diff options
| -rw-r--r-- | Documentation/sched-design-CFS.txt | 119 |
1 files changed, 119 insertions, 0 deletions
diff --git a/Documentation/sched-design-CFS.txt b/Documentation/sched-design-CFS.txt new file mode 100644 index 00000000000..16feebb7bdc --- /dev/null +++ b/Documentation/sched-design-CFS.txt @@ -0,0 +1,119 @@ + +This is the CFS scheduler. + +80% of CFS's design can be summed up in a single sentence: CFS basically +models an "ideal, precise multi-tasking CPU" on real hardware. + +"Ideal multi-tasking CPU" is a (non-existent :-)) CPU that has 100% +physical power and which can run each task at precise equal speed, in +parallel, each at 1/nr_running speed. For example: if there are 2 tasks +running then it runs each at 50% physical power - totally in parallel. + +On real hardware, we can run only a single task at once, so while that +one task runs, the other tasks that are waiting for the CPU are at a +disadvantage - the current task gets an unfair amount of CPU time. In +CFS this fairness imbalance is expressed and tracked via the per-task +p->wait_runtime (nanosec-unit) value. "wait_runtime" is the amount of +time the task should now run on the CPU for it to become completely fair +and balanced. + +( small detail: on 'ideal' hardware, the p->wait_runtime value would + always be zero - no task would ever get 'out of balance' from the + 'ideal' share of CPU time. ) + +CFS's task picking logic is based on this p->wait_runtime value and it +is thus very simple: it always tries to run the task with the largest +p->wait_runtime value. In other words, CFS tries to run the task with +the 'gravest need' for more CPU time. So CFS always tries to split up +CPU time between runnable tasks as close to 'ideal multitasking +hardware' as possible. + +Most of the rest of CFS's design just falls out of this really simple +concept, with a few add-on embellishments like nice levels, +multiprocessing and various algorithm variants to recognize sleepers. + +In practice it works like this: the system runs a task a bit, and when +the task schedules (or a scheduler tick happens) the task's CPU usage is +'accounted for': the (small) time it just spent using the physical CPU +is deducted from p->wait_runtime. [minus the 'fair share' it would have +gotten anyway]. Once p->wait_runtime gets low enough so that another +task becomes the 'leftmost task' of the time-ordered rbtree it maintains +(plus a small amount of 'granularity' distance relative to the leftmost +task so that we do not over-schedule tasks and trash the cache) then the +new leftmost task is picked and the current task is preempted. + +The rq->fair_clock value tracks the 'CPU time a runnable task would have +fairly gotten, had it been runnable during that time'. So by using +rq->fair_clock values we can accurately timestamp and measure the +'expected CPU time' a task should have gotten. All runnable tasks are +sorted in the rbtree by the "rq->fair_clock - p->wait_runtime" key, and +CFS picks the 'leftmost' task and sticks to it. As the system progresses +forwards, newly woken tasks are put into the tree more and more to the +right - slowly but surely giving a chance for every task to become the +'leftmost task' and thus get on the CPU within a deterministic amount of +time. + +Some implementation details: + + - the introduction of Scheduling Classes: an extensible hierarchy of + scheduler modules. These modules encapsulate scheduling policy + details and are handled by the scheduler core without the core + code assuming about them too much. + + - sched_fair.c implements the 'CFS desktop scheduler': it is a + replacement for the vanilla scheduler's SCHED_OTHER interactivity + code. + + I'd like to give credit to Con Kolivas for the general approach here: + he has proven via RSDL/SD that 'fair scheduling' is possible and that + it results in better desktop scheduling. Kudos Con! + + The CFS patch uses a completely different approach and implementation + from RSDL/SD. My goal was to make CFS's interactivity quality exceed + that of RSDL/SD, which is a high standard to meet :-) Testing + feedback is welcome to decide this one way or another. [ and, in any + case, all of SD's logic could be added via a kernel/sched_sd.c module + as well, if Con is interested in such an approach. ] + + CFS's design is quite radical: it does not use runqueues, it uses a + time-ordered rbtree to build a 'timeline' of future task execution, + and thus has no 'array switch' artifacts (by which both the vanilla + scheduler and RSDL/SD are affected). + + CFS uses nanosecond granularity accounting and does not rely on any + jiffies or other HZ detail. Thus the CFS scheduler has no notion of + 'timeslices' and has no heuristics whatsoever. There is only one + central tunable: + + /proc/sys/kernel/sched_granularity_ns + + which can be used to tune the scheduler from 'desktop' (low + latencies) to 'server' (good batching) workloads. It defaults to a + setting suitable for desktop workloads. SCHED_BATCH is handled by the + CFS scheduler module too. + + Due to its design, the CFS scheduler is not prone to any of the + 'attacks' that exist today against the heuristics of the stock + scheduler: fiftyp.c, thud.c, chew.c, ring-test.c, massive_intr.c all + work fine and do not impact interactivity and produce the expected + behavior. + + the CFS scheduler has a much stronger handling of nice levels and + SCHED_BATCH: both types of workloads should be isolated much more + agressively than under the vanilla scheduler. + + ( another detail: due to nanosec accounting and timeline sorting, + sched_yield() support is very simple under CFS, and in fact under + CFS sched_yield() behaves much better than under any other + scheduler i have tested so far. ) + + - sched_rt.c implements SCHED_FIFO and SCHED_RR semantics, in a simpler + way than the vanilla scheduler does. It uses 100 runqueues (for all + 100 RT priority levels, instead of 140 in the vanilla scheduler) + and it needs no expired array. + + - reworked/sanitized SMP load-balancing: the runqueue-walking + assumptions are gone from the load-balancing code now, and + iterators of the scheduling modules are used. The balancing code got + quite a bit simpler as a result. + |