ipcs : display all the IPC fascilities information

crond : CRON scheduler

2 files changed, 1131 insertions, 0 deletions

diff --git a/toys/pending/crond.c b/toys/pending/crond.c
new file mode 100644
index 00000000..1b5e4194
--- /dev/null
+++ b/toys/pending/crond.c
@@ -0,0 +1,681 @@
+/* crond.c - daemon to execute scheduled commands.
+ *
+ * Copyright 2014 Ranjan Kumar <ranjankumar.bth@gmail.com>
+ *
+ * No Standard
+
+USE_CROND(NEWTOY(crond, "fbSl#<0=8d#<0L:c:[-bf][-LS][-ld]", TOYFLAG_USR|TOYFLAG_SBIN|TOYFLAG_NEEDROOT))
+
+config CROND
+  bool "crond"
+  default n
+  help
+    usage: crond [-fbS] [-l N] [-d N] [-L LOGFILE] [-c DIR]
+
+    A daemon to execute scheduled commands.
+
+    -b Background (default)
+    -c crontab dir
+    -d Set log level, log to stderr
+    -f Foreground
+    -l Set log level. 0 is the most verbose, default 8
+    -S Log to syslog (default)
+    -L Log to file
+*/
+
+#define FOR_crond
+#include "toys.h"
+
+GLOBALS(
+  char *crontabs_dir;
+  char *logfile;
+  int loglevel_d;
+  int loglevel;
+
+  time_t crontabs_dir_mtime;
+  uint8_t flagd;
+)
+
+typedef struct _var {
+  struct _var *next, *prev;
+  char *name, *val;
+} VAR;
+
+typedef struct _job {
+  struct _job *next, *prev;
+  char min[60], hour[24], dom[31], mon[12], dow[7], *cmd;
+  int isrunning, needstart, mailsize;
+  pid_t pid;
+} JOB;
+
+typedef struct _cronfile {
+  struct _cronfile *next, *prev;
+  struct double_list *job, *var;
+  char *username, *mailto;
+  int invalid;
+} CRONFILE;
+
+static char days[]={"sun""mon""tue""wed""thu""fri""sat"};
+static char months[]={"jan""feb""mar""apr""may""jun""jul"
+  "aug""sep""oct""nov""dec"};
+CRONFILE *gclist;
+
+#define LOG_EXIT 0
+#define LOG_LEVEL5 5
+#define LOG_LEVEL7 7
+#define LOG_LEVEL8 8
+#define LOG_LEVEL9 9 // warning
+#define LOG_ERROR 20
+
+static void loginfo(uint8_t loglevel, char *msg, ...)
+{
+  va_list s, d;
+
+  va_start(s, msg);
+  va_copy(d, s);
+  if (loglevel >= TT.loglevel) {
+    int used;
+    char *smsg;
+
+    if (!TT.flagd && TT.logfile) {
+      int fd = open(TT.logfile, O_WRONLY | O_CREAT | O_APPEND, 0666);
+      if (fd >=0 && fd != 2) {
+        dup2(fd, 2);
+        close(fd);
+      } else if (fd < 0) perror_msg("'%s", TT.logfile);
+    }
+    used = vsnprintf(NULL, 0, msg, d);
+    smsg = xzalloc(++used);
+    vsnprintf(smsg, used, msg, s);
+    if (TT.flagd || TT.logfile) {
+      fflush(NULL);
+      smsg[used-1] = '\n';
+      writeall((loglevel > 8) ? 2 : 1, smsg, used);
+    } else syslog((loglevel > 8) ? LOG_ERR : LOG_INFO, "%s", smsg);
+    free(smsg);
+  }
+  va_end(d);
+  va_end(s);
+  if (!loglevel) exit(20);
+}
+
+/*
+ * Names can also be used for the 'month' and 'day of week' fields
+ * (First three letters of the particular day or month).
+ */
+static int getindex(char *src, int size)
+{
+  int i;
+  char *field = (size == 12) ? months : days;
+
+  // strings are not allowed for min, hour and dom fields.
+  if (!(size == 7 || size == 12)) return -1;
+
+  for (i = 0; field[i]; i += 3) {
+    if (!strncasecmp(src, &field[i], 3))
+      return (i/3);
+  }
+  return -1;
+}
+
+// set elements of minute, hour, day of month, month and day of week arrays.
+static void fillarray(char *dst, int start, int end, int skip)
+{
+  int sk = 1;
+
+  if (end < 0) {
+    dst[start] = 1;
+    return;
+  }
+  if (!skip) skip = 1;
+  do {
+    if (!--sk) {
+      dst[start] = 1;
+      sk = skip;
+    }
+  } while (start++ != end);
+}
+
+static long getval(char *num, long low, long high)
+{
+  long val = strtol(num, &num, 10);
+
+  if (*num || (val < low) || (val > high)) return -1;
+  return val;
+}
+
+//static int parse_and_fillarray(char *dst, int size, char *src)
+static int parse_and_fillarray(char *dst, int min, int max, char *src)
+{
+  int start, end, skip = 0;
+  char *ptr = strchr(src, '/');
+
+  if (ptr) {
+    *ptr++ = 0;
+    if ((skip = getval(ptr, min, (min ? max: max-1))) < 0) goto ERROR;
+  }
+
+  if (*src == '-' || *src == ',') goto ERROR;
+  if (*src == '*') {
+    if (*(src+1)) goto ERROR;
+    fillarray(dst, 0, max-1, skip);
+  } else {
+    for (;;) {
+      char *ctoken = strsep(&src, ","), *dtoken;
+
+      if (!ctoken) break;
+      if (!*ctoken) goto ERROR;
+
+      // Get start position.
+      dtoken = strsep(&ctoken, "-");
+      if (isdigit(*dtoken)) {
+        if ((start = getval(dtoken, min, (min ? max : max-1))) < 0) goto ERROR;
+        start = min ? (start-1) : start;
+      } else if ((start = getindex(dtoken, max)) < 0) goto ERROR;
+
+      // Get end position.
+      if (!ctoken) end = -1; // e.g. N1,N2,N3
+      else if (*ctoken) {// e.g. N-M
+        if (isdigit(*ctoken)) {
+          if ((end = getval(ctoken, min, (min ? max : max-1))) < 0) goto ERROR;
+          end = min ? (end-1) : end;
+        } else if ((end = getindex(ctoken, max)) < 0) goto ERROR;
+        if (end == start) end = -1;
+      } else goto ERROR; // error condition 'N-'
+      fillarray(dst, start, end, skip);
+    }
+  }
+
+  if (TT.flagd && (TT.loglevel <= 5)) {
+    for (start = 0; start < max; start++)
+      fprintf(stderr, "%d", (unsigned char)dst[start]);
+    fputc('\n', stderr);
+  }
+  return 0;
+ERROR:
+  loginfo(LOG_LEVEL9, "parse error at %s", src);
+  return -1;
+}
+
+static char *omitspace(char *line)
+{
+  while (*line == ' ' || *line == '\t') line++;
+  return line;
+}
+
+static void parse_line(char *line, CRONFILE *cfile)
+{
+  int count = 0;
+  char *name, *val, *tokens[5] = {0,};
+  VAR *v;
+  JOB *j;
+
+  line = omitspace(line);
+  if (!*line || *line == '#') return;
+
+  /*
+   * TODO: Enhancement to support 8 special strings
+   * @reboot -> Run once at startup.
+   * @yearly -> Run once a year (0 0 1 1 *).
+   * @annually -> Same as above.
+   * @monthly -> Run once a month (0 0 1 * *).
+   * @weekly -> Run once a week (0 0 * * 0).
+   * @daily -> Run once a day (0 0 * * *).
+   * @midnight -> same as above.
+   * @hourly -> Run once an hour (0 * * * *).
+   */
+  if (*line == '@') return;
+  if (TT.flagd) loginfo(LOG_LEVEL5, "user:%s entry:%s", cfile->username, line);
+  while (count<5) {
+    int len = strcspn(line, " \t");
+
+    if (line[len]) line[len++] = '\0';
+    tokens[count++] = line;
+    line += len;
+    line = omitspace(line);
+    if (!*line) break;
+  }
+
+  switch (count) {
+    case 1: // form SHELL=/bin/sh
+      name = tokens[0];
+      if ((val = strchr(name, '='))) *val++ = 0;
+      if (!val || !*val) return;
+      break;
+    case 2: // form SHELL =/bin/sh or SHELL= /bin/sh
+      name = tokens[0];
+      if ((val = strchr(name, '='))) {
+        *val = 0;
+        val = tokens[1];
+      } else {
+        if (*(tokens[1]) != '=') return;
+        val = tokens[1] + 1;
+      }
+      if (!*val) return;
+      break;
+    case 3: // NAME = VAL
+      name = tokens[0];
+      val = tokens[2];
+      if (*(tokens[1]) != '=') return;
+      break;
+    case 5:
+      // don't have any cmd to execute.
+      if (!*line) return;
+      j = xzalloc(sizeof(JOB));
+
+      if (parse_and_fillarray(j->min, 0, sizeof(j->min), tokens[0]))
+        goto STOP_PARSING;
+      if (parse_and_fillarray(j->hour, 0, sizeof(j->hour), tokens[1]))
+        goto STOP_PARSING;
+      if (parse_and_fillarray(j->dom, 1, sizeof(j->dom), tokens[2]))
+        goto STOP_PARSING;
+      if (parse_and_fillarray(j->mon, 1, sizeof(j->mon), tokens[3]))
+        goto STOP_PARSING;
+      if (parse_and_fillarray(j->dow, 0, sizeof(j->dow), tokens[4]))
+        goto STOP_PARSING;
+      j->cmd = xstrdup(line);
+
+      if (TT.flagd) loginfo(LOG_LEVEL5, " command:%s", j->cmd);
+      dlist_add_nomalloc((struct double_list **)&cfile->job, (struct double_list *)j);
+      return;
+STOP_PARSING:
+      free(j);
+      return;
+    default: return;
+  }
+  if (!strcmp(name, "MAILTO")) cfile->mailto = xstrdup(val);
+  else {
+    v = xzalloc(sizeof(VAR));
+    v->name = xstrdup(name);
+    v->val = xstrdup(val);
+    dlist_add_nomalloc((struct double_list **)&cfile->var, (struct double_list *)v);
+  }
+}
+
+static void free_jobs(JOB **jlist)
+{
+  JOB *j = dlist_pop(jlist);
+  free(j->cmd);
+  free(j);
+}
+
+static void free_cronfile(CRONFILE **list)
+{
+  CRONFILE *l = dlist_pop(list);
+  VAR *v, *vnode = (VAR *)l->var;
+
+  if (l->username != l->mailto) free(l->mailto);
+  free(l->username);
+  while (vnode && (v = dlist_pop(&vnode))) {
+    free(v->name);
+    free(v->val);
+    free(v);
+  }
+  free(l);
+}
+
+/*
+ * Iterate all cronfiles to identify the completed jobs and freed them.
+ * If all jobs got completed for a cronfile, freed cronfile too.
