[LITMUS^RT] How to stop all the rt-tasks manually

[Sanjib Das]

Dear Glenn and Björn,

Thank you for the suggestions. Process actually turns into an
uninterruptible state.
Actually I am trying to simulate with Experiment scripts from Jonathan
Herman

https://github.com/hermanjl/experiment-scripts


I am running the script on actual machine, 4 cores.
Generated task file of P-EDF and C-EDF runs without stalling up to 2500
tasks with 2 cpus.
./gen_exps.py -f tasks=2500 cpus=2 P-EDF, release_mater=default
./gen_exps.py -f tasks=2500 cpus=2 C-EDF , release_mater=default  where
cluster was default=L2

With both of them if 5000 tasks is generated to run CPU seems stopped
working. (Even I can not open terminal, may be its was still running)

But wherein ,G-EDF only working with release_master=False.
even with tasks ./gen_exps.py -f tasks=500 cpus=2 G-EDF, ran execution
stalls.

Suggestions about task generation and also task execution and assumption of
execution time is badly necessary.

Thanks in advance
Sanjib



On Sun, Feb 9, 2014 at 3:38 PM, Glenn Elliott <gelliott at cs.unc.edu> wrote:

>
> On Feb 9, 2014, at 8:57 AM, Sanjib Das <cnt.sanjib at googlemail.com> wrote:
>
> > Hello Björn,
> >
> > No i haven't tried with killall.
> > what will be parameter for the killall command,
> > killall -g rtspin doesnt work
> > also
> > killall -g rcu_preempt
> >
> > how should I try it ?
> >
> > Thanks in advance
> >
> > Sanjib
> > On 09.02.2014 13:54, Björn Brandenburg wrote:
> >> On 09 Feb 2014, at 05:11, Sanjib Das <cnt.sanjib at googlemail.com> wrote:
> >>
> >>> Dear Concern,
> >>>
> >>> I am wondering if there is any way to kill all the running real-time
> task manually.
> >> Have you tried killall?
> >>
> >> - Björn
> >>
> >>
> >>> I am trying to run G-EDF and C-EDF with different tasks and because of
> wrong parameter I found "rcu_preempt detected stalls on cpus/tasks", as a
> result my rt tasks were still running.
> >>>
> >>> "
> >>> [ 2072.588370] Could not switch plugin: -16
> >>> [ 2107.177026] Could not switch plugin: -16
> >>> [ 2120.735386] Could not switch plugin: -16
> >>> [ 2131.718955] Could not switch plugin: -16
> >>> [ 2173.427350] INFO: rcu_preempt detected stalls on CPUs/tasks: {}
> (detected by 1, t=600017 jiffies, g=53906, c=53905, q=0)
> >>> [ 2173.427371] INFO: Stall ended before state dump start
> >>> [ 2189.814720] Could not switch plugin: -16
> >>> [ 2353.473222] INFO: rcu_preempt detected stalls on CPUs/tasks: {}
> (detected by 1, t=780022 jiffies, g=53906, c=53905, q=0)
> >>> [ 2353.473241] INFO: Stall ended before state dump start
> >>> [ 2533.519115] INFO: rcu_preempt detected stalls on CPUs/tasks: {}
> (detected by 1, t=960027 jiffies, g=53906, c=53905, q=0)
> >>> [ 2533.519131] INFO: Stall ended before state dump start
> >>>
> >>> "
> >>>
> >>>
> >>> Though I ran the command using sudo,
> >>> "
> >>> Error: Cannot write to /proc/litmus/active_plugin. Do you have root
> privileges?
> >>> Error: Setting new plugin failed.
> >>> Are there any active real-time tasks?
> >>> sdas at sdas-lab:~/littest/liblitmus$
> >>> "
> >>>
> >>> Any suggestions will be very helpful.
> >>>
> >>> Thanks in advance
>
>
> "sudo pkill -9 rtspin" should work, unless the tasks are stuck in an
> uninterruptible state.  If that pkill command won't work, then I think you
> have no choice but to reboot or even a hard reset.  You shouldn't kill
> rcu_preempt--it's a kernel demon.
>
> -Glenn
> _______________________________________________
> litmus-dev mailing list
> litmus-dev at lists.litmus-rt.org
> https://lists.litmus-rt.org/listinfo/litmus-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.litmus-rt.org/pipermail/litmus-dev/attachments/20140210/1b21460f/attachment.html>