Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20230605125248.279921-1-brgl@bgdev.pl> <CAL_JsqKczF9yYHWjqneBv-y+Qv+O7AkX4gwVG87+aPPazKxtDw@mail.gmail.com>
 <22a21176-f2e2-bce5-2223-97cb095162c5@gmail.com> <98a35030-6dd3-795c-4381-1db4e94a18e4@gmail.com>
In-Reply-To: <98a35030-6dd3-795c-4381-1db4e94a18e4@gmail.com>
From:   Rob Herring <robh+dt@kernel.org>
Date:   Mon, 12 Jun 2023 09:42:43 -0600
Message-ID: <CAL_JsqKDQy+tq7uDJ9impTr+uetkb7vz-K4BYX8bZBkpvZ=8uQ@mail.gmail.com>
Subject: Re: [PATCH] gpiolib: demote the hogging log messages to debug
To:     Frank Rowand <frowand.list@gmail.com>
Cc:     Bartosz Golaszewski <brgl@bgdev.pl>,
        Linus Walleij <linus.walleij@linaro.org>,
        Kent Gibson <warthog618@gmail.com>,
        Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
        linux-gpio@vger.kernel.org, linux-kernel@vger.kernel.org,
        devicetree@vger.kernel.org,
        Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Sun, Jun 11, 2023 at 6:48=E2=80=AFAM Frank Rowand <frowand.list@gmail.co=
m> wrote:
>
> On 6/11/23 07:39, Frank Rowand wrote:
> > On 6/9/23 08:47, Rob Herring wrote:
> >> On Mon, Jun 5, 2023 at 6:53=E2=80=AFAM Bartosz Golaszewski <brgl@bgdev=
.pl> wrote:
> >>>
> >>> From: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
> >>>
> >>> Drivers should be silent when they work correctly. There's no reason =
to
> >>> emit info messages when GPIO lines are hogged. Demote the message to
> >>> debug.
> >>>
> >>> Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
> >>> Suggested-by: Kent Gibson <warthog618@gmail.com>
> >>> ---
> >>>  drivers/gpio/gpiolib.c |  2 +-
> >>>  drivers/of/unittest.c  | 16 ++++++++--------
> >>>  2 files changed, 9 insertions(+), 9 deletions(-)
> >>>
> >>> diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
> >>> index a7220e04a93e..e4515bda8915 100644
> >>> --- a/drivers/gpio/gpiolib.c
> >>> +++ b/drivers/gpio/gpiolib.c
> >>> @@ -4243,7 +4243,7 @@ int gpiod_hog(struct gpio_desc *desc, const cha=
r *name,
> >>>         /* Mark GPIO as hogged so it can be identified and removed la=
ter */
> >>>         set_bit(FLAG_IS_HOGGED, &desc->flags);
> >>>
> >>> -       gpiod_info(desc, "hogged as %s%s\n",
> >>> +       gpiod_dbg(desc, "hogged as %s%s\n",
> >>>                 (dflags & GPIOD_FLAGS_BIT_DIR_OUT) ? "output" : "inpu=
t",
> >>>                 (dflags & GPIOD_FLAGS_BIT_DIR_OUT) ?
> >>>                   (dflags & GPIOD_FLAGS_BIT_DIR_VAL) ? "/high" : "/lo=
w" : "");
> >>> diff --git a/drivers/of/unittest.c b/drivers/of/unittest.c
> >>> index 2191c0136531..0060334a98a7 100644
> >>> --- a/drivers/of/unittest.c
> >>> +++ b/drivers/of/unittest.c
> >>> @@ -1849,19 +1849,19 @@ static void __init of_unittest_overlay_gpio(v=
oid)
> >>>          * driver is registered
> >>>          */
> >>>
> >>> -       EXPECT_BEGIN(KERN_INFO,
> >>> +       EXPECT_BEGIN(KERN_DEBUG,
> >>>                      "gpio-<<int>> (line-B-input): hogged as input\n"=
);
> >>
> >> As debug messages are normally off, I think you can just remove these.
> >
> > This patch is an example of exactly why the message level is the first =
parameter
> > passed to EXPECT_*().  The test results are then _always_ valid, not ju=
st
> > _normally_.
>
> One should never say never.  One should never say always. :-)
>
> Yes, there is still the exception where debug can be enabled independentl=
y
> for drivers/gpio/gpiolib.c  vs for drivers/of/unittest.c.  And dynamic
> debug can make things even more wonky.

If we turned on debug messages for drivers/of/, the unittest would be
hopelessly broken.

Debug messages are special compared to all the other levels as they
are normally off whereas the rest are always on. For the unittest we
should assume they are off.

Rob