This reintroduces a concept removed by: commit d6cb41cc44 ("mm, hugetlb:
remove hugepages_treat_as_movable sysctl")
This sysctl provides flexibility between ZONE_MOVABLE use cases:
1) onlining memory in ZONE_MOVABLE to maintain hotplug compatibility
2) onlining memory in ZONE_MOVABLE to make hugepage allocate reliable
When ZONE_MOVABLE is used to make huge page allocation more reliable,
disallowing gigantic pages memory in this region is pointless. If hotplug
is not a requirement, we can loosen the restrictions to allow 1GB gigantic
pages in ZONE_MOVABLE.
Since 1GB can be difficult to migrate / has impacts on compaction /
defragmentation, we don't enable this by default. Notably, 1GB pages can
only be migrated if another 1GB page is available - so hot-unplug will
fail if such a page cannot be found.
However, since there are scenarios where gigantic pages are migratable, we
should allow use of these on movable regions.
When not valid 1GB is available for migration, hot-unplug will retry
indefinitely (or until interrupted). For example:
echo 0 > node0/hugepages/..-1GB/nr_hugepages # clear node0 1GB pages
echo 1 > node1/hugepages/..-1GB/nr_hugepages # reserve node1 1GB page
./alloc_huge_node1 & # Allocate a 1GB page on node1
./node1_offline & # attempt to offline all node1 memory
echo 1 > node0/hugepages/..-1GB/nr_hugepages # reserve node0 1GB page
In this example, node1_offline will block indefinitely until the final
step, when a node0 1GB page is made available.
Note: Boot-time CMA is not possible for driver-managed hotplug memory, as
CMA requires the memory to be registered as SystemRAM at boot time.
Additionally, 1GB huge pages are not supported by THP.
Link: https://lkml.kernel.org/r/20251221125603.2364174-1-gourry@gourry.net
Signed-off-by: Gregory Price <gourry@gourry.net>
Suggested-by: David Rientjes <rientjes@google.com>
Link: https://lore.kernel.org/all/20180201193132.Hk7vI_xaU%25akpm@linux-foundation.org/
Acked-by: David Hildenbrand (Red Hat) <david@kernel.org>
Acked-by: David Rientjes <rientjes@google.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Michal Hocko <mhocko@suse.com>
Cc: "David Hildenbrand (Red Hat)" <david@kernel.org>
Cc: Gregory Price <gourry@gourry.net>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Liam Howlett <liam.howlett@oracle.com>
Cc: Lorenzo Stoakes <lorenzo.stoakes@oracle.com>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Muchun Song <muchun.song@linux.dev>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Suren Baghdasaryan <surenb@google.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Laptop mode was introduced to save battery, by delaying and consolidating
writes and thereby maximize the time rotating hard drives wouldn't have to
spin.
Luckily, rotating hard drives, with their high spin-up times and power
draw, are a thing of the past for battery-powered devices. Reclaim has
also since changed to not write single filesystem pages anymore, and
regular filesystem writeback is lumpy by design.
The juice doesn't appear worth the squeeze anymore. The footprint of the
feature is small, but nevertheless it's a complicating factor in mm,
block, filesystems. Developers don't think about it, and it likely hasn't
been tested with new reclaim and writeback changes in years.
Let's sunset it. Keep the sysctl with a deprecation warning around for a
few more cycles, but remove all functionality behind it.
[akpm@linux-foundation.org: fix Documentation/admin-guide/laptops/index.rst]
Link: https://lkml.kernel.org/r/20251216185201.GH905277@cmpxchg.org
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Suggested-by: Christoph Hellwig <hch@infradead.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Acked-by: Jens Axboe <axboe@kernel.dk>
Reviewed-by: Shakeel Butt <shakeel.butt@linux.dev>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Deepanshu Kartikey <kartikey406@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Removing them solves an issue where they were incorrectly considered as
not implemented by the check-sysctl-docs script
Signed-off-by: Joel Granados <joel.granados@kernel.org>
Pull MM updates from Andrew Morton:
- "Add folio_mk_pte()" from Matthew Wilcox simplifies the act of
creating a pte which addresses the first page in a folio and reduces
the amount of plumbing which architecture must implement to provide
this.
- "Misc folio patches for 6.16" from Matthew Wilcox is a shower of
largely unrelated folio infrastructure changes which clean things up
and better prepare us for future work.
- "memory,x86,acpi: hotplug memory alignment advisement" from Gregory
Price adds early-init code to prevent x86 from leaving physical
memory unused when physical address regions are not aligned to memory
block size.
- "mm/compaction: allow more aggressive proactive compaction" from
Michal Clapinski provides some tuning of the (sadly, hard-coded (more
sadly, not auto-tuned)) thresholds for our invokation of proactive
compaction. In a simple test case, the reduction of a guest VM's
memory consumption was dramatic.
- "Minor cleanups and improvements to swap freeing code" from Kemeng
Shi provides some code cleaups and a small efficiency improvement to
this part of our swap handling code.
- "ptrace: introduce PTRACE_SET_SYSCALL_INFO API" from Dmitry Levin
adds the ability for a ptracer to modify syscalls arguments. At this
time we can alter only "system call information that are used by
strace system call tampering, namely, syscall number, syscall
arguments, and syscall return value.
This series should have been incorporated into mm.git's "non-MM"
branch, but I goofed.
- "fs/proc: extend the PAGEMAP_SCAN ioctl to report guard regions" from
Andrei Vagin extends the info returned by the PAGEMAP_SCAN ioctl
against /proc/pid/pagemap. This permits CRIU to more efficiently get
at the info about guard regions.
- "Fix parameter passed to page_mapcount_is_type()" from Gavin Shan
implements that fix. No runtime effect is expected because
validate_page_before_insert() happens to fix up this error.
- "kernel/events/uprobes: uprobe_write_opcode() rewrite" from David
Hildenbrand basically brings uprobe text poking into the current
decade. Remove a bunch of hand-rolled implementation in favor of
using more current facilities.
- "mm/ptdump: Drop assumption that pxd_val() is u64" from Anshuman
Khandual provides enhancements and generalizations to the pte dumping
code. This might be needed when 128-bit Page Table Descriptors are
enabled for ARM.
- "Always call constructor for kernel page tables" from Kevin Brodsky
ensures that the ctor/dtor is always called for kernel pgtables, as
it already is for user pgtables.
This permits the addition of more functionality such as "insert hooks
to protect page tables". This change does result in various
architectures performing unnecesary work, but this is fixed up where
it is anticipated to occur.
- "Rust support for mm_struct, vm_area_struct, and mmap" from Alice
Ryhl adds plumbing to permit Rust access to core MM structures.
- "fix incorrectly disallowed anonymous VMA merges" from Lorenzo
Stoakes takes advantage of some VMA merging opportunities which we've
been missing for 15 years.
- "mm/madvise: batch tlb flushes for MADV_DONTNEED and MADV_FREE" from
SeongJae Park optimizes process_madvise()'s TLB flushing.
Instead of flushing each address range in the provided iovec, we
batch the flushing across all the iovec entries. The syscall's cost
was approximately halved with a microbenchmark which was designed to
load this particular operation.
- "Track node vacancy to reduce worst case allocation counts" from
Sidhartha Kumar makes the maple tree smarter about its node
preallocation.
stress-ng mmap performance increased by single-digit percentages and
the amount of unnecessarily preallocated memory was dramaticelly
reduced.
- "mm/gup: Minor fix, cleanup and improvements" from Baoquan He removes
a few unnecessary things which Baoquan noted when reading the code.
- ""Enhance sysfs handling for memory hotplug in weighted interleave"
from Rakie Kim "enhances the weighted interleave policy in the memory
management subsystem by improving sysfs handling, fixing memory
leaks, and introducing dynamic sysfs updates for memory hotplug
support". Fixes things on error paths which we are unlikely to hit.
- "mm/damon: auto-tune DAMOS for NUMA setups including tiered memory"
from SeongJae Park introduces new DAMOS quota goal metrics which
eliminate the manual tuning which is required when utilizing DAMON
for memory tiering.
- "mm/vmalloc.c: code cleanup and improvements" from Baoquan He
provides cleanups and small efficiency improvements which Baoquan
found via code inspection.
- "vmscan: enforce mems_effective during demotion" from Gregory Price
changes reclaim to respect cpuset.mems_effective during demotion when
possible. because presently, reclaim explicitly ignores
cpuset.mems_effective when demoting, which may cause the cpuset
settings to violated.
This is useful for isolating workloads on a multi-tenant system from
certain classes of memory more consistently.
- "Clean up split_huge_pmd_locked() and remove unnecessary folio
pointers" from Gavin Guo provides minor cleanups and efficiency gains
in in the huge page splitting and migrating code.
- "Use kmem_cache for memcg alloc" from Huan Yang creates a slab cache
for `struct mem_cgroup', yielding improved memory utilization.
- "add max arg to swappiness in memory.reclaim and lru_gen" from
Zhongkun He adds a new "max" argument to the "swappiness=" argument
for memory.reclaim MGLRU's lru_gen.
This directs proactive reclaim to reclaim from only anon folios
rather than file-backed folios.
- "kexec: introduce Kexec HandOver (KHO)" from Mike Rapoport is the
first step on the path to permitting the kernel to maintain existing
VMs while replacing the host kernel via file-based kexec. At this
time only memblock's reserve_mem is preserved.
- "mm: Introduce for_each_valid_pfn()" from David Woodhouse provides
and uses a smarter way of looping over a pfn range. By skipping
ranges of invalid pfns.
- "sched/numa: Skip VMA scanning on memory pinned to one NUMA node via
cpuset.mems" from Libo Chen removes a lot of pointless VMA scanning
when a task is pinned a single NUMA mode.
Dramatic performance benefits were seen in some real world cases.
- "JFS: Implement migrate_folio for jfs_metapage_aops" from Shivank
Garg addresses a warning which occurs during memory compaction when
using JFS.
- "move all VMA allocation, freeing and duplication logic to mm" from
Lorenzo Stoakes moves some VMA code from kernel/fork.c into the more
appropriate mm/vma.c.
- "mm, swap: clean up swap cache mapping helper" from Kairui Song
provides code consolidation and cleanups related to the folio_index()
function.
- "mm/gup: Cleanup memfd_pin_folios()" from Vishal Moola does that.
- "memcg: Fix test_memcg_min/low test failures" from Waiman Long
addresses some bogus failures which are being reported by the
test_memcontrol selftest.
- "eliminate mmap() retry merge, add .mmap_prepare hook" from Lorenzo
Stoakes commences the deprecation of file_operations.mmap() in favor
of the new file_operations.mmap_prepare().
The latter is more restrictive and prevents drivers from messing with
things in ways which, amongst other problems, may defeat VMA merging.
- "memcg: decouple memcg and objcg stocks"" from Shakeel Butt decouples
the per-cpu memcg charge cache from the objcg's one.
This is a step along the way to making memcg and objcg charging
NMI-safe, which is a BPF requirement.
- "mm/damon: minor fixups and improvements for code, tests, and
documents" from SeongJae Park is yet another batch of miscellaneous
DAMON changes. Fix and improve minor problems in code, tests and
documents.
- "memcg: make memcg stats irq safe" from Shakeel Butt converts memcg
stats to be irq safe. Another step along the way to making memcg
charging and stats updates NMI-safe, a BPF requirement.
- "Let unmap_hugepage_range() and several related functions take folio
instead of page" from Fan Ni provides folio conversions in the
hugetlb code.
* tag 'mm-stable-2025-05-31-14-50' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm: (285 commits)
mm: pcp: increase pcp->free_count threshold to trigger free_high
mm/hugetlb: convert use of struct page to folio in __unmap_hugepage_range()
mm/hugetlb: refactor __unmap_hugepage_range() to take folio instead of page
mm/hugetlb: refactor unmap_hugepage_range() to take folio instead of page
mm/hugetlb: pass folio instead of page to unmap_ref_private()
memcg: objcg stock trylock without irq disabling
memcg: no stock lock for cpu hot-unplug
memcg: make __mod_memcg_lruvec_state re-entrant safe against irqs
memcg: make count_memcg_events re-entrant safe against irqs
memcg: make mod_memcg_state re-entrant safe against irqs
memcg: move preempt disable to callers of memcg_rstat_updated
memcg: memcg_rstat_updated re-entrant safe against irqs
mm: khugepaged: decouple SHMEM and file folios' collapse
selftests/eventfd: correct test name and improve messages
alloc_tag: check mem_profiling_support in alloc_tag_init
Docs/damon: update titles and brief introductions to explain DAMOS
selftests/damon/_damon_sysfs: read tried regions directories in order
mm/damon/tests/core-kunit: add a test for damos_set_filters_default_reject()
mm/damon/paddr: remove unused variable, folio_list, in damon_pa_stat()
mm/damon/sysfs-schemes: fix wrong comment on damons_sysfs_quota_goal_metric_strs
...
On our HDFS servers with 12 HDDs per server, a HDFS datanode[0] startup
involves scanning all files and caching their metadata (including dentries
and inodes) in memory. Each HDD contains approximately 2 million files,
resulting in a total of ~20 million cached dentries after initialization.
To minimize dentry reclamation, we set vfs_cache_pressure to 1. Despite
this configuration, memory pressure conditions can still trigger
reclamation of up to 50% of cached dentries, reducing the cache from 20
million to approximately 10 million entries. During the subsequent cache
rebuild period, any HDFS datanode restart operation incurs substantial
latency penalties until full cache recovery completes.
To maintain service stability, we need to preserve more dentries during
memory reclamation. The current minimum reclaim ratio (1/100 of total
dentries) remains too aggressive for our workload. This patch introduces
vfs_cache_pressure_denom for more granular cache pressure control. The
configuration [vfs_cache_pressure=1, vfs_cache_pressure_denom=10000]
effectively maintains the full 20 million dentry cache under memory
pressure, preventing datanode restart performance degradation.
Link: https://hadoop.apache.org/docs/r1.2.1/hdfs_design.html#NameNode+and+DataNodes [0]
Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
Link: https://lore.kernel.org/20250511083624.9305-1-laoar.shao@gmail.com
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Christian Brauner <brauner@kernel.org>
The page allocator groups requests by migratetype to stave off
fragmentation. However, in practice this is routinely defeated by the
fact that it gives up *before* invoking reclaim and compaction - which may
well produce suitable pages. As a result, fragmentation of physical
memory is a common ongoing process in many load scenarios.
Fragmentation deteriorates compaction's ability to produce huge pages.
Depending on the lifetime of the fragmenting allocations, those effects
can be long-lasting or even permanent, requiring drastic measures like
forcible idle states or even reboots as the only reliable ways to recover
the address space for THP production.
In a kernel build test with supplemental THP pressure, the THP allocation
rate steadily declines over 15 runs:
thp_fault_alloc
61988
56474
57258
50187
52388
55409
52925
47648
43669
40621
36077
41721
36685
34641
33215
This is a hurdle in adopting THP in any environment where hosts are shared
between multiple overlapping workloads (cloud environments), and rarely
experience true idle periods. To make THP a reliable and predictable
optimization, there needs to be a stronger guarantee to avoid such
fragmentation.
Introduce defrag_mode. When enabled, reclaim/compaction is invoked to its
full extent *before* falling back. Specifically, ALLOC_NOFRAGMENT is
enforced on the allocator fastpath and the reclaiming slowpath.
For now, fallbacks are permitted to avert OOMs. There is a plan to add
defrag_mode=2 to prefer OOMs over fragmentation, but this requires
additional prep work in compaction and the reserve management to make it
ready for all possible allocation contexts.
The following test results are from a kernel build with periodic bursts of
THP allocations, over 15 runs:
vanilla defrag_mode=1
@claimer[unmovable]: 189 103
@claimer[movable]: 92 103
@claimer[reclaimable]: 207 61
@pollute[unmovable from movable]: 25 0
@pollute[unmovable from reclaimable]: 28 0
@pollute[movable from unmovable]: 38835 0
@pollute[movable from reclaimable]: 147136 0
@pollute[reclaimable from unmovable]: 178 0
@pollute[reclaimable from movable]: 33 0
@steal[unmovable from movable]: 11 0
@steal[unmovable from reclaimable]: 5 0
@steal[reclaimable from unmovable]: 107 0
@steal[reclaimable from movable]: 90 0
@steal[movable from reclaimable]: 354 0
@steal[movable from unmovable]: 130 0
Both types of polluting fallbacks are eliminated in this workload.
Interestingly, whole block conversions are reduced as well. This is
because once a block is claimed for a type, its empty space remains
available for future allocations, instead of being padded with fallbacks;
this allows the native type to group up instead of spreading out to new
blocks. The assumption in the allocator has been that pollution from
movable allocations is less harmful than from other types, since they can
be reclaimed or migrated out should the space be needed. However, since
fallbacks occur *before* reclaim/compaction is invoked, movable pollution
will still cause non-movable allocations to spread out and claim more
blocks.
Without fragmentation, THP rates hold steady with defrag_mode=1:
thp_fault_alloc
32478
20725
45045
32130
14018
21711
40791
29134
34458
45381
28305
17265
22584
28454
30850
While the downward trend is eliminated, the keen reader will of course
notice that the baseline rate is much smaller than the vanilla kernel's to
begin with. This is due to deficiencies in how reclaim and compaction are
currently driven: ALLOC_NOFRAGMENT increases the extent to which smaller
allocations are competing with THPs for pageblocks, while making no effort
themselves to reclaim or compact beyond their own request size. This
effect already exists with the current usage of ALLOC_NOFRAGMENT, but is
amplified by defrag_mode insisting on whole block stealing much more
strongly.
Subsequent patches will address defrag_mode reclaim strategy to raise the
THP success baseline above the vanilla kernel.
Link: https://lkml.kernel.org/r/20250313210647.1314586-4-hannes@cmpxchg.org
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Zi Yan <ziy@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
For now, the feature of hugetlb_free_vmemmap is not compatible with the
feature of memory_hotplug.memmap_on_memory, and hugetlb_free_vmemmap takes
precedence over memory_hotplug.memmap_on_memory. However, someone wants
to make memory_hotplug.memmap_on_memory takes precedence over
hugetlb_free_vmemmap since memmap_on_memory makes it more likely to
succeed memory hotplug in close-to-OOM situations. So the decision of
making hugetlb_free_vmemmap take precedence is not wise and elegant.
The proper approach is to have hugetlb_vmemmap.c do the check whether the
section which the HugeTLB pages belong to can be optimized. If the
section's vmemmap pages are allocated from the added memory block itself,
hugetlb_free_vmemmap should refuse to optimize the vmemmap, otherwise, do
the optimization. Then both kernel parameters are compatible. So this
patch introduces VmemmapSelfHosted to mask any non-optimizable vmemmap
pages. The hugetlb_vmemmap can use this flag to detect if a vmemmap page
can be optimized.
[songmuchun@bytedance.com: walk vmemmap page tables to avoid false-positive]
Link: https://lkml.kernel.org/r/20220620110616.12056-3-songmuchun@bytedance.com
Link: https://lkml.kernel.org/r/20220617135650.74901-3-songmuchun@bytedance.com
Signed-off-by: Muchun Song <songmuchun@bytedance.com>
Co-developed-by: Oscar Salvador <osalvador@suse.de>
Signed-off-by: Oscar Salvador <osalvador@suse.de>
Acked-by: David Hildenbrand <david@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Paul E. McKenney <paulmck@kernel.org>
Cc: Xiongchun Duan <duanxiongchun@bytedance.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
so it will be consistent with code mm directory and with
Documentation/admin-guide/mm and won't be confused with virtual machines.
Signed-off-by: Mike Rapoport <rppt@linux.ibm.com>
Suggested-by: Matthew Wilcox <willy@infradead.org>
Tested-by: Ira Weiny <ira.weiny@intel.com>
Acked-by: Jonathan Corbet <corbet@lwn.net>
Acked-by: Wu XiangCheng <bobwxc@email.cn>
We must add hugetlb_free_vmemmap=on (or "off") to the boot cmdline and
reboot the server to enable or disable the feature of optimizing vmemmap
pages associated with HugeTLB pages. However, rebooting usually takes a
long time. So add a sysctl to enable or disable the feature at runtime
without rebooting. Why we need this? There are 3 use cases.
1) The feature of minimizing overhead of struct page associated with
each HugeTLB is disabled by default without passing
"hugetlb_free_vmemmap=on" to the boot cmdline. When we (ByteDance)
deliver the servers to the users who want to enable this feature, they
have to configure the grub (change boot cmdline) and reboot the
servers, whereas rebooting usually takes a long time (we have thousands
of servers). It's a very bad experience for the users. So we need a
approach to enable this feature after rebooting. This is a use case in
our practical environment.
2) Some use cases are that HugeTLB pages are allocated 'on the fly'
instead of being pulled from the HugeTLB pool, those workloads would be
affected with this feature enabled. Those workloads could be
identified by the characteristics of they never explicitly allocating
huge pages with 'nr_hugepages' but only set 'nr_overcommit_hugepages'
and then let the pages be allocated from the buddy allocator at fault
time. We can confirm it is a real use case from the commit
099730d674. For those workloads, the page fault time could be ~2x
slower than before. We suspect those users want to disable this
feature if the system has enabled this before and they don't think the
memory savings benefit is enough to make up for the performance drop.
3) If the workload which wants vmemmap pages to be optimized and the
workload which wants to set 'nr_overcommit_hugepages' and does not want
the extera overhead at fault time when the overcommitted pages be
allocated from the buddy allocator are deployed in the same server.
The user could enable this feature and set 'nr_hugepages' and
'nr_overcommit_hugepages', then disable the feature. In this case, the
overcommited HugeTLB pages will not encounter the extra overhead at
fault time.
Link: https://lkml.kernel.org/r/20220512041142.39501-5-songmuchun@bytedance.com
Signed-off-by: Muchun Song <songmuchun@bytedance.com>
Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Luis Chamberlain <mcgrof@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Iurii Zaikin <yzaikin@google.com>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: David Hildenbrand <david@redhat.com>
Cc: Masahiro Yamada <masahiroy@kernel.org>
Cc: Xiongchun Duan <duanxiongchun@bytedance.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
The proactive compaction[1] gets triggered for every 500msec and run
compaction on the node for COMPACTION_HPAGE_ORDER (usually order-9) pages
based on the value set to sysctl.compaction_proactiveness. Triggering the
compaction for every 500msec in search of COMPACTION_HPAGE_ORDER pages is
not needed for all applications, especially on the embedded system
usecases which may have few MB's of RAM. Enabling the proactive
compaction in its state will endup in running almost always on such
systems.
Other side, proactive compaction can still be very much useful for getting
a set of higher order pages in some controllable manner(controlled by
using the sysctl.compaction_proactiveness). So, on systems where enabling
the proactive compaction always may proove not required, can trigger the
same from user space on write to its sysctl interface. As an example, say
app launcher decide to launch the memory heavy application which can be
launched fast if it gets more higher order pages thus launcher can prepare
the system in advance by triggering the proactive compaction from
userspace.
This triggering of proactive compaction is done on a write to
sysctl.compaction_proactiveness by user.
[1]https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit?id=facdaa917c4d5a376d09d25865f5a863f906234a
[akpm@linux-foundation.org: tweak vm.rst, per Mike]
Link: https://lkml.kernel.org/r/1627653207-12317-1-git-send-email-charante@codeaurora.org
Signed-off-by: Charan Teja Reddy <charante@codeaurora.org>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Acked-by: Rafael Aquini <aquini@redhat.com>
Cc: Mike Rapoport <rppt@kernel.org>
Cc: Luis Chamberlain <mcgrof@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Iurii Zaikin <yzaikin@google.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Nitin Gupta <nigupta@nvidia.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Khalid Aziz <khalid.aziz@oracle.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Vinayak Menon <vinmenon@codeaurora.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This introduces a new sysctl vm.percpu_pagelist_high_fraction. It is
similar to the old vm.percpu_pagelist_fraction. The old sysctl increased
both pcp->batch and pcp->high with the higher pcp->high potentially
reducing zone->lock contention. However, the higher pcp->batch value also
potentially increased allocation latency while the PCP was refilled. This
sysctl only adjusts pcp->high so that zone->lock contention is potentially
reduced but allocation latency during a PCP refill remains the same.
# grep -E "high:|batch" /proc/zoneinfo | tail -2
high: 649
batch: 63
# sysctl vm.percpu_pagelist_high_fraction=8
# grep -E "high:|batch" /proc/zoneinfo | tail -2
high: 35071
batch: 63
# sysctl vm.percpu_pagelist_high_fraction=64
high: 4383
batch: 63
# sysctl vm.percpu_pagelist_high_fraction=0
high: 649
batch: 63
[mgorman@techsingularity.net: fix documentation]
Link: https://lkml.kernel.org/r/20210528151010.GQ30378@techsingularity.net
Link: https://lkml.kernel.org/r/20210525080119.5455-7-mgorman@techsingularity.net
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Hillf Danton <hdanton@sina.com>
Cc: Michal Hocko <mhocko@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Patch series "Calculate pcp->high based on zone sizes and active CPUs", v2.
The per-cpu page allocator (PCP) is meant to reduce contention on the zone
lock but the sizing of batch and high is archaic and neither takes the
zone size into account or the number of CPUs local to a zone. With larger
zones and more CPUs per node, the contention is getting worse.
Furthermore, the fact that vm.percpu_pagelist_fraction adjusts both batch
and high values means that the sysctl can reduce zone lock contention but
also increase allocation latencies.
This series disassociates pcp->high from pcp->batch and then scales
pcp->high based on the size of the local zone with limited impact to
reclaim and accounting for active CPUs but leaves pcp->batch static. It
also adapts the number of pages that can be on the pcp list based on
recent freeing patterns.
The motivation is partially to adjust to larger memory sizes but is also
driven by the fact that large batches of page freeing via release_pages()
often shows zone contention as a major part of the problem. Another is a
bug report based on an older kernel where a multi-terabyte process can
takes several minutes to exit. A workaround was to use
vm.percpu_pagelist_fraction to increase the pcp->high value but testing
indicated that a production workload could not use the same values because
of an increase in allocation latencies. Unfortunately, I cannot reproduce
this test case myself as the multi-terabyte machines are in active use but
it should alleviate the problem.
The series aims to address both and partially acts as a pre-requisite.
pcp only works with order-0 which is useless for SLUB (when using high
orders) and THP (unconditionally). To store high-order pages on PCP, the
pcp->high values need to be increased first.
This patch (of 6):
The vm.percpu_pagelist_fraction is used to increase the batch and high
limits for the per-cpu page allocator (PCP). The intent behind the sysctl
is to reduce zone lock acquisition when allocating/freeing pages but it
has a problem. While it can decrease contention, it can also increase
latency on the allocation side due to unreasonably large batch sizes.
This leads to games where an administrator adjusts
percpu_pagelist_fraction on the fly to work around contention and
allocation latency problems.
This series aims to alleviate the problems with zone lock contention while
avoiding the allocation-side latency problems. For the purposes of
review, it's easier to remove this sysctl now and reintroduce a similar
sysctl later in the series that deals only with pcp->high.
Link: https://lkml.kernel.org/r/20210525080119.5455-1-mgorman@techsingularity.net
Link: https://lkml.kernel.org/r/20210525080119.5455-2-mgorman@techsingularity.net
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Hillf Danton <hdanton@sina.com>
Cc: Michal Hocko <mhocko@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
I went to go add a new RECLAIM_* mode for the zone_reclaim_mode sysctl.
Like a good kernel developer, I also went to go update the
documentation. I noticed that the bits in the documentation didn't
match the bits in the #defines.
The VM never explicitly checks the RECLAIM_ZONE bit. The bit is,
however implicitly checked when checking 'node_reclaim_mode==0'. The
RECLAIM_ZONE #define was removed in a cleanup. That, by itself is fine.
But, when the bit was removed (bit 0) the _other_ bit locations also got
changed. That's not OK because the bit values are documented to mean
one specific thing. Users surely do not expect the meaning to change
from kernel to kernel.
The end result is that if someone had a script that did:
sysctl vm.zone_reclaim_mode=1
it would have gone from enabling node reclaim for clean unmapped pages
to writing out pages during node reclaim after the commit in question.
That's not great.
Put the bits back the way they were and add a comment so something like
this is a bit harder to do again. Update the documentation to make it
clear that the first bit is ignored.
Link: https://lkml.kernel.org/r/20210219172555.FF0CDF23@viggo.jf.intel.com
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Fixes: 648b5cf368 ("mm/vmscan: remove unused RECLAIM_OFF/RECLAIM_ZONE")
Reviewed-by: Ben Widawsky <ben.widawsky@intel.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
Acked-by: David Rientjes <rientjes@google.com>
Acked-by: Christoph Lameter <cl@linux.com>
Cc: Alex Shi <alex.shi@linux.alibaba.com>
Cc: Daniel Wagner <dwagner@suse.de>
Cc: "Tobin C. Harding" <tobin@kernel.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Huang Ying <ying.huang@intel.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Qian Cai <cai@lca.pw>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pull documentation fixes from Jonathan Corbet:
"A small set of late-arriving, small documentation fixes"
* tag 'docs-5.11-2' of git://git.lwn.net/linux:
docs: admin-guide: Fix default value of max_map_count in sysctl/vm.rst
Documentation/submitting-patches: Document the SoB chain
Documentation: process: Correct numbering
docs: submitting-patches: Trivial - fix grammatical error
Since the default value of sysctl_max_map_count is defined as
DEFAULT_MAX_MAP_COUNT from mm/util.c
int sysctl_max_map_count __read_mostly = DEFAULT_MAX_MAP_COUNT;
DEFAULT_MAX_MAP_COUNT is defined as 65530 (65535-5) in include/linux/mm.h
#define MAPCOUNT_ELF_CORE_MARGIN (5)
#define DEFAULT_MAX_MAP_COUNT (USHRT_MAX - MAPCOUNT_ELF_CORE_MARGIN)
Signed-off-by: Fengfei Xi <xi.fengfei@h3c.com>
Link: https://lore.kernel.org/r/20201210082134.36957-1-xi.fengfei@h3c.com
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Merge misc updates from Andrew Morton:
- a few random little subsystems
- almost all of the MM patches which are staged ahead of linux-next
material. I'll trickle to post-linux-next work in as the dependents
get merged up.
Subsystems affected by this patch series: kthread, kbuild, ide, ntfs,
ocfs2, arch, and mm (slab-generic, slab, slub, dax, debug, pagecache,
gup, swap, shmem, memcg, pagemap, mremap, hmm, vmalloc, documentation,
kasan, pagealloc, memory-failure, hugetlb, vmscan, z3fold, compaction,
oom-kill, migration, cma, page-poison, userfaultfd, zswap, zsmalloc,
uaccess, zram, and cleanups).
* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (200 commits)
mm: cleanup kstrto*() usage
mm: fix fall-through warnings for Clang
mm: slub: convert sysfs sprintf family to sysfs_emit/sysfs_emit_at
mm: shmem: convert shmem_enabled_show to use sysfs_emit_at
mm:backing-dev: use sysfs_emit in macro defining functions
mm: huge_memory: convert remaining use of sprintf to sysfs_emit and neatening
mm: use sysfs_emit for struct kobject * uses
mm: fix kernel-doc markups
zram: break the strict dependency from lzo
zram: add stat to gather incompressible pages since zram set up
zram: support page writeback
mm/process_vm_access: remove redundant initialization of iov_r
mm/zsmalloc.c: rework the list_add code in insert_zspage()
mm/zswap: move to use crypto_acomp API for hardware acceleration
mm/zswap: fix passing zero to 'PTR_ERR' warning
mm/zswap: make struct kernel_param_ops definitions const
userfaultfd/selftests: hint the test runner on required privilege
userfaultfd/selftests: fix retval check for userfaultfd_open()
userfaultfd/selftests: always dump something in modes
userfaultfd: selftests: make __{s,u}64 format specifiers portable
...
For some applications, we need to allocate almost all memory as hugepages.
However, on a running system, higher-order allocations can fail if the
memory is fragmented. Linux kernel currently does on-demand compaction as
we request more hugepages, but this style of compaction incurs very high
latency. Experiments with one-time full memory compaction (followed by
hugepage allocations) show that kernel is able to restore a highly
fragmented memory state to a fairly compacted memory state within <1 sec
for a 32G system. Such data suggests that a more proactive compaction can
help us allocate a large fraction of memory as hugepages keeping
allocation latencies low.
For a more proactive compaction, the approach taken here is to define a
new sysctl called 'vm.compaction_proactiveness' which dictates bounds for
external fragmentation which kcompactd tries to maintain.
The tunable takes a value in range [0, 100], with a default of 20.
Note that a previous version of this patch [1] was found to introduce too
many tunables (per-order extfrag{low, high}), but this one reduces them to
just one sysctl. Also, the new tunable is an opaque value instead of
asking for specific bounds of "external fragmentation", which would have
been difficult to estimate. The internal interpretation of this opaque
value allows for future fine-tuning.
Currently, we use a simple translation from this tunable to [low, high]
"fragmentation score" thresholds (low=100-proactiveness, high=low+10%).
The score for a node is defined as weighted mean of per-zone external
fragmentation. A zone's present_pages determines its weight.
To periodically check per-node score, we reuse per-node kcompactd threads,
which are woken up every 500 milliseconds to check the same. If a node's
score exceeds its high threshold (as derived from user-provided
proactiveness value), proactive compaction is started until its score
reaches its low threshold value. By default, proactiveness is set to 20,
which implies threshold values of low=80 and high=90.
This patch is largely based on ideas from Michal Hocko [2]. See also the
LWN article [3].
Performance data
================
System: x64_64, 1T RAM, 80 CPU threads.
Kernel: 5.6.0-rc3 + this patch
echo madvise | sudo tee /sys/kernel/mm/transparent_hugepage/enabled
echo madvise | sudo tee /sys/kernel/mm/transparent_hugepage/defrag
Before starting the driver, the system was fragmented from a userspace
program that allocates all memory and then for each 2M aligned section,
frees 3/4 of base pages using munmap. The workload is mainly anonymous
userspace pages, which are easy to move around. I intentionally avoided
unmovable pages in this test to see how much latency we incur when
hugepage allocations hit direct compaction.
1. Kernel hugepage allocation latencies
With the system in such a fragmented state, a kernel driver then allocates
as many hugepages as possible and measures allocation latency:
(all latency values are in microseconds)
- With vanilla 5.6.0-rc3
percentile latency
–––––––––– –––––––
5 7894
10 9496
25 12561
30 15295
40 18244
50 21229
60 27556
75 30147
80 31047
90 32859
95 33799
Total 2M hugepages allocated = 383859 (749G worth of hugepages out of 762G
total free => 98% of free memory could be allocated as hugepages)
- With 5.6.0-rc3 + this patch, with proactiveness=20
sysctl -w vm.compaction_proactiveness=20
percentile latency
–––––––––– –––––––
5 2
10 2
25 3
30 3
40 3
50 4
60 4
75 4
80 4
90 5
95 429
Total 2M hugepages allocated = 384105 (750G worth of hugepages out of 762G
total free => 98% of free memory could be allocated as hugepages)
2. JAVA heap allocation
In this test, we first fragment memory using the same method as for (1).
Then, we start a Java process with a heap size set to 700G and request the
heap to be allocated with THP hugepages. We also set THP to madvise to
allow hugepage backing of this heap.
/usr/bin/time
java -Xms700G -Xmx700G -XX:+UseTransparentHugePages -XX:+AlwaysPreTouch
The above command allocates 700G of Java heap using hugepages.
- With vanilla 5.6.0-rc3
17.39user 1666.48system 27:37.89elapsed
- With 5.6.0-rc3 + this patch, with proactiveness=20
8.35user 194.58system 3:19.62elapsed
Elapsed time remains around 3:15, as proactiveness is further increased.
Note that proactive compaction happens throughout the runtime of these
workloads. The situation of one-time compaction, sufficient to supply
hugepages for following allocation stream, can probably happen for more
extreme proactiveness values, like 80 or 90.
In the above Java workload, proactiveness is set to 20. The test starts
with a node's score of 80 or higher, depending on the delay between the
fragmentation step and starting the benchmark, which gives more-or-less
time for the initial round of compaction. As t he benchmark consumes
hugepages, node's score quickly rises above the high threshold (90) and
proactive compaction starts again, which brings down the score to the low
threshold level (80). Repeat.
bpftrace also confirms proactive compaction running 20+ times during the
runtime of this Java benchmark. kcompactd threads consume 100% of one of
the CPUs while it tries to bring a node's score within thresholds.
Backoff behavior
================
Above workloads produce a memory state which is easy to compact. However,
if memory is filled with unmovable pages, proactive compaction should
essentially back off. To test this aspect:
- Created a kernel driver that allocates almost all memory as hugepages
followed by freeing first 3/4 of each hugepage.
- Set proactiveness=40
- Note that proactive_compact_node() is deferred maximum number of times
with HPAGE_FRAG_CHECK_INTERVAL_MSEC of wait between each check
(=> ~30 seconds between retries).
[1] https://patchwork.kernel.org/patch/11098289/
[2] https://lore.kernel.org/linux-mm/20161230131412.GI13301@dhcp22.suse.cz/
[3] https://lwn.net/Articles/817905/
Signed-off-by: Nitin Gupta <nigupta@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Tested-by: Oleksandr Natalenko <oleksandr@redhat.com>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Reviewed-by: Khalid Aziz <khalid.aziz@oracle.com>
Reviewed-by: Oleksandr Natalenko <oleksandr@redhat.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Khalid Aziz <khalid.aziz@oracle.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Nitin Gupta <ngupta@nitingupta.dev>
Cc: Oleksandr Natalenko <oleksandr@redhat.com>
Link: http://lkml.kernel.org/r/20200616204527.19185-1-nigupta@nvidia.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The stuff under sysctl describes /sys interface from userspace
point of view. So, add it to the admin-guide and remove the
:orphan: from its index file.
Signed-off-by: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>
