3PHY Abstraction Layer
   4(Updated 2008-04-08)
   8 Most network devices consist of set of registers which provide an interface
   9 to a MAC layer, which communicates with the physical connection through a
  10 PHY.  The PHY concerns itself with negotiating link parameters with the link
  11 partner on the other side of the network connection (typically, an ethernet
  12 cable), and provides a register interface to allow drivers to determine what
  13 settings were chosen, and to configure what settings are allowed.
  15 While these devices are distinct from the network devices, and conform to a
  16 standard layout for the registers, it has been common practice to integrate
  17 the PHY management code with the network driver.  This has resulted in large
  18 amounts of redundant code.  Also, on embedded systems with multiple (and
  19 sometimes quite different) ethernet controllers connected to the same 
  20 management bus, it is difficult to ensure safe use of the bus.
  22 Since the PHYs are devices, and the management busses through which they are
  23 accessed are, in fact, busses, the PHY Abstraction Layer treats them as such.
  24 In doing so, it has these goals:
  26   1) Increase code-reuse
  27   2) Increase overall code-maintainability
  28   3) Speed development time for new network drivers, and for new systems
  30 Basically, this layer is meant to provide an interface to PHY devices which
  31 allows network driver writers to write as little code as possible, while
  32 still providing a full feature set.
  34The MDIO bus
  36 Most network devices are connected to a PHY by means of a management bus.
  37 Different devices use different busses (though some share common interfaces).
  38 In order to take advantage of the PAL, each bus interface needs to be
  39 registered as a distinct device.
  41 1) read and write functions must be implemented.  Their prototypes are:
  43     int write(struct mii_bus *bus, int mii_id, int regnum, u16 value);
  44     int read(struct mii_bus *bus, int mii_id, int regnum);
  46   mii_id is the address on the bus for the PHY, and regnum is the register
  47   number.  These functions are guaranteed not to be called from interrupt
  48   time, so it is safe for them to block, waiting for an interrupt to signal
  49   the operation is complete
  51 2) A reset function is necessary.  This is used to return the bus to an
  52   initialized state.
  54 3) A probe function is needed.  This function should set up anything the bus
  55   driver needs, setup the mii_bus structure, and register with the PAL using
  56   mdiobus_register.  Similarly, there's a remove function to undo all of
  57   that (use mdiobus_unregister).
  59 4) Like any driver, the device_driver structure must be configured, and init
  60   exit functions are used to register the driver.
  62 5) The bus must also be declared somewhere as a device, and registered.
  64 As an example for how one driver implemented an mdio bus driver, see
  65 drivers/net/ethernet/freescale/fsl_pq_mdio.c and an associated DTS file
  66 for one of the users. (e.g. "git grep fsl,.*-mdio arch/powerpc/boot/dts/")
  68Connecting to a PHY
  70 Sometime during startup, the network driver needs to establish a connection
  71 between the PHY device, and the network device.  At this time, the PHY's bus
  72 and drivers need to all have been loaded, so it is ready for the connection.
  73 At this point, there are several ways to connect to the PHY:
  75 1) The PAL handles everything, and only calls the network driver when
  76   the link state changes, so it can react.
  78 2) The PAL handles everything except interrupts (usually because the
  79   controller has the interrupt registers).
  81 3) The PAL handles everything, but checks in with the driver every second,
  82   allowing the network driver to react first to any changes before the PAL
  83   does.
  85 4) The PAL serves only as a library of functions, with the network device
  86   manually calling functions to update status, and configure the PHY
  89Letting the PHY Abstraction Layer do Everything
  91 If you choose option 1 (The hope is that every driver can, but to still be
  92 useful to drivers that can't), connecting to the PHY is simple:
  94 First, you need a function to react to changes in the link state.  This
  95 function follows this protocol:
  97   static void adjust_link(struct net_device *dev);
  99 Next, you need to know the device name of the PHY connected to this device. 
 100 The name will look something like, "0:00", where the first number is the
 101 bus id, and the second is the PHY's address on that bus.  Typically,
 102 the bus is responsible for making its ID unique.
 104 Now, to connect, just call this function:
 106   phydev = phy_connect(dev, phy_name, &adjust_link, flags, interface);
 108 phydev is a pointer to the phy_device structure which represents the PHY.  If
 109 phy_connect is successful, it will return the pointer.  dev, here, is the
 110 pointer to your net_device.  Once done, this function will have started the
 111 PHY's software state machine, and registered for the PHY's interrupt, if it
 112 has one.  The phydev structure will be populated with information about the
 113 current state, though the PHY will not yet be truly operational at this
 114 point.
