Both PHYs are 4-port PHY that are 10/100/1000BASE-T, 100BASE-FX, 1000BASE-X
and triple-speed copper SFP capable, can communicate with the MAC via
SGMII, QSGMII or 1000BASE-X, supports downshifting and can set the blinking
pattern of each of its 4 LEDs, supports SyncE as well as HP Auto-MDIX
detection.
VSC8574 supports WOL and VSC8584 supports hardware offloading of MACsec.
This patch series add support for 10/100/1000BASE-T, SGMII/QSGMII link with
the MAC, downshifting, HP Auto-MDIX detection and blinking pattern for
their 4 LEDs.
They have also an internal Intel 8051 microcontroller whose firmware needs
to be patched when the PHY is reset. If the 8051's firmware has the
expected CRC, its patching can be skipped. The microcontroller can be
accessed from any port of the PHY, though the CRC function can only be done
through the PHY that is the base PHY of the package (internal address 0)
due to a limitation of the firmware.
The GPIO register bank is a set of registers that are common to all PHYs in
the package. So any modification in any register of this bank affects all
PHYs of the package.
If the PHYs haven't been reset before booting the Linux kernel and were
configured to use interrupts for e.g. link status updates, it is required
to clear the interrupts mask register of all PHYs before being able to use
interrupts with any PHY. The first PHY of the package that will be init
will take care of clearing all PHYs interrupts mask registers. Thus, we
need to keep track of the init sequence in the package, if it's already
been done or if it's to be done.
Most of the init sequence of a PHY of the package is common to all PHYs in
the package, thus we use the SMI broadcast feature which enables us to
propagate a write in one register of one PHY to all PHYs in the package.
We also introduce a new development board called PCB120 which exists in
variants for VSC8584 and VSC8574 (and that's the only difference to the
best of my knowledge).
I suggest patches 1 to 4 go through net tree and patches 5 to 7 go through
MIPS tree. Patches going through net tree and those going through MIPS tree
do not depend on one another.
This patch series depends on two patch series though:
"mscc: ocelot: add support for SerDes muxing configuration"
(https://lore.kernel.org/lkml/cover.ff40d591b548a6da31716e6e600f11a303e0e643.1536912834.git-series.quentin.schulz@bootlin.com/)
"Various improvements to Microsemi PHY driver"
(https://lore.kernel.org/lkml/cover.616d15610d44a0e3d463acd8119859f243163ad2.1536913944.git-series.quentin.schulz@bootlin.com/)
specifically patch 2/5 which defines constants that are used in this patch
series.
Thanks,
Quentin
Quentin Schulz (7):
dt-bindings: net: vsc8531: add two additional LED modes for VSC8584
net: phy: mscc: add support for VSC8584 PHY
net: phy: mscc: split config_init in two functions for VSC8584
net: phy: mscc: add support for VSC8574 PHY
MIPS: mscc: ocelot: add GPIO4 pinmuxing DT node
MIPS: mscc: add DT for Ocelot PCB120
MIPS: mscc: add PCB120 to the ocelot fitImage
arch/mips/boot/dts/mscc/Makefile | 2 +-
arch/mips/boot/dts/mscc/ocelot.dtsi | 5 +-
arch/mips/boot/dts/mscc/ocelot_pcb120.dts | 100 ++-
arch/mips/generic/Kconfig | 6 +-
arch/mips/generic/Platform | 2 +-
arch/mips/generic/board-ocelot.its.S | 40 +-
arch/mips/generic/board-ocelot_pcb123.its.S | 23 +-
drivers/net/phy/mscc.c | 1019 ++++++++++++++++++++-
include/dt-bindings/net/mscc-phy-vsc8531.h | 2 +-
9 files changed, 1171 insertions(+), 28 deletions(-)
create mode 100644 arch/mips/boot/dts/mscc/ocelot_pcb120.dts
create mode 100644 arch/mips/generic/board-ocelot.its.S
delete mode 100644 arch/mips/generic/board-ocelot_pcb123.its.S
base-commit: d9cca8eef36bb8918c9ed28574b79b7674fd36f6
--
git-series 0.9.1
The VSC8584 (and most likely other PHYs in the same generation) has two
additional LED modes that can be picked, so let's add them.
Signed-off-by: Quentin Schulz <[email protected]>
---
include/dt-bindings/net/mscc-phy-vsc8531.h | 2 ++
1 file changed, 2 insertions(+)
diff --git a/include/dt-bindings/net/mscc-phy-vsc8531.h b/include/dt-bindings/net/mscc-phy-vsc8531.h
index 697161f..9eb2ec2 100644
--- a/include/dt-bindings/net/mscc-phy-vsc8531.h
+++ b/include/dt-bindings/net/mscc-phy-vsc8531.h
@@ -18,9 +18,11 @@
#define VSC8531_LINK_100_1000_ACTIVITY 4
#define VSC8531_LINK_10_1000_ACTIVITY 5
#define VSC8531_LINK_10_100_ACTIVITY 6
+#define VSC8584_LINK_100FX_1000X_ACTIVITY 7
#define VSC8531_DUPLEX_COLLISION 8
#define VSC8531_COLLISION 9
#define VSC8531_ACTIVITY 10
+#define VSC8584_100FX_1000X_ACTIVITY 11
#define VSC8531_AUTONEG_FAULT 12
#define VSC8531_SERIAL_MODE 13
#define VSC8531_FORCE_LED_OFF 14
--
git-series 0.9.1
Part of the config init is common between the VSC8584 and the VSC8574,
so to prepare the upcoming support for VSC8574, separate config_init
PHY-specific code to config_pre_init function which is set in the probe
function of the PHY and used in config_init.
Signed-off-by: Quentin Schulz <[email protected]>
---
drivers/net/phy/mscc.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/net/phy/mscc.c b/drivers/net/phy/mscc.c
index b450489..69cc3cf 100644
--- a/drivers/net/phy/mscc.c
+++ b/drivers/net/phy/mscc.c
@@ -355,6 +355,7 @@ struct vsc8531_private {
u64 *stats;
int nstats;
bool pkg_init;
+ int (*config_pre_init)(struct mii_bus *bus, int phy);
};
#ifdef CONFIG_OF_MDIO
@@ -1298,7 +1299,7 @@ static int vsc8584_config_init(struct phy_device *phydev)
*/
if (!vsc8584_is_pkg_init(phydev, base_addr,
val & PHY_ADDR_REVERSED ? 1 : 0)) {
- ret = vsc8584_config_pre_init(phydev->mdio.bus, base_addr);
+ ret = vsc8531->config_pre_init(phydev->mdio.bus, base_addr);
if (ret)
goto err;
}
@@ -1486,6 +1487,7 @@ static int vsc8584_probe(struct phy_device *phydev)
phydev->priv = vsc8531;
+ vsc8531->config_pre_init = vsc8584_config_pre_init;
vsc8531->nleds = 4;
vsc8531->supp_led_modes = VSC8584_SUPP_LED_MODES;
vsc8531->hw_stats = vsc8584_hw_stats;
--
git-series 0.9.1
The Ocelot PCB120 evaluation board is different from the PCB123 in that
it has 4 external VSC8584 (or VSC8574) PHYs.
It uses the SoC's second MDIO bus for external PHYs which have a
reversed address on the bus (i.e. PHY4 is on address 3, PHY5 is on
address 2, PHY6 on 1 and PHY7 on 0).
Here is how the PHYs are connected to the switch ports:
port 0: phy0 (internal)
port 1: phy1 (internal)
port 2: phy2 (internal)
port 3: phy3 (internal)
port 4: phy7
port 5: phy4
port 6: phy6
port 9: phy5
Signed-off-by: Quentin Schulz <[email protected]>
---
arch/mips/boot/dts/mscc/Makefile | 2 +-
arch/mips/boot/dts/mscc/ocelot_pcb120.dts | 100 +++++++++++++++++++++++-
2 files changed, 101 insertions(+), 1 deletion(-)
create mode 100644 arch/mips/boot/dts/mscc/ocelot_pcb120.dts
diff --git a/arch/mips/boot/dts/mscc/Makefile b/arch/mips/boot/dts/mscc/Makefile
index 9a9bb7e..ec6f5b2 100644
--- a/arch/mips/boot/dts/mscc/Makefile
+++ b/arch/mips/boot/dts/mscc/Makefile
@@ -1,3 +1,3 @@
-dtb-$(CONFIG_MSCC_OCELOT) += ocelot_pcb123.dtb
+dtb-$(CONFIG_MSCC_OCELOT) += ocelot_pcb123.dtb ocelot_pcb120.dtb
obj-$(CONFIG_BUILTIN_DTB) += $(addsuffix .o, $(dtb-y))
diff --git a/arch/mips/boot/dts/mscc/ocelot_pcb120.dts b/arch/mips/boot/dts/mscc/ocelot_pcb120.dts
new file mode 100644
index 0000000..8eb03a5
--- /dev/null
+++ b/arch/mips/boot/dts/mscc/ocelot_pcb120.dts
@@ -0,0 +1,100 @@
+// SPDX-License-Identifier: (GPL-2.0 OR MIT)
+/* Copyright (c) 2017 Microsemi Corporation */
+
+/dts-v1/;
+
+#include <dt-bindings/interrupt-controller/irq.h>
+#include <dt-bindings/phy/phy-ocelot-serdes.h>
+#include "ocelot.dtsi"
+
+/ {
+ compatible = "mscc,ocelot-pcb120", "mscc,ocelot";
+
+ chosen {
+ stdout-path = "serial0:115200n8";
+ };
+
+ memory@0 {
+ device_type = "memory";
+ reg = <0x0 0x0e000000>;
+ };
+};
+
+&mdio0 {
+ status = "okay";
+};
+
+&mdio1 {
+ status = "okay";
+ pinctrl-names = "default";
+ pinctrl-0 = <&miim1>, <&gpio4>;
+
+ phy7: ethernet-phy@0 {
+ reg = <0>;
+ interrupts = <4 IRQ_TYPE_LEVEL_HIGH>;
+ interrupt-parent = <&gpio>;
+ };
+ phy6: ethernet-phy@1 {
+ reg = <1>;
+ interrupts = <4 IRQ_TYPE_LEVEL_HIGH>;
+ interrupt-parent = <&gpio>;
+ };
+ phy5: ethernet-phy@2 {
+ reg = <2>;
+ interrupts = <4 IRQ_TYPE_LEVEL_HIGH>;
+ interrupt-parent = <&gpio>;
+ };
+ phy4: ethernet-phy@3 {
+ reg = <3>;
+ interrupts = <4 IRQ_TYPE_LEVEL_HIGH>;
+ interrupt-parent = <&gpio>;
+ };
+};
+
+&port0 {
+ phy-handle = <&phy0>;
+};
+
+&port1 {
+ phy-handle = <&phy1>;
+};
+
+&port2 {
+ phy-handle = <&phy2>;
+};
+
+&port3 {
+ phy-handle = <&phy3>;
+};
+
+&port4 {
+ phy-handle = <&phy7>;
+ phy-mode = "sgmii";
+ phys = <&serdes 4 SERDES1G_2>;
+};
+
+&port5 {
+ phy-handle = <&phy4>;
+ phy-mode = "sgmii";
+ phys = <&serdes 5 SERDES1G_5>;
+};
+
+&port6 {
+ phy-handle = <&phy6>;
+ phy-mode = "sgmii";
+ phys = <&serdes 6 SERDES1G_3>;
+};
+
+&port9 {
+ phy-handle = <&phy5>;
+ phy-mode = "sgmii";
+ phys = <&serdes 9 SERDES1G_4>;
+};
+
+&uart0 {
+ status = "okay";
+};
+
+&uart2 {
+ status = "okay";
+};
--
git-series 0.9.1
In order to use GPIO4 as a GPIO, we need to mux it in this mode so let's
declare a new pinctrl DT node for it.
Signed-off-by: Quentin Schulz <[email protected]>
---
arch/mips/boot/dts/mscc/ocelot.dtsi | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/arch/mips/boot/dts/mscc/ocelot.dtsi b/arch/mips/boot/dts/mscc/ocelot.dtsi
index 8ce317c..b5c4c74 100644
--- a/arch/mips/boot/dts/mscc/ocelot.dtsi
+++ b/arch/mips/boot/dts/mscc/ocelot.dtsi
@@ -182,6 +182,11 @@
interrupts = <13>;
#interrupt-cells = <2>;
+ gpio4: gpio4 {
+ pins = "GPIO_4";
+ function = "gpio";
+ };
+
uart_pins: uart-pins {
pins = "GPIO_6", "GPIO_7";
function = "uart";
--
git-series 0.9.1
The VSC8574 PHY is a 4-ports PHY that is 10/100/1000BASE-T, 100BASE-FX,
1000BASE-X and triple-speed copper SFP capable, can communicate with
the MAC via SGMII, QSGMII or 1000BASE-X, supports WOL, downshifting and
can set the blinking pattern of each of its 4 LEDs, supports SyncE as
well as HP Auto-MDIX detection.
This adds support for 10/100/1000BASE-T, SGMII/QSGMII link with the MAC,
WOL, downshifting, HP Auto-MDIX detection and blinking pattern for its 4
LEDs.
The VSC8574 has also an internal Intel 8051 microcontroller whose
firmware needs to be patched when the PHY is reset. If the 8051's
firmware has the expected CRC, its patching can be skipped. The
microcontroller can be accessed from any port of the PHY, though the CRC
function can only be done through the PHY that is the base PHY of the
package (internal address 0) due to a limitation of the firmware.
The GPIO register bank is a set of registers that are common to all PHYs
in the package. So any modification in any register of this bank affects
all PHYs of the package.
If the PHYs haven't been reset before booting the Linux kernel and were
configured to use interrupts for e.g. link status updates, it is
required to clear the interrupts mask register of all PHYs before being
able to use interrupts with any PHY. The first PHY of the package that
will be init will take care of clearing all PHYs interrupts mask
registers. Thus, we need to keep track of the init sequence in the
package, if it's already been done or if it's to be done.
Most of the init sequence of a PHY of the package is common to all PHYs
in the package, thus we use the SMI broadcast feature which enables us
to propagate a write in one register of one PHY to all PHYs in the
package.
