GSOC 2013 project " Kernel Size Reduction for Embedded System "

[Kevin Day]

On Apr 8, 2013, at 7:34 PM, Alfred Perlstein <bright at mu.org> wrote:
> However, until a bunch of embedded folks come forward and state what they are really willing to sacrifice, then we won't really have anything to go on, and it will be guessing at what will work for a space that not all of us are familiar with.
> 
> So I'm hoping some people can make the tough call to give direction here, otherwise nothing good will come of it.
> 
> Has anyone actually done this?  Or maybe compared against another embedded OS?


Ages ago we had to make things work in 16 or 32MB of total system memory on i386. 

For the most part, disabling every compiled-in option/driver we didn't need was 90% of the effort. Which options/drivers is going to be totally application dependent, so that really can't be done for you.

As for the rest, there isn't any large low hanging fruit that can get culled from the kernel easily. The base kernel isn't modular enough to trim out individual syscalls or anything, and doing so wouldn't have made a huge dent.

There are a lot of ways FreeBSD could be more embedded friendly (being able turn on/off parts of userland depending on licenses is a huge one), but producing a trimmed kernel isn't something I'd rank very highly. If building a kernel with everything modularized as possible isn't small enough, FreeBSD probably isn't going to work for you for other reasons.