Systemd/src/cgtop/cgtop.c

1100 lines
36 KiB
C
Raw Normal View History

2012-01-22 18:21:15 +01:00
/*-*- Mode: C; c-basic-offset: 8; indent-tabs-mode: nil -*-*/
/***
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
2012-01-22 18:21:15 +01:00
(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.
2012-01-22 18:21:15 +01:00
You should have received a copy of the GNU Lesser General Public License
2012-01-22 18:21:15 +01:00
along with systemd; If not, see <http://www.gnu.org/licenses/>.
***/
#include <alloca.h>
#include <errno.h>
2012-01-22 18:21:15 +01:00
#include <getopt.h>
#include <signal.h>
#include <stdint.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
2012-01-22 18:21:15 +01:00
2015-09-21 14:04:45 +02:00
#include "sd-bus.h"
#include "alloc-util.h"
2015-09-21 14:04:45 +02:00
#include "bus-error.h"
#include "bus-util.h"
#include "cgroup-util.h"
#include "fd-util.h"
#include "fileio.h"
#include "hashmap.h"
#include "parse-util.h"
#include "path-util.h"
#include "process-util.h"
#include "stdio-util.h"
#include "terminal-util.h"
2015-09-21 14:04:45 +02:00
#include "unit-name.h"
#include "util.h"
2012-01-22 18:21:15 +01:00
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;
2012-01-22 18:21:15 +01:00
unsigned cpu_iteration;
nsec_t cpu_usage;
nsec_t cpu_timestamp;
2012-01-22 18:21:15 +01:00
double cpu_fraction;
uint64_t memory;
unsigned io_iteration;
uint64_t io_input, io_output;
nsec_t io_timestamp;
2012-01-22 18:21:15 +01:00
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;
2012-01-22 18:21:15 +01:00
static usec_t arg_delay = 1*USEC_PER_SEC;
2015-09-21 14:04:45 +02:00
static char* arg_machine = NULL;
enum {
COUNT_PIDS,
COUNT_USERSPACE_PROCESSES,
COUNT_ALL_PROCESSES,
} arg_count = COUNT_PIDS;
static bool arg_recursive = true;
2012-01-22 18:21:15 +01:00
static enum {
ORDER_PATH,
ORDER_TASKS,
ORDER_CPU,
ORDER_MEMORY,
ORDER_IO,
2012-01-22 18:21:15 +01:00
} arg_order = ORDER_CPU;
static enum {
CPU_PERCENT,
CPU_TIME,
} arg_cpu_type = CPU_PERCENT;
2012-01-22 18:21:15 +01:00
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) {
2012-01-22 18:21:15 +01:00
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;
2012-01-22 18:21:15 +01:00
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;
2012-01-22 18:21:15 +01:00
r = cg_enumerate_processes(controller, path, &f);
if (r == -ENOENT)
return 0;
if (r < 0)
return r;
2012-01-22 18:21:15 +01:00
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++;
}
2012-01-22 18:21:15 +01:00
if (g->n_tasks > 0)
g->n_tasks_valid = true;
2012-01-22 18:21:15 +01:00
} 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;
core: unified cgroup hierarchy support This patch set adds full support the new unified cgroup hierarchy logic of modern kernels. A new kernel command line option "systemd.unified_cgroup_hierarchy=1" is added. If specified the unified hierarchy is mounted to /sys/fs/cgroup instead of a tmpfs. No further hierarchies are mounted. The kernel command line option defaults to off. We can turn it on by default as soon as the kernel's APIs regarding this are stabilized (but even then downstream distros might want to turn this off, as this will break any tools that access cgroupfs directly). It is possibly to choose for each boot individually whether the unified or the legacy hierarchy is used. nspawn will by default provide the legacy hierarchy to containers if the host is using it, and the unified otherwise. However it is possible to run containers with the unified hierarchy on a legacy host and vice versa, by setting the $UNIFIED_CGROUP_HIERARCHY environment variable for nspawn to 1 or 0, respectively. The unified hierarchy provides reliable cgroup empty notifications for the first time, via inotify. To make use of this we maintain one manager-wide inotify fd, and each cgroup to it. This patch also removes cg_delete() which is unused now. On kernel 4.2 only the "memory" controller is compatible with the unified hierarchy, hence that's the only controller systemd exposes when booted in unified heirarchy mode. This introduces a new enum for enumerating supported controllers, plus a related enum for the mask bits mapping to it. The core is changed to make use of this everywhere. This moves PID 1 into a new "init.scope" implicit scope unit in the root slice. This is necessary since on the unified hierarchy cgroups may either contain subgroups or processes but not both. PID 1 hence has to move out of the root cgroup (strictly speaking the root cgroup is the only one where processes and subgroups are still allowed, but in order to support containers nicey, we move PID 1 into the new scope in all cases.) This new unit is also used on legacy hierarchy setups. It's actually pretty useful on all systems, as it can then be used to filter journal messages coming from PID 1, and so on. The root slice ("-.slice") is now implicitly created and started (and does not require a unit file on disk anymore), since that's where "init.scope" is located and the slice needs to be started before the scope can. To check whether we are in unified or legacy hierarchy mode we use statfs() on /sys/fs/cgroup. If the .f_type field reports tmpfs we are in legacy mode, if it reports cgroupfs we are in unified mode. This patch set carefuly makes sure that cgls and cgtop continue to work as desired. When invoking nspawn as a service it will implicitly create two subcgroups in the cgroup it is using, one to move the nspawn process into, the other to move the actual container processes into. This is done because of the requirement that cgroups may either contain processes or other subgroups.
