alpha requires percpu variables in modules to be defined as weak so that the
compiler generates GOT based external references for them. This puts two
extra restrictions on percpu variable definitions. The symbol must be
globally unique even when static and a static percpu variable can't be
defined inside a function. DEBUG_FORCE_WEAK_PER_CPU exists to give generic
code build coverage for these restrictions without building for alpha.
MEM_ALLOC_PROFILING defines a static percpu counter at each allocation call
site and thus can't be built with weak percpu definitions, so it depends on
!DEBUG_FORCE_WEAK_PER_CPU. As allmodconfig enables DEBUG_FORCE_WEAK_PER_CPU,
this knocks MEM_ALLOC_PROFILING out of allmodconfig build coverage.
allmodconfig coverage for MEM_ALLOC_PROFILING is worth more than build
coverage for restrictions which only matter to alpha module builds. Drop
DEBUG_FORCE_WEAK_PER_CPU. Restriction violations will now show up only on
alpha builds.
Link: https://lore.kernel.org/178656406317.2437052.7257990869957704195@slm.duckdns.org
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Suren Baghdasaryan <surenb@google.com>
Acked-by: Gabriele Monaco <gmonaco@redhat.com> [include/rv/da_monitor.h]
Cc: Dennis Zhou <dennis@kernel.org>
Cc: Kent Overstreet <kent.overstreet@linux.dev>
Cc: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
sort_folio() has a shortcut for moving folios that are no longer evictable
but are still sitting on a generation list. However, this shortcut is
buggy. It does not follow the PG_lru usage convention, and it has a more
serious issue.
Unevictable folios are not threaded on lists[LRU_UNEVICTABLE], so that
folio->lru can be reused to hold folio->mlock_count (see the comment in
lruvec_init()). Hence lruvec_add_folio() skips the list_add() for them,
and every other place that turns a folio unevictable initialises
mlock_count explicitly: lru_add() sets it to 0, __mlock_folio() and
__mlock_new_folio() set it to !!folio_test_mlocked(folio). sort_folio()
sets nothing, and the lru_gen_del_folio() right above it may have already
poisoned folio->lru via list_del(), so mlock_count ends up aliasing
LIST_POISON2, which reads as 0x122, i.e. 290. The result is user
visible. On munlock, __munlock_folio() decrements that bogus count, finds
it still non-zero and bails out before clearing PG_mlocked, so the folio
remains unevictable and the Mlocked accounting stays inflated until the
folio is freed.
The shortcut also touches the LRU flags in the wrong order. It calls
lru_gen_del_folio() while PG_lru is still set, so a concurrent
folio_test_clear_lru() (e.g. compaction, folio_isolate_lru()) can succeed
on a folio that has already been taken off the generation list, which may
lead to unexpected behavior.
So fix it by isolating them as common folios and letting the generic
shrink path cull them. This matches the classical LRU behavior, and there
should be no visible effect on the generic eviction or isolation behavior.
There is no performance concern either, such a folio goes through this
once, and then it is off the generation lists for good.
Link: https://lore.kernel.org/20260812-mglru-mlock-fix-v2-1-a3fec5853c08@tencent.com
Fixes: ac35a49023 ("mm: multi-gen LRU: minimal implementation")
Signed-off-by: Kairui Song <kasong@tencent.com>
Reviewed-by: Barry Song <baohua@kernel.org>
Reviewed-by: Baolin Wang <baolin.wang@linux.alibaba.com>
Cc: Axel Rasmussen <axelrasmussen@google.com>
Cc: Brian Geffon <bgeffon@google.com>
Cc: David Hildenbrand <david@kernel.org>
Cc: Jan Alexander Steffens (heftig) <heftig@archlinux.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Lorenzo Stoakes <ljs@kernel.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Oleksandr Natalenko <oleksandr@natalenko.name>
Cc: Shakeel Butt <shakeel.butt@linux.dev>
Cc: Steven Barrett <steven@liquorix.net>
Cc: Suleiman Souhlal <suleiman@google.com>
Cc: Wei Xu <weixugc@google.com>
Cc: Yuanchu Xie <yuanchu@google.com>
Cc: Yu Zhao <yuzhao@google.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "mm/khugepaged: several cleanups", v4.
The following changes stem from a number of reviews during my khugepaged
mTHP support series [1]. Some of these are minor code cleanups, issues or
reviews that we decided to deferred to a followup series, or in the case
of the more major patch of the series, changes [2] Lance Yang attempted
while my series was in-flight and we decided to wait till later to try.
The first 3 patches introduce helper functions to increase code reuse and
readability. This includes a per-scan state clearing function, extracting
the young page check into a helper, and a count_collapse_event() function
to reduce a repetative pattern used across mTHP collapse.
The 4th patch was the byproduct of me throwing Claude at all the comments
in khugepaged verifying and looking for any outdated info.
The 5th patch is based on Lance Yang's commit series [2] trying to extract
the PTE state checking into a helper function. This required a bit of
rewriting due to differences after mTHP collapse was introduced. I also
took into account the changes requested during his patches review cycle.
The remaining 2 patches were review points during my mTHP series that we
agreed can be deferred to a later series.
Thank you to those whos reviews and work I leveraged to achieve these
cleanups.
This patch (of 6):
Extract the repeated clearing of node_load, alloc_nmask, and
mthp_present_ptes into a helper to reduce duplication in
collapse_scan_pmd() and collapse_scan_file(). Althought file scans do not
current use the bitmap, they will in the future, and clearing it now is
harmless.
Link: https://lore.kernel.org/20260811-khugepaged_pte_refactor-v4-0-ddac39d61c4a@linux.dev
Link: https://lore.kernel.org/20260811-khugepaged_pte_refactor-v4-1-ddac39d61c4a@linux.dev
Link: https://lore.kernel.org/all/20260605161422.213817-1-npache@redhat.com/ [1]
Link: https://lore.kernel.org/all/20251008043748.45554-1-lance.yang@linux.dev/ [2]
Signed-off-by: Nico Pache (Red Hat) <nico.pache@linux.dev>
Reviewed-by: Baolin Wang <baolin.wang@linux.alibaba.com>
Acked-by: Usama Arif <usama.arif@linux.dev>
Acked-by: David Hildenbrand (Arm) <david@kernel.org>
Reviewed-by: Lorenzo Stoakes (ARM) <ljs@kernel.org>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Reviewed-by: Pedro Falcato <pfalcato@suse.de>
Reviewed-by: Lance Yang <lance.yang@linux.dev>
Cc: Barry Song <baohua@kernel.org>
Cc: Dev Jain <dev.jain@arm.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Liam R. Howlett <liam@infradead.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Ryan Roberts <ryan.roberts@arm.com>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Vlastimil Babka <vbabka@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
lru_gen_look_around() feeds its local 'young' counter into
suitable_to_scan(), which decides whether the current PMD is added to
the bloom filter and checked again on the next aging round.
The folio triggering the look-around is processed at function entry:
test_and_clear_young_ptes_notify() clears the accessed bits of the nr PTEs
it maps, and the function bails out if none of them is young. The loop
that follows therefore never recounts this folio, since its accessed bits
are already cleared. Every other young folio the loop finds is accounted
as a batch (young += nr), where nr is the number of consecutive PTEs it
maps. The triggering folio, however, still contributes a fixed young = 1
regardless of its size -- a leftover from before PTE batching. A large
triggering folio is thus accounted inconsistently with the rest of the
window.
Initialize young to nr so the triggering folio is accounted the same way
as any other young folio batch in the loop.
Note this is a deliberate overestimate, not a measured value. The
test-and-clear helper only reports whether any of the nr PTEs is young,
not how many were accessed, so the true number of accessed PTEs in a large
folio is unknown and can be smaller than nr. Counting the full batch is
intentional: the mm core tracks accessed/dirty state per folio, not per
page, so a per-page count is neither obtainable nor meaningful. The only
consumer is suitable_to_scan(), and the bloom filter it feeds tolerates
error. Overestimating is also the safe direction: at worst a PMD that saw
little access is rescanned, whereas underestimating could skip rescanning
a PMD whose folios are still hot and reclaim them incorrectly. (nr here
is the PTE batch size, not necessarily folio_nr_pages().)
Link: https://lore.kernel.org/20260813061019.49806-1-hui.zhu@linux.dev
Link: https://lore.kernel.org/20260812065933.103627-1-hui.zhu@linux.dev
Fixes: 56e5b60b21 ("mm: support batched checking of the young flag for MGLRU")
Signed-off-by: Hui Zhu <zhuhui@kylinos.cn>
Reviewed-by: Baolin Wang <baolin.wang@linux.alibaba.com>
Reviewed-by: Barry Song <baohua@kernel.org>
Cc: Axel Rasmussen <axelrasmussen@google.com>
Cc: David Hildenbrand <david@kernel.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Kairui Song <kasong@tencent.com>
Cc: Lorenzo Stoakes <ljs@kernel.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Shakeel Butt <shakeel.butt@linux.dev>
Cc: Wei Xu <weixugc@google.com>
Cc: Yuanchu Xie <yuanchu@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
quarantine_size tracks the total number of bytes stored in
global_quarantine[]. It is incremented when per-CPU quarantine objects
are moved into the global quarantine and decremented when a global batch
is evicted by kasan_quarantine_reduce().
kasan_quarantine_remove_cache() also removes objects from the global
quarantine. qlist_move_cache() rebuilds the source batch and updates its
.bytes field, but quarantine_size is not adjusted accordingly.
As a result, quarantine_size remains over-counted by the size of the
removed objects. The stale accounting accumulates across cache removals.
Once the inflated value exceeds quarantine_max_size,
kasan_quarantine_reduce() can evict a batch even though the actual number
of bytes in global_quarantine[] is still below quarantine_max_size,
shortening the quarantine window.
Fix the accounting by recording each batch's size before
qlist_move_cache() and subtracting the number of bytes actually removed
from quarantine_size while holding quarantine_lock.
A KUnit reproducer used during testing observed the over-count grow by
4698864 bytes after one kasan_quarantine_remove_cache() call with the fix
reverted. With this change applied, the over-count did not grow.
Link: https://lore.kernel.org/20260811073332.1351893-1-sh_def@163.com
Fixes: 64abdcb243 ("kasan: eliminate long stalls during quarantine reduction")
Signed-off-by: Hui Su <sh_def@163.com>
Reported-by: Sashiko <sashiko-bot@kernel.org>
Closes: https://sashiko.dev/#/patchset/20260808031459.3032812-1-sh_def%40163.com
Reviewed-by: Andrey Ryabinin <ryabinin.a.a@gmail.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Andrey Konovalov <andreyknvl@gmail.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Vincenzo Frascino <vincenzo.frascino@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
zs_lookup_class_index() lets zram recompression decide whether a newly
compressed object would use a smaller size class. It currently classifies
the payload size directly, while zs_malloc() adds ZS_HANDLE_SIZE before
selecting the class.
This makes lookup disagree with allocation near size-class boundaries.
With 4 KiB pages, CONFIG_ZSMALLOC_CHAIN_SIZE=8, and 64-bit handles, a
1025-to-1024-byte recompression appears to move from class 64 to class 62
although both allocations use class 64. Conversely, a 1049-to-1025-byte
recompression appears to stay in class 64 although the allocations move
from class 65 to class 64.
As a result, zram can accept replacements with no allocation benefit or
reject ones that would save memory, potentially marking the object
incompressible.
Factor size-class selection into lookup_size_class(), account for the
handle there, and use the helper for both lookup and allocation.
Link: https://lore.kernel.org/20260809115518.3791787-1-xialonglong2025@163.com
Fixes: 7c2af309ab ("zram: add size class equals check into recompression")
Assisted-by: Codex:gpt-5.6-sol
Signed-off-by: Longlong Xia <xialonglong@kylinos.cn>
Reviewed-by: Sergey Senozhatsky <senozhatsky@chromium.org>
Cc: Minchan Kim <minchan@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
The slot lock is a bit operation on the whole __lock word, which flags and
ac_time alias as two u32s. On little-endian the lock bit lands in the
position ZRAM_ENTRY_LOCK reserves in flags, so the aliasing works out. On
64-bit big-endian it lands in ac_time instead: with
ZRAM_TRACK_ENTRY_ACTIME enabled, storing the access time from
mark_slot_accessed() or slot_free() wipes out the held lock bit, letting
another CPU take the same slot lock; an access time value with that bit
set makes the slot look locked forever.
Shift the lock bit into the flags half of the word on big-endian 64-bit.
Link: https://lore.kernel.org/20260810202241.2436603-1-devnexen@gmail.com
Fixes: 2e8ff2f51d ("zram: use u32 for entry ac_time tracking")
Signed-off-by: David Carlier <devnexen@gmail.com>
Reviewed-by: Sergey Senozhatsky <senozhatsky@chromium.org>
Cc: Minchan Kim <minchan@kernel.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "mm: kmemleak: default min_unref_scans to 2 for verbose
kernels", v2.
When CONFIG_DEBUG_KMEMLEAK_VERBOSE is set, which means the host is in
auto scan mode, set min_unref_scans to 2, avoiding false positives.
CONFIG_DEBUG_KMEMLEAK_VERBOSE depends on CONFIG_DEBUG_KMEMLEAK_AUTO_SCAN,
so a kernel built with it already runs the scan thread periodically and
the user has asked for detailed leak reports.
The confirming second scan comes for free there, so default
min_unref_scans to 2 in that case and keep it at 1 everywhere else.
CONFIG_DEBUG_KMEMLEAK_VERBOSE defaults to n, so nothing changes for
kernels that do not opt in.
The other two patches bring the documentation and the selftest comments
in line with the new conditional default.
PS: A similar patch (v1 of this patchset) is applied to Meta's kernel,
in real production hosts.
This patch (of 3):
min_unref_scans defers reporting an object as leaked until it has stayed
unreferenced for that many consecutive scans, filtering out objects that
are only transiently unreferenced during a scan.
It defaults to 1, which reports on the first unreferenced scan.
CONFIG_DEBUG_KMEMLEAK_VERBOSE depends on CONFIG_DEBUG_KMEMLEAK_AUTO_SCAN,
so a kernel built with it runs the scan thread periodically and the user
has opted into detailed leak reporting. A second confirming scan then
happens on its own.
Default min_unref_scans to 2 there to suppress transient false positives,
and keep it at 1 otherwise, where a manually triggered scan is expected to
report immediately. The value stays writable through the module
parameter.
CONFIG_DEBUG_KMEMLEAK_VERBOSE defaults to n, so this does not change the
default for kernels that do not opt in.
Link: https://lore.kernel.org/20260731-kmemleak_hardened-v2-0-7b9689ac77cb@debian.org
Link: https://lore.kernel.org/20260731-kmemleak_hardened-v2-1-7b9689ac77cb@debian.org
Signed-off-by: Breno Leitao <leitao@debian.org>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Cc: David Hildenbrand <david@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Liam R. Howlett <liam@infradead.org>
Cc: Lorenzo Stoakes <ljs@kernel.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Vlastimil Babka <vbabka@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Add a functional test for the min_unref_scans kmemleak module parameter.
Using samples/kmemleak's helper module it checks that min_unref_scans=1
reports an orphan on the first scan, min_unref_scans=2 reports nothing on
the first scan but does on the second, and that the parameter reads back
what was written.
It counts only the helper module's own orphans (matched by their
[kmemleak_test] backtrace, with the module kept loaded so the symbols
resolve) so unrelated leaks already present on the system do not perturb
the result. The test skips when run as non-root, without
CONFIG_DEBUG_KMEMLEAK / CONFIG_SAMPLE_KMEMLEAK, on a kernel without the
parameter, or when the helper yields no detectable orphan.
Link: https://lore.kernel.org/20260713-catalin_pto-v1-4-5b93b1131089@debian.org
Signed-off-by: Breno Leitao <leitao@debian.org>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Cc: David Hildenbrand <david@kernel.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Liam R. Howlett <liam@infradead.org>
Cc: Lorenzo Stoakes <ljs@kernel.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Vlastimil Babka <vbabka@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
kmemleak reports an object the first scan it is found unreferenced. Its
mark phase runs without stopping the rest of the kernel and without a
write barrier, so a live object whose only reference is briefly invisible
during a concurrent RCU update -- e.g. a VMA moved between maple tree
nodes, or a page-cache xa_node -- can be seen as unreferenced for that one
scan. Because an object is flagged as reported only once, such a
transient race turns into a permanent false positive.
Track how many consecutive scans each object has been seen unreferenced
and only report it once that reaches min_unref_scans, a new module
parameter. It defaults to 1, leaving the behaviour unchanged; setting it
higher (e.g. 2) still reports a genuine leak, one scan later, while an
object referenced again before the threshold restarts its run and is never
reported.
min_unref_scans can be set at boot with kmemleak.min_unref_scans=<n> or at
run-time via /sys/module/kmemleak/parameters/min_unref_scans.
Link: https://lore.kernel.org/20260713-catalin_pto-v1-2-5b93b1131089@debian.org
Signed-off-by: Breno Leitao <leitao@debian.org>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Cc: David Hildenbrand <david@kernel.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Liam R. Howlett <liam@infradead.org>
Cc: Lorenzo Stoakes <ljs@kernel.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Vlastimil Babka <vbabka@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "mm: kmemleak: reduce transient false positives by confirming
leaks".
This series combines two kmemleak enhancements that were originally
submitted separately but both required rebasing after commit 79c37ae373
("mm/kmemleak: fix checksum computation for per-cpu objects").
The first feature introduces a second scan to confirm suspected leaks:
https://lore.kernel.org/all/20260709173347.689607-1-catalin.marinas@arm.com/
The second feature adds a module parameter controlling the minimum number
of consecutive unreferenced scans before a leak is reported, as discussed
in:
https://lore.kernel.org/all/20260626-kmemleak_twice-v1-0-ab28f7cc0971@debian.org/
Changes from v1:
Now that commit 79c37ae373 is upstream, the selftest includes an additional
priming phase scan as requested by Catalin.
Additionally, I've factored out the leak-detection conditional into a helper
function to be more digestible for the reader's eye.
This 4-patch series resolves all outstanding kmemleak issues I've been
tracking.
This patch (of 4):
The kmemleak marking phase is not atomic. While the object graph is
traversed, the kernel can modify pointers, free objects or allocate new
ones. If a reference to an object is moved from one location to another,
kmemleak scanning may miss it. We have explicit annotations like
kmemleak_transient_leak() but identifying and maintaining them is not
trivial.
Given that such transient leaks are short-lived, rather than just
reporting such objects as leaks, do another scan to confirm the suspected
objects. If no new leaks are found during the first scan, skip the
confirmation one.
Link: https://lore.kernel.org/20260713-catalin_pto-v1-0-5b93b1131089@debian.org
Link: https://lore.kernel.org/20260713-catalin_pto-v1-1-5b93b1131089@debian.org
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Assisted-by: Claude:claude-opus-4-8
Signed-off-by: Breno Leitao <leitao@debian.org>
Cc: David Hildenbrand <david@kernel.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Liam R. Howlett <liam@infradead.org>
Cc: Lorenzo Stoakes <ljs@kernel.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Vlastimil Babka <vbabka@kernel.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
kasan_quarantine_remove_cache() first invokes per_cpu_remove_cache() on
all online CPUs. Each callback moves objects belonging to the cache from
cpu_quarantine to the CPU's shrink_qlist, where they can later be freed
from task context.
kmem_cache_destroy() invokes the quarantine removal path while holding
cpus_read_lock(), but kmem_cache_shrink() does not. The latter can
therefore race with CPU offlining as follows:
kmem_cache_shrink() CPU hotplug
------------------- -----------
on_each_cpu()
CPU1 moves objects to
CPU1's shrink_qlist
on_each_cpu() returns
CPU1 goes offline
kasan_cpu_offline()
drains cpu_quarantine
leaves shrink_qlist untouched
for_each_online_cpu()
skips CPU1
The objects left on CPU1's shrink_qlist are not returned to the slab
allocator. This may prevent kmem_cache_shrink() from releasing slabs that
would otherwise become empty. If CPU1 remains offline, a later
kmem_cache_destroy() also skips the list and can report that the cache
still contains objects.
An intermittent occurrence was observed with a virtio-9p filesystem. The
mount and umount commands both returned 0, but the kernel logged the
following during the userspace-triggered teardown:
[ 2994.380134][ T111] BUG 9p-fcall-cache-1 (Tainted: G B ): Objects remaining on __kmem_cache_shutdown()
[ 2994.381140][ T111] Object 0xff11000004361118 @offset=4376
[ 2994.381607][ T111] Allocated in p9_fcall_init+0x201/0x400 age=19564 cpu=1 pid=104
[ 2994.382591][ T111] p9_fcall_init+0x201/0x400
[ 2994.382810][ T111] p9_tag_alloc+0x12f/0x700
[ 2994.382982][ T111] p9_client_prepare_req+0x102/0x3e0
[ 2994.383165][ T111] p9_client_rpc+0x1ab/0xa50
[ 2994.383334][ T111] p9_client_getattr_dotl+0xb0/0x1a0
[ 2994.383515][ T111] v9fs_vfs_getattr_dotl+0x115/0x360
[ 2994.383719][ T111] vfs_getattr_nosec+0x22c/0x3a0
[ 2994.383910][ T111] vfs_statx+0xd7/0x170
[ 2994.384062][ T111] vfs_fstatat+0x45/0x80
[ 2994.384215][ T111] __do_sys_newfstatat+0x84/0xe0
[ 2994.384386][ T111] do_syscall_64+0x115/0x6a0
[ 2994.384566][ T111] entry_SYSCALL_64_after_hwframe+0x77/0x7f
[ 2994.399720][ T111] WARNING: mm/slub.c:1244 at __kmem_cache_shutdown+0x363/0x500, CPU#0: busybox/111
[ 2994.405655][ T111] Call Trace:
[ 2994.406325][ T111] kmem_cache_destroy+0x73/0x1b0
[ 2994.406630][ T111] p9_client_destroy+0x271/0x3c0
[ 2994.407210][ T111] v9fs_session_close+0x3c/0x260
[ 2994.407409][ T111] v9fs_kill_super+0x48/0x90
[ 2994.407584][ T111] deactivate_locked_super+0xa3/0x160
[ 2994.407778][ T111] cleanup_mnt+0x1dd/0x3e0
Thus, a successful umount left objects in the 9p fcall cache and prevented
the cache from being destroyed cleanly.
Per-CPU shrink_qlist storage exists for every possible CPU, and each list
is protected by its own raw spinlock. Iterate over possible CPUs so that
a list populated before its CPU went offline is drained as well.
for_each_possible_cpu() can do more work than for_each_online_cpu(), but
this change only affects CONFIG_KASAN_GENERIC kernels. The extra work is
limited to cache shrink and cache destruction paths and does not affect
the normal allocation/free fast path. It adds one raw-spinlock-protected
scan of each possible CPU's shrink list. These lists are normally empty;
a non-empty list is traversed to remove objects belonging to the cache
being shrunk or destroyed.
Link: https://lore.kernel.org/20260808031459.3032812-1-sh_def@163.com
Fixes: 07d067e4f2 ("kasan: fix sleeping function called from invalid context on RT kernel")
Signed-off-by: Hui Su <sh_def@163.com>
Reviewed-by: Andrey Ryabinin <ryabinin.a.a@gmail.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Andrey Konovalov <andreyknvl@gmail.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Vincenzo Frascino <vincenzo.frascino@arm.com>
Cc: "Zhang, Qiang1" <qiang1.zhang@intel.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
zswap_shrinker_count() calls mem_cgroup_flush_stats(), which takes the
global cgroup rstat lock synchronously. On machines with many CPUs and
NUMA nodes, this creates severe lock contention in the kswapd reclaim
path:
- Multiple kswapd threads (one per NUMA node) run concurrently.
- do_shrink_slab() invokes zswap_shrinker_count() for each
memcg-aware shrinker pass.
- Each call flushes the full cgroup rstat hierarchy under the global
lock.
On AMD EPYC 9684X machines (96 cores, 192 threads, 12 NUMA nodes) running
production workloads with zswap enabled, perf shows 2.88% of kernel cycles
in osq_lock contention from this path:
2.88% [k] osq_lock
--__mutex_lock.constprop.0
--__cgroup_rstat_lock
--cgroup_rstat_flush_locked
--cgroup_rstat_flush
--zswap_shrinker_count
do_shrink_slab
shrink_slab
shrink_node
balance_pgdat
kswapd
84% of kswapd kernel cycles are spent in
shrink_slab -> zswap_shrinker_count -> cgroup_rstat_flush, not in actual
page reclaim (shrink_lruvec).
Controlled A/B on identical hardware and workload:
shrinker=Y: 2.88% osq_lock, memory PSI 1.58%
shrinker=N: 0.00% osq_lock, memory PSI 0.57%
eBPF-based rstat lock wait measurement across 8 production metals
confirms the contention splits cleanly along shrinker enablement:
shrinker=Y: 50-250x more contended lock acquisitions (248/s vs 1.1/s)
shrinker=N: baseline lock wait (0.0017 s/s vs 1.04 s/s)
zswap_shrinker_count() only produces a heuristic estimate, scaled by
compression ratio via mult_frac(). The actual writeback happens in
zswap_shrinker_scan(). Slightly stale stats are acceptable here.
Switch to mem_cgroup_flush_stats_ratelimited(), which only flushes if
the periodic 2-second flusher is one full cycle late. This matches the
approach already used in prepare_scan_control() (mm/vmscan.c) for the
same reclaim path.
After applying this patch, rstat flush latency and lock wait time on
shrinker=Y machines dropped to the same level as shrinker=N controls,
while the zswap shrinker continues to function (pool size remains
bounded under the max_pool_percent cap).
Previously discussed:
- Chengming Zhou (Dec 2023): rstat contention from
zswap_shrinker_count [1]
- Shakeel Butt (Aug 2024): zswap_shrinker_count still uses sync
flush [2]
- Yosry Ahmed (Aug 2024): suggested eliminating in-kernel
flushers [3]
- Jesper Dangaard Brouer (Sep 2024): cgroup/rstat V11 patch [4]
Link: https://lore.kernel.org/20260702180908.150136-1-yunzhao@cloudflare.com
Link: https://lore.kernel.org/linux-mm/20231206103935.3440502-1-zhouchengming@bytedance.com/ [1]
Link: https://lore.kernel.org/linux-mm/CALvZod7LFxLCxVpOFH8b2Ppm8T40HPGMKQwX_=NPCWB_mFW+oQ@mail.gmail.com/ [2]
Link: https://lore.kernel.org/linux-mm/CAJD7tkYvFyOSX+rP_FKGBhxvZiCDxtpsNp-c5CGOA-4Bq9oXSg@mail.gmail.com/ [3]
Link: https://lore.kernel.org/linux-mm/172616070094.2055617.17676042522679701515.stgit@firesoul/ [4]
Suggested-by: Jesper Dangaard Brouer <hawk@kernel.org>
Signed-off-by: Jesper Dangaard Brouer <hawk@kernel.org>
Signed-off-by: Yunzhao Li <yunzhao@cloudflare.com>
Tested-by: Yunzhao Li <yunzhao@cloudflare.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Jesper Dangaard Brouer <hawk@kernel.org>
Acked-by: Nhat Pham <nphamcs@gmail.com>
Cc: Chengming Zhou <chengming.zhou@linux.dev>
Cc: Shakeel Butt <shakeel.butt@linux.dev>
Cc: Yosry Ahmed <yosry@kernel.org>
Cc: Yunzhao Li <yunzhao@cloudflare.com>
Cc: Sourav Panda <souravpanda@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Block I/O benefits from batching just as much as remote file systems.
Extend struct swap_iocb to support building a bio on the fly as well, and
rewrite the block based swap code for it. This especially benefits
submit_bio based drivers that do not have the block plugging available,
but also saves allocating extra bios for blk-mq drivers.
Add a pre-allocated bio to struct swap_iocb in a union with kiocb used for
file system based swap so that struct swap_iocb can be used for all swap
I/O, and initialize the pool for it unconditionally.
Various low-level bdev and fs functions are now replaced with a unified
can_merge/add/submit scheme.
Note that the block based swap code now uses the same memcg-based check
previously added for file system based swap as well.
Link: https://lore.kernel.org/20260713093350.2154226-4-hch@lst.de
Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Baolin Wang <baolin.wang@linux.alibaba.com>
Cc: Baoquan He <baoquan.he@linux.dev>
Cc: Barry Song <baohua@kernel.org>
Cc: Chris Li <chrisl@kernel.org>
Cc: Kairui Song <kasong@tencent.com>
Cc: Kemeng Shi <shikemeng@huaweicloud.com>
Cc: Nhat Pham <nphamcs@gmail.com>
Cc: Youngjun Park <youngjun.park@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Generalize the context currently provided by double pointers to struct
swap_iocb to an on-stack context. This cleans up the code and prepares
for adding more fields and supporting batching multiple folios into a
single bio for block-based swap as well.
This new swap_io_ctx is required for all functions using it, the old way
of allowing a NULL iocb for some callers is removed to keep the interface
consistent. To reduce code duplication caused by this, a new
swap_cache_read_folio_sync helper is added to consolidate the code to call
swap_cache_read_folio with a local swap_io_ctx.
The unpug helpers are renamed to use the submit wording as they are
generalized.
Link: https://lore.kernel.org/20260713093350.2154226-3-hch@lst.de
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Nhat Pham <nphamcs@gmail.com>
Cc: Baolin Wang <baolin.wang@linux.alibaba.com>
Cc: Baoquan He <baoquan.he@linux.dev>
Cc: Barry Song <baohua@kernel.org>
Cc: Chris Li <chrisl@kernel.org>
Cc: Kairui Song <kasong@tencent.com>
Cc: Kemeng Shi <shikemeng@huaweicloud.com>
Cc: Youngjun Park <youngjun.park@lge.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
shmem_fallocate() validates offset + len with inode_newsize_ok(), but then
rounds that end offset up to a page boundary before entering the
preallocation loop.
For a valid request ending at MAX_LFS_FILESIZE, such as offset = 0 and len
= LLONG_MAX, adding PAGE_SIZE - 1 to the validated end can overflow the
signed loff_t used for the rounded end calculation. If that wrapped value
is then converted into a page index, shmem_fallocate() can enter the folio
allocation loop with an invalid range.
Use check_add_overflow() when calculating the page-aligned end, and fail
before entering the allocation loop if the rounded end cannot be
represented.
Link: https://lore.kernel.org/1929a466735dcbb9438936ff50b7a4fc2332a8a4.1785377919.git.zhilinz@nebusec.ai
Fixes: e2d12e22c5 ("tmpfs: support fallocate preallocation")
Signed-off-by: Zhiling Zou <zhilinz@nebusec.ai>
Reported-by: Vega <vega@nebusec.ai>
Reviewed-by: Baolin Wang <baolin.wang@linux.alibaba.com>
Cc: Hugh Dickins <hughd@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Add the following 2 scenarios to the allocinfo ioctl kselftest:
1. Validate size based filtering
2. Validate lineno based filtering
The first test uses "do_init_module" as the candidate function for the
test. This is because the associated site will only allocate memory when
a kernel module is loaded. The return value of get_content_id() changes
every time modules are loaded or unloaded. Hence, as long as
get_content_id() values at the start and the end of the test are the same,
the memory allocated by the do_init_module call site should also remain
the same. Consequently, the test can assume consistency between the value
returned by the ioctl and the procfs resulting in less flakiness.
Link: https://lore.kernel.org/e5171926b48802531284c1cb5f04734017141341.1783532853.git.abhishekbapat@google.com
Signed-off-by: Abhishek Bapat <abhishekbapat@google.com>
Tested-by: Hao Ge <hao.ge@linux.dev>
Acked-by: Hao Ge <hao.ge@linux.dev>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Kent Overstreet <kent.overstreet@linux.dev>
Cc: Sourav Panda <souravpanda@google.com>
Cc: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "alloc_tag: introduce IOCTL-based filtering for MAP", v8.
Currently, memory allocation profiling data is primarily exposed through
/proc/allocinfo. While useful for manual inspection, this text-based
interface poses challenges for production monitoring and large-scale
analysis:
1. Userspace must parse large amounts of text to extract specific
fields.
2. To find specific tags, userspace must read the entire dataset,
requiring many context switches and high data copying.
3. The kernel currently aggregates per-CPU counters for every allocation
size, even those the user intends to filter out immediately.
This series introduces a new IOCTL-based binary interface for allocinfo
that supports kernel-side filtering. By allowing the user to specify a
filter mask, we significantly reduce the work performed in-kernel and the
amount of data transferred to userspace. The IOCTL mechanism was chosen
for allocinfo to address the per-CPU counter aggregation bottleneck. A
traditional read() operation must report the total allocation count and
sizes for every code tag in the system. Doing so requires iterating
across all CPUs to sum their per-CPU counters for thousands of tags, which
introduces substantial runtime overhead.
