C Programming

C: Nanosleep Function Usage

The function nanosleep is a method of the UNIX system. Nanosleep’s purpose is to suspend or pause the execution of a particular program for a defined period. This article will help you understand and give you an enhanced understanding of “How to utilize nanosleep() function.” One more function is “sleep”, but we will make use of nanosleep() because we can provide nanoseconds for pause/sleep time.


The above figure states the syntax of the nanosleep function and it is defined in the <time.h> header file.

RQTP: RQTP is a pointer to timespec which indicates the interval of time for which the user wants to suspend or pause the thread/program.

RMTP: RMTP is a pointer to timespec that indicates that the function has stored the period that remains in the interval.

The structure timespec is utilized to identify nanosecond-level time intervals.

Purpose of Using nanosleep() in C

Nanosleep is a Portable Operating System Interface. It is a system-compliant call to suspend a particular thread of a program’s execution for a specific amount of time. Similar functions are also available for the same purpose. Sleep is one of those processes that takes seconds to suspend the program, but it is said to deliver low-resolution suspension. Therefore, the nanosleep function grants permission to the user to provide the sleep time in nanoseconds for better precision.

Previously, the nanosleep() method was used to handle pauses of up to 2 MS when called from the scheduled threads, but it would require more precision to handle time-critical hardware or applications.

Return Value

  1. If the program has been executed successfully, it will return 0.
  2. If the program has been executed unsuccessfully or has failed and been interrupted, it will return -1.


  1. EFAULT: EFAULT type of error occurs if there is any problem in copying information from userspace.
  2. EINTR: EINTR type of error occurs when there is an interruption in the pause by a signal which was delivered to the thread.
  3. EINVAL: If the value of nanoseconds in the struct timespec is not in the range of 0 to 999999999 or has a negative value it will throw this error.

If the span determined in RQTP is anything other than a precise difference of the granularity hidden clock, it will be gathered. Besides, there might be a postponement later if the rest of the work is finished before the CPU is allowed to execute the calling string once more.

Since the method nanosleep does not function for a relative time span, it tends to be risky if the method is called repeatedly after facing hindrance or interruptions by signals, because the time between signal interruptions and the restart call will cause a slight shift when the sleep finishes. Use clock nanosleep (2) with an outright time value to stay away from this issue.

Nanosleep() should quantify time with the opponent’s REALTIME clock, as per POSIX.1. Linux, once again, utilizes the CLOCK MONOTONIC clock to monitor time. This is presumably immaterial because the POSIX.1 clock settime (2) particularly expresses that spasmodic changes in CLOCK REALTIME ought not to influence nanosleep().

If we set the value of the REALTIME clock via settime(2). This shall not have any effect on the programs which are blocked and waiting in queue for a relative time based on this clock.

Example in C

First of all, we had to initialize the <time.h> library which has the structure of a timespec request time pointer and a timespec remaining time pointer. There are two pointers that store the amount of time the user wants to suspend the program and also the remaining time that the stoppage interval time has left.

After that, we start our main body, and we have to create two timespec objects that will contain our request and remaining time. We could assign any value to these two objects, but in our case, we have chosen 3 seconds and 500 nanoseconds.

Now, we will pass the addresses of the created objects to nanosleep, as you can observe in line number 10. We will also check if the program was successful or failed by observing the return value of the method nanosleep.

The above program will print the following output if it is executed successfully:

If we change the response value to 1 the program execution will fail and produce the following error as output.

Now, if we want to run the following code on our GCC terminal. We will first save our file as main.c and then use the following command on your terminal to run the program: “gcc-Wall main.c-o”. A wall means to enable all the warning messages while executing our program.


The current execution of nanosleep() depends on the typical bit clock component, which has a goal of 1/HZ s. Along these lines, nanosleep() stops consistently for the predefined time, but it can take up to 10 ms longer than indicated until the interaction becomes runnable once more. For a similar explanation, the value returned in the event of a conveyed signal in *rmtp and is normally adjusted to the following larger difference of 1/HZ s.


It is normal to suspend the execution of a string for some time to survey the situation with an eye towards non-intruding work. Countless real necessities can be met with a straightforward extension to sleep() that gives a better goal.

In the POSIX.1-1990 norm and SVR4, it is feasible to execute such an everyday practice, except the recurrence of wakeup is restricted by the goal of the alarm() and sleep() functions. It is likely to write such a standard in 4.3 BSD while utilizing no static stockpiling and sparing no framework offices. Even though it is feasible to compose a function with comparable usefulness to sleep() utilizing the rest of the timer_* () capacities, such a capacity requires the utilization of signs and the booking of some significant number. This volume of IEEE Std 1003.1-2001 needs that nanosleep() be non-meddlesome if the signs work.

The nanosleep() work will return a value of 0 on progress and – 1 on failure, or again whenever interfered with. This last option case is not quite the same as sleep(). This was done in light of the fact that the leftover time is returned utilizing a contention structure pointer, RMTP, rather than as a way to bring approval back.


The focus of this research was to help you develop a better grasp of the method nanosleep(). To have a good grip of methods like nanosleep it is necessary to illustrate them with the simplest example. We have tried our best to provide the best of information such as bugs, rationale, examples, errors, and synopsis. So that you can continue improving the interpretability and reusability of your code. We have gone over a simple syntax explanation. The article will assist you in quickly gaining a thorough interpretation of how to use nanosleep () as a method. To make much better use of the method, considerations as variables have been addressed and well explained for the users.

About the author

Aqsa Yasin

I am a self-motivated information technology professional with a passion for writing. I am a technical writer and love to write for all Linux flavors and Windows.