linux.git/mm/memory-failure.c, branch v6.16

mm/vmscan: fix hwpoisoned large folio handling in shrink_folio_list

2025-07-20T02:26:15+00:00

In shrink_folio_list(), the hwpoisoned folio may be large folio, which
can't be handled by unmap_poisoned_folio().  For THP, try_to_unmap_one()
must be passed with TTU_SPLIT_HUGE_PMD to split huge PMD first and then
retry.  Without TTU_SPLIT_HUGE_PMD, we will trigger null-ptr deref of
pvmw.pte.  Even we passed TTU_SPLIT_HUGE_PMD, we will trigger a
WARN_ON_ONCE due to the page isn't in swapcache.

Since UCE is rare in real world, and race with reclaimation is more rare,
just skipping the hwpoisoned large folio is enough.  memory_failure() will
handle it if the UCE is triggered again.

This happens when memory reclaim for large folio races with
memory_failure(), and will lead to kernel panic.  The race is as
follows:

cpu0      cpu1
 shrink_folio_list memory_failure
  TestSetPageHWPoison
  unmap_poisoned_folio
  --> trigger BUG_ON due to
  unmap_poisoned_folio couldn't
   handle large folio

[tujinjiang@huawei.com: add comment to unmap_poisoned_folio()]
  Link: https://lkml.kernel.org/r/69fd4e00-1b13-d5f7-1c82-705c7d977ea4@huawei.com
Link: https://lkml.kernel.org/r/20250627125747.3094074-2-tujinjiang@huawei.com
Signed-off-by: Jinjiang Tu 
Fixes: 1b0449544c64 ("mm/vmscan: don't try to reclaim hwpoison folio")
Reported-by: syzbot+3b220254df55d8ca8a61@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/all/68412d57.050a0220.2461cf.000e.GAE@google.com/
Acked-by: David Hildenbrand 
Reviewed-by: Miaohe Lin 
Acked-by: Zi Yan 
Reviewed-by: Oscar Salvador 
Cc: Kefeng Wang 
Cc: Michal Hocko 
Cc: Oscar Salvador 
Cc: 
Signed-off-by: Andrew Morton

mm: memory-failure: enhance comments for return value of memory_failure()

2025-03-18T05:07:05+00:00

The comments for the return value of memory_failure are not complete,
supplement the comments.

Link: https://lkml.kernel.org/r/20250312112852.82415-4-xueshuai@linux.alibaba.com
Signed-off-by: Shuai Xue 
Reviewed-by: Jarkko Sakkinen 
Reviewed-by: Jonathan Cameron 
Reviewed-by: Yazen Ghannam 
Reviewed-by: Jane Chu 
Acked-by: Miaohe Lin 
Tested-by: Tony Luck 
Cc: Baolin Wang 
Cc: Borislav Betkov 
Cc: Catalin Marinas 
Cc: Dave Hansen 
Cc: "H. Peter Anvin" 
Cc: Ingo Molnar 
Cc: Josh Poimboeuf 
Cc: Naoya Horiguchi 
Cc: Peter Zijlstra 
Cc: Ruidong Tian 
Cc: Thomas Gleinxer 
Signed-off-by: Andrew Morton

mm/hwpoison: do not send SIGBUS to processes with recovered clean pages

2025-03-18T05:07:05+00:00

When an uncorrected memory error is consumed there is a race between the
CMCI from the memory controller reporting an uncorrected error with a UCNA
signature, and the core reporting and SRAR signature machine check when
the data is about to be consumed.

- Background: why *UN*corrected errors tied to *C*MCI in Intel platform [1]

Prior to Icelake memory controllers reported patrol scrub events that
detected a previously unseen uncorrected error in memory by signaling a
broadcast machine check with an SRAO (Software Recoverable Action
Optional) signature in the machine check bank.  This was overkill because
it's not an urgent problem that no core is on the verge of consuming that
bad data.  It's also found that multi SRAO UCE may cause nested MCE
interrupts and finally become an IERR.

Hence, Intel downgrades the machine check bank signature of patrol scrub
from SRAO to UCNA (Uncorrected, No Action required), and signal changed to
#CMCI.  Just to add to the confusion, Linux does take an action (in
uc_decode_notifier()) to try to offline the page despite the UC*NA*
signature name.

- Background: why #CMCI and #MCE race when poison is consuming in Intel platform [1]

Having decided that CMCI/UCNA is the best action for patrol scrub errors,
the memory controller uses it for reads too.  But the memory controller is
executing asynchronously from the core, and can't tell the difference
between a "real" read and a speculative read.  So it will do CMCI/UCNA if
an error is found in any read.

