sched_4bsd.c revision 1.17
1 /* $NetBSD: sched_4bsd.c,v 1.17 2008/04/14 09:39:31 yamt Exp $ */ 2 3 /*- 4 * Copyright (c) 1999, 2000, 2004, 2006, 2007, 2008 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * This code is derived from software contributed to The NetBSD Foundation 8 * by Jason R. Thorpe of the Numerical Aerospace Simulation Facility, 9 * NASA Ames Research Center, by Charles M. Hannum, Andrew Doran, and 10 * Daniel Sieger. 11 * 12 * Redistribution and use in source and binary forms, with or without 13 * modification, are permitted provided that the following conditions 14 * are met: 15 * 1. Redistributions of source code must retain the above copyright 16 * notice, this list of conditions and the following disclaimer. 17 * 2. Redistributions in binary form must reproduce the above copyright 18 * notice, this list of conditions and the following disclaimer in the 19 * documentation and/or other materials provided with the distribution. 20 * 3. All advertising materials mentioning features or use of this software 21 * must display the following acknowledgement: 22 * This product includes software developed by the NetBSD 23 * Foundation, Inc. and its contributors. 24 * 4. Neither the name of The NetBSD Foundation nor the names of its 25 * contributors may be used to endorse or promote products derived 26 * from this software without specific prior written permission. 27 * 28 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 29 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 30 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 31 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 32 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 33 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 34 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 35 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 36 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 37 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 38 * POSSIBILITY OF SUCH DAMAGE. 39 */ 40 41 /*- 42 * Copyright (c) 1982, 1986, 1990, 1991, 1993 43 * The Regents of the University of California. All rights reserved. 44 * (c) UNIX System Laboratories, Inc. 45 * All or some portions of this file are derived from material licensed 46 * to the University of California by American Telephone and Telegraph 47 * Co. or Unix System Laboratories, Inc. and are reproduced herein with 48 * the permission of UNIX System Laboratories, Inc. 49 * 50 * Redistribution and use in source and binary forms, with or without 51 * modification, are permitted provided that the following conditions 52 * are met: 53 * 1. Redistributions of source code must retain the above copyright 54 * notice, this list of conditions and the following disclaimer. 55 * 2. Redistributions in binary form must reproduce the above copyright 56 * notice, this list of conditions and the following disclaimer in the 57 * documentation and/or other materials provided with the distribution. 58 * 3. Neither the name of the University nor the names of its contributors 59 * may be used to endorse or promote products derived from this software 60 * without specific prior written permission. 61 * 62 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 63 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 64 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 65 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 66 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 67 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 68 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 69 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 70 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 71 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 72 * SUCH DAMAGE. 