[PATCH 1/3] ARM: OMAP2+: Fix realtime_counter_init warning in timer.c

[Jon Hunter]

On 11/28/2012 12:09 AM, Santosh Shilimkar wrote:
> On Wednesday 28 November 2012 07:45 AM, Jon Hunter wrote:
>> In commit fa6d79d (ARM: OMAP: Add initialisation for the real-time
>> counter), the function realtime_counter_init() was added. However, if
>> the kernel configuration option CONFIG_SOC_OMAP5 is not selected then
>> the following compiler warning is observed.
>>
>>    CC      arch/arm/mach-omap2/timer.o
>>    arch/arm/mach-omap2/timer.c:489:20: warning: ‘realtime_counter_init’
>>    defined but not used [-Wunused-function]
>>
>> Commit fa6d79d also introduced the kernel configuration option
>> CONFIG_SOC_HAS_REALTIME_COUNTER. If this option is not selected then the
>> a stub function for realtime_counter_init() is defined.
>>
>> The option CONFIG_SOC_HAS_REALTIME_COUNTER and stub function are not
>> really needed because ...
>>
>> 1. For non-OMAP5 devices, there is no realtime counter and so
>>     realtime_counter_init() function is not used.
>> 2. For OMAP5 devices, CONFIG_SOC_HAS_REALTIME_COUNTER is always selected
>>     and cannot be disabled, so the stub function for
>> realtime_counter_init()
>>     is never used.
>>
>> Fix this warning by removing the kernel configuration option
>> CONFIG_SOC_HAS_REALTIME_COUNTER and stub function, and only include
>> the function realtime_counter_init() if CONFIG_SOC_OMAP5 is selected.
>>
>> Cc: Santosh Shilimkar <santosh.shilimkar at ti.com>
>>
>> Reported-by: Tony Lindgren <tony at atomide.com>
>> Signed-off-by: Jon Hunter <jon-hunter at ti.com>
>> ---
> The #ifdef was avoided because the real-time counter can be used on
> other future SOCs. And the those SOCs just select
> SOC_HAS_REALTIME_COUNTER. And that stub was added because OMAP5 can
> work without real-time counter configuration enabled using 32K counter.
> But since we are any way have that SOC_HAS_REALTIME_COUNTER always
> set for SOC which wants to use it, we can actually remove the stub
> and hence avoid the warning. Let me know if below patch is ok
> with you ? attached the same for mailer issues
>
> From e000aa13e47e29fbe3473bfd0277cb057c3160cc Mon Sep 17 00:00:00 2001
> From: Santosh Shilimkar <santosh.shilimkar at ti.com>
> Date: Wed, 28 Nov 2012 11:28:57 +0530
> Subject: [PATCH] ARM: OMAP2+: Fix realtime_counter_init warning in timer.c
> MIME-Version: 1.0
> Content-Type: text/plain; charset=UTF-8
> Content-Transfer-Encoding: 8bit
> 
> In commit fa6d79d (ARM: OMAP: Add initialisation for the real-time
> counter), the function realtime_counter_init() was added. However, if
> the kernel configuration option CONFIG_SOC_OMAP5 is not selected then
> the following compiler warning is observed.
> 
>     CC      arch/arm/mach-omap2/timer.o
>     arch/arm/mach-omap2/timer.c:489:20: warning: ‘realtime_counter_init’
>     defined but not used [-Wunused-function]
> 
> It is because of the stub init function which was added for the cases
> where realtime_counter_init() is called with
> !CONFIG_SOC_HAS_REALTIME_COUNTER.
> This is actually not necessary since the SOC which need this feature
> will explicitly select the configuration.
> 
> So just drop the unused stub to avoid the build warning.
> 
> Patch is made after seeing Jon's patch which was wrapping the
> real-time counter code under needed SOC #ifdef
> 
> Cc: Jon Hunter <jon-hunter at ti.com>
> 
> Reported-by: Tony Lindgren <tony at atomide.com>
> Signed-off-by: Santosh Shilimkar <santosh.shilimkar at ti.com>
> ---
>  arch/arm/mach-omap2/timer.c |    3 ---
>  1 file changed, 3 deletions(-)
> 
> diff --git a/arch/arm/mach-omap2/timer.c b/arch/arm/mach-omap2/timer.c
> index 69e4663..79d8e6b 100644
> --- a/arch/arm/mach-omap2/timer.c
> +++ b/arch/arm/mach-omap2/timer.c
> @@ -428,9 +428,6 @@ static void __init realtime_counter_init(void)
> 
>      iounmap(base);
>  }
> -#else
> -static inline void __init realtime_counter_init(void)
> -{}
>  #endif
> 
>  #define OMAP_SYS_TIMER_INIT(name, clkev_nr, clkev_src,            \

Hi Santosh,

Thanks for the feedback. Unfortunately, the above is not enough. If I
disable SOC_OMAP5 and leave SOC_HAS_REALTIME_COUNTER enabled, then I
still see the warning message. So I could see that people could still
hit this with randconfig.

Understand the desire to make this a selectable configuration, but given
to date only OMAP5 uses it, it was simpler and cleaner just to remove
the option. However, we could keep the option and just do something like ...

diff --git a/arch/arm/mach-omap2/Kconfig b/arch/arm/mach-omap2/Kconfig
index 813c267..500f2f6 100644
--- a/arch/arm/mach-omap2/Kconfig
+++ b/arch/arm/mach-omap2/Kconfig
@@ -26,6 +26,8 @@ config SOC_HAS_OMAP2_SDRC

 config SOC_HAS_REALTIME_COUNTER
        bool "Real time free running counter"
+       depends on SOC_OMAP5
+       default y

 config ARCH_OMAP2
        bool "TI OMAP2"
@@ -79,7 +81,6 @@ config SOC_OMAP5
        select ARM_GIC
        select CPU_V7
        select HAVE_SMP
-       select SOC_HAS_REALTIME_COUNTER
        select COMMON_CLK

I have tested this and this resolves the warning.

Cheers
Jon