Signed-off-by: Quentin Schulz <[email protected]>
---
drivers/net/phy/mscc.c | 303 ++++++++++++++++++++++++++++++++++++++++++-
1 file changed, 303 insertions(+)
diff --git a/drivers/net/phy/mscc.c b/drivers/net/phy/mscc.c
index 69cc3cf..2289d0a 100644
--- a/drivers/net/phy/mscc.c
+++ b/drivers/net/phy/mscc.c
@@ -65,6 +65,8 @@ enum rgmii_rx_clock_delay {
#define MEDIA_OP_MODE_AMS_COPPER_100BASEFX 7
#define MEDIA_OP_MODE_POS 8
+#define MSCC_PHY_EXT_PHY_CNTL_2 24
+
#define MII_VSC85XX_INT_MASK 25
#define MII_VSC85XX_INT_MASK_MASK 0xa000
#define MII_VSC85XX_INT_MASK_WOL 0x0040
@@ -151,6 +153,7 @@ enum rgmii_rx_clock_delay {
#define DW8051_CLK_EN 0x0010
#define MICRO_CLK_EN 0x0008
#define MICRO_CLK_DIVIDE(x) ((x) >> 1)
+#define MSCC_DW8051_VLD_MASK 0xf1ff
/* x Address in range 1-4 */
#define MSCC_TRAP_ROM_ADDR(x) ((x) * 2 + 1)
@@ -184,7 +187,9 @@ enum rgmii_rx_clock_delay {
#define PROC_CMD_SGMII_MAC 0x0030
#define PROC_CMD_QSGMII_MAC 0x0020
#define PROC_CMD_NO_MAC_CONF 0x0000
+#define PROC_CMD_1588_DEFAULT_INIT 0x0010
#define PROC_CMD_NOP 0x000f
+#define PROC_CMD_PHY_INIT 0x000a
#define PROC_CMD_CRC16 0x0008
#define PROC_CMD_FIBER_MEDIA_CONF 0x0001
#define PROC_CMD_MCB_ACCESS_MAC_CONF 0x0000
@@ -198,6 +203,9 @@ enum rgmii_rx_clock_delay {
/* Test page Registers */
#define MSCC_PHY_TEST_PAGE_5 5
#define MSCC_PHY_TEST_PAGE_8 8
+#define MSCC_PHY_TEST_PAGE_9 9
+#define MSCC_PHY_TEST_PAGE_20 20
+#define MSCC_PHY_TEST_PAGE_24 24
/* Token ring page Registers */
#define MSCC_PHY_TR_CNTL 16
@@ -211,6 +219,7 @@ enum rgmii_rx_clock_delay {
#define PHY_ID_VSC8531 0x00070570
#define PHY_ID_VSC8540 0x00070760
#define PHY_ID_VSC8541 0x00070770
+#define PHY_ID_VSC8574 0x000704a0
#define PHY_ID_VSC8584 0x000707c0
#define MSCC_VDDMAC_1500 1500
@@ -258,6 +267,10 @@ enum rgmii_rx_clock_delay {
#define MSCC_VSC8584_REVB_INT8051_FW_START_ADDR 0xe800
#define MSCC_VSC8584_REVB_INT8051_FW_CRC 0xfb48
+#define MSCC_VSC8574_REVB_INT8051_FW "mscc_vsc8574_revb_int8051_29e8.bin"
+#define MSCC_VSC8574_REVB_INT8051_FW_START_ADDR 0x4000
+#define MSCC_VSC8574_REVB_INT8051_FW_CRC 0x29e8
+
#define VSC8584_REVB 0x0001
#define MSCC_DEV_REV_MASK GENMASK(3, 0)
@@ -1084,6 +1097,243 @@ static int vsc8584_patch_fw(struct mii_bus *bus, int phy,
}
/* bus->mdio_lock should be locked when using this function */
+static bool vsc8574_is_serdes_init(struct mii_bus *bus, int phy)
+{
+ u16 reg;
+ bool ret;
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS,
+ MSCC_PHY_PAGE_EXTENDED_GPIO);
+
+ reg = __mdiobus_read(bus, phy, MSCC_TRAP_ROM_ADDR(1));
+ if (reg != 0x3eb7) {
+ ret = false;
+ goto out;
+ }
+
+ reg = __mdiobus_read(bus, phy, MSCC_PATCH_RAM_ADDR(1));
+ if (reg != 0x4012) {
+ ret = false;
+ goto out;
+ }
+
+ reg = __mdiobus_read(bus, phy, MSCC_INT_MEM_CNTL);
+ if (reg != EN_PATCH_RAM_TRAP_ADDR(1)) {
+ ret = false;
+ goto out;
+ }
+
+ reg = __mdiobus_read(bus, phy, MSCC_DW8051_CNTL_STATUS);
+ if ((MICRO_NSOFT_RESET | RUN_FROM_INT_ROM | DW8051_CLK_EN |
+ MICRO_CLK_EN) != (reg & MSCC_DW8051_VLD_MASK)) {
+ ret = false;
+ goto out;
+ }
+
+ ret = true;
+out:
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_STANDARD);
+
+ return ret;
+}
+
+/* bus->mdio_lock should be locked when using this function */
+static int vsc8574_config_pre_init(struct mii_bus *bus, int phy)
+{
+ struct device *dev = &bus->mdio_map[phy]->dev;
+ const struct firmware *fw;
+ u16 crc, reg;
+ bool serdes_init;
+ int ret;
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_STANDARD);
+
+ /* all writes below this line are broadcasted to all PHYs */
+ reg = __mdiobus_read(bus, phy, MSCC_PHY_EXT_CNTL_STATUS);
+ reg |= SMI_BROADCAST_WR_EN;
+ __mdiobus_write(bus, phy, MSCC_PHY_EXT_CNTL_STATUS, reg);
+
+ __mdiobus_write(bus, phy, MII_VSC85XX_INT_MASK, 0);
+
+ /* The below register writes are tweaking analog and electrical
+ * configuration that were determined through characterization by PHY
+ * engineers. These don't mean anything more than "these are the best
+ * values".
+ */
+ __mdiobus_write(bus, phy, MSCC_PHY_EXT_PHY_CNTL_2, 0x0040);
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_TEST);
+
+ __mdiobus_write(bus, phy, MSCC_PHY_TEST_PAGE_20, 0x4320);
+ __mdiobus_write(bus, phy, MSCC_PHY_TEST_PAGE_24, 0x0c00);
+ __mdiobus_write(bus, phy, MSCC_PHY_TEST_PAGE_9, 0x18ca);
+ __mdiobus_write(bus, phy, MSCC_PHY_TEST_PAGE_5, 0x1b20);
+
+ reg = __mdiobus_read(bus, phy, MSCC_PHY_TEST_PAGE_8);
+ reg |= 0x8000;
+ __mdiobus_write(bus, phy, MSCC_PHY_TEST_PAGE_8, reg);
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_TR);
+
+ vsc8584_csr_write(bus, phy, 0x8fae, 0x000401bd);
+ vsc8584_csr_write(bus, phy, 0x8fac, 0x000f000f);
+ vsc8584_csr_write(bus, phy, 0x97a0, 0x00a0f147);
+ vsc8584_csr_write(bus, phy, 0x8fe4, 0x00052f54);
+ vsc8584_csr_write(bus, phy, 0x9792, 0x0027303d);
+ vsc8584_csr_write(bus, phy, 0x87fe, 0x00000704);
+ vsc8584_csr_write(bus, phy, 0x8fe0, 0x00060150);
+ vsc8584_csr_write(bus, phy, 0x8f82, 0x0012b00a);
+ vsc8584_csr_write(bus, phy, 0x8f80, 0x00000d74);
+ vsc8584_csr_write(bus, phy, 0x82e0, 0x00000012);
+ vsc8584_csr_write(bus, phy, 0x83a2, 0x00050208);
+ vsc8584_csr_write(bus, phy, 0x83b2, 0x00009186);
+ vsc8584_csr_write(bus, phy, 0x8fb0, 0x000e3700);
+ vsc8584_csr_write(bus, phy, 0x9688, 0x00049f81);
+ vsc8584_csr_write(bus, phy, 0x8fd2, 0x0000ffff);
+ vsc8584_csr_write(bus, phy, 0x968a, 0x00039fa2);
+ vsc8584_csr_write(bus, phy, 0x9690, 0x0020640b);
+ vsc8584_csr_write(bus, phy, 0x8258, 0x00002220);
+ vsc8584_csr_write(bus, phy, 0x825a, 0x00002a20);
+ vsc8584_csr_write(bus, phy, 0x825c, 0x00003060);
+ vsc8584_csr_write(bus, phy, 0x825e, 0x00003fa0);
+ vsc8584_csr_write(bus, phy, 0x83a6, 0x0000e0f0);
+ vsc8584_csr_write(bus, phy, 0x8f92, 0x00001489);
+ vsc8584_csr_write(bus, phy, 0x96a2, 0x00007000);
+ vsc8584_csr_write(bus, phy, 0x96a6, 0x00071448);
+ vsc8584_csr_write(bus, phy, 0x96a0, 0x00eeffdd);
+ vsc8584_csr_write(bus, phy, 0x8fe8, 0x0091b06c);
+ vsc8584_csr_write(bus, phy, 0x8fea, 0x00041600);
+ vsc8584_csr_write(bus, phy, 0x96b0, 0x00eeff00);
+ vsc8584_csr_write(bus, phy, 0x96b2, 0x00007000);
+ vsc8584_csr_write(bus, phy, 0x96b4, 0x00000814);
+ vsc8584_csr_write(bus, phy, 0x8f90, 0x00688980);
+ vsc8584_csr_write(bus, phy, 0x83a4, 0x0000d8f0);
+ vsc8584_csr_write(bus, phy, 0x8fc0, 0x00000400);
+ vsc8584_csr_write(bus, phy, 0x87fa, 0x0050100f);
+ vsc8584_csr_write(bus, phy, 0x8796, 0x00000003);
+ vsc8584_csr_write(bus, phy, 0x87f8, 0x00c3ff98);
+ vsc8584_csr_write(bus, phy, 0x8fa4, 0x0018292a);
+ vsc8584_csr_write(bus, phy, 0x968c, 0x00d2c46f);
+ vsc8584_csr_write(bus, phy, 0x97a2, 0x00000620);
+ vsc8584_csr_write(bus, phy, 0x96a4, 0x0013132f);
+ vsc8584_csr_write(bus, phy, 0x96a8, 0x00000000);
+ vsc8584_csr_write(bus, phy, 0x8ffc, 0x00c0a028);
+ vsc8584_csr_write(bus, phy, 0x8fec, 0x00901c09);
+ vsc8584_csr_write(bus, phy, 0x8fee, 0x0004a6a1);
+ vsc8584_csr_write(bus, phy, 0x8ffe, 0x00b01807);
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS,
+ MSCC_PHY_PAGE_EXTENDED_2);
+
+ __mdiobus_write(bus, phy, MSCC_PHY_CU_PMD_TX_CNTL, 0x028e);
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_TR);
+
+ vsc8584_csr_write(bus, phy, 0x8486, 0x0008a518);
+ vsc8584_csr_write(bus, phy, 0x8488, 0x006dc696);
+ vsc8584_csr_write(bus, phy, 0x848a, 0x00000912);
+ vsc8584_csr_write(bus, phy, 0x848e, 0x00000db6);
+ vsc8584_csr_write(bus, phy, 0x849c, 0x00596596);
+ vsc8584_csr_write(bus, phy, 0x849e, 0x00000514);
+ vsc8584_csr_write(bus, phy, 0x84a2, 0x00410280);
+ vsc8584_csr_write(bus, phy, 0x84a4, 0x00000000);
+ vsc8584_csr_write(bus, phy, 0x84a6, 0x00000000);
+ vsc8584_csr_write(bus, phy, 0x84a8, 0x00000000);
+ vsc8584_csr_write(bus, phy, 0x84aa, 0x00000000);
+ vsc8584_csr_write(bus, phy, 0x84ae, 0x007df7dd);
+ vsc8584_csr_write(bus, phy, 0x84b0, 0x006d95d4);
+ vsc8584_csr_write(bus, phy, 0x84b2, 0x00492410);
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_TEST);
+
+ reg = __mdiobus_read(bus, phy, MSCC_PHY_TEST_PAGE_8);
+ reg &= ~0x8000;
+ __mdiobus_write(bus, phy, MSCC_PHY_TEST_PAGE_8, reg);
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS,
+ MSCC_PHY_PAGE_STANDARD);
+
+ /* end of write broadcasting */
+ reg = __mdiobus_read(bus, phy, MSCC_PHY_EXT_CNTL_STATUS);
+ reg &= ~SMI_BROADCAST_WR_EN;
+ __mdiobus_write(bus, phy, MSCC_PHY_EXT_CNTL_STATUS, reg);
+
+ ret = request_firmware(&fw, MSCC_VSC8574_REVB_INT8051_FW, dev);
+ if (ret) {
+ dev_err(dev, "failed to load firmware %s, ret: %d\n",
+ MSCC_VSC8574_REVB_INT8051_FW, ret);
+ return ret;
+ }
+
+ /* Add one byte to size for the one added by the patch_fw function */
+ ret = vsc8584_get_fw_crc(bus, phy,
+ MSCC_VSC8574_REVB_INT8051_FW_START_ADDR,
+ fw->size + 1, &crc);
+ if (ret)
+ goto out;
+
+ if (crc == MSCC_VSC8574_REVB_INT8051_FW_CRC) {
+ serdes_init = vsc8574_is_serdes_init(bus, phy);
+
+ if (!serdes_init) {
+ ret = vsc8584_micro_assert_reset(bus, phy);
+ if (ret) {
+ dev_err(dev,
+ "%s: failed to assert reset of micro\n",
+ __func__);
+ return ret;
+ }
+ }
+ } else {
+ dev_dbg(dev, "FW CRC is not the expected one, patching FW\n");
+
+ serdes_init = false;
+
+ if (vsc8584_patch_fw(bus, phy, fw))
+ dev_warn(dev,
+ "failed to patch FW, expect non-optimal device\n");
+ }
+
+ if (!serdes_init) {
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS,
+ MSCC_PHY_PAGE_EXTENDED_GPIO);
+
+ __mdiobus_write(bus, phy, MSCC_TRAP_ROM_ADDR(1), 0x3eb7);
+ __mdiobus_write(bus, phy, MSCC_PATCH_RAM_ADDR(1), 0x4012);
+ __mdiobus_write(bus, phy, MSCC_INT_MEM_CNTL,
+ EN_PATCH_RAM_TRAP_ADDR(1));
+
+ vsc8584_micro_deassert_reset(bus, phy, false);
+
+ /* Add one byte to size for the one added by the patch_fw
+ * function
+ */
+ ret = vsc8584_get_fw_crc(bus, phy,
+ MSCC_VSC8574_REVB_INT8051_FW_START_ADDR,
+ fw->size + 1, &crc);
+ if (ret)
+ goto out;
+
+ if (crc != MSCC_VSC8574_REVB_INT8051_FW_CRC)
+ dev_warn(dev,
+ "FW CRC after patching is not the expected one, expect non-optimal device\n");
+ }
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS,
+ MSCC_PHY_PAGE_EXTENDED_GPIO);
+
+ ret = vsc8584_cmd(bus, phy, PROC_CMD_1588_DEFAULT_INIT |
+ PROC_CMD_PHY_INIT);
+
+out:
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_STANDARD);
+
+ release_firmware(fw);
+
+ return ret;
+}
+
+/* bus->mdio_lock should be locked when using this function */
static int vsc8584_config_pre_init(struct mii_bus *bus, int phy)
{
struct device *dev = &bus->mdio_map[phy]->dev;
@@ -1469,6 +1719,33 @@ static int vsc85xx_read_status(struct phy_device *phydev)
return genphy_read_status(phydev);
}
+static int vsc8574_probe(struct phy_device *phydev)
+{
+ struct vsc8531_private *vsc8531;
+ u32 default_mode[4] = {VSC8531_LINK_1000_ACTIVITY,
+ VSC8531_LINK_100_ACTIVITY, VSC8531_LINK_ACTIVITY,
+ VSC8531_DUPLEX_COLLISION};
+
+ vsc8531 = devm_kzalloc(&phydev->mdio.dev, sizeof(*vsc8531), GFP_KERNEL);
+ if (!vsc8531)
+ return -ENOMEM;
+
+ phydev->priv = vsc8531;
+
+ vsc8531->config_pre_init = vsc8574_config_pre_init;
+ vsc8531->nleds = 4;
+ vsc8531->supp_led_modes = VSC8584_SUPP_LED_MODES;
+ vsc8531->hw_stats = vsc8584_hw_stats;
+ vsc8531->nstats = ARRAY_SIZE(vsc8584_hw_stats);
+ vsc8531->stats = devm_kzalloc(&phydev->mdio.dev,
+ sizeof(u64) * vsc8531->nstats,
+ GFP_KERNEL);
+ if (!vsc8531->stats)
+ return -ENOMEM;
+
+ return vsc85xx_dt_led_modes_get(phydev, default_mode);
+}
+
static int vsc8584_probe(struct phy_device *phydev)
{
struct vsc8531_private *vsc8531;
@@ -1631,6 +1908,31 @@ static struct phy_driver vsc85xx_driver[] = {
.get_stats = &vsc85xx_get_stats,
},
{
+ .phy_id = PHY_ID_VSC8574,
+ .name = "Microsemi GE VSC8574 SyncE",
+ .phy_id_mask = 0xfffffff0,
+ .features = PHY_GBIT_FEATURES,
+ .flags = PHY_HAS_INTERRUPT,
+ .soft_reset = &genphy_soft_reset,
+ .config_init = &vsc8584_config_init,
+ .config_aneg = &vsc85xx_config_aneg,
+ .aneg_done = &genphy_aneg_done,
+ .read_status = &vsc85xx_read_status,
+ .ack_interrupt = &vsc85xx_ack_interrupt,
+ .config_intr = &vsc85xx_config_intr,
+ .did_interrupt = &vsc8584_did_interrupt,
+ .suspend = &genphy_suspend,
+ .resume = &genphy_resume,
+ .probe = &vsc8574_probe,
+ .set_wol = &vsc85xx_wol_set,
+ .get_wol = &vsc85xx_wol_get,
+ .get_tunable = &vsc85xx_get_tunable,
+ .set_tunable = &vsc85xx_set_tunable,
+ .get_sset_count = &vsc85xx_get_sset_count,
+ .get_strings = &vsc85xx_get_strings,
+ .get_stats = &vsc85xx_get_stats,
+},
+{
.phy_id = PHY_ID_VSC8584,
.name = "Microsemi GE VSC8584 SyncE",
.phy_id_mask = 0xfffffff0,
@@ -1663,6 +1965,7 @@ static struct mdio_device_id __maybe_unused vsc85xx_tbl[] = {
{ PHY_ID_VSC8531, 0xfffffff0, },
{ PHY_ID_VSC8540, 0xfffffff0, },
{ PHY_ID_VSC8541, 0xfffffff0, },
+ { PHY_ID_VSC8574, 0xfffffff0, },
{ PHY_ID_VSC8584, 0xfffffff0, },
{ }
};
--
git-series 0.9.1
PCB120 and PCB123 are both development boards based on Microsemi Ocelot
so let's use the same fitImage for both.
