Re: [PATCH 11/13] rcu/nocb: Only cancel nocb timer if not polling

2021-03-10 Thread Frederic Weisbecker
On Tue, Mar 02, 2021 at 05:22:29PM -0800, Paul E. McKenney wrote:
> On Tue, Feb 23, 2021 at 01:10:09AM +0100, Frederic Weisbecker wrote:
> > No need to disarm the nocb_timer if rcu_nocb is polling because it
> > shouldn't be armed either.
> > 
> > Signed-off-by: Frederic Weisbecker 
> > Cc: Josh Triplett 
> > Cc: Lai Jiangshan 
> > Cc: Joel Fernandes 
> > Cc: Neeraj Upadhyay 
> > Cc: Boqun Feng 
> 
> OK, so it does make sense to move that del_timer() under the following
> "if" statement, then.  ;-)

Right, probably I should have handled that in the beginning of the patchset
instead of the end but heh, my mind is never that clear.

Thanks.


Re: [PATCH 11/13] rcu/nocb: Only cancel nocb timer if not polling

2021-03-03 Thread Paul E. McKenney
On Tue, Feb 23, 2021 at 01:10:09AM +0100, Frederic Weisbecker wrote:
> No need to disarm the nocb_timer if rcu_nocb is polling because it
> shouldn't be armed either.
> 
> Signed-off-by: Frederic Weisbecker 
> Cc: Josh Triplett 
> Cc: Lai Jiangshan 
> Cc: Joel Fernandes 
> Cc: Neeraj Upadhyay 
> Cc: Boqun Feng 

OK, so it does make sense to move that del_timer() under the following
"if" statement, then.  ;-)

> ---
>  kernel/rcu/tree_plugin.h | 14 +++---
>  1 file changed, 7 insertions(+), 7 deletions(-)
> 
> diff --git a/kernel/rcu/tree_plugin.h b/kernel/rcu/tree_plugin.h
> index 9da67b0d3997..d8b50ff40e4b 100644
> --- a/kernel/rcu/tree_plugin.h
> +++ b/kernel/rcu/tree_plugin.h
> @@ -2186,18 +2186,18 @@ static void nocb_gp_wait(struct rcu_data *my_rdp)
>   my_rdp->nocb_gp_gp = needwait_gp;
>   my_rdp->nocb_gp_seq = needwait_gp ? wait_gp_seq : 0;
>   if (bypass) {
> - raw_spin_lock_irqsave(_rdp->nocb_gp_lock, flags);
> - // Avoid race with first bypass CB.
> - if (my_rdp->nocb_defer_wakeup > RCU_NOCB_WAKE_NOT) {
> - WRITE_ONCE(my_rdp->nocb_defer_wakeup, 
> RCU_NOCB_WAKE_NOT);
> - del_timer(_rdp->nocb_timer);
> - }
>   if (!rcu_nocb_poll) {
> + raw_spin_lock_irqsave(_rdp->nocb_gp_lock, flags);
> + // Avoid race with first bypass CB.
> + if (my_rdp->nocb_defer_wakeup > RCU_NOCB_WAKE_NOT) {
> + WRITE_ONCE(my_rdp->nocb_defer_wakeup, 
> RCU_NOCB_WAKE_NOT);
> + del_timer(_rdp->nocb_timer);
> + }
>   // At least one child with non-empty ->nocb_bypass, so 
> set
>   // timer in order to avoid stranding its callbacks.
>   mod_timer(_rdp->nocb_bypass_timer, j + 2);
> + raw_spin_unlock_irqrestore(_rdp->nocb_gp_lock, 
> flags);
>   }
> - raw_spin_unlock_irqrestore(_rdp->nocb_gp_lock, flags);
>   }
>   if (rcu_nocb_poll) {
>   /* Polling, so trace if first poll in the series. */
> -- 
> 2.25.1
> 


[PATCH 11/13] rcu/nocb: Only cancel nocb timer if not polling

2021-02-22 Thread Frederic Weisbecker
No need to disarm the nocb_timer if rcu_nocb is polling because it
shouldn't be armed either.

Signed-off-by: Frederic Weisbecker 
Cc: Josh Triplett 
Cc: Lai Jiangshan 
Cc: Joel Fernandes 
Cc: Neeraj Upadhyay 
Cc: Boqun Feng 
---
 kernel/rcu/tree_plugin.h | 14 +++---
 1 file changed, 7 insertions(+), 7 deletions(-)

diff --git a/kernel/rcu/tree_plugin.h b/kernel/rcu/tree_plugin.h
index 9da67b0d3997..d8b50ff40e4b 100644
--- a/kernel/rcu/tree_plugin.h
+++ b/kernel/rcu/tree_plugin.h
@@ -2186,18 +2186,18 @@ static void nocb_gp_wait(struct rcu_data *my_rdp)
my_rdp->nocb_gp_gp = needwait_gp;
my_rdp->nocb_gp_seq = needwait_gp ? wait_gp_seq : 0;
if (bypass) {
-   raw_spin_lock_irqsave(_rdp->nocb_gp_lock, flags);
-   // Avoid race with first bypass CB.
-   if (my_rdp->nocb_defer_wakeup > RCU_NOCB_WAKE_NOT) {
-   WRITE_ONCE(my_rdp->nocb_defer_wakeup, 
RCU_NOCB_WAKE_NOT);
-   del_timer(_rdp->nocb_timer);
-   }
if (!rcu_nocb_poll) {
+   raw_spin_lock_irqsave(_rdp->nocb_gp_lock, flags);
+   // Avoid race with first bypass CB.
+   if (my_rdp->nocb_defer_wakeup > RCU_NOCB_WAKE_NOT) {
+   WRITE_ONCE(my_rdp->nocb_defer_wakeup, 
RCU_NOCB_WAKE_NOT);
+   del_timer(_rdp->nocb_timer);
+   }
// At least one child with non-empty ->nocb_bypass, so 
set
// timer in order to avoid stranding its callbacks.
mod_timer(_rdp->nocb_bypass_timer, j + 2);
+   raw_spin_unlock_irqrestore(_rdp->nocb_gp_lock, 
flags);
}
-   raw_spin_unlock_irqrestore(_rdp->nocb_gp_lock, flags);
}
if (rcu_nocb_poll) {
/* Polling, so trace if first poll in the series. */
-- 
2.25.1