Re: [PATCH v2 2/3] list_debug: Introduce inline wrappers for debug checks

From: Marco Elver
Date: Fri Aug 04 2023 - 14:09:35 EST


On Fri, 4 Aug 2023 at 19:59, Steven Rostedt <rostedt@xxxxxxxxxxx> wrote:
>
> On Fri, 4 Aug 2023 13:57:57 -0400
> Steven Rostedt <rostedt@xxxxxxxxxxx> wrote:
>
> > On Fri, 4 Aug 2023 19:49:48 +0200
> > Marco Elver <elver@xxxxxxxxxx> wrote:
> >
> > > > I've been guilty of this madness myself, but I have learned the errors of
> > > > my ways, and have been avoiding doing so in any new code I write.
> > >
> > > That's fair. We can call them __list_*_valid() (inline), and
> > > __list_*_valid_or_report() ?
> >
> > __list_*_valid_check() ?

Well, in patch 3/3, the inline function will also do a reduced set of
checking, so "valid_check" is also misleading because both will do
checks.

The key distinguishing thing between the inline and non-inline version
is that the non-inline version will check more things, and also
produce reports.

So I can see

1. __list_*_valid_or_report()
2. __list_*_full_valid()

To be appropriate. Preference?

> I have to admit, I think the main reason kernel developers default to using
> these useless underscores is because kernel developers are notoriously
> lousy at naming. ;-)

Heh, naming is hard. ;-)