subr_evcnt.c revision 1.16 1 /* $NetBSD: subr_evcnt.c,v 1.16 2021/04/15 00:37:31 rin Exp $ */
2
3 /*
4 * Copyright (c) 1996, 2000 Christopher G. Demetriou
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 * 3. All advertising materials mentioning features or use of this software
16 * must display the following acknowledgement:
17 * This product includes software developed for the
18 * NetBSD Project. See http://www.NetBSD.org/ for
19 * information about NetBSD.
20 * 4. The name of the author may not be used to endorse or promote products
21 * derived from this software without specific prior written permission.
22 *
23 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
24 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
25 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
26 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
27 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
28 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
29 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
30 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
31 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
32 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
33 *
34 * --(license Id: LICENSE.proto,v 1.1 2000/06/13 21:40:26 cgd Exp )--
35 */
36
37 /*
38 * Copyright (c) 1992, 1993
39 * The Regents of the University of California. All rights reserved.
40 *
41 * This software was developed by the Computer Systems Engineering group
42 * at Lawrence Berkeley Laboratory under DARPA contract BG 91-66 and
43 * contributed to Berkeley.
44 *
45 * All advertising materials mentioning features or use of this software
46 * must display the following acknowledgement:
47 * This product includes software developed by the University of
48 * California, Lawrence Berkeley Laboratories.
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 * from: Header: subr_autoconf.c,v 1.12 93/02/01 19:31:48 torek Exp (LBL)
75 *
76 * @(#)subr_autoconf.c 8.3 (Berkeley) 5/17/94
77 */
78
79 #include <sys/cdefs.h>
80 __KERNEL_RCSID(0, "$NetBSD: subr_evcnt.c,v 1.16 2021/04/15 00:37:31 rin Exp $");
81
82 #include <sys/param.h>
83 #include <sys/evcnt.h>
84 #include <sys/kmem.h>
85 #include <sys/mutex.h>
86 #include <sys/sysctl.h>
87 #include <sys/systm.h>
88
89 /*
90 * Everything related to __HAVE_LEGACY_INTRCNT can disappear once
91 * no more ports are using old-style intrcnt/intrnames interrupt
92 * accounting. The follow files have __HAVE_LEGACY_INTRCNT code:
93 *
94 * sys/kern/init_main.c
95 * sys/kern/subr_evcnt.c
96 * sys/sys/evcnt.h
97 * sys/arch/<port>/include/types.h
98 */
99 #ifdef _RUMPKERNEL
100 /* RUMP doesn't need/want to know about intrcnts */
101 #undef __HAVE_LEGACY_INTRCNT
102 #endif
103
104 #ifdef __HAVE_LEGACY_INTRCNT
105 static void evcnt_update_intrcnt(void);
106 #endif
107
108 /* list of all events */
109 struct evcntlist allevents = TAILQ_HEAD_INITIALIZER(allevents);
110 static kmutex_t evcnt_lock __cacheline_aligned;
111 static bool init_done;
112 static uint32_t evcnt_generation;
113
114 /*
115 * We need a dummy object to stuff into the evcnt link set to
116 * ensure that there always is at least one object in the set.
117 */
118 static struct evcnt dummy_static_evcnt;
119 __link_set_add_bss(evcnts, dummy_static_evcnt);
120
121 /*
122 * Initialize event counters. This does the attach procedure for
123 * each of the static event counters in the "evcnts" link set.
124 */
125 void
126 evcnt_init(void)
127 {
128 __link_set_decl(evcnts, struct evcnt);
129 struct evcnt * const *evp;
130
131 KASSERT(!init_done);
132
133 mutex_init(&evcnt_lock, MUTEX_DEFAULT, IPL_NONE);
134
135 init_done = true;
136
137 __link_set_foreach(evp, evcnts) {
138 if (*evp == &dummy_static_evcnt)
139 continue;
140 evcnt_attach_static(*evp);
141 }
142 }
143
144 /*
145 * Attach a statically-initialized event. The type and string pointers
146 * are already set up.
