Re: [lwip-users] process model

[zakaria jouilil]

Up!!

(seems like nobody saw my message :(  )

2014-03-21 2:33 GMT+01:00 zakaria jouilil <address@hidden>:

Hi everyone.

First of all, thank you for all your answers, your community is very very active and it's really nice to know that. I solved my problem with callback easily (it was in fact simple) and the discussion continued about the problematic about thread, multi-threading (it interests me also but I think that I will come back to it only if I master the basic concepts of lwip). It's why I decided to open a new discussion here. It's because I didn't receive yet all the answers for my previous questions (for example, which points can we configure in this lwip stack?).

 

As the title suggests, I will talk about the process model of the lwip stack which was present in Adam Dunkel’s documentation.

From what I understand, I think that the process model can be represented like that:

http://image.noelshack.com/fichiers/2014/12/1395364195-sans-titre.png

(I apologize for the quality of my scheme, I did it with paint)

 

I think that a scheme is always better to understand than a long speech, so correct me please if you see that I was wrong with this representation.

 

So, when I was looking at the scheme, I found a lot of questions that I want to ask you about:

-       You see in this scheme, that “all” protocols reside in the same place: between the high level and the low level. It’s what I mean when I talked about “static protocol structure” in my previous message. And I want to know if the protocols have to follow this structure or can we get them out.

I read that the lwip was mainly designed for simple OS which don’t support swapping out process. In this case, the answer to my question is negative but I need to be sure about it

-       Second, always about the internal structure of the stack. About the internal service of a protocol, will the protocols offer a single service or is it possible to offer multiple services? (In this case, how can I model these services?). Then, how can we give an access to a service which is external to the stack? Finally, I want also know if the communication between protocols must always be bidirectional?

-       In the low level (which is the provider level), can we have one or many access point? Because for the high level, it’s clear that we will have many access points (since protocols offers different services) but for the low level, what can we say?

 

 

And that’s all.

Sorry, if I ask to many questions, it’s because I really meet some difficulties to understand all the concepts here.