Annotation of src/sys/kern/sched_4bsd.c, Revision 1.7.4.1
1.7.4.1 ! mjf 1: /* $NetBSD: sched_4bsd.c,v 1.8 2007/11/06 00:42:43 ad Exp $ */
1.2 yamt 2:
3: /*-
4: * Copyright (c) 1999, 2000, 2004, 2006, 2007 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>
1.7.4.1 ! mjf 78: __KERNEL_RCSID(0, "$NetBSD: sched_4bsd.c,v 1.8 2007/11/06 00:42:43 ad Exp $");
1.2 yamt 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>
1.5 ad 99: #include <sys/intr.h>
1.2 yamt 100:
101: #include <uvm/uvm_extern.h>
102:
103: /*
104: * Run queues.
105: *
1.7.4.1 ! mjf 106: * We maintain bitmasks of non-empty queues in order speed up finding
! 107: * the first runnable process. Since there can be (by definition) few
! 108: * real time LWPs in the the system, we maintain them on a linked list,
! 109: * sorted by priority.
1.2 yamt 110: */
111:
1.7.4.1 ! mjf 112: #define PPB_SHIFT 5
! 113: #define PPB_MASK 31
! 114:
! 115: #define NUM_Q (NPRI_KERNEL + NPRI_USER)
! 116: #define NUM_PPB (1 << PPB_SHIFT)
! 117: #define NUM_B (NUM_Q / NUM_PPB)
! 118:
1.2 yamt 119: typedef struct runqueue {
1.7.4.1 ! mjf 120: TAILQ_HEAD(, lwp) rq_fixedpri; /* realtime, kthread */
! 121: u_int rq_count; /* total # jobs */
! 122: uint32_t rq_bitmap[NUM_B]; /* bitmap of queues */
! 123: TAILQ_HEAD(, lwp) rq_queue[NUM_Q]; /* user+kernel */
1.2 yamt 124: } runqueue_t;
1.7.4.1 ! mjf 125:
1.2 yamt 126: static runqueue_t global_queue;
127:
128: static void updatepri(struct lwp *);
129: static void resetpriority(struct lwp *);
130:
1.6 rmind 131: fixpt_t decay_cpu(fixpt_t, fixpt_t);
132:
1.2 yamt 133: extern unsigned int sched_pstats_ticks; /* defined in kern_synch.c */
134:
135: /* The global scheduler state */
136: kmutex_t sched_mutex;
137:
138: /* Number of hardclock ticks per sched_tick() */
139: int rrticks;
140:
1.7.4.1 ! mjf 141: const int schedppq = 1;
! 142:
1.2 yamt 143: /*
144: * Force switch among equal priority processes every 100ms.
145: * Called from hardclock every hz/10 == rrticks hardclock ticks.
1.5 ad 146: *
147: * There's no need to lock anywhere in this routine, as it's
148: * CPU-local and runs at IPL_SCHED (called from clock interrupt).
1.2 yamt 149: */
150: /* ARGSUSED */
151: void
152: sched_tick(struct cpu_info *ci)
153: {
154: struct schedstate_percpu *spc = &ci->ci_schedstate;
155:
156: spc->spc_ticks = rrticks;
157:
1.7 rmind 158: if (CURCPU_IDLE_P())
159: return;
160:
161: if (spc->spc_flags & SPCF_SEENRR) {
162: /*
163: * The process has already been through a roundrobin
164: * without switching and may be hogging the CPU.
165: * Indicate that the process should yield.
166: */
167: spc->spc_flags |= SPCF_SHOULDYIELD;
168: } else
169: spc->spc_flags |= SPCF_SEENRR;
170:
171: cpu_need_resched(ci, 0);
1.2 yamt 172: }
173:
1.7.4.1 ! mjf 174: /*
! 175: * Why PRIO_MAX - 2? From setpriority(2):
! 176: *
! 177: * prio is a value in the range -20 to 20. The default priority is
! 178: * 0; lower priorities cause more favorable scheduling. A value of
! 179: * 19 or 20 will schedule a process only when nothing at priority <=
! 180: * 0 is runnable.
! 181: *
! 182: * This gives estcpu influence over 18 priority levels, and leaves nice
! 183: * with 40 levels. One way to think about it is that nice has 20 levels
! 184: * either side of estcpu's 18.
! 185: */
1.2 yamt 186: #define ESTCPU_SHIFT 11
1.7.4.1 ! mjf 187: #define ESTCPU_MAX ((PRIO_MAX - 2) << ESTCPU_SHIFT)
! 188: #define ESTCPU_ACCUM (1 << (ESTCPU_SHIFT - 1))
1.2 yamt 189: #define ESTCPULIM(e) min((e), ESTCPU_MAX)
190:
191: /*
192: * Constants for digital decay and forget:
1.7.4.1 ! mjf 193: * 90% of (l_estcpu) usage in 5 * loadav time
! 194: * 95% of (l_pctcpu) usage in 60 seconds (load insensitive)
1.2 yamt 195: * Note that, as ps(1) mentions, this can let percentages
196: * total over 100% (I've seen 137.9% for 3 processes).
197: *
1.7.4.1 ! mjf 198: * Note that hardclock updates l_estcpu and l_cpticks independently.
1.2 yamt 199: *
1.7.4.1 ! mjf 200: * We wish to decay away 90% of l_estcpu in (5 * loadavg) seconds.
1.2 yamt 201: * That is, the system wants to compute a value of decay such
202: * that the following for loop:
203: * for (i = 0; i < (5 * loadavg); i++)
1.7.4.1 ! mjf 204: * l_estcpu *= decay;
1.2 yamt 205: * will compute
1.7.4.1 ! mjf 206: * l_estcpu *= 0.1;
1.2 yamt 207: * for all values of loadavg:
208: *
209: * Mathematically this loop can be expressed by saying:
210: * decay ** (5 * loadavg) ~= .1
211: *
212: * The system computes decay as:
213: * decay = (2 * loadavg) / (2 * loadavg + 1)
214: *
215: * We wish to prove that the system's computation of decay
216: * will always fulfill the equation:
217: * decay ** (5 * loadavg) ~= .1
218: *
219: * If we compute b as:
220: * b = 2 * loadavg
221: * then
222: * decay = b / (b + 1)
223: *
224: * We now need to prove two things:
225: * 1) Given factor ** (5 * loadavg) ~= .1, prove factor == b/(b+1)
226: * 2) Given b/(b+1) ** power ~= .1, prove power == (5 * loadavg)
227: *
228: * Facts:
229: * For x close to zero, exp(x) =~ 1 + x, since
230: * exp(x) = 0! + x**1/1! + x**2/2! + ... .
231: * therefore exp(-1/b) =~ 1 - (1/b) = (b-1)/b.
232: * For x close to zero, ln(1+x) =~ x, since
233: * ln(1+x) = x - x**2/2 + x**3/3 - ... -1 < x < 1
234: * therefore ln(b/(b+1)) = ln(1 - 1/(b+1)) =~ -1/(b+1).
235: * ln(.1) =~ -2.30
236: *
237: * Proof of (1):
238: * Solve (factor)**(power) =~ .1 given power (5*loadav):
239: * solving for factor,
240: * ln(factor) =~ (-2.30/5*loadav), or
241: * factor =~ exp(-1/((5/2.30)*loadav)) =~ exp(-1/(2*loadav)) =
242: * exp(-1/b) =~ (b-1)/b =~ b/(b+1). QED
243: *
244: * Proof of (2):
245: * Solve (factor)**(power) =~ .1 given factor == (b/(b+1)):
246: * solving for power,
247: * power*ln(b/(b+1)) =~ -2.30, or
248: * power =~ 2.3 * (b + 1) = 4.6*loadav + 2.3 =~ 5*loadav. QED
249: *
250: * Actual power values for the implemented algorithm are as follows:
251: * loadav: 1 2 3 4
252: * power: 5.68 10.32 14.94 19.55
253: */
254:
255: /* calculations for digital decay to forget 90% of usage in 5*loadav sec */
256: #define loadfactor(loadav) (2 * (loadav))
257:
1.6 rmind 258: fixpt_t
1.2 yamt 259: decay_cpu(fixpt_t loadfac, fixpt_t estcpu)
260: {
261:
262: if (estcpu == 0) {
263: return 0;
264: }
265:
266: #if !defined(_LP64)
267: /* avoid 64bit arithmetics. */
268: #define FIXPT_MAX ((fixpt_t)((UINTMAX_C(1) << sizeof(fixpt_t) * CHAR_BIT) - 1))
269: if (__predict_true(loadfac <= FIXPT_MAX / ESTCPU_MAX)) {
270: return estcpu * loadfac / (loadfac + FSCALE);
271: }
272: #endif /* !defined(_LP64) */
273:
274: return (uint64_t)estcpu * loadfac / (loadfac + FSCALE);
275: }
276:
277: /*
1.7.4.1 ! mjf 278: * For all load averages >= 1 and max l_estcpu of (255 << ESTCPU_SHIFT),
! 279: * sleeping for at least seven times the loadfactor will decay l_estcpu to
1.2 yamt 280: * less than (1 << ESTCPU_SHIFT).
281: *
282: * note that our ESTCPU_MAX is actually much smaller than (255 << ESTCPU_SHIFT).
283: */
284: static fixpt_t
285: decay_cpu_batch(fixpt_t loadfac, fixpt_t estcpu, unsigned int n)
286: {
287:
288: if ((n << FSHIFT) >= 7 * loadfac) {
289: return 0;
290: }
291:
292: while (estcpu != 0 && n > 1) {
293: estcpu = decay_cpu(loadfac, estcpu);
294: n--;
295: }
296:
297: return estcpu;
298: }
299:
300: /*
301: * sched_pstats_hook:
302: *
303: * Periodically called from sched_pstats(); used to recalculate priorities.
304: */
305: void
1.6 rmind 306: sched_pstats_hook(struct lwp *l)
1.2 yamt 307: {
1.7.4.1 ! mjf 308: fixpt_t loadfac;
! 309: int sleeptm;
1.2 yamt 310:
1.7.4.1 ! mjf 311: /*
! 312: * If the LWP has slept an entire second, stop recalculating
! 313: * its priority until it wakes up.
! 314: */
! 315: if (l->l_stat == LSSLEEP || l->l_stat == LSSTOP ||
! 316: l->l_stat == LSSUSPENDED) {
! 317: l->l_slptime++;
! 318: sleeptm = 1;
! 319: } else {
! 320: sleeptm = 0x7fffffff;
! 321: }
! 322:
! 323: if (l->l_slptime <= sleeptm) {
! 324: loadfac = 2 * (averunnable.ldavg[0]);
! 325: l->l_estcpu = decay_cpu(loadfac, l->l_estcpu);
1.6 rmind 326: resetpriority(l);
1.7.4.1 ! mjf 327: }
1.2 yamt 328: }
329:
330: /*
331: * Recalculate the priority of a process after it has slept for a while.
332: */
333: static void
334: updatepri(struct lwp *l)
335: {
336: fixpt_t loadfac;
337:
1.3 ad 338: KASSERT(lwp_locked(l, NULL));
1.2 yamt 339: KASSERT(l->l_slptime > 1);
340:
341: loadfac = loadfactor(averunnable.ldavg[0]);
342:
343: l->l_slptime--; /* the first time was done in sched_pstats */
1.7.4.1 ! mjf 344: l->l_estcpu = decay_cpu_batch(loadfac, l->l_estcpu, l->l_slptime);
1.2 yamt 345: resetpriority(l);
346: }
347:
348: static void
349: runqueue_init(runqueue_t *rq)
350: {
351: int i;
352:
1.7.4.1 ! mjf 353: for (i = 0; i < NUM_Q; i++)
! 354: TAILQ_INIT(&rq->rq_queue[i]);
! 355: for (i = 0; i < NUM_B; i++)
! 356: rq->rq_bitmap[i] = 0;
! 357: TAILQ_INIT(&rq->rq_fixedpri);
! 358: rq->rq_count = 0;
1.2 yamt 359: }
360:
361: static void
362: runqueue_enqueue(runqueue_t *rq, struct lwp *l)
363: {
1.7.4.1 ! mjf 364: pri_t pri;
! 365: lwp_t *l2;
1.2 yamt 366:
367: KASSERT(lwp_locked(l, l->l_cpu->ci_schedstate.spc_mutex));
368:
1.7.4.1 ! mjf 369: pri = lwp_eprio(l);
! 370: rq->rq_count++;
! 371:
! 372: if (pri >= PRI_KTHREAD) {
! 373: TAILQ_FOREACH(l2, &rq->rq_fixedpri, l_runq) {
! 374: if (lwp_eprio(l2) < pri) {
! 375: TAILQ_INSERT_BEFORE(l2, l, l_runq);
! 376: return;
! 377: }
! 378: }
! 379: TAILQ_INSERT_TAIL(&rq->rq_fixedpri, l, l_runq);
! 380: return;
! 381: }
! 382:
! 383: rq->rq_bitmap[pri >> PPB_SHIFT] |=
! 384: (0x80000000U >> (pri & PPB_MASK));