 116 flags is a u32 which can optionally contain phy-specific flags.
 117 This is useful if the system has put hardware restrictions on
 118 the PHY/controller, of which the PHY needs to be aware.
 120 interface is a u32 which specifies the connection type used
 121 between the controller and the PHY.  Examples are GMII, MII,
 122 RGMII, and SGMII.  For a full list, see include/linux/phy.h
 124 Now just make sure that phydev->supported and phydev->advertising have any
 125 values pruned from them which don't make sense for your controller (a 10/100
 126 controller may be connected to a gigabit capable PHY, so you would need to
 127 mask off SUPPORTED_1000baseT*).  See include/linux/ethtool.h for definitions
 128 for these bitfields. Note that you should not SET any bits, or the PHY may
 129 get put into an unsupported state.
 131 Lastly, once the controller is ready to handle network traffic, you call
 132 phy_start(phydev).  This tells the PAL that you are ready, and configures the
 133 PHY to connect to the network.  If you want to handle your own interrupts,
 134 just set phydev->irq to PHY_IGNORE_INTERRUPT before you call phy_start.
 135 Similarly, if you don't want to use interrupts, set phydev->irq to PHY_POLL.
 137 When you want to disconnect from the network (even if just briefly), you call
 138 phy_stop(phydev).
 140Keeping Close Tabs on the PAL
 142 It is possible that the PAL's built-in state machine needs a little help to
 143 keep your network device and the PHY properly in sync.  If so, you can
 144 register a helper function when connecting to the PHY, which will be called
 145 every second before the state machine reacts to any changes.  To do this, you
 146 need to manually call phy_attach() and phy_prepare_link(), and then call
 147 phy_start_machine() with the second argument set to point to your special
 148 handler.
 150 Currently there are no examples of how to use this functionality, and testing
 151 on it has been limited because the author does not have any drivers which use
 152 it (they all use option 1).  So Caveat Emptor.
 154Doing it all yourself
 156 There's a remote chance that the PAL's built-in state machine cannot track
 157 the complex interactions between the PHY and your network device.  If this is
 158 so, you can simply call phy_attach(), and not call phy_start_machine or
 159 phy_prepare_link().  This will mean that phydev->state is entirely yours to
 160 handle (phy_start and phy_stop toggle between some of the states, so you
 161 might need to avoid them).
 163 An effort has been made to make sure that useful functionality can be
 164 accessed without the state-machine running, and most of these functions are
 165 descended from functions which did not interact with a complex state-machine.
 166 However, again, no effort has been made so far to test running without the
 167 state machine, so tryer beware.
 169 Here is a brief rundown of the functions:
 171 int phy_read(struct phy_device *phydev, u16 regnum);
 172 int phy_write(struct phy_device *phydev, u16 regnum, u16 val);
 174   Simple read/write primitives.  They invoke the bus's read/write function
 175   pointers.
 177 void phy_print_status(struct phy_device *phydev);
 179   A convenience function to print out the PHY status neatly.
 181 int phy_start_interrupts(struct phy_device *phydev);
 182 int phy_stop_interrupts(struct phy_device *phydev);
 184   Requests the IRQ for the PHY interrupts, then enables them for
 185   start, or disables then frees them for stop.
 187 struct phy_device * phy_attach(struct net_device *dev, const char *phy_id,
 188                 u32 flags, phy_interface_t interface);
 190   Attaches a network device to a particular PHY, binding the PHY to a generic
 191   driver if none was found during bus initialization.  Passes in
 192   any phy-specific flags as needed.
 194 int phy_start_aneg(struct phy_device *phydev);
 196   Using variables inside the phydev structure, either configures advertising
 197   and resets autonegotiation, or disables autonegotiation, and configures
 198   forced settings.
 200 static inline int phy_read_status(struct phy_device *phydev);
 202   Fills the phydev structure with up-to-date information about the current
 203   settings in the PHY.
 205 int phy_ethtool_sset(struct phy_device *phydev, struct ethtool_cmd *cmd);
 206 int phy_ethtool_gset(struct phy_device *phydev, struct ethtool_cmd *cmd);
 208   Ethtool convenience functions.
 210 int phy_mii_ioctl(struct phy_device *phydev,
 211                 struct mii_ioctl_data *mii_data, int cmd);
 213   The MII ioctl.  Note that this function will completely screw up the state
 214   machine if you write registers like BMCR, BMSR, ADVERTISE, etc.  Best to
 215   use this only to write registers which are not standard, and don't set off
 216   a renegotiation.
