* [RFC PATCH net-next 0/4] Add support for LAN937x T1 Phy
@ 2022-02-28 14:05 Arun Ramadoss
2022-02-28 14:05 ` [RFC PATCH net-next 1/4] net: phy: used the genphy_soft_reset for phy reset in Lan87xx Arun Ramadoss
` (3 more replies)
0 siblings, 4 replies; 11+ messages in thread
From: Arun Ramadoss @ 2022-02-28 14:05 UTC (permalink / raw)
To: netdev, linux-kernel
Cc: Andrew Lunn, Heiner Kallweit, Russell King, Jakub Kicinski,
David S . Miller
LAN937x is a Multi-port 100Base-T1 Switch and it internally uses LAN87xx
T1 Phy. This series of patch update the intialization routine for the
LAN87xx phy and also add LAN937x part support. Added the T1 Phy
master-slave configuration through ethtool.
Arun Ramadoss (4):
net: phy: used the genphy_soft_reset for phy reset in Lan87xx
net: phy: updated the initialization routine for LAN87xx
net: phy: added the LAN937x phy support
net: phy: added master-slave config and cable diagnostics for Lan937x
drivers/net/phy/microchip_t1.c | 373 +++++++++++++++++++++++++++------
1 file changed, 308 insertions(+), 65 deletions(-)
--
2.33.0
^ permalink raw reply [flat|nested] 11+ messages in thread* [RFC PATCH net-next 1/4] net: phy: used the genphy_soft_reset for phy reset in Lan87xx 2022-02-28 14:05 [RFC PATCH net-next 0/4] Add support for LAN937x T1 Phy Arun Ramadoss @ 2022-02-28 14:05 ` Arun Ramadoss 2022-03-02 3:19 ` Andrew Lunn 2022-02-28 14:05 ` [RFC PATCH net-next 2/4] net: phy: updated the initialization routine for LAN87xx Arun Ramadoss ` (2 subsequent siblings) 3 siblings, 1 reply; 11+ messages in thread From: Arun Ramadoss @ 2022-02-28 14:05 UTC (permalink / raw) To: netdev, linux-kernel Cc: Andrew Lunn, Heiner Kallweit, Russell King, Jakub Kicinski, David S . Miller Replaced current code for soft resetting phy to genphy_soft_reset function. And added the macro for LAN87xx Phy ID. Signed-off-by: Prasanna Vengateshan <prasanna.vengateshan@microchip.com> Signed-off-by: Arun Ramadoss <arun.ramadoss@microchip.com> --- drivers/net/phy/microchip_t1.c | 32 ++++++++++++-------------------- 1 file changed, 12 insertions(+), 20 deletions(-) diff --git a/drivers/net/phy/microchip_t1.c b/drivers/net/phy/microchip_t1.c index bc50224d43dd..ece21c1e5716 100644 --- a/drivers/net/phy/microchip_t1.c +++ b/drivers/net/phy/microchip_t1.c @@ -9,6 +9,9 @@ #include <linux/ethtool.h> #include <linux/ethtool_netlink.h> +#define LAN87XX_PHY_ID 0x0007c150 +#define MICROCHIP_PHY_ID_MASK 0xfffffff0 + /* External Register Control Register */ #define LAN87XX_EXT_REG_CTL (0x14) #define LAN87XX_EXT_REG_CTL_RD_CTL (0x1000) @@ -197,20 +200,10 @@ static int lan87xx_phy_init(struct phy_device *phydev) if (rc < 0) return rc; - /* Soft Reset the SMI block */ - rc = access_ereg_modify_changed(phydev, PHYACC_ATTR_BANK_SMI, - 0x00, 0x8000, 0x8000); - if (rc < 0) - return rc; - - /* Check to see if the self-clearing bit is cleared */ - usleep_range(1000, 2000); - rc = access_ereg(phydev, PHYACC_ATTR_MODE_READ, - PHYACC_ATTR_BANK_SMI, 0x00, 0); + /* phy Soft reset */ + rc = genphy_soft_reset(phydev); if (rc < 0) return rc; - if ((rc & 0x8000) != 0) - return -ETIMEDOUT; /* PHY Initialization */ for (i = 0; i < ARRAY_SIZE(init); i++) { @@ -273,6 +266,9 @@ static int lan87xx_config_init(struct phy_device *phydev) { int rc = lan87xx_phy_init(phydev); + if (rc < 0) + phydev_err(phydev, "failed to initialize phy\n"); + return rc < 0 ? rc : 0; } @@ -506,18 +502,14 @@ static int lan87xx_cable_test_get_status(struct phy_device *phydev, static struct phy_driver microchip_t1_phy_driver[] = { { - .phy_id = 0x0007c150, - .phy_id_mask = 0xfffffff0, - .name = "Microchip LAN87xx T1", + .phy_id = LAN87XX_PHY_ID, + .phy_id_mask = MICROCHIP_PHY_ID_MASK, + .name = "LAN87xx T1", .flags = PHY_POLL_CABLE_TEST, - .features = PHY_BASIC_T1_FEATURES, - .config_init = lan87xx_config_init, - .config_intr = lan87xx_phy_config_intr, .handle_interrupt = lan87xx_handle_interrupt, - .suspend = genphy_suspend, .resume = genphy_resume, .cable_test_start = lan87xx_cable_test_start, @@ -528,7 +520,7 @@ static struct phy_driver microchip_t1_phy_driver[] = { module_phy_driver(microchip_t1_phy_driver); static struct mdio_device_id __maybe_unused microchip_t1_tbl[] = { - { 0x0007c150, 0xfffffff0 }, + { LAN87XX_PHY_ID, MICROCHIP_PHY_ID_MASK}, { } }; -- 2.33.0 ^ permalink raw reply related [flat|nested] 11+ messages in thread
