1 This document describes the device tree bindings associated with the
2 keystone network coprocessor(NetCP) driver support.
3
4 The network coprocessor (NetCP) is a hardware accelerator that processes
5 Ethernet packets. NetCP has a gigabit Ethernet (GbE) subsystem with a ethernet
6 switch sub-module to send and receive packets. NetCP also includes a packet
7 accelerator (PA) module to perform packet classification operations such as
8 header matching, and packet modification operations such as checksum
9 generation. NetCP can also optionally include a Security Accelerator (SA)
10 capable of performing IPSec operations on ingress/egress packets.
11
12 Keystone II SoC's also have a 10 Gigabit Ethernet Subsystem (XGbE) which
13 includes a 3-port Ethernet switch sub-module capable of 10Gb/s and 1Gb/s rates
14 per Ethernet port.
15
16 Keystone NetCP driver has a plug-in module architecture where each of the NetCP
17 sub-modules exist as a loadable kernel module which plug in to the netcp core.
18 These sub-modules are represented as "netcp-devices" in the dts bindings. It is
19 mandatory to have the ethernet switch sub-module for the ethernet interface to
20 be operational. Any other sub-module like the PA is optional.
21
22 NetCP Ethernet SubSystem Layout:
23
24 -----------------------------
25 NetCP subsystem(10G or 1G)
26 -----------------------------
27 |
28 |-> NetCP Devices -> |
29 | |-> GBE/XGBE Switch
30 | |
31 | |-> Packet Accelerator
32 | |
33 | |-> Security Accelerator
34 |
35 |
36 |
37 |-> NetCP Interfaces -> |
38 |-> Ethernet Port 0
39 |
40 |-> Ethernet Port 1
41 |
42 |-> Ethernet Port 2
43 |
44 |-> Ethernet Port 3
45
46
47 NetCP subsystem properties:
48 Required properties:
49 - compatible: Should be "ti,netcp-1.0"
50 - clocks: phandle to the reference clocks for the subsystem.
51 - dma-id: Navigator packet dma instance id.
52 - ranges: address range of NetCP (includes, Ethernet SS, PA and SA)
53
54 Optional properties:
55 - reg: register location and the size for the following register
56 regions in the specified order.
57 - Efuse MAC address register
58 - dma-coherent: Present if dma operations are coherent
59 - big-endian: Keystone devices can be operated in a mode where the DSP is in
60 the big endian mode. In such cases enable this option. This
61 option should also be enabled if the ARM is operated in
62 big endian mode with the DSP in little endian.
63
64 NetCP device properties: Device specification for NetCP sub-modules.
65 1Gb/10Gb (gbe/xgbe) ethernet switch sub-module specifications.
66 Required properties:
67 - label: Must be "netcp-gbe" for 1Gb & "netcp-xgbe" for 10Gb.
68 - compatible: Must be one of below:-
69 "ti,netcp-gbe" for 1GbE on NetCP 1.4
70 "ti,netcp-gbe-5" for 1GbE N NetCP 1.5 (N=5)
71 "ti,netcp-gbe-9" for 1GbE N NetCP 1.5 (N=9)
72 "ti,netcp-gbe-2" for 1GbE N NetCP 1.5 (N=2)
73 "ti,netcp-xgbe" for 10 GbE
74
75 - reg: register location and the size for the following register
76 regions in the specified order.
77 - switch subsystem registers
78 - sgmii port3/4 module registers (only for NetCP 1.4)
79 - switch module registers
80 - serdes registers (only for 10G)
81
82 NetCP 1.4 ethss, here is the order
83 index #0 - switch subsystem registers
84 index #1 - sgmii port3/4 module registers
85 index #2 - switch module registers
86
87 NetCP 1.5 ethss 9 port, 5 port and 2 port
88 index #0 - switch subsystem registers
89 index #1 - switch module registers
90 index #2 - serdes registers
91
92 - tx-channel: the navigator packet dma channel name for tx.
93 - tx-queue: the navigator queue number associated with the tx dma channel.
94 - interfaces: specification for each of the switch port to be registered as a
95 network interface in the stack.
96 -- slave-port: Switch port number, 0 based numbering.
97 -- link-interface: type of link interface, supported options are
98 - mac<->mac auto negotiate mode: 0
99 - mac<->phy mode: 1
100 - mac<->mac forced mode: 2
101 - mac<->fiber mode: 3
102 - mac<->phy mode with no mdio: 4
103 - 10Gb mac<->phy mode : 10
104 - 10Gb mac<->mac forced mode : 11
105 ----phy-handle: phandle to PHY device
106
107 - cpts: sub-node time synchronization (CPTS) submodule configuration
108 -- clocks: CPTS reference clock. Should point on cpts-refclk-mux clock.
109 -- clock-names: should be "cpts"
110 -- cpts-refclk-mux: multiplexer clock definition sub-node for CPTS reference (RFTCLK) clock
111 --- #clock-cells: should be 0
112 --- clocks: list of CPTS reference (RFTCLK) clock's parents as defined in Data manual
113 --- ti,mux-tbl: array of multiplexer indexes as defined in Data manual
114 --- assigned-clocks: should point on cpts-refclk-mux clock
115 --- assigned-clock-parents: should point on required RFTCLK clock parent to be selected
116 -- cpts_clock_mult: (optional) Numerator to convert input clock ticks
117 into nanoseconds
118 -- cpts_clock_shift: (optional) Denominator to convert input clock ticks into
119 nanoseconds.
120 Mult and shift will be calculated basing on CPTS
121 rftclk frequency if both cpts_clock_shift and
122 cpts_clock_mult properties are not provided.
123
124 Optional properties:
125 - enable-ale: NetCP driver keeps the address learning feature in the ethernet
126 switch module disabled. This attribute is to enable the address
127 learning.
128 - secondary-slave-ports: specification for each of the switch port not be
129 registered as a network interface. NetCP driver
130 will only initialize these ports and attach PHY
131 driver to them if needed.
132
133 NetCP interface properties: Interface specification for NetCP sub-modules.
134 Required properties:
135 - rx-channel: the navigator packet dma channel name for rx.
136 - rx-queue: the navigator queue number associated with rx dma channel.
137 - rx-pool: specifies the number of descriptors to be used & the region-id
138 for creating the rx descriptor pool.
139 - tx-pool: specifies the number of descriptors to be used & the region-id
140 for creating the tx descriptor pool.
141 - rx-queue-depth: number of descriptors in each of the free descriptor
142 queue (FDQ) for the pktdma Rx flow. There can be at
143 present a maximum of 4 queues per Rx flow.
144 - rx-buffer-size: the buffer size for each of the Rx flow FDQ.
145 - tx-completion-queue: the navigator queue number where the descriptors are
146 recycled after Tx DMA completion.
147
148 Optional properties:
149 - efuse-mac: If this is 1, then the MAC address for the interface is
150 obtained from the device efuse mac address register.
151 If this is 2, the two DWORDs occupied by the MAC address
152 are swapped. The netcp driver will swap the two DWORDs
153 back to the proper order when this property is set to 2
154 when it obtains the mac address from efuse.
155 - "netcp-device label": phandle to the device specification for each of NetCP
156 sub-module attached to this interface.
157
158 The MAC address will be determined using the optional properties defined in
159 ethernet.txt and only if efuse-mac is set to 0. If all of the optional MAC
160 address properties are not present, then the driver will use a random MAC
161 address.
162
163 Example binding:
164
165 netcp: netcp@2000000 {
166 reg = <0x2620110 0x8>;
167 reg-names = "efuse";
168 compatible = "ti,netcp-1.0";
169 #address-cells = <1>;
170 #size-cells = <1>;
171 ranges = <0 0x2000000 0xfffff>;
172 clocks = <&papllclk>, <&clkcpgmac>, <&chipclk12>;
173 dma-coherent;
174 /* big-endian; */
175 dma-id = <0>;
176
177 netcp-devices {
178 #address-cells = <1>;
179 #size-cells = <1>;
180 ranges;
181 gbe@90000 {
182 label = "netcp-gbe";
183 reg = <0x90000 0x300>, <0x90400 0x400>, <0x90800 0x700>;
184 /* enable-ale; */
185 tx-queue = <648>;
186 tx-channel = <8>;
187
188 cpts {
189 clocks = <&cpts_refclk_mux>;
190 clock-names = "cpts";
191
192 cpts_refclk_mux: cpts-refclk-mux {
193 #clock-cells = <0>;
194 clocks = <&chipclk12>, <&chipclk13>,
195 <&timi0>, <&timi1>,
196 <&tsipclka>, <&tsrefclk>,
197 <&tsipclkb>;
198 ti,mux-tbl = <0x0>, <0x1>, <0x2>,
199 <0x3>, <0x4>, <0x8>, <0xC>;
200 assigned-clocks = <&cpts_refclk_mux>;
201 assigned-clock-parents = <&chipclk12>;
202 };
203 };
204
205 interfaces {
206 gbe0: interface-0 {
207 slave-port = <0>;
208 link-interface = <4>;
209 };
210 gbe1: interface-1 {
211 slave-port = <1>;
212 link-interface = <4>;
213 };
214 };
215
216 secondary-slave-ports {
217 port-2 {
218 slave-port = <2>;
219 link-interface = <2>;
220 };
221 port-3 {
222 slave-port = <3>;
223 link-interface = <2>;
224 };
225 };
226 };
227 };
228
229 netcp-interfaces {
230 interface-0 {
231 rx-channel = <22>;
232 rx-pool = <1024 12>;
233 tx-pool = <1024 12>;
234 rx-queue-depth = <128 128 0 0>;
235 rx-buffer-size = <1518 4096 0 0>;
236 rx-queue = <8704>;
237 tx-completion-queue = <8706>;
238 efuse-mac = <1>;
239 netcp-gbe = <&gbe0>;
240
241 };
242 interface-1 {
243 rx-channel = <23>;
244 rx-pool = <1024 12>;
245 tx-pool = <1024 12>;
246 rx-queue-depth = <128 128 0 0>;
247 rx-buffer-size = <1518 4096 0 0>;
248 rx-queue = <8705>;
249 tx-completion-queue = <8707>;
250 efuse-mac = <0>;
251 local-mac-address = [02 18 31 7e 3e 6f];
252 netcp-gbe = <&gbe1>;
253 };
254 };
255 };
256
257 CPTS board configuration - select external CPTS RFTCLK:
258
259 &tsrefclk{
260 clock-frequency = <500000000>;
261 };
262
263 &cpts_refclk_mux {
264 assigned-clock-parents = <&tsrefclk>;
265 };
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.