Magellan Linux

  • Back to homepage
    • /[alx-svn]/alx-src/tags/kernel26-2.6.12-alx-r9/Documentation/specialix.txt

    Contents of /alx-src/tags/kernel26-2.6.12-alx-r9/Documentation/specialix.txt

    Parent Directory Parent Directory | Revision Log Revision Log

    Revision 630 - (show annotations) (download)
    Wed Mar 4 11:03:09 2009 UTC (15 years, 2 months ago) by niro
    File MIME type: text/plain
    File size: 15155 byte(s) 
    Tag kernel26-2.6.12-alx-r9
    1
    2 specialix.txt -- specialix IO8+ multiport serial driver readme.
    3
    4
    5
    6 Copyright (C) 1997 Roger Wolff (R.E.Wolff@BitWizard.nl)
    7
    8 Specialix pays for the development and support of this driver.
    9 Please DO contact io8-linux@specialix.co.uk if you require
    10 support.
    11
    12 This driver was developed in the BitWizard linux device
    13 driver service. If you require a linux device driver for your
    14 product, please contact devices@BitWizard.nl for a quote.
    15
    16 This code is firmly based on the riscom/8 serial driver,
    17 written by Dmitry Gorodchanin. The specialix IO8+ card
    18 programming information was obtained from the CL-CD1865 Data
    19 Book, and Specialix document number 6200059: IO8+ Hardware
    20 Functional Specification, augmented by document number 6200088:
    21 Merak Hardware Functional Specification. (IO8+/PCI is also
    22 called Merak)
    23
    24
    25 This program is free software; you can redistribute it and/or
    26 modify it under the terms of the GNU General Public License as
    27 published by the Free Software Foundation; either version 2 of
    28 the License, or (at your option) any later version.
    29
    30 This program is distributed in the hope that it will be
    31 useful, but WITHOUT ANY WARRANTY; without even the implied
    32 warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR
    33 PURPOSE. See the GNU General Public License for more details.
    34
    35 You should have received a copy of the GNU General Public
    36 License along with this program; if not, write to the Free
    37 Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139,
    38 USA.
    39
    40
    41 Intro
    42 =====
    43
    44
    45 This file contains some random information, that I like to have online
    46 instead of in a manual that can get lost. Ever misplace your Linux
    47 kernel sources? And the manual of one of the boards in your computer?
    48
    49
    50 Addresses and interrupts
    51 ========================
    52
    53 Address dip switch settings:
    54 The dip switch sets bits 2-9 of the IO address.
    55
    56 9 8 7 6 5 4 3 2
    57 +-----------------+
    58 0 | X X X X X X X |
    59 | | = IoBase = 0x100
    60 1 | X |
    61 +-----------------+ ------ RS232 connectors ---->
    62
    63 | | |
    64 edge connector
    65 | | |
    66 V V V
    67
    68 Base address 0x100 caused a conflict in one of my computers once. I
    69 haven't the foggiest why. My Specialix card is now at 0x180. My
    70 other computer runs just fine with the Specialix card at 0x100....
    71 The card occupies 4 addresses, but actually only two are really used.
    72
    73 The PCI version doesn't have any dip switches. The BIOS assigns
    74 an IO address.
    75
    76 The driver now still autoprobes at 0x100, 0x180, 0x250 and 0x260. If
    77 that causes trouble for you, please report that. I'll remove
    78 autoprobing then.
    79
    80 The driver will tell the card what IRQ to use, so you don't have to
    81 change any jumpers to change the IRQ. Just use a command line
    82 argument (irq=xx) to the insmod program to set the interrupt.
    83
    84 The BIOS assigns the IRQ on the PCI version. You have no say in what
    85 IRQ to use in that case.
    86
    87 If your specialix cards are not at the default locations, you can use
    88 the kernel command line argument "specialix=io0,irq0,io1,irq1...".
    89 Here "io0" is the io address for the first card, and "irq0" is the
    90 irq line that the first card should use. And so on.
