| 1 |
========================= |
| 2 |
NXP SJA1105 switch driver |
| 3 |
========================= |
| 4 |
|
| 5 |
Overview |
| 6 |
======== |
| 7 |
|
| 8 |
The NXP SJA1105 is a family of 10 SPI-managed automotive switches: |
| 9 |
|
| 10 |
- SJA1105E: First generation, no TTEthernet |
| 11 |
- SJA1105T: First generation, TTEthernet |
| 12 |
- SJA1105P: Second generation, no TTEthernet, no SGMII |
| 13 |
- SJA1105Q: Second generation, TTEthernet, no SGMII |
| 14 |
- SJA1105R: Second generation, no TTEthernet, SGMII |
| 15 |
- SJA1105S: Second generation, TTEthernet, SGMII |
| 16 |
- SJA1110A: Third generation, TTEthernet, SGMII, integrated 100base-T1 and |
| 17 |
100base-TX PHYs |
| 18 |
- SJA1110B: Third generation, TTEthernet, SGMII, 100base-T1, 100base-TX |
| 19 |
- SJA1110C: Third generation, TTEthernet, SGMII, 100base-T1, 100base-TX |
| 20 |
- SJA1110D: Third generation, TTEthernet, SGMII, 100base-T1 |
| 21 |
|
| 22 |
Being automotive parts, their configuration interface is geared towards |
| 23 |
set-and-forget use, with minimal dynamic interaction at runtime. They |
| 24 |
require a static configuration to be composed by software and packed |
| 25 |
with CRC and table headers, and sent over SPI. |
| 26 |
|
| 27 |
The static configuration is composed of several configuration tables. Each |
| 28 |
table takes a number of entries. Some configuration tables can be (partially) |
| 29 |
reconfigured at runtime, some not. Some tables are mandatory, some not: |
| 30 |
|
| 31 |
============================= ================== ============================= |
| 32 |
Table Mandatory Reconfigurable |
| 33 |
============================= ================== ============================= |
| 34 |
Schedule no no |
| 35 |
Schedule entry points if Scheduling no |
| 36 |
VL Lookup no no |
| 37 |
VL Policing if VL Lookup no |
| 38 |
VL Forwarding if VL Lookup no |
| 39 |
L2 Lookup no no |
| 40 |
L2 Policing yes no |
| 41 |
VLAN Lookup yes yes |
| 42 |
L2 Forwarding yes partially (fully on P/Q/R/S) |
| 43 |
MAC Config yes partially (fully on P/Q/R/S) |
| 44 |
Schedule Params if Scheduling no |
| 45 |
Schedule Entry Points Params if Scheduling no |
| 46 |
VL Forwarding Params if VL Forwarding no |
| 47 |
L2 Lookup Params no partially (fully on P/Q/R/S) |
| 48 |
L2 Forwarding Params yes no |
| 49 |
Clock Sync Params no no |
| 50 |
AVB Params no no |
| 51 |
General Params yes partially |
| 52 |
Retagging no yes |
| 53 |
xMII Params yes no |
| 54 |
SGMII no yes |
| 55 |
============================= ================== ============================= |
| 56 |
|
| 57 |
|
| 58 |
Also the configuration is write-only (software cannot read it back from the |
| 59 |
switch except for very few exceptions). |
| 60 |
|
| 61 |
The driver creates a static configuration at probe time, and keeps it at |
| 62 |
all times in memory, as a shadow for the hardware state. When required to |
| 63 |
change a hardware setting, the static configuration is also updated. |
| 64 |
If that changed setting can be transmitted to the switch through the dynamic |
| 65 |
reconfiguration interface, it is; otherwise the switch is reset and |
| 66 |
reprogrammed with the updated static configuration. |
| 67 |
|
| 68 |
Switching features |
| 69 |
================== |
| 70 |
|
| 71 |
The driver supports the configuration of L2 forwarding rules in hardware for |
| 72 |
port bridging. The forwarding, broadcast and flooding domain between ports can |
| 73 |
be restricted through two methods: either at the L2 forwarding level (isolate |
| 74 |
one bridge's ports from another's) or at the VLAN port membership level |
| 75 |
(isolate ports within the same bridge). The final forwarding decision taken by |
| 76 |
the hardware is a logical AND of these two sets of rules. |
| 77 |
|
| 78 |
The hardware tags all traffic internally with a port-based VLAN (pvid), or it |
| 79 |
decodes the VLAN information from the 802.1Q tag. Advanced VLAN classification |
| 80 |
is not possible. Once attributed a VLAN tag, frames are checked against the |
| 81 |
port's membership rules and dropped at ingress if they don't match any VLAN. |
| 82 |
This behavior is available when switch ports join a bridge with |
| 83 |
``vlan_filtering 1``. |
| 84 |
|
| 85 |
Normally the hardware is not configurable with respect to VLAN awareness, but |
| 86 |
by changing what TPID the switch searches 802.1Q tags for, the semantics of a |
| 87 |
bridge with ``vlan_filtering 0`` can be kept (accept all traffic, tagged or |
| 88 |
untagged), and therefore this mode is also supported. |
| 89 |
|
| 90 |
Segregating the switch ports in multiple bridges is supported (e.g. 2 + 2), but |
| 91 |
all bridges should have the same level of VLAN awareness (either both have |
| 92 |
``vlan_filtering`` 0, or both 1). |
| 93 |
|
| 94 |
Topology and loop detection through STP is supported. |
| 95 |
|
| 96 |
Offloads |
| 97 |
======== |
| 98 |
|
| 99 |
Time-aware scheduling |
| 100 |
--------------------- |
| 101 |
|
| 102 |
The switch supports a variation of the enhancements for scheduled traffic |
| 103 |
specified in IEEE 802.1Q-2018 (formerly 802.1Qbv). This means it can be used to |
| 104 |
ensure deterministic latency for priority traffic that is sent in-band with its |
| 105 |
gate-open event in the network schedule. |
| 106 |
|
| 107 |
This capability can be managed through the tc-taprio offload ('flags 2'). The |
| 108 |
difference compared to the software implementation of taprio is that the latter |
| 109 |
would only be able to shape traffic originated from the CPU, but not |
| 110 |
autonomously forwarded flows. |
| 111 |
|
| 112 |
The device has 8 traffic classes, and maps incoming frames to one of them based |
| 113 |
on the VLAN PCP bits (if no VLAN is present, the port-based default is used). |
| 114 |
As described in the previous sections, depending on the value of |
| 115 |
``vlan_filtering``, the EtherType recognized by the switch as being VLAN can |
| 116 |
either be the typical 0x8100 or a custom value used internally by the driver |
| 117 |
for tagging. Therefore, the switch ignores the VLAN PCP if used in standalone |
| 118 |
or bridge mode with ``vlan_filtering=0``, as it will not recognize the 0x8100 |
| 119 |
EtherType. In these modes, injecting into a particular TX queue can only be |
| 120 |
done by the DSA net devices, which populate the PCP field of the tagging header |
| 121 |
on egress. Using ``vlan_filtering=1``, the behavior is the other way around: |
| 122 |
offloaded flows can be steered to TX queues based on the VLAN PCP, but the DSA |
| 123 |
net devices are no longer able to do that. To inject frames into a hardware TX |
| 124 |
queue with VLAN awareness active, it is necessary to create a VLAN |
| 125 |
sub-interface on the DSA conduit port, and send normal (0x8100) VLAN-tagged |
| 126 |
