Re: [PATCH] libxfs: Optimize the loop for xfs_bitmap_empty

[hejianet <hejianet@xxxxxxxxx>]

Hi Brian
Thanks, I will resend the patch V2

B.R.
Justin

å 11/10/15 8:54 PM, Brian Foster åé:
> On Mon, Nov 09, 2015 at 10:31:22PM +0800, Jia He wrote:
> > If there is any non zero bit in a long bitmap, it can jump out of the for
> > loop and finish the function as soon as possible.
> >
> > Signed-off-by: Jia He <hejianet@xxxxxxxxx>
> > Cc: Dave Chinner <david@xxxxxxxxxxxxx>
> > Cc: Brian Foster <bfoster@xxxxxxxxxx>
> > ---
> >   fs/xfs/libxfs/xfs_bit.c | 2 ++
> >   1 file changed, 2 insertions(+)
> >
> > diff --git a/fs/xfs/libxfs/xfs_bit.c b/fs/xfs/libxfs/xfs_bit.c
> > index 0e8885a..84614b0 100644
> > --- a/fs/xfs/libxfs/xfs_bit.c
> > +++ b/fs/xfs/libxfs/xfs_bit.c
> > @@ -36,6 +36,8 @@ xfs_bitmap_empty(uint *map, uint size)
> >   
> >   	for (i = 0; i < size; i++) {
> >                 ret |= map[i];
> > +               if (ret != 0)
> > +                       return 0;
> >         }
> >   
> Seems Ok to me, but if we're going to do this, why not just kill ret
> entirely? For example, check 'if (map[i] != 0) return 0;' in the loop
> and unconditionally return 1 if we make it to the end.
>
> Brian
>
> >         return (ret == 0);
> > --
> > 2.5.0
> >
> > _______________________________________________
> > xfs mailing list
> > xfs@xxxxxxxxxxx
> > http://oss.sgi.com/mailman/listinfo/xfs
> _______________________________________________
> xfs mailing list
> xfs@xxxxxxxxxxx
> http://oss.sgi.com/mailman/listinfo/xfs