Blame view
kernel/rcutree.c
64.7 KB
64db4cfff
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 |
/* * Read-Copy Update mechanism for mutual exclusion * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by * the Free Software Foundation; either version 2 of the License, or * (at your option) any later version. * * This program is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. * * You should have received a copy of the GNU General Public License * along with this program; if not, write to the Free Software * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. * * Copyright IBM Corporation, 2008 * * Authors: Dipankar Sarma <dipankar@in.ibm.com> * Manfred Spraul <manfred@colorfullife.com> * Paul E. McKenney <paulmck@linux.vnet.ibm.com> Hierarchical version * * Based on the original work by Paul McKenney <paulmck@us.ibm.com> * and inputs from Rusty Russell, Andrea Arcangeli and Andi Kleen. * * For detailed explanation of Read-Copy Update mechanism see - |
a71fca58b
|
28 |
* Documentation/RCU |
64db4cfff
|
29 30 31 32 33 34 35 36 37 |
*/ #include <linux/types.h> #include <linux/kernel.h> #include <linux/init.h> #include <linux/spinlock.h> #include <linux/smp.h> #include <linux/rcupdate.h> #include <linux/interrupt.h> #include <linux/sched.h> |
c1dc0b9c0
|
38 |
#include <linux/nmi.h> |
8826f3b03
|
39 |
#include <linux/atomic.h> |
64db4cfff
|
40 |
#include <linux/bitops.h> |
9984de1a5
|
41 |
#include <linux/export.h> |
64db4cfff
|
42 43 44 45 46 47 48 |
#include <linux/completion.h> #include <linux/moduleparam.h> #include <linux/percpu.h> #include <linux/notifier.h> #include <linux/cpu.h> #include <linux/mutex.h> #include <linux/time.h> |
bbad93798
|
49 |
#include <linux/kernel_stat.h> |
a26ac2455
|
50 51 |
#include <linux/wait.h> #include <linux/kthread.h> |
268bb0ce3
|
52 |
#include <linux/prefetch.h> |
64db4cfff
|
53 |
|
9f77da9f4
|
54 |
#include "rcutree.h" |
29c00b4a1
|
55 56 57 |
#include <trace/events/rcu.h> #include "rcu.h" |
9f77da9f4
|
58 |
|
64db4cfff
|
59 |
/* Data structures. */ |
b668c9cf3
|
60 |
static struct lock_class_key rcu_node_class[NUM_RCU_LVLS]; |
88b91c7ca
|
61 |
|
4300aa642
|
62 |
#define RCU_STATE_INITIALIZER(structname) { \ |
e99033c5c
|
63 |
.level = { &structname##_state.node[0] }, \ |
64db4cfff
|
64 65 66 67 |
.levelcnt = { \ NUM_RCU_LVL_0, /* root of hierarchy. */ \ NUM_RCU_LVL_1, \ NUM_RCU_LVL_2, \ |
cf244dc01
|
68 69 |
NUM_RCU_LVL_3, \ NUM_RCU_LVL_4, /* == MAX_RCU_LVLS */ \ |
64db4cfff
|
70 |
}, \ |
83f5b01ff
|
71 |
.signaled = RCU_GP_IDLE, \ |
64db4cfff
|
72 73 |
.gpnum = -300, \ .completed = -300, \ |
e99033c5c
|
74 75 |
.onofflock = __RAW_SPIN_LOCK_UNLOCKED(&structname##_state.onofflock), \ .fqslock = __RAW_SPIN_LOCK_UNLOCKED(&structname##_state.fqslock), \ |
64db4cfff
|
76 77 |
.n_force_qs = 0, \ .n_force_qs_ngp = 0, \ |
4300aa642
|
78 |
.name = #structname, \ |
64db4cfff
|
79 |
} |
e99033c5c
|
80 |
struct rcu_state rcu_sched_state = RCU_STATE_INITIALIZER(rcu_sched); |
d6714c22b
|
81 |
DEFINE_PER_CPU(struct rcu_data, rcu_sched_data); |
64db4cfff
|
82 |
|
e99033c5c
|
83 |
struct rcu_state rcu_bh_state = RCU_STATE_INITIALIZER(rcu_bh); |
6258c4fb5
|
84 |
DEFINE_PER_CPU(struct rcu_data, rcu_bh_data); |
b1f77b058
|
85 |
|
27f4d2805
|
86 |
static struct rcu_state *rcu_state; |
b0d304172
|
87 88 89 90 91 92 93 94 95 |
/* * The rcu_scheduler_active variable transitions from zero to one just * before the first task is spawned. So when this variable is zero, RCU * can assume that there is but one task, allowing RCU to (for example) * optimized synchronize_sched() to a simple barrier(). When this variable * is one, RCU must actually do all the hard work required to detect real * grace periods. This variable is also used to suppress boot-time false * positives from lockdep-RCU error checking. */ |
bbad93798
|
96 97 |
int rcu_scheduler_active __read_mostly; EXPORT_SYMBOL_GPL(rcu_scheduler_active); |
b0d304172
|
98 99 100 101 102 103 104 105 106 107 108 109 110 |
/* * The rcu_scheduler_fully_active variable transitions from zero to one * during the early_initcall() processing, which is after the scheduler * is capable of creating new tasks. So RCU processing (for example, * creating tasks for RCU priority boosting) must be delayed until after * rcu_scheduler_fully_active transitions from zero to one. We also * currently delay invocation of any RCU callbacks until after this point. * * It might later prove better for people registering RCU callbacks during * early boot to take responsibility for these callbacks, but one step at * a time. */ static int rcu_scheduler_fully_active __read_mostly; |
a46e0899e
|
111 |
#ifdef CONFIG_RCU_BOOST |
b1f77b058
|
112 |
/* |
a26ac2455
|
113 114 115 116 |
* Control variables for per-CPU and per-rcu_node kthreads. These * handle all flavors of RCU. */ static DEFINE_PER_CPU(struct task_struct *, rcu_cpu_kthread_task); |
d71df90ea
|
117 |
DEFINE_PER_CPU(unsigned int, rcu_cpu_kthread_status); |
15ba0ba86
|
118 |
DEFINE_PER_CPU(int, rcu_cpu_kthread_cpu); |
5ece5bab3
|
119 |
DEFINE_PER_CPU(unsigned int, rcu_cpu_kthread_loops); |
d71df90ea
|
120 |
DEFINE_PER_CPU(char, rcu_cpu_has_work); |
a26ac2455
|
121 |
|
a46e0899e
|
122 |
#endif /* #ifdef CONFIG_RCU_BOOST */ |
0f962a5e7
|
123 |
static void rcu_node_kthread_setaffinity(struct rcu_node *rnp, int outgoingcpu); |
a46e0899e
|
124 125 |
static void invoke_rcu_core(void); static void invoke_rcu_callbacks(struct rcu_state *rsp, struct rcu_data *rdp); |
a26ac2455
|
126 |
|
a26ac2455
|
127 |
/* |
4a2986568
|
128 129 130 131 132 133 134 135 136 137 138 139 |
* Track the rcutorture test sequence number and the update version * number within a given test. The rcutorture_testseq is incremented * on every rcutorture module load and unload, so has an odd value * when a test is running. The rcutorture_vernum is set to zero * when rcutorture starts and is incremented on each rcutorture update. * These variables enable correlating rcutorture output with the * RCU tracing information. */ unsigned long rcutorture_testseq; unsigned long rcutorture_vernum; /* |
fc2219d49
|
140 141 142 143 144 145 146 147 148 149 |
* Return true if an RCU grace period is in progress. The ACCESS_ONCE()s * permit this function to be invoked without holding the root rcu_node * structure's ->lock, but of course results can be subject to change. */ static int rcu_gp_in_progress(struct rcu_state *rsp) { return ACCESS_ONCE(rsp->completed) != ACCESS_ONCE(rsp->gpnum); } /* |
d6714c22b
|
150 |
* Note a quiescent state. Because we do not need to know |
b1f77b058
|
151 |
* how many quiescent states passed, just if there was at least |
d6714c22b
|
152 |
* one since the start of the grace period, this just sets a flag. |
e4cc1f22b
|
153 |
* The caller must have disabled preemption. |
b1f77b058
|
154 |
*/ |
d6714c22b
|
155 |
void rcu_sched_qs(int cpu) |
b1f77b058
|
156 |
{ |
25502a6c1
|
157 |
struct rcu_data *rdp = &per_cpu(rcu_sched_data, cpu); |
f41d911f8
|
158 |
|
e4cc1f22b
|
159 |
rdp->passed_quiesce_gpnum = rdp->gpnum; |
c3422bea5
|
160 |
barrier(); |
e4cc1f22b
|
161 |
if (rdp->passed_quiesce == 0) |
d4c08f2ac
|
162 |
trace_rcu_grace_period("rcu_sched", rdp->gpnum, "cpuqs"); |
e4cc1f22b
|
163 |
rdp->passed_quiesce = 1; |
b1f77b058
|
164 |
} |
d6714c22b
|
165 |
void rcu_bh_qs(int cpu) |
b1f77b058
|
166 |
{ |
25502a6c1
|
167 |
struct rcu_data *rdp = &per_cpu(rcu_bh_data, cpu); |
f41d911f8
|
168 |
|
e4cc1f22b
|
169 |
rdp->passed_quiesce_gpnum = rdp->gpnum; |
c3422bea5
|
170 |
barrier(); |
e4cc1f22b
|
171 |
if (rdp->passed_quiesce == 0) |
d4c08f2ac
|
172 |
trace_rcu_grace_period("rcu_bh", rdp->gpnum, "cpuqs"); |
e4cc1f22b
|
173 |
rdp->passed_quiesce = 1; |
b1f77b058
|
174 |
} |
64db4cfff
|
175 |
|
25502a6c1
|
176 177 178 |
/* * Note a context switch. This is a quiescent state for RCU-sched, * and requires special handling for preemptible RCU. |
e4cc1f22b
|
179 |
* The caller must have disabled preemption. |
25502a6c1
|
180 181 182 |
*/ void rcu_note_context_switch(int cpu) { |
300df91ca
|
183 |
trace_rcu_utilization("Start context switch"); |
25502a6c1
|
184 185 |
rcu_sched_qs(cpu); rcu_preempt_note_context_switch(cpu); |
300df91ca
|
186 |
trace_rcu_utilization("End context switch"); |
25502a6c1
|
187 |
} |
29ce83100
|
188 |
EXPORT_SYMBOL_GPL(rcu_note_context_switch); |
25502a6c1
|
189 |
|
64db4cfff
|
190 |
#ifdef CONFIG_NO_HZ |
90a4d2c01
|
191 192 |
DEFINE_PER_CPU(struct rcu_dynticks, rcu_dynticks) = { .dynticks_nesting = 1, |
23b5c8fa0
|
193 |
.dynticks = ATOMIC_INIT(1), |
90a4d2c01
|
194 |
}; |
64db4cfff
|
195 |
#endif /* #ifdef CONFIG_NO_HZ */ |
e0f23060a
|
196 |
static int blimit = 10; /* Maximum callbacks per rcu_do_batch. */ |
64db4cfff
|
197 198 |
static int qhimark = 10000; /* If this many pending, ignore blimit. */ static int qlowmark = 100; /* Once only this many pending, use blimit. */ |
3d76c0829
|
199 200 201 |
module_param(blimit, int, 0); module_param(qhimark, int, 0); module_param(qlowmark, int, 0); |
a00e0d714
|
202 |
int rcu_cpu_stall_suppress __read_mostly; |
f2e0dd709
|
203 |
module_param(rcu_cpu_stall_suppress, int, 0644); |
742734eea
|
204 |
|
64db4cfff
|
205 |
static void force_quiescent_state(struct rcu_state *rsp, int relaxed); |
a157229ca
|
206 |
static int rcu_pending(int cpu); |
64db4cfff
|
207 208 |
/* |
d6714c22b
|
209 |
* Return the number of RCU-sched batches processed thus far for debug & stats. |
64db4cfff
|
210 |
*/ |
d6714c22b
|
211 |
long rcu_batches_completed_sched(void) |
64db4cfff
|
212 |
{ |
d6714c22b
|
213 |
return rcu_sched_state.completed; |
64db4cfff
|
214 |
} |
d6714c22b
|
215 |
EXPORT_SYMBOL_GPL(rcu_batches_completed_sched); |
64db4cfff
|
216 217 218 219 220 221 222 223 224 225 226 |
/* * Return the number of RCU BH batches processed thus far for debug & stats. */ long rcu_batches_completed_bh(void) { return rcu_bh_state.completed; } EXPORT_SYMBOL_GPL(rcu_batches_completed_bh); /* |
bf66f18e7
|
227 228 229 230 231 232 233 234 235 |
* Force a quiescent state for RCU BH. */ void rcu_bh_force_quiescent_state(void) { force_quiescent_state(&rcu_bh_state, 0); } EXPORT_SYMBOL_GPL(rcu_bh_force_quiescent_state); /* |
4a2986568
|
236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 |
* Record the number of times rcutorture tests have been initiated and * terminated. This information allows the debugfs tracing stats to be * correlated to the rcutorture messages, even when the rcutorture module * is being repeatedly loaded and unloaded. In other words, we cannot * store this state in rcutorture itself. */ void rcutorture_record_test_transition(void) { rcutorture_testseq++; rcutorture_vernum = 0; } EXPORT_SYMBOL_GPL(rcutorture_record_test_transition); /* * Record the number of writer passes through the current rcutorture test. * This is also used to correlate debugfs tracing stats with the rcutorture * messages. */ void rcutorture_record_progress(unsigned long vernum) { rcutorture_vernum++; } EXPORT_SYMBOL_GPL(rcutorture_record_progress); /* |
bf66f18e7
|
261 262 263 264 265 266 267 268 269 |
* Force a quiescent state for RCU-sched. */ void rcu_sched_force_quiescent_state(void) { force_quiescent_state(&rcu_sched_state, 0); } EXPORT_SYMBOL_GPL(rcu_sched_force_quiescent_state); /* |
64db4cfff
|
270 271 272 273 274 275 276 277 278 279 280 281 282 283 |
* Does the CPU have callbacks ready to be invoked? */ static int cpu_has_callbacks_ready_to_invoke(struct rcu_data *rdp) { return &rdp->nxtlist != rdp->nxttail[RCU_DONE_TAIL]; } /* * Does the current CPU require a yet-as-unscheduled grace period? */ static int cpu_needs_another_gp(struct rcu_state *rsp, struct rcu_data *rdp) { |
fc2219d49
|
284 |
return *rdp->nxttail[RCU_DONE_TAIL] && !rcu_gp_in_progress(rsp); |
64db4cfff
|
285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 |
} /* * Return the root node of the specified rcu_state structure. */ static struct rcu_node *rcu_get_root(struct rcu_state *rsp) { return &rsp->node[0]; } #ifdef CONFIG_SMP /* * If the specified CPU is offline, tell the caller that it is in * a quiescent state. Otherwise, whack it with a reschedule IPI. * Grace periods can end up waiting on an offline CPU when that * CPU is in the process of coming online -- it will be added to the * rcu_node bitmasks before it actually makes it online. The same thing * can happen while a CPU is in the process of coming online. Because this * race is quite rare, we check for it after detecting that the grace * period has been delayed rather than checking each and every CPU * each and every time we start a new grace period. */ static int rcu_implicit_offline_qs(struct rcu_data *rdp) { /* * If the CPU is offline, it is in a quiescent state. We can * trust its state not to change because interrupts are disabled. */ if (cpu_is_offline(rdp->cpu)) { |
d4c08f2ac
|
315 |
trace_rcu_fqs(rdp->rsp->name, rdp->gpnum, rdp->cpu, "ofl"); |
64db4cfff
|
316 317 318 |
rdp->offline_fqs++; return 1; } |
6cc68793e
|
319 320 |
/* If preemptible RCU, no point in sending reschedule IPI. */ if (rdp->preemptible) |
f41d911f8
|
321 |
return 0; |
64db4cfff
|
322 323 324 325 326 327 328 329 330 331 332 333 |
/* The CPU is online, so send it a reschedule IPI. */ if (rdp->cpu != smp_processor_id()) smp_send_reschedule(rdp->cpu); else set_need_resched(); rdp->resched_ipi++; return 0; } #endif /* #ifdef CONFIG_SMP */ #ifdef CONFIG_NO_HZ |
64db4cfff
|
334 335 336 337 338 339 340 341 342 343 344 345 346 |
/** * rcu_enter_nohz - inform RCU that current CPU is entering nohz * * Enter nohz mode, in other words, -leave- the mode in which RCU * read-side critical sections can occur. (Though RCU read-side * critical sections can occur in irq handlers in nohz mode, a possibility * handled by rcu_irq_enter() and rcu_irq_exit()). */ void rcu_enter_nohz(void) { unsigned long flags; struct rcu_dynticks *rdtp; |
64db4cfff
|
347 348 |
local_irq_save(flags); rdtp = &__get_cpu_var(rcu_dynticks); |
23b5c8fa0
|
349 350 351 352 |
if (--rdtp->dynticks_nesting) { local_irq_restore(flags); return; } |
d4c08f2ac
|
353 |
trace_rcu_dyntick("Start"); |
23b5c8fa0
|
354 355 356 357 358 |
/* CPUs seeing atomic_inc() must see prior RCU read-side crit sects */ smp_mb__before_atomic_inc(); /* See above. */ atomic_inc(&rdtp->dynticks); smp_mb__after_atomic_inc(); /* Force ordering with next sojourn. */ WARN_ON_ONCE(atomic_read(&rdtp->dynticks) & 0x1); |
64db4cfff
|
359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 |
local_irq_restore(flags); } /* * rcu_exit_nohz - inform RCU that current CPU is leaving nohz * * Exit nohz mode, in other words, -enter- the mode in which RCU * read-side critical sections normally occur. */ void rcu_exit_nohz(void) { unsigned long flags; struct rcu_dynticks *rdtp; local_irq_save(flags); rdtp = &__get_cpu_var(rcu_dynticks); |
23b5c8fa0
|
375 376 377 378 379 380 381 382 383 |
if (rdtp->dynticks_nesting++) { local_irq_restore(flags); return; } smp_mb__before_atomic_inc(); /* Force ordering w/previous sojourn. */ atomic_inc(&rdtp->dynticks); /* CPUs seeing atomic_inc() must see later RCU read-side crit sects */ smp_mb__after_atomic_inc(); /* See above. */ WARN_ON_ONCE(!(atomic_read(&rdtp->dynticks) & 0x1)); |
d4c08f2ac
|
384 |
trace_rcu_dyntick("End"); |
64db4cfff
|
385 |
local_irq_restore(flags); |
64db4cfff
|
386 387 388 389 390 391 392 393 394 395 396 397 |
} /** * rcu_nmi_enter - inform RCU of entry to NMI context * * If the CPU was idle with dynamic ticks active, and there is no * irq handler running, this updates rdtp->dynticks_nmi to let the * RCU grace-period handling know that the CPU is active. */ void rcu_nmi_enter(void) { struct rcu_dynticks *rdtp = &__get_cpu_var(rcu_dynticks); |
23b5c8fa0
|
398 399 |
if (rdtp->dynticks_nmi_nesting == 0 && (atomic_read(&rdtp->dynticks) & 0x1)) |
64db4cfff
|
400 |
return; |
23b5c8fa0
|
401 402 403 404 405 406 |
rdtp->dynticks_nmi_nesting++; smp_mb__before_atomic_inc(); /* Force delay from prior write. */ atomic_inc(&rdtp->dynticks); /* CPUs seeing atomic_inc() must see later RCU read-side crit sects */ smp_mb__after_atomic_inc(); /* See above. */ WARN_ON_ONCE(!(atomic_read(&rdtp->dynticks) & 0x1)); |
64db4cfff
|
407 408 409 410 411 412 413 414 415 416 417 418 |
} /** * rcu_nmi_exit - inform RCU of exit from NMI context * * If the CPU was idle with dynamic ticks active, and there is no * irq handler running, this updates rdtp->dynticks_nmi to let the * RCU grace-period handling know that the CPU is no longer active. */ void rcu_nmi_exit(void) { struct rcu_dynticks *rdtp = &__get_cpu_var(rcu_dynticks); |
23b5c8fa0
|
419 420 |
if (rdtp->dynticks_nmi_nesting == 0 || --rdtp->dynticks_nmi_nesting != 0) |
64db4cfff
|
421 |
return; |
23b5c8fa0
|
422 423 424 425 426 |
/* CPUs seeing atomic_inc() must see prior RCU read-side crit sects */ smp_mb__before_atomic_inc(); /* See above. */ atomic_inc(&rdtp->dynticks); smp_mb__after_atomic_inc(); /* Force delay to next write. */ WARN_ON_ONCE(atomic_read(&rdtp->dynticks) & 0x1); |
64db4cfff
|
427 428 429 430 431 432 433 434 435 436 |
} /** * rcu_irq_enter - inform RCU of entry to hard irq context * * If the CPU was idle with dynamic ticks active, this updates the * rdtp->dynticks to let the RCU handling know that the CPU is active. */ void rcu_irq_enter(void) { |
23b5c8fa0
|
437 |
rcu_exit_nohz(); |
64db4cfff
|
438 439 440 441 442 443 444 445 446 447 448 |
} /** * rcu_irq_exit - inform RCU of exit from hard irq context * * If the CPU was idle with dynamic ticks active, update the rdp->dynticks * to put let the RCU handling be aware that the CPU is going back to idle * with no ticks. */ void rcu_irq_exit(void) { |
23b5c8fa0
|
449 |
rcu_enter_nohz(); |
64db4cfff
|
450 |
} |
64db4cfff
|
451 452 453 |
#ifdef CONFIG_SMP /* |
64db4cfff
|
454 455 |
* Snapshot the specified CPU's dynticks counter so that we can later * credit them with an implicit quiescent state. Return 1 if this CPU |
1eba8f843
|
456 |
* is in dynticks idle mode, which is an extended quiescent state. |
64db4cfff
|
457 458 459 |
*/ static int dyntick_save_progress_counter(struct rcu_data *rdp) { |
23b5c8fa0
|
460 461 |
rdp->dynticks_snap = atomic_add_return(0, &rdp->dynticks->dynticks); return 0; |
64db4cfff
|
462 463 464 465 466 467 468 469 470 471 |
} /* * Return true if the specified CPU has passed through a quiescent * state by virtue of being in or having passed through an dynticks * idle state since the last call to dyntick_save_progress_counter() * for this same CPU. */ static int rcu_implicit_dynticks_qs(struct rcu_data *rdp) { |
7eb4f4553
|
472 473 |
unsigned int curr; unsigned int snap; |
64db4cfff
|
474 |
|
7eb4f4553
|
475 476 |
curr = (unsigned int)atomic_add_return(0, &rdp->dynticks->dynticks); snap = (unsigned int)rdp->dynticks_snap; |
64db4cfff
|
477 478 479 480 481 482 483 484 485 |
/* * If the CPU passed through or entered a dynticks idle phase with * no active irq/NMI handlers, then we can safely pretend that the CPU * already acknowledged the request to pass through a quiescent * state. Either way, that CPU cannot possibly be in an RCU * read-side critical section that started before the beginning * of the current RCU grace period. */ |
7eb4f4553
|
486 |
if ((curr & 0x1) == 0 || UINT_CMP_GE(curr, snap + 2)) { |
d4c08f2ac
|
487 |
trace_rcu_fqs(rdp->rsp->name, rdp->gpnum, rdp->cpu, "dti"); |
64db4cfff
|
488 489 490 491 492 493 494 495 496 497 498 |
rdp->dynticks_fqs++; return 1; } /* Go check for the CPU being offline. */ return rcu_implicit_offline_qs(rdp); } #endif /* #ifdef CONFIG_SMP */ #else /* #ifdef CONFIG_NO_HZ */ |
64db4cfff
|
499 |
#ifdef CONFIG_SMP |
64db4cfff
|
500 501 502 503 504 505 506 507 508 509 510 511 512 |
static int dyntick_save_progress_counter(struct rcu_data *rdp) { return 0; } static int rcu_implicit_dynticks_qs(struct rcu_data *rdp) { return rcu_implicit_offline_qs(rdp); } #endif /* #ifdef CONFIG_SMP */ #endif /* #else #ifdef CONFIG_NO_HZ */ |
742734eea
|
513 |
int rcu_cpu_stall_suppress __read_mostly; |
c68de2097
|
514 |
|
64db4cfff
|
515 516 517 518 519 520 521 522 523 524 525 |
static void record_gp_stall_check_time(struct rcu_state *rsp) { rsp->gp_start = jiffies; rsp->jiffies_stall = jiffies + RCU_SECONDS_TILL_STALL_CHECK; } static void print_other_cpu_stall(struct rcu_state *rsp) { int cpu; long delta; unsigned long flags; |
9bc8b5586
|
526 |
int ndetected; |
64db4cfff
|
527 |
struct rcu_node *rnp = rcu_get_root(rsp); |
64db4cfff
|
528 529 |
/* Only let one CPU complain about others per time interval. */ |
1304afb22
|
530 |
raw_spin_lock_irqsave(&rnp->lock, flags); |
64db4cfff
|
531 |
delta = jiffies - rsp->jiffies_stall; |
fc2219d49
|
532 |
if (delta < RCU_STALL_RAT_DELAY || !rcu_gp_in_progress(rsp)) { |
1304afb22
|
533 |
raw_spin_unlock_irqrestore(&rnp->lock, flags); |
64db4cfff
|
534 535 536 |
return; } rsp->jiffies_stall = jiffies + RCU_SECONDS_TILL_STALL_RECHECK; |
a0b6c9a78
|
537 538 539 540 541 |
/* * Now rat on any tasks that got kicked up to the root rcu_node * due to CPU offlining. */ |
9bc8b5586
|
542 |
ndetected = rcu_print_task_stall(rnp); |
1304afb22
|
543 |
raw_spin_unlock_irqrestore(&rnp->lock, flags); |
64db4cfff
|
544 |
|
8cdd32a91
|
545 546 547 548 549 |
