| 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
| ||
|
Date: Sat, 15 Sep 2012 17:23:03 +0200
From: Fabio Baltieri <fabio.baltieri@...il.com>
To: Bryan Wu <bryan.wu@...onical.com>
Cc: rpurdie@...ys.net, linux-kernel@...r.kernel.org,
akpm@...ux-foundation.org, linux-leds@...r.kernel.org,
broonie@...nsource.wolfsonmicro.com, shuahkhan@...il.com,
raph@...com, tpiepho@...escale.com
Subject: Re: [PATCH 1/2] leds: use workqueue in led_set_brightness() API
internally
Hello Bryan,
On Fri, Sep 14, 2012 at 03:53:02PM +0800, Bryan Wu wrote:
> The API function led_set_brightness() and __led_set_brightness will
> call .brightness_set() function provided by led class drivers. So
> .brightness_set() function will run in atomic context, which requires
> led class drivers use workqueue in .brightness_set().
>
> Finally, all the led class driver implemented their own workqueue in
> .brightness_set(). For those missing this, we will face runtime warning
> or error when running it in atomic context.
>
> This patch adds a workqueue in led_set_brightness() internally. LED
> class driver doesn't need to care about duplicated workqueue stuff in
> their own driver.
>
> Also this patch unified the led_set_brightness() and __led_set_brightness()
I like the idea of a semplification of led driver's code and the removal
of the confusion between the two variants of led_set_brightness.
Despite that, due to the lockless nature of the soft-blink code in the
led subsystem I expect it to break quite easility with this kind of
modification.
> Signed-off-by: Bryan Wu <bryan.wu@...onical.com>
> ---
> drivers/leds/led-class.c | 23 ++++++++++++-----------
> drivers/leds/led-core.c | 15 +++++++--------
> drivers/leds/leds.h | 11 ++---------
> drivers/leds/ledtrig-backlight.c | 8 ++++----
> drivers/leds/ledtrig-default-on.c | 2 +-
> drivers/leds/ledtrig-gpio.c | 6 +++---
> drivers/leds/ledtrig-heartbeat.c | 2 +-
> drivers/leds/ledtrig-oneshot.c | 4 ++--
> drivers/leds/ledtrig-transient.c | 8 ++++----
> include/linux/leds.h | 12 +++++++-----
> 10 files changed, 43 insertions(+), 48 deletions(-)
>
> diff --git a/drivers/leds/led-class.c b/drivers/leds/led-class.c
> index 48cce18..7a3e886 100644
> --- a/drivers/leds/led-class.c
> +++ b/drivers/leds/led-class.c
> @@ -53,7 +53,7 @@ static ssize_t led_brightness_store(struct device *dev,
>
> if (state == LED_OFF)
> led_trigger_remove(led_cdev);
> - __led_set_brightness(led_cdev, state);
> + led_set_brightness(led_cdev, state);
>
> return size;
> }
> @@ -82,7 +82,7 @@ static void led_timer_function(unsigned long data)
> unsigned long delay;
>
> if (!led_cdev->blink_delay_on || !led_cdev->blink_delay_off) {
> - __led_set_brightness(led_cdev, LED_OFF);
> + led_set_brightness(led_cdev, LED_OFF);
> return;
> }
>
> @@ -105,7 +105,7 @@ static void led_timer_function(unsigned long data)
> delay = led_cdev->blink_delay_off;
> }
>
> - __led_set_brightness(led_cdev, brightness);
> + led_set_brightness(led_cdev, brightness);
>
> /* Return in next iteration if led is in one-shot mode and we are in
> * the final blink state so that the led is toggled each delay_on +
> @@ -124,14 +124,15 @@ static void led_timer_function(unsigned long data)
> mod_timer(&led_cdev->blink_timer, jiffies + msecs_to_jiffies(delay));
> }
>
> -static void set_brightness_delayed(struct work_struct *ws)
> +static void led_set_brightness_work(struct work_struct *ws)
> {
> struct led_classdev *led_cdev =
> container_of(ws, struct led_classdev, set_brightness_work);
>
> - led_stop_software_blink(led_cdev);
> + if (led_cdev->blink_delay_on || led_cdev->blink_delay_off)
> + led_stop_software_blink(led_cdev);
This actually breaks soft-blink, as led_set_brightness_work is called by
the soft-blink code itself when soft blink is active. You may probabily
want replicate the actual set code in led_timer_function() or to keep it
in an internal function.
