|Subject:||[lwip-users] Lwip problem when PSH,ACK sent to a listening socket.|
|Date:||Fri, 11 Mar 2005 16:04:31 -0000|
Some information first.
LwIP Version 7.2
ST10168 Embed Controller
No operating, using callback interface.
We had LwIP up and running in our application for +6months now but only using minor interfaces
Ie telnet, NTP etc.
Our problem came about whilst testing for system robustness/recovery.
The system comprises of a windows 2000 Sever, running our server application.
This server application connects to 100's of Lwip (Embedded Controllers).
In order to ensure that Lwip (Embedded Controllers) are alive the server application
polls the Lwip (Embedded Controllers) they in turn reply with there status.
This enables us to detect when Lwip (Embedded Controllers) have gone off line, Ie network, power issues etc.
This is where we found the problem.
In the attached trace.txt is the TCP trace of the server connecting to one Lwip (Embedded Controller).
After connection the Lwip (Embedded Controller) power is removed and reapplied. Thus causing the Lwip (Embedded Controller)
To return to a listening socket state. But the server does not know this and therefore polls the Lwip (Embedded Controller). Now because the Lwip (Embedded Controller) did not reply to the poll, the server resets its connection, and waits 60secs. After the a lapsed time the server tries to re-connect but Lwip (Embedded Controller) will not allow it.
The Lwip (Embedded Controller) socket is now useable.
Why should this be ?
Please see trace for more detail TCP communications
Tel: 0113 230 8820
Fax: 0113 279 9229
<<Lee Trueman (address@hidden).vcf>>
Description: Text document
Lee Trueman (LTrueman@coe.co.uk).vcf
Description: Binary data
|[Prev in Thread]||Current Thread||[Next in Thread]|