diff mbox series

rcu: Remove impossible wakeup rcu GP kthread action from rcu_report_qs_rdp()

Message ID 20230117021955.1967316-1-qiang1.zhang@intel.com
State New
Headers
Series rcu: Remove impossible wakeup rcu GP kthread action from rcu_report_qs_rdp() |

Commit Message

Zqiang Jan. 17, 2023, 2:19 a.m. UTC 
  When inovke rcu_report_qs_rdp(), if current CPU's rcu_data structure's ->
grpmask has not been cleared from the corresponding rcu_node structure's
->qsmask, after that will clear and report quiescent state, but in this
time, this also means that current grace period is not end, the current
grace period is ongoing, because the rcu_gp_in_progress() currently return
true, so for non-offloaded rdp, invoke rcu_accelerate_cbs() is impossible
to return true.

This commit therefore remove impossible rcu_gp_kthread_wake() calling.

Signed-off-by: Zqiang <qiang1.zhang@intel.com>
---
 kernel/rcu/tree.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

Comments

Paul E. McKenney Jan. 17, 2023, 4:32 a.m. UTC | #1 
On Tue, Jan 17, 2023 at 10:19:55AM +0800, Zqiang wrote:
> When inovke rcu_report_qs_rdp(), if current CPU's rcu_data structure's ->
> grpmask has not been cleared from the corresponding rcu_node structure's
> ->qsmask, after that will clear and report quiescent state, but in this
> time, this also means that current grace period is not end, the current
> grace period is ongoing, because the rcu_gp_in_progress() currently return
> true, so for non-offloaded rdp, invoke rcu_accelerate_cbs() is impossible
> to return true.
> 
> This commit therefore remove impossible rcu_gp_kthread_wake() calling.
> 
> Signed-off-by: Zqiang <qiang1.zhang@intel.com>
> ---
>  kernel/rcu/tree.c | 5 +----
>  1 file changed, 1 insertion(+), 4 deletions(-)
> 
> diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
> index b2c204529478..477eb1a374e5 100644
> --- a/kernel/rcu/tree.c
> +++ b/kernel/rcu/tree.c
> @@ -1956,7 +1956,6 @@ rcu_report_qs_rdp(struct rcu_data *rdp)
>  {
>  	unsigned long flags;
>  	unsigned long mask;
> -	bool needwake = false;
>  	bool needacc = false;
>  	struct rcu_node *rnp;
>  
> @@ -1988,7 +1987,7 @@ rcu_report_qs_rdp(struct rcu_data *rdp)
>  		 * NOCB kthreads have their own way to deal with that...
>  		 */
>  		if (!rcu_rdp_is_offloaded(rdp)) {
> -			needwake = rcu_accelerate_cbs(rnp, rdp);
> +			rcu_accelerate_cbs(rnp, rdp);

If it is impossible, we should use WARN_ON_ONCE() or similar.  Just
in case the system disagrees on the impossibility.  ;-)

