Re: [PATCH] fix __percpu annotation in asm-generic

[Luc Van Oostenryck]

On Mon, Dec 02, 2019 at 02:07:18PM -0500, Dennis Zhou wrote:
> On Sat, Nov 30, 2019 at 01:00:37AM +0100, Luc Van Oostenryck wrote:
> > On Fri, Nov 29, 2019 at 06:11:59PM +0000, Christopher Lameter wrote:
> > > On Wed, 27 Nov 2019, Luc Van Oostenryck wrote:
> > > 
> > > > 1) it would strip any address space, not just __percpu, so:
> > > >    it would need to be combined with __verify_pcpu_ptr() or,
> > > >    * a better name should be used,
> > > 
> > > typeof_cast_kernel() to express the fact that it creates a kernel pointer
> > > and ignored the attributes??
> > 
> > typeof_strip_address_space() would, I think, express this better. 
> > It's not obvious at all to me that 'kernel' in 'typeof_cast_kernel()'
> > relates to the (default) kernel address space.
> > Maybe it's just me. I don't know.
> > 
> 
> I think typeof_cast_kernel() or typeof_force_kernel() are reasonable
> names. I kind of like the idea of cast/force over strip because we're
> really still moving address spaces even if it is moving it back.

Well, 'typeof_cast_kernel()' somehow conveys the idea but sounds
a bit weird as the macro doesn't contain a cast (expression).

> Thanks for debugging this. I'm still inclined to have a macro for either
> cast/force. I do agree it could be misused, but it's no different doing
> it in a macro than by just adding __force __kernel.

I'm glad to help making the kernel type-clean (with the goal of
catching more bugs earlier) but I admit that I absolutely detest
these layers of ugly macros. 

I'm working on a nicer implementation but it's not yet ready.

Best regards,
-- Luc