[lwip-users] Re: Sending too slow.

[Kelvin Lawson]

Hi Mateusz,

Have you tried setting TCP_NODELAY?

According to do_write() in api_msg.c:

      /* This is the Nagle algorithm: inhibit the sending of new TCP
   segments when new outgoing data arrives from the user if any
   previously transmitted data on the connection remains
   unacknowledged. */
   if(err == ERR_OK && (msg->conn->pcb.tcp->unacked == NULL || 
(msg->conn->pcb.tcp->flags & TF_NODELAY)) ) {
  tcp_output(msg->conn->pcb.tcp);
      }

I'm not sure that this is correct - it seems to be doing only part of 
the Nagle algorithm. As I understand it, it should buffer up any 
outgoing data until all outstanding data is acked, but also permit a 
send when a full packets' worth of data is ready to send. Judging from 
the statement above, we do wait for outstanding data to be acked, but 
don't permit a send when a full packet's worth is ready.

Kelvin.


Mateusz Plocinski wrote:
> Hi,
> I'm currently working on building audioserver based on AT91SAM7X256 
> microcontroller and freertos+lwip stack.
> I'm new to lwIP and I encountered one problem while sending packets. When I send a frame (for example 32 KB), lwip sends it as 1.5KB packets and thats normal for TCPIP. But every packet it waits for acknowledgement from client (PC, WinXP). 
> And PC sends it after 200ms!
> I have to send about 100KB per second or more and waiting for ack 200ms is not 
> possible. As far as I know, I have a right to send the whole window size 
> (windows XP sends 65KB window requests) without waiting for ack.
> My question: is it possible to make lwIP stack send packets faster without 
> waiting for ack every single packet?
>
> We, at our company have written a stack for AVR atmega128 which is able to send 
> about 16mbits/s without problem (ack's come every few packets but Windows sends 
> them much faster than every 200ms when it gets more data). But this stack is 
> buggy and I would like to use lwIP on ARM7)
> Can you tell me if fast sending is possible under lwIP or if this is only 
> developed for slow, every-packet-acked connection?
> Please help ;]
>
> Mateusz Plocinski