xref: /openbmc/linux/mm/Kconfig (revision 0b6f15824cc7e431a9706c78bfb9cb3011477ad3)
1# SPDX-License-Identifier: GPL-2.0-only
2
3menu "Memory Management options"
4
5#
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.
8#
9config ARCH_NO_SWAP
10	bool
11
12config ZPOOL
13	bool
14
15menuconfig SWAP
16	bool "Support for paging of anonymous memory (swap)"
17	depends on MMU && BLOCK && !ARCH_NO_SWAP
18	default y
19	help
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.
24
25config ZSWAP
26	bool "Compressed cache for swap pages"
27	depends on SWAP
28	select FRONTSWAP
29	select CRYPTO
30	select ZPOOL
31	help
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.
38
39config ZSWAP_DEFAULT_ON
40	bool "Enable the compressed cache for swap pages by default"
41	depends on ZSWAP
42	help
43	  If selected, the compressed cache for swap pages will be enabled
44	  at boot, otherwise it will be disabled.
45
46	  The selection made here can be overridden by using the kernel
47	  command line 'zswap.enabled=' option.
48
49config ZSWAP_EXCLUSIVE_LOADS_DEFAULT_ON
50	bool "Invalidate zswap entries when pages are loaded"
51	depends on ZSWAP
52	help
53	  If selected, exclusive loads for zswap will be enabled at boot,
54	  otherwise it will be disabled.
55
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.
59
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.
64
65choice
66	prompt "Default compressor"
67	depends on ZSWAP
68	default ZSWAP_COMPRESSOR_DEFAULT_LZO
69	help
70	  Selects the default compression algorithm for the compressed cache
71	  for swap pages.
72
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/
77
78	  If in doubt, select 'LZO'.
79
80	  The selection made here can be overridden by using the kernel
81	  command line 'zswap.compressor=' option.
82
83config ZSWAP_COMPRESSOR_DEFAULT_DEFLATE
84	bool "Deflate"
85	select CRYPTO_DEFLATE
86	help
87	  Use the Deflate algorithm as the default compression algorithm.
88
89config ZSWAP_COMPRESSOR_DEFAULT_LZO
90	bool "LZO"
91	select CRYPTO_LZO
92	help
93	  Use the LZO algorithm as the default compression algorithm.
94
95config ZSWAP_COMPRESSOR_DEFAULT_842
96	bool "842"
97	select CRYPTO_842
98	help
99	  Use the 842 algorithm as the default compression algorithm.
100
101config ZSWAP_COMPRESSOR_DEFAULT_LZ4
102	bool "LZ4"
103	select CRYPTO_LZ4
104	help
105	  Use the LZ4 algorithm as the default compression algorithm.
106
107config ZSWAP_COMPRESSOR_DEFAULT_LZ4HC
108	bool "LZ4HC"
109	select CRYPTO_LZ4HC
110	help
111	  Use the LZ4HC algorithm as the default compression algorithm.
112
113config ZSWAP_COMPRESSOR_DEFAULT_ZSTD
114	bool "zstd"
115	select CRYPTO_ZSTD
116	help
117	  Use the zstd algorithm as the default compression algorithm.
118endchoice
119
120config ZSWAP_COMPRESSOR_DEFAULT
121       string
122       depends on ZSWAP
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
129       default ""
130
131choice
132	prompt "Default allocator"
133	depends on ZSWAP
134	default ZSWAP_ZPOOL_DEFAULT_ZBUD
135	help
136	  Selects the default allocator for the compressed cache for
137	  swap pages.
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.
141
142	  The selection made here can be overridden by using the kernel
143	  command line 'zswap.zpool=' option.
144
145config ZSWAP_ZPOOL_DEFAULT_ZBUD
146	bool "zbud"
147	select ZBUD
148	help
149	  Use the zbud allocator as the default allocator.
150
151config ZSWAP_ZPOOL_DEFAULT_Z3FOLD
152	bool "z3fold"
153	select Z3FOLD
154	help
155	  Use the z3fold allocator as the default allocator.
156
157config ZSWAP_ZPOOL_DEFAULT_ZSMALLOC
158	bool "zsmalloc"
159	select ZSMALLOC
160	help
161	  Use the zsmalloc allocator as the default allocator.
162endchoice
163
164config ZSWAP_ZPOOL_DEFAULT
165       string
166       depends on ZSWAP
167       default "zbud" if ZSWAP_ZPOOL_DEFAULT_ZBUD
168       default "z3fold" if ZSWAP_ZPOOL_DEFAULT_Z3FOLD
169       default "zsmalloc" if ZSWAP_ZPOOL_DEFAULT_ZSMALLOC
170       default ""
171
172config ZBUD
173	tristate "2:1 compression allocator (zbud)"
174	depends on ZSWAP
175	help
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.
181
182config Z3FOLD
183	tristate "3:1 compression allocator (z3fold)"
184	depends on ZSWAP
185	help
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
189	  still there.
190
191config ZSMALLOC
192	tristate
193	prompt "N:1 compression allocator (zsmalloc)" if ZSWAP
194	depends on MMU
195	help
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.
199
200config ZSMALLOC_STAT
201	bool "Export zsmalloc statistics"
202	depends on ZSMALLOC
203	select DEBUG_FS
204	help
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.
208	  If unsure, say N.
209
210config ZSMALLOC_CHAIN_SIZE
211	int "Maximum number of physical pages per-zspage"
212	default 8
213	range 4 16
214	depends on ZSMALLOC
215	help
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.
220
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
225	  characteristics.
226
227	  For more information, see zsmalloc documentation.
228
229menu "SLAB allocator options"
230
231choice
232	prompt "Choose SLAB allocator"
233	default SLUB
234	help
235	   This option allows to select a slab allocator.
236
237config SLAB_DEPRECATED
238	bool "SLAB (DEPRECATED)"
239	depends on !PREEMPT_RT
240	help
241	  Deprecated and scheduled for removal in a few cycles. Replaced by
242	  SLUB.
243
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.
247
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.
251
252config SLUB
253	bool "SLUB (Unqueued Allocator)"
254	help
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
260	   a slab allocator.
261
262endchoice
263
264config SLAB
265	bool
266	default y
267	depends on SLAB_DEPRECATED
268
269config SLUB_TINY
270	bool "Configure SLUB for minimal memory footprint"
271	depends on SLUB && EXPERT
272	select SLAB_MERGE_DEFAULT
273	help
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
278	   16MB RAM.
279
280	   If unsure, say N.
281
282config SLAB_MERGE_DEFAULT
283	bool "Allow slab caches to be merged"
284	default y
285	depends on SLAB || SLUB
286	help
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
295	  command line.
296
297config SLAB_FREELIST_RANDOM
298	bool "Randomize slab freelist"
299	depends on SLAB || (SLUB && !SLUB_TINY)
300	help
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.
304
305config SLAB_FREELIST_HARDENED
306	bool "Harden slab freelist metadata"
307	depends on SLAB || (SLUB && !SLUB_TINY)
308	help
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
314	  CONFIG_SLUB.
315
316config SLUB_STATS
317	default n
318	bool "Enable SLUB performance statistics"
319	depends on SLUB && SYSFS && !SLUB_TINY
320	help
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
328
329config SLUB_CPU_PARTIAL
330	default y
331	depends on SLUB && SMP && !SLUB_TINY
332	bool "SLUB per cpu partial cache"
333	help
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.
339
340endmenu # SLAB allocator options
341
342config SHUFFLE_PAGE_ALLOCATOR
343	bool "Page allocator randomization"
344	default SLAB_FREELIST_RANDOM && ACPI_NUMA
345	help
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
355	  on x86.
356
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.
363
364	  Say Y if unsure.
365
366config COMPAT_BRK
367	bool "Disable heap randomization"
368	default y
369	help
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.
375
376	  On non-ancient distros (post-2000 ones) N is usually a safe choice.
377
378config MMAP_ALLOW_UNINITIALIZED
379	bool "Allow mmapped anonymous memory to be uninitialized"
380	depends on EXPERT && !MMU
381	default n
382	help
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.
389
390	  This is taken advantage of by uClibc's malloc(), and also by
391	  ELF-FDPIC binfmt's brk and stack allocator.
392
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.
397
398	  See Documentation/admin-guide/mm/nommu-mmap.rst for more information.
399
400config SELECT_MEMORY_MODEL
401	def_bool y
402	depends on ARCH_SELECT_MEMORY_MODEL
403
404choice
405	prompt "Memory model"
406	depends on SELECT_MEMORY_MODEL
407	default SPARSEMEM_MANUAL if ARCH_SPARSEMEM_DEFAULT
408	default FLATMEM_MANUAL
409	help
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.
414
415config FLATMEM_MANUAL
416	bool "Flat Memory"
417	depends on !ARCH_SPARSEMEM_ENABLE || ARCH_FLATMEM_ENABLE
418	help
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.
423
424	  For systems that have holes in their physical address
425	  spaces and for features like NUMA and memory hotplug,
426	  choose "Sparse Memory".
427
428	  If unsure, choose this option (Flat Memory) over any other.
429
430config SPARSEMEM_MANUAL
431	bool "Sparse Memory"
432	depends on ARCH_SPARSEMEM_ENABLE
433	help
434	  This will be the only option for some systems, including
435	  memory hot-plug systems.  This is normal.
436
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.
440
441	  If unsure, choose "Flat Memory" over this option.
442
443endchoice
444
445config SPARSEMEM
446	def_bool y
447	depends on (!SELECT_MEMORY_MODEL && ARCH_SPARSEMEM_ENABLE) || SPARSEMEM_MANUAL
448
449config FLATMEM
450	def_bool y
451	depends on !SPARSEMEM || FLATMEM_MANUAL
452
453#
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.
459#
460# This option will also potentially produce smaller runtime code
461# with gcc 3.4 and later.
462#
463config SPARSEMEM_STATIC
464	bool
465
466#
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.
470#
471config SPARSEMEM_EXTREME
472	def_bool y
473	depends on SPARSEMEM && !SPARSEMEM_STATIC
474
475config SPARSEMEM_VMEMMAP_ENABLE
476	bool
477
478config SPARSEMEM_VMEMMAP
479	bool "Sparse Memory virtual memmap"
480	depends on SPARSEMEM && SPARSEMEM_VMEMMAP_ENABLE
481	default y
482	help
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.
486#
487# Select this config option from the architecture Kconfig, if it is preferred
488# to enable the feature of HugeTLB/dev_dax vmemmap optimization.
489#
490config ARCH_WANT_OPTIMIZE_DAX_VMEMMAP
491	bool
492
493config ARCH_WANT_OPTIMIZE_HUGETLB_VMEMMAP
494	bool
495
496config HAVE_MEMBLOCK_PHYS_MAP
497	bool
498
499config HAVE_FAST_GUP
500	depends on MMU
501	bool
502
503# Don't discard allocated memory used to track "memory" and "reserved" memblocks
504# after early boot, so it can still be used to test for validity of memory.
505# Also, memblocks are updated with memory hot(un)plug.
506config ARCH_KEEP_MEMBLOCK
507	bool
508
509# Keep arch NUMA mapping infrastructure post-init.
510config NUMA_KEEP_MEMINFO
511	bool
512
513config MEMORY_ISOLATION
514	bool
515
516# IORESOURCE_SYSTEM_RAM regions in the kernel resource tree that are marked
517# IORESOURCE_EXCLUSIVE cannot be mapped to user space, for example, via
518# /dev/mem.
519config EXCLUSIVE_SYSTEM_RAM
520	def_bool y
521	depends on !DEVMEM || STRICT_DEVMEM
522
523#
524# Only be set on architectures that have completely implemented memory hotplug
525# feature. If you are not sure, don't touch it.
526#
527config HAVE_BOOTMEM_INFO_NODE
528	def_bool n
529
530config ARCH_ENABLE_MEMORY_HOTPLUG
531	bool
532
533config ARCH_ENABLE_MEMORY_HOTREMOVE
534	bool
535
536# eventually, we can have this option just 'select SPARSEMEM'
537menuconfig MEMORY_HOTPLUG
538	bool "Memory hotplug"
539	select MEMORY_ISOLATION
540	depends on SPARSEMEM
541	depends on ARCH_ENABLE_MEMORY_HOTPLUG
542	depends on 64BIT
543	select NUMA_KEEP_MEMINFO if NUMA
544
545if MEMORY_HOTPLUG
546
547config MEMORY_HOTPLUG_DEFAULT_ONLINE
548	bool "Online the newly added memory blocks by default"
549	depends on MEMORY_HOTPLUG
550	help
551	  This option sets the default policy setting for memory hotplug
552	  onlining policy (/sys/devices/system/memory/auto_online_blocks) which
553	  determines what happens to newly added memory regions. Policy setting
554	  can always be changed at runtime.
555	  See Documentation/admin-guide/mm/memory-hotplug.rst for more information.
556
557	  Say Y here if you want all hot-plugged memory blocks to appear in
558	  'online' state by default.
559	  Say N here if you want the default policy to keep all hot-plugged
560	  memory blocks in 'offline' state.
561
562config MEMORY_HOTREMOVE
563	bool "Allow for memory hot remove"
564	select HAVE_BOOTMEM_INFO_NODE if (X86_64 || PPC64)
565	depends on MEMORY_HOTPLUG && ARCH_ENABLE_MEMORY_HOTREMOVE
566	depends on MIGRATION
567
568config MHP_MEMMAP_ON_MEMORY
569	def_bool y
570	depends on MEMORY_HOTPLUG && SPARSEMEM_VMEMMAP
571	depends on ARCH_MHP_MEMMAP_ON_MEMORY_ENABLE
572
573endif # MEMORY_HOTPLUG
574
575# Heavily threaded applications may benefit from splitting the mm-wide
576# page_table_lock, so that faults on different parts of the user address
577# space can be handled with less contention: split it at this NR_CPUS.
578# Default to 4 for wider testing, though 8 might be more appropriate.
579# ARM's adjust_pte (unused if VIPT) depends on mm-wide page_table_lock.
580# PA-RISC 7xxx's spinlock_t would enlarge struct page from 32 to 44 bytes.
581# SPARC32 allocates multiple pte tables within a single page, and therefore
582# a per-page lock leads to problems when multiple tables need to be locked
583# at the same time (e.g. copy_page_range()).
584# DEBUG_SPINLOCK and DEBUG_LOCK_ALLOC spinlock_t also enlarge struct page.