/* * OK, time to rat on our buddy... * See Documentation/RCU/stallwarn.txt for info on how to debug * RCU CPU stall warnings. */ |
4300aa642
|
550 551 |
printk(KERN_ERR "INFO: %s detected stalls on CPUs/tasks: {", rsp->name); |
a0b6c9a78
|
552 |
rcu_for_each_leaf_node(rsp, rnp) { |
3acd9eb31
|
553 |
raw_spin_lock_irqsave(&rnp->lock, flags); |
9bc8b5586
|
554 |
ndetected += rcu_print_task_stall(rnp); |
3acd9eb31
|
555 |
raw_spin_unlock_irqrestore(&rnp->lock, flags); |
a0b6c9a78
|
556 |
if (rnp->qsmask == 0) |
64db4cfff
|
557 |
continue; |
a0b6c9a78
|
558 |
for (cpu = 0; cpu <= rnp->grphi - rnp->grplo; cpu++) |
9bc8b5586
|
559 |
if (rnp->qsmask & (1UL << cpu)) { |
a0b6c9a78
|
560 |
printk(" %d", rnp->grplo + cpu); |
9bc8b5586
|
561 562 |
ndetected++; } |
64db4cfff
|
563 |
} |
4300aa642
|
564 565 |
printk("} (detected by %d, t=%ld jiffies) ", |
64db4cfff
|
566 |
smp_processor_id(), (long)(jiffies - rsp->gp_start)); |
9bc8b5586
|
567 568 569 570 |
if (ndetected == 0) printk(KERN_ERR "INFO: Stall ended before state dump start "); else if (!trigger_all_cpu_backtrace()) |
4627e240d
|
571 |
dump_stack(); |
c1dc0b9c0
|
572 |
|
1ed509a22
|
573 574 575 |
/* If so configured, complain about tasks blocking the grace period. */ rcu_print_detail_task_stall(rsp); |
64db4cfff
|
576 577 578 579 580 581 582 |
force_quiescent_state(rsp, 0); /* Kick them all. */ } static void print_cpu_stall(struct rcu_state *rsp) { unsigned long flags; struct rcu_node *rnp = rcu_get_root(rsp); |
8cdd32a91
|
583 584 585 586 587 |
/* * OK, time to rat on ourselves... * See Documentation/RCU/stallwarn.txt for info on how to debug * RCU CPU stall warnings. */ |
4300aa642
|
588 589 590 |
printk(KERN_ERR "INFO: %s detected stall on CPU %d (t=%lu jiffies) ", rsp->name, smp_processor_id(), jiffies - rsp->gp_start); |
4627e240d
|
591 592 |
if (!trigger_all_cpu_backtrace()) dump_stack(); |
c1dc0b9c0
|
593 |
|
1304afb22
|
594 |
raw_spin_lock_irqsave(&rnp->lock, flags); |
20133cfce
|
595 |
if (ULONG_CMP_GE(jiffies, rsp->jiffies_stall)) |
64db4cfff
|
596 597 |
rsp->jiffies_stall = jiffies + RCU_SECONDS_TILL_STALL_RECHECK; |
1304afb22
|
598 |
raw_spin_unlock_irqrestore(&rnp->lock, flags); |
c1dc0b9c0
|
599 |
|
64db4cfff
|
600 601 602 603 604 |
set_need_resched(); /* kick ourselves to get things going. */ } static void check_cpu_stall(struct rcu_state *rsp, struct rcu_data *rdp) { |
bad6e1393
|
605 606 |
unsigned long j; unsigned long js; |
64db4cfff
|
607 |
struct rcu_node *rnp; |
742734eea
|
608 |
if (rcu_cpu_stall_suppress) |
c68de2097
|
609 |
return; |
bad6e1393
|
610 611 |
j = ACCESS_ONCE(jiffies); js = ACCESS_ONCE(rsp->jiffies_stall); |
64db4cfff
|
612 |
rnp = rdp->mynode; |
bad6e1393
|
613 |
if ((ACCESS_ONCE(rnp->qsmask) & rdp->grpmask) && ULONG_CMP_GE(j, js)) { |
64db4cfff
|
614 615 616 |
/* We haven't checked in, so go dump stack. */ print_cpu_stall(rsp); |
bad6e1393
|
617 618 |
} else if (rcu_gp_in_progress(rsp) && ULONG_CMP_GE(j, js + RCU_STALL_RAT_DELAY)) { |
64db4cfff
|
619 |
|
bad6e1393
|
620 |
/* They had a few time units to dump stack, so complain. */ |
64db4cfff
|
621 622 623 |
print_other_cpu_stall(rsp); } } |
c68de2097
|
624 625 |
static int rcu_panic(struct notifier_block *this, unsigned long ev, void *ptr) { |
742734eea
|
626 |
rcu_cpu_stall_suppress = 1; |
c68de2097
|
627 628 |
return NOTIFY_DONE; } |
53d84e004
|
629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 |
/** * rcu_cpu_stall_reset - prevent further stall warnings in current grace period * * Set the stall-warning timeout way off into the future, thus preventing * any RCU CPU stall-warning messages from appearing in the current set of * RCU grace periods. * * The caller must disable hard irqs. */ void rcu_cpu_stall_reset(void) { rcu_sched_state.jiffies_stall = jiffies + ULONG_MAX / 2; rcu_bh_state.jiffies_stall = jiffies + ULONG_MAX / 2; rcu_preempt_stall_reset(); } |
c68de2097
|
644 645 646 647 648 649 650 651 |
static struct notifier_block rcu_panic_block = { .notifier_call = rcu_panic, }; static void __init check_cpu_stall_init(void) { atomic_notifier_chain_register(&panic_notifier_list, &rcu_panic_block); } |
64db4cfff
|
652 653 654 |
/* * Update CPU-local rcu_data state to record the newly noticed grace period. * This is used both when we started the grace period and when we notice |
9160306e6
|
655 656 657 |
* that someone else started the grace period. The caller must hold the * ->lock of the leaf rcu_node structure corresponding to the current CPU, * and must have irqs disabled. |
64db4cfff
|
658 |
*/ |
9160306e6
|
659 660 661 |
static void __note_new_gpnum(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp) { if (rdp->gpnum != rnp->gpnum) { |
121dfc4b3
|
662 663 664 665 666 |
/* * If the current grace period is waiting for this CPU, * set up to detect a quiescent state, otherwise don't * go looking for one. */ |
9160306e6
|
667 |
rdp->gpnum = rnp->gpnum; |
d4c08f2ac
|
668 |
trace_rcu_grace_period(rsp->name, rdp->gpnum, "cpustart"); |
121dfc4b3
|
669 670 |
if (rnp->qsmask & rdp->grpmask) { rdp->qs_pending = 1; |
e4cc1f22b
|
671 |
rdp->passed_quiesce = 0; |
121dfc4b3
|
672 673 |
} else rdp->qs_pending = 0; |
9160306e6
|
674 675 |
} } |
64db4cfff
|
676 677 |
static void note_new_gpnum(struct rcu_state *rsp, struct rcu_data *rdp) { |
9160306e6
|
678 679 680 681 682 683 |
unsigned long flags; struct rcu_node *rnp; local_irq_save(flags); rnp = rdp->mynode; if (rdp->gpnum == ACCESS_ONCE(rnp->gpnum) || /* outside lock. */ |
1304afb22
|
684 |
!raw_spin_trylock(&rnp->lock)) { /* irqs already off, so later. */ |
9160306e6
|
685 686 687 688 |
local_irq_restore(flags); return; } __note_new_gpnum(rsp, rnp, rdp); |
1304afb22
|
689 |
raw_spin_unlock_irqrestore(&rnp->lock, flags); |
64db4cfff
|
690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 |
} /* * Did someone else start a new RCU grace period start since we last * checked? Update local state appropriately if so. Must be called * on the CPU corresponding to rdp. */ static int check_for_new_grace_period(struct rcu_state *rsp, struct rcu_data *rdp) { unsigned long flags; int ret = 0; local_irq_save(flags); if (rdp->gpnum != rsp->gpnum) { note_new_gpnum(rsp, rdp); ret = 1; } local_irq_restore(flags); return ret; } /* |
d09b62dfa
|
713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 |
* Advance this CPU's callbacks, but only if the current grace period * has ended. This may be called only from the CPU to whom the rdp * belongs. In addition, the corresponding leaf rcu_node structure's * ->lock must be held by the caller, with irqs disabled. */ static void __rcu_process_gp_end(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp) { /* Did another grace period end? */ if (rdp->completed != rnp->completed) { /* Advance callbacks. No harm if list empty. */ rdp->nxttail[RCU_DONE_TAIL] = rdp->nxttail[RCU_WAIT_TAIL]; rdp->nxttail[RCU_WAIT_TAIL] = rdp->nxttail[RCU_NEXT_READY_TAIL]; rdp->nxttail[RCU_NEXT_READY_TAIL] = rdp->nxttail[RCU_NEXT_TAIL]; /* Remember that we saw this grace-period completion. */ rdp->completed = rnp->completed; |
d4c08f2ac
|
731 |
trace_rcu_grace_period(rsp->name, rdp->gpnum, "cpuend"); |
20377f32d
|
732 733 |
/* |
5ff8e6f05
|
734 |
* If we were in an extended quiescent state, we may have |
121dfc4b3
|
735 |
* missed some grace periods that others CPUs handled on |
5ff8e6f05
|
736 |
* our behalf. Catch up with this state to avoid noting |
121dfc4b3
|
737 738 739 |
* spurious new grace periods. If another grace period * has started, then rnp->gpnum will have advanced, so * we will detect this later on. |
5ff8e6f05
|
740 |
*/ |
121dfc4b3
|
741 |
if (ULONG_CMP_LT(rdp->gpnum, rdp->completed)) |
5ff8e6f05
|
742 743 744 |
rdp->gpnum = rdp->completed; /* |
121dfc4b3
|
745 746 |
* If RCU does not need a quiescent state from this CPU, * then make sure that this CPU doesn't go looking for one. |
20377f32d
|
747 |
*/ |
121dfc4b3
|
748 |
if ((rnp->qsmask & rdp->grpmask) == 0) |
20377f32d
|
749 |
rdp->qs_pending = 0; |
d09b62dfa
|
750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 |
} } /* * Advance this CPU's callbacks, but only if the current grace period * has ended. This may be called only from the CPU to whom the rdp * belongs. */ static void rcu_process_gp_end(struct rcu_state *rsp, struct rcu_data *rdp) { unsigned long flags; struct rcu_node *rnp; local_irq_save(flags); rnp = rdp->mynode; if (rdp->completed == ACCESS_ONCE(rnp->completed) || /* outside lock. */ |
1304afb22
|
767 |
!raw_spin_trylock(&rnp->lock)) { /* irqs already off, so later. */ |
d09b62dfa
|
768 769 770 771 |
local_irq_restore(flags); return; } __rcu_process_gp_end(rsp, rnp, rdp); |
1304afb22
|
772 |
raw_spin_unlock_irqrestore(&rnp->lock, flags); |
d09b62dfa
|
773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 |
} /* * Do per-CPU grace-period initialization for running CPU. The caller * must hold the lock of the leaf rcu_node structure corresponding to * this CPU. */ static void rcu_start_gp_per_cpu(struct rcu_state *rsp, struct rcu_node *rnp, struct rcu_data *rdp) { /* Prior grace period ended, so advance callbacks for current CPU. */ __rcu_process_gp_end(rsp, rnp, rdp); /* * Because this CPU just now started the new grace period, we know * that all of its callbacks will be covered by this upcoming grace * period, even the ones that were registered arbitrarily recently. * Therefore, advance all outstanding callbacks to RCU_WAIT_TAIL. * * Other CPUs cannot be sure exactly when the grace period started. * Therefore, their recently registered callbacks must pass through * an additional RCU_NEXT_READY stage, so that they will be handled * by the next RCU grace period. */ rdp->nxttail[RCU_NEXT_READY_TAIL] = rdp->nxttail[RCU_NEXT_TAIL]; rdp->nxttail[RCU_WAIT_TAIL] = rdp->nxttail[RCU_NEXT_TAIL]; |
9160306e6
|
799 800 801 |
/* Set state so that this CPU will detect the next quiescent state. */ __note_new_gpnum(rsp, rnp, rdp); |
d09b62dfa
|
802 803 804 |
} /* |
64db4cfff
|
805 806 807 808 809 810 811 812 813 |
* Start a new RCU grace period if warranted, re-initializing the hierarchy * in preparation for detecting the next grace period. The caller must hold * the root node's ->lock, which is released before return. Hard irqs must * be disabled. */ static void rcu_start_gp(struct rcu_state *rsp, unsigned long flags) __releases(rcu_get_root(rsp)->lock) { |
394f99a90
|
814 |
struct rcu_data *rdp = this_cpu_ptr(rsp->rda); |
64db4cfff
|
815 |
struct rcu_node *rnp = rcu_get_root(rsp); |
64db4cfff
|
816 |
|
037067a1b
|
817 |
if (!rcu_scheduler_fully_active || |
afe24b122
|
818 819 820 821 822 823 824 825 826 827 |
!cpu_needs_another_gp(rsp, rdp)) { /* * Either the scheduler hasn't yet spawned the first * non-idle task or this CPU does not need another * grace period. Either way, don't start a new grace * period. */ raw_spin_unlock_irqrestore(&rnp->lock, flags); return; } |
b32e9eb6a
|
828 |
|
afe24b122
|
829 |
if (rsp->fqs_active) { |
b32e9eb6a
|
830 |
/* |
afe24b122
|
831 832 |
* This CPU needs a grace period, but force_quiescent_state() * is running. Tell it to start one on this CPU's behalf. |
b32e9eb6a
|
833 |
*/ |
afe24b122
|
834 835 |
rsp->fqs_need_gp = 1; raw_spin_unlock_irqrestore(&rnp->lock, flags); |
64db4cfff
|
836 837 838 839 840 |
return; } /* Advance to a new grace period and initialize state. */ rsp->gpnum++; |
d4c08f2ac
|
841 |
trace_rcu_grace_period(rsp->name, rsp->gpnum, "start"); |
c3422bea5
|
842 |
WARN_ON_ONCE(rsp->signaled == RCU_GP_INIT); |
64db4cfff
|
843 844 |
rsp->signaled = RCU_GP_INIT; /* Hold off force_quiescent_state. */ rsp->jiffies_force_qs = jiffies + RCU_JIFFIES_TILL_FORCE_QS; |
64db4cfff
|
845 |
record_gp_stall_check_time(rsp); |
64db4cfff
|
846 |
|
64db4cfff
|
847 848 |
/* Special-case the common single-level case. */ if (NUM_RCU_NODES == 1) { |
b0e165c03
|
849 |
rcu_preempt_check_blocked_tasks(rnp); |
28ecd5802
|
850 |
rnp->qsmask = rnp->qsmaskinit; |
de078d875
|
851 |
rnp->gpnum = rsp->gpnum; |
d09b62dfa
|
852 |
rnp->completed = rsp->completed; |
c12172c02
|
853 |
rsp->signaled = RCU_SIGNAL_INIT; /* force_quiescent_state OK. */ |
d09b62dfa
|
854 |
rcu_start_gp_per_cpu(rsp, rnp, rdp); |
27f4d2805
|
855 |
rcu_preempt_boost_start_gp(rnp); |
d4c08f2ac
|
856 857 858 |
trace_rcu_grace_period_init(rsp->name, rnp->gpnum, rnp->level, rnp->grplo, rnp->grphi, rnp->qsmask); |
1304afb22
|
859 |
raw_spin_unlock_irqrestore(&rnp->lock, flags); |
64db4cfff
|
860 861 |
return; } |
1304afb22
|
862 |
raw_spin_unlock(&rnp->lock); /* leave irqs disabled. */ |
64db4cfff
|
863 864 865 |
/* Exclude any concurrent CPU-hotplug operations. */ |
1304afb22
|
866 |
raw_spin_lock(&rsp->onofflock); /* irqs already disabled. */ |
64db4cfff
|
867 868 |
/* |
b835db1f9
|
869 870 871 872 873 874 875 876 877 |
* Set the quiescent-state-needed bits in all the rcu_node * structures for all currently online CPUs in breadth-first * order, starting from the root rcu_node structure. This * operation relies on the layout of the hierarchy within the * rsp->node[] array. Note that other CPUs will access only * the leaves of the hierarchy, which still indicate that no * grace period is in progress, at least until the corresponding * leaf node has been initialized. In addition, we have excluded * CPU-hotplug operations. |
64db4cfff
|
878 879 880 881 |
* * Note that the grace period cannot complete until we finish * the initialization process, as there will be at least one * qsmask bit set in the root node until that time, namely the |
b835db1f9
|
882 883 |
* one corresponding to this CPU, due to the fact that we have * irqs disabled. |
64db4cfff
|
884 |
*/ |
a0b6c9a78
|
885 |
rcu_for_each_node_breadth_first(rsp, rnp) { |
1304afb22
|
886 |
raw_spin_lock(&rnp->lock); /* irqs already disabled. */ |
b0e165c03
|
887 |
rcu_preempt_check_blocked_tasks(rnp); |
49e291266
|
888 |
rnp->qsmask = rnp->qsmaskinit; |
de078d875
|
889 |
rnp->gpnum = rsp->gpnum; |
d09b62dfa
|
890 891 892 |
rnp->completed = rsp->completed; if (rnp == rdp->mynode) rcu_start_gp_per_cpu(rsp, rnp, rdp); |
27f4d2805
|
893 |
rcu_preempt_boost_start_gp(rnp); |
d4c08f2ac
|
894 895 896 |
trace_rcu_grace_period_init(rsp->name, rnp->gpnum, rnp->level, rnp->grplo, rnp->grphi, rnp->qsmask); |
1304afb22
|
897 |
raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ |
64db4cfff
|
898 |
} |
83f5b01ff
|
899 |
rnp = rcu_get_root(rsp); |
1304afb22
|
900 |
raw_spin_lock(&rnp->lock); /* irqs already disabled. */ |
64db4cfff
|
901 |
rsp->signaled = RCU_SIGNAL_INIT; /* force_quiescent_state now OK. */ |
1304afb22
|
902 903 |
raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ raw_spin_unlock_irqrestore(&rsp->onofflock, flags); |
64db4cfff
|
904 905 906 |
} /* |
d3f6bad39
|
907 908 909 910 911 |
* Report a full set of quiescent states to the specified rcu_state * data structure. This involves cleaning up after the prior grace * period and letting rcu_start_gp() start up the next grace period * if one is needed. Note that the caller must hold rnp->lock, as * required by rcu_start_gp(), which will release it. |
f41d911f8
|
912 |
*/ |
d3f6bad39
|
913 |
static void rcu_report_qs_rsp(struct rcu_state *rsp, unsigned long flags) |
fc2219d49
|
914 |
__releases(rcu_get_root(rsp)->lock) |
f41d911f8
|
915 |
{ |
15ba0ba86
|
916 |
unsigned long gp_duration; |
afe24b122
|
917 918 |
struct rcu_node *rnp = rcu_get_root(rsp); struct rcu_data *rdp = this_cpu_ptr(rsp->rda); |
15ba0ba86
|
919 |
|
fc2219d49
|
920 |
WARN_ON_ONCE(!rcu_gp_in_progress(rsp)); |
0bbcc529f
|
921 922 923 924 925 926 |
/* * Ensure that all grace-period and pre-grace-period activity * is seen before the assignment to rsp->completed. */ smp_mb(); /* See above block comment. */ |
15ba0ba86
|
927 928 929 |
gp_duration = jiffies - rsp->gp_start; if (gp_duration > rsp->gp_max) rsp->gp_max = gp_duration; |
afe24b122
|
930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 |
/* * We know the grace period is complete, but to everyone else * it appears to still be ongoing. But it is also the case * that to everyone else it looks like there is nothing that * they can do to advance the grace period. It is therefore * safe for us to drop the lock in order to mark the grace * period as completed in all of the rcu_node structures. * * But if this CPU needs another grace period, it will take * care of this while initializing the next grace period. * We use RCU_WAIT_TAIL instead of the usual RCU_DONE_TAIL * because the callbacks have not yet been advanced: Those * callbacks are waiting on the grace period that just now * completed. */ if (*rdp->nxttail[RCU_WAIT_TAIL] == NULL) { raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ /* * Propagate new ->completed value to rcu_node structures * so that other CPUs don't have to wait until the start * of the next grace period to process their callbacks. */ rcu_for_each_node_breadth_first(rsp, rnp) { raw_spin_lock(&rnp->lock); /* irqs already disabled. */ rnp->completed = rsp->gpnum; raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ } rnp = rcu_get_root(rsp); raw_spin_lock(&rnp->lock); /* irqs already disabled. */ } rsp->completed = rsp->gpnum; /* Declare the grace period complete. */ |
d4c08f2ac
|
964 |
trace_rcu_grace_period(rsp->name, rsp->completed, "end"); |
83f5b01ff
|
965 |
rsp->signaled = RCU_GP_IDLE; |
f41d911f8
|
966 967 968 969 |
rcu_start_gp(rsp, flags); /* releases root node's rnp->lock. */ } /* |
d3f6bad39
|
970 971 972 973 974 975 |
* Similar to rcu_report_qs_rdp(), for which it is a helper function. * Allows quiescent states for a group of CPUs to be reported at one go * to the specified rcu_node structure, though all the CPUs in the group * must be represented by the same rcu_node structure (which need not be * a leaf rcu_node structure, though it often will be). That structure's * lock must be held upon entry, and it is released before return. |
64db4cfff
|
976 977 |
*/ static void |
d3f6bad39
|
978 979 |
rcu_report_qs_rnp(unsigned long mask, struct rcu_state *rsp, struct rcu_node *rnp, unsigned long flags) |
64db4cfff
|
980 981 |
__releases(rnp->lock) { |
28ecd5802
|
982 |
struct rcu_node *rnp_c; |
64db4cfff
|
983 984 985 986 987 |
/* Walk up the rcu_node hierarchy. */ for (;;) { if (!(rnp->qsmask & mask)) { /* Our bit has already been cleared, so done. */ |
1304afb22
|
988 |
raw_spin_unlock_irqrestore(&rnp->lock, flags); |
64db4cfff
|
989 990 991 |
return; } rnp->qsmask &= ~mask; |
d4c08f2ac
|
992 993 994 995 |
trace_rcu_quiescent_state_report(rsp->name, rnp->gpnum, mask, rnp->qsmask, rnp->level, rnp->grplo, rnp->grphi, !!rnp->gp_tasks); |
27f4d2805
|
996 |
if (rnp->qsmask != 0 || rcu_preempt_blocked_readers_cgp(rnp)) { |
64db4cfff
|
997 998 |
/* Other bits still set at this level, so done. */ |
1304afb22
|
999 |
raw_spin_unlock_irqrestore(&rnp->lock, flags); |
64db4cfff
|
1000 1001 1002 1003 1004 1005 1006 1007 1008 |
return; } mask = rnp->grpmask; if (rnp->parent == NULL) { /* No more levels. Exit loop holding root lock. */ break; } |
1304afb22
|
1009 |
raw_spin_unlock_irqrestore(&rnp->lock, flags); |
28ecd5802
|
1010 |
rnp_c = rnp; |
64db4cfff
|
1011 |
rnp = rnp->parent; |
1304afb22
|
1012 |
raw_spin_lock_irqsave(&rnp->lock, flags); |
28ecd5802
|
1013 |
WARN_ON_ONCE(rnp_c->qsmask); |
64db4cfff
|
1014 1015 1016 1017 |
} /* * Get here if we are the last CPU to pass through a quiescent |
d3f6bad39
|
1018 |
* state for this grace period. Invoke rcu_report_qs_rsp() |
f41d911f8
|
1019 |
* to clean up and start the next grace period if one is needed. |
64db4cfff
|
1020 |
*/ |
d3f6bad39
|
1021 |
rcu_report_qs_rsp(rsp, flags); /* releases rnp->lock. */ |
64db4cfff
|
1022 1023 1024 |
} /* |
d3f6bad39
|
1025 1026 1027 1028 1029 1030 1031 |
* Record a quiescent state for the specified CPU to that CPU's rcu_data * structure. This must be either called from the specified CPU, or * called when the specified CPU is known to be offline (and when it is * also known that no other CPU is concurrently trying to help the offline * CPU). The lastcomp argument is used to make sure we are still in the * grace period of interest. We don't want to end the current grace period * based on quiescent states detected in an earlier grace period! |
64db4cfff
|
1032 1033 |
*/ static void |
e4cc1f22b
|
1034 |
rcu_report_qs_rdp(int cpu, struct rcu_state *rsp, struct rcu_data *rdp, long lastgp) |
64db4cfff
|
1035 1036 1037 1038 1039 1040 |
{ unsigned long flags; unsigned long mask; struct rcu_node *rnp; rnp = rdp->mynode; |
1304afb22
|
1041 |
raw_spin_lock_irqsave(&rnp->lock, flags); |
e4cc1f22b
|
1042 |
if (lastgp != rnp->gpnum || rnp->completed == rnp->gpnum) { |
64db4cfff
|
1043 1044 |
/* |
e4cc1f22b
|
1045 1046 1047 1048 |
* The grace period in which this quiescent state was * recorded has ended, so don't report it upwards. * We will instead need a new quiescent state that lies * within the current grace period. |
64db4cfff
|
1049 |
*/ |
e4cc1f22b
|
1050 |
rdp->passed_quiesce = 0; /* need qs for new gp. */ |
1304afb22
|
1051 |
raw_spin_unlock_irqrestore(&rnp->lock, flags); |
64db4cfff
|
1052 1053 1054 1055 |
return; } mask = rdp->grpmask; if ((rnp->qsmask & mask) == 0) { |
1304afb22
|
1056 |
raw_spin_unlock_irqrestore(&rnp->lock, flags); |
64db4cfff
|
1057 1058 1059 1060 1061 1062 1063 |
