1Memory Resource Controller
   3NOTE: The Memory Resource Controller has generically been referred to as the
   4      memory controller in this document. Do not confuse memory controller
   5      used here with the memory controller that is used in hardware.
   7(For editors)
   8In this document:
   9      When we mention a cgroup (cgroupfs's directory) with memory controller,
  10      we call it "memory cgroup". When you see git-log and source code, you'll
  11      see patch's title and function names tend to use "memcg".
  12      In this document, we avoid using it.
  14Benefits and Purpose of the memory controller
  16The memory controller isolates the memory behaviour of a group of tasks
  17from the rest of the system. The article on LWN [12] mentions some probable
  18uses of the memory controller. The memory controller can be used to
  20a. Isolate an application or a group of applications
  21   Memory hungry applications can be isolated and limited to a smaller
  22   amount of memory.
  23b. Create a cgroup with limited amount of memory, this can be used
  24   as a good alternative to booting with mem=XXXX.
  25c. Virtualization solutions can control the amount of memory they want
  26   to assign to a virtual machine instance.
  27d. A CD/DVD burner could control the amount of memory used by the
  28   rest of the system to ensure that burning does not fail due to lack
  29   of available memory.
  30e. There are several other use cases, find one or use the controller just
  31   for fun (to learn and hack on the VM subsystem).
  33Current Status: linux-2.6.34-mmotm(development version of 2010/April)
  36 - accounting anonymous pages, file caches, swap caches usage and limiting them.
  37 - private LRU and reclaim routine. (system's global LRU and private LRU
  38   work independently from each other)
  39 - optionally, memory+swap usage can be accounted and limited.
  40 - hierarchical accounting
  41 - soft limit
  42 - moving(recharging) account at moving a task is selectable.
  43 - usage threshold notifier
  44 - oom-killer disable knob and oom-notifier
  45 - Root cgroup has no limit controls.
  47 Kernel memory support is work in progress, and the current version provides
  48 basically functionality. (See Section 2.7)
  50Brief summary of control files.
  52 tasks                           # attach a task(thread) and show list of threads
  53 cgroup.procs                    # show list of processes
  54 cgroup.event_control            # an interface for event_fd()
  55 memory.usage_in_bytes           # show current res_counter usage for memory
  56                                 (See 5.5 for details)
  57 memory.memsw.usage_in_bytes     # show current res_counter usage for memory+Swap
  58                                 (See 5.5 for details)
  59 memory.limit_in_bytes           # set/show limit of memory usage
  60 memory.memsw.limit_in_bytes     # set/show limit of memory+Swap usage
  61 memory.failcnt                  # show the number of memory usage hits limits
  62 memory.memsw.failcnt            # show the number of memory+Swap hits limits
  63 memory.max_usage_in_bytes       # show max memory usage recorded
  64 memory.memsw.max_usage_in_bytes # show max memory+Swap usage recorded
  65 memory.soft_limit_in_bytes      # set/show soft limit of memory usage
  66 memory.stat                     # show various statistics
  67 memory.use_hierarchy            # set/show hierarchical account enabled
  68 memory.force_empty              # trigger forced move charge to parent
  69 memory.swappiness               # set/show swappiness parameter of vmscan
  70                                 (See sysctl's vm.swappiness)
  71 memory.move_charge_at_immigrate # set/show controls of moving charges
  72 memory.oom_control              # set/show oom controls.
  73 memory.numa_stat                # show the number of memory usage per numa node
  75 memory.kmem.tcp.limit_in_bytes  # set/show hard limit for tcp buf memory
  76 memory.kmem.tcp.usage_in_bytes  # show current tcp buf memory allocation
  781. History
  80The memory controller has a long history. A request for comments for the memory
  81controller was posted by Balbir Singh [1]. At the time the RFC was posted
  82there were several implementations for memory control. The goal of the
  83RFC was to build consensus and agreement for the minimal features required
  84for memory control. The first RSS controller was posted by Balbir Singh[2]
  85in Feb 2007. Pavel Emelianov [3][4][5] has since posted three versions of the
  86RSS controller. At OLS, at the resource management BoF, everyone suggested
  87that we handle both page cache and RSS together. Another request was raised
  88to allow user space handling of OOM. The current memory controller is
  89at version 6; it combines both mapped (RSS) and unmapped Page
  90Cache Control [11].
