xref: /openbmc/linux/mm/Kconfig (revision 42c06a0e8ebe95b81e5fb41c6556ff22d9255b0c)
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 CRYPTO
29	select ZPOOL
30	help
31	  A lightweight compressed cache for swap pages.  It takes
32	  pages that are in the process of being swapped out and attempts to
33	  compress them into a dynamically allocated RAM-based memory pool.
34	  This can result in a significant I/O reduction on swap device and,
35	  in the case where decompressing from RAM is faster than swap device
36	  reads, can also improve workload performance.
37
38config ZSWAP_DEFAULT_ON
39	bool "Enable the compressed cache for swap pages by default"
40	depends on ZSWAP
41	help
42	  If selected, the compressed cache for swap pages will be enabled
43	  at boot, otherwise it will be disabled.
44
45	  The selection made here can be overridden by using the kernel
46	  command line 'zswap.enabled=' option.
47
48config ZSWAP_EXCLUSIVE_LOADS_DEFAULT_ON
49	bool "Invalidate zswap entries when pages are loaded"
50	depends on ZSWAP
51	help
52	  If selected, exclusive loads for zswap will be enabled at boot,
53	  otherwise it will be disabled.
54
55	  If exclusive loads are enabled, when a page is loaded from zswap,
56	  the zswap entry is invalidated at once, as opposed to leaving it
57	  in zswap until the swap entry is freed.
58
59	  This avoids having two copies of the same page in memory
60	  (compressed and uncompressed) after faulting in a page from zswap.
61	  The cost is that if the page was never dirtied and needs to be
62	  swapped out again, it will be re-compressed.
63
64choice
65	prompt "Default compressor"
66	depends on ZSWAP
67	default ZSWAP_COMPRESSOR_DEFAULT_LZO
68	help
69	  Selects the default compression algorithm for the compressed cache
70	  for swap pages.
71
72	  For an overview what kind of performance can be expected from
73	  a particular compression algorithm please refer to the benchmarks
74	  available at the following LWN page:
75	  https://lwn.net/Articles/751795/
76
77	  If in doubt, select 'LZO'.
78
79	  The selection made here can be overridden by using the kernel
80	  command line 'zswap.compressor=' option.
81
82config ZSWAP_COMPRESSOR_DEFAULT_DEFLATE
83	bool "Deflate"
84	select CRYPTO_DEFLATE
85	help
86	  Use the Deflate algorithm as the default compression algorithm.
87
88config ZSWAP_COMPRESSOR_DEFAULT_LZO
89	bool "LZO"
90	select CRYPTO_LZO
91	help
92	  Use the LZO algorithm as the default compression algorithm.
93
94config ZSWAP_COMPRESSOR_DEFAULT_842
95	bool "842"
96	select CRYPTO_842
97	help
98	  Use the 842 algorithm as the default compression algorithm.
99
100config ZSWAP_COMPRESSOR_DEFAULT_LZ4
101	bool "LZ4"
102	select CRYPTO_LZ4
103	help
104	  Use the LZ4 algorithm as the default compression algorithm.
105
106config ZSWAP_COMPRESSOR_DEFAULT_LZ4HC
107	bool "LZ4HC"
108	select CRYPTO_LZ4HC
109	help
110	  Use the LZ4HC algorithm as the default compression algorithm.
111
112config ZSWAP_COMPRESSOR_DEFAULT_ZSTD
113	bool "zstd"
114	select CRYPTO_ZSTD
115	help
116	  Use the zstd algorithm as the default compression algorithm.
117endchoice
118
119config ZSWAP_COMPRESSOR_DEFAULT
120       string
121       depends on ZSWAP
122       default "deflate" if ZSWAP_COMPRESSOR_DEFAULT_DEFLATE
123       default "lzo" if ZSWAP_COMPRESSOR_DEFAULT_LZO
124       default "842" if ZSWAP_COMPRESSOR_DEFAULT_842
125       default "lz4" if ZSWAP_COMPRESSOR_DEFAULT_LZ4
126       default "lz4hc" if ZSWAP_COMPRESSOR_DEFAULT_LZ4HC
127       default "zstd" if ZSWAP_COMPRESSOR_DEFAULT_ZSTD
128       default ""
129
130choice
131	prompt "Default allocator"
132	depends on ZSWAP
133	default ZSWAP_ZPOOL_DEFAULT_ZBUD
134	help
135	  Selects the default allocator for the compressed cache for
136	  swap pages.
137	  The default is 'zbud' for compatibility, however please do
138	  read the description of each of the allocators below before
139	  making a right choice.
140
141	  The selection made here can be overridden by using the kernel
142	  command line 'zswap.zpool=' option.
143
144config ZSWAP_ZPOOL_DEFAULT_ZBUD
145	bool "zbud"
146	select ZBUD
147	help
148	  Use the zbud allocator as the default allocator.
149
150config ZSWAP_ZPOOL_DEFAULT_Z3FOLD
151	bool "z3fold"
152	select Z3FOLD
153	help
154	  Use the z3fold allocator as the default allocator.
155
156config ZSWAP_ZPOOL_DEFAULT_ZSMALLOC
157	bool "zsmalloc"
158	select ZSMALLOC
159	help
160	  Use the zsmalloc allocator as the default allocator.
161endchoice
162
163config ZSWAP_ZPOOL_DEFAULT
164       string
165       depends on ZSWAP
166       default "zbud" if ZSWAP_ZPOOL_DEFAULT_ZBUD
167       default "z3fold" if ZSWAP_ZPOOL_DEFAULT_Z3FOLD
168       default "zsmalloc" if ZSWAP_ZPOOL_DEFAULT_ZSMALLOC
169       default ""
170
171config ZBUD
172	tristate "2:1 compression allocator (zbud)"
173	depends on ZSWAP
174	help
175	  A special purpose allocator for storing compressed pages.
176	  It is designed to store up to two compressed pages per physical
177	  page.  While this design limits storage density, it has simple and
178	  deterministic reclaim properties that make it preferable to a higher
179	  density approach when reclaim will be used.