} else { rdp->qs_pending = 0; /* * This GP can't end until cpu checks in, so all of our * callbacks can be processed during the next GP. */ |
64db4cfff
|
1064 |
rdp->nxttail[RCU_NEXT_READY_TAIL] = rdp->nxttail[RCU_NEXT_TAIL]; |
d3f6bad39
|
1065 |
rcu_report_qs_rnp(mask, rsp, rnp, flags); /* rlses rnp->lock */ |
64db4cfff
|
1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 |
} } /* * Check to see if there is a new grace period of which this CPU * is not yet aware, and if so, set up local rcu_data state for it. * Otherwise, see if this CPU has just passed through its first * quiescent state for this grace period, and record that fact if so. */ static void rcu_check_quiescent_state(struct rcu_state *rsp, struct rcu_data *rdp) { /* If there is now a new grace period, record and return. */ if (check_for_new_grace_period(rsp, rdp)) return; /* * Does this CPU still need to do its part for current grace period? * If no, return and let the other CPUs do their part as well. */ if (!rdp->qs_pending) return; /* * Was there a quiescent state since the beginning of the grace * period? If no, then exit and wait for the next call. */ |
e4cc1f22b
|
1093 |
if (!rdp->passed_quiesce) |
64db4cfff
|
1094 |
return; |
d3f6bad39
|
1095 1096 1097 1098 |
/* * Tell RCU we are done (but rcu_report_qs_rdp() will be the * judge of that). */ |
e4cc1f22b
|
1099 |
rcu_report_qs_rdp(rdp->cpu, rsp, rdp, rdp->passed_quiesce_gpnum); |
64db4cfff
|
1100 1101 1102 1103 1104 |
} #ifdef CONFIG_HOTPLUG_CPU /* |
29494be71
|
1105 1106 1107 |
* Move a dying CPU's RCU callbacks to online CPU's callback list. * Synchronization is not required because this function executes * in stop_machine() context. |
e74f4c456
|
1108 |
*/ |
29494be71
|
1109 |
static void rcu_send_cbs_to_online(struct rcu_state *rsp) |
e74f4c456
|
1110 1111 |
{ int i; |
29494be71
|
1112 1113 |
/* current DYING CPU is cleared in the cpu_online_mask */ int receive_cpu = cpumask_any(cpu_online_mask); |
394f99a90
|
1114 |
struct rcu_data *rdp = this_cpu_ptr(rsp->rda); |
29494be71
|
1115 |
struct rcu_data *receive_rdp = per_cpu_ptr(rsp->rda, receive_cpu); |
e74f4c456
|
1116 1117 1118 |
if (rdp->nxtlist == NULL) return; /* irqs disabled, so comparison is stable. */ |
29494be71
|
1119 1120 1121 1122 1123 1124 |
*receive_rdp->nxttail[RCU_NEXT_TAIL] = rdp->nxtlist; receive_rdp->nxttail[RCU_NEXT_TAIL] = rdp->nxttail[RCU_NEXT_TAIL]; receive_rdp->qlen += rdp->qlen; receive_rdp->n_cbs_adopted += rdp->qlen; rdp->n_cbs_orphaned += rdp->qlen; |
e74f4c456
|
1125 1126 1127 |
rdp->nxtlist = NULL; for (i = 0; i < RCU_NEXT_SIZE; i++) rdp->nxttail[i] = &rdp->nxtlist; |
e74f4c456
|
1128 |
rdp->qlen = 0; |
e74f4c456
|
1129 1130 1131 |
} /* |
64db4cfff
|
1132 1133 |
* Remove the outgoing CPU from the bitmasks in the rcu_node hierarchy * and move all callbacks from the outgoing CPU to the current one. |
a26ac2455
|
1134 1135 |
* There can only be one CPU hotplug operation at a time, so no other * CPU can be attempting to update rcu_cpu_kthread_task. |
64db4cfff
|
1136 1137 1138 |
*/ static void __rcu_offline_cpu(int cpu, struct rcu_state *rsp) { |
64db4cfff
|
1139 |
unsigned long flags; |
64db4cfff
|
1140 |
unsigned long mask; |
d9a3da069
|
1141 |
int need_report = 0; |
394f99a90
|
1142 |
struct rcu_data *rdp = per_cpu_ptr(rsp->rda, cpu); |
64db4cfff
|
1143 |
struct rcu_node *rnp; |
a26ac2455
|
1144 |
|
f8b7fc6b5
|
1145 |
rcu_stop_cpu_kthread(cpu); |
64db4cfff
|
1146 1147 |
/* Exclude any attempts to start a new grace period. */ |
1304afb22
|
1148 |
raw_spin_lock_irqsave(&rsp->onofflock, flags); |
64db4cfff
|
1149 1150 |
/* Remove the outgoing CPU from the masks in the rcu_node hierarchy. */ |
28ecd5802
|
1151 |
rnp = rdp->mynode; /* this is the outgoing CPU's rnp. */ |
64db4cfff
|
1152 1153 |
mask = rdp->grpmask; /* rnp->grplo is constant. */ do { |
1304afb22
|
1154 |
raw_spin_lock(&rnp->lock); /* irqs already disabled. */ |
64db4cfff
|
1155 1156 |
rnp->qsmaskinit &= ~mask; if (rnp->qsmaskinit != 0) { |
b668c9cf3
|
1157 |
if (rnp != rdp->mynode) |
1304afb22
|
1158 |
raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ |
d4c08f2ac
|
1159 1160 1161 1162 1163 |
else trace_rcu_grace_period(rsp->name, rnp->gpnum + 1 - !!(rnp->qsmask & mask), "cpuofl"); |
64db4cfff
|
1164 1165 |
break; } |
d4c08f2ac
|
1166 1167 1168 1169 1170 |
if (rnp == rdp->mynode) { trace_rcu_grace_period(rsp->name, rnp->gpnum + 1 - !!(rnp->qsmask & mask), "cpuofl"); |
d9a3da069
|
1171 |
need_report = rcu_preempt_offline_tasks(rsp, rnp, rdp); |
d4c08f2ac
|
1172 |
} else |
1304afb22
|
1173 |
raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ |
64db4cfff
|
1174 |
mask = rnp->grpmask; |
64db4cfff
|
1175 1176 |
rnp = rnp->parent; } while (rnp != NULL); |
64db4cfff
|
1177 |
|
b668c9cf3
|
1178 1179 1180 |
/* * We still hold the leaf rcu_node structure lock here, and * irqs are still disabled. The reason for this subterfuge is |
d3f6bad39
|
1181 1182 |
* because invoking rcu_report_unblock_qs_rnp() with ->onofflock * held leads to deadlock. |
b668c9cf3
|
1183 |
*/ |
1304afb22
|
1184 |
raw_spin_unlock(&rsp->onofflock); /* irqs remain disabled. */ |
b668c9cf3
|
1185 |
rnp = rdp->mynode; |
d9a3da069
|
1186 |
if (need_report & RCU_OFL_TASKS_NORM_GP) |
d3f6bad39
|
1187 |
rcu_report_unblock_qs_rnp(rnp, flags); |
b668c9cf3
|
1188 |
else |
1304afb22
|
1189 |
raw_spin_unlock_irqrestore(&rnp->lock, flags); |
d9a3da069
|
1190 1191 |
if (need_report & RCU_OFL_TASKS_EXP_GP) rcu_report_exp_rnp(rsp, rnp); |
1217ed1ba
|
1192 |
rcu_node_kthread_setaffinity(rnp, -1); |
64db4cfff
|
1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 |
} /* * Remove the specified CPU from the RCU hierarchy and move any pending * callbacks that it might have to the current CPU. This code assumes * that at least one CPU in the system will remain running at all times. * Any attempt to offline -all- CPUs is likely to strand RCU callbacks. */ static void rcu_offline_cpu(int cpu) { |
d6714c22b
|
1203 |
__rcu_offline_cpu(cpu, &rcu_sched_state); |
64db4cfff
|
1204 |
__rcu_offline_cpu(cpu, &rcu_bh_state); |
33f76148c
|
1205 |
rcu_preempt_offline_cpu(cpu); |
64db4cfff
|
1206 1207 1208 |
} #else /* #ifdef CONFIG_HOTPLUG_CPU */ |
29494be71
|
1209 |
static void rcu_send_cbs_to_online(struct rcu_state *rsp) |
e74f4c456
|
1210 1211 |
{ } |
64db4cfff
|
1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 |
static void rcu_offline_cpu(int cpu) { } #endif /* #else #ifdef CONFIG_HOTPLUG_CPU */ /* * Invoke any RCU callbacks that have made it to the end of their grace * period. Thottle as specified by rdp->blimit. */ |
37c72e56f
|
1222 |
static void rcu_do_batch(struct rcu_state *rsp, struct rcu_data *rdp) |
64db4cfff
|
1223 1224 1225 |
{ unsigned long flags; struct rcu_head *next, *list, **tail; |
29c00b4a1
|
1226 |
int bl, count; |
64db4cfff
|
1227 1228 |
/* If no callbacks are ready, just return.*/ |
29c00b4a1
|
1229 |
if (!cpu_has_callbacks_ready_to_invoke(rdp)) { |
72fe701b7
|
1230 1231 |
trace_rcu_batch_start(rsp->name, 0, 0); trace_rcu_batch_end(rsp->name, 0); |
64db4cfff
|
1232 |
return; |
29c00b4a1
|
1233 |
} |
64db4cfff
|
1234 1235 1236 1237 1238 1239 |
/* * Extract the list of ready callbacks, disabling to prevent * races with call_rcu() from interrupt handlers. */ local_irq_save(flags); |
29c00b4a1
|
1240 |
bl = rdp->blimit; |
72fe701b7
|
1241 |
trace_rcu_batch_start(rsp->name, rdp->qlen, bl); |
64db4cfff
|
1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 |
list = rdp->nxtlist; rdp->nxtlist = *rdp->nxttail[RCU_DONE_TAIL]; *rdp->nxttail[RCU_DONE_TAIL] = NULL; tail = rdp->nxttail[RCU_DONE_TAIL]; for (count = RCU_NEXT_SIZE - 1; count >= 0; count--) if (rdp->nxttail[count] == rdp->nxttail[RCU_DONE_TAIL]) rdp->nxttail[count] = &rdp->nxtlist; local_irq_restore(flags); /* Invoke callbacks. */ count = 0; while (list) { next = list->next; prefetch(next); |
551d55a94
|
1256 |
debug_rcu_head_unqueue(list); |
d4c08f2ac
|
1257 |
__rcu_reclaim(rsp->name, list); |
64db4cfff
|
1258 |
list = next; |
29c00b4a1
|
1259 |
if (++count >= bl) |
64db4cfff
|
1260 1261 1262 1263 |
break; } local_irq_save(flags); |
72fe701b7
|
1264 |
trace_rcu_batch_end(rsp->name, count); |
64db4cfff
|
1265 1266 1267 |
/* Update count, and requeue any remaining callbacks. */ rdp->qlen -= count; |
269dcc1c2
|
1268 |
rdp->n_cbs_invoked += count; |
64db4cfff
|
1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 |
if (list != NULL) { *tail = rdp->nxtlist; rdp->nxtlist = list; for (count = 0; count < RCU_NEXT_SIZE; count++) if (&rdp->nxtlist == rdp->nxttail[count]) rdp->nxttail[count] = tail; else break; } /* Reinstate batch limit if we have worked down the excess. */ if (rdp->blimit == LONG_MAX && rdp->qlen <= qlowmark) rdp->blimit = blimit; |
37c72e56f
|
1282 1283 1284 1285 1286 1287 |
/* Reset ->qlen_last_fqs_check trigger if enough CBs have drained. */ if (rdp->qlen == 0 && rdp->qlen_last_fqs_check != 0) { rdp->qlen_last_fqs_check = 0; rdp->n_force_qs_snap = rsp->n_force_qs; } else if (rdp->qlen < rdp->qlen_last_fqs_check - qhimark) rdp->qlen_last_fqs_check = rdp->qlen; |
64db4cfff
|
1288 |
local_irq_restore(flags); |
e0f23060a
|
1289 |
/* Re-invoke RCU core processing if there are callbacks remaining. */ |
64db4cfff
|
1290 |
if (cpu_has_callbacks_ready_to_invoke(rdp)) |
a46e0899e
|
1291 |
invoke_rcu_core(); |
64db4cfff
|
1292 1293 1294 1295 1296 |
} /* * Check to see if this CPU is in a non-context-switch quiescent state * (user mode or idle loop for rcu, non-softirq execution for rcu_bh). |
e0f23060a
|
1297 |
* Also schedule RCU core processing. |
64db4cfff
|
1298 1299 1300 1301 1302 1303 1304 |
* * This function must be called with hardirqs disabled. It is normally * invoked from the scheduling-clock interrupt. If rcu_pending returns * false, there is no point in invoking rcu_check_callbacks(). */ void rcu_check_callbacks(int cpu, int user) { |
300df91ca
|
1305 |
trace_rcu_utilization("Start scheduler-tick"); |
64db4cfff
|
1306 |
if (user || |
a68260483
|
1307 1308 |
(idle_cpu(cpu) && rcu_scheduler_active && !in_softirq() && hardirq_count() <= (1 << HARDIRQ_SHIFT))) { |
64db4cfff
|
1309 1310 1311 1312 1313 |
/* * Get here if this CPU took its interrupt from user * mode or from the idle loop, and if this is not a * nested interrupt. In this case, the CPU is in |
d6714c22b
|
1314 |
* a quiescent state, so note it. |
64db4cfff
|
1315 1316 |
* * No memory barrier is required here because both |
d6714c22b
|
1317 1318 1319 |
* rcu_sched_qs() and rcu_bh_qs() reference only CPU-local * variables that other CPUs neither access nor modify, * at least not while the corresponding CPU is online. |
64db4cfff
|
1320 |
*/ |
d6714c22b
|
1321 1322 |
rcu_sched_qs(cpu); rcu_bh_qs(cpu); |
64db4cfff
|
1323 1324 1325 1326 1327 1328 1329 |
} else if (!in_softirq()) { /* * Get here if this CPU did not take its interrupt from * softirq, in other words, if it is not interrupting * a rcu_bh read-side critical section. This is an _bh |
d6714c22b
|
1330 |
* critical section, so note it. |
64db4cfff
|
1331 |
*/ |
d6714c22b
|
1332 |
rcu_bh_qs(cpu); |
64db4cfff
|
1333 |
} |
f41d911f8
|
1334 |
rcu_preempt_check_callbacks(cpu); |
d21670aca
|
1335 |
if (rcu_pending(cpu)) |
a46e0899e
|
1336 |
invoke_rcu_core(); |
300df91ca
|
1337 |
trace_rcu_utilization("End scheduler-tick"); |
64db4cfff
|
1338 1339 1340 1341 1342 1343 1344 |
} #ifdef CONFIG_SMP /* * Scan the leaf rcu_node structures, processing dyntick state for any that * have not yet encountered a quiescent state, using the function specified. |
27f4d2805
|
1345 1346 |
* Also initiate boosting for any threads blocked on the root rcu_node. * |
ee47eb9f4
|
1347 |
* The caller must have suppressed start of new grace periods. |
64db4cfff
|
1348 |
*/ |
45f014c52
|
1349 |
static void force_qs_rnp(struct rcu_state *rsp, int (*f)(struct rcu_data *)) |
64db4cfff
|
1350 1351 1352 1353 1354 |
{ unsigned long bit; int cpu; unsigned long flags; unsigned long mask; |
a0b6c9a78
|
1355 |
struct rcu_node *rnp; |
64db4cfff
|
1356 |
|
a0b6c9a78
|
1357 |
rcu_for_each_leaf_node(rsp, rnp) { |
64db4cfff
|
1358 |
mask = 0; |
1304afb22
|
1359 |
raw_spin_lock_irqsave(&rnp->lock, flags); |
ee47eb9f4
|
1360 |
if (!rcu_gp_in_progress(rsp)) { |
1304afb22
|
1361 |
raw_spin_unlock_irqrestore(&rnp->lock, flags); |
0f10dc826
|
1362 |
return; |
64db4cfff
|
1363 |
} |
a0b6c9a78
|
1364 |
if (rnp->qsmask == 0) { |
1217ed1ba
|
1365 |
rcu_initiate_boost(rnp, flags); /* releases rnp->lock */ |
64db4cfff
|
1366 1367 |
continue; } |
a0b6c9a78
|
1368 |
cpu = rnp->grplo; |
64db4cfff
|
1369 |
bit = 1; |
a0b6c9a78
|
1370 |
for (; cpu <= rnp->grphi; cpu++, bit <<= 1) { |
394f99a90
|
1371 1372 |
if ((rnp->qsmask & bit) != 0 && f(per_cpu_ptr(rsp->rda, cpu))) |
64db4cfff
|
1373 1374 |
mask |= bit; } |
45f014c52
|
1375 |
if (mask != 0) { |
64db4cfff
|
1376 |
|
d3f6bad39
|
1377 1378 |
/* rcu_report_qs_rnp() releases rnp->lock. */ rcu_report_qs_rnp(mask, rsp, rnp, flags); |
64db4cfff
|
1379 1380 |
continue; } |
1304afb22
|
1381 |
raw_spin_unlock_irqrestore(&rnp->lock, flags); |
64db4cfff
|
1382 |
} |
27f4d2805
|
1383 |
rnp = rcu_get_root(rsp); |
1217ed1ba
|
1384 1385 1386 1387 |
if (rnp->qsmask == 0) { raw_spin_lock_irqsave(&rnp->lock, flags); rcu_initiate_boost(rnp, flags); /* releases rnp->lock. */ } |
64db4cfff
|
1388 1389 1390 1391 1392 1393 1394 1395 1396 |
} /* * Force quiescent states on reluctant CPUs, and also detect which * CPUs are in dyntick-idle mode. */ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) { unsigned long flags; |
64db4cfff
|
1397 |
struct rcu_node *rnp = rcu_get_root(rsp); |
64db4cfff
|
1398 |
|
300df91ca
|
1399 1400 1401 |
trace_rcu_utilization("Start fqs"); if (!rcu_gp_in_progress(rsp)) { trace_rcu_utilization("End fqs"); |
64db4cfff
|
1402 |
return; /* No grace period in progress, nothing to force. */ |
300df91ca
|
1403 |
} |
1304afb22
|
1404 |
if (!raw_spin_trylock_irqsave(&rsp->fqslock, flags)) { |
64db4cfff
|
1405 |
rsp->n_force_qs_lh++; /* Inexact, can lose counts. Tough! */ |
300df91ca
|
1406 |
trace_rcu_utilization("End fqs"); |
64db4cfff
|
1407 1408 |
return; /* Someone else is already on the job. */ } |
20133cfce
|
1409 |
if (relaxed && ULONG_CMP_GE(rsp->jiffies_force_qs, jiffies)) |
f96e9232e
|
1410 |
goto unlock_fqs_ret; /* no emergency and done recently. */ |
64db4cfff
|
1411 |
rsp->n_force_qs++; |
1304afb22
|
1412 |
raw_spin_lock(&rnp->lock); /* irqs already disabled */ |
64db4cfff
|
1413 |
rsp->jiffies_force_qs = jiffies + RCU_JIFFIES_TILL_FORCE_QS; |
560d4bc0d
|
1414 |
if(!rcu_gp_in_progress(rsp)) { |
64db4cfff
|
1415 |
rsp->n_force_qs_ngp++; |
1304afb22
|
1416 |
raw_spin_unlock(&rnp->lock); /* irqs remain disabled */ |
f96e9232e
|
1417 |
goto unlock_fqs_ret; /* no GP in progress, time updated. */ |
64db4cfff
|
1418 |
} |
07079d535
|
1419 |
rsp->fqs_active = 1; |
f3a8b5c6a
|
1420 |
switch (rsp->signaled) { |
83f5b01ff
|
1421 |
case RCU_GP_IDLE: |
64db4cfff
|
1422 |
case RCU_GP_INIT: |
83f5b01ff
|
1423 |
break; /* grace period idle or initializing, ignore. */ |
64db4cfff
|
1424 1425 |
case RCU_SAVE_DYNTICK: |
64db4cfff
|
1426 1427 |
if (RCU_SIGNAL_INIT != RCU_SAVE_DYNTICK) break; /* So gcc recognizes the dead code. */ |
f261414f0
|
1428 |
raw_spin_unlock(&rnp->lock); /* irqs remain disabled */ |
64db4cfff
|
1429 |
/* Record dyntick-idle state. */ |
45f014c52
|
1430 |
force_qs_rnp(rsp, dyntick_save_progress_counter); |
1304afb22
|
1431 |
raw_spin_lock(&rnp->lock); /* irqs already disabled */ |
ee47eb9f4
|
1432 |
if (rcu_gp_in_progress(rsp)) |
64db4cfff
|
1433 |
rsp->signaled = RCU_FORCE_QS; |
ee47eb9f4
|
1434 |
break; |
64db4cfff
|
1435 1436 1437 1438 |
case RCU_FORCE_QS: /* Check dyntick-idle state, send IPI to laggarts. */ |
1304afb22
|
1439 |
raw_spin_unlock(&rnp->lock); /* irqs remain disabled */ |
45f014c52
|
1440 |
force_qs_rnp(rsp, rcu_implicit_dynticks_qs); |
64db4cfff
|
1441 1442 |
/* Leave state in case more forcing is required. */ |
1304afb22
|
1443 |
raw_spin_lock(&rnp->lock); /* irqs already disabled */ |
f96e9232e
|
1444 |
break; |
64db4cfff
|
1445 |
} |
07079d535
|
1446 |
rsp->fqs_active = 0; |
46a1e34ed
|
1447 |
if (rsp->fqs_need_gp) { |
1304afb22
|
1448 |
raw_spin_unlock(&rsp->fqslock); /* irqs remain disabled */ |
46a1e34ed
|
1449 1450 |
rsp->fqs_need_gp = 0; rcu_start_gp(rsp, flags); /* releases rnp->lock */ |
300df91ca
|
1451 |
trace_rcu_utilization("End fqs"); |
46a1e34ed
|
1452 1453 |
return; } |
1304afb22
|
1454 |
raw_spin_unlock(&rnp->lock); /* irqs remain disabled */ |
f96e9232e
|
1455 |
unlock_fqs_ret: |
1304afb22
|
1456 |
raw_spin_unlock_irqrestore(&rsp->fqslock, flags); |
300df91ca
|
1457 |
trace_rcu_utilization("End fqs"); |
64db4cfff
|
1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 |
} #else /* #ifdef CONFIG_SMP */ static void force_quiescent_state(struct rcu_state *rsp, int relaxed) { set_need_resched(); } #endif /* #else #ifdef CONFIG_SMP */ /* |
e0f23060a
|
1470 1471 1472 |
* This does the RCU core processing work for the specified rcu_state * and rcu_data structures. This may be called only from the CPU to * whom the rdp belongs. |
64db4cfff
|
1473 1474 1475 1476 1477 |
*/ static void __rcu_process_callbacks(struct rcu_state *rsp, struct rcu_data *rdp) { unsigned long flags; |
2e5975580
|
1478 |
WARN_ON_ONCE(rdp->beenonline == 0); |
64db4cfff
|
1479 1480 1481 1482 |
/* * If an RCU GP has gone long enough, go check for dyntick * idle CPUs and, if needed, send resched IPIs. */ |
20133cfce
|
1483 |
if (ULONG_CMP_LT(ACCESS_ONCE(rsp->jiffies_force_qs), jiffies)) |
64db4cfff
|
1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 |
force_quiescent_state(rsp, 1); /* * Advance callbacks in response to end of earlier grace * period that some other CPU ended. */ rcu_process_gp_end(rsp, rdp); /* Update RCU state based on any recent quiescent states. */ rcu_check_quiescent_state(rsp, rdp); /* Does this CPU require a not-yet-started grace period? */ if (cpu_needs_another_gp(rsp, rdp)) { |
1304afb22
|
1497 |
raw_spin_lock_irqsave(&rcu_get_root(rsp)->lock, flags); |
64db4cfff
|
1498 1499 1500 1501 |
rcu_start_gp(rsp, flags); /* releases above lock */ } /* If there are callbacks ready, invoke them. */ |
09223371d
|
1502 |
if (cpu_has_callbacks_ready_to_invoke(rdp)) |
a46e0899e
|
1503 |
invoke_rcu_callbacks(rsp, rdp); |
09223371d
|
1504 |
} |
64db4cfff
|
1505 |
/* |
e0f23060a
|
1506 |
* Do RCU core processing for the current CPU. |
64db4cfff
|
1507 |
*/ |
09223371d
|
1508 |
static void rcu_process_callbacks(struct softirq_action *unused) |
64db4cfff
|
1509 |
{ |
300df91ca
|
1510 |
trace_rcu_utilization("Start RCU core"); |
d6714c22b
|
1511 1512 |
__rcu_process_callbacks(&rcu_sched_state, &__get_cpu_var(rcu_sched_data)); |
64db4cfff
|
1513 |
__rcu_process_callbacks(&rcu_bh_state, &__get_cpu_var(rcu_bh_data)); |
f41d911f8
|
1514 |
rcu_preempt_process_callbacks(); |
300df91ca
|
1515 |
trace_rcu_utilization("End RCU core"); |
64db4cfff
|
1516 |
} |
a26ac2455
|
1517 |
/* |
e0f23060a
|
1518 1519 1520 1521 1522 |
* Schedule RCU callback invocation. If the specified type of RCU * does not support RCU priority boosting, just do a direct call, * otherwise wake up the per-CPU kernel kthread. Note that because we * are running on the current CPU with interrupts disabled, the * rcu_cpu_kthread_task cannot disappear out from under us. |
a26ac2455
|
1523 |
*/ |
a46e0899e
|
1524 |
static void invoke_rcu_callbacks(struct rcu_state *rsp, struct rcu_data *rdp) |
a26ac2455
|
1525 |
{ |
b0d304172
|
1526 1527 |
if (unlikely(!ACCESS_ONCE(rcu_scheduler_fully_active))) return; |
a46e0899e
|
1528 1529 |
if (likely(!rsp->boost)) { rcu_do_batch(rsp, rdp); |
a26ac2455
|
1530 1531 |
return; } |
a46e0899e
|
1532 |
invoke_rcu_callbacks_kthread(); |
a26ac2455
|
1533 |
} |
a46e0899e
|
1534 |
static void invoke_rcu_core(void) |
09223371d
|
1535 1536 1537 |
{ raise_softirq(RCU_SOFTIRQ); } |
64db4cfff
|
1538 1539 1540 1541 1542 1543 |
static void __call_rcu(struct rcu_head *head, void (*func)(struct rcu_head *rcu), struct rcu_state *rsp) { unsigned long flags; struct rcu_data *rdp; |
551d55a94
|
1544 |
debug_rcu_head_queue(head); |
64db4cfff
|
1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 |
head->func = func; head->next = NULL; smp_mb(); /* Ensure RCU update seen before callback registry. */ /* * Opportunistically note grace-period endings and beginnings. * Note that we might see a beginning right after we see an * end, but never vice versa, since this CPU has to pass through * a quiescent state betweentimes. */ local_irq_save(flags); |
394f99a90
|
1557 |
rdp = this_cpu_ptr(rsp->rda); |
64db4cfff
|
1558 1559 1560 1561 |
/* Add the callback to our list. */ *rdp->nxttail[RCU_NEXT_TAIL] = head; rdp->nxttail[RCU_NEXT_TAIL] = &head->next; |
2655d57ef
|
1562 |
rdp->qlen++; |
d4c08f2ac
|
1563 1564 1565 1566 1567 |
if (__is_kfree_rcu_offset((unsigned long)func)) trace_rcu_kfree_callback(rsp->name, head, (unsigned long)func, rdp->qlen); else trace_rcu_callback(rsp->name, head, rdp->qlen); |
2655d57ef
|
1568 1569 1570 1571 1572 |
/* If interrupts were disabled, don't dive into RCU core. */ if (irqs_disabled_flags(flags)) { local_irq_restore(flags); return; } |
64db4cfff
|
1573 |
|
37c72e56f
|
1574 1575 1576 1577 1578 1579 1580 |
