Loading...
1# SPDX-License-Identifier: GPL-2.0-only
2
3menu "Memory Management options"
4
5config SELECT_MEMORY_MODEL
6 def_bool y
7 depends on ARCH_SELECT_MEMORY_MODEL
8
9choice
10 prompt "Memory model"
11 depends on SELECT_MEMORY_MODEL
12 default DISCONTIGMEM_MANUAL if ARCH_DISCONTIGMEM_DEFAULT
13 default SPARSEMEM_MANUAL if ARCH_SPARSEMEM_DEFAULT
14 default FLATMEM_MANUAL
15 help
16 This option allows you to change some of the ways that
17 Linux manages its memory internally. Most users will
18 only have one option here selected by the architecture
19 configuration. This is normal.
20
21config FLATMEM_MANUAL
22 bool "Flat Memory"
23 depends on !(ARCH_DISCONTIGMEM_ENABLE || ARCH_SPARSEMEM_ENABLE) || ARCH_FLATMEM_ENABLE
24 help
25 This option is best suited for non-NUMA systems with
26 flat address space. The FLATMEM is the most efficient
27 system in terms of performance and resource consumption
28 and it is the best option for smaller systems.
29
30 For systems that have holes in their physical address
31 spaces and for features like NUMA and memory hotplug,
32 choose "Sparse Memory".
33
34 If unsure, choose this option (Flat Memory) over any other.
35
36config DISCONTIGMEM_MANUAL
37 bool "Discontiguous Memory"
38 depends on ARCH_DISCONTIGMEM_ENABLE
39 help
40 This option provides enhanced support for discontiguous
41 memory systems, over FLATMEM. These systems have holes
42 in their physical address spaces, and this option provides
43 more efficient handling of these holes.
44
45 Although "Discontiguous Memory" is still used by several
46 architectures, it is considered deprecated in favor of
47 "Sparse Memory".
48
49 If unsure, choose "Sparse Memory" over this option.
50
51config SPARSEMEM_MANUAL
52 bool "Sparse Memory"
53 depends on ARCH_SPARSEMEM_ENABLE
54 help
55 This will be the only option for some systems, including
56 memory hot-plug systems. This is normal.
57
58 This option provides efficient support for systems with
59 holes is their physical address space and allows memory
60 hot-plug and hot-remove.
61
62 If unsure, choose "Flat Memory" over this option.
63
64endchoice
65
66config DISCONTIGMEM
67 def_bool y
68 depends on (!SELECT_MEMORY_MODEL && ARCH_DISCONTIGMEM_ENABLE) || DISCONTIGMEM_MANUAL
69
70config SPARSEMEM
71 def_bool y
72 depends on (!SELECT_MEMORY_MODEL && ARCH_SPARSEMEM_ENABLE) || SPARSEMEM_MANUAL
73
74config FLATMEM
75 def_bool y
76 depends on (!DISCONTIGMEM && !SPARSEMEM) || FLATMEM_MANUAL
77
78config FLAT_NODE_MEM_MAP
79 def_bool y
80 depends on !SPARSEMEM
81
82#
83# Both the NUMA code and DISCONTIGMEM use arrays of pg_data_t's
84# to represent different areas of memory. This variable allows
85# those dependencies to exist individually.
86#
87config NEED_MULTIPLE_NODES
88 def_bool y
89 depends on DISCONTIGMEM || NUMA
90
91#
92# SPARSEMEM_EXTREME (which is the default) does some bootmem
93# allocations when sparse_init() is called. If this cannot
94# be done on your architecture, select this option. However,
95# statically allocating the mem_section[] array can potentially
96# consume vast quantities of .bss, so be careful.
97#
98# This option will also potentially produce smaller runtime code
99# with gcc 3.4 and later.
100#
101config SPARSEMEM_STATIC
102 bool
103
104#
105# Architecture platforms which require a two level mem_section in SPARSEMEM
106# must select this option. This is usually for architecture platforms with
107# an extremely sparse physical address space.
108#
109config SPARSEMEM_EXTREME
110 def_bool y
111 depends on SPARSEMEM && !SPARSEMEM_STATIC
112
113config SPARSEMEM_VMEMMAP_ENABLE
114 bool
115
116config SPARSEMEM_VMEMMAP
117 bool "Sparse Memory virtual memmap"
118 depends on SPARSEMEM && SPARSEMEM_VMEMMAP_ENABLE
119 default y
120 help
121 SPARSEMEM_VMEMMAP uses a virtually mapped memmap to optimise
122 pfn_to_page and page_to_pfn operations. This is the most
123 efficient option when sufficient kernel resources are available.
124
125config HAVE_MEMBLOCK_PHYS_MAP
126 bool
127
128config HAVE_FAST_GUP
129 depends on MMU
130 bool
131
132# Don't discard allocated memory used to track "memory" and "reserved" memblocks
133# after early boot, so it can still be used to test for validity of memory.
134# Also, memblocks are updated with memory hot(un)plug.
135config ARCH_KEEP_MEMBLOCK
136 bool
137
138# Keep arch NUMA mapping infrastructure post-init.
139config NUMA_KEEP_MEMINFO
140 bool
141
142config MEMORY_ISOLATION
143 bool
144
145#
146# Only be set on architectures that have completely implemented memory hotplug
147# feature. If you are not sure, don't touch it.
148#
149config HAVE_BOOTMEM_INFO_NODE
150 def_bool n
151
152# eventually, we can have this option just 'select SPARSEMEM'
153config MEMORY_HOTPLUG
154 bool "Allow for memory hot-add"
155 depends on SPARSEMEM || X86_64_ACPI_NUMA
156 depends on ARCH_ENABLE_MEMORY_HOTPLUG
157 depends on 64BIT || BROKEN
158 select NUMA_KEEP_MEMINFO if NUMA
159
160config MEMORY_HOTPLUG_SPARSE
161 def_bool y
162 depends on SPARSEMEM && MEMORY_HOTPLUG
163
164config MEMORY_HOTPLUG_DEFAULT_ONLINE
165 bool "Online the newly added memory blocks by default"
166 depends on MEMORY_HOTPLUG
167 help
168 This option sets the default policy setting for memory hotplug
169 onlining policy (/sys/devices/system/memory/auto_online_blocks) which
170 determines what happens to newly added memory regions. Policy setting
171 can always be changed at runtime.
