Re: [PATCH v4 5/5] drm/msm/dp: Implement hpd_notify()

From: Bjorn Andersson
Date: Mon May 02 2022 - 19:01:23 EST


On Mon 02 May 15:29 PDT 2022, Bjorn Andersson wrote:

> On Mon 02 May 13:59 PDT 2022, Kuogee Hsieh wrote:
>
> >
> > On 5/2/2022 9:53 AM, Bjorn Andersson wrote:
> > > The Qualcomm DisplayPort driver contains traces of the necessary
> > > plumbing to hook up USB HPD, in the form of the dp_hpd module and the
> > > dp_usbpd_cb struct. Use this as basis for implementing the
> > > hpd_notify() callback, by amending the dp_hpd module with the
> > > missing logic.
> > >
> > > Overall the solution is similar to what's done downstream, but upstream
> > > all the code to disect the HPD notification lives on the calling side of
> > > drm_connector_oob_hotplug_event().
> > >
> > > drm_connector_oob_hotplug_event() performs the lookup of the
> > > drm_connector based on fwnode, hence the need to assign the fwnode in
> > > dp_drm_connector_init().
> > >
> > > Signed-off-by: Bjorn Andersson <bjorn.andersson@xxxxxxxxxx>
> > > ---
> > >
> > > Changes since v3:
> > > - Implements hpd_notify instead of oob_hotplug_event
> > > - Rebased on new cleanup patch from Dmitry
> > > - Set hpd_state to ST_MAINLINK_READY when dp_display_usbpd_configure() succeeds
> > >
> > > drivers/gpu/drm/msm/dp/dp_display.c | 26 ++++++++++++++++++++++++++
> > > drivers/gpu/drm/msm/dp/dp_display.h | 1 +
> > > drivers/gpu/drm/msm/dp/dp_drm.c | 3 +++
> > > drivers/gpu/drm/msm/dp/dp_drm.h | 2 ++
> > > 4 files changed, 32 insertions(+)
> > >
> > > diff --git a/drivers/gpu/drm/msm/dp/dp_display.c b/drivers/gpu/drm/msm/dp/dp_display.c
> > > index b447446d75e9..080294ac6144 100644
> > > --- a/drivers/gpu/drm/msm/dp/dp_display.c
> > > +++ b/drivers/gpu/drm/msm/dp/dp_display.c
> > > @@ -83,6 +83,8 @@ struct dp_display_private {
> > > bool hpd_irq_on;
> > > bool audio_supported;
> > > + bool connected;
> > > +
> > > struct drm_device *drm_dev;
> > > struct platform_device *pdev;
> > > struct dentry *root;
> > > @@ -1271,6 +1273,7 @@ static int dp_display_probe(struct platform_device *pdev)
> > > if (!desc)
> > > return -EINVAL;
> > > + dp->dp_display.dev = &pdev->dev;
> > > dp->pdev = pdev;
> > > dp->name = "drm_dp";
> > > dp->dp_display.connector_type = desc->connector_type;
> > > @@ -1760,3 +1763,26 @@ void dp_bridge_mode_set(struct drm_bridge *drm_bridge,
> > > dp_display->dp_mode.h_active_low =
> > > !!(dp_display->dp_mode.drm_mode.flags & DRM_MODE_FLAG_NHSYNC);
> > > }
> > > +
> > > +void dp_bridge_hpd_notify(struct drm_bridge *bridge,
> > > + enum drm_connector_status status)
> > > +{
> > > + struct msm_dp_bridge *dp_bridge = to_dp_bridge(bridge);
> > > + struct msm_dp *dp = dp_bridge->dp_display;
> > > + struct dp_display_private *dp_display = container_of(dp, struct dp_display_private, dp_display);
> > > + int ret;
> > > +
> > > + drm_dbg_dp(dp_display->drm_dev, "status: %d connected: %d\n", status, dp_display->connected);
> > > +
> > > + if (!dp_display->connected && status == connector_status_connected) {
> > > + dp_display->connected = true;
> > > + ret = dp_display_usbpd_configure(dp_display);
> > > + if (!ret)
> > > + dp_display->hpd_state = ST_MAINLINK_READY;
> > > + } else if (status != connector_status_connected) {
> > > + dp_display->connected = false;
> > > + dp_display_notify_disconnect(dp_display);
> > > + } else {
> > > + dp_display_usbpd_attention(dp_display);
> > > + }
> > > +}
> >
> > I would assume dp_bridge_hpd_notify() will server same purpose as
> > dp_display_irq_handler() if hpd_notification is enabled.
> >
>
> I agree with this statement.
>
> > In that case, should dp_bridge_hpd_notify() add
> > EV_HPD_PLUG_INT/EV_IRQ_HPD_INT/EV_HPD_UNPLUG_INT
> >
>
> I tried this originally, but couldn't get it to work and expected that
> as the downstream driver doesn't do this, there was some good reason for
> me not to do it either.
>
> > into event q to kick off corresponding
> > dp_hpd_plug_handle()/dp_irq_hpd_handle()/dp_hpd_unplug_handle()?
> >
>
> But since then the driver has been cleaned up significantly, so I
> decided to give it a test again.
> Unfortunately it still doesn't work, but now it's easier to trace.
>
> Replacing the 3 cases with relevant calls to dp_add_event() results in
> us inserting a EV_HPD_UNPLUG_INT event really early, before things has
> been brought up. This will result in dp_hpd_unplug_handle() trying to
> disable the dp_catalog_hpd_config_intr(), which will crash as the
> hardware isn't yet clocked up.
>

As I sent that I realized and have now confirmed that if I get a HPD
connect before dp_display_host_init() things will not be in an
appropriate state to either and the board will crash...

I will dig a little bit more to see what we can do about this.

Regards,
Bjorn

> Further more, this points out the main difference between the normal HPD
> code and the USB HPD code; dp_catalog_hpd_config_intr() will enable the
> plug/unplug interrupts, which it shouldn't do for USB-controlled.
>
>
> So it seems we need two code paths after all.
>
> > By the way, I am going to test this patch out.
> >
> > Any patches I have to pull in before apply this serial patches?
> >
>
> The patches applies on Dmitry's msm-next-staging, which I've merged on
> top of linux-next together with a number of pending patches to get the
> DPU up on SM8350 and a pmic_glink driver which I'm about to post.
>
> But to validate that it doesn't affect your non-USB case, Dmitry's
> branch should be sufficient.
>
> Thanks,
> Bjorn
>
> > > diff --git a/drivers/gpu/drm/msm/dp/dp_display.h b/drivers/gpu/drm/msm/dp/dp_display.h
> > > index 4f9fe4d7610b..2d2614bc5a14 100644
> > > --- a/drivers/gpu/drm/msm/dp/dp_display.h
> > > +++ b/drivers/gpu/drm/msm/dp/dp_display.h
> > > @@ -11,6 +11,7 @@
> > > #include "disp/msm_disp_snapshot.h"
> > > struct msm_dp {
> > > + struct device *dev;
> > > struct drm_device *drm_dev;
> > > struct device *codec_dev;
> > > struct drm_bridge *bridge;
> > > diff --git a/drivers/gpu/drm/msm/dp/dp_drm.c b/drivers/gpu/drm/msm/dp/dp_drm.c
> > > index 62d58b9c4647..821cfd37b1fb 100644
> > > --- a/drivers/gpu/drm/msm/dp/dp_drm.c
> > > +++ b/drivers/gpu/drm/msm/dp/dp_drm.c
> > > @@ -68,6 +68,7 @@ static const struct drm_bridge_funcs dp_bridge_ops = {
> > > .mode_valid = dp_bridge_mode_valid,
> > > .get_modes = dp_bridge_get_modes,
> > > .detect = dp_bridge_detect,
> > > + .hpd_notify = dp_bridge_hpd_notify,
> > > };
> > > struct drm_bridge *dp_bridge_init(struct msm_dp *dp_display, struct drm_device *dev,
> > > @@ -138,6 +139,8 @@ struct drm_connector *dp_drm_connector_init(struct msm_dp *dp_display)
> > > if (IS_ERR(connector))
> > > return connector;
> > > + connector->fwnode = fwnode_handle_get(dev_fwnode(dp_display->dev));
> > > +
> > > drm_connector_attach_encoder(connector, dp_display->encoder);
> > > return connector;
> > > diff --git a/drivers/gpu/drm/msm/dp/dp_drm.h b/drivers/gpu/drm/msm/dp/dp_drm.h
> > > index f4b1ed1e24f7..3b7480a86844 100644
> > > --- a/drivers/gpu/drm/msm/dp/dp_drm.h
> > > +++ b/drivers/gpu/drm/msm/dp/dp_drm.h
> > > @@ -32,5 +32,7 @@ enum drm_mode_status dp_bridge_mode_valid(struct drm_bridge *bridge,
> > > void dp_bridge_mode_set(struct drm_bridge *drm_bridge,
> > > const struct drm_display_mode *mode,
> > > const struct drm_display_mode *adjusted_mode);
> > > +void dp_bridge_hpd_notify(struct drm_bridge *bridge,
> > > + enum drm_connector_status status);
> > > #endif /* _DP_DRM_H_ */