Re: [PATCH] mm: Do not use double negation for testing page flags
From: Minchan Kim
Date: Thu Mar 09 2017 - 01:42:36 EST
Hi Vlastimil,
On Wed, Mar 08, 2017 at 08:51:23AM +0100, Vlastimil Babka wrote:
> On 03/08/2017 06:25 AM, Minchan Kim wrote:
> > Hi Anshuman,
> >
> > On Tue, Mar 07, 2017 at 09:31:18PM +0530, Anshuman Khandual wrote:
> >> On 03/07/2017 12:06 PM, Minchan Kim wrote:
> >>> With the discussion[1], I found it seems there are every PageFlags
> >>> functions return bool at this moment so we don't need double
> >>> negation any more.
> >>> Although it's not a problem to keep it, it makes future users
> >>> confused to use dobule negation for them, too.
> >>>
> >>> Remove such possibility.
> >>
> >> A quick search of '!!Page' in the source tree does not show any other
> >> place having this double negation. So I guess this is all which need
> >> to be fixed.
> >
> > Yeb. That's the why my patch includes only khugepagd part but my
> > concern is PageFlags returns int type not boolean so user might
> > be confused easily and tempted to use dobule negation.
> >
> > Other side is they who create new custom PageXXX(e.g., PageMovable)
> > should keep it in mind that they should return 0 or 1 although
> > fucntion prototype's return value is int type.
>
> > It shouldn't be
> > documented nowhere.
>
> Was this double negation intentional? :P
Nice catch!
It seems you have a crystal ball. ;-)
>
> > Although we can add a little description
> > somewhere in page-flags.h, I believe changing to boolean is more
> > clear/not-error-prone so Chen's work is enough worth, I think.
>
> Agree, unless some arches benefit from the int by performance
> for some reason (no idea if it's possible).
>
> Anyway, to your original patch:
>
> Acked-by: Vlastimil Babka <vbabka@xxxxxxx>
Thanks!