Signed-off-by: Quentin Schulz <[email protected]>
---
arch/mips/generic/Kconfig | 6 +--
arch/mips/generic/Platform | 2 +-
arch/mips/generic/board-ocelot.its.S | 40 ++++++++++++++++++++++-
arch/mips/generic/board-ocelot_pcb123.its.S | 23 +-------------
4 files changed, 44 insertions(+), 27 deletions(-)
create mode 100644 arch/mips/generic/board-ocelot.its.S
delete mode 100644 arch/mips/generic/board-ocelot_pcb123.its.S
diff --git a/arch/mips/generic/Kconfig b/arch/mips/generic/Kconfig
index 08e33c6..fd60198 100644
--- a/arch/mips/generic/Kconfig
+++ b/arch/mips/generic/Kconfig
@@ -65,11 +65,11 @@ config FIT_IMAGE_FDT_XILFPGA
Enable this to include the FDT for the MIPSfpga platform
from Imagination Technologies in the FIT kernel image.
-config FIT_IMAGE_FDT_OCELOT_PCB123
- bool "Include FDT for Microsemi Ocelot PCB123"
+config FIT_IMAGE_FDT_OCELOT
+ bool "Include FDT for Microsemi Ocelot development platforms"
select MSCC_OCELOT
help
- Enable this to include the FDT for the Ocelot PCB123 platform
+ Enable this to include the FDT for the Ocelot development platforms
from Microsemi in the FIT kernel image.
This requires u-boot on the platform.
diff --git a/arch/mips/generic/Platform b/arch/mips/generic/Platform
index 879cb80..eaa19d1 100644
--- a/arch/mips/generic/Platform
+++ b/arch/mips/generic/Platform
@@ -16,5 +16,5 @@ all-$(CONFIG_MIPS_GENERIC) := vmlinux.gz.itb
its-y := vmlinux.its.S
its-$(CONFIG_FIT_IMAGE_FDT_BOSTON) += board-boston.its.S
its-$(CONFIG_FIT_IMAGE_FDT_NI169445) += board-ni169445.its.S
-its-$(CONFIG_FIT_IMAGE_FDT_OCELOT_PCB123) += board-ocelot_pcb123.its.S
+its-$(CONFIG_FIT_IMAGE_FDT_OCELOT) += board-ocelot.its.S
its-$(CONFIG_FIT_IMAGE_FDT_XILFPGA) += board-xilfpga.its.S
diff --git a/arch/mips/generic/board-ocelot.its.S b/arch/mips/generic/board-ocelot.its.S
new file mode 100644
index 0000000..3da2398
--- /dev/null
+++ b/arch/mips/generic/board-ocelot.its.S
@@ -0,0 +1,40 @@
+/* SPDX-License-Identifier: (GPL-2.0 OR MIT) */
+/ {
+ images {
+ fdt@ocelot_pcb123 {
+ description = "MSCC Ocelot PCB123 Device Tree";
+ data = /incbin/("boot/dts/mscc/ocelot_pcb123.dtb");
+ type = "flat_dt";
+ arch = "mips";
+ compression = "none";
+ hash@0 {
+ algo = "sha1";
+ };
+ };
+
+ fdt@ocelot_pcb120 {
+ description = "MSCC Ocelot PCB120 Device Tree";
+ data = /incbin/("boot/dts/mscc/ocelot_pcb120.dtb");
+ type = "flat_dt";
+ arch = "mips";
+ compression = "none";
+ hash@0 {
+ algo = "sha1";
+ };
+ };
+ };
+
+ configurations {
+ conf@ocelot_pcb123 {
+ description = "Ocelot Linux kernel";
+ kernel = "kernel@0";
+ fdt = "fdt@ocelot_pcb123";
+ };
+
+ conf@ocelot_pcb120 {
+ description = "Ocelot Linux kernel";
+ kernel = "kernel@0";
+ fdt = "fdt@ocelot_pcb120";
+ };
+ };
+};
diff --git a/arch/mips/generic/board-ocelot_pcb123.its.S b/arch/mips/generic/board-ocelot_pcb123.its.S
deleted file mode 100644
index 5a7d5e1..0000000
--- a/arch/mips/generic/board-ocelot_pcb123.its.S
+++ /dev/null
@@ -1,23 +0,0 @@
-/* SPDX-License-Identifier: (GPL-2.0 OR MIT) */
-/ {
- images {
- fdt@ocelot_pcb123 {
- description = "MSCC Ocelot PCB123 Device Tree";
- data = /incbin/("boot/dts/mscc/ocelot_pcb123.dtb");
- type = "flat_dt";
- arch = "mips";
- compression = "none";
- hash@0 {
- algo = "sha1";
- };
- };
- };
-
- configurations {
- conf@ocelot_pcb123 {
- description = "Ocelot Linux kernel";
- kernel = "kernel@0";
- fdt = "fdt@ocelot_pcb123";
- };
- };
-};
--
git-series 0.9.1
The VSC8584 PHY is a 4-ports PHY that is 10/100/1000BASE-T, 100BASE-FX,
1000BASE-X and triple-speed copper SFP capable, can communicate with the
MAC via SGMII, QSGMII or 1000BASE-X, supports downshifting and can set
the blinking pattern of each of its 4 LEDs, supports hardware offloading
of MACsec and supports SyncE as well as HP Auto-MDIX detection.
This adds support for 10/100/1000BASE-T, SGMII/QSGMII link with the MAC,
downshifting, HP Auto-MDIX detection and blinking pattern for its 4
LEDs.
The VSC8584 has also an internal Intel 8051 microcontroller whose
firmware needs to be patched when the PHY is reset. If the 8051's
firmware has the expected CRC, its patching can be skipped. The
microcontroller can be accessed from any port of the PHY, though the CRC
function can only be done through the PHY that is the base PHY of the
package (internal address 0) due to a limitation of the firmware.
The GPIO register bank is a set of registers that are common to all PHYs
in the package. So any modification in any register of this bank affects
all PHYs of the package.
If the PHYs haven't been reset before booting the Linux kernel and were
configured to use interrupts for e.g. link status updates, it is
required to clear the interrupts mask register of all PHYs before being
able to use interrupts with any PHY. The first PHY of the package that
will be init will take care of clearing all PHYs interrupts mask
registers. Thus, we need to keep track of the init sequence in the
package, if it's already been done or if it's to be done.
Most of the init sequence of a PHY of the package is common to all PHYs
in the package, thus we use the SMI broadcast feature which enables us
to propagate a write in one register of one PHY to all PHYs in the
package.
The revA of the VSC8584 PHY (which is not and will not be publicly
released) should NOT patch the firmware of the microcontroller or it'll
make things worse, the easiest way is just to not support it.
Signed-off-by: Quentin Schulz <[email protected]>
---
drivers/net/phy/mscc.c | 714 ++++++++++++++++++++++++++++++++++++++++++-
1 file changed, 714 insertions(+)
diff --git a/drivers/net/phy/mscc.c b/drivers/net/phy/mscc.c
index 24f4754..b450489 100644
--- a/drivers/net/phy/mscc.c
+++ b/drivers/net/phy/mscc.c
@@ -6,6 +6,8 @@
* Copyright (c) 2016 Microsemi Corporation
*/
+#include <linux/firmware.h>
+#include <linux/jiffies.h>
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/mdio.h>
@@ -32,6 +34,10 @@ enum rgmii_rx_clock_delay {
#define DISABLE_HP_AUTO_MDIX_MASK 0x0080
#define DISABLE_PAIR_SWAP_CORR_MASK 0x0020
#define DISABLE_POLARITY_CORR_MASK 0x0010
+#define PARALLEL_DET_IGNORE_ADVERTISED 0x0008
+
+#define MSCC_PHY_EXT_CNTL_STATUS 22
+#define SMI_BROADCAST_WR_EN 0x0001
#define MSCC_PHY_ERR_RX_CNT 19
#define MSCC_PHY_ERR_FALSE_CARRIER_CNT 20
@@ -44,7 +50,20 @@ enum rgmii_rx_clock_delay {
#define MAC_IF_SELECTION_RMII 1
#define MAC_IF_SELECTION_RGMII 2
#define MAC_IF_SELECTION_POS 11
+#define VSC8584_MAC_IF_SELECTION_MASK 0x1000
+#define VSC8584_MAC_IF_SELECTION_SGMII 0
+#define VSC8584_MAC_IF_SELECTION_1000BASEX 1
+#define VSC8584_MAC_IF_SELECTION_POS 12
#define FAR_END_LOOPBACK_MODE_MASK 0x0008
+#define MEDIA_OP_MODE_MASK 0x0700
+#define MEDIA_OP_MODE_COPPER 0
+#define MEDIA_OP_MODE_SERDES 1
+#define MEDIA_OP_MODE_1000BASEX 2
+#define MEDIA_OP_MODE_100BASEFX 3
+#define MEDIA_OP_MODE_AMS_COPPER_SERDES 5
+#define MEDIA_OP_MODE_AMS_COPPER_1000BASEX 6
+#define MEDIA_OP_MODE_AMS_COPPER_100BASEFX 7
+#define MEDIA_OP_MODE_POS 8
#define MII_VSC85XX_INT_MASK 25
#define MII_VSC85XX_INT_MASK_MASK 0xa000
@@ -67,6 +86,13 @@ enum rgmii_rx_clock_delay {
#define MSCC_PHY_PAGE_STANDARD 0x0000 /* Standard registers */
#define MSCC_PHY_PAGE_EXTENDED 0x0001 /* Extended registers */
#define MSCC_PHY_PAGE_EXTENDED_2 0x0002 /* Extended reg - page 2 */
+#define MSCC_PHY_PAGE_EXTENDED_3 0x0003 /* Extended reg - page 3 */
+#define MSCC_PHY_PAGE_EXTENDED_4 0x0004 /* Extended reg - page 4 */
+/* Extended reg - GPIO; this is a bank of registers that are shared for all PHYs
+ * in the same package.
