Re: [PATCH 4/4] sched_ext: Make scx_locked_rq() shared

[Tejun Heo <tj@kernel.org>]

On Wed, Jun 04, 2025 at 08:48:56AM -1000, Tejun Heo wrote:
> On Wed, Jun 04, 2025 at 04:33:14PM +0200, Andrea Righi wrote:
> > scx_locked_rq() is used both from ext.c and ext_idle.c, so make it
> > public and declare its prototype in ext.h.
> > 
> > No functional changes.
> > 
> > Signed-off-by: Andrea Righi <arighi@nvidia.com>
> > ---
> >  kernel/sched/ext.c | 2 +-
> >  kernel/sched/ext.h | 2 ++
> >  2 files changed, 3 insertions(+), 1 deletion(-)
> > 
> > diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c
> > index 3e483138dff60..941603ec67e27 100644
> > --- a/kernel/sched/ext.c
> > +++ b/kernel/sched/ext.c
> > @@ -1265,7 +1265,7 @@ static inline void update_locked_rq(struct rq *rq)
> >   * Return the rq currently locked from an scx callback, or NULL if no rq is
> >   * locked.
> >   */
> > -static inline struct rq *scx_locked_rq(void)
> > +struct rq *scx_locked_rq(void)
> >  {
> >  	return __this_cpu_read(locked_rq);
> 
> Can you rename locked_rq to scx_locked_rq_var (or something else), expose it
> and then make scx_locked_rq() an inline function in ext.h. Alternatively, I
> think it'd be fine to drop the wrapper and let each user do
> __this_cpu_read(scx_locked_rq) too.

BTW, if you update this patch, no need to resend 1-3. Just send the updated
patch as a reply to this one.

Thanks.

-- 
tejun