GSoC 2012

Joel Sherrill joel.sherrill at
Wed Mar 21 11:24:17 UTC 2012


There is a port of an older LWIP version in my ftp directory of It was already old enough when I received it that the porting part wouldn't work directly with what was at that time the newest LWIP. I have heard hints of more recent LWIP uses with RTEMS but nothing submitted. The port has to end up in the official LWIP code. 

Other requirements include it being very desirable to be able to use the same driver with either stack. And to be able to use vanilla existing LWIP drivers with RTEMS LWIP port but not the BSD stack.

I updated the TCP/IP stack update since you read it to reflect what has been done with that effort. The FreeBSD 8.2 USB stack does work with RTEMS and we are making progress of the TCP/IP portion of that code base. 

LWIP port and running should not take all summer. It exists in an older, unmerged form. Adding on working on some of the stack upgrade activities such as how to update the existing drivers that are only in RTEMS and how to have a driver which works with both are good activities.

Look again at the TCP/IP upgrade description and see what smaller activities you could combine with LWIP port upgrade.  The new stack effort is in rtems-libbsd.git.

Ask questions as needed. This is active work and documentation and status reports trails code.


Kevin Polulak <kpolulak at> wrote:

>Interesting. It sounds like the project isn't necessarily tied to just one library but is rather about developing a package format for use with any external third party software. Or does RTEMS already have a pseudo-packaging system that the *Kit projects will use? Either way, it's definitely do-able.
>Actually, I was initially attracted to the FreeBSD TCP/IP stack project. However, like the wiki says, it's way too large to complete in a single summer. Anyhow, I was looking through some of the other open projects and I happened to stumble upon the LWIP project. Since LWIP is a lightweight TCP/IP stack, this might just be conceivable.
>I spent the last weekend reading Adam Dunkel's masters thesis. LWIP is very basic (compared to something like FreeBSD's stack, that is) and easy to understand. I really enjoy network programming so this is something I could really sink my teeth into.
>In his thesis, Dunkel mentions that a port of LWIP generally only requires an implementation of the operating system emulation layer. I've only briefly taken a look at the current RTEMS implementation of LWIP though. What exactly needs to be updated? Is it actually broken or just not utilizing some of the newer features of RTEMS (whatever they may be)? I'll dig through the code a little more once I've completed my Hello World example.
>Speaking of which, I plan to work on my Hello World example later tonight. I'll submit it as soon as I'm done. I want to quick get things rolling so I can become more familiar with the RTEMS source by fixing some bugs and submitting a few patches. :)
>- Kevin Polulak (soh_cah_toa)

More information about the devel mailing list