On 2021-08-25 10:49, Stephen Boyd wrote:

Quoting Kuogee Hsieh (2021-08-24 16:29:35)

quoted

Add voltage and pre emphesis swing tables so that voltage and

Is it "pre-emphasis"?

quoted

pre emphsis swing level can be configured base on link rate.

This one is also different.

quoted

Signed-off-by: Kuogee Hsieh <redacted>

Presumably

Fixes: aff188feb5e1 ("phy: qcom-qmp: add support for sm8250-usb3-dp 
phy")

quoted

---
 drivers/phy/qualcomm/phy-qcom-qmp.c | 95 
++++++++++++++++++++++++++++++++-----
 1 file changed, 82 insertions(+), 13 deletions(-)

diff --git a/drivers/phy/qualcomm/phy-qcom-qmp.c 

b/drivers/phy/qualcomm/phy-qcom-qmp.c
index 31036aa..52bab6e 100644

--- a/drivers/phy/qualcomm/phy-qcom-qmp.c
+++ b/drivers/phy/qualcomm/phy-qcom-qmp.c

@@ -1916,7 +1916,7 @@ static const struct qmp_phy_init_tbl 

qmp_v4_dp_tx_tbl[] = {
        QMP_PHY_INIT_CFG(QSERDES_V4_TX_RES_CODE_LANE_OFFSET_RX, 0x11),
        QMP_PHY_INIT_CFG(QSERDES_V4_TX_TX_BAND, 0x4),
        QMP_PHY_INIT_CFG(QSERDES_V4_TX_TX_POL_INV, 0x0a),
-       QMP_PHY_INIT_CFG(QSERDES_V4_TX_TX_DRV_LVL, 0x2a),
+       QMP_PHY_INIT_CFG(QSERDES_V4_TX_TX_DRV_LVL, 0x22),

Is 0x22 the better "default"? Can that be described in the commit text?

quoted

        QMP_PHY_INIT_CFG(QSERDES_V4_TX_TX_EMP_POST1_LVL, 0x20),
 };

@@ -3727,6 +3727,81 @@ static int qcom_qmp_v3_dp_phy_calibrate(struct 

qmp_phy *qphy)

        return 0;
 }

Nitpick: Newline here please.

quoted

+/*
+ * 0x20 deducted from tables
+ *
+ * swing_value |= DP_PHY_TXn_TX_DRV_LVL_MUX_EN;
+ * pre_emphasis_value |= DP_PHY_TXn_TX_EMP_POST1_LVL_MUX_EN;
+*/
+static const u8 qmp_dp_v4_pre_emphasis_hbr3_hbr2[4][4] = {
+       /* p0    p1    p2    p3 */
+       { 0x00, 0x0c, 0x15, 0x1b },     /* s0 */
+       { 0x02, 0x0e, 0x16, 0xff },     /* s1 */
+       { 0x02, 0x11, 0xff, 0xff },     /* s2 */
+       { 0x04, 0xff, 0xff, 0xff }      /* s3 */
+};
+
+static const u8 qmp_dp_v4_voltage_swing_hbr3_hbr2[4][4] = {

This looks the same as qmp_dp_v3_voltage_swing_hbr3_hbr2. Can that be
used?

to avoid confuse, i like to keep them separated.

quoted

+       /* p0    p1    p2    p3 */
+       { 0x02, 0x12, 0x16, 0x1a },     /* s0 */
+       { 0x09, 0x19, 0x1f, 0xff },     /* s1 */
+       { 0x10, 0x1f, 0xff, 0xff },     /* s2 */
+       { 0x1f, 0xff, 0xff, 0xff }      /* s3 */
+};
+
+static const u8 qmp_dp_v4_pre_emphasis_hbr_rbr[4][4] = {
+       /* p0    p1    p2    p3 */
+       { 0x00, 0x0e, 0x15, 0x1b },     /* s0 */
+       { 0x00, 0x0e, 0x15, 0xff },     /* s1 */
+       { 0x00, 0x0e, 0xff, 0xff },     /* s2 */
+       { 0x04, 0xff, 0xff, 0xff }      /* s3 */
+};
+
+static const u8 qmp_dp_v4_voltage_swing_hbr_rbr[4][4] = {
+       /* p0    p1    p2    p3 */
+       { 0x08, 0x0f, 0x16, 0x1f },     /* s0 */
+       { 0x11, 0x1e, 0x1f, 0xff },     /* s1 */
+       { 0x16, 0x1f, 0xff, 0xff },     /* s2 */
+       { 0x1f, 0xff, 0xff, 0xff }      /* s3 */

Do these comments add any value? Can we drop them?

quoted

+};
+
+static int qcom_qmp_v4_phy_configure_dp_swing(struct qmp_phy *qphy,
+               unsigned int drv_lvl_reg, unsigned int emp_post_reg)
+{
+       const struct phy_configure_opts_dp *dp_opts = &qphy->dp_opts;
+       unsigned int v_level = 0, p_level = 0;
+       u8 voltage_swing_cfg, pre_emphasis_cfg;
+       int i;
+
+       for (i = 0; i < dp_opts->lanes; i++) {
+               v_level = max(v_level, dp_opts->voltage[i]);
+               p_level = max(p_level, dp_opts->pre[i]);
+       }
+
+

