net: phy: correctly model the PHY voltage supply in DT
Subsystems (12)
cover net: phy: correctly model the PHY voltage supply in DT
From: Bartosz Golaszewski <[email protected]>
PHY devices on an MDIO bus can have their own regulators. This is currently mostly modeled using the semi-standard phy-supply property on the MAC node. This seems to be conceptually wrong though, as the MAC shouldn't need to care about enabling the PHY regulator explicitly. Instead this should be done by the core PHY/MDIO code.
This series introduces a new DT property at the PHY node level in mdio.yaml and adds support for PHY regulator, then uses the new property on pumpkin boards. It also addresses several issues I noticed when working on this feature.
First four patches are just cosmetic improvements in source files we'll modify later in this series.
Patches 5 and 6 modify the way PHY reset handling works. Currently the probe() callback needs to be implemented to take the PHY out of reset but PHY drivers without probe() can have resets defined as well.
Patches 7-11 address an issue where we probe the PHY for ID without deasserting its reset signal. We delay the ID read until after the logical device is created and resets have been configured.
Last four patches add the new DT property, implement support for PHY regulator in phy and mdio code and set the new property in the DT file for MediaTek's pumpkin boards.
Bartosz Golaszewski (15):
net: phy: arrange headers in mdio_bus.c alphabetically
net: phy: arrange headers in mdio_device.c alphabetically
net: phy: arrange headers in phy_device.c alphabetically
net: mdio: add a forward declaration for reset_control to mdio.h
net: phy: reset the PHY even if probe() is not implemented
net: phy: mdio: reset MDIO devices even if probe() is not implemented
net: phy: split out the PHY driver request out of phy_device_create()
net: phy: check the PHY presence in get_phy_id()
net: phy: delay PHY driver probe until PHY registration
net: phy: simplify phy_device_create()
net: phy: drop get_phy_device()
dt-bindings: mdio: add phy-supply property to ethernet phy node
net: phy: mdio: add support for PHY supply regulator
net: phy: add PHY regulator support
ARM64: dts: mediatek: add a phy regulator to pumpkin-common.dtsi
.../devicetree/bindings/net/mdio.yaml | 4 + .../boot/dts/mediatek/pumpkin-common.dtsi | 1 + drivers/net/dsa/ocelot/felix_vsc9959.c | 3 +- drivers/net/ethernet/amd/xgbe/xgbe-phy-v2.c | 5 +- .../net/ethernet/hisilicon/hns/hns_dsaf_mac.c | 2 +- drivers/net/ethernet/socionext/netsec.c | 3 +- drivers/net/phy/fixed_phy.c | 2 +- drivers/net/phy/mdio-xgene.c | 2 +- drivers/net/phy/mdio_bus.c | 55 +++-- drivers/net/phy/mdio_device.c | 51 ++++- drivers/net/phy/nxp-tja11xx.c | 2 +- drivers/net/phy/phy_device.c | 216 ++++++++++-------- drivers/net/phy/sfp.c | 2 +- drivers/of/of_mdio.c | 11 +- include/linux/mdio.h | 4 + include/linux/phy.h | 21 +- 16 files changed, 240 insertions(+), 144 deletions(-)
Signature · 1 line
2.26.1
PATCH 2/15 net: phy: arrange headers in mdio_device.c alphabetically
From: Bartosz Golaszewski <[email protected]>
Keeping the headers in alphabetical order is better for readability and allows to easily see if given header is already included.
--- drivers/net/phy/mdio_device.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/phy/mdio_device.c +1 -1
@@ -6,6 +6,7 @@#define pr_fmt (fmt) KBUILD_MODNAME ": " fmt +#include <linux/delay.h> #include <linux/errno.h> #include <linux/gpio.h> #include <linux/gpio/consumer.h>
@@ -20,7 +21,6 @@#include <linux/slab.h> #include <linux/string.h> #include <linux/unistd.h> -#include <linux/delay.h> void mdio_device_free (struct mdio_device *mdiodev) {
Signature · 1 line
2.26.1
PATCH 3/15 net: phy: arrange headers in phy_device.c alphabetically
In reply to Bartosz GolaszewskiFrom: Bartosz Golaszewski <[email protected]>
Keeping the headers in alphabetical order is better for readability and allows to easily see if given header is already included.
--- drivers/net/phy/phy_device.c | 28 ++++++++++++++-------------- 1 file changed, 14 insertions(+), 14 deletions(-)
drivers/net/phy/phy_device.c +14 -14
@@ -9,28 +9,28 @@#define pr_fmt (fmt) KBUILD_MODNAME ": " fmt -#include <linux/kernel.h> -#include <linux/string.h> -#include <linux/errno.h> -#include <linux/unistd.h> -#include <linux/slab.h> -#include <linux/interrupt.h> -#include <linux/init.h> +#include <linux/bitmap.h> #include <linux/delay.h> -#include <linux/netdevice.h> +#include <linux/errno.h> #include <linux/etherdevice.h> -#include <linux/skbuff.h> +#include <linux/ethtool.h> +#include <linux/init.h> +#include <linux/interrupt.h> +#include <linux/io.h> +#include <linux/kernel.h> +#include <linux/mdio.h> +#include <linux/mii.h> #include <linux/mm.h> #include <linux/module.h> -#include <linux/mii.h> -#include <linux/ethtool.h> -#include <linux/bitmap.h> +#include <linux/netdevice.h> #include <linux/phy.h> #include <linux/phy_led_triggers.h> #include <linux/sfp.h> -#include <linux/mdio.h> -#include <linux/io.h> +#include <linux/skbuff.h> +#include <linux/slab.h> +#include <linux/string.h> #include <linux/uaccess.h> +#include <linux/unistd.h> MODULE_DESCRIPTION ("PHY library" );MODULE_AUTHOR ("Andy Fleming" );
Signature · 1 line
2.26.1
PATCH 4/15 net: mdio: add a forward declaration for reset_control to mdio.h
In reply to Bartosz GolaszewskiFrom: Bartosz Golaszewski <[email protected]>
This header refers to struct reset_control but doesn't include any reset header. The structure definition is probably somehow indirectly pulled in since no warnings are reported but for the sake of correctness add the forward declaration for struct reset_control.
--- include/linux/mdio.h | 1 + 1 file changed, 1 insertion(+)
include/linux/mdio.h +1 -0
@@ -17,6 +17,7 @@#define MII_REGADDR_C45_MASK GENMASK(15, 0)struct gpio_desc ; +struct reset_control ;struct mii_bus ;/* Multiple levels of nesting are possible. However typically this is
Signature · 1 line
2.26.1
PATCH 11/15 net: phy: drop get_phy_device()
In reply to Bartosz GolaszewskiFrom: Bartosz Golaszewski <[email protected]>
get_phy_device() has now become just a wrapper for phy_device_create() with the phy_id argument set to PHY_ID_NONE. Let's remove this function treewide and replace it with opencoded phy_device_create(). This has the advantage of being more explicit about the PHY not having the ID set when being created.
--- drivers/net/dsa/ocelot/felix_vsc9959.c | 3 ++- drivers/net/ethernet/amd/xgbe/xgbe-phy-v2.c | 5 +++-- drivers/net/ethernet/hisilicon/hns/hns_dsaf_mac.c | 2 +- drivers/net/ethernet/socionext/netsec.c | 3 ++- drivers/net/phy/fixed_phy.c | 2 +- drivers/net/phy/mdio-xgene.c | 2 +- drivers/net/phy/mdio_bus.c | 2 +- drivers/net/phy/phy_device.c | 14 -------------- drivers/net/phy/sfp.c | 2 +- drivers/of/of_mdio.c | 11 +++++++---- include/linux/phy.h | 7 ------- 11 files changed, 19 insertions(+), 34 deletions(-)
drivers/net/dsa/ocelot/felix_vsc9959.c +2 -1
@@ -1183,7 +1183,8 @@ static int vsc9959_mdio_bus_alloc(struct ocelot *ocelot)if (ocelot_port->phy_mode ==PHY_INTERFACE_MODE_USXGMII ) is_c45 = true; - pcs =get_phy_device (felix->imdio, port, is_c45); + pcs =phy_device_create (felix->imdio, port, +PHY_ID_NONE , is_c45);if (IS_ERR (pcs))continue ;
drivers/net/ethernet/amd/xgbe/xgbe-phy-v2.c +3 -2
@@ -1017,8 +1017,9 @@ static int xgbe_phy_find_phy_device(struct xgbe_prv_data *pdata) }/* Create and connect to the PHY device */ - phydev =get_phy_device (phy_data->mii, phy_data->mdio_addr, - (phy_data->phydev_mode ==XGBE_MDIO_MODE_CL45 )); + phydev =phy_device_create (phy_data->mii, phy_data->mdio_addr, +PHY_ID_NONE , + phy_data->phydev_mode ==XGBE_MDIO_MODE_CL45 );if (IS_ERR (phydev)) {netdev_err (pdata->netdev,"get_phy_device failed\n" );return -ENODEV ;
drivers/net/ethernet/hisilicon/hns/hns_dsaf_mac.c +1 -1
@@ -703,7 +703,7 @@ hns_mac_register_phydev(struct mii_bus *mdio, struct hns_mac_cb *mac_cb,else return -ENODATA ; - phy =get_phy_device (mdio, addr, is_c45); + phy =phy_device_create (mdio, addr,PHY_ID_NONE , is_c45);if (!phy ||IS_ERR (phy))return -EIO ;
drivers/net/ethernet/socionext/netsec.c +2 -1
@@ -1935,7 +1935,8 @@ static int netsec_register_mdio(struct netsec_priv *priv, u32 phy_addr)return ret; } - priv->phydev =get_phy_device (bus, phy_addr, false); + priv->phydev =phy_device_create (bus, phy_addr, +PHY_ID_NONE , false);if (IS_ERR (priv->phydev)) { ret =PTR_ERR (priv->phydev);dev_err (priv->dev,"get_phy_device err(%d)\n" , ret);
drivers/net/phy/fixed_phy.c +1 -1
@@ -254,7 +254,7 @@ static struct phy_device *__fixed_phy_register(unsigned int irq,return ERR_PTR (ret); } - phy =get_phy_device (fmb->mii_bus, phy_addr, false); + phy =phy_device_create (fmb->mii_bus, phy_addr,PHY_ID_NONE , false);if (IS_ERR (phy)) {fixed_phy_del (phy_addr);return ERR_PTR (-EINVAL );
drivers/net/phy/mdio-xgene.c +1 -1
@@ -264,7 +264,7 @@ struct phy_device *xgene_enet_phy_register(struct mii_bus *bus, int phy_addr) {struct phy_device *phy_dev; - phy_dev =get_phy_device (bus, phy_addr, false); + phy_dev =phy_device_create (bus, phy_addr,PHY_ID_NONE , false);if (!phy_dev ||IS_ERR (phy_dev))return NULL ;
