1ec8f24b7SThomas Gleixner# SPDX-License-Identifier: GPL-2.0-only 259e0b520SChristoph Hellwig 359e0b520SChristoph Hellwigmenu "Memory Management options" 459e0b520SChristoph Hellwig 5e1785e85SDave Hansenconfig SELECT_MEMORY_MODEL 6e1785e85SDave Hansen def_bool y 7a8826eebSKees Cook depends on ARCH_SELECT_MEMORY_MODEL 8e1785e85SDave Hansen 93a9da765SDave Hansenchoice 103a9da765SDave Hansen prompt "Memory model" 11e1785e85SDave Hansen depends on SELECT_MEMORY_MODEL 12d41dee36SAndy Whitcroft default SPARSEMEM_MANUAL if ARCH_SPARSEMEM_DEFAULT 13e1785e85SDave Hansen default FLATMEM_MANUAL 14d66d109dSMike Rapoport help 15d66d109dSMike Rapoport This option allows you to change some of the ways that 16d66d109dSMike Rapoport Linux manages its memory internally. Most users will 17d66d109dSMike Rapoport only have one option here selected by the architecture 18d66d109dSMike Rapoport configuration. This is normal. 193a9da765SDave Hansen 20e1785e85SDave Hansenconfig FLATMEM_MANUAL 213a9da765SDave Hansen bool "Flat Memory" 22bb1c50d3SMike Rapoport depends on !ARCH_SPARSEMEM_ENABLE || ARCH_FLATMEM_ENABLE 233a9da765SDave Hansen help 24d66d109dSMike Rapoport This option is best suited for non-NUMA systems with 25d66d109dSMike Rapoport flat address space. The FLATMEM is the most efficient 26d66d109dSMike Rapoport system in terms of performance and resource consumption 27d66d109dSMike Rapoport and it is the best option for smaller systems. 283a9da765SDave Hansen 29d66d109dSMike Rapoport For systems that have holes in their physical address 30d66d109dSMike Rapoport spaces and for features like NUMA and memory hotplug, 31dd33d29aSRandy Dunlap choose "Sparse Memory". 32d41dee36SAndy Whitcroft 33d41dee36SAndy Whitcroft If unsure, choose this option (Flat Memory) over any other. 343a9da765SDave Hansen 35d41dee36SAndy Whitcroftconfig SPARSEMEM_MANUAL 36d41dee36SAndy Whitcroft bool "Sparse Memory" 37d41dee36SAndy Whitcroft depends on ARCH_SPARSEMEM_ENABLE 38d41dee36SAndy Whitcroft help 39d41dee36SAndy Whitcroft This will be the only option for some systems, including 40d66d109dSMike Rapoport memory hot-plug systems. This is normal. 41d41dee36SAndy Whitcroft 42d66d109dSMike Rapoport This option provides efficient support for systems with 43d66d109dSMike Rapoport holes is their physical address space and allows memory 44d66d109dSMike Rapoport hot-plug and hot-remove. 45d41dee36SAndy Whitcroft 46d66d109dSMike Rapoport If unsure, choose "Flat Memory" over this option. 47d41dee36SAndy Whitcroft 483a9da765SDave Hansenendchoice 493a9da765SDave Hansen 50d41dee36SAndy Whitcroftconfig SPARSEMEM 51d41dee36SAndy Whitcroft def_bool y 521a83e175SRussell King depends on (!SELECT_MEMORY_MODEL && ARCH_SPARSEMEM_ENABLE) || SPARSEMEM_MANUAL 53d41dee36SAndy Whitcroft 54e1785e85SDave Hansenconfig FLATMEM 55e1785e85SDave Hansen def_bool y 56bb1c50d3SMike Rapoport depends on !SPARSEMEM || FLATMEM_MANUAL 57d41dee36SAndy Whitcroft 5893b7504eSDave Hansen# 593e347261SBob Picco# SPARSEMEM_EXTREME (which is the default) does some bootmem 60c89ab04fSMike Rapoport# allocations when sparse_init() is called. If this cannot 613e347261SBob Picco# be done on your architecture, select this option. However, 623e347261SBob Picco# statically allocating the mem_section[] array can potentially 633e347261SBob Picco# consume vast quantities of .bss, so be careful. 643e347261SBob Picco# 653e347261SBob Picco# This option will also potentially produce smaller runtime code 663e347261SBob Picco# with gcc 3.4 and later. 673e347261SBob Picco# 683e347261SBob Piccoconfig SPARSEMEM_STATIC 699ba16087SJan Beulich bool 703e347261SBob Picco 713e347261SBob Picco# 7244c09201SMatt LaPlante# Architecture platforms which require a two level mem_section in SPARSEMEM 73802f192eSBob Picco# must select this option. This is usually for architecture platforms with 74802f192eSBob Picco# an extremely sparse physical address space. 75802f192eSBob Picco# 763e347261SBob Piccoconfig SPARSEMEM_EXTREME 773e347261SBob Picco def_bool y 783e347261SBob Picco depends on SPARSEMEM && !SPARSEMEM_STATIC 794c21e2f2SHugh Dickins 8029c71111SAndy Whitcroftconfig SPARSEMEM_VMEMMAP_ENABLE 819ba16087SJan Beulich bool 8229c71111SAndy Whitcroft 8329c71111SAndy Whitcroftconfig SPARSEMEM_VMEMMAP 84a5ee6daaSGeoff Levand bool "Sparse Memory virtual memmap" 85a5ee6daaSGeoff Levand depends on SPARSEMEM && SPARSEMEM_VMEMMAP_ENABLE 86a5ee6daaSGeoff Levand default y 87a5ee6daaSGeoff Levand help 88a5ee6daaSGeoff Levand SPARSEMEM_VMEMMAP uses a virtually mapped memmap to optimise 89a5ee6daaSGeoff Levand pfn_to_page and page_to_pfn operations. This is the most 90a5ee6daaSGeoff Levand efficient option when sufficient kernel resources are available. 9129c71111SAndy Whitcroft 9270210ed9SPhilipp Hachtmannconfig HAVE_MEMBLOCK_PHYS_MAP 936341e62bSChristoph Jaeger bool 9470210ed9SPhilipp Hachtmann 9567a929e0SChristoph Hellwigconfig HAVE_FAST_GUP 96050a9adcSChristoph Hellwig depends on MMU 976341e62bSChristoph Jaeger bool 982667f50eSSteve Capper 99781eb2cdSKefeng Wangconfig HOLES_IN_ZONE 100781eb2cdSKefeng Wang bool 101781eb2cdSKefeng Wang 10252219aeaSDavid Hildenbrand# Don't discard allocated memory used to track "memory" and "reserved" memblocks 10352219aeaSDavid Hildenbrand# after early boot, so it can still be used to test for validity of memory. 10452219aeaSDavid Hildenbrand# Also, memblocks are updated with memory hot(un)plug. 105350e88baSMike Rapoportconfig ARCH_KEEP_MEMBLOCK 1066341e62bSChristoph Jaeger bool 107c378ddd5STejun Heo 1081e5d8e1eSDan Williams# Keep arch NUMA mapping infrastructure post-init. 1091e5d8e1eSDan Williamsconfig NUMA_KEEP_MEMINFO 1101e5d8e1eSDan Williams bool 1111e5d8e1eSDan Williams 112ee6f509cSMinchan Kimconfig MEMORY_ISOLATION 1136341e62bSChristoph Jaeger bool 114ee6f509cSMinchan Kim 11546723bfaSYasuaki Ishimatsu# 11646723bfaSYasuaki Ishimatsu# Only be set on architectures that have completely implemented memory hotplug 11746723bfaSYasuaki Ishimatsu# feature. If you are not sure, don't touch it. 11846723bfaSYasuaki Ishimatsu# 11946723bfaSYasuaki Ishimatsuconfig HAVE_BOOTMEM_INFO_NODE 12046723bfaSYasuaki Ishimatsu def_bool n 12146723bfaSYasuaki Ishimatsu 12291024b3cSAnshuman Khandualconfig ARCH_ENABLE_MEMORY_HOTPLUG 12391024b3cSAnshuman Khandual bool 12491024b3cSAnshuman Khandual 1253947be19SDave Hansen# eventually, we can have this option just 'select SPARSEMEM' 1263947be19SDave Hansenconfig