linux-kernel - Re: [PATCH] cpufreq: schedutil: Examine the correct CPU when we update util

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

Message-ID: <20171102111022.GB4240@vireshk-i7>
Date:   Thu, 2 Nov 2017 16:40:23 +0530
From:   Viresh Kumar <viresh.kumar@...aro.org>
To:     Chris Redpath <chris.redpath@....com>
Cc:     linux-kernel@...r.kernel.org, linux-pm@...r.kernel.org,
        morten.rasmussen@....com, dietmar.eggemann@....com,
        "Rafael J . Wysocki" <rjw@...ysocki.net>,
        Ingo Molnar <mingo@...hat.com>,
        Peter Zijlstra <peterz@...radead.org>,
        "4 . 9+" <stable@...r.kernel.org>
Subject: Re: [PATCH] cpufreq: schedutil: Examine the correct CPU when we
 update util

On 02-11-17, 10:54, Chris Redpath wrote:
> Since:
> 
> 4296f23ed cpufreq: schedutil: Fix per-CPU structure initialization in
>  sugov_start()

Good catch but you pointed out to the wrong commit really.

674e75411fc2 ("sched: cpufreq: Allow remote cpufreq callbacks"

is the real culprit as that's where the 'cpu' field is added.

> We lost the value of sg_cpu->cpu which is assigned during
> sugov_register. The memset in sugov_start overwrites it with zero.
> 
> This leads to always looking at the utilization of CPU0 instead of
> the one we just updated when we do a utilization update callback.
> 
> Let's fix this by consolidating the initialization code into
> sugov_start().
> 
> Fixes: 4296f23ed49a ("cpufreq: schedutil: Fix per-CPU structure initialization in sugov_start()")

Fixes: 674e75411fc2 ("sched: cpufreq: Allow remote cpufreq callbacks")

> Signed-off-by: Chris Redpath <chris.redpath@....com>
> Reviewed-by: Patrick Bellasi <patrick.bellasi@....com>
> Reviewed-by: Brendan Jackman <brendan.jackman@....com>
> Cc: Rafael J. Wysocki <rjw@...ysocki.net>
> Cc: Viresh Kumar <viresh.kumar@...aro.org>
> Cc: Ingo Molnar <mingo@...hat.com>
> Cc: Peter Zijlstra <peterz@...radead.org>
> Cc: 4.9+ <stable@...r.kernel.org>

This wouldn't be required now as it got merged in 4.14 only and Rafael
should be able to get that as part of 4.14 itself.

> ---
>  kernel/sched/cpufreq_schedutil.c | 6 +-----
>  1 file changed, 1 insertion(+), 5 deletions(-)
> 
> diff --git a/kernel/sched/cpufreq_schedutil.c b/kernel/sched/cpufreq_schedutil.c
> index 6c1a7fcfa2a7..eeb7e0a1d861 100644
> --- a/kernel/sched/cpufreq_schedutil.c
> +++ b/kernel/sched/cpufreq_schedutil.c
> @@ -731,6 +731,7 @@ static int sugov_start(struct cpufreq_policy *policy)
>  		sg_cpu->sg_policy = sg_policy;
>  		sg_cpu->flags = SCHED_CPUFREQ_RT;
>  		sg_cpu->iowait_boost_max = policy->cpuinfo.max_freq;
> +		sg_cpu->cpu = cpu;

Maybe just move this below memset.

>  	}
>  
>  	for_each_cpu(cpu, policy->cpus) {
> @@ -793,11 +794,6 @@ struct cpufreq_governor *cpufreq_default_governor(void)
>  
>  static int __init sugov_register(void)
>  {
> -	int cpu;
> -
> -	for_each_possible_cpu(cpu)
> -		per_cpu(sugov_cpu, cpu).cpu = cpu;
> -
>  	return cpufreq_register_governor(&schedutil_gov);
>  }
>  fs_initcall(sugov_register);

Diff looks fine though.

-- 
viresh