[En-Nut-Discussion] Ethernut newbie with a lot of questions

[Greg Wood]

Hello everyone,
 
I got an ethernut board (v1.3) a while ago and have just now gotten
around to creating some useful applications with it.  In only a few
hours I was able to get a simple application up and running.  I think
the creators of ethernut have done a fantastic job of making an easy to
use system.  The work I have done has left me with more questions than
when I started.  
 
Assuming the ethernut board is run behind a switch and not a hub how
much processor load does the stack and OS exert?  Suppose that a simple
web server and maybe a telnet session are running.  Is there a time
limit that should be observed when a thread is blocking?  What is
reasonable number of simultaneous connected sockets?
 
What are the consequences if ethernut is run behind a hub and not a
switch?  
 
What is the best way to implement ISRs?  I'm done a lot of programming
with WINAVR, can I still use SIGNAL functions?  Are there any special
considerations when doing this?  Is there any new delays created in
responding to an ISR from the OS/stack?
 
Are the streaming functions, specifically printf, interrupt based when
used with the uart?
 
I'm somewhat new to real time OSes.  Can someone point me to a good
source if information about programming with Nut/OS?  I've read the
ethernut software manual and it left me wanting more.  The nut/API
reference does not have sufficient explanation for some things to make
sense to me.  I'd like to see some example programs that take advantage
of some of the features availabe in Nut/Os.  In the past I've used
global variables to pass info in and out of ISRs, what is the prefered
method to do this in Nut/Os?  
 
I've got an application in mind in which several ethernet devices will
exist on a robot and be connected via an ethernet radio to a LAN.  I'd
like to have the oem radio module mounted atop my electronics which will
likely be an ethernut compatible setup.  I want to avoid having to go
radio->magnetics->switch<-magenets<-ethernut.  It seems to me that if
everything will be build onto a PCB there is no need for magnetics.  Is
there a chipset that implements a simple 4 or 5 port 10mbps switch that
would allow me to avoid the magnetics for the devices that are on the
pcb together, and also build the switch onto the PCB?
 
Is there some software out there that will allow me to send information
to specific IPs/ports via TCP and/or UDP?  Basically the equivalent of a
serial terminal program that supports binary data.  For serial
applications I've used windmill and brays terminal.  
 
Is it possible to port ethernut over to other members of the AVR family
such as the new fangled 100 pin devices?
 
Is there a way to manually read/send files from urom.c?  I wrote a
simple CGI and it seems to me like it would be nice if you could just
send out a file to fill in parts of a CGI page.
 
I'd like to see some ethernut applications.  If someone could post some
links to ethernut based projects that would be great.  
 
Anyone written a DNP3 driver for ethernut?
 
Anyone interfaced an ethernut 1.3 board to a SD/MMC card?  How about
other storage devices?  
 
How about an ethernet based bootloader?  Ponyprog is painfully slow.
How can I use my STK500 compatible programmer with my ethernut board?
 
Can Nut/Os be run on a stripped down system?  I.e. just the ATMEGA128?
I'm thinking at least the external SRAM is required.
 
Sorry for all the questions.  I'm sure some of them are answered within
the ethernut documentation.  If you can refer me to certain documents
thats great.  I've got a lot to learn about ethernut and ethernet and
have gotten a little tired of reading and searching for information the
last 3 days.  
 
Regards,
 
Greg Wood
 
 

-- 
Internal Virus Database is out-of-date.
Checked by AVG Free Edition.
Version: 7.1.375 / Virus Database: 268.2.0/276 - Release Date:
07/03/2006