xiaomi-sm8450/android_kernel_xiaomi_sm8450
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Documentation/serial: Delete obsolete driver documentation

Browse Source 
These serial drivers were removed in kernel v3.1, so we can drop their
documentation files and references to their magic numbers and
parameters.

There are still references to these old drivers in
Documentation/devices.txt but I'm afraid they can't be removed.

Signed-off-by: Jean Delvare <jdelvare@suse.de>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Jiri Slaby <jslaby@suse.cz>
Cc: Rob Landley <rob@landley.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This commit is contained in:
Jean Delvare
2014-03-31 14:54:25 +02:00
committed by Greg Kroah-Hartman
parent d758c9c1b3
commit 3053075cea
9 changed files with 0 additions and 866 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
Documentation/serial/00-INDEX
View File

@@ -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.

98
Documentation/serial/digiepca.txt
View File

@@ -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NOTE: This driver is obsolete.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

36
Documentation/serial/riscom8.txt
View File

@@ -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.

383
Documentation/serial/specialix.txt
View File

@@ -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?

294
Documentation/serial/sx.txt
View File

@@ -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)