Deadlock with sigsuspend() - c++

I am trying to synchronize a father and children, the following code is not working (apparently usr_interrupt++ is not atomic). Semaphores does not seems to help either.
#include <sys/types.h>
#include <sys/ipc.h>
#include <sys/shm.h>
#include <sys/stat.h>
#include <unistd.h>
#include <cstdlib>
#include <iostream>
#include <unistd.h>
#include <cstring>
#include <string>
#include <semaphore.h>
#include <fcntl.h>
using namespace std;
/* When a SIGUSR1 signal arrives, set this variable. */
volatile sig_atomic_t usr_interrupt;
sem_t *mutex;
char* SEM_NAME;
void
synch_signal (int sig)
{
// sem_wait(mutex);
usr_interrupt++;
// sem_post(mutex);
}
/* The child process executes this function. */
void
child_function (void)
{
/* Perform initialization. */
cerr << "I'm here!!! My pid is " << (int)getpid() << " my usr_int=" << usr_interrupt << endl;
/* Let parent know you're done. */
kill (getppid (), SIGUSR1);
/* Continue with execution. */
cerr << "Bye, now...." << endl;
exit(0);
}
int
main (void)
{
usr_interrupt = 0;
string s_sem_name = "lir";
SEM_NAME = new char[s_sem_name.size()+1];
memcpy(SEM_NAME, s_sem_name.c_str(), s_sem_name.size());
SEM_NAME[s_sem_name.size()] = '\0';
mutex = sem_open (SEM_NAME,O_CREAT,0644,1);
if(mutex == SEM_FAILED) {
perror("unable to create semaphore");
sem_unlink(SEM_NAME);
exit(-1);
}
struct sigaction usr_action;
sigset_t mask, oldmask;
pid_t child_id, child_id2;
/* Set up the mask of signals to temporarily block. */
sigemptyset (&mask);
sigaddset (&mask, SIGUSR1);
/* Establish the signal handler.*/
usr_action.sa_handler = synch_signal;
usr_action.sa_flags = 0;
sigaction (SIGUSR1, &usr_action, NULL);
/* Create the 2 children processes. */
child_id = fork ();
if (child_id == 0)
child_function ();
child_id2 = fork();
if (child_id2 == 0)
child_function ();
/* Wait for a signal to arrive. */
sigprocmask (SIG_BLOCK, &mask, &oldmask);
while (usr_interrupt != 2) {
sigsuspend (&oldmask);
}
sigprocmask (SIG_UNBLOCK, &mask, NULL);
/* Now continue execution. */
puts ("That's all, folks!");
return 0;
}
Can anyone suggest a fix? (I cannot use threads)
Best,
-- Liron

You can't count signals. Two signals of the same type has the same semantic meaning as one signal of that type. You could use two different signal types like USR1 and USR2. But honestly, you shouldn't use signals as a communication mechanism. Use something sensible like a pipe.

Related

Valgrind with signal handling

In this code, I am trying to ipc with signals. Parent process will send SIGUSR1 signal to his children and waits for a SIGUSR1 signal for conformation of the signal arrive at the child end.
So I have;
Parent --> SIGUSR1 --> CHILD[i]
Parent <--- SIGUSR1 <--- CHILD[i]
this kind of communication between parent child processes. In normal executing, I have never face with a deadlock scenario. But when I run with valgrind, sometimes child process canot receive SIGUSR1 signal from parent process, which is start signal for communication.
Is the valgrind uses some mechanisms to avoid signal communication between processes ?
Parent process :
#include <sys/types.h>
#include <sys/wait.h>
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <string.h>
#include <signal.h>
#include <fcntl.h>
#include <errno.h>
#define CHILD_SIZE 80
sigset_t sigusr_mask;
pid_t children[CHILD_SIZE];
void SIGUSR1_handler(int sig)
{
write(1,"test.\n",6);
}
void init_signal()
{
sigset_t blck_msk;
sigemptyset(&blck_msk);
sigaddset(&blck_msk, SIGUSR1);
sigfillset(&sigusr_mask);
sigdelset(&sigusr_mask, SIGUSR1);
struct sigaction usr1_act;
memset(&usr1_act, 0, sizeof(usr1_act));
usr1_act.sa_handler = &SIGUSR1_handler;
sigaction(SIGUSR1, &usr1_act, NULL);
if(-1== sigprocmask(SIG_BLOCK, &blck_msk,NULL))
exit(-1);
}
void create_children(const char* processPath, int round_)
{
for(int i = 0; i < CHILD_SIZE; ++i)
{
pid_t pid_c = fork();
if(pid_c > 0) {
children[i] = pid_c;
}
else if(pid_c == 0)
{
char* arg[1024] = {"childProcess", (char*)0};
execve(processPath, arg, NULL);
}
}
}
// for the synchronization
void wait_childs()
{
for (int i = 0; i < CHILD_SIZE; ++i)
{
// send SIGUSR1 to the children.
kill(children[i], SIGUSR1);
printf("%d : %s\n", children[i],strerror(errno));
// wait until children reply by SIGUSR1
sigsuspend(&sigusr_mask);
}
}
int main(int argc, char const *argv[])
{
init_signal();
create_children("childProcess", 1);
wait_childs();
return (0);
}
Child process :
#include <sys/types.h>
#include <sys/wait.h>
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <string.h>
#include <signal.h>
#include <errno.h>
#include "utils.h"
sigset_t sigusr_mask;
void SIGUSR1_handler(int sig) {
write(1,"usr1\n",4);
}
void init_signal()
{
sigfillset(&sigusr_mask);
sigdelset(&sigusr_mask, SIGUSR1);
struct sigaction usr1_act;
memset(&usr1_act, 0, sizeof(usr1_act));
usr1_act.sa_handler = &SIGUSR1_handler;
sigaction(SIGUSR1, &usr1_act, NULL);
}
int main(int argc, char const *argv[])
{
init_signal();
printf("child %d waiting \n",getpid());
// wait until parent starts the communication.
sigsuspend(&sigusr_mask);
// after parent sends SIGUSR1, send reply.
kill(getppid(), SIGUSR1);
return 0;
}

Does daemon() have any disadvantages?

Does using the C function daemon() have any security or stability disadvantages for a linux daemon compared to using explicit functions like fork(), setsid(), umask(), etc. (beside being unable to set all daemon parameters)?
I was wondering why I should write
#include <sys/types.h>
#include <sys/stat.h>
#include <cstdio>
#include <cstdlib>
#include <fcntl.h>
#include <cerrno>
#include <unistd.h>
#include <syslog.h>
#include <string>
int main()
{
//Set our Logging Mask and open the Log
setlogmask(LOG_UPTO(LOG_NOTICE));
openlog(DAEMON_NAME, LOG_CONS | LOG_NDELAY | LOG_PERROR | LOG_PID, LOG_USER);
syslog(LOG_INFO, "Entering Daemon");
pid_t pid, sid;
//Fork the Parent Process
pid = fork();
if (pid < 0)
exit(EXIT_FAILURE);
//We got a good pid, Close the Parent Process
if (pid > 0)
exit(EXIT_SUCCESS);
//Change File Mask
umask(0);
//Create a new Signature Id for our child
sid = setsid();
if (sid < 0)
exit(EXIT_FAILURE);
//Change Directory
//If we cant find the directory we exit with failure.
if ((chdir("/")) < 0)
exit(EXIT_FAILURE);
//Close Standard File Descriptors
close(STDIN_FILENO);
close(STDOUT_FILENO);
close(STDERR_FILENO);
while (true)
{
sleep(5);
//Do something
}
closelog ();
}
instead of
#include <unistd.h>
int main()
{
daemon(0, 0);
while (true)
{
//Do something
sleep(5);
}
}
Per the manpage, it's not in POSIX so you're always taking a risk regarding its existence.
Otherwise, no.

setitimer signalling seems only to work after fork

I would like to use the following C++ code to wait for a predefined amount of time (in this example always 2 seconds), but still be interruptible by a signal (that's why I don't use sleep):
#include <unistd.h>
#include <stdlib.h>
#include <sys/wait.h>
#include <sys/types.h>
#include <sys/time.h>
#include <signal.h>
#include <iostream>
using namespace std;
int measure() {
itimerval idle;
sigset_t sigset;
int sig;
idle.it_value.tv_sec = 2;
idle.it_value.tv_usec = 0;
setitimer(ITIMER_REAL, &idle, NULL); // TODO: check return value
sigemptyset(&sigset);
sigaddset(&sigset, SIGALRM); // TODO return values
sigaddset(&sigset, SIGUSR1);
sigprocmask(SIG_BLOCK, &sigset, NULL); // TODO return value?
sigwait(&sigset, &sig); // TODO check return value
while(sig != SIGUSR1) {
cout << "Hohoho" << endl;
idle.it_value.tv_sec = 2;
idle.it_value.tv_usec = 0;
setitimer(ITIMER_REAL, &idle, NULL); // TODO: check return value
sigwait(&sigset, &sig); // TODO check return value
}
cout << "Done with measurements." << endl;
return 0;
}
int main(int argc, char **argv) {
//if(fork() != 0) exit(0);
//if(fork() == 0) exit(0);
return measure();
}
I would expect this code to print "Hohoho" every 2 seconds until it receives SIGUSR1. Then it prints "Done with measurements." and exits. The second part works as expected. However, I see no "Hohoho", so it seems to me that the SIGALRM from setitimer somehow is not received. The strange thing is that if I do a fork before, the program works as expected. More specifically, if I uncomment either one of the two fork commands at the end, it works. Hence it does not depend on if it's the parent or child process, but somehow the fork event matters. Can someone explain to me what's going on and how to fix my code?
Thanks a lot,
Lutz
(1) Your setitimer is failing because you haven't set it correctly. Struct itimerval contains two structs of type timeval. You are only setting one and thereby picking up whatever garbage was in local storage when idle was declared.
struct itimerval {
struct timeval it_interval; /* next value */
struct timeval it_value; /* current value */
};
struct timeval {
time_t tv_sec; /* seconds */
suseconds_t tv_usec; /* microseconds */
};
If you want a repeating timer every 2 seconds then set the 2nd set to repeat with the same values.
idle.it_value.tv_sec = 2;
idle.it_value.tv_usec = 0;
idle.it_interval.tv_sec = 2;
idle.it_interval.tv_usec = 0;

