| Message ID | 20250206-hotplug-drm-bridge-v6-10-9d6f2c9c3058@bootlin.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | Add support for hot-pluggable DRM bridges | expand |
On Thu, Feb 06, 2025 at 07:14:25PM +0100, Luca Ceresoli wrote: > devm_drm_of_get_bridge(), which is based on graph links, is the recommended > function to get a pointer to the following bridge. > > This is valid even for panels, for which the recommended device tree > description is via graph links and not (or not only) panel subnodes of a > panel controller (e.g. "dsi@1234" controller node with a "panel@0" > subnode). > > However there are drivers supporting the panel subnode description in I'd say, name them here (at least the one which we looked upon). > addition to the graph links. For those drivers add a _by_node variant that > takes the node of the target node. > > Suggested-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> > Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com> >
On Fri, 7 Feb 2025 04:53:32 +0200 Dmitry Baryshkov <dmitry.baryshkov@linaro.org> wrote: > On Thu, Feb 06, 2025 at 07:14:25PM +0100, Luca Ceresoli wrote: > > devm_drm_of_get_bridge(), which is based on graph links, is the recommended > > function to get a pointer to the following bridge. > > > > This is valid even for panels, for which the recommended device tree > > description is via graph links and not (or not only) panel subnodes of a > > panel controller (e.g. "dsi@1234" controller node with a "panel@0" > > subnode). > > > > However there are drivers supporting the panel subnode description in > > I'd say, name them here (at least the one which we looked upon). Sure, adding the samsung-dsim to the commit message. I'm not aware of other drivers doing the same, but I will do a quick search. Luca
On Thu, Feb 06, 2025 at 07:14:25PM +0100, Luca Ceresoli wrote: > devm_drm_of_get_bridge(), which is based on graph links, is the recommended > function to get a pointer to the following bridge. > > This is valid even for panels, for which the recommended device tree > description is via graph links and not (or not only) panel subnodes of a > panel controller (e.g. "dsi@1234" controller node with a "panel@0" > subnode). > > However there are drivers supporting the panel subnode description in > addition to the graph links. For those drivers add a _by_node variant that > takes the node of the target node. > > Suggested-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> > Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com> > > --- > > This patch was added in v6. > --- > drivers/gpu/drm/drm_bridge.c | 30 ++++++++++++++++++++++++++++++ > include/drm/drm_bridge.h | 8 ++++++++ > 2 files changed, 38 insertions(+) > > diff --git a/drivers/gpu/drm/drm_bridge.c b/drivers/gpu/drm/drm_bridge.c > index 2aa17fbe538b86066c4e68f0d0e8046e9ca9b965..b0834b8644284e5f7751cec81724af849b4180e7 100644 > --- a/drivers/gpu/drm/drm_bridge.c > +++ b/drivers/gpu/drm/drm_bridge.c > @@ -1407,6 +1407,36 @@ struct drm_bridge *devm_drm_of_get_bridge(struct device *dev, > } > EXPORT_SYMBOL(devm_drm_of_get_bridge); > > +/** > + * devm_drm_of_get_bridge_by_node - Return bridge for a given OF node > + * @dev: device to tie the bridge lifetime to > + * @bridge_node: device node of the remote bridge > + * > + * Given a bridge DT node, returns the associated bridge if any. This > + * should be used in addition to devm_drm_of_get_bridge() when the regular > + * graph link search is not enough, e.g. for drivers that need to support > + * panels described only as subnodes. > + * > + * RETURNS: > + * A pointer to the bridge if successful, or an error pointer otherwise. > + */ > +struct drm_bridge *devm_drm_of_get_bridge_by_node(struct device *dev, > + struct device_node *bridge_node) > +{ > + struct drm_bridge *bridge; > + int ret; > + > + if (!bridge_node) > + return ERR_PTR(-EINVAL); > + > + bridge = of_drm_find_bridge(bridge_node); > + if (!bridge) > + return ERR_PTR(-ENODEV); > + > + return bridge; > +} > +EXPORT_SYMBOL(devm_drm_of_get_bridge_by_node); > + A bridge is a KMS-facing structure, there's no reason to tie it to the lifetime of the device. Maxime
diff --git a/drivers/gpu/drm/drm_bridge.c b/drivers/gpu/drm/drm_bridge.c index 2aa17fbe538b86066c4e68f0d0e8046e9ca9b965..b0834b8644284e5f7751cec81724af849b4180e7 100644 --- a/drivers/gpu/drm/drm_bridge.c +++ b/drivers/gpu/drm/drm_bridge.c @@ -1407,6 +1407,36 @@ struct drm_bridge *devm_drm_of_get_bridge(struct device *dev, } EXPORT_SYMBOL(devm_drm_of_get_bridge); +/** + * devm_drm_of_get_bridge_by_node - Return bridge for a given OF node + * @dev: device to tie the bridge lifetime to + * @bridge_node: device node of the remote bridge + * + * Given a bridge DT node, returns the associated bridge if any. This + * should be used in addition to devm_drm_of_get_bridge() when the regular + * graph link search is not enough, e.g. for drivers that need to support + * panels described only as subnodes. + * + * RETURNS: + * A pointer to the bridge if successful, or an error pointer otherwise. + */ +struct drm_bridge *devm_drm_of_get_bridge_by_node(struct device *dev, + struct device_node *bridge_node) +{ + struct drm_bridge *bridge; + int ret; + + if (!bridge_node) + return ERR_PTR(-EINVAL); + + bridge = of_drm_find_bridge(bridge_node); + if (!bridge) + return ERR_PTR(-ENODEV); + + return bridge; +} +EXPORT_SYMBOL(devm_drm_of_get_bridge_by_node); + /** * drmm_of_get_bridge - Return next bridge in the chain * @drm: device to tie the bridge lifetime to diff --git a/include/drm/drm_bridge.h b/include/drm/drm_bridge.h index 496dbbd2ad7edff7f091adfbe62de1e33ef0cf07..1561347c4991dac6022319774510f9560c9283c3 100644 --- a/include/drm/drm_bridge.h +++ b/include/drm/drm_bridge.h @@ -1088,6 +1088,8 @@ static inline int drm_panel_bridge_set_orientation(struct drm_connector *connect #if defined(CONFIG_OF) && defined(CONFIG_DRM_PANEL_BRIDGE) struct drm_bridge *devm_drm_of_get_bridge(struct device *dev, struct device_node *node, u32 port, u32 endpoint); +struct drm_bridge *devm_drm_of_get_bridge_by_node(struct device *dev, + struct device_node *bridge_node); struct drm_bridge *drmm_of_get_bridge(struct drm_device *drm, struct device_node *node, u32 port, u32 endpoint); #else @@ -1099,6 +1101,12 @@ static inline struct drm_bridge *devm_drm_of_get_bridge(struct device *dev, return ERR_PTR(-ENODEV); } +static inline struct drm_bridge *devm_drm_of_get_bridge_by_node(struct device *dev, + struct device_node *bridge_node) +{ + return ERR_PTR(-ENODEV); +} + static inline struct drm_bridge *drmm_of_get_bridge(struct drm_device *drm, struct device_node *node, u32 port,
devm_drm_of_get_bridge(), which is based on graph links, is the recommended function to get a pointer to the following bridge. This is valid even for panels, for which the recommended device tree description is via graph links and not (or not only) panel subnodes of a panel controller (e.g. "dsi@1234" controller node with a "panel@0" subnode). However there are drivers supporting the panel subnode description in addition to the graph links. For those drivers add a _by_node variant that takes the node of the target node. Suggested-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com> --- This patch was added in v6. --- drivers/gpu/drm/drm_bridge.c | 30 ++++++++++++++++++++++++++++++ include/drm/drm_bridge.h | 8 ++++++++ 2 files changed, 38 insertions(+)