  922. Memory Control
  94Memory is a unique resource in the sense that it is present in a limited
  95amount. If a task requires a lot of CPU processing, the task can spread
  96its processing over a period of hours, days, months or years, but with
  97memory, the same physical memory needs to be reused to accomplish the task.
  99The memory controller implementation has been divided into phases. These
 1021. Memory controller
 1032. mlock(2) controller
 1043. Kernel user memory accounting and slab control
 1054. user mappings length controller
 107The memory controller is the first controller developed.
 1092.1. Design
 111The core of the design is a counter called the res_counter. The res_counter
 112tracks the current memory usage and limit of the group of processes associated
 113with the controller. Each cgroup has a memory controller specific data
 114structure (mem_cgroup) associated with it.
 1162.2. Accounting
 118                +--------------------+
 119                |  mem_cgroup     |
 120                |  (res_counter)     |
 121                +--------------------+
 122                 /            ^      \
 123                /             |       \
 124           +---------------+  |        +---------------+
 125           | mm_struct     |  |....    | mm_struct     |
 126           |               |  |        |               |
 127           +---------------+  |        +---------------+
 128                              |
 129                              + --------------+
 130                                              |
 131           +---------------+           +------+--------+
 132           | page          +---------->  page_cgroup|
 133           |               |           |               |
 134           +---------------+           +---------------+
 136             (Figure 1: Hierarchy of Accounting)
 139Figure 1 shows the important aspects of the controller
 1411. Accounting happens per cgroup
 1422. Each mm_struct knows about which cgroup it belongs to
 1433. Each page has a pointer to the page_cgroup, which in turn knows the
 144   cgroup it belongs to
 146The accounting is done as follows: mem_cgroup_charge() is invoked to setup
 147the necessary data structures and check if the cgroup that is being charged
 148is over its limit. If it is then reclaim is invoked on the cgroup.
 149More details can be found in the reclaim section of this document.
 150If everything goes well, a page meta-data-structure called page_cgroup is
 151updated. page_cgroup has its own LRU on cgroup.
 152(*) page_cgroup structure is allocated at boot/memory-hotplug time.
 1542.2.1 Accounting details
 156All mapped anon pages (RSS) and cache pages (Page Cache) are accounted.
 157Some pages which are never reclaimable and will not be on the global LRU
 158are not accounted. We just account pages under usual VM management.
 160RSS pages are accounted at page_fault unless they've already been accounted
 161for earlier. A file page will be accounted for as Page Cache when it's
 162inserted into inode (radix-tree). While it's mapped into the page tables of
 163processes, duplicate accounting is carefully avoided.
 165A RSS page is unaccounted when it's fully unmapped. A PageCache page is
 166unaccounted when it's removed from radix-tree. Even if RSS pages are fully
 167unmapped (by kswapd), they may exist as SwapCache in the system until they
 168are really freed. Such SwapCaches also also accounted.
 169A swapped-in page is not accounted until it's mapped.
 171Note: The kernel does swapin-readahead and read multiple swaps at once.
 172This means swapped-in pages may contain pages for other tasks than a task
 173causing page fault. So, we avoid accounting at swap-in I/O.
 175At page migration, accounting information is kept.
 177Note: we just account pages-on-LRU because our purpose is to control amount
 178of used pages; not-on-LRU pages tend to be out-of-control from VM view.
 1802.3 Shared Page Accounting
 182Shared pages are accounted on the basis of the first touch approach. The
 183cgroup that first touches a page is accounted for the page. The principle
 184behind this approach is that a cgroup that aggressively uses a shared
 185page will eventually get charged for it (once it is uncharged from
 186the cgroup that brought it in -- this will happen on memory pressure).
 188Exception: If CONFIG_CGROUP_CGROUP_MEM_RES_CTLR_SWAP is not used.
 189When you do swapoff and make swapped-out pages of shmem(tmpfs) to
 190be backed into memory in force, charges for pages are accounted against the
 191caller of swapoff rather than the users of shmem.
 1942.4 Swap Extension (CONFIG_CGROUP_MEM_RES_CTLR_SWAP)
 196Swap Extension allows you to record charge for swap. A swapped-in page is
 197charged back to original page allocator if possible.
 199When swap is accounted, following files are added.
 200 - memory.memsw.usage_in_bytes.
 201 - memory.memsw.limit_in_bytes.
 203memsw means memory+swap. Usage of memory+swap is limited by
 206Example: Assume a system with 4G of swap. A task which allocates 6G of memory
 207(by mistake) under 2G memory limitation will use all swap.
 208In this case, setting memsw.limit_in_bytes=3G will prevent bad use of swap.
 209By using memsw limit, you can avoid system OOM which can be caused by swap
 212* why 'memory+swap' rather than swap.
 213The global LRU(kswapd) can swap out arbitrary pages. Swap-out means
 214to move account from memory to swap...there is no change in usage of
 215memory+swap. In other words, when we want to limit the usage of swap without
 216affecting global LRU, memory+swap limit is better than just limiting swap from
 217OS point of view.
 219* What happens when a cgroup hits memory.memsw.limit_in_bytes
 220When a cgroup hits memory.memsw.limit_in_bytes, it's useless to do swap-out
 221in this cgroup. Then, swap-out will not be done by cgroup routine and file
 222caches are dropped. But as mentioned above, global LRU can do swapout memory
 223from it for sanity of the system's memory management state. You can't forbid
 224it by cgroup.
 2262.5 Reclaim
 228Each cgroup maintains a per cgroup LRU which has the same structure as
 229global VM. When a cgroup goes over its limit, we first try
 230to reclaim memory from the cgroup so as to make space for the new
 231pages that the cgroup has touched. If the reclaim is unsuccessful,
 232an OOM routine is invoked to select and kill the bulkiest task in the
 233cgroup. (See 10. OOM Control below.)
 235The reclaim algorithm has not been modified for cgroups, except that
 236pages that are selected for reclaiming come from the per cgroup LRU
 239NOTE: Reclaim does not work for the root cgroup, since we cannot set any
 240limits on the root cgroup.
 242Note2: When panic_on_oom is set to "2", the whole system will panic.
 244When oom event notifier is registered, event will be delivered.
 245(See oom_control section)
 2472.6 Locking
 249   lock_page_cgroup()/unlock_page_cgroup() should not be called under
 250   mapping->tree_lock.
 252   Other lock order is following:
 253   PG_locked.
 254   mm->page_table_lock
 255       zone->lru_lock
 256          lock_page_cgroup.
 257  In many cases, just lock_page_cgroup() is called.
 258  per-zone-per-cgroup LRU (cgroup's private LRU) is just guarded by
 259  zone->lru_lock, it has no lock of its own.
 2612.7 Kernel Memory Extension (CONFIG_CGROUP_MEM_RES_CTLR_KMEM)
 263With the Kernel memory extension, the Memory Controller is able to limit
 264the amount of kernel memory used by the system. Kernel memory is fundamentally
 265different than user memory, since it can't be swapped out, which makes it
 266possible to DoS the system by consuming too much of this precious resource.
 268Kernel memory limits are not imposed for the root cgroup. Usage for the root
 269cgroup may or may not be accounted.
 271Currently no soft limit is implemented for kernel memory. It is future work
 272to trigger slab reclaim when those limits are reached.
 2742.7.1 Current Kernel Memory resources accounted
 276* sockets memory pressure: some sockets protocols have memory pressure
 277thresholds. The Memory Controller allows them to be controlled individually
 278per cgroup, instead of globally.
 280* tcp memory pressure: sockets memory pressure for the tcp protocol.
 2823. User Interface
 2840. Configuration
 286a. Enable CONFIG_CGROUPS
 289d. Enable CONFIG_CGROUP_MEM_RES_CTLR_SWAP (to use swap extension)
 2911. Prepare the cgroups (see cgroups.txt, Why are cgroups needed?)
 292# mount -t tmpfs none /sys/fs/cgroup
 293# mkdir /sys/fs/cgroup/memory
 294# mount -t cgroup none /sys/fs/cgroup/memory -o memory
 2962. Make the new group and move bash into it
 297# mkdir /sys/fs/cgroup/memory/0
 298# echo $$ > /sys/fs/cgroup/memory/0/tasks
 300Since now we're in the 0 cgroup, we can alter the memory limit:
 301# echo 4M > /sys/fs/cgroup/memory/0/memory.limit_in_bytes
 303NOTE: We can use a suffix (k, K, m, M, g or G) to indicate values in kilo,
 304mega or gigabytes. (Here, Kilo, Mega, Giga are Kibibytes, Mebibytes, Gibibytes.)