+ */
+static void remove_completed_jobs()
+{
+  CRONFILE *lstart, *list = gclist;
+
+  lstart = list;
+  while (list) {
+    int delete = 1;
+    JOB *jstart, *jlist = (JOB *)list->job;
+
+    list->invalid = 1;
+    jstart = jlist;
+    while (jlist) {
+      jlist->isrunning = 0;
+      if (jlist->pid > 0) {
+        jlist->isrunning = 1;
+        delete = 0;
+        jlist = jlist->next;
+      } else {
+        if (jlist == jstart) { // if 1st node has to delete.
+          jstart = jstart->next;
+          free_jobs(&jlist);
+          continue;
+        } else free_jobs(&jlist);
+      }
+      if (jlist == jstart) break;
+    }
+    list->job = (struct double_list *)jlist;
+
+    if (delete) {
+      if (lstart == list) {
+        lstart = lstart->next;
+        free_cronfile(&list);
+        continue;
+      } else free_cronfile(&list);
+    }
+    list = list->next;
+    if (lstart == list) break;
+  }
+  gclist = list;
+}
+
+// Scan cronfiles and prepare the list of cronfiles with their jobs.
+static void scan_cronfiles()
+{
+  DIR *dp;
+  struct dirent *entry;
+
+  remove_completed_jobs();
+  if (chdir(TT.crontabs_dir)) loginfo(LOG_EXIT, "chdir(%s)", TT.crontabs_dir);
+  if (!(dp = opendir("."))) loginfo(LOG_EXIT, "chdir(%s)", ".");
+
+  while ((entry = readdir(dp))) {
+    int fd;
+    char *line;
+    CRONFILE *cfile;
+
+    if (entry->d_name[0] == '.' && (!entry->d_name[1] ||
+          (entry->d_name[1] == '.' && !entry->d_name[2]))) 
+      continue;
+
+    if (!getpwnam(entry->d_name)) {
+      loginfo(LOG_LEVEL7, "ignoring file '%s' (no such user)", entry->d_name);
+      continue;
+    }
+    if ((fd = open(entry->d_name, O_RDONLY)) < 0) continue;
+
+    // one node for each user
+    cfile = xzalloc(sizeof(CRONFILE));
+    cfile->username = xstrdup(entry->d_name);
+
+    for (; (line = get_line(fd)); free(line))
+      parse_line(line, cfile);
+
+    // If there is no job for a cron, remove the VAR list.
+    if (!cfile->job) {
+      VAR *v, *vnode = (VAR *)cfile->var;
+
+      free(cfile->username);
+      if (cfile->mailto) free(cfile->mailto);
+
+      while (vnode && (v = dlist_pop(&vnode))) {
+        free(v->name);
+        free(v->val);
+        free(v);
+      }
+      free(cfile);
+    } else {
+      if (!cfile->mailto) cfile->mailto = cfile->username;
+      dlist_add_nomalloc((struct double_list **)&gclist,
+          (struct double_list *)cfile);
+    }
+    close(fd);
+  }
+  closedir(dp);
+}
+
+/*
+ * Set env variables, if any in the cronfile. Execute given job with the given
+ * SHELL or Default SHELL and send an e-mail with respect to every successfully
+ * completed job (as per the given param 'prog').
+ */
+static void do_fork(CRONFILE *cfile, JOB *job, int fd, char *prog)
+{
+  pid_t pid = vfork();
+
+  if (pid == 0) {
+    VAR *v, *vstart = (VAR *)cfile->var;
+    struct passwd *pwd = getpwnam(cfile->username);
+
+    if (!pwd) loginfo(LOG_LEVEL9, "can't get uid for %s", cfile->username);
+    else {
+      char *file = "/bin/sh";
+
+      if (setenv("USER", pwd->pw_name, 1)) _exit(1);
+      for (v = vstart; v;) {
+        if (!strcmp("SHELL", v->name)) file = v->val;
+        if (setenv(v->name, v->val, 1)) _exit(1);
+        if ((v=v->next) == vstart) break;
+      }
+      if (!getenv("HOME")) {
+        if (setenv("HOME", pwd->pw_dir, 1))
+          _exit(1);
+      }
+      xsetuser(pwd);
+      if (chdir(pwd->pw_dir)) loginfo(LOG_LEVEL9, "chdir(%s)", pwd->pw_dir);
+      if (prog) file = prog;
+      if (TT.flagd) loginfo(LOG_LEVEL5, "child running %s", file);
+
+      if (fd >= 0) {
+        int newfd = prog ? 0 : 1;
+        if (fd != newfd) {
+          dup2(fd, newfd);
+          close(fd);
+        }
+        dup2(1, 2);
+      }
+      setpgrp();
+      execlp(file, file, (prog ? "-ti" : "-c"), (prog ? NULL : job->cmd), (char *) NULL);
+      loginfo(LOG_ERROR, "can't execute '%s' for user %s", file, cfile->username);
+
+      if (!prog) dprintf(1, "Exec failed: %s -c %s\n", file, job->cmd);
+      _exit(EXIT_SUCCESS);
+    }
+  }
+  if (pid < 0) {
+    loginfo(LOG_ERROR, "can't vfork");
+    pid = 0;
+  }
+  if (fd >=0) close(fd);
+  job->pid = pid;
+}
+
+// Send an e-mail for each successfully completed jobs.
