Rob
I apologize for thge very late reply
1. memset. You're right there's a problem. IMHO this is the fix
memset(myGlobals.device[deviceId].packetQueue, 0,
sizeof(PacketInformation) * (CONST_PACKET_QUEUE_LENGTH+1));
2. for the calloc nothing really changes when reverting parameters.
Regards Luca
----
Luca Deri <[EMAIL PROTECTED]>
You must be the change you want to see in the world
Mahatma Gandhi
On Aug 5, 2008, at 11:27 PM, Rob Landrito wrote:
> Greetings, I've been trying to make a few modifications to the ntop
> source and started to get mysterious floating point exceptions in
> areas unrelated to my additions. Some debugging in gdb led me to some
> odd initializations of myGlobals.device[id].packetQueue. (or maybe
> not so odd, but I'm not a very experienced coder). The following
> modifications seemed to clear things up...
>
> --- initialize.c 2008-08-05 10:07:44.000000000 -0400
> +++ initialize.c.new 2008-08-05 17:16:55.000000000 -0400
> @@ -887,5 +887,5 @@
> createMutex(&myGlobals.device[deviceId].packetProcessMutex);
> createMutex(&myGlobals.device[deviceId].packetQueueMutex);
> - memset(&myGlobals.device[deviceId].packetQueue, 0,
> sizeof(PacketInformation));
> + memset(myGlobals.device[deviceId].packetQueue, 0,
> sizeof(PacketInformation));
> myGlobals.device[deviceId].packetQueueLen = 0;
> myGlobals.device[deviceId].maxPacketQueueLen = 0;
> @@ -901,9 +901,9 @@
> if(!myGlobals.device[deviceId].ipPorts)
> myGlobals.device[deviceId].ipPorts =
> - (PortCounter**)calloc(sizeof(PortCounter*), MAX_IP_PORT);
> + (PortCounter**)calloc(MAX_IP_PORT, sizeof(PortCounter*));
>
> if(!myGlobals.device[deviceId].packetQueue)
> myGlobals.device[deviceId].packetQueue =
> - (PacketInformation*)calloc(sizeof(PacketInformation),
> (CONST_PACKET_QUEUE_LENGTH+1));
> + (PacketInformation*)calloc((CONST_PACKET_QUEUE_LENGTH+1),
> sizeof(PacketInformation));
> }
>
>
> Does the patch seem reasonable?
>
> --
> :wq!
> _______________________________________________
> Ntop-dev mailing list
> [email protected]
> http://listgateway.unipi.it/mailman/listinfo/ntop-dev
_______________________________________________
Ntop-dev mailing list
[email protected]
http://listgateway.unipi.it/mailman/listinfo/ntop-dev
