[PATCH v3 0/2] Pass down hot plug CONNECTOR ID to user-space

[PATCH v3 0/2] Pass down hot plug CONNECTOR ID to user-space

[Marius Vlad]

Patch series addresses a shortcoming where we're sending a hot plug event
without passing the actual CONNECTOR that caused it. This takes into
consideration both the polling path and the HPD (Hot Plug Detect) path.

v3: Address comments from Dmitry

- guard connector status write with mode_config.mutex
- avoid setting up the connector status and immediately unset it. Do the
  unset in drm_kms_helper_hotplug_event/drm_kms_helper_connector_hotplug_event

v2: Address comments from Daniel

- split patch into 2, one that introduces a bool to track connector
  connection status change and a patch that uses that to be able to send
  hot plug events with the proper CONNECTOR ID to udev and further pass
  that down to user-space
- nuke out mutex when iterating connector list
- fix typo

v2 is at https://lore.kernel.org/dri-devel/20250729165708.9947-1-marius.vlad@collabora.com/

Marius Vlad (2):
  drm: Introduce a new connector status
  drm: Propagate connector status change

 drivers/gpu/drm/drm_connector.c    |  1 +
 drivers/gpu/drm/drm_probe_helper.c | 40 ++++++++++++++++++++++++++----
 drivers/gpu/drm/drm_sysfs.c        |  1 +
 include/drm/drm_connector.h        |  3 +++
 4 files changed, 40 insertions(+), 5 deletions(-)

-- 
2.47.2

[PATCH 1/2] drm: Introduce a new connector status

[Marius Vlad]

This patch introduces a new boolean variable used to track connector's
connect/disconnect status and it is being used on both polling and
the HPD (Hot Plug Detect) paths.

A subsequent patch would make use of this connector status to propagate
per-connector udev hot plug events. This allows user-space to receive
the connector's ID, rather than having a generic hot-plug event for all
connectors, or in the HPD path, just the first one found with a
connection status change.

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
---
 drivers/gpu/drm/drm_connector.c    |  1 +
 drivers/gpu/drm/drm_probe_helper.c | 18 ++++++++++++++++++
 drivers/gpu/drm/drm_sysfs.c        |  1 +
 include/drm/drm_connector.h        |  3 +++
 4 files changed, 23 insertions(+)

diff --git a/drivers/gpu/drm/drm_connector.c b/drivers/gpu/drm/drm_connector.c
index 272d6254ea47..3c6628ee3096 100644
--- a/drivers/gpu/drm/drm_connector.c
+++ b/drivers/gpu/drm/drm_connector.c
@@ -274,6 +274,7 @@ static int drm_connector_init_only(struct drm_device *dev,
 
 	/* provide ddc symlink in sysfs */
 	connector->ddc = ddc;
+	connector->status_changed = false;
 
 	INIT_LIST_HEAD(&connector->head);
 	INIT_LIST_HEAD(&connector->global_connector_list_entry);
diff --git a/drivers/gpu/drm/drm_probe_helper.c b/drivers/gpu/drm/drm_probe_helper.c
index 09b12c30df69..a865d5aa6f73 100644
--- a/drivers/gpu/drm/drm_probe_helper.c
+++ b/drivers/gpu/drm/drm_probe_helper.c
@@ -629,6 +629,9 @@ int drm_helper_probe_single_connector_modes(struct drm_connector *connector,
 			mod_delayed_work(system_wq,
 					 &dev->mode_config.output_poll_work,
 					 0);
+		mutex_lock(&dev->mode_config.mutex);
+		connector->status_changed = true;
+		mutex_unlock(&dev->mode_config.mutex);
 	}
 
 	/*
@@ -732,6 +735,17 @@ EXPORT_SYMBOL(drm_helper_probe_single_connector_modes);
  */
 void drm_kms_helper_hotplug_event(struct drm_device *dev)
 {
+	struct drm_connector *connector;
+	struct drm_connector_list_iter conn_iter;
+
+	drm_connector_list_iter_begin(dev, &conn_iter);
+	drm_for_each_connector_iter(connector, &conn_iter) {
+		mutex_lock(&dev->mode_config.mutex);
+		connector->status_changed = false;
+		mutex_unlock(&dev->mode_config.mutex);
+	}
+	drm_connector_list_iter_end(&conn_iter);
+
 	drm_sysfs_hotplug_event(dev);
 	drm_client_dev_hotplug(dev);
 }
@@ -748,6 +762,10 @@ void drm_kms_helper_connector_hotplug_event(struct drm_connector *connector)
 {
 	struct drm_device *dev = connector->dev;
 
+	mutex_lock(&dev->mode_config.mutex);
+	connector->status_changed = false;
+	mutex_unlock(&dev->mode_config.mutex);
+
 	drm_sysfs_connector_hotplug_event(connector);
 	drm_client_dev_hotplug(dev);
 }
diff --git a/drivers/gpu/drm/drm_sysfs.c b/drivers/gpu/drm/drm_sysfs.c
index b01ffa4d6509..bd9161490116 100644
--- a/drivers/gpu/drm/drm_sysfs.c
+++ b/drivers/gpu/drm/drm_sysfs.c
@@ -199,6 +199,7 @@ static ssize_t status_store(struct device *device,
 		return ret;
 
 	old_force = connector->force;
+	connector->status_changed = true;
 
 	if (sysfs_streq(buf, "detect"))
 		connector->force = 0;
diff --git a/include/drm/drm_connector.h b/include/drm/drm_connector.h
index 8f34f4b8183d..e4310df3d55c 100644
--- a/include/drm/drm_connector.h
+++ b/include/drm/drm_connector.h
@@ -2146,6 +2146,9 @@ struct drm_connector {
 	/** @force: a DRM_FORCE_<foo> state for forced mode sets */
 	enum drm_connector_force force;
 
