summaryrefslogtreecommitdiffstats
path: root/Documentation/serial
diff options
context:
space:
mode:
authorTakashi Iwai <tiwai@suse.de>2014-05-05 16:54:33 +0200
committerTakashi Iwai <tiwai@suse.de>2014-05-05 16:54:33 +0200
commit59991da498122f779434c4e0beac8a69e0322938 (patch)
tree23e46bc7f355c6d65f3246f756cd8ef84cfc3030 /Documentation/serial
parentALSA: fm801: introduce fm801_ac97_is_ready()/fm801_ac97_is_valid() helpers (diff)
parentALSA: hda - hdmi: Set converter channel count even without sink (diff)
downloadlinux-59991da498122f779434c4e0beac8a69e0322938.tar.xz
linux-59991da498122f779434c4e0beac8a69e0322938.zip
Merge branch 'for-linus' into for-next
... for applying the further HDMI fixes.
Diffstat (limited to 'Documentation/serial')
-rw-r--r--Documentation/serial/00-INDEX8
-rw-r--r--Documentation/serial/digiepca.txt98
-rw-r--r--Documentation/serial/riscom8.txt36
-rw-r--r--Documentation/serial/specialix.txt383
-rw-r--r--Documentation/serial/sx.txt294
5 files changed, 0 insertions, 819 deletions
diff --git a/Documentation/serial/00-INDEX b/Documentation/serial/00-INDEX
index f9c6b5ed03e7..8021a9f29fc5 100644
--- a/Documentation/serial/00-INDEX
+++ b/Documentation/serial/00-INDEX
@@ -2,23 +2,15 @@
- this file.
README.cycladesZ
- info on Cyclades-Z firmware loading.
-digiepca.txt
- - info on Digi Intl. {PC,PCI,EISA}Xx and Xem series cards.
driver
- intro to the low level serial driver.
moxa-smartio
- file with info on installing/using Moxa multiport serial driver.
n_gsm.txt
- GSM 0710 tty multiplexer howto.
-riscom8.txt
- - notes on using the RISCom/8 multi-port serial driver.
rocket.txt
- info on the Comtrol RocketPort multiport serial driver.
serial-rs485.txt
- info about RS485 structures and support in the kernel.
-specialix.txt
- - info on hardware/driver for specialix IO8+ multiport serial card.
-sx.txt
- - info on the Specialix SX/SI multiport serial driver.
tty.txt
- guide to the locking policies of the tty layer.
diff --git a/Documentation/serial/digiepca.txt b/Documentation/serial/digiepca.txt
deleted file mode 100644
index f2560e22f2c9..000000000000
--- a/Documentation/serial/digiepca.txt
+++ /dev/null
@@ -1,98 +0,0 @@
-NOTE: This driver is obsolete. Digi provides a 2.6 driver (dgdm) at
-http://www.digi.com for PCI cards. They no longer maintain this driver,
-and have no 2.6 driver for ISA cards.
-
-This driver requires a number of user-space tools. They can be acquired from
-http://www.digi.com, but only works with 2.4 kernels.
-
-
-The Digi Intl. epca driver.
-----------------------------
-The Digi Intl. epca driver for Linux supports the following boards:
-
-Digi PC/Xem, PC/Xr, PC/Xe, PC/Xi, PC/Xeve
-Digi EISA/Xem, PCI/Xem, PCI/Xr
-
-Limitations:
-------------
-Currently the driver only autoprobes for supported PCI boards.
-
-The Linux MAKEDEV command does not support generating the Digiboard
-Devices. Users executing digiConfig to setup EISA and PC series cards
-will have their device nodes automatically constructed (cud?? for ~CLOCAL,
-and ttyD?? for CLOCAL). Users wishing to boot their board from the LILO
-prompt, or those users booting PCI cards may use buildDIGI to construct
-the necessary nodes.
-
-Notes:
-------
-This driver may be configured via LILO. For users who have already configured
-their driver using digiConfig, configuring from LILO will override previous
-settings. Multiple boards may be configured by issuing multiple LILO command
-lines. For examples see the bottom of this document.
-
-Device names start at 0 and continue up. Beware of this as previous Digi
-drivers started device names with 1.