* Re: [RFC PATCH net-next 1/4] net: phy: used the genphy_soft_reset for phy reset in Lan87xx 2022-02-28 14:05 ` [RFC PATCH net-next 1/4] net: phy: used the genphy_soft_reset for phy reset in Lan87xx Arun Ramadoss @ 2022-03-02 3:19 ` Andrew Lunn 2022-03-02 10:04 ` Arun.Ramadoss 0 siblings, 1 reply; 11+ messages in thread From: Andrew Lunn @ 2022-03-02 3:19 UTC (permalink / raw) To: Arun Ramadoss Cc: netdev, linux-kernel, Heiner Kallweit, Russell King, Jakub Kicinski, David S . Miller On Mon, Feb 28, 2022 at 07:35:07PM +0530, Arun Ramadoss wrote: > Replaced current code for soft resetting phy to genphy_soft_reset > function. And added the macro for LAN87xx Phy ID. Hi Arun Please don't mix multiple things in one patch. Looking at the actual path, you have: > +#define LAN87XX_PHY_ID 0x0007c150 > +#define MICROCHIP_PHY_ID_MASK 0xfffffff0 Part of macros for PHY ID. > + > /* External Register Control Register */ > #define LAN87XX_EXT_REG_CTL (0x14) > #define LAN87XX_EXT_REG_CTL_RD_CTL (0x1000) > @@ -197,20 +200,10 @@ static int lan87xx_phy_init(struct phy_device *phydev) > if (rc < 0) > return rc; > > - /* Soft Reset the SMI block */ > - rc = access_ereg_modify_changed(phydev, PHYACC_ATTR_BANK_SMI, > - 0x00, 0x8000, 0x8000); > - if (rc < 0) > - return rc; > - > - /* Check to see if the self-clearing bit is cleared */ > - usleep_range(1000, 2000); > - rc = access_ereg(phydev, PHYACC_ATTR_MODE_READ, > - PHYACC_ATTR_BANK_SMI, 0x00, 0); > + /* phy Soft reset */ > + rc = genphy_soft_reset(phydev); > if (rc < 0) > return rc; > - if ((rc & 0x8000) != 0) > - return -ETIMEDOUT; Soft reset. > > /* PHY Initialization */ > for (i = 0; i < ARRAY_SIZE(init); i++) { > @@ -273,6 +266,9 @@ static int lan87xx_config_init(struct phy_device *phydev) > { > int rc = lan87xx_phy_init(phydev); > > + if (rc < 0) > + phydev_err(phydev, "failed to initialize phy\n"); > + A new error message. > return rc < 0 ? rc : 0; > } > > @@ -506,18 +502,14 @@ static int lan87xx_cable_test_get_status(struct phy_device *phydev, > > static struct phy_driver microchip_t1_phy_driver[] = { > { > - .phy_id = 0x0007c150, > - .phy_id_mask = 0xfffffff0, > - .name = "Microchip LAN87xx T1", > + .phy_id = LAN87XX_PHY_ID, > + .phy_id_mask = MICROCHIP_PHY_ID_MASK, 2nd part of the PHY ID macros. > + .name = "LAN87xx T1", A change in name. > .flags = PHY_POLL_CABLE_TEST, > - > .features = PHY_BASIC_T1_FEATURES, > - > .config_init = lan87xx_config_init, > - > .config_intr = lan87xx_phy_config_intr, > .handle_interrupt = lan87xx_handle_interrupt, > - White space changes. You can also use PHY_ID_MATCH_MODEL(). Andrew ^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [RFC PATCH net-next 1/4] net: phy: used the genphy_soft_reset for phy reset in Lan87xx 2022-03-02 3:19 ` Andrew Lunn @ 2022-03-02 10:04 ` Arun.Ramadoss 0 siblings, 0 replies; 11+ messages in thread From: Arun.Ramadoss @ 2022-03-02 10:04 UTC (permalink / raw) To: andrew; +Cc: linux, netdev, linux-kernel, davem, hkallweit1, kuba On Wed, 2022-03-02 at 04:19 +0100, Andrew Lunn wrote: > EXTERNAL EMAIL: Do not click links or open attachments unless you > know the content is safe > > On Mon, Feb 28, 2022 at 07:35:07PM +0530, Arun Ramadoss wrote: > > Replaced current code for soft resetting phy to genphy_soft_reset > > function. And added the macro for LAN87xx Phy ID. > > Hi Arun > > Please don't mix multiple things in one patch. > > Looking at the actual path, you have: > > > +#define LAN87XX_PHY_ID 0x0007c150 > > +#define MICROCHIP_PHY_ID_MASK 0xfffffff0 > > Part of macros for PHY ID. > > > + > > /* External Register Control Register */ > > #define LAN87XX_EXT_REG_CTL (0x14) > > #define LAN87XX_EXT_REG_CTL_RD_CTL (0x1000) > > @@ -197,20 +200,10 @@ static int lan87xx_phy_init(struct phy_device > > *phydev) > > if (rc < 0) > > return rc; > > > > - /* Soft Reset the SMI block */ > > - rc = access_ereg_modify_changed(phydev, PHYACC_ATTR_BANK_SMI, > > - 0x00, 0x8000, 0x8000); > > - if (rc < 0) > > - return rc; > > - > > - /* Check to see if the self-clearing bit is cleared */ > > - usleep_range(1000, 2000); > > - rc = access_ereg(phydev, PHYACC_ATTR_MODE_READ, > > - PHYACC_ATTR_BANK_SMI, 0x00, 0); > > + /* phy Soft reset */ > > + rc = genphy_soft_reset(phydev); > > if (rc < 0) > > return rc; > > - if ((rc & 0x8000) != 0) > > - return -ETIMEDOUT; > > Soft reset. > > > > > /* PHY Initialization */ > > for (i = 0; i < ARRAY_SIZE(init); i++) { > > @@ -273,6 +266,9 @@ static int lan87xx_config_init(struct > > phy_device *phydev) > > { > > int rc = lan87xx_phy_init(phydev); > > > > + if (rc < 0) > > + phydev_err(phydev, "failed to initialize phy\n"); > > + > > A new error message. > > > return rc < 0 ? rc : 0; > > } > > > > @@ -506,18 +502,14 @@ static int > > lan87xx_cable_test_get_status(struct phy_device *phydev, > > > > static struct phy_driver microchip_t1_phy_driver[] = { > > { > > - .phy_id = 0x0007c150, > > - .phy_id_mask = 0xfffffff0, > > - .name = "Microchip LAN87xx T1", > > + .phy_id = LAN87XX_PHY_ID, > > + .phy_id_mask = MICROCHIP_PHY_ID_MASK, > > 2nd part of the PHY ID macros. > > > + .name = "LAN87xx T1", > > A change in name. > > > .flags = PHY_POLL_CABLE_TEST, > > - > > .features = PHY_BASIC_T1_FEATURES, > > - > > .config_init = lan87xx_config_init, > > - > > .config_intr = lan87xx_phy_config_intr, > > .handle_interrupt = lan87xx_handle_interrupt, > > - > > White space changes. > > You can also use PHY_ID_MATCH_MODEL(). > > Andrew Thanks for the comment. I will keep one change per patch and split this patch & resend. ^ permalink raw reply [flat|nested] 11+ messages in thread