+	/** @status_changed: if the old status doesn't match current connection status */
+	bool status_changed;
+
 	/**
 	 * @edid_override: Override EDID set via debugfs.
 	 *
-- 
2.47.2

Re: [PATCH 1/2] drm: Introduce a new connector status

[Dmitry Baryshkov]

On Tue, Sep 23, 2025 at 11:36:35AM +0300, Marius Vlad wrote:
> This patch introduces a new boolean variable used to track connector's
> connect/disconnect status and it is being used on both polling and
> the HPD (Hot Plug Detect) paths.

Please see Documentation/process/submitting-patches.rst, it has special
paragraph about "This patch".

> 
> A subsequent patch would make use of this connector status to propagate
> per-connector udev hot plug events. This allows user-space to receive
> the connector's ID, rather than having a generic hot-plug event for all
> connectors, or in the HPD path, just the first one found with a
> connection status change.

It's not clear from the commit message, what is the expected behaviour.
The flag has to be set when we've detected the status change - e.g.
monitor being plugged or unplugged. When is it expected to be cleared?

> 
> Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
> ---
>  drivers/gpu/drm/drm_connector.c    |  1 +
>  drivers/gpu/drm/drm_probe_helper.c | 18 ++++++++++++++++++
>  drivers/gpu/drm/drm_sysfs.c        |  1 +
>  include/drm/drm_connector.h        |  3 +++
>  4 files changed, 23 insertions(+)
> 
> diff --git a/drivers/gpu/drm/drm_connector.c b/drivers/gpu/drm/drm_connector.c
> index 272d6254ea47..3c6628ee3096 100644
> --- a/drivers/gpu/drm/drm_connector.c
> +++ b/drivers/gpu/drm/drm_connector.c
> @@ -274,6 +274,7 @@ static int drm_connector_init_only(struct drm_device *dev,
>  
>  	/* provide ddc symlink in sysfs */
>  	connector->ddc = ddc;
> +	connector->status_changed = false;
>  
>  	INIT_LIST_HEAD(&connector->head);
>  	INIT_LIST_HEAD(&connector->global_connector_list_entry);
> diff --git a/drivers/gpu/drm/drm_probe_helper.c b/drivers/gpu/drm/drm_probe_helper.c
> index 09b12c30df69..a865d5aa6f73 100644
> --- a/drivers/gpu/drm/drm_probe_helper.c
> +++ b/drivers/gpu/drm/drm_probe_helper.c
> @@ -629,6 +629,9 @@ int drm_helper_probe_single_connector_modes(struct drm_connector *connector,
>  			mod_delayed_work(system_wq,
>  					 &dev->mode_config.output_poll_work,
>  					 0);
> +		mutex_lock(&dev->mode_config.mutex);
> +		connector->status_changed = true;
> +		mutex_unlock(&dev->mode_config.mutex);
>  	}
>  
>  	/*
> @@ -732,6 +735,17 @@ EXPORT_SYMBOL(drm_helper_probe_single_connector_modes);
>   */
>  void drm_kms_helper_hotplug_event(struct drm_device *dev)
>  {
> +	struct drm_connector *connector;
> +	struct drm_connector_list_iter conn_iter;
> +
> +	drm_connector_list_iter_begin(dev, &conn_iter);
> +	drm_for_each_connector_iter(connector, &conn_iter) {
> +		mutex_lock(&dev->mode_config.mutex);
> +		connector->status_changed = false;
> +		mutex_unlock(&dev->mode_config.mutex);
> +	}
> +	drm_connector_list_iter_end(&conn_iter);
> +
>  	drm_sysfs_hotplug_event(dev);
>  	drm_client_dev_hotplug(dev);
>  }
> @@ -748,6 +762,10 @@ void drm_kms_helper_connector_hotplug_event(struct drm_connector *connector)
>  {
>  	struct drm_device *dev = connector->dev;
>  
> +	mutex_lock(&dev->mode_config.mutex);
> +	connector->status_changed = false;
> +	mutex_unlock(&dev->mode_config.mutex);
> +
>  	drm_sysfs_connector_hotplug_event(connector);
>  	drm_client_dev_hotplug(dev);
>  }
> diff --git a/drivers/gpu/drm/drm_sysfs.c b/drivers/gpu/drm/drm_sysfs.c
> index b01ffa4d6509..bd9161490116 100644
> --- a/drivers/gpu/drm/drm_sysfs.c
> +++ b/drivers/gpu/drm/drm_sysfs.c
> @@ -199,6 +199,7 @@ static ssize_t status_store(struct device *device,
>  		return ret;
>  
>  	old_force = connector->force;
> +	connector->status_changed = true;
>  
>  	if (sysfs_streq(buf, "detect"))
>  		connector->force = 0;
> diff --git a/include/drm/drm_connector.h b/include/drm/drm_connector.h
> index 8f34f4b8183d..e4310df3d55c 100644
> --- a/include/drm/drm_connector.h
> +++ b/include/drm/drm_connector.h
> @@ -2146,6 +2146,9 @@ struct drm_connector {
>  	/** @force: a DRM_FORCE_<foo> state for forced mode sets */
>  	enum drm_connector_force force;
>  
> +	/** @status_changed: if the old status doesn't match current connection status */
> +	bool status_changed;
> +
>  	/**
>  	 * @edid_override: Override EDID set via debugfs.
>  	 *
> -- 
> 2.47.2
> 

-- 
With best wishes
Dmitry

Re: [PATCH 1/2] drm: Introduce a new connector status

[Marius Vlad]

[-- Attachment #1: Type: text/plain, Size: 4772 bytes --]

Hi Dmitry,
On Tue, Sep 23, 2025 at 06:53:01PM +0300, Dmitry Baryshkov wrote:
> On Tue, Sep 23, 2025 at 11:36:35AM +0300, Marius Vlad wrote:
> > This patch introduces a new boolean variable used to track connector's
> > connect/disconnect status and it is being used on both polling and
> > the HPD (Hot Plug Detect) paths.
> 
> Please see Documentation/process/submitting-patches.rst, it has special
> paragraph about "This patch".
Yes, removed that.
> 
> > 
> > A subsequent patch would make use of this connector status to propagate
> > per-connector udev hot plug events. This allows user-space to receive
> > the connector's ID, rather than having a generic hot-plug event for all
> > connectors, or in the HPD path, just the first one found with a
> > connection status change.
> 
> It's not clear from the commit message, what is the expected behaviour.
> The flag has to be set when we've detected the status change - e.g.
> monitor being plugged or unplugged. When is it expected to be cleared?
Flag should be cleared when firing up KMS uevents and set when detecting
connected/disconnected events. Added a note in the commit desc for v4.
> 
> > 
> > Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
> > ---
> >  drivers/gpu/drm/drm_connector.c    |  1 +
> >  drivers/gpu/drm/drm_probe_helper.c | 18 ++++++++++++++++++
> >  drivers/gpu/drm/drm_sysfs.c        |  1 +
> >  include/drm/drm_connector.h        |  3 +++
> >  4 files changed, 23 insertions(+)
> > 
> > diff --git a/drivers/gpu/drm/drm_connector.c b/drivers/gpu/drm/drm_connector.c
> > index 272d6254ea47..3c6628ee3096 100644
> > --- a/drivers/gpu/drm/drm_connector.c
> > +++ b/drivers/gpu/drm/drm_connector.c
> > @@ -274,6 +274,7 @@ static int drm_connector_init_only(struct drm_device *dev,
> >  
> >  	/* provide ddc symlink in sysfs */
> >  	connector->ddc = ddc;
> > +	connector->status_changed = false;
> >  
> >  	INIT_LIST_HEAD(&connector->head);
> >  	INIT_LIST_HEAD(&connector->global_connector_list_entry);
> > diff --git a/drivers/gpu/drm/drm_probe_helper.c b/drivers/gpu/drm/drm_probe_helper.c
> > index 09b12c30df69..a865d5aa6f73 100644
> > --- a/drivers/gpu/drm/drm_probe_helper.c
> > +++ b/drivers/gpu/drm/drm_probe_helper.c
> > @@ -629,6 +629,9 @@ int drm_helper_probe_single_connector_modes(struct drm_connector *connector,
> >  			mod_delayed_work(system_wq,
> >  					 &dev->mode_config.output_poll_work,
> >  					 0);
> > +		mutex_lock(&dev->mode_config.mutex);
> > +		connector->status_changed = true;
> > +		mutex_unlock(&dev->mode_config.mutex);
> >  	}
> >  
> >  	/*
> > @@ -732,6 +735,17 @@ EXPORT_SYMBOL(drm_helper_probe_single_connector_modes);
> >   */
> >  void drm_kms_helper_hotplug_event(struct drm_device *dev)
> >  {
> > +	struct drm_connector *connector;
> > +	struct drm_connector_list_iter conn_iter;
> > +
> > +	drm_connector_list_iter_begin(dev, &conn_iter);
> > +	drm_for_each_connector_iter(connector, &conn_iter) {
> > +		mutex_lock(&dev->mode_config.mutex);
> > +		connector->status_changed = false;
> > +		mutex_unlock(&dev->mode_config.mutex);
> > +	}
> > +	drm_connector_list_iter_end(&conn_iter);
> > +
> >  	drm_sysfs_hotplug_event(dev);
> >  	drm_client_dev_hotplug(dev);
> >  }
> > @@ -748,6 +762,10 @@ void drm_kms_helper_connector_hotplug_event(struct drm_connector *connector)
> >  {
> >  	struct drm_device *dev = connector->dev;
> >  
> > +	mutex_lock(&dev->mode_config.mutex);
> > +	connector->status_changed = false;
> > +	mutex_unlock(&dev->mode_config.mutex);
> > +
> >  	drm_sysfs_connector_hotplug_event(connector);
> >  	drm_client_dev_hotplug(dev);
> >  }
> > diff --git a/drivers/gpu/drm/drm_sysfs.c b/drivers/gpu/drm/drm_sysfs.c
> > index b01ffa4d6509..bd9161490116 100644
> > --- a/drivers/gpu/drm/drm_sysfs.c
> > +++ b/drivers/gpu/drm/drm_sysfs.c
> > @@ -199,6 +199,7 @@ static ssize_t status_store(struct device *device,
> >  		return ret;
> >  
> >  	old_force = connector->force;
> > +	connector->status_changed = true;
> >  
> >  	if (sysfs_streq(buf, "detect"))
> >  		connector->force = 0;
> > diff --git a/include/drm/drm_connector.h b/include/drm/drm_connector.h
> > index 8f34f4b8183d..e4310df3d55c 100644
> > --- a/include/drm/drm_connector.h
> > +++ b/include/drm/drm_connector.h
> > @@ -2146,6 +2146,9 @@ struct drm_connector {
> >  	/** @force: a DRM_FORCE_<foo> state for forced mode sets */
> >  	enum drm_connector_force force;
> >  
> > +	/** @status_changed: if the old status doesn't match current connection status */
> > +	bool status_changed;
> > +
> >  	/**
> >  	 * @edid_override: Override EDID set via debugfs.
> >  	 *
> > -- 
> > 2.47.2
> > 
> 
> -- 
> With best wishes
> Dmitry

