blob: 0b4352557dd54643c05ccc52e6c48de6b42a5a06 [file] [log] [blame]
Thomas Gleixnerec8f24b2019-05-19 13:07:45 +01001# SPDX-License-Identifier: GPL-2.0-only
Christoph Hellwig59e0b522018-07-31 13:39:35 +02002
3menu "Memory Management options"
4
Dave Hansene1785e82005-06-23 00:07:49 -07005config SELECT_MEMORY_MODEL
6 def_bool y
Kees Cooka8826ee2013-01-16 18:54:17 -08007 depends on ARCH_SELECT_MEMORY_MODEL
Dave Hansene1785e82005-06-23 00:07:49 -07008
Dave Hansen3a9da762005-06-23 00:07:42 -07009choice
10 prompt "Memory model"
Dave Hansene1785e82005-06-23 00:07:49 -070011 depends on SELECT_MEMORY_MODEL
12 default DISCONTIGMEM_MANUAL if ARCH_DISCONTIGMEM_DEFAULT
Andy Whitcroftd41dee32005-06-23 00:07:54 -070013 default SPARSEMEM_MANUAL if ARCH_SPARSEMEM_DEFAULT
Dave Hansene1785e82005-06-23 00:07:49 -070014 default FLATMEM_MANUAL
Mike Rapoportd66d1092019-05-13 17:23:05 -070015 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.
Dave Hansen3a9da762005-06-23 00:07:42 -070020
Dave Hansene1785e82005-06-23 00:07:49 -070021config FLATMEM_MANUAL
Dave Hansen3a9da762005-06-23 00:07:42 -070022 bool "Flat Memory"
Anton Blanchardc898ec12006-01-06 00:12:07 -080023 depends on !(ARCH_DISCONTIGMEM_ENABLE || ARCH_SPARSEMEM_ENABLE) || ARCH_FLATMEM_ENABLE
Dave Hansen3a9da762005-06-23 00:07:42 -070024 help
Mike Rapoportd66d1092019-05-13 17:23:05 -070025 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.
Dave Hansen3a9da762005-06-23 00:07:42 -070029
Mike Rapoportd66d1092019-05-13 17:23:05 -070030 For systems that have holes in their physical address
31 spaces and for features like NUMA and memory hotplug,
32 choose "Sparse Memory"
Andy Whitcroftd41dee32005-06-23 00:07:54 -070033
34 If unsure, choose this option (Flat Memory) over any other.
Dave Hansen3a9da762005-06-23 00:07:42 -070035
Dave Hansene1785e82005-06-23 00:07:49 -070036config DISCONTIGMEM_MANUAL
Dave Hansenf3519f92005-09-16 19:27:54 -070037 bool "Discontiguous Memory"
Dave Hansen3a9da762005-06-23 00:07:42 -070038 depends on ARCH_DISCONTIGMEM_ENABLE
39 help
Dave Hansen785dcd42005-06-23 00:07:50 -070040 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
Mike Rapoportd66d1092019-05-13 17:23:05 -070043 more efficient handling of these holes.
Dave Hansen785dcd42005-06-23 00:07:50 -070044
Mike Rapoportd66d1092019-05-13 17:23:05 -070045 Although "Discontiguous Memory" is still used by several
46 architectures, it is considered deprecated in favor of
47 "Sparse Memory".
Dave Hansen785dcd42005-06-23 00:07:50 -070048
Mike Rapoportd66d1092019-05-13 17:23:05 -070049 If unsure, choose "Sparse Memory" over this option.
Dave Hansen3a9da762005-06-23 00:07:42 -070050
Andy Whitcroftd41dee32005-06-23 00:07:54 -070051config 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
Mike Rapoportd66d1092019-05-13 17:23:05 -070056 memory hot-plug systems. This is normal.
Andy Whitcroftd41dee32005-06-23 00:07:54 -070057
Mike Rapoportd66d1092019-05-13 17:23:05 -070058 This option provides efficient support for systems with
59 holes is their physical address space and allows memory
60 hot-plug and hot-remove.
Andy Whitcroftd41dee32005-06-23 00:07:54 -070061
Mike Rapoportd66d1092019-05-13 17:23:05 -070062 If unsure, choose "Flat Memory" over this option.
Andy Whitcroftd41dee32005-06-23 00:07:54 -070063
Dave Hansen3a9da762005-06-23 00:07:42 -070064endchoice
65
Dave Hansene1785e82005-06-23 00:07:49 -070066config DISCONTIGMEM
67 def_bool y
68 depends on (!SELECT_MEMORY_MODEL && ARCH_DISCONTIGMEM_ENABLE) || DISCONTIGMEM_MANUAL
69
Andy Whitcroftd41dee32005-06-23 00:07:54 -070070config SPARSEMEM
71 def_bool y
Russell King1a83e172009-10-26 16:50:12 -070072 depends on (!SELECT_MEMORY_MODEL && ARCH_SPARSEMEM_ENABLE) || SPARSEMEM_MANUAL
Andy Whitcroftd41dee32005-06-23 00:07:54 -070073
Dave Hansene1785e82005-06-23 00:07:49 -070074config FLATMEM
75 def_bool y
Andy Whitcroftd41dee32005-06-23 00:07:54 -070076 depends on (!DISCONTIGMEM && !SPARSEMEM) || FLATMEM_MANUAL
77
78config FLAT_NODE_MEM_MAP
79 def_bool y
80 depends on !SPARSEMEM
Dave Hansene1785e82005-06-23 00:07:49 -070081
Dave Hansen93b75042005-06-23 00:07:47 -070082#
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
Andy Whitcroftaf705362005-06-23 00:07:53 -070090
91config HAVE_MEMORY_PRESENT
92 def_bool y
Andy Whitcroftd41dee32005-06-23 00:07:54 -070093 depends on ARCH_HAVE_MEMORY_PRESENT || SPARSEMEM
Bob Picco802f1922005-09-03 15:54:26 -070094
95#
Bob Picco3e347262005-09-03 15:54:28 -070096# SPARSEMEM_EXTREME (which is the default) does some bootmem
Matt LaPlante84eb8d02006-10-03 22:53:09 +020097# allocations when memory_present() is called. If this cannot
Bob Picco3e347262005-09-03 15:54:28 -070098# be done on your architecture, select this option. However,
99# statically allocating the mem_section[] array can potentially
100# consume vast quantities of .bss, so be careful.
101#
102# This option will also potentially produce smaller runtime code
103# with gcc 3.4 and later.
104#
105config SPARSEMEM_STATIC
Jan Beulich9ba16082008-10-15 22:01:38 -0700106 bool
Bob Picco3e347262005-09-03 15:54:28 -0700107
108#
Matt LaPlante44c09202006-10-03 22:34:14 +0200109# Architecture platforms which require a two level mem_section in SPARSEMEM
Bob Picco802f1922005-09-03 15:54:26 -0700110# must select this option. This is usually for architecture platforms with
111# an extremely sparse physical address space.
112#
Bob Picco3e347262005-09-03 15:54:28 -0700113config SPARSEMEM_EXTREME
114 def_bool y
115 depends on SPARSEMEM && !SPARSEMEM_STATIC
Hugh Dickins4c21e2f2005-10-29 18:16:40 -0700116
Andy Whitcroft29c71112007-10-16 01:24:14 -0700117config SPARSEMEM_VMEMMAP_ENABLE
Jan Beulich9ba16082008-10-15 22:01:38 -0700118 bool
Andy Whitcroft29c71112007-10-16 01:24:14 -0700119
120config SPARSEMEM_VMEMMAP
Geoff Levanda5ee6da2007-12-17 16:19:53 -0800121 bool "Sparse Memory virtual memmap"
122 depends on SPARSEMEM && SPARSEMEM_VMEMMAP_ENABLE
123 default y
124 help
125 SPARSEMEM_VMEMMAP uses a virtually mapped memmap to optimise
126 pfn_to_page and page_to_pfn operations. This is the most
127 efficient option when sufficient kernel resources are available.
Andy Whitcroft29c71112007-10-16 01:24:14 -0700128
Tejun Heo7c0caeb2011-07-14 11:43:42 +0200129config HAVE_MEMBLOCK_NODE_MAP
Christoph Jaeger6341e622014-12-20 15:41:11 -0500130 bool
Tejun Heo7c0caeb2011-07-14 11:43:42 +0200131
Philipp Hachtmann70210ed2014-01-29 18:16:01 +0100132config HAVE_MEMBLOCK_PHYS_MAP
Christoph Jaeger6341e622014-12-20 15:41:11 -0500133 bool
Philipp Hachtmann70210ed2014-01-29 18:16:01 +0100134
Christoph Hellwig67a929e2019-07-11 20:57:14 -0700135config HAVE_FAST_GUP
Christoph Hellwig050a9ad2019-07-11 20:57:21 -0700136 depends on MMU
Christoph Jaeger6341e622014-12-20 15:41:11 -0500137 bool
Steve Capper2667f502014-10-09 15:29:14 -0700138
Mike Rapoport350e88b2019-05-13 17:22:59 -0700139config ARCH_KEEP_MEMBLOCK
Christoph Jaeger6341e622014-12-20 15:41:11 -0500140 bool
Tejun Heoc378ddd2011-07-14 11:46:03 +0200141
Minchan Kimee6f5092012-07-31 16:43:50 -0700142config MEMORY_ISOLATION
Christoph Jaeger6341e622014-12-20 15:41:11 -0500143 bool
Minchan Kimee6f5092012-07-31 16:43:50 -0700144
Yasuaki Ishimatsu46723bf2013-02-22 16:33:00 -0800145#
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
Dave Hansen3947be12005-10-29 18:16:54 -0700152# eventually, we can have this option just 'select SPARSEMEM'
153config MEMORY_HOTPLUG
154 bool "Allow for memory hot-add"
Keith Manntheyec69acb2006-09-30 23:27:05 -0700155 depends on SPARSEMEM || X86_64_ACPI_NUMA
Stephen Rothwell40b31362013-05-21 13:49:35 +1000156 depends on ARCH_ENABLE_MEMORY_HOTPLUG
Dave Hansen3947be12005-10-29 18:16:54 -0700157
Keith Manntheyec69acb2006-09-30 23:27:05 -0700158config MEMORY_HOTPLUG_SPARSE
159 def_bool y
160 depends on SPARSEMEM && MEMORY_HOTPLUG
161
Vitaly Kuznetsov8604d9e2016-05-19 17:13:03 -0700162config MEMORY_HOTPLUG_DEFAULT_ONLINE
163 bool "Online the newly added memory blocks by default"
Vitaly Kuznetsov8604d9e2016-05-19 17:13:03 -0700164 depends on MEMORY_HOTPLUG
165 help
166 This option sets the default policy setting for memory hotplug
167 onlining policy (/sys/devices/system/memory/auto_online_blocks) which
168 determines what happens to newly added memory regions. Policy setting
169 can always be changed at runtime.
Mauro Carvalho Chehabcb1aaeb2019-06-07 15:54:32 -0300170 See Documentation/admin-guide/mm/memory-hotplug.rst for more information.
Vitaly Kuznetsov8604d9e2016-05-19 17:13:03 -0700171
172 Say Y here if you want all hot-plugged memory blocks to appear in
173 'online' state by default.
174 Say N here if you want the default policy to keep all hot-plugged
175 memory blocks in 'offline' state.
