Re: Unexpected switches to in-band

[Philippe Gerum <rpm@xenomai.org>]

Giulio Moro <giulio@bela.io> writes:

>> Quick check: does "evl check" run without any output (as expected)?
>
> Yes
>
> Further update: tested v6.16-evl-rebase and it seems to have the same in-band switching issue. So far the only good one was v6.1.y-cip-evl-rebase.
>
>> Meanwhile, I've been trying to reproduce on x86 and armv8 (odroid c4) using the posted recipe, no luck so far.
>
> Thanks for giving it a try. Wanna have a look at/try my config? https://github.com/giuliomoro/linux/blob/v6.12.y-evl-rebase/.config

I could not reproduce this issue yet, neither on x86 nor arm64, trying a
bunch of configurations of linux-evl/v6.12.49. At this point, the best
approach would be to track down the latency spot on the TI board using
the 'latspot' (f)tracepoint which latmus triggers each time a new
latency maximum is reached, so that we might find out the reason for
taking those minor faults which are causing in-band switches. There is
some documentation about its usage at [1].

[1] https://v4.xenomai.org/core/commands/index.html

-- 
Philippe.