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH 1/2] drm: Introduce a new connector status

[Ian Forbes]

[-- Attachment #1: Type: text/plain, Size: 785 bytes --]

On Tue, Sep 23, 2025 at 3:36 AM Marius Vlad <marius.vlad@collabora.com> wrote:
>
> This patch introduces a new boolean variable used to track connector's
> connect/disconnect status and it is being used on both polling and
> the HPD (Hot Plug Detect) paths.
>
> A subsequent patch would make use of this connector status to propagate
> per-connector udev hot plug events. This allows user-space to receive
> the connector's ID, rather than having a generic hot-plug event for all
> connectors, or in the HPD path, just the first one found with a
> connection status change.
>
> Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
> ---

How do you see this working with virtual connectors which can hotplug
without a connected -> disconnected -> connected cycle?

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 5414 bytes --]

Re: [PATCH 1/2] drm: Introduce a new connector status

[Marius Vlad]

Hi Ian,

On Tue, Sep 23, 2025 at 12:34:58PM -0500, Ian Forbes wrote:
> On Tue, Sep 23, 2025 at 3:36 AM Marius Vlad <marius.vlad@collabora.com> wrote:
> >
> > This patch introduces a new boolean variable used to track connector's
> > connect/disconnect status and it is being used on both polling and
> > the HPD (Hot Plug Detect) paths.
> >
> > A subsequent patch would make use of this connector status to propagate
> > per-connector udev hot plug events. This allows user-space to receive
> > the connector's ID, rather than having a generic hot-plug event for all
> > connectors, or in the HPD path, just the first one found with a
> > connection status change.
> >
> > Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
> > ---
> 
> How do you see this working with virtual connectors which can hotplug
> without a connected -> disconnected -> connected cycle?

tbh I don't see how those hot plug events are generated on virtual
connectors. Do you have a particular driver in mind, or how do you
"simulate" that hotplug cycle?

Pushed today a v4 of this patches series which adds polling in the vkms
driver and with that in I'm able to use sysfs and simulate a hot plug
event, like you'd have regular connectors. I've CC'ed on that v4 if 
you'd like to have another go.

Re: [PATCH 1/2] drm: Introduce a new connector status

[kernel test robot]

Hello,

kernel test robot noticed "WARNING:possible_recursive_locking_detected" on:

commit: 432c7653afa4afe08e59d608db7cbbc321c52149 ("[PATCH 1/2] drm: Introduce a new connector status")
url: https://github.com/intel-lab-lkp/linux/commits/Marius-Vlad/drm-Introduce-a-new-connector-status/20250923-163922
base: git://anongit.freedesktop.org/drm/drm-misc drm-misc-next
patch link: https://lore.kernel.org/all/20250923083636.4749-2-marius.vlad@collabora.com/
patch subject: [PATCH 1/2] drm: Introduce a new connector status

in testcase: boot

config: x86_64-randconfig-074-20250924
compiler: gcc-14
test machine: qemu-system-x86_64 -enable-kvm -cpu SandyBridge -smp 2 -m 16G

(please refer to attached dmesg/kmsg for entire log/backtrace)



If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <oliver.sang@intel.com>
| Closes: https://lore.kernel.org/oe-lkp/202509251410.fdfbcac3-lkp@intel.com


[   19.416092][    T1] WARNING: possible recursive locking detected
[   19.416406][    T1] 6.17.0-rc2-00290-g432c7653afa4 #1 Not tainted
[   19.416406][    T1] --------------------------------------------
[   19.416406][    T1] swapper/0/1 is trying to acquire lock:
[ 19.416406][ T1] ffff88813b9c84d0 (&dev->mode_config.mutex){+.+.}-{4:4}, at: drm_helper_probe_single_connector_modes (drivers/gpu/drm/drm_probe_helper.c:633) 
[   19.416406][    T1]
[   19.416406][    T1] but task is already holding lock:
[ 19.416406][ T1] ffff88813b9c84d0 (&dev->mode_config.mutex){+.+.}-{4:4}, at: drm_client_modeset_probe (drivers/gpu/drm/drm_client_modeset.c:869) 
[   19.416406][    T1]
[   19.416406][    T1] other info that might help us debug this:
[   19.416406][    T1]  Possible unsafe locking scenario:
[   19.416406][    T1]
[   19.416406][    T1]        CPU0
[   19.416406][    T1]        ----
[   19.416406][    T1]   lock(&dev->mode_config.mutex);
[   19.416406][    T1]
[   19.416406][    T1]  *** DEADLOCK ***
[   19.416406][    T1]
[   19.416406][    T1]  May be due to missing lock nesting notation
[   19.416406][    T1]
[   19.416406][    T1] 6 locks held by swapper/0/1:
[ 19.416406][ T1] #0: ffff88813b9c8320 (&dev->clientlist_mutex){+.+.}-{4:4}, at: drm_client_register (include/linux/list.h:169 drivers/gpu/drm/drm_client.c:128) 
[ 19.416406][ T1] #1: ffff888108a882a0 (&helper->lock){+.+.}-{4:4}, at: drm_fb_helper_initial_config (drivers/gpu/drm/drm_fb_helper.c:1917) 
[ 19.416406][ T1] #2: ffff888108a88098 (&client->modeset_mutex){+.+.}-{4:4}, at: drm_client_modeset_probe (drivers/gpu/drm/drm_client_modeset.c:867) 
[ 19.416406][ T1] #3: ffff88813b9c84d0 (&dev->mode_config.mutex){+.+.}-{4:4}, at: drm_client_modeset_probe (drivers/gpu/drm/drm_client_modeset.c:869) 
[ 19.416406][ T1] #4: ffffc9000001fa78 (crtc_ww_class_acquire){+.+.}-{0:0}, at: drm_helper_probe_single_connector_modes (drivers/gpu/drm/drm_probe_helper.c:572) 
[ 19.416406][ T1] #5: ffffc9000001faa8 (crtc_ww_class_mutex){+.+.}-{4:4}, at: drm_helper_probe_single_connector_modes (drivers/gpu/drm/drm_probe_helper.c:572) 
[   19.416406][    T1]
[   19.416406][    T1] stack backtrace:
[   19.416406][    T1] CPU: 0 UID: 0 PID: 1 Comm: swapper/0 Not tainted 6.17.0-rc2-00290-g432c7653afa4 #1 PREEMPT(none)
[   19.416406][    T1] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.3-debian-1.16.3-2 04/01/2014
[   19.416406][    T1] Call Trace:
[   19.416406][    T1]  <TASK>
[ 19.416406][ T1] dump_stack_lvl (lib/dump_stack.c:122) 
[ 19.416406][ T1] print_deadlock_bug.cold (kernel/locking/lockdep.c:3044) 
[ 19.416406][ T1] validate_chain (kernel/locking/lockdep.c:3898) 
[ 19.416406][ T1] __lock_acquire (kernel/locking/lockdep.c:5237 (discriminator 1)) 
[ 19.416406][ T1] lock_acquire (kernel/locking/lockdep.c:470 kernel/locking/lockdep.c:5870) 
[ 19.416406][ T1] ? drm_helper_probe_single_connector_modes (drivers/gpu/drm/drm_probe_helper.c:633) 
[ 19.416406][ T1] ? sched_show_task (kernel/sched/core.c:8912) 
[ 19.416406][ T1] ? look_up_lock_class (kernel/locking/lockdep.c:933 (discriminator 28)) 
[ 19.416406][ T1] __mutex_lock (arch/x86/include/asm/atomic.h:23 include/linux/atomic/atomic-arch-fallback.h:457 include/linux/jump_label.h:262 include/trace/events/lock.h:95 kernel/locking/mutex.c:600 kernel/locking/mutex.c:760) 
[ 19.416406][ T1] ? drm_helper_probe_single_connector_modes (drivers/gpu/drm/drm_probe_helper.c:633) 
[ 19.416406][ T1] ? look_up_lock_class (kernel/locking/lockdep.c:933 (discriminator 28)) 
[ 19.416406][ T1] ? mutex_lock_io_nested (kernel/locking/mutex.c:759) 
[ 19.416406][ T1] ? drm_print_bits (drivers/gpu/drm/drm_print.c:334) 
[ 19.416406][ T1] ? ww_mutex_lock (kernel/locking/mutex.c:887) 
[ 19.416406][ T1] ? drm_helper_probe_single_connector_modes (drivers/gpu/drm/drm_probe_helper.c:633) 
[ 19.416406][ T1] drm_helper_probe_single_connector_modes (drivers/gpu/drm/drm_probe_helper.c:633) 
[ 19.416406][ T1] ? __drm_helper_update_and_validate (drivers/gpu/drm/drm_probe_helper.c:561) 
[ 19.416406][ T1] ? __kmalloc_noprof (include/trace/events/kmem.h:54 (discriminator 2) mm/slub.c:4366 (discriminator 2) mm/slub.c:4377 (discriminator 2)) 
[ 19.416406][ T1] drm_client_modeset_probe (drivers/gpu/drm/drm_client_modeset.c:869 (discriminator 1)) 
[ 19.416406][ T1] ? drm_client_firmware_config (drivers/gpu/drm/drm_client_modeset.c:818) 
[ 19.416406][ T1] __drm_fb_helper_initial_config_and_unlock (drivers/gpu/drm/drm_fb_helper.c:1830) 
[ 19.416406][ T1] drm_fbdev_client_hotplug (drivers/gpu/drm/clients/drm_fbdev_client.c:53) 
