Why some system calls shield SIGPROF signal under strace - c

When I strace this code
void printMsg();
int main() {
signal(SIGPROF, printMsg);
struct itimerval tick;
memset(&tick, 0, sizeof(tick));
tick.it_value.tv_sec = 1; // sec
tick.it_value.tv_usec = 0; // micro sec.
tick.it_interval.tv_sec = 0;
tick.it_interval.tv_usec = 0;
setitimer(ITIMER_PROF, &tick, NULL);
while(1) {
;
}
return 0;
}
void printMsg() {
printf("%s","Hello World!!\n");
}
I got the SIGPROF signal after 1 second as expected
...
05:54:10 setitimer(ITIMER_PROF, {it_interval={0, 0}, it_value={1, 0}}, NULL) = 0
05:54:11 --- SIGPROF {si_signo=SIGPROF, si_code=SI_KERNEL} ---
...
But when I add a system call like write(2, "", 0) or read(2, "", 0) in while(1) and strace again, it looks like SIGPROF can not be fired. However, time(0) in while(1) can trigger SIGPROF properly.
Btw, I use this code to emulate following PHP script, which ignored time limit under PHP-FPM SAPI,
<?php
set_time_limit(5); // PHP uses setitimer(ITIMER_PROF) to implement this function
while (true) {
flush(); // PHP uses write(<fd>, "", 0) to implement this function
}

The code trigger the SIGPROF using ITIMER_PROF, which is counting when the process is executing, or when the system is running on behalf of the process.
When the process goes into 'read' or 'write' calls, it is not 'running'. The process goes into IO Wait (and the system will allocate the CPU to other processes). When the process is in IOWait, the timer is not moving. When there is a system call that perform processing (mostly in system space) - e.g. time() is a tight loop, the timer will be moving.
See more about process states in https://www.tecmint.com/linux-process-management/
If you want to measure 'clock' time, consider using timer_create with CLOCK_REALTIME or CLOCK_MONOTONIC

Related

to interrupt a sleeping thread when main thread is to exit

I have a main thread which create child threads to do various task. There is a child thread which is tasked to report on the status every 100s
My current mechanism of stopping the thread is to observe a global boolean. Somewhat like this
Child thread
void* ReportThread(bool* operation)
{
while(*operation)
{
// do its reporting task
// ........
int counter = 0;
while( counter < 100 && operation )
{
// let it sleep for 1 seconds and wake up to check
sleep(1);
sleepCounter += 1;
}
}
}
Parent (Main) Thread:
bool operation = false;
int main(){
pthread_t tid;
err = pthread_create(&tid), NULL, &ReportThread, &operation);
printf("Please input esc to end operation \n");
while ((ch = getchar()) != 27);
operation =true;
pthread_join(tid,NULL);
return 0;
}
The problem:
It seem that using sleep(n). The number of seconds seem very inconsistent. When the program is stopped, this thread takes a while maybe 10 second to actually stop
Is there a way to interrupt a thread to sleep? I heard you could use signal. I am coding in linux
Can I just simply just use a pthread_cancel(tid) instead of pthread_join(tid)?
Regards
This part
while( counter < 100 || operation )
{
// let it sleep for 1 seconds and wake up to check
sleep(1);
sleepCounter += 1;
}
is wrong.
First I assume that sleepCounter += 1; is really a typo and that it should be:
while( counter < 100 || operation )
{
// let it sleep for 1 seconds and wake up to check
sleep(1);
counter += 1;
}
Then the problem is that even if operation is set to false by some other thread, the while will not finish until counter reach 100.
The code should be
while( counter < 100 && operation )
{
// let it sleep for 1 seconds and wake up to check
sleep(1);
counter += 1;
}
Further, in main you never set operation to false. Another typo?
You don't need two while loops. And if you want to set a timer, use time functions for it, because sleep is a cancellation point and it is not guaranteed that sleep actually sleeps that amount of time.
Example:
void* ReportThread(void *args)
{
time_t start = time(NULL);
time_t now;
bool *operation = (bool*) args;
while (*operation) { //while active
now = time(NULL); //get current time
if (now - start >= 100) { //if the threshold is exceeded
start = now; //reset timer
//and probably do other stuff
}
sleep(1); //sleep for one second
}
return NULL;
}
The example above has a max lag of one second, that means if you set operation to false right at that moment when the thread entered the sleep state, you have to wait until sleep returns, only then it will recognize the modified state. The example also has the advantage, that you can easily modify the threshold value (since it depends on the 'real' time, instead of a counter and a non accurate sleep time).
Btw. the variable operation should be either an atomic boolean or protected by a mutex (since it is accessed from different threads).
To answer the questions of your problem:
should be answered by the example above
since i mentioned it before, sleep is a cancellation point, that means it gets interrupted if the process handles a signal (see man pthreads - section Cancellation points).
see man pthread_cancel - section Notes
On Linux, cancellation is implemented using signals. Under the NPTL threading implementation, the first real-time signal (i.e., signal 32) is used for this purpose. On LinuxThreads, the second real-time signal is used, if real-time signals are available, otherwise SIGUSR2 is used.
You cannot use pthread_cancel over pthread_join! You have to use pthread_join in either case (described in detail in the man page).
I don't know if this will fix all your problems, but it's a bit too much for a comment. One problem, your ReportThread function signature is wrong. It should be:
void* ReportThread(void* args);
And then in that function you need to do something like:
void* ReportThread(void* args)
{
bool* operation = (bool*)args;
while(*operation)
{
...
}
}
I'm not sure how it's working right now, but your compiler should at least be issuing a warning trying to convert a bool* type to a bool.
Also be aware of race conditions on operation

socket communication time cost too long

I'm running a piece of program on a Linux development board, and I found that when the CPU load is low, it works alright, but if the CPU load gets to a high peak, it will take a much longer time.
Here is how it looks like: There are 2 programs running on board. Comsumer application has multiple threads, and they will call func1 to request some info from the producer process. The producer is a daemon process that will feed info back to the Consumer process.
The sample code looks like this :
Comsumer:
static int send_to_service(msg_t *cmd)
{
int cnt = send(fd, cmd, sizeof(msg_t), MSG_WAITALL);
if (cnt != sizeof(msg_t)) {
log(L_ERROR, "send failed");
return -1;
}
return 0;
}
static int func1(int aaa)
{
struct_t msg = {...};
msg.a = ...;
msg.b = ...;
gettimeofday(time1, NULL);
send_to_service(&msg);
...
}
Producer:
while(1) {
gettimeofday(time2, NULL);
int ret = select(maxfd+1, &fd_list, NULL, NULL, NULL);
gettimeofday(time3, NULL);
for (int fd = 0; fd <= maxfd; fd ++) {
if (!FD_ISSET(fd, &fd_list))
continue;
...
}
}
The time difference between time1 and time3 can be 30ms+ more during CPU high load time. And it does not happen all the time, only once in a while.
I tried the single process way earlier, that way the Consumer calls the driver and get the info directly. That worked well. Now I have to add another process to the system to get the same info, so I have to use a daemon process to feed to both process. The performance is not as good as the single process method. even if there is only one consumer.
The system I use is Linux version 4.14.74, I'm not sure about the socket type and network, both consumer processes are within the same system waiting to get image info. I just used "send, recv and select" system provided.

Catching SIGTERM, and sleep preventing it working

I have some code written in C (working on ubuntu 17):
void sig_stop(int sig_num) {
/* Some cleanup that needs to be done */
}
void some_routine(const char *array[], const int length) {
/* Initialization */
signal(SIGTERM, sig_stop);
while (true) {
/* Some function */
/* I have this sleep to minimize the load on the CPU
as I don't need to check the conditions here
all the time. */
sleep(5);
}
}
Whenever I include the 5 minute sleep (sleep(5)), it appears sig_stop isn't called. However, when I comment out the sleep(5), the sig_stop cleanup works just fine. Have I got something wrong with my understanding of how to catch SIGTERM?
If I can't use the sleep function, is there a better way to "sleep" the program" so that it only runs the loop every x minutes or in such a way that minimizes the CPU load?
sleep() and signals
sleep() should not prevent the signal from being caught and the signal handler being executed. From the manpage for sleep() (emphasis mine):
sleep() causes the calling thread to sleep either until the number of real-time seconds specified in seconds have elapsed or until a signal arrives which is not ignored.
Take the following example ...
#include <signal.h>
#include <stdio.h>
#include <time.h>
#include <unistd.h>
static volatile sig_atomic_t flag = 0;
static void sig_stop(int signum) { flag = 1; }
int main(void) {
int secs_remaining = 0;
signal(SIGTERM, sig_stop);
while (!flag) {
printf("Sleeping at time %d\n", time(NULL));
secs_remaining = sleep(5);
}
printf(
"Flag raised. Exiting at time %d. sleep() was interrupted %d seconds "
"early ...\n",
time(NULL), secs_remaining);
return 0;
}
Note that - in the case where it was interrupted by a signal - sleep() returns the number of seconds left to sleep. E.g., if it is interrupted 3 seconds early it will return 3. It will return 0 if it is not interrupted.
Compile as gcc -o test test.c and run. Then from another terminal run
pkill -15 test
You will see output similar to the following ...
Sleeping at time 1532273709
Flag raised. Exiting at time 1532273711. sleep() was interrupted 2 seconds early ...
By the way ... sleep(x) sleeps for x seconds - not minutes.
signal() vs sigaction()
Due to portability issues associated with signal(), it is often recommended to use sigaction() instead. The use of sigaction() would be something like the following.
int main(void) {
struct sigaction sa;
sa.sa_flags = 0;
sigemptyset(&sa.sa_mask);
sa.sa_handler = sig_stop;
if (sigaction(SIGTERM, &sa, NULL) == -1) {
perror("sigaction");
return 1;
}
// Etc.
}
As you can see the usage of sigaction() is a little more verbose than that of signal(). Perhaps that's why people still sometimes use signal().

