Re: [Ltib] Regarding nanosleep implemation problem in LTIB

[Peter Barada]

On 02/04/2012 12:40 AM, lakshminarayana babu wrote:
> I am using motorola MCF5485 processor board. We have ported  linux 
> 2.6.25 kernel downloaded from LTIB.   .  I have written a small 
> application program using "nanosleep". If i  give a delay of 100 
> milliseconds it is giving 120 miliseconds. Like this whatever may the 
> amount i am giving in this function it is adding 20 milliseconds 
> extra.  I tried with "usleep" and "sleep" also. In both the cases 
>  same problem. Can you please help me out.
1) replying to a digest email in the list is not good - you've hijacked 
an email thread and leave a lot of us trying to figure out the context 
of your questions.

2) using nansleep from a userspace application either has the userspace 
trying to sleep (while the scheduler is still running), or it invkes a 
system call to do the same, either of which can easily add at least one 
tick to the delay; sleeps of that short a duration should only be done 
within the kernel - sounds like you're trying to do in userspace 
application what should be done in the kernel (i.e. hardware driver?). 
What is it you're really trying to do that requires a sleep of that 
short a duration???

--
Peter Barada
address@hidden