159e0b520SChristoph Hellwig 259e0b520SChristoph Hellwigmenu "Memory Management options" 359e0b520SChristoph Hellwig 4e1785e85SDave Hansenconfig SELECT_MEMORY_MODEL 5e1785e85SDave Hansen def_bool y 6a8826eebSKees Cook depends on ARCH_SELECT_MEMORY_MODEL 7e1785e85SDave Hansen 83a9da765SDave Hansenchoice 93a9da765SDave Hansen prompt "Memory model" 10e1785e85SDave Hansen depends on SELECT_MEMORY_MODEL 11e1785e85SDave Hansen default DISCONTIGMEM_MANUAL if ARCH_DISCONTIGMEM_DEFAULT 12d41dee36SAndy Whitcroft default SPARSEMEM_MANUAL if ARCH_SPARSEMEM_DEFAULT 13e1785e85SDave Hansen default FLATMEM_MANUAL 143a9da765SDave Hansen 15e1785e85SDave Hansenconfig FLATMEM_MANUAL 163a9da765SDave Hansen bool "Flat Memory" 17c898ec16SAnton Blanchard depends on !(ARCH_DISCONTIGMEM_ENABLE || ARCH_SPARSEMEM_ENABLE) || ARCH_FLATMEM_ENABLE 183a9da765SDave Hansen help 193a9da765SDave Hansen This option allows you to change some of the ways that 203a9da765SDave Hansen Linux manages its memory internally. Most users will 213a9da765SDave Hansen only have one option here: FLATMEM. This is normal 223a9da765SDave Hansen and a correct option. 233a9da765SDave Hansen 24d41dee36SAndy Whitcroft Some users of more advanced features like NUMA and 25d41dee36SAndy Whitcroft memory hotplug may have different options here. 2618f65332SGeert Uytterhoeven DISCONTIGMEM is a more mature, better tested system, 27d41dee36SAndy Whitcroft but is incompatible with memory hotplug and may suffer 28d41dee36SAndy Whitcroft decreased performance over SPARSEMEM. If unsure between 29d41dee36SAndy Whitcroft "Sparse Memory" and "Discontiguous Memory", choose 30d41dee36SAndy Whitcroft "Discontiguous Memory". 31d41dee36SAndy Whitcroft 32d41dee36SAndy Whitcroft If unsure, choose this option (Flat Memory) over any other. 333a9da765SDave Hansen 34e1785e85SDave Hansenconfig DISCONTIGMEM_MANUAL 35f3519f91SDave Hansen bool "Discontiguous Memory" 363a9da765SDave Hansen depends on ARCH_DISCONTIGMEM_ENABLE 373a9da765SDave Hansen help 38785dcd44SDave Hansen This option provides enhanced support for discontiguous 39785dcd44SDave Hansen memory systems, over FLATMEM. These systems have holes 40785dcd44SDave Hansen in their physical address spaces, and this option provides 41785dcd44SDave Hansen more efficient handling of these holes. However, the vast 42785dcd44SDave Hansen majority of hardware has quite flat address spaces, and 43ad3d0a38SPhilipp Marek can have degraded performance from the extra overhead that 44785dcd44SDave Hansen this option imposes. 45785dcd44SDave Hansen 46785dcd44SDave Hansen Many NUMA configurations will have this as the only option. 47785dcd44SDave Hansen 483a9da765SDave Hansen If unsure, choose "Flat Memory" over this option. 493a9da765SDave Hansen 50d41dee36SAndy Whitcroftconfig SPARSEMEM_MANUAL 51d41dee36SAndy Whitcroft bool "Sparse Memory" 52d41dee36SAndy Whitcroft depends on ARCH_SPARSEMEM_ENABLE 53d41dee36SAndy Whitcroft help 54d41dee36SAndy Whitcroft This will be the only option for some systems, including 55d41dee36SAndy Whitcroft memory hotplug systems. This is normal. 56d41dee36SAndy Whitcroft 57d41dee36SAndy Whitcroft For many other systems, this will be an alternative to 58f3519f91SDave Hansen "Discontiguous Memory". This option provides some potential 59d41dee36SAndy Whitcroft performance benefits, along with decreased code complexity, 60d41dee36SAndy Whitcroft but it is newer, and more experimental. 61d41dee36SAndy Whitcroft 62d41dee36SAndy Whitcroft If unsure, choose "Discontiguous Memory" or "Flat Memory" 63d41dee36SAndy Whitcroft over this option. 64d41dee36SAndy Whitcroft 653a9da765SDave Hansenendchoice 663a9da765SDave Hansen 67e1785e85SDave Hansenconfig DISCONTIGMEM 68e1785e85SDave Hansen def_bool y 69e1785e85SDave Hansen depends on (!SELECT_MEMORY_MODEL && ARCH_DISCONTIGMEM_ENABLE) || DISCONTIGMEM_MANUAL 70e1785e85SDave Hansen 71d41dee36SAndy Whitcroftconfig SPARSEMEM 72d41dee36SAndy Whitcroft def_bool y 731a83e175SRussell King depends on (!SELECT_MEMORY_MODEL && ARCH_SPARSEMEM_ENABLE) || SPARSEMEM_MANUAL 74d41dee36SAndy Whitcroft 75e1785e85SDave Hansenconfig FLATMEM 76e1785e85SDave Hansen def_bool y 77d41dee36SAndy Whitcroft depends on (!DISCONTIGMEM && !SPARSEMEM) || FLATMEM_MANUAL 78d41dee36SAndy Whitcroft 79d41dee36SAndy Whitcroftconfig FLAT_NODE_MEM_MAP 80d41dee36SAndy Whitcroft def_bool y 81d41dee36SAndy Whitcroft depends on !SPARSEMEM 82e1785e85SDave Hansen 8393b7504eSDave Hansen# 8493b7504eSDave Hansen# Both the NUMA code and DISCONTIGMEM use arrays of pg_data_t's 8593b7504eSDave Hansen# to represent different areas of memory. This variable allows 8693b7504eSDave Hansen# those dependencies to exist individually. 8793b7504eSDave Hansen# 8893b7504eSDave Hansenconfig NEED_MULTIPLE_NODES 8993b7504eSDave Hansen def_bool y 9093b7504eSDave Hansen depends on DISCONTIGMEM || NUMA 91af705362SAndy Whitcroft 92af705362SAndy Whitcroftconfig HAVE_MEMORY_PRESENT 93af705362SAndy Whitcroft def_bool y 94d41dee36SAndy Whitcroft depends on ARCH_HAVE_MEMORY_PRESENT || SPARSEMEM 95802f192eSBob Picco 96802f192eSBob Picco# 973e347261SBob Picco# SPARSEMEM_EXTREME (which is the default) does some bootmem 983e347261SBob Picco# allocations when memory_present() is called. If this cannot 993e347261SBob Picco# be done on your architecture, select this option. However, 1003e347261SBob Picco# statically allocating the mem_section[] array can potentially 1013e347261SBob Picco# consume vast quantities of .bss, so be careful. 