linux/Documentation/cgroups/cgroups.txt
<<
>>
Prefs
   1                                CGROUPS
   2                                -------
   3
   4Written by Paul Menage <menage@google.com> based on
   5Documentation/cgroups/cpusets.txt
   6
   7Original copyright statements from cpusets.txt:
   8Portions Copyright (C) 2004 BULL SA.
   9Portions Copyright (c) 2004-2006 Silicon Graphics, Inc.
  10Modified by Paul Jackson <pj@sgi.com>
  11Modified by Christoph Lameter <clameter@sgi.com>
  12
  13CONTENTS:
  14=========
  15
  161. Control Groups
  17  1.1 What are cgroups ?
  18  1.2 Why are cgroups needed ?
  19  1.3 How are cgroups implemented ?
  20  1.4 What does notify_on_release do ?
  21  1.5 What does clone_children do ?
  22  1.6 How do I use cgroups ?
  232. Usage Examples and Syntax
  24  2.1 Basic Usage
  25  2.2 Attaching processes
  26  2.3 Mounting hierarchies by name
  27  2.4 Notification API
  283. Kernel API
  29  3.1 Overview
  30  3.2 Synchronization
  31  3.3 Subsystem API
  324. Extended attributes usage
  335. Questions
  34
  351. Control Groups
  36=================
  37
  381.1 What are cgroups ?
  39----------------------
  40
  41Control Groups provide a mechanism for aggregating/partitioning sets of
  42tasks, and all their future children, into hierarchical groups with
  43specialized behaviour.
  44
  45Definitions:
  46
  47A *cgroup* associates a set of tasks with a set of parameters for one
  48or more subsystems.
  49
  50A *subsystem* is a module that makes use of the task grouping
  51facilities provided by cgroups to treat groups of tasks in
  52particular ways. A subsystem is typically a "resource controller" that
  53schedules a resource or applies per-cgroup limits, but it may be
  54anything that wants to act on a group of processes, e.g. a
  55virtualization subsystem.
  56
  57A *hierarchy* is a set of cgroups arranged in a tree, such that
  58every task in the system is in exactly one of the cgroups in the
  59hierarchy, and a set of subsystems; each subsystem has system-specific
  60state attached to each cgroup in the hierarchy.  Each hierarchy has
  61an instance of the cgroup virtual filesystem associated with it.
  62
  63At any one time there may be multiple active hierarchies of task
  64cgroups. Each hierarchy is a partition of all tasks in the system.
  65
  66User-level code may create and destroy cgroups by name in an
  67instance of the cgroup virtual file system, specify and query to
  68which cgroup a task is assigned, and list the task PIDs assigned to
  69a cgroup. Those creations and assignments only affect the hierarchy
  70associated with that instance of the cgroup file system.
  71
  72On their own, the only use for cgroups is for simple job
  73tracking. The intention is that other subsystems hook into the generic
  74cgroup support to provide new attributes for cgroups, such as
  75accounting/limiting the resources which processes in a cgroup can
  76access. For example, cpusets (see Documentation/cgroups/cpusets.txt) allow
  77you to associate a set of CPUs and a set of memory nodes with the
  78tasks in each cgroup.
  79
  801.2 Why are cgroups needed ?
  81----------------------------
  82
  83There are multiple efforts to provide process aggregations in the
  84Linux kernel, mainly for resource-tracking purposes. Such efforts
  85include cpusets, CKRM/ResGroups, UserBeanCounters, and virtual server
  86namespaces. These all require the basic notion of a
  87grouping/partitioning of processes, with newly forked processes ending
  88up in the same group (cgroup) as their parent process.
  89
  90The kernel cgroup patch provides the minimum essential kernel
  91mechanisms required to efficiently implement such groups. It has
  92minimal impact on the system fast paths, and provides hooks for
  93specific subsystems such as cpusets to provide additional behaviour as
  94desired.
  95
  96Multiple hierarchy support is provided to allow for situations where
  97the division of tasks into cgroups is distinctly different for
  98different subsystems - having parallel hierarchies allows each
  99hierarchy to be a natural division of tasks, without having to handle
 100complex combinations of tasks that would be present if several
 101unrelated subsystems needed to be forced into the same tree of
 102cgroups.
