1 /*- 2 * Copyright (c) 2001 Jake Burkholder <jake@FreeBSD.org> 3 * All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 14 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 15 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 16 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 17 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 18 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 19 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 20 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 21 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 22 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 23 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 24 * SUCH DAMAGE. 25 */ 26 27 28 #include <sys/cdefs.h> 29 __FBSDID("$FreeBSD$"); 30 31 #include "opt_sched.h" 32 33 #include <sys/param.h> 34 #include <sys/systm.h> 35 #include <sys/kdb.h> 36 #include <sys/kernel.h> 37 #include <sys/ktr.h> 38 #include <sys/lock.h> 39 #include <sys/mutex.h> 40 #include <sys/proc.h> 41 #include <sys/queue.h> 42 #include <sys/sched.h> 43 #include <sys/smp.h> 44 #include <sys/sysctl.h> 45 46 #include <machine/cpu.h> 47 48 /* Uncomment this to enable logging of critical_enter/exit. */ 49 #if 0 50 #define KTR_CRITICAL KTR_SCHED 51 #else 52 #define KTR_CRITICAL 0 53 #endif 54 55 #ifdef FULL_PREEMPTION 56 #ifndef PREEMPTION 57 #error "The FULL_PREEMPTION option requires the PREEMPTION option" 58 #endif 59 #endif 60 61 CTASSERT((RQB_BPW * RQB_LEN) == RQ_NQS); 62 63 /* 64 * kern.sched.preemption allows user space to determine if preemption support 65 * is compiled in or not. It is not currently a boot or runtime flag that 66 * can be changed. 67 */ 68 #ifdef PREEMPTION 69 static int kern_sched_preemption = 1; 70 #else 71 static int kern_sched_preemption = 0; 72 #endif 73 SYSCTL_INT(_kern_sched, OID_AUTO, preemption, CTLFLAG_RD, 74 &kern_sched_preemption, 0, "Kernel preemption enabled"); 75 76 /* 77 * Support for scheduler stats exported via kern.sched.stats. All stats may 78 * be reset with kern.sched.stats.reset = 1. Stats may be defined elsewhere 79 * with SCHED_STAT_DEFINE(). 80 */ 81 #ifdef SCHED_STATS 82 SYSCTL_NODE(_kern_sched, OID_AUTO, stats, CTLFLAG_RW, 0, "switch stats"); 83 84 /* Switch reasons from mi_switch(). */ 85 DPCPU_DEFINE(long, sched_switch_stats[SWT_COUNT]); 86 SCHED_STAT_DEFINE_VAR(uncategorized, 87 &DPCPU_NAME(sched_switch_stats[SWT_NONE]), ""); 88 SCHED_STAT_DEFINE_VAR(preempt, 89 &DPCPU_NAME(sched_switch_stats[SWT_PREEMPT]), ""); 90 SCHED_STAT_DEFINE_VAR(owepreempt, 91 &DPCPU_NAME(sched_switch_stats[SWT_OWEPREEMPT]), ""); 92 SCHED_STAT_DEFINE_VAR(turnstile, 93 &DPCPU_NAME(sched_switch_stats[SWT_TURNSTILE]), ""); 94 SCHED_STAT_DEFINE_VAR(sleepq, 95 &DPCPU_NAME(sched_switch_stats[SWT_SLEEPQ]), ""); 96 SCHED_STAT_DEFINE_VAR(sleepqtimo, 97 &DPCPU_NAME(sched_switch_stats[SWT_SLEEPQTIMO]), ""); 98 SCHED_STAT_DEFINE_VAR(relinquish, 99 &DPCPU_NAME(sched_switch_stats[SWT_RELINQUISH]), ""); 100 SCHED_STAT_DEFINE_VAR(needresched, 101 &DPCPU_NAME(sched_switch_stats[SWT_NEEDRESCHED]), ""); 102 SCHED_STAT_DEFINE_VAR(idle, 103 &DPCPU_NAME(sched_switch_stats[SWT_IDLE]), ""); 104 SCHED_STAT_DEFINE_VAR(iwait, 105 &DPCPU_NAME(sched_switch_stats[SWT_IWAIT]), ""); 106 SCHED_STAT_DEFINE_VAR(suspend, 107 &DPCPU_NAME(sched_switch_stats[SWT_SUSPEND]), ""); 108 SCHED_STAT_DEFINE_VAR(remotepreempt, 109 &DPCPU_NAME(sched_switch_stats[SWT_REMOTEPREEMPT]), ""); 110 SCHED_STAT_DEFINE_VAR(remotewakeidle, 111 &DPCPU_NAME(sched_switch_stats[SWT_REMOTEWAKEIDLE]), ""); 112 113 static int 114 sysctl_stats_reset(SYSCTL_HANDLER_ARGS) 115 { 116 struct sysctl_oid *p; 117 uintptr_t counter; 118 int error; 119 int val; 120 int i; 121 122 val = 0; 123 error = sysctl_handle_int(oidp, &val, 0, req); 124 if (error != 0 || req->newptr == NULL) 125 return (error); 126 if (val == 0) 127 return (0); 128 /* 129 * Traverse the list of children of _kern_sched_stats and reset each 130 * to 0. Skip the reset entry. 131 */ 132 SLIST_FOREACH(p, oidp->oid_parent, oid_link) { 133 if (p == oidp || p->oid_arg1 == NULL) 134 continue; 135 counter = (uintptr_t)p->oid_arg1; 136 for (i = 0; i <= mp_maxid; i++) { 137 if (CPU_ABSENT(i)) 138 continue; 139 *(long *)(dpcpu_off[i] + counter) = 0; 140 } 141 } 142 return (0); 143 } 144 145 SYSCTL_PROC(_kern_sched_stats, OID_AUTO, reset, CTLTYPE_INT | CTLFLAG_WR, NULL, 146 0, sysctl_stats_reset, "I", "Reset scheduler statistics"); 147 #endif 148 149 /************************************************************************ 150 * Functions that manipulate runnability from a thread perspective. * 151 ************************************************************************/ 152 /* 153 * Select the thread that will be run next. 154 */ 155 struct thread * 156 choosethread(void) 157 { 158 struct thread *td; 159 160 retry: 161 td = sched_choose(); 162 163 /* 164 * If we are in panic, only allow system threads, 165 * plus the one we are running in, to be run. 166 */ 167 if (panicstr && ((td->td_proc->p_flag & P_SYSTEM) == 0 && 168 (td->td_flags & TDF_INPANIC) == 0)) { 169 /* note that it is no longer on the run queue */ 170 TD_SET_CAN_RUN(td); 171 goto retry; 172 } 173 174 TD_SET_RUNNING(td); 175 return (td); 176 } 177 178 /* 179 * Kernel thread preemption implementation. Critical sections mark 180 * regions of code in which preemptions are not allowed. 181 */ 182 void 183 critical_enter(void) 184 { 185 struct thread *td; 186 187 td = curthread; 188 td->td_critnest++; 189 CTR4(KTR_CRITICAL, "critical_enter by thread %p (%ld, %s) to %d", td, 190 (long)td->td_proc->p_pid, td->td_name, td->td_critnest); 191 } 192 193 void 194 critical_exit(void) 195 { 196 struct thread *td; 197 int flags; 198 199 td = curthread; 200 KASSERT(td->td_critnest != 0, 201 ("critical_exit: td_critnest == 0")); 202 203 if (td->td_critnest == 1) { 204 td->td_critnest = 0; 205 if (td->td_owepreempt) { 206 td->td_critnest = 1; 207 thread_lock(td); 208 td->td_critnest--; 209 flags = SW_INVOL | SW_PREEMPT; 210 if (TD_IS_IDLETHREAD(td)) 211 flags |= SWT_IDLE; 212 else 213 flags |= SWT_OWEPREEMPT; 214 mi_switch(flags, NULL); 215 thread_unlock(td); 216 } 217 } else 218 td->td_critnest--; 219 220 CTR4(KTR_CRITICAL, "critical_exit by thread %p (%ld, %s) to %d", td, 221 (long)td->td_proc->p_pid, td->td_name, td->td_critnest); 222 } 223 224 /************************************************************************ 225 * SYSTEM RUN QUEUE manipulations and tests * 226 ************************************************************************/ 227 /* 228 * Initialize a run structure. 229 */ 230 void 231 runq_init(struct runq *rq) 232 { 233 int i; 234 235 bzero(rq, sizeof *rq); 236 for (i = 0; i < RQ_NQS; i++) 237 TAILQ_INIT(&rq->rq_queues[i]); 238 } 239 240 /* 241 * Clear the status bit of the queue corresponding to priority level pri, 242 * indicating that it is empty. 243 */ 244 static __inline void 245 runq_clrbit(struct runq *rq, int pri) 246 { 247 struct rqbits *rqb; 248 249 rqb = &rq->rq_status; 250 CTR4(KTR_RUNQ, "runq_clrbit: bits=%#x %#x bit=%#x word=%d", 251 rqb->rqb_bits[RQB_WORD(pri)], 252 rqb->rqb_bits[RQB_WORD(pri)] & ~RQB_BIT(pri), 253 RQB_BIT(pri), RQB_WORD(pri)); 254 rqb->rqb_bits[RQB_WORD(pri)] &= ~RQB_BIT(pri); 255 } 256 257 /* 258 * Find the index of the first non-empty run queue. This is done by 259 * scanning the status bits, a set bit indicates a non-empty queue. 260 */ 261 static __inline int 262 runq_findbit(struct runq *rq) 263 { 264 struct rqbits *rqb; 265 int pri; 266 int i; 267 268 rqb = &rq->rq_status; 269 for (i = 0; i < RQB_LEN; i++) 270 if (rqb->rqb_bits[i]) { 271 pri = RQB_FFS(rqb->rqb_bits[i]) + (i << RQB_L2BPW); 272 CTR3(KTR_RUNQ, "runq_findbit: bits=%#x i=%d pri=%d", 273 rqb->rqb_bits[i], i, pri); 274 return (pri); 275 } 276 277 return (-1); 278 } 279 280 static __inline int 281 runq_findbit_from(struct runq *rq, u_char pri) 282 { 283 struct rqbits *rqb; 284 rqb_word_t mask; 285 int i; 286 287 /* 288 * Set the mask for the first word so we ignore priorities before 'pri'. 289 */ 290 mask = (rqb_word_t)-1 << (pri & (RQB_BPW - 1)); 291 rqb = &rq->rq_status; 292 again: 293 for (i = RQB_WORD(pri); i < RQB_LEN; mask = -1, i++) { 294 mask = rqb->rqb_bits[i] & mask; 295 if (mask == 0) 296 continue; 297 pri = RQB_FFS(mask) + (i << RQB_L2BPW); 298 CTR3(KTR_RUNQ, "runq_findbit_from: bits=%#x i=%d pri=%d", 299 mask, i, pri); 300 return (pri); 301 } 302 if (pri == 0) 303 return (-1); 304 /* 305 * Wrap back around to the beginning of the list just once so we 306 * scan the whole thing. 307 */ 308 pri = 0; 309 goto again; 310 } 311 312 /* 313 * Set the status bit of the queue corresponding to priority level pri, 314 * indicating that it is non-empty. 315 */ 316 static __inline void 317 runq_setbit(struct runq *rq, int pri) 318 { 319 struct rqbits *rqb; 320 321 rqb = &rq->rq_status; 322 CTR4(KTR_RUNQ, "runq_setbit: bits=%#x %#x bit=%#x word=%d", 323 rqb->rqb_bits[RQB_WORD(pri)], 324 rqb->rqb_bits[RQB_WORD(pri)] | RQB_BIT(pri), 325 RQB_BIT(pri), RQB_WORD(pri)); 326 rqb->rqb_bits[RQB_WORD(pri)] |= RQB_BIT(pri); 327 } 328 329 /* 330 * Add the thread to the queue specified by its priority, and set the 331 * corresponding status bit. 332 */ 333 void 334 runq_add(struct runq *rq, struct thread *td, int flags) 335 { 336 struct rqhead *rqh; 337 int pri; 338 339 pri = td->td_priority / RQ_PPQ; 340 td->td_rqindex = pri; 341 runq_setbit(rq, pri); 342 rqh = &rq->rq_queues[pri]; 343 CTR4(KTR_RUNQ, "runq_add: td=%p pri=%d %d rqh=%p", 344 td, td->td_priority, pri, rqh); 345 if (flags & SRQ_PREEMPTED) { 346 TAILQ_INSERT_HEAD(rqh, td, td_runq); 347 } else { 348 TAILQ_INSERT_TAIL(rqh, td, td_runq); 349 } 350 } 351 352 void 353 runq_add_pri(struct runq *rq, struct thread *td, u_char pri, int flags) 354 { 355 struct rqhead *rqh; 356 357 KASSERT(pri < RQ_NQS, ("runq_add_pri: %d out of range", pri)); 358 td->td_rqindex = pri; 359 runq_setbit(rq, pri); 360 rqh = &rq->rq_queues[pri]; 361 CTR4(KTR_RUNQ, "runq_add_pri: td=%p pri=%d idx=%d rqh=%p", 362 td, td->td_priority, pri, rqh); 363 if (flags & SRQ_PREEMPTED) { 364 TAILQ_INSERT_HEAD(rqh, td, td_runq); 365 } else { 366 TAILQ_INSERT_TAIL(rqh, td, td_runq); 367 } 368 } 369 /* 370 * Return true if there are runnable processes of any priority on the run 371 * queue, false otherwise. Has no side effects, does not modify the run 372 * queue structure. 373 */ 374 int 375 runq_check(struct runq *rq) 376 { 377 struct rqbits *rqb; 378 int i; 379 380 rqb = &rq->rq_status; 381 for (i = 0; i < RQB_LEN; i++) 382 if (rqb->rqb_bits[i]) { 383 CTR2(KTR_RUNQ, "runq_check: bits=%#x i=%d", 384 rqb->rqb_bits[i], i); 385 return (1); 386 } 387 CTR0(KTR_RUNQ, "runq_check: empty"); 388 389 return (0); 390 } 391 392 /* 393 * Find the highest priority process on the run queue. 394 */ 395 struct thread * 396 runq_choose_fuzz(struct runq *rq, int fuzz) 397 { 398 struct rqhead *rqh; 399 struct thread *td; 400 int pri; 401 402 while ((pri = runq_findbit(rq)) != -1) { 403 rqh = &rq->rq_queues[pri]; 404 /* fuzz == 1 is normal.. 0 or less are ignored */ 405 if (fuzz > 1) { 406 /* 407 * In the first couple of entries, check if 408 * there is one for our CPU as a preference. 409 */ 410 int count = fuzz; 411 int cpu = PCPU_GET(cpuid); 412 struct thread *td2; 413 td2 = td = TAILQ_FIRST(rqh); 414 415 while (count-- && td2) { 416 if (td2->td_lastcpu == cpu) { 417 td = td2; 418 break; 419 } 420 td2 = TAILQ_NEXT(td2, td_runq); 421 } 422 } else 423 td = TAILQ_FIRST(rqh); 424 KASSERT(td != NULL, ("runq_choose_fuzz: no proc on busy queue")); 425 CTR3(KTR_RUNQ, 426 "runq_choose_fuzz: pri=%d thread=%p rqh=%p", pri, td, rqh); 427 return (td); 428 } 429 CTR1(KTR_RUNQ, "runq_choose_fuzz: idleproc pri=%d", pri); 430 431 return (NULL); 432 } 433 434 /* 435 * Find the highest priority process on the run queue. 436 */ 437 struct thread * 438 runq_choose(struct runq *rq) 439 { 440 struct rqhead *rqh; 441 struct thread *td; 442 int pri; 443 444 while ((pri = runq_findbit(rq)) != -1) { 445 rqh = &rq->rq_queues[pri]; 446 td = TAILQ_FIRST(rqh); 447 KASSERT(td != NULL, ("runq_choose: no thread on busy queue")); 448 CTR3(KTR_RUNQ, 449 "runq_choose: pri=%d thread=%p rqh=%p", pri, td, rqh); 450 return (td); 451 } 452 CTR1(KTR_RUNQ, "runq_choose: idlethread pri=%d", pri); 453 454 return (NULL); 455 } 456 457 struct thread * 458 runq_choose_from(struct runq *rq, u_char idx) 459 { 460 struct rqhead *rqh; 461 struct thread *td; 462 int pri; 463 464 if ((pri = runq_findbit_from(rq, idx)) != -1) { 465 rqh = &rq->rq_queues[pri]; 466 td = TAILQ_FIRST(rqh); 467 KASSERT(td != NULL, ("runq_choose: no thread on busy queue")); 468 CTR4(KTR_RUNQ, 469 "runq_choose_from: pri=%d thread=%p idx=%d rqh=%p", 470 pri, td, td->td_rqindex, rqh); 471 return (td); 472 } 473 CTR1(KTR_RUNQ, "runq_choose_from: idlethread pri=%d", pri); 474 475 return (NULL); 476 } 477 /* 478 * Remove the thread from the queue specified by its priority, and clear the 479 * corresponding status bit if the queue becomes empty. 480 * Caller must set state afterwards. 481 */ 482 void 483 runq_remove(struct runq *rq, struct thread *td) 484 { 485 486 runq_remove_idx(rq, td, NULL); 487 } 488 489 void 490 runq_remove_idx(struct runq *rq, struct thread *td, u_char *idx) 491 { 492 struct rqhead *rqh; 493 u_char pri; 494 495 KASSERT(td->td_flags & TDF_INMEM, 496 ("runq_remove_idx: thread swapped out")); 497 pri = td->td_rqindex; 498 KASSERT(pri < RQ_NQS, ("runq_remove_idx: Invalid index %d\n", pri)); 499 rqh = &rq->rq_queues[pri]; 500 CTR4(KTR_RUNQ, "runq_remove_idx: td=%p, pri=%d %d rqh=%p", 501 td, td->td_priority, pri, rqh); 502 TAILQ_REMOVE(rqh, td, td_runq); 503 if (TAILQ_EMPTY(rqh)) { 504 CTR0(KTR_RUNQ, "runq_remove_idx: empty"); 505 runq_clrbit(rq, pri); 506 if (idx != NULL && *idx == pri) 507 *idx = (pri + 1) % RQ_NQS; 508 } 509 } 510