diff options
Diffstat (limited to 'public/x11-systemd.html')
| -rw-r--r-- | public/x11-systemd.html | 188 |
1 files changed, 188 insertions, 0 deletions
diff --git a/public/x11-systemd.html b/public/x11-systemd.html new file mode 100644 index 0000000..0f751ca --- /dev/null +++ b/public/x11-systemd.html @@ -0,0 +1,188 @@ +<!DOCTYPE html> +<html lang="en"> +<head> + <meta charset="utf-8"> + <title>My X11 setup with systemd — Luke Shumaker</title> + <link rel="stylesheet" href="assets/style.css"> + <link rel="alternate" type="application/atom+xml" href="./index.atom" name="web log entries"/> +</head> +<body> +<header><a href="/">Luke Shumaker</a> » <a href=/blog>blog</a> » x11-systemd</header> +<article> +<h1 id="my-x11-setup-with-systemd">My X11 setup with systemd</h1> +<p>Somewhere along the way, I decided to use systemd user sessions to manage the various parts of my X11 environment would be a good idea. If that was a good idea or not... we'll see.</p> +<p>I've sort-of been running this setup as my daily-driver for <a href="https://lukeshu.com/git/dotfiles.git/commit/?id=a9935b7a12a522937d91cb44a0e138132b555e16">a bit over a year</a>, continually tweaking it though.</p> +<p>My setup is substantially different than the one on <a href="https://wiki.archlinux.org/index.php/Systemd/User">ArchWiki</a>, because the ArchWiki solution assumes that there is only ever one X server for a user; I like the ability to run <code>Xorg</code> on my real monitor, and also have <code>Xvnc</code> running headless, or start my desktop environment on a remote X server. Though, I would like to figure out how to use systemd socket activation for the X server, as the ArchWiki solution does.</p> +<p>This means that all of my graphical units take <code>DISPLAY</code> as an <code>@</code> argument. To get this to all work out, this goes in each <code>.service</code> file, unless otherwise noted:</p> +<pre><code>[Unit] +After=X11@%i.target +Requisite=X11@%i.target +[Service] +Environment=DISPLAY=%I</code></pre> +<p>We'll get to <code>X11@.target</code> later, what it says is "I should only be running if X11 is running".</p> +<p>I eschew complex XDMs or <code>startx</code> wrapper scripts, opting for the more simple <code>xinit</code>, which I either run on login for some boxes (my media station), or type <code>xinit</code> when I want X11 on others (most everything else). Essentially, what <code>xinit</code> does is run <code>~/.xserverrc</code> (or <code>/etc/X11/xinit/xserverrc</code>) to start the server, then once the server is started (which it takes a substantial amount of magic to detect) it runs run <code>~/.xinitrc</code> (or <code>/etc/X11/xinit/xinitrc</code>) to start the clients. Once <code>.xinitrc</code> finishes running, it stops the X server and exits. Now, when I say "run", I don't mean execute, it passes each file to the system shell (<code>/bin/sh</code>) as input.</p> +<p>Xorg requires a TTY to run on; if we log in to a TTY with <code>logind</code>, it will give us the <code>XDG_VTNR</code> variable to tell us which one we have, so I pass this to <code>X</code> in <a href="https://lukeshu.com/git/dotfiles.git/tree/.config/X11/serverrc">my <code>.xserverrc</code></a>:</p> +<pre><code>#!/hint/sh +if [ -z "$XDG_VTNR" ]; then + exec /usr/bin/X -nolisten tcp "$@" +else + exec /usr/bin/X -nolisten tcp "$@" vt$XDG_VTNR +fi</code></pre> +<p>This was the default for <a href="https://projects.archlinux.org/svntogit/packages.git/commit/trunk/xserverrc?h=packages/xorg-xinit&id=f9f5de58df03aae6c8a8c8231a83327d19b943a1">a while</a> in Arch, to support <code>logind</code>, but was <a href="https://projects.archlinux.org/svntogit/packages.git/commit/trunk/xserverrc?h=packages/xorg-xinit&id=5a163ddd5dae300e7da4b027e28c37ad3b535804">later removed</a> in part because <code>startx</code> (which calls <code>xinit</code>) started adding it as an argument as well, so <code>vt$XDG_VTNR</code> was being listed as an argument twice, which is an error. IMO, that was a problem in <code>startx</code>, and they shouldn't have removed it from the default system <code>xserverrc</code>, but that's just me. So I copy/pasted it into my user <code>xserverrc</code>.</p> +<p>That's the boring part, though. Where the magic starts happening is in <a href="https://lukeshu.com/git/dotfiles.git/tree/.config/X11/clientrc">my <code>.xinitrc</code></a>:</p> +<pre><code>#!