Re: powerpc allmodconfig

[David Miller <davem@davemloft.net>]

From: Johannes Berg <johannes@sipsolutions.net>
Date: Thu, 16 Oct 2008 16:57:19 +0200

> On Wed, 2008-10-15 at 22:02 -0700, David Miller wrote:
> > 
> > 
> > > net/sched/sch_generic.c: In function 'dev_watchdog':
> > > net/sched/sch_generic.c:224: warning: unused variable 'drivername'
> > 
> > Sucky, if WARN_ONCE() evaluates to nothing the sprintf() string buffer
> > on the stack looks unused.
> 
> I've complained about this to Arjan before, we actually lose all
> messages passed to WARN() or WARN_ONCE() on platforms that use bug traps
> for warnings too.

Ok I see how that works, yes, it should be fixed.

If the platform defines a __WARN (which powerpc does) the
whole format string and printf args go unevaluated, it's
because of the following sequence in asm-generic/bug.h:

#ifndef __WARN
#ifndef __ASSEMBLY__
extern void warn_on_slowpath(const char *file, const int line);
extern void warn_slowpath(const char *file, const int line,
		const char *fmt, ...) __attribute__((format(printf, 3, 4)));
#define WANT_WARN_ON_SLOWPATH
#endif
#define __WARN() warn_on_slowpath(__FILE__, __LINE__)
#define __WARN_printf(arg...) warn_slowpath(__FILE__, __LINE__, arg)
#else
#define __WARN_printf(arg...) __WARN()
#endif