drivers/net/phy/mdio_bus.c +1 -1
@@ -742,7 +742,7 @@ struct phy_device *mdiobus_scan(struct mii_bus *bus, int addr)struct phy_device *phydev;int err; - phydev =get_phy_device (bus, addr, false); + phydev =phy_device_create (bus, addr,PHY_ID_NONE , false);if (IS_ERR (phydev))return phydev;
drivers/net/phy/phy_device.c +0 -14
@@ -817,20 +817,6 @@ static int phy_device_read_id(struct phy_device *phydev) phydev->is_c45, &phydev->c45_ids); } -/** -* get_phy_device - create a phy_device withoug PHY ID -* @bus: the target MII bus -* @addr: PHY address on the MII bus -* @is_c45: If true the PHY uses the 802.3 clause 45 protocol -* -* Allocates a new phy_device for @addr on the @bus. -*/ -struct phy_device *get_phy_device (struct mii_bus *bus,int addr,bool is_c45) -{ -return phy_device_create (bus, addr,PHY_ID_NONE , is_c45); -} -EXPORT_SYMBOL (get_phy_device); -/** * phy_device_register - Register the phy device on the MDIO bus * @phydev: phy_device structure to be added to the MDIO bus
drivers/net/phy/sfp.c +1 -1
@@ -1431,7 +1431,7 @@ static int sfp_sm_probe_phy(struct sfp *sfp, bool is_c45)struct phy_device *phy;int err; - phy =get_phy_device (sfp->i2c_mii,SFP_PHY_ADDR , is_c45); + phy =phy_device_create (sfp->i2c_mii,SFP_PHY_ADDR ,PHY_ID_NONE , is_c45);if (phy ==ERR_PTR (-ENODEV ))return PTR_ERR (phy);if (IS_ERR (phy)) {
drivers/of/of_mdio.c +7 -4
@@ -120,10 +120,13 @@ static int of_mdiobus_register_phy(struct mii_bus *mdio, is_c45 =of_device_is_compatible (child,"ethernet-phy-ieee802.3-c45" ); -if (!is_c45 && !of_get_phy_id (child, &phy_id)) - phy =phy_device_create (mdio, addr, phy_id,0 ); -else - phy =get_phy_device (mdio, addr, is_c45); +if (!is_c45) { + rc =of_get_phy_id (child, &phy_id); +if (rc) + phy_id =PHY_ID_NONE ; + } + + phy =phy_device_create (mdio, addr, phy_id, is_c45);if (IS_ERR (phy)) {if (mii_ts)unregister_mii_timestamper (mii_ts);
include/linux/phy.h +0 -7
@@ -1215,16 +1215,9 @@ int phy_modify_paged(struct phy_device *phydev, int page, u32 regnum,struct phy_device *phy_device_create (struct mii_bus *bus,int addr,u32 phy_id,bool is_c45);#if IS_ENABLED (CONFIG_PHYLIB ) -struct phy_device *get_phy_device (struct mii_bus *bus,int addr,bool is_c45);int phy_device_register (struct phy_device *phy);void phy_device_free (struct phy_device *phydev);#else -static inline -struct phy_device *get_phy_device (struct mii_bus *bus,int addr,bool is_c45) -{ -return NULL ; -} -static inline int phy_device_register (struct phy_device *phy) {return 0 ;
Signature · 1 line
2.26.1
PATCH 12/15 dt-bindings: mdio: add phy-supply property to ethernet phy node
In reply to Bartosz GolaszewskiFrom: Bartosz Golaszewski <[email protected]>
The phy-supply property is often added to MAC nodes but this is wrong conceptually. These supplies should be part of the PHY node on the MDIO bus. Add phy-supply property at PHY level to mdio.yaml.
--- Documentation/devicetree/bindings/net/mdio.yaml | 4 ++++ 1 file changed, 4 insertions(+)
Documentation/devicetree/bindings/net/mdio.yaml +4 -0
@@ -90,6 +90,10 @@ patternProperties: Delay after the reset was deasserted in microseconds. If this property is missing the delay will be skipped. + phy-supply: + description: + Phandle to the regulator that provides the supply voltage to the PHY. + required: - reg
Signature · 1 line
2.26.1
PATCH 8/15 net: phy: check the PHY presence in get_phy_id()
In reply to Bartosz GolaszewskiFrom: Bartosz Golaszewski <[email protected]>
get_phy_id() is only called from get_phy_device() so the check for the 0x1fffffff value can be pulled into the former. This way it'll be easier to remove get_phy_device() later on.
--- drivers/net/phy/phy_device.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
drivers/net/phy/phy_device.c +4 -4
@@ -806,6 +806,10 @@ static int get_phy_id(struct mii_bus *bus, int addr, u32 *phy_id, *phy_id |= phy_reg; +/* If the phy_id is mostly Fs, there is no device there */ +if ((*phy_id &0x1fffffff ) ==0x1fffffff ) +return -ENODEV ; +return 0 ; }
@@ -832,10 +836,6 @@ struct phy_device *get_phy_device(struct mii_bus *bus, int addr, bool is_c45)if (r)return ERR_PTR (r); -/* If the phy_id is mostly Fs, there is no device there */ -if ((phy_id &0x1fffffff ) ==0x1fffffff ) -return ERR_PTR (-ENODEV ); -return phy_device_create (bus, addr, phy_id, is_c45, &c45_ids); }EXPORT_SYMBOL (get_phy_device);
Signature · 1 line
2.26.1
PATCH 14/15 net: phy: add PHY regulator support
In reply to Bartosz GolaszewskiFrom: Bartosz Golaszewski <[email protected]>
The MDIO sub-system now supports PHY regulators. Let's reuse the code to extend this support over to the PHY device.
--- drivers/net/phy/phy_device.c | 49 ++++++++++++++++++++++++++++-------- include/linux/phy.h | 10 ++++++++ 2 files changed, 48 insertions(+), 11 deletions(-)
drivers/net/phy/phy_device.c +38 -11
@@ -827,7 +827,12 @@ int phy_device_register(struct phy_device *phydev) err =mdiobus_register_device (&phydev->mdio);if (err) -return err; + goto err_out; + +/* Enable the PHY regulator */ + err =phy_device_power_on (phydev); +if (err) + goto err_unregister_mdio;/* Deassert the reset signal */ phy_device_reset (phydev,0 );
@@ -846,22 +851,25 @@ int phy_device_register(struct phy_device *phydev) err =phy_scan_fixups (phydev);if (err) {phydev_err (phydev,"failed to initialize\n" ); - goto out; + goto err_reset; } err =device_add (&phydev->mdio.dev);if (err) {phydev_err (phydev,"failed to add\n" ); - goto out; + goto err_reset; }return 0 ; - out: +err_reset:/* Assert the reset signal */ phy_device_reset (phydev,1 ); - +/* Disable the PHY regulator */ +phy_device_power_off (phydev); +err_unregister_mdio:mdiobus_unregister_device (&phydev->mdio); +err_out:return err; }EXPORT_SYMBOL (phy_device_register);
@@ -883,6 +891,8 @@ void phy_device_remove(struct phy_device *phydev)/* Assert the reset signal */ phy_device_reset (phydev,1 ); +/* Disable the PHY regulator */ +phy_device_power_off (phydev);mdiobus_unregister_device (&phydev->mdio); }
@@ -1064,6 +1074,11 @@ int phy_init_hw(struct phy_device *phydev) {int ret =0 ; +/* Enable the PHY regulator */ + ret =phy_device_power_on (phydev); +if (ret) +return ret; +/* Deassert the reset signal */ phy_device_reset (phydev,0 );
@@ -1644,6 +1659,8 @@ void phy_detach(struct phy_device *phydev)/* Assert the reset signal */ phy_device_reset (phydev,1 ); +/* Disable the PHY regulator */ +phy_device_power_off (phydev); }EXPORT_SYMBOL (phy_detach);
@@ -2684,13 +2701,18 @@ static int phy_probe(struct device *dev)mutex_lock (&phydev->lock); +/* Enable the PHY regulator */ + err =phy_device_power_on (phydev); +if (err) + goto out; +/* Deassert the reset signal */ phy_device_reset (phydev,0 );if (phydev->drv->probe) { err = phydev->drv->probe (phydev);if (err) - goto out; + goto out_reset; }/* Start out supporting everything. Eventually,
@@ -2708,7 +2730,7 @@ static int phy_probe(struct device *dev) }if (err) - goto out; + goto out_reset;if (!linkmode_test_bit (ETHTOOL_LINK_MODE_Autoneg_BIT, phydev->supported))
@@ -2751,11 +2773,16 @@ static int phy_probe(struct device *dev)/* Set the state to READY by default */ phydev->state =PHY_READY ; -out: -/* Assert the reset signal */ -if (err) -phy_device_reset (phydev,1 ); +mutex_unlock (&phydev->lock); + +return 0 ; +out_reset: +/* Assert the reset signal */ +phy_device_reset (phydev,1 ); +/* Disable the PHY regulator */ +phy_device_power_off (phydev); +out:mutex_unlock (&phydev->lock);return err;
include/linux/phy.h +10 -0
@@ -1291,6 +1291,16 @@ static inline void phy_device_reset(struct phy_device *phydev, int value)mdio_device_reset (&phydev->mdio, value); } +static inline int phy_device_power_on (struct phy_device *phydev) +{ +return mdio_device_power_on (&phydev->mdio); +} + +static inline int phy_device_power_off (struct phy_device *phydev) +{ +return mdio_device_power_off (&phydev->mdio); +} + #define phydev_err(_phydev, format, args...) \ dev_err(&_phydev->mdio.dev, format, ##args)
Signature · 1 line
2.26.1
PATCH 15/15 ARM64: dts: mediatek: add a phy regulator to pumpkin-common.dtsi
In reply to Bartosz GolaszewskiFrom: Bartosz Golaszewski <[email protected]>
Add the appropriate supply to the PHY child-node on the MDIO bus.
--- arch/arm64/boot/dts/mediatek/pumpkin-common.dtsi | 1 + 1 file changed, 1 insertion(+)
arch/arm64/boot/dts/mediatek/pumpkin-common.dtsi +1 -0
@@ -206,6 +206,7 @@ mdio { eth_phy: ethernet-phy@0 { reg = <0>; + phy-supply = <&mt6392_vmch_reg>; }; }; };
Signature · 1 line
2.26.1
PATCH 13/15 net: phy: mdio: add support for PHY supply regulator
In reply to Bartosz GolaszewskiFrom: Bartosz Golaszewski <[email protected]>
Currently many MAC drivers control the regulator supplying the PHY but this is conceptually wrong. The regulator should be defined as a property of the PHY node on the MDIO bus and controlled by the MDIO sub-system.
Add support for an optional PHY regulator which will be enabled before optional deasserting of the reset signal.
