[PATCH 12/38] powerpc/booke: Provide exception macros with interrupt name

[Kumar Gala]

On Mar 21, 2012, at 1:19 PM, Scott Wood wrote:

> On 03/21/2012 01:04 PM, Kumar Gala wrote:
>> 
>> On Feb 28, 2012, at 6:09 PM, Alexander Graf wrote:
>> 
>>> From: Scott Wood <scottwood at freescale.com>
>>> 
>>> DO_KVM will need to identify the particular exception type.
>>> 
>>> There is an existing set of arbitrary numbers that Linux passes,
>>> but it's an undocumented mess that sort of corresponds to server/classic
>>> exception vectors but not really.
>> 
>> So what do the new names correspond to?
> 
> The names are Linux-defined.  The values are IVOR numbers.
> 
>> What header is defining MACHINE_CHECK, BOOKE_INTERRUPT_EXTERNAL, etc
>> (asm/kvm_asm.h)?
> 
> Yes, it's asm/kvm_asm.h at the moment.
> 
> It's actually BOOKE_INTERRUPT_MACHINE_CHECK, etc.  The exception macros
> paste on BOOKE_INTERRUPT_ when they use it, to keep the macro users from
> having to wrap lines even more often.
> 
>> If so we really should move these out of
>> asm/kvm_asm.h and into something a bit more appropriate.
> 
> Yes, that was one of the things I had been planning to fix post-RFC.
> Any preference what header it should be in?  asm/reg_booke.h seems to be
> the least bad option of the existing headers -- or perhaps move
> head_booke.h to arch/powerpc/include/asm.

asm/reg_booke.h seems the least painful right now.  head_booke.h is only used on 32-bit so that's not the best choice at this point.  We could create an <asm/head_booke.h> in addition to the one in arch/powerpc/kernel/

> Long-term it would also be nice for this to replace, rather than
> supplement, the current numbers, though as Ben pointed out there are a
> number of places throughout the code that will have to be fixed for that
> to happen.

Yeah, not too worried about that.  More just wanting it to be clear what one has to 'update' if adding support for a new exception.

- k