Description of problem:

customer wants to know which function is high accuracy cause when they calling nanosleep, the function will weak up after the time they specified

in customer's environment, they called nanosleep sleep 125ms, and they observed this function weak up at 131ms ( exceeded about 5% ) 

according the man page of nanosleep, it's clearly described this function may still be a delay
~~~
 If the interval specified in request is not an exact multiple of the granularity underlying clock (see time(7)), then the interval will be rounded up to the next multiple.  Furthermore, after the sleep completes, there may still be a delay before the CPU becomes free to once again execute the calling thread.
~~~

can we provide a high accuracy function to replace "nanosleep" ?

Version-Release

Red Hat Enterprise Linux 6.7

How reproducible:

We can reproduce this issue by following code
~~~
[root@rhel6-repo test]# cat test.c 
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>

#if _POSIX_C_SOURCE >= 199309L
#include <time.h>   // for nanosleep
#else
#include <unistd.h> // for usleep
#endif

void sleep_ms(int milliseconds) // cross-platform sleep function
{
#if _POSIX_C_SOURCE >= 199309L
    struct timespec ts;
    ts.tv_sec = milliseconds / 1000;
    ts.tv_nsec = (milliseconds % 1000) * 1000000;
    nanosleep(&ts, NULL);
#else
    usleep(milliseconds * 1000);
#endif
}

int main()
{
	int i;

	for (i = 0; i < 8; i++) {
		printf("%d\n", i);
		sleep_ms(125);
	}
	return 0;
}
---------------------------------------------------
[root@rhel6-repo test]# time ./a.out 
0
1
2
3
4
5
6
7

real	0m1.002s
user	0m0.000s
sys	0m0.000s
~~~

Actual results:
0m1.002s

Expected results:
0m1.000s

Request:
1. can nanosleep improve?
2. any other systemcall can do the same thing and sleep correctly?

many thanks !!