Nitpick: Drop extra newline.

quoted

+       if (dp_opts->link_rate <= 2700) {
+               voltage_swing_cfg = 
qmp_dp_v4_voltage_swing_hbr_rbr[v_level][p_level];
+               pre_emphasis_cfg = 
qmp_dp_v4_pre_emphasis_hbr_rbr[v_level][p_level];
+       } else {
+               voltage_swing_cfg = 
qmp_dp_v4_voltage_swing_hbr3_hbr2[v_level][p_level];
+               pre_emphasis_cfg = 
qmp_dp_v4_pre_emphasis_hbr3_hbr2[v_level][p_level];
+       }
+
+       /* TODO: Move check to config check */
+       if (voltage_swing_cfg == 0xFF && pre_emphasis_cfg == 0xFF)
+               return -EINVAL;
+
+       /* Enable MUX to use Cursor values from these registers */
+       voltage_swing_cfg |= DP_PHY_TXn_TX_DRV_LVL_MUX_EN;
+       pre_emphasis_cfg |= DP_PHY_TXn_TX_EMP_POST1_LVL_MUX_EN;
+
+       writel(voltage_swing_cfg, qphy->tx + drv_lvl_reg);
+       writel(pre_emphasis_cfg, qphy->tx + emp_post_reg);
+       writel(voltage_swing_cfg, qphy->tx2 + drv_lvl_reg);
+       writel(pre_emphasis_cfg, qphy->tx2 + emp_post_reg);

This is copy/pasted from qcom_qmp_phy_configure_dp_swing() right? How
about making a function

static int
__qcom_qmp_phy_configure_dp_swing(struct qmp_phy *qphy,
				  unsigned int drv_lvl_reg,
				  unsigned int emp_post_reg,
				  const u8 **voltage_rbr_hbr,
				  const u8 **pre_emphasis_rbr_hbr,
				  const u8 **voltage_hbr3_hbr2,
				  const u8 **pre_emphasis_hbr3_hbr2)

that does the same stuff but allows the tables to be different.

quoted

+
+       return 0;
+}

 static void qcom_qmp_v4_phy_dp_aux_init(struct qmp_phy *qphy)
 {

@@ -3757,14 +3832,7 @@ static void qcom_qmp_v4_phy_dp_aux_init(struct 

qmp_phy *qphy)

 static void qcom_qmp_v4_phy_configure_dp_tx(struct qmp_phy *qphy)
 {
-       /* Program default values before writing proper values */
-       writel(0x27, qphy->tx + QSERDES_V4_TX_TX_DRV_LVL);
-       writel(0x27, qphy->tx2 + QSERDES_V4_TX_TX_DRV_LVL);
-
-       writel(0x20, qphy->tx + QSERDES_V4_TX_TX_EMP_POST1_LVL);
-       writel(0x20, qphy->tx2 + QSERDES_V4_TX_TX_EMP_POST1_LVL);
-
-       qcom_qmp_phy_configure_dp_swing(qphy,
+       qcom_qmp_v4_phy_configure_dp_swing(qphy,
                        QSERDES_V4_TX_TX_DRV_LVL,
                        QSERDES_V4_TX_TX_EMP_POST1_LVL);
 }

@@ -3885,6 +3953,9 @@ static int 

qcom_qmp_v4_phy_configure_dp_phy(struct qmp_phy *qphy)
        writel(drvr1_en, qphy->tx2 + QSERDES_V4_TX_HIGHZ_DRVR_EN);
        writel(bias1_en, qphy->tx2 + 
QSERDES_V4_TX_TRANSCEIVER_BIAS_EN);

+       writel(0x0a, qphy->tx + QSERDES_V4_TX_TX_POL_INV);
+       writel(0x0a, qphy->tx2 + QSERDES_V4_TX_TX_POL_INV);

Is this mentioned in the commit text? Is this fixing the sequence?
It doesn't look like we're adding tables.

add these since they are missed from HPG sequence.

quoted

+
        writel(0x18, qphy->pcs + QSERDES_DP_PHY_CFG);
        udelay(2000);
        writel(0x19, qphy->pcs + QSERDES_DP_PHY_CFG);

@@ -3896,11 +3967,9 @@ static int 

qcom_qmp_v4_phy_configure_dp_phy(struct qmp_phy *qphy)
                        10000))
                return -ETIMEDOUT;

-       writel(0x0a, qphy->tx + QSERDES_V4_TX_TX_POL_INV);
-       writel(0x0a, qphy->tx2 + QSERDES_V4_TX_TX_POL_INV);

-       writel(0x27, qphy->tx + QSERDES_V4_TX_TX_DRV_LVL);
-       writel(0x27, qphy->tx2 + QSERDES_V4_TX_TX_DRV_LVL);
+       writel(0x22, qphy->tx + QSERDES_V4_TX_TX_DRV_LVL);
+       writel(0x22, qphy->tx2 + QSERDES_V4_TX_TX_DRV_LVL);

        writel(0x20, qphy->tx + QSERDES_V4_TX_TX_EMP_POST1_LVL);
        writel(0x20, qphy->tx2 + QSERDES_V4_TX_TX_EMP_POST1_LVL);