Blame view
kernel/sched_fair.c
101 KB
bf0f6f24a sched: cfs core, ... |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 |
/* * Completely Fair Scheduling (CFS) Class (SCHED_NORMAL/SCHED_BATCH) * * Copyright (C) 2007 Red Hat, Inc., Ingo Molnar <mingo@redhat.com> * * Interactivity improvements by Mike Galbraith * (C) 2007 Mike Galbraith <efault@gmx.de> * * Various enhancements by Dmitry Adamushko. * (C) 2007 Dmitry Adamushko <dmitry.adamushko@gmail.com> * * Group scheduling enhancements by Srivatsa Vaddagiri * Copyright IBM Corporation, 2007 * Author: Srivatsa Vaddagiri <vatsa@linux.vnet.ibm.com> * * Scaled math optimizations by Thomas Gleixner * Copyright (C) 2007, Thomas Gleixner <tglx@linutronix.de> |
218050855 sched: adaptive s... |
18 19 20 |
* * Adaptive scheduling granularity, math enhancements by Peter Zijlstra * Copyright (C) 2007 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com> |
bf0f6f24a sched: cfs core, ... |
21 |
*/ |
9745512ce sched: latencytop... |
22 |
#include <linux/latencytop.h> |
1983a922a sched: Make tunab... |
23 |
#include <linux/sched.h> |
9745512ce sched: latencytop... |
24 |
|
bf0f6f24a sched: cfs core, ... |
25 |
/* |
218050855 sched: adaptive s... |
26 |
* Targeted preemption latency for CPU-bound tasks: |
864616ee6 sched: Comment up... |
27 |
* (default: 6ms * (1 + ilog(ncpus)), units: nanoseconds) |
bf0f6f24a sched: cfs core, ... |
28 |
* |
218050855 sched: adaptive s... |
29 |
* NOTE: this latency value is not the same as the concept of |
d274a4cee sched: update com... |
30 31 32 |
* 'timeslice length' - timeslices in CFS are of variable length * and have no persistent notion like in traditional, time-slice * based scheduling concepts. |
bf0f6f24a sched: cfs core, ... |
33 |
* |
d274a4cee sched: update com... |
34 35 |
* (to see the precise effective timeslice length of your workload, * run vmstat and monitor the context-switches (cs) field) |
bf0f6f24a sched: cfs core, ... |
36 |
*/ |
21406928a sched: Tweak sche... |
37 38 |
unsigned int sysctl_sched_latency = 6000000ULL; unsigned int normalized_sysctl_sched_latency = 6000000ULL; |
2bd8e6d42 sched: use consta... |
39 40 |
/* |
1983a922a sched: Make tunab... |
41 42 43 44 45 46 47 48 49 50 51 52 |
* The initial- and re-scaling of tunables is configurable * (default SCHED_TUNABLESCALING_LOG = *(1+ilog(ncpus)) * * Options are: * SCHED_TUNABLESCALING_NONE - unscaled, always *1 * SCHED_TUNABLESCALING_LOG - scaled logarithmical, *1+ilog(ncpus) * SCHED_TUNABLESCALING_LINEAR - scaled linear, *ncpus */ enum sched_tunable_scaling sysctl_sched_tunable_scaling = SCHED_TUNABLESCALING_LOG; /* |
b2be5e96d sched: reintroduc... |
53 |
* Minimal preemption granularity for CPU-bound tasks: |
864616ee6 sched: Comment up... |
54 |
* (default: 0.75 msec * (1 + ilog(ncpus)), units: nanoseconds) |
2bd8e6d42 sched: use consta... |
55 |
*/ |
0bf377bbb sched: Improve la... |
56 57 |
unsigned int sysctl_sched_min_granularity = 750000ULL; unsigned int normalized_sysctl_sched_min_granularity = 750000ULL; |
218050855 sched: adaptive s... |
58 59 |
/* |
b2be5e96d sched: reintroduc... |
60 61 |
* is kept at sysctl_sched_latency / sysctl_sched_min_granularity */ |
0bf377bbb sched: Improve la... |
62 |
static unsigned int sched_nr_latency = 8; |
b2be5e96d sched: reintroduc... |
63 64 |
/* |
2bba22c50 sched: Turn off c... |
65 |
* After fork, child runs first. If set to 0 (default) then |
b2be5e96d sched: reintroduc... |
66 |
* parent will (try to) run first. |
218050855 sched: adaptive s... |
67 |
*/ |
2bba22c50 sched: Turn off c... |
68 |
unsigned int sysctl_sched_child_runs_first __read_mostly; |
bf0f6f24a sched: cfs core, ... |
69 70 |
/* |
1799e35d5 sched: add /proc/... |
71 72 73 74 75 76 77 78 |
* sys_sched_yield() compat mode * * This option switches the agressive yield implementation of the * old scheduler back on. */ unsigned int __read_mostly sysctl_sched_compat_yield; /* |
bf0f6f24a sched: cfs core, ... |
79 |
* SCHED_OTHER wake-up granularity. |
172e082a9 sched: Re-tune th... |
80 |
* (default: 1 msec * (1 + ilog(ncpus)), units: nanoseconds) |
bf0f6f24a sched: cfs core, ... |
81 82 83 84 85 |
* * This option delays the preemption effects of decoupled workloads * and reduces their over-scheduling. Synchronous workloads will still * have immediate wakeup/sleep latencies. */ |
172e082a9 sched: Re-tune th... |
86 |
unsigned int sysctl_sched_wakeup_granularity = 1000000UL; |
0bcdcf28c sched: Fix missin... |
87 |
unsigned int normalized_sysctl_sched_wakeup_granularity = 1000000UL; |
bf0f6f24a sched: cfs core, ... |
88 |
|
da84d9617 sched: reintroduc... |
89 |
const_debug unsigned int sysctl_sched_migration_cost = 500000UL; |
a4c2f00f5 sched: fair sched... |
90 |
static const struct sched_class fair_sched_class; |
bf0f6f24a sched: cfs core, ... |
91 92 93 |
/************************************************************** * CFS operations on generic schedulable entities: */ |
62160e3f4 sched: track cfs_... |
94 |
#ifdef CONFIG_FAIR_GROUP_SCHED |
bf0f6f24a sched: cfs core, ... |
95 |
|
62160e3f4 sched: track cfs_... |
96 |
/* cpu runqueue to which this cfs_rq is attached */ |
bf0f6f24a sched: cfs core, ... |
97 98 |
static inline struct rq *rq_of(struct cfs_rq *cfs_rq) { |
62160e3f4 sched: track cfs_... |
99 |
return cfs_rq->rq; |
bf0f6f24a sched: cfs core, ... |
100 |
} |
62160e3f4 sched: track cfs_... |
101 102 |
/* An entity is a task if it doesn't "own" a runqueue */ #define entity_is_task(se) (!se->my_q) |
bf0f6f24a sched: cfs core, ... |
103 |
|
8f48894fc sched: Add debug ... |
104 105 106 107 108 109 110 |
static inline struct task_struct *task_of(struct sched_entity *se) { #ifdef CONFIG_SCHED_DEBUG WARN_ON_ONCE(!entity_is_task(se)); #endif return container_of(se, struct task_struct, se); } |
b758149c0 sched: prepatory ... |
111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 |
/* Walk up scheduling entities hierarchy */ #define for_each_sched_entity(se) \ for (; se; se = se->parent) static inline struct cfs_rq *task_cfs_rq(struct task_struct *p) { return p->se.cfs_rq; } /* runqueue on which this entity is (to be) queued */ static inline struct cfs_rq *cfs_rq_of(struct sched_entity *se) { return se->cfs_rq; } /* runqueue "owned" by this group */ static inline struct cfs_rq *group_cfs_rq(struct sched_entity *grp) { return grp->my_q; } /* Given a group's cfs_rq on one cpu, return its corresponding cfs_rq on * another cpu ('this_cpu') */ static inline struct cfs_rq *cpu_cfs_rq(struct cfs_rq *cfs_rq, int this_cpu) { return cfs_rq->tg->cfs_rq[this_cpu]; } /* Iterate thr' all leaf cfs_rq's on a runqueue */ #define for_each_leaf_cfs_rq(rq, cfs_rq) \ list_for_each_entry_rcu(cfs_rq, &rq->leaf_cfs_rq_list, leaf_cfs_rq_list) /* Do the two (enqueued) entities belong to the same group ? */ static inline int is_same_group(struct sched_entity *se, struct sched_entity *pse) { if (se->cfs_rq == pse->cfs_rq) return 1; return 0; } static inline struct sched_entity *parent_entity(struct sched_entity *se) { return se->parent; } |
464b75273 sched: re-instate... |
158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 |
/* return depth at which a sched entity is present in the hierarchy */ static inline int depth_se(struct sched_entity *se) { int depth = 0; for_each_sched_entity(se) depth++; return depth; } static void find_matching_se(struct sched_entity **se, struct sched_entity **pse) { int se_depth, pse_depth; /* * preemption test can be made between sibling entities who are in the * same cfs_rq i.e who have a common parent. Walk up the hierarchy of * both tasks until we find their ancestors who are siblings of common * parent. */ /* First walk up until both entities are at same depth */ se_depth = depth_se(*se); pse_depth = depth_se(*pse); while (se_depth > pse_depth) { se_depth--; *se = parent_entity(*se); } while (pse_depth > se_depth) { pse_depth--; *pse = parent_entity(*pse); } while (!is_same_group(*se, *pse)) { *se = parent_entity(*se); *pse = parent_entity(*pse); } } |
8f48894fc sched: Add debug ... |
200 201 202 203 204 205 |
#else /* !CONFIG_FAIR_GROUP_SCHED */ static inline struct task_struct *task_of(struct sched_entity *se) { return container_of(se, struct task_struct, se); } |
bf0f6f24a sched: cfs core, ... |
206 |
|
62160e3f4 sched: track cfs_... |
207 208 209 |
static inline struct rq *rq_of(struct cfs_rq *cfs_rq) { return container_of(cfs_rq, struct rq, cfs); |
bf0f6f24a sched: cfs core, ... |
210 211 212 |
} #define entity_is_task(se) 1 |
b758149c0 sched: prepatory ... |
213 214 |
#define for_each_sched_entity(se) \ for (; se; se = NULL) |
bf0f6f24a sched: cfs core, ... |
215 |
|
b758149c0 sched: prepatory ... |
216 |
static inline struct cfs_rq *task_cfs_rq(struct task_struct *p) |
bf0f6f24a sched: cfs core, ... |
217 |
{ |
b758149c0 sched: prepatory ... |
218 |
return &task_rq(p)->cfs; |
bf0f6f24a sched: cfs core, ... |
219 |
} |
b758149c0 sched: prepatory ... |
220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 |
static inline struct cfs_rq *cfs_rq_of(struct sched_entity *se) { struct task_struct *p = task_of(se); struct rq *rq = task_rq(p); return &rq->cfs; } /* runqueue "owned" by this group */ static inline struct cfs_rq *group_cfs_rq(struct sched_entity *grp) { return NULL; } static inline struct cfs_rq *cpu_cfs_rq(struct cfs_rq *cfs_rq, int this_cpu) { return &cpu_rq(this_cpu)->cfs; } #define for_each_leaf_cfs_rq(rq, cfs_rq) \ for (cfs_rq = &rq->cfs; cfs_rq; cfs_rq = NULL) static inline int is_same_group(struct sched_entity *se, struct sched_entity *pse) { return 1; } static inline struct sched_entity *parent_entity(struct sched_entity *se) { return NULL; } |
464b75273 sched: re-instate... |
252 253 254 255 |
static inline void find_matching_se(struct sched_entity **se, struct sched_entity **pse) { } |
b758149c0 sched: prepatory ... |
256 |
#endif /* CONFIG_FAIR_GROUP_SCHED */ |
bf0f6f24a sched: cfs core, ... |
257 258 259 260 |
/************************************************************** * Scheduling class tree data structure manipulation methods: */ |
0702e3ebc sched: cleanup: f... |
261 |
static inline u64 max_vruntime(u64 min_vruntime, u64 vruntime) |
02e0431a3 sched: better min... |
262 |
{ |
368059a97 sched: max_vrunti... |
263 264 |
s64 delta = (s64)(vruntime - min_vruntime); if (delta > 0) |
02e0431a3 sched: better min... |
265 266 267 268 |
min_vruntime = vruntime; return min_vruntime; } |
0702e3ebc sched: cleanup: f... |
269 |
static inline u64 min_vruntime(u64 min_vruntime, u64 vruntime) |
b0ffd246e sched: clean up m... |
270 271 272 273 274 275 276 |
{ s64 delta = (s64)(vruntime - min_vruntime); if (delta < 0) min_vruntime = vruntime; return min_vruntime; } |
54fdc5816 sched: Account fo... |
277 278 279 280 281 |
static inline int entity_before(struct sched_entity *a, struct sched_entity *b) { return (s64)(a->vruntime - b->vruntime) < 0; } |
0702e3ebc sched: cleanup: f... |
282 |
static inline s64 entity_key(struct cfs_rq *cfs_rq, struct sched_entity *se) |
9014623c0 sched: handle vru... |
283 |
{ |
30cfdcfc5 sched: do not kee... |
284 |
return se->vruntime - cfs_rq->min_vruntime; |
9014623c0 sched: handle vru... |
285 |
} |
1af5f730f sched: more accur... |
286 287 288 289 290 291 292 293 294 295 296 |
static void update_min_vruntime(struct cfs_rq *cfs_rq) { u64 vruntime = cfs_rq->min_vruntime; if (cfs_rq->curr) vruntime = cfs_rq->curr->vruntime; if (cfs_rq->rb_leftmost) { struct sched_entity *se = rb_entry(cfs_rq->rb_leftmost, struct sched_entity, run_node); |
e17036dac sched: fix update... |
297 |
if (!cfs_rq->curr) |
1af5f730f sched: more accur... |
298 299 300 301 302 303 304 |
vruntime = se->vruntime; else vruntime = min_vruntime(vruntime, se->vruntime); } cfs_rq->min_vruntime = max_vruntime(cfs_rq->min_vruntime, vruntime); } |
bf0f6f24a sched: cfs core, ... |
305 306 307 |
/* * Enqueue an entity into the rb-tree: */ |
0702e3ebc sched: cleanup: f... |
308 |
static void __enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se) |
bf0f6f24a sched: cfs core, ... |
309 310 311 312 |
{ struct rb_node **link = &cfs_rq->tasks_timeline.rb_node; struct rb_node *parent = NULL; struct sched_entity *entry; |
9014623c0 sched: handle vru... |
313 |
s64 key = entity_key(cfs_rq, se); |
bf0f6f24a sched: cfs core, ... |
314 315 316 317 318 319 320 321 322 323 324 325 |
int leftmost = 1; /* * Find the right place in the rbtree: */ while (*link) { parent = *link; entry = rb_entry(parent, struct sched_entity, run_node); /* * We dont care about collisions. Nodes with * the same key stay together. */ |
9014623c0 sched: handle vru... |
326 |
if (key < entity_key(cfs_rq, entry)) { |
bf0f6f24a sched: cfs core, ... |
327 328 329 330 331 332 333 334 335 336 337 |
link = &parent->rb_left; } else { link = &parent->rb_right; leftmost = 0; } } /* * Maintain a cache of leftmost tree entries (it is frequently * used): */ |
1af5f730f sched: more accur... |
338 |
if (leftmost) |
57cb499df sched: remove set... |
339 |
cfs_rq->rb_leftmost = &se->run_node; |
bf0f6f24a sched: cfs core, ... |
340 341 342 |
rb_link_node(&se->run_node, parent, link); rb_insert_color(&se->run_node, &cfs_rq->tasks_timeline); |
bf0f6f24a sched: cfs core, ... |
343 |
} |
0702e3ebc sched: cleanup: f... |
344 |
static void __dequeue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se) |
bf0f6f24a sched: cfs core, ... |
345 |
{ |
3fe69747d sched: min_vrunti... |
346 347 |
if (cfs_rq->rb_leftmost == &se->run_node) { struct rb_node *next_node; |
3fe69747d sched: min_vrunti... |
348 349 350 |
next_node = rb_next(&se->run_node); cfs_rq->rb_leftmost = next_node; |
3fe69747d sched: min_vrunti... |
351 |
} |
e9acbff64 sched: introduce ... |
352 |
|
bf0f6f24a sched: cfs core, ... |
353 |
rb_erase(&se->run_node, &cfs_rq->tasks_timeline); |
bf0f6f24a sched: cfs core, ... |
354 |
} |
bf0f6f24a sched: cfs core, ... |
355 356 |
static struct sched_entity *__pick_next_entity(struct cfs_rq *cfs_rq) { |
f4b6755fb sched: cleanup fa... |
357 358 359 360 361 362 |
struct rb_node *left = cfs_rq->rb_leftmost; if (!left) return NULL; return rb_entry(left, struct sched_entity, run_node); |
bf0f6f24a sched: cfs core, ... |
363 |
} |
f4b6755fb sched: cleanup fa... |
364 |
static struct sched_entity *__pick_last_entity(struct cfs_rq *cfs_rq) |
aeb73b040 sched: clean up n... |
365 |
{ |
7eee3e677 sched: clean up _... |
366 |
struct rb_node *last = rb_last(&cfs_rq->tasks_timeline); |
aeb73b040 sched: clean up n... |
367 |
|
70eee74b7 sched: remove dup... |
368 369 |
if (!last) return NULL; |
7eee3e677 sched: clean up _... |
370 371 |
return rb_entry(last, struct sched_entity, run_node); |
aeb73b040 sched: clean up n... |
372 |
} |
bf0f6f24a sched: cfs core, ... |
373 374 375 |
/************************************************************** * Scheduling class statistics methods: */ |
b2be5e96d sched: reintroduc... |
376 |
#ifdef CONFIG_SCHED_DEBUG |
acb4a848d sched: Update nor... |
377 |
int sched_proc_update_handler(struct ctl_table *table, int write, |
8d65af789 sysctl: remove "s... |
378 |
void __user *buffer, size_t *lenp, |
b2be5e96d sched: reintroduc... |
379 380 |
loff_t *ppos) { |
8d65af789 sysctl: remove "s... |
381 |
int ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos); |
acb4a848d sched: Update nor... |
382 |
int factor = get_update_sysctl_factor(); |
b2be5e96d sched: reintroduc... |
383 384 385 386 387 388 |
if (ret || !write) return ret; sched_nr_latency = DIV_ROUND_UP(sysctl_sched_latency, sysctl_sched_min_granularity); |
acb4a848d sched: Update nor... |
389 390 391 392 393 394 395 |
#define WRT_SYSCTL(name) \ (normalized_sysctl_##name = sysctl_##name / (factor)) WRT_SYSCTL(sched_min_granularity); WRT_SYSCTL(sched_latency); WRT_SYSCTL(sched_wakeup_granularity); WRT_SYSCTL(sched_shares_ratelimit); #undef WRT_SYSCTL |
b2be5e96d sched: reintroduc... |
396 397 398 |
return 0; } #endif |
647e7cac2 sched: vslice fix... |
399 400 |
/* |
f9c0b0950 sched: revert bac... |
401 |
* delta /= w |
a7be37ac8 sched: revert the... |
402 403 404 405 |
*/ static inline unsigned long calc_delta_fair(unsigned long delta, struct sched_entity *se) { |
f9c0b0950 sched: revert bac... |
406 407 |
if (unlikely(se->load.weight != NICE_0_LOAD)) delta = calc_delta_mine(delta, NICE_0_LOAD, &se->load); |
a7be37ac8 sched: revert the... |
408 409 410 411 412 |
return delta; } /* |
647e7cac2 sched: vslice fix... |
413 414 415 416 417 418 419 |
* The idea is to set a period in which each task runs once. * * When there are too many tasks (sysctl_sched_nr_latency) we have to stretch * this period because otherwise the slices get too small. * * p = (nr <= nl) ? l : l*nr/nl */ |
4d78e7b65 sched: new task p... |
420 421 422 |
static u64 __sched_period(unsigned long nr_running) { u64 period = sysctl_sched_latency; |
b2be5e96d sched: reintroduc... |
423 |
unsigned long nr_latency = sched_nr_latency; |
4d78e7b65 sched: new task p... |
424 425 |
if (unlikely(nr_running > nr_latency)) { |
4bf0b7715 sched: remove do_... |
426 |
period = sysctl_sched_min_granularity; |
4d78e7b65 sched: new task p... |
427 |
period *= nr_running; |
4d78e7b65 sched: new task p... |
428 429 430 431 |
} return period; } |
647e7cac2 sched: vslice fix... |
432 433 434 435 |
/* * We calculate the wall-time slice from the period by taking a part * proportional to the weight. * |
f9c0b0950 sched: revert bac... |
436 |
* s = p*P[w/rw] |
647e7cac2 sched: vslice fix... |
437 |
*/ |
6d0f0ebd0 sched: simplify a... |
438 |
static u64 sched_slice(struct cfs_rq *cfs_rq, struct sched_entity *se) |
218050855 sched: adaptive s... |
439 |
{ |
0a582440f sched: fix sched_... |
440 |
u64 slice = __sched_period(cfs_rq->nr_running + !se->on_rq); |
f9c0b0950 sched: revert bac... |
441 |
|
0a582440f sched: fix sched_... |
442 |
for_each_sched_entity(se) { |
6272d68cc sched: sched_slic... |
443 |
struct load_weight *load; |
3104bf03a sched: Fix out of... |
444 |
struct load_weight lw; |
6272d68cc sched: sched_slic... |
445 446 447 |
cfs_rq = cfs_rq_of(se); load = &cfs_rq->load; |
f9c0b0950 sched: revert bac... |
448 |
|
0a582440f sched: fix sched_... |
449 |
if (unlikely(!se->on_rq)) { |
3104bf03a sched: Fix out of... |
450 |
lw = cfs_rq->load; |
0a582440f sched: fix sched_... |
451 452 453 454 455 456 457 |
update_load_add(&lw, se->load.weight); load = &lw; } slice = calc_delta_mine(slice, se->load.weight, load); } return slice; |
bf0f6f24a sched: cfs core, ... |
458 |
} |
647e7cac2 sched: vslice fix... |
459 |
/* |
ac884dec6 sched: fair-group... |
460 |
* We calculate the vruntime slice of a to be inserted task |
647e7cac2 sched: vslice fix... |
461 |
* |
f9c0b0950 sched: revert bac... |
462 |
* vs = s/w |
647e7cac2 sched: vslice fix... |
463 |
*/ |
f9c0b0950 sched: revert bac... |
464 |
static u64 sched_vslice(struct cfs_rq *cfs_rq, struct sched_entity *se) |
67e9fb2a3 sched: add vslice |
465 |
{ |
f9c0b0950 sched: revert bac... |
466 |
return calc_delta_fair(sched_slice(cfs_rq, se), se); |
a7be37ac8 sched: revert the... |
467 468 469 |
} /* |
bf0f6f24a sched: cfs core, ... |
470 471 472 473 |
* Update the current task's runtime statistics. Skip current tasks that * are not in our scheduling class. */ static inline void |
8ebc91d93 sched: remove sta... |
474 475 |
__update_curr(struct cfs_rq *cfs_rq, struct sched_entity *curr, unsigned long delta_exec) |
bf0f6f24a sched: cfs core, ... |
476 |
{ |
bbdba7c0e sched: remove wai... |
477 |
unsigned long delta_exec_weighted; |
bf0f6f24a sched: cfs core, ... |
478 |
|
41acab885 sched: Implement ... |
479 480 |
schedstat_set(curr->statistics.exec_max, max((u64)delta_exec, curr->statistics.exec_max)); |
bf0f6f24a sched: cfs core, ... |
481 482 |
curr->sum_exec_runtime += delta_exec; |
7a62eabc4 sched: debug: upd... |
483 |
schedstat_add(cfs_rq, exec_clock, delta_exec); |
a7be37ac8 sched: revert the... |
484 |
delta_exec_weighted = calc_delta_fair(delta_exec, curr); |
88ec22d3e sched: Remove the... |
485 |
|
e9acbff64 sched: introduce ... |
486 |
curr->vruntime += delta_exec_weighted; |
1af5f730f sched: more accur... |
487 |
update_min_vruntime(cfs_rq); |
bf0f6f24a sched: cfs core, ... |
488 |
} |
b7cc08965 sched: remove the... |
489 |
static void update_curr(struct cfs_rq *cfs_rq) |
bf0f6f24a sched: cfs core, ... |
490 |
{ |
429d43bcc sched: cleanup: s... |
491 |
struct sched_entity *curr = cfs_rq->curr; |
305e6835e sched: Do not acc... |
492 |
u64 now = rq_of(cfs_rq)->clock_task; |
bf0f6f24a sched: cfs core, ... |
493 494 495 496 497 498 499 500 501 502 |
unsigned long delta_exec; if (unlikely(!curr)) return; /* * Get the amount of time the current task was running * since the last time we changed load (this cannot * overflow on 32 bits): */ |
8ebc91d93 sched: remove sta... |
503 |
delta_exec = (unsigned long)(now - curr->exec_start); |
34f28ecd0 sched: optimize u... |
504 505 |
if (!delta_exec) return; |
bf0f6f24a sched: cfs core, ... |
506 |
|
8ebc91d93 sched: remove sta... |
507 508 |
__update_curr(cfs_rq, curr, delta_exec); curr->exec_start = now; |
d842de871 sched: cpu accoun... |
509 510 511 |
if (entity_is_task(curr)) { struct task_struct *curtask = task_of(curr); |
f977bb493 perf_counter, sch... |
512 |
trace_sched_stat_runtime(curtask, delta_exec, curr->vruntime); |
d842de871 sched: cpu accoun... |
513 |
cpuacct_charge(curtask, delta_exec); |
f06febc96 timers: fix itime... |
514 |
account_group_exec_runtime(curtask, delta_exec); |
d842de871 sched: cpu accoun... |
515 |
} |
bf0f6f24a sched: cfs core, ... |
516 517 518 |
} static inline void |
5870db5b8 sched: remove the... |
519 |
update_stats_wait_start(struct cfs_rq *cfs_rq, struct sched_entity *se) |
bf0f6f24a sched: cfs core, ... |
520 |
{ |
41acab885 sched: Implement ... |
521 |
schedstat_set(se->statistics.wait_start, rq_of(cfs_rq)->clock); |
bf0f6f24a sched: cfs core, ... |
522 |
} |
bf0f6f24a sched: cfs core, ... |
523 524 525 |
/* * Task is being enqueued - update stats: */ |
d2417e5a3 sched: remove the... |
526 |
static void update_stats_enqueue(struct cfs_rq *cfs_rq, struct sched_entity *se) |
bf0f6f24a sched: cfs core, ... |
527 |
{ |
bf0f6f24a sched: cfs core, ... |
528 529 530 531 |
/* * Are we enqueueing a waiting task? (for current tasks * a dequeue/enqueue event is a NOP) */ |
429d43bcc sched: cleanup: s... |
532 |
if (se != cfs_rq->curr) |
5870db5b8 sched: remove the... |
533 |
update_stats_wait_start(cfs_rq, se); |
bf0f6f24a sched: cfs core, ... |
534 |
} |
bf0f6f24a sched: cfs core, ... |
535 |
static void |
9ef0a9615 sched: remove the... |
536 |
update_stats_wait_end(struct cfs_rq *cfs_rq, struct sched_entity *se) |
bf0f6f24a sched: cfs core, ... |
537 |
{ |
41acab885 sched: Implement ... |
538 539 540 541 542 |
schedstat_set(se->statistics.wait_max, max(se->statistics.wait_max, rq_of(cfs_rq)->clock - se->statistics.wait_start)); schedstat_set(se->statistics.wait_count, se->statistics.wait_count + 1); schedstat_set(se->statistics.wait_sum, se->statistics.wait_sum + rq_of(cfs_rq)->clock - se->statistics.wait_start); |
768d0c272 sched: Add wait, ... |
543 544 545 |
#ifdef CONFIG_SCHEDSTATS if (entity_is_task(se)) { trace_sched_stat_wait(task_of(se), |
41acab885 sched: Implement ... |
546 |
rq_of(cfs_rq)->clock - se->statistics.wait_start); |
768d0c272 sched: Add wait, ... |
547 548 |
} #endif |
41acab885 sched: Implement ... |
549 |
schedstat_set(se->statistics.wait_start, 0); |
bf0f6f24a sched: cfs core, ... |
550 551 552 |
} static inline void |
19b6a2e37 sched: remove the... |
553 |
update_stats_dequeue(struct cfs_rq *cfs_rq, struct sched_entity *se) |
bf0f6f24a sched: cfs core, ... |
554 |
{ |
bf0f6f24a sched: cfs core, ... |
555 556 557 558 |
/* * Mark the end of the wait period if dequeueing a * waiting task: */ |
429d43bcc sched: cleanup: s... |
559 |
if (se != cfs_rq->curr) |
9ef0a9615 sched: remove the... |
560 |
update_stats_wait_end(cfs_rq, se); |
bf0f6f24a sched: cfs core, ... |
561 562 563 564 565 566 |
} /* * We are picking a new current task - update its stats: */ static inline void |
79303e9e0 sched: remove the... |
567 |
update_stats_curr_start(struct cfs_rq *cfs_rq, struct sched_entity *se) |
bf0f6f24a sched: cfs core, ... |
568 569 570 571 |
{ /* * We are starting a new run period: */ |
305e6835e sched: Do not acc... |
572 |
se->exec_start = rq_of(cfs_rq)->clock_task; |
bf0f6f24a sched: cfs core, ... |
573 |
} |
bf0f6f24a sched: cfs core, ... |
574 575 576 |
/************************************************** * Scheduling class queueing methods: */ |
c09595f63 sched: revert rev... |
577 578 579 580 581 582 583 584 585 586 587 588 |
#if defined CONFIG_SMP && defined CONFIG_FAIR_GROUP_SCHED static void add_cfs_task_weight(struct cfs_rq *cfs_rq, unsigned long weight) { cfs_rq->task_weight += weight; } #else static inline void add_cfs_task_weight(struct cfs_rq *cfs_rq, unsigned long weight) { } #endif |
30cfdcfc5 sched: do not kee... |
589 590 591 592 |
static void account_entity_enqueue(struct cfs_rq *cfs_rq, struct sched_entity *se) { update_load_add(&cfs_rq->load, se->load.weight); |
c09595f63 sched: revert rev... |
593 594 |
if (!parent_entity(se)) inc_cpu_load(rq_of(cfs_rq), se->load.weight); |
b87f17242 sched: maintain o... |
595 |
if (entity_is_task(se)) { |
c09595f63 sched: revert rev... |
596 |
add_cfs_task_weight(cfs_rq, se->load.weight); |
b87f17242 sched: maintain o... |
597 598 |
list_add(&se->group_node, &cfs_rq->tasks); } |
30cfdcfc5 sched: do not kee... |
599 600 601 602 603 604 605 606 |
cfs_rq->nr_running++; se->on_rq = 1; } static void account_entity_dequeue(struct cfs_rq *cfs_rq, struct sched_entity *se) { update_load_sub(&cfs_rq->load, se->load.weight); |
c09595f63 sched: revert rev... |
607 608 |
if (!parent_entity(se)) dec_cpu_load(rq_of(cfs_rq), se->load.weight); |
b87f17242 sched: maintain o... |
609 |
if (entity_is_task(se)) { |
c09595f63 sched: revert rev... |
610 |
add_cfs_task_weight(cfs_rq, -se->load.weight); |
b87f17242 sched: maintain o... |
611 612 |
list_del_init(&se->group_node); } |
30cfdcfc5 sched: do not kee... |
613 614 615 |
cfs_rq->nr_running--; se->on_rq = 0; } |
2396af69b sched: remove the... |
616 |
static void enqueue_sleeper(struct cfs_rq *cfs_rq, struct sched_entity *se) |
bf0f6f24a sched: cfs core, ... |
617 |
{ |
bf0f6f24a sched: cfs core, ... |
618 |
#ifdef CONFIG_SCHEDSTATS |
e414314cc sched: Fix latenc... |
619 620 621 622 |
struct task_struct *tsk = NULL; if (entity_is_task(se)) tsk = task_of(se); |
41acab885 sched: Implement ... |
623 624 |
if (se->statistics.sleep_start) { u64 delta = rq_of(cfs_rq)->clock - se->statistics.sleep_start; |
bf0f6f24a sched: cfs core, ... |
625 626 627 |
if ((s64)delta < 0) delta = 0; |
41acab885 sched: Implement ... |
628 629 |
if (unlikely(delta > se->statistics.sleep_max)) se->statistics.sleep_max = delta; |
bf0f6f24a sched: cfs core, ... |
630 |
|
41acab885 sched: Implement ... |
631 632 |
se->statistics.sleep_start = 0; se->statistics.sum_sleep_runtime += delta; |
9745512ce sched: latencytop... |
633 |
|
768d0c272 sched: Add wait, ... |
634 |
if (tsk) { |
e414314cc sched: Fix latenc... |
635 |
account_scheduler_latency(tsk, delta >> 10, 1); |
768d0c272 sched: Add wait, ... |
636 637 |
trace_sched_stat_sleep(tsk, delta); } |
bf0f6f24a sched: cfs core, ... |
638 |
} |
41acab885 sched: Implement ... |
639 640 |
if (se->statistics.block_start) { u64 delta = rq_of(cfs_rq)->clock - se->statistics.block_start; |
bf0f6f24a sched: cfs core, ... |
641 642 643 |
if ((s64)delta < 0) delta = 0; |
41acab885 sched: Implement ... |
644 645 |
if (unlikely(delta > se->statistics.block_max)) se->statistics.block_max = delta; |
bf0f6f24a sched: cfs core, ... |
646 |
|
41acab885 sched: Implement ... |
647 648 |
se->statistics.block_start = 0; se->statistics.sum_sleep_runtime += delta; |
30084fbd1 sched: fix profil... |
649 |
|
e414314cc sched: Fix latenc... |
650 |
if (tsk) { |
8f0dfc34e sched: Provide io... |
651 |
if (tsk->in_iowait) { |
41acab885 sched: Implement ... |
652 653 |
se->statistics.iowait_sum += delta; se->statistics.iowait_count++; |
768d0c272 sched: Add wait, ... |
654 |
trace_sched_stat_iowait(tsk, delta); |
8f0dfc34e sched: Provide io... |
655 |
} |
e414314cc sched: Fix latenc... |
656 657 658 659 660 661 662 663 664 665 666 |
/* * Blocking time is in units of nanosecs, so shift by * 20 to get a milliseconds-range estimation of the * amount of time that the task spent sleeping: */ if (unlikely(prof_on == SLEEP_PROFILING)) { profile_hits(SLEEP_PROFILING, (void *)get_wchan(tsk), delta >> 20); } account_scheduler_latency(tsk, delta >> 10, 0); |
30084fbd1 sched: fix profil... |
667 |
} |
bf0f6f24a sched: cfs core, ... |
668 669 670 |
} #endif } |
ddc972975 sched debug: chec... |
671 672 673 674 675 676 677 678 679 680 681 682 |
static void check_spread(struct cfs_rq *cfs_rq, struct sched_entity *se) { #ifdef CONFIG_SCHED_DEBUG s64 d = se->vruntime - cfs_rq->min_vruntime; if (d < 0) d = -d; if (d > 3*sysctl_sched_latency) schedstat_inc(cfs_rq, nr_spread_over); #endif } |
bf0f6f24a sched: cfs core, ... |
683 |
static void |
aeb73b040 sched: clean up n... |
684 685 |
place_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int initial) { |
1af5f730f sched: more accur... |
686 |
u64 vruntime = cfs_rq->min_vruntime; |
94dfb5e75 sched: add tree b... |
687 |
|
2cb8600e6 sched: documentat... |
688 689 690 691 692 693 |
/* * The 'current' period is already promised to the current tasks, * however the extra weight of the new task will slow them down a * little, place the new task so that it fits in the slot that * stays open at the end. */ |
94dfb5e75 sched: add tree b... |
694 |
if (initial && sched_feat(START_DEBIT)) |
f9c0b0950 sched: revert bac... |
695 |
vruntime += sched_vslice(cfs_rq, se); |
aeb73b040 sched: clean up n... |
696 |
|
a2e7a7eb2 sched: Remove unn... |
697 |
/* sleeps up to a single latency don't count. */ |
5ca9880c6 sched: Remove FAI... |
698 |
if (!initial) { |
a2e7a7eb2 sched: Remove unn... |
699 |
unsigned long thresh = sysctl_sched_latency; |
a7be37ac8 sched: revert the... |
700 |
|
a2e7a7eb2 sched: Remove unn... |
701 |
/* |
a2e7a7eb2 sched: Remove unn... |
702 703 704 705 706 |
* Halve their sleep time's effect, to allow * for a gentler effect of sleepers: */ if (sched_feat(GENTLE_FAIR_SLEEPERS)) thresh >>= 1; |
51e0304ce sched: Implement ... |
707 |
|
a2e7a7eb2 sched: Remove unn... |
708 |
vruntime -= thresh; |
aeb73b040 sched: clean up n... |
709 |
} |
b5d9d734a sched: Ensure tha... |
710 711 |
/* ensure we never gain time by being placed backwards. */ vruntime = max_vruntime(se->vruntime, vruntime); |
67e9fb2a3 sched: add vslice |
712 |
se->vruntime = vruntime; |
aeb73b040 sched: clean up n... |
713 714 715 |
} static void |
88ec22d3e sched: Remove the... |
716 |
enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags) |
bf0f6f24a sched: cfs core, ... |
717 718 |
{ /* |
88ec22d3e sched: Remove the... |
719 720 721 |
* Update the normalized vruntime before updating min_vruntime * through callig update_curr(). */ |
371fd7e7a sched: Add enqueu... |
722 |
if (!(flags & ENQUEUE_WAKEUP) || (flags & ENQUEUE_WAKING)) |
88ec22d3e sched: Remove the... |
723 724 725 |
se->vruntime += cfs_rq->min_vruntime; /* |
a2a2d6807 sched: cleanup, m... |
726 |
* Update run-time statistics of the 'current'. |
bf0f6f24a sched: cfs core, ... |
727 |
*/ |
b7cc08965 sched: remove the... |
728 |
update_curr(cfs_rq); |
a992241de sched: fix normal... |
729 |
account_entity_enqueue(cfs_rq, se); |
bf0f6f24a sched: cfs core, ... |
730 |
|
88ec22d3e sched: Remove the... |
731 |
if (flags & ENQUEUE_WAKEUP) { |
aeb73b040 sched: clean up n... |
732 |
place_entity(cfs_rq, se, 0); |
2396af69b sched: remove the... |
733 |
enqueue_sleeper(cfs_rq, se); |
e9acbff64 sched: introduce ... |
734 |
} |
bf0f6f24a sched: cfs core, ... |
735 |
|
d2417e5a3 sched: remove the... |
736 |
update_stats_enqueue(cfs_rq, se); |
ddc972975 sched debug: chec... |
737 |
check_spread(cfs_rq, se); |
83b699ed2 sched: revert rec... |
738 739 |
if (se != cfs_rq->curr) __enqueue_entity(cfs_rq, se); |
bf0f6f24a sched: cfs core, ... |
740 |
} |
a571bbeaf sched: fix buddie... |
741 |
static void __clear_buddies(struct cfs_rq *cfs_rq, struct sched_entity *se) |
2002c6959 sched: release bu... |
742 |
{ |
de69a80be sched: Stop buddi... |
743 |
if (!se || cfs_rq->last == se) |
2002c6959 sched: release bu... |
744 |
cfs_rq->last = NULL; |
de69a80be sched: Stop buddi... |
745 |
if (!se || cfs_rq->next == se) |
2002c6959 sched: release bu... |
746 747 |
cfs_rq->next = NULL; } |
a571bbeaf sched: fix buddie... |
748 749 750 751 752 |
static void clear_buddies(struct cfs_rq *cfs_rq, struct sched_entity *se) { for_each_sched_entity(se) __clear_buddies(cfs_rq_of(se), se); } |
bf0f6f24a sched: cfs core, ... |
753 |
static void |
371fd7e7a sched: Add enqueu... |
754 |
dequeue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags) |
bf0f6f24a sched: cfs core, ... |
755 |
{ |
a2a2d6807 sched: cleanup, m... |
756 757 758 759 |
/* * Update run-time statistics of the 'current'. */ update_curr(cfs_rq); |
19b6a2e37 sched: remove the... |
760 |
update_stats_dequeue(cfs_rq, se); |
371fd7e7a sched: Add enqueu... |
761 |
if (flags & DEQUEUE_SLEEP) { |
67e9fb2a3 sched: add vslice |
762 |
#ifdef CONFIG_SCHEDSTATS |
bf0f6f24a sched: cfs core, ... |
763 764 765 766 |
if (entity_is_task(se)) { struct task_struct *tsk = task_of(se); if (tsk->state & TASK_INTERRUPTIBLE) |
41acab885 sched: Implement ... |
767 |
se->statistics.sleep_start = rq_of(cfs_rq)->clock; |
bf0f6f24a sched: cfs core, ... |
768 |
if (tsk->state & TASK_UNINTERRUPTIBLE) |
41acab885 sched: Implement ... |
769 |
se->statistics.block_start = rq_of(cfs_rq)->clock; |
bf0f6f24a sched: cfs core, ... |
770 |
} |
db36cc7d6 sched: clean up s... |
771 |
#endif |
67e9fb2a3 sched: add vslice |
772 |
} |
2002c6959 sched: release bu... |
773 |
clear_buddies(cfs_rq, se); |
4793241be sched: backward l... |
774 |
|
83b699ed2 sched: revert rec... |
775 |
if (se != cfs_rq->curr) |
30cfdcfc5 sched: do not kee... |
776 777 |
__dequeue_entity(cfs_rq, se); account_entity_dequeue(cfs_rq, se); |
1af5f730f sched: more accur... |
778 |
update_min_vruntime(cfs_rq); |
88ec22d3e sched: Remove the... |
779 780 781 782 783 784 |
/* * Normalize the entity after updating the min_vruntime because the * update can refer to the ->curr item and we need to reflect this * movement in our normalized position. */ |
371fd7e7a sched: Add enqueu... |
785 |
if (!(flags & DEQUEUE_SLEEP)) |
88ec22d3e sched: Remove the... |
786 |
se->vruntime -= cfs_rq->min_vruntime; |
bf0f6f24a sched: cfs core, ... |
787 788 789 790 791 |
} /* * Preempt the current task with a newly woken task if needed: */ |
7c92e54f6 sched: simplify _... |
792 |
static void |
2e09bf556 sched: wakeup gra... |
793 |
check_preempt_tick(struct cfs_rq *cfs_rq, struct sched_entity *curr) |
bf0f6f24a sched: cfs core, ... |
794 |
{ |
116978308 sched: fix ideal_... |
795 |
unsigned long ideal_runtime, delta_exec; |
6d0f0ebd0 sched: simplify a... |
796 |
ideal_runtime = sched_slice(cfs_rq, curr); |
116978308 sched: fix ideal_... |
797 |
delta_exec = curr->sum_exec_runtime - curr->prev_sum_exec_runtime; |
a9f3e2b54 sched: clear budd... |
798 |
if (delta_exec > ideal_runtime) { |
bf0f6f24a sched: cfs core, ... |
799 |
resched_task(rq_of(cfs_rq)->curr); |
a9f3e2b54 sched: clear budd... |
800 801 802 803 804 |
/* * The current task ran long enough, ensure it doesn't get * re-elected due to buddy favours. */ clear_buddies(cfs_rq, curr); |
f685ceaca sched: Strengthen... |
805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 |
return; } /* * Ensure that a task that missed wakeup preemption by a * narrow margin doesn't have to wait for a full slice. * This also mitigates buddy induced latencies under load. */ if (!sched_feat(WAKEUP_PREEMPT)) return; if (delta_exec < sysctl_sched_min_granularity) return; if (cfs_rq->nr_running > 1) { struct sched_entity *se = __pick_next_entity(cfs_rq); s64 delta = curr->vruntime - se->vruntime; if (delta > ideal_runtime) resched_task(rq_of(cfs_rq)->curr); |
a9f3e2b54 sched: clear budd... |
825 |
} |
bf0f6f24a sched: cfs core, ... |
826 |
} |
83b699ed2 sched: revert rec... |
827 |
static void |
8494f412e sched: remove the... |
828 |
set_next_entity(struct cfs_rq *cfs_rq, struct sched_entity *se) |
bf0f6f24a sched: cfs core, ... |
829 |
{ |
83b699ed2 sched: revert rec... |
830 831 832 833 834 835 836 837 838 839 |
/* 'current' is not kept within the tree. */ if (se->on_rq) { /* * Any task has to be enqueued before it get to execute on * a CPU. So account for the time it spent waiting on the * runqueue. */ update_stats_wait_end(cfs_rq, se); __dequeue_entity(cfs_rq, se); } |
79303e9e0 sched: remove the... |
840 |
update_stats_curr_start(cfs_rq, se); |
429d43bcc sched: cleanup: s... |
841 |
cfs_rq->curr = se; |
eba1ed4b7 sched: debug: tra... |
842 843 844 845 846 847 |
#ifdef CONFIG_SCHEDSTATS /* * Track our maximum slice length, if the CPU's load is at * least twice that of our own weight (i.e. dont track it * when there are only lesser-weight tasks around): */ |
495eca494 sched: clean up s... |
848 |
if (rq_of(cfs_rq)->load.weight >= 2*se->load.weight) { |
41acab885 sched: Implement ... |
849 |
se->statistics.slice_max = max(se->statistics.slice_max, |
eba1ed4b7 sched: debug: tra... |
850 851 852 |
se->sum_exec_runtime - se->prev_sum_exec_runtime); } #endif |
4a55b4503 sched: improve pr... |
853 |
se->prev_sum_exec_runtime = se->sum_exec_runtime; |
bf0f6f24a sched: cfs core, ... |
854 |
} |
3f3a49048 sched: virtual ti... |
855 856 |
static int wakeup_preempt_entity(struct sched_entity *curr, struct sched_entity *se); |
f4b6755fb sched: cleanup fa... |
857 |
