1 /*- 2 * Copyright (c) 2003-2004 Sean M. Kelly <smkelly@FreeBSD.org> 3 * All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 14 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 15 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 16 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 17 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 18 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 19 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 20 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 21 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 22 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 23 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 24 * SUCH DAMAGE. 25 */ 26 27 /* 28 * Software watchdog daemon. 29 */ 30 31 #include <sys/types.h> 32 __FBSDID("$FreeBSD$"); 33 34 #include <sys/mman.h> 35 #include <sys/param.h> 36 #include <sys/rtprio.h> 37 #include <sys/stat.h> 38 #include <sys/time.h> 39 #include <sys/watchdog.h> 40 41 #include <err.h> 42 #include <errno.h> 43 #include <fcntl.h> 44 #include <libutil.h> 45 #include <math.h> 46 #include <paths.h> 47 #include <signal.h> 48 #include <stdio.h> 49 #include <stdlib.h> 50 #include <string.h> 51 #include <strings.h> 52 #include <sysexits.h> 53 #include <unistd.h> 54 55 static void parseargs(int, char *[]); 56 static void sighandler(int); 57 static void watchdog_loop(void); 58 static int watchdog_init(void); 59 static int watchdog_onoff(int onoff); 60 static int watchdog_patpat(u_int timeout); 61 static void usage(void); 62 63 static int debugging = 0; 64 static int end_program = 0; 65 static const char *pidfile = _PATH_VARRUN "watchdogd.pid"; 66 static u_int timeout = WD_TO_16SEC; 67 static u_int passive = 0; 68 static int is_daemon = 0; 69 static int fd = -1; 70 static int nap = 1; 71 static char *test_cmd = NULL; 72 73 /* 74 * Ask malloc() to map minimum-sized chunks of virtual address space at a time, 75 * so that mlockall() won't needlessly wire megabytes of unused memory into the 76 * process. This must be done using the malloc_conf string so that it gets set 77 * up before the first allocation, which happens before entry to main(). 78 */ 79 const char * malloc_conf = "lg_chunk:0"; 80 81 /* 82 * Periodically pat the watchdog, preventing it from firing. 83 */ 84 int 85 main(int argc, char *argv[]) 86 { 87 struct rtprio rtp; 88 struct pidfh *pfh; 89 pid_t otherpid; 90 91 if (getuid() != 0) 92 errx(EX_SOFTWARE, "not super user"); 93 94 parseargs(argc, argv); 95 96 rtp.type = RTP_PRIO_REALTIME; 97 rtp.prio = 0; 98 if (rtprio(RTP_SET, 0, &rtp) == -1) 99 err(EX_OSERR, "rtprio"); 100 101 if (watchdog_init() == -1) 102 errx(EX_SOFTWARE, "unable to initialize watchdog"); 103 104 if (is_daemon) { 105 if (watchdog_onoff(1) == -1) 106 err(EX_OSERR, "patting the dog"); 107 108 pfh = pidfile_open(pidfile, 0600, &otherpid); 109 if (pfh == NULL) { 110 if (errno == EEXIST) { 111 errx(EX_SOFTWARE, "%s already running, pid: %d", 112 getprogname(), otherpid); 113 } 114 warn("Cannot open or create pidfile"); 115 } 116 117 if (debugging == 0 && daemon(0, 0) == -1) { 118 watchdog_onoff(0); 119 pidfile_remove(pfh); 120 err(EX_OSERR, "daemon"); 121 } 122 123 signal(SIGHUP, SIG_IGN); 124 signal(SIGINT, sighandler); 125 signal(SIGTERM, sighandler); 126 127 pidfile_write(pfh); 128 if (madvise(0, 0, MADV_PROTECT) != 0) 129 warn("madvise failed"); 130 if (mlockall(MCL_CURRENT | MCL_FUTURE) != 0) 131 warn("mlockall failed"); 132 133 watchdog_loop(); 134 135 /* exiting */ 136 pidfile_remove(pfh); 137 return (EX_OK); 138 } else { 139 if (passive) 140 timeout |= WD_PASSIVE; 141 else 142 timeout |= WD_ACTIVE; 143 if (watchdog_patpat(timeout) < 0) 144 err(EX_OSERR, "patting the dog"); 145 return (EX_OK); 146 } 147 } 148 149 /* 150 * Catch signals and begin shutdown process. 151 */ 152 static void 153 sighandler(int signum) 154 { 155 156 if (signum == SIGINT || signum == SIGTERM) 157 end_program = 1; 158 } 159 160 /* 161 * Open the watchdog device. 162 */ 163 static int 164 watchdog_init(void) 165 { 166 167 fd = open("/dev/" _PATH_WATCHDOG, O_RDWR); 168 if (fd >= 0) 169 return (0); 170 warn("Could not open watchdog device"); 171 return (-1); 172 } 173 174 /* 175 * Main program loop which is iterated every second. 176 */ 177 static void 178 watchdog_loop(void) 179 { 180 struct stat sb; 181 int failed; 182 183 while (end_program != 2) { 184 failed = 0; 185 186 if (test_cmd != NULL) 187 failed = system(test_cmd); 188 else 189 failed = stat("/etc", &sb); 190 191 if (failed == 0) 192 watchdog_patpat(timeout|WD_ACTIVE); 193 sleep(nap); 194 195 if (end_program != 0) { 196 if (watchdog_onoff(0) == 0) { 197 end_program = 2; 198 } else { 199 warnx("Could not stop the watchdog, not exiting"); 200 end_program = 0; 201 } 202 } 203 } 204 } 205 206 /* 207 * Reset the watchdog timer. This function must be called periodically 208 * to keep the watchdog from firing. 209 */ 210 static int 211 watchdog_patpat(u_int t) 212 { 213 214 return ioctl(fd, WDIOCPATPAT, &t); 215 } 216 217 /* 218 * Toggle the kernel's watchdog. This routine is used to enable and 219 * disable the watchdog. 220 */ 221 static int 222 watchdog_onoff(int onoff) 223 { 224 225 if (onoff) 226 return watchdog_patpat((timeout|WD_ACTIVE)); 227 else 228 return watchdog_patpat(0); 229 } 230 231 /* 232 * Tell user how to use the program. 233 */ 234 static void 235 usage(void) 236 { 237 if (is_daemon) 238 fprintf(stderr, "usage: watchdogd [-d] [-e cmd] [-I file] [-s sleep] [-t timeout]\n"); 239 else 240 fprintf(stderr, "usage: watchdog [-d] [-t timeout]\n"); 241 exit(EX_USAGE); 242 } 243 244 /* 245 * Handle the few command line arguments supported. 246 */ 247 static void 248 parseargs(int argc, char *argv[]) 249 { 250 int c; 251 char *p; 252 double a; 253 254 c = strlen(argv[0]); 255 if (argv[0][c - 1] == 'd') 256 is_daemon = 1; 257 while ((c = getopt(argc, argv, 258 is_daemon ? "I:de:s:t:?" : "dt:?")) != -1) { 259 switch (c) { 260 case 'I': 261 pidfile = optarg; 262 break; 263 case 'd': 264 debugging = 1; 265 break; 266 case 'e': 267 test_cmd = strdup(optarg); 268 break; 269 #ifdef notyet 270 case 'p': 271 passive = 1; 272 break; 273 #endif 274 case 's': 275 p = NULL; 276 errno = 0; 277 nap = strtol(optarg, &p, 0); 278 if ((p != NULL && *p != '\0') || errno != 0) 279 errx(EX_USAGE, "-s argument is not a number"); 280 break; 281 case 't': 282 p = NULL; 283 errno = 0; 284 a = strtod(optarg, &p); 285 if ((p != NULL && *p != '\0') || errno != 0) 286 errx(EX_USAGE, "-t argument is not a number"); 287 if (a < 0) 288 errx(EX_USAGE, "-t argument must be positive"); 289 if (a == 0) 290 timeout = WD_TO_NEVER; 291 else 292 timeout = flsll(a * 1e9); 293 if (debugging) 294 printf("Timeout is 2^%d nanoseconds\n", 295 timeout); 296 break; 297 case '?': 298 default: 299 usage(); 300 /* NOTREACHED */ 301 } 302 } 303 if (argc != optind) 304 errx(EX_USAGE, "extra arguments."); 305 if (is_daemon && timeout < WD_TO_1SEC) 306 errx(EX_USAGE, "-t argument is less than one second."); 307 } 308