Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH v3 6/7] trace: events: add devfreq trace event file
To:     Lukasz Luba <l.luba@partner.samsung.com>,
        Steven Rostedt <rostedt@goodmis.org>
Cc:     linux-kernel@vger.kernel.org, linux-pm@vger.kernel.org,
        b.zolnierkie@samsung.com, myungjoo.ham@samsung.com,
        kyungmin.park@samsung.com, m.szyprowski@samsung.com,
        s.nawrocki@samsung.com, joel@joelfernandes.org,
        chris.diamand@arm.com, mka@chromium.org, mingo@redhat.com
From:   Chanwoo Choi <cw00.choi@samsung.com>
Organization: Samsung Electronics
Message-ID: <3e44ca8a-a72b-8bdd-e702-00c927bd3c06@samsung.com>
Date:   Thu, 14 Feb 2019 14:01:53 +0900
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101
        Thunderbird/60.4.0
MIME-Version: 1.0
In-Reply-To: <fb5d5359-7663-f119-f28f-7be820e1108d@partner.samsung.com>
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Content-Type: text/plain; charset="utf-8"
CMS-TYPE: 101P
DLP-Filter: Pass
References: <1550010238-24002-1-git-send-email-l.luba@partner.samsung.com>
        <CGME20190212222437eucas1p198db6fca1f1ba3056d93c57327dd48ed@eucas1p1.samsung.com>
        <1550010238-24002-7-git-send-email-l.luba@partner.samsung.com>
        <20190212181415.1156e542@gandalf.local.home>
        <fb5d5359-7663-f119-f28f-7be820e1108d@partner.samsung.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Hi Lukasz,

On 19. 2. 13. 오후 10:35, Lukasz Luba wrote:
> Hi Steven,
> 
> On 2/13/19 12:14 AM, Steven Rostedt wrote:
>> On Tue, 12 Feb 2019 23:23:57 +0100
>> Lukasz Luba <l.luba@partner.samsung.com> wrote:
>>
>>> The patch adds a new file for with trace events for devfreq
>>> framework. They are used for performance analysis of the framework.
>>> It also contains updates in MAINTAINERS file adding new entry for
>>> devfreq maintainers.
>>>
>>> Signed-off-by: Lukasz Luba <l.luba@partner.samsung.com>
>>> ---
>>>   MAINTAINERS                    |  1 +
>>>   include/trace/events/devfreq.h | 39 +++++++++++++++++++++++++++++++++++++++
>>>   2 files changed, 40 insertions(+)
>>>   create mode 100644 include/trace/events/devfreq.h
>>>
>>> diff --git a/MAINTAINERS b/MAINTAINERS
>>> index 9919840..c042fda 100644
>>> --- a/MAINTAINERS
>>> +++ b/MAINTAINERS
>>> @@ -4447,6 +4447,7 @@ S:	Maintained
>>>   F:	drivers/devfreq/
>>>   F:	include/linux/devfreq.h
>>>   F:	Documentation/devicetree/bindings/devfreq/
>>> +F:	include/trace/events/devfreq.h
>>>   
>>>   DEVICE FREQUENCY EVENT (DEVFREQ-EVENT)
>>>   M:	Chanwoo Choi <cw00.choi@samsung.com>
>>> diff --git a/include/trace/events/devfreq.h b/include/trace/events/devfreq.h
>>> new file mode 100644
>>> index 0000000..fec9304
>>> --- /dev/null
>>> +++ b/include/trace/events/devfreq.h
>>> @@ -0,0 +1,39 @@
>>> +/* SPDX-License-Identifier: GPL-2.0 */
>>> +#undef TRACE_SYSTEM
>>> +#define TRACE_SYSTEM devfreq
>>> +
>>> +#if !defined(_TRACE_DEVFREQ_H) || defined(TRACE_HEADER_MULTI_READ)
>>> +#define _TRACE_DEVFREQ_H
>>> +
>>> +#include <linux/devfreq.h>
>>> +#include <linux/tracepoint.h>
>>> +
>>> +TRACE_EVENT(devfreq_monitor,
>>> +	TP_PROTO(const char *dev_name, unsigned long freq,
>>> +		 unsigned int polling_ms, unsigned long busy_time,
>>> +		 unsigned long total_time),
>>> +
>>> +	TP_ARGS(dev_name, freq, polling_ms, busy_time, total_time),
>>> +
>>> +	TP_STRUCT__entry(
>>> +		__string(dev_name, dev_name)
>>> +		__field(unsigned long, freq)
>>> +		__field(unsigned int, polling_ms)
>>> +		__field(unsigned int, load)
>>> +	),
>>> +
>>> +	TP_fast_assign(
>>> +		__assign_str(dev_name, dev_name);
>>> +		__entry->freq = freq;
>>> +		__entry->polling_ms = polling_ms;
>>> +		__entry->load = (100 * busy_time) / total_time;
>>
>> How critical is the code that this trace event is called in. If it is
>> not that critical (it is a slow path), then this is fine, but if this
>> is not a slow path (something triggered 100 HZ or less), then I would
>> recommend moving the above calculation to TP_printk(). A divide is a
>> slow operation, and is best to do that in the post processing. The
>> current location does the divide at the time of the tracepoint is
>> called.
> I wasn't aware of these two stages, good to know.
> I will move it to TP_printk().
>>
>> I would also have a check to make sure that total_time is not zero
>> here, otherwise that could be bad.
>>
>> 		__entry->busy_time = busy_time;
>> 		__entry->total_time = total_time;
>>
> 
>>> +	),
>>> +
>>> +	TP_printk("dev_name=%s freq=%lu polling_ms=%u load=%u",
>>
>>
>>> +		__get_str(dev_name), __entry->freq, __entry->polling_ms,
>>
>> 		__entry->total_time == 0 ? 100 :
>> 			__entry->busy_time / __entry->total_time)
> Thank you for the review.
> I will add this check.
> 
> Regards,
> Lukasz
>>
>> -- Steve
>>
>>> +		__entry->load)
>>> +);
>>> +#endif /* _TRACE_DEVFREQ_H */
>>> +
>>> +/* This part must be outside protection */
>>> +#include <trace/define_trace.h>
>>
>>
>>
> 
> 

I agree that trace point is necessary for devfreq framework.
Reviewed-by: Chanwoo Choi <cw00.choi@samsung.com>


-- 
Best Regards,
Chanwoo Choi
Samsung Electronics