 103
 104At one extreme, each resource controller or subsystem could be in a
 105separate hierarchy; at the other extreme, all subsystems
 106would be attached to the same hierarchy.
 107
 108As an example of a scenario (originally proposed by vatsa@in.ibm.com)
 109that can benefit from multiple hierarchies, consider a large
 110university server with various users - students, professors, system
 111tasks etc. The resource planning for this server could be along the
 112following lines:
 113
 114       CPU :          "Top cpuset"
 115                       /       \
 116               CPUSet1         CPUSet2
 117                  |               |
 118               (Professors)    (Students)
 119
 120               In addition (system tasks) are attached to topcpuset (so
 121               that they can run anywhere) with a limit of 20%
 122
 123       Memory : Professors (50%), Students (30%), system (20%)
 124
 125       Disk : Professors (50%), Students (30%), system (20%)
 126
 127       Network : WWW browsing (20%), Network File System (60%), others (20%)
 128                               / \
 129               Professors (15%)  students (5%)
 130
 131Browsers like Firefox/Lynx go into the WWW network class, while (k)nfsd goes
 132into the NFS network class.
 133
 134At the same time Firefox/Lynx will share an appropriate CPU/Memory class
 135depending on who launched it (prof/student).
 136
 137With the ability to classify tasks differently for different resources
 138(by putting those resource subsystems in different hierarchies),
 139the admin can easily set up a script which receives exec notifications
 140and depending on who is launching the browser he can
 141
 142    # echo browser_pid > /sys/fs/cgroup/<restype>/<userclass>/tasks
 143
 144With only a single hierarchy, he now would potentially have to create
 145a separate cgroup for every browser launched and associate it with
 146appropriate network and other resource class.  This may lead to
 147proliferation of such cgroups.
 148
 149Also let's say that the administrator would like to give enhanced network
 150access temporarily to a student's browser (since it is night and the user
 151wants to do online gaming :))  OR give one of the student's simulation
 152apps enhanced CPU power.
 153
 154With ability to write PIDs directly to resource classes, it's just a
 155matter of:
 156
 157       # echo pid > /sys/fs/cgroup/network/<new_class>/tasks
 158       (after some time)
 159       # echo pid > /sys/fs/cgroup/network/<orig_class>/tasks
 160
 161Without this ability, the administrator would have to split the cgroup into
 162multiple separate ones and then associate the new cgroups with the
 163new resource classes.
 164
 165
 166
 1671.3 How are cgroups implemented ?
 168---------------------------------
 169
 170Control Groups extends the kernel as follows:
 171
 172 - Each task in the system has a reference-counted pointer to a
 173   css_set.
 174
 175 - A css_set contains a set of reference-counted pointers to
 176   cgroup_subsys_state objects, one for each cgroup subsystem
 177   registered in the system. There is no direct link from a task to
 178   the cgroup of which it's a member in each hierarchy, but this
 179   can be determined by following pointers through the
 180   cgroup_subsys_state objects. This is because accessing the
 181   subsystem state is something that's expected to happen frequently
 182   and in performance-critical code, whereas operations that require a
 183   task's actual cgroup assignments (in particular, moving between
 184   cgroups) are less common. A linked list runs through the cg_list
 185   field of each task_struct using the css_set, anchored at
 186   css_set->tasks.
 187
 188 - A cgroup hierarchy filesystem can be mounted for browsing and
 189   manipulation from user space.
 190
 191 - You can list all the tasks (by PID) attached to any cgroup.
 192
 193The implementation of cgroups requires a few, simple hooks
 194into the rest of the kernel, none in performance-critical paths:
 195
 196 - in init/main.c, to initialize the root cgroups and initial
 197   css_set at system boot.
 198
 199 - in fork and exit, to attach and detach a task from its css_set.
 200
 201In addition, a new file system of type "cgroup" may be mounted, to
 202enable browsing and modifying the cgroups presently known to the
 203kernel.  When mounting a cgroup hierarchy, you may specify a
 204comma-separated list of subsystems to mount as the filesystem mount
 205options.  By default, mounting the cgroup filesystem attempts to
 206mount a hierarchy containing all registered subsystems.