--- drivers/net/phy/mdio_bus.c | 21 ++++++++++++++++++++ drivers/net/phy/mdio_device.c | 36 +++++++++++++++++++++++++++++++++++ include/linux/mdio.h | 3 +++ 3 files changed, 60 insertions(+)
drivers/net/phy/mdio_bus.c +21 -0
@@ -27,6 +27,7 @@#include <linux/of_gpio.h> #include <linux/of_mdio.h> #include <linux/phy.h> +#include <linux/regulator/consumer.h> #include <linux/reset.h> #include <linux/skbuff.h> #include <linux/slab.h>
@@ -67,6 +68,22 @@ static int mdiobus_register_reset(struct mdio_device *mdiodev)return 0 ; } +static int mdiobus_register_regulator (struct mdio_device *mdiodev) +{ +struct regulator *phy_supply; + + phy_supply =regulator_get_optional (&mdiodev->dev,"phy" ); +if (IS_ERR (phy_supply)) { +if (PTR_ERR (phy_supply) == -EPROBE_DEFER ) +return -EPROBE_DEFER ; + phy_supply =NULL ; + } + + mdiodev->phy_supply = phy_supply; + +return 0 ; +} +int mdiobus_register_device (struct mdio_device *mdiodev) {int err;
@@ -83,6 +100,10 @@ int mdiobus_register_device(struct mdio_device *mdiodev)if (err)return err; + err =mdiobus_register_regulator (mdiodev); +if (err) +return err; +/* Assert the reset signal */ mdio_device_reset (mdiodev,1 ); }
drivers/net/phy/mdio_device.c +36 -0
@@ -17,6 +17,7 @@#include <linux/mii.h> #include <linux/module.h> #include <linux/phy.h> +#include <linux/regulator/consumer.h> #include <linux/reset.h> #include <linux/slab.h> #include <linux/string.h>
@@ -136,6 +137,32 @@ void mdio_device_reset(struct mdio_device *mdiodev, int value) }EXPORT_SYMBOL (mdio_device_reset); +int mdio_device_power_on (struct mdio_device *mdiodev) +{ +int ret =0 ; + +if (mdiodev->phy_supply) +/* TODO We may need a delay here just like reset_assert_delay +* but since no user currently needs it, I'm not adding it just +* yet. +*/ + ret =regulator_enable (mdiodev->phy_supply); + +return ret; +} +EXPORT_SYMBOL (mdio_device_power_on); + +int mdio_device_power_off (struct mdio_device *mdiodev) +{ +int ret =0 ; + +if (mdiodev->phy_supply) + ret =regulator_disable (mdiodev->phy_supply); + +return ret; +} +EXPORT_SYMBOL (mdio_device_power_off); +/** * mdio_probe - probe an MDIO device * @dev: device to probe
@@ -150,6 +177,11 @@ static int mdio_probe(struct device *dev)struct mdio_driver *mdiodrv =to_mdio_driver (drv);int err =0 ; +/* Enable the power supply */ + err =mdio_device_power_on (mdiodev); +if (err) +return err; +/* Deassert the reset signal */ mdio_device_reset (mdiodev,0 );
@@ -158,6 +190,8 @@ static int mdio_probe(struct device *dev)if (err) {/* Assert the reset signal */ mdio_device_reset (mdiodev,1 ); +/* Disable the power supply */ +mdio_device_power_off (mdiodev); } }
@@ -175,6 +209,8 @@ static int mdio_remove(struct device *dev)/* Assert the reset signal */ mdio_device_reset (mdiodev,1 ); +/* Disable the power supply */ +mdio_device_power_off (mdiodev);return 0 ; }
include/linux/mdio.h +3 -0
@@ -46,6 +46,7 @@ struct mdio_device {int flags;struct gpio_desc *reset_gpio;struct reset_control *reset_ctrl; +struct regulator *phy_supply;unsigned int reset_assert_delay;unsigned int reset_deassert_delay; };
@@ -92,6 +93,8 @@ struct mdio_device *mdio_device_create(struct mii_bus *bus, int addr);int mdio_device_register (struct mdio_device *mdiodev);void mdio_device_remove (struct mdio_device *mdiodev);void mdio_device_reset (struct mdio_device *mdiodev,int value); +int mdio_device_power_on (struct mdio_device *mdiodev); +int mdio_device_power_off (struct mdio_device *mdiodev);int mdio_driver_register (struct mdio_driver *drv);void mdio_driver_unregister (struct mdio_driver *drv);int mdio_device_bus_match (struct device *dev,struct device_driver *drv);
Signature · 1 line
2.26.1
PATCH 5/15 net: phy: reset the PHY even if probe() is not implemented
In reply to Bartosz GolaszewskiFrom: Bartosz Golaszewski <[email protected]>
Currently we only call phy_device_reset() if the PHY driver implements the probe() callback. This is not mandatory and many drivers (e.g. realtek) don't need probe() for most devices but still can have reset GPIOs defined. There's no reason to depend on the presence of probe() here so pull the reset code out of the if clause.
--- drivers/net/phy/phy_device.c | 23 ++++++++++++----------- 1 file changed, 12 insertions(+), 11 deletions(-)
drivers/net/phy/phy_device.c +12 -11
@@ -2690,16 +2690,13 @@ static int phy_probe(struct device *dev)mutex_lock (&phydev->lock); -if (phydev->drv->probe) { -/* Deassert the reset signal */ -phy_device_reset (phydev,0 ); +/* Deassert the reset signal */ +phy_device_reset (phydev,0 ); +if (phydev->drv->probe) { err = phydev->drv->probe (phydev); -if (err) { -/* Assert the reset signal */ -phy_device_reset (phydev,1 ); +if (err) goto out; - } }/* Start out supporting everything. Eventually,
@@ -2761,6 +2758,10 @@ static int phy_probe(struct device *dev) phydev->state =PHY_READY ; out: +/* Assert the reset signal */ +if (err) +phy_device_reset (phydev,1 ); +mutex_unlock (&phydev->lock);return err;
@@ -2779,12 +2780,12 @@ static int phy_remove(struct device *dev)sfp_bus_del_upstream (phydev->sfp_bus); phydev->sfp_bus =NULL ; -if (phydev->drv && phydev->drv->remove) { +if (phydev->drv && phydev->drv->remove) phydev->drv->remove (phydev); -/* Assert the reset signal */ -phy_device_reset (phydev,1 ); - } +/* Assert the reset signal */ +phy_device_reset (phydev,1 ); + phydev->drv =NULL ;return 0 ;
Signature · 1 line
2.26.1
PATCH 10/15 net: phy: simplify phy_device_create()
In reply to Bartosz GolaszewskiFrom: Bartosz Golaszewski <[email protected]>
The last argument passed to phy_device_create() (c45_ids) is never used in current mainline outside of the core PHY code - it can only be configured when reading the PHY ID from phy_device_read_id().
Let's drop this argument treewide.
--- drivers/net/phy/nxp-tja11xx.c | 2 +- drivers/net/phy/phy_device.c | 8 +++----- drivers/of/of_mdio.c | 2 +- include/linux/phy.h | 3 +-- 4 files changed, 6 insertions(+), 9 deletions(-)
drivers/net/phy/nxp-tja11xx.c +1 -1
@@ -507,7 +507,7 @@ static void tja1102_p1_register(struct work_struct *work) }/* Real PHY ID of Port 1 is 0 */ - phy =phy_device_create (bus, addr,PHY_ID_TJA1102 , false,NULL ); + phy =phy_device_create (bus, addr,PHY_ID_TJA1102 , false);if (IS_ERR (phy)) {dev_err (dev,"Can't create PHY device for Port 1: %i\n" , addr);
drivers/net/phy/phy_device.c +3 -5
@@ -613,8 +613,7 @@ static int phy_request_driver_module(struct phy_device *phydev) }struct phy_device *phy_device_create (struct mii_bus *bus,int addr,u32 phy_id, -bool is_c45, -struct phy_c45_device_ids *c45_ids) +bool is_c45) {struct phy_device *dev;struct mdio_device *mdiodev;
@@ -647,8 +646,7 @@ struct phy_device *phy_device_create(struct mii_bus *bus, int addr, u32 phy_id, dev->is_c45 = is_c45; dev->phy_id = phy_id; -if (c45_ids) - dev->c45_ids = *c45_ids; + dev->irq = bus->irq[addr];dev_set_name (&mdiodev->dev,PHY_ID_FMT , bus->id, addr);
@@ -829,7 +827,7 @@ static int phy_device_read_id(struct phy_device *phydev)*/ struct phy_device *get_phy_device (struct mii_bus *bus,int addr,bool is_c45) { -return phy_device_create (bus, addr,PHY_ID_NONE , is_c45,NULL ); +return phy_device_create (bus, addr,PHY_ID_NONE , is_c45); }EXPORT_SYMBOL (get_phy_device);
drivers/of/of_mdio.c +1 -1
@@ -121,7 +121,7 @@ static int of_mdiobus_register_phy(struct mii_bus *mdio, "ethernet-phy-ieee802.3-c45");if (!is_c45 && !of_get_phy_id (child, &phy_id)) - phy =phy_device_create (mdio, addr, phy_id,0 ,NULL ); + phy =phy_device_create (mdio, addr, phy_id,0 );else phy =get_phy_device (mdio, addr, is_c45);if (IS_ERR (phy)) {
include/linux/phy.h +1 -2
@@ -1213,8 +1213,7 @@ int phy_modify_paged(struct phy_device *phydev, int page, u32 regnum, u16 mask, u16 set);struct phy_device *phy_device_create (struct mii_bus *bus,int addr,u32 phy_id, -bool is_c45, -struct phy_c45_device_ids *c45_ids); +bool is_c45);#if IS_ENABLED (CONFIG_PHYLIB )struct phy_device *get_phy_device (struct mii_bus *bus,int addr,bool is_c45);int phy_device_register (struct phy_device *phy);
Signature · 1 line
2.26.1
PATCH 9/15 net: phy: delay PHY driver probe until PHY registration
In reply to Bartosz GolaszewskiFrom: Bartosz Golaszewski <[email protected]>
Currently the PHY ID is read without taking the PHY out of reset. This can only work if no resets are defined. This change delays the ID read until we're actually registering the PHY device - this is needed because earlier (when creating the device) we don't have a struct device yet with resets already configured.
While we could use the of_ helpers for GPIO and resets, we will be adding PHY regulator support layer on and there are no regulator APIs that work without struct device.
This means that phy_device_create() now only instantiates the device but doesn't request the relevant driver. If no phy_id is passed to phy_device_create() (for that we introduce a new define: PHY_ID_NONE) then the ID will be read inside phy_device_register().
