[PATCH] iio: magnetometer: Only declare ACPI table when ACPI is enable

[PATCH] iio: magnetometer: Only declare ACPI table when ACPI is enable

[Matthias Kaehlcke]

Don't inflate the kernel size with data that isn't used. The conditional
declaration also fixes the following warning when building with clang:

drivers/iio/magnetometer/ak8975.c:704:36: error: variable 'ak_acpi_match'
  is not needed and will not be emitted
  [-Werror,-Wunneeded-internal-declaration]

Signed-off-by: Matthias Kaehlcke <mka@chromium.org>
---
 drivers/iio/magnetometer/ak8975.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/iio/magnetometer/ak8975.c b/drivers/iio/magnetometer/ak8975.c
index 825369fb1c57..4ff883942f7b 100644
--- a/drivers/iio/magnetometer/ak8975.c
+++ b/drivers/iio/magnetometer/ak8975.c
@@ -784,6 +784,7 @@ static const struct iio_info ak8975_info = {
 	.driver_module = THIS_MODULE,
 };
 
+#ifdef CONFIG_ACPI
 static const struct acpi_device_id ak_acpi_match[] = {
 	{"AK8975", AK8975},
 	{"AK8963", AK8963},
@@ -793,6 +794,7 @@ static const struct acpi_device_id ak_acpi_match[] = {
 	{ },
 };
 MODULE_DEVICE_TABLE(acpi, ak_acpi_match);
+#endif
 
 static const char *ak8975_match_acpi_device(struct device *dev,
 					    enum asahi_compass_chipset *chipset)
-- 
2.13.1.611.g7e3b11ae1-goog

Re: [PATCH] iio: magnetometer: Only declare ACPI table when ACPI is enable

[Linus Walleij]

On Tue, Jun 27, 2017 at 2:25 AM, Matthias Kaehlcke <mka@chromium.org> wrote:

> Don't inflate the kernel size with data that isn't used. The conditional
> declaration also fixes the following warning when building with clang:
>
> drivers/iio/magnetometer/ak8975.c:704:36: error: variable 'ak_acpi_match'
>   is not needed and will not be emitted
>   [-Werror,-Wunneeded-internal-declaration]
>
> Signed-off-by: Matthias Kaehlcke <mka@chromium.org>

Acked-by: Linus Walleij <linus.walleij@linaro.org>

Yours,
Linus Walleij

Re: [PATCH] iio: magnetometer: Only declare ACPI table when ACPI is enable

[Jonathan Cameron]

On Thu, 29 Jun 2017 14:51:56 +0200
Linus Walleij <linus.walleij@linaro.org> wrote:

> On Tue, Jun 27, 2017 at 2:25 AM, Matthias Kaehlcke <mka@chromium.org> wrote:
> 
> > Don't inflate the kernel size with data that isn't used. The conditional
> > declaration also fixes the following warning when building with clang:
> >
> > drivers/iio/magnetometer/ak8975.c:704:36: error: variable 'ak_acpi_match'
> >   is not needed and will not be emitted
> >   [-Werror,-Wunneeded-internal-declaration]
> >
> > Signed-off-by: Matthias Kaehlcke <mka@chromium.org>  
> 
> Acked-by: Linus Walleij <linus.walleij@linaro.org>
> 
> Yours,
> Linus Walleij
Applied to the togreg branch of iio.git and pushed out as testing
for the autobuilders to play with it.

I'm curious though... Can the compiler not optimize this away? 
I'd always assumed the main reason for these warnings was that the
presence of unused data might represent a bug rather than it actually
making any difference to the size...

Re: [PATCH] iio: magnetometer: Only declare ACPI table when ACPI is enable

[Matthias Kaehlcke]

El Sat, Jul 01, 2017 at 01:11:12PM +0100 Jonathan Cameron ha dit:

> On Thu, 29 Jun 2017 14:51:56 +0200
> Linus Walleij <linus.walleij@linaro.org> wrote:
> 
> > On Tue, Jun 27, 2017 at 2:25 AM, Matthias Kaehlcke <mka@chromium.org> wrote:
> > 
> > > Don't inflate the kernel size with data that isn't used. The conditional
> > > declaration also fixes the following warning when building with clang:
> > >
> > > drivers/iio/magnetometer/ak8975.c:704:36: error: variable 'ak_acpi_match'
> > >   is not needed and will not be emitted
> > >   [-Werror,-Wunneeded-internal-declaration]
> > >
> > > Signed-off-by: Matthias Kaehlcke <mka@chromium.org>  
> > 
> > Acked-by: Linus Walleij <linus.walleij@linaro.org>
> > 
> > Yours,
> > Linus Walleij
> Applied to the togreg branch of iio.git and pushed out as testing
> for the autobuilders to play with it.

Thanks!

> I'm curious though... Can the compiler not optimize this away? 
> I'd always assumed the main reason for these warnings was that the
> presence of unused data might represent a bug rather than it actually
> making any difference to the size...

clang actually optimizes it away: "variable 'ak_acpi_match' is not
needed and will not be emitted", my argument about inflating the
kernel size is not valid.