Hello Sanjay,

One suggestion - its probably better to precede the output lines with cpu no. As currently output from child processes are intermixed as below and difficult to co-relate.

jiffies are : 10000 usecs
found 4 cpu(s)
duration: 120 secs, #sleep: 1200, delay: 100000 us
duration: 120 secs, #sleep: 1200, delay: 100000 us
duration: 120 secs, #sleep: 1200, delay: 100000 us
duration: 120 secs, #sleep: 1200, delay: 100000 us
counter value 4072718528
test duration: 120.057526 secs
deviation 0.000479
counter value 3914063589
test duration: 120.057335 secs
counter value 4015937716
test duration: 120.057430 secs
deviation 0.000479
deviation 0.000478
counter value 411037603
test duration: 120.062716 secs
deviation 0.000523



--
Thanks,
-Meraj


On Fri, May 2, 2014 at 11:23 PM, Sanjay Singh Rawat <sanjay.rawat@linaro.org> wrote:
On Wednesday 30 April 2014 02:04 PM, Mohammad Merajul Islam Molla wrote:
Hello Sanjay,
As far I know, if option argument is 0, the parent will wait for

Looks like waiting is done if childs exit. I checked for the offlined CPU case, if there are no child processes, waitpid returns -1. Setting errno as "no child processes"

specified child pid to terminate, its not for immediate return as in
case of WNOHANG. This is probably the intended use of the code (author
will be able to confirm). Changing 0 to WNOHANG macro will change the
meaning of the code.
Also, from header file -
#define WNOHANG         0x00000001

sorry i referred wrong document

WNOHANG is defined as 1, not zero.
Which makes me think of two cases below -
1. If the real intended purpose is to not to wait infinitely, replacing
[...]