Re: [PATCH v2 1/2] tracing/events: make __string() more general

[Li Zefan <lizf@cn.fujitsu.com>]

>> Ah, it's pseudo code.. Should be construct_str(__entry->cmd, rq),
>> this helper function convert rq->cmd (unsigned char[]) to a string of
>> hex dump.
>>
>> No source string is available, so I'm not calling assign_str() in
>> TP_fast_assign.
> 
> OK, I think I have a better understanding of what you are doing.
> 
> But, I don't like changing the interface for string by adding that 
> "strlen" to all users. I find that very ugly.
> 

Yeah, it does look somewhat ugly..

> Could we instead perhaps add a __dynamic_array? That would take a length, 
> and append itself to after the strings?
> 

I thought about this actually. ;)

What troubled me is to give this new macro a better name.

We have __array() which is fixed-size, but __dynamic_array() is not
a dynamic-size version of __array(). How about __dyn_string() or
__dynamic_string()?

> stage 1:
> 
> #define __dynamic_array(item, len)	int __dyn_loc_##item;
> 
> stage 2:
> 
> #define __dynamic_array(item, len)	int item;
> 
> stage 3:
> 
> #define __get_dynamic_array(item)	\
> 	((void *)__entry + __entry->__str_loc_##field
> 
> stage 4:
> 
> #define __dynamic_array(item, len)	\
> 	__str_offsets.item = __str_size + \
> 		offsetof(typeof(*entry), __str_data);	\
> 	__str_size += len;
> 

Then we can make __string() reuse __dynamic_array().

> 
> And so on. You should be able to do something like this, and make it more 
> generic.
>