towards the switch, with the VLAN PCP bits set appropriately. |
| 127 |
|
| 128 |
Management traffic (having DMAC 01-80-C2-xx-xx-xx or 01-19-1B-xx-xx-xx) is the |
| 129 |
notable exception: the switch always treats it with a fixed priority and |
| 130 |
disregards any VLAN PCP bits even if present. The traffic class for management |
| 131 |
traffic has a value of 7 (highest priority) at the moment, which is not |
| 132 |
configurable in the driver. |
| 133 |
|
| 134 |
Below is an example of configuring a 500 us cyclic schedule on egress port |
| 135 |
``swp5``. The traffic class gate for management traffic (7) is open for 100 us, |
| 136 |
and the gates for all other traffic classes are open for 400 us:: |
| 137 |
|
| 138 |
#!/bin/bash |
| 139 |
|
| 140 |
set -e -u -o pipefail |
| 141 |
|
| 142 |
NSEC_PER_SEC="1000000000" |
| 143 |
|
| 144 |
gatemask() { |
| 145 |
local tc_list="$1" |
| 146 |
local mask=0 |
| 147 |
|
| 148 |
for tc in ${tc_list}; do |
| 149 |
mask=$((${mask} | (1 << ${tc}))) |
| 150 |
done |
| 151 |
|
| 152 |
printf "%02x" ${mask} |
| 153 |
} |
| 154 |
|
| 155 |
if ! systemctl is-active --quiet ptp4l; then |
| 156 |
echo "Please start the ptp4l service" |
| 157 |
exit |
| 158 |
fi |
| 159 |
|
| 160 |
now=$(phc_ctl /dev/ptp1 get | gawk '/clock time is/ { print $5; }') |
| 161 |
# Phase-align the base time to the start of the next second. |
| 162 |
sec=$(echo "${now}" | gawk -F. '{ print $1; }') |
| 163 |
base_time="$(((${sec} + 1) * ${NSEC_PER_SEC}))" |
| 164 |
|
| 165 |
tc qdisc add dev swp5 parent root handle 100 taprio \ |
| 166 |
num_tc 8 \ |
| 167 |
map 0 1 2 3 5 6 7 \ |
| 168 |
queues 1@0 1@1 1@2 1@3 1@4 1@5 1@6 1@7 \ |
| 169 |
base-time ${base_time} \ |
| 170 |
sched-entry S $(gatemask 7) 100000 \ |
| 171 |
sched-entry S $(gatemask "0 1 2 3 4 5 6") 400000 \ |
| 172 |
flags 2 |
| 173 |
|
| 174 |
It is possible to apply the tc-taprio offload on multiple egress ports. There |
| 175 |
are hardware restrictions related to the fact that no gate event may trigger |
| 176 |
simultaneously on two ports. The driver checks the consistency of the schedules |
| 177 |
against this restriction and errors out when appropriate. Schedule analysis is |
| 178 |
needed to avoid this, which is outside the scope of the document. |
| 179 |
|
| 180 |
Routing actions (redirect, trap, drop) |
| 181 |
-------------------------------------- |
| 182 |
|
| 183 |
The switch is able to offload flow-based redirection of packets to a set of |
| 184 |
destination ports specified by the user. Internally, this is implemented by |
| 185 |
making use of Virtual Links, a TTEthernet concept. |
| 186 |
|
| 187 |
The driver supports 2 types of keys for Virtual Links: |
| 188 |
|
| 189 |
- VLAN-aware virtual links: these match on destination MAC address, VLAN ID and |
| 190 |
VLAN PCP. |
| 191 |
- VLAN-unaware virtual links: these match on destination MAC address only. |
| 192 |
|
| 193 |
The VLAN awareness state of the bridge (vlan_filtering) cannot be changed while |
| 194 |
there are virtual link rules installed. |
| 195 |
|
| 196 |
Composing multiple actions inside the same rule is supported. When only routing |
| 197 |
actions are requested, the driver creates a "non-critical" virtual link. When |
| 198 |
the action list also contains tc-gate (more details below), the virtual link |
| 199 |
becomes "time-critical" (draws frame buffers from a reserved memory partition, |
| 200 |
etc). |
