linux-kernel - Re: [PATCH 3/7] devfreq: add clearing transitions stats in sysfs

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]

Date:   Fri, 15 Nov 2019 11:47:12 +0900
From:   Chanwoo Choi <cw00.choi@...sung.com>
To:     Bartlomiej Zolnierkiewicz <b.zolnierkie@...sung.com>,
        Kamil Konieczny <k.konieczny@...sung.com>
Cc:     Krzysztof Kozlowski <krzk@...nel.org>,
        Kyungmin Park <kyungmin.park@...sung.com>,
        linux-kernel@...r.kernel.org, linux-pm@...r.kernel.org,
        Marek Szyprowski <m.szyprowski@...sung.com>,
        MyungJoo Ham <myungjoo.ham@...sung.com>
Subject: Re: [PATCH 3/7] devfreq: add clearing transitions stats in sysfs

Hi Bartlomiej,

On 11/15/19 3:23 AM, Bartlomiej Zolnierkiewicz wrote:
> 
> Hi Chanwoo,
> 
> On 11/13/19 10:41 AM, Chanwoo Choi wrote:
>> Hi,
>>
>> If user only want to use the transitions stats information
>> from now, the user just read the sysfs twice and then
>> can calculate them between the read data. It is enough
> 
> IOW you are suggesting that user should invest his valuable time
> into actually doing such calculations (or implementing some extra
> script to do the data parsing and calculations automatically)
> instead of doing simple write to special sysfs file?

Actually, I wouldn't like to add the new node for the debugging.
If there are some requirement for the debugging, we better to
add the debugfs for devfreq (devfreq has not yet developed
the debugfs. I'll do that). If some node is necessary for the devfreq
core/device operation, Surely, I'll agree.

On the initial version of devfreq, it contained the 'trans_stat'.
In order to keep the compatibility, just could not move 'trans_stat'
to under debugfs path.

If there are any requirement for resetting for transition stats,
I prefer to use the following style without any extra debugging node.

	For resetting for transition stats as following,
	echo 0 > /sys/class/devfreq/devfreq0/trans_stat

> 
> Also similar patch for cpufreq has been applied not so long ago:
> 
> commit ee7930ee27fe5240398cc302fa8eb4454725f188
> Author: Markus Mayer <mmayer@...adcom.com>
> Date:   Mon Nov 7 10:02:23 2016 -0800
> 
>     cpufreq: stats: New sysfs attribute for clearing statistics
>     
>     Allow CPUfreq statistics to be cleared by writing anything to
>     /sys/.../cpufreq/stats/reset.
>     
>     Signed-off-by: Markus Mayer <mmayer@...adcom.com>
>     Acked-by: Viresh Kumar <viresh.kumar@...aro.org>
>     Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@...el.com>
> ...
> 
> Best regards,
> --
> Bartlomiej Zolnierkiewicz
> Samsung R&D Institute Poland
> Samsung Electronics
> 
>> to get the existing sysfs information. 
>>
>> And I don't know the any other reason. So, I can't agree this patch.
>> So, Not ack.
>>
>> Regards,
>> Chanwoo Choi
>>
>>
>> On 11/13/19 6:13 PM, Kamil Konieczny wrote:
>>> Add new function trans_reset in sysfs for clearing transition
>>> table and time in states devfreq statistics.> 
>>> Signed-off-by: Kamil Konieczny <k.konieczny@...sung.com>
>>> ---
>>>  drivers/devfreq/devfreq.c | 26 ++++++++++++++++++++++++++
>>>  1 file changed, 26 insertions(+)
>>>
>>> diff --git a/drivers/devfreq/devfreq.c b/drivers/devfreq/devfreq.c
>>> index ac04b5baef70..0a88055d1362 100644
>>> --- a/drivers/devfreq/devfreq.c
>>> +++ b/drivers/devfreq/devfreq.c
>>> @@ -1445,6 +1445,31 @@ static ssize_t trans_stat_show(struct device *dev,
>>>  }
>>>  static DEVICE_ATTR_RO(trans_stat);
>>>  
>>> +static void defvreq_stats_clear_table(struct devfreq *devfreq)
>>> +{
>>> +	unsigned int count = devfreq->profile->max_state;
>>> +
>>> +	spin_lock(&devfreq->stats_lock);
>>> +	memset(devfreq->time_in_state, 0, count * sizeof(u64));
>>> +	memset(devfreq->trans_table, 0, count * count * sizeof(int));
>>> +	devfreq->last_stat_updated = get_jiffies_64();
>>> +	devfreq->total_trans = 0;
>>> +	spin_unlock(&devfreq->stats_lock);
>>> +}
>>> +
>>> +static ssize_t trans_reset_store(struct device *dev,
>>> +				 struct device_attribute *attr,
>>> +				 const char *buf,
>>> +				 size_t count)
>>> +{
>>> +	struct devfreq *devfreq = to_devfreq(dev);
>>> +
>>> +	defvreq_stats_clear_table(devfreq);
>>> +
>>> +	return count;
>>> +}
>>> +static DEVICE_ATTR_WO(trans_reset);
>>> +
>>>  static struct attribute *devfreq_attrs[] = {
>>>  	&dev_attr_governor.attr,
>>>  	&dev_attr_available_governors.attr,
>>> @@ -1455,6 +1480,7 @@ static struct attribute *devfreq_attrs[] = {
>>>  	&dev_attr_min_freq.attr,
>>>  	&dev_attr_max_freq.attr,
>>>  	&dev_attr_trans_stat.attr,
>>> +	&dev_attr_trans_reset.attr,
>>>  	NULL,
>>>  };
>>>  ATTRIBUTE_GROUPS(devfreq);
> 
> 
> 


-- 
Best Regards,
Chanwoo Choi
Samsung Electronics