~ [ source navigation ] ~ [ diff markup ] ~ [ identifier search ] ~

TOMOYO Linux Cross Reference
Linux/Documentation/scheduler/sched-design-CFS.rst

Version: ~ [ linux-6.12-rc7 ] ~ [ linux-6.11.7 ] ~ [ linux-6.10.14 ] ~ [ linux-6.9.12 ] ~ [ linux-6.8.12 ] ~ [ linux-6.7.12 ] ~ [ linux-6.6.60 ] ~ [ linux-6.5.13 ] ~ [ linux-6.4.16 ] ~ [ linux-6.3.13 ] ~ [ linux-6.2.16 ] ~ [ linux-6.1.116 ] ~ [ linux-6.0.19 ] ~ [ linux-5.19.17 ] ~ [ linux-5.18.19 ] ~ [ linux-5.17.15 ] ~ [ linux-5.16.20 ] ~ [ linux-5.15.171 ] ~ [ linux-5.14.21 ] ~ [ linux-5.13.19 ] ~ [ linux-5.12.19 ] ~ [ linux-5.11.22 ] ~ [ linux-5.10.229 ] ~ [ linux-5.9.16 ] ~ [ linux-5.8.18 ] ~ [ linux-5.7.19 ] ~ [ linux-5.6.19 ] ~ [ linux-5.5.19 ] ~ [ linux-5.4.285 ] ~ [ linux-5.3.18 ] ~ [ linux-5.2.21 ] ~ [ linux-5.1.21 ] ~ [ linux-5.0.21 ] ~ [ linux-4.20.17 ] ~ [ linux-4.19.323 ] ~ [ linux-4.18.20 ] ~ [ linux-4.17.19 ] ~ [ linux-4.16.18 ] ~ [ linux-4.15.18 ] ~ [ linux-4.14.336 ] ~ [ linux-4.13.16 ] ~ [ linux-4.12.14 ] ~ [ linux-4.11.12 ] ~ [ linux-4.10.17 ] ~ [ linux-4.9.337 ] ~ [ linux-4.4.302 ] ~ [ linux-3.10.108 ] ~ [ linux-2.6.32.71 ] ~ [ linux-2.6.0 ] ~ [ linux-2.4.37.11 ] ~ [ unix-v6-master ] ~ [ ccs-tools-1.8.12 ] ~ [ policy-sample ] ~
Architecture: ~ [ i386 ] ~ [ alpha ] ~ [ m68k ] ~ [ mips ] ~ [ ppc ] ~ [ sparc ] ~ [ sparc64 ] ~

