Hi... On Wed, Nov 24, 2010 at 14:56, Parmenides <mobile.parmenides@xxxxxxxxx> wrote: > Hi Santosa, > > My confusion is clarified by your reply. Furthermore, there are some > details that deserves more considerations. Glad I could offer some humble help... > 2010/11/24 Mulyadi Santosa <mulyadi.santosa@xxxxxxxxx>: ... > we can see that only if the next process is a normal process and in > TASK_INTERRUPTIBLE or TASK_STOOPED state, the recalc_task_prio() get > its running to update the dynamic priority of next. IMHO you already found the right code snippet to prove just that...congrats.. > But, for another kernel control path, namely try_to_wake_up() -> > activate_task() -> recalc_task_prio(), it seems that there is no any > conditions set in this path to ensure that a real-time process' > dynamic priority can not be updated. Of course, there may be some > other kernel control path neglected by me. If so, the goal to maintain > the dynamic priority fixed for a real-time process can not be achieved > readily. > I found this function "rt_policy" in kernel/sched.c. I think you can trace back on which functions that call this. I believe, this macro will lead you on code snippet that might prove that for real time process, time slice/priority is always the same. -- regards, Mulyadi Santosa Freelance Linux trainer and consultant blog: the-hydra.blogspot.com training: mulyaditraining.blogspot.com -- To unsubscribe from this list: send an email with "unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx Please read the FAQ at http://kernelnewbies.org/FAQ
