./tools/notsd-move

6 files changed, 1649 insertions, 0 deletions

diff --git a/src/systemd-cgtop/GNUmakefile b/src/systemd-cgtop/GNUmakefile
new file mode 120000
index 0000000000..54fdd42278
--- /dev/null
+++ b/src/systemd-cgtop/GNUmakefile
@@ -0,0 +1 @@
+../../GNUmakefile
\ No newline at end of file
diff --git a/src/systemd-cgtop/Makefile b/src/systemd-cgtop/Makefile
new file mode 100644
index 0000000000..abebe7f3d0
--- /dev/null
+++ b/src/systemd-cgtop/Makefile
@@ -0,0 +1,33 @@
+#  -*- Mode: makefile; indent-tabs-mode: t -*-
+#
+#  This file is part of systemd.
+#
+#  Copyright 2010-2012 Lennart Poettering
+#  Copyright 2010-2012 Kay Sievers
+#  Copyright 2013 Zbigniew Jędrzejewski-Szmek
+#  Copyright 2013 David Strauss
+#  Copyright 2016 Luke Shumaker
+#
+#  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 $(dir $(lastword $(MAKEFILE_LIST)))/../../config.mk
+include $(topsrcdir)/build-aux/Makefile.head.mk
+
+bin_PROGRAMS += systemd-cgtop
+systemd_cgtop_SOURCES = \
+	src/cgtop/cgtop.c
+
+systemd_cgtop_LDADD = \
+	libsystemd-shared.la
+
+include $(topsrcdir)/build-aux/Makefile.tail.mk
diff --git a/src/systemd-cgtop/cgtop.c b/src/systemd-cgtop/cgtop.c
new file mode 100644
index 0000000000..53c8f1b848
--- /dev/null
+++ b/src/systemd-cgtop/cgtop.c
@@ -0,0 +1,1159 @@
+/***
+  This file is part of systemd.
+
+  Copyright 2012 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 <alloca.h>
+#include <errno.h>
+#include <getopt.h>
+#include <signal.h>
+#include <stdint.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+
+#include <systemd/sd-bus.h>
+
+#include "sd-bus/bus-error.h"
+#include "sd-bus/bus-util.h"
+#include "systemd-basic/alloc-util.h"
+#include "systemd-basic/cgroup-util.h"
+#include "systemd-basic/fd-util.h"
+#include "systemd-basic/fileio.h"
+#include "systemd-basic/hashmap.h"
+#include "systemd-basic/parse-util.h"
+#include "systemd-basic/path-util.h"
+#include "systemd-basic/process-util.h"
+#include "systemd-basic/stdio-util.h"
+#include "systemd-basic/terminal-util.h"
+#include "systemd-basic/unit-name.h"
+#include "systemd-basic/util.h"
+
+typedef struct Group {
+        char *path;
+
+        bool n_tasks_valid:1;
+        bool cpu_valid:1;
+        bool memory_valid:1;
+        bool io_valid:1;
+
+        uint64_t n_tasks;
+
+        unsigned cpu_iteration;
+        nsec_t cpu_usage;
+        nsec_t cpu_timestamp;
+        double cpu_fraction;
+
+        uint64_t memory;
+
+        unsigned io_iteration;
+        uint64_t io_input, io_output;
+        nsec_t io_timestamp;
+        uint64_t io_input_bps, io_output_bps;
+} Group;
+
+static unsigned arg_depth = 3;
+static unsigned arg_iterations = (unsigned) -1;
+static bool arg_batch = false;
+static bool arg_raw = false;
+static usec_t arg_delay = 1*USEC_PER_SEC;
+static char* arg_machine = NULL;
+static char* arg_root = NULL;
+static bool arg_recursive = true;
+
+static enum {
+        COUNT_PIDS,
+        COUNT_USERSPACE_PROCESSES,
+        COUNT_ALL_PROCESSES,
+} arg_count = COUNT_PIDS;
+
+static enum {
+        ORDER_PATH,
+        ORDER_TASKS,
+        ORDER_CPU,
+        ORDER_MEMORY,
+        ORDER_IO,
+} arg_order = ORDER_CPU;
+
+static enum {
+        CPU_PERCENT,
+        CPU_TIME,
+} arg_cpu_type = CPU_PERCENT;
+
+static void group_free(Group *g) {
+        assert(g);
+
+        free(g->path);
+        free(g);
+}
+
+static void group_hashmap_clear(Hashmap *h) {
+        Group *g;
+
+        while ((g = hashmap_steal_first(h)))
+                group_free(g);
+}
+
+static void group_hashmap_free(Hashmap *h) {
+        group_hashmap_clear(h);
+        hashmap_free(h);
+}
+
+static const char *maybe_format_bytes(char *buf, size_t l, bool is_valid, uint64_t t) {
+        if (!is_valid)
+                return "-";
+        if (arg_raw) {
+                snprintf(buf, l, "%jd", t);
+                return buf;
+        }
+        return format_bytes(buf, l, t);
+}
+
+static int process(
+                const char *controller,
+                const char *path,
+                Hashmap *a,
+                Hashmap *b,
+                unsigned iteration,
+                Group **ret) {
+
+        Group *g;
+        int r;
+
+        assert(controller);
+        assert(path);
+        assert(a);
+
+        g = hashmap_get(a, path);
+        if (!g) {
+                g = hashmap_get(b, path);
+                if (!g) {
+                        g = new0(Group, 1);
+                        if (!g)
+                                return -ENOMEM;
+
+                        g->path = strdup(path);
+                        if (!g->path) {
+                                group_free(g);
+                                return -ENOMEM;
+                        }
+
+                        r = hashmap_put(a, g->path, g);
+                        if (r < 0) {
+                                group_free(g);
+                                return r;
+                        }
+                } else {
+                        r = hashmap_move_one(a, b, path);
+                        if (r < 0)
+                                return r;
+
+                        g->cpu_valid = g->memory_valid = g->io_valid = g->n_tasks_valid = false;
+                }
+        }
+
+        if (streq(controller, SYSTEMD_CGROUP_CONTROLLER) && IN_SET(arg_count, COUNT_ALL_PROCESSES, COUNT_USERSPACE_PROCESSES)) {
+                _cleanup_fclose_ FILE *f = NULL;
+                pid_t pid;
+
+                r = cg_enumerate_processes(controller, path, &f);
+                if (r == -ENOENT)
+                        return 0;
+                if (r < 0)
+                        return r;
+
+                g->n_tasks = 0;
+                while (cg_read_pid(f, &pid) > 0) {
+
+                        if (arg_count == COUNT_USERSPACE_PROCESSES && is_kernel_thread(pid) > 0)
+                                continue;
+
+                        g->n_tasks++;
+                }
+
+                if (g->n_tasks > 0)
+                        g->n_tasks_valid = true;
+
+        } else if (streq(controller, "pids") && arg_count == COUNT_PIDS) {
+                _cleanup_free_ char *p = NULL, *v = NULL;
+
+                r = cg_get_path(controller, path, "pids.current", &p);
+                if (r < 0)
+                        return r;
+
+                r = read_one_line_file(p, &v);
+                if (r == -ENOENT)
+                        return 0;
+                if (r < 0)
+                        return r;
+
+                r = safe_atou64(v, &g->n_tasks);
+                if (r < 0)
+                        return r;
+
+                if (g->n_tasks > 0)
+                        g->n_tasks_valid = true;
+
+        } else if (streq(controller, "cpu") || streq(controller, "cpuacct")) {
+                _cleanup_free_ char *p = NULL, *v = NULL;
+                uint64_t new_usage;
+                nsec_t timestamp;
+
+                if (cg_all_unified() > 0) {
+                        const char *keys[] = { "usage_usec", NULL };