147 */
148 void
149 evcnt_attach_static(struct evcnt *ev)
150 {
151 int len;
152
153 KASSERTMSG(init_done,
154 "%s: evcnt non initialized: group=<%s> name=<%s>",
155 __func__, ev->ev_group, ev->ev_name);
156
157 len = strlen(ev->ev_group);
158 #ifdef DIAGNOSTIC
159 if (len == 0 || len >= EVCNT_STRING_MAX) /* ..._MAX includes NUL */
160 panic("evcnt_attach_static: group length (%s)", ev->ev_group);
161 #endif
162 ev->ev_grouplen = len;
163
164 len = strlen(ev->ev_name);
165 #ifdef DIAGNOSTIC
166 if (len == 0 || len >= EVCNT_STRING_MAX) /* ..._MAX includes NUL */
167 panic("evcnt_attach_static: name length (%s)", ev->ev_name);
168 #endif
169 ev->ev_namelen = len;
170
171 mutex_enter(&evcnt_lock);
172 TAILQ_INSERT_TAIL(&allevents, ev, ev_list);
173 mutex_exit(&evcnt_lock);
174 }
175
176 /*
177 * Attach a dynamically-initialized event. Zero it, set up the type
178 * and string pointers and then act like it was statically initialized.
179 */
180 void
181 evcnt_attach_dynamic_nozero(struct evcnt *ev, int type,
182 const struct evcnt *parent, const char *group, const char *name)
183 {
184
185 ev->ev_type = type;
186 ev->ev_parent = parent;
187 ev->ev_group = group;
188 ev->ev_name = name;
189 evcnt_attach_static(ev);
190 }
191 /*
192 * Attach a dynamically-initialized event. Zero it, set up the type
193 * and string pointers and then act like it was statically initialized.
194 */
195 void
196 evcnt_attach_dynamic(struct evcnt *ev, int type, const struct evcnt *parent,
197 const char *group, const char *name)
198 {
199
200 memset(ev, 0, sizeof *ev);
201 evcnt_attach_dynamic_nozero(ev, type, parent, group, name);
202 }
203
204 /*
205 * Detach an event.
206 */
207 void
208 evcnt_detach(struct evcnt *ev)
209 {
210
211 mutex_enter(&evcnt_lock);
212 TAILQ_REMOVE(&allevents, ev, ev_list);
213 evcnt_generation++;
214 mutex_exit(&evcnt_lock);
215 }
216
217 struct xevcnt_sysctl {
218 struct evcnt_sysctl evs;
219 char ev_strings[2*EVCNT_STRING_MAX];
220 };
221
222 static size_t
223 sysctl_fillevcnt(const struct evcnt *ev, struct xevcnt_sysctl *xevs,
224 size_t *copylenp)
225 {
226 const bool allowaddr = get_expose_address(curproc);
227 const size_t copylen = offsetof(struct evcnt_sysctl, ev_strings)
228 + ev->ev_grouplen + 1 + ev->ev_namelen + 1;
229 const size_t len = roundup2(copylen, sizeof(uint64_t));
230
231 if (xevs != NULL) {
232 xevs->evs.ev_count = ev->ev_count;
233 COND_SET_VALUE(xevs->evs.ev_addr, PTRTOUINT64(ev), allowaddr);
234 COND_SET_VALUE(xevs->evs.ev_parent, PTRTOUINT64(ev->ev_parent),
235 allowaddr);
236 xevs->evs.ev_type = ev->ev_type;
237 xevs->evs.ev_grouplen = ev->ev_grouplen;
238 xevs->evs.ev_namelen = ev->ev_namelen;
239 xevs->evs.ev_len = len / sizeof(uint64_t);
240 strcpy(xevs->evs.ev_strings, ev->ev_group);
241 strcpy(xevs->evs.ev_strings + ev->ev_grouplen + 1, ev->ev_name);
242 }
243
244 *copylenp = copylen;
245 return len;
246 }
247
248 static int
249 sysctl_doevcnt(SYSCTLFN_ARGS)
250 {
251 struct xevcnt_sysctl *xevs0 = NULL, *xevs;
252 const struct evcnt *ev;
253 int error;
254 int retries;
255 size_t needed, len;
256 char *dp;
257
258 if (namelen == 1 && name[0] == CTL_QUERY)
259 return (sysctl_query(SYSCTLFN_CALL(rnode)));
260
261 if (namelen != 2)
262 return (EINVAL);
263
264 /*
265 * We can filter on the type of evcnt.
