src/journal/journald.c

   1 /*-*- Mode: C; c-basic-offset: 8; indent-tabs-mode: nil -*-*/
   2
   3 /***
   4   This file is part of systemd.
   5
   6   Copyright 2011 Lennart Poettering
   7
   8   systemd is free software; you can redistribute it and/or modify it
   9   under the terms of the GNU General Public License as published by
  10   the Free Software Foundation; either version 2 of the License, or
  11   (at your option) any later version.
  12
  13   systemd is distributed in the hope that it will be useful, but
  14   WITHOUT ANY WARRANTY; without even the implied warranty of
  15   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
  16   General Public License for more details.
  17
  18   You should have received a copy of the GNU General Public License
  19   along with systemd; If not, see <http://www.gnu.org/licenses/>.
  20 ***/
  21
  22 #include <sys/epoll.h>
  23 #include <sys/socket.h>
  24 #include <errno.h>
  25 #include <sys/signalfd.h>
  26 #include <unistd.h>
  27 #include <fcntl.h>
  28 #include <sys/acl.h>
  29 #include <acl/libacl.h>
  30 #include <stddef.h>
  31 #include <sys/ioctl.h>
  32 #include <linux/sockios.h>
  33 #include <sys/statvfs.h>
  34
  35 #include "hashmap.h"
  36 #include "journal-file.h"
  37 #include "sd-daemon.h"
  38 #include "socket-util.h"
  39 #include "acl-util.h"
  40 #include "cgroup-util.h"
  41 #include "list.h"
  42 #include "journal-rate-limit.h"
  43 #include "sd-journal.h"
  44 #include "journal-internal.h"
  45
  46 #define USER_JOURNALS_MAX 1024
  47 #define STDOUT_STREAMS_MAX 4096
  48
  49 #define DEFAULT_RATE_LIMIT_INTERVAL (10*USEC_PER_SEC)
  50 #define DEFAULT_RATE_LIMIT_BURST 200
  51
  52 #define RECHECK_AVAILABLE_SPACE_USEC (30*USEC_PER_SEC)
  53
  54 #define RECHECK_VAR_AVAILABLE_USEC (30*USEC_PER_SEC)
  55
  56 #define SYSLOG_TIMEOUT_USEC (5*USEC_PER_SEC)
  57
  58 typedef struct StdoutStream StdoutStream;
  59
  60 typedef struct Server {
  61         int epoll_fd;
  62         int signal_fd;
  63         int syslog_fd;
  64         int native_fd;
  65         int stdout_fd;
  66
  67         JournalFile *runtime_journal;
  68         JournalFile *system_journal;
  69         Hashmap *user_journals;
  70
  71         uint64_t seqnum;
  72
  73         char *buffer;
  74         size_t buffer_size;
  75
  76         JournalRateLimit *rate_limit;
  77
  78         JournalMetrics metrics;
  79         uint64_t max_use;
  80         bool compress;
  81
  82         uint64_t cached_available_space;
  83         usec_t cached_available_space_timestamp;
  84
  85         uint64_t var_available_timestamp;
  86
  87         LIST_HEAD(StdoutStream, stdout_streams);
  88         unsigned n_stdout_streams;
  89 } Server;
  90
  91 typedef enum StdoutStreamState {
  92         STDOUT_STREAM_TAG,
  93         STDOUT_STREAM_PRIORITY,
  94         STDOUT_STREAM_PRIORITY_PREFIX,
  95         STDOUT_STREAM_TEE_CONSOLE,
  96         STDOUT_STREAM_RUNNING
  97 } StdoutStreamState;
  98
  99 struct StdoutStream {
 100         Server *server;
 101         StdoutStreamState state;
 102
 103         int fd;
 104
 105         struct ucred ucred;
 106
 107         char *tag;
 108         int priority;
 109         bool priority_prefix:1;
 110         bool tee_console:1;
 111
 112         char buffer[LINE_MAX+1];
 113         size_t length;
 114
 115         LIST_FIELDS(StdoutStream, stdout_stream);
 116 };
 117
 118 static int server_flush_to_var(Server *s);
 119
 120 static uint64_t available_space(Server *s) {
 121         char ids[33];
 122         sd_id128_t machine;
 123         char *p;
 124         const char *f;
 125         struct statvfs ss;
 126         uint64_t sum = 0, avail = 0, ss_avail = 0;
 127         int r;
 128         DIR *d;
 129         usec_t ts = now(CLOCK_MONOTONIC);
 130
 131         if (s->cached_available_space_timestamp + RECHECK_AVAILABLE_SPACE_USEC > ts)
 132                 return s->cached_available_space;
 133
 134         r = sd_id128_get_machine(&machine);
 135         if (r < 0)
 136                 return 0;
 137
 138         if (s->system_journal)
 139                 f = "/var/log/journal/";
 140         else
 141                 f = "/run/log/journal/";
 142
 143         p = strappend(f, sd_id128_to_string(machine, ids));
 144         if (!p)
 145                 return 0;
 146
 147         d = opendir(p);
 148         free(p);
 149
 150         if (!d)
 151                 return 0;
 152
 153         if (fstatvfs(dirfd(d), &ss) < 0)
 154                 goto finish;
 155
 156         for (;;) {
 157                 struct stat st;
 158                 struct dirent buf, *de;
 159                 int k;
 160
 161                 k = readdir_r(d, &buf, &de);
 162                 if (k != 0) {
 163                         r = -k;
 164                         goto finish;
 165                 }
 166
 167                 if (!de)
 168                         break;
 169
 170                 if (!dirent_is_file_with_suffix(de, ".journal"))
 171                         continue;
 172
 173                 if (fstatat(dirfd(d), de->d_name, &st, AT_SYMLINK_NOFOLLOW) < 0)
 174                         continue;
 175
 176                 sum += (uint64_t) st.st_blocks * (uint64_t) st.st_blksize;
 177         }
 178
 179         avail = sum >= s->max_use ? 0 : s->max_use - sum;
 180
 181         ss_avail = ss.f_bsize * ss.f_bavail;
 182
 183         ss_avail = ss_avail < s->metrics.keep_free ? 0 : ss_avail - s->metrics.keep_free;
 184
 185         if (ss_avail < avail)
 186                 avail = ss_avail;
 187
 188         s->cached_available_space = avail;
 189         s->cached_available_space_timestamp = ts;
 190
 191 finish:
 192         closedir(d);
 193
 194         return avail;
 195 }
 196
 197 static void fix_perms(JournalFile *f, uid_t uid) {
 198         acl_t acl;
 199         acl_entry_t entry;
 200         acl_permset_t permset;
 201         int r;
 202
 203         assert(f);
 204
 205         r = fchmod_and_fchown(f->fd, 0640, 0, 0);
 206         if (r < 0)
 207                 log_warning("Failed to fix access mode/rights on %s, ignoring: %s", f->path, strerror(-r));
 208
 209         if (uid <= 0)
 210                 return;
 211
 212         acl = acl_get_fd(f->fd);
 213         if (!acl) {
 214                 log_warning("Failed to read ACL on %s, ignoring: %m", f->path);
 215                 return;
 216         }
 217
 218         r = acl_find_uid(acl, uid, &entry);
 219         if (r <= 0) {
 220
 221                 if (acl_create_entry(&acl, &entry) < 0 ||
 222                     acl_set_tag_type(entry, ACL_USER) < 0 ||
 223                     acl_set_qualifier(entry, &uid) < 0) {
 224                         log_warning("Failed to patch ACL on %s, ignoring: %m", f->path);
 225                         goto finish;
 226                 }
 227         }
 228
 229         if (acl_get_permset(entry, &permset) < 0 ||
 230             acl_add_perm(permset, ACL_READ) < 0 ||
 231             acl_calc_mask(&acl) < 0) {
 232                 log_warning("Failed to patch ACL on %s, ignoring: %m", f->path);
 233                 goto finish;
 234         }
 235
 236         if (acl_set_fd(f->fd, acl) < 0)
 237                 log_warning("Failed to set ACL on %s, ignoring: %m", f->path);
 238
 239 finish:
 240         acl_free(acl);
 241 }
 242
 243 static JournalFile* find_journal(Server *s, uid_t uid) {
 244         char *p;
 245         int r;
