linux-kernel - Re: [PATCH] iio: health/afe4403: mark suspend/resume functions __maybe

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Mon, 15 Feb 2016 10:20:49 +0100
From:	Arnd Bergmann <arnd@...db.de>
To:	Daniel Baluta <daniel.baluta@...il.com>
Cc:	Jonathan Cameron <jic23@...nel.org>,
	linux-arm-kernel@...ts.infradead.org,
	"Andrew F . Davis" <afd@...com>, Hartmut Knaack <knaack.h@....de>,
	Lars-Peter Clausen <lars@...afoo.de>,
	Peter Meerwald <pmeerw@...erw.net>,
	"linux-iio@...r.kernel.org" <linux-iio@...r.kernel.org>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] iio: health/afe4403: mark suspend/resume functions __maybe_unused

On Monday 15 February 2016 11:18:20 Daniel Baluta wrote:
> On Mon, Feb 15, 2016 at 11:02 AM, Arnd Bergmann <arnd@...db.de> wrote:
> > The newly added afe4403 driver implements suspend/resume using the
> > SIMPLE_DEV_PM_OPS() macro, which leaves out references to the actual
> > functions when CONFIG_PM is disabled, causing a harmless warning:
> >
> > health/afe4403.c:509:12: error: 'afe4403_suspend' defined but not used
> > health/afe4403.c:530:12: error: 'afe4403_resume' defined but not used
> >
> > This marks the functions as __maybe_unused so we don't get those
> > warnings.
> >
> Or we could put _suspend and _resume under CONFIG_PM_SLEEP.

That never works reliably. I still have a backlog of ~50 patches
of drivers that got this wrong in the past.

> We have:
> 
> :~/w/iio/drivers/iio$ ack-grep SIMPLE_DEV_PM_OPS | wc -l
> 31
> 
> and only 2 drivers using __maybe_unused.
> 
> :~/w/iio/drivers/iio$ ack-grep maybe_unused
> gyro/itg3200_core.c
> 354:static int __maybe_unused itg3200_suspend(struct device *dev)
> 365:static int __maybe_unused itg3200_resume(struct device *dev)
> 
> adc/imx7d_adc.c
> 553:static int __maybe_unused imx7d_adc_suspend(struct device *dev)
> 566:static int __maybe_unused imx7d_adc_resume(struct device *dev)
> 
> For consistency, we should use CONFIG_PM_SLEEP and get rid of __maybe_unused.

I hope to eventually kill off all the CONFIG_PM_SLEEP checks and then
make it work automatically by redefining SIMPLE_DEV_PM_OPS so it
leaves an unused reference to the functions.

	Arnd