Re: [PATCH v4 3/3] media: atomisp: Only use trace_printk if allowed

[Steven Rostedt <rostedt@goodmis.org>]

On Thu, 20 Aug 2020 19:49:59 -0700
Joe Perches <joe@perches.com> wrote:

> Perhaps make trace_printk dependent on #define DEBUG?

This is basically what Nicolas's patch series does in this very patch!

And no, I hate it. We are currently discussing ways of not having to
modify the config in order to allow trace_printk() to be used.

We don't want to burden the developer to take a config, add a bunch of
trace_printks() and find that it's compiled out!

Thus, this is a NAK.

-- Steve


> 
> Something like:
> ---
>  include/linux/kernel.h | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/include/linux/kernel.h b/include/linux/kernel.h
> index 500def620d8f..6ca8f958df73 100644
> --- a/include/linux/kernel.h
> +++ b/include/linux/kernel.h
> @@ -717,6 +717,7 @@ do {									\
>   * let gcc optimize the rest.
>   */
>  
> +#ifdef DEBUG
>  #define trace_printk(fmt, ...)				\
>  do {							\
>  	char _______STR[] = __stringify((__VA_ARGS__));	\
> @@ -725,6 +726,12 @@ do {							\
>  	else						\
>  		trace_puts(fmt);			\
>  } while (0)
> +#else
> +#define trace_printk(fmt, ...)						\
> +do {									\
> +	__trace_printk_check_format(fmt, ##args);			\
> +} while (0)
> +#endif
>  
>  #define do_trace_printk(fmt, args...)					\
>  do {									\
>