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 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".
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 ARCH_DISCARD_MEMBLOCK
143 config MEMORY_ISOLATION
147 boolean "Enable to assign a node which has only movable memory"
148 depends on HAVE_MEMBLOCK
149 depends on NO_BOOTMEM
154 Allow a node to have only movable memory. Pages used by the kernel,
155 such as direct mapping pages cannot be migrated. So the corresponding
156 memory device cannot be hotplugged. This option allows users to
157 online all the memory of a node as movable memory so that the whole
158 node can be hotplugged. Users who don't use the memory hotplug
159 feature are fine with this option on since they don't online memory
162 Say Y here if you want to hotplug a whole node.
163 Say N here if you want kernel to use memory on all nodes evenly.
166 # Only be set on architectures that have completely implemented memory hotplug
167 # feature. If you are not sure, don't touch it.
169 config HAVE_BOOTMEM_INFO_NODE
172 # eventually, we can have this option just 'select SPARSEMEM'
173 config MEMORY_HOTPLUG
174 bool "Allow for memory hot-add"
175 depends on SPARSEMEM || X86_64_ACPI_NUMA
176 depends on ARCH_ENABLE_MEMORY_HOTPLUG
177 depends on (IA64 || X86 || PPC_BOOK3S_64 || SUPERH || S390)
179 config MEMORY_HOTPLUG_SPARSE
181 depends on SPARSEMEM && MEMORY_HOTPLUG
183 config MEMORY_HOTREMOVE
184 bool "Allow for memory hot remove"
185 select MEMORY_ISOLATION
186 select HAVE_BOOTMEM_INFO_NODE if (X86_64 || PPC64)
187 depends on MEMORY_HOTPLUG && ARCH_ENABLE_MEMORY_HOTREMOVE
191 # If we have space for more page flags then we can enable additional
192 # optimizations and functionality.
194 # Regular Sparsemem takes page flag bits for the sectionid if it does not
195 # use a virtual memmap. Disable extended page flags for 32 bit platforms
196 # that require the use of a sectionid in the page flags.
198 config PAGEFLAGS_EXTENDED
200 depends on 64BIT || SPARSEMEM_VMEMMAP || !SPARSEMEM
202 # Heavily threaded applications may benefit from splitting the mm-wide
203 # page_table_lock, so that faults on different parts of the user address
204 # space can be handled with less contention: split it at this NR_CPUS.
205 # Default to 4 for wider testing, though 8 might be more appropriate.
206 # ARM's adjust_pte (unused if VIPT) depends on mm-wide page_table_lock.
207 # PA-RISC 7xxx's spinlock_t would enlarge struct page from 32 to 44 bytes.
208 # DEBUG_SPINLOCK and DEBUG_LOCK_ALLOC spinlock_t also enlarge struct page.
210 config SPLIT_PTLOCK_CPUS
212 default "999999" if ARM && !CPU_CACHE_VIPT
213 default "999999" if PARISC && !PA20
214 default "999999" if DEBUG_SPINLOCK || DEBUG_LOCK_ALLOC
218 # support for memory balloon compaction
219 config BALLOON_COMPACTION
220 bool "Allow for balloon memory compaction/migration"
222 depends on COMPACTION && VIRTIO_BALLOON
224 Memory fragmentation introduced by ballooning might reduce
225 significantly the number of 2MB contiguous memory blocks that can be
226 used within a guest, thus imposing performance penalties associated
227 with the reduced number of transparent huge pages that could be used
228 by the guest workload. Allowing the compaction & migration for memory
229 pages enlisted as being part of memory balloon devices avoids the
230 scenario aforementioned and helps improving memory defragmentation.
233 # support for memory compaction
235 bool "Allow for memory compaction"
240 Allows the compaction of memory for the allocation of huge pages.
243 # support for page migration
246 bool "Page migration"
248 depends on (NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION || CMA) && MMU
250 Allows the migration of the physical location of pages of processes
251 while the virtual addresses are not changed. This is useful in
252 two situations. The first is on NUMA systems to put pages nearer
253 to the processors accessing. The second is when allocating huge
254 pages as migration can relocate pages to satisfy a huge page
255 allocation instead of reclaiming.