Trying to prevent race conditions with pselect(), but signal won't interupt

I am currently trying to implement a server in C++ using sockets. I am trying to prevent race conditions by blocking the SIGINT signal until it is stuck in the blocking pselect. From there, it should be exiting, changing my loop variable, and then quitting the thread. From my attempts at getting this working, it appears that it reaches the pselect(), but it does not get interrupted using my code. Any help is appreciated.
Listener.h:
#include <sys/socket.h>
#include <netinet/in.h>
#include <arpa/inet.h>
class CListener
{
public:
CListener();
void quitListener(void);
private:
void* InitListener(void);
static void* StartListenerThread(void* context);
static bool mbListening;
pthread_t mtThreadID;
};
Listener.cpp
#include <iostream>
#include <unistd.h>
#include <string.h>
#include <sys/types.h>
#include <pthread.h>
#include <signal.h>
#include <errno.h>
#include "Listener.h"
bool CListener::mbListening = true;
CListener::CListener()
{
mbListening = true;
mtThreadID = 0;
pthread_create(&mtThreadID, NULL, &CListener::StartListenerThread, this);
}
void* CListener::StartListenerThread(void* context)
{
return ((CListener*)context)->InitListener();
}
void* CListener::InitListener()
{
sigset_t tSignalSet;
sigset_t tOriginalSignalSet;
sigemptyset(&tSignalSet);
sigaddset(&tSignalSet, SIGINT);
sigprocmask(SIG_BLOCK, &tSignalSet, &tOriginalSignalSet);
FD_ZERO(&tConnectionSet);
FD_SET(0, &tConnectionSet);
while(mbListening)
{
tSelectSet = tConnectionSet;
std::cout << "Reached pselect\n";
nReadyConnections = pselect(nSelectSocket+1, &tSelectSet,
NULL, NULL, NULL, &tOriginalSignalSet);
std::cout << "Broke out of pselect\n";
if(nReadyConnections < 0 && errno == EINTR)
{
mbListening = false;
}
}
pthread_exit(NULL);
return NULL;
}
void CListener::quitListener()
{
raise(SIGINT);
}
As long as I copied everything correctly fingers crossed you should just be able to run:
CListener tListener = CListener();
usleep(20000);
tListener.quitListener();
and the outputs should be displayed in terminal. My end goal is that I can allow for pselect to be interrupted without breaking any processing that may come after and allowing the thread to close gracefully. (blocking at pselect > recieve SIGINT > interupt pselect > return to loop > finish up and exit)
I've solved my own problem. Since I am generating a thread, I needed to add functions that allow for that, as well as adding a signal handler, like shown below.
void CListener::quitListener()
{
pthread_kill(mtThreadID,SIGINT);
}
void CListener::installSIGINTHandler()
{
signal(SIGINT, CListener::SIGINTHandler);
}
void CListener::SIGINTHandler(int signo)
{
mbListening = false;
}
And needed to change the sig mask setup to this:
pthread_sigmask(SIG_BLOCK, &tSignalSet, &tOriginalSignalSet);

UNIX/OSX version of semtimedop

GLibC has a method semtimedop which allows you to perform an operation (a semaphore acquire in this case) which times out after a certain amount of time. Win32 also provides WaitForSingleObject which provides similar functionalty.
As far as I can see there is no equivalent on OSX or other Unices. Can you suggest either the equivalent for semtimedop or a workaround to terminate a semop after a certain amount of time cleanly.
You can break out of a semop() call (and most other blocking calls) by getting a signal, such as one caused by alarm().
untested example:
#include <signal.h>
#include <stdio.h>
#include <sys/types.h>
#include <sys/ipc.h>
#include <sys/sem.h>
volatile int alarm_triggered = 0;
void alarm_handler(int sig)
{
alarm_triggered = 1;
}
int main(int argc, char **argv)
{
int rc;
/* set up signal handler */
signal(SIGALRM, alarm_handler);
/* ... */
alarm(30); /* 30 second timeout */
rc = semop(...);
if (rc == -1 && errno == EINTR)
{
if (alarm_triggered)
{
/* timed out! */
}
}
alarm(0); /* disable alarm */
/* ... */
}