* [RFC PATCH net-next 2/4] net: phy: updated the initialization routine for LAN87xx 2022-02-28 14:05 [RFC PATCH net-next 0/4] Add support for LAN937x T1 Phy Arun Ramadoss 2022-02-28 14:05 ` [RFC PATCH net-next 1/4] net: phy: used the genphy_soft_reset for phy reset in Lan87xx Arun Ramadoss @ 2022-02-28 14:05 ` Arun Ramadoss 2022-02-28 14:05 ` [RFC PATCH net-next 3/4] net: phy: added the LAN937x phy support Arun Ramadoss 2022-02-28 14:05 ` [RFC PATCH net-next 4/4] net: phy: added master-slave config and cable diagnostics for Lan937x Arun Ramadoss 3 siblings, 0 replies; 11+ messages in thread From: Arun Ramadoss @ 2022-02-28 14:05 UTC (permalink / raw) To: netdev, linux-kernel Cc: Andrew Lunn, Heiner Kallweit, Russell King, Jakub Kicinski, David S . Miller The new initialization sequence is the improvement to the existing init routine. Init routine does soft reset, run init script and set Hw_init. Added the new access_ereg_clr_poll_timeout() for polling phy bank write. Signed-off-by: Prasanna Vengateshan <prasanna.vengateshan@microchip.com> Signed-off-by: Arun Ramadoss <arun.ramadoss@microchip.com> --- drivers/net/phy/microchip_t1.c | 219 ++++++++++++++++++++++++++------- 1 file changed, 177 insertions(+), 42 deletions(-) diff --git a/drivers/net/phy/microchip_t1.c b/drivers/net/phy/microchip_t1.c index ece21c1e5716..33325e5bd884 100644 --- a/drivers/net/phy/microchip_t1.c +++ b/drivers/net/phy/microchip_t1.c @@ -40,6 +40,7 @@ #define PHYACC_ATTR_MODE_READ 0 #define PHYACC_ATTR_MODE_WRITE 1 #define PHYACC_ATTR_MODE_MODIFY 2 +#define PHYACC_ATTR_MODE_POLL 3 #define PHYACC_ATTR_BANK_SMI 0 #define PHYACC_ATTR_BANK_MISC 1 @@ -53,6 +54,28 @@ #define LAN87XX_CABLE_TEST_OPEN 1 #define LAN87XX_CABLE_TEST_SAME_SHORT 2 +/* T1 Registers */ +#define T1_AFE_PORT_CFG1_REG 0x0B +#define T1_POWER_DOWN_CONTROL_REG 0x1A +#define T1_SLV_FD_MULT_CFG_REG 0x18 +#define T1_CDR_CFG_PRE_LOCK_REG 0x05 +#define T1_CDR_CFG_POST_LOCK_REG 0x06 +#define T1_LCK_STG2_MUFACT_CFG_REG 0x1A +#define T1_LCK_STG3_MUFACT_CFG_REG 0x1B +#define T1_POST_LCK_MUFACT_CFG_REG 0x1C +#define T1_TX_RX_FIFO_CFG_REG 0x02 +#define T1_TX_LPF_FIR_CFG_REG 0x55 +#define T1_SQI_CONFIG_REG 0x2E +#define T1_MDIO_CONTROL2_REG 0x10 +#define T1_INTERRUPT_SOURCE_REG 0x18 +#define T1_INTERRUPT2_SOURCE_REG 0x08 +#define T1_EQ_FD_STG1_FRZ_CFG 0x69 +#define T1_EQ_FD_STG2_FRZ_CFG 0x6A +#define T1_EQ_FD_STG3_FRZ_CFG 0x6B +#define T1_EQ_FD_STG4_FRZ_CFG 0x6C +#define T1_EQ_WT_FD_LCK_FRZ_CFG 0x6D +#define T1_PST_EQ_LCK_STG1_FRZ_CFG 0x6E + #define DRIVER_AUTHOR "Nisar Sayed <nisar.sayed@microchip.com>" #define DRIVER_DESC "Microchip LAN87XX T1 PHY driver" @@ -120,6 +143,18 @@ static int access_ereg_modify_changed(struct phy_device *phydev, return rc; } +static int access_ereg_clr_poll_timeout(struct phy_device *phydev, u8 bank, + u8 offset, u16 mask, u16 clr) +{ + int val; + + if (bank != PHYACC_ATTR_BANK_SMI) + return -EINVAL; + + return phy_read_poll_timeout(phydev, offset, val, (val & mask) == clr, + 150, 30000, true); +} + static int lan87xx_config_rgmii_delay(struct phy_device *phydev) { int rc; @@ -160,46 +195,146 @@ static int lan87xx_config_rgmii_delay(struct phy_device *phydev) static int lan87xx_phy_init(struct phy_device *phydev) { static const struct access_ereg_val init[] = { - /* TX Amplitude = 5 */ - {PHYACC_ATTR_MODE_MODIFY, PHYACC_ATTR_BANK_AFE, 0x0B, - 0x000A, 0x001E}, - /* Clear SMI interrupts */ - {PHYACC_ATTR_MODE_READ, PHYACC_ATTR_BANK_SMI, 0x18, - 0, 0}, - /* Clear MISC interrupts */ - {PHYACC_ATTR_MODE_READ, PHYACC_ATTR_BANK_MISC, 0x08, - 0, 0}, - /* Turn on TC10 Ring Oscillator (ROSC) */ - {PHYACC_ATTR_MODE_MODIFY, PHYACC_ATTR_BANK_MISC, 0x20, - 0x0020, 0x0020}, - /* WUR Detect Length to 1.2uS, LPC Detect Length to 1.09uS */ - {PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_PCS, 0x20, - 0x283C, 0}, - /* Wake_In Debounce Length to 39uS, Wake_Out Length to 79uS */ - {PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_MISC, 0x21, - 0x274F, 0}, - /* Enable Auto Wake Forward to Wake_Out, ROSC on, Sleep, - * and Wake_In to wake PHY - */ - {PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_MISC, 0x20, - 0x80A7, 0}, - /* Enable