Re: [kvm-unit-tests PATCH 05/14] page: add page alignment checker

[Peter Xu <peterx@redhat.com>]

On Thu, Oct 20, 2016 at 02:30:26PM +0200, Andrew Jones wrote:
> On Thu, Oct 20, 2016 at 02:23:14PM +0200, Andrew Jones wrote:
> > On Fri, Oct 14, 2016 at 08:40:43PM +0800, Peter Xu wrote:
> > > Signed-off-by: Peter Xu <peterx@redhat.com>
> > > ---
> > >  lib/asm-generic/page.h | 5 +++++
> > >  1 file changed, 5 insertions(+)
> > > 
> > > diff --git a/lib/asm-generic/page.h b/lib/asm-generic/page.h
> > > index 7b8a08b..cbdc8f6 100644
> > > --- a/lib/asm-generic/page.h
> > > +++ b/lib/asm-generic/page.h
> 
> Another comment. IS_ALIGNED shouldn't be here, it should be
> in lib/libcflat.h under the definition of ALIGN there.

Moving over.

> 
> drew
> 
> > > @@ -19,6 +19,11 @@
> > >  
> > >  #define PAGE_ALIGN(addr)	ALIGN(addr, PAGE_SIZE)
> > >  
> > > +#define IS_ALIGNED(x, a)	(((x) & ((typeof(x))(a) - 1)) == 0)
> > > +#define PAGE_ALIGNED_4K(x)	IS_ALIGNED((x), (0x1000))
> > > +#define PAGE_ALIGNED_2M(x)	IS_ALIGNED((x), (0x200000))
> > > +#define PAGE_ALIGNED_1G(x)	IS_ALIGNED((x), (0x40000000))
> > 
> > I like IS_ALIGNED, but not the others. The others are strangely
> > named because it's half asking if an address is page aligned and
> > half asking if it's aligned to a given boundary. Why not just
> > ask the boundary, like ALIGNED_2M(x)? Anyway, my preference would
> > be to replace them with SZ_4K, SZ_2M, SZ_1G definitions. Users would
> > then do
> > 
> >  IS_ALIGNED(addr, SZ_2M)
> > 
> > or whatever, which isn't much more typing.

Yeah these looks more clean. Will take them. Thanks!

-- peterx