[PATCH 2/4] drm/panel/samsung-dsicm: Keep up with refcounting

[Anusha Srivatsa]

On Fri, Jun 6, 2025 at 7:03 AM Maxime Ripard <mripard at kernel.org> wrote:

> On Wed, Jun 04, 2025 at 10:45:11PM -0500, Anusha Srivatsa wrote:
> > Put the panel reference back when driver is no
> > longer using it.
> >
> > Signed-off-by: Anusha Srivatsa <asrivats at redhat.com>
> > ---
> >  drivers/gpu/drm/bridge/samsung-dsim.c | 1 +
> >  1 file changed, 1 insertion(+)
> >
> > diff --git a/drivers/gpu/drm/bridge/samsung-dsim.c
> b/drivers/gpu/drm/bridge/samsung-dsim.c
> > index
> 0014c497e3fe7d8349a119dbdda30d65d816cccf..3667855ff0d6d1b608c579573de657af7fd14388
> 100644
> > --- a/drivers/gpu/drm/bridge/samsung-dsim.c
> > +++ b/drivers/gpu/drm/bridge/samsung-dsim.c
> > @@ -1748,6 +1748,7 @@ static int samsung_dsim_host_attach(struct
> mipi_dsi_host *host,
> >                       dsi->out_bridge = ERR_PTR(-EINVAL);
> >       }
> >
> > +     drm_panel_put(panel);
> >       of_node_put(remote);
> >
> >       if (IS_ERR(dsi->out_bridge)) {
>
> Explaining in your commit log why you think it's a good idea to put it
> there would be really nice. In particular, it looks super odd to me that
> you would put the panel reference in attach.
>
Also, your patch doesn't work, and you have a reference inbalance. You
> have one taken by the panel driver, put in remove. You have one in
> drm_panel_add, put in drm_panel_remove. Which reference do you put here?
>
> This applies to your other patches too.
>
> Yes, I should get the ref in of_drm_find_panel() to put it here. With that
said, all callers
of of_drm_find_panel() should be converted instead of the small batch that
this series addresses.

Anusha

> Maxime
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/dri-devel/attachments/20250609/cc097ed4/attachment-0001.htm>