Re: [lwip-users] problems with pbuf alloc

[bingsen xie]

I found a problem similar to this too.
The LWIP is also v0.5.3 with uCOS/II.
My target board is run the example "tcpecho",
And In my host PC, a client send the data to "tepecho"
all the time.The PC client application also receive
the data send back from "tcpecho".But some strange
things occur. If the PC Client's sending packet size
is 40 byte, when the total sended packet 
size is 77.96k, the tcpecho will stop working, the
connectiong is closed. 
I don't know why.

Hi group,
>  
> I am using LWIP v0.5.3 with uCOS/II.
>  
> I have made a small program using the RAW API that
> listens for an
> incoming TCP connection and then it sends 1000 Bytes
> of random data when
> it receives a "GET" command from the remote host.
> The remote host sends the "GET" command every 100ms.
> This should
> simulate somehow a streaming data transfer.
>  
> After a certain amount of packets the transmission
> simply stops. When I
> print out some debug information I can observe that
> the transmission
> stops due to a problem with the PBUF ALLOCATION.
> Below I have printed the debug information of the
> two last requested
> data packets. The packet (1) was o.k. but for packet
> (2) the tcp_output
> could not allocate a PBUF.
>  
> Is there anything known about such problems or does
> anybody have a
> solution for this.
>  
> For me it seems that deallocated PBUF memory is not
> used anymore for the
> allocation of new PBUFs and therefore program is
> running out of memory.
>  


_________________________________________________________
Do You Yahoo!? 
嫌邮箱太小？雅虎电邮自助扩容！
http://cn.rd.yahoo.com/mail_cn/tag/10m/*http://cn.mail.yahoo.com/event/10m.html