Re: [lwip-users] TCP Previous segment lost - lwip down to crawl

[Simon Goldschmidt]

m8847 <address@hidden> wrote:
>  I have this problem where lwip (1.4) suddenly goes into a state where 
>  throughput is very low. 
>  [..]
>  The log is reduced, but all messages between these two nodes are not 
>  filtered.
>  10.0.0.123 is where I am logging below.

As always, it would help to mention which IP address is lwIP and which is the 
other and (and what OS the other end is running)! Also, please give pcap files 
a correct extension so that they can be opened automatically.

Then, why does the pcap only include packets from .123 to .121, not the respond 
packets? And last, please attach longer traces so that we can see a bit back in 
time (also to allow wireshark to better decode the log).

>From the small excerpt you sent, it seems like lwIP decided to send zero 
>window probes. This might be triggered by an actual zero window received or by 
>a bug in 1.4.0 that will be fixed in 1.4.1 (see 
>https://savannah.nongnu.org/bugs/index.php?34517). In that case, low resources 
>might well trigger the bug sooner than if you had plenty of ram...

Please send us a full log.

Simon
-- 
Empfehlen Sie GMX DSL Ihren Freunden und Bekannten und wir
belohnen Sie mit bis zu 50,- Euro! https://freundschaftswerbung.gmx.de