[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 5 May 2017 01:53:24 +0000
From: "Li, Fei" <fei.li@...el.com>
To: Daniel Lezcano <daniel.lezcano@...aro.org>
CC: Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
"Shi, FengX" <fengx.shi@...el.com>,
"rjw@...ysocki.net" <rjw@...ysocki.net>,
"linux-pm@...r.kernel.org" <linux-pm@...r.kernel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"Koul, Vinod" <vinod.koul@...el.com>
Subject: RE: [PATCH V3] cpuidle: check dev before usage in
cpuidle_use_deepest_state
On Thursday, May 4, 2017 6:16 PM Daniel Lezcano wrote:
> >
> > preempt_disable();
> > dev = cpuidle_get_device();
> > - dev->use_deepest_state = enable;
> > + if (dev)
> > + dev->use_deepest_state = enable;
>
>
> This change is acceptable as a hot fix but the question is 'why is this
> function called if there is no cpuidle driver registered?'
It's called as below:
clamp_idle_injection_func
-> play_idle
-> cpuidle_use_deepest_state
In play_idle, if no cpuidle driver registered, cpu_idle_poll can be called.
It seems to go too far if not calling play_idle in case of no cpuidle driver
registered.
At present no other suitable and simple fix seen.
>
> > preempt_enable();
> > }
> >
> > --
> > 1.9.1
> >
> > Best Regards,
> > Fei
>
> --
>
> <http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs
>
> Follow Linaro: <http://www.facebook.com/pages/Linaro> Facebook |
> <http://twitter.com/#!/linaroorg> Twitter |
> <http://www.linaro.org/linaro-blog/> Blog
Best Regards,
Fei
Powered by blists - more mailing lists