1023e347261SBob Picco# 1033e347261SBob Picco# This option will also potentially produce smaller runtime code 1043e347261SBob Picco# with gcc 3.4 and later. 1053e347261SBob Picco# 1063e347261SBob Piccoconfig SPARSEMEM_STATIC 1079ba16087SJan Beulich bool 1083e347261SBob Picco 1093e347261SBob Picco# 11044c09201SMatt LaPlante# Architecture platforms which require a two level mem_section in SPARSEMEM 111802f192eSBob Picco# must select this option. This is usually for architecture platforms with 112802f192eSBob Picco# an extremely sparse physical address space. 113802f192eSBob Picco# 1143e347261SBob Piccoconfig SPARSEMEM_EXTREME 1153e347261SBob Picco def_bool y 1163e347261SBob Picco depends on SPARSEMEM && !SPARSEMEM_STATIC 1174c21e2f2SHugh Dickins 11829c71111SAndy Whitcroftconfig SPARSEMEM_VMEMMAP_ENABLE 1199ba16087SJan Beulich bool 12029c71111SAndy Whitcroft 12129c71111SAndy Whitcroftconfig SPARSEMEM_VMEMMAP 122a5ee6daaSGeoff Levand bool "Sparse Memory virtual memmap" 123a5ee6daaSGeoff Levand depends on SPARSEMEM && SPARSEMEM_VMEMMAP_ENABLE 124a5ee6daaSGeoff Levand default y 125a5ee6daaSGeoff Levand help 126a5ee6daaSGeoff Levand SPARSEMEM_VMEMMAP uses a virtually mapped memmap to optimise 127a5ee6daaSGeoff Levand pfn_to_page and page_to_pfn operations. This is the most 128a5ee6daaSGeoff Levand efficient option when sufficient kernel resources are available. 12929c71111SAndy Whitcroft 1307c0caeb8STejun Heoconfig HAVE_MEMBLOCK_NODE_MAP 1316341e62bSChristoph Jaeger bool 1327c0caeb8STejun Heo 13370210ed9SPhilipp Hachtmannconfig HAVE_MEMBLOCK_PHYS_MAP 1346341e62bSChristoph Jaeger bool 13570210ed9SPhilipp Hachtmann 136e585513bSKirill A. Shutemovconfig HAVE_GENERIC_GUP 1376341e62bSChristoph Jaeger bool 1382667f50eSSteve Capper 139c378ddd5STejun Heoconfig ARCH_DISCARD_MEMBLOCK 1406341e62bSChristoph Jaeger bool 141c378ddd5STejun Heo 142ee6f509cSMinchan Kimconfig MEMORY_ISOLATION 1436341e62bSChristoph Jaeger bool 144ee6f509cSMinchan Kim 14546723bfaSYasuaki Ishimatsu# 14646723bfaSYasuaki Ishimatsu# Only be set on architectures that have completely implemented memory hotplug 14746723bfaSYasuaki Ishimatsu# feature. If you are not sure, don't touch it. 14846723bfaSYasuaki Ishimatsu# 14946723bfaSYasuaki Ishimatsuconfig HAVE_BOOTMEM_INFO_NODE 15046723bfaSYasuaki Ishimatsu def_bool n 15146723bfaSYasuaki Ishimatsu 1523947be19SDave Hansen# eventually, we can have this option just 'select SPARSEMEM' 1533947be19SDave Hansenconfig MEMORY_HOTPLUG 1543947be19SDave Hansen bool "Allow for memory hot-add" 155ec69acbbSKeith Mannthey depends on SPARSEMEM || X86_64_ACPI_NUMA 15640b31360SStephen Rothwell depends on ARCH_ENABLE_MEMORY_HOTPLUG 1573947be19SDave Hansen 158ec69acbbSKeith Manntheyconfig MEMORY_HOTPLUG_SPARSE 159ec69acbbSKeith Mannthey def_bool y 160ec69acbbSKeith Mannthey depends on SPARSEMEM && MEMORY_HOTPLUG 161ec69acbbSKeith Mannthey 1628604d9e5SVitaly Kuznetsovconfig MEMORY_HOTPLUG_DEFAULT_ONLINE 1638604d9e5SVitaly Kuznetsov bool "Online the newly added memory blocks by default" 1648604d9e5SVitaly Kuznetsov default n 1658604d9e5SVitaly Kuznetsov depends on MEMORY_HOTPLUG 1668604d9e5SVitaly Kuznetsov help 1678604d9e5SVitaly Kuznetsov This option sets the default policy setting for memory hotplug 1688604d9e5SVitaly Kuznetsov onlining policy (/sys/devices/system/memory/auto_online_blocks) which 1698604d9e5SVitaly Kuznetsov determines what happens to newly added memory regions. Policy setting 1708604d9e5SVitaly Kuznetsov can always be changed at runtime. 1718604d9e5SVitaly Kuznetsov See Documentation/memory-hotplug.txt for more information. 1728604d9e5SVitaly Kuznetsov 1738604d9e5SVitaly Kuznetsov Say Y here if you want all hot-plugged memory blocks to appear in 1748604d9e5SVitaly Kuznetsov 'online' state by default. 1758604d9e5SVitaly Kuznetsov Say N here if you want the default policy to keep all hot-plugged 1768604d9e5SVitaly Kuznetsov memory blocks in 'offline' state. 1778604d9e5SVitaly Kuznetsov 1780c0e6195SKAMEZAWA Hiroyukiconfig MEMORY_HOTREMOVE 1790c0e6195SKAMEZAWA Hiroyuki bool "Allow for memory hot remove" 18046723bfaSYasuaki Ishimatsu select MEMORY_ISOLATION 181f7e3334aSNathan Fontenot select HAVE_BOOTMEM_INFO_NODE if (X86_64 || PPC64) 1820c0e6195SKAMEZAWA Hiroyuki depends on MEMORY_HOTPLUG && ARCH_ENABLE_MEMORY_HOTREMOVE 1830c0e6195SKAMEZAWA Hiroyuki depends on MIGRATION 1840c0e6195SKAMEZAWA Hiroyuki 1854c21e2f2SHugh Dickins# Heavily threaded applications may benefit from splitting the mm-wide 1864c21e2f2SHugh Dickins# page_table_lock, so that faults on different parts of the user address 1874c21e2f2SHugh Dickins# space can be handled with less contention: split it at this NR_CPUS. 1884c21e2f2SHugh Dickins# Default to 4 for wider testing, though 8 might be more appropriate. 1894c21e2f2SHugh Dickins# ARM's adjust_pte (unused if VIPT) depends on mm-wide page_table_lock. 1907b6ac9dfSHugh Dickins# PA-RISC 7xxx's spinlock_t would enlarge struct page from 32 to 44 bytes. 