--- drivers/net/phy/phy_device.c | 47 +++++++++++++++++++----------------- include/linux/phy.h | 1 + 2 files changed, 26 insertions(+), 22 deletions(-)
drivers/net/phy/phy_device.c +25 -22
@@ -658,12 +658,6 @@ struct phy_device *phy_device_create(struct mii_bus *bus, int addr, u32 phy_id,INIT_DELAYED_WORK (&dev->state_queue, phy_state_machine);device_initialize (&mdiodev->dev); - ret =phy_request_driver_module (dev); -if (ret) { -phy_device_free (dev); - dev =ERR_PTR (ret); - } -return dev; }EXPORT_SYMBOL (phy_device_create);
@@ -813,30 +807,29 @@ static int get_phy_id(struct mii_bus *bus, int addr, u32 *phy_id,return 0 ; } +static int phy_device_read_id (struct phy_device *phydev) +{ +struct mdio_device *mdiodev = &phydev->mdio; + + phydev->c45_ids.devices_in_package =0 ; +memset (phydev->c45_ids.device_ids,0xff , +sizeof (phydev->c45_ids.device_ids)); + +return get_phy_id (mdiodev->bus, mdiodev->addr, &phydev->phy_id, + phydev->is_c45, &phydev->c45_ids); +} +/** -* get_phy_device - reads the specified PHY device and returns its @phy_device -* struct +* get_phy_device - create a phy_device withoug PHY ID * @bus: the target MII bus * @addr: PHY address on the MII bus * @is_c45: If true the PHY uses the 802.3 clause 45 protocol * -* Description: Reads the ID registers of the PHY at @addr on the -* @bus, then allocates and returns the phy_device to represent it. +* Allocates a new phy_device for @addr on the @bus. */ struct phy_device *get_phy_device (struct mii_bus *bus,int addr,bool is_c45) { -struct phy_c45_device_ids c45_ids; -u32 phy_id =0 ; -int r; - - c45_ids.devices_in_package =0 ; -memset (c45_ids.device_ids,0xff ,sizeof (c45_ids.device_ids)); - - r =get_phy_id (bus, addr, &phy_id, is_c45, &c45_ids); -if (r) -return ERR_PTR (r); - -return phy_device_create (bus, addr, phy_id, is_c45, &c45_ids); +return phy_device_create (bus, addr,PHY_ID_NONE , is_c45,NULL ); }EXPORT_SYMBOL (get_phy_device);
@@ -855,6 +848,16 @@ int phy_device_register(struct phy_device *phydev)/* Deassert the reset signal */ phy_device_reset (phydev,0 ); +if (phydev->phy_id ==PHY_ID_NONE ) { + err =phy_device_read_id (phydev); +if (err) + goto err_unregister_mdio; + } + + err =phy_request_driver_module (phydev); +if (err) + goto err_unregister_mdio; +/* Run all of the fixups for this PHY */ err =phy_scan_fixups (phydev);if (err) {
include/linux/phy.h +1 -0
@@ -742,6 +742,7 @@ struct phy_driver {#define PHY_ANY_ID "MATCH ANY PHY"#define PHY_ANY_UID 0xffffffff +#define PHY_ID_NONE 0#define PHY_ID_MATCH_EXACT (id) .phy_id = (id), .phy_id_mask = GENMASK(31, 0)#define PHY_ID_MATCH_MODEL (id) .phy_id = (id), .phy_id_mask = GENMASK(31, 4)
Signature · 1 line
2.26.1
PATCH 7/15 net: phy: split out the PHY driver request out of phy_device_create()
In reply to Bartosz GolaszewskiFrom: Bartosz Golaszewski <[email protected]>
Move the code requesting the PHY driver module out of phy_device_create() into a separate helper. This will be later reused when we delay the module loading.
--- drivers/net/phy/phy_device.c | 71 ++++++++++++++++++++---------------- 1 file changed, 39 insertions(+), 32 deletions(-)
drivers/net/phy/phy_device.c +39 -32
@@ -558,7 +558,7 @@ static const struct device_type mdio_bus_phy_type = { .pm = MDIO_BUS_PHY_PM_OPS, }; -static int phy_request_driver_module (struct phy_device *dev,u32 phy_id) +static int phy_do_request_driver_module (struct phy_device *dev,u32 phy_id) {int ret;
@@ -578,6 +578,40 @@ static int phy_request_driver_module(struct phy_device *dev, u32 phy_id)return 0 ; } +static int phy_request_driver_module (struct phy_device *phydev) +{ +int ret; + +/* Request the appropriate module unconditionally; don't +* bother trying to do so only if it isn't already loaded, +* because that gets complicated. A hotplug event would have +* done an unconditional modprobe anyway. +* We don't do normal hotplug because it won't work for MDIO +* -- because it relies on the device staying around for long +* enough for the driver to get loaded. With MDIO, the NIC +* driver will get bored and give up as soon as it finds that +* there's no driver _already_ loaded. +*/ +if (phydev->is_c45) { +const int num_ids =ARRAY_SIZE (phydev->c45_ids.device_ids); +int i; + +for (i =1 ; i < num_ids; i++) { +if (phydev->c45_ids.device_ids[i] ==0xffffffff ) +continue ; + + ret =phy_do_request_driver_module (phydev, + phydev->c45_ids.device_ids[i]); +if (ret) +break ; + } + }else { + ret =phy_do_request_driver_module (phydev, phydev->phy_id); + } + +return ret; +} +struct phy_device *phy_device_create (struct mii_bus *bus,int addr,u32 phy_id,bool is_c45,struct phy_c45_device_ids *c45_ids)
@@ -622,38 +656,11 @@ struct phy_device *phy_device_create(struct mii_bus *bus, int addr, u32 phy_id,mutex_init (&dev->lock);INIT_DELAYED_WORK (&dev->state_queue, phy_state_machine); +device_initialize (&mdiodev->dev); -/* Request the appropriate module unconditionally; don't -* bother trying to do so only if it isn't already loaded, -* because that gets complicated. A hotplug event would have -* done an unconditional modprobe anyway. -* We don't do normal hotplug because it won't work for MDIO -* -- because it relies on the device staying around for long -* enough for the driver to get loaded. With MDIO, the NIC -* driver will get bored and give up as soon as it finds that -* there's no driver _already_ loaded. -*/ -if (is_c45 && c45_ids) { -const int num_ids =ARRAY_SIZE (c45_ids->device_ids); -int i; - -for (i =1 ; i < num_ids; i++) { -if (c45_ids->device_ids[i] ==0xffffffff ) -continue ; - - ret =phy_request_driver_module (dev, - c45_ids->device_ids[i]); -if (ret) -break ; - } - }else { - ret =phy_request_driver_module (dev, phy_id); - } - -if (!ret) { -device_initialize (&mdiodev->dev); - }else { -kfree (dev); + ret =phy_request_driver_module (dev); +if (ret) { +phy_device_free (dev); dev =ERR_PTR (ret); }
Signature · 1 line
2.26.1
PATCH 6/15 net: phy: mdio: reset MDIO devices even if probe() is not implemented
In reply to Bartosz GolaszewskiFrom: Bartosz Golaszewski <[email protected]>
Similarily to PHY drivers - there's no reason to require probe() to be implemented in order to call mdio_device_reset(). MDIO devices can have resets defined without needing to do anything in probe().
--- drivers/net/phy/mdio_device.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-)
drivers/net/phy/mdio_device.c +6 -7
@@ -150,10 +150,10 @@ static int mdio_probe(struct device *dev)struct mdio_driver *mdiodrv =to_mdio_driver (drv);int err =0 ; -if (mdiodrv->probe) { -/* Deassert the reset signal */ -mdio_device_reset (mdiodev,0 ); +/* Deassert the reset signal */ +mdio_device_reset (mdiodev,0 ); +if (mdiodrv->probe) { err = mdiodrv->probe (mdiodev);if (err) {/* Assert the reset signal */
@@ -170,12 +170,11 @@ static int mdio_remove(struct device *dev)struct device_driver *drv = mdiodev->dev.driver;struct mdio_driver *mdiodrv =to_mdio_driver (drv); -if (mdiodrv->remove) { +if (mdiodrv->remove) mdiodrv->remove (mdiodev); -/* Assert the reset signal */ -mdio_device_reset (mdiodev,1 ); - } +/* Assert the reset signal */ +mdio_device_reset (mdiodev,1 );return 0 ; }
Signature · 1 line
2.26.1
PATCH 1/15 net: phy: arrange headers in mdio_bus.c alphabetically
In reply to Bartosz GolaszewskiFrom: Bartosz Golaszewski <[email protected]>
Keeping the headers in alphabetical order is better for readability and allows to easily see if given header is already included.
--- drivers/net/phy/mdio_bus.c | 32 ++++++++++++++++---------------- 1 file changed, 16 insertions(+), 16 deletions(-)
drivers/net/phy/mdio_bus.c +16 -16
@@ -8,32 +8,32 @@#define pr_fmt (fmt) KBUILD_MODNAME ": " fmt -#include <linux/kernel.h> -#include <linux/string.h> -#include <linux/errno.h> -#include <linux/unistd.h> -#include <linux/slab.h> -#include <linux/interrupt.h> -#include <linux/init.h> #include <linux/delay.h> #include <linux/device.h> +#include <linux/errno.h> +#include <linux/etherdevice.h> +#include <linux/ethtool.h> #include <linux/gpio.h> #include <linux/gpio/consumer.h> +#include <linux/init.h> +#include <linux/interrupt.h> +#include <linux/io.h> +#include <linux/kernel.h> +#include <linux/mii.h> +#include <linux/mm.h> +#include <linux/module.h> +#include <linux/netdevice.h> #include <linux/of_device.h> -#include <linux/of_mdio.h> #include <linux/of_gpio.h> -#include <linux/netdevice.h> -#include <linux/etherdevice.h> +#include <linux/of_mdio.h> +#include <linux/phy.h> #include <linux/reset.h> #include <linux/skbuff.h> +#include <linux/slab.h> #include <linux/spinlock.h> -#include <linux/mm.h> -#include <linux/module.h> -#include <linux/mii.h> -#include <linux/ethtool.h> -#include <linux/phy.h> -#include <linux/io.h> +#include <linux/string.h> #include <linux/uaccess.h> +#include <linux/unistd.h> #define CREATE_TRACE_POINTS #include <trace/events/mdio.h>
Signature · 1 line
2.26.1
Re: [PATCH 01/15] net: phy: arrange headers in mdio_bus.c alphabetically
Quoted 6 lines · Bartosz Golaszewski
From: Bartosz Golaszewski <[email protected]>
Keeping the headers in alphabetical order is better for readability and allows to easily see if given header is already included.
Signed-off-by Bartosz Golaszewski <[email protected]>
Andrew
Re: [PATCH 02/15] net: phy: arrange headers in mdio_device.c alphabetically
In reply to Bartosz GolaszewskiQuoted 6 lines · Bartosz Golaszewski
From: Bartosz Golaszewski <[email protected]>
Keeping the headers in alphabetical order is better for readability and allows to easily see if given header is already included.
Signed-off-by Bartosz Golaszewski <[email protected]>
Andrew
Re: [PATCH 03/15] net: phy: arrange headers in phy_device.c alphabetically
In reply to Bartosz GolaszewskiQuoted 6 lines · Bartosz Golaszewski
From: Bartosz Golaszewski <[email protected]>
Keeping the headers in alphabetical order is better for readability and allows to easily see if given header is already included.
Signed-off-by Bartosz Golaszewski <[email protected]>
Andrew
Re: [PATCH 04/15] net: mdio: add a forward declaration for reset_control to mdio.h
In reply to Bartosz GolaszewskiQuoted 8 lines · Bartosz Golaszewski
From: Bartosz Golaszewski <[email protected]>
This header refers to struct
reset_controlbut doesn't include any reset header. The structure definition is probably somehow indirectly pulled in since no warnings are reported but for the sake of correctness add the forward declaration for structreset_control.Signed-off-by Bartosz Golaszewski <[email protected]>
Andrew
Re: [PATCH 05/15] net: phy: reset the PHY even if probe() is not implemented
In reply to Bartosz GolaszewskiQuoted 9 lines · Bartosz Golaszewski
From: Bartosz Golaszewski <[email protected]>
Currently we only call
phy_device_reset()if the PHY driver implements theprobe()callback. This is not mandatory and many drivers (e.g. realtek) don't needprobe()for most devices but still can have reset GPIOs defined. There's no reason to depend on the presence ofprobe()here so pull the reset code out of the if clause.Signed-off-by Bartosz Golaszewski <[email protected]>
Andrew
Re: [PATCH 06/15] net: phy: mdio: reset MDIO devices even if probe() is not implemented
In reply to Bartosz GolaszewskiQuoted 7 lines · Bartosz Golaszewski
From: Bartosz Golaszewski <[email protected]>
Similarily to PHY drivers - there's no reason to require
probe()to be implemented in order to callmdio_device_reset(). MDIO devices can have resets defined without needing to do anything inprobe().Signed-off-by Bartosz Golaszewski <[email protected]>
I would be surprised if there is any MDIO driver which don't have a probe callback.
