summaryrefslogtreecommitdiff
blob: 64082d0dc6b64140636d1f10ee209e4fa9deecf3 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
/*-*- Mode: C; c-basic-offset: 8; indent-tabs-mode: nil -*-*/

/***
  This file is part of systemd.

  Copyright 2010 Lennart Poettering

  systemd is free software; you can redistribute it and/or modify it
  under the terms of the GNU General Public License as published by
  the Free Software Foundation; either version 2 of the License, or
  (at your option) any later version.

  systemd is distributed in the hope that it will be useful, but
  WITHOUT ANY WARRANTY; without even the implied warranty of
  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
  General Public License for more details.

  You should have received a copy of the GNU General Public License
  along with systemd; If not, see <http://www.gnu.org/licenses/>.
***/

#include <errno.h>
#include <assert.h>
#include <unistd.h>
#include <sys/types.h>
#include <signal.h>
#include <sys/mount.h>
#include <fcntl.h>

#include "cgroup.h"
#include "cgroup-util.h"
#include "log.h"

int cgroup_bonding_realize(CGroupBonding *b) {
        int r;

        assert(b);
        assert(b->path);
        assert(b->controller);

        if (b->realized)
                return 0;

        if ((r = cg_create(b->controller, b->path)) < 0)
                return r;

        b->realized = true;

        if (b->ours)
                cg_trim(b->controller, b->path, false);

        return 0;
}

int cgroup_bonding_realize_list(CGroupBonding *first) {
        CGroupBonding *b;
        int r;

        LIST_FOREACH(by_unit, b, first)
                if ((r = cgroup_bonding_realize(b)) < 0 && b->essential)
                        return r;

        return 0;
}

void cgroup_bonding_free(CGroupBonding *b) {
        assert(b);

        if (b->unit) {
                CGroupBonding *f;

                LIST_REMOVE(CGroupBonding, by_unit, b->unit->meta.cgroup_bondings, b);

                if (streq(b->controller, SYSTEMD_CGROUP_CONTROLLER)) {
                        assert_se(f = hashmap_get(b->unit->meta.manager->cgroup_bondings, b->path));
                        LIST_REMOVE(CGroupBonding, by_path, f, b);

                        if (f)
                                hashmap_replace(b->unit->meta.manager->cgroup_bondings, b->path, f);
                        else
                                hashmap_remove(b->unit->meta.manager->cgroup_bondings, b->path);
                }
        }

        if (b->realized && b->ours) {

                if (cgroup_bonding_is_empty(b) > 0)
                        cg_delete(b->controller, b->path);
                else
                        cg_trim(b->controller, b->path, false);
        }

        free(b->controller);
        free(b->path);
        free(b);
}

void cgroup_bonding_free_list(CGroupBonding *first) {
        CGroupBonding *b, *n;

        LIST_FOREACH_SAFE(by_unit, b, n, first)
                cgroup_bonding_free(b);
}

void cgroup_bonding_trim(CGroupBonding *b, bool delete_root) {
        assert(b);

        if (b->realized && b->ours)
                cg_trim(b->controller, b->path, delete_root);
}

void cgroup_bonding_trim_list(CGroupBonding *first, bool delete_root) {
        CGroupBonding *b;

        LIST_FOREACH(by_unit, b, first)
                cgroup_bonding_trim(b, delete_root);
}

int cgroup_bonding_install(CGroupBonding *b, pid_t pid) {
        int r;

        assert(b);
        assert(pid >= 0);

        if ((r = cg_create_and_attach(b->controller, b->path, pid)) < 0)
                return r;

        b->realized = true;
        return 0;
}

int cgroup_bonding_install_list(CGroupBonding *first, pid_t pid) {
        CGroupBonding *b;
        int r;

        LIST_FOREACH(by_unit, b, first)
                if ((r = cgroup_bonding_install(b, pid)) < 0 && b->essential)
                        return r;

        return 0;
}

int cgroup_bonding_kill(CGroupBonding *b, int sig, Set *s) {
        assert(b);
        assert(sig >= 0);

        /* Don't kill cgroups that aren't ours */
        if (!b->realized || !b->ours)
                return 0;

        return cg_kill_recursive(b->controller, b->path, sig, true, false, s);
}

int cgroup_bonding_kill_list(CGroupBonding *first, int sig, Set *s) {
        CGroupBonding *b;
        Set *allocated_set = NULL;
        int ret = -EAGAIN, r;

        if (!s)
                if (!(s = allocated_set = set_new(trivial_hash_func, trivial_compare_func)))
                        return -ENOMEM;

        LIST_FOREACH(by_unit, b, first) {
                if ((r = cgroup_bonding_kill(b, sig, s)) < 0) {
                        if (r == -EAGAIN || r == -ESRCH)
                                continue;

                        ret = r;
                        goto finish;
                }

                if (ret < 0 || r > 0)
                        ret = r;
        }

finish:
        if (allocated_set)
                set_free(allocated_set);

        return ret;
}

/* Returns 1 if the group is empty, 0 if it is not, -EAGAIN if we
 * cannot know */
int cgroup_bonding_is_empty(CGroupBonding *b) {
        int r;

        assert(b);

        if ((r = cg_is_empty_recursive(b->controller, b->path, true)) < 0)
                return r;

        /* If it is empty it is empty */
        if (r > 0)
                return 1;

        /* It's not only us using this cgroup, so we just don't know */
        return b->ours ? 0 : -EAGAIN;
}

int cgroup_bonding_is_empty_list(CGroupBonding *first) {
        CGroupBonding *b;

        LIST_FOREACH(by_unit, b, first) {
                int r;

                if ((r = cgroup_bonding_is_empty(b)) < 0) {
                        /* If this returned -EAGAIN, then we don't know if the
                         * group is empty, so let's see if another group can
                         * tell us */

                        if (r != -EAGAIN)
                                return r;
                } else
                        return r;
        }

        return -EAGAIN;
}

int manager_setup_cgroup(Manager *m) {
        char *current = NULL, *path = NULL;
        int r;
        char suffix[32];

        assert(m);

        /* 1. Determine hierarchy */
        if ((r = cg_get_by_pid(SYSTEMD_CGROUP_CONTROLLER, 0, &current)) < 0)
                goto finish;

        if (m->running_as == MANAGER_SYSTEM)
                strcpy(suffix, "/system");
        else {
                snprintf(suffix, sizeof(suffix), "/systemd-%lu", (unsigned long) getpid());
                char_array_0(suffix);
        }

        free(m->cgroup_hierarchy);
        if (endswith(current, suffix)) {
                /* We probably got reexecuted and can continue to use our root cgroup */
                m->cgroup_hierarchy = current;
                current = NULL;

        } else {
                /* We need a new root cgroup */
                m->cgroup_hierarchy = NULL;
                if (asprintf(&m->cgroup_hierarchy, "%s%s", streq(current, "/") ? "" : current, suffix) < 0) {
                        r = -ENOMEM;
                        goto finish;
                }
        }

        /* 2. Show data */
        if ((r = cg_get_path(SYSTEMD_CGROUP_CONTROLLER, m->cgroup_hierarchy, NULL, &path)) < 0)
                goto finish;

        log_debug("Using cgroup controller " SYSTEMD_CGROUP_CONTROLLER ". File system hierarchy is at %s.", path);

        /* 3. Install agent */
        if ((r = cg_install_release_agent(SYSTEMD_CGROUP_CONTROLLER, SYSTEMD_CGROUP_AGENT_PATH)) < 0)
                log_warning("Failed to install release agent, ignoring: %s", strerror(-r));
        else if (r > 0)
                log_debug("Installed release agent.");
        else
                log_debug("Release agent already installed.");

        /* 4. Realize the group */
        if ((r = cg_create_and_attach(SYSTEMD_CGROUP_CONTROLLER, m->cgroup_hierarchy, 0)) < 0) {
                log_error("Failed to create root cgroup hierarchy: %s", strerror(-r));
                goto finish;
        }

