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_GUP
143 config ARCH_DISCARD_MEMBLOCK
149 config MEMORY_ISOLATION
153 # Only be set on architectures that have completely implemented memory hotplug
154 # feature. If you are not sure, don't touch it.
156 config HAVE_BOOTMEM_INFO_NODE
159 # eventually, we can have this option just 'select SPARSEMEM'
160 config MEMORY_HOTPLUG
161 bool "Allow for memory hot-add"
162 depends on SPARSEMEM || X86_64_ACPI_NUMA
163 depends on ARCH_ENABLE_MEMORY_HOTPLUG
165 config MEMORY_HOTPLUG_SPARSE
167 depends on SPARSEMEM && MEMORY_HOTPLUG
169 config MEMORY_HOTPLUG_DEFAULT_ONLINE
170 bool "Online the newly added memory blocks by default"
172 depends on MEMORY_HOTPLUG
174 This option sets the default policy setting for memory hotplug
175 onlining policy (/sys/devices/system/memory/auto_online_blocks) which
176 determines what happens to newly added memory regions. Policy setting
177 can always be changed at runtime.
178 See Documentation/memory-hotplug.txt for more information.
180 Say Y here if you want all hot-plugged memory blocks to appear in
181 'online' state by default.
182 Say N here if you want the default policy to keep all hot-plugged
183 memory blocks in 'offline' state.
185 config MEMORY_HOTREMOVE
186 bool "Allow for memory hot remove"
187 select MEMORY_ISOLATION
188 select HAVE_BOOTMEM_INFO_NODE if (X86_64 || PPC64)
189 depends on MEMORY_HOTPLUG && ARCH_ENABLE_MEMORY_HOTREMOVE
192 # Heavily threaded applications may benefit from splitting the mm-wide
193 # page_table_lock, so that faults on different parts of the user address
194 # space can be handled with less contention: split it at this NR_CPUS.
195 # Default to 4 for wider testing, though 8 might be more appropriate.
196 # ARM's adjust_pte (unused if VIPT) depends on mm-wide page_table_lock.
197 # PA-RISC 7xxx's spinlock_t would enlarge struct page from 32 to 44 bytes.
198 # DEBUG_SPINLOCK and DEBUG_LOCK_ALLOC spinlock_t also enlarge struct page.
200 config SPLIT_PTLOCK_CPUS
202 default "999999" if !MMU
203 default "999999" if ARM && !CPU_CACHE_VIPT
204 default "999999" if PARISC && !PA20
207 config ARCH_ENABLE_SPLIT_PMD_PTLOCK
211 # support for memory balloon
212 config MEMORY_BALLOON
216 # support for memory balloon compaction
217 config BALLOON_COMPACTION
218 bool "Allow for balloon memory compaction/migration"
220 depends on COMPACTION && MEMORY_BALLOON
222 Memory fragmentation introduced by ballooning might reduce
223 significantly the number of 2MB contiguous memory blocks that can be
224 used within a guest, thus imposing performance penalties associated
225 with the reduced number of transparent huge pages that could be used
226 by the guest workload. Allowing the compaction & migration for memory
227 pages enlisted as being part of memory balloon devices avoids the
228 scenario aforementioned and helps improving memory defragmentation.
231 # support for memory compaction
233 bool "Allow for memory compaction"
238 Compaction is the only memory management component to form
239 high order (larger physically contiguous) memory blocks
240 reliably. The page allocator relies on compaction heavily and
241 the lack of the feature can lead to unexpected OOM killer
242 invocations for high order memory requests. You shouldn't
243 disable this option unless there really is a strong reason for
244 it and then we would be really interested to hear about that at
248 # support for page migration
251 bool "Page migration"
253 depends on (NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION || CMA) && MMU
255 Allows the migration of the physical location of pages of processes
256 while the virtual addresses are not changed. This is useful in
257 two situations. The first is on NUMA systems to put pages nearer
258 to the processors accessing. The second is when allocating huge
259 pages as migration can relocate pages to satisfy a huge page
260 allocation instead of reclaiming.
262 config ARCH_ENABLE_HUGEPAGE_MIGRATION
265 config ARCH_ENABLE_THP_MIGRATION
268 config PHYS_ADDR_T_64BIT
269 def_bool 64BIT || ARCH_PHYS_ADDR_T_64BIT
272 bool "Enable bounce buffers"
274 depends on BLOCK && MMU && (ZONE_DMA || HIGHMEM)
276 Enable bounce buffers for devices that cannot access
277 the full range of memory available to the CPU. Enabled
278 by default when ZONE_DMA or HIGHMEM is selected, but you
279 may say n to override this.
281 # On the 'tile' arch, USB OHCI needs the bounce pool since tilegx will often
282 # have more than 4GB of memory, but we don't currently use the IOTLB to present
283 # a 32-bit address to OHCI. So we need to use a bounce pool instead.