176
KAMEZAWA Hiroyuki0c0e6192007-10-16 01:26:12 -0700177config MEMORY_HOTREMOVE
178 bool "Allow for memory hot remove"
Yasuaki Ishimatsu46723bf2013-02-22 16:33:00 -0800179 select MEMORY_ISOLATION
Nathan Fontenotf7e33342013-09-27 10:18:09 -0500180 select HAVE_BOOTMEM_INFO_NODE if (X86_64 || PPC64)
KAMEZAWA Hiroyuki0c0e6192007-10-16 01:26:12 -0700181 depends on MEMORY_HOTPLUG && ARCH_ENABLE_MEMORY_HOTREMOVE
182 depends on MIGRATION
183
Hugh Dickins4c21e2f2005-10-29 18:16:40 -0700184# Heavily threaded applications may benefit from splitting the mm-wide
185# page_table_lock, so that faults on different parts of the user address
186# space can be handled with less contention: split it at this NR_CPUS.
187# Default to 4 for wider testing, though 8 might be more appropriate.
188# ARM's adjust_pte (unused if VIPT) depends on mm-wide page_table_lock.
Hugh Dickins7b6ac9d2005-11-23 13:37:37 -0800189# PA-RISC 7xxx's spinlock_t would enlarge struct page from 32 to 44 bytes.
Hugh Dickinsa70caa82009-12-14 17:59:02 -0800190# DEBUG_SPINLOCK and DEBUG_LOCK_ALLOC spinlock_t also enlarge struct page.
Hugh Dickins4c21e2f2005-10-29 18:16:40 -0700191#
192config SPLIT_PTLOCK_CPUS
193 int
Kirill A. Shutemov91645502014-04-07 15:37:14 -0700194 default "999999" if !MMU
Hugh Dickinsa70caa82009-12-14 17:59:02 -0800195 default "999999" if ARM && !CPU_CACHE_VIPT
196 default "999999" if PARISC && !PA20
Hugh Dickins4c21e2f2005-10-29 18:16:40 -0700197 default "4"
Christoph Lameter7cbe34c2006-01-08 01:00:49 -0800198
Kirill A. Shutemove009bb32013-11-14 14:31:07 -0800199config ARCH_ENABLE_SPLIT_PMD_PTLOCK
Christoph Jaeger6341e622014-12-20 15:41:11 -0500200 bool
Kirill A. Shutemove009bb32013-11-14 14:31:07 -0800201
Christoph Lameter7cbe34c2006-01-08 01:00:49 -0800202#
Konstantin Khlebnikov09316c02014-10-09 15:29:32 -0700203# support for memory balloon
204config MEMORY_BALLOON
Christoph Jaeger6341e622014-12-20 15:41:11 -0500205 bool
Konstantin Khlebnikov09316c02014-10-09 15:29:32 -0700206
207#
Rafael Aquini18468d92012-12-11 16:02:38 -0800208# support for memory balloon compaction
209config BALLOON_COMPACTION
210 bool "Allow for balloon memory compaction/migration"
211 def_bool y
Konstantin Khlebnikov09316c02014-10-09 15:29:32 -0700212 depends on COMPACTION && MEMORY_BALLOON
Rafael Aquini18468d92012-12-11 16:02:38 -0800213 help
214 Memory fragmentation introduced by ballooning might reduce
215 significantly the number of 2MB contiguous memory blocks that can be
216 used within a guest, thus imposing performance penalties associated
217 with the reduced number of transparent huge pages that could be used
218 by the guest workload. Allowing the compaction & migration for memory
219 pages enlisted as being part of memory balloon devices avoids the
220 scenario aforementioned and helps improving memory defragmentation.
221
222#
Mel Gormane9e96b32010-05-24 14:32:21 -0700223# support for memory compaction
224config COMPACTION
225 bool "Allow for memory compaction"
Rik van Riel05106e62012-10-08 16:33:03 -0700226 def_bool y
Mel Gormane9e96b32010-05-24 14:32:21 -0700227 select MIGRATION
Andrea Arcangeli33a93872011-01-25 15:07:25 -0800228 depends on MMU
Mel Gormane9e96b32010-05-24 14:32:21 -0700229 help
Michal Hockob32eaf72016-08-25 15:17:05 -0700230 Compaction is the only memory management component to form
231 high order (larger physically contiguous) memory blocks
232 reliably. The page allocator relies on compaction heavily and
233 the lack of the feature can lead to unexpected OOM killer
234 invocations for high order memory requests. You shouldn't
235 disable this option unless there really is a strong reason for
236 it and then we would be really interested to hear about that at
237 linux-mm@kvack.org.
Mel Gormane9e96b32010-05-24 14:32:21 -0700238
239#
Christoph Lameter7cbe34c2006-01-08 01:00:49 -0800240# support for page migration
241#
242config MIGRATION
Christoph Lameterb20a3502006-03-22 00:09:12 -0800243 bool "Page migration"
Christoph Lameter6c5240a2006-06-23 02:03:37 -0700244 def_bool y
Chen Gangde32a812013-09-12 15:14:08 -0700245 depends on (NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION || CMA) && MMU
Christoph Lameterb20a3502006-03-22 00:09:12 -0800246 help
247 Allows the migration of the physical location of pages of processes
Mel Gormane9e96b32010-05-24 14:32:21 -0700248 while the virtual addresses are not changed. This is useful in
249 two situations. The first is on NUMA systems to put pages nearer
250 to the processors accessing. The second is when allocating huge
251 pages as migration can relocate pages to satisfy a huge page
252 allocation instead of reclaiming.