  1 .. _sched_design_CFS:
  2 
  3 =============
  4 CFS Scheduler
  5 =============
  6 
  7 
  8 1.  OVERVIEW
  9 ============
 10 
 11 CFS stands for "Completely Fair Scheduler," and is the "desktop" process
 12 scheduler implemented by Ingo Molnar and merged in Linux 2.6.23. When
 13 originally merged, it was the replacement for the previous vanilla
 14 scheduler's SCHED_OTHER interactivity code. Nowadays, CFS is making room
 15 for EEVDF, for which documentation can be found in
 16 Documentation/scheduler/sched-eevdf.rst.
 17 
 18 80% of CFS's design can be summed up in a single sentence: CFS basically models
 19 an "ideal, precise multi-tasking CPU" on real hardware.
 20 
 21 "Ideal multi-tasking CPU" is a (non-existent  :-)) CPU that has 100% physical
 22 power and which can run each task at precise equal speed, in parallel, each at
 23 1/nr_running speed.  For example: if there are 2 tasks running, then it runs
 24 each at 50% physical power --- i.e., actually in parallel.
 25 
 26 On real hardware, we can run only a single task at once, so we have to
 27 introduce the concept of "virtual runtime."  The virtual runtime of a task
 28 specifies when its next timeslice would start execution on the ideal
 29 multi-tasking CPU described above.  In practice, the virtual runtime of a task
 30 is its actual runtime normalized to the total number of running tasks.
 31 
 32 
 33 
 34 2.  FEW IMPLEMENTATION DETAILS
 35 ==============================
 36 
 37 In CFS the virtual runtime is expressed and tracked via the per-task
 38 p->se.vruntime (nanosec-unit) value.  This way, it's possible to accurately
 39 timestamp and measure the "expected CPU time" a task should have gotten.
 40 
 41    Small detail: on "ideal" hardware, at any time all tasks would have the same
 42    p->se.vruntime value --- i.e., tasks would execute simultaneously and no task
 43    would ever get "out of balance" from the "ideal" share of CPU time.
 44 
 45 CFS's task picking logic is based on this p->se.vruntime value and it is thus
 46 very simple: it always tries to run the task with the smallest p->se.vruntime
 47 value (i.e., the task which executed least so far).  CFS always tries to split
 48 up CPU time between runnable tasks as close to "ideal multitasking hardware" as
 49 possible.
 50 
 51 Most of the rest of CFS's design just falls out of this really simple concept,
 52 with a few add-on embellishments like nice levels, multiprocessing and various
 53 algorithm variants to recognize sleepers.
 54 
 55 
 56 
 57 3.  THE RBTREE
 58 ==============
 59 
 60 CFS's design is quite radical: it does not use the old data structures for the
 61 runqueues, but it uses a time-ordered rbtree to build a "timeline" of future
 62 task execution, and thus has no "array switch" artifacts (by which both the
 63 previous vanilla scheduler and RSDL/SD are affected).
 64 
 65 CFS also maintains the rq->cfs.min_vruntime value, which is a monotonic
 66 increasing value tracking the smallest vruntime among all tasks in the
 67 runqueue.  The total amount of work done by the system is tracked using
 68 min_vruntime; that value is used to place newly activated entities on the left
 69 side of the tree as much as possible.
 70 
 71 The total number of running tasks in the runqueue is accounted through the
 72 rq->cfs.load value, which is the sum of the weights of the tasks queued on the
 73 runqueue.
 74 
 75 CFS maintains a time-ordered rbtree, where all runnable tasks are sorted by the
 76 p->se.vruntime key. CFS picks the "leftmost" task from this tree and sticks to it.
 77 As the system progresses forwards, the executed tasks are put into the tree
 78 more and more to the right --- slowly but surely giving a chance for every task
 79 to become the "leftmost task" and thus get on the CPU within a deterministic
 80 amount of time.
 81 
 82 Summing up, CFS works like this: it runs a task a bit, and when the task
 83 schedules (or a scheduler tick happens) the task's CPU usage is "accounted
 84 for": the (small) time it just spent using the physical CPU is added to
 85 p->se.vruntime.  Once p->se.vruntime gets high enough so that another task
 86 becomes the "leftmost task" of the time-ordered rbtree it maintains (plus a
 87 small amount of "granularity" distance relative to the leftmost task so that we
 88 do not over-schedule tasks and trash the cache), then the new leftmost task is
 89 picked and the current task is preempted.
 90 
 91 
 92 
 93 4.  SOME FEATURES OF CFS
 94 ========================
 95 
 96 CFS uses nanosecond granularity accounting and does not rely on any jiffies or
 97 other HZ detail.  Thus the CFS scheduler has no notion of "timeslices" in the
 98 way the previous scheduler had, and has no heuristics whatsoever.  There is
 99 only one central tunable (you have to switch on CONFIG_SCHED_DEBUG):
100 
101    /sys/kernel/debug/sched/base_slice_ns
102 
103 which can be used to tune the scheduler from "desktop" (i.e., low latencies) to
104 "server" (i.e., good batching) workloads.  It defaults to a setting suitable
105 for desktop workloads.  SCHED_BATCH is handled by the CFS scheduler module too.
106 
107 In case CONFIG_HZ results in base_slice_ns < TICK_NSEC, the value of
108 base_slice_ns will have little to no impact on the workloads.
109 
110 Due to its design, the CFS scheduler is not prone to any of the "attacks" that
111 exist today against the heuristics of the stock scheduler: fiftyp.c, thud.c,
112 chew.c, ring-test.c, massive_intr.c all work fine and do not impact
113 interactivity and produce the expected behavior.
114 
115 The CFS scheduler has a much stronger handling of nice levels and SCHED_BATCH
116 than the previous vanilla scheduler: both types of workloads are isolated much
