sched_m2.c revision 1.24.6.2 1 /* $NetBSD: sched_m2.c,v 1.24.6.2 2008/10/10 22:34:14 skrll Exp $ */
2
3 /*
4 * Copyright (c) 2007, 2008 Mindaugas Rasiukevicius <rmind at NetBSD org>
5 * All rights reserved.
6 *
7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions
9 * are met:
10 * 1. Redistributions of source code must retain the above copyright
11 * notice, this list of conditions and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution.
15 *
16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
20 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26 * SUCH DAMAGE.
27 */
28
29 /*
30 * TODO:
31 * - Implementation of fair share queue;
32 * - Support for NUMA;
33 */
34
35 #include <sys/cdefs.h>
36 __KERNEL_RCSID(0, "$NetBSD: sched_m2.c,v 1.24.6.2 2008/10/10 22:34:14 skrll Exp $");
37
38 #include <sys/param.h>
39
40 #include <sys/bitops.h>
41 #include <sys/cpu.h>
42 #include <sys/callout.h>
43 #include <sys/errno.h>
44 #include <sys/kernel.h>
45 #include <sys/kmem.h>
46 #include <sys/lwp.h>
47 #include <sys/mutex.h>
48 #include <sys/pool.h>
49 #include <sys/proc.h>
50 #include <sys/pset.h>
51 #include <sys/resource.h>
52 #include <sys/resourcevar.h>
53 #include <sys/sched.h>
54 #include <sys/syscallargs.h>
55 #include <sys/sysctl.h>
56 #include <sys/types.h>
57
58 /*
59 * Priority related defintions.
60 */
61 #define PRI_TS_COUNT (NPRI_USER)
62 #define PRI_RT_COUNT (PRI_COUNT - PRI_TS_COUNT)
63 #define PRI_HTS_RANGE (PRI_TS_COUNT / 10)
64
65 #define PRI_HIGHEST_TS (MAXPRI_USER)
66
67 /*
68 * Time-slices and priorities.
69 */
70 static u_int min_ts; /* Minimal time-slice */
71 static u_int max_ts; /* Maximal time-slice */
72 static u_int rt_ts; /* Real-time time-slice */
73 static u_int ts_map[PRI_COUNT]; /* Map of time-slices */
74 static pri_t high_pri[PRI_COUNT]; /* Map for priority increase */
75
76 static void sched_precalcts(void);
77
78 /*
79 * Initialization and setup.
80 */
81
82 void
83 sched_rqinit(void)
84 {
85 struct cpu_info *ci = curcpu();
86
87 if (hz < 100) {
88 panic("sched_rqinit: value of HZ is too low\n");
89 }
90
91 /* Default timing ranges */
92 min_ts = mstohz(20); /* ~20 ms */
93 max_ts = mstohz(150); /* ~150 ms */
94 rt_ts = mstohz(100); /* ~100 ms */
95 sched_precalcts();
96
97 /* Attach the primary CPU here */
98 sched_cpuattach(ci);
99
100 sched_lwp_fork(NULL, &lwp0);
101 sched_newts(&lwp0);
102 }
103
104 /* Pre-calculate the time-slices for the priorities */
105 static void
106 sched_precalcts(void)
107 {
108 pri_t p;
109
110 /* Time-sharing range */
111 for (p = 0; p <= PRI_HIGHEST_TS; p++) {
112 ts_map[p] = max_ts -
113 (p * 100 / (PRI_TS_COUNT - 1) * (max_ts - min_ts) / 100);
114 high_pri[p] = (PRI_HIGHEST_TS - PRI_HTS_RANGE) +
115 ((p * PRI_HTS_RANGE) / (PRI_TS_COUNT - 1));
116 }
117
118 /* Real-time range */
119 for (p = (PRI_HIGHEST_TS + 1); p < PRI_COUNT; p++) {
120 ts_map[p] = rt_ts;
121 high_pri[p] = p;
122 }
123 }
124
125 /*
126 * Hooks.
127 */
128
129 void
130 sched_proc_fork(struct proc *parent, struct proc *child)
131 {
132 struct lwp *l;
133
134 LIST_FOREACH(l, &child->p_lwps, l_sibling) {
135 lwp_lock(l);
136 sched_newts(l);
137 lwp_unlock(l);
138 }
139 }
140
141 void
142 sched_proc_exit(struct proc *child, struct proc *parent)
143 {
144
145 }
146
147 void
148 sched_lwp_fork(struct lwp *l1, struct lwp *l2)
149 {
150
151 }
152
153 void
154 sched_lwp_collect(struct lwp *l)
155 {
156
157 }
158
159 void
160 sched_setrunnable(struct lwp *l)
161 {
162
163 }
164
165 void
166 sched_schedclock(struct lwp *l)
167 {
168
169 }
170
171 /*
172 * Priorities and time-slice.