Thus:

1) Core is clever and thinks address A is needed soon, issues a speculative read.
2) Core finds it is going to use address A soon after sending the read request
3) The CMCI from the memory controller is in a race with MCE from the core
   that will soon try to retire the load from address A.

Quite often (because speculation has got better) the CMCI from the memory
controller is delivered before the core is committed to the instruction
reading address A, so the interrupt is taken, and Linux offlines the page
(marking it as poison).

- Why user process is killed for instr case

Commit 046545a661af ("mm/hwpoison: fix error page recovered but reported
"not recovered"") tries to fix noise message "Memory error not recovered"
and skips duplicate SIGBUSs due to the race.  But it also introduced a bug
that kill_accessing_process() return -EHWPOISON for instr case, as result,
kill_me_maybe() send a SIGBUS to user process.

If the CMCI wins that race, the page is marked poisoned when
uc_decode_notifier() calls memory_failure().  For dirty pages,
memory_failure() invokes try_to_unmap() with the TTU_HWPOISON flag,
converting the PTE to a hwpoison entry.  As a result,
kill_accessing_process():

- call walk_page_range() and return 1 regardless of whether
  try_to_unmap() succeeds or fails,
- call kill_proc() to make sure a SIGBUS is sent
- return -EHWPOISON to indicate that SIGBUS is already sent to the
  process and kill_me_maybe() doesn't have to send it again.

However, for clean pages, the TTU_HWPOISON flag is cleared, leaving the
PTE unchanged and not converted to a hwpoison entry.  Conversely, for
clean pages where PTE entries are not marked as hwpoison,
kill_accessing_process() returns -EFAULT, causing kill_me_maybe() to send
a SIGBUS.

Console log looks like this:

    Memory failure: 0x827ca68: corrupted page was clean: dropped without side effects
    Memory failure: 0x827ca68: recovery action for clean LRU page: Recovered
    Memory failure: 0x827ca68: already hardware poisoned
    mce: Memory error not recovered

To fix it, return 0 for "corrupted page was clean", preventing an
unnecessary SIGBUS to user process.

[1] https://lore.kernel.org/lkml/20250217063335.22257-1-xueshuai@linux.alibaba.com/T/#mba94f1305b3009dd340ce4114d3221fe810d1871
Link: https://lkml.kernel.org/r/20250312112852.82415-3-xueshuai@linux.alibaba.com
Fixes: 046545a661af ("mm/hwpoison: fix error page recovered but reported "not recovered"")
Signed-off-by: Shuai Xue 
Tested-by: Tony Luck 
Acked-by: Miaohe Lin 
Cc: Baolin Wang 
Cc: Borislav Betkov 
Cc: Catalin Marinas 
Cc: Dave Hansen 
Cc: "H. Peter Anvin" 
Cc: Ingo Molnar 
Cc: Jane Chu 
Cc: Jarkko Sakkinen 
Cc: Jonathan Cameron 
Cc: Josh Poimboeuf 
Cc: Naoya Horiguchi 
Cc: Peter Zijlstra 
Cc: Ruidong Tian 
Cc: Thomas Gleinxer 
Cc: Yazen Ghannam 
Cc: 
Signed-off-by: Andrew Morton

fs/dax: properly refcount fs dax pages

2025-03-18T05:06:41+00:00

Currently fs dax pages are considered free when the refcount drops to one
and their refcounts are not increased when mapped via PTEs or decreased
when unmapped.  This requires special logic in mm paths to detect that
these pages should not be properly refcounted, and to detect when the
refcount drops to one instead of zero.

On the other hand get_user_pages(), etc.  will properly refcount fs dax
pages by taking a reference and dropping it when the page is unpinned.

Tracking this special behaviour requires extra PTE bits (eg.  pte_devmap)
and introduces rules that are potentially confusing and specific to FS DAX
pages.  To fix this, and to possibly allow removal of the special PTE bits
in future, convert the fs dax page refcounts to be zero based and instead
take a reference on the page each time it is mapped as is currently the
case for normal pages.

This may also allow a future clean-up to remove the pgmap refcounting that
is currently done in mm/gup.c.

Link: https://lkml.kernel.org/r/c7d886ad7468a20452ef6e0ddab6cfe220874e7c.1740713401.git-series.apopple@nvidia.com
Signed-off-by: Alistair Popple 
Reviewed-by: Dan Williams 
Tested-by: Alison Schofield 
Acked-by: David Hildenbrand 
Cc: Alexander Gordeev 
Cc: Asahi Lina 
Cc: Balbir Singh 
Cc: Bjorn Helgaas 
Cc: Catalin Marinas 
Cc: Christian Borntraeger 
Cc: Christoph Hellwig 
Cc: Chunyan Zhang 
Cc: "Darrick J. Wong" 
Cc: Dave Chinner 
Cc: Dave Hansen 
Cc: Dave Jiang 
Cc: Gerald Schaefer 
Cc: Heiko Carstens 
Cc: Huacai Chen 
Cc: Ira Weiny 
Cc: Jan Kara 
Cc: Jason Gunthorpe 
Cc: Jason Gunthorpe 
Cc: John Hubbard 
Cc: linmiaohe 
Cc: Logan Gunthorpe 
Cc: Matthew Wilcow (Oracle) 
Cc: Michael "Camp Drill Sergeant" Ellerman 
Cc: Nicholas Piggin 
Cc: Peter Xu 
Cc: Sven Schnelle 
Cc: Ted Ts'o 
Cc: Vasily Gorbik 
Cc: Vishal Verma 
Cc: Vivek Goyal 
Cc: WANG Xuerui 
Cc: Will Deacon 
Signed-off-by: Andrew Morton

mm: memory-failure: update ttu flag inside unmap_poisoned_folio

2025-03-06T05:36:13+00:00

Patch series "mm: memory_failure: unmap poisoned folio during migrate
properly", v3.

Fix two bugs during folio migration if the folio is poisoned.


This patch (of 3):

Commit 6da6b1d4a7df ("mm/hwpoison: convert TTU_IGNORE_HWPOISON to
TTU_HWPOISON") introduce TTU_HWPOISON to replace TTU_IGNORE_HWPOISON in
order to stop send SIGBUS signal when accessing an error page after a
memory error on a clean folio.  However during page migration, anon folio
must be set with TTU_HWPOISON during unmap_*().  For pagecache we need
some policy just like the one in hwpoison_user_mappings to set this flag. 
So move this policy from hwpoison_user_mappings to unmap_poisoned_folio to
handle this warning properly.

Warning will be produced during unamp poison folio with the following log:

  ------------[ cut here ]------------
  WARNING: CPU: 1 PID: 365 at mm/rmap.c:1847 try_to_unmap_one+0x8fc/0xd3c
  Modules linked in:
  CPU: 1 UID: 0 PID: 365 Comm: bash Tainted: G        W          6.13.0-rc1-00018-gacdb4bbda7ab #42
  Tainted: [W]=WARN
  Hardware name: QEMU QEMU Virtual Machine, BIOS 0.0.0 02/06/2015
  pstate: 20400005 (nzCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
  pc : try_to_unmap_one+0x8fc/0xd3c
  lr : try_to_unmap_one+0x3dc/0xd3c
  Call trace:
   try_to_unmap_one+0x8fc/0xd3c (P)
   try_to_unmap_one+0x3dc/0xd3c (L)
   rmap_walk_anon+0xdc/0x1f8
   rmap_walk+0x3c/0x58
   try_to_unmap+0x88/0x90
   unmap_poisoned_folio+0x30/0xa8
   do_migrate_range+0x4a0/0x568
   offline_pages+0x5a4/0x670
   memory_block_action+0x17c/0x374
   memory_subsys_offline+0x3c/0x78
   device_offline+0xa4/0xd0
   state_store+0x8c/0xf0
   dev_attr_store+0x18/0x2c
   sysfs_kf_write+0x44/0x54
   kernfs_fop_write_iter+0x118/0x1a8
   vfs_write+0x3a8/0x4bc
   ksys_write+0x6c/0xf8
   __arm64_sys_write+0x1c/0x28
   invoke_syscall+0x44/0x100
   el0_svc_common.constprop.0+0x40/0xe0
   do_el0_svc+0x1c/0x28
   el0_svc+0x30/0xd0
   el0t_64_sync_handler+0xc8/0xcc
   el0t_64_sync+0x198/0x19c
  ---[ end trace 0000000000000000 ]---