172 See Documentation/admin-guide/mm/memory-hotplug.rst for more information.
173
174 Say Y here if you want all hot-plugged memory blocks to appear in
175 'online' state by default.
176 Say N here if you want the default policy to keep all hot-plugged
177 memory blocks in 'offline' state.
178
179config MEMORY_HOTREMOVE
180 bool "Allow for memory hot remove"
181 select MEMORY_ISOLATION
182 select HAVE_BOOTMEM_INFO_NODE if (X86_64 || PPC64)
183 depends on MEMORY_HOTPLUG && ARCH_ENABLE_MEMORY_HOTREMOVE
184 depends on MIGRATION
185
186# Heavily threaded applications may benefit from splitting the mm-wide
187# page_table_lock, so that faults on different parts of the user address
188# space can be handled with less contention: split it at this NR_CPUS.
189# Default to 4 for wider testing, though 8 might be more appropriate.
190# ARM's adjust_pte (unused if VIPT) depends on mm-wide page_table_lock.
191# PA-RISC 7xxx's spinlock_t would enlarge struct page from 32 to 44 bytes.
192# SPARC32 allocates multiple pte tables within a single page, and therefore
193# a per-page lock leads to problems when multiple tables need to be locked
194# at the same time (e.g. copy_page_range()).
195# DEBUG_SPINLOCK and DEBUG_LOCK_ALLOC spinlock_t also enlarge struct page.
196#
197config SPLIT_PTLOCK_CPUS
198 int
199 default "999999" if !MMU
200 default "999999" if ARM && !CPU_CACHE_VIPT
201 default "999999" if PARISC && !PA20
202 default "999999" if SPARC32
203 default "4"
204
205config ARCH_ENABLE_SPLIT_PMD_PTLOCK
206 bool
207
208#
209# support for memory balloon
210config MEMORY_BALLOON
211 bool
212
213#
214# support for memory balloon compaction
215config BALLOON_COMPACTION
216 bool "Allow for balloon memory compaction/migration"
217 def_bool y
218 depends on COMPACTION && MEMORY_BALLOON
219 help
220 Memory fragmentation introduced by ballooning might reduce
221 significantly the number of 2MB contiguous memory blocks that can be
222 used within a guest, thus imposing performance penalties associated
223 with the reduced number of transparent huge pages that could be used
224 by the guest workload. Allowing the compaction & migration for memory
225 pages enlisted as being part of memory balloon devices avoids the
226 scenario aforementioned and helps improving memory defragmentation.
227
228#
229# support for memory compaction
230config COMPACTION
231 bool "Allow for memory compaction"
232 def_bool y
233 select MIGRATION
234 depends on MMU
235 help
236 Compaction is the only memory management component to form
237 high order (larger physically contiguous) memory blocks
238 reliably. The page allocator relies on compaction heavily and
239 the lack of the feature can lead to unexpected OOM killer
240 invocations for high order memory requests. You shouldn't
241 disable this option unless there really is a strong reason for
242 it and then we would be really interested to hear about that at
243 linux-mm@kvack.org.
244
245#
246# support for free page reporting
247config PAGE_REPORTING
248 bool "Free page reporting"
249 def_bool n
250 help
251 Free page reporting allows for the incremental acquisition of
252 free pages from the buddy allocator for the purpose of reporting
253 those pages to another entity, such as a hypervisor, so that the
254 memory can be freed within the host for other uses.
255
256#
257# support for page migration
258#
259config MIGRATION
260 bool "Page migration"
261 def_bool y
262 depends on (NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION || CMA) && MMU
263 help
264 Allows the migration of the physical location of pages of processes
265 while the virtual addresses are not changed. This is useful in
266 two situations. The first is on NUMA systems to put pages nearer
267 to the processors accessing. The second is when allocating huge
268 pages as migration can relocate pages to satisfy a huge page
269 allocation instead of reclaiming.
270
271config ARCH_ENABLE_HUGEPAGE_MIGRATION
272 bool
273
274config ARCH_ENABLE_THP_MIGRATION
275 bool
276
277config CONTIG_ALLOC
278 def_bool (MEMORY_ISOLATION && COMPACTION) || CMA
279
280config PHYS_ADDR_T_64BIT
281 def_bool 64BIT
282
283config BOUNCE
284 bool "Enable bounce buffers"
285 default y
286 depends on BLOCK && MMU && (ZONE_DMA || HIGHMEM)
287 help
288 Enable bounce buffers for devices that cannot access
289 the full range of memory available to the CPU. Enabled
290 by default when ZONE_DMA or HIGHMEM is selected, but you
291 may say n to override this.
292
293config VIRT_TO_BUS
294 bool
295 help
296 An architecture should select this if it implements the
297 deprecated interface virt_to_bus(). All new architectures
298 should probably not select this.
299
300
301config MMU_NOTIFIER
302 bool
303 select SRCU
304 select INTERVAL_TREE
305
306config KSM
307 bool "Enable KSM for page merging"
308 depends on MMU
309 select XXHASH
310 help
311 Enable Kernel Samepage Merging: KSM periodically scans those areas
312 of an application's address space that an app has advised may be
313 mergeable. When it finds pages of identical content, it replaces
314 the many instances by a single page with that content, so
315 saving memory until one or another app needs to modify the content.
316 Recommended for use with KVM, or with other duplicative applications.
317 See Documentation/vm/ksm.rst for more information: KSM is inactive
318 until a program has madvised that an area is MADV_MERGEABLE, and
319 root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set).
320
321config DEFAULT_MMAP_MIN_ADDR
322 int "Low address space to protect from user allocation"
323 depends on MMU
324 default 4096
325 help
326 This is the portion of low virtual memory which should be protected
327 from userspace allocation. Keeping a user from writing to low pages
328 can help reduce the impact of kernel NULL pointer bugs.
329
330 For most ia64, ppc64 and x86 users with lots of address space
331 a value of 65536 is reasonable and should cause no problems.
332 On arm and other archs it should not be higher than 32768.
333 Programs which use vm86 functionality or have some need to map
334 this low address space will need CAP_SYS_RAWIO or disable this
335 protection by setting the value to 0.
336
337 This value can be changed after boot using the
338 /proc/sys/vm/mmap_min_addr tunable.
339
340config ARCH_SUPPORTS_MEMORY_FAILURE
341 bool
342
343config MEMORY_FAILURE
344 depends on MMU
345 depends on ARCH_SUPPORTS_MEMORY_FAILURE
346 bool "Enable recovery from hardware memory errors"
347 select MEMORY_ISOLATION
348 select RAS
349 help
350 Enables code to recover from some memory failures on systems
351 with MCA recovery. This allows a system to continue running
352 even when some of its memory has uncorrected errors. This requires
353 special hardware support and typically ECC memory.
354
355config HWPOISON_INJECT
356 tristate "HWPoison pages injector"
357 depends on MEMORY_FAILURE && DEBUG_KERNEL && PROC_FS
358 select PROC_PAGE_MONITOR
359
360config NOMMU_INITIAL_TRIM_EXCESS
361 int "Turn on mmap() excess space trimming before booting"
362 depends on !MMU
363 default 1
364 help
365 The NOMMU mmap() frequently needs to allocate large contiguous chunks
366 of memory on which to store mappings, but it can only ask the system
367 allocator for chunks in 2^N*PAGE_SIZE amounts - which is frequently
368 more than it requires. To deal with this, mmap() is able to trim off
369 the excess and return it to the allocator.
370
371 If trimming is enabled, the excess is trimmed off and returned to the
372 system allocator, which can cause extra fragmentation, particularly
373 if there are a lot of transient processes.
374
375 If trimming is disabled, the excess is kept, but not used, which for
376 long-term mappings means that the space is wasted.
377
378 Trimming can be dynamically controlled through a sysctl option
379 (/proc/sys/vm/nr_trim_pages) which specifies the minimum number of
380 excess pages there must be before trimming should occur, or zero if
381 no trimming is to occur.
382
383 This option specifies the initial value of this option. The default
384 of 1 says that all excess pages should be trimmed.
385
386 See Documentation/mm/nommu-mmap.rst for more information.
387
388config TRANSPARENT_HUGEPAGE
389 bool "Transparent Hugepage Support"
390 depends on HAVE_ARCH_TRANSPARENT_HUGEPAGE
391 select COMPACTION
392 select XARRAY_MULTI
393 help
394 Transparent Hugepages allows the kernel to use huge pages and
395 huge tlb transparently to the applications whenever possible.
396 This feature can improve computing performance to certain
397 applications by speeding up page faults during memory
398 allocation, by reducing the number of tlb misses and by speeding
399 up the pagetable walking.
400
401 If memory constrained on embedded, you may want to say N.
402
403choice
404 prompt "Transparent Hugepage Support sysfs defaults"
405 depends on TRANSPARENT_HUGEPAGE
406 default TRANSPARENT_HUGEPAGE_ALWAYS
407 help
408 Selects the sysfs defaults for Transparent Hugepage Support.
409
410 config TRANSPARENT_HUGEPAGE_ALWAYS
411 bool "always"
412 help
413 Enabling Transparent Hugepage always, can increase the
414 memory footprint of applications without a guaranteed
415 benefit but it will work automatically for all applications.
416
417 config TRANSPARENT_HUGEPAGE_MADVISE
418 bool "madvise"
419 help
420 Enabling Transparent Hugepage madvise, will only provide a
421 performance improvement benefit to the applications using
422 madvise(MADV_HUGEPAGE) but it won't risk to increase the
423 memory footprint of applications without a guaranteed
424 benefit.
425endchoice
426
427config ARCH_WANTS_THP_SWAP
428 def_bool n
429
430config THP_SWAP
431 def_bool y
432 depends on TRANSPARENT_HUGEPAGE && ARCH_WANTS_THP_SWAP && SWAP
433 help
434 Swap transparent huge pages in one piece, without splitting.
435 XXX: For now, swap cluster backing transparent huge page
436 will be split after swapout.
437
438 For selection by architectures with reasonable THP sizes.
439
440#
441# UP and nommu archs use km based percpu allocator
442#
443config NEED_PER_CPU_KM
444 depends on !SMP
445 bool
446 default y
447
448config CLEANCACHE
449 bool "Enable cleancache driver to cache clean pages if tmem is present"
450 help
451 Cleancache can be thought of as a page-granularity victim cache
452 for clean pages that the kernel's pageframe replacement algorithm
453 (PFRA) would like to keep around, but can't since there isn't enough
454 memory. So when the PFRA "evicts" a page, it first attempts to use
455 cleancache code to put the data contained in that page into
456 "transcendent memory", memory that is not directly accessible or
457 addressable by the kernel and is of unknown and possibly
458 time-varying size. And when a cleancache-enabled
459 filesystem wishes to access a page in a file on disk, it first
460 checks cleancache to see if it already contains it; if it does,
461 the page is copied into the kernel and a disk access is avoided.
462 When a transcendent memory driver is available (such as zcache or
463 Xen transcendent memory), a significant I/O reduction
464 may be achieved. When none is available, all cleancache calls
465 are reduced to a single pointer-compare-against-NULL resulting
466 in a negligible performance hit.
467
468 If unsure, say Y to enable cleancache
469
470config FRONTSWAP
471 bool "Enable frontswap to cache swap pages if tmem is present"
472 depends on SWAP
473 help
474 Frontswap is so named because it can be thought of as the opposite
475 of a "backing" store for a swap device. The data is stored into
476 "transcendent memory", memory that is not directly accessible or
477 addressable by the kernel and is of unknown and possibly
478 time-varying size. When space in transcendent memory is available,
479 a significant swap I/O reduction may be achieved. When none is
480 available, all frontswap calls are reduced to a single pointer-
481 compare-against-NULL resulting in a negligible performance hit
482 and swap data is stored as normal on the matching swap device.
483
484 If unsure, say Y to enable frontswap.
485
486config CMA
487 bool "Contiguous Memory Allocator"
488 depends on MMU
489 select MIGRATION
490 select MEMORY_ISOLATION
491 help
492 This enables the Contiguous Memory Allocator which allows other
493 subsystems to allocate big physically-contiguous blocks of memory.
494 CMA reserves a region of memory and allows only movable pages to
495 be allocated from it. This way, the kernel can use the memory for
496 pagecache and when a subsystem requests for contiguous area, the
497 allocated pages are migrated away to serve the contiguous request.
498
499 If unsure, say "n".
500
501config CMA_DEBUG
502 bool "CMA debug messages (DEVELOPMENT)"
503 depends on DEBUG_KERNEL && CMA
504 help
505 Turns on debug messages in CMA. This produces KERN_DEBUG
506 messages for every CMA call as well as various messages while
507 processing calls such as dma_alloc_from_contiguous().
508 This option does not affect warning and error messages.
509
510config CMA_DEBUGFS
511 bool "CMA debugfs interface"
512 depends on CMA && DEBUG_FS
513 help
514 Turns on the DebugFS interface for CMA.
515
516config CMA_AREAS
517 int "Maximum count of the CMA areas"
518 depends on CMA
519 default 7
520 help
521 CMA allows to create CMA areas for particular purpose, mainly,
522 used as device private area. This parameter sets the maximum
523 number of CMA area in the system.
524
525 If unsure, leave the default value "7".
526
527config MEM_SOFT_DIRTY
528 bool "Track memory changes"
529 depends on CHECKPOINT_RESTORE && HAVE_ARCH_SOFT_DIRTY && PROC_FS
530 select PROC_PAGE_MONITOR
531 help
532 This option enables memory changes tracking by introducing a
533 soft-dirty bit on pte-s. This bit it set when someone writes
534 into a page just as regular dirty bit, but unlike the latter
535 it can be cleared by hands.
536
537 See Documentation/admin-guide/mm/soft-dirty.rst for more details.
538
539config ZSWAP
540 bool "Compressed cache for swap pages (EXPERIMENTAL)"
541 depends on FRONTSWAP && CRYPTO=y
542 select ZPOOL
543 help
544 A lightweight compressed cache for swap pages. It takes
545 pages that are in the process of being swapped out and attempts to
546 compress them into a dynamically allocated RAM-based memory pool.
547 This can result in a significant I/O reduction on swap device and,
548 in the case where decompressing from RAM is faster that swap device
549 reads, can also improve workload performance.
550
551 This is marked experimental because it is a new feature (as of
552 v3.11) that interacts heavily with memory reclaim. While these
553 interactions don't cause any known issues on simple memory setups,
554 they have not be fully explored on the large set of potential
555 configurations and workloads that exist.
556
557choice
558 prompt "Compressed cache for swap pages default compressor"
559 depends on ZSWAP
560 default ZSWAP_COMPRESSOR_DEFAULT_LZO
561 help
562 Selects the default compression algorithm for the compressed cache
563 for swap pages.
564
565 For an overview what kind of performance can be expected from
566 a particular compression algorithm please refer to the benchmarks
567 available at the following LWN page:
568 https://lwn.net/Articles/751795/
569
570 If in doubt, select 'LZO'.
571
572 The selection made here can be overridden by using the kernel
573 command line 'zswap.compressor=' option.
574
575config ZSWAP_COMPRESSOR_DEFAULT_DEFLATE
576 bool "Deflate"
577 select CRYPTO_DEFLATE
578 help
579 Use the Deflate algorithm as the default compression algorithm.
580
581config ZSWAP_COMPRESSOR_DEFAULT_LZO
582 bool "LZO"
583 select CRYPTO_LZO
584 help
585 Use the LZO algorithm as the default compression algorithm.
586
587config ZSWAP_COMPRESSOR_DEFAULT_842
588 bool "842"
589 select CRYPTO_842
590 help
591 Use the 842 algorithm as the default compression algorithm.
592
593config ZSWAP_COMPRESSOR_DEFAULT_LZ4
594 bool "LZ4"
595 select CRYPTO_LZ4
596 help
597 Use the LZ4 algorithm as the default compression algorithm.
598
599config ZSWAP_COMPRESSOR_DEFAULT_LZ4HC
600 bool "LZ4HC"
601 select CRYPTO_LZ4HC
602 help
603 Use the LZ4HC algorithm as the default compression algorithm.
604
605config ZSWAP_COMPRESSOR_DEFAULT_ZSTD
606 bool "zstd"
607 select CRYPTO_ZSTD
608 help
609 Use the zstd algorithm as the default compression algorithm.
610endchoice
611
612config ZSWAP_COMPRESSOR_DEFAULT
613 string
614 depends on ZSWAP
615 default "deflate" if ZSWAP_COMPRESSOR_DEFAULT_DEFLATE
616 default "lzo" if ZSWAP_COMPRESSOR_DEFAULT_LZO
617 default "842" if ZSWAP_COMPRESSOR_DEFAULT_842
618 default "lz4" if ZSWAP_COMPRESSOR_DEFAULT_LZ4
619 default "lz4hc" if ZSWAP_COMPRESSOR_DEFAULT_LZ4HC
620 default "zstd" if ZSWAP_COMPRESSOR_DEFAULT_ZSTD
621 default ""
622
623choice
624 prompt "Compressed cache for swap pages default allocator"
625 depends on ZSWAP
626 default ZSWAP_ZPOOL_DEFAULT_ZBUD
627 help
628 Selects the default allocator for the compressed cache for
629 swap pages.
