man2/madvise.2

   1 .\" Copyright (C) 2001 David Gómez <davidge@jazzfree.com>
   2 .\"
   3 .\" SPDX-License-Identifier: Linux-man-pages-copyleft
   4 .\"
   5 .\" Based on comments from mm/filemap.c. Last modified on 10-06-2001
   6 .\" Modified, 25 Feb 2002, Michael Kerrisk, <mtk.manpages@gmail.com>
   7 .\"     Added notes on MADV_DONTNEED
   8 .\" 2010-06-19, mtk, Added documentation of MADV_MERGEABLE and
   9 .\"     MADV_UNMERGEABLE
  10 .\" 2010-06-15, Andi Kleen, Add documentation of MADV_HWPOISON.
  11 .\" 2010-06-19, Andi Kleen, Add documentation of MADV_SOFT_OFFLINE.
  12 .\" 2011-09-18, Doug Goldstein <cardoe@cardoe.com>
  13 .\"     Document MADV_HUGEPAGE and MADV_NOHUGEPAGE
  14 .\"
  15 .TH MADVISE 2 (date) "Linux man-pages (unreleased)"
  16 .SH NAME
  17 madvise \- give advice about use of memory
  18 .SH LIBRARY
  19 Standard C library
  20 .RI ( libc ", " \-lc )
  21 .SH SYNOPSIS
  22 .nf
  23 .B #include <sys/mman.h>
  24 .PP
  25 .BI "int madvise(void *" addr ", size_t " length ", int " advice );
  26 .fi
  27 .PP
  28 .RS -4
  29 Feature Test Macro Requirements for glibc (see
  30 .BR feature_test_macros (7)):
  31 .RE
  32 .PP
  33 .BR madvise ():
  34 .nf
  35     Since glibc 2.19:
  36         _DEFAULT_SOURCE
  37     Up to and including glibc 2.19:
  38         _BSD_SOURCE
  39 .fi
  40 .SH DESCRIPTION
  41 The
  42 .BR madvise ()
  43 system call is used to give advice or directions to the kernel
  44 about the address range beginning at address
  45 .I addr
  46 and with size
  47 .IR length .
  48 .BR madvise ()
  49 only operates on whole pages, therefore
  50 .I addr
  51 must be page-aligned.
  52 The value of
  53 .I length
  54 is rounded up to a multiple of page size.
  55 In most cases,
  56 the goal of such advice is to improve system or application performance.
  57 .PP
  58 Initially, the system call supported a set of "conventional"
  59 .I advice
  60 values, which are also available on several other implementations.
  61 (Note, though, that
  62 .BR madvise ()
  63 is not specified in POSIX.)
  64 Subsequently, a number of Linux-specific
  65 .I advice
  66 values have been added.
  67 .\"
  68 .\" ======================================================================
  69 .\"
  70 .SS Conventional advice values
  71 The
  72 .I advice
  73 values listed below
  74 allow an application to tell the kernel how it expects to use
  75 some mapped or shared memory areas, so that the kernel can choose
  76 appropriate read-ahead and caching techniques.
  77 These
  78 .I advice
  79 values do not influence the semantics of the application
  80 (except in the case of
  81 .BR MADV_DONTNEED ),
  82 but may influence its performance.
  83 All of the
  84 .I advice
  85 values listed here have analogs in the POSIX-specified
  86 .BR posix_madvise (3)
  87 function, and the values have the same meanings, with the exception of
  88 .BR MADV_DONTNEED .
  89 .PP
  90 The advice is indicated in the
  91 .I advice
  92 argument, which is one of the following:
  93 .TP
  94 .B MADV_NORMAL
  95 No special treatment.
  96 This is the default.
  97 .TP
  98 .B MADV_RANDOM
  99 Expect page references in random order.
 100 (Hence, read ahead may be less useful than normally.)
 101 .TP
 102 .B MADV_SEQUENTIAL
 103 Expect page references in sequential order.
 104 (Hence, pages in the given range can be aggressively read ahead,
 105 and may be freed soon after they are accessed.)
 106 .TP
 107 .B MADV_WILLNEED
 108 Expect access in the near future.
 109 (Hence, it might be a good idea to read some pages ahead.)
 110 .TP
 111 .B MADV_DONTNEED
 112 Do not expect access in the near future.
 113 (For the time being, the application is finished with the given range,
 114 so the kernel can free resources associated with it.)
 115 .IP
 116 After a successful
 117 .B MADV_DONTNEED
 118 operation,
 119 the semantics of memory access in the specified region are changed:
 120 subsequent accesses of pages in the range will succeed, but will result
 121 in either repopulating the memory contents from the
 122 up-to-date contents of the underlying mapped file
 123 (for shared file mappings, shared anonymous mappings,
 124 and shmem-based techniques such as System V shared memory segments)
 125 or zero-fill-on-demand pages for anonymous private mappings.
 126 .IP
 127 Note that, when applied to shared mappings,
 128 .B MADV_DONTNEED
 129 might not lead to immediate freeing of the pages in the range.
 130 The kernel is free to delay freeing the pages until an appropriate moment.
 131 The resident set size (RSS) of the calling process will be immediately
 132 reduced however.
 133 .IP
 134 .B MADV_DONTNEED
 135 cannot be applied to locked pages, or
 136 .B VM_PFNMAP
 137 pages.
 138 (Pages marked with the kernel-internal
 139 .B VM_PFNMAP
 140 .\" http://lwn.net/Articles/162860/
 141 flag are special memory areas that are not managed
 142 by the virtual memory subsystem.
 143 Such pages are typically created by device drivers that
 144 map the pages into user space.)
 145 .IP
 146 Support for Huge TLB pages was added in Linux v5.18.
 147 Addresses within a mapping backed by Huge TLB pages must be aligned
 148 to the underlying Huge TLB page size,
 149 and the range length is rounded up
 150 to a multiple of the underlying Huge TLB page size.
 151 .\"
 152 .\" ======================================================================
 153 .\"
 154 .SS Linux-specific advice values
 155 The following Linux-specific
 156 .I advice
 157 values have no counterparts in the POSIX-specified
 158 .BR posix_madvise (3),
 159 and may or may not have counterparts in the
 160 .BR madvise ()
 161 interface available on other implementations.
 162 Note that some of these operations change the semantics of memory accesses.
 163 .TP
 164 .BR MADV_REMOVE " (since Linux 2.6.16)"
 165 .\" commit f6b3ec238d12c8cc6cc71490c6e3127988460349
 166 Free up a given range of pages
 167 and its associated backing store.
 168 This is equivalent to punching a hole in the corresponding
 169 range of the backing store (see
 170 .BR fallocate (2)).
 171 Subsequent accesses in the specified address range will see
 172 data with a value of zero.
 173 .\" Databases want to use this feature to drop a section of their
 174 .\" bufferpool (shared memory segments) - without writing back to
 175 .\" disk/swap space.  This feature is also useful for supporting
 176 .\" hot-plug memory on UML.
 177 .IP
 178 The specified address range must be mapped shared and writable.
 179 This flag cannot be applied to locked pages, or
 180 .B VM_PFNMAP
 181 pages.
 182 .IP
 183 In the initial implementation, only
 184 .BR tmpfs (5)
 185 supported
 186 .BR MADV_REMOVE ;
 187 but since Linux 3.5,
 188 .\" commit 3f31d07571eeea18a7d34db9af21d2285b807a17
 189 any filesystem which supports the
 190 .BR fallocate (2)
 191 .B FALLOC_FL_PUNCH_HOLE
 192 mode also supports
 193 .BR MADV_REMOVE .
 194 Filesystems which do not support
 195 .B MADV_REMOVE
 196 fail with the error
 197 .BR EOPNOTSUPP .
