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"
32 A lightweight compressed cache for swap pages. It takes
33 pages that are in the process of being swapped out and attempts to
34 compress them into a dynamically allocated RAM-based memory pool.
35 This can result in a significant I/O reduction on swap device and,
36 in the case where decompressing from RAM is faster than swap device
37 reads, can also improve workload performance.
39 config ZSWAP_DEFAULT_ON
40 bool "Enable the compressed cache for swap pages by default"
43 If selected, the compressed cache for swap pages will be enabled
44 at boot, otherwise it will be disabled.
46 The selection made here can be overridden by using the kernel
47 command line 'zswap.enabled=' option.
50 prompt "Default compressor"
52 default ZSWAP_COMPRESSOR_DEFAULT_LZO
54 Selects the default compression algorithm for the compressed cache
57 For an overview what kind of performance can be expected from
58 a particular compression algorithm please refer to the benchmarks
59 available at the following LWN page:
60 https://lwn.net/Articles/751795/
62 If in doubt, select 'LZO'.
64 The selection made here can be overridden by using the kernel
65 command line 'zswap.compressor=' option.
67 config ZSWAP_COMPRESSOR_DEFAULT_DEFLATE
71 Use the Deflate algorithm as the default compression algorithm.
73 config ZSWAP_COMPRESSOR_DEFAULT_LZO
77 Use the LZO algorithm as the default compression algorithm.
79 config ZSWAP_COMPRESSOR_DEFAULT_842
83 Use the 842 algorithm as the default compression algorithm.
85 config ZSWAP_COMPRESSOR_DEFAULT_LZ4
89 Use the LZ4 algorithm as the default compression algorithm.
91 config ZSWAP_COMPRESSOR_DEFAULT_LZ4HC
95 Use the LZ4HC algorithm as the default compression algorithm.
97 config ZSWAP_COMPRESSOR_DEFAULT_ZSTD
101 Use the zstd algorithm as the default compression algorithm.
104 config ZSWAP_COMPRESSOR_DEFAULT
107 default "deflate" if ZSWAP_COMPRESSOR_DEFAULT_DEFLATE
108 default "lzo" if ZSWAP_COMPRESSOR_DEFAULT_LZO
109 default "842" if ZSWAP_COMPRESSOR_DEFAULT_842
110 default "lz4" if ZSWAP_COMPRESSOR_DEFAULT_LZ4
111 default "lz4hc" if ZSWAP_COMPRESSOR_DEFAULT_LZ4HC
112 default "zstd" if ZSWAP_COMPRESSOR_DEFAULT_ZSTD
116 prompt "Default allocator"
118 default ZSWAP_ZPOOL_DEFAULT_ZBUD
120 Selects the default allocator for the compressed cache for
122 The default is 'zbud' for compatibility, however please do
123 read the description of each of the allocators below before
124 making a right choice.
126 The selection made here can be overridden by using the kernel
127 command line 'zswap.zpool=' option.
129 config ZSWAP_ZPOOL_DEFAULT_ZBUD
133 Use the zbud allocator as the default allocator.
135 config ZSWAP_ZPOOL_DEFAULT_Z3FOLD
139 Use the z3fold allocator as the default allocator.
141 config ZSWAP_ZPOOL_DEFAULT_ZSMALLOC
145 Use the zsmalloc allocator as the default allocator.
148 config ZSWAP_ZPOOL_DEFAULT
151 default "zbud" if ZSWAP_ZPOOL_DEFAULT_ZBUD
152 default "z3fold" if ZSWAP_ZPOOL_DEFAULT_Z3FOLD
153 default "zsmalloc" if ZSWAP_ZPOOL_DEFAULT_ZSMALLOC
157 tristate "2:1 compression allocator (zbud)"
160 A special purpose allocator for storing compressed pages.
161 It is designed to store up to two compressed pages per physical
162 page. While this design limits storage density, it has simple and
163 deterministic reclaim properties that make it preferable to a higher
164 density approach when reclaim will be used.
167 tristate "3:1 compression allocator (z3fold)"
170 A special purpose allocator for storing compressed pages.
171 It is designed to store up to three compressed pages per physical
172 page. It is a ZBUD derivative so the simplicity and determinism are
177 prompt "N:1 compression allocator (zsmalloc)" if ZSWAP
180 zsmalloc is a slab-based memory allocator designed to store
181 pages of various compression levels efficiently. It achieves
182 the highest storage density with the least amount of fragmentation.
185 bool "Export zsmalloc statistics"
189 This option enables code in the zsmalloc to collect various
190 statistics about what's happening in zsmalloc and exports that
191 information to userspace via debugfs.
194 menu "SLAB allocator options"
197 prompt "Choose SLAB allocator"
200 This option allows to select a slab allocator.
204 depends on !PREEMPT_RT
205 select HAVE_HARDENED_USERCOPY_ALLOCATOR
207 The regular slab allocator that is established and known to work
208 well in all environments. It organizes cache hot objects in
209 per cpu and per node queues.
212 bool "SLUB (Unqueued Allocator)"
213 select HAVE_HARDENED_USERCOPY_ALLOCATOR
215 SLUB is a slab allocator that minimizes cache line usage
216 instead of managing queues of cached objects (SLAB approach).
217 Per cpu caching is realized using slabs of objects instead
218 of queues of objects. SLUB can use memory efficiently
219 and has enhanced diagnostics. SLUB is the default choice for
222 config SLOB_DEPRECATED
224 bool "SLOB (Simple Allocator - DEPRECATED)"
225 depends on !PREEMPT_RT
227 Deprecated and scheduled for removal in a few cycles. SLUB
228 recommended as replacement. CONFIG_SLUB_TINY can be considered
229 on systems with 16MB or less RAM.
231 If you need SLOB to stay, please contact linux-mm@kvack.org and
232 people listed in the SLAB ALLOCATOR section of MAINTAINERS file,
235 SLOB replaces the stock allocator with a drastically simpler
236 allocator. SLOB is generally more space efficient but
237 does not perform as well on large systems.
244 depends on SLOB_DEPRECATED
247 bool "Configure SLUB for minimal memory footprint"
248 depends on SLUB && EXPERT
249 select SLAB_MERGE_DEFAULT
251 Configures the SLUB allocator in a way to achieve minimal memory
252 footprint, sacrificing scalability, debugging and other features.
253 This is intended only for the smallest system that had used the
254 SLOB allocator and is not recommended for systems with more than
259 config SLAB_MERGE_DEFAULT
260 bool "Allow slab caches to be merged"
262 depends on SLAB || SLUB
264 For reduced kernel memory fragmentation, slab caches can be
265 merged when they share the same size and other characteristics.
266 This carries a risk of kernel heap overflows being able to
267 overwrite objects from merged caches (and more easily control
268 cache layout), which makes such heap attacks easier to exploit
269 by attackers. By keeping caches unmerged, these kinds of exploits
270 can usually only damage objects in the same cache. To disable
271 merging at runtime, "slab_nomerge" can be passed on the kernel
274 config SLAB_FREELIST_RANDOM
275 bool "Randomize slab freelist"
276 depends on SLAB || (SLUB && !SLUB_TINY)
278 Randomizes the freelist order used on creating new pages. This
279 security feature reduces the predictability of the kernel slab
280 allocator against heap overflows.
282 config SLAB_FREELIST_HARDENED
283 bool "Harden slab freelist metadata"
284 depends on SLAB || (SLUB && !SLUB_TINY)
286 Many kernel heap attacks try to target slab cache metadata and
287 other infrastructure. This options makes minor performance
288 sacrifices to harden the kernel slab allocator against common
289 freelist exploit methods. Some slab implementations have more
290 sanity-checking than others. This option is most effective with
295 bool "Enable SLUB performance statistics"
296 depends on SLUB && SYSFS && !SLUB_TINY
298 SLUB statistics are useful to debug SLUBs allocation behavior in
299 order find ways to optimize the allocator. This should never be
300 enabled for production use since keeping statistics slows down
301 the allocator by a few percentage points. The slabinfo command
302 supports the determination of the most active slabs to figure
303 out which slabs are relevant to a particular load.
304 Try running: slabinfo -DA
306 config SLUB_CPU_PARTIAL
308 depends on SLUB && SMP && !SLUB_TINY
309 bool "SLUB per cpu partial cache"
311 Per cpu partial caches accelerate objects allocation and freeing
312 that is local to a processor at the price of more indeterminism
313 in the latency of the free. On overflow these caches will be cleared
314 which requires the taking of locks that may cause latency spikes.
315 Typically one would choose no for a realtime system.
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_ORDER - 1" i.e,
331 10th order of pages is selected based on cache utilization
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 config HAVE_MEMBLOCK_PHYS_MAP
471 # Don't discard allocated memory used to track "memory" and "reserved" memblocks
472 # after early boot, so it can still be used to test for validity of memory.
473 # Also, memblocks are updated with memory hot(un)plug.
474 config ARCH_KEEP_MEMBLOCK
477 # Keep arch NUMA mapping infrastructure post-init.
478 config NUMA_KEEP_MEMINFO
481 config MEMORY_ISOLATION
484 # IORESOURCE_SYSTEM_RAM regions in the kernel resource tree that are marked
485 # IORESOURCE_EXCLUSIVE cannot be mapped to user space, for example, via
487 config EXCLUSIVE_SYSTEM_RAM
489 depends on !DEVMEM || STRICT_DEVMEM
492 # Only be set on architectures that have completely implemented memory hotplug
493 # feature. If you are not sure, don't touch it.
495 config HAVE_BOOTMEM_INFO_NODE
498 config ARCH_ENABLE_MEMORY_HOTPLUG
501 config ARCH_ENABLE_MEMORY_HOTREMOVE
504 # eventually, we can have this option just 'select SPARSEMEM'
505 menuconfig MEMORY_HOTPLUG
506 bool "Memory hotplug"
507 select MEMORY_ISOLATION
509 depends on ARCH_ENABLE_MEMORY_HOTPLUG
511 select NUMA_KEEP_MEMINFO if NUMA
515 config MEMORY_HOTPLUG_DEFAULT_ONLINE
516 bool "Online the newly added memory blocks by default"
517 depends on MEMORY_HOTPLUG
519 This option sets the default policy setting for memory hotplug
520 onlining policy (/sys/devices/system/memory/auto_online_blocks) which
521 determines what happens to newly added memory regions. Policy setting
522 can always be changed at runtime.
523 See Documentation/admin-guide/mm/memory-hotplug.rst for more information.
525 Say Y here if you want all hot-plugged memory blocks to appear in
526 'online' state by default.
527 Say N here if you want the default policy to keep all hot-plugged
528 memory blocks in 'offline' state.
530 config MEMORY_HOTREMOVE
531 bool "Allow for memory hot remove"
532 select HAVE_BOOTMEM_INFO_NODE if (X86_64 || PPC64)
533 depends on MEMORY_HOTPLUG && ARCH_ENABLE_MEMORY_HOTREMOVE
536 config MHP_MEMMAP_ON_MEMORY
538 depends on MEMORY_HOTPLUG && SPARSEMEM_VMEMMAP
539 depends on ARCH_MHP_MEMMAP_ON_MEMORY_ENABLE
541 endif # MEMORY_HOTPLUG
543 # Heavily threaded applications may benefit from splitting the mm-wide
544 # page_table_lock, so that faults on different parts of the user address
545 # space can be handled with less contention: split it at this NR_CPUS.
546 # Default to 4 for wider testing, though 8 might be more appropriate.
547 # ARM's adjust_pte (unused if VIPT) depends on mm-wide page_table_lock.
548 # PA-RISC 7xxx's spinlock_t would enlarge struct page from 32 to 44 bytes.
549 # SPARC32 allocates multiple pte tables within a single page, and therefore
550 # a per-page lock leads to problems when multiple tables need to be locked
551 # at the same time (e.g. copy_page_range()).
552 # DEBUG_SPINLOCK and DEBUG_LOCK_ALLOC spinlock_t also enlarge struct page.
554 config SPLIT_PTLOCK_CPUS
556 default "999999" if !MMU
557 default "999999" if ARM && !CPU_CACHE_VIPT
558 default "999999" if PARISC && !PA20
559 default "999999" if SPARC32
562 config ARCH_ENABLE_SPLIT_PMD_PTLOCK
566 # support for memory balloon
567 config MEMORY_BALLOON
571 # support for memory balloon compaction
572 config BALLOON_COMPACTION
573 bool "Allow for balloon memory compaction/migration"
575 depends on COMPACTION && MEMORY_BALLOON
577 Memory fragmentation introduced by ballooning might reduce
578 significantly the number of 2MB contiguous memory blocks that can be
579 used within a guest, thus imposing performance penalties associated
580 with the reduced number of transparent huge pages that could be used
581 by the guest workload. Allowing the compaction & migration for memory
582 pages enlisted as being part of memory balloon devices avoids the
583 scenario aforementioned and helps improving memory defragmentation.
586 # support for memory compaction
588 bool "Allow for memory compaction"
593 Compaction is the only memory management component to form
594 high order (larger physically contiguous) memory blocks
595 reliably. The page allocator relies on compaction heavily and
596 the lack of the feature can lead to unexpected OOM killer
597 invocations for high order memory requests. You shouldn't
598 disable this option unless there really is a strong reason for
599 it and then we would be really interested to hear about that at
602 config COMPACT_UNEVICTABLE_DEFAULT
604 depends on COMPACTION
605 default 0 if PREEMPT_RT
609 # support for free page reporting
610 config PAGE_REPORTING
611 bool "Free page reporting"
614 Free page reporting allows for the incremental acquisition of
615 free pages from the buddy allocator for the purpose of reporting
616 those pages to another entity, such as a hypervisor, so that the
617 memory can be freed within the host for other uses.
620 # support for page migration
623 bool "Page migration"
625 depends on (NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION || CMA) && MMU
627 Allows the migration of the physical location of pages of processes
628 while the virtual addresses are not changed. This is useful in
629 two situations. The first is on NUMA systems to put pages nearer
630 to the processors accessing. The second is when allocating huge
631 pages as migration can relocate pages to satisfy a huge page
632 allocation instead of reclaiming.
634 config DEVICE_MIGRATION
635 def_bool MIGRATION && ZONE_DEVICE
637 config ARCH_ENABLE_HUGEPAGE_MIGRATION
640 config ARCH_ENABLE_THP_MIGRATION
643 config HUGETLB_PAGE_SIZE_VARIABLE
646 Allows the pageblock_order value to be dynamic instead of just standard
647 HUGETLB_PAGE_ORDER when there are multiple HugeTLB page sizes available
650 Note that the pageblock_order cannot exceed MAX_ORDER - 1 and will be
651 clamped down to MAX_ORDER - 1.
654 def_bool (MEMORY_ISOLATION && COMPACTION) || CMA
656 config PHYS_ADDR_T_64BIT
660 bool "Enable bounce buffers"
662 depends on BLOCK && MMU && HIGHMEM
664 Enable bounce buffers for devices that cannot access the full range of
665 memory available to the CPU. Enabled by default when HIGHMEM is
666 selected, but you may say n to override this.
674 bool "Enable KSM for page merging"
678 Enable Kernel Samepage Merging: KSM periodically scans those areas
679 of an application's address space that an app has advised may be
680 mergeable. When it finds pages of identical content, it replaces
681 the many instances by a single page with that content, so
682 saving memory until one or another app needs to modify the content.
683 Recommended for use with KVM, or with other duplicative applications.
684 See Documentation/mm/ksm.rst for more information: KSM is inactive
685 until a program has madvised that an area is MADV_MERGEABLE, and
686 root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set).
688 config DEFAULT_MMAP_MIN_ADDR
689 int "Low address space to protect from user allocation"
693 This is the portion of low virtual memory which should be protected
694 from userspace allocation. Keeping a user from writing to low pages
695 can help reduce the impact of kernel NULL pointer bugs.
697 For most ia64, ppc64 and x86 users with lots of address space
698 a value of 65536 is reasonable and should cause no problems.
699 On arm and other archs it should not be higher than 32768.
700 Programs which use vm86 functionality or have some need to map
701 this low address space will need CAP_SYS_RAWIO or disable this
702 protection by setting the value to 0.
704 This value can be changed after boot using the
705 /proc/sys/vm/mmap_min_addr tunable.
707 config ARCH_SUPPORTS_MEMORY_FAILURE
710 config MEMORY_FAILURE
712 depends on ARCH_SUPPORTS_MEMORY_FAILURE
713 bool "Enable recovery from hardware memory errors"
714 select MEMORY_ISOLATION
717 Enables code to recover from some memory failures on systems
718 with MCA recovery. This allows a system to continue running
719 even when some of its memory has uncorrected errors. This requires
720 special hardware support and typically ECC memory.
722 config HWPOISON_INJECT
723 tristate "HWPoison pages injector"
724 depends on MEMORY_FAILURE && DEBUG_KERNEL && PROC_FS
725 select PROC_PAGE_MONITOR
727 config NOMMU_INITIAL_TRIM_EXCESS
728 int "Turn on mmap() excess space trimming before booting"
732 The NOMMU mmap() frequently needs to allocate large contiguous chunks
733 of memory on which to store mappings, but it can only ask the system
734 allocator for chunks in 2^N*PAGE_SIZE amounts - which is frequently
735 more than it requires. To deal with this, mmap() is able to trim off
736 the excess and return it to the allocator.
738 If trimming is enabled, the excess is trimmed off and returned to the
739 system allocator, which can cause extra fragmentation, particularly
740 if there are a lot of transient processes.
742 If trimming is disabled, the excess is kept, but not used, which for
743 long-term mappings means that the space is wasted.
745 Trimming can be dynamically controlled through a sysctl option
746 (/proc/sys/vm/nr_trim_pages) which specifies the minimum number of
747 excess pages there must be before trimming should occur, or zero if
748 no trimming is to occur.
750 This option specifies the initial value of this option. The default
751 of 1 says that all excess pages should be trimmed.
753 See Documentation/admin-guide/mm/nommu-mmap.rst for more information.
755 config ARCH_WANT_GENERAL_HUGETLB
758 config ARCH_WANTS_THP_SWAP
761 menuconfig TRANSPARENT_HUGEPAGE
762 bool "Transparent Hugepage Support"
763 depends on HAVE_ARCH_TRANSPARENT_HUGEPAGE && !PREEMPT_RT
767 Transparent Hugepages allows the kernel to use huge pages and
768 huge tlb transparently to the applications whenever possible.
769 This feature can improve computing performance to certain
770 applications by speeding up page faults during memory
771 allocation, by reducing the number of tlb misses and by speeding
772 up the pagetable walking.
774 If memory constrained on embedded, you may want to say N.
776 if TRANSPARENT_HUGEPAGE
779 prompt "Transparent Hugepage Support sysfs defaults"
780 depends on TRANSPARENT_HUGEPAGE
781 default TRANSPARENT_HUGEPAGE_ALWAYS
783 Selects the sysfs defaults for Transparent Hugepage Support.
785 config TRANSPARENT_HUGEPAGE_ALWAYS
788 Enabling Transparent Hugepage always, can increase the
789 memory footprint of applications without a guaranteed
790 benefit but it will work automatically for all applications.
