[v2] drm: Allow determining if current task is output poll worker

Submitted by Lukas Wunner on Feb. 14, 2018, 7:41 a.m.

Details

Message ID 3549ce32e7f1467102e70d3e9cbf70c46bfe108e.1518593424.git.lukas@wunner.de
State New
Headers show
Series "Fix deadlock on runtime suspend in DRM drivers" ( rev: 3 ) in Nouveau

Not browsing as part of any series.

Commit Message

Lukas Wunner Feb. 14, 2018, 7:41 a.m.
Introduce a helper to determine if the current task is an output poll
worker.

This allows us to fix a long-standing deadlock in several DRM drivers
wherein the ->runtime_suspend callback waits for the output poll worker
to finish and the worker in turn calls a ->detect callback which waits
for runtime suspend to finish.  The ->detect callback is invoked from
multiple call sites and waiting for runtime suspend to finish is the
correct thing to do except if it's executing in the context of the
worker.

v2: Expand kerneldoc to specifically mention deadlock between
    output poll worker and autosuspend worker as use case. (Lyude)

Cc: Dave Airlie <airlied@redhat.com>
Cc: Ben Skeggs <bskeggs@redhat.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Lyude Paul <lyude@redhat.com>
Signed-off-by: Lukas Wunner <lukas@wunner.de>
---
 drivers/gpu/drm/drm_probe_helper.c | 20 ++++++++++++++++++++
 include/drm/drm_crtc_helper.h      |  1 +
 2 files changed, 21 insertions(+)

Patch hide | download patch | download mbox

diff --git a/drivers/gpu/drm/drm_probe_helper.c b/drivers/gpu/drm/drm_probe_helper.c
index 6dc2dde5b672..7a6b2dc08913 100644
--- a/drivers/gpu/drm/drm_probe_helper.c
+++ b/drivers/gpu/drm/drm_probe_helper.c
@@ -654,6 +654,26 @@  static void output_poll_execute(struct work_struct *work)
 		schedule_delayed_work(delayed_work, DRM_OUTPUT_POLL_PERIOD);
 }
 
+/**
+ * drm_kms_helper_is_poll_worker - is %current task an output poll worker?
+ *
+ * Determine if %current task is an output poll worker.  This can be used
+ * to select distinct code paths for output polling versus other contexts.
+ *
+ * One use case is to avoid a deadlock between the output poll worker and
+ * the autosuspend worker wherein the latter waits for polling to finish
+ * upon calling drm_kms_helper_poll_disable(), while the former waits for
+ * runtime suspend to finish upon calling pm_runtime_get_sync() in a
+ * connector ->detect hook.
+ */
+bool drm_kms_helper_is_poll_worker(void)
+{
+	struct work_struct *work = current_work();
+
+	return work && work->func == output_poll_execute;
+}
+EXPORT_SYMBOL(drm_kms_helper_is_poll_worker);
+
 /**
  * drm_kms_helper_poll_disable - disable output polling
  * @dev: drm_device
diff --git a/include/drm/drm_crtc_helper.h b/include/drm/drm_crtc_helper.h
index 76e237bd989b..6914633037a5 100644
--- a/include/drm/drm_crtc_helper.h
+++ b/include/drm/drm_crtc_helper.h
@@ -77,5 +77,6 @@  void drm_kms_helper_hotplug_event(struct drm_device *dev);
 
 void drm_kms_helper_poll_disable(struct drm_device *dev);
 void drm_kms_helper_poll_enable(struct drm_device *dev);
+bool drm_kms_helper_is_poll_worker(void);
 
 #endif

Comments

I think your idea of having the extra kerneldoc as a seperate patch to make
this easier to backport should work fine :). Thanks for the good work!

Reviewed-by: Lyude Paul <lyude@redhat.com>

On Wed, 2018-02-14 at 08:41 +0100, Lukas Wunner wrote:
> Introduce a helper to determine if the current task is an output poll
> worker.
> 
> This allows us to fix a long-standing deadlock in several DRM drivers
> wherein the ->runtime_suspend callback waits for the output poll worker
> to finish and the worker in turn calls a ->detect callback which waits
> for runtime suspend to finish.  The ->detect callback is invoked from
> multiple call sites and waiting for runtime suspend to finish is the
> correct thing to do except if it's executing in the context of the
> worker.
> 
> v2: Expand kerneldoc to specifically mention deadlock between
>     output poll worker and autosuspend worker as use case. (Lyude)
> 
> Cc: Dave Airlie <airlied@redhat.com>
> Cc: Ben Skeggs <bskeggs@redhat.com>
> Cc: Alex Deucher <alexander.deucher@amd.com>
> Reviewed-by: Lyude Paul <lyude@redhat.com>
> Signed-off-by: Lukas Wunner <lukas@wunner.de>
> ---
>  drivers/gpu/drm/drm_probe_helper.c | 20 ++++++++++++++++++++
>  include/drm/drm_crtc_helper.h      |  1 +
>  2 files changed, 21 insertions(+)
> 
> diff --git a/drivers/gpu/drm/drm_probe_helper.c
> b/drivers/gpu/drm/drm_probe_helper.c
> index 6dc2dde5b672..7a6b2dc08913 100644
> --- a/drivers/gpu/drm/drm_probe_helper.c
> +++ b/drivers/gpu/drm/drm_probe_helper.c
> @@ -654,6 +654,26 @@ static void output_poll_execute(struct work_struct
> *work)
>  		schedule_delayed_work(delayed_work,
> DRM_OUTPUT_POLL_PERIOD);
>  }
>  
> +/**
> + * drm_kms_helper_is_poll_worker - is %current task an output poll worker?
> + *
> + * Determine if %current task is an output poll worker.  This can be used
> + * to select distinct code paths for output polling versus other contexts.
> + *
> + * One use case is to avoid a deadlock between the output poll worker and
> + * the autosuspend worker wherein the latter waits for polling to finish
> + * upon calling drm_kms_helper_poll_disable(), while the former waits for
> + * runtime suspend to finish upon calling pm_runtime_get_sync() in a
> + * connector ->detect hook.
> + */
> +bool drm_kms_helper_is_poll_worker(void)
> +{
> +	struct work_struct *work = current_work();
> +
> +	return work && work->func == output_poll_execute;
> +}
> +EXPORT_SYMBOL(drm_kms_helper_is_poll_worker);
> +
>  /**
>   * drm_kms_helper_poll_disable - disable output polling
>   * @dev: drm_device
> diff --git a/include/drm/drm_crtc_helper.h b/include/drm/drm_crtc_helper.h
> index 76e237bd989b..6914633037a5 100644
> --- a/include/drm/drm_crtc_helper.h
> +++ b/include/drm/drm_crtc_helper.h
> @@ -77,5 +77,6 @@ void drm_kms_helper_hotplug_event(struct drm_device *dev);
>  
>  void drm_kms_helper_poll_disable(struct drm_device *dev);
>  void drm_kms_helper_poll_enable(struct drm_device *dev);
> +bool drm_kms_helper_is_poll_worker(void);
>  
>  #endif