Nouveau - Dec 2018 - [WIP PATCH 06/15] drm/i915: Keep malloc references to MST ports

On Thu, Dec 13, 2018 at 08:25:35PM -0500, Lyude Paul
wrote:
> So that the ports stay around until we've destroyed the connectors, in
> order to ensure that we don't pass an invalid pointer to any MST
helpers
> once we introduce the new MST VCPI helpers.
> 
> Signed-off-by: Lyude Paul <lyude at redhat.com>
> ---
>  drivers/gpu/drm/i915/intel_connector.c | 4 ++++
>  drivers/gpu/drm/i915/intel_dp_mst.c    | 2 ++
>  2 files changed, 6 insertions(+)
> 
> diff --git a/drivers/gpu/drm/i915/intel_connector.c
b/drivers/gpu/drm/i915/intel_connector.c
> index 18e370f607bc..37d2c644f4b8 100644
> --- a/drivers/gpu/drm/i915/intel_connector.c
> +++ b/drivers/gpu/drm/i915/intel_connector.c
> @@ -95,6 +95,10 @@ void intel_connector_destroy(struct drm_connector
*connector)
>  	intel_panel_fini(&intel_connector->panel);
>  
>  	drm_connector_cleanup(connector);
> +
> +	if (intel_connector->port)
We set this as the first thing in intel_dp_add_mst_connector, so this
check isn't doing anything.
> +		drm_dp_mst_put_port_malloc(intel_connector->port);
> +
>  	kfree(connector);
>  }
>  
> diff --git a/drivers/gpu/drm/i915/intel_dp_mst.c
b/drivers/gpu/drm/i915/intel_dp_mst.c
> index f05427b74e34..4d6ced34d465 100644
> --- a/drivers/gpu/drm/i915/intel_dp_mst.c
> +++ b/drivers/gpu/drm/i915/intel_dp_mst.c
> @@ -484,6 +484,8 @@ static struct drm_connector
*intel_dp_add_mst_connector(struct drm_dp_mst_topolo
>  	if (ret)
>  		goto err;
>  
> +	drm_dp_mst_get_port_malloc(port);
Needs to be moved up where we assing intel_connector->port, or it'll
underflow on cleanup on error paths.
> +
>  	return connector;
>  
>  err:
> -- 
> 2.19.2
> 
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch

On Fri, 2018-12-14 at 10:32 +0100, Daniel Vetter wrote:
> On Thu, Dec 13, 2018 at 08:25:35PM -0500, Lyude Paul wrote:
> > So that the ports stay around until we've destroyed the
connectors, in
> > order to ensure that we don't pass an invalid pointer to any MST
helpers
> > once we introduce the new MST VCPI helpers.
> > 
> > Signed-off-by: Lyude Paul <lyude at redhat.com>
> > ---
> >  drivers/gpu/drm/i915/intel_connector.c | 4 ++++
> >  drivers/gpu/drm/i915/intel_dp_mst.c    | 2 ++
> >  2 files changed, 6 insertions(+)
> > 
> > diff --git a/drivers/gpu/drm/i915/intel_connector.c
> > b/drivers/gpu/drm/i915/intel_connector.c
> > index 18e370f607bc..37d2c644f4b8 100644
> > --- a/drivers/gpu/drm/i915/intel_connector.c
> > +++ b/drivers/gpu/drm/i915/intel_connector.c
> > @@ -95,6 +95,10 @@ void intel_connector_destroy(struct drm_connector
> > *connector)
> >  	intel_panel_fini(&intel_connector->panel);
> >  
> >  	drm_connector_cleanup(connector);
> > +
> > +	if (intel_connector->port)
> 
> We set this as the first thing in intel_dp_add_mst_connector, so this
> check isn't doing anything.
I think this comment might be a mistake? intel_connector_destroy() is shared
by all of the intel connectors, so some may not have a value in
intel_connector->port. I can move it to it's own destroy callback for MST
if
you would prefer though.
> 
> > +		drm_dp_mst_put_port_malloc(intel_connector->port);
> > +
> >  	kfree(connector);
> >  }
> >  
> > diff --git a/drivers/gpu/drm/i915/intel_dp_mst.c
> > b/drivers/gpu/drm/i915/intel_dp_mst.c
> > index f05427b74e34..4d6ced34d465 100644
> > --- a/drivers/gpu/drm/i915/intel_dp_mst.c
> > +++ b/drivers/gpu/drm/i915/intel_dp_mst.c
> > @@ -484,6 +484,8 @@ static struct drm_connector
> > *intel_dp_add_mst_connector(struct drm_dp_mst_topolo
> >  	if (ret)
> >  		goto err;
> >  
> > +	drm_dp_mst_get_port_malloc(port);
> 
> Needs to be moved up where we assing intel_connector->port, or it'll
> underflow on cleanup on error paths.
> 
> > +
> >  	return connector;
> >  
> >  err:
> > -- 
> > 2.19.2
> > 
-- 
Cheers,
	Lyude Paul

On Tue, Dec 18, 2018 at 04:52:24PM -0500, Lyude Paul
wrote:
> On Fri, 2018-12-14 at 10:32 +0100, Daniel Vetter wrote:
> > On Thu, Dec 13, 2018 at 08:25:35PM -0500, Lyude Paul wrote:
> > > So that the ports stay around until we've destroyed the
connectors, in
> > > order to ensure that we don't pass an invalid pointer to any
MST helpers
> > > once we introduce the new MST VCPI helpers.
> > > 
> > > Signed-off-by: Lyude Paul <lyude at redhat.com>
> > > ---
> > >  drivers/gpu/drm/i915/intel_connector.c | 4 ++++
> > >  drivers/gpu/drm/i915/intel_dp_mst.c    | 2 ++
> > >  2 files changed, 6 insertions(+)
> > > 
> > > diff --git a/drivers/gpu/drm/i915/intel_connector.c
> > > b/drivers/gpu/drm/i915/intel_connector.c
> > > index 18e370f607bc..37d2c644f4b8 100644
> > > --- a/drivers/gpu/drm/i915/intel_connector.c
> > > +++ b/drivers/gpu/drm/i915/intel_connector.c
> > > @@ -95,6 +95,10 @@ void intel_connector_destroy(struct
drm_connector
> > > *connector)
> > >  	intel_panel_fini(&intel_connector->panel);
> > >  
> > >  	drm_connector_cleanup(connector);
> > > +
> > > +	if (intel_connector->port)
> > 
> > We set this as the first thing in intel_dp_add_mst_connector, so this
> > check isn't doing anything.
> 
> I think this comment might be a mistake? intel_connector_destroy() is
shared
> by all of the intel connectors, so some may not have a value in
> intel_connector->port. I can move it to it's own destroy callback
for MST if
> you would prefer though.
Oops, didn't look at the patch carefully enough. I think a
intel_dp_mst_connector_destroy would be useful in this case. But a bit a
bikeshed, so up to you.
-Daniel
> > 
> > > +		drm_dp_mst_put_port_malloc(intel_connector->port);
> > > +
> > >  	kfree(connector);
> > >  }
> > >  
> > > diff --git a/drivers/gpu/drm/i915/intel_dp_mst.c
> > > b/drivers/gpu/drm/i915/intel_dp_mst.c
> > > index f05427b74e34..4d6ced34d465 100644
> > > --- a/drivers/gpu/drm/i915/intel_dp_mst.c
> > > +++ b/drivers/gpu/drm/i915/intel_dp_mst.c
> > > @@ -484,6 +484,8 @@ static struct drm_connector
> > > *intel_dp_add_mst_connector(struct drm_dp_mst_topolo
> > >  	if (ret)
> > >  		goto err;
> > >  
> > > +	drm_dp_mst_get_port_malloc(port);
> > 
> > Needs to be moved up where we assing intel_connector->port, or
it'll
> > underflow on cleanup on error paths.
> > 
> > > +
> > >  	return connector;
> > >  
> > >  err:
> > > -- 
> > > 2.19.2
> > > 
> -- 
> Cheers,
> 	Lyude Paul
> 
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch