Re: [PATCH] drivers/misc/kgdbts.c: remove eprintk

[Julia Lawall <julia.lawall@lip6.fr>]

On Mon, 5 Nov 2012, Arnd Bergmann wrote:

> On Sunday 04 November 2012, Julia Lawall wrote:
> > >
> > > I don't see yet where that KERN_WARNING gets added. Looking at
> > > warn_slowpath_common, there are two or three lines that get printed at
> > > KERN_WARNING level, followed by the format that got passed into WARN(),
> > > which may or may not include a printk level, but I don't see one getting
> > > added.
> >
> > OK, I agree.  There are lots of KERN_WARNINGs, but not on the string that
> > was passed in.  Still, maybe it is not so good to pass a KERN_XXX for some
> > other XXX to WARN.
>
> Given that most users don't pass anything here, and that those that do pass
> something are somewhat inconsistent, could we make the messages always get
> printed at KERN_WARNING level from WARN(), and kill off the instances
> that already pass a level?

OK, I could try proposing that, and if someone doesn't think it is the
right thing to do, they can ignore the patch.

Concretely, KERN_WARNING should be added in the printk called from WARN,
and all KERN information should be removed from the calls?

thanks,
julia