WUP Auto Fwd, Enable Wake on MDI, Wakeup Debouncer - * to 128 uS - */ - {PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_MISC, 0x24, - 0xF110, 0}, - /* Enable HW Init */ - {PHYACC_ATTR_MODE_MODIFY, PHYACC_ATTR_BANK_SMI, 0x1A, - 0x0100, 0x0100}, + /* TXPD/TXAMP6 Configs */ + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_AFE, + T1_AFE_PORT_CFG1_REG, 0x002D, 0 }, + /* HW_Init Hi and Force_ED */ + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_SMI, + T1_POWER_DOWN_CONTROL_REG, 0x0308, 0 }, + /* Equalizer Full Duplex Freeze - T1 Slave */ + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_EQ_FD_STG1_FRZ_CFG, 0x0002, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_EQ_FD_STG2_FRZ_CFG, 0x0002, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_EQ_FD_STG3_FRZ_CFG, 0x0002, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_EQ_FD_STG4_FRZ_CFG, 0x0002, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_EQ_WT_FD_LCK_FRZ_CFG, 0x0002, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_PST_EQ_LCK_STG1_FRZ_CFG, 0x0002, 0 }, + /* Slave Full Duplex Multi Configs */ + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_SLV_FD_MULT_CFG_REG, 0x0D53, 0 }, + /* CDR Pre and Post Lock Configs */ + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_CDR_CFG_PRE_LOCK_REG, 0x0AB2, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_CDR_CFG_POST_LOCK_REG, 0x0AB3, 0 }, + /* Lock Stage 2-3 Multi Factor Config */ + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_LCK_STG2_MUFACT_CFG_REG, 0x0AEA, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_LCK_STG3_MUFACT_CFG_REG, 0x0AEB, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_POST_LCK_MUFACT_CFG_REG, 0x0AEB, 0 }, + /* Pointer delay */ + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_RX_FIFO_CFG_REG, 0x1C00, 0 }, + /* Tx iir edits */ + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1000, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1861, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1061, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1922, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1122, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1983, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1183, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1944, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1144, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x18c5, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x10c5, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1846, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1046, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1807, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1007, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1808, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1008, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1809, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1009, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x180A, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x100A, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x180B, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x100B, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x180C, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x100C, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x180D, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x100D, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x180E, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x100E, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x180F, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x100F, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1810, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1010, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1811, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1011, 0 }, + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_TX_LPF_FIR_CFG_REG, 0x1000, 0 }, + /* SQI enable */ + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_DSP, + T1_SQI_CONFIG_REG, 0x9572, 0 }, + /* Flag LPS and WUR as idle errors */ + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_SMI, + T1_MDIO_CONTROL2_REG, 0x0014, 0 }, + /* HW_Init toggle, undo force ED, TXPD off */ + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_SMI, + T1_POWER_DOWN_CONTROL_REG, 0x0200, 0 }, + /* Reset PCS to trigger hardware initialization */ + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_SMI, + T1_MDIO_CONTROL2_REG, 0x0094, 0 }, + /* Poll till Hardware is initialized */ + { PHYACC_ATTR_MODE_POLL, PHYACC_ATTR_BANK_SMI, + T1_MDIO_CONTROL2_REG, 0x0080, 0 }, + /* Tx AMP - 0x06 */ + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_AFE, + T1_AFE_PORT_CFG1_REG, 0x000C, 0 }, + /* Read INTERRUPT_SOURCE Register */ + { PHYACC_ATTR_MODE_READ, PHYACC_ATTR_BANK_SMI, + T1_INTERRUPT_SOURCE_REG, 0, 0 }, + /* Read INTERRUPT_SOURCE Register */ + { PHYACC_ATTR_MODE_READ, PHYACC_ATTR_BANK_MISC, + T1_INTERRUPT2_SOURCE_REG, 0, 0 }, + /* HW_Init Hi */ + { PHYACC_ATTR_MODE_WRITE, PHYACC_ATTR_BANK_SMI, + T1_POWER_DOWN_CONTROL_REG, 0x0300, 0 }, }; int rc, i; - /* Start manual initialization procedures in Managed Mode */ - rc = access_ereg_modify_changed(phydev, PHYACC_ATTR_BANK_SMI, - 0x1a, 0x0000, 0x0100); - if (rc < 0) - return rc; - /* phy Soft reset */ rc = genphy_soft_reset(phydev); if (rc < 0) @@ -207,11 +342,11 @@ static int lan87xx_phy_init(struct phy_device *phydev) /* PHY Initialization */ for (i = 0; i < ARRAY_SIZE(init); i++) { - if (init[i].mode == PHYACC_ATTR_MODE_MODIFY) { - rc = access_ereg_modify_changed(phydev, init[i].bank, - init[i].offset, - init[i].val, - init[i].mask); + if (init[i].mode == PHYACC_ATTR_MODE_POLL) { + rc = access_ereg_clr_poll_timeout(phydev, init[i].bank, + init[i].offset, + init[i].val, + init[i].mask); } else { rc = access_ereg(phydev, init[i].mode, init[i].bank, init[i].offset, init[i].val); -- 2.33.0 ^ permalink raw reply related [flat|nested] 11+ messages in thread
* [RFC PATCH net-next 3/4] net: phy: added the LAN937x phy support 2022-02-28 14:05 [RFC PATCH net-next 0/4] Add support for LAN937x T1 Phy Arun Ramadoss 2022-02-28 14:05 ` [RFC PATCH net-next 1/4] net: phy: used the genphy_soft_reset for phy reset in Lan87xx Arun Ramadoss 2022-02-28 14:05 ` [RFC PATCH net-next 2/4] net: phy: updated the initialization routine for LAN87xx Arun Ramadoss @ 2022-02-28 14:05 ` Arun Ramadoss 2022-03-02 3:22 ` Andrew Lunn 2022-02-28 14:05 ` [RFC PATCH net-next 4/4] net: phy: added master-slave config and cable diagnostics for Lan937x Arun Ramadoss 3 siblings, 1 reply; 11+ messages in thread From: Arun Ramadoss @ 2022-02-28 14:05 UTC (permalink / raw) To: netdev, linux-kernel Cc: Andrew Lunn, Heiner Kallweit, Russell King, Jakub Kicinski, David S . Miller LAN937x T1 Phy is based on LAN87xx Phy, so reusing the init script of the Lan87xx. There is a workaround in accessing the DSP bank register for Lan937x Phy. Whenever there is a bank switch to DSP registers, then we need a dummy read access before proceeding to the actual register access. Signed-off-by: Prasanna Vengateshan <prasanna.vengateshan@microchip.com> Signed-off-by: Arun Ramadoss <arun.ramadoss@microchip.com> --- drivers/net/phy/microchip_t1.c | 47 +++++++++++++++++++++++++++++++--- 1 file changed, 44 insertions(+), 3 deletions(-) diff --git a/drivers/net/phy/microchip_t1.c b/drivers/net/phy/microchip_t1.c index 33325e5bd884..634a1423182a 100644 --- a/drivers/net/phy/microchip_t1.c +++ b/drivers/net/phy/microchip_t1.c @@ -10,6 +10,7 @@ #include <linux/ethtool_netlink.h> #define LAN87XX_PHY_ID 0x0007c150 +#define LAN937X_T1_PHY_ID 0x0007c181 #define MICROCHIP_PHY_ID_MASK 0xfffffff0 /* External Register Control Register */ @@ -76,8 +77,12 @@ #define T1_EQ_WT_FD_LCK_FRZ_CFG 0x6D #define T1_PST_EQ_LCK_STG1_FRZ_CFG 0x6E +#define T1_REG_BANK_SEL_MASK 0x7 +#define T1_REG_BANK_SEL 8 +#define T1_REG_ADDR_MASK 0xFF + #define DRIVER_AUTHOR "Nisar Sayed <nisar.sayed@microchip.com>" -#define DRIVER_DESC "Microchip LAN87XX T1 PHY driver" +#define DRIVER_DESC "Microchip LAN87XX/LAN937x T1 PHY driver" struct access_ereg_val { u8 mode; @@ -115,6 +120,32 @@ static int access_ereg(struct phy_device *phydev, u8 mode, u8 bank, ereg |= (bank << 8) | offset; + /* DSP bank access workaround for lan937x */ + if (phydev->phy_id == LAN937X_T1_PHY_ID) { + u8 prev_bank; + u16 val; + + /* Read previous selected bank */ + rc = phy_read(phydev, LAN87XX_EXT_REG_CTL); + if (rc < 0) + return rc; + + /* store the prev_bank */ + prev_bank = (rc >> T1_REG_BANK_SEL) & T1_REG_BANK_SEL_MASK; + + if (bank != prev_bank && bank == PHYACC_ATTR_BANK_DSP) { + val = ereg & ~T1_REG_ADDR_MASK; + + val &= ~LAN87XX_EXT_REG_CTL_WR_CTL; + val |= LAN87XX_EXT_REG_CTL_RD_CTL; + + /* access twice for DSP bank change,dummy access */ + rc = phy_write(phydev, LAN87XX_EXT_REG_CTL, val); + if (rc < 0) + return rc; + } + } + rc = phy_write(phydev, LAN87XX_EXT_REG_CTL, ereg); if (rc < 0) return rc; @@ -397,7 +428,7 @@ static irqreturn_t lan87xx_handle_interrupt(struct phy_device *phydev) return IRQ_HANDLED; } -static int lan87xx_config_init(struct phy_device *phydev) +static int lan_phy_config_init(struct phy_device *phydev) { int rc = lan87xx_phy_init(phydev); @@ -642,13 +673,22 @@ static struct phy_driver microchip_t1_phy_driver[] = { .name = "LAN87xx T1", .flags = PHY_POLL_CABLE_TEST, .features = PHY_BASIC_T1_FEATURES, - .config_init = lan87xx_config_init, + .config_init = lan_phy_config_init, .config_intr = lan87xx_phy_config_intr, .handle_interrupt = lan87xx_handle_interrupt, .suspend = genphy_suspend, .resume = genphy_resume, .cable_test_start = lan87xx_cable_test_start, .cable_test_get_status = lan87xx_cable_test_get_status, + }, + { + .phy_id = LAN937X_T1_PHY_ID, + .phy_id_mask = MICROCHIP_PHY_ID_MASK, + .name = "LAN937x T1", + .features = PHY_BASIC_T1_FEATURES, + .config_init = lan_phy_config_init, + .suspend = genphy_suspend, + .resume = genphy_resume, } }; @@ -656,6 +696,7 @@ module_phy_driver(microchip_t1_phy_driver); static struct mdio_device_id __maybe_unused microchip_t1_tbl[] = { { LAN87XX_PHY_ID, MICROCHIP_PHY_ID_MASK}, + { LAN937X_T1_PHY_ID, MICROCHIP_PHY_ID_MASK}, { } }; -- 2.33.0 ^ permalink raw reply related [flat|nested] 11+ messages in thread
* Re: [RFC PATCH net-next 3/4] net: phy: added the LAN937x phy support 2022-02-28 14:05 ` [RFC PATCH net-next 3/4] net: phy: added the LAN937x phy support Arun Ramadoss @ 2022-03-02 3:22 ` Andrew Lunn 2022-03-02 9:57 ` Arun.Ramadoss 0 siblings, 1 reply; 11+ messages in thread From: Andrew Lunn @ 2022-03-02 3:22 UTC (permalink / raw) To: Arun Ramadoss Cc: netdev, linux-kernel, Heiner Kallweit, Russell King, Jakub Kicinski, David S . Miller On Mon, Feb 28, 2022 at 07:35:09PM +0530, Arun Ramadoss wrote: > LAN937x T1 Phy is based on LAN87xx Phy, so reusing the init script of > the Lan87xx. There is a workaround in accessing the DSP bank register > for Lan937x Phy. Whenever there is a bank switch to DSP registers, then > we need a dummy read access before proceeding to the actual register > access. > > Signed-off-by: Prasanna Vengateshan <prasanna.vengateshan@microchip.com> > Signed-off-by: Arun Ramadoss <arun.ramadoss@microchip.com> > --- > drivers/net/phy/microchip_t1.c | 47 +++++++++++++++++++++++++++++++--- > 1 file changed, 44 insertions(+), 3 deletions(-) > > diff --git a/drivers/net/phy/microchip_t1.c b/drivers/net/phy/microchip_t1.c > index 33325e5bd884..634a1423182a 100644 > --- a/drivers/net/phy/microchip_t1.c > +++ b/drivers/net/phy/microchip_t1.c > @@ -10,6 +10,7 @@ > #include <linux/ethtool_netlink.h> > > #define LAN87XX_PHY_ID 0x0007c150 > +#define LAN937X_T1_PHY_ID 0x0007c181 I guess the last 1 is meaningless, given the mask? Andrew ^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [RFC PATCH net-next 3/4] net: phy: added the LAN937x phy support 2022-03-02 3:22 ` Andrew Lunn @ 2022-03-02 9:57 ` Arun.Ramadoss 0 siblings, 0 replies; 11+ messages in thread From: Arun.Ramadoss @ 2022-03-02 9:57 UTC (permalink / raw) To: andrew; +Cc: linux, netdev, linux-kernel, davem, hkallweit1, kuba On Wed, 2022-03-02 at 04:22 +0100, Andrew Lunn wrote: > EXTERNAL EMAIL: Do not click links or open attachments unless you > know the content is safe > > On Mon, Feb 28, 2022 at 07:35:09PM +0530, Arun Ramadoss wrote: > > LAN937x T1 Phy is based on LAN87xx Phy, so reusing the init script > > of > > the Lan87xx. There is a workaround in accessing the DSP bank > > register > > for Lan937x Phy. Whenever there is a bank switch to DSP registers, > > then > > we need a dummy read access before proceeding to the actual > > register > > access. > > > > Signed-off-by: Prasanna Vengateshan < > > prasanna.vengateshan@microchip.com> > > Signed-off-by: Arun Ramadoss <arun.ramadoss@microchip.com> > > --- > > drivers/net/phy/microchip_t1.c | 47 > > +++++++++++++++++++++++++++++++--- > > 1 file changed, 44 insertions(+), 3 deletions(-) > > > > diff --git a/drivers/net/phy/microchip_t1.c > > b/drivers/net/phy/microchip_t1.c > > index 33325e5bd884..634a1423182a 100644 > > --- a/drivers/net/phy/microchip_t1.c > > +++ b/drivers/net/phy/microchip_t1.c > > @@ -10,6 +10,7 @@ > > #include <linux/ethtool_netlink.h> > > > > #define LAN87XX_PHY_ID 0x0007c150 > > +#define LAN937X_T1_PHY_ID 0x0007c181 > > I guess the last 1 is meaningless, given the mask? Thanks Andrew for the comment. I will make 3:0 bits of phy_id to 0. > > Andrew ^ permalink raw reply [flat|nested] 11+ messages in thread
* [RFC PATCH net-next 4/4] net: phy: added master-slave config and cable diagnostics for Lan937x 2022-02-28 14:05 [RFC PATCH net-next 0/4] Add support for LAN937x T1 Phy Arun Ramadoss ` (2 preceding siblings ...) 2022-02-28 14:05 ` [RFC PATCH net-next 3/4] net: phy: added the LAN937x phy support Arun Ramadoss @ 2022-02-28 14:05 ` Arun Ramadoss 2022-02-28 14:12 ` Russell King (Oracle) 3 siblings, 1 reply; 11+ messages in thread From: Arun Ramadoss @ 2022-02-28 14:05 UTC (permalink / raw) To: netdev, linux-kernel Cc: Andrew Lunn, Heiner Kallweit, Russell King, Jakub Kicinski, David S . Miller To configure the Lan937x T1 phy as master or slave using the ethtool -s <dev> master-slave <forced-master/forced-slave>, the config_aneg and read status functions are added. And for the cable-diagnostics, used the lan87xx routines. Signed-off-by: Prasanna Vengateshan <prasanna.vengateshan@microchip.com> Signed-off-by: Arun Ramadoss <arun.ramadoss@microchip.com> --- drivers/net/phy/microchip_t1.c | 75 ++++++++++++++++++++++++++++++++++ 1 file changed, 75 insertions(+) diff --git a/drivers/net/phy/microchip_t1.c b/drivers/net/phy/microchip_t1.c index 634a1423182a..3a0d4c4fab0a 100644 --- a/drivers/net/phy/microchip_t1.c +++ b/drivers/net/phy/microchip_t1.c @@ -81,6 +81,9 @@ #define T1_REG_BANK_SEL 8 #define T1_REG_ADDR_MASK 0xFF +#define T1_MODE_STAT_REG 0x11 +#define T1_LINK_UP_MSK BIT(0) + #define DRIVER_AUTHOR "Nisar Sayed <nisar.sayed@microchip.com>" #define DRIVER_DESC "Microchip LAN87XX/LAN937x T1 PHY driver" @@ -435,6 +438,11 @@ static int lan_phy_config_init(struct phy_device *phydev) if (rc < 0) phydev_err(phydev, "failed to initialize phy\n"); + phydev->duplex = DUPLEX_FULL; + phydev->speed = SPEED_100; + phydev->pause = 0; + phydev->asym_pause = 0; + return rc < 0 ? rc : 0; } @@ -666,6 +674,69 @@ static int lan87xx_cable_test_get_status(struct phy_device *phydev, return 0; } +static int lan937x_read_status(struct phy_device *phydev) +{ + int rc; + + rc = phy_read(phydev, T1_MODE_STAT_REG); + if (rc < 0) + return rc; + + if (rc & T1_LINK_UP_MSK) + phydev->link = 1; + else + phydev->link = 0; + + phydev->master_slave_get = MASTER_SLAVE_CFG_UNKNOWN; + phydev->master_slave_state = MASTER_SLAVE_STATE_UNKNOWN; + + rc = phy_read(phydev, MII_CTRL1000); + if (rc < 0) + return rc; + + if (rc & CTL1000_AS_MASTER) + phydev->master_slave_get = MASTER_SLAVE_CFG_MASTER_FORCE; + else + phydev->master_slave_get = MASTER_SLAVE_CFG_SLAVE_FORCE; + + rc = phy_read(phydev, MII_STAT1000); + if (rc < 0) + return rc; + + if (rc & LPA_1000MSRES) + phydev->master_slave_state = MASTER_SLAVE_STATE_MASTER; + else + phydev->master_slave_state = MASTER_SLAVE_STATE_SLAVE; + + return 0; +} + +static int lan937x_config_aneg(struct phy_device *phydev) +{ + int rc; + u16 ctl = 0; + + switch (phydev->master_slave_set) { + case MASTER_SLAVE_CFG_MASTER_FORCE: + ctl |= CTL1000_AS_MASTER; + break; + case MASTER_SLAVE_CFG_SLAVE_FORCE: + break; + case MASTER_SLAVE_CFG_UNKNOWN: + case MASTER_SLAVE_CFG_UNSUPPORTED: + return 0; + default: + phydev_warn(phydev, "Unsupported Master/Slave mode\n"); + return -EOPNOTSUPP; + } + + rc = phy_modify_changed(phydev, MII_CTRL1000, CTL1000_AS_MASTER, ctl); + if (rc == 1) + rc = genphy_soft_reset(phydev); + + return rc; +} + static struct phy_driver microchip_t1_phy_driver[] = { { .phy_id = LAN87XX_PHY_ID, @@ -689,6 +760,10 @@ static struct phy_driver microchip_t1_phy_driver[] = { .config_init = lan_phy_config_init, .suspend = genphy_suspend, .resume = genphy_resume, + .config_aneg = lan937x_config_aneg, + .read_status = lan937x_read_status, + .cable_test_start = lan87xx_cable_test_start, + .cable_test_get_status = lan87xx_cable_test_get_status, } }; -- 2.33.0 ^ permalink raw reply related [flat|nested] 11+ messages in thread
* Re: [RFC PATCH net-next 4/4] net: phy: added master-slave config and cable diagnostics for Lan937x 2022-02-28 14:05 ` [RFC PATCH net-next 4/4] net: phy: added master-slave config and cable diagnostics for Lan937x Arun Ramadoss @ 2022-02-28 14:12 ` Russell King (Oracle) 2022-03-01 16:59 ` Arun.Ramadoss 0 siblings, 1 reply; 11+ messages in thread From: Russell King (Oracle) @ 2022-02-28 14:12 UTC (permalink / raw) To: Arun Ramadoss Cc: netdev, linux-kernel, Andrew Lunn, Heiner Kallweit, Jakub Kicinski, David S . Miller On Mon, Feb 28, 2022 at 07:35:10PM +0530, Arun Ramadoss wrote: > To configure the Lan937x T1 phy as master or slave using the ethtool -s > <dev> master-slave <forced-master/forced-slave>, the config_aneg and > read status functions are added. And for the cable-diagnostics, used the > lan87xx routines. > > Signed-off-by: Prasanna Vengateshan <prasanna.vengateshan@microchip.com> > Signed-off-by: Arun Ramadoss <arun.ramadoss@microchip.com> > --- > drivers/net/phy/microchip_t1.c | 75 ++++++++++++++++++++++++++++++++++ > 1 file changed, 75 insertions(+) > > diff --git a/drivers/net/phy/microchip_t1.c b/drivers/net/phy/microchip_t1.c > index 634a1423182a..3a0d4c4fab0a 100644 > --- a/drivers/net/phy/microchip_t1.c > +++ b/drivers/net/phy/microchip_t1.c > @@ -81,6 +81,9 @@ > #define T1_REG_BANK_SEL 8 > #define T1_REG_ADDR_MASK 0xFF > > +#define T1_MODE_STAT_REG 0x11 > +#define T1_LINK_UP_MSK BIT(0) > + > #define DRIVER_AUTHOR "Nisar Sayed <nisar.sayed@microchip.com>" > #define DRIVER_DESC "Microchip LAN87XX/LAN937x T1 PHY driver" > > @@ -435,6 +438,11 @@ static int lan_phy_config_init(struct phy_device *phydev) > if (rc < 0) > phydev_err(phydev, "failed to initialize phy\n"); > > + phydev->duplex = DUPLEX_FULL; > + phydev->speed = SPEED_100; > + phydev->pause = 0; > + phydev->asym_pause = 0; Shouldn't this be done in lan937x_read_status()? Have you tested this patch with various invocations of ethtool -s ? E.g. autoneg on, autoneg off at various forced speeds, both suitable for the PHY and unsuitable? Are all these sensibly handled? Thanks. -- RMK's Patch system: https://www.armlinux.org.uk/developer/patches/ FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last! ^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: [RFC PATCH net-next 4/4] net: phy: added master-slave config and cable diagnostics for Lan937x 2022-02-28 14:12 ` Russell King (Oracle) @ 2022-03-01 16:59 ` Arun.Ramadoss 0 siblings, 0 replies; 11+ messages in thread From: Arun.Ramadoss @ 2022-03-01 16:59 UTC (permalink / raw) To: linux; +Cc: netdev, linux-kernel, davem, hkallweit1, andrew, kuba On Mon, 2022-02-28 at 14:12 +0000, Russell King (Oracle) wrote: > EXTERNAL EMAIL: Do not click links or open attachments unless you > know the content is safe > > On Mon, Feb 28, 2022 at 07:35:10PM +0530, Arun Ramadoss wrote: > > To configure the Lan937x T1 phy as master or slave using the > > ethtool -s > > <dev> master-slave <forced-master/forced-slave>, the config_aneg > > and > > read status functions are added. And for the cable-diagnostics, > > used the > > lan87xx routines. > > > > Signed-off-by: Prasanna Vengateshan < > > prasanna.vengateshan@microchip.com> > > Signed-off-by: Arun Ramadoss <arun.ramadoss@microchip.com> > > --- > > drivers/net/phy/microchip_t1.c | 75 > > ++++++++++++++++++++++++++++++++++ > > 1 file changed, 75 insertions(+) > > > > diff --git a/drivers/net/phy/microchip_t1.c > > b/drivers/net/phy/microchip_t1.c > > index 634a1423182a..3a0d4c4fab0a 100644 > > --- a/drivers/net/phy/microchip_t1.c > > +++ b/drivers/net/phy/microchip_t1.c > > @@ -81,6 +81,9 @@ > > #define T1_REG_BANK_SEL 8 > > #define T1_REG_ADDR_MASK 0xFF > > > > +#define T1_MODE_STAT_REG 0x11 > > +#define T1_LINK_UP_MSK BIT(0) > > + > > #define DRIVER_AUTHOR "Nisar Sayed < > > nisar.sayed@microchip.com>" > > #define DRIVER_DESC "Microchip LAN87XX/LAN937x T1 PHY driver" > > > > @@ -435,6 +438,11 @@ static int lan_phy_config_init(struct > > phy_device *phydev) > > if (rc < 0) > > phydev_err(phydev, "failed to initialize phy\n"); > > > > + phydev->duplex = DUPLEX_FULL; > > + phydev->speed = SPEED_100; > > + phydev->pause = 0; > > + phydev->asym_pause = 0; > > Shouldn't this be done in lan937x_read_status()? > > Have you tested this patch with various invocations of ethtool -s ? > E.g. autoneg on, autoneg off at various forced speeds, both suitable > for the PHY and unsuitable? Are all these sensibly handled? > Thanks for the comment. LAN937x is 100BaseT1 Phy and it doesn't provide option for changing speed and autonegotiation. Since it is fixed and not going to change using ethtool -s command, I added it in the config_init routine. I took the drivers/net/phy/nxp-tja11xx as reference in which speed is initialised in config_init routine. I tested the patch using the ethtool -s for changing the master/slave configuration. For other change like speed/duplex setting, driver will not make change. When I execute ethtool <dev>, it returns the default value 100Mbps, Full duplex. If lan937x_read_status( ) is the suitable place, then I will move it there. > Thanks. > > -- > RMK's Patch system: https://www.armlinux.org.uk/developer/patches/ > FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last! ^ permalink raw reply [flat|nested] 11+ messages in thread
end of thread, other threads:[~2022-03-02 10:04 UTC | newest] Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2022-02-28 14:05 [RFC PATCH net-next 0/4] Add support for LAN937x T1 Phy Arun Ramadoss 2022-02-28 14:05 ` [RFC PATCH net-next 1/4] net: phy: used the genphy_soft_reset for phy reset in Lan87xx Arun Ramadoss 2022-03-02 3:19 ` Andrew Lunn 2022-03-02 10:04 ` Arun.Ramadoss 2022-02-28 14:05 ` [RFC PATCH net-next 2/4] net: phy: updated the initialization routine for LAN87xx Arun Ramadoss 2022-02-28 14:05 ` [RFC PATCH net-next 3/4] net: phy: added the LAN937x phy support Arun Ramadoss 2022-03-02 3:22 ` Andrew Lunn 2022-03-02 9:57 ` Arun.Ramadoss 2022-02-28 14:05 ` [RFC PATCH net-next 4/4] net: phy: added master-slave config and cable diagnostics for Lan937x Arun Ramadoss 2022-02-28 14:12 ` Russell King (Oracle) 2022-03-01 16:59 ` Arun.Ramadoss
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).