792 config TRANSPARENT_HUGEPAGE_MADVISE
795 Enabling Transparent Hugepage madvise, will only provide a
796 performance improvement benefit to the applications using
797 madvise(MADV_HUGEPAGE) but it won't risk to increase the
798 memory footprint of applications without a guaranteed
804 depends on TRANSPARENT_HUGEPAGE && ARCH_WANTS_THP_SWAP && SWAP && 64BIT
806 Swap transparent huge pages in one piece, without splitting.
807 XXX: For now, swap cluster backing transparent huge page
808 will be split after swapout.
810 For selection by architectures with reasonable THP sizes.
812 config READ_ONLY_THP_FOR_FS
813 bool "Read-only THP for filesystems (EXPERIMENTAL)"
814 depends on TRANSPARENT_HUGEPAGE && SHMEM
817 Allow khugepaged to put read-only file-backed pages in THP.
819 This is marked experimental because it is a new feature. Write
820 support of file THPs will be developed in the next few release
823 endif # TRANSPARENT_HUGEPAGE
826 # UP and nommu archs use km based percpu allocator
828 config NEED_PER_CPU_KM
829 depends on !SMP || !MMU
833 config NEED_PER_CPU_EMBED_FIRST_CHUNK
836 config NEED_PER_CPU_PAGE_FIRST_CHUNK
839 config USE_PERCPU_NUMA_NODE_ID
842 config HAVE_SETUP_PER_CPU_AREA
849 bool "Contiguous Memory Allocator"
852 select MEMORY_ISOLATION
854 This enables the Contiguous Memory Allocator which allows other
855 subsystems to allocate big physically-contiguous blocks of memory.
856 CMA reserves a region of memory and allows only movable pages to
857 be allocated from it. This way, the kernel can use the memory for
858 pagecache and when a subsystem requests for contiguous area, the
859 allocated pages are migrated away to serve the contiguous request.
864 bool "CMA debug messages (DEVELOPMENT)"
865 depends on DEBUG_KERNEL && CMA
867 Turns on debug messages in CMA. This produces KERN_DEBUG
868 messages for every CMA call as well as various messages while
869 processing calls such as dma_alloc_from_contiguous().
870 This option does not affect warning and error messages.
873 bool "CMA debugfs interface"
874 depends on CMA && DEBUG_FS
876 Turns on the DebugFS interface for CMA.
879 bool "CMA information through sysfs interface"
880 depends on CMA && SYSFS
882 This option exposes some sysfs attributes to get information
886 int "Maximum count of the CMA areas"
891 CMA allows to create CMA areas for particular purpose, mainly,
892 used as device private area. This parameter sets the maximum
893 number of CMA area in the system.
895 If unsure, leave the default value "7" in UMA and "19" in NUMA.
897 config MEM_SOFT_DIRTY
898 bool "Track memory changes"
899 depends on CHECKPOINT_RESTORE && HAVE_ARCH_SOFT_DIRTY && PROC_FS
900 select PROC_PAGE_MONITOR
902 This option enables memory changes tracking by introducing a
903 soft-dirty bit on pte-s. This bit it set when someone writes
904 into a page just as regular dirty bit, but unlike the latter
905 it can be cleared by hands.
907 See Documentation/admin-guide/mm/soft-dirty.rst for more details.
909 config GENERIC_EARLY_IOREMAP
912 config STACK_MAX_DEFAULT_SIZE_MB
913 int "Default maximum user stack size for 32-bit processes (MB)"
916 depends on STACK_GROWSUP && (!64BIT || COMPAT)
918 This is the maximum stack size in Megabytes in the VM layout of 32-bit
919 user processes when the stack grows upwards (currently only on parisc
920 arch) when the RLIMIT_STACK hard limit is unlimited.