 207
 208If an active hierarchy with exactly the same set of subsystems already
 209exists, it will be reused for the new mount. If no existing hierarchy
 210matches, and any of the requested subsystems are in use in an existing
 211hierarchy, the mount will fail with -EBUSY. Otherwise, a new hierarchy
 212is activated, associated with the requested subsystems.
 213
 214It's not currently possible to bind a new subsystem to an active
 215cgroup hierarchy, or to unbind a subsystem from an active cgroup
 216hierarchy. This may be possible in future, but is fraught with nasty
 217error-recovery issues.
 218
 219When a cgroup filesystem is unmounted, if there are any
 220child cgroups created below the top-level cgroup, that hierarchy
 221will remain active even though unmounted; if there are no
 222child cgroups then the hierarchy will be deactivated.
 223
 224No new system calls are added for cgroups - all support for
 225querying and modifying cgroups is via this cgroup file system.
 226
 227Each task under /proc has an added file named 'cgroup' displaying,
 228for each active hierarchy, the subsystem names and the cgroup name
 229as the path relative to the root of the cgroup file system.
 230
 231Each cgroup is represented by a directory in the cgroup file system
 232containing the following files describing that cgroup:
 233
 234 - tasks: list of tasks (by PID) attached to that cgroup.  This list
 235   is not guaranteed to be sorted.  Writing a thread ID into this file
 236   moves the thread into this cgroup.
 237 - cgroup.procs: list of thread group IDs in the cgroup.  This list is
 238   not guaranteed to be sorted or free of duplicate TGIDs, and userspace
 239   should sort/uniquify the list if this property is required.
 240   Writing a thread group ID into this file moves all threads in that
 241   group into this cgroup.
 242 - notify_on_release flag: run the release agent on exit?
 243 - release_agent: the path to use for release notifications (this file
 244   exists in the top cgroup only)
 245
 246Other subsystems such as cpusets may add additional files in each
 247cgroup dir.
 248
 249New cgroups are created using the mkdir system call or shell
 250command.  The properties of a cgroup, such as its flags, are
 251modified by writing to the appropriate file in that cgroups
 252directory, as listed above.
 253
 254The named hierarchical structure of nested cgroups allows partitioning
 255a large system into nested, dynamically changeable, "soft-partitions".
 256
 257The attachment of each task, automatically inherited at fork by any
 258children of that task, to a cgroup allows organizing the work load
 259on a system into related sets of tasks.  A task may be re-attached to
 260any other cgroup, if allowed by the permissions on the necessary
 261cgroup file system directories.
 262
 263When a task is moved from one cgroup to another, it gets a new
 264css_set pointer - if there's an already existing css_set with the
 265desired collection of cgroups then that group is reused, otherwise a new
 266css_set is allocated. The appropriate existing css_set is located by
 267looking into a hash table.
 268
 269To allow access from a cgroup to the css_sets (and hence tasks)
 270that comprise it, a set of cg_cgroup_link objects form a lattice;
 271each cg_cgroup_link is linked into a list of cg_cgroup_links for
 272a single cgroup on its cgrp_link_list field, and a list of
 273cg_cgroup_links for a single css_set on its cg_link_list.
 274
 275Thus the set of tasks in a cgroup can be listed by iterating over
 276each css_set that references the cgroup, and sub-iterating over
 277each css_set's task set.
 278
 279The use of a Linux virtual file system (vfs) to represent the
 280cgroup hierarchy provides for a familiar permission and name space
 281for cgroups, with a minimum of additional kernel code.
 282
 2831.4 What does notify_on_release do ?
 284------------------------------------
 285
 286If the notify_on_release flag is enabled (1) in a cgroup, then
 287whenever the last task in the cgroup leaves (exits or attaches to
 288some other cgroup) and the last child cgroup of that cgroup
 289is removed, then the kernel runs the command specified by the contents
 290of the "release_agent" file in that hierarchy's root directory,
 291supplying the pathname (relative to the mount point of the cgroup
 292file system) of the abandoned cgroup.  This enables automatic
 293removal of abandoned cgroups.  The default value of
 294notify_on_release in the root cgroup at system boot is disabled
 295(0).  The default value of other cgroups at creation is the current
 296value of their parents' notify_on_release settings. The default value of
 297a cgroup hierarchy's release_agent path is empty.
