along with systemd; If not, see <http://www.gnu.org/licenses/>.
***/
+#include <ctype.h>
+#include <errno.h>
+#include <limits.h>
+#include <linux/oom.h>
+#include <sched.h>
+#include <signal.h>
#include <stdbool.h>
-#include <sys/types.h>
-#include <string.h>
#include <stdio.h>
-#include <assert.h>
-#include <errno.h>
-#include <unistd.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/personality.h>
+#include <sys/prctl.h>
+#include <sys/types.h>
#include <sys/wait.h>
-#include <signal.h>
-#include <ctype.h>
-
+#include <syslog.h>
+#include <unistd.h>
+#ifdef HAVE_VALGRIND_VALGRIND_H
+#include <valgrind/valgrind.h>
+#endif
+
+#include "alloc-util.h"
+//#include "architecture.h"
+#include "escape.h"
+#include "fd-util.h"
#include "fileio.h"
-#include "util.h"
+#include "fs-util.h"
+//#include "ioprio.h"
#include "log.h"
-#include "signal-util.h"
+#include "macro.h"
+#include "missing.h"
#include "process-util.h"
+#include "signal-util.h"
+//#include "stat-util.h"
+#include "string-table.h"
+#include "string-util.h"
+#include "user-util.h"
+#include "util.h"
int get_process_state(pid_t pid) {
const char *p;
assert(pid >= 0);
p = procfs_file_alloca(pid, "stat");
+
r = read_one_line_file(p, &line);
+ if (r == -ENOENT)
+ return -ESRCH;
if (r < 0)
return r;
int get_process_cmdline(pid_t pid, size_t max_length, bool comm_fallback, char **line) {
_cleanup_fclose_ FILE *f = NULL;
- char *r = NULL, *k;
+ bool space = false;
+ char *k, *ans = NULL;
const char *p;
int c;
assert(line);
assert(pid >= 0);
+ /* Retrieves a process' command line. Replaces unprintable characters while doing so by whitespace (coalescing
+ * multiple sequential ones into one). If max_length is != 0 will return a string of the specified size at most
+ * (the trailing NUL byte does count towards the length here!), abbreviated with a "..." ellipsis. If
+ * comm_fallback is true and the process has no command line set (the case for kernel threads), or has a
+ * command line that resolves to the empty string will return the "comm" name of the process instead.
+ *
+ * Returns -ESRCH if the process doesn't exist, and -ENOENT if the process has no command line (and
+ * comm_fallback is false). Returns 0 and sets *line otherwise. */
+
p = procfs_file_alloca(pid, "cmdline");
f = fopen(p, "re");
- if (!f)
+ if (!f) {
+ if (errno == ENOENT)
+ return -ESRCH;
return -errno;
+ }
+
+ if (max_length == 1) {
+
+ /* If there's only room for one byte, return the empty string */
+ ans = new0(char, 1);
+ if (!ans)
+ return -ENOMEM;
- if (max_length == 0) {
+ *line = ans;
+ return 0;
+
+ } else if (max_length == 0) {
size_t len = 0, allocated = 0;
while ((c = getc(f)) != EOF) {
- if (!GREEDY_REALLOC(r, allocated, len+2)) {
- free(r);
+ if (!GREEDY_REALLOC(ans, allocated, len+3)) {
+ free(ans);
return -ENOMEM;
}
- r[len++] = isprint(c) ? c : ' ';
- }
+ if (isprint(c)) {
+ if (space) {
+ ans[len++] = ' ';
+ space = false;
+ }
+
+ ans[len++] = c;
+ } else if (len > 0)
+ space = true;
+ }
if (len > 0)
- r[len-1] = 0;
+ ans[len] = '\0';
+ else
+ ans = mfree(ans);
} else {
- bool space = false;
+ bool dotdotdot = false;
