1========================== 2Memory Resource Controller 3========================== 4 5.. caution:: 6 This document is hopelessly outdated and it asks for a complete 7 rewrite. It still contains a useful information so we are keeping it 8 here but make sure to check the current code if you need a deeper 9 understanding. 10 11.. note:: 12 The Memory Resource Controller has generically been referred to as the 13 memory controller in this document. Do not confuse memory controller 14 used here with the memory controller that is used in hardware. 15 16.. hint:: 17 When we mention a cgroup (cgroupfs's directory) with memory controller, 18 we call it "memory cgroup". When you see git-log and source code, you'll 19 see patch's title and function names tend to use "memcg". 20 In this document, we avoid using it. 21 22Benefits and Purpose of the memory controller 23============================================= 24 25The memory controller isolates the memory behaviour of a group of tasks 26from the rest of the system. The article on LWN [12]_ mentions some probable 27uses of the memory controller. The memory controller can be used to 28 29a. Isolate an application or a group of applications 30 Memory-hungry applications can be isolated and limited to a smaller 31 amount of memory. 32b. Create a cgroup with a limited amount of memory; this can be used 33 as a good alternative to booting with mem=XXXX. 34c. Virtualization solutions can control the amount of memory they want 35 to assign to a virtual machine instance. 36d. A CD/DVD burner could control the amount of memory used by the 37 rest of the system to ensure that burning does not fail due to lack 38 of available memory. 39e. There are several other use cases; find one or use the controller just 40 for fun (to learn and hack on the VM subsystem). 41 42Current Status: linux-2.6.34-mmotm(development version of 2010/April) 43 44Features: 45 46 - accounting anonymous pages, file caches, swap caches usage and limiting them. 47 - pages are linked to per-memcg LRU exclusively, and there is no global LRU. 48 - optionally, memory+swap usage can be accounted and limited. 49 - hierarchical accounting 50 - soft limit 51 - moving (recharging) account at moving a task is selectable. 52 - usage threshold notifier 53 - memory pressure notifier 54 - oom-killer disable knob and oom-notifier 55 - Root cgroup has no limit controls. 56 57 Kernel memory support is a work in progress, and the current version provides 58 basically functionality. (See :ref:`section 2.7 59 <cgroup-v1-memory-kernel-extension>`) 60 61Brief summary of control files. 62 63==================================== ========================================== 64 tasks attach a task(thread) and show list of 65 threads 66 cgroup.procs show list of processes 67 cgroup.event_control an interface for event_fd() 68 This knob is not available on CONFIG_PREEMPT_RT systems. 69 memory.usage_in_bytes show current usage for memory 70 (See 5.5 for details) 71 memory.memsw.usage_in_bytes show current usage for memory+Swap 72 (See 5.5 for details) 73 memory.limit_in_bytes set/show limit of memory usage 74 memory.memsw.limit_in_bytes set/show limit of memory+Swap usage 75 memory.failcnt show the number of memory usage hits limits 76 memory.memsw.failcnt show the number of memory+Swap hits limits 77 memory.max_usage_in_bytes show max memory usage recorded 78 memory.memsw.max_usage_in_bytes show max memory+Swap usage recorded 79 memory.soft_limit_in_bytes set/show soft limit of memory usage 80 This knob is not available on CONFIG_PREEMPT_RT systems. 81 This knob is deprecated and shouldn't be 82 used. 83 memory.stat show various statistics 84 memory.use_hierarchy set/show hierarchical account enabled 85 This knob is deprecated and shouldn't be 86 used. 87 memory.force_empty trigger forced page reclaim 88 memory.pressure_level set memory pressure notifications 89 This knob is deprecated and shouldn't be 90 used. 91 memory.swappiness set/show swappiness parameter of vmscan 92 (See sysctl's vm.swappiness) 93 memory.move_charge_at_immigrate This knob is deprecated. 94 memory.oom_control set/show oom controls. 95 This knob is deprecated and shouldn't be 96 used. 97 memory.numa_stat show the number of memory usage per numa 98 node 99 memory.kmem.limit_in_bytes Deprecated knob to set and read the kernel 100 memory hard limit. Kernel hard limit is not 101 supported since 5.16. Writing any value to 102 do file will not have any effect same as if 103 nokmem kernel parameter was specified. 104 Kernel memory is still charged and reported 105 by memory.kmem.usage_in_bytes. 