Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Thu, 13 Apr 2023 08:32:41 -0700
From:   Josh Poimboeuf <jpoimboe@kernel.org>
To:     Miroslav Benes <mbenes@suse.cz>
Cc:     x86@kernel.org, linux-kernel@vger.kernel.org,
        Peter Zijlstra <peterz@infradead.org>
Subject: Re: [PATCH v2 8/9] objtool: Detect missing __noreturn annotations
Message-ID: <20230413153241.7px5x2w6jvkpmes7@treble>
References: <cover.1681325924.git.jpoimboe@kernel.org>
 <0f630a0eb4585ab4114e4eecaa6f166a1fd81d49.1681325924.git.jpoimboe@kernel.org>
 <alpine.LSU.2.21.2304131613490.27633@pobox.suse.cz>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
In-Reply-To: <alpine.LSU.2.21.2304131613490.27633@pobox.suse.cz>
Precedence: bulk

On Thu, Apr 13, 2023 at 04:19:10PM +0200, Miroslav Benes wrote:
> > --- a/tools/objtool/check.c
> > +++ b/tools/objtool/check.c
> > @@ -4485,7 +4485,8 @@ static int validate_sls(struct objtool_file *file)
> >  
> >  static int validate_reachable_instructions(struct objtool_file *file)
> >  {
> > -	struct instruction *insn;
> > +	struct instruction *insn, *prev_insn;
> > +	struct symbol *call_dest;
> >  	int warnings = 0;
> >  
> >  	if (file->ignore_unreachables)
> > @@ -4495,6 +4496,17 @@ static int validate_reachable_instructions(struct objtool_file *file)
> >  		if (insn->visited || ignore_unreachable_insn(file, insn))
> >  			continue;
> >  
> > +		prev_insn = prev_insn_same_sec(file, insn);
> > +		if (prev_insn && prev_insn->dead_end) {
> > +			call_dest = insn_call_dest(prev_insn);
> > +			if (call_dest) {
> > +				WARN_INSN(insn, "%s() is missing a __noreturn annotation",
> > +					  call_dest->name);
> > +				warnings++;
> > +				continue;
> 
> A nit but this and
> 
> > +			}
> > +		}
> > +
> >  		WARN_INSN(insn, "unreachable instruction");
> >  		warnings++;
> 
> this makes me thinking. Wouldn't it be confusing to anyone that there is 
> no correspondence between warnings and a number of actual reported 
> warnings through WARN_INSN()? In the future when there would be a usage 
> for warnings. It does not really matter now.

True, maybe we need WARN_INSN_ONCE_PER_FUNC() or so ;-)

-- 
Josh