Newsgroups: comp.unix.sco.announce,comp.answers,news.answers
Path: senator-bedfellow.mit.edu!bloom-beacon.mit.edu!howland.erols.net!torn!enews.sgi.com!coop.net!world!apl
From: [email protected] (Tony Lawrence)
Subject: comp.unix.sco Technical FAQ (5/7)
Message-ID: <[email protected]>
Approved: [email protected]
Date: Thu, 12 Oct 2000 23:32:26 GMT
Organization: http://www.aplawrence.com
Keywords: FAQ SCO Xenix Unix Frequently Asked Questions
Followup-To: comp.unix.sco.misc
Lines: 727
Xref: senator-bedfellow.mit.edu comp.unix.sco.announce:2481 comp.answers:42688 news.answers:193882

Archive-Name: comp.unix.sco Technical FAQ 5/7
Posting-Frequency: Monthly (mid month)
Last-modified: Oct 10



  comp.unix.sco Technical FAQ 5/7

  Questions and Answers about Serial Communications and UUCP

  FAQ Starting Page http://aplawrence.com/SCOFAQ/index.html

  These FAQS were developed and maintained for years by
  [email protected] (Stephen M. Dunn). Steve no longer has the time to
  maintain them, and has asked me to take them over. Please remember the
  debt all of us owe to Steve for his efforts- I myself spent many hours
  learning from these very documents, and I'm sure many of us can say
  similar things.

  Because Steve has not been able to maintain these for a while now,
  some of the information herein is outdated. I am working to correct
  that, but it's a lot to catch up on, so if you spot something, please
  let me know. For the moment, I'm just marking some of it as probably
  being useless; as I have time, I'll check further to be certain before
  I remove anything.

  Suggestion: Use my Search to find what you are looking for.

  My serial connections are losing characters

  Here are three possibilities. First, check the NCLIST kernel
  parameter. This governs how many CLIST structures are allocated; these
  are used to buffer input and output. If this is too low, then at times
  of high serial I/O demands, your system will run out of CLIST
  structures and start discarding characters. Note that there is a limit
  as to how many CLIST structures may be allocated to an individual
  process, regardless of how many are available systemwide. This is done
  to prevent one misbehaving program from monopolizing all of the CLIST
  structures. Look for the tunable parameter TTHOG (only available in
  newer Unix systems), which controls this limit.

  The next possibility is that you may have an old UART (8250 or 16450).
  See the next answer for more info.

  And you may also have flow control problems. The devices at either end
  of a serial cable must agree on what flow control is being used or
  else you can end up with data loss, unexplained pauses in the data
  stream, etc. See the man pages for stty and termio for more
  information on the available settings.

  [Table of Contents]
    _________________________________________________________________

  What do the terms UART, 8250, 16450 and 16550 mean?

  UART means Universal Asynchronous Receiver/Transmitter. This is a chip
  which receives and transmits data serially; each serial port you have
  will use one, though it is possible that several may be integrated
  into one chip.

  8250, 16450 and 16550 are all common types of UARTs. The 8250 is an
  old chip which cannot run at high speed. The 16450 is similar to the
  8250 except that it supports data communications at higher speeds.
  Both of these chips generate an interrupt for every character that is
  sent or received, which basically tells the CPU either "Here is some
  data for you" or "Feed me!" This is all very well, except that at high
  speed, the number of interrupts (nearly 4000 per port per second at 38
  400 bps) can overwhelm a CPU, bringing system performance way down.
  Also, if the CPU is busy servicing another interrupt at the time, the
  serial port's interrupt may not be serviced in time, which will cause
  a character to be lost.

  The 16550 is pin-compatible with the 16450 and, by default, runs in
  16450 mode. This makes it compatible with software which is not
  16550-aware. If your software is 16550-aware, it can turn on a special
  mode in which the 16550 buffers all data with 16-byte internal
  buffers. This not only allows the CPU to deal with far more bytes at a
  time, increasing efficiency, but also means that if the CPU can't
  service the interrupt before the next character comes in, there's
  still space in the buffer for it.

  16550 support was introduced in Xenix 2.3.4, ODT 1.1 and Unix 3.2.2.
  If you have these, or later, versions, your operating system will
  automatically detect 16550-equipped ports and will enable their
  buffering. A third-party serial driver called FAS includes 16550
  awareness in its feature set; you may wish to investigate this as
  well. FAS can be found at ftp://ftp.fu-berlin.de/pub/unix/driver/fas/.

  Note that the above is not really applicable to intelligent multiport
  serial cards. While these cards may well use 16550s, it is the
  processor on the serial card which is responsible for dealing with the
  serial ports it controls, and the main CPU has nothing to do with the
  UARTs.

  [Table of Contents]
    _________________________________________________________________

  How do I adjust my 16550's trigger level?

  The answer is different for Unix and Xenix, but much of the
  information is the same, so it's been grouped together here. We will
  deal with the common information first, and the specific details for
  Xenix and Unix.

  By default, Xenix and Unix set the 16550's trigger level to 14. This
  means that once fourteen characters have been received, the 16550 will
  generate an interrupt (it will also generate an interrupt if the
  buffer is not full but serial data flow has stopped, so the system
  doesn't always have to wait until the trigger level is reached). This
  give the system two character times in which to begin to clear the
  buffer; at high speed on a highly loaded system, this may not be
  enough, and you may still lose characters even though you have a
  16550. On the other hand, this value should generally be set as high
  as possible to reduce the number of interrupts generated; servicing an
  interrupt is quite costly in terms of CPU time.

  There is an array in the kernel called sio_fifoctl[]. It is a 16-byte
  array with control values for different minor numbers. To find which
  array element will be used for a particular serial port, AND the minor
  number of the port with 0x0F (for example, /dev/tty2A has a minor
  number of 136 and /dev/tty2a of 8; either one ANDed with 0x0F yields
  8, so sio_fifoctl[8] controls this port).

  There are four different values you may wish to use for the entries in
  sio_fifoctl[]. A value of 0x0F sets the trigger value to 1; 0x4F sets
  it to 4; 0x8F sets it to 8; 0xCF sets it to 14 (these values are
  determined by the 16550 itself, not by SCO, and other values will not
  set the trigger level to intermediate values).

  In Xenix, this parameter is set by patching the disk image of the
  kernel (/xenix) using adb (the info on how to find adb is elsewhere in
  this FAQ). The following is a sample adb session to change the trigger
  level of /dev/tty2A to 8 from 14 (the line numbers in parentheses are
  for the explanation below); the asterisks are adb's prompt and should
  not be typed in:

   1. cp /xenix /xenix.save
   2. adb -w /xenix -
   3. * sio_fifoctl+8/x
   4. sio_fifoctl+0x8: 0xcfcf
   5. * sio_fifoctl+0x8/w 0xcf8f
   6. sio_fifoctl+0x8: 0xcfcf= 0xcf8f
   7. * $q

  Line 1 makes a backup, and line 2 runs adb in write mode. Line 3 tells
  adb to print the current value of sio_fifoctl[8]. Line 4 is adb's
  reply, which includes two bytes from this array (the rightmost one is
  the value for sio_fifoctl[8], and the leftmost is for sio_fifoctl[9]).
  You must look at these carefully, as one half will have to be changed
  while the other will have to be left alone. In line 5, we write 0xCF8F
  into this location; note that the value for sio_fifoctl[9] is left
  unchanged at 0xCF. Line 6 is adb's reply giving the old and new
  values. Line 7 quits adb.

  For Unix, there is a table at the end of the text file
  /etc/conf/pack.d/sio/space.c which gives the same array. It is
  formatted in the same manner (to find the appropriate value, AND the
  minor device number with 0x0F).

  If you run Unix, check the man page for the sar command to see if you
  have the -g option to check for serial I/O overruns. If so, try
  running it. If you see overruns, this indicates that your trigger
  level is set too high and the system doesn't have adequate time to
  service the 16550. The cure is to turn the trigger level down one
  notch and try again.

  The information for Xenix in this answer is taken from the
  comp.unix.xenix.sco FAQ, maintained by Chip Rosenthal. The copyright
  for that document reads:

  This collection is Copyright 1992-1994, Unicom Systems Development,
  Inc. All rights reserved. Permission granted to reproduce and
  distribute this document provided this notice remains intact and any
  changes to the document are clearly marked. We have tried to review
  all information, but cannot guarantee it for any particular purpose.
  We do not offer any warranties or representations, nor do we accept
  any liability for any damage resulting from the use or misuse of
  information or procedures in this document.
  [Table of Contents]
    _________________________________________________________________

  I can transfer small files via UUCP but large files won't go

  I can transfer small files via UUCP but large ones go really slowly

  This may indicate a flow control problem. uucico, the program that
  actually performs UUCP transfers, turns off all hardware and software
  flow control on the port it is using (prior to 3.2v4.0, which allows
  you to specify what it should use). If your modem's buffers are too
  small, then the stream of data involved in a large file transfer may
  overflow them, causing transmission errors. This may just cause your
  throughput to go way down, or it may result in a total inability to
  transfer larger files. Run uucico with debugging (you may find
  /usr/lib/uucp/uutry to be useful) and watch for "alarm" messages. If
  you see these, it's an indication that some characters are likely
  being dropped during transmission.

  If you are using a serial port on a multiport serial card such as one
  from Equinox or Digi, your board may have shipped with a utility that
  lets you permanently set flow control on your port. Another
  alternative is to reduce uucico's window size.

  Under 3.2v4.0, uucico does not turn off flow control; it is left at
  whatever setting the dialer used. If you are using a compiled dialer
  and have the source and the development system, you can write in
  whatever stty settings you desire. Under 3.2v4.1, if you are using
  atdialer, you can specify stty settings in /etc/default/atdial*. Note
  that it has been reported that cu, even under 3.2v4.x, turns on
  XON/XOFF flow control and, in doing so, disables hardware flow
  control.

  In 3.2v4.x, a new stty setting has been added to perform bidirectional
  RTS/CTS flow control. This is CRTSFL. CTSFLOW and RTSFLOW do not
  perform proper bidirectional flow control; they allow the modem to
  signal the computer to stop sending, but not vice versa. Depending on
  what modem you have, you may need CRTSFL if your system can't accept
  characters as quickly as your modem can produce them.

  [Table of Contents]
    _________________________________________________________________

  How do I get better UUCP throughput?

  You may or may not be able to. Here are some rules of thumb which may
  or may apply to your situation:

  If you and the remote site both use Telebit modems, enable UUCP
  spoofing. This enables the modems to maintain the UUCP protocol
  between them, and means that your CPU's response time to individual
  UUCP packets is not as critical. Some sites can achieve higher
  throughput with a Telebit link than with a direct connection at the
  same baud rate.

  If you and the remote site are both running versions of UUCP which
  allow you to specify different protocols (SCO Unix 3.2v4.0 and above,
  for example), look through your release notes to find the
  highest-performing protocol that's applicable to your situation. For
  example, the standard g protocol has its own error detection and
  correction logic. If you are running over a guaranteed error-free
  link, this is unnecessary overhead, and eliminating it by selecting a
  protocol that relies on an error-free link can speed up transfers. Be
  careful that you don't specify such a protocol for a non-error-free
  link, or else your transmissions will occasionally be corrupted. Note
  that just because you are using error-correcting modems does not
  guarantee that your data will be error-free; there is a possibility of
  lost characters and flow control problems in the serial ports at both
  ends of the connection. While the link between the modems may well be
  error-free, the end-to-end link from one uucico to another may not.
  You should usually use g or G for modem links, and reserve t, e, f,
  etc. for network connections.