! 385: TAILQ_INSERT_TAIL(&rq->rq_queue[pri], l, l_runq);
1.2 yamt 386: }
387:
388: static void
389: runqueue_dequeue(runqueue_t *rq, struct lwp *l)
390: {
1.7.4.1 ! mjf 391: pri_t pri;
1.2 yamt 392:
393: KASSERT(lwp_locked(l, l->l_cpu->ci_schedstate.spc_mutex));
394:
1.7.4.1 ! mjf 395: pri = lwp_eprio(l);
! 396: rq->rq_count--;
! 397:
! 398: if (pri >= PRI_KTHREAD) {
! 399: TAILQ_REMOVE(&rq->rq_fixedpri, l, l_runq);
! 400: return;
! 401: }
! 402:
! 403: TAILQ_REMOVE(&rq->rq_queue[pri], l, l_runq);
! 404: if (TAILQ_EMPTY(&rq->rq_queue[pri]))
! 405: rq->rq_bitmap[pri >> PPB_SHIFT] ^=
! 406: (0x80000000U >> (pri & PPB_MASK));
1.2 yamt 407: }
408:
1.7.4.1 ! mjf 409: #if (NUM_B != 3) || (NUM_Q != 96)
! 410: #error adjust runqueue_nextlwp
! 411: #endif
! 412:
1.2 yamt 413: static struct lwp *
414: runqueue_nextlwp(runqueue_t *rq)
415: {
1.7.4.1 ! mjf 416: pri_t pri;
1.2 yamt 417:
1.7.4.1 ! mjf 418: KASSERT(rq->rq_count != 0);
! 419:
! 420: if (!TAILQ_EMPTY(&rq->rq_fixedpri))
! 421: return TAILQ_FIRST(&rq->rq_fixedpri);
! 422:
! 423: if (rq->rq_bitmap[2] != 0)
! 424: pri = 96 - ffs(rq->rq_bitmap[2]);
! 425: else if (rq->rq_bitmap[1] != 0)
! 426: pri = 64 - ffs(rq->rq_bitmap[1]);
! 427: else
! 428: pri = 32 - ffs(rq->rq_bitmap[0]);
! 429: return TAILQ_FIRST(&rq->rq_queue[pri]);
1.2 yamt 430: }
431:
432: #if defined(DDB)
433: static void
434: runqueue_print(const runqueue_t *rq, void (*pr)(const char *, ...))
435: {
1.7.4.1 ! mjf 436: CPU_INFO_ITERATOR cii;
! 437: struct cpu_info *ci;
! 438: lwp_t *l;
! 439: int i;
1.2 yamt 440:
1.7.4.1 ! mjf 441: printf("PID\tLID\tPRI\tIPRI\tEPRI\tLWP\t\t NAME\n");
! 442:
! 443: TAILQ_FOREACH(l, &rq->rq_fixedpri, l_runq) {
! 444: (*pr)("%d\t%d\%d\t%d\t%d\t%016lx %s\n",
! 445: l->l_proc->p_pid, l->l_lid, (int)l->l_priority,
! 446: (int)l->l_inheritedprio, lwp_eprio(l),
! 447: (long)l, l->l_proc->p_comm);
! 448: }
! 449:
! 450: for (i = NUM_Q - 1; i >= 0; i--) {
! 451: TAILQ_FOREACH(l, &rq->rq_queue[i], l_runq) {
! 452: (*pr)("%d\t%d\t%d\t%d\t%d\t%016lx %s\n",
! 453: l->l_proc->p_pid, l->l_lid, (int)l->l_priority,
! 454: (int)l->l_inheritedprio, lwp_eprio(l),
! 455: (long)l, l->l_proc->p_comm);
1.2 yamt 456: }
457: }
1.7.4.1 ! mjf 458:
! 459: printf("CPUIDX\tRESCHED\tCURPRI\tFLAGS\n");
! 460: for (CPU_INFO_FOREACH(cii, ci)) {
! 461: printf("%d\t%d\t%d\t%04x\n", (int)ci->ci_index,
! 462: (int)ci->ci_want_resched,
! 463: (int)ci->ci_schedstate.spc_curpriority,
! 464: (int)ci->ci_schedstate.spc_flags);
! 465: }
! 466:
! 467: printf("NEXTLWP\n%016lx\n", (long)sched_nextlwp());
1.2 yamt 468: }
469: #endif /* defined(DDB) */
470:
471: /*
472: * Initialize the (doubly-linked) run queues
473: * to be empty.
