nanosleep : High Resolution Sleeping

A process that voluntarily suspends the execution for a specified time is said to sleep. Whenever a thread or process sleeps, it will allow other thread or process to utilize CPU time and do useful operation.

Unix/Linux system provides mechanisms where a process or a thread can suspend the execution for the specified time and resumes executions after the time has expired.


The sleep function causes the calling thread to be suspended either until the specified number of seconds has elapsed or until the calling thread catches the signal.

#include <unistd.h>

int sleep(unsigned seconds)

The sleep function returns 0 if the requested time has elapsed or the amount of unslept time if interrupted by signal or a high priority interrupt. The disadvantage of sleep function is, it expects arguments only in seconds. Since the most of the embedded systems requires timing in terms of microseconds or nanoseconds.


The usleep function causes the calling thread to be suspended either the specified number of microsecond has elapsed or until the calling thread catches the signal. The sleep may be lengthened slightly by any system activity or by the time spent processing the call or by the granularity of system timers.

#include <unistd.h>

int usleep (useconds_t usec)

the usleep function returns 0 on sucess and -1 on error.



Interrupted by a signal;


usec is not smaller than 1000000. (On systems where that is considered an error.)

The disadvantage of usleep() function is, if  the function is interrupted by a signal, it will not return for how much time this function has slept and how much time is remaining.


nanosleep() suspends the execution of the calling thread until either at least the time specified in *req has elapsed, or the delivery of a signal that triggers the invocation of a handler in the calling thread or that terminates the process.

If the call is interrupted by a signal handler, nanosleep() returns -1, sets errno to EINTR, and writes the remaining time into the structure pointed to by rem unless rem is NULL. The value of *rem can then be used to call nanosleep() again and complete the specified
pause (but see NOTES).

#include <time.h>

int nanosleep(const struct timespec *req, struct timespec *rem);

The structure time-spec is used to specify intervals of time with nanosecond precision. It is defined as follows:

struct timespec {
          time_t tv_sec; /* seconds */
          long tv_nsec; /* nanoseconds */

The value of the nanoseconds field must be in the range 0 to 999999999.

Compared to sleep() and usleep(), nanosleep() has the following advantages: it provides a higher resolution for specifying the sleep interval; POSIX.1 explicitly specifies that it does not interact with signals; and it makes the task of resuming a sleep that has been interrupted by a signal handler easier.

On successfully sleeping for the requested interval, nanosleep() returns 0. If the call is interrupted by a signal handler or encounters an error, then it returns -1, with errno set to indicate the error.

EFAULT Problem with copying information from user space.

EINTR The pause has been interrupted by a signal that was delivered to the thread (see signal(7)). The remaining sleep time has been written into *rem so that the thread can easily call nanosleep() again and continue with the pause.

EINVAL The value in the tv_nsec field was not in the range 0 to 999999999 or tv_sec was negative.

In embedded system application, a wrapper function can be implemented which will take the argument in terms of second, microsecond or nanosecond, and nanosleep() and it’s functionality will be wrapped inside the function. 

#include <stdio.h>

#include <time.h>

enum SleepType = {Second = 0, miliSecond, microSecond, nanoSeconds}

void MyWrappedSleepFunction(SleepType  sleepType, long sleepTime)


     struct timespec tv;



           case Second :


  tv.tv_sec = sleepTime ;

                tv.tv_nsec  = 0;



           case miliSecond:


  tv.tv_sec = 0;

                tv.tv_nsec  = sleepTime * 1000 * 1000;



           case microSecond:


  tv.tv_sec = 0;

                tv.tv_nsec  = sleepTime * 1000;



           case nanoSeconds:


  tv.tv_sec = 0;

                tv.tv_nsec  = sleepTime;




   while ((nanosleep(&tv, &tv) == -1) && (errno == EINTR))


         /* wait here till the specified time gets over */



int main()


MyWrappedSleepFunction(Second , 2);

MyWrappedSleepFunction(miliSecond , 30);

MyWrappedSleepFunction(microSecond , 10);

MyWrappedSleepFunction(nanoSeconds , 5);

return 0;


I hope, you would find this post useful and hope you will try to implement this wrapper function in your application.

Please share your thoughts and suggestions in the comments and stay tuned for the next post. Till then enjoy clicking embedded.


Update U-Boot Environment Variables stored in Flash from Linux

Hi folks,

How are you all?

Many of us wondering that, Is it possible to get U-boot environment variables in user space? Is it possible to set U-boot environment variable from user space? Answer to these questions is YES. It is possible to set and retrieve U-boot environment variables from user space through command-line or any application.

In this article I will demonstrate how to set and get U-boot environment from Linux user space through an application or shell command. To explain this I will take a reference of AM335x EVM.

In Linux source code, there is a board file specific to board or processor (in my case it is board-am335xevm.c) which contains a structure of memory partitions, In that memory partions will be a separate memory partition of Uboot environment variables where these variables will be stored. refer the below image.


Build fw_printenv / fw_setenv tool

  • first of all cd to top of the u-boot source directory
  • Issue a command
    host# make HOSTCC=arm-arago-linux-gnueabi-gcc env

    This must create a fw_printenv and fw_setenv cross compiled binaries in <uboot-src>/tools/env directory. You must see these binaries along with fw_env.config file. You must copy these two binaries to /bin folder and fw_env.config file to /etc directory of the target system.

Test fw_printenv / fw_setenv tool

The executable fw_printenv will dump the entire environment space, individual variables can be read by:

target# ./fw_printenv autoload

To set an environment variable use fw_setenv. This example will set the variable autoload to no, use fw_printenv to read back the change.

target# ./fw_setenv autoload no