 246         JournalFile *f;
 247         char ids[33];
 248         sd_id128_t machine;
 249
 250         assert(s);
 251
 252         /* We split up user logs only on /var, not on /run. If the
 253          * runtime file is open, we write to it exclusively, in order
 254          * to guarantee proper order as soon as we flush /run to
 255          * /var and close the runtime file. */
 256
 257         if (s->runtime_journal)
 258                 return s->runtime_journal;
 259
 260         if (uid <= 0)
 261                 return s->system_journal;
 262
 263         r = sd_id128_get_machine(&machine);
 264         if (r < 0)
 265                 return s->system_journal;
 266
 267         f = hashmap_get(s->user_journals, UINT32_TO_PTR(uid));
 268         if (f)
 269                 return f;
 270
 271         if (asprintf(&p, "/var/log/journal/%s/user-%lu.journal", sd_id128_to_string(machine, ids), (unsigned long) uid) < 0)
 272                 return s->system_journal;
 273
 274         while (hashmap_size(s->user_journals) >= USER_JOURNALS_MAX) {
 275                 /* Too many open? Then let's close one */
 276                 f = hashmap_steal_first(s->user_journals);
 277                 assert(f);
 278                 journal_file_close(f);
 279         }
 280
 281         r = journal_file_open(p, O_RDWR|O_CREAT, 0640, s->system_journal, &f);
 282         free(p);
 283
 284         if (r < 0)
 285                 return s->system_journal;
 286
 287         fix_perms(f, uid);
 288         f->metrics = s->metrics;
 289         f->compress = s->compress;
 290
 291         r = hashmap_put(s->user_journals, UINT32_TO_PTR(uid), f);
 292         if (r < 0) {
 293                 journal_file_close(f);
 294                 return s->system_journal;
 295         }
 296
 297         return f;
 298 }
 299
 300 static void server_vacuum(Server *s) {
 301         Iterator i;
 302         void *k;
 303         char *p;
 304         char ids[33];
 305         sd_id128_t machine;
 306         int r;
 307         JournalFile *f;
 308
 309         log_info("Rotating...");
 310
 311         if (s->runtime_journal) {
 312                 r = journal_file_rotate(&s->runtime_journal);
 313                 if (r < 0)
 314                         log_error("Failed to rotate %s: %s", s->runtime_journal->path, strerror(-r));
 315         }
 316
 317         if (s->system_journal) {
 318                 r = journal_file_rotate(&s->system_journal);
 319                 if (r < 0)
 320                         log_error("Failed to rotate %s: %s", s->system_journal->path, strerror(-r));
 321         }
 322
 323         HASHMAP_FOREACH_KEY(f, k, s->user_journals, i) {
 324                 r = journal_file_rotate(&f);
 325                 if (r < 0)
 326                         log_error("Failed to rotate %s: %s", f->path, strerror(-r));
 327                 else
 328                         hashmap_replace(s->user_journals, k, f);
 329         }
 330
 331         log_info("Vacuuming...");
 332
 333         r = sd_id128_get_machine(&machine);
 334         if (r < 0) {
 335                 log_error("Failed to get machine ID: %s", strerror(-r));
 336                 return;
 337         }
 338
 339         if (asprintf(&p, "/var/log/journal/%s", sd_id128_to_string(machine, ids)) < 0) {
 340                 log_error("Out of memory.");
 341                 return;
 342         }
 343
 344         r = journal_directory_vacuum(p, s->max_use, s->metrics.keep_free);
 345         if (r < 0 && r != -ENOENT)
 346                 log_error("Failed to vacuum %s: %s", p, strerror(-r));
 347         free(p);
 348
 349         if (asprintf(&p, "/run/log/journal/%s", ids) < 0) {
 350                 log_error("Out of memory.");
 351                 return;
 352         }
 353
 354         r = journal_directory_vacuum(p, s->max_use, s->metrics.keep_free);
 355         if (r < 0 && r != -ENOENT)
 356                 log_error("Failed to vacuum %s: %s", p, strerror(-r));
 357         free(p);
 358
 359         s->cached_available_space_timestamp = 0;
 360 }
 361
 362 static char *shortened_cgroup_path(pid_t pid) {
 363         int r;
 364         char *process_path, *init_path, *path;
 365
 366         assert(pid > 0);
 367
 368         r = cg_get_by_pid(SYSTEMD_CGROUP_CONTROLLER, pid, &process_path);
 369         if (r < 0)
 370                 return NULL;
 371
 372         r = cg_get_by_pid(SYSTEMD_CGROUP_CONTROLLER, 1, &init_path);
 373         if (r < 0) {
 374                 free(process_path);
 375                 return NULL;
 376         }
 377
 378         if (streq(init_path, "/"))
 379                 init_path[0] = 0;
 380
 381         if (startswith(process_path, init_path)) {
 382                 char *p;
 383
 384                 p = strdup(process_path + strlen(init_path));
 385                 if (!p) {
 386                         free(process_path);
 387                         free(init_path);
 388                         return NULL;
 389                 }
 390                 path = p;
 391         } else {
 392                 path = process_path;
 393                 process_path = NULL;
 394         }
 395
 396         free(process_path);
 397         free(init_path);
 398
 399         return path;
 400 }
 401
 402 static void dispatch_message_real(Server *s,
 403                              struct iovec *iovec, unsigned n, unsigned m,
 404                              struct ucred *ucred,
 405                              struct timeval *tv) {
 406
 407         char *pid = NULL, *uid = NULL, *gid = NULL,
 408                 *source_time = NULL, *boot_id = NULL, *machine_id = NULL,
 409                 *comm = NULL, *cmdline = NULL, *hostname = NULL,
 410                 *audit_session = NULL, *audit_loginuid = NULL,
 411                 *exe = NULL, *cgroup = NULL;
 412
 413         char idbuf[33];
 414         sd_id128_t id;
 415         int r;
 416         char *t;
 417         uid_t loginuid = 0, realuid = 0;
 418         JournalFile *f;
 419         bool vacuumed = false;
 420
 421         assert(s);
 422         assert(iovec);
 423         assert(n > 0);
 424         assert(n + 13 <= m);
 425
 426         if (ucred) {
 427                 uint32_t session;
 428                 char *path;
 429
 430                 realuid = ucred->uid;
 431
 432                 if (asprintf(&pid, "_PID=%lu", (unsigned long) ucred->pid) >= 0)
 433                         IOVEC_SET_STRING(iovec[n++], pid);
 434
 435                 if (asprintf(&uid, "_UID=%lu", (unsigned long) ucred->uid) >= 0)
 436                         IOVEC_SET_STRING(iovec[n++], uid);
 437
 438                 if (asprintf(&gid, "_GID=%lu", (unsigned long) ucred->gid) >= 0)
 439                         IOVEC_SET_STRING(iovec[n++], gid);
 440
 441                 r = get_process_comm(ucred->pid, &t);
 442                 if (r >= 0) {
 443                         comm = strappend("_COMM=", t);
 444                         if (comm)
 445                                 IOVEC_SET_STRING(iovec[n++], comm);
 446                         free(t);
 447                 }
 448
 449                 r = get_process_exe(ucred->pid, &t);
 450                 if (r >= 0) {
 451                         exe = strappend("_EXE=", t);
 452                         if (comm)
 453                                 IOVEC_SET_STRING(iovec[n++], exe);
 454                         free(t);
 455                 }
 456
 457                 r = get_process_cmdline(ucred->pid, LINE_MAX, false, &t);
 458                 if (r >= 0) {
 459                         cmdline = strappend("_CMDLINE=", t);
 460                         if (cmdline)
 461                                 IOVEC_SET_STRING(iovec[n++], cmdline);
 462                         free(t);
 463                 }
 464
 465                 r = audit_session_from_pid(ucred->pid, &session);
 466                 if (r >= 0)
 467                         if (asprintf(&audit_session, "_AUDIT_SESSION=%lu", (unsigned long) session) >= 0)
 468                                 IOVEC_SET_STRING(iovec[n++], audit_session);
 469
 470                 r = audit_loginuid_from_pid(ucred->pid, &loginuid);
 471                 if (r >= 0)
 472                         if (asprintf(&audit_loginuid, "_AUDIT_LOGINUID=%lu", (unsigned long) loginuid) >= 0)
 473                                 IOVEC_SET_STRING(iovec[n++], audit_loginuid);
 474
 475                 path = shortened_cgroup_path(ucred->pid);
 476                 if (path) {
 477                         cgroup = strappend("_SYSTEMD_CGROUP=", path);
 478                         if (cgroup)
 479                                 IOVEC_SET_STRING(iovec[n++], cgroup);
 480
 481                         free(path);
 482                 }
 483         }
 484
 485         if (tv) {
 486                 if (asprintf(&source_time, "_SOURCE_REALTIME_TIMESTAMP=%llu",
 487                              (unsigned long long) timeval_load(tv)) >= 0)
 488                         IOVEC_SET_STRING(iovec[n++], source_time);
 489         }
 490
 491         /* Note that strictly speaking storing the boot id here is
 492          * redundant since the entry includes this in-line
 493          * anyway. However, we need this indexed, too. */
 494         r = sd_id128_get_boot(&id);
 495         if (r >= 0)
 496                 if (asprintf(&boot_id, "_BOOT_ID=%s", sd_id128_to_string(id, idbuf)) >= 0)
 497                         IOVEC_SET_STRING(iovec[n++], boot_id);
 498
 499         r = sd_id128_get_machine(&id);
 500         if (r >= 0)
 501                 if (asprintf(&machine_id, "_MACHINE_ID=%s", sd_id128_to_string(id, idbuf)) >= 0)
 502                         IOVEC_SET_STRING(iovec[n++], machine_id);
 503
 504         t = gethostname_malloc();
 505         if (t) {
 506                 hostname = strappend("_HOSTNAME=", t);
 507                 if (hostname)
 508                         IOVEC_SET_STRING(iovec[n++], hostname);
 509                 free(t);
 510         }
 511
 512         assert(n <= m);
 513
 514         server_flush_to_var(s);
 515
 516 retry:
 517         f = find_journal(s, realuid == 0 ? 0 : loginuid);
 518         if (!f)
 519                 log_warning("Dropping message, as we can't find a place to store the data.");
 520         else {
 521                 r = journal_file_append_entry(f, NULL, iovec, n, &s->seqnum, NULL, NULL);
 522
 523                 if (r == -E2BIG && !vacuumed) {
 524                         log_info("Allocation limit reached.");
 525
 526                         server_vacuum(s);
 527                         vacuumed = true;
 528
 529                         log_info("Retrying write.");
 530                         goto retry;
 531                 }
 532
 533                 if (r < 0)
 534                         log_error("Failed to write entry, ignoring: %s", strerror(-r));
 535         }
 536
 537         free(pid);
 538         free(uid);
 539         free(gid);
 540         free(comm);
 541         free(exe);
 542         free(cmdline);
 543         free(source_time);
 544         free(boot_id);
 545         free(machine_id);
 546         free(hostname);
 547         free(audit_session);
 548         free(audit_loginuid);
 549         free(cgroup);
 550 }
 551
 552 static void dispatch_message(Server *s,
 553                              struct iovec *iovec, unsigned n, unsigned m,
 554                              struct ucred *ucred,
 555                              struct timeval *tv,
 556                              int priority) {
 557         int rl;
 558         char *path = NULL, *c;
 559
 560         assert(s);
 561         assert(iovec || n == 0);
 562
 563         if (n == 0)
 564                 return;
 565
 566         if (!ucred)
 567                 goto finish;
 568
 569         path = shortened_cgroup_path(ucred->pid);
 570         if (!path)
 571                 goto finish;
 572
 573         /* example: /user/lennart/3/foobar
 574          *          /system/dbus.service/foobar
 575          *
 576          * So let's cut of everything past the third /, since that is
 577          * wher user directories start */
 578
 579         c = strchr(path, '/');
 580         if (c) {
 581                 c = strchr(c+1, '/');
 582                 if (c) {
 583                         c = strchr(c+1, '/');
 584                         if (c)
 585                                 *c = 0;
 586                 }
 587         }
 588
 589         rl = journal_rate_limit_test(s->rate_limit, path, priority, available_space(s));
 590
 591         if (rl == 0) {
 592                 free(path);
 593                 return;
 594         }
 595
 596         if (rl > 1) {
 597                 int j = 0;
 598                 char suppress_message[LINE_MAX];
 599                 struct iovec suppress_iovec[15];
 600
 601                 /* Write a suppression message if we suppressed something */
 602
 603                 snprintf(suppress_message, sizeof(suppress_message), "MESSAGE=Suppressed %u messages from %s", rl - 1, path);
 604                 char_array_0(suppress_message);
 605
 606                 IOVEC_SET_STRING(suppress_iovec[j++], "PRIORITY=5");
 607                 IOVEC_SET_STRING(suppress_iovec[j++], suppress_message);
 608
 609                 dispatch_message_real(s, suppress_iovec, j, ELEMENTSOF(suppress_iovec), NULL, NULL);
 610         }
 611
 612         free(path);
 613
 614 finish:
 615         dispatch_message_real(s, iovec, n, m, ucred, tv);
 616 }
 617
 618 static void process_syslog_message(Server *s, const char *buf, struct ucred *ucred, struct timeval *tv) {
 619         char *message = NULL, *syslog_priority = NULL, *syslog_facility = NULL;
 620         struct iovec iovec[16];
 621         unsigned n = 0;
 622         int priority = LOG_USER | LOG_INFO;
 623
 624         assert(s);
 625         assert(buf);
 626
 627         parse_syslog_priority((char**) &buf, &priority);
 628         skip_syslog_date((char**) &buf);
 629
 630         if (asprintf(&syslog_priority, "PRIORITY=%i", priority & LOG_PRIMASK) >= 0)
 631                 IOVEC_SET_STRING(iovec[n++], syslog_priority);
 632
 633         if (asprintf(&syslog_facility, "SYSLOG_FACILITY=%i", LOG_FAC(priority)) >= 0)
 634                 IOVEC_SET_STRING(iovec[n++], syslog_facility);
 635
 636         message = strappend("MESSAGE=", buf);
 637         if (message)
 638                 IOVEC_SET_STRING(iovec[n++], message);
 639
 640         dispatch_message(s, iovec, n, ELEMENTSOF(iovec), ucred, tv, priority & LOG_PRIMASK);
 641
 642         free(message);
 643         free(syslog_facility);
 644         free(syslog_priority);
 645 }
 646
 647 static bool valid_user_field(const char *p, size_t l) {
 648         const char *a;
 649
 650         /* We kinda enforce POSIX syntax recommendations for
 651            environment variables here, but make a couple of additional
 652            requirements.
 653
 654            http://pubs.opengroup.org/onlinepubs/000095399/basedefs/xbd_chap08.html */
 655
 656         /* No empty field names */
 657         if (l <= 0)
 658                 return false;
 659
 660         /* Don't allow names longer than 64 chars */
 661         if (l > 64)
 662                 return false;
 663
 664         /* Variables starting with an underscore are protected */
 665         if (p[0] == '_')
 666                 return false;
 667
 668         /* Don't allow digits as first character */
 669         if (p[0] >= '0' && p[0] <= '9')
 670                 return false;
 671
 672         /* Only allow A-Z0-9 and '_' */
 673         for (a = p; a < p + l; a++)
 674                 if (!((*a >= 'A' && *a <= 'Z') ||
 675                       (*a >= '0' && *a <= '9') ||
 676                       *a == '_'))
 677                         return false;
 678
 679         return true;
 680 }
 681
 682 static void process_native_message(Server *s, const void *buffer, size_t buffer_size, struct ucred *ucred, struct timeval *tv) {
 683         struct iovec *iovec = NULL;
 684         unsigned n = 0, m = 0, j;
 685         const char *p;
 686         size_t remaining;
 687         int priority = LOG_INFO;
 688
 689         assert(s);
 690         assert(buffer || n == 0);
 691
 692         p = buffer;
 693         remaining = buffer_size;
 694
 695         while (remaining > 0) {
 696                 const char *e, *q;
 697
 698                 e = memchr(p, '\n', remaining);
 699
 700                 if (!e) {
 701                         /* Trailing noise, let's ignore it, and flush what we collected */
 702                         log_debug("Received message with trailing noise, ignoring.");
 703                         break;
 704                 }
 705
 706                 if (e == p) {
 707                         /* Entry separator */
 708                         dispatch_message(s, iovec, n, m, ucred, tv, priority);
 709                         n = 0;
 710                         priority = LOG_INFO;
 711
 712                         p++;
 713                         remaining--;
 714                         continue;
 715                 }
 716
 717                 if (*p == '.' || *p == '#') {
 718                         /* Ignore control commands for now, and
 719                          * comments too. */
 720                         remaining -= (e - p) + 1;
 721                         p = e + 1;
 722                         continue;
 723                 }
 724
 725                 /* A property follows */
 726
 727                 if (n+13 >= m) {
 728                         struct iovec *c;
 729                         unsigned u;
 730
 731                         u = MAX((n+13U) * 2U, 4U);
 732                         c = realloc(iovec, u * sizeof(struct iovec));
 733                         if (!c) {
 734                                 log_error("Out of memory");
 735                                 break;
 736                         }
 737
 738                         iovec = c;
 739                         m = u;
 740                 }
 741
 742                 q = memchr(p, '=', e - p);
 743                 if (q) {
 744                         if (valid_user_field(p, q - p)) {
 745                                 /* If the field name starts with an
 746                                  * underscore, skip the variable,
 747                                  * since that indidates a trusted
 748                                  * field */
 749                                 iovec[n].iov_base = (char*) p;
 750                                 iovec[n].iov_len = e - p;
 751                                 n++;
 752
 753                                 /* We need to determine the priority
 754                                  * of this entry for the rate limiting
 755                                  * logic */
 756                                 if (e - p == 10 &&
 757                                     memcmp(p, "PRIORITY=", 10) == 0 &&
 758                                     p[10] >= '0' &&
 759                                     p[10] <= '9')
 760                                         priority = p[10] - '0';
 761                         }
 762
 763                         remaining -= (e - p) + 1;
 764                         p = e + 1;
 765                         continue;
 766                 } else {
 767                         uint64_t l;
 768                         char *k;
 769
 770                         if (remaining < e - p + 1 + sizeof(uint64_t) + 1) {
 771                                 log_debug("Failed to parse message, ignoring.");
 772                                 break;
 773                         }
 774
 775                         memcpy(&l, e + 1, sizeof(uint64_t));
 776                         l = le64toh(l);
 777
 778                         if (remaining < e - p + 1 + sizeof(uint64_t) + l + 1 ||
 779                             e[1+sizeof(uint64_t)+l] != '\n') {
 780                                 log_debug("Failed to parse message, ignoring.");
 781                                 break;
 782                         }
 783
 784                         k = malloc((e - p) + 1 + l);
 785                         if (!k) {
 786                                 log_error("Out of memory");
 787                                 break;
 788                         }
 789
 790                         memcpy(k, p, e - p);
 791                         k[e - p] = '=';
 792                         memcpy(k + (e - p) + 1, e + 1 + sizeof(uint64_t), l);
 793
 794                         if (valid_user_field(p, e - p)) {
 795                                 iovec[n].iov_base = k;
 796                                 iovec[n].iov_len = (e - p) + 1 + l;
 797                                 n++;
 798                         } else
 799                                 free(k);
 800
 801                         remaining -= (e - p) + 1 + sizeof(uint64_t) + l + 1;
 802                         p = e + 1 + sizeof(uint64_t) + l + 1;
 803                 }
 804         }
 805
 806         dispatch_message(s, iovec, n, m, ucred, tv, priority);
 807
 808         for (j = 0; j < n; j++)
 809                 if (iovec[j].iov_base < buffer ||
 810                     (const uint8_t*) iovec[j].iov_base >= (const uint8_t*) buffer + buffer_size)
 811                         free(iovec[j].iov_base);
 812 }
 813
 814 static int stdout_stream_log(StdoutStream *s, const char *p, size_t l) {
 815         struct iovec iovec[15];
 816         char *message = NULL, *syslog_priority = NULL;
 817         unsigned n = 0;
 818         size_t tag_len;
 819         int priority;
 820
 821         assert(s);
 822         assert(p);
 823
 824         priority = s->priority;
 825
 826         if (s->priority_prefix &&
 827             l > 3 &&
 828             p[0] == '<' &&
 829             p[1] >= '0' && p[1] <= '7' &&
 830             p[2] == '>') {
 831
 832                 priority = p[1] - '0';
 833                 p += 3;
 834                 l -= 3;
 835         }
 836
 837         if (l <= 0)
 838                 return 0;
 839
 840         if (asprintf(&syslog_priority, "PRIORITY=%i", priority) >= 0)
 841                 IOVEC_SET_STRING(iovec[n++], syslog_priority);
 842
 843         tag_len = s->tag ? strlen(s->tag) + 2: 0;
 844         message = malloc(8 + tag_len + l);
 845         if (message) {
 846                 memcpy(message, "MESSAGE=", 8);
 847
 848                 if (s->tag) {
 849                         memcpy(message+8, s->tag, tag_len-2);
 850                         memcpy(message+8+tag_len-2, ": ", 2);
 851                 }
 852
 853                 memcpy(message+8+tag_len, p, l);
 854                 iovec[n].iov_base = message;
 855                 iovec[n].iov_len = 8+tag_len+l;
 856                 n++;
 857         }
 858
 859         dispatch_message(s->server, iovec, n, ELEMENTSOF(iovec), &s->ucred, NULL, priority);
 860
 861         if (s->tee_console) {
 862                 int console;
 863
 864                 console = open_terminal("/dev/console", O_WRONLY|O_NOCTTY|O_CLOEXEC);
 865                 if (console >= 0) {
 866                         n = 0;
 867                         if (s->tag) {
 868                                 IOVEC_SET_STRING(iovec[n++], s->tag);
 869                                 IOVEC_SET_STRING(iovec[n++], ": ");
 870                         }
 871
 872                         iovec[n].iov_base = (void*) p;
 873                         iovec[n].iov_len = l;
 874                         n++;
 875
 876                         IOVEC_SET_STRING(iovec[n++], (char*) "\n");
 877
 878                         writev(console, iovec, n);
 879                 }
 880         }
 881
 882         free(message);
 883         free(syslog_priority);
 884
 885         return 0;
 886 }
 887
 888 static int stdout_stream_line(StdoutStream *s, const char *p, size_t l) {
 889         assert(s);
 890         assert(p);
