On Fri, Jan 11, 2008 at 10:45:30AM +0100, Roel Kluin wrote:
quoted text
> Paul Mundt wrote:
> > On Fri, Jan 11, 2008 at 04:09:45AM +0100, Peter Stuge wrote:
> >> On Thu, Jan 10, 2008 at 10:03:58PM +0100, Roel Kluin wrote:
> >>> -#define DEBUG(x,args...)	printk(__FUNCTION__ ": " x,##args)
> >>> +#define DEBUG(x, args...)	printk("%s: ", __func__, x, ##args)
> >> Can this really be expected to work when x contains conversions?
> >>
> >> How about:
> >>
> >> #define DEBUG(x, args...) printk("%s: " x, __func__, ##args)
> >>
> > How about throwing out hand-rolled debug printk wrappers for the
> > brain-damage they are and using the ones the kernel provides instead?
> > 
> Should it be done like this?
> --
> Replace printk wrapper - with a syntax error - by pr_debug
> 
> Signed-off-by: Roel Kluin <12o3l@tiscali.nl>

Close. But in this case #define DEBUG is already instrumented by the
subsystem-wide debug option if it's enabled, so it's preferable to use
that and just drop the special debug Kconfig option completely.

Take a look at how CONFIG_PCMCIA_DEBUG is handled.

With DEBUG()->pr_debug() conversion here you've silently dropped the
__func__ prefixing. Note that dev_dbg() is usually preferred when you can
get a hold of a struct device pointer, as it takes care of prettifying
the output with the driver name and so on, rather than the convention of
adding a prefix. If you can't get at the struct device pointer, you'll
probably just want to insert the __func__ prefixing manually at the
callsites.
--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/