lwip-users
[Top][All Lists]
Advanced

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

Re: [lwip-users] recv and sent callbacks usage


From: narke
Subject: Re: [lwip-users] recv and sent callbacks usage
Date: Thu, 11 Aug 2011 21:47:43 +0800

On 11 August 2011 21:15, Kieran Mansley <address@hidden> wrote:
> On Thu, 2011-08-11 at 21:11 +0800, narke wrote:
>> Hi,
>>
>> I have a few questions about the raw api:
>>
>> 1. When a connection received 3 segments, how many times the recv
>> callback will be called?
>> 2. After continuously  did 3 times of tcp_write, how many times the
>> sent callback will be got called after all the data have been sent out
>> and acknowledged?
>
> I think both of the above are undefined by the API.  I.e. you shouldn't
> rely on any particular behaviour.  You may get called once, you may get
> called 3 times, you may get called more.

For the 2nd question, if I called tcp_write 3 times, each time with a
100 bytes pbuf, and I when I got the first call of sent callback with
a len argument equals to 100, should I prepare myself to resent the
last two 100 bytes pbuf again?

In above example, is that possible I got a sent callback with a len
that is not a multiple of 100?


>
>> In the sent callback, how do I understand and use
>> the 'len' argument?
>
> It tells you how many bytes have been sent and acknowledged, and so you
> no longer need to maintain the next "len" bytes of data as valid.
>

Why I need to keep maintain this 'len' of bytes after I pass it to
tcp_write?  I don't need to worry about resenting things, right? It
should be a job of lwIP.



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



-- 
Life is the only flaw in an otherwise perfect nonexistence
    -- Schopenhauer

narke
public key at http://subkeys.pgp.net:11371 (address@hidden)



reply via email to

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