266 */
267 const int filter = name[0];
268 if (filter != EVCNT_TYPE_ANY
269 && filter != EVCNT_TYPE_MISC
270 && filter != EVCNT_TYPE_INTR
271 && filter != EVCNT_TYPE_TRAP)
272 return (EINVAL);
273
274 const u_int count = name[1];
275 if (count != KERN_EVCNT_COUNT_ANY
276 && count != KERN_EVCNT_COUNT_NONZERO)
277 return (EINVAL);
278
279 sysctl_unlock();
280
281 if (oldp != NULL && xevs0 == NULL)
282 xevs0 = kmem_zalloc(sizeof(*xevs0), KM_SLEEP);
283
284 retries = 100;
285 retry:
286 dp = oldp;
287 len = (oldp != NULL) ? *oldlenp : 0;
288 xevs = xevs0;
289 error = 0;
290 needed = 0;
291
292 mutex_enter(&evcnt_lock);
293 #ifdef __HAVE_LEGACY_INTRCNT
294 evcnt_update_intrcnt();
295 #endif
296 TAILQ_FOREACH(ev, &allevents, ev_list) {
297 if (filter != EVCNT_TYPE_ANY && filter != ev->ev_type)
298 continue;
299 if (count == KERN_EVCNT_COUNT_NONZERO && ev->ev_count == 0)
300 continue;
301
302 /*
303 * Prepare to copy. If xevs is NULL, fillevcnt will just
304 * how big the item is.
305 */
306 size_t copylen;
307 const size_t elem_size = sysctl_fillevcnt(ev, xevs, ©len);
308 needed += elem_size;
309
310 if (len < elem_size) {
311 xevs = NULL;
312 continue;
313 }
314
315 KASSERT(xevs != NULL);
316 KASSERT(xevs->evs.ev_grouplen != 0);
317 KASSERT(xevs->evs.ev_namelen != 0);
318 KASSERT(xevs->evs.ev_strings[0] != 0);
319
320 const uint32_t last_generation = evcnt_generation;
321 mutex_exit(&evcnt_lock);
322
323 /*
324 * Only copy the actual number of bytes, not the rounded
325 * number. If we did the latter we'd have to zero them
326 * first or we'd leak random kernel memory.
327 */
328 error = copyout(xevs, dp, copylen);
329
330 mutex_enter(&evcnt_lock);
331 if (error)
332 break;
333
334 if (__predict_false(last_generation != evcnt_generation)) {
335 /*
336 * This sysctl node is only for statistics.
337 * Retry; if the queue keeps changing, then
338 * bail out.
339 */
340 if (--retries == 0) {
341 error = EAGAIN;
342 break;
343 }
344 mutex_exit(&evcnt_lock);
345 goto retry;
346 }
347
348 /*
349 * Now we deal with the pointer/len since we aren't going to
350 * toss their values away.
351 */
352 dp += elem_size;
353 len -= elem_size;
354 }
355 mutex_exit(&evcnt_lock);
356
357 if (xevs0 != NULL)
358 kmem_free(xevs0, sizeof(*xevs0));
359
360 sysctl_relock();
361
362 *oldlenp = needed;
363 if (oldp == NULL)
364 *oldlenp += 1024;
365
366 return (error);
367 }
368
369
370
371 SYSCTL_SETUP(sysctl_evcnt_setup, "sysctl kern.evcnt subtree setup")
372 {
373
374 sysctl_createv(clog, 0, NULL, NULL,
375 CTLFLAG_PERMANENT,
376 CTLTYPE_STRUCT, "evcnt",
377 SYSCTL_DESCR("Kernel evcnt information"),
378 sysctl_doevcnt, 0, NULL, 0,
379 CTL_KERN, KERN_EVCNT, CTL_EOL);
380 }
381
382 #ifdef __HAVE_LEGACY_INTRCNT
383 extern u_int intrcnt[], eintrcnt[];
384 extern char intrnames[];
385 static size_t nintr;
386 struct evcnt *intr_evcnts;
387 /*
388 * Remove the following when the last intrcnt/intrnames user is cleaned up.
389 */
390 void
391 evcnt_attach_legacy_intrcnt(void)
392 {
393 size_t i;
394 const char *cp;
395
396 nintr = ((intptr_t)eintrcnt - (intptr_t)intrcnt) / sizeof(intrcnt[0]);
397 intr_evcnts = kmem_alloc(sizeof(struct evcnt) * nintr, KM_SLEEP);
398 for (cp = intrnames, i = 0; i < nintr; i++) {
399 evcnt_attach_dynamic(&intr_evcnts[i], EVCNT_TYPE_INTR,
400 NULL, "cpu", cp);
401 cp += strlen(cp) + 1;
402 }
403 }
404
405 static void
406 evcnt_update_intrcnt(void)
407 {
408 size_t i;
409
410 KASSERT(nintr > 0);
411 KASSERT(intr_evcnts != NULL);
412
413 for (i = 0; i < nintr; i++) {
414 intr_evcnts[i].ev_count = intrcnt[i];
415 }
416 }
417 #endif
418