        /* 5. And pin it, so that it cannot be unmounted */
        if (m->pin_cgroupfs_fd >= 0)
                close_nointr_nofail(m->pin_cgroupfs_fd);

        if ((m->pin_cgroupfs_fd = open(path, O_RDONLY|O_CLOEXEC|O_DIRECTORY|O_NOCTTY|O_NONBLOCK)) < 0) {
                r = -errno;
                goto finish;
        }

        log_debug("Created root group.");

finish:
        free(current);
        free(path);

        return r;
}

void manager_shutdown_cgroup(Manager *m, bool delete) {
        assert(m);

        if (delete && m->cgroup_hierarchy)
                cg_delete(SYSTEMD_CGROUP_CONTROLLER, m->cgroup_hierarchy);

        if (m->pin_cgroupfs_fd >= 0) {
                close_nointr_nofail(m->pin_cgroupfs_fd);
                m->pin_cgroupfs_fd = -1;
        }

        free(m->cgroup_hierarchy);
        m->cgroup_hierarchy = NULL;
}

int cgroup_notify_empty(Manager *m, const char *group) {
        CGroupBonding *l, *b;

        assert(m);
        assert(group);

        if (!(l = hashmap_get(m->cgroup_bondings, group)))
                return 0;

        LIST_FOREACH(by_path, b, l) {
                int t;

                if (!b->unit)
                        continue;

                if ((t = cgroup_bonding_is_empty_list(b)) < 0) {

                        /* If we don't know, we don't know */
                        if (t != -EAGAIN)
                                log_warning("Failed to check whether cgroup is empty: %s", strerror(errno));

                        continue;
                }

                if (t > 0)
                        if (UNIT_VTABLE(b->unit)->cgroup_notify_empty)
                                UNIT_VTABLE(b->unit)->cgroup_notify_empty(b->unit);
        }

        return 0;
}

Unit* cgroup_unit_by_pid(Manager *m, pid_t pid) {
        CGroupBonding *l, *b;
        char *group = NULL;

        assert(m);

        if (pid <= 1)
                return NULL;

        if (cg_get_by_pid(SYSTEMD_CGROUP_CONTROLLER, pid, &group) < 0)
                return NULL;

        l = hashmap_get(m->cgroup_bondings, group);

        if (!l) {
                char *slash;

                while ((slash = strrchr(group, '/'))) {
                        if (slash == group)
                                break;

                        *slash = 0;

                        if ((l = hashmap_get(m->cgroup_bondings, group)))
                                break;
                }
        }

        free(group);

        LIST_FOREACH(by_path, b, l) {

                if (!b->unit)
                        continue;

                if (b->ours)
                        return b->unit;
        }

        return NULL;
}

CGroupBonding *cgroup_bonding_find_list(CGroupBonding *first, const char *controller) {
        CGroupBonding *b;

        assert(controller);

        LIST_FOREACH(by_unit, b, first)
                if (streq(b->controller, controller))
                        return b;

        return NULL;
}

char *cgroup_bonding_to_string(CGroupBonding *b) {
        char *r;

        assert(b);

        if (asprintf(&r, "%s:%s", b->controller, b->path) < 0)
                return NULL;

        return r;
}

pid_t cgroup_bonding_search_main_pid(CGroupBonding *b) {
        FILE *f;
        pid_t pid = 0, npid;
        int r;

        assert(b);

        if (!b->ours)
                return 0;

        if ((r = cg_enumerate_processes(b->controller, b->path, &f)) < 0)
                return 0;

        while ((r = cg_read_pid(f, &npid)) > 0)  {

                if (npid == pid)
                        continue;

                if (pid != 0) {
                        /* Dang, there's more than one PID in this
                         * group, so we don't know what process is the
                         * main process. */
                        pid = 0;
                        break;
                }

                pid = npid;
        }

        fclose(f);

        return pid;
}

pid_t cgroup_bonding_search_main_pid_list(CGroupBonding *first) {
        CGroupBonding *b;
        pid_t pid;

        /* Try to find a main pid from this cgroup, but checking if
         * there's only one PID in the cgroup and returning it. Later
         * on we might want to add additional, smarter heuristics
         * here. */

        LIST_FOREACH(by_unit, b, first)
                if ((pid = cgroup_bonding_search_main_pid(b)) != 0)
                        return pid;

        return 0;

}