 306NOTE: We can write "-1" to reset the *.limit_in_bytes(unlimited).
 307NOTE: We cannot set limits on the root cgroup any more.
 309# cat /sys/fs/cgroup/memory/0/memory.limit_in_bytes
 312We can check the usage:
 313# cat /sys/fs/cgroup/memory/0/memory.usage_in_bytes
 316A successful write to this file does not guarantee a successful set of
 317this limit to the value written into the file. This can be due to a
 318number of factors, such as rounding up to page boundaries or the total
 319availability of memory on the system. The user is required to re-read
 320this file after a write to guarantee the value committed by the kernel.
 322# echo 1 > memory.limit_in_bytes
 323# cat memory.limit_in_bytes
 326The memory.failcnt field gives the number of times that the cgroup limit was
 329The memory.stat file gives accounting information. Now, the number of
 330caches, RSS and Active pages/Inactive pages are shown.
 3324. Testing
 334For testing features and implementation, see memcg_test.txt.
 336Performance test is also important. To see pure memory controller's overhead,
 337testing on tmpfs will give you good numbers of small overheads.
 338Example: do kernel make on tmpfs.
 340Page-fault scalability is also important. At measuring parallel
 341page fault test, multi-process test may be better than multi-thread
 342test because it has noise of shared objects/status.
 344But the above two are testing extreme situations.
 345Trying usual test under memory controller is always helpful.
 3474.1 Troubleshooting
 349Sometimes a user might find that the application under a cgroup is
 350terminated by OOM killer. There are several causes for this:
 3521. The cgroup limit is too low (just too low to do anything useful)
 3532. The user is using anonymous memory and swap is turned off or too low
 355A sync followed by echo 1 > /proc/sys/vm/drop_caches will help get rid of
 356some of the pages cached in the cgroup (page cache pages).
 358To know what happens, disable OOM_Kill by 10. OOM Control(see below) and
 359seeing what happens will be helpful.
 3614.2 Task migration
 363When a task migrates from one cgroup to another, its charge is not
 364carried forward by default. The pages allocated from the original cgroup still
 365remain charged to it, the charge is dropped when the page is freed or
 368You can move charges of a task along with task migration.
 369See 8. "Move charges at task migration"
 3714.3 Removing a cgroup
 373A cgroup can be removed by rmdir, but as discussed in sections 4.1 and 4.2, a
 374cgroup might have some charge associated with it, even though all
 375tasks have migrated away from it. (because we charge against pages, not
 376against tasks.)
 378Such charges are freed or moved to their parent. At moving, both of RSS
 379and CACHES are moved to parent.
 380rmdir() may return -EBUSY if freeing/moving fails. See 5.1 also.
 382Charges recorded in swap information is not updated at removal of cgroup.
 383Recorded information is discarded and a cgroup which uses swap (swapcache)
 384will be charged as a new owner of it.
 3875. Misc. interfaces.
 3895.1 force_empty
 390  memory.force_empty interface is provided to make cgroup's memory usage empty.
 391  You can use this interface only when the cgroup has no tasks.
 392  When writing anything to this
 394  # echo 0 > memory.force_empty
 396  Almost all pages tracked by this memory cgroup will be unmapped and freed.
 397  Some pages cannot be freed because they are locked or in-use. Such pages are
 398  moved to parent and this cgroup will be empty. This may return -EBUSY if
 399  VM is too busy to free/move all pages immediately.
 401  Typical use case of this interface is that calling this before rmdir().
 402  Because rmdir() moves all pages to parent, some out-of-use page caches can be
 403  moved to the parent. If you want to avoid that, force_empty will be useful.
 4055.2 stat file
 407memory.stat file includes following statistics
 409# per-memory cgroup local status
 410cache           - # of bytes of page cache memory.
 411rss             - # of bytes of anonymous and swap cache memory.
 412mapped_file     - # of bytes of mapped file (includes tmpfs/shmem)
 413pgpgin          - # of charging events to the memory cgroup. The charging
 414                event happens each time a page is accounted as either mapped
 415                anon page(RSS) or cache page(Page Cache) to the cgroup.
 416pgpgout         - # of uncharging events to the memory cgroup. The uncharging
 417                event happens each time a page is unaccounted from the cgroup.
 418swap            - # of bytes of swap usage
 419inactive_anon   - # of bytes of anonymous memory and swap cache memory on
 420                LRU list.
 421active_anon     - # of bytes of anonymous and swap cache memory on active
 422                inactive LRU list.
 423inactive_file   - # of bytes of file-backed memory on inactive LRU list.
 424active_file     - # of bytes of file-backed memory on active LRU list.
 425unevictable     - # of bytes of memory that cannot be reclaimed (mlocked etc).
 427# status considering hierarchy (see memory.use_hierarchy settings)
 429hierarchical_memory_limit - # of bytes of memory limit with regard to hierarchy
 430                        under which the memory cgroup is
 431hierarchical_memsw_limit - # of bytes of memory+swap limit with regard to
 432                        hierarchy under which memory cgroup is.
 434total_cache             - sum of all children's "cache"
 435total_rss               - sum of all children's "rss"
 436total_mapped_file       - sum of all children's "cache"
 437total_pgpgin            - sum of all children's "pgpgin"
 438total_pgpgout           - sum of all children's "pgpgout"
 439total_swap              - sum of all children's "swap"
 440total_inactive_anon     - sum of all children's "inactive_anon"
 441total_active_anon       - sum of all children's "active_anon"
 442total_inactive_file     - sum of all children's "inactive_file"
 443total_active_file       - sum of all children's "active_file"
 444total_unevictable       - sum of all children's "unevictable"
 446# The following additional stats are dependent on CONFIG_DEBUG_VM.
 448recent_rotated_anon     - VM internal parameter. (see mm/vmscan.c)
 449recent_rotated_file     - VM internal parameter. (see mm/vmscan.c)
 450recent_scanned_anon     - VM internal parameter. (see mm/vmscan.c)
 451recent_scanned_file     - VM internal parameter. (see mm/vmscan.c)
 454        recent_rotated means recent frequency of LRU rotation.
 455        recent_scanned means recent # of scans to LRU.
 456        showing for better debug please see the code for meanings.
 459        Only anonymous and swap cache memory is listed as part of 'rss' stat.
 460        This should not be confused with the true 'resident set size' or the
 461        amount of physical memory used by the cgroup.
 462        'rss + file_mapped" will give you resident set size of cgroup.
 463        (Note: file and shmem may be shared among other cgroups. In that case,
 464         file_mapped is accounted only when the memory cgroup is owner of page
 465         cache.)
 4675.3 swappiness
 469Similar to /proc/sys/vm/swappiness, but affecting a hierarchy of groups only.
 471Following cgroups' swappiness can't be changed.
 472- root cgroup (uses /proc/sys/vm/swappiness).
 473- a cgroup which uses hierarchy and it has other cgroup(s) below it.
 474- a cgroup which uses hierarchy and not the root of hierarchy.
 4765.4 failcnt
 478A memory cgroup provides memory.failcnt and memory.memsw.failcnt files.
 479This failcnt(== failure count) shows the number of times that a usage counter
 480hit its limit. When a memory cgroup hits a limit, failcnt increases and
 481memory under it will be reclaimed.
 483You can reset failcnt by writing 0 to failcnt file.
 484# echo 0 > .../memory.failcnt
 4865.5 usage_in_bytes
 488For efficiency, as other kernel components, memory cgroup uses some optimization
 489to avoid unnecessary cacheline false sharing. usage_in_bytes is affected by the
 490method and doesn't show 'exact' value of memory(and swap) usage, it's an fuzz
 491value for efficient access. (Of course, when necessary, it's synchronized.)
 492If you want to know more exact memory usage, you should use RSS+CACHE(+SWAP)
 493value in memory.stat(see 5.2).
 4955.6 numa_stat
 497This is similar to numa_maps but operates on a per-memcg basis.  This is
 498useful for providing visibility into the numa locality information within
 499an memcg since the pages are allowed to be allocated from any physical
 500node.  One of the usecases is evaluating application performance by
 501combining this information with the application's cpu allocation.
 503We export "total", "file", "anon" and "unevictable" pages per-node for
 504each memcg.  The ouput format of memory.numa_stat is:
 506total=<total pages> N0=<node 0 pages> N1=<node 1 pages> ...
 507file=<total file pages> N0=<node 0 pages> N1=<node 1 pages> ...
 508anon=<total anon pages> N0=<node 0 pages> N1=<node 1 pages> ...
 509unevictable=<total anon pages> N0=<node 0 pages> N1=<node 1 pages> ...
 511And we have total = file + anon + unevictable.
 5136. Hierarchy support
 515The memory controller supports a deep hierarchy and hierarchical accounting.
 516The hierarchy is created by creating the appropriate cgroups in the
 517cgroup filesystem. Consider for example, the following cgroup filesystem
 520               root
 521             /  |   \
 522            /   |    \
 523           a    b     c
 524                      | \
 525                      |  \
 526                      d   e
 528In the diagram above, with hierarchical accounting enabled, all memory
 529usage of e, is accounted to its ancestors up until the root (i.e, c and root),
 530that has memory.use_hierarchy enabled. If one of the ancestors goes over its
 531limit, the reclaim algorithm reclaims from the tasks in the ancestor and the
 532children of the ancestor.
 5346.1 Enabling hierarchical accounting and reclaim
 536A memory cgroup by default disables the hierarchy feature. Support
 537can be enabled by writing 1 to memory.use_hierarchy file of the root cgroup
 539# echo 1 > memory.use_hierarchy
 541The feature can be disabled by
 543# echo 0 > memory.use_hierarchy
 545NOTE1: Enabling/disabling will fail if either the cgroup already has other
 546       cgroups created below it, or if the parent cgroup has use_hierarchy
 547       enabled.
 549NOTE2: When panic_on_oom is set to "2", the whole system will panic in
 550       case of an OOM event in any cgroup.
 5527. Soft limits
 554Soft limits allow for greater sharing of memory. The idea behind soft limits
 555is to allow control groups to use as much of the memory as needed, provided
 557a. There is no memory contention
 558b. They do not exceed their hard limit
 560When the system detects memory contention or low memory, control groups
 561are pushed back to their soft limits. If the soft limit of each control
 562group is very high, they are pushed back as much as possible to make
 563sure that one control group does not starve the others of memory.
 565Please note that soft limits is a best effort feature, it comes with
 566no guarantees, but it does its best to make sure that when memory is
 567heavily contended for, memory is allocated based on the soft limit
 568hints/setup. Currently soft limit based reclaim is setup such that
 569it gets invoked from balance_pgdat (kswapd).
 5717.1 Interface
 573Soft limits can be setup by using the following commands (in this example we
 574assume a soft limit of 256 MiB)
 576# echo 256M > memory.soft_limit_in_bytes
 578If we want to change this to 1G, we can at any time use
 580# echo 1G > memory.soft_limit_in_bytes
 582NOTE1: Soft limits take effect over a long period of time, since they involve
 583       reclaiming memory for balancing between memory cgroups
 584NOTE2: It is recommended to set the soft limit always below the hard limit,
 585       otherwise the hard limit will take precedence.
 5878. Move charges at task migration
 589Users can move charges associated with a task along with task migration, that
 590is, uncharge task's pages from the old cgroup and charge them to the new cgroup.
 591This feature is not supported in !CONFIG_MMU environments because of lack of
 592page tables.
 5948.1 Interface
 596This feature is disabled by default. It can be enabled(and disabled again) by
 597writing to memory.move_charge_at_immigrate of the destination cgroup.
 599If you want to enable it:
 601# echo (some positive value) > memory.move_charge_at_immigrate
 603Note: Each bits of move_charge_at_immigrate has its own meaning about what type
 604      of charges should be moved. See 8.2 for details.
 605Note: Charges are moved only when you move mm->owner, IOW, a leader of a thread
 606      group.
 607Note: If we cannot find enough space for the task in the destination cgroup, we
 608      try to make space by reclaiming memory. Task migration may fail if we
 609      cannot make enough space.
 610Note: It can take several seconds if you move charges much.
 612And if you want disable it again:
 614# echo 0 > memory.move_charge_at_immigrate
 6168.2 Type of charges which can be move
 618Each bits of move_charge_at_immigrate has its own meaning about what type of
 619charges should be moved. But in any cases, it must be noted that an account of
 620a page or a swap can be moved only when it is charged to the task's current(old)
 621memory cgroup.
 623  bit | what type of charges would be moved ?
 624 -----+------------------------------------------------------------------------
 625   0  | A charge of an anonymous page(or swap of it) used by the target task.
 626      | Those pages and swaps must be used only by the target task. You must
 627      | enable Swap Extension(see 2.4) to enable move of swap charges.
 628 -----+------------------------------------------------------------------------
 629   1  | A charge of file pages(normal file, tmpfs file(e.g. ipc shared memory)
 630      | and swaps of tmpfs file) mmapped by the target task. Unlike the case of
 631      | anonymous pages, file pages(and swaps) in the range mmapped by the task
 632      | will be moved even if the task hasn't done page fault, i.e. they might
 633      | not be the task's "RSS", but other task's "RSS" that maps the same file.
 634      | And mapcount of the page is ignored(the page can be moved even if
 635      | page_mapcount(page) > 1). You must enable Swap Extension(see 2.4) to
 636      | enable move of swap charges.
 6388.3 TODO
 640- Implement madvise(2) to let users decide the vma to be moved or not to be
 641  moved.
 642- All of moving charge operations are done under cgroup_mutex. It's not good
 643  behavior to hold the mutex too long, so we may need some trick.
 6459. Memory thresholds
 647Memory cgroup implements memory thresholds using cgroups notification
 648API (see cgroups.txt). It allows to register multiple memory and memsw
 649thresholds and gets notifications when it crosses.
 651To register a threshold application need:
 652- create an eventfd using eventfd(2);
 653- open memory.usage_in_bytes or memory.memsw.usage_in_bytes;
 654- write string like "<event_fd> <fd of memory.usage_in_bytes> <threshold>" to
 655  cgroup.event_control.
 657Application will be notified through eventfd when memory usage crosses
 658threshold in any direction.
 660It's applicable for root and non-root cgroup.
 66210. OOM Control
 664memory.oom_control file is for OOM notification and other controls.
 666Memory cgroup implements OOM notifier using cgroup notification
 667API (See cgroups.txt). It allows to register multiple OOM notification
 668delivery and gets notification when OOM happens.
 670To register a notifier, application need:
 671 - create an eventfd using eventfd(2)
 672 - open memory.oom_control file
 673 - write string like "<event_fd> <fd of memory.oom_control>" to
 674   cgroup.event_control
 676Application will be notified through eventfd when OOM happens.
 677OOM notification doesn't work for root cgroup.
 679You can disable OOM-killer by writing "1" to memory.oom_control file, as:
 681        #echo 1 > memory.oom_control
 683This operation is only allowed to the top cgroup of sub-hierarchy.
 684If OOM-killer is disabled, tasks under cgroup will hang/sleep
 685in memory cgroup's OOM-waitqueue when they request accountable memory.
 687For running them, you have to relax the memory cgroup's OOM status by
 688        * enlarge limit or reduce usage.
 689To reduce usage,
 690        * kill some tasks.
 691        * move some tasks to other group with account migration.
 692        * remove some files (on tmpfs?)
 694Then, stopped tasks will work again.
 696At reading, current status of OOM is shown.
 697        oom_kill_disable 0 or 1 (if 1, oom-killer is disabled)
 698        under_oom        0 or 1 (if 1, the memory cgroup is under OOM, tasks may
 699                                 be stopped.)
 70111. TODO
 7031. Add support for accounting huge pages (as a separate controller)
 7042. Make per-cgroup scanner reclaim not-shared pages first
 7053. Teach controller to account for shared-pages
 7064. Start reclamation in the background when the limit is
 707   not yet hit but the usage is getting closer
 711Overall, the memory controller has been a stable controller and has been
 712commented and discussed quite extensively in the community.
 7161. Singh, Balbir. RFC: Memory Controller,
 7172. Singh, Balbir. Memory Controller (RSS Control),
 7193. Emelianov, Pavel. Resource controllers based on process cgroups
 7214. Emelianov, Pavel. RSS controller based on process cgroups (v2)
 7235. Emelianov, Pavel. RSS controller based on process cgroups (v3)
 7256. Menage, Paul. Control Groups v10,
 7267. Vaidyanathan, Srinivasan, Control Groups: Pagecache accounting and control
 727   subsystem (v3),
 7288. Singh, Balbir. RSS controller v2 test results (lmbench),
 7309. Singh, Balbir. RSS controller v2 AIM9 results
 73210. Singh, Balbir. Memory controller v6 test results,
 73411. Singh, Balbir. Memory controller introduction (v6),
 73612. Corbet, Jonathan, Controlling memory use in cgroups,
 738 kindly hosted by Redpill Linpro AS, provider of Linux consulting and operations services since 1995.