1ec8f24b7SThomas Gleixner# SPDX-License-Identifier: GPL-2.0-only 259e0b520SChristoph Hellwig 359e0b520SChristoph Hellwigmenu "Memory Management options" 459e0b520SChristoph Hellwig 57b42f104SJohannes Weiner# 67b42f104SJohannes Weiner# For some reason microblaze and nios2 hard code SWAP=n. Hopefully we can 77b42f104SJohannes Weiner# add proper SWAP support to them, in which case this can be remove. 87b42f104SJohannes Weiner# 97b42f104SJohannes Weinerconfig ARCH_NO_SWAP 107b42f104SJohannes Weiner bool 117b42f104SJohannes Weiner 12b3fbd58fSJohannes Weinerconfig ZPOOL 13b3fbd58fSJohannes Weiner bool 14b3fbd58fSJohannes Weiner 15519bcb79SJohannes Weinermenuconfig SWAP 167b42f104SJohannes Weiner bool "Support for paging of anonymous memory (swap)" 177b42f104SJohannes Weiner depends on MMU && BLOCK && !ARCH_NO_SWAP 187b42f104SJohannes Weiner default y 197b42f104SJohannes Weiner help 207b42f104SJohannes Weiner This option allows you to choose whether you want to have support 217b42f104SJohannes Weiner for so called swap devices or swap files in your kernel that are 227b42f104SJohannes Weiner used to provide more virtual memory than the actual RAM present 237b42f104SJohannes Weiner in your computer. If unsure say Y. 247b42f104SJohannes Weiner 25519bcb79SJohannes Weinerconfig ZSWAP 26fcab9b44SDavid Heidelberg bool "Compressed cache for swap pages" 27b3fbd58fSJohannes Weiner depends on SWAP 28519bcb79SJohannes Weiner select FRONTSWAP 29b3fbd58fSJohannes Weiner select CRYPTO 30519bcb79SJohannes Weiner select ZPOOL 31519bcb79SJohannes Weiner help 32519bcb79SJohannes Weiner A lightweight compressed cache for swap pages. It takes 33519bcb79SJohannes Weiner pages that are in the process of being swapped out and attempts to 34519bcb79SJohannes Weiner compress them into a dynamically allocated RAM-based memory pool. 35519bcb79SJohannes Weiner This can result in a significant I/O reduction on swap device and, 361a44131dSSophia Gabriella in the case where decompressing from RAM is faster than swap device 37519bcb79SJohannes Weiner reads, can also improve workload performance. 38519bcb79SJohannes Weiner 39b3fbd58fSJohannes Weinerconfig ZSWAP_DEFAULT_ON 40b3fbd58fSJohannes Weiner bool "Enable the compressed cache for swap pages by default" 41b3fbd58fSJohannes Weiner depends on ZSWAP 42b3fbd58fSJohannes Weiner help 43b3fbd58fSJohannes Weiner If selected, the compressed cache for swap pages will be enabled 44b3fbd58fSJohannes Weiner at boot, otherwise it will be disabled. 45b3fbd58fSJohannes Weiner 46b3fbd58fSJohannes Weiner The selection made here can be overridden by using the kernel 47b3fbd58fSJohannes Weiner command line 'zswap.enabled=' option. 48b3fbd58fSJohannes Weiner 49519bcb79SJohannes Weinerchoice 50b3fbd58fSJohannes Weiner prompt "Default compressor" 51519bcb79SJohannes Weiner depends on ZSWAP 52519bcb79SJohannes Weiner default ZSWAP_COMPRESSOR_DEFAULT_LZO 53519bcb79SJohannes Weiner help 54519bcb79SJohannes Weiner Selects the default compression algorithm for the compressed cache 55519bcb79SJohannes Weiner for swap pages. 56519bcb79SJohannes Weiner 57519bcb79SJohannes Weiner For an overview what kind of performance can be expected from 58519bcb79SJohannes Weiner a particular compression algorithm please refer to the benchmarks 59519bcb79SJohannes Weiner available at the following LWN page: 60519bcb79SJohannes Weiner https://lwn.net/Articles/751795/ 61519bcb79SJohannes Weiner 62519bcb79SJohannes Weiner If in doubt, select 'LZO'. 63519bcb79SJohannes Weiner 64519bcb79SJohannes Weiner The selection made here can be overridden by using the kernel 65519bcb79SJohannes Weiner command line 'zswap.compressor=' option. 66519bcb79SJohannes Weiner 67519bcb79SJohannes Weinerconfig ZSWAP_COMPRESSOR_DEFAULT_DEFLATE 68519bcb79SJohannes Weiner bool "Deflate" 69519bcb79SJohannes Weiner select CRYPTO_DEFLATE 70519bcb79SJohannes Weiner help 71519bcb79SJohannes Weiner Use the Deflate algorithm as the default compression algorithm. 72519bcb79SJohannes Weiner 73519bcb79SJohannes Weinerconfig ZSWAP_COMPRESSOR_DEFAULT_LZO 74519bcb79SJohannes Weiner bool "LZO" 75519bcb79SJohannes Weiner select CRYPTO_LZO 76519bcb79SJohannes Weiner help 77519bcb79SJohannes Weiner Use the LZO algorithm as the default compression algorithm. 78519bcb79SJohannes Weiner 79519bcb79SJohannes Weinerconfig ZSWAP_COMPRESSOR_DEFAULT_842 80519bcb79SJohannes Weiner bool "842" 81519bcb79SJohannes Weiner select CRYPTO_842 82519bcb79SJohannes Weiner help 83519bcb79SJohannes Weiner Use the 842 algorithm as the default compression algorithm. 84519bcb79SJohannes Weiner 85519bcb79SJohannes Weinerconfig ZSWAP_COMPRESSOR_DEFAULT_LZ4 86519bcb79SJohannes Weiner bool "LZ4" 87519bcb79SJohannes Weiner select CRYPTO_LZ4 88519bcb79SJohannes Weiner help 89519bcb79SJohannes Weiner Use the LZ4 algorithm as the default compression algorithm. 90519bcb79SJohannes Weiner 91519bcb79SJohannes Weinerconfig ZSWAP_COMPRESSOR_DEFAULT_LZ4HC 92519bcb79SJohannes Weiner bool "LZ4HC" 93519bcb79SJohannes Weiner select CRYPTO_LZ4HC 94519bcb79SJohannes Weiner help 95519bcb79SJohannes Weiner Use the LZ4HC algorithm as the default compression algorithm. 96519bcb79SJohannes Weiner 97519bcb79SJohannes Weinerconfig ZSWAP_COMPRESSOR_DEFAULT_ZSTD 98519bcb79SJohannes Weiner bool "zstd" 99519bcb79SJohannes Weiner select CRYPTO_ZSTD 100519bcb79SJohannes Weiner help 101519bcb79SJohannes Weiner Use the zstd algorithm as the default compression algorithm. 102519bcb79SJohannes Weinerendchoice 103519bcb79SJohannes Weiner 104519bcb79SJohannes Weinerconfig ZSWAP_COMPRESSOR_DEFAULT 105519bcb79SJohannes Weiner string 106519bcb79SJohannes Weiner depends on ZSWAP 107519bcb79SJohannes Weiner default "deflate" if ZSWAP_COMPRESSOR_DEFAULT_DEFLATE 108519bcb79SJohannes Weiner default "lzo" if ZSWAP_COMPRESSOR_DEFAULT_LZO 109519bcb79SJohannes Weiner default "842" if ZSWAP_COMPRESSOR_DEFAULT_842 110519bcb79SJohannes Weiner default "lz4" if ZSWAP_COMPRESSOR_DEFAULT_LZ4 111519bcb79SJohannes Weiner default "lz4hc" if ZSWAP_COMPRESSOR_DEFAULT_LZ4HC 112519bcb79SJohannes Weiner default "zstd" if ZSWAP_COMPRESSOR_DEFAULT_ZSTD 113519bcb79SJohannes Weiner default "" 114519bcb79SJohannes Weiner 115519bcb79SJohannes Weinerchoice 116b3fbd58fSJohannes Weiner prompt "Default allocator" 117519bcb79SJohannes Weiner depends on ZSWAP 118519bcb79SJohannes Weiner default ZSWAP_ZPOOL_DEFAULT_ZBUD 119519bcb79SJohannes Weiner help 120519bcb79SJohannes Weiner Selects the default allocator for the compressed cache for 121519bcb79SJohannes Weiner swap pages. 122519bcb79SJohannes Weiner The default is 'zbud' for compatibility, however please do 123519bcb79SJohannes Weiner read the description of each of the allocators below before 124519bcb79SJohannes Weiner making a right choice. 125519bcb79SJohannes Weiner 126519bcb79SJohannes Weiner The selection made here can be overridden by using the kernel 127519bcb79SJohannes Weiner command line 'zswap.zpool=' option. 128519bcb79SJohannes Weiner 129519bcb79SJohannes Weinerconfig ZSWAP_ZPOOL_DEFAULT_ZBUD 130519bcb79SJohannes Weiner bool "zbud" 131519bcb79SJohannes Weiner select ZBUD 132519bcb79SJohannes Weiner help 133519bcb79SJohannes Weiner Use the zbud allocator as the default allocator. 134519bcb79SJohannes Weiner 135519bcb79SJohannes Weinerconfig ZSWAP_ZPOOL_DEFAULT_Z3FOLD 136519bcb79SJohannes Weiner bool "z3fold" 137519bcb79SJohannes Weiner select Z3FOLD 138519bcb79SJohannes Weiner help 139519bcb79SJohannes Weiner Use the z3fold allocator as the default allocator. 140519bcb79SJohannes Weiner 141519bcb79SJohannes Weinerconfig ZSWAP_ZPOOL_DEFAULT_ZSMALLOC 142519bcb79SJohannes Weiner bool "zsmalloc" 143519bcb79SJohannes Weiner select ZSMALLOC 144519bcb79SJohannes Weiner help 145519bcb79SJohannes Weiner Use the zsmalloc allocator as the default allocator. 146519bcb79SJohannes Weinerendchoice 147519bcb79SJohannes Weiner 148519bcb79SJohannes Weinerconfig ZSWAP_ZPOOL_DEFAULT 149519bcb79SJohannes Weiner string 150519bcb79SJohannes Weiner depends on ZSWAP 151519bcb79SJohannes Weiner default "zbud" if ZSWAP_ZPOOL_DEFAULT_ZBUD 152519bcb79SJohannes Weiner default "z3fold" if ZSWAP_ZPOOL_DEFAULT_Z3FOLD 153519bcb79SJohannes Weiner default "zsmalloc" if ZSWAP_ZPOOL_DEFAULT_ZSMALLOC 154519bcb79SJohannes Weiner default "" 155519bcb79SJohannes Weiner 156519bcb79SJohannes Weinerconfig ZBUD 157b3fbd58fSJohannes Weiner tristate "2:1 compression allocator (zbud)" 158b3fbd58fSJohannes Weiner depends on ZSWAP 159519bcb79SJohannes Weiner help 160519bcb79SJohannes Weiner A special purpose allocator for storing compressed pages. 161519bcb79SJohannes Weiner It is designed to store up to two compressed pages per physical 162519bcb79SJohannes Weiner page. While this design limits storage density, it has simple and 163519bcb79SJohannes Weiner deterministic reclaim properties that make it preferable to a higher 164519bcb79SJohannes Weiner density approach when reclaim will be used. 165519bcb79SJohannes Weiner 166519bcb79SJohannes Weinerconfig Z3FOLD 167b3fbd58fSJohannes Weiner tristate "3:1 compression allocator (z3fold)" 168b3fbd58fSJohannes Weiner depends on ZSWAP 169519bcb79SJohannes Weiner help 170519bcb79SJohannes Weiner A special purpose allocator for storing compressed pages. 171519bcb79SJohannes Weiner It is designed to store up to three compressed pages per physical 172519bcb79SJohannes Weiner page. It is a ZBUD derivative so the simplicity and determinism are 173519bcb79SJohannes Weiner still there. 174519bcb79SJohannes Weiner 175519bcb79SJohannes Weinerconfig ZSMALLOC 176b3fbd58fSJohannes Weiner tristate 177b3fbd58fSJohannes Weiner prompt "N:1 compression allocator (zsmalloc)" if ZSWAP 178519bcb79SJohannes Weiner depends on MMU 179519bcb79SJohannes Weiner help 180519bcb79SJohannes Weiner zsmalloc is a slab-based memory allocator designed to store 181b3fbd58fSJohannes Weiner pages of various compression levels efficiently. It achieves 182b3fbd58fSJohannes Weiner the highest storage density with the least amount of fragmentation. 183519bcb79SJohannes Weiner 184519bcb79SJohannes Weinerconfig ZSMALLOC_STAT 185519bcb79SJohannes Weiner bool "Export zsmalloc statistics" 186519bcb79SJohannes Weiner depends on ZSMALLOC 187519bcb79SJohannes Weiner select DEBUG_FS 188519bcb79SJohannes Weiner help 189519bcb79SJohannes Weiner This option enables code in the zsmalloc to collect various 190519bcb79SJohannes Weiner statistics about what's happening in zsmalloc and exports that 191519bcb79SJohannes Weiner information to userspace via debugfs. 192519bcb79SJohannes Weiner If unsure, say N. 193519bcb79SJohannes Weiner 1944ff93b29SSergey Senozhatskyconfig ZSMALLOC_CHAIN_SIZE 1954ff93b29SSergey Senozhatsky int "Maximum number of physical pages per-zspage" 196b46402faSSergey Senozhatsky default 8 1974ff93b29SSergey Senozhatsky range 4 16 1984ff93b29SSergey Senozhatsky depends on ZSMALLOC 1994ff93b29SSergey Senozhatsky help 2004ff93b29SSergey Senozhatsky This option sets the upper limit on the number of physical pages 2014ff93b29SSergey Senozhatsky that a zmalloc page (zspage) can consist of. The optimal zspage 2024ff93b29SSergey Senozhatsky chain size is calculated for each size class during the 2034ff93b29SSergey Senozhatsky initialization of the pool. 2044ff93b29SSergey Senozhatsky 2054ff93b29SSergey Senozhatsky Changing this option can alter the characteristics of size classes, 2064ff93b29SSergey Senozhatsky such as the number of pages per zspage and the number of objects 2074ff93b29SSergey Senozhatsky per zspage. This can also result in different configurations of 2084ff93b29SSergey Senozhatsky the pool, as zsmalloc merges size classes with similar 2094ff93b29SSergey Senozhatsky characteristics. 2104ff93b29SSergey Senozhatsky 2114ff93b29SSergey Senozhatsky For more information, see zsmalloc documentation. 2124ff93b29SSergey Senozhatsky 213519bcb79SJohannes Weinermenu "SLAB allocator options" 214519bcb79SJohannes Weiner 2157b42f104SJohannes Weinerchoice 2167b42f104SJohannes Weiner prompt "Choose SLAB allocator" 2177b42f104SJohannes Weiner default SLUB 2187b42f104SJohannes Weiner help 2197b42f104SJohannes Weiner This option allows to select a slab allocator. 2207b42f104SJohannes Weiner 2217b42f104SJohannes Weinerconfig SLAB 2227b42f104SJohannes Weiner bool "SLAB" 2237b42f104SJohannes Weiner depends on !PREEMPT_RT 2247b42f104SJohannes Weiner select HAVE_HARDENED_USERCOPY_ALLOCATOR 2257b42f104SJohannes Weiner help 2267b42f104SJohannes Weiner The regular slab allocator that is established and known to work 2277b42f104SJohannes Weiner well in all environments. It organizes cache hot objects in 2287b42f104SJohannes Weiner per cpu and per node queues. 2297b42f104SJohannes Weiner 2307b42f104SJohannes Weinerconfig SLUB 2317b42f104SJohannes Weiner bool "SLUB (Unqueued Allocator)" 2327b42f104SJohannes Weiner select HAVE_HARDENED_USERCOPY_ALLOCATOR 2337b42f104SJohannes Weiner help 2347b42f104SJohannes Weiner SLUB is a slab allocator that minimizes cache line usage 2357b42f104SJohannes Weiner instead of managing queues of cached objects (SLAB approach). 2367b42f104SJohannes Weiner Per cpu caching is realized using slabs of objects instead 2377b42f104SJohannes Weiner of queues of objects. SLUB can use memory efficiently 2387b42f104SJohannes Weiner and has enhanced diagnostics. SLUB is the default choice for 2397b42f104SJohannes Weiner a slab allocator. 2407b42f104SJohannes Weiner 241149b6fa2SVlastimil Babkaconfig SLOB_DEPRECATED 2427b42f104SJohannes Weiner depends on EXPERT 243149b6fa2SVlastimil Babka bool "SLOB (Simple Allocator - DEPRECATED)" 2447b42f104SJohannes Weiner depends on !PREEMPT_RT 2457b42f104SJohannes Weiner help 246149b6fa2SVlastimil Babka Deprecated and scheduled for removal in a few cycles. SLUB 247149b6fa2SVlastimil Babka recommended as replacement. CONFIG_SLUB_TINY can be considered 248149b6fa2SVlastimil Babka on systems with 16MB or less RAM. 249149b6fa2SVlastimil Babka 250149b6fa2SVlastimil Babka If you need SLOB to stay, please contact linux-mm@kvack.org and 251149b6fa2SVlastimil Babka people listed in the SLAB ALLOCATOR section of MAINTAINERS file, 252149b6fa2SVlastimil Babka with your use case. 253149b6fa2SVlastimil Babka 2547b42f104SJohannes Weiner SLOB replaces the stock allocator with a drastically simpler 2557b42f104SJohannes Weiner allocator. SLOB is generally more space efficient but 2567b42f104SJohannes Weiner does not perform as well on large systems. 2577b42f104SJohannes Weiner 2587b42f104SJohannes Weinerendchoice 2597b42f104SJohannes Weiner 260149b6fa2SVlastimil Babkaconfig SLOB 261149b6fa2SVlastimil Babka bool 262149b6fa2SVlastimil Babka default y 263149b6fa2SVlastimil Babka depends on SLOB_DEPRECATED 264149b6fa2SVlastimil Babka 265e240e53aSVlastimil Babkaconfig SLUB_TINY 266e240e53aSVlastimil Babka bool "Configure SLUB for minimal memory footprint" 267e240e53aSVlastimil Babka depends on SLUB && EXPERT 268e240e53aSVlastimil Babka select SLAB_MERGE_DEFAULT 269e240e53aSVlastimil Babka help 270e240e53aSVlastimil Babka Configures the SLUB allocator in a way to achieve minimal memory 271e240e53aSVlastimil Babka footprint, sacrificing scalability, debugging and other features. 272e240e53aSVlastimil Babka This is intended only for the smallest system that had used the 273e240e53aSVlastimil Babka SLOB allocator and is not recommended for systems with more than 274e240e53aSVlastimil Babka 16MB RAM. 275e240e53aSVlastimil Babka 276e240e53aSVlastimil Babka If unsure, say N. 277e240e53aSVlastimil Babka 2787b42f104SJohannes Weinerconfig SLAB_MERGE_DEFAULT 2797b42f104SJohannes Weiner bool "Allow slab caches to be merged" 2807b42f104SJohannes Weiner default y 2817b42f104SJohannes Weiner depends on SLAB || SLUB 2827b42f104SJohannes Weiner help 2837b42f104SJohannes Weiner For reduced kernel memory fragmentation, slab caches can be 2847b42f104SJohannes Weiner merged when they share the same size and other characteristics. 2857b42f104SJohannes Weiner This carries a risk of kernel heap overflows being able to 2867b42f104SJohannes Weiner overwrite objects from merged caches (and more easily control 2877b42f104SJohannes Weiner cache layout), which makes such heap attacks easier to exploit 2887b42f104SJohannes Weiner by attackers. By keeping caches unmerged, these kinds of exploits 2897b42f104SJohannes Weiner can usually only damage objects in the same cache. To disable 2907b42f104SJohannes Weiner merging at runtime, "slab_nomerge" can be passed on the kernel 2917b42f104SJohannes Weiner command line. 2927b42f104SJohannes Weiner 2937b42f104SJohannes Weinerconfig SLAB_FREELIST_RANDOM 2947b42f104SJohannes Weiner bool "Randomize slab freelist" 295e240e53aSVlastimil Babka depends on SLAB || (SLUB && !SLUB_TINY) 2967b42f104SJohannes Weiner help 2977b42f104SJohannes Weiner Randomizes the freelist order used on creating new pages. This 2987b42f104SJohannes Weiner security feature reduces the predictability of the kernel slab 2997b42f104SJohannes Weiner allocator against heap overflows. 3007b42f104SJohannes Weiner 3017b42f104SJohannes Weinerconfig SLAB_FREELIST_HARDENED 3027b42f104SJohannes Weiner bool "Harden slab freelist metadata" 303e240e53aSVlastimil Babka depends on SLAB || (SLUB && !SLUB_TINY) 3047b42f104SJohannes Weiner help 3057b42f104SJohannes Weiner Many kernel heap attacks try to target slab cache metadata and 3067b42f104SJohannes Weiner other infrastructure. This options makes minor performance 3077b42f104SJohannes Weiner sacrifices to harden the kernel slab allocator against common 3087b42f104SJohannes Weiner freelist exploit methods. Some slab implementations have more 3097b42f104SJohannes Weiner sanity-checking than others. This option is most effective with 3107b42f104SJohannes Weiner CONFIG_SLUB. 3117b42f104SJohannes Weiner 3120710d012SVlastimil Babkaconfig SLUB_STATS 3130710d012SVlastimil Babka default n 3140710d012SVlastimil Babka bool "Enable SLUB performance statistics" 315e240e53aSVlastimil Babka depends on SLUB && SYSFS && !SLUB_TINY 3160710d012SVlastimil Babka help 3170710d012SVlastimil Babka SLUB statistics are useful to debug SLUBs allocation behavior in 3180710d012SVlastimil Babka order find ways to optimize the allocator. This should never be 3190710d012SVlastimil Babka enabled for production use since keeping statistics slows down 3200710d012SVlastimil Babka the allocator by a few percentage points. The slabinfo command 3210710d012SVlastimil Babka supports the determination of the most active slabs to figure 3220710d012SVlastimil Babka out which slabs are relevant to a particular load. 3230710d012SVlastimil Babka Try running: slabinfo -DA 3240710d012SVlastimil Babka 325519bcb79SJohannes Weinerconfig SLUB_CPU_PARTIAL 326519bcb79SJohannes Weiner default y 327e240e53aSVlastimil Babka depends on SLUB && SMP && !SLUB_TINY 328519bcb79SJohannes Weiner bool "SLUB per cpu partial cache" 329519bcb79SJohannes Weiner help 330519bcb79SJohannes Weiner Per cpu partial caches accelerate objects allocation and freeing 331519bcb79SJohannes Weiner that is local to a processor at the price of more indeterminism 332519bcb79SJohannes Weiner in the latency of the free. On overflow these caches will be cleared 333519bcb79SJohannes Weiner which requires the taking of locks that may cause latency spikes. 334519bcb79SJohannes Weiner Typically one would choose no for a realtime system. 335519bcb79SJohannes Weiner 336519bcb79SJohannes Weinerendmenu # SLAB allocator options 337519bcb79SJohannes Weiner 3387b42f104SJohannes Weinerconfig SHUFFLE_PAGE_ALLOCATOR 3397b42f104SJohannes Weiner bool "Page allocator randomization" 3407b42f104SJohannes Weiner default SLAB_FREELIST_RANDOM && ACPI_NUMA 3417b42f104SJohannes Weiner help 3427b42f104SJohannes Weiner Randomization of the page allocator improves the average 3437b42f104SJohannes Weiner utilization of a direct-mapped memory-side-cache. See section 3447b42f104SJohannes Weiner 5.2.27 Heterogeneous Memory Attribute Table (HMAT) in the ACPI 3457b42f104SJohannes Weiner 6.2a specification for an example of how a platform advertises 3467b42f104SJohannes Weiner the presence of a memory-side-cache. There are also incidental 3477b42f104SJohannes Weiner security benefits as it reduces the predictability of page 3487b42f104SJohannes Weiner allocations to compliment SLAB_FREELIST_RANDOM, but the 34923baf831SKirill A. Shutemov default granularity of shuffling on the MAX_ORDER i.e, 10th 35023baf831SKirill A. Shutemov order of pages is selected based on cache utilization benefits 35123baf831SKirill A. Shutemov on x86. 3527b42f104SJohannes Weiner 3537b42f104SJohannes Weiner While the randomization improves cache utilization it may 3547b42f104SJohannes Weiner negatively impact workloads on platforms without a cache. For 3557b42f104SJohannes Weiner this reason, by default, the randomization is enabled only 3567b42f104SJohannes Weiner after runtime detection of a direct-mapped memory-side-cache. 3577b42f104SJohannes Weiner Otherwise, the randomization may be force enabled with the 3587b42f104SJohannes Weiner 'page_alloc.shuffle' kernel command line parameter. 3597b42f104SJohannes Weiner 3607b42f104SJohannes Weiner Say Y if unsure. 3617b42f104SJohannes Weiner 3620710d012SVlastimil Babkaconfig COMPAT_BRK 3630710d012SVlastimil Babka bool "Disable heap randomization" 3640710d012SVlastimil Babka default y 3650710d012SVlastimil Babka help 3660710d012SVlastimil Babka Randomizing heap placement makes heap exploits harder, but it 3670710d012SVlastimil Babka also breaks ancient binaries (including anything libc5 based). 3680710d012SVlastimil Babka This option changes the bootup default to heap randomization 3690710d012SVlastimil Babka disabled, and can be overridden at runtime by setting 3700710d012SVlastimil Babka /proc/sys/kernel/randomize_va_space to 2. 3710710d012SVlastimil Babka 3720710d012SVlastimil Babka On non-ancient distros (post-2000 ones) N is usually a safe choice. 3730710d012SVlastimil Babka 3740710d012SVlastimil Babkaconfig MMAP_ALLOW_UNINITIALIZED 3750710d012SVlastimil Babka bool "Allow mmapped anonymous memory to be uninitialized" 3760710d012SVlastimil Babka depends on EXPERT && !MMU 3770710d012SVlastimil Babka default n 3780710d012SVlastimil Babka help 3790710d012SVlastimil Babka Normally, and according to the Linux spec, anonymous memory obtained 3800710d012SVlastimil Babka from mmap() has its contents cleared before it is passed to 3810710d012SVlastimil Babka userspace. Enabling this config option allows you to request that 3820710d012SVlastimil Babka mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus 3830710d012SVlastimil Babka providing a huge performance boost. If this option is not enabled, 3840710d012SVlastimil Babka then the flag will be ignored. 3850710d012SVlastimil Babka 3860710d012SVlastimil Babka This is taken advantage of by uClibc's malloc(), and also by 3870710d012SVlastimil Babka ELF-FDPIC binfmt's brk and stack allocator. 3880710d012SVlastimil Babka 3890710d012SVlastimil Babka Because of the obvious security issues, this option should only be 3900710d012SVlastimil Babka enabled on embedded devices where you control what is run in 3910710d012SVlastimil Babka userspace. Since that isn't generally a problem on no-MMU systems, 3920710d012SVlastimil Babka it is normally safe to say Y here. 3930710d012SVlastimil Babka 3940710d012SVlastimil Babka See Documentation/admin-guide/mm/nommu-mmap.rst for more information. 3950710d012SVlastimil Babka 396e1785e85SDave Hansenconfig SELECT_MEMORY_MODEL 397e1785e85SDave Hansen def_bool y 398a8826eebSKees Cook depends on ARCH_SELECT_MEMORY_MODEL 399e1785e85SDave Hansen 4003a9da765SDave Hansenchoice 4013a9da765SDave Hansen prompt "Memory model" 402e1785e85SDave Hansen depends on SELECT_MEMORY_MODEL 403d41dee36SAndy Whitcroft default SPARSEMEM_MANUAL if ARCH_SPARSEMEM_DEFAULT 404e1785e85SDave Hansen default FLATMEM_MANUAL 405d66d109dSMike Rapoport help 406d66d109dSMike Rapoport This option allows you to change some of the ways that 407d66d109dSMike Rapoport Linux manages its memory internally. Most users will 408d66d109dSMike Rapoport only have one option here selected by the architecture 409d66d109dSMike Rapoport configuration. This is normal. 4103a9da765SDave Hansen 411e1785e85SDave Hansenconfig FLATMEM_MANUAL 4123a9da765SDave Hansen bool "Flat Memory" 413bb1c50d3SMike Rapoport depends on !ARCH_SPARSEMEM_ENABLE || ARCH_FLATMEM_ENABLE 4143a9da765SDave Hansen help 415d66d109dSMike Rapoport This option is best suited for non-NUMA systems with 416d66d109dSMike Rapoport flat address space. The FLATMEM is the most efficient 417d66d109dSMike Rapoport system in terms of performance and resource consumption 418d66d109dSMike Rapoport and it is the best option for smaller systems. 4193a9da765SDave Hansen 420d66d109dSMike Rapoport For systems that have holes in their physical address 421d66d109dSMike Rapoport spaces and for features like NUMA and memory hotplug, 422dd33d29aSRandy Dunlap choose "Sparse Memory". 423d41dee36SAndy Whitcroft 424d41dee36SAndy Whitcroft If unsure, choose this option (Flat Memory) over any other. 4253a9da765SDave Hansen 426d41dee36SAndy Whitcroftconfig SPARSEMEM_MANUAL 427d41dee36SAndy Whitcroft bool "Sparse Memory" 428d41dee36SAndy Whitcroft depends on ARCH_SPARSEMEM_ENABLE 429d41dee36SAndy Whitcroft help 430d41dee36SAndy Whitcroft This will be the only option for some systems, including 431d66d109dSMike Rapoport memory hot-plug systems. This is normal. 432d41dee36SAndy Whitcroft 433d66d109dSMike Rapoport This option provides efficient support for systems with 434d66d109dSMike Rapoport holes is their physical address space and allows memory 435d66d109dSMike Rapoport hot-plug and hot-remove. 436d41dee36SAndy Whitcroft 437d66d109dSMike Rapoport If unsure, choose "Flat Memory" over this option. 438d41dee36SAndy Whitcroft 4393a9da765SDave Hansenendchoice 4403a9da765SDave Hansen 441d41dee36SAndy Whitcroftconfig SPARSEMEM 442d41dee36SAndy Whitcroft def_bool y 4431a83e175SRussell King depends on (!SELECT_MEMORY_MODEL && ARCH_SPARSEMEM_ENABLE) || SPARSEMEM_MANUAL 444d41dee36SAndy Whitcroft 445e1785e85SDave Hansenconfig FLATMEM 446e1785e85SDave Hansen def_bool y 447bb1c50d3SMike Rapoport depends on !SPARSEMEM || FLATMEM_MANUAL 448d41dee36SAndy Whitcroft 44993b7504eSDave Hansen# 4503e347261SBob Picco# SPARSEMEM_EXTREME (which is the default) does some bootmem 451c89ab04fSMike Rapoport# allocations when sparse_init() is called. If this cannot 4523e347261SBob Picco# be done on your architecture, select this option. However, 4533e347261SBob Picco# statically allocating the mem_section[] array can potentially 4543e347261SBob Picco# consume vast quantities of .bss, so be careful. 