2015-09-01 19:22:36 +02:00
} else if (streq(controller, "cpuacct") && cg_unified() <= 0) {
_cleanup_free_ char *p = NULL, *v = NULL;
2012-01-22 18:21:15 +01:00
uint64_t new_usage;
nsec_t timestamp;
2012-01-22 18:21:15 +01:00
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;
2012-01-22 18:21:15 +01:00
if (r < 0)
return r;
r = safe_atou64(v, &new_usage);
if (r < 0)
return r;
timestamp = now_nsec(CLOCK_MONOTONIC);
2012-01-22 18:21:15 +01:00
if (g->cpu_iteration == iteration - 1 &&
(nsec_t) new_usage > g->cpu_usage) {
2012-01-22 18:21:15 +01:00
nsec_t x, y;
2012-01-22 18:21:15 +01:00
x = timestamp - g->cpu_timestamp;
if (x < 1)
x = 1;
2012-01-22 18:21:15 +01:00
y = (nsec_t) new_usage - g->cpu_usage;
g->cpu_fraction = (double) y / (double) x;
g->cpu_valid = true;
2012-01-22 18:21:15 +01:00
}
g->cpu_usage = (nsec_t) new_usage;
g->cpu_timestamp = timestamp;
2012-01-22 18:21:15 +01:00
g->cpu_iteration = iteration;
} else if (streq(controller, "memory")) {
_cleanup_free_ char *p = NULL, *v = NULL;
2012-01-22 18:21:15 +01:00
core: unified cgroup hierarchy support This patch set adds full support the new unified cgroup hierarchy logic of modern kernels. A new kernel command line option "systemd.unified_cgroup_hierarchy=1" is added. If specified the unified hierarchy is mounted to /sys/fs/cgroup instead of a tmpfs. No further hierarchies are mounted. The kernel command line option defaults to off. We can turn it on by default as soon as the kernel's APIs regarding this are stabilized (but even then downstream distros might want to turn this off, as this will break any tools that access cgroupfs directly). It is possibly to choose for each boot individually whether the unified or the legacy hierarchy is used. nspawn will by default provide the legacy hierarchy to containers if the host is using it, and the unified otherwise. However it is possible to run containers with the unified hierarchy on a legacy host and vice versa, by setting the $UNIFIED_CGROUP_HIERARCHY environment variable for nspawn to 1 or 0, respectively. The unified hierarchy provides reliable cgroup empty notifications for the first time, via inotify. To make use of this we maintain one manager-wide inotify fd, and each cgroup to it. This patch also removes cg_delete() which is unused now. On kernel 4.2 only the "memory" controller is compatible with the unified hierarchy, hence that's the only controller systemd exposes when booted in unified heirarchy mode. This introduces a new enum for enumerating supported controllers, plus a related enum for the mask bits mapping to it. The core is changed to make use of this everywhere. This moves PID 1 into a new "init.scope" implicit scope unit in the root slice. This is necessary since on the unified hierarchy cgroups may either contain subgroups or processes but not both. PID 1 hence has to move out of the root cgroup (strictly speaking the root cgroup is the only one where processes and subgroups are still allowed, but in order to support containers nicey, we move PID 1 into the new scope in all cases.) This new unit is also used on legacy hierarchy setups. It's actually pretty useful on all systems, as it can then be used to filter journal messages coming from PID 1, and so on. The root slice ("-.slice") is now implicitly created and started (and does not require a unit file on disk anymore), since that's where "init.scope" is located and the slice needs to be started before the scope can. To check whether we are in unified or legacy hierarchy mode we use statfs() on /sys/fs/cgroup. If the .f_type field reports tmpfs we are in legacy mode, if it reports cgroupfs we are in unified mode. This patch set carefuly makes sure that cgls and cgtop continue to work as desired. When invoking nspawn as a service it will implicitly create two subcgroups in the cgroup it is using, one to move the nspawn process into, the other to move the actual container processes into. This is done because of the requirement that cgroups may either contain processes or other subgroups.
