cpu_*(), #define, and &

cpu_*(), #define, and &

[George Dunlap]

I hope I'm not bikeshedding here
(http://catb.org/jargon/html/B/bikeshedding.html), but I'm not really
happy with the cpu_*() macros in cpumask.h adding ampersands before
the arguments.  In C (as opposed to C++), passing a non-pointer
generally means that no values are going to be changed.  Other than
Linux similarity, is there a good reason to do this in a macro, rather
than just having the caller provide the &?  Would anyone object to me
submitting a patch to change that?

(A patch to change it looks to be rather big and boring, so I want to
talk about it first before doing it...)

 -George

Re: cpu_*(), #define, and &

[Keir Fraser]

On 21/09/2009 17:49, "George Dunlap" <dunlapg@umich.edu> wrote:

> I hope I'm not bikeshedding here
> (http://catb.org/jargon/html/B/bikeshedding.html), but I'm not really
> happy with the cpu_*() macros in cpumask.h adding ampersands before
> the arguments.  In C (as opposed to C++), passing a non-pointer
> generally means that no values are going to be changed.  Other than
> Linux similarity, is there a good reason to do this in a macro, rather
> than just having the caller provide the &?  Would anyone object to me
> submitting a patch to change that?
> 
> (A patch to change it looks to be rather big and boring, so I want to
> talk about it first before doing it...)

I somewhat agree about the pointless macro-ification in cpumask.h, but I
don't care enough about this to diverge from the original Linux definitions.
The cpumask macros which modify the given mask are named to make it pretty
obvious. We'll live with it; it's not really that annoying.

 -- Keir