180
181config Z3FOLD
182	tristate "3:1 compression allocator (z3fold)"
183	depends on ZSWAP
184	help
185	  A special purpose allocator for storing compressed pages.
186	  It is designed to store up to three compressed pages per physical
187	  page. It is a ZBUD derivative so the simplicity and determinism are
188	  still there.
189
190config ZSMALLOC
191	tristate
192	prompt "N:1 compression allocator (zsmalloc)" if ZSWAP
193	depends on MMU
194	help
195	  zsmalloc is a slab-based memory allocator designed to store
196	  pages of various compression levels efficiently. It achieves
197	  the highest storage density with the least amount of fragmentation.
198
199config ZSMALLOC_STAT
200	bool "Export zsmalloc statistics"
201	depends on ZSMALLOC
202	select DEBUG_FS
203	help
204	  This option enables code in the zsmalloc to collect various
205	  statistics about what's happening in zsmalloc and exports that
206	  information to userspace via debugfs.
207	  If unsure, say N.
208
209config ZSMALLOC_CHAIN_SIZE
210	int "Maximum number of physical pages per-zspage"
211	default 8
212	range 4 16
213	depends on ZSMALLOC
214	help
215	  This option sets the upper limit on the number of physical pages
216	  that a zmalloc page (zspage) can consist of. The optimal zspage
217	  chain size is calculated for each size class during the
218	  initialization of the pool.
219
220	  Changing this option can alter the characteristics of size classes,
221	  such as the number of pages per zspage and the number of objects
222	  per zspage. This can also result in different configurations of
223	  the pool, as zsmalloc merges size classes with similar
224	  characteristics.
225
226	  For more information, see zsmalloc documentation.
227
228menu "SLAB allocator options"
229
230choice
231	prompt "Choose SLAB allocator"
232	default SLUB
233	help
234	   This option allows to select a slab allocator.
235
236config SLAB_DEPRECATED
237	bool "SLAB (DEPRECATED)"
238	depends on !PREEMPT_RT
239	help
240	  Deprecated and scheduled for removal in a few cycles. Replaced by
241	  SLUB.
242
243	  If you cannot migrate to SLUB, please contact linux-mm@kvack.org
244	  and the people listed in the SLAB ALLOCATOR section of MAINTAINERS
245	  file, explaining why.
246
247	  The regular slab allocator that is established and known to work
248	  well in all environments. It organizes cache hot objects in
249	  per cpu and per node queues.
250
251config SLUB
252	bool "SLUB (Unqueued Allocator)"
253	help
254	   SLUB is a slab allocator that minimizes cache line usage
255	   instead of managing queues of cached objects (SLAB approach).
256	   Per cpu caching is realized using slabs of objects instead
257	   of queues of objects. SLUB can use memory efficiently
258	   and has enhanced diagnostics. SLUB is the default choice for
259	   a slab allocator.
260
261endchoice
262
263config SLAB
264	bool
265	default y
266	depends on SLAB_DEPRECATED
267
268config SLUB_TINY
269	bool "Configure SLUB for minimal memory footprint"
270	depends on SLUB && EXPERT
271	select SLAB_MERGE_DEFAULT
272	help
273	   Configures the SLUB allocator in a way to achieve minimal memory
274	   footprint, sacrificing scalability, debugging and other features.
275	   This is intended only for the smallest system that had used the
276	   SLOB allocator and is not recommended for systems with more than
277	   16MB RAM.
278
279	   If unsure, say N.
280
281config SLAB_MERGE_DEFAULT
282	bool "Allow slab caches to be merged"
283	default y
284	depends on SLAB || SLUB
285	help
286	  For reduced kernel memory fragmentation, slab caches can be
287	  merged when they share the same size and other characteristics.
288	  This carries a risk of kernel heap overflows being able to
289	  overwrite objects from merged caches (and more easily control
290	  cache layout), which makes such heap attacks easier to exploit
291	  by attackers. By keeping caches unmerged, these kinds of exploits
292	  can usually only damage objects in the same cache. To disable
293	  merging at runtime, "slab_nomerge" can be passed on the kernel
294	  command line.
295
296config SLAB_FREELIST_RANDOM
297	bool "Randomize slab freelist"
298	depends on SLAB || (SLUB && !SLUB_TINY)
299	help
300	  Randomizes the freelist order used on creating new pages. This
301	  security feature reduces the predictability of the kernel slab
302	  allocator against heap overflows.
303
304config SLAB_FREELIST_HARDENED
305	bool "Harden slab freelist metadata"
306	depends on SLAB || (SLUB && !SLUB_TINY)
307	help
308	  Many kernel heap attacks try to target slab cache metadata and
309	  other infrastructure. This options makes minor performance
310	  sacrifices to harden the kernel slab allocator against common
311	  freelist exploit methods. Some slab implementations have more
312	  sanity-checking than others. This option is most effective with
313	  CONFIG_SLUB.
314
315config SLUB_STATS
316	default n
317	bool "Enable SLUB performance statistics"
318	depends on SLUB && SYSFS && !SLUB_TINY
319	help
320	  SLUB statistics are useful to debug SLUBs allocation behavior in
321	  order find ways to optimize the allocator. This should never be
322	  enabled for production use since keeping statistics slows down
323	  the allocator by a few percentage points. The slabinfo command
324	  supports the determination of the most active slabs to figure
325	  out which slabs are relevant to a particular load.