[ 19.416406][ T1] drm_client_register (drivers/gpu/drm/drm_client.c:141) 
[ 19.416406][ T1] drm_fbdev_client_setup (drivers/gpu/drm/clients/drm_fbdev_client.c:167) 
[ 19.416406][ T1] drm_client_setup (drivers/gpu/drm/clients/drm_client_setup.c:47) 
[ 19.416406][ T1] vkms_create (drivers/gpu/drm/vkms/vkms_drv.c:203) 
[ 19.416406][ T1] ? drm_display_helper_module_init (drivers/gpu/drm/vkms/vkms_drv.c:213) 
[ 19.416406][ T1] vkms_init (drivers/gpu/drm/vkms/vkms_drv.c:221) 
[ 19.416406][ T1] do_one_initcall (init/main.c:1269) 
[ 19.416406][ T1] ? trace_event_raw_event_initcall_level (init/main.c:1260) 
[ 19.416406][ T1] do_initcalls (init/main.c:1330 (discriminator 3) init/main.c:1347 (discriminator 3)) 
[ 19.416406][ T1] kernel_init_freeable (init/main.c:1583) 
[ 19.416406][ T1] ? rest_init (init/main.c:1461) 
[ 19.416406][ T1] kernel_init (init/main.c:1471) 
[ 19.416406][ T1] ? rest_init (init/main.c:1461) 
[ 19.416406][ T1] ret_from_fork (arch/x86/kernel/process.c:154) 
[ 19.416406][ T1] ? rest_init (init/main.c:1461) 
[ 19.416406][ T1] ret_from_fork_asm (arch/x86/entry/entry_64.S:255) 
[   19.416406][    T1]  </TASK>
[  984.236393][   T23] INFO: task swapper/0:1 blocked for more than 491 seconds.
[  984.237354][   T23]       Not tainted 6.17.0-rc2-00290-g432c7653afa4 #1
[  984.238322][   T23] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[  984.239375][   T23] task:swapper/0       state:D stack:0     pid:1     tgid:1     ppid:0      task_flags:0x0140 flags:0x00004000
[  984.240834][   T23] Call Trace:
[  984.241228][   T23]  <TASK>
[ 984.241634][ T23] __schedule (kernel/sched/core.c:5357 kernel/sched/core.c:6961) 
[ 984.242202][ T23] ? io_schedule_timeout (kernel/sched/core.c:6817) 
[ 984.242846][ T23] ? lock_acquire (include/trace/events/lock.h:24 (discriminator 2) kernel/locking/lockdep.c:5831 (discriminator 2)) 
[ 984.243441][ T23] schedule (kernel/sched/core.c:7044 kernel/sched/core.c:7058) 
[ 984.243975][ T23] schedule_preempt_disabled (kernel/sched/core.c:7116) 
[ 984.244639][ T23] __mutex_lock (kernel/locking/mutex.c:183 kernel/locking/mutex.c:678 kernel/locking/mutex.c:760) 
[ 984.245258][ T23] ? drm_helper_probe_single_connector_modes (drivers/gpu/drm/drm_probe_helper.c:633) 
[ 984.246160][ T23] ? mutex_lock_io_nested (kernel/locking/mutex.c:759) 
[ 984.246876][ T23] ? drm_print_bits (drivers/gpu/drm/drm_print.c:334) 
[ 984.247486][ T23] ? ww_mutex_lock (kernel/locking/mutex.c:887) 
[ 984.248132][ T23] ? ww_mutex_lock (kernel/locking/mutex.c:887) 
[ 984.248681][ T23] ? drm_helper_probe_single_connector_modes (drivers/gpu/drm/drm_probe_helper.c:633) 
[ 984.249557][ T23] drm_helper_probe_single_connector_modes (drivers/gpu/drm/drm_probe_helper.c:633) 
[ 984.250457][ T23] ? __drm_helper_update_and_validate (drivers/gpu/drm/drm_probe_helper.c:561) 
[ 984.251275][ T23] ? __kmalloc_noprof (include/trace/events/kmem.h:54 (discriminator 2) mm/slub.c:4366 (discriminator 2) mm/slub.c:4377 (discriminator 2)) 
[ 984.251990][ T23] drm_client_modeset_probe (drivers/gpu/drm/drm_client_modeset.c:869 (discriminator 1)) 
[ 984.252659][ T23] ? drm_client_firmware_config (drivers/gpu/drm/drm_client_modeset.c:818) 
[ 984.253433][ T23] __drm_fb_helper_initial_config_and_unlock (drivers/gpu/drm/drm_fb_helper.c:1830) 
[ 984.254309][ T23] drm_fbdev_client_hotplug (drivers/gpu/drm/clients/drm_fbdev_client.c:53) 
[ 984.255079][ T23] drm_client_register (drivers/gpu/drm/drm_client.c:141) 
[ 984.255737][ T23] drm_fbdev_client_setup (drivers/gpu/drm/clients/drm_fbdev_client.c:167) 
[ 984.256474][ T23] drm_client_setup (drivers/gpu/drm/clients/drm_client_setup.c:47) 
[ 984.257080][ T23] vkms_create (drivers/gpu/drm/vkms/vkms_drv.c:203) 
[ 984.257646][ T23] ? drm_display_helper_module_init (drivers/gpu/drm/vkms/vkms_drv.c:213) 
[ 984.258477][ T23] vkms_init (drivers/gpu/drm/vkms/vkms_drv.c:221) 
[ 984.258979][ T23] do_one_initcall (init/main.c:1269) 
[ 984.259580][ T23] ? trace_event_raw_event_initcall_level (init/main.c:1260) 
[ 984.260457][ T23] do_initcalls (init/main.c:1330 (discriminator 3) init/main.c:1347 (discriminator 3)) 
[ 984.261031][ T23] kernel_init_freeable (init/main.c:1583) 
[ 984.261730][ T23] ? rest_init (init/main.c:1461) 
[ 984.262250][ T23] kernel_init (init/main.c:1471) 
[ 984.262775][ T23] ? rest_init (init/main.c:1461) 
[ 984.263366][ T23] ret_from_fork (arch/x86/kernel/process.c:154) 
[ 984.263966][ T23] ? rest_init (init/main.c:1461) 
[ 984.264445][ T23] ret_from_fork_asm (arch/x86/entry/entry_64.S:255) 
