/proc/stat file problem for cpu usage

arunka · 10-03-2005, 05:47 AM

hi all
i am having linux kernel version 2.4.21-138 (suse 8),
some times i observe that in /proc/stat file the count of cpu_idle
field will be geeting decremented if i cat the /proc/stat file two times.
since these are the statistics that will be from the system uptime,
i believe either the value should be same or incremented.
i think it is a bug in kernel code...
Also can any one let me know how top calculates the CPU usage??

Thanks in advance...

-Arun KA

bulliver · 10-04-2005, 04:48 AM

This is from man (5) proc re: /proc/stat:

Quote:

The amount of time, measured in units of USER_HZ (1/100ths of a second on most architecures), that the system spent in
user mode, user mode with low priority (nice), system mode, and the idle task, respectively. The last value should be
USER_HZ times the second entry in the uptime pseudo-file.

This is from proc/sysinfo.c, which is included in the procps package. This is where the source for 'top' includes...

Quote:

* Some values in /proc are expressed in units of 1/HZ seconds, where HZ
* is the kernel clock tick rate. One of these units is called a jiffy.
* The HZ value used in the kernel may vary according to hacker desire.
* According to Linus Torvalds, this is not true. He considers the values
* in /proc as being in architecture-dependant units that have no relation
* to the kernel clock tick rate. Examination of the kernel source code
* reveals that opinion as wishful thinking.
*
* In any case, we need the HZ constant as used in /proc. (the real HZ value
* may differ, but we don't care) There are several ways we could get HZ:
*
* 1. Include the kernel header file. If it changes, recompile this library.
* 2. Use the sysconf() function. When HZ changes, recompile the C library!
* 3. Ask the kernel. This is obviously correct...
*
* Linus Torvalds won't let us ask the kernel, because he thinks we should
* not know the HZ value. Oh well, we don't have to listen to him.
* Someone smuggled out the HZ value. :-)

I recommend you read this source code to see what is going on. You can get the source here:
http://procps.sourceforge.net/

There are however, easier ways to get this info. See man (2) sysinfo, for a system call.

danibel.bonetti · 06-20-2009, 10:49 PM

Hi, try this:

Code:

#include <stdio.h>
#include <stdlib.h>
#include <sys/time.h>
#include <unistd.h>

double time_so_far()
{
 struct timeval tp;

 if (gettimeofday(&tp, (struct timezone *) NULL) == -1)
   perror("gettimeofday");
 return ((double) (tp.tv_sec)) +
   (((double) tp.tv_usec) * 0.000001 );
}


int main() {

	FILE *f1;
	double ti, tf;
	
	char c[10];
	int i1,i2,i3,i4,i5,i6;

	ti = time_so_far();
	f1 = fopen("/proc/stat", "r");
	fscanf(f1, "%s\t%d\t%d\t%d\n", c, &i1, &i2, &i3);
	fclose(f1);

	usleep(1000000);

	tf = time_so_far();
	f1 = fopen("/proc/stat", "r");
	fscanf(f1, "%s\t%d\t%d\t%d\n", c, &i4, &i5, &i6);
	fclose(f1);

	int t = (i4+i5+i6)-(i1+i2+i3);
	printf("cpu usage: %.1f%%\n", (t / ((tf-ti) * 100)) * 100);

	return 0;
}

eyas · 04-23-2014, 01:08 PM

it's work carefully