[PATCH v10 18/21] drm/mediatek: Add mt8195 Embedded DisplayPort driver

[CK Hu]

Hi, Rex:

On Wed, 2022-06-08 at 16:43 +0800, Rex-BC Chen wrote:
> On Wed, 2022-06-08 at 10:23 +0800, CK Hu wrote:
> > Hi, Rex:
> > 
> > On Tue, 2022-06-07 at 20:24 +0800, Rex-BC Chen wrote:
> > > On Tue, 2022-06-07 at 14:21 +0800, CK Hu wrote:
> > > > Hi, Rex:
> > > > 
> > > > On Mon, 2022-05-23 at 12:47 +0200, Guillaume Ranquet wrote:
> > > > > From: Markus Schneider-Pargmann <msp at baylibre.com>
> > > > > 
> > > > > This patch adds a DisplayPort driver for the Mediatek mt8195
> > > > > SoC.
> > > > > 
> > > > > It supports the mt8195, the embedded DisplayPort units. It
> > > > > offers
> > > > > DisplayPort 1.4 with up to 4 lanes.
> > > > > 
> > > > > The driver creates a child device for the phy. The child
> > > > > device
> > > > > will
> > > > > never exist without the parent being active. As they are
> > > > > sharing
> > > > > a
> > > > > register range, the parent passes a regmap pointer to the
> > > > > child
> > > > > so
> > > > > that
> > > > > both can work with the same register range. The phy driver
> > > > > sets
> > > > > device
> > > > > data that is read by the parent to get the phy device that
> > > > > can
> > > > > be
> > > > > used
> > > > > to control the phy properties.
> > > > > 
> > > > > This driver is based on an initial version by
> > > > > Jason-JH.Lin <jason-jh.lin at mediatek.com>.
> > > > > 
> > > > > Signed-off-by: Markus Schneider-Pargmann <msp at baylibre.com>
> > > > > Signed-off-by: Guillaume Ranquet <granquet at baylibre.com>
> > > > > ---
> > > > 
> > > > [snip]
> > > > 
> > > > > +
> > > > > +static irqreturn_t mtk_dp_hpd_event_thread(int hpd, void
> > > > > *dev)
> > > > > +{
> > > > > +	struct mtk_dp *mtk_dp = dev;
> > > > > +	int event;
> > > > > +	u8 buf[DP_RECEIVER_CAP_SIZE] = {};
> > > > > +
> > > > > +	event = mtk_dp_plug_state(mtk_dp) ?
> > > > > connector_status_connected
> > > > > :
> > > > > +						  connector_sta
> > > > > tus_disc
> > > > > onnected;
> > > > > +
> > > > > +	if (event < 0)
> > > > 
> > > > event is always > 0, isn't it?
> > > > 
> > > 
> > > Hello CK,
> > > 
> > > ok, I will move this to dp patch.
> > > 
> > > > > +		return IRQ_HANDLED;
> > > > > +
> > > > > +	if (mtk_dp->drm_dev) {
> > > > > +		dev_info(mtk_dp->dev,
> > > > > "drm_helper_hpd_irq_event\n");
> > > > > +		drm_helper_hpd_irq_event(mtk_dp->bridge.dev);
> > > > 
> > > > I think this ISR would come once. If bridge has not attached,
> > > > the
> > > > drm
> > > > core would lost this event. Maybe you should enable eDP
> > > > hardware
> > > > after
> > > > bridge attached or send this event when attached.
> > > > 
> > > 
> > > for edp patch, I will move it to (mtk_dp_bridge_attach).
> > > for dp patch, I will add it back.
> > 
> > I find out that mtk_dp_poweron() is in top of
> > mtk_dp_bridge_attach().
> > If move mtk_dp_poweron() to bottom of mtk_dp_bridge_attach(),
> > mtk_dp-
> > > drm_dev would not be NULL here. So we could drop this checking.
> > > 
> 
> Hello CK,
> 
> If we failed to setup phy(ret!=0), we alos need to deattach this
> bridge.
> I don't think  it's a good idea just for remove this.

OK, move mtk_dp_hwirq_enable() out of mtk_dp_poweron() and to the
bottom of mtk_dp_bridge_attach(). irq is not part of power.

> 
> > > > > +	}
> > > > > +
> > > > > +	if (mtk_dp->train_info.cable_state_change) {
> > > > 
> > > > Executing this thread imply cable_state_change = true, so drop
> > > > cable_state_change.
> > > > 
> > > 
> > > In mtk_dp_hpd_isr_handler(), there is another irq
> > > "MTK_DP_HPD_INTERRUPT" which means the sink devices give a
> > > interrupt
> > > to
> > > source device. it's not about connected status, so I think we
> > > still
> > > need this.
> > 
> > In bottom of mtk_dp_hpd_isr_handler(), the code is:
> > 
> > +	train_info->cable_state_change = true;
> > +
> > +	return IRQ_WAKE_THREAD;
> > 
> > This thread is called only when return IRQ_WAKE_THREAD, and before
> > return IRQ_WAKE_THREAD, train_info->cable_state_change is always
> > set
> > to
> > true. So in this thread, train_info->cable_state_change must be
> > true.
> > 
> 
> As mentioned, this irq handler function is not only for connected
> status.
> 
> this could be return if this irq is interrupt from sink device.
> +	if (!(train_info->irq_status &
> +	      (MTK_DP_HPD_CONNECT | MTK_DP_HPD_DISCONNECT)))
> +		return IRQ_HANDLED;

According to [1], return IRQ_WAKE_THREAD to wake up thread. So return
IRQ_HANDLED would not wake up thread.

[1] 
https://www.kernel.org/doc/htmldocs/kernel-api/API-request-threaded-irq.html

Regards,
CK

> 
> BRs,
> Bo-Chen
> > Regards,
> > CK
> > 
> > > 
> > > > > +		mtk_dp->train_info.cable_state_change = false;
> > > > > +
> > > > > +		mtk_dp->train_state =
> > > > > MTK_DP_TRAIN_STATE_STARTUP;
> > > > > +
> > > > > +		if (!mtk_dp->train_info.cable_plugged_in ||
> > > > > +		    !mtk_dp_plug_state(mtk_dp)) {
> > > > 
> > > > I do not like two variable to present one thing. If
> > > > 
> > > > mtk_dp->train_info.cable_plugged_in = false
> > > > and
> > > > mtk_dp_plug_state(mtk_dp) = ture
> > > > 
> > > > What does this mean? I think this mean 'now' is connected
> > > > because
> > > > cable_plugged_in is old information and mtk_dp_plug_state() is
> > > > current
> > > > information.
> > > > 
> > > > But I would like to keep cable_plugged_in and drop
> > > > mtk_dp_plug_state()
> > > > because cable_plugged_in would be changed in isr and it would
> > > > be
> > > > the
> > > > same as mtk_dp_plug_state().
> > > > 
> > > > Regards,
> > > > CK
> > > > 
> > > 
> > > ok, I will drop this.
> > > 
> > > BRs,
> > > Rex
> > > 
> > > > > +			mtk_dp_video_mute(mtk_dp, true);
> > > > > +
> > > > > +			mtk_dp_initialize_priv_data(mtk_dp);
> > > > > +			mtk_dp_set_idle_pattern(mtk_dp, true);
> > > > > +			if (mtk_dp->has_fec)
> > > > > +				mtk_dp_fec_enable(mtk_dp,
> > > > > false);
> > > > > +
> > > > > +			mtk_dp_update_bits(mtk_dp,
> > > > > MTK_DP_TOP_PWR_STATE,
> > > > > +					   DP_PWR_STATE_BANDGAP
> > > > > _TPLL,
> > > > > +					   DP_PWR_STATE_MASK);
> > > > > +		} else {
> > > > > +			mtk_dp_update_bits(mtk_dp,
> > > > > MTK_DP_TOP_PWR_STATE,
> > > > > +					   DP_PWR_STATE_BANDGAP
> > > > > _TPLL_LA
> > > > > NE,
> > > > > +					   DP_PWR_STATE_MASK);
> > > > > +			drm_dp_read_dpcd_caps(&mtk_dp->aux,
> > > > > buf);
> > > > > +			mtk_dp->train_info.link_rate =
> > > > > +				min_t(int, mtk_dp-
> > > > > > max_linkrate,
> > > > > 
> > > > > +				      buf[mtk_dp-
> > > > > > max_linkrate]);
> > > > > 
> > > > > +			mtk_dp->train_info.lane_count =
> > > > > +				min_t(int, mtk_dp->max_lanes,
> > > > > +				      drm_dp_max_lane_count(buf
> > > > > ));
> > > > > +		}
> > > > > +	}
> > > > > +
> > > > > +	if (mtk_dp->train_info.irq_status &
> > > > > MTK_DP_HPD_INTERRUPT) {
> > > > > +		dev_dbg(mtk_dp->dev, "MTK_DP_HPD_INTERRUPT\n");
> > > > > +		mtk_dp->train_info.irq_status &=
> > > > > ~MTK_DP_HPD_INTERRUPT;
> > > > > +		mtk_dp_hpd_sink_event(mtk_dp);
> > > > > +	}
> > > > > +
> > > > > +	return IRQ_HANDLED;
> > > > > +}
> > > > > +
> > > > 
> > > > 
> > > 
> > > 
> > 
> > 
> 
>