MEMORY_HOTPLUG 1273947be19SDave Hansen bool "Allow for memory hot-add" 128b30c5927SDavid Hildenbrand select MEMORY_ISOLATION 129ec69acbbSKeith Mannthey depends on SPARSEMEM || X86_64_ACPI_NUMA 13040b31360SStephen Rothwell depends on ARCH_ENABLE_MEMORY_HOTPLUG 131b59d02edSMichal Hocko depends on 64BIT || BROKEN 1321e5d8e1eSDan Williams select NUMA_KEEP_MEMINFO if NUMA 1333947be19SDave Hansen 134ec69acbbSKeith Manntheyconfig MEMORY_HOTPLUG_SPARSE 135ec69acbbSKeith Mannthey def_bool y 136ec69acbbSKeith Mannthey depends on SPARSEMEM && MEMORY_HOTPLUG 137ec69acbbSKeith Mannthey 1388604d9e5SVitaly Kuznetsovconfig MEMORY_HOTPLUG_DEFAULT_ONLINE 1398604d9e5SVitaly Kuznetsov bool "Online the newly added memory blocks by default" 1408604d9e5SVitaly Kuznetsov depends on MEMORY_HOTPLUG 1418604d9e5SVitaly Kuznetsov help 1428604d9e5SVitaly Kuznetsov This option sets the default policy setting for memory hotplug 1438604d9e5SVitaly Kuznetsov onlining policy (/sys/devices/system/memory/auto_online_blocks) which 1448604d9e5SVitaly Kuznetsov determines what happens to newly added memory regions. Policy setting 1458604d9e5SVitaly Kuznetsov can always be changed at runtime. 146cb1aaebeSMauro Carvalho Chehab See Documentation/admin-guide/mm/memory-hotplug.rst for more information. 1478604d9e5SVitaly Kuznetsov 1488604d9e5SVitaly Kuznetsov Say Y here if you want all hot-plugged memory blocks to appear in 1498604d9e5SVitaly Kuznetsov 'online' state by default. 1508604d9e5SVitaly Kuznetsov Say N here if you want the default policy to keep all hot-plugged 1518604d9e5SVitaly Kuznetsov memory blocks in 'offline' state. 1528604d9e5SVitaly Kuznetsov 15391024b3cSAnshuman Khandualconfig ARCH_ENABLE_MEMORY_HOTREMOVE 15491024b3cSAnshuman Khandual bool 15591024b3cSAnshuman Khandual 1560c0e6195SKAMEZAWA Hiroyukiconfig MEMORY_HOTREMOVE 1570c0e6195SKAMEZAWA Hiroyuki bool "Allow for memory hot remove" 158f7e3334aSNathan Fontenot select HAVE_BOOTMEM_INFO_NODE if (X86_64 || PPC64) 1590c0e6195SKAMEZAWA Hiroyuki depends on MEMORY_HOTPLUG && ARCH_ENABLE_MEMORY_HOTREMOVE 1600c0e6195SKAMEZAWA Hiroyuki depends on MIGRATION 1610c0e6195SKAMEZAWA Hiroyuki 162a08a2ae3SOscar Salvadorconfig MHP_MEMMAP_ON_MEMORY 163a08a2ae3SOscar Salvador def_bool y 164a08a2ae3SOscar Salvador depends on MEMORY_HOTPLUG && SPARSEMEM_VMEMMAP 165a08a2ae3SOscar Salvador depends on ARCH_MHP_MEMMAP_ON_MEMORY_ENABLE 166a08a2ae3SOscar Salvador 1674c21e2f2SHugh Dickins# Heavily threaded applications may benefit from splitting the mm-wide 1684c21e2f2SHugh Dickins# page_table_lock, so that faults on different parts of the user address 1694c21e2f2SHugh Dickins# space can be handled with less contention: split it at this NR_CPUS. 1704c21e2f2SHugh Dickins# Default to 4 for wider testing, though 8 might be more appropriate. 1714c21e2f2SHugh Dickins# ARM's adjust_pte (unused if VIPT) depends on mm-wide page_table_lock. 1727b6ac9dfSHugh Dickins# PA-RISC 7xxx's spinlock_t would enlarge struct page from 32 to 44 bytes. 17360bccaa6SWill Deacon# SPARC32 allocates multiple pte tables within a single page, and therefore 17460bccaa6SWill Deacon# a per-page lock leads to problems when multiple tables need to be locked 17560bccaa6SWill Deacon# at the same time (e.g. copy_page_range()). 176a70caa8bSHugh Dickins# DEBUG_SPINLOCK and DEBUG_LOCK_ALLOC spinlock_t also enlarge struct page. 1774c21e2f2SHugh Dickins# 1784c21e2f2SHugh Dickinsconfig SPLIT_PTLOCK_CPUS 1794c21e2f2SHugh Dickins int 1809164550eSKirill A. Shutemov default "999999" if !MMU 181a70caa8bSHugh Dickins default "999999" if ARM && !CPU_CACHE_VIPT 182a70caa8bSHugh Dickins default "999999" if PARISC && !PA20 18360bccaa6SWill Deacon default "999999" if SPARC32 1844c21e2f2SHugh Dickins default "4" 1857cbe34cfSChristoph Lameter 186e009bb30SKirill A. Shutemovconfig ARCH_ENABLE_SPLIT_PMD_PTLOCK 1876341e62bSChristoph Jaeger bool 188e009bb30SKirill A. Shutemov 1897cbe34cfSChristoph Lameter# 19009316c09SKonstantin Khlebnikov# support for memory balloon 19109316c09SKonstantin Khlebnikovconfig MEMORY_BALLOON 1926341e62bSChristoph Jaeger bool 19309316c09SKonstantin Khlebnikov 19409316c09SKonstantin Khlebnikov# 19518468d93SRafael Aquini# support for memory balloon compaction 19618468d93SRafael Aquiniconfig BALLOON_COMPACTION 19718468d93SRafael Aquini bool "Allow for balloon memory compaction/migration" 19818468d93SRafael Aquini def_bool y 19909316c09SKonstantin Khlebnikov depends on COMPACTION && MEMORY_BALLOON 20018468d93SRafael Aquini help 20118468d93SRafael Aquini Memory fragmentation introduced by ballooning might reduce 20218468d93SRafael Aquini significantly the number of 2MB contiguous memory blocks that can be 20318468d93SRafael Aquini used within a guest, thus imposing performance penalties associated 20418468d93SRafael Aquini with the reduced number of transparent huge pages that could be used 20518468d93SRafael Aquini by the guest workload. Allowing the compaction & migration for memory 20618468d93SRafael Aquini pages enlisted as being part of memory balloon devices avoids the 20718468d93SRafael Aquini scenario aforementioned and helps improving memory defragmentation. 20818468d93SRafael Aquini 20918468d93SRafael Aquini# 210e9e96b39SMel Gorman# support for memory compaction 211e9e96b39SMel Gormanconfig COMPACTION 212e9e96b39SMel Gorman bool "Allow for memory compaction" 21305106e6aSRik van Riel def_bool y 214e9e96b39SMel Gorman select MIGRATION 21533a93877SAndrea Arcangeli depends on MMU 216e9e96b39SMel Gorman help 217b32eaf71SMichal Hocko Compaction is the only memory management component to form 218b32eaf71SMichal Hocko high order (larger physically contiguous) memory blocks 219b32eaf71SMichal Hocko reliably. The page allocator relies on compaction heavily and 220b32eaf71SMichal Hocko the lack of the feature can lead to unexpected OOM killer 221b32eaf71SMichal Hocko invocations for high order memory requests. You shouldn't 222b32eaf71SMichal Hocko disable this option unless there really is a strong reason for 223b32eaf71SMichal Hocko it and then we would be really interested to hear about that at 224b32eaf71SMichal Hocko linux-mm@kvack.org. 225e9e96b39SMel Gorman 226e9e96b39SMel Gorman# 22736e66c55SAlexander Duyck# support for free page reporting 22836e66c55SAlexander Duyckconfig PAGE_REPORTING 22936e66c55SAlexander Duyck bool "Free page reporting" 23036e66c55SAlexander Duyck def_bool n 23136e66c55SAlexander Duyck help 23236e66c55SAlexander Duyck Free page reporting allows for the incremental acquisition of 23336e66c55SAlexander Duyck free pages from the buddy allocator for the purpose of reporting 23436e66c55SAlexander Duyck those pages to another entity, such as a hypervisor, so that the 23536e66c55SAlexander Duyck memory can be freed within the host for other uses. 23636e66c55SAlexander Duyck 23736e66c55SAlexander Duyck# 2387cbe34cfSChristoph Lameter# support for page migration 2397cbe34cfSChristoph Lameter# 2407cbe34cfSChristoph Lameterconfig MIGRATION 241b20a3503SChristoph Lameter bool "Page migration" 2426c5240aeSChristoph Lameter def_bool y 243de32a817SChen Gang depends on (NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION || CMA) && MMU 244b20a3503SChristoph Lameter help 245b20a3503SChristoph Lameter Allows the migration of the physical location of pages of processes 246e9e96b39SMel Gorman while the virtual addresses are not changed. This is useful in 247e9e96b39SMel Gorman two situations. The first is on NUMA systems to put pages nearer 248e9e96b39SMel Gorman to the processors accessing. The second is when allocating huge 249e9e96b39SMel Gorman pages as migration can relocate pages to satisfy a huge page 250e9e96b39SMel Gorman allocation instead of reclaiming. 2516550e07fSGreg Kroah-Hartman 252c177c81eSNaoya Horiguchiconfig ARCH_ENABLE_HUGEPAGE_MIGRATION 2536341e62bSChristoph Jaeger bool 254c177c81eSNaoya Horiguchi 2559c670ea3SNaoya Horiguchiconfig ARCH_ENABLE_THP_MIGRATION 2569c670ea3SNaoya Horiguchi bool 2579c670ea3SNaoya Horiguchi 2584bfb68a0SAnshuman Khandualconfig HUGETLB_PAGE_SIZE_VARIABLE 2594bfb68a0SAnshuman Khandual def_bool n 2604bfb68a0SAnshuman Khandual help 2614bfb68a0SAnshuman Khandual Allows the pageblock_order value to be dynamic instead of just standard 2624bfb68a0SAnshuman Khandual HUGETLB_PAGE_ORDER when there are multiple HugeTLB page sizes available 2634bfb68a0SAnshuman Khandual on a platform. 2644bfb68a0SAnshuman Khandual 2658df995f6SAlexandre Ghiticonfig CONTIG_ALLOC 2668df995f6SAlexandre Ghiti def_bool (MEMORY_ISOLATION && COMPACTION) || CMA 2678df995f6SAlexandre Ghiti 268600715dcSJeremy Fitzhardingeconfig PHYS_ADDR_T_64BIT 269d4a451d5SChristoph Hellwig def_bool 64BIT 270600715dcSJeremy Fitzhardinge 2712a7326b5SChristoph Lameterconfig BOUNCE 2729ca24e2eSVinayak Menon bool "Enable bounce buffers" 2739ca24e2eSVinayak Menon default y 274ce288e05SChristoph Hellwig depends on BLOCK && MMU && HIGHMEM 2759ca24e2eSVinayak Menon help 276ce288e05SChristoph Hellwig Enable bounce buffers for devices that cannot access the full range of 277ce288e05SChristoph Hellwig memory available to the CPU. Enabled by default when HIGHMEM is 278ce288e05SChristoph Hellwig selected, but you may say n to override this. 2792a7326b5SChristoph Lameter 280f057eac0SStephen Rothwellconfig VIRT_TO_BUS 2814febd95aSStephen Rothwell bool 2824febd95aSStephen Rothwell help 2834febd95aSStephen Rothwell An architecture should select this if it implements the 2844febd95aSStephen Rothwell deprecated interface virt_to_bus(). All new architectures 2854febd95aSStephen Rothwell should probably not select this. 2864febd95aSStephen Rothwell 287cddb8a5cSAndrea Arcangeli 288cddb8a5cSAndrea Arcangeliconfig MMU_NOTIFIER 289cddb8a5cSAndrea Arcangeli bool 29083fe27eaSPranith Kumar select SRCU 29199cb252fSJason Gunthorpe select INTERVAL_TREE 292fc4d5c29SDavid Howells 293f8af4da3SHugh Dickinsconfig KSM 294f8af4da3SHugh Dickins bool "Enable KSM for page merging" 295f8af4da3SHugh Dickins depends on MMU 29659e1a2f4STimofey Titovets select XXHASH 297f8af4da3SHugh Dickins help 298f8af4da3SHugh Dickins Enable Kernel Samepage Merging: KSM periodically scans those areas 299f8af4da3SHugh Dickins of an application's address space that an app has advised may be 300f8af4da3SHugh Dickins mergeable. When it finds pages of identical content, it replaces 301d0f209f6SHugh Dickins the many instances by a single page with that content, so 302f8af4da3SHugh Dickins saving memory until one or another app needs to modify the content. 303f8af4da3SHugh Dickins Recommended for use with KVM, or with other duplicative applications. 304ad56b738SMike Rapoport See Documentation/vm/ksm.rst for more information: KSM is inactive 305c73602adSHugh Dickins until a program has madvised that an area is MADV_MERGEABLE, and 306c73602adSHugh Dickins root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set). 307f8af4da3SHugh Dickins 308e0a94c2aSChristoph Lameterconfig DEFAULT_MMAP_MIN_ADDR 309e0a94c2aSChristoph Lameter int "Low address space to protect from user allocation" 3106e141546SDavid Howells depends on MMU 311e0a94c2aSChristoph Lameter default 4096 312e0a94c2aSChristoph Lameter help 313e0a94c2aSChristoph Lameter This is the portion of low virtual memory which should be protected 314e0a94c2aSChristoph Lameter from userspace allocation. Keeping a user from writing to low pages 315e0a94c2aSChristoph Lameter can help reduce the impact of kernel NULL pointer bugs. 316e0a94c2aSChristoph Lameter 317e0a94c2aSChristoph Lameter For most ia64, ppc64 and x86 users with lots of address space 318e0a94c2aSChristoph Lameter a value of 65536 is reasonable and should cause no problems. 319e0a94c2aSChristoph Lameter On arm and other archs it should not be higher than 32768. 320788084abSEric Paris Programs which use vm86 functionality or have some need to map 321788084abSEric Paris this low address space will need CAP_SYS_RAWIO or disable this 322788084abSEric Paris protection by setting the value to 0. 323e0a94c2aSChristoph Lameter 324e0a94c2aSChristoph Lameter This value can be changed after boot using the 325e0a94c2aSChristoph Lameter /proc/sys/vm/mmap_min_addr tunable. 326e0a94c2aSChristoph Lameter 327d949f36fSLinus Torvaldsconfig ARCH_SUPPORTS_MEMORY_FAILURE 328d949f36fSLinus Torvalds bool 329e0a94c2aSChristoph Lameter 3306a46079cSAndi Kleenconfig MEMORY_FAILURE 3316a46079cSAndi Kleen depends on MMU 332d949f36fSLinus Torvalds depends on ARCH_SUPPORTS_MEMORY_FAILURE 3336a46079cSAndi Kleen bool "Enable recovery from hardware memory errors" 334ee6f509cSMinchan Kim select MEMORY_ISOLATION 33597f0b134SXie XiuQi select RAS 3366a46079cSAndi Kleen help 3376a46079cSAndi Kleen Enables code to recover from some memory failures on systems 3386a46079cSAndi Kleen with MCA recovery. This allows a system to continue running 3396a46079cSAndi Kleen even when some of its memory has uncorrected errors. This requires 3406a46079cSAndi Kleen special hardware support and typically ECC memory. 