922 A sane initial value is 100 MB.
924 config DEFERRED_STRUCT_PAGE_INIT
925 bool "Defer initialisation of struct pages to kthreads"
927 depends on !NEED_PER_CPU_KM
931 Ordinarily all struct pages are initialised during early boot in a
932 single thread. On very large machines this can take a considerable
933 amount of time. If this option is set, large machines will bring up
934 a subset of memmap at boot and then initialise the rest in parallel.
935 This has a potential performance impact on tasks running early in the
936 lifetime of the system until these kthreads finish the
939 config PAGE_IDLE_FLAG
941 select PAGE_EXTENSION if !64BIT
943 This adds PG_idle and PG_young flags to 'struct page'. PTE Accessed
944 bit writers can set the state of the bit in the flags so that PTE
945 Accessed bit readers may avoid disturbance.
947 config IDLE_PAGE_TRACKING
948 bool "Enable idle page tracking"
949 depends on SYSFS && MMU
950 select PAGE_IDLE_FLAG
952 This feature allows to estimate the amount of user pages that have
953 not been touched during a given period of time. This information can
954 be useful to tune memory cgroup limits and/or for job placement
955 within a compute cluster.
957 See Documentation/admin-guide/mm/idle_page_tracking.rst for
960 config ARCH_HAS_CACHE_LINE_SIZE
963 config ARCH_HAS_CURRENT_STACK_POINTER
966 In support of HARDENED_USERCOPY performing stack variable lifetime
967 checking, an architecture-agnostic way to find the stack pointer
968 is needed. Once an architecture defines an unsigned long global
969 register alias named "current_stack_pointer", this config can be
972 config ARCH_HAS_PTE_DEVMAP
975 config ARCH_HAS_ZONE_DMA_SET
979 bool "Support DMA zone" if ARCH_HAS_ZONE_DMA_SET
980 default y if ARM64 || X86
983 bool "Support DMA32 zone" if ARCH_HAS_ZONE_DMA_SET
988 bool "Device memory (pmem, HMM, etc...) hotplug support"
989 depends on MEMORY_HOTPLUG
990 depends on MEMORY_HOTREMOVE
991 depends on SPARSEMEM_VMEMMAP
992 depends on ARCH_HAS_PTE_DEVMAP
996 Device memory hotplug support allows for establishing pmem,
997 or other device driver discovered memory regions, in the
998 memmap. This allows pfn_to_page() lookups of otherwise
999 "device-physical" addresses which is needed for using a DAX
1000 mapping in an O_DIRECT operation, among other things.
1002 If FS_DAX is enabled, then say Y.
1005 # Helpers to mirror range of the CPU page tables of a process into device page
1012 config GET_FREE_REGION
1013 depends on SPARSEMEM
1016 config DEVICE_PRIVATE
1017 bool "Unaddressable device memory (GPU memory, ...)"
1018 depends on ZONE_DEVICE
1019 select GET_FREE_REGION
1022 Allows creation of struct pages to represent unaddressable device
1023 memory; i.e., memory that is only accessible from the device (or
1024 group of devices). You likely also want to select HMM_MIRROR.
1029 config ARCH_USES_HIGH_VMA_FLAGS
1031 config ARCH_HAS_PKEYS
1034 config ARCH_USES_PG_ARCH_X
1037 Enable the definition of PG_arch_x page flags with x > 1. Only
1038 suitable for 64-bit architectures with CONFIG_FLATMEM or
1039 CONFIG_SPARSEMEM_VMEMMAP enabled, otherwise there may not be
1040 enough room for additional bits in page->flags.
1042 config VM_EVENT_COUNTERS
1044 bool "Enable VM event counters for /proc/vmstat" if EXPERT
1046 VM event counters are needed for event counts to be shown.
1047 This option allows the disabling of the VM event counters
1048 on EXPERT systems. /proc/vmstat will only show page counts
1049 if VM event counters are disabled.
1052 bool "Collect percpu memory statistics"
1054 This feature collects and exposes statistics via debugfs. The
1055 information includes global and per chunk statistics, which can
1056 be used to help understand percpu memory usage.