static struct sched_entity *pick_next_entity(struct cfs_rq *cfs_rq) |
aa2ac2522 sched: fix overlo... |
858 |
{ |
f4b6755fb sched: cleanup fa... |
859 |
struct sched_entity *se = __pick_next_entity(cfs_rq); |
f685ceaca sched: Strengthen... |
860 |
struct sched_entity *left = se; |
f4b6755fb sched: cleanup fa... |
861 |
|
f685ceaca sched: Strengthen... |
862 863 |
if (cfs_rq->next && wakeup_preempt_entity(cfs_rq->next, left) < 1) se = cfs_rq->next; |
aa2ac2522 sched: fix overlo... |
864 |
|
f685ceaca sched: Strengthen... |
865 866 867 868 869 870 871 |
/* * Prefer last buddy, try to return the CPU to a preempted task. */ if (cfs_rq->last && wakeup_preempt_entity(cfs_rq->last, left) < 1) se = cfs_rq->last; clear_buddies(cfs_rq, se); |
4793241be sched: backward l... |
872 873 |
return se; |
aa2ac2522 sched: fix overlo... |
874 |
} |
ab6cde269 sched: remove the... |
875 |
static void put_prev_entity(struct cfs_rq *cfs_rq, struct sched_entity *prev) |
bf0f6f24a sched: cfs core, ... |
876 877 878 879 880 881 |
{ /* * If still on the runqueue then deactivate_task() * was not called and update_curr() has to be done: */ if (prev->on_rq) |
b7cc08965 sched: remove the... |
882 |
update_curr(cfs_rq); |
bf0f6f24a sched: cfs core, ... |
883 |
|
ddc972975 sched debug: chec... |
884 |
check_spread(cfs_rq, prev); |
30cfdcfc5 sched: do not kee... |
885 |
if (prev->on_rq) { |
5870db5b8 sched: remove the... |
886 |
update_stats_wait_start(cfs_rq, prev); |
30cfdcfc5 sched: do not kee... |
887 888 889 |
/* Put 'current' back into the tree. */ __enqueue_entity(cfs_rq, prev); } |
429d43bcc sched: cleanup: s... |
890 |
cfs_rq->curr = NULL; |
bf0f6f24a sched: cfs core, ... |
891 |
} |
8f4d37ec0 sched: high-res p... |
892 893 |
static void entity_tick(struct cfs_rq *cfs_rq, struct sched_entity *curr, int queued) |
bf0f6f24a sched: cfs core, ... |
894 |
{ |
bf0f6f24a sched: cfs core, ... |
895 |
/* |
30cfdcfc5 sched: do not kee... |
896 |
* Update run-time statistics of the 'current'. |
bf0f6f24a sched: cfs core, ... |
897 |
*/ |
30cfdcfc5 sched: do not kee... |
898 |
update_curr(cfs_rq); |
bf0f6f24a sched: cfs core, ... |
899 |
|
8f4d37ec0 sched: high-res p... |
900 901 902 903 904 |
#ifdef CONFIG_SCHED_HRTICK /* * queued ticks are scheduled to match the slice, so don't bother * validating it and just reschedule. */ |
983ed7a66 sched: add static... |
905 906 907 908 |
if (queued) { resched_task(rq_of(cfs_rq)->curr); return; } |
8f4d37ec0 sched: high-res p... |
909 910 911 912 913 914 915 |
/* * don't let the period tick interfere with the hrtick preemption */ if (!sched_feat(DOUBLE_TICK) && hrtimer_active(&rq_of(cfs_rq)->hrtick_timer)) return; #endif |
ce6c13113 sched: disable fo... |
916 |
if (cfs_rq->nr_running > 1 || !sched_feat(WAKEUP_PREEMPT)) |
2e09bf556 sched: wakeup gra... |
917 |
check_preempt_tick(cfs_rq, curr); |
bf0f6f24a sched: cfs core, ... |
918 919 920 921 922 |
} /************************************************** * CFS operations on tasks: */ |
8f4d37ec0 sched: high-res p... |
923 924 925 |
#ifdef CONFIG_SCHED_HRTICK static void hrtick_start_fair(struct rq *rq, struct task_struct *p) { |
8f4d37ec0 sched: high-res p... |
926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 |
struct sched_entity *se = &p->se; struct cfs_rq *cfs_rq = cfs_rq_of(se); WARN_ON(task_rq(p) != rq); if (hrtick_enabled(rq) && cfs_rq->nr_running > 1) { u64 slice = sched_slice(cfs_rq, se); u64 ran = se->sum_exec_runtime - se->prev_sum_exec_runtime; s64 delta = slice - ran; if (delta < 0) { if (rq->curr == p) resched_task(p); return; } /* * Don't schedule slices shorter than 10000ns, that just * doesn't make sense. Rely on vruntime for fairness. */ |
31656519e sched, x86: clean... |
946 |
if (rq->curr != p) |
157124c11 sched: fix warnin... |
947 |
delta = max_t(s64, 10000LL, delta); |
8f4d37ec0 sched: high-res p... |
948 |
|
31656519e sched, x86: clean... |
949 |
hrtick_start(rq, delta); |
8f4d37ec0 sched: high-res p... |
950 951 |
} } |
a4c2f00f5 sched: fair sched... |
952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 |
/* * called from enqueue/dequeue and updates the hrtick when the * current task is from our class and nr_running is low enough * to matter. */ static void hrtick_update(struct rq *rq) { struct task_struct *curr = rq->curr; if (curr->sched_class != &fair_sched_class) return; if (cfs_rq_of(&curr->se)->nr_running < sched_nr_latency) hrtick_start_fair(rq, curr); } |
55e12e5e7 sched: make sched... |
968 |
#else /* !CONFIG_SCHED_HRTICK */ |
8f4d37ec0 sched: high-res p... |
969 970 971 972 |
static inline void hrtick_start_fair(struct rq *rq, struct task_struct *p) { } |
a4c2f00f5 sched: fair sched... |
973 974 975 976 |
static inline void hrtick_update(struct rq *rq) { } |
8f4d37ec0 sched: high-res p... |
977 |
#endif |
bf0f6f24a sched: cfs core, ... |
978 979 980 981 982 |
/* * The enqueue_task method is called before nr_running is * increased. Here we update the fair scheduling stats and * then put the task into the rbtree: */ |
ea87bb785 sched: Extend enq... |
983 |
static void |
371fd7e7a sched: Add enqueu... |
984 |
enqueue_task_fair(struct rq *rq, struct task_struct *p, int flags) |
bf0f6f24a sched: cfs core, ... |
985 986 |
{ struct cfs_rq *cfs_rq; |
62fb18513 sched: revert loa... |
987 |
struct sched_entity *se = &p->se; |
bf0f6f24a sched: cfs core, ... |
988 989 |
for_each_sched_entity(se) { |
62fb18513 sched: revert loa... |
990 |
if (se->on_rq) |
bf0f6f24a sched: cfs core, ... |
991 992 |
break; cfs_rq = cfs_rq_of(se); |
88ec22d3e sched: Remove the... |
993 994 |
enqueue_entity(cfs_rq, se, flags); flags = ENQUEUE_WAKEUP; |
bf0f6f24a sched: cfs core, ... |
995 |
} |
8f4d37ec0 sched: high-res p... |
996 |
|
a4c2f00f5 sched: fair sched... |
997 |
hrtick_update(rq); |
bf0f6f24a sched: cfs core, ... |
998 999 1000 1001 1002 1003 1004 |
} /* * The dequeue_task method is called before nr_running is * decreased. We remove the task from the rbtree and * update the fair scheduling stats: */ |
371fd7e7a sched: Add enqueu... |
1005 |
static void dequeue_task_fair(struct rq *rq, struct task_struct *p, int flags) |
bf0f6f24a sched: cfs core, ... |
1006 1007 |
{ struct cfs_rq *cfs_rq; |
62fb18513 sched: revert loa... |
1008 |
struct sched_entity *se = &p->se; |
bf0f6f24a sched: cfs core, ... |
1009 1010 1011 |
for_each_sched_entity(se) { cfs_rq = cfs_rq_of(se); |
371fd7e7a sched: Add enqueu... |
1012 |
dequeue_entity(cfs_rq, se, flags); |
bf0f6f24a sched: cfs core, ... |
1013 |
/* Don't dequeue parent if it has other entities besides us */ |
62fb18513 sched: revert loa... |
1014 |
if (cfs_rq->load.weight) |
bf0f6f24a sched: cfs core, ... |
1015 |
break; |
371fd7e7a sched: Add enqueu... |
1016 |
flags |= DEQUEUE_SLEEP; |
bf0f6f24a sched: cfs core, ... |
1017 |
} |
8f4d37ec0 sched: high-res p... |
1018 |
|
a4c2f00f5 sched: fair sched... |
1019 |
hrtick_update(rq); |
bf0f6f24a sched: cfs core, ... |
1020 1021 1022 |
} /* |
1799e35d5 sched: add /proc/... |
1023 1024 1025 |
* sched_yield() support is very simple - we dequeue and enqueue. * * If compat_yield is turned on then we requeue to the end of the tree. |
bf0f6f24a sched: cfs core, ... |
1026 |
*/ |
4530d7ab0 sched: simplify s... |
1027 |
static void yield_task_fair(struct rq *rq) |
bf0f6f24a sched: cfs core, ... |
1028 |
{ |
db292ca30 sched: default to... |
1029 1030 1031 |
struct task_struct *curr = rq->curr; struct cfs_rq *cfs_rq = task_cfs_rq(curr); struct sched_entity *rightmost, *se = &curr->se; |
bf0f6f24a sched: cfs core, ... |
1032 1033 |
/* |
1799e35d5 sched: add /proc/... |
1034 1035 1036 1037 |
* Are we the only task in the tree? */ if (unlikely(cfs_rq->nr_running == 1)) return; |
2002c6959 sched: release bu... |
1038 |
clear_buddies(cfs_rq, se); |
db292ca30 sched: default to... |
1039 |
if (likely(!sysctl_sched_compat_yield) && curr->policy != SCHED_BATCH) { |
3e51f33fc sched: add option... |
1040 |
update_rq_clock(rq); |
1799e35d5 sched: add /proc/... |
1041 |
/* |
a2a2d6807 sched: cleanup, m... |
1042 |
* Update run-time statistics of the 'current'. |
1799e35d5 sched: add /proc/... |
1043 |
*/ |
2b1e315dd sched: yield fix |
1044 |
update_curr(cfs_rq); |
1799e35d5 sched: add /proc/... |
1045 1046 1047 1048 1049 |
return; } /* * Find the rightmost entry in the rbtree: |
bf0f6f24a sched: cfs core, ... |
1050 |
*/ |
2b1e315dd sched: yield fix |
1051 |
rightmost = __pick_last_entity(cfs_rq); |
1799e35d5 sched: add /proc/... |
1052 1053 1054 |
/* * Already in the rightmost position? */ |
54fdc5816 sched: Account fo... |
1055 |
if (unlikely(!rightmost || entity_before(rightmost, se))) |
1799e35d5 sched: add /proc/... |
1056 1057 1058 1059 |
return; /* * Minimally necessary key value to be last in the tree: |
2b1e315dd sched: yield fix |
1060 1061 |
* Upon rescheduling, sched_class::put_prev_task() will place * 'current' within the tree based on its new key value. |
1799e35d5 sched: add /proc/... |
1062 |
*/ |
30cfdcfc5 sched: do not kee... |
1063 |
se->vruntime = rightmost->vruntime + 1; |
bf0f6f24a sched: cfs core, ... |
1064 |
} |
e7693a362 sched: de-SCHED_O... |
1065 |
#ifdef CONFIG_SMP |
098fb9db2 sched: clean up w... |
1066 |
|
88ec22d3e sched: Remove the... |
1067 1068 1069 1070 1071 1072 1073 |
static void task_waking_fair(struct rq *rq, struct task_struct *p) { struct sched_entity *se = &p->se; struct cfs_rq *cfs_rq = cfs_rq_of(se); se->vruntime -= cfs_rq->min_vruntime; } |
bb3469ac9 sched: hierarchic... |
1074 |
#ifdef CONFIG_FAIR_GROUP_SCHED |
f5bfb7d9f sched: bias effec... |
1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 |
/* * effective_load() calculates the load change as seen from the root_task_group * * Adding load to a group doesn't make a group heavier, but can cause movement * of group shares between cpus. Assuming the shares were perfectly aligned one * can calculate the shift in shares. * * The problem is that perfectly aligning the shares is rather expensive, hence * we try to avoid doing that too often - see update_shares(), which ratelimits * this change. * * We compensate this by not only taking the current delta into account, but * also considering the delta between when the shares were last adjusted and * now. * * We still saw a performance dip, some tracing learned us that between * cgroup:/ and cgroup:/foo balancing the number of affine wakeups increased * significantly. Therefore try to bias the error in direction of failing * the affine wakeup. * */ |
f1d239f73 sched: incrementa... |
1096 1097 |
static long effective_load(struct task_group *tg, int cpu, long wl, long wg) |
bb3469ac9 sched: hierarchic... |
1098 |
{ |
4be9daaa1 sched: fix task_h... |
1099 |
struct sched_entity *se = tg->se[cpu]; |
f1d239f73 sched: incrementa... |
1100 1101 1102 1103 1104 |
if (!tg->parent) return wl; /* |
f5bfb7d9f sched: bias effec... |
1105 1106 1107 1108 1109 |
* By not taking the decrease of shares on the other cpu into * account our error leans towards reducing the affine wakeups. */ if (!wl && sched_feat(ASYM_EFF_LOAD)) return wl; |
4be9daaa1 sched: fix task_h... |
1110 |
for_each_sched_entity(se) { |
cb5ef42a0 sched: optimize e... |
1111 |
long S, rw, s, a, b; |
940959e93 sched: fixlet for... |
1112 1113 1114 1115 1116 1117 1118 1119 1120 |
long more_w; /* * Instead of using this increment, also add the difference * between when the shares were last updated and now. */ more_w = se->my_q->load.weight - se->my_q->rq_weight; wl += more_w; wg += more_w; |
4be9daaa1 sched: fix task_h... |
1121 1122 1123 |
S = se->my_q->tg->shares; s = se->my_q->shares; |
f1d239f73 sched: incrementa... |
1124 |
rw = se->my_q->rq_weight; |
bb3469ac9 sched: hierarchic... |
1125 |
|
cb5ef42a0 sched: optimize e... |
1126 1127 |
a = S*(rw + wl); b = S*rw + s*wg; |
4be9daaa1 sched: fix task_h... |
1128 |
|
940959e93 sched: fixlet for... |
1129 1130 1131 1132 |
wl = s*(a-b); if (likely(b)) wl /= b; |
83378269a sched: correct wa... |
1133 1134 1135 1136 1137 1138 1139 |
/* * Assume the group is already running and will * thus already be accounted for in the weight. * * That is, moving shares between CPUs, does not * alter the group weight. */ |
4be9daaa1 sched: fix task_h... |
1140 |
wg = 0; |
4be9daaa1 sched: fix task_h... |
1141 |
} |
bb3469ac9 sched: hierarchic... |
1142 |
|
4be9daaa1 sched: fix task_h... |
1143 |
return wl; |
bb3469ac9 sched: hierarchic... |
1144 |
} |
4be9daaa1 sched: fix task_h... |
1145 |
|
bb3469ac9 sched: hierarchic... |
1146 |
#else |
4be9daaa1 sched: fix task_h... |
1147 |
|
83378269a sched: correct wa... |
1148 1149 |
static inline unsigned long effective_load(struct task_group *tg, int cpu, unsigned long wl, unsigned long wg) |
4be9daaa1 sched: fix task_h... |
1150 |
{ |
83378269a sched: correct wa... |
1151 |
return wl; |
bb3469ac9 sched: hierarchic... |
1152 |
} |
4be9daaa1 sched: fix task_h... |
1153 |
|
bb3469ac9 sched: hierarchic... |
1154 |
#endif |
c88d59108 sched: Merge sele... |
1155 |
static int wake_affine(struct sched_domain *sd, struct task_struct *p, int sync) |
098fb9db2 sched: clean up w... |
1156 |
{ |
c88d59108 sched: Merge sele... |
1157 1158 |
unsigned long this_load, load; int idx, this_cpu, prev_cpu; |
098fb9db2 sched: clean up w... |
1159 |
unsigned long tl_per_task; |
c88d59108 sched: Merge sele... |
1160 |
struct task_group *tg; |
83378269a sched: correct wa... |
1161 |
unsigned long weight; |
b3137bc8e sched: stop wake_... |
1162 |
int balanced; |
098fb9db2 sched: clean up w... |
1163 |
|
c88d59108 sched: Merge sele... |
1164 1165 1166 1167 1168 |
idx = sd->wake_idx; this_cpu = smp_processor_id(); prev_cpu = task_cpu(p); load = source_load(prev_cpu, idx); this_load = target_load(this_cpu, idx); |
098fb9db2 sched: clean up w... |
1169 1170 |
/* |
b3137bc8e sched: stop wake_... |
1171 1172 1173 1174 |
* If sync wakeup then subtract the (maximum possible) * effect of the currently running task from the load * of the current CPU: */ |
f3b577dec rcu: apply RCU pr... |
1175 |
rcu_read_lock(); |
83378269a sched: correct wa... |
1176 1177 1178 |
if (sync) { tg = task_group(current); weight = current->se.load.weight; |
c88d59108 sched: Merge sele... |
1179 |
this_load += effective_load(tg, this_cpu, -weight, -weight); |
83378269a sched: correct wa... |
1180 1181 |
load += effective_load(tg, prev_cpu, 0, -weight); } |
b3137bc8e sched: stop wake_... |
1182 |
|
83378269a sched: correct wa... |
1183 1184 |
tg = task_group(p); weight = p->se.load.weight; |
b3137bc8e sched: stop wake_... |
1185 |
|
71a29aa7b sched: Deal with ... |
1186 1187 |
/* * In low-load situations, where prev_cpu is idle and this_cpu is idle |
c88d59108 sched: Merge sele... |
1188 1189 1190 |
* due to the sync cause above having dropped this_load to 0, we'll * always have an imbalance, but there's really nothing you can do * about that, so that's good too. |
71a29aa7b sched: Deal with ... |
1191 1192 1193 1194 |
* * Otherwise check if either cpus are near enough in load to allow this * task to be woken on this_cpu. */ |
e51fd5e22 sched: Fix wake_a... |
1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 |
if (this_load) { unsigned long this_eff_load, prev_eff_load; this_eff_load = 100; this_eff_load *= power_of(prev_cpu); this_eff_load *= this_load + effective_load(tg, this_cpu, weight, weight); prev_eff_load = 100 + (sd->imbalance_pct - 100) / 2; prev_eff_load *= power_of(this_cpu); prev_eff_load *= load + effective_load(tg, prev_cpu, 0, weight); balanced = this_eff_load <= prev_eff_load; } else balanced = true; |
f3b577dec rcu: apply RCU pr... |
1210 |
rcu_read_unlock(); |
b3137bc8e sched: stop wake_... |
1211 1212 |
/* |
4ae7d5cef sched: improve af... |
1213 1214 1215 |
* If the currently running task will sleep within * a reasonable amount of time then attract this newly * woken task: |
098fb9db2 sched: clean up w... |
1216 |
*/ |
2fb7635c4 sched: sync wakeu... |
1217 1218 |
if (sync && balanced) return 1; |
098fb9db2 sched: clean up w... |
1219 |
|
41acab885 sched: Implement ... |
1220 |
schedstat_inc(p, se.statistics.nr_wakeups_affine_attempts); |
098fb9db2 sched: clean up w... |
1221 |
tl_per_task = cpu_avg_load_per_task(this_cpu); |
c88d59108 sched: Merge sele... |
1222 1223 1224 |
if (balanced || (this_load <= load && this_load + target_load(prev_cpu, idx) <= tl_per_task)) { |
098fb9db2 sched: clean up w... |
1225 1226 1227 1228 1229 |
/* * This domain has SD_WAKE_AFFINE and * p is cache cold in this domain, and * there is no bad imbalance. */ |
c88d59108 sched: Merge sele... |
1230 |
schedstat_inc(sd, ttwu_move_affine); |
41acab885 sched: Implement ... |
1231 |
schedstat_inc(p, se.statistics.nr_wakeups_affine); |
098fb9db2 sched: clean up w... |
1232 1233 1234 1235 1236 |
return 1; } return 0; } |
aaee1203c sched: Move sched... |
1237 1238 1239 1240 1241 |
/* * find_idlest_group finds and returns the least busy CPU group within the * domain. */ static struct sched_group * |
78e7ed53c sched: Tweak wake... |
1242 |
find_idlest_group(struct sched_domain *sd, struct task_struct *p, |
5158f4e44 sched: Clean up t... |
1243 |
int this_cpu, int load_idx) |
e7693a362 sched: de-SCHED_O... |
1244 |
{ |
b3bd3de66 gcc-4.6: kernel/*... |
1245 |
struct sched_group *idlest = NULL, *group = sd->groups; |
aaee1203c sched: Move sched... |
1246 |
unsigned long min_load = ULONG_MAX, this_load = 0; |
aaee1203c sched: Move sched... |
1247 |
int imbalance = 100 + (sd->imbalance_pct-100)/2; |
e7693a362 sched: de-SCHED_O... |
1248 |
|
aaee1203c sched: Move sched... |
1249 1250 1251 1252 |
do { unsigned long load, avg_load; int local_group; int i; |
e7693a362 sched: de-SCHED_O... |
1253 |
|
aaee1203c sched: Move sched... |
1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 |
/* Skip over this group if it has no CPUs allowed */ if (!cpumask_intersects(sched_group_cpus(group), &p->cpus_allowed)) continue; local_group = cpumask_test_cpu(this_cpu, sched_group_cpus(group)); /* Tally up the load of all CPUs in the group */ avg_load = 0; for_each_cpu(i, sched_group_cpus(group)) { /* Bias balancing toward cpus of our domain */ if (local_group) load = source_load(i, load_idx); else load = target_load(i, load_idx); avg_load += load; } /* Adjust by relative CPU power of the group */ avg_load = (avg_load * SCHED_LOAD_SCALE) / group->cpu_power; if (local_group) { this_load = avg_load; |
aaee1203c sched: Move sched... |
1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 |
} else if (avg_load < min_load) { min_load = avg_load; idlest = group; } } while (group = group->next, group != sd->groups); if (!idlest || 100*this_load < imbalance*min_load) return NULL; return idlest; } /* * find_idlest_cpu - find the idlest cpu among the cpus in group. */ static int find_idlest_cpu(struct sched_group *group, struct task_struct *p, int this_cpu) { unsigned long load, min_load = ULONG_MAX; int idlest = -1; int i; /* Traverse only the allowed CPUs */ for_each_cpu_and(i, sched_group_cpus(group), &p->cpus_allowed) { load = weighted_cpuload(i); if (load < min_load || (load == min_load && i == this_cpu)) { min_load = load; idlest = i; |