3416a46079cSAndi Kleen 342cae681fcSAndi Kleenconfig HWPOISON_INJECT 343413f9efbSAndi Kleen tristate "HWPoison pages injector" 34427df5068SAndi Kleen depends on MEMORY_FAILURE && DEBUG_KERNEL && PROC_FS 345478c5ffcSWu Fengguang select PROC_PAGE_MONITOR 346cae681fcSAndi Kleen 347fc4d5c29SDavid Howellsconfig NOMMU_INITIAL_TRIM_EXCESS 348fc4d5c29SDavid Howells int "Turn on mmap() excess space trimming before booting" 349fc4d5c29SDavid Howells depends on !MMU 350fc4d5c29SDavid Howells default 1 351fc4d5c29SDavid Howells help 352fc4d5c29SDavid Howells The NOMMU mmap() frequently needs to allocate large contiguous chunks 353fc4d5c29SDavid Howells of memory on which to store mappings, but it can only ask the system 354fc4d5c29SDavid Howells allocator for chunks in 2^N*PAGE_SIZE amounts - which is frequently 355fc4d5c29SDavid Howells more than it requires. To deal with this, mmap() is able to trim off 356fc4d5c29SDavid Howells the excess and return it to the allocator. 357fc4d5c29SDavid Howells 358fc4d5c29SDavid Howells If trimming is enabled, the excess is trimmed off and returned to the 359fc4d5c29SDavid Howells system allocator, which can cause extra fragmentation, particularly 360fc4d5c29SDavid Howells if there are a lot of transient processes. 361fc4d5c29SDavid Howells 362fc4d5c29SDavid Howells If trimming is disabled, the excess is kept, but not used, which for 363fc4d5c29SDavid Howells long-term mappings means that the space is wasted. 364fc4d5c29SDavid Howells 365fc4d5c29SDavid Howells Trimming can be dynamically controlled through a sysctl option 366fc4d5c29SDavid Howells (/proc/sys/vm/nr_trim_pages) which specifies the minimum number of 367fc4d5c29SDavid Howells excess pages there must be before trimming should occur, or zero if 368fc4d5c29SDavid Howells no trimming is to occur. 369fc4d5c29SDavid Howells 370fc4d5c29SDavid Howells This option specifies the initial value of this option. The default 371fc4d5c29SDavid Howells of 1 says that all excess pages should be trimmed. 372fc4d5c29SDavid Howells 373dd19d293SStephen Kitt See Documentation/admin-guide/mm/nommu-mmap.rst for more information. 374bbddff05STejun Heo 3754c76d9d1SAndrea Arcangeliconfig TRANSPARENT_HUGEPAGE 37613ece886SAndrea Arcangeli bool "Transparent Hugepage Support" 37715626062SGerald Schaefer depends on HAVE_ARCH_TRANSPARENT_HUGEPAGE 3785d689240SAndrea Arcangeli select COMPACTION 3793a08cd52SMatthew Wilcox select XARRAY_MULTI 3804c76d9d1SAndrea Arcangeli help 3814c76d9d1SAndrea Arcangeli Transparent Hugepages allows the kernel to use huge pages and 3824c76d9d1SAndrea Arcangeli huge tlb transparently to the applications whenever possible. 3834c76d9d1SAndrea Arcangeli This feature can improve computing performance to certain 3844c76d9d1SAndrea Arcangeli applications by speeding up page faults during memory 3854c76d9d1SAndrea Arcangeli allocation, by reducing the number of tlb misses and by speeding 3864c76d9d1SAndrea Arcangeli up the pagetable walking. 3874c76d9d1SAndrea Arcangeli 3884c76d9d1SAndrea Arcangeli If memory constrained on embedded, you may want to say N. 3894c76d9d1SAndrea Arcangeli 39013ece886SAndrea Arcangelichoice 39113ece886SAndrea Arcangeli prompt "Transparent Hugepage Support sysfs defaults" 39213ece886SAndrea Arcangeli depends on TRANSPARENT_HUGEPAGE 39313ece886SAndrea Arcangeli default TRANSPARENT_HUGEPAGE_ALWAYS 39413ece886SAndrea Arcangeli help 39513ece886SAndrea Arcangeli Selects the sysfs defaults for Transparent Hugepage Support. 39613ece886SAndrea Arcangeli 39713ece886SAndrea Arcangeli config TRANSPARENT_HUGEPAGE_ALWAYS 39813ece886SAndrea Arcangeli bool "always" 39913ece886SAndrea Arcangeli help 40013ece886SAndrea Arcangeli Enabling Transparent Hugepage always, can increase the 40113ece886SAndrea Arcangeli memory footprint of applications without a guaranteed 40213ece886SAndrea Arcangeli benefit but it will work automatically for all applications. 40313ece886SAndrea Arcangeli 40413ece886SAndrea Arcangeli config TRANSPARENT_HUGEPAGE_MADVISE 40513ece886SAndrea Arcangeli bool "madvise" 40613ece886SAndrea Arcangeli help 40713ece886SAndrea Arcangeli Enabling Transparent Hugepage madvise, will only provide a 40813ece886SAndrea Arcangeli performance improvement benefit to the applications using 40913ece886SAndrea Arcangeli madvise(MADV_HUGEPAGE) but it won't risk to increase the 41013ece886SAndrea Arcangeli memory footprint of applications without a guaranteed 41113ece886SAndrea Arcangeli benefit. 41213ece886SAndrea Arcangeliendchoice 41313ece886SAndrea Arcangeli 41438d8b4e6SHuang Yingconfig ARCH_WANTS_THP_SWAP 41538d8b4e6SHuang Ying def_bool n 41638d8b4e6SHuang Ying 41738d8b4e6SHuang Yingconfig THP_SWAP 41838d8b4e6SHuang Ying def_bool y 41914fef284SHuang Ying depends on TRANSPARENT_HUGEPAGE && ARCH_WANTS_THP_SWAP && SWAP 42038d8b4e6SHuang Ying help 42138d8b4e6SHuang Ying Swap transparent huge pages in one piece, without splitting. 42214fef284SHuang Ying XXX: For now, swap cluster backing transparent huge page 42314fef284SHuang Ying will be split after swapout. 42438d8b4e6SHuang Ying 42538d8b4e6SHuang Ying For selection by architectures with reasonable THP sizes. 42638d8b4e6SHuang Ying 427e496cf3dSKirill A. Shutemov# 428bbddff05STejun Heo# UP and nommu archs use km based percpu allocator 429bbddff05STejun Heo# 430bbddff05STejun Heoconfig NEED_PER_CPU_KM 431bbddff05STejun Heo depends on !SMP 432bbddff05STejun Heo bool 433bbddff05STejun Heo default y 434077b1f83SDan Magenheimer 435077b1f83SDan Magenheimerconfig CLEANCACHE 436077b1f83SDan Magenheimer bool "Enable cleancache driver to cache clean pages if tmem is present" 437077b1f83SDan Magenheimer help 438077b1f83SDan Magenheimer Cleancache can be thought of as a page-granularity victim cache 439077b1f83SDan Magenheimer for clean pages that the kernel's pageframe replacement algorithm 440077b1f83SDan Magenheimer (PFRA) would like to keep around, but can't since there isn't enough 441077b1f83SDan Magenheimer memory. So when the PFRA "evicts" a page, it first attempts to use 442140a1ef2SMichael Witten cleancache code to put the data contained in that page into 443077b1f83SDan Magenheimer "transcendent memory", memory that is not directly accessible or 444077b1f83SDan Magenheimer addressable by the kernel and is of unknown and possibly 445077b1f83SDan Magenheimer time-varying size. And when a cleancache-enabled 446077b1f83SDan Magenheimer filesystem wishes to access a page in a file on disk, it first 447077b1f83SDan Magenheimer checks cleancache to see if it already contains it; if it does, 448077b1f83SDan Magenheimer the page is copied into the kernel and a disk access is avoided. 