size_t left;
- r = new(char, max_length);
- if (!r)
+ ans = new(char, max_length);
+ if (!ans)
return -ENOMEM;
- k = r;
+ k = ans;
left = max_length;
while ((c = getc(f)) != EOF) {
if (isprint(c)) {
+
if (space) {
- if (left <= 4)
+ if (left <= 2) {
+ dotdotdot = true;
break;
+ }
*(k++) = ' ';
left--;
space = false;
}
- if (left <= 4)
+ if (left <= 1) {
+ dotdotdot = true;
break;
+ }
*(k++) = (char) c;
left--;
- } else
+ } else if (k > ans)
space = true;
}
- if (left <= 4) {
- size_t n = MIN(left-1, 3U);
- memcpy(k, "...", n);
- k[n] = 0;
+ if (dotdotdot) {
+ if (max_length <= 4) {
+ k = ans;
+ left = max_length;
+ } else {
+ k = ans + max_length - 4;
+ left = 4;
+
+ /* Eat up final spaces */
+ while (k > ans && isspace(k[-1])) {
+ k--;
+ left++;
+ }
+ }
+
+ strncpy(k, "...", left-1);
+ k[left-1] = 0;
} else
*k = 0;
}
/* Kernel threads have no argv[] */
- if (isempty(r)) {
+ if (isempty(ans)) {
_cleanup_free_ char *t = NULL;
int h;
- free(r);
+ free(ans);
if (!comm_fallback)
return -ENOENT;
if (h < 0)
return h;
- r = strjoin("[", t, "]", NULL);
- if (!r)
+ if (max_length == 0)
+ ans = strjoin("[", t, "]");
+ else {
+ size_t l;
+
+ l = strlen(t);
+
+ if (l + 3 <= max_length)
+ ans = strjoin("[", t, "]");
+ else if (max_length <= 6) {
+
+ ans = new(char, max_length);
+ if (!ans)
+ return -ENOMEM;
+
+ memcpy(ans, "[...]", max_length-1);
+ ans[max_length-1] = 0;
+ } else {
+ char *e;
+
+ t[max_length - 6] = 0;
+
+ /* Chop off final spaces */
+ e = strchr(t, 0);
+ while (e > t && isspace(e[-1]))
+ e--;
+ *e = 0;
+
+ ans = strjoin("[", t, "...]");
+ }
+ }
+ if (!ans)
return -ENOMEM;
}
- *line = r;
+ *line = ans;
return 0;
}
+#if 0 /// UNNEEDED by elogind
+int rename_process(const char name[]) {
+ static size_t mm_size = 0;
+ static char *mm = NULL;
+ bool truncated = false;
+ size_t l;
+
+ /* This is a like a poor man's setproctitle(). It changes the comm field, argv[0], and also the glibc's
+ * internally used name of the process. For the first one a limit of 16 chars applies; to the second one in
+ * many cases one of 10 (i.e. length of "/sbin/init") — however if we have CAP_SYS_RESOURCES it is unbounded;
+ * to the third one 7 (i.e. the length of "systemd". If you pass a longer string it will likely be
+ * truncated.
+ *
+ * Returns 0 if a name was set but truncated, > 0 if it was set but not truncated. */
+
+ if (isempty(name))
+ return -EINVAL; /* let's not confuse users unnecessarily with an empty name */
+
+ l = strlen(name);
+
+ /* First step, change the comm field. */
+ (void) prctl(PR_SET_NAME, name);
+ if (l > 15) /* Linux process names can be 15 chars at max */
+ truncated = true;
+
+ /* Second step, change glibc's ID of the process name. */
+ if (program_invocation_name) {
+ size_t k;
+
+ k = strlen(program_invocation_name);
+ strncpy(program_invocation_name, name, k);
+ if (l > k)
+ truncated = true;
+ }
+
+ /* Third step, completely replace the argv[] array the kernel maintains for us. This requires privileges, but
+ * has the advantage that the argv[] array is exactly what we want it to be, and not filled up with zeros at