 891
 892         while (l > 0 && strchr(WHITESPACE, *p)) {
 893                 l--;
 894                 p++;
 895         }
 896
 897         while (l > 0 && strchr(WHITESPACE, *(p+l-1)))
 898                 l--;
 899
 900         switch (s->state) {
 901
 902         case STDOUT_STREAM_TAG:
 903
 904                 if (l > 0) {
 905                         s->tag = strndup(p, l);
 906                         if (!s->tag) {
 907                                 log_error("Out of memory");
 908                                 return -EINVAL;
 909                         }
 910                 }
 911
 912                 s->state = STDOUT_STREAM_PRIORITY;
 913                 return 0;
 914
 915         case STDOUT_STREAM_PRIORITY:
 916                 if (l != 1 || *p < '0' || *p > '7') {
 917                         log_warning("Failed to parse log priority line.");
 918                         return -EINVAL;
 919                 }
 920
 921                 s->priority = *p - '0';
 922                 s->state = STDOUT_STREAM_PRIORITY_PREFIX;
 923                 return 0;
 924
 925         case STDOUT_STREAM_PRIORITY_PREFIX:
 926                 if (l != 1 || *p < '0' || *p > '1') {
 927                         log_warning("Failed to parse priority prefix line.");
 928                         return -EINVAL;
 929                 }
 930
 931                 s->priority_prefix = *p - '0';
 932                 s->state = STDOUT_STREAM_TEE_CONSOLE;
 933                 return 0;
 934
 935         case STDOUT_STREAM_TEE_CONSOLE:
 936                 if (l != 1 || *p < '0' || *p > '1') {
 937                         log_warning("Failed to parse tee to console line.");
 938                         return -EINVAL;
 939                 }
 940
 941                 s->tee_console = *p - '0';
 942                 s->state = STDOUT_STREAM_RUNNING;
 943                 return 0;
 944
 945         case STDOUT_STREAM_RUNNING:
 946                 return stdout_stream_log(s, p, l);
 947         }
 948
 949         assert_not_reached("Unknown stream state");
 950 }
 951
 952 static int stdout_stream_scan(StdoutStream *s, bool force_flush) {
 953         char *p;
 954         size_t remaining;
 955         int r;
 956
 957         assert(s);
 958
 959         p = s->buffer;
 960         remaining = s->length;
 961         for (;;) {
 962                 char *end;
 963                 size_t skip;
 964
 965                 end = memchr(p, '\n', remaining);
 966                 if (!end) {
 967                         if (remaining >= LINE_MAX) {
 968                                 end = p + LINE_MAX;
 969                                 skip = LINE_MAX;
 970                         } else
 971                                 break;
 972                 } else
 973                         skip = end - p + 1;
 974
 975                 r = stdout_stream_line(s, p, end - p);
 976                 if (r < 0)
 977                         return r;
 978
 979                 remaining -= skip;
 980                 p += skip;
 981         }
 982
 983         if (force_flush && remaining > 0) {
 984                 r = stdout_stream_line(s, p, remaining);
 985                 if (r < 0)
 986                         return r;
 987
 988                 p += remaining;
 989                 remaining = 0;
 990         }
 991
 992         if (p > s->buffer) {
 993                 memmove(s->buffer, p, remaining);
 994                 s->length = remaining;
 995         }
 996
 997         return 0;
 998 }
 999
1000 static int stdout_stream_process(StdoutStream *s) {
1001         ssize_t l;
1002         int r;
1003
1004         assert(s);
1005
1006         l = read(s->fd, s->buffer+s->length, sizeof(s->buffer)-1-s->length);
1007         if (l < 0) {
1008
1009                 if (errno == EAGAIN)
1010                         return 0;
1011
1012                 log_warning("Failed to read from stream: %m");
1013                 return -errno;
1014         }
1015
1016         if (l == 0) {
1017                 r = stdout_stream_scan(s, true);
1018                 if (r < 0)
1019                         return r;
1020
1021                 return 0;
1022         }
1023
1024         s->length += l;
1025         r = stdout_stream_scan(s, false);
1026         if (r < 0)
1027                 return r;
1028
1029         return 1;
1030
1031 }
1032
1033 static void stdout_stream_free(StdoutStream *s) {
1034         assert(s);
1035
1036         if (s->server) {
1037                 assert(s->server->n_stdout_streams > 0);
1038                 s->server->n_stdout_streams --;
1039                 LIST_REMOVE(StdoutStream, stdout_stream, s->server->stdout_streams, s);
1040         }
1041
1042         if (s->fd >= 0) {
1043                 if (s->server)
1044                         epoll_ctl(s->server->epoll_fd, EPOLL_CTL_DEL, s->fd, NULL);
1045
1046                 close_nointr_nofail(s->fd);
1047         }
1048
1049         free(s->tag);
1050         free(s);
1051 }
1052
1053 static int stdout_stream_new(Server *s) {
1054         StdoutStream *stream;
1055         int fd, r;
1056         socklen_t len;
1057         struct epoll_event ev;
1058
1059         assert(s);
1060
1061         fd = accept4(s->stdout_fd, NULL, NULL, SOCK_NONBLOCK|SOCK_CLOEXEC);
1062         if (fd < 0) {
1063                 if (errno == EAGAIN)
1064                         return 0;
1065
1066                 log_error("Failed to accept stdout connection: %m");
1067                 return -errno;
1068         }
1069
1070         if (s->n_stdout_streams >= STDOUT_STREAMS_MAX) {
1071                 log_warning("Too many stdout streams, refusing connection.");
1072                 close_nointr_nofail(fd);
1073                 return 0;
1074         }
1075
1076         stream = new0(StdoutStream, 1);
1077         if (!stream) {
1078                 log_error("Out of memory.");
1079                 close_nointr_nofail(fd);
1080                 return -ENOMEM;
1081         }
1082
1083         stream->fd = fd;
1084
1085         len = sizeof(stream->ucred);
1086         if (getsockopt(fd, SOL_SOCKET, SO_PEERCRED, &stream->ucred, &len) < 0) {
1087                 log_error("Failed to determine peer credentials: %m");
1088                 r = -errno;
1089                 goto fail;
1090         }
1091
1092         if (shutdown(fd, SHUT_WR) < 0) {
1093                 log_error("Failed to shutdown writing side of socket: %m");
1094                 r = -errno;
1095                 goto fail;
1096         }
1097
1098         zero(ev);
1099         ev.data.ptr = stream;
1100         ev.events = EPOLLIN;
1101         if (epoll_ctl(s->epoll_fd, EPOLL_CTL_ADD, fd, &ev) < 0) {
1102                 log_error("Failed to add stream to event loop: %m");
1103                 r = -errno;
1104                 goto fail;
1105         }
1106
1107         stream->server = s;
1108         LIST_PREPEND(StdoutStream, stdout_stream, s->stdout_streams, stream);
1109         s->n_stdout_streams ++;
1110
1111         return 0;
1112
1113 fail:
1114         stdout_stream_free(stream);
1115         return r;
1116 }
1117
1118 static int system_journal_open(Server *s) {
1119         int r;
1120         char *fn;
1121         sd_id128_t machine;
1122         char ids[33];
1123
1124         r = sd_id128_get_machine(&machine);
1125         if (r < 0)
1126                 return r;
1127
1128         sd_id128_to_string(machine, ids);
1129
1130         if (!s->system_journal) {
1131
1132                 /* First try to create the machine path, but not the prefix */
1133                 fn = strappend("/var/log/journal/", ids);
1134                 if (!fn)
1135                         return -ENOMEM;
1136                 (void) mkdir(fn, 0755);
1137                 free(fn);
1138
1139                 /* The create the system journal file */
1140                 fn = join("/var/log/journal/", ids, "/system.journal", NULL);
1141                 if (!fn)
1142                         return -ENOMEM;
1143
1144                 r = journal_file_open(fn, O_RDWR|O_CREAT, 0640, NULL, &s->system_journal);
1145                 free(fn);
1146
1147                 if (r >= 0) {
1148                         s->system_journal->metrics = s->metrics;
1149                         s->system_journal->compress = s->compress;
1150
1151                         fix_perms(s->system_journal, 0);
1152                 } else if (r < 0) {
1153
1154                         if (r == -ENOENT)
1155                                 r = 0;
1156                         else {
1157                                 log_error("Failed to open system journal: %s", strerror(-r));
1158                                 return r;
1159                         }
1160                 }
1161         }
1162
1163         if (!s->runtime_journal) {
1164
1165                 fn = join("/run/log/journal/", ids, "/system.journal", NULL);
1166                 if (!fn)
1167                         return -ENOMEM;
1168
1169                 if (s->system_journal) {
1170
1171                         /* Try to open the runtime journal, but only
1172                          * if it already exists, so that we can flush
1173                          * it into the system journal */
1174
1175                         r = journal_file_open(fn, O_RDWR, 0640, NULL, &s->runtime_journal);
1176                         free(fn);
1177
1178                         if (r < 0) {
1179
1180                                 if (r == -ENOENT)
1181                                         r = 0;
1182                                 else {
1183                                         log_error("Failed to open runtime journal: %s", strerror(-r));
1184                                         return r;
1185                                 }
1186                         }
1187
1188                 } else {
1189
1190                         /* OK, we really need the runtime journal, so create
1191                          * it if necessary. */
1192
1193                         (void) mkdir_parents(fn, 0755);
1194                         r = journal_file_open(fn, O_RDWR|O_CREAT, 0640, NULL, &s->runtime_journal);
1195                         free(fn);
1196
1197                         if (r < 0) {
1198                                 log_error("Failed to open runtime journal: %s", strerror(-r));
1199                                 return r;
1200                         }
1201                 }
1202
1203                 if (s->runtime_journal) {
1204                         s->runtime_journal->metrics = s->metrics;
1205                         s->runtime_journal->compress = s->compress;
1206
1207                         fix_perms(s->runtime_journal, 0);
1208                 }
1209         }
1210
1211         return r;
1212 }
1213
1214 static int server_flush_to_var(Server *s) {
1215         char path[] = "/run/log/journal/xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx";
1216         Object *o = NULL;
1217         int r;
1218         sd_id128_t machine;
1219         sd_journal *j;
1220         usec_t ts;
1221
1222         assert(s);
1223
1224         if (!s->runtime_journal)
1225                 return 0;
1226
1227         ts = now(CLOCK_MONOTONIC);
1228         if (s->var_available_timestamp + RECHECK_VAR_AVAILABLE_USEC > ts)
1229                 return 0;
1230
1231         s->var_available_timestamp = ts;
1232
1233         system_journal_open(s);
1234
1235         if (!s->system_journal)
1236                 return 0;
1237
1238         r = sd_id128_get_machine(&machine);
1239         if (r < 0) {
1240                 log_error("Failed to get machine id: %s", strerror(-r));
1241                 return r;
1242         }
1243
1244         r = sd_journal_open(&j, SD_JOURNAL_RUNTIME_ONLY);
1245         if (r < 0) {
1246                 log_error("Failed to read runtime journal: %s", strerror(-r));
1247                 return r;
1248         }
1249
1250         SD_JOURNAL_FOREACH(j) {
1251                 JournalFile *f;
1252
1253                 f = j->current_file;
1254                 assert(f && f->current_offset > 0);
1255
1256                 r = journal_file_move_to_object(f, OBJECT_ENTRY, f->current_offset, &o);
1257                 if (r < 0) {
1258                         log_error("Can't read entry: %s", strerror(-r));
1259                         goto finish;
1260                 }
1261
1262                 r = journal_file_copy_entry(f, s->system_journal, o, f->current_offset, NULL, NULL, NULL);
1263                 if (r == -E2BIG) {
1264                         log_info("Allocation limit reached.");
1265
1266                         journal_file_post_change(s->system_journal);
1267                         server_vacuum(s);
1268
1269                         r = journal_file_copy_entry(f, s->system_journal, o, f->current_offset, NULL, NULL, NULL);
1270                 }
1271
1272                 if (r < 0) {
1273                         log_error("Can't write entry: %s", strerror(-r));
1274                         goto finish;
1275                 }
1276         }
1277
1278 finish:
1279         journal_file_post_change(s->system_journal);
1280
1281         journal_file_close(s->runtime_journal);
1282         s->runtime_journal = NULL;
1283
1284         if (r >= 0) {
1285                 sd_id128_to_string(machine, path + 17);
1286                 rm_rf(path, false, true, false);
1287         }
1288
1289         return r;
1290 }
1291
1292 static void forward_syslog(Server *s, const void *buffer, size_t length, struct ucred *ucred, struct timeval *tv) {
1293         struct msghdr msghdr;
1294         struct iovec iovec;
1295         struct cmsghdr *cmsg;
1296         union {
1297                 struct cmsghdr cmsghdr;
1298                 uint8_t buf[CMSG_SPACE(sizeof(struct ucred)) +
1299                             CMSG_SPACE(sizeof(struct timeval))];
1300         } control;
1301         union sockaddr_union sa;
1302
1303         assert(s);
1304
1305         zero(msghdr);
1306
1307         zero(iovec);
1308         iovec.iov_base = (void*) buffer;
1309         iovec.iov_len = length;
1310         msghdr.msg_iov = &iovec;
1311         msghdr.msg_iovlen = 1;
1312
1313         zero(sa);
1314         sa.un.sun_family = AF_UNIX;
1315         strncpy(sa.un.sun_path, "/run/systemd/syslog", sizeof(sa.un.sun_path));
1316         msghdr.msg_name = &sa;
1317         msghdr.msg_namelen = offsetof(union sockaddr_union, un.sun_path) + strlen(sa.un.sun_path);
1318
1319         zero(control);
1320         msghdr.msg_control = &control;
1321         msghdr.msg_controllen = sizeof(control);
1322
1323         cmsg = CMSG_FIRSTHDR(&msghdr);
1324         cmsg->cmsg_level = SOL_SOCKET;
1325         cmsg->cmsg_type = SCM_CREDENTIALS;
1326         cmsg->cmsg_len = CMSG_LEN(sizeof(struct ucred));
1327         memcpy(CMSG_DATA(cmsg), ucred, sizeof(struct ucred));
1328         msghdr.msg_controllen = cmsg->cmsg_len;
1329
1330         /* Forward the syslog message we received via /dev/log to
1331          * /run/systemd/syslog. Unfortunately we currently can't set
1332          * the SO_TIMESTAMP auxiliary data, and hence we don't. */
1333
1334         if (sendmsg(s->syslog_fd, &msghdr, MSG_NOSIGNAL) >= 0)
1335                 return;
1336
1337         if (errno == ESRCH) {
1338                 struct ucred u;
1339
1340                 /* Hmm, presumably the sender process vanished
1341                  * by now, so let's fix it as good as we
1342                  * can, and retry */
1343
1344                 u = *ucred;
1345                 u.pid = getpid();
1346                 memcpy(CMSG_DATA(cmsg), &u, sizeof(struct ucred));
1347
1348                 if (sendmsg(s->syslog_fd, &msghdr, MSG_NOSIGNAL) >= 0)
1349                         return;
1350         }
1351
1352         log_debug("Failed to forward syslog message: %m");
1353 }
1354
1355 static int process_event(Server *s, struct epoll_event *ev) {
1356         assert(s);
1357
1358         if (ev->data.fd == s->signal_fd) {
1359                 struct signalfd_siginfo sfsi;
1360                 ssize_t n;
1361
1362                 if (ev->events != EPOLLIN) {
1363                         log_info("Got invalid event from epoll.");
1364                         return -EIO;
1365                 }
1366
1367                 n = read(s->signal_fd, &sfsi, sizeof(sfsi));
1368                 if (n != sizeof(sfsi)) {
1369
1370                         if (n >= 0)
1371                                 return -EIO;
1372
1373                         if (errno == EINTR || errno == EAGAIN)
1374                                 return 0;
1375
1376                         return -errno;
1377                 }
1378
1379                 if (sfsi.ssi_signo == SIGUSR1) {
1380                         server_flush_to_var(s);
1381                         return 0;
1382                 }
1383
1384                 log_debug("Received SIG%s", signal_to_string(sfsi.ssi_signo));
1385                 return 0;
1386
1387         } else if (ev->data.fd == s->native_fd ||
1388                    ev->data.fd == s->syslog_fd) {
1389
1390                 if (ev->events != EPOLLIN) {
1391                         log_info("Got invalid event from epoll.");
1392                         return -EIO;
1393                 }
1394
1395                 for (;;) {
1396                         struct msghdr msghdr;
1397                         struct iovec iovec;
1398                         struct ucred *ucred = NULL;
1399                         struct timeval *tv = NULL;
1400                         struct cmsghdr *cmsg;
1401                         union {
1402                                 struct cmsghdr cmsghdr;
1403                                 uint8_t buf[CMSG_SPACE(sizeof(struct ucred)) +
1404                                             CMSG_SPACE(sizeof(struct timeval))];
1405                         } control;
1406                         ssize_t n;
1407                         int v;
1408
1409                         if (ioctl(ev->data.fd, SIOCINQ, &v) < 0) {
1410                                 log_error("SIOCINQ failed: %m");
1411                                 return -errno;
1412                         }
1413
1414                         if (v <= 0)
1415                                 return 1;
1416
1417                         if (s->buffer_size < (size_t) v) {
1418                                 void *b;
1419                                 size_t l;
1420
1421                                 l = MAX(LINE_MAX + (size_t) v, s->buffer_size * 2);
1422                                 b = realloc(s->buffer, l+1);
1423
1424                                 if (!b) {
1425                                         log_error("Couldn't increase buffer.");
1426                                         return -ENOMEM;
1427                                 }
1428
1429                                 s->buffer_size = l;
1430                                 s->buffer = b;
1431                         }
1432
1433                         zero(iovec);
1434                         iovec.iov_base = s->buffer;
1435                         iovec.iov_len = s->buffer_size;
1436
1437                         zero(control);
1438                         zero(msghdr);
1439                         msghdr.msg_iov = &iovec;
1440                         msghdr.msg_iovlen = 1;
1441                         msghdr.msg_control = &control;
1442                         msghdr.msg_controllen = sizeof(control);
1443
1444                         n = recvmsg(ev->data.fd, &msghdr, MSG_DONTWAIT);
1445                         if (n < 0) {
1446
1447                                 if (errno == EINTR || errno == EAGAIN)
1448                                         return 1;
1449
1450                                 log_error("recvmsg() failed: %m");
1451                                 return -errno;
1452                         }
1453
1454                         for (cmsg = CMSG_FIRSTHDR(&msghdr); cmsg; cmsg = CMSG_NXTHDR(&msghdr, cmsg)) {
1455
1456                                 if (cmsg->cmsg_level == SOL_SOCKET &&
1457                                     cmsg->cmsg_type == SCM_CREDENTIALS &&
1458                                     cmsg->cmsg_len == CMSG_LEN(sizeof(struct ucred)))
1459                                         ucred = (struct ucred*) CMSG_DATA(cmsg);
1460                                 else if (cmsg->cmsg_level == SOL_SOCKET &&
1461                                          cmsg->cmsg_type == SO_TIMESTAMP &&
1462                                          cmsg->cmsg_len == CMSG_LEN(sizeof(struct timeval)))
1463                                         tv = (struct timeval*) CMSG_DATA(cmsg);
1464                         }
1465
1466                         if (ev->data.fd == s->syslog_fd) {
1467                                 char *e;
1468
1469                                 e = memchr(s->buffer, '\n', n);
1470                                 if (e)
1471                                         *e = 0;
1472                                 else
1473                                         s->buffer[n] = 0;
1474
1475                                 forward_syslog(s, s->buffer, n, ucred, tv);
1476                                 process_syslog_message(s, strstrip(s->buffer), ucred, tv);
1477                         } else
1478                                 process_native_message(s, s->buffer, n, ucred, tv);
1479                 }
1480
1481                 return 1;
1482
1483         } else if (ev->data.fd == s->stdout_fd) {
1484
1485                 if (ev->events != EPOLLIN) {
1486                         log_info("Got invalid event from epoll.");
1487                         return -EIO;
1488                 }
1489
1490                 stdout_stream_new(s);
1491                 return 1;
1492
1493         } else {
1494                 StdoutStream *stream;
1495
1496                 if ((ev->events|EPOLLIN|EPOLLHUP) != (EPOLLIN|EPOLLHUP)) {
1497                         log_info("Got invalid event from epoll.");
1498                         return -EIO;
1499                 }
1500
1501                 /* If it is none of the well-known fds, it must be an
1502                  * stdout stream fd. Note that this is a bit ugly here
1503                  * (since we rely that none of the well-known fds
1504                  * could be interpreted as pointer), but nonetheless
1505                  * safe, since the well-known fds would never get an
1506                  * fd > 4096, i.e. beyond the first memory page */
1507
1508                 stream = ev->data.ptr;
1509
1510                 if (stdout_stream_process(stream) <= 0)
1511                         stdout_stream_free(stream);
1512
1513                 return 1;
1514         }
1515
1516         log_error("Unknown event.");
1517         return 0;
1518 }
1519
1520 static int open_syslog_socket(Server *s) {
1521         union sockaddr_union sa;
1522         int one, r;
1523         struct epoll_event ev;
1524         struct timeval tv;
1525
1526         assert(s);
1527
1528         if (s->syslog_fd < 0) {
1529
1530                 s->syslog_fd = socket(AF_UNIX, SOCK_DGRAM|SOCK_CLOEXEC, 0);
1531                 if (s->syslog_fd < 0) {
1532                         log_error("socket() failed: %m");
1533                         return -errno;
1534                 }
1535
1536                 zero(sa);
1537                 sa.un.sun_family = AF_UNIX;
1538                 strncpy(sa.un.sun_path, "/dev/log", sizeof(sa.un.sun_path));
1539
1540                 unlink(sa.un.sun_path);
1541
1542                 r = bind(s->syslog_fd, &sa.sa, offsetof(union sockaddr_union, un.sun_path) + strlen(sa.un.sun_path));
1543                 if (r < 0) {
1544                         log_error("bind() failed: %m");
1545                         return -errno;
1546                 }
1547
1548                 chmod(sa.un.sun_path, 0666);
1549         }
1550
1551         one = 1;
1552         r = setsockopt(s->syslog_fd, SOL_SOCKET, SO_PASSCRED, &one, sizeof(one));
1553         if (r < 0) {
1554                 log_error("SO_PASSCRED failed: %m");
1555                 return -errno;
1556         }
1557
1558         one = 1;
1559         r = setsockopt(s->syslog_fd, SOL_SOCKET, SO_TIMESTAMP, &one, sizeof(one));
1560         if (r < 0) {
1561                 log_error("SO_TIMESTAMP failed: %m");
1562                 return -errno;
1563         }
1564
1565         /* Since we use the same socket for forwarding this to some
1566          * other syslog implementation, make sure we don't hang
1567          * forever */
1568         timeval_store(&tv, SYSLOG_TIMEOUT_USEC);
1569         if (setsockopt(s->syslog_fd, SOL_SOCKET, SO_SNDTIMEO, &tv, sizeof(tv)) < 0) {
1570                 log_error("SO_SNDTIMEO failed: %m");
1571                 return -errno;
1572         }
1573
1574         zero(ev);
1575         ev.events = EPOLLIN;
1576         ev.data.fd = s->syslog_fd;
1577         if (epoll_ctl(s->epoll_fd, EPOLL_CTL_ADD, s->syslog_fd, &ev) < 0) {
1578                 log_error("Failed to add syslog server fd to epoll object: %m");
1579                 return -errno;
1580         }
1581
1582         return 0;
1583 }
1584
1585 static int open_native_socket(Server*s) {
1586         union sockaddr_union sa;
1587         int one, r;
1588         struct epoll_event ev;
1589
1590         assert(s);
1591
1592         if (s->native_fd < 0) {
1593
1594                 s->native_fd = socket(AF_UNIX, SOCK_DGRAM|SOCK_CLOEXEC, 0);
1595                 if (s->native_fd < 0) {
1596                         log_error("socket() failed: %m");
1597                         return -errno;
1598                 }
1599
1600                 zero(sa);
1601                 sa.un.sun_family = AF_UNIX;
1602                 strncpy(sa.un.sun_path, "/run/systemd/journal", sizeof(sa.un.sun_path));
1603
1604                 unlink(sa.un.sun_path);
1605
1606                 r = bind(s->native_fd, &sa.sa, offsetof(union sockaddr_union, un.sun_path) + strlen(sa.un.sun_path));
1607                 if (r < 0) {
1608                         log_error("bind() failed: %m");
1609                         return -errno;
1610                 }
1611
1612                 chmod(sa.un.sun_path, 0666);
1613         }
1614
1615         one = 1;
1616         r = setsockopt(s->native_fd, SOL_SOCKET, SO_PASSCRED, &one, sizeof(one));
1617         if (r < 0) {
1618                 log_error("SO_PASSCRED failed: %m");
1619                 return -errno;
1620         }
1621
1622         one = 1;
1623         r = setsockopt(s->native_fd, SOL_SOCKET, SO_TIMESTAMP, &one, sizeof(one));
1624         if (r < 0) {
1625                 log_error("SO_TIMESTAMP failed: %m");
1626                 return -errno;
1627         }
1628
1629         zero(ev);
1630         ev.events = EPOLLIN;
1631         ev.data.fd = s->native_fd;
1632         if (epoll_ctl(s->epoll_fd, EPOLL_CTL_ADD, s->native_fd, &ev) < 0) {
1633                 log_error("Failed to add native server fd to epoll object: %m");
1634                 return -errno;
1635         }
1636
1637         return 0;
1638 }
1639
1640 static int open_stdout_socket(Server *s) {
1641         union sockaddr_union sa;
1642         int r;
1643         struct epoll_event ev;
1644
1645         assert(s);
1646
1647         if (s->stdout_fd < 0) {
1648
1649                 s->stdout_fd = socket(AF_UNIX, SOCK_STREAM|SOCK_CLOEXEC, 0);
1650                 if (s->stdout_fd < 0) {
1651                         log_error("socket() failed: %m");
1652                         return -errno;
1653                 }
1654
1655                 zero(sa);
1656                 sa.un.sun_family = AF_UNIX;
1657                 strncpy(sa.un.sun_path, "/run/systemd/stdout", sizeof(sa.un.sun_path));
1658
1659                 unlink(sa.un.sun_path);
1660
1661                 r = bind(s->stdout_fd, &sa.sa, offsetof(union sockaddr_union, un.sun_path) + strlen(sa.un.sun_path));
1662                 if (r < 0) {
1663                         log_error("bind() failed: %m");
1664                         return -errno;
1665                 }
1666
1667                 chmod(sa.un.sun_path, 0666);
1668
1669                 if (listen(s->stdout_fd, SOMAXCONN) < 0) {
1670                         log_error("liste() failed: %m");
1671                         return -errno;
1672                 }
1673         }
1674
1675         zero(ev);
1676         ev.events = EPOLLIN;
1677         ev.data.fd = s->stdout_fd;
1678         if (epoll_ctl(s->epoll_fd, EPOLL_CTL_ADD, s->stdout_fd, &ev) < 0) {
1679                 log_error("Failed to add stdout server fd to epoll object: %m");
1680                 return -errno;
1681         }
1682
1683         return 0;
1684 }
1685
1686 static int open_signalfd(Server *s) {
1687         sigset_t mask;
1688         struct epoll_event ev;
1689
1690         assert(s);
1691
1692         assert_se(sigemptyset(&mask) == 0);
1693         sigset_add_many(&mask, SIGINT, SIGTERM, SIGUSR1, -1);
1694         assert_se(sigprocmask(SIG_SETMASK, &mask, NULL) == 0);
1695
1696         s->signal_fd = signalfd(-1, &mask, SFD_NONBLOCK|SFD_CLOEXEC);
1697         if (s->signal_fd < 0) {
1698                 log_error("signalfd(): %m");
1699                 return -errno;
1700         }
1701
1702         zero(ev);
1703         ev.events = EPOLLIN;
1704         ev.data.fd = s->signal_fd;
1705
1706         if (epoll_ctl(s->epoll_fd, EPOLL_CTL_ADD, s->signal_fd, &ev) < 0) {
1707                 log_error("epoll_ctl(): %m");
1708                 return -errno;
1709         }
1710
1711         return 0;
1712 }
1713
1714 static int server_init(Server *s) {
1715         int n, r, fd;
1716
1717         assert(s);
1718
1719         zero(*s);
1720         s->syslog_fd = s->native_fd = s->stdout_fd = s->signal_fd = s->epoll_fd = -1;
1721         s->metrics.max_size = DEFAULT_MAX_SIZE;
1722         s->metrics.min_size = DEFAULT_MIN_SIZE;
1723         s->metrics.keep_free = DEFAULT_KEEP_FREE;
1724         s->max_use = DEFAULT_MAX_USE;
1725         s->compress = true;
1726
1727         s->user_journals = hashmap_new(trivial_hash_func, trivial_compare_func);
1728         if (!s->user_journals) {
1729                 log_error("Out of memory.");
1730                 return -ENOMEM;
1731         }
1732
1733         s->epoll_fd = epoll_create1(EPOLL_CLOEXEC);
1734         if (s->epoll_fd < 0) {
1735                 log_error("Failed to create epoll object: %m");
1736                 return -errno;
1737         }
1738
1739         n = sd_listen_fds(true);
1740         if (n < 0) {
1741                 log_error("Failed to read listening file descriptors from environment: %s", strerror(-n));
1742                 return n;
1743         }
1744
1745         for (fd = SD_LISTEN_FDS_START; fd < SD_LISTEN_FDS_START + n; fd++) {
1746
1747                 if (sd_is_socket_unix(fd, SOCK_DGRAM, -1, "/run/systemd/native", 0) > 0) {
1748
1749                         if (s->native_fd >= 0) {
1750                                 log_error("Too many native sockets passed.");
1751                                 return -EINVAL;
1752                         }
1753
1754                         s->native_fd = fd;
1755
1756                 } else if (sd_is_socket_unix(fd, SOCK_STREAM, 1, "/run/systemd/stdout", 0) > 0) {
1757
1758                         if (s->stdout_fd >= 0) {
1759                                 log_error("Too many stdout sockets passed.");
1760                                 return -EINVAL;
1761                         }
1762
1763                         s->stdout_fd = fd;
1764
1765                 } else if (sd_is_socket_unix(fd, SOCK_DGRAM, -1, "/dev/log", 0) > 0) {
1766
1767                         if (s->syslog_fd >= 0) {
1768                                 log_error("Too many /dev/log sockets passed.");
1769                                 return -EINVAL;
1770                         }
1771
1772                         s->syslog_fd = fd;
1773
1774                 } else {
1775                         log_error("Unknown socket passed.");
1776                         return -EINVAL;
1777                 }
1778         }
1779
1780         r = open_syslog_socket(s);
1781         if (r < 0)
1782                 return r;
1783
1784         r = open_native_socket(s);
1785         if (r < 0)
1786                 return r;
1787
1788         r = open_stdout_socket(s);
1789         if (r < 0)
1790                 return r;
1791
1792         r = system_journal_open(s);
1793         if (r < 0)
1794                 return r;
1795
1796         r = open_signalfd(s);
1797         if (r < 0)
1798                 return r;
1799
1800         s->rate_limit = journal_rate_limit_new(DEFAULT_RATE_LIMIT_INTERVAL, DEFAULT_RATE_LIMIT_BURST);
1801         if (!s->rate_limit)
1802                 return -ENOMEM;
1803
1804         return 0;
1805 }
1806
1807 static void server_done(Server *s) {
1808         JournalFile *f;
1809         assert(s);
1810
1811         while (s->stdout_streams)
1812                 stdout_stream_free(s->stdout_streams);
1813
1814         if (s->system_journal)
1815                 journal_file_close(s->system_journal);
1816
1817         if (s->runtime_journal)
1818                 journal_file_close(s->runtime_journal);
1819
1820         while ((f = hashmap_steal_first(s->user_journals)))
1821                 journal_file_close(f);
1822
1823         hashmap_free(s->user_journals);
1824
1825         if (s->epoll_fd >= 0)
1826                 close_nointr_nofail(s->epoll_fd);
1827
1828         if (s->signal_fd >= 0)
1829                 close_nointr_nofail(s->signal_fd);
1830
1831         if (s->syslog_fd >= 0)
1832                 close_nointr_nofail(s->syslog_fd);
1833
1834         if (s->native_fd >= 0)
1835                 close_nointr_nofail(s->native_fd);
1836
1837         if (s->stdout_fd >= 0)
1838                 close_nointr_nofail(s->stdout_fd);
1839
1840         if (s->rate_limit)
1841                 journal_rate_limit_free(s->rate_limit);
1842
1843         free(s->buffer);
1844 }
1845
1846 int main(int argc, char *argv[]) {
1847         Server server;
1848         int r;
1849
1850         /* if (getppid() != 1) { */
1851         /*         log_error("This program should be invoked by init only."); */
1852         /*         return EXIT_FAILURE; */
1853         /* } */
1854
1855         if (argc > 1) {
1856                 log_error("This program does not take arguments.");
1857                 return EXIT_FAILURE;
1858         }
1859
1860         log_set_target(LOG_TARGET_CONSOLE);
1861         log_parse_environment();
1862         log_open();
1863
1864         umask(0022);
1865
1866         r = server_init(&server);
1867         if (r < 0)
1868                 goto finish;
1869
1870         log_debug("systemd-journald running as pid %lu", (unsigned long) getpid());
1871
1872         sd_notify(false,
1873                   "READY=1\n"
1874                   "STATUS=Processing requests...");
1875
1876         server_vacuum(&server);
1877         server_flush_to_var(&server);
1878
1879         for (;;) {
1880                 struct epoll_event event;
1881
1882                 r = epoll_wait(server.epoll_fd, &event, 1, -1);
1883                 if (r < 0) {
1884
1885                         if (errno == EINTR)
1886                                 continue;
1887
1888                         log_error("epoll_wait() failed: %m");
1889                         r = -errno;
1890                         goto finish;
1891                 } else if (r == 0)
1892                         break;
1893
1894                 r = process_event(&server, &event);
1895                 if (r < 0)
1896                         goto finish;
1897                 else if (r == 0)
1898                         break;
1899         }
1900
1901         log_debug("systemd-journald stopped as pid %lu", (unsigned long) getpid());
1902
1903 finish:
1904         sd_notify(false,
1905                   "STATUS=Shutting down...");
1906
1907         server_done(&server);
1908
1909         return r < 0 ? EXIT_FAILURE : EXIT_SUCCESS;
1910 }