Re: [PATCH 4/5] drm/msm/dp: report IRQ_HPD as an IRQ-only notification

[Yongxing Mou]

On 7/12/2026 6:55 PM, Dmitry Baryshkov wrote:
> On Mon, Jun 29, 2026 at 10:48:06PM +0800, Yongxing Mou wrote:
> > MST reuses the SST connector bridge to propagate HPD IRQ events through
> > the bridge chain.
> >
> > For IRQ_HPD notifications there is no connector state transition to
> > report. Use connector_status_unknown together with
> > DRM_CONNECTOR_DP_IRQ_HPD so that the bridge connector framework treats
> > them as IRQ-only notifications and forwards them without modifying
> > connector state.
> >
> > The DP driver handles IRQ_HPD events based on
> > DRM_CONNECTOR_DP_IRQ_HPD rather than connector status transitions.
> >
> > Signed-off-by: Yongxing Mou <yongxing.mou@xxxxxxxxxxxxxxxx>
> > ---
> >   drivers/gpu/drm/msm/dp/dp_display.c   | 22 +++++++++-------------
> >   drivers/soc/qcom/pmic_glink_altmode.c | 14 +++++++++-----
>
> And which tree (and why) would be able to merge this patch?
Emm , I will split the patch and reorganize it.
> >   2 files changed, 18 insertions(+), 18 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> > index bc93b566fbca..4ee391cc7165 100644
> > --- a/drivers/gpu/drm/msm/dp/dp_display.c
> > +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> > @@ -1119,14 +1119,10 @@ static irqreturn_t msm_dp_display_irq_thread(int irq, void *dev_id)
> >   		drm_bridge_hpd_notify(dp->msm_dp_display.bridge,
> >   				      connector_status_connected);
> >   
> > -	/* Send HPD as connected and distinguish it in the notifier */
> > -	if (hpd_isr_status & DP_DP_IRQ_HPD_INT_MASK) {
> > -		if (dp->msm_dp_display.mst_active)
> > -			msm_dp_irq_hpd_handle(dp);
> > -		else
> > -			drm_bridge_hpd_notify(dp->msm_dp_display.bridge,
> > -					      connector_status_connected);
> > -	}
> > +	if (hpd_isr_status & DP_DP_IRQ_HPD_INT_MASK)
> > +		drm_bridge_hpd_notify_extra(dp->msm_dp_display.bridge,
> > +					    connector_status_unknown,
>
> It's _not_ unknown.
Same as my reply to 2/5 -- the status is indeed not unknown here, and
using connector_status_unknown as a sentinel in v1 is not appropriate.
I will fix this in v2 along the direction discussed on 2/5.
> > +					    DRM_CONNECTOR_DP_IRQ_HPD);
>
> This should be fixed in the MST patchset. You should not be doing any
> actual handling in the ISR routing.
>
> >   
> >   	ret = IRQ_HANDLED;
> >   
> > @@ -1781,11 +1777,11 @@ void msm_dp_bridge_hpd_notify(struct drm_bridge *bridge,
> >   	drm_dbg_dp(dp->drm_dev, "type=%d link hpd_link_status=0x%x, status=%d\n",
> >   		   msm_dp_display->connector_type, hpd_link_status, status);
> >   
> > -	if (status == connector_status_connected) {
> > -		if (hpd_link_status == ISR_IRQ_HPD_PULSE_COUNT ||
> > -		    extra_status == DRM_CONNECTOR_DP_IRQ_HPD) {
> > -			msm_dp_irq_hpd_handle(dp);
> > -		} else if (hpd_link_status == ISR_HPD_REPLUG_COUNT) {
> > +	if (extra_status == DRM_CONNECTOR_DP_IRQ_HPD ||
> > +	    hpd_link_status == ISR_IRQ_HPD_PULSE_COUNT) {
> > +		msm_dp_irq_hpd_handle(dp);
>
> ANd here you missed the case when IRQ_HPD is being reported together
> with the first HPD event.
Yes, will fix it next patchset.
> > +	} else if (status == connector_status_connected) {
> > +		if (hpd_link_status == ISR_HPD_REPLUG_COUNT) {
> >   			msm_dp_hpd_unplug_handle(dp);
> >   			msm_dp_hpd_plug_handle(dp);
> >   		} else {
> > diff --git a/drivers/soc/qcom/pmic_glink_altmode.c b/drivers/soc/qcom/pmic_glink_altmode.c
> > index 946eb20b8f83..28ab8cbb5ef9 100644
> > --- a/drivers/soc/qcom/pmic_glink_altmode.c
> > +++ b/drivers/soc/qcom/pmic_glink_altmode.c
> > @@ -373,11 +373,15 @@ static void pmic_glink_altmode_worker(struct work_struct *work)
> >   		else
> >   			conn_status = connector_status_disconnected;
> >   
> > -		drm_aux_hpd_bridge_notify_extra(&alt_port->bridge->dev,
> > -						conn_status,
> > -						alt_port->hpd_irq ?
> > -						DRM_CONNECTOR_DP_IRQ_HPD :
> > -						DRM_CONNECTOR_NO_EXTRA_STATUS);
> > +		if (alt_port->hpd_irq) {
> > +			drm_aux_hpd_bridge_notify_extra(&alt_port->bridge->dev,
> > +							connector_status_unknown,
>
> And this is completely wrong. The AltMode driver (btw, you also missed
> the normal altmode driver for DP) doesn't know if it should handle the
> events in some way. It should report the HPD and IRQ_HPD as is.
Will fix it next patchset .
> > +							DRM_CONNECTOR_DP_IRQ_HPD);
> > +		} else {
> > +			drm_aux_hpd_bridge_notify_extra(&alt_port->bridge->dev,
> > +							conn_status,
> > +							DRM_CONNECTOR_NO_EXTRA_STATUS);
> > +		}
> >   	} else if (alt_port->mux_ctrl == MUX_CTRL_STATE_TUNNELING) {
> >   		if (alt_port->svid == USB_TYPEC_TBT_SID)
> >   			pmic_glink_altmode_enable_tbt(altmode, alt_port);
> >
> > --
> > 2.43.0