+                        _cleanup_free_ char *val = NULL;
+
+                        if (!streq(controller, "cpu"))
+                                return 0;
+
+                        r = cg_get_keyed_attribute("cpu", path, "cpu.stat", keys, &val);
+                        if (r == -ENOENT)
+                                return 0;
+                        if (r < 0)
+                                return r;
+
+                        r = safe_atou64(val, &new_usage);
+                        if (r < 0)
+                                return r;
+
+                        new_usage *= NSEC_PER_USEC;
+                } else {
+                        if (!streq(controller, "cpuacct"))
+                                return 0;
+
+                        r = cg_get_path(controller, path, "cpuacct.usage", &p);
+                        if (r < 0)
+                                return r;
+
+                        r = read_one_line_file(p, &v);
+                        if (r == -ENOENT)
+                                return 0;
+                        if (r < 0)
+                                return r;
+
+                        r = safe_atou64(v, &new_usage);
+                        if (r < 0)
+                                return r;
+                }
+
+                timestamp = now_nsec(CLOCK_MONOTONIC);
+
+                if (g->cpu_iteration == iteration - 1 &&
+                    (nsec_t) new_usage > g->cpu_usage) {
+
+                        nsec_t x, y;
+
+                        x = timestamp - g->cpu_timestamp;
+                        if (x < 1)
+                                x = 1;
+
+                        y = (nsec_t) new_usage - g->cpu_usage;
+                        g->cpu_fraction = (double) y / (double) x;
+                        g->cpu_valid = true;
+                }
+
+                g->cpu_usage = (nsec_t) new_usage;
+                g->cpu_timestamp = timestamp;
+                g->cpu_iteration = iteration;
+
+        } else if (streq(controller, "memory")) {
+                _cleanup_free_ char *p = NULL, *v = NULL;
+
+                if (cg_all_unified() <= 0)
+                        r = cg_get_path(controller, path, "memory.usage_in_bytes", &p);
+                else
+                        r = cg_get_path(controller, path, "memory.current", &p);
+                if (r < 0)
+                        return r;
+
+                r = read_one_line_file(p, &v);
+                if (r == -ENOENT)
+                        return 0;
+                if (r < 0)
+                        return r;
+
+                r = safe_atou64(v, &g->memory);
+                if (r < 0)
+                        return r;
+
+                if (g->memory > 0)
+                        g->memory_valid = true;
+
+        } else if ((streq(controller, "io") && cg_all_unified() > 0) ||
+                   (streq(controller, "blkio") && cg_all_unified() <= 0)) {
+                _cleanup_fclose_ FILE *f = NULL;
+                _cleanup_free_ char *p = NULL;
+                bool unified = cg_all_unified() > 0;
+                uint64_t wr = 0, rd = 0;
+                nsec_t timestamp;
+
+                r = cg_get_path(controller, path, unified ? "io.stat" : "blkio.io_service_bytes", &p);
+                if (r < 0)
+                        return r;
+
+                f = fopen(p, "re");
+                if (!f) {
+                        if (errno == ENOENT)
+                                return 0;
+                        return -errno;
+                }
+
+                for (;;) {
+                        char line[LINE_MAX], *l;
+                        uint64_t k, *q;
+
+                        if (!fgets(line, sizeof(line), f))
+                                break;
+
+                        /* Trim and skip the device */
+                        l = strstrip(line);
+                        l += strcspn(l, WHITESPACE);
+                        l += strspn(l, WHITESPACE);
+
+                        if (unified) {
+                                while (!isempty(l)) {
+                                        if (sscanf(l, "rbytes=%" SCNu64, &k))
+                                                rd += k;
+                                        else if (sscanf(l, "wbytes=%" SCNu64, &k))
+                                                wr += k;
+
+                                        l += strcspn(l, WHITESPACE);
+                                        l += strspn(l, WHITESPACE);
+                                }
+                        } else {
+                                if (first_word(l, "Read")) {
+                                        l += 4;
+                                        q = &rd;
+                                } else if (first_word(l, "Write")) {
+                                        l += 5;
+                                        q = &wr;
+                                } else
+                                        continue;
+
+                                l += strspn(l, WHITESPACE);
+                                r = safe_atou64(l, &k);
+                                if (r < 0)
+                                        continue;
+
+                                *q += k;
+                        }
+                }
+
+                timestamp = now_nsec(CLOCK_MONOTONIC);
+
+                if (g->io_iteration == iteration - 1) {
+                        uint64_t x, yr, yw;
+
+                        x = (uint64_t) (timestamp - g->io_timestamp);
+                        if (x < 1)
+                                x = 1;
+
+                        if (rd > g->io_input)
+                                yr = rd - g->io_input;
+                        else
+                                yr = 0;
+
+                        if (wr > g->io_output)
+                                yw = wr - g->io_output;
+                        else
+                                yw = 0;
+
+                        if (yr > 0 || yw > 0) {
+                                g->io_input_bps = (yr * 1000000000ULL) / x;
+                                g->io_output_bps = (yw * 1000000000ULL) / x;
+                                g->io_valid = true;
+                        }
+                }
+
+                g->io_input = rd;
+                g->io_output = wr;
+                g->io_timestamp = timestamp;
+                g->io_iteration = iteration;
+        }
+
+        if (ret)
+                *ret = g;
+
+        return 0;
+}
+
+static int refresh_one(
+                const char *controller,
+                const char *path,
+                Hashmap *a,
+                Hashmap *b,
+                unsigned iteration,
+                unsigned depth,
+                Group **ret) {
+
+        _cleanup_closedir_ DIR *d = NULL;
+        Group *ours = NULL;
+        int r;
+
+        assert(controller);
+        assert(path);
+        assert(a);
+
+        if (depth > arg_depth)
+                return 0;
+
+        r = process(controller, path, a, b, iteration, &ours);
