Re: [PATCH] tiny_rcu: directly force QS when call_rcu_[bh|sched]() on idle_task

2014-12-09 Thread Paul E. McKenney
On Tue, Dec 09, 2014 at 05:53:34PM +0800, Lai Jiangshan wrote: > For RCU in UP, context-switch = QS = GP, thus we can force a > context-switch when any call_rcu_[bh|sched]() is happened on idle_task. > After doing so, rcu_idle/irq_enter/exit() are useless, so we can simply > make these functions

[PATCH] tiny_rcu: directly force QS when call_rcu_[bh|sched]() on idle_task

2014-12-09 Thread Lai Jiangshan
For RCU in UP, context-switch = QS = GP, thus we can force a context-switch when any call_rcu_[bh|sched]() is happened on idle_task. After doing so, rcu_idle/irq_enter/exit() are useless, so we can simply make these functions empty. More important, this change does not change the functionality

[PATCH] tiny_rcu: directly force QS when call_rcu_[bh|sched]() on idle_task

2014-12-09 Thread Lai Jiangshan
For RCU in UP, context-switch = QS = GP, thus we can force a context-switch when any call_rcu_[bh|sched]() is happened on idle_task. After doing so, rcu_idle/irq_enter/exit() are useless, so we can simply make these functions empty. More important, this change does not change the functionality

Re: [PATCH] tiny_rcu: directly force QS when call_rcu_[bh|sched]() on idle_task

2014-12-09 Thread Paul E. McKenney
On Tue, Dec 09, 2014 at 05:53:34PM +0800, Lai Jiangshan wrote: For RCU in UP, context-switch = QS = GP, thus we can force a context-switch when any call_rcu_[bh|sched]() is happened on idle_task. After doing so, rcu_idle/irq_enter/exit() are useless, so we can simply make these functions