 198 .IP
 199 Support for the Huge TLB filesystem was added in Linux v4.3.
 200 .TP
 201 .BR MADV_DONTFORK " (since Linux 2.6.16)"
 202 .\" commit f822566165dd46ff5de9bf895cfa6c51f53bb0c4
 203 .\" See http://lwn.net/Articles/171941/
 204 Do not make the pages in this range available to the child after a
 205 .BR fork (2).
 206 This is useful to prevent copy-on-write semantics from changing
 207 the physical location of a page if the parent writes to it after a
 208 .BR fork (2).
 209 (Such page relocations cause problems for hardware that
 210 DMAs into the page.)
 211 .\" [PATCH] madvise MADV_DONTFORK/MADV_DOFORK
 212 .\" Currently, copy-on-write may change the physical address of
 213 .\" a page even if the user requested that the page is pinned in
 214 .\" memory (either by mlock or by get_user_pages).  This happens
 215 .\" if the process forks meanwhile, and the parent writes to that
 216 .\" page.  As a result, the page is orphaned: in case of
 217 .\" get_user_pages, the application will never see any data hardware
 218 .\" DMA's into this page after the COW.  In case of mlock'd memory,
 219 .\" the parent is not getting the realtime/security benefits of mlock.
 220 .\"
 221 .\" In particular, this affects the Infiniband modules which do DMA from
 222 .\" and into user pages all the time.
 223 .\"
 224 .\" This patch adds madvise options to control whether memory range is
 225 .\" inherited across fork. Useful e.g. for when hardware is doing DMA
 226 .\" from/into these pages.  Could also be useful to an application
 227 .\" wanting to speed up its forks by cutting large areas out of
 228 .\" consideration.
 229 .\"
 230 .\" SEE ALSO: http://lwn.net/Articles/171941/
 231 .\" "Tweaks to madvise() and posix_fadvise()", 14 Feb 2006
 232 .TP
 233 .BR MADV_DOFORK " (since Linux 2.6.16)"
 234 Undo the effect of
 235 .BR MADV_DONTFORK ,
 236 restoring the default behavior, whereby a mapping is inherited across
 237 .BR fork (2).
 238 .TP
 239 .BR MADV_HWPOISON " (since Linux 2.6.32)"
 240 .\" commit 9893e49d64a4874ea67849ee2cfbf3f3d6817573
 241 Poison the pages in the range specified by
 242 .I addr
 243 and
 244 .I length
 245 and handle subsequent references to those pages
 246 like a hardware memory corruption.
 247 This operation is available only for privileged
 248 .RB ( CAP_SYS_ADMIN )
 249 processes.
 250 This operation may result in the calling process receiving a
 251 .B SIGBUS
 252 and the page being unmapped.
 253 .IP
 254 This feature is intended for testing of memory error-handling code;
 255 it is available only if the kernel was configured with
 256 .BR CONFIG_MEMORY_FAILURE .
 257 .TP
 258 .BR MADV_MERGEABLE " (since Linux 2.6.32)"
 259 .\" commit f8af4da3b4c14e7267c4ffb952079af3912c51c5
 260 Enable Kernel Samepage Merging (KSM) for the pages in the range specified by
 261 .I addr
 262 and
 263 .IR length .
 264 The kernel regularly scans those areas of user memory that have
 265 been marked as mergeable,
 266 looking for pages with identical content.
 267 These are replaced by a single write-protected page (which is automatically
 268 copied if a process later wants to update the content of the page).
 269 KSM merges only private anonymous pages (see
 270 .BR mmap (2)).
 271 .IP
 272 The KSM feature is intended for applications that generate many
 273 instances of the same data (e.g., virtualization systems such as KVM).
 274 It can consume a lot of processing power; use with care.
 275 See the Linux kernel source file
 276 .I Documentation/admin\-guide/mm/ksm.rst
 277 for more details.
