On Thu, 2022-06-09 at 10:30 +0800, CK Hu wrote:

Hi, Rex:

On Wed, 2022-06-08 at 18:26 +0800, Rex-BC Chen wrote:

quoted

On Tue, 2022-06-07 at 15:47 +0800, CK Hu wrote:

quoted

Hi, Rex:

On Mon, 2022-05-23 at 12:47 +0200, Guillaume Ranquet wrote:

quoted

From: Markus Schneider-Pargmann <msp@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 [off-list ref].

Signed-off-by: Markus Schneider-Pargmann <msp@baylibre.com>
Signed-off-by: Guillaume Ranquet <redacted>
---

[snip]

quoted

+
+static int mtk_dp_hpd_sink_event(struct mtk_dp *mtk_dp)
+{
+	ssize_t ret;
+	u8 sink_count;
+	bool locked;
+	u8 link_status[DP_LINK_STATUS_SIZE] = {};
+	u32 sink_count_reg = DP_SINK_COUNT_ESI;
+	u32 link_status_reg = DP_LANE0_1_STATUS;
+
+	ret = drm_dp_dpcd_readb(&mtk_dp->aux, sink_count_reg,
&sink_count);
+	if (ret < 0) {
+		drm_err(mtk_dp->drm_dev, "Read sink count
failed:
%ld\n", ret);
+		return ret;
+	}
+
+	ret = drm_dp_dpcd_read(&mtk_dp->aux, link_status_reg,
link_status,
+			       sizeof(link_status));
+	if (!ret) {
+		drm_err(mtk_dp->drm_dev, "Read link status
failed:
%ld\n",
+			ret);
+		return ret;
+	}
+
+	locked = drm_dp_channel_eq_ok(link_status,
+				      mtk_dp-

quoted

train_info.lane_count);

+	if (!locked && mtk_dp->train_state >
MTK_DP_TRAIN_STATE_TRAINING_PRE)

Before enter this function, mtk_dp->train_state is set to
MTK_DP_TRAIN_STATE_STARTUP, so this never happen, drop this.

The interrupt from sink device could come any time. Why it's
impossible?

I still ask this question: "For eDP, when does disconnect happened?"
If
it indeed happen, I still do not like to change state here. When
disconnect, the training flow would result in fail finally and need
not
to change state here, but the training flow would block for a while
when disconnect. You could add some check point to check train_info-

quoted

cable_plugged_in to quickly break out the training flow.

Regards,
CK

Hello CK,

This function is to handle "sink interrupt" and we need to read the
status from sink device via aux channel.
edp spec is just add some extension rule from dp spec, this is define
in dp spec.

The basic spec is still "dp", and these are all define in dp spec.
Therefore, I think we should keep this.

BRs,
Bo-Chen

quoted

quoted

quoted

+		mtk_dp->train_state =
MTK_DP_TRAIN_STATE_TRAINING_PRE;
+
+	if (link_status[1] & DP_REMOTE_CONTROL_COMMAND_PENDING)
+		drm_dp_dpcd_writeb(&mtk_dp->aux,
DP_DEVICE_SERVICE_IRQ_VECTOR,
+				   DP_REMOTE_CONTROL_COMMAND_PE
NDING);
+
+	if (DP_GET_SINK_COUNT(sink_count) &&
+	    (link_status[2] &
DP_DOWNSTREAM_PORT_STATUS_CHANGED)) {
+		mtk_dp->train_info.check_cap_count = 0;
+		mtk_dp->train_state =
MTK_DP_TRAIN_STATE_CHECKEDID;

Why change state from MTK_DP_TRAIN_STATE_STARTUP to
MTK_DP_TRAIN_STATE_CHECKEDID? In mtk_dp_train_handler(),
mtk_dp_parse_capabilities() is true then change to
MTK_DP_TRAIN_STATE_CHECKEDID. Give a reason why these two are
different.

Regards,
CK

I will drop this and drop state of MTK_DP_TRAIN_STATE_CHECKEDID.
MTK_DP_TRAIN_STATE_CHECKEDID is only used for audio.
We can check enable status in another place.

BRs,
Bo-Chen

quoted

quoted

+		msleep(20);
+	}
+
+	return 0;
+}
+