191a70caa8bSHugh Dickins# DEBUG_SPINLOCK and DEBUG_LOCK_ALLOC spinlock_t also enlarge struct page. 1924c21e2f2SHugh Dickins# 1934c21e2f2SHugh Dickinsconfig SPLIT_PTLOCK_CPUS 1944c21e2f2SHugh Dickins int 1959164550eSKirill A. Shutemov default "999999" if !MMU 196a70caa8bSHugh Dickins default "999999" if ARM && !CPU_CACHE_VIPT 197a70caa8bSHugh Dickins default "999999" if PARISC && !PA20 1984c21e2f2SHugh Dickins default "4" 1997cbe34cfSChristoph Lameter 200e009bb30SKirill A. Shutemovconfig ARCH_ENABLE_SPLIT_PMD_PTLOCK 2016341e62bSChristoph Jaeger bool 202e009bb30SKirill A. Shutemov 2037cbe34cfSChristoph Lameter# 20409316c09SKonstantin Khlebnikov# support for memory balloon 20509316c09SKonstantin Khlebnikovconfig MEMORY_BALLOON 2066341e62bSChristoph Jaeger bool 20709316c09SKonstantin Khlebnikov 20809316c09SKonstantin Khlebnikov# 20918468d93SRafael Aquini# support for memory balloon compaction 21018468d93SRafael Aquiniconfig BALLOON_COMPACTION 21118468d93SRafael Aquini bool "Allow for balloon memory compaction/migration" 21218468d93SRafael Aquini def_bool y 21309316c09SKonstantin Khlebnikov depends on COMPACTION && MEMORY_BALLOON 21418468d93SRafael Aquini help 21518468d93SRafael Aquini Memory fragmentation introduced by ballooning might reduce 21618468d93SRafael Aquini significantly the number of 2MB contiguous memory blocks that can be 21718468d93SRafael Aquini used within a guest, thus imposing performance penalties associated 21818468d93SRafael Aquini with the reduced number of transparent huge pages that could be used 21918468d93SRafael Aquini by the guest workload. Allowing the compaction & migration for memory 22018468d93SRafael Aquini pages enlisted as being part of memory balloon devices avoids the 22118468d93SRafael Aquini scenario aforementioned and helps improving memory defragmentation. 22218468d93SRafael Aquini 22318468d93SRafael Aquini# 224e9e96b39SMel Gorman# support for memory compaction 225e9e96b39SMel Gormanconfig COMPACTION 226e9e96b39SMel Gorman bool "Allow for memory compaction" 22705106e6aSRik van Riel def_bool y 228e9e96b39SMel Gorman select MIGRATION 22933a93877SAndrea Arcangeli depends on MMU 230e9e96b39SMel Gorman help 231b32eaf71SMichal Hocko Compaction is the only memory management component to form 232b32eaf71SMichal Hocko high order (larger physically contiguous) memory blocks 233b32eaf71SMichal Hocko reliably. The page allocator relies on compaction heavily and 234b32eaf71SMichal Hocko the lack of the feature can lead to unexpected OOM killer 235b32eaf71SMichal Hocko invocations for high order memory requests. You shouldn't 236b32eaf71SMichal Hocko disable this option unless there really is a strong reason for 237b32eaf71SMichal Hocko it and then we would be really interested to hear about that at 238b32eaf71SMichal Hocko linux-mm@kvack.org. 239e9e96b39SMel Gorman 240e9e96b39SMel Gorman# 2417cbe34cfSChristoph Lameter# support for page migration 2427cbe34cfSChristoph Lameter# 2437cbe34cfSChristoph Lameterconfig MIGRATION 244b20a3503SChristoph Lameter bool "Page migration" 2456c5240aeSChristoph Lameter def_bool y 246de32a817SChen Gang depends on (NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION || CMA) && MMU 247b20a3503SChristoph Lameter help 248b20a3503SChristoph Lameter Allows the migration of the physical location of pages of processes 249e9e96b39SMel Gorman while the virtual addresses are not changed. This is useful in 250e9e96b39SMel Gorman two situations. The first is on NUMA systems to put pages nearer 251e9e96b39SMel Gorman to the processors accessing. The second is when allocating huge 252e9e96b39SMel Gorman pages as migration can relocate pages to satisfy a huge page 253e9e96b39SMel Gorman allocation instead of reclaiming. 2546550e07fSGreg Kroah-Hartman 255c177c81eSNaoya Horiguchiconfig ARCH_ENABLE_HUGEPAGE_MIGRATION 2566341e62bSChristoph Jaeger bool 257c177c81eSNaoya Horiguchi 2589c670ea3SNaoya Horiguchiconfig ARCH_ENABLE_THP_MIGRATION 2599c670ea3SNaoya Horiguchi bool 2609c670ea3SNaoya Horiguchi 261600715dcSJeremy Fitzhardingeconfig PHYS_ADDR_T_64BIT 262d4a451d5SChristoph Hellwig def_bool 64BIT 263600715dcSJeremy Fitzhardinge 2642a7326b5SChristoph Lameterconfig BOUNCE 2659ca24e2eSVinayak Menon bool "Enable bounce buffers" 2669ca24e2eSVinayak Menon default y 2672a7326b5SChristoph Lameter depends on BLOCK && MMU && (ZONE_DMA || HIGHMEM) 2689ca24e2eSVinayak Menon help 2699ca24e2eSVinayak Menon Enable bounce buffers for devices that cannot access 2709ca24e2eSVinayak Menon the full range of memory available to the CPU. Enabled 2719ca24e2eSVinayak Menon by default when ZONE_DMA or HIGHMEM is selected, but you 2729ca24e2eSVinayak Menon may say n to override this. 2732a7326b5SChristoph Lameter 2746225e937SChristoph Lameterconfig NR_QUICK 2756225e937SChristoph Lameter int 2766225e937SChristoph Lameter depends on QUICKLIST 2776225e937SChristoph Lameter default "1" 278f057eac0SStephen Rothwell 279f057eac0SStephen Rothwellconfig VIRT_TO_BUS 2804febd95aSStephen Rothwell bool 2814febd95aSStephen Rothwell help 2824febd95aSStephen Rothwell An architecture should select this if it implements the 2834febd95aSStephen Rothwell deprecated interface virt_to_bus(). All new architectures 2844febd95aSStephen Rothwell should probably not select this. 2854febd95aSStephen Rothwell 286cddb8a5cSAndrea Arcangeli 287cddb8a5cSAndrea Arcangeliconfig MMU_NOTIFIER 288cddb8a5cSAndrea Arcangeli bool 28983fe27eaSPranith Kumar select SRCU 290fc4d5c29SDavid Howells 291f8af4da3SHugh Dickinsconfig KSM 292f8af4da3SHugh Dickins bool "Enable KSM for page merging" 293f8af4da3SHugh Dickins depends on MMU 294f8af4da3SHugh Dickins help 295f8af4da3SHugh Dickins Enable Kernel Samepage Merging: KSM periodically scans those areas 296f8af4da3SHugh Dickins of an application's address space that an app has advised may be 297f8af4da3SHugh Dickins mergeable. When it finds pages of identical content, it replaces 298d0f209f6SHugh Dickins the many instances by a single page with that content, so 299f8af4da3SHugh Dickins saving memory until one or another app needs to modify the content. 