On linux 2.6.9, jdk 1.6
Local host: mortimer/172.30.6.12
Open: (datagram, server, /172.30.6.12:5678 => /0:0:0:0:0:0:0:0:1234)
Received: (datagram, server, /172.30.6.12:5678 => /0:0:0:0:0:0:0:0:1234),
HeapBuffer[pos=0 lim=1 cap=2048: 00]
Received: (datagram, server, /172.30.6.12:5678 => /0:0:0:0:0:0:0:0:1234),
HeapBuffer[pos=0 lim=2 cap=2048: 00 00]
Received: (datagram, server, /172.30.6.12:5678 => /0:0:0:0:0:0:0:0:1234),
HeapBuffer[pos=0 lim=3 cap=2048: 00 00 00]
Received: (datagram, server, /172.30.6.12:5678 => /0:0:0:0:0:0:0:0:1234),
HeapBuffer[pos=0 lim=4 cap=2048: 00 00 00 00]
unbound
Closed: (datagram, server, /172.30.6.12:5678 => /0:0:0:0:0:0:0:0:1234)
Adding a virtual interface (eth0:0) doesn't make a difference.
Maarten
On 9/18/07, Maarten Bosteels <[EMAIL PROTECTED]> wrote:
>
> Hello Trustin,
>
> I ran the code on Windows XP jdk 1.6 , and this is the output I got:
>
> Local host: kweenie/172.29.100.104
> Open: (datagram, server, /172.29.100.104:5678 => /0.0.0.0:1234)
> Received: (datagram, server, /172.29.100.104:5678 => /0.0.0.0:1234),
> HeapBuffer[pos=0 lim=2 cap=2048: 00 00]
> Received: (datagram, server, /172.29.100.104:5678 => /0.0.0.0:1234),
> HeapBuffer[pos=0 lim=1 cap=2048: 00]
> Received: (datagram, server, /172.29.100.104:5678 => /0.0.0.0:1234),
> HeapBuffer[pos=0 lim=3 cap=2048: 00 00 00]
> Received: (datagram, server, /172.29.100.104:5678 => /0.0.0.0:1234),
> HeapBuffer[pos=0 lim=3 cap=2048: 00 00 00]
> Received: (datagram, server, /172.29.100.104:5678 => /0.0.0.0:1234),
> HeapBuffer[pos=0 lim=3 cap=2048: 00 00 00]
> Received: (datagram, server, /172.29.100.104:5678 => /0.0.0.0:1234),
> HeapBuffer[pos=0 lim=3 cap=2048: 00 00 00]
> Received: (datagram, server, /172.29.100.104:5678 => /0.0.0.0:1234),
> HeapBuffer[pos=0 lim=4 cap=2048: 00 00 00 00]
> Received: (datagram, server, /172.29.100.104:5678 => /0.0.0.0:1234),
> HeapBuffer[pos=0 lim=4 cap=2048: 00 00 00 00]
> Received: (datagram, server, /172.29.100.104:5678 => /0.0.0.0:1234),
> HeapBuffer[pos=0 lim=4 cap=2048: 00 00 00 00]
> Received: (datagram, server, /172.29.100.104:5678 => /0.0.0.0:1234),
> HeapBuffer[pos=0 lim=4 cap=2048: 00 00 00 00]
> Closed: (datagram, server, /172.29.100.104:5678 => /0.0.0.0:1234)
> unbound
>
> I changed the code a little bit, to be able to distinguish the messages.
> Also added a sleep before the unbind, (I think there's one line missing in
> your output :-)
>
> f1.getSession().write(ByteBuffer.allocate(1)).await();
> f1.getSession().write(ByteBuffer.allocate(2)).await();
>
> f2.getSession().write(ByteBuffer.allocate(3)).await();
> f2.getSession().write(ByteBuffer.allocate(4)).await();
>
> Thread.sleep(2000);
> acceptor.unbind();
> System.out.println ("unbound");
> System.exit(0);
>
> The messages sent to the broadcast address are received multiple times:
> twice per enabled network connection.
> When I enable all 4 network interfaces (LAN, wireless LAN and two virtual
> ones for VMWare) I receive them 8 times !
>
> Maarten
>
> On 9/17/07, Trustin Lee <[EMAIL PROTECTED]> wrote:
> >
> > Hi,
> >
> > I made a lot of changes to the current datagram implementation in
> > trunk. DatagramAcceptor now collaborates with DatagramConnector to
> > create connected datagram socket which listens to the remote address
> > of the received packet. The following is simplified description of
> > flow:
> >
> > * DatagramAcceptor's underlying DatagramChannel.receive is invoked
> > * Once data is received, new connected datagram session is created
> > using DatagramConnector.connect(receivedRemoteAddress,
> > acceptor.getLocalAddress()). If there's already a session instance
> > associated with the remote address, it's reused.
> >
> > The advantages of this implementation are:
> >
> > * The acceptor-side datagram sessions can perform much better in
> > multi-core environment.
> > * There's no need for IoSessionRecycler on the acceptor-side anymore.
> > You have full control over disconnection; probably using sessionIdle
> > event in most cases.
> > * Code duplication can be reduced once again, and most part of the
> > existing acceptor code can be extracted into a generic acceptor
> > implementation.
> >
> > The disadvantage is that it forces you to set reuseAddress to true,
> > because multiple datagram channels are bound to the same local
> > address.
> >
> > In Linux, connected datagram channels have priority over unconnected
> > channels, so most traffic will go directly into DatagramConnector's
> > NIOProcessors. The only exception I've found is broadcast packets. It
> > is always received from the unconnected channel in DatagramAcceptor.
> > DatagramAcceptor immediately forwards the event to an appropriate
> > session that DatagramConnector is managing, or creates a new session
> > using connect().
> >
> > I think this behavior won't have any big differences among different
> > platforms (e.g. Windows and SUNOS). As long as multiple datagram
> > channels can be bound to the same local address, this implementation
> > should work fine. To make this sure, please run the following test
> > code in your machine, and let me know what the result is:
> >
> > -------- CODE BEGINS --------
> > package net.gleamynode.tmp;
> >
> > import java.net.InetAddress;
> > import java.net.InetSocketAddress;
> >
> > import org.apache.mina.common.ByteBuffer;
> > import org.apache.mina.common.ConnectFuture;
> > import org.apache.mina.common.IoHandlerAdapter ;
> > import org.apache.mina.common.IoSession;
> > import org.apache.mina.transport.socket.nio.DatagramAcceptor;
> > import org.apache.mina.transport.socket.nio.DatagramConnector;
> >
> > public class Main {
> >
> > private static final int SERVER_PORT = 1234;
> > private static final int CLIENT_PORT = 5678;
> >
> > public static void main(String args[]) throws Exception {
> > DatagramAcceptor acceptor = new DatagramAcceptor();
> > acceptor.setLocalAddress(new InetSocketAddress(SERVER_PORT));
> > acceptor.setHandler(new IoHandlerAdapter() {
> > @Override
> > public void sessionOpened(IoSession session) {
> > System.out.println("Open: " + session);
> > }
> >
> > @Override
> > public void sessionClosed(IoSession session) {
> > System.out.println("Closed: " + session);
> > }
> >
> > @Override
> > public void messageReceived(IoSession session, Object o) {
> > System.out.println("Received: " + session + ", " + o);
> > }
> > });
> > acceptor.bind();
> >
> > DatagramConnector connector = new DatagramConnector();
> > connector.getSessionConfig().setReuseAddress(true);
> > connector.setHandler(new IoHandlerAdapter() {
> > @Override
> > public void messageReceived(IoSession session, Object o) {
> > System.out.println("ERR?");
> > }
> > });
> >
> > // Print the local host.
> > System.out.println("Local host: " + InetAddress.getLocalHost());
> >
> > // Try point-to-point.
> > ConnectFuture f1 = connector.connect(
> > new InetSocketAddress(
> > InetAddress.getLocalHost(), SERVER_PORT),
> > new InetSocketAddress(
> > InetAddress.getLocalHost(), CLIENT_PORT));
> > f1.await();
> >
> >
> > f1.getSession().write(ByteBuffer.allocate(1)).await();
> > f1.getSession().write(ByteBuffer.allocate(1)).await();
> >
> >
> > // Try broadcast.
> > ConnectFuture f2 = connector.connect(
> > new InetSocketAddress(
> > "255.255.255.255", SERVER_PORT),
> > new InetSocketAddress(
> > InetAddress.getLocalHost(), CLIENT_PORT));
> >
> > f2.await();
> > f2.getSession().write(ByteBuffer.allocate(1)).await();
> > f2.getSession().write(ByteBuffer.allocate(1)).await();
> >
> > acceptor.unbind();
> >
> > System.exit(0);
> > }
> > }
> > -------- CODE ENDS --------
> >
> > The expected result is:
> >
> > -------- RESULT BEGINS --------
> > Local host: primary/127.0.0.1
> > Open: (datagram, server, /127.0.0.1:5678 => /0:0:0:0:0:0:0:0:1234)
> > Received: (datagram, server, /127.0.0.1:5678 =>
> > /0:0:0:0:0:0:0:0:1234), HeapBuffer[pos=0 lim=1 cap=2048: 00]
> > Received: (datagram, server, /127.0.0.1:5678 =>
> > /0:0:0:0:0:0:0:0:1234), HeapBuffer[pos=0 lim=1 cap=2048: 00]
> > Received: (datagram, server, /127.0.0.1:5678 =>
> > /0:0:0:0:0:0:0:0:1234), HeapBuffer[pos=0 lim=1 cap=2048: 00]
> > Closed: (datagram, server, /127.0.0.1:5678 => /0:0:0:0:0:0:0:0:1234)
> > -------- RESULT ENDS --------
> >
> > If this test fails on any platform that runs Java 5, I am doomed to
> > revert my changes. :)
> >
> > Trustin
> > --
> > what we call human nature is actually human habit
> > --
> > http://gleamynode.net/
> > --
> > PGP Key ID: 0x0255ECA6
> >
>
>