+        if (r < 0)
+                return r;
+
+        r = cg_enumerate_subgroups(controller, path, &d);
+        if (r == -ENOENT)
+                return 0;
+        if (r < 0)
+                return r;
+
+        for (;;) {
+                _cleanup_free_ char *fn = NULL, *p = NULL;
+                Group *child = NULL;
+
+                r = cg_read_subgroup(d, &fn);
+                if (r < 0)
+                        return r;
+                if (r == 0)
+                        break;
+
+                p = strjoin(path, "/", fn, NULL);
+                if (!p)
+                        return -ENOMEM;
+
+                path_kill_slashes(p);
+
+                r = refresh_one(controller, p, a, b, iteration, depth + 1, &child);
+                if (r < 0)
+                        return r;
+
+                if (arg_recursive &&
+                    IN_SET(arg_count, COUNT_ALL_PROCESSES, COUNT_USERSPACE_PROCESSES) &&
+                    child &&
+                    child->n_tasks_valid &&
+                    streq(controller, SYSTEMD_CGROUP_CONTROLLER)) {
+
+                        /* Recursively sum up processes */
+
+                        if (ours->n_tasks_valid)
+                                ours->n_tasks += child->n_tasks;
+                        else {
+                                ours->n_tasks = child->n_tasks;
+                                ours->n_tasks_valid = true;
+                        }
+                }
+        }
+
+        if (ret)
+                *ret = ours;
+
+        return 1;
+}
+
+static int refresh(const char *root, Hashmap *a, Hashmap *b, unsigned iteration) {
+        int r;
+
+        assert(a);
+
+        r = refresh_one(SYSTEMD_CGROUP_CONTROLLER, root, a, b, iteration, 0, NULL);
+        if (r < 0)
+                return r;
+        r = refresh_one("cpu", root, a, b, iteration, 0, NULL);
+        if (r < 0)
+                return r;
+        r = refresh_one("cpuacct", root, a, b, iteration, 0, NULL);
+        if (r < 0)
+                return r;
+        r = refresh_one("memory", root, a, b, iteration, 0, NULL);
+        if (r < 0)
+                return r;
+        r = refresh_one("io", root, a, b, iteration, 0, NULL);
+        if (r < 0)
+                return r;
+        r = refresh_one("blkio", root, a, b, iteration, 0, NULL);
+        if (r < 0)
+                return r;
+        r = refresh_one("pids", root, a, b, iteration, 0, NULL);
+        if (r < 0)
+                return r;
+
+        return 0;
+}
+
+static int group_compare(const void*a, const void *b) {
+        const Group *x = *(Group**)a, *y = *(Group**)b;
+
+        if (arg_order != ORDER_TASKS || arg_recursive) {
+                /* Let's make sure that the parent is always before
+                 * the child. Except when ordering by tasks and
+                 * recursive summing is off, since that is actually
+                 * not accumulative for all children. */
+
+                if (path_startswith(y->path, x->path))
+                        return -1;
+                if (path_startswith(x->path, y->path))
+                        return 1;
+        }
+
+        switch (arg_order) {
+
+        case ORDER_PATH:
+                break;
+
+        case ORDER_CPU:
+                if (arg_cpu_type == CPU_PERCENT) {
+                        if (x->cpu_valid && y->cpu_valid) {
+                                if (x->cpu_fraction > y->cpu_fraction)
+                                        return -1;
+                                else if (x->cpu_fraction < y->cpu_fraction)
+                                        return 1;
+                        } else if (x->cpu_valid)
+                                return -1;
+                        else if (y->cpu_valid)
+                                return 1;
+                } else {
+                        if (x->cpu_usage > y->cpu_usage)
+                                return -1;
+                        else if (x->cpu_usage < y->cpu_usage)
+                                return 1;
+                }
+
+                break;
+
+        case ORDER_TASKS:
+                if (x->n_tasks_valid && y->n_tasks_valid) {
+                        if (x->n_tasks > y->n_tasks)
+                                return -1;
+                        else if (x->n_tasks < y->n_tasks)
+                                return 1;
+                } else if (x->n_tasks_valid)
+                        return -1;
+                else if (y->n_tasks_valid)
+                        return 1;
+
+                break;
+
+        case ORDER_MEMORY:
+                if (x->memory_valid && y->memory_valid) {
+                        if (x->memory > y->memory)
+                                return -1;
+                        else if (x->memory < y->memory)
+                                return 1;
+                } else if (x->memory_valid)
+                        return -1;
+                else if (y->memory_valid)
+                        return 1;
+
+                break;
+
+        case ORDER_IO:
+                if (x->io_valid && y->io_valid) {
+                        if (x->io_input_bps + x->io_output_bps > y->io_input_bps + y->io_output_bps)
+                                return -1;
+                        else if (x->io_input_bps + x->io_output_bps < y->io_input_bps + y->io_output_bps)
+                                return 1;
+                } else if (x->io_valid)
+                        return -1;
+                else if (y->io_valid)
+                        return 1;
+        }
+
+        return path_compare(x->path, y->path);
+}
+
+static void display(Hashmap *a) {
+        Iterator i;
+        Group *g;
+        Group **array;
+        signed path_columns;
+        unsigned rows, n = 0, j, maxtcpu = 0, maxtpath = 3; /* 3 for ellipsize() to work properly */
+        char buffer[MAX3(21, FORMAT_BYTES_MAX, FORMAT_TIMESPAN_MAX)];
+
+        assert(a);
+
+        if (!terminal_is_dumb())
+                fputs(ANSI_HOME_CLEAR, stdout);
+
+        array = alloca(sizeof(Group*) * hashmap_size(a));
+
+        HASHMAP_FOREACH(g, a, i)
+                if (g->n_tasks_valid || g->cpu_valid || g->memory_valid || g->io_valid)
+                        array[n++] = g;
+
+        qsort_safe(array, n, sizeof(Group*), group_compare);
+
+        /* Find the longest names in one run */
+        for (j = 0; j < n; j++) {
+                unsigned cputlen, pathtlen;
+
+                format_timespan(buffer, sizeof(buffer), (usec_t) (array[j]->cpu_usage / NSEC_PER_USEC), 0);
+                cputlen = strlen(buffer);
+                maxtcpu = MAX(maxtcpu, cputlen);
+
+                pathtlen = strlen(array[j]->path);
+                maxtpath = MAX(maxtpath, pathtlen);
+        }
+
+        if (arg_cpu_type == CPU_PERCENT)
+                xsprintf(buffer, "%6s", "%CPU");
+        else
+                xsprintf(buffer, "%*s", maxtcpu, "CPU Time");
+
+        rows = lines();
+        if (rows <= 10)
+                rows = 10;