-
-PCI boards are auto-detected and configured by the driver. PCI boards will
-be allocated device numbers (internally) beginning with the lowest PCI slot
-first. In other words a PCI card in slot 3 will always have higher device
-nodes than a PCI card in slot 1.
-
-LILO config examples:
----------------------
-Using LILO's APPEND command, a string of comma separated identifiers or
-integers can be used to configure supported boards. The six values in order
-are:
-
- Enable/Disable this card or Override,
- Type of card: PC/Xe (AccelePort) (0), PC/Xeve (1), PC/Xem or PC/Xr (2),
- EISA/Xem (3), PC/64Xe (4), PC/Xi (5),
- Enable/Disable alternate pin arrangement,
- Number of ports on this card,
- I/O Port where card is configured (in HEX if using string identifiers),
- Base of memory window (in HEX if using string identifiers),
-
-NOTE : PCI boards are auto-detected and configured. Do not attempt to
-configure PCI boards with the LILO append command. If you wish to override
-previous configuration data (As set by digiConfig), but you do not wish to
-configure any specific card (Example if there are PCI cards in the system)
-the following override command will accomplish this:
--> append="digi=2"
-
-Samples:
- append="digiepca=E,PC/Xe,D,16,200,D0000"
- or
- append="digi=1,0,0,16,512,851968"
-
-Supporting Tools:
------------------
-Supporting tools include digiDload, digiConfig, buildPCI, and ditty. See
-drivers/char/README.epca for more details. Note,
-this driver REQUIRES that digiDload be executed prior to it being used.
-Failure to do this will result in an ENODEV error.
-
-Documentation:
---------------
-Complete documentation for this product may be found in the tool package.
-
-Sources of information and support:
------------------------------------
-Digi Intl. support site for this product:
-
--> http://www.digi.com
-
-Acknowledgments:
-----------------
-Much of this work (And even text) was derived from a similar document
-supporting the original public domain DigiBoard driver Copyright (C)
-1994,1995 Troy De Jongh. Many thanks to Christoph Lameter
-(christoph@lameter.com) and Mike McLagan (mike.mclagan@linux.org) who authored
-and contributed to the original document.
-
-Changelog:
-----------
-10-29-04: Update status of driver, remove dead links in document
- James Nelson <james4765@gmail.com>
-
-2000 (?) Original Document
diff --git a/Documentation/serial/riscom8.txt b/Documentation/serial/riscom8.txt
deleted file mode 100644
index 14f61fdad7ca..000000000000
--- a/Documentation/serial/riscom8.txt
+++ /dev/null
@@ -1,36 +0,0 @@
-* NOTE - this is an unmaintained driver. The original author cannot be located.
-
-SDL Communications is now SBS Technologies, and does not have any
-information on these ancient ISA cards on their website.
-
-James Nelson <james4765@gmail.com> - 12-12-2004
-
- This is the README for RISCom/8 multi-port serial driver
- (C) 1994-1996 D.Gorodchanin
- See file LICENSE for terms and conditions.
-
-NOTE: English is not my native language.
- I'm sorry for any mistakes in this text.
-
-Misc. notes for RISCom/8 serial driver, in no particular order :)
-
-1) This driver can support up to 4 boards at time.
- Use string "riscom8=0xXXX,0xXXX,0xXXX,0xXXX" at LILO prompt, for
- setting I/O base addresses for boards. If you compile driver
- as module use modprobe options "iobase=0xXXX iobase1=0xXXX iobase2=..."
-
-2) The driver partially supports famous 'setserial' program, you can use almost
- any of its options, excluding port & irq settings.
-
-3) There are some misc. defines at the beginning of riscom8.c, please read the
- comments and try to change some of them in case of problems.
-
-4) I consider the current state of the driver as BETA.
-
-5) SDL Communications WWW page is http://www.sdlcomm.com.
-
-6) You can use the MAKEDEV program to create RISCom/8 /dev/ttyL* entries.
-
-7) Minor numbers for first board are 0-7, for second 8-15, etc.