300f8af4da3SHugh Dickins Recommended for use with KVM, or with other duplicative applications. 301ad56b738SMike Rapoport See Documentation/vm/ksm.rst for more information: KSM is inactive 302c73602adSHugh Dickins until a program has madvised that an area is MADV_MERGEABLE, and 303c73602adSHugh Dickins root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set). 304f8af4da3SHugh Dickins 305e0a94c2aSChristoph Lameterconfig DEFAULT_MMAP_MIN_ADDR 306e0a94c2aSChristoph Lameter int "Low address space to protect from user allocation" 3076e141546SDavid Howells depends on MMU 308e0a94c2aSChristoph Lameter default 4096 309e0a94c2aSChristoph Lameter help 310e0a94c2aSChristoph Lameter This is the portion of low virtual memory which should be protected 311e0a94c2aSChristoph Lameter from userspace allocation. Keeping a user from writing to low pages 312e0a94c2aSChristoph Lameter can help reduce the impact of kernel NULL pointer bugs. 313e0a94c2aSChristoph Lameter 314e0a94c2aSChristoph Lameter For most ia64, ppc64 and x86 users with lots of address space 315e0a94c2aSChristoph Lameter a value of 65536 is reasonable and should cause no problems. 316e0a94c2aSChristoph Lameter On arm and other archs it should not be higher than 32768. 317788084abSEric Paris Programs which use vm86 functionality or have some need to map 318788084abSEric Paris this low address space will need CAP_SYS_RAWIO or disable this 319788084abSEric Paris protection by setting the value to 0. 320e0a94c2aSChristoph Lameter 321e0a94c2aSChristoph Lameter This value can be changed after boot using the 322e0a94c2aSChristoph Lameter /proc/sys/vm/mmap_min_addr tunable. 323e0a94c2aSChristoph Lameter 324d949f36fSLinus Torvaldsconfig ARCH_SUPPORTS_MEMORY_FAILURE 325d949f36fSLinus Torvalds bool 326e0a94c2aSChristoph Lameter 3276a46079cSAndi Kleenconfig MEMORY_FAILURE 3286a46079cSAndi Kleen depends on MMU 329d949f36fSLinus Torvalds depends on ARCH_SUPPORTS_MEMORY_FAILURE 3306a46079cSAndi Kleen bool "Enable recovery from hardware memory errors" 331ee6f509cSMinchan Kim select MEMORY_ISOLATION 33297f0b134SXie XiuQi select RAS 3336a46079cSAndi Kleen help 3346a46079cSAndi Kleen Enables code to recover from some memory failures on systems 3356a46079cSAndi Kleen with MCA recovery. This allows a system to continue running 3366a46079cSAndi Kleen even when some of its memory has uncorrected errors. This requires 3376a46079cSAndi Kleen special hardware support and typically ECC memory. 3386a46079cSAndi Kleen 339cae681fcSAndi Kleenconfig HWPOISON_INJECT 340413f9efbSAndi Kleen tristate "HWPoison pages injector" 34127df5068SAndi Kleen depends on MEMORY_FAILURE && DEBUG_KERNEL && PROC_FS 342478c5ffcSWu Fengguang select PROC_PAGE_MONITOR 343cae681fcSAndi Kleen 344fc4d5c29SDavid Howellsconfig NOMMU_INITIAL_TRIM_EXCESS 345fc4d5c29SDavid Howells int "Turn on mmap() excess space trimming before booting" 346fc4d5c29SDavid Howells depends on !MMU 347fc4d5c29SDavid Howells default 1 348fc4d5c29SDavid Howells help 349fc4d5c29SDavid Howells The NOMMU mmap() frequently needs to allocate large contiguous chunks 350fc4d5c29SDavid Howells of memory on which to store mappings, but it can only ask the system 351fc4d5c29SDavid Howells allocator for chunks in 2^N*PAGE_SIZE amounts - which is frequently 352fc4d5c29SDavid Howells more than it requires. To deal with this, mmap() is able to trim off 353fc4d5c29SDavid Howells the excess and return it to the allocator. 354fc4d5c29SDavid Howells 355fc4d5c29SDavid Howells If trimming is enabled, the excess is trimmed off and returned to the 356fc4d5c29SDavid Howells system allocator, which can cause extra fragmentation, particularly 357fc4d5c29SDavid Howells if there are a lot of transient processes. 358fc4d5c29SDavid Howells 359fc4d5c29SDavid Howells If trimming is disabled, the excess is kept, but not used, which for 360fc4d5c29SDavid Howells long-term mappings means that the space is wasted. 361fc4d5c29SDavid Howells 362fc4d5c29SDavid Howells Trimming can be dynamically controlled through a sysctl option 363fc4d5c29SDavid Howells (/proc/sys/vm/nr_trim_pages) which specifies the minimum number of 364fc4d5c29SDavid Howells excess pages there must be before trimming should occur, or zero if 365fc4d5c29SDavid Howells no trimming is to occur. 366fc4d5c29SDavid Howells 367fc4d5c29SDavid Howells This option specifies the initial value of this option. The default 368fc4d5c29SDavid Howells of 1 says that all excess pages should be trimmed. 369fc4d5c29SDavid Howells 370fc4d5c29SDavid Howells See Documentation/nommu-mmap.txt for more information. 371bbddff05STejun Heo 3724c76d9d1SAndrea Arcangeliconfig TRANSPARENT_HUGEPAGE 37313ece886SAndrea Arcangeli bool "Transparent Hugepage Support" 37415626062SGerald Schaefer depends on HAVE_ARCH_TRANSPARENT_HUGEPAGE 3755d689240SAndrea Arcangeli select COMPACTION 3763a08cd52SMatthew Wilcox select XARRAY_MULTI 3774c76d9d1SAndrea Arcangeli help 3784c76d9d1SAndrea Arcangeli Transparent Hugepages allows the kernel to use huge pages and 3794c76d9d1SAndrea Arcangeli huge tlb transparently to the applications whenever possible. 3804c76d9d1SAndrea Arcangeli This feature can improve computing performance to certain 3814c76d9d1SAndrea Arcangeli applications by speeding up page faults during memory 3824c76d9d1SAndrea Arcangeli allocation, by reducing the number of tlb misses and by speeding 3834c76d9d1SAndrea Arcangeli up the pagetable walking. 