4553e347261SBob Picco# 4563e347261SBob Picco# This option will also potentially produce smaller runtime code 4573e347261SBob Picco# with gcc 3.4 and later. 4583e347261SBob Picco# 4593e347261SBob Piccoconfig SPARSEMEM_STATIC 4609ba16087SJan Beulich bool 4613e347261SBob Picco 4623e347261SBob Picco# 46344c09201SMatt LaPlante# Architecture platforms which require a two level mem_section in SPARSEMEM 464802f192eSBob Picco# must select this option. This is usually for architecture platforms with 465802f192eSBob Picco# an extremely sparse physical address space. 466802f192eSBob Picco# 4673e347261SBob Piccoconfig SPARSEMEM_EXTREME 4683e347261SBob Picco def_bool y 4693e347261SBob Picco depends on SPARSEMEM && !SPARSEMEM_STATIC 4704c21e2f2SHugh Dickins 47129c71111SAndy Whitcroftconfig SPARSEMEM_VMEMMAP_ENABLE 4729ba16087SJan Beulich bool 47329c71111SAndy Whitcroft 47429c71111SAndy Whitcroftconfig SPARSEMEM_VMEMMAP 475a5ee6daaSGeoff Levand bool "Sparse Memory virtual memmap" 476a5ee6daaSGeoff Levand depends on SPARSEMEM && SPARSEMEM_VMEMMAP_ENABLE 477a5ee6daaSGeoff Levand default y 478a5ee6daaSGeoff Levand help 479a5ee6daaSGeoff Levand SPARSEMEM_VMEMMAP uses a virtually mapped memmap to optimise 480a5ee6daaSGeoff Levand pfn_to_page and page_to_pfn operations. This is the most 481a5ee6daaSGeoff Levand efficient option when sufficient kernel resources are available. 482*0b376f1eSAneesh Kumar K.V# 483*0b376f1eSAneesh Kumar K.V# Select this config option from the architecture Kconfig, if it is preferred 484*0b376f1eSAneesh Kumar K.V# to enable the feature of HugeTLB/dev_dax vmemmap optimization. 485*0b376f1eSAneesh Kumar K.V# 486*0b376f1eSAneesh Kumar K.Vconfig ARCH_WANT_OPTIMIZE_VMEMMAP 487*0b376f1eSAneesh Kumar K.V bool 48829c71111SAndy Whitcroft 48970210ed9SPhilipp Hachtmannconfig HAVE_MEMBLOCK_PHYS_MAP 4906341e62bSChristoph Jaeger bool 49170210ed9SPhilipp Hachtmann 49267a929e0SChristoph Hellwigconfig HAVE_FAST_GUP 493050a9adcSChristoph Hellwig depends on MMU 4946341e62bSChristoph Jaeger bool 4952667f50eSSteve Capper 49652219aeaSDavid Hildenbrand# Don't discard allocated memory used to track "memory" and "reserved" memblocks 49752219aeaSDavid Hildenbrand# after early boot, so it can still be used to test for validity of memory. 49852219aeaSDavid Hildenbrand# Also, memblocks are updated with memory hot(un)plug. 499350e88baSMike Rapoportconfig ARCH_KEEP_MEMBLOCK 5006341e62bSChristoph Jaeger bool 501c378ddd5STejun Heo 5021e5d8e1eSDan Williams# Keep arch NUMA mapping infrastructure post-init. 5031e5d8e1eSDan Williamsconfig NUMA_KEEP_MEMINFO 5041e5d8e1eSDan Williams bool 5051e5d8e1eSDan Williams 506ee6f509cSMinchan Kimconfig MEMORY_ISOLATION 5076341e62bSChristoph Jaeger bool 508ee6f509cSMinchan Kim 509a9e7b8d4SDavid Hildenbrand# IORESOURCE_SYSTEM_RAM regions in the kernel resource tree that are marked 510a9e7b8d4SDavid Hildenbrand# IORESOURCE_EXCLUSIVE cannot be mapped to user space, for example, via 511a9e7b8d4SDavid Hildenbrand# /dev/mem. 512a9e7b8d4SDavid Hildenbrandconfig EXCLUSIVE_SYSTEM_RAM 513a9e7b8d4SDavid Hildenbrand def_bool y 514a9e7b8d4SDavid Hildenbrand depends on !DEVMEM || STRICT_DEVMEM 515a9e7b8d4SDavid Hildenbrand 51646723bfaSYasuaki Ishimatsu# 51746723bfaSYasuaki Ishimatsu# Only be set on architectures that have completely implemented memory hotplug 51846723bfaSYasuaki Ishimatsu# feature. If you are not sure, don't touch it. 51946723bfaSYasuaki Ishimatsu# 52046723bfaSYasuaki Ishimatsuconfig HAVE_BOOTMEM_INFO_NODE 52146723bfaSYasuaki Ishimatsu def_bool n 52246723bfaSYasuaki Ishimatsu 52391024b3cSAnshuman Khandualconfig ARCH_ENABLE_MEMORY_HOTPLUG 52491024b3cSAnshuman Khandual bool 52591024b3cSAnshuman Khandual 526519bcb79SJohannes Weinerconfig ARCH_ENABLE_MEMORY_HOTREMOVE 527519bcb79SJohannes Weiner bool 528519bcb79SJohannes Weiner 5293947be19SDave Hansen# eventually, we can have this option just 'select SPARSEMEM' 530519bcb79SJohannes Weinermenuconfig MEMORY_HOTPLUG 531519bcb79SJohannes Weiner bool "Memory hotplug" 532b30c5927SDavid Hildenbrand select MEMORY_ISOLATION 53371b6f2ddSDavid Hildenbrand depends on SPARSEMEM 53440b31360SStephen Rothwell depends on ARCH_ENABLE_MEMORY_HOTPLUG 5357ec58a2bSDavid Hildenbrand depends on 64BIT 5361e5d8e1eSDan Williams select NUMA_KEEP_MEMINFO if NUMA 5373947be19SDave Hansen 538519bcb79SJohannes Weinerif MEMORY_HOTPLUG 539519bcb79SJohannes Weiner 5408604d9e5SVitaly Kuznetsovconfig MEMORY_HOTPLUG_DEFAULT_ONLINE 5418604d9e5SVitaly Kuznetsov bool "Online the newly added memory blocks by default" 5428604d9e5SVitaly Kuznetsov depends on MEMORY_HOTPLUG 5438604d9e5SVitaly Kuznetsov help 5448604d9e5SVitaly Kuznetsov This option sets the default policy setting for memory hotplug 5458604d9e5SVitaly Kuznetsov onlining policy (/sys/devices/system/memory/auto_online_blocks) which 5468604d9e5SVitaly Kuznetsov determines what happens to newly added memory regions. Policy setting 5478604d9e5SVitaly Kuznetsov can always be changed at runtime. 548cb1aaebeSMauro Carvalho Chehab See Documentation/admin-guide/mm/memory-hotplug.rst for more information. 5498604d9e5SVitaly Kuznetsov 5508604d9e5SVitaly Kuznetsov Say Y here if you want all hot-plugged memory blocks to appear in 5518604d9e5SVitaly Kuznetsov 'online' state by default. 5528604d9e5SVitaly Kuznetsov Say N here if you want the default policy to keep all hot-plugged 5538604d9e5SVitaly Kuznetsov memory blocks in 'offline' state. 5548604d9e5SVitaly Kuznetsov 5550c0e6195SKAMEZAWA Hiroyukiconfig MEMORY_HOTREMOVE 5560c0e6195SKAMEZAWA Hiroyuki bool "Allow for memory hot remove" 557f7e3334aSNathan Fontenot select HAVE_BOOTMEM_INFO_NODE if (X86_64 || PPC64) 5580c0e6195SKAMEZAWA Hiroyuki depends on MEMORY_HOTPLUG && ARCH_ENABLE_MEMORY_HOTREMOVE 5590c0e6195SKAMEZAWA Hiroyuki depends on MIGRATION 5600c0e6195SKAMEZAWA Hiroyuki 561a08a2ae3SOscar Salvadorconfig MHP_MEMMAP_ON_MEMORY 562a08a2ae3SOscar Salvador def_bool y 563a08a2ae3SOscar Salvador depends on MEMORY_HOTPLUG && SPARSEMEM_VMEMMAP 564a08a2ae3SOscar Salvador depends on ARCH_MHP_MEMMAP_ON_MEMORY_ENABLE 565a08a2ae3SOscar Salvador 566519bcb79SJohannes Weinerendif # MEMORY_HOTPLUG 567519bcb79SJohannes Weiner 5684c21e2f2SHugh Dickins# Heavily threaded applications may benefit from splitting the mm-wide 5694c21e2f2SHugh Dickins# page_table_lock, so that faults on different parts of the user address 5704c21e2f2SHugh Dickins# space can be handled with less contention: split it at this NR_CPUS. 5714c21e2f2SHugh Dickins# Default to 4 for wider testing, though 8 might be more appropriate. 5724c21e2f2SHugh Dickins# ARM's adjust_pte (unused if VIPT) depends on mm-wide page_table_lock. 5737b6ac9dfSHugh Dickins# PA-RISC 7xxx's spinlock_t would enlarge struct page from 32 to 44 bytes. 57460bccaa6SWill Deacon# SPARC32 allocates multiple pte tables within a single page, and therefore 57560bccaa6SWill Deacon# a per-page lock leads to problems when multiple tables need to be locked 57660bccaa6SWill Deacon# at the same time (e.g. copy_page_range()). 577a70caa8bSHugh Dickins# DEBUG_SPINLOCK and DEBUG_LOCK_ALLOC spinlock_t also enlarge struct page. 5784c21e2f2SHugh Dickins# 5794c21e2f2SHugh Dickinsconfig SPLIT_PTLOCK_CPUS 5804c21e2f2SHugh Dickins int 5819164550eSKirill A. Shutemov default "999999" if !MMU 582a70caa8bSHugh Dickins default "999999" if ARM && !CPU_CACHE_VIPT 583a70caa8bSHugh Dickins default "999999" if PARISC && !PA20 58460bccaa6SWill Deacon default "999999" if SPARC32 5854c21e2f2SHugh Dickins default "4" 5867cbe34cfSChristoph Lameter 587e009bb30SKirill A. Shutemovconfig ARCH_ENABLE_SPLIT_PMD_PTLOCK 5886341e62bSChristoph Jaeger bool 589e009bb30SKirill A. Shutemov 5907cbe34cfSChristoph Lameter# 59109316c09SKonstantin Khlebnikov# support for memory balloon 59209316c09SKonstantin Khlebnikovconfig MEMORY_BALLOON 5936341e62bSChristoph Jaeger bool 59409316c09SKonstantin Khlebnikov 59509316c09SKonstantin Khlebnikov# 59618468d93SRafael Aquini# support for memory balloon compaction 59718468d93SRafael Aquiniconfig BALLOON_COMPACTION 59818468d93SRafael Aquini bool "Allow for balloon memory compaction/migration" 59918468d93SRafael Aquini def_bool y 60009316c09SKonstantin Khlebnikov depends on COMPACTION && MEMORY_BALLOON 60118468d93SRafael Aquini help 60218468d93SRafael Aquini Memory fragmentation introduced by ballooning might reduce 60318468d93SRafael Aquini significantly the number of 2MB contiguous memory blocks that can be 60418468d93SRafael Aquini used within a guest, thus imposing performance penalties associated 60518468d93SRafael Aquini with the reduced number of transparent huge pages that could be used 60618468d93SRafael Aquini by the guest workload. Allowing the compaction & migration for memory 60718468d93SRafael Aquini pages enlisted as being part of memory balloon devices avoids the 60818468d93SRafael Aquini scenario aforementioned and helps improving memory defragmentation. 