+
+        if (on_tty()) {
+                const char *on, *off;
+
+                path_columns = columns() - 36 - strlen(buffer);
+                if (path_columns < 10)
+                        path_columns = 10;
+
+                on = ansi_highlight_underline();
+                off = ansi_underline();
+
+                printf("%s%s%-*s%s %s%7s%s %s%s%s %s%8s%s %s%8s%s %s%8s%s%s\n",
+                       ansi_underline(),
+                       arg_order == ORDER_PATH ? on : "", path_columns, "Control Group",
+                       arg_order == ORDER_PATH ? off : "",
+                       arg_order == ORDER_TASKS ? on : "", arg_count == COUNT_PIDS ? "Tasks" : arg_count == COUNT_USERSPACE_PROCESSES ? "Procs" : "Proc+",
+                       arg_order == ORDER_TASKS ? off : "",
+                       arg_order == ORDER_CPU ? on : "", buffer,
+                       arg_order == ORDER_CPU ? off : "",
+                       arg_order == ORDER_MEMORY ? on : "", "Memory",
+                       arg_order == ORDER_MEMORY ? off : "",
+                       arg_order == ORDER_IO ? on : "", "Input/s",
+                       arg_order == ORDER_IO ? off : "",
+                       arg_order == ORDER_IO ? on : "", "Output/s",
+                       arg_order == ORDER_IO ? off : "",
+                       ansi_normal());
+        } else
+                path_columns = maxtpath;
+
+        for (j = 0; j < n; j++) {
+                _cleanup_free_ char *ellipsized = NULL;
+                const char *path;
+
+                if (on_tty() && j + 6 > rows)
+                        break;
+
+                g = array[j];
+
+                path = isempty(g->path) ? "/" : g->path;
+                ellipsized = ellipsize(path, path_columns, 33);
+                printf("%-*s", path_columns, ellipsized ?: path);
+
+                if (g->n_tasks_valid)
+                        printf(" %7" PRIu64, g->n_tasks);
+                else
+                        fputs("       -", stdout);
+
+                if (arg_cpu_type == CPU_PERCENT) {
+                        if (g->cpu_valid)
+                                printf(" %6.1f", g->cpu_fraction*100);
+                        else
+                                fputs("      -", stdout);
+                } else
+                        printf(" %*s", maxtcpu, format_timespan(buffer, sizeof(buffer), (usec_t) (g->cpu_usage / NSEC_PER_USEC), 0));
+
+                printf(" %8s", maybe_format_bytes(buffer, sizeof(buffer), g->memory_valid, g->memory));
+                printf(" %8s", maybe_format_bytes(buffer, sizeof(buffer), g->io_valid, g->io_input_bps));
+                printf(" %8s", maybe_format_bytes(buffer, sizeof(buffer), g->io_valid, g->io_output_bps));
+
+                putchar('\n');
+        }
+}
+
+static void help(void) {
+        printf("%s [OPTIONS...] [CGROUP]\n\n"
+               "Show top control groups by their resource usage.\n\n"
+               "  -h --help           Show this help\n"
+               "     --version        Show package version\n"
+               "  -p --order=path     Order by path\n"
+               "  -t --order=tasks    Order by number of tasks/processes\n"
+               "  -c --order=cpu      Order by CPU load (default)\n"
+               "  -m --order=memory   Order by memory load\n"
+               "  -i --order=io       Order by IO load\n"
+               "  -r --raw            Provide raw (not human-readable) numbers\n"
+               "     --cpu=percentage Show CPU usage as percentage (default)\n"
+               "     --cpu=time       Show CPU usage as time\n"
+               "  -P                  Count userspace processes instead of tasks (excl. kernel)\n"
+               "  -k                  Count all processes instead of tasks (incl. kernel)\n"
+               "     --recursive=BOOL Sum up process count recursively\n"
+               "  -d --delay=DELAY    Delay between updates\n"
+               "  -n --iterations=N   Run for N iterations before exiting\n"
+               "  -b --batch          Run in batch mode, accepting no input\n"
+               "     --depth=DEPTH    Maximum traversal depth (default: %u)\n"
+               "  -M --machine=       Show container\n"
+               , program_invocation_short_name, arg_depth);
+}
+
+static int parse_argv(int argc, char *argv[]) {
+
+        enum {
+                ARG_VERSION = 0x100,
+                ARG_DEPTH,
+                ARG_CPU_TYPE,
+                ARG_ORDER,
+                ARG_RECURSIVE,
+        };
+
+        static const struct option options[] = {
+                { "help",         no_argument,       NULL, 'h'           },
+                { "version",      no_argument,       NULL, ARG_VERSION   },
+                { "delay",        required_argument, NULL, 'd'           },
+                { "iterations",   required_argument, NULL, 'n'           },
+                { "batch",        no_argument,       NULL, 'b'           },
+                { "raw",          no_argument,       NULL, 'r'           },
+                { "depth",        required_argument, NULL, ARG_DEPTH     },
+                { "cpu",          optional_argument, NULL, ARG_CPU_TYPE  },
+                { "order",        required_argument, NULL, ARG_ORDER     },
+                { "recursive",    required_argument, NULL, ARG_RECURSIVE },
+                { "machine",      required_argument, NULL, 'M'           },
+                {}
+        };
+
+        bool recursive_unset = false;
+        int c, r;
+
+        assert(argc >= 1);
+        assert(argv);
+
+        while ((c = getopt_long(argc, argv, "hptcmin:brd:kPM:", options, NULL)) >= 0)
+
+                switch (c) {
+
+                case 'h':
+                        help();
+                        return 0;
+
+                case ARG_VERSION:
+                        return version();
+
+                case ARG_CPU_TYPE:
+                        if (optarg) {
+                                if (streq(optarg, "time"))
+                                        arg_cpu_type = CPU_TIME;
+                                else if (streq(optarg, "percentage"))
+                                        arg_cpu_type = CPU_PERCENT;
+                                else {
+                                        log_error("Unknown argument to --cpu=: %s", optarg);
+                                        return -EINVAL;
+                                }
+                        } else
+                                arg_cpu_type = CPU_TIME;
+
+                        break;
+
+                case ARG_DEPTH:
+                        r = safe_atou(optarg, &arg_depth);
+                        if (r < 0) {
+                                log_error("Failed to parse depth parameter.");
+                                return -EINVAL;
+                        }
+
+                        break;
+
+                case 'd':
+                        r = parse_sec(optarg, &arg_delay);