  If you are running over a link that uses MNP, V.42 or V.42bis, set
  your window size as high as possible to ensure that the modems have as
  much data to work with as possible. For further tips on high-speed
  modems, see the section on maximizing serial throughput. For
  information on changing uucico's window size, see below.

  If the site with which you are communicating sometimes drops packets,
  try adjusting your packet delay to a lower number. If you are
  operating over a link with high latency (i.e. it takes a long time for
  a packet to get from one site to the other - an example would be a
  satellite link), you may need to increase your packet delay. For more
  details, see below.

  See the note below on how to get the UUCP Internals FAQ.

  [Table of Contents]
    _________________________________________________________________

  How do I change uucico's window size?

  For OpenServer Release 5, see the man page for Configuration. For
  earlier versions, read on.

  You need to have adb or _fst to do this. See the information on
  getting these, which can be found in section 1. First, make a backup
  copy of /usr/lib/uucp/uucico! This is very important. Also, note its
  ownership and permissions. To change the window size to 7, run the
  following command from the Bourne shell:


  adb -w /usr/lib/uucp/uucico - << EOF
  $d
  _windows/w 7
  $q
  EOF

  When you have done this, reset the ownership and permissions to what
  they were originally; this process will change them on you.

  Note that specifying a window size greater than seven may break
  uucico, as the g protocol is designed for a window size of no more
  than seven. Also, the documentation and/or scopatches that ship with
  some SCO products list the second line of the above as "%d". This, at
  least on the copy of adb I have, will fail; $d is correct. If you are
  using Unix 3.2v4, leave the underscore off the symbol name.

  On Xenix 2.3.4, you may be tempted to use the scopatch command to make
  this change. Don't bother; there are at least three bugs in the short
  script which does this work, and it's just as easy to do it manually
  as it is to fix the script.

  [Table of Contents]
    _________________________________________________________________

  I increased my window size but nothing changed

  When UUCP is negotiating a connection, each side will tell the other
  what window size to use when sending. Therefore, if your window size
  is 7 and the remote uses 3, you will be sending files using a window
  size of 3, but the other side may send with a window size of up to 7.
  Note that the UUCP on the other side may not support the window size
  you specify, and may send with a smaller window than you requested.
  While this should not cause problems, it may provide lower performance
  than you'd like.

  If your connection does not have data compression, error correction,
  or long transmission delays, and the two sites involved have
  sufficient CPU power to respond quickly to incoming packets, a window
  size of 2 or 3 should be sufficient to achieve streaming data flow. In
  this case, a larger window size will not provide any benefit.

  If you have a Telebit modem, see the Telebit notes below.

  [Table of Contents]
    _________________________________________________________________

  I increased my window size and UUCP broke!

  If your modem, or the modem on a remote site, has small data buffers,
  a large window size may cause the modem's buffers to overrun. If the
  site that broke polls your system, you may wish to create a separate
  copy of uucico with a smaller window size, and specify it as the login
  shell in /etc/passwd. If you are the polling site, you may have to
  restore your original uucico. Alternatively, you could try a window
  size of 4, then 5, then 6, then 7, and see at what point UUCP stops
  working. You will then know the maximum window size you can use. You
  may also wish to ask the sysadmin at the remote site if there is
  anything they can do, such as forcing flow control to be used on their
  modem, that may remedy the problem. You may, however, have no choice
  but to return to using a window size of 3. If you have a Telebit
  modem, see the notes below relating to Telebit modems.

  [Table of Contents]
    _________________________________________________________________

  UUCP frequently has to resend packets

  If you are using a modem with error correction and/or data
  compression, or if you are making a long-distance connection, there
  are delays inherent in your connection which may cause UUCP to timeout
  and resend packets. You may need to increase the packet timeout.

  See also the section above on problems in which UUCP can transfer
  small files but not larger ones.

  [Table of Contents]
    _________________________________________________________________

  How do I change uucico's packet timeout?

  For OpenServer Release 5, see the man page for Configuration. For
  earlier versions, read on.

  This is very similar to changing the window size; once again, you will
  need to have a copy of adb, and you must first make a backup copy of
  uucico. Then run the following Bourne shell command:


  adb -w /usr/lib/uucp/uucico - << EOF
  $d
  _pktime/w 5
  $q
  EOF

  Replace 5 with the desired number of seconds. Note that you should set
  this parameter to the smallest value that works reliably. If you
  specify too large a value, it will reduce throughput if you encounter
  a bad line or other conditions that require packets to be resent. If
  you are running Unix 3.2v4, leave the underscore out of the symbol.
  Once again, restore the permissions and ownership once you're done.

  [Table of Contents]
    _________________________________________________________________

  What special considerations are there regarding my Telebit modem?

  If you are using Telebit UUCP spoofing (S111=30), the modem spoofing
  firmware will only negotiate a 3-window connection, so changing the
  window size to 7 will not do anything for spoofed connections. It
  will, however, be effective for non-spoofed connections (i.e. if you
  connect to a site which is not using a Telebit modem).

  Some people have reported problems with Telebit modems and UUCP-g
  spoofing when they change their window and/or packet sizes.
  Apparently, Telebits can handle requests for increased window sizes
  without error (though they will only use a maximum window size of 3),
  but will not work correctly with large packets (probably above the SCO
  default of 64 bytes).

  There is a bug in version LA 5.00W of the Telebit Worldblazer firmware
  that causes uucico to fail in startup when the answering uucico is set
  for windows=7.

  [Table of Contents]
    _________________________________________________________________

  What are all the V.something codes, MNP, HST, etc.?

  Here are the most common data modem signalling standards. Note that
  the descriptions are not complete technical descriptions, and only
  cover the major feature of the standard (i.e. no discussion is made of
  fallback data rates etc.)

    * Bell 103 - North American 0-300 bps
    * Bell 212A - North American 1200 bps
    * V.22 - International 1200 bps; not generally used in North America
    * V.22bis - 2400 bps
    * V.32 - 4800 & 9600 bps
    * V.32bis - 4800 - 14 400 bps
    * V.32terbo - Vendor standard (no formal recognition), 16.8 & 19.2
      kbps
    * V.FC - Vendor standard (no formal recognition), up to 28.8 kbps
    * V.34 - International standard, up to 28.8 kbps; this has been
      extended to 33.6 kbps
    * HST - USRobotics' proprietary High-Speed Transfer, 9600 - 16 800
      bps
    * PEP - Telebit's proprietary standard; also Turbo PEP
    * X2 - USRobotics' asymmetrical high-speed modulation; up to 56 kbps
      downstream (limited to 53.3 kbps by legislation), V.34 upstream.
      See http://x2.usr.com/
    * K56flex - same idea as X2, different implementation

  There are also numerous standard for error correction and data
  transmission. MNP (Microcom Networking Protocol) is a family of
  protocols with various levels. MNP levels 1 through 4 denote error
  correction schemes of increasing sophistication. MNP level 4 is the
  most common MNP error correction level; while there are higher levels,
  they are not terribly widespread. You may never see your modem report
  an MNP level 4 connection if you have data compression enabled; it
  will report level 5 instead. MNP level 5 is usually considered a 2:1
  compressor, meaning that it will generally compress your data by up to
  a factor of 2 (though it can exceed this on some data and not reach it
  on other data). Note that it does not check to ensure that it can
  actually compress the data; if you send precompressed data through it,
  it will actually increase the amount of data which must be transferred
  between modems, thereby decreasing throughput.

  V.42 is another error correction standard; unlike MNP, it is
  non-proprietary. Its primary error-control protocol is called LAP-M.
  It also includes a provision for falling back to MNP level 4 error
  correction should the remote modem not support V.42. V.42bis is the
  corresponding data compression standard. It is both more efficient
  than MNP level 5, providing up to a 4:1 rate (even higher on
  particularly repetitive data), and more intelligent, in that it can
  recognize whether or not it can compress data and send the data
  uncompressed if this makes the most sense.

  For further information on modems, see the newsgroup comp.dcom.modems.

  [Table of Contents]
    _________________________________________________________________

  How do I maximize serial throughput?

  This answer assumes that you have an error-correcting and/or
  data-compressing modem. The rule of thumb here is to feed data to the
  modem as fast as you can. The more data it has to work with, the more
  efficiently it can compress it. Not only that, but there is a
  more-or-less fixed overhead involved in bundling data into packets,
  which is how the modems transmit it, so when error correction and/or
  data compression is in use, you can reduce this overhead to a minimum
  by ensuring that the modem has as much data as possible to put into
  each packet. The usual rule of thumb is to feed the modem four times
  faster than the fastest connection rate it supports. If you have a
  V.32 modem, which supports connections of up to 9600 bps, you should
  communicate with it at 38 400 bps (assuming your hardware supports
  this rate). If you are using error correction but not data
  compression, the next speed up (19 200 bps, in this case) is generally
  sufficient.

  The modem will have an internal data buffer of some size; it could be
  as large as a couple of kilobytes on better modems. If you are sending
  it data faster than it can transmit it (which you should be), this
  buffer will fill up over time. You will need to have handshaking
  between the modem and the computer so that the modem can signal to the
  computer to stop sending when the buffer is nearly full, and to start
  sending again once there is room in the buffer. This can be done in
  software (usually using the ASCII XON and XOFF characters), or in
  hardware (usually using the RS-232 CTS and RTS lines). The use of
  software handshaking requires less wires between the computer and the
  modem, but will interfere with the transmission of binary data. The
  use of hardware handshaking requires additional wiring, but will not
  interfere with binary data. There are advocates of both methods; I
  personally prefer hardware handshaking. Whichever method you choose,
  you must make sure that you configure both the modem and the computer
  (via stty settings) to use the same protocol.

  [Table of Contents]
    _________________________________________________________________

  My data-compressing modem doesn't work much faster than my old modem

  There are a couple of possible reasons for this. Firstly, you may be
  transferring data that has already been compressed and is therefore
  not compressible (compressed news batches, for example). If your modem
  connects to the remote site using V.42bis data compression, this will
  not adversely affect throughput. If, however, you are using an MNP
  level 5 connection, you will actually lose throughput when sending
  compressed data. If the majority of your transfers are already
  compressed, you should disable MNP level 5 data compression. Note that
  using MNP level 3 or 4 error correction, or V.42 error correction,
  without any compression will increase throughput regardless of the
  type of data being transmitted.

  Also, due to the way data compression and error correction in modems
  work, there are delays involved while the two modems agree on whether
  or not each packet of data was received correctly. If possible, you
  and the remote site should try to increase uucico's window size; this
  will cause more data to be transmitted at once, which will not only
  improve the efficiency of your modem's data compression, but will also
  often mean that the response to a previous packet arrives before the
  window has expired. I'm afraid a complete discussion of packetization
  and sliding window protocols is beyond the scope of this FAQ; just
  trust me on this one if you don't quite follow me. Further discussion
  can sometimes be found in the newsgroup comp.mail.uucp.

  [Table of Contents]
    _________________________________________________________________

  A BSD-based machine can't connect to mine via UUCP

  BSD-based UUCP sends data with even parity; SCO expects no parity. Add
  "" P_ZERO to the chat script in the BSD machine's Systems or L.sys
  file. For example:

  xnxbox Any uucp 19200 5553333 "" P_ZERO "" \r in:--in: nuucp

  [Table of Contents]
    _________________________________________________________________

  Where can I get more information on UUCP protocols?

  There is a FAQ posted to the news group comp.mail.uucp which goes into
  a good deal of detail on the format of UUCP files, the handshake
  process at the start of a connection, what some of the debugging
  output from uucico -x? means, and the internals of numerous UUCP
  protocols including the common g protocol and several others.

  If you do not get this group, you can find the FAQ at
  ftp://rtfm.mit.edu/pub/usenet-by-hierarchy/comp/mail/uucp/UUCP_Interna
  ls_Frequently_Asked_Questions. If you do not have anonymous ftp
  access, send email to [email protected]; include the words
  "help" and "index" in separate lines in the body of your message.

  [Table of Contents]
    _________________________________________________________________

  I edited my gettydefs and things broke

  You can check gettydefs using /etc/getty -c /etc/gettydefs; this
  command will print out everything you could possibly want to know, and
  then some. Common mistakes include not leaving a blank line between
  entries or having a cycle of settings but forgetting to close the loop
  (e.g. if 1 fails, use 2; if 2 fails, use 3; if 3 fails, you forgot to
  tell it to go back to 1).

  [Table of Contents]
    _________________________________________________________________

  I can't get shared dial-in and dial-out working

  Here are a few things to keep in mind. Under Unix, the entries in
  /etc/inittab and /usr/lib/uucp/Devices must match exactly; if one says
  "ttyi3" while the other says "/dev/ttyi3", getty won't consider them
  to be the same port, and your modem will not be sent the undialer
  string.

  Serial ports to be used by UUCP (or even just initialized with an
  undialer) should generally be owned by user uucp and group uucp. If
  you set it this way but it keeps being put back to something else,
  chances are you're suffering from the problem mentioned above.

  If you have several entries in /usr/lib/uucp/Devices for a port
  (specifying different dialers for different speeds or whatever), only
  the first one will actually be used to reset the port. Make sure the
  first one listed is the one you want - that's usually marked ACU, but
  local customs may vary.

  [Table of Contents]
    _________________________________________________________________

  What are some common settings for my modem?

  For the real answers, read your modem's manual. If you want to discuss
  modems, see comp.dcom.modems. Having said that, here are some common
  setup strings for various features on many modems; see your manual to
  determine which one(s) apply to your modem. Note that not all modems
  will accept the following commands; even the manual for one modem is
  too large to fit, let alone all possible modems.