585#
586config SPLIT_PTLOCK_CPUS
587	int
588	default "999999" if !MMU
589	default "999999" if ARM && !CPU_CACHE_VIPT
590	default "999999" if PARISC && !PA20
591	default "999999" if SPARC32
592	default "4"
593
594config ARCH_ENABLE_SPLIT_PMD_PTLOCK
595	bool
596
597#
598# support for memory balloon
599config MEMORY_BALLOON
600	bool
601
602#
603# support for memory balloon compaction
604config BALLOON_COMPACTION
605	bool "Allow for balloon memory compaction/migration"
606	def_bool y
607	depends on COMPACTION && MEMORY_BALLOON
608	help
609	  Memory fragmentation introduced by ballooning might reduce
610	  significantly the number of 2MB contiguous memory blocks that can be
611	  used within a guest, thus imposing performance penalties associated
612	  with the reduced number of transparent huge pages that could be used
613	  by the guest workload. Allowing the compaction & migration for memory
614	  pages enlisted as being part of memory balloon devices avoids the
615	  scenario aforementioned and helps improving memory defragmentation.
616
617#
618# support for memory compaction
619config COMPACTION
620	bool "Allow for memory compaction"
621	def_bool y
622	select MIGRATION
623	depends on MMU
624	help
625	  Compaction is the only memory management component to form
626	  high order (larger physically contiguous) memory blocks
627	  reliably. The page allocator relies on compaction heavily and
628	  the lack of the feature can lead to unexpected OOM killer
629	  invocations for high order memory requests. You shouldn't
630	  disable this option unless there really is a strong reason for
631	  it and then we would be really interested to hear about that at
632	  linux-mm@kvack.org.
633
634config COMPACT_UNEVICTABLE_DEFAULT
635	int
636	depends on COMPACTION
637	default 0 if PREEMPT_RT
638	default 1
639
640#
641# support for free page reporting
642config PAGE_REPORTING
643	bool "Free page reporting"
644	def_bool n
645	help
646	  Free page reporting allows for the incremental acquisition of
647	  free pages from the buddy allocator for the purpose of reporting
648	  those pages to another entity, such as a hypervisor, so that the
649	  memory can be freed within the host for other uses.
650
651#
652# support for page migration
653#
654config MIGRATION
655	bool "Page migration"
656	def_bool y
657	depends on (NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION || CMA) && MMU
658	help
659	  Allows the migration of the physical location of pages of processes
660	  while the virtual addresses are not changed. This is useful in
661	  two situations. The first is on NUMA systems to put pages nearer
662	  to the processors accessing. The second is when allocating huge
663	  pages as migration can relocate pages to satisfy a huge page
664	  allocation instead of reclaiming.
665
666config DEVICE_MIGRATION
667	def_bool MIGRATION && ZONE_DEVICE
668
669config ARCH_ENABLE_HUGEPAGE_MIGRATION
670	bool
671
672config ARCH_ENABLE_THP_MIGRATION
673	bool
674
675config HUGETLB_PAGE_SIZE_VARIABLE
676	def_bool n
677	help
678	  Allows the pageblock_order value to be dynamic instead of just standard
679	  HUGETLB_PAGE_ORDER when there are multiple HugeTLB page sizes available
680	  on a platform.
681
682	  Note that the pageblock_order cannot exceed MAX_ORDER and will be
683	  clamped down to MAX_ORDER.
684
685config CONTIG_ALLOC
686	def_bool (MEMORY_ISOLATION && COMPACTION) || CMA
687
688config PHYS_ADDR_T_64BIT
689	def_bool 64BIT
690
691config BOUNCE
692	bool "Enable bounce buffers"
693	default y
694	depends on BLOCK && MMU && HIGHMEM
695	help
696	  Enable bounce buffers for devices that cannot access the full range of
697	  memory available to the CPU. Enabled by default when HIGHMEM is
698	  selected, but you may say n to override this.
699
700config MMU_NOTIFIER
701	bool
702	select INTERVAL_TREE
703
704config KSM
705	bool "Enable KSM for page merging"
706	depends on MMU
707	select XXHASH
708	help
709	  Enable Kernel Samepage Merging: KSM periodically scans those areas
710	  of an application's address space that an app has advised may be
711	  mergeable.  When it finds pages of identical content, it replaces
712	  the many instances by a single page with that content, so
713	  saving memory until one or another app needs to modify the content.
714	  Recommended for use with KVM, or with other duplicative applications.
715	  See Documentation/mm/ksm.rst for more information: KSM is inactive
716	  until a program has madvised that an area is MADV_MERGEABLE, and
717	  root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set).
718
719config DEFAULT_MMAP_MIN_ADDR
720	int "Low address space to protect from user allocation"
721	depends on MMU
722	default 4096
723	help
724	  This is the portion of low virtual memory which should be protected
725	  from userspace allocation.  Keeping a user from writing to low pages
726	  can help reduce the impact of kernel NULL pointer bugs.
