1.. SPDX-License-Identifier: GPL-2.0
2
3=========================
43Com Vortex device driver
5=========================
6
7Andrew Morton
8
930 April 2000
10
11
12This document describes the usage and errata of the 3Com "Vortex" device
13driver for Linux, 3c59x.c.
14
15The driver was written by Donald Becker <becker@scyld.com>
16
17Don is no longer the prime maintainer of this version of the driver.
18Please report problems to one or more of:
19
20- Andrew Morton
21- Netdev mailing list <netdev@vger.kernel.org>
22- Linux kernel mailing list <linux-kernel@vger.kernel.org>
23
24Please note the 'Reporting and Diagnosing Problems' section at the end
25of this file.
26
27
28Since kernel 2.3.99-pre6, this driver incorporates the support for the
293c575-series Cardbus cards which used to be handled by 3c575_cb.c.
30
31This driver supports the following hardware:
32
33	- 3c590 Vortex 10Mbps
34	- 3c592 EISA 10Mbps Demon/Vortex
35	- 3c597 EISA Fast Demon/Vortex
36	- 3c595 Vortex 100baseTx
37	- 3c595 Vortex 100baseT4
38	- 3c595 Vortex 100base-MII
39	- 3c900 Boomerang 10baseT
40	- 3c900 Boomerang 10Mbps Combo
41	- 3c900 Cyclone 10Mbps TPO
42	- 3c900 Cyclone 10Mbps Combo
43	- 3c900 Cyclone 10Mbps TPC
44	- 3c900B-FL Cyclone 10base-FL
45	- 3c905 Boomerang 100baseTx
46	- 3c905 Boomerang 100baseT4
47	- 3c905B Cyclone 100baseTx
48	- 3c905B Cyclone 10/100/BNC
49	- 3c905B-FX Cyclone 100baseFx
50	- 3c905C Tornado
51	- 3c920B-EMB-WNM (ATI Radeon 9100 IGP)
52	- 3c980 Cyclone
53	- 3c980C Python-T
54	- 3cSOHO100-TX Hurricane
55	- 3c555 Laptop Hurricane
56	- 3c556 Laptop Tornado
57	- 3c556B Laptop Hurricane
58	- 3c575 [Megahertz] 10/100 LAN  CardBus
59	- 3c575 Boomerang CardBus
60	- 3CCFE575BT Cyclone CardBus
61	- 3CCFE575CT Tornado CardBus
62	- 3CCFE656 Cyclone CardBus
63	- 3CCFEM656B Cyclone+Winmodem CardBus
64	- 3CXFEM656C Tornado+Winmodem CardBus
65	- 3c450 HomePNA Tornado
66	- 3c920 Tornado
67	- 3c982 Hydra Dual Port A
68	- 3c982 Hydra Dual Port B
69	- 3c905B-T4
70	- 3c920B-EMB-WNM Tornado
71
72Module parameters
73=================
74
75There are several parameters which may be provided to the driver when
76its module is loaded.  These are usually placed in ``/etc/modprobe.d/*.conf``
77configuration files.  Example::
78
79    options 3c59x debug=3 rx_copybreak=300
80
81If you are using the PCMCIA tools (cardmgr) then the options may be
82placed in /etc/pcmcia/config.opts::
83
84    module "3c59x" opts "debug=3 rx_copybreak=300"
85
86
87The supported parameters are:
88
89debug=N
90
91  Where N is a number from 0 to 7.  Anything above 3 produces a lot
92  of output in your system logs.  debug=1 is default.
93
94options=N1,N2,N3,...
95
96  Each number in the list provides an option to the corresponding
97  network card.  So if you have two 3c905's and you wish to provide
98  them with option 0x204 you would use::
99
100    options=0x204,0x204
101
102  The individual options are composed of a number of bitfields which
103  have the following meanings:
104
105  Possible media type settings
106
107	==	=================================
108	0	10baseT
109	1	10Mbs AUI
110	2	undefined
111	3	10base2 (BNC)
112	4	100base-TX
113	5	100base-FX
114	6	MII (Media Independent Interface)
115	7	Use default setting from EEPROM
116	8       Autonegotiate
117	9       External MII
118	10      Use default setting from EEPROM
119	==	=================================
120
121  When generating a value for the 'options' setting, the above media
122  selection values may be OR'ed (or added to) the following:
123
124  ======  =============================================
125  0x8000  Set driver debugging level to 7
126  0x4000  Set driver debugging level to 2
127  0x0400  Enable Wake-on-LAN
128  0x0200  Force full duplex mode.
129  0x0010  Bus-master enable bit (Old Vortex cards only)
130  ======  =============================================
131
132  For example::
133
134    insmod 3c59x options=0x204
135
136  will force full-duplex 100base-TX, rather than allowing the usual
137  autonegotiation.
138
139global_options=N
140
141  Sets the ``options`` parameter for all 3c59x NICs in the machine.
142  Entries in the ``options`` array above will override any setting of
143  this.
144
145full_duplex=N1,N2,N3...
146
147  Similar to bit 9 of 'options'.  Forces the corresponding card into
148  full-duplex mode.  Please use this in preference to the ``options``
149  parameter.
150
151  In fact, please don't use this at all! You're better off getting
152  autonegotiation working properly.
153
154global_full_duplex=N1
155
156  Sets full duplex mode for all 3c59x NICs in the machine.  Entries
157  in the ``full_duplex`` array above will override any setting of this.
158
159flow_ctrl=N1,N2,N3...
160
161  Use 802.3x MAC-layer flow control.  The 3com cards only support the
162  PAUSE command, which means that they will stop sending packets for a
163  short period if they receive a PAUSE frame from the link partner.
164
165  The driver only allows flow control on a link which is operating in
166  full duplex mode.
167
168  This feature does not appear to work on the 3c905 - only 3c905B and
169  3c905C have been tested.
170
171  The 3com cards appear to only respond to PAUSE frames which are
172  sent to the reserved destination address of 01:80:c2:00:00:01.  They
173  do not honour PAUSE frames which are sent to the station MAC address.
174
175rx_copybreak=M
176
177  The driver preallocates 32 full-sized (1536 byte) network buffers
178  for receiving.  When a packet arrives, the driver has to decide
179  whether to leave the packet in its full-sized buffer, or to allocate
180  a smaller buffer and copy the packet across into it.
181
182  This is a speed/space tradeoff.
183
184  The value of rx_copybreak is used to decide when to make the copy.
185  If the packet size is less than rx_copybreak, the packet is copied.
186  The default value for rx_copybreak is 200 bytes.
187
188max_interrupt_work=N
189
190  The driver's interrupt service routine can handle many receive and
191  transmit packets in a single invocation.  It does this in a loop.
192  The value of max_interrupt_work governs how many times the interrupt
193  service routine will loop.  The default value is 32 loops.  If this
194  is exceeded the interrupt service routine gives up and generates a
195  warning message "eth0: Too much work in interrupt".
196
197hw_checksums=N1,N2,N3,...
198
199  Recent 3com NICs are able to generate IPv4, TCP and UDP checksums
200  in hardware.  Linux has used the Rx checksumming for a long time.
201  The "zero copy" patch which is planned for the 2.4 kernel series
202  allows you to make use of the NIC's DMA scatter/gather and transmit
203  checksumming as well.
204
205  The driver is set up so that, when the zerocopy patch is applied,
206  all Tornado and Cyclone devices will use S/G and Tx checksums.
207
208  This module parameter has been provided so you can override this
209  decision.  If you think that Tx checksums are causing a problem, you
210  may disable the feature with ``hw_checksums=0``.
211
212  If you think your NIC should be performing Tx checksumming and the
213  driver isn't enabling it, you can force the use of hardware Tx
214  checksumming with ``hw_checksums=1``.
215
216  The driver drops a message in the logfiles to indicate whether or
217  not it is using hardware scatter/gather and hardware Tx checksums.
218
219  Scatter/gather and hardware checksums provide considerable
220  performance improvement for the sendfile() system call, but a small
221  decrease in throughput for send().  There is no effect upon receive
222  efficiency.
223
224compaq_ioaddr=N,
225compaq_irq=N,
226compaq_device_id=N
227
228  "Variables to work-around the Compaq PCI BIOS32 problem"....
229
230watchdog=N
231
232  Sets the time duration (in milliseconds) after which the kernel
233  decides that the transmitter has become stuck and needs to be reset.
234  This is mainly for debugging purposes, although it may be advantageous
235  to increase this value on LANs which have very high collision rates.
236  The default value is 5000 (5.0 seconds).
237
238enable_wol=N1,N2,N3,...
239
240  Enable Wake-on-LAN support for the relevant interface.  Donald
241  Becker's ``ether-wake`` application may be used to wake suspended
242  machines.
243
244  Also enables the NIC's power management support.
245
246global_enable_wol=N
247
248  Sets enable_wol mode for all 3c59x NICs in the machine.  Entries in
249  the ``enable_wol`` array above will override any setting of this.
250
251Media selection
252---------------
253
254A number of the older NICs such as the 3c590 and 3c900 series have
25510base2 and AUI interfaces.
256
257Prior to January, 2001 this driver would autoeselect the 10base2 or AUI
258port if it didn't detect activity on the 10baseT port.  It would then
259get stuck on the 10base2 port and a driver reload was necessary to
260switch back to 10baseT.  This behaviour could not be prevented with a
261module option override.
262
263Later (current) versions of the driver _do_ support locking of the
264media type.  So if you load the driver module with
265
266	modprobe 3c59x options=0
267
268it will permanently select the 10baseT port.  Automatic selection of
269other media types does not occur.
270
271
272Transmit error, Tx status register 82
273-------------------------------------
274
275This is a common error which is almost always caused by another host on
276the same network being in full-duplex mode, while this host is in
277half-duplex mode.  You need to find that other host and make it run in
278half-duplex mode or fix this host to run in full-duplex mode.
279
280As a last resort, you can force the 3c59x driver into full-duplex mode
281with
282
283	options 3c59x full_duplex=1
284
285but this has to be viewed as a workaround for broken network gear and
286should only really be used for equipment which cannot autonegotiate.
287
288
289Additional resources
290--------------------
291
292Details of the device driver implementation are at the top of the source file.
293
294Additional documentation is available at Don Becker's Linux Drivers site:
295
296     http://www.scyld.com/vortex.html
297
298Donald Becker's driver development site:
299
300     http://www.scyld.com/network.html
301
302Donald's vortex-diag program is useful for inspecting the NIC's state:
303
304     http://www.scyld.com/ethercard_diag.html
305
306Donald's mii-diag program may be used for inspecting and manipulating
307the NIC's Media Independent Interface subsystem:
308
309     http://www.scyld.com/ethercard_diag.html#mii-diag
310
311Donald's wake-on-LAN page:
312
313     http://www.scyld.com/wakeonlan.html
314
3153Com's DOS-based application for setting up the NICs EEPROMs:
316
317	ftp://ftp.3com.com/pub/nic/3c90x/3c90xx2.exe
318
319
320Autonegotiation notes
321---------------------
322
323  The driver uses a one-minute heartbeat for adapting to changes in
324  the external LAN environment if link is up and 5 seconds if link is down.
325  This means that when, for example, a machine is unplugged from a hubbed
326  10baseT LAN plugged into a  switched 100baseT LAN, the throughput
327  will be quite dreadful for up to sixty seconds.  Be patient.
328
329  Cisco interoperability note from Walter Wong <wcw+@CMU.EDU>:
330
331  On a side note, adding HAS_NWAY seems to share a problem with the
332  Cisco 6509 switch.  Specifically, you need to change the spanning
333  tree parameter for the port the machine is plugged into to 'portfast'
334  mode.  Otherwise, the negotiation fails.  This has been an issue
335  we've noticed for a while but haven't had the time to track down.
336
337  Cisco switches    (Jeff Busch <jbusch@deja.com>)
338
339    My "standard config" for ports to which PC's/servers connect directly::
340
341	interface FastEthernet0/N
342	description machinename
343	load-interval 30
344	spanning-tree portfast
345
346    If autonegotiation is a problem, you may need to specify "speed
347    100" and "duplex full" as well (or "speed 10" and "duplex half").
348
349    WARNING: DO NOT hook up hubs/switches/bridges to these
350    specially-configured ports! The switch will become very confused.
