On Tue, 2014-02-04 at 17:13 +0100, Daniel Vetter wrote: > On Fri, Jan 24, 2014 at 10:15 AM, Imre Deak <imre.deak@xxxxxxxxx> wrote: > > Atm we try to remove the connector's i2c sysfs entry too late in the > > encoder's destroy callback. By that time the kobject used as the parent > > for all connector sysfs entries is already removed when we do an early > > removal of all connector sysfs entries in intel_modeset_cleanup(). Fix > > this by adding an early_destory encoder callback, where we remove the > > encoder's i2c adapter. > > > > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=70523 > > > > Signed-off-by: Imre Deak <imre.deak@xxxxxxxxx> > > Ok, I guess with Greg's clarification this seems to be the correct > fix. But I'm not too happy about the ->early_destroy since that > inversion of control is usually a bad sign for wrong layering. Imo > it'd be better to push all the encoder clean down into each encoders > ->destroy callback and then move the dp aux cleanup at the right > place. So essentially we'd need to push the > intel_panel_destroy_backlight and drm_sysfs_connector_remove calls > down. That also allows us to only cleanup the backlight on edp/lvds. > > Comments or too insane an idea? Agreed about ->early_destroy being hacky, it was the fast and usual solution for reordering something in drm :) I'll check this later / discuss with Jani about reworking this based on your suggestion. --Imre
Attachment:
signature.asc
Description: This is a digitally signed message part
_______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx
