Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Sun, 17 Feb 2019 12:37:18 +0100 (CET)
From:   Julia Lawall <julia.lawall@lip6.fr>
To:     Markus Elfring <Markus.Elfring@web.de>
cc:     Wen Yang <yellowriver2010@hotmail.com>,
        Gilles Muller <Gilles.Muller@lip6.fr>,
        Nicolas Palix <nicolas.palix@imag.fr>,
        Michal Marek <michal.lkml@markovi.net>,
        Masahiro Yamada <yamada.masahiro@socionext.com>,
        Wen Yang <wen.yang99@zte.com.cn>,
        Cheng Shengyu <cheng.shengyu@zte.com.cn>,
        kernel-janitors@vger.kernel.org,
        LKML <linux-kernel@vger.kernel.org>,
        Coccinelle <cocci@systeme.lip6.fr>
Subject: Re: [PATCH v6] coccinelle: semantic code search for missing
 put_device()
In-Reply-To: <8e7ba7c0-b7fe-a1f0-d28b-0c716ecbcfdb@web.de>
Message-ID: <alpine.DEB.2.21.1902171235570.2444@hadrien>
References: <HK0PR02MB36344E2B29CEB195892F6420B2610@HK0PR02MB3634.apcprd02.prod.outlook.com> <8e7ba7c0-b7fe-a1f0-d28b-0c716ecbcfdb@web.de>
User-Agent: Alpine 2.21 (DEB 202 2017-01-01)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


On Sun, 17 Feb 2019, Markus Elfring wrote:

> > +@search exists@
> > +local idexpression id;
> > +expression x,e,e1;
> > +position p1,p2;
> > +type T,T1,T2;
> > +@@
> > +
> > +id = of_find_device_by_node@p1(x)
> > +... when != e = id
>
> I suggest to increase your software development attention also for
> another implementation detail.
> Source code analysis triggers challenges for safe data flow handling.
> the semantic patch language supports search specifications for
> the exclusion of specific assignments.
>
> Does this SmPL code contain a questionable order for the source
> and target metavariables?
> Can the following variant be more appropriate?
>
> + ... when != id = e

This is possible, but I think unlikely.

>
>
> > +if (id == NULL || ...) { ... return ...; }
> > +... when != put_device(&id->dev)
> > +    when != platform_device_put(id)
> > +    when != of_dev_put(id)
> > +    when != if (id) { ... put_device(&id->dev) ... }
> > +    when != e1 = (T)id
>
> Would you like to avoid that the return value from the shown function call
> gets overwritten in the variable before it was used once at least
> (when a bit of extra C code is tolerated before a null pointer check)?

Indeed there should be a put then too, but again, it seems unlikely.

julia


>
> Regards,
> Markus
>