1059 bool "Enable infrastructure for get_user_pages()-related unit tests"
1062 Provides /sys/kernel/debug/gup_test, which in turn provides a way
1063 to make ioctl calls that can launch kernel-based unit tests for
1064 the get_user_pages*() and pin_user_pages*() family of API calls.
1066 These tests include benchmark testing of the _fast variants of
1067 get_user_pages*() and pin_user_pages*(), as well as smoke tests of
1068 the non-_fast variants.
1070 There is also a sub-test that allows running dump_page() on any
1071 of up to eight pages (selected by command line args) within the
1072 range of user-space addresses. These pages are either pinned via
1073 pin_user_pages*(), or pinned via get_user_pages*(), as specified
1074 by other command line arguments.
1076 See tools/testing/selftests/vm/gup_test.c
1078 comment "GUP_TEST needs to have DEBUG_FS enabled"
1079 depends on !GUP_TEST && !DEBUG_FS
1081 config GUP_GET_PXX_LOW_HIGH
1084 config ARCH_HAS_PTE_SPECIAL
1088 # Some architectures require a special hugepage directory format that is
1089 # required to support multiple hugepage sizes. For example a4fe3ce76
1090 # "powerpc/mm: Allow more flexible layouts for hugepage pagetables"
1091 # introduced it on powerpc. This allows for a more flexible hugepage
1092 # pagetable layouts.
1094 config ARCH_HAS_HUGEPD
1097 config MAPPING_DIRTY_HELPERS
1103 config KMAP_LOCAL_NON_LINEAR_PTE_ARRAY
1106 # struct io_mapping based helper. Selected by drivers that need them
1112 bool "Enable memfd_secret() system call" if EXPERT
1113 depends on ARCH_HAS_SET_DIRECT_MAP
1115 Enable the memfd_secret() system call with the ability to create
1116 memory areas visible only in the context of the owning process and
1117 not mapped to other processes and other kernel page tables.
1119 config ANON_VMA_NAME
1120 bool "Anonymous VMA name support"
1121 depends on PROC_FS && ADVISE_SYSCALLS && MMU
1124 Allow naming anonymous virtual memory areas.
1126 This feature allows assigning names to virtual memory areas. Assigned
1127 names can be later retrieved from /proc/pid/maps and /proc/pid/smaps
1128 and help identifying individual anonymous memory areas.
1129 Assigning a name to anonymous virtual memory area might prevent that
1130 area from being merged with adjacent virtual memory areas due to the
1131 difference in their name.
1134 bool "Enable userfaultfd() system call"
1137 Enable the userfaultfd() system call that allows to intercept and
1138 handle page faults in userland.
1140 config HAVE_ARCH_USERFAULTFD_WP
1143 Arch has userfaultfd write protection support
1145 config HAVE_ARCH_USERFAULTFD_MINOR
1148 Arch has userfaultfd minor fault support
1150 config PTE_MARKER_UFFD_WP
1151 bool "Userfaultfd write protection support for shmem/hugetlbfs"
1153 depends on HAVE_ARCH_USERFAULTFD_WP
1156 Allows to create marker PTEs for userfaultfd write protection
1157 purposes. It is required to enable userfaultfd write protection on
1158 file-backed memory types like shmem and hugetlbfs.
1162 bool "Multi-Gen LRU"
1164 # make sure folio->flags has enough spare bits
1165 depends on 64BIT || !SPARSEMEM || SPARSEMEM_VMEMMAP
1167 A high performance LRU implementation to overcommit memory. See
1168 Documentation/admin-guide/mm/multigen_lru.rst for details.
1170 config LRU_GEN_ENABLED
1171 bool "Enable by default"
1174 This option enables the multi-gen LRU by default.
1176 config LRU_GEN_STATS
1177 bool "Full stats for debugging"
1180 Do not enable this option unless you plan to look at historical stats
1181 from evicted generations for debugging purpose.
1183 This option has a per-memcg and per-node memory overhead.
1186 source "mm/damon/Kconfig"