Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.37 as permitted sender) client-ip=23.128.96.37;
Date:   Thu, 2 Nov 2023 15:09:57 +0000
From:   Sakari Ailus <sakari.ailus@linux.intel.com>
To:     Hans de Goede <hdegoede@redhat.com>
Cc:     Alexandra Diupina <adiupina@astralinux.ru>,
        Mauro Carvalho Chehab <mchehab@kernel.org>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Andy Shevchenko <andy@kernel.org>,
        Uwe =?iso-8859-1?Q?Kleine-K=F6nig?= 
        <u.kleine-koenig@pengutronix.de>, linux-media@vger.kernel.org,
        linux-staging@lists.linux.dev, linux-kernel@vger.kernel.org,
        lvc-project@linuxtesting.org
Subject: Re: [PATCH 1/4] Remove redundant return value check
Message-ID: <ZUO7xUuZmIVp06sw@kekkonen.localdomain>
References: <20231102141135.369-1-adiupina@astralinux.ru>
 <b2aa6ccb-0a9a-449f-1341-fd71d5c545fd@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <b2aa6ccb-0a9a-449f-1341-fd71d5c545fd@redhat.com>
Precedence: bulk

Hi Hans, Alexandra,

On Thu, Nov 02, 2023 at 03:21:04PM +0100, Hans de Goede wrote:
> Hi Alexandra,
> 
> On 11/2/23 15:11, Alexandra Diupina wrote:
> > media_entity_pads_init() will not return 0 only if the
> > 2nd parameter >= MEDIA_ENTITY_MAX_PADS (512), but 1 is
> > passed, so checking the return value is redundant
> 
> Generally speaking functions which can fail should always be
> error-checked even if their invocation today happen to be
> such that they will not fail.
> 
> Either the invocation or the function itself my change
> causing it to fail in the future. Which is why we want
> to keep the error checks.
> 
> But maybe media_entity_pads_init() is special and
> does not need to be error checked.
> 
> Sakari, Laurent do you have any opinion on this ?
> 
> > Found by Linux Verification Center (linuxtesting.org) with SVACE.
> 
> This feels like a false positive of the tool to me,
> but lets wait and see what Sakari or Laurent have
> to say.

I agree with Hans: this function today may not fail with the parameters
passed to it but it may happen in the future. In general it's good to check
a return value of a function that returns one: if that function is changed,
no-one will go through the callers as long as the arguments and the return
value remain the same.

-- 
Regards,

Sakari Ailus