Re: [PATCH] coccinelle: api: detect unnecessary le16_to_cpu

From: Sebastian Reichel <sebastian.reichel@collabora.co.uk>
To: Andy Shevchenko <andy.shevchenko@gmail.com>
Cc: Julia Lawall <Julia.Lawall@lip6.fr>,
	kernel-janitors@vger.kernel.org,
	Gilles Muller <Gilles.Muller@lip6.fr>,
	Nicolas Palix <nicolas.palix@imag.fr>,
	Michal Marek <mmarek@suse.com>,
	cocci@systeme.lip6.fr,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	Benjamin Tissoires <benjamin.tissoires@redhat.com>,
	Bastien Nocera <hadess@hadess.net>,
	Stephen Just <stephenjust@gmail.com>,
	"Rafael J . Wysocki" <rafael.j.wysocki@intel.com>,
	Len Brown <lenb@kernel.org>,
	Robert Moore <robert.moore@intel.com>,
	Lv Zheng <lv.zheng@intel.com>,
	Mika Westerberg <mika.westerberg@linux.intel.com>,
	"linux-acpi@vger.kernel.org" <linux-acpi@vger.kernel.org>,
	devel@acpica.org,
	"linux-pm@vger.kernel.org" <linux-pm@vger.kernel.org>
Subject: Re: [PATCH] coccinelle: api: detect unnecessary le16_to_cpu
Date: Mon, 3 Jul 2017 20:20:34 +0200	[thread overview]
Message-ID: <20170703182033.5stv5eq33et63coh@earth> (raw)
In-Reply-To: <CAHp75VeLV_ySCRN3B=MSoZdyrbXXqqEXrgC+hJCDoLLCJRComQ@mail.gmail.com>

[-- Attachment #1: Type: text/plain, Size: 1158 bytes --]

Hi,

On Mon, Jul 03, 2017 at 08:33:53PM +0300, Andy Shevchenko wrote:
> On Mon, Jul 3, 2017 at 8:14 PM, Sebastian Reichel
> <sebastian.reichel@collabora.co.uk> wrote:
> > On Mon, Jul 03, 2017 at 07:37:59PM +0300, Andy Shevchenko wrote:
> >> On Mon, Jul 3, 2017 at 4:36 PM, Sebastian Reichel
> >> <sebastian.reichel@collabora.co.uk> wrote:
> >> > On Sat, Jul 01, 2017 at 09:28:10PM +0200, Julia Lawall wrote:
> >>
> >> >  * drivers/gpio/gpio-pca953x.c (line 190-192)
> >>
> >> It has double conversion there:
> >> 1. LE CPU: Read as LE and converted to LE (no-op), so, just u16
> >> 2. BE CPU: Read as BE and converted to LE, makes it __le16
> >>
> >> Looks like the conversion is not needed, only get_unaligned() is necessary.
> >>
> >> P.S. What about lines 244-245 there? I think they are no-op.
> >> Interesting that those two parts were added in quite different
> >> commits.
> >
> > val[0] = (u16)ret & 0xFF;
> > val[1] = (u16)ret >> 8;
> >
> > looks like cpu_to_be16()?
> 
> cpu_to_le16(). No-op on LE CPU.

uhm yes of course.

> Perhaps they should be replaced by put_unaligned().

Makes sense to me.

-- Sebastian

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]