Quoting Chris Wilson (2017-11-28 13:02:09) > Quoting Chris Wilson (2017-11-28 11:01:47) > > @@ -1266,6 +1271,13 @@ static void i915_driver_unregister(struct drm_i915_private *dev_priv) > > intel_fbdev_unregister(dev_priv); > > intel_audio_deinit(dev_priv); > > > > + /* > > + * After flushing the fbdev (incl. a late async config which will > > + * have delayed queuing of a hotplug event), then flush the hotplug > > + * events. > > + */ > > + drm_kms_helper_poll_fini(&dev_priv->drm); > > As discussed on irc, I'm unsure as to whether it should be fbdev > followed by kms_helper, or the other way around. In general, I think we > lack some serialisation to control the async hotplug event generation > here... But I think this is sufficiently to make CI happy for the > time being, we will see. Fwiw, this did clear all the WARN_ON(!intel_irq_enabled()) for one CI run. It may not be perfect, it may be "good enough". -Chris _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
