1.. SPDX-License-Identifier: GPL-2.0
2
3===================================
4DEC EtherWORKS Ethernet De4x5 cards
5===================================
6
7    Originally,   this  driver  was    written  for the  Digital   Equipment
8    Corporation series of EtherWORKS Ethernet cards:
9
10	 - DE425 TP/COAX EISA
11	 - DE434 TP PCI
12	 - DE435 TP/COAX/AUI PCI
13	 - DE450 TP/COAX/AUI PCI
14	 - DE500 10/100 PCI Fasternet
15
16    but it  will  now attempt  to  support all  cards which   conform to the
17    Digital Semiconductor   SROM   Specification.    The  driver   currently
18    recognises the following chips:
19
20	 - DC21040  (no SROM)
21	 - DC21041[A]
22	 - DC21140[A]
23	 - DC21142
24	 - DC21143
25
26    So far the driver is known to work with the following cards:
27
28	 - KINGSTON
29	 - Linksys
30	 - ZNYX342
31	 - SMC8432
32	 - SMC9332 (w/new SROM)
33	 - ZNYX31[45]
34	 - ZNYX346 10/100 4 port (can act as a 10/100 bridge!)
35
36    The driver has been tested on a relatively busy network using the DE425,
37    DE434, DE435 and DE500 cards and benchmarked with 'ttcp': it transferred
38    16M of data to a DECstation 5000/200 as follows::
39
40		  TCP           UDP
41	       TX     RX     TX     RX
42      DE425   1030k  997k   1170k  1128k
43      DE434   1063k  995k   1170k  1125k
44      DE435   1063k  995k   1170k  1125k
45      DE500   1063k  998k   1170k  1125k  in 10Mb/s mode
46
47    All  values are typical (in   kBytes/sec) from a  sample  of 4 for  each
48    measurement. Their error is +/-20k on a quiet (private) network and also
49    depend on what load the CPU has.
50
51----------------------------------------------------------------------------
52
53    The ability to load this  driver as a loadable  module has been included
54    and used extensively  during the driver development  (to save those long
55    reboot sequences).  Loadable module support  under PCI and EISA has been
56    achieved by letting the driver autoprobe as if it were compiled into the
57    kernel. Do make sure  you're not sharing  interrupts with anything  that
58    cannot accommodate  interrupt  sharing!
59
60    To utilise this ability, you have to do 8 things:
61
62    0) have a copy of the loadable modules code installed on your system.
63    1) copy de4x5.c from the  /linux/drivers/net directory to your favourite
64       temporary directory.
65    2) for fixed  autoprobes (not  recommended),  edit the source code  near
66       line 5594 to reflect the I/O address  you're using, or assign these when
67       loading by::
68
69		   insmod de4x5 io=0xghh           where g = bus number
70							hh = device number
71
72       .. note::
73
74	   autoprobing for modules is now supported by default. You may just
75	   use::
76
77		   insmod de4x5
78
79	   to load all available boards. For a specific board, still use
80	   the 'io=?' above.
81    3) compile  de4x5.c, but include -DMODULE in  the command line to ensure
82       that the correct bits are compiled (see end of source code).
83    4) if you are wanting to add a new  card, goto 5. Otherwise, recompile a
84       kernel with the de4x5 configuration turned off and reboot.
85    5) insmod de4x5 [io=0xghh]
86    6) run the net startup bits for your new eth?? interface(s) manually
87       (usually /etc/rc.inet[12] at boot time).
88    7) enjoy!
89
90    To unload a module, turn off the associated interface(s)
91    'ifconfig eth?? down' then 'rmmod de4x5'.
92
93    Automedia detection is included so that in  principle you can disconnect
94    from, e.g.  TP, reconnect  to BNC  and  things will still work  (after a
95    pause while the   driver figures out   where its media went).  My tests
96    using ping showed that it appears to work....
97
98    By  default,  the driver will  now   autodetect any  DECchip based card.
99    Should you have a need to restrict the driver to DIGITAL only cards, you
100    can compile with a  DEC_ONLY define, or if  loading as a module, use the
101    'dec_only=1'  parameter.
102
103    I've changed the timing routines to  use the kernel timer and scheduling
104    functions  so that the  hangs  and other assorted problems that occurred
105    while autosensing the  media  should be gone.  A  bonus  for the DC21040
106    auto  media sense algorithm is  that it can now  use one that is more in
107    line with the  rest (the DC21040  chip doesn't  have a hardware  timer).
108    The downside is the 1 'jiffies' (10ms) resolution.
109
110    IEEE 802.3u MII interface code has  been added in anticipation that some
111    products may use it in the future.
112
113    The SMC9332 card  has a non-compliant SROM  which needs fixing -  I have
114    patched this  driver to detect it  because the SROM format used complies
115    to a previous DEC-STD format.
116
117    I have removed the buffer copies needed for receive on Intels.  I cannot
118    remove them for   Alphas since  the  Tulip hardware   only does longword
119    aligned  DMA transfers  and  the  Alphas get   alignment traps with  non
120    longword aligned data copies (which makes them really slow). No comment.
121
122    I  have added SROM decoding  routines to make this  driver work with any
123    card that  supports the Digital  Semiconductor SROM spec. This will help
124    all  cards running the dc2114x  series chips in particular.  Cards using
125    the dc2104x  chips should run correctly with  the basic  driver.  I'm in
126    debt to <mjacob@feral.com> for the  testing and feedback that helped get
127    this feature working.  So far we have  tested KINGSTON, SMC8432, SMC9332
128    (with the latest SROM complying  with the SROM spec  V3: their first was
129    broken), ZNYX342  and  LinkSys. ZNYX314 (dual  21041  MAC) and  ZNYX 315
130    (quad 21041 MAC)  cards also  appear  to work despite their  incorrectly
131    wired IRQs.
132
133    I have added a temporary fix for interrupt problems when some SCSI cards
134    share the same interrupt as the DECchip based  cards. The problem occurs
135    because  the SCSI card wants to  grab the interrupt  as a fast interrupt
136    (runs the   service routine with interrupts turned   off) vs.  this card
137    which really needs to run the service routine with interrupts turned on.
138    This driver will  now   add the interrupt service   routine  as  a  fast
139    interrupt if it   is bounced from the   slow interrupt.  THIS IS NOT   A
140    RECOMMENDED WAY TO RUN THE DRIVER  and has been done  for a limited time
141    until  people   sort  out their  compatibility    issues and the  kernel
142    interrupt  service code  is  fixed.   YOU  SHOULD SEPARATE OUT  THE FAST
143    INTERRUPT CARDS FROM THE SLOW INTERRUPT CARDS to ensure that they do not
144    run on the same interrupt. PCMCIA/CardBus is another can of worms...
145
146    Finally, I think  I have really  fixed  the module  loading problem with
147    more than one DECchip based  card.  As a  side effect, I don't mess with
148    the  device structure any  more which means that  if more than 1 card in
149    2.0.x is    installed (4  in   2.1.x),  the  user   will have   to  edit
150    linux/drivers/net/Space.c  to make room for  them. Hence, module loading
151    is  the preferred way to use   this driver, since  it  doesn't have this
152    limitation.
153
154    Where SROM media  detection is used and  full duplex is specified in the
155    SROM,  the feature is  ignored unless  lp->params.fdx  is set at compile
156    time  OR during  a   module load  (insmod  de4x5   args='eth??:fdx' [see
157    below]).  This is because there  is no way  to automatically detect full
158    duplex   links  except through   autonegotiation.    When I  include the
159    autonegotiation feature in  the SROM autoconf  code, this detection will
160    occur automatically for that case.
161
162    Command line  arguments are  now allowed, similar to  passing  arguments
163    through LILO. This will allow a per adapter board set  up of full duplex
164    and media. The only lexical constraints are:  the board name (dev->name)
165    appears in  the list before its parameters.  The list of parameters ends
166    either at the end of the parameter list or with another board name.  The
167    following parameters are allowed:
168
169	    =========  ===============================================
170	    fdx        for full duplex
171	    autosense  to set the media/speed; with the following
172		       sub-parameters:
173		       TP, TP_NW, BNC, AUI, BNC_AUI, 100Mb, 10Mb, AUTO
174	    =========  ===============================================
175
176    Case sensitivity is important  for  the sub-parameters. They *must*   be
177    upper case. Examples::
178
179	insmod de4x5 args='eth1:fdx autosense=BNC eth0:autosense=100Mb'.
180
181    For a compiled in driver, in linux/drivers/net/CONFIG, place e.g.::
182
183	DE4X5_OPTS = -DDE4X5_PARM='"eth0:fdx autosense=AUI eth2:autosense=TP"'
184
185    Yes,  I know full duplex  isn't permissible on BNC  or AUI; they're just
186    examples. By default, full duplex is turned  off and AUTO is the default
187    autosense setting. In  reality, I expect only the  full duplex option to
188    be used. Note the use of single quotes in the two examples above and the
189    lack of commas to separate items.
190