+ * the end. This is the best option for changing /proc/self/cmdline. */
+ if (mm_size < l+1) {
+ size_t nn_size;
+ char *nn;
+
+ /* Let's not bother with this if we don't have euid == 0. Strictly speaking if people do weird stuff
+ * with capabilities this could work even for euid != 0, but our own code generally doesn't do that,
+ * hence let's use this as quick bypass check, to avoid calling mmap() if PR_SET_MM_ARG_START fails
+ * with EPERM later on anyway. After all geteuid() is dead cheap to call, but mmap() is not. */
+ if (geteuid() != 0) {
+ log_debug("Skipping PR_SET_MM_ARG_START, as we don't have privileges.");
+ goto use_saved_argv;
+ }
+
+ nn_size = PAGE_ALIGN(l+1);
+ nn = mmap(NULL, nn_size, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0);
+ if (nn == MAP_FAILED) {
+ log_debug_errno(errno, "mmap() failed: %m");
+ goto use_saved_argv;
+ }
+
+ strncpy(nn, name, nn_size);
+
+ /* Now, let's tell the kernel about this new memory */
+ if (prctl(PR_SET_MM, PR_SET_MM_ARG_START, (unsigned long) nn, 0, 0) < 0) {
+ log_debug_errno(errno, "PR_SET_MM_ARG_START failed, proceeding without: %m");
+ (void) munmap(nn, nn_size);
+ goto use_saved_argv;
+ }
+
+ /* And update the end pointer to the new end, too. If this fails, we don't really know what to do, it's
+ * pretty unlikely that we can rollback, hence we'll just accept the failure, and continue. */
+ if (prctl(PR_SET_MM, PR_SET_MM_ARG_END, (unsigned long) nn + l + 1, 0, 0) < 0)
+ log_debug_errno(errno, "PR_SET_MM_ARG_END failed, proceeding without: %m");
+
+ if (mm)
+ (void) munmap(mm, mm_size);
+
+ mm = nn;
+ mm_size = nn_size;
+ } else
+ strncpy(mm, name, mm_size);
+
+use_saved_argv:
+ /* Fourth step: in all cases we'll also update the original argv[], so that our own code gets it right too if
+ * it still looks here */
+
+ if (saved_argc > 0) {
+ int i;
+
+ if (saved_argv[0]) {
+ size_t k;
+
+ k = strlen(saved_argv[0]);
+ strncpy(saved_argv[0], name, k);
+ if (l > k)
+ truncated = true;
+ }
+
+ for (i = 1; i < saved_argc; i++) {
+ if (!saved_argv[i])
+ break;
+
+ memzero(saved_argv[i], strlen(saved_argv[i]));
+ }
+ }
+
+ return !truncated;
+}
+#endif // 0
+
int is_kernel_thread(pid_t pid) {
const char *p;
size_t count;
bool eof;
FILE *f;
- if (pid == 0)
+ if (pid == 0 || pid == 1) /* pid 1, and we ourselves certainly aren't a kernel thread */
return 0;
- assert(pid > 0);
+ assert(pid > 1);
p = procfs_file_alloca(pid, "cmdline");
f = fopen(p, "re");
- if (!f)
+ if (!f) {
+ if (errno == ENOENT)
+ return -ESRCH;
return -errno;
+ }
count = fread(&c, 1, 1, f);
eof = feof(f);
return 0;
}
-/// UNNEEDED by elogind
-#if 0
+#if 0 /// UNNEEDED by elogind
int get_process_capeff(pid_t pid, char **capeff) {
const char *p;
+ int r;
assert(capeff);
assert(pid >= 0);
p = procfs_file_alloca(pid, "status");
- return get_status_field(p, "\nCapEff:", capeff);
+ r = get_proc_field(p, "CapEff", WHITESPACE, capeff);
+ if (r == -ENOENT)
+ return -ESRCH;
+
+ return r;
}
#endif // 0
assert(name);
r = readlink_malloc(proc_file, name);
+ if (r == -ENOENT)
+ return -ESRCH;
if (r < 0)
- return r == -ENOENT ? -ESRCH : r;
+ return r;
return 0;
}
return 0;
}
-/// UNNEEDED by elogind
-#if 0
+#if 0 /// UNNEEDED by elogind
static int get_process_id(pid_t pid, const char *field, uid_t *uid) {
_cleanup_fclose_ FILE *f = NULL;
char line[LINE_MAX];
assert(field);
assert(uid);
- if (pid == 0)
- return getuid();
-
p = procfs_file_alloca(pid, "status");
f = fopen(p, "re");
- if (!f)
+ if (!f) {
+ if (errno == ENOENT)
+ return -ESRCH;
return -errno;
+ }
FOREACH_LINE(line, f, return -errno) {
char *l;
p = procfs_file_alloca(pid, "environ");
f = fopen(p, "re");
- if (!f)
+ if (!f) {
+ if (errno == ENOENT)
+ return -ESRCH;
return -errno;
+ }
while ((c = fgetc(f)) != EOF) {
if (!GREEDY_REALLOC(outcome, allocated, sz + 5))
sz += cescape_char(c, outcome + sz);
}
- outcome[sz] = '\0';
+ if (!outcome) {
+ outcome = strdup("");
+ if (!outcome)
+ return -ENOMEM;
+ } else
+ outcome[sz] = '\0';
+
*env = outcome;
outcome = NULL;
return 0;
}
-#endif // 0
-int get_parent_of_pid(pid_t pid, pid_t *_ppid) {
+int get_process_ppid(pid_t pid, pid_t *_ppid) {
int r;
_cleanup_free_ char *line = NULL;
long unsigned ppid;
p = procfs_file_alloca(pid, "stat");
r = read_one_line_file(p, &line);
+ if (r == -ENOENT)
+ return -ESRCH;
if (r < 0)
return r;
return 0;
}
+#endif // 0
int wait_for_terminate(pid_t pid, siginfo_t *status) {
siginfo_t dummy;
if (errno == EINTR)
continue;
- return -errno;
+ return negative_errno();
}
return 0;
return -EPROTO;
}
-/// UNNEEDED by elogind
-#if 0
+#if 0 /// UNNEEDED by elogind
+void sigkill_wait(pid_t pid) {
+ assert(pid > 1);
+
+ if (kill(pid, SIGKILL) > 0)
+ (void) wait_for_terminate(pid, NULL);
+}
+
+void sigkill_waitp(pid_t *pid) {
+ if (!pid)
+ return;
+ if (*pid <= 1)
+ return;
+
+ sigkill_wait(*pid);
+}
+
int kill_and_sigcont(pid_t pid, int sig) {
int r;
r = kill(pid, sig) < 0 ? -errno : 0;
- if (r >= 0)
- kill(pid, SIGCONT);
+ /* If this worked, also send SIGCONT, unless we already just sent a SIGCONT, or SIGKILL was sent which isn't
+ * affected by a process being suspended anyway. */
+ if (r >= 0 && !IN_SET(sig, SIGCONT, SIGKILL))
+ (void) kill(pid, SIGCONT);
return r;
}
path = procfs_file_alloca(pid, "environ");
f = fopen(path, "re");
- if (!f)
+ if (!f) {
+ if (errno == ENOENT)
+ return -ESRCH;
return -errno;
+ }
l = strlen(field);
r = 0;
}
line[i] = 0;
- if (memcmp(line, field, l) == 0 && line[l] == '=') {
+ if (strneq(line, field, l) && line[l] == '=') {
value = strdup(line + l + 1);
if (!value)
return -ENOMEM;
bool pid_is_unwaited(pid_t pid) {
/* Checks whether a PID is still valid at all, including a zombie */
- if (pid <= 0)
+ if (pid < 0)
return false;
+ if (pid <= 1) /* If we or PID 1 would be dead and have been waited for, this code would not be running */
+ return true;
+
if (kill(pid, 0) >= 0)
return true;
/* Checks whether a PID is still valid and not a zombie */
- if (pid <= 0)
+ if (pid < 0)
return false;
+ if (pid <= 1) /* If we or PID 1 would be a zombie, this code would not be running */
+ return true;
+
r = get_process_state(pid);
- if (r == -ENOENT || r == 'Z')
+ if (r == -ESRCH || r == 'Z')
return false;
return true;
}
+
+#if 0 /// UNNEEDED by elogind
+int pid_from_same_root_fs(pid_t pid) {
+ const char *root;
+
+ if (pid < 0)
+ return 0;
+
+ root = procfs_file_alloca(pid, "root");
+
+ return files_same(root, "/proc/1/root", 0);
+}
+#endif // 0
+
+bool is_main_thread(void) {
+ static thread_local int cached = 0;
+
+ if (_unlikely_(cached == 0))
+ cached = getpid() == gettid() ? 1 : -1;
+
+ return cached > 0;
+}
+
+#if 0 /// UNNEEDED by elogind
+noreturn void freeze(void) {
+
+ log_close();
+
+ /* Make sure nobody waits for us on a socket anymore */
+ close_all_fds(NULL, 0);
+
+ sync();
+
+ for (;;)
+ pause();
+}
+
+bool oom_score_adjust_is_valid(int oa) {
+ return oa >= OOM_SCORE_ADJ_MIN && oa <= OOM_SCORE_ADJ_MAX;
+}
+
+unsigned long personality_from_string(const char *p) {
+ int architecture;
+
+ if (!p)
+ return PERSONALITY_INVALID;
+
+ /* Parse a personality specifier. We use our own identifiers that indicate specific ABIs, rather than just
+ * hints regarding the register size, since we want to keep things open for multiple locally supported ABIs for
+ * the same register size. */
+
+ architecture = architecture_from_string(p);
+ if (architecture < 0)
+ return PERSONALITY_INVALID;
+
+ if (architecture == native_architecture())
+ return PER_LINUX;
+#ifdef SECONDARY_ARCHITECTURE
+ if (architecture == SECONDARY_ARCHITECTURE)
+ return PER_LINUX32;
+#endif
+
+ return PERSONALITY_INVALID;
+}
+
+const char* personality_to_string(unsigned long p) {
+ int architecture = _ARCHITECTURE_INVALID;
+
+ if (p == PER_LINUX)
+ architecture = native_architecture();
+#ifdef SECONDARY_ARCHITECTURE
+ else if (p == PER_LINUX32)
+ architecture = SECONDARY_ARCHITECTURE;
+#endif
+
+ if (architecture < 0)
+ return NULL;
+
+ return architecture_to_string(architecture);
+}
+
+void valgrind_summary_hack(void) {
+#ifdef HAVE_VALGRIND_VALGRIND_H
+ if (getpid() == 1 && RUNNING_ON_VALGRIND) {
+ pid_t pid;
+ pid = raw_clone(SIGCHLD);
+ if (pid < 0)
+ log_emergency_errno(errno, "Failed to fork off valgrind helper: %m");
+ else if (pid == 0)
+ exit(EXIT_SUCCESS);
+ else {
+ log_info("Spawned valgrind helper as PID "PID_FMT".", pid);
+ (void) wait_for_terminate(pid, NULL);
+ }
+ }
+#endif
+}
+
+int pid_compare_func(const void *a, const void *b) {
+ const pid_t *p = a, *q = b;
+
+ /* Suitable for usage in qsort() */
+
+ if (*p < *q)
+ return -1;
+ if (*p > *q)
+ return 1;
+ return 0;
+}
+
+static const char *const ioprio_class_table[] = {
+ [IOPRIO_CLASS_NONE] = "none",
+ [IOPRIO_CLASS_RT] = "realtime",
+ [IOPRIO_CLASS_BE] = "best-effort",
+ [IOPRIO_CLASS_IDLE] = "idle"
+};
+
+DEFINE_STRING_TABLE_LOOKUP_WITH_FALLBACK(ioprio_class, int, INT_MAX);
+
+static const char *const sigchld_code_table[] = {
+ [CLD_EXITED] = "exited",
+ [CLD_KILLED] = "killed",
+ [CLD_DUMPED] = "dumped",
+ [CLD_TRAPPED] = "trapped",
+ [CLD_STOPPED] = "stopped",
+ [CLD_CONTINUED] = "continued",
+};
+
+DEFINE_STRING_TABLE_LOOKUP(sigchld_code, int);
+
+static const char* const sched_policy_table[] = {
+ [SCHED_OTHER] = "other",
+ [SCHED_BATCH] = "batch",
+ [SCHED_IDLE] = "idle",
+ [SCHED_FIFO] = "fifo",
+ [SCHED_RR] = "rr"
+};
+
+DEFINE_STRING_TABLE_LOOKUP_WITH_FALLBACK(sched_policy, int, INT_MAX);
+#endif // 0