+                        if (r < 0 || arg_delay <= 0) {
+                                log_error("Failed to parse delay parameter.");
+                                return -EINVAL;
+                        }
+
+                        break;
+
+                case 'n':
+                        r = safe_atou(optarg, &arg_iterations);
+                        if (r < 0) {
+                                log_error("Failed to parse iterations parameter.");
+                                return -EINVAL;
+                        }
+
+                        break;
+
+                case 'b':
+                        arg_batch = true;
+                        break;
+
+                case 'r':
+                        arg_raw = true;
+                        break;
+
+                case 'p':
+                        arg_order = ORDER_PATH;
+                        break;
+
+                case 't':
+                        arg_order = ORDER_TASKS;
+                        break;
+
+                case 'c':
+                        arg_order = ORDER_CPU;
+                        break;
+
+                case 'm':
+                        arg_order = ORDER_MEMORY;
+                        break;
+
+                case 'i':
+                        arg_order = ORDER_IO;
+                        break;
+
+                case ARG_ORDER:
+                        if (streq(optarg, "path"))
+                                arg_order = ORDER_PATH;
+                        else if (streq(optarg, "tasks"))
+                                arg_order = ORDER_TASKS;
+                        else if (streq(optarg, "cpu"))
+                                arg_order = ORDER_CPU;
+                        else if (streq(optarg, "memory"))
+                                arg_order = ORDER_MEMORY;
+                        else if (streq(optarg, "io"))
+                                arg_order = ORDER_IO;
+                        else {
+                                log_error("Invalid argument to --order=: %s", optarg);
+                                return -EINVAL;
+                        }
+                        break;
+
+                case 'k':
+                        arg_count = COUNT_ALL_PROCESSES;
+                        break;
+
+                case 'P':
+                        arg_count = COUNT_USERSPACE_PROCESSES;
+                        break;
+
+                case ARG_RECURSIVE:
+                        r = parse_boolean(optarg);
+                        if (r < 0) {
+                                log_error("Failed to parse --recursive= argument: %s", optarg);
+                                return r;
+                        }
+
+                        arg_recursive = r;
+                        recursive_unset = r == 0;
+                        break;
+
+                case 'M':
+                        arg_machine = optarg;
+                        break;
+
+                case '?':
+                        return -EINVAL;
+
+                default:
+                        assert_not_reached("Unhandled option");
+                }
+
+        if (optind == argc-1) {
+                if (arg_machine) {
+                        log_error("Specifying a control group path together with the -M option is not allowed");
+                        return -EINVAL;
+                }
+                arg_root = argv[optind];
+        } else if (optind < argc) {
+                log_error("Too many arguments.");
+                return -EINVAL;
+        }
+
+        if (recursive_unset && arg_count == COUNT_PIDS) {
+                log_error("Non-recursive counting is only supported when counting processes, not tasks. Use -P or -k.");
+                return -EINVAL;
+        }
+
+        return 1;
+}
+
+static const char* counting_what(void) {
+        if (arg_count == COUNT_PIDS)
+                return "tasks";
+        else if (arg_count == COUNT_ALL_PROCESSES)
+                return "all processes (incl. kernel)";
+        else
+                return "userspace processes (excl. kernel)";
+}
+
+static int get_cgroup_root(char **ret) {
+        _cleanup_(sd_bus_error_free) sd_bus_error error = SD_BUS_ERROR_NULL;
+        _cleanup_(sd_bus_flush_close_unrefp) sd_bus *bus = NULL;
+        _cleanup_free_ char *unit = NULL, *path = NULL;
+        const char *m;
+        int r;
+
+        if (arg_root) {
+                char *aux;
+
+                aux = strdup(arg_root);
+                if (!aux)
+                        return log_oom();
+
+                *ret = aux;
+                return 0;
+        }
+
+        if (!arg_machine) {
+                r = cg_get_root_path(ret);
+                if (r < 0)
+                        return log_error_errno(r, "Failed to get root control group path: %m");
+
+                return 0;
+        }
+
+        m = strjoina("/run/systemd/machines/", arg_machine);
+        r = parse_env_file(m, NEWLINE, "SCOPE", &unit, NULL);
+        if (r < 0)
+                return log_error_errno(r, "Failed to load machine data: %m");
+
+        path = unit_dbus_path_from_name(unit);
+        if (!path)
+                return log_oom();
+
+        r = bus_connect_transport_systemd(BUS_TRANSPORT_LOCAL, NULL, false, &bus);
+        if (r < 0)
+                return log_error_errno(r, "Failed to create bus connection: %m");
+
+        r = sd_bus_get_property_string(
+                        bus,
+                        "org.freedesktop.systemd1",
+                        path,
+                        unit_dbus_interface_from_name(unit),
+                        "ControlGroup",
+                        &error,
+                        ret);
+        if (r < 0)
+                return log_error_errno(r, "Failed to query unit control group path: %s", bus_error_message(&error, r));
+
+        return 0;
+}
+
+int main(int argc, char *argv[]) {
+        int r;
+        Hashmap *a = NULL, *b = NULL;
+        unsigned iteration = 0;
+        usec_t last_refresh = 0;
+        bool quit = false, immediate_refresh = false;
+        _cleanup_free_ char *root = NULL;
+        CGroupMask mask;
+
+        log_parse_environment();
+        log_open();
+
+        r = cg_mask_supported(&mask);
+        if (r < 0) {
+                log_error_errno(r, "Failed to determine supported controllers: %m");
+                goto finish;
+        }
+
+        arg_count = (mask & CGROUP_MASK_PIDS) ? COUNT_PIDS : COUNT_USERSPACE_PROCESSES;
+
+        r = parse_argv(argc, argv);
+        if (r <= 0)
+                goto finish;
+
+        r = get_cgroup_root(&root);
+        if (r < 0) {
+                log_error_errno(r, "Failed to get root control group path: %m");
+                goto finish;
+        }
+
+        a = hashmap_new(&string_hash_ops);
+        b = hashmap_new(&string_hash_ops);
+        if (!a || !b) {
+                r = log_oom();
+                goto finish;
+        }
+
+        signal(SIGWINCH, columns_lines_cache_reset);
+
+        if (arg_iterations == (unsigned) -1)
