mm/Kconfig

   1 config SELECT_MEMORY_MODEL
   2         def_bool y
   3         depends on ARCH_SELECT_MEMORY_MODEL
   4
   5 choice
   6         prompt "Memory model"
   7         depends on SELECT_MEMORY_MODEL
   8         default DISCONTIGMEM_MANUAL if ARCH_DISCONTIGMEM_DEFAULT
   9         default SPARSEMEM_MANUAL if ARCH_SPARSEMEM_DEFAULT
  10         default FLATMEM_MANUAL
  11
  12 config FLATMEM_MANUAL
  13         bool "Flat Memory"
  14         depends on !(ARCH_DISCONTIGMEM_ENABLE || ARCH_SPARSEMEM_ENABLE) || ARCH_FLATMEM_ENABLE
  15         help
  16           This option allows you to change some of the ways that
  17           Linux manages its memory internally.  Most users will
  18           only have one option here: FLATMEM.  This is normal
  19           and a correct option.
  20
  21           Some users of more advanced features like NUMA and
  22           memory hotplug may have different options here.
  23           DISCONTIGMEM is a more mature, better tested system,
  24           but is incompatible with memory hotplug and may suffer
  25           decreased performance over SPARSEMEM.  If unsure between
  26           "Sparse Memory" and "Discontiguous Memory", choose
  27           "Discontiguous Memory".
  28
  29           If unsure, choose this option (Flat Memory) over any other.
  30
  31 config DISCONTIGMEM_MANUAL
  32         bool "Discontiguous Memory"
  33         depends on ARCH_DISCONTIGMEM_ENABLE
  34         help
  35           This option provides enhanced support for discontiguous
  36           memory systems, over FLATMEM.  These systems have holes
  37           in their physical address spaces, and this option provides
  38           more efficient handling of these holes.  However, the vast
  39           majority of hardware has quite flat address spaces, and
  40           can have degraded performance from the extra overhead that
  41           this option imposes.
  42
  43           Many NUMA configurations will have this as the only option.
  44
  45           If unsure, choose "Flat Memory" over this option.
  46
  47 config SPARSEMEM_MANUAL
  48         bool "Sparse Memory"
  49         depends on ARCH_SPARSEMEM_ENABLE
  50         help
  51           This will be the only option for some systems, including
  52           memory hotplug systems.  This is normal.
  53
  54           For many other systems, this will be an alternative to
  55           "Discontiguous Memory".  This option provides some potential
  56           performance benefits, along with decreased code complexity,
  57           but it is newer, and more experimental.
  58
  59           If unsure, choose "Discontiguous Memory" or "Flat Memory"
  60           over this option.
  61
  62 endchoice
  63
  64 config DISCONTIGMEM
  65         def_bool y
  66         depends on (!SELECT_MEMORY_MODEL && ARCH_DISCONTIGMEM_ENABLE) || DISCONTIGMEM_MANUAL
  67
  68 config SPARSEMEM
  69         def_bool y
  70         depends on (!SELECT_MEMORY_MODEL && ARCH_SPARSEMEM_ENABLE) || SPARSEMEM_MANUAL
  71
  72 config FLATMEM
  73         def_bool y
  74         depends on (!DISCONTIGMEM && !SPARSEMEM) || FLATMEM_MANUAL
  75
  76 config FLAT_NODE_MEM_MAP
  77         def_bool y
  78         depends on !SPARSEMEM
  79
  80 #
  81 # Both the NUMA code and DISCONTIGMEM use arrays of pg_data_t's
  82 # to represent different areas of memory.  This variable allows
  83 # those dependencies to exist individually.
  84 #
  85 config NEED_MULTIPLE_NODES
  86         def_bool y
  87         depends on DISCONTIGMEM || NUMA
  88
  89 config HAVE_MEMORY_PRESENT
  90         def_bool y
  91         depends on ARCH_HAVE_MEMORY_PRESENT || SPARSEMEM
  92
  93 #
  94 # SPARSEMEM_EXTREME (which is the default) does some bootmem
  95 # allocations when memory_present() is called.  If this cannot
  96 # be done on your architecture, select this option.  However,
  97 # statically allocating the mem_section[] array can potentially
  98 # consume vast quantities of .bss, so be careful.
  99 #
 100 # This option will also potentially produce smaller runtime code
 101 # with gcc 3.4 and later.
