[Intel-gfx] [PATCH] drm/i915: Cancel outstanding modeset workers before suspend

[Daniel Vetter]

On Tue, Aug 06, 2013 at 01:24:02PM +0100, Chris Wilson wrote:
> Upon resume we will do a complete restoration of the mode and so reset
> all tasks, but during suspend (and unload) we need to make sure that no
> workers run concurrently with our suspend code. Or worse just after.
> 
> The issue was first raised whilst tackling a suspend issue with Takashi
> Iwai (http://lists.freedesktop.org/archives/intel-gfx/2012-April/016738.html)
> and then it was independently rediscovered by Chuanshen Lui.
> 
> v2: Rebase for the lost year.
> 
> Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> Cc: Takashi Iwai <tiwai at suse.de>
> Cc: "Liu, Chuansheng" <chuansheng.liu at intel.com>
> ---
>  drivers/gpu/drm/i915/i915_drv.h      |  1 +
>  drivers/gpu/drm/i915/intel_display.c | 21 ++++++++++++++++-----
>  2 files changed, 17 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
> index 64b5d55..cb2c59d 100644
> --- a/drivers/gpu/drm/i915/i915_drv.h
> +++ b/drivers/gpu/drm/i915/i915_drv.h
> @@ -2130,6 +2130,7 @@ extern void intel_modeset_init_hw(struct drm_device *dev);
>  extern void intel_modeset_suspend_hw(struct drm_device *dev);
>  extern void intel_modeset_init(struct drm_device *dev);
>  extern void intel_modeset_gem_init(struct drm_device *dev);
> +extern void intel_modeset_quiesce(struct drm_device *dev);
>  extern void intel_modeset_cleanup(struct drm_device *dev);
>  extern int intel_modeset_vga_set_state(struct drm_device *dev, bool state);
>  extern void intel_modeset_setup_hw_state(struct drm_device *dev,
> diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
> index 7c7325e..0584480 100644
> --- a/drivers/gpu/drm/i915/intel_display.c
> +++ b/drivers/gpu/drm/i915/intel_display.c
> @@ -9891,6 +9891,7 @@ void intel_modeset_init_hw(struct drm_device *dev)
>  
>  void intel_modeset_suspend_hw(struct drm_device *dev)
>  {
> +	intel_modeset_quiesce(dev);
>  	intel_suspend_hw(dev);
>  }
>  
> @@ -10324,9 +10325,19 @@ void intel_modeset_gem_init(struct drm_device *dev)
>  	intel_modeset_setup_hw_state(dev, false);
>  }
>  
> -void intel_modeset_cleanup(struct drm_device *dev)
> +void intel_modeset_quiesce(struct drm_device *dev)
>  {
>  	struct drm_i915_private *dev_priv = dev->dev_private;
> +
> +	cancel_work_sync(&dev_priv->hotplug_work);
> +	cancel_work_sync(&dev_priv->rps.work);
> +
> +	/* catch all required for dev_priv->wq */
> +	flush_scheduled_work();

This only flushes the system workqueue, so the changed comment about
dev_priv->wq is misleading. Can you please change it back?

Thanks, Daniel

> +}
> +
> +void intel_modeset_cleanup(struct drm_device *dev)
> +{
>  	struct drm_crtc *crtc;
>  	struct intel_crtc *intel_crtc;
>  
> @@ -10336,7 +10347,10 @@ void intel_modeset_cleanup(struct drm_device *dev)
>  	 * experience fancy races otherwise.
>  	 */
>  	drm_irq_uninstall(dev);
> -	cancel_work_sync(&dev_priv->hotplug_work);
> +
> +	/* flush any delayed tasks or pending work */
> +	intel_modeset_quiesce(dev);
> +
>  	/*
>  	 * Due to the hpd irq storm handling the hotplug work can re-arm the
>  	 * poll handlers. Hence disable polling after hpd handling is shut down.
> @@ -10364,9 +10378,6 @@ void intel_modeset_cleanup(struct drm_device *dev)
>  
>  	mutex_unlock(&dev->struct_mutex);
>  
> -	/* flush any delayed tasks or pending work */
> -	flush_scheduled_work();
> -
>  	/* destroy backlight, if any, before the connectors */
>  	intel_panel_destroy_backlight(dev);
>  
> -- 
> 1.8.4.rc1
> 
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx

-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch