novena: Fix ethernet PHY reset sequence
This patch fixes conflict between PHY pins becoming outputs after reset and
imx6 still driving the pins. It also fixes the reset timing as recommended by
the PHY datasheet.
Signed-off-by: Nikolay Dimitrov <picmaster@mail.bg>
Cc: Stefano Babic <sbabic@denx.de>
Cc: Sean Cross <xobs@kosagi.com>
Cc: Marek Vasut <marex@denx.de>
Reviewed-by: Fabio Estevam <fabio.estevam@freescale.com>
diff --git a/board/kosagi/novena/novena_spl.c b/board/kosagi/novena/novena_spl.c
index a8389d9..c4155dd 100644
--- a/board/kosagi/novena/novena_spl.c
+++ b/board/kosagi/novena/novena_spl.c
@@ -41,6 +41,10 @@
PAD_CTL_PUS_100K_UP | PAD_CTL_SPEED_MED | \
PAD_CTL_DSE_40ohm | PAD_CTL_HYS)
+#define ENET_PHY_CFG_PAD_CTRL \
+ (PAD_CTL_PKE | PAD_CTL_PUE | \
+ PAD_CTL_PUS_22K_UP | PAD_CTL_HYS)
+
#define RGMII_PAD_CTRL \
(PAD_CTL_PKE | PAD_CTL_PUE | \
PAD_CTL_PUS_100K_UP | PAD_CTL_DSE_40ohm | PAD_CTL_HYS)
@@ -98,18 +102,20 @@
MX6_PAD_RGMII_TD3__RGMII_TD3 | MUX_PAD_CTRL(RGMII_PAD_CTRL),
MX6_PAD_RGMII_TX_CTL__RGMII_TX_CTL | MUX_PAD_CTRL(RGMII_PAD_CTRL),
MX6_PAD_ENET_REF_CLK__ENET_TX_CLK | MUX_PAD_CTRL(ENET_PAD_CTRL),
- /* pin 35 - 1 (PHY_AD2) on reset */
- MX6_PAD_RGMII_RXC__GPIO6_IO30 | MUX_PAD_CTRL(NO_PAD_CTRL),
- /* pin 32 - 1 - (MODE0) all */
- MX6_PAD_RGMII_RD0__GPIO6_IO25 | MUX_PAD_CTRL(NO_PAD_CTRL),
- /* pin 31 - 1 - (MODE1) all */
- MX6_PAD_RGMII_RD1__GPIO6_IO27 | MUX_PAD_CTRL(NO_PAD_CTRL),
- /* pin 28 - 1 - (MODE2) all */
- MX6_PAD_RGMII_RD2__GPIO6_IO28 | MUX_PAD_CTRL(NO_PAD_CTRL),
- /* pin 27 - 1 - (MODE3) all */
- MX6_PAD_RGMII_RD3__GPIO6_IO29 | MUX_PAD_CTRL(NO_PAD_CTRL),
- /* pin 33 - 1 - (CLK125_EN) 125Mhz clockout enabled */
- MX6_PAD_RGMII_RX_CTL__GPIO6_IO24 | MUX_PAD_CTRL(NO_PAD_CTRL),
+
+ /* pin 35, PHY_AD2 */
+ MX6_PAD_RGMII_RXC__GPIO6_IO30 | MUX_PAD_CTRL(ENET_PHY_CFG_PAD_CTRL),
+ /* pin 32, MODE0 */
+ MX6_PAD_RGMII_RD0__GPIO6_IO25 | MUX_PAD_CTRL(ENET_PHY_CFG_PAD_CTRL),
+ /* pin 31, MODE1 */
+ MX6_PAD_RGMII_RD1__GPIO6_IO27 | MUX_PAD_CTRL(ENET_PHY_CFG_PAD_CTRL),
+ /* pin 28, MODE2 */
+ MX6_PAD_RGMII_RD2__GPIO6_IO28 | MUX_PAD_CTRL(ENET_PHY_CFG_PAD_CTRL),
+ /* pin 27, MODE3 */
+ MX6_PAD_RGMII_RD3__GPIO6_IO29 | MUX_PAD_CTRL(ENET_PHY_CFG_PAD_CTRL),
+ /* pin 33, CLK125_EN */
+ MX6_PAD_RGMII_RX_CTL__GPIO6_IO24 | MUX_PAD_CTRL(ENET_PHY_CFG_PAD_CTRL),
+
/* pin 42 PHY nRST */
MX6_PAD_EIM_D23__GPIO3_IO23 | MUX_PAD_CTRL(NO_PAD_CTRL),
};
@@ -127,15 +133,37 @@
{
imx_iomux_v3_setup_multiple_pads(enet_pads1, ARRAY_SIZE(enet_pads1));
+ /* Assert Ethernet PHY nRST */
gpio_direction_output(IMX_GPIO_NR(3, 23), 0);
- gpio_direction_output(IMX_GPIO_NR(6, 30), 1);
- gpio_direction_output(IMX_GPIO_NR(6, 25), 1);
- gpio_direction_output(IMX_GPIO_NR(6, 27), 1);
- gpio_direction_output(IMX_GPIO_NR(6, 28), 1);
- gpio_direction_output(IMX_GPIO_NR(6, 29), 1);
- gpio_direction_output(IMX_GPIO_NR(6, 24), 1);
+ /*
+ * Use imx6 internal pull-ups to drive PHY mode pins during PHY reset
+ * de-assertion. The intention is to use weak signal drivers (pull-ups)
+ * to prevent the conflict between PHY pins becoming outputs after
+ * reset and imx6 still driving the pins. The issue is described in PHY
+ * datasheet, p.14
+ */
+ gpio_direction_input(IMX_GPIO_NR(6, 30)); /* PHY_AD2 = 1 */
+ gpio_direction_input(IMX_GPIO_NR(6, 25)); /* MODE0 = 1 */
+ gpio_direction_input(IMX_GPIO_NR(6, 27)); /* MODE1 = 1 */
+ gpio_direction_input(IMX_GPIO_NR(6, 28)); /* MODE2 = 1 */
+ gpio_direction_input(IMX_GPIO_NR(6, 29)); /* MODE3 = 1 */
+ gpio_direction_input(IMX_GPIO_NR(6, 24)); /* CLK125_EN = 1 */
+
+ /* Following reset timing (p.53, fig.8 from the PHY datasheet) */
+ mdelay(10);
+
+ /* De-assert Ethernet PHY nRST */
+ gpio_set_value(IMX_GPIO_NR(3, 23), 1);
+
+ /* PHY is now configured, connect FEC to the pads */
imx_iomux_v3_setup_multiple_pads(enet_pads2, ARRAY_SIZE(enet_pads2));
+
+ /*
+ * PHY datasheet recommends on p.53 to wait at least 100us after reset
+ * before using MII, so we enforce the delay here
+ */
+ udelay(100);
}
/*