Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.38 as permitted sender) client-ip=23.128.96.38;
Date:   Thu, 14 Dec 2023 08:18:01 +0800
From:   Kent Gibson <warthog618@gmail.com>
To:     Andy Shevchenko <andy@kernel.org>
Cc:     Bartosz Golaszewski <brgl@bgdev.pl>, linux-kernel@vger.kernel.org,
        linux-gpio@vger.kernel.org, linus.walleij@linaro.org
Subject: Re: [PATCH 1/4] gpiolib: cdev: relocate debounce_period_us from
 struct gpio_desc
Message-ID: <ZXpJueTnmtUIecCd@rigel>
References: <20231212054253.50094-2-warthog618@gmail.com>
 <ZXm3rayrcvfO1t1Z@smile.fi.intel.com>
 <ZXm_WsIpgIyOUNHt@rigel>
 <CAMRc=Mfri8K4ZqcHb_eQY6gi+q_-uBZc2wiMrrb-+a7Tric3FA@mail.gmail.com>
 <ZXnU3tMYCc2Rw8Qv@rigel>
 <ZXnX8jPHxRLW8lhi@smile.fi.intel.com>
 <CAMRc=Mfj_4YvQVP=UWkULBwJniDDjapttU+qSgqfN5ZWNgikKw@mail.gmail.com>
 <ZXnb-ks_noYLWZZ2@smile.fi.intel.com>
 <CAMRc=MfaHKKKNkXW0L1FPjCH4VvG22Vn9q8z9tupZCtCEKZU2g@mail.gmail.com>
 <ZXoO8B0N3S49GnvX@smile.fi.intel.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <ZXoO8B0N3S49GnvX@smile.fi.intel.com>
Precedence: bulk

On Wed, Dec 13, 2023 at 10:07:12PM +0200, Andy Shevchenko wrote:
> On Wed, Dec 13, 2023 at 08:03:44PM +0100, Bartosz Golaszewski wrote:
>
> ...
>
> > > > > > > > > > +static struct supinfo supinfo;
> > > > > > > > >
> > > > > > > > > Why supinfo should be a struct to begin with? Seems to me as an unneeded
> > > > > > > > > complication.
> > > > > > >
> > > > > > > I think we should keep it as a struct but defined the following way:
> > > > > > >
> > > > > > > struct {
> > > > > > >     spinlock_t lock;
> > > > > > >     struct rb_root tree;
> > > > > > > } supinfo;
> > > > > >
> > > > > > That is what I meant be merging the struct definition with the variable
> > > > > > definition.  Or is there some other way to completely do away with the
> > > > > > struct that I'm missing?
> > > > >
> > > > > Look at the top of gpiolib.c:
> > > > >
> > > > > static DEFINE_MUTEX(gpio_lookup_lock);
> > > > > static LIST_HEAD(gpio_lookup_list);
> > > > >
> > > > > In the similar way you can simply do
> > > > >
> > > > > static DEFINE_SPINLOCK(gpio_sup_lock);
> > > > > static struct rb_root gpio_sup_tree;
> > > >
> > > > The fact that this has been like this, doesn't mean it's the only
> > > > right way. IMO putting these into the same structure makes logical
> > > > sense.
> > >
> > > I disagree on the struct like this on the grounds:
> > > - it's global

I dislike having the global at all - and now you want two :-(.

> > > - it's one time use

Its not about how many times it is instanciated, it is about
maintainability.

> > > - it adds complications for no benefit

It provides a placeholder for collective documentation and clarifies
scope for the reader.
How is it more complicated?

> > > - it makes code uglier and longer
> > >

What, swapping an underscore for a period?

And you would hope the generated code is essentially the same.

> >
> > It boils down to supinfo.lock vs supinfo_lock. I do prefer the former
> > but it's a weak opinion, I won't die on that hill.
>
> Me neither, just showing rationale from my side.
>

I can't recall the last time I intentionally used separate globals over a
struct, so if there are no strong opinions otherwise I'll leave it as a
struct.

Cheers,
Kent.