[PATCH v13 02/18] drm: bridge: panel: Support nodrm case for drmm_panel_bridge_add

[Maxime Ripard]

On Mon, Feb 27, 2023 at 05:58:03PM +0530, Jagan Teki wrote:
> On Mon, Feb 27, 2023 at 5:41 PM Maxime Ripard <maxime at cerno.tech> wrote:
> >
> > On Mon, Feb 27, 2023 at 05:09:09PM +0530, Jagan Teki wrote:
> > > drmm_panel_bridge_add DRM-managed action helper is useful for the bridge
> > > which automatically removes the bridge when drm pointer is cleaned.
> > >
> > > Supporting the same on non-component bridges like host DSI bridge requires
> > > a drm pointer which is indeed available only when a panel-bridge is found.
> > >
> > > For these use cases, the caller would call the drmm_panel_bridge_add by
> > > passing NULL to drm pointer.
> > >
> > > So, assign the bridge->dev to drm pointer for those cases.
> > >
> > > Cc: Maxime Ripard <mripard at kernel.org>
> > > Cc: Laurent Pinchart <Laurent.pinchart at ideasonboard.com>
> > > Cc: Linus Walleij <linus.walleij at linaro.org>
> > > Cc: Maarten Lankhorst <maarten.lankhorst at linux.intel.com>
> > > Signed-off-by: Jagan Teki <jagan at amarulasolutions.com>
> > > ---
> > > Changes for v13:
> > > - new patch
> > >
> > > Note: use case on
> > > "[PATCH v13 04/18] drm: exynos: dsi: Switch to DSI panel or bridge find helper"
> > >
> > >  drivers/gpu/drm/bridge/panel.c | 7 +++++++
> > >  1 file changed, 7 insertions(+)
> > >
> > > diff --git a/drivers/gpu/drm/bridge/panel.c b/drivers/gpu/drm/bridge/panel.c
> > > index d4b112911a99..45a0c6671000 100644
> > > --- a/drivers/gpu/drm/bridge/panel.c
> > > +++ b/drivers/gpu/drm/bridge/panel.c
> > > @@ -402,6 +402,13 @@ struct drm_bridge *drmm_panel_bridge_add(struct drm_device *drm,
> > >       if (IS_ERR(bridge))
> > >               return bridge;
> > >
> > > +     /*
> > > +      * For non-component bridges, like host DSI bridge the DRM pointer
> > > +      * can be available only when a panel-bridge is found.
> > > +      */
> > > +     if (!drm)
> > > +             drm = bridge->dev;
> > > +
> >
> > Why can't the caller use bridge->dev?
> 
> The first step of drmm_panel_bridge_add is to find the panel-bridge,
> so we can only get bridge->dev after this step. The caller doesn't
> know anything about the panel bridge it just sends a panel pointer to
> find the panel-bridge

Ah, yes, indeed. This is still a hack we don't need.

> then assigns bridge->dev to drm for DRM action
>
> Please check this patch about the caller,
> https://patchwork.kernel.org/project/dri-devel/patch/20230227113925.875425-5-jagan@amarulasolutions.com/

Because we've already been over this. You can't call that function from
DSI's attach. So you should change that to what I already pointed you
to, and then you'll have the drm device pointer available.

> > Also, where did the devm_drm_of_dsi_get_bridge go? I thought you were
> > going to convert it to a drm-managed version?
> 
> Look like your suggestion to can't use devm_drm_of_dsi_get_bridge and
> call the DRM-action from the driver, that is the reason I have removed
> this and done the same as your previous inputs.
> https://lore.kernel.org/all/20230203110437.otzl2azscbujigv6@houat/

You can't use devm. You can and should definitely use drmm.

Maxime
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 228 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20230227/7c4ff45b/attachment.sig>