picocom

Minimal dumb-terminal emulator

by Nick Patavalis ([email protected])

The latest release can be downloaded from:

https://github.com/npat-efault/picocom/releases

As its name suggests, picocom is a minimal dumb-terminal emulation program. It is, in principle, very much like minicom, only it's "pico" instead of "mini"!

It was designed to serve as a simple, manual, modem configuration, testing, and debugging tool. It has also served (quite well) as a low-tech serial communications program to allow access to all types of devices that provide serial consoles. It could also prove useful in many other similar tasks. It can be used in embedded systems, since its memory footprint is rather small (approximately 40K, when stripped and minimally configured).

Picocom runs and is primarily tested on Linux. With no, or with minor, modifications it will run (and most of its features will work) on any Unix-like system with a reasonably POSIX-compatible termios(3) interface. Patches to support idiosyncrasies of specific Unix-like operating systems are very welcome.

For a description of picocom's operation, its command line options, and usage examples, see the manual page included in the source distribution as "picocom.1", and also html-ized as "picocom.1.html".

People who have contributed to picocom, by offering feature implementations, bug-fixes, corrections, and suggestions are listed in the "CONTRIBUTORS" file.

Please feel free to send comments, requests for new features (no promises, though!), bug-fixes and rants, to the author's email address shown at the top of this file.

Compilation / Installation

Change into picocom's source directory and say:

make

This will be enough to compile picocom for most modern Unix-like systems. If you want, you can then strip the resulting binary like this:

strip picocom

Striping the binary is not required, it just reduces its size by a few kilobytes. Then you can copy the picocom binary, as well as the man-page, to wherever you put your binaries and man-pages. For example:

cp picocom ~/bin
cp picocom.1 ~/man/man1

Again, this is not strictly necessary. You can run picocom and read its man-page directly from the source directory.

If something goes wrong and picocom can't compile cleanly, or if it's lacking a feature you need, take a look at the included Makefile. It's very simple and easy to understand. It allows you to select compile-time options and enable or disable some compile-time features by commenting in or out the respective lines. Once you edit the Makefile, to recompile say:

make clean
make

If your system's default make(1) command is not GNU Make (or compatible enough), find out how you can run GNU Make on your system. For example:

gmake clean
gmake

Alternatively, you might have to make some trivial edits to the Makefile for it to work with your system's make(1) command.

Using picocom

If your computer is a PC and has the standard on-board RS-233 ports (usually accessible as two male DB9 connectors at the back) then under Linux these are accessed through device nodes most likely named: /dev/ttyS0 and /dev/ttyS1. If your computer has no on-board serial ports, then you will need a USB-to-Serial adapter (or something similar). Once inserted to a USB port and recognized by Linux, a device node is created for each serial port accessed through the adapter(s). These nodes are most likely named /dev/ttyUSB0, /dev/ttyUSB1, and so on. For other systems and other Unix-like OSes you will have to consult their documentation as to how the serial port device nodes are named. Lets assume your serial port is accessed through a device node named /dev/ttyS0.

You can start picocom with its default option values (default serial port settings) like this:

picocom /dev/ttyS0

If you have not installed the picocom binary to a suitable place, then you can run it directly from the source distribution directory like this:

./picocom /dev/ttyS0

If this fails with a message like:

FATAL: cannot open /dev/ttyS0: Permission denied

This means that you do not have permissions to access the serial port's device node. To overcome this you can run picocom as root:

sudo picocom /dev/ttyS0

Alternatively, and preferably, you can add yourself to the user-group that your system has for allowing access to serial ports. For most Unix-like systems this group is called "dialout". Consult you system's documentation to find out how you can do this (as it differs form system to system). On most Linux systems you can do it like this:

sudo usermod -a -G dialout username

You will need to log-out and then log-in back again for this change to take effect.

You can explicitly set one or more of the serial port settings to the desired values using picocom's command line options. For example, to set the baud-rate to 115200bps (the default is 9600bps), and enable hardware flow-control (RTS/CTS handshake) you can say:

picocom -b 115200 -f h /dev/ttyS0

or:

picocom --baud 115200 --flow h /dev/ttyS0

To see all available options run picocom like this:

picocom --help

Once picocom starts, it initializes the serial port and prints the message:

Terminal is ready

From now on, every character you type is sent to the serial port, and every character received from the serial port is sent ro your terminal. Including control and special characters. Assuming that there is nothing connected to the other end of your serial port, to respond to the characters you send to it (e.g. echo them back to you), then nothing that you type in picocom will appear on your terminal. This is normal.

To exit picocom you have to type:

C-a, C-x

Which means you have to type [Control-A] followed by [Control-X]. You can do this by pressing and holding down the [Control] key, then pressing (and releasing) the [A] key and then pressing (and releasing) the [X] key (while you still keep [Control] held down).

This C-a is called the "escape character". It is used to inform picocom that the next character typed is to be interpreted as a command to picocom itself (in this case the exit command) and not to be sent-down to the serial port. There are several other commands (other than C-a, C-x), all prefixed by C-a.

Next you should take a look at the very detailed picocom manual page. It can be accessed like this (assuming you are inside the picocom distribution source directory):

man ./picocom.1

or (assuming you have installed the manual page to a suitable place):

man picocom

Thanks for using picocom

Custom Bash completion

Starting with release 3.2, picocom includes support for custom Bash-shell completion. With this you can press the [TAB] key, and the bash shell will complete command-line option names and values and propose valid selections for both. This makes the experience of using picocom more pleasant.

Custom completion works only with recent versions of the Bash shell (>= 4.3). It is in no way mandatory in order to use picocom. Here's how you can enable it, if you wish.

To manually enable custom completion support you need to source the file (custom completion script):

<picocom source dir>/bash_completion/picocom

Assuming you are inside the picocom source directory, you can do it like this:

. ./bash_completion/picocom

This will enable custom completion support for the current shell session only. Use it for a while and see if you like it.

To enable support automatically for all Bash-shell sessions, you have the following options:

1. If you are running a relatively modern Debian or Ubuntu or other Debian-based distribution, and you have package bash-completion installed, you can simply copy the custom completion script to the directory:

   /etc/bash_completion.d/
   

   Obviously, you need to be root to do this. Assuming you are inside the picocom source directory, something like this will do it:

   sudo cp ./bash_completion/picocom /etc/bash_completion.d/
   

   This will enable custom completion support for picocom, globaly (for all Bash-shell users).

   NOTICE: If you have another version of picocom already installed, there may already be a picocom completion script in /etc/bash_completion.d. The command above will obviously overwrite it with the new one. So be careful if this is not what you want.

   For other distributions and operating systems you have to check their documentation to see if they provide a similar mechanism for automatically sourcing custom completion scripts.

2. If you want to automatically enable support only for the current user, you must arange for your user's .bashrc to source the custom completion script. There are, obviously, many ways to do this, so the following is only a suggestion:

   Create a directory to keep the custom completion scripts

   mkdir ~/.bash_completion.d
   

   Copy the picocom completion script to the directory you created. Assuming you are inside the picocom source directory:

   cp ./bash_completion/picocom ~/.bash_completion.d
   

   Add the following (or similar) to the end of your .bashrc

   # Source custom bash completions
   if [ -d "$HOME"/.bash_completion.d ]; then
       for c in "$HOME"/.bash_completion.d/*; do
           [ -r "$c" ] && . "$c"
       done
   fi
   

   From now on every new shell session you start will load (source) all the custom completion scripts you have put in ~/.bash_completion.d

A low-tech terminal server

You can use picocom to patch-together a very simple, very low-tech, terminal server.

The situation is like this: You have, in your lab, a box with several serial ports on it, where you connect the console ports of embedded devices, development boards, etc. Let's call it "termbox". You want to access these console ports remotely.

If you provide shell-access to termbox for your users, then it's as simple as having the users say (from their remote workstations):

$ ssh -t user@termbox picocom -b 115200 /dev/ttyS0

Or make a convenient script/alias for this. Remember the -t switch which instructs ssh to create a pseudo-tty, otherwise picocom won't work.

What if you don't want to give users shell-access to termbox? Then you can use picocom in a setup like the one described below. Just remember, there are countless variations to this theme, the one below is just one of them. Also, keep in mind that some of the commands shown may have small differences from system to system; more so if you go from Linux to other Unix-like systems.

Login to termbox and create a user called termbox:

$ sudo useradd -r -m termbox

The -r means "system account", and the -m means do make the home-directory. Mostly we need this account's home-directory as a convenient place to keep stuff; so it doesn't need a login shell or a password.

Switch to the termbox account and create a bin directory in its home-dir.

$ sudo su termbox
$ cd ~
$ mkdir bin

Copy the picocom binary in ~termbox/bin (if you don't have it globally installed):

$ cp /path/to/picocom ./bin

For every serial port you want to provide access to, create a file named after the port and put it in ~termbox/bin. It should look like this:

$ cat ./bin/ttyS0
#!/bin/sh
exec /home/termbox/bin/picocom \
  --send-cmd '' \
  --receive-cmd '' \
  -b 115200 \
  /dev/ttyS0

And make it executable:

$ chmod +x ./bin/ttyS0

Repeat accordingly for every other port. Now the contents of ~termbox/bin should look like this:

$ ls -l ./bin
-rwxrwxr-x 1 termbox termbox 102128 Aug 29 13:56 picocom*
-rwxrwxr-x 1 termbox termbox    108 Aug 29 14:07 ttyS0*
-rwxrwxr-x 1 termbox termbox    108 Aug 29 14:07 ttyS1*
... and so on ...

Exit the termbox account:

$ exit

Now, for every serial port, create a user account named after the port, like this:

$ sudo useradd -r -g dialout -d ~termbox -M -s ~termbox/bin/ttyS0 ttyS0

Observe that we make dialout the default group for this account, so the account has access to the serial ports. Also observe that we make the script we just wrote (~termbox/bin/ttyS0) the login-shell for the account. The -d option instructs useradd to use /home/termbox as the user's home directory, and the -M switch instructs it not to create the home-directory. We don't really need a home directory for the ttyS0 account, since picocom will not read or write any files; but we provide one, regardless, because some systems need a valid home-directory to cd-into on login (else they choke). We could as well have used / as the home directory, or we could have let useradd create the usual /home/ttyS0.

Then set a password for the newly created account:

$ sudo passwd ttyS0
Enter new UNIX password: ******
Retype new UNIX password: ******

Repeat (create user account, set password) for every port you want to give access to.

You 're set. All a user has to do to remotely access the console connected to termbox's /dev/ttyS0 port, is:

ssh ttyS0@termbox

Some interesting points:

• If the default port settings you specified as command-line arguments to picocom in ~termbox/bin/ttySx do not match the settings of the device connected to the port, the user can easily change them from within picocom, using picocom commands.

• If a second user tries to remotely access the same port, at the same time, picocom won't let him (picocom will find the port locked and exit).

• In the example ~termbox/bin/ttySx scripts we have completely disabled the send- and receive-file picocom commands. This guarantees that picocom won't execute any external commands. If you want, you can enable the commands by providing specific file-upload and file-download programs as the arguments to the --send-cmd and --receive-cmd picocom command-line options (provided, of-course, that you trust these programs). Picocom (starting with release 2.0) does not use /bin/sh to execute the file-upload and file-download programs and will not let the user inject shell-commands when supplying additional arguments to them.

• If you allow send- and receive-file operations as described above, you will, most likely, also need a way for your users to put files on termbox, and get files back from it. There are many ways to arrange for this, but they are beyond the scope of this simple example.

Again, this is only one possible setup. There are countless other variations and elaborations you can try. Be creative!

Some notes on custom baudrate support

Custom baudrate support gives you the ability to set arbitrary baudrate values (like 1234, or 42000, etc) to a serial port, provided that the underlying driver can handle this. Since release 2.0, picocom can be compiled with custom baudrate support for some systems. Since release 3.1 picocom is compiled with support enabled by default on some systems (like Linux, kernels > 2.6, on x86 and x86_64, modern Intel Macs, and some BSDs). In any case, you can explicitly ask for support to be enabled by compiling picocom like this:

CPPFLAGS=-DUSE_CUSTOM_BAUD make clean
CPPFLAGS=-DUSE_CUSTOM_BAUD make

If custom baudrate support is not available for your system, the compilation will fail. Similarly, you can ask for support to be disabled by compiling like:

CPPFLAGS=-DNO_CUSTOM_BAUD make clean
CPPFLAGS=-DNO_CUSTOM_BAUD make

(or you can comment in or out the respective lines in the Makefile)

When picocom is compiled with custom baudrate support on Linux, it uses a new set of ioctl's (TCGETS2, TCSETSF2 vs TCGETS, TCSETSF, etc) to access the serial ports. It is not impossible that some systems or some serial devices may not accept these new ioctl's (though they should). In order to be able to use picocom even in this case, and without recompiling it, you can disable the custom baudrate support at runtime, and force picocom to use the "old" ioctls. To do this (starting with release 3.2) just define the environment variable NO_CUSTOM_BAUD before running picocom. Something like this:

NO_CUSTOM_BAUD=1 picocom ...

This only applies to Linux, and to picocom binaries that have been compiled with custom baudrate support.

To see if your binary has been compiled with custom baudrate support, and / or if it has detected the NO_CUSTOM_BAUD variable, run it with the --help option, and take a look at the first few lines of output.