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

[CK Hu]

Hi, Rex:

On Tue, 2022-06-07 at 20:44 +0800, Rex-BC Chen wrote:
> On Tue, 2022-06-07 at 14:44 +0800, CK Hu wrote:
> > Hi, Rex:
> > 
> > On Mon, 2022-05-23 at 12:47 +0200, Guillaume Ranquet wrote:
> > > From: Markus Schneider-Pargmann <msp@xxxxxxxxxxxx>
> > > 
> > > 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@xxxxxxxxxxxx>.
> > > 
> > > Signed-off-by: Markus Schneider-Pargmann <msp@xxxxxxxxxxxx>
> > > Signed-off-by: Guillaume Ranquet <granquet@xxxxxxxxxxxx>
> > > ---
> > 
> > [snip]
> > 
> > > +
> > > +static int mtk_dp_train_handler(struct mtk_dp *mtk_dp)
> > > +{
> > > +	bool training_done = false;
> > > +	short max_retry = 50;
> > > +	int ret = 0;
> > > +
> > > +	do {
> > > +		switch (mtk_dp->train_state) {
> > > +		case MTK_DP_TRAIN_STATE_STARTUP:
> > 
> > mtk_dp->train_state is initialized as MTK_DP_TRAIN_STATE_STARTUP
> > even
> > though HPD ISR does not exist. Does this mean HPD ISR is redundant?
> > If
> > HPD ISR is not redundant, create a new state
> > MTK_DP_TRAIN_STATE_NONE
> > for init state.
> > 
> 
> Hello CK,
> 
> I think we don't need MTK_DP_TRAIN_STATE_NONE.
> Because it's "DP_TRAIN_STATE" not "DP_STATE", I think it's ok if we
> start this state machine with "MTK_DP_TRAIN_STATE_STARTUP".

The initial state is MTK_DP_TRAIN_STATE_STARTUP, and HPD thread would
change state from MTK_DP_TRAIN_STATE_STARTUP to
MTK_DP_TRAIN_STATE_STARTUP, this is redundant. So drop the state change
in HPD thread.

> 
> > > +			mtk_dp_state_handler(mtk_dp);
> > > +			mtk_dp->train_state =
> > > MTK_DP_TRAIN_STATE_CHECKCAP;
> > > +			break;
> > > +
> > > +		case MTK_DP_TRAIN_STATE_CHECKCAP:
> > > +			if (mtk_dp_parse_capabilities(mtk_dp)) {
> > > +				mtk_dp->train_info.check_cap_count = 0;
> > > +				mtk_dp->train_state =
> > > MTK_DP_TRAIN_STATE_CHECKEDID;
> > > +			} else {
> > > +				mtk_dp->train_info.check_cap_count++;
> > > +
> > > +				if (mtk_dp->train_info.check_cap_count
> > > > 
> > > 
> > > +				    MTK_DP_CHECK_SINK_CAP_TIMEOUT_COUNT
> > > ) {
> > > +					mtk_dp-
> > > > train_info.check_cap_count = 0;
> > > 
> > > +					mtk_dp->train_state =
> > > MTK_DP_TRAIN_STATE_DPIDLE;
> > > +					ret = -ETIMEDOUT;
> > > +				}
> > > +			}
> > > +			break;
> > > +
> > > +		case MTK_DP_TRAIN_STATE_CHECKEDID:
> > > +			mtk_dp->train_state =
> > > MTK_DP_TRAIN_STATE_TRAINING_PRE;
> > 
> > MTK_DP_TRAIN_STATE_CHECKEDID is a redundant state, drop it.
> > 
> > > +			break;
> > > +
> > > +		case MTK_DP_TRAIN_STATE_TRAINING_PRE:
> > > +			mtk_dp_state_handler(mtk_dp);
> > > +			mtk_dp->train_state =
> > > MTK_DP_TRAIN_STATE_TRAINING;
> > > +			break;
> > > +
> > > +		case MTK_DP_TRAIN_STATE_TRAINING:
> > > +			ret = mtk_dp_train_start(mtk_dp);
> > > +			if (ret == 0) {
> > > +				mtk_dp_video_mute(mtk_dp, true);
> > > +				mtk_dp->train_state =
> > > MTK_DP_TRAIN_STATE_NORMAL;
> > > +				mtk_dp_fec_enable(mtk_dp, mtk_dp-
> > > > has_fec);
> > > 
> > > +			} else if (ret != -EAGAIN) {
> > > +				mtk_dp->train_state =
> > > MTK_DP_TRAIN_STATE_DPIDLE;
> > > +			}
> > > +			break;
> > > +		case MTK_DP_TRAIN_STATE_NORMAL:
> > > +			mtk_dp_state_handler(mtk_dp);
> > > +			training_done = true;
> > > +			break;
> > > +		case MTK_DP_TRAIN_STATE_DPIDLE:
> > 
> > When would this case happen?
> > 
> > Regards,
> > CK
> 
> Yes, if it's disconnected if we are still training for dp.
> or failed to training min spec RBR.

I mean, every time state change to MTK_DP_TRAIN_STATE_DPIDLE, it would
jump out of this loop and would not get into this loop, so this case
would never get in. This is redundant, so remove this.

Regards,
CK

> 
> BRs,
> Rex
> > 
> > > +			break;
> > > +		default:
> > > +			break;
> > > +		}
> > > +
> > > +		if (ret) {
> > > +			if (ret == -EAGAIN)
> > > +				continue;
> > > +			/*
> > > +			 * If we get any other error number, it doesn't
> > > +			 * make any sense to keep iterating.
> > > +			 */
> > > +			break;
> > > +		}
> > > +	} while (!training_done || --max_retry);
> > > +
> > > +	return ret;
> > > +}
> > 
> > 
> 
>