474: */
475: void
476: sched_rqinit()
477: {
478:
479: runqueue_init(&global_queue);
480: mutex_init(&sched_mutex, MUTEX_SPIN, IPL_SCHED);
481: /* Initialize the lock pointer for lwp0 */
482: lwp0.l_mutex = &curcpu()->ci_schedstate.spc_lwplock;
483: }
484:
485: void
486: sched_cpuattach(struct cpu_info *ci)
487: {
488: runqueue_t *rq;
489:
490: ci->ci_schedstate.spc_mutex = &sched_mutex;
491: rq = kmem_zalloc(sizeof(*rq), KM_NOSLEEP);
492: runqueue_init(rq);
493: ci->ci_schedstate.spc_sched_info = rq;
494: }
495:
496: void
497: sched_setup()
498: {
499:
500: rrticks = hz / 10;
501: }
502:
503: void
504: sched_setrunnable(struct lwp *l)
505: {
506:
507: if (l->l_slptime > 1)
508: updatepri(l);
509: }
510:
511: bool
512: sched_curcpu_runnable_p(void)
513: {
1.4 ad 514: struct schedstate_percpu *spc;
1.7.4.1 ! mjf 515: struct cpu_info *ci;
! 516: int bits;
1.2 yamt 517:
1.7.4.1 ! mjf 518: ci = curcpu();
! 519: spc = &ci->ci_schedstate;
! 520: #ifndef __HAVE_FAST_SOFTINTS
! 521: bits = ci->ci_data.cpu_softints;
! 522: bits |= ((runqueue_t *)spc->spc_sched_info)->rq_count;
! 523: #else
! 524: bits = ((runqueue_t *)spc->spc_sched_info)->rq_count;
! 525: #endif
1.4 ad 526: if (__predict_true((spc->spc_flags & SPCF_OFFLINE) == 0))
1.7.4.1 ! mjf 527: bits |= global_queue.rq_count;
! 528: return bits != 0;
1.2 yamt 529: }
530:
531: void
1.7.4.1 ! mjf 532: sched_nice(struct proc *p, int n)
1.2 yamt 533: {
1.7.4.1 ! mjf 534: struct lwp *l;
! 535:
! 536: KASSERT(mutex_owned(&p->p_smutex));
1.2 yamt 537:
1.7.4.1 ! mjf 538: p->p_nice = n;
! 539: LIST_FOREACH(l, &p->p_lwps, l_sibling) {
! 540: lwp_lock(l);
! 541: resetpriority(l);
! 542: lwp_unlock(l);
! 543: }
1.2 yamt 544: }
545:
546: /*
1.7.4.1 ! mjf 547: * Recompute the priority of an LWP. Arrange to reschedule if
! 548: * the resulting priority is better than that of the current LWP.