 278 .IP
 279 The
 280 .B MADV_MERGEABLE
 281 and
 282 .B MADV_UNMERGEABLE
 283 operations are available only if the kernel was configured with
 284 .BR CONFIG_KSM .
 285 .TP
 286 .BR MADV_UNMERGEABLE " (since Linux 2.6.32)"
 287 Undo the effect of an earlier
 288 .B MADV_MERGEABLE
 289 operation on the specified address range;
 290 KSM unmerges whatever pages it had merged in the address range specified by
 291 .I addr
 292 and
 293 .IR length .
 294 .TP
 295 .BR MADV_SOFT_OFFLINE " (since Linux 2.6.33)"
 296 .\" commit afcf938ee0aac4ef95b1a23bac704c6fbeb26de6
 297 Soft offline the pages in the range specified by
 298 .I addr
 299 and
 300 .IR length .
 301 The memory of each page in the specified range is preserved
 302 (i.e., when next accessed, the same content will be visible,
 303 but in a new physical page frame),
 304 and the original page is offlined
 305 (i.e., no longer used, and taken out of normal memory management).
 306 The effect of the
 307 .B MADV_SOFT_OFFLINE
 308 operation is invisible to (i.e., does not change the semantics of)
 309 the calling process.
 310 .IP
 311 This feature is intended for testing of memory error-handling code;
 312 it is available only if the kernel was configured with
 313 .BR CONFIG_MEMORY_FAILURE .
 314 .TP
 315 .BR MADV_HUGEPAGE " (since Linux 2.6.38)"
 316 .\" commit 0af4e98b6b095c74588af04872f83d333c958c32
 317 .\" http://lwn.net/Articles/358904/
 318 .\" https://lwn.net/Articles/423584/
 319 Enable Transparent Huge Pages (THP) for pages in the range specified by
 320 .I addr
 321 and
 322 .IR length .
 323 The kernel will regularly scan the areas marked as huge page candidates
 324 to replace them with huge pages.
 325 The kernel will also allocate huge pages directly when the region is
 326 naturally aligned to the huge page size (see
 327 .BR posix_memalign (2)).
 328 .IP
 329 This feature is primarily aimed at applications that use large mappings of
 330 data and access large regions of that memory at a time (e.g., virtualization
 331 systems such as QEMU).
 332 It can very easily waste memory (e.g., a 2\ MB mapping that only ever accesses
 333 1 byte will result in 2\ MB of wired memory instead of one 4\ KB page).
 334 See the Linux kernel source file
 335 .I Documentation/admin\-guide/mm/transhuge.rst
 336 for more details.
 337 .IP
 338 Most common kernels configurations provide
 339 .BR MADV_HUGEPAGE -style
 340 behavior by default, and thus
 341 .B MADV_HUGEPAGE
 342 is normally not necessary.
 343 It is mostly intended for embedded systems, where
 344 .BR MADV_HUGEPAGE -style
 345 behavior may not be enabled by default in the kernel.
 346 On such systems,
 347 this flag can be used in order to selectively enable THP.
 348 Whenever
 349 .B MADV_HUGEPAGE
 350 is used, it should always be in regions of memory with
 351 an access pattern that the developer knows in advance won't risk
 352 to increase the memory footprint of the application when transparent
 353 hugepages are enabled.
 354 .IP
 355 .\" commit 99cb0dbd47a15d395bf3faa78dc122bc5efe3fc0
 356 Since Linux 5.4,
 357 automatic scan of eligible areas and replacement by huge pages works with
 358 private anonymous pages (see
 359 .BR mmap (2)),
 360 shmem pages,
 361 and file-backed pages.
 362 For all memory types,
 363 memory may only be replaced by huge pages on hugepage-aligned boundaries.
 364 For file-mapped memory
 365 \(emincluding tmpfs (see
 366 .BR tmpfs (2))\(em
 367 the mapping must also be naturally hugepage-aligned within the file.
 368 Additionally,
 369 for file-backed,
 370 non-tmpfs memory,
 371 the file must not be open for write and the mapping must be executable.
