Received-SPF: pass (google.com: domain of linux-kernel+bounces-2320-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
Message-ID: <59d9ac4e1f7475f51672222af1fc90d7ec23b5da.camel@perches.com>
Subject: Re: [PATCH v4 01/15] scripts: checkpatch: Add __aligned to the list
 of attribute notes
From: Joe Perches <joe@perches.com>
To: Marcelo Schmitt <marcelo.schmitt@analog.com>, apw@canonical.com, 
 dwaipayanray1@gmail.com, lukas.bulwahn@gmail.com, paul.cercueil@analog.com,
  Michael.Hennerich@analog.com, lars@metafoo.de, jic23@kernel.org, 
 robh+dt@kernel.org, krzysztof.kozlowski+dt@linaro.org, conor+dt@kernel.org,
  dan.carpenter@linaro.org, dlechner@baylibre.com,
 marcelo.schmitt1@gmail.com,  Andrew Morton <akpm@linux-foundation.org>
Cc: linux-iio@vger.kernel.org, devicetree@vger.kernel.org, 
	linux-kernel@vger.kernel.org
Date: Sat, 16 Dec 2023 10:07:24 -0800
In-Reply-To: <1c5c93ecbd8c46a338b22a4ef52e51648e333c01.1702746240.git.marcelo.schmitt1@gmail.com>
References: <cover.1702746240.git.marcelo.schmitt1@gmail.com>
	 <1c5c93ecbd8c46a338b22a4ef52e51648e333c01.1702746240.git.marcelo.schmitt1@gmail.com>
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.48.4 (3.48.4-1.fc38) 
Precedence: bulk
MIME-Version: 1.0

On Sat, 2023-12-16 at 14:45 -0300, Marcelo Schmitt wrote:
> Checkpatch presumes attributes marked with __aligned(alignment) are part
> of a function declaration and throws a warning stating that those
> compiler attributes should have an identifier name which is not correct.
> Add __aligned compiler attributes to the list of attribute notes
> so they don't cause warnings anymore.
>=20
> Signed-off-by: Marcelo Schmitt <marcelo.schmitt@analog.com>
> ---
> Any expression that evaluates to an integer that is a power of 2 can be
> within __aligned parenthesis.
> I can't see how we could use a regex to check code meets such constraint =
(if possible).

You can't because if a #define is used for the alignment
value it is not necessarily available to a patch fragment
or even a file if the #define is in an #include.

> Some additional exotic uses of __aligned are:
> drivers/net/wireless/quantenna/qtnfmac/bus.h:72:   char bus_priv[] __alig=
ned(sizeof(void *));
> include/linux/netdevice.h:225:} __aligned(4 * sizeof(unsigned long));=20

Right, then there are random uses like:

drivers/firmware/qcom/qcom_qseecom_uefisecapp.c:		size_t __aligned;						\
drivers/firmware/qcom/qcom_qseecom_uefisecapp.c:			*__offset =3D __aligned;=
					\

so there's always some false positive/negative issue
with checkpatch.

Anyway:

Acked-by: Joe Perches <joe@perches.com>

>=20
> The regex
> 			__aligned\s*\(.*\)
> seems to match all use cases.=20
>=20
> We might not catch invalid arguments to __aligned, but it looks like maki=
ng
> checkpath confidently report those wouldn't be feasible anyway.
>=20
> The patch that would trigger the mentioned warning in now
> patch number 13 (iio: adc: Add support for AD7091R-8).
>=20
>  scripts/checkpatch.pl | 1 +
>  1 file changed, 1 insertion(+)
>=20
> diff --git a/scripts/checkpatch.pl b/scripts/checkpatch.pl
> index 25fdb7fda112..d56c98146da3 100755
> --- a/scripts/checkpatch.pl
> +++ b/scripts/checkpatch.pl
> @@ -512,6 +512,7 @@ our $Attribute	=3D qr{
>  			__ro_after_init|
>  			__kprobes|
>  			$InitAttribute|
> +			__aligned\s*\(.*\)|
>  			____cacheline_aligned|
>  			____cacheline_aligned_in_smp|
>  			____cacheline_internodealigned_in_smp|