e7693a362 sched: de-SCHED_O... |
1308 1309 |
} } |
aaee1203c sched: Move sched... |
1310 1311 |
return idlest; } |
e7693a362 sched: de-SCHED_O... |
1312 |
|
aaee1203c sched: Move sched... |
1313 |
/* |
a50bde513 sched: Cleanup se... |
1314 1315 |
* Try and locate an idle CPU in the sched_domain. */ |
99bd5e2f2 sched: Fix select... |
1316 |
static int select_idle_sibling(struct task_struct *p, int target) |
a50bde513 sched: Cleanup se... |
1317 1318 1319 |
{ int cpu = smp_processor_id(); int prev_cpu = task_cpu(p); |
99bd5e2f2 sched: Fix select... |
1320 |
struct sched_domain *sd; |
a50bde513 sched: Cleanup se... |
1321 1322 1323 |
int i; /* |
99bd5e2f2 sched: Fix select... |
1324 1325 |
* If the task is going to be woken-up on this cpu and if it is * already idle, then it is the right target. |
a50bde513 sched: Cleanup se... |
1326 |
*/ |
99bd5e2f2 sched: Fix select... |
1327 1328 1329 1330 1331 1332 1333 1334 |
if (target == cpu && idle_cpu(cpu)) return cpu; /* * If the task is going to be woken-up on the cpu where it previously * ran and if it is currently idle, then it the right target. */ if (target == prev_cpu && idle_cpu(prev_cpu)) |
fe3bcfe1f sched: More gener... |
1335 |
return prev_cpu; |
a50bde513 sched: Cleanup se... |
1336 1337 |
/* |
99bd5e2f2 sched: Fix select... |
1338 |
* Otherwise, iterate the domains and find an elegible idle cpu. |
a50bde513 sched: Cleanup se... |
1339 |
*/ |
99bd5e2f2 sched: Fix select... |
1340 1341 |
for_each_domain(target, sd) { if (!(sd->flags & SD_SHARE_PKG_RESOURCES)) |
fe3bcfe1f sched: More gener... |
1342 |
break; |
99bd5e2f2 sched: Fix select... |
1343 1344 1345 1346 1347 1348 |
for_each_cpu_and(i, sched_domain_span(sd), &p->cpus_allowed) { if (idle_cpu(i)) { target = i; break; } |
a50bde513 sched: Cleanup se... |
1349 |
} |
99bd5e2f2 sched: Fix select... |
1350 1351 1352 1353 1354 1355 1356 1357 |
/* * Lets stop looking for an idle sibling when we reached * the domain that spans the current cpu and prev_cpu. */ if (cpumask_test_cpu(cpu, sched_domain_span(sd)) && cpumask_test_cpu(prev_cpu, sched_domain_span(sd))) break; |
a50bde513 sched: Cleanup se... |
1358 1359 1360 1361 1362 1363 |
} return target; } /* |
aaee1203c sched: Move sched... |
1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 |
* sched_balance_self: balance the current task (running on cpu) in domains * that have the 'flag' flag set. In practice, this is SD_BALANCE_FORK and * SD_BALANCE_EXEC. * * Balance, ie. select the least loaded group. * * Returns the target CPU number, or the same CPU if no balancing is needed. * * preempt must be disabled. */ |
0017d7350 sched: Fix TASK_W... |
1374 1375 |
static int select_task_rq_fair(struct rq *rq, struct task_struct *p, int sd_flag, int wake_flags) |
aaee1203c sched: Move sched... |
1376 |
{ |
29cd8bae3 sched: Fix SD_POW... |
1377 |
struct sched_domain *tmp, *affine_sd = NULL, *sd = NULL; |
c88d59108 sched: Merge sele... |
1378 1379 1380 |
int cpu = smp_processor_id(); int prev_cpu = task_cpu(p); int new_cpu = cpu; |
99bd5e2f2 sched: Fix select... |
1381 |
int want_affine = 0; |
29cd8bae3 sched: Fix SD_POW... |
1382 |
int want_sd = 1; |
5158f4e44 sched: Clean up t... |
1383 |
int sync = wake_flags & WF_SYNC; |
c88d59108 sched: Merge sele... |
1384 |
|
0763a660a sched: Rename sel... |
1385 |
if (sd_flag & SD_BALANCE_WAKE) { |
beac4c7e4 sched: Remove AFF... |
1386 |
if (cpumask_test_cpu(cpu, &p->cpus_allowed)) |
c88d59108 sched: Merge sele... |
1387 1388 1389 |
want_affine = 1; new_cpu = prev_cpu; } |
aaee1203c sched: Move sched... |
1390 1391 |
for_each_domain(cpu, tmp) { |
e4f428884 sched: Select_tas... |
1392 1393 |
if (!(tmp->flags & SD_LOAD_BALANCE)) continue; |
aaee1203c sched: Move sched... |
1394 |
/* |
ae154be1f sched: Weaken SD_... |
1395 1396 |
* If power savings logic is enabled for a domain, see if we * are not overloaded, if so, don't balance wider. |
aaee1203c sched: Move sched... |
1397 |
*/ |
59abf0264 sched: Add SD_PRE... |
1398 |
if (tmp->flags & (SD_POWERSAVINGS_BALANCE|SD_PREFER_LOCAL)) { |
ae154be1f sched: Weaken SD_... |
1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 |
unsigned long power = 0; unsigned long nr_running = 0; unsigned long capacity; int i; for_each_cpu(i, sched_domain_span(tmp)) { power += power_of(i); nr_running += cpu_rq(i)->cfs.nr_running; } capacity = DIV_ROUND_CLOSEST(power, SCHED_LOAD_SCALE); |
59abf0264 sched: Add SD_PRE... |
1410 1411 1412 1413 |
if (tmp->flags & SD_POWERSAVINGS_BALANCE) nr_running /= 2; if (nr_running < capacity) |
29cd8bae3 sched: Fix SD_POW... |
1414 |
want_sd = 0; |
ae154be1f sched: Weaken SD_... |
1415 |
} |
aaee1203c sched: Move sched... |
1416 |
|
fe3bcfe1f sched: More gener... |
1417 |
/* |
99bd5e2f2 sched: Fix select... |
1418 1419 |
* If both cpu and prev_cpu are part of this domain, * cpu is a valid SD_WAKE_AFFINE target. |
fe3bcfe1f sched: More gener... |
1420 |
*/ |
99bd5e2f2 sched: Fix select... |
1421 1422 1423 1424 |
if (want_affine && (tmp->flags & SD_WAKE_AFFINE) && cpumask_test_cpu(prev_cpu, sched_domain_span(tmp))) { affine_sd = tmp; want_affine = 0; |
c88d59108 sched: Merge sele... |
1425 |
} |
29cd8bae3 sched: Fix SD_POW... |
1426 1427 |
if (!want_sd && !want_affine) break; |
0763a660a sched: Rename sel... |
1428 |
if (!(tmp->flags & sd_flag)) |
c88d59108 sched: Merge sele... |
1429 |
continue; |
29cd8bae3 sched: Fix SD_POW... |
1430 1431 1432 |
if (want_sd) sd = tmp; } |
8b911acdf sched: Fix select... |
1433 |
#ifdef CONFIG_FAIR_GROUP_SCHED |
29cd8bae3 sched: Fix SD_POW... |
1434 1435 1436 1437 1438 |
if (sched_feat(LB_SHARES_UPDATE)) { /* * Pick the largest domain to update shares over */ tmp = sd; |
669c55e9f sched: Pre-comput... |
1439 |
if (affine_sd && (!tmp || affine_sd->span_weight > sd->span_weight)) |
29cd8bae3 sched: Fix SD_POW... |
1440 |
tmp = affine_sd; |
0017d7350 sched: Fix TASK_W... |
1441 1442 |
if (tmp) { raw_spin_unlock(&rq->lock); |
29cd8bae3 sched: Fix SD_POW... |
1443 |
update_shares(tmp); |
0017d7350 sched: Fix TASK_W... |
1444 1445 |
raw_spin_lock(&rq->lock); } |
c88d59108 sched: Merge sele... |
1446 |
} |
8b911acdf sched: Fix select... |
1447 |
#endif |
aaee1203c sched: Move sched... |
1448 |
|
8b911acdf sched: Fix select... |
1449 |
if (affine_sd) { |
99bd5e2f2 sched: Fix select... |
1450 1451 1452 1453 |
if (cpu == prev_cpu || wake_affine(affine_sd, p, sync)) return select_idle_sibling(p, cpu); else return select_idle_sibling(p, prev_cpu); |
8b911acdf sched: Fix select... |
1454 |
} |
e7693a362 sched: de-SCHED_O... |
1455 |
|
aaee1203c sched: Move sched... |
1456 |
while (sd) { |
5158f4e44 sched: Clean up t... |
1457 |
int load_idx = sd->forkexec_idx; |
aaee1203c sched: Move sched... |
1458 |
struct sched_group *group; |
c88d59108 sched: Merge sele... |
1459 |
int weight; |
098fb9db2 sched: clean up w... |
1460 |
|
0763a660a sched: Rename sel... |
1461 |
if (!(sd->flags & sd_flag)) { |
aaee1203c sched: Move sched... |
1462 1463 1464 |
sd = sd->child; continue; } |
098fb9db2 sched: clean up w... |
1465 |
|
5158f4e44 sched: Clean up t... |
1466 1467 |
if (sd_flag & SD_BALANCE_WAKE) load_idx = sd->wake_idx; |
098fb9db2 sched: clean up w... |
1468 |
|
5158f4e44 sched: Clean up t... |
1469 |
group = find_idlest_group(sd, p, cpu, load_idx); |
aaee1203c sched: Move sched... |
1470 1471 1472 1473 |
if (!group) { sd = sd->child; continue; } |
4ae7d5cef sched: improve af... |
1474 |
|
d7c33c493 sched: Fix task a... |
1475 |
new_cpu = find_idlest_cpu(group, p, cpu); |
aaee1203c sched: Move sched... |
1476 1477 1478 1479 |
if (new_cpu == -1 || new_cpu == cpu) { /* Now try balancing at a lower domain level of cpu */ sd = sd->child; continue; |
e7693a362 sched: de-SCHED_O... |
1480 |
} |
aaee1203c sched: Move sched... |
1481 1482 1483 |
/* Now try balancing at a lower domain level of new_cpu */ cpu = new_cpu; |
669c55e9f sched: Pre-comput... |
1484 |
weight = sd->span_weight; |
aaee1203c sched: Move sched... |
1485 1486 |
sd = NULL; for_each_domain(cpu, tmp) { |
669c55e9f sched: Pre-comput... |
1487 |
if (weight <= tmp->span_weight) |
aaee1203c sched: Move sched... |
1488 |
break; |
0763a660a sched: Rename sel... |
1489 |
if (tmp->flags & sd_flag) |
aaee1203c sched: Move sched... |
1490 1491 1492 |
sd = tmp; } /* while loop will break here if sd == NULL */ |
e7693a362 sched: de-SCHED_O... |
1493 |
} |
c88d59108 sched: Merge sele... |
1494 |
return new_cpu; |
e7693a362 sched: de-SCHED_O... |
1495 1496 |
} #endif /* CONFIG_SMP */ |
e52fb7c09 sched: prefer wakers |
1497 1498 |
static unsigned long wakeup_gran(struct sched_entity *curr, struct sched_entity *se) |
0bbd3336e sched: fix wakeup... |
1499 1500 1501 1502 |
{ unsigned long gran = sysctl_sched_wakeup_granularity; /* |
e52fb7c09 sched: prefer wakers |
1503 1504 |
* Since its curr running now, convert the gran from real-time * to virtual-time in his units. |
13814d42e sched: Remove ASY... |
1505 1506 1507 1508 1509 1510 1511 1512 1513 |
* * By using 'se' instead of 'curr' we penalize light tasks, so * they get preempted easier. That is, if 'se' < 'curr' then * the resulting gran will be larger, therefore penalizing the * lighter, if otoh 'se' > 'curr' then the resulting gran will * be smaller, again penalizing the lighter task. * * This is especially important for buddies when the leftmost * task is higher priority than the buddy. |
0bbd3336e sched: fix wakeup... |
1514 |
*/ |
13814d42e sched: Remove ASY... |
1515 1516 |
if (unlikely(se->load.weight != NICE_0_LOAD)) gran = calc_delta_fair(gran, se); |
0bbd3336e sched: fix wakeup... |
1517 1518 1519 1520 1521 |
return gran; } /* |
464b75273 sched: re-instate... |
1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 |
* Should 'se' preempt 'curr'. * * |s1 * |s2 * |s3 * g * |<--->|c * * w(c, s1) = -1 * w(c, s2) = 0 * w(c, s3) = 1 * */ static int wakeup_preempt_entity(struct sched_entity *curr, struct sched_entity *se) { s64 gran, vdiff = curr->vruntime - se->vruntime; if (vdiff <= 0) return -1; |
e52fb7c09 sched: prefer wakers |
1542 |
gran = wakeup_gran(curr, se); |
464b75273 sched: re-instate... |
1543 1544 1545 1546 1547 |
if (vdiff > gran) return 1; return 0; } |
02479099c sched: fix buddie... |
1548 1549 |
static void set_last_buddy(struct sched_entity *se) { |
6bc912b71 sched: SCHED_OTHE... |
1550 1551 1552 1553 |
if (likely(task_of(se)->policy != SCHED_IDLE)) { for_each_sched_entity(se) cfs_rq_of(se)->last = se; } |
02479099c sched: fix buddie... |
1554 1555 1556 1557 |
} static void set_next_buddy(struct sched_entity *se) { |
6bc912b71 sched: SCHED_OTHE... |
1558 1559 1560 1561 |
if (likely(task_of(se)->policy != SCHED_IDLE)) { for_each_sched_entity(se) cfs_rq_of(se)->next = se; } |
02479099c sched: fix buddie... |
1562 |
} |
464b75273 sched: re-instate... |
1563 |
/* |
bf0f6f24a sched: cfs core, ... |
1564 1565 |
* Preempt the current task with a newly woken task if needed: */ |
5a9b86f64 sched: Rename fla... |
1566 |
static void check_preempt_wakeup(struct rq *rq, struct task_struct *p, int wake_flags) |
bf0f6f24a sched: cfs core, ... |
1567 1568 |
{ struct task_struct *curr = rq->curr; |
8651a86c3 sched: group sche... |
1569 |
struct sched_entity *se = &curr->se, *pse = &p->se; |
03e89e457 sched: fix wakeup... |
1570 |
struct cfs_rq *cfs_rq = task_cfs_rq(curr); |
f685ceaca sched: Strengthen... |
1571 |
int scale = cfs_rq->nr_running >= sched_nr_latency; |
bf0f6f24a sched: cfs core, ... |
1572 |
|
3a7e73a2e sched: Clean up c... |
1573 1574 |
if (unlikely(rt_prio(p->prio))) goto preempt; |
aa2ac2522 sched: fix overlo... |
1575 |
|
d95f98d06 sched: fix fair p... |
1576 1577 |
if (unlikely(p->sched_class != &fair_sched_class)) return; |
4ae7d5cef sched: improve af... |
1578 1579 |
if (unlikely(se == pse)) return; |
f685ceaca sched: Strengthen... |
1580 |
if (sched_feat(NEXT_BUDDY) && scale && !(wake_flags & WF_FORK)) |
3cb63d527 sched: Complete b... |
1581 |
set_next_buddy(pse); |
57fdc26d4 sched: fixup budd... |
1582 |
|
aec0a5142 sched: call resch... |
1583 1584 1585 1586 1587 1588 |
/* * We can come here with TIF_NEED_RESCHED already set from new task * wake up path. */ if (test_tsk_need_resched(curr)) return; |
91c234b4e sched: do not wak... |
1589 |
/* |
6bc912b71 sched: SCHED_OTHE... |
1590 |
* Batch and idle tasks do not preempt (their preemption is driven by |
91c234b4e sched: do not wak... |
1591 1592 |
* the tick): */ |
6bc912b71 sched: SCHED_OTHE... |
1593 |
if (unlikely(p->policy != SCHED_NORMAL)) |
91c234b4e sched: do not wak... |
1594 |
return; |
bf0f6f24a sched: cfs core, ... |
1595 |
|
6bc912b71 sched: SCHED_OTHE... |
1596 |
/* Idle tasks are by definition preempted by everybody. */ |
3a7e73a2e sched: Clean up c... |
1597 1598 |
if (unlikely(curr->policy == SCHED_IDLE)) goto preempt; |
bf0f6f24a sched: cfs core, ... |
1599 |
|
ad4b78bbc sched: Add new wa... |
1600 1601 |
if (!sched_feat(WAKEUP_PREEMPT)) return; |
3a7e73a2e sched: Clean up c... |
1602 |
update_curr(cfs_rq); |
464b75273 sched: re-instate... |
1603 |
find_matching_se(&se, &pse); |
002f128b4 sched: remove red... |
1604 |
BUG_ON(!pse); |
3a7e73a2e sched: Clean up c... |
1605 1606 |
if (wakeup_preempt_entity(se, pse) == 1) goto preempt; |
464b75273 sched: re-instate... |
1607 |
|
3a7e73a2e sched: Clean up c... |
1608 |
return; |
a65ac745e sched: Move updat... |
1609 |
|
3a7e73a2e sched: Clean up c... |
1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 |
preempt: resched_task(curr); /* * Only set the backward buddy when the current task is still * on the rq. This can happen when a wakeup gets interleaved * with schedule on the ->pre_schedule() or idle_balance() * point, either of which can * drop the rq lock. * * Also, during early boot the idle thread is in the fair class, * for obvious reasons its a bad idea to schedule back to it. */ if (unlikely(!se->on_rq || curr == rq->idle)) return; if (sched_feat(LAST_BUDDY) && scale && entity_is_task(se)) set_last_buddy(se); |
bf0f6f24a sched: cfs core, ... |
1626 |
} |
fb8d47240 sched: remove the... |
1627 |
static struct task_struct *pick_next_task_fair(struct rq *rq) |
bf0f6f24a sched: cfs core, ... |
1628 |
{ |
8f4d37ec0 sched: high-res p... |
1629 |
struct task_struct *p; |
bf0f6f24a sched: cfs core, ... |
1630 1631 |
struct cfs_rq *cfs_rq = &rq->cfs; struct sched_entity *se; |
36ace27e3 sched: Optimize b... |
1632 |
if (!cfs_rq->nr_running) |
bf0f6f24a sched: cfs core, ... |
1633 1634 1635 |
return NULL; do { |
9948f4b2a sched: remove the... |
1636 |
se = pick_next_entity(cfs_rq); |
f4b6755fb sched: cleanup fa... |
1637 |
set_next_entity(cfs_rq, se); |
bf0f6f24a sched: cfs core, ... |
1638 1639 |
cfs_rq = group_cfs_rq(se); } while (cfs_rq); |
8f4d37ec0 sched: high-res p... |
1640 1641 1642 1643 |
p = task_of(se); hrtick_start_fair(rq, p); return p; |
bf0f6f24a sched: cfs core, ... |
1644 1645 1646 1647 1648 |
} /* * Account for a descheduled task: */ |
31ee529cc sched: remove the... |
1649 |
static void put_prev_task_fair(struct rq *rq, struct task_struct *prev) |
bf0f6f24a sched: cfs core, ... |
1650 1651 1652 1653 1654 1655 |
{ struct sched_entity *se = &prev->se; struct cfs_rq *cfs_rq; for_each_sched_entity(se) { cfs_rq = cfs_rq_of(se); |
ab6cde269 sched: remove the... |
1656 |
put_prev_entity(cfs_rq, se); |
bf0f6f24a sched: cfs core, ... |
1657 1658 |
} } |
681f3e685 sched: isolate SM... |
1659 |
#ifdef CONFIG_SMP |
bf0f6f24a sched: cfs core, ... |
1660 1661 1662 |
/************************************************** * Fair scheduling class load-balancing methods: */ |
1e3c88bde sched: Move load ... |
1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 |
/* * pull_task - move a task from a remote runqueue to the local runqueue. * Both runqueues must be locked. */ static void pull_task(struct rq *src_rq, struct task_struct *p, struct rq *this_rq, int this_cpu) { deactivate_task(src_rq, p, 0); set_task_cpu(p, this_cpu); activate_task(this_rq, p, 0); check_preempt_curr(this_rq, p, 0); |
fab476228 sched: Force bala... |
1674 1675 1676 1677 |
/* re-arm NEWIDLE balancing when moving tasks */ src_rq->avg_idle = this_rq->avg_idle = 2*sysctl_sched_migration_cost; this_rq->idle_stamp = 0; |
1e3c88bde sched: Move load ... |
1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 |
} /* * can_migrate_task - may task p from runqueue rq be migrated to this_cpu? */ static int can_migrate_task(struct task_struct *p, struct rq *rq, int this_cpu, struct sched_domain *sd, enum cpu_idle_type idle, int *all_pinned) { int tsk_cache_hot = 0; /* * We do not migrate tasks that are: * 1) running (obviously), or * 2) cannot be migrated to this CPU due to cpus_allowed, or * 3) are cache-hot on their current CPU. */ if (!cpumask_test_cpu(this_cpu, &p->cpus_allowed)) { |
41acab885 sched: Implement ... |
1696 |
schedstat_inc(p, se.statistics.nr_failed_migrations_affine); |
1e3c88bde sched: Move load ... |
1697 1698 1699 1700 1701 |
return 0; } *all_pinned = 0; if (task_running(rq, p)) { |
41acab885 sched: Implement ... |
1702 |
schedstat_inc(p, se.statistics.nr_failed_migrations_running); |
1e3c88bde sched: Move load ... |
1703 1704 1705 1706 1707 1708 1709 1710 |
return 0; } /* * Aggressive migration if: * 1) task is cache cold, or * 2) too many balance attempts have failed. */ |
305e6835e sched: Do not acc... |
1711 |
tsk_cache_hot = task_hot(p, rq->clock_task, sd); |
1e3c88bde sched: Move load ... |
1712 1713 1714 1715 1716 |
if (!tsk_cache_hot || sd->nr_balance_failed > sd->cache_nice_tries) { #ifdef CONFIG_SCHEDSTATS if (tsk_cache_hot) { schedstat_inc(sd, lb_hot_gained[idle]); |
41acab885 sched: Implement ... |
1717 |
schedstat_inc(p, se.statistics.nr_forced_migrations); |
1e3c88bde sched: Move load ... |
1718 1719 1720 1721 1722 1723 |
} #endif return 1; } if (tsk_cache_hot) { |
41acab885 sched: Implement ... |
1724 |
schedstat_inc(p, se.statistics.nr_failed_migrations_hot); |
1e3c88bde sched: Move load ... |
1725 1726 1727 1728 |
return 0; } return 1; } |
897c395f4 sched: Remove rq_... |
1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 |
/* * move_one_task tries to move exactly one task from busiest to this_rq, as * part of active balancing operations within "domain". * Returns 1 if successful and 0 otherwise. * * Called with both runqueues locked. */ static int move_one_task(struct rq *this_rq, int this_cpu, struct rq *busiest, struct sched_domain *sd, enum cpu_idle_type idle) { struct task_struct *p, *n; struct cfs_rq *cfs_rq; int pinned = 0; for_each_leaf_cfs_rq(busiest, cfs_rq) { list_for_each_entry_safe(p, n, &cfs_rq->tasks, se.group_node) { if (!can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) continue; pull_task(busiest, p, this_rq, this_cpu); /* * Right now, this is only the second place pull_task() * is called, so we can safely collect pull_task() * stats here rather than inside pull_task(). */ schedstat_inc(sd, lb_gained[idle]); return 1; } } return 0; } |
1e3c88bde sched: Move load ... |
1764 1765 1766 1767 |
static unsigned long balance_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, unsigned long max_load_move, struct sched_domain *sd, enum cpu_idle_type idle, int *all_pinned, |
ee00e66ff sched: Remove rq_... |
1768 |
int *this_best_prio, struct cfs_rq *busiest_cfs_rq) |
1e3c88bde sched: Move load ... |
1769 1770 |
{ int loops = 0, pulled = 0, pinned = 0; |
1e3c88bde sched: Move load ... |
1771 |
long rem_load_move = max_load_move; |
ee00e66ff sched: Remove rq_... |
1772 |
struct task_struct *p, *n; |
1e3c88bde sched: Move load ... |
1773 1774 1775 1776 1777 |
if (max_load_move == 0) goto out; pinned = 1; |
ee00e66ff sched: Remove rq_... |
1778 1779 1780 |
list_for_each_entry_safe(p, n, &busiest_cfs_rq->tasks, se.group_node) { if (loops++ > sysctl_sched_nr_migrate) break; |
1e3c88bde sched: Move load ... |
1781 |
|
ee00e66ff sched: Remove rq_... |
1782 1783 1784 |
if ((p->se.load.weight >> 1) > rem_load_move || !can_migrate_task(p, busiest, this_cpu, sd, idle, &pinned)) continue; |
1e3c88bde sched: Move load ... |
1785 |
|
ee00e66ff sched: Remove rq_... |
1786 1787 1788 |
pull_task(busiest, p, this_rq, this_cpu); pulled++; rem_load_move -= p->se.load.weight; |
1e3c88bde sched: Move load ... |
1789 1790 |
#ifdef CONFIG_PREEMPT |
ee00e66ff sched: Remove rq_... |
1791 1792 1793 1794 1795 1796 1797 |
/* * NEWIDLE balancing is a source of latency, so preemptible * kernels will stop after the first task is pulled to minimize * the critical section. */ if (idle == CPU_NEWLY_IDLE) break; |
1e3c88bde sched: Move load ... |
1798 |
#endif |
ee00e66ff sched: Remove rq_... |
1799 1800 1801 1802 1803 1804 |
/* * We only want to steal up to the prescribed amount of * weighted load. */ if (rem_load_move <= 0) break; |
1e3c88bde sched: Move load ... |
1805 1806 |
if (p->prio < *this_best_prio) *this_best_prio = p->prio; |
1e3c88bde sched: Move load ... |
1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 |
} out: /* * Right now, this is one of only two places pull_task() is called, * so we can safely collect pull_task() stats here rather than * inside pull_task(). */ schedstat_add(sd, lb_gained[idle], pulled); if (all_pinned) *all_pinned = pinned; return max_load_move - rem_load_move; } |
230059de7 sched: Remove fro... |
1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 |
#ifdef CONFIG_FAIR_GROUP_SCHED static unsigned long load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, unsigned long max_load_move, struct sched_domain *sd, enum cpu_idle_type idle, int *all_pinned, int *this_best_prio) { long rem_load_move = max_load_move; int busiest_cpu = cpu_of(busiest); struct task_group *tg; rcu_read_lock(); update_h_load(busiest_cpu); list_for_each_entry_rcu(tg, &task_groups, list) { struct cfs_rq *busiest_cfs_rq = tg->cfs_rq[busiest_cpu]; unsigned long busiest_h_load = busiest_cfs_rq->h_load; unsigned long busiest_weight = busiest_cfs_rq->load.weight; u64 rem_load, moved_load; /* * empty group */ if (!busiest_cfs_rq->task_weight) continue; rem_load = (u64)rem_load_move * busiest_weight; rem_load = div_u64(rem_load, busiest_h_load + 1); moved_load = balance_tasks(this_rq, this_cpu, busiest, rem_load, sd, idle, all_pinned, this_best_prio, busiest_cfs_rq); if (!moved_load) continue; moved_load *= busiest_h_load; moved_load = div_u64(moved_load, busiest_weight + 1); rem_load_move -= moved_load; if (rem_load_move < 0) break; } rcu_read_unlock(); return max_load_move - rem_load_move; } #else static unsigned long load_balance_fair(struct rq *this_rq, int this_cpu, struct rq *busiest, unsigned long max_load_move, struct sched_domain *sd, enum cpu_idle_type idle, int *all_pinned, int *this_best_prio) { return balance_tasks(this_rq, this_cpu, busiest, max_load_move, sd, idle, all_pinned, this_best_prio, &busiest->cfs); } #endif |
1e3c88bde sched: Move load ... |
1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 |
/* * move_tasks tries to move up to max_load_move weighted load from busiest to * this_rq, as part of a balancing operation within domain "sd". * Returns 1 if successful and 0 otherwise. * * Called with both runqueues locked. */ static int move_tasks(struct rq *this_rq, int this_cpu, struct rq *busiest, unsigned long max_load_move, struct sched_domain *sd, enum cpu_idle_type idle, int *all_pinned) { |
3d45fd804 sched: Remove the... |
1892 |
unsigned long total_load_moved = 0, load_moved; |
1e3c88bde sched: Move load ... |
1893 1894 1895 |
int this_best_prio = this_rq->curr->prio; do { |
3d45fd804 sched: Remove the... |
1896 |
load_moved = load_balance_fair(this_rq, this_cpu, busiest, |
1e3c88bde sched: Move load ... |
1897 1898 |
max_load_move - total_load_moved, sd, idle, all_pinned, &this_best_prio); |
3d45fd804 sched: Remove the... |
1899 1900 |
total_load_moved += load_moved; |
1e3c88bde sched: Move load ... |
1901 1902 1903 1904 1905 1906 1907 1908 1909 |
#ifdef CONFIG_PREEMPT /* * NEWIDLE balancing is a source of latency, so preemptible * kernels will stop after the first task is pulled to minimize * the critical section. */ if (idle == CPU_NEWLY_IDLE && this_rq->nr_running) break; |
baa8c1102 sched: Add a lock... |
1910 1911 1912 1913 |
if (raw_spin_is_contended(&this_rq->lock) || raw_spin_is_contended(&busiest->lock)) break; |
1e3c88bde sched: Move load ... |
1914 |
#endif |
3d45fd804 sched: Remove the... |
1915 |
} while (load_moved && max_load_move > total_load_moved); |
1e3c88bde sched: Move load ... |
1916 1917 1918 |
return total_load_moved > 0; } |
1e3c88bde sched: Move load ... |
1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 |
/********** Helpers for find_busiest_group ************************/ /* * sd_lb_stats - Structure to store the statistics of a sched_domain * during load balancing. */ struct sd_lb_stats { struct sched_group *busiest; /* Busiest group in this sd */ struct sched_group *this; /* Local group in this sd */ unsigned long total_load; /* Total load of all groups in sd */ unsigned long total_pwr; /* Total power of all groups in sd */ unsigned long avg_load; /* Average load across all groups in sd */ /** Statistics of this group */ unsigned long this_load; unsigned long this_load_per_task; unsigned long this_nr_running; |
fab476228 sched: Force bala... |
1935 |
unsigned long this_has_capacity; |
1e3c88bde sched: Move load ... |
1936 1937 1938 1939 1940 |
/* Statistics of the busiest group */ unsigned long max_load; unsigned long busiest_load_per_task; unsigned long busiest_nr_running; |
dd5feea14 sched: Fix SCHED_... |
1941 |
unsigned long busiest_group_capacity; |
fab476228 sched: Force bala... |
1942 |
unsigned long busiest_has_capacity; |
1e3c88bde sched: Move load ... |
1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 |
int group_imb; /* Is there imbalance in this sd */ #if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) int power_savings_balance; /* Is powersave balance needed for this sd */ struct sched_group *group_min; /* Least loaded group in sd */ struct sched_group *group_leader; /* Group which relieves group_min */ unsigned long min_load_per_task; /* load_per_task in group_min */ unsigned long leader_nr_running; /* Nr running of group_leader */ unsigned long min_nr_running; /* Nr running of group_min */ #endif }; /* * sg_lb_stats - stats of a sched_group required for load_balancing */ struct sg_lb_stats { unsigned long avg_load; /*Avg load across the CPUs of the group */ unsigned long group_load; /* Total load over the CPUs of the group */ unsigned long sum_nr_running; /* Nr tasks running in the group */ unsigned long sum_weighted_load; /* Weighted load of group's tasks */ unsigned long group_capacity; int group_imb; /* Is there an imbalance in the group ? */ |
fab476228 sched: Force bala... |
1965 |
int group_has_capacity; /* Is there extra capacity in the group? */ |
1e3c88bde sched: Move load ... |
1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 |
}; /** * group_first_cpu - Returns the first cpu in the cpumask of a sched_group. * @group: The group whose first cpu is to be returned. */ static inline unsigned int group_first_cpu(struct sched_group *group) { return cpumask_first(sched_group_cpus(group)); } /** * get_sd_load_idx - Obtain the load index for a given sched domain. * @sd: The sched_domain whose load_idx is to be obtained. * @idle: The Idle status of the CPU for whose sd load_icx is obtained. */ static inline int get_sd_load_idx(struct sched_domain *sd, enum cpu_idle_type idle) { int load_idx; switch (idle) { case CPU_NOT_IDLE: load_idx = sd->busy_idx; break; case CPU_NEWLY_IDLE: load_idx = sd->newidle_idx; break; default: load_idx = sd->idle_idx; break; } return load_idx; } #if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) /** * init_sd_power_savings_stats - Initialize power savings statistics for * the given sched_domain, during load balancing. * * @sd: Sched domain whose power-savings statistics are to be initialized. * @sds: Variable containing the statistics for sd. * @idle: Idle status of the CPU at which we're performing load-balancing. */ static inline void init_sd_power_savings_stats(struct sched_domain *sd, struct sd_lb_stats *sds, enum cpu_idle_type idle) { /* * Busy processors will not participate in power savings * balance. */ if (idle == CPU_NOT_IDLE || !(sd->flags & SD_POWERSAVINGS_BALANCE)) sds->power_savings_balance = 0; else { sds->power_savings_balance = 1; sds->min_nr_running = ULONG_MAX; sds->leader_nr_running = 0; } } /** * update_sd_power_savings_stats - Update the power saving stats for a * sched_domain while performing load balancing. * * @group: sched_group belonging to the sched_domain under consideration. * @sds: Variable containing the statistics of the sched_domain * @local_group: Does group contain the CPU for which we're performing * load balancing ? * @sgs: Variable containing the statistics of the group. */ static inline void update_sd_power_savings_stats(struct sched_group *group, struct sd_lb_stats *sds, int local_group, struct sg_lb_stats *sgs) { if (!sds->power_savings_balance) return; /* * If the local group is idle or completely loaded * no need to do power savings balance at this domain */ if (local_group && (sds->this_nr_running >= sgs->group_capacity || !sds->this_nr_running)) sds->power_savings_balance = 0; /* * If a group is already running at full capacity or idle, * don't include that group in power savings calculations */ if (!sds->power_savings_balance || sgs->sum_nr_running >= sgs->group_capacity || !sgs->sum_nr_running) return; /* * Calculate the group which has the least non-idle load. * This is the group from where we need to pick up the load * for saving power */ if ((sgs->sum_nr_running < sds->min_nr_running) || (sgs->sum_nr_running == sds->min_nr_running && group_first_cpu(group) > group_first_cpu(sds->group_min))) { sds->group_min = group; sds->min_nr_running = sgs->sum_nr_running; sds->min_load_per_task = sgs->sum_weighted_load / sgs->sum_nr_running; } /* * Calculate the group which is almost near its * capacity but still has some space to pick up some load * from other group and save more power */ if (sgs->sum_nr_running + 1 > sgs->group_capacity) return; if (sgs->sum_nr_running > sds->leader_nr_running || (sgs->sum_nr_running == sds->leader_nr_running && group_first_cpu(group) < group_first_cpu(sds->group_leader))) { sds->group_leader = group; sds->leader_nr_running = sgs->sum_nr_running; } } /** * check_power_save_busiest_group - see if there is potential for some power-savings balance * @sds: Variable containing the statistics of the sched_domain * under consideration. * @this_cpu: Cpu at which we're currently performing load-balancing. * @imbalance: Variable to store the imbalance. * * Description: * Check if we have potential to perform some power-savings balance. * If yes, set the busiest group to be the least loaded group in the * sched_domain, so that it's CPUs can be put to idle. * * Returns 1 if there is potential to perform power-savings balance. * Else returns 0. */ static inline int check_power_save_busiest_group(struct sd_lb_stats *sds, int this_cpu, unsigned long *imbalance) { if (!sds->power_savings_balance) return 0; if (sds->this != sds->group_leader || sds->group_leader == sds->group_min) return 0; *imbalance = sds->min_load_per_task; sds->busiest = sds->group_min; return 1; } #else /* CONFIG_SCHED_MC || CONFIG_SCHED_SMT */ static inline void init_sd_power_savings_stats(struct sched_domain *sd, struct sd_lb_stats *sds, enum cpu_idle_type idle) { return; } static inline void update_sd_power_savings_stats(struct sched_group *group, struct sd_lb_stats *sds, int local_group, struct sg_lb_stats *sgs) { return; } static inline int check_power_save_busiest_group(struct sd_lb_stats *sds, int this_cpu, unsigned long *imbalance) { return 0; } #endif /* CONFIG_SCHED_MC || CONFIG_SCHED_SMT */ unsigned long default_scale_freq_power(struct sched_domain *sd, int cpu) { return SCHED_LOAD_SCALE; } unsigned long __weak arch_scale_freq_power(struct sched_domain *sd, int cpu) { return default_scale_freq_power(sd, cpu); } unsigned long default_scale_smt_power(struct sched_domain *sd, int cpu) { |
669c55e9f sched: Pre-comput... |
2157 |
unsigned long weight = sd->span_weight; |
1e3c88bde sched: Move load ... |
2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 |
unsigned long smt_gain = sd->smt_gain; smt_gain /= weight; return smt_gain; } unsigned long __weak arch_scale_smt_power(struct sched_domain *sd, int cpu) { return default_scale_smt_power(sd, cpu); } unsigned long scale_rt_power(int cpu) { struct rq *rq = cpu_rq(cpu); u64 total, available; |
1e3c88bde sched: Move load ... |
2174 |
total = sched_avg_period() + (rq->clock - rq->age_stamp); |
aa4838085 sched: Remove irq... |
2175 2176 2177 2178 2179 2180 2181 |
if (unlikely(total < rq->rt_avg)) { /* Ensures that power won't end up being negative */ available = 0; } else { available = total - rq->rt_avg; } |
1e3c88bde sched: Move load ... |
2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 |
if (unlikely((s64)total < SCHED_LOAD_SCALE)) total = SCHED_LOAD_SCALE; total >>= SCHED_LOAD_SHIFT; return div_u64(available, total); } static void update_cpu_power(struct sched_domain *sd, int cpu) { |
669c55e9f sched: Pre-comput... |
2193 |
unsigned long weight = sd->span_weight; |
1e3c88bde sched: Move load ... |
2194 2195 |
unsigned long power = SCHED_LOAD_SCALE; struct sched_group *sdg = sd->groups; |
1e3c88bde sched: Move load ... |
2196 2197 2198 2199 2200 2201 2202 2203 |
if ((sd->flags & SD_SHARE_CPUPOWER) && weight > 1) { if (sched_feat(ARCH_POWER)) power *= arch_scale_smt_power(sd, cpu); else power *= default_scale_smt_power(sd, cpu); power >>= SCHED_LOAD_SHIFT; } |
9d5efe05e sched: Fix capaci... |
2204 2205 2206 2207 2208 2209 2210 2211 |
sdg->cpu_power_orig = power; if (sched_feat(ARCH_POWER)) power *= arch_scale_freq_power(sd, cpu); else power *= default_scale_freq_power(sd, cpu); power >>= SCHED_LOAD_SHIFT; |
1e3c88bde sched: Move load ... |
2212 2213 2214 2215 2216 |
power *= scale_rt_power(cpu); power >>= SCHED_LOAD_SHIFT; if (!power) power = 1; |
e51fd5e22 sched: Fix wake_a... |
2217 |
cpu_rq(cpu)->cpu_power = power; |
1e3c88bde sched: Move load ... |
2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 |
sdg->cpu_power = power; } static void update_group_power(struct sched_domain *sd, int cpu) { struct sched_domain *child = sd->child; struct sched_group *group, *sdg = sd->groups; unsigned long power; if (!child) { update_cpu_power(sd, cpu); return; } power = 0; group = child->groups; do { power += group->cpu_power; group = group->next; } while (group != child->groups); sdg->cpu_power = power; } |
9d5efe05e sched: Fix capaci... |
2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 |
/* * Try and fix up capacity for tiny siblings, this is needed when * things like SD_ASYM_PACKING need f_b_g to select another sibling * which on its own isn't powerful enough. * * See update_sd_pick_busiest() and check_asym_packing(). */ static inline int fix_small_capacity(struct sched_domain *sd, struct sched_group *group) { /* * Only siblings can have significantly less than SCHED_LOAD_SCALE */ if (sd->level != SD_LV_SIBLING) return 0; /* * If ~90% of the cpu_power is still there, we're good. */ |
694f5a111 sched: Fix fix_sm... |
2261 |
if (group->cpu_power * 32 > group->cpu_power_orig * 29) |
9d5efe05e sched: Fix capaci... |
2262 2263 2264 2265 |
return 1; return 0; } |
1e3c88bde sched: Move load ... |
2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 |
/** * update_sg_lb_stats - Update sched_group's statistics for load balancing. * @sd: The sched_domain whose statistics are to be updated. * @group: sched_group whose statistics are to be updated. * @this_cpu: Cpu for which load balance is currently performed. * @idle: Idle status of this_cpu * @load_idx: Load index of sched_domain of this_cpu for load calc. * @sd_idle: Idle status of the sched_domain containing group. * @local_group: Does group contain this_cpu. * @cpus: Set of cpus considered for load balancing. * @balance: Should we balance. * @sgs: variable to hold the statistics for this group. */ static inline void update_sg_lb_stats(struct sched_domain *sd, struct sched_group *group, int this_cpu, enum cpu_idle_type idle, int load_idx, int *sd_idle, int local_group, const struct cpumask *cpus, int *balance, struct sg_lb_stats *sgs) { |
2582f0eba sched: Set group_... |
2285 |
unsigned long load, max_cpu_load, min_cpu_load, max_nr_running; |
1e3c88bde sched: Move load ... |
2286 2287 |
int i; unsigned int balance_cpu = -1, first_idle_cpu = 0; |
dd5feea14 sched: Fix SCHED_... |
2288 |
unsigned long avg_load_per_task = 0; |
1e3c88bde sched: Move load ... |
2289 |
|
871e35bc9 sched: Fix the pl... |
2290 |
if (local_group) |
1e3c88bde sched: Move load ... |
2291 |
balance_cpu = group_first_cpu(group); |
1e3c88bde sched: Move load ... |
2292 2293 |
/* Tally up the load of all CPUs in the group */ |
1e3c88bde sched: Move load ... |
2294 2295 |
max_cpu_load = 0; min_cpu_load = ~0UL; |
2582f0eba sched: Set group_... |
2296 |
max_nr_running = 0; |
1e3c88bde sched: Move load ... |
2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 |
for_each_cpu_and(i, sched_group_cpus(group), cpus) { struct rq *rq = cpu_rq(i); if (*sd_idle && rq->nr_running) *sd_idle = 0; /* Bias balancing toward cpus of our domain */ if (local_group) { if (idle_cpu(i) && !first_idle_cpu) { first_idle_cpu = 1; balance_cpu = i; } load = target_load(i, load_idx); } else { load = source_load(i, load_idx); |
2582f0eba sched: Set group_... |
2314 |
if (load > max_cpu_load) { |
1e3c88bde sched: Move load ... |
2315 |
max_cpu_load = load; |
2582f0eba sched: Set group_... |
2316 2317 |
max_nr_running = rq->nr_running; } |
1e3c88bde sched: Move load ... |
2318 2319 2320 2321 2322 2323 2324 |
if (min_cpu_load > load) min_cpu_load = load; } sgs->group_load += load; sgs->sum_nr_running += rq->nr_running; sgs->sum_weighted_load += weighted_cpuload(i); |
1e3c88bde sched: Move load ... |
2325 2326 2327 2328 2329 2330 2331 2332 |
} /* * First idle cpu or the first cpu(busiest) in this sched group * is eligible for doing load balancing at this and above * domains. In the newly idle case, we will allow all the cpu's * to do the newly idle load balance. */ |
bbc8cb5ba sched: Reduce upd... |
2333 2334 2335 2336 2337 2338 |
if (idle != CPU_NEWLY_IDLE && local_group) { if (balance_cpu != this_cpu) { *balance = 0; return; } update_group_power(sd, this_cpu); |
1e3c88bde sched: Move load ... |
2339 2340 2341 2342 |
} /* Adjust by relative CPU power of the group */ sgs->avg_load = (sgs->group_load * SCHED_LOAD_SCALE) / group->cpu_power; |
1e3c88bde sched: Move load ... |
2343 2344 2345 2346 2347 2348 2349 2350 2351 |
/* * Consider the group unbalanced when the imbalance is larger * than the average weight of two tasks. * * APZ: with cgroup the avg task weight can vary wildly and * might not be a suitable number - should we keep a * normalized nr_running number somewhere that negates * the hierarchy? */ |
dd5feea14 sched: Fix SCHED_... |
2352 2353 |
if (sgs->sum_nr_running) avg_load_per_task = sgs->sum_weighted_load / sgs->sum_nr_running; |
1e3c88bde sched: Move load ... |
2354 |
|
2582f0eba sched: Set group_... |
2355 |
if ((max_cpu_load - min_cpu_load) > 2*avg_load_per_task && max_nr_running > 1) |
1e3c88bde sched: Move load ... |
2356 |
sgs->group_imb = 1; |
2582f0eba sched: Set group_... |
2357 |
sgs->group_capacity = DIV_ROUND_CLOSEST(group->cpu_power, SCHED_LOAD_SCALE); |
9d5efe05e sched: Fix capaci... |
2358 2359 |
if (!sgs->group_capacity) sgs->group_capacity = fix_small_capacity(sd, group); |
fab476228 sched: Force bala... |
2360 2361 2362 |
if (sgs->group_capacity > sgs->sum_nr_running) sgs->group_has_capacity = 1; |
1e3c88bde sched: Move load ... |
2363 2364 2365 |
} /** |
532cb4c40 sched: Add asymme... |
2366 2367 2368 2369 |
* update_sd_pick_busiest - return 1 on busiest group * @sd: sched_domain whose statistics are to be checked * @sds: sched_domain statistics * @sg: sched_group candidate to be checked for being the busiest |
b6b122944 sched: Fix commen... |
2370 2371 |
* @sgs: sched_group statistics * @this_cpu: the current cpu |
532cb4c40 sched: Add asymme... |
2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 |
* * Determine if @sg is a busier group than the previously selected * busiest group. */ static bool update_sd_pick_busiest(struct sched_domain *sd, struct sd_lb_stats *sds, struct sched_group *sg, struct sg_lb_stats *sgs, int this_cpu) { if (sgs->avg_load <= sds->max_load) return false; if (sgs->sum_nr_running > sgs->group_capacity) return true; if (sgs->group_imb) return true; /* * ASYM_PACKING needs to move all the work to the lowest * numbered CPUs in the group, therefore mark all groups * higher than ourself as busy. */ if ((sd->flags & SD_ASYM_PACKING) && sgs->sum_nr_running && this_cpu < group_first_cpu(sg)) { if (!sds->busiest) return true; if (group_first_cpu(sds->busiest) > group_first_cpu(sg)) return true; } return false; } /** |
1e3c88bde sched: Move load ... |
2409 2410 2411 2412 |
* update_sd_lb_stats - Update sched_group's statistics for load balancing. * @sd: sched_domain whose statistics are to be updated. * @this_cpu: Cpu for which load balance is currently performed. * @idle: Idle status of this_cpu |
532cb4c40 sched: Add asymme... |
2413 |
* @sd_idle: Idle status of the sched_domain containing sg. |
1e3c88bde sched: Move load ... |
2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 |
* @cpus: Set of cpus considered for load balancing. * @balance: Should we balance. * @sds: variable to hold the statistics for this sched_domain. */ static inline void update_sd_lb_stats(struct sched_domain *sd, int this_cpu, enum cpu_idle_type idle, int *sd_idle, const struct cpumask *cpus, int *balance, struct sd_lb_stats *sds) { struct sched_domain *child = sd->child; |
532cb4c40 sched: Add asymme... |
2424 |
struct sched_group *sg = sd->groups; |
1e3c88bde sched: Move load ... |
2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 |
struct sg_lb_stats sgs; int load_idx, prefer_sibling = 0; if (child && child->flags & SD_PREFER_SIBLING) prefer_sibling = 1; init_sd_power_savings_stats(sd, sds, idle); load_idx = get_sd_load_idx(sd, idle); do { int local_group; |
532cb4c40 sched: Add asymme... |
2436 |
local_group = cpumask_test_cpu(this_cpu, sched_group_cpus(sg)); |
1e3c88bde sched: Move load ... |
2437 |
memset(&sgs, 0, sizeof(sgs)); |
532cb4c40 sched: Add asymme... |
2438 |
update_sg_lb_stats(sd, sg, this_cpu, idle, load_idx, sd_idle, |
1e3c88bde sched: Move load ... |
2439 |
local_group, cpus, balance, &sgs); |
8f190fb3f sched: Assume *ba... |
2440 |
if (local_group && !(*balance)) |
1e3c88bde sched: Move load ... |
2441 2442 2443 |
return; sds->total_load += sgs.group_load; |
532cb4c40 sched: Add asymme... |
2444 |
sds->total_pwr += sg->cpu_power; |
1e3c88bde sched: Move load ... |
2445 2446 2447 |
/* * In case the child domain prefers tasks go to siblings |
532cb4c40 sched: Add asymme... |
2448 |
* first, lower the sg capacity to one so that we'll try |
75dd321d7 sched: Drop group... |
2449 2450 2451 2452 2453 2454 |
* and move all the excess tasks away. We lower the capacity * of a group only if the local group has the capacity to fit * these excess tasks, i.e. nr_running < group_capacity. The * extra check prevents the case where you always pull from the * heaviest group when it is already under-utilized (possible * with a large weight task outweighs the tasks on the system). |
1e3c88bde sched: Move load ... |
2455 |
*/ |
75dd321d7 sched: Drop group... |
2456 |
if (prefer_sibling && !local_group && sds->this_has_capacity) |
1e3c88bde sched: Move load ... |
2457 2458 2459 2460 |
sgs.group_capacity = min(sgs.group_capacity, 1UL); if (local_group) { sds->this_load = sgs.avg_load; |
532cb4c40 sched: Add asymme... |
2461 |
sds->this = sg; |
1e3c88bde sched: Move load ... |
2462 2463 |
sds->this_nr_running = sgs.sum_nr_running; sds->this_load_per_task = sgs.sum_weighted_load; |
fab476228 sched: Force bala... |
2464 |
sds->this_has_capacity = sgs.group_has_capacity; |
532cb4c40 sched: Add asymme... |
2465 |
} else if (update_sd_pick_busiest(sd, sds, sg, &sgs, this_cpu)) { |
1e3c88bde sched: Move load ... |
2466 |
sds->max_load = sgs.avg_load; |
532cb4c40 sched: Add asymme... |
2467 |
sds->busiest = sg; |
1e3c88bde sched: Move load ... |
2468 |
sds->busiest_nr_running = sgs.sum_nr_running; |
dd5feea14 sched: Fix SCHED_... |
2469 |
sds->busiest_group_capacity = sgs.group_capacity; |
1e3c88bde sched: Move load ... |
2470 |
sds->busiest_load_per_task = sgs.sum_weighted_load; |
fab476228 sched: Force bala... |
2471 |
sds->busiest_has_capacity = sgs.group_has_capacity; |
1e3c88bde sched: Move load ... |
2472 2473 |
sds->group_imb = sgs.group_imb; } |
532cb4c40 sched: Add asymme... |
2474 2475 2476 2477 |
update_sd_power_savings_stats(sg, sds, local_group, &sgs); sg = sg->next; } while (sg != sd->groups); } |
2ec57d448 sched: Fix spelli... |
2478 |
int __weak arch_sd_sibling_asym_packing(void) |
532cb4c40 sched: Add asymme... |
2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 |
{ return 0*SD_ASYM_PACKING; } /** * check_asym_packing - Check to see if the group is packed into the * sched doman. * * This is primarily intended to used at the sibling level. Some * cores like POWER7 prefer to use lower numbered SMT threads. In the * case of POWER7, it can move to lower SMT modes only when higher * threads are idle. When in lower SMT modes, the threads will * perform better since they share less core resources. Hence when we * have idle threads, we want them to be the higher ones. * * This packing function is run on idle threads. It checks to see if * the busiest CPU in this domain (core in the P7 case) has a higher * CPU number than the packing function is being run on. Here we are * assuming lower CPU number will be equivalent to lower a SMT thread * number. * |
b6b122944 sched: Fix commen... |
2500 2501 2502 |
* Returns 1 when packing is required and a task should be moved to * this CPU. The amount of the imbalance is returned in *imbalance. * |
532cb4c40 sched: Add asymme... |
2503 2504 2505 2506 |
* @sd: The sched_domain whose packing is to be checked. * @sds: Statistics of the sched_domain which is to be packed * @this_cpu: The cpu at whose sched_domain we're performing load-balance. * @imbalance: returns amount of imbalanced due to packing. |
532cb4c40 sched: Add asymme... |
2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 |
*/ static int check_asym_packing(struct sched_domain *sd, struct sd_lb_stats *sds, int this_cpu, unsigned long *imbalance) { int busiest_cpu; if (!(sd->flags & SD_ASYM_PACKING)) return 0; if (!sds->busiest) return 0; busiest_cpu = group_first_cpu(sds->busiest); if (this_cpu > busiest_cpu) return 0; *imbalance = DIV_ROUND_CLOSEST(sds->max_load * sds->busiest->cpu_power, SCHED_LOAD_SCALE); return 1; |
1e3c88bde sched: Move load ... |
2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 |
} /** * fix_small_imbalance - Calculate the minor imbalance that exists * amongst the groups of a sched_domain, during * load balancing. * @sds: Statistics of the sched_domain whose imbalance is to be calculated. * @this_cpu: The cpu at whose sched_domain we're performing load-balance. * @imbalance: Variable to store the imbalance. */ static inline void fix_small_imbalance(struct sd_lb_stats *sds, int this_cpu, unsigned long *imbalance) { unsigned long tmp, pwr_now = 0, pwr_move = 0; unsigned int imbn = 2; |
dd5feea14 sched: Fix SCHED_... |
2542 |
unsigned long scaled_busy_load_per_task; |
1e3c88bde sched: Move load ... |
2543 2544 2545 2546 2547 2548 2549 2550 2551 |
if (sds->this_nr_running) { sds->this_load_per_task /= sds->this_nr_running; if (sds->busiest_load_per_task > sds->this_load_per_task) imbn = 1; } else sds->this_load_per_task = cpu_avg_load_per_task(this_cpu); |
dd5feea14 sched: Fix SCHED_... |
2552 2553 2554 2555 2556 2557 |
scaled_busy_load_per_task = sds->busiest_load_per_task * SCHED_LOAD_SCALE; scaled_busy_load_per_task /= sds->busiest->cpu_power; if (sds->max_load - sds->this_load + scaled_busy_load_per_task >= (scaled_busy_load_per_task * imbn)) { |
1e3c88bde sched: Move load ... |
2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 |
*imbalance = sds->busiest_load_per_task; return; } /* * OK, we don't have enough imbalance to justify moving tasks, * however we may be able to increase total CPU power used by * moving them. */ pwr_now += sds->busiest->cpu_power * min(sds->busiest_load_per_task, sds->max_load); pwr_now += sds->this->cpu_power * min(sds->this_load_per_task, sds->this_load); pwr_now /= SCHED_LOAD_SCALE; /* Amount of load we'd subtract */ tmp = (sds->busiest_load_per_task * SCHED_LOAD_SCALE) / sds->busiest->cpu_power; if (sds->max_load > tmp) pwr_move += sds->busiest->cpu_power * min(sds->busiest_load_per_task, sds->max_load - tmp); /* Amount of load we'd add */ if (sds->max_load * sds->busiest->cpu_power < sds->busiest_load_per_task * SCHED_LOAD_SCALE) tmp = (sds->max_load * sds->busiest->cpu_power) / sds->this->cpu_power; else tmp = (sds->busiest_load_per_task * SCHED_LOAD_SCALE) / sds->this->cpu_power; pwr_move += sds->this->cpu_power * min(sds->this_load_per_task, sds->this_load + tmp); pwr_move /= SCHED_LOAD_SCALE; /* Move if we gain throughput */ if (pwr_move > pwr_now) *imbalance = sds->busiest_load_per_task; } /** * calculate_imbalance - Calculate the amount of imbalance present within the * groups of a given sched_domain during load balance. * @sds: statistics of the sched_domain whose imbalance is to be calculated. * @this_cpu: Cpu for which currently load balance is being performed. * @imbalance: The variable to store the imbalance. */ static inline void calculate_imbalance(struct sd_lb_stats *sds, int this_cpu, unsigned long *imbalance) { |
dd5feea14 sched: Fix SCHED_... |
2608 2609 2610 2611 2612 2613 2614 |
unsigned long max_pull, load_above_capacity = ~0UL; sds->busiest_load_per_task /= sds->busiest_nr_running; if (sds->group_imb) { sds->busiest_load_per_task = min(sds->busiest_load_per_task, sds->avg_load); } |
1e3c88bde sched: Move load ... |
2615 2616 2617 2618 2619 2620 2621 2622 2623 |
/* * In the presence of smp nice balancing, certain scenarios can have * max load less than avg load(as we skip the groups at or below * its cpu_power, while calculating max_load..) */ if (sds->max_load < sds->avg_load) { *imbalance = 0; return fix_small_imbalance(sds, this_cpu, imbalance); } |
dd5feea14 sched: Fix SCHED_... |
2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 |
if (!sds->group_imb) { /* * Don't want to pull so many tasks that a group would go idle. */ load_above_capacity = (sds->busiest_nr_running - sds->busiest_group_capacity); load_above_capacity *= (SCHED_LOAD_SCALE * SCHED_LOAD_SCALE); load_above_capacity /= sds->busiest->cpu_power; } /* * We're trying to get all the cpus to the average_load, so we don't * want to push ourselves above the average load, nor do we wish to * reduce the max loaded cpu below the average load. At the same time, * we also don't want to reduce the group load below the group capacity * (so that we can implement power-savings policies etc). Thus we look * for the minimum possible imbalance. * Be careful of negative numbers as they'll appear as very large values * with unsigned longs. */ max_pull = min(sds->max_load - sds->avg_load, load_above_capacity); |
1e3c88bde sched: Move load ... |
2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 |
/* How much load to actually move to equalise the imbalance */ *imbalance = min(max_pull * sds->busiest->cpu_power, (sds->avg_load - sds->this_load) * sds->this->cpu_power) / SCHED_LOAD_SCALE; /* * if *imbalance is less than the average load per runnable task * there is no gaurantee that any tasks will be moved so we'll have * a think about bumping its value to force at least one task to be * moved */ if (*imbalance < sds->busiest_load_per_task) return fix_small_imbalance(sds, this_cpu, imbalance); } |
fab476228 sched: Force bala... |
2663 |
|
1e3c88bde sched: Move load ... |
2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 |
/******* find_busiest_group() helpers end here *********************/ /** * find_busiest_group - Returns the busiest group within the sched_domain * if there is an imbalance. If there isn't an imbalance, and * the user has opted for power-savings, it returns a group whose * CPUs can be put to idle by rebalancing those tasks elsewhere, if * such a group exists. * * Also calculates the amount of weighted load which should be moved * to restore balance. * * @sd: The sched_domain whose busiest group is to be returned. * @this_cpu: The cpu for which load balancing is currently being performed. * @imbalance: Variable which stores amount of weighted load which should * be moved to restore balance/put a group to idle. * @idle: The idle status of this_cpu. * @sd_idle: The idleness of sd * @cpus: The set of CPUs under consideration for load-balancing. * @balance: Pointer to a variable indicating if this_cpu * is the appropriate cpu to perform load balancing at this_level. * * Returns: - the busiest group if imbalance exists. * - If no imbalance and user has opted for power-savings balance, * return the least loaded group whose CPUs can be * put to idle by rebalancing its tasks onto our group. */ static struct sched_group * find_busiest_group(struct sched_domain *sd, int this_cpu, unsigned long *imbalance, enum cpu_idle_type idle, int *sd_idle, const struct cpumask *cpus, int *balance) { struct sd_lb_stats sds; memset(&sds, 0, sizeof(sds)); /* * Compute the various statistics relavent for load balancing at * this level. */ update_sd_lb_stats(sd, this_cpu, idle, sd_idle, cpus, balance, &sds); /* Cases where imbalance does not exist from POV of this_cpu */ /* 1) this_cpu is not the appropriate cpu to perform load balancing * at this level. * 2) There is no busy sibling group to pull from. * 3) This group is the busiest group. * 4) This group is more busy than the avg busieness at this * sched_domain. * 5) The imbalance is within the specified limit. |
fab476228 sched: Force bala... |
2715 2716 2717 2718 2719 |
* * Note: when doing newidle balance, if the local group has excess * capacity (i.e. nr_running < group_capacity) and the busiest group * does not have any capacity, we force a load balance to pull tasks * to the local group. In this case, we skip past checks 3, 4 and 5. |
1e3c88bde sched: Move load ... |
2720 |
*/ |
8f190fb3f sched: Assume *ba... |
2721 |
if (!(*balance)) |
1e3c88bde sched: Move load ... |
2722 |
goto ret; |
532cb4c40 sched: Add asymme... |
2723 2724 2725 |
if ((idle == CPU_IDLE || idle == CPU_NEWLY_IDLE) && check_asym_packing(sd, &sds, this_cpu, imbalance)) return sds.busiest; |
1e3c88bde sched: Move load ... |
2726 2727 |
if (!sds.busiest || sds.busiest_nr_running == 0) goto out_balanced; |
fab476228 sched: Force bala... |
2728 2729 2730 2731 |
/* SD_BALANCE_NEWIDLE trumps SMP nice when underutilized */ if (idle == CPU_NEWLY_IDLE && sds.this_has_capacity && !sds.busiest_has_capacity) goto force_balance; |
1e3c88bde sched: Move load ... |
2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 |
if (sds.this_load >= sds.max_load) goto out_balanced; sds.avg_load = (SCHED_LOAD_SCALE * sds.total_load) / sds.total_pwr; if (sds.this_load >= sds.avg_load) goto out_balanced; if (100 * sds.max_load <= sd->imbalance_pct * sds.this_load) goto out_balanced; |
fab476228 sched: Force bala... |
2742 |
force_balance: |
1e3c88bde sched: Move load ... |
2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 |
/* Looks like there is an imbalance. Compute it */ calculate_imbalance(&sds, this_cpu, imbalance); return sds.busiest; out_balanced: /* * There is no obvious imbalance. But check if we can do some balancing * to save power. */ if (check_power_save_busiest_group(&sds, this_cpu, imbalance)) return sds.busiest; ret: *imbalance = 0; return NULL; } /* * find_busiest_queue - find the busiest runqueue among the cpus in group. */ static struct rq * |
9d5efe05e sched: Fix capaci... |
2763 2764 2765 |
find_busiest_queue(struct sched_domain *sd, struct sched_group *group, enum cpu_idle_type idle, unsigned long imbalance, const struct cpumask *cpus) |
1e3c88bde sched: Move load ... |
2766 2767 2768 2769 2770 2771 2772 2773 2774 |
{ struct rq *busiest = NULL, *rq; unsigned long max_load = 0; int i; for_each_cpu(i, sched_group_cpus(group)) { unsigned long power = power_of(i); unsigned long capacity = DIV_ROUND_CLOSEST(power, SCHED_LOAD_SCALE); unsigned long wl; |
9d5efe05e sched: Fix capaci... |
2775 2776 |
if (!capacity) capacity = fix_small_capacity(sd, group); |
1e3c88bde sched: Move load ... |
2777 2778 2779 2780 |
if (!cpumask_test_cpu(i, cpus)) continue; rq = cpu_rq(i); |
6e40f5bbb Merge branch 'sch... |
2781 |
wl = weighted_cpuload(i); |
1e3c88bde sched: Move load ... |
2782 |
|
6e40f5bbb Merge branch 'sch... |
2783 2784 2785 2786 |
/* * When comparing with imbalance, use weighted_cpuload() * which is not scaled with the cpu power. */ |
1e3c88bde sched: Move load ... |
2787 2788 |
if (capacity && rq->nr_running == 1 && wl > imbalance) continue; |
6e40f5bbb Merge branch 'sch... |
2789 2790 2791 2792 2793 2794 2795 |
/* * For the load comparisons with the other cpu's, consider * the weighted_cpuload() scaled with the cpu power, so that * the load can be moved away from the cpu that is potentially * running at a lower capacity. */ wl = (wl * SCHED_LOAD_SCALE) / power; |
1e3c88bde sched: Move load ... |
2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 |
if (wl > max_load) { max_load = wl; busiest = rq; } } return busiest; } /* * Max backoff if we encounter pinned tasks. Pretty arbitrary value, but * so long as it is large enough. */ #define MAX_PINNED_INTERVAL 512 /* Working cpumask for load_balance and load_balance_newidle. */ static DEFINE_PER_CPU(cpumask_var_t, load_balance_tmpmask); |
532cb4c40 sched: Add asymme... |
2813 2814 |
static int need_active_balance(struct sched_domain *sd, int sd_idle, int idle, int busiest_cpu, int this_cpu) |
1af3ed3dd sched: Unify load... |
2815 2816 |
{ if (idle == CPU_NEWLY_IDLE) { |
532cb4c40 sched: Add asymme... |
2817 2818 2819 2820 2821 2822 2823 2824 |
/* * ASYM_PACKING needs to force migrate tasks from busy but * higher numbered CPUs in order to pack all tasks in the * lowest numbered CPUs. */ if ((sd->flags & SD_ASYM_PACKING) && busiest_cpu > this_cpu) return 1; |
1af3ed3dd sched: Unify load... |
2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 |
/* * The only task running in a non-idle cpu can be moved to this * cpu in an attempt to completely freeup the other CPU * package. * * The package power saving logic comes from * find_busiest_group(). If there are no imbalance, then * f_b_g() will return NULL. However when sched_mc={1,2} then * f_b_g() will select a group from which a running task may be * pulled to this cpu in order to make the other package idle. * If there is no opportunity to make a package idle and if * there are no imbalance, then f_b_g() will return NULL and no * action will be taken in load_balance_newidle(). * * Under normal task pull operation due to imbalance, there * will be more than one task in the source run queue and * move_tasks() will succeed. ld_moved will be true and this * active balance code will not be triggered. */ if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) return 0; if (sched_mc_power_savings < POWERSAVINGS_BALANCE_WAKEUP) return 0; } return unlikely(sd->nr_balance_failed > sd->cache_nice_tries+2); } |
969c79215 sched: replace mi... |
2854 |
static int active_load_balance_cpu_stop(void *data); |
1e3c88bde sched: Move load ... |
2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 |
/* * Check this_cpu to ensure it is balanced within domain. Attempt to move * tasks if there is an imbalance. */ static int load_balance(int this_cpu, struct rq *this_rq, struct sched_domain *sd, enum cpu_idle_type idle, int *balance) { int ld_moved, all_pinned = 0, active_balance = 0, sd_idle = 0; struct sched_group *group; unsigned long imbalance; struct rq *busiest; unsigned long flags; struct cpumask *cpus = __get_cpu_var(load_balance_tmpmask); cpumask_copy(cpus, cpu_active_mask); /* * When power savings policy is enabled for the parent domain, idle * sibling can pick up load irrespective of busy siblings. In this case, * let the state of idle sibling percolate up as CPU_IDLE, instead of * portraying it as CPU_NOT_IDLE. */ if (idle != CPU_NOT_IDLE && sd->flags & SD_SHARE_CPUPOWER && !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) sd_idle = 1; schedstat_inc(sd, lb_count[idle]); redo: update_shares(sd); group = find_busiest_group(sd, this_cpu, &imbalance, idle, &sd_idle, cpus, balance); if (*balance == 0) goto out_balanced; if (!group) { schedstat_inc(sd, lb_nobusyg[idle]); goto out_balanced; } |
9d5efe05e sched: Fix capaci... |
2896 |
busiest = find_busiest_queue(sd, group, idle, imbalance, cpus); |
1e3c88bde sched: Move load ... |
2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 |
if (!busiest) { schedstat_inc(sd, lb_nobusyq[idle]); goto out_balanced; } BUG_ON(busiest == this_rq); schedstat_add(sd, lb_imbalance[idle], imbalance); ld_moved = 0; if (busiest->nr_running > 1) { /* * Attempt to move tasks. If find_busiest_group has found * an imbalance but busiest->nr_running <= 1, the group is * still unbalanced. ld_moved simply stays zero, so it is * correctly treated as an imbalance. */ local_irq_save(flags); double_rq_lock(this_rq, busiest); ld_moved = move_tasks(this_rq, this_cpu, busiest, imbalance, sd, idle, &all_pinned); double_rq_unlock(this_rq, busiest); local_irq_restore(flags); /* * some other cpu did the load balance for us. */ if (ld_moved && this_cpu != smp_processor_id()) resched_cpu(this_cpu); /* All tasks on this runqueue were pinned by CPU affinity */ if (unlikely(all_pinned)) { cpumask_clear_cpu(cpu_of(busiest), cpus); if (!cpumask_empty(cpus)) goto redo; goto out_balanced; } } if (!ld_moved) { schedstat_inc(sd, lb_failed[idle]); |
58b26c4c0 sched: Increment ... |
2938 2939 2940 2941 2942 2943 2944 2945 |
/* * Increment the failure counter only on periodic balance. * We do not want newidle balance, which can be very * frequent, pollute the failure counter causing * excessive cache_hot migrations and active balances. */ if (idle != CPU_NEWLY_IDLE) sd->nr_balance_failed++; |
1e3c88bde sched: Move load ... |
2946 |
|
532cb4c40 sched: Add asymme... |
2947 2948 |
if (need_active_balance(sd, sd_idle, idle, cpu_of(busiest), this_cpu)) { |
1e3c88bde sched: Move load ... |
2949 |
raw_spin_lock_irqsave(&busiest->lock, flags); |
969c79215 sched: replace mi... |
2950 2951 2952 |
/* don't kick the active_load_balance_cpu_stop, * if the curr task on busiest cpu can't be * moved to this_cpu |
1e3c88bde sched: Move load ... |
2953 2954 2955 2956 2957 2958 2959 2960 |
*/ if (!cpumask_test_cpu(this_cpu, &busiest->curr->cpus_allowed)) { raw_spin_unlock_irqrestore(&busiest->lock, flags); all_pinned = 1; goto out_one_pinned; } |
969c79215 sched: replace mi... |
2961 2962 2963 2964 2965 |
/* * ->active_balance synchronizes accesses to * ->active_balance_work. Once set, it's cleared * only after active load balance is finished. */ |
1e3c88bde sched: Move load ... |
2966 2967 2968 2969 2970 2971 |
if (!busiest->active_balance) { busiest->active_balance = 1; busiest->push_cpu = this_cpu; active_balance = 1; } raw_spin_unlock_irqrestore(&busiest->lock, flags); |
969c79215 sched: replace mi... |
2972 |
|
1e3c88bde sched: Move load ... |
2973 |
if (active_balance) |
969c79215 sched: replace mi... |
2974 2975 2976 |
stop_one_cpu_nowait(cpu_of(busiest), active_load_balance_cpu_stop, busiest, &busiest->active_balance_work); |
1e3c88bde sched: Move load ... |
2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 |
/* * We've kicked active balancing, reset the failure * counter. */ sd->nr_balance_failed = sd->cache_nice_tries+1; } } else sd->nr_balance_failed = 0; if (likely(!active_balance)) { /* We were unbalanced, so reset the balancing interval */ sd->balance_interval = sd->min_interval; } else { /* * If we've begun active balancing, start to back off. This * case may not be covered by the all_pinned logic if there * is only 1 task on the busy runqueue (because we don't call * move_tasks). */ if (sd->balance_interval < sd->max_interval) sd->balance_interval *= 2; } if (!ld_moved && !sd_idle && sd->flags & SD_SHARE_CPUPOWER && !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) ld_moved = -1; goto out; out_balanced: schedstat_inc(sd, lb_balanced[idle]); sd->nr_balance_failed = 0; out_one_pinned: /* tune up the balancing interval */ if ((all_pinned && sd->balance_interval < MAX_PINNED_INTERVAL) || (sd->balance_interval < sd->max_interval)) sd->balance_interval *= 2; if (!sd_idle && sd->flags & SD_SHARE_CPUPOWER && !test_sd_parent(sd, SD_POWERSAVINGS_BALANCE)) ld_moved = -1; else ld_moved = 0; out: if (ld_moved) update_shares(sd); return ld_moved; } /* |
1e3c88bde sched: Move load ... |
3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 |
* idle_balance is called by schedule() if this_cpu is about to become * idle. Attempts to pull tasks from other CPUs. */ static void idle_balance(int this_cpu, struct rq *this_rq) { struct sched_domain *sd; int pulled_task = 0; unsigned long next_balance = jiffies + HZ; this_rq->idle_stamp = this_rq->clock; if (this_rq->avg_idle < sysctl_sched_migration_cost) return; |
f492e12ef sched: Remove loa... |
3043 3044 3045 3046 |
/* * Drop the rq->lock, but keep IRQ/preempt disabled. */ raw_spin_unlock(&this_rq->lock); |
1e3c88bde sched: Move load ... |
3047 3048 |
for_each_domain(this_cpu, sd) { unsigned long interval; |
f492e12ef sched: Remove loa... |
3049 |
int balance = 1; |
1e3c88bde sched: Move load ... |
3050 3051 3052 |
if (!(sd->flags & SD_LOAD_BALANCE)) continue; |
f492e12ef sched: Remove loa... |
3053 |
if (sd->flags & SD_BALANCE_NEWIDLE) { |
1e3c88bde sched: Move load ... |
3054 |
/* If we've pulled tasks over stop searching: */ |
f492e12ef sched: Remove loa... |
3055 3056 3057 |
pulled_task = load_balance(this_cpu, this_rq, sd, CPU_NEWLY_IDLE, &balance); } |
1e3c88bde sched: Move load ... |
3058 3059 3060 3061 |
interval = msecs_to_jiffies(sd->balance_interval); if (time_after(next_balance, sd->last_balance + interval)) next_balance = sd->last_balance + interval; |
fab476228 sched: Force bala... |
3062 |
if (pulled_task) |
1e3c88bde sched: Move load ... |
3063 |
break; |
1e3c88bde sched: Move load ... |
3064 |
} |
f492e12ef sched: Remove loa... |
3065 3066 |
raw_spin_lock(&this_rq->lock); |
1e3c88bde sched: Move load ... |
3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 |
if (pulled_task || time_after(jiffies, this_rq->next_balance)) { /* * We are going idle. next_balance may be set based on * a busy processor. So reset next_balance. */ this_rq->next_balance = next_balance; } } /* |
969c79215 sched: replace mi... |
3077 3078 3079 3080 |
* active_load_balance_cpu_stop is run by cpu stopper. It pushes * running tasks off the busiest CPU onto idle CPUs. It requires at * least 1 task to be running on each physical CPU where possible, and * avoids physical / logical imbalances. |
1e3c88bde sched: Move load ... |
3081 |
*/ |
969c79215 sched: replace mi... |
3082 |
static int active_load_balance_cpu_stop(void *data) |
1e3c88bde sched: Move load ... |
3083 |
{ |
969c79215 sched: replace mi... |
3084 3085 |
struct rq *busiest_rq = data; int busiest_cpu = cpu_of(busiest_rq); |
1e3c88bde sched: Move load ... |
3086 |
int target_cpu = busiest_rq->push_cpu; |
969c79215 sched: replace mi... |
3087 |
struct rq *target_rq = cpu_rq(target_cpu); |
1e3c88bde sched: Move load ... |
3088 |
struct sched_domain *sd; |
969c79215 sched: replace mi... |
3089 3090 3091 3092 3093 3094 3095 |
raw_spin_lock_irq(&busiest_rq->lock); /* make sure the requested cpu hasn't gone down in the meantime */ if (unlikely(busiest_cpu != smp_processor_id() || !busiest_rq->active_balance)) goto out_unlock; |
1e3c88bde sched: Move load ... |
3096 3097 3098 |
/* Is there any task to move? */ if (busiest_rq->nr_running <= 1) |
969c79215 sched: replace mi... |
3099 |
goto out_unlock; |
1e3c88bde sched: Move load ... |
3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 |
/* * This condition is "impossible", if it occurs * we need to fix it. Originally reported by * Bjorn Helgaas on a 128-cpu setup. */ BUG_ON(busiest_rq == target_rq); /* move a task from busiest_rq to target_rq */ double_lock_balance(busiest_rq, target_rq); |
1e3c88bde sched: Move load ... |
3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 |
/* Search for an sd spanning us and the target CPU. */ for_each_domain(target_cpu, sd) { if ((sd->flags & SD_LOAD_BALANCE) && cpumask_test_cpu(busiest_cpu, sched_domain_span(sd))) break; } if (likely(sd)) { schedstat_inc(sd, alb_count); if (move_one_task(target_rq, target_cpu, busiest_rq, sd, CPU_IDLE)) schedstat_inc(sd, alb_pushed); else schedstat_inc(sd, alb_failed); } double_unlock_balance(busiest_rq, target_rq); |
969c79215 sched: replace mi... |
3128 3129 3130 3131 |
out_unlock: busiest_rq->active_balance = 0; raw_spin_unlock_irq(&busiest_rq->lock); return 0; |
1e3c88bde sched: Move load ... |
3132 3133 3134 |
} #ifdef CONFIG_NO_HZ |
83cd4fe27 sched: Change noh... |
3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 |
static DEFINE_PER_CPU(struct call_single_data, remote_sched_softirq_cb); static void trigger_sched_softirq(void *data) { raise_softirq_irqoff(SCHED_SOFTIRQ); } static inline void init_sched_softirq_csd(struct call_single_data *csd) { csd->func = trigger_sched_softirq; csd->info = NULL; csd->flags = 0; csd->priv = 0; } /* * idle load balancing details * - One of the idle CPUs nominates itself as idle load_balancer, while * entering idle. * - This idle load balancer CPU will also go into tickless mode when * it is idle, just like all other idle CPUs * - When one of the busy CPUs notice that there may be an idle rebalancing * needed, they will kick the idle load balancer, which then does idle * load balancing for all the idle CPUs. */ |
1e3c88bde sched: Move load ... |
3161 3162 |
static struct { atomic_t load_balancer; |
83cd4fe27 sched: Change noh... |
3163 3164 3165 3166 3167 3168 |
atomic_t first_pick_cpu; atomic_t second_pick_cpu; cpumask_var_t idle_cpus_mask; cpumask_var_t grp_idle_mask; unsigned long next_balance; /* in jiffy units */ } nohz ____cacheline_aligned; |
1e3c88bde sched: Move load ... |
3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221 |
int get_nohz_load_balancer(void) { return atomic_read(&nohz.load_balancer); } #if defined(CONFIG_SCHED_MC) || defined(CONFIG_SCHED_SMT) /** * lowest_flag_domain - Return lowest sched_domain containing flag. * @cpu: The cpu whose lowest level of sched domain is to * be returned. * @flag: The flag to check for the lowest sched_domain * for the given cpu. * * Returns the lowest sched_domain of a cpu which contains the given flag. */ static inline struct sched_domain *lowest_flag_domain(int cpu, int flag) { struct sched_domain *sd; for_each_domain(cpu, sd) if (sd && (sd->flags & flag)) break; return sd; } /** * for_each_flag_domain - Iterates over sched_domains containing the flag. * @cpu: The cpu whose domains we're iterating over. * @sd: variable holding the value of the power_savings_sd * for cpu. * @flag: The flag to filter the sched_domains to be iterated. * * Iterates over all the scheduler domains for a given cpu that has the 'flag' * set, starting from the lowest sched_domain to the highest. */ #define for_each_flag_domain(cpu, sd, flag) \ for (sd = lowest_flag_domain(cpu, flag); \ (sd && (sd->flags & flag)); sd = sd->parent) /** * is_semi_idle_group - Checks if the given sched_group is semi-idle. * @ilb_group: group to be checked for semi-idleness * * Returns: 1 if the group is semi-idle. 0 otherwise. * * We define a sched_group to be semi idle if it has atleast one idle-CPU * and atleast one non-idle CPU. This helper function checks if the given * sched_group is semi-idle or not. */ static inline int is_semi_idle_group(struct sched_group *ilb_group) { |
83cd4fe27 sched: Change noh... |
3222 |
cpumask_and(nohz.grp_idle_mask, nohz.idle_cpus_mask, |
1e3c88bde sched: Move load ... |
3223 3224 3225 3226 3227 3228 |
sched_group_cpus(ilb_group)); /* * A sched_group is semi-idle when it has atleast one busy cpu * and atleast one idle cpu. */ |
83cd4fe27 sched: Change noh... |
3229 |
if (cpumask_empty(nohz.grp_idle_mask)) |
1e3c88bde sched: Move load ... |
3230 |
return 0; |
83cd4fe27 sched: Change noh... |
3231 |
if (cpumask_equal(nohz.grp_idle_mask, sched_group_cpus(ilb_group))) |
1e3c88bde sched: Move load ... |
3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 |
return 0; return 1; } /** * find_new_ilb - Finds the optimum idle load balancer for nomination. * @cpu: The cpu which is nominating a new idle_load_balancer. * * Returns: Returns the id of the idle load balancer if it exists, * Else, returns >= nr_cpu_ids. * * This algorithm picks the idle load balancer such that it belongs to a * semi-idle powersavings sched_domain. The idea is to try and avoid * completely idle packages/cores just for the purpose of idle load balancing * when there are other idle cpu's which are better suited for that job. */ static int find_new_ilb(int cpu) { struct sched_domain *sd; struct sched_group *ilb_group; /* * Have idle load balancer selection from semi-idle packages only * when power-aware load balancing is enabled */ if (!(sched_smt_power_savings || sched_mc_power_savings)) goto out_done; /* * Optimize for the case when we have no idle CPUs or only one * idle CPU. Don't walk the sched_domain hierarchy in such cases */ |
83cd4fe27 sched: Change noh... |
3264 |
if (cpumask_weight(nohz.idle_cpus_mask) < 2) |
1e3c88bde sched: Move load ... |
3265 3266 3267 3268 3269 3270 3271 |
goto out_done; for_each_flag_domain(cpu, sd, SD_POWERSAVINGS_BALANCE) { ilb_group = sd->groups; do { if (is_semi_idle_group(ilb_group)) |
83cd4fe27 sched: Change noh... |
3272 |
return cpumask_first(nohz.grp_idle_mask); |
1e3c88bde sched: Move load ... |
3273 3274 3275 3276 3277 3278 3279 |
ilb_group = ilb_group->next; } while (ilb_group != sd->groups); } out_done: |
83cd4fe27 sched: Change noh... |
3280 |
return nr_cpu_ids; |
1e3c88bde sched: Move load ... |
3281 3282 3283 3284 |
} #else /* (CONFIG_SCHED_MC || CONFIG_SCHED_SMT) */ static inline int find_new_ilb(int call_cpu) { |
83cd4fe27 sched: Change noh... |
3285 |
return nr_cpu_ids; |
1e3c88bde sched: Move load ... |
3286 3287 3288 3289 |
} #endif /* |
83cd4fe27 sched: Change noh... |
3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 |
* Kick a CPU to do the nohz balancing, if it is time for it. We pick the * nohz_load_balancer CPU (if there is one) otherwise fallback to any idle * CPU (if there is one). */ static void nohz_balancer_kick(int cpu) { int ilb_cpu; nohz.next_balance++; ilb_cpu = get_nohz_load_balancer(); if (ilb_cpu >= nr_cpu_ids) { ilb_cpu = cpumask_first(nohz.idle_cpus_mask); if (ilb_cpu >= nr_cpu_ids) return; } if (!cpu_rq(ilb_cpu)->nohz_balance_kick) { struct call_single_data *cp; cpu_rq(ilb_cpu)->nohz_balance_kick = 1; cp = &per_cpu(remote_sched_softirq_cb, cpu); __smp_call_function_single(ilb_cpu, cp, 0); } return; } /* |
