Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <20200624114127.3016-1-a.hajda@samsung.com> <CGME20200624114136eucas1p1c84f81b1d78e2dbad7ac1b762f0a4b4f@eucas1p1.samsung.com>
 <20200624114127.3016-4-a.hajda@samsung.com> <2203e0c2-016b-4dbe-452d-63c857f06dd1@arm.com>
In-Reply-To: <2203e0c2-016b-4dbe-452d-63c857f06dd1@arm.com>
From:   Andy Shevchenko <andy.shevchenko@gmail.com>
Date:   Wed, 24 Jun 2020 15:55:57 +0300
Message-ID: <CAHp75VfpP1cGK3FvTL0hBudRY2N_7GpXYRuUHUCipz7X2sMLmQ@mail.gmail.com>
Subject: Re: [RESEND PATCH v5 3/5] drivers core: allow probe_err accept
 integer and pointer types
To:     Robin Murphy <robin.murphy@arm.com>
Cc:     Andrzej Hajda <a.hajda@samsung.com>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Jernej Skrabec <jernej.skrabec@siol.net>,
        "Rafael J. Wysocki" <rafael@kernel.org>,
        Jonas Karlman <jonas@kwiboo.se>,
        Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        "open list:DRM DRIVERS" <dri-devel@lists.freedesktop.org>,
        Russell King - ARM Linux <linux@armlinux.org.uk>,
        Neil Armstrong <narmstrong@baylibre.com>,
        Mark Brown <broonie@kernel.org>,
        Laurent Pinchart <Laurent.pinchart@ideasonboard.com>,
        Daniel Vetter <daniel@ffwll.ch>,
        linux-arm Mailing List <linux-arm-kernel@lists.infradead.org>,
        Marek Szyprowski <m.szyprowski@samsung.com>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Wed, Jun 24, 2020 at 3:37 PM Robin Murphy <robin.murphy@arm.com> wrote:
> On 2020-06-24 12:41, Andrzej Hajda wrote:
> > Many resource acquisition functions return error value encapsulated in
> > pointer instead of integer value. To simplify coding we can use macro
> > which will accept both types of error.
> > With this patch user can use:
> >       probe_err(dev, ptr, ...)
> > instead of:
> >       probe_err(dev, PTR_ERR(ptr), ...)
> > Without loosing old functionality:
> >       probe_err(dev, err, ...)
>
> Personally I'm not convinced that simplification has much value, and I'd
> say it *does* have a significant downside. This:
>
>         if (IS_ERR(x))
>                 do_something_with(PTR_ERR(x));
>
> is a familiar and expected pattern when reading/reviewing code, and at a
> glance is almost certainly doing the right thing. If I see this, on the
> other hand:
>
>         if (IS_ERR(x))
>                 do_something_with(x);

I don't consider your arguments strong enough. You are appealing to
one pattern vs. new coming *pattern* just with a different name and
actually much less characters to parse. We have a lot of clean ups
like this, have you protested against them? JFYI: they are now
*established patterns* and saved a ton of LOCs in some of which even
were typos.

> my immediate instinct is to be suspicious, and now I've got to go off
> and double-check that if do_something_with() really expects a pointer
> it's also robust against PTR_ERR values. Off-hand I can't think of any
> APIs that work that way in the areas with which I'm familiar, so it
> would be a pretty unusual and non-obvious thing.

-- 
With Best Regards,
Andy Shevchenko