linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Boqun Feng <boqun.feng@gmail.com>
To: Marc Kleine-Budde <mkl@pengutronix.de>
Cc: Andrea Righi <andrea.righi@canonical.com>,
	Pavel Machek <pavel@ucw.cz>, Dan Murphy <dmurphy@ti.com>,
	linux-leds@vger.kernel.org, linux-kernel@vger.kernel.org,
	kernel@pengutronix.de, schuchmann@schleissheimer.de
Subject: Re: [PATCH] leds: trigger: fix potential deadlock with libata
Date: Sun, 7 Mar 2021 10:02:32 +0800	[thread overview]
Message-ID: <YEQ0ONQCwVUd0wmc@boqun-archlinux> (raw)
In-Reply-To: <20210306203954.ya5oqgkdalcw45c4@pengutronix.de>

On Sat, Mar 06, 2021 at 09:39:54PM +0100, Marc Kleine-Budde wrote:
> Hello *,
> 
> On 02.11.2020 11:41:52, Andrea Righi wrote:
> > We have the following potential deadlock condition:
> > 
> >  ========================================================
> >  WARNING: possible irq lock inversion dependency detected
> >  5.10.0-rc2+ #25 Not tainted
> >  --------------------------------------------------------
> >  swapper/3/0 just changed the state of lock:
> >  ffff8880063bd618 (&host->lock){-...}-{2:2}, at: ata_bmdma_interrupt+0x27/0x200
> >  but this lock took another, HARDIRQ-READ-unsafe lock in the past:
> >   (&trig->leddev_list_lock){.+.?}-{2:2}
> > 
> >  and interrupts could create inverse lock ordering between them.
> 
> [...]
> 
> > ---
> >  drivers/leds/led-triggers.c | 5 +++--
> >  1 file changed, 3 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/leds/led-triggers.c b/drivers/leds/led-triggers.c
> > index 91da90cfb11d..16d1a93a10a8 100644
> > --- a/drivers/leds/led-triggers.c
> > +++ b/drivers/leds/led-triggers.c
> > @@ -378,14 +378,15 @@ void led_trigger_event(struct led_trigger *trig,
> >  			enum led_brightness brightness)
> >  {
> >  	struct led_classdev *led_cdev;
> > +	unsigned long flags;
> >  
> >  	if (!trig)
> >  		return;
> >  
> > -	read_lock(&trig->leddev_list_lock);
> > +	read_lock_irqsave(&trig->leddev_list_lock, flags);
> >  	list_for_each_entry(led_cdev, &trig->led_cdevs, trig_list)
> >  		led_set_brightness(led_cdev, brightness);
> > -	read_unlock(&trig->leddev_list_lock);
> > +	read_unlock_irqrestore(&trig->leddev_list_lock, flags);
> >  }
> >  EXPORT_SYMBOL_GPL(led_trigger_event);
> 
> meanwhile this patch hit v5.10.x stable and caused a performance
> degradation on our use case:
> 
> It's an embedded ARM system, 4x Cortex A53, with an SPI attached CAN
> controller. CAN stands for Controller Area Network and here used to
> connect to some automotive equipment. Over CAN an ISOTP (a CAN-specific
> Transport Protocol) transfer is running. With this patch, we see CAN
> frames delayed for ~6ms, the usual gap between CAN frames is 240µs.
> 
> Reverting this patch, restores the old performance.
> 
> What is the best way to solve this dilemma? Identify the critical path
> in our use case? Is there a way we can get around the irqsave in
> led_trigger_event()?
> 

Probably, we can change from rwlock to rcu here, POC code as follow,
only compile tested. Marc, could you see whether this help the
performance on your platform? Please note that I haven't test it in a
running kernel and I'm not that familir with led subsystem, so use it
with caution ;-)

(While at it, I think maybe we miss the leddev_list_lock in net/mac80211
in the patch)

