Jan 112013
 

Customising keyboard maps in Linux is somewhat … confused with lots of different tools and layers to perform the same task. There are a number of tools for performing some form of keyboard mapping, but the most common ones have some disadvantages :-

  •  xkb (which is the modern X way), and xmodmap (which is deprecated but conveniently has a very simple syntax for dealing with a single key) both work fine for ordinary keys but cannot do anything with “unusual” keys not passed into X. Just look online for just how many people have trouble with multimedia keys not being recognised.
  • The PS/2-specific tools of dumpkeys, loadkeys, and setkeycodes which work fine, but are somewhat reluctant to help out with USB keyboards.
  • Plus the desktop environment you are using may well have its own idea of how the keyboard will be used (GNOME has a nasty tendency to grab the menu key away from me).

There is fortunately another way which is rather difficult to find information about. Which is the reason behind this posting of course.

This “other method” is to use the generic input system to perform the keyboard mapping which has certain advantages over other methods. Most of the information to do this came from a README file contained within the source code.

The Example Keyboard

To demonstrate keyboard mapping, it is helpful to have an example keyboard with custom mappings to play with. Many of the keyboards I use this for are rather complex with many mappings, but I also have a mini keyboard with relatively few mappings :-

Original Key new function
Esc Lock screen
`/~ Esc
Caps Lock Control
Insert Delete
Delete `/~

No great mystery as to why I want my keyboard mapped this way – I’m just fussy about keyboards.

The Basic “Tool”

In fact there is just one tool – /lib/udev/keymap – which performs all of the relevant tasks. Before it can do anything, it needs to be provided with the path of the relevant input device. This is easiest done from the console (rather than in X) as root. The easiest way of identifying the device is to unplug the keyboard, reboot the machine, and :-

# ls /dev/input/e* > /var/tmp/old.list
[Plug in keyboard]
# ls /dev/input/e* > /var/tmp/new.list
# diff /var/tmp/old.list /var/tmp/old.list
> /dev/input/event13
> /dev/input/event14

If you are lucky, there will be just one new input device. If not, you will have to try each one in turn. The first job is to record the keycode of each key to be customised in turn. To do this, it is necessary to run keymap with the input device and the “-i” option, and each keystroke will result in some output :-

# /lib/udev/keymap /dev/input/event13 -i
Press ESC to finish, or Control-C if this device is not your primary keyboard
scan code: 0x70029   key code: esc
# /lib/udev/keymap /dev/input/event13 -i
Press ESC to finish, or Control-C if this device is not your primary keyboard
scan code: 0x70035   key code: grave
scan code: 0x70039   key code: capslock
scan code: 0x70049   key code: insert
scan code: 0x7004C   key code: delete

A key can be mapped temporarily using keymap. But before that a list of possible key names is useful to have; there is one to be found in /usr/include/ :-

# grep KEY_ /usr/include/linux/input.h | less

The relevant name would be the part that follows the “KEY_” converted to lower-case.

# /lib/udev/keymap /dev/input/event13 0x70035 esc

But that is rather a temporary solution; it is better by far to create a file containing the necessary mappings to be automatically applied :-

# cat /tmp/custom-filco.map
0x70029 screenlock
#	Original: key code: esc
0x70035 esc
#	Original: key code: grave
0x70039 leftctrl
#	Original: key code: capslock
0x70049 delete
#	Original: key code: insert
0x7004C insert
#	Original: key code: delete

Making The Mappings Permanent

The first step is to obtain some details to uniquely (or as much as possible) identify the keyboard. Run :-

# udevadm info --export-db > /tmp/udev-db.txt

And look through the output for the input device you previously used. Look for a ID_VENDOR_ID and ID_MODEL_ID that you can use.

Next add a rule to /lib/udev/rules.d/95-keymap.rules along the lines of :-

ENV{ID_VENDOR_ID}=="04d9", ENV{ID_MODEL_ID}=="2011", RUN+="keymap $name custom-filco.map"

Once this is working you may want to add it to your version of custom-filco.map as a comment to preserve it for use after upgrades; alternatively you may wish to create a new file that will not get overwritten.

Before activating the new rule, remember to copy /tmp/custom-filco.map into /lib/keymaps/custom-filco.map. And again keep another copy in a safe place to preserve.

As to how to activate, a reboot is probably the simplest way.

Dec 312012
 

This has been updated for Awesome 3.5

I have long been interested in tiling window managers, which would seem to offer the flexibility I crave. But switching is painful when you want your machine to be as functional as much as possible. On at least two occasions I have switched … and very quickly switched back when problems were encountered.

So just for the fun of it, I have decided to give it another go … in a way that can be gradually worked into functionality.

But first, what exactly is a tiling window manager? Well a conventional window manager works with overlapping windows where windows can stack on top of each other and overlap. A tiling window manager on the other hand uses tiling to maximise the size of every displayed window.

Or in other words, if you want to make the maximum use of space on your screen, a stacking window manager usually asks you to manually move and resize windows yourself whereas a tiling window manager does it all automatically. Although with a tiling window manager, you will need to rotate amongst different tiling layouts to suit.

Tiling window managers tend to be keyboard driven rather than mouse driven, but this is merely an aspect of their implementation.

A Testing Environment

The key to setting up any window manager gradually over time without losing an existing working environment, is to create a separate display which a new window manager can control. This can be quite easily accomplished with a nesting X server, which is provided by the xserver-xephyr package.

As to the choice of window manager, the title of this post gives away the secret – Awesome. Of course the same method works for other window managers too.

The first step is to install the required packages :-

sudo apt-get install xserver-xephyr awesome awesome-extra

The next step is to create a simple shell script to start the nested X server. This basically starts the nested X server with a sensible screen size, and sets the $DISPLAY environment variable so that any further X programs are directed to communicate with the nested X server. The nested X server simply starts up as a standard window within which Awesome operates. The script looks like :-

#!/bin/sh

Xephyr -ac -reset -screen 1280x1000 :1 &
export DISPLAY=:1.0
sleep 5
awesome --config ~/src/site/etc/awesome/config.lua

The sleep command in there is to wait until X has settled down before trying to manage it … without this, it is possible that awesome will fail to open the display (on a slower laptop it worked without; on a faster desktop it did not). The parameter to awesome is to set the preferred location for the configuration file (which I will describe next), although you can leave it out to use the default.

Don’t start it yet! Before doing that, copy the default configuration file to the specified location so you have something to work with :-

cp /etc/xdg/awesome/rc.lua ~/src/site/etc/awesome/config.lua

It’s All In The Repository

The custom configuration file I specified just happens to point to a directory and file within the working copy of a Subversion repository. Using a repository for configuration files has all sorts of benefits; not least of which is to reverse “poor” choices made along the way. Whilst describing how to get one set up is beyond the purpose of this post, it is well worth doing.

Of course if you are working with a repository, it is important to remember to commit changes at every stage :-

% svn add ~/src/site/etc/awesome 
A         /home/mike/src/site/etc/awesome
A         /home/mike/src/site/etc/awesome/config.lua
% svn commit ~/src/site/etc/awesome -m "Awesome: Initial Configuration from default"
Adding         src/site/etc/awesome
Adding         src/site/etc/awesome/config.lua
Transmitting file data .
Committed revision 166.

I won’t be documenting every single commit, but they are all in there!

It is also helpful (at this stage) to create a symlink between the default configuration directory (~/.config/awesome) and the working copy of your repository. The configuration directory can probably be changed, but one thing at a time!

Status Bar Size

One of the annoyances I remember from the last time I played with Awesome, was the tiny size of the status bar. It is all very well maximising the size of the main screen area, but some of us have high resolution screens and tired old eyes! Within the config.lua file (or whatever you named your file), look for :-

mywibox[s] = awful.wibox({ position = "top", screen = s})

And change it to add a height parameter :-

mywibox[s] = awful.wibox({ position = "top", screen = s, height=32 })

The exact choice of size is of course down to personal preferences! However changing the size also means changing the font would be wise :-

awesome.font = "LMSans10 10"

Making the font too much larger will result in some display oddities … at least for now.

Adding A Title Bar

In theory, there is no need for a title bar with a tiling window manager, but I guess I am a little too conventional as a window without a title bar still looks a little weird to me. Plus my shell prompts stick useful bits of information into the window title bar and I would really rather hang on to that extra information. Still, drawing title bars on windows turns out to be really simple … just find the existing code to do so and :-

local titlebars_enabled = true

Some appearance tweaks and indeed messing around with the unnecessarily large number of buttons could be usefully accomplished, but for now this is fine.

 

Themes

Of course messing around with the appearance of a window manager is probably best dealt with using themes (if they are supported). The default configuration file I worked with included enabling beautiful via :-

require("beautiful")
beautiful.init("/usr/share/awesome/themes/default/theme.lua")

It is easier to tinker with themes with a configuration file you can edit, so I copied this theme.lua to the configuration directory named as default-theme.lua, and changed the configuration :-

-- beautiful.init("/usr/share/awesome/themes/default/theme.lua")
beautiful.init(awful.util.getdir("config") .. "/default-theme.lua")

I also removed the previous section’s font setting and set the same font within the new configuration file :-

theme.font = "LMSans10 10"

To cope with the larger font, I also increased the width of each menu item :-

theme.menu_width  = "300"

Although this width does depend on having an appropriately sized screen, and not having too many levels to your menu (if it gets too wide, the extra levels don’t appear neatly).

And finally I commented out the command used to set the background … having a background image for a tiling window manager (which is designed to utilise the whole screen) is a touch unnecessary. And I don’t particularly like the default image :-

--theme.wallpaper_cmd = ...

For now I will leave the colours alone … the defaults may not be perfect but are not sufficiently irritating to spend time on right now. But the details of how to change colours of various components can be found in the theme file we have created.

Disabling Layouts

Personally I believe that you can have too much of a good thing, and choosing between the rather large number of different possible layouts that Awesome makes available is one of those. To reduce the number of possibilities, simply comment out some of the possibilities when setting the “layouts” variable in the configuration file :-

layouts =
{
    awful.layout.suit.floating,
    awful.layout.suit.tile,
    --awful.layout.suit.tile.left,
    --awful.layout.suit.tile.bottom,
    --awful.layout.suit.tile.top,
    awful.layout.suit.fair,
    --awful.layout.suit.fair.horizontal,
    awful.layout.suit.spiral,
    --awful.layout.suit.spiral.dwindle,
    awful.layout.suit.max,
    --awful.layout.suit.max.fullscreen,
    --awful.layout.suit.magnifier
}

Which you command out does not matter so much, and indeed you may very well change you mind at a later point.

A Menu

Whilst the menu built into Awesome has undoubtedly many advantages, I already have a somewhat adequate menu built up over years that relies on myGtkMenu. Whilst changing this menu could well be sensible, it does also seem sensible to keep it around during the transitional period. To do so is simply a matter of replacing certain bindings :-

 -- {{{ Mouse bindings
 root.buttons(awful.util.table.join(
-    awful.button({ }, 3, function () mymainmenu:toggle() end),
+    --awful.button({ }, 3, function () mymainmenu:toggle() end),
+    awful.button({ }, 3, function () awful.util.spawn("/opt/bin/myGtkMenu /home/mike/lib/root.menu") end),
     awful.button({ }, 4, awful.tag.viewnext),
     awful.button({ }, 5, awful.tag.viewprev)
 ))

If you hadn’t guessed, I’ve started using diffs to show the changes where it makes sense; if you cannot read diffs, this page is probably not for you.

That bit replaces the root window, but not the status bar menu :-

 mylauncher = awful.widget.launcher({ image = image(beautiful.awesome_icon),
-                                     menu = mymainmenu })
+--                                     menu = mymainmenu })
+                                       command = "/opt/bin/myGtkMenu /home/mike/lib/root.menu" })

My own menu has some issues in that it doesn’t yet have anything to control awesome itself. But is good enough to get started with.

Final Word … Or Not!

And that is it! Or at least it is enough to be getting on with.

Nov 242012
 

NTP is one of those strange services that are so vital to the operation of an organisation’s network; if the servers around the network get their time in a muddle, all sorts of strange things can start happening. Besides which most people expect their computers to be able to tell the right time.

But often it is one of the unloved services. After all no user is going to ask about the health of the NTP service. And if you are a senior manager involved in IT, do you know who manages your NTP infrastructure ? If so, have you ever asked them to explain the design of the NTP infrastructure ? If not, you may find a nasty surprise – your network’s NTP infrastructure may rely on whatever servers can be scavenged and with the minimum investment of time.

Of course, NTP is pretty reliable and in most circumstances extremely resilient. NTP has built in safeguards against against confused time servers sending wildly inappropriate time adjustments, and even in the event of a total NTP failure, servers should be able to keep reasonable time for at least a while. Even with a minimal of investment, an NTP infrastructure can often run merrily in the background for years without an issue.

Not that it is a good idea to ignore NTP for years. It is better by far to spend a little time and money on a yearly basis to keep things fresh – perhaps a little server, and a day’s time each year.

That was quite a long rambling introduction to the NTP “glitch” that I learned about this week, but perhaps goes some way to explaining why such a glitch occurred.