 102 #
 103 config SPARSEMEM_STATIC
 104         bool
 105
 106 #
 107 # Architecture platforms which require a two level mem_section in SPARSEMEM
 108 # must select this option. This is usually for architecture platforms with
 109 # an extremely sparse physical address space.
 110 #
 111 config SPARSEMEM_EXTREME
 112         def_bool y
 113         depends on SPARSEMEM && !SPARSEMEM_STATIC
 114
 115 config SPARSEMEM_VMEMMAP_ENABLE
 116         bool
 117
 118 config SPARSEMEM_ALLOC_MEM_MAP_TOGETHER
 119         def_bool y
 120         depends on SPARSEMEM && X86_64
 121
 122 config SPARSEMEM_VMEMMAP
 123         bool "Sparse Memory virtual memmap"
 124         depends on SPARSEMEM && SPARSEMEM_VMEMMAP_ENABLE
 125         default y
 126         help
 127          SPARSEMEM_VMEMMAP uses a virtually mapped memmap to optimise
 128          pfn_to_page and page_to_pfn operations.  This is the most
 129          efficient option when sufficient kernel resources are available.
 130
 131 config HAVE_MEMBLOCK
 132         boolean
 133
 134 config HAVE_MEMBLOCK_NODE_MAP
 135         boolean
 136
 137 config HAVE_MEMBLOCK_PHYS_MAP
 138         boolean
 139
 140 config HAVE_GENERIC_RCU_GUP
 141         boolean
 142
 143 config ARCH_DISCARD_MEMBLOCK
 144         boolean
 145
 146 config NO_BOOTMEM
 147         boolean
 148
 149 config MEMORY_ISOLATION
 150         boolean
 151
 152 config MOVABLE_NODE
 153         boolean "Enable to assign a node which has only movable memory"
 154         depends on HAVE_MEMBLOCK
 155         depends on NO_BOOTMEM
 156         depends on X86_64
 157         depends on NUMA
 158         default n
 159         help
 160           Allow a node to have only movable memory.  Pages used by the kernel,
 161           such as direct mapping pages cannot be migrated.  So the corresponding
 162           memory device cannot be hotplugged.  This option allows the following
 163           two things:
 164           - When the system is booting, node full of hotpluggable memory can
 165           be arranged to have only movable memory so that the whole node can
 166           be hot-removed. (need movable_node boot option specified).
 167           - After the system is up, the option allows users to online all the
 168           memory of a node as movable memory so that the whole node can be
 169           hot-removed.
 170
 171           Users who don't use the memory hotplug feature are fine with this
 172           option on since they don't specify movable_node boot option or they
 173           don't online memory as movable.
 174
 175           Say Y here if you want to hotplug a whole node.
 176           Say N here if you want kernel to use memory on all nodes evenly.
 177
 178 #
 179 # Only be set on architectures that have completely implemented memory hotplug
 180 # feature. If you are not sure, don't touch it.
 181 #
 182 config HAVE_BOOTMEM_INFO_NODE
 183         def_bool n
 184
 185 # eventually, we can have this option just 'select SPARSEMEM'
 186 config MEMORY_HOTPLUG
 187         bool "Allow for memory hot-add"
 188         depends on SPARSEMEM || X86_64_ACPI_NUMA
 189         depends on ARCH_ENABLE_MEMORY_HOTPLUG
 190         depends on (IA64 || X86 || PPC_BOOK3S_64 || SUPERH || S390)
 191
 192 config MEMORY_HOTPLUG_SPARSE
 193         def_bool y
 194         depends on SPARSEMEM && MEMORY_HOTPLUG
 195
 196 config MEMORY_HOTREMOVE
 197         bool "Allow for memory hot remove"
 198         select MEMORY_ISOLATION
 199         select HAVE_BOOTMEM_INFO_NODE if (X86_64 || PPC64)
 200         depends on MEMORY_HOTPLUG && ARCH_ENABLE_MEMORY_HOTREMOVE
 201         depends on MIGRATION
 202
 203 #
 204 # If we have space for more page flags then we can enable additional
 205 # optimizations and functionality.
 206 #
 207 # Regular Sparsemem takes page flag bits for the sectionid if it does not
 208 # use a virtual memmap. Disable extended page flags for 32 bit platforms
 209 # that require the use of a sectionid in the page flags.