449077b1f83SDan Magenheimer When a transcendent memory driver is available (such as zcache or 450077b1f83SDan Magenheimer Xen transcendent memory), a significant I/O reduction 451077b1f83SDan Magenheimer may be achieved. When none is available, all cleancache calls 452077b1f83SDan Magenheimer are reduced to a single pointer-compare-against-NULL resulting 453077b1f83SDan Magenheimer in a negligible performance hit. 454077b1f83SDan Magenheimer 455077b1f83SDan Magenheimer If unsure, say Y to enable cleancache 45627c6aec2SDan Magenheimer 45727c6aec2SDan Magenheimerconfig FRONTSWAP 45827c6aec2SDan Magenheimer bool "Enable frontswap to cache swap pages if tmem is present" 45927c6aec2SDan Magenheimer depends on SWAP 46027c6aec2SDan Magenheimer help 46127c6aec2SDan Magenheimer Frontswap is so named because it can be thought of as the opposite 46227c6aec2SDan Magenheimer of a "backing" store for a swap device. The data is stored into 46327c6aec2SDan Magenheimer "transcendent memory", memory that is not directly accessible or 46427c6aec2SDan Magenheimer addressable by the kernel and is of unknown and possibly 46527c6aec2SDan Magenheimer time-varying size. When space in transcendent memory is available, 46627c6aec2SDan Magenheimer a significant swap I/O reduction may be achieved. When none is 46727c6aec2SDan Magenheimer available, all frontswap calls are reduced to a single pointer- 46827c6aec2SDan Magenheimer compare-against-NULL resulting in a negligible performance hit 46927c6aec2SDan Magenheimer and swap data is stored as normal on the matching swap device. 47027c6aec2SDan Magenheimer 47127c6aec2SDan Magenheimer If unsure, say Y to enable frontswap. 472f825c736SAneesh Kumar K.V 473f825c736SAneesh Kumar K.Vconfig CMA 474f825c736SAneesh Kumar K.V bool "Contiguous Memory Allocator" 475aca52c39SMike Rapoport depends on MMU 476f825c736SAneesh Kumar K.V select MIGRATION 477f825c736SAneesh Kumar K.V select MEMORY_ISOLATION 478f825c736SAneesh Kumar K.V help 479f825c736SAneesh Kumar K.V This enables the Contiguous Memory Allocator which allows other 480f825c736SAneesh Kumar K.V subsystems to allocate big physically-contiguous blocks of memory. 481f825c736SAneesh Kumar K.V CMA reserves a region of memory and allows only movable pages to 482f825c736SAneesh Kumar K.V be allocated from it. This way, the kernel can use the memory for 483f825c736SAneesh Kumar K.V pagecache and when a subsystem requests for contiguous area, the 484f825c736SAneesh Kumar K.V allocated pages are migrated away to serve the contiguous request. 485f825c736SAneesh Kumar K.V 486f825c736SAneesh Kumar K.V If unsure, say "n". 487f825c736SAneesh Kumar K.V 488f825c736SAneesh Kumar K.Vconfig CMA_DEBUG 489f825c736SAneesh Kumar K.V bool "CMA debug messages (DEVELOPMENT)" 490f825c736SAneesh Kumar K.V depends on DEBUG_KERNEL && CMA 491f825c736SAneesh Kumar K.V help 492f825c736SAneesh Kumar K.V Turns on debug messages in CMA. This produces KERN_DEBUG 493f825c736SAneesh Kumar K.V messages for every CMA call as well as various messages while 494f825c736SAneesh Kumar K.V processing calls such as dma_alloc_from_contiguous(). 495f825c736SAneesh Kumar K.V This option does not affect warning and error messages. 496bf550fc9SAlexander Graf 49728b24c1fSSasha Levinconfig CMA_DEBUGFS 49828b24c1fSSasha Levin bool "CMA debugfs interface" 49928b24c1fSSasha Levin depends on CMA && DEBUG_FS 50028b24c1fSSasha Levin help 50128b24c1fSSasha Levin Turns on the DebugFS interface for CMA. 50228b24c1fSSasha Levin 50343ca106fSMinchan Kimconfig CMA_SYSFS 50443ca106fSMinchan Kim bool "CMA information through sysfs interface" 50543ca106fSMinchan Kim depends on CMA && SYSFS 50643ca106fSMinchan Kim help 50743ca106fSMinchan Kim This option exposes some sysfs attributes to get information 50843ca106fSMinchan Kim from CMA. 50943ca106fSMinchan Kim 510a254129eSJoonsoo Kimconfig CMA_AREAS 511a254129eSJoonsoo Kim int "Maximum count of the CMA areas" 512a254129eSJoonsoo Kim depends on CMA 513b7176c26SBarry Song default 19 if NUMA 514a254129eSJoonsoo Kim default 7 515a254129eSJoonsoo Kim help 516a254129eSJoonsoo Kim CMA allows to create CMA areas for particular purpose, mainly, 517a254129eSJoonsoo Kim used as device private area. This parameter sets the maximum 518a254129eSJoonsoo Kim number of CMA area in the system. 519a254129eSJoonsoo Kim 520b7176c26SBarry Song If unsure, leave the default value "7" in UMA and "19" in NUMA. 521a254129eSJoonsoo Kim 522af8d417aSDan Streetmanconfig MEM_SOFT_DIRTY 523af8d417aSDan Streetman bool "Track memory changes" 524af8d417aSDan Streetman depends on CHECKPOINT_RESTORE && HAVE_ARCH_SOFT_DIRTY && PROC_FS 525af8d417aSDan Streetman select PROC_PAGE_MONITOR 5264e2e2770SSeth Jennings help 527af8d417aSDan Streetman This option enables memory changes tracking by introducing a 528af8d417aSDan Streetman soft-dirty bit on pte-s. This bit it set when someone writes 529af8d417aSDan Streetman into a page just as regular dirty bit, but unlike the latter 530af8d417aSDan Streetman it can be cleared by hands. 531af8d417aSDan Streetman 5321ad1335dSMike Rapoport See Documentation/admin-guide/mm/soft-dirty.rst for more details. 5334e2e2770SSeth Jennings 5342b281117SSeth Jenningsconfig ZSWAP 5352b281117SSeth Jennings bool "Compressed cache for swap pages (EXPERIMENTAL)" 5362b281117SSeth Jennings depends on FRONTSWAP && CRYPTO=y 53712d79d64SDan Streetman select ZPOOL 5382b281117SSeth Jennings help 5392b281117SSeth Jennings A lightweight compressed cache for swap pages. It takes 5402b281117SSeth Jennings pages that are in the process of being swapped out and attempts to 5412b281117SSeth Jennings compress them into a dynamically allocated RAM-based memory pool. 5422b281117SSeth Jennings This can result in a significant I/O reduction on swap device and, 5432b281117SSeth Jennings in the case where decompressing from RAM is faster that swap device 5442b281117SSeth Jennings reads, can also improve workload performance. 