Greg Kroah-Hartman6550e072006-06-12 17:11:31 -0700253
Naoya Horiguchic177c812014-06-04 16:05:35 -0700254config ARCH_ENABLE_HUGEPAGE_MIGRATION
Christoph Jaeger6341e622014-12-20 15:41:11 -0500255 bool
Naoya Horiguchic177c812014-06-04 16:05:35 -0700256
Naoya Horiguchi9c670ea2017-09-08 16:10:53 -0700257config ARCH_ENABLE_THP_MIGRATION
258 bool
259
Alexandre Ghiti8df995f2019-05-13 17:19:00 -0700260config CONTIG_ALLOC
261 def_bool (MEMORY_ISOLATION && COMPACTION) || CMA
262
Jeremy Fitzhardinge600715d2008-09-11 01:31:45 -0700263config PHYS_ADDR_T_64BIT
Christoph Hellwigd4a451d2018-04-03 16:24:20 +0200264 def_bool 64BIT
Jeremy Fitzhardinge600715d2008-09-11 01:31:45 -0700265
Christoph Lameter2a7326b2007-07-17 04:03:37 -0700266config BOUNCE
Vinayak Menon9ca24e22013-04-29 15:08:55 -0700267 bool "Enable bounce buffers"
268 default y
Christoph Lameter2a7326b2007-07-17 04:03:37 -0700269 depends on BLOCK && MMU && (ZONE_DMA || HIGHMEM)
Vinayak Menon9ca24e22013-04-29 15:08:55 -0700270 help
271 Enable bounce buffers for devices that cannot access
272 the full range of memory available to the CPU. Enabled
273 by default when ZONE_DMA or HIGHMEM is selected, but you
274 may say n to override this.
Christoph Lameter2a7326b2007-07-17 04:03:37 -0700275
Christoph Lameter6225e932007-05-06 14:49:50 -0700276config NR_QUICK
277 int
278 depends on QUICKLIST
279 default "1"
Stephen Rothwellf057eac2007-07-15 23:40:05 -0700280
281config VIRT_TO_BUS
Stephen Rothwell4febd952013-03-07 15:48:16 +1100282 bool
283 help
284 An architecture should select this if it implements the
285 deprecated interface virt_to_bus(). All new architectures
286 should probably not select this.
287
Andrea Arcangelicddb8a52008-07-28 15:46:29 -0700288
289config MMU_NOTIFIER
290 bool
Pranith Kumar83fe27e2014-12-05 11:24:45 -0500291 select SRCU
David Howellsfc4d5c22009-05-06 16:03:05 -0700292
Hugh Dickinsf8af4da2009-09-21 17:01:57 -0700293config KSM
294 bool "Enable KSM for page merging"
295 depends on MMU
Timofey Titovets59e1a2f42018-12-28 00:34:05 -0800296 select XXHASH
Hugh Dickinsf8af4da2009-09-21 17:01:57 -0700297 help
298 Enable Kernel Samepage Merging: KSM periodically scans those areas
299 of an application's address space that an app has advised may be
300 mergeable. When it finds pages of identical content, it replaces
Hugh Dickinsd0f209f2009-12-14 17:59:34 -0800301 the many instances by a single page with that content, so
Hugh Dickinsf8af4da2009-09-21 17:01:57 -0700302 saving memory until one or another app needs to modify the content.
303 Recommended for use with KVM, or with other duplicative applications.
Mike Rapoportad56b732018-03-21 21:22:47 +0200304 See Documentation/vm/ksm.rst for more information: KSM is inactive
Hugh Dickinsc73602a2009-10-07 16:32:22 -0700305 until a program has madvised that an area is MADV_MERGEABLE, and
306 root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set).
Hugh Dickinsf8af4da2009-09-21 17:01:57 -0700307
Christoph Lametere0a94c22009-06-03 16:04:31 -0400308config DEFAULT_MMAP_MIN_ADDR
309 int "Low address space to protect from user allocation"
David Howells6e141542009-12-15 19:27:45 +0000310 depends on MMU
Christoph Lametere0a94c22009-06-03 16:04:31 -0400311 default 4096
312 help
313 This is the portion of low virtual memory which should be protected
314 from userspace allocation. Keeping a user from writing to low pages
315 can help reduce the impact of kernel NULL pointer bugs.
316
317 For most ia64, ppc64 and x86 users with lots of address space
318 a value of 65536 is reasonable and should cause no problems.
319 On arm and other archs it should not be higher than 32768.
Eric Paris788084a2009-07-31 12:54:11 -0400320 Programs which use vm86 functionality or have some need to map
321 this low address space will need CAP_SYS_RAWIO or disable this
322 protection by setting the value to 0.
Christoph Lametere0a94c22009-06-03 16:04:31 -0400323
324 This value can be changed after boot using the
325 /proc/sys/vm/mmap_min_addr tunable.
326
Linus Torvaldsd949f362009-09-26 09:35:07 -0700327config ARCH_SUPPORTS_MEMORY_FAILURE
328 bool
Christoph Lametere0a94c22009-06-03 16:04:31 -0400329
Andi Kleen6a460792009-09-16 11:50:15 +0200330config MEMORY_FAILURE
331 depends on MMU
Linus Torvaldsd949f362009-09-26 09:35:07 -0700332 depends on ARCH_SUPPORTS_MEMORY_FAILURE
Andi Kleen6a460792009-09-16 11:50:15 +0200333 bool "Enable recovery from hardware memory errors"
Minchan Kimee6f5092012-07-31 16:43:50 -0700334 select MEMORY_ISOLATION
Xie XiuQi97f0b132015-06-24 16:57:36 -0700335 select RAS
Andi Kleen6a460792009-09-16 11:50:15 +0200336 help
337 Enables code to recover from some memory failures on systems
338 with MCA recovery. This allows a system to continue running
339 even when some of its memory has uncorrected errors. This requires
340 special hardware support and typically ECC memory.
341
Andi Kleencae681f2009-09-16 11:50:17 +0200342config HWPOISON_INJECT
Andi Kleen413f9ef2009-12-16 12:20:00 +0100343 tristate "HWPoison pages injector"
Andi Kleen27df5062009-12-21 19:56:42 +0100344 depends on MEMORY_FAILURE && DEBUG_KERNEL && PROC_FS
Wu Fengguang478c5ff2009-12-16 12:19:59 +0100345 select PROC_PAGE_MONITOR
Andi Kleencae681f2009-09-16 11:50:17 +0200346
David Howellsfc4d5c22009-05-06 16:03:05 -0700347config NOMMU_INITIAL_TRIM_EXCESS
348 int "Turn on mmap() excess space trimming before booting"
349 depends on !MMU
350 default 1
351 help
352 The NOMMU mmap() frequently needs to allocate large contiguous chunks
353 of memory on which to store mappings, but it can only ask the system
354 allocator for chunks in 2^N*PAGE_SIZE amounts - which is frequently
355 more than it requires. To deal with this, mmap() is able to trim off
356 the excess and return it to the allocator.
