The sysrq_handle_sched_ext_reset() callback has already been
executed in the rcu read critical section within __handle_sysrq(),
this commit therefore remove redundant rcu_read_lock/unlock() in
the sysrq_handle_sched_ext_reset().

Signed-off-by: Zqiang <qiang.zhang@linux.dev>
---
 kernel/sched/ext.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c
index 45a2668284bc..0f3e9d088151 100644
--- a/kernel/sched/ext.c
+++ b/kernel/sched/ext.c
@@ -7953,13 +7953,11 @@ static void sysrq_handle_sched_ext_reset(u8 key)
 {
 	struct scx_sched *sch;
 
-	rcu_read_lock();
 	sch = rcu_dereference(scx_root);
 	if (likely(sch))
 		scx_disable(sch, SCX_EXIT_SYSRQ);
 	else
 		pr_info("sched_ext: BPF schedulers not loaded\n");
-	rcu_read_unlock();
 }
 
 static const struct sysrq_key_op sysrq_sched_ext_reset_op = {
-- 
2.17.1

Hello,

Applied to sched_ext/for-7.2 with Andrea's suggested commit message
rewording and both Reviewed-by tags.

Thanks.

--
tejun

> 
> Hello,
> 
> Applied to sched_ext/for-7.2 with Andrea's suggested commit message
> rewording and both Reviewed-by tags.

Thanks both (Andrea's suggested commit and Cheng-Yang Chou's review) and tejun.

Thanks
Zqiang

> 
> Thanks.
> 
> --
> tejun
>

Hi Zqiang,

On Thu, Apr 30, 2026 at 04:45:43PM +0800, Zqiang wrote:
> The sysrq_handle_sched_ext_reset() callback has already been
> executed in the rcu read critical section within __handle_sysrq(),
> this commit therefore remove redundant rcu_read_lock/unlock() in
> the sysrq_handle_sched_ext_reset().
> 
> Signed-off-by: Zqiang <qiang.zhang@linux.dev>

Looks good to me.

Reviewed-by: Andrea Righi <arighi@nvidia.com>

Thanks,
-Andrea

> ---
>  kernel/sched/ext.c | 2 --
>  1 file changed, 2 deletions(-)
> 
> diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c
> index 45a2668284bc..0f3e9d088151 100644
> --- a/kernel/sched/ext.c
> +++ b/kernel/sched/ext.c
> @@ -7953,13 +7953,11 @@ static void sysrq_handle_sched_ext_reset(u8 key)
>  {
>  	struct scx_sched *sch;
>  
> -	rcu_read_lock();
>  	sch = rcu_dereference(scx_root);
>  	if (likely(sch))
>  		scx_disable(sch, SCX_EXIT_SYSRQ);
>  	else
>  		pr_info("sched_ext: BPF schedulers not loaded\n");
> -	rcu_read_unlock();
>  }
>  
>  static const struct sysrq_key_op sysrq_sched_ext_reset_op = {
> -- 
> 2.17.1
>

On Mon, May 04, 2026 at 07:17:16AM +0200, Andrea Righi wrote:
> Hi Zqiang,
> 
> On Thu, Apr 30, 2026 at 04:45:43PM +0800, Zqiang wrote:
> > The sysrq_handle_sched_ext_reset() callback has already been
> > executed in the rcu read critical section within __handle_sysrq(),
> > this commit therefore remove redundant rcu_read_lock/unlock() in
> > the sysrq_handle_sched_ext_reset().

Actually, minor nits in the commit message:

1) s/has already been executed/is executed/
2) s/remove/removes/

Or maybe rephrase as following?

  sysrq_handle_sched_ext_reset() is called from __handle_sysrq(), which already
  holds rcu_read_lock() while invoking the sysrq handler. Remove the redundant
  rcu_read_lock/unlock() pair.

Thanks,
-Andrea

> > 
> > Signed-off-by: Zqiang <qiang.zhang@linux.dev>
> 
> Looks good to me.
> 
> Reviewed-by: Andrea Righi <arighi@nvidia.com>
> 
> Thanks,
> -Andrea
> 
> > ---
> >  kernel/sched/ext.c | 2 --
> >  1 file changed, 2 deletions(-)
> > 
> > diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c
> > index 45a2668284bc..0f3e9d088151 100644
> > --- a/kernel/sched/ext.c
> > +++ b/kernel/sched/ext.c
> > @@ -7953,13 +7953,11 @@ static void sysrq_handle_sched_ext_reset(u8 key)
> >  {
> >  	struct scx_sched *sch;
> >  
> > -	rcu_read_lock();
> >  	sch = rcu_dereference(scx_root);
> >  	if (likely(sch))
> >  		scx_disable(sch, SCX_EXIT_SYSRQ);
> >  	else
> >  		pr_info("sched_ext: BPF schedulers not loaded\n");
> > -	rcu_read_unlock();
> >  }
> >  
> >  static const struct sysrq_key_op sysrq_sched_ext_reset_op = {
> > -- 
> > 2.17.1
> >

Hi Zqiang,

On Thu, Apr 30, 2026 at 04:45:43PM +0800, Zqiang wrote:
> The sysrq_handle_sched_ext_reset() callback has already been
> executed in the rcu read critical section within __handle_sysrq(),
> this commit therefore remove redundant rcu_read_lock/unlock() in

Nit: s/remove/removes

Looks good!
Reviewed-by: Cheng-Yang Chou <yphbchou0911@gmail.com>

> the sysrq_handle_sched_ext_reset().
> 
> Signed-off-by: Zqiang <qiang.zhang@linux.dev>
> ---
>  kernel/sched/ext.c | 2 --
>  1 file changed, 2 deletions(-)
> 
> diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c
> index 45a2668284bc..0f3e9d088151 100644
> --- a/kernel/sched/ext.c
> +++ b/kernel/sched/ext.c
> @@ -7953,13 +7953,11 @@ static void sysrq_handle_sched_ext_reset(u8 key)
>  {
>  	struct scx_sched *sch;
>  
> -	rcu_read_lock();
>  	sch = rcu_dereference(scx_root);
>  	if (likely(sch))
>  		scx_disable(sch, SCX_EXIT_SYSRQ);
>  	else
>  		pr_info("sched_ext: BPF schedulers not loaded\n");
> -	rcu_read_unlock();
>  }
>  
>  static const struct sysrq_key_op sysrq_sched_ext_reset_op = {
> -- 
> 2.17.1
> 
> 

-- 
Cheers,
Cheng-Yang