-
-22 Apr 1996.
diff --git a/Documentation/serial/specialix.txt b/Documentation/serial/specialix.txt
deleted file mode 100644
index 6eb6f3a3331c..000000000000
--- a/Documentation/serial/specialix.txt
+++ /dev/null
@@ -1,383 +0,0 @@
-
- specialix.txt -- specialix IO8+ multiport serial driver readme.
-
-
-
- Copyright (C) 1997 Roger Wolff (R.E.Wolff@BitWizard.nl)
-
- Specialix pays for the development and support of this driver.
- Please DO contact io8-linux@specialix.co.uk if you require
- support.
-
- This driver was developed in the BitWizard linux device
- driver service. If you require a linux device driver for your
- product, please contact devices@BitWizard.nl for a quote.
-
- This code is firmly based on the riscom/8 serial driver,
- written by Dmitry Gorodchanin. The specialix IO8+ card
- programming information was obtained from the CL-CD1865 Data
- Book, and Specialix document number 6200059: IO8+ Hardware
- Functional Specification, augmented by document number 6200088:
- Merak Hardware Functional Specification. (IO8+/PCI is also
- called Merak)
-
-
- This program is free software; you can redistribute it and/or
- modify it under the terms of the GNU General Public License as
- published by the Free Software Foundation; either version 2 of
- the License, or (at your option) any later version.
-
- This program is distributed in the hope that it will be
- useful, but WITHOUT ANY WARRANTY; without even the implied
- warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR
- PURPOSE. See the GNU General Public License for more details.
-
- You should have received a copy of the GNU General Public
- License along with this program; if not, write to the Free
- Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139,
- USA.
-
-
-Intro
-=====
-
-
-This file contains some random information, that I like to have online
-instead of in a manual that can get lost. Ever misplace your Linux
-kernel sources? And the manual of one of the boards in your computer?
-
-
-Addresses and interrupts
-========================
-
-Address dip switch settings:
-The dip switch sets bits 2-9 of the IO address.
-
- 9 8 7 6 5 4 3 2
- +-----------------+
- 0 | X X X X X X X |
- | | = IoBase = 0x100
- 1 | X |
- +-----------------+ ------ RS232 connectors ---->
-
- | | |
- edge connector
- | | |
- V V V
-
-Base address 0x100 caused a conflict in one of my computers once. I
-haven't the foggiest why. My Specialix card is now at 0x180. My
-other computer runs just fine with the Specialix card at 0x100....
-The card occupies 4 addresses, but actually only two are really used.
-
-The PCI version doesn't have any dip switches. The BIOS assigns
-an IO address.
-
-The driver now still autoprobes at 0x100, 0x180, 0x250 and 0x260. If
-that causes trouble for you, please report that. I'll remove
-autoprobing then.
-
-The driver will tell the card what IRQ to use, so you don't have to
-change any jumpers to change the IRQ. Just use a command line
-argument (irq=xx) to the insmod program to set the interrupt.
-
-The BIOS assigns the IRQ on the PCI version. You have no say in what
-IRQ to use in that case.
-
-If your specialix cards are not at the default locations, you can use
-the kernel command line argument "specialix=io0,irq0,io1,irq1...".
-Here "io0" is the io address for the first card, and "irq0" is the
-irq line that the first card should use. And so on.
-
-Examples.
-
-You use the driver as a module and have three cards at 0x100, 0x250
-and 0x180. And some way or another you want them detected in that
-order. Moreover irq 12 is taken (e.g. by your PS/2 mouse).
-
- insmod specialix.o iobase=0x100,0x250,0x180 irq=9,11,15
-
-The same three cards, but now in the kernel would require you to
-add
-
- specialix=0x100,9,0x250,11,0x180,15
-
-to the command line. This would become
-
- append="specialix=0x100,9,0x250,11,0x180,15"
-
-in your /etc/lilo.conf file if you use lilo.
-
-The Specialix driver is slightly odd: It allows you to have the second
-or third card detected without having a first card. This has
-advantages and disadvantages. A slot that isn't filled by an ISA card,
-might be filled if a PCI card is detected. Thus if you have an ISA
-card at 0x250 and a PCI card, you would get:
-
-sx0: specialix IO8+ Board at 0x100 not found.
-sx1: specialix IO8+ Board at 0x180 not found.
-sx2: specialix IO8+ board detected at 0x250, IRQ 12, CD1865 Rev. B.
-sx3: specialix IO8+ Board at 0x260 not found.
-sx0: specialix IO8+ board detected at 0xd800, IRQ 9, CD1865 Rev. B.
-
-This would happen if you don't give any probe hints to the driver.
-If you would specify:
-
- specialix=0x250,11
-
-you'd get the following messages:
-
-sx0: specialix IO8+ board detected at 0x250, IRQ 11, CD1865 Rev. B.
-sx1: specialix IO8+ board detected at 0xd800, IRQ 9, CD1865 Rev. B.
-
-ISA probing is aborted after the IO address you gave is exhausted, and
-the PCI card is now detected as the second card. The ISA card is now
-also forced to IRQ11....
-
-
-Baud rates
-==========
-
-The rev 1.2 and below boards use a CL-CD1864. These chips can only
-do 64kbit. The rev 1.3 and newer boards use a CL-CD1865. These chips
-are officially capable of 115k2.
-
-The Specialix card uses a 25MHz crystal (in times two mode, which in
-fact is a divided by two mode). This is not enough to reach the rated
-115k2 on all ports at the same time. With this clock rate you can only
-do 37% of this rate. This means that at 115k2 on all ports you are
-going to lose characters (The chip cannot handle that many incoming
-bits at this clock rate.) (Yes, you read that correctly: there is a
-limit to the number of -=bits=- per second that the chip can handle.)
-
-If you near the "limit" you will first start to see a graceful
-degradation in that the chip cannot keep the transmitter busy at all
-times. However with a central clock this slow, you can also get it to
-miss incoming characters. The driver will print a warning message when
-you are outside the official specs. The messages usually show up in
-the file /var/log/messages .
-
-The specialix card cannot reliably do 115k2. If you use it, you have
-to do "extensive testing" (*) to verify if it actually works.
-
-When "mgetty" communicates with my modem at 115k2 it reports:
-got: +++[0d]ATQ0V1H0[0d][0d][8a]O[cb][0d][8a]
- ^^^^ ^^^^ ^^^^
-
-The three characters that have the "^^^" under them have suffered a
-bit error in the highest bit. In conclusion: I've tested it, and found
-that it simply DOESN'T work for me. I also suspect that this is also
-caused by the baud rate being just a little bit out of tune.
-
-I upgraded the crystal to 66Mhz on one of my Specialix cards. Works
-great! Contact me for details. (Voids warranty, requires a steady hand
-and more such restrictions....)
-
-
-(*) Cirrus logic CD1864 databook, page 40.
-
-
-Cables for the Specialix IO8+
-=============================
-
-The pinout of the connectors on the IO8+ is:
-
- pin short direction long name
- name
- Pin 1 DCD input Data Carrier Detect
- Pin 2 RXD input Receive
- Pin 3 DTR/RTS output Data Terminal Ready/Ready To Send
- Pin 4 GND - Ground
- Pin 5 TXD output Transmit
- Pin 6 CTS input Clear To Send
-
-
- -- 6 5 4 3 2 1 --
- | |
- | |
- | |
- | |
- +----- -----+
- |__________|
- clip
-
- Front view of an RJ12 connector. Cable moves "into" the paper.
- (the plug is ready to plug into your mouth this way...)
-
-
- NULL cable. I don't know who is going to use these except for
- testing purposes, but I tested the cards with this cable. (It
- took quite a while to figure out, so I'm not going to delete
- it. So there! :-)
-
-
- This end goes This end needs
- straight into the some twists in
- RJ12 plug. the wiring.
- IO8+ RJ12 IO8+ RJ12
- 1 DCD white -
- - - 1 DCD
- 2 RXD black 5 TXD
- 3 DTR/RTS red 6 CTS
- 4 GND green 4 GND
- 5 TXD yellow 2 RXD
- 6 CTS blue 3 DTR/RTS
-
-
- Same NULL cable, but now sorted on the second column.
