[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20210711044419.7b1c00f8@gmail.com>
Date: Sun, 11 Jul 2021 04:44:19 +0000
From: Vincent Pelletier <plr.vincent@...il.com>
To: Guenter Roeck <linux@...ck-us.net>
Cc: Jean Delvare <jdelvare@...e.com>, Jonathan Corbet <corbet@....net>,
Support Opensource <support.opensource@...semi.com>,
Lee Jones <lee.jones@...aro.org>, linux-hwmon@...r.kernel.org,
linux-doc@...r.kernel.org, linux-kernel@...r.kernel.org,
"Opensource [Steve Twiss]" <stwiss.opensource@...semi.com>
Subject: Re: [PATCH v3 2/3] hwmon: da9063: HWMON driver
On Sun, 11 Jul 2021 02:55:02 +0000, Vincent Pelletier <plr.vincent@...il.com> wrote:
> On Sat, 10 Jul 2021 09:08:13 -0700, Guenter Roeck <linux@...ck-us.net> wrote:
> > Unnecessary include.
> [...]
> > I don't immediately see where this include is needed. Is this a
> > leftover ?
> [...]
> > Same here.
>
> Are there ways to systematically tell which includes are useless
> besides commenting them out all and uncommenting until it compiles ?
> (if that is even a good idea)
I tried this, just to get a baseline: the module compiles with just
linux/hwmon.h
linux/mfd/da9063/core.h
linux/module.h
linux/platform_device.h
linux/regmap.h
Beyond what you suggested this also gets rid of:
- seems reasonable:
- linux/delay.h
- linux/init.h
- linux/slab.h
- looks suspicious to me:
- linux/err.h, which means the error constants are indirectly
imported. Removing it feels brittle.
- linux/kernel.h, although to my surprise a lot of c files do not
include it.
By default I'll drop the former and keep the latter in the
next version, please let me know if another combination is preferred.
--
Vincent Pelletier
GPG fingerprint 983A E8B7 3B91 1598 7A92 3845 CAC9 3691 4257 B0C1
Powered by blists - more mailing lists