+ */
+#define MSCC_PHY_PAGE_EXTENDED_GPIO 0x0010 /* Extended reg - GPIO */
+#define MSCC_PHY_PAGE_TEST 0x2a30 /* Test reg */
#define MSCC_PHY_PAGE_TR 0x52b5 /* Token ring registers */
/* Extended Page 1 Registers */
@@ -79,13 +105,21 @@ enum rgmii_rx_clock_delay {
#define FORCE_MDI_CROSSOVER_MDI 0x0008
#define MSCC_PHY_ACTIPHY_CNTL 20
+#define PHY_ADDR_REVERSED 0x0200
#define DOWNSHIFT_CNTL_MASK 0x001C
#define DOWNSHIFT_EN 0x0010
#define DOWNSHIFT_CNTL_POS 2
#define MSCC_PHY_EXT_PHY_CNTL_4 23
+#define PHY_CNTL_4_ADDR_POS 11
+
+#define MSCC_PHY_VERIPHY_CNTL_2 25
+
+#define MSCC_PHY_VERIPHY_CNTL_3 26
/* Extended Page 2 Registers */
+#define MSCC_PHY_CU_PMD_TX_CNTL 16
+
#define MSCC_PHY_RGMII_CNTL 20
#define RGMII_RX_CLK_DELAY_MASK 0x0070
#define RGMII_RX_CLK_DELAY_POS 4
@@ -101,6 +135,70 @@ enum rgmii_rx_clock_delay {
#define SECURE_ON_ENABLE 0x8000
#define SECURE_ON_PASSWD_LEN_4 0x4000
+/* Extended Page 3 Registers */
+#define MSCC_PHY_SERDES_TX_VALID_CNT 21
+#define MSCC_PHY_SERDES_TX_CRC_ERR_CNT 22
+#define MSCC_PHY_SERDES_RX_VALID_CNT 28
+#define MSCC_PHY_SERDES_RX_CRC_ERR_CNT 29
+
+/* Extended page GPIO Registers */
+#define MSCC_DW8051_CNTL_STATUS 0
+#define MICRO_NSOFT_RESET 0x8000
+#define RUN_FROM_INT_ROM 0x4000
+#define AUTOINC_ADDR 0x2000
+#define PATCH_RAM_CLK 0x1000
+#define MICRO_PATCH_EN 0x0080
+#define DW8051_CLK_EN 0x0010
+#define MICRO_CLK_EN 0x0008
+#define MICRO_CLK_DIVIDE(x) ((x) >> 1)
+
+/* x Address in range 1-4 */
+#define MSCC_TRAP_ROM_ADDR(x) ((x) * 2 + 1)
+#define MSCC_PATCH_RAM_ADDR(x) (((x) + 1) * 2)
+#define MSCC_INT_MEM_ADDR 11
+
+#define MSCC_INT_MEM_CNTL 12
+#define READ_SFR 0x6000
+#define READ_PRAM 0x4000
+#define READ_ROM 0x2000
+#define READ_RAM 0x0000
+#define INT_MEM_WRITE_EN 0x1000
+#define EN_PATCH_RAM_TRAP_ADDR(x) (0x0100 << ((x) - 1))
+#define INT_MEM_DATA_M 0x00ff
+#define INT_MEM_DATA(x) (INT_MEM_DATA_M & (x))
+
+#define MSCC_PHY_PROC_CMD 18
+#define PROC_CMD_NCOMPLETED 0x8000
+#define PROC_CMD_FAILED 0x4000
+#define PROC_CMD_SGMII_PORT(x) ((x) << 8)
+#define PROC_CMD_FIBER_PORT(x) (0x0100 << (x) % 4)
+#define PROC_CMD_QSGMII_PORT 0x0c00
+#define PROC_CMD_RST_CONF_PORT 0x0080
+#define PROC_CMD_RECONF_PORT 0x0000
+#define PROC_CMD_READ_MOD_WRITE_PORT 0x0040
+#define PROC_CMD_WRITE 0x0040
+#define PROC_CMD_READ 0x0000
+#define PROC_CMD_FIBER_DISABLE 0x0020
+#define PROC_CMD_FIBER_100BASE_FX 0x0010
+#define PROC_CMD_FIBER_1000BASE_X 0x0000
+#define PROC_CMD_SGMII_MAC 0x0030
+#define PROC_CMD_QSGMII_MAC 0x0020
+#define PROC_CMD_NO_MAC_CONF 0x0000
+#define PROC_CMD_NOP 0x000f
+#define PROC_CMD_CRC16 0x0008
+#define PROC_CMD_FIBER_MEDIA_CONF 0x0001
+#define PROC_CMD_MCB_ACCESS_MAC_CONF 0x0000
+#define PROC_CMD_NCOMPLETED_TIMEOUT_MS 500
+
+#define MSCC_PHY_MAC_CFG_FASTLINK 19
+#define MAC_CFG_MASK 0xc000
+#define MAC_CFG_SGMII 0x0000
+#define MAC_CFG_QSGMII 0x4000
+
+/* Test page Registers */
+#define MSCC_PHY_TEST_PAGE_5 5
+#define MSCC_PHY_TEST_PAGE_8 8
+
/* Token ring page Registers */
#define MSCC_PHY_TR_CNTL 16
#define TR_WRITE 0x8000
@@ -113,6 +211,7 @@ enum rgmii_rx_clock_delay {
#define PHY_ID_VSC8531 0x00070570
#define PHY_ID_VSC8540 0x00070760
#define PHY_ID_VSC8541 0x00070770
+#define PHY_ID_VSC8584 0x000707c0
#define MSCC_VDDMAC_1500 1500
#define MSCC_VDDMAC_1800 1800
@@ -122,6 +221,24 @@ enum rgmii_rx_clock_delay {
#define DOWNSHIFT_COUNT_MAX 5
#define MAX_LEDS 4
+
+#define VSC8584_SUPP_LED_MODES (BIT(VSC8531_LINK_ACTIVITY) | \
+ BIT(VSC8531_LINK_1000_ACTIVITY) | \
+ BIT(VSC8531_LINK_100_ACTIVITY) | \
+ BIT(VSC8531_LINK_10_ACTIVITY) | \
+ BIT(VSC8531_LINK_100_1000_ACTIVITY) | \
+ BIT(VSC8531_LINK_10_1000_ACTIVITY) | \
+ BIT(VSC8531_LINK_10_100_ACTIVITY) | \
+ BIT(VSC8584_LINK_100FX_1000X_ACTIVITY) | \
+ BIT(VSC8531_DUPLEX_COLLISION) | \
+ BIT(VSC8531_COLLISION) | \
+ BIT(VSC8531_ACTIVITY) | \
+ BIT(VSC8584_100FX_1000X_ACTIVITY) | \
+ BIT(VSC8531_AUTONEG_FAULT) | \
+ BIT(VSC8531_SERIAL_MODE) | \
+ BIT(VSC8531_FORCE_LED_OFF) | \
+ BIT(VSC8531_FORCE_LED_ON))
+
#define VSC85XX_SUPP_LED_MODES (BIT(VSC8531_LINK_ACTIVITY) | \
BIT(VSC8531_LINK_1000_ACTIVITY) | \
BIT(VSC8531_LINK_100_ACTIVITY) | \
@@ -137,6 +254,13 @@ enum rgmii_rx_clock_delay {
BIT(VSC8531_FORCE_LED_OFF) | \
BIT(VSC8531_FORCE_LED_ON))
+#define MSCC_VSC8584_REVB_INT8051_FW "mscc_vsc8584_revb_int8051_fb48.bin"
+#define MSCC_VSC8584_REVB_INT8051_FW_START_ADDR 0xe800
+#define MSCC_VSC8584_REVB_INT8051_FW_CRC 0xfb48
+
+#define VSC8584_REVB 0x0001
+#define MSCC_DEV_REV_MASK GENMASK(3, 0)
+
struct vsc85xx_hw_stat {
const char *string;
u8 reg;
@@ -173,6 +297,55 @@ static struct vsc85xx_hw_stat vsc85xx_hw_stats[] = {
},
};
+static struct vsc85xx_hw_stat vsc8584_hw_stats[] = {
+ {
+ .string = "phy_receive_errors",
+ .reg = MSCC_PHY_ERR_RX_CNT,
+ .page = MSCC_PHY_PAGE_STANDARD,
+ .mask = ERR_CNT_MASK,
+ }, {
+ .string = "phy_false_carrier",
+ .reg = MSCC_PHY_ERR_FALSE_CARRIER_CNT,
+ .page = MSCC_PHY_PAGE_STANDARD,
+ .mask = ERR_CNT_MASK,
+ }, {
+ .string = "phy_cu_media_link_disconnect",
+ .reg = MSCC_PHY_ERR_LINK_DISCONNECT_CNT,
+ .page = MSCC_PHY_PAGE_STANDARD,
+ .mask = ERR_CNT_MASK,
+ }, {
+ .string = "phy_cu_media_crc_good_count",
+ .reg = MSCC_PHY_CU_MEDIA_CRC_VALID_CNT,
+ .page = MSCC_PHY_PAGE_EXTENDED,
+ .mask = VALID_CRC_CNT_CRC_MASK,
+ }, {
+ .string = "phy_cu_media_crc_error_count",
+ .reg = MSCC_PHY_EXT_PHY_CNTL_4,
+ .page = MSCC_PHY_PAGE_EXTENDED,
+ .mask = ERR_CNT_MASK,
+ }, {
+ .string = "phy_serdes_tx_good_pkt_count",
+ .reg = MSCC_PHY_SERDES_TX_VALID_CNT,
+ .page = MSCC_PHY_PAGE_EXTENDED_3,
+ .mask = VALID_CRC_CNT_CRC_MASK,
+ }, {
+ .string = "phy_serdes_tx_bad_crc_count",
+ .reg = MSCC_PHY_SERDES_TX_CRC_ERR_CNT,
+ .page = MSCC_PHY_PAGE_EXTENDED_3,
+ .mask = ERR_CNT_MASK,
+ }, {
+ .string = "phy_serdes_rx_good_pkt_count",
+ .reg = MSCC_PHY_SERDES_RX_VALID_CNT,
+ .page = MSCC_PHY_PAGE_EXTENDED_3,
+ .mask = VALID_CRC_CNT_CRC_MASK,
+ }, {
+ .string = "phy_serdes_rx_bad_crc_count",
+ .reg = MSCC_PHY_SERDES_RX_CRC_ERR_CNT,
+ .page = MSCC_PHY_PAGE_EXTENDED_3,
+ .mask = ERR_CNT_MASK,
+ },
+};
+
struct vsc8531_private {
int rate_magic;
u16 supp_led_modes;
@@ -181,6 +354,7 @@ struct vsc8531_private {
struct vsc85xx_hw_stat *hw_stats;
u64 *stats;
int nstats;
+ bool pkg_init;
};
#ifdef CONFIG_OF_MDIO
@@ -730,6 +904,481 @@ static int vsc85xx_eee_init_seq_set(struct phy_device *phydev)
return rc;
}
+/* bus->mdio_lock should be locked when using this function */
+static void vsc8584_csr_write(struct mii_bus *bus, int phy, u16 addr, u32 val)
+{
+ __mdiobus_write(bus, phy, MSCC_PHY_TR_MSB, val >> 16);
+ __mdiobus_write(bus, phy, MSCC_PHY_TR_LSB, val & GENMASK(15, 0));
+ __mdiobus_write(bus, phy, MSCC_PHY_TR_CNTL, TR_WRITE | TR_ADDR(addr));
+}
+
+/* bus->mdio_lock should be locked when using this function */
+static int vsc8584_cmd(struct mii_bus *bus, int phy, u16 val)
+{
+ unsigned long deadline;
+ u16 reg_val;
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS,
+ MSCC_PHY_PAGE_EXTENDED_GPIO);
+
+ __mdiobus_write(bus, phy, MSCC_PHY_PROC_CMD, PROC_CMD_NCOMPLETED | val);
+
+ deadline = jiffies + msecs_to_jiffies(PROC_CMD_NCOMPLETED_TIMEOUT_MS);
+ do {
+ reg_val = __mdiobus_read(bus, phy, MSCC_PHY_PROC_CMD);
+ } while (time_before(jiffies, deadline) &&
+ (reg_val & PROC_CMD_NCOMPLETED) &&
+ !(reg_val & PROC_CMD_FAILED));
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_STANDARD);
+
+ if (reg_val & PROC_CMD_FAILED)
+ return -EIO;
+
+ if (reg_val & PROC_CMD_NCOMPLETED)
+ return -ETIMEDOUT;
+
+ return 0;
+}
+
+/* bus->mdio_lock should be locked when using this function */
+static int vsc8584_micro_deassert_reset(struct mii_bus *bus, int phy,
+ bool patch_en)
+{
+ u32 enable, release;
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS,
+ MSCC_PHY_PAGE_EXTENDED_GPIO);
+
+ enable = RUN_FROM_INT_ROM | MICRO_CLK_EN | DW8051_CLK_EN;
+ release = MICRO_NSOFT_RESET | RUN_FROM_INT_ROM | DW8051_CLK_EN |
+ MICRO_CLK_EN;
+
+ if (patch_en) {
+ enable |= MICRO_PATCH_EN;
+ release |= MICRO_PATCH_EN;
+
+ /* Clear all patches */
+ __mdiobus_write(bus, phy, MSCC_INT_MEM_CNTL, READ_RAM);
+ }
+
+ /* Enable 8051 Micro clock; CLEAR/SET patch present; disable PRAM clock
+ * override and addr. auto-incr; operate at 125 MHz
+ */
+ __mdiobus_write(bus, phy, MSCC_DW8051_CNTL_STATUS, enable);
+ /* Release 8051 Micro SW reset */
+ __mdiobus_write(bus, phy, MSCC_DW8051_CNTL_STATUS, release);
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_STANDARD);
+
+ return 0;
+}
+
+/* bus->mdio_lock should be locked when using this function */
+static int vsc8584_micro_assert_reset(struct mii_bus *bus, int phy)
+{
+ int ret;
+ u16 reg;
+
+ ret = vsc8584_cmd(bus, phy, PROC_CMD_NOP);
+ if (ret)
+ return ret;
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS,
+ MSCC_PHY_PAGE_EXTENDED_GPIO);
+
+ reg = __mdiobus_read(bus, phy, MSCC_INT_MEM_CNTL);
+ reg &= ~EN_PATCH_RAM_TRAP_ADDR(4);
+ __mdiobus_write(bus, phy, MSCC_INT_MEM_CNTL, reg);
+
+ __mdiobus_write(bus, phy, MSCC_TRAP_ROM_ADDR(4), 0x005b);
+ __mdiobus_write(bus, phy, MSCC_PATCH_RAM_ADDR(4), 0x005b);
+
+ reg = __mdiobus_read(bus, phy, MSCC_INT_MEM_CNTL);
+ reg |= EN_PATCH_RAM_TRAP_ADDR(4);
+ __mdiobus_write(bus, phy, MSCC_INT_MEM_CNTL, reg);
+
+ __mdiobus_write(bus, phy, MSCC_PHY_PROC_CMD, PROC_CMD_NOP);
+
+ reg = __mdiobus_read(bus, phy, MSCC_DW8051_CNTL_STATUS);
+ reg &= ~MICRO_NSOFT_RESET;
+ __mdiobus_write(bus, phy, MSCC_DW8051_CNTL_STATUS, reg);
+
+ __mdiobus_write(bus, phy, MSCC_PHY_PROC_CMD,
+ PROC_CMD_MCB_ACCESS_MAC_CONF |
+ PROC_CMD_SGMII_PORT(0) | PROC_CMD_NO_MAC_CONF |
+ PROC_CMD_READ);
+
+ reg = __mdiobus_read(bus, phy, MSCC_INT_MEM_CNTL);
+ reg &= ~EN_PATCH_RAM_TRAP_ADDR(4);
+ __mdiobus_write(bus, phy, MSCC_INT_MEM_CNTL, reg);
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_STANDARD);
+
+ return 0;
+}
+
+/* bus->mdio_lock should be locked when using this function */
+static int vsc8584_get_fw_crc(struct mii_bus *bus, int phy, u16 start,
+ u16 size, u16 *crc)
+{
+ int ret;
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_EXTENDED);
+
+ __mdiobus_write(bus, phy, MSCC_PHY_VERIPHY_CNTL_2, start);
+ __mdiobus_write(bus, phy, MSCC_PHY_VERIPHY_CNTL_3, size);
+
+ /* Start Micro command */
+ ret = vsc8584_cmd(bus, phy, PROC_CMD_CRC16);
+ if (ret)
+ goto out;
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_EXTENDED);
+
+ *crc = __mdiobus_read(bus, phy, MSCC_PHY_VERIPHY_CNTL_2);
+
+out:
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_STANDARD);
+
+ return ret;
+}
+
+/* bus->mdio_lock should be locked when using this function */
+static int vsc8584_patch_fw(struct mii_bus *bus, int phy,
+ const struct firmware *fw)
+{
+ int i, ret;
+
+ ret = vsc8584_micro_assert_reset(bus, phy);
+ if (ret) {
+ dev_err(&bus->mdio_map[phy]->dev,
+ "%s: failed to assert reset of micro\n", __func__);
+ return ret;
+ }
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS,
+ MSCC_PHY_PAGE_EXTENDED_GPIO);
+
+ /* Hold 8051 Micro in SW Reset, Enable auto incr address and patch clock
+ * Disable the 8051 Micro clock
+ */
+ __mdiobus_write(bus, phy, MSCC_DW8051_CNTL_STATUS, RUN_FROM_INT_ROM |
+ AUTOINC_ADDR | PATCH_RAM_CLK | MICRO_CLK_EN |
+ MICRO_CLK_DIVIDE(2));
+ __mdiobus_write(bus, phy, MSCC_INT_MEM_CNTL, READ_PRAM |
+ INT_MEM_WRITE_EN | INT_MEM_DATA(2));
+ __mdiobus_write(bus, phy, MSCC_INT_MEM_ADDR, 0x0000);
+
+ for (i = 0; i < fw->size; i++)
+ __mdiobus_write(bus, phy, MSCC_INT_MEM_CNTL, READ_PRAM |
+ INT_MEM_WRITE_EN | fw->data[i]);
+
+ /* Clear internal memory access */
+ __mdiobus_write(bus, phy, MSCC_INT_MEM_CNTL, READ_RAM);
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_STANDARD);
+
+ return 0;
+}
+
+/* bus->mdio_lock should be locked when using this function */
+static int vsc8584_config_pre_init(struct mii_bus *bus, int phy)
+{
+ struct device *dev = &bus->mdio_map[phy]->dev;
+ const struct firmware *fw;
+ u16 crc, reg;
+ int ret;
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_STANDARD);
+
+ /* all writes below this line are broadcasted to all PHYs */
+ reg = __mdiobus_read(bus, phy, MSCC_PHY_EXT_CNTL_STATUS);
+ reg |= SMI_BROADCAST_WR_EN;
+ __mdiobus_write(bus, phy, MSCC_PHY_EXT_CNTL_STATUS, reg);
+
+ __mdiobus_write(bus, phy, MII_VSC85XX_INT_MASK, 0);
+
+ reg = __mdiobus_read(bus, phy, MSCC_PHY_BYPASS_CONTROL);
+ reg |= PARALLEL_DET_IGNORE_ADVERTISED;
+ __mdiobus_write(bus, phy, MSCC_PHY_BYPASS_CONTROL, reg);
+
+ /* The below register writes are tweaking analog and electrical
+ * configuration that were determined through characterization by PHY
+ * engineers. These don't mean anything more than "these are the best
+ * values".
