QNX neutrino TCP/IP crash due to duplicate IP

Hi All,

I have a QNX neutrino based Embedded system running on NS486 processor based target board. I have a windows application to connect to this device using TCP/IP socket program. Now i am facing a problem for which i need some help.

The problem is that, when my QNX embedded system is remotely connected and if in the same subnet, one more PC or device is configured with the same IP address of my QNX device, then the QNX device will crash TCP/IP and i’m no more able to ping to the QNX device.

Let me know why this happens and what is the solution to this problem.
Is this problem with QNX neutrino TCP/IP stack? how to solve this?


Are you sure the device tcp/ip crashed or is it just your switch routing the packages to the other pc with same ip even after removing the cable? This is by the way the default behaviour of a switch, you would need to ping from your device to your host to remove the lock in the switch. If tcp/ip really crashes this is not the default behaviour but i never ran into this.

Thanks for the reply. If the QNX device is not connected remotely using windows Application, then if another device in the same subnet has the same IP as of the QNX device, the ping to QNX device fails. This can be recovered by removing the QNX device network cable from the switch and reconnecting back to switch or restarting the switch. After this i can ping to QNX device. But if the QNX device is remotely connected using windows application (The application receives video frames from the QNX device) and in this condition another device with the duplicate IP is added to the network, the QNX device hangs and only power cycling can get it back to working. Let me know your comments about this problem.
I would like to know if QNX4 and QNX neutrino TCP/IP stack has any problems?

Thats what i meant with the behaviour of the switch.

The only thing i can think of is your frames trying to get send out will queue up and make your supplier hang up and end up in an overflow.
Generally there are some problems with ever part of code, so has tcp/ip, but i did not ran in any situation til now (lucky me).

Thanks Again, Even if i am not remotely connecting the QNX device, sometime it doesn’t respond to ping command. The only way to overcome this issue is to reboot the unit. How do i check if the TCP/IP stack is working properly during this scenerio?
In my Case the TCP/IP is spwaned as a separate process.
Please suggest.

One more thing i would like to know is that, from the QNX device end, how will i detect the switch lock due to duplicate IP address?
I don’t have provision to run the ping command from the QNX device, in this case how do i remove the lock?

If you have no alternate path (such as a serial console) then there is no way to check.

Luckily under QNX, this is always the case. You can use the HA manager to automatically restart the tcpip stack should it crash or hang.

bot alert