60918468d93SRafael Aquini 61018468d93SRafael Aquini# 611e9e96b39SMel Gorman# support for memory compaction 612e9e96b39SMel Gormanconfig COMPACTION 613e9e96b39SMel Gorman bool "Allow for memory compaction" 61405106e6aSRik van Riel def_bool y 615e9e96b39SMel Gorman select MIGRATION 61633a93877SAndrea Arcangeli depends on MMU 617e9e96b39SMel Gorman help 618b32eaf71SMichal Hocko Compaction is the only memory management component to form 619b32eaf71SMichal Hocko high order (larger physically contiguous) memory blocks 620b32eaf71SMichal Hocko reliably. The page allocator relies on compaction heavily and 621b32eaf71SMichal Hocko the lack of the feature can lead to unexpected OOM killer 622b32eaf71SMichal Hocko invocations for high order memory requests. You shouldn't 623b32eaf71SMichal Hocko disable this option unless there really is a strong reason for 624b32eaf71SMichal Hocko it and then we would be really interested to hear about that at 625b32eaf71SMichal Hocko linux-mm@kvack.org. 626e9e96b39SMel Gorman 627c7e0b3d0SThomas Gleixnerconfig COMPACT_UNEVICTABLE_DEFAULT 628c7e0b3d0SThomas Gleixner int 629c7e0b3d0SThomas Gleixner depends on COMPACTION 630c7e0b3d0SThomas Gleixner default 0 if PREEMPT_RT 631c7e0b3d0SThomas Gleixner default 1 632c7e0b3d0SThomas Gleixner 633e9e96b39SMel Gorman# 63436e66c55SAlexander Duyck# support for free page reporting 63536e66c55SAlexander Duyckconfig PAGE_REPORTING 63636e66c55SAlexander Duyck bool "Free page reporting" 63736e66c55SAlexander Duyck def_bool n 63836e66c55SAlexander Duyck help 63936e66c55SAlexander Duyck Free page reporting allows for the incremental acquisition of 64036e66c55SAlexander Duyck free pages from the buddy allocator for the purpose of reporting 64136e66c55SAlexander Duyck those pages to another entity, such as a hypervisor, so that the 64236e66c55SAlexander Duyck memory can be freed within the host for other uses. 64336e66c55SAlexander Duyck 64436e66c55SAlexander Duyck# 6457cbe34cfSChristoph Lameter# support for page migration 6467cbe34cfSChristoph Lameter# 6477cbe34cfSChristoph Lameterconfig MIGRATION 648b20a3503SChristoph Lameter bool "Page migration" 6496c5240aeSChristoph Lameter def_bool y 650de32a817SChen Gang depends on (NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION || CMA) && MMU 651b20a3503SChristoph Lameter help 652b20a3503SChristoph Lameter Allows the migration of the physical location of pages of processes 653e9e96b39SMel Gorman while the virtual addresses are not changed. This is useful in 654e9e96b39SMel Gorman two situations. The first is on NUMA systems to put pages nearer 655e9e96b39SMel Gorman to the processors accessing. The second is when allocating huge 656e9e96b39SMel Gorman pages as migration can relocate pages to satisfy a huge page 657e9e96b39SMel Gorman allocation instead of reclaiming. 6586550e07fSGreg Kroah-Hartman 65976cbbeadSChristoph Hellwigconfig DEVICE_MIGRATION 660d90a25f8SChristoph Hellwig def_bool MIGRATION && ZONE_DEVICE 66176cbbeadSChristoph Hellwig 662c177c81eSNaoya Horiguchiconfig ARCH_ENABLE_HUGEPAGE_MIGRATION 6636341e62bSChristoph Jaeger bool 664c177c81eSNaoya Horiguchi 6659c670ea3SNaoya Horiguchiconfig ARCH_ENABLE_THP_MIGRATION 6669c670ea3SNaoya Horiguchi bool 6679c670ea3SNaoya Horiguchi 6684bfb68a0SAnshuman Khandualconfig HUGETLB_PAGE_SIZE_VARIABLE 6694bfb68a0SAnshuman Khandual def_bool n 6704bfb68a0SAnshuman Khandual help 6714bfb68a0SAnshuman Khandual Allows the pageblock_order value to be dynamic instead of just standard 6724bfb68a0SAnshuman Khandual HUGETLB_PAGE_ORDER when there are multiple HugeTLB page sizes available 6734bfb68a0SAnshuman Khandual on a platform. 6744bfb68a0SAnshuman Khandual 67523baf831SKirill A. Shutemov Note that the pageblock_order cannot exceed MAX_ORDER and will be 67623baf831SKirill A. Shutemov clamped down to MAX_ORDER. 677b3d40a2bSDavid Hildenbrand 6788df995f6SAlexandre Ghiticonfig CONTIG_ALLOC 6798df995f6SAlexandre Ghiti def_bool (MEMORY_ISOLATION && COMPACTION) || CMA 6808df995f6SAlexandre Ghiti 681600715dcSJeremy Fitzhardingeconfig PHYS_ADDR_T_64BIT 682d4a451d5SChristoph Hellwig def_bool 64BIT 683600715dcSJeremy Fitzhardinge 6842a7326b5SChristoph Lameterconfig BOUNCE 6859ca24e2eSVinayak Menon bool "Enable bounce buffers" 6869ca24e2eSVinayak Menon default y 687ce288e05SChristoph Hellwig depends on BLOCK && MMU && HIGHMEM 6889ca24e2eSVinayak Menon help 689ce288e05SChristoph Hellwig Enable bounce buffers for devices that cannot access the full range of 690ce288e05SChristoph Hellwig memory available to the CPU. Enabled by default when HIGHMEM is 691ce288e05SChristoph Hellwig selected, but you may say n to override this. 6922a7326b5SChristoph Lameter 693cddb8a5cSAndrea Arcangeliconfig MMU_NOTIFIER 694cddb8a5cSAndrea Arcangeli bool 69583fe27eaSPranith Kumar select SRCU 69699cb252fSJason Gunthorpe select INTERVAL_TREE 697fc4d5c29SDavid Howells 698f8af4da3SHugh Dickinsconfig KSM 699f8af4da3SHugh Dickins bool "Enable KSM for page merging" 700f8af4da3SHugh Dickins depends on MMU 70159e1a2f4STimofey Titovets select XXHASH 702f8af4da3SHugh Dickins help 703f8af4da3SHugh Dickins Enable Kernel Samepage Merging: KSM periodically scans those areas 704f8af4da3SHugh Dickins of an application's address space that an app has advised may be 705f8af4da3SHugh Dickins mergeable. When it finds pages of identical content, it replaces 706d0f209f6SHugh Dickins the many instances by a single page with that content, so 707f8af4da3SHugh Dickins saving memory until one or another app needs to modify the content. 708f8af4da3SHugh Dickins Recommended for use with KVM, or with other duplicative applications. 709ee65728eSMike Rapoport See Documentation/mm/ksm.rst for more information: KSM is inactive 710c73602adSHugh Dickins until a program has madvised that an area is MADV_MERGEABLE, and 711c73602adSHugh Dickins root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set). 712f8af4da3SHugh Dickins 713e0a94c2aSChristoph Lameterconfig DEFAULT_MMAP_MIN_ADDR 714e0a94c2aSChristoph Lameter int "Low address space to protect from user allocation" 7156e141546SDavid Howells depends on MMU 716e0a94c2aSChristoph Lameter default 4096 717e0a94c2aSChristoph Lameter help 718e0a94c2aSChristoph Lameter This is the portion of low virtual memory which should be protected 719e0a94c2aSChristoph Lameter from userspace allocation. Keeping a user from writing to low pages 720e0a94c2aSChristoph Lameter can help reduce the impact of kernel NULL pointer bugs. 721e0a94c2aSChristoph Lameter 722e0a94c2aSChristoph Lameter For most ia64, ppc64 and x86 users with lots of address space 723e0a94c2aSChristoph Lameter a value of 65536 is reasonable and should cause no problems. 724e0a94c2aSChristoph Lameter On arm and other archs it should not be higher than 32768. 725788084abSEric Paris Programs which use vm86 functionality or have some need to map 726788084abSEric Paris this low address space will need CAP_SYS_RAWIO or disable this 727788084abSEric Paris protection by setting the value to 0. 728e0a94c2aSChristoph Lameter 729e0a94c2aSChristoph Lameter This value can be changed after boot using the 730e0a94c2aSChristoph Lameter /proc/sys/vm/mmap_min_addr tunable. 731e0a94c2aSChristoph Lameter 732d949f36fSLinus Torvaldsconfig ARCH_SUPPORTS_MEMORY_FAILURE 733d949f36fSLinus Torvalds bool 734e0a94c2aSChristoph Lameter 7356a46079cSAndi Kleenconfig MEMORY_FAILURE 7366a46079cSAndi Kleen depends on MMU 737d949f36fSLinus Torvalds depends on ARCH_SUPPORTS_MEMORY_FAILURE 7386a46079cSAndi Kleen bool "Enable recovery from hardware memory errors" 739ee6f509cSMinchan Kim select MEMORY_ISOLATION 74097f0b134SXie XiuQi select RAS 7416a46079cSAndi Kleen help 7426a46079cSAndi Kleen Enables code to recover from some memory failures on systems 7436a46079cSAndi Kleen with MCA recovery. This allows a system to continue running 7446a46079cSAndi Kleen even when some of its memory has uncorrected errors. This requires 7456a46079cSAndi Kleen special hardware support and typically ECC memory. 