Date: Tue, 5 May 2015 15:04:05 +0200
From: Nicholas Mc Guire <der.herr@hofr.at>
To: SF Markus Elfring <elfring@users.sourceforge.net>
Cc: Nicholas Mc Guire <hofrat@osadl.org>, cocci@systeme.lip6.fr,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH RFC] Coccinelle: Check for return not matching function
	signature
Message-ID: <20150505130405.GA1956@opentech.at>
References: <1430820761-28122-1-git-send-email-hofrat@osadl.org> <5548B87B.4060103@users.sourceforge.net>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <5548B87B.4060103@users.sourceforge.net>
User-Agent: Mutt/1.5.18 (2008-05-17)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1760
Lines: 63

On Tue, 05 May 2015, SF Markus Elfring wrote:

> > Check if the signature of a function and the return value type match.
> 
> Is this a task that is usually performed by a compiler?
> 
> 
> > In many cases this mismatch will have no side-effects
> > but in some cases it may lead to hard to locate problems
> 
> It is another software development challenge to find concrete
> open issues there.
> 
> 
> > - and for readability and code understanding it is also helpful
> > when types match.
> 
> How would you like to check for compatible data types here?
>

coccinelle knows the type so all you need to do is comare them in
phython .
 
> 
> > The output is a bit lengthy - not sure if that is too much
> > but it seemed useful to me to see the non-matching types explicitly
> > in the warning message.
> 
> How do you think about to import the result list into a database table?
> 

working on that "re-cycling" your parameter count example
top 10:
    488 ssize_t != int
    195 int != unsigned int
    183 long != int
    113 int != u32
     55 int != unsigned long
     48 int != u8
     45 int != u16
     44 unsigned int != int
     37 int != s32
     30 int != long


> 
> > +if T1 != T2:
> > +   print "%s:%s,%s WARNING: return of wrong type (%s != %s)" % (p[0].file,fn,p[0].line,T1,T2)
> 
> Is such a check a bit too simple?
>
Nop - why ?
Cocci "knwow" C so the assignment of types is reliable - 
flaging a s32 != int is fine with respect to readability
even if they are technically the same.

thx!
hofrat
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/