[En-Nut-Discussion] devnut_m3n branch

[Ulrich Prinz]

Hi Uwe,

> Yes, I need to use -mcm3-gcc in nutconf
> 
> ...
>     uprinz> Third: I 99% of my time developed on STM3210E-EVAL and
>     uprinz> STM3210C-EVAL kits. I used the COMsticks only very little. If
>     uprinz> you look at the date of the .conf files you see, what
>     uprinz> happened. I have a couple of these neat sticks so I can put one
>     uprinz> to my station and update the configuration for you.
> 
> Yes, nut/conf/stm32-comStick.conf is outdated
I'll fix that right after I found the sticks. I moved to another room in
my house...
> 
> Using the right compiler (-mcm3-gcc) and the right conf file
> (nut/conf/stm3210c-eval.conf) looks much better.
> 
Happy to hear that!

> However either I have to remove "-Wstrict-prototypes -Werror" from
> nut/Makevars.cm3-gcc to get around the strict aliasing warnings. Or I
> rewrite nut/arch/cm3/cmsis/cortex_init.c as in the patch below. I think,
> this is similar to the "AT (ldadr)" section in 
> http://ftp.gnu.org/old-gnu/Manuals/ld-2.9.1/html_node/ld_21.html
> Please bear with me if the cortex_init patch is totaly bogus!

Ok, I studied lots of linker docs to get the one running for Cortex.
This is something for one who killed his Mother... I don't know why
gcc-4.5.0 and gcc-4.6.0 do not claim about that while your 4.4.x version
does. But I'll apply the patch and keep it if it doesn't break anything.
> 
> Furthermore nut/app/led_key/pbtest.c has some bad line.
> 
> With the changes applied, I can compile the library and the examples.
> 
> Thanks so long!
> 
Please let me know if something runs on the stick and keep me updated on
your progress.

And thanks for your patience.

Ulrich