/* * Force the grace period if too many callbacks or too long waiting. * Enforce hysteresis, and don't invoke force_quiescent_state() * if some other CPU has recently done so. Also, don't bother * invoking force_quiescent_state() if the newly enqueued callback * is the only one waiting for a grace period to complete. */ |
2655d57ef
|
1581 |
if (unlikely(rdp->qlen > rdp->qlen_last_fqs_check + qhimark)) { |
b52573d27
|
1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 |
/* Are we ignoring a completed grace period? */ rcu_process_gp_end(rsp, rdp); check_for_new_grace_period(rsp, rdp); /* Start a new grace period if one not already started. */ if (!rcu_gp_in_progress(rsp)) { unsigned long nestflag; struct rcu_node *rnp_root = rcu_get_root(rsp); raw_spin_lock_irqsave(&rnp_root->lock, nestflag); rcu_start_gp(rsp, nestflag); /* rlses rnp_root->lock */ } else { /* Give the grace period a kick. */ rdp->blimit = LONG_MAX; if (rsp->n_force_qs == rdp->n_force_qs_snap && *rdp->nxttail[RCU_DONE_TAIL] != head) force_quiescent_state(rsp, 0); rdp->n_force_qs_snap = rsp->n_force_qs; rdp->qlen_last_fqs_check = rdp->qlen; } |
20133cfce
|
1603 |
} else if (ULONG_CMP_LT(ACCESS_ONCE(rsp->jiffies_force_qs), jiffies)) |
64db4cfff
|
1604 1605 1606 1607 1608 |
force_quiescent_state(rsp, 1); local_irq_restore(flags); } /* |
d6714c22b
|
1609 |
* Queue an RCU-sched callback for invocation after a grace period. |
64db4cfff
|
1610 |
*/ |
d6714c22b
|
1611 |
void call_rcu_sched(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) |
64db4cfff
|
1612 |
{ |
d6714c22b
|
1613 |
__call_rcu(head, func, &rcu_sched_state); |
64db4cfff
|
1614 |
} |
d6714c22b
|
1615 |
EXPORT_SYMBOL_GPL(call_rcu_sched); |
64db4cfff
|
1616 1617 1618 1619 1620 1621 1622 1623 1624 |
/* * Queue an RCU for invocation after a quicker grace period. */ void call_rcu_bh(struct rcu_head *head, void (*func)(struct rcu_head *rcu)) { __call_rcu(head, func, &rcu_bh_state); } EXPORT_SYMBOL_GPL(call_rcu_bh); |
6ebb237be
|
1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 |
/** * synchronize_sched - wait until an rcu-sched grace period has elapsed. * * Control will return to the caller some time after a full rcu-sched * grace period has elapsed, in other words after all currently executing * rcu-sched read-side critical sections have completed. These read-side * critical sections are delimited by rcu_read_lock_sched() and * rcu_read_unlock_sched(), and may be nested. Note that preempt_disable(), * local_irq_disable(), and so on may be used in place of * rcu_read_lock_sched(). * * This means that all preempt_disable code sequences, including NMI and * hardware-interrupt handlers, in progress on entry will have completed * before this primitive returns. However, this does not guarantee that * softirq handlers will have completed, since in some kernels, these * handlers can run in process context, and can block. * * This primitive provides the guarantees made by the (now removed) * synchronize_kernel() API. In contrast, synchronize_rcu() only * guarantees that rcu_read_lock() sections will have completed. * In "classic RCU", these two guarantees happen to be one and * the same, but can differ in realtime RCU implementations. */ void synchronize_sched(void) { |
6ebb237be
|
1650 1651 |
if (rcu_blocking_is_gp()) return; |
2c42818e9
|
1652 |
wait_rcu_gp(call_rcu_sched); |
6ebb237be
|
1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 |
} EXPORT_SYMBOL_GPL(synchronize_sched); /** * synchronize_rcu_bh - wait until an rcu_bh grace period has elapsed. * * Control will return to the caller some time after a full rcu_bh grace * period has elapsed, in other words after all currently executing rcu_bh * read-side critical sections have completed. RCU read-side critical * sections are delimited by rcu_read_lock_bh() and rcu_read_unlock_bh(), * and may be nested. */ void synchronize_rcu_bh(void) { |
6ebb237be
|
1667 1668 |
if (rcu_blocking_is_gp()) return; |
2c42818e9
|
1669 |
wait_rcu_gp(call_rcu_bh); |
6ebb237be
|
1670 1671 |
} EXPORT_SYMBOL_GPL(synchronize_rcu_bh); |
64db4cfff
|
1672 1673 1674 1675 1676 1677 1678 1679 1680 |
/* * Check to see if there is any immediate RCU-related work to be done * by the current CPU, for the specified type of RCU, returning 1 if so. * The checks are in order of increasing expense: checks that can be * carried out against CPU-local state are performed first. However, * we must check for CPU stalls first, else we might not get a chance. */ static int __rcu_pending(struct rcu_state *rsp, struct rcu_data *rdp) { |
2f51f9884
|
1681 |
struct rcu_node *rnp = rdp->mynode; |
64db4cfff
|
1682 1683 1684 1685 1686 1687 |
rdp->n_rcu_pending++; /* Check for CPU stalls, if enabled. */ check_cpu_stall(rsp, rdp); /* Is the RCU core waiting for a quiescent state from this CPU? */ |
5c51dd734
|
1688 1689 |
if (rcu_scheduler_fully_active && rdp->qs_pending && !rdp->passed_quiesce) { |
d25eb9442
|
1690 1691 1692 1693 1694 1695 |
/* * If force_quiescent_state() coming soon and this CPU * needs a quiescent state, and this is either RCU-sched * or RCU-bh, force a local reschedule. */ |
d21670aca
|
1696 |
rdp->n_rp_qs_pending++; |
6cc68793e
|
1697 |
if (!rdp->preemptible && |
d25eb9442
|
1698 1699 1700 |
ULONG_CMP_LT(ACCESS_ONCE(rsp->jiffies_force_qs) - 1, jiffies)) set_need_resched(); |
e4cc1f22b
|
1701 |
} else if (rdp->qs_pending && rdp->passed_quiesce) { |
d21670aca
|
1702 |
rdp->n_rp_report_qs++; |
64db4cfff
|
1703 |
return 1; |
7ba5c840e
|
1704 |
} |
64db4cfff
|
1705 1706 |
/* Does this CPU have callbacks ready to invoke? */ |
7ba5c840e
|
1707 1708 |
if (cpu_has_callbacks_ready_to_invoke(rdp)) { rdp->n_rp_cb_ready++; |
64db4cfff
|
1709 |
return 1; |
7ba5c840e
|
1710 |
} |
64db4cfff
|
1711 1712 |
/* Has RCU gone idle with this CPU needing another grace period? */ |
7ba5c840e
|
1713 1714 |
if (cpu_needs_another_gp(rsp, rdp)) { rdp->n_rp_cpu_needs_gp++; |
64db4cfff
|
1715 |
return 1; |
7ba5c840e
|
1716 |
} |
64db4cfff
|
1717 1718 |
/* Has another RCU grace period completed? */ |
2f51f9884
|
1719 |
if (ACCESS_ONCE(rnp->completed) != rdp->completed) { /* outside lock */ |
7ba5c840e
|
1720 |
rdp->n_rp_gp_completed++; |
64db4cfff
|
1721 |
return 1; |
7ba5c840e
|
1722 |
} |
64db4cfff
|
1723 1724 |
/* Has a new RCU grace period started? */ |
2f51f9884
|
1725 |
if (ACCESS_ONCE(rnp->gpnum) != rdp->gpnum) { /* outside lock */ |
7ba5c840e
|
1726 |
rdp->n_rp_gp_started++; |
64db4cfff
|
1727 |
return 1; |
7ba5c840e
|
1728 |
} |
64db4cfff
|
1729 1730 |
/* Has an RCU GP gone long enough to send resched IPIs &c? */ |
fc2219d49
|
1731 |
if (rcu_gp_in_progress(rsp) && |
20133cfce
|
1732 |
ULONG_CMP_LT(ACCESS_ONCE(rsp->jiffies_force_qs), jiffies)) { |
7ba5c840e
|
1733 |
rdp->n_rp_need_fqs++; |
64db4cfff
|
1734 |
return 1; |
7ba5c840e
|
1735 |
} |
64db4cfff
|
1736 1737 |
/* nothing to do */ |
7ba5c840e
|
1738 |
rdp->n_rp_need_nothing++; |
64db4cfff
|
1739 1740 1741 1742 1743 1744 1745 1746 |
return 0; } /* * Check to see if there is any immediate RCU-related work to be done * by the current CPU, returning 1 if so. This function is part of the * RCU implementation; it is -not- an exported member of the RCU API. */ |
a157229ca
|
1747 |
static int rcu_pending(int cpu) |
64db4cfff
|
1748 |
{ |
d6714c22b
|
1749 |
return __rcu_pending(&rcu_sched_state, &per_cpu(rcu_sched_data, cpu)) || |
f41d911f8
|
1750 1751 |
__rcu_pending(&rcu_bh_state, &per_cpu(rcu_bh_data, cpu)) || rcu_preempt_pending(cpu); |
64db4cfff
|
1752 1753 1754 1755 1756 |
} /* * Check to see if any future RCU-related work will need to be done * by the current CPU, even if none need be done immediately, returning |
8bd93a2c5
|
1757 |
* 1 if so. |
64db4cfff
|
1758 |
*/ |
8bd93a2c5
|
1759 |
static int rcu_needs_cpu_quick_check(int cpu) |
64db4cfff
|
1760 1761 |
{ /* RCU callbacks either ready or pending? */ |
d6714c22b
|
1762 |
return per_cpu(rcu_sched_data, cpu).nxtlist || |
f41d911f8
|
1763 1764 |
per_cpu(rcu_bh_data, cpu).nxtlist || rcu_preempt_needs_cpu(cpu); |
64db4cfff
|
1765 |
} |
d0ec774cb
|
1766 1767 1768 1769 |
static DEFINE_PER_CPU(struct rcu_head, rcu_barrier_head) = {NULL}; static atomic_t rcu_barrier_cpu_count; static DEFINE_MUTEX(rcu_barrier_mutex); static struct completion rcu_barrier_completion; |
d0ec774cb
|
1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 |
static void rcu_barrier_callback(struct rcu_head *notused) { if (atomic_dec_and_test(&rcu_barrier_cpu_count)) complete(&rcu_barrier_completion); } /* * Called with preemption disabled, and from cross-cpu IRQ context. */ static void rcu_barrier_func(void *type) { int cpu = smp_processor_id(); struct rcu_head *head = &per_cpu(rcu_barrier_head, cpu); void (*call_rcu_func)(struct rcu_head *head, void (*func)(struct rcu_head *head)); atomic_inc(&rcu_barrier_cpu_count); call_rcu_func = type; call_rcu_func(head, rcu_barrier_callback); } |
d0ec774cb
|
1791 1792 1793 1794 |
/* * Orchestrate the specified type of RCU barrier, waiting for all * RCU callbacks of the specified type to complete. */ |
e74f4c456
|
1795 1796 |
static void _rcu_barrier(struct rcu_state *rsp, void (*call_rcu_func)(struct rcu_head *head, |
d0ec774cb
|
1797 1798 1799 |
void (*func)(struct rcu_head *head))) { BUG_ON(in_interrupt()); |
e74f4c456
|
1800 |
/* Take mutex to serialize concurrent rcu_barrier() requests. */ |
d0ec774cb
|
1801 1802 1803 1804 1805 1806 1807 1808 1809 |
mutex_lock(&rcu_barrier_mutex); init_completion(&rcu_barrier_completion); /* * Initialize rcu_barrier_cpu_count to 1, then invoke * rcu_barrier_func() on each CPU, so that each CPU also has * incremented rcu_barrier_cpu_count. Only then is it safe to * decrement rcu_barrier_cpu_count -- otherwise the first CPU * might complete its grace period before all of the other CPUs * did their increment, causing this function to return too |
2d999e03b
|
1810 1811 1812 |
* early. Note that on_each_cpu() disables irqs, which prevents * any CPUs from coming online or going offline until each online * CPU has queued its RCU-barrier callback. |
d0ec774cb
|
1813 1814 1815 1816 1817 1818 1819 |
*/ atomic_set(&rcu_barrier_cpu_count, 1); on_each_cpu(rcu_barrier_func, (void *)call_rcu_func, 1); if (atomic_dec_and_test(&rcu_barrier_cpu_count)) complete(&rcu_barrier_completion); wait_for_completion(&rcu_barrier_completion); mutex_unlock(&rcu_barrier_mutex); |
d0ec774cb
|
1820 |
} |
d0ec774cb
|
1821 1822 1823 1824 1825 1826 |
/** * rcu_barrier_bh - Wait until all in-flight call_rcu_bh() callbacks complete. */ void rcu_barrier_bh(void) { |
e74f4c456
|
1827 |
_rcu_barrier(&rcu_bh_state, call_rcu_bh); |
d0ec774cb
|
