Re: [B.A.T.M.A.N.] [PATCH 1/2] batman-adv: add routing debug log accessible via debugfs

[Sven Eckelmann <sven.eckelmann@gmx.de>]

[-- Attachment #1: Type: Text/Plain, Size: 1520 bytes --]

Marek Lindner wrote:
>  /*
>   *  Vis
> @@ -160,3 +148,20 @@ int choose_orig(void *data, int32_t size);
>  int is_my_mac(uint8_t *addr);
>  int is_bcast(uint8_t *addr);
>  int is_mcast(uint8_t *addr);
> +
> +#ifdef CONFIG_BATMAN_ADV_DEBUG
> +extern int debug;
> +
> +extern int bat_debug_type(int type);
> +extern int debug_log(struct bat_priv *bat_priv, char *fmt, ...);
> +
> +#define bat_dbg(type, bat_priv, fmt, arg...) do {		\
> +		if (bat_debug_type(type))			\
> +			debug_log(bat_priv, fmt, ## arg);	\
> +	}							\
> +	while (0)
> +#else /* !CONFIG_BATMAN_ADV_DEBUG */
> +#define bat_dbg(type, bat_priv, fmt, arg...) do {	\
> +	}						\
> +	while (0)
> +#endif

Can you please change that (at least the 'null' function) from a define to an 
static inline function? Otherwise there are a lot of "unused" warning when 
compiling without debug support.

static inline void bat_dbg(char type __attribute__((unused)),
			   struct bat_priv *bat_priv __attribute__((unused)),
			   char *fmt __attribute__((unused)), ...)
{ }

I will send a include guard patch later which is needed to get this working.

To you suggestion to add more information from kernel log to the debug log - 
it seems to be interesting, but we must ensure that those information are 
specific to a mesh device. I will sent a patch which modifies all printks 
later. Maybe we could discuss where it would be a benefit and where it is not 
needed (~54 printks are for discussion).

Best regards,
	Sven

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 836 bytes --]