[En-Nut-Discussion] Nut/OS GPIO API Initial Design and Current Status

[Ulrich Prinz]

Hi Harald,

Am 14.10.2012 14:54, schrieb Harald Kipp:
> Hi Ulrich,
>
> On 13.10.2012 15:17, Ulrich Prinz wrote:
>
>> You have to see, that STM32 was a one man show and nobody gave any
>> complaint. So I had the option just to do it as I guess was the best
>> way.
>
> I think I did complain a little bit, but I'm not sure. I just had the
> "feeling", that something started to go wrong. Only after people in this
> list had discussions about the rationale of GPIO extensions, I
> discovered the misconception, that I'm complaining about now.
>
thiago and I agreed to a certain way of implementation and I think that 
I didn't break any compatibility with the GPIO thing. What could be is, 
that I added additional GPIO setup options in the wrong file. Setting an 
AVRs GPIO to PULLUP10K should not be possible and result in a compiler 
error.

>
> Off topic: In Nut/OS history, this happened quite often. The system
> timer code had been completely changed several times, until it reached
> its current state. Many years(!) later, we discovered, that Nut/OS
> context switching is extremely slow, compared to other OSes. Someone
> (was it Nathan?) posted, that this is caused by the last changes in
> timer processing. Note the present tense. It was never corrected,
> because the following discussion never came to any result.

Not coming to a result is one thing, never correcting it is bad option.
I know that context switching is slow and I'd like to see it going much 
more faster.

Context switching is ugly and it slows down the CortexM3 a lot. I will 
have a look into it again, when I start with a private project near future.

Ulrich