Andrew
Russell King - ARM Linux admin
maintainerRe: [PATCH 13/15] net: phy: mdio: add support for PHY supply regulator
In reply to Bartosz GolaszewskiQuoted 8 lines · Bartosz Golaszewski
From: Bartosz Golaszewski <[email protected]>
Currently many MAC drivers control the regulator supplying the PHY but this is conceptually wrong. The regulator should be defined as a property of the PHY node on the MDIO bus and controlled by the MDIO sub-system.
Add support for an optional PHY regulator which will be enabled before optional deasserting of the reset signal.
I wonder if this is the right place for this - MDIO devices do not have to be PHYs - they can be switches, and using "phy-supply" for a switch doesn't seem logical.
However, I can see the utility of having having a supply provided for all mdio devices, so it seems to me to be a naming issue. Andrew?
Signature · 2 lines
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/ FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!
Russell King - ARM Linux admin
maintainerRe: [PATCH 14/15] net: phy: add PHY regulator support
In reply to Bartosz GolaszewskiQuoted 81 lines · Bartosz Golaszewski
From: Bartosz Golaszewski <[email protected]>
The MDIO sub-system now supports PHY regulators. Let's reuse the code to extend this support over to the PHY device.
Signed-off-by Bartosz Golaszewski <[email protected]>--- drivers/net/phy/phy_device.c | 49 ++++++++++++++++++++++++++++-------- include/linux/phy.h | 10 ++++++++ 2 files changed, 48 insertions(+), 11 deletions(-)drivers/net/phy/phy_device.c +20
-5@@ -827,7 +827,12 @@ int phy_device_register(struct phy_device *phydev) err =mdiobus_register_device (&phydev->mdio);if (err) -return err; + goto err_out; + +/* Enable the PHY regulator */ + err =phy_device_power_on (phydev); +if (err) + goto err_unregister_mdio;/* Deassert the reset signal */ phy_device_reset (phydev,0 );@@ -846,22 +851,25 @@ int phy_device_register(struct phy_device *phydev) err =phy_scan_fixups (phydev);if (err) {phydev_err (phydev,"failed to initialize\n" ); - goto out; + goto err_reset; } err =device_add (&phydev->mdio.dev);if (err) {phydev_err (phydev,"failed to add\n" ); - goto out; + goto err_reset; }return 0 ; - out: +err_reset:/* Assert the reset signal */ phy_device_reset (phydev,1 ); - +/* Disable the PHY regulator */ +phy_device_power_off (phydev); +err_unregister_mdio:mdiobus_unregister_device (&phydev->mdio); +err_out:return err; }EXPORT_SYMBOL (phy_device_register);@@ -883,6 +891,8 @@ void phy_device_remove(struct phy_device *phydev)/* Assert the reset signal */ phy_device_reset (phydev,1 ); +/* Disable the PHY regulator */ +phy_device_power_off (phydev);mdiobus_unregister_device (&phydev->mdio); }@@ -1064,6 +1074,11 @@ int phy_init_hw(struct phy_device *phydev) {int ret =0 ; +/* Enable the PHY regulator */ + ret =phy_device_power_on (phydev); +if (ret) +return ret; +/* Deassert the reset signal */ phy_device_reset (phydev,0 );
drivers/net/phy/phy_device.c void phy_detach(struct phy_device *phydev) · line 1663
Quoted 5 lines
drivers/net/phy/phy_device.c +2
-0@@ -1644,6 +1659,8 @@ void phy_detach(struct phy_device *phydev)/* Assert the reset signal */ phy_device_reset (phydev,1 ); +/* Disable the PHY regulator */ +phy_device_power_off (phydev);
This is likely to cause issues for some PHY drivers. Note that we have some PHY drivers which register a temperature sensor in the probe function, which means they can be accessed independently of the lifetime of the PHY bound to the network driver (which may only be while the network device is "up".) We certainly do not want hwmon failing just because the network device is down.
That's kind of worked around for the reset stuff, because there are two layers to that: the mdio device layer reset support which knows nothing of the PHY binding state to the network driver, and the phylib reset support, but it is not nice.
Signature · 2 lines
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/ FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!
Andrew Lunn
maintainerRe: [PATCH 09/15] net: phy: delay PHY driver probe until PHY registration
In reply to Bartosz GolaszewskiQuoted 11 lines · Bartosz Golaszewski
From: Bartosz Golaszewski <[email protected]>
Currently the PHY ID is read without taking the PHY out of reset. This can only work if no resets are defined. This change delays the ID read until we're actually registering the PHY device - this is needed because earlier (when creating the device) we don't have a struct device yet with resets already configured.
While we could use the
of_helpers for GPIO and resets, we will be adding PHY regulator support layer on and there are no regulator APIs that work without struct device.
The PHY subsystem cannot be the first to run into this problem, that you need a device structure to make use of the regulator API, but you need the regulator API to probe the device. How do other subsystems work around this?
Maybe it is time to add a lower level API to the regulator framework?
Andrew
Re: [PATCH 09/15] net: phy: delay PHY driver probe until PHY registration
Quoted 4 lines · Andrew Lunn
The PHY subsystem cannot be the first to run into this problem, that you need a device structure to make use of the regulator API, but you need the regulator API to probe the device. How do other subsystems work around this?
If the bus includes power management for the devices on the bus the controller is generally responsible for that rather than the devices, the devices access this via facilities provided by the bus if needed. If the device is enumerated by firmware prior to being physically enumerable then the bus will generally instantiate the device model device and then arrange to wait for the physical device to appear and get joined up with the device model device, typically in such situations the physical device might appear and disappear dynamically at runtime based on what the driver is doing anyway.
Quoted 1 line
Maybe it is time to add a lower level API to the regulator framework?
I don't see any need for that here, this is far from the only thing that's keyed off a struct device and having the device appear and disappear at runtime can make things like runtime PM look really messy to userspace.
We could use a pre-probe stage in the device model for hotpluggable buses in embedded contexts where you might need to bring things out of reset or power them up before they'll appear on the bus for enumeration but buses have mostly handled that at their level.
kernel test robot
build botRe: [PATCH 11/15] net: phy: drop get_phy_device()
In reply to Bartosz GolaszewskiI love your patch! Perhaps something to improve:
Hi Bartosz,
I love your patch! Perhaps something to improve:
[auto build test WARNING on robh/for-next]
[also build test WARNING on linus/master v5.8-rc2]
[cannot apply to arm/for-next next-20200622]
[If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Bartosz-Golaszewski/net-phy-correctly-model-the-PHY-voltage-supply-in-DT/20200622-174726 base: https://git.kernel.org/pub/scm/linux/kernel/git/robh/linux.git for-next config: x86_64-allyesconfig (attached as .config) compiler: clang version 11.0.0 (https://github.com/llvm/llvm-project 1d4c87335d5236ea1f35937e1014980ba961ae34)
reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# install x86_64 cross compiling tool for clang build
# apt-get install binutils-x86-64-linux-gnu
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross ARCH=x86_64
If you fix the issue, kindly add following tag as appropriate
All warnings (new ones prefixed by >>, old ones prefixed by <<):
>> drivers/of/of_mdio.c:123:6: warning: variable 'phy_id' is used uninitialized whenever 'if' condition is false [-Wsometimes-uninitialized]
if (!is_c45) {
^~~~~~~drivers/of/of_mdio.c:129:38: note: uninitialized use occurs here
phy = phy_device_create(mdio, addr, phy_id, is_c45);
^~~~~~drivers/of/of_mdio.c:123:2: note: remove the 'if' if its condition is always true if (!is_c45) {
^~~~~~~~~~~~~drivers/of/of_mdio.c:114:12: note: initialize the variable 'phy_id' to silence this warning u32 phy_id;
^
= 0
1 warning generated.
vim +123 drivers/of/of_mdio.c
106
107 static int of_mdiobus_register_phy(struct mii_bus *mdio,
108 struct device_node *child, u32 addr)
109 {
110 struct mii_timestamper *mii_ts;
111 struct phy_device *phy;
112 bool is_c45;
113 int rc;
114 u32 phy_id;
115
116 mii_ts = of_find_mii_timestamper(child);
117 if (IS_ERR(mii_ts))
118 return PTR_ERR(mii_ts);
119
120 is_c45 = of_device_is_compatible(child,
121 "ethernet-phy-ieee802.3-c45");
122
> 123 if (!is_c45) { 124 rc = of_get_phy_id(child, &phy_id);
125 if (rc)
126 phy_id = PHY_ID_NONE;
127 }
128
129 phy = phy_device_create(mdio, addr, phy_id, is_c45);
130 if (IS_ERR(phy)) {
131 if (mii_ts)
132 unregister_mii_timestamper(mii_ts);
133 return PTR_ERR(phy);
134 }
135
136 rc = of_mdiobus_phy_device_register(mdio, phy, child, addr);
137 if (rc) {
138 if (mii_ts)
139 unregister_mii_timestamper(mii_ts);
140 phy_device_free(phy);
141 return rc;
142 }
143
144 /* phy->mii_ts may already be defined by the PHY driver. A
145 * mii_timestamper probed via the device tree will still have
146 * precedence.
147 */
148 if (mii_ts)
149 phy->mii_ts = mii_ts;
150
151 return 0;
152 }
153
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org
Re: [PATCH 13/15] net: phy: mdio: add support for PHY supply regulator
In reply to Russell King - ARM Linux adminpon., 22 cze 2020 o 15:25 Russell King - ARM Linux admin <[email protected]> napisał(a):
Quoted 11 lines · Bartosz Golaszewski
Quoted 8 lines
From: Bartosz Golaszewski <[email protected]>
Currently many MAC drivers control the regulator supplying the PHY but this is conceptually wrong. The regulator should be defined as a property of the PHY node on the MDIO bus and controlled by the MDIO sub-system.
Add support for an optional PHY regulator which will be enabled before optional deasserting of the reset signal.
Quoted 7 lines
I wonder if this is the right place for this - MDIO devices do not have to be PHYs - they can be switches, and using "phy-supply" for a switch doesn't seem logical.
However, I can see the utility of having having a supply provided for all mdio devices, so it seems to me to be a naming issue. Andrew?
I followed the example of the phy reset retrieved in mdiobus_register_reset(). I'm happy to change it to some other name.
