1 config SELECT_MEMORY_MODEL
3 depends on ARCH_SELECT_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
14 depends on !(ARCH_DISCONTIGMEM_ENABLE || ARCH_SPARSEMEM_ENABLE) || ARCH_FLATMEM_ENABLE
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
21 Some users of more advanced features like NUMA and
22 memory hotplug may have different options here.
23 DISCONTIGMEM is a 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".
29 If unsure, choose this option (Flat Memory) over any other.
31 config DISCONTIGMEM_MANUAL
32 bool "Discontiguous Memory"
33 depends on ARCH_DISCONTIGMEM_ENABLE
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
43 Many NUMA configurations will have this as the only option.
45 If unsure, choose "Flat Memory" over this option.
47 config SPARSEMEM_MANUAL
49 depends on ARCH_SPARSEMEM_ENABLE
51 This will be the only option for some systems, including
52 memory hotplug systems. This is normal.
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.
59 If unsure, choose "Discontiguous Memory" or "Flat Memory"
66 depends on (!SELECT_MEMORY_MODEL && ARCH_DISCONTIGMEM_ENABLE) || DISCONTIGMEM_MANUAL
70 depends on (!SELECT_MEMORY_MODEL && ARCH_SPARSEMEM_ENABLE) || SPARSEMEM_MANUAL
74 depends on (!DISCONTIGMEM && !SPARSEMEM) || FLATMEM_MANUAL
76 config FLAT_NODE_MEM_MAP
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.
85 config NEED_MULTIPLE_NODES
87 depends on DISCONTIGMEM || NUMA
89 config HAVE_MEMORY_PRESENT
91 depends on ARCH_HAVE_MEMORY_PRESENT || SPARSEMEM
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.
100 # This option will also potentially produce smaller runtime code
101 # with gcc 3.4 and later.
103 config SPARSEMEM_STATIC
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.
111 config SPARSEMEM_EXTREME
113 depends on SPARSEMEM && !SPARSEMEM_STATIC
115 config SPARSEMEM_VMEMMAP_ENABLE
118 config SPARSEMEM_ALLOC_MEM_MAP_TOGETHER
120 depends on SPARSEMEM && X86_64
122 config SPARSEMEM_VMEMMAP
123 bool "Sparse Memory virtual memmap"
124 depends on SPARSEMEM && SPARSEMEM_VMEMMAP_ENABLE
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.
134 config HAVE_MEMBLOCK_NODE_MAP
137 config HAVE_MEMBLOCK_PHYS_MAP
140 config HAVE_GENERIC_RCU_GUP
143 config ARCH_DISCARD_MEMBLOCK
149 config MEMORY_ISOLATION
153 bool "Enable to assign a node which has only movable memory"
154 depends on HAVE_MEMBLOCK
155 depends on NO_BOOTMEM
160 Allow a node to have only movable memory. Pages used by the kernel,
161 such as direct mapping pages cannot be migrated. So the corresponding
162 memory device cannot be hotplugged. This option allows the following
164 - When the system is booting, node full of hotpluggable memory can
165 be arranged to have only movable memory so that the whole node can
166 be hot-removed. (need movable_node boot option specified).
167 - After the system is up, the option allows users to online all the
168 memory of a node as movable memory so that the whole node can be
171 Users who don't use the memory hotplug feature are fine with this
172 option on since they don't specify movable_node boot option or they
173 don't online memory as movable.
175 Say Y here if you want to hotplug a whole node.
176 Say N here if you want kernel to use memory on all nodes evenly.
179 # Only be set on architectures that have completely implemented memory hotplug
180 # feature. If you are not sure, don't touch it.
182 config HAVE_BOOTMEM_INFO_NODE
185 # eventually, we can have this option just 'select SPARSEMEM'
186 config MEMORY_HOTPLUG
187 bool "Allow for memory hot-add"
188 depends on SPARSEMEM || X86_64_ACPI_NUMA
189 depends on ARCH_ENABLE_MEMORY_HOTPLUG
191 config MEMORY_HOTPLUG_SPARSE
193 depends on SPARSEMEM && MEMORY_HOTPLUG
195 config MEMORY_HOTPLUG_DEFAULT_ONLINE
196 bool "Online the newly added memory blocks by default"
198 depends on MEMORY_HOTPLUG
200 This option sets the default policy setting for memory hotplug
201 onlining policy (/sys/devices/system/memory/auto_online_blocks) which
202 determines what happens to newly added memory regions. Policy setting
203 can always be changed at runtime.
204 See Documentation/memory-hotplug.txt for more information.
206 Say Y here if you want all hot-plugged memory blocks to appear in
207 'online' state by default.
208 Say N here if you want the default policy to keep all hot-plugged
209 memory blocks in 'offline' state.
211 config MEMORY_HOTREMOVE
212 bool "Allow for memory hot remove"
213 select MEMORY_ISOLATION
214 select HAVE_BOOTMEM_INFO_NODE if (X86_64 || PPC64)
215 depends on MEMORY_HOTPLUG && ARCH_ENABLE_MEMORY_HOTREMOVE
218 # Heavily threaded applications may benefit from splitting the mm-wide
219 # page_table_lock, so that faults on different parts of the user address
220 # space can be handled with less contention: split it at this NR_CPUS.
221 # Default to 4 for wider testing, though 8 might be more appropriate.
222 # ARM's adjust_pte (unused if VIPT) depends on mm-wide page_table_lock.
223 # PA-RISC 7xxx's spinlock_t would enlarge struct page from 32 to 44 bytes.
224 # DEBUG_SPINLOCK and DEBUG_LOCK_ALLOC spinlock_t also enlarge struct page.
226 config SPLIT_PTLOCK_CPUS
228 default "999999" if !MMU
229 default "999999" if ARM && !CPU_CACHE_VIPT
230 default "999999" if PARISC && !PA20
233 config ARCH_ENABLE_SPLIT_PMD_PTLOCK
237 # support for memory balloon
238 config MEMORY_BALLOON
242 # support for memory balloon compaction
243 config BALLOON_COMPACTION
244 bool "Allow for balloon memory compaction/migration"
246 depends on COMPACTION && MEMORY_BALLOON
248 Memory fragmentation introduced by ballooning might reduce
249 significantly the number of 2MB contiguous memory blocks that can be
250 used within a guest, thus imposing performance penalties associated
251 with the reduced number of transparent huge pages that could be used
252 by the guest workload. Allowing the compaction & migration for memory
253 pages enlisted as being part of memory balloon devices avoids the
254 scenario aforementioned and helps improving memory defragmentation.
257 # support for memory compaction
259 bool "Allow for memory compaction"
264 Allows the compaction of memory for the allocation of huge pages.
267 # support for page migration
270 bool "Page migration"
272 depends on (NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION || CMA) && MMU
274 Allows the migration of the physical location of pages of processes
275 while the virtual addresses are not changed. This is useful in
276 two situations. The first is on NUMA systems to put pages nearer
277 to the processors accessing. The second is when allocating huge
278 pages as migration can relocate pages to satisfy a huge page
279 allocation instead of reclaiming.
281 config ARCH_ENABLE_HUGEPAGE_MIGRATION
284 config PHYS_ADDR_T_64BIT
285 def_bool 64BIT || ARCH_PHYS_ADDR_T_64BIT
288 bool "Enable bounce buffers"
290 depends on BLOCK && MMU && (ZONE_DMA || HIGHMEM)
292 Enable bounce buffers for devices that cannot access
293 the full range of memory available to the CPU. Enabled
294 by default when ZONE_DMA or HIGHMEM is selected, but you
295 may say n to override this.
297 # On the 'tile' arch, USB OHCI needs the bounce pool since tilegx will often
298 # have more than 4GB of memory, but we don't currently use the IOTLB to present
299 # a 32-bit address to OHCI. So we need to use a bounce pool instead.
300 config NEED_BOUNCE_POOL
302 default y if TILE && USB_OHCI_HCD
313 An architecture should select this if it implements the
314 deprecated interface virt_to_bus(). All new architectures
315 should probably not select this.
323 bool "Enable KSM for page merging"
326 Enable Kernel Samepage Merging: KSM periodically scans those areas
327 of an application's address space that an app has advised may be
328 mergeable. When it finds pages of identical content, it replaces
329 the many instances by a single page with that content, so
330 saving memory until one or another app needs to modify the content.