1.2 yamt 549: */
550: static void
551: resetpriority(struct lwp *l)
552: {
1.7.4.1 ! mjf 553: pri_t pri;
1.2 yamt 554: struct proc *p = l->l_proc;
555:
1.7.4.1 ! mjf 556: KASSERT(lwp_locked(l, NULL));
1.2 yamt 557:
1.7.4.1 ! mjf 558: if (l->l_class != SCHED_OTHER)
1.2 yamt 559: return;
560:
1.7.4.1 ! mjf 561: /* See comments above ESTCPU_SHIFT definition. */
! 562: pri = (PRI_KERNEL - 1) - (l->l_estcpu >> ESTCPU_SHIFT) - p->p_nice;
! 563: pri = imax(pri, 0);
! 564: if (pri != l->l_priority)
! 565: lwp_changepri(l, pri);
1.2 yamt 566: }
567:
568: /*
569: * We adjust the priority of the current process. The priority of a process
1.7.4.1 ! mjf 570: * gets worse as it accumulates CPU time. The CPU usage estimator (l_estcpu)
1.2 yamt 571: * is increased here. The formula for computing priorities (in kern_synch.c)
1.7.4.1 ! mjf 572: * will compute a different value each time l_estcpu increases. This can
1.2 yamt 573: * cause a switch, but unless the priority crosses a PPQ boundary the actual
574: * queue will not change. The CPU usage estimator ramps up quite quickly
575: * when the process is running (linearly), and decays away exponentially, at
576: * a rate which is proportionally slower when the system is busy. The basic
577: * principle is that the system will 90% forget that the process used a lot
578: * of CPU time in 5 * loadav seconds. This causes the system to favor
579: * processes which haven't run much recently, and to round-robin among other
580: * processes.
581: */
582:
583: void
584: sched_schedclock(struct lwp *l)
585: {
1.7.4.1 ! mjf 586:
! 587: if (l->l_class != SCHED_OTHER)
! 588: return;
1.2 yamt 589:
590: KASSERT(!CURCPU_IDLE_P());
1.7.4.1 ! mjf 591: l->l_estcpu = ESTCPULIM(l->l_estcpu + ESTCPU_ACCUM);
1.2 yamt 592: lwp_lock(l);
593: resetpriority(l);
594: lwp_unlock(l);
595: }
596:
597: /*
598: * sched_proc_fork:
599: *
600: * Inherit the parent's scheduler history.
601: */
602: void
603: sched_proc_fork(struct proc *parent, struct proc *child)
604: {
1.7.4.1 ! mjf 605: lwp_t *pl;
1.2 yamt 606:
1.3 ad 607: KASSERT(mutex_owned(&parent->p_smutex));
1.2 yamt 608:
1.7.4.1 ! mjf 609: pl = LIST_FIRST(&parent->p_lwps);
! 610: child->p_estcpu_inherited = pl->l_estcpu;
1.2 yamt 611: child->p_forktime = sched_pstats_ticks;
612: }
613:
614: /*
615: * sched_proc_exit:
616: *
617: * Chargeback parents for the sins of their children.
618: */
619: void
620: sched_proc_exit(struct proc *parent, struct proc *child)
621: {
622: fixpt_t loadfac = loadfactor(averunnable.ldavg[0]);
623: fixpt_t estcpu;
1.7.4.1 ! mjf 624: lwp_t *pl, *cl;
1.2 yamt 625:
626: /* XXX Only if parent != init?? */
627:
1.7.4.1 ! mjf 628: mutex_enter(&parent->p_smutex);
! 629: pl = LIST_FIRST(&parent->p_lwps);
! 630: cl = LIST_FIRST(&child->p_lwps);
1.2 yamt 631: estcpu = decay_cpu_batch(loadfac, child->p_estcpu_inherited,
632: sched_pstats_ticks - child->p_forktime);
1.7.4.1 ! mjf 633: if (cl->l_estcpu > estcpu) {
! 634: lwp_lock(pl);
! 635: pl->l_estcpu = ESTCPULIM(pl->l_estcpu + cl->l_estcpu - estcpu);
! 636: lwp_unlock(pl);
! 637: }
! 638: mutex_exit(&parent->p_smutex);
1.2 yamt 639: }
640:
641: void
642: sched_enqueue(struct lwp *l, bool ctxswitch)
643: {
644:
645: if ((l->l_flag & LW_BOUND) != 0)
646: runqueue_enqueue(l->l_cpu->ci_schedstate.spc_sched_info, l);
647: else
648: runqueue_enqueue(&global_queue, l);
649: }
650:
651: /*
652: * XXXSMP When LWP dispatch (cpu_switch()) is changed to use sched_dequeue(),
653: * drop of the effective priority level from kernel to user needs to be
654: * moved here from userret(). The assignment in userret() is currently
655: * done unlocked.
