1 ==========================
2 Memory Resource Controller
3 ==========================
6 This document is hopelessly outdated and it asks for a complete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
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.
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.
22 Benefits and Purpose of the memory controller
23 =============================================
25 The memory controller isolates the memory behaviour of a group of tasks
26 from the rest of the system. The article on LWN [12]_ mentions some probable
27 uses of the memory controller. The memory controller can be used to
29 a. Isolate an application or a group of applications
30 Memory-hungry applications can be isolated and limited to a smaller
32 b. Create a cgroup with a limited amount of memory; this can be used
33 as a good alternative to booting with mem=XXXX.
34 c. Virtualization solutions can control the amount of memory they want
35 to assign to a virtual machine instance.
36 d. 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
39 e. There are several other use cases; find one or use the controller just
40 for fun (to learn and hack on the VM subsystem).
42 Current Status: linux-2.6.34-mmotm(development version of 2010/April)
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
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.
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>`)
61 Brief summary of control files.
63 ==================================== ==========================================
64 tasks attach a task(thread) and show list of
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
71 memory.memsw.usage_in_bytes show current usage for memory+Swap
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 memory.stat show various statistics
82 memory.use_hierarchy set/show hierarchical account enabled
83 This knob is deprecated and shouldn't be
85 memory.force_empty trigger forced page reclaim
86 memory.pressure_level set memory pressure notifications
87 memory.swappiness set/show swappiness parameter of vmscan
88 (See sysctl's vm.swappiness)
89 memory.move_charge_at_immigrate set/show controls of moving charges
90 This knob is deprecated and shouldn't be
92 memory.oom_control set/show oom controls.
93 memory.numa_stat show the number of memory usage per numa
95 memory.kmem.limit_in_bytes This knob is deprecated and writing to
96 it will return -ENOTSUPP.
97 memory.kmem.usage_in_bytes show current kernel memory allocation
98 memory.kmem.failcnt show the number of kernel memory usage
100 memory.kmem.max_usage_in_bytes show max kernel memory usage recorded
102 memory.kmem.tcp.limit_in_bytes set/show hard limit for tcp buf memory
103 memory.kmem.tcp.usage_in_bytes show current tcp buf memory allocation
104 memory.kmem.tcp.failcnt show the number of tcp buf memory usage
106 memory.kmem.tcp.max_usage_in_bytes show max tcp buf memory usage recorded
107 ==================================== ==========================================
112 The memory controller has a long history. A request for comments for the memory
113 controller was posted by Balbir Singh [1]_. At the time the RFC was posted
114 there were several implementations for memory control. The goal of the
115 RFC was to build consensus and agreement for the minimal features required
116 for memory control. The first RSS controller was posted by Balbir Singh [2]_
117 in Feb 2007. Pavel Emelianov [3]_ [4]_ [5]_ has since posted three versions
118 of the RSS controller. At OLS, at the resource management BoF, everyone
119 suggested that we handle both page cache and RSS together. Another request was
120 raised to allow user space handling of OOM. The current memory controller is
121 at version 6; it combines both mapped (RSS) and unmapped Page
127 Memory is a unique resource in the sense that it is present in a limited
128 amount. If a task requires a lot of CPU processing, the task can spread
129 its processing over a period of hours, days, months or years, but with
130 memory, the same physical memory needs to be reused to accomplish the task.
132 The memory controller implementation has been divided into phases. These
136 2. mlock(2) controller
137 3. Kernel user memory accounting and slab control
138 4. user mappings length controller
140 The memory controller is the first controller developed.
145 The core of the design is a counter called the page_counter. The
146 page_counter tracks the current memory usage and limit of the group of
147 processes associated with the controller. Each cgroup has a memory controller
148 specific data structure (mem_cgroup) associated with it.
154 :caption: Figure 1: Hierarchy of Accounting
156 +--------------------+
159 +--------------------+
162 +---------------+ | +---------------+
163 | mm_struct | |.... | mm_struct |
165 +---------------+ | +---------------+
169 +---------------+ +------+--------+
170 | page +----------> page_cgroup|
172 +---------------+ +---------------+
176 Figure 1 shows the important aspects of the controller
178 1. Accounting happens per cgroup
179 2. Each mm_struct knows about which cgroup it belongs to
180 3. Each page has a pointer to the page_cgroup, which in turn knows the
183 The accounting is done as follows: mem_cgroup_charge_common() is invoked to
184 set up the necessary data structures and check if the cgroup that is being
185 charged is over its limit. If it is, then reclaim is invoked on the cgroup.
186 More details can be found in the reclaim section of this document.
187 If everything goes well, a page meta-data-structure called page_cgroup is
188 updated. page_cgroup has its own LRU on cgroup.
189 (*) page_cgroup structure is allocated at boot/memory-hotplug time.
191 2.2.1 Accounting details
192 ------------------------
194 All mapped anon pages (RSS) and cache pages (Page Cache) are accounted.
195 Some pages which are never reclaimable and will not be on the LRU
196 are not accounted. We just account pages under usual VM management.
198 RSS pages are accounted at page_fault unless they've already been accounted
199 for earlier. A file page will be accounted for as Page Cache when it's
200 inserted into inode (radix-tree). While it's mapped into the page tables of
201 processes, duplicate accounting is carefully avoided.
203 An RSS page is unaccounted when it's fully unmapped. A PageCache page is
204 unaccounted when it's removed from radix-tree. Even if RSS pages are fully
205 unmapped (by kswapd), they may exist as SwapCache in the system until they
206 are really freed. Such SwapCaches are also accounted.
207 A swapped-in page is accounted after adding into swapcache.
209 Note: The kernel does swapin-readahead and reads multiple swaps at once.
210 Since page's memcg recorded into swap whatever memsw enabled, the page will
211 be accounted after swapin.
213 At page migration, accounting information is kept.
215 Note: we just account pages-on-LRU because our purpose is to control amount
216 of used pages; not-on-LRU pages tend to be out-of-control from VM view.
218 2.3 Shared Page Accounting
219 --------------------------
221 Shared pages are accounted on the basis of the first touch approach. The
222 cgroup that first touches a page is accounted for the page. The principle
223 behind this approach is that a cgroup that aggressively uses a shared
224 page will eventually get charged for it (once it is uncharged from
225 the cgroup that brought it in -- this will happen on memory pressure).
227 But see :ref:`section 8.2 <cgroup-v1-memory-movable-charges>` when moving a
228 task to another cgroup, its pages may be recharged to the new cgroup, if
229 move_charge_at_immigrate has been chosen.
232 --------------------------------------
234 Swap usage is always recorded for each of cgroup. Swap Extension allows you to
237 When CONFIG_SWAP is enabled, following files are added.
239 - memory.memsw.usage_in_bytes.
240 - memory.memsw.limit_in_bytes.
242 memsw means memory+swap. Usage of memory+swap is limited by
243 memsw.limit_in_bytes.
245 Example: Assume a system with 4G of swap. A task which allocates 6G of memory
246 (by mistake) under 2G memory limitation will use all swap.
247 In this case, setting memsw.limit_in_bytes=3G will prevent bad use of swap.
248 By using the memsw limit, you can avoid system OOM which can be caused by swap
251 2.4.1 why 'memory+swap' rather than swap
252 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
254 The global LRU(kswapd) can swap out arbitrary pages. Swap-out means
255 to move account from memory to swap...there is no change in usage of
256 memory+swap. In other words, when we want to limit the usage of swap without
257 affecting global LRU, memory+swap limit is better than just limiting swap from
260 2.4.2. What happens when a cgroup hits memory.memsw.limit_in_bytes
261 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
263 When a cgroup hits memory.memsw.limit_in_bytes, it's useless to do swap-out
264 in this cgroup. Then, swap-out will not be done by cgroup routine and file
265 caches are dropped. But as mentioned above, global LRU can do swapout memory
266 from it for sanity of the system's memory management state. You can't forbid
272 Each cgroup maintains a per cgroup LRU which has the same structure as
273 global VM. When a cgroup goes over its limit, we first try
274 to reclaim memory from the cgroup so as to make space for the new
275 pages that the cgroup has touched. If the reclaim is unsuccessful,
276 an OOM routine is invoked to select and kill the bulkiest task in the
277 cgroup. (See :ref:`10. OOM Control <cgroup-v1-memory-oom-control>` below.)
279 The reclaim algorithm has not been modified for cgroups, except that
280 pages that are selected for reclaiming come from the per-cgroup LRU
284 Reclaim does not work for the root cgroup, since we cannot set any
285 limits on the root cgroup.
288 When panic_on_oom is set to "2", the whole system will panic.
290 When oom event notifier is registered, event will be delivered.
291 (See :ref:`oom_control <cgroup-v1-memory-oom-control>` section)
296 Lock order is as follows::
298 Page lock (PG_locked bit of page->flags)
299 mm->page_table_lock or split pte_lock
300 folio_memcg_lock (memcg->move_lock)
301 mapping->i_pages lock
304 Per-node-per-memcgroup LRU (cgroup's private LRU) is guarded by
305 lruvec->lru_lock; PG_lru bit of page->flags is cleared before
306 isolating a page from its LRU under lruvec->lru_lock.
308 .. _cgroup-v1-memory-kernel-extension:
310 2.7 Kernel Memory Extension
311 -----------------------------------------------
313 With the Kernel memory extension, the Memory Controller is able to limit
314 the amount of kernel memory used by the system. Kernel memory is fundamentally
315 different than user memory, since it can't be swapped out, which makes it
316 possible to DoS the system by consuming too much of this precious resource.
318 Kernel memory accounting is enabled for all memory cgroups by default. But
319 it can be disabled system-wide by passing cgroup.memory=nokmem to the kernel
320 at boot time. In this case, kernel memory will not be accounted at all.
322 Kernel memory limits are not imposed for the root cgroup. Usage for the root
323 cgroup may or may not be accounted. The memory used is accumulated into
324 memory.kmem.usage_in_bytes, or in a separate counter when it makes sense.
325 (currently only for tcp).
327 The main "kmem" counter is fed into the main counter, so kmem charges will
328 also be visible from the user counter.
330 Currently no soft limit is implemented for kernel memory. It is future work
331 to trigger slab reclaim when those limits are reached.
333 2.7.1 Current Kernel Memory resources accounted
334 -----------------------------------------------
337 every process consumes some stack pages. By accounting into
338 kernel memory, we prevent new processes from being created when the kernel
339 memory usage is too high.
342 pages allocated by the SLAB or SLUB allocator are tracked. A copy
343 of each kmem_cache is created every time the cache is touched by the first time
344 from inside the memcg. The creation is done lazily, so some objects can still be
345 skipped while the cache is being created. All objects in a slab page should
346 belong to the same memcg. This only fails to hold when a task is migrated to a
347 different memcg during the page allocation by the cache.
349 sockets memory pressure:
350 some sockets protocols have memory pressure
351 thresholds. The Memory Controller allows them to be controlled individually
352 per cgroup, instead of globally.
355 sockets memory pressure for the tcp protocol.
357 2.7.2 Common use cases
358 ----------------------
360 Because the "kmem" counter is fed to the main user counter, kernel memory can
361 never be limited completely independently of user memory. Say "U" is the user
362 limit, and "K" the kernel limit. There are three possible ways limits can be
365 U != 0, K = unlimited:
366 This is the standard memcg limitation mechanism already present before kmem
367 accounting. Kernel memory is completely ignored.
370 Kernel memory is a subset of the user memory. This setup is useful in
371 deployments where the total amount of memory per-cgroup is overcommitted.
372 Overcommitting kernel memory limits is definitely not recommended, since the
373 box can still run out of non-reclaimable memory.
374 In this case, the admin could set up K so that the sum of all groups is
375 never greater than the total memory, and freely set U at the cost of his
379 In the current implementation, memory reclaim will NOT be triggered for
380 a cgroup when it hits K while staying below U, which makes this setup
384 Since kmem charges will also be fed to the user counter and reclaim will be
385 triggered for the cgroup for both kinds of memory. This setup gives the
386 admin a unified view of memory, and it is also useful for people who just
387 want to track kernel memory usage.
392 To use the user interface:
394 1. Enable CONFIG_CGROUPS and CONFIG_MEMCG options
395 2. Prepare the cgroups (see :ref:`Why are cgroups needed?
396 <cgroups-why-needed>` for the background information)::
398 # mount -t tmpfs none /sys/fs/cgroup
399 # mkdir /sys/fs/cgroup/memory
400 # mount -t cgroup none /sys/fs/cgroup/memory -o memory
402 3. Make the new group and move bash into it::
404 # mkdir /sys/fs/cgroup/memory/0
405 # echo $$ > /sys/fs/cgroup/memory/0/tasks
407 4. Since now we're in the 0 cgroup, we can alter the memory limit::
409 # echo 4M > /sys/fs/cgroup/memory/0/memory.limit_in_bytes
411 The limit can now be queried::
413 # cat /sys/fs/cgroup/memory/0/memory.limit_in_bytes
417 We can use a suffix (k, K, m, M, g or G) to indicate values in kilo,
418 mega or gigabytes. (Here, Kilo, Mega, Giga are Kibibytes, Mebibytes,
422 We can write "-1" to reset the ``*.limit_in_bytes(unlimited)``.
425 We cannot set limits on the root cgroup any more.
428 We can check the usage::
430 # cat /sys/fs/cgroup/memory/0/memory.usage_in_bytes
433 A successful write to this file does not guarantee a successful setting of
434 this limit to the value written into the file. This can be due to a
435 number of factors, such as rounding up to page boundaries or the total
436 availability of memory on the system. The user is required to re-read
437 this file after a write to guarantee the value committed by the kernel::
439 # echo 1 > memory.limit_in_bytes
440 # cat memory.limit_in_bytes
443 The memory.failcnt field gives the number of times that the cgroup limit was
446 The memory.stat file gives accounting information. Now, the number of
447 caches, RSS and Active pages/Inactive pages are shown.
452 For testing features and implementation, see memcg_test.txt.
454 Performance test is also important. To see pure memory controller's overhead,
455 testing on tmpfs will give you good numbers of small overheads.
456 Example: do kernel make on tmpfs.
458 Page-fault scalability is also important. At measuring parallel
459 page fault test, multi-process test may be better than multi-thread
460 test because it has noise of shared objects/status.
462 But the above two are testing extreme situations.
463 Trying usual test under memory controller is always helpful.
465 .. _cgroup-v1-memory-test-troubleshoot:
470 Sometimes a user might find that the application under a cgroup is
471 terminated by the OOM killer. There are several causes for this:
473 1. The cgroup limit is too low (just too low to do anything useful)
474 2. The user is using anonymous memory and swap is turned off or too low
476 A sync followed by echo 1 > /proc/sys/vm/drop_caches will help get rid of
477 some of the pages cached in the cgroup (page cache pages).
479 To know what happens, disabling OOM_Kill as per :ref:`"10. OOM Control"
480 <cgroup-v1-memory-oom-control>` (below) and seeing what happens will be
483 .. _cgroup-v1-memory-test-task-migration:
488 When a task migrates from one cgroup to another, its charge is not
489 carried forward by default. The pages allocated from the original cgroup still
490 remain charged to it, the charge is dropped when the page is freed or
493 You can move charges of a task along with task migration.
494 See :ref:`8. "Move charges at task migration" <cgroup-v1-memory-move-charges>`
496 4.3 Removing a cgroup
497 ---------------------
499 A cgroup can be removed by rmdir, but as discussed in :ref:`sections 4.1
500 <cgroup-v1-memory-test-troubleshoot>` and :ref:`4.2
501 <cgroup-v1-memory-test-task-migration>`, a cgroup might have some charge
502 associated with it, even though all tasks have migrated away from it. (because
503 we charge against pages, not against tasks.)
505 We move the stats to parent, and no change on the charge except uncharging
508 Charges recorded in swap information is not updated at removal of cgroup.
509 Recorded information is discarded and a cgroup which uses swap (swapcache)
510 will be charged as a new owner of it.
517 memory.force_empty interface is provided to make cgroup's memory usage empty.
518 When writing anything to this::
520 # echo 0 > memory.force_empty
522 the cgroup will be reclaimed and as many pages reclaimed as possible.
524 The typical use case for this interface is before calling rmdir().
525 Though rmdir() offlines memcg, but the memcg may still stay there due to
526 charged file caches. Some out-of-use page caches may keep charged until
527 memory pressure happens. If you want to avoid that, force_empty will be useful.
532 memory.stat file includes following statistics:
534 * per-memory cgroup local status
536 =============== ===============================================================
537 cache # of bytes of page cache memory.
538 rss # of bytes of anonymous and swap cache memory (includes
539 transparent hugepages).
540 rss_huge # of bytes of anonymous transparent hugepages.
541 mapped_file # of bytes of mapped file (includes tmpfs/shmem)
542 pgpgin # of charging events to the memory cgroup. The charging
543 event happens each time a page is accounted as either mapped
544 anon page(RSS) or cache page(Page Cache) to the cgroup.
545 pgpgout # of uncharging events to the memory cgroup. The uncharging
546 event happens each time a page is unaccounted from the
548 swap # of bytes of swap usage
549 dirty # of bytes that are waiting to get written back to the disk.
550 writeback # of bytes of file/anon cache that are queued for syncing to
552 inactive_anon # of bytes of anonymous and swap cache memory on inactive
554 active_anon # of bytes of anonymous and swap cache memory on active
556 inactive_file # of bytes of file-backed memory and MADV_FREE anonymous
557 memory (LazyFree pages) on inactive LRU list.
558 active_file # of bytes of file-backed memory on active LRU list.
559 unevictable # of bytes of memory that cannot be reclaimed (mlocked etc).
560 =============== ===============================================================
562 * status considering hierarchy (see memory.use_hierarchy settings):
564 ========================= ===================================================
565 hierarchical_memory_limit # of bytes of memory limit with regard to
567 under which the memory cgroup is
568 hierarchical_memsw_limit # of bytes of memory+swap limit with regard to
569 hierarchy under which memory cgroup is.
571 total_<counter> # hierarchical version of <counter>, which in
572 addition to the cgroup's own value includes the
573 sum of all hierarchical children's values of
574 <counter>, i.e. total_cache
575 ========================= ===================================================
577 * additional vm parameters (depends on CONFIG_DEBUG_VM):
579 ========================= ========================================
580 recent_rotated_anon VM internal parameter. (see mm/vmscan.c)
581 recent_rotated_file VM internal parameter. (see mm/vmscan.c)
582 recent_scanned_anon VM internal parameter. (see mm/vmscan.c)
583 recent_scanned_file VM internal parameter. (see mm/vmscan.c)
584 ========================= ========================================
587 recent_rotated means recent frequency of LRU rotation.
588 recent_scanned means recent # of scans to LRU.
589 showing for better debug please see the code for meanings.
592 Only anonymous and swap cache memory is listed as part of 'rss' stat.
593 This should not be confused with the true 'resident set size' or the
594 amount of physical memory used by the cgroup.
596 'rss + mapped_file" will give you resident set size of cgroup.
598 (Note: file and shmem may be shared among other cgroups. In that case,
599 mapped_file is accounted only when the memory cgroup is owner of page
605 Overrides /proc/sys/vm/swappiness for the particular group. The tunable
606 in the root cgroup corresponds to the global swappiness setting.
608 Please note that unlike during the global reclaim, limit reclaim
609 enforces that 0 swappiness really prevents from any swapping even if
610 there is a swap storage available. This might lead to memcg OOM killer
611 if there are no file pages to reclaim.
616 A memory cgroup provides memory.failcnt and memory.memsw.failcnt files.
617 This failcnt(== failure count) shows the number of times that a usage counter
618 hit its limit. When a memory cgroup hits a limit, failcnt increases and
619 memory under it will be reclaimed.
621 You can reset failcnt by writing 0 to failcnt file::
623 # echo 0 > .../memory.failcnt
628 For efficiency, as other kernel components, memory cgroup uses some optimization
629 to avoid unnecessary cacheline false sharing. usage_in_bytes is affected by the
630 method and doesn't show 'exact' value of memory (and swap) usage, it's a fuzz
631 value for efficient access. (Of course, when necessary, it's synchronized.)
632 If you want to know more exact memory usage, you should use RSS+CACHE(+SWAP)
633 value in memory.stat(see 5.2).
638 This is similar to numa_maps but operates on a per-memcg basis. This is
639 useful for providing visibility into the numa locality information within
640 an memcg since the pages are allowed to be allocated from any physical
641 node. One of the use cases is evaluating application performance by
642 combining this information with the application's CPU allocation.
644 Each memcg's numa_stat file includes "total", "file", "anon" and "unevictable"
645 per-node page counts including "hierarchical_<counter>" which sums up all
646 hierarchical children's values in addition to the memcg's own value.
648 The output format of memory.numa_stat is::
650 total=<total pages> N0=<node 0 pages> N1=<node 1 pages> ...
651 file=<total file pages> N0=<node 0 pages> N1=<node 1 pages> ...
652 anon=<total anon pages> N0=<node 0 pages> N1=<node 1 pages> ...
653 unevictable=<total anon pages> N0=<node 0 pages> N1=<node 1 pages> ...
654 hierarchical_<counter>=<counter pages> N0=<node 0 pages> N1=<node 1 pages> ...
656 The "total" count is sum of file + anon + unevictable.
661 The memory controller supports a deep hierarchy and hierarchical accounting.
662 The hierarchy is created by creating the appropriate cgroups in the
663 cgroup filesystem. Consider for example, the following cgroup filesystem
674 In the diagram above, with hierarchical accounting enabled, all memory
675 usage of e, is accounted to its ancestors up until the root (i.e, c and root).
676 If one of the ancestors goes over its limit, the reclaim algorithm reclaims
677 from the tasks in the ancestor and the children of the ancestor.
679 6.1 Hierarchical accounting and reclaim
680 ---------------------------------------
682 Hierarchical accounting is enabled by default. Disabling the hierarchical
683 accounting is deprecated. An attempt to do it will result in a failure
684 and a warning printed to dmesg.
686 For compatibility reasons writing 1 to memory.use_hierarchy will always pass::
688 # echo 1 > memory.use_hierarchy
693 Soft limits allow for greater sharing of memory. The idea behind soft limits
694 is to allow control groups to use as much of the memory as needed, provided
696 a. There is no memory contention
697 b. They do not exceed their hard limit
699 When the system detects memory contention or low memory, control groups
700 are pushed back to their soft limits. If the soft limit of each control
701 group is very high, they are pushed back as much as possible to make
702 sure that one control group does not starve the others of memory.
704 Please note that soft limits is a best-effort feature; it comes with
705 no guarantees, but it does its best to make sure that when memory is
706 heavily contended for, memory is allocated based on the soft limit
707 hints/setup. Currently soft limit based reclaim is set up such that
708 it gets invoked from balance_pgdat (kswapd).
713 Soft limits can be setup by using the following commands (in this example we
714 assume a soft limit of 256 MiB)::
716 # echo 256M > memory.soft_limit_in_bytes
718 If we want to change this to 1G, we can at any time use::
720 # echo 1G > memory.soft_limit_in_bytes
723 Soft limits take effect over a long period of time, since they involve
724 reclaiming memory for balancing between memory cgroups
727 It is recommended to set the soft limit always below the hard limit,
728 otherwise the hard limit will take precedence.
730 .. _cgroup-v1-memory-move-charges:
732 8. Move charges at task migration (DEPRECATED!)
733 ===============================================
737 It's expensive and unreliable! It's better practice to launch workload
738 tasks directly from inside their target cgroup. Use dedicated workload
739 cgroups to allow fine-grained policy adjustments without having to
740 move physical pages between control domains.
742 Users can move charges associated with a task along with task migration, that
743 is, uncharge task's pages from the old cgroup and charge them to the new cgroup.
744 This feature is not supported in !CONFIG_MMU environments because of lack of
750 This feature is disabled by default. It can be enabled (and disabled again) by
751 writing to memory.move_charge_at_immigrate of the destination cgroup.
753 If you want to enable it::
755 # echo (some positive value) > memory.move_charge_at_immigrate
758 Each bits of move_charge_at_immigrate has its own meaning about what type
759 of charges should be moved. See :ref:`section 8.2
760 <cgroup-v1-memory-movable-charges>` for details.
763 Charges are moved only when you move mm->owner, in other words,
764 a leader of a thread group.
767 If we cannot find enough space for the task in the destination cgroup, we
768 try to make space by reclaiming memory. Task migration may fail if we
769 cannot make enough space.
772 It can take several seconds if you move charges much.
774 And if you want disable it again::
776 # echo 0 > memory.move_charge_at_immigrate
778 .. _cgroup-v1-memory-movable-charges:
780 8.2 Type of charges which can be moved
781 --------------------------------------
783 Each bit in move_charge_at_immigrate has its own meaning about what type of
784 charges should be moved. But in any case, it must be noted that an account of
785 a page or a swap can be moved only when it is charged to the task's current
788 +---+--------------------------------------------------------------------------+
789 |bit| what type of charges would be moved ? |
790 +===+==========================================================================+
791 | 0 | A charge of an anonymous page (or swap of it) used by the target task. |
792 | | You must enable Swap Extension (see 2.4) to enable move of swap charges. |
793 +---+--------------------------------------------------------------------------+
794 | 1 | A charge of file pages (normal file, tmpfs file (e.g. ipc shared memory) |
795 | | and swaps of tmpfs file) mmapped by the target task. Unlike the case of |
796 | | anonymous pages, file pages (and swaps) in the range mmapped by the task |
797 | | will be moved even if the task hasn't done page fault, i.e. they might |
798 | | not be the task's "RSS", but other task's "RSS" that maps the same file. |
799 | | And mapcount of the page is ignored (the page can be moved even if |
800 | | page_mapcount(page) > 1). You must enable Swap Extension (see 2.4) to |
801 | | enable move of swap charges. |
802 +---+--------------------------------------------------------------------------+
807 - All of moving charge operations are done under cgroup_mutex. It's not good
808 behavior to hold the mutex too long, so we may need some trick.
813 Memory cgroup implements memory thresholds using the cgroups notification
814 API (see cgroups.txt). It allows to register multiple memory and memsw
815 thresholds and gets notifications when it crosses.
817 To register a threshold, an application must:
819 - create an eventfd using eventfd(2);
820 - open memory.usage_in_bytes or memory.memsw.usage_in_bytes;
821 - write string like "<event_fd> <fd of memory.usage_in_bytes> <threshold>" to
822 cgroup.event_control.
824 Application will be notified through eventfd when memory usage crosses
825 threshold in any direction.
827 It's applicable for root and non-root cgroup.
829 .. _cgroup-v1-memory-oom-control:
834 memory.oom_control file is for OOM notification and other controls.
836 Memory cgroup implements OOM notifier using the cgroup notification
837 API (See cgroups.txt). It allows to register multiple OOM notification
838 delivery and gets notification when OOM happens.
840 To register a notifier, an application must:
842 - create an eventfd using eventfd(2)
843 - open memory.oom_control file
844 - write string like "<event_fd> <fd of memory.oom_control>" to
847 The application will be notified through eventfd when OOM happens.
848 OOM notification doesn't work for the root cgroup.
850 You can disable the OOM-killer by writing "1" to memory.oom_control file, as:
852 #echo 1 > memory.oom_control
854 If OOM-killer is disabled, tasks under cgroup will hang/sleep
855 in memory cgroup's OOM-waitqueue when they request accountable memory.
857 For running them, you have to relax the memory cgroup's OOM status by
859 * enlarge limit or reduce usage.
864 * move some tasks to other group with account migration.
865 * remove some files (on tmpfs?)
867 Then, stopped tasks will work again.
869 At reading, current status of OOM is shown.
871 - oom_kill_disable 0 or 1
872 (if 1, oom-killer is disabled)
874 (if 1, the memory cgroup is under OOM, tasks may be stopped.)
875 - oom_kill integer counter
876 The number of processes belonging to this cgroup killed by any
882 The pressure level notifications can be used to monitor the memory
883 allocation cost; based on the pressure, applications can implement
884 different strategies of managing their memory resources. The pressure
885 levels are defined as following:
887 The "low" level means that the system is reclaiming memory for new
888 allocations. Monitoring this reclaiming activity might be useful for
889 maintaining cache level. Upon notification, the program (typically
890 "Activity Manager") might analyze vmstat and act in advance (i.e.
891 prematurely shutdown unimportant services).
893 The "medium" level means that the system is experiencing medium memory
894 pressure, the system might be making swap, paging out active file caches,
895 etc. Upon this event applications may decide to further analyze
896 vmstat/zoneinfo/memcg or internal memory usage statistics and free any
897 resources that can be easily reconstructed or re-read from a disk.
899 The "critical" level means that the system is actively thrashing, it is
900 about to out of memory (OOM) or even the in-kernel OOM killer is on its
901 way to trigger. Applications should do whatever they can to help the
902 system. It might be too late to consult with vmstat or any other
903 statistics, so it's advisable to take an immediate action.
905 By default, events are propagated upward until the event is handled, i.e. the
906 events are not pass-through. For example, you have three cgroups: A->B->C. Now
907 you set up an event listener on cgroups A, B and C, and suppose group C
908 experiences some pressure. In this situation, only group C will receive the
909 notification, i.e. groups A and B will not receive it. This is done to avoid
910 excessive "broadcasting" of messages, which disturbs the system and which is
911 especially bad if we are low on memory or thrashing. Group B, will receive
912 notification only if there are no event listers for group C.