106 memory.kmem.usage_in_bytes show current kernel memory allocation 107 memory.kmem.failcnt show the number of kernel memory usage 108 hits limits 109 memory.kmem.max_usage_in_bytes show max kernel memory usage recorded 110 111 memory.kmem.tcp.limit_in_bytes set/show hard limit for tcp buf memory 112 This knob is deprecated and shouldn't be 113 used. 114 memory.kmem.tcp.usage_in_bytes show current tcp buf memory allocation 115 This knob is deprecated and shouldn't be 116 used. 117 memory.kmem.tcp.failcnt show the number of tcp buf memory usage 118 hits limits 119 This knob is deprecated and shouldn't be 120 used. 121 memory.kmem.tcp.max_usage_in_bytes show max tcp buf memory usage recorded 122 This knob is deprecated and shouldn't be 123 used. 124==================================== ========================================== 125 1261. History 127========== 128 129The memory controller has a long history. A request for comments for the memory 130controller was posted by Balbir Singh [1]_. At the time the RFC was posted 131there were several implementations for memory control. The goal of the 132RFC was to build consensus and agreement for the minimal features required 133for memory control. The first RSS controller was posted by Balbir Singh [2]_ 134in Feb 2007. Pavel Emelianov [3]_ [4]_ [5]_ has since posted three versions 135of the RSS controller. At OLS, at the resource management BoF, everyone 136suggested that we handle both page cache and RSS together. Another request was 137raised to allow user space handling of OOM. The current memory controller is 138at version 6; it combines both mapped (RSS) and unmapped Page 139Cache Control [11]_. 140 1412. Memory Control 142================= 143 144Memory is a unique resource in the sense that it is present in a limited 145amount. If a task requires a lot of CPU processing, the task can spread 146its processing over a period of hours, days, months or years, but with 147memory, the same physical memory needs to be reused to accomplish the task. 148 149The memory controller implementation has been divided into phases. These 150are: 151 1521. Memory controller 1532. mlock(2) controller 1543. Kernel user memory accounting and slab control 1554. user mappings length controller 156 157The memory controller is the first controller developed. 158 1592.1. Design 160----------- 161 162The core of the design is a counter called the page_counter. The 163page_counter tracks the current memory usage and limit of the group of 164processes associated with the controller. Each cgroup has a memory controller 165specific data structure (mem_cgroup) associated with it. 166 1672.2. Accounting 168--------------- 169 170.. code-block:: 171 :caption: Figure 1: Hierarchy of Accounting 172 173 +--------------------+ 174 | mem_cgroup | 175 | (page_counter) | 176 +--------------------+ 177 / ^ \ 178 / | \ 179 +---------------+ | +---------------+ 180 | mm_struct | |.... | mm_struct | 181 | | | | | 182 +---------------+ | +---------------+ 183 | 184 + --------------+ 185 | 186 +---------------+ +------+--------+ 187 | page +----------> page_cgroup| 188 | | | | 189 +---------------+ +---------------+ 190 191 192 193Figure 1 shows the important aspects of the controller 194 1951. Accounting happens per cgroup 1962. Each mm_struct knows about which cgroup it belongs to 1973. Each page has a pointer to the page_cgroup, which in turn knows the 198 cgroup it belongs to 199 200The accounting is done as follows: mem_cgroup_charge_common() is invoked to 201set up the necessary data structures and check if the cgroup that is being 202charged is over its limit. If it is, then reclaim is invoked on the cgroup. 203More details can be found in the reclaim section of this document. 204If everything goes well, a page meta-data-structure called page_cgroup is 205updated. page_cgroup has its own LRU on cgroup. 206(*) page_cgroup structure is allocated at boot/memory-hotplug time. 207 2082.2.1 Accounting details 209------------------------ 210 211All mapped anon pages (RSS) and cache pages (Page Cache) are accounted. 212Some pages which are never reclaimable and will not be on the LRU 213are not accounted. We just account pages under usual VM management. 214 215RSS pages are accounted at page_fault unless they've already been accounted 216for earlier. A file page will be accounted for as Page Cache when it's 217inserted into inode (xarray). While it's mapped into the page tables of 218processes, duplicate accounting is carefully avoided. 219 220An RSS page is unaccounted when it's fully unmapped. A PageCache page is 221unaccounted when it's removed from xarray. Even if RSS pages are fully 222unmapped (by kswapd), they may exist as SwapCache in the system until they 223are really freed. Such SwapCaches are also accounted. 224A swapped-in page is accounted after adding into swapcache. 225 226Note: The kernel does swapin-readahead and reads multiple swaps at once. 227Since page's memcg recorded into swap whatever memsw enabled, the page will 228be accounted after swapin. 229 230At page migration, accounting information is kept. 231 232Note: we just account pages-on-LRU because our purpose is to control amount 233of used pages; not-on-LRU pages tend to be out-of-control from VM view. 234 2352.3 Shared Page Accounting 236-------------------------- 237 238Shared pages are accounted on the basis of the first touch approach. The 239cgroup that first touches a page is accounted for the page. The principle 240behind this approach is that a cgroup that aggressively uses a shared 241page will eventually get charged for it (once it is uncharged from 242the cgroup that brought it in -- this will happen on memory pressure). 243 2442.4 Swap Extension 245-------------------------------------- 246 247Swap usage is always recorded for each of cgroup. Swap Extension allows you to 248read and limit it. 249 250When CONFIG_SWAP is enabled, following files are added. 251 252 - memory.memsw.usage_in_bytes. 253 - memory.memsw.limit_in_bytes. 254 255memsw means memory+swap. Usage of memory+swap is limited by 256memsw.limit_in_bytes. 257 258Example: Assume a system with 4G of swap. A task which allocates 6G of memory 259(by mistake) under 2G memory limitation will use all swap. 260In this case, setting memsw.limit_in_bytes=3G will prevent bad use of swap. 261By using the memsw limit, you can avoid system OOM which can be caused by swap 262shortage. 263 2642.4.1 why 'memory+swap' rather than swap 265~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 266 267The global LRU(kswapd) can swap out arbitrary pages. Swap-out means 268to move account from memory to swap...there is no change in usage of 269memory+swap. In other words, when we want to limit the usage of swap without 270affecting global LRU, memory+swap limit is better than just limiting swap from 271an OS point of view. 272 2732.4.2. What happens when a cgroup hits memory.memsw.limit_in_bytes 274~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 275 276When a cgroup hits memory.memsw.limit_in_bytes, it's useless to do swap-out 277in this cgroup. Then, swap-out will not be done by cgroup routine and file 278caches are dropped. But as mentioned above, global LRU can do swapout memory 279from it for sanity of the system's memory management state. You can't forbid 280it by cgroup. 281 2822.5 Reclaim 283----------- 284 285Each cgroup maintains a per cgroup LRU which has the same structure as 286global VM. When a cgroup goes over its limit, we first try 287to reclaim memory from the cgroup so as to make space for the new 288pages that the cgroup has touched. If the reclaim is unsuccessful, 289an OOM routine is invoked to select and kill the bulkiest task in the 290cgroup. (See :ref:`10. OOM Control <cgroup-v1-memory-oom-control>` below.) 291 292The reclaim algorithm has not been modified for cgroups, except that 293pages that are selected for reclaiming come from the per-cgroup LRU 294list. 295 296.. note:: 297 Reclaim does not work for the root cgroup, since we cannot set any 298 limits on the root cgroup. 299 300.. note:: 301 When panic_on_oom is set to "2", the whole system will panic. 302 303When oom event notifier is registered, event will be delivered. 304(See :ref:`oom_control <cgroup-v1-memory-oom-control>` section) 305 3062.6 Locking 307----------- 308 309Lock order is as follows:: 310 311 folio_lock 312 mm->page_table_lock or split pte_lock 313 folio_memcg_lock (memcg->move_lock) 314 mapping->i_pages lock 315 lruvec->lru_lock. 316 317Per-node-per-memcgroup LRU (cgroup's private LRU) is guarded by 318lruvec->lru_lock; the folio LRU flag is cleared before 319isolating a page from its LRU under lruvec->lru_lock. 320 321.. _cgroup-v1-memory-kernel-extension: 322 3232.7 Kernel Memory Extension 324----------------------------------------------- 325 326With the Kernel memory extension, the Memory Controller is able to limit 327the amount of kernel memory used by the system. Kernel memory is fundamentally 328different than user memory, since it can't be swapped out, which makes it 329possible to DoS the system by consuming too much of this precious resource. 330 331Kernel memory accounting is enabled for all memory cgroups by default. But 332it can be disabled system-wide by passing cgroup.memory=nokmem to the kernel 333at boot time. In this case, kernel memory will not be accounted at all. 334 335Kernel memory limits are not imposed for the root cgroup. Usage for the root 336cgroup may or may not be accounted. The memory used is accumulated into 337memory.kmem.usage_in_bytes, or in a separate counter when it makes sense. 338(currently only for tcp). 339 340The main "kmem" counter is fed into the main counter, so kmem charges will 341also be visible from the user counter. 342 343Currently no soft limit is implemented for kernel memory. It is future work 344to trigger slab reclaim when those limits are reached. 345 3462.7.1 Current Kernel Memory resources accounted 347----------------------------------------------- 348 349stack pages: 350 every process consumes some stack pages. By accounting into 351 kernel memory, we prevent new processes from being created when the kernel 352 memory usage is too high. 353 354slab pages: 355 pages allocated by the SLAB or SLUB allocator are tracked. A copy 356 of each kmem_cache is created every time the cache is touched by the first time 357 from inside the memcg. The creation is done lazily, so some objects can still be 358 skipped while the cache is being created. All objects in a slab page should 359 belong to the same memcg. This only fails to hold when a task is migrated to a 360 different memcg during the page allocation by the cache. 361 362sockets memory pressure: 363 some sockets protocols have memory pressure 364 thresholds. The Memory Controller allows them to be controlled individually 365 per cgroup, instead of globally. 366 367tcp memory pressure: 368 sockets memory pressure for the tcp protocol. 369 3702.7.2 Common use cases 371---------------------- 372 373Because the "kmem" counter is fed to the main user counter, kernel memory can 374never be limited completely independently of user memory. Say "U" is the user 375limit, and "K" the kernel limit. There are three possible ways limits can be 376set: 377 378U != 0, K = unlimited: 379 This is the standard memcg limitation mechanism already present before kmem 380 accounting. Kernel memory is completely ignored. 381 382U != 0, K < U: 383 Kernel memory is a subset of the user memory. This setup is useful in 384 deployments where the total amount of memory per-cgroup is overcommitted. 385 Overcommitting kernel memory limits is definitely not recommended, since the 386 box can still run out of non-reclaimable memory. 387 In this case, the admin could set up K so that the sum of all groups is 388 never greater than the total memory, and freely set U at the cost of his 389 QoS. 390 391 .. warning:: 392 In the current implementation, memory reclaim will NOT be triggered for 393 a cgroup when it hits K while staying below U, which makes this setup 394 impractical. 395 396U != 0, K >= U: 397 Since kmem charges will also be fed to the user counter and reclaim will be 398 triggered for the cgroup for both kinds of memory. This setup gives the 399 admin a unified view of memory, and it is also useful for people who just 400 want to track kernel memory usage. 401 4023. User Interface 403================= 404 405To use the user interface: 406 4071. Enable CONFIG_CGROUPS and CONFIG_MEMCG options 4082. Prepare the cgroups (see :ref:`Why are cgroups needed? 409 <cgroups-why-needed>` for the background information):: 410 411 # mount -t tmpfs none /sys/fs/cgroup 412 # mkdir /sys/fs/cgroup/memory 413 # mount -t cgroup none /sys/fs/cgroup/memory -o memory 414 4153. Make the new group and move bash into it:: 416 417 # mkdir /sys/fs/cgroup/memory/0 418 # echo $$ > /sys/fs/cgroup/memory/0/tasks 419 4204. Since now we're in the 0 cgroup, we can alter the memory limit:: 421 422 # echo 4M > /sys/fs/cgroup/memory/0/memory.limit_in_bytes 423 424 The limit can now be queried:: 425 426 # cat /sys/fs/cgroup/memory/0/memory.limit_in_bytes 427 4194304 428 429.. note:: 430 We can use a suffix (k, K, m, M, g or G) to indicate values in kilo, 431 mega or gigabytes. (Here, Kilo, Mega, Giga are Kibibytes, Mebibytes, 432 Gibibytes.) 433 434.. note:: 435 We can write "-1" to reset the ``*.limit_in_bytes(unlimited)``. 436 437.. note:: 438 We cannot set limits on the root cgroup any more. 439 440 441We can check the usage:: 442 443 # cat /sys/fs/cgroup/memory/0/memory.usage_in_bytes 444 1216512 445 446A successful write to this file does not guarantee a successful setting of 447this limit to the value written into the file. This can be due to a 448number of factors, such as rounding up to page boundaries or the total 449availability of memory on the system. The user is required to re-read 450this file after a write to guarantee the value committed by the kernel:: 451 452 # echo 1 > memory.limit_in_bytes 453 # cat memory.limit_in_bytes 454 4096 455 456The memory.failcnt field gives the number of times that the cgroup limit was 457exceeded. 458 459The memory.stat file gives accounting information. Now, the number of 460caches, RSS and Active pages/Inactive pages are shown. 461 4624. Testing 463========== 464 465For testing features and implementation, see memcg_test.txt. 466 467Performance test is also important. To see pure memory controller's overhead, 468testing on tmpfs will give you good numbers of small overheads. 469Example: do kernel make on tmpfs. 470 471Page-fault scalability is also important. At measuring parallel 472page fault test, multi-process test may be better than multi-thread 473test because it has noise of shared objects/status. 474 475But the above two are testing extreme situations. 476Trying usual test under memory controller is always helpful. 477 478.. _cgroup-v1-memory-test-troubleshoot: 479 4804.1 Troubleshooting 481------------------- 482 483Sometimes a user might find that the application under a cgroup is 484terminated by the OOM killer. There are several causes for this: 485 4861. The cgroup limit is too low (just too low to do anything useful) 4872. The user is using anonymous memory and swap is turned off or too low 488 489A sync followed by echo 1 > /proc/sys/vm/drop_caches will help get rid of 490some of the pages cached in the cgroup (page cache pages). 491 492To know what happens, disabling OOM_Kill as per :ref:`"10. OOM Control" 493<cgroup-v1-memory-oom-control>` (below) and seeing what happens will be 494helpful. 495 496.. _cgroup-v1-memory-test-task-migration: 497 4984.2 Task migration 499------------------ 500 501When a task migrates from one cgroup to another, its charge is not 502carried forward by default. The pages allocated from the original cgroup still 503remain charged to it, the charge is dropped when the page is freed or 504reclaimed. 505 506You can move charges of a task along with task migration. 507See :ref:`8. "Move charges at task migration" <cgroup-v1-memory-move-charges>` 508 5094.3 Removing a cgroup 510--------------------- 511 512A cgroup can be removed by rmdir, but as discussed in :ref:`sections 4.1 513<cgroup-v1-memory-test-troubleshoot>` and :ref:`4.2 514<cgroup-v1-memory-test-task-migration>`, a cgroup might have some charge 515associated with it, even though all tasks have migrated away from it. (because 516we charge against pages, not against tasks.) 517 518We move the stats to parent, and no change on the charge except uncharging 519from the child. 520 521Charges recorded in swap information is not updated at removal of cgroup. 522Recorded information is discarded and a cgroup which uses swap (swapcache) 523will be charged as a new owner of it. 524 5255. Misc. interfaces 526=================== 527 5285.1 force_empty 529--------------- 530 memory.force_empty interface is provided to make cgroup's memory usage empty. 531 When writing anything to this:: 532 533 # echo 0 > memory.force_empty 534 535 the cgroup will be reclaimed and as many pages reclaimed as possible. 536 537 The typical use case for this interface is before calling rmdir(). 538 Though rmdir() offlines memcg, but the memcg may still stay there due to 539 charged file caches. Some out-of-use page caches may keep charged until 540 memory pressure happens. If you want to avoid that, force_empty will be useful. 541 5425.2 stat file 543------------- 544 545memory.stat file includes following statistics: 546 547 * per-memory cgroup local status 548 549 =============== =============================================================== 550 cache # of bytes of page cache memory. 551 rss # of bytes of anonymous and swap cache memory (includes 552 transparent hugepages). 553 rss_huge # of bytes of anonymous transparent hugepages. 554 mapped_file # of bytes of mapped file (includes tmpfs/shmem) 555 pgpgin # of charging events to the memory cgroup. The charging 556 event happens each time a page is accounted as either mapped 557 anon page(RSS) or cache page(Page Cache) to the cgroup. 558 pgpgout # of uncharging events to the memory cgroup. The uncharging 559 event happens each time a page is unaccounted from the 560 cgroup. 561 swap # of bytes of swap usage 562 swapcached # of bytes of swap cached in memory 563 dirty # of bytes that are waiting to get written back to the disk. 564 writeback # of bytes of file/anon cache that are queued for syncing to 565 disk. 566 inactive_anon # of bytes of anonymous and swap cache memory on inactive 567 LRU list. 568 active_anon # of bytes of anonymous and swap cache memory on active 569 LRU list. 570 inactive_file # of bytes of file-backed memory and MADV_FREE anonymous 571 memory (LazyFree pages) on inactive LRU list. 572 active_file # of bytes of file-backed memory on active LRU list. 573 unevictable # of bytes of memory that cannot be reclaimed (mlocked etc). 574 =============== =============================================================== 575 576 * status considering hierarchy (see memory.use_hierarchy settings): 577 578 ========================= =================================================== 579 hierarchical_memory_limit # of bytes of memory limit with regard to 580 hierarchy 581 under which the memory cgroup is 582 hierarchical_memsw_limit # of bytes of memory+swap limit with regard to 583 hierarchy under which memory cgroup is. 584 585 total_<counter> # hierarchical version of <counter>, which in 586 addition to the cgroup's own value includes the 587 sum of all hierarchical children's values of 588 <counter>, i.e. total_cache 589 ========================= =================================================== 590 591 * additional vm parameters (depends on CONFIG_DEBUG_VM): 592 593 ========================= ======================================== 594 recent_rotated_anon VM internal parameter. (see mm/vmscan.c) 595 recent_rotated_file VM internal parameter. (see mm/vmscan.c) 596 recent_scanned_anon VM internal parameter. (see mm/vmscan.c) 597 recent_scanned_file VM internal parameter. (see mm/vmscan.c) 598 ========================= ======================================== 599 600.. hint:: 601 recent_rotated means recent frequency of LRU rotation. 