3844c76d9d1SAndrea Arcangeli 3854c76d9d1SAndrea Arcangeli If memory constrained on embedded, you may want to say N. 3864c76d9d1SAndrea Arcangeli 38713ece886SAndrea Arcangelichoice 38813ece886SAndrea Arcangeli prompt "Transparent Hugepage Support sysfs defaults" 38913ece886SAndrea Arcangeli depends on TRANSPARENT_HUGEPAGE 39013ece886SAndrea Arcangeli default TRANSPARENT_HUGEPAGE_ALWAYS 39113ece886SAndrea Arcangeli help 39213ece886SAndrea Arcangeli Selects the sysfs defaults for Transparent Hugepage Support. 39313ece886SAndrea Arcangeli 39413ece886SAndrea Arcangeli config TRANSPARENT_HUGEPAGE_ALWAYS 39513ece886SAndrea Arcangeli bool "always" 39613ece886SAndrea Arcangeli help 39713ece886SAndrea Arcangeli Enabling Transparent Hugepage always, can increase the 39813ece886SAndrea Arcangeli memory footprint of applications without a guaranteed 39913ece886SAndrea Arcangeli benefit but it will work automatically for all applications. 40013ece886SAndrea Arcangeli 40113ece886SAndrea Arcangeli config TRANSPARENT_HUGEPAGE_MADVISE 40213ece886SAndrea Arcangeli bool "madvise" 40313ece886SAndrea Arcangeli help 40413ece886SAndrea Arcangeli Enabling Transparent Hugepage madvise, will only provide a 40513ece886SAndrea Arcangeli performance improvement benefit to the applications using 40613ece886SAndrea Arcangeli madvise(MADV_HUGEPAGE) but it won't risk to increase the 40713ece886SAndrea Arcangeli memory footprint of applications without a guaranteed 40813ece886SAndrea Arcangeli benefit. 40913ece886SAndrea Arcangeliendchoice 41013ece886SAndrea Arcangeli 41138d8b4e6SHuang Yingconfig ARCH_WANTS_THP_SWAP 41238d8b4e6SHuang Ying def_bool n 41338d8b4e6SHuang Ying 41438d8b4e6SHuang Yingconfig THP_SWAP 41538d8b4e6SHuang Ying def_bool y 41614fef284SHuang Ying depends on TRANSPARENT_HUGEPAGE && ARCH_WANTS_THP_SWAP && SWAP 41738d8b4e6SHuang Ying help 41838d8b4e6SHuang Ying Swap transparent huge pages in one piece, without splitting. 41914fef284SHuang Ying XXX: For now, swap cluster backing transparent huge page 42014fef284SHuang Ying will be split after swapout. 42138d8b4e6SHuang Ying 42238d8b4e6SHuang Ying For selection by architectures with reasonable THP sizes. 42338d8b4e6SHuang Ying 424e496cf3dSKirill A. Shutemovconfig TRANSPARENT_HUGE_PAGECACHE 425e496cf3dSKirill A. Shutemov def_bool y 426953c66c2SAneesh Kumar K.V depends on TRANSPARENT_HUGEPAGE 427e496cf3dSKirill A. Shutemov 428e496cf3dSKirill A. Shutemov# 429bbddff05STejun Heo# UP and nommu archs use km based percpu allocator 430bbddff05STejun Heo# 431bbddff05STejun Heoconfig NEED_PER_CPU_KM 432bbddff05STejun Heo depends on !SMP 433bbddff05STejun Heo bool 434bbddff05STejun Heo default y 435077b1f83SDan Magenheimer 436077b1f83SDan Magenheimerconfig CLEANCACHE 437077b1f83SDan Magenheimer bool "Enable cleancache driver to cache clean pages if tmem is present" 438077b1f83SDan Magenheimer default n 439077b1f83SDan Magenheimer help 440077b1f83SDan Magenheimer Cleancache can be thought of as a page-granularity victim cache 441077b1f83SDan Magenheimer for clean pages that the kernel's pageframe replacement algorithm 442077b1f83SDan Magenheimer (PFRA) would like to keep around, but can't since there isn't enough 443077b1f83SDan Magenheimer memory. So when the PFRA "evicts" a page, it first attempts to use 444140a1ef2SMichael Witten cleancache code to put the data contained in that page into 445077b1f83SDan Magenheimer "transcendent memory", memory that is not directly accessible or 446077b1f83SDan Magenheimer addressable by the kernel and is of unknown and possibly 447077b1f83SDan Magenheimer time-varying size. And when a cleancache-enabled 448077b1f83SDan Magenheimer filesystem wishes to access a page in a file on disk, it first 449077b1f83SDan Magenheimer checks cleancache to see if it already contains it; if it does, 450077b1f83SDan Magenheimer the page is copied into the kernel and a disk access is avoided. 451077b1f83SDan Magenheimer When a transcendent memory driver is available (such as zcache or 452077b1f83SDan Magenheimer Xen transcendent memory), a significant I/O reduction 453077b1f83SDan Magenheimer may be achieved. When none is available, all cleancache calls 454077b1f83SDan Magenheimer are reduced to a single pointer-compare-against-NULL resulting 455077b1f83SDan Magenheimer in a negligible performance hit. 456077b1f83SDan Magenheimer 457077b1f83SDan Magenheimer If unsure, say Y to enable cleancache 45827c6aec2SDan Magenheimer 45927c6aec2SDan Magenheimerconfig FRONTSWAP 46027c6aec2SDan Magenheimer bool "Enable frontswap to cache swap pages if tmem is present" 46127c6aec2SDan Magenheimer depends on SWAP 46227c6aec2SDan Magenheimer default n 46327c6aec2SDan Magenheimer help 46427c6aec2SDan Magenheimer Frontswap is so named because it can be thought of as the opposite 46527c6aec2SDan Magenheimer of a "backing" store for a swap device. The data is stored into 46627c6aec2SDan Magenheimer "transcendent memory", memory that is not directly accessible or 46727c6aec2SDan Magenheimer addressable by the kernel and is of unknown and possibly 46827c6aec2SDan Magenheimer time-varying size. When space in transcendent memory is available, 46927c6aec2SDan Magenheimer a significant swap I/O reduction may be achieved. When none is 47027c6aec2SDan Magenheimer available, all frontswap calls are reduced to a single pointer- 47127c6aec2SDan Magenheimer compare-against-NULL resulting in a negligible performance hit 47227c6aec2SDan Magenheimer and swap data is stored as normal on the matching swap device. 