xref: /openbmc/linux/mm/Kconfig (revision 3c6152940584290668b35fa0800026f6a1ae05fe)
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
340config RANDOM_KMALLOC_CACHES
341	default n
342	depends on SLUB && !SLUB_TINY
343	bool "Randomize slab caches for normal kmalloc"
344	help
345	  A hardening feature that creates multiple copies of slab caches for
346	  normal kmalloc allocation and makes kmalloc randomly pick one based
347	  on code address, which makes the attackers more difficult to spray
348	  vulnerable memory objects on the heap for the purpose of exploiting
349	  memory vulnerabilities.
350
351	  Currently the number of copies is set to 16, a reasonably large value
352	  that effectively diverges the memory objects allocated for different
353	  subsystems or modules into different caches, at the expense of a
354	  limited degree of memory and CPU overhead that relates to hardware and
355	  system workload.
356
357endmenu # SLAB allocator options
358
359config SHUFFLE_PAGE_ALLOCATOR
360	bool "Page allocator randomization"
361	default SLAB_FREELIST_RANDOM && ACPI_NUMA
362	help
363	  Randomization of the page allocator improves the average
364	  utilization of a direct-mapped memory-side-cache. See section
365	  5.2.27 Heterogeneous Memory Attribute Table (HMAT) in the ACPI
366	  6.2a specification for an example of how a platform advertises
367	  the presence of a memory-side-cache. There are also incidental
368	  security benefits as it reduces the predictability of page
369	  allocations to compliment SLAB_FREELIST_RANDOM, but the
370	  default granularity of shuffling on the MAX_ORDER i.e, 10th
371	  order of pages is selected based on cache utilization benefits
372	  on x86.
373
374	  While the randomization improves cache utilization it may
375	  negatively impact workloads on platforms without a cache. For
376	  this reason, by default, the randomization is enabled only
377	  after runtime detection of a direct-mapped memory-side-cache.
378	  Otherwise, the randomization may be force enabled with the
379	  'page_alloc.shuffle' kernel command line parameter.
380
381	  Say Y if unsure.
382
383config COMPAT_BRK
384	bool "Disable heap randomization"
385	default y
386	help
387	  Randomizing heap placement makes heap exploits harder, but it
388	  also breaks ancient binaries (including anything libc5 based).
389	  This option changes the bootup default to heap randomization
390	  disabled, and can be overridden at runtime by setting
391	  /proc/sys/kernel/randomize_va_space to 2.
392
393	  On non-ancient distros (post-2000 ones) N is usually a safe choice.
394
395config MMAP_ALLOW_UNINITIALIZED
396	bool "Allow mmapped anonymous memory to be uninitialized"
397	depends on EXPERT && !MMU
398	default n
399	help
400	  Normally, and according to the Linux spec, anonymous memory obtained
401	  from mmap() has its contents cleared before it is passed to
402	  userspace.  Enabling this config option allows you to request that
403	  mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus
404	  providing a huge performance boost.  If this option is not enabled,
405	  then the flag will be ignored.
406
407	  This is taken advantage of by uClibc's malloc(), and also by
408	  ELF-FDPIC binfmt's brk and stack allocator.
409
410	  Because of the obvious security issues, this option should only be
411	  enabled on embedded devices where you control what is run in
412	  userspace.  Since that isn't generally a problem on no-MMU systems,
413	  it is normally safe to say Y here.
414
415	  See Documentation/admin-guide/mm/nommu-mmap.rst for more information.
416
417config SELECT_MEMORY_MODEL
418	def_bool y
419	depends on ARCH_SELECT_MEMORY_MODEL
420
421choice
422	prompt "Memory model"
423	depends on SELECT_MEMORY_MODEL
424	default SPARSEMEM_MANUAL if ARCH_SPARSEMEM_DEFAULT
425	default FLATMEM_MANUAL
426	help
427	  This option allows you to change some of the ways that
428	  Linux manages its memory internally. Most users will
429	  only have one option here selected by the architecture
430	  configuration. This is normal.
431
432config FLATMEM_MANUAL
433	bool "Flat Memory"
434	depends on !ARCH_SPARSEMEM_ENABLE || ARCH_FLATMEM_ENABLE
435	help
436	  This option is best suited for non-NUMA systems with
437	  flat address space. The FLATMEM is the most efficient
438	  system in terms of performance and resource consumption
439	  and it is the best option for smaller systems.
