[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <CAJZ5v0ivZnvVbY4WyM1SQO1TY=G8yb4=C0kWkVQPas5=NX1Sog@mail.gmail.com>
Date: Mon, 28 Sep 2020 16:10:36 +0200
From: "Rafael J. Wysocki" <rafael@...nel.org>
To: zhuguangqing83 <zhuguangqing83@...il.com>
Cc: "Rafael J. Wysocki" <rafael@...nel.org>,
"Rafael J. Wysocki" <rjw@...ysocki.net>,
Daniel Lezcano <daniel.lezcano@...aro.org>,
Linux PM <linux-pm@...r.kernel.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
zhuguangqing <zhuguangqing@...omi.com>
Subject: Re: [PATCH] cpuidle: change #ifdef for the declaration of cpuidle_enter_s2idle()
On Sat, Sep 26, 2020 at 3:30 AM zhuguangqing83 <zhuguangqing83@...il.com> wrote:
>
>
> > On Thu, Sep 24, 2020 at 10:01 AM <zhuguangqing83@...il.com> wrote:
> > >
> > > From: zhuguangqing <zhuguangqing@...omi.com>
> > >
> > > Currently, if CONFIG_SUSPEND=n and CONFIG_CPU_IDLE=y, the function
> > > cpuidle_enter_s2idle() is declared but not defined, it may cause error
> > > when cpuidle_enter_s2idle() is called.
> > >
> > > If CONFIG_SUSPEND=y and CONFIG_CPU_IDLE=n, the function
> > > cpuidle_enter_s2idle() is defined as "return -ENODEV;" which is not
> > > supposed to be.
> > >
> > > Change #ifdef CONFIG_CPU_IDLE to #ifdef CONFIG_SUSPEND for
> > > cpuidle_enter_s2idle() in cpuidle.h, which is consistent with its
> > > defination in cpuidle.c.
> >
> > Well, what about the case when CONFIG_SUSPEND is set, but CONFIG_CPU_IDLE
> > isn't?
> >
>
> When CONFIG_SUSPEND is set, but CONFIG_CPU_IDLE isn't, the function
> cpuidle_enter_s2idle() is defined in cpuidle.c, but the defination in
> cpuidle.c is not used actually because CONFIG_CPU_IDLE isn't set, we
> only use its defination as "return -ENODEV;" in cpuidle.h.
Actually, if CONFIG_CPU_IDLE is not set, cpuidle.c is not compiled at
all AFAICS, but after the $subject patch the compiler will be looking
for cpuidle_enter_s2idle() in that case, won't it? [The static inline
stub is only present for CONFIG_SUSPEND unset.]
> > > Signed-off-by: zhuguangqing <zhuguangqing@...omi.com>
> > > ---
> > > include/linux/cpuidle.h | 14 +++++++++-----
> > > 1 file changed, 9 insertions(+), 5 deletions(-)
> > >
> > > diff --git a/include/linux/cpuidle.h b/include/linux/cpuidle.h
> > > index 6175c77bf25e..2aa8cead1727 100644
> > > --- a/include/linux/cpuidle.h
> > > +++ b/include/linux/cpuidle.h
> > > @@ -216,22 +216,26 @@ static inline struct cpuidle_device
> > *cpuidle_get_device(void) {return NULL; }
> > > extern int cpuidle_find_deepest_state(struct cpuidle_driver *drv,
> > > struct cpuidle_device *dev,
> > > u64 latency_limit_ns);
> > > -extern int cpuidle_enter_s2idle(struct cpuidle_driver *drv,
> > > - struct cpuidle_device *dev);
> > > extern void cpuidle_use_deepest_state(u64 latency_limit_ns);
> > > #else
> > > static inline int cpuidle_find_deepest_state(struct cpuidle_driver *drv,
> > > struct cpuidle_device *dev,
> > > u64 latency_limit_ns)
> > > {return -ENODEV; }
> > > -static inline int cpuidle_enter_s2idle(struct cpuidle_driver *drv,
> > > - struct cpuidle_device *dev)
> > > -{return -ENODEV; }
> > > static inline void cpuidle_use_deepest_state(u64 latency_limit_ns)
> > > {
> > > }
> > > #endif
> > >
> > > +#ifdef CONFIG_SUSPEND
> > > +extern int cpuidle_enter_s2idle(struct cpuidle_driver *drv,
> > > + struct cpuidle_device *dev);
> > > +#else
> > > +static inline int cpuidle_enter_s2idle(struct cpuidle_driver *drv,
> > > + struct cpuidle_device *dev)
> > > +{return -ENODEV; }
> > > +#endif
> > > +
> > > /* kernel/sched/idle.c */
> > > extern void sched_idle_set_state(struct cpuidle_state *idle_state);
> > > extern void default_idle_call(void);
> > > --
> > > 2.17.1
> > >
>
Powered by blists - more mailing lists