The IOCTL interface allows userspace to push selective filtering criteria
directly into the kernel before the per-CPU counter aggregation. The
kernel aggregates per-CPU counters only for a small subset of tags that
match the filter. This results in significant performance improvement.
Beyond fast filtered retrieval, the IOCTL foundation allows introducing a
context capture mechanism in the future to capture the context for
specific allocations.
Performance measurements were conducted on an Intel Xeon Platinum 8481C
(224 CPUs) with caches dropped before each run.
The IOCTL mechanism shows a ~20x performance improvement for
filtered queries. The kernel avoids the expensive per-CPU counter
aggregation (alloc_tag_read) for any tags that fail the initial string
or location filters.
Scenario 1: Specific File Filtering (arch/x86/events/rapl.c)
1. Traditional (cat /proc/allocinfo | grep): 22ms (sys)
2. IOCTL Interface: 1ms (sys)
Scenario 2: Compound Filtering (Filename + Size)
1. Traditional: (cat ... | grep | awk): 21ms (sys)
2. IOCTL Interface: 1ms (sys)
Scenario 3: Size-Based Filtering (min_size = 1MB)
1. Traditional: (cat ... | awk): 21ms (sys)
2. IOCTL Interface: 14ms (sys)
This patch (of 6):
Add the following ioctl commands for /proc/allocinfo file:
ALLOCINFO_IOC_CONTENT_ID - gets content identifier which can be used to
check whether the file content has changed specifically due to module
load/unload. Every time a module is loaded / unloaded, the returned value
will be different. By comparing the identifier value at the beginning and
at the end of the content retrieval operation, users can validate
retrieved information for consistency.
ALLOCINFO_IOC_GET_AT - gets the record at the specified position. This is
the position of a record in /proc/allocinfo.
ALLOCINFO_IOC_GET_NEXT - gets the record next to the last retrieved one.
If no records were previously retrieved, returns the first record.
Note, function file and module names often have the same prefixes,
therefore when filtering for them, we compare the last 64 characters to
minimize the chances of name collisions.
[akpm@linux-foundation.org: include compat.h, per Suren]
Closes: https://lore.kernel.org/oe-kbuild-all/202607091820.qbjlGhKK-lkp@intel.com/
Link: https://lore.kernel.org/cover.1783532853.git.abhishekbapat@google.com
Link: https://lore.kernel.org/15596de2607ef13e7c77c6d74763f4ae992ec475.1783532853.git.abhishekbapat@google.com
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Abhishek Bapat <abhishekbapat@google.com>
Acked-by: Hao Ge <hao.ge@linux.dev>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Kent Overstreet <kent.overstreet@linux.dev>
Cc: Sourav Panda <souravpanda@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Commit 2b6a3f061f ("mm: declare VMA flags by bit") significantly
refactored the header file include/linux/mm.h. In that step, it
introduced a typo in an ifdef, referring to a non-existing config option
STACK_GROWS_UP, whereas the actual config option is called STACK_GROWSUP.
Commit 40a4af52e0 ("mm: fix CONFIG_STACK_GROWSUP typo in mm.h") fixed
this typo in the mm.h header file, but did not update the copy of the code
in tools/testing/vma/include/dup.h. Update this copy as well.
Commit message adapted from the above-referenced fix to mm.h.
Link: https://lore.kernel.org/20260611012258.432043-1-enelsonmoore@gmail.com
Signed-off-by: Ethan Nelson-Moore <enelsonmoore@gmail.com>
Reviewed-by: Lorenzo Stoakes <ljs@kernel.org>
Cc: Alice Ryhl <aliceryhl@google.com>
Cc: Jann Horn <jannh@google.com>
Cc: Liam R. Howlett <liam@infradead.org>
Cc: Pedro Falcato <pfalcato@suse.de>
Cc: Vlastimil Babka <vbabka@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Currently, hugetlb_cma reservation only supports absolute sizes (e.g.,
hugetlb_cma=2G or hugetlb_cma=0:1G,1:1G). This can be restrictive in
heterogeneous environments or when deploying common kernel command lines
across machines with different memory capacities.
Add support for percentage-based hugetlb_cma reservation (e.g.,
hugetlb_cma=20% or hugetlb_cma=0:20%,1:10%).
The percentage is calculated against the total memory (for global
settings) or against the node-specific memory (for node-specific settings)
using memblock APIs during early boot.
Link: https://lore.kernel.org/20260807040003.2156630-1-souravpanda@google.com
Signed-off-by: Sourav Panda <souravpanda@google.com>
Acked-by: Usama Arif <usama.arif@linux.dev>
Cc: David Hildenbrand <david@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Frank van der Linden <fvdl@google.com>
Cc: Greg Thelen <gthelen@google.com>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
