cpu/board: Initial import of samr21-xpro

[thomaseichinger]

This PR provides basic support for the samr21-xpro board including

• UART I/O
• hwtimer
• GPIO/EXIT

This PR is meant to provide a basic implementation for this board int RIOT's main repository to speed up distributed development and make the progress more visible. Development is coordinated via #1646 and documentation can be found in the wiki

Caution: Code will most certainly contain bugs.

This code also contains contributions of @Troels51.

[haukepetersen]

Travis fails for some test, can you blacklist them:

• net_if, pnet
• posix_semaphore
• pthread_rwlock
• queue_fairness
• thread_cooperation
• vtimer_msg_diff

[haukepetersen]

You could throw out these two lines

[haukepetersen]

Looking good, there are some minor things I see in some files, but I can't comment on them because github is not able to display them for too many changes in this PR... How about we merge this PR and I just create another one with some fixes?

So ACK when the comments above are addressed.

[haukepetersen]

still one unread varialbe problem in the uart driver says Travis, can you address this?

[thomaseichinger]

it is actually used, but inside the #ifdef UART_0_EN...

[haukepetersen]

ok. Then I guess we don't have to care about that one -> ACK

[N8Fear]

Travis fails because of unknown license headers: therefore NACK.

I'm not sure why cppcheck doesn't realize that the variable is used (it occurs for other boards at the same place, too). Maybe we should just suppress the warning at this place.

[Kijewski]

+ * 3. The name of Atmel may not be used to endorse or promote products derived
+ * from this software without specific prior written permission.
+ *
+ * 4. This software may only be redistributed and used in connection with an
+ * Atmel microcontroller product.

These sections are not GPL conforming. We decided to ignore this fact for header files, or didn't we?

[OlegHahm]

Only noticed now: Congrats to the PR's nice number. ;)

[thomaseichinger]

updated license pattern and reenabled some tests because of RAM savings due to reduction of stack sizes.

[thomaseichinger]

ok, rebased squashed and Travis is happy.
@N8Fear any further objections?
@OlegHahm stack sizes are fine with you?

[N8Fear]

@thomaseichinger: Travis isn't happy anymore (qemu-i386 fails, don't know how that relates to this PR).

[thomaseichinger]

Download server for qemu-i386 toolchain was down. Kicked Travis again.

[thomaseichinger]

ok Travis failure fixed with #1710 and #1711. Last run.

[N8Fear]

Yeah - just seen that. Maybe you can try to look what causes:
cpu/samd21/periph/uart.c:88: style (unreadVariable): Variable 'baud_calculated' is assigned a value that is never used.
and either fix or supress the warning.
Otherwise there are no further objections from me: ACK.

[thomaseichinger]

@N8Fear This implementation will see a major refactory with a more generic implementation soon while this PR should provide the a starting point to simplify contributions by various people. If your are ok with it, I'd leave it as is for now since the variable is actually used. Getting it bug free and beautiful is the job of upcoming PRs.

[N8Fear]

I don't see it as a show stopper... ;-)
I just like to reduce the warnings cppcheck throws with the goal to be able to enable it in a way that causes hard fails instead of the current "warnings" (exitcode 0 -> 1).

[OlegHahm]

Works for me with default and ipc_pingpong. I updated the Wiki page.

ACK

[thomaseichinger]

So let's do this.