284 config NEED_BOUNCE_POOL
286 default y if TILE && USB_OHCI_HCD
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.
306 bool "Enable KSM for page merging"
309 Enable Kernel Samepage Merging: KSM periodically scans those areas
310 of an application's address space that an app has advised may be
311 mergeable. When it finds pages of identical content, it replaces
312 the many instances by a single page with that content, so
313 saving memory until one or another app needs to modify the content.
314 Recommended for use with KVM, or with other duplicative applications.
315 See Documentation/vm/ksm.txt for more information: KSM is inactive
316 until a program has madvised that an area is MADV_MERGEABLE, and
317 root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set).
319 config DEFAULT_MMAP_MIN_ADDR
320 int "Low address space to protect from user allocation"
324 This is the portion of low virtual memory which should be protected
325 from userspace allocation. Keeping a user from writing to low pages
326 can help reduce the impact of kernel NULL pointer bugs.
328 For most ia64, ppc64 and x86 users with lots of address space
329 a value of 65536 is reasonable and should cause no problems.
330 On arm and other archs it should not be higher than 32768.
331 Programs which use vm86 functionality or have some need to map
332 this low address space will need CAP_SYS_RAWIO or disable this
333 protection by setting the value to 0.
335 This value can be changed after boot using the
336 /proc/sys/vm/mmap_min_addr tunable.
338 config ARCH_SUPPORTS_MEMORY_FAILURE
341 config MEMORY_FAILURE
343 depends on ARCH_SUPPORTS_MEMORY_FAILURE
344 bool "Enable recovery from hardware memory errors"
345 select MEMORY_ISOLATION
348 Enables code to recover from some memory failures on systems
349 with MCA recovery. This allows a system to continue running
350 even when some of its memory has uncorrected errors. This requires
351 special hardware support and typically ECC memory.
353 config HWPOISON_INJECT
354 tristate "HWPoison pages injector"
355 depends on MEMORY_FAILURE && DEBUG_KERNEL && PROC_FS
356 select PROC_PAGE_MONITOR
358 config NOMMU_INITIAL_TRIM_EXCESS
359 int "Turn on mmap() excess space trimming before booting"
363 The NOMMU mmap() frequently needs to allocate large contiguous chunks
364 of memory on which to store mappings, but it can only ask the system
365 allocator for chunks in 2^N*PAGE_SIZE amounts - which is frequently
366 more than it requires. To deal with this, mmap() is able to trim off
367 the excess and return it to the allocator.
369 If trimming is enabled, the excess is trimmed off and returned to the
370 system allocator, which can cause extra fragmentation, particularly
371 if there are a lot of transient processes.
373 If trimming is disabled, the excess is kept, but not used, which for
374 long-term mappings means that the space is wasted.
376 Trimming can be dynamically controlled through a sysctl option
377 (/proc/sys/vm/nr_trim_pages) which specifies the minimum number of
378 excess pages there must be before trimming should occur, or zero if
379 no trimming is to occur.
381 This option specifies the initial value of this option. The default
382 of 1 says that all excess pages should be trimmed.
384 See Documentation/nommu-mmap.txt for more information.
386 config TRANSPARENT_HUGEPAGE
387 bool "Transparent Hugepage Support"
388 depends on HAVE_ARCH_TRANSPARENT_HUGEPAGE
390 select RADIX_TREE_MULTIORDER
392 Transparent Hugepages allows the kernel to use huge pages and
393 huge tlb transparently to the applications whenever possible.
394 This feature can improve computing performance to certain
395 applications by speeding up page faults during memory
396 allocation, by reducing the number of tlb misses and by speeding
397 up the pagetable walking.
399 If memory constrained on embedded, you may want to say N.
402 prompt "Transparent Hugepage Support sysfs defaults"
403 depends on TRANSPARENT_HUGEPAGE
404 default TRANSPARENT_HUGEPAGE_ALWAYS
406 Selects the sysfs defaults for Transparent Hugepage Support.
408 config TRANSPARENT_HUGEPAGE_ALWAYS
411 Enabling Transparent Hugepage always, can increase the
412 memory footprint of applications without a guaranteed
413 benefit but it will work automatically for all applications.
