lkml.org 
[lkml]   [2011]   [Apr]   [1]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
Date
SubjectRe: Re: [PATCH] regulator: recursive locking detected
Regrettably, this patch came out only in half. Please disregard, updated
patch will be sent instantly. Sorry for my spamming!

BR,
Robert

On Fri, 2011-04-01 at 11:33 +0200, Robert ROSENGREN wrote:
> "possible recursive locking detected"-warnings are issued when a
> regulator has specified supply regulator. Both when enabling and
> disabling regulators uses recursive call chains for notify the supply
> regulators. This is due to locking mutexes of the same lock class,
> i.e. the locks reside in the regulator_dev struct.
>
> Since this is valid behavior for the regulators, this patch changes the
> mutex lock into nested, as suggested in lockdep-design.txt.
>
> Signed-off-by: Robert Rosengren <robert.rosengren@stericsson.com>
> Acked-by: Linus Walleij <linus.walleij@linaro.org>
> ---
> drivers/regulator/core.c | 20 ++++++++++++--------
> 1 files changed, 12 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c
> index 3ffc697..2a66c6b 100644
> --- a/drivers/regulator/core.c
> +++ b/drivers/regulator/core.c
> @@ -87,7 +87,7 @@ static int _regulator_get_voltage(struct regulator_dev *rdev);
> static int _regulator_get_current_limit(struct regulator_dev *rdev);
> static unsigned int _regulator_get_mode(struct regulator_dev *rdev);
> static void _notifier_call_chain(struct regulator_dev *rdev,
> - unsigned long event, void *data);
> + unsigned long event, void *data, int lock_sublevel);
> static int _regulator_do_set_voltage(struct regulator_dev *rdev,
> int min_uV, int max_uV);
>
> @@ -1407,7 +1407,7 @@ static int _regulator_disable(struct regulator_dev *rdev,
> trace_regulator_disable_complete(rdev_get_name(rdev));
>
> _notifier_call_chain(rdev, REGULATOR_EVENT_DISABLE,
> - NULL);
> + NULL, 0);
> }
>
> /* decrease our supplies ref count and disable if required */
> @@ -1477,7 +1477,7 @@ static int _regulator_force_disable(struct regulator_dev *rdev,
> }
> /* notify other consumers that power has been forced off */
> _notifier_call_chain(rdev, REGULATOR_EVENT_FORCE_DISABLE |
> - REGULATOR_EVENT_DISABLE, NULL);
> + REGULATOR_EVENT_DISABLE, NULL, 0);
> }
>
> /* decrease our supplies ref count and disable if required */
> @@ -1699,7 +1699,7 @@ static int _regulator_do_set_voltage(struct regulator_dev *rdev,
>
> if (ret == 0)
> _notifier_call_chain(rdev, REGULATOR_EVENT_VOLTAGE_CHANGE,
> - NULL);
> + NULL, 0);
>
> trace_regulator_set_voltage_complete(rdev_get_name(rdev), selector);
>
> @@ -2167,19 +2167,23 @@ EXPORT_SYMBOL_GPL(regulator_unregister_notifier);
>
> /* notify regulator consumers and downstream regulator consumers.
> * Note mutex must be held by caller.
> + * lock_sublevel should always be 0, only used for recursive calls.
> */
> static void _notifier_call_chain(struct regulator_dev *rdev,
> - unsigned long event, void *data)
> + unsigned long event, void *data, int lock_sublevel)
> {
> struct regulator_dev *_rdev;
>
> /* call rdev chain first */
> blocking_notifier_call_chain(&rdev->notifier, event, NULL);
>
> + /* increase sublevel before stepping into nested regulators */
> + lock_sublevel++;
> +
> /* now notify regulator we supply */
> list_for_each_entry(_rdev, &rdev->supply_list, slist) {
> - mutex_lock(&_rdev->mutex);
> - _notifier_call_chain(_rdev, event, data);
> + mutex_lock_nested(&_rdev->mutex, lock_sublevel);
> + _notifier_call_chain(_rdev, event, data, lock_sublevel);
> mutex_unlock(&_rdev->mutex);
> }
> }
> @@ -2333,7 +2337,7 @@ EXPORT_SYMBOL_GPL(regulator_bulk_free);
> int regulator_notifier_call_chain(struct regulator_dev *rdev,
> unsigned long event, void *data)
> {
> - _notifier_call_chain(rdev, event, data);
> + _notifier_call_chain(rdev, event, data, 0);
> return NOTIFY_DONE;
>
> }
> --
> 1.7.4.1
>

[unhandled content-type:application/ms-tnef]
\
 
 \ /
  Last update: 2011-04-01 12:31    [W:0.043 / U:0.712 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site