Lines Matching +full:memory +full:- +full:to +full:- +full:memory
1 # SPDX-License-Identifier: GPL-2.0-only
3 menu "Memory Management options"
10 prompt "Memory model"
16 This option allows you to change some of the ways that
17 Linux manages its memory internally. Most users will
22 bool "Flat Memory"
25 This option is best suited for non-NUMA systems with
31 spaces and for features like NUMA and memory hotplug,
32 choose "Sparse Memory".
34 If unsure, choose this option (Flat Memory) over any other.
37 bool "Discontiguous Memory"
41 memory systems, over FLATMEM. These systems have holes
45 Although "Discontiguous Memory" is still used by several
47 "Sparse Memory".
49 If unsure, choose "Sparse Memory" over this option.
52 bool "Sparse Memory"
56 memory hot-plug systems. This is normal.
59 holes is their physical address space and allows memory
60 hot-plug and hot-remove.
62 If unsure, choose "Flat Memory" over this option.
71 MEMORY_MONITOR is a monitor of some memory reclaim method.
80 File-LRU is a mechanism that put file page in global lru list,
88 Add more attributes in memory cgroup, these attribute is used
89 to show information, shrink memory, swapin page and so on.
92 bool "Enable zswapd thread to reclaim anon pages in background"
99 zram will exchanged to eswap by a certain percentage.
108 bool "Memory reclaim delay accounting"
111 Memory reclaim delay accounting. Never use it as a kernel module.
131 # to represent different areas of memory. This variable allows
132 # those dependencies to exist individually.
164 bool "Sparse Memory virtual memmap"
168 SPARSEMEM_VMEMMAP uses a virtually mapped memmap to optimise
182 # Don't discard allocated memory used to track "memory" and "reserved" memblocks
183 # after early boot, so it can still be used to test for validity of memory.
184 # Also, memblocks are updated with memory hot(un)plug.
188 # Keep arch NUMA mapping infrastructure post-init.
196 # Only be set on architectures that have completely implemented memory hotplug
204 bool "Allow for memory hot-add"
216 bool "Online the newly added memory blocks by default"
219 This option sets the default policy setting for memory hotplug
220 onlining policy (/sys/devices/system/memory/auto_online_blocks) which
221 determines what happens to newly added memory regions. Policy setting
223 See Documentation/admin-guide/mm/memory-hotplug.rst for more information.
225 Say Y here if you want all hot-plugged memory blocks to appear in
227 Say N here if you want the default policy to keep all hot-plugged
228 memory blocks in 'offline' state.
231 bool "Allow for memory hot remove"
236 # Heavily threaded applications may benefit from splitting the mm-wide
239 # Default to 4 for wider testing, though 8 might be more appropriate.
240 # ARM's adjust_pte (unused if VIPT) depends on mm-wide page_table_lock.
241 # PA-RISC 7xxx's spinlock_t would enlarge struct page from 32 to 44 bytes.
243 # a per-page lock leads to problems when multiple tables need to be locked
259 # support for memory balloon
264 # support for memory balloon compaction
266 bool "Allow for balloon memory compaction/migration"
270 Memory fragmentation introduced by ballooning might reduce
271 significantly the number of 2MB contiguous memory blocks that can be
274 by the guest workload. Allowing the compaction & migration for memory
275 pages enlisted as being part of memory balloon devices avoids the
276 scenario aforementioned and helps improving memory defragmentation.
279 # support for memory compaction
281 bool "Allow for memory compaction"
286 Compaction is the only memory management component to form
287 high order (larger physically contiguous) memory blocks
289 the lack of the feature can lead to unexpected OOM killer
290 invocations for high order memory requests. You shouldn't
292 it and then we would be really interested to hear about that at
293 linux-mm@kvack.org.
303 those pages to another entity, such as a hypervisor, so that the
304 memory can be freed within the host for other uses.
316 two situations. The first is on NUMA systems to put pages nearer
317 to the processors accessing. The second is when allocating huge
318 pages as migration can relocate pages to satisfy a huge page
339 the full range of memory available to the CPU. Enabled
341 may say n to override this.
365 saving memory until one or another app needs to modify the content.
369 root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set).
372 int "Low address space to protect from user allocation"
376 This is the portion of low virtual memory which should be protected
377 from userspace allocation. Keeping a user from writing to low pages
383 Programs which use vm86 functionality or have some need to map
385 protection by setting the value to 0.
396 bool "Enable recovery from hardware memory errors"
400 Enables code to recover from some memory failures on systems
401 with MCA recovery. This allows a system to continue running
402 even when some of its memory has uncorrected errors. This requires
403 special hardware support and typically ECC memory.
415 The NOMMU mmap() frequently needs to allocate large contiguous chunks
416 of memory on which to store mappings, but it can only ask the system
417 allocator for chunks in 2^N*PAGE_SIZE amounts - which is frequently
418 more than it requires. To deal with this, mmap() is able to trim off
419 the excess and return it to the allocator.
421 If trimming is enabled, the excess is trimmed off and returned to the
426 long-term mappings means that the space is wasted.
431 no trimming is to occur.
436 See Documentation/admin-guide/mm/nommu-mmap.rst for more information.
444 Transparent Hugepages allows the kernel to use huge pages and
445 huge tlb transparently to the applications whenever possible.