+ */
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS,
+ MSCC_PHY_PAGE_EXTENDED_3);
+
+ __mdiobus_write(bus, phy, MSCC_PHY_SERDES_TX_CRC_ERR_CNT, 0x2000);
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_TEST);
+
+ __mdiobus_write(bus, phy, MSCC_PHY_TEST_PAGE_5, 0x1f20);
+
+ reg = __mdiobus_read(bus, phy, MSCC_PHY_TEST_PAGE_8);
+ reg |= 0x8000;
+ __mdiobus_write(bus, phy, MSCC_PHY_TEST_PAGE_8, reg);
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_TR);
+
+ __mdiobus_write(bus, phy, MSCC_PHY_TR_CNTL, TR_WRITE | TR_ADDR(0x2fa4));
+
+ reg = __mdiobus_read(bus, phy, MSCC_PHY_TR_MSB);
+ reg &= ~0x007f;
+ reg |= 0x0019;
+ __mdiobus_write(bus, phy, MSCC_PHY_TR_MSB, reg);
+
+ __mdiobus_write(bus, phy, MSCC_PHY_TR_CNTL, TR_WRITE | TR_ADDR(0x0fa4));
+
+ vsc8584_csr_write(bus, phy, 0x07fa, 0x0050100f);
+ vsc8584_csr_write(bus, phy, 0x1688, 0x00049f81);
+ vsc8584_csr_write(bus, phy, 0x0f90, 0x00688980);
+ vsc8584_csr_write(bus, phy, 0x03a4, 0x0000d8f0);
+ vsc8584_csr_write(bus, phy, 0x0fc0, 0x00000400);
+ vsc8584_csr_write(bus, phy, 0x0f82, 0x0012b002);
+ vsc8584_csr_write(bus, phy, 0x1686, 0x00000004);
+ vsc8584_csr_write(bus, phy, 0x168c, 0x00d2c46f);
+ vsc8584_csr_write(bus, phy, 0x17a2, 0x00000620);
+ vsc8584_csr_write(bus, phy, 0x16a0, 0x00eeffdd);
+ vsc8584_csr_write(bus, phy, 0x16a6, 0x00071448);
+ vsc8584_csr_write(bus, phy, 0x16a4, 0x0013132f);
+ vsc8584_csr_write(bus, phy, 0x16a8, 0x00000000);
+ vsc8584_csr_write(bus, phy, 0x0ffc, 0x00c0a028);
+ vsc8584_csr_write(bus, phy, 0x0fe8, 0x0091b06c);
+ vsc8584_csr_write(bus, phy, 0x0fea, 0x00041600);
+ vsc8584_csr_write(bus, phy, 0x0f80, 0x00fffaff);
+ vsc8584_csr_write(bus, phy, 0x0fec, 0x00901809);
+ vsc8584_csr_write(bus, phy, 0x0ffe, 0x00b01007);
+ vsc8584_csr_write(bus, phy, 0x16b0, 0x00eeff00);
+ vsc8584_csr_write(bus, phy, 0x16b2, 0x00007000);
+ vsc8584_csr_write(bus, phy, 0x16b4, 0x00000814);
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS,
+ MSCC_PHY_PAGE_EXTENDED_2);
+
+ __mdiobus_write(bus, phy, MSCC_PHY_CU_PMD_TX_CNTL, 0x028e);
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_TR);
+
+ vsc8584_csr_write(bus, phy, 0x0486, 0x0008a518);
+ vsc8584_csr_write(bus, phy, 0x0488, 0x006dc696);
+ vsc8584_csr_write(bus, phy, 0x048a, 0x00000912);
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_TEST);
+
+ reg = __mdiobus_read(bus, phy, MSCC_PHY_TEST_PAGE_8);
+ reg &= ~0x8000;
+ __mdiobus_write(bus, phy, MSCC_PHY_TEST_PAGE_8, reg);
+
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_STANDARD);
+
+ /* end of write broadcasting */
+ reg = __mdiobus_read(bus, phy, MSCC_PHY_EXT_CNTL_STATUS);
+ reg &= ~SMI_BROADCAST_WR_EN;
+ __mdiobus_write(bus, phy, MSCC_PHY_EXT_CNTL_STATUS, reg);
+
+ ret = request_firmware(&fw, MSCC_VSC8584_REVB_INT8051_FW, dev);
+ if (ret) {
+ dev_err(dev, "failed to load firmware %s, ret: %d\n",
+ MSCC_VSC8584_REVB_INT8051_FW, ret);
+ return ret;
+ }
+
+ /* Add one byte to size for the one added by the patch_fw function */
+ ret = vsc8584_get_fw_crc(bus, phy,
+ MSCC_VSC8584_REVB_INT8051_FW_START_ADDR,
+ fw->size + 1, &crc);
+ if (ret)
+ goto out;
+
+ if (crc != MSCC_VSC8584_REVB_INT8051_FW_CRC) {
+ dev_dbg(dev, "FW CRC is not the expected one, patching FW\n");
+ if (vsc8584_patch_fw(bus, phy, fw))
+ dev_warn(dev,
+ "failed to patch FW, expect non-optimal device\n");
+ }
+
+ vsc8584_micro_deassert_reset(bus, phy, false);
+
+ /* Add one byte to size for the one added by the patch_fw function */
+ ret = vsc8584_get_fw_crc(bus, phy,
+ MSCC_VSC8584_REVB_INT8051_FW_START_ADDR,
+ fw->size + 1, &crc);
+ if (ret)
+ goto out;
+
+ if (crc != MSCC_VSC8584_REVB_INT8051_FW_CRC)
+ dev_warn(dev,
+ "FW CRC after patching is not the expected one, expect non-optimal device\n");
+
+ ret = vsc8584_micro_assert_reset(bus, phy);
+ if (ret)
+ goto out;
+
+ vsc8584_micro_deassert_reset(bus, phy, true);
+
+out:
+ __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_STANDARD);
+
+ release_firmware(fw);
+
+ return ret;
+}
+
+/* Check if one PHY has already done the init of the parts common to all PHYs
+ * in the Quad PHY package.
+ */
+static bool vsc8584_is_pkg_init(struct phy_device *phydev, int phy0,
+ bool reversed)
+{
+ struct mdio_device **map = phydev->mdio.bus->mdio_map;
+ struct vsc8531_private *vsc8531;
+ struct phy_device *phy;
+ int i, addr;
+
+ /* VSC8584 is a Quad PHY */
+ for (i = 0; i < 4; i++) {
+ if (reversed)
+ addr = phy0 - i;
+ else
+ addr = phy0 + i;
+
+ phy = container_of(map[addr], struct phy_device, mdio);
+
+ if ((phy->phy_id & phydev->drv->phy_id_mask) !=
+ (phydev->drv->phy_id & phydev->drv->phy_id_mask))
+ continue;
+
+ vsc8531 = phy->priv;
+
+ if (vsc8531 && vsc8531->pkg_init)
+ return true;
+ }
+
+ return false;
+}
+
+static int vsc8584_config_init(struct phy_device *phydev)
+{
+ struct vsc8531_private *vsc8531 = phydev->priv;
+ u16 addr, val, base_addr;
+ int ret, i;
+
+ phydev->mdix_ctrl = ETH_TP_MDI_AUTO;
+
+ mutex_lock(&phydev->mdio.bus->mdio_lock);
+
+ __mdiobus_write(phydev->mdio.bus, phydev->mdio.addr,
+ MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_EXTENDED);
+ addr = __mdiobus_read(phydev->mdio.bus, phydev->mdio.addr,
+ MSCC_PHY_EXT_PHY_CNTL_4);
+ addr >>= PHY_CNTL_4_ADDR_POS;
+
+ val = __mdiobus_read(phydev->mdio.bus, phydev->mdio.addr,
+ MSCC_PHY_ACTIPHY_CNTL);
+ if (val & PHY_ADDR_REVERSED)
+ base_addr = phydev->mdio.addr + addr;
+ else
+ base_addr = phydev->mdio.addr - addr;
+
+ /* Some parts of the init sequence are identical for every PHY in the
+ * package. Some parts are modifying the GPIO register bank which is a
+ * set of registers that are affecting all PHYs, a few resetting the
+ * microprocessor common to all PHYs. The CRC check responsible of the
+ * checking the firmware within the 8051 microprocessor can only be
+ * accessed via the PHY whose internal address in the package is 0.
+ * All PHYs' interrupts mask register has to be zeroed before enabling
+ * any PHY's interrupt in this register.
+ * For all these reasons, we need to do the init sequence once and only
+ * once whatever is the first PHY in the package that is initialized and
+ * do the correct init sequence for all PHYs that are package-critical
+ * in this pre-init function.
+ */
+ if (!vsc8584_is_pkg_init(phydev, base_addr,
+ val & PHY_ADDR_REVERSED ? 1 : 0)) {
+ ret = vsc8584_config_pre_init(phydev->mdio.bus, base_addr);
+ if (ret)
+ goto err;
+ }
+
+ vsc8531->pkg_init = true;
+
+ __mdiobus_write(phydev->mdio.bus, base_addr, MSCC_EXT_PAGE_ACCESS,
+ MSCC_PHY_PAGE_EXTENDED_GPIO);
+
+ val = __mdiobus_read(phydev->mdio.bus, base_addr,
+ MSCC_PHY_MAC_CFG_FASTLINK);
+ val &= ~MAC_CFG_MASK;
+ if (phydev->interface == PHY_INTERFACE_MODE_QSGMII)
+ val |= MAC_CFG_QSGMII;
+ else
+ val |= MAC_CFG_SGMII;
+
+ ret = __mdiobus_write(phydev->mdio.bus, base_addr,
+ MSCC_PHY_MAC_CFG_FASTLINK, val);
+ if (ret)
+ goto err;
+
+ val = PROC_CMD_MCB_ACCESS_MAC_CONF | PROC_CMD_RST_CONF_PORT |
+ PROC_CMD_READ_MOD_WRITE_PORT;
+ if (phydev->interface == PHY_INTERFACE_MODE_QSGMII)
+ val |= PROC_CMD_QSGMII_MAC;
+ else
+ val |= PROC_CMD_SGMII_MAC;
+
+ ret = vsc8584_cmd(phydev->mdio.bus, base_addr, val);
+ if (ret)
+ goto err;
+
+ usleep_range(10000, 20000);
+
+ /* Disable SerDes for 100Base-FX */
+ ret = vsc8584_cmd(phydev->mdio.bus, base_addr,
+ PROC_CMD_FIBER_MEDIA_CONF |
+ PROC_CMD_FIBER_PORT(addr) | PROC_CMD_FIBER_DISABLE |
+ PROC_CMD_READ_MOD_WRITE_PORT |
+ PROC_CMD_RST_CONF_PORT | PROC_CMD_FIBER_100BASE_FX);
+ if (ret)
+ goto err;
+
+ /* Disable SerDes for 1000Base-X */
+ ret = vsc8584_cmd(phydev->mdio.bus, base_addr,
+ PROC_CMD_FIBER_MEDIA_CONF |
+ PROC_CMD_FIBER_PORT(addr) | PROC_CMD_FIBER_DISABLE |
+ PROC_CMD_READ_MOD_WRITE_PORT |
+ PROC_CMD_RST_CONF_PORT | PROC_CMD_FIBER_1000BASE_X);
+ if (ret)
+ goto err;
+
+ mutex_unlock(&phydev->mdio.bus->mdio_lock);
+
+ phy_write(phydev, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_STANDARD);
+
+ val = phy_read(phydev, MSCC_PHY_EXT_PHY_CNTL_1);
+ val &= ~(MEDIA_OP_MODE_MASK | VSC8584_MAC_IF_SELECTION_MASK);
+ val |= MEDIA_OP_MODE_COPPER | (VSC8584_MAC_IF_SELECTION_SGMII <<
+ VSC8584_MAC_IF_SELECTION_POS);
+ ret = phy_write(phydev, MSCC_PHY_EXT_PHY_CNTL_1, val);
+
+ ret = genphy_soft_reset(phydev);
+ if (ret)
+ return ret;
+
+ for (i = 0; i < vsc8531->nleds; i++) {
+ ret = vsc85xx_led_cntl_set(phydev, i, vsc8531->leds_mode[i]);
+ if (ret)
+ return ret;
+ }
+
+ return genphy_config_init(phydev);
+
+err:
+ mutex_unlock(&phydev->mdio.bus->mdio_lock);
+ return ret;
+}
+
static int vsc85xx_config_init(struct phy_device *phydev)
{
int rc, i;
@@ -760,6 +1409,16 @@ static int vsc85xx_config_init(struct phy_device *phydev)
return genphy_config_init(phydev);
}
+static int vsc8584_did_interrupt(struct phy_device *phydev)
+{
+ int rc = 0;
+
+ if (phydev->interrupts == PHY_INTERRUPT_ENABLED)
+ rc = phy_read(phydev, MII_VSC85XX_INT_STATUS);
+
+ return (rc < 0) ? 0 : rc & MII_VSC85XX_INT_MASK_MASK;
+}
+
static int vsc85xx_ack_interrupt(struct phy_device *phydev)
{
int rc = 0;
@@ -809,6 +1468,37 @@ static int vsc85xx_read_status(struct phy_device *phydev)
return genphy_read_status(phydev);
}
+static int vsc8584_probe(struct phy_device *phydev)
+{
+ struct vsc8531_private *vsc8531;
+ u32 default_mode[4] = {VSC8531_LINK_1000_ACTIVITY,
+ VSC8531_LINK_100_ACTIVITY, VSC8531_LINK_ACTIVITY,
+ VSC8531_DUPLEX_COLLISION};
+
+ if ((phydev->phy_id & MSCC_DEV_REV_MASK) != VSC8584_REVB) {
+ dev_err(&phydev->mdio.dev, "Only VSC8584 revB is supported.\n");
+ return -ENOTSUPP;
+ }
+
+ vsc8531 = devm_kzalloc(&phydev->mdio.dev, sizeof(*vsc8531), GFP_KERNEL);
+ if (!vsc8531)
+ return -ENOMEM;
+
+ phydev->priv = vsc8531;
+
+ vsc8531->nleds = 4;
+ vsc8531->supp_led_modes = VSC8584_SUPP_LED_MODES;
+ vsc8531->hw_stats = vsc8584_hw_stats;
+ vsc8531->nstats = ARRAY_SIZE(vsc8584_hw_stats);
+ vsc8531->stats = devm_kzalloc(&phydev->mdio.dev,
+ sizeof(u64) * vsc8531->nstats,
+ GFP_KERNEL);
+ if (!vsc8531->stats)
+ return -ENOMEM;
+
+ return vsc85xx_dt_led_modes_get(phydev, default_mode);
+}
+
static int vsc85xx_probe(struct phy_device *phydev)
{
struct vsc8531_private *vsc8531;
@@ -937,6 +1627,29 @@ static struct phy_driver vsc85xx_driver[] = {
.get_sset_count = &vsc85xx_get_sset_count,
.get_strings = &vsc85xx_get_strings,
.get_stats = &vsc85xx_get_stats,
+},
+{
+ .phy_id = PHY_ID_VSC8584,
+ .name = "Microsemi GE VSC8584 SyncE",
+ .phy_id_mask = 0xfffffff0,
+ .features = PHY_GBIT_FEATURES,
+ .flags = PHY_HAS_INTERRUPT,
+ .soft_reset = &genphy_soft_reset,
+ .config_init = &vsc8584_config_init,
+ .config_aneg = &vsc85xx_config_aneg,
+ .aneg_done = &genphy_aneg_done,
+ .read_status = &vsc85xx_read_status,
+ .ack_interrupt = &vsc85xx_ack_interrupt,
+ .config_intr = &vsc85xx_config_intr,
+ .did_interrupt = &vsc8584_did_interrupt,
+ .suspend = &genphy_suspend,
+ .resume = &genphy_resume,
+ .probe = &vsc8584_probe,
+ .get_tunable = &vsc85xx_get_tunable,
+ .set_tunable = &vsc85xx_set_tunable,
+ .get_sset_count = &vsc85xx_get_sset_count,
+ .get_strings = &vsc85xx_get_strings,
+ .get_stats = &vsc85xx_get_stats,
}
};
@@ -948,6 +1661,7 @@ static struct mdio_device_id __maybe_unused vsc85xx_tbl[] = {
{ PHY_ID_VSC8531, 0xfffffff0, },
{ PHY_ID_VSC8540, 0xfffffff0, },
{ PHY_ID_VSC8541, 0xfffffff0, },
+ { PHY_ID_VSC8584, 0xfffffff0, },
{ }
};
--
git-series 0.9.1
On Fri, Sep 14, 2018 at 11:44:22AM +0200, Quentin Schulz wrote:
> The VSC8584 (and most likely other PHYs in the same generation) has two
> additional LED modes that can be picked, so let's add them.
>
> Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Andrew Lunn <[email protected]>
Andrew
> Most of the init sequence of a PHY of the package is common to all PHYs
> in the package, thus we use the SMI broadcast feature which enables us
> to propagate a write in one register of one PHY to all PHYs in the
> package.
Hi Quinten
Could you say a bit more about the broadcast. Does the SMI broadcast
go to all PHY everywhere on an MDIO bus, or only all PHYs within one
package? I'm just thinking about the case you need two of these
packages to cover 8 switch ports.
Thanks
Andrew
Hi Andrew,
On Fri, Sep 14, 2018 at 03:18:46PM +0200, Andrew Lunn wrote:
> > Most of the init sequence of a PHY of the package is common to all PHYs
> > in the package, thus we use the SMI broadcast feature which enables us
> > to propagate a write in one register of one PHY to all PHYs in the
> > package.
>
> Hi Quinten
>
> Could you say a bit more about the broadcast. Does the SMI broadcast
> go to all PHY everywhere on an MDIO bus, or only all PHYs within one
> package? I'm just thinking about the case you need two of these
> packages to cover 8 switch ports.
>
Ah sorry, that wasn't very explicit. That's a feature on the PHY side so
my wildest guess is that it wouldn't impact any other PHY outside of
this package. Affecting any other PHY on the bus is counter-intuitive to
me but I'll ask the HW engineers for confirmation.
Thanks,
Quentin
Hi,
On 14/09/2018 11:44:26+0200, Quentin Schulz wrote:
> In order to use GPIO4 as a GPIO, we need to mux it in this mode so let's
> declare a new pinctrl DT node for it.
>
> Signed-off-by: Quentin Schulz <[email protected]>
> ---
> arch/mips/boot/dts/mscc/ocelot.dtsi | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/arch/mips/boot/dts/mscc/ocelot.dtsi b/arch/mips/boot/dts/mscc/ocelot.dtsi
> index 8ce317c..b5c4c74 100644
> --- a/arch/mips/boot/dts/mscc/ocelot.dtsi
> +++ b/arch/mips/boot/dts/mscc/ocelot.dtsi
> @@ -182,6 +182,11 @@
> interrupts = <13>;
> #interrupt-cells = <2>;
>
> + gpio4: gpio4 {
> + pins = "GPIO_4";
> + function = "gpio";
> + };
> +
For a GPIO, I would do that in the board dts because it is not used
directly in the dtsi.
> uart_pins: uart-pins {
> pins = "GPIO_6", "GPIO_7";
> function = "uart";
> --
> git-series 0.9.1
--
Alexandre Belloni, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
On 14/09/2018 11:44:27+0200, Quentin Schulz wrote:
> The Ocelot PCB120 evaluation board is different from the PCB123 in that
> it has 4 external VSC8584 (or VSC8574) PHYs.
>
> It uses the SoC's second MDIO bus for external PHYs which have a
> reversed address on the bus (i.e. PHY4 is on address 3, PHY5 is on
> address 2, PHY6 on 1 and PHY7 on 0).
>
> Here is how the PHYs are connected to the switch ports:
> port 0: phy0 (internal)
> port 1: phy1 (internal)
> port 2: phy2 (internal)
> port 3: phy3 (internal)
> port 4: phy7
> port 5: phy4
> port 6: phy6
> port 9: phy5
>
> Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Alexandre Belloni <[email protected]>
> ---
> arch/mips/boot/dts/mscc/Makefile | 2 +-
> arch/mips/boot/dts/mscc/ocelot_pcb120.dts | 100 +++++++++++++++++++++++-
> 2 files changed, 101 insertions(+), 1 deletion(-)
> create mode 100644 arch/mips/boot/dts/mscc/ocelot_pcb120.dts
>
> diff --git a/arch/mips/boot/dts/mscc/Makefile b/arch/mips/boot/dts/mscc/Makefile
> index 9a9bb7e..ec6f5b2 100644
> --- a/arch/mips/boot/dts/mscc/Makefile
> +++ b/arch/mips/boot/dts/mscc/Makefile
> @@ -1,3 +1,3 @@
> -dtb-$(CONFIG_MSCC_OCELOT) += ocelot_pcb123.dtb
> +dtb-$(CONFIG_MSCC_OCELOT) += ocelot_pcb123.dtb ocelot_pcb120.dtb
>
> obj-$(CONFIG_BUILTIN_DTB) += $(addsuffix .o, $(dtb-y))
> diff --git a/arch/mips/boot/dts/mscc/ocelot_pcb120.dts b/arch/mips/boot/dts/mscc/ocelot_pcb120.dts
> new file mode 100644
> index 0000000..8eb03a5
> --- /dev/null
> +++ b/arch/mips/boot/dts/mscc/ocelot_pcb120.dts
> @@ -0,0 +1,100 @@
> +// SPDX-License-Identifier: (GPL-2.0 OR MIT)
> +/* Copyright (c) 2017 Microsemi Corporation */
> +
> +/dts-v1/;
> +
> +#include <dt-bindings/interrupt-controller/irq.h>
> +#include <dt-bindings/phy/phy-ocelot-serdes.h>
> +#include "ocelot.dtsi"
> +
> +/ {
> + compatible = "mscc,ocelot-pcb120", "mscc,ocelot";
> +
> + chosen {
> + stdout-path = "serial0:115200n8";
> + };
> +
> + memory@0 {
> + device_type = "memory";
> + reg = <0x0 0x0e000000>;
> + };
> +};
> +
> +&mdio0 {
> + status = "okay";
> +};
> +
> +&mdio1 {
> + status = "okay";
> + pinctrl-names = "default";
> + pinctrl-0 = <&miim1>, <&gpio4>;
> +
> + phy7: ethernet-phy@0 {
> + reg = <0>;
> + interrupts = <4 IRQ_TYPE_LEVEL_HIGH>;
> + interrupt-parent = <&gpio>;
> + };
> + phy6: ethernet-phy@1 {
> + reg = <1>;
> + interrupts = <4 IRQ_TYPE_LEVEL_HIGH>;
> + interrupt-parent = <&gpio>;
> + };
> + phy5: ethernet-phy@2 {
> + reg = <2>;
> + interrupts = <4 IRQ_TYPE_LEVEL_HIGH>;
> + interrupt-parent = <&gpio>;
> + };
> + phy4: ethernet-phy@3 {
> + reg = <3>;
> + interrupts = <4 IRQ_TYPE_LEVEL_HIGH>;
> + interrupt-parent = <&gpio>;
> + };
> +};
> +
> +&port0 {
> + phy-handle = <&phy0>;
> +};
> +
> +&port1 {
> + phy-handle = <&phy1>;
> +};
> +
> +&port2 {
> + phy-handle = <&phy2>;
> +};
> +
> +&port3 {
> + phy-handle = <&phy3>;
> +};
> +
> +&port4 {
> + phy-handle = <&phy7>;
> + phy-mode = "sgmii";
> + phys = <&serdes 4 SERDES1G_2>;
> +};
> +
> +&port5 {
> + phy-handle = <&phy4>;
> + phy-mode = "sgmii";
> + phys = <&serdes 5 SERDES1G_5>;
> +};
> +
> +&port6 {
> + phy-handle = <&phy6>;
> + phy-mode = "sgmii";
> + phys = <&serdes 6 SERDES1G_3>;
> +};
> +
> +&port9 {
> + phy-handle = <&phy5>;
> + phy-mode = "sgmii";
> + phys = <&serdes 9 SERDES1G_4>;
> +};
> +
> +&uart0 {
> + status = "okay";
> +};
> +
> +&uart2 {
> + status = "okay";
> +};
> --
> git-series 0.9.1
--
Alexandre Belloni, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
On 14/09/2018 11:44:28+0200, Quentin Schulz wrote:
> PCB120 and PCB123 are both development boards based on Microsemi Ocelot
> so let's use the same fitImage for both.
>
> Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Alexandre Belloni <[email protected]>
> ---
> arch/mips/generic/Kconfig | 6 +--
> arch/mips/generic/Platform | 2 +-
> arch/mips/generic/board-ocelot.its.S | 40 ++++++++++++++++++++++-
> arch/mips/generic/board-ocelot_pcb123.its.S | 23 +-------------
> 4 files changed, 44 insertions(+), 27 deletions(-)
> create mode 100644 arch/mips/generic/board-ocelot.its.S
> delete mode 100644 arch/mips/generic/board-ocelot_pcb123.its.S
>
> diff --git a/arch/mips/generic/Kconfig b/arch/mips/generic/Kconfig
> index 08e33c6..fd60198 100644
> --- a/arch/mips/generic/Kconfig
> +++ b/arch/mips/generic/Kconfig
> @@ -65,11 +65,11 @@ config FIT_IMAGE_FDT_XILFPGA
> Enable this to include the FDT for the MIPSfpga platform
> from Imagination Technologies in the FIT kernel image.
