[riot-devel] Timers (caution: long mail, abstract available)

Oleg Hahm oliver.hahm at inria.fr
Wed Sep 17 10:53:46 CEST 2014

Dear Abhinav,

> Thank you for bringing this topic up. I would like to add a couple of
> comments to your email as a user or RIOT without having contributed to
> the development.

thanks for your input. Feedback from your perspective, i.e., someone who's not
"living" inside RIOT for a loooong time, is highly appreciated and very
> Firstly, I found the timer implementation in RIOT one of the most
> confusing part and it took me a while to figure out how to use the
> different timers. So your suggestion of a more straightforward timer
> implementation would be much easier to start using.

Probably even a better documentation of the current implementation could be an
enhancement of the situation.

> Secondly, there is still an issue with timers that I am unable to solve.
> I see that you do not want to use timers to schedule tasks because of
> your tickles scheduler concept. However, there are some low-level
> peripherals drivers that still need timers to schedule certain events.
> For example I am currently using the HAL library from STM and several
> peripherals (e.g I2C) uses timeouts from within interrupt routines to
> stop a peripheral if it does not get a response from a peer device.
> Therefore, for such peripheral drivers, it is essential to get access to
> a timer of some kind that has a higher interrupt priority level then all
> other interrupts (e.g. the SYSTICK). 

Interestingly that's exactly what I've just discussed with a colleague some
minutes ago. The problem with interrupt priority levels is that not all
architectures support this. But I agree that we have to find a (generic)
solution for this.

fs_dprintk (FS_DEBUG_INIT, "Ha! Initialized OK!\n");
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://lists.riot-os.org/pipermail/devel/attachments/20140917/376928f1/attachment-0001.sig>

More information about the devel mailing list