Re: [PATCH v3 19/40] KVM: x86: Swap the include order between x86.h and mmu.h

[Yosry Ahmed]

On Mon, Jun 01, 2026 at 07:55:28AM -0700, Sean Christopherson wrote:
> On Sat, May 30, 2026, Yosry Ahmed wrote:
> > On Fri, May 29, 2026 at 03:22:02PM -0700, Sean Christopherson wrote:
> > > Invert the include ordering between x86.h and mmu.h, so that x86.h is the
> > > "top-level" include for KVM x86.
> > 
> > You're also silently moving mmu_is_nested().
> 
> I'll explicitly call that out.
> 
> > Aside from that, I thought top-level include means that other headers
> > will include it, and it will include fewer headers. Seems like this is
> > doing the opposite?
> 
> Yeah, I'm probably using confusing terminology.  I could quite figure out how
> to concisely describe this.  I like my pyramid visualation, so about:
> 
>   Invert the include ordering between x86.h and mmu.h, and move
>   mmu_is_nested() to mmu.h where it belongs (mmu_is_nested()'s placement in
>   x86.h was solely responsible for the existing ordering), so that x86.h is
>   the top of KVM x86's "include pyramid".

Not to be pedantic, but I still can't quite figure out this
analogy/visualization.

How about just spelling it out:

  so that x86.h is not included by most headers (but includes them).