440
441	  For systems that have holes in their physical address
442	  spaces and for features like NUMA and memory hotplug,
443	  choose "Sparse Memory".
444
445	  If unsure, choose this option (Flat Memory) over any other.
446
447config SPARSEMEM_MANUAL
448	bool "Sparse Memory"
449	depends on ARCH_SPARSEMEM_ENABLE
450	help
451	  This will be the only option for some systems, including
452	  memory hot-plug systems.  This is normal.
453
454	  This option provides efficient support for systems with
455	  holes is their physical address space and allows memory
456	  hot-plug and hot-remove.
457
458	  If unsure, choose "Flat Memory" over this option.
459
460endchoice
461
462config SPARSEMEM
463	def_bool y
464	depends on (!SELECT_MEMORY_MODEL && ARCH_SPARSEMEM_ENABLE) || SPARSEMEM_MANUAL
465
466config FLATMEM
467	def_bool y
468	depends on !SPARSEMEM || FLATMEM_MANUAL
469
470#
471# SPARSEMEM_EXTREME (which is the default) does some bootmem
472# allocations when sparse_init() is called.  If this cannot
473# be done on your architecture, select this option.  However,
474# statically allocating the mem_section[] array can potentially
475# consume vast quantities of .bss, so be careful.
476#
477# This option will also potentially produce smaller runtime code
478# with gcc 3.4 and later.
479#
480config SPARSEMEM_STATIC
481	bool
482
483#
484# Architecture platforms which require a two level mem_section in SPARSEMEM
485# must select this option. This is usually for architecture platforms with
486# an extremely sparse physical address space.
487#
488config SPARSEMEM_EXTREME
489	def_bool y
490	depends on SPARSEMEM && !SPARSEMEM_STATIC
491
492config SPARSEMEM_VMEMMAP_ENABLE
493	bool
494
495config SPARSEMEM_VMEMMAP
496	bool "Sparse Memory virtual memmap"
497	depends on SPARSEMEM && SPARSEMEM_VMEMMAP_ENABLE
498	default y
499	help
500	  SPARSEMEM_VMEMMAP uses a virtually mapped memmap to optimise
501	  pfn_to_page and page_to_pfn operations.  This is the most
502	  efficient option when sufficient kernel resources are available.
503#
504# Select this config option from the architecture Kconfig, if it is preferred
505# to enable the feature of HugeTLB/dev_dax vmemmap optimization.
506#
507config ARCH_WANT_OPTIMIZE_VMEMMAP
508	bool
509
510config HAVE_MEMBLOCK_PHYS_MAP
511	bool
512
513config HAVE_FAST_GUP
514	depends on MMU
515	bool
516
517# Don't discard allocated memory used to track "memory" and "reserved" memblocks
518# after early boot, so it can still be used to test for validity of memory.
519# Also, memblocks are updated with memory hot(un)plug.
520config ARCH_KEEP_MEMBLOCK
521	bool
522
523# Keep arch NUMA mapping infrastructure post-init.
524config NUMA_KEEP_MEMINFO
525	bool
526
527config MEMORY_ISOLATION
528	bool
529
530# IORESOURCE_SYSTEM_RAM regions in the kernel resource tree that are marked
531# IORESOURCE_EXCLUSIVE cannot be mapped to user space, for example, via
532# /dev/mem.
533config EXCLUSIVE_SYSTEM_RAM
534	def_bool y
535	depends on !DEVMEM || STRICT_DEVMEM
536
537#
538# Only be set on architectures that have completely implemented memory hotplug
539# feature. If you are not sure, don't touch it.
540#
541config HAVE_BOOTMEM_INFO_NODE
542	def_bool n
543
544config ARCH_ENABLE_MEMORY_HOTPLUG
545	bool
546
547config ARCH_ENABLE_MEMORY_HOTREMOVE
548	bool
549
550# eventually, we can have this option just 'select SPARSEMEM'
551menuconfig MEMORY_HOTPLUG
552	bool "Memory hotplug"
553	select MEMORY_ISOLATION
554	depends on SPARSEMEM
555	depends on ARCH_ENABLE_MEMORY_HOTPLUG
556	depends on 64BIT
557	select NUMA_KEEP_MEMINFO if NUMA
558
559if MEMORY_HOTPLUG
560
561config MEMORY_HOTPLUG_DEFAULT_ONLINE
562	bool "Online the newly added memory blocks by default"
563	depends on MEMORY_HOTPLUG
564	help
565	  This option sets the default policy setting for memory hotplug
566	  onlining policy (/sys/devices/system/memory/auto_online_blocks) which
567	  determines what happens to newly added memory regions. Policy setting
568	  can always be changed at runtime.
569	  See Documentation/admin-guide/mm/memory-hotplug.rst for more information.
570
571	  Say Y here if you want all hot-plugged memory blocks to appear in
572	  'online' state by default.
573	  Say N here if you want the default policy to keep all hot-plugged
574	  memory blocks in 'offline' state.
575
576config MEMORY_HOTREMOVE
577	bool "Allow for memory hot remove"
578	select HAVE_BOOTMEM_INFO_NODE if (X86_64 || PPC64)
579	depends on MEMORY_HOTPLUG && ARCH_ENABLE_MEMORY_HOTREMOVE
580	depends on MIGRATION
581
582config MHP_MEMMAP_ON_MEMORY
583	def_bool y
584	depends on MEMORY_HOTPLUG && SPARSEMEM_VMEMMAP
585	depends on ARCH_MHP_MEMMAP_ON_MEMORY_ENABLE
586
587endif # MEMORY_HOTPLUG
588
589# Heavily threaded applications may benefit from splitting the mm-wide
590# page_table_lock, so that faults on different parts of the user address
591# space can be handled with less contention: split it at this NR_CPUS.
592# Default to 4 for wider testing, though 8 might be more appropriate.
593# ARM's adjust_pte (unused if VIPT) depends on mm-wide page_table_lock.
594# PA-RISC 7xxx's spinlock_t would enlarge struct page from 32 to 44 bytes.
595# SPARC32 allocates multiple pte tables within a single page, and therefore
596# a per-page lock leads to problems when multiple tables need to be locked
597# at the same time (e.g. copy_page_range()).
598# DEBUG_SPINLOCK and DEBUG_LOCK_ALLOC spinlock_t also enlarge struct page.
599#
600config SPLIT_PTLOCK_CPUS
601	int
602	default "999999" if !MMU
603	default "999999" if ARM && !CPU_CACHE_VIPT
604	default "999999" if PARISC && !PA20
605	default "999999" if SPARC32
606	default "4"
607
608config ARCH_ENABLE_SPLIT_PMD_PTLOCK
609	bool
610
611#
612# support for memory balloon
613config MEMORY_BALLOON
614	bool
615
616#
617# support for memory balloon compaction
618config BALLOON_COMPACTION
619	bool "Allow for balloon memory compaction/migration"
620	def_bool y
621	depends on COMPACTION && MEMORY_BALLOON
622	help
623	  Memory fragmentation introduced by ballooning might reduce
624	  significantly the number of 2MB contiguous memory blocks that can be
625	  used within a guest, thus imposing performance penalties associated
626	  with the reduced number of transparent huge pages that could be used
627	  by the guest workload. Allowing the compaction & migration for memory
628	  pages enlisted as being part of memory balloon devices avoids the
629	  scenario aforementioned and helps improving memory defragmentation.
630
631#
632# support for memory compaction
633config COMPACTION
634	bool "Allow for memory compaction"
635	def_bool y
636	select MIGRATION
637	depends on MMU
638	help
639	  Compaction is the only memory management component to form
640	  high order (larger physically contiguous) memory blocks
641	  reliably. The page allocator relies on compaction heavily and
642	  the lack of the feature can lead to unexpected OOM killer
643	  invocations for high order memory requests. You shouldn't
644	  disable this option unless there really is a strong reason for
645	  it and then we would be really interested to hear about that at
646	  linux-mm@kvack.org.
647
648config COMPACT_UNEVICTABLE_DEFAULT
649	int
650	depends on COMPACTION
651	default 0 if PREEMPT_RT
652	default 1
653
654#
655# support for free page reporting
656config PAGE_REPORTING
657	bool "Free page reporting"
658	def_bool n
659	help
660	  Free page reporting allows for the incremental acquisition of
661	  free pages from the buddy allocator for the purpose of reporting
662	  those pages to another entity, such as a hypervisor, so that the
663	  memory can be freed within the host for other uses.
664
665#
666# support for page migration
667#
668config MIGRATION
669	bool "Page migration"
670	def_bool y
671	depends on (NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION || CMA) && MMU
672	help
673	  Allows the migration of the physical location of pages of processes
674	  while the virtual addresses are not changed. This is useful in
675	  two situations. The first is on NUMA systems to put pages nearer
676	  to the processors accessing. The second is when allocating huge
677	  pages as migration can relocate pages to satisfy a huge page
678	  allocation instead of reclaiming.
679
680config DEVICE_MIGRATION
681	def_bool MIGRATION && ZONE_DEVICE
682
683config ARCH_ENABLE_HUGEPAGE_MIGRATION
684	bool
685
686config ARCH_ENABLE_THP_MIGRATION
687	bool
688
689config HUGETLB_PAGE_SIZE_VARIABLE
690	def_bool n
691	help
692	  Allows the pageblock_order value to be dynamic instead of just standard
693	  HUGETLB_PAGE_ORDER when there are multiple HugeTLB page sizes available
694	  on a platform.
695
696	  Note that the pageblock_order cannot exceed MAX_ORDER and will be
697	  clamped down to MAX_ORDER.
698
699config CONTIG_ALLOC
700	def_bool (MEMORY_ISOLATION && COMPACTION) || CMA
701
702config PHYS_ADDR_T_64BIT
703	def_bool 64BIT
704
705config BOUNCE
706	bool "Enable bounce buffers"
707	default y
708	depends on BLOCK && MMU && HIGHMEM
709	help
710	  Enable bounce buffers for devices that cannot access the full range of
711	  memory available to the CPU. Enabled by default when HIGHMEM is
712	  selected, but you may say n to override this.
713
714config MMU_NOTIFIER
715	bool
716	select INTERVAL_TREE
717
718config KSM
719	bool "Enable KSM for page merging"
720	depends on MMU
721	select XXHASH
722	help
723	  Enable Kernel Samepage Merging: KSM periodically scans those areas
724	  of an application's address space that an app has advised may be
725	  mergeable.  When it finds pages of identical content, it replaces
726	  the many instances by a single page with that content, so
727	  saving memory until one or another app needs to modify the content.
728	  Recommended for use with KVM, or with other duplicative applications.
729	  See Documentation/mm/ksm.rst for more information: KSM is inactive
730	  until a program has madvised that an area is MADV_MERGEABLE, and
731	  root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set).
732
733config DEFAULT_MMAP_MIN_ADDR
734	int "Low address space to protect from user allocation"
735	depends on MMU
736	default 4096
737	help
738	  This is the portion of low virtual memory which should be protected
739	  from userspace allocation.  Keeping a user from writing to low pages
740	  can help reduce the impact of kernel NULL pointer bugs.
741
742	  For most ia64, ppc64 and x86 users with lots of address space
743	  a value of 65536 is reasonable and should cause no problems.
744	  On arm and other archs it should not be higher than 32768.
745	  Programs which use vm86 functionality or have some need to map
746	  this low address space will need CAP_SYS_RAWIO or disable this
747	  protection by setting the value to 0.
748
749	  This value can be changed after boot using the
750	  /proc/sys/vm/mmap_min_addr tunable.
751
752config ARCH_SUPPORTS_MEMORY_FAILURE
753	bool
754
755config MEMORY_FAILURE
756	depends on MMU
757	depends on ARCH_SUPPORTS_MEMORY_FAILURE
758	bool "Enable recovery from hardware memory errors"
759	select MEMORY_ISOLATION
760	select RAS
761	help
762	  Enables code to recover from some memory failures on systems
763	  with MCA recovery. This allows a system to continue running
764	  even when some of its memory has uncorrected errors. This requires
765	  special hardware support and typically ECC memory.
766
767config HWPOISON_INJECT
768	tristate "HWPoison pages injector"
769	depends on MEMORY_FAILURE && DEBUG_KERNEL && PROC_FS
770	select PROC_PAGE_MONITOR
771
772config NOMMU_INITIAL_TRIM_EXCESS
773	int "Turn on mmap() excess space trimming before booting"
774	depends on !MMU
775	default 1
776	help
777	  The NOMMU mmap() frequently needs to allocate large contiguous chunks
778	  of memory on which to store mappings, but it can only ask the system
779	  allocator for chunks in 2^N*PAGE_SIZE amounts - which is frequently
780	  more than it requires.  To deal with this, mmap() is able to trim off
781	  the excess and return it to the allocator.
782
783	  If trimming is enabled, the excess is trimmed off and returned to the
784	  system allocator, which can cause extra fragmentation, particularly
785	  if there are a lot of transient processes.
786
787	  If trimming is disabled, the excess is kept, but not used, which for
788	  long-term mappings means that the space is wasted.
789
790	  Trimming can be dynamically controlled through a sysctl option
791	  (/proc/sys/vm/nr_trim_pages) which specifies the minimum number of
792	  excess pages there must be before trimming should occur, or zero if
793	  no trimming is to occur.
794
795	  This option specifies the initial value of this option.  The default
796	  of 1 says that all excess pages should be trimmed.
797
798	  See Documentation/admin-guide/mm/nommu-mmap.rst for more information.
799
800config ARCH_WANT_GENERAL_HUGETLB
801	bool
802
803config ARCH_WANTS_THP_SWAP
804	def_bool n
805
806menuconfig TRANSPARENT_HUGEPAGE
807	bool "Transparent Hugepage Support"
808	depends on HAVE_ARCH_TRANSPARENT_HUGEPAGE && !PREEMPT_RT
809	select COMPACTION
810	select XARRAY_MULTI
811	help
812	  Transparent Hugepages allows the kernel to use huge pages and
813	  huge tlb transparently to the applications whenever possible.
814	  This feature can improve computing performance to certain
815	  applications by speeding up page faults during memory
816	  allocation, by reducing the number of tlb misses and by speeding
817	  up the pagetable walking.
818
819	  If memory constrained on embedded, you may want to say N.
820
821if TRANSPARENT_HUGEPAGE
822
823choice
824	prompt "Transparent Hugepage Support sysfs defaults"
825	depends on TRANSPARENT_HUGEPAGE
826	default TRANSPARENT_HUGEPAGE_ALWAYS
827	help
828	  Selects the sysfs defaults for Transparent Hugepage Support.
829
830	config TRANSPARENT_HUGEPAGE_ALWAYS
831		bool "always"
832	help
833	  Enabling Transparent Hugepage always, can increase the
834	  memory footprint of applications without a guaranteed
835	  benefit but it will work automatically for all applications.
836
837	config TRANSPARENT_HUGEPAGE_MADVISE
838		bool "madvise"
839	help
840	  Enabling Transparent Hugepage madvise, will only provide a
841	  performance improvement benefit to the applications using
842	  madvise(MADV_HUGEPAGE) but it won't risk to increase the
843	  memory footprint of applications without a guaranteed
844	  benefit.
845endchoice
846
847config THP_SWAP
848	def_bool y
849	depends on TRANSPARENT_HUGEPAGE && ARCH_WANTS_THP_SWAP && SWAP && 64BIT
850	help
851	  Swap transparent huge pages in one piece, without splitting.
852	  XXX: For now, swap cluster backing transparent huge page
853	  will be split after swapout.
854
855	  For selection by architectures with reasonable THP sizes.
856
857config READ_ONLY_THP_FOR_FS
858	bool "Read-only THP for filesystems (EXPERIMENTAL)"
859	depends on TRANSPARENT_HUGEPAGE && SHMEM
860
861	help
862	  Allow khugepaged to put read-only file-backed pages in THP.
863
864	  This is marked experimental because it is a new feature. Write
865	  support of file THPs will be developed in the next few release
866	  cycles.
867
868endif # TRANSPARENT_HUGEPAGE
869
870#
871# UP and nommu archs use km based percpu allocator
872#
873config NEED_PER_CPU_KM
874	depends on !SMP || !MMU
875	bool
876	default y
877
878config NEED_PER_CPU_EMBED_FIRST_CHUNK
879	bool
880
881config NEED_PER_CPU_PAGE_FIRST_CHUNK
882	bool
883
884config USE_PERCPU_NUMA_NODE_ID
885	bool
886
887config HAVE_SETUP_PER_CPU_AREA
888	bool
889
890config FRONTSWAP
891	bool
892
893config CMA
894	bool "Contiguous Memory Allocator"
895	depends on MMU
896	select MIGRATION
897	select MEMORY_ISOLATION
898	help
899	  This enables the Contiguous Memory Allocator which allows other
900	  subsystems to allocate big physically-contiguous blocks of memory.
901	  CMA reserves a region of memory and allows only movable pages to
902	  be allocated from it. This way, the kernel can use the memory for
903	  pagecache and when a subsystem requests for contiguous area, the
904	  allocated pages are migrated away to serve the contiguous request.
905
906	  If unsure, say "n".
907
908config CMA_DEBUG
909	bool "CMA debug messages (DEVELOPMENT)"
910	depends on DEBUG_KERNEL && CMA
911	help
912	  Turns on debug messages in CMA.  This produces KERN_DEBUG
913	  messages for every CMA call as well as various messages while
914	  processing calls such as dma_alloc_from_contiguous().
915	  This option does not affect warning and error messages.
916
917config CMA_DEBUGFS
918	bool "CMA debugfs interface"
919	depends on CMA && DEBUG_FS
920	help
921	  Turns on the DebugFS interface for CMA.
922
923config CMA_SYSFS
924	bool "CMA information through sysfs interface"
925	depends on CMA && SYSFS
926	help
927	  This option exposes some sysfs attributes to get information
928	  from CMA.
929
930config CMA_AREAS
931	int "Maximum count of the CMA areas"
932	depends on CMA
933	default 19 if NUMA
934	default 7
935	help
936	  CMA allows to create CMA areas for particular purpose, mainly,
937	  used as device private area. This parameter sets the maximum
938	  number of CMA area in the system.
939
940	  If unsure, leave the default value "7" in UMA and "19" in NUMA.
941
942config MEM_SOFT_DIRTY
943	bool "Track memory changes"
944	depends on CHECKPOINT_RESTORE && HAVE_ARCH_SOFT_DIRTY && PROC_FS
945	select PROC_PAGE_MONITOR
946	help
947	  This option enables memory changes tracking by introducing a
948	  soft-dirty bit on pte-s. This bit it set when someone writes
949	  into a page just as regular dirty bit, but unlike the latter
950	  it can be cleared by hands.
951
952	  See Documentation/admin-guide/mm/soft-dirty.rst for more details.
953
954config GENERIC_EARLY_IOREMAP
955	bool
956
957config STACK_MAX_DEFAULT_SIZE_MB
958	int "Default maximum user stack size for 32-bit processes (MB)"
959	default 100
960	range 8 2048
961	depends on STACK_GROWSUP && (!64BIT || COMPAT)
962	help
963	  This is the maximum stack size in Megabytes in the VM layout of 32-bit
964	  user processes when the stack grows upwards (currently only on parisc
965	  arch) when the RLIMIT_STACK hard limit is unlimited.
966
967	  A sane initial value is 100 MB.
968
969config DEFERRED_STRUCT_PAGE_INIT
970	bool "Defer initialisation of struct pages to kthreads"
971	depends on SPARSEMEM
972	depends on !NEED_PER_CPU_KM
973	depends on 64BIT
974	select PADATA
975	help
976	  Ordinarily all struct pages are initialised during early boot in a
977	  single thread. On very large machines this can take a considerable
978	  amount of time. If this option is set, large machines will bring up
979	  a subset of memmap at boot and then initialise the rest in parallel.
980	  This has a potential performance impact on tasks running early in the
981	  lifetime of the system until these kthreads finish the
982	  initialisation.
983
984config PAGE_IDLE_FLAG
985	bool
986	select PAGE_EXTENSION if !64BIT
987	help
988	  This adds PG_idle and PG_young flags to 'struct page'.  PTE Accessed
989	  bit writers can set the state of the bit in the flags so that PTE
990	  Accessed bit readers may avoid disturbance.
991
992config IDLE_PAGE_TRACKING
993	bool "Enable idle page tracking"
994	depends on SYSFS && MMU
995	select PAGE_IDLE_FLAG
996	help
997	  This feature allows to estimate the amount of user pages that have
998	  not been touched during a given period of time. This information can
999	  be useful to tune memory cgroup limits and/or for job placement
1000	  within a compute cluster.
1001
1002	  See Documentation/admin-guide/mm/idle_page_tracking.rst for
1003	  more details.
1004
1005config ARCH_HAS_CACHE_LINE_SIZE
1006	bool
1007
1008config ARCH_HAS_CURRENT_STACK_POINTER
1009	bool
1010	help
1011	  In support of HARDENED_USERCOPY performing stack variable lifetime
1012	  checking, an architecture-agnostic way to find the stack pointer
1013	  is needed. Once an architecture defines an unsigned long global
1014	  register alias named "current_stack_pointer", this config can be
1015	  selected.
1016
1017config ARCH_HAS_PTE_DEVMAP
1018	bool
1019
1020config ARCH_HAS_ZONE_DMA_SET
1021	bool
1022
1023config ZONE_DMA
1024	bool "Support DMA zone" if ARCH_HAS_ZONE_DMA_SET
1025	default y if ARM64 || X86
1026
1027config ZONE_DMA32
1028	bool "Support DMA32 zone" if ARCH_HAS_ZONE_DMA_SET
1029	depends on !X86_32
1030	default y if ARM64
1031
1032config ZONE_DEVICE
1033	bool "Device memory (pmem, HMM, etc...) hotplug support"
1034	depends on MEMORY_HOTPLUG
1035	depends on MEMORY_HOTREMOVE
1036	depends on SPARSEMEM_VMEMMAP
1037	depends on ARCH_HAS_PTE_DEVMAP
1038	select XARRAY_MULTI
1039
1040	help
1041	  Device memory hotplug support allows for establishing pmem,
1042	  or other device driver discovered memory regions, in the
1043	  memmap. This allows pfn_to_page() lookups of otherwise
1044	  "device-physical" addresses which is needed for using a DAX
1045	  mapping in an O_DIRECT operation, among other things.
1046
1047	  If FS_DAX is enabled, then say Y.
1048
1049#
1050# Helpers to mirror range of the CPU page tables of a process into device page
1051# tables.
1052#
1053config HMM_MIRROR
1054	bool
1055	depends on MMU
1056
1057config GET_FREE_REGION
1058	depends on SPARSEMEM
1059	bool
1060
1061config DEVICE_PRIVATE
1062	bool "Unaddressable device memory (GPU memory, ...)"
1063	depends on ZONE_DEVICE
1064	select GET_FREE_REGION
1065
1066	help
1067	  Allows creation of struct pages to represent unaddressable device
1068	  memory; i.e., memory that is only accessible from the device (or
1069	  group of devices). You likely also want to select HMM_MIRROR.
1070
1071config VMAP_PFN
1072	bool
1073
1074config ARCH_USES_HIGH_VMA_FLAGS
1075	bool
1076config ARCH_HAS_PKEYS
1077	bool
1078
1079config ARCH_USES_PG_ARCH_X
1080	bool
1081	help
1082	  Enable the definition of PG_arch_x page flags with x > 1. Only
1083	  suitable for 64-bit architectures with CONFIG_FLATMEM or
1084	  CONFIG_SPARSEMEM_VMEMMAP enabled, otherwise there may not be
1085	  enough room for additional bits in page->flags.
1086
1087config VM_EVENT_COUNTERS
1088	default y
1089	bool "Enable VM event counters for /proc/vmstat" if EXPERT
1090	help
1091	  VM event counters are needed for event counts to be shown.
1092	  This option allows the disabling of the VM event counters
1093	  on EXPERT systems.  /proc/vmstat will only show page counts
1094	  if VM event counters are disabled.
1095
1096config PERCPU_STATS
1097	bool "Collect percpu memory statistics"
1098	help
1099	  This feature collects and exposes statistics via debugfs. The
1100	  information includes global and per chunk statistics, which can
1101	  be used to help understand percpu memory usage.
1102
1103config GUP_TEST
1104	bool "Enable infrastructure for get_user_pages()-related unit tests"
1105	depends on DEBUG_FS
1106	help
1107	  Provides /sys/kernel/debug/gup_test, which in turn provides a way
1108	  to make ioctl calls that can launch kernel-based unit tests for
1109	  the get_user_pages*() and pin_user_pages*() family of API calls.
1110
1111	  These tests include benchmark testing of the _fast variants of
1112	  get_user_pages*() and pin_user_pages*(), as well as smoke tests of
1113	  the non-_fast variants.
1114
1115	  There is also a sub-test that allows running dump_page() on any
1116	  of up to eight pages (selected by command line args) within the
1117	  range of user-space addresses. These pages are either pinned via
1118	  pin_user_pages*(), or pinned via get_user_pages*(), as specified
1119	  by other command line arguments.
1120
1121	  See tools/testing/selftests/mm/gup_test.c
1122
1123comment "GUP_TEST needs to have DEBUG_FS enabled"
1124	depends on !GUP_TEST && !DEBUG_FS
1125
1126config GUP_GET_PXX_LOW_HIGH
1127	bool
1128
1129config DMAPOOL_TEST
1130	tristate "Enable a module to run time tests on dma_pool"
1131	depends on HAS_DMA
1132	help
1133	  Provides a test module that will allocate and free many blocks of
1134	  various sizes and report how long it takes. This is intended to
1135	  provide a consistent way to measure how changes to the
1136	  dma_pool_alloc/free routines affect performance.
1137
1138config ARCH_HAS_PTE_SPECIAL
1139	bool
1140
1141#
1142# Some architectures require a special hugepage directory format that is
1143# required to support multiple hugepage sizes. For example a4fe3ce76
1144# "powerpc/mm: Allow more flexible layouts for hugepage pagetables"
1145# introduced it on powerpc.  This allows for a more flexible hugepage
1146# pagetable layouts.
1147#
1148config ARCH_HAS_HUGEPD
1149	bool
1150
1151config MAPPING_DIRTY_HELPERS
1152        bool
1153
1154config KMAP_LOCAL
1155	bool
1156
1157config KMAP_LOCAL_NON_LINEAR_PTE_ARRAY
1158	bool
1159
1160# struct io_mapping based helper.  Selected by drivers that need them
1161config IO_MAPPING
1162	bool
1163
1164config SECRETMEM
1165	default y
1166	bool "Enable memfd_secret() system call" if EXPERT
1167	depends on ARCH_HAS_SET_DIRECT_MAP
1168	help
1169	  Enable the memfd_secret() system call with the ability to create
1170	  memory areas visible only in the context of the owning process and
1171	  not mapped to other processes and other kernel page tables.
1172
1173config ANON_VMA_NAME
1174	bool "Anonymous VMA name support"
1175	depends on PROC_FS && ADVISE_SYSCALLS && MMU
1176
1177	help
1178	  Allow naming anonymous virtual memory areas.
1179
1180	  This feature allows assigning names to virtual memory areas. Assigned
1181	  names can be later retrieved from /proc/pid/maps and /proc/pid/smaps
1182	  and help identifying individual anonymous memory areas.
1183	  Assigning a name to anonymous virtual memory area might prevent that
1184	  area from being merged with adjacent virtual memory areas due to the
1185	  difference in their name.
1186
1187config USERFAULTFD
1188	bool "Enable userfaultfd() system call"
1189	depends on MMU
1190	help
1191	  Enable the userfaultfd() system call that allows to intercept and
1192	  handle page faults in userland.
1193
1194config HAVE_ARCH_USERFAULTFD_WP
1195	bool
1196	help
1197	  Arch has userfaultfd write protection support
1198
1199config HAVE_ARCH_USERFAULTFD_MINOR
1200	bool
1201	help
1202	  Arch has userfaultfd minor fault support
1203
1204config PTE_MARKER_UFFD_WP
1205	bool "Userfaultfd write protection support for shmem/hugetlbfs"
1206	default y
1207	depends on HAVE_ARCH_USERFAULTFD_WP
1208
1209	help
1210	  Allows to create marker PTEs for userfaultfd write protection
1211	  purposes.  It is required to enable userfaultfd write protection on
1212	  file-backed memory types like shmem and hugetlbfs.
1213
1214# multi-gen LRU {
1215config LRU_GEN
1216	bool "Multi-Gen LRU"
1217	depends on MMU
1218	# make sure folio->flags has enough spare bits
1219	depends on 64BIT || !SPARSEMEM || SPARSEMEM_VMEMMAP
1220	help
1221	  A high performance LRU implementation to overcommit memory. See
1222	  Documentation/admin-guide/mm/multigen_lru.rst for details.
1223
1224config LRU_GEN_ENABLED
1225	bool "Enable by default"
1226	depends on LRU_GEN
1227	help
1228	  This option enables the multi-gen LRU by default.
1229
1230config LRU_GEN_STATS
1231	bool "Full stats for debugging"
1232	depends on LRU_GEN
1233	help
1234	  Do not enable this option unless you plan to look at historical stats
1235	  from evicted generations for debugging purpose.
1236
1237	  This option has a per-memcg and per-node memory overhead.
1238# }
1239
1240config ARCH_SUPPORTS_PER_VMA_LOCK
1241       def_bool n
1242
1243config PER_VMA_LOCK
1244	def_bool y
1245	depends on ARCH_SUPPORTS_PER_VMA_LOCK && MMU && SMP
1246	help
1247	  Allow per-vma locking during page fault handling.
1248
1249	  This feature allows locking each virtual memory area separately when
1250	  handling page faults instead of taking mmap_lock.
1251
1252config LOCK_MM_AND_FIND_VMA
1253	bool
1254	depends on !STACK_GROWSUP
1255
1256source "mm/damon/Kconfig"
1257
1258endmenu
1259