727
728	  For most ia64, ppc64 and x86 users with lots of address space
729	  a value of 65536 is reasonable and should cause no problems.
730	  On arm and other archs it should not be higher than 32768.
731	  Programs which use vm86 functionality or have some need to map
732	  this low address space will need CAP_SYS_RAWIO or disable this
733	  protection by setting the value to 0.
734
735	  This value can be changed after boot using the
736	  /proc/sys/vm/mmap_min_addr tunable.
737
738config ARCH_SUPPORTS_MEMORY_FAILURE
739	bool
740
741config MEMORY_FAILURE
742	depends on MMU
743	depends on ARCH_SUPPORTS_MEMORY_FAILURE
744	bool "Enable recovery from hardware memory errors"
745	select MEMORY_ISOLATION
746	select RAS
747	help
748	  Enables code to recover from some memory failures on systems
749	  with MCA recovery. This allows a system to continue running
750	  even when some of its memory has uncorrected errors. This requires
751	  special hardware support and typically ECC memory.
752
753config HWPOISON_INJECT
754	tristate "HWPoison pages injector"
755	depends on MEMORY_FAILURE && DEBUG_KERNEL && PROC_FS
756	select PROC_PAGE_MONITOR
757
758config NOMMU_INITIAL_TRIM_EXCESS
759	int "Turn on mmap() excess space trimming before booting"
760	depends on !MMU
761	default 1
762	help
763	  The NOMMU mmap() frequently needs to allocate large contiguous chunks
764	  of memory on which to store mappings, but it can only ask the system
765	  allocator for chunks in 2^N*PAGE_SIZE amounts - which is frequently
766	  more than it requires.  To deal with this, mmap() is able to trim off
767	  the excess and return it to the allocator.
768
769	  If trimming is enabled, the excess is trimmed off and returned to the
770	  system allocator, which can cause extra fragmentation, particularly
771	  if there are a lot of transient processes.
772
773	  If trimming is disabled, the excess is kept, but not used, which for
774	  long-term mappings means that the space is wasted.
775
776	  Trimming can be dynamically controlled through a sysctl option
777	  (/proc/sys/vm/nr_trim_pages) which specifies the minimum number of
778	  excess pages there must be before trimming should occur, or zero if
779	  no trimming is to occur.
780
781	  This option specifies the initial value of this option.  The default
782	  of 1 says that all excess pages should be trimmed.
783
784	  See Documentation/admin-guide/mm/nommu-mmap.rst for more information.
785
786config ARCH_WANT_GENERAL_HUGETLB
787	bool
788
789config ARCH_WANTS_THP_SWAP
790	def_bool n
791
792menuconfig TRANSPARENT_HUGEPAGE
793	bool "Transparent Hugepage Support"
794	depends on HAVE_ARCH_TRANSPARENT_HUGEPAGE && !PREEMPT_RT
795	select COMPACTION
796	select XARRAY_MULTI
797	help
798	  Transparent Hugepages allows the kernel to use huge pages and
799	  huge tlb transparently to the applications whenever possible.
800	  This feature can improve computing performance to certain
801	  applications by speeding up page faults during memory
802	  allocation, by reducing the number of tlb misses and by speeding
803	  up the pagetable walking.
804
805	  If memory constrained on embedded, you may want to say N.
806
807if TRANSPARENT_HUGEPAGE
808
809choice
810	prompt "Transparent Hugepage Support sysfs defaults"
811	depends on TRANSPARENT_HUGEPAGE
812	default TRANSPARENT_HUGEPAGE_ALWAYS
813	help
814	  Selects the sysfs defaults for Transparent Hugepage Support.
815
816	config TRANSPARENT_HUGEPAGE_ALWAYS
817		bool "always"
818	help
819	  Enabling Transparent Hugepage always, can increase the
820	  memory footprint of applications without a guaranteed
821	  benefit but it will work automatically for all applications.
822
823	config TRANSPARENT_HUGEPAGE_MADVISE
824		bool "madvise"
825	help
826	  Enabling Transparent Hugepage madvise, will only provide a
827	  performance improvement benefit to the applications using
828	  madvise(MADV_HUGEPAGE) but it won't risk to increase the
829	  memory footprint of applications without a guaranteed
830	  benefit.
831endchoice
832
833config THP_SWAP
834	def_bool y
835	depends on TRANSPARENT_HUGEPAGE && ARCH_WANTS_THP_SWAP && SWAP && 64BIT
836	help
837	  Swap transparent huge pages in one piece, without splitting.
838	  XXX: For now, swap cluster backing transparent huge page
839	  will be split after swapout.
840
841	  For selection by architectures with reasonable THP sizes.
842
843config READ_ONLY_THP_FOR_FS
844	bool "Read-only THP for filesystems (EXPERIMENTAL)"
845	depends on TRANSPARENT_HUGEPAGE && SHMEM
846
847	help
848	  Allow khugepaged to put read-only file-backed pages in THP.
849
850	  This is marked experimental because it is a new feature. Write
851	  support of file THPs will be developed in the next few release
852	  cycles.
