Re: [PATCH V2 1/4] x86/kvm: Resolve some missing-initializers warnings

[Paolo Bonzini <pbonzini@redhat.com>]

Il 31/07/2014 18:35, Rustad, Mark D ha scritto:
> I agree it is ugly. .name = NULL would be enough to silence it. Would
> that be better? At the moment I am thinking of this as a test case
> for the other 1000 { } and {0} initializers in the kernel that are
> throwing warnings. I know we both agree that the compiler really
> shouldn't be warning on them, but they currently make a lot noise.
> 
> How would you feel about a macro called something like ZERO_ENTRY
> defined something like:
> 
> #define ZERO_ENTRY DIAG_PUSH DIAG_IGNORE(missing-field-initializers)
> { } DIAG_POP
> 
> Where the DIAG_ macros pretty much do what you think. I have another
> patch series that Jeff hasn't gotten to yet that defines such macros.
> Of course they get put to good use.
> 
> At this point, I'll put the terminator back the way it was, but I
> would still like your opinion on the macro approach to addressing all
> of these terminators.

If you get such a macro in include/linux, I will of course accept its usage.

Paolo