357
358 If trimming is enabled, the excess is trimmed off and returned to the
359 system allocator, which can cause extra fragmentation, particularly
360 if there are a lot of transient processes.
361
362 If trimming is disabled, the excess is kept, but not used, which for
363 long-term mappings means that the space is wasted.
364
365 Trimming can be dynamically controlled through a sysctl option
366 (/proc/sys/vm/nr_trim_pages) which specifies the minimum number of
367 excess pages there must be before trimming should occur, or zero if
368 no trimming is to occur.
369
370 This option specifies the initial value of this option. The default
371 of 1 says that all excess pages should be trimmed.
372
373 See Documentation/nommu-mmap.txt for more information.
Tejun Heobbddff02010-09-03 18:22:48 +0200374
Andrea Arcangeli4c76d9d2011-01-13 15:46:39 -0800375config TRANSPARENT_HUGEPAGE
Andrea Arcangeli13ece882011-01-13 15:47:07 -0800376 bool "Transparent Hugepage Support"
Gerald Schaefer15626062012-10-08 16:30:04 -0700377 depends on HAVE_ARCH_TRANSPARENT_HUGEPAGE
Andrea Arcangeli5d689242011-01-13 15:47:07 -0800378 select COMPACTION
Matthew Wilcox3a08cd52018-09-22 16:14:30 -0400379 select XARRAY_MULTI
Andrea Arcangeli4c76d9d2011-01-13 15:46:39 -0800380 help
381 Transparent Hugepages allows the kernel to use huge pages and
382 huge tlb transparently to the applications whenever possible.
383 This feature can improve computing performance to certain
384 applications by speeding up page faults during memory
385 allocation, by reducing the number of tlb misses and by speeding
386 up the pagetable walking.
387
388 If memory constrained on embedded, you may want to say N.
389
Andrea Arcangeli13ece882011-01-13 15:47:07 -0800390choice
391 prompt "Transparent Hugepage Support sysfs defaults"
392 depends on TRANSPARENT_HUGEPAGE
393 default TRANSPARENT_HUGEPAGE_ALWAYS
394 help
395 Selects the sysfs defaults for Transparent Hugepage Support.
396
397 config TRANSPARENT_HUGEPAGE_ALWAYS
398 bool "always"
399 help
400 Enabling Transparent Hugepage always, can increase the
401 memory footprint of applications without a guaranteed
402 benefit but it will work automatically for all applications.
403
404 config TRANSPARENT_HUGEPAGE_MADVISE
405 bool "madvise"
406 help
407 Enabling Transparent Hugepage madvise, will only provide a
408 performance improvement benefit to the applications using
409 madvise(MADV_HUGEPAGE) but it won't risk to increase the
410 memory footprint of applications without a guaranteed
411 benefit.
412endchoice
413
Huang Ying38d8b4e2017-07-06 15:37:18 -0700414config ARCH_WANTS_THP_SWAP
415 def_bool n
416
417config THP_SWAP
418 def_bool y
Huang Ying14fef282018-08-17 15:49:41 -0700419 depends on TRANSPARENT_HUGEPAGE && ARCH_WANTS_THP_SWAP && SWAP
Huang Ying38d8b4e2017-07-06 15:37:18 -0700420 help
421 Swap transparent huge pages in one piece, without splitting.
Huang Ying14fef282018-08-17 15:49:41 -0700422 XXX: For now, swap cluster backing transparent huge page
423 will be split after swapout.
Huang Ying38d8b4e2017-07-06 15:37:18 -0700424
425 For selection by architectures with reasonable THP sizes.
426
Kirill A. Shutemove496cf32016-07-26 15:26:35 -0700427config TRANSPARENT_HUGE_PAGECACHE
428 def_bool y
Aneesh Kumar K.V953c66c2016-12-12 16:44:32 -0800429 depends on TRANSPARENT_HUGEPAGE
Kirill A. Shutemove496cf32016-07-26 15:26:35 -0700430
431#
Tejun Heobbddff02010-09-03 18:22:48 +0200432# UP and nommu archs use km based percpu allocator
433#
434config NEED_PER_CPU_KM
435 depends on !SMP
436 bool
437 default y
Dan Magenheimer077b1f82011-05-26 10:01:36 -0600438
439config CLEANCACHE
440 bool "Enable cleancache driver to cache clean pages if tmem is present"
Dan Magenheimer077b1f82011-05-26 10:01:36 -0600441 help
442 Cleancache can be thought of as a page-granularity victim cache
443 for clean pages that the kernel's pageframe replacement algorithm
444 (PFRA) would like to keep around, but can't since there isn't enough
445 memory. So when the PFRA "evicts" a page, it first attempts to use
Michael Witten140a1ef2011-06-10 03:57:26 +0000446 cleancache code to put the data contained in that page into
Dan Magenheimer077b1f82011-05-26 10:01:36 -0600447 "transcendent memory", memory that is not directly accessible or
448 addressable by the kernel and is of unknown and possibly
449 time-varying size. And when a cleancache-enabled
450 filesystem wishes to access a page in a file on disk, it first
451 checks cleancache to see if it already contains it; if it does,
452 the page is copied into the kernel and a disk access is avoided.
453 When a transcendent memory driver is available (such as zcache or
454 Xen transcendent memory), a significant I/O reduction
455 may be achieved. When none is available, all cleancache calls
456 are reduced to a single pointer-compare-against-NULL resulting
457 in a negligible performance hit.
