Help: /bin/ping problem with Red Hat

[Vivek Sawant]

I am having a peculiar problem with the /bin/ping program. I am running
Red Hat with Kernel 2.4.18-19.7.xsmp. Ping to a non-responsive (down)
host does not terminate when invoked from my python application even
though I use the timeout (-w) flag. Details:

My Python application has two threads: a main thread and another created
by the main thread. The second thread invokes the following ping command
using os.system () of Python.

/bin/ping -c 3 -s 548 -p "0003a8e4" -w 3 -l 3  <destination>

The stdout and stderr of this command has been redirected to a file.
For a destination that is down (does not respond to pings), the above
command never terminates while it is supposed to after 3 seconds (-w 3).
If I cut and paste the above command at the command line, it terminates
after 3 seconds as expected even though the destination does not
echo back any of the packets.

If I kill the above ping command using 'kill -9', the rest of my app
logic behaves like the ping has terminated. Also, for the destinations
that respond to ping, the app is behaving normally. So, I don't think
there is any problem with my prorgam logic.

While the ping thread is pinging, the main thread periodically reads a 
file to see if the user has signalled termination of application. This
thread gets periodically scheduled even when the ping thread is stuck as
described above.

Does anyone have any insight about what's going on here. I am hopelessly
stuck :-(

Thanks.

\vivek