630 The default is 'zbud' for compatibility, however please do
631 read the description of each of the allocators below before
632 making a right choice.
633
634 The selection made here can be overridden by using the kernel
635 command line 'zswap.zpool=' option.
636
637config ZSWAP_ZPOOL_DEFAULT_ZBUD
638 bool "zbud"
639 select ZBUD
640 help
641 Use the zbud allocator as the default allocator.
642
643config ZSWAP_ZPOOL_DEFAULT_Z3FOLD
644 bool "z3fold"
645 select Z3FOLD
646 help
647 Use the z3fold allocator as the default allocator.
648
649config ZSWAP_ZPOOL_DEFAULT_ZSMALLOC
650 bool "zsmalloc"
651 select ZSMALLOC
652 help
653 Use the zsmalloc allocator as the default allocator.
654endchoice
655
656config ZSWAP_ZPOOL_DEFAULT
657 string
658 depends on ZSWAP
659 default "zbud" if ZSWAP_ZPOOL_DEFAULT_ZBUD
660 default "z3fold" if ZSWAP_ZPOOL_DEFAULT_Z3FOLD
661 default "zsmalloc" if ZSWAP_ZPOOL_DEFAULT_ZSMALLOC
662 default ""
663
664config ZSWAP_DEFAULT_ON
665 bool "Enable the compressed cache for swap pages by default"
666 depends on ZSWAP
667 help
668 If selected, the compressed cache for swap pages will be enabled
669 at boot, otherwise it will be disabled.
670
671 The selection made here can be overridden by using the kernel
672 command line 'zswap.enabled=' option.
673
674config ZPOOL
675 tristate "Common API for compressed memory storage"
676 help
677 Compressed memory storage API. This allows using either zbud or
678 zsmalloc.
679
680config ZBUD
681 tristate "Low (Up to 2x) density storage for compressed pages"
682 help
683 A special purpose allocator for storing compressed pages.
684 It is designed to store up to two compressed pages per physical
685 page. While this design limits storage density, it has simple and
686 deterministic reclaim properties that make it preferable to a higher
687 density approach when reclaim will be used.
688
689config Z3FOLD
690 tristate "Up to 3x density storage for compressed pages"
691 depends on ZPOOL
692 help
693 A special purpose allocator for storing compressed pages.
694 It is designed to store up to three compressed pages per physical
695 page. It is a ZBUD derivative so the simplicity and determinism are
696 still there.
697
698config ZSMALLOC
699 tristate "Memory allocator for compressed pages"
700 depends on MMU
701 help
702 zsmalloc is a slab-based memory allocator designed to store
703 compressed RAM pages. zsmalloc uses virtual memory mapping
704 in order to reduce fragmentation. However, this results in a
705 non-standard allocator interface where a handle, not a pointer, is
706 returned by an alloc(). This handle must be mapped in order to
707 access the allocated space.
708
709config ZSMALLOC_PGTABLE_MAPPING
710 bool "Use page table mapping to access object in zsmalloc"
711 depends on ZSMALLOC=y
712 help
713 By default, zsmalloc uses a copy-based object mapping method to
714 access allocations that span two pages. However, if a particular
715 architecture (ex, ARM) performs VM mapping faster than copying,
716 then you should select this. This causes zsmalloc to use page table
717 mapping rather than copying for object mapping.
718
719 You can check speed with zsmalloc benchmark:
720 https://github.com/spartacus06/zsmapbench
721
722config ZSMALLOC_STAT
723 bool "Export zsmalloc statistics"
724 depends on ZSMALLOC
725 select DEBUG_FS
726 help
727 This option enables code in the zsmalloc to collect various
728 statistics about whats happening in zsmalloc and exports that
729 information to userspace via debugfs.
730 If unsure, say N.
731
732config GENERIC_EARLY_IOREMAP
733 bool
734
735config MAX_STACK_SIZE_MB
736 int "Maximum user stack size for 32-bit processes (MB)"
737 default 80
738 range 8 2048
739 depends on STACK_GROWSUP && (!64BIT || COMPAT)
740 help
741 This is the maximum stack size in Megabytes in the VM layout of 32-bit
742 user processes when the stack grows upwards (currently only on parisc
743 arch). The stack will be located at the highest memory address minus
744 the given value, unless the RLIMIT_STACK hard limit is changed to a
745 smaller value in which case that is used.
746
747 A sane initial value is 80 MB.
748
749config DEFERRED_STRUCT_PAGE_INIT
750 bool "Defer initialisation of struct pages to kthreads"
751 depends on SPARSEMEM
752 depends on !NEED_PER_CPU_KM
753 depends on 64BIT
754 select PADATA
755 help
756 Ordinarily all struct pages are initialised during early boot in a
757 single thread. On very large machines this can take a considerable
758 amount of time. If this option is set, large machines will bring up
759 a subset of memmap at boot and then initialise the rest in parallel.
760 This has a potential performance impact on tasks running early in the
761 lifetime of the system until these kthreads finish the
762 initialisation.
763
764config IDLE_PAGE_TRACKING
765 bool "Enable idle page tracking"
766 depends on SYSFS && MMU
767 select PAGE_EXTENSION if !64BIT
768 help
769 This feature allows to estimate the amount of user pages that have
770 not been touched during a given period of time. This information can
771 be useful to tune memory cgroup limits and/or for job placement
772 within a compute cluster.