1e3c88bde sched: Move load ... |
3319 3320 |
* This routine will try to nominate the ilb (idle load balancing) * owner among the cpus whose ticks are stopped. ilb owner will do the idle |
83cd4fe27 sched: Change noh... |
3321 |
* load balancing on behalf of all those cpus. |
1e3c88bde sched: Move load ... |
3322 |
* |
83cd4fe27 sched: Change noh... |
3323 3324 3325 |
* When the ilb owner becomes busy, we will not have new ilb owner until some * idle CPU wakes up and goes back to idle or some busy CPU tries to kick * idle load balancing by kicking one of the idle CPUs. |
1e3c88bde sched: Move load ... |
3326 |
* |
83cd4fe27 sched: Change noh... |
3327 3328 3329 |
* Ticks are stopped for the ilb owner as well, with busy CPU kicking this * ilb owner CPU in future (when there is a need for idle load balancing on * behalf of all idle CPUs). |
1e3c88bde sched: Move load ... |
3330 |
*/ |
83cd4fe27 sched: Change noh... |
3331 |
void select_nohz_load_balancer(int stop_tick) |
1e3c88bde sched: Move load ... |
3332 3333 3334 3335 |
{ int cpu = smp_processor_id(); if (stop_tick) { |
1e3c88bde sched: Move load ... |
3336 3337 |
if (!cpu_active(cpu)) { if (atomic_read(&nohz.load_balancer) != cpu) |
83cd4fe27 sched: Change noh... |
3338 |
return; |
1e3c88bde sched: Move load ... |
3339 3340 3341 3342 3343 |
/* * If we are going offline and still the leader, * give up! */ |
83cd4fe27 sched: Change noh... |
3344 3345 |
if (atomic_cmpxchg(&nohz.load_balancer, cpu, nr_cpu_ids) != cpu) |
1e3c88bde sched: Move load ... |
3346 |
BUG(); |
83cd4fe27 sched: Change noh... |
3347 |
return; |
1e3c88bde sched: Move load ... |
3348 |
} |
83cd4fe27 sched: Change noh... |
3349 |
cpumask_set_cpu(cpu, nohz.idle_cpus_mask); |
1e3c88bde sched: Move load ... |
3350 |
|
83cd4fe27 sched: Change noh... |
3351 3352 3353 3354 |
if (atomic_read(&nohz.first_pick_cpu) == cpu) atomic_cmpxchg(&nohz.first_pick_cpu, cpu, nr_cpu_ids); if (atomic_read(&nohz.second_pick_cpu) == cpu) atomic_cmpxchg(&nohz.second_pick_cpu, cpu, nr_cpu_ids); |
1e3c88bde sched: Move load ... |
3355 |
|
83cd4fe27 sched: Change noh... |
3356 |
if (atomic_read(&nohz.load_balancer) >= nr_cpu_ids) { |
1e3c88bde sched: Move load ... |
3357 |
int new_ilb; |
83cd4fe27 sched: Change noh... |
3358 3359 3360 3361 |
/* make me the ilb owner */ if (atomic_cmpxchg(&nohz.load_balancer, nr_cpu_ids, cpu) != nr_cpu_ids) return; |
1e3c88bde sched: Move load ... |
3362 3363 3364 3365 3366 3367 |
/* * Check to see if there is a more power-efficient * ilb. */ new_ilb = find_new_ilb(cpu); if (new_ilb < nr_cpu_ids && new_ilb != cpu) { |
83cd4fe27 sched: Change noh... |
3368 |
atomic_set(&nohz.load_balancer, nr_cpu_ids); |
1e3c88bde sched: Move load ... |
3369 |
resched_cpu(new_ilb); |
83cd4fe27 sched: Change noh... |
3370 |
return; |
1e3c88bde sched: Move load ... |
3371 |
} |
83cd4fe27 sched: Change noh... |
3372 |
return; |
1e3c88bde sched: Move load ... |
3373 3374 |
} } else { |
83cd4fe27 sched: Change noh... |
3375 3376 |
if (!cpumask_test_cpu(cpu, nohz.idle_cpus_mask)) return; |
1e3c88bde sched: Move load ... |
3377 |
|
83cd4fe27 sched: Change noh... |
3378 |
cpumask_clear_cpu(cpu, nohz.idle_cpus_mask); |
1e3c88bde sched: Move load ... |
3379 3380 |
if (atomic_read(&nohz.load_balancer) == cpu) |
83cd4fe27 sched: Change noh... |
3381 3382 |
if (atomic_cmpxchg(&nohz.load_balancer, cpu, nr_cpu_ids) != cpu) |
1e3c88bde sched: Move load ... |
3383 3384 |
BUG(); } |
83cd4fe27 sched: Change noh... |
3385 |
return; |
1e3c88bde sched: Move load ... |
3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 |
} #endif static DEFINE_SPINLOCK(balancing); /* * It checks each scheduling domain to see if it is due to be balanced, * and initiates a balancing operation if so. * * Balancing parameters are set up in arch_init_sched_domains. */ static void rebalance_domains(int cpu, enum cpu_idle_type idle) { int balance = 1; struct rq *rq = cpu_rq(cpu); unsigned long interval; struct sched_domain *sd; /* Earliest time when we have to do rebalance again */ unsigned long next_balance = jiffies + 60*HZ; int update_next_balance = 0; int need_serialize; for_each_domain(cpu, sd) { if (!(sd->flags & SD_LOAD_BALANCE)) continue; interval = sd->balance_interval; if (idle != CPU_IDLE) interval *= sd->busy_factor; /* scale ms to jiffies */ interval = msecs_to_jiffies(interval); if (unlikely(!interval)) interval = 1; if (interval > HZ*NR_CPUS/10) interval = HZ*NR_CPUS/10; need_serialize = sd->flags & SD_SERIALIZE; if (need_serialize) { if (!spin_trylock(&balancing)) goto out; } if (time_after_eq(jiffies, sd->last_balance + interval)) { if (load_balance(cpu, rq, sd, idle, &balance)) { /* * We've pulled tasks over so either we're no * longer idle, or one of our SMT siblings is * not idle. */ idle = CPU_NOT_IDLE; } sd->last_balance = jiffies; } if (need_serialize) spin_unlock(&balancing); out: if (time_after(next_balance, sd->last_balance + interval)) { next_balance = sd->last_balance + interval; update_next_balance = 1; } /* * Stop the load balance at this level. There is another * CPU in our sched group which is doing load balancing more * actively. */ if (!balance) break; } /* * next_balance will be updated only when there is a need. * When the cpu is attached to null domain for ex, it will not be * updated. */ if (likely(update_next_balance)) rq->next_balance = next_balance; } |
83cd4fe27 sched: Change noh... |
3466 |
#ifdef CONFIG_NO_HZ |
1e3c88bde sched: Move load ... |
3467 |
/* |
83cd4fe27 sched: Change noh... |
3468 |
* In CONFIG_NO_HZ case, the idle balance kickee will do the |
1e3c88bde sched: Move load ... |
3469 3470 |
* rebalancing for all the cpus for whom scheduler ticks are stopped. */ |
83cd4fe27 sched: Change noh... |
3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 |
static void nohz_idle_balance(int this_cpu, enum cpu_idle_type idle) { struct rq *this_rq = cpu_rq(this_cpu); struct rq *rq; int balance_cpu; if (idle != CPU_IDLE || !this_rq->nohz_balance_kick) return; for_each_cpu(balance_cpu, nohz.idle_cpus_mask) { if (balance_cpu == this_cpu) continue; /* * If this cpu gets work to do, stop the load balancing * work being done for other cpus. Next load * balancing owner will pick it up. */ if (need_resched()) { this_rq->nohz_balance_kick = 0; break; } raw_spin_lock_irq(&this_rq->lock); |
5343bdb8f sched: Update rq-... |
3495 |
update_rq_clock(this_rq); |
83cd4fe27 sched: Change noh... |
3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506 3507 3508 3509 3510 3511 3512 3513 3514 3515 3516 3517 3518 3519 3520 3521 3522 3523 3524 3525 3526 3527 3528 |
update_cpu_load(this_rq); raw_spin_unlock_irq(&this_rq->lock); rebalance_domains(balance_cpu, CPU_IDLE); rq = cpu_rq(balance_cpu); if (time_after(this_rq->next_balance, rq->next_balance)) this_rq->next_balance = rq->next_balance; } nohz.next_balance = this_rq->next_balance; this_rq->nohz_balance_kick = 0; } /* * Current heuristic for kicking the idle load balancer * - first_pick_cpu is the one of the busy CPUs. It will kick * idle load balancer when it has more than one process active. This * eliminates the need for idle load balancing altogether when we have * only one running process in the system (common case). * - If there are more than one busy CPU, idle load balancer may have * to run for active_load_balance to happen (i.e., two busy CPUs are * SMT or core siblings and can run better if they move to different * physical CPUs). So, second_pick_cpu is the second of the busy CPUs * which will kick idle load balancer as soon as it has any load. */ static inline int nohz_kick_needed(struct rq *rq, int cpu) { unsigned long now = jiffies; int ret; int first_pick_cpu, second_pick_cpu; if (time_before(now, nohz.next_balance)) return 0; |
f6c3f1686 sched: Fix nohz b... |
3529 |
if (rq->idle_at_tick) |
83cd4fe27 sched: Change noh... |
3530 3531 3532 3533 3534 3535 3536 3537 3538 3539 3540 3541 3542 3543 3544 3545 3546 3547 3548 3549 3550 3551 3552 3553 3554 3555 3556 3557 3558 3559 3560 |
return 0; first_pick_cpu = atomic_read(&nohz.first_pick_cpu); second_pick_cpu = atomic_read(&nohz.second_pick_cpu); if (first_pick_cpu < nr_cpu_ids && first_pick_cpu != cpu && second_pick_cpu < nr_cpu_ids && second_pick_cpu != cpu) return 0; ret = atomic_cmpxchg(&nohz.first_pick_cpu, nr_cpu_ids, cpu); if (ret == nr_cpu_ids || ret == cpu) { atomic_cmpxchg(&nohz.second_pick_cpu, cpu, nr_cpu_ids); if (rq->nr_running > 1) return 1; } else { ret = atomic_cmpxchg(&nohz.second_pick_cpu, nr_cpu_ids, cpu); if (ret == nr_cpu_ids || ret == cpu) { if (rq->nr_running) return 1; } } return 0; } #else static void nohz_idle_balance(int this_cpu, enum cpu_idle_type idle) { } #endif /* * run_rebalance_domains is triggered when needed from the scheduler tick. * Also triggered for nohz idle balancing (with nohz_balancing_kick set). */ |
1e3c88bde sched: Move load ... |
3561 3562 3563 3564 3565 3566 3567 3568 |
static void run_rebalance_domains(struct softirq_action *h) { int this_cpu = smp_processor_id(); struct rq *this_rq = cpu_rq(this_cpu); enum cpu_idle_type idle = this_rq->idle_at_tick ? CPU_IDLE : CPU_NOT_IDLE; rebalance_domains(this_cpu, idle); |
1e3c88bde sched: Move load ... |
3569 |
/* |
83cd4fe27 sched: Change noh... |
3570 |
* If this cpu has a pending nohz_balance_kick, then do the |
1e3c88bde sched: Move load ... |
3571 3572 3573 |
* balancing on behalf of the other idle cpus whose ticks are * stopped. */ |
83cd4fe27 sched: Change noh... |
3574 |
nohz_idle_balance(this_cpu, idle); |
1e3c88bde sched: Move load ... |
3575 3576 3577 3578 |
} static inline int on_null_domain(int cpu) { |
90a6501f9 sched, rcu: Fix r... |
3579 |
return !rcu_dereference_sched(cpu_rq(cpu)->sd); |
1e3c88bde sched: Move load ... |
3580 3581 3582 3583 |
} /* * Trigger the SCHED_SOFTIRQ if it is time to do periodic load balancing. |
1e3c88bde sched: Move load ... |
3584 3585 3586 |
*/ static inline void trigger_load_balance(struct rq *rq, int cpu) { |
1e3c88bde sched: Move load ... |
3587 3588 3589 3590 |
/* Don't need to rebalance while attached to NULL domain */ if (time_after_eq(jiffies, rq->next_balance) && likely(!on_null_domain(cpu))) raise_softirq(SCHED_SOFTIRQ); |
83cd4fe27 sched: Change noh... |
3591 3592 3593 3594 |
#ifdef CONFIG_NO_HZ else if (nohz_kick_needed(rq, cpu) && likely(!on_null_domain(cpu))) nohz_balancer_kick(cpu); #endif |
1e3c88bde sched: Move load ... |
3595 |
} |
0bcdcf28c sched: Fix missin... |
3596 3597 3598 3599 3600 3601 3602 3603 3604 |
static void rq_online_fair(struct rq *rq) { update_sysctl(); } static void rq_offline_fair(struct rq *rq) { update_sysctl(); } |
1e3c88bde sched: Move load ... |
3605 3606 3607 3608 3609 3610 3611 3612 |
#else /* CONFIG_SMP */ /* * on UP we do not need to balance between CPUs: */ static inline void idle_balance(int cpu, struct rq *rq) { } |
55e12e5e7 sched: make sched... |
3613 |
#endif /* CONFIG_SMP */ |
e1d1484f7 sched: reduce bal... |
3614 |
|
bf0f6f24a sched: cfs core, ... |
3615 3616 3617 |
/* * scheduler tick hitting a task of our scheduling class: */ |
8f4d37ec0 sched: high-res p... |
3618 |
static void task_tick_fair(struct rq *rq, struct task_struct *curr, int queued) |
bf0f6f24a sched: cfs core, ... |
3619 3620 3621 3622 3623 3624 |
{ struct cfs_rq *cfs_rq; struct sched_entity *se = &curr->se; for_each_sched_entity(se) { cfs_rq = cfs_rq_of(se); |
8f4d37ec0 sched: high-res p... |
3625 |
entity_tick(cfs_rq, se, queued); |
bf0f6f24a sched: cfs core, ... |
3626 3627 3628 3629 |
} } /* |
cd29fe6f2 sched: Sanitize f... |
3630 3631 3632 |
* called on fork with the child task as argument from the parent's context * - child not yet on the tasklist * - preemption disabled |
bf0f6f24a sched: cfs core, ... |
3633 |
*/ |
cd29fe6f2 sched: Sanitize f... |
3634 |
static void task_fork_fair(struct task_struct *p) |
bf0f6f24a sched: cfs core, ... |
3635 |
{ |
cd29fe6f2 sched: Sanitize f... |
3636 |
struct cfs_rq *cfs_rq = task_cfs_rq(current); |
429d43bcc sched: cleanup: s... |
3637 |
struct sched_entity *se = &p->se, *curr = cfs_rq->curr; |
00bf7bfc2 sched: fix: move ... |
3638 |
int this_cpu = smp_processor_id(); |
cd29fe6f2 sched: Sanitize f... |
3639 3640 |
struct rq *rq = this_rq(); unsigned long flags; |
05fa785cf sched: Convert rq... |
3641 |
raw_spin_lock_irqsave(&rq->lock, flags); |
bf0f6f24a sched: cfs core, ... |
3642 |
|
861d034ee sched: Fix rq->cl... |
3643 |
update_rq_clock(rq); |
b0a0f667a sched: suppress R... |
3644 3645 |
if (unlikely(task_cpu(p) != this_cpu)) { rcu_read_lock(); |
cd29fe6f2 sched: Sanitize f... |
3646 |
__set_task_cpu(p, this_cpu); |
b0a0f667a sched: suppress R... |
3647 3648 |
rcu_read_unlock(); } |
bf0f6f24a sched: cfs core, ... |
3649 |
|
7109c4429 sched: call updat... |
3650 |
update_curr(cfs_rq); |
cd29fe6f2 sched: Sanitize f... |
3651 |
|
b5d9d734a sched: Ensure tha... |
3652 3653 |
if (curr) se->vruntime = curr->vruntime; |
aeb73b040 sched: clean up n... |
3654 |
place_entity(cfs_rq, se, 1); |
4d78e7b65 sched: new task p... |
3655 |
|
cd29fe6f2 sched: Sanitize f... |
3656 |
if (sysctl_sched_child_runs_first && curr && entity_before(curr, se)) { |
87fefa381 sched: optimize t... |
3657 |
/* |
edcb60a30 sched: kernel/sch... |
3658 3659 3660 |
* Upon rescheduling, sched_class::put_prev_task() will place * 'current' within the tree based on its new key value. */ |
4d78e7b65 sched: new task p... |
3661 |
swap(curr->vruntime, se->vruntime); |
aec0a5142 sched: call resch... |
3662 |
resched_task(rq->curr); |
4d78e7b65 sched: new task p... |
3663 |
} |
bf0f6f24a sched: cfs core, ... |
3664 |
|
88ec22d3e sched: Remove the... |
3665 |
se->vruntime -= cfs_rq->min_vruntime; |
05fa785cf sched: Convert rq... |
3666 |
raw_spin_unlock_irqrestore(&rq->lock, flags); |
bf0f6f24a sched: cfs core, ... |
3667 |
} |
cb4698450 sched: RT-balance... |
3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678 3679 3680 3681 3682 3683 |
/* * Priority of the task has changed. Check to see if we preempt * the current task. */ static void prio_changed_fair(struct rq *rq, struct task_struct *p, int oldprio, int running) { /* * Reschedule if we are currently running on this runqueue and * our priority decreased, or if we are not currently running on * this runqueue and our priority is higher than the current's */ if (running) { if (p->prio > oldprio) resched_task(rq->curr); } else |
15afe09bf sched: wakeup pre... |
3684 |
check_preempt_curr(rq, p, 0); |
cb4698450 sched: RT-balance... |
3685 3686 3687 3688 3689 3690 3691 3692 3693 3694 3695 3696 3697 3698 3699 3700 |
} /* * We switched to the sched_fair class. */ static void switched_to_fair(struct rq *rq, struct task_struct *p, int running) { /* * We were most likely switched from sched_rt, so * kick off the schedule if running, otherwise just see * if we can still preempt the current task. */ if (running) resched_task(rq->curr); else |
15afe09bf sched: wakeup pre... |
3701 |
check_preempt_curr(rq, p, 0); |
cb4698450 sched: RT-balance... |
3702 |
} |
83b699ed2 sched: revert rec... |
3703 3704 3705 3706 3707 3708 3709 3710 3711 3712 3713 3714 |
/* Account for a task changing its policy or group. * * This routine is mostly called to set cfs_rq->curr field when a task * migrates between groups/classes. */ static void set_curr_task_fair(struct rq *rq) { struct sched_entity *se = &rq->curr->se; for_each_sched_entity(se) set_next_entity(cfs_rq_of(se), se); } |
810b38179 sched: retain vru... |
3715 |
#ifdef CONFIG_FAIR_GROUP_SCHED |
88ec22d3e sched: Remove the... |
3716 |
static void moved_group_fair(struct task_struct *p, int on_rq) |
810b38179 sched: retain vru... |
3717 3718 3719 3720 |
{ struct cfs_rq *cfs_rq = task_cfs_rq(p); update_curr(cfs_rq); |
88ec22d3e sched: Remove the... |
3721 3722 |
if (!on_rq) place_entity(cfs_rq, &p->se, 1); |
810b38179 sched: retain vru... |
3723 3724 |
} #endif |
6d686f456 sched: Don't expo... |
3725 |
static unsigned int get_rr_interval_fair(struct rq *rq, struct task_struct *task) |
0d721cead sched: Simplify s... |
3726 3727 |
{ struct sched_entity *se = &task->se; |
0d721cead sched: Simplify s... |
3728 3729 3730 3731 3732 3733 |
unsigned int rr_interval = 0; /* * Time slice is 0 for SCHED_OTHER tasks that are on an otherwise * idle runqueue: */ |
0d721cead sched: Simplify s... |
3734 3735 |
if (rq->cfs.load.weight) rr_interval = NS_TO_JIFFIES(sched_slice(&rq->cfs, se)); |
0d721cead sched: Simplify s... |
3736 3737 3738 |
return rr_interval; } |
bf0f6f24a sched: cfs core, ... |
3739 3740 3741 |
/* * All the scheduling class methods: */ |
5522d5d5f sched: mark sched... |
3742 3743 |
static const struct sched_class fair_sched_class = { .next = &idle_sched_class, |
bf0f6f24a sched: cfs core, ... |
3744 3745 3746 |
.enqueue_task = enqueue_task_fair, .dequeue_task = dequeue_task_fair, .yield_task = yield_task_fair, |
2e09bf556 sched: wakeup gra... |
3747 |
.check_preempt_curr = check_preempt_wakeup, |
bf0f6f24a sched: cfs core, ... |
3748 3749 3750 |
.pick_next_task = pick_next_task_fair, .put_prev_task = put_prev_task_fair, |
681f3e685 sched: isolate SM... |
3751 |
#ifdef CONFIG_SMP |
4ce72a2c0 sched: add CONFIG... |
3752 |
.select_task_rq = select_task_rq_fair, |
0bcdcf28c sched: Fix missin... |
3753 3754 |
.rq_online = rq_online_fair, .rq_offline = rq_offline_fair, |
88ec22d3e sched: Remove the... |
3755 3756 |
.task_waking = task_waking_fair, |
681f3e685 sched: isolate SM... |
3757 |
#endif |
bf0f6f24a sched: cfs core, ... |
3758 |
|
83b699ed2 sched: revert rec... |
3759 |
.set_curr_task = set_curr_task_fair, |
bf0f6f24a sched: cfs core, ... |
3760 |
.task_tick = task_tick_fair, |
cd29fe6f2 sched: Sanitize f... |
3761 |
.task_fork = task_fork_fair, |
cb4698450 sched: RT-balance... |
3762 3763 3764 |
.prio_changed = prio_changed_fair, .switched_to = switched_to_fair, |
810b38179 sched: retain vru... |
3765 |
|
0d721cead sched: Simplify s... |
3766 |
.get_rr_interval = get_rr_interval_fair, |
810b38179 sched: retain vru... |
3767 3768 3769 |
#ifdef CONFIG_FAIR_GROUP_SCHED .moved_group = moved_group_fair, #endif |
bf0f6f24a sched: cfs core, ... |
3770 3771 3772 |
}; #ifdef CONFIG_SCHED_DEBUG |
5cef9eca3 sched: remove the... |
3773 |
static void print_cfs_stats(struct seq_file *m, int cpu) |
bf0f6f24a sched: cfs core, ... |
3774 |
{ |
bf0f6f24a sched: cfs core, ... |
3775 |
struct cfs_rq *cfs_rq; |
5973e5b95 sched: fix: don't... |
3776 |
rcu_read_lock(); |
c3b64f1e4 sched: clean up s... |
3777 |
for_each_leaf_cfs_rq(cpu_rq(cpu), cfs_rq) |
5cef9eca3 sched: remove the... |
3778 |
print_cfs_rq(m, cpu, cfs_rq); |
5973e5b95 sched: fix: don't... |
3779 |
rcu_read_unlock(); |
bf0f6f24a sched: cfs core, ... |
3780 3781 |
} #endif |