From: Chanwoo Choi <cw00.choi@samsung.com>
To: Yue Hu <zbestahu@gmail.com>
Cc: myungjoo.ham@samsung.com, kyungmin.park@samsung.com,
linux-pm@vger.kernel.org, huyue2@yulong.com
Subject: Re: [PATCH] PM / devfreq: Drop the name check to request module in try_then_request_governor()
Date: Wed, 31 Jul 2019 15:02:29 +0900 [thread overview]
Message-ID: <2e16d9c4-3207-6f0d-9210-1948f64b20cc@samsung.com> (raw)
In-Reply-To: <20190731135736.00006a55.zbestahu@gmail.com>
On 19. 7. 31. 오후 2:57, Yue Hu wrote:
> On Wed, 31 Jul 2019 14:55:39 +0900
> Chanwoo Choi <cw00.choi@samsung.com> wrote:
>
>> On 19. 7. 31. 오후 2:38, Yue Hu wrote:
>>> On Wed, 31 Jul 2019 09:33:06 +0900
>>> Chanwoo Choi <cw00.choi@samsung.com> wrote:
>>>
>>>> On 19. 7. 30. 오후 7:08, Yue Hu wrote:
>>>>> From: Yue Hu <huyue2@yulong.com>
>>>>>
>>>>> No need to check specific governor name of `simple_ondemand` to request
>>>>> module, let's change the name string to `simpleondemand` to keep the
>>>>> consistency on loading module if needed.
>>>>
>>>> NACK.
>>>>
>>>> hmm.... It is impossible to change the devfreq governor name
>>>> because there are many reason.
>>>>
>>>> The devfreq governor could be changed through the sysfs interface
>>>> on runtime. For a long time, many users or platforms change
>>>> the devfreq governor with the defined governor name through sysfs.
>>>> If it is just changed, it breaks ABI interface and cannot support
>>>> the compatibility. It is very critical problem. Please drop it.
>>>
>>> Yes, needs update also if using sysfs. it's problem indeed.
>>
>> No, It is impossible to update it. You have to change all kind of
>> platform in the world. We never know the all use-case in the world.
>> As I said, it break the ABI interface.
>>
>>>
>>>>
>>>>
>>>> Maybe, you didn't check the usage of devfreq device driver
>>>> in the mainline kernel. Almost devfreq device using simple_ondemand
>>>> governor have to add the governor name with devfreq_add_device().
>>>> If changed the governor name, it cause the fault of device driver
>>>> using the devfreq framework with simple_ondemand.
>>>
>>> Currently, seems no devfreq users use the simple_ondemand directly in
>>> mainline kernel.
>>
>> You can find them in the mainline kernel as following:
>>
>> drivers/gpu/drm/panfrost/panfrost_devfreq.c:160:&panfrost_devfreq_profile, "simple_ondemand", NULL);
>> drivers/gpu/drm/msm/msm_gpu.c:98: &msm_devfreq_profile, "simple_ondemand", NULL);
>
> drm related code is already updated as below link:
> https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=67fe62dcf713c36f4766c0218cc14796ee9536e1
>
>>
>> drivers/scsi/ufs/ufshcd.c:1333: DEVFREQ_GOV_SIMPLE_ONDEMAND,
>> drivers/devfreq/tegra20-devfreq.c:176: DEVFREQ_GOV_SIMPLE_ONDEMAND, NULL);
>> drivers/devfreq/rk3399_dmc.c:452: DEVFREQ_GOV_SIMPLE_ONDEMAND,
>> drivers/devfreq/exynos-bus.c:437: DEVFREQ_GOV_SIMPLE_ONDEMAND,
>>
>>>
>>> Maybe we can rename the governor file name to governor_simpleondemand.c,
>>> just not compatible to module name compared with this change.
>>
>> The file name was already 'drivers/devfreq/governor_simpleondemand.c'.
>
> Sorry for the typo error. I mean governor_simple_ondemand.c?
Actually, it is not necessary because there are no benefit.
>
> Thanks.
>
>>
>>
>>>
>>> Thanks.
>>>
>>>>
>>>>
>>>>>
>>>>> Signed-off-by: Yue Hu <huyue2@yulong.com>
>>>>> ---
>>>>> drivers/devfreq/devfreq.c | 6 +-----
>>>>> include/linux/devfreq.h | 2 +-
>>>>> 2 files changed, 2 insertions(+), 6 deletions(-)
>>>>>
>>>>> diff --git a/drivers/devfreq/devfreq.c b/drivers/devfreq/devfreq.c
>>>>> index 784c08e..baff682 100644
>>>>> --- a/drivers/devfreq/devfreq.c
>>>>> +++ b/drivers/devfreq/devfreq.c
>>>>> @@ -246,11 +246,7 @@ static struct devfreq_governor *try_then_request_governor(const char *name)
>>>>> if (IS_ERR(governor)) {
>>>>> mutex_unlock(&devfreq_list_lock);
>>>>>
>>>>> - if (!strncmp(name, DEVFREQ_GOV_SIMPLE_ONDEMAND,
>>>>> - DEVFREQ_NAME_LEN))
>>>>> - err = request_module("governor_%s", "simpleondemand");
>>>>> - else
>>>>> - err = request_module("governor_%s", name);
>>>>> + err = request_module("governor_%s", name);
>>>>> /* Restore previous state before return */
>>>>> mutex_lock(&devfreq_list_lock);
>>>>> if (err)
>>>>> diff --git a/include/linux/devfreq.h b/include/linux/devfreq.h
>>>>> index 2bae9ed..41e8792 100644
>>>>> --- a/include/linux/devfreq.h
>>>>> +++ b/include/linux/devfreq.h
>>>>> @@ -17,7 +17,7 @@
>>>>> #define DEVFREQ_NAME_LEN 16
>>>>>
>>>>> /* DEVFREQ governor name */
>>>>> -#define DEVFREQ_GOV_SIMPLE_ONDEMAND "simple_ondemand"
>>>>> +#define DEVFREQ_GOV_SIMPLE_ONDEMAND "simpleondemand"
>>>>> #define DEVFREQ_GOV_PERFORMANCE "performance"
>>>>> #define DEVFREQ_GOV_POWERSAVE "powersave"
>>>>> #define DEVFREQ_GOV_USERSPACE "userspace"
>>>>>
>>>>
>>>>
>>>
>>>
>>>
>>
>>
>
>
>
--
Best Regards,
Chanwoo Choi
Samsung Electronics
prev parent reply other threads:[~2019-07-31 5:59 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <CGME20190730100844epcas4p14dfa39fff2636e89c94033f154240db0@epcas4p1.samsung.com>
2019-07-30 10:08 ` [PATCH] PM / devfreq: Drop the name check to request module in try_then_request_governor() Yue Hu
2019-07-31 0:33 ` Chanwoo Choi
2019-07-31 5:38 ` Yue Hu
2019-07-31 5:55 ` Chanwoo Choi
2019-07-31 5:57 ` Yue Hu
2019-07-31 6:02 ` Chanwoo Choi [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=2e16d9c4-3207-6f0d-9210-1948f64b20cc@samsung.com \
--to=cw00.choi@samsung.com \
--cc=huyue2@yulong.com \
--cc=kyungmin.park@samsung.com \
--cc=linux-pm@vger.kernel.org \
--cc=myungjoo.ham@samsung.com \
--cc=zbestahu@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).