Messages in this thread Patch in this message |  | | | Subject | [RFC][PATCH 11/11] sched: add sched_dl documentation. | | From | Raistlin <> | | Date | Sun, 28 Feb 2010 20:28:12 +0100 |
| |
Add in Documentation/scheduler/ some hints about the design choices, the usage and the future possible developments of the sched_dl scheduling class and of the SCHED_DEADLINE policy.
Signed-off-by: Dario Faggioli <raistlin@linux.it> --- Documentation/scheduler/sched-deadline.txt | 188 ++++++++++++++++++++++++++++ init/Kconfig | 1 + 2 files changed, 189 insertions(+), 0 deletions(-) diff --git a/Documentation/scheduler/sched-deadline.txt b/Documentation/scheduler/sched-deadline.txt new file mode 100644 index 0000000..1ff0e1e --- /dev/null +++ b/Documentation/scheduler/sched-deadline.txt @@ -0,0 +1,188 @@ + Deadline Task and Group Scheduling + ---------------------------------- + +CONTENTS +======== + +0. WARNING +1. Overview + 1.1 Task scheduling + 1.2 Group scheduling +2. The interface + 2.1 System wide settings + 2.2 Task interface + 2.3 Group interface + 2.4 Default behavior +3. Future plans + + +0. WARNING +========== + + Fiddling with these settings can result in an unpredictable or even unstable + system behavior. As for -rt (group) scheduling, it is assumed that root + knows what he is doing. + + +1. Overview +=========== + + The SCHED_DEADLINE policy contained inside the sched_dl scheduling class is + basically an implementation of the Earliest Deadline First (EDF) scheduling + algorithm, augmented with a mechanism (called Constant Bandwidth Server, CBS) + that make it possible to isolate the behaviour of tasks between each other. + + +1.1 Task scheduling +------------------- + + The typical -deadline task will be made up of a computation phase (instance) + which is activated on a periodic or sporadic fashion. The expected (maximum) + duration of such computation is called the task's runtime; the time interval + by which each instance need to be completed is called the task's relative + deadline. The task's absolute deadline is dynamically calculated as the + time instant a task (better, an instance) activates plus the relative + deadline. + + The EDF algorithms selects the task with the smallest absolute deadline as + the one to be executed first, while the CBS ensures each task to run for + at most the its runtime every (relative) deadline length time interval, + avoiding any interference between different tasks (bandwidth isolation). + Thanks to this feature, also tasks that do not strictly comply with the + computational model sketched above can effectively use the new policy. + IOW, there are no limitations on what kind of task can exploit this new + scheduling discipline, even if it must be said that it is particularly + suited for periodic or sporadic tasks that need guarantees on their + timing behaviour, e.g., multimedia, streaming, control applications, etc. + + +1.2 Group scheduling +---------------- + + In order of -deaadline scheduling to be effective and useful, it is important + that some method of having the allocation of the available CPU bandwidth to + tasks and task groups under control. + This is usually called "admission control" and if it is not performed at all, + no guarantee can be given on the actual scheduling of the -deadline tasks. + + Since when RT-throttling has been introduced each task group have a bandwidth + associated to itself, calculated as a certain amount of runtime over a + period. Moreover, to make it possible to manipulate such bandwidth, + readable/writable controls have been added to both procfs (for system + wide settings) and cgroupfs (for per-group settings). + Therefore, the same interface is being used for controlling the bandwidth + distrubution to -deadline tasks and task groups, i.e., new controls but + with similar names, equivalent meaning and with the same usage paradigm are + added. + + The main differences between deadline bandwidth management and RT-throttling + is that -deadline tasks have bandwidth on their own (while -rt ones doesn't!), + and thus we don't need a throttling mechanism in the groups, which can be + used nothing more than for admission control of tasks. + + This means that what we check is the sum of the bandwidth of all the tasks + belonging to the group stays, on each CPU, below the bandwidth of the group + itself. + + +2. The Interface +================ + +2.1 System wide settings +------------------------ + +The system wide settings are configured under the /proc virtual file system: + + The per-group controls that are added to the cgroupfs virtual file system are: + * /proc/sys/kernel/sched_dl_runtime_us, + * /proc/sys/kernel/sched_dl_period_us, + * /proc/sys/kernel/sched_dl_total_bw. + + The first two accepts (if written) and provides (if read) the new runtime and + period, respectively, for each CPU. + The last one accepts (if written) the index of one online CPU, and it provides + (if read) the total amount of bandwidth currently alloceted on that CPU. + + Settings are checked against the following limit: + + * for the whole system, on each CPU: + rt_runtime / rt_period + dl_runtime + dl_period <= 100% + + +2.2 Task interface +------------------ + + Specifying a periodic/sporadic task that executes for a given amount of + runtime at each instance, and that is scheduled according to the usrgency of + their own timing constraints needs, in general, a way of declaring: + - a (maximum/typical) instance execution time, + - a minimum interval between consecutive instances, + - a time constraint by which each instance must be completed. + + Therefore: + * a new struct sched_param_ex, containing all the necessary fields is + provided; + * the new scheduling related syscalls that manipulate it, i.e., + sched_setscheduler_ex(), sched_setparam_ex() and sched_getparam_ex() + are implemented. + + +2.3 Group Interface +------------------- + + The per-group controls that are added to the cgroupfs virtual file system are: + * /cgroup/<cgroup>/cpu.dl_runtime_us, + * /cgroup/<cgroup>/cpu.dl_period_us, + * /cgroup/<cgroup>/cpu.dl_total_bw. + + The first two accepts (if written) and provides (if read) the new runtime and + period, respectively, of the group for each CPU. + The last one accepts (if written) the index of one online CPU, and it provides + (if read) the total amount of bandwidth currently alloceted inside the group + on that CPU. + + Group settings are checked against the following limits: + + * for the root group {r}, on each CPU: + dl_runtime_{r} / dl_period_{r} <= dl_runtime / dl_period + + * for each group {i}, subgroup of group {j}, on each CPU: + dl_runtime_{i} / dl_period_{i} < 100% + \Sum_{i} dl_runtime_{i} / dl_period_{i} <= dl_runtime_{j} / dl_period_{j} + + For more information on working with control groups, + Documentation/cgroups/cgroups.txt should be read. + + +2.4 Default behavior +--------------------- + +The default values for system wide and root group dl_runtime and dl_period are +500000 over 1000000. This means -deadline tasks and task groups can use at +most 5% bandwidth on each CPU. + +When a -deadline task fork a child, its dl_runtime is set to 0, which means +someone must call sched_setscheduler_ex() on it, or it won't even start. + +When a new group is created, its dl_runtime is 0, which means someone must +(try to) increase it before tasks can be added to the group. + + +3. Future plans +=============== + +Still Missing parts: + + - bandwidth reclaiming mechanisms, i.e., methods that avoid stopping the + tasks until their next deadline when overrunning. There are at least + three of them that are very simple, and patches are on their way; + + - migration of tasks throughout push and pull (as in -rt) to make it + possible to deploy global-EDF scheduling. Patches are ready, they're + just being tested and adapted to this last version; + + - refinements in deadline inheritance, especially regarding the possibility + of retaining bandwidth isolation among non-interacting tasks. This is + being studied from both theoretical and practical point of views, and + hopefully we can have some demonstrative code soon. + diff --git a/init/Kconfig b/init/Kconfig index de57415..377caed 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -486,6 +486,7 @@ config DEADLINE_GROUP_SCHED tasks (and other groups) can be added to it only up to such "bandwidth cap", which might be useful for avoiding or controlling oversubscription. + See Documentation/scheduler/sched-deadline.txt for more. choice depends on GROUP_SCHED -- 1.7.0 -- <<This happens because I choose it to happen!>> (Raistlin Majere) ---------------------------------------------------------------------- Dario Faggioli, ReTiS Lab, Scuola Superiore Sant'Anna, Pisa (Italy) http://blog.linux.it/raistlin / raistlin@ekiga.net / dario.faggioli@jabber.org [unhandled content-type:application/pgp-signature] |  |