Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
Pre-existing issues: - [High] The DSI device registered with `mipi_dsi_device_register_full()` is never unregistered, causing a memory and device node leak that prevents driver reload. -- --- Patch [3]: [PATCH v2 03/12] drm/bridge: ti-dlpc3433: don't depend on DRM_PANEL --- commit 1c4393105a1871917fd54292e2edbfa8634d5ce0 Author: Luca Ceresoli <[email protected]> drm/bridge: ti-dlpc3433: don't depend on DRM_PANEL This driver does not use the drm_panel APIs at all, so no need to select DRM_PANEL. > diff --git a/drivers/gpu/drm/bridge/Kconfig b/drivers/gpu/drm/bridge/Kconfig > index 958b4f687d4dd..043596801a65b 100644 > --- a/drivers/gpu/drm/bridge/Kconfig > +++ b/drivers/gpu/drm/bridge/Kconfig > @@ -428,7 +428,7 @@ config DRM_TOSHIBA_TC358775 > > config DRM_TI_DLPC3433 > tristate "TI DLPC3433 Display controller" > - depends on DRM && DRM_PANEL > + depends on DRM [Severity: High] This is a pre-existing issue, but I noticed a device node leak in the driver itself. In dlpc_host_attach(), the DSI device is registered manually: drivers/gpu/drm/bridge/ti-dlpc3433.c:dlpc_host_attach() ... dlpc->dsi = mipi_dsi_device_register_full(host, &info); ... However, the remove callback fails to unregister it: drivers/gpu/drm/bridge/ti-dlpc3433.c:dlpc3433_remove() static void dlpc3433_remove(struct i2c_client *client) { struct dlpc *dlpc = i2c_get_clientdata(client); drm_bridge_remove(&dlpc->bridge); of_node_put(dlpc->host_node); } Should there be a call to mipi_dsi_device_unregister(dlpc->dsi) here to prevent leaking the device node unconditionally upon unbind? Alternatively, could this be migrated to devm_mipi_dsi_device_register_full() so it is managed by devres? > depends on OF > select DRM_MIPI_DSI > help -- Sashiko AI review ยท https://sashiko.dev/#/patchset/20260825-drm-bridge-drm_panel-cleanup-v2-0-c6c88ba7f...@bootlin.com?part=3
