linux-kernel - Re: [PATCH v5 10/16] gpio: devres: Add devm_gpiod_get_parent

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CACRpkdY_2WzAnK01bQdMF69KsDvHHu9TXuyRoBcmiQMziux=eQ@mail.gmail.com>
Date:   Tue, 19 Nov 2019 15:43:47 +0100
From:   Linus Walleij <linus.walleij@...aro.org>
To:     Matti Vaittinen <matti.vaittinen@...rohmeurope.com>
Cc:     Matti Vaittinen <mazziesaccount@...il.com>,
        Lee Jones <lee.jones@...aro.org>,
        Jacek Anaszewski <jacek.anaszewski@...il.com>,
        Pavel Machek <pavel@....cz>, Dan Murphy <dmurphy@...com>,
        Rob Herring <robh+dt@...nel.org>,
        Mark Rutland <mark.rutland@....com>,
        Liam Girdwood <lgirdwood@...il.com>,
        Mark Brown <broonie@...nel.org>,
        Jonathan Corbet <corbet@....net>,
        Michael Turquette <mturquette@...libre.com>,
        Stephen Boyd <sboyd@...nel.org>,
        Bartosz Golaszewski <bgolaszewski@...libre.com>,
        Alessandro Zummo <a.zummo@...ertech.it>,
        Alexandre Belloni <alexandre.belloni@...tlin.com>,
        Mauro Carvalho Chehab <mchehab+samsung@...nel.org>,
        Jeff Kirsher <jeffrey.t.kirsher@...el.com>,
        Wolfram Sang <wsa+renesas@...g-engineering.com>,
        Marek Szyprowski <m.szyprowski@...sung.com>,
        Heiner Kallweit <hkallweit1@...il.com>,
        Nicholas Mc Guire <hofrat@...dl.org>,
        Phil Edworthy <phil.edworthy@...esas.com>,
        Linux LED Subsystem <linux-leds@...r.kernel.org>,
        "open list:OPEN FIRMWARE AND FLATTENED DEVICE TREE BINDINGS" 
        <devicetree@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        Linux Doc Mailing List <linux-doc@...r.kernel.org>,
        linux-clk <linux-clk@...r.kernel.org>,
        "open list:GPIO SUBSYSTEM" <linux-gpio@...r.kernel.org>,
        linux-rtc@...r.kernel.org
Subject: Re: [PATCH v5 10/16] gpio: devres: Add devm_gpiod_get_parent_array

On Mon, Nov 18, 2019 at 7:58 AM Matti Vaittinen
<matti.vaittinen@...rohmeurope.com> wrote:

> Bunch of MFD sub-devices which are instantiated by MFD do not have
> own device-tree nodes but have (for example) the GPIO consumer
> information in parent device's DT node. Add resource managed
> devm_gpiod_get_array() for such devices so that they can get the
> consumer information from parent DT while still binding the GPIO
> reservation life-time to this sub-device life time.
>
> If devm_gpiod_get_array is used as such - then unloading and then
> re-loading the child device fails as the GPIOs reserved during first
> load are not freed when driver for sub-device is unload (if parent
> stays there).
>
> Signed-off-by: Matti Vaittinen <matti.vaittinen@...rohmeurope.com>
(...)
> +static struct gpio_descs *__must_check
> +__devm_gpiod_get_array(struct device *gpiodev,
> +                      struct device *managed,
> +                      const char *con_id,
> +                      enum gpiod_flags flags)

I'm opposed to functions named __underscore_something()
so find a proper name for this function.
devm_gpiod_get_array_common() works if nothing else.

> @@ -292,19 +284,62 @@ struct gpio_descs *__must_check devm_gpiod_get_array(struct device *dev,
>         if (!dr)
>                 return ERR_PTR(-ENOMEM);
>
> -       descs = gpiod_get_array(dev, con_id, flags);
> +       descs = gpiod_get_array(gpiodev, con_id, flags);
>         if (IS_ERR(descs)) {
>                 devres_free(dr);
>                 return descs;
>         }
>
>         *dr = descs;
> -       devres_add(dev, dr);
> +       if (managed)
> +               devres_add(managed, dr);
> +       else
> +               devres_add(gpiodev, dr);

So we only get managed resources if the "managed" device is
passed in.

> +/**
> + * devm_gpiod_get_array - Resource-managed gpiod_get_array()

And this function is supposed to be resource managed for sure.

> + * @dev:       GPIO consumer
> + * @con_id:    function within the GPIO consumer
> + * @flags:     optional GPIO initialization flags
> + *
> + * Managed gpiod_get_array(). GPIO descriptors returned from this function are
> + * automatically disposed on driver detach. See gpiod_get_array() for detailed
> + * information about behavior and return values.
> + */
> +struct gpio_descs *__must_check devm_gpiod_get_array(struct device *dev,
> +                                                    const char *con_id,
> +                                                    enum gpiod_flags flags)
> +{
> +       return __devm_gpiod_get_array(dev, NULL, con_id, flags);

So what is this? NULL?

Doesn't that mean you just removed all resource management for this
call?

Or am I reading it wrong?

Yours,
Linus Walleij