47327c6aec2SDan Magenheimer 47427c6aec2SDan Magenheimer If unsure, say Y to enable frontswap. 475f825c736SAneesh Kumar K.V 476f825c736SAneesh Kumar K.Vconfig CMA 477f825c736SAneesh Kumar K.V bool "Contiguous Memory Allocator" 478*aca52c39SMike Rapoport depends on MMU 479f825c736SAneesh Kumar K.V select MIGRATION 480f825c736SAneesh Kumar K.V select MEMORY_ISOLATION 481f825c736SAneesh Kumar K.V help 482f825c736SAneesh Kumar K.V This enables the Contiguous Memory Allocator which allows other 483f825c736SAneesh Kumar K.V subsystems to allocate big physically-contiguous blocks of memory. 484f825c736SAneesh Kumar K.V CMA reserves a region of memory and allows only movable pages to 485f825c736SAneesh Kumar K.V be allocated from it. This way, the kernel can use the memory for 486f825c736SAneesh Kumar K.V pagecache and when a subsystem requests for contiguous area, the 487f825c736SAneesh Kumar K.V allocated pages are migrated away to serve the contiguous request. 488f825c736SAneesh Kumar K.V 489f825c736SAneesh Kumar K.V If unsure, say "n". 490f825c736SAneesh Kumar K.V 491f825c736SAneesh Kumar K.Vconfig CMA_DEBUG 492f825c736SAneesh Kumar K.V bool "CMA debug messages (DEVELOPMENT)" 493f825c736SAneesh Kumar K.V depends on DEBUG_KERNEL && CMA 494f825c736SAneesh Kumar K.V help 495f825c736SAneesh Kumar K.V Turns on debug messages in CMA. This produces KERN_DEBUG 496f825c736SAneesh Kumar K.V messages for every CMA call as well as various messages while 497f825c736SAneesh Kumar K.V processing calls such as dma_alloc_from_contiguous(). 498f825c736SAneesh Kumar K.V This option does not affect warning and error messages. 499bf550fc9SAlexander Graf 50028b24c1fSSasha Levinconfig CMA_DEBUGFS 50128b24c1fSSasha Levin bool "CMA debugfs interface" 50228b24c1fSSasha Levin depends on CMA && DEBUG_FS 50328b24c1fSSasha Levin help 50428b24c1fSSasha Levin Turns on the DebugFS interface for CMA. 50528b24c1fSSasha Levin 506a254129eSJoonsoo Kimconfig CMA_AREAS 507a254129eSJoonsoo Kim int "Maximum count of the CMA areas" 508a254129eSJoonsoo Kim depends on CMA 509a254129eSJoonsoo Kim default 7 510a254129eSJoonsoo Kim help 511a254129eSJoonsoo Kim CMA allows to create CMA areas for particular purpose, mainly, 512a254129eSJoonsoo Kim used as device private area. This parameter sets the maximum 513a254129eSJoonsoo Kim number of CMA area in the system. 514a254129eSJoonsoo Kim 515a254129eSJoonsoo Kim If unsure, leave the default value "7". 516a254129eSJoonsoo Kim 517af8d417aSDan Streetmanconfig MEM_SOFT_DIRTY 518af8d417aSDan Streetman bool "Track memory changes" 519af8d417aSDan Streetman depends on CHECKPOINT_RESTORE && HAVE_ARCH_SOFT_DIRTY && PROC_FS 520af8d417aSDan Streetman select PROC_PAGE_MONITOR 5214e2e2770SSeth Jennings help 522af8d417aSDan Streetman This option enables memory changes tracking by introducing a 523af8d417aSDan Streetman soft-dirty bit on pte-s. This bit it set when someone writes 524af8d417aSDan Streetman into a page just as regular dirty bit, but unlike the latter 525af8d417aSDan Streetman it can be cleared by hands. 526af8d417aSDan Streetman 5271ad1335dSMike Rapoport See Documentation/admin-guide/mm/soft-dirty.rst for more details. 5284e2e2770SSeth Jennings 5292b281117SSeth Jenningsconfig ZSWAP 5302b281117SSeth Jennings bool "Compressed cache for swap pages (EXPERIMENTAL)" 5312b281117SSeth Jennings depends on FRONTSWAP && CRYPTO=y 5322b281117SSeth Jennings select CRYPTO_LZO 53312d79d64SDan Streetman select ZPOOL 5342b281117SSeth Jennings default n 5352b281117SSeth Jennings help 5362b281117SSeth Jennings A lightweight compressed cache for swap pages. It takes 5372b281117SSeth Jennings pages that are in the process of being swapped out and attempts to 5382b281117SSeth Jennings compress them into a dynamically allocated RAM-based memory pool. 5392b281117SSeth Jennings This can result in a significant I/O reduction on swap device and, 5402b281117SSeth Jennings in the case where decompressing from RAM is faster that swap device 5412b281117SSeth Jennings reads, can also improve workload performance. 5422b281117SSeth Jennings 5432b281117SSeth Jennings This is marked experimental because it is a new feature (as of 5442b281117SSeth Jennings v3.11) that interacts heavily with memory reclaim. While these 5452b281117SSeth Jennings interactions don't cause any known issues on simple memory setups, 5462b281117SSeth Jennings they have not be fully explored on the large set of potential 5472b281117SSeth Jennings configurations and workloads that exist. 5482b281117SSeth Jennings 549af8d417aSDan Streetmanconfig ZPOOL 550af8d417aSDan Streetman tristate "Common API for compressed memory storage" 551af8d417aSDan Streetman default n 5520f8975ecSPavel Emelyanov help 553af8d417aSDan Streetman Compressed memory storage API. This allows using either zbud or 554af8d417aSDan Streetman zsmalloc. 5550f8975ecSPavel Emelyanov 556af8d417aSDan Streetmanconfig ZBUD 5579a001fc1SVitaly Wool tristate "Low (Up to 2x) density storage for compressed pages" 558af8d417aSDan Streetman default n 559af8d417aSDan Streetman help 560af8d417aSDan Streetman A special purpose allocator for storing compressed pages. 