 372 .IP
 373 The VMA must not be marked
 374 .BR VM_NOHUGEPAGE ,
 375 .BR VM_HUGETLB ,
 376 .BR VM_IO ,
 377 .BR VM_DONTEXPAND ,
 378 .BR VM_MIXEDMAP ,
 379 or
 380 .BR VM_PFNMAP ,
 381 nor can it be stack memory or backed by a DAX-enabled device
 382 (unless the DAX device is hot-plugged as System RAM).
 383 The process must also not have
 384 .B PR_SET_THP_DISABLE
 385 set (see
 386 .BR prctl (2)).
 387 .IP
 388 The
 389 .B MADV_HUGEPAGE
 390 and
 391 .B MADV_NOHUGEPAGE
 392 operations are available only if the kernel was configured with
 393 .B CONFIG_TRANSPARENT_HUGEPAGE
 394 and file/shmem memory is only supported if the kernel was configured with
 395 .BR CONFIG_READ_ONLY_THP_FOR_FS .
 396 .TP
 397 .BR MADV_NOHUGEPAGE " (since Linux 2.6.38)"
 398 Ensures that memory in the address range specified by
 399 .I addr
 400 and
 401 .I length
 402 will not be backed by transparent hugepages.
 403 .TP
 404 .BR MADV_DONTDUMP " (since Linux 3.4)"
 405 .\" commit 909af768e88867016f427264ae39d27a57b6a8ed
 406 .\" commit accb61fe7bb0f5c2a4102239e4981650f9048519
 407 Exclude from a core dump those pages in the range specified by
 408 .I addr
 409 and
 410 .IR length .
 411 This is useful in applications that have large areas of memory
 412 that are known not to be useful in a core dump.
 413 The effect of
 414 .B MADV_DONTDUMP
 415 takes precedence over the bit mask that is set via the
 416 .I /proc/[pid]/coredump_filter
 417 file (see
 418 .BR core (5)).
 419 .TP
 420 .BR MADV_DODUMP " (since Linux 3.4)"
 421 Undo the effect of an earlier
 422 .BR MADV_DONTDUMP .
 423 .TP
 424 .BR MADV_FREE " (since Linux 4.5)"
 425 The application no longer requires the pages in the range specified by
 426 .I addr
 427 and
 428 .IR len .
 429 The kernel can thus free these pages,
 430 but the freeing could be delayed until memory pressure occurs.
 431 For each of the pages that has been marked to be freed
 432 but has not yet been freed,
 433 the free operation will be canceled if the caller writes into the page.
 434 After a successful
 435 .B MADV_FREE
 436 operation, any stale data (i.e., dirty, unwritten pages) will be lost
 437 when the kernel frees the pages.
 438 However, subsequent writes to pages in the range will succeed
 439 and then kernel cannot free those dirtied pages,
 440 so that the caller can always see just written data.
 441 If there is no subsequent write,
 442 the kernel can free the pages at any time.
 443 Once pages in the range have been freed, the caller will
 444 see zero-fill-on-demand pages upon subsequent page references.
 445 .IP
 446 The
 447 .B MADV_FREE
 448 operation
 449 can be applied only to private anonymous pages (see
 450 .BR mmap (2)).
 451 In Linux before version 4.12,
 452 .\" commit 93e06c7a645343d222c9a838834a51042eebbbf7
 453 when freeing pages on a swapless system,
 454 the pages in the given range are freed instantly,
 455 regardless of memory pressure.
 456 .TP
 457 .BR MADV_WIPEONFORK " (since Linux 4.14)"
 458 .\" commit d2cd9ede6e193dd7d88b6d27399e96229a551b19
 459 Present the child process with zero-filled memory in this range after a
 460 .BR fork (2).
 461 This is useful in forking servers in order to ensure
 462 that sensitive per-process data
 463 (for example, PRNG seeds, cryptographic secrets, and so on)
 464 is not handed to child processes.