    91
    92 Examples.
    93
    94 You use the driver as a module and have three cards at 0x100, 0x250
    95 and 0x180. And some way or another you want them detected in that
    96 order. Moreover irq 12 is taken (e.g. by your PS/2 mouse).
    97
    98 insmod specialix.o iobase=0x100,0x250,0x180 irq=9,11,15
    99
    100 The same three cards, but now in the kernel would require you to
    101 add
    102
    103 specialix=0x100,9,0x250,11,0x180,15
    104
    105 to the command line. This would become
    106
    107 append="specialix=0x100,9,0x250,11,0x180,15"
    108
    109 in your /etc/lilo.conf file if you use lilo.
    110
    111 The Specialix driver is slightly odd: It allows you to have the second
    112 or third card detected without having a first card. This has
    113 advantages and disadvantages. A slot that isn't filled by an ISA card,
    114 might be filled if a PCI card is detected. Thus if you have an ISA
    115 card at 0x250 and a PCI card, you would get:
    116
    117 sx0: specialix IO8+ Board at 0x100 not found.
    118 sx1: specialix IO8+ Board at 0x180 not found.
    119 sx2: specialix IO8+ board detected at 0x250, IRQ 12, CD1865 Rev. B.
    120 sx3: specialix IO8+ Board at 0x260 not found.
    121 sx0: specialix IO8+ board detected at 0xd800, IRQ 9, CD1865 Rev. B.
    122
    123 This would happen if you don't give any probe hints to the driver.
    124 If you would specify:
    125
    126 specialix=0x250,11
    127
    128 you'd get the following messages:
    129
    130 sx0: specialix IO8+ board detected at 0x250, IRQ 11, CD1865 Rev. B.
    131 sx1: specialix IO8+ board detected at 0xd800, IRQ 9, CD1865 Rev. B.
    132
    133 ISA probing is aborted after the IO address you gave is exhausted, and
    134 the PCI card is now detected as the second card. The ISA card is now
    135 also forced to IRQ11....
    136
    137
    138 Baud rates
    139 ==========
    140
    141 The rev 1.2 and below boards use a CL-CD1864. These chips can only
    142 do 64kbit. The rev 1.3 and newer boards use a CL-CD1865. These chips
    143 are officially capable of 115k2.
    144
    145 The Specialix card uses a 25MHz crystal (in times two mode, which in
    146 fact is a divided by two mode). This is not enough to reach the rated
    147 115k2 on all ports at the same time. With this clock rate you can only
    148 do 37% of this rate. This means that at 115k2 on all ports you are
    149 going to lose characters (The chip cannot handle that many incoming
    150 bits at this clock rate.) (Yes, you read that correctly: there is a
    151 limit to the number of -=bits=- per second that the chip can handle.)
    152
    153 If you near the "limit" you will first start to see a graceful
    154 degradation in that the chip cannot keep the transmitter busy at all
    155 times. However with a central clock this slow, you can also get it to
    156 miss incoming characters. The driver will print a warning message when
    157 you are outside the official specs. The messages usually show up in
    158 the file /var/log/messages .
    159
    160 The specialix card cannot reliably do 115k2. If you use it, you have
    161 to do "extensive testing" (*) to verify if it actually works.
    162
    163 When "mgetty" communicates with my modem at 115k2 it reports:
    164 got: +++[0d]ATQ0V1H0[0d][0d][8a]O[cb][0d][8a]
    165 ^^^^ ^^^^ ^^^^
    166
    167 The three characters that have the "^^^" under them have suffered a
    168 bit error in the highest bit. In conclusion: I've tested it, and found
    169 that it simply DOESN'T work for me. I also suspect that this is also
    170 caused by the baud rate being just a little bit out of tune.
    171
    172 I upgraded the crystal to 66Mhz on one of my Specialix cards. Works
    173 great! Contact me for details. (Voids warranty, requires a steady hand
    174 and more such restrictions....)
    175
    176
    177 (*) Cirrus logic CD1864 databook, page 40.
