[U-Boot] [PATCH v2] drivers/net/phy: add fixed-phy / fixed-link support

From: Hannes Schmelzer hannes.schmelzer@br-automation.com
This patch adds support for having a "fixed-link" to some other MAC (like some embedded switch-device).
For this purpose we introduce a new phy-driver, called "Fixed PHY".
Fixed PHY works only with CONFIG_DM_ETH enabled, since the fixed-link is described with a subnode below ethernet interface.
Most ethernet drivers (unfortunately not all are following same scheme for searching/attaching phys) are calling "phy_connect(...)" for getting a phy-device. At this point we link in, we search here for a subnode called "fixed- link", once found we start phy_device_create(...) with the special phy- id PHY_FIXED_ID (0xa5a55a5a).
During init the "Fixed PHY" driver has registered with this id and now gets probed, during probe we get all the details about fixed-link out of dts, later on the phy reports this values.
Signed-off-by: Hannes Schmelzer hannes.schmelzer@br-automation.com
Signed-off-by: Hannes Schmelzer oe5hpm@oevsv.at ---
Changes in v2: - drop unary debug prints - simplify 'speed-property' check - make fixed-phy selectable trough Kconfig - add doc/device-tree-bindings/net/fixed-link.txt from linux kernel
doc/device-tree-bindings/net/fixed-link.txt | 30 +++++++++++ drivers/net/phy/Kconfig | 9 ++++ drivers/net/phy/Makefile | 1 + drivers/net/phy/fixed.c | 82 +++++++++++++++++++++++++++++ drivers/net/phy/phy.c | 23 ++++++-- include/phy.h | 3 ++ 6 files changed, 145 insertions(+), 3 deletions(-) create mode 100644 doc/device-tree-bindings/net/fixed-link.txt create mode 100644 drivers/net/phy/fixed.c
diff --git a/doc/device-tree-bindings/net/fixed-link.txt b/doc/device-tree-bindings/net/fixed-link.txt new file mode 100644 index 0000000..5829bd8 --- /dev/null +++ b/doc/device-tree-bindings/net/fixed-link.txt @@ -0,0 +1,30 @@ +Fixed link Device Tree binding +------------------------------ + +Some Ethernet MACs have a "fixed link", and are not connected to a +normal MDIO-managed PHY device. For those situations, a Device Tree +binding allows to describe a "fixed link". + +Such a fixed link situation is described by creating a 'fixed-link' +sub-node of the Ethernet MAC device node, with the following +properties: + +* 'speed' (integer, mandatory), to indicate the link speed. Accepted + values are 10, 100 and 1000 +* 'full-duplex' (boolean, optional), to indicate that full duplex is + used. When absent, half duplex is assumed. +* 'pause' (boolean, optional), to indicate that pause should be + enabled. +* 'asym-pause' (boolean, optional), to indicate that asym_pause should + be enabled. + +Examples: + +ethernet@0 { + ... + fixed-link { + speed = <1000>; + full-duplex; + }; + ... +}; diff --git a/drivers/net/phy/Kconfig b/drivers/net/phy/Kconfig index 1d514e9..6a4a940 100644 --- a/drivers/net/phy/Kconfig +++ b/drivers/net/phy/Kconfig @@ -90,4 +90,13 @@ config PHY_VITESSE config PHY_XILINX bool "Xilinx Ethernet PHYs support"
+config PHY_FIXED + bool "Fixed-Link PHY" + help + Fixed PHY is used for having a 'fixed-link' to another MAC with a direct + connection (MII, RGMII, ...). + There is nothing like autoneogation and so + on, the link is always up with fixed speed and fixed duplex-setting. + More information: doc/device-tree-bindings/net/fixed-link.txt + endif #PHYLIB diff --git a/drivers/net/phy/Makefile b/drivers/net/phy/Makefile index d372971..88c00a5 100644 --- a/drivers/net/phy/Makefile +++ b/drivers/net/phy/Makefile @@ -28,3 +28,4 @@ obj-$(CONFIG_PHY_TI) += ti.o obj-$(CONFIG_PHY_XILINX) += xilinx_phy.o obj-$(CONFIG_PHY_VITESSE) += vitesse.o obj-$(CONFIG_PHY_MSCC) += mscc.o +obj-$(CONFIG_PHY_FIXED) += fixed.o diff --git a/drivers/net/phy/fixed.c b/drivers/net/phy/fixed.c new file mode 100644 index 0000000..2f4bb7a --- /dev/null +++ b/drivers/net/phy/fixed.c @@ -0,0 +1,82 @@ +/* + * Fixed-Link phy + * + * Copyright 2017 Bernecker & Rainer Industrieelektronik GmbH + * + * SPDX-License-Identifier: GPL-2.0+ + */ + +#include <config.h> +#include <common.h> +#include <phy.h> +#include <dm.h> +#include <fdt_support.h> + +DECLARE_GLOBAL_DATA_PTR; + +int fixedphy_probe(struct phy_device *phydev) +{ + struct fixed_link *priv; + int ofnode = phydev->addr; + u32 val; + + /* check for mandatory properties within fixed-link node */ + val = fdt_getprop_u32_default_node(gd->fdt_blob, + ofnode, 0, "speed", ~0UL); + if (val != SPEED_10 && val != SPEED_1000 && val != SPEED_1000) { + printf("ERROR: no/invalid speed given in fixed-link node!"); + return -1; + } + + priv = malloc(sizeof(*priv)); + if (!priv) + return -ENOMEM; + memset(priv, 0, sizeof(*priv)); + + phydev->priv = priv; + phydev->addr = 0; + + priv->link_speed = val; + priv->duplex = fdtdec_get_bool(gd->fdt_blob, ofnode, "full-duplex"); + priv->pause = fdtdec_get_bool(gd->fdt_blob, ofnode, "pause"); + priv->asym_pause = fdtdec_get_bool(gd->fdt_blob, ofnode, "asym-pause"); + + /* set this for preventing stack doing some unary reset */ + phydev->flags |= PHY_FLAG_BROKEN_RESET; + + return 0; +} + +int fixedphy_startup(struct phy_device *phydev) +{ + struct fixed_link *priv = phydev->priv; + + phydev->asym_pause = priv->asym_pause; + phydev->pause = priv->pause; + phydev->duplex = priv->duplex; + phydev->speed = priv->link_speed; + phydev->link = 1; + + return 0; +} + +int fixedphy_shutdown(struct phy_device *phydev) +{ + return 0; +} + +static struct phy_driver fixedphy_driver = { + .uid = PHY_FIXED_ID, + .mask = 0xffffffff, + .name = "Fixed PHY", + .features = PHY_GBIT_FEATURES | SUPPORTED_MII, + .probe = fixedphy_probe, + .startup = fixedphy_startup, + .shutdown = fixedphy_shutdown, +}; + +int phy_fixed_init(void) +{ + phy_register(&fixedphy_driver); + return 0; +} diff --git a/drivers/net/phy/phy.c b/drivers/net/phy/phy.c index 8db6574..a569d0b 100644 --- a/drivers/net/phy/phy.c +++ b/drivers/net/phy/phy.c @@ -515,7 +515,9 @@ int phy_init(void) #ifdef CONFIG_PHY_MSCC phy_mscc_init(); #endif - +#ifdef CONFIG_PHY_FIXED + phy_fixed_init(); +#endif return 0; }
@@ -854,9 +856,24 @@ struct phy_device *phy_connect(struct mii_dev *bus, int addr, struct eth_device *dev, phy_interface_t interface) #endif { - struct phy_device *phydev; + struct phy_device *phydev = NULL; +#if defined(CONFIG_DM_ETH) && defined(CONFIG_PHY_FIXED) + int sn; + const char *name; + sn = fdt_first_subnode(gd->fdt_blob, dev->of_offset); + while (sn > 0) { + name = fdt_get_name(gd->fdt_blob, sn, NULL); + if (name != NULL && strcmp(name, "fixed-link") == 0) { + phydev = phy_device_create(bus, + sn, PHY_FIXED_ID, interface); + break; + } + sn = fdt_next_subnode(gd->fdt_blob, sn); + } +#endif + if (phydev == NULL) + phydev = phy_find_by_mask(bus, 1 << addr, interface);
- phydev = phy_find_by_mask(bus, 1 << addr, interface); if (phydev) phy_connect_dev(phydev, dev); else diff --git a/include/phy.h b/include/phy.h index 5477496..f5fdc04 100644 --- a/include/phy.h +++ b/include/phy.h @@ -15,6 +15,8 @@ #include <linux/ethtool.h> #include <linux/mdio.h>
+#define PHY_FIXED_ID 0xa5a55a5a + #define PHY_MAX_ADDR 32
#define PHY_FLAG_BROKEN_RESET (1 << 0) /* soft reset not supported */ @@ -267,6 +269,7 @@ int phy_ti_init(void); int phy_vitesse_init(void); int phy_xilinx_init(void); int phy_mscc_init(void); +int phy_fixed_init(void);
int board_phy_config(struct phy_device *phydev); int get_phy_id(struct mii_dev *bus, int addr, int devad, u32 *phy_id);

On Tue, Mar 21, 2017 at 4:30 AM, Hannes Schmelzer oe5hpm@oevsv.at wrote:
From: Hannes Schmelzer hannes.schmelzer@br-automation.com
This patch adds support for having a "fixed-link" to some other MAC (like some embedded switch-device).
For this purpose we introduce a new phy-driver, called "Fixed PHY".
Fixed PHY works only with CONFIG_DM_ETH enabled, since the fixed-link is described with a subnode below ethernet interface.
Most ethernet drivers (unfortunately not all are following same scheme for searching/attaching phys) are calling "phy_connect(...)" for getting a phy-device. At this point we link in, we search here for a subnode called "fixed- link", once found we start phy_device_create(...) with the special phy- id PHY_FIXED_ID (0xa5a55a5a).
During init the "Fixed PHY" driver has registered with this id and now gets probed, during probe we get all the details about fixed-link out of dts, later on the phy reports this values.
Signed-off-by: Hannes Schmelzer hannes.schmelzer@br-automation.com
Signed-off-by: Hannes Schmelzer oe5hpm@oevsv.at
Changes in v2:
- drop unary debug prints
- simplify 'speed-property' check
- make fixed-phy selectable trough Kconfig
- add doc/device-tree-bindings/net/fixed-link.txt from linux kernel
doc/device-tree-bindings/net/fixed-link.txt | 30 +++++++++++ drivers/net/phy/Kconfig | 9 ++++ drivers/net/phy/Makefile | 1 + drivers/net/phy/fixed.c | 82 +++++++++++++++++++++++++++++ drivers/net/phy/phy.c | 23 ++++++-- include/phy.h | 3 ++ 6 files changed, 145 insertions(+), 3 deletions(-) create mode 100644 doc/device-tree-bindings/net/fixed-link.txt create mode 100644 drivers/net/phy/fixed.c
diff --git a/doc/device-tree-bindings/net/fixed-link.txt b/doc/device-tree-bindings/net/fixed-link.txt new file mode 100644 index 0000000..5829bd8 --- /dev/null +++ b/doc/device-tree-bindings/net/fixed-link.txt @@ -0,0 +1,30 @@ +Fixed link Device Tree binding +------------------------------
+Some Ethernet MACs have a "fixed link", and are not connected to a +normal MDIO-managed PHY device. For those situations, a Device Tree +binding allows to describe a "fixed link".
+Such a fixed link situation is described by creating a 'fixed-link' +sub-node of the Ethernet MAC device node, with the following +properties:
+* 'speed' (integer, mandatory), to indicate the link speed. Accepted
- values are 10, 100 and 1000
+* 'full-duplex' (boolean, optional), to indicate that full duplex is
- used. When absent, half duplex is assumed.
+* 'pause' (boolean, optional), to indicate that pause should be
- enabled.
+* 'asym-pause' (boolean, optional), to indicate that asym_pause should
- be enabled.
+Examples:
+ethernet@0 {
...
fixed-link {
speed = <1000>;
full-duplex;
};
...
+}; diff --git a/drivers/net/phy/Kconfig b/drivers/net/phy/Kconfig index 1d514e9..6a4a940 100644 --- a/drivers/net/phy/Kconfig +++ b/drivers/net/phy/Kconfig @@ -90,4 +90,13 @@ config PHY_VITESSE config PHY_XILINX bool "Xilinx Ethernet PHYs support"
+config PHY_FIXED
bool "Fixed-Link PHY"
depends on DM_ETH
help
Fixed PHY is used for having a 'fixed-link' to another MAC with a direct
connection (MII, RGMII, ...).
There is nothing like autoneogation and so
on, the link is always up with fixed speed and fixed duplex-setting.
More information: doc/device-tree-bindings/net/fixed-link.txt
endif #PHYLIB diff --git a/drivers/net/phy/Makefile b/drivers/net/phy/Makefile index d372971..88c00a5 100644 --- a/drivers/net/phy/Makefile +++ b/drivers/net/phy/Makefile @@ -28,3 +28,4 @@ obj-$(CONFIG_PHY_TI) += ti.o obj-$(CONFIG_PHY_XILINX) += xilinx_phy.o obj-$(CONFIG_PHY_VITESSE) += vitesse.o obj-$(CONFIG_PHY_MSCC) += mscc.o +obj-$(CONFIG_PHY_FIXED) += fixed.o diff --git a/drivers/net/phy/fixed.c b/drivers/net/phy/fixed.c new file mode 100644 index 0000000..2f4bb7a --- /dev/null +++ b/drivers/net/phy/fixed.c @@ -0,0 +1,82 @@ +/*
- Fixed-Link phy
- Copyright 2017 Bernecker & Rainer Industrieelektronik GmbH
- SPDX-License-Identifier: GPL-2.0+
- */
+#include <config.h> +#include <common.h> +#include <phy.h> +#include <dm.h> +#include <fdt_support.h>
+DECLARE_GLOBAL_DATA_PTR;
+int fixedphy_probe(struct phy_device *phydev) +{
struct fixed_link *priv;
int ofnode = phydev->addr;
u32 val;
/* check for mandatory properties within fixed-link node */
val = fdt_getprop_u32_default_node(gd->fdt_blob,
ofnode, 0, "speed", ~0UL);
Why not just default to 0 here? Avoid any 64-bit issues.