351
352
353Reporting and diagnosing problems
354---------------------------------
355
356Maintainers find that accurate and complete problem reports are
357invaluable in resolving driver problems.  We are frequently not able to
358reproduce problems and must rely on your patience and efforts to get to
359the bottom of the problem.
360
361If you believe you have a driver problem here are some of the
362steps you should take:
363
364- Is it really a driver problem?
365
366   Eliminate some variables: try different cards, different
367   computers, different cables, different ports on the switch/hub,
368   different versions of the kernel or of the driver, etc.
369
370- OK, it's a driver problem.
371
372   You need to generate a report.  Typically this is an email to the
373   maintainer and/or netdev@vger.kernel.org.  The maintainer's
374   email address will be in the driver source or in the MAINTAINERS file.
375
376- The contents of your report will vary a lot depending upon the
377  problem.  If it's a kernel crash then you should refer to the
378  admin-guide/reporting-bugs.rst file.
379
380  But for most problems it is useful to provide the following:
381
382   - Kernel version, driver version
383
384   - A copy of the banner message which the driver generates when
385     it is initialised.  For example:
386
387     eth0: 3Com PCI 3c905C Tornado at 0xa400,  00:50:da:6a:88:f0, IRQ 19
388     8K byte-wide RAM 5:3 Rx:Tx split, autoselect/Autonegotiate interface.
389     MII transceiver found at address 24, status 782d.
390     Enabling bus-master transmits and whole-frame receives.
391
392     NOTE: You must provide the ``debug=2`` modprobe option to generate
393     a full detection message.  Please do this::
394
395	modprobe 3c59x debug=2
396
397   - If it is a PCI device, the relevant output from 'lspci -vx', eg::
398
399       00:09.0 Ethernet controller: 3Com Corporation 3c905C-TX [Fast Etherlink] (rev 74)
400	       Subsystem: 3Com Corporation: Unknown device 9200
401	       Flags: bus master, medium devsel, latency 32, IRQ 19
402	       I/O ports at a400 [size=128]
403	       Memory at db000000 (32-bit, non-prefetchable) [size=128]
404	       Expansion ROM at <unassigned> [disabled] [size=128K]
405	       Capabilities: [dc] Power Management version 2
406       00: b7 10 00 92 07 00 10 02 74 00 00 02 08 20 00 00
407       10: 01 a4 00 00 00 00 00 db 00 00 00 00 00 00 00 00
408       20: 00 00 00 00 00 00 00 00 00 00 00 00 b7 10 00 10
409       30: 00 00 00 00 dc 00 00 00 00 00 00 00 05 01 0a 0a
410
411   - A description of the environment: 10baseT? 100baseT?
412     full/half duplex? switched or hubbed?
413
414   - Any additional module parameters which you may be providing to the driver.
415
416   - Any kernel logs which are produced.  The more the merrier.
417     If this is a large file and you are sending your report to a
418     mailing list, mention that you have the logfile, but don't send
419     it.  If you're reporting direct to the maintainer then just send
420     it.
421
422     To ensure that all kernel logs are available, add the
423     following line to /etc/syslog.conf::
424
425	 kern.* /var/log/messages
426
427     Then restart syslogd with::
428
429	 /etc/rc.d/init.d/syslog restart
430
431     (The above may vary, depending upon which Linux distribution you use).
432
433    - If your problem is reproducible then that's great.  Try the
434      following:
435
436      1) Increase the debug level.  Usually this is done via:
437
438	 a) modprobe driver debug=7
439	 b) In /etc/modprobe.d/driver.conf:
440	    options driver debug=7
441
442      2) Recreate the problem with the higher debug level,
443	 send all logs to the maintainer.
444
445      3) Download you card's diagnostic tool from Donald
446	 Becker's website <http://www.scyld.com/ethercard_diag.html>.
447	 Download mii-diag.c as well.  Build these.
448
449	 a) Run 'vortex-diag -aaee' and 'mii-diag -v' when the card is
450	    working correctly.  Save the output.
451
452	 b) Run the above commands when the card is malfunctioning.  Send
453	    both sets of output.
454
455Finally, please be patient and be prepared to do some work.  You may
456end up working on this problem for a week or more as the maintainer
457asks more questions, asks for more tests, asks for patches to be
458applied, etc.  At the end of it all, the problem may even remain
459unresolved.
460