Re: [lwip-users] raw interface, detecting retries

[lwip]

Moin Simon,

many thanks. I should have seen that.

I was puzzled by some code in contrib-0.7.1-20040319\port\msvc6\httpd.c.
http_poll checks for retries there :-(

Friedrich


----- Original Message ----- 
From: "Goldschmidt Simon" <address@hidden>
To: "Mailing list for lwIP users" <address@hidden>
Sent: Wednesday, February 20, 2008 12:36 PM
Subject: RE: [lwip-users] raw interface, detecting retries


Hi,

TCP retransmissions are taken care of internally, there is no need to care about this in your application code. Once the (possibly 
retransmitted) data is acked, tcp_sent will be called. As to tcp_poll, I don't remember exactly, but I think it is always called, no 
matter which state the tcp_pcb is in...

Simon

-----Ursprüngliche Nachricht-----
Von: address@hidden 
[mailto:address@hidden Im Auftrag von address@hidden
Gesendet: Mittwoch, 20. Februar 2008 12:22
An: address@hidden
Betreff: [lwip-users] raw interface, detecting retries

Moin all,

I use the raw interface to xmit some lengthy data over TCP. My applications 
works fine so far.

I have difficulties to implement retries in case of lost datagrams.

tcp_write sends data without copying it. The data is kept until tcp_sent callback notes success. Thereafter more data is prepared 
and xmitted.

What happens if the data gets lost? I expect tcp_sent will not be called as the other side did not acked it. After a while the 
tcp_poll callback will be called, which in turn could retransmit the data.

Nevertheless I see tcp_poll callback even when everything works fine.

How can I differentiate whether a retransmit is required or more data should be 
send?

Did I miss something in the docs? Any help apreciated.

Friedrich



_______________________________________________
lwip-users mailing list
address@hidden
http://lists.nongnu.org/mailman/listinfo/lwip-users


_______________________________________________
lwip-users mailing list
address@hidden
http://lists.nongnu.org/mailman/listinfo/lwip-users