458
459 If unsure, say Y to enable cleancache
Dan Magenheimer27c6aec2012-04-09 17:10:34 -0600460
461config FRONTSWAP
462 bool "Enable frontswap to cache swap pages if tmem is present"
463 depends on SWAP
Dan Magenheimer27c6aec2012-04-09 17:10:34 -0600464 help
465 Frontswap is so named because it can be thought of as the opposite
466 of a "backing" store for a swap device. The data is stored into
467 "transcendent memory", memory that is not directly accessible or
468 addressable by the kernel and is of unknown and possibly
469 time-varying size. When space in transcendent memory is available,
470 a significant swap I/O reduction may be achieved. When none is
471 available, all frontswap calls are reduced to a single pointer-
472 compare-against-NULL resulting in a negligible performance hit
473 and swap data is stored as normal on the matching swap device.
474
475 If unsure, say Y to enable frontswap.
Aneesh Kumar K.Vf825c732013-07-02 11:15:15 +0530476
477config CMA
478 bool "Contiguous Memory Allocator"
Mike Rapoportaca52c32018-10-30 15:07:44 -0700479 depends on MMU
Aneesh Kumar K.Vf825c732013-07-02 11:15:15 +0530480 select MIGRATION
481 select MEMORY_ISOLATION
482 help
483 This enables the Contiguous Memory Allocator which allows other
484 subsystems to allocate big physically-contiguous blocks of memory.
485 CMA reserves a region of memory and allows only movable pages to
486 be allocated from it. This way, the kernel can use the memory for
487 pagecache and when a subsystem requests for contiguous area, the
488 allocated pages are migrated away to serve the contiguous request.
489
490 If unsure, say "n".
491
492config CMA_DEBUG
493 bool "CMA debug messages (DEVELOPMENT)"
494 depends on DEBUG_KERNEL && CMA
495 help
496 Turns on debug messages in CMA. This produces KERN_DEBUG
497 messages for every CMA call as well as various messages while
498 processing calls such as dma_alloc_from_contiguous().
499 This option does not affect warning and error messages.
Alexander Grafbf550fc2013-08-29 00:41:59 +0200500
Sasha Levin28b24c12015-04-14 15:44:57 -0700501config CMA_DEBUGFS
502 bool "CMA debugfs interface"
503 depends on CMA && DEBUG_FS
504 help
505 Turns on the DebugFS interface for CMA.
506
Joonsoo Kima2541292014-08-06 16:05:25 -0700507config CMA_AREAS
508 int "Maximum count of the CMA areas"
509 depends on CMA
510 default 7
511 help
512 CMA allows to create CMA areas for particular purpose, mainly,
513 used as device private area. This parameter sets the maximum
514 number of CMA area in the system.
515
516 If unsure, leave the default value "7".
517
Dan Streetmanaf8d4172014-08-06 16:08:36 -0700518config MEM_SOFT_DIRTY
519 bool "Track memory changes"
520 depends on CHECKPOINT_RESTORE && HAVE_ARCH_SOFT_DIRTY && PROC_FS
521 select PROC_PAGE_MONITOR
Seth Jennings4e2e2772013-07-10 16:04:55 -0700522 help
Dan Streetmanaf8d4172014-08-06 16:08:36 -0700523 This option enables memory changes tracking by introducing a
524 soft-dirty bit on pte-s. This bit it set when someone writes
525 into a page just as regular dirty bit, but unlike the latter
526 it can be cleared by hands.
527
Mike Rapoport1ad13352018-04-18 11:07:49 +0300528 See Documentation/admin-guide/mm/soft-dirty.rst for more details.
Seth Jennings4e2e2772013-07-10 16:04:55 -0700529
Seth Jennings2b281112013-07-10 16:05:03 -0700530config ZSWAP
531 bool "Compressed cache for swap pages (EXPERIMENTAL)"
532 depends on FRONTSWAP && CRYPTO=y
533 select CRYPTO_LZO
Dan Streetman12d79d62014-08-06 16:08:40 -0700534 select ZPOOL
Seth Jennings2b281112013-07-10 16:05:03 -0700535 help
536 A lightweight compressed cache for swap pages. It takes
537 pages that are in the process of being swapped out and attempts to
538 compress them into a dynamically allocated RAM-based memory pool.
539 This can result in a significant I/O reduction on swap device and,
540 in the case where decompressing from RAM is faster that swap device
541 reads, can also improve workload performance.
542
543 This is marked experimental because it is a new feature (as of
544 v3.11) that interacts heavily with memory reclaim. While these
545 interactions don't cause any known issues on simple memory setups,
546 they have not be fully explored on the large set of potential
547 configurations and workloads that exist.
548
Dan Streetmanaf8d4172014-08-06 16:08:36 -0700549config ZPOOL
550 tristate "Common API for compressed memory storage"
Pavel Emelyanov0f8975e2013-07-03 15:01:20 -0700551 help
Dan Streetmanaf8d4172014-08-06 16:08:36 -0700552 Compressed memory storage API. This allows using either zbud or
553 zsmalloc.
Pavel Emelyanov0f8975e2013-07-03 15:01:20 -0700554
Dan Streetmanaf8d4172014-08-06 16:08:36 -0700555config ZBUD
Vitaly Wool9a001fc2016-05-20 16:58:30 -0700556 tristate "Low (Up to 2x) density storage for compressed pages"
Dan Streetmanaf8d4172014-08-06 16:08:36 -0700557 help
558 A special purpose allocator for storing compressed pages.
559 It is designed to store up to two compressed pages per physical
560 page. While this design limits storage density, it has simple and
561 deterministic reclaim properties that make it preferable to a higher
562 density approach when reclaim will be used.
Minchan Kimbcf16472014-01-30 15:45:50 -0800563
Vitaly Wool9a001fc2016-05-20 16:58:30 -0700564config Z3FOLD
565 tristate "Up to 3x density storage for compressed pages"
566 depends on ZPOOL
Vitaly Wool9a001fc2016-05-20 16:58:30 -0700567 help
568 A special purpose allocator for storing compressed pages.
569 It is designed to store up to three compressed pages per physical
570 page. It is a ZBUD derivative so the simplicity and determinism are
571 still there.
