Subject:
|
Re: Setting LNP host address
|
Newsgroups:
|
lugnet.robotics.rcx.legos
|
Date:
|
Fri, 24 May 2002 02:58:41 GMT
|
Viewed:
|
2209 times
|
| |
|
|
At first, it seems logical that firmdl3 would be the best place to make the
change. However, it was very easy to allow the Host ID to be changed at
runtime.
Whether changed by a user program or changed via the RCX buttons/lcd, it is
easy enough to do. For statically linked programs, the Host ID can be set
near the top of main(). For dynamically linked programs, the code would be
in program.c allowing the user to view/set it while no programs are running.
(similar to viewing free memory now)
I like this idea since it allows the user to change the ID without reloading
anything. Like at a competition where a computer may not be readily
available, etc...
// Joe
I am wondering if legOS supports assigning of lnp ports to specific
processes (or tasks). I guess I will have to look around in the code ...
maybe it can already be done.
"Stephen M Moraco" <stephen_moraco@agilent.com> wrote in message
news:GwKLEt.Jq5@lugnet.com...
> In lugnet.robotics.rcx.legos, Ted Hess writes:
> > I've experimented with a couple of choices which were relatively easy to
> > implement. Let me know if we should keep any these:
> >
> > 1. You can choose to set the host address at build time with a • command-line
> > option to make.
> > <snip>
> >
> > 2. I have added a new protocol message to program.c and added a new • command
> > to dll that will dynamically change the RCX host address when invoked.
> > <snip>
> >
> > Do you all think it is still important to be able to set the LNP host
> > address during system image download (firmdl3)? These features seem to • fit
> > the need and modifying firmdl3 to modify the image is a bit messy.
>
> Joe Woolley prototyped the firmdl3 mods, I think, and has now a user
> program which can change the Id. This now makes 4 approaches we
> have to consider ;-)
>
> My thinking on firmdl3 use was that it is a direct replacement for
> compile time set that we have today. It seems to me that we want a
> fixed id in each bot so that only one is listening to
> downloads/messages, so they can talk "amongst themselves", etc.
> (yeah, I know, this is probably obvious :) However, I came to this
> thinking rather simply. In looking for other patterns it seems to me
> that host id is one of those things that would normally be a set of
> dip-switches on the back panel of equipment, for example.
>
> When I first heard that we ask users to recompile legOS for each
> id, that struck me as obviously wrong. It also struct me that setting
> it at kernel download time was the next best approach. It also occurred
> to me that we might have a concern for being able to load the exact • same
> kernel in different bots and rebuilding the kernel made it easier to
> be not the case.
>
> One final set of thoughts. Helping us build an "intuition" of how to • use
> the tools. I come to firmdl3 being the logical place for the ID set • for a
> couple of reasons: (1) one must set the id each time a kernel is • downloaded,
> (2) I don't think of a program that needs to specify the target id for • a
> download as the logical place to set the host id of the bot to which i
> then must download as the intended target. That is to say, firmdl3
> intuitively made more sense... I also acknowledge the need to change
> the id once it has been set. I don't think this eliminates firmdl3 as
> the tool from which to do this, however.
>
> I know this is kind of a stream of conciousness writeup (barely a
> natural flow... ;-) but does this make sense? Is there anything here
> to guide us? I simply offer it as more to think about and discuss
> before we make the choice.
>
> Regards,
> Stephen
> --
|
|
Message is in Reply To:
 | | Re: Setting LNP host address
|
| (...) Joe Woolley prototyped the firmdl3 mods, I think, and has now a user program which can change the Id. This now makes 4 approaches we have to consider ;-) My thinking on firmdl3 use was that it is a direct replacement for compile time set that (...) (23 years ago, 23-May-02, to lugnet.robotics.rcx.legos)
|
4 Messages in This Thread:
  
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