 210 #
 211 config PAGEFLAGS_EXTENDED
 212         def_bool y
 213         depends on 64BIT || SPARSEMEM_VMEMMAP || !SPARSEMEM
 214
 215 # Heavily threaded applications may benefit from splitting the mm-wide
 216 # page_table_lock, so that faults on different parts of the user address
 217 # space can be handled with less contention: split it at this NR_CPUS.
 218 # Default to 4 for wider testing, though 8 might be more appropriate.
 219 # ARM's adjust_pte (unused if VIPT) depends on mm-wide page_table_lock.
 220 # PA-RISC 7xxx's spinlock_t would enlarge struct page from 32 to 44 bytes.
 221 # DEBUG_SPINLOCK and DEBUG_LOCK_ALLOC spinlock_t also enlarge struct page.
 222 #
 223 config SPLIT_PTLOCK_CPUS
 224         int
 225         default "999999" if !MMU
 226         default "999999" if ARM && !CPU_CACHE_VIPT
 227         default "999999" if PARISC && !PA20
 228         default "4"
 229
 230 config ARCH_ENABLE_SPLIT_PMD_PTLOCK
 231         boolean
 232
 233 #
 234 # support for memory balloon
 235 config MEMORY_BALLOON
 236         boolean
 237
 238 #
 239 # support for memory balloon compaction
 240 config BALLOON_COMPACTION
 241         bool "Allow for balloon memory compaction/migration"
 242         def_bool y
 243         depends on COMPACTION && MEMORY_BALLOON
 244         help
 245           Memory fragmentation introduced by ballooning might reduce
 246           significantly the number of 2MB contiguous memory blocks that can be
 247           used within a guest, thus imposing performance penalties associated
 248           with the reduced number of transparent huge pages that could be used
 249           by the guest workload. Allowing the compaction & migration for memory
 250           pages enlisted as being part of memory balloon devices avoids the
 251           scenario aforementioned and helps improving memory defragmentation.
 252
 253 #
 254 # support for memory compaction
 255 config COMPACTION
 256         bool "Allow for memory compaction"
 257         def_bool y
 258         select MIGRATION
 259         depends on MMU
 260         help
 261           Allows the compaction of memory for the allocation of huge pages.
 262
 263 #
 264 # support for page migration
 265 #
 266 config MIGRATION
 267         bool "Page migration"
 268         def_bool y
 269         depends on (NUMA || ARCH_ENABLE_MEMORY_HOTREMOVE || COMPACTION || CMA) && MMU
 270         help
 271           Allows the migration of the physical location of pages of processes
 272           while the virtual addresses are not changed. This is useful in
 273           two situations. The first is on NUMA systems to put pages nearer
 274           to the processors accessing. The second is when allocating huge
 275           pages as migration can relocate pages to satisfy a huge page
 276           allocation instead of reclaiming.
 277
 278 config ARCH_ENABLE_HUGEPAGE_MIGRATION
 279         boolean
 280
 281 config PHYS_ADDR_T_64BIT
 282         def_bool 64BIT || ARCH_PHYS_ADDR_T_64BIT
 283
 284 config ZONE_DMA_FLAG
 285         int
 286         default "0" if !ZONE_DMA
 287         default "1"
 288
 289 config BOUNCE
 290         bool "Enable bounce buffers"
 291         default y
 292         depends on BLOCK && MMU && (ZONE_DMA || HIGHMEM)
 293         help
 294           Enable bounce buffers for devices that cannot access
 295           the full range of memory available to the CPU. Enabled
 296           by default when ZONE_DMA or HIGHMEM is selected, but you
 297           may say n to override this.
 298
 299 # On the 'tile' arch, USB OHCI needs the bounce pool since tilegx will often
 300 # have more than 4GB of memory, but we don't currently use the IOTLB to present
 301 # a 32-bit address to OHCI.  So we need to use a bounce pool instead.
 302 #
 303 # We also use the bounce pool to provide stable page writes for jbd.  jbd
 304 # initiates buffer writeback without locking the page or setting PG_writeback,
 305 # and fixing that behavior (a second time; jbd2 doesn't have this problem) is
 306 # a major rework effort.  Instead, use the bounce buffer to snapshot pages
 307 # (until jbd goes away).  The only jbd user is ext3.
