Hello, On Wed Aug 12, 2026 at 3:26 PM CEST, Maxime Ripard wrote: > On Wed, Aug 12, 2026 at 03:04:18PM +0200, Osama Abdelkader wrote: >> Use the managed MIPI DSI attach helper in BOE panel drivers and drop the >> corresponding manual detach calls from remove paths. >> >> Signed-off-by: Osama Abdelkader <[email protected]> >> >> --- >> v3: >> - Move the changelog after the `---`. >> >> v2: >> - break long description lines. >> - add Signed-off-by tag >> >> drivers/gpu/drm/panel/panel-boe-bf060y8m-aj0.c | 8 +------- >> drivers/gpu/drm/panel/panel-boe-himax8279d.c | 8 +------- >> drivers/gpu/drm/panel/panel-boe-td4320.c | 8 +------- >> drivers/gpu/drm/panel/panel-boe-th101mb31ig002-28a.c | 4 +--- >> drivers/gpu/drm/panel/panel-boe-tv101wum-ll2.c | 8 +------- >> drivers/gpu/drm/panel/panel-boe-tv101wum-nl6.c | 7 +------ >> 6 files changed, 6 insertions(+), 37 deletions(-) >> >> diff --git a/drivers/gpu/drm/panel/panel-boe-bf060y8m-aj0.c >> b/drivers/gpu/drm/panel/panel-boe-bf060y8m-aj0.c >> index 84c21c62a43e..6a74523a6614 100644 >> --- a/drivers/gpu/drm/panel/panel-boe-bf060y8m-aj0.c >> +++ b/drivers/gpu/drm/panel/panel-boe-bf060y8m-aj0.c >> @@ -359,7 +359,7 @@ static int boe_bf060y8m_aj0_probe(struct mipi_dsi_device >> *dsi) >> >> drm_panel_add(&boe->panel); >> >> - ret = mipi_dsi_attach(dsi); >> + ret = devm_mipi_dsi_attach(&dsi->dev, dsi); >> if (ret < 0) { >> dev_err(dev, "Failed to attach to DSI host: %d\n", ret); >> return ret; >> @@ -371,12 +371,6 @@ static int boe_bf060y8m_aj0_probe(struct >> mipi_dsi_device *dsi) >> static void boe_bf060y8m_aj0_remove(struct mipi_dsi_device *dsi) >> { >> struct boe_bf060y8m_aj0 *boe = mipi_dsi_get_drvdata(dsi); >> - int ret; >> - >> - ret = mipi_dsi_detach(dsi); >> - if (ret < 0) >> - dev_err(&dsi->dev, "Failed to detach from DSI host: %d\n", ret); >> - >> drm_panel_remove(&boe->panel); >> } > > mipi_dsi_detach is now called after drm_panel_remove. Your commit log > should mention it, and explain why it's not a concern.
Indeed. Or switch from drm_panel_add() to devm_drm_panel_add(), which would avoid the issue mentioned by Maxime and cleanup even more the code. Luca -- Luca Ceresoli, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