853
854endif # TRANSPARENT_HUGEPAGE
855
856#
857# UP and nommu archs use km based percpu allocator
858#
859config NEED_PER_CPU_KM
860	depends on !SMP || !MMU
861	bool
862	default y
863
864config NEED_PER_CPU_EMBED_FIRST_CHUNK
865	bool
866
867config NEED_PER_CPU_PAGE_FIRST_CHUNK
868	bool
869
870config USE_PERCPU_NUMA_NODE_ID
871	bool
872
873config HAVE_SETUP_PER_CPU_AREA
874	bool
875
876config FRONTSWAP
877	bool
878
879config CMA
880	bool "Contiguous Memory Allocator"
881	depends on MMU
882	select MIGRATION
883	select MEMORY_ISOLATION
884	help
885	  This enables the Contiguous Memory Allocator which allows other
886	  subsystems to allocate big physically-contiguous blocks of memory.
887	  CMA reserves a region of memory and allows only movable pages to
888	  be allocated from it. This way, the kernel can use the memory for
889	  pagecache and when a subsystem requests for contiguous area, the
890	  allocated pages are migrated away to serve the contiguous request.
891
892	  If unsure, say "n".
893
894config CMA_DEBUG
895	bool "CMA debug messages (DEVELOPMENT)"
896	depends on DEBUG_KERNEL && CMA
897	help
898	  Turns on debug messages in CMA.  This produces KERN_DEBUG
899	  messages for every CMA call as well as various messages while
900	  processing calls such as dma_alloc_from_contiguous().
901	  This option does not affect warning and error messages.
902
903config CMA_DEBUGFS
904	bool "CMA debugfs interface"
905	depends on CMA && DEBUG_FS
906	help
907	  Turns on the DebugFS interface for CMA.
908
909config CMA_SYSFS
910	bool "CMA information through sysfs interface"
911	depends on CMA && SYSFS
912	help
913	  This option exposes some sysfs attributes to get information
914	  from CMA.
915
916config CMA_AREAS
917	int "Maximum count of the CMA areas"
918	depends on CMA
919	default 19 if NUMA
920	default 7
921	help
922	  CMA allows to create CMA areas for particular purpose, mainly,
923	  used as device private area. This parameter sets the maximum
924	  number of CMA area in the system.
925
926	  If unsure, leave the default value "7" in UMA and "19" in NUMA.
927
928config MEM_SOFT_DIRTY
929	bool "Track memory changes"
930	depends on CHECKPOINT_RESTORE && HAVE_ARCH_SOFT_DIRTY && PROC_FS
931	select PROC_PAGE_MONITOR
932	help
933	  This option enables memory changes tracking by introducing a
934	  soft-dirty bit on pte-s. This bit it set when someone writes
935	  into a page just as regular dirty bit, but unlike the latter
936	  it can be cleared by hands.
937
938	  See Documentation/admin-guide/mm/soft-dirty.rst for more details.
939
940config GENERIC_EARLY_IOREMAP
941	bool
942
943config STACK_MAX_DEFAULT_SIZE_MB
944	int "Default maximum user stack size for 32-bit processes (MB)"
945	default 100
946	range 8 2048
947	depends on STACK_GROWSUP && (!64BIT || COMPAT)
948	help
949	  This is the maximum stack size in Megabytes in the VM layout of 32-bit
950	  user processes when the stack grows upwards (currently only on parisc
951	  arch) when the RLIMIT_STACK hard limit is unlimited.
952
953	  A sane initial value is 100 MB.
954
955config DEFERRED_STRUCT_PAGE_INIT
956	bool "Defer initialisation of struct pages to kthreads"
957	depends on SPARSEMEM
958	depends on !NEED_PER_CPU_KM
959	depends on 64BIT
960	select PADATA
961	help
962	  Ordinarily all struct pages are initialised during early boot in a
963	  single thread. On very large machines this can take a considerable
964	  amount of time. If this option is set, large machines will bring up
965	  a subset of memmap at boot and then initialise the rest in parallel.
966	  This has a potential performance impact on tasks running early in the
967	  lifetime of the system until these kthreads finish the
968	  initialisation.
969
970config PAGE_IDLE_FLAG
971	bool
972	select PAGE_EXTENSION if !64BIT
973	help
974	  This adds PG_idle and PG_young flags to 'struct page'.  PTE Accessed
975	  bit writers can set the state of the bit in the flags so that PTE
976	  Accessed bit readers may avoid disturbance.
977
978config IDLE_PAGE_TRACKING
979	bool "Enable idle page tracking"
980	depends on SYSFS && MMU
981	select PAGE_IDLE_FLAG
982	help
983	  This feature allows to estimate the amount of user pages that have
984	  not been touched during a given period of time. This information can
985	  be useful to tune memory cgroup limits and/or for job placement
986	  within a compute cluster.