Bart
Re: [PATCH 14/15] net: phy: add PHY regulator support
In reply to Russell King - ARM Linux adminpon., 22 cze 2020 o 15:29 Russell King - ARM Linux admin <[email protected]> napisał(a):
Quoted 1 line
[snip!]
Quoted 8 lines
This is likely to cause issues for some PHY drivers. Note that we have some PHY drivers which register a temperature sensor in the probe function, which means they can be accessed independently of the lifetime of the PHY bound to the network driver (which may only be while the network device is "up".) We certainly do not want hwmon failing just because the network device is down.
Quoted 5 lines
That's kind of worked around for the reset stuff, because there are two layers to that: the mdio device layer reset support which knows nothing of the PHY binding state to the network driver, and the phylib reset support, but it is not nice.
Regulators are reference counted so if the hwmon driver enables it using mdio_device_power_on() it will stay on even after the PHY driver calls phy_device_power_off(), right? Am I missing something?
Bart
Russell King - ARM Linux admin
maintainerRe: [PATCH 14/15] net: phy: add PHY regulator support
Quoted 20 lines · Bartosz Golaszewski
pon., 22 cze 2020 o 15:29 Russell King - ARM Linux admin <[email protected]> napisał(a):
Quoted 1 line
[snip!]
Quoted 8 lines
This is likely to cause issues for some PHY drivers. Note that we have some PHY drivers which register a temperature sensor in the probe function, which means they can be accessed independently of the lifetime of the PHY bound to the network driver (which may only be while the network device is "up".) We certainly do not want hwmon failing just because the network device is down.
Quoted 5 lines
That's kind of worked around for the reset stuff, because there are two layers to that: the mdio device layer reset support which knows nothing of the PHY binding state to the network driver, and the phylib reset support, but it is not nice.
Quoted 3 lines
Regulators are reference counted so if the hwmon driver enables it using
mdio_device_power_on()it will stay on even after the PHY driver callsphy_device_power_off(), right? Am I missing something?
If that is true, you will need to audit the PHY drivers to add that.
Signature · 2 lines
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/ FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!
Re: [PATCH 14/15] net: phy: add PHY regulator support
wt., 23 cze 2020 o 11:43 Russell King - ARM Linux admin <[email protected]> napisał(a):
Quoted 22 lines · Bartosz Golaszewski
Quoted 20 lines
pon., 22 cze 2020 o 15:29 Russell King - ARM Linux admin <[email protected]> napisał(a):
Quoted 1 line
[snip!]
Quoted 13 lines
This is likely to cause issues for some PHY drivers. Note that we have some PHY drivers which register a temperature sensor in the probe function, which means they can be accessed independently of the lifetime of the PHY bound to the network driver (which may only be while the network device is "up".) We certainly do not want hwmon failing just because the network device is down.
That's kind of worked around for the reset stuff, because there are two layers to that: the mdio device layer reset support which knows nothing of the PHY binding state to the network driver, and the phylib reset support, but it is not nice.
Quoted 6 lines
Quoted 3 lines
Regulators are reference counted so if the hwmon driver enables it using
mdio_device_power_on()it will stay on even after the PHY driver callsphy_device_power_off(), right? Am I missing something?If that is true, you will need to audit the PHY drivers to add that.
This change doesn't have any effect on devices which don't have a regulator assigned in DT though. The one I'm adding in the last patch is the first to use this.
Bart
Russell King - ARM Linux admin
maintainerRe: [PATCH 14/15] net: phy: add PHY regulator support
Quoted 28 lines · Bartosz Golaszewski
wt., 23 cze 2020 o 11:43 Russell King - ARM Linux admin <[email protected]> napisał(a):
Quoted 26 lines · Bartosz Golaszewski
Quoted 23 lines
pon., 22 cze 2020 o 15:29 Russell King - ARM Linux admin <[email protected]> napisał(a):
Quoted 1 line
[snip!]
Quoted 8 lines
This is likely to cause issues for some PHY drivers. Note that we have some PHY drivers which register a temperature sensor in the probe function, which means they can be accessed independently of the lifetime of the PHY bound to the network driver (which may only be while the network device is "up".) We certainly do not want hwmon failing just because the network device is down.
Quoted 5 lines
That's kind of worked around for the reset stuff, because there are two layers to that: the mdio device layer reset support which knows nothing of the PHY binding state to the network driver, and the phylib reset support, but it is not nice.
Regulators are reference counted so if the hwmon driver enables it using
mdio_device_power_on()it will stay on even after the PHY driver callsphy_device_power_off(), right? Am I missing something?
Quoted 6 lines
Quoted 2 lines
If that is true, you will need to audit the PHY drivers to add that.
This change doesn't have any effect on devices which don't have a regulator assigned in DT though. The one I'm adding in the last patch is the first to use this.
It's quality of implementation.
Should we wait for someone else to make use of the new regulator support that has been added with a PHY that uses hwmon, and they don't realise that it breaks hwmon on it, and several kernel versions go by without it being noticed. It will only be a noticable issue when the associated network device is down, and that network device driver detaches from the PHY, so _is_ likely not to be noticed.
Or should we do a small amount of work now to properly implement regulator support, which includes a trivial grep for "hwmon" amongst the PHY drivers, and add the necessary call to avoid the regulator being shut off.
Signature · 2 lines
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/ FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!
Re: [PATCH 14/15] net: phy: add PHY regulator support
wt., 23 cze 2020 o 11:56 Russell King - ARM Linux admin <[email protected]> napisał(a):
Quoted 46 lines · Bartosz Golaszewski
Quoted 28 lines
wt., 23 cze 2020 o 11:43 Russell King - ARM Linux admin <[email protected]> napisał(a):
Quoted 26 lines · Bartosz Golaszewski
Quoted 23 lines
pon., 22 cze 2020 o 15:29 Russell King - ARM Linux admin <[email protected]> napisał(a):
Quoted 1 line
[snip!]
Quoted 8 lines
This is likely to cause issues for some PHY drivers. Note that we have some PHY drivers which register a temperature sensor in the probe function, which means they can be accessed independently of the lifetime of the PHY bound to the network driver (which may only be while the network device is "up".) We certainly do not want hwmon failing just because the network device is down.
Quoted 5 lines
That's kind of worked around for the reset stuff, because there are two layers to that: the mdio device layer reset support which knows nothing of the PHY binding state to the network driver, and the phylib reset support, but it is not nice.
Regulators are reference counted so if the hwmon driver enables it using
mdio_device_power_on()it will stay on even after the PHY driver callsphy_device_power_off(), right? Am I missing something?Quoted 6 lines
Quoted 2 lines
If that is true, you will need to audit the PHY drivers to add that.
This change doesn't have any effect on devices which don't have a regulator assigned in DT though. The one I'm adding in the last patch is the first to use this.
It's quality of implementation.
Should we wait for someone else to make use of the new regulator support that has been added with a PHY that uses hwmon, and they don't realise that it breaks hwmon on it, and several kernel versions go by without it being noticed. It will only be a noticable issue when the associated network device is down, and that network device driver detaches from the PHY, so _is_ likely not to be noticed.
Quoted 5 lines
Or should we do a small amount of work now to properly implement regulator support, which includes a trivial grep for "hwmon" amongst the PHY drivers, and add the necessary call to avoid the regulator being shut off.
I'm not sure what the correct approach is here. Provide some helper that, when called, would increase the regulator's reference count even more to keep it enabled from the moment hwmon is registered to when the driver is detached?
Bart
Re: [PATCH 01/15] net: phy: arrange headers in mdio_bus.c alphabetically
In reply to Bartosz GolaszewskiQuoted 6 lines · Bartosz Golaszewski
From: Bartosz Golaszewski <[email protected]>
Keeping the headers in alphabetical order is better for readability and allows to easily see if given header is already included.
Signed-off-by Bartosz Golaszewski <[email protected]>
Signature · 1 line
Florian
Re: [PATCH 02/15] net: phy: arrange headers in mdio_device.c alphabetically
In reply to Bartosz GolaszewskiQuoted 6 lines · Bartosz Golaszewski
From: Bartosz Golaszewski <[email protected]>
Keeping the headers in alphabetical order is better for readability and allows to easily see if given header is already included.
Signed-off-by Bartosz Golaszewski <[email protected]>
Signature · 1 line
Florian
Re: [PATCH 03/15] net: phy: arrange headers in phy_device.c alphabetically
In reply to Bartosz GolaszewskiQuoted 6 lines · Bartosz Golaszewski
From: Bartosz Golaszewski <[email protected]>
Keeping the headers in alphabetical order is better for readability and allows to easily see if given header is already included.
Signed-off-by Bartosz Golaszewski <[email protected]>
Signature · 1 line
Florian
Re: [PATCH 04/15] net: mdio: add a forward declaration for reset_control to mdio.h
In reply to Bartosz GolaszewskiQuoted 17 lines · Bartosz Golaszewski
From: Bartosz Golaszewski <[email protected]>
This header refers to struct
reset_controlbut doesn't include any reset header. The structure definition is probably somehow indirectly pulled in since no warnings are reported but for the sake of correctness add the forward declaration for structreset_control.Signed-off-by Bartosz Golaszewski <[email protected]>--- include/linux/mdio.h | 1 + 1 file changed, 1 insertion(+)include/linux/mdio.h +0
-0
include/linux/mdio.h line 21
Quoted 5 lines
include/linux/mdio.h +1
-0@@ -17,6 +17,7 @@#define MII_REGADDR_C45_MASK GENMASK(15, 0)struct gpio_desc ; +struct reset_control ;struct mii_bus ;
You wrote 3 patches to sort the headers alphabetically, do you mind doing the same thing for forward declarations as well?
Signature · 1 line
Florian
Re: [PATCH 05/15] net: phy: reset the PHY even if probe() is not implemented
In reply to Bartosz GolaszewskiQuoted 9 lines · Bartosz Golaszewski
From: Bartosz Golaszewski <[email protected]>
Currently we only call
phy_device_reset()if the PHY driver implements theprobe()callback. This is not mandatory and many drivers (e.g. realtek) don't needprobe()for most devices but still can have reset GPIOs defined. There's no reason to depend on the presence ofprobe()here so pull the reset code out of the if clause.Signed-off-by Bartosz Golaszewski <[email protected]>
OK, but now let's imagine that a PHY device has two or more reset lines, one of them is going to be managed by the core PHY library and the rest is going to be under the responsibility of the PHY driver, that does not sound intuitive or convenient at all. This is a hypothetical case, but it could conceivable happen, so how about adding a flag to the driver that says "let me manage it a all"?
Signature · 1 line
Florian
Re: [PATCH 06/15] net: phy: mdio: reset MDIO devices even if probe() is not implemented
In reply to Bartosz GolaszewskiQuoted 7 lines · Bartosz Golaszewski
From: Bartosz Golaszewski <[email protected]>
Similarily to PHY drivers - there's no reason to require
probe()to be implemented in order to callmdio_device_reset(). MDIO devices can have resets defined without needing to do anything inprobe().Signed-off-by Bartosz Golaszewski <[email protected]>
Same comment as patch #5, I would prefer that we seek a solution that allows MDIO drivers to manage multiple reset lines (would that exist) on their own instead of this one size fits all approach. Thank you.
Signature · 1 line
Florian
Re: [PATCH 12/15] dt-bindings: mdio: add phy-supply property to ethernet phy node
In reply to Bartosz GolaszewskiQuoted 16 lines · Bartosz Golaszewski
From: Bartosz Golaszewski <[email protected]>
The phy-supply property is often added to MAC nodes but this is wrong conceptually. These supplies should be part of the PHY node on the MDIO bus. Add phy-supply property at PHY level to
mdio.yaml.Signed-off-by Bartosz Golaszewski <[email protected]>--- Documentation/devicetree/bindings/net/mdio.yaml | 4 ++++ 1 file changed, 4 insertions(+)Documentation/devicetree/bindings/net/mdio.yaml +0
-0
Documentation/devicetree/bindings/net/mdio.yaml patternProperties: · line 95
Quoted 6 lines
Documentation/devicetree/bindings/net/mdio.yaml +3
-0@@ -90,6 +90,10 @@ patternProperties: Delay after the reset was deasserted in microseconds. If this property is missing the delay will be skipped. + phy-supply: + description: + Phandle to the regulator that provides the supply voltage to the PHY.
I do not see how you can come up with a generic name here, there could be PHYs supporting different voltages (3.3V, 1.8V, 1.5V) depending on their operation mode/strapping, there can also be different parts of the PHY being powered by different regulators, the analog part could be on an always-on island such that Wake-on-LAN from the PHY could be done, and the digital part could be on a complete different island.
Signature · 1 line
Florian
Re: [PATCH 09/15] net: phy: delay PHY driver probe until PHY registration
In reply to Mark BrownQuoted 7 lines · Mark Brown
Quoted 4 lines · Andrew Lunn
The PHY subsystem cannot be the first to run into this problem, that you need a device structure to make use of the regulator API, but you need the regulator API to probe the device. How do other subsystems work around this?
Quoted 9 lines
If the bus includes power management for the devices on the bus the controller is generally responsible for that rather than the devices, the devices access this via facilities provided by the bus if needed. If the device is enumerated by firmware prior to being physically enumerable then the bus will generally instantiate the device model device and then arrange to wait for the physical device to appear and get joined up with the device model device, typically in such situations the physical device might appear and disappear dynamically at runtime based on what the driver is doing anyway.
In premise there is nothing that prevents the MDIO bus from taking care of the regulators, resets, prior to probing the PHY driver, what is complicated here is that we do need to issue a read of the actual PHY to know its 32-bit unique identifier and match it with an appropriate driver. The way that we have worked around this with if you do not wish such a hardware access to be made, is to provide an Ethernet PHY node compatible string that encodes that 32-bit OUI directly. In premise the same challenges exist with PCI devices/endpoints as well as USB, would they have reset or regulator typically attached to them.
Quoted 8 lines
Quoted 2 lines
Maybe it is time to add a lower level API to the regulator framework?
I don't see any need for that here, this is far from the only thing that's keyed off a struct device and having the device appear and disappear at runtime can make things like runtime PM look really messy to userspace.
Quoted 5 lines
We could use a pre-probe stage in the device model for hotpluggable buses in embedded contexts where you might need to bring things out of reset or power them up before they'll appear on the bus for enumeration but buses have mostly handled that at their level.
That sounds like a better solution, are there any subsystems currently implementing that, or would this be a generic Linux device driver model addition that needs to be done?
Signature · 1 line
Florian
Re: [PATCH 09/15] net: phy: delay PHY driver probe until PHY registration
Quoted 12 lines · Florian Fainelli
Quoted 9 lines · Mark Brown
If the bus includes power management for the devices on the bus the controller is generally responsible for that rather than the devices, the devices access this via facilities provided by the bus if needed. If the device is enumerated by firmware prior to being physically enumerable then the bus will generally instantiate the device model device and then arrange to wait for the physical device to appear and get joined up with the device model device, typically in such situations the physical device might appear and disappear dynamically at runtime based on what the driver is doing anyway.
Quoted 9 lines
In premise there is nothing that prevents the MDIO bus from taking care of the regulators, resets, prior to probing the PHY driver, what is complicated here is that we do need to issue a read of the actual PHY to know its 32-bit unique identifier and match it with an appropriate driver. The way that we have worked around this with if you do not wish such a hardware access to be made, is to provide an Ethernet PHY node compatible string that encodes that 32-bit OUI directly. In premise the same challenges exist with PCI devices/endpoints as well as USB, would they have reset or regulator typically attached to them.
That all sounds very normal and is covered by both cases I describe?
Quoted 8 lines
Quoted 4 lines
We could use a pre-probe stage in the device model for hotpluggable buses in embedded contexts where you might need to bring things out of reset or power them up before they'll appear on the bus for enumeration but buses have mostly handled that at their level.
That sounds like a better solution, are there any subsystems currently implementing that, or would this be a generic Linux device driver model addition that needs to be done?
Like I say I'm suggesting doing something at the device model level.
Re: [PATCH 09/15] net: phy: delay PHY driver probe until PHY registration
śr., 24 cze 2020 o 11:43 Mark Brown <[email protected]> napisał(a):
Quoted 31 lines · Florian Fainelli
Quoted 12 lines · Mark Brown
Quoted 10 lines
If the bus includes power management for the devices on the bus the controller is generally responsible for that rather than the devices, the devices access this via facilities provided by the bus if needed. If the device is enumerated by firmware prior to being physically enumerable then the bus will generally instantiate the device model device and then arrange to wait for the physical device to appear and get joined up with the device model device, typically in such situations the physical device might appear and disappear dynamically at runtime based on what the driver is doing anyway.
Quoted 9 lines
In premise there is nothing that prevents the MDIO bus from taking care of the regulators, resets, prior to probing the PHY driver, what is complicated here is that we do need to issue a read of the actual PHY to know its 32-bit unique identifier and match it with an appropriate driver. The way that we have worked around this with if you do not wish such a hardware access to be made, is to provide an Ethernet PHY node compatible string that encodes that 32-bit OUI directly. In premise the same challenges exist with PCI devices/endpoints as well as USB, would they have reset or regulator typically attached to them.
That all sounds very normal and is covered by both cases I describe?
Quoted 4 lines
Quoted 4 lines
We could use a pre-probe stage in the device model for hotpluggable buses in embedded contexts where you might need to bring things out of reset or power them up before they'll appear on the bus for enumeration but buses have mostly handled that at their level.
Quoted 5 lines
Quoted 3 lines
That sounds like a better solution, are there any subsystems currently implementing that, or would this be a generic Linux device driver model addition that needs to be done?
Like I say I'm suggesting doing something at the device model level.
I didn't expect to open such a can of worms...
This has evolved into several new concepts being proposed vs my use-case which is relatively simple. The former will probably take several months of development, reviews and discussions and it will block supporting the phy supply on pumpkin boards upstream. I would prefer not to redo what other MAC drivers do (phy-supply property on the MAC node, controlling it from the MAC driver itself) if we've already established it's wrong.
Is there any compromise we could reach to add support for a basic, common use-case of a single regulator supplying a PHY that needs enabling before reading its ID short-term (just like we currently support a single reset or reset-gpios property for PHYs) and introducing a whole new concept to the device model for more advanced (but currently mostly hypothetical) cases long-term?
Bart
Re: [PATCH 09/15] net: phy: delay PHY driver probe until PHY registration
Quoted 40 lines · Bartosz Golaszewski
śr., 24 cze 2020 o 11:43 Mark Brown <[email protected]> napisał(a):
Quoted 31 lines · Florian Fainelli
Quoted 12 lines · Mark Brown
Quoted 10 lines
If the bus includes power management for the devices on the bus the controller is generally responsible for that rather than the devices, the devices access this via facilities provided by the bus if needed. If the device is enumerated by firmware prior to being physically enumerable then the bus will generally instantiate the device model device and then arrange to wait for the physical device to appear and get joined up with the device model device, typically in such situations the physical device might appear and disappear dynamically at runtime based on what the driver is doing anyway.
Quoted 9 lines
In premise there is nothing that prevents the MDIO bus from taking care of the regulators, resets, prior to probing the PHY driver, what is complicated here is that we do need to issue a read of the actual PHY to know its 32-bit unique identifier and match it with an appropriate driver. The way that we have worked around this with if you do not wish such a hardware access to be made, is to provide an Ethernet PHY node compatible string that encodes that 32-bit OUI directly. In premise the same challenges exist with PCI devices/endpoints as well as USB, would they have reset or regulator typically attached to them.
That all sounds very normal and is covered by both cases I describe?
Quoted 4 lines
Quoted 4 lines
We could use a pre-probe stage in the device model for hotpluggable buses in embedded contexts where you might need to bring things out of reset or power them up before they'll appear on the bus for enumeration but buses have mostly handled that at their level.
Quoted 5 lines
Quoted 3 lines
That sounds like a better solution, are there any subsystems currently implementing that, or would this be a generic Linux device driver model addition that needs to be done?
Like I say I'm suggesting doing something at the device model level.
I didn't expect to open such a can of worms...
Quoted 7 lines
This has evolved into several new concepts being proposed vs my use-case which is relatively simple. The former will probably take several months of development, reviews and discussions and it will block supporting the phy supply on pumpkin boards upstream. I would prefer not to redo what other MAC drivers do (phy-supply property on the MAC node, controlling it from the MAC driver itself) if we've already established it's wrong.
You are not new to Linux development, so none of this should come as a surprise to you. Your proposed solution has clearly short comings and is a hack, especially around the PHY_ID_NONE business to get a phy_device only then to have the real PHY device ID. You should also now that "I need it now because my product deliverable depends on it" has never been received as a valid argument to coerce people into accepting a solution for which there are at review time known deficiencies to the proposed approach.
Quoted 7 lines
Is there any compromise we could reach to add support for a basic, common use-case of a single regulator supplying a PHY that needs enabling before reading its ID short-term (just like we currently support a single reset or reset-gpios property for PHYs) and introducing a whole new concept to the device model for more advanced (but currently mostly hypothetical) cases long-term?
The pre-probe use case is absolutely not hypothetical, and I would need it for pcie-brcmstb.c at some point which is a PCIe root complex driver with multiple regulators that need to be turned on *prior* to enumerating the PCIe bus and creating pci_device instances. It is literally the same thing as what you are trying to do, just in a different subsystem, therefore I am happy to test and review your patches.
Signature · 1 line
Florian
Re: [PATCH 05/15] net: phy: reset the PHY even if probe() is not implemented
In reply to Florian Fainelliwt., 23 cze 2020 o 21:14 Florian Fainelli <[email protected]> napisał(a):
Quoted 10 lines · Bartosz Golaszewski
Quoted 8 lines
From: Bartosz Golaszewski <[email protected]>
Currently we only call
phy_device_reset()if the PHY driver implements theprobe()callback. This is not mandatory and many drivers (e.g. realtek) don't needprobe()for most devices but still can have reset GPIOs defined. There's no reason to depend on the presence ofprobe()here so pull the reset code out of the if clause.
Quoted 8 lines
Quoted 1 line
Signed-off-by Bartosz Golaszewski <[email protected]>OK, but now let's imagine that a PHY device has two or more reset lines, one of them is going to be managed by the core PHY library and the rest is going to be under the responsibility of the PHY driver, that does not sound intuitive or convenient at all. This is a hypothetical case, but it could conceivable happen, so how about adding a flag to the driver that says "let me manage it a all"?
This sounds good as a new feature idea but doesn't seem to be related to what this patch is trying to do. The only thing it does is improve the current behavior. I'll note your point for the future work on the pre-probe stage.
Bartosz
Re: [PATCH 09/15] net: phy: delay PHY driver probe until PHY registration
In reply to Florian FainelliQuoted 1 line · Florian Fainelli
[snip!]
Quoted 9 lines
Quoted 8 lines
This has evolved into several new concepts being proposed vs my use-case which is relatively simple. The former will probably take several months of development, reviews and discussions and it will block supporting the phy supply on pumpkin boards upstream. I would prefer not to redo what other MAC drivers do (phy-supply property on the MAC node, controlling it from the MAC driver itself) if we've already established it's wrong.
Quoted 9 lines
You are not new to Linux development, so none of this should come as a surprise to you. Your proposed solution has clearly short comings and is a hack, especially around the
PHY_ID_NONEbusiness to get aphy_deviceonly then to have the real PHY device ID. You should also now that "I need it now because my product deliverable depends on it" has never been received as a valid argument to coerce people into accepting a solution for which there are at review time known deficiencies to the proposed approach.
Don't get me wrong, I understand that full well. On the other hand a couple years ago I put a significant amount of work into the concept of early platform device drivers for linux clocksource, clock and interrupt drivers. Every reviewer had his own preferred approach and after something like three completely different submissions and several conversations at conferences I simply gave up due to all the bikeshedding. It just wasn't moving forward and frankly: I expect any changes to the core driver model to follow a similar path of most resistance.
I will give it a shot but at some point getting the job done is better than not getting it done just because the solution isn't perfect. IMO this approach is still slightly more correct than controlling the PHY's supply from the MAC driver.
Bartosz
Russell King - ARM Linux admin
maintainerRe: [PATCH 09/15] net: phy: delay PHY driver probe until PHY registration
In reply to Florian FainelliQuoted 11 lines · Florian Fainelli
Quoted 9 lines · Bartosz Golaszewski
I didn't expect to open such a can of worms...
This has evolved into several new concepts being proposed vs my use-case which is relatively simple. The former will probably take several months of development, reviews and discussions and it will block supporting the phy supply on pumpkin boards upstream. I would prefer not to redo what other MAC drivers do (phy-supply property on the MAC node, controlling it from the MAC driver itself) if we've already established it's wrong.
Quoted 8 lines
You are not new to Linux development, so none of this should come as a surprise to you. Your proposed solution has clearly short comings and is a hack, especially around the
PHY_ID_NONEbusiness to get aphy_deviceonly then to have the real PHY device ID. You should also now that "I need it now because my product deliverable depends on it" has never been received as a valid argument to coerce people into accepting a solution for which there are at review time known deficiencies to the proposed approach.
It /is/ a generic issue. The same problem exists for AMBA Primecell devices, and that code has an internal deferred device list that it manages. See drivers/amba/bus.c, amba_deferred_retry_func(), amba_device_try_add(), and amba_device_add().
As we see more devices gain this property, it needs to be addressed in a generic way, rather than coming up with multiple bus specific implementations.
Maybe struct bus_type needs a method to do the preparation to add a device (such as reading IDs etc), which is called by device_add(). If that method returns -EPROBE_DEFER, the device gets added to a deferred list, which gets retried when drivers are successfully probed. Possible maybe?
Signature · 2 lines
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/ FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!
Russell King - ARM Linux admin
maintainerRe: [PATCH 14/15] net: phy: add PHY regulator support
In reply to Bartosz GolaszewskiQuoted 54 lines · Bartosz Golaszewski
wt., 23 cze 2020 o 11:56 Russell King - ARM Linux admin <[email protected]> napisał(a):
Quoted 46 lines · Bartosz Golaszewski
Quoted 28 lines
wt., 23 cze 2020 o 11:43 Russell King - ARM Linux admin <[email protected]> napisał(a):
Quoted 26 lines · Bartosz Golaszewski
Quoted 23 lines
pon., 22 cze 2020 o 15:29 Russell King - ARM Linux admin <[email protected]> napisał(a):
Quoted 1 line
[snip!]
Quoted 8 lines
This is likely to cause issues for some PHY drivers. Note that we have some PHY drivers which register a temperature sensor in the probe function, which means they can be accessed independently of the lifetime of the PHY bound to the network driver (which may only be while the network device is "up".) We certainly do not want hwmon failing just because the network device is down.
Quoted 5 lines
That's kind of worked around for the reset stuff, because there are two layers to that: the mdio device layer reset support which knows nothing of the PHY binding state to the network driver, and the phylib reset support, but it is not nice.
Regulators are reference counted so if the hwmon driver enables it using
mdio_device_power_on()it will stay on even after the PHY driver callsphy_device_power_off(), right? Am I missing something?Quoted 6 lines
Quoted 2 lines
If that is true, you will need to audit the PHY drivers to add that.
This change doesn't have any effect on devices which don't have a regulator assigned in DT though. The one I'm adding in the last patch is the first to use this.
It's quality of implementation.
Should we wait for someone else to make use of the new regulator support that has been added with a PHY that uses hwmon, and they don't realise that it breaks hwmon on it, and several kernel versions go by without it being noticed. It will only be a noticable issue when the associated network device is down, and that network device driver detaches from the PHY, so _is_ likely not to be noticed.
Quoted 5 lines
Or should we do a small amount of work now to properly implement regulator support, which includes a trivial grep for "hwmon" amongst the PHY drivers, and add the necessary call to avoid the regulator being shut off.
Quoted 4 lines
I'm not sure what the correct approach is here. Provide some helper that, when called, would increase the regulator's reference count even more to keep it enabled from the moment hwmon is registered to when the driver is detached?
I think a PHY driver needs the utility to control this. We need to be careful here with naming, because phylib is not the only code in the kernel that uses the phy_ prefix.
If we had runtime PM support for PHYs, with regulator support hooked into runtime PM, then we already have standard interfaces that drivers can use to control whether the device gets powered down.
Signature · 2 lines
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/ FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!
Russell King - ARM Linux admin
maintainerRe: [PATCH 14/15] net: phy: add PHY regulator support
Quoted 60 lines · Russell King - ARM Linux admin
Quoted 54 lines · Bartosz Golaszewski
wt., 23 cze 2020 o 11:56 Russell King - ARM Linux admin <[email protected]> napisał(a):
Quoted 46 lines · Bartosz Golaszewski
Quoted 28 lines
wt., 23 cze 2020 o 11:43 Russell King - ARM Linux admin <[email protected]> napisał(a):
Quoted 26 lines · Bartosz Golaszewski
Quoted 23 lines
pon., 22 cze 2020 o 15:29 Russell King - ARM Linux admin <[email protected]> napisał(a):
Quoted 1 line
[snip!]
Quoted 8 lines
This is likely to cause issues for some PHY drivers. Note that we have some PHY drivers which register a temperature sensor in the probe function, which means they can be accessed independently of the lifetime of the PHY bound to the network driver (which may only be while the network device is "up".) We certainly do not want hwmon failing just because the network device is down.
Quoted 5 lines
That's kind of worked around for the reset stuff, because there are two layers to that: the mdio device layer reset support which knows nothing of the PHY binding state to the network driver, and the phylib reset support, but it is not nice.
Regulators are reference counted so if the hwmon driver enables it using
mdio_device_power_on()it will stay on even after the PHY driver callsphy_device_power_off(), right? Am I missing something?Quoted 6 lines
Quoted 2 lines
If that is true, you will need to audit the PHY drivers to add that.
This change doesn't have any effect on devices which don't have a regulator assigned in DT though. The one I'm adding in the last patch is the first to use this.
It's quality of implementation.
Should we wait for someone else to make use of the new regulator support that has been added with a PHY that uses hwmon, and they don't realise that it breaks hwmon on it, and several kernel versions go by without it being noticed. It will only be a noticable issue when the associated network device is down, and that network device driver detaches from the PHY, so _is_ likely not to be noticed.
Quoted 5 lines
Or should we do a small amount of work now to properly implement regulator support, which includes a trivial grep for "hwmon" amongst the PHY drivers, and add the necessary call to avoid the regulator being shut off.
Quoted 4 lines
I'm not sure what the correct approach is here. Provide some helper that, when called, would increase the regulator's reference count even more to keep it enabled from the moment hwmon is registered to when the driver is detached?
Quoted 7 lines
I think a PHY driver needs the utility to control this. We need to be careful here with naming, because phylib is not the only code in the kernel that uses the
phy_prefix.If we had runtime PM support for PHYs, with regulator support hooked into runtime PM, then we already have standard interfaces that drivers can use to control whether the device gets powered down.
Other ideas:
- using genpd outside of the SoC to provide power domain management. This is already hooked into runtime PM, but would need their agreement, a genpd provider written, and runtime PM added to phylib.
- if we're going for some core driver model approach, then the driver model only knows when devices are bound and unbound to their driver, it knows nothing of phylib's attach/detach from the network interface. If we want to shut off power when a PHY is not attached, we would likely need some kind of interface to do that.
Signature · 2 lines
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/ FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!
Re: [PATCH 09/15] net: phy: delay PHY driver probe until PHY registration
In reply to Russell King - ARM Linux adminQuoted 30 lines · Russell King - ARM Linux admin
Quoted 11 lines · Florian Fainelli
Quoted 9 lines · Bartosz Golaszewski
I didn't expect to open such a can of worms...
This has evolved into several new concepts being proposed vs my use-case which is relatively simple. The former will probably take several months of development, reviews and discussions and it will block supporting the phy supply on pumpkin boards upstream. I would prefer not to redo what other MAC drivers do (phy-supply property on the MAC node, controlling it from the MAC driver itself) if we've already established it's wrong.
Quoted 8 lines
You are not new to Linux development, so none of this should come as a surprise to you. Your proposed solution has clearly short comings and is a hack, especially around the
PHY_ID_NONEbusiness to get aphy_deviceonly then to have the real PHY device ID. You should also now that "I need it now because my product deliverable depends on it" has never been received as a valid argument to coerce people into accepting a solution for which there are at review time known deficiencies to the proposed approach.It /is/ a generic issue. The same problem exists for AMBA Primecell devices, and that code has an internal deferred device list that it manages. See
drivers/amba/bus.c,amba_deferred_retry_func(),amba_device_try_add(), andamba_device_add().As we see more devices gain this property, it needs to be addressed in a generic way, rather than coming up with multiple bus specific implementations.
Quoted 5 lines
Maybe struct
bus_typeneeds a method to do the preparation to add a device (such as reading IDs etc), which is called bydevice_add(). If that method returns -EPROBE_DEFER, the device gets added to a deferred list, which gets retried when drivers are successfully probed. Possible maybe?
FWIW that would be ideal for solving an ordering a problem we have in the IOMMU subsystem too (which we currently sort-of-handle by deferring driver probe from dma_configure(), but it really needs to be done earlier and not depend on drivers being present at all).
Robin.