blob: d0c3571165bcbc63fd54721345066542a35e25e9 [file] [log] [blame]
Mike Frysinger1f1e7742009-09-09 14:41:22 -04001-----------------------
2 Ethernet Driver Guide
3-----------------------
4
5The networking stack in Das U-Boot is designed for multiple network devices
6to be easily added and controlled at runtime. This guide is meant for people
7who wish to review the net driver stack with an eye towards implementing your
8own ethernet device driver. Here we will describe a new pseudo 'APE' driver.
9
10------------------
11 Driver Functions
12------------------
13
14All functions you will be implementing in this document have the return value
15meaning of 0 for success and non-zero for failure.
16
17 ----------
18 Register
19 ----------
20
21When U-Boot initializes, it will call the common function eth_initialize().
22This will in turn call the board-specific board_eth_init() (or if that fails,
23the cpu-specific cpu_eth_init()). These board-specific functions can do random
24system handling, but ultimately they will call the driver-specific register
25function which in turn takes care of initializing that particular instance.
26
27Keep in mind that you should code the driver to avoid storing state in global
Ben Warren99dbd4e2009-10-05 00:02:51 -070028data as someone might want to hook up two of the same devices to one board.
29Any such information that is specific to an interface should be stored in a
30private, driver-defined data structure and pointed to by eth->priv (see below).
Mike Frysinger1f1e7742009-09-09 14:41:22 -040031
32So the call graph at this stage would look something like:
33board_init()
34 eth_initialize()
35 board_eth_init() / cpu_eth_init()
36 driver_register()
37 initialize eth_device
38 eth_register()
39
40At this point in time, the only thing you need to worry about is the driver's
41register function. The pseudo code would look something like:
42int ape_register(bd_t *bis, int iobase)
43{
44 struct ape_priv *priv;
45 struct eth_device *dev;
46
47 priv = malloc(sizeof(*priv));
48 if (priv == NULL)
49 return 1;
50
51 dev = malloc(sizeof(*dev));
52 if (dev == NULL) {
53 free(priv);
54 return 1;
55 }
56
57 /* setup whatever private state you need */
58
59 memset(dev, 0, sizeof(*dev));
60 sprintf(dev->name, "APE");
61
62 /* if your device has dedicated hardware storage for the
63 * MAC, read it and initialize dev->enetaddr with it
64 */
65 ape_mac_read(dev->enetaddr);
66
67 dev->iobase = iobase;
68 dev->priv = priv;
69 dev->init = ape_init;
70 dev->halt = ape_halt;
71 dev->send = ape_send;
72 dev->recv = ape_recv;
73
74 eth_register(dev);
75
76#ifdef CONFIG_CMD_MII)
77 miiphy_register(dev->name, ape_mii_read, ape_mii_write);
78#endif
79
Ben Warren99dbd4e2009-10-05 00:02:51 -070080 return 1;
Mike Frysinger1f1e7742009-09-09 14:41:22 -040081}
82
83The exact arguments needed to initialize your device are up to you. If you
84need to pass more/less arguments, that's fine. You should also add the
Ben Warren99dbd4e2009-10-05 00:02:51 -070085prototype for your new register function to include/netdev.h.
86
87The return value for this function should be as follows:
88< 0 - failure (hardware failure, not probe failure)
89>=0 - number of interfaces detected
90
Wolfgang Denk49467752009-10-28 00:49:47 +010091You might notice that many drivers seem to use xxx_initialize() rather than
Ben Warren99dbd4e2009-10-05 00:02:51 -070092xxx_register(). This is the old naming convention and should be avoided as it
93causes confusion with the driver-specific init function.
Mike Frysinger1f1e7742009-09-09 14:41:22 -040094
95Other than locating the MAC address in dedicated hardware storage, you should
96not touch the hardware in anyway. That step is handled in the driver-specific
97init function. Remember that we are only registering the device here, we are
98not checking its state or doing random probing.
99
100 -----------
101 Callbacks
102 -----------
103
104Now that we've registered with the ethernet layer, we can start getting some
105real work done. You will need four functions:
106 int ape_init(struct eth_device *dev, bd_t *bis);
107 int ape_send(struct eth_device *dev, volatile void *packet, int length);
108 int ape_recv(struct eth_device *dev);
109 int ape_halt(struct eth_device *dev);
110
111The init function checks the hardware (probing/identifying) and gets it ready
112for send/recv operations. You often do things here such as resetting the MAC
113and/or PHY, and waiting for the link to autonegotiate. You should also take
114the opportunity to program the device's MAC address with the dev->enetaddr
115member. This allows the rest of U-Boot to dynamically change the MAC address
116and have the new settings be respected.
117
118The send function does what you think -- transmit the specified packet whose
119size is specified by length (in bytes). You should not return until the
120transmission is complete, and you should leave the state such that the send
121function can be called multiple times in a row.
122
123The recv function should process packets as long as the hardware has them
124readily available before returning. i.e. you should drain the hardware fifo.
Mike Frysingere5c5d9e2009-11-01 22:39:56 -0500125For each packet you receive, you should call the NetReceive() function on it
126along with the packet length. The common code sets up packet buffers for you
127already in the .bss (NetRxPackets), so there should be no need to allocate your
128own. This doesn't mean you must use the NetRxPackets array however; you're
129free to call the NetReceive() function with any buffer you wish. So the pseudo
130code here would look something like:
Mike Frysinger1f1e7742009-09-09 14:41:22 -0400131int ape_recv(struct eth_device *dev)
132{
133 int length, i = 0;
134 ...
135 while (packets_are_available()) {
136 ...
137 length = ape_get_packet(&NetRxPackets[i]);
138 ...
139 NetReceive(&NetRxPackets[i], length);
140 ...
141 if (++i >= PKTBUFSRX)
142 i = 0;
143 ...
144 }
145 ...
146 return 0;
147}
148
149The halt function should turn off / disable the hardware and place it back in
Mike Frysingere5c5d9e2009-11-01 22:39:56 -0500150its reset state. It can be called at any time (before any call to the related
151init function), so make sure it can handle this sort of thing.
Mike Frysinger1f1e7742009-09-09 14:41:22 -0400152
153So the call graph at this stage would look something like:
154some net operation (ping / tftp / whatever...)
155 eth_init()
156 dev->init()
157 eth_send()
158 dev->send()
159 eth_rx()
160 dev->recv()
161 eth_halt()
162 dev->halt()
163
164-----------------------------
165 CONFIG_MII / CONFIG_CMD_MII
166-----------------------------
167
168If your device supports banging arbitrary values on the MII bus (pretty much
169every device does), you should add support for the mii command. Doing so is
170fairly trivial and makes debugging mii issues a lot easier at runtime.
171
172After you have called eth_register() in your driver's register function, add
173a call to miiphy_register() like so:
174#if defined(CONFIG_MII) || defined(CONFIG_CMD_MII)
175 miiphy_register(dev->name, mii_read, mii_write);
176#endif
177
178And then define the mii_read and mii_write functions if you haven't already.
179Their syntax is straightforward:
180 int mii_read(char *devname, uchar addr, uchar reg, ushort *val);
181 int mii_write(char *devname, uchar addr, uchar reg, ushort val);
182
183The read function should read the register 'reg' from the phy at address 'addr'
184and store the result in the pointer 'val'. The implementation for the write
185function should logically follow.