Re: [PATCH] Correct printk %pF to work on all architectures

On Wed, 3 Sep 2008, James Bottomley wrote:

Oh ... because Arjan has a patch to export
dereference_function_descriptor.  I suppose I could make him do the
heavy lifting, but it seemed sensible to make it easy for him (and me)
by putting it in a header.

http://marc.info/?l=linux-kernel&m=121976793429869

Ahh.

NOW it all starts to make sense.

Or perhaps not sense, but I at least understand why people want to move it 
around. The kernel.h location kind of goes together with that 
core_kernel_text() thing, although it seems to be more of a "random 
collection of routines" thing than anything else (but hey, that's the very 
definition of "kernel.h" for you).

The module.h location still seems to be more of a "oh, both 
kernel/extable.c and lib/vsprintf.c already included <linux/module.h>" and 
it's a bit sad, since it really has nothing at all to do with modules.

Grr. It does seem like we don't have any kind of "abi" header file. 
<linux/kernel.h> and <asm/processor.h> has various random things.

So yea, there doesn't seem to be any _obvious_ place that makes sense.

		Linus

Not-very-strong-opinion: How about <asm/sections.h>? That does seem to be 
where we already hide things like "in_kernel_text()" at least on powerpc. 
In fact, since we already always have a generic version, the patch would 
actually be something like

 - in <asm-generic/sections.h>, just do

	#define dereference_function_descriptor(p) (p)

 - in architectures that want to override it

	#undef dereference_function_descriptor

   followed by 

	static inline void *dereference_function_descriptor(..) ..
	or
	#define dereference_function_descriptor my_fn_dereference

   since they all include the generic one as a base

Hmm? I do admit that "<asm/sections.h>" doesn't really strike me as a very 
natural name for this, but kernel/extable.c does already include it for 
other reasons, and it's at least no worse than module.h.