static inline void task_state(struct seq_file *m, struct pid_namespace *ns,
struct pid *pid, struct task_struct *p)
{
struct user_namespace *user_ns = seq_user_ns(m);
struct group_info *group_info;
int g;
struct fdtable *fdt = NULL;
const struct cred *cred;
pid_t ppid, tpid;
rcu_read_lock();
ppid = pid_alive(p) ?
task_tgid_nr_ns(rcu_dereference(p->real_parent), ns) : 0;
tpid = 0;
if (pid_alive(p)) {
struct task_struct *tracer = ptrace_parent(p);
if (tracer)
tpid = task_pid_nr_ns(tracer, ns);
}
cred = get_task_cred(p);
seq_printf(m,
"State:\t%s\n"
"Tgid:\t%d\n"
"Pid:\t%d\n"
"PPid:\t%d\n"
"TracerPid:\t%d\n"
"Uid:\t%d\t%d\t%d\t%d\n"
"Gid:\t%d\t%d\t%d\t%d\n",
get_task_state(p),
task_tgid_nr_ns(p, ns),
pid_nr_ns(pid, ns),
ppid, tpid,
from_kuid_munged(user_ns, cred->uid),
from_kuid_munged(user_ns, cred->euid),
from_kuid_munged(user_ns, cred->suid),
from_kuid_munged(user_ns, cred->fsuid),
from_kgid_munged(user_ns, cred->gid),
from_kgid_munged(user_ns, cred->egid),
from_kgid_munged(user_ns, cred->sgid),
from_kgid_munged(user_ns, cred->fsgid));
task_lock(p);
if (p->files)
fdt = files_fdtable(p->files);
seq_printf(m,
"FDSize:\t%d\n"
"Groups:\t",
fdt ? fdt->max_fds : 0);
rcu_read_unlock();
group_info = cred->group_info;
task_unlock(p);
for (g = 0; g < group_info->ngroups; g++)
seq_printf(m, "%d ",
from_kgid_munged(user_ns, GROUP_AT(group_info, g)));
put_cred(cred);
seq_putc(m, '\n');
}
static inline void task_state(struct seq_file *m, struct pid_namespace *ns,
struct pid *pid, struct task_struct *p)
{
struct group_info *group_info;
int g;
struct fdtable *fdt = NULL;
const struct cred *cred;
pid_t ppid, tpid, vpid;
rcu_read_lock();
ppid = pid_alive(p) ? ve_task_ppid_nr_ns(p, ns) : 0;
tpid = 0;
if (pid_alive(p)) {
struct task_struct *tracer = tracehook_tracer_task(p);
if (tracer)
tpid = task_pid_nr_ns(tracer, ns);
}
vpid = task_virtual_pid(p);
cred = get_task_cred(p);
seq_printf(m,
"State:\t%s\n"
"Tgid:\t%d\n"
"Pid:\t%d\n"
"PPid:\t%d\n"
"TracerPid:\t%d\n"
"Uid:\t%d\t%d\t%d\t%d\n"
"Gid:\t%d\t%d\t%d\t%d\n",
get_task_state(p),
task_tgid_nr_ns(p, ns),
pid_nr_ns(pid, ns),
ppid, tpid,
cred->uid, cred->euid, cred->suid, cred->fsuid,
cred->gid, cred->egid, cred->sgid, cred->fsgid);
task_utrace_proc_status(m, p);
task_lock(p);
if (p->files)
fdt = files_fdtable(p->files);
seq_printf(m,
"FDSize:\t%d\n"
"Groups:\t",
fdt ? fdt->max_fds : 0);
rcu_read_unlock();
group_info = cred->group_info;
task_unlock(p);
for (g = 0; g < min(group_info->ngroups, NGROUPS_SMALL); g++)
seq_printf(m, "%d ", GROUP_AT(group_info, g));
put_cred(cred);
seq_printf(m, "\n");
seq_printf(m, "envID:\t%d\nVPid:\t%d\n",
p->ve_task_info.owner_env->veid, vpid);
seq_printf(m, "StopState:\t%u\n", p->stopped_state);
}
static inline void task_state(struct seq_file *m, struct pid_namespace *ns,
struct pid *pid, struct task_struct *p)
{
struct group_info *group_info;
int g;
struct fdtable *fdt = NULL;
pid_t ppid, tpid;
rcu_read_lock();
ppid = pid_alive(p) ?
task_tgid_nr_ns(rcu_dereference(p->real_parent), ns) : 0;
tpid = 0;
if (pid_alive(p)) {
struct task_struct *tracer = tracehook_tracer_task(p);
if (tracer)
tpid = task_pid_nr_ns(tracer, ns);
}
seq_printf(m,
"State:\t%s\n"
"Tgid:\t%d\n"
"Pid:\t%d\n"
"PPid:\t%d\n"
"TracerPid:\t%d\n"
"Uid:\t%d\t%d\t%d\t%d\n"
"Gid:\t%d\t%d\t%d\t%d\n",
get_task_state(p),
task_tgid_nr_ns(p, ns),
pid_nr_ns(pid, ns),
ppid, tpid,
p->uid, p->euid, p->suid, p->fsuid,
p->gid, p->egid, p->sgid, p->fsgid);
task_lock(p);
if (p->files)
fdt = files_fdtable(p->files);
seq_printf(m,
"FDSize:\t%d\n"
"Groups:\t",
fdt ? fdt->max_fds : 0);
rcu_read_unlock();
group_info = p->group_info;
get_group_info(group_info);
task_unlock(p);
for (g = 0; g < min(group_info->ngroups, NGROUPS_SMALL); g++)
seq_printf(m, "%d ", GROUP_AT(group_info, g));
put_group_info(group_info);
seq_printf(m, "\n");
}
/* count current worker and jobs */
void count_current_worker(int restart) {
int x;
gm_log( GM_LOG_TRACE3, "count_current_worker()\n");
gm_log( GM_LOG_TRACE3, "done jobs: shm[SHM_JOBS_DONE] = %d\n", shm[SHM_JOBS_DONE]);
/* shm states:
* 0 -> undefined
* -1 -> free
* <-1 -> used but idle
* > 1 -> used and working
*/
/* check if status worker died */
if( shm[SHM_STATUS_WORKER_PID] != -1 && pid_alive(shm[SHM_STATUS_WORKER_PID]) == FALSE ) {
gm_log( GM_LOG_TRACE, "removed stale status worker, old pid: %d\n", shm[SHM_STATUS_WORKER_PID] );
shm[SHM_STATUS_WORKER_PID] = -1;
}
gm_log( GM_LOG_TRACE3, "status worker: shm[SHM_STATUS_WORKER_PID] = %d\n", shm[SHM_STATUS_WORKER_PID]);
/* check all known worker */
current_number_of_workers = 0;
current_number_of_jobs = 0;
for(x=SHM_SHIFT; x < mod_gm_opt->max_worker+SHM_SHIFT; x++) {
/* verify worker is alive */
gm_log( GM_LOG_TRACE3, "worker slot: shm[%d] = %d\n", x, shm[x]);
if( shm[x] != -1 && pid_alive(shm[x]) == FALSE ) {
gm_log( GM_LOG_TRACE, "removed stale worker %d, old pid: %d\n", x, shm[x]);
shm[x] = -1;
/* immediately start new worker, otherwise the fork rate cannot be guaranteed */
if(restart == GM_ENABLED) {
make_new_child(GM_WORKER_MULTI);
current_number_of_workers++;
}
}
if(shm[x] != -1) {
current_number_of_workers++;
}
if(shm[x] > 0) {
current_number_of_jobs++;
}
}
shm[SHM_WORKER_TOTAL] = current_number_of_workers; /* total worker */
shm[SHM_WORKER_RUNNING] = current_number_of_jobs; /* running worker */
gm_log( GM_LOG_TRACE3, "worker: %d - running: %d\n", current_number_of_workers, current_number_of_jobs);
return;
}
/*
* Check that the processes holding locks are still alive.
*/
static bool check_lock(lock_t *_lock)
{
pid_t pid;
/* We don't care about unlocked or locking-in-progress */
if (_lock->lock != LOCKED)
return FALSE;
/* First the easy case. If it's held by a dead pid, release it. */
pid = _lock->owner;
/* if we're in the process of unlocking, it can show up as LOCKED
* but with no owner. Just bail, we'll try again next time around.
*/
if (pid == 0)
return FALSE;
if (pid_alive(pid) == FALSE) {
if (errno != ESRCH)
return TRUE;
debugf("Found a lock held by dead pid %d. Freeing.\n", pid);
unlock(_lock);
return TRUE;
}
return FALSE;
}
static void check_lock(lock_t *_lock)
{
pid_t pid;
if (_lock->lock != LOCKED)
return;
/* First the easy case. If it's held by a dead pid, release it. */
pid = _lock->owner;
if (pid_alive(pid) == -1) {
if (errno != ESRCH)
return;
debugf("Found a lock held by dead pid %d. Freeing.\n", pid);
unlock(_lock);
return;
}
/* If a pid has had a lock a long time, something is up. */
if (_lock->contention > STEAL_THRESHOLD) {
debugf("pid %d has held lock for too long. Releasing, and killing.\n", pid);
kill_pid(pid);
unlock(_lock);
return;
}
return;
}
/* send kill to all forked processes */
void kill_child_checks(void) {
int retval;
pid_t pid;
signal(SIGINT, SIG_IGN);
pid = getpid();
if(current_child_pid > 0 && current_child_pid != pid) {
gm_log( GM_LOG_TRACE, "kill_child_checks(): send SIGINT to %d\n", current_child_pid);
kill(-current_child_pid, SIGINT);
kill(current_child_pid, SIGINT);
sleep(1);
if(waitpid(current_child_pid,&retval,WNOHANG)!=0) {
signal(SIGINT, SIG_DFL);
return;
}
if(pid_alive(current_child_pid)) {
gm_log( GM_LOG_TRACE, "kill_child_checks(): send SIGKILL to %d\n", current_child_pid);
kill(current_child_pid, SIGKILL);
}
}
gm_log( GM_LOG_TRACE, "send SIGINT to %d\n", pid);
kill(0, SIGINT);
signal(SIGINT, SIG_DFL);
return;
}
/* Main function for TLB walkthrough
* Check accessed page every Memmap_wakeupsIinterval ms
*/
int Moca_MonitorThread(void * arg)
{
task_data data;
moca_task t;
struct task_struct * task;
//Init tlb walk data
int pos;
unsigned long long lastwake=0;
MOCA_DEBUG_PRINT("Moca monitor thread alive \n");
while(!kthread_should_stop())
{
pos=0;
while((t=Moca_NextTask(&pos)))
{
data=t->data;
task=(struct task_struct *)(t->key);
MOCA_DEBUG_PRINT("Moca monitor thread testing task %p\n", task);
if(pid_alive(task) && task->sched_info.last_arrival >= lastwake)
{
lastwake=task->sched_info.last_arrival;
MOCA_DEBUG_PRINT("Moca monitor thread found task %p\n",task);
Moca_MonitorPage(data);
}
}
Moca_UpdateClock();
MOCA_DEBUG_PRINT("Moca monitor thread going to sleep for %d\n",
Moca_wakeupInterval);
msleep(Moca_wakeupInterval);
}
MOCA_DEBUG_PRINT("Moca monitor thread finished\n");
return 0;
}
// Add pid to the monitored process if pid is a monitored process
moca_task Moca_AddTaskIfNeeded(struct task_struct *t)
{
moca_task ret=NULL;
if(t && pid_alive(t) && t->real_parent && Moca_ShouldMonitorTask(t))
ret=Moca_AddTask(t);
return ret;
}
/*
* Accumulate raw cputime values of dead tasks (sig->[us]time) and live
* tasks (sum on group iteration) belonging to @tsk's group.
*/
void thread_group_cputime(struct task_struct *tsk, struct task_cputime *times)
{
struct signal_struct *sig = tsk->signal;
cputime_t utime, stime;
struct task_struct *t;
times->utime = sig->utime;
times->stime = sig->stime;
times->sum_exec_runtime = sig->sum_sched_runtime;
rcu_read_lock();
/* make sure we can trust tsk->thread_group list */
if (!likely(pid_alive(tsk)))
goto out;
t = tsk;
do {
task_cputime(tsk, &utime, &stime);
times->utime += utime;
times->stime += stime;
times->sum_exec_runtime += task_sched_runtime(t);
} while_each_thread(tsk, t);
out:
rcu_read_unlock();
}
void init_child(int childno)
{
cpu_set_t set;
pid_t pid = getpid();
char childname[17];
this_child = childno;
set_seed(childno);
shm->kill_count[childno] = 0;
shm->num_mappings[childno] = 0;
shm->mappings[childno] = zmalloc(sizeof(struct map));
INIT_LIST_HEAD(&shm->mappings[childno]->list);
setup_page_maps();
if (sched_getaffinity(pid, sizeof(set), &set) == 0) {
CPU_ZERO(&set);
CPU_SET(childno, &set);
sched_setaffinity(pid, sizeof(set), &set);
}
shm->child_syscall_count[childno] = 0;
memset(childname, 0, sizeof(childname));
sprintf(childname, "trinity-c%d", childno);
prctl(PR_SET_NAME, (unsigned long) &childname);
oom_score_adj(500);
/* Wait for parent to set our pidslot */
while (shm->pids[childno] != getpid()) {
int ret = 0;
/* Make sure parent is actually alive to wait for us. */
ret = pid_alive(shm->mainpid);
if (ret != 0) {
shm->exit_reason = EXIT_SHM_CORRUPTION;
outputerr(BUGTXT "parent (%d) went away!\n", shm->mainpid);
sleep(20000);
}
}
/* Wait for all the children to start up. */
while (shm->ready == FALSE)
sleep(1);
set_make_it_fail();
if (rand() % 100 < 50)
use_fpu();
mask_signals_child();
disable_coredumps();
}
static inline char * task_state(struct task_struct *p, char *buffer)
{
struct group_info *group_info;
int g;
struct fdtable *fdt = NULL;
read_lock(&tasklist_lock);
buffer += sprintf(buffer,
"State:\t%s\n"
"SleepAVG:\t%lu%%\n"
"Tgid:\t%d\n"
"Pid:\t%d\n"
"PPid:\t%d\n"
"TracerPid:\t%d\n"
"Uid:\t%d\t%d\t%d\t%d\n"
"Gid:\t%d\t%d\t%d\t%d\n",
get_task_state(p),
(p->sleep_avg/1024)*100/(1020000000/1024),
p->tgid,
p->pid, pid_alive(p) ? p->group_leader->real_parent->tgid : 0,
pid_alive(p) && p->ptrace ? p->parent->pid : 0,
p->uid, p->euid, p->suid, p->fsuid,
p->gid, p->egid, p->sgid, p->fsgid);
read_unlock(&tasklist_lock);
task_lock(p);
rcu_read_lock();
if (p->files)
fdt = files_fdtable(p->files);
buffer += sprintf(buffer,
"FDSize:\t%d\n"
"Groups:\t",
fdt ? fdt->max_fds : 0);
rcu_read_unlock();
group_info = p->group_info;
get_group_info(group_info);
task_unlock(p);
for (g = 0; g < min(group_info->ngroups,NGROUPS_SMALL); g++)
buffer += sprintf(buffer, "%d ", GROUP_AT(group_info,g));
put_group_info(group_info);
buffer += sprintf(buffer, "\n");
return buffer;
}
static inline void task_state(struct seq_file *m, struct pid *pid,
struct task_struct *p)
{
struct group_info *group_info;
int g;
struct fdtable *fdt = NULL;
rcu_read_lock();
seq_printf(m,
"State:\t%s\n"
"SleepAVG:\t%lu%%\n"
"Tgid:\t%d\n"
"Pid:\t%d\n"
"PPid:\t%d\n"
"TracerPid:\t%d\n"
"Uid:\t%d\t%d\t%d\t%d\n"
"Gid:\t%d\t%d\t%d\t%d\n",
get_task_state(p),
(p->sleep_avg/1024)*100/(1020000000/1024),
p->tgid, p->pid,
pid_alive(p) ? rcu_dereference(p->real_parent)->tgid : 0,
pid_alive(p) && p->ptrace ? rcu_dereference(p->parent)->pid : 0,
p->uid, p->euid, p->suid, p->fsuid,
p->gid, p->egid, p->sgid, p->fsgid);
task_lock(p);
if (p->files)
fdt = files_fdtable(p->files);
seq_printf(m,
"FDSize:\t%d\n"
"Groups:\t",
fdt ? fdt->max_fds : 0);
rcu_read_unlock();
group_info = p->group_info;
get_group_info(group_info);
task_unlock(p);
for (g = 0; g < min(group_info->ngroups, NGROUPS_SMALL); g++)
seq_printf(m, "%d ", GROUP_AT(group_info, g));
put_group_info(group_info);
seq_printf(m, "\n");
}
/* This is a special case for things like execve, which would replace our
* child process with something unknown to us. We use a 'throwaway' process
* to do the execve in, and let it run for a max of a second before we kill it
*/
static void do_extrafork(struct syscallrecord *rec)
{
pid_t pid = 0;
pid_t extrapid;
extrapid = fork();
if (extrapid == 0) {
/* grand-child */
char childname[]="trinity-subchild";
prctl(PR_SET_NAME, (unsigned long) &childname);
__do_syscall(rec, GOING_AWAY);
/* if this was for eg. an successful execve, we should never get here.
* if it failed though... */
_exit(EXIT_SUCCESS);
}
/* misc failure. */
if (extrapid == -1) {
//debugf("Couldn't fork grandchild: %s\n", strerror(errno));
return;
}
/* small pause to let grandchild do some work. */
if (pid_alive(extrapid) == TRUE)
usleep(100);
/* We take the rec lock here even though we don't obviously use it.
* The reason, is that the grandchild is using it. */
lock(&rec->lock);
while (pid == 0) {
int childstatus;
pid = waitpid(extrapid, &childstatus, WUNTRACED | WCONTINUED | WNOHANG);
if (pid_alive(extrapid) == TRUE)
kill(extrapid, SIGKILL);
usleep(1000);
}
unlock(&rec->lock);
}
void bacct_add_tsk(struct taskstats *stats, struct task_struct *tsk)
{
const struct cred *tcred;
struct timespec uptime, ts;
u64 ac_etime;
BUILD_BUG_ON(TS_COMM_LEN < TASK_COMM_LEN);
do_posix_clock_monotonic_gettime(&uptime);
ts = timespec_sub(uptime, tsk->start_time);
ac_etime = timespec_to_ns(&ts);
do_div(ac_etime, NSEC_PER_USEC);
stats->ac_etime = ac_etime;
stats->ac_btime = get_seconds() - ts.tv_sec;
if (thread_group_leader(tsk)) {
stats->ac_exitcode = tsk->exit_code;
if (tsk->flags & PF_FORKNOEXEC)
stats->ac_flag |= AFORK;
}
if (tsk->flags & PF_SUPERPRIV)
stats->ac_flag |= ASU;
if (tsk->flags & PF_DUMPCORE)
stats->ac_flag |= ACORE;
if (tsk->flags & PF_SIGNALED)
stats->ac_flag |= AXSIG;
stats->ac_nice = task_nice(tsk);
stats->ac_sched = tsk->policy;
stats->ac_pid = tsk->pid;
rcu_read_lock();
tcred = __task_cred(tsk);
stats->ac_uid = tcred->uid;
stats->ac_gid = tcred->gid;
stats->ac_ppid = pid_alive(tsk) ?
rcu_dereference(tsk->real_parent)->tgid : 0;
rcu_read_unlock();
stats->ac_utime = cputime_to_msecs(tsk->utime) * USEC_PER_MSEC;
stats->ac_stime = cputime_to_msecs(tsk->stime) * USEC_PER_MSEC;
stats->ac_utimescaled =
cputime_to_msecs(tsk->utimescaled) * USEC_PER_MSEC;
stats->ac_stimescaled =
cputime_to_msecs(tsk->stimescaled) * USEC_PER_MSEC;
stats->ac_minflt = tsk->min_flt;
stats->ac_majflt = tsk->maj_flt;
strncpy(stats->ac_comm, tsk->comm, sizeof(stats->ac_comm));
}
static void timer_function(unsigned long par)
{
ushort cpu_share;
if (unlikely(!pid_alive(check_task))) {
del_timer(&check_timer);
printk(KERN_INFO "sendsig: cannot find pid %i. Is the process still active? Timer removed\n", pid);
return;
}
cpu_share = thread_group_cpu_share(check_task);
if (cpu_share >= max_cpu_share) {
count_check++;
printk(KERN_INFO "sendsig: current cpu share over limit of %i (check #%i)\n",
max_cpu_share, count_check);
/* the ratio is: if the process has a cpu share higher than
max_cpu_share for more than max_checks * wait_timeout seconds, then
we'll send the signal sig_to_send to it
*/
if (count_check >= max_checks) {
/*
sending the signal to the process
*/
signal_send(check_task);
/*
remove the timer
*/
del_timer(&check_timer);
printk(KERN_INFO "sendsig: sent signal to process %i, timer removed\n", pid);
return;
}
} else {
/*
if the process is being good, let's reset its counter
*/
count_check = 0;
}
/*
update the timer
*/
mod_timer(&check_timer, jiffies + wait_timeout * HZ);
return;
}
static void restart_lost_procs(void)
{
svc_t *svc;
for (svc = svc_iterator(1); svc; svc = svc_iterator(0)) {
if (svc->pid > 0 && pid_alive(svc->pid))
continue;
/* Only restart lost daemons, not task/run/inetd services */
if (SVC_TYPE_SERVICE != svc->type) {
svc->pid = 0;
continue;
}
service_start(svc);
}
}
static void post_fork(struct syscallrecord *rec)
{
pid_t pid;
pid = rec->retval;
if (pid == 0) {
// child
sleep(1);
_exit(EXIT_SUCCESS);
} else {
__unused__ int ret;
while (pid_alive(pid)) {
int status;
ret = waitpid(pid, &status, WUNTRACED | WCONTINUED | WNOHANG);
}
}
}
static inline void task_state(struct seq_file *m, struct pid_namespace *ns,
struct pid *pid, struct task_struct *p)
{
struct group_info *group_info;
int g;
struct fdtable *fdt = NULL;
const struct cred *cred;
pid_t ppid, tpid;
rcu_read_lock();
ppid = pid_alive(p) ?
task_tgid_nr_ns(rcu_dereference(p->real_parent), ns) : 0;
tpid = 0;
if (pid_alive(p)) {
<<<<<<< HEAD
struct task_struct *tracer = ptrace_parent(p);
=======
<<<<<<< HEAD
static void post_fork(int childno)
{
pid_t pid;
pid = shm->syscall[childno].retval;
if (pid == 0) {
// child
sleep(1);
_exit(EXIT_SUCCESS);
} else {
__unused__ int ret;
while (pid_alive(pid)) {
int status;
ret = waitpid(pid, &status, WUNTRACED | WCONTINUED | WNOHANG);
}
}
}
int monitor_fn(void* unused)
{
tm->target_task = get_pid_task(tm->target_pid, PIDTYPE_PID);
while(tm->target_task && pid_alive(tm->target_task)
&& !kthread_should_stop()){
set_current_state(TASK_INTERRUPTIBLE);
schedule_timeout(HZ);
pr_info("taskmonitor: pid %d\tusr %d\tsys %d\n", target,
(int)tm->target_task->utime,
(int)tm->target_task->stime);
}
if(tm->target_task)
put_task_struct(tm->target_task);
pr_warn("monitor_fn: target task is no longer alive !\n");
return 0;
}
void lock(lock_t *_lock)
{
while (_lock->lock == LOCKED) {
_lock->contention++;
usleep(1);
if (_lock->contention > STEAL_THRESHOLD) {
pid_t pid = _lock->owner;
if (pid_alive(pid) == FALSE) {
output(0, "[%d] more than %d attempts to get lock. pid %d looks dead, stealing.\n",
getpid(), STEAL_THRESHOLD, pid);
goto steal;
}
}
}
steal:
_lock->contention = 0;
_lock->lock = LOCKED;
_lock->owner = getpid();
}
/* Used by wait */
static void wait_checker(struct work_struct *w)
{
struct delayed_work * w_del;
struct wait_struct * w_stru;
int i;
w_del = to_delayed_work(w);
w_stru = container_of(w_del, struct wait_struct, w_delayed);
wake_up_interruptible(&waiter);
for (i = 0; i < w_stru->w_size; i++) {
if (!pid_alive(w_stru->w_pids[i])) {
w_stru->w_is_finished = i;
wake_up_interruptible(&waiter);
return;
}
}
/* On relance la boucle */
queue_delayed_work(syndicate, &(wOs->w_dw), TICKS_DELAY);
}
void dump_childnos(void)
{
unsigned int i, j = 0;
char string[512], *sptr = string;
sptr += sprintf(sptr, "## pids: (%u active)\n", shm->running_childs);
for (i = 0; i < max_children; i += 8) {
sptr += sprintf(sptr, "%u-%u: ", i, i + 7);
for (j = 0; j < 8; j++) {
struct childdata *child;
if (i + j > max_children)
break;
child = shm->children[i + j];
if (pids[child->num] == EMPTY_PIDSLOT) {
sptr += sprintf(sptr, "[empty] ");
} else {
pid_t pid = pids[child->num];
if (pid_is_valid(pid) == FALSE)
sptr += sprintf(sptr, "%s", ANSI_RED);
if (pid_alive(pid) == FALSE)
sptr += sprintf(sptr, "%s", ANSI_RED);
sptr += sprintf(sptr, "%u %s", pid, ANSI_RESET);
}
}
sptr += sprintf(sptr, "\n");
*sptr = '\0';
outputerr("%s", string);
sptr = string;
}
}
static void fork_children(void)
{
int pidslot;
static char childname[17];
/* Generate children*/
while (shm->running_childs < shm->max_children) {
int pid = 0;
/* Find a space for it in the pid map */
pidslot = find_pid_slot(EMPTY_PIDSLOT);
if (pidslot == PIDSLOT_NOT_FOUND) {
printf("[%d] ## Pid map was full!\n", getpid());
dump_pid_slots();
exit(EXIT_FAILURE);
}
(void)alarm(0);
fflush(stdout);
pid = fork();
if (pid != 0)
shm->pids[pidslot] = pid;
else {
/* Child process. */
int ret = 0;
mask_signals_child();
memset(childname, 0, sizeof(childname));
sprintf(childname, "trinity-child%d", pidslot);
prctl(PR_SET_NAME, (unsigned long) &childname);
oom_score_adj(500);
/* Wait for parent to set our pidslot */
while (shm->pids[pidslot] != getpid()) {
/* Make sure parent is actually alive to wait for us. */
ret = pid_alive(shm->parentpid);
if (ret != 0) {
shm->exit_reason = EXIT_SHM_CORRUPTION;
printf("[%d] " BUGTXT "parent (%d) went away!\n", getpid(), shm->parentpid);
sleep(20000);
}
}
init_child(pidslot);
ret = child_process(pidslot);
output(1, "child %d exiting\n", getpid());
_exit(ret);
}
shm->running_childs++;
debugf("[%d] Created child %d in pidslot %d [total:%d/%d]\n",
getpid(), shm->pids[pidslot], pidslot,
shm->running_childs, shm->max_children);
if (shm->exit_reason != STILL_RUNNING)
return;
}
debugf("[%d] created enough children\n", getpid());
}
/* main tests */
int main (int argc, char **argv, char **env) {
argc = argc; argv = argv; env = env;
int status, chld, rc;
int tests = 125;
int rrc;
char cmd[150];
char *result, *error, *message, *output;
plan(tests);
mod_gm_opt = malloc(sizeof(mod_gm_opt_t));
set_default_options(mod_gm_opt);
#ifdef EMBEDDEDPERL
char p1[150];
snprintf(p1, 150, "--p1_file=worker/mod_gearman_p1.pl");
parse_args_line(mod_gm_opt, p1, 0);
init_embedded_perl(env);
#endif
char options[150];
snprintf(options, 150, "--server=127.0.0.1:%d", GEARMAND_TEST_PORT);
ok(parse_args_line(mod_gm_opt, options, 0) == 0, "parse_args_line()");
mod_gm_opt->debug_level = GM_LOG_ERROR;
worker_logfile = my_tmpfile();
