linux-kernel - Re: [PATCH][RFC] PM / s2idle: Clear _TIF_POLLING

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20200616033618.GA20959@chenyu-office.sh.intel.com>
Date:   Tue, 16 Jun 2020 11:36:18 +0800
From:   Chen Yu <yu.c.chen@...el.com>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     "Rafael J. Wysocki" <rafael@...nel.org>,
        Daniel Lezcano <daniel.lezcano@...aro.org>,
        Ingo Molnar <mingo@...nel.org>,
        Len Brown <len.brown@...el.com>, linux-pm@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH][RFC] PM / s2idle: Clear _TIF_POLLING_NRFLAG before
 suspend to idle

On Mon, Jun 15, 2020 at 09:31:54PM +0200, Peter Zijlstra wrote:
> On Mon, Jun 15, 2020 at 08:40:41PM +0200, Peter Zijlstra wrote:
> 
> > > @@ -186,8 +187,10 @@ int cpuidle_enter_s2idle(struct cpuidle_driver *drv, struct cpuidle_device *dev)
> > >  	 * be frozen safely.
> > >  	 */
> > >  	index = find_deepest_state(drv, dev, U64_MAX, 0, true);
> > > -	if (index > 0)
> > > +	if (index > 0) {
> > > +		__current_clr_polling();
> > >  		enter_s2idle_proper(drv, dev, index);
> > > +	}
> > >  
> > >  	return index;
> > >  }
> > 
> > So how is that commit 08e237fa56a1 not suffient? That makes
> > mwait_idle_with_hints() DTRT for this 'functionally challenged' piece of
> > hardware.
> > 
> > AFAICT intel_enter_s2idle() uses mwait_idle_with_hints().
> > 
> > What am I missing?
> 
> What's missing is that cpuidle_enter_s2idle() doesn't properly match
> call_cpuidle().
>
Right.
> Something like so then. Your version is racy, if someone already set
> TIF_NEED_RESCHED you just clear POLLING and go to sleep.
> 
Got it, I'll test the patch below.

Thanks,
Chenyu
> ---
> 
> diff --git a/drivers/cpuidle/cpuidle.c b/drivers/cpuidle/cpuidle.c
> index c149d9e20dfd..81bee8d03c6d 100644
> --- a/drivers/cpuidle/cpuidle.c
> +++ b/drivers/cpuidle/cpuidle.c
> @@ -133,8 +133,8 @@ int cpuidle_find_deepest_state(struct cpuidle_driver *drv,
>  }
>  
>  #ifdef CONFIG_SUSPEND
> -static void enter_s2idle_proper(struct cpuidle_driver *drv,
> -				struct cpuidle_device *dev, int index)
> +static void s2idle_enter(struct cpuidle_driver *drv,
> +			 struct cpuidle_device *dev, int index)
>  {
>  	ktime_t time_start, time_end;
>  
> @@ -168,6 +168,15 @@ static void enter_s2idle_proper(struct cpuidle_driver *drv,
>  	dev->states_usage[index].s2idle_usage++;
>  }
>  
> +static int call_s2idle(struct cpuidle_driver *drv, struct cpuidle_device *dev,
> +		       int index)
> +{
> +	if (!current_clr_polling_and_test())
> +		s2idle_enter(drv, dev, index);
> +
> +	return index;
> +}
> +
>  /**
>   * cpuidle_enter_s2idle - Enter an idle state suitable for suspend-to-idle.
>   * @drv: cpuidle driver for the given CPU.
> @@ -187,7 +196,7 @@ int cpuidle_enter_s2idle(struct cpuidle_driver *drv, struct cpuidle_device *dev)
>  	 */
>  	index = find_deepest_state(drv, dev, U64_MAX, 0, true);
>  	if (index > 0)
> -		enter_s2idle_proper(drv, dev, index);
> +		call_s2idle(drv, dev, index);
>  
>  	return index;
>  }