173 */
174
175 void
176 sched_nice(struct proc *p, int prio)
177 {
178
179 }
180
181 /* Recalculate the time-slice */
182 void
183 sched_newts(struct lwp *l)
184 {
185
186 l->l_sched.timeslice = ts_map[lwp_eprio(l)];
187 }
188
189 void
190 sched_slept(struct lwp *l)
191 {
192
193 /*
194 * If thread is in time-sharing queue and batch flag is not marked,
195 * increase the the priority, and run with the lower time-quantum.
196 */
197 if (l->l_priority < PRI_HIGHEST_TS && (l->l_flag & LW_BATCH) == 0) {
198 KASSERT(l->l_class == SCHED_OTHER);
199 l->l_priority++;
200 }
201 }
202
203 void
204 sched_wakeup(struct lwp *l)
205 {
206
207 /* If thread was sleeping a second or more - set a high priority */
208 if (l->l_slptime >= 1)
209 l->l_priority = high_pri[l->l_priority];
210 }
211
212 void
213 sched_pstats_hook(struct lwp *l, int batch)
214 {
215 pri_t prio;
216
217 /*
218 * Estimate threads on time-sharing queue only, however,
219 * exclude the highest priority for performance purposes.
220 */
221 KASSERT(lwp_locked(l, NULL));
222 if (l->l_priority >= PRI_HIGHEST_TS)
223 return;
224 KASSERT(l->l_class == SCHED_OTHER);
225
226 /* If it is CPU-bound not a first time - decrease the priority */
227 prio = l->l_priority;
228 if (batch && prio != 0)
229 prio--;
230
231 /* If thread was not ran a second or more - set a high priority */
232 if (l->l_stat == LSRUN) {
233 if (l->l_rticks && (hardclock_ticks - l->l_rticks >= hz))
234 prio = high_pri[prio];
235 /* Re-enqueue the thread if priority has changed */
236 if (prio != l->l_priority)
237 lwp_changepri(l, prio);
238 } else {
239 /* In other states, change the priority directly */
240 l->l_priority = prio;
241 }
242 }
243
244 void
245 sched_oncpu(lwp_t *l)
246 {
247 struct schedstate_percpu *spc = &l->l_cpu->ci_schedstate;
248
249 /* Update the counters */
250 KASSERT(l->l_sched.timeslice >= min_ts);
251 KASSERT(l->l_sched.timeslice <= max_ts);
252 spc->spc_ticks = l->l_sched.timeslice;
253 }
254
255 /*
256 * Time-driven events.
257 */
258
259 /*
260 * Called once per time-quantum. This routine is CPU-local and runs at
261 * IPL_SCHED, thus the locking is not needed.
262 */
263 void
264 sched_tick(struct cpu_info *ci)
265 {
266 struct schedstate_percpu *spc = &ci->ci_schedstate;
267 struct lwp *l = curlwp;
268
269 if (__predict_false(CURCPU_IDLE_P()))
270 return;
271
272 switch (l->l_class) {
273 case SCHED_FIFO:
274 /*
275 * Update the time-quantum, and continue running,
276 * if thread runs on FIFO real-time policy.
277 */
278 KASSERT(l->l_priority > PRI_HIGHEST_TS);
279 spc->spc_ticks = l->l_sched.timeslice;
280 return;
281 case SCHED_OTHER:
282 /*
283 * If thread is in time-sharing queue, decrease the priority,
284 * and run with a higher time-quantum.
285 */
286 KASSERT(l->l_priority <= PRI_HIGHEST_TS);
287 if (l->l_priority != 0)
288 l->l_priority--;
289 break;
290 }
291
292 /*
293 * If there are higher priority threads or threads in the same queue,
294 * mark that thread should yield, otherwise, continue running.
295 */
296 if (lwp_eprio(l) <= spc->spc_maxpriority || l->l_target_cpu) {
297 spc->spc_flags |= SPCF_SHOULDYIELD;
298 cpu_need_resched(ci, 0);
299 } else
300 spc->spc_ticks = l->l_sched.timeslice;
301 }
302
303 /*
304 * Sysctl nodes and initialization.
305 */
306
307 static int
308 sysctl_sched_rtts(SYSCTLFN_ARGS)
309 {
310 struct sysctlnode node;
311 int rttsms = hztoms(rt_ts);
312
313 node = *rnode;
314 node.sysctl_data = &rttsms;
315 return sysctl_lookup(SYSCTLFN_CALL(&node));
316 }
317
318 static int
319 sysctl_sched_mints(SYSCTLFN_ARGS)
320 {
321 struct sysctlnode node;
322 struct cpu_info *ci;
323 int error, newsize;
324 CPU_INFO_ITERATOR cii;
325
326 node = *rnode;
327 node.sysctl_data = &newsize;
328
329 newsize = hztoms(min_ts);
330 error = sysctl_lookup(SYSCTLFN_CALL(&node));
331 if (error || newp == NULL)
332 return error;
333
334 newsize = mstohz(newsize);
335 if (newsize < 1 || newsize > hz || newsize >= max_ts)
336 return EINVAL;
337
338 /* It is safe to do this in such order */
339 for (CPU_INFO_FOREACH(cii, ci))
340 spc_lock(ci);
341
342 min_ts = newsize;
343 sched_precalcts();
344
345 for (CPU_INFO_FOREACH(cii, ci))
346 spc_unlock(ci);
347
348 return 0;
349 }
350
351 static int
352 sysctl_sched_maxts(SYSCTLFN_ARGS)
353 {
354 struct sysctlnode node;
355 struct cpu_info *ci;
356 int error, newsize;
357 CPU_INFO_ITERATOR cii;
358
359 node = *rnode;
360 node.sysctl_data = &newsize;
361
362 newsize = hztoms(max_ts);
363 error = sysctl_lookup(SYSCTLFN_CALL(&node));
364 if (error || newp == NULL)
365 return error;
366
367 newsize = mstohz(newsize);
368 if (newsize < 10 || newsize > hz || newsize <= min_ts)
369 return EINVAL;
370
371 /* It is safe to do this in such order */
372 for (CPU_INFO_FOREACH(cii, ci))
373 spc_lock(ci);
374
375 max_ts = newsize;
376 sched_precalcts();
377
378 for (CPU_INFO_FOREACH(cii, ci))
379 spc_unlock(ci);
380
381 return 0;
382 }
383
384 SYSCTL_SETUP(sysctl_sched_m2_setup, "sysctl sched setup")
385 {
386 const struct sysctlnode *node = NULL;
387
388 sysctl_createv(clog, 0, NULL, NULL,
389 CTLFLAG_PERMANENT,
390 CTLTYPE_NODE, "kern", NULL,
391 NULL, 0, NULL, 0,
392 CTL_KERN, CTL_EOL);
393 sysctl_createv(clog, 0, NULL, &node,
394 CTLFLAG_PERMANENT,
395 CTLTYPE_NODE, "sched",
396 SYSCTL_DESCR("Scheduler options"),
397 NULL, 0, NULL, 0,
398 CTL_KERN, CTL_CREATE, CTL_EOL);
399
400 if (node == NULL)
401 return;
402
403 sysctl_createv(NULL, 0, &node, NULL,
404 CTLFLAG_PERMANENT,
405 CTLTYPE_STRING, "name", NULL,
406 NULL, 0, __UNCONST("M2"), 0,
407 CTL_CREATE, CTL_EOL);
408 sysctl_createv(NULL, 0, &node, NULL,
409 CTLFLAG_PERMANENT,
410 CTLTYPE_INT, "rtts",
411 SYSCTL_DESCR("Round-robin time quantum (in miliseconds)"),
412 sysctl_sched_rtts, 0, NULL, 0,
413 CTL_CREATE, CTL_EOL);
414 sysctl_createv(NULL, 0, &node, NULL,
415 CTLFLAG_PERMANENT | CTLFLAG_READWRITE,
416 CTLTYPE_INT, "maxts",
417 SYSCTL_DESCR("Maximal time quantum (in miliseconds)"),
418 sysctl_sched_maxts, 0, &max_ts, 0,
419 CTL_CREATE, CTL_EOL);
420 sysctl_createv(NULL, 0, &node, NULL,
421 CTLFLAG_PERMANENT | CTLFLAG_READWRITE,
422 CTLTYPE_INT, "mints",
423 SYSCTL_DESCR("Minimal time quantum (in miliseconds)"),
424 sysctl_sched_mints, 0, &min_ts, 0,
425 CTL_CREATE, CTL_EOL);
426 }
427