RTEMS socket programming (select usage etc.)

[Richard M. Barry]

My suggestion is:

Use a single (server) task to read and write from the socket, and then use message queues to asynchronously communicate with the server from your other tasks. 

Richard

-----Original Message-----
From: Jonas Moberg [mailto:jonas.moberg at solidsoftware.net]
Sent: 13 November 2003 12:48
To: rtems-users at rtems.com
Subject: RTEMS socket programming (select usage etc.)


In the Network documentation I read the following information:

"The RTEMS network package provides almost a complete set of BSD network
services. The network functions work like their BSD counterparts with
the following exceptions:
A given socket can be read or written by only one task at a time. "

Would this state that one can use only one task for reading and writing
or one task that reads and
another task that writes?
If we don't want to do a traditional server that reads, processes data
and then writes to a socket
by one thread what, is then the recommended way?
We would like to asyncronously read and write from several tasks.
The rtems select implementation says that usage should be avoided if
possible.

Solid Software AB
Jonas Moberg