| 201 |
|
| 202 |
The 3 routing actions that are supported are "trap", "drop" and "redirect". |
| 203 |
|
| 204 |
Example 1: send frames received on swp2 with a DA of 42:be:24:9b:76:20 to the |
| 205 |
CPU and to swp3. This type of key (DA only) when the port's VLAN awareness |
| 206 |
state is off:: |
| 207 |
|
| 208 |
tc qdisc add dev swp2 clsact |
| 209 |
tc filter add dev swp2 ingress flower skip_sw dst_mac 42:be:24:9b:76:20 \ |
| 210 |
action mirred egress redirect dev swp3 \ |
| 211 |
action trap |
| 212 |
|
| 213 |
Example 2: drop frames received on swp2 with a DA of 42:be:24:9b:76:20, a VID |
| 214 |
of 100 and a PCP of 0:: |
| 215 |
|
| 216 |
tc filter add dev swp2 ingress protocol 802.1Q flower skip_sw \ |
| 217 |
dst_mac 42:be:24:9b:76:20 vlan_id 100 vlan_prio 0 action drop |
| 218 |
|
| 219 |
Time-based ingress policing |
| 220 |
--------------------------- |
| 221 |
|
| 222 |
The TTEthernet hardware abilities of the switch can be constrained to act |
| 223 |
similarly to the Per-Stream Filtering and Policing (PSFP) clause specified in |
| 224 |
IEEE 802.1Q-2018 (formerly 802.1Qci). This means it can be used to perform |
| 225 |
tight timing-based admission control for up to 1024 flows (identified by a |
| 226 |
tuple composed of destination MAC address, VLAN ID and VLAN PCP). Packets which |
| 227 |
are received outside their expected reception window are dropped. |
| 228 |
|
| 229 |
This capability can be managed through the offload of the tc-gate action. As |
| 230 |
routing actions are intrinsic to virtual links in TTEthernet (which performs |
| 231 |
explicit routing of time-critical traffic and does not leave that in the hands |
| 232 |
of the FDB, flooding etc), the tc-gate action may never appear alone when |
| 233 |
asking sja1105 to offload it. One (or more) redirect or trap actions must also |
| 234 |
follow along. |
| 235 |
|
| 236 |
Example: create a tc-taprio schedule that is phase-aligned with a tc-gate |
| 237 |
schedule (the clocks must be synchronized by a 1588 application stack, which is |
| 238 |
outside the scope of this document). No packet delivered by the sender will be |
| 239 |
dropped. Note that the reception window is larger than the transmission window |
| 240 |
(and much more so, in this example) to compensate for the packet propagation |
| 241 |
delay of the link (which can be determined by the 1588 application stack). |
| 242 |
|
| 243 |
Receiver (sja1105):: |
| 244 |
|
| 245 |
tc qdisc add dev swp2 clsact |
| 246 |
now=$(phc_ctl /dev/ptp1 get | awk '/clock time is/ {print $5}') && \ |
| 247 |
sec=$(echo $now | awk -F. '{print $1}') && \ |
| 248 |
base_time="$(((sec + 2) * 1000000000))" && \ |
| 249 |
echo "base time ${base_time}" |
| 250 |
tc filter add dev swp2 ingress flower skip_sw \ |
| 251 |
dst_mac 42:be:24:9b:76:20 \ |
| 252 |
action gate base-time ${base_time} \ |
| 253 |
sched-entry OPEN 60000 -1 -1 \ |
| 254 |
sched-entry CLOSE 40000 -1 -1 \ |
| 255 |
action trap |
| 256 |
|
| 257 |
Sender:: |
| 258 |
|
| 259 |
now=$(phc_ctl /dev/ptp0 get | awk '/clock time is/ {print $5}') && \ |
| 260 |
sec=$(echo $now | awk -F. '{print $1}') && \ |
| 261 |
base_time="$(((sec + 2) * 1000000000))" && \ |
| 262 |
echo "base time ${base_time}" |
| 263 |
tc qdisc add dev eno0 parent root taprio \ |
| 264 |
num_tc 8 \ |
| 265 |
map 0 1 2 3 4 5 6 7 \ |
| 266 |
queues 1@0 1@1 1@2 1@3 1@4 1@5 1@6 1@7 \ |
| 267 |
base-time ${base_time} \ |
| 268 |
sched-entry S 01 50000 \ |
| 269 |
sched-entry S 00 50000 \ |
| 270 |
flags 2 |
| 271 |
|
| 272 |
The engine used to schedule the ingress gate operations is the same that the |
| 273 |
one used for the tc-taprio offload. Therefore, the restrictions regarding the |
| 274 |
fact that no two gate actions (either tc-gate or tc-taprio gates) may fire at |
| 275 |
the same time (during the same 200 ns slot) still apply. |
| 276 |
|
| 277 |
To come in handy, it is possible to share time-triggered virtual links across |
| 278 |
more than 1 ingress port, via flow blocks. In this case, the restriction of |
| 279 |
firing at the same time does not apply because there is a single schedule in |
| 280 |
the system, that of the shared virtual link:: |
| 281 |
|
| 282 |
tc qdisc add dev swp2 ingress_block 1 clsact |
| 283 |
tc qdisc add dev swp3 ingress_block 1 clsact |
| 284 |
tc filter add block 1 flower skip_sw dst_mac 42:be:24:9b:76:20 \ |
| 285 |
action gate index 2 \ |
| 286 |
base-time 0 \ |
| 287 |
sched-entry OPEN 50000000 -1 -1 \ |
| 288 |
sched-entry CLOSE 50000000 -1 -1 \ |
| 289 |
action trap |
| 290 |
|
| 291 |
Hardware statistics for each flow are also available ("pkts" counts the number |
| 292 |
of dropped frames, which is a sum of frames dropped due to timing violations, |
| 293 |
lack of destination ports and MTU enforcement checks). Byte-level counters are |
| 294 |
not available. |
| 295 |
|
| 296 |
Limitations |
| 297 |
=========== |
| 298 |
|
| 299 |
The SJA1105 switch family always performs VLAN processing. When configured as |
| 300 |
VLAN-unaware, frames carry a different VLAN tag internally, depending on |
| 301 |
whether the port is standalone or under a VLAN-unaware bridge. |
| 302 |
|
| 303 |
The virtual link keys are always fixed at {MAC DA, VLAN ID, VLAN PCP}, but the |
| 304 |
driver asks for the VLAN ID and VLAN PCP when the port is under a VLAN-aware |
| 305 |
bridge. Otherwise, it fills in the VLAN ID and PCP automatically, based on |
| 306 |
whether the port is standalone or in a VLAN-unaware bridge, and accepts only |
| 307 |
"VLAN-unaware" tc-flower keys (MAC DA). |
| 308 |
|
| 309 |
The existing tc-flower keys that are offloaded using virtual links are no |
| 310 |
longer operational after one of the following happens: |
| 311 |
|
| 312 |
- port was standalone and joins a bridge (VLAN-aware or VLAN-unaware) |
| 313 |
- port is part of a bridge whose VLAN awareness state changes |
| 314 |
- port was part of a bridge and becomes standalone |
| 315 |
- port was standalone, but another port joins a VLAN-aware bridge and this |
| 316 |
changes the global VLAN awareness state of the bridge |
| 317 |
|
| 318 |
The driver cannot veto all these operations, and it cannot update/remove the |
| 319 |
existing tc-flower filters either. So for proper operation, the tc-flower |
| 320 |
filters should be installed only after the forwarding configuration of the port |
| 321 |
has been made, and removed by user space before making any changes to it. |
| 322 |
|
| 323 |
Device Tree bindings and board design |
| 324 |
===================================== |
| 325 |
|
| 326 |
This section references ``Documentation/devicetree/bindings/net/dsa/nxp,sja1105.yaml`` |
| 327 |
and aims to showcase some potential switch caveats. |
| 328 |
|
| 329 |
RMII PHY role and out-of-band signaling |
| 330 |
--------------------------------------- |
| 331 |
|
| 332 |
