Received-SPF: pass (google.com: domain of linux-kernel+bounces-212788-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
Precedence: bulk
MIME-Version: 1.0
References: <20240610112700.80819-1-wsa+renesas@sang-engineering.com>
 <20240610112700.80819-2-wsa+renesas@sang-engineering.com> <CAMRc=MfZ11U+kAh1+K=DxtJ=QL+cY7Q_sBN4sQDF-RNgjpV0QA@mail.gmail.com>
 <jvnvx7a4pn6evrp5ehfrt4qsiuprq6ogvrue2a3uupwtydmgcm@2rvat7ibvgb4>
In-Reply-To: <jvnvx7a4pn6evrp5ehfrt4qsiuprq6ogvrue2a3uupwtydmgcm@2rvat7ibvgb4>
From: Bartosz Golaszewski <brgl@bgdev.pl>
Date: Thu, 13 Jun 2024 10:17:27 +0200
Message-ID: <CAMRc=Mc4__0zzJZG3BPnmbua88SLuEbX=Wk=EZnKH5HQvB+JPg@mail.gmail.com>
Subject: Re: [PATCH v9 1/1] gpio: add sloppy logic analyzer using polling
To: Wolfram Sang <wsa+renesas@sang-engineering.com>, Bartosz Golaszewski <brgl@bgdev.pl>, 
	linux-renesas-soc@vger.kernel.org, Jonathan Corbet <corbet@lwn.net>, 
	Linus Walleij <linus.walleij@linaro.org>, Kent Gibson <warthog618@gmail.com>, 
	linux-doc@vger.kernel.org, linux-kernel@vger.kernel.org, 
	linux-gpio@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Wed, Jun 12, 2024 at 6:03=E2=80=AFPM Wolfram Sang
<wsa+renesas@sang-engineering.com> wrote:
>
> Hi Bart,
>
> > I really dislike drivers being called in an ambiguous way like
> > "simple" or - in this case "sloppy". I understand why it is - in fact
> > - sloppy but can we call it anything else? Like
> > "gpio-logic-analyzer.c"?
>
> Sure, we can if you prefer. I named it like this to make the limitations
> super-clear. And even with that in place, I still got a private email
> where someone wanted to build a 400MHz-RPi-based logic analyzer device
> with it. Which would not only have the latency problems, but also
> likely have a max sampling speed of whopping 400kHz.
>
> > > +Note that you must provide a name for every GPIO specified. Currentl=
y a
> > > +maximum of 8 probes are supported. 32 are likely possible but are no=
t
> > > +implemented yet.
> > > +
> >
> > What happens on non-DT systems? Can you still create an analyzer in a
> > different way? Can I maybe interest you in configfs for the purpose of
> > device configuration like what gpio-sim and the upcoming gpio-virtuser
> > does?
>
> Frankly, I'd like to leave this to the person needing it. I've been
> working on this for way too long already and am not up to major changes
> anymore. Minor stuff, okay, I'll go one or two more rounds.
>
> The GPIO analyzer is a debug tool aimed for development boards in remote
> labs, and all boards I have access to use DT. Furthermore, debugfs is nic=
e
> because it is clear there is no stable ABI. It has been useful as-is in
> the past. That's what I am offering. If that's not enough, no hard
> feelings, but someone else needs to continue then.
>
> All the best,

I see. I don't want to stop it from going upstream. On second thought
though: are you sure drivers/gpio/ is the right place for it? This
directory is for GPIO *providers* - mostly drivers for real HW and
some "virtual" GPIO providers (gpio-aggregator, gpio-sim, gpio-latch).
In general: modules that export their own gpio_chip. This module is a
consumer of GPIOs exclusively. May I suggest moving it over to
drivers/misc/? I think it's a better place for it and I wouldn't mind
keeping the "sloppy" in the name then.

Bart