>
> -config FIT_IMAGE_FDT_OCELOT_PCB123
> - bool "Include FDT for Microsemi Ocelot PCB123"
> +config FIT_IMAGE_FDT_OCELOT
> + bool "Include FDT for Microsemi Ocelot development platforms"
> select MSCC_OCELOT
> help
> - Enable this to include the FDT for the Ocelot PCB123 platform
> + Enable this to include the FDT for the Ocelot development platforms
> from Microsemi in the FIT kernel image.
> This requires u-boot on the platform.
>
> diff --git a/arch/mips/generic/Platform b/arch/mips/generic/Platform
> index 879cb80..eaa19d1 100644
> --- a/arch/mips/generic/Platform
> +++ b/arch/mips/generic/Platform
> @@ -16,5 +16,5 @@ all-$(CONFIG_MIPS_GENERIC) := vmlinux.gz.itb
> its-y := vmlinux.its.S
> its-$(CONFIG_FIT_IMAGE_FDT_BOSTON) += board-boston.its.S
> its-$(CONFIG_FIT_IMAGE_FDT_NI169445) += board-ni169445.its.S
> -its-$(CONFIG_FIT_IMAGE_FDT_OCELOT_PCB123) += board-ocelot_pcb123.its.S
> +its-$(CONFIG_FIT_IMAGE_FDT_OCELOT) += board-ocelot.its.S
> its-$(CONFIG_FIT_IMAGE_FDT_XILFPGA) += board-xilfpga.its.S
> diff --git a/arch/mips/generic/board-ocelot.its.S b/arch/mips/generic/board-ocelot.its.S
> new file mode 100644
> index 0000000..3da2398
> --- /dev/null
> +++ b/arch/mips/generic/board-ocelot.its.S
> @@ -0,0 +1,40 @@
> +/* SPDX-License-Identifier: (GPL-2.0 OR MIT) */
> +/ {
> + images {
> + fdt@ocelot_pcb123 {
> + description = "MSCC Ocelot PCB123 Device Tree";
> + data = /incbin/("boot/dts/mscc/ocelot_pcb123.dtb");
> + type = "flat_dt";
> + arch = "mips";
> + compression = "none";
> + hash@0 {
> + algo = "sha1";
> + };
> + };
> +
> + fdt@ocelot_pcb120 {
> + description = "MSCC Ocelot PCB120 Device Tree";
> + data = /incbin/("boot/dts/mscc/ocelot_pcb120.dtb");
> + type = "flat_dt";
> + arch = "mips";
> + compression = "none";
> + hash@0 {
> + algo = "sha1";
> + };
> + };
> + };
> +
> + configurations {
> + conf@ocelot_pcb123 {
> + description = "Ocelot Linux kernel";
> + kernel = "kernel@0";
> + fdt = "fdt@ocelot_pcb123";
> + };
> +
> + conf@ocelot_pcb120 {
> + description = "Ocelot Linux kernel";
> + kernel = "kernel@0";
> + fdt = "fdt@ocelot_pcb120";
> + };
> + };
> +};
> diff --git a/arch/mips/generic/board-ocelot_pcb123.its.S b/arch/mips/generic/board-ocelot_pcb123.its.S
> deleted file mode 100644
> index 5a7d5e1..0000000
> --- a/arch/mips/generic/board-ocelot_pcb123.its.S
> +++ /dev/null
> @@ -1,23 +0,0 @@
> -/* SPDX-License-Identifier: (GPL-2.0 OR MIT) */
> -/ {
> - images {
> - fdt@ocelot_pcb123 {
> - description = "MSCC Ocelot PCB123 Device Tree";
> - data = /incbin/("boot/dts/mscc/ocelot_pcb123.dtb");
> - type = "flat_dt";
> - arch = "mips";
> - compression = "none";
> - hash@0 {
> - algo = "sha1";
> - };
> - };
> - };
> -
> - configurations {
> - conf@ocelot_pcb123 {
> - description = "Ocelot Linux kernel";
> - kernel = "kernel@0";
> - fdt = "fdt@ocelot_pcb123";
> - };
> - };
> -};
> --
> git-series 0.9.1
--
Alexandre Belloni, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
Hi Alexandre,
On Fri, Sep 14, 2018 at 04:54:46PM +0200, Alexandre Belloni wrote:
> Hi,
>
> On 14/09/2018 11:44:26+0200, Quentin Schulz wrote:
> > In order to use GPIO4 as a GPIO, we need to mux it in this mode so let's
> > declare a new pinctrl DT node for it.
> >
> > Signed-off-by: Quentin Schulz <[email protected]>
> > ---
> > arch/mips/boot/dts/mscc/ocelot.dtsi | 5 +++++
> > 1 file changed, 5 insertions(+)
> >
> > diff --git a/arch/mips/boot/dts/mscc/ocelot.dtsi b/arch/mips/boot/dts/mscc/ocelot.dtsi
> > index 8ce317c..b5c4c74 100644
> > --- a/arch/mips/boot/dts/mscc/ocelot.dtsi
> > +++ b/arch/mips/boot/dts/mscc/ocelot.dtsi
> > @@ -182,6 +182,11 @@
> > interrupts = <13>;
> > #interrupt-cells = <2>;
> >
> > + gpio4: gpio4 {
> > + pins = "GPIO_4";
> > + function = "gpio";
> > + };
> > +
>
> For a GPIO, I would do that in the board dts because it is not used
> directly in the dtsi.
>
And the day we've two boards using this pinctrl we move it to a dtsi. Is
that the plan?
Thanks,
Quentin
Hi Andrew,
On Fri, Sep 14, 2018 at 03:29:30PM +0200, Quentin Schulz wrote:
> Hi Andrew,
>
> On Fri, Sep 14, 2018 at 03:18:46PM +0200, Andrew Lunn wrote:
> > > Most of the init sequence of a PHY of the package is common to all PHYs
> > > in the package, thus we use the SMI broadcast feature which enables us
> > > to propagate a write in one register of one PHY to all PHYs in the
> > > package.
> >
> > Hi Quinten
> >
> > Could you say a bit more about the broadcast. Does the SMI broadcast
> > go to all PHY everywhere on an MDIO bus, or only all PHYs within one
> > package? I'm just thinking about the case you need two of these
> > packages to cover 8 switch ports.
> >
>
> Ah sorry, that wasn't very explicit. That's a feature on the PHY side so
> my wildest guess is that it wouldn't impact any other PHY outside of
> this package. Affecting any other PHY on the bus is counter-intuitive to
> me but I'll ask the HW engineers for confirmation.
>
Confirmed by HW engineers, it only impacts PHYs in the same package.
Quentin
> Confirmed by HW engineers, it only impacts PHYs in the same package.
Hi Quentin
Thanks for checking. As you said, it would be counter intuitive,
meaning a lot of confusion if it actually did happen.
Maybe you can add "in package" before broadcast in the commit message
and the code comments.
Andrew
On Fri, Sep 14, 2018 at 06:26:38PM +0200, Quentin Schulz wrote:
> Hi Alexandre,
>
> On Fri, Sep 14, 2018 at 04:54:46PM +0200, Alexandre Belloni wrote:
> > Hi,
> >
> > On 14/09/2018 11:44:26+0200, Quentin Schulz wrote:
> > > In order to use GPIO4 as a GPIO, we need to mux it in this mode so let's
> > > declare a new pinctrl DT node for it.
> > >
> > > Signed-off-by: Quentin Schulz <[email protected]>
> > > ---
> > > arch/mips/boot/dts/mscc/ocelot.dtsi | 5 +++++
> > > 1 file changed, 5 insertions(+)
> > >
> > > diff --git a/arch/mips/boot/dts/mscc/ocelot.dtsi b/arch/mips/boot/dts/mscc/ocelot.dtsi
> > > index 8ce317c..b5c4c74 100644
> > > --- a/arch/mips/boot/dts/mscc/ocelot.dtsi
> > > +++ b/arch/mips/boot/dts/mscc/ocelot.dtsi
> > > @@ -182,6 +182,11 @@
> > > interrupts = <13>;
> > > #interrupt-cells = <2>;
> > >
> > > + gpio4: gpio4 {
> > > + pins = "GPIO_4";
> > > + function = "gpio";
> > > + };
> > > +
> >
> > For a GPIO, I would do that in the board dts because it is not used
> > directly in the dtsi.
> >
>
> And the day we've two boards using this pinctrl we move it to a dtsi. Is
> that the plan?
Hi Quentin
gpio4 appears to be pretty arbitrary. Could a different design use a
different gpio? It me, this seems like a board property.
Andrew
> struct vsc8531_private {
> int rate_magic;
> u16 supp_led_modes;
> @@ -181,6 +354,7 @@ struct vsc8531_private {
> struct vsc85xx_hw_stat *hw_stats;
> u64 *stats;
> int nstats;
> + bool pkg_init;
> +/* bus->mdio_lock should be locked when using this function */
> +static int vsc8584_cmd(struct mii_bus *bus, int phy, u16 val)
> +{
> + unsigned long deadline;
> + u16 reg_val;
> +
> + __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS,
> + MSCC_PHY_PAGE_EXTENDED_GPIO);
> +
> + __mdiobus_write(bus, phy, MSCC_PHY_PROC_CMD, PROC_CMD_NCOMPLETED | val);
Hi Quentin
All the __mdiobus_write() look a bit ugly. Maybe add bus and base_addr
to the vsc8531_private structure. Then add helpers
phy_write_base_phy(priv, reg, val) and phy_read_base_phy(priv, reg).
You could also add in:
if (unlikely(!mutex_is_locked(&priv->bus->mdio_lock))) {
dev_err(bus->dev, "MDIO bus lock not held!\n");
dump_stack();
}
Having such code in the mv88e6xxx driver has found a few bugs for me.
Andrew
On 09/14/2018 02:44 AM, Quentin Schulz wrote:
> Part of the config init is common between the VSC8584 and the VSC8574,
> so to prepare the upcoming support for VSC8574, separate config_init
> PHY-specific code to config_pre_init function which is set in the probe
> function of the PHY and used in config_init.
>
> Signed-off-by: Quentin Schulz <[email protected]>
> ---
> drivers/net/phy/mscc.c | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/net/phy/mscc.c b/drivers/net/phy/mscc.c
> index b450489..69cc3cf 100644
> --- a/drivers/net/phy/mscc.c
> +++ b/drivers/net/phy/mscc.c
> @@ -355,6 +355,7 @@ struct vsc8531_private {
> u64 *stats;
> int nstats;
> bool pkg_init;
> + int (*config_pre_init)(struct mii_bus *bus, int phy);
Is not this overkill given that you have a reference to the phy_device,
you could check for the for phy_id to know which exact type you have and
call the appropriate pre_init function?
unsigned int phy might be more appropriate.
--
Florian
On 14/09/2018 18:26:38+0200, Quentin Schulz wrote:
> Hi Alexandre,
>
> On Fri, Sep 14, 2018 at 04:54:46PM +0200, Alexandre Belloni wrote:
> > Hi,
> >
> > On 14/09/2018 11:44:26+0200, Quentin Schulz wrote:
> > > In order to use GPIO4 as a GPIO, we need to mux it in this mode so let's
> > > declare a new pinctrl DT node for it.
> > >
> > > Signed-off-by: Quentin Schulz <[email protected]>
> > > ---
> > > arch/mips/boot/dts/mscc/ocelot.dtsi | 5 +++++
> > > 1 file changed, 5 insertions(+)
> > >
> > > diff --git a/arch/mips/boot/dts/mscc/ocelot.dtsi b/arch/mips/boot/dts/mscc/ocelot.dtsi
> > > index 8ce317c..b5c4c74 100644
> > > --- a/arch/mips/boot/dts/mscc/ocelot.dtsi
> > > +++ b/arch/mips/boot/dts/mscc/ocelot.dtsi
> > > @@ -182,6 +182,11 @@
> > > interrupts = <13>;
> > > #interrupt-cells = <2>;
> > >
> > > + gpio4: gpio4 {
> > > + pins = "GPIO_4";
> > > + function = "gpio";
> > > + };
> > > +
> >
> > For a GPIO, I would do that in the board dts because it is not used
> > directly in the dtsi.
> >
>
> And the day we've two boards using this pinctrl we move it to a dtsi. Is
> that the plan?
>
Not really, at least not for gpios. I've included the pinctrl for the
uart, i2c and spi because they are the only option if you are to use
those peripherals. Else, I've would have left the pinctrl to the board
file. From my point of view, the gpios are too board specific to be in a
soc dtsi.
--
Alexandre Belloni, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
On 09/14/2018 02:44 AM, Quentin Schulz wrote:
> The VSC8574 PHY is a 4-ports PHY that is 10/100/1000BASE-T, 100BASE-FX,
> 1000BASE-X and triple-speed copper SFP capable, can communicate with
> the MAC via SGMII, QSGMII or 1000BASE-X, supports WOL, downshifting and
> can set the blinking pattern of each of its 4 LEDs, supports SyncE as
> well as HP Auto-MDIX detection.
>
> This adds support for 10/100/1000BASE-T, SGMII/QSGMII link with the MAC,
> WOL, downshifting, HP Auto-MDIX detection and blinking pattern for its 4
> LEDs.
>
> The VSC8574 has also an internal Intel 8051 microcontroller whose
> firmware needs to be patched when the PHY is reset. If the 8051's
> firmware has the expected CRC, its patching can be skipped. The
> microcontroller can be accessed from any port of the PHY, though the CRC
> function can only be done through the PHY that is the base PHY of the
> package (internal address 0) due to a limitation of the firmware.
>
> The GPIO register bank is a set of registers that are common to all PHYs
> in the package. So any modification in any register of this bank affects
> all PHYs of the package.
>
> If the PHYs haven't been reset before booting the Linux kernel and were
> configured to use interrupts for e.g. link status updates, it is
> required to clear the interrupts mask register of all PHYs before being
> able to use interrupts with any PHY. The first PHY of the package that
> will be init will take care of clearing all PHYs interrupts mask
> registers. Thus, we need to keep track of the init sequence in the
> package, if it's already been done or if it's to be done.
>
> Most of the init sequence of a PHY of the package is common to all PHYs
> in the package, thus we use the SMI broadcast feature which enables us
> to propagate a write in one register of one PHY to all PHYs in the
> package.
>
> Signed-off-by: Quentin Schulz <[email protected]>
> ---
[snip]
> + reg = __mdiobus_read(bus, phy, MSCC_PHY_TEST_PAGE_8);
> + reg |= 0x8000;
Having a define would be nice here? This looks like a write enable?
> + __mdiobus_write(bus, phy, MSCC_PHY_TEST_PAGE_8, reg);
> +
> + __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_TR);
> +
> + vsc8584_csr_write(bus, phy, 0x8fae, 0x000401bd);
Just make this an array of address + value pairs and blast it to the
PHY, having them be inlined here is both error prone and does not scale
well at all.
[snip]
> + vsc8584_csr_write(bus, phy, 0x84a8, 0x00000000);
> + vsc8584_csr_write(bus, phy, 0x84aa, 0x00000000);
> + vsc8584_csr_write(bus, phy, 0x84ae, 0x007df7dd);
> + vsc8584_csr_write(bus, phy, 0x84b0, 0x006d95d4);
> + vsc8584_csr_write(bus, phy, 0x84b2, 0x00492410);
Likewise
[snip]
--
Florian
On 09/14/18 02:44, Quentin Schulz wrote:
> The VSC8584 (and most likely other PHYs in the same generation) has two
> additional LED modes that can be picked, so let's add them.
>
> Signed-off-by: Quentin Schulz <[email protected]>
Reviewed-by: Florian Fainelli <[email protected]>
--
Florian
Just as a drive-by comment this seems vaguely related to the Vitesse
DSA switch I merged in drivers/net/dsa/vitesse-vsc73xx.c
The VSC* product name handily gives away the origin in Vitesse's
product line.
The VSC73xx also have the 8051 CPU and internal RAM, but are
accessed (typically) over SPI, and AFAICT this thing is talking over
MDIO.
The Vitesse 73xx however also supports a WAN port and VLANs
which makes it significantly different, falling into switch class I
guess.
These VSC85*4's does have an SPI interface as well, according
to the data sheet but I assume your target boards don't even
connect it?
When it comes to 8051 code we have quite a lot of this in the kernel
these days, I suspect the 8051 snippets in this code could be
disassembled and put into linux-firmware in source form, but
that is maybe a bit overly ambitious. We have done that for a few
USB to serial controllers using the EzUSB 8051 though:
https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/keyspan_pda
https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/usbdux
These can rebuild their firmware using the as31 assembler.
https://github.com/nitsky/as31
Yours,
Linus Walleij
Hi Andrew,
On Fri, Sep 14, 2018 at 07:02:21PM +0200, Andrew Lunn wrote:
> On Fri, Sep 14, 2018 at 06:26:38PM +0200, Quentin Schulz wrote:
> > Hi Alexandre,
> >
> > On Fri, Sep 14, 2018 at 04:54:46PM +0200, Alexandre Belloni wrote:
> > > Hi,
> > >
> > > On 14/09/2018 11:44:26+0200, Quentin Schulz wrote:
> > > > In order to use GPIO4 as a GPIO, we need to mux it in this mode so let's
> > > > declare a new pinctrl DT node for it.
> > > >
> > > > Signed-off-by: Quentin Schulz <[email protected]>
> > > > ---
> > > > arch/mips/boot/dts/mscc/ocelot.dtsi | 5 +++++
> > > > 1 file changed, 5 insertions(+)
> > > >
> > > > diff --git a/arch/mips/boot/dts/mscc/ocelot.dtsi b/arch/mips/boot/dts/mscc/ocelot.dtsi
> > > > index 8ce317c..b5c4c74 100644
> > > > --- a/arch/mips/boot/dts/mscc/ocelot.dtsi
> > > > +++ b/arch/mips/boot/dts/mscc/ocelot.dtsi
> > > > @@ -182,6 +182,11 @@
> > > > interrupts = <13>;
> > > > #interrupt-cells = <2>;
> > > >
> > > > + gpio4: gpio4 {
> > > > + pins = "GPIO_4";
> > > > + function = "gpio";
> > > > + };
> > > > +
> > >
> > > For a GPIO, I would do that in the board dts because it is not used
> > > directly in the dtsi.
> > >
> >
> > And the day we've two boards using this pinctrl we move it to a dtsi. Is
> > that the plan?
>
> Hi Quentin
>
> gpio4 appears to be pretty arbitrary. Could a different design use a
> different gpio? It me, this seems like a board property.
>
Right now, I don't see why it couldn't be.
Quentin
Hi Alexandre,
On Fri, Sep 14, 2018 at 08:02:22PM +0200, Alexandre Belloni wrote:
> On 14/09/2018 18:26:38+0200, Quentin Schulz wrote:
> > Hi Alexandre,
> >
> > On Fri, Sep 14, 2018 at 04:54:46PM +0200, Alexandre Belloni wrote:
> > > Hi,
> > >
> > > On 14/09/2018 11:44:26+0200, Quentin Schulz wrote:
> > > > In order to use GPIO4 as a GPIO, we need to mux it in this mode so let's
> > > > declare a new pinctrl DT node for it.
> > > >
> > > > Signed-off-by: Quentin Schulz <[email protected]>
> > > > ---
> > > > arch/mips/boot/dts/mscc/ocelot.dtsi | 5 +++++
> > > > 1 file changed, 5 insertions(+)
> > > >
> > > > diff --git a/arch/mips/boot/dts/mscc/ocelot.dtsi b/arch/mips/boot/dts/mscc/ocelot.dtsi
> > > > index 8ce317c..b5c4c74 100644
> > > > --- a/arch/mips/boot/dts/mscc/ocelot.dtsi
> > > > +++ b/arch/mips/boot/dts/mscc/ocelot.dtsi
> > > > @@ -182,6 +182,11 @@
> > > > interrupts = <13>;
> > > > #interrupt-cells = <2>;
> > > >
> > > > + gpio4: gpio4 {
> > > > + pins = "GPIO_4";
> > > > + function = "gpio";
> > > > + };
> > > > +
> > >
> > > For a GPIO, I would do that in the board dts because it is not used
> > > directly in the dtsi.
> > >
> >
> > And the day we've two boards using this pinctrl we move it to a dtsi. Is
> > that the plan?
> >
>
> Not really, at least not for gpios. I've included the pinctrl for the
> uart, i2c and spi because they are the only option if you are to use
> those peripherals. Else, I've would have left the pinctrl to the board
> file. From my point of view, the gpios are too board specific to be in a
> soc dtsi.
>
Understood, will move it to the board file.
Thanks,
Quentin
Hi Florian,
On Fri, Sep 14, 2018 at 10:57:08AM -0700, Florian Fainelli wrote:
> On 09/14/2018 02:44 AM, Quentin Schulz wrote:
> > Part of the config init is common between the VSC8584 and the VSC8574,
> > so to prepare the upcoming support for VSC8574, separate config_init
> > PHY-specific code to config_pre_init function which is set in the probe
> > function of the PHY and used in config_init.
> >
> > Signed-off-by: Quentin Schulz <[email protected]>
> > ---
> > drivers/net/phy/mscc.c | 4 +++-
> > 1 file changed, 3 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/net/phy/mscc.c b/drivers/net/phy/mscc.c
> > index b450489..69cc3cf 100644
> > --- a/drivers/net/phy/mscc.c
> > +++ b/drivers/net/phy/mscc.c
> > @@ -355,6 +355,7 @@ struct vsc8531_private {
> > u64 *stats;
> > int nstats;
> > bool pkg_init;
> > + int (*config_pre_init)(struct mii_bus *bus, int phy);
>
> Is not this overkill given that you have a reference to the phy_device,
> you could check for the for phy_id to know which exact type you have and
> call the appropriate pre_init function?
>
Agreed. It just seemed "cleaner" to me to set config_pre_init in
separate probe functions which are pretty straightforward and short.
Thus not complexifying an already-not-so-straightforward config_init
function.
Anyway, I'll use the phy_id as suggested so that we don't overload the
vsc8531_private structure (since it's also only called once, in
config_init).
Thanks,
Quentin
Hi Andrew,
On Fri, Sep 14, 2018 at 07:27:54PM +0200, Andrew Lunn wrote:
>
> > struct vsc8531_private {
> > int rate_magic;
> > u16 supp_led_modes;
> > @@ -181,6 +354,7 @@ struct vsc8531_private {
> > struct vsc85xx_hw_stat *hw_stats;
> > u64 *stats;
> > int nstats;
> > + bool pkg_init;
>
> > +/* bus->mdio_lock should be locked when using this function */
> > +static int vsc8584_cmd(struct mii_bus *bus, int phy, u16 val)
> > +{
> > + unsigned long deadline;
> > + u16 reg_val;
> > +
> > + __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS,
> > + MSCC_PHY_PAGE_EXTENDED_GPIO);
> > +
> > + __mdiobus_write(bus, phy, MSCC_PHY_PROC_CMD, PROC_CMD_NCOMPLETED | val);
>
> Hi Quentin
>
> All the __mdiobus_write() look a bit ugly.
I agree :)
> Maybe add bus and base_addr
> to the vsc8531_private structure. Then add helpers
> phy_write_base_phy(priv, reg, val) and phy_read_base_phy(priv, reg).
>
ACK.
> You could also add in:
>
> if (unlikely(!mutex_is_locked(&priv->bus->mdio_lock))) {
> dev_err(bus->dev, "MDIO bus lock not held!\n");
> dump_stack();
> }
>
> Having such code in the mv88e6xxx driver has found a few bugs for me.
>
ACK.
Thanks,
Quentin
Hi Andrew,
On Fri, Sep 14, 2018 at 06:58:24PM +0200, Andrew Lunn wrote:
> > Confirmed by HW engineers, it only impacts PHYs in the same package.
>
> Hi Quentin
>
> Thanks for checking. As you said, it would be counter intuitive,
> meaning a lot of confusion if it actually did happen.
>
> Maybe you can add "in package" before broadcast in the commit message
> and the code comments.
>
ACK.
Quentin
Hi Florian,
On Fri, Sep 14, 2018 at 01:26:06PM -0700, Florian Fainelli wrote:
> On 09/14/2018 02:44 AM, Quentin Schulz wrote:
> > The VSC8574 PHY is a 4-ports PHY that is 10/100/1000BASE-T, 100BASE-FX,
> > 1000BASE-X and triple-speed copper SFP capable, can communicate with
> > the MAC via SGMII, QSGMII or 1000BASE-X, supports WOL, downshifting and
> > can set the blinking pattern of each of its 4 LEDs, supports SyncE as
> > well as HP Auto-MDIX detection.
> >
> > This adds support for 10/100/1000BASE-T, SGMII/QSGMII link with the MAC,
> > WOL, downshifting, HP Auto-MDIX detection and blinking pattern for its 4
> > LEDs.
> >
> > The VSC8574 has also an internal Intel 8051 microcontroller whose
> > firmware needs to be patched when the PHY is reset. If the 8051's
> > firmware has the expected CRC, its patching can be skipped. The
> > microcontroller can be accessed from any port of the PHY, though the CRC
> > function can only be done through the PHY that is the base PHY of the
> > package (internal address 0) due to a limitation of the firmware.
> >
> > The GPIO register bank is a set of registers that are common to all PHYs
> > in the package. So any modification in any register of this bank affects
> > all PHYs of the package.
> >
> > If the PHYs haven't been reset before booting the Linux kernel and were
> > configured to use interrupts for e.g. link status updates, it is
> > required to clear the interrupts mask register of all PHYs before being
> > able to use interrupts with any PHY. The first PHY of the package that
> > will be init will take care of clearing all PHYs interrupts mask
> > registers. Thus, we need to keep track of the init sequence in the
> > package, if it's already been done or if it's to be done.
> >
> > Most of the init sequence of a PHY of the package is common to all PHYs
> > in the package, thus we use the SMI broadcast feature which enables us
> > to propagate a write in one register of one PHY to all PHYs in the
> > package.
> >
> > Signed-off-by: Quentin Schulz <[email protected]>
> > ---
>
> [snip]
>
> > + reg = __mdiobus_read(bus, phy, MSCC_PHY_TEST_PAGE_8);
> > + reg |= 0x8000;
>
> Having a define would be nice here? This looks like a write enable?
>
I had asked for the meaning of this bit in this register before but we
couldn't find documentation for it. I'll ask again and let you know.
> > + __mdiobus_write(bus, phy, MSCC_PHY_TEST_PAGE_8, reg);
> > +
> > + __mdiobus_write(bus, phy, MSCC_EXT_PAGE_ACCESS, MSCC_PHY_PAGE_TR);
> > +
> > + vsc8584_csr_write(bus, phy, 0x8fae, 0x000401bd);
>
> Just make this an array of address + value pairs and blast it to the
> PHY, having them be inlined here is both error prone and does not scale
> well at all.
Right. Turned out it was a great idea as the below values were
mistakenly adding 0x8000 to the register (which was fine since in
vsc8584_csr_write does a 0x8000 | reg).
Thanks,
Quentin
Hi Florian,
On Fri, Sep 14, 2018 at 01:26:06PM -0700, Florian Fainelli wrote:
> On 09/14/2018 02:44 AM, Quentin Schulz wrote:
> > The VSC8574 PHY is a 4-ports PHY that is 10/100/1000BASE-T, 100BASE-FX,
> > 1000BASE-X and triple-speed copper SFP capable, can communicate with
> > the MAC via SGMII, QSGMII or 1000BASE-X, supports WOL, downshifting and
> > can set the blinking pattern of each of its 4 LEDs, supports SyncE as
> > well as HP Auto-MDIX detection.
> >
> > This adds support for 10/100/1000BASE-T, SGMII/QSGMII link with the MAC,
> > WOL, downshifting, HP Auto-MDIX detection and blinking pattern for its 4
> > LEDs.
> >
> > The VSC8574 has also an internal Intel 8051 microcontroller whose
> > firmware needs to be patched when the PHY is reset. If the 8051's
> > firmware has the expected CRC, its patching can be skipped. The
> > microcontroller can be accessed from any port of the PHY, though the CRC
> > function can only be done through the PHY that is the base PHY of the
> > package (internal address 0) due to a limitation of the firmware.
> >
> > The GPIO register bank is a set of registers that are common to all PHYs
> > in the package. So any modification in any register of this bank affects
> > all PHYs of the package.
> >
> > If the PHYs haven't been reset before booting the Linux kernel and were
> > configured to use interrupts for e.g. link status updates, it is
> > required to clear the interrupts mask register of all PHYs before being
> > able to use interrupts with any PHY. The first PHY of the package that
> > will be init will take care of clearing all PHYs interrupts mask
> > registers. Thus, we need to keep track of the init sequence in the
> > package, if it's already been done or if it's to be done.
> >
> > Most of the init sequence of a PHY of the package is common to all PHYs
> > in the package, thus we use the SMI broadcast feature which enables us
> > to propagate a write in one register of one PHY to all PHYs in the
> > package.
> >
> > Signed-off-by: Quentin Schulz <[email protected]>
> > ---
>
> [snip]
>
> > + reg = __mdiobus_read(bus, phy, MSCC_PHY_TEST_PAGE_8);
> > + reg |= 0x8000;
>
> Having a define would be nice here? This looks like a write enable?
>
Same as for the SerDes muxing patch series, I'm sending a new version
without the suggested change. I've requested some information on this
bit but can't guarantee I'll get anything (and if anything, worthy).
I'll try not to forget to add a constant if and when I get a clue on
what this bit does.
Thanks,
Quentin