 465 .IP
 466 The
 467 .B MADV_WIPEONFORK
 468 operation can be applied only to private anonymous pages (see
 469 .BR mmap (2)).
 470 .IP
 471 Within the child created by
 472 .BR fork (2),
 473 the
 474 .B MADV_WIPEONFORK
 475 setting remains in place on the specified address range.
 476 This setting is cleared during
 477 .BR execve (2).
 478 .TP
 479 .BR MADV_KEEPONFORK " (since Linux 4.14)"
 480 .\" commit d2cd9ede6e193dd7d88b6d27399e96229a551b19
 481 Undo the effect of an earlier
 482 .BR MADV_WIPEONFORK .
 483 .TP
 484 .BR MADV_COLD " (since Linux 5.4)"
 485 .\" commit 9c276cc65a58faf98be8e56962745ec99ab87636
 486 Deactivate a given range of pages.
 487 This will make the pages a more probable
 488 reclaim target should there be a memory pressure.
 489 This is a nondestructive operation.
 490 The advice might be ignored for some pages in the range when it is not
 491 applicable.
 492 .TP
 493 .BR MADV_PAGEOUT " (since Linux 5.4)"
 494 .\" commit 1a4e58cce84ee88129d5d49c064bd2852b481357
 495 Reclaim a given range of pages.
 496 This is done to free up memory occupied by these pages.
 497 If a page is anonymous, it will be swapped out.
 498 If a page is file-backed and dirty, it will be written back to the backing
 499 storage.
 500 The advice might be ignored for some pages in the range when it is not
 501 applicable.
 502 .TP
 503 .BR MADV_POPULATE_READ " (since Linux 5.14)"
 504 "Populate (prefault) page tables readable,
 505 faulting in all pages in the range just as if manually reading from each page;
 506 however,
 507 avoid the actual memory access that would have been performed after handling
 508 the fault.
 509 .IP
 510 In contrast to
 511 .BR MAP_POPULATE ,
 512 .B MADV_POPULATE_READ
 513 does not hide errors,
 514 can be applied to (parts of) existing mappings and will always populate
 515 (prefault) page tables readable.
 516 One example use case is prefaulting a file mapping,
 517 reading all file content from disk;
 518 however,
 519 pages won't be dirtied and consequently won't have to be written back to disk
 520 when evicting the pages from memory.
 521 .IP
 522 Depending on the underlying mapping,
 523 map the shared zeropage,
 524 preallocate memory or read the underlying file;
 525 files with holes might or might not preallocate blocks.
 526 If populating fails,
 527 a
 528 .B SIGBUS
 529 signal is not generated; instead, an error is returned.
 530 .IP
 531 If
 532 .B MADV_POPULATE_READ
 533 succeeds,
 534 all page tables have been populated (prefaulted) readable once.
 535 If
 536 .B MADV_POPULATE_READ
 537 fails,
 538 some page tables might have been populated.
 539 .IP
 540 .B MADV_POPULATE_READ
 541 cannot be applied to mappings without read permissions
 542 and special mappings,
 543 for example,
 544 mappings marked with kernel-internal flags such as
 545 .B VM_PFNMAP
 546 or
 547 .BR VM_IO ,
 548 or secret memory regions created using
 549 .BR memfd_secret(2) .
 550 .IP
 551 Note that with
 552 .BR MADV_POPULATE_READ ,
 553 the process can be killed at any moment when the system runs out of memory.
 554 .TP
 555 .BR MADV_POPULATE_WRITE " (since Linux 5.14)"
 556 Populate (prefault) page tables writable,
 557 faulting in all pages in the range just as if manually writing to each
 558 each page;
 559 however,
 560 avoid the actual memory access that would have been performed after handling
 561 the fault.
 562 .IP
 563 In contrast to
 564 .BR MAP_POPULATE ,
 565 MADV_POPULATE_WRITE does not hide errors,
 566 can be applied to (parts of) existing mappings and will always populate
 567 (prefault) page tables writable.
 568 One example use case is preallocating memory,
 569 breaking any CoW (Copy on Write).
 570 .IP
 571 Depending on the underlying mapping,
 572 preallocate memory or read the underlying file;
 573 files with holes will preallocate blocks.
 574 If populating fails,
 575 a
 576 .B SIGBUS
 577 signal is not generated; instead, an error is returned.
 578 .IP
 579 If
 580 .B MADV_POPULATE_WRITE
 581 succeeds,
 582 all page tables have been populated (prefaulted) writable once.
 583 If
 584 .B MADV_POPULATE_WRITE
 585 fails,
 586 some page tables might have been populated.
 587 .IP
 588 .B MADV_POPULATE_WRITE
 589 cannot be applied to mappings without write permissions
 590 and special mappings,
 591 for example,
 592 mappings marked with kernel-internal flags such as
 593 .B VM_PFNMAP
 594 or
 595 .BR VM_IO ,
 596 or secret memory regions created using
 597 .BR memfd_secret(2) .
 598 .IP
 599 Note that with
 600 .BR MADV_POPULATE_WRITE ,
 601 the process can be killed at any moment when the system runs out of memory.
 602 .SH RETURN VALUE
 603 On success,
 604 .BR madvise ()
 605 returns zero.
 606 On error, it returns \-1 and
 607 .I errno
 608 is set to indicate the error.
 609 .SH ERRORS
 610 .TP
 611 .B EACCES
 612 .I advice
 613 is
 614 .BR MADV_REMOVE ,
 615 but the specified address range is not a shared writable mapping.
 616 .TP
 617 .B EAGAIN
 618 A kernel resource was temporarily unavailable.
 619 .TP
 620 .B EBADF
 621 The map exists, but the area maps something that isn't a file.
 622 .TP
 623 .B EFAULT
 624 .I advice
 625 is
 626 .B MADV_POPULATE_READ
 627 or
 628 .BR MADV_POPULATE_WRITE ,
 629 and populating (prefaulting) page tables failed because a
 630 .B SIGBUS
 631 would have been generated on actual memory access and the reason is not a
 632 HW poisoned page
 633 (HW poisoned pages can,
 634 for example,
 635 be created using the
 636 .B MADV_HWPOISON
 637 flag described elsewhere in this page).
 638 .TP
 639 .B EINVAL
 640 .I addr
 641 is not page-aligned or
 642 .I length
 643 is negative.
 644 .\" .I length
 645 .\" is zero,
 646 .TP
 647 .B EINVAL
 648 .I advice
 649 is not a valid.
 650 .TP
 651 .B EINVAL
 652 .I advice
 653 is
 654 .B MADV_COLD
 655 or
 656 .B MADV_PAGEOUT
 657 and the specified address range includes locked, Huge TLB pages, or
 658 .B VM_PFNMAP
 659 pages.
 660 .TP
 661 .B EINVAL
 662 .I advice
 663 is
 664 .B MADV_DONTNEED
 665 or
 666 .B MADV_REMOVE
 667 and the specified address range includes locked, Huge TLB pages, or
 668 .B VM_PFNMAP
 669 pages.
 670 .TP
 671 .B EINVAL
 672 .I advice
 673 is
 674 .B MADV_MERGEABLE
 675 or
 676 .BR MADV_UNMERGEABLE ,
 677 but the kernel was not configured with
 678 .BR CONFIG_KSM .
 679 .TP
 680 .B EINVAL
 681 .I advice
 682 is
 683 .B MADV_FREE
 684 or
 685 .B MADV_WIPEONFORK
 686 but the specified address range includes file, Huge TLB,
 687 .BR MAP_SHARED ,
 688 or
 689 .B VM_PFNMAP
 690 ranges.