415 config TRANSPARENT_HUGEPAGE_MADVISE
418 Enabling Transparent Hugepage madvise, will only provide a
419 performance improvement benefit to the applications using
420 madvise(MADV_HUGEPAGE) but it won't risk to increase the
421 memory footprint of applications without a guaranteed
425 config ARCH_WANTS_THP_SWAP
430 depends on TRANSPARENT_HUGEPAGE && ARCH_WANTS_THP_SWAP
432 Swap transparent huge pages in one piece, without splitting.
433 XXX: For now this only does clustered swap space allocation.
435 For selection by architectures with reasonable THP sizes.
437 config TRANSPARENT_HUGE_PAGECACHE
439 depends on TRANSPARENT_HUGEPAGE
442 # UP and nommu archs use km based percpu allocator
444 config NEED_PER_CPU_KM
450 bool "Enable cleancache driver to cache clean pages if tmem is present"
453 Cleancache can be thought of as a page-granularity victim cache
454 for clean pages that the kernel's pageframe replacement algorithm
455 (PFRA) would like to keep around, but can't since there isn't enough
456 memory. So when the PFRA "evicts" a page, it first attempts to use
457 cleancache code to put the data contained in that page into
458 "transcendent memory", memory that is not directly accessible or
459 addressable by the kernel and is of unknown and possibly
460 time-varying size. And when a cleancache-enabled
461 filesystem wishes to access a page in a file on disk, it first
462 checks cleancache to see if it already contains it; if it does,
463 the page is copied into the kernel and a disk access is avoided.
464 When a transcendent memory driver is available (such as zcache or
465 Xen transcendent memory), a significant I/O reduction
466 may be achieved. When none is available, all cleancache calls
467 are reduced to a single pointer-compare-against-NULL resulting
468 in a negligible performance hit.
470 If unsure, say Y to enable cleancache
473 bool "Enable frontswap to cache swap pages if tmem is present"
477 Frontswap is so named because it can be thought of as the opposite
478 of a "backing" store for a swap device. The data is stored into
479 "transcendent memory", memory that is not directly accessible or
480 addressable by the kernel and is of unknown and possibly
481 time-varying size. When space in transcendent memory is available,
482 a significant swap I/O reduction may be achieved. When none is
483 available, all frontswap calls are reduced to a single pointer-
484 compare-against-NULL resulting in a negligible performance hit
485 and swap data is stored as normal on the matching swap device.
487 If unsure, say Y to enable frontswap.
490 bool "Contiguous Memory Allocator"
491 depends on HAVE_MEMBLOCK && MMU
493 select MEMORY_ISOLATION
495 This enables the Contiguous Memory Allocator which allows other
496 subsystems to allocate big physically-contiguous blocks of memory.
497 CMA reserves a region of memory and allows only movable pages to
498 be allocated from it. This way, the kernel can use the memory for
499 pagecache and when a subsystem requests for contiguous area, the
500 allocated pages are migrated away to serve the contiguous request.
505 bool "CMA debug messages (DEVELOPMENT)"
506 depends on DEBUG_KERNEL && CMA
508 Turns on debug messages in CMA. This produces KERN_DEBUG
509 messages for every CMA call as well as various messages while
510 processing calls such as dma_alloc_from_contiguous().
511 This option does not affect warning and error messages.
514 bool "CMA debugfs interface"
515 depends on CMA && DEBUG_FS
517 Turns on the DebugFS interface for CMA.
520 int "Maximum count of the CMA areas"
524 CMA allows to create CMA areas for particular purpose, mainly,
525 used as device private area. This parameter sets the maximum
526 number of CMA area in the system.
528 If unsure, leave the default value "7".
530 config MEM_SOFT_DIRTY
531 bool "Track memory changes"
532 depends on CHECKPOINT_RESTORE && HAVE_ARCH_SOFT_DIRTY && PROC_FS
533 select PROC_PAGE_MONITOR
535 This option enables memory changes tracking by introducing a
536 soft-dirty bit on pte-s. This bit it set when someone writes
537 into a page just as regular dirty bit, but unlike the latter
538 it can be cleared by hands.
540 See Documentation/vm/soft-dirty.txt for more details.
543 bool "Compressed cache for swap pages (EXPERIMENTAL)"
544 depends on FRONTSWAP && CRYPTO=y
549 A lightweight compressed cache for swap pages. It takes
550 pages that are in the process of being swapped out and attempts to
551 compress them into a dynamically allocated RAM-based memory pool.
552 This can result in a significant I/O reduction on swap device and,
553 in the case where decompressing from RAM is faster that swap device
554 reads, can also improve workload performance.
556 This is marked experimental because it is a new feature (as of
557 v3.11) that interacts heavily with memory reclaim. While these
558 interactions don't cause any known issues on simple memory setups,
559 they have not be fully explored on the large set of potential
560 configurations and workloads that exist.
563 tristate "Common API for compressed memory storage"
566 Compressed memory storage API. This allows using either zbud or
570 tristate "Low (Up to 2x) density storage for compressed pages"
573 A special purpose allocator for storing compressed pages.
574 It is designed to store up to two compressed pages per physical
575 page. While this design limits storage density, it has simple and
576 deterministic reclaim properties that make it preferable to a higher
577 density approach when reclaim will be used.