331 Recommended for use with KVM, or with other duplicative applications.
332 See Documentation/vm/ksm.txt for more information: KSM is inactive
333 until a program has madvised that an area is MADV_MERGEABLE, and
334 root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set).
336 config DEFAULT_MMAP_MIN_ADDR
337 int "Low address space to protect from user allocation"
341 This is the portion of low virtual memory which should be protected
342 from userspace allocation. Keeping a user from writing to low pages
343 can help reduce the impact of kernel NULL pointer bugs.
345 For most ia64, ppc64 and x86 users with lots of address space
346 a value of 65536 is reasonable and should cause no problems.
347 On arm and other archs it should not be higher than 32768.
348 Programs which use vm86 functionality or have some need to map
349 this low address space will need CAP_SYS_RAWIO or disable this
350 protection by setting the value to 0.
352 This value can be changed after boot using the
353 /proc/sys/vm/mmap_min_addr tunable.
355 config ARCH_SUPPORTS_MEMORY_FAILURE
358 config MEMORY_FAILURE
360 depends on ARCH_SUPPORTS_MEMORY_FAILURE
361 bool "Enable recovery from hardware memory errors"
362 select MEMORY_ISOLATION
365 Enables code to recover from some memory failures on systems
366 with MCA recovery. This allows a system to continue running
367 even when some of its memory has uncorrected errors. This requires
368 special hardware support and typically ECC memory.
370 config HWPOISON_INJECT
371 tristate "HWPoison pages injector"
372 depends on MEMORY_FAILURE && DEBUG_KERNEL && PROC_FS
373 select PROC_PAGE_MONITOR
375 config NOMMU_INITIAL_TRIM_EXCESS
376 int "Turn on mmap() excess space trimming before booting"
380 The NOMMU mmap() frequently needs to allocate large contiguous chunks
381 of memory on which to store mappings, but it can only ask the system
382 allocator for chunks in 2^N*PAGE_SIZE amounts - which is frequently
383 more than it requires. To deal with this, mmap() is able to trim off
384 the excess and return it to the allocator.
386 If trimming is enabled, the excess is trimmed off and returned to the
387 system allocator, which can cause extra fragmentation, particularly
388 if there are a lot of transient processes.
390 If trimming is disabled, the excess is kept, but not used, which for
391 long-term mappings means that the space is wasted.
393 Trimming can be dynamically controlled through a sysctl option
394 (/proc/sys/vm/nr_trim_pages) which specifies the minimum number of
395 excess pages there must be before trimming should occur, or zero if
396 no trimming is to occur.
398 This option specifies the initial value of this option. The default
399 of 1 says that all excess pages should be trimmed.
401 See Documentation/nommu-mmap.txt for more information.
403 config TRANSPARENT_HUGEPAGE
404 bool "Transparent Hugepage Support"
405 depends on HAVE_ARCH_TRANSPARENT_HUGEPAGE
407 select RADIX_TREE_MULTIORDER
409 Transparent Hugepages allows the kernel to use huge pages and
410 huge tlb transparently to the applications whenever possible.
411 This feature can improve computing performance to certain
412 applications by speeding up page faults during memory
413 allocation, by reducing the number of tlb misses and by speeding
414 up the pagetable walking.
416 If memory constrained on embedded, you may want to say N.
419 prompt "Transparent Hugepage Support sysfs defaults"
420 depends on TRANSPARENT_HUGEPAGE
421 default TRANSPARENT_HUGEPAGE_ALWAYS
423 Selects the sysfs defaults for Transparent Hugepage Support.
425 config TRANSPARENT_HUGEPAGE_ALWAYS
428 Enabling Transparent Hugepage always, can increase the
429 memory footprint of applications without a guaranteed
430 benefit but it will work automatically for all applications.
432 config TRANSPARENT_HUGEPAGE_MADVISE
435 Enabling Transparent Hugepage madvise, will only provide a
436 performance improvement benefit to the applications using
437 madvise(MADV_HUGEPAGE) but it won't risk to increase the
438 memory footprint of applications without a guaranteed
443 # UP and nommu archs use km based percpu allocator
445 config NEED_PER_CPU_KM
451 bool "Enable cleancache driver to cache clean pages if tmem is present"
454 Cleancache can be thought of as a page-granularity victim cache
455 for clean pages that the kernel's pageframe replacement algorithm
456 (PFRA) would like to keep around, but can't since there isn't enough
457 memory. So when the PFRA "evicts" a page, it first attempts to use
458 cleancache code to put the data contained in that page into
459 "transcendent memory", memory that is not directly accessible or
460 addressable by the kernel and is of unknown and possibly
461 time-varying size. And when a cleancache-enabled
462 filesystem wishes to access a page in a file on disk, it first
463 checks cleancache to see if it already contains it; if it does,
464 the page is copied into the kernel and a disk access is avoided.
465 When a transcendent memory driver is available (such as zcache or
466 Xen transcendent memory), a significant I/O reduction
467 may be achieved. When none is available, all cleancache calls
468 are reduced to a single pointer-compare-against-NULL resulting
469 in a negligible performance hit.
471 If unsure, say Y to enable cleancache
474 bool "Enable frontswap to cache swap pages if tmem is present"
478 Frontswap is so named because it can be thought of as the opposite
479 of a "backing" store for a swap device. The data is stored into
480 "transcendent memory", memory that is not directly accessible or
481 addressable by the kernel and is of unknown and possibly
482 time-varying size. When space in transcendent memory is available,
483 a significant swap I/O reduction may be achieved. When none is
484 available, all frontswap calls are reduced to a single pointer-
485 compare-against-NULL resulting in a negligible performance hit
486 and swap data is stored as normal on the matching swap device.
488 If unsure, say Y to enable frontswap.
491 bool "Contiguous Memory Allocator"
492 depends on HAVE_MEMBLOCK && MMU
494 select MEMORY_ISOLATION
496 This enables the Contiguous Memory Allocator which allows other
497 subsystems to allocate big physically-contiguous blocks of memory.
498 CMA reserves a region of memory and allows only movable pages to
499 be allocated from it. This way, the kernel can use the memory for
500 pagecache and when a subsystem requests for contiguous area, the
501 allocated pages are migrated away to serve the contiguous request.
506 bool "CMA debug messages (DEVELOPMENT)"
507 depends on DEBUG_KERNEL && CMA
509 Turns on debug messages in CMA. This produces KERN_DEBUG
510 messages for every CMA call as well as various messages while
511 processing calls such as dma_alloc_from_contiguous().
512 This option does not affect warning and error messages.
515 bool "CMA debugfs interface"
516 depends on CMA && DEBUG_FS
518 Turns on the DebugFS interface for CMA.
521 int "Maximum count of the CMA areas"
525 CMA allows to create CMA areas for particular purpose, mainly,
526 used as device private area. This parameter sets the maximum
527 number of CMA area in the system.
529 If unsure, leave the default value "7".
531 config MEM_SOFT_DIRTY
532 bool "Track memory changes"
533 depends on CHECKPOINT_RESTORE && HAVE_ARCH_SOFT_DIRTY && PROC_FS
534 select PROC_PAGE_MONITOR
536 This option enables memory changes tracking by introducing a
537 soft-dirty bit on pte-s. This bit it set when someone writes
538 into a page just as regular dirty bit, but unlike the latter
539 it can be cleared by hands.
541 See Documentation/vm/soft-dirty.txt for more details.
544 bool "Compressed cache for swap pages (EXPERIMENTAL)"
545 depends on FRONTSWAP && CRYPTO=y
550 A lightweight compressed cache for swap pages. It takes
551 pages that are in the process of being swapped out and attempts to
552 compress them into a dynamically allocated RAM-based memory pool.
553 This can result in a significant I/O reduction on swap device and,
554 in the case where decompressing from RAM is faster that swap device
555 reads, can also improve workload performance.
557 This is marked experimental because it is a new feature (as of
558 v3.11) that interacts heavily with memory reclaim. While these
559 interactions don't cause any known issues on simple memory setups,
560 they have not be fully explored on the large set of potential
561 configurations and workloads that exist.
564 tristate "Common API for compressed memory storage"
567 Compressed memory storage API. This allows using either zbud or
571 tristate "Low (Up to 2x) density storage for compressed pages"
574 A special purpose allocator for storing compressed pages.
575 It is designed to store up to two compressed pages per physical
576 page. While this design limits storage density, it has simple and
577 deterministic reclaim properties that make it preferable to a higher
578 density approach when reclaim will be used.
581 tristate "Up to 3x density storage for compressed pages"
585 A special purpose allocator for storing compressed pages.
586 It is designed to store up to three compressed pages per physical
587 page. It is a ZBUD derivative so the simplicity and determinism are
591 tristate "Memory allocator for compressed pages"
595 zsmalloc is a slab-based memory allocator designed to store
596 compressed RAM pages. zsmalloc uses virtual memory mapping
597 in order to reduce fragmentation. However, this results in a
598 non-standard allocator interface where a handle, not a pointer, is
599 returned by an alloc(). This handle must be mapped in order to
600 access the allocated space.
602 config PGTABLE_MAPPING
603 bool "Use page table mapping to access object in zsmalloc"
606 By default, zsmalloc uses a copy-based object mapping method to
607 access allocations that span two pages. However, if a particular
608 architecture (ex, ARM) performs VM mapping faster than copying,
609 then you should select this. This causes zsmalloc to use page table
610 mapping rather than copying for object mapping.
612 You can check speed with zsmalloc benchmark:
613 https://github.com/spartacus06/zsmapbench
616 bool "Export zsmalloc statistics"
620 This option enables code in the zsmalloc to collect various
621 statistics about whats happening in zsmalloc and exports that
622 information to userspace via debugfs.
625 config GENERIC_EARLY_IOREMAP
628 config MAX_STACK_SIZE_MB
629 int "Maximum user stack size for 32-bit processes (MB)"
633 depends on STACK_GROWSUP && (!64BIT || COMPAT)
635 This is the maximum stack size in Megabytes in the VM layout of 32-bit
636 user processes when the stack grows upwards (currently only on parisc
637 and metag arch). The stack will be located at the highest memory
638 address minus the given value, unless the RLIMIT_STACK hard limit is
639 changed to a smaller value in which case that is used.
641 A sane initial value is 80 MB.
643 # For architectures that support deferred memory initialisation
644 config ARCH_SUPPORTS_DEFERRED_STRUCT_PAGE_INIT
647 config DEFERRED_STRUCT_PAGE_INIT
648 bool "Defer initialisation of struct pages to kthreads"
650 depends on ARCH_SUPPORTS_DEFERRED_STRUCT_PAGE_INIT
651 depends on MEMORY_HOTPLUG
654 Ordinarily all struct pages are initialised during early boot in a
655 single thread. On very large machines this can take a considerable
656 amount of time. If this option is set, large machines will bring up
657 a subset of memmap at boot and then initialise the rest in parallel
658 by starting one-off "pgdatinitX" kernel thread for each node X. This
659 has a potential performance impact on processes running early in the
660 lifetime of the system until these kthreads finish the
663 config IDLE_PAGE_TRACKING
664 bool "Enable idle page tracking"
665 depends on SYSFS && MMU
666 select PAGE_EXTENSION if !64BIT
668 This feature allows to estimate the amount of user pages that have
669 not been touched during a given period of time. This information can
670 be useful to tune memory cgroup limits and/or for job placement
671 within a compute cluster.
673 See Documentation/vm/idle_page_tracking.txt for more details.
676 bool "Device memory (pmem, etc...) hotplug support" if EXPERT
677 depends on MEMORY_HOTPLUG
678 depends on MEMORY_HOTREMOVE
679 depends on SPARSEMEM_VMEMMAP
680 depends on X86_64 #arch_add_memory() comprehends device memory
683 Device memory hotplug support allows for establishing pmem,
684 or other device driver discovered memory regions, in the
685 memmap. This allows pfn_to_page() lookups of otherwise
686 "device-physical" addresses which is needed for using a DAX
687 mapping in an O_DIRECT operation, among other things.
689 If FS_DAX is enabled, then say Y.
694 config ARCH_USES_HIGH_VMA_FLAGS
696 config ARCH_HAS_PKEYS