1 # SPDX-License-Identifier: GPL-2.0-only
3 menu "Memory Management options"
6 # For some reason microblaze and nios2 hard code SWAP=n. Hopefully we can
7 # add proper SWAP support to them, in which case this can be remove.
16 bool "Support for paging of anonymous memory (swap)"
17 depends on MMU && BLOCK && !ARCH_NO_SWAP
20 This option allows you to choose whether you want to have support
21 for so called swap devices or swap files in your kernel that are
22 used to provide more virtual memory than the actual RAM present
23 in your computer. If unsure say Y.
26 bool "Compressed cache for swap pages"
31 A lightweight compressed cache for swap pages. It takes
32 pages that are in the process of being swapped out and attempts to
33 compress them into a dynamically allocated RAM-based memory pool.
34 This can result in a significant I/O reduction on swap device and,
35 in the case where decompressing from RAM is faster than swap device
36 reads, can also improve workload performance.
38 config ZSWAP_DEFAULT_ON
39 bool "Enable the compressed cache for swap pages by default"
42 If selected, the compressed cache for swap pages will be enabled
43 at boot, otherwise it will be disabled.
45 The selection made here can be overridden by using the kernel
46 command line 'zswap.enabled=' option.
48 config ZSWAP_SHRINKER_DEFAULT_ON
49 bool "Shrink the zswap pool on memory pressure"
53 If selected, the zswap shrinker will be enabled, and the pages
54 stored in the zswap pool will become available for reclaim (i.e
55 written back to the backing swap device) on memory pressure.
57 This means that zswap writeback could happen even if the pool is
58 not yet full, or the cgroup zswap limit has not been reached,
59 reducing the chance that cold pages will reside in the zswap pool
60 and consume memory indefinitely.
63 prompt "Default compressor"
65 default ZSWAP_COMPRESSOR_DEFAULT_LZO
67 Selects the default compression algorithm for the compressed cache
70 For an overview what kind of performance can be expected from
71 a particular compression algorithm please refer to the benchmarks
72 available at the following LWN page:
73 https://lwn.net/Articles/751795/
75 If in doubt, select 'LZO'.
77 The selection made here can be overridden by using the kernel
78 command line 'zswap.compressor=' option.
80 config ZSWAP_COMPRESSOR_DEFAULT_DEFLATE
84 Use the Deflate algorithm as the default compression algorithm.
86 config ZSWAP_COMPRESSOR_DEFAULT_LZO
90 Use the LZO algorithm as the default compression algorithm.
92 config ZSWAP_COMPRESSOR_DEFAULT_842
96 Use the 842 algorithm as the default compression algorithm.
98 config ZSWAP_COMPRESSOR_DEFAULT_LZ4
102 Use the LZ4 algorithm as the default compression algorithm.
104 config ZSWAP_COMPRESSOR_DEFAULT_LZ4HC
108 Use the LZ4HC algorithm as the default compression algorithm.
110 config ZSWAP_COMPRESSOR_DEFAULT_ZSTD
114 Use the zstd algorithm as the default compression algorithm.
117 config ZSWAP_COMPRESSOR_DEFAULT
120 default "deflate" if ZSWAP_COMPRESSOR_DEFAULT_DEFLATE
121 default "lzo" if ZSWAP_COMPRESSOR_DEFAULT_LZO
122 default "842" if ZSWAP_COMPRESSOR_DEFAULT_842
123 default "lz4" if ZSWAP_COMPRESSOR_DEFAULT_LZ4
124 default "lz4hc" if ZSWAP_COMPRESSOR_DEFAULT_LZ4HC
125 default "zstd" if ZSWAP_COMPRESSOR_DEFAULT_ZSTD
129 prompt "Default allocator"
131 default ZSWAP_ZPOOL_DEFAULT_ZSMALLOC if MMU
132 default ZSWAP_ZPOOL_DEFAULT_ZBUD
134 Selects the default allocator for the compressed cache for
136 The default is 'zbud' for compatibility, however please do
137 read the description of each of the allocators below before
138 making a right choice.
140 The selection made here can be overridden by using the kernel
141 command line 'zswap.zpool=' option.
143 config ZSWAP_ZPOOL_DEFAULT_ZBUD
147 Use the zbud allocator as the default allocator.
149 config ZSWAP_ZPOOL_DEFAULT_Z3FOLD
153 Use the z3fold allocator as the default allocator.
155 config ZSWAP_ZPOOL_DEFAULT_ZSMALLOC
159 Use the zsmalloc allocator as the default allocator.
162 config ZSWAP_ZPOOL_DEFAULT
165 default "zbud" if ZSWAP_ZPOOL_DEFAULT_ZBUD
166 default "z3fold" if ZSWAP_ZPOOL_DEFAULT_Z3FOLD
167 default "zsmalloc" if ZSWAP_ZPOOL_DEFAULT_ZSMALLOC
171 tristate "2:1 compression allocator (zbud)"
174 A special purpose allocator for storing compressed pages.
175 It is designed to store up to two compressed pages per physical
176 page. While this design limits storage density, it has simple and
177 deterministic reclaim properties that make it preferable to a higher
178 density approach when reclaim will be used.