One of the effects is that blink_delay_{on,off} are erroneously set to 0
when setting 'timer' or 'oneshot' as current trigger.
> - __led_set_brightness(led_cdev, led_cdev->delayed_set_value);
> + led_cdev->brightness_set(led_cdev, led_cdev->brightness);
Is it ok to access led_cdev->brightness without locks? I'm afraid that
this may race in SMP if another CPU is running led_set_brightness.
The lockless solution would be to make sure the work is canceled before
setting the new value, through that may not be kind to hardirq context.
> }
>
> /**
> @@ -141,7 +142,7 @@ static void set_brightness_delayed(struct work_struct *ws)
> void led_classdev_suspend(struct led_classdev *led_cdev)
> {
> led_cdev->flags |= LED_SUSPENDED;
> - led_cdev->brightness_set(led_cdev, 0);
> + led_set_brightness(led_cdev, 0);
> }
> EXPORT_SYMBOL_GPL(led_classdev_suspend);
>
> @@ -151,7 +152,7 @@ EXPORT_SYMBOL_GPL(led_classdev_suspend);
> */
> void led_classdev_resume(struct led_classdev *led_cdev)
> {
> - led_cdev->brightness_set(led_cdev, led_cdev->brightness);
> + led_set_brightness(led_cdev, led_cdev->brightness);
> led_cdev->flags &= ~LED_SUSPENDED;
> }
> EXPORT_SYMBOL_GPL(led_classdev_resume);
> @@ -201,7 +202,7 @@ int led_classdev_register(struct device *parent, struct led_classdev *led_cdev)
>
> led_update_brightness(led_cdev);
>
> - INIT_WORK(&led_cdev->set_brightness_work, set_brightness_delayed);
> + INIT_WORK(&led_cdev->set_brightness_work, led_set_brightness_work);
>
> init_timer(&led_cdev->blink_timer);
> led_cdev->blink_timer.function = led_timer_function;
> @@ -233,12 +234,12 @@ void led_classdev_unregister(struct led_classdev *led_cdev)
> up_write(&led_cdev->trigger_lock);
> #endif
>
> - cancel_work_sync(&led_cdev->set_brightness_work);
> -
> /* Stop blinking */
> led_stop_software_blink(led_cdev);
> led_set_brightness(led_cdev, LED_OFF);
>
> + flush_work(&led_cdev->set_brightness_work);
> +
Is it possible that this code runs with led_set_brightness_work()
already scheduled on another CPU? In that case flush_work would not
wait for it.
Fabio
> device_unregister(led_cdev->dev);
>
> down_write(&leds_list_lock);
> diff --git a/drivers/leds/led-core.c b/drivers/leds/led-core.c
> index ce8921a..c212262 100644
> --- a/drivers/leds/led-core.c
> +++ b/drivers/leds/led-core.c
> @@ -45,7 +45,7 @@ static void led_set_software_blink(struct led_classdev *led_cdev,
>
> /* never off - just set to brightness */
> if (!delay_off) {
> - __led_set_brightness(led_cdev, led_cdev->blink_brightness);
> + led_set_brightness(led_cdev, led_cdev->blink_brightness);
> return;
> }
>
> @@ -114,13 +114,12 @@ EXPORT_SYMBOL_GPL(led_stop_software_blink);
> void led_set_brightness(struct led_classdev *led_cdev,
> enum led_brightness brightness)
> {
> - /* delay brightness setting if need to stop soft-blink timer */
> - if (led_cdev->blink_delay_on || led_cdev->blink_delay_off) {
> - led_cdev->delayed_set_value = brightness;
> - schedule_work(&led_cdev->set_brightness_work);
> - return;
> - }
> + if (brightness > led_cdev->max_brightness)
> + brightness = led_cdev->max_brightness;
>
> - __led_set_brightness(led_cdev, brightness);
> + led_cdev->brightness = brightness;
> +
> + if (!(led_cdev->flags & LED_SUSPENDED))
> + schedule_work(&led_cdev->set_brightness_work);
> }
> EXPORT_SYMBOL(led_set_brightness);
> diff --git a/drivers/leds/leds.h b/drivers/leds/leds.h
