On 05/16/2017 12:33 PM, Steven Rostedt wrote:
> On Tue, 16 May 2017 23:14:23 +0530
> "Naveen N. Rao" <naveen.n....@linux.vnet.ibm.com> wrote:
>
>> On 2017/05/16 07:31PM, Naveen N. Rao wrote:
>>> On 2017/05/15 10:20PM, Steven Rostedt wrote:
>>>> On Sun, 14 May 2017 01:01:02 +0530
>>>> "Naveen N. Rao" <naveen.n....@linux.vnet.ibm.com> wrote:
>>>>
>>>
>>> [snip]
>>>
>>>>> diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
>>>>> index c4536c449021..3f2aed4ad1ed 100644
>>>>> --- a/kernel/trace/trace.c
>>>>> +++ b/kernel/trace/trace.c
>>>>> @@ -7550,6 +7550,7 @@ static int instance_rmdir(const char *name)
>>>>> }
>>>>>
>>>>> tracing_set_nop(tr);
>>>>> + clear_ftrace_function_probes(tr);
>>>>> event_trace_del_tracer(tr);
>>>>> ftrace_clear_pids(tr);
>>>>> ftrace_destroy_function_files(tr);
>>>>> diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h
>>>>> index 291a1bca5748..98e0845f7235 100644
>>>>> --- a/kernel/trace/trace.h
>>>>> +++ b/kernel/trace/trace.h
>>>>> @@ -980,6 +980,7 @@ register_ftrace_function_probe(char *glob, struct
>>>>> trace_array *tr,
>>>>> extern int
>>>>> unregister_ftrace_function_probe_func(char *glob, struct trace_array *tr,
>>>>> struct ftrace_probe_ops *ops);
>>>>> +extern void clear_ftrace_function_probes(struct trace_array *tr);
>>>>
>>>> This needs to have a stub function when CONFIG_DYNAMIC_FTRACE is not
>>>> defined. Otherwise we have:
>>>>
>>>> kernel/trace/trace.c:7553:2: error: implicit declaration of function
>>>> 'clear_ftrace_function_probes' [-Werror=implicit-function-declaration]
>>>> clear_ftrace_function_probes(tr);
>>>> ^~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>
>> The prototype in trace.h is actually guarded by:
>> #if defined(CONFIG_FUNCTION_TRACER) && defined(CONFIG_DYNAMIC_FTRACE)
>>
>> So, I will guard the call to clear_ftrace_function_probes() in trace.c
>> with the same.
>>
>
> No, the proper thing to do is to make a stub function in the else part
> of the #if that the prototype was declared in.
>
> static inline void clear_ftrace_function_probes(struct trace_array *tr)
> {
> }
>
> Please, lets limit the #ifdef ugliness from the code.
>
> -- Steve
>
>
Hi Steve,
Are you good with this patch series. I am looking to see which
of these should go into 4.13-rc1 - please ack individual patches
that should be pulled into 4.13-rc1
thanks,
-- Shuah
