diff mbox series

[2/2] drm/i915: Extend hotplug detect retry on TypeC connectors to 5 seconds

Message ID 20200330095425.29113-2-imre.deak@intel.com (mailing list archive)
State New, archived
Headers show
Series [1/2] drm/i915: Add a retry counter for hotplug detect retries | expand

Commit Message

Imre Deak March 30, 2020, 9:54 a.m. UTC 
On TypeC ports if a sink deasserts/reasserts its HPD signal, generating
a hotplug interrupt without the sink getting unplugged/replugged from
the connector, there can be an up to 3 seconds delay until the AUX
channel gets functional. To avoid detection failures this delay causes
retry the detection for 5 seconds.

I noticed this on ICL/TGL RVPs and a DELL XPS 13 7390 ICL laptop.

References: https://gitlab.freedesktop.org/drm/intel/issues/1067
Signed-off-by: Imre Deak <imre.deak@intel.com>
---
 drivers/gpu/drm/i915/display/intel_ddi.c | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)

Comments

Souza, Jose March 30, 2020, 7:13 p.m. UTC | #1 
On Mon, 2020-03-30 at 12:54 +0300, Imre Deak wrote:
> On TypeC ports if a sink deasserts/reasserts its HPD signal,
> generating
> a hotplug interrupt without the sink getting unplugged/replugged from
> the connector, there can be an up to 3 seconds delay until the AUX
> channel gets functional. To avoid detection failures this delay
> causes
> retry the detection for 5 seconds.

5 seconds? would it be 5 tries?

Other than that both patches looks good.


