This is the mail archive of the ecos-discuss@sources.redhat.com mailing list for the eCos project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

graceful termination of TCP/IP stack


I am testing my system, pinging it to death until I 
get "out of mbuf" conditions (panic bug repaired) and a 
watchdog I set in my network thread quits getting set.  

I added the watchdog in another task because once the out of mbufs condition has gotten out of control, I can not recover by killing all of my pings.

So, upon watchdog, I 

1) suspend my network task
2) close my network task sockets (not ping sockets)
3) kill my network task 
4) create network task
5) resume network task

My network task does restart, my socket and bind 
commands return OK, but the connect TCP fails.

I think the TCP/IP stack task I have read about is still
running.  I only close my sockets before killing my 
network task. 

thank you for any help on gracefully terminating the TCP/IP,

regards,

Dan Ash
SonicBox


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]