    178
    179
    180 Cables for the Specialix IO8+
    181 =============================
    182
    183 The pinout of the connectors on the IO8+ is:
    184
    185 pin short direction long name
    186 name
    187 Pin 1 DCD input Data Carrier Detect
    188 Pin 2 RXD input Receive
    189 Pin 3 DTR/RTS output Data Terminal Ready/Ready To Send
    190 Pin 4 GND - Ground
    191 Pin 5 TXD output Transmit
    192 Pin 6 CTS input Clear To Send
    193
    194
    195 -- 6 5 4 3 2 1 --
    196 | |
    197 | |
    198 | |
    199 | |
    200 +----- -----+
    201 |__________|
    202 clip
    203
    204 Front view of an RJ12 connector. Cable moves "into" the paper.
    205 (the plug is ready to plug into your mouth this way...)
    206
    207
    208 NULL cable. I don't know who is going to use these except for
    209 testing purposes, but I tested the cards with this cable. (It
    210 took quite a while to figure out, so I'm not going to delete
    211 it. So there! :-)
    212
    213
    214 This end goes This end needs
    215 straight into the some twists in
    216 RJ12 plug. the wiring.
    217 IO8+ RJ12 IO8+ RJ12
    218 1 DCD white -
    219 - - 1 DCD
    220 2 RXD black 5 TXD
    221 3 DTR/RTS red 6 CTS
    222 4 GND green 4 GND
    223 5 TXD yellow 2 RXD
    224 6 CTS blue 3 DTR/RTS
    225
    226
    227 Same NULL cable, but now sorted on the second column.
    228
    229 1 DCD white -
    230 - - 1 DCD
    231 5 TXD yellow 2 RXD
    232 6 CTS blue 3 DTR/RTS
    233 4 GND green 4 GND
    234 2 RXD black 5 TXD
    235 3 DTR/RTS red 6 CTS
    236
    237
    238
    239 This is a modem cable usable for hardware handshaking:
    240 RJ12 DB25 DB9
    241 1 DCD white 8 DCD 1 DCD
    242 2 RXD black 3 RXD 2 RXD
    243 3 DTR/RTS red 4 RTS 7 RTS
    244 4 GND green 7 GND 5 GND
    245 5 TXD yellow 2 TXD 3 TXD
    246 6 CTS blue 5 CTS 8 CTS
    247 +---- 6 DSR 6 DSR
    248 +---- 20 DTR 4 DTR
    249
    250 This is a modem cable usable for software handshaking:
    251 It allows you to reset the modem using the DTR ioctls.
    252 I (REW) have never tested this, "but xxxxxxxxxxxxx
    253 says that it works." If you test this, please
    254 tell me and I'll fill in your name on the xxx's.
    255
    256 RJ12 DB25 DB9
    257 1 DCD white 8 DCD 1 DCD
    258 2 RXD black 3 RXD 2 RXD
    259 3 DTR/RTS red 20 DTR 4 DTR
    260 4 GND green 7 GND 5 GND
    261 5 TXD yellow 2 TXD 3 TXD
    262 6 CTS blue 5 CTS 8 CTS
    263 +---- 6 DSR 6 DSR
    264 +---- 4 RTS 7 RTS
    265
    266 I bought a 6 wire flat cable. It was colored as indicated.
    267 Check that yours is the same before you trust me on this.
    268
    269
    270 Hardware handshaking issues.
    271 ============================
    272
    273 The driver can be compiled in two different ways. The default
    274 ("Specialix DTR/RTS pin is RTS" is off) the pin behaves as DTR when
    275 hardware handshaking is off. It behaves as the RTS hardware
    276 handshaking signal when hardware handshaking is selected.
    277
    278 When you use this, you have to use the appropriate cable. The
    279 cable will either be compatible with hardware handshaking or with
    280 software handshaking. So switching on the fly is not really an
    281 option.
    282
    283 I actually prefer to use the "Specialix DTR/RTS pin is RTS" option.
    284 This makes the DTR/RTS pin always an RTS pin, and ioctls to
    285 change DTR are always ignored. I have a cable that is configured
    286 for this.
