On Thu, 2014-08-07 at 19:01 +0300, Dan Carpenter wrote:
> On Thu, Aug 07, 2014 at 09:01:36PM +0530, Srikrishan Malik wrote:
> > On Wed, Aug 06, 2014 at 11:18:13PM +0300, Dan Carpenter wrote:
> > > That looks silly before and after. Everything is indented in a funny
> > > way.
> >
> > Is this better:
> >
> > static const ldlm_policy_data_t lookup_policy = {
> > .l_inodebits = { MDS_INODELOCK_LOOKUP }
> > };
> >
>
> That is indented too far.
>
> Honestly, I think it looks best on one line but in terms of real life we
> can't ignore checkpatch warnings because eventually someone else will
> try to "fix" it to not be on one line.
>
> This function has the silly thing where the types are in one column and
> the variables are in another. But then over time inevitably we add more
> variables and nothing is lined up any more.
>
> I think it's best to move this const variable block to the very front of
> the list.
>
> req doesn't need to be initialized.
>
> rc is normally the last variable declared.
>
> lvb_type should be initialized to LVB_T_NONE instead of zero.
>
> __u64 should be u64.
>
> All those changes could be done as one patch titled, "cleanup variable
> declarations in mdc_enqueue()". There may be other cleanups you could
> do as well. Look hard.
I think it looks odd to mix named and unnamed
initializers for the typedef and its members.
ldlm_policy_data_t is a union and it could be
explicit instead of a typedef.
Perhaps:
static const union ldlm_policy_data lookup_policy = {
.l_inodebits = {
.bits = MDS_INODELOCK_LOOKUP,
},
};
or maybe use another DECLARE_<foo> macro indirection.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
