We are running QNX 6.3 on x86 CPUs and using io-net with both TCP as well as QNET on a 8 node network out of which only 
3 of them are QNX nodes that use QNET and the rest are non-qnx nodes using TCP. After running this system for 48+ hours,
 we suddenly see the io-net taking close to 70% of CPU and then take more and hang the total system. Any idea as to what
 might cause the io-net to be such a hog.