Re: [linux-usb-devel] [PATCH 7/9] USB usbfs: destroy submitted urbs only on the disconnected interface

[Oliver Neukum <oliver@neukum.org>]

Am Donnerstag, 15. April 2004 10:05 schrieb Duncan Sands:
> On Wednesday 14 April 2004 22:39, Oliver Neukum wrote:
> > > > I would prefer a real WARN_ON() so that the imbedded people compiling
> > > > for size are not affected.
> > >
> > > What do you mean?  How is a real WARN_ON() better?
> >
> > WARN_ON can be defined away to make a smaller kernel. Code that does
> > not use it takes away that option.
>
> Hi Oliver, I thought you meant that CONFIG_EMBEDDED made WARN_ON go away
> (or something like that).  If you just mean that it is easy to redefine
> WARN_ON by hand, then all I can say is: it is also easy to redefine warn by
> hand!  Anyway, I made you the following patch:

Yes, but I don't trust gcc to optimise away the 'if' if you redefine warn().

But there is another point. The embedded people deserve a single switch
to remove assertion checks. The purpose of macros like WARN_ON() is
easy and _central_ choice of debugging output vs. kernel size.

	Regards
		Oliver