 308 config NEED_BOUNCE_POOL
 309         bool
 310         default y if (TILE && USB_OHCI_HCD) || (BLK_DEV_INTEGRITY && JBD)
 311
 312 config NR_QUICK
 313         int
 314         depends on QUICKLIST
 315         default "2" if AVR32
 316         default "1"
 317
 318 config VIRT_TO_BUS
 319         bool
 320         help
 321           An architecture should select this if it implements the
 322           deprecated interface virt_to_bus().  All new architectures
 323           should probably not select this.
 324
 325
 326 config MMU_NOTIFIER
 327         bool
 328
 329 config KSM
 330         bool "Enable KSM for page merging"
 331         depends on MMU
 332         help
 333           Enable Kernel Samepage Merging: KSM periodically scans those areas
 334           of an application's address space that an app has advised may be
 335           mergeable.  When it finds pages of identical content, it replaces
 336           the many instances by a single page with that content, so
 337           saving memory until one or another app needs to modify the content.
 338           Recommended for use with KVM, or with other duplicative applications.
 339           See Documentation/vm/ksm.txt for more information: KSM is inactive
 340           until a program has madvised that an area is MADV_MERGEABLE, and
 341           root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set).
 342
 343 config DEFAULT_MMAP_MIN_ADDR
 344         int "Low address space to protect from user allocation"
 345         depends on MMU
 346         default 4096
 347         help
 348           This is the portion of low virtual memory which should be protected
 349           from userspace allocation.  Keeping a user from writing to low pages
 350           can help reduce the impact of kernel NULL pointer bugs.
 351
 352           For most ia64, ppc64 and x86 users with lots of address space
 353           a value of 65536 is reasonable and should cause no problems.
 354           On arm and other archs it should not be higher than 32768.
 355           Programs which use vm86 functionality or have some need to map
 356           this low address space will need CAP_SYS_RAWIO or disable this
 357           protection by setting the value to 0.
 358
 359           This value can be changed after boot using the
 360           /proc/sys/vm/mmap_min_addr tunable.
 361
 362 config ARCH_SUPPORTS_MEMORY_FAILURE
 363         bool
 364
 365 config MEMORY_FAILURE
 366         depends on MMU
 367         depends on ARCH_SUPPORTS_MEMORY_FAILURE
 368         bool "Enable recovery from hardware memory errors"
 369         select MEMORY_ISOLATION
 370         help
 371           Enables code to recover from some memory failures on systems
 372           with MCA recovery. This allows a system to continue running
 373           even when some of its memory has uncorrected errors. This requires
 374           special hardware support and typically ECC memory.
 375
 376 config HWPOISON_INJECT
 377         tristate "HWPoison pages injector"
 378         depends on MEMORY_FAILURE && DEBUG_KERNEL && PROC_FS
 379         select PROC_PAGE_MONITOR
 380
 381 config NOMMU_INITIAL_TRIM_EXCESS
 382         int "Turn on mmap() excess space trimming before booting"
 383         depends on !MMU
 384         default 1
 385         help
 386           The NOMMU mmap() frequently needs to allocate large contiguous chunks
 387           of memory on which to store mappings, but it can only ask the system
 388           allocator for chunks in 2^N*PAGE_SIZE amounts - which is frequently
 389           more than it requires.  To deal with this, mmap() is able to trim off
 390           the excess and return it to the allocator.
 391
 392           If trimming is enabled, the excess is trimmed off and returned to the
 393           system allocator, which can cause extra fragmentation, particularly
 394           if there are a lot of transient processes.
 395
 396           If trimming is disabled, the excess is kept, but not used, which for
 397           long-term mappings means that the space is wasted.
 398
 399           Trimming can be dynamically controlled through a sysctl option
 400           (/proc/sys/vm/nr_trim_pages) which specifies the minimum number of
 401           excess pages there must be before trimming should occur, or zero if
 402           no trimming is to occur.
 403
 404           This option specifies the initial value of this option.  The default
 405           of 1 says that all excess pages should be trimmed.
 406
 407           See Documentation/nommu-mmap.txt for more information.
 408
 409 config TRANSPARENT_HUGEPAGE
 410         bool "Transparent Hugepage Support"
 411         depends on HAVE_ARCH_TRANSPARENT_HUGEPAGE
 412         select COMPACTION
 413         help
 414           Transparent Hugepages allows the kernel to use huge pages and
 415           huge tlb transparently to the applications whenever possible.