2015-09-01 19:22:36 +02:00
if (cg_unified() <= 0)
r = cg_get_path(controller, path, "memory.usage_in_bytes", &p);
else
r = cg_get_path(controller, path, "memory.current", &p);
2012-01-22 18:21:15 +01:00
if (r < 0)
return r;
r = read_one_line_file(p, &v);
if (r == -ENOENT)
return 0;
2012-01-22 18:21:15 +01:00
if (r < 0)
return r;
r = safe_atou64(v, &g->memory);
if (r < 0)
return r;
if (g->memory > 0)
g->memory_valid = true;
core: unified cgroup hierarchy support This patch set adds full support the new unified cgroup hierarchy logic of modern kernels. A new kernel command line option "systemd.unified_cgroup_hierarchy=1" is added. If specified the unified hierarchy is mounted to /sys/fs/cgroup instead of a tmpfs. No further hierarchies are mounted. The kernel command line option defaults to off. We can turn it on by default as soon as the kernel's APIs regarding this are stabilized (but even then downstream distros might want to turn this off, as this will break any tools that access cgroupfs directly). It is possibly to choose for each boot individually whether the unified or the legacy hierarchy is used. nspawn will by default provide the legacy hierarchy to containers if the host is using it, and the unified otherwise. However it is possible to run containers with the unified hierarchy on a legacy host and vice versa, by setting the $UNIFIED_CGROUP_HIERARCHY environment variable for nspawn to 1 or 0, respectively. The unified hierarchy provides reliable cgroup empty notifications for the first time, via inotify. To make use of this we maintain one manager-wide inotify fd, and each cgroup to it. This patch also removes cg_delete() which is unused now. On kernel 4.2 only the "memory" controller is compatible with the unified hierarchy, hence that's the only controller systemd exposes when booted in unified heirarchy mode. This introduces a new enum for enumerating supported controllers, plus a related enum for the mask bits mapping to it. The core is changed to make use of this everywhere. This moves PID 1 into a new "init.scope" implicit scope unit in the root slice. This is necessary since on the unified hierarchy cgroups may either contain subgroups or processes but not both. PID 1 hence has to move out of the root cgroup (strictly speaking the root cgroup is the only one where processes and subgroups are still allowed, but in order to support containers nicey, we move PID 1 into the new scope in all cases.) This new unit is also used on legacy hierarchy setups. It's actually pretty useful on all systems, as it can then be used to filter journal messages coming from PID 1, and so on. The root slice ("-.slice") is now implicitly created and started (and does not require a unit file on disk anymore), since that's where "init.scope" is located and the slice needs to be started before the scope can. To check whether we are in unified or legacy hierarchy mode we use statfs() on /sys/fs/cgroup. If the .f_type field reports tmpfs we are in legacy mode, if it reports cgroupfs we are in unified mode. This patch set carefuly makes sure that cgls and cgtop continue to work as desired. When invoking nspawn as a service it will implicitly create two subcgroups in the cgroup it is using, one to move the nspawn process into, the other to move the actual container processes into. This is done because of the requirement that cgroups may either contain processes or other subgroups.