-
- 1 DCD white -
- - - 1 DCD
- 5 TXD yellow 2 RXD
- 6 CTS blue 3 DTR/RTS
- 4 GND green 4 GND
- 2 RXD black 5 TXD
- 3 DTR/RTS red 6 CTS
-
-
-
- This is a modem cable usable for hardware handshaking:
- RJ12 DB25 DB9
- 1 DCD white 8 DCD 1 DCD
- 2 RXD black 3 RXD 2 RXD
- 3 DTR/RTS red 4 RTS 7 RTS
- 4 GND green 7 GND 5 GND
- 5 TXD yellow 2 TXD 3 TXD
- 6 CTS blue 5 CTS 8 CTS
- +---- 6 DSR 6 DSR
- +---- 20 DTR 4 DTR
-
- This is a modem cable usable for software handshaking:
- It allows you to reset the modem using the DTR ioctls.
- I (REW) have never tested this, "but xxxxxxxxxxxxx
- says that it works." If you test this, please
- tell me and I'll fill in your name on the xxx's.
-
- RJ12 DB25 DB9
- 1 DCD white 8 DCD 1 DCD
- 2 RXD black 3 RXD 2 RXD
- 3 DTR/RTS red 20 DTR 4 DTR
- 4 GND green 7 GND 5 GND
- 5 TXD yellow 2 TXD 3 TXD
- 6 CTS blue 5 CTS 8 CTS
- +---- 6 DSR 6 DSR
- +---- 4 RTS 7 RTS
-
- I bought a 6 wire flat cable. It was colored as indicated.
- Check that yours is the same before you trust me on this.
-
-
-Hardware handshaking issues.
-============================
-
-The driver can be told to operate in two different ways. The default
-behaviour is specialix.sx_rtscts = 0 where the pin behaves as DTR when
-hardware handshaking is off. It behaves as the RTS hardware
-handshaking signal when hardware handshaking is selected.
-
-When you use this, you have to use the appropriate cable. The
-cable will either be compatible with hardware handshaking or with
-software handshaking. So switching on the fly is not really an
-option.
-
-I actually prefer to use the "specialix.sx_rtscts=1" option.
-This makes the DTR/RTS pin always an RTS pin, and ioctls to
-change DTR are always ignored. I have a cable that is configured
-for this.
-
-
-Ports and devices
-=================
-
-Port 0 is the one furthest from the card-edge connector.
-
-Devices:
-
-You should make the devices as follows:
-
-bash
-cd /dev
-for i in 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 \
- 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
-do
- echo -n "$i "
- mknod /dev/ttyW$i c 75 $i
- mknod /dev/cuw$i c 76 $i
-done
-echo ""
-
-If your system doesn't come with these devices preinstalled, bug your
-linux-vendor about this. They have had ample time to get this
-implemented by now.
-
-You cannot have more than 4 boards in one computer. The card only
-supports 4 different interrupts. If you really want this, contact me
-about this and I'll give you a few tips (requires soldering iron)....
-
-If you have enough PCI slots, you can probably use more than 4 PCI
-versions of the card though....
-
-The PCI version of the card cannot adhere to the mechanical part of
-the PCI spec because the 8 serial connectors are simply too large. If
-it doesn't fit in your computer, bring back the card.
-
-
-------------------------------------------------------------------------
-
-
- Fixed bugs and restrictions:
- - During initialization, interrupts are blindly turned on.
- Having a shadow variable would cause an extra memory
- access on every IO instruction.
- - The interrupt (on the card) should be disabled when we
- don't allocate the Linux end of the interrupt. This allows
- a different driver/card to use it while all ports are not in
- use..... (a la standard serial port)
- == An extra _off variant of the sx_in and sx_out macros are
- now available. They don't set the interrupt enable bit.
- These are used during initialization. Normal operation uses
- the old variant which enables the interrupt line.
- - RTS/DTR issue needs to be implemented according to
- specialix' spec.
- I kind of like the "determinism" of the current
- implementation. Compile time flag?
- == Ok. Compile time flag! Default is how Specialix likes it.
- == Now a config time flag! Gets saved in your config file. Neat!
- - Can you set the IO address from the lilo command line?
- If you need this, bug me about it, I'll make it.
- == Hah! No bugging needed. Fixed! :-)
- - Cirrus logic hasn't gotten back to me yet why the CD1865 can
- and the CD1864 can't do 115k2. I suspect that this is
- because the CD1864 is not rated for 33MHz operation.
- Therefore the CD1864 versions of the card can't do 115k2 on
- all ports just like the CD1865 versions. The driver does
- not block 115k2 on CD1864 cards.
- == I called the Cirrus Logic representative here in Holland.
- The CD1864 databook is identical to the CD1865 databook,
- except for an extra warning at the end. Similar Bit errors
- have been observed in testing at 115k2 on both an 1865 and
- a 1864 chip. I see no reason why I would prohibit 115k2 on
- 1864 chips and not do it on 1865 chips. Actually there is
- reason to prohibit it on BOTH chips. I print a warning.
- If you use 115k2, you're on your own.
- - A spiky CD may send spurious HUPs. Also in CLOCAL???
- -- A fix for this turned out to be counter productive.
- Different fix? Current behaviour is acceptable?
- -- Maybe the current implementation is correct. If anybody
- gets bitten by this, please report, and it will get fixed.
-
- -- Testing revealed that when in CLOCAL, the problem doesn't
- occur. As warned for in the CD1865 manual, the chip may
- send modem intr's on a spike. We could filter those out,
- but that would be a cludge anyway (You'd still risk getting
- a spurious HUP when two spikes occur.).....
-
-
-
- Bugs & restrictions:
- - This is a difficult card to autoprobe.
- You have to WRITE to the address register to even
- read-probe a CD186x register. Disable autodetection?
- -- Specialix: any suggestions?
-
-
diff --git a/Documentation/serial/sx.txt b/Documentation/serial/sx.txt
deleted file mode 100644
index cb4efa0fb5cc..000000000000
--- a/Documentation/serial/sx.txt
+++ /dev/null
@@ -1,294 +0,0 @@
-
- sx.txt -- specialix SX/SI multiport serial driver readme.
-
-
-
- Copyright (C) 1997 Roger Wolff (R.E.Wolff@BitWizard.nl)
-
- Specialix pays for the development and support of this driver.
- Please DO contact support@specialix.co.uk if you require
- support.
-
- This driver was developed in the BitWizard linux device
- driver service. If you require a linux device driver for your
- product, please contact devices@BitWizard.nl for a quote.
-
- (History)
- There used to be an SI driver by Simon Allan. This is a complete
- rewrite from scratch. Just a few lines-of-code have been snatched.
-
- (Sources)
- Specialix document number 6210028: SX Host Card and Download Code
- Software Functional Specification.
-
- (Copying)
- This program is free software; you can redistribute it and/or
- modify it under the terms of the GNU General Public License as
- published by the Free Software Foundation; either version 2 of
- the License, or (at your option) any later version.
-
- This program is distributed in the hope that it will be
- useful, but WITHOUT ANY WARRANTY; without even the implied
- warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR
- PURPOSE. See the GNU General Public License for more details.
-
- You should have received a copy of the GNU General Public
- License along with this program; if not, write to the Free
- Software Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139,
- USA.
-
- (Addendum)
- I'd appreciate it that if you have fixes, that you send them
- to me first.
-
-
-Introduction
-============
-
-This file contains some random information, that I like to have online
-instead of in a manual that can get lost. Ever misplace your Linux
-kernel sources? And the manual of one of the boards in your computer?
-
-
-Theory of operation
-===================
-
-An important thing to know is that the driver itself doesn't have the
-firmware for the card. This means that you need the separate package
-"sx_firmware". For now you can get the source at
-
- ftp://ftp.bitwizard.nl/specialix/sx_firmware_<version>.tgz
-
-The firmware load needs a "misc" device, so you'll need to enable the
-"Support for user misc device modules" in your kernel configuration.
-The misc device needs to be called "/dev/specialix_sxctl". It needs
-misc major 10, and minor number 167 (assigned by HPA). The section
-on creating device files below also creates this device.
-
-After loading the sx.o module into your kernel, the driver will report
-the number of cards detected, but because it doesn't have any
-firmware, it will not be able to determine the number of ports. Only
-when you then run "sx_firmware" will the firmware be downloaded and
-the rest of the driver initialized. At that time the sx_firmware
-program will report the number of ports installed.