181 tristate "3:1 compression allocator (z3fold)"
184 A special purpose allocator for storing compressed pages.
185 It is designed to store up to three compressed pages per physical
186 page. It is a ZBUD derivative so the simplicity and determinism are
191 prompt "N:1 compression allocator (zsmalloc)" if ZSWAP
194 zsmalloc is a slab-based memory allocator designed to store
195 pages of various compression levels efficiently. It achieves
196 the highest storage density with the least amount of fragmentation.
199 bool "Export zsmalloc statistics"
203 This option enables code in the zsmalloc to collect various
204 statistics about what's happening in zsmalloc and exports that
205 information to userspace via debugfs.
208 config ZSMALLOC_CHAIN_SIZE
209 int "Maximum number of physical pages per-zspage"
214 This option sets the upper limit on the number of physical pages
215 that a zmalloc page (zspage) can consist of. The optimal zspage
216 chain size is calculated for each size class during the
217 initialization of the pool.
219 Changing this option can alter the characteristics of size classes,
220 such as the number of pages per zspage and the number of objects
221 per zspage. This can also result in different configurations of
222 the pool, as zsmalloc merges size classes with similar
225 For more information, see zsmalloc documentation.
227 menu "Slab allocator options"
233 bool "Configure for minimal memory footprint"
235 select SLAB_MERGE_DEFAULT
237 Configures the slab allocator in a way to achieve minimal memory
238 footprint, sacrificing scalability, debugging and other features.
239 This is intended only for the smallest system that had used the
240 SLOB allocator and is not recommended for systems with more than
245 config SLAB_MERGE_DEFAULT
246 bool "Allow slab caches to be merged"
249 For reduced kernel memory fragmentation, slab caches can be
250 merged when they share the same size and other characteristics.
251 This carries a risk of kernel heap overflows being able to
252 overwrite objects from merged caches (and more easily control
253 cache layout), which makes such heap attacks easier to exploit
254 by attackers. By keeping caches unmerged, these kinds of exploits
255 can usually only damage objects in the same cache. To disable
256 merging at runtime, "slab_nomerge" can be passed on the kernel
259 config SLAB_FREELIST_RANDOM
260 bool "Randomize slab freelist"
261 depends on !SLUB_TINY
263 Randomizes the freelist order used on creating new pages. This
264 security feature reduces the predictability of the kernel slab
265 allocator against heap overflows.
267 config SLAB_FREELIST_HARDENED
268 bool "Harden slab freelist metadata"
269 depends on !SLUB_TINY
271 Many kernel heap attacks try to target slab cache metadata and
272 other infrastructure. This options makes minor performance
273 sacrifices to harden the kernel slab allocator against common
274 freelist exploit methods.
278 bool "Enable performance statistics"
279 depends on SYSFS && !SLUB_TINY
281 The statistics are useful to debug slab allocation behavior in
282 order find ways to optimize the allocator. This should never be
283 enabled for production use since keeping statistics slows down
284 the allocator by a few percentage points. The slabinfo command
285 supports the determination of the most active slabs to figure
286 out which slabs are relevant to a particular load.
287 Try running: slabinfo -DA
289 config SLUB_CPU_PARTIAL
291 depends on SMP && !SLUB_TINY
292 bool "Enable per cpu partial caches"
294 Per cpu partial caches accelerate objects allocation and freeing
295 that is local to a processor at the price of more indeterminism
296 in the latency of the free. On overflow these caches will be cleared
297 which requires the taking of locks that may cause latency spikes.
298 Typically one would choose no for a realtime system.
300 config RANDOM_KMALLOC_CACHES
302 depends on !SLUB_TINY
303 bool "Randomize slab caches for normal kmalloc"
305 A hardening feature that creates multiple copies of slab caches for
306 normal kmalloc allocation and makes kmalloc randomly pick one based
307 on code address, which makes the attackers more difficult to spray
308 vulnerable memory objects on the heap for the purpose of exploiting
309 memory vulnerabilities.
311 Currently the number of copies is set to 16, a reasonably large value
312 that effectively diverges the memory objects allocated for different
313 subsystems or modules into different caches, at the expense of a
314 limited degree of memory and CPU overhead that relates to hardware and
317 endmenu # Slab allocator options
319 config SHUFFLE_PAGE_ALLOCATOR
320 bool "Page allocator randomization"
321 default SLAB_FREELIST_RANDOM && ACPI_NUMA
323 Randomization of the page allocator improves the average
324 utilization of a direct-mapped memory-side-cache. See section
325 5.2.27 Heterogeneous Memory Attribute Table (HMAT) in the ACPI
326 6.2a specification for an example of how a platform advertises
327 the presence of a memory-side-cache. There are also incidental
328 security benefits as it reduces the predictability of page
329 allocations to compliment SLAB_FREELIST_RANDOM, but the
330 default granularity of shuffling on the MAX_PAGE_ORDER i.e, 10th
331 order of pages is selected based on cache utilization benefits
334 While the randomization improves cache utilization it may
335 negatively impact workloads on platforms without a cache. For
336 this reason, by default, the randomization is enabled only
337 after runtime detection of a direct-mapped memory-side-cache.