7466a46079cSAndi Kleen 747cae681fcSAndi Kleenconfig HWPOISON_INJECT 748413f9efbSAndi Kleen tristate "HWPoison pages injector" 74927df5068SAndi Kleen depends on MEMORY_FAILURE && DEBUG_KERNEL && PROC_FS 750478c5ffcSWu Fengguang select PROC_PAGE_MONITOR 751cae681fcSAndi Kleen 752fc4d5c29SDavid Howellsconfig NOMMU_INITIAL_TRIM_EXCESS 753fc4d5c29SDavid Howells int "Turn on mmap() excess space trimming before booting" 754fc4d5c29SDavid Howells depends on !MMU 755fc4d5c29SDavid Howells default 1 756fc4d5c29SDavid Howells help 757fc4d5c29SDavid Howells The NOMMU mmap() frequently needs to allocate large contiguous chunks 758fc4d5c29SDavid Howells of memory on which to store mappings, but it can only ask the system 759fc4d5c29SDavid Howells allocator for chunks in 2^N*PAGE_SIZE amounts - which is frequently 760fc4d5c29SDavid Howells more than it requires. To deal with this, mmap() is able to trim off 761fc4d5c29SDavid Howells the excess and return it to the allocator. 762fc4d5c29SDavid Howells 763fc4d5c29SDavid Howells If trimming is enabled, the excess is trimmed off and returned to the 764fc4d5c29SDavid Howells system allocator, which can cause extra fragmentation, particularly 765fc4d5c29SDavid Howells if there are a lot of transient processes. 766fc4d5c29SDavid Howells 767fc4d5c29SDavid Howells If trimming is disabled, the excess is kept, but not used, which for 768fc4d5c29SDavid Howells long-term mappings means that the space is wasted. 769fc4d5c29SDavid Howells 770fc4d5c29SDavid Howells Trimming can be dynamically controlled through a sysctl option 771fc4d5c29SDavid Howells (/proc/sys/vm/nr_trim_pages) which specifies the minimum number of 772fc4d5c29SDavid Howells excess pages there must be before trimming should occur, or zero if 773fc4d5c29SDavid Howells no trimming is to occur. 774fc4d5c29SDavid Howells 775fc4d5c29SDavid Howells This option specifies the initial value of this option. The default 776fc4d5c29SDavid Howells of 1 says that all excess pages should be trimmed. 777fc4d5c29SDavid Howells 778dd19d293SStephen Kitt See Documentation/admin-guide/mm/nommu-mmap.rst for more information. 779bbddff05STejun Heo 780519bcb79SJohannes Weinerconfig ARCH_WANT_GENERAL_HUGETLB 781519bcb79SJohannes Weiner bool 782519bcb79SJohannes Weiner 783519bcb79SJohannes Weinerconfig ARCH_WANTS_THP_SWAP 784519bcb79SJohannes Weiner def_bool n 785519bcb79SJohannes Weiner 786519bcb79SJohannes Weinermenuconfig TRANSPARENT_HUGEPAGE 78713ece886SAndrea Arcangeli bool "Transparent Hugepage Support" 788554b0f3cSSebastian Andrzej Siewior depends on HAVE_ARCH_TRANSPARENT_HUGEPAGE && !PREEMPT_RT 7895d689240SAndrea Arcangeli select COMPACTION 7903a08cd52SMatthew Wilcox select XARRAY_MULTI 7914c76d9d1SAndrea Arcangeli help 7924c76d9d1SAndrea Arcangeli Transparent Hugepages allows the kernel to use huge pages and 7934c76d9d1SAndrea Arcangeli huge tlb transparently to the applications whenever possible. 7944c76d9d1SAndrea Arcangeli This feature can improve computing performance to certain 7954c76d9d1SAndrea Arcangeli applications by speeding up page faults during memory 7964c76d9d1SAndrea Arcangeli allocation, by reducing the number of tlb misses and by speeding 7974c76d9d1SAndrea Arcangeli up the pagetable walking. 7984c76d9d1SAndrea Arcangeli 7994c76d9d1SAndrea Arcangeli If memory constrained on embedded, you may want to say N. 8004c76d9d1SAndrea Arcangeli 801519bcb79SJohannes Weinerif TRANSPARENT_HUGEPAGE 802519bcb79SJohannes Weiner 80313ece886SAndrea Arcangelichoice 80413ece886SAndrea Arcangeli prompt "Transparent Hugepage Support sysfs defaults" 80513ece886SAndrea Arcangeli depends on TRANSPARENT_HUGEPAGE 80613ece886SAndrea Arcangeli default TRANSPARENT_HUGEPAGE_ALWAYS 80713ece886SAndrea Arcangeli help 80813ece886SAndrea Arcangeli Selects the sysfs defaults for Transparent Hugepage Support. 80913ece886SAndrea Arcangeli 81013ece886SAndrea Arcangeli config TRANSPARENT_HUGEPAGE_ALWAYS 81113ece886SAndrea Arcangeli bool "always" 81213ece886SAndrea Arcangeli help 81313ece886SAndrea Arcangeli Enabling Transparent Hugepage always, can increase the 81413ece886SAndrea Arcangeli memory footprint of applications without a guaranteed 81513ece886SAndrea Arcangeli benefit but it will work automatically for all applications. 81613ece886SAndrea Arcangeli 81713ece886SAndrea Arcangeli config TRANSPARENT_HUGEPAGE_MADVISE 81813ece886SAndrea Arcangeli bool "madvise" 81913ece886SAndrea Arcangeli help 82013ece886SAndrea Arcangeli Enabling Transparent Hugepage madvise, will only provide a 82113ece886SAndrea Arcangeli performance improvement benefit to the applications using 82213ece886SAndrea Arcangeli madvise(MADV_HUGEPAGE) but it won't risk to increase the 82313ece886SAndrea Arcangeli memory footprint of applications without a guaranteed 82413ece886SAndrea Arcangeli benefit. 82513ece886SAndrea Arcangeliendchoice 82613ece886SAndrea Arcangeli 82738d8b4e6SHuang Yingconfig THP_SWAP 82838d8b4e6SHuang Ying def_bool y 829dad6a5ebSHugh Dickins depends on TRANSPARENT_HUGEPAGE && ARCH_WANTS_THP_SWAP && SWAP && 64BIT 83038d8b4e6SHuang Ying help 83138d8b4e6SHuang Ying Swap transparent huge pages in one piece, without splitting. 83214fef284SHuang Ying XXX: For now, swap cluster backing transparent huge page 83314fef284SHuang Ying will be split after swapout. 83438d8b4e6SHuang Ying 83538d8b4e6SHuang Ying For selection by architectures with reasonable THP sizes. 83638d8b4e6SHuang Ying 837519bcb79SJohannes Weinerconfig READ_ONLY_THP_FOR_FS 838519bcb79SJohannes Weiner bool "Read-only THP for filesystems (EXPERIMENTAL)" 839519bcb79SJohannes Weiner depends on TRANSPARENT_HUGEPAGE && SHMEM 840519bcb79SJohannes Weiner 841519bcb79SJohannes Weiner help 842519bcb79SJohannes Weiner Allow khugepaged to put read-only file-backed pages in THP. 843519bcb79SJohannes Weiner 844519bcb79SJohannes Weiner This is marked experimental because it is a new feature. Write 845519bcb79SJohannes Weiner support of file THPs will be developed in the next few release 846519bcb79SJohannes Weiner cycles. 847519bcb79SJohannes Weiner 848519bcb79SJohannes Weinerendif # TRANSPARENT_HUGEPAGE 849519bcb79SJohannes Weiner 850e496cf3dSKirill A. Shutemov# 851bbddff05STejun Heo# UP and nommu archs use km based percpu allocator 852bbddff05STejun Heo# 853bbddff05STejun Heoconfig NEED_PER_CPU_KM 8543583521aSVladimir Murzin depends on !SMP || !MMU 855bbddff05STejun Heo bool 856bbddff05STejun Heo default y 857077b1f83SDan Magenheimer 8587ecd19cfSKefeng Wangconfig NEED_PER_CPU_EMBED_FIRST_CHUNK 8597ecd19cfSKefeng Wang bool 8607ecd19cfSKefeng Wang 8617ecd19cfSKefeng Wangconfig NEED_PER_CPU_PAGE_FIRST_CHUNK 8627ecd19cfSKefeng Wang bool 8637ecd19cfSKefeng Wang 8647ecd19cfSKefeng Wangconfig USE_PERCPU_NUMA_NODE_ID 8657ecd19cfSKefeng Wang bool 8667ecd19cfSKefeng Wang 8677ecd19cfSKefeng Wangconfig HAVE_SETUP_PER_CPU_AREA 8687ecd19cfSKefeng Wang bool 8697ecd19cfSKefeng Wang 87027c6aec2SDan Magenheimerconfig FRONTSWAP 8716e61dde8SChristoph Hellwig bool 872f825c736SAneesh Kumar K.V 873f825c736SAneesh Kumar K.Vconfig CMA 874f825c736SAneesh Kumar K.V bool "Contiguous Memory Allocator" 875aca52c39SMike Rapoport depends on MMU 876f825c736SAneesh Kumar K.V select MIGRATION 877f825c736SAneesh Kumar K.V select MEMORY_ISOLATION 878f825c736SAneesh Kumar K.V help 879f825c736SAneesh Kumar K.V This enables the Contiguous Memory Allocator which allows other 880f825c736SAneesh Kumar K.V subsystems to allocate big physically-contiguous blocks of memory. 881f825c736SAneesh Kumar K.V CMA reserves a region of memory and allows only movable pages to 882f825c736SAneesh Kumar K.V be allocated from it. This way, the kernel can use the memory for 883f825c736SAneesh Kumar K.V pagecache and when a subsystem requests for contiguous area, the 884f825c736SAneesh Kumar K.V allocated pages are migrated away to serve the contiguous request. 885f825c736SAneesh Kumar K.V 886f825c736SAneesh Kumar K.V If unsure, say "n". 887f825c736SAneesh Kumar K.V 888f825c736SAneesh Kumar K.Vconfig CMA_DEBUG 889f825c736SAneesh Kumar K.V bool "CMA debug messages (DEVELOPMENT)" 890f825c736SAneesh Kumar K.V depends on DEBUG_KERNEL && CMA 891f825c736SAneesh Kumar K.V help 892f825c736SAneesh Kumar K.V Turns on debug messages in CMA. This produces KERN_DEBUG 893f825c736SAneesh Kumar K.V messages for every CMA call as well as various messages while 894f825c736SAneesh Kumar K.V processing calls such as dma_alloc_from_contiguous(). 895f825c736SAneesh Kumar K.V This option does not affect warning and error messages. 896bf550fc9SAlexander Graf 89728b24c1fSSasha Levinconfig CMA_DEBUGFS 89828b24c1fSSasha Levin bool "CMA debugfs interface" 89928b24c1fSSasha Levin depends on CMA && DEBUG_FS 90028b24c1fSSasha Levin help 90128b24c1fSSasha Levin Turns on the DebugFS interface for CMA. 90228b24c1fSSasha Levin 90343ca106fSMinchan Kimconfig CMA_SYSFS 90443ca106fSMinchan Kim bool "CMA information through sysfs interface" 90543ca106fSMinchan Kim depends on CMA && SYSFS 90643ca106fSMinchan Kim help 90743ca106fSMinchan Kim This option exposes some sysfs attributes to get information 90843ca106fSMinchan Kim from CMA. 90943ca106fSMinchan Kim 910a254129eSJoonsoo Kimconfig CMA_AREAS 911a254129eSJoonsoo Kim int "Maximum count of the CMA areas" 912a254129eSJoonsoo Kim depends on CMA 913b7176c26SBarry Song default 19 if NUMA 914a254129eSJoonsoo Kim default 7 915a254129eSJoonsoo Kim help 916a254129eSJoonsoo Kim CMA allows to create CMA areas for particular purpose, mainly, 917a254129eSJoonsoo Kim used as device private area. This parameter sets the maximum 918a254129eSJoonsoo Kim number of CMA area in the system. 