446 This feature can improve computing performance to certain
447 applications by speeding up page faults during memory
451 If memory constrained on embedded, you may want to say N.
464 memory footprint of applications without a guaranteed
471 performance improvement benefit to the applications using
472 madvise(MADV_HUGEPAGE) but it won't risk to increase the
473 memory footprint of applications without a guaranteed
499 bool "Enable cleancache driver to cache clean pages if tmem is present"
501 Cleancache can be thought of as a page-granularity victim cache
503 (PFRA) would like to keep around, but can't since there isn't enough
504 memory. So when the PFRA "evicts" a page, it first attempts to use
505 cleancache code to put the data contained in that page into
506 "transcendent memory", memory that is not directly accessible or
508 time-varying size. And when a cleancache-enabled
509 filesystem wishes to access a page in a file on disk, it first
510 checks cleancache to see if it already contains it; if it does,
512 When a transcendent memory driver is available (such as zcache or
513 Xen transcendent memory), a significant I/O reduction
515 are reduced to a single pointer-compare-against-NULL resulting
518 If unsure, say Y to enable cleancache
521 bool "Enable frontswap to cache swap pages if tmem is present"
526 "transcendent memory", memory that is not directly accessible or
528 time-varying size. When space in transcendent memory is available,
530 available, all frontswap calls are reduced to a single pointer-
531 compare-against-NULL resulting in a negligible performance hit
534 If unsure, say Y to enable frontswap.
537 bool "Contiguous Memory Allocator"
542 This enables the Contiguous Memory Allocator which allows other
543 subsystems to allocate big physically-contiguous blocks of memory.
544 CMA reserves a region of memory and allows only movable pages to
545 be allocated from it. This way, the kernel can use the memory for
547 allocated pages are migrated away to serve the contiguous request.
572 CMA allows to create CMA areas for particular purpose, mainly,
582 If enabled, it will add MIGRATE_CMA to pcp lists and movable
583 allocations with __GFP_CMA flag will use cma areas prior to
589 bool "Track memory changes"
593 This option enables memory changes tracking by introducing a
594 soft-dirty bit on pte-s. This bit it set when someone writes
598 See Documentation/admin-guide/mm/soft-dirty.rst for more details.
606 pages that are in the process of being swapped out and attempts to
607 compress them into a dynamically allocated RAM-based memory pool.
613 v3.11) that interacts heavily with memory reclaim. While these
614 interactions don't cause any known issues on simple memory setups,
627 a particular compression algorithm please refer to the benchmarks
736 tristate "Common API for compressed memory storage"
738 Compressed memory storage API. This allows using either zbud or
742 tristate "Low (Up to 2x) density storage for compressed pages"
745 It is designed to store up to two compressed pages per physical
747 deterministic reclaim properties that make it preferable to a higher
751 tristate "Up to 3x density storage for compressed pages"
755 It is designed to store up to three compressed pages per physical
760 tristate "Memory allocator for compressed pages"
763 zsmalloc is a slab-based memory allocator designed to store
764 compressed RAM pages. zsmalloc uses virtual memory mapping
765 in order to reduce fragmentation. However, this results in a
766 non-standard allocator interface where a handle, not a pointer, is
767 returned by an alloc(). This handle must be mapped in order to
775 This option enables code in the zsmalloc to collect various
777 information to userspace via debugfs.
784 int "Maximum user stack size for 32-bit processes (MB)"
789 This is the maximum stack size in Megabytes in the VM layout of 32-bit
791 arch). The stack will be located at the highest memory address minus
792 the given value, unless the RLIMIT_STACK hard limit is changed to a
798 bool "Defer initialisation of struct pages to kthreads"
817 This feature allows to estimate the amount of user pages that have
819 be useful to tune memory cgroup limits and/or for job placement
822 See Documentation/admin-guide/mm/idle_page_tracking.rst for
832 bool "Device memory (pmem, HMM, etc...) hotplug support"
840 Device memory hotplug support allows for establishing pmem,
841 or other device driver discovered memory regions, in the
843 "device-physical" addresses which is needed for using a DAX
852 # Helpers to mirror range of the CPU page tables of a process into device page
860 bool "Unaddressable device memory (GPU memory, ...)"
865 Allows creation of struct pages to represent unaddressable device
866 memory; i.e., memory that is only accessible from the device (or
867 group of devices). You likely also want to select HMM_MIRROR.
881 bool "Collect percpu memory statistics"
885 be used to help understand percpu memory usage.
899 bool "Read-only THP for filesystems (EXPERIMENTAL)"
903 Allow khugepaged to put read-only file-backed pages in THP.
914 # required to support multiple hugepage sizes. For example a4fe3ce76
930 Allow naming anonymous virtual memory areas.
932 This feature allows assigning names to virtual memory areas. Assigned
934 and help identifying individual anonymous memory areas.
935 Assigning a name to anonymous virtual memory area might prevent that
936 area from being merged with adjacent virtual memory areas due to the
939 # For lmkd to trigger in-kernel lowmem info
942 bool "Low Memory Killer"
948 bool "Low Memory Killer Debug"
963 # Use rss_threshold to monitoring RSS
966 bool "Enable /proc/<pid>/rss and /proc/<pid>/rss_threshold to monitoring RSS"
970 Set a threshold to monitoring RSS in per pid
973 bool "Purgeable memory feature"
981 bool "Purgeable memory debug"
985 Debug info for purgeable memory
988 bool "Purgeable memory feature for ashmem"