572
Minchan Kimbcf16472014-01-30 15:45:50 -0800573config ZSMALLOC
Minchan Kimd867f202014-06-04 16:11:10 -0700574 tristate "Memory allocator for compressed pages"
Minchan Kimbcf16472014-01-30 15:45:50 -0800575 depends on MMU
Minchan Kimbcf16472014-01-30 15:45:50 -0800576 help
577 zsmalloc is a slab-based memory allocator designed to store
578 compressed RAM pages. zsmalloc uses virtual memory mapping
579 in order to reduce fragmentation. However, this results in a
580 non-standard allocator interface where a handle, not a pointer, is
581 returned by an alloc(). This handle must be mapped in order to
582 access the allocated space.
583
584config PGTABLE_MAPPING
585 bool "Use page table mapping to access object in zsmalloc"
586 depends on ZSMALLOC
587 help
588 By default, zsmalloc uses a copy-based object mapping method to
589 access allocations that span two pages. However, if a particular
590 architecture (ex, ARM) performs VM mapping faster than copying,
591 then you should select this. This causes zsmalloc to use page table
592 mapping rather than copying for object mapping.
593
Ben Hutchings2216ee82014-03-10 15:49:46 -0700594 You can check speed with zsmalloc benchmark:
595 https://github.com/spartacus06/zsmapbench
Mark Salter9e5c33d2014-04-07 15:39:48 -0700596
Ganesh Mahendran0f050d92015-02-12 15:00:54 -0800597config ZSMALLOC_STAT
598 bool "Export zsmalloc statistics"
599 depends on ZSMALLOC
600 select DEBUG_FS
601 help
602 This option enables code in the zsmalloc to collect various
603 statistics about whats happening in zsmalloc and exports that
604 information to userspace via debugfs.
605 If unsure, say N.
606
Mark Salter9e5c33d2014-04-07 15:39:48 -0700607config GENERIC_EARLY_IOREMAP
608 bool
Helge Deller042d27a2014-04-30 23:26:02 +0200609
610config MAX_STACK_SIZE_MB
611 int "Maximum user stack size for 32-bit processes (MB)"
612 default 80
Helge Deller042d27a2014-04-30 23:26:02 +0200613 range 8 2048
614 depends on STACK_GROWSUP && (!64BIT || COMPAT)
615 help
616 This is the maximum stack size in Megabytes in the VM layout of 32-bit
617 user processes when the stack grows upwards (currently only on parisc
James Hogan5f171572017-10-24 16:52:32 +0100618 arch). The stack will be located at the highest memory address minus
619 the given value, unless the RLIMIT_STACK hard limit is changed to a
620 smaller value in which case that is used.
Helge Deller042d27a2014-04-30 23:26:02 +0200621
622 A sane initial value is 80 MB.
Mel Gorman3a80a7f2015-06-30 14:57:02 -0700623
Mel Gorman3a80a7f2015-06-30 14:57:02 -0700624config DEFERRED_STRUCT_PAGE_INIT
Vlastimil Babka1ce22102016-02-05 15:36:21 -0800625 bool "Defer initialisation of struct pages to kthreads"
Mike Rapoportd39f8fb2018-08-17 15:47:07 -0700626 depends on SPARSEMEM
Pavel Tatashinab1e8d82018-05-18 16:09:13 -0700627 depends on !NEED_PER_CPU_KM
Pasha Tatashin889c6952018-09-20 12:22:30 -0700628 depends on 64BIT
Mel Gorman3a80a7f2015-06-30 14:57:02 -0700629 help
630 Ordinarily all struct pages are initialised during early boot in a
631 single thread. On very large machines this can take a considerable
632 amount of time. If this option is set, large machines will bring up
633 a subset of memmap at boot and then initialise the rest in parallel
Vlastimil Babka1ce22102016-02-05 15:36:21 -0800634 by starting one-off "pgdatinitX" kernel thread for each node X. This
635 has a potential performance impact on processes running early in the
636 lifetime of the system until these kthreads finish the
637 initialisation.
Dan Williams033fbae2015-08-09 15:29:06 -0400638
Vladimir Davydov33c3fc72015-09-09 15:35:45 -0700639config IDLE_PAGE_TRACKING
640 bool "Enable idle page tracking"
641 depends on SYSFS && MMU
642 select PAGE_EXTENSION if !64BIT
643 help
644 This feature allows to estimate the amount of user pages that have
645 not been touched during a given period of time. This information can
646 be useful to tune memory cgroup limits and/or for job placement
647 within a compute cluster.
648
Mike Rapoport1ad13352018-04-18 11:07:49 +0300649 See Documentation/admin-guide/mm/idle_page_tracking.rst for
650 more details.
Vladimir Davydov33c3fc72015-09-09 15:35:45 -0700651
Oliver O'Halloran65f7d042017-06-28 11:32:31 +1000652# arch_add_memory() comprehends device memory
653config ARCH_HAS_ZONE_DEVICE
654 bool
655
Dan Williams033fbae2015-08-09 15:29:06 -0400656config ZONE_DEVICE
Jérôme Glisse5042db42017-09-08 16:11:43 -0700657 bool "Device memory (pmem, HMM, etc...) hotplug support"
Dan Williams033fbae2015-08-09 15:29:06 -0400658 depends on MEMORY_HOTPLUG
659 depends on MEMORY_HOTREMOVE
Dan Williams99490f12016-03-17 14:19:58 -0700660 depends on SPARSEMEM_VMEMMAP
Oliver O'Halloran65f7d042017-06-28 11:32:31 +1000661 depends on ARCH_HAS_ZONE_DEVICE
Matthew Wilcox3a08cd52018-09-22 16:14:30 -0400662 select XARRAY_MULTI
Dan Williams033fbae2015-08-09 15:29:06 -0400663
664 help
665 Device memory hotplug support allows for establishing pmem,
666 or other device driver discovered memory regions, in the
667 memmap. This allows pfn_to_page() lookups of otherwise
668 "device-physical" addresses which is needed for using a DAX
669 mapping in an O_DIRECT operation, among other things.