561af8d417aSDan Streetman It is designed to store up to two compressed pages per physical 562af8d417aSDan Streetman page. While this design limits storage density, it has simple and 563af8d417aSDan Streetman deterministic reclaim properties that make it preferable to a higher 564af8d417aSDan Streetman density approach when reclaim will be used. 565bcf1647dSMinchan Kim 5669a001fc1SVitaly Woolconfig Z3FOLD 5679a001fc1SVitaly Wool tristate "Up to 3x density storage for compressed pages" 5689a001fc1SVitaly Wool depends on ZPOOL 5699a001fc1SVitaly Wool default n 5709a001fc1SVitaly Wool help 5719a001fc1SVitaly Wool A special purpose allocator for storing compressed pages. 5729a001fc1SVitaly Wool It is designed to store up to three compressed pages per physical 5739a001fc1SVitaly Wool page. It is a ZBUD derivative so the simplicity and determinism are 5749a001fc1SVitaly Wool still there. 5759a001fc1SVitaly Wool 576bcf1647dSMinchan Kimconfig ZSMALLOC 577d867f203SMinchan Kim tristate "Memory allocator for compressed pages" 578bcf1647dSMinchan Kim depends on MMU 579bcf1647dSMinchan Kim default n 580bcf1647dSMinchan Kim help 581bcf1647dSMinchan Kim zsmalloc is a slab-based memory allocator designed to store 582bcf1647dSMinchan Kim compressed RAM pages. zsmalloc uses virtual memory mapping 583bcf1647dSMinchan Kim in order to reduce fragmentation. However, this results in a 584bcf1647dSMinchan Kim non-standard allocator interface where a handle, not a pointer, is 585bcf1647dSMinchan Kim returned by an alloc(). This handle must be mapped in order to 586bcf1647dSMinchan Kim access the allocated space. 587bcf1647dSMinchan Kim 588bcf1647dSMinchan Kimconfig PGTABLE_MAPPING 589bcf1647dSMinchan Kim bool "Use page table mapping to access object in zsmalloc" 590bcf1647dSMinchan Kim depends on ZSMALLOC 591bcf1647dSMinchan Kim help 592bcf1647dSMinchan Kim By default, zsmalloc uses a copy-based object mapping method to 593bcf1647dSMinchan Kim access allocations that span two pages. However, if a particular 594bcf1647dSMinchan Kim architecture (ex, ARM) performs VM mapping faster than copying, 595bcf1647dSMinchan Kim then you should select this. This causes zsmalloc to use page table 596bcf1647dSMinchan Kim mapping rather than copying for object mapping. 597bcf1647dSMinchan Kim 5982216ee85SBen Hutchings You can check speed with zsmalloc benchmark: 5992216ee85SBen Hutchings https://github.com/spartacus06/zsmapbench 6009e5c33d7SMark Salter 6010f050d99SGanesh Mahendranconfig ZSMALLOC_STAT 6020f050d99SGanesh Mahendran bool "Export zsmalloc statistics" 6030f050d99SGanesh Mahendran depends on ZSMALLOC 6040f050d99SGanesh Mahendran select DEBUG_FS 6050f050d99SGanesh Mahendran help 6060f050d99SGanesh Mahendran This option enables code in the zsmalloc to collect various 6070f050d99SGanesh Mahendran statistics about whats happening in zsmalloc and exports that 6080f050d99SGanesh Mahendran information to userspace via debugfs. 6090f050d99SGanesh Mahendran If unsure, say N. 6100f050d99SGanesh Mahendran 6119e5c33d7SMark Salterconfig GENERIC_EARLY_IOREMAP 6129e5c33d7SMark Salter bool 613042d27acSHelge Deller 614042d27acSHelge Dellerconfig MAX_STACK_SIZE_MB 615042d27acSHelge Deller int "Maximum user stack size for 32-bit processes (MB)" 616042d27acSHelge Deller default 80 617042d27acSHelge Deller range 8 2048 618042d27acSHelge Deller depends on STACK_GROWSUP && (!64BIT || COMPAT) 619042d27acSHelge Deller help 620042d27acSHelge Deller This is the maximum stack size in Megabytes in the VM layout of 32-bit 621042d27acSHelge Deller user processes when the stack grows upwards (currently only on parisc 6225f171577SJames Hogan arch). The stack will be located at the highest memory address minus 6235f171577SJames Hogan the given value, unless the RLIMIT_STACK hard limit is changed to a 6245f171577SJames Hogan smaller value in which case that is used. 625042d27acSHelge Deller 626042d27acSHelge Deller A sane initial value is 80 MB. 6273a80a7faSMel Gorman 6283a80a7faSMel Gormanconfig DEFERRED_STRUCT_PAGE_INIT 6291ce22103SVlastimil Babka bool "Defer initialisation of struct pages to kthreads" 6303a80a7faSMel Gorman default n 631d39f8fb4SMike Rapoport depends on SPARSEMEM 632ab1e8d89SPavel Tatashin depends on !NEED_PER_CPU_KM 633889c695dSPasha Tatashin depends on 64BIT 6343a80a7faSMel Gorman help 6353a80a7faSMel Gorman Ordinarily all struct pages are initialised during early boot in a 6363a80a7faSMel Gorman single thread. On very large machines this can take a considerable 6373a80a7faSMel Gorman amount of time. If this option is set, large machines will bring up 6383a80a7faSMel Gorman a subset of memmap at boot and then initialise the rest in parallel 6391ce22103SVlastimil Babka by starting one-off "pgdatinitX" kernel thread for each node X. This 6401ce22103SVlastimil Babka has a potential performance impact on processes running early in the 6411ce22103SVlastimil Babka lifetime of the system until these kthreads finish the 6421ce22103SVlastimil Babka initialisation. 643033fbae9SDan Williams 64433c3fc71SVladimir Davydovconfig IDLE_PAGE_TRACKING 64533c3fc71SVladimir Davydov bool "Enable idle page tracking" 64633c3fc71SVladimir Davydov depends on SYSFS && MMU 64733c3fc71SVladimir Davydov select PAGE_EXTENSION if !64BIT 64833c3fc71SVladimir Davydov help 64933c3fc71SVladimir Davydov This feature allows to estimate the amount of user pages that have 65033c3fc71SVladimir Davydov not been touched during a given period of time. This information can 65133c3fc71SVladimir Davydov be useful to tune memory cgroup limits and/or for job placement 65233c3fc71SVladimir Davydov within a compute cluster. 65333c3fc71SVladimir Davydov 6541ad1335dSMike Rapoport See Documentation/admin-guide/mm/idle_page_tracking.rst for 6551ad1335dSMike Rapoport more details. 65633c3fc71SVladimir Davydov 65765f7d049SOliver O'Halloran# arch_add_memory() comprehends device memory 65865f7d049SOliver O'Halloranconfig ARCH_HAS_ZONE_DEVICE 65965f7d049SOliver O'Halloran bool 66065f7d049SOliver O'Halloran 661033fbae9SDan Williamsconfig ZONE_DEVICE 6625042db43SJérôme Glisse bool "Device memory (pmem, HMM, etc...) hotplug support" 663033fbae9SDan Williams depends on MEMORY_HOTPLUG 664033fbae9SDan Williams depends on MEMORY_HOTREMOVE 66599490f16SDan Williams depends on SPARSEMEM_VMEMMAP 66665f7d049SOliver O'Halloran depends on ARCH_HAS_ZONE_DEVICE 6673a08cd52SMatthew Wilcox select XARRAY_MULTI 668033fbae9SDan Williams 669033fbae9SDan Williams help 670033fbae9SDan Williams Device memory hotplug support allows for establishing pmem, 671033fbae9SDan Williams or other device driver discovered memory regions, in the 672033fbae9SDan Williams memmap. This allows pfn_to_page() lookups of otherwise 673033fbae9SDan Williams "device-physical" addresses which is needed for using a DAX 674033fbae9SDan Williams mapping in an O_DIRECT operation, among other things. 675033fbae9SDan Williams 676033fbae9SDan Williams If FS_DAX is enabled, then say Y. 67706a660adSLinus Torvalds 678133ff0eaSJérôme Glisseconfig ARCH_HAS_HMM 679133ff0eaSJérôme Glisse bool 680133ff0eaSJérôme Glisse default y 681133ff0eaSJérôme Glisse depends on (X86_64 || PPC64) 682133ff0eaSJérôme Glisse depends on ZONE_DEVICE 683133ff0eaSJérôme Glisse depends on MMU && 64BIT 684133ff0eaSJérôme Glisse depends on MEMORY_HOTPLUG 685133ff0eaSJérôme Glisse depends on MEMORY_HOTREMOVE 686133ff0eaSJérôme Glisse depends on SPARSEMEM_VMEMMAP 687133ff0eaSJérôme Glisse 6886b368cd4SJérôme Glisseconfig MIGRATE_VMA_HELPER 6896b368cd4SJérôme Glisse bool 6906b368cd4SJérôme Glisse 691e7638488SDan Williamsconfig DEV_PAGEMAP_OPS 692e7638488SDan Williams bool 693e7638488SDan Williams 694133ff0eaSJérôme Glisseconfig HMM 695133ff0eaSJérôme Glisse bool 6966b368cd4SJérôme Glisse select MIGRATE_VMA_HELPER 697133ff0eaSJérôme Glisse 698c0b12405SJérôme Glisseconfig HMM_MIRROR 699c0b12405SJérôme Glisse bool "HMM mirror CPU page table into a device page table" 700c0b12405SJérôme Glisse depends on ARCH_HAS_HMM 701c0b12405SJérôme Glisse select MMU_NOTIFIER 702c0b12405SJérôme Glisse select HMM 703c0b12405SJérôme Glisse help 704c0b12405SJérôme Glisse Select HMM_MIRROR if you want to mirror range of the CPU page table of a 705c0b12405SJérôme Glisse process into a device page table. Here, mirror means "keep synchronized". 706c0b12405SJérôme Glisse Prerequisites: the device must provide the ability to write-protect its 707c0b12405SJérôme Glisse page tables (at PAGE_SIZE granularity), and must be able to recover from 708c0b12405SJérôme Glisse the resulting potential page faults. 709c0b12405SJérôme Glisse 7105042db43SJérôme Glisseconfig DEVICE_PRIVATE 7115042db43SJérôme Glisse bool "Unaddressable device memory (GPU memory, ...)" 7125042db43SJérôme Glisse depends on ARCH_HAS_HMM 713df6ad698SJérôme Glisse select HMM 714e7638488SDan Williams select DEV_PAGEMAP_OPS 7155042db43SJérôme Glisse 7165042db43SJérôme Glisse help 7175042db43SJérôme Glisse Allows creation of struct pages to represent unaddressable device 7185042db43SJérôme Glisse memory; i.e., memory that is only accessible from the device (or 7195042db43SJérôme Glisse group of devices). You likely also want to select HMM_MIRROR. 7205042db43SJérôme Glisse 721df6ad698SJérôme Glisseconfig DEVICE_PUBLIC 722df6ad698SJérôme Glisse bool "Addressable device memory (like GPU memory)" 723df6ad698SJérôme Glisse depends on ARCH_HAS_HMM 724df6ad698SJérôme Glisse select HMM 725e7638488SDan Williams select DEV_PAGEMAP_OPS 726df6ad698SJérôme Glisse 727df6ad698SJérôme Glisse help 728df6ad698SJérôme Glisse Allows creation of struct pages to represent addressable device 729df6ad698SJérôme Glisse memory; i.e., memory that is accessible from both the device and 730df6ad698SJérôme Glisse the CPU 731df6ad698SJérôme Glisse 7328025e5ddSJan Karaconfig FRAME_VECTOR 7338025e5ddSJan Kara bool 73463c17fb8SDave Hansen 73563c17fb8SDave Hansenconfig ARCH_USES_HIGH_VMA_FLAGS 73663c17fb8SDave Hansen bool 73766d37570SDave Hansenconfig ARCH_HAS_PKEYS 73866d37570SDave Hansen bool 73930a5b536SDennis Zhou 74030a5b536SDennis Zhouconfig PERCPU_STATS 74130a5b536SDennis Zhou bool "Collect percpu memory statistics" 74230a5b536SDennis Zhou default n 74330a5b536SDennis Zhou help 74430a5b536SDennis Zhou This feature collects and exposes statistics via debugfs. The 74530a5b536SDennis Zhou information includes global and per chunk statistics, which can 74630a5b536SDennis Zhou be used to help understand percpu memory usage. 74764c349f4SKirill A. Shutemov 74864c349f4SKirill A. Shutemovconfig GUP_BENCHMARK 74964c349f4SKirill A. Shutemov bool "Enable infrastructure for get_user_pages_fast() benchmarking" 75064c349f4SKirill A. Shutemov default n 75164c349f4SKirill A. Shutemov help 75264c349f4SKirill A. Shutemov Provides /sys/kernel/debug/gup_benchmark that helps with testing 75364c349f4SKirill A. Shutemov performance of get_user_pages_fast(). 75464c349f4SKirill A. Shutemov 75564c349f4SKirill A. Shutemov See tools/testing/selftests/vm/gup_benchmark.c 7563010a5eaSLaurent Dufour 7573010a5eaSLaurent Dufourconfig ARCH_HAS_PTE_SPECIAL 7583010a5eaSLaurent Dufour bool 75959e0b520SChristoph Hellwig 76059e0b520SChristoph Hellwigendmenu 761