257 config ARCH_ENABLE_HUGEPAGE_MIGRATION
260 config PHYS_ADDR_T_64BIT
261 def_bool 64BIT || ARCH_PHYS_ADDR_T_64BIT
265 default "0" if !ZONE_DMA
269 bool "Enable bounce buffers"
271 depends on BLOCK && MMU && (ZONE_DMA || HIGHMEM)
273 Enable bounce buffers for devices that cannot access
274 the full range of memory available to the CPU. Enabled
275 by default when ZONE_DMA or HIGHMEM is selected, but you
276 may say n to override this.
278 # On the 'tile' arch, USB OHCI needs the bounce pool since tilegx will often
279 # have more than 4GB of memory, but we don't currently use the IOTLB to present
280 # a 32-bit address to OHCI. So we need to use a bounce pool instead.
282 # We also use the bounce pool to provide stable page writes for jbd. jbd
283 # initiates buffer writeback without locking the page or setting PG_writeback,
284 # and fixing that behavior (a second time; jbd2 doesn't have this problem) is
285 # a major rework effort. Instead, use the bounce buffer to snapshot pages
286 # (until jbd goes away). The only jbd user is ext3.
287 config NEED_BOUNCE_POOL
289 default y if (TILE && USB_OHCI_HCD) || (BLK_DEV_INTEGRITY && JBD)
300 An architecture should select this if it implements the
301 deprecated interface virt_to_bus(). All new architectures
302 should probably not select this.
309 bool "Enable KSM for page merging"
312 Enable Kernel Samepage Merging: KSM periodically scans those areas
313 of an application's address space that an app has advised may be
314 mergeable. When it finds pages of identical content, it replaces
315 the many instances by a single page with that content, so
316 saving memory until one or another app needs to modify the content.
317 Recommended for use with KVM, or with other duplicative applications.
318 See Documentation/vm/ksm.txt for more information: KSM is inactive
319 until a program has madvised that an area is MADV_MERGEABLE, and
320 root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set).
322 config DEFAULT_MMAP_MIN_ADDR
323 int "Low address space to protect from user allocation"
327 This is the portion of low virtual memory which should be protected
328 from userspace allocation. Keeping a user from writing to low pages
329 can help reduce the impact of kernel NULL pointer bugs.
331 For most ia64, ppc64 and x86 users with lots of address space
332 a value of 65536 is reasonable and should cause no problems.
333 On arm and other archs it should not be higher than 32768.
334 Programs which use vm86 functionality or have some need to map
335 this low address space will need CAP_SYS_RAWIO or disable this
336 protection by setting the value to 0.
338 This value can be changed after boot using the
339 /proc/sys/vm/mmap_min_addr tunable.
341 config ARCH_SUPPORTS_MEMORY_FAILURE
344 config MEMORY_FAILURE
346 depends on ARCH_SUPPORTS_MEMORY_FAILURE
347 bool "Enable recovery from hardware memory errors"
348 select MEMORY_ISOLATION
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.
355 config HWPOISON_INJECT
356 tristate "HWPoison pages injector"
357 depends on MEMORY_FAILURE && DEBUG_KERNEL && PROC_FS
358 select PROC_PAGE_MONITOR
360 config NOMMU_INITIAL_TRIM_EXCESS
361 int "Turn on mmap() excess space trimming before booting"
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.
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.
375 If trimming is disabled, the excess is kept, but not used, which for
376 long-term mappings means that the space is wasted.
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.
383 This option specifies the initial value of this option. The default
384 of 1 says that all excess pages should be trimmed.
386 See Documentation/nommu-mmap.txt for more information.
388 config TRANSPARENT_HUGEPAGE
389 bool "Transparent Hugepage Support"
390 depends on HAVE_ARCH_TRANSPARENT_HUGEPAGE
393 Transparent Hugepages allows the kernel to use huge pages and
394 huge tlb transparently to the applications whenever possible.
395 This feature can improve computing performance to certain
396 applications by speeding up page faults during memory
397 allocation, by reducing the number of tlb misses and by speeding
398 up the pagetable walking.
400 If memory constrained on embedded, you may want to say N.
403 prompt "Transparent Hugepage Support sysfs defaults"
404 depends on TRANSPARENT_HUGEPAGE
405 default TRANSPARENT_HUGEPAGE_ALWAYS
407 Selects the sysfs defaults for Transparent Hugepage Support.
409 config TRANSPARENT_HUGEPAGE_ALWAYS
412 Enabling Transparent Hugepage always, can increase the
413 memory footprint of applications without a guaranteed
414 benefit but it will work automatically for all applications.