if (val != SPEED_10 && val != SPEED_1000 && val != SPEED_1000) {
printf("ERROR: no/invalid speed given in fixed-link node!");
return -1;
Can you return an appropriate error code here? -EINVAL?
}
priv = malloc(sizeof(*priv));
if (!priv)
return -ENOMEM;
memset(priv, 0, sizeof(*priv));
phydev->priv = priv;
phydev->addr = 0;
priv->link_speed = val;
priv->duplex = fdtdec_get_bool(gd->fdt_blob, ofnode, "full-duplex");
priv->pause = fdtdec_get_bool(gd->fdt_blob, ofnode, "pause");
priv->asym_pause = fdtdec_get_bool(gd->fdt_blob, ofnode, "asym-pause");
/* set this for preventing stack doing some unary reset */
Please clarify this comment.
phydev->flags |= PHY_FLAG_BROKEN_RESET;
return 0;
+}
+int fixedphy_startup(struct phy_device *phydev) +{
struct fixed_link *priv = phydev->priv;
phydev->asym_pause = priv->asym_pause;
phydev->pause = priv->pause;
phydev->duplex = priv->duplex;
phydev->speed = priv->link_speed;
phydev->link = 1;
return 0;
+}
+int fixedphy_shutdown(struct phy_device *phydev) +{
return 0;
+}
+static struct phy_driver fixedphy_driver = {
.uid = PHY_FIXED_ID,
.mask = 0xffffffff,
.name = "Fixed PHY",
.features = PHY_GBIT_FEATURES | SUPPORTED_MII,
.probe = fixedphy_probe,
.startup = fixedphy_startup,
.shutdown = fixedphy_shutdown,
+};
+int phy_fixed_init(void) +{
phy_register(&fixedphy_driver);
return 0;
+} diff --git a/drivers/net/phy/phy.c b/drivers/net/phy/phy.c index 8db6574..a569d0b 100644 --- a/drivers/net/phy/phy.c +++ b/drivers/net/phy/phy.c @@ -515,7 +515,9 @@ int phy_init(void) #ifdef CONFIG_PHY_MSCC phy_mscc_init(); #endif
+#ifdef CONFIG_PHY_FIXED
phy_fixed_init();
+#endif return 0; }
@@ -854,9 +856,24 @@ struct phy_device *phy_connect(struct mii_dev *bus, int addr, struct eth_device *dev, phy_interface_t interface) #endif {
struct phy_device *phydev;
struct phy_device *phydev = NULL;
+#if defined(CONFIG_DM_ETH) && defined(CONFIG_PHY_FIXED)
Rather than check for DM_ETH here, just add the dependency in the Kconfig.
int sn;
const char *name;
sn = fdt_first_subnode(gd->fdt_blob, dev->of_offset);
while (sn > 0) {
name = fdt_get_name(gd->fdt_blob, sn, NULL);
if (name != NULL && strcmp(name, "fixed-link") == 0) {
phydev = phy_device_create(bus,
sn, PHY_FIXED_ID, interface);
break;
}
sn = fdt_next_subnode(gd->fdt_blob, sn);
}
+#endif
if (phydev == NULL)
phydev = phy_find_by_mask(bus, 1 << addr, interface);
phydev = phy_find_by_mask(bus, 1 << addr, interface); if (phydev) phy_connect_dev(phydev, dev); else
diff --git a/include/phy.h b/include/phy.h index 5477496..f5fdc04 100644 --- a/include/phy.h +++ b/include/phy.h @@ -15,6 +15,8 @@ #include <linux/ethtool.h> #include <linux/mdio.h>
+#define PHY_FIXED_ID 0xa5a55a5a
#define PHY_MAX_ADDR 32
#define PHY_FLAG_BROKEN_RESET (1 << 0) /* soft reset not supported */ @@ -267,6 +269,7 @@ int phy_ti_init(void); int phy_vitesse_init(void); int phy_xilinx_init(void); int phy_mscc_init(void); +int phy_fixed_init(void);
int board_phy_config(struct phy_device *phydev); int get_phy_id(struct mii_dev *bus, int addr, int devad, u32 *phy_id); -- 2.7.4

On 03/21/2017 07:26 PM, Joe Hershberger wrote: Hi Joe,
many thanks for your input, i will do some rework and send v3. But a few words about my doing below.
On Tue, Mar 21, 2017 at 4:30 AM, Hannes Schmelzer oe5hpm@oevsv.at wrote:
From: Hannes Schmelzer hannes.schmelzer@br-automation.com
This patch adds support for having a "fixed-link" to some other MAC (like some embedded switch-device).
(....)
priv->duplex = fdtdec_get_bool(gd->fdt_blob, ofnode, "full-duplex");
priv->pause = fdtdec_get_bool(gd->fdt_blob, ofnode, "pause");
priv->asym_pause = fdtdec_get_bool(gd->fdt_blob, ofnode, "asym-pause");
/* set this for preventing stack doing some unary reset */
Please clarify this comment.
During startup phy_reset(...) is called from phy_connect_dev(...) to reset the connected phy. This isn't done here due to 2 points: a) reset of a fixed link is unary b) reset would fail since the code wants to write mii registers to not physically existing device.
OK?
cheers, Hannes

On Tue, Mar 21, 2017 at 2:34 PM, Hannes Schmelzer hannes@schmelzer.or.at wrote:
On 03/21/2017 07:26 PM, Joe Hershberger wrote: Hi Joe,
many thanks for your input, i will do some rework and send v3. But a few words about my doing below.
On Tue, Mar 21, 2017 at 4:30 AM, Hannes Schmelzer oe5hpm@oevsv.at wrote:
From: Hannes Schmelzer hannes.schmelzer@br-automation.com
This patch adds support for having a "fixed-link" to some other MAC (like some embedded switch-device).
(....)
priv->duplex = fdtdec_get_bool(gd->fdt_blob, ofnode,
"full-duplex");
priv->pause = fdtdec_get_bool(gd->fdt_blob, ofnode, "pause");
priv->asym_pause = fdtdec_get_bool(gd->fdt_blob, ofnode,
"asym-pause");
/* set this for preventing stack doing some unary reset */
Please clarify this comment.
During startup phy_reset(...) is called from phy_connect_dev(...) to reset the connected phy. This isn't done here due to 2 points: a) reset of a fixed link is unary
Are you meaning to say unnecessary? Or do you mean unary (i.e. only _one_ reset?)
b) reset would fail since the code wants to write mii registers to not physically existing device.
This is a much better reason to put in the comment.
Maybe change the comment to "Fixed-link phy must not be reset by core phy code"
OK?
cheers, Hannes

On 03/21/2017 09:00 PM, Joe Hershberger wrote:
On Tue, Mar 21, 2017 at 2:34 PM, Hannes Schmelzer hannes@schmelzer.or.at wrote:
On 03/21/2017 07:26 PM, Joe Hershberger wrote: Hi Joe,
many thanks for your input, i will do some rework and send v3. But a few words about my doing below.
On Tue, Mar 21, 2017 at 4:30 AM, Hannes Schmelzer oe5hpm@oevsv.at wrote:
From: Hannes Schmelzer hannes.schmelzer@br-automation.com
This patch adds support for having a "fixed-link" to some other MAC (like some embedded switch-device).
(....)
priv->duplex = fdtdec_get_bool(gd->fdt_blob, ofnode,
"full-duplex");
priv->pause = fdtdec_get_bool(gd->fdt_blob, ofnode, "pause");
priv->asym_pause = fdtdec_get_bool(gd->fdt_blob, ofnode,
"asym-pause");
/* set this for preventing stack doing some unary reset */
Please clarify this comment.
During startup phy_reset(...) is called from phy_connect_dev(...) to reset the connected phy. This isn't done here due to 2 points: a) reset of a fixed link is unary
Are you meaning to say unnecessary? Or do you mean unary (i.e. only _one_ reset?)
Yes, thats what i wanted to say ;-)
b) reset would fail since the code wants to write mii registers to not physically existing device.
This is a much better reason to put in the comment.
Maybe change the comment to "Fixed-link phy must not be reset by core phy code"
Okay, changing the comment and sending v4.
participants (3)
-
Hannes Schmelzer
-
Hannes Schmelzer
-
Joe Hershberger