Hi all,

I need your help finding out what could be the problems causing the io-pkt-v4-hc crash occationally when I issue a "
ifconfig en0 down" command on QNX 6.5.0. All clues are appreciated!

Some details of the setting up.(Please let me know if you need more information):

There are two QNX PCs(works as 1+1 backup). Each QNX PC has two NICs (en0 and en1); one of the two NIC (en1) on each PC 
is used to connect with the other PC's en1 running mqueue with qnet protocol. The other NICs (en0) on each PC are 
configured with the same MAC/IP addresses (one PC en0 is down and another PC en0 is up) to work as 1+1 backup. Both en0 
are connected with a third PC to provide service. I am running rp-pppoe 3.20(built from source) server on both QNX PCs 
and pppoe client on the third PC. I need to test hot swap between two QNX PCs  (basically, move the established pppoe 
sessions from one PC to another without redial, both PCs have the same pppoe session information before moving).

It works well most of the time. After I "ifconfig en0 down" one PC and "ifconfig en0 up" another PC, the pppoe sessions 
on the second PC take over by redoing the LCP negotiation automatically. But some time, when issuing the ifconfig en0 
down on one PC and issuing the ifconfig en0 up on another(down is about 300ms earlier than the up command), the one 
which ifconfig en0 down is issued reports "bad file descriptor" for the socket or "no such process" errors. And the io-
pkt-v4-hc disappeared from the "ps -A" list. When I checked by pidin mem, neither io-pkt-v4-hc nor devn-rtl.so/devnp-
shim.so/devn-rtl.so/lsm-qnet.so exist. nicinfo is empty. ifconfig list is empty too. I have to reload all NIC drivers or
 restart the OS to recover this.

It does not happend all the time, some time it happens if I swap once; some time I need to do swap more than 100 times 
to make it happen. 

Thanks a lot in advance!

Yours
Jay