5 Each CPU has a "base" scheduling domain (struct sched_domain). The domain
6 hierarchy is built from these base domains via the ->parent pointer. ->parent
7 MUST be NULL terminated, and domain structures should be per-CPU as they are
10 Each scheduling domain spans a number of CPUs (stored in the ->span field).
11 A domain's span MUST be a superset of it child's span (this restriction could
12 be relaxed if the need arises), and a base domain for CPU i MUST span at least
13 i. The top domain for each CPU will generally span all CPUs in the system
14 although strictly it doesn't have to, but this could lead to a case where some
15 CPUs will never be given tasks to run unless the CPUs allowed mask is
16 explicitly set. A sched domain's span means "balance process load among these
19 Each scheduling domain must have one or more CPU groups (struct sched_group)
20 which are organised as a circular one way linked list from the ->groups
21 pointer. The union of cpumasks of these groups MUST be the same as the
22 domain's span. The intersection of cpumasks from any two of these groups
23 MUST be the empty set. The group pointed to by the ->groups pointer MUST
24 contain the CPU to which the domain belongs. Groups may be shared among
25 CPUs as they contain read only data after they have been set up.
27 Balancing within a sched domain occurs between groups. That is, each group
28 is treated as one entity. The load of a group is defined as the sum of the
29 load of each of its member CPUs, and only when the load of a group becomes
30 out of balance are tasks moved between groups.
32 In kernel/sched/core.c, trigger_load_balance() is run periodically on each CPU
33 through scheduler_tick(). It raises a softirq after the next regularly scheduled
34 rebalancing event for the current runqueue has arrived. The actual load
35 balancing workhorse, run_rebalance_domains()->rebalance_domains(), is then run
36 in softirq context (SCHED_SOFTIRQ).
38 The latter function takes two arguments: the current CPU and whether it was idle
39 at the time the scheduler_tick() happened and iterates over all sched domains
40 our CPU is on, starting from its base domain and going up the ->parent chain.
41 While doing that, it checks to see if the current domain has exhausted its
42 rebalance interval. If so, it runs load_balance() on that domain. It then checks
43 the parent sched_domain (if it exists), and the parent of the parent and so
46 Initially, load_balance() finds the busiest group in the current sched domain.
47 If it succeeds, it looks for the busiest runqueue of all the CPUs' runqueues in
48 that group. If it manages to find such a runqueue, it locks both our initial
49 CPU's runqueue and the newly found busiest one and starts moving tasks from it
50 to our runqueue. The exact number of tasks amounts to an imbalance previously
51 computed while iterating over this sched domain's groups.
53 Implementing sched domains
54 ==========================
56 The "base" domain will "span" the first level of the hierarchy. In the case
57 of SMT, you'll span all siblings of the physical CPU, with each group being
60 In SMP, the parent of the base domain will span all physical CPUs in the
61 node. Each group being a single physical CPU. Then with NUMA, the parent
62 of the SMP domain will span the entire machine, with each group having the
63 cpumask of a node. Or, you could do multi-level NUMA or Opteron, for example,
64 might have just one domain covering its one NUMA level.
66 The implementor should read comments in include/linux/sched.h:
67 struct sched_domain fields, SD_FLAG_*, SD_*_INIT to get an idea of
68 the specifics and what to tune.
70 Architectures may retain the regular override the default SD_*_INIT flags
71 while using the generic domain builder in kernel/sched/core.c if they wish to
72 retain the traditional SMT->SMP->NUMA topology (or some subset of that). This
73 can be done by #define'ing ARCH_HASH_SCHED_TUNE.
75 Alternatively, the architecture may completely override the generic domain
76 builder by #define'ing ARCH_HASH_SCHED_DOMAIN, and exporting your
77 arch_init_sched_domains function. This function will attach domains to all
78 CPUs using cpu_attach_domain.
80 The sched-domains debugging infrastructure can be enabled by enabling
81 CONFIG_SCHED_DEBUG. This enables an error checking parse of the sched domains
82 which should catch most possible errors (described above). It also prints out
83 the domain structure in a visual format.