/hint/sh + +if [ -z "$XDG_RUNTIME_DIR" ]; then + printf "XDG_RUNTIME_DIR isn't set\n" >&2 + exit 6 +fi + +_DISPLAY="$(systemd-escape -- "$DISPLAY")" +trap "rm -f $(printf '%q' "${XDG_RUNTIME_DIR}/x11-wm@${_DISPLAY}")" EXIT +mkfifo "${XDG_RUNTIME_DIR}/x11-wm@${_DISPLAY}" + +cat < "${XDG_RUNTIME_DIR}/x11-wm@${_DISPLAY}" & +systemctl --user start "X11@${_DISPLAY}.target" & +wait +systemctl --user stop "X11@${_DISPLAY}.target"</code></pre> +<p>There are two contracts/interfaces here: the <code>X11@DISPLAY.target</code> systemd target, and the <code>${XDG_RUNTIME_DIR}/x11-wm@DISPLAY</code> named pipe. The systemd <code>.target</code> should be pretty self explanatory; the most important part is that it starts the window manager. The named pipe is just a hacky way of blocking until the window manager exits ("traditional" <code>.xinitrc</code> files end with the line <code>exec your-window-manager</code>, so this mimics that behavior). It works by assuming that the window manager will open the pipe at startup, and keep it open (without necessarily writing anything to it); when the window manager exits, the pipe will get closed, sending EOF to the <code>wait</code>ed-for <code>cat</code>, allowing it to exit, letting the script resume. The window manager (WMII) is made to have the pipe opened by executing it this way in <a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user/wmii@.service">its <code>.service</code> file</a>:</p> +<pre><code>ExecStart=/usr/bin/env bash -c 'exec 8>${XDG_RUNTIME_DIR}/x11-wm@%I; exec wmii'</code></pre> +<p>which just opens the file on file descriptor 8, then launches the window manager normally. The only further logic required by the window manager with regard to the pipe is that in the window manager <a href="https://lukeshu.com/git/dotfiles.git/tree/.config/wmii-hg/config.sh">configuration</a>, I should close that file descriptor after forking any process that isn't "part of" the window manager:</p> +<pre><code>runcmd() ( + ... + exec 8>&- # xinit/systemd handshake + ... +)</code></pre> +<p>So, back to the <code>X11@DISPLAY.target</code>; I configure what it "does" with symlinks in the <code>.requires</code> and <code>.wants</code> directories:</p> +<ul class="tree"> +<li> +<p><a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user">.config/systemd/user/</a></p> +<ul> +<li><a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user/X11@.target">X11@.target</a></li> +<li><a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user/X11@.target.requires">X11@.target.requires</a>/ +<ul> +<li>wmii@.service -> ../<a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user/wmii@.service">wmii@.service</a></li> +</ul></li> +<li><a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user/X11@.target.wants">X11@.target.wants</a>/ +<ul> +<li>xmodmap@.service -> ../<a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user/xmodmap@.service">xmodmap@.service</a></li> +<li>xresources-dpi@.service -> ../<a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user/xresources-dpi@.service">xresources-dpi@.service</a></li> +<li>xresources@.service -> ../<a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user/xresources@.service">xresources@.service</a></li> +</ul></li> +</ul> +</li> +</ul> +<p>The <code>.requires</code> directory is how I configure which window manager it starts. This would allow me to configure different window managers on different displays, by creating a <code>.requires</code> directory with the <code>DISPLAY</code> included, e.g. <code>X11@:2.requires</code>.</p> +<p>The <code>.wants</code> directory is for general X display setup; it's analogous to <code>/etc/X11/xinit/xinitrc.d/</code>. All of the files in it are simple <code>Type=oneshot</code> service files. The <a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user/xmodmap@.service">xmodmap</a> and <a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user/xresources@.service">xresources</a> files are pretty boring, they're just systemd versions of the couple lines that just about every traditional <code>.xinitrc</code> contains, the biggest difference being that they look at <a href="https://lukeshu.com/git/dotfiles.git/tree/.config/X11/modmap"><code>~/.config/X11/modmap</code></a> and <a href="https://lukeshu.com/git/dotfiles.git/tree/.config/X11/resources"><code>~/.config/X11/resources</code></a> instead of the traditional locations <code>~/.xmodmap</code> and <code>~/.Xresources</code>.</p> +<p>What's possibly of note is <a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user/xresources-dpi@.service"><code>xresources-dpi@.service</code></a>. In X11, there are two sources of DPI information, the X display resolution, and the XRDB <code>Xft.dpi</code> setting. It isn't defined which takes precedence (to my knowledge), and even if it were (is), application authors wouldn't be arsed to actually do the right thing. For years, Firefox (well, Iceweasel) happily listened to the X display resolution, but recently it decided to only look at <code>Xft.dpi</code>, which objectively seems a little silly, since the X display resolution is always present, but <code>Xft.dpi</code> isn't. Anyway, Mozilla's change drove me to to create a <a href="https://lukeshu.com/git/dotfiles/tree/.local/bin/xrdb-set-dpi">script</a> to make the <code>Xft.dpi</code> setting match the X display resolution. Disclaimer: I have no idea if it works if the X server has multiple displays (with possibly varying resolution).</p> +<pre><code>#!/usr/bin/env bash +dpi=$(LC_ALL=C xdpyinfo|sed -rn 's/^\s*resolution:\s*(.