Re: [Xen-devel] [PATCH v2 04/16] xen/x86: p2m-pod: Fix coding style

[Julien Grall <julien.grall@xxxxxxx>]

On 04/10/17 06:38, Jan Beulich wrote:
> > > > Julien Grall <julien.grall@xxxxxxx> 09/28/17 9:30 PM >>>
> > On 09/22/2017 10:15 AM, Jan Beulich wrote:
> > > > > > On 21.09.17 at 14:40, <julien.grall@xxxxxxx> wrote:
> > > > Also take the opportunity to:
> > > >       - move from 1 << * to 1UL << *.
> > > >       - use unsigned when possible
> > > >       - move from unsigned int -> unsigned long for some induction
> > > >       variables
> > > I don't understand this last point, btw - the largest order page the
> > > code needs to deal with right now is 1Gb, so there's no risk of
> > > overflow (yet). But you've got George's and Andrew's ack, so no
> > > need to revise this...
> > The last one result from the existing 1UL << in the code. We have place
> > where the induction variable is unsigned int but the shift unsigned long.
> In which case I would have suggested to change the shift to use 1U, since ...
Looking at "<< order" within Xen we seems to use a mix of 1UL <<, 1U <<, 
1 <<. The variant 1UL << seems to be predominant.
At the end we really want to be consistent with the rest of the code base.

> > Similarly the code is using a mix of 1 << and 1UL <<. I moved to UL
> > because even if the code only support up to 1GB superpage at the moment,
> > it would be pain to find all the places the day we decide to use bigger one.
> ... I doubt this would be too hard. But in the end it's George to judge anyway.
order is the number of MFN that is unsigned long and therefore would 
make sense to use 1UL <<. So I am not sure why you are pushing for 1U << 
here...
Cheers,

--
Julien Grall

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel