Re: [PATCH 16/63] i8259: Rename TYPE_KVM_I8259 to TYPE_KVM_PIC

[Eduardo Habkost]

On Thu, Sep 03, 2020 at 02:53:52PM +0200, Philippe Mathieu-Daudé wrote:
> On 9/3/20 12:42 AM, Eduardo Habkost wrote:
> > This will make the type name constant consistent with the name of
> > the type checking macro.
> > 
> > Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
> > ---
> > Cc: "Michael S. Tsirkin" <mst@redhat.com>
> > Cc: Marcel Apfelbaum <marcel.apfelbaum@gmail.com>
> > Cc: Paolo Bonzini <pbonzini@redhat.com>
> > Cc: Richard Henderson <rth@twiddle.net>
> > Cc: Eduardo Habkost <ehabkost@redhat.com>
> > Cc: qemu-devel@nongnu.org
> > ---
> >  hw/i386/kvm/i8259.c | 10 +++++-----
> >  1 file changed, 5 insertions(+), 5 deletions(-)
> > 
> > diff --git a/hw/i386/kvm/i8259.c b/hw/i386/kvm/i8259.c
> > index 3f8bf69e9c..687c0cd536 100644
> > --- a/hw/i386/kvm/i8259.c
> > +++ b/hw/i386/kvm/i8259.c
> > @@ -19,10 +19,10 @@
> >  #include "sysemu/kvm.h"
> >  #include "qom/object.h"
> >  
> > -#define TYPE_KVM_I8259 "kvm-i8259"
> > +#define TYPE_KVM_PIC "kvm-i8259"
> 
> I disagree with this patch, as we have various KVM INTC and only one
> KVM_I8259.
> 
> TYPE_KVM_ARM_GIC and TYPE_KVM_S390_FLIC are kind of TYPE_KVM_INTC ...
> 
> Can we rename it KVM_I8259_PIC?

I'm inclined to agree, but I'm not completely sure.

Why is it OK to have a macro named KVM_PIC, a struct named
KVMPICClass, a struct named PICCommonState, but not OK to have a
constant named TYPE_KVM_PIC?  What about the TYPE_PIC_COMMON
constant?

All these symbols are internal to the i8259 code and aren't
expected to be unique globally.  Are TYPE_* names more special
and expected to be unique globally?  If yes, why?

-- 
Eduardo