+static void sendmail(CRONFILE *cfile, JOB *job)
+{
+  pid_t pid = job->pid;
+  int mailfd;
+  struct stat sb;
+
+  job->pid = 0;
+  if (pid <=0 || job->mailsize <=0) {
+    job->isrunning = 0;
+    job->needstart = 1;
+    return;
+  }
+  snprintf(toybuf, sizeof(toybuf), "/var/spool/cron/cron.%s.%d",
+      cfile->username, (int)pid);
+
+  mailfd = open(toybuf, O_RDONLY);
+  unlink(toybuf);
+  if (mailfd < 0) return;
+
+  if (fstat(mailfd, &sb) == -1 || sb.st_uid != 0 || sb.st_nlink != 0
+      || sb.st_size == job->mailsize || !S_ISREG(sb.st_mode)) {
+    xclose(mailfd);
+    return;
+  }
+  job->mailsize = 0;
+  do_fork(cfile, job, mailfd, "sendmail");
+}
+
+// Count the number of jobs, which are not completed.
+static int count_running_jobs()
+{
+  CRONFILE *cfile = gclist;
+  JOB *job, *jstart;
+  int count = 0;
+
+  while (cfile) {
+    job = jstart = (JOB *)cfile->job;
+    while (job) {
+      int ret;
+
+      if (!job->isrunning || job->pid<=0) goto NEXT_JOB;
+      job->isrunning = 0;
+      ret = waitpid(job->pid, NULL, WNOHANG);
+      if (ret < 0 || ret == job->pid) {
+        sendmail(cfile, job);
+        if (job->pid) count += (job->isrunning=1);
+        else {
+          job->isrunning = 0;
+          job->needstart = 1;
+        }
+      }
+      else count += (job->isrunning=1);
+
+NEXT_JOB:
+      if ((job = job->next) == jstart) break;
+    }
+    if ((cfile = cfile->next) == gclist) break;
+  }
+  return count;
+}
+
+// Execute jobs one by one and prepare for the e-mail sending.
+static void execute_jobs(void)
+{
+  CRONFILE *cfile = gclist;
+  JOB *job, *jstart;
+
+  while (cfile) {
+    job = jstart = (JOB *)cfile->job;
+    while (job) {
+      if (job->needstart) {
+        job->needstart = 0;
+        if (job->pid < 0) {
+          int mailfd = -1;
+
+          job->mailsize = job->pid = 0;
+          snprintf(toybuf, sizeof(toybuf), "/var/spool/cron/cron.%s.%d",
+              cfile->username, getpid());
+          if ((mailfd = open(toybuf, O_CREAT|O_TRUNC|O_WRONLY|O_EXCL|O_APPEND,
+                  0600)) < 0) {
+            loginfo(LOG_ERROR, "can't create mail file %s for user %s, "
+                "discarding output", toybuf, cfile->username);
+          } else {
+            dprintf(mailfd, "To: %s\nSubject: cron: %s\n\n", cfile->mailto, job->cmd);
+            job->mailsize = lseek(mailfd, 0, SEEK_CUR);
+          }
+          do_fork(cfile, job, mailfd, NULL);
+          if (mailfd >= 0) {
+            if (job->pid <= 0) unlink(toybuf);
+            else {
+              char *mailfile = xmprintf("/var/spool/cron/cron.%s.%d",
+                  cfile->username, (int)job->pid);
+              rename(toybuf, mailfile);
+              free(mailfile);
+            }
+          }
+          loginfo(LOG_LEVEL8, "USER %s pid %3d cmd %s", 
+              cfile->username, job->pid, job->cmd);
+          if (job->pid < 0) job->needstart = 1;
+          else job->isrunning = 1;
+        }
+      }
+      if ((job = job->next) == jstart) break;
+    }
+    if ((cfile = cfile->next) == gclist) break;
+  }
+}
+
+// Identify jobs, which needs to be started at the given time interval.
+static void schedule_jobs(time_t ctime, time_t ptime)
+{
+  time_t tm = ptime-ptime%60;
+
+  for (; tm <= ctime; tm += 60) {
+    struct tm *lt;
+    CRONFILE *cfile = gclist;
+    JOB *job, *jstart;
+
+    if (tm <= ptime) continue;
+    lt = localtime(&tm);
+
+    while (cfile) {
+      if (TT.flagd) loginfo(LOG_LEVEL5, "file %s:", cfile->username);
+      if (cfile->invalid) goto NEXT_CRONFILE;
+      job = jstart = (JOB *)cfile->job;
+
+      while (job) {
+        if (TT.flagd) loginfo(LOG_LEVEL5, " line %s", job->cmd);
+
+        if (job->min[lt->tm_min] && job->hour[lt->tm_hour]
+            && (job->dom[lt->tm_mday] || job->dow[lt->tm_wday])
+            && job->mon[lt->tm_mon-1]) {
+          if (TT.flagd)
+            loginfo(LOG_LEVEL5, " job: %d %s\n", (int)job->pid, job->cmd);
+          if (job->pid > 0) {
+            loginfo(LOG_LEVEL8, "user %s: process already running: %s",
+                cfile->username, job->cmd);
+          } else if (!job->pid) {
+            job->pid = -1;
+            job->needstart = 1;
+            job->isrunning = 0;
+          }
+        }
+        if ((job = job->next) == jstart) break;
+      }
+NEXT_CRONFILE:
+      if ((cfile = cfile->next) == gclist) break;
+    }
+  }
+}
+
+void crond_main(void)
+{
+  time_t ctime, ptime;
+  int sleepfor = 60;
+  struct stat sb;
+
+  TT.flagd = (toys.optflags & FLAG_d);
+
+  // Setting default params.
+  if (TT.flagd) TT.loglevel = TT.loglevel_d;
+  if (!(toys.optflags & (FLAG_f | FLAG_b))) toys.optflags |= FLAG_b;
+  if (!(toys.optflags & (FLAG_S | FLAG_L))) toys.optflags |= FLAG_S;
+
+  if ((toys.optflags & FLAG_c)
+      && (TT.crontabs_dir[strlen(TT.crontabs_dir)-1] != '/'))
+    TT.crontabs_dir = xmprintf("%s/", TT.crontabs_dir);
+
+  if (!TT.crontabs_dir) TT.crontabs_dir = xstrdup("/var/spool/cron/crontabs/");
+  if (toys.optflags & FLAG_b) daemon(0,0);
+
+  if (!TT.flagd && !TT.logfile)
+    openlog(toys.which->name, LOG_CONS | LOG_PID, LOG_CRON);
+
+  // Set default shell once.
+  if (setenv("SHELL", "/bin/sh", 1)) error_exit("Can't set default shell");
+  xchdir(TT.crontabs_dir);
+  loginfo(LOG_LEVEL8, "crond started, log level %d", TT.loglevel);
+
+  if (stat(TT.crontabs_dir, &sb)) sb.st_mtime = 0;
+  TT.crontabs_dir_mtime = sb.st_mtime;
+  scan_cronfiles();
+  ctime = time(NULL);
+
+  while (1) {
+    long tdiff;
+
+    ptime = ctime;
+    sleep(sleepfor - (ptime%sleepfor) +1);
+    tdiff =(long) ((ctime = time(NULL)) - ptime);
+
+    if (stat(TT.crontabs_dir, &sb)) sb.st_mtime = 0;
+    if (TT.crontabs_dir_mtime != sb.st_mtime) {
+      TT.crontabs_dir_mtime = sb.st_mtime;
+      scan_cronfiles();
+    }
+
+    if (TT.flagd) loginfo(LOG_LEVEL5, "wakeup diff=%ld\n", tdiff);
+    if (tdiff < -60 * 60 || tdiff > 60 * 60)
+      loginfo(LOG_LEVEL9, "time disparity of %ld minutes detected", tdiff / 60);
+    else if (tdiff > 0) {
+      schedule_jobs(ctime, ptime);
+      execute_jobs();
+      if (count_running_jobs()) sleepfor = 10;
+      else sleepfor = 60;
+    }
+  }
+}
diff --git a/toys/pending/ipcs.c b/toys/pending/ipcs.c
new file mode 100644
index 00000000..d974a519
--- /dev/null
+++ b/toys/pending/ipcs.c
@@ -0,0 +1,450 @@
+/* ipcs.c - provide information on ipc facilities
+ *
+ * Copyright 2014 Sandeep Sharma <sandeep.jack2756@gmail.com>
+ *
+ * see http://pubs.opengroup.org/onlinepubs/9699919799/utilities/ipcs.html
+
+USE_IPCS(NEWTOY(ipcs, "acptulsqmi#", TOYFLAG_USR|TOYFLAG_BIN))
+
+config IPCS
+  bool "ipcs"
+  default n
+  help
+    usage: ipcs [[-smq] -i shmid] | [[-asmq] [-tcplu]]
+
+    -i Show specific resource
+    Resource specification:
+    -a All (default)
+    -m Shared memory segments
+    -q Message queues
+    -s Semaphore arrays
+    Output format:
+    -c Creator
+    -l Limits
+    -p Pid
+    -t Time
+    -u Summary
+*/
+
+#define FOR_ipcs
+#include "toys.h"
+#include <sys/ipc.h>
+#include <sys/shm.h>
+#include <sys/sem.h>
+#include <sys/msg.h>
+
+GLOBALS(
+  int id;
+)
+
+//used many times, so good to paste it
+#define flag(x) (toys.optflags & FLAG_ ## x)
+
+union semun { //man says declare it yourself
+  int              val;
+  struct semid_ds *buf;
+  unsigned short  *array;
+  struct seminfo  *__buf;
+};
+
+static void show_msg_id(void)
+{
+  struct msqid_ds buf;
+  int ret;
+
+  if ((ret = msgctl(TT.id, IPC_STAT, &buf)) < 0) {
+    perror_msg("msgctl");
+    return;
+  }
+
+#define ipcperm buf.msg_perm
+
+  printf("\nMessage Queue msqid=%d\n"
+      "uid=%d\tgid=%d\tcuid=%d\tcgid=%d\tmode=%#o\n",
+      TT.id, ipcperm.uid, ipcperm.gid, ipcperm.cuid, ipcperm.cgid,ipcperm.mode);
+  printf ("cbytes=%ld\tqbytes=%ld\tqnum=%ld\tlspid=%d\tlrpid=%d\n",
+      (long) buf.msg_cbytes, (long) buf.msg_qbytes,
+      (long) buf.msg_qnum, buf.msg_lspid, buf.msg_lrpid);
+
+  printf("send_time=%-26.24s\nrcv_time=%-26.24s\nchange_time=%-26.24s\n\n",
+      buf.msg_stime ? ctime(&buf.msg_stime) : "Not set",
+      buf.msg_rtime ? ctime(&buf.msg_rtime) : "Not set",
+      buf.msg_ctime ? ctime(&buf.msg_ctime) : "Not set");
+#undef ipcperm
+}
+
+static void show_sem_id(void)
+{
+  struct semid_ds buf;
+  union semun n;
+  int ret, i;
+
+  n.buf = &buf;
+  if ((ret = semctl(TT.id, 0, IPC_STAT, n)) < 0) {
+    perror_msg("semctl");
+    return;
+  }
+
+#define ipcperm buf.sem_perm
+  printf("\nSemaphore Array semid=%d\n"
+      "uid=%d\t gid=%d\t cuid=%d\t cgid=%d\n"
+      "mode=%#o, access_perms=%#o\n"
+      "nsems = %ld\n"
+      "otime = %-26.24s\n",
+      TT.id,
+      ipcperm.uid, ipcperm.gid, ipcperm.cuid, ipcperm.cgid,
+      ipcperm.mode, ipcperm.mode & 0777,
+      (long) buf.sem_nsems,
+      buf.sem_otime ? ctime(&buf.sem_otime) : "Not set");
+  printf("ctime = %-26.24s\n"
+      "%-10s %-10s %-10s %-10s %-10s\n",
+      ctime(&buf.sem_ctime),
+      "semnum", "value", "ncount", "zcount", "pid");
+#undef ipcperm
+
+  for (i = 0; i < buf.sem_nsems; i++) {
+    int val, nc, zc, pid;
+    val = semctl(TT.id, i, GETVAL, n);
+    nc = semctl(TT.id, i, GETNCNT, n);
+    zc = semctl(TT.id, i, GETZCNT, n);
+    pid = semctl(TT.id, i, GETPID, n);
+    if (val < 0 || nc < 0 || zc < 0 || pid < 0)
+      perror_exit("semctl");
+    printf("%-10d %-10d %-10d %-10d %-10d\n", i, val, nc, zc, pid);
+  }
+  xputc('\n');
+}
+
+static void show_shm_id(void)
+{
+  struct shmid_ds buf;
+  int ret;
+
+  if ((ret = shmctl(TT.id, IPC_STAT, &buf)) < 0) {
+    perror_msg("shmctl");
+    return;
+  }
+
+#define ipcperm buf.shm_perm
+
+  printf("\nShared memory Segment shmid=%d\n"
+      "uid=%d\tgid=%d\tcuid=%d\tcgid=%d\n"
+      "mode=%#o\taccess_perms=%#o\n"
+      "bytes=%ld\tlpid=%d\tcpid=%d\tnattch=%ld\n",
+      TT.id,
+      ipcperm.uid, ipcperm.gid, ipcperm.cuid, ipcperm.cgid,
+      ipcperm.mode, (ipcperm.mode & 0777),
+      (long) buf.shm_segsz, buf.shm_lpid, buf.shm_cpid,
+      (long) buf.shm_nattch);
+  printf("att_time=%-26.24s\n",
+      buf.shm_atime ? ctime(&buf.shm_atime) : "Not set");
+  printf("det_time=%-26.24s\n",
+      buf.shm_dtime ? ctime(&buf.shm_dtime) : "Not set");
+  printf("change_time=%-26.24s\n\n", ctime(&buf.shm_ctime));
+#undef ipcperm
+}
+
+static void shm_array(void)
+{
+  struct shm_info shm_buf;
+  struct shminfo ipc_buf;
+  struct shmid_ds buf;
+  int max_nr, i, shmid;
+  struct passwd *pw;
+  struct group *gr;
+
+  if ((max_nr = shmctl(0, SHM_INFO, (struct shmid_ds*)&shm_buf)) < 0) {
+    perror_msg("kernel not configured for shared memory");
+    return;
+  }
+
+  if (flag(u)) {
+    printf("------ Shared Memory Status --------\n");
+    printf("segments allocated %d\n"
+        "pages allocated %ld\n"
+        "pages resident  %ld\n"
+        "pages swapped   %ld\n"
+        "Swap performance: %ld attempts\t%ld successes\n",
+        shm_buf.used_ids,
+        shm_buf.shm_tot,
+        shm_buf.shm_rss,
+        shm_buf.shm_swp,
+        shm_buf.swap_attempts, shm_buf.swap_successes);
+    return;
+  }
+  if (flag(l)) {
+    if ((shmctl(0, 3, (struct shmid_ds*)&ipc_buf)) < 0) return; //IPC_INFO
+    printf("------ Shared Memory Limits --------\n");
+    printf("max number of segments = %lu\n"
+        "max seg size (kbytes) = %lu\n"
+        "max total shared memory (pages) = %lu\n"
+        "min seg size (bytes) = %lu\n",
+        (unsigned long) ipc_buf.shmmni,
+        (unsigned long) (ipc_buf.shmmax >> 10),
+        (unsigned long) ipc_buf.shmall,
+        (unsigned long) ipc_buf.shmmin);
+    return;
+  }
+
+  if (flag(t)) {
+    printf("------ Shared Memory Attach/Detach/Change Times --------\n");
+    printf("%-10s %-10s %-20s %-20s %-20s\n",
+        "shmid", "owner", "attached", "detached", "changed");
+  } else if (flag(p)) {
+    printf("------ Shared Memory Creator/Last-op --------\n");
+    printf("%-10s %-10s %-10s %-10s\n",
+        "shmid", "owner", "cpid", "lpid");
+  } else if (flag(c)) {
+    printf("------ Shared Memory Segment Creators/Owners --------\n");
+    printf("%-10s %-10s %-10s %-10s %-10s %-10s\n",
+        "shmid", "perms", "cuid", "cgid", "uid", "gid");
+  } else {
+    printf("------ Shared Memory Segments --------\n");
+    printf("%-10s %-10s %-10s %-10s %-10s %-10s %-12s\n",
+        "key", "shmid", "owner", "perms", "bytes", "nattch",
+        "status");
+  }
+
+  for (i = 0; i <= max_nr; i++) {
+    if ((shmid = shmctl(i, SHM_STAT, &buf)) < 0 ) continue;
+    if (flag(t)) {
+      if ((pw = getpwuid(buf.shm_perm.uid)))
+        printf("%-10d %-10.10s", shmid, pw->pw_name);
+      else printf("%-10d %-10.10d", shmid, buf.shm_perm.uid);
+      printf(" %-20.16s", buf.shm_atime
+          ? ctime(&buf.shm_atime) + 4 : "Not set");
+      printf(" %-20.16s", buf.shm_dtime
+          ? ctime(&buf.shm_dtime) + 4 : "Not set");
+      printf(" %-20.16s\n", buf.shm_ctime
+          ? ctime(&buf.shm_ctime) + 4 : "Not set");
+    } else if (flag(p)) {
+      if ((pw = getpwuid(buf.shm_perm.uid)))
+        printf("%-10d %-10.10s", shmid, pw->pw_name);
+      else printf("%-10d %-10.10d", shmid, buf.shm_perm.uid);
+      printf(" %-10d %-10d\n", buf.shm_cpid, buf.shm_lpid);
+    } else if (flag(c)) {
+      printf("%-10d %-10o", shmid, buf.shm_perm.mode & 0777);
+      if ((pw = getpwuid(buf.shm_perm.cuid))) printf(" %-10s", pw->pw_name);
+      else printf(" %-10d", buf.shm_perm.cuid);
+      if ((gr = getgrgid(buf.shm_perm.cgid))) printf(" %-10s", gr->gr_name);
+      else printf(" %-10d", buf.shm_perm.cgid);
+      if ((pw = getpwuid(buf.shm_perm.uid))) printf(" %-10s", pw->pw_name);
+      else printf(" %-10d", buf.shm_perm.uid);
+      if ((gr = getgrgid(buf.shm_perm.gid))) printf(" %-10s\n", gr->gr_name);
+      else printf(" %-10d\n", buf.shm_perm.gid);
+    } else {
+      printf("0x%08x ", buf.shm_perm.__key);
+      if ((pw = getpwuid(buf.shm_perm.uid)))
+        printf("%-10d %-10.10s", shmid, pw->pw_name);
+      else printf("%-10d %-10.10d", shmid, buf.shm_perm.uid);
+      printf(" %-10o %-10lu %-10ld %-6s %-6s\n", buf.shm_perm.mode & 0777,
+          (unsigned long) buf.shm_segsz,
+          (long) buf.shm_nattch,
+          buf.shm_perm.mode & SHM_DEST ? "dest" : " ",
+          buf.shm_perm.mode & SHM_LOCKED ? "locked" : " ");
+    }
+  }
+}
+
+static void sem_array(void)
+{
+  struct seminfo info_buf;
+  struct semid_ds buf;
+  union semun u;
+  int max_nr, i,semid;
+  struct passwd *pw;
+  struct group *gr;
+
+  u.array = (unsigned short *)&info_buf;
+  if ((max_nr = semctl(0, 0, SEM_INFO, u)) < 0) {
+    perror_msg("kernel is not configured for semaphores");
+    return;
+  }
+
+
+  if (flag(u)) {
+    printf("------ Semaphore Status --------\n");
+    printf("used arrays = %d\n"
+        "allocated semaphores = %d\n",
+        info_buf.semusz, info_buf.semaem);
+    return;
+  } 
+  if (flag(l)) {
+    printf("------ Semaphore Limits --------\n");
+    u.array = (unsigned short *)&info_buf;
+    if ((semctl(0, 0, 3, u)) < 0) //IPC_INFO
+      return;
+    printf("max number of arrays = %d\n"
+        "max semaphores per array = %d\n"
+        "max semaphores system wide = %d\n"
+        "max ops per semop call = %d\n"
+        "semaphore max value = %d\n",
+        info_buf.semmni,
+        info_buf.semmsl,
+        info_buf.semmns, info_buf.semopm, info_buf.semvmx);
+    return;
+  }
+
+  if (flag(t)) {
+    printf("------ Semaphore Operation/Change Times --------\n");
+    printf("%-8s %-10s %-26.24s %-26.24s\n",
+        "shmid", "owner", "last-op", "last-changed");
+  } else if (flag(c)) {
+    printf("------ Semaphore %s --------\n", "Arrays Creators/Owners");
+    printf("%-10s %-10s %-10s %-10s %-10s %-10s\n",
+        "semid", "perms", "cuid", "cgid", "uid", "gid");
+
+  } else if (flag(p)){
+    return;
+  } else {
+    printf("------ Semaphore %s --------\n", "Arrays");
+    printf("%-10s %-10s %-10s %-10s %-10s\n",
+        "key", "semid", "owner", "perms", "nsems");
+  }
+
+  for (i = 0; i <= max_nr; i++) {
+    u.buf = &buf;
+    if ((semid = semctl(i, 0, SEM_STAT, u)) < 0) continue;
+    pw = getpwuid(buf.sem_perm.uid);
+    if (flag(t)) {
+      if (pw) printf("%-8d %-10.10s", semid, pw->pw_name);
+      else printf("%-8d %-10d", semid, buf.sem_perm.uid);
+
+      printf("  %-26.24s", buf.sem_otime
+          ? ctime(&buf.sem_otime) : "Not set");
+      printf(" %-26.24s\n", buf.sem_ctime
+          ? ctime(&buf.sem_ctime) : "Not set");
+    } else if (flag(c)) {
+      printf("%-10d %-10o", semid, buf.sem_perm.mode & 0777);
+      if ((pw = getpwuid(buf.sem_perm.cuid))) printf(" %-10s", pw->pw_name);
+      else printf(" %-10d", buf.sem_perm.cuid);
+      if ((gr = getgrgid(buf.sem_perm.cgid))) printf(" %-10s", gr->gr_name);
+      else printf(" %-10d", buf.sem_perm.cgid);
+      if ((pw = getpwuid(buf.sem_perm.uid))) printf(" %-10s", pw->pw_name);
+      else printf(" %-10d", buf.sem_perm.uid);
+      if ((gr = getgrgid(buf.sem_perm.gid))) printf(" %-10s\n", gr->gr_name);
+      else printf(" %-10d\n", buf.sem_perm.gid);
+    } else {
+      printf("0x%08x ", buf.sem_perm.__key);
+      if (pw) printf("%-10d %-10.9s", semid, pw->pw_name);
+      else printf("%-10d %-9d", semid, buf.sem_perm.uid);
+      printf(" %-10o %-10ld\n", buf.sem_perm.mode & 0777,
+          (long) buf.sem_nsems);
+    }
+  }
+}
+
+static void msg_array(void)
+{
+  struct msginfo info_buf;
+  struct msqid_ds buf;
+  int max_nr, i, msqid;
+  struct passwd *pw;
+  struct group *gr;
+
+  if ((max_nr = msgctl(0, MSG_INFO, (struct msqid_ds*)&info_buf)) < 0) {
+    perror_msg("kernel not configured for message queue");
+    return;
+  }
+
+  if (flag(u)) {
+    printf("------ Message%s --------\n", "s: Status");
+    printf("allocated queues = %d\n"
+        "used headers = %d\n"
+        "used space = %d bytes\n",
+        info_buf.msgpool, info_buf.msgmap, info_buf.msgtql);
+    return;
+  }
+  if (flag(l)) {
+    if ((msgctl(0, 3, (struct msqid_ds*)&info_buf)) < 0) return; //IPC_INFO
+    printf("------ Messages: Limits --------\n");
+    printf("max queues system wide = %d\n"
+        "max size of message (bytes) = %d\n"
+        "default max size of queue (bytes) = %d\n",
+        info_buf.msgmni, info_buf.msgmax, info_buf.msgmnb);
+    return;
+  }
+
+  if (flag(t)) {
+    printf("------ Message%s --------\n", " Queues Send/Recv/Change Times");
+    printf("%-8s %-10s %-20s %-20s %-20s\n",
+        "msqid", "owner", "send", "recv", "change");
+  } else if (flag(p)) {
+    printf("------ Message%s --------\n", " Queues PIDs");
+    printf("%-10s %-10s %-10s %-10s\n",
+        "msqid", "owner", "lspid", "lrpid");
+  } else if (flag(c)) {
+    printf("------ Message%s --------\n", " Queues: Creators/Owners");
+    printf("%-10s %-10s %-10s %-10s %-10s %-10s\n",
+        "msqid", "perms", "cuid", "cgid", "uid", "gid");
+  } else {
+    printf("------ Message%s --------\n", " Queues");
+    printf("%-10s %-10s %-10s %-10s %-12s %-12s\n",
+        "key", "msqid", "owner", "perms", "used-bytes", "messages");
+  }
+
+  for (i = 0; i <= max_nr; i++) {
+    if ((msqid = msgctl(i, MSG_STAT, &buf)) < 0 ) continue;
+    pw = getpwuid(buf.msg_perm.uid);
+    if (flag(t)) {
+      if (pw) printf("%-8d %-10.10s", msqid, pw->pw_name);
+      else printf("%-8d %-10d", msqid, buf.msg_perm.uid);
+      printf(" %-20.16s", buf.msg_stime
+          ? ctime(&buf.msg_stime) + 4 : "Not set");
+      printf(" %-20.16s", buf.msg_rtime
+          ? ctime(&buf.msg_rtime) + 4 : "Not set");
+      printf(" %-20.16s\n", buf.msg_ctime
+          ? ctime(&buf.msg_ctime) + 4 : "Not set");
+    } else if (flag(p)) {
+      if (pw) printf("%-8d %-10.10s", msqid, pw->pw_name);
+      else printf("%-8d %-10d", msqid, buf.msg_perm.uid);
+      printf("  %5d     %5d\n", buf.msg_lspid, buf.msg_lrpid);
+    } else if (flag(c)) {
+      printf("%-10d %-10o", msqid, buf.msg_perm.mode & 0777);
+      if ((pw = getpwuid(buf.msg_perm.cuid))) printf(" %-10s", pw->pw_name);
+      else printf(" %-10d", buf.msg_perm.cuid);
+      if ((gr = getgrgid(buf.msg_perm.cgid))) printf(" %-10s", gr->gr_name);
+      else printf(" %-10d", buf.msg_perm.cgid);
+      if ((pw = getpwuid(buf.msg_perm.uid))) printf(" %-10s", pw->pw_name);
+      else printf(" %-10d", buf.msg_perm.uid);
+      if ((gr = getgrgid(buf.msg_perm.gid))) printf(" %-10s\n", gr->gr_name);
+      else printf(" %-10d\n", buf.msg_perm.gid);
+    } else {
+      printf("0x%08x ", buf.msg_perm.__key);
+      if (pw) printf("%-10d %-10.10s", msqid, pw->pw_name);
+      else printf("%-10d %-10d", msqid, buf.msg_perm.uid);
+      printf(" %-10o %-12ld %-12ld\n", buf.msg_perm.mode & 0777,
+          (long) buf.msg_cbytes, (long) buf.msg_qnum);
+    }
+  }
+}
+
+void ipcs_main(void)
+{
+  if (flag(i)) {
+    if (flag(m)) show_shm_id();
+    else if (flag(s)) show_sem_id();
+    else if (flag(q)) show_msg_id();
+    else {
+      toys.exithelp++;
+      error_exit(NULL);
+    }
+    return;
+  }
+
+  if (!(flag(m) || flag(s) || flag(q)) || flag(a)) toys.optflags |= (FLAG_m|FLAG_s|FLAG_q);
+
+  xputc('\n');
+  if (flag(m)) {
+    shm_array();
+    xputc('\n');
+  }
+  if (flag(s)) {
+    sem_array();
+    xputc('\n');
+  }
+  if (flag(q)) {
+    msg_array();
+    xputc('\n');
+  }
+}