987
988	  See Documentation/admin-guide/mm/idle_page_tracking.rst for
989	  more details.
990
991config ARCH_HAS_CACHE_LINE_SIZE
992	bool
993
994config ARCH_HAS_CURRENT_STACK_POINTER
995	bool
996	help
997	  In support of HARDENED_USERCOPY performing stack variable lifetime
998	  checking, an architecture-agnostic way to find the stack pointer
999	  is needed. Once an architecture defines an unsigned long global
1000	  register alias named "current_stack_pointer", this config can be
1001	  selected.
1002
1003config ARCH_HAS_PTE_DEVMAP
1004	bool
1005
1006config ARCH_HAS_ZONE_DMA_SET
1007	bool
1008
1009config ZONE_DMA
1010	bool "Support DMA zone" if ARCH_HAS_ZONE_DMA_SET
1011	default y if ARM64 || X86
1012
1013config ZONE_DMA32
1014	bool "Support DMA32 zone" if ARCH_HAS_ZONE_DMA_SET
1015	depends on !X86_32
1016	default y if ARM64
1017
1018config ZONE_DEVICE
1019	bool "Device memory (pmem, HMM, etc...) hotplug support"
1020	depends on MEMORY_HOTPLUG
1021	depends on MEMORY_HOTREMOVE
1022	depends on SPARSEMEM_VMEMMAP
1023	depends on ARCH_HAS_PTE_DEVMAP
1024	select XARRAY_MULTI
1025
1026	help
1027	  Device memory hotplug support allows for establishing pmem,
1028	  or other device driver discovered memory regions, in the
1029	  memmap. This allows pfn_to_page() lookups of otherwise
1030	  "device-physical" addresses which is needed for using a DAX
1031	  mapping in an O_DIRECT operation, among other things.
1032
1033	  If FS_DAX is enabled, then say Y.
1034
1035#
1036# Helpers to mirror range of the CPU page tables of a process into device page
1037# tables.
1038#
1039config HMM_MIRROR
1040	bool
1041	depends on MMU
1042
1043config GET_FREE_REGION
1044	depends on SPARSEMEM
1045	bool
1046
1047config DEVICE_PRIVATE
1048	bool "Unaddressable device memory (GPU memory, ...)"
1049	depends on ZONE_DEVICE
1050	select GET_FREE_REGION
1051
1052	help
1053	  Allows creation of struct pages to represent unaddressable device
1054	  memory; i.e., memory that is only accessible from the device (or
1055	  group of devices). You likely also want to select HMM_MIRROR.
1056
1057config VMAP_PFN
1058	bool
1059
1060config ARCH_USES_HIGH_VMA_FLAGS
1061	bool
1062config ARCH_HAS_PKEYS
1063	bool
1064
1065config ARCH_USES_PG_ARCH_X
1066	bool
1067	help
1068	  Enable the definition of PG_arch_x page flags with x > 1. Only
1069	  suitable for 64-bit architectures with CONFIG_FLATMEM or
1070	  CONFIG_SPARSEMEM_VMEMMAP enabled, otherwise there may not be
1071	  enough room for additional bits in page->flags.
1072
1073config VM_EVENT_COUNTERS
1074	default y
1075	bool "Enable VM event counters for /proc/vmstat" if EXPERT
1076	help
1077	  VM event counters are needed for event counts to be shown.
1078	  This option allows the disabling of the VM event counters
1079	  on EXPERT systems.  /proc/vmstat will only show page counts
1080	  if VM event counters are disabled.
1081
1082config PERCPU_STATS
1083	bool "Collect percpu memory statistics"
1084	help
1085	  This feature collects and exposes statistics via debugfs. The
1086	  information includes global and per chunk statistics, which can
1087	  be used to help understand percpu memory usage.
1088
1089config GUP_TEST
1090	bool "Enable infrastructure for get_user_pages()-related unit tests"
1091	depends on DEBUG_FS
1092	help
1093	  Provides /sys/kernel/debug/gup_test, which in turn provides a way
1094	  to make ioctl calls that can launch kernel-based unit tests for
1095	  the get_user_pages*() and pin_user_pages*() family of API calls.
1096
1097	  These tests include benchmark testing of the _fast variants of
1098	  get_user_pages*() and pin_user_pages*(), as well as smoke tests of
1099	  the non-_fast variants.
1100
1101	  There is also a sub-test that allows running dump_page() on any
1102	  of up to eight pages (selected by command line args) within the
1103	  range of user-space addresses. These pages are either pinned via
1104	  pin_user_pages*(), or pinned via get_user_pages*(), as specified
1105	  by other command line arguments.