326	  Try running: slabinfo -DA
327
328config SLUB_CPU_PARTIAL
329	default y
330	depends on SLUB && SMP && !SLUB_TINY
331	bool "SLUB per cpu partial cache"
332	help
333	  Per cpu partial caches accelerate objects allocation and freeing
334	  that is local to a processor at the price of more indeterminism
335	  in the latency of the free. On overflow these caches will be cleared
336	  which requires the taking of locks that may cause latency spikes.
337	  Typically one would choose no for a realtime system.
338
339endmenu # SLAB allocator options
340
341config SHUFFLE_PAGE_ALLOCATOR
342	bool "Page allocator randomization"
343	default SLAB_FREELIST_RANDOM && ACPI_NUMA
344	help
345	  Randomization of the page allocator improves the average
346	  utilization of a direct-mapped memory-side-cache. See section
347	  5.2.27 Heterogeneous Memory Attribute Table (HMAT) in the ACPI
348	  6.2a specification for an example of how a platform advertises
349	  the presence of a memory-side-cache. There are also incidental
350	  security benefits as it reduces the predictability of page
351	  allocations to compliment SLAB_FREELIST_RANDOM, but the
352	  default granularity of shuffling on the MAX_ORDER i.e, 10th
353	  order of pages is selected based on cache utilization benefits
354	  on x86.
355
356	  While the randomization improves cache utilization it may
357	  negatively impact workloads on platforms without a cache. For
358	  this reason, by default, the randomization is enabled only
359	  after runtime detection of a direct-mapped memory-side-cache.
360	  Otherwise, the randomization may be force enabled with the
361	  'page_alloc.shuffle' kernel command line parameter.
362
363	  Say Y if unsure.
364
365config COMPAT_BRK
366	bool "Disable heap randomization"
367	default y
368	help
369	  Randomizing heap placement makes heap exploits harder, but it
370	  also breaks ancient binaries (including anything libc5 based).
371	  This option changes the bootup default to heap randomization
372	  disabled, and can be overridden at runtime by setting
373	  /proc/sys/kernel/randomize_va_space to 2.
374
375	  On non-ancient distros (post-2000 ones) N is usually a safe choice.
376
377config MMAP_ALLOW_UNINITIALIZED
378	bool "Allow mmapped anonymous memory to be uninitialized"
379	depends on EXPERT && !MMU
380	default n
381	help
382	  Normally, and according to the Linux spec, anonymous memory obtained
383	  from mmap() has its contents cleared before it is passed to
384	  userspace.  Enabling this config option allows you to request that
385	  mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus
386	  providing a huge performance boost.  If this option is not enabled,
387	  then the flag will be ignored.
388
389	  This is taken advantage of by uClibc's malloc(), and also by
390	  ELF-FDPIC binfmt's brk and stack allocator.
391
392	  Because of the obvious security issues, this option should only be
393	  enabled on embedded devices where you control what is run in
394	  userspace.  Since that isn't generally a problem on no-MMU systems,
395	  it is normally safe to say Y here.
396
397	  See Documentation/admin-guide/mm/nommu-mmap.rst for more information.
398
399config SELECT_MEMORY_MODEL
400	def_bool y
401	depends on ARCH_SELECT_MEMORY_MODEL
402
403choice
404	prompt "Memory model"
405	depends on SELECT_MEMORY_MODEL
406	default SPARSEMEM_MANUAL if ARCH_SPARSEMEM_DEFAULT
407	default FLATMEM_MANUAL
408	help
409	  This option allows you to change some of the ways that
410	  Linux manages its memory internally. Most users will
411	  only have one option here selected by the architecture
412	  configuration. This is normal.
413
414config FLATMEM_MANUAL
415	bool "Flat Memory"
416	depends on !ARCH_SPARSEMEM_ENABLE || ARCH_FLATMEM_ENABLE
417	help
418	  This option is best suited for non-NUMA systems with
419	  flat address space. The FLATMEM is the most efficient
420	  system in terms of performance and resource consumption
421	  and it is the best option for smaller systems.
422
423	  For systems that have holes in their physical address
424	  spaces and for features like NUMA and memory hotplug,
425	  choose "Sparse Memory".
426
427	  If unsure, choose this option (Flat Memory) over any other.
428
429config SPARSEMEM_MANUAL
430	bool "Sparse Memory"
431	depends on ARCH_SPARSEMEM_ENABLE
432	help
433	  This will be the only option for some systems, including
434	  memory hot-plug systems.  This is normal.
435
436	  This option provides efficient support for systems with
437	  holes is their physical address space and allows memory
438	  hot-plug and hot-remove.
439
440	  If unsure, choose "Flat Memory" over this option.
441
442endchoice
443
444config SPARSEMEM
445	def_bool y
446	depends on (!SELECT_MEMORY_MODEL && ARCH_SPARSEMEM_ENABLE) || SPARSEMEM_MANUAL
447
448config FLATMEM
449	def_bool y
450	depends on !SPARSEMEM || FLATMEM_MANUAL
451
452#
453# SPARSEMEM_EXTREME (which is the default) does some bootmem
454# allocations when sparse_init() is called.  If this cannot
455# be done on your architecture, select this option.  However,
456# statically allocating the mem_section[] array can potentially
457# consume vast quantities of .bss, so be careful.
458#
459# This option will also potentially produce smaller runtime code
460# with gcc 3.4 and later.
461#
462config SPARSEMEM_STATIC
463	bool
464
465#
466# Architecture platforms which require a two level mem_section in SPARSEMEM
467# must select this option. This is usually for architecture platforms with
468# an extremely sparse physical address space.
469#
470config SPARSEMEM_EXTREME
471	def_bool y
472	depends on SPARSEMEM && !SPARSEMEM_STATIC
473
474config SPARSEMEM_VMEMMAP_ENABLE
475	bool
476
477config SPARSEMEM_VMEMMAP
478	bool "Sparse Memory virtual memmap"
479	depends on SPARSEMEM && SPARSEMEM_VMEMMAP_ENABLE
480	default y
481	help
482	  SPARSEMEM_VMEMMAP uses a virtually mapped memmap to optimise
483	  pfn_to_page and page_to_pfn operations.  This is the most
484	  efficient option when sufficient kernel resources are available.
485#
486# Select this config option from the architecture Kconfig, if it is preferred
487# to enable the feature of HugeTLB/dev_dax vmemmap optimization.
488#
489config ARCH_WANT_OPTIMIZE_DAX_VMEMMAP
490	bool
491
492config ARCH_WANT_OPTIMIZE_HUGETLB_VMEMMAP
493	bool
494
495config HAVE_MEMBLOCK_PHYS_MAP
496	bool
497
498config HAVE_FAST_GUP
499	depends on MMU
500	bool
501
502# Don't discard allocated memory used to track "memory" and "reserved" memblocks
503# after early boot, so it can still be used to test for validity of memory.
504# Also, memblocks are updated with memory hot(un)plug.
505config ARCH_KEEP_MEMBLOCK
506	bool
507
508# Keep arch NUMA mapping infrastructure post-init.
509config NUMA_KEEP_MEMINFO
510	bool
511
512config MEMORY_ISOLATION
513	bool
514
515# IORESOURCE_SYSTEM_RAM regions in the kernel resource tree that are marked
516# IORESOURCE_EXCLUSIVE cannot be mapped to user space, for example, via
517# /dev/mem.
518config EXCLUSIVE_SYSTEM_RAM
519	def_bool y
520	depends on !DEVMEM || STRICT_DEVMEM
521
522#
523# Only be set on architectures that have completely implemented memory hotplug
524# feature. If you are not sure, don't touch it.
525#
526config HAVE_BOOTMEM_INFO_NODE
527	def_bool n
528
529config ARCH_ENABLE_MEMORY_HOTPLUG
530	bool
531
532config ARCH_ENABLE_MEMORY_HOTREMOVE
533	bool
534
535# eventually, we can have this option just 'select SPARSEMEM'
536menuconfig MEMORY_HOTPLUG
537	bool "Memory hotplug"
538	select MEMORY_ISOLATION
539	depends on SPARSEMEM
540	depends on ARCH_ENABLE_MEMORY_HOTPLUG
541	depends on 64BIT
542	select NUMA_KEEP_MEMINFO if NUMA
543
544if MEMORY_HOTPLUG
545
546config MEMORY_HOTPLUG_DEFAULT_ONLINE
547	bool "Online the newly added memory blocks by default"
548	depends on MEMORY_HOTPLUG
549	help
550	  This option sets the default policy setting for memory hotplug
551	  onlining policy (/sys/devices/system/memory/auto_online_blocks) which
552	  determines what happens to newly added memory regions. Policy setting
553	  can always be changed at runtime.
554	  See Documentation/admin-guide/mm/memory-hotplug.rst for more information.
555
556	  Say Y here if you want all hot-plugged memory blocks to appear in
557	  'online' state by default.
558	  Say N here if you want the default policy to keep all hot-plugged
559	  memory blocks in 'offline' state.
560
561config MEMORY_HOTREMOVE
562	bool "Allow for memory hot remove"
563	select HAVE_BOOTMEM_INFO_NODE if (X86_64 || PPC64)
564	depends on MEMORY_HOTPLUG && ARCH_ENABLE_MEMORY_HOTREMOVE
565	depends on MIGRATION
566
567config MHP_MEMMAP_ON_MEMORY
568	def_bool y
569	depends on MEMORY_HOTPLUG && SPARSEMEM_VMEMMAP
570	depends on ARCH_MHP_MEMMAP_ON_MEMORY_ENABLE
571
572endif # MEMORY_HOTPLUG
573
574# Heavily threaded applications may benefit from splitting the mm-wide
575# page_table_lock, so that faults on different parts of the user address
576# space can be handled with less contention: split it at this NR_CPUS.
577# Default to 4 for wider testing, though 8 might be more appropriate.
578# ARM's adjust_pte (unused if VIPT) depends on mm-wide page_table_lock.
579# PA-RISC 7xxx's spinlock_t would enlarge struct page from 32 to 44 bytes.
580# SPARC32 allocates multiple pte tables within a single page, and therefore
581# a per-page lock leads to problems when multiple tables need to be locked
582# at the same time (e.g. copy_page_range()).
583# DEBUG_SPINLOCK and DEBUG_LOCK_ALLOC spinlock_t also enlarge struct page.
584#
585config SPLIT_PTLOCK_CPUS
586	int
587	default "999999" if !MMU
588	default "999999" if ARM && !CPU_CACHE_VIPT
589	default "999999" if PARISC && !PA20
590	default "999999" if SPARC32
591	default "4"
592
593config ARCH_ENABLE_SPLIT_PMD_PTLOCK
594	bool
595
596#
597# support for memory balloon
598config MEMORY_BALLOON
599	bool
600
601#
602# support for memory balloon compaction
603config BALLOON_COMPACTION
604	bool "Allow for balloon memory compaction/migration"
605	def_bool y
606	depends on COMPACTION && MEMORY_BALLOON
607	help
608	  Memory fragmentation introduced by ballooning might reduce
609	  significantly the number of 2MB contiguous memory blocks that can be
610	  used within a guest, thus imposing performance penalties associated
611	  with the reduced number of transparent huge pages that could be used
612	  by the guest workload. Allowing the compaction & migration for memory
613	  pages enlisted as being part of memory balloon devices avoids the
614	  scenario aforementioned and helps improving memory defragmentation.
615
616#
617# support for memory compaction
618config COMPACTION
619	bool "Allow for memory compaction"
620	def_bool y
621	select MIGRATION
622	depends on MMU
623	help
624	  Compaction is the only memory management component to form
625	  high order (larger physically contiguous) memory blocks
626	  reliably. The page allocator relies on compaction heavily and
627	  the lack of the feature can lead to unexpected OOM killer
628	  invocations for high order memory requests. You shouldn't
629	  disable this option unless there really is a strong reason for
630	  it and then we would be really interested to hear about that at
631	  linux-mm@kvack.org.
632
633config COMPACT_UNEVICTABLE_DEFAULT
634	int
635	depends on COMPACTION
636	default 0 if PREEMPT_RT
637	default 1
638
639#
640# support for free page reporting
641config PAGE_REPORTING
642	bool "Free page reporting"
643	def_bool n
644	help
645	  Free page reporting allows for the incremental acquisition of
646	  free pages from the buddy allocator for the purpose of reporting
647	  those pages to another entity, such as a hypervisor, so that the
648	  memory can be freed within the host for other uses.
649
650#
651# support for page migration
652#
653config MIGRATION
654	bool "Page migration"
655	def_bool y
656	depends on (NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION || CMA) && MMU
657	help
658	  Allows the migration of the physical location of pages of processes
659	  while the virtual addresses are not changed. This is useful in
660	  two situations. The first is on NUMA systems to put pages nearer
661	  to the processors accessing. The second is when allocating huge
662	  pages as migration can relocate pages to satisfy a huge page
663	  allocation instead of reclaiming.
664
665config DEVICE_MIGRATION
666	def_bool MIGRATION && ZONE_DEVICE
667
668config ARCH_ENABLE_HUGEPAGE_MIGRATION
669	bool
670
671config ARCH_ENABLE_THP_MIGRATION
672	bool
673
674config HUGETLB_PAGE_SIZE_VARIABLE
675	def_bool n
676	help
677	  Allows the pageblock_order value to be dynamic instead of just standard
678	  HUGETLB_PAGE_ORDER when there are multiple HugeTLB page sizes available
679	  on a platform.
680
681	  Note that the pageblock_order cannot exceed MAX_ORDER and will be
682	  clamped down to MAX_ORDER.
683
684config CONTIG_ALLOC
685	def_bool (MEMORY_ISOLATION && COMPACTION) || CMA
686
687config PHYS_ADDR_T_64BIT
688	def_bool 64BIT
689
690config BOUNCE
691	bool "Enable bounce buffers"
692	default y
693	depends on BLOCK && MMU && HIGHMEM
694	help
695	  Enable bounce buffers for devices that cannot access the full range of
696	  memory available to the CPU. Enabled by default when HIGHMEM is
697	  selected, but you may say n to override this.
698
699config MMU_NOTIFIER
700	bool
701	select INTERVAL_TREE
702
703config KSM
704	bool "Enable KSM for page merging"
705	depends on MMU
706	select XXHASH
707	help
708	  Enable Kernel Samepage Merging: KSM periodically scans those areas
709	  of an application's address space that an app has advised may be
710	  mergeable.  When it finds pages of identical content, it replaces
711	  the many instances by a single page with that content, so
712	  saving memory until one or another app needs to modify the content.
713	  Recommended for use with KVM, or with other duplicative applications.
714	  See Documentation/mm/ksm.rst for more information: KSM is inactive
715	  until a program has madvised that an area is MADV_MERGEABLE, and
716	  root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set).
717
718config DEFAULT_MMAP_MIN_ADDR
719	int "Low address space to protect from user allocation"
720	depends on MMU
721	default 4096
722	help
723	  This is the portion of low virtual memory which should be protected
724	  from userspace allocation.  Keeping a user from writing to low pages
725	  can help reduce the impact of kernel NULL pointer bugs.
726
727	  For most ia64, ppc64 and x86 users with lots of address space
728	  a value of 65536 is reasonable and should cause no problems.
729	  On arm and other archs it should not be higher than 32768.
730	  Programs which use vm86 functionality or have some need to map
731	  this low address space will need CAP_SYS_RAWIO or disable this
732	  protection by setting the value to 0.
733
734	  This value can be changed after boot using the
735	  /proc/sys/vm/mmap_min_addr tunable.
736
737config ARCH_SUPPORTS_MEMORY_FAILURE
738	bool
739
740config MEMORY_FAILURE
741	depends on MMU
742	depends on ARCH_SUPPORTS_MEMORY_FAILURE
743	bool "Enable recovery from hardware memory errors"
744	select MEMORY_ISOLATION
745	select RAS
746	help
747	  Enables code to recover from some memory failures on systems
748	  with MCA recovery. This allows a system to continue running
749	  even when some of its memory has uncorrected errors. This requires
750	  special hardware support and typically ECC memory.
751
752config HWPOISON_INJECT
753	tristate "HWPoison pages injector"
754	depends on MEMORY_FAILURE && DEBUG_KERNEL && PROC_FS
755	select PROC_PAGE_MONITOR
756
757config NOMMU_INITIAL_TRIM_EXCESS
758	int "Turn on mmap() excess space trimming before booting"
759	depends on !MMU
760	default 1
761	help
762	  The NOMMU mmap() frequently needs to allocate large contiguous chunks
763	  of memory on which to store mappings, but it can only ask the system
764	  allocator for chunks in 2^N*PAGE_SIZE amounts - which is frequently
765	  more than it requires.  To deal with this, mmap() is able to trim off
766	  the excess and return it to the allocator.
767
768	  If trimming is enabled, the excess is trimmed off and returned to the
769	  system allocator, which can cause extra fragmentation, particularly
770	  if there are a lot of transient processes.
771
772	  If trimming is disabled, the excess is kept, but not used, which for
773	  long-term mappings means that the space is wasted.
774
775	  Trimming can be dynamically controlled through a sysctl option
776	  (/proc/sys/vm/nr_trim_pages) which specifies the minimum number of
777	  excess pages there must be before trimming should occur, or zero if
778	  no trimming is to occur.
779
780	  This option specifies the initial value of this option.  The default
781	  of 1 says that all excess pages should be trimmed.
782
783	  See Documentation/admin-guide/mm/nommu-mmap.rst for more information.
784
785config ARCH_WANT_GENERAL_HUGETLB
786	bool
787
788config ARCH_WANTS_THP_SWAP
789	def_bool n
790
791menuconfig TRANSPARENT_HUGEPAGE
792	bool "Transparent Hugepage Support"
793	depends on HAVE_ARCH_TRANSPARENT_HUGEPAGE && !PREEMPT_RT
794	select COMPACTION
795	select XARRAY_MULTI
796	help
797	  Transparent Hugepages allows the kernel to use huge pages and
798	  huge tlb transparently to the applications whenever possible.
799	  This feature can improve computing performance to certain
800	  applications by speeding up page faults during memory
801	  allocation, by reducing the number of tlb misses and by speeding
802	  up the pagetable walking.
803
804	  If memory constrained on embedded, you may want to say N.
805
806if TRANSPARENT_HUGEPAGE
807
808choice
809	prompt "Transparent Hugepage Support sysfs defaults"
810	depends on TRANSPARENT_HUGEPAGE
811	default TRANSPARENT_HUGEPAGE_ALWAYS
812	help
813	  Selects the sysfs defaults for Transparent Hugepage Support.
814
815	config TRANSPARENT_HUGEPAGE_ALWAYS
816		bool "always"
817	help
818	  Enabling Transparent Hugepage always, can increase the
819	  memory footprint of applications without a guaranteed
820	  benefit but it will work automatically for all applications.
821
822	config TRANSPARENT_HUGEPAGE_MADVISE
823		bool "madvise"
824	help
825	  Enabling Transparent Hugepage madvise, will only provide a
826	  performance improvement benefit to the applications using
827	  madvise(MADV_HUGEPAGE) but it won't risk to increase the
828	  memory footprint of applications without a guaranteed
829	  benefit.
830endchoice
831
832config THP_SWAP
833	def_bool y
834	depends on TRANSPARENT_HUGEPAGE && ARCH_WANTS_THP_SWAP && SWAP && 64BIT
835	help
836	  Swap transparent huge pages in one piece, without splitting.
837	  XXX: For now, swap cluster backing transparent huge page
838	  will be split after swapout.
839
840	  For selection by architectures with reasonable THP sizes.
841
842config READ_ONLY_THP_FOR_FS
843	bool "Read-only THP for filesystems (EXPERIMENTAL)"
844	depends on TRANSPARENT_HUGEPAGE && SHMEM
845
846	help
847	  Allow khugepaged to put read-only file-backed pages in THP.
848
849	  This is marked experimental because it is a new feature. Write
850	  support of file THPs will be developed in the next few release
851	  cycles.
852
853endif # TRANSPARENT_HUGEPAGE
854
855#
856# UP and nommu archs use km based percpu allocator
857#
858config NEED_PER_CPU_KM
859	depends on !SMP || !MMU
860	bool
861	default y
862
863config NEED_PER_CPU_EMBED_FIRST_CHUNK
864	bool
865
866config NEED_PER_CPU_PAGE_FIRST_CHUNK
867	bool
868
869config USE_PERCPU_NUMA_NODE_ID
870	bool
871
872config HAVE_SETUP_PER_CPU_AREA
873	bool
874
875config CMA
876	bool "Contiguous Memory Allocator"
877	depends on MMU
878	select MIGRATION
879	select MEMORY_ISOLATION
880	help
881	  This enables the Contiguous Memory Allocator which allows other
882	  subsystems to allocate big physically-contiguous blocks of memory.
883	  CMA reserves a region of memory and allows only movable pages to
884	  be allocated from it. This way, the kernel can use the memory for
885	  pagecache and when a subsystem requests for contiguous area, the
886	  allocated pages are migrated away to serve the contiguous request.
887
888	  If unsure, say "n".
889
890config CMA_DEBUG
891	bool "CMA debug messages (DEVELOPMENT)"
892	depends on DEBUG_KERNEL && CMA
893	help
894	  Turns on debug messages in CMA.  This produces KERN_DEBUG
895	  messages for every CMA call as well as various messages while
896	  processing calls such as dma_alloc_from_contiguous().
897	  This option does not affect warning and error messages.
898
899config CMA_DEBUGFS
900	bool "CMA debugfs interface"
901	depends on CMA && DEBUG_FS
902	help
903	  Turns on the DebugFS interface for CMA.
904
905config CMA_SYSFS
906	bool "CMA information through sysfs interface"
907	depends on CMA && SYSFS
908	help
909	  This option exposes some sysfs attributes to get information
910	  from CMA.
911
912config CMA_AREAS
913	int "Maximum count of the CMA areas"
914	depends on CMA
915	default 19 if NUMA
916	default 7
917	help
918	  CMA allows to create CMA areas for particular purpose, mainly,
919	  used as device private area. This parameter sets the maximum
920	  number of CMA area in the system.
921
922	  If unsure, leave the default value "7" in UMA and "19" in NUMA.
923
924config MEM_SOFT_DIRTY
925	bool "Track memory changes"
926	depends on CHECKPOINT_RESTORE && HAVE_ARCH_SOFT_DIRTY && PROC_FS
927	select PROC_PAGE_MONITOR
928	help
929	  This option enables memory changes tracking by introducing a
930	  soft-dirty bit on pte-s. This bit it set when someone writes
931	  into a page just as regular dirty bit, but unlike the latter
932	  it can be cleared by hands.
933
934	  See Documentation/admin-guide/mm/soft-dirty.rst for more details.
935
936config GENERIC_EARLY_IOREMAP
937	bool
938
939config STACK_MAX_DEFAULT_SIZE_MB
940	int "Default maximum user stack size for 32-bit processes (MB)"
941	default 100
942	range 8 2048
943	depends on STACK_GROWSUP && (!64BIT || COMPAT)
944	help
945	  This is the maximum stack size in Megabytes in the VM layout of 32-bit
946	  user processes when the stack grows upwards (currently only on parisc
947	  arch) when the RLIMIT_STACK hard limit is unlimited.
948
949	  A sane initial value is 100 MB.
950
951config DEFERRED_STRUCT_PAGE_INIT
952	bool "Defer initialisation of struct pages to kthreads"
953	depends on SPARSEMEM
954	depends on !NEED_PER_CPU_KM
955	depends on 64BIT
956	select PADATA
957	help
958	  Ordinarily all struct pages are initialised during early boot in a
959	  single thread. On very large machines this can take a considerable
960	  amount of time. If this option is set, large machines will bring up
961	  a subset of memmap at boot and then initialise the rest in parallel.
962	  This has a potential performance impact on tasks running early in the
963	  lifetime of the system until these kthreads finish the
964	  initialisation.
965
966config PAGE_IDLE_FLAG
967	bool
968	select PAGE_EXTENSION if !64BIT
969	help
970	  This adds PG_idle and PG_young flags to 'struct page'.  PTE Accessed
971	  bit writers can set the state of the bit in the flags so that PTE
972	  Accessed bit readers may avoid disturbance.
973
974config IDLE_PAGE_TRACKING
975	bool "Enable idle page tracking"
976	depends on SYSFS && MMU
977	select PAGE_IDLE_FLAG
978	help
979	  This feature allows to estimate the amount of user pages that have
980	  not been touched during a given period of time. This information can
981	  be useful to tune memory cgroup limits and/or for job placement
982	  within a compute cluster.
983
984	  See Documentation/admin-guide/mm/idle_page_tracking.rst for
985	  more details.
986
987config ARCH_HAS_CACHE_LINE_SIZE
988	bool
989
990config ARCH_HAS_CURRENT_STACK_POINTER
991	bool
992	help
993	  In support of HARDENED_USERCOPY performing stack variable lifetime
994	  checking, an architecture-agnostic way to find the stack pointer
995	  is needed. Once an architecture defines an unsigned long global
996	  register alias named "current_stack_pointer", this config can be
997	  selected.
998
999config ARCH_HAS_PTE_DEVMAP
1000	bool
1001
1002config ARCH_HAS_ZONE_DMA_SET
1003	bool
1004
1005config ZONE_DMA
1006	bool "Support DMA zone" if ARCH_HAS_ZONE_DMA_SET
1007	default y if ARM64 || X86
1008
1009config ZONE_DMA32
1010	bool "Support DMA32 zone" if ARCH_HAS_ZONE_DMA_SET
1011	depends on !X86_32
1012	default y if ARM64
1013
1014config ZONE_DEVICE
1015	bool "Device memory (pmem, HMM, etc...) hotplug support"
1016	depends on MEMORY_HOTPLUG
1017	depends on MEMORY_HOTREMOVE
1018	depends on SPARSEMEM_VMEMMAP
1019	depends on ARCH_HAS_PTE_DEVMAP
1020	select XARRAY_MULTI
1021
1022	help
1023	  Device memory hotplug support allows for establishing pmem,
1024	  or other device driver discovered memory regions, in the
1025	  memmap. This allows pfn_to_page() lookups of otherwise
1026	  "device-physical" addresses which is needed for using a DAX
1027	  mapping in an O_DIRECT operation, among other things.
1028
1029	  If FS_DAX is enabled, then say Y.
1030
1031#
1032# Helpers to mirror range of the CPU page tables of a process into device page
1033# tables.
1034#
1035config HMM_MIRROR
1036	bool
1037	depends on MMU
1038
1039config GET_FREE_REGION
1040	depends on SPARSEMEM
1041	bool
1042
1043config DEVICE_PRIVATE
1044	bool "Unaddressable device memory (GPU memory, ...)"
1045	depends on ZONE_DEVICE
1046	select GET_FREE_REGION
1047
1048	help
1049	  Allows creation of struct pages to represent unaddressable device
1050	  memory; i.e., memory that is only accessible from the device (or
1051	  group of devices). You likely also want to select HMM_MIRROR.
1052
1053config VMAP_PFN
1054	bool
1055
1056config ARCH_USES_HIGH_VMA_FLAGS
1057	bool
1058config ARCH_HAS_PKEYS
1059	bool
1060
1061config ARCH_USES_PG_ARCH_X
1062	bool
1063	help
1064	  Enable the definition of PG_arch_x page flags with x > 1. Only
1065	  suitable for 64-bit architectures with CONFIG_FLATMEM or
1066	  CONFIG_SPARSEMEM_VMEMMAP enabled, otherwise there may not be
1067	  enough room for additional bits in page->flags.
1068
1069config VM_EVENT_COUNTERS
1070	default y
1071	bool "Enable VM event counters for /proc/vmstat" if EXPERT
1072	help
1073	  VM event counters are needed for event counts to be shown.
1074	  This option allows the disabling of the VM event counters
1075	  on EXPERT systems.  /proc/vmstat will only show page counts
1076	  if VM event counters are disabled.
1077
1078config PERCPU_STATS
1079	bool "Collect percpu memory statistics"
1080	help
1081	  This feature collects and exposes statistics via debugfs. The
1082	  information includes global and per chunk statistics, which can
1083	  be used to help understand percpu memory usage.
1084
1085config GUP_TEST
1086	bool "Enable infrastructure for get_user_pages()-related unit tests"
1087	depends on DEBUG_FS
1088	help
1089	  Provides /sys/kernel/debug/gup_test, which in turn provides a way
1090	  to make ioctl calls that can launch kernel-based unit tests for
1091	  the get_user_pages*() and pin_user_pages*() family of API calls.
1092
1093	  These tests include benchmark testing of the _fast variants of
1094	  get_user_pages*() and pin_user_pages*(), as well as smoke tests of
1095	  the non-_fast variants.
1096
1097	  There is also a sub-test that allows running dump_page() on any
1098	  of up to eight pages (selected by command line args) within the
1099	  range of user-space addresses. These pages are either pinned via
1100	  pin_user_pages*(), or pinned via get_user_pages*(), as specified
1101	  by other command line arguments.
1102
1103	  See tools/testing/selftests/mm/gup_test.c
1104
1105comment "GUP_TEST needs to have DEBUG_FS enabled"
1106	depends on !GUP_TEST && !DEBUG_FS
1107
1108config GUP_GET_PXX_LOW_HIGH
1109	bool
1110
1111config DMAPOOL_TEST
1112	tristate "Enable a module to run time tests on dma_pool"
1113	depends on HAS_DMA
1114	help
1115	  Provides a test module that will allocate and free many blocks of
1116	  various sizes and report how long it takes. This is intended to
1117	  provide a consistent way to measure how changes to the
1118	  dma_pool_alloc/free routines affect performance.
1119
1120config ARCH_HAS_PTE_SPECIAL
1121	bool
1122
1123#
1124# Some architectures require a special hugepage directory format that is
1125# required to support multiple hugepage sizes. For example a4fe3ce76
1126# "powerpc/mm: Allow more flexible layouts for hugepage pagetables"
1127# introduced it on powerpc.  This allows for a more flexible hugepage
1128# pagetable layouts.
1129#
1130config ARCH_HAS_HUGEPD
1131	bool
1132
1133config MAPPING_DIRTY_HELPERS
1134        bool
1135
1136config KMAP_LOCAL
1137	bool
1138
1139config KMAP_LOCAL_NON_LINEAR_PTE_ARRAY
1140	bool
1141
1142# struct io_mapping based helper.  Selected by drivers that need them
1143config IO_MAPPING
1144	bool
1145
1146config MEMFD_CREATE
1147	bool "Enable memfd_create() system call" if EXPERT
1148
1149config SECRETMEM
1150	default y
1151	bool "Enable memfd_secret() system call" if EXPERT
1152	depends on ARCH_HAS_SET_DIRECT_MAP
1153	help
1154	  Enable the memfd_secret() system call with the ability to create
1155	  memory areas visible only in the context of the owning process and
1156	  not mapped to other processes and other kernel page tables.
1157
1158config ANON_VMA_NAME
1159	bool "Anonymous VMA name support"
1160	depends on PROC_FS && ADVISE_SYSCALLS && MMU
1161
1162	help
1163	  Allow naming anonymous virtual memory areas.
1164
1165	  This feature allows assigning names to virtual memory areas. Assigned
1166	  names can be later retrieved from /proc/pid/maps and /proc/pid/smaps
1167	  and help identifying individual anonymous memory areas.
1168	  Assigning a name to anonymous virtual memory area might prevent that
1169	  area from being merged with adjacent virtual memory areas due to the
1170	  difference in their name.
1171
1172config USERFAULTFD
1173	bool "Enable userfaultfd() system call"
1174	depends on MMU
1175	help
1176	  Enable the userfaultfd() system call that allows to intercept and
1177	  handle page faults in userland.
1178
1179config HAVE_ARCH_USERFAULTFD_WP
1180	bool
1181	help
1182	  Arch has userfaultfd write protection support
1183
1184config HAVE_ARCH_USERFAULTFD_MINOR
1185	bool
1186	help
1187	  Arch has userfaultfd minor fault support
1188
1189config PTE_MARKER_UFFD_WP
1190	bool "Userfaultfd write protection support for shmem/hugetlbfs"
1191	default y
1192	depends on HAVE_ARCH_USERFAULTFD_WP
1193
1194	help
1195	  Allows to create marker PTEs for userfaultfd write protection
1196	  purposes.  It is required to enable userfaultfd write protection on
1197	  file-backed memory types like shmem and hugetlbfs.
1198
1199# multi-gen LRU {
1200config LRU_GEN
1201	bool "Multi-Gen LRU"
1202	depends on MMU
1203	# make sure folio->flags has enough spare bits
1204	depends on 64BIT || !SPARSEMEM || SPARSEMEM_VMEMMAP
1205	help
1206	  A high performance LRU implementation to overcommit memory. See
1207	  Documentation/admin-guide/mm/multigen_lru.rst for details.
1208
1209config LRU_GEN_ENABLED
1210	bool "Enable by default"
1211	depends on LRU_GEN
1212	help
1213	  This option enables the multi-gen LRU by default.
1214
1215config LRU_GEN_STATS
1216	bool "Full stats for debugging"
1217	depends on LRU_GEN
1218	help
1219	  Do not enable this option unless you plan to look at historical stats
1220	  from evicted generations for debugging purpose.
1221
1222	  This option has a per-memcg and per-node memory overhead.
1223# }
1224
1225config ARCH_SUPPORTS_PER_VMA_LOCK
1226       def_bool n
1227
1228config PER_VMA_LOCK
1229	def_bool y
1230	depends on ARCH_SUPPORTS_PER_VMA_LOCK && MMU && SMP
1231	help
1232	  Allow per-vma locking during page fault handling.
1233
1234	  This feature allows locking each virtual memory area separately when
1235	  handling page faults instead of taking mmap_lock.
1236
1237config LOCK_MM_AND_FIND_VMA
1238	bool
1239	depends on !STACK_GROWSUP
1240
1241source "mm/damon/Kconfig"
1242
1243endmenu
1244