* Copyright (C) 2013-2021 Canonical, Ltd.
* Copyright (C) 2021-2025 Colin Ian King.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version 2
* of the License, or (at your option) any later version.
*
* This program 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 General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
*
*/
#include "stress-ng.h"
static const stress_help_t help[] = {
{ NULL, "kill N", "start N workers killing with SIGUSR1" },
{ NULL, "kill-ops N", "stop after N kill bogo operations" },
{ NULL, NULL, NULL }
};
* stress on sched_kill()
* stress system by rapid kills
*/
static int stress_kill(stress_args_t *args)
{
uint64_t udelay = 5000;
pid_t pid;
const pid_t ppid = getpid();
int ret;
double duration = 0.0, count = 0.0, rate;
if (stress_sighandler(args->name, SIGUSR1, SIG_IGN, NULL) < 0)
return EXIT_FAILURE;
pid = fork();
if (pid == 0) {
stress_set_proc_state(args->name, STRESS_STATE_RUN);
VOID_RET(int, stress_sighandler(args->name, SIGUSR1, stress_sighandler_nop, NULL));
while (stress_continue(args)) {
if (kill(ppid, 0) < 0)
break;
pause();
}
stress_set_proc_state(args->name, STRESS_STATE_WAIT);
_exit(0);
}
stress_set_proc_state(args->name, STRESS_STATE_SYNC_WAIT);
stress_sync_start_wait(args);
stress_set_proc_state(args->name, STRESS_STATE_RUN);
do {
pid_t bad_pid;
double t;
* With many kill stressors we get into a state
* where they all hammer on kill system calls and
* this stops the parent from getting scheduling
* time to spawn off the rest of the kill stressors
* causing some lag in getting all the stressors
* running. Ease this pressure off to being with
* with some small sleeps that shrink to zero over
* time. The alternative was to re-nice all the
* processes, but even then one gets the child
* stressors all contending and causing a bottle
* neck. Any simpler and/or better solutions would
* be appreciated!
*/
if (udelay >= 1000) {
(void)shim_usleep(udelay);
udelay -= 500;
}
t = stress_time_now();
ret = kill(args->pid, SIGUSR1);
if (LIKELY(ret == 0)) {
const int saved_errno = errno;
duration += stress_time_now() - t;
count += 1.0;
errno = saved_errno;
} else if (UNLIKELY((ret < 0) && (g_opt_flags & OPT_FLAGS_VERIFY))) {
pr_fail("%s: kill PID %" PRIdMAX " with SIGUSR1 failed, errno=%d (%s)\n",
args->name, (intmax_t)args->pid, errno, strerror(errno));
}
t = stress_time_now();
ret = kill(args->pid, 0);
if (LIKELY(ret == 0)) {
const int saved_errno = errno;
duration += stress_time_now() - t;
count += 1.0;
errno = saved_errno;
} else if (UNLIKELY((ret < 0) && (g_opt_flags & OPT_FLAGS_VERIFY))) {
pr_fail("%s: kill PID %" PRIdMAX " with signal 0 failed, errno=%d (%s)\n",
args->name, (intmax_t)args->pid, errno, strerror(errno));
}
* Zero signal can be used to see if process exists,
* -1 pid means signal sent to every process caller has
* permission to send to
*/
t = stress_time_now();
ret = kill(-1, 0);
if (LIKELY(ret == 0)) {
const int saved_errno = errno;
duration += stress_time_now() - t;
count += 1.0;
errno = saved_errno;
} else if (UNLIKELY((ret < 0) && (g_opt_flags & OPT_FLAGS_VERIFY))) {
pr_fail("%s: kill PID -1 with signal 0 failed, errno=%d (%s)\n",
args->name, errno, strerror(errno));
}
* Exercise the kernel by sending illegal signal numbers,
* should return -EINVAL
*/
VOID_RET(int, kill(args->pid, -1));
VOID_RET(int, kill(args->pid, INT_MIN));
VOID_RET(int, kill(0, INT_MIN));
* Exercise the kernel by sending illegal pid INT_MIN,
* should return -ESRCH but not sure if that is portable
*/
VOID_RET(int, kill(INT_MIN, 0));
* Send child process some signals to keep it busy
*/
if (LIKELY(pid > 1)) {
VOID_RET(int, kill(pid, 0));
#if defined(SIGSTOP) && \
defined(SIGCONT)
VOID_RET(int, kill(pid, SIGSTOP));
VOID_RET(int, kill(pid, SIGCONT));
#endif
VOID_RET(int, kill(pid, SIGUSR1));
}
bad_pid = stress_get_unused_pid_racy(false);
VOID_RET(int, kill(bad_pid, 0));
stress_bogo_inc(args);
} while (stress_continue(args));
stress_set_proc_state(args->name, STRESS_STATE_DEINIT);
if (pid != -1) {
int status;
VOID_RET(int, kill(pid, SIGKILL));
VOID_RET(pid_t, waitpid(pid, &status, 0));
}
rate = (duration > 0.0) ? count / duration : 0.0;
stress_metrics_set(args, 0, "kill calls per sec",
rate, STRESS_METRIC_HARMONIC_MEAN);
return EXIT_SUCCESS;
}
const stressor_info_t stress_kill_info = {
.stressor = stress_kill,
.classifier = CLASS_INTERRUPT | CLASS_SCHEDULER | CLASS_OS,
.verify = VERIFY_OPTIONAL,
.help = help
};