Received-SPF: softfail (google.com: domain of transitioning linux-kernel-owner@vger.kernel.org does not designate 23.128.96.19 as permitted sender) client-ip=23.128.96.19;
MIME-Version: 1.0
References: <20220409034849.3717231-1-zheyuma97@gmail.com> <20220410165425.6c2f60e7@jic23-huawei>
 <CACRpkdaFgB55HHR8a3vyVbZphu5fpguutBYemyVvGz=tcn6j+A@mail.gmail.com>
In-Reply-To: <CACRpkdaFgB55HHR8a3vyVbZphu5fpguutBYemyVvGz=tcn6j+A@mail.gmail.com>
From:   Ulf Hansson <ulf.hansson@linaro.org>
Date:   Thu, 21 Apr 2022 16:07:28 +0200
Message-ID: <CAPDyKFr50QirUvkUy+b665=2prMkGz=EXtbyc+Bty1woT=qO3w@mail.gmail.com>
Subject: Re: [PATCH] iio: magnetometer: ak8974: Fix the error handling of ak8974_probe()
To:     Linus Walleij <linus.walleij@linaro.org>
Cc:     Jonathan Cameron <jic23@kernel.org>,
        Zheyu Ma <zheyuma97@gmail.com>, lars@metafoo.de,
        linux-iio@vger.kernel.org, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Thu, 21 Apr 2022 at 01:14, Linus Walleij <linus.walleij@linaro.org> wrote:
>
> On Sun, Apr 10, 2022 at 5:46 PM Jonathan Cameron <jic23@kernel.org> wrote:
> > On Sat,  9 Apr 2022 11:48:48 +0800
> > Zheyu Ma <zheyuma97@gmail.com> wrote:
> >
> > > When the driver fail at devm_regmap_init_i2c(), we will get the
> > > following splat:
> > >
> > > [  106.797388] WARNING: CPU: 4 PID: 413 at drivers/regulator/core.c:2257 _regulator_put+0x3ec/0x4e0
> > > [  106.802183] RIP: 0010:_regulator_put+0x3ec/0x4e0
> > > [  106.811237] Call Trace:
> > > [  106.811515]  <TASK>
> > > [  106.811695]  regulator_bulk_free+0x82/0xe0
> > > [  106.812032]  devres_release_group+0x319/0x3d0
> > > [  106.812425]  i2c_device_probe+0x766/0x940
> > >
> > > Fix this by disabling the regulators at the error path.
> > >
> > > Signed-off-by: Zheyu Ma <zheyuma97@gmail.com>
> > + CC Linus W as it's his driver.
> >
> > Fix looks correct to me, though the handling of runtime pm in here is
> > probably more complex than it needs to be (and hence this odd error
> > handling for this one place in the probe).
>
> At the time I discussed how to do runtime pm with Ulf Hansson a lot
> and I think it was the state of the art at that time. It might have
> changed since.

It still looks good to me. One could switch the order of the calls to
pm_runtime_disable() and pm_runtime_put_noidle(), but it's not a big
thing.

Whether it looks complicated or not, that's a different story. :-)

Note that some drivers are always being built with CONFIG_PM being
set, which allows some simplifications. Although, I don't think that's
the case here, right?

>
> > > ---
> > >  drivers/iio/magnetometer/ak8974.c | 1 +
> > >  1 file changed, 1 insertion(+)
> > >
> > > diff --git a/drivers/iio/magnetometer/ak8974.c b/drivers/iio/magnetometer/ak8974.c
> > > index e54feacfb980..84bbf7ccc887 100644
> > > --- a/drivers/iio/magnetometer/ak8974.c
> > > +++ b/drivers/iio/magnetometer/ak8974.c
> > > @@ -862,6 +862,7 @@ static int ak8974_probe(struct i2c_client *i2c,
> > >               dev_err(&i2c->dev, "failed to allocate register map\n");
> > >               pm_runtime_put_noidle(&i2c->dev);
> > >               pm_runtime_disable(&i2c->dev);
> > > +             regulator_bulk_disable(ARRAY_SIZE(ak8974->regs), ak8974->regs);
>
> Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
>
> Yours,
> Linus Walleij

Kind regards
Uffe