Re: [PATCH 2/7] x86: make some apic symbols init

[Henrik Kretzschmar <henne@nachtwindheim.de>]

Am 03.03.2011 22:19, schrieb Cyrill Gorcunov:
> On 03/03/2011 11:04 PM, Henrik Kretzschmar wrote:
>> apic_force_enable(), apic_verify() and the variable
>> force_enable_local_apic are only used by init code
>> and now get marked as such.
>>
>> Global __initdata variables may better be initialized,
>> since they are in the data section and not in the bss.
>>
>> Signed-off-by: Henrik Kretzschmar <henne@nachtwindheim.de>
>> ---
>>  arch/x86/include/asm/apic.h |    2 +-
>>  arch/x86/kernel/apic/apic.c |    6 +++---
>>  2 files changed, 4 insertions(+), 4 deletions(-)
>>
>> diff --git a/arch/x86/include/asm/apic.h b/arch/x86/include/asm/apic.h
>> index dbd558c..afe69e1 100644
>> --- a/arch/x86/include/asm/apic.h
>> +++ b/arch/x86/include/asm/apic.h
>> @@ -240,7 +240,7 @@ extern void setup_boot_APIC_clock(void);
>>  extern void setup_secondary_APIC_clock(void);
>>  extern int APIC_init_uniprocessor(void);
>>  extern void enable_NMI_through_LVT0(void);
>> -extern int apic_force_enable(unsigned long addr);
>> +extern int apic_force_enable(unsigned long addr) __init;
> 
> Nope, we either should _check_ all the prototipes and fix them
> either left them untouched. This will confuse code readers with "for
> what reason some functions have __init, some -- not". So I rather
> would fix this nit in different patch later which would address
> all of them (if this would not break someone's queue).
> 
Ack, I'll fix it.

>>   */
>> -static int force_enable_local_apic;
>> +static int force_enable_local_apic __initdata = 0;
> 
> Hmm, I fail to see why we need to set it to 0.

Ack, I'll fix this also.