From: Andy Shevchenko <andy.shevchenko@gmail.com>
To: Ahmad Fatoum <a.fatoum@pengutronix.de>
Cc: Linus Walleij <linus.walleij@linaro.org>,
Bartosz Golaszewski <bgolaszewski@baylibre.com>,
Grygorii Strashko <grygorii.strashko@ti.com>,
Thierry Reding <treding@nvidia.com>,
Sascha Hauer <kernel@pengutronix.de>,
"open list:GPIO SUBSYSTEM" <linux-gpio@vger.kernel.org>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] gpio: don't use same lockdep class for all devm_gpiochip_add_data users
Date: Sat, 1 Aug 2020 14:51:46 +0300 [thread overview]
Message-ID: <CAHp75VcnGX8Bjcxx6sVkp1wzeLb38eBqUhrPWqjnzc-7jXShGw@mail.gmail.com> (raw)
In-Reply-To: <20200731123835.8003-1-a.fatoum@pengutronix.de>
On Fri, Jul 31, 2020 at 3:40 PM Ahmad Fatoum <a.fatoum@pengutronix.de> wrote:
>
> Commit 959bc7b22bd2 ("gpio: Automatically add lockdep keys") documents
> in its commits message its intention to "create a unique class key for
> each driver".
>
> It does so by having gpiochip_add_data add in-place the definition of
> two static lockdep classes for LOCKDEP use. That way, every caller of
> the macro adds their gpiochip with unique lockdep classes.
>
> There are many indirect callers of gpiochip_add_data, however, via
> use of devm_gpiochip_add_data. devm_gpiochip_add_data has external
> linkage and all its users will share the same lockdep classes, which
> probably is not intended.
>
> Fix this by replicating the gpio_chip_add_data statics-in-macro for
> the devm_ version as well.
I ran into similar issues in another driver (not GPIO) and I agree with the fix.
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
> Fixes: 959bc7b22bd2 ("gpio: Automatically add lockdep keys")
> Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
> ---
> This doesn't fix any particular problem I ran into, but the code
> looked buggy, at least to my lockdep-user-not-developer eyes.
> ---
> drivers/gpio/gpiolib-devres.c | 13 ++++++++-----
> include/linux/gpio/driver.h | 13 +++++++++++--
> 2 files changed, 19 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/gpio/gpiolib-devres.c b/drivers/gpio/gpiolib-devres.c
> index 5c91c4365da1..7dbce4c4ebdf 100644
> --- a/drivers/gpio/gpiolib-devres.c
> +++ b/drivers/gpio/gpiolib-devres.c
> @@ -487,10 +487,12 @@ static void devm_gpio_chip_release(struct device *dev, void *res)
> }
>
> /**
> - * devm_gpiochip_add_data() - Resource managed gpiochip_add_data()
> + * devm_gpiochip_add_data_with_key() - Resource managed gpiochip_add_data_with_key()
> * @dev: pointer to the device that gpio_chip belongs to.
> * @gc: the GPIO chip to register
> * @data: driver-private data associated with this chip
> + * @lock_key: lockdep class for IRQ lock
> + * @request_key: lockdep class for IRQ request
> *
> * Context: potentially before irqs will work
> *
> @@ -501,8 +503,9 @@ static void devm_gpio_chip_release(struct device *dev, void *res)
> * gc->base is invalid or already associated with a different chip.
> * Otherwise it returns zero as a success code.
> */
> -int devm_gpiochip_add_data(struct device *dev, struct gpio_chip *gc,
> - void *data)
> +int devm_gpiochip_add_data_with_key(struct device *dev, struct gpio_chip *gc, void *data,
> + struct lock_class_key *lock_key,
> + struct lock_class_key *request_key)
> {
> struct gpio_chip **ptr;
> int ret;
> @@ -512,7 +515,7 @@ int devm_gpiochip_add_data(struct device *dev, struct gpio_chip *gc,
> if (!ptr)
> return -ENOMEM;
>
> - ret = gpiochip_add_data(gc, data);
> + ret = gpiochip_add_data_with_key(gc, data, lock_key, request_key);
> if (ret < 0) {
> devres_free(ptr);
> return ret;
> @@ -523,4 +526,4 @@ int devm_gpiochip_add_data(struct device *dev, struct gpio_chip *gc,
>
> return 0;
> }
> -EXPORT_SYMBOL_GPL(devm_gpiochip_add_data);
> +EXPORT_SYMBOL_GPL(devm_gpiochip_add_data_with_key);
> diff --git a/include/linux/gpio/driver.h b/include/linux/gpio/driver.h
> index c4f272af7af5..e6217d8e2e9f 100644
> --- a/include/linux/gpio/driver.h
> +++ b/include/linux/gpio/driver.h
> @@ -509,8 +509,16 @@ extern int gpiochip_add_data_with_key(struct gpio_chip *gc, void *data,
> gpiochip_add_data_with_key(gc, data, &lock_key, \
> &request_key); \
> })
> +#define devm_gpiochip_add_data(dev, gc, data) ({ \
> + static struct lock_class_key lock_key; \
> + static struct lock_class_key request_key; \
> + devm_gpiochip_add_data_with_key(dev, gc, data, &lock_key, \
> + &request_key); \
> + })
> #else
> #define gpiochip_add_data(gc, data) gpiochip_add_data_with_key(gc, data, NULL, NULL)
> +#define devm_gpiochip_add_data(dev, gc, data) \
> + devm_gpiochip_add_data_with_key(dev, gc, data, NULL, NULL)
> #endif /* CONFIG_LOCKDEP */
>
> static inline int gpiochip_add(struct gpio_chip *gc)
> @@ -518,8 +526,9 @@ static inline int gpiochip_add(struct gpio_chip *gc)
> return gpiochip_add_data(gc, NULL);
> }
> extern void gpiochip_remove(struct gpio_chip *gc);
> -extern int devm_gpiochip_add_data(struct device *dev, struct gpio_chip *gc,
> - void *data);
> +extern int devm_gpiochip_add_data_with_key(struct device *dev, struct gpio_chip *gc, void *data,
> + struct lock_class_key *lock_key,
> + struct lock_class_key *request_key);
>
> extern struct gpio_chip *gpiochip_find(void *data,
> int (*match)(struct gpio_chip *gc, void *data));
> --
> 2.27.0
>
--
With Best Regards,
Andy Shevchenko
next prev parent reply other threads:[~2020-08-01 11:52 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-07-31 12:38 [PATCH] gpio: don't use same lockdep class for all devm_gpiochip_add_data users Ahmad Fatoum
2020-07-31 15:19 ` Bartosz Golaszewski
2020-08-01 11:51 ` Andy Shevchenko [this message]
2020-08-03 23:23 ` Linus Walleij
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=CAHp75VcnGX8Bjcxx6sVkp1wzeLb38eBqUhrPWqjnzc-7jXShGw@mail.gmail.com \
--to=andy.shevchenko@gmail.com \
--cc=a.fatoum@pengutronix.de \
--cc=bgolaszewski@baylibre.com \
--cc=grygorii.strashko@ti.com \
--cc=kernel@pengutronix.de \
--cc=linus.walleij@linaro.org \
--cc=linux-gpio@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=treding@nvidia.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).