    287
    288
    289 Ports and devices
    290 =================
    291
    292 Port 0 is the one furthest from the card-edge connector.
    293
    294 Devices:
    295
    296 You should make the devices as follows:
    297
    298 bash
    299 cd /dev
    300 for i in 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 \
    301 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
    302 do
    303 echo -n "$i "
    304 mknod /dev/ttyW$i c 75 $i
    305 mknod /dev/cuw$i c 76 $i
    306 done
    307 echo ""
    308
    309 If your system doesn't come with these devices preinstalled, bug your
    310 linux-vendor about this. They have had ample time to get this
    311 implemented by now.
    312
    313 You cannot have more than 4 boards in one computer. The card only
    314 supports 4 different interrupts. If you really want this, contact me
    315 about this and I'll give you a few tips (requires soldering iron)....
    316
    317 If you have enough PCI slots, you can probably use more than 4 PCI
    318 versions of the card though....
    319
    320 The PCI version of the card cannot adhere to the mechanical part of
    321 the PCI spec because the 8 serial connectors are simply too large. If
    322 it doesn't fit in your computer, bring back the card.
    323
    324
    325 ------------------------------------------------------------------------
    326
    327
    328 Fixed bugs and restrictions:
    329 - During initialization, interrupts are blindly turned on.
    330 Having a shadow variable would cause an extra memory
    331 access on every IO instruction.
    332 - The interrupt (on the card) should be disabled when we
    333 don't allocate the Linux end of the interrupt. This allows
    334 a different driver/card to use it while all ports are not in
    335 use..... (a la standard serial port)
    336 == An extra _off variant of the sx_in and sx_out macros are
    337 now available. They don't set the interrupt enable bit.
    338 These are used during initialization. Normal operation uses
    339 the old variant which enables the interrupt line.
    340 - RTS/DTR issue needs to be implemented according to
    341 specialix' spec.
    342 I kind of like the "determinism" of the current
    343 implementation. Compile time flag?
    344 == Ok. Compile time flag! Default is how Specialix likes it.
    345 == Now a config time flag! Gets saved in your config file. Neat!
    346 - Can you set the IO address from the lilo command line?
    347 If you need this, bug me about it, I'll make it.
    348 == Hah! No bugging needed. Fixed! :-)
    349 - Cirrus logic hasn't gotten back to me yet why the CD1865 can
    350 and the CD1864 can't do 115k2. I suspect that this is
    351 because the CD1864 is not rated for 33MHz operation.
    352 Therefore the CD1864 versions of the card can't do 115k2 on
    353 all ports just like the CD1865 versions. The driver does
    354 not block 115k2 on CD1864 cards.
    355 == I called the Cirrus Logic representative here in Holland.
    356 The CD1864 databook is identical to the CD1865 databook,
    357 except for an extra warning at the end. Similar Bit errors
    358 have been observed in testing at 115k2 on both an 1865 and
    359 a 1864 chip. I see no reason why I would prohibit 115k2 on
    360 1864 chips and not do it on 1865 chips. Actually there is
    361 reason to prohibit it on BOTH chips. I print a warning.
    362 If you use 115k2, you're on your own.
    363 - A spiky CD may send spurious HUPs. Also in CLOCAL???
    364 -- A fix for this turned out to be counter productive.
    365 Different fix? Current behaviour is acceptable?
    366 -- Maybe the current implementation is correct. If anybody
    367 gets bitten by this, please report, and it will get fixed.
    368
    369 -- Testing revealed that when in CLOCAL, the problem doesn't
    370 occur. As warned for in the CD1865 manual, the chip may
    371 send modem intr's on a spike. We could filter those out,
    372 but that would be a cludge anyway (You'd still risk getting
    373 a spurious HUP when two spikes occur.).....
    374
    375
    376
    377 Bugs & restrictions:
    378 - This is a difficult card to autoprobe.
    379 You have to WRITE to the address register to even
    380 read-probe a CD186x register. Disable autodetection?
    381 -- Specialix: any suggestions?
    382 - Arbitrary baud rates are not implemented yet.
    383 If you need this, bug me about it.
    384
    385