Reviewed-by: José Roberto de Souza <jose.souza@intel.com>

> 
> I noticed this on ICL/TGL RVPs and a DELL XPS 13 7390 ICL laptop.
> 
> References: https://gitlab.freedesktop.org/drm/intel/issues/1067
> Signed-off-by: Imre Deak <imre.deak@intel.com>
> ---
>  drivers/gpu/drm/i915/display/intel_ddi.c | 12 +++++++++++-
>  1 file changed, 11 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/i915/display/intel_ddi.c
> b/drivers/gpu/drm/i915/display/intel_ddi.c
> index 4f508bf70f3b..2d947ff83488 100644
> --- a/drivers/gpu/drm/i915/display/intel_ddi.c
> +++ b/drivers/gpu/drm/i915/display/intel_ddi.c
> @@ -4371,7 +4371,10 @@ static enum intel_hotplug_state
>  intel_ddi_hotplug(struct intel_encoder *encoder,
>  		  struct intel_connector *connector)
>  {
> +	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
>  	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
> +	enum phy phy = intel_port_to_phy(i915, encoder->port);
> +	bool is_tc = intel_phy_is_tc(i915, phy);
>  	struct drm_modeset_acquire_ctx ctx;
>  	enum intel_hotplug_state state;
>  	int ret;
> @@ -4414,8 +4417,15 @@ intel_ddi_hotplug(struct intel_encoder
> *encoder,
>  	 * valid EDID. To solve this schedule another detection cycle
> if this
>  	 * time around we didn't detect any change in the sink's
> connection
>  	 * status.
> +	 *
> +	 * Type-c connectors which get their HPD signal deasserted then
> +	 * reasserted, without unplugging/replugging the sink from the
> +	 * connector, introduce a delay until the AUX channel
> communication
> +	 * becomes functional. Retry the detection for 5 seconds on
> type-c
> +	 * connectors to account for this delay.
>  	 */
> -	if (state == INTEL_HOTPLUG_UNCHANGED && !connector-
> >hotplug_retries &&
> +	if (state == INTEL_HOTPLUG_UNCHANGED &&
> +	    connector->hotplug_retries < (is_tc ? 5 : 1) &&
>  	    !dig_port->dp.is_mst)
>  		state = INTEL_HOTPLUG_RETRY;
>
Gupta, Anshuman April 1, 2020, 11:20 a.m. UTC | #2 
On 2020-03-30 at 15:24:25 +0530, Imre Deak wrote:
> On TypeC ports if a sink deasserts/reasserts its HPD signal, generating
> a hotplug interrupt without the sink getting unplugged/replugged from
> the connector, there can be an up to 3 seconds delay until the AUX
> channel gets functional. To avoid detection failures this delay causes
> retry the detection for 5 seconds.
> 
> I noticed this on ICL/TGL RVPs and a DELL XPS 13 7390 ICL laptop.
> 
> References: https://gitlab.freedesktop.org/drm/intel/issues/1067
> Signed-off-by: Imre Deak <imre.deak@intel.com>
> ---
>  drivers/gpu/drm/i915/display/intel_ddi.c | 12 +++++++++++-
>  1 file changed, 11 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/i915/display/intel_ddi.c b/drivers/gpu/drm/i915/display/intel_ddi.c
> index 4f508bf70f3b..2d947ff83488 100644
> --- a/drivers/gpu/drm/i915/display/intel_ddi.c
> +++ b/drivers/gpu/drm/i915/display/intel_ddi.c
> @@ -4371,7 +4371,10 @@ static enum intel_hotplug_state
>  intel_ddi_hotplug(struct intel_encoder *encoder,
>  		  struct intel_connector *connector)
>  {
> +	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
>  	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
> +	enum phy phy = intel_port_to_phy(i915, encoder->port);
> +	bool is_tc = intel_phy_is_tc(i915, phy);
>  	struct drm_modeset_acquire_ctx ctx;
>  	enum intel_hotplug_state state;
>  	int ret;
> @@ -4414,8 +4417,15 @@ intel_ddi_hotplug(struct intel_encoder *encoder,
>  	 * valid EDID. To solve this schedule another detection cycle if this
>  	 * time around we didn't detect any change in the sink's connection
>  	 * status.
> +	 *
> +	 * Type-c connectors which get their HPD signal deasserted then
> +	 * reasserted, without unplugging/replugging the sink from the
> +	 * connector, introduce a delay until the AUX channel communication
> +	 * becomes functional. Retry the detection for 5 seconds on type-c
> +	 * connectors to account for this delay.
>  	 */
> -	if (state == INTEL_HOTPLUG_UNCHANGED && !connector->hotplug_retries &&
> +	if (state == INTEL_HOTPLUG_UNCHANGED &&
> +	    connector->hotplug_retries < (is_tc ? 5 : 1) &&
I had observed that intel_dp_detect may race between user spece invoked
get connector call and intel_encoder_hotplug(), that may leave connector status
to be UNCHANGED in actual hotplug flow as intel_dp_detect() already called from 
drm_helper_probe_single_connector_modes(), this may results in 5 retries for 
type-C ports for normal HPD assertion.
Please correct me if i am wrong.
Thanks,
Anshuman Gupta.
>  	    !dig_port->dp.is_mst)
>  		state = INTEL_HOTPLUG_RETRY;
>  
> -- 
> 2.23.1
> 
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Imre Deak April 1, 2020, 11:54 a.m. UTC | #3 
On Wed, Apr 01, 2020 at 04:50:23PM +0530, Anshuman Gupta wrote:
> On 2020-03-30 at 15:24:25 +0530, Imre Deak wrote:
> > On TypeC ports if a sink deasserts/reasserts its HPD signal, generating
> > a hotplug interrupt without the sink getting unplugged/replugged from
> > the connector, there can be an up to 3 seconds delay until the AUX
> > channel gets functional. To avoid detection failures this delay causes
> > retry the detection for 5 seconds.
> > 
> > I noticed this on ICL/TGL RVPs and a DELL XPS 13 7390 ICL laptop.
> > 
> > References: https://gitlab.freedesktop.org/drm/intel/issues/1067
> > Signed-off-by: Imre Deak <imre.deak@intel.com>
> > ---
> >  drivers/gpu/drm/i915/display/intel_ddi.c | 12 +++++++++++-
> >  1 file changed, 11 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/display/intel_ddi.c b/drivers/gpu/drm/i915/display/intel_ddi.c
> > index 4f508bf70f3b..2d947ff83488 100644
> > --- a/drivers/gpu/drm/i915/display/intel_ddi.c
> > +++ b/drivers/gpu/drm/i915/display/intel_ddi.c
> > @@ -4371,7 +4371,10 @@ static enum intel_hotplug_state
> >  intel_ddi_hotplug(struct intel_encoder *encoder,
> >  		  struct intel_connector *connector)
> >  {
> > +	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
> >  	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
> > +	enum phy phy = intel_port_to_phy(i915, encoder->port);
> > +	bool is_tc = intel_phy_is_tc(i915, phy);
> >  	struct drm_modeset_acquire_ctx ctx;
> >  	enum intel_hotplug_state state;
> >  	int ret;
> > @@ -4414,8 +4417,15 @@ intel_ddi_hotplug(struct intel_encoder *encoder,
> >  	 * valid EDID. To solve this schedule another detection cycle if this
> >  	 * time around we didn't detect any change in the sink's connection
> >  	 * status.
> > +	 *
> > +	 * Type-c connectors which get their HPD signal deasserted then
> > +	 * reasserted, without unplugging/replugging the sink from the
> > +	 * connector, introduce a delay until the AUX channel communication
> > +	 * becomes functional. Retry the detection for 5 seconds on type-c
> > +	 * connectors to account for this delay.
> >  	 */
> > -	if (state == INTEL_HOTPLUG_UNCHANGED && !connector->hotplug_retries &&
> > +	if (state == INTEL_HOTPLUG_UNCHANGED &&
> > +	    connector->hotplug_retries < (is_tc ? 5 : 1) &&
>
> I had observed that intel_dp_detect may race between user spece
> invoked get connector call and intel_encoder_hotplug(), that may leave
> connector status to be UNCHANGED in actual hotplug flow as
> intel_dp_detect() already called from
> drm_helper_probe_single_connector_modes(), this may results in 5
> retries for type-C ports for normal HPD assertion.  Please correct me
> if i am wrong.

Yes, it's possible the retries will be unneccessary, but I don't think
we can do much about avoiding a race between a hotplug event and a
detect call.

> Thanks,
> Anshuman Gupta.
> >  	    !dig_port->dp.is_mst)
> >  		state = INTEL_HOTPLUG_RETRY;
> >  
> > -- 
> > 2.23.1
> > 
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx@lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Imre Deak April 6, 2020, 3:02 p.m. UTC | #4 
On Mon, Mar 30, 2020 at 10:13:02PM +0300, Souza, Jose wrote:
> On Mon, 2020-03-30 at 12:54 +0300, Imre Deak wrote:
> > On TypeC ports if a sink deasserts/reasserts its HPD signal,
> > generating
> > a hotplug interrupt without the sink getting unplugged/replugged from
> > the connector, there can be an up to 3 seconds delay until the AUX
> > channel gets functional. To avoid detection failures this delay
> > causes
> > retry the detection for 5 seconds.
> 
> 5 seconds? would it be 5 tries?

Yes, 5 tries with a 1 second delay in-between them.

Thanks for the review, patchset pushed to -dinq.

> Other than that both patches looks good.
> 
> Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
> 
> > 
> > I noticed this on ICL/TGL RVPs and a DELL XPS 13 7390 ICL laptop.
> > 
> > References: https://gitlab.freedesktop.org/drm/intel/issues/1067
> > Signed-off-by: Imre Deak <imre.deak@intel.com>
> > ---
> >  drivers/gpu/drm/i915/display/intel_ddi.c | 12 +++++++++++-
> >  1 file changed, 11 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/display/intel_ddi.c
> > b/drivers/gpu/drm/i915/display/intel_ddi.c
> > index 4f508bf70f3b..2d947ff83488 100644
> > --- a/drivers/gpu/drm/i915/display/intel_ddi.c
> > +++ b/drivers/gpu/drm/i915/display/intel_ddi.c
> > @@ -4371,7 +4371,10 @@ static enum intel_hotplug_state
> >  intel_ddi_hotplug(struct intel_encoder *encoder,
> >  		  struct intel_connector *connector)
> >  {
> > +	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
> >  	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
> > +	enum phy phy = intel_port_to_phy(i915, encoder->port);
> > +	bool is_tc = intel_phy_is_tc(i915, phy);
> >  	struct drm_modeset_acquire_ctx ctx;
> >  	enum intel_hotplug_state state;
> >  	int ret;
> > @@ -4414,8 +4417,15 @@ intel_ddi_hotplug(struct intel_encoder
> > *encoder,
> >  	 * valid EDID. To solve this schedule another detection cycle
> > if this
> >  	 * time around we didn't detect any change in the sink's
> > connection
> >  	 * status.
> > +	 *
> > +	 * Type-c connectors which get their HPD signal deasserted then
> > +	 * reasserted, without unplugging/replugging the sink from the
> > +	 * connector, introduce a delay until the AUX channel
> > communication
> > +	 * becomes functional. Retry the detection for 5 seconds on
> > type-c
> > +	 * connectors to account for this delay.
> >  	 */
> > -	if (state == INTEL_HOTPLUG_UNCHANGED && !connector-
> > >hotplug_retries &&
> > +	if (state == INTEL_HOTPLUG_UNCHANGED &&
> > +	    connector->hotplug_retries < (is_tc ? 5 : 1) &&
> >  	    !dig_port->dp.is_mst)
> >  		state = INTEL_HOTPLUG_RETRY;
> >
diff mbox series

Patch

diff --git a/drivers/gpu/drm/i915/display/intel_ddi.c b/drivers/gpu/drm/i915/display/intel_ddi.c
index 4f508bf70f3b..2d947ff83488 100644
--- a/drivers/gpu/drm/i915/display/intel_ddi.c
+++ b/drivers/gpu/drm/i915/display/intel_ddi.c
@@ -4371,7 +4371,10 @@  static enum intel_hotplug_state
 intel_ddi_hotplug(struct intel_encoder *encoder,
 		  struct intel_connector *connector)
 {
+	struct drm_i915_private *i915 = to_i915(encoder->base.dev);
 	struct intel_digital_port *dig_port = enc_to_dig_port(encoder);
+	enum phy phy = intel_port_to_phy(i915, encoder->port);
+	bool is_tc = intel_phy_is_tc(i915, phy);
 	struct drm_modeset_acquire_ctx ctx;
 	enum intel_hotplug_state state;
 	int ret;
@@ -4414,8 +4417,15 @@  intel_ddi_hotplug(struct intel_encoder *encoder,
 	 * valid EDID. To solve this schedule another detection cycle if this
 	 * time around we didn't detect any change in the sink's connection
 	 * status.
+	 *
+	 * Type-c connectors which get their HPD signal deasserted then
+	 * reasserted, without unplugging/replugging the sink from the
+	 * connector, introduce a delay until the AUX channel communication
+	 * becomes functional. Retry the detection for 5 seconds on type-c
+	 * connectors to account for this delay.
 	 */
-	if (state == INTEL_HOTPLUG_UNCHANGED && !connector->hotplug_retries &&
+	if (state == INTEL_HOTPLUG_UNCHANGED &&
+	    connector->hotplug_retries < (is_tc ? 5 : 1) &&
 	    !dig_port->dp.is_mst)
 		state = INTEL_HOTPLUG_RETRY;