    * Locked DTE rate - this causes your modem to always talk to your
      serial port at the same rate, regardless of the data transfer rate
      between modems. Try AT\J0, AT&Qn (n is a number which depends on
      other settings as well), (Multitech) AT$SB38400 or (USRobotics)
      AT&B1.
    * Hardware flow control - this uses out-of-band signalling on the
      RTS and CTS lines to perform flow control and will not interfere
      with the transmission of binary data, as software flow control
      will. Try AT&K3 or AT&E4.
    * Software flow control - particularly on interactive (cu)
      connections, you may wish to use software flow control in some
      cases. This is usually done with the same command as hardware flow
      control, but with a different number (e.g. AT&K4). One
      disadvantage is that if you get garbage characters (due to line
      noise, modem disconnection, etc.), you will sometimes get a flow
      control character as part of the garbage; the system doesn't know
      it's garbage and so it may appear to hang the port.
    * Correct use of carrier detect - the DCD line is used to indicate
      to the computer that the modem has established carrier. This
      setting is almost universally done with AT&C1.
    * Error control - this feature, if negotiated between modems, causes
      the modems to take care of retransmitting data when errors crop up
      on the phone line. Try AT&Qn (again, n depends on other settings),
      AT\Nn (ditto), or AT&E1. See the discussion elsewhere in this FAQ
      to help determine what, if any, error control you should use.
      Generally, though not always, you will want to enable all error
      control methods your modem supports, in order to be able to
      negotiate connections with other modems, but may need adjustment
      in some circumstances.
    * Data compression - this feature, if negotiated between modems,
      causes the modems to attempt to compress the data before sending,
      improving throughput. Try AT&Qn (n depends on other settings) and
      AT%C1; some modems also use registers such as S36 to fine-tune
      this. See the discussion elsewhere in this FAQ for more
      information on when you might or might not want to use data
      compression.
    * Correct use of DTR - the DTR line, when dropped by the computer,
      should cause the modem to hang up (and, according to some
      opinions, reset entirely). This feature is almost universally done
      with AT&Dn, but the supported values of n may vary from modem to
      modem. Try AT&D2 to hang up and AT&D3 to hang up and reset.
    * Writing settings - Two cautions first. First, you do NOT want to
      do this every time you initialize the modem. The EEPROMs used in
      most modems to store this information have a lifetime of something
      like 10 000 writes. At 30 reinitializations a day, you'll burn out
      your EEPROM in around a year. Second, if you use this to do the
      initial setup of the modem, document your settings somewhere so
      that if your modem does die, you'll have the settings you used
      somewhere (I'd recommend in a comment in your
      /usr/lib/uucp/Dialers file). See AT&W; some modems have several
      settings available, written to with AT&W0, AT&W1, etc. See AT&Y on
      some modems to determine which settings are used upon reset, and
      ATZn for resetting the modem.
    * To Echo or Not To Echo - you can configure your modem to echo
      commands back, or not to echo commands back, with ATE1 and ATE0,
      respectively. You can control the sending of result messages with
      ATQ1 and ATQ0; some modems also use ATQ2 to disable result codes
      on incoming connections.
    * Enabling busy detection - ATXn controls both the range of result
      codes the modem sends (such as enabling the display of connect
      speed rather than simply CONNECT) and what phone company signals
      the modem detects (e.g. busy, dialtone). Most people use ATX4;
      check your manual for other options.
    * Auto-answer - use ATS0=n, where n is the number of rings to allow
      to pass before answering. ATS0=0 disables auto-answer. You may
      wish to disable auto-answer while calling out or at other times
      when you don't wish the modem to answer.
    * Escape code - most modems use a pause, followed by three plus
      signs, followed by a pause, as a signal from the computer that the
      modem should go into command mode while on-line. This may cause
      problems in some cases. To disable this escape sequence, use
      ATS2=128.

  [Table of Contents]
    _________________________________________________________________

  Will a Winmodem work?

  No. It requires software drivers that aren't available (though SOME
  Winmodems can be made to work on Linux systems with special modules).

  Get a real modem. Preferably, get an external modem, because:
    * External modems are only a few dollars more than internal.
    * Internal modems take up a slot that you might need someday for
      something else.
    * Modems can get so confused that only shutting them off will clear
      their problem. You don't want to shut off your server when that
      happens, do you?
    * If lightning comes down your phone line (it happens) and you have
      an internal modem, your whole computer is likely to be fried. With
      an external modem, you'll lose the modem, but probably not
      anything else.
    * You can see what's going on with an external modem. You can see
      when the modem is receiving, transmitting, handshaking- much
      easier to diagnose problems.
    * You can share an external modem between multiple computers either
      manually or by an A/B box- again this facilitates testing and is
      sometimes very convenient.
    * When you upgrade your computer, swapping the modem is much easier.
    * When you finally get your cable modem, DSL line or T1 :-), you
      might actually be able to sell that external modem, but used
      internal cards are near worthless.

  [Table of Contents]
    _________________________________________________________________

  Why can't I get my dialin modem to work above 9600 baud?

  Because you need the modem speed in /etc/inittab to match the DTE rate
  that the modem is set for. If the DTE rate is fixed, the modem can
  connect with another modem at any speed. See
  http://aplawrence.com/Unixart/hispeed.htmli for a full description of
  how to do this.

  [Table of Contents]
    _________________________________________________________________

  I've set the modem for 38400 in inittab, but stty shows it at some other speed

  You have a line in /usr/lib/uucp/Devices that references that modem
  line at some other speed, and the getty program is looking at that.
  See http://aplawrence.com/Unixart/hispeed.html for a more detailed
  explanation.

  [Table of Contents]
    _________________________________________________________________