Re: [PATCH] KVM: Replaces 'unsigned' with 'unsigned int' in the codebase

From: Roman Storozhenko
Date: Tue Jun 27 2017 - 09:31:49 EST


On Tue, Jun 27, 2017 at 05:21:30AM -0700, Joe Perches wrote:
> On Tue, 2017-06-27 at 15:06 +0300, Roman Storozhenko wrote:
> > On Tue, Jun 27, 2017 at 12:01:25PM +0200, Paolo Bonzini wrote:
> > > On 27/06/2017 11:54, Roman Storozhenko wrote:
> > > > Signed-off-by: Roman Storozhenko <romeusmeister@xxxxxxxxx>
> > > > ---
> > > > virt/kvm/arm/arm.c | 2 +-
> > > > virt/kvm/coalesced_mmio.c | 2 +-
> > > > virt/kvm/eventfd.c | 10 ++++++----
> > > > virt/kvm/irqchip.c | 7 ++++---
> > > > virt/kvm/kvm_main.c | 16 ++++++++--------
> > > > 5 files changed, 20 insertions(+), 17 deletions(-)
> > >
> > > This change is pointless. Contributors to KVM should know what
> > > "unsigned" means.
> >
> > Paolo, thanks that you mentioned this. But I have a question - is this just
> > useless or this is an error?
>
> (Not Paulo and my 2c) Neither really.
>
> > I saw many places in the codebase where 'unsigned int' is used. That why I
> > decided to make the codebase more standartized from the style point of
> > view.
>
> In virt/kvm, there are 16 lines with unsigned, 160 with unsigned int
>
> <shrug>, Both statements are correct, yes, it's kinda pointless,
> and, yes, it does standardize the declarations.
>
> It's entirely up the the maintainers (Paulo and Radim) to apply
> or reject this style-only trivial patch. The compiler doesn't care.
>

Joe, thank you for your detailed answer.
Anyway I hope that this patch will be accepted despite it is useless
from the compiler's point of view. I think that the codebase style
matters too.