-
-In contrast with many other multi port serial cards, some of the data
-structures are only allocated when the card knows the number of ports
-that are connected. This means we won't waste memory for 120 port
-descriptor structures when you only have 8 ports. If you experience
-problems due to this, please report them: I haven't seen any.
-
-
-Interrupts
-==========
-
-A multi port serial card, would generate a horrendous amount of
-interrupts if it would interrupt the CPU for every received
-character. Even more than 10 years ago, the trick not to use
-interrupts but to poll the serial cards was invented.
-
-The SX card allow us to do this two ways. First the card limits its
-own interrupt rate to a rate that won't overwhelm the CPU. Secondly,
-we could forget about the cards interrupt completely and use the
-internal timer for this purpose.
-
-Polling the card can take up to a few percent of your CPU. Using the
-interrupts would be better if you have most of the ports idle. Using
-timer-based polling is better if your card almost always has work to
-do. You save the separate interrupt in that case.
-
-In any case, it doesn't really matter all that much.
-
-The most common problem with interrupts is that for ISA cards in a PCI
-system the BIOS has to be told to configure that interrupt as "legacy
-ISA". Otherwise the card can pull on the interrupt line all it wants
-but the CPU won't see this.
-
-If you can't get the interrupt to work, remember that polling mode is
-more efficient (provided you actually use the card intensively).
-
-
-Allowed Configurations
-======================
-
-Some configurations are disallowed. Even though at a glance they might
-seem to work, they are known to lockup the bus between the host card
-and the device concentrators. You should respect the drivers decision
-not to support certain configurations. It's there for a reason.
-
-Warning: Seriously technical stuff ahead. Executive summary: Don't use
-SX cards except configured at a 64k boundary. Skip the next paragraph.
-
-The SX cards can theoretically be placed at a 32k boundary. So for
-instance you can put an SX card at 0xc8000-0xd7fff. This is not a
-"recommended configuration". ISA cards have to tell the bus controller
-how they like their timing. Due to timing issues they have to do this
-based on which 64k window the address falls into. This means that the
-32k window below and above the SX card have to use exactly the same
-timing as the SX card. That reportedly works for other SX cards. But
-you're still left with two useless 32k windows that should not be used
-by anybody else.
-
-
-Configuring the driver
-======================
-
-PCI cards are always detected. The driver auto-probes for ISA cards at
-some sensible addresses. Please report if the auto-probe causes trouble
-in your system, or when a card isn't detected.
-
-I'm afraid I haven't implemented "kernel command line parameters" yet.
-This means that if the default doesn't work for you, you shouldn't use
-the compiled-into-the-kernel version of the driver. Use a module
-instead. If you convince me that you need this, I'll make it for
-you. Deal?
-
-I'm afraid that the module parameters are a bit clumsy. If you have a
-better idea, please tell me.
-
-You can specify several parameters:
-
- sx_poll: number of jiffies between timer-based polls.
-
- Set this to "0" to disable timer based polls.
- Initialization of cards without a working interrupt
- will fail.
-
- Set this to "1" if you want a polling driver.
- (on Intel: 100 polls per second). If you don't use
- fast baud rates, you might consider a value like "5".
- (If you don't know how to do the math, use 1).
-
- sx_slowpoll: Number of jiffies between timer-based polls.
- Set this to "100" to poll once a second.
- This should get the card out of a stall if the driver
- ever misses an interrupt. I've never seen this happen,
- and if it does, that's a bug. Tell me.
-
- sx_maxints: Number of interrupts to request from the card.
- The card normally limits interrupts to about 100 per
- second to offload the host CPU. You can increase this
- number to reduce latency on the card a little.
- Note that if you give a very high number you can overload
- your CPU as well as the CPU on the host card. This setting
- is inaccurate and not recommended for SI cards (But it
- works).