+                arg_iterations = on_tty() ? 0 : 1;
+
+        while (!quit) {
+                Hashmap *c;
+                usec_t t;
+                char key;
+                char h[FORMAT_TIMESPAN_MAX];
+
+                t = now(CLOCK_MONOTONIC);
+
+                if (t >= last_refresh + arg_delay || immediate_refresh) {
+
+                        r = refresh(root, a, b, iteration++);
+                        if (r < 0) {
+                                log_error_errno(r, "Failed to refresh: %m");
+                                goto finish;
+                        }
+
+                        group_hashmap_clear(b);
+
+                        c = a;
+                        a = b;
+                        b = c;
+
+                        last_refresh = t;
+                        immediate_refresh = false;
+                }
+
+                display(b);
+
+                if (arg_iterations && iteration >= arg_iterations)
+                        break;
+
+                if (!on_tty()) /* non-TTY: Empty newline as delimiter between polls */
+                        fputs("\n", stdout);
+                fflush(stdout);
+
+                if (arg_batch)
+                        (void) usleep(last_refresh + arg_delay - t);
+                else {
+                        r = read_one_char(stdin, &key, last_refresh + arg_delay - t, NULL);
+                        if (r == -ETIMEDOUT)
+                                continue;
+                        if (r < 0) {
+                                log_error_errno(r, "Couldn't read key: %m");
+                                goto finish;
+                        }
+                }
+
+                if (on_tty()) { /* TTY: Clear any user keystroke */
+                        fputs("\r \r", stdout);
+                        fflush(stdout);
+                }
+
+                if (arg_batch)
+                        continue;
+
+                switch (key) {
+
+                case ' ':
+                        immediate_refresh = true;
+                        break;
+
+                case 'q':
+                        quit = true;
+                        break;
+
+                case 'p':
+                        arg_order = ORDER_PATH;
+                        break;
+
+                case 't':
+                        arg_order = ORDER_TASKS;
+                        break;
+
+                case 'c':
+                        arg_order = ORDER_CPU;
+                        break;
+
+                case 'm':
+                        arg_order = ORDER_MEMORY;
+                        break;
+
+                case 'i':
+                        arg_order = ORDER_IO;
+                        break;
+
+                case '%':
+                        arg_cpu_type = arg_cpu_type == CPU_TIME ? CPU_PERCENT : CPU_TIME;
+                        break;
+
+                case 'k':
+                        arg_count = arg_count != COUNT_ALL_PROCESSES ? COUNT_ALL_PROCESSES : COUNT_PIDS;
+                        fprintf(stdout, "\nCounting: %s.", counting_what());
+                        fflush(stdout);
+                        sleep(1);
+                        break;
+
+                case 'P':
+                        arg_count = arg_count != COUNT_USERSPACE_PROCESSES ? COUNT_USERSPACE_PROCESSES : COUNT_PIDS;
+                        fprintf(stdout, "\nCounting: %s.", counting_what());
+                        fflush(stdout);
+                        sleep(1);
+                        break;
+
+                case 'r':
+                        if (arg_count == COUNT_PIDS)
+                                fprintf(stdout, "\n\aCannot toggle recursive counting, not available in task counting mode.");
+                        else {
+                                arg_recursive = !arg_recursive;
+                                fprintf(stdout, "\nRecursive process counting: %s", yes_no(arg_recursive));
+                        }
+                        fflush(stdout);
+                        sleep(1);
+                        break;
+
+                case '+':
+                        if (arg_delay < USEC_PER_SEC)
+                                arg_delay += USEC_PER_MSEC*250;
+                        else
+                                arg_delay += USEC_PER_SEC;
+
+                        fprintf(stdout, "\nIncreased delay to %s.", format_timespan(h, sizeof(h), arg_delay, 0));
+                        fflush(stdout);
+                        sleep(1);
+                        break;
+
+                case '-':
+                        if (arg_delay <= USEC_PER_MSEC*500)
+                                arg_delay = USEC_PER_MSEC*250;
+                        else if (arg_delay < USEC_PER_MSEC*1250)
+                                arg_delay -= USEC_PER_MSEC*250;
+                        else
+                                arg_delay -= USEC_PER_SEC;
+
+                        fprintf(stdout, "\nDecreased delay to %s.", format_timespan(h, sizeof(h), arg_delay, 0));
+                        fflush(stdout);
+                        sleep(1);
+                        break;
+
+                case '?':
+                case 'h':
+
+#define ON ANSI_HIGHLIGHT
+#define OFF ANSI_NORMAL
+
+                        fprintf(stdout,
+                                "\t<" ON "p" OFF "> By path; <" ON "t" OFF "> By tasks/procs; <" ON "c" OFF "> By CPU; <" ON "m" OFF "> By memory; <" ON "i" OFF "> By I/O\n"
+                                "\t<" ON "+" OFF "> Inc. delay; <" ON "-" OFF "> Dec. delay; <" ON "%%" OFF "> Toggle time; <" ON "SPACE" OFF "> Refresh\n"
+                                "\t<" ON "P" OFF "> Toggle count userspace processes; <" ON "k" OFF "> Toggle count all processes\n"
+                                "\t<" ON "r" OFF "> Count processes recursively; <" ON "q" OFF "> Quit");
+                        fflush(stdout);
+                        sleep(3);
+                        break;
+
+                default:
+                        if (key < ' ')
+                                fprintf(stdout, "\nUnknown key '\\x%x'. Ignoring.", key);
+                        else
+                                fprintf(stdout, "\nUnknown key '%c'. Ignoring.", key);
+                        fflush(stdout);
+                        sleep(1);
+                        break;
+                }
+        }
+
+        r = 0;
+
+finish:
+        group_hashmap_free(a);
+        group_hashmap_free(b);
+
+        return r < 0 ? EXIT_FAILURE : EXIT_SUCCESS;
+}
diff --git a/src/systemd-cgtop/systemd-cgtop.completion.bash b/src/systemd-cgtop/systemd-cgtop.completion.bash
new file mode 100644
index 0000000000..f1ed22fd55
--- /dev/null
+++ b/src/systemd-cgtop/systemd-cgtop.completion.bash
@@ -0,0 +1,62 @@
+# systemd-cgtop(1) completion                  -*- shell-script -*-
+#
+# This file is part of systemd.
+#
+# Copyright 2014 Thomas H.P. Andersen
+#
+# 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
+# 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/>.
+
+__contains_word() {
+        local w word=$1; shift
+        for w in "$@"; do
+                [[ $w = "$word" ]] && return
+        done
+}
+
+__get_machines() {
+        local a b
+        machinectl list --no-legend --no-pager | { while read a b; do echo " $a"; done; };
+}
+
+_systemd_cgtop() {
+        local cur=${COMP_WORDS[COMP_CWORD]} prev=${COMP_WORDS[COMP_CWORD-1]}
+        local comps
+
+        local -A OPTS=(
+               [STANDALONE]='-h --help --version -p -t -c -m -i -b --batch -r --raw -k -P'
+               [ARG]='--cpu --depth -M --machine --recursive -n --iterations -d --delay --order'
+               )
+
+        _init_completion || return
+
+        if __contains_word "$prev" ${OPTS[ARG]}; then
+                case $prev in
+                        --machine|-M)
+                                comps=$( __get_machines )
+                        ;;
+                        --recursive)
+                                comps='yes no'
+                        ;;
+                        --order)
+                                comps='path tasks cpu memory io'
+                        ;;
+                esac
+                COMPREPLY=( $(compgen -W '$comps' -- "$cur") )
+                return 0
+        fi
+
+        COMPREPLY=( $(compgen -W '${OPTS[*]}' -- "$cur") )
+}
+
+complete -F _systemd_cgtop systemd-cgtop
diff --git a/src/systemd-cgtop/systemd-cgtop.completion.zsh b/src/systemd-cgtop/systemd-cgtop.completion.zsh
new file mode 100644
index 0000000000..f6e1b2422a
--- /dev/null
+++ b/src/systemd-cgtop/systemd-cgtop.completion.zsh
@@ -0,0 +1,17 @@
+#compdef systemd-cgtop
+
+local curcontext="$curcontext" state lstate line
+_arguments \
+    {-h,--help}'[Show this help]' \
+    '--version[Print version and exit]' \
+    '(-c -m -i -t)-p[Order by path]' \
+    '(-c -p -m -i)-t[Order by number of tasks]' \
+    '(-m -p -i -t)-c[Order by CPU load]' \
+    '(-c -p -i -t)-m[Order by memory load]' \
+    '(-c -m -p -t)-i[Order by IO load]' \
+    {-d+,--delay=}'[Specify delay]:delay:' \
+    {-n+,--iterations=}'[Run for N iterations before exiting]:number of iterations:' \
+    {-b,--batch}'[Run in batch mode, accepting no input]' \
+    '--depth=[Maximum traversal depth]:maximum depth:'
+
+#vim: set ft=zsh sw=4 ts=4 et
diff --git a/src/systemd-cgtop/systemd-cgtop.xml b/src/systemd-cgtop/systemd-cgtop.xml
new file mode 100644
index 0000000000..be13631239
--- /dev/null
+++ b/src/systemd-cgtop/systemd-cgtop.xml
@@ -0,0 +1,377 @@
+<?xml version='1.0'?> <!--*- Mode: nxml; nxml-child-indent: 2; indent-tabs-mode: nil -*-->
+<!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN"
+  "http://www.oasis-open.org/docbook/xml/4.2/docbookx.dtd">
+
+<!--
+  This file is part of systemd.
+
+  Copyright 2012 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/>.
+-->
+
+<refentry id="systemd-cgtop"
+    xmlns:xi="http://www.w3.org/2001/XInclude">
+
+  <refentryinfo>
+    <title>systemd-cgtop</title>
+    <productname>systemd</productname>
+
+    <authorgroup>
+      <author>
+        <contrib>Developer</contrib>
+        <firstname>Lennart</firstname>
+        <surname>Poettering</surname>
+        <email>lennart@poettering.net</email>
+      </author>
+    </authorgroup>
+  </refentryinfo>
+
+  <refmeta>
+    <refentrytitle>systemd-cgtop</refentrytitle>
+    <manvolnum>1</manvolnum>
+  </refmeta>
+
+  <refnamediv>
+    <refname>systemd-cgtop</refname>
+    <refpurpose>Show top control groups by their resource usage</refpurpose>
+  </refnamediv>
+
+  <refsynopsisdiv>
+    <cmdsynopsis>
+      <command>systemd-cgtop</command>
+      <arg choice="opt" rep="repeat">OPTIONS</arg>
+      <arg choice="opt">GROUP</arg>
+    </cmdsynopsis>
+  </refsynopsisdiv>
+
+  <refsect1>
+    <title>Description</title>
+
+    <para><command>systemd-cgtop</command> shows the top control
+    groups of the local Linux control group hierarchy, ordered by
+    their CPU, memory, or disk I/O load. The display is refreshed in
+    regular intervals (by default every 1s), similar in style to
+    <citerefentry project='man-pages'><refentrytitle>top</refentrytitle><manvolnum>1</manvolnum></citerefentry>.
+    If a control group path is specified, shows only the services of
+    the specified control group.</para>
+
+    <para>If <command>systemd-cgtop</command> is not connected to a
+    tty, no column headers are printed and the default is to only run
+    one iteration.  The <varname>--iterations=</varname> argument, if
+    given, is honored. This mode is suitable for scripting.</para>
+
+    <para>Resource usage is only accounted for control groups in the
+    relevant hierarchy, i.e. CPU usage is only accounted for control
+    groups in the <literal>cpuacct</literal> hierarchy, memory usage
+    only for those in <literal>memory</literal> and disk I/O usage for
+    those in <literal>blkio</literal>. If resource monitoring for
+    these resources is required, it is recommended to add the
+    <varname>CPUAccounting=1</varname>,
+    <varname>MemoryAccounting=1</varname> and
+    <varname>BlockIOAccounting=1</varname> settings in the unit files
+    in question. See
+    <citerefentry><refentrytitle>systemd.resource-control</refentrytitle><manvolnum>5</manvolnum></citerefentry>
+    for details.</para>
+
+    <para>The CPU load value can be between 0 and 100 times the number of
+    processors the system has. For example, if the system has 8 processors,
+    the CPU load value is going to be between 0% and 800%. The number of
+    processors can be found in <literal>/proc/cpuinfo</literal>.</para>
+
+    <para>To emphasize this: unless
+    <literal>CPUAccounting=1</literal>,
+    <literal>MemoryAccounting=1</literal> and
+    <literal>BlockIOAccounting=1</literal> are enabled for the
+    services in question, no resource accounting will be available for
+    system services and the data shown by
+    <command>systemd-cgtop</command> will be incomplete.</para>
+  </refsect1>
+
+  <refsect1>
+    <title>Options</title>
+
+    <para>The following options are understood:</para>
+
+    <variablelist>
+      <varlistentry>
+        <term><option>-p</option></term>
+        <term><option>--order=path</option></term>
+
+        <listitem><para>Order by control group
+        path name.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><option>-t</option></term>
+        <term><option>--order=tasks</option></term>
+
+        <listitem><para>Order by number of tasks/processes in the control group.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><option>-c</option></term>
+        <term><option>--order=cpu</option></term>
+
+        <listitem><para>Order by CPU load.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><option>-m</option></term>
+        <term><option>--order=memory</option></term>
+
+        <listitem><para>Order by memory usage.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><option>-i</option></term>
+        <term><option>--order=io</option></term>
+
+        <listitem><para>Order by disk I/O load.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><option>-b</option></term>
+        <term><option>--batch</option></term>
+
+        <listitem><para>Run in "batch" mode: do not accept input and
+        run until the iteration limit set with
+        <option>--iterations=</option> is exhausted or until killed.
+        This mode could be useful for sending output from
+        <command>systemd-cgtop</command> to other programs or to a
+        file.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><option>-r</option></term>
+        <term><option>--raw</option></term>
+
+	<listitem><para>Format byte counts (as in memory usage and I/O metrics)
+	with raw numeric values rather than human-readable
+        numbers.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><option>--cpu=percentage</option></term>
+        <term><option>--cpu=time</option></term>
+
+        <listitem><para>Controls whether the CPU usage is shown as
+        percentage or time. By default, the CPU usage is shown as
+        percentage. This setting may also be toggled at runtime by
+        pressing the <keycap>%</keycap> key.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><option>-P</option></term>
+
+        <listitem><para>Count only userspace processes instead of all
+        tasks. By default, all tasks are counted: each kernel thread
+        and each userspace thread individually. With this setting,
+        kernel threads are excluded from the counting and each
+        userspace process only counts as one, regardless how many
+        threads it consists of. This setting may also be toggled at
+        runtime by pressing the <keycap>P</keycap> key. This option
+        may not be combined with
+        <option>-k</option>.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><option>-k</option></term>
+
+        <listitem><para>Count only userspace processes and kernel
+        threads instead of all tasks. By default, all tasks are
+        counted: each kernel thread and each userspace thread
+        individually. With this setting, kernel threads are included in
+        the counting and each userspace process only counts as on one,
+        regardless how many threads it consists of. This setting may
+        also be toggled at runtime by pressing the <keycap>k</keycap>
+        key. This option may not be combined with
+        <option>-P</option>.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><option>--recursive=</option></term>
+
+        <listitem><para>Controls whether the number of processes shown
+        for a control group shall include all processes that are
+        contained in any of the child control groups as well. Takes a
+        boolean argument, which defaults to <literal>yes</literal>. If
+        enabled, the processes in child control groups are included, if
+        disabled, only the processes in the control group itself are
+        counted. This setting may also be toggled at runtime by
+        pressing the <keycap>r</keycap> key. Note that this setting
+        only applies to process counting, i.e. when the
+        <option>-P</option> or <option>-k</option> options are
+        used. It has not effect if all tasks are counted, in which
+        case the counting is always recursive.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><option>-n</option></term>
+        <term><option>--iterations=</option></term>
+
+        <listitem><para>Perform only this many iterations. A value of
+        0 indicates that the program should run
+        indefinitely.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><option>-d</option></term>
+        <term><option>--delay=</option></term>
+
+        <listitem><para>Specify refresh delay in seconds (or if one of
+        <literal>ms</literal>, <literal>us</literal>,
+        <literal>min</literal> is specified as unit in this time
+        unit). This setting may also be increased and decreased at
+        runtime by pressing the <keycap>+</keycap> and
+        <keycap>-</keycap> keys.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><option>--depth=</option></term>
+
+        <listitem><para>Maximum control group tree traversal depth.
+        Specifies how deep <command>systemd-cgtop</command> shall
+        traverse the control group hierarchies. If 0 is specified,
+        only the root group is monitored. For 1, only the first level
+        of control groups is monitored, and so on. Defaults to
+        3.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><option>-M <replaceable>MACHINE</replaceable></option></term>
+        <term><option>--machine=<replaceable>MACHINE</replaceable></option></term>
+
+        <listitem><para>Limit control groups shown to the part
+        corresponding to the container
+        <replaceable>MACHINE</replaceable>.
+        This option may not be used when a control group path is specified.</para></listitem>
+      </varlistentry>
+
+      <xi:include href="standard-options.xml" xpointer="help" />
+      <xi:include href="standard-options.xml" xpointer="version" />
+    </variablelist>
+
+  </refsect1>
+
+  <refsect1>
+    <title>Keys</title>
+
+    <para><command>systemd-cgtop</command> is an interactive tool and
+    may be controlled via user input using the following keys:</para>
+
+    <variablelist>
+      <varlistentry>
+        <term><keycap>h</keycap></term>
+
+        <listitem><para>Shows a short help text.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><keycap function="space"/></term>
+
+        <listitem><para>Immediately refresh output.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><keycap>q</keycap></term>
+
+        <listitem><para>Terminate the program.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><keycap>p</keycap></term>
+        <term><keycap>t</keycap></term>
+        <term><keycap>c</keycap></term>
+        <term><keycap>m</keycap></term>
+        <term><keycap>i</keycap></term>
+
+        <listitem><para>Sort the control groups by path, number of
+        tasks, CPU load, memory usage, or I/O load, respectively.  This
+        setting may also be controlled using the
+        <option>--order=</option> command line
+        switch.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><keycap>%</keycap></term>
+
+        <listitem><para>Toggle between showing CPU time as time or
+        percentage. This setting may also be controlled using the
+        <option>--cpu=</option> command line switch.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><keycap>+</keycap></term>
+        <term><keycap>-</keycap></term>
+
+        <listitem><para>Increase or decrease refresh delay,
+        respectively. This setting may also be controlled using the
+        <option>--delay=</option> command line
+        switch.</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><keycap>P</keycap></term>
+
+        <listitem><para>Toggle between counting all tasks, or only
+        userspace processes. This setting may also be controlled using
+        the <option>-P</option> command line switch (see
+        above).</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><keycap>k</keycap></term>
+
+        <listitem><para>Toggle between counting all tasks, or only
+        userspace processes and kernel threads. This setting may also
+        be controlled using the <option>-k</option> command line
+        switch (see above).</para></listitem>
+      </varlistentry>
+
+      <varlistentry>
+        <term><keycap>r</keycap></term>
+
+        <listitem><para>Toggle between recursively including or
+        excluding processes in child control groups in control group
+        process counts. This setting may also be controlled using the
+        <option>--recursive=</option> command line switch. This key is
+        not available if all tasks are counted, it is only available
+        if processes are counted, as enabled with the
+        <keycap>P</keycap> or <keycap>k</keycap>
+        keys.</para></listitem>
+      </varlistentry>
+
+    </variablelist>
+  </refsect1>
+
+  <refsect1>
+    <title>Exit status</title>
+
+    <para>On success, 0 is returned, a non-zero failure code
+    otherwise.</para>
+  </refsect1>
+
+  <refsect1>
+    <title>See Also</title>
+    <para>
+      <citerefentry><refentrytitle>systemd</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
+      <citerefentry><refentrytitle>systemctl</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
+      <citerefentry><refentrytitle>systemd-cgls</refentrytitle><manvolnum>1</manvolnum></citerefentry>,
+      <citerefentry><refentrytitle>systemd.resource-control</refentrytitle><manvolnum>5</manvolnum></citerefentry>,
+      <citerefentry project='man-pages'><refentrytitle>top</refentrytitle><manvolnum>1</manvolnum></citerefentry>
+    </para>
+  </refsect1>
+
+</refentry>