 298
 2991.5 What does clone_children do ?
 300---------------------------------
 301
 302If the clone_children flag is enabled (1) in a cgroup, then all
 303cgroups created beneath will call the post_clone callbacks for each
 304subsystem of the newly created cgroup. Usually when this callback is
 305implemented for a subsystem, it copies the values of the parent
 306subsystem, this is the case for the cpuset.
 307
 3081.6 How do I use cgroups ?
 309--------------------------
 310
 311To start a new job that is to be contained within a cgroup, using
 312the "cpuset" cgroup subsystem, the steps are something like:
 313
 314 1) mount -t tmpfs cgroup_root /sys/fs/cgroup
 315 2) mkdir /sys/fs/cgroup/cpuset
 316 3) mount -t cgroup -ocpuset cpuset /sys/fs/cgroup/cpuset
 317 4) Create the new cgroup by doing mkdir's and write's (or echo's) in
 318    the /sys/fs/cgroup virtual file system.
 319 5) Start a task that will be the "founding father" of the new job.
 320 6) Attach that task to the new cgroup by writing its PID to the
 321    /sys/fs/cgroup/cpuset/tasks file for that cgroup.
 322 7) fork, exec or clone the job tasks from this founding father task.
 323
 324For example, the following sequence of commands will setup a cgroup
 325named "Charlie", containing just CPUs 2 and 3, and Memory Node 1,
 326and then start a subshell 'sh' in that cgroup:
 327
 328  mount -t tmpfs cgroup_root /sys/fs/cgroup
 329  mkdir /sys/fs/cgroup/cpuset
 330  mount -t cgroup cpuset -ocpuset /sys/fs/cgroup/cpuset
 331  cd /sys/fs/cgroup/cpuset
 332  mkdir Charlie
 333  cd Charlie
 334  /bin/echo 2-3 > cpuset.cpus
 335  /bin/echo 1 > cpuset.mems
 336  /bin/echo $$ > tasks
 337  sh
 338  # The subshell 'sh' is now running in cgroup Charlie
 339  # The next line should display '/Charlie'
 340  cat /proc/self/cgroup
 341
 3422. Usage Examples and Syntax
 343============================
 344
 3452.1 Basic Usage
 346---------------
 347
 348Creating, modifying, using cgroups can be done through the cgroup
 349virtual filesystem.
 350
 351To mount a cgroup hierarchy with all available subsystems, type:
 352# mount -t cgroup xxx /sys/fs/cgroup
 353
 354The "xxx" is not interpreted by the cgroup code, but will appear in
 355/proc/mounts so may be any useful identifying string that you like.
 356
 357Note: Some subsystems do not work without some user input first.  For instance,
 358if cpusets are enabled the user will have to populate the cpus and mems files
 359for each new cgroup created before that group can be used.
 360
 361As explained in section `1.2 Why are cgroups needed?' you should create
 362different hierarchies of cgroups for each single resource or group of
 363resources you want to control. Therefore, you should mount a tmpfs on
 364/sys/fs/cgroup and create directories for each cgroup resource or resource
 365group.
 366
 367# mount -t tmpfs cgroup_root /sys/fs/cgroup
 368# mkdir /sys/fs/cgroup/rg1
 369
 370To mount a cgroup hierarchy with just the cpuset and memory
 371subsystems, type:
 372# mount -t cgroup -o cpuset,memory hier1 /sys/fs/cgroup/rg1
 373
 374While remounting cgroups is currently supported, it is not recommend
 375to use it. Remounting allows changing bound subsystems and
 376release_agent. Rebinding is hardly useful as it only works when the
 377hierarchy is empty and release_agent itself should be replaced with
 378conventional fsnotify. The support for remounting will be removed in
 379the future.
 380
 381To Specify a hierarchy's release_agent:
 382# mount -t cgroup -o cpuset,release_agent="/sbin/cpuset_release_agent" \
 383  xxx /sys/fs/cgroup/rg1
 384
 385Note that specifying 'release_agent' more than once will return failure.
 386
 387Note that changing the set of subsystems is currently only supported
 388when the hierarchy consists of a single (root) cgroup. Supporting
 389the ability to arbitrarily bind/unbind subsystems from an existing
 390cgroup hierarchy is intended to be implemented in the future.
 391
 392Then under /sys/fs/cgroup/rg1 you can find a tree that corresponds to the
 393tree of the cgroups in the system. For instance, /sys/fs/cgroup/rg1
 394is the cgroup that holds the whole system.
 395
 396If you want to change the value of release_agent:
 397# echo "/sbin/new_release_agent" > /sys/fs/cgroup/rg1/release_agent
 398
 399It can also be changed via remount.
 400
 401If you want to create a new cgroup under /sys/fs/cgroup/rg1:
 402# cd /sys/fs/cgroup/rg1
 403# mkdir my_cgroup
 404
 405Now you want to do something with this cgroup.
 406# cd my_cgroup
 407
 408In this directory you can find several files:
 409# ls
 410cgroup.procs notify_on_release tasks
 411(plus whatever files added by the attached subsystems)
 412
 413Now attach your shell to this cgroup:
 414# /bin/echo $$ > tasks
 415
 416You can also create cgroups inside your cgroup by using mkdir in this
 417directory.
 418# mkdir my_sub_cs
 419
 420To remove a cgroup, just use rmdir:
 421# rmdir my_sub_cs
 422
 423This will fail if the cgroup is in use (has cgroups inside, or
 424has processes attached, or is held alive by other subsystem-specific
 425reference).
 426
 4272.2 Attaching processes
 428-----------------------
 429
 430# /bin/echo PID > tasks
 431
 432Note that it is PID, not PIDs. You can only attach ONE task at a time.
 433If you have several tasks to attach, you have to do it one after another:
 434
 435# /bin/echo PID1 > tasks
 436# /bin/echo PID2 > tasks
 437        ...
 438# /bin/echo PIDn > tasks
 439
 440You can attach the current shell task by echoing 0:
 441
 442# echo 0 > tasks
 443
 444You can use the cgroup.procs file instead of the tasks file to move all
 445threads in a threadgroup at once. Echoing the PID of any task in a
 446threadgroup to cgroup.procs causes all tasks in that threadgroup to be
 447be attached to the cgroup. Writing 0 to cgroup.procs moves all tasks
 448in the writing task's threadgroup.
 449
 450Note: Since every task is always a member of exactly one cgroup in each
 451mounted hierarchy, to remove a task from its current cgroup you must
 452move it into a new cgroup (possibly the root cgroup) by writing to the
 453new cgroup's tasks file.
 454
 455Note: Due to some restrictions enforced by some cgroup subsystems, moving
 456a process to another cgroup can fail.
 457
 4582.3 Mounting hierarchies by name
 459--------------------------------
 460
 461Passing the name=<x> option when mounting a cgroups hierarchy
 462associates the given name with the hierarchy.  This can be used when
 463mounting a pre-existing hierarchy, in order to refer to it by name
 464rather than by its set of active subsystems.  Each hierarchy is either
 465nameless, or has a unique name.
 466
 467The name should match [\w.-]+
 468
 469When passing a name=<x> option for a new hierarchy, you need to
 470specify subsystems manually; the legacy behaviour of mounting all
 471subsystems when none are explicitly specified is not supported when
 472you give a subsystem a name.
 473
 474The name of the subsystem appears as part of the hierarchy description
 475in /proc/mounts and /proc/<pid>/cgroups.
 476
 4772.4 Notification API
 478--------------------
 479
 480There is mechanism which allows to get notifications about changing
 481status of a cgroup.
 482
 483To register a new notification handler you need to:
 484 - create a file descriptor for event notification using eventfd(2);
 485 - open a control file to be monitored (e.g. memory.usage_in_bytes);
 486 - write "<event_fd> <control_fd> <args>" to cgroup.event_control.
 487   Interpretation of args is defined by control file implementation;
 488
 489eventfd will be woken up by control file implementation or when the
 490cgroup is removed.
 491
 492To unregister a notification handler just close eventfd.
 493
 494NOTE: Support of notifications should be implemented for the control
 495file. See documentation for the subsystem.
 496
 4973. Kernel API
 498=============
 499
 5003.1 Overview
 501------------
 502
 503Each kernel subsystem that wants to hook into the generic cgroup
 504system needs to create a cgroup_subsys object. This contains
 505various methods, which are callbacks from the cgroup system, along
 506with a subsystem ID which will be assigned by the cgroup system.
 507
 508Other fields in the cgroup_subsys object include:
 509
 510- subsys_id: a unique array index for the subsystem, indicating which
 511  entry in cgroup->subsys[] this subsystem should be managing.
 512
 513- name: should be initialized to a unique subsystem name. Should be
 514  no longer than MAX_CGROUP_TYPE_NAMELEN.
 515
 516- early_init: indicate if the subsystem needs early initialization
 517  at system boot.
 518
 519Each cgroup object created by the system has an array of pointers,
 520indexed by subsystem ID; this pointer is entirely managed by the
 521subsystem; the generic cgroup code will never touch this pointer.
 522
 5233.2 Synchronization
 524-------------------
 525
 526There is a global mutex, cgroup_mutex, used by the cgroup
 527system. This should be taken by anything that wants to modify a
 528cgroup. It may also be taken to prevent cgroups from being
 529modified, but more specific locks may be more appropriate in that
 530situation.
 531
 532See kernel/cgroup.c for more details.
 533
 534Subsystems can take/release the cgroup_mutex via the functions
 535cgroup_lock()/cgroup_unlock().
 536
 537Accessing a task's cgroup pointer may be done in the following ways:
 538- while holding cgroup_mutex
 539- while holding the task's alloc_lock (via task_lock())
 540- inside an rcu_read_lock() section via rcu_dereference()
 541
 5423.3 Subsystem API
 543-----------------
 544
 545Each subsystem should:
 546
 547- add an entry in linux/cgroup_subsys.h
 548- define a cgroup_subsys object called <name>_subsys
 549
 550If a subsystem can be compiled as a module, it should also have in its
 551module initcall a call to cgroup_load_subsys(), and in its exitcall a
 552call to cgroup_unload_subsys(). It should also set its_subsys.module =
 553THIS_MODULE in its .c file.
 554
 555Each subsystem may export the following methods. The only mandatory
 556methods are create/destroy. Any others that are null are presumed to
 557be successful no-ops.
 558
 559struct cgroup_subsys_state *create(struct cgroup *cgrp)
 560(cgroup_mutex held by caller)
 561
 562Called to create a subsystem state object for a cgroup. The
 563subsystem should allocate its subsystem state object for the passed
 564cgroup, returning a pointer to the new object on success or a
 565negative error code. On success, the subsystem pointer should point to
 566a structure of type cgroup_subsys_state (typically embedded in a
 567larger subsystem-specific object), which will be initialized by the
 568cgroup system. Note that this will be called at initialization to
 569create the root subsystem state for this subsystem; this case can be
 570identified by the passed cgroup object having a NULL parent (since
 571it's the root of the hierarchy) and may be an appropriate place for
 572initialization code.
 573
 574void destroy(struct cgroup *cgrp)
 575(cgroup_mutex held by caller)
 576
 577The cgroup system is about to destroy the passed cgroup; the subsystem
 578should do any necessary cleanup and free its subsystem state
 579object. By the time this method is called, the cgroup has already been
 580unlinked from the file system and from the child list of its parent;
 581cgroup->parent is still valid. (Note - can also be called for a
 582newly-created cgroup if an error occurs after this subsystem's
 583create() method has been called for the new cgroup).
