Re: [patch 13/17] Use WARN() in drivers/base/

[Arjan van de Ven <arjan@infradead.org>]

On Fri, 11 Jul 2008 12:20:11 -0700
Andrew Morton <akpm@linux-foundation.org> wrote:

> On Tue, 8 Jul 2008 09:53:07 -0700 Arjan van de Ven
> <arjan@infradead.org> wrote:
> 
> > Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>Index:
> > linux.trees.git/drivers/base/core.c
> 
> A number of these patches had mangled signed-off-by: lines.
> 
> Please try to be consistent in the presence and placement of the ^---
> line at the end of the changelog.
> 
> I verified that all three copies of "Use WARN() in fs/" were the same.
> 
> I've decided that I don't like the whole thing :(  This:
> 
> #define WARN(condition, format...)
> ({					\ int __ret_warn_on
> = !!(condition);				\ if
> (unlikely(__ret_warn_on))					\
> __WARN_printf(format);					\
> unlikely(__ret_warn_on);					\ })
> 
> is not a WARN().  It is a WARN_ON() function.  The use of this name
> now leaves us no sensible name under which to implement
> 

I'm totally open to a better name.
Having a condition in there is really nice, it means we can fold the
if() into it in many cases. Just like BUG_ON() did.