							Thanx, Paul

>  		} else if (!rcu_segcblist_completely_offloaded(&rdp->cblist)) {
>  			/*
>  			 * ...but NOCB kthreads may miss or delay callbacks acceleration
> @@ -2000,8 +1999,6 @@ rcu_report_qs_rdp(struct rcu_data *rdp)
>  		rcu_disable_urgency_upon_qs(rdp);
>  		rcu_report_qs_rnp(mask, rnp, rnp->gp_seq, flags);
>  		/* ^^^ Released rnp->lock */
> -		if (needwake)
> -			rcu_gp_kthread_wake();
>  
>  		if (needacc) {
>  			rcu_nocb_lock_irqsave(rdp, flags);
> -- 
> 2.25.1
>
Zqiang Jan. 17, 2023, 7:34 a.m. UTC | #2 
On Tue, Jan 17, 2023 at 10:19:55AM +0800, Zqiang wrote:
> When inovke rcu_report_qs_rdp(), if current CPU's rcu_data structure's ->
> grpmask has not been cleared from the corresponding rcu_node structure's
> ->qsmask, after that will clear and report quiescent state, but in this
> time, this also means that current grace period is not end, the current
> grace period is ongoing, because the rcu_gp_in_progress() currently return
> true, so for non-offloaded rdp, invoke rcu_accelerate_cbs() is impossible
> to return true.
> 
> This commit therefore remove impossible rcu_gp_kthread_wake() calling.
> 
> Signed-off-by: Zqiang <qiang1.zhang@intel.com>
> ---
>  kernel/rcu/tree.c | 5 +----
>  1 file changed, 1 insertion(+), 4 deletions(-)
> 
> diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
> index b2c204529478..477eb1a374e5 100644
> --- a/kernel/rcu/tree.c
> +++ b/kernel/rcu/tree.c
> @@ -1956,7 +1956,6 @@ rcu_report_qs_rdp(struct rcu_data *rdp)
>  {
>  	unsigned long flags;
>  	unsigned long mask;
> -	bool needwake = false;
>  	bool needacc = false;
>  	struct rcu_node *rnp;
>  
> @@ -1988,7 +1987,7 @@ rcu_report_qs_rdp(struct rcu_data *rdp)
>  		 * NOCB kthreads have their own way to deal with that...
>  		 */
>  		if (!rcu_rdp_is_offloaded(rdp)) {
> -			needwake = rcu_accelerate_cbs(rnp, rdp);
> +			rcu_accelerate_cbs(rnp, rdp);
>
>If it is impossible, we should use WARN_ON_ONCE() or similar.  Just
>in case the system disagrees on the impossibility.  ;-)

Thanks for suggestion, I will resend v2 😊.

>
>							Thanx, Paul

>  		} else if (!rcu_segcblist_completely_offloaded(&rdp->cblist)) {
>  			/*
>  			 * ...but NOCB kthreads may miss or delay callbacks acceleration
> @@ -2000,8 +1999,6 @@ rcu_report_qs_rdp(struct rcu_data *rdp)
>  		rcu_disable_urgency_upon_qs(rdp);
>  		rcu_report_qs_rnp(mask, rnp, rnp->gp_seq, flags);
>  		/* ^^^ Released rnp->lock */
> -		if (needwake)
> -			rcu_gp_kthread_wake();
>  
>  		if (needacc) {
>  			rcu_nocb_lock_irqsave(rdp, flags);
> -- 
> 2.25.1
>
diff mbox series

Patch

diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
index b2c204529478..477eb1a374e5 100644
--- a/kernel/rcu/tree.c
+++ b/kernel/rcu/tree.c
@@ -1956,7 +1956,6 @@  rcu_report_qs_rdp(struct rcu_data *rdp)
 {
 	unsigned long flags;
 	unsigned long mask;
-	bool needwake = false;
 	bool needacc = false;
 	struct rcu_node *rnp;
 
@@ -1988,7 +1987,7 @@  rcu_report_qs_rdp(struct rcu_data *rdp)
 		 * NOCB kthreads have their own way to deal with that...
 		 */
 		if (!rcu_rdp_is_offloaded(rdp)) {
-			needwake = rcu_accelerate_cbs(rnp, rdp);
+			rcu_accelerate_cbs(rnp, rdp);
 		} else if (!rcu_segcblist_completely_offloaded(&rdp->cblist)) {
 			/*
 			 * ...but NOCB kthreads may miss or delay callbacks acceleration
@@ -2000,8 +1999,6 @@  rcu_report_qs_rdp(struct rcu_data *rdp)
 		rcu_disable_urgency_upon_qs(rdp);
 		rcu_report_qs_rnp(mask, rnp, rnp->gp_seq, flags);
 		/* ^^^ Released rnp->lock */
-		if (needwake)
-			rcu_gp_kthread_wake();
 
 		if (needacc) {
 			rcu_nocb_lock_irqsave(rdp, flags);