On Tue, 29 Apr 2025 at 18:35, Maxime Ripard <mrip...@kernel.org> wrote:
>
> On Tue, Apr 15, 2025 at 12:03:23PM +0300, Dmitry Baryshkov wrote:
> > On 14/04/2025 17:47, Maxime Ripard wrote:
> > > Hi,
> > >
> > > On Mon, Apr 07, 2025 at 06:11:01PM +0300, Dmitry Baryshkov wrote:
> > > > In order to make sure that CEC adapters or notifiers are unregistered
> > > > and CEC-related data is properly destroyed make drm_connector_cleanup()
> > > > call CEC's unregister() callback.
> > > >
> > > > Signed-off-by: Dmitry Baryshkov <dmitry.barysh...@linaro.org>
> > > > Signed-off-by: Dmitry Baryshkov <dmitry.barysh...@oss.qualcomm.com>
> > > > ---
> > > > drivers/gpu/drm/drm_connector.c | 9 +++++++++
> > > > 1 file changed, 9 insertions(+)
> > > >
> > > > diff --git a/drivers/gpu/drm/drm_connector.c
> > > > b/drivers/gpu/drm/drm_connector.c
> > > > index
> > > > ba08fbd973829e49ea977251c4f0fb6d96ade631..ae9c02ef9ab102db03c2824683ece37cfbcd3300
> > > > 100644
> > > > --- a/drivers/gpu/drm/drm_connector.c
> > > > +++ b/drivers/gpu/drm/drm_connector.c
> > > > @@ -743,6 +743,13 @@ void drm_connector_cec_phys_addr_set(struct
> > > > drm_connector *connector)
> > > > }
> > > > EXPORT_SYMBOL(drm_connector_cec_phys_addr_set);
> > > > +static void drm_connector_cec_unregister(struct drm_connector
> > > > *connector)
> > > > +{
> > > > + if (connector->cec.funcs &&
> > > > + connector->cec.funcs->unregister)
> > > > + connector->cec.funcs->unregister(connector);
> > > > +}
> > > > +
> > > > /**
> > > > * drm_connector_cleanup - cleans up an initialised connector
> > > > * @connector: connector to cleanup
> > > > @@ -763,6 +770,8 @@ void drm_connector_cleanup(struct drm_connector
> > > > *connector)
> > > > platform_device_unregister(connector->hdmi_audio.codec_pdev);
> > > > + drm_connector_cec_unregister(connector);
> > > > +
> > >
> > > Actually, since we know that the HDMI connector is drm-managed, why
> > > can't we make the call to connector->cec.funcs->unregister a drm-managed
> > > action registered by drm_connector_hdmi_cec_register?
> >
> > I haven't settled yet in my mind whether we can/should also use this
> > infrastructure for drm_dp_cec management. So, at this point, I'd prefer to
> > keep a non-managed unregister function. Once we settle on something for
> > drm_dp_cec, we can switch to drmm.
>
> I'd rather do the opposite. Let's go for drmm for now, and if we need to
> change it for DP, we can always change it.
>
> "Nothing is so permanent as a temporary solution", so I'd rather have
> the natural and consistent one for now :)
Ack, I'll change that for the next version.
--
With best wishes
Dmitry