2015-09-01 19:22:36 +02:00
} else if (streq(controller, "blkio") && cg_unified() <= 0) {
_cleanup_fclose_ FILE *f = NULL;
_cleanup_free_ char *p = NULL;
2012-01-22 18:21:15 +01:00
uint64_t wr = 0, rd = 0;
nsec_t timestamp;
2012-01-22 18:21:15 +01:00
r = cg_get_path(controller, path, "blkio.io_service_bytes", &p);
if (r < 0)
return r;
f = fopen(p, "re");
if (!f) {
if (errno == ENOENT)
return 0;
2012-01-22 18:21:15 +01:00
return -errno;
}
2012-01-22 18:21:15 +01:00
for (;;) {
char line[LINE_MAX], *l;
uint64_t k, *q;
if (!fgets(line, sizeof(line), f))
break;
l = strstrip(line);
l += strcspn(l, WHITESPACE);
l += strspn(l, WHITESPACE);
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);
2012-01-22 18:21:15 +01:00
if (g->io_iteration == iteration - 1) {
uint64_t x, yr, yw;
x = (uint64_t) (timestamp - g->io_timestamp);
if (x < 1)
x = 1;
2012-01-22 18:21:15 +01:00
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;
2012-01-22 18:21:15 +01:00
if (yr > 0 || yw > 0) {
2012-01-22 18:21:15 +01:00
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;
2012-01-22 18:21:15 +01:00
g->io_iteration = iteration;
}
if (ret)
*ret = g;
2012-01-22 18:21:15 +01:00
return 0;
}
static int refresh_one(
const char *controller,
const char *path,
Hashmap *a,
Hashmap *b,
unsigned iteration,
unsigned depth,
Group **ret) {
2012-01-22 18:21:15 +01:00
_cleanup_closedir_ DIR *d = NULL;
Group *ours;
2012-01-22 18:21:15 +01:00
int r;
assert(controller);
assert(path);
assert(a);
if (depth > arg_depth)
return 0;
r = process(controller, path, a, b, iteration, &ours);
2012-01-22 18:21:15 +01:00
if (r < 0)
return r;
r = cg_enumerate_subgroups(controller, path, &d);
if (r == -ENOENT)
return 0;
if (r < 0)
2012-01-22 18:21:15 +01:00
return r;
for (;;) {
_cleanup_free_ char *fn = NULL, *p = NULL;
Group *child = NULL;
2012-01-22 18:21:15 +01:00
r = cg_read_subgroup(d, &fn);
if (r < 0)
return r;
if (r == 0)
break;
2012-01-22 18:21:15 +01:00
p = strjoin(path, "/", fn, NULL);
if (!p)
return -ENOMEM;
2012-01-22 18:21:15 +01:00
path_kill_slashes(p);
r = refresh_one(controller, p, a, b, iteration, depth + 1, &child);
2012-01-22 18:21:15 +01:00
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;
}
}
2012-01-22 18:21:15 +01:00
}
if (ret)
*ret = ours;
return 1;
2012-01-22 18:21:15 +01:00
}
static int refresh(const char *root, Hashmap *a, Hashmap *b, unsigned iteration) {
2012-01-22 18:21:15 +01:00
int r;
assert(a);
r = refresh_one(SYSTEMD_CGROUP_CONTROLLER, root, a, b, iteration, 0, NULL);
2012-01-22 18:21:15 +01:00
if (r < 0)
return r;
r = refresh_one("cpuacct", root, a, b, iteration, 0, NULL);
2012-01-22 18:21:15 +01:00
if (r < 0)
return r;
r = refresh_one("memory", root, a, b, iteration, 0, NULL);
2012-01-22 18:21:15 +01:00
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;
2012-01-22 18:21:15 +01:00
}
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;
2012-01-22 18:21:15 +01:00
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)
2012-01-22 18:21:15 +01:00
return -1;
else if (y->cpu_valid)
2012-01-22 18:21:15 +01:00
return 1;
} else {
if (x->cpu_usage > y->cpu_usage)
return -1;
else if (x->cpu_usage < y->cpu_usage)
return 1;
}
2012-01-22 18:21:15 +01:00
break;
2012-01-22 18:21:15 +01:00
case ORDER_TASKS:
2012-01-22 18:21:15 +01:00
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:
2012-01-22 18:21:15 +01:00
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:
2012-01-22 18:21:15 +01:00
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);
2012-01-22 18:21:15 +01:00
}
static void display(Hashmap *a) {
2012-01-22 18:21:15 +01:00
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)];
2012-01-22 18:21:15 +01:00
assert(a);
if (on_tty())
fputs(ANSI_HOME_CLEAR, stdout);
2012-01-22 18:21:15 +01:00
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);
2012-01-22 18:21:15 +01:00
/* 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;
2012-01-22 18:21:15 +01:00
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;
2012-01-22 18:21:15 +01:00
for (j = 0; j < n; j++) {
_cleanup_free_ char *ellipsized = NULL;
const char *path;
2012-01-22 18:21:15 +01:00
if (on_tty() && j + 6 > rows)
2012-01-22 18:21:15 +01:00
break;
g = array[j];
path = isempty(g->path) ? "/" : g->path;
ellipsized = ellipsize(path, path_columns, 33);
printf("%-*s", path_columns, ellipsized ?: path);
2012-01-22 18:21:15 +01:00
if (g->n_tasks_valid)
printf(" %7" PRIu64, g->n_tasks);
2012-01-22 18:21:15 +01:00
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));
2012-01-22 18:21:15 +01:00
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));
2012-01-22 18:21:15 +01:00
putchar('\n');
}
}
static void help(void) {
2012-01-22 18:21:15 +01:00
printf("%s [OPTIONS...]\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"
2015-09-21 14:04:45 +02:00
" -M --machine= Show container\n"
, program_invocation_short_name, arg_depth);
2012-07-26 23:09:02 +02:00
}
2012-01-22 18:21:15 +01:00
static int parse_argv(int argc, char *argv[]) {
enum {
2012-07-26 23:09:02 +02:00
ARG_VERSION = 0x100,
ARG_DEPTH,
ARG_CPU_TYPE,
ARG_ORDER,
ARG_RECURSIVE,
2012-01-22 18:21:15 +01:00
};
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 },
2015-09-21 14:04:45 +02:00
{ "machine", required_argument, NULL, 'M' },
{}
2012-01-22 18:21:15 +01:00
};
bool recursive_unset = false;
int c, r;
2012-01-22 18:21:15 +01:00
assert(argc >= 1);
assert(argv);
2015-09-21 14:04:45 +02:00
while ((c = getopt_long(argc, argv, "hptcmin:brd:kPM:", options, NULL)) >= 0)
2012-01-22 18:21:15 +01:00
switch (c) {
case 'h':
help();
return 0;
2012-01-22 18:21:15 +01:00
2012-07-26 23:09:02 +02:00
case ARG_VERSION:
return version();
2012-07-26 23:09:02 +02:00
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;
2012-01-22 18:21:15 +01:00
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);
2012-01-22 18:21:15 +01:00
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;
2012-01-22 18:21:15 +01:00
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;
2015-09-21 14:04:45 +02:00
case 'M':
arg_machine = optarg;
break;
2012-01-22 18:21:15 +01:00
case '?':
return -EINVAL;
default:
assert_not_reached("Unhandled option");
2012-01-22 18:21:15 +01:00
}
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;
}
2012-01-22 18:21:15 +01:00
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)";
}
2015-09-21 14:04:45 +02:00
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;
2015-09-21 14:04:45 +02:00
_cleanup_free_ char *unit = NULL, *path = NULL;
const char *m;
int r;
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);
2015-09-21 14:04:45 +02:00
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;
}
2012-01-22 18:21:15 +01:00
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;
2012-01-22 18:21:15 +01:00
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;
2012-01-22 18:21:15 +01:00
r = parse_argv(argc, argv);
if (r <= 0)
goto finish;
2015-09-21 14:04:45 +02:00
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);
2012-01-22 18:21:15 +01:00
if (!a || !b) {
r = log_oom();
2012-01-22 18:21:15 +01:00
goto finish;
}
signal(SIGWINCH, columns_lines_cache_reset);
2012-10-18 23:50:26 +02:00
if (arg_iterations == (unsigned) -1)
arg_iterations = on_tty() ? 0 : 1;
2012-01-22 18:21:15 +01:00
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");
2012-01-22 18:21:15 +01:00
goto finish;
}
2012-01-22 18:21:15 +01:00
group_hashmap_clear(b);
c = a;
a = b;
b = c;
last_refresh = t;
immediate_refresh = false;
}
display(b);
2012-01-22 18:21:15 +01:00
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;
}
2012-01-22 18:21:15 +01:00
}
if (on_tty()) { /* TTY: Clear any user keystroke */
fputs("\r \r", stdout);
fflush(stdout);
}
2012-01-22 18:21:15 +01:00
if (arg_batch)
continue;
2012-01-22 18:21:15 +01:00
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;
2012-01-22 18:21:15 +01:00
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));
2012-01-22 18:21:15 +01:00
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));
2012-01-22 18:21:15 +01:00
fflush(stdout);
sleep(1);
break;
case '?':
case 'h':
#define ON ANSI_HIGHLIGHT
#define OFF ANSI_NORMAL
2012-01-22 18:21:15 +01:00
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");
2012-01-22 18:21:15 +01:00
fflush(stdout);
sleep(3);
break;
default:
if (key < ' ')
fprintf(stdout, "\nUnknown key '\\x%x'. Ignoring.", key);
else
fprintf(stdout, "\nUnknown key '%c'. Ignoring.", key);
2012-01-22 18:21:15 +01:00
fflush(stdout);
sleep(1);
break;
}
}
r = 0;
finish:
group_hashmap_free(a);
group_hashmap_free(b);
return r < 0 ? EXIT_FAILURE : EXIT_SUCCESS;
2012-01-22 18:21:15 +01:00
}