919a254129eSJoonsoo Kim 920b7176c26SBarry Song If unsure, leave the default value "7" in UMA and "19" in NUMA. 921a254129eSJoonsoo Kim 922af8d417aSDan Streetmanconfig MEM_SOFT_DIRTY 923af8d417aSDan Streetman bool "Track memory changes" 924af8d417aSDan Streetman depends on CHECKPOINT_RESTORE && HAVE_ARCH_SOFT_DIRTY && PROC_FS 925af8d417aSDan Streetman select PROC_PAGE_MONITOR 9264e2e2770SSeth Jennings help 927af8d417aSDan Streetman This option enables memory changes tracking by introducing a 928af8d417aSDan Streetman soft-dirty bit on pte-s. This bit it set when someone writes 929af8d417aSDan Streetman into a page just as regular dirty bit, but unlike the latter 930af8d417aSDan Streetman it can be cleared by hands. 931af8d417aSDan Streetman 9321ad1335dSMike Rapoport See Documentation/admin-guide/mm/soft-dirty.rst for more details. 9334e2e2770SSeth Jennings 9349e5c33d7SMark Salterconfig GENERIC_EARLY_IOREMAP 9359e5c33d7SMark Salter bool 936042d27acSHelge Deller 93722ee3ea5SHelge Dellerconfig STACK_MAX_DEFAULT_SIZE_MB 93822ee3ea5SHelge Deller int "Default maximum user stack size for 32-bit processes (MB)" 93922ee3ea5SHelge Deller default 100 940042d27acSHelge Deller range 8 2048 941042d27acSHelge Deller depends on STACK_GROWSUP && (!64BIT || COMPAT) 942042d27acSHelge Deller help 943042d27acSHelge Deller This is the maximum stack size in Megabytes in the VM layout of 32-bit 944042d27acSHelge Deller user processes when the stack grows upwards (currently only on parisc 94522ee3ea5SHelge Deller arch) when the RLIMIT_STACK hard limit is unlimited. 946042d27acSHelge Deller 94722ee3ea5SHelge Deller A sane initial value is 100 MB. 9483a80a7faSMel Gorman 9493a80a7faSMel Gormanconfig DEFERRED_STRUCT_PAGE_INIT 9501ce22103SVlastimil Babka bool "Defer initialisation of struct pages to kthreads" 951d39f8fb4SMike Rapoport depends on SPARSEMEM 952ab1e8d89SPavel Tatashin depends on !NEED_PER_CPU_KM 953889c695dSPasha Tatashin depends on 64BIT 954e4443149SDaniel Jordan select PADATA 9553a80a7faSMel Gorman help 9563a80a7faSMel Gorman Ordinarily all struct pages are initialised during early boot in a 9573a80a7faSMel Gorman single thread. On very large machines this can take a considerable 9583a80a7faSMel Gorman amount of time. If this option is set, large machines will bring up 959e4443149SDaniel Jordan a subset of memmap at boot and then initialise the rest in parallel. 960e4443149SDaniel Jordan This has a potential performance impact on tasks running early in the 9611ce22103SVlastimil Babka lifetime of the system until these kthreads finish the 9621ce22103SVlastimil Babka initialisation. 963033fbae9SDan Williams 9641c676e0dSSeongJae Parkconfig PAGE_IDLE_FLAG 9651c676e0dSSeongJae Park bool 9661c676e0dSSeongJae Park select PAGE_EXTENSION if !64BIT 9671c676e0dSSeongJae Park help 9681c676e0dSSeongJae Park This adds PG_idle and PG_young flags to 'struct page'. PTE Accessed 9691c676e0dSSeongJae Park bit writers can set the state of the bit in the flags so that PTE 9701c676e0dSSeongJae Park Accessed bit readers may avoid disturbance. 9711c676e0dSSeongJae Park 97233c3fc71SVladimir Davydovconfig IDLE_PAGE_TRACKING 97333c3fc71SVladimir Davydov bool "Enable idle page tracking" 97433c3fc71SVladimir Davydov depends on SYSFS && MMU 9751c676e0dSSeongJae Park select PAGE_IDLE_FLAG 97633c3fc71SVladimir Davydov help 97733c3fc71SVladimir Davydov This feature allows to estimate the amount of user pages that have 97833c3fc71SVladimir Davydov not been touched during a given period of time. This information can 97933c3fc71SVladimir Davydov be useful to tune memory cgroup limits and/or for job placement 98033c3fc71SVladimir Davydov within a compute cluster. 98133c3fc71SVladimir Davydov 9821ad1335dSMike Rapoport See Documentation/admin-guide/mm/idle_page_tracking.rst for 9831ad1335dSMike Rapoport more details. 98433c3fc71SVladimir Davydov 985c2280be8SAnshuman Khandualconfig ARCH_HAS_CACHE_LINE_SIZE 986c2280be8SAnshuman Khandual bool 987c2280be8SAnshuman Khandual 9882792d84eSKees Cookconfig ARCH_HAS_CURRENT_STACK_POINTER 9892792d84eSKees Cook bool 9902792d84eSKees Cook help 9912792d84eSKees Cook In support of HARDENED_USERCOPY performing stack variable lifetime 9922792d84eSKees Cook checking, an architecture-agnostic way to find the stack pointer 9932792d84eSKees Cook is needed. Once an architecture defines an unsigned long global 9942792d84eSKees Cook register alias named "current_stack_pointer", this config can be 9952792d84eSKees Cook selected. 9962792d84eSKees Cook 99717596731SRobin Murphyconfig ARCH_HAS_PTE_DEVMAP 99865f7d049SOliver O'Halloran bool 99965f7d049SOliver O'Halloran 100063703f37SKefeng Wangconfig ARCH_HAS_ZONE_DMA_SET 100163703f37SKefeng Wang bool 100263703f37SKefeng Wang 100363703f37SKefeng Wangconfig ZONE_DMA 100463703f37SKefeng Wang bool "Support DMA zone" if ARCH_HAS_ZONE_DMA_SET 100563703f37SKefeng Wang default y if ARM64 || X86 100663703f37SKefeng Wang 100763703f37SKefeng Wangconfig ZONE_DMA32 100863703f37SKefeng Wang bool "Support DMA32 zone" if ARCH_HAS_ZONE_DMA_SET 100963703f37SKefeng Wang depends on !X86_32 101063703f37SKefeng Wang default y if ARM64 101163703f37SKefeng Wang 1012033fbae9SDan Williamsconfig ZONE_DEVICE 10135042db43SJérôme Glisse bool "Device memory (pmem, HMM, etc...) hotplug support" 1014033fbae9SDan Williams depends on MEMORY_HOTPLUG 1015033fbae9SDan Williams depends on MEMORY_HOTREMOVE 101699490f16SDan Williams depends on SPARSEMEM_VMEMMAP 101717596731SRobin Murphy depends on ARCH_HAS_PTE_DEVMAP 10183a08cd52SMatthew Wilcox select XARRAY_MULTI 1019033fbae9SDan Williams 1020033fbae9SDan Williams help 1021033fbae9SDan Williams Device memory hotplug support allows for establishing pmem, 1022033fbae9SDan Williams or other device driver discovered memory regions, in the 1023033fbae9SDan Williams memmap. This allows pfn_to_page() lookups of otherwise 1024033fbae9SDan Williams "device-physical" addresses which is needed for using a DAX 1025033fbae9SDan Williams mapping in an O_DIRECT operation, among other things. 1026033fbae9SDan Williams 1027033fbae9SDan Williams If FS_DAX is enabled, then say Y. 102806a660adSLinus Torvalds 10299c240a7bSChristoph Hellwig# 10309c240a7bSChristoph Hellwig# Helpers to mirror range of the CPU page tables of a process into device page 10319c240a7bSChristoph Hellwig# tables. 10329c240a7bSChristoph Hellwig# 1033c0b12405SJérôme Glisseconfig HMM_MIRROR 10349c240a7bSChristoph Hellwig bool 1035f442c283SChristoph Hellwig depends on MMU 1036c0b12405SJérôme Glisse 103714b80582SDan Williamsconfig GET_FREE_REGION 103814b80582SDan Williams depends on SPARSEMEM 103914b80582SDan Williams bool 104014b80582SDan Williams 10415042db43SJérôme Glisseconfig DEVICE_PRIVATE 10425042db43SJérôme Glisse bool "Unaddressable device memory (GPU memory, ...)" 10437328d9ccSChristoph Hellwig depends on ZONE_DEVICE 104414b80582SDan Williams select GET_FREE_REGION 10455042db43SJérôme Glisse 10465042db43SJérôme Glisse help 10475042db43SJérôme Glisse Allows creation of struct pages to represent unaddressable device 10485042db43SJérôme Glisse memory; i.e., memory that is only accessible from the device (or 10495042db43SJérôme Glisse group of devices). You likely also want to select HMM_MIRROR. 10505042db43SJérôme Glisse 10513e9a9e25SChristoph Hellwigconfig VMAP_PFN 10523e9a9e25SChristoph Hellwig bool 10533e9a9e25SChristoph Hellwig 105463c17fb8SDave Hansenconfig ARCH_USES_HIGH_VMA_FLAGS 105563c17fb8SDave Hansen bool 105666d37570SDave Hansenconfig ARCH_HAS_PKEYS 105766d37570SDave Hansen bool 105830a5b536SDennis Zhou 1059b0284cd2SCatalin Marinasconfig ARCH_USES_PG_ARCH_X 1060b0284cd2SCatalin Marinas bool 1061b0284cd2SCatalin Marinas help 1062b0284cd2SCatalin Marinas Enable the definition of PG_arch_x page flags with x > 1. Only 1063b0284cd2SCatalin Marinas suitable for 64-bit architectures with CONFIG_FLATMEM or 1064b0284cd2SCatalin Marinas CONFIG_SPARSEMEM_VMEMMAP enabled, otherwise there may not be 1065b0284cd2SCatalin Marinas enough room for additional bits in page->flags. 1066b0284cd2SCatalin Marinas 10670710d012SVlastimil Babkaconfig VM_EVENT_COUNTERS 10680710d012SVlastimil Babka default y 10690710d012SVlastimil Babka bool "Enable VM event counters for /proc/vmstat" if EXPERT 10700710d012SVlastimil Babka help 10710710d012SVlastimil Babka VM event counters are needed for event counts to be shown. 10720710d012SVlastimil Babka This option allows the disabling of the VM event counters 10730710d012SVlastimil Babka on EXPERT systems. /proc/vmstat will only show page counts 10740710d012SVlastimil Babka if VM event counters are disabled. 10750710d012SVlastimil Babka 107630a5b536SDennis Zhouconfig PERCPU_STATS 107730a5b536SDennis Zhou bool "Collect percpu memory statistics" 107830a5b536SDennis Zhou help 107930a5b536SDennis Zhou This feature collects and exposes statistics via debugfs. The 108030a5b536SDennis Zhou information includes global and per chunk statistics, which can 108130a5b536SDennis Zhou be used to help understand percpu memory usage. 108264c349f4SKirill A. Shutemov 10839c84f229SJohn Hubbardconfig GUP_TEST 10849c84f229SJohn Hubbard bool "Enable infrastructure for get_user_pages()-related unit tests" 1085d0de8241SBarry Song depends on DEBUG_FS 108664c349f4SKirill A. Shutemov help 10879c84f229SJohn Hubbard Provides /sys/kernel/debug/gup_test, which in turn provides a way 10889c84f229SJohn Hubbard to make ioctl calls that can launch kernel-based unit tests for 10899c84f229SJohn Hubbard the get_user_pages*() and pin_user_pages*() family of API calls. 109064c349f4SKirill A. Shutemov 10919c84f229SJohn Hubbard These tests include benchmark testing of the _fast variants of 10929c84f229SJohn Hubbard get_user_pages*() and pin_user_pages*(), as well as smoke tests of 10939c84f229SJohn Hubbard the non-_fast variants. 10949c84f229SJohn Hubbard 1095f4f9bda4SJohn Hubbard There is also a sub-test that allows running dump_page() on any 1096f4f9bda4SJohn Hubbard of up to eight pages (selected by command line args) within the 1097f4f9bda4SJohn Hubbard range of user-space addresses. These pages are either pinned via 1098f4f9bda4SJohn Hubbard pin_user_pages*(), or pinned via get_user_pages*(), as specified 1099f4f9bda4SJohn Hubbard by other command line arguments. 1100f4f9bda4SJohn Hubbard 1101baa489faSSeongJae Park See tools/testing/selftests/mm/gup_test.c 11023010a5eaSLaurent Dufour 1103d0de8241SBarry Songcomment "GUP_TEST needs to have DEBUG_FS enabled" 1104d0de8241SBarry Song depends on !GUP_TEST && !DEBUG_FS 11053010a5eaSLaurent Dufour 11066ca297d4SPeter Zijlstraconfig GUP_GET_PXX_LOW_HIGH 110739656e83SChristoph Hellwig bool 110839656e83SChristoph Hellwig 1109def85743SKeith Buschconfig DMAPOOL_TEST 1110def85743SKeith Busch tristate "Enable a module to run time tests on dma_pool" 1111def85743SKeith Busch depends on HAS_DMA 1112def85743SKeith Busch help 1113def85743SKeith Busch Provides a test module that will allocate and free many blocks of 1114def85743SKeith Busch various sizes and report how long it takes. This is intended to 1115def85743SKeith Busch provide a consistent way to measure how changes to the 1116def85743SKeith Busch dma_pool_alloc/free routines affect performance. 1117def85743SKeith Busch 11183010a5eaSLaurent Dufourconfig ARCH_HAS_PTE_SPECIAL 11193010a5eaSLaurent Dufour bool 112059e0b520SChristoph Hellwig 1121cbd34da7SChristoph Hellwig# 1122cbd34da7SChristoph Hellwig# Some architectures require a special hugepage directory format that is 1123cbd34da7SChristoph Hellwig# required to support multiple hugepage sizes. For example a4fe3ce76 1124cbd34da7SChristoph Hellwig# "powerpc/mm: Allow more flexible layouts for hugepage pagetables" 1125cbd34da7SChristoph Hellwig# introduced it on powerpc. This allows for a more flexible hugepage 1126cbd34da7SChristoph Hellwig# pagetable layouts. 1127cbd34da7SChristoph Hellwig# 1128cbd34da7SChristoph Hellwigconfig ARCH_HAS_HUGEPD 1129cbd34da7SChristoph Hellwig bool 1130cbd34da7SChristoph Hellwig 1131c5acad84SThomas Hellstromconfig MAPPING_DIRTY_HELPERS 1132c5acad84SThomas Hellstrom bool 1133c5acad84SThomas Hellstrom 1134298fa1adSThomas Gleixnerconfig KMAP_LOCAL 1135298fa1adSThomas Gleixner bool 1136298fa1adSThomas Gleixner 1137825c43f5SArd Biesheuvelconfig KMAP_LOCAL_NON_LINEAR_PTE_ARRAY 1138825c43f5SArd Biesheuvel bool 1139825c43f5SArd Biesheuvel 11401fbaf8fcSChristoph Hellwig# struct io_mapping based helper. Selected by drivers that need them 11411fbaf8fcSChristoph Hellwigconfig IO_MAPPING 11421fbaf8fcSChristoph Hellwig bool 11431507f512SMike Rapoport 11441507f512SMike Rapoportconfig SECRETMEM 114574947724SLukas Bulwahn default y 114674947724SLukas Bulwahn bool "Enable memfd_secret() system call" if EXPERT 114774947724SLukas Bulwahn depends on ARCH_HAS_SET_DIRECT_MAP 114874947724SLukas Bulwahn help 114974947724SLukas Bulwahn Enable the memfd_secret() system call with the ability to create 115074947724SLukas Bulwahn memory areas visible only in the context of the owning process and 115174947724SLukas Bulwahn not mapped to other processes and other kernel page tables. 11521507f512SMike Rapoport 11539a10064fSColin Crossconfig ANON_VMA_NAME 11549a10064fSColin Cross bool "Anonymous VMA name support" 11559a10064fSColin Cross depends on PROC_FS && ADVISE_SYSCALLS && MMU 11569a10064fSColin Cross 11579a10064fSColin Cross help 11589a10064fSColin Cross Allow naming anonymous virtual memory areas. 11599a10064fSColin Cross 11609a10064fSColin Cross This feature allows assigning names to virtual memory areas. Assigned 11619a10064fSColin Cross names can be later retrieved from /proc/pid/maps and /proc/pid/smaps 11629a10064fSColin Cross and help identifying individual anonymous memory areas. 11639a10064fSColin Cross Assigning a name to anonymous virtual memory area might prevent that 11649a10064fSColin Cross area from being merged with adjacent virtual memory areas due to the 11659a10064fSColin Cross difference in their name. 11669a10064fSColin Cross 1167430529b5SPeter Xuconfig USERFAULTFD 1168430529b5SPeter Xu bool "Enable userfaultfd() system call" 1169430529b5SPeter Xu depends on MMU 1170430529b5SPeter Xu help 1171430529b5SPeter Xu Enable the userfaultfd() system call that allows to intercept and 1172430529b5SPeter Xu handle page faults in userland. 1173430529b5SPeter Xu 1174430529b5SPeter Xuconfig HAVE_ARCH_USERFAULTFD_WP 1175430529b5SPeter Xu bool 1176430529b5SPeter Xu help 1177430529b5SPeter Xu Arch has userfaultfd write protection support 1178430529b5SPeter Xu 1179430529b5SPeter Xuconfig HAVE_ARCH_USERFAULTFD_MINOR 1180430529b5SPeter Xu bool 1181430529b5SPeter Xu help 1182430529b5SPeter Xu Arch has userfaultfd minor fault support 1183430529b5SPeter Xu 11841db9dbc2SPeter Xuconfig PTE_MARKER_UFFD_WP 118581e0f15fSPeter Xu bool "Userfaultfd write protection support for shmem/hugetlbfs" 118681e0f15fSPeter Xu default y 118781e0f15fSPeter Xu depends on HAVE_ARCH_USERFAULTFD_WP 11881db9dbc2SPeter Xu 11891db9dbc2SPeter Xu help 11901db9dbc2SPeter Xu Allows to create marker PTEs for userfaultfd write protection 11911db9dbc2SPeter Xu purposes. It is required to enable userfaultfd write protection on 11921db9dbc2SPeter Xu file-backed memory types like shmem and hugetlbfs. 11931db9dbc2SPeter Xu 1194ac35a490SYu Zhao# multi-gen LRU { 1195ec1c86b2SYu Zhaoconfig LRU_GEN 1196ec1c86b2SYu Zhao bool "Multi-Gen LRU" 1197ec1c86b2SYu Zhao depends on MMU 1198ec1c86b2SYu Zhao # make sure folio->flags has enough spare bits 1199ec1c86b2SYu Zhao depends on 64BIT || !SPARSEMEM || SPARSEMEM_VMEMMAP 1200ec1c86b2SYu Zhao help 120107017acbSYu Zhao A high performance LRU implementation to overcommit memory. See 120207017acbSYu Zhao Documentation/admin-guide/mm/multigen_lru.rst for details. 1203ec1c86b2SYu Zhao 1204354ed597SYu Zhaoconfig LRU_GEN_ENABLED 1205354ed597SYu Zhao bool "Enable by default" 1206354ed597SYu Zhao depends on LRU_GEN 1207354ed597SYu Zhao help 1208354ed597SYu Zhao This option enables the multi-gen LRU by default. 1209354ed597SYu Zhao 1210ac35a490SYu Zhaoconfig LRU_GEN_STATS 1211ac35a490SYu Zhao bool "Full stats for debugging" 1212ac35a490SYu Zhao depends on LRU_GEN 1213ac35a490SYu Zhao help 1214ac35a490SYu Zhao Do not enable this option unless you plan to look at historical stats 1215ac35a490SYu Zhao from evicted generations for debugging purpose. 1216ac35a490SYu Zhao 1217ac35a490SYu Zhao This option has a per-memcg and per-node memory overhead. 1218ac35a490SYu Zhao# } 1219ac35a490SYu Zhao 12200b6cc04fSSuren Baghdasaryanconfig ARCH_SUPPORTS_PER_VMA_LOCK 12210b6cc04fSSuren Baghdasaryan def_bool n 12220b6cc04fSSuren Baghdasaryan 12230b6cc04fSSuren Baghdasaryanconfig PER_VMA_LOCK 12240b6cc04fSSuren Baghdasaryan def_bool y 12250b6cc04fSSuren Baghdasaryan depends on ARCH_SUPPORTS_PER_VMA_LOCK && MMU && SMP 12260b6cc04fSSuren Baghdasaryan help 12270b6cc04fSSuren Baghdasaryan Allow per-vma locking during page fault handling. 12280b6cc04fSSuren Baghdasaryan 12290b6cc04fSSuren Baghdasaryan This feature allows locking each virtual memory area separately when 12300b6cc04fSSuren Baghdasaryan handling page faults instead of taking mmap_lock. 12310b6cc04fSSuren Baghdasaryan 12322224d848SSeongJae Parksource "mm/damon/Kconfig" 12332224d848SSeongJae Park 123459e0b520SChristoph Hellwigendmenu 1235