Re: Unexpected switches to in-band

["Łukasz Majewski" <lukma@nabladev.com>]

HI Giulio,

> > 2. After the "first" occurrence of the in-band switch in latmus,
> > when it is started again, the issue cannot be reproduced any more.
> > The conclusion here is that the issue is with "startup" or "fresh"
> > configuration of memory subsystem with dovetail  
> 
> Do you mean that only the first instance of latmus that is started
> after a reboot is affected by this issue, while successive instances
> are not?

Yes. That is what I do observe.

Just to mention - I do run latmus only on cpu2:
latmus -c 2 -p 1000 -P 99

Test programs are run on other cores as well

> 
> > With guidance (and huge help) from Phillipe I'm now ftracing the
> > evl core with latmus to check what can cause the issue.  
> 
> Would you be able to explain your setup here to see if I can
> reproduce?

Basically I do run in one ssh connected console:

evl trace -e 2 
latmus -c 2 -p 1000 -P 99
...
...



On the other ssh connected console:
<test_program_10GiB_vector> &
sleep 5;
<test_program_10GiB_vector> &

Then in latmus console I do see the switch to in-band notification.

From dmesg also I can read:
[   94.685063] EVL: timer-responder:745 switching in-band [pid=747,
excpt=14, user_pc=0x797b51227f20]


And te 94.685063 timestamp is crucial here. Unfortunately the ftrace
output doesn't match with it - i.e. there is no operation "around" this
time (there are ones 70 ms before and ~20 ms after).
This I'm investigating now.

> 
> Thanks,
> Giulio



-- 
Best regards,

Lukasz Majewski

--
Nabla Software Engineering GmbH
HRB 40522 Augsburg
Phone: +49 821 45592596
E-Mail: office@nabladev.com
Geschftsfhrer : Stefano Babic