*) dots per inch$/\1/p') +xrdb -merge <<<"Xft.dpi: ${dpi}"</code></pre> +<p>Since we want XRDB to be set up before any other programs launch, we give both of the <code>xresources</code> units <code>Before=X11@%i.target</code> (instead of <code>After=</code> like everything else). Also, two programs writing to <code>xrdb</code> at the same time has the same problem as two programs writing to the same file; one might trash the other's changes. So, I stuck <code>Conflicts=xresources@:i.service</code> into <code>xresources-dpi.service</code>.</p> +<p>And that's the "core" of my X11 systemd setup. But, you generally want more things running than just the window manager, like a desktop notification daemon, a system panel, and an X composition manager (unless your window manager is bloated and has a composition manager built in). Since these things are probably window-manager specific, I've stuck them in a directory <code>wmii@.service.wants</code>:</p> +<ul class="tree"> +<li> +<p><a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user">.config/systemd/user/</a></p> +<ul> +<li><a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user/wmii@.service.wants">wmii@.service.wants</a>/ +<ul> +<li>dunst@.service -> ../<a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user/dunst@.service">dunst@.service</a> # a notification daemon</li> +<li>lxpanel@.service -> ../<a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user/lxpanel@.service">lxpanel@.service</a> # a system panel</li> +<li>rbar@97_acpi.service -> ../<a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user/rbar@.service">rbar@.service</a> # wmii stuff</li> +<li>rbar@99_clock.service -> ../<a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user/rbar@.service">rbar@.service</a> # wmii stuff</li> +<li>xcompmgr@.service -> ../<a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user/xcompmgr@.service">xcompmgr@.service</a> # an X composition manager</li> +</ul></li> +</ul> +</li> +</ul> +<p>For the window manager <code>.service</code>, I <em>could</em> just say <code>Type=simple</code> and call it a day (and I did for a while). But, I like to have <code>lxpanel</code> show up on all of my WMII tags (desktops), so I have <a href="https://lukeshu.com/git/dotfiles.git/tree/.config/wmii-hg/config.sh">my WMII configuration</a> stick this in the WMII <a href="https://lukeshu.com/git/dotfiles.git/tree/.config/wmii-hg/rules"><code>/rules</code></a>:</p> +<pre><code>/panel/ tags=/.*/ floating=always</code></pre> +<p>Unfortunately, for this to work, <code>lxpanel</code> must be started <em>after</em> that gets inserted into WMII's rules. That wasn't a problem pre-systemd, because <code>lxpanel</code> was started by my WMII configuration, so ordering was simple. For systemd to get this right, I must have a way of notifying systemd that WMII's fully started, and it's safe to start <code>lxpanel</code>. So, I stuck this in <a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user/wmii@.service">my WMII <code>.service</code> file</a>:</p> +<pre><code># This assumes that you write READY=1 to $NOTIFY_SOCKET in wmiirc +Type=notify +NotifyAccess=all</code></pre> +<p>and this in <a href="https://lukeshu.com/git/dotfiles.git/tree/.config/wmii-hg/wmiirc">my WMII configuration</a>:</p> +<pre><code>systemd-notify --ready || true</code></pre> +<p>Now, this setup means that <code>NOTIFY_SOCKET</code> is set for all the children of <code>wmii</code>; I'd rather not have it leak into the applications that I start from the window manager, so I also stuck <code>unset NOTIFY_SOCKET</code> after forking a process that isn't part of the window manager:</p> +<pre><code>runcmd() ( + ... + unset NOTIFY_SOCKET # systemd + ... + exec 8>&- # xinit/systemd handshake + ... +)</code></pre> +<p>Unfortunately, because of a couple of <a href="https://github.com/systemd/systemd/issues/2739">bugs</a> and <a href="https://github.com/systemd/systemd/issues/2737">race conditions</a> in systemd, <code>systemd-notify</code> isn't reliable. If systemd can't receive the <code>READY=1</code> signal from my WMII configuration, there are two consequences:</p> +<ol type="1"> +<li><code>lxpanel</code> will never start, because it will always be waiting for <code>wmii</code> to be ready, which will never happen.</li> +<li>After a couple of minutes, systemd will consider <code>wmii</code> to be timed out, which is a failure, so then it will kill <code>wmii</code>, and exit my X11 session. That's no good!</li> +</ol> +<p>Using <code>socat</code> to send the message to systemd instead of <code>systemd-notify</code> "should" always work, because it tries to read from both ends of the bi-directional stream, and I can't imagine that getting EOF from the <code>UNIX-SENDTO</code> end will ever be faster than the systemd manager from handling the datagram that got sent. Which is to say, "we work around the race condition by being slow and shitty."</p> +<pre><code>socat STDIO UNIX-SENDTO:"$NOTIFY_SOCKET" <<<READY=1 || true</code></pre> +<p>But, I don't like that. I'd rather write my WMII configuration to the world as I wish it existed, and have workarounds encapsulated elsewhere; <a href="http://blog.robertelder.org/interfaces-most-important-software-engineering-concept/">"If you have to cut corners in your project, do it inside the implementation, and wrap a very good interface around it."