Hi Dimitry,
Thanks for the review.
On Wed, Dec 11, 2024 at 5:44 PM Dmitry Baryshkov
<dmitry.baryshkov@xxxxxxxxxx> wrote:
>
> On Mon, Dec 09, 2024 at 02:46:32PM +0800, Xin Ji wrote:
> > When user enabled HDCP feature, upper layer will set HDCP content
> > to DRM_MODE_CONTENT_PROTECTION_DESIRED. Next, anx7625 will update
> > HDCP content to DRM_MODE_CONTENT_PROTECTION_ENABLED if down stream
> > support HDCP feature.
> >
> > However once HDCP content turn to DRM_MODE_CONTENT_PROTECTION_ENABLED
> > upper layer will not update the HDCP content to
> > DRM_MODE_CONTENT_PROTECTION_UNDESIRED until monitor disconnect.
>
> What is "upper layer"? Is it a kernel or a userspace?
I think Xin meant userspace, but sounds like there are some
misunderstanding around the HDCP status.
>
> >From drm_hdcp_update_content_protection() documentation:
>
> No uevent for DESIRED->UNDESIRED or ENABLED->UNDESIRED,
> as userspace is triggering such state change and kernel performs it without
> fail.This function update the new state of the property into the connector's
> state and generate an uevent to notify the userspace.
>
>
> >
> > So when user dynamic change the display resolution, anx7625 driver must
> > call drm_hdcp_update_content_protection() to update HDCP content to
> > DRM_MODE_CONTENT_PROTECTION_UNDESIRED in bridge interface
> > .atomic_disable().
>
> Why?
>
> >
> > Signed-off-by: Xin Ji <xji@xxxxxxxxxxxxxxxx>
> > ---
> > drivers/gpu/drm/bridge/analogix/anx7625.c | 25 ++++++++++++++++++-----
> > 1 file changed, 20 insertions(+), 5 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/bridge/analogix/anx7625.c b/drivers/gpu/drm/bridge/analogix/anx7625.c
> > index a2675b121fe4..a75f519ddcb8 100644
> > --- a/drivers/gpu/drm/bridge/analogix/anx7625.c
> > +++ b/drivers/gpu/drm/bridge/analogix/anx7625.c
> > @@ -861,6 +861,22 @@ static int anx7625_hdcp_disable(struct anx7625_data *ctx)
> > TX_HDCP_CTRL0, ~HARD_AUTH_EN & 0xFF);
> > }
> >
> > +static void anx7625_hdcp_disable_and_update_cp(struct anx7625_data *ctx)
> > +{
> > + struct device *dev = ctx->dev;
> > +
> > + if (!ctx->connector)
> > + return;
> > +
> > + anx7625_hdcp_disable(ctx);
> > +
> > + ctx->hdcp_cp = DRM_MODE_CONTENT_PROTECTION_UNDESIRED;
> > + drm_hdcp_update_content_protection(ctx->connector,
> > + ctx->hdcp_cp);
> > +
> > + dev_dbg(dev, "update CP to UNDESIRE\n");
> > +}
> > +
> > static int anx7625_hdcp_enable(struct anx7625_data *ctx)
> > {
> > u8 bcap;
> > @@ -2165,11 +2181,8 @@ static int anx7625_connector_atomic_check(struct anx7625_data *ctx,
> > dev_err(dev, "current CP is not ENABLED\n");
> > return -EINVAL;
> > }
> > - anx7625_hdcp_disable(ctx);
> > - ctx->hdcp_cp = DRM_MODE_CONTENT_PROTECTION_UNDESIRED;
> > - drm_hdcp_update_content_protection(ctx->connector,
> > - ctx->hdcp_cp);
> > - dev_dbg(dev, "update CP to UNDESIRE\n");
> > +
> > + anx7625_hdcp_disable_and_update_cp(ctx);
>
> No. atomic_check() MAY NOT perform any changes to the hardware. It might
> be just a probe from userspace to check if the mode or a particular
> option can be set in a particular way. There is no guarantee that
> userspace will even try to commit it.
So, we should move the hdcp status update from .atomic_check() to
.atomic_enable() and .atomic_disable(), right? That is, enable HDCP
for the chip at .atomic_enable() if it is DESIRED and disable it at
.atomic_disable() if we enabled it previously.
Maybe we can keep some of the checks in .atomic_check(), but I doubt
if those logics actually make sense.
>
> > }
> >
> > if (cp == DRM_MODE_CONTENT_PROTECTION_ENABLED) {
> > @@ -2449,6 +2462,8 @@ static void anx7625_bridge_atomic_disable(struct drm_bridge *bridge,
> >
> > dev_dbg(dev, "drm atomic disable\n");
> >
> > + anx7625_hdcp_disable_and_update_cp(ctx);
> > +
> > ctx->connector = NULL;
> > anx7625_dp_stop(ctx);
> >
> > --
> > 2.25.1
> >
>
> --
> With best wishes
> Dmitry
Regards,
Pin-yen
