of warning for Linus since he wouldn't've been cc'd on the
> earlier email.  This branch is based on v6.15-rc6 but contains a
> commit which depends on 97dfbbd135cb which was merged during the
> current merge window.  You might want to do a rebase or ask Andrew to
> do it in order to prevent a bisection hazard.
> 
> 0day-ci found the problem here:
> https://lore.kernel.org/linux-mm/202506022027.IYQzZghL-lkp@intel.com/

I think it's OK?  I did the merge and 

	d9736929445e iov: remove copy_page_from_iter_atomic()

lands later than

	97dfbbd135cb highmem: add folio_test_partial_kmap()

Re: [GIT PULL] Additional MM updates for 6.16-rc1

[Matthew Wilcox]

On Mon, Jun 02, 2025 at 12:17:51PM -0700, Andrew Morton wrote:
> On Mon, 2 Jun 2025 16:05:38 +0100 Matthew Wilcox <willy@infradead.org> wrote:
> 
> > On Sun, Jun 01, 2025 at 03:12:22PM -0700, Andrew Morton wrote:
> > > Linus, please merge this second batch of MM updates for the 6.16-rcX cycle.
> > 
> > A word of warning for Linus since he wouldn't've been cc'd on the
> > earlier email.  This branch is based on v6.15-rc6 but contains a
> > commit which depends on 97dfbbd135cb which was merged during the
> > current merge window.  You might want to do a rebase or ask Andrew to
> > do it in order to prevent a bisection hazard.
> > 
> > 0day-ci found the problem here:
> > https://lore.kernel.org/linux-mm/202506022027.IYQzZghL-lkp@intel.com/
> 
> I think it's OK?  I did the merge and 
> 
> 	d9736929445e iov: remove copy_page_from_iter_atomic()
> 
> lands later than
> 
> 	97dfbbd135cb highmem: add folio_test_partial_kmap()

Yes, but if 'git bisect' lands you on there being a problem in 0b43b8bc8ef8
you won't be able to build it because 0b43b8bc8ef8 has d9736929445e as
an ancestor but not 97dfbbd135cb .  You'll need to add a cherry-pick of
97dfbbd135cb which is aggravating and we try to avoid this.

Re: [GIT PULL] Additional MM updates for 6.16-rc1

[Linus Torvalds]

On Mon, 2 Jun 2025 at 08:05, Matthew Wilcox <willy@infradead.org> wrote:
>
> A word of warning for Linus since he wouldn't've been cc'd on the
> earlier email.  This branch is based on v6.15-rc6 but contains a
> commit which depends on 97dfbbd135cb which was merged during the
> current merge window.  You might want to do a rebase or ask Andrew to
> do it in order to prevent a bisection hazard.

Gah.

If I rebase the thing, I lose the pgp signature.

This is not optimal, but let's hope nobody hits the bisect issue.

            Linus

Re: [GIT PULL] Additional MM updates for 6.16-rc1

[Geert Uytterhoeven]

Hi Andrew et al,

On Mon, 2 Jun 2025 at 17:55, Andrew Morton <akpm@linux-foundation.org> wrote:
> - The 2 patch series "mm/damon: build-enable essential DAMON components
>   by default" from SeongJae Park reworks DAMON Kconfig to make it easier
>   to enable CONFIG_DAMON.

... or, make it harder to disable it?

Given no single defconfig file in v6.15 enables CONFIG_DAMON, I find
it hard to believe defaulting DAMON to "y" is the right thing to do...
(Yes, I have read the rationale in commit 28615e6eed152f2f
("mm/damon/Kconfig: enable CONFIG_DAMON by default")).

Thanks!

Gr{oetje,eeting}s,

                        Geert


--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [GIT PULL] Additional MM updates for 6.16-rc1

[Andrew Morton]

On Tue, 10 Jun 2025 11:05:40 +0200 Geert Uytterhoeven <geert@linux-m68k.org> wrote:

> Hi Andrew et al,
> 
> On Mon, 2 Jun 2025 at 17:55, Andrew Morton <akpm@linux-foundation.org> wrote:
> > - The 2 patch series "mm/damon: build-enable essential DAMON components
> >   by default" from SeongJae Park reworks DAMON Kconfig to make it easier
> >   to enable CONFIG_DAMON.
> 
> ... or, make it harder to disable it?
> 
> Given no single defconfig file in v6.15 enables CONFIG_DAMON, I find
> it hard to believe defaulting DAMON to "y" is the right thing to do...
> (Yes, I have read the rationale in commit 28615e6eed152f2f
> ("mm/damon/Kconfig: enable CONFIG_DAMON by default")).
> 

So what do you recommend?  Editing all the defconfigs seems a bit lame.

