linux-kernel - Re: [PATCH] sched/cputime: Mark function as __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: Windows password security audit tool. GUI, reports in PDF.

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

Message-ID: <20200819073253.GV2674@hirez.programming.kicks-ass.net>
Date:   Wed, 19 Aug 2020 09:32:53 +0200
From:   peterz@...radead.org
To:     Nick Desaulniers <ndesaulniers@...gle.com>
Cc:     Alex Dewar <alex.dewar90@...il.com>,
        Ingo Molnar <mingo@...hat.com>,
        Juri Lelli <juri.lelli@...hat.com>,
        Vincent Guittot <vincent.guittot@...aro.org>,
        Dietmar Eggemann <dietmar.eggemann@....com>,
        Steven Rostedt <rostedt@...dmis.org>,
        Ben Segall <bsegall@...gle.com>, Mel Gorman <mgorman@...e.de>,
        LKML <linux-kernel@...r.kernel.org>,
        clang-built-linux <clang-built-linux@...glegroups.com>
Subject: Re: [PATCH] sched/cputime: Mark function as __maybe_unused

On Tue, Aug 18, 2020 at 01:02:26PM -0700, Nick Desaulniers wrote:
> On Tue, Aug 18, 2020 at 12:57 PM Alex Dewar <alex.dewar90@...il.com> wrote:
> >
> > On Tue, Aug 18, 2020 at 11:13:10AM -0700, Nick Desaulniers wrote:
> > > On Tue, Aug 18, 2020 at 10:04 AM Alex Dewar <alex.dewar90@...il.com> wrote:
> > > >
> > > > Depending on config options, account_other_time() may not be called
> > > > anywhere. Add __maybe_unused flag to fix clang warning.
> > >
> > > Just curious, would moving this definition to be within an existing
> > > preprocessor guard for a particular config also fix the issue? If so,
> > > prefer that. If not, __maybe_unused is the way to go.
> >
> > I don't think that'd work here: it's used within an "#ifdef
> > CONFIG_IRQ_TIME_ACCOUNTING" block and a separate "#ifdef
> > CONFIG_VIRT_CPU_ACCOUNTING_GEN" one. We could do:
> >         #if defined(CONFIG_IRQ_TIME_ACCOUNTING) ||
> >             defined(CONFIG_VIRT_CPU_ACCOUNTING)
> >                 ...
> > ... but that might be a bit ugly.
> 
> Yeah, ok, in that case it's fine.  One issue with __maybe_unused is
> that this function will stick around forever if all call sites get
> removed.  But when the preprocessor checks start getting hairy,
> __maybe_unused is maybe simpler.

Urgh, also, I have vague memories that we made it 'static inline' to
avoid all this. Should we instead mark it __always_inline ?

Also; I detest the placement of __maybe_unused, it is not a return
value, but a git-grep showed me it is wide-spread :/