773
774 See Documentation/admin-guide/mm/idle_page_tracking.rst for
775 more details.
776
777config ARCH_HAS_PTE_DEVMAP
778 bool
779
780config ZONE_DEVICE
781 bool "Device memory (pmem, HMM, etc...) hotplug support"
782 depends on MEMORY_HOTPLUG
783 depends on MEMORY_HOTREMOVE
784 depends on SPARSEMEM_VMEMMAP
785 depends on ARCH_HAS_PTE_DEVMAP
786 select XARRAY_MULTI
787
788 help
789 Device memory hotplug support allows for establishing pmem,
790 or other device driver discovered memory regions, in the
791 memmap. This allows pfn_to_page() lookups of otherwise
792 "device-physical" addresses which is needed for using a DAX
793 mapping in an O_DIRECT operation, among other things.
794
795 If FS_DAX is enabled, then say Y.
796
797config DEV_PAGEMAP_OPS
798 bool
799
800#
801# Helpers to mirror range of the CPU page tables of a process into device page
802# tables.
803#
804config HMM_MIRROR
805 bool
806 depends on MMU
807
808config DEVICE_PRIVATE
809 bool "Unaddressable device memory (GPU memory, ...)"
810 depends on ZONE_DEVICE
811 select DEV_PAGEMAP_OPS
812
813 help
814 Allows creation of struct pages to represent unaddressable device
815 memory; i.e., memory that is only accessible from the device (or
816 group of devices). You likely also want to select HMM_MIRROR.
817
818config FRAME_VECTOR
819 bool
820
821config ARCH_USES_HIGH_VMA_FLAGS
822 bool
823config ARCH_HAS_PKEYS
824 bool
825
826config PERCPU_STATS
827 bool "Collect percpu memory statistics"
828 help
829 This feature collects and exposes statistics via debugfs. The
830 information includes global and per chunk statistics, which can
831 be used to help understand percpu memory usage.
832
833config GUP_BENCHMARK
834 bool "Enable infrastructure for get_user_pages_fast() benchmarking"
835 help
836 Provides /sys/kernel/debug/gup_benchmark that helps with testing
837 performance of get_user_pages_fast().
838
839 See tools/testing/selftests/vm/gup_benchmark.c
840
841config GUP_GET_PTE_LOW_HIGH
842 bool
843
844config READ_ONLY_THP_FOR_FS
845 bool "Read-only THP for filesystems (EXPERIMENTAL)"
846 depends on TRANSPARENT_HUGEPAGE && SHMEM
847
848 help
849 Allow khugepaged to put read-only file-backed pages in THP.
850
851 This is marked experimental because it is a new feature. Write
852 support of file THPs will be developed in the next few release
853 cycles.
854
855config ARCH_HAS_PTE_SPECIAL
856 bool
857
858#
859# Some architectures require a special hugepage directory format that is
860# required to support multiple hugepage sizes. For example a4fe3ce76
861# "powerpc/mm: Allow more flexible layouts for hugepage pagetables"
862# introduced it on powerpc. This allows for a more flexible hugepage
863# pagetable layouts.
864#
865config ARCH_HAS_HUGEPD
866 bool
867
868config MAPPING_DIRTY_HELPERS
869 bool
870
871endmenu
1config SELECT_MEMORY_MODEL
2 def_bool y
3 depends on EXPERIMENTAL || ARCH_SELECT_MEMORY_MODEL
4
5choice
6 prompt "Memory model"
7 depends on SELECT_MEMORY_MODEL
8 default DISCONTIGMEM_MANUAL if ARCH_DISCONTIGMEM_DEFAULT
9 default SPARSEMEM_MANUAL if ARCH_SPARSEMEM_DEFAULT
10 default FLATMEM_MANUAL
11
12config FLATMEM_MANUAL
13 bool "Flat Memory"
14 depends on !(ARCH_DISCONTIGMEM_ENABLE || ARCH_SPARSEMEM_ENABLE) || ARCH_FLATMEM_ENABLE
15 help
16 This option allows you to change some of the ways that
17 Linux manages its memory internally. Most users will
18 only have one option here: FLATMEM. This is normal
19 and a correct option.
20
21 Some users of more advanced features like NUMA and
22 memory hotplug may have different options here.
23 DISCONTIGMEM is an more mature, better tested system,
24 but is incompatible with memory hotplug and may suffer
25 decreased performance over SPARSEMEM. If unsure between
26 "Sparse Memory" and "Discontiguous Memory", choose
27 "Discontiguous Memory".
28
29 If unsure, choose this option (Flat Memory) over any other.
30
31config DISCONTIGMEM_MANUAL
32 bool "Discontiguous Memory"
33 depends on ARCH_DISCONTIGMEM_ENABLE
34 help
35 This option provides enhanced support for discontiguous
36 memory systems, over FLATMEM. These systems have holes
37 in their physical address spaces, and this option provides
38 more efficient handling of these holes. However, the vast
39 majority of hardware has quite flat address spaces, and
40 can have degraded performance from the extra overhead that
41 this option imposes.
42
43 Many NUMA configurations will have this as the only option.
44
45 If unsure, choose "Flat Memory" over this option.
46
47config SPARSEMEM_MANUAL
48 bool "Sparse Memory"
49 depends on ARCH_SPARSEMEM_ENABLE
50 help
51 This will be the only option for some systems, including
52 memory hotplug systems. This is normal.
53
54 For many other systems, this will be an alternative to
55 "Discontiguous Memory". This option provides some potential
56 performance benefits, along with decreased code complexity,
57 but it is newer, and more experimental.
58
59 If unsure, choose "Discontiguous Memory" or "Flat Memory"
60 over this option.