 691 .TP
 692 .B EINVAL
 693 .I advice
 694 is
 695 .B MADV_POPULATE_READ
 696 or
 697 .BR MADV_POPULATE_WRITE ,
 698 but the specified address range includes ranges with insufficient permissions
 699 or special mappings,
 700 for example,
 701 mappings marked with kernel-internal flags such a
 702 .B VM_IO
 703 or
 704 .BR VM_PFNMAP ,
 705 or secret memory regions created using
 706 .BR memfd_secret(2) .
 707 .TP
 708 .B EIO
 709 (for
 710 .BR MADV_WILLNEED )
 711 Paging in this area would exceed the process's
 712 maximum resident set size.
 713 .TP
 714 .B ENOMEM
 715 (for
 716 .BR MADV_WILLNEED )
 717 Not enough memory: paging in failed.
 718 .TP
 719 .B ENOMEM
 720 Addresses in the specified range are not currently
 721 mapped, or are outside the address space of the process.
 722 .TP
 723 .B ENOMEM
 724 .I advice
 725 is
 726 .B MADV_POPULATE_READ
 727 or
 728 .BR MADV_POPULATE_WRITE ,
 729 and populating (prefaulting) page tables failed because there was not enough
 730 memory.
 731 .TP
 732 .B EPERM
 733 .I advice
 734 is
 735 .BR MADV_HWPOISON ,
 736 but the caller does not have the
 737 .B CAP_SYS_ADMIN
 738 capability.
 739 .TP
 740 .B EHWPOISON
 741 .I advice
 742 is
 743 .B MADV_POPULATE_READ
 744 or
 745 .BR MADV_POPULATE_WRITE ,
 746 and populating (prefaulting) page tables failed because a HW poisoned page
 747 (HW poisoned pages can,
 748 for example,
 749 be created using the
 750 .B MADV_HWPOISON
 751 flag described elsewhere in this page)
 752 was encountered.
 753 .SH VERSIONS
 754 Since Linux 3.18,
 755 .\" commit d3ac21cacc24790eb45d735769f35753f5b56ceb
 756 support for this system call is optional,
 757 depending on the setting of the
 758 .B CONFIG_ADVISE_SYSCALLS
 759 configuration option.
 760 .SH STANDARDS
 761 .BR madvise ()
 762 is not specified by any standards.
 763 Versions of this system call, implementing a wide variety of
 764 .I advice
 765 values, exist on many other implementations.
 766 Other implementations typically implement at least the flags listed
 767 above under
 768 .IR "Conventional advice flags" ,
 769 albeit with some variation in semantics.
 770 .PP
 771 POSIX.1-2001 describes
 772 .BR posix_madvise (3)
 773 with constants
 774 .BR POSIX_MADV_NORMAL ,
 775 .BR POSIX_MADV_RANDOM ,
 776 .BR POSIX_MADV_SEQUENTIAL ,
 777 .BR POSIX_MADV_WILLNEED ,
 778 and
 779 .BR POSIX_MADV_DONTNEED ,
 780 and so on, with behavior close to the similarly named flags listed above.
 781 .SH NOTES
 782 .SS Linux notes
 783 The Linux implementation requires that the address
 784 .I addr
 785 be page-aligned, and allows
 786 .I length
 787 to be zero.
 788 If there are some parts of the specified address range
 789 that are not mapped, the Linux version of
 790 .BR madvise ()
 791 ignores them and applies the call to the rest (but returns
 792 .B ENOMEM
 793 from the system call, as it should).
 794 .\" .SH HISTORY
 795 .\" The
 796 .\" .BR madvise ()
 797 .\" function first appeared in 4.4BSD.
 798 .SH SEE ALSO
 799 .BR getrlimit (2),
 800 .BR memfd_secret (2),
 801 .BR mincore (2),
 802 .BR mmap (2),
 803 .BR mprotect (2),
 804 .BR msync (2),
 805 .BR munmap (2),
 806 .BR prctl (2),
 807 .BR process_madvise (2),
 808 .BR posix_madvise (3),
 809 .BR core (5)