338 Otherwise, the randomization may be force enabled with the
339 'page_alloc.shuffle' kernel command line parameter.
344 bool "Disable heap randomization"
347 Randomizing heap placement makes heap exploits harder, but it
348 also breaks ancient binaries (including anything libc5 based).
349 This option changes the bootup default to heap randomization
350 disabled, and can be overridden at runtime by setting
351 /proc/sys/kernel/randomize_va_space to 2.
353 On non-ancient distros (post-2000 ones) N is usually a safe choice.
355 config MMAP_ALLOW_UNINITIALIZED
356 bool "Allow mmapped anonymous memory to be uninitialized"
357 depends on EXPERT && !MMU
360 Normally, and according to the Linux spec, anonymous memory obtained
361 from mmap() has its contents cleared before it is passed to
362 userspace. Enabling this config option allows you to request that
363 mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus
364 providing a huge performance boost. If this option is not enabled,
365 then the flag will be ignored.
367 This is taken advantage of by uClibc's malloc(), and also by
368 ELF-FDPIC binfmt's brk and stack allocator.
370 Because of the obvious security issues, this option should only be
371 enabled on embedded devices where you control what is run in
372 userspace. Since that isn't generally a problem on no-MMU systems,
373 it is normally safe to say Y here.
375 See Documentation/admin-guide/mm/nommu-mmap.rst for more information.
377 config SELECT_MEMORY_MODEL
379 depends on ARCH_SELECT_MEMORY_MODEL
382 prompt "Memory model"
383 depends on SELECT_MEMORY_MODEL
384 default SPARSEMEM_MANUAL if ARCH_SPARSEMEM_DEFAULT
385 default FLATMEM_MANUAL
387 This option allows you to change some of the ways that
388 Linux manages its memory internally. Most users will
389 only have one option here selected by the architecture
390 configuration. This is normal.
392 config FLATMEM_MANUAL
394 depends on !ARCH_SPARSEMEM_ENABLE || ARCH_FLATMEM_ENABLE
396 This option is best suited for non-NUMA systems with
397 flat address space. The FLATMEM is the most efficient
398 system in terms of performance and resource consumption
399 and it is the best option for smaller systems.
401 For systems that have holes in their physical address
402 spaces and for features like NUMA and memory hotplug,
403 choose "Sparse Memory".
405 If unsure, choose this option (Flat Memory) over any other.
407 config SPARSEMEM_MANUAL
409 depends on ARCH_SPARSEMEM_ENABLE
411 This will be the only option for some systems, including
412 memory hot-plug systems. This is normal.
414 This option provides efficient support for systems with
415 holes is their physical address space and allows memory
416 hot-plug and hot-remove.
418 If unsure, choose "Flat Memory" over this option.
424 depends on (!SELECT_MEMORY_MODEL && ARCH_SPARSEMEM_ENABLE) || SPARSEMEM_MANUAL
428 depends on !SPARSEMEM || FLATMEM_MANUAL
431 # SPARSEMEM_EXTREME (which is the default) does some bootmem
432 # allocations when sparse_init() is called. If this cannot
433 # be done on your architecture, select this option. However,
434 # statically allocating the mem_section[] array can potentially
435 # consume vast quantities of .bss, so be careful.
437 # This option will also potentially produce smaller runtime code
438 # with gcc 3.4 and later.
440 config SPARSEMEM_STATIC
444 # Architecture platforms which require a two level mem_section in SPARSEMEM
445 # must select this option. This is usually for architecture platforms with
446 # an extremely sparse physical address space.
448 config SPARSEMEM_EXTREME
450 depends on SPARSEMEM && !SPARSEMEM_STATIC
452 config SPARSEMEM_VMEMMAP_ENABLE
455 config SPARSEMEM_VMEMMAP
456 bool "Sparse Memory virtual memmap"
457 depends on SPARSEMEM && SPARSEMEM_VMEMMAP_ENABLE
460 SPARSEMEM_VMEMMAP uses a virtually mapped memmap to optimise
461 pfn_to_page and page_to_pfn operations. This is the most
462 efficient option when sufficient kernel resources are available.
464 # Select this config option from the architecture Kconfig, if it is preferred
465 # to enable the feature of HugeTLB/dev_dax vmemmap optimization.