Re: [GIT PULL] Additional MM updates for 6.16-rc1

[Geert Uytterhoeven]

Hi Andrew,

On Tue, 10 Jun 2025 at 17:54, Andrew Morton <akpm@linux-foundation.org> wrote:
> On Tue, 10 Jun 2025 11:05:40 +0200 Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> > On Mon, 2 Jun 2025 at 17:55, Andrew Morton <akpm@linux-foundation.org> wrote:
> > > - The 2 patch series "mm/damon: build-enable essential DAMON components
> > >   by default" from SeongJae Park reworks DAMON Kconfig to make it easier
> > >   to enable CONFIG_DAMON.
> >
> > ... or, make it harder to disable it?
> >
> > Given no single defconfig file in v6.15 enables CONFIG_DAMON, I find
> > it hard to believe defaulting DAMON to "y" is the right thing to do...
> > (Yes, I have read the rationale in commit 28615e6eed152f2f
> > ("mm/damon/Kconfig: enable CONFIG_DAMON by default")).
>
> So what do you recommend?  Editing all the defconfigs seems a bit lame.

Just revert the commit?
Distros are (usually) not using defconfig files anyway.
Thanks!

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [GIT PULL] Additional MM updates for 6.16-rc1

[SeongJae Park]

Hi Geert,

On Tue, 10 Jun 2025 18:41:43 +0200 Geert Uytterhoeven <geert@linux-m68k.org> wrote:

> Hi Andrew,
> 
> On Tue, 10 Jun 2025 at 17:54, Andrew Morton <akpm@linux-foundation.org> wrote:
> > On Tue, 10 Jun 2025 11:05:40 +0200 Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> > > On Mon, 2 Jun 2025 at 17:55, Andrew Morton <akpm@linux-foundation.org> wrote:
> > > > - The 2 patch series "mm/damon: build-enable essential DAMON components
> > > >   by default" from SeongJae Park reworks DAMON Kconfig to make it easier
> > > >   to enable CONFIG_DAMON.
> > >
> > > ... or, make it harder to disable it?
> > >
> > > Given no single defconfig file in v6.15 enables CONFIG_DAMON, I find
> > > it hard to believe defaulting DAMON to "y" is the right thing to do...
> > > (Yes, I have read the rationale in commit 28615e6eed152f2f
> > > ("mm/damon/Kconfig: enable CONFIG_DAMON by default")).

I'm not clearly seeing what is your concern.  Could I ask you more
elaborations?  I'm not that familiar with distros' config setup process and
defconfigs, so this might be just a silly question, but please bear in mind
with me.  I'd like to clearly understand this to avoid making similar problems
later.

> >
> > So what do you recommend?  Editing all the defconfigs seems a bit lame.
> 
> Just revert the commit?
> Distros are (usually) not using defconfig files anyway.

Reverting the commit is not a big problem for me.  But, it would be nice if I
could better understand and agree your concerns, and learn how I could do
better next time.


Thanks,
SJ

[...]

Re: [GIT PULL] Additional MM updates for 6.16-rc1

[Linus Torvalds]

On Tue, 10 Jun 2025 at 09:41, Geert Uytterhoeven <geert@linux-m68k.org> wrote:
>
> Just revert the commit?

Yes. We don't do "default y" for random features.

EVERY SINGLE DEVELOPER thinks that *their* feature is so important
that everybody else should enable it.

And if that feature wasn't enabled before, they are completely wrong
99% of the time.

There is a very real reason why 'default' defaults to 'n'.

So we do *not* use "default y" for features that aren't universal.

The only time we should use 'default y' is if some old feature that
used to be unconditional gets split up into a new Kconfig variable and
not using 'default y' means that people *lose* functionality.

Or if the feature is some critical security thing, or is some hardware
thing that has become so universal that you find it on basically every
single machine.

Or if that feature cures cancer or brings world peace.

Then you can enable it by default.

I have reverted that 'default y' thing, because I see no reason to
believe that DAEMON cures cancer.

               Linus

Re: [GIT PULL] Additional MM updates for 6.16-rc1

[SeongJae Park]

On Tue, 10 Jun 2025 10:27:46 -0700 Linus Torvalds <torvalds@linux-foundation.org> wrote:

> On Tue, 10 Jun 2025 at 09:41, Geert Uytterhoeven <geert@linux-m68k.org> wrote:
> >
> > Just revert the commit?
> 
> Yes. We don't do "default y" for random features.
> 
> EVERY SINGLE DEVELOPER thinks that *their* feature is so important
> that everybody else should enable it.
> 
> And if that feature wasn't enabled before, they are completely wrong
> 99% of the time.
> 
> There is a very real reason why 'default' defaults to 'n'.
> 
> So we do *not* use "default y" for features that aren't universal.
> 
> The only time we should use 'default y' is if some old feature that
> used to be unconditional gets split up into a new Kconfig variable and
> not using 'default y' means that people *lose* functionality.
> 
> Or if the feature is some critical security thing, or is some hardware
> thing that has become so universal that you find it on basically every
> single machine.
> 
> Or if that feature cures cancer or brings world peace.
> 
> Then you can enable it by default.
> 
> I have reverted that 'default y' thing, because I see no reason to
> believe that DAEMON cures cancer.

Makes sense, and thank you for explaining these all.  I will keep thse in my
mind to avoid making same mistakes in future.


Thanks,
SJ

[...]