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.
49 config ZSWAP_EXCLUSIVE_LOADS_DEFAULT_ON
50 bool "Invalidate zswap entries when pages are loaded"
53 If selected, exclusive loads for zswap will be enabled at boot,
54 otherwise it will be disabled.
56 If exclusive loads are enabled, when a page is loaded from zswap,
57 the zswap entry is invalidated at once, as opposed to leaving it
58 in zswap until the swap entry is freed.
60 This avoids having two copies of the same page in memory
61 (compressed and uncompressed) after faulting in a page from zswap.
62 The cost is that if the page was never dirtied and needs to be
63 swapped out again, it will be re-compressed.
66 prompt "Default compressor"
68 default ZSWAP_COMPRESSOR_DEFAULT_LZO
70 Selects the default compression algorithm for the compressed cache
73 For an overview what kind of performance can be expected from
74 a particular compression algorithm please refer to the benchmarks
75 available at the following LWN page:
76 https://lwn.net/Articles/751795/
78 If in doubt, select 'LZO'.
80 The selection made here can be overridden by using the kernel
81 command line 'zswap.compressor=' option.
83 config ZSWAP_COMPRESSOR_DEFAULT_DEFLATE
87 Use the Deflate algorithm as the default compression algorithm.
89 config ZSWAP_COMPRESSOR_DEFAULT_LZO
93 Use the LZO algorithm as the default compression algorithm.
95 config ZSWAP_COMPRESSOR_DEFAULT_842
99 Use the 842 algorithm as the default compression algorithm.
101 config ZSWAP_COMPRESSOR_DEFAULT_LZ4
105 Use the LZ4 algorithm as the default compression algorithm.
107 config ZSWAP_COMPRESSOR_DEFAULT_LZ4HC
111 Use the LZ4HC algorithm as the default compression algorithm.
113 config ZSWAP_COMPRESSOR_DEFAULT_ZSTD
117 Use the zstd algorithm as the default compression algorithm.
120 config ZSWAP_COMPRESSOR_DEFAULT
123 default "deflate" if ZSWAP_COMPRESSOR_DEFAULT_DEFLATE
124 default "lzo" if ZSWAP_COMPRESSOR_DEFAULT_LZO
125 default "842" if ZSWAP_COMPRESSOR_DEFAULT_842
126 default "lz4" if ZSWAP_COMPRESSOR_DEFAULT_LZ4
127 default "lz4hc" if ZSWAP_COMPRESSOR_DEFAULT_LZ4HC
128 default "zstd" if ZSWAP_COMPRESSOR_DEFAULT_ZSTD
132 prompt "Default allocator"
134 default ZSWAP_ZPOOL_DEFAULT_ZBUD
136 Selects the default allocator for the compressed cache for
138 The default is 'zbud' for compatibility, however please do
139 read the description of each of the allocators below before
140 making a right choice.
142 The selection made here can be overridden by using the kernel
143 command line 'zswap.zpool=' option.
145 config ZSWAP_ZPOOL_DEFAULT_ZBUD
149 Use the zbud allocator as the default allocator.
151 config ZSWAP_ZPOOL_DEFAULT_Z3FOLD
155 Use the z3fold allocator as the default allocator.
157 config ZSWAP_ZPOOL_DEFAULT_ZSMALLOC
161 Use the zsmalloc allocator as the default allocator.
164 config ZSWAP_ZPOOL_DEFAULT
167 default "zbud" if ZSWAP_ZPOOL_DEFAULT_ZBUD
168 default "z3fold" if ZSWAP_ZPOOL_DEFAULT_Z3FOLD
169 default "zsmalloc" if ZSWAP_ZPOOL_DEFAULT_ZSMALLOC
173 tristate "2:1 compression allocator (zbud)"
176 A special purpose allocator for storing compressed pages.
177 It is designed to store up to two compressed pages per physical
178 page. While this design limits storage density, it has simple and
179 deterministic reclaim properties that make it preferable to a higher
180 density approach when reclaim will be used.
183 tristate "3:1 compression allocator (z3fold)"
186 A special purpose allocator for storing compressed pages.
187 It is designed to store up to three compressed pages per physical
188 page. It is a ZBUD derivative so the simplicity and determinism are
193 prompt "N:1 compression allocator (zsmalloc)" if ZSWAP
196 zsmalloc is a slab-based memory allocator designed to store
197 pages of various compression levels efficiently. It achieves
198 the highest storage density with the least amount of fragmentation.
201 bool "Export zsmalloc statistics"
205 This option enables code in the zsmalloc to collect various
206 statistics about what's happening in zsmalloc and exports that
207 information to userspace via debugfs.