656: */
657: void
658: sched_dequeue(struct lwp *l)
659: {
660:
661: if ((l->l_flag & LW_BOUND) != 0)
662: runqueue_dequeue(l->l_cpu->ci_schedstate.spc_sched_info, l);
663: else
664: runqueue_dequeue(&global_queue, l);
665: }
666:
667: struct lwp *
668: sched_nextlwp(void)
669: {
1.4 ad 670: struct schedstate_percpu *spc;
1.7.4.1 ! mjf 671: runqueue_t *rq;
1.2 yamt 672: lwp_t *l1, *l2;
673:
1.4 ad 674: spc = &curcpu()->ci_schedstate;
675:
1.2 yamt 676: /* For now, just pick the highest priority LWP. */
1.7.4.1 ! mjf 677: rq = spc->spc_sched_info;
! 678: l1 = NULL;
! 679: if (rq->rq_count != 0)
! 680: l1 = runqueue_nextlwp(rq);
! 681:
! 682: rq = &global_queue;
! 683: if (__predict_false((spc->spc_flags & SPCF_OFFLINE) != 0) ||
! 684: rq->rq_count == 0)
1.4 ad 685: return l1;
1.7.4.1 ! mjf 686: l2 = runqueue_nextlwp(rq);
1.2 yamt 687:
688: if (l1 == NULL)
689: return l2;
690: if (l2 == NULL)
691: return l1;
1.7.4.1 ! mjf 692: if (lwp_eprio(l2) > lwp_eprio(l1))
1.2 yamt 693: return l2;
694: else
695: return l1;
696: }
697:
1.6 rmind 698: struct cpu_info *
699: sched_takecpu(struct lwp *l)
700: {
701:
702: return l->l_cpu;
703: }
704:
705: void
706: sched_wakeup(struct lwp *l)
707: {
708:
709: }
710:
711: void
712: sched_slept(struct lwp *l)
713: {
714:
715: }
716:
1.2 yamt 717: void
1.7.4.1 ! mjf 718: sched_lwp_fork(struct lwp *l1, struct lwp *l2)
1.2 yamt 719: {
720:
1.7.4.1 ! mjf 721: l2->l_estcpu = l1->l_estcpu;
1.2 yamt 722: }
723:
724: void
725: sched_lwp_exit(struct lwp *l)
726: {
727:
728: }
729:
1.7.4.1 ! mjf 730: void
! 731: sched_lwp_collect(struct lwp *t)
! 732: {
! 733: lwp_t *l;
! 734:
! 735: /* Absorb estcpu value of collected LWP. */
! 736: l = curlwp;
! 737: lwp_lock(l);
! 738: l->l_estcpu += t->l_estcpu;
! 739: lwp_unlock(l);
! 740: }
! 741:
1.5 ad 742: /*
743: * sysctl setup. XXX This should be split with kern_synch.c.
744: */
1.2 yamt 745: SYSCTL_SETUP(sysctl_sched_setup, "sysctl kern.sched subtree setup")
746: {
747: const struct sysctlnode *node = NULL;
748:
749: sysctl_createv(clog, 0, NULL, NULL,
750: CTLFLAG_PERMANENT,
751: CTLTYPE_NODE, "kern", NULL,
752: NULL, 0, NULL, 0,
753: CTL_KERN, CTL_EOL);
754: sysctl_createv(clog, 0, NULL, &node,
755: CTLFLAG_PERMANENT,
756: CTLTYPE_NODE, "sched",
757: SYSCTL_DESCR("Scheduler options"),
758: NULL, 0, NULL, 0,
759: CTL_KERN, CTL_CREATE, CTL_EOL);
760:
1.5 ad 761: KASSERT(node != NULL);
762:
763: sysctl_createv(clog, 0, &node, NULL,
764: CTLFLAG_PERMANENT,
765: CTLTYPE_STRING, "name", NULL,
766: NULL, 0, __UNCONST("4.4BSD"), 0,
767: CTL_CREATE, CTL_EOL);
768: sysctl_createv(clog, 0, &node, NULL,
769: CTLFLAG_READWRITE,
770: CTLTYPE_INT, "timesoftints",
771: SYSCTL_DESCR("Track CPU time for soft interrupts"),
772: NULL, 0, &softint_timing, 0,
773: CTL_CREATE, CTL_EOL);
1.2 yamt 774: }
775:
776: #if defined(DDB)
777: void
778: sched_print_runqueue(void (*pr)(const char *, ...))
779: {
780:
781: runqueue_print(&global_queue, pr);
782: }
783: #endif /* defined(DDB) */
CVSweb <webmaster@jp.NetBSD.org>