416 config TRANSPARENT_HUGEPAGE_MADVISE
419 Enabling Transparent Hugepage madvise, will only provide a
420 performance improvement benefit to the applications using
421 madvise(MADV_HUGEPAGE) but it won't risk to increase the
422 memory footprint of applications without a guaranteed
426 config CROSS_MEMORY_ATTACH
427 bool "Cross Memory Support"
431 Enabling this option adds the system calls process_vm_readv and
432 process_vm_writev which allow a process with the correct privileges
433 to directly read from or write to to another process's address space.
434 See the man page for more details.
437 # UP and nommu archs use km based percpu allocator
439 config NEED_PER_CPU_KM
445 bool "Enable cleancache driver to cache clean pages if tmem is present"
448 Cleancache can be thought of as a page-granularity victim cache
449 for clean pages that the kernel's pageframe replacement algorithm
450 (PFRA) would like to keep around, but can't since there isn't enough
451 memory. So when the PFRA "evicts" a page, it first attempts to use
452 cleancache code to put the data contained in that page into
453 "transcendent memory", memory that is not directly accessible or
454 addressable by the kernel and is of unknown and possibly
455 time-varying size. And when a cleancache-enabled
456 filesystem wishes to access a page in a file on disk, it first
457 checks cleancache to see if it already contains it; if it does,
458 the page is copied into the kernel and a disk access is avoided.
459 When a transcendent memory driver is available (such as zcache or
460 Xen transcendent memory), a significant I/O reduction
461 may be achieved. When none is available, all cleancache calls
462 are reduced to a single pointer-compare-against-NULL resulting
463 in a negligible performance hit.
465 If unsure, say Y to enable cleancache
468 bool "Enable frontswap to cache swap pages if tmem is present"
472 Frontswap is so named because it can be thought of as the opposite
473 of a "backing" store for a swap device. The data is stored into
474 "transcendent memory", memory that is not directly accessible or
475 addressable by the kernel and is of unknown and possibly
476 time-varying size. When space in transcendent memory is available,
477 a significant swap I/O reduction may be achieved. When none is
478 available, all frontswap calls are reduced to a single pointer-
479 compare-against-NULL resulting in a negligible performance hit
480 and swap data is stored as normal on the matching swap device.
482 If unsure, say Y to enable frontswap.
485 bool "Contiguous Memory Allocator"
486 depends on HAVE_MEMBLOCK && MMU
488 select MEMORY_ISOLATION
490 This enables the Contiguous Memory Allocator which allows other
491 subsystems to allocate big physically-contiguous blocks of memory.
492 CMA reserves a region of memory and allows only movable pages to
493 be allocated from it. This way, the kernel can use the memory for
494 pagecache and when a subsystem requests for contiguous area, the
495 allocated pages are migrated away to serve the contiguous request.
500 bool "CMA debug messages (DEVELOPMENT)"
501 depends on DEBUG_KERNEL && CMA
503 Turns on debug messages in CMA. This produces KERN_DEBUG
504 messages for every CMA call as well as various messages while
505 processing calls such as dma_alloc_from_contiguous().
506 This option does not affect warning and error messages.
512 A special purpose allocator for storing compressed pages.
513 It is designed to store up to two compressed pages per physical
514 page. While this design limits storage density, it has simple and
515 deterministic reclaim properties that make it preferable to a higher
516 density approach when reclaim will be used.
519 bool "Compressed cache for swap pages (EXPERIMENTAL)"
520 depends on FRONTSWAP && CRYPTO=y
525 A lightweight compressed cache for swap pages. It takes
526 pages that are in the process of being swapped out and attempts to
527 compress them into a dynamically allocated RAM-based memory pool.
528 This can result in a significant I/O reduction on swap device and,
529 in the case where decompressing from RAM is faster that swap device
530 reads, can also improve workload performance.
532 This is marked experimental because it is a new feature (as of
533 v3.11) that interacts heavily with memory reclaim. While these
534 interactions don't cause any known issues on simple memory setups,
535 they have not be fully explored on the large set of potential
536 configurations and workloads that exist.
538 config MEM_SOFT_DIRTY
539 bool "Track memory changes"
540 depends on CHECKPOINT_RESTORE && HAVE_ARCH_SOFT_DIRTY
541 select PROC_PAGE_MONITOR
543 This option enables memory changes tracking by introducing a
544 soft-dirty bit on pte-s. This bit it set when someone writes
545 into a page just as regular dirty bit, but unlike the latter
546 it can be cleared by hands.
548 See Documentation/vm/soft-dirty.txt for more details.