602 recent_scanned means recent # of scans to LRU. 603 showing for better debug please see the code for meanings. 604 605.. note:: 606 Only anonymous and swap cache memory is listed as part of 'rss' stat. 607 This should not be confused with the true 'resident set size' or the 608 amount of physical memory used by the cgroup. 609 610 'rss + mapped_file" will give you resident set size of cgroup. 611 612 (Note: file and shmem may be shared among other cgroups. In that case, 613 mapped_file is accounted only when the memory cgroup is owner of page 614 cache.) 615 6165.3 swappiness 617-------------- 618 619Overrides /proc/sys/vm/swappiness for the particular group. The tunable 620in the root cgroup corresponds to the global swappiness setting. 621 622Please note that unlike during the global reclaim, limit reclaim 623enforces that 0 swappiness really prevents from any swapping even if 624there is a swap storage available. This might lead to memcg OOM killer 625if there are no file pages to reclaim. 626 6275.4 failcnt 628----------- 629 630A memory cgroup provides memory.failcnt and memory.memsw.failcnt files. 631This failcnt(== failure count) shows the number of times that a usage counter 632hit its limit. When a memory cgroup hits a limit, failcnt increases and 633memory under it will be reclaimed. 634 635You can reset failcnt by writing 0 to failcnt file:: 636 637 # echo 0 > .../memory.failcnt 638 6395.5 usage_in_bytes 640------------------ 641 642For efficiency, as other kernel components, memory cgroup uses some optimization 643to avoid unnecessary cacheline false sharing. usage_in_bytes is affected by the 644method and doesn't show 'exact' value of memory (and swap) usage, it's a fuzz 645value for efficient access. (Of course, when necessary, it's synchronized.) 646If you want to know more exact memory usage, you should use RSS+CACHE(+SWAP) 647value in memory.stat(see 5.2). 648 6495.6 numa_stat 650------------- 651 652This is similar to numa_maps but operates on a per-memcg basis. This is 653useful for providing visibility into the numa locality information within 654an memcg since the pages are allowed to be allocated from any physical 655node. One of the use cases is evaluating application performance by 656combining this information with the application's CPU allocation. 657 658Each memcg's numa_stat file includes "total", "file", "anon" and "unevictable" 659per-node page counts including "hierarchical_<counter>" which sums up all 660hierarchical children's values in addition to the memcg's own value. 661 662The output format of memory.numa_stat is:: 663 664 total=<total pages> N0=<node 0 pages> N1=<node 1 pages> ... 665 file=<total file pages> N0=<node 0 pages> N1=<node 1 pages> ... 666 anon=<total anon pages> N0=<node 0 pages> N1=<node 1 pages> ... 667 unevictable=<total anon pages> N0=<node 0 pages> N1=<node 1 pages> ... 668 hierarchical_<counter>=<counter pages> N0=<node 0 pages> N1=<node 1 pages> ... 669 670The "total" count is sum of file + anon + unevictable. 671 6726. Hierarchy support 673==================== 674 675The memory controller supports a deep hierarchy and hierarchical accounting. 676The hierarchy is created by creating the appropriate cgroups in the 677cgroup filesystem. Consider for example, the following cgroup filesystem 678hierarchy:: 679 680 root 681 / | \ 682 / | \ 683 a b c 684 | \ 685 | \ 686 d e 687 688In the diagram above, with hierarchical accounting enabled, all memory 689usage of e, is accounted to its ancestors up until the root (i.e, c and root). 690If one of the ancestors goes over its limit, the reclaim algorithm reclaims 691from the tasks in the ancestor and the children of the ancestor. 692 6936.1 Hierarchical accounting and reclaim 694--------------------------------------- 695 696Hierarchical accounting is enabled by default. Disabling the hierarchical 697accounting is deprecated. An attempt to do it will result in a failure 698and a warning printed to dmesg. 699 700For compatibility reasons writing 1 to memory.use_hierarchy will always pass:: 701 702 # echo 1 > memory.use_hierarchy 703 7047. Soft limits (DEPRECATED) 705=========================== 706 707THIS IS DEPRECATED! 708 709Soft limits allow for greater sharing of memory. The idea behind soft limits 710is to allow control groups to use as much of the memory as needed, provided 711 712a. There is no memory contention 713b. They do not exceed their hard limit 714 715When the system detects memory contention or low memory, control groups 716are pushed back to their soft limits. If the soft limit of each control 717group is very high, they are pushed back as much as possible to make 718sure that one control group does not starve the others of memory. 