117 more aggressively.
118 
119 SMP load-balancing has been reworked/sanitized: the runqueue-walking
120 assumptions are gone from the load-balancing code now, and iterators of the
121 scheduling modules are used.  The balancing code got quite a bit simpler as a
122 result.
123 
124 
125 
126 5. Scheduling policies
127 ======================
128 
129 CFS implements three scheduling policies:
130 
131   - SCHED_NORMAL (traditionally called SCHED_OTHER): The scheduling
132     policy that is used for regular tasks.
133 
134   - SCHED_BATCH: Does not preempt nearly as often as regular tasks
135     would, thereby allowing tasks to run longer and make better use of
136     caches but at the cost of interactivity. This is well suited for
137     batch jobs.
138 
139   - SCHED_IDLE: This is even weaker than nice 19, but its not a true
140     idle timer scheduler in order to avoid to get into priority
141     inversion problems which would deadlock the machine.
142 
143 SCHED_FIFO/_RR are implemented in sched/rt.c and are as specified by
144 POSIX.
145 
146 The command chrt from util-linux-ng 2.13.1.1 can set all of these except
147 SCHED_IDLE.
148 
149 
150 
151 6.  SCHEDULING CLASSES
152 ======================
153 
154 The new CFS scheduler has been designed in such a way to introduce "Scheduling
155 Classes," an extensible hierarchy of scheduler modules.  These modules
156 encapsulate scheduling policy details and are handled by the scheduler core
157 without the core code assuming too much about them.
158 
159 sched/fair.c implements the CFS scheduler described above.
160 
161 sched/rt.c implements SCHED_FIFO and SCHED_RR semantics, in a simpler way than
162 the previous vanilla scheduler did.  It uses 100 runqueues (for all 100 RT
163 priority levels, instead of 140 in the previous scheduler) and it needs no
164 expired array.
165 
166 Scheduling classes are implemented through the sched_class structure, which
167 contains hooks to functions that must be called whenever an interesting event
168 occurs.
169 
170 This is the (partial) list of the hooks:
171 
172  - enqueue_task(...)
173 
174    Called when a task enters a runnable state.
175    It puts the scheduling entity (task) into the red-black tree and
176    increments the nr_running variable.
177 
178  - dequeue_task(...)
179 
180    When a task is no longer runnable, this function is called to keep the
181    corresponding scheduling entity out of the red-black tree.  It decrements
182    the nr_running variable.
183 
184  - yield_task(...)
185 
186    This function is basically just a dequeue followed by an enqueue, unless the
187    compat_yield sysctl is turned on; in that case, it places the scheduling
188    entity at the right-most end of the red-black tree.
189 
190  - wakeup_preempt(...)
191 
192    This function checks if a task that entered the runnable state should
193    preempt the currently running task.
194 
195  - pick_next_task(...)
196 
197    This function chooses the most appropriate task eligible to run next.
198 
199  - set_next_task(...)
200 
201    This function is called when a task changes its scheduling class, changes
202    its task group or is scheduled.
203 
204  - task_tick(...)
205 
206    This function is mostly called from time tick functions; it might lead to
207    process switch.  This drives the running preemption.
208 
209 
210 
211 
212 7.  GROUP SCHEDULER EXTENSIONS TO CFS
213 =====================================
214 
215 Normally, the scheduler operates on individual tasks and strives to provide
216 fair CPU time to each task.  Sometimes, it may be desirable to group tasks and
217 provide fair CPU time to each such task group.  For example, it may be
218 desirable to first provide fair CPU time to each user on the system and then to
219 each task belonging to a user.
220 
221 CONFIG_CGROUP_SCHED strives to achieve exactly that.  It lets tasks to be
222 grouped and divides CPU time fairly among such groups.
223 
224 CONFIG_RT_GROUP_SCHED permits to group real-time (i.e., SCHED_FIFO and
225 SCHED_RR) tasks.
226 
227 CONFIG_FAIR_GROUP_SCHED permits to group CFS (i.e., SCHED_NORMAL and
228 SCHED_BATCH) tasks.
229 
230    These options need CONFIG_CGROUPS to be defined, and let the administrator
231    create arbitrary groups of tasks, using the "cgroup" pseudo filesystem.  See
232    Documentation/admin-guide/cgroup-v1/cgroups.rst for more information about this filesystem.
233 
234 When CONFIG_FAIR_GROUP_SCHED is defined, a "cpu.shares" file is created for each
235 group created using the pseudo filesystem.  See example steps below to create
236 task groups and modify their CPU share using the "cgroups" pseudo filesystem::
237 
238         # mount -t tmpfs cgroup_root /sys/fs/cgroup
239         # mkdir /sys/fs/cgroup/cpu
240         # mount -t cgroup -ocpu none /sys/fs/cgroup/cpu
241         # cd /sys/fs/cgroup/cpu
242 
243         # mkdir multimedia      # create "multimedia" group of tasks
244         # mkdir browser         # create "browser" group of tasks
245 
246         # #Configure the multimedia group to receive twice the CPU bandwidth
247         # #that of browser group
248 
249         # echo 2048 > multimedia/cpu.shares
250         # echo 1024 > browser/cpu.shares
251 
252         # firefox &     # Launch firefox and move it to "browser" group
253         # echo <firefox_pid> > browser/tasks
254 
255         # #Launch gmplayer (or your favourite movie player)
256         # echo <movie_player_pid> > multimedia/tasks

~ [ source navigation ] ~ [ diff markup ] ~ [ identifier search ] ~

kernel.org | git.kernel.org | LWN.net | Project Home | SVN repository | Mail admin

Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.

sflogo.php