Re: [PATCH 7/8] xen/mmu: Introduce IDENTITY_FRAME_BIT

[Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>]

On Tue, Jan 04, 2011 at 04:26:15PM +0000, Ian Campbell wrote:
> On Thu, 2010-12-30 at 19:48 +0000, Konrad Rzeszutek Wilk wrote:
> > @@ -52,9 +53,12 @@ static inline unsigned long pfn_to_mfn(unsigned long pfn)
> >  
> >  	mfn = get_phys_to_machine(pfn);
> >  
> > -	if (mfn != INVALID_P2M_ENTRY)
> > +	if (mfn != INVALID_P2M_ENTRY) {
> >  		mfn &= ~FOREIGN_FRAME_BIT;
> >  
> > +		if (mfn & IDENTITY_FRAME_BIT)
> > +			mfn &= ~IDENTITY_FRAME_BIT;
> > +	}
> 
> I don't think the inner-if buys us anything here and the whole thing is
> equivalent to:
> 	if (mfn != INVALID_P2M_ENTRY)
> 		mfn &= ~(FOREIGN_FRAME_BIT|IDENTITY_FRAME_BIT);
> 
> Not sure if the FOREIGN_FRAME_BIT|IDENTITY_FRAME_BIT construct gets
> enough use to be worthy of a #define FRAME_TYPE_MASK etc.

Yeah, that is much better.
> 
> >  	return mfn;
> >  }
> >  
> > diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c
> > index d98bd43..d470435 100644
> > --- a/arch/x86/xen/mmu.c
> > +++ b/arch/x86/xen/mmu.c
> > @@ -425,10 +425,10 @@ unsigned long get_phys_to_machine(unsigned long pfn)
> >  	 * would be wrong.
> >  	 */
> >  	if (p2m_top[topidx] == p2m_mid_identity)
> > -		return pfn;
> > +		return pfn | IDENTITY_FRAME_BIT;
> 
> It's probably worth defining IDENTITY_FRAME(m) in the pattern of
> FOREIGN_FRAME(m).

<nods>