> index 4c50365..745bf06 100644
> --- a/drivers/leds/leds.h
> +++ b/drivers/leds/leds.h
> @@ -17,15 +17,8 @@
> #include <linux/rwsem.h>
> #include <linux/leds.h>
>
> -static inline void __led_set_brightness(struct led_classdev *led_cdev,
> - enum led_brightness value)
> -{
> - if (value > led_cdev->max_brightness)
> - value = led_cdev->max_brightness;
> - led_cdev->brightness = value;
> - if (!(led_cdev->flags & LED_SUSPENDED))
> - led_cdev->brightness_set(led_cdev, value);
> -}
> +void led_set_brightness(struct led_classdev *led_cdev,
> + enum led_brightness value);
>
> static inline int led_get_brightness(struct led_classdev *led_cdev)
> {
> diff --git a/drivers/leds/ledtrig-backlight.c b/drivers/leds/ledtrig-backlight.c
> index b941685..e272686 100644
> --- a/drivers/leds/ledtrig-backlight.c
> +++ b/drivers/leds/ledtrig-backlight.c
> @@ -46,9 +46,9 @@ static int fb_notifier_callback(struct notifier_block *p,
>
> if ((n->old_status == UNBLANK) ^ n->invert) {
> n->brightness = led->brightness;
> - __led_set_brightness(led, LED_OFF);
> + led_set_brightness(led, LED_OFF);
> } else {
> - __led_set_brightness(led, n->brightness);
> + led_set_brightness(led, n->brightness);
> }
>
> n->old_status = new_status;
> @@ -87,9 +87,9 @@ static ssize_t bl_trig_invert_store(struct device *dev,
>
> /* After inverting, we need to update the LED. */
> if ((n->old_status == BLANK) ^ n->invert)
> - __led_set_brightness(led, LED_OFF);
> + led_set_brightness(led, LED_OFF);
> else
> - __led_set_brightness(led, n->brightness);
> + led_set_brightness(led, n->brightness);
>
> return num;
> }
> diff --git a/drivers/leds/ledtrig-default-on.c b/drivers/leds/ledtrig-default-on.c
> index eac1f1b..a4ef54b 100644
> --- a/drivers/leds/ledtrig-default-on.c
> +++ b/drivers/leds/ledtrig-default-on.c
> @@ -19,7 +19,7 @@
>
> static void defon_trig_activate(struct led_classdev *led_cdev)
> {
> - __led_set_brightness(led_cdev, led_cdev->max_brightness);
> + led_set_brightness(led_cdev, led_cdev->max_brightness);
> }
>
> static struct led_trigger defon_led_trigger = {
> diff --git a/drivers/leds/ledtrig-gpio.c b/drivers/leds/ledtrig-gpio.c
> index ba215dc..f057c10 100644
> --- a/drivers/leds/ledtrig-gpio.c
> +++ b/drivers/leds/ledtrig-gpio.c
> @@ -54,12 +54,12 @@ static void gpio_trig_work(struct work_struct *work)
>
> if (tmp) {
> if (gpio_data->desired_brightness)
> - __led_set_brightness(gpio_data->led,
> + led_set_brightness(gpio_data->led,
> gpio_data->desired_brightness);
> else
> - __led_set_brightness(gpio_data->led, LED_FULL);
> + led_set_brightness(gpio_data->led, LED_FULL);
> } else {
> - __led_set_brightness(gpio_data->led, LED_OFF);
> + led_set_brightness(gpio_data->led, LED_OFF);
> }
> }
>
> diff --git a/drivers/leds/ledtrig-heartbeat.c b/drivers/leds/ledtrig-heartbeat.c
> index 1edc746..a019fbb 100644
> --- a/drivers/leds/ledtrig-heartbeat.c
> +++ b/drivers/leds/ledtrig-heartbeat.c
> @@ -74,7 +74,7 @@ static void led_heartbeat_function(unsigned long data)
> break;
> }
>
> - __led_set_brightness(led_cdev, brightness);
> + led_set_brightness(led_cdev, brightness);
> mod_timer(&heartbeat_data->timer, jiffies + delay);
> }
>
> diff --git a/drivers/leds/ledtrig-oneshot.c b/drivers/leds/ledtrig-oneshot.c
> index 2c029aa..2b10b28 100644
> --- a/drivers/leds/ledtrig-oneshot.c
> +++ b/drivers/leds/ledtrig-oneshot.c