 219PHY Device Drivers
 221 With the PHY Abstraction Layer, adding support for new PHYs is
 222 quite easy.  In some cases, no work is required at all!  However,
 223 many PHYs require a little hand-holding to get up-and-running.
 225Generic PHY driver
 227 If the desired PHY doesn't have any errata, quirks, or special
 228 features you want to support, then it may be best to not add
 229 support, and let the PHY Abstraction Layer's Generic PHY Driver
 230 do all of the work.  
 232Writing a PHY driver
 234 If you do need to write a PHY driver, the first thing to do is
 235 make sure it can be matched with an appropriate PHY device.
 236 This is done during bus initialization by reading the device's
 237 UID (stored in registers 2 and 3), then comparing it to each
 238 driver's phy_id field by ANDing it with each driver's
 239 phy_id_mask field.  Also, it needs a name.  Here's an example:
 241   static struct phy_driver dm9161_driver = {
 242         .phy_id         = 0x0181b880,
 243         .name           = "Davicom DM9161E",
 244         .phy_id_mask    = 0x0ffffff0,
 245         ...
 246   }
 248 Next, you need to specify what features (speed, duplex, autoneg,
 249 etc) your PHY device and driver support.  Most PHYs support
 250 PHY_BASIC_FEATURES, but you can look in include/mii.h for other
 251 features.
 253 Each driver consists of a number of function pointers:
 255   config_init: configures PHY into a sane state after a reset.
 256     For instance, a Davicom PHY requires descrambling disabled.
 257   probe: Does any setup needed by the driver
 258   suspend/resume: power management
 259   config_aneg: Changes the speed/duplex/negotiation settings
 260   read_status: Reads the current speed/duplex/negotiation settings
 261   ack_interrupt: Clear a pending interrupt
 262   config_intr: Enable or disable interrupts
 263   remove: Does any driver take-down
 265 Of these, only config_aneg and read_status are required to be
 266 assigned by the driver code.  The rest are optional.  Also, it is
 267 preferred to use the generic phy driver's versions of these two
 268 functions if at all possible: genphy_read_status and
 269 genphy_config_aneg.  If this is not possible, it is likely that
 270 you only need to perform some actions before and after invoking
 271 these functions, and so your functions will wrap the generic
 272 ones.
 274 Feel free to look at the Marvell, Cicada, and Davicom drivers in
 275 drivers/net/phy/ for examples (the lxt and qsemi drivers have
 276 not been tested as of this writing)
 278Board Fixups
 280 Sometimes the specific interaction between the platform and the PHY requires
 281 special handling.  For instance, to change where the PHY's clock input is,
 282 or to add a delay to account for latency issues in the data path.  In order
 283 to support such contingencies, the PHY Layer allows platform code to register
 284 fixups to be run when the PHY is brought up (or subsequently reset).
 286 When the PHY Layer brings up a PHY it checks to see if there are any fixups
 287 registered for it, matching based on UID (contained in the PHY device's phy_id
 288 field) and the bus identifier (contained in phydev->dev.bus_id).  Both must
 289 match, however two constants, PHY_ANY_ID and PHY_ANY_UID, are provided as
 290 wildcards for the bus ID and UID, respectively.
 292 When a match is found, the PHY layer will invoke the run function associated
 293 with the fixup.  This function is passed a pointer to the phy_device of
 294 interest.  It should therefore only operate on that PHY.
 296 The platform code can either register the fixup using phy_register_fixup():
 298        int phy_register_fixup(const char *phy_id,
 299                u32 phy_uid, u32 phy_uid_mask,
 300                int (*run)(struct phy_device *));
 302 Or using one of the two stubs, phy_register_fixup_for_uid() and
 303 phy_register_fixup_for_id():
 305 int phy_register_fixup_for_uid(u32 phy_uid, u32 phy_uid_mask,
 306                int (*run)(struct phy_device *));
 307 int phy_register_fixup_for_id(const char *phy_id,
 308                int (*run)(struct phy_device *));
 310 The stubs set one of the two matching criteria, and set the other one to
 311 match anything.
 313 kindly hosted by Redpill Linpro AS, provider of Linux consulting and operations services since 1995.