linux-kernel - Re: [PATCH v6 08/19] PM / devfreq: tegra30: Ensure that target freq won't overflow

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: <f4398556-12e0-9ba8-b59b-19a38764312f@gmail.com>
Date:   Wed, 21 Aug 2019 02:19:52 +0300
From:   Dmitry Osipenko <digetx@...il.com>
To:     Chanwoo Choi <cw00.choi@...sung.com>
Cc:     Thierry Reding <thierry.reding@...il.com>,
        MyungJoo Ham <myungjoo.ham@...sung.com>,
        Kyungmin Park <kyungmin.park@...sung.com>,
        Jonathan Hunter <jonathanh@...dia.com>,
        Tomeu Vizoso <tomeu.vizoso@...labora.com>,
        linux-pm@...r.kernel.org, linux-tegra@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v6 08/19] PM / devfreq: tegra30: Ensure that target freq
 won't overflow

Hello Chanwoo,

20.08.2019 3:23, Chanwoo Choi пишет:
> On 19. 8. 12. 오전 6:23, Dmitry Osipenko wrote:
>> We already had few integer overflow bugs, let's limit the freq for
>> consistency.
>>
>> Signed-off-by: Dmitry Osipenko <digetx@...il.com>
>> ---
>>  drivers/devfreq/tegra30-devfreq.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/devfreq/tegra30-devfreq.c b/drivers/devfreq/tegra30-devfreq.c
>> index 70dce58212a4..ca499368ee81 100644
>> --- a/drivers/devfreq/tegra30-devfreq.c
>> +++ b/drivers/devfreq/tegra30-devfreq.c
>> @@ -430,7 +430,7 @@ static unsigned long actmon_update_target(struct tegra_devfreq *tegra,
>>  	target_freq = dev->avg_count / ACTMON_SAMPLING_PERIOD + dev->boost_freq;
>>  	target_freq = tegra_actmon_account_cpu_freq(tegra, dev, target_freq);
>>  
>> -	return target_freq;
>> +	return min(target_freq, tegra->max_freq);
> 
> Once again, did you meet this case sometimes?

This case happens at least whenever CPU's freq accounting returns
KHZ_MAX, please see actmon_emc_ratios[] and actmon_cpu_to_emc_rate().

> Usually, we can prevent the overflow of target_freq
> when calculating the target frequency or this style.
> 
> I think that if the overflow of target frequency happen frequently,
> it might have the problem of calculation way. 

It is possible to get sampled avg_count from hardware that results in
target_freq > max_freq because translating number of memory events to
memory frequency has some calculation error, although it is negligible.

In fact, it shouldn't matter that the returned target_freq > max_freq
with the current driver's code structure and in the commit's message I
wrote that this change is done for consistency. Hence it should be okay
to drop this patch as well.