timer_settime calling handler function in pthread on uClinux

I've got the following function that gets called from a pthread_create. This function does some work, sets a timer, does some other work and then waits for the timer to expire before doing the loop again. However, on the first run of the timer, after it expires the program quits and I'm not totally sure why. It should never leave the infinite while loop. The main thread accesses nothing from this thread and vice versa (for now).
My guess is I might not have something setup correctly with the thread, or the timer is not calling the handler function correctly. Perhaps changing the IDLE global variable from the thread causes a problem.
I would like to call the handler without signals, hence the use of SIGEV_THREAD_ID. I'm using the SIGUSRx signals in the main thread anyway. Any thoughts about what I've started here what could be wrong?
#ifndef sigev_notify_thread_id
#define sigev_notify_thread_id _sigev_un._tid
#endif
volatile sig_atomic_t IDLE = 0;
timer_t timer_id;
struct sigevent sev;
void handler() {
printf("Timer expired.\n");
IDLE = 0;
}
void *thread_worker() {
struct itimerspec ts;
/* setup the handler for timer event */
memset (&sev, 0, sizeof(struct sigevent));
sev.sigev_notify = SIGEV_THREAD_ID;
sev.sigev_value.sival_ptr = NULL;
sev.sigev_notify_function = handler;
sev.sigev_notify_attributes = NULL;
sev.sigev_signo = SIGRTMIN + 1;
sev.sigev_notify_thread_id = syscall(SYS_gettid);
/* setup "idle" timer */
ts.it_value.tv_sec = 55;
ts.it_value.tv_nsec = 0;
ts.it_interval.tv_sec = 0;
ts.it_interval.tv_nsec = 0;
if (timer_create(0, &sev, &timer_id) == -1) {
printf("timer_create failed: %d: %s\n", errno, strerror(errno));
exit(3);
}
while (1) {
// do work here before timer gets started that takes 5 seconds
while (IDLE); /* wait here until timer_id expires */
/* setup timer */
if (timer_settime(timer_id, 0, &ts, NULL) == -1) {
printf("timer_settime failed: %d\n", errno);
exit(3);
}
IDLE = 1;
// do work here while timer is running but that does not take 10 seconds
}
}
As far as I can tell, you haven't installed a signal handler for SIGUSR1, so by the default action it kills the process when it's acted upon.
In any case, the whole thing strikes me as extraordinarily bad design:
The while loop will give you 100% cpu load while waiting for the timer to expire.
This is not the way you use SIGEV_THREAD_ID, and in fact SIGEV_THREAD_ID isn't really setup to be usable by applications. Rather it's for the libc to use internally for implementing SIGEV_THREAD.
You really don't want to be using signals. They're messy.
If you have threads, why aren't you just calling clock_nanosleep in a loop? Timers are mainly useful when you can't do this, e.g. when you can't use threads.

Trouble with Interval Timer

Our current project is based on extending more by including scroll. In order to do this, a timer interval has to be set for a certain period. The part I'm not sure about is where the loop for the alarm signal should be. All the examples I've seen have the timer values in the main, then explicitly call the signal handler via pause() in an infinite while loop.
My code is a bit different, since the functionality requirements go like
print first screen of text after getting terminal dimensions
print prompt
if prompt = space, print another screen of text //WORKS
if prompe = q, restore original terminal settings & quit program //WORKS
if prompt = ENTER, initialize scroll at 1 line every 2 seconds //DOESN'T WORK
if prompt == f/s, increase/decrease scroll speed by 20% //DOESN'T WORK
The read in buffer, file pointer and itimerval struct are all global variables to avoid passing as arguments through a chain of functions.
The main function of the program is
void processInput(FILE *fp){
void printLine(int); //prints a single line of text
signal(SIGPROF, printLine);
int c;
//print first screen of text, check for more text to display
info(); //print prompt at bottom of screen
FILE *fterm= fopen("/dev/tty", "r");
while ((c=getc(fterm)) != EOF){
if (c== '\n'){
setTimer(2);
//four more conditionals like this in basic similarity
}
}
My setTimer function has a base interval of 2 seconds, and changes that by plus/minus 20% based on f/s input from the user.
void setTimer(int direction){
int speed=2000000; //2 seconds
int change= 400000; //400 milliseconds, 20% of 2 seconds
if (direction == 1) //slow down by 20%
speed+= change;
if (direction == 0)
speed -= change;
timer.it_value.tv_sec=2;
timer.it_value.tv_usec=0;
timer.it_interval.tv_sec=0;
timer.it_interval.tv_usec= speed;
setitimer(ITIMER_PROF, &timer, NULL);
}
First question: should I use SIGALRM vs SIGPROF, and alter the ITIMER_XXXX variable accordingly?
Second, where should I put in the loop to trigger the signal? I tried
while(1)
pause();
in several of the conditionals, but it had the effect of stopping the execution and ignoring any input.
Without knowing the details of your requirements, couldn't you do this more easily using
select()?
Set your initial select timeout to 2 seconds and adjust according to f/s input, meanwhile if there is any standard input before the timeout you process it.
More or less valid general outline:
int retval;
fd_set rfds;
int input = fileno(fterm);
struct timeval tv, delay;
delay.tv_sec = 2;
delay.tv_usec = 0;
while (true)
{
FD_ZERO(&rfds);
FD_SET(input, &rfds);
tv.tv_sec = delay.tv_sec;
tv.tv_usec = delay.tv_usec;
retval = select(input + 1, &rfds, NULL, NULL, &tv);
if (retval == -1)
perror("select()");
else
if (retval)
{
if (FD_ISSET(input, &rfds))
{
command = readInput(...);
switch(command)
{
case 'q' ... cleanup & quit
case 's' ... calc 20% off delay values
case etc ...
case default...error handling
}
}
}
else //timeout
printLine();
}
Working with pause() is dangerous because it's not an atomic operation ... your program could be interrupted by the OS causing you to "lose" the arrival of a signal. Additionally, when pause() itself returns because of the arrival of a signal, it will simply call pause() again. That means you're going to have to-do all your work inside of a signal handler, which may not be the best thing, i.e., if you're inside the signal handler when the next signal goes off, you can end up with some unpredictable behavior if you haven't planned for that sort of event.
A better approach would be to-do the following:
1) Setup a signal mask that blocks SIGPROF at the start of your program.
2) Rather than using a signal handler to-do your heavy lifting, use sigwait(), and set it up with a sigset_t that contains a mask for SIGPROF.
3) Setup the main flow of your program the following way:
sigset_t sigset;
sigemptyset(&sigset);
sigaddset(&sigset, SIGPROF);
sigprocmask(SIG_BLOCK, &sigset, NULL); //block the SIGPROF signal
//process your input
//if you need to, initialize your timer and set it to go off
while(SOME_FLAG_IS_TRUE) //maybe this loops forever ... depends on what you want?
{
sigwait(&sigset, &signal_num);
if (signal_num != SIGPROF)
continue;
//process input ...
//... setup new interval timer with correct timeout ...
//... repeat loop and/or exit loop or set flag to exit loop
}
That should always catch the signal from the interval timer since sigwait() will properly return after waiting for a signal to arrive to your process, and the SIGPROF signal is always blocked, meaning you can't "lose" signals ... instead at least one of them will be queued up and waiting for the next call to sigwait() to be detected just in case one arrives while you're processing something in your while-loop.
Hope this helps,
Jason

Resources