719 720Please note that soft limits is a best-effort feature; it comes with 721no guarantees, but it does its best to make sure that when memory is 722heavily contended for, memory is allocated based on the soft limit 723hints/setup. Currently soft limit based reclaim is set up such that 724it gets invoked from balance_pgdat (kswapd). 725 7267.1 Interface 727------------- 728 729Soft limits can be setup by using the following commands (in this example we 730assume a soft limit of 256 MiB):: 731 732 # echo 256M > memory.soft_limit_in_bytes 733 734If we want to change this to 1G, we can at any time use:: 735 736 # echo 1G > memory.soft_limit_in_bytes 737 738.. note:: 739 Soft limits take effect over a long period of time, since they involve 740 reclaiming memory for balancing between memory cgroups 741 742.. note:: 743 It is recommended to set the soft limit always below the hard limit, 744 otherwise the hard limit will take precedence. 745 746.. _cgroup-v1-memory-move-charges: 747 7488. Move charges at task migration (DEPRECATED!) 749=============================================== 750 751THIS IS DEPRECATED! 752 753Reading memory.move_charge_at_immigrate will always return 0 and writing 754to it will always return -EINVAL. 755 7569. Memory thresholds 757==================== 758 759Memory cgroup implements memory thresholds using the cgroups notification 760API (see cgroups.txt). It allows to register multiple memory and memsw 761thresholds and gets notifications when it crosses. 762 763To register a threshold, an application must: 764 765- create an eventfd using eventfd(2); 766- open memory.usage_in_bytes or memory.memsw.usage_in_bytes; 767- write string like "<event_fd> <fd of memory.usage_in_bytes> <threshold>" to 768 cgroup.event_control. 769 770Application will be notified through eventfd when memory usage crosses 771threshold in any direction. 772 773It's applicable for root and non-root cgroup. 774 775.. _cgroup-v1-memory-oom-control: 776 77710. OOM Control (DEPRECATED) 778============================ 779 780THIS IS DEPRECATED! 781 782memory.oom_control file is for OOM notification and other controls. 783 784Memory cgroup implements OOM notifier using the cgroup notification 785API (See cgroups.txt). It allows to register multiple OOM notification 786delivery and gets notification when OOM happens. 787 788To register a notifier, an application must: 789 790 - create an eventfd using eventfd(2) 791 - open memory.oom_control file 792 - write string like "<event_fd> <fd of memory.oom_control>" to 793 cgroup.event_control 794 795The application will be notified through eventfd when OOM happens. 796OOM notification doesn't work for the root cgroup. 797 798You can disable the OOM-killer by writing "1" to memory.oom_control file, as: 799 800 #echo 1 > memory.oom_control 801 802If OOM-killer is disabled, tasks under cgroup will hang/sleep 803in memory cgroup's OOM-waitqueue when they request accountable memory. 804 805For running them, you have to relax the memory cgroup's OOM status by 806 807 * enlarge limit or reduce usage. 808 809To reduce usage, 810 811 * kill some tasks. 812 * move some tasks to other group with account migration. 813 * remove some files (on tmpfs?) 814 815Then, stopped tasks will work again. 816 817At reading, current status of OOM is shown. 818 819 - oom_kill_disable 0 or 1 820 (if 1, oom-killer is disabled) 821 - under_oom 0 or 1 822 (if 1, the memory cgroup is under OOM, tasks may be stopped.) 823 - oom_kill integer counter 824 The number of processes belonging to this cgroup killed by any 825 kind of OOM killer. 826 82711. Memory Pressure (DEPRECATED) 828================================ 829 830THIS IS DEPRECATED! 831 832The pressure level notifications can be used to monitor the memory 833allocation cost; based on the pressure, applications can implement 834different strategies of managing their memory resources. The pressure 835levels are defined as following: 836 837The "low" level means that the system is reclaiming memory for new 838allocations. Monitoring this reclaiming activity might be useful for 839maintaining cache level. Upon notification, the program (typically 840"Activity Manager") might analyze vmstat and act in advance (i.e. 841prematurely shutdown unimportant services). 842 843The "medium" level means that the system is experiencing medium memory 844pressure, the system might be making swap, paging out active file caches, 845etc. Upon this event applications may decide to further analyze 846vmstat/zoneinfo/memcg or internal memory usage statistics and free any 847resources that can be easily reconstructed or re-read from a disk. 848 849The "critical" level means that the system is actively thrashing, it is 850about to out of memory (OOM) or even the in-kernel OOM killer is on its 851way to trigger. Applications should do whatever they can to help the 852system. It might be too late to consult with vmstat or any other 853statistics, so it's advisable to take an immediate action. 