467 config ARCH_WANT_OPTIMIZE_DAX_VMEMMAP
470 config ARCH_WANT_OPTIMIZE_HUGETLB_VMEMMAP
473 config HAVE_MEMBLOCK_PHYS_MAP
480 # Don't discard allocated memory used to track "memory" and "reserved" memblocks
481 # after early boot, so it can still be used to test for validity of memory.
482 # Also, memblocks are updated with memory hot(un)plug.
483 config ARCH_KEEP_MEMBLOCK
486 # Keep arch NUMA mapping infrastructure post-init.
487 config NUMA_KEEP_MEMINFO
490 config MEMORY_ISOLATION
493 # IORESOURCE_SYSTEM_RAM regions in the kernel resource tree that are marked
494 # IORESOURCE_EXCLUSIVE cannot be mapped to user space, for example, via
496 config EXCLUSIVE_SYSTEM_RAM
498 depends on !DEVMEM || STRICT_DEVMEM
501 # Only be set on architectures that have completely implemented memory hotplug
502 # feature. If you are not sure, don't touch it.
504 config HAVE_BOOTMEM_INFO_NODE
507 config ARCH_ENABLE_MEMORY_HOTPLUG
510 config ARCH_ENABLE_MEMORY_HOTREMOVE
513 # eventually, we can have this option just 'select SPARSEMEM'
514 menuconfig MEMORY_HOTPLUG
515 bool "Memory hotplug"
516 select MEMORY_ISOLATION
518 depends on ARCH_ENABLE_MEMORY_HOTPLUG
520 select NUMA_KEEP_MEMINFO if NUMA
524 config MEMORY_HOTPLUG_DEFAULT_ONLINE
525 bool "Online the newly added memory blocks by default"
526 depends on MEMORY_HOTPLUG
528 This option sets the default policy setting for memory hotplug
529 onlining policy (/sys/devices/system/memory/auto_online_blocks) which
530 determines what happens to newly added memory regions. Policy setting
531 can always be changed at runtime.
532 See Documentation/admin-guide/mm/memory-hotplug.rst for more information.
534 Say Y here if you want all hot-plugged memory blocks to appear in
535 'online' state by default.
536 Say N here if you want the default policy to keep all hot-plugged
537 memory blocks in 'offline' state.
539 config MEMORY_HOTREMOVE
540 bool "Allow for memory hot remove"
541 select HAVE_BOOTMEM_INFO_NODE if (X86_64 || PPC64)
542 depends on MEMORY_HOTPLUG && ARCH_ENABLE_MEMORY_HOTREMOVE
545 config MHP_MEMMAP_ON_MEMORY
547 depends on MEMORY_HOTPLUG && SPARSEMEM_VMEMMAP
548 depends on ARCH_MHP_MEMMAP_ON_MEMORY_ENABLE
550 endif # MEMORY_HOTPLUG
552 config ARCH_MHP_MEMMAP_ON_MEMORY_ENABLE
555 # Heavily threaded applications may benefit from splitting the mm-wide
556 # page_table_lock, so that faults on different parts of the user address
557 # space can be handled with less contention: split it at this NR_CPUS.
558 # Default to 4 for wider testing, though 8 might be more appropriate.
559 # ARM's adjust_pte (unused if VIPT) depends on mm-wide page_table_lock.
560 # PA-RISC 7xxx's spinlock_t would enlarge struct page from 32 to 44 bytes.
561 # SPARC32 allocates multiple pte tables within a single page, and therefore
562 # a per-page lock leads to problems when multiple tables need to be locked
563 # at the same time (e.g. copy_page_range()).
564 # DEBUG_SPINLOCK and DEBUG_LOCK_ALLOC spinlock_t also enlarge struct page.
566 config SPLIT_PTLOCK_CPUS
568 default "999999" if !MMU
569 default "999999" if ARM && !CPU_CACHE_VIPT
570 default "999999" if PARISC && !PA20
571 default "999999" if SPARC32
574 config ARCH_ENABLE_SPLIT_PMD_PTLOCK
578 # support for memory balloon
579 config MEMORY_BALLOON
583 # support for memory balloon compaction
584 config BALLOON_COMPACTION
585 bool "Allow for balloon memory compaction/migration"
587 depends on COMPACTION && MEMORY_BALLOON
589 Memory fragmentation introduced by ballooning might reduce
590 significantly the number of 2MB contiguous memory blocks that can be
591 used within a guest, thus imposing performance penalties associated
592 with the reduced number of transparent huge pages that could be used
593 by the guest workload. Allowing the compaction & migration for memory
594 pages enlisted as being part of memory balloon devices avoids the
595 scenario aforementioned and helps improving memory defragmentation.