Regards,
Boqun
------->8
diff --git a/drivers/leds/led-triggers.c b/drivers/leds/led-triggers.c
index 4e7b78a84149..ae68ccab6cc9 100644
--- a/drivers/leds/led-triggers.c
+++ b/drivers/leds/led-triggers.c
@@ -171,10 +171,12 @@ int led_trigger_set(struct led_classdev *led_cdev, struct led_trigger *trig)
 
 	/* Remove any existing trigger */
 	if (led_cdev->trigger) {
-		write_lock_irqsave(&led_cdev->trigger->leddev_list_lock, flags);
-		list_del(&led_cdev->trig_list);
-		write_unlock_irqrestore(&led_cdev->trigger->leddev_list_lock,
+		spin_lock_irqsave(&led_cdev->trigger->leddev_list_lock, flags);
+		list_del_rcu(&led_cdev->trig_list);
+		spin_unlock_irqrestore(&led_cdev->trigger->leddev_list_lock,
 			flags);
+		/* Wait for the readers gone */
+		synchronize_rcu();
 		cancel_work_sync(&led_cdev->set_brightness_work);
 		led_stop_software_blink(led_cdev);
 		if (led_cdev->trigger->deactivate)
@@ -186,9 +188,9 @@ int led_trigger_set(struct led_classdev *led_cdev, struct led_trigger *trig)
 		led_set_brightness(led_cdev, LED_OFF);
 	}
 	if (trig) {
-		write_lock_irqsave(&trig->leddev_list_lock, flags);
-		list_add_tail(&led_cdev->trig_list, &trig->led_cdevs);
-		write_unlock_irqrestore(&trig->leddev_list_lock, flags);
+		spin_lock_irqsave(&trig->leddev_list_lock, flags);
+		list_add_tail_rcu(&led_cdev->trig_list, &trig->led_cdevs);
+		spin_unlock_irqrestore(&trig->leddev_list_lock, flags);
 		led_cdev->trigger = trig;
 
 		if (trig->activate)
@@ -223,9 +225,12 @@ int led_trigger_set(struct led_classdev *led_cdev, struct led_trigger *trig)
 		trig->deactivate(led_cdev);
 err_activate:
 
-	write_lock_irqsave(&led_cdev->trigger->leddev_list_lock, flags);
-	list_del(&led_cdev->trig_list);
-	write_unlock_irqrestore(&led_cdev->trigger->leddev_list_lock, flags);
+	spin_lock_irqsave(&led_cdev->trigger->leddev_list_lock, flags);
+	list_del_rcu(&led_cdev->trig_list);
+	spin_unlock_irqrestore(&led_cdev->trigger->leddev_list_lock, flags);
+
+	/* XXX could use call_rcu() here? */
+	synchronize_rcu();
 	led_cdev->trigger = NULL;
 	led_cdev->trigger_data = NULL;
 	led_set_brightness(led_cdev, LED_OFF);
@@ -285,7 +290,7 @@ int led_trigger_register(struct led_trigger *trig)
 	struct led_classdev *led_cdev;
 	struct led_trigger *_trig;
 
-	rwlock_init(&trig->leddev_list_lock);
+	spin_lock_init(&trig->leddev_list_lock);
 	INIT_LIST_HEAD(&trig->led_cdevs);
 
 	down_write(&triggers_list_lock);
@@ -378,15 +383,14 @@ void led_trigger_event(struct led_trigger *trig,
 			enum led_brightness brightness)
 {
 	struct led_classdev *led_cdev;
-	unsigned long flags;
 
 	if (!trig)
 		return;
 
-	read_lock_irqsave(&trig->leddev_list_lock, flags);
-	list_for_each_entry(led_cdev, &trig->led_cdevs, trig_list)
+	rcu_read_lock();
+	list_for_each_entry_rcu(led_cdev, &trig->led_cdevs, trig_list)
 		led_set_brightness(led_cdev, brightness);
-	read_unlock_irqrestore(&trig->leddev_list_lock, flags);
+	rcu_read_unlock();
 }
 EXPORT_SYMBOL_GPL(led_trigger_event);
 