914 There are three optional modes that specify different propagation behavior:
916 - "default": this is the default behavior specified above. This mode is the
917 same as omitting the optional mode parameter, preserved by backwards
920 - "hierarchy": events always propagate up to the root, similar to the default
921 behavior, except that propagation continues regardless of whether there are
922 event listeners at each level, with the "hierarchy" mode. In the above
923 example, groups A, B, and C will receive notification of memory pressure.
925 - "local": events are pass-through, i.e. they only receive notifications when
926 memory pressure is experienced in the memcg for which the notification is
927 registered. In the above example, group C will receive notification if
928 registered for "local" notification and the group experiences memory
929 pressure. However, group B will never receive notification, regardless if
930 there is an event listener for group C or not, if group B is registered for
933 The level and event notification mode ("hierarchy" or "local", if necessary) are
934 specified by a comma-delimited string, i.e. "low,hierarchy" specifies
935 hierarchical, pass-through, notification for all ancestor memcgs. Notification
936 that is the default, non pass-through behavior, does not specify a mode.
937 "medium,local" specifies pass-through notification for the medium level.
939 The file memory.pressure_level is only used to setup an eventfd. To
940 register a notification, an application must:
942 - create an eventfd using eventfd(2);
943 - open memory.pressure_level;
944 - write string as "<event_fd> <fd of memory.pressure_level> <level[,mode]>"
945 to cgroup.event_control.
947 Application will be notified through eventfd when memory pressure is at
948 the specific level (or higher). Read/write operations to
949 memory.pressure_level are no implemented.
953 Here is a small script example that makes a new cgroup, sets up a
954 memory limit, sets up a notification in the cgroup and then makes child
955 cgroup experience a critical pressure::
957 # cd /sys/fs/cgroup/memory/
960 # cgroup_event_listener memory.pressure_level low,hierarchy &
961 # echo 8000000 > memory.limit_in_bytes
962 # echo 8000000 > memory.memsw.limit_in_bytes
964 # dd if=/dev/zero | read x
966 (Expect a bunch of notifications, and eventually, the oom-killer will
972 1. Make per-cgroup scanner reclaim not-shared pages first
973 2. Teach controller to account for shared-pages
974 3. Start reclamation in the background when the limit is
975 not yet hit but the usage is getting closer
980 Overall, the memory controller has been a stable controller and has been
981 commented and discussed quite extensively in the community.
986 .. [1] Singh, Balbir. RFC: Memory Controller, http://lwn.net/Articles/206697/
987 .. [2] Singh, Balbir. Memory Controller (RSS Control),
988 http://lwn.net/Articles/222762/
989 .. [3] Emelianov, Pavel. Resource controllers based on process cgroups
990 https://lore.kernel.org/r/45ED7DEC.7010403@sw.ru
991 .. [4] Emelianov, Pavel. RSS controller based on process cgroups (v2)
992 https://lore.kernel.org/r/461A3010.90403@sw.ru
993 .. [5] Emelianov, Pavel. RSS controller based on process cgroups (v3)
994 https://lore.kernel.org/r/465D9739.8070209@openvz.org
996 6. Menage, Paul. Control Groups v10, http://lwn.net/Articles/236032/
997 7. Vaidyanathan, Srinivasan, Control Groups: Pagecache accounting and control
998 subsystem (v3), http://lwn.net/Articles/235534/
999 8. Singh, Balbir. RSS controller v2 test results (lmbench),
1000 https://lore.kernel.org/r/464C95D4.7070806@linux.vnet.ibm.com
1001 9. Singh, Balbir. RSS controller v2 AIM9 results
1002 https://lore.kernel.org/r/464D267A.50107@linux.vnet.ibm.com
1003 10. Singh, Balbir. Memory controller v6 test results,
1004 https://lore.kernel.org/r/20070819094658.654.84837.sendpatchset@balbir-laptop
1006 .. [11] Singh, Balbir. Memory controller introduction (v6),
1007 https://lore.kernel.org/r/20070817084228.26003.12568.sendpatchset@balbir-laptop
1008 .. [12] Corbet, Jonathan, Controlling memory use in cgroups,
1009 http://lwn.net/Articles/243795/