linux-kernel - Re: [bug report] checkpatch: if statement does not need to be enclosed in parentheses

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: <20170409085531.GA26418@rhlx01.hs-esslingen.de>
Date:   Sun, 9 Apr 2017 10:55:31 +0200
From:   Andreas Mohr <andi@...as.de>
To:     Joe Perches <joe@...ches.com>
Cc:     Alfonso Lima <alfonsolimaastor@...il.com>, apw@...onical.com,
        linux-kernel@...r.kernel.org
Subject: Re: [bug report] checkpatch: if statement does not need to be
 enclosed in parentheses

On Sat, Apr 08, 2017 at 10:33:19AM -0700, Joe Perches wrote:
> On Sat, 2017-04-08 at 17:07 +0100, Alfonso Lima wrote:
> > Hi,
> > 
> > In current linux-next, if you run the script on this piece of
> > code:
> > 
> > #define FOO(a)				\
> > 	if (a) {			\
> > 		something();		\
> > 		something_else();	\
> > 	}
> > 
> > You should get:
> > 
> > ERROR: Macros with complex values should be enclosed in parentheses
> > #1: FILE: foo.c:1:
> > +#define FOO(a)				\
> > +	if (a) {			\
> > +		something();		\
> > +		something_else();	\
> > +	}
> > 
> > We could silence checkpatch.pl using "do {} while ()" around the
> > if statement. However, the "if () {}" statement should be
> > enough. If someone could confirm this, I'll go and fix it.
> 
> Multiple if/else use is the reason do {} while (0) is suggested.
> 
> 	if (bar())
> 		FOO(a);
> 	else
> 		baz(b);
>  

So, if I read this right checkpatch.pl is generating a correct complaint
in this case after all?

However, then we seem to have the following issues remaining:
- "Macros with complex values should be enclosed in parentheses" != "do {} while (0) is suggested" --> this is an imprecision which may cause confusion
- *further* cases of this issue occurring may just as well suggest again
  "fixing" (i.e., "damaging") checkpatch.pl rather than
  simply obeying (following through on) its correct diagnosis
  --> checkpatch.pl behaviour is confusing enough that
  it may prompt incorrect core handling "repair actions",
  and this danger will remain permanently until
  checkpatch.pl reporting is improved to be
  less confusing/insufficient about its diagnosis

Or perhaps this imprecision here
again simply is due to
the very frequently observable case of
saying "thou shan't do this" without then also
giving some tried-and-true *reason(s)* for such advice.
("ERROR: Macros with complex values should be enclosed in parentheses")

HTH,

Andreas Mohr