598 # support for memory compaction
600 bool "Allow for memory compaction"
605 Compaction is the only memory management component to form
606 high order (larger physically contiguous) memory blocks
607 reliably. The page allocator relies on compaction heavily and
608 the lack of the feature can lead to unexpected OOM killer
609 invocations for high order memory requests. You shouldn't
610 disable this option unless there really is a strong reason for
611 it and then we would be really interested to hear about that at
614 config COMPACT_UNEVICTABLE_DEFAULT
616 depends on COMPACTION
617 default 0 if PREEMPT_RT
621 # support for free page reporting
622 config PAGE_REPORTING
623 bool "Free page reporting"
625 Free page reporting allows for the incremental acquisition of
626 free pages from the buddy allocator for the purpose of reporting
627 those pages to another entity, such as a hypervisor, so that the
628 memory can be freed within the host for other uses.
631 # support for page migration
634 bool "Page migration"
636 depends on (NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION || CMA) && MMU
638 Allows the migration of the physical location of pages of processes
639 while the virtual addresses are not changed. This is useful in
640 two situations. The first is on NUMA systems to put pages nearer
641 to the processors accessing. The second is when allocating huge
642 pages as migration can relocate pages to satisfy a huge page
643 allocation instead of reclaiming.
645 config DEVICE_MIGRATION
646 def_bool MIGRATION && ZONE_DEVICE
648 config ARCH_ENABLE_HUGEPAGE_MIGRATION
651 config ARCH_ENABLE_THP_MIGRATION
654 config HUGETLB_PAGE_SIZE_VARIABLE
657 Allows the pageblock_order value to be dynamic instead of just standard
658 HUGETLB_PAGE_ORDER when there are multiple HugeTLB page sizes available
661 Note that the pageblock_order cannot exceed MAX_PAGE_ORDER and will be
662 clamped down to MAX_PAGE_ORDER.
665 def_bool (MEMORY_ISOLATION && COMPACTION) || CMA
667 config PCP_BATCH_SCALE_MAX
668 int "Maximum scale factor of PCP (Per-CPU pageset) batch allocate/free"
672 In page allocator, PCP (Per-CPU pageset) is refilled and drained in
673 batches. The batch number is scaled automatically to improve page
674 allocation/free throughput. But too large scale factor may hurt
675 latency. This option sets the upper limit of scale factor to limit
678 config PHYS_ADDR_T_64BIT
682 bool "Enable bounce buffers"
684 depends on BLOCK && MMU && HIGHMEM
686 Enable bounce buffers for devices that cannot access the full range of
687 memory available to the CPU. Enabled by default when HIGHMEM is
688 selected, but you may say n to override this.
695 bool "Enable KSM for page merging"
699 Enable Kernel Samepage Merging: KSM periodically scans those areas
700 of an application's address space that an app has advised may be
701 mergeable. When it finds pages of identical content, it replaces
702 the many instances by a single page with that content, so
703 saving memory until one or another app needs to modify the content.
704 Recommended for use with KVM, or with other duplicative applications.
705 See Documentation/mm/ksm.rst for more information: KSM is inactive
706 until a program has madvised that an area is MADV_MERGEABLE, and
707 root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set).
709 config DEFAULT_MMAP_MIN_ADDR
710 int "Low address space to protect from user allocation"
714 This is the portion of low virtual memory which should be protected
715 from userspace allocation. Keeping a user from writing to low pages
716 can help reduce the impact of kernel NULL pointer bugs.
718 For most ppc64 and x86 users with lots of address space
719 a value of 65536 is reasonable and should cause no problems.
720 On arm and other archs it should not be higher than 32768.
721 Programs which use vm86 functionality or have some need to map
722 this low address space will need CAP_SYS_RAWIO or disable this
723 protection by setting the value to 0.
725 This value can be changed after boot using the
726 /proc/sys/vm/mmap_min_addr tunable.
728 config ARCH_SUPPORTS_MEMORY_FAILURE
731 config MEMORY_FAILURE
733 depends on ARCH_SUPPORTS_MEMORY_FAILURE
734 bool "Enable recovery from hardware memory errors"
735 select MEMORY_ISOLATION
738 Enables code to recover from some memory failures on systems
739 with MCA recovery. This allows a system to continue running
740 even when some of its memory has uncorrected errors. This requires
741 special hardware support and typically ECC memory.
743 config HWPOISON_INJECT
744 tristate "HWPoison pages injector"
745 depends on MEMORY_FAILURE && DEBUG_KERNEL && PROC_FS
746 select PROC_PAGE_MONITOR
748 config NOMMU_INITIAL_TRIM_EXCESS
749 int "Turn on mmap() excess space trimming before booting"
753 The NOMMU mmap() frequently needs to allocate large contiguous chunks
754 of memory on which to store mappings, but it can only ask the system
755 allocator for chunks in 2^N*PAGE_SIZE amounts - which is frequently
756 more than it requires. To deal with this, mmap() is able to trim off
757 the excess and return it to the allocator.
759 If trimming is enabled, the excess is trimmed off and returned to the
760 system allocator, which can cause extra fragmentation, particularly
761 if there are a lot of transient processes.
