[PATCH] ssb: mark ssb_bus_register as __maybe_unused

[PATCH] ssb: mark ssb_bus_register as __maybe_unused

[Arnd Bergmann]

The SoC variant of the ssb code is now optional like the other
ones, which means we can build the framwork without any
front-end, but that results in a warning:

drivers/ssb/main.c:616:12: warning: 'ssb_bus_register' defined but not used [-Wunused-function]

This annotates the ssb_bus_register function as __maybe_unused to
shut up the warning. A configuration like this will not work on
any hardware of course, but we still want this to silently build
without warnings if the configuration is allowed in the first
place.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Fixes: 845da6e58e19 ("ssb: add Kconfig entry for compiling SoC related code")
----
This showed up on ARM at91_defconfig and sama5_defconfig, so we should probably
change those configurations as well. I can't see why they enable CONFIG_SSB
when they do not use any drivers for this framework.

diff --git a/drivers/ssb/main.c b/drivers/ssb/main.c
index 5d1e9a0fc389..e2ff6b5b2094 100644
--- a/drivers/ssb/main.c
+++ b/drivers/ssb/main.c
@@ -613,9 +613,10 @@ out:
 	return err;
 }
 
-static int ssb_bus_register(struct ssb_bus *bus,
-			    ssb_invariants_func_t get_invariants,
-			    unsigned long baseaddr)
+static int __maybe_unused
+ssb_bus_register(struct ssb_bus *bus,
+		 ssb_invariants_func_t get_invariants,
+		 unsigned long baseaddr)
 {
 	int err;
 

[PATCH] ARM: at91/defconfig: remove CONFIG_SSB from Atmel defconfigs

[Nicolas Ferre]

This "Sonics Silicon Backplane" support is not needed on Atmel SoCs: remove it.

Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
---
Arnd,

Here is the patch that fixes what you highlighted in your patch "[PATCH] ssb:
mark ssb_bus_register as __maybe_unused". Indeed, there is no reason for
enabling this option.

Thanks for the heads up, bye,
   Nicolas.

 arch/arm/configs/at91_dt_defconfig | 1 -
 arch/arm/configs/sama5_defconfig   | 1 -
 2 files changed, 2 deletions(-)

diff --git a/arch/arm/configs/at91_dt_defconfig b/arch/arm/configs/at91_dt_defconfig
index 1b1e5acd76e2..e4b1be66b3f5 100644
--- a/arch/arm/configs/at91_dt_defconfig
+++ b/arch/arm/configs/at91_dt_defconfig
@@ -125,7 +125,6 @@ CONFIG_POWER_RESET=y
 # CONFIG_HWMON is not set
 CONFIG_WATCHDOG=y
 CONFIG_AT91SAM9X_WATCHDOG=y
-CONFIG_SSB=m
 CONFIG_MFD_ATMEL_HLCDC=y
 CONFIG_REGULATOR=y
 CONFIG_REGULATOR_FIXED_VOLTAGE=y
diff --git a/arch/arm/configs/sama5_defconfig b/arch/arm/configs/sama5_defconfig
index a0c57ac88b27..63f7e6ce649a 100644
--- a/arch/arm/configs/sama5_defconfig
+++ b/arch/arm/configs/sama5_defconfig
@@ -129,7 +129,6 @@ CONFIG_GPIO_SYSFS=y
 CONFIG_POWER_SUPPLY=y
 CONFIG_POWER_RESET=y
 # CONFIG_HWMON is not set
-CONFIG_SSB=m
 CONFIG_MFD_ATMEL_FLEXCOM=y
 CONFIG_REGULATOR=y
 CONFIG_REGULATOR_FIXED_VOLTAGE=y
-- 
2.1.3

Re: [PATCH] ssb: mark ssb_bus_register as __maybe_unused

[Michael Büsch]

[-- Attachment #1: Type: text/plain, Size: 707 bytes --]

On Tue, 03 Nov 2015 16:05:51 +0100
Arnd Bergmann <arnd@arndb.de> wrote:

> The SoC variant of the ssb code is now optional like the other
> ones, which means we can build the framwork without any
> front-end, but that results in a warning:
> 
> drivers/ssb/main.c:616:12: warning: 'ssb_bus_register' defined but not used [-Wunused-function]
> 
> This annotates the ssb_bus_register function as __maybe_unused to
> shut up the warning. A configuration like this will not work on
> any hardware of course, but we still want this to silently build
> without warnings if the configuration is allowed in the first
> place.


Is there a simple way to disallow this configuration?

-- 
Michael

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: [PATCH] ssb: mark ssb_bus_register as __maybe_unused

[Arnd Bergmann]

On Tuesday 03 November 2015 17:27:21 Michael Büsch wrote:
> On Tue, 03 Nov 2015 16:05:51 +0100
> Arnd Bergmann <arnd@arndb.de> wrote:
> 
> > The SoC variant of the ssb code is now optional like the other
> > ones, which means we can build the framwork without any
> > front-end, but that results in a warning:
> > 
> > drivers/ssb/main.c:616:12: warning: 'ssb_bus_register' defined but not used [-Wunused-function]
> > 
> > This annotates the ssb_bus_register function as __maybe_unused to
> > shut up the warning. A configuration like this will not work on
> > any hardware of course, but we still want this to silently build
> > without warnings if the configuration is allowed in the first
> > place.
> 
> 
> Is there a simple way to disallow this configuration?

I could not come up with a simple one. We could turn 'CONFIG_SSB' into
a silent option and have it selected by each bus specific driver,
but then we also have to change all the device drivers (usb and
wireless I guess) to use 'depends on' rather than 'select'.

	Arnd

Re: [PATCH] ssb: mark ssb_bus_register as __maybe_unused

[Michael Büsch]

[-- Attachment #1: Type: text/plain, Size: 1450 bytes --]

On Tue, 03 Nov 2015 17:42:26 +0100
Arnd Bergmann <arnd@arndb.de> wrote:

> On Tuesday 03 November 2015 17:27:21 Michael Büsch wrote:
> > On Tue, 03 Nov 2015 16:05:51 +0100
> > Arnd Bergmann <arnd@arndb.de> wrote:
> >   
> > > The SoC variant of the ssb code is now optional like the other
> > > ones, which means we can build the framwork without any
> > > front-end, but that results in a warning:
> > > 
> > > drivers/ssb/main.c:616:12: warning: 'ssb_bus_register' defined but not used [-Wunused-function]
> > > 
> > > This annotates the ssb_bus_register function as __maybe_unused to
> > > shut up the warning. A configuration like this will not work on
> > > any hardware of course, but we still want this to silently build
> > > without warnings if the configuration is allowed in the first
> > > place.  
> > 
> > 
> > Is there a simple way to disallow this configuration?  
> 
> I could not come up with a simple one. We could turn 'CONFIG_SSB' into
> a silent option and have it selected by each bus specific driver,
> but then we also have to change all the device drivers (usb and
> wireless I guess) to use 'depends on' rather than 'select'.

The other way around?
The drivers already select SSB. However I think we should have SSB
selectable by user for the embedded case.

But well, I can live with this patch then. Kalle, you might apply it
with

Acked-by: Michael Buesch <m@bues.ch>

-- 
Michael

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]