In the RMII spec, the 50 MHz clock signals are either driven by the MAC or by |
| 333 |
an external oscillator (but not by the PHY). |
| 334 |
But the spec is rather loose and devices go outside it in several ways. |
| 335 |
Some PHYs go against the spec and may provide an output pin where they source |
| 336 |
the 50 MHz clock themselves, in an attempt to be helpful. |
| 337 |
On the other hand, the SJA1105 is only binary configurable - when in the RMII |
| 338 |
MAC role it will also attempt to drive the clock signal. To prevent this from |
| 339 |
happening it must be put in RMII PHY role. |
| 340 |
But doing so has some unintended consequences. |
| 341 |
In the RMII spec, the PHY can transmit extra out-of-band signals via RXD[1:0]. |
| 342 |
These are practically some extra code words (/J/ and /K/) sent prior to the |
| 343 |
preamble of each frame. The MAC does not have this out-of-band signaling |
| 344 |
mechanism defined by the RMII spec. |
| 345 |
So when the SJA1105 port is put in PHY role to avoid having 2 drivers on the |
| 346 |
clock signal, inevitably an RMII PHY-to-PHY connection is created. The SJA1105 |
| 347 |
emulates a PHY interface fully and generates the /J/ and /K/ symbols prior to |
| 348 |
frame preambles, which the real PHY is not expected to understand. So the PHY |
| 349 |
simply encodes the extra symbols received from the SJA1105-as-PHY onto the |
| 350 |
100Base-Tx wire. |
| 351 |
On the other side of the wire, some link partners might discard these extra |
| 352 |
symbols, while others might choke on them and discard the entire Ethernet |
| 353 |
frames that follow along. This looks like packet loss with some link partners |
| 354 |
but not with others. |
| 355 |
The take-away is that in RMII mode, the SJA1105 must be let to drive the |
| 356 |
reference clock if connected to a PHY. |
| 357 |
|
| 358 |
RGMII fixed-link and internal delays |
| 359 |
------------------------------------ |
| 360 |
|
| 361 |
As mentioned in the bindings document, the second generation of devices has |
| 362 |
tunable delay lines as part of the MAC, which can be used to establish the |
| 363 |
correct RGMII timing budget. |
| 364 |
When powered up, these can shift the Rx and Tx clocks with a phase difference |
| 365 |
between 73.8 and 101.7 degrees. |
| 366 |
The catch is that the delay lines need to lock onto a clock signal with a |
| 367 |
stable frequency. This means that there must be at least 2 microseconds of |
| 368 |
silence between the clock at the old vs at the new frequency. Otherwise the |
| 369 |
lock is lost and the delay lines must be reset (powered down and back up). |
| 370 |
In RGMII the clock frequency changes with link speed (125 MHz at 1000 Mbps, 25 |
| 371 |
MHz at 100 Mbps and 2.5 MHz at 10 Mbps), and link speed might change during the |
| 372 |
AN process. |
| 373 |
In the situation where the switch port is connected through an RGMII fixed-link |
| 374 |
to a link partner whose link state life cycle is outside the control of Linux |
| 375 |
(such as a different SoC), then the delay lines would remain unlocked (and |
| 376 |
inactive) until there is manual intervention (ifdown/ifup on the switch port). |
| 377 |
The take-away is that in RGMII mode, the switch's internal delays are only |
| 378 |
reliable if the link partner never changes link speeds, or if it does, it does |
| 379 |
so in a way that is coordinated with the switch port (practically, both ends of |
| 380 |
the fixed-link are under control of the same Linux system). |
| 381 |
As to why would a fixed-link interface ever change link speeds: there are |
| 382 |
Ethernet controllers out there which come out of reset in 100 Mbps mode, and |
| 383 |
their driver inevitably needs to change the speed and clock frequency if it's |
| 384 |
required to work at gigabit. |
| 385 |
|
| 386 |
MDIO bus and PHY management |
| 387 |
--------------------------- |
| 388 |
|
| 389 |
The SJA1105 does not have an MDIO bus and does not perform in-band AN either. |
| 390 |
Therefore there is no link state notification coming from the switch device. |
| 391 |
A board would need to hook up the PHYs connected to the switch to any other |
| 392 |
MDIO bus available to Linux within the system (e.g. to the DSA conduit's MDIO |
| 393 |
bus). Link state management then works by the driver manually keeping in sync |
| 394 |
(over SPI commands) the MAC link speed with the settings negotiated by the PHY. |
| 395 |
|
| 396 |
By comparison, the SJA1110 supports an MDIO slave access point over which its |
| 397 |
internal 100base-T1 PHYs can be accessed from the host. This is, however, not |
| 398 |
used by the driver, instead the internal 100base-T1 and 100base-TX PHYs are |
| 399 |
accessed through SPI commands, modeled in Linux as virtual MDIO buses. |
| 400 |
|
| 401 |
The microcontroller attached to the SJA1110 port 0 also has an MDIO controller |
| 402 |
operating in master mode, however the driver does not support this either, |
| 403 |
since the microcontroller gets disabled when the Linux driver operates. |
| 404 |
Discrete PHYs connected to the switch ports should have their MDIO interface |
| 405 |
attached to an MDIO controller from the host system and not to the switch, |
| 406 |
similar to SJA1105. |
| 407 |
|
| 408 |
Port compatibility matrix |
| 409 |
------------------------- |
| 410 |
|
| 411 |
The SJA1105 port compatibility matrix is: |
| 412 |
|
| 413 |
===== ============== ============== ============== |
| 414 |
Port SJA1105E/T SJA1105P/Q SJA1105R/S |
| 415 |
===== ============== ============== ============== |
| 416 |
0 xMII xMII xMII |
| 417 |
1 xMII xMII xMII |
| 418 |
2 xMII xMII xMII |
| 419 |
3 xMII xMII xMII |
| 420 |
4 xMII xMII SGMII |
| 421 |
===== ============== ============== ============== |
| 422 |
|
| 423 |
|
| 424 |
The SJA1110 port compatibility matrix is: |
| 425 |
|
| 426 |
===== ============== ============== ============== ============== |
| 427 |
Port SJA1110A SJA1110B SJA1110C SJA1110D |
| 428 |
===== ============== ============== ============== ============== |
| 429 |
0 RevMII (uC) RevMII (uC) RevMII (uC) RevMII (uC) |
| 430 |
1 100base-TX 100base-TX 100base-TX |
| 431 |
or SGMII SGMII |
| 432 |
2 xMII xMII xMII xMII |
| 433 |
or SGMII or SGMII |
| 434 |
3 xMII xMII xMII |
| 435 |
or SGMII or SGMII SGMII |
| 436 |
or 2500base-X or 2500base-X or 2500base-X |
| 437 |
4 SGMII SGMII SGMII SGMII |
| 438 |
or 2500base-X or 2500base-X or 2500base-X or 2500base-X |
| 439 |
5 100base-T1 100base-T1 100base-T1 100base-T1 |
| 440 |
6 100base-T1 100base-T1 100base-T1 100base-T1 |
| 441 |
7 100base-T1 100base-T1 100base-T1 100base-T1 |
| 442 |
8 100base-T1 100base-T1 n/a n/a |
| 443 |
9 100base-T1 100base-T1 n/a n/a |
| 444 |
10 100base-T1 n/a n/a n/a |
| 445 |
===== ============== ============== ============== ============== |