1828 1829 1830 1831 1832 1833 1834 1835 |
} EXPORT_SYMBOL_GPL(rcu_barrier_bh); /** * rcu_barrier_sched - Wait for in-flight call_rcu_sched() callbacks. */ void rcu_barrier_sched(void) { |
e74f4c456
|
1836 |
_rcu_barrier(&rcu_sched_state, call_rcu_sched); |
d0ec774cb
|
1837 1838 |
} EXPORT_SYMBOL_GPL(rcu_barrier_sched); |
64db4cfff
|
1839 |
/* |
27569620c
|
1840 |
* Do boot-time initialization of a CPU's per-CPU RCU data. |
64db4cfff
|
1841 |
*/ |
27569620c
|
1842 1843 |
static void __init rcu_boot_init_percpu_data(int cpu, struct rcu_state *rsp) |
64db4cfff
|
1844 1845 1846 |
{ unsigned long flags; int i; |
394f99a90
|
1847 |
struct rcu_data *rdp = per_cpu_ptr(rsp->rda, cpu); |
27569620c
|
1848 1849 1850 |
struct rcu_node *rnp = rcu_get_root(rsp); /* Set up local state, ensuring consistent view of global state. */ |
1304afb22
|
1851 |
raw_spin_lock_irqsave(&rnp->lock, flags); |
27569620c
|
1852 1853 1854 1855 1856 1857 1858 1859 1860 |
rdp->grpmask = 1UL << (cpu - rdp->mynode->grplo); rdp->nxtlist = NULL; for (i = 0; i < RCU_NEXT_SIZE; i++) rdp->nxttail[i] = &rdp->nxtlist; rdp->qlen = 0; #ifdef CONFIG_NO_HZ rdp->dynticks = &per_cpu(rcu_dynticks, cpu); #endif /* #ifdef CONFIG_NO_HZ */ rdp->cpu = cpu; |
d4c08f2ac
|
1861 |
rdp->rsp = rsp; |
1304afb22
|
1862 |
raw_spin_unlock_irqrestore(&rnp->lock, flags); |
27569620c
|
1863 1864 1865 1866 1867 1868 1869 |
} /* * Initialize a CPU's per-CPU RCU data. Note that only one online or * offline event can be happening at a given time. Note also that we * can accept some slop in the rsp->completed access due to the fact * that this CPU cannot possibly have any RCU callbacks in flight yet. |
64db4cfff
|
1870 |
*/ |
e4fa4c970
|
1871 |
static void __cpuinit |
6cc68793e
|
1872 |
rcu_init_percpu_data(int cpu, struct rcu_state *rsp, int preemptible) |
64db4cfff
|
1873 1874 |
{ unsigned long flags; |
64db4cfff
|
1875 |
unsigned long mask; |
394f99a90
|
1876 |
struct rcu_data *rdp = per_cpu_ptr(rsp->rda, cpu); |
64db4cfff
|
1877 1878 1879 |
struct rcu_node *rnp = rcu_get_root(rsp); /* Set up local state, ensuring consistent view of global state. */ |
1304afb22
|
1880 |
raw_spin_lock_irqsave(&rnp->lock, flags); |
64db4cfff
|
1881 |
rdp->beenonline = 1; /* We have now been online. */ |
6cc68793e
|
1882 |
rdp->preemptible = preemptible; |
37c72e56f
|
1883 1884 |
rdp->qlen_last_fqs_check = 0; rdp->n_force_qs_snap = rsp->n_force_qs; |
64db4cfff
|
1885 |
rdp->blimit = blimit; |
1304afb22
|
1886 |
raw_spin_unlock(&rnp->lock); /* irqs remain disabled. */ |
64db4cfff
|
1887 1888 1889 1890 1891 1892 1893 |
/* * A new grace period might start here. If so, we won't be part * of it, but that is OK, as we are currently in a quiescent state. */ /* Exclude any attempts to start a new GP on large systems. */ |
1304afb22
|
1894 |
raw_spin_lock(&rsp->onofflock); /* irqs already disabled. */ |
64db4cfff
|
1895 1896 1897 1898 1899 1900 |
/* Add CPU to rcu_node bitmasks. */ rnp = rdp->mynode; mask = rdp->grpmask; do { /* Exclude any attempts to start a new GP on small systems. */ |
1304afb22
|
1901 |
raw_spin_lock(&rnp->lock); /* irqs already disabled. */ |
64db4cfff
|
1902 1903 |
rnp->qsmaskinit |= mask; mask = rnp->grpmask; |
d09b62dfa
|
1904 |
if (rnp == rdp->mynode) { |
06ae115a1
|
1905 1906 1907 1908 1909 1910 |
/* * If there is a grace period in progress, we will * set up to wait for it next time we run the * RCU core code. */ rdp->gpnum = rnp->completed; |
d09b62dfa
|
1911 |
rdp->completed = rnp->completed; |
06ae115a1
|
1912 1913 |
rdp->passed_quiesce = 0; rdp->qs_pending = 0; |
e4cc1f22b
|
1914 |
rdp->passed_quiesce_gpnum = rnp->gpnum - 1; |
d4c08f2ac
|
1915 |
trace_rcu_grace_period(rsp->name, rdp->gpnum, "cpuonl"); |
d09b62dfa
|
1916 |
} |
1304afb22
|
1917 |
raw_spin_unlock(&rnp->lock); /* irqs already disabled. */ |
64db4cfff
|
1918 1919 |
rnp = rnp->parent; } while (rnp != NULL && !(rnp->qsmaskinit & mask)); |
1304afb22
|
1920 |
raw_spin_unlock_irqrestore(&rsp->onofflock, flags); |
64db4cfff
|
1921 |
} |
d72bce0e6
|
1922 |
static void __cpuinit rcu_prepare_cpu(int cpu) |
64db4cfff
|
1923 |
{ |
f41d911f8
|
1924 1925 1926 |
rcu_init_percpu_data(cpu, &rcu_sched_state, 0); rcu_init_percpu_data(cpu, &rcu_bh_state, 0); rcu_preempt_init_percpu_data(cpu); |
64db4cfff
|
1927 1928 1929 |
} /* |
f41d911f8
|
1930 |
* Handle CPU online/offline notification events. |
64db4cfff
|
1931 |
*/ |
9f680ab41
|
1932 1933 |
static int __cpuinit rcu_cpu_notify(struct notifier_block *self, unsigned long action, void *hcpu) |
64db4cfff
|
1934 1935 |
{ long cpu = (long)hcpu; |
27f4d2805
|
1936 |
struct rcu_data *rdp = per_cpu_ptr(rcu_state->rda, cpu); |
a26ac2455
|
1937 |
struct rcu_node *rnp = rdp->mynode; |
64db4cfff
|
1938 |
|
300df91ca
|
1939 |
trace_rcu_utilization("Start CPU hotplug"); |
64db4cfff
|
1940 1941 1942 |
switch (action) { case CPU_UP_PREPARE: case CPU_UP_PREPARE_FROZEN: |
d72bce0e6
|
1943 1944 |
rcu_prepare_cpu(cpu); rcu_prepare_kthreads(cpu); |
a26ac2455
|
1945 1946 |
break; case CPU_ONLINE: |
0f962a5e7
|
1947 1948 |
case CPU_DOWN_FAILED: rcu_node_kthread_setaffinity(rnp, -1); |
e3995a25f
|
1949 |
rcu_cpu_kthread_setrt(cpu, 1); |
0f962a5e7
|
1950 1951 1952 |
break; case CPU_DOWN_PREPARE: rcu_node_kthread_setaffinity(rnp, cpu); |
e3995a25f
|
1953 |
rcu_cpu_kthread_setrt(cpu, 0); |
64db4cfff
|
1954 |
break; |
d0ec774cb
|
1955 1956 1957 |
case CPU_DYING: case CPU_DYING_FROZEN: /* |
2d999e03b
|
1958 1959 1960 |
* The whole machine is "stopped" except this CPU, so we can * touch any data without introducing corruption. We send the * dying CPU's callbacks to an arbitrarily chosen online CPU. |
d0ec774cb
|
1961 |
*/ |
29494be71
|
1962 1963 1964 |
rcu_send_cbs_to_online(&rcu_bh_state); rcu_send_cbs_to_online(&rcu_sched_state); rcu_preempt_send_cbs_to_online(); |
d0ec774cb
|
1965 |
break; |
64db4cfff
|
1966 1967 1968 1969 1970 1971 1972 1973 1974 |
case CPU_DEAD: case CPU_DEAD_FROZEN: case CPU_UP_CANCELED: case CPU_UP_CANCELED_FROZEN: rcu_offline_cpu(cpu); break; default: break; } |
300df91ca
|
1975 |
trace_rcu_utilization("End CPU hotplug"); |
64db4cfff
|
1976 1977 1978 1979 |
return NOTIFY_OK; } /* |
bbad93798
|
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 |
* This function is invoked towards the end of the scheduler's initialization * process. Before this is called, the idle task might contain * RCU read-side critical sections (during which time, this idle * task is booting the system). After this function is called, the * idle tasks are prohibited from containing RCU read-side critical * sections. This function also enables RCU lockdep checking. */ void rcu_scheduler_starting(void) { WARN_ON(num_online_cpus() != 1); WARN_ON(nr_context_switches() > 0); rcu_scheduler_active = 1; } /* |
64db4cfff
|
1995 1996 1997 1998 1999 2000 2001 |
* Compute the per-level fanout, either using the exact fanout specified * or balancing the tree, depending on CONFIG_RCU_FANOUT_EXACT. */ #ifdef CONFIG_RCU_FANOUT_EXACT static void __init rcu_init_levelspread(struct rcu_state *rsp) { int i; |
0209f6490
|
2002 |
for (i = NUM_RCU_LVLS - 1; i > 0; i--) |
64db4cfff
|
2003 |
rsp->levelspread[i] = CONFIG_RCU_FANOUT; |
0209f6490
|
2004 |
rsp->levelspread[0] = RCU_FANOUT_LEAF; |
64db4cfff
|
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 |
} #else /* #ifdef CONFIG_RCU_FANOUT_EXACT */ static void __init rcu_init_levelspread(struct rcu_state *rsp) { int ccur; int cprv; int i; cprv = NR_CPUS; for (i = NUM_RCU_LVLS - 1; i >= 0; i--) { ccur = rsp->levelcnt[i]; rsp->levelspread[i] = (cprv + ccur - 1) / ccur; cprv = ccur; } } #endif /* #else #ifdef CONFIG_RCU_FANOUT_EXACT */ /* * Helper function for rcu_init() that initializes one rcu_state structure. */ |
394f99a90
|
2025 2026 |
static void __init rcu_init_one(struct rcu_state *rsp, struct rcu_data __percpu *rda) |
64db4cfff
|
2027 |
{ |
b6407e863
|
2028 2029 2030 2031 |
static char *buf[] = { "rcu_node_level_0", "rcu_node_level_1", "rcu_node_level_2", "rcu_node_level_3" }; /* Match MAX_RCU_LVLS */ |
64db4cfff
|
2032 2033 2034 2035 |
int cpustride = 1; int i; int j; struct rcu_node *rnp; |
b6407e863
|
2036 |
BUILD_BUG_ON(MAX_RCU_LVLS > ARRAY_SIZE(buf)); /* Fix buf[] init! */ |
64db4cfff
|
2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 |
/* Initialize the level-tracking arrays. */ for (i = 1; i < NUM_RCU_LVLS; i++) rsp->level[i] = rsp->level[i - 1] + rsp->levelcnt[i - 1]; rcu_init_levelspread(rsp); /* Initialize the elements themselves, starting from the leaves. */ for (i = NUM_RCU_LVLS - 1; i >= 0; i--) { cpustride *= rsp->levelspread[i]; rnp = rsp->level[i]; for (j = 0; j < rsp->levelcnt[i]; j++, rnp++) { |
1304afb22
|
2049 |
raw_spin_lock_init(&rnp->lock); |
b6407e863
|
2050 2051 |
lockdep_set_class_and_name(&rnp->lock, &rcu_node_class[i], buf[i]); |
f41d911f8
|
2052 |
rnp->gpnum = 0; |
64db4cfff
|
2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 |
rnp->qsmask = 0; rnp->qsmaskinit = 0; rnp->grplo = j * cpustride; rnp->grphi = (j + 1) * cpustride - 1; if (rnp->grphi >= NR_CPUS) rnp->grphi = NR_CPUS - 1; if (i == 0) { rnp->grpnum = 0; rnp->grpmask = 0; rnp->parent = NULL; } else { rnp->grpnum = j % rsp->levelspread[i - 1]; rnp->grpmask = 1UL << rnp->grpnum; rnp->parent = rsp->level[i - 1] + j / rsp->levelspread[i - 1]; } rnp->level = i; |
12f5f524c
|
2070 |
INIT_LIST_HEAD(&rnp->blkd_tasks); |
64db4cfff
|
2071 2072 |
} } |
0c34029ab
|
2073 |
|
394f99a90
|
2074 |
rsp->rda = rda; |
0c34029ab
|
2075 2076 |
rnp = rsp->level[NUM_RCU_LVLS - 1]; for_each_possible_cpu(i) { |
4a90a0681
|
2077 |
while (i > rnp->grphi) |
0c34029ab
|
2078 |
rnp++; |
394f99a90
|
2079 |
per_cpu_ptr(rsp->rda, i)->mynode = rnp; |
0c34029ab
|
2080 2081 |
rcu_boot_init_percpu_data(i, rsp); } |
64db4cfff
|
2082 |
} |
9f680ab41
|
2083 |
void __init rcu_init(void) |
64db4cfff
|
2084 |
{ |
017c42613
|
2085 |
int cpu; |
9f680ab41
|
2086 |
|
f41d911f8
|
2087 |
rcu_bootup_announce(); |
394f99a90
|
2088 2089 |
rcu_init_one(&rcu_sched_state, &rcu_sched_data); rcu_init_one(&rcu_bh_state, &rcu_bh_data); |
f41d911f8
|
2090 |
__rcu_init_preempt(); |
09223371d
|
2091 |
open_softirq(RCU_SOFTIRQ, rcu_process_callbacks); |
9f680ab41
|
2092 2093 2094 2095 2096 2097 2098 |
/* * We don't need protection against CPU-hotplug here because * this is called early in boot, before either interrupts * or the scheduler are operational. */ cpu_notifier(rcu_cpu_notify, 0); |
017c42613
|
2099 2100 |
for_each_online_cpu(cpu) rcu_cpu_notify(NULL, CPU_UP_PREPARE, (void *)(long)cpu); |
c68de2097
|
2101 |
check_cpu_stall_init(); |
64db4cfff
|
2102 |
} |
1eba8f843
|
2103 |
#include "rcutree_plugin.h" |