[En-Nut-Discussion] Patch hanging: Set Idle Heartbeat

[Harald Kipp]

Hi Uwe,

On 17.05.2010 15:44, Uwe Bonnes wrote:
> again a patch sent to the mailing list and then revised to sourceforge is
> hanging for some time. Is there someting wrong with the patch? Should I
> proceed in another way?

Thanks for your patch and sorry for the delay.

However, I'm irritated. When anonymously loading the patch from SF, I
get an incomplete file. When trying to load it while logged in, I got a
"connection refused". Probably an SF server problem. Trying again and
again I may have got a complete patch file, but how to know?

What confuses me most is the comment "No longer use cbi/sbi". Both
patches look the same and do contain cbi/sbi. May be you can send your
patch to my email address to make sure I get the latest complete one?

One comment: Although used a lot in nutinit.c, "PORT = value" notation
should be avoided. See
http://www.ethernut.de/en/documents/programming-style-guide.html
But that wouldn't make a patch unacceptable. :-)

Btw. sbi/cbi works fine with memory mapped ports. The compiler will take
care to use the optimal assembly instructions.

Thanks again,

Harald




P.S.: For those, who are wondering what's the background of the
heartbeat implementation: My company manufactures an Ethernut 2 based
custom board, which contains an external hardware watchdog. If a
specific GPIO port is not toggled within a certain time, this hardware
will pull the reset line. While this gives extra safety for the final
application, it is almost impossible to run "normal" Nut/OS
applications. A quick solution was to add GPIO toggling to the idle thread.