[LITMUS^RT] issue with P-RES

Laurent Pautet laurent.pautet at telecom-paristech.fr
Sat Feb 22 12:24:08 CET 2020


Dear all,

I am testing P-RES with the following scenario

setsched Linux
setsched P-FP

resctl -n 100 -c 1 -t table-driven -m 100 '[0, 50)'
resctl -n 101 -c 1 -t table-driven -m 100 '[50, 100)'

rtspin -w -r 100 40 100 5 &
rtspin -w  -r 101 40 100 5 &

release_ts
wait

None of the tasks execute for its full timeslot (their budget is 40 and 
not 50).

However, in the actual scheduling, reservation 101 does not start at 
t=50 but at t=40 and that is not what was specified ('[50, 100)').

I understand that a workaround would be to force the task to execute for 
the full timeslot (by specifying a budget of 50).

But is there a wrong configuration in my kernel ? Is this a well-known 
issue ? If there is another workaround, I am interested in (offset ?)

Best regards,

-- 
--  Laurent

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.litmus-rt.org/pipermail/litmus-dev/attachments/20200222/d711141d/attachment.html>


More information about the litmus-dev mailing list