73 * 74 * @(#)kern_synch.c 8.9 (Berkeley) 5/19/95 75 */ 76 77 #include <sys/cdefs.h> 78 __KERNEL_RCSID(0, "$NetBSD: sched_4bsd.c,v 1.17 2008/04/14 09:39:31 yamt Exp $"); 79 80 #include "opt_ddb.h" 81 #include "opt_lockdebug.h" 82 #include "opt_perfctrs.h" 83 84 #define __MUTEX_PRIVATE 85 86 #include <sys/param.h> 87 #include <sys/systm.h> 88 #include <sys/callout.h> 89 #include <sys/cpu.h> 90 #include <sys/proc.h> 91 #include <sys/kernel.h> 92 #include <sys/signalvar.h> 93 #include <sys/resourcevar.h> 94 #include <sys/sched.h> 95 #include <sys/sysctl.h> 96 #include <sys/kauth.h> 97 #include <sys/lockdebug.h> 98 #include <sys/kmem.h> 99 #include <sys/intr.h> 100 101 #include <uvm/uvm_extern.h> 102 103 static void updatepri(struct lwp *); 104 static void resetpriority(struct lwp *); 105 106 extern unsigned int sched_pstats_ticks; /* defined in kern_synch.c */ 107 108 /* Number of hardclock ticks per sched_tick() */ 109 static int rrticks; 110 111 /* 112 * Force switch among equal priority processes every 100ms. 113 * Called from hardclock every hz/10 == rrticks hardclock ticks. 114 * 115 * There's no need to lock anywhere in this routine, as it's 116 * CPU-local and runs at IPL_SCHED (called from clock interrupt). 117 */ 118 /* ARGSUSED */ 119 void 120 sched_tick(struct cpu_info *ci) 121 { 122 struct schedstate_percpu *spc = &ci->ci_schedstate; 123 124 spc->spc_ticks = rrticks; 125 126 if (CURCPU_IDLE_P()) { 127 cpu_need_resched(ci, 0); 128 return; 129 } 130 131 if (spc->spc_flags & SPCF_SEENRR) { 132 /* 133 * The process has already been through a roundrobin 134 * without switching and may be hogging the CPU. 135 * Indicate that the process should yield. 136 */ 137 spc->spc_flags |= SPCF_SHOULDYIELD; 138 cpu_need_resched(ci, 0); 139 } else 140 spc->spc_flags |= SPCF_SEENRR; 141 } 142 143 /* 144 * Why PRIO_MAX - 2? From setpriority(2): 145 * 146 * prio is a value in the range -20 to 20. The default priority is 147 * 0; lower priorities cause more favorable scheduling. A value of 148 * 19 or 20 will schedule a process only when nothing at priority <= 149 * 0 is runnable. 150 * 151 * This gives estcpu influence over 18 priority levels, and leaves nice 152 * with 40 levels. One way to think about it is that nice has 20 levels 153 * either side of estcpu's 18. 154 */ 155 #define ESTCPU_SHIFT 11 156 #define ESTCPU_MAX ((PRIO_MAX - 2) << ESTCPU_SHIFT) 157 #define ESTCPU_ACCUM (1 << (ESTCPU_SHIFT - 1)) 158 #define ESTCPULIM(e) min((e), ESTCPU_MAX) 159 160 /* 161 * Constants for digital decay and forget: 162 * 90% of (l_estcpu) usage in 5 * loadav time 163 * 95% of (l_pctcpu) usage in 60 seconds (load insensitive) 164 * Note that, as ps(1) mentions, this can let percentages 165 * total over 100% (I've seen 137.9% for 3 processes). 166 * 167 * Note that hardclock updates l_estcpu and l_cpticks independently. 168 * 169 * We wish to decay away 90% of l_estcpu in (5 * loadavg) seconds. 170 * That is, the system wants to compute a value of decay such 171 * that the following for loop: 172 * for (i = 0; i < (5 * loadavg); i++) 173 * l_estcpu *= decay; 174 * will compute 175 * l_estcpu *= 0.1; 176 * for all values of loadavg: 177 * 178 * Mathematically this loop can be expressed by saying: 179 * decay ** (5 * loadavg) ~= .1 180 * 181 * The system computes decay as: 182 * decay = (2 * loadavg) / (2 * loadavg + 1) 183 * 184 * We wish to prove that the system's computation of decay 185 * will always fulfill the equation: 186 * decay ** (5 * loadavg) ~= .1 187 * 188 * If we compute b as: 189 * b = 2 * loadavg 190 * then 191 * decay = b / (b + 1) 192 * 193 * We now need to prove two things: 194 * 1) Given factor ** (5 * loadavg) ~= .1, prove factor == b/(b+1) 195 * 2) Given b/(b+1) ** power ~= .1, prove power == (5 * loadavg) 196 * 197 * Facts: 198 * For x close to zero, exp(x) =~ 1 + x, since 199 * exp(x) = 0! + x**1/1! + x**2/2! + ... . 200 * therefore exp(-1/b) =~ 1 - (1/b) = (b-1)/b. 201 * For x close to zero, ln(1+x) =~ x, since 202 * ln(1+x) = x - x**2/2 + x**3/3 - ... -1 < x < 1 203 * therefore ln(b/(b+1)) = ln(1 - 1/(b+1)) =~ -1/(b+1). 