A number of organisations reported that their network had started reporting a time way back in the year 2000. It turns out that :-

  • The USN(aval)O(observatory) had a server that for 51 minutes reported the year as 2000 rather than 2012.
  • A number of organisations with an insufficient number of clock sources (i.e. just the erroneous USNO one) attempted to synchronise to the year 2000 causing the NTP daemon to stop.
  • Some “clever” servers noticed that NTP had stopped, and restarted it. Because most default NTP startup scripts set the clock on startup, these servers were suddenly sent back in time to the year 2000.

And a cascade of relative minor issues, becomes a major issue.

Reading around, the recommendations to prevent this sort of thing happening :-

  1. Use an appropriate number of time sources for your main NTP servers; various suggestions have been made ranging from 5 (probably too few) to 8 (perhaps about right) to 20 (possibly overkill).
  2. Have an appropriate number of main NTP servers for your servers (and other equipment) to synchronise their time with. Anything less than 3 is inadequate; more than 4 is recommended.
  3. Prevent your main NTP servers from setting their time when NTP is restarted and monitor the time on each server regularly.
  4. And a personal recommendation: Restart all your NTP daemons regularly – perhaps daily – to get them to check with the DNS for any updated NTP server names.
  5. And as suggested above, regularly review your NTP infrastructure.
Oct 172012
 

I have recently become interested in the amount of entropy available in Linux and decided to spend some time poking around on my Debian workstation. Specifically looking to increase the amount of entropy available to improve the speed of random number generation. There are a variety of different ways of accomplishing this including hardware devices (some of which cost rather too much for a simple experiment).

Eh?

Linux has a device (/dev/random) which makes available random numbers to software packages that really need access to a high quality source of random numbers. Any decently written cryptographic software will use /dev/random (and not /dev/urandom which does not generate “proper” random numbers of quality) to implement encryption.

Using poor quality random numbers can potentially result in encryption not being secure. Or perhaps more realisticallybecause Linux waits until there is sufficient entropy available before releasing numbers through /dev/random, software reading from that device may be subject to random stalling. Not necessarily long enough to cause a major problem, but perhaps enough to have an effect on performance.

Especially for a server in a virtualised environment!

Adding Entropy The Software Way (haveged)

HAVEGED is a way of using processor flutter to add entropy to the Linux /dev/random device. It can be installed relatively easily with :-

apt-get install haveged
/etc/init.d/haveged start

As soon as this was running the amount of entropy available (cat /proc/sys/kernel/random/entropy_avail) jumped from several hundred to close to 4,000.

Now does this increased entropy have an effect on performance? Copying a CD-sized ISO image file using ssh :-

Default entropy 29.496
With HAVEGED 28.636

A 2% improvement in performance is hardly a dramatic improvement, but every little bit helps and it may well have a more dramatic effect on a server which regularly exhausts entropy.

Checking The Randomness

But hang on … more important than performance is the randomness of the numbers generated. And you cannot mess with the generation of random numbers without checking the results. The first part of checking the randomness is making sure you have the right tools installed :-

apt-get install rng-tools

Once installed you can test the current set of random numbers :-

dd if=/dev/random bs=1k count=32768 iflag=fullblock| rngtest

This produces a whole bunch of output, but the key bits of output are the “FIPS 140-2 failures” and “FIPS 140-2 successes”; if you have too many failures something is wrong. For the record my failure rate is 0.05% with haveged running (without: tests ongoing).

Links

… to more information.

Jun 302012
 

Warning: This page details a shell script that I’ve produced for my own amusement; it isn’t a product. It hasn’t been tested in lots of environments, and it will take some hacking to get it to work for you. If you’re looking for something to use, move along; if you’re looking for ideas to improve a real wallpaper setting program, you might want to read on.

So elsewhere I’ve admitted to driving a stake through the heart of GNOME’s wallpaper plugin to allow my own wallpaper script to work. Well, I could hardly do that and not announce it could I? So here goes :-

  1. It doesn’t actually set the wallpaper; it lets hsetroot do that.
  2. It requires a parameter to determine which directory to choose – i.e. ~/lib/backgrounds/one~/lib/backgrounds/two, etc.
  3. It uses xrandr to pick out the “regions” of the default screen.
  4. It puts portrait images on my portrait monitor, and landscape images on my landscape monitor by overlaying them onto an overall image the size of both monitors added together.
  5. It waits a set duration, and then repeats.

If you’re still interested in getting a copy it’s available at http://zonky.org/src/set-random-background.