if(!ok(worker_logfile != NULL, "created temp logile: %s", worker_logfile)) {
diag("could not create temp logfile");
exit( EXIT_FAILURE );
}
/* first fire up a gearmand server and one worker */
start_gearmand((void*)NULL);
sleep(2);
start_worker((void*)NULL);
sleep(2);
/* wait one second and catch died procs */
while((chld = waitpid(-1, &status, WNOHANG)) != -1 && chld > 0) {
diag( "waitpid() %d exited with %d\n", chld, status);
status = 0;
}
if(!ok(gearmand_pid > 0, "'gearmand started with port %d and pid: %d", GEARMAND_TEST_PORT, gearmand_pid)) {
diag("make sure gearmand is in your PATH. Common locations are /usr/sbin or /usr/local/sbin");
exit( EXIT_FAILURE );
}
if(!ok(pid_alive(gearmand_pid) == TRUE, "gearmand alive")) {
check_logfile("/tmp/gearmand.log", 3);
kill(gearmand_pid, SIGTERM);
kill(worker_pid, SIGTERM);
exit( EXIT_FAILURE );
}
if(!ok(worker_pid > 0, "worker started with pid: %d", worker_pid))
diag("could not start worker");
if(!ok(pid_alive(worker_pid) == TRUE, "worker alive")) {
check_logfile(worker_logfile, 3);
kill(gearmand_pid, SIGTERM);
kill(worker_pid, SIGTERM);
exit( EXIT_FAILURE );
}
skip(gearmand_pid <= 0 || worker_pid <= 0,
tests-3, /* Number of tests to skip */
"Skipping all tests, no need to go on without gearmand or worker");
/* create server / clients */
mod_gm_opt->transportmode = GM_ENCODE_ONLY;
create_modules();
/* send big job */
send_big_jobs(GM_ENCODE_ONLY);
//diag_queues();
wait_for_empty_queue("eventhandler", 20);
wait_for_empty_queue("service", 20);
//diag_queues();
do_result_work(1);
//diag_queues();
wait_for_empty_queue(GM_DEFAULT_RESULT_QUEUE, 5);
/*****************************************
* test check
*/
//diag_queues();
test_servicecheck(GM_ENCODE_ONLY, "./t/crit.pl");
//diag_queues();
wait_for_empty_queue("eventhandler", 20);
wait_for_empty_queue("service", 5);
//diag_queues();
do_result_work(1);
//diag_queues();
wait_for_empty_queue(GM_DEFAULT_RESULT_QUEUE, 5);
//diag_queues();
like(last_result, "test plugin CRITICAL", "stdout output from ./t/crit.pl");
like(last_result, "some errors on stderr", "stderr output from ./t/crit.pl");
/*****************************************
* test check2
*/
//diag_queues();
test_servicecheck(GM_ENCODE_ONLY, "./t/both");
//diag_queues();
wait_for_empty_queue("eventhandler", 20);
wait_for_empty_queue("service", 5);
//diag_queues();
do_result_work(1);
//diag_queues();
wait_for_empty_queue(GM_DEFAULT_RESULT_QUEUE, 5);
like(last_result, "stdout output", "stdout output from ./t/both");
like(last_result, "stderr output", "stderr output from ./t/both");
/* try to send some data with base64 only */
//diag_queues();
test_eventhandler(GM_ENCODE_ONLY);
//diag_queues();
test_servicecheck(GM_ENCODE_ONLY, NULL);
//diag_queues();
wait_for_empty_queue("eventhandler", 20);
wait_for_empty_queue("service", 5);
//diag_queues();
do_result_work(1);
//diag_queues();
wait_for_empty_queue(GM_DEFAULT_RESULT_QUEUE, 5);
sleep(1);
kill(worker_pid, SIGTERM);
waitpid(worker_pid, &status, 0);
ok(status == 0, "worker (%d) exited with exit code %d", worker_pid, real_exit_code(status));
status = 0;
check_no_worker_running(worker_logfile);
check_logfile(worker_logfile, 0);
char * test_keys[] = {
"12345",
"test",
"test key 123",
"me make you loooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooong key"
};
/* ignore some signals for now */
signal(SIGTERM, SIG_IGN);
int i;
for(i=0;i<4;i++) {
mod_gm_opt->transportmode = GM_ENCODE_AND_ENCRYPT;
start_worker((void *)test_keys[i]);
mod_gm_crypt_init( test_keys[i] );
ok(1, "initialized with key: %s", test_keys[i]);
test_eventhandler(GM_ENCODE_AND_ENCRYPT);
test_servicecheck(GM_ENCODE_AND_ENCRYPT, NULL);
wait_for_empty_queue("eventhandler", 20);
wait_for_empty_queue("service", 5);
do_result_work(1);
wait_for_empty_queue(GM_DEFAULT_RESULT_QUEUE, 5);
sleep(1);
kill(worker_pid, SIGTERM);
waitpid(worker_pid, &status, 0);
ok(status == 0, "worker (%d) exited with exit code %d", worker_pid, real_exit_code(status));
status = 0;
check_no_worker_running(worker_logfile);
check_logfile(worker_logfile, 0);
}
/*****************************************
* send_gearman
*/
snprintf(cmd, 150, "./send_gearman --server=127.0.0.1:%d --key=testtest --host=test --service=test --message=test --returncode=0", GEARMAND_TEST_PORT);
rrc = real_exit_code(run_check(cmd, &result, &error));
cmp_ok(rrc, "==", 0, "cmd '%s' returned rc %d", cmd, rrc);
like(result, "^\\s*$", "output from ./send_gearman");
free(result);
free(error);
/*****************************************
* send_multi
*/
snprintf(cmd, 150, "./send_multi --server=127.0.0.1:%d --host=blah < t/data/send_multi.txt", GEARMAND_TEST_PORT);