5452b281117SSeth Jennings 5462b281117SSeth Jennings This is marked experimental because it is a new feature (as of 5472b281117SSeth Jennings v3.11) that interacts heavily with memory reclaim. While these 5482b281117SSeth Jennings interactions don't cause any known issues on simple memory setups, 5492b281117SSeth Jennings they have not be fully explored on the large set of potential 5502b281117SSeth Jennings configurations and workloads that exist. 5512b281117SSeth Jennings 552bb8b93b5SMaciej S. Szmigierochoice 553bb8b93b5SMaciej S. Szmigiero prompt "Compressed cache for swap pages default compressor" 554bb8b93b5SMaciej S. Szmigiero depends on ZSWAP 555bb8b93b5SMaciej S. Szmigiero default ZSWAP_COMPRESSOR_DEFAULT_LZO 556bb8b93b5SMaciej S. Szmigiero help 557bb8b93b5SMaciej S. Szmigiero Selects the default compression algorithm for the compressed cache 558bb8b93b5SMaciej S. Szmigiero for swap pages. 559bb8b93b5SMaciej S. Szmigiero 560bb8b93b5SMaciej S. Szmigiero For an overview what kind of performance can be expected from 561bb8b93b5SMaciej S. Szmigiero a particular compression algorithm please refer to the benchmarks 562bb8b93b5SMaciej S. Szmigiero available at the following LWN page: 563bb8b93b5SMaciej S. Szmigiero https://lwn.net/Articles/751795/ 564bb8b93b5SMaciej S. Szmigiero 565bb8b93b5SMaciej S. Szmigiero If in doubt, select 'LZO'. 566bb8b93b5SMaciej S. Szmigiero 567bb8b93b5SMaciej S. Szmigiero The selection made here can be overridden by using the kernel 568bb8b93b5SMaciej S. Szmigiero command line 'zswap.compressor=' option. 569bb8b93b5SMaciej S. Szmigiero 570bb8b93b5SMaciej S. Szmigieroconfig ZSWAP_COMPRESSOR_DEFAULT_DEFLATE 571bb8b93b5SMaciej S. Szmigiero bool "Deflate" 572bb8b93b5SMaciej S. Szmigiero select CRYPTO_DEFLATE 573bb8b93b5SMaciej S. Szmigiero help 574bb8b93b5SMaciej S. Szmigiero Use the Deflate algorithm as the default compression algorithm. 575bb8b93b5SMaciej S. Szmigiero 576bb8b93b5SMaciej S. Szmigieroconfig ZSWAP_COMPRESSOR_DEFAULT_LZO 577bb8b93b5SMaciej S. Szmigiero bool "LZO" 578bb8b93b5SMaciej S. Szmigiero select CRYPTO_LZO 579bb8b93b5SMaciej S. Szmigiero help 580bb8b93b5SMaciej S. Szmigiero Use the LZO algorithm as the default compression algorithm. 581bb8b93b5SMaciej S. Szmigiero 582bb8b93b5SMaciej S. Szmigieroconfig ZSWAP_COMPRESSOR_DEFAULT_842 583bb8b93b5SMaciej S. Szmigiero bool "842" 584bb8b93b5SMaciej S. Szmigiero select CRYPTO_842 585bb8b93b5SMaciej S. Szmigiero help 586bb8b93b5SMaciej S. Szmigiero Use the 842 algorithm as the default compression algorithm. 587bb8b93b5SMaciej S. Szmigiero 588bb8b93b5SMaciej S. Szmigieroconfig ZSWAP_COMPRESSOR_DEFAULT_LZ4 589bb8b93b5SMaciej S. Szmigiero bool "LZ4" 590bb8b93b5SMaciej S. Szmigiero select CRYPTO_LZ4 591bb8b93b5SMaciej S. Szmigiero help 592bb8b93b5SMaciej S. Szmigiero Use the LZ4 algorithm as the default compression algorithm. 593bb8b93b5SMaciej S. Szmigiero 594bb8b93b5SMaciej S. Szmigieroconfig ZSWAP_COMPRESSOR_DEFAULT_LZ4HC 595bb8b93b5SMaciej S. Szmigiero bool "LZ4HC" 596bb8b93b5SMaciej S. Szmigiero select CRYPTO_LZ4HC 597bb8b93b5SMaciej S. Szmigiero help 598bb8b93b5SMaciej S. Szmigiero Use the LZ4HC algorithm as the default compression algorithm. 599bb8b93b5SMaciej S. Szmigiero 600bb8b93b5SMaciej S. Szmigieroconfig ZSWAP_COMPRESSOR_DEFAULT_ZSTD 601bb8b93b5SMaciej S. Szmigiero bool "zstd" 602bb8b93b5SMaciej S. Szmigiero select CRYPTO_ZSTD 603bb8b93b5SMaciej S. Szmigiero help 604bb8b93b5SMaciej S. Szmigiero Use the zstd algorithm as the default compression algorithm. 605bb8b93b5SMaciej S. Szmigieroendchoice 606bb8b93b5SMaciej S. Szmigiero 607bb8b93b5SMaciej S. Szmigieroconfig ZSWAP_COMPRESSOR_DEFAULT 608bb8b93b5SMaciej S. Szmigiero string 609bb8b93b5SMaciej S. Szmigiero depends on ZSWAP 610bb8b93b5SMaciej S. Szmigiero default "deflate" if ZSWAP_COMPRESSOR_DEFAULT_DEFLATE 611bb8b93b5SMaciej S. Szmigiero default "lzo" if ZSWAP_COMPRESSOR_DEFAULT_LZO 612bb8b93b5SMaciej S. Szmigiero default "842" if ZSWAP_COMPRESSOR_DEFAULT_842 613bb8b93b5SMaciej S. Szmigiero default "lz4" if ZSWAP_COMPRESSOR_DEFAULT_LZ4 614bb8b93b5SMaciej S. Szmigiero default "lz4hc" if ZSWAP_COMPRESSOR_DEFAULT_LZ4HC 615bb8b93b5SMaciej S. Szmigiero default "zstd" if ZSWAP_COMPRESSOR_DEFAULT_ZSTD 616bb8b93b5SMaciej S. Szmigiero default "" 617bb8b93b5SMaciej S. Szmigiero 618bb8b93b5SMaciej S. Szmigierochoice 619bb8b93b5SMaciej S. Szmigiero prompt "Compressed cache for swap pages default allocator" 620bb8b93b5SMaciej S. Szmigiero depends on ZSWAP 621bb8b93b5SMaciej S. Szmigiero default ZSWAP_ZPOOL_DEFAULT_ZBUD 622bb8b93b5SMaciej S. Szmigiero help 623bb8b93b5SMaciej S. Szmigiero Selects the default allocator for the compressed cache for 624bb8b93b5SMaciej S. Szmigiero swap pages. 625bb8b93b5SMaciej S. Szmigiero The default is 'zbud' for compatibility, however please do 626bb8b93b5SMaciej S. Szmigiero read the description of each of the allocators below before 627bb8b93b5SMaciej S. Szmigiero making a right choice. 628bb8b93b5SMaciej S. Szmigiero 629bb8b93b5SMaciej S. Szmigiero The selection made here can be overridden by using the kernel 630bb8b93b5SMaciej S. Szmigiero command line 'zswap.zpool=' option. 631bb8b93b5SMaciej S. Szmigiero 632bb8b93b5SMaciej S. Szmigieroconfig ZSWAP_ZPOOL_DEFAULT_ZBUD 633bb8b93b5SMaciej S. Szmigiero bool "zbud" 634bb8b93b5SMaciej S. Szmigiero select ZBUD 635bb8b93b5SMaciej S. Szmigiero help 636bb8b93b5SMaciej S. Szmigiero Use the zbud allocator as the default allocator. 637bb8b93b5SMaciej S. Szmigiero 638bb8b93b5SMaciej S. Szmigieroconfig ZSWAP_ZPOOL_DEFAULT_Z3FOLD 639bb8b93b5SMaciej S. Szmigiero bool "z3fold" 640bb8b93b5SMaciej S. Szmigiero select Z3FOLD 641bb8b93b5SMaciej S. Szmigiero help 642bb8b93b5SMaciej S. Szmigiero Use the z3fold allocator as the default allocator. 643bb8b93b5SMaciej S. Szmigiero 644bb8b93b5SMaciej S. Szmigieroconfig ZSWAP_ZPOOL_DEFAULT_ZSMALLOC 645bb8b93b5SMaciej S. Szmigiero bool "zsmalloc" 646bb8b93b5SMaciej S. Szmigiero select ZSMALLOC 647bb8b93b5SMaciej S. Szmigiero help 648bb8b93b5SMaciej S. Szmigiero Use the zsmalloc allocator as the default allocator. 