iot-lab_M3: definitions for implementation of peripheral timer driver

[thomaseichinger]

accompanies PR OlegHahm/thirdparty_cpu#26 by providing definitions for the peripheral timer driver.

[LudwigKnuepfer]

Is this in a working condition?

[thomaseichinger]

yes it is

[LudwigKnuepfer]

@authmillenon can you please review?

[LudwigKnuepfer]

@authmillenon (I saw you had the hardware to test this)

[miri64]

@thomaseichinger which branches do I need of what repository?

[thomaseichinger]

@authmillenon you'll need
thirdparty_boards:iot-lab_M3-periph_timer_driver_defs
thirdparty_cpu:periph_timer

[miri64]

Maybe not related to this PR, but:

RIOTBOARD=~/Repositories/RIOT-OS/thirdparty_boards/ RIOTCPU=~/Repositories/RIOT-OS/thirdparty_cpu/ BOARD=iot-lab_M3 make -C ~/Repositories/RIOT-OS/RIOT/tests/test_hwtimer -B clean all flash term

make: Entering directory `/home/martine/Repositories/RIOT-OS/RIOT/tests/test_hwtimer'
rm -rf /home/martine/Repositories/RIOT-OS/RIOT/tests/test_hwtimer/bin/iot-lab_M3/

Compiling.... main.c

Building application test_hwtimer for iot-lab_M3 w/ MCU stm32f103rey6.
[…]
make[3]: Entering directory `/home/martine/Repositories/RIOT-OS/thirdparty_cpu/cortex_common'
crash.c: In function 'core_panic':
crash.c:64:1: warning: 'noreturn' function does return [enabled by default]
 }
 ^
make[3]: Leaving directory `/home/martine/Repositories/RIOT-OS/thirdparty_cpu/cortex_common'
[…]
make[1]: Leaving directory `/home/martine/Repositories/RIOT-OS/RIOT'
arm-none-eabi-gcc: error: /home/martine/Repositories/RIOT-OS/RIOT/tests/test_hwtimer/bin/iot-lab_M3/syscalls.o: No such file or directory
make: *** [all] Error 1
make: Leaving directory `/home/martine/Repositories/RIOT-OS/RIOT/tests/test_hwtimer'

[thomaseichinger]

hm works for me, I rebased both branches could you try again?

[thomaseichinger]

ok, now it is hard faulting, I'll investigate and come back to you

[thomaseichinger]

@authmillenon You will need to add

puts("hwtimer set.");
+while(1);
return 0;

to test_hwtimer/main.c because thread exit seems to be still broken on stm32f103rey6

[haukepetersen]

The crash on return of main is known, it's fixed in the cortexm-common code in master...

[thomaseichinger]

@authmillenon could you try again? Don't know why jenkins fails.

[miri64]

It seems to fail for all thirdparty_cpu PRs (it can't find RIOTBASE in RIOTCPU for some reason…). I have to look into it.

[miri64]

Still cannot find syscalls.o Nevermind, I just forgot to rebase thirdparty_cpu

[miri64]

Cannot flash it (since my udev is not configured correctly anymore… would be nice if we could document this somewhere), but at least all tests/test_hwtimer*/ are building except with a few warnings:

[…]
make[2]: Entering directory `/home/martine/Repositories/RIOT-OS/thirdparty_cpu/stm32f103rey6'
cpu.c: In function 'cpu_switch_context_exit':
cpu.c:84:1: warning: 'noreturn' function does return [enabled by default]
 }
 ^
[…]
make[3]: Entering directory `/home/martine/Repositories/RIOT-OS/thirdparty_cpu/cortex_common'
crash.c: In function 'core_panic':
crash.c:64:1: warning: 'noreturn' function does return [enabled by default]
 }
 ^
[…]

[miri64]

Sorry, hit the wrong button.

[thomaseichinger]

what's the problem with udev? which kind of configuration does it need?

[miri64]

Sorry, was not a udev problem: see OlegHahm/thirdparty_cpu#32

[miri64]

Appart from the afformentioned problem with crash on return (which is not part of this PR imho), all test applications run.

[miri64]

Docstring?

[haukepetersen]

actually no, these comments were only meant for better readability of the code and not really for doxygen...

[miri64]

If above comments are addressed, I will give my ACK :)

[thomaseichinger]

updated addressing @authmillenon's comments

[miri64]

ACK, squash and go.

[thomaseichinger]

squashed will merge when @haukepetersen approves OlegHahm/thirdparty_cpu#26

[haukepetersen]

But a copyright + doxygen header wouldn't hurt here!

[haukepetersen]

Conditional ACK if you add the doxygen and copyright headers.

[thomaseichinger]

updated addressing @haukepetersen's comments

[haukepetersen]

looking good. ACK

[thomaseichinger]

reworded second commit with tag and GO