rrc = real_exit_code(run_check(cmd, &result, &error));
cmp_ok(rrc, "==", 0, "cmd '%s' returned rc %d", cmd, rrc);
like(result, "send_multi OK: 2 check_multi child checks submitted", "output from ./send_multi");
free(result);
free(error);
/*****************************************
* check_gearman
*/
snprintf(cmd, 150, "./check_gearman -H 127.0.0.1:%d -s check -a -q worker_test", GEARMAND_TEST_PORT);
rrc = real_exit_code(run_check(cmd, &result, &error));
cmp_ok(rrc, "==", 0, "cmd '%s' returned rc %d", cmd, rrc);
like(result, "check_gearman OK - sending background job succeded", "output from ./check_gearman");
/* cleanup */
free(result);
free(error);
free_client(&client);
free_worker(&worker);
/* shutdown gearmand */
rc = send2gearmandadmin("shutdown\n", "127.0.0.1", GEARMAND_TEST_PORT, &output, &message);
ok(rc == 0, "rc of send2gearmandadmin %d", rc);
like(output, "OK", "output contains OK");
free(message);
free(output);
/* wait 5 seconds to shutdown */
for(i=0;i<=5;i++) {
waitpid(gearmand_pid, &status, WNOHANG);
if(pid_alive(gearmand_pid) == FALSE) {
todo();
ok(status == 0, "gearmand (%d) exited with: %d", gearmand_pid, real_exit_code(status));
endtodo;
break;
}
sleep(1);
}
if(pid_alive(gearmand_pid) == TRUE) {
/* kill it the hard way */
kill(gearmand_pid, SIGTERM);
waitpid(gearmand_pid, &status, 0);
ok(status == 0, "gearmand (%d) exited with exit code %d", gearmand_pid, real_exit_code(status));
status = 0;
ok(false, "gearmand had to be killed!");
}
todo();
check_logfile("/tmp/gearmand.log", status != 0 ? 2 : 0);
endtodo;
status = 0;
kill(worker_pid, SIGTERM);
waitpid(worker_pid, &status, 0);
ok(status == 0, "worker (%d) exited with exit code %d", worker_pid, real_exit_code(status));
check_no_worker_running(worker_logfile);
status = 0;
#ifdef EMBEDDEDPERL
deinit_embedded_perl(0);
#endif
free(last_result);
free(worker_logfile);
endskip;
mod_gm_free_opt(mod_gm_opt);
return exit_status();
}
static void handle_child(pid_t childpid, int childstatus)
{
switch (childpid) {
case 0:
//debugf("Nothing changed. children:%d\n", shm->running_childs);
break;
case -1:
if (shm->exit_reason != STILL_RUNNING)
return;
if (errno == ECHILD) {
unsigned int i;
bool seen = FALSE;
debugf("All children exited!\n");
for_each_child(i) {
struct childdata *child;
child = shm->children[i];
if (child->pid != EMPTY_PIDSLOT) {
if (pid_alive(child->pid) == -1) {
debugf("Removing %d from pidmap\n", child->pid);
child->pid = EMPTY_PIDSLOT;
shm->running_childs--;
} else {
debugf("%d looks still alive! ignoring.\n", child->pid);
}
seen = TRUE;
}
}
if (seen == FALSE)
shm->running_childs = 0;
break;
}
output(0, "error! (%s)\n", strerror(errno));
break;
default:
debugf("Something happened to pid %d\n", childpid);
if (WIFEXITED(childstatus)) {
int childno;
childno = find_childno(childpid);
if (childno != CHILD_NOT_FOUND) {
debugf("Child %d exited after %ld operations.\n",
childpid, shm->children[childno]->syscall.op_nr);
reap_child(childpid);
}
break;
} else if (WIFSIGNALED(childstatus)) {
handle_childsig(childpid, childstatus, FALSE);
} else if (WIFSTOPPED(childstatus)) {
handle_childsig(childpid, childstatus, TRUE);
} else if (WIFCONTINUED(childstatus)) {
break;
} else {
output(0, "erk, wtf\n");
}
}
static void fork_children(void)
{
int pidslot;
static char childname[17];
/* Generate children*/
while (shm->running_childs < shm->max_children) {
int pid = 0;
int fd;
if (shm->spawn_no_more == TRUE)
return;
/* a new child means a new seed, or the new child
* will do the same syscalls as the one in the pidslot it's replacing.
* (special case startup, or we reseed unnecessarily)
*/
if (shm->ready == TRUE)
reseed();
/* Find a space for it in the pid map */
pidslot = find_pid_slot(EMPTY_PIDSLOT);
if (pidslot == PIDSLOT_NOT_FOUND) {
outputerr("## Pid map was full!\n");
dump_pid_slots();
exit(EXIT_FAILURE);
}
if (logging == TRUE) {
fd = fileno(shm->logfiles[pidslot]);
if (ftruncate(fd, 0) == 0)
lseek(fd, 0, SEEK_SET);
}
(void)alarm(0);
fflush(stdout);
pid = fork();
if (pid != 0)
shm->pids[pidslot] = pid;
else {
/* Child process. */
int ret = 0;
mask_signals_child();
memset(childname, 0, sizeof(childname));
sprintf(childname, "trinity-child%d", pidslot);
prctl(PR_SET_NAME, (unsigned long) &childname);
oom_score_adj(500);
/* Wait for parent to set our pidslot */
while (shm->pids[pidslot] != getpid()) {
/* Make sure parent is actually alive to wait for us. */
ret = pid_alive(shm->mainpid);
if (ret != 0) {
shm->exit_reason = EXIT_SHM_CORRUPTION;
outputerr(BUGTXT "parent (%d) went away!\n", shm->mainpid);
sleep(20000);
}
}
/* Wait for all the children to start up. */
while (shm->ready == FALSE)
sleep(1);
init_child(pidslot);
ret = child_process(pidslot);
output(1, "child exiting.\n");
_exit(ret);
}
shm->running_childs++;
debugf("Created child %d in pidslot %d [total:%d/%d]\n",
shm->pids[pidslot], pidslot,
shm->running_childs, shm->max_children);
if (shm->exit_reason != STILL_RUNNING)
return;
}
shm->ready = TRUE;
debugf("created enough children\n");
}
static void handle_child(pid_t childpid, int childstatus)
{
unsigned int i;
int slot;
switch (childpid) {
case 0:
//debugf("Nothing changed. children:%d\n", shm->running_childs);
break;
case -1:
if (shm->exit_reason != STILL_RUNNING)
return;
if (errno == ECHILD) {
debugf("All children exited!\n");
for_each_pidslot(i) {
if (shm->pids[i] != EMPTY_PIDSLOT) {
if (pid_alive(shm->pids[i]) == -1) {
debugf("Removing %d from pidmap\n", shm->pids[i]);
shm->pids[i] = EMPTY_PIDSLOT;
shm->running_childs--;
} else {
debugf("%d looks still alive! ignoring.\n", shm->pids[i]);
}
}
}
break;
}
output(0, "error! (%s)\n", strerror(errno));
break;
default:
debugf("Something happened to pid %d\n", childpid);
if (WIFEXITED(childstatus)) {
slot = find_pid_slot(childpid);
if (slot == PIDSLOT_NOT_FOUND) {
/* If we reaped it, it wouldn't show up, so check that. */
if (shm->last_reaped != childpid) {
outputerr("## Couldn't find pid slot for %d\n", childpid);
shm->exit_reason = EXIT_LOST_PID_SLOT;
dump_pid_slots();
}
} else {
debugf("Child %d exited after %ld syscalls.\n", childpid, shm->child_syscall_count[slot]);
reap_child(childpid);
}
break;
} else if (WIFSIGNALED(childstatus)) {
switch (WTERMSIG(childstatus)) {
case SIGALRM:
debugf("got a alarm signal from pid %d\n", childpid);
break;
case SIGFPE:
case SIGSEGV:
case SIGKILL:
case SIGPIPE:
case SIGABRT:
debugf("got a signal from pid %d (%s)\n", childpid, strsignal(WTERMSIG(childstatus)));
reap_child(childpid);
break;
default:
debugf("** Child got an unhandled signal (%d)\n", WTERMSIG(childstatus));
break;
}
break;
} else if (WIFSTOPPED(childstatus)) {
switch (WSTOPSIG(childstatus)) {
case SIGALRM:
debugf("got an alarm signal from pid %d\n", childpid);
break;
case SIGSTOP:
debugf("Sending PTRACE_DETACH (and then KILL)\n");
ptrace(PTRACE_DETACH, childpid, NULL, NULL);
kill(childpid, SIGKILL);
reap_child(childpid);
break;
case SIGFPE:
case SIGSEGV:
case SIGKILL:
case SIGPIPE:
case SIGABRT:
debugf("Child %d was stopped by %s\n", childpid, strsignal(WTERMSIG(childstatus)));
reap_child(childpid);
break;
default:
debugf("Child %d was stopped by unhandled signal (%s).\n", childpid, strsignal(WSTOPSIG(childstatus)));
break;
}
break;
} else if (WIFCONTINUED(childstatus)) {
break;
} else {
output(0, "erk, wtf\n");
}
}
static inline void task_state(struct seq_file *m, struct pid_namespace *ns,
struct pid *pid, struct task_struct *p)
{
struct user_namespace *user_ns = seq_user_ns(m);
struct group_info *group_info;
int g;
struct task_struct *tracer;
const struct cred *cred;
pid_t ppid, tpid = 0, tgid, ngid;
unsigned int max_fds = 0;
rcu_read_lock();
ppid = pid_alive(p) ?
task_tgid_nr_ns(rcu_dereference(p->real_parent), ns) : 0;
tracer = ptrace_parent(p);
if (tracer)
tpid = task_pid_nr_ns(tracer, ns);
tgid = task_tgid_nr_ns(p, ns);
ngid = task_numa_group_id(p);
cred = get_task_cred(p);
task_lock(p);
if (p->files)
max_fds = files_fdtable(p->files)->max_fds;
task_unlock(p);
rcu_read_unlock();
seq_printf(m,
"State:\t%s\n"
"Tgid:\t%d\n"
"Ngid:\t%d\n"
"Pid:\t%d\n"
"PPid:\t%d\n"
"TracerPid:\t%d\n"
"Uid:\t%d\t%d\t%d\t%d\n"
"Gid:\t%d\t%d\t%d\t%d\n"
"FDSize:\t%d\nGroups:\t",
get_task_state(p),
tgid, ngid, pid_nr_ns(pid, ns), ppid, tpid,
from_kuid_munged(user_ns, cred->uid),
from_kuid_munged(user_ns, cred->euid),
from_kuid_munged(user_ns, cred->suid),
from_kuid_munged(user_ns, cred->fsuid),
from_kgid_munged(user_ns, cred->gid),
from_kgid_munged(user_ns, cred->egid),
from_kgid_munged(user_ns, cred->sgid),
from_kgid_munged(user_ns, cred->fsgid),
max_fds);
group_info = cred->group_info;
for (g = 0; g < group_info->ngroups; g++)
seq_printf(m, "%d ",
from_kgid_munged(user_ns, GROUP_AT(group_info, g)));
put_cred(cred);
#ifdef CONFIG_PID_NS
seq_puts(m, "\nNStgid:");
for (g = ns->level; g <= pid->level; g++)
seq_printf(m, "\t%d",
task_tgid_nr_ns(p, pid->numbers[g].ns));
seq_puts(m, "\nNSpid:");
for (g = ns->level; g <= pid->level; g++)
seq_printf(m, "\t%d",
task_pid_nr_ns(p, pid->numbers[g].ns));
seq_puts(m, "\nNSpgid:");
for (g = ns->level; g <= pid->level; g++)
seq_printf(m, "\t%d",
task_pgrp_nr_ns(p, pid->numbers[g].ns));
seq_puts(m, "\nNSsid:");
for (g = ns->level; g <= pid->level; g++)
seq_printf(m, "\t%d",
task_session_nr_ns(p, pid->numbers[g].ns));
#endif
seq_putc(m, '\n');
}
