Hi Maxime, On Wed, 9 Jul 2025 09:32:28 +0200 Maxime Ripard <[email protected]> wrote:
... > > @@ -677,7 +683,6 @@ static int sti_hda_bind(struct device *dev, struct > > device *master, void *data) > > struct drm_encoder *encoder; > > struct sti_hda_connector *connector; > > struct drm_connector *drm_connector; > > - struct drm_bridge *bridge; > > int err; > > > > /* Set the drm device handle */ > > @@ -693,13 +698,7 @@ static int sti_hda_bind(struct device *dev, struct > > device *master, void *data) > > > > connector->hda = hda; > > > > - bridge = devm_kzalloc(dev, sizeof(*bridge), GFP_KERNEL); > > - if (!bridge) > > - return -ENOMEM; > > - > > - bridge->driver_private = hda; > > - bridge->funcs = &sti_hda_bridge_funcs; > > - drm_bridge_attach(encoder, bridge, NULL, 0); > > + drm_bridge_attach(encoder, &hda->bridge, NULL, 0); > > It's not entirely related, but the connector is also allocated right > before and could be moved into the structure instead of storing a > pointer. > > Either way, > Reviewed-by: Maxime Ripard <[email protected]> Given this patch as-is is fixing a (potentially invisible) bug due to using the old bridge allocation policy, and the connector allocation change is an improvements but not a fix, I'll apply as is. BTW, as a side effect, this series is removing two users of bridge->driver_private :) Luca -- Luca Ceresoli, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