649bb8b93b5SMaciej S. Szmigieroendchoice 650bb8b93b5SMaciej S. Szmigiero 651bb8b93b5SMaciej S. Szmigieroconfig ZSWAP_ZPOOL_DEFAULT 652bb8b93b5SMaciej S. Szmigiero string 653bb8b93b5SMaciej S. Szmigiero depends on ZSWAP 654bb8b93b5SMaciej S. Szmigiero default "zbud" if ZSWAP_ZPOOL_DEFAULT_ZBUD 655bb8b93b5SMaciej S. Szmigiero default "z3fold" if ZSWAP_ZPOOL_DEFAULT_Z3FOLD 656bb8b93b5SMaciej S. Szmigiero default "zsmalloc" if ZSWAP_ZPOOL_DEFAULT_ZSMALLOC 657bb8b93b5SMaciej S. Szmigiero default "" 658bb8b93b5SMaciej S. Szmigiero 659bb8b93b5SMaciej S. Szmigieroconfig ZSWAP_DEFAULT_ON 660bb8b93b5SMaciej S. Szmigiero bool "Enable the compressed cache for swap pages by default" 661bb8b93b5SMaciej S. Szmigiero depends on ZSWAP 662bb8b93b5SMaciej S. Szmigiero help 663bb8b93b5SMaciej S. Szmigiero If selected, the compressed cache for swap pages will be enabled 664bb8b93b5SMaciej S. Szmigiero at boot, otherwise it will be disabled. 665bb8b93b5SMaciej S. Szmigiero 666bb8b93b5SMaciej S. Szmigiero The selection made here can be overridden by using the kernel 667bb8b93b5SMaciej S. Szmigiero command line 'zswap.enabled=' option. 668bb8b93b5SMaciej S. Szmigiero 669af8d417aSDan Streetmanconfig ZPOOL 670af8d417aSDan Streetman tristate "Common API for compressed memory storage" 6710f8975ecSPavel Emelyanov help 672af8d417aSDan Streetman Compressed memory storage API. This allows using either zbud or 673af8d417aSDan Streetman zsmalloc. 6740f8975ecSPavel Emelyanov 675af8d417aSDan Streetmanconfig ZBUD 6769a001fc1SVitaly Wool tristate "Low (Up to 2x) density storage for compressed pages" 6772a03085cSMiaohe Lin depends on ZPOOL 678af8d417aSDan Streetman help 679af8d417aSDan Streetman A special purpose allocator for storing compressed pages. 680af8d417aSDan Streetman It is designed to store up to two compressed pages per physical 681af8d417aSDan Streetman page. While this design limits storage density, it has simple and 682af8d417aSDan Streetman deterministic reclaim properties that make it preferable to a higher 683af8d417aSDan Streetman density approach when reclaim will be used. 684bcf1647dSMinchan Kim 6859a001fc1SVitaly Woolconfig Z3FOLD 6869a001fc1SVitaly Wool tristate "Up to 3x density storage for compressed pages" 6879a001fc1SVitaly Wool depends on ZPOOL 6889a001fc1SVitaly Wool help 6899a001fc1SVitaly Wool A special purpose allocator for storing compressed pages. 6909a001fc1SVitaly Wool It is designed to store up to three compressed pages per physical 6919a001fc1SVitaly Wool page. It is a ZBUD derivative so the simplicity and determinism are 6929a001fc1SVitaly Wool still there. 6939a001fc1SVitaly Wool 694bcf1647dSMinchan Kimconfig ZSMALLOC 695d867f203SMinchan Kim tristate "Memory allocator for compressed pages" 696bcf1647dSMinchan Kim depends on MMU 697bcf1647dSMinchan Kim help 698bcf1647dSMinchan Kim zsmalloc is a slab-based memory allocator designed to store 699bcf1647dSMinchan Kim compressed RAM pages. zsmalloc uses virtual memory mapping 700bcf1647dSMinchan Kim in order to reduce fragmentation. However, this results in a 701bcf1647dSMinchan Kim non-standard allocator interface where a handle, not a pointer, is 702bcf1647dSMinchan Kim returned by an alloc(). This handle must be mapped in order to 703bcf1647dSMinchan Kim access the allocated space. 704bcf1647dSMinchan Kim 7050f050d99SGanesh Mahendranconfig ZSMALLOC_STAT 7060f050d99SGanesh Mahendran bool "Export zsmalloc statistics" 7070f050d99SGanesh Mahendran depends on ZSMALLOC 7080f050d99SGanesh Mahendran select DEBUG_FS 7090f050d99SGanesh Mahendran help 7100f050d99SGanesh Mahendran This option enables code in the zsmalloc to collect various 71101ab1edeSColin Ian King statistics about what's happening in zsmalloc and exports that 7120f050d99SGanesh Mahendran information to userspace via debugfs. 7130f050d99SGanesh Mahendran If unsure, say N. 7140f050d99SGanesh Mahendran 7159e5c33d7SMark Salterconfig GENERIC_EARLY_IOREMAP 7169e5c33d7SMark Salter bool 717042d27acSHelge Deller 71822ee3ea5SHelge Dellerconfig STACK_MAX_DEFAULT_SIZE_MB 71922ee3ea5SHelge Deller int "Default maximum user stack size for 32-bit processes (MB)" 72022ee3ea5SHelge Deller default 100 721042d27acSHelge Deller range 8 2048 722042d27acSHelge Deller depends on STACK_GROWSUP && (!64BIT || COMPAT) 723042d27acSHelge Deller help 724042d27acSHelge Deller This is the maximum stack size in Megabytes in the VM layout of 32-bit 725042d27acSHelge Deller user processes when the stack grows upwards (currently only on parisc 72622ee3ea5SHelge Deller arch) when the RLIMIT_STACK hard limit is unlimited. 727042d27acSHelge Deller 72822ee3ea5SHelge Deller A sane initial value is 100 MB. 7293a80a7faSMel Gorman 7303a80a7faSMel Gormanconfig DEFERRED_STRUCT_PAGE_INIT 7311ce22103SVlastimil Babka bool "Defer initialisation of struct pages to kthreads" 732d39f8fb4SMike Rapoport depends on SPARSEMEM 733ab1e8d89SPavel Tatashin depends on !NEED_PER_CPU_KM 734889c695dSPasha Tatashin depends on 64BIT 735e4443149SDaniel Jordan select PADATA 7363a80a7faSMel Gorman help 7373a80a7faSMel Gorman Ordinarily all struct pages are initialised during early boot in a 7383a80a7faSMel Gorman single thread. On very large machines this can take a considerable 7393a80a7faSMel Gorman amount of time. If this option is set, large machines will bring up 740e4443149SDaniel Jordan a subset of memmap at boot and then initialise the rest in parallel. 741e4443149SDaniel Jordan This has a potential performance impact on tasks running early in the 7421ce22103SVlastimil Babka lifetime of the system until these kthreads finish the 7431ce22103SVlastimil Babka initialisation. 744033fbae9SDan Williams 74533c3fc71SVladimir Davydovconfig IDLE_PAGE_TRACKING 74633c3fc71SVladimir Davydov bool "Enable idle page tracking" 74733c3fc71SVladimir Davydov depends on SYSFS && MMU 74833c3fc71SVladimir Davydov select PAGE_EXTENSION if !64BIT 74933c3fc71SVladimir Davydov help 75033c3fc71SVladimir Davydov This feature allows to estimate the amount of user pages that have 75133c3fc71SVladimir Davydov not been touched during a given period of time. This information can 75233c3fc71SVladimir Davydov be useful to tune memory cgroup limits and/or for job placement 75333c3fc71SVladimir Davydov within a compute cluster. 75433c3fc71SVladimir Davydov 7551ad1335dSMike Rapoport See Documentation/admin-guide/mm/idle_page_tracking.rst for 7561ad1335dSMike Rapoport more details. 