 416           This feature can improve computing performance to certain
 417           applications by speeding up page faults during memory
 418           allocation, by reducing the number of tlb misses and by speeding
 419           up the pagetable walking.
 420
 421           If memory constrained on embedded, you may want to say N.
 422
 423 choice
 424         prompt "Transparent Hugepage Support sysfs defaults"
 425         depends on TRANSPARENT_HUGEPAGE
 426         default TRANSPARENT_HUGEPAGE_ALWAYS
 427         help
 428           Selects the sysfs defaults for Transparent Hugepage Support.
 429
 430         config TRANSPARENT_HUGEPAGE_ALWAYS
 431                 bool "always"
 432         help
 433           Enabling Transparent Hugepage always, can increase the
 434           memory footprint of applications without a guaranteed
 435           benefit but it will work automatically for all applications.
 436
 437         config TRANSPARENT_HUGEPAGE_MADVISE
 438                 bool "madvise"
 439         help
 440           Enabling Transparent Hugepage madvise, will only provide a
 441           performance improvement benefit to the applications using
 442           madvise(MADV_HUGEPAGE) but it won't risk to increase the
 443           memory footprint of applications without a guaranteed
 444           benefit.
 445 endchoice
 446
 447 #
 448 # UP and nommu archs use km based percpu allocator
 449 #
 450 config NEED_PER_CPU_KM
 451         depends on !SMP
 452         bool
 453         default y
 454
 455 config CLEANCACHE
 456         bool "Enable cleancache driver to cache clean pages if tmem is present"
 457         default n
 458         help
 459           Cleancache can be thought of as a page-granularity victim cache
 460           for clean pages that the kernel's pageframe replacement algorithm
 461           (PFRA) would like to keep around, but can't since there isn't enough
 462           memory.  So when the PFRA "evicts" a page, it first attempts to use
 463           cleancache code to put the data contained in that page into
 464           "transcendent memory", memory that is not directly accessible or
 465           addressable by the kernel and is of unknown and possibly
 466           time-varying size.  And when a cleancache-enabled
 467           filesystem wishes to access a page in a file on disk, it first
 468           checks cleancache to see if it already contains it; if it does,
 469           the page is copied into the kernel and a disk access is avoided.
 470           When a transcendent memory driver is available (such as zcache or
 471           Xen transcendent memory), a significant I/O reduction
 472           may be achieved.  When none is available, all cleancache calls
 473           are reduced to a single pointer-compare-against-NULL resulting
 474           in a negligible performance hit.
 475
 476           If unsure, say Y to enable cleancache
 477
 478 config FRONTSWAP
 479         bool "Enable frontswap to cache swap pages if tmem is present"
 480         depends on SWAP
 481         default n
 482         help
 483           Frontswap is so named because it can be thought of as the opposite
 484           of a "backing" store for a swap device.  The data is stored into
 485           "transcendent memory", memory that is not directly accessible or
 486           addressable by the kernel and is of unknown and possibly
 487           time-varying size.  When space in transcendent memory is available,
 488           a significant swap I/O reduction may be achieved.  When none is
 489           available, all frontswap calls are reduced to a single pointer-
 490           compare-against-NULL resulting in a negligible performance hit
 491           and swap data is stored as normal on the matching swap device.
 492
 493           If unsure, say Y to enable frontswap.
 494
 495 config CMA
 496         bool "Contiguous Memory Allocator"
 497         depends on HAVE_MEMBLOCK && MMU
 498         select MIGRATION
 499         select MEMORY_ISOLATION
 500         help
 501           This enables the Contiguous Memory Allocator which allows other
 502           subsystems to allocate big physically-contiguous blocks of memory.
 503           CMA reserves a region of memory and allows only movable pages to
 504           be allocated from it. This way, the kernel can use the memory for
 505           pagecache and when a subsystem requests for contiguous area, the
 506           allocated pages are migrated away to serve the contiguous request.
 507
 508           If unsure, say "n".
 509
 510 config CMA_DEBUG
 511         bool "CMA debug messages (DEVELOPMENT)"
 512         depends on DEBUG_KERNEL && CMA
 513         help
 514           Turns on debug messages in CMA.  This produces KERN_DEBUG
 515           messages for every CMA call as well as various messages while
 516           processing calls such as dma_alloc_from_contiguous().
 517           This option does not affect warning and error messages.
 518
 519 config CMA_AREAS
 520         int "Maximum count of the CMA areas"
 521         depends on CMA
 522         default 7
 523         help
 524           CMA allows to create CMA areas for particular purpose, mainly,
 525           used as device private area. This parameter sets the maximum
 526           number of CMA area in the system.
 527
 528           If unsure, leave the default value "7".
 529
 530 config MEM_SOFT_DIRTY
 531         bool "Track memory changes"
 532         depends on CHECKPOINT_RESTORE && HAVE_ARCH_SOFT_DIRTY && PROC_FS
 533         select PROC_PAGE_MONITOR
 534         help
 535           This option enables memory changes tracking by introducing a
 536           soft-dirty bit on pte-s. This bit it set when someone writes
 537           into a page just as regular dirty bit, but unlike the latter
 538           it can be cleared by hands.
 539
 540           See Documentation/vm/soft-dirty.txt for more details.
 541
 542 config ZSWAP
 543         bool "Compressed cache for swap pages (EXPERIMENTAL)"
 544         depends on FRONTSWAP && CRYPTO=y
 545         select CRYPTO_LZO
 546         select ZPOOL
 547         default n
 548         help
 549           A lightweight compressed cache for swap pages.  It takes
 550           pages that are in the process of being swapped out and attempts to
 551           compress them into a dynamically allocated RAM-based memory pool.
 552           This can result in a significant I/O reduction on swap device and,
 553           in the case where decompressing from RAM is faster that swap device
 554           reads, can also improve workload performance.
 555
 556           This is marked experimental because it is a new feature (as of
 557           v3.11) that interacts heavily with memory reclaim.  While these
 558           interactions don't cause any known issues on simple memory setups,
 559           they have not be fully explored on the large set of potential
 560           configurations and workloads that exist.
 561
 562 config ZPOOL
 563         tristate "Common API for compressed memory storage"
 564         default n
 565         help
 566           Compressed memory storage API.  This allows using either zbud or
 567           zsmalloc.
 568
 569 config ZBUD
 570         tristate "Low density storage for compressed pages"
 571         default n
 572         help
 573           A special purpose allocator for storing compressed pages.
 574           It is designed to store up to two compressed pages per physical
 575           page.  While this design limits storage density, it has simple and
 576           deterministic reclaim properties that make it preferable to a higher
 577           density approach when reclaim will be used.
 578
 579 config ZSMALLOC
 580         tristate "Memory allocator for compressed pages"
 581         depends on MMU
 582         default n
 583         help
 584           zsmalloc is a slab-based memory allocator designed to store
 585           compressed RAM pages.  zsmalloc uses virtual memory mapping
 586           in order to reduce fragmentation.  However, this results in a
 587           non-standard allocator interface where a handle, not a pointer, is
 588           returned by an alloc().  This handle must be mapped in order to
 589           access the allocated space.
 590
 591 config PGTABLE_MAPPING
 592         bool "Use page table mapping to access object in zsmalloc"
 593         depends on ZSMALLOC
 594         help
 595           By default, zsmalloc uses a copy-based object mapping method to
 596           access allocations that span two pages. However, if a particular
 597           architecture (ex, ARM) performs VM mapping faster than copying,
 598           then you should select this. This causes zsmalloc to use page table
 599           mapping rather than copying for object mapping.
 600
 601           You can check speed with zsmalloc benchmark:
 602           https://github.com/spartacus06/zsmapbench
 603
 604 config GENERIC_EARLY_IOREMAP
 605         bool
 606
 607 config MAX_STACK_SIZE_MB
 608         int "Maximum user stack size for 32-bit processes (MB)"
 609         default 80
 610         range 8 256 if METAG
 611         range 8 2048
 612         depends on STACK_GROWSUP && (!64BIT || COMPAT)
 613         help
 614           This is the maximum stack size in Megabytes in the VM layout of 32-bit
 615           user processes when the stack grows upwards (currently only on parisc
 616           and metag arch). The stack will be located at the highest memory
 617           address minus the given value, unless the RLIMIT_STACK hard limit is
 618           changed to a smaller value in which case that is used.
 619
 620           A sane initial value is 80 MB.