-
- sx_irqmask: The mask of allowable IRQs to use. I suggest you set
- this to 0 (disable IRQs all together) and use polling if
- the assignment of IRQs becomes problematic. This is defined
- as the sum of (1 << irq) 's that you want to allow. So
- sx_irqmask of 8 (1 << 3) specifies that only irq 3 may
- be used by the SX driver. If you want to specify to the
- driver: "Either irq 11 or 12 is ok for you to use", then
- specify (1 << 11) | (1 << 12) = 0x1800 .
-
- sx_debug: You can enable different sorts of debug traces with this.
- At "-1" all debugging traces are active. You'll get several
- times more debugging output than you'll get characters
- transmitted.
-
-
-Baud rates
-==========
-
-Theoretically new SXDCs should be capable of more than 460k
-baud. However the line drivers usually give up before that. Also the
-CPU on the card may not be able to handle 8 channels going at full
-blast at that speed. Moreover, the buffers are not large enough to
-allow operation with 100 interrupts per second. You'll have to realize
-that the card has a 256 byte buffer, so you'll have to increase the
-number of interrupts per second if you have more than 256*100 bytes
-per second to transmit. If you do any performance testing in this
-area, I'd be glad to hear from you...
-
-(Psst Linux users..... I think the Linux driver is more efficient than
-the driver for other OSes. If you can and want to benchmark them
-against each other, be my guest, and report your findings...... :-)
-
-
-Ports and devices
-=================
-
-Port 0 is the top connector on the module closest to the host
-card. Oh, the ports on the SXDCs and TAs are labelled from 1 to 8
-instead of from 0 to 7, as they are numbered by linux. I'm stubborn in
-this: I know for sure that I wouldn't be able to calculate which port
-is which anymore if I would change that....
-
-
-Devices:
-
-You should make the device files as follows:
-
-#!/bin/sh
-# (I recommend that you cut-and-paste this into a file and run that)
-cd /dev
-t=0
-mknod specialix_sxctl c 10 167
-while [ $t -lt 64 ]
- do
- echo -n "$t "
- mknod ttyX$t c 32 $t
- mknod cux$t c 33 $t
- t=`expr $t + 1`
-done
-echo ""
-rm /etc/psdevtab
-ps > /dev/null
-
-
-This creates 64 devices. If you have more, increase the constant on
-the line with "while". The devices start at 0, as is customary on
-Linux. Specialix seems to like starting the numbering at 1.
-
-If your system doesn't come with these devices pre-installed, bug your
-linux-vendor about this. They should have these devices
-"pre-installed" before the new millennium. The "ps" stuff at the end
-is to "tell" ps that the new devices exist.
-
-Officially the maximum number of cards per computer is 4. This driver
-however supports as many cards in one machine as you want. You'll run
-out of interrupts after a few, but you can switch to polled operation
-then. At about 256 ports (More than 8 cards), we run out of minor
-device numbers. Sorry. I suggest you buy a second computer.... (Or
-switch to RIO).
-
-------------------------------------------------------------------------
-
-
- Fixed bugs and restrictions:
- - Hangup processing.
- -- Done.
-
- - the write path in generic_serial (lockup / oops).
- -- Done (Ugly: not the way I want it. Copied from serial.c).
-
- - write buffer isn't flushed at close.
- -- Done. I still seem to lose a few chars at close.
- Sorry. I think that this is a firmware issue. (-> Specialix)
-
- - drain hardware before changing termios
- - Change debug on the fly.
- - ISA free irq -1. (no firmware loaded).
- - adding c8000 as a probe address. Added warning.
- - Add a RAMtest for the RAM on the card.c
- - Crash when opening a port "way" of the number of allowed ports.
- (for example opening port 60 when there are only 24 ports attached)
- - Sometimes the use-count strays a bit. After a few hours of
- testing the use count is sometimes "3". If you are not like
- me and can remember what you did to get it that way, I'd
- appreciate an Email. Possibly fixed. Tell me if anyone still
- sees this.
- - TAs don't work right if you don't connect all the modem control
- signals. SXDCs do. T225 firmware problem -> Specialix.
- (Mostly fixed now, I think. Tell me if you encounter this!)
-
- Bugs & restrictions:
-
- - Arbitrary baud rates. Requires firmware update. (-> Specialix)
-
- - Low latency (mostly firmware, -> Specialix)
-
-
-