1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282
|
/*
Copyright 2024 Northern.tech AS
This file is part of CFEngine 3 - written and maintained by Northern.tech AS.
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; version 3.
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., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA
To the extent this program is licensed as part of the Enterprise
versions of CFEngine, the applicable Commercial Open Source License
(COSL) may apply to this file if you as a licensee so wish it. See
included file COSL.txt.
*/
#include <platform.h>
#include <logging.h>
#include <process_lib.h>
#include <process_unix_priv.h>
#include <misc_lib.h>
#define SLEEP_POLL_TIMEOUT_NS 10000000
/*
* Wait until process specified by #pid is stopped due to SIGSTOP signal.
*
* @returns true if process has come to stop during #timeout_ns nanoseconds,
* false if the process cannot be found or failed to stop during #timeout_ns
* nanoseconds.
*
* FIXME: Only timeouts < 1s are supported
*/
static bool ProcessWaitUntilStopped(pid_t pid, long timeout_ns)
{
while (timeout_ns > 0)
{
switch (GetProcessState(pid))
{
case PROCESS_STATE_RUNNING:
break; /* retry in a while */
case PROCESS_STATE_STOPPED:
return true;
case PROCESS_STATE_ZOMBIE:
/* There is not much we can do by waiting a zombie process. It
* will never change to a stopped state. */
return false;
case PROCESS_STATE_DOES_NOT_EXIST:
return false;
}
struct timespec ts = {
.tv_sec = 0,
.tv_nsec = MIN(SLEEP_POLL_TIMEOUT_NS, timeout_ns),
};
while (nanosleep(&ts, &ts) < 0)
{
if (errno != EINTR)
{
ProgrammingError("Invalid timeout for nanosleep");
}
}
timeout_ns = MAX(0, timeout_ns - SLEEP_POLL_TIMEOUT_NS);
}
return false;
}
/*
* Currently only timeouts < 1s are supported
*/
static bool ProcessWaitUntilExited(pid_t pid, long timeout_ns)
{
assert(timeout_ns < 1000000000);
while (timeout_ns > 0)
{
switch (GetProcessState(pid))
{
case PROCESS_STATE_RUNNING:
break; /* retry in a while */
case PROCESS_STATE_DOES_NOT_EXIST:
return true;
case PROCESS_STATE_ZOMBIE:
/* There is not much we can do by waiting a zombie process. It's
the responsibility of the caller to reap the child so we're
considering it has already exited. */
return true;
case PROCESS_STATE_STOPPED:
/* Almost the same case with a zombie process, but it will
* respond only to signals that can't be caught. */
return false;
}
struct timespec ts = {
.tv_sec = 0,
.tv_nsec = MIN(SLEEP_POLL_TIMEOUT_NS, timeout_ns),
};
Log(LOG_LEVEL_DEBUG,
"PID %jd still alive after signalling, waiting for %lu ms...",
(intmax_t) pid, ts.tv_nsec / 1000000);
while (nanosleep(&ts, &ts) < 0)
{
if (errno != EINTR)
{
ProgrammingError("Invalid timeout for nanosleep");
}
}
timeout_ns = MAX(0, timeout_ns - SLEEP_POLL_TIMEOUT_NS);
}
return false;
}
/* A timeout (in nanoseconds) to wait for process to stop (pause) or exit.
* Note that it's important that it does not overflow 32 bits; no more than
* nine 9s in a row, i.e. one second. */
#define STOP_WAIT_TIMEOUT 999999999L
/*
* Safely kill process by checking that the process is the right one by matching
* process start time.
*
* The algorithm:
*
* 1. Check that the process has the same start time as stored in lock. If it
* is not, return, as we know for sure this is a wrong process. (This step
* is an optimization to avoid sending SIGSTOP/SIGCONT to wrong processes).
*
* 2. Send SIGSTOP to the process.
*
* 3. Poll process state until it is stopped.
*
* Now the process is stopped, so we may examine it and not be afraid that it
* will exit and another one with the same PID will appear.
*
* 4. Check that the process has the same start time as provided.
* If it is, send the signal to the process.
*
* 5. Send SIGCONT to the process, so it may continue.
*
*
* Returns 0 on success, -1 on error. Error code is signalled through errno.
*
* ERRORS
*
* EINVAL An invalid signal was specified.
* EPERM The process does not have permission to send the signal.
* ESRCH The pid does not exist or its start time does not match expected one.
*/
static int SafeKill(pid_t pid, time_t expected_start_time, int signal)
{
/* Preliminary check: in case process start time is different already, we
* are sure we don't want to STOP it or kill it. */
time_t pid_start_time = GetProcessStartTime(pid);
if (pid_start_time != expected_start_time)
{
errno = ESRCH;
return -1;
}
/* Now to avoid race conditions we need to stop process so it won't exit
* voluntarily while we are working on it */
if (kill(pid, SIGSTOP) < 0)
{
return -1;
}
if (!ProcessWaitUntilStopped(pid, STOP_WAIT_TIMEOUT))
{
/* Ensure the process is started again in case of timeout or error, so
* we don't leave SIGSTOP'ed processes around on overloaded or
* misconfigured machine */
kill(pid, SIGCONT);
errno = ESRCH;
return -1;
}
/* Here process has stopped, so we may interrogate it without race conditions */
pid_start_time = GetProcessStartTime(pid);
if (pid_start_time != expected_start_time)
{
/* This is a wrong process, let it continue */
kill(pid, SIGCONT);
errno = ESRCH;
return -1;
}
/* We've got a right process, signal it and let it continue */
int ret = kill(pid, signal);
int saved_errno = errno;
/*
* We don't check return value of SIGCONT, as the process may have been
* terminated already by previous kill. Moreover, what would we do with the
* return code?
*/
kill(pid, SIGCONT);
errno = saved_errno;
return ret;
}
static int Kill(pid_t pid, time_t process_start_time, int signal)
{
if (process_start_time == PROCESS_START_TIME_UNKNOWN)
{
/* We don't know when the process has started, do a plain kill(2) */
return kill(pid, signal);
}
else
{
return SafeKill(pid, process_start_time, signal);
}
}
bool GracefulTerminate(pid_t pid, time_t process_start_time)
{
/* We can't allow to kill ourselves. First it does not make sense, and
* second, once SafeKill() sends SIGSTOP, we will just freeze forever. */
if (pid == getpid())
{
Log(LOG_LEVEL_WARNING,
"Ignoring request to kill ourself (pid %jd)!",
(intmax_t) pid);
return false;
}
if (Kill(pid, process_start_time, SIGINT) < 0)
{
/* If we failed to kill the process return error. If the process
* doesn't even exist (errno==ESRCH), again return error, we shouldn't
* have signalled the PID in the first place. */
return false;
}
if (ProcessWaitUntilExited(pid, STOP_WAIT_TIMEOUT))
{
return true;
}
if (Kill(pid, process_start_time, SIGTERM) < 0)
{
return errno == ESRCH;
}
if (ProcessWaitUntilExited(pid, STOP_WAIT_TIMEOUT))
{
return true;
}
if (Kill(pid, process_start_time, SIGKILL) < 0)
{
return errno == ESRCH;
}
return true;
}
|