Skip to contentSkip to conversation

drm: kirin: Add a mutex to avoid fb initialization race

RFC PATCH Superseded John Stultz 2 participants 4 messages Feb 23, 2017 – Feb 25, 2017

Subsystems (3)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

RFC PATCH 1/1 drm: kirin: Add a mutex to avoid fb initialization race

In some cases I've been seeing a race where two framebuffers would be initialized, as kirin_fbdev_output_poll_changed() might get called quickly in succession, resulting in the fb initialization happening twice. This could cause the system to boot up with a blank screen.

This patch adds a simple mutex to serialize it and seems to avoid the race.

Obviously I suspect this patch isn't the best solution, but I wanted to send it out as something concrete to discuss the bug.

Suggestions or feedback for a better solution would be greatly appreciated.

Cc Xinliang Liu <[email protected]>
Cc Rongrong Zou <[email protected]>
Cc Xinwei Kong <[email protected]>
Cc Chen Feng <[email protected]>
Cc David Airlie <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Sean Paul <[email protected]>
---
 drivers/gpu/drm/hisilicon/kirin/kirin_drm_drv.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/gpu/drm/hisilicon/kirin/kirin_drm_drv.c +3 -0
index ebd5f4f..80c607f 100644
@@ -50,11 +50,13 @@ static int kirin_drm_kms_cleanup(struct drm_device *dev)
 	return 0;
 }
 
+static DEFINE_MUTEX(fb_lock);
 #ifdef CONFIG_DRM_FBDEV_EMULATION
 static void kirin_fbdev_output_poll_changed(struct drm_device *dev)
 {
 	struct kirin_drm_private *priv = dev->dev_private;
 
+	mutex_lock(&fb_lock);
 	if (priv->fbdev) {
 		drm_fbdev_cma_hotplug_event(priv->fbdev);
 	} else {
@@ -64,6 +66,7 @@ static void kirin_fbdev_output_poll_changed(struct drm_device *dev)
 		if (IS_ERR(priv->fbdev))
 			priv->fbdev = NULL;
 	}
+	mutex_unlock(&fb_lock);
 }
 #endif
 
Signature · 6 lines
2.7.4

_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel
#1 lore raw

Hi John,

Quoted 4 lines · John Stultz

In some cases I've been seeing a race where two framebuffers would be initialized, as kirin_fbdev_output_poll_changed() might get called quickly in succession, resulting in the fb initialization happening twice. This could cause the system

I might understand this race. This because two places call drm_helper_hpd_irq_event might cause the race:
One place is here
static int kirin_drm_kms_init(struct drm_device *dev)
{
...

    /* force detection after connectors init */
    (void)drm_helper_hpd_irq_event(dev);

...
}

another is the adv7533 interrupt thread handler
static int adv7511_irq_process(struct adv7511 *adv7511, bool process_hpd) {
...

    if (process_hpd && irq0 & ADV7511_INT0_HPD && adv7511->bridge.encoder)
        drm_helper_hpd_irq_event(adv7511->connector.dev);

...
}

right?

I don't get a better way to fix this yet , I like to put fb_lock into kirin_drm_private. And it seems hard to fix this in the core drm_helper_hpd_irq_event.

-xinliang

Quoted 51 lines

to boot up with a blank screen.

This patch adds a simple mutex to serialize it and seems to avoid the race.

Obviously I suspect this patch isn't the best solution, but I wanted to send it out as something concrete to discuss the bug.

Suggestions or feedback for a better solution would be greatly appreciated.

Cc Xinliang Liu <[email protected]>
Cc Rongrong Zou <[email protected]>
Cc Xinwei Kong <[email protected]>
Cc Chen Feng <[email protected]>
Cc David Airlie <[email protected]>
Cc Daniel Vetter <[email protected]>
Cc Sean Paul <[email protected]>
---
  drivers/gpu/drm/hisilicon/kirin/kirin_drm_drv.c | 3 +++
  1 file changed, 3 insertions(+)
drivers/gpu/drm/hisilicon/kirin/kirin_drm_drv.c +3 -0
index ebd5f4f..80c607f 100644
@@ -50,11 +50,13 @@ static int kirin_drm_kms_cleanup(struct drm_device *dev)
  	return 0;
  }
 
+static DEFINE_MUTEX(fb_lock);
  #ifdef CONFIG_DRM_FBDEV_EMULATION
  static void kirin_fbdev_output_poll_changed(struct drm_device *dev)
  {
  	struct kirin_drm_private *priv = dev->dev_private;
 
+	mutex_lock(&fb_lock);
  	if (priv->fbdev) {
  		drm_fbdev_cma_hotplug_event(priv->fbdev);
  	} else {
@@ -64,6 +66,7 @@ static void kirin_fbdev_output_poll_changed(struct drm_device *dev)
  		if (IS_ERR(priv->fbdev))
  			priv->fbdev = NULL;
  	}
+	mutex_unlock(&fb_lock);
  }
  #endif
 
#2 lore raw
Quoted 27 lines · liuxinliang

Hi John,

Quoted 5 lines · John Stultz

In some cases I've been seeing a race where two framebuffers would be initialized, as kirin_fbdev_output_poll_changed() might get called quickly in succession, resulting in the fb initialization happening twice. This could cause the system

I might understand this race. This because two places call drm_helper_hpd_irq_event might cause the race:
One place is here
static int kirin_drm_kms_init(struct drm_device *dev) {
...

    /* force detection after connectors init */
    (void)drm_helper_hpd_irq_event(dev);

...
}

another is the adv7533 interrupt thread handler
static int adv7511_irq_process(struct adv7511 *adv7511, bool process_hpd) {
...

    if (process_hpd && irq0 & ADV7511_INT0_HPD && adv7511->bridge.encoder)
        drm_helper_hpd_irq_event(adv7511->connector.dev);
Quoted 7 lines

...
}

right?

I don't get a better way to fix this yet , I like to put fb_lock into kirin_drm_private.

Ok. I've moved the mutex to the kirin_drm_private structure.

Anything else you'd like to see before I resend?

thanks
-john
_______________________________________________
dri-devel mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/dri-devel

#3 lore raw
Quoted 30 lines · John Stultz
Quoted 28 lines · liuxinliang

Hi John,

Quoted 4 lines · John Stultz

In some cases I've been seeing a race where two framebuffers would be initialized, as kirin_fbdev_output_poll_changed() might get called quickly in succession, resulting in the fb initialization happening twice. This could cause the system

I might understand this race. This because two places call drm_helper_hpd_irq_event might cause the race:
One place is here
static int kirin_drm_kms_init(struct drm_device *dev) {
...

     /* force detection after connectors init */
     (void)drm_helper_hpd_irq_event(dev);

...
}

another is the adv7533 interrupt thread handler
static int adv7511_irq_process(struct adv7511 *adv7511, bool process_hpd) {
...

     if (process_hpd && irq0 & ADV7511_INT0_HPD && adv7511->bridge.encoder)
         drm_helper_hpd_irq_event(adv7511->connector.dev);

...
}

Quoted 7 lines
Quoted 4 lines

right?

I don't get a better way to fix this yet , I like to put fb_lock into kirin_drm_private.

Ok. I've moved the mutex to the kirin_drm_private structure.

Anything else you'd like to see before I resend?

Thanks, no more then.

Best,
-xinliang

Quoted 6 lines

thanks
-john

.