763 If trimming is disabled, the excess is kept, but not used, which for
764 long-term mappings means that the space is wasted.
766 Trimming can be dynamically controlled through a sysctl option
767 (/proc/sys/vm/nr_trim_pages) which specifies the minimum number of
768 excess pages there must be before trimming should occur, or zero if
769 no trimming is to occur.
771 This option specifies the initial value of this option. The default
772 of 1 says that all excess pages should be trimmed.
774 See Documentation/admin-guide/mm/nommu-mmap.rst for more information.
776 config ARCH_WANT_GENERAL_HUGETLB
779 config ARCH_WANTS_THP_SWAP
782 menuconfig TRANSPARENT_HUGEPAGE
783 bool "Transparent Hugepage Support"
784 depends on HAVE_ARCH_TRANSPARENT_HUGEPAGE && !PREEMPT_RT
788 Transparent Hugepages allows the kernel to use huge pages and
789 huge tlb transparently to the applications whenever possible.
790 This feature can improve computing performance to certain
791 applications by speeding up page faults during memory
792 allocation, by reducing the number of tlb misses and by speeding
793 up the pagetable walking.
795 If memory constrained on embedded, you may want to say N.
797 if TRANSPARENT_HUGEPAGE
800 prompt "Transparent Hugepage Support sysfs defaults"
801 depends on TRANSPARENT_HUGEPAGE
802 default TRANSPARENT_HUGEPAGE_ALWAYS
804 Selects the sysfs defaults for Transparent Hugepage Support.
806 config TRANSPARENT_HUGEPAGE_ALWAYS
809 Enabling Transparent Hugepage always, can increase the
810 memory footprint of applications without a guaranteed
811 benefit but it will work automatically for all applications.
813 config TRANSPARENT_HUGEPAGE_MADVISE
816 Enabling Transparent Hugepage madvise, will only provide a
817 performance improvement benefit to the applications using
818 madvise(MADV_HUGEPAGE) but it won't risk to increase the
819 memory footprint of applications without a guaranteed
822 config TRANSPARENT_HUGEPAGE_NEVER
825 Disable Transparent Hugepage by default. It can still be
826 enabled at runtime via sysfs.
831 depends on TRANSPARENT_HUGEPAGE && ARCH_WANTS_THP_SWAP && SWAP && 64BIT
833 Swap transparent huge pages in one piece, without splitting.
834 XXX: For now, swap cluster backing transparent huge page
835 will be split after swapout.
837 For selection by architectures with reasonable THP sizes.
839 config READ_ONLY_THP_FOR_FS
840 bool "Read-only THP for filesystems (EXPERIMENTAL)"
841 depends on TRANSPARENT_HUGEPAGE && SHMEM
844 Allow khugepaged to put read-only file-backed pages in THP.
846 This is marked experimental because it is a new feature. Write
847 support of file THPs will be developed in the next few release
850 endif # TRANSPARENT_HUGEPAGE
853 # UP and nommu archs use km based percpu allocator
855 config NEED_PER_CPU_KM
856 depends on !SMP || !MMU
860 config NEED_PER_CPU_EMBED_FIRST_CHUNK
863 config NEED_PER_CPU_PAGE_FIRST_CHUNK
866 config USE_PERCPU_NUMA_NODE_ID
869 config HAVE_SETUP_PER_CPU_AREA
873 bool "Contiguous Memory Allocator"
876 select MEMORY_ISOLATION
878 This enables the Contiguous Memory Allocator which allows other
879 subsystems to allocate big physically-contiguous blocks of memory.
880 CMA reserves a region of memory and allows only movable pages to
881 be allocated from it. This way, the kernel can use the memory for
882 pagecache and when a subsystem requests for contiguous area, the
883 allocated pages are migrated away to serve the contiguous request.
888 bool "CMA debugfs interface"
889 depends on CMA && DEBUG_FS
891 Turns on the DebugFS interface for CMA.
894 bool "CMA information through sysfs interface"
895 depends on CMA && SYSFS
897 This option exposes some sysfs attributes to get information
901 int "Maximum count of the CMA areas"
906 CMA allows to create CMA areas for particular purpose, mainly,
907 used as device private area. This parameter sets the maximum
908 number of CMA area in the system.
910 If unsure, leave the default value "8" in UMA and "20" in NUMA.
912 config MEM_SOFT_DIRTY
913 bool "Track memory changes"
914 depends on CHECKPOINT_RESTORE && HAVE_ARCH_SOFT_DIRTY && PROC_FS
915 select PROC_PAGE_MONITOR
917 This option enables memory changes tracking by introducing a
918 soft-dirty bit on pte-s. This bit it set when someone writes
919 into a page just as regular dirty bit, but unlike the latter
920 it can be cleared by hands.
922 See Documentation/admin-guide/mm/soft-dirty.rst for more details.
924 config GENERIC_EARLY_IOREMAP
927 config STACK_MAX_DEFAULT_SIZE_MB
928 int "Default maximum user stack size for 32-bit processes (MB)"
931 depends on STACK_GROWSUP && (!64BIT || COMPAT)
933 This is the maximum stack size in Megabytes in the VM layout of 32-bit
934 user processes when the stack grows upwards (currently only on parisc
935 arch) when the RLIMIT_STACK hard limit is unlimited.
937 A sane initial value is 100 MB.
939 config DEFERRED_STRUCT_PAGE_INIT
940 bool "Defer initialisation of struct pages to kthreads"
942 depends on !NEED_PER_CPU_KM
946 Ordinarily all struct pages are initialised during early boot in a
947 single thread. On very large machines this can take a considerable
948 amount of time. If this option is set, large machines will bring up
949 a subset of memmap at boot and then initialise the rest in parallel.
950 This has a potential performance impact on tasks running early in the
951 lifetime of the system until these kthreads finish the
954 config PAGE_IDLE_FLAG
956 select PAGE_EXTENSION if !64BIT
958 This adds PG_idle and PG_young flags to 'struct page'. PTE Accessed
959 bit writers can set the state of the bit in the flags so that PTE
960 Accessed bit readers may avoid disturbance.
962 config IDLE_PAGE_TRACKING
963 bool "Enable idle page tracking"
964 depends on SYSFS && MMU
965 select PAGE_IDLE_FLAG
967 This feature allows to estimate the amount of user pages that have
968 not been touched during a given period of time. This information can
969 be useful to tune memory cgroup limits and/or for job placement
970 within a compute cluster.
972 See Documentation/admin-guide/mm/idle_page_tracking.rst for
975 # Architectures which implement cpu_dcache_is_aliasing() to query
976 # whether the data caches are aliased (VIVT or VIPT with dcache
977 # aliasing) need to select this.
978 config ARCH_HAS_CPU_CACHE_ALIASING
981 config ARCH_HAS_CACHE_LINE_SIZE
984 config ARCH_HAS_CURRENT_STACK_POINTER
987 In support of HARDENED_USERCOPY performing stack variable lifetime
988 checking, an architecture-agnostic way to find the stack pointer
989 is needed. Once an architecture defines an unsigned long global
990 register alias named "current_stack_pointer", this config can be
993 config ARCH_HAS_PTE_DEVMAP
996 config ARCH_HAS_ZONE_DMA_SET
1000 bool "Support DMA zone" if ARCH_HAS_ZONE_DMA_SET
1001 default y if ARM64 || X86
1004 bool "Support DMA32 zone" if ARCH_HAS_ZONE_DMA_SET
1009 bool "Device memory (pmem, HMM, etc...) hotplug support"
1010 depends on MEMORY_HOTPLUG
1011 depends on MEMORY_HOTREMOVE
1012 depends on SPARSEMEM_VMEMMAP
1013 depends on ARCH_HAS_PTE_DEVMAP
1017 Device memory hotplug support allows for establishing pmem,
1018 or other device driver discovered memory regions, in the
1019 memmap. This allows pfn_to_page() lookups of otherwise
1020 "device-physical" addresses which is needed for using a DAX
1021 mapping in an O_DIRECT operation, among other things.
1023 If FS_DAX is enabled, then say Y.
1026 # Helpers to mirror range of the CPU page tables of a process into device page
1033 config GET_FREE_REGION
1034 depends on SPARSEMEM
1037 config DEVICE_PRIVATE
1038 bool "Unaddressable device memory (GPU memory, ...)"
1039 depends on ZONE_DEVICE
1040 select GET_FREE_REGION
1043 Allows creation of struct pages to represent unaddressable device
1044 memory; i.e., memory that is only accessible from the device (or
1045 group of devices). You likely also want to select HMM_MIRROR.
1050 config ARCH_USES_HIGH_VMA_FLAGS
1052 config ARCH_HAS_PKEYS
1055 config ARCH_USES_PG_ARCH_X
1058 Enable the definition of PG_arch_x page flags with x > 1. Only
1059 suitable for 64-bit architectures with CONFIG_FLATMEM or
1060 CONFIG_SPARSEMEM_VMEMMAP enabled, otherwise there may not be
1061 enough room for additional bits in page->flags.
1063 config VM_EVENT_COUNTERS
1065 bool "Enable VM event counters for /proc/vmstat" if EXPERT
1067 VM event counters are needed for event counts to be shown.
1068 This option allows the disabling of the VM event counters
1069 on EXPERT systems. /proc/vmstat will only show page counts
1070 if VM event counters are disabled.
1073 bool "Collect percpu memory statistics"
1075 This feature collects and exposes statistics via debugfs. The
1076 information includes global and per chunk statistics, which can
1077 be used to help understand percpu memory usage.
