Re: [PATCH] kernel: context_tracking: use extern function instead of static inline function for user_enter/exit()

[Chen Gang <gang.chen@asianux.com>]

On 10/23/2013 06:47 PM, Steven Rostedt wrote:
> On Wed, 2013-10-23 at 18:38 +0800, Chen Gang wrote:
> 
>>
>> Signed-off-by: Chen Gang <gang.chen@asianux.com>
>> ---
>>  include/linux/context_tracking.h |   14 ++------------
>>  kernel/context_tracking.c        |   12 ++++++++++++
>>  2 files changed, 14 insertions(+), 12 deletions(-)
>>
>> diff --git a/include/linux/context_tracking.h b/include/linux/context_tracking.h
>> index 1581587..bf220f7 100644
>> --- a/include/linux/context_tracking.h
>> +++ b/include/linux/context_tracking.h
>> @@ -10,23 +10,13 @@
>>  #ifdef CONFIG_CONTEXT_TRACKING
>>  extern void context_tracking_cpu_set(int cpu);
>>  
>> +extern void user_enter(void);
>> +extern void user_exit(void);
>>  extern void context_tracking_user_enter(void);
>>  extern void context_tracking_user_exit(void);
>>  extern void __context_tracking_task_switch(struct task_struct *prev,
>>  					   struct task_struct *next);
>>  
>> -static inline void user_enter(void)
>> -{
>> -	if (static_key_false(&context_tracking_enabled))
>> -		context_tracking_user_enter();
>> -
>> -}
>> -static inline void user_exit(void)
>> -{
>> -	if (static_key_false(&context_tracking_enabled))
>> -		context_tracking_user_exit();
>> -}
>> -
>>  static inline enum ctx_state exception_enter(void)
>>  {
>>  	enum ctx_state prev_ctx;
>> diff --git a/kernel/context_tracking.c b/kernel/context_tracking.c
>> index 013161f..c03d0d6 100644
>> --- a/kernel/context_tracking.c
>> +++ b/kernel/context_tracking.c
>> @@ -29,6 +29,18 @@ EXPORT_SYMBOL_GPL(context_tracking_enabled);
>>  DEFINE_PER_CPU(struct context_tracking, context_tracking);
>>  EXPORT_SYMBOL_GPL(context_tracking);
>>  
>> +void user_enter(void)
>> +{
>> +	if (static_key_false(&context_tracking_enabled))
>> +		context_tracking_user_enter();
>> +}
>> +
>> +void user_exit(void)
>> +{
>> +	if (static_key_false(&context_tracking_enabled))
>> +		context_tracking_user_exit();
>> +}
>> +
>>  void context_tracking_cpu_set(int cpu)
>>  {
>>  	if (!per_cpu(context_tracking.active, cpu)) {
> 
> Ug, you just put an unneeded function call into a fast path because of a
> failure with a particular arch.
> 
> If you need the assembly to call this code, then make yourself a wrapper
> function to call that does the user_enter/exit() calls.
> 
> void arch_user_enter(void)
> {
> 	user_enter();
> }
> 

OK, thanks. That sounds reasonable to me. I will send related patch to
arm guys. :-)


Thanks.
-- 
Chen Gang