854 855By default, events are propagated upward until the event is handled, i.e. the 856events are not pass-through. For example, you have three cgroups: A->B->C. Now 857you set up an event listener on cgroups A, B and C, and suppose group C 858experiences some pressure. In this situation, only group C will receive the 859notification, i.e. groups A and B will not receive it. This is done to avoid 860excessive "broadcasting" of messages, which disturbs the system and which is 861especially bad if we are low on memory or thrashing. Group B, will receive 862notification only if there are no event listeners for group C. 863 864There are three optional modes that specify different propagation behavior: 865 866 - "default": this is the default behavior specified above. This mode is the 867 same as omitting the optional mode parameter, preserved by backwards 868 compatibility. 869 870 - "hierarchy": events always propagate up to the root, similar to the default 871 behavior, except that propagation continues regardless of whether there are 872 event listeners at each level, with the "hierarchy" mode. In the above 873 example, groups A, B, and C will receive notification of memory pressure. 874 875 - "local": events are pass-through, i.e. they only receive notifications when 876 memory pressure is experienced in the memcg for which the notification is 877 registered. In the above example, group C will receive notification if 878 registered for "local" notification and the group experiences memory 879 pressure. However, group B will never receive notification, regardless if 880 there is an event listener for group C or not, if group B is registered for 881 local notification. 882 883The level and event notification mode ("hierarchy" or "local", if necessary) are 884specified by a comma-delimited string, i.e. "low,hierarchy" specifies 885hierarchical, pass-through, notification for all ancestor memcgs. Notification 886that is the default, non pass-through behavior, does not specify a mode. 887"medium,local" specifies pass-through notification for the medium level. 888 889The file memory.pressure_level is only used to setup an eventfd. To 890register a notification, an application must: 891 892- create an eventfd using eventfd(2); 893- open memory.pressure_level; 894- write string as "<event_fd> <fd of memory.pressure_level> <level[,mode]>" 895 to cgroup.event_control. 896 897Application will be notified through eventfd when memory pressure is at 898the specific level (or higher). Read/write operations to 899memory.pressure_level are no implemented. 900 901Test: 902 903 Here is a small script example that makes a new cgroup, sets up a 904 memory limit, sets up a notification in the cgroup and then makes child 905 cgroup experience a critical pressure:: 906 907 # cd /sys/fs/cgroup/memory/ 908 # mkdir foo 909 # cd foo 910 # cgroup_event_listener memory.pressure_level low,hierarchy & 911 # echo 8000000 > memory.limit_in_bytes 912 # echo 8000000 > memory.memsw.limit_in_bytes 913 # echo $$ > tasks 914 # dd if=/dev/zero | read x 915 916 (Expect a bunch of notifications, and eventually, the oom-killer will 917 trigger.) 918 91912. TODO 920======== 921 9221. Make per-cgroup scanner reclaim not-shared pages first 9232. Teach controller to account for shared-pages 9243. Start reclamation in the background when the limit is 925 not yet hit but the usage is getting closer 926 927Summary 928======= 929 930Overall, the memory controller has been a stable controller and has been 931commented and discussed quite extensively in the community. 932 933References 934========== 935 936.. [1] Singh, Balbir. RFC: Memory Controller, http://lwn.net/Articles/206697/ 937.. [2] Singh, Balbir. Memory Controller (RSS Control), 938 http://lwn.net/Articles/222762/ 939.. [3] Emelianov, Pavel. Resource controllers based on process cgroups 940 https://lore.kernel.org/r/45ED7DEC.7010403@sw.ru 941.. [4] Emelianov, Pavel. RSS controller based on process cgroups (v2) 942 https://lore.kernel.org/r/461A3010.90403@sw.ru 943.. [5] Emelianov, Pavel. RSS controller based on process cgroups (v3) 944 https://lore.kernel.org/r/465D9739.8070209@openvz.org 945 9466. Menage, Paul. Control Groups v10, http://lwn.net/Articles/236032/ 9477. Vaidyanathan, Srinivasan, Control Groups: Pagecache accounting and control 948 subsystem (v3), http://lwn.net/Articles/235534/ 9498. Singh, Balbir. RSS controller v2 test results (lmbench), 950 https://lore.kernel.org/r/464C95D4.7070806@linux.vnet.ibm.com 9519. Singh, Balbir. RSS controller v2 AIM9 results 952 https://lore.kernel.org/r/464D267A.50107@linux.vnet.ibm.com 95310. Singh, Balbir. Memory controller v6 test results, 954 https://lore.kernel.org/r/20070819094658.654.84837.sendpatchset@balbir-laptop 955 956.. [11] Singh, Balbir. Memory controller introduction (v6), 957 https://lore.kernel.org/r/20070817084228.26003.12568.sendpatchset@balbir-laptop 958.. [12] Corbet, Jonathan, Controlling memory use in cgroups, 959 http://lwn.net/Articles/243795/ 960