75733c3fc71SVladimir Davydov 758c2280be8SAnshuman Khandualconfig ARCH_HAS_CACHE_LINE_SIZE 759c2280be8SAnshuman Khandual bool 760c2280be8SAnshuman Khandual 76117596731SRobin Murphyconfig ARCH_HAS_PTE_DEVMAP 76265f7d049SOliver O'Halloran bool 76365f7d049SOliver O'Halloran 764*63703f37SKefeng Wangconfig ARCH_HAS_ZONE_DMA_SET 765*63703f37SKefeng Wang bool 766*63703f37SKefeng Wang 767*63703f37SKefeng Wangconfig ZONE_DMA 768*63703f37SKefeng Wang bool "Support DMA zone" if ARCH_HAS_ZONE_DMA_SET 769*63703f37SKefeng Wang default y if ARM64 || X86 770*63703f37SKefeng Wang 771*63703f37SKefeng Wangconfig ZONE_DMA32 772*63703f37SKefeng Wang bool "Support DMA32 zone" if ARCH_HAS_ZONE_DMA_SET 773*63703f37SKefeng Wang depends on !X86_32 774*63703f37SKefeng Wang default y if ARM64 775*63703f37SKefeng Wang 776033fbae9SDan Williamsconfig ZONE_DEVICE 7775042db43SJérôme Glisse bool "Device memory (pmem, HMM, etc...) hotplug support" 778033fbae9SDan Williams depends on MEMORY_HOTPLUG 779033fbae9SDan Williams depends on MEMORY_HOTREMOVE 78099490f16SDan Williams depends on SPARSEMEM_VMEMMAP 78117596731SRobin Murphy depends on ARCH_HAS_PTE_DEVMAP 7823a08cd52SMatthew Wilcox select XARRAY_MULTI 783033fbae9SDan Williams 784033fbae9SDan Williams help 785033fbae9SDan Williams Device memory hotplug support allows for establishing pmem, 786033fbae9SDan Williams or other device driver discovered memory regions, in the 787033fbae9SDan Williams memmap. This allows pfn_to_page() lookups of otherwise 788033fbae9SDan Williams "device-physical" addresses which is needed for using a DAX 789033fbae9SDan Williams mapping in an O_DIRECT operation, among other things. 790033fbae9SDan Williams 791033fbae9SDan Williams If FS_DAX is enabled, then say Y. 79206a660adSLinus Torvalds 793e7638488SDan Williamsconfig DEV_PAGEMAP_OPS 794e7638488SDan Williams bool 795e7638488SDan Williams 7969c240a7bSChristoph Hellwig# 7979c240a7bSChristoph Hellwig# Helpers to mirror range of the CPU page tables of a process into device page 7989c240a7bSChristoph Hellwig# tables. 7999c240a7bSChristoph Hellwig# 800c0b12405SJérôme Glisseconfig HMM_MIRROR 8019c240a7bSChristoph Hellwig bool 802f442c283SChristoph Hellwig depends on MMU 803c0b12405SJérôme Glisse 8045042db43SJérôme Glisseconfig DEVICE_PRIVATE 8055042db43SJérôme Glisse bool "Unaddressable device memory (GPU memory, ...)" 8067328d9ccSChristoph Hellwig depends on ZONE_DEVICE 807e7638488SDan Williams select DEV_PAGEMAP_OPS 8085042db43SJérôme Glisse 8095042db43SJérôme Glisse help 8105042db43SJérôme Glisse Allows creation of struct pages to represent unaddressable device 8115042db43SJérôme Glisse memory; i.e., memory that is only accessible from the device (or 8125042db43SJérôme Glisse group of devices). You likely also want to select HMM_MIRROR. 8135042db43SJérôme Glisse 8143e9a9e25SChristoph Hellwigconfig VMAP_PFN 8153e9a9e25SChristoph Hellwig bool 8163e9a9e25SChristoph Hellwig 81763c17fb8SDave Hansenconfig ARCH_USES_HIGH_VMA_FLAGS 81863c17fb8SDave Hansen bool 81966d37570SDave Hansenconfig ARCH_HAS_PKEYS 82066d37570SDave Hansen bool 82130a5b536SDennis Zhou 82230a5b536SDennis Zhouconfig PERCPU_STATS 82330a5b536SDennis Zhou bool "Collect percpu memory statistics" 82430a5b536SDennis Zhou help 82530a5b536SDennis Zhou This feature collects and exposes statistics via debugfs. The 82630a5b536SDennis Zhou information includes global and per chunk statistics, which can 82730a5b536SDennis Zhou be used to help understand percpu memory usage. 82864c349f4SKirill A. Shutemov 8299c84f229SJohn Hubbardconfig GUP_TEST 8309c84f229SJohn Hubbard bool "Enable infrastructure for get_user_pages()-related unit tests" 831d0de8241SBarry Song depends on DEBUG_FS 83264c349f4SKirill A. Shutemov help 8339c84f229SJohn Hubbard Provides /sys/kernel/debug/gup_test, which in turn provides a way 8349c84f229SJohn Hubbard to make ioctl calls that can launch kernel-based unit tests for 8359c84f229SJohn Hubbard the get_user_pages*() and pin_user_pages*() family of API calls. 83664c349f4SKirill A. Shutemov 8379c84f229SJohn Hubbard These tests include benchmark testing of the _fast variants of 8389c84f229SJohn Hubbard get_user_pages*() and pin_user_pages*(), as well as smoke tests of 8399c84f229SJohn Hubbard the non-_fast variants. 8409c84f229SJohn Hubbard 841f4f9bda4SJohn Hubbard There is also a sub-test that allows running dump_page() on any 842f4f9bda4SJohn Hubbard of up to eight pages (selected by command line args) within the 843f4f9bda4SJohn Hubbard range of user-space addresses. These pages are either pinned via 844f4f9bda4SJohn Hubbard pin_user_pages*(), or pinned via get_user_pages*(), as specified 845f4f9bda4SJohn Hubbard by other command line arguments. 846f4f9bda4SJohn Hubbard 8479c84f229SJohn Hubbard See tools/testing/selftests/vm/gup_test.c 8483010a5eaSLaurent Dufour 849d0de8241SBarry Songcomment "GUP_TEST needs to have DEBUG_FS enabled" 850d0de8241SBarry Song depends on !GUP_TEST && !DEBUG_FS 8513010a5eaSLaurent Dufour 85239656e83SChristoph Hellwigconfig GUP_GET_PTE_LOW_HIGH 85339656e83SChristoph Hellwig bool 85439656e83SChristoph Hellwig 85599cb0dbdSSong Liuconfig READ_ONLY_THP_FOR_FS 85699cb0dbdSSong Liu bool "Read-only THP for filesystems (EXPERIMENTAL)" 857396bcc52SMatthew Wilcox (Oracle) depends on TRANSPARENT_HUGEPAGE && SHMEM 85899cb0dbdSSong Liu 85999cb0dbdSSong Liu help 86099cb0dbdSSong Liu Allow khugepaged to put read-only file-backed pages in THP. 86199cb0dbdSSong Liu 86299cb0dbdSSong Liu This is marked experimental because it is a new feature. Write 86399cb0dbdSSong Liu support of file THPs will be developed in the next few release 86499cb0dbdSSong Liu cycles. 86599cb0dbdSSong Liu 8663010a5eaSLaurent Dufourconfig ARCH_HAS_PTE_SPECIAL 8673010a5eaSLaurent Dufour bool 86859e0b520SChristoph Hellwig 869cbd34da7SChristoph Hellwig# 870cbd34da7SChristoph Hellwig# Some architectures require a special hugepage directory format that is 871cbd34da7SChristoph Hellwig# required to support multiple hugepage sizes. For example a4fe3ce76 872cbd34da7SChristoph Hellwig# "powerpc/mm: Allow more flexible layouts for hugepage pagetables" 873cbd34da7SChristoph Hellwig# introduced it on powerpc. This allows for a more flexible hugepage 874cbd34da7SChristoph Hellwig# pagetable layouts. 875cbd34da7SChristoph Hellwig# 876cbd34da7SChristoph Hellwigconfig ARCH_HAS_HUGEPD 877cbd34da7SChristoph Hellwig bool 878cbd34da7SChristoph Hellwig 879c5acad84SThomas Hellstromconfig MAPPING_DIRTY_HELPERS 880c5acad84SThomas Hellstrom bool 881c5acad84SThomas Hellstrom 882298fa1adSThomas Gleixnerconfig KMAP_LOCAL 883298fa1adSThomas Gleixner bool 884298fa1adSThomas Gleixner 8851fbaf8fcSChristoph Hellwig# struct io_mapping based helper. Selected by drivers that need them 8861fbaf8fcSChristoph Hellwigconfig IO_MAPPING 8871fbaf8fcSChristoph Hellwig bool 88859e0b520SChristoph Hellwigendmenu 889