580 tristate "Up to 3x density storage for compressed pages"
584 A special purpose allocator for storing compressed pages.
585 It is designed to store up to three compressed pages per physical
586 page. It is a ZBUD derivative so the simplicity and determinism are
590 tristate "Memory allocator for compressed pages"
594 zsmalloc is a slab-based memory allocator designed to store
595 compressed RAM pages. zsmalloc uses virtual memory mapping
596 in order to reduce fragmentation. However, this results in a
597 non-standard allocator interface where a handle, not a pointer, is
598 returned by an alloc(). This handle must be mapped in order to
599 access the allocated space.
601 config PGTABLE_MAPPING
602 bool "Use page table mapping to access object in zsmalloc"
605 By default, zsmalloc uses a copy-based object mapping method to
606 access allocations that span two pages. However, if a particular
607 architecture (ex, ARM) performs VM mapping faster than copying,
608 then you should select this. This causes zsmalloc to use page table
609 mapping rather than copying for object mapping.
611 You can check speed with zsmalloc benchmark:
612 https://github.com/spartacus06/zsmapbench
615 bool "Export zsmalloc statistics"
619 This option enables code in the zsmalloc to collect various
620 statistics about whats happening in zsmalloc and exports that
621 information to userspace via debugfs.
624 config GENERIC_EARLY_IOREMAP
627 config MAX_STACK_SIZE_MB
628 int "Maximum user stack size for 32-bit processes (MB)"
632 depends on STACK_GROWSUP && (!64BIT || COMPAT)
634 This is the maximum stack size in Megabytes in the VM layout of 32-bit
635 user processes when the stack grows upwards (currently only on parisc
636 and metag arch). The stack will be located at the highest memory
637 address minus the given value, unless the RLIMIT_STACK hard limit is
638 changed to a smaller value in which case that is used.
640 A sane initial value is 80 MB.
642 # For architectures that support deferred memory initialisation
643 config ARCH_SUPPORTS_DEFERRED_STRUCT_PAGE_INIT
646 config DEFERRED_STRUCT_PAGE_INIT
647 bool "Defer initialisation of struct pages to kthreads"
649 depends on ARCH_SUPPORTS_DEFERRED_STRUCT_PAGE_INIT
650 depends on NO_BOOTMEM && MEMORY_HOTPLUG
652 depends on !NEED_PER_CPU_KM
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.
675 # arch_add_memory() comprehends device memory
676 config ARCH_HAS_ZONE_DEVICE
680 bool "Device memory (pmem, HMM, etc...) hotplug support"
681 depends on MEMORY_HOTPLUG
682 depends on MEMORY_HOTREMOVE
683 depends on SPARSEMEM_VMEMMAP
684 depends on ARCH_HAS_ZONE_DEVICE
685 select RADIX_TREE_MULTIORDER
688 Device memory hotplug support allows for establishing pmem,
689 or other device driver discovered memory regions, in the
690 memmap. This allows pfn_to_page() lookups of otherwise
691 "device-physical" addresses which is needed for using a DAX
692 mapping in an O_DIRECT operation, among other things.
694 If FS_DAX is enabled, then say Y.
699 depends on (X86_64 || PPC64)
700 depends on ZONE_DEVICE
701 depends on MMU && 64BIT
702 depends on MEMORY_HOTPLUG
703 depends on MEMORY_HOTREMOVE
704 depends on SPARSEMEM_VMEMMAP
706 config MIGRATE_VMA_HELPER
711 select MIGRATE_VMA_HELPER
714 bool "HMM mirror CPU page table into a device page table"
715 depends on ARCH_HAS_HMM
719 Select HMM_MIRROR if you want to mirror range of the CPU page table of a
720 process into a device page table. Here, mirror means "keep synchronized".
721 Prerequisites: the device must provide the ability to write-protect its
722 page tables (at PAGE_SIZE granularity), and must be able to recover from
723 the resulting potential page faults.
725 config DEVICE_PRIVATE
726 bool "Unaddressable device memory (GPU memory, ...)"
727 depends on ARCH_HAS_HMM
731 Allows creation of struct pages to represent unaddressable device
732 memory; i.e., memory that is only accessible from the device (or
733 group of devices). You likely also want to select HMM_MIRROR.
736 bool "Addressable device memory (like GPU memory)"
737 depends on ARCH_HAS_HMM
741 Allows creation of struct pages to represent addressable device
742 memory; i.e., memory that is accessible from both the device and
748 config ARCH_USES_HIGH_VMA_FLAGS
750 config ARCH_HAS_PKEYS
754 bool "Collect percpu memory statistics"
757 This feature collects and exposes statistics via debugfs. The
758 information includes global and per chunk statistics, which can
759 be used to help understand percpu memory usage.