lwip-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [lwip-users] Once UDP fails to receive then resending fails as well


From: Patrick Klos
Subject: Re: [lwip-users] Once UDP fails to receive then resending fails as well
Date: Thu, 4 May 2017 10:09:18 -0400
User-agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:45.0) Gecko/20100101 Thunderbird/45.8.0

On 5/3/2017 4:57 PM, digitalwave wrote:
hi,

I am working on tm4c1294+FreeRTOS+lwip1.4.1 with daemons running on HP-unix

The communication modules had worked fine in previous network environment. 
One day network guy set up a new network env and it stops communication
after runnig 3~4 hours.

I just took over this project and there is no one who is in charge of
daemons, so it's hard to check where the problems come from.

The network guy said it's same as the previous network environment, but it
seems not for me.

Every second, MCU sends a request of server time and receive a response from
a daemon. It runs well for 3~4 hours then what happens is once the MCU
requests the server time, MCU doesn't receive anything.

Even though MCU re-sends the same packet to the server it fails to send at
lwip_sendto().

Obtaining socket fd and close it works properly. Because of UDP , closing
the socket with fd doesn't physically disconnect the connection I guess.

The daemon doesn't spit out any errors although MCU doesn't receive the
response. it usually says something on log file if something happens.

Weird thing is I can't trace down packets using "Wireshark". I don't see any
packets between the server and MCU even if they communicate each other well
before the error happens.

The network guy is not really supportive so I try to figure things out by
myself.

How can I narrow down the issue? and where is the problem comes from?

Regards,
Jin

Two quick suggestions:
  1. Like Simon pointed out, if you're using Wireshark over a switch, that won't be too helpful.  See if you can run tcpdump on the HP-unix system(s)?
  2. What changed about the network?  Did your network guy change subnet sizes? If so, don't forget that the subnet mask needs to given to LwIP during init time.  Give us an explanation of how the network changed.  There might be a clue in there?

What kind of traffic pattern does your device create?  Is there a long time between data spurts?  Maybe one of the HP-unix systems has a firewall that's closing down the path when it doesn't see traffic for a while?

Patrick Klos
Klos Technologies, Inc.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]