/*-*- Mode: C; c-basic-offset: 8; indent-tabs-mode: nil -*-*/

/***
  This file is part of systemd.

  Copyright 2010 Lennart Poettering

  systemd is free software; you can redistribute it and/or modify it
  under the terms of the GNU Lesser General Public License as published by
  the Free Software Foundation; either version 2.1 of the License, or
  (at your option) any later version.

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

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

#include <errno.h>
#include <signal.h>
#include <string.h>
#include <stdlib.h>
#include <sys/types.h>

#include "cgroup-util.h"
#include "set.h"
#include "path-util.h"

int cg_enumerate_processes(const char *controller, const char *path, FILE **_f) {
        _cleanup_free_ char *fs = NULL;
        FILE *f;
        int r;

        assert(_f);

        r = cg_get_path(controller, path, "cgroup.procs", &fs);
        if (r < 0)
                return r;

        f = fopen(fs, "re");
        if (!f)
                return -errno;

        *_f = f;
        return 0;
}

int cg_read_pid(FILE *f, pid_t *_pid) {
        unsigned long ul;

        /* Note that the cgroup.procs might contain duplicates! See
         * cgroups.txt for details. */

        assert(f);
        assert(_pid);

        errno = 0;
        if (fscanf(f, "%lu", &ul) != 1) {

                if (feof(f))
                        return 0;

                return errno ? -errno : -EIO;
        }

        if (ul <= 0)
                return -EIO;

        *_pid = (pid_t) ul;
        return 1;
}

int cg_kill(const char *controller, const char *path, int sig, bool sigcont, bool ignore_self, Set *s) {
        _cleanup_set_free_ Set *allocated_set = NULL;
        bool done = false;
        int r, ret = 0;
        pid_t my_pid;

        assert(sig >= 0);

        /* This goes through the tasks list and kills them all. This
         * is repeated until no further processes are added to the
         * tasks list, to properly handle forking processes */

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

        my_pid = getpid();

        do {
                _cleanup_fclose_ FILE *f = NULL;
                pid_t pid = 0;
                done = true;

                r = cg_enumerate_processes(controller, path, &f);
                if (r < 0) {
                        if (ret >= 0 && r != -ENOENT)
                                return r;

                        return ret;
                }

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

                        if (ignore_self && pid == my_pid)
                                continue;

                        if (set_get(s, LONG_TO_PTR(pid)) == LONG_TO_PTR(pid))
                                continue;

                        /* If we haven't killed this process yet, kill
                         * it */
                        if (kill(pid, sig) < 0) {
                                if (ret >= 0 && errno != ESRCH)
                                        ret = -errno;
                        } else if (ret == 0) {

                                if (sigcont)
                                        kill(pid, SIGCONT);

                                ret = 1;
                        }

                        done = false;

                        r = set_put(s, LONG_TO_PTR(pid));
                        if (r < 0) {
                                if (ret >= 0)
                                        return r;

                                return ret;
                        }
                }

                if (r < 0) {
                        if (ret >= 0)
                                return r;

                        return ret;
                }

                /* To avoid racing against processes which fork
                 * quicker than we can kill them we repeat this until
                 * no new pids need to be killed. */

        } while (!done);

        return ret;
}

static const char *normalize_controller(const char *controller) {

        assert(controller);

        if (streq(controller, SYSTEMD_CGROUP_CONTROLLER))
                return "systemd";
        else if (startswith(controller, "name="))
                return controller + 5;
        else
                return controller;
}

static int join_path(const char *controller, const char *path, const char *suffix, char **fs) {
        char *t = NULL;

        if (!isempty(controller)) {
                if (!isempty(path) && !isempty(suffix))
                        t = strjoin("/sys/fs/cgroup/", controller, "/", path, "/", suffix, NULL);
                else if (!isempty(path))
                        t = strjoin("/sys/fs/cgroup/", controller, "/", path, NULL);
                else if (!isempty(suffix))
                        t = strjoin("/sys/fs/cgroup/", controller, "/", suffix, NULL);
                else
                        t = strappend("/sys/fs/cgroup/", controller);
        } else {
                if (!isempty(path) && !isempty(suffix))
                        t = strjoin(path, "/", suffix, NULL);
                else if (!isempty(path))
                        t = strdup(path);
                else
                        return -EINVAL;
        }

        if (!t)
                return -ENOMEM;

        path_kill_slashes(t);

        *fs = t;
        return 0;
}

int cg_get_path(const char *controller, const char *path, const char *suffix, char **fs) {
        const char *p;
        static __thread bool good = false;

        assert(fs);

        if (controller && !cg_controller_is_valid(controller, true))
                return -EINVAL;

        if (_unlikely_(!good)) {
                int r;

                r = path_is_mount_point("/sys/fs/cgroup", false);
                if (r <= 0)
                        return r < 0 ? r : -ENOENT;

                /* Cache this to save a few stat()s */
                good = true;
        }

        p = controller ? normalize_controller(controller) : NULL;

        return join_path(p, path, suffix, fs);
}


#define CONTROLLER_VALID                        \
        "0123456789"                            \
        "abcdefghijklmnopqrstuvwxyz"            \
        "ABCDEFGHIJKLMNOPQRSTUVWXYZ"            \
        "_"

bool cg_controller_is_valid(const char *p, bool allow_named) {
        const char *t, *s;

        if (!p)
                return false;

        if (allow_named) {
                s = startswith(p, "name=");
                if (s)
                        p = s;
        }

        if (*p == 0 || *p == '_')
                return false;

        for (t = p; *t; t++)
                if (!strchr(CONTROLLER_VALID, *t))
                        return false;

        if (t - p > FILENAME_MAX)
                return false;

        return true;
}