1106
1107	  See tools/testing/selftests/mm/gup_test.c
1108
1109comment "GUP_TEST needs to have DEBUG_FS enabled"
1110	depends on !GUP_TEST && !DEBUG_FS
1111
1112config GUP_GET_PXX_LOW_HIGH
1113	bool
1114
1115config DMAPOOL_TEST
1116	tristate "Enable a module to run time tests on dma_pool"
1117	depends on HAS_DMA
1118	help
1119	  Provides a test module that will allocate and free many blocks of
1120	  various sizes and report how long it takes. This is intended to
1121	  provide a consistent way to measure how changes to the
1122	  dma_pool_alloc/free routines affect performance.
1123
1124config ARCH_HAS_PTE_SPECIAL
1125	bool
1126
1127#
1128# Some architectures require a special hugepage directory format that is
1129# required to support multiple hugepage sizes. For example a4fe3ce76
1130# "powerpc/mm: Allow more flexible layouts for hugepage pagetables"
1131# introduced it on powerpc.  This allows for a more flexible hugepage
1132# pagetable layouts.
1133#
1134config ARCH_HAS_HUGEPD
1135	bool
1136
1137config MAPPING_DIRTY_HELPERS
1138        bool
1139
1140config KMAP_LOCAL
1141	bool
1142
1143config KMAP_LOCAL_NON_LINEAR_PTE_ARRAY
1144	bool
1145
1146# struct io_mapping based helper.  Selected by drivers that need them
1147config IO_MAPPING
1148	bool
1149
1150config MEMFD_CREATE
1151	bool "Enable memfd_create() system call" if EXPERT
1152
1153config SECRETMEM
1154	default y
1155	bool "Enable memfd_secret() system call" if EXPERT
1156	depends on ARCH_HAS_SET_DIRECT_MAP
1157	help
1158	  Enable the memfd_secret() system call with the ability to create
1159	  memory areas visible only in the context of the owning process and
1160	  not mapped to other processes and other kernel page tables.
1161
1162config ANON_VMA_NAME
1163	bool "Anonymous VMA name support"
1164	depends on PROC_FS && ADVISE_SYSCALLS && MMU
1165
1166	help
1167	  Allow naming anonymous virtual memory areas.
1168
1169	  This feature allows assigning names to virtual memory areas. Assigned
1170	  names can be later retrieved from /proc/pid/maps and /proc/pid/smaps
1171	  and help identifying individual anonymous memory areas.
1172	  Assigning a name to anonymous virtual memory area might prevent that
1173	  area from being merged with adjacent virtual memory areas due to the
1174	  difference in their name.
1175
1176config USERFAULTFD
1177	bool "Enable userfaultfd() system call"
1178	depends on MMU
1179	help
1180	  Enable the userfaultfd() system call that allows to intercept and
1181	  handle page faults in userland.
1182
1183config HAVE_ARCH_USERFAULTFD_WP
1184	bool
1185	help
1186	  Arch has userfaultfd write protection support
1187
1188config HAVE_ARCH_USERFAULTFD_MINOR
1189	bool
1190	help
1191	  Arch has userfaultfd minor fault support
1192
1193config PTE_MARKER_UFFD_WP
1194	bool "Userfaultfd write protection support for shmem/hugetlbfs"
1195	default y
1196	depends on HAVE_ARCH_USERFAULTFD_WP
1197
1198	help
1199	  Allows to create marker PTEs for userfaultfd write protection
1200	  purposes.  It is required to enable userfaultfd write protection on
1201	  file-backed memory types like shmem and hugetlbfs.
1202
1203# multi-gen LRU {
1204config LRU_GEN
1205	bool "Multi-Gen LRU"
1206	depends on MMU
1207	# make sure folio->flags has enough spare bits
1208	depends on 64BIT || !SPARSEMEM || SPARSEMEM_VMEMMAP
1209	help
1210	  A high performance LRU implementation to overcommit memory. See
1211	  Documentation/admin-guide/mm/multigen_lru.rst for details.
1212
1213config LRU_GEN_ENABLED
1214	bool "Enable by default"
1215	depends on LRU_GEN
1216	help
1217	  This option enables the multi-gen LRU by default.
1218
1219config LRU_GEN_STATS
1220	bool "Full stats for debugging"
1221	depends on LRU_GEN
1222	help
1223	  Do not enable this option unless you plan to look at historical stats
1224	  from evicted generations for debugging purpose.
1225
1226	  This option has a per-memcg and per-node memory overhead.
1227# }
1228
1229config ARCH_SUPPORTS_PER_VMA_LOCK
1230       def_bool n
1231
1232config PER_VMA_LOCK
1233	def_bool y
1234	depends on ARCH_SUPPORTS_PER_VMA_LOCK && MMU && SMP
1235	help
1236	  Allow per-vma locking during page fault handling.
1237
1238	  This feature allows locking each virtual memory area separately when
1239	  handling page faults instead of taking mmap_lock.
1240
1241config LOCK_MM_AND_FIND_VMA
1242	bool
1243	depends on !STACK_GROWSUP
1244
1245source "mm/damon/Kconfig"
1246
1247endmenu
1248