> @@ -63,9 +63,9 @@ static ssize_t led_invert_store(struct device *dev,
> oneshot_data->invert = !!state;
>
> if (oneshot_data->invert)
> - __led_set_brightness(led_cdev, LED_FULL);
> + led_set_brightness(led_cdev, LED_FULL);
> else
> - __led_set_brightness(led_cdev, LED_OFF);
> + led_set_brightness(led_cdev, LED_OFF);
>
> return size;
> }
> diff --git a/drivers/leds/ledtrig-transient.c b/drivers/leds/ledtrig-transient.c
> index 398f104..83179f4 100644
> --- a/drivers/leds/ledtrig-transient.c
> +++ b/drivers/leds/ledtrig-transient.c
> @@ -41,7 +41,7 @@ static void transient_timer_function(unsigned long data)
> struct transient_trig_data *transient_data = led_cdev->trigger_data;
>
> transient_data->activate = 0;
> - __led_set_brightness(led_cdev, transient_data->restore_state);
> + led_set_brightness(led_cdev, transient_data->restore_state);
> }
>
> static ssize_t transient_activate_show(struct device *dev,
> @@ -72,7 +72,7 @@ static ssize_t transient_activate_store(struct device *dev,
> if (state == 0 && transient_data->activate == 1) {
> del_timer(&transient_data->timer);
> transient_data->activate = state;
> - __led_set_brightness(led_cdev, transient_data->restore_state);
> + led_set_brightness(led_cdev, transient_data->restore_state);
> return size;
> }
>
> @@ -80,7 +80,7 @@ static ssize_t transient_activate_store(struct device *dev,
> if (state == 1 && transient_data->activate == 0 &&
> transient_data->duration != 0) {
> transient_data->activate = state;
> - __led_set_brightness(led_cdev, transient_data->state);
> + led_set_brightness(led_cdev, transient_data->state);
> transient_data->restore_state =
> (transient_data->state == LED_FULL) ? LED_OFF : LED_FULL;
> mod_timer(&transient_data->timer,
> @@ -203,7 +203,7 @@ static void transient_trig_deactivate(struct led_classdev *led_cdev)
>
> if (led_cdev->activated) {
> del_timer_sync(&transient_data->timer);
> - __led_set_brightness(led_cdev, transient_data->restore_state);
> + led_set_brightness(led_cdev, transient_data->restore_state);
> device_remove_file(led_cdev->dev, &dev_attr_activate);
> device_remove_file(led_cdev->dev, &dev_attr_duration);
> device_remove_file(led_cdev->dev, &dev_attr_state);
> diff --git a/include/linux/leds.h b/include/linux/leds.h
> index 5676197..d7f5697 100644
> --- a/include/linux/leds.h
> +++ b/include/linux/leds.h
> @@ -43,10 +43,15 @@ struct led_classdev {
> #define LED_BLINK_ONESHOT_STOP (1 << 18)
> #define LED_BLINK_INVERT (1 << 19)
>
> - /* Set LED brightness level */
> - /* Must not sleep, use a workqueue if needed */
> + /*
> + * Set LED brightness level
> + * use a workqueue internally, so driver writer doesn't need to
> + * care about using duplicated workqueue in brightness_set()
> + */
> void (*brightness_set)(struct led_classdev *led_cdev,
> enum led_brightness brightness);
> + struct work_struct set_brightness_work;
> +
> /* Get LED brightness level */
> enum led_brightness (*brightness_get)(struct led_classdev *led_cdev);
>
> @@ -70,9 +75,6 @@ struct led_classdev {
> struct timer_list blink_timer;
> int blink_brightness;
>
> - struct work_struct set_brightness_work;
> - int delayed_set_value;
> -
> #ifdef CONFIG_LEDS_TRIGGERS
> /* Protects the trigger data below */
> struct rw_semaphore trigger_lock;
> --
> 1.7.11.4
>
--
Fabio Baltieri
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists