linux-kernel - Re: [PATCH v1 2/2] mfd: intel_soc_pmic_chtwc: Switch from __maybe_unused to pm_sleep

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: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <YrBAbYsSB3SZPQvt@smile.fi.intel.com>
Date:   Mon, 20 Jun 2022 12:39:57 +0300
From:   Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
To:     Hans de Goede <hdegoede@...hat.com>
Cc:     linux-kernel@...r.kernel.org, Andy Shevchenko <andy@...nel.org>,
        Lee Jones <lee.jones@...aro.org>
Subject: Re: [PATCH v1 2/2] mfd: intel_soc_pmic_chtwc: Switch from
 __maybe_unused to pm_sleep_ptr() etc

On Mon, Jun 20, 2022 at 10:38:34AM +0200, Hans de Goede wrote:
> On 6/16/22 19:10, Andy Shevchenko wrote:
> > Letting the compiler remove these functions when the kernel is built
> > without CONFIG_PM_SLEEP support is simpler and less heavier for builds
> > than the use of __maybe_unused attributes.
> > 
> > Signed-off-by: Andy Shevchenko <andriy.shevchenko@...ux.intel.com>
> 
> I don't see how this helps, __maybe_unused only means the compiler should not
> warn AFAIK it is still allowed to remove static globals which are unused
> even of they are marked __maybe_unused ?

__maybe_unused will force compiler to compile and linker to drop the section.
The proposed approach saves resources on build, i.e. it's already compiler that
decides not to compile the code.

> And DEFINE_SIMPLE_DEV_PM_OPS does not mark the dev_pm_ops as
> __maybe_unused where as pm_sleep_ptr() will not reference the struct
> when CONFIG_PM is not set.
> 
> But I guess that the new PTR_IF magic:
> 
> #define PTR_IF(cond, ptr)       ((cond) ? (ptr) : NULL)
> 
> Still counts as referencing the struct so we don't get
> an unused warning and since cond is a const 0 value the compiler
> can optimize things away ?

See above, it's not olny about warning and compiler.


-- 
With Best Regards,
Andy Shevchenko