1080 bool "Enable infrastructure for get_user_pages()-related unit tests"
1083 Provides /sys/kernel/debug/gup_test, which in turn provides a way
1084 to make ioctl calls that can launch kernel-based unit tests for
1085 the get_user_pages*() and pin_user_pages*() family of API calls.
1087 These tests include benchmark testing of the _fast variants of
1088 get_user_pages*() and pin_user_pages*(), as well as smoke tests of
1089 the non-_fast variants.
1091 There is also a sub-test that allows running dump_page() on any
1092 of up to eight pages (selected by command line args) within the
1093 range of user-space addresses. These pages are either pinned via
1094 pin_user_pages*(), or pinned via get_user_pages*(), as specified
1095 by other command line arguments.
1097 See tools/testing/selftests/mm/gup_test.c
1099 comment "GUP_TEST needs to have DEBUG_FS enabled"
1100 depends on !GUP_TEST && !DEBUG_FS
1102 config GUP_GET_PXX_LOW_HIGH
1106 tristate "Enable a module to run time tests on dma_pool"
1109 Provides a test module that will allocate and free many blocks of
1110 various sizes and report how long it takes. This is intended to
1111 provide a consistent way to measure how changes to the
1112 dma_pool_alloc/free routines affect performance.
1114 config ARCH_HAS_PTE_SPECIAL
1118 # Some architectures require a special hugepage directory format that is
1119 # required to support multiple hugepage sizes. For example a4fe3ce76
1120 # "powerpc/mm: Allow more flexible layouts for hugepage pagetables"
1121 # introduced it on powerpc. This allows for a more flexible hugepage
1122 # pagetable layouts.
1124 config ARCH_HAS_HUGEPD
1127 config MAPPING_DIRTY_HELPERS
1133 config KMAP_LOCAL_NON_LINEAR_PTE_ARRAY
1136 # struct io_mapping based helper. Selected by drivers that need them
1141 bool "Enable memfd_create() system call" if EXPERT
1145 bool "Enable memfd_secret() system call" if EXPERT
1146 depends on ARCH_HAS_SET_DIRECT_MAP
1148 Enable the memfd_secret() system call with the ability to create
1149 memory areas visible only in the context of the owning process and
1150 not mapped to other processes and other kernel page tables.
1152 config ANON_VMA_NAME
1153 bool "Anonymous VMA name support"
1154 depends on PROC_FS && ADVISE_SYSCALLS && MMU
1157 Allow naming anonymous virtual memory areas.
1159 This feature allows assigning names to virtual memory areas. Assigned
1160 names can be later retrieved from /proc/pid/maps and /proc/pid/smaps
1161 and help identifying individual anonymous memory areas.
1162 Assigning a name to anonymous virtual memory area might prevent that
1163 area from being merged with adjacent virtual memory areas due to the
1164 difference in their name.
1166 config HAVE_ARCH_USERFAULTFD_WP
1169 Arch has userfaultfd write protection support
1171 config HAVE_ARCH_USERFAULTFD_MINOR
1174 Arch has userfaultfd minor fault support
1176 menuconfig USERFAULTFD
1177 bool "Enable userfaultfd() system call"
1180 Enable the userfaultfd() system call that allows to intercept and
1181 handle page faults in userland.
1184 config PTE_MARKER_UFFD_WP
1185 bool "Userfaultfd write protection support for shmem/hugetlbfs"
1187 depends on HAVE_ARCH_USERFAULTFD_WP
1190 Allows to create marker PTEs for userfaultfd write protection
1191 purposes. It is required to enable userfaultfd write protection on
1192 file-backed memory types like shmem and hugetlbfs.
1197 bool "Multi-Gen LRU"
1199 # make sure folio->flags has enough spare bits
1200 depends on 64BIT || !SPARSEMEM || SPARSEMEM_VMEMMAP
1202 A high performance LRU implementation to overcommit memory. See
1203 Documentation/admin-guide/mm/multigen_lru.rst for details.
1205 config LRU_GEN_ENABLED
1206 bool "Enable by default"
1209 This option enables the multi-gen LRU by default.
1211 config LRU_GEN_STATS
1212 bool "Full stats for debugging"
1215 Do not enable this option unless you plan to look at historical stats
1216 from evicted generations for debugging purpose.
1218 This option has a per-memcg and per-node memory overhead.
1220 config LRU_GEN_WALKS_MMU
1222 depends on LRU_GEN && ARCH_HAS_HW_PTE_YOUNG
1225 config ARCH_SUPPORTS_PER_VMA_LOCK
1230 depends on ARCH_SUPPORTS_PER_VMA_LOCK && MMU && SMP
1232 Allow per-vma locking during page fault handling.
1234 This feature allows locking each virtual memory area separately when
1235 handling page faults instead of taking mmap_lock.
1237 config LOCK_MM_AND_FIND_VMA
1239 depends on !STACK_GROWSUP
1241 config IOMMU_MM_DATA
1244 source "mm/damon/Kconfig"