[  984.265087][   T23]  </TASK>
[  984.265460][   T23] INFO: task swapper/0:1 is blocked on a mutex likely owned by task swapper/0:1.
[  984.266569][   T23] task:swapper/0       state:D stack:0     pid:1     tgid:1     ppid:0      task_flags:0x0140 flags:0x00004000
[  984.267995][   T23] Call Trace:
[  984.268440][   T23]  <TASK>
[ 984.268818][ T23] __schedule (kernel/sched/core.c:5357 kernel/sched/core.c:6961) 
[ 984.269392][ T23] ? io_schedule_timeout (kernel/sched/core.c:6817) 
[ 984.270076][ T23] ? lock_acquire (include/trace/events/lock.h:24 (discriminator 2) kernel/locking/lockdep.c:5831 (discriminator 2)) 
[ 984.270650][ T23] schedule (kernel/sched/core.c:7044 kernel/sched/core.c:7058) 
[ 984.271187][ T23] schedule_preempt_disabled (kernel/sched/core.c:7116) 
[ 984.271938][ T23] __mutex_lock (kernel/locking/mutex.c:183 kernel/locking/mutex.c:678 kernel/locking/mutex.c:760) 
[ 984.272512][ T23] ? drm_helper_probe_single_connector_modes (drivers/gpu/drm/drm_probe_helper.c:633) 
[ 984.273417][ T23] ? mutex_lock_io_nested (kernel/locking/mutex.c:759) 
[ 984.274100][ T23] ? drm_print_bits (drivers/gpu/drm/drm_print.c:334) 
[ 984.274700][ T23] ? ww_mutex_lock (kernel/locking/mutex.c:887) 
[ 984.275321][ T23] ? ww_mutex_lock (kernel/locking/mutex.c:887) 
[ 984.275900][ T23] ? drm_helper_probe_single_connector_modes (drivers/gpu/drm/drm_probe_helper.c:633) 
[ 984.276788][ T23] drm_helper_probe_single_connector_modes (drivers/gpu/drm/drm_probe_helper.c:633) 
[ 984.277636][ T23] ? __drm_helper_update_and_validate (drivers/gpu/drm/drm_probe_helper.c:561) 
[ 984.278471][ T23] ? __kmalloc_noprof (include/trace/events/kmem.h:54 (discriminator 2) mm/slub.c:4366 (discriminator 2) mm/slub.c:4377 (discriminator 2)) 
[ 984.279094][ T23] drm_client_modeset_probe (drivers/gpu/drm/drm_client_modeset.c:869 (discriminator 1)) 
[ 984.279852][ T23] ? drm_client_firmware_config (drivers/gpu/drm/drm_client_modeset.c:818) 
[ 984.280602][ T23] __drm_fb_helper_initial_config_and_unlock (drivers/gpu/drm/drm_fb_helper.c:1830) 
[ 984.281537][ T23] drm_fbdev_client_hotplug (drivers/gpu/drm/clients/drm_fbdev_client.c:53) 
[ 984.282238][ T23] drm_client_register (drivers/gpu/drm/drm_client.c:141) 
[ 984.282887][ T23] drm_fbdev_client_setup (drivers/gpu/drm/clients/drm_fbdev_client.c:167) 
[ 984.283611][ T23] drm_client_setup (drivers/gpu/drm/clients/drm_client_setup.c:47) 
[ 984.284232][ T23] vkms_create (drivers/gpu/drm/vkms/vkms_drv.c:203) 
[ 984.284846][ T23] ? drm_display_helper_module_init (drivers/gpu/drm/vkms/vkms_drv.c:213) 
[ 984.285609][ T23] vkms_init (drivers/gpu/drm/vkms/vkms_drv.c:221) 
[ 984.286136][ T23] do_one_initcall (init/main.c:1269) 
[ 984.286773][ T23] ? trace_event_raw_event_initcall_level (init/main.c:1260) 
[ 984.287581][ T23] do_initcalls (init/main.c:1330 (discriminator 3) init/main.c:1347 (discriminator 3)) 
[ 984.288212][ T23] kernel_init_freeable (init/main.c:1583) 
[ 984.288856][ T23] ? rest_init (init/main.c:1461) 
[ 984.289412][ T23] kernel_init (init/main.c:1471) 
[ 984.289983][ T23] ? rest_init (init/main.c:1461) 
[ 984.290537][ T23] ret_from_fork (arch/x86/kernel/process.c:154) 
[ 984.291124][ T23] ? rest_init (init/main.c:1461) 
[ 984.291735][ T23] ret_from_fork_asm (arch/x86/entry/entry_64.S:255) 
[  984.292371][   T23]  </TASK>
[  984.292826][   T23] INFO: lockdep is turned off.
[ 1045.677903][    C0] BUG: workqueue lockup - pool cpus=0 node=0 flags=0x0 nice=0 stuck for 59s!
[ 1045.680667][    C0] Showing busy workqueues and worker pools:
[ 1045.684192][    C0] workqueue events_power_efficient: flags=0x80
[ 1045.684971][    C0]   pwq 2: cpus=0 node=0 flags=0x0 nice=0 active=3 refcnt=4
[ 1045.684987][    C0]     pending: neigh_managed_work, neigh_periodic_work, do_cache_clean
[ 1045.685083][    C0] Showing backtraces of running workers in stalled CPU-bound worker pools:
[ 1076.396346][    C0] BUG: workqueue lockup - pool cpus=0 node=0 flags=0x0 nice=0 stuck for 89s!
[ 1076.400410][    C0] Showing busy workqueues and worker pools:
[ 1076.401145][    C0] workqueue events: flags=0x0
[ 1076.401765][    C0]   pwq 2: cpus=0 node=0 flags=0x0 nice=0 active=1 refcnt=2
[ 1076.401781][    C0]     pending: stop_one_cpu_nowait_workfn
[ 1076.401795][    C0] workqueue events_power_efficient: flags=0x80
[ 1076.404125][    C0]   pwq 2: cpus=0 node=0 flags=0x0 nice=0 active=3 refcnt=4
[ 1076.404141][    C0]     pending: neigh_managed_work, neigh_periodic_work, do_cache_clean
[ 1076.411517][    C0] Showing backtraces of running workers in stalled CPU-bound worker pools:



The kernel config and materials to reproduce are available at:
https://download.01.org/0day-ci/archive/20250925/202509251410.fdfbcac3-lkp@intel.com



-- 
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki

[PATCH 2/2] drm: Propagate connector status change

[Marius Vlad]

On the HPD (Hot Plug Detect) path this change makes use of the connector
status to notify all connectors, rather than just first one found that
suffered a status change.

Similarly on the polling side, this also takes into consideration
sending per-connector udev hot plug events.

Signed-off-by: Marius Vlad <marius.vlad@collabora.com>
---
 drivers/gpu/drm/drm_probe_helper.c | 22 +++++++++++++++++-----
 1 file changed, 17 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/drm_probe_helper.c b/drivers/gpu/drm/drm_probe_helper.c
index a865d5aa6f73..98afab9f15e9 100644
--- a/drivers/gpu/drm/drm_probe_helper.c
+++ b/drivers/gpu/drm/drm_probe_helper.c
@@ -862,8 +862,14 @@ static void output_poll_execute(struct work_struct *work)
 	mutex_unlock(&dev->mode_config.mutex);
 
 out:
-	if (changed)
-		drm_kms_helper_hotplug_event(dev);
+	if (changed) {
+		drm_connector_list_iter_begin(dev, &conn_iter);
+		drm_for_each_connector_iter(connector, &conn_iter) {
+			if (connector->status_changed)
+				drm_kms_helper_connector_hotplug_event(connector);
+		}
+		drm_connector_list_iter_end(&conn_iter);
+	}
 
 	if (repoll)
 		schedule_delayed_work(delayed_work, DRM_OUTPUT_POLL_PERIOD);
@@ -1125,10 +1131,16 @@ bool drm_helper_hpd_irq_event(struct drm_device *dev)
 	drm_connector_list_iter_end(&conn_iter);
 	mutex_unlock(&dev->mode_config.mutex);
 
-	if (changed == 1)
+	if (changed == 1) {
 		drm_kms_helper_connector_hotplug_event(first_changed_connector);
-	else if (changed > 0)
-		drm_kms_helper_hotplug_event(dev);
+	} else if (changed > 0) {
+		drm_connector_list_iter_begin(dev, &conn_iter);
+		drm_for_each_connector_iter(connector, &conn_iter) {
+			if (connector->status_changed)
+				drm_kms_helper_connector_hotplug_event(connector);
+		}
+		drm_connector_list_iter_end(&conn_iter);
+	}
 
 	if (first_changed_connector)
 		drm_connector_put(first_changed_connector);
-- 
2.47.2