NanoX was Re: RTEMS Shell

[Rosimildo da Silva]

From: "Joel Sherrill" <joel.sherrill at OARcorp.com>
To: <amimjf at connectfree.co.uk>; "Rosimildo da Silva"
<rdasilva at connecttel.com>
Cc: <rtems-users at oarcorp.com>
Sent: Monday, March 18, 2002 8:18 AM
Subject: NanoX was Re: RTEMS Shell


> Rosimildo.. please provide advice on NanoX and GUIs.


Ok, From what I did two years ago. These days, I am not spending
much time with embedded systems, but I am starting to see the light
at the end of the tunnel, again..... :-)




> amimjf at connectfree.co.uk wrote:
> >
> > >Dear ppl,
> > >
> > >I'm still trying to understand how RTEMS works.. I've downloaded the
> > >4.5.0 and compiled it fine.. After compilation, I get a bunch of files
> > >in the build/BSP/samples directory. I'm built it for the LEON.. and I
> > >tested the software with TSIM..
> > >
> > >Now it occurs to me, that there seems to be lack of a shell.. The files
> > >run fine.. Even the multi-thread sample runs fine.. So, this tells me
> > >that the OS is working.. Just one problem.. there's no shell..
> > >
> > >Is there a shell for the RTEMS??
> >
> > do you mean a unix/linux type bash/csh/sh shell ? or a GUI shell ?
> >
> > rtems is an embedded operating system its not like unix where you use
the shell to copy files around, edit them for configureation, etc. So there
is no real need for a 'shell' as such.
> >
> > Of course people have created / ported shells to rtems one of them is
even quite small, i think its called 'busybox' or 'blackbox' or somehting
like that, i cant remeber exactly.
>
> The "shell" included with RTEMS is a mini-POSIX-ish shell that is very
> basic.
> Deeply embedded systems just don't usually require that type of
> interaction.
> But there is a telnetd that allows remote access to that shell and even
> login support. :)
>
> RTEMS is generally a statically linked environment in which your
> system is more akin to a single UNIX executable.  There is work on
> dynamically loading application pieces but that does not change
> the overall structure.  RTEMS is logically a single process,
> multi-thread environment.
>
> [NOTE: There is a notion of a psuedo-process which allows a
> group of threads to have their own copy of data which UNIX
> would consider "process data" but there is no memory protection
> at all.]
>
> > >What kind of filesystem does the RTEMS use??
> >
> > any you want, (or can port accross), its does have a flash filesystem
and a ram filesystem.
>
> And a TFTP filesystem and now a FAT12/16/32 filesystem.
>
> > >What's the filesystem organization??
> >
> > well, as i said rtems isent unix (RNU !), its does not store parts of
itself in files, the device access and other stuff is based around unix
(/dev etc)
>
> Right.  Code that has been ported from UNIX/POSIX to RTEMS has
> expectations of
> filesystem existence and placement (e.g. /etc/passwd and /etc/hosts).
> Chris
> Johns provided a utility to generate the network stack related files
> from
> RTEMS configuration information so porting other net utilities is
> easier.
>
> > >Can RTEMS support rom filesystem??
> >
> > yes,.
> >
> > >I plan to put in NanoX in the future.. How should I go about placing it
> > >in and starting it up? (that's why I asked about the rom filesystem).
>
> The usual trick is to include a tar of the initial filesystem somehow
> in the program image and sue that to populate the RAM filesystem.
>
> Rosimildo.. do you have any advice on NanoX/MicroWindows?

I had some examples on my site, bu they are too old. Technically,
you do need a file system to run Nano-X, but you might need it
as soon as you need to store "Fonts", and some configuration.
On RTEMS, NanoX runs as a process linked along with the
rest of the system, so there is no concept of "client" and "server".

I personally prefer the Nano-X API, but if you have a team that
knows the Windows API ( SDK kind ), you better of with
MicroWindows.

Hopefully one day, I'll find time to get my old RTEMS PC from the
garage and update these demos, but the days are too short. :-)

ROsimildo.