[En-Nut-Discussion] Porting Nut/OS to atmel's SAM3U (cortex-m3)

[Curtis Maloney]

On 04/09/10 01:49, Harald Kipp wrote:
> Hi all,
>
> On 01.02.2010 10:29, Nikolay Zamotaev wrote:
>
>> I posted patch done against svn revision r2881
>> https://sourceforge.net/tracker/?func=detail&atid=410689&aid=2943586&group_id=34079
>
> This patch had been posted about 2 months ago. As far as I can see,
> nobody took care.

> Cons:

> * I assume, that it is not very well tested by anyone except the author
> himself.
> * Even Nikolay himself admits, that it may be partly buggy.
> * In the past I often applied patches without further checking and later
> it took months for discovering and removing the bugs (e.g. register
> usage of SAM7 context switching).

Isn't this the joy of open source?  Once it's there where everyone can 
easily get at it, we get a LOT more eyes on the problem.

I know my employer, for one, is eager to run Nut/OS on some Cortex-M3 kit...

> * I'm working towards a next stable release and don't like to have big
> changes right now.
> Pros:
> * The patch contains very few changes on existing files.
> * We may declare Cortex support as beta, even in a stable release.

The sooner it's rolled in, the sooner it gets more testing, and the 
sooner it becomes "release" quality.

> What's your opinion?
>
> What worries me most is, that more changes on the trunk may make the
> patch unusable. Because of its size, it may later require a lot of
> effort to fix this.

As you say, it doesn't change many existing files, meaning it should 
have little to no impact on others.  Since the M3 is such a rigidly 
defined platform, supporting it for one vendor gets you the majority of 
the work for supporting all vendors.

> Last not least: Respect, Nikolay!

I'll second that!

--
Curtis Maloney