Google summer of code

[Joel Sherrill]

Kliakhandler Kosta wrote:
>
> Thanks for your comments!
>
> Quoting Joel Sherrill <joel.sherrill at oarcorp.com>:
>
> > Kliakhandler Kosta wrote:
> >>
> >> Hello,
> >>
> >> I'm interested in working on RTEMS for Google summer of code. I'm 
> >> particularly interested in working on tinyRTEMS and porting it to 
> >> an lpc2138 board (which I posseess). I never used it (RTEMS), but I 
> >> am developing an application on the mentioned board and although I 
> >> didn't plan to use an rtos yet, the project will in general benefit 
> >> from it.
> >>
> > OK. There has been a lot of discussion on TinyRTEMS this week so I
> > am not going to repeat any of that.  It's Friday and I am tired. :)
>
> I joined the list only a couple of days ago so I didn't see it, but I 
> will definitely take a read it when I find the archives... (while 
> writing this email I found it :)
>
I did too and realized that the archives at rtems.org/ml/ go back
to 1997 while the pipermail/ ones are only back to 2006.
>
>
> >
> > Some suggestions.  You are proposing an lpc2138 BSP.  So I see
> > a few obvious tasks in your proposal:
> >
> > + BSP for real hardware
> > + BSP runs on Skyeye simulator
> >
>
> AFAIK skyeye doesn't run lpc21xx yet - if so, that will probably be 
> another task.. I haven't worked with skyeye either yet - Is there a 
> significant difference (for rtems or otherwise) for code which runs on 
> skyeye as opposed to code which runs on HW?
>
>
Skyeye is a simulator -- that means it pretends to be the CPU in question
and is supposed to be a 100% accurate functional version of that CPU and
peripherals.  So the goal is that executables interchangeably on both.
>
> > From a TinyRTEMS perspective, there are lots of ideas and you will 
> have to
> > decide which tickle your fancy. How much memory is available on your
> > target hardware?
>
> I'm thinking about what I want to do right now. I was hoping for some 
> guidance about which features seemed most useful to the developers.
>
The recent archives touched on that.... elimination of filesystem
infrastructure for device node only support. dropping c library reentrancy
support, etc.  Those are partially in and not clean.

One way to view the goal is that any program like hello world shouldn't
have  code in it that it isn't actually executing.  There may be paths
it doesn't hit but whole functions that are unused is bad.  Sometimes
those paths themselves may represent unusable code in a configuration.
That needs to be modularized so it drops out when not needed.

As much configuration as possible must be deferred to the application
developer.  See
 http://www.rtems.org/onlinedocs//doc-current/share/rtems/html/c_user/c_user00405.html
for documentation.
>
>
> > Do you have any display option?  If so, consider adding 
> Microwindows/Nano-X
> > support to your proposal.
>
> I have an lcd character screen - I don't know whether it qualifies as 
> a "display", but I certainly will consider it.
>
I have had some of those.  2x16 lines of characters isn't GUI friendly. :)
>
> >>
> >> If possible, I would like some guidence regarding what is 
> >> important, times, and general advice regarding writing a proposal 
> >> for this project.
> >>
> > The student proposal deadline is approaching soon but you still have 
> time
> > to make a proposal.  I think there are guidelines at the SOC site on 
> what
> > outline to follow.  SOC has a schedule and your proposal should 
> track that
> > with activities in stages that make sense based upon that.
>
> I realized that - I meant it in an rtems centric way.. I realise that 
> times for developing the various projects differ between developers, 
> but I was hoping for some suggestions on the projects to take, so that 
> my proposal wouldn't turn out unacheivable due to a massive amount of 
> work on the one hand, or "unsatisfactory" because I proposed things 
> which were too small... I guess what's left to do is to write a 
> proposal and hope I'll have time to revise it according to suggestions...
>

TinyRTEMS is a set of small activities which together keep us moving
in a direction.  So you can propose any number of steps.  A BSP and
Skyeye support is also doable and gives you a concrete configuration
to aim for.
>
> >
> > Maybe one of the others can post some helpful links.
> >>
> >> Thanks in advance,
> >> Kosta.
> >>
> >> --
> >> Kliakhandler Kosta.
> >> Website: http://www.kosta.tk
> >>
> >
> >
> > --
> > Joel Sherrill, Ph.D.             Director of Research & Development
> > joel.sherrill at OARcorp.com        On-Line Applications Research
> > Ask me about RTEMS: a free RTOS  Huntsville AL 35805
> >   Support Available             (256) 722-9985
>


-- 
Joel Sherrill, Ph.D.             Director of Research & Development
joel.sherrill at OARcorp.com        On-Line Applications Research
Ask me about RTEMS: a free RTOS  Huntsville AL 35805
   Support Available             (256) 722-9985