670
671 If FS_DAX is enabled, then say Y.
Linus Torvalds06a660a2015-09-11 16:42:39 -0700672
Jérôme Glisse2c8fc3d2019-05-13 17:22:37 -0700673config ARCH_HAS_HMM_MIRROR
674 bool
675 default y
676 depends on (X86_64 || PPC64)
677 depends on MMU && 64BIT
678
679config ARCH_HAS_HMM_DEVICE
680 bool
681 default y
682 depends on (X86_64 || PPC64)
683 depends on MEMORY_HOTPLUG
684 depends on MEMORY_HOTREMOVE
685 depends on SPARSEMEM_VMEMMAP
686 depends on ARCH_HAS_ZONE_DEVICE
687 select XARRAY_MULTI
688
Jérôme Glisse133ff0e2017-09-08 16:11:23 -0700689config ARCH_HAS_HMM
690 bool
691 default y
692 depends on (X86_64 || PPC64)
693 depends on ZONE_DEVICE
694 depends on MMU && 64BIT
695 depends on MEMORY_HOTPLUG
696 depends on MEMORY_HOTREMOVE
697 depends on SPARSEMEM_VMEMMAP
698
Jérôme Glisse6b368cd2017-09-08 16:12:32 -0700699config MIGRATE_VMA_HELPER
700 bool
701
Dan Williamse76384882018-05-16 11:46:08 -0700702config DEV_PAGEMAP_OPS
703 bool
704
Jérôme Glisse133ff0e2017-09-08 16:11:23 -0700705config HMM
706 bool
Jérôme Glisse734fb892019-05-13 17:19:45 -0700707 select MMU_NOTIFIER
Jérôme Glisse6b368cd2017-09-08 16:12:32 -0700708 select MIGRATE_VMA_HELPER
Jérôme Glisse133ff0e2017-09-08 16:11:23 -0700709
Jérôme Glissec0b12402017-09-08 16:11:27 -0700710config HMM_MIRROR
711 bool "HMM mirror CPU page table into a device page table"
712 depends on ARCH_HAS_HMM
Jérôme Glissec0b12402017-09-08 16:11:27 -0700713 select HMM
714 help
715 Select HMM_MIRROR if you want to mirror range of the CPU page table of a
716 process into a device page table. Here, mirror means "keep synchronized".
717 Prerequisites: the device must provide the ability to write-protect its
718 page tables (at PAGE_SIZE granularity), and must be able to recover from
719 the resulting potential page faults.
720
Jérôme Glisse5042db42017-09-08 16:11:43 -0700721config DEVICE_PRIVATE
722 bool "Unaddressable device memory (GPU memory, ...)"
723 depends on ARCH_HAS_HMM
Jérôme Glissedf6ad692017-09-08 16:12:24 -0700724 select HMM
Dan Williamse76384882018-05-16 11:46:08 -0700725 select DEV_PAGEMAP_OPS
Jérôme Glisse5042db42017-09-08 16:11:43 -0700726
727 help
728 Allows creation of struct pages to represent unaddressable device
729 memory; i.e., memory that is only accessible from the device (or
730 group of devices). You likely also want to select HMM_MIRROR.
731
Jérôme Glissedf6ad692017-09-08 16:12:24 -0700732config DEVICE_PUBLIC
733 bool "Addressable device memory (like GPU memory)"
734 depends on ARCH_HAS_HMM
735 select HMM
Dan Williamse76384882018-05-16 11:46:08 -0700736 select DEV_PAGEMAP_OPS
Jérôme Glissedf6ad692017-09-08 16:12:24 -0700737
738 help
739 Allows creation of struct pages to represent addressable device
740 memory; i.e., memory that is accessible from both the device and
741 the CPU
742
Jan Kara8025e5d2015-07-13 11:55:44 -0300743config FRAME_VECTOR
744 bool
Dave Hansen63c17fb2016-02-12 13:02:08 -0800745
746config ARCH_USES_HIGH_VMA_FLAGS
747 bool
Dave Hansen66d37572016-02-12 13:02:32 -0800748config ARCH_HAS_PKEYS
749 bool
Dennis Zhou30a5b532017-06-19 19:28:31 -0400750
751config PERCPU_STATS
752 bool "Collect percpu memory statistics"
Dennis Zhou30a5b532017-06-19 19:28:31 -0400753 help
754 This feature collects and exposes statistics via debugfs. The
755 information includes global and per chunk statistics, which can
756 be used to help understand percpu memory usage.
Kirill A. Shutemov64c349f2017-11-17 15:31:22 -0800757
758config GUP_BENCHMARK
759 bool "Enable infrastructure for get_user_pages_fast() benchmarking"
Kirill A. Shutemov64c349f2017-11-17 15:31:22 -0800760 help
761 Provides /sys/kernel/debug/gup_benchmark that helps with testing
762 performance of get_user_pages_fast().
763
764 See tools/testing/selftests/vm/gup_benchmark.c
Laurent Dufour3010a5e2018-06-07 17:06:08 -0700765
Christoph Hellwig39656e82019-07-11 20:56:49 -0700766config GUP_GET_PTE_LOW_HIGH
767 bool
768
Laurent Dufour3010a5e2018-06-07 17:06:08 -0700769config ARCH_HAS_PTE_SPECIAL
770 bool
Christoph Hellwig59e0b522018-07-31 13:39:35 +0200771
Christoph Hellwigcbd34da2019-07-11 20:57:28 -0700772#
773# Some architectures require a special hugepage directory format that is
774# required to support multiple hugepage sizes. For example a4fe3ce76
775# "powerpc/mm: Allow more flexible layouts for hugepage pagetables"
776# introduced it on powerpc. This allows for a more flexible hugepage
777# pagetable layouts.
778#
779config ARCH_HAS_HUGEPD
780 bool
781
Christoph Hellwig59e0b522018-07-31 13:39:35 +0200782endmenu