linux-media.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Helen Koike <helen.koike@collabora.com>
To: Mauro Carvalho Chehab <mchehab+samsung@kernel.org>,
	Chen-Yu Tsai <wens@kernel.org>
Cc: Ezequiel Garcia <ezequiel@collabora.com>,
	Hans Verkuil <hans.verkuil@cisco.com>,
	kernel@collabora.com,
	Linux Media Mailing List <linux-media@vger.kernel.org>
Subject: Re: [PATCH 2/2] media: Don't hide any menu if "ancillary drivers autoselect" is enabled
Date: Thu, 25 Jul 2019 16:00:02 -0300	[thread overview]
Message-ID: <4517f590-dcda-5bf0-70dd-3cffc5d71ba3@collabora.com> (raw)
In-Reply-To: <20190725154111.7fc7e335@coco.lan>



On 7/25/19 3:41 PM, Mauro Carvalho Chehab wrote:
> Em Fri, 26 Jul 2019 01:29:58 +0800
> Chen-Yu Tsai <wens@kernel.org> escreveu:
> 
>> On Fri, Jul 26, 2019 at 1:06 AM Ezequiel Garcia <ezequiel@collabora.com> wrote:
>>>
>>> On Thu, 2019-07-25 at 12:57 -0300, Mauro Carvalho Chehab wrote:  
>>>> Em Mon, 15 Jul 2019 18:23:16 -0300
>>>> Ezequiel Garcia <ezequiel@collabora.com> escreveu:
>>>>  
>>>>> Many users have been complaining about not being able to find
>>>>> certain menu options. One such example are camera sensor drivers
>>>>> (e.g IMX219, OV5645, etc) which are common on embedded platforms
>>>>> and not always ancillary devices.
>>>>>
>>>>> The problem with MEDIA_SUBDRV_AUTOSELECT seems to be related
>>>>> to the fact that it uses the "visible" kbuild syntax to hide
>>>>> entire group of drivers.
>>>>>
>>>>> This is not obvious and, as explained above, not always desired.
>>>>>
>>>>> To fix the problem, drop the "visible" and stop hiding any menu
>>>>> options. Users skilled enough to configure their kernel are expected
>>>>> to be skilled enough to know what (not) to configure anyway.
>>>>>
>>>>> Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
>>>>> ---
>>>>>  drivers/media/dvb-frontends/Kconfig | 1 -
>>>>>  drivers/media/i2c/Kconfig           | 1 -
>>>>>  drivers/media/spi/Kconfig           | 1 -
>>>>>  drivers/media/tuners/Kconfig        | 1 -
>>>>>  4 files changed, 4 deletions(-)
>>>>>
>>>>> diff --git a/drivers/media/dvb-frontends/Kconfig b/drivers/media/dvb-frontends/Kconfig
>>>>> index dc43749177df..2d1fea3bf546 100644
>>>>> --- a/drivers/media/dvb-frontends/Kconfig
>>>>> +++ b/drivers/media/dvb-frontends/Kconfig
>>>>> @@ -1,5 +1,4 @@
>>>>>  menu "Customise DVB Frontends"
>>>>> -   visible if !MEDIA_SUBDRV_AUTOSELECT || COMPILE_TEST || EXPERT
>>>>>
>>>>>  comment "Multistandard (satellite) frontends"
>>>>>     depends on DVB_CORE
>>>>> diff --git a/drivers/media/i2c/Kconfig b/drivers/media/i2c/Kconfig
>>>>> index 79ce9ec6fc1b..475072bb67d6 100644
>>>>> --- a/drivers/media/i2c/Kconfig
>>>>> +++ b/drivers/media/i2c/Kconfig
>>>>> @@ -23,7 +23,6 @@ config VIDEO_IR_I2C
>>>>>  #
>>>>>
>>>>>  menu "I2C Encoders, decoders, sensors and other helper chips"
>>>>> -   visible if !MEDIA_SUBDRV_AUTOSELECT || COMPILE_TEST || EXPERT  
>>>>
>>>> Hmm... Hans picked this patch, but IMO it doesn't make sense
>>>> for PC consumer people to see the hundreds of extra options
>>>> that making those menus visible will produce.
>>>>
>>>> This was added because in the past we had lots of issues with
>>>> people desktop/laptop settings with all those things enabled.
>>>>
>>>> In any case, if the desktop/laptop user is smart enough to
>>>> go though it, he can simply disable MEDIA_SUBDRV_AUTOSELECT and
>>>> manually select what he wants, so I really miss the point of
>>>> making those stuff always visible.
>>>>
>>>> Now, from this patch's comments, it seems that you want this
>>>> to be visible if CONFIG_EMBEDDED. So, I won't complain if you
>>>> replace the changes on this patch to:
>>>>
>>>>       menu "foo"
>>>>           visible if !MEDIA_SUBDRV_AUTOSELECT || !EMBEDDED || COMPILE_TEST || EXPERT
>>>>
>>>> In other words, for the normal guy that just wants to build the
>>>> latest media stuff for his PC camera or TV device to work, he won't
>>>> need to dig into hundreds of things that won't make any difference
>>>> if he enables, except for making the Kernel bigger.
>>>>  
>>>
>>> Well, I think the real value of MEDIA_SUBDRV_AUTOSELECT is the autoselection,
>>> not the hidden part. I'm really missing to see what hiding anything gives you.
>>>
>>> In other words, this option gets useful when driver authors select ancillary
>>> drivers such as:
>>>
>>> config VIDEO_USBVISION
>>>         tristate "USB video devices based on Nogatech NT1003/1004/1005"
>>>         depends on I2C && VIDEO_V4L2
>>>         select VIDEO_TUNER
>>>         select VIDEO_SAA711X if MEDIA_SUBDRV_AUTOSELECT
>>>
>>> What's so confusing about having these drivers visible? Compared to the
>>> rest of the zillion menu options, what's more confusing about seeing these?
>>>
>>> Now, while I would agree with EMBEDDED, the problem with that is that
>>> many "embedded" platforms don't enable EMBEDDED. So, it's not that useful.
>>>
>>> Finally, let me give an example of why hiding the menus is so bad.
>>> Normally, to enable a symbol, we use the search tool.
>>>
>>> Now, when MEDIA_SUBDRV_AUTOSELECT=y, the search tool will _not_ take you
>>> there and there's no indication why.  
>>
>> As someone who has done so in the past year, I agree it's confusing.
>> I had to dig through the Kconfig files to figure out which knobs to
>> turn to get the OV5640 option out. The description says "auto-selecting",

I had this same problem.

> 
> Well, the text and/or the help message can be changed, if it is not
> clear enough, but this option was added because we had too many issues
> with users trying to build drivers for their devices without being
> able to do that, because selecting thousands of devices is something
> that an average PC user has troubles.
> 
> I'm all to improve it, provided that we don't make harder for non-devs
> to build the Kernel.
> 
>> which does not equal hiding everything. You could still have drivers
>> auto-selected (or not) based on a Kconfig option without hiding things.
>>

Another idea is to separate in two options, MEDIA_SUBDRV_AUTOSELECT and
MEDIA_SUBDRV_ADVANCED, where the first one autoselects things and the later
enables more options.

Helen

>> ChenYu
> 
> 
> 
> Thanks,
> Mauro
> 

  reply	other threads:[~2019-07-25 19:00 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-07-15 21:23 [PATCH 0/2] Some cleanups for ancillary drivers autoselect Ezequiel Garcia
2019-07-15 21:23 ` [PATCH 1/2] media: Don't default-enable "ancillary driver autoselect" if EMBEDDED Ezequiel Garcia
2019-07-15 21:23 ` [PATCH 2/2] media: Don't hide any menu if "ancillary drivers autoselect" is enabled Ezequiel Garcia
2019-07-25 15:57   ` Mauro Carvalho Chehab
2019-07-25 16:01     ` Hans Verkuil
2019-07-25 17:05     ` Ezequiel Garcia
2019-07-25 17:29       ` Chen-Yu Tsai
2019-07-25 18:41         ` Mauro Carvalho Chehab
2019-07-25 19:00           ` Helen Koike [this message]
2019-07-25 19:16             ` Mauro Carvalho Chehab
2019-07-25 23:55           ` Ezequiel Garcia
2019-07-26  2:09             ` Mauro Carvalho Chehab
2019-07-26  2:54               ` Ezequiel Garcia
2019-07-27  0:17               ` Ezequiel Garcia
2019-07-27  1:39                 ` Mauro Carvalho Chehab
2019-07-25 17:45       ` Mauro Carvalho Chehab

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=4517f590-dcda-5bf0-70dd-3cffc5d71ba3@collabora.com \
    --to=helen.koike@collabora.com \
    --cc=ezequiel@collabora.com \
    --cc=hans.verkuil@cisco.com \
    --cc=kernel@collabora.com \
    --cc=linux-media@vger.kernel.org \
    --cc=mchehab+samsung@kernel.org \
    --cc=wens@kernel.org \
    /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).