204 * ln(.1) =~ -2.30 205 * 206 * Proof of (1): 207 * Solve (factor)**(power) =~ .1 given power (5*loadav): 208 * solving for factor, 209 * ln(factor) =~ (-2.30/5*loadav), or 210 * factor =~ exp(-1/((5/2.30)*loadav)) =~ exp(-1/(2*loadav)) = 211 * exp(-1/b) =~ (b-1)/b =~ b/(b+1). QED 212 * 213 * Proof of (2): 214 * Solve (factor)**(power) =~ .1 given factor == (b/(b+1)): 215 * solving for power, 216 * power*ln(b/(b+1)) =~ -2.30, or 217 * power =~ 2.3 * (b + 1) = 4.6*loadav + 2.3 =~ 5*loadav. QED 218 * 219 * Actual power values for the implemented algorithm are as follows: 220 * loadav: 1 2 3 4 221 * power: 5.68 10.32 14.94 19.55 222 */ 223 224 /* calculations for digital decay to forget 90% of usage in 5*loadav sec */ 225 #define loadfactor(loadav) (2 * (loadav)) 226 227 static fixpt_t 228 decay_cpu(fixpt_t loadfac, fixpt_t estcpu) 229 { 230 231 if (estcpu == 0) { 232 return 0; 233 } 234 235 #if !defined(_LP64) 236 /* avoid 64bit arithmetics. */ 237 #define FIXPT_MAX ((fixpt_t)((UINTMAX_C(1) << sizeof(fixpt_t) * CHAR_BIT) - 1)) 238 if (__predict_true(loadfac <= FIXPT_MAX / ESTCPU_MAX)) { 239 return estcpu * loadfac / (loadfac + FSCALE); 240 } 241 #endif /* !defined(_LP64) */ 242 243 return (uint64_t)estcpu * loadfac / (loadfac + FSCALE); 244 } 245 246 /* 247 * For all load averages >= 1 and max l_estcpu of (255 << ESTCPU_SHIFT), 248 * sleeping for at least seven times the loadfactor will decay l_estcpu to 249 * less than (1 << ESTCPU_SHIFT). 250 * 251 * note that our ESTCPU_MAX is actually much smaller than (255 << ESTCPU_SHIFT). 252 */ 253 static fixpt_t 254 decay_cpu_batch(fixpt_t loadfac, fixpt_t estcpu, unsigned int n) 255 { 256 257 if ((n << FSHIFT) >= 7 * loadfac) { 258 return 0; 259 } 260 261 while (estcpu != 0 && n > 1) { 262 estcpu = decay_cpu(loadfac, estcpu); 263 n--; 264 } 265 266 return estcpu; 267 } 268 269 /* 270 * sched_pstats_hook: 271 * 272 * Periodically called from sched_pstats(); used to recalculate priorities. 273 */ 274 void 275 sched_pstats_hook(struct lwp *l) 276 { 277 fixpt_t loadfac; 278 int sleeptm; 279 280 /* 281 * If the LWP has slept an entire second, stop recalculating 282 * its priority until it wakes up. 283 */ 284 if (l->l_stat == LSSLEEP || l->l_stat == LSSTOP || 285 l->l_stat == LSSUSPENDED) { 286 l->l_slptime++; 287 sleeptm = 1; 288 } else { 289 sleeptm = 0x7fffffff; 290 } 291 292 if (l->l_slptime <= sleeptm) { 293 loadfac = 2 * (averunnable.ldavg[0]); 294 l->l_estcpu = decay_cpu(loadfac, l->l_estcpu); 295 resetpriority(l); 296 } 297 } 298 299 /* 300 * Recalculate the priority of a process after it has slept for a while. 301 */ 302 static void 303 updatepri(struct lwp *l) 304 { 305 fixpt_t loadfac; 306 307 KASSERT(lwp_locked(l, NULL)); 308 KASSERT(l->l_slptime > 1); 309 310 loadfac = loadfactor(averunnable.ldavg[0]); 311 312 l->l_slptime--; /* the first time was done in sched_pstats */ 313 l->l_estcpu = decay_cpu_batch(loadfac, l->l_estcpu, l->l_slptime); 314 resetpriority(l); 315 } 316 317 void 318 sched_rqinit(void) 319 { 320 321 } 322 323 void 324 sched_setrunnable(struct lwp *l) 325 { 326 327 if (l->l_slptime > 1) 328 updatepri(l); 329 } 330 331 void 332 sched_nice(struct proc *p, int n) 333 { 334 struct lwp *l; 335 336 KASSERT(mutex_owned(&p->p_smutex)); 337 338 p->p_nice = n; 339 LIST_FOREACH(l, &p->p_lwps, l_sibling) { 340 lwp_lock(l); 341 resetpriority(l); 342 lwp_unlock(l); 343 } 344 } 345 346 /* 347 * Recompute the priority of an LWP. Arrange to reschedule if 348 * the resulting priority is better than that of the current LWP. 349 */ 350 static void 351 resetpriority(struct lwp *l) 352 { 353 pri_t pri; 354 struct proc *p = l->l_proc; 355 356 KASSERT(lwp_locked(l, NULL)); 357 358 if (l->l_class != SCHED_OTHER) 359 return; 360 361 /* See comments above ESTCPU_SHIFT definition. */ 362 pri = (PRI_KERNEL - 1) - (l->l_estcpu >> ESTCPU_SHIFT) - p->p_nice; 363 pri = imax(pri, 0); 364 if (pri != l->l_priority) 365 lwp_changepri(l, pri); 366 } 367 368 /* 369 * We adjust the priority of the current process. The priority of a process 370 * gets worse as it accumulates CPU time. The CPU usage estimator (l_estcpu) 371 * is increased here. The formula for computing priorities (in kern_synch.c) 372 * will compute a different value each time l_estcpu increases. This can 373 * cause a switch, but unless the priority crosses a PPQ boundary the actual 374 * queue will not change. The CPU usage estimator ramps up quite quickly 375 * when the process is running (linearly), and decays away exponentially, at 376 * a rate which is proportionally slower when the system is busy. The basic 377 * principle is that the system will 90% forget that the process used a lot 378 * of CPU time in 5 * loadav seconds. This causes the system to favor 379 * processes which haven't run much recently, and to round-robin among other 380 * processes. 381 */ 382 383 void 384 sched_schedclock(struct lwp *l) 385 { 386 387 if (l->l_class != SCHED_OTHER) 388 return; 389 390 KASSERT(!CURCPU_IDLE_P()); 391 l->l_estcpu = ESTCPULIM(l->l_estcpu + ESTCPU_ACCUM); 392 lwp_lock(l); 393 resetpriority(l); 394 lwp_unlock(l); 395 } 396 397 /* 398 * sched_proc_fork: 399 * 400 * Inherit the parent's scheduler history. 401 */ 402 void 403 sched_proc_fork(struct proc *parent, struct proc *child) 404 { 405 lwp_t *pl; 406 407 KASSERT(mutex_owned(&parent->p_smutex)); 408 409 pl = LIST_FIRST(&parent->p_lwps); 410 child->p_estcpu_inherited = pl->l_estcpu; 411 child->p_forktime = sched_pstats_ticks; 412 } 413 414 /* 415 * sched_proc_exit: 416 * 417 * Chargeback parents for the sins of their children. 418 */ 419 void 420 sched_proc_exit(struct proc *parent, struct proc *child) 421 { 422 fixpt_t loadfac = loadfactor(averunnable.ldavg[0]); 423 fixpt_t estcpu; 424 lwp_t *pl, *cl; 425 426 /* XXX Only if parent != init?? */ 427 428 mutex_enter(&parent->p_smutex); 429 pl = LIST_FIRST(&parent->p_lwps); 430 cl = LIST_FIRST(&child->p_lwps); 431 estcpu = decay_cpu_batch(loadfac, child->p_estcpu_inherited, 432 sched_pstats_ticks - child->p_forktime); 433 if (cl->l_estcpu > estcpu) { 434 lwp_lock(pl); 435 pl->l_estcpu = ESTCPULIM(pl->l_estcpu + cl->l_estcpu - estcpu); 436 lwp_unlock(pl); 437 } 438 mutex_exit(&parent->p_smutex); 439 } 440 441 void 442 sched_wakeup(struct lwp *l) 443 { 444 445 l->l_cpu = sched_takecpu(l); 446 } 447 448 void 449 sched_slept(struct lwp *l) 450 { 451 452 } 453 454 void 455 sched_lwp_fork(struct lwp *l1, struct lwp *l2) 456 { 457 458 l2->l_estcpu = l1->l_estcpu; 459 } 460 461 void 462 sched_lwp_exit(struct lwp *l) 463 { 464 465 } 466 467 void 468 sched_lwp_collect(struct lwp *t) 469 { 470 lwp_t *l; 471 472 /* Absorb estcpu value of collected LWP. */ 473 l = curlwp; 474 lwp_lock(l); 475 l->l_estcpu += t->l_estcpu; 476 lwp_unlock(l); 477 } 478 479 void 480 sched_oncpu(lwp_t *l) 481 { 482 483 } 484 485 void 486 sched_newts(lwp_t *l) 487 { 488 489 } 490 491 /* 492 * Sysctl nodes and initialization. 493 */ 494 495 static int 496 sysctl_sched_rtts(SYSCTLFN_ARGS) 497 { 498 struct sysctlnode node; 499 int rttsms = hztoms(rrticks); 500 501 node = *rnode; 502 node.sysctl_data = &rttsms; 503 return sysctl_lookup(SYSCTLFN_CALL(&node)); 504 } 505 506 SYSCTL_SETUP(sysctl_sched_4bsd_setup, "sysctl sched setup") 507 { 508 const struct sysctlnode *node = NULL; 509 510 sysctl_createv(clog, 0, NULL, NULL, 511 CTLFLAG_PERMANENT, 512 CTLTYPE_NODE, "kern", NULL, 513 NULL, 0, NULL, 0, 514 CTL_KERN, CTL_EOL); 515 sysctl_createv(clog, 0, NULL, &node, 516 CTLFLAG_PERMANENT, 517 CTLTYPE_NODE, "sched", 518 SYSCTL_DESCR("Scheduler options"), 519 NULL, 0, NULL, 0, 520 CTL_KERN, CTL_CREATE, CTL_EOL); 521 522 if (node == NULL) 523 return; 524 525 rrticks = hz / 10; 526 527 sysctl_createv(NULL, 0, &node, NULL, 528 CTLFLAG_PERMANENT, 529 CTLTYPE_STRING, "name", NULL, 530 NULL, 0, __UNCONST("4.4BSD"), 0, 531 CTL_CREATE, CTL_EOL); 532 sysctl_createv(NULL, 0, &node, NULL, 533 CTLFLAG_PERMANENT, 534 CTLTYPE_INT, "rtts", 535 SYSCTL_DESCR("Round-robin time quantum (in miliseconds)"), 536 sysctl_sched_rtts, 0, NULL, 0, 537 CTL_CREATE, CTL_EOL); 538 } 539
Indexes created Tue Sep 23 07:09:52 GMT 2025