61
62endchoice
63
64config DISCONTIGMEM
65 def_bool y
66 depends on (!SELECT_MEMORY_MODEL && ARCH_DISCONTIGMEM_ENABLE) || DISCONTIGMEM_MANUAL
67
68config SPARSEMEM
69 def_bool y
70 depends on (!SELECT_MEMORY_MODEL && ARCH_SPARSEMEM_ENABLE) || SPARSEMEM_MANUAL
71
72config FLATMEM
73 def_bool y
74 depends on (!DISCONTIGMEM && !SPARSEMEM) || FLATMEM_MANUAL
75
76config FLAT_NODE_MEM_MAP
77 def_bool y
78 depends on !SPARSEMEM
79
80#
81# Both the NUMA code and DISCONTIGMEM use arrays of pg_data_t's
82# to represent different areas of memory. This variable allows
83# those dependencies to exist individually.
84#
85config NEED_MULTIPLE_NODES
86 def_bool y
87 depends on DISCONTIGMEM || NUMA
88
89config HAVE_MEMORY_PRESENT
90 def_bool y
91 depends on ARCH_HAVE_MEMORY_PRESENT || SPARSEMEM
92
93#
94# SPARSEMEM_EXTREME (which is the default) does some bootmem
95# allocations when memory_present() is called. If this cannot
96# be done on your architecture, select this option. However,
97# statically allocating the mem_section[] array can potentially
98# consume vast quantities of .bss, so be careful.
99#
100# This option will also potentially produce smaller runtime code
101# with gcc 3.4 and later.
102#
103config SPARSEMEM_STATIC
104 bool
105
106#
107# Architecture platforms which require a two level mem_section in SPARSEMEM
108# must select this option. This is usually for architecture platforms with
109# an extremely sparse physical address space.
110#
111config SPARSEMEM_EXTREME
112 def_bool y
113 depends on SPARSEMEM && !SPARSEMEM_STATIC
114
115config SPARSEMEM_VMEMMAP_ENABLE
116 bool
117
118config SPARSEMEM_ALLOC_MEM_MAP_TOGETHER
119 def_bool y
120 depends on SPARSEMEM && X86_64
121
122config SPARSEMEM_VMEMMAP
123 bool "Sparse Memory virtual memmap"
124 depends on SPARSEMEM && SPARSEMEM_VMEMMAP_ENABLE
125 default y
126 help
127 SPARSEMEM_VMEMMAP uses a virtually mapped memmap to optimise
128 pfn_to_page and page_to_pfn operations. This is the most
129 efficient option when sufficient kernel resources are available.
130
131config HAVE_MEMBLOCK
132 boolean
133
134# eventually, we can have this option just 'select SPARSEMEM'
135config MEMORY_HOTPLUG
136 bool "Allow for memory hot-add"
137 depends on SPARSEMEM || X86_64_ACPI_NUMA
138 depends on HOTPLUG && ARCH_ENABLE_MEMORY_HOTPLUG
139 depends on (IA64 || X86 || PPC_BOOK3S_64 || SUPERH || S390)
140
141config MEMORY_HOTPLUG_SPARSE
142 def_bool y
143 depends on SPARSEMEM && MEMORY_HOTPLUG
144
145config MEMORY_HOTREMOVE
146 bool "Allow for memory hot remove"
147 depends on MEMORY_HOTPLUG && ARCH_ENABLE_MEMORY_HOTREMOVE
148 depends on MIGRATION
149
150#
151# If we have space for more page flags then we can enable additional
152# optimizations and functionality.
153#
154# Regular Sparsemem takes page flag bits for the sectionid if it does not
155# use a virtual memmap. Disable extended page flags for 32 bit platforms
156# that require the use of a sectionid in the page flags.
157#
158config PAGEFLAGS_EXTENDED
159 def_bool y
160 depends on 64BIT || SPARSEMEM_VMEMMAP || !SPARSEMEM
161
162# Heavily threaded applications may benefit from splitting the mm-wide
163# page_table_lock, so that faults on different parts of the user address
164# space can be handled with less contention: split it at this NR_CPUS.
165# Default to 4 for wider testing, though 8 might be more appropriate.
166# ARM's adjust_pte (unused if VIPT) depends on mm-wide page_table_lock.
167# PA-RISC 7xxx's spinlock_t would enlarge struct page from 32 to 44 bytes.
168# DEBUG_SPINLOCK and DEBUG_LOCK_ALLOC spinlock_t also enlarge struct page.
169#
170config SPLIT_PTLOCK_CPUS
171 int
172 default "999999" if ARM && !CPU_CACHE_VIPT
173 default "999999" if PARISC && !PA20
174 default "999999" if DEBUG_SPINLOCK || DEBUG_LOCK_ALLOC
175 default "4"
176
177#
178# support for memory compaction
179config COMPACTION
180 bool "Allow for memory compaction"
181 select MIGRATION
182 depends on MMU
183 help
184 Allows the compaction of memory for the allocation of huge pages.
185
186#
187# support for page migration
188#
189config MIGRATION
190 bool "Page migration"
191 def_bool y
192 depends on NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION
193 help
194 Allows the migration of the physical location of pages of processes
195 while the virtual addresses are not changed. This is useful in
196 two situations. The first is on NUMA systems to put pages nearer
197 to the processors accessing. The second is when allocating huge
198 pages as migration can relocate pages to satisfy a huge page
199 allocation instead of reclaiming.
200
201config PHYS_ADDR_T_64BIT
202 def_bool 64BIT || ARCH_PHYS_ADDR_T_64BIT
203
204config ZONE_DMA_FLAG
205 int
206 default "0" if !ZONE_DMA
207 default "1"
208
209config BOUNCE
210 def_bool y
211 depends on BLOCK && MMU && (ZONE_DMA || HIGHMEM)
212
213config NR_QUICK
214 int
215 depends on QUICKLIST
216 default "2" if AVR32
217 default "1"
218
219config VIRT_TO_BUS
220 def_bool y
221 depends on !ARCH_NO_VIRT_TO_BUS
222
223config MMU_NOTIFIER
224 bool
225
226config KSM
227 bool "Enable KSM for page merging"
228 depends on MMU
229 help
230 Enable Kernel Samepage Merging: KSM periodically scans those areas
231 of an application's address space that an app has advised may be
232 mergeable. When it finds pages of identical content, it replaces
233 the many instances by a single page with that content, so
234 saving memory until one or another app needs to modify the content.
235 Recommended for use with KVM, or with other duplicative applications.
236 See Documentation/vm/ksm.txt for more information: KSM is inactive
237 until a program has madvised that an area is MADV_MERGEABLE, and
238 root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set).
239
240config DEFAULT_MMAP_MIN_ADDR
241 int "Low address space to protect from user allocation"
242 depends on MMU
243 default 4096
244 help
245 This is the portion of low virtual memory which should be protected
246 from userspace allocation. Keeping a user from writing to low pages
247 can help reduce the impact of kernel NULL pointer bugs.
248
249 For most ia64, ppc64 and x86 users with lots of address space
250 a value of 65536 is reasonable and should cause no problems.
251 On arm and other archs it should not be higher than 32768.
252 Programs which use vm86 functionality or have some need to map
253 this low address space will need CAP_SYS_RAWIO or disable this
254 protection by setting the value to 0.
255
256 This value can be changed after boot using the
257 /proc/sys/vm/mmap_min_addr tunable.
258
259config ARCH_SUPPORTS_MEMORY_FAILURE
260 bool
261
262config MEMORY_FAILURE
263 depends on MMU
264 depends on ARCH_SUPPORTS_MEMORY_FAILURE
265 bool "Enable recovery from hardware memory errors"
266 help
267 Enables code to recover from some memory failures on systems
268 with MCA recovery. This allows a system to continue running
269 even when some of its memory has uncorrected errors. This requires
270 special hardware support and typically ECC memory.
271
272config HWPOISON_INJECT
273 tristate "HWPoison pages injector"
274 depends on MEMORY_FAILURE && DEBUG_KERNEL && PROC_FS
275 select PROC_PAGE_MONITOR
276
277config NOMMU_INITIAL_TRIM_EXCESS
278 int "Turn on mmap() excess space trimming before booting"
279 depends on !MMU
280 default 1
281 help
282 The NOMMU mmap() frequently needs to allocate large contiguous chunks
283 of memory on which to store mappings, but it can only ask the system
284 allocator for chunks in 2^N*PAGE_SIZE amounts - which is frequently
285 more than it requires. To deal with this, mmap() is able to trim off
286 the excess and return it to the allocator.
287
288 If trimming is enabled, the excess is trimmed off and returned to the
289 system allocator, which can cause extra fragmentation, particularly
290 if there are a lot of transient processes.
291
292 If trimming is disabled, the excess is kept, but not used, which for
293 long-term mappings means that the space is wasted.
294
295 Trimming can be dynamically controlled through a sysctl option
296 (/proc/sys/vm/nr_trim_pages) which specifies the minimum number of
297 excess pages there must be before trimming should occur, or zero if
298 no trimming is to occur.
299
300 This option specifies the initial value of this option. The default
301 of 1 says that all excess pages should be trimmed.
302
303 See Documentation/nommu-mmap.txt for more information.
304
305config TRANSPARENT_HUGEPAGE
306 bool "Transparent Hugepage Support"
307 depends on X86 && MMU
308 select COMPACTION
309 help
310 Transparent Hugepages allows the kernel to use huge pages and
311 huge tlb transparently to the applications whenever possible.
312 This feature can improve computing performance to certain
313 applications by speeding up page faults during memory
314 allocation, by reducing the number of tlb misses and by speeding
315 up the pagetable walking.
316
317 If memory constrained on embedded, you may want to say N.
318
319choice
320 prompt "Transparent Hugepage Support sysfs defaults"
321 depends on TRANSPARENT_HUGEPAGE
322 default TRANSPARENT_HUGEPAGE_ALWAYS
323 help
324 Selects the sysfs defaults for Transparent Hugepage Support.
325
326 config TRANSPARENT_HUGEPAGE_ALWAYS
327 bool "always"
328 help
329 Enabling Transparent Hugepage always, can increase the
330 memory footprint of applications without a guaranteed
331 benefit but it will work automatically for all applications.
332
333 config TRANSPARENT_HUGEPAGE_MADVISE
334 bool "madvise"
335 help
336 Enabling Transparent Hugepage madvise, will only provide a
337 performance improvement benefit to the applications using
338 madvise(MADV_HUGEPAGE) but it won't risk to increase the
339 memory footprint of applications without a guaranteed
340 benefit.
341endchoice
342
343#
344# UP and nommu archs use km based percpu allocator
345#
346config NEED_PER_CPU_KM
347 depends on !SMP
348 bool
349 default y
350
351config CLEANCACHE
352 bool "Enable cleancache driver to cache clean pages if tmem is present"
353 default n
354 help
355 Cleancache can be thought of as a page-granularity victim cache
356 for clean pages that the kernel's pageframe replacement algorithm
357 (PFRA) would like to keep around, but can't since there isn't enough
358 memory. So when the PFRA "evicts" a page, it first attempts to use
359 cleancache code to put the data contained in that page into
360 "transcendent memory", memory that is not directly accessible or
361 addressable by the kernel and is of unknown and possibly
362 time-varying size. And when a cleancache-enabled
363 filesystem wishes to access a page in a file on disk, it first
364 checks cleancache to see if it already contains it; if it does,
365 the page is copied into the kernel and a disk access is avoided.
366 When a transcendent memory driver is available (such as zcache or
367 Xen transcendent memory), a significant I/O reduction
368 may be achieved. When none is available, all cleancache calls
369 are reduced to a single pointer-compare-against-NULL resulting
370 in a negligible performance hit.
371
372 If unsure, say Y to enable cleancache