 584
 585int pre_destroy(struct cgroup *cgrp);
 586
 587Called before checking the reference count on each subsystem. This may
 588be useful for subsystems which have some extra references even if
 589there are not tasks in the cgroup. If pre_destroy() returns error code,
 590rmdir() will fail with it. From this behavior, pre_destroy() can be
 591called multiple times against a cgroup.
 592
 593int can_attach(struct cgroup *cgrp, struct cgroup_taskset *tset)
 594(cgroup_mutex held by caller)
 595
 596Called prior to moving one or more tasks into a cgroup; if the
 597subsystem returns an error, this will abort the attach operation.
 598@tset contains the tasks to be attached and is guaranteed to have at
 599least one task in it.
 600
 601If there are multiple tasks in the taskset, then:
 602  - it's guaranteed that all are from the same thread group
 603  - @tset contains all tasks from the thread group whether or not
 604    they're switching cgroups
 605  - the first task is the leader
 606
 607Each @tset entry also contains the task's old cgroup and tasks which
 608aren't switching cgroup can be skipped easily using the
 609cgroup_taskset_for_each() iterator. Note that this isn't called on a
 610fork. If this method returns 0 (success) then this should remain valid
 611while the caller holds cgroup_mutex and it is ensured that either
 612attach() or cancel_attach() will be called in future.
 613
 614void cancel_attach(struct cgroup *cgrp, struct cgroup_taskset *tset)
 615(cgroup_mutex held by caller)
 616
 617Called when a task attach operation has failed after can_attach() has succeeded.
 618A subsystem whose can_attach() has some side-effects should provide this
 619function, so that the subsystem can implement a rollback. If not, not necessary.
 620This will be called only about subsystems whose can_attach() operation have
 621succeeded. The parameters are identical to can_attach().
 622
 623void attach(struct cgroup *cgrp, struct cgroup_taskset *tset)
 624(cgroup_mutex held by caller)
 625
 626Called after the task has been attached to the cgroup, to allow any
 627post-attachment activity that requires memory allocations or blocking.
 628The parameters are identical to can_attach().
 629
 630void fork(struct task_struct *task)
 631
 632Called when a task is forked into a cgroup.
 633
 634void exit(struct task_struct *task)
 635
 636Called during task exit.
 637
 638void post_clone(struct cgroup *cgrp)
 639(cgroup_mutex held by caller)
 640
 641Called during cgroup_create() to do any parameter
 642initialization which might be required before a task could attach.  For
 643example, in cpusets, no task may attach before 'cpus' and 'mems' are set
 644up.
 645
 646void bind(struct cgroup *root)
 647(cgroup_mutex held by caller)
 648
 649Called when a cgroup subsystem is rebound to a different hierarchy
 650and root cgroup. Currently this will only involve movement between
 651the default hierarchy (which never has sub-cgroups) and a hierarchy
 652that is being created/destroyed (and hence has no sub-cgroups).
 653
 6544. Extended attribute usage
 655===========================
 656
 657cgroup filesystem supports certain types of extended attributes in its
 658directories and files.  The current supported types are:
 659        - Trusted (XATTR_TRUSTED)
 660        - Security (XATTR_SECURITY)
 661
 662Both require CAP_SYS_ADMIN capability to set.
 663
 664Like in tmpfs, the extended attributes in cgroup filesystem are stored
 665using kernel memory and it's advised to keep the usage at minimum.  This
 666is the reason why user defined extended attributes are not supported, since
 667any user can do it and there's no limit in the value size.
 668
 669The current known users for this feature are SELinux to limit cgroup usage
 670in containers and systemd for assorted meta data like main PID in a cgroup
 671(systemd creates a cgroup per service).
 672
 6735. Questions
 674============
 675
 676Q: what's up with this '/bin/echo' ?
 677A: bash's builtin 'echo' command does not check calls to write() against
 678   errors. If you use it in the cgroup file system, you won't be
 679   able to tell whether a command succeeded or failed.
 680
 681Q: When I attach processes, only the first of the line gets really attached !
 682A: We can only return one error code per call to write(). So you should also
 683   put only ONE PID.
 684
 685
lxr.linux.no kindly hosted by Redpill Linpro AS, provider of Linux consulting and operations services since 1995.