[mawupeng1@huawei.com: unmap_poisoned_folio(): remove shadowed local `mapping', per Miaohe]
  Link: https://lkml.kernel.org/r/20250219060653.3849083-1-mawupeng1@huawei.com
Link: https://lkml.kernel.org/r/20250217014329.3610326-1-mawupeng1@huawei.com
Link: https://lkml.kernel.org/r/20250217014329.3610326-2-mawupeng1@huawei.com
Fixes: 6da6b1d4a7df ("mm/hwpoison: convert TTU_IGNORE_HWPOISON to TTU_HWPOISON")
Signed-off-by: Ma Wupeng 
Suggested-by: David Hildenbrand 
Acked-by: David Hildenbrand 
Acked-by: Miaohe Lin 
Cc: Ma Wupeng 
Cc: Michal Hocko 
Cc: Naoya Horiguchi 
Cc: Oscar Salvador 
Cc: 
Signed-off-by: Andrew Morton

treewide: const qualify ctl_tables where applicable

2025-01-28T12:48:37+00:00

Add the const qualifier to all the ctl_tables in the tree except for
watchdog_hardlockup_sysctl, memory_allocation_profiling_sysctls,
loadpin_sysctl_table and the ones calling register_net_sysctl (./net,
drivers/inifiniband dirs). These are special cases as they use a
registration function with a non-const qualified ctl_table argument or
modify the arrays before passing them on to the registration function.

Constifying ctl_table structs will prevent the modification of
proc_handler function pointers as the arrays would reside in .rodata.
This is made possible after commit 78eb4ea25cd5 ("sysctl: treewide:
constify the ctl_table argument of proc_handlers") constified all the
proc_handlers.

Created this by running an spatch followed by a sed command:
Spatch:
    virtual patch

    @
    depends on !(file in "net")
    disable optional_qualifier
    @

    identifier table_name != {
      watchdog_hardlockup_sysctl,
      iwcm_ctl_table,
      ucma_ctl_table,
      memory_allocation_profiling_sysctls,
      loadpin_sysctl_table
    };
    @@

    + const
    struct ctl_table table_name [] = { ... };

sed:
    sed --in-place \
      -e "s/struct ctl_table .table = &uts_kern/const struct ctl_table *table = \&uts_kern/" \
      kernel/utsname_sysctl.c

Reviewed-by: Song Liu 
Acked-by: Steven Rostedt (Google)  # for kernel/trace/
Reviewed-by: Martin K. Petersen  # SCSI
Reviewed-by: Darrick J. Wong  # xfs
Acked-by: Jani Nikula 
Acked-by: Corey Minyard 
Acked-by: Wei Liu 
Acked-by: Thomas Gleixner 
Reviewed-by: Bill O'Donnell 
Acked-by: Baoquan He 
Acked-by: Ashutosh Dixit 
Acked-by: Anna Schumaker 
Signed-off-by: Joel Granados

mm/memory-failure: replace sprintf() with sysfs_emit()

2024-11-11T08:26:46+00:00

As Documentation/filesystems/sysfs.rst suggested, show() should only use
sysfs_emit() or sysfs_emit_at() when formatting the value to be returned
to user space.

Link: https://lkml.kernel.org/r/20241029101853.37890-1-zhangguopeng@kylinos.cn
Signed-off-by: zhangguopeng 
Acked-by: Miaohe Lin 
Cc: Naoya Horiguchi 
Signed-off-by: Andrew Morton

mm: mass constification of folio/page pointers

2024-11-07T22:38:07+00:00

Now that page_pgoff() takes const pointers, we can constify the pointers
to a lot of functions.

Link: https://lkml.kernel.org/r/20241005200121.3231142-5-willy@infradead.org
Signed-off-by: Matthew Wilcox (Oracle) 
Signed-off-by: Andrew Morton

mm: renovate page_address_in_vma()

2024-11-07T22:38:07+00:00

This function doesn't modify any of its arguments, so if we make a few
other functions take const pointers, we can make page_address_in_vma()
take const pointers too.  All of its callers have the containing folio
already, so pass that in as an argument instead of recalculating it.  Also
add kernel-doc

Link: https://lkml.kernel.org/r/20241005200121.3231142-4-willy@infradead.org
Signed-off-by: Matthew Wilcox (Oracle) 
Signed-off-by: Andrew Morton

mm: convert page_to_pgoff() to page_pgoff()

2024-11-07T22:38:07+00:00

Patch series "page->index removals in mm", v2.

As part of shrinking struct page, we need to stop using page->index.  This
patchset gets rid of most of the remaining references to page->index in
mm, as well as increasing the number of functions which take a const
folio/page pointer.  It shrinks the text segment of mm by a few hundred
bytes in my test config, probably mostly from removing calls to
compound_head() in page_to_pgoff().


This patch (of 7):

Change the function signature to pass in the folio as all three callers
have it.  This removes a reference to page->index, which we're trying to
get rid of.  And add kernel-doc.

Link: https://lkml.kernel.org/r/20241005200121.3231142-1-willy@infradead.org
Link: https://lkml.kernel.org/r/20241005200121.3231142-2-willy@infradead.org
Signed-off-by: Matthew Wilcox (Oracle) 
Signed-off-by: Andrew Morton