@@ -397,20 +401,19 @@ static void led_trigger_blink_setup(struct led_trigger *trig,
 			     int invert)
 {
 	struct led_classdev *led_cdev;
-	unsigned long flags;
 
 	if (!trig)
 		return;
 
-	read_lock_irqsave(&trig->leddev_list_lock, flags);
-	list_for_each_entry(led_cdev, &trig->led_cdevs, trig_list) {
+	rcu_read_lock();
+	list_for_each_entry_rcu(led_cdev, &trig->led_cdevs, trig_list) {
 		if (oneshot)
 			led_blink_set_oneshot(led_cdev, delay_on, delay_off,
 					      invert);
 		else
 			led_blink_set(led_cdev, delay_on, delay_off);
 	}
-	read_unlock_irqrestore(&trig->leddev_list_lock, flags);
+	rcu_read_unlock();
 }
 
 void led_trigger_blink(struct led_trigger *trig,
diff --git a/include/linux/leds.h b/include/linux/leds.h
index 6a8d6409c993..5acc0e8a9f18 100644
--- a/include/linux/leds.h
+++ b/include/linux/leds.h
@@ -356,7 +356,7 @@ struct led_trigger {
 	struct led_hw_trigger_type *trigger_type;
 
 	/* LEDs under control by this trigger (for simple triggers) */
-	rwlock_t	  leddev_list_lock;
+	spinlock_t	  leddev_list_lock;
 	struct list_head  led_cdevs;
 
 	/* Link to next registered trigger */
diff --git a/net/mac80211/led.c b/net/mac80211/led.c
index b275c8853074..5ec5070fe210 100644
--- a/net/mac80211/led.c
+++ b/net/mac80211/led.c
@@ -283,10 +283,10 @@ static void tpt_trig_timer(struct timer_list *t)
 		}
 	}
 
-	read_lock(&local->tpt_led.leddev_list_lock);
+	rcu_read_lock();
 	list_for_each_entry(led_cdev, &local->tpt_led.led_cdevs, trig_list)
 		led_blink_set(led_cdev, &on, &off);
-	read_unlock(&local->tpt_led.leddev_list_lock);
+	rcu_read_unlock();
 }
 
 const char *
@@ -349,10 +349,10 @@ static void ieee80211_stop_tpt_led_trig(struct ieee80211_local *local)
 	tpt_trig->running = false;
 	del_timer_sync(&tpt_trig->timer);
 
-	read_lock(&local->tpt_led.leddev_list_lock);
+	rcu_read_lock();
 	list_for_each_entry(led_cdev, &local->tpt_led.led_cdevs, trig_list)
 		led_set_brightness(led_cdev, LED_OFF);
-	read_unlock(&local->tpt_led.leddev_list_lock);
+	rcu_read_unlock();
 }
 
 void ieee80211_mod_tpt_led_trig(struct ieee80211_local *local,

  reply	other threads:[~2021-03-07  2:09 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-11-02 10:41 [PATCH] leds: trigger: fix potential deadlock with libata Andrea Righi
2020-11-25 12:46 ` Pavel Machek
2020-11-25 14:01   ` Boqun Feng
2020-11-25 14:15   ` Andrea Righi
2020-11-25 15:20     ` Andrea Righi
2021-03-06 20:39 ` Marc Kleine-Budde
2021-03-07  2:02   ` Boqun Feng [this message]
2021-03-07  7:14     ` Andrea Righi
2021-03-08 14:56     ` Marc Kleine-Budde
2021-03-08 15:05       ` Marc Kleine-Budde
2021-03-07 16:13   ` Pavel Machek
2021-03-07 19:04     ` Hans de Goede
2021-03-07 16:26   ` Pavel Machek

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=YEQ0ONQCwVUd0wmc@boqun-archlinux \
    --to=boqun.feng@gmail.com \
    --cc=andrea.righi@canonical.com \
    --cc=dmurphy@ti.com \
    --cc=kernel@pengutronix.de \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-leds@vger.kernel.org \
    --cc=mkl@pengutronix.de \
    --cc=pavel@ucw.cz \
    --cc=schuchmann@schleissheimer.de \
    /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).