210 config ZSMALLOC_CHAIN_SIZE
211 int "Maximum number of physical pages per-zspage"
216 This option sets the upper limit on the number of physical pages
217 that a zmalloc page (zspage) can consist of. The optimal zspage
218 chain size is calculated for each size class during the
219 initialization of the pool.
221 Changing this option can alter the characteristics of size classes,
222 such as the number of pages per zspage and the number of objects
223 per zspage. This can also result in different configurations of
224 the pool, as zsmalloc merges size classes with similar
227 For more information, see zsmalloc documentation.
229 menu "SLAB allocator options"
232 prompt "Choose SLAB allocator"
235 This option allows to select a slab allocator.
237 config SLAB_DEPRECATED
238 bool "SLAB (DEPRECATED)"
239 depends on !PREEMPT_RT
241 Deprecated and scheduled for removal in a few cycles. Replaced by
244 If you cannot migrate to SLUB, please contact linux-mm@kvack.org
245 and the people listed in the SLAB ALLOCATOR section of MAINTAINERS
246 file, explaining why.
248 The regular slab allocator that is established and known to work
249 well in all environments. It organizes cache hot objects in
250 per cpu and per node queues.
253 bool "SLUB (Unqueued Allocator)"
255 SLUB is a slab allocator that minimizes cache line usage
256 instead of managing queues of cached objects (SLAB approach).
257 Per cpu caching is realized using slabs of objects instead
258 of queues of objects. SLUB can use memory efficiently
259 and has enhanced diagnostics. SLUB is the default choice for
267 depends on SLAB_DEPRECATED
270 bool "Configure SLUB for minimal memory footprint"
271 depends on SLUB && EXPERT
272 select SLAB_MERGE_DEFAULT
274 Configures the SLUB allocator in a way to achieve minimal memory
275 footprint, sacrificing scalability, debugging and other features.
276 This is intended only for the smallest system that had used the
277 SLOB allocator and is not recommended for systems with more than
282 config SLAB_MERGE_DEFAULT
283 bool "Allow slab caches to be merged"
285 depends on SLAB || SLUB
287 For reduced kernel memory fragmentation, slab caches can be
288 merged when they share the same size and other characteristics.
289 This carries a risk of kernel heap overflows being able to
290 overwrite objects from merged caches (and more easily control
291 cache layout), which makes such heap attacks easier to exploit
292 by attackers. By keeping caches unmerged, these kinds of exploits
293 can usually only damage objects in the same cache. To disable
294 merging at runtime, "slab_nomerge" can be passed on the kernel
297 config SLAB_FREELIST_RANDOM
298 bool "Randomize slab freelist"
299 depends on SLAB || (SLUB && !SLUB_TINY)
301 Randomizes the freelist order used on creating new pages. This
302 security feature reduces the predictability of the kernel slab
303 allocator against heap overflows.
305 config SLAB_FREELIST_HARDENED
306 bool "Harden slab freelist metadata"
307 depends on SLAB || (SLUB && !SLUB_TINY)
309 Many kernel heap attacks try to target slab cache metadata and
310 other infrastructure. This options makes minor performance
311 sacrifices to harden the kernel slab allocator against common
312 freelist exploit methods. Some slab implementations have more
313 sanity-checking than others. This option is most effective with
318 bool "Enable SLUB performance statistics"
319 depends on SLUB && SYSFS && !SLUB_TINY
321 SLUB statistics are useful to debug SLUBs allocation behavior in
322 order find ways to optimize the allocator. This should never be
323 enabled for production use since keeping statistics slows down
324 the allocator by a few percentage points. The slabinfo command
325 supports the determination of the most active slabs to figure
326 out which slabs are relevant to a particular load.
327 Try running: slabinfo -DA
329 config SLUB_CPU_PARTIAL
331 depends on SLUB && SMP && !SLUB_TINY
332 bool "SLUB per cpu partial cache"
334 Per cpu partial caches accelerate objects allocation and freeing
335 that is local to a processor at the price of more indeterminism
336 in the latency of the free. On overflow these caches will be cleared
337 which requires the taking of locks that may cause latency spikes.
338 Typically one would choose no for a realtime system.
340 endmenu # SLAB allocator options
342 config SHUFFLE_PAGE_ALLOCATOR
343 bool "Page allocator randomization"
344 default SLAB_FREELIST_RANDOM && ACPI_NUMA
346 Randomization of the page allocator improves the average
347 utilization of a direct-mapped memory-side-cache. See section
348 5.2.27 Heterogeneous Memory Attribute Table (HMAT) in the ACPI
349 6.2a specification for an example of how a platform advertises
350 the presence of a memory-side-cache. There are also incidental
351 security benefits as it reduces the predictability of page
352 allocations to compliment SLAB_FREELIST_RANDOM, but the
353 default granularity of shuffling on the MAX_ORDER i.e, 10th
354 order of pages is selected based on cache utilization benefits
357 While the randomization improves cache utilization it may
358 negatively impact workloads on platforms without a cache. For
359 this reason, by default, the randomization is enabled only
360 after runtime detection of a direct-mapped memory-side-cache.
361 Otherwise, the randomization may be force enabled with the
362 'page_alloc.shuffle' kernel command line parameter.
367 bool "Disable heap randomization"
370 Randomizing heap placement makes heap exploits harder, but it
371 also breaks ancient binaries (including anything libc5 based).
372 This option changes the bootup default to heap randomization
373 disabled, and can be overridden at runtime by setting
374 /proc/sys/kernel/randomize_va_space to 2.
376 On non-ancient distros (post-2000 ones) N is usually a safe choice.
378 config MMAP_ALLOW_UNINITIALIZED
379 bool "Allow mmapped anonymous memory to be uninitialized"
380 depends on EXPERT && !MMU
383 Normally, and according to the Linux spec, anonymous memory obtained
384 from mmap() has its contents cleared before it is passed to
385 userspace. Enabling this config option allows you to request that
386 mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus
387 providing a huge performance boost. If this option is not enabled,
388 then the flag will be ignored.
390 This is taken advantage of by uClibc's malloc(), and also by
391 ELF-FDPIC binfmt's brk and stack allocator.
393 Because of the obvious security issues, this option should only be
394 enabled on embedded devices where you control what is run in
395 userspace. Since that isn't generally a problem on no-MMU systems,
396 it is normally safe to say Y here.
398 See Documentation/admin-guide/mm/nommu-mmap.rst for more information.
400 config SELECT_MEMORY_MODEL
402 depends on ARCH_SELECT_MEMORY_MODEL
405 prompt "Memory model"
406 depends on SELECT_MEMORY_MODEL
407 default SPARSEMEM_MANUAL if ARCH_SPARSEMEM_DEFAULT
408 default FLATMEM_MANUAL
410 This option allows you to change some of the ways that
411 Linux manages its memory internally. Most users will
412 only have one option here selected by the architecture
413 configuration. This is normal.
415 config FLATMEM_MANUAL
417 depends on !ARCH_SPARSEMEM_ENABLE || ARCH_FLATMEM_ENABLE
419 This option is best suited for non-NUMA systems with
420 flat address space. The FLATMEM is the most efficient
421 system in terms of performance and resource consumption
422 and it is the best option for smaller systems.
424 For systems that have holes in their physical address
425 spaces and for features like NUMA and memory hotplug,
426 choose "Sparse Memory".
428 If unsure, choose this option (Flat Memory) over any other.
430 config SPARSEMEM_MANUAL
432 depends on ARCH_SPARSEMEM_ENABLE
434 This will be the only option for some systems, including
435 memory hot-plug systems. This is normal.
437 This option provides efficient support for systems with
438 holes is their physical address space and allows memory
439 hot-plug and hot-remove.
441 If unsure, choose "Flat Memory" over this option.
447 depends on (!SELECT_MEMORY_MODEL && ARCH_SPARSEMEM_ENABLE) || SPARSEMEM_MANUAL
451 depends on !SPARSEMEM || FLATMEM_MANUAL
454 # SPARSEMEM_EXTREME (which is the default) does some bootmem
455 # allocations when sparse_init() is called. If this cannot
456 # be done on your architecture, select this option. However,
457 # statically allocating the mem_section[] array can potentially
458 # consume vast quantities of .bss, so be careful.
460 # This option will also potentially produce smaller runtime code
461 # with gcc 3.4 and later.
463 config SPARSEMEM_STATIC
467 # Architecture platforms which require a two level mem_section in SPARSEMEM
468 # must select this option. This is usually for architecture platforms with
469 # an extremely sparse physical address space.
471 config SPARSEMEM_EXTREME
473 depends on SPARSEMEM && !SPARSEMEM_STATIC
475 config SPARSEMEM_VMEMMAP_ENABLE
478 config SPARSEMEM_VMEMMAP
479 bool "Sparse Memory virtual memmap"
480 depends on SPARSEMEM && SPARSEMEM_VMEMMAP_ENABLE
483 SPARSEMEM_VMEMMAP uses a virtually mapped memmap to optimise
484 pfn_to_page and page_to_pfn operations. This is the most
485 efficient option when sufficient kernel resources are available.
487 # Select this config option from the architecture Kconfig, if it is preferred
488 # to enable the feature of HugeTLB/dev_dax vmemmap optimization.
490 config ARCH_WANT_OPTIMIZE_VMEMMAP
493 config HAVE_MEMBLOCK_PHYS_MAP
500 # Don't discard allocated memory used to track "memory" and "reserved" memblocks
501 # after early boot, so it can still be used to test for validity of memory.
502 # Also, memblocks are updated with memory hot(un)plug.
503 config ARCH_KEEP_MEMBLOCK
506 # Keep arch NUMA mapping infrastructure post-init.
507 config NUMA_KEEP_MEMINFO
510 config MEMORY_ISOLATION
513 # IORESOURCE_SYSTEM_RAM regions in the kernel resource tree that are marked
514 # IORESOURCE_EXCLUSIVE cannot be mapped to user space, for example, via
516 config EXCLUSIVE_SYSTEM_RAM
518 depends on !DEVMEM || STRICT_DEVMEM
521 # Only be set on architectures that have completely implemented memory hotplug
522 # feature. If you are not sure, don't touch it.
524 config HAVE_BOOTMEM_INFO_NODE
527 config ARCH_ENABLE_MEMORY_HOTPLUG
530 config ARCH_ENABLE_MEMORY_HOTREMOVE
533 # eventually, we can have this option just 'select SPARSEMEM'
534 menuconfig MEMORY_HOTPLUG
535 bool "Memory hotplug"
536 select MEMORY_ISOLATION
538 depends on ARCH_ENABLE_MEMORY_HOTPLUG
540 select NUMA_KEEP_MEMINFO if NUMA
544 config MEMORY_HOTPLUG_DEFAULT_ONLINE
545 bool "Online the newly added memory blocks by default"
546 depends on MEMORY_HOTPLUG
548 This option sets the default policy setting for memory hotplug
549 onlining policy (/sys/devices/system/memory/auto_online_blocks) which
550 determines what happens to newly added memory regions. Policy setting
551 can always be changed at runtime.
552 See Documentation/admin-guide/mm/memory-hotplug.rst for more information.
554 Say Y here if you want all hot-plugged memory blocks to appear in
555 'online' state by default.
556 Say N here if you want the default policy to keep all hot-plugged
557 memory blocks in 'offline' state.
559 config MEMORY_HOTREMOVE
560 bool "Allow for memory hot remove"
561 select HAVE_BOOTMEM_INFO_NODE if (X86_64 || PPC64)
562 depends on MEMORY_HOTPLUG && ARCH_ENABLE_MEMORY_HOTREMOVE
565 config MHP_MEMMAP_ON_MEMORY
567 depends on MEMORY_HOTPLUG && SPARSEMEM_VMEMMAP
568 depends on ARCH_MHP_MEMMAP_ON_MEMORY_ENABLE
570 endif # MEMORY_HOTPLUG
572 # Heavily threaded applications may benefit from splitting the mm-wide
573 # page_table_lock, so that faults on different parts of the user address
574 # space can be handled with less contention: split it at this NR_CPUS.
575 # Default to 4 for wider testing, though 8 might be more appropriate.
576 # ARM's adjust_pte (unused if VIPT) depends on mm-wide page_table_lock.
577 # PA-RISC 7xxx's spinlock_t would enlarge struct page from 32 to 44 bytes.
578 # SPARC32 allocates multiple pte tables within a single page, and therefore
579 # a per-page lock leads to problems when multiple tables need to be locked
580 # at the same time (e.g. copy_page_range()).
581 # DEBUG_SPINLOCK and DEBUG_LOCK_ALLOC spinlock_t also enlarge struct page.
583 config SPLIT_PTLOCK_CPUS
585 default "999999" if !MMU
586 default "999999" if ARM && !CPU_CACHE_VIPT
587 default "999999" if PARISC && !PA20
588 default "999999" if SPARC32
591 config ARCH_ENABLE_SPLIT_PMD_PTLOCK
595 # support for memory balloon
596 config MEMORY_BALLOON
600 # support for memory balloon compaction
601 config BALLOON_COMPACTION
602 bool "Allow for balloon memory compaction/migration"
604 depends on COMPACTION && MEMORY_BALLOON
606 Memory fragmentation introduced by ballooning might reduce
607 significantly the number of 2MB contiguous memory blocks that can be
608 used within a guest, thus imposing performance penalties associated
609 with the reduced number of transparent huge pages that could be used
610 by the guest workload. Allowing the compaction & migration for memory
611 pages enlisted as being part of memory balloon devices avoids the
612 scenario aforementioned and helps improving memory defragmentation.
615 # support for memory compaction
617 bool "Allow for memory compaction"
622 Compaction is the only memory management component to form
623 high order (larger physically contiguous) memory blocks
624 reliably. The page allocator relies on compaction heavily and
625 the lack of the feature can lead to unexpected OOM killer
626 invocations for high order memory requests. You shouldn't
627 disable this option unless there really is a strong reason for
628 it and then we would be really interested to hear about that at
631 config COMPACT_UNEVICTABLE_DEFAULT
633 depends on COMPACTION
634 default 0 if PREEMPT_RT
638 # support for free page reporting
639 config PAGE_REPORTING
640 bool "Free page reporting"
643 Free page reporting allows for the incremental acquisition of
644 free pages from the buddy allocator for the purpose of reporting
645 those pages to another entity, such as a hypervisor, so that the
646 memory can be freed within the host for other uses.
649 # support for page migration
652 bool "Page migration"
654 depends on (NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION || CMA) && MMU
656 Allows the migration of the physical location of pages of processes
657 while the virtual addresses are not changed. This is useful in
658 two situations. The first is on NUMA systems to put pages nearer
659 to the processors accessing. The second is when allocating huge
660 pages as migration can relocate pages to satisfy a huge page
661 allocation instead of reclaiming.
663 config DEVICE_MIGRATION
664 def_bool MIGRATION && ZONE_DEVICE
666 config ARCH_ENABLE_HUGEPAGE_MIGRATION
669 config ARCH_ENABLE_THP_MIGRATION
672 config HUGETLB_PAGE_SIZE_VARIABLE
675 Allows the pageblock_order value to be dynamic instead of just standard
676 HUGETLB_PAGE_ORDER when there are multiple HugeTLB page sizes available
679 Note that the pageblock_order cannot exceed MAX_ORDER and will be
680 clamped down to MAX_ORDER.
683 def_bool (MEMORY_ISOLATION && COMPACTION) || CMA
685 config PHYS_ADDR_T_64BIT
689 bool "Enable bounce buffers"
691 depends on BLOCK && MMU && HIGHMEM
693 Enable bounce buffers for devices that cannot access the full range of
694 memory available to the CPU. Enabled by default when HIGHMEM is
695 selected, but you may say n to override this.
702 bool "Enable KSM for page merging"
706 Enable Kernel Samepage Merging: KSM periodically scans those areas
707 of an application's address space that an app has advised may be
708 mergeable. When it finds pages of identical content, it replaces
709 the many instances by a single page with that content, so
710 saving memory until one or another app needs to modify the content.
711 Recommended for use with KVM, or with other duplicative applications.
712 See Documentation/mm/ksm.rst for more information: KSM is inactive
713 until a program has madvised that an area is MADV_MERGEABLE, and
714 root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set).
716 config DEFAULT_MMAP_MIN_ADDR
717 int "Low address space to protect from user allocation"
721 This is the portion of low virtual memory which should be protected
722 from userspace allocation. Keeping a user from writing to low pages
723 can help reduce the impact of kernel NULL pointer bugs.
725 For most ia64, ppc64 and x86 users with lots of address space
726 a value of 65536 is reasonable and should cause no problems.
727 On arm and other archs it should not be higher than 32768.
728 Programs which use vm86 functionality or have some need to map
729 this low address space will need CAP_SYS_RAWIO or disable this
730 protection by setting the value to 0.
732 This value can be changed after boot using the
733 /proc/sys/vm/mmap_min_addr tunable.
735 config ARCH_SUPPORTS_MEMORY_FAILURE
738 config MEMORY_FAILURE
740 depends on ARCH_SUPPORTS_MEMORY_FAILURE
741 bool "Enable recovery from hardware memory errors"
742 select MEMORY_ISOLATION
745 Enables code to recover from some memory failures on systems
746 with MCA recovery. This allows a system to continue running
747 even when some of its memory has uncorrected errors. This requires
748 special hardware support and typically ECC memory.
750 config HWPOISON_INJECT
751 tristate "HWPoison pages injector"
752 depends on MEMORY_FAILURE && DEBUG_KERNEL && PROC_FS
753 select PROC_PAGE_MONITOR
755 config NOMMU_INITIAL_TRIM_EXCESS
756 int "Turn on mmap() excess space trimming before booting"
760 The NOMMU mmap() frequently needs to allocate large contiguous chunks
761 of memory on which to store mappings, but it can only ask the system
762 allocator for chunks in 2^N*PAGE_SIZE amounts - which is frequently
763 more than it requires. To deal with this, mmap() is able to trim off
764 the excess and return it to the allocator.
766 If trimming is enabled, the excess is trimmed off and returned to the
767 system allocator, which can cause extra fragmentation, particularly
768 if there are a lot of transient processes.
770 If trimming is disabled, the excess is kept, but not used, which for
771 long-term mappings means that the space is wasted.
773 Trimming can be dynamically controlled through a sysctl option
774 (/proc/sys/vm/nr_trim_pages) which specifies the minimum number of
775 excess pages there must be before trimming should occur, or zero if
776 no trimming is to occur.
778 This option specifies the initial value of this option. The default
779 of 1 says that all excess pages should be trimmed.
781 See Documentation/admin-guide/mm/nommu-mmap.rst for more information.
783 config ARCH_WANT_GENERAL_HUGETLB
786 config ARCH_WANTS_THP_SWAP
789 menuconfig TRANSPARENT_HUGEPAGE
790 bool "Transparent Hugepage Support"
791 depends on HAVE_ARCH_TRANSPARENT_HUGEPAGE && !PREEMPT_RT
795 Transparent Hugepages allows the kernel to use huge pages and
796 huge tlb transparently to the applications whenever possible.
797 This feature can improve computing performance to certain
798 applications by speeding up page faults during memory
799 allocation, by reducing the number of tlb misses and by speeding
800 up the pagetable walking.
802 If memory constrained on embedded, you may want to say N.
804 if TRANSPARENT_HUGEPAGE
807 prompt "Transparent Hugepage Support sysfs defaults"
808 depends on TRANSPARENT_HUGEPAGE
809 default TRANSPARENT_HUGEPAGE_ALWAYS
811 Selects the sysfs defaults for Transparent Hugepage Support.
813 config TRANSPARENT_HUGEPAGE_ALWAYS
816 Enabling Transparent Hugepage always, can increase the
817 memory footprint of applications without a guaranteed
818 benefit but it will work automatically for all applications.
820 config TRANSPARENT_HUGEPAGE_MADVISE
823 Enabling Transparent Hugepage madvise, will only provide a
824 performance improvement benefit to the applications using
825 madvise(MADV_HUGEPAGE) but it won't risk to increase the
826 memory footprint of applications without a guaranteed
832 depends on TRANSPARENT_HUGEPAGE && ARCH_WANTS_THP_SWAP && SWAP && 64BIT
834 Swap transparent huge pages in one piece, without splitting.
835 XXX: For now, swap cluster backing transparent huge page
836 will be split after swapout.
838 For selection by architectures with reasonable THP sizes.
840 config READ_ONLY_THP_FOR_FS
841 bool "Read-only THP for filesystems (EXPERIMENTAL)"
842 depends on TRANSPARENT_HUGEPAGE && SHMEM
845 Allow khugepaged to put read-only file-backed pages in THP.
847 This is marked experimental because it is a new feature. Write
848 support of file THPs will be developed in the next few release
851 endif # TRANSPARENT_HUGEPAGE
854 # UP and nommu archs use km based percpu allocator
856 config NEED_PER_CPU_KM
857 depends on !SMP || !MMU
861 config NEED_PER_CPU_EMBED_FIRST_CHUNK
864 config NEED_PER_CPU_PAGE_FIRST_CHUNK
867 config USE_PERCPU_NUMA_NODE_ID
870 config HAVE_SETUP_PER_CPU_AREA
877 bool "Contiguous Memory Allocator"
880 select MEMORY_ISOLATION
882 This enables the Contiguous Memory Allocator which allows other
883 subsystems to allocate big physically-contiguous blocks of memory.
884 CMA reserves a region of memory and allows only movable pages to
885 be allocated from it. This way, the kernel can use the memory for
886 pagecache and when a subsystem requests for contiguous area, the
887 allocated pages are migrated away to serve the contiguous request.
892 bool "CMA debug messages (DEVELOPMENT)"
893 depends on DEBUG_KERNEL && CMA
895 Turns on debug messages in CMA. This produces KERN_DEBUG
896 messages for every CMA call as well as various messages while
897 processing calls such as dma_alloc_from_contiguous().
898 This option does not affect warning and error messages.
901 bool "CMA debugfs interface"
902 depends on CMA && DEBUG_FS
904 Turns on the DebugFS interface for CMA.
907 bool "CMA information through sysfs interface"
908 depends on CMA && SYSFS
910 This option exposes some sysfs attributes to get information
914 int "Maximum count of the CMA areas"
919 CMA allows to create CMA areas for particular purpose, mainly,
920 used as device private area. This parameter sets the maximum
921 number of CMA area in the system.
923 If unsure, leave the default value "7" in UMA and "19" in NUMA.
925 config MEM_SOFT_DIRTY
926 bool "Track memory changes"
927 depends on CHECKPOINT_RESTORE && HAVE_ARCH_SOFT_DIRTY && PROC_FS
928 select PROC_PAGE_MONITOR
930 This option enables memory changes tracking by introducing a
931 soft-dirty bit on pte-s. This bit it set when someone writes
932 into a page just as regular dirty bit, but unlike the latter
933 it can be cleared by hands.
935 See Documentation/admin-guide/mm/soft-dirty.rst for more details.
937 config GENERIC_EARLY_IOREMAP
940 config STACK_MAX_DEFAULT_SIZE_MB
941 int "Default maximum user stack size for 32-bit processes (MB)"
944 depends on STACK_GROWSUP && (!64BIT || COMPAT)
946 This is the maximum stack size in Megabytes in the VM layout of 32-bit
947 user processes when the stack grows upwards (currently only on parisc
948 arch) when the RLIMIT_STACK hard limit is unlimited.
950 A sane initial value is 100 MB.
952 config DEFERRED_STRUCT_PAGE_INIT
953 bool "Defer initialisation of struct pages to kthreads"
955 depends on !NEED_PER_CPU_KM
959 Ordinarily all struct pages are initialised during early boot in a
960 single thread. On very large machines this can take a considerable
961 amount of time. If this option is set, large machines will bring up
962 a subset of memmap at boot and then initialise the rest in parallel.
963 This has a potential performance impact on tasks running early in the
964 lifetime of the system until these kthreads finish the
967 config PAGE_IDLE_FLAG
969 select PAGE_EXTENSION if !64BIT
971 This adds PG_idle and PG_young flags to 'struct page'. PTE Accessed
972 bit writers can set the state of the bit in the flags so that PTE
973 Accessed bit readers may avoid disturbance.
975 config IDLE_PAGE_TRACKING
976 bool "Enable idle page tracking"
977 depends on SYSFS && MMU
978 select PAGE_IDLE_FLAG
980 This feature allows to estimate the amount of user pages that have
981 not been touched during a given period of time. This information can
982 be useful to tune memory cgroup limits and/or for job placement
983 within a compute cluster.
985 See Documentation/admin-guide/mm/idle_page_tracking.rst for
988 config ARCH_HAS_CACHE_LINE_SIZE
991 config ARCH_HAS_CURRENT_STACK_POINTER
994 In support of HARDENED_USERCOPY performing stack variable lifetime
995 checking, an architecture-agnostic way to find the stack pointer
996 is needed. Once an architecture defines an unsigned long global
997 register alias named "current_stack_pointer", this config can be
1000 config ARCH_HAS_PTE_DEVMAP
1003 config ARCH_HAS_ZONE_DMA_SET
1007 bool "Support DMA zone" if ARCH_HAS_ZONE_DMA_SET
1008 default y if ARM64 || X86
1011 bool "Support DMA32 zone" if ARCH_HAS_ZONE_DMA_SET
1016 bool "Device memory (pmem, HMM, etc...) hotplug support"
1017 depends on MEMORY_HOTPLUG
1018 depends on MEMORY_HOTREMOVE
1019 depends on SPARSEMEM_VMEMMAP
1020 depends on ARCH_HAS_PTE_DEVMAP
1024 Device memory hotplug support allows for establishing pmem,
1025 or other device driver discovered memory regions, in the
1026 memmap. This allows pfn_to_page() lookups of otherwise
1027 "device-physical" addresses which is needed for using a DAX
1028 mapping in an O_DIRECT operation, among other things.
1030 If FS_DAX is enabled, then say Y.
1033 # Helpers to mirror range of the CPU page tables of a process into device page
1040 config GET_FREE_REGION
1041 depends on SPARSEMEM
1044 config DEVICE_PRIVATE
1045 bool "Unaddressable device memory (GPU memory, ...)"
1046 depends on ZONE_DEVICE
1047 select GET_FREE_REGION
1050 Allows creation of struct pages to represent unaddressable device
1051 memory; i.e., memory that is only accessible from the device (or
1052 group of devices). You likely also want to select HMM_MIRROR.
1057 config ARCH_USES_HIGH_VMA_FLAGS
1059 config ARCH_HAS_PKEYS
1062 config ARCH_USES_PG_ARCH_X
1065 Enable the definition of PG_arch_x page flags with x > 1. Only
1066 suitable for 64-bit architectures with CONFIG_FLATMEM or
1067 CONFIG_SPARSEMEM_VMEMMAP enabled, otherwise there may not be
1068 enough room for additional bits in page->flags.
1070 config VM_EVENT_COUNTERS
1072 bool "Enable VM event counters for /proc/vmstat" if EXPERT
1074 VM event counters are needed for event counts to be shown.
1075 This option allows the disabling of the VM event counters
1076 on EXPERT systems. /proc/vmstat will only show page counts
1077 if VM event counters are disabled.
1080 bool "Collect percpu memory statistics"
1082 This feature collects and exposes statistics via debugfs. The
1083 information includes global and per chunk statistics, which can
1084 be used to help understand percpu memory usage.
1087 bool "Enable infrastructure for get_user_pages()-related unit tests"
1090 Provides /sys/kernel/debug/gup_test, which in turn provides a way
1091 to make ioctl calls that can launch kernel-based unit tests for
1092 the get_user_pages*() and pin_user_pages*() family of API calls.
1094 These tests include benchmark testing of the _fast variants of
1095 get_user_pages*() and pin_user_pages*(), as well as smoke tests of
1096 the non-_fast variants.
1098 There is also a sub-test that allows running dump_page() on any
1099 of up to eight pages (selected by command line args) within the
1100 range of user-space addresses. These pages are either pinned via
1101 pin_user_pages*(), or pinned via get_user_pages*(), as specified
1102 by other command line arguments.
1104 See tools/testing/selftests/mm/gup_test.c
1106 comment "GUP_TEST needs to have DEBUG_FS enabled"
1107 depends on !GUP_TEST && !DEBUG_FS
1109 config GUP_GET_PXX_LOW_HIGH
1113 tristate "Enable a module to run time tests on dma_pool"
1116 Provides a test module that will allocate and free many blocks of
1117 various sizes and report how long it takes. This is intended to
1118 provide a consistent way to measure how changes to the
1119 dma_pool_alloc/free routines affect performance.
1121 config ARCH_HAS_PTE_SPECIAL
1125 # Some architectures require a special hugepage directory format that is
1126 # required to support multiple hugepage sizes. For example a4fe3ce76
1127 # "powerpc/mm: Allow more flexible layouts for hugepage pagetables"
1128 # introduced it on powerpc. This allows for a more flexible hugepage
1129 # pagetable layouts.
1131 config ARCH_HAS_HUGEPD
1134 config MAPPING_DIRTY_HELPERS
1140 config KMAP_LOCAL_NON_LINEAR_PTE_ARRAY
1143 # struct io_mapping based helper. Selected by drivers that need them
1149 bool "Enable memfd_secret() system call" if EXPERT
1150 depends on ARCH_HAS_SET_DIRECT_MAP
1152 Enable the memfd_secret() system call with the ability to create
1153 memory areas visible only in the context of the owning process and
1154 not mapped to other processes and other kernel page tables.
1156 config ANON_VMA_NAME
1157 bool "Anonymous VMA name support"
1158 depends on PROC_FS && ADVISE_SYSCALLS && MMU
1161 Allow naming anonymous virtual memory areas.
1163 This feature allows assigning names to virtual memory areas. Assigned
1164 names can be later retrieved from /proc/pid/maps and /proc/pid/smaps
1165 and help identifying individual anonymous memory areas.
1166 Assigning a name to anonymous virtual memory area might prevent that
1167 area from being merged with adjacent virtual memory areas due to the
1168 difference in their name.
1171 bool "Enable userfaultfd() system call"
1174 Enable the userfaultfd() system call that allows to intercept and
1175 handle page faults in userland.
1177 config HAVE_ARCH_USERFAULTFD_WP
1180 Arch has userfaultfd write protection support
1182 config HAVE_ARCH_USERFAULTFD_MINOR
1185 Arch has userfaultfd minor fault support
1187 config PTE_MARKER_UFFD_WP
1188 bool "Userfaultfd write protection support for shmem/hugetlbfs"
1190 depends on HAVE_ARCH_USERFAULTFD_WP
1193 Allows to create marker PTEs for userfaultfd write protection
1194 purposes. It is required to enable userfaultfd write protection on
1195 file-backed memory types like shmem and hugetlbfs.
1199 bool "Multi-Gen LRU"
1201 # make sure folio->flags has enough spare bits
1202 depends on 64BIT || !SPARSEMEM || SPARSEMEM_VMEMMAP
1204 A high performance LRU implementation to overcommit memory. See
1205 Documentation/admin-guide/mm/multigen_lru.rst for details.
1207 config LRU_GEN_ENABLED
1208 bool "Enable by default"
1211 This option enables the multi-gen LRU by default.
1213 config LRU_GEN_STATS
1214 bool "Full stats for debugging"
1217 Do not enable this option unless you plan to look at historical stats
1218 from evicted generations for debugging purpose.
1220 This option has a per-memcg and per-node memory overhead.
1223 config ARCH_SUPPORTS_PER_VMA_LOCK
1227 bool "Enable per-vma locking during page fault handling."
1228 depends on ARCH_SUPPORTS_PER_VMA_LOCK && MMU && SMP
1231 Allow per-vma locking during page fault handling.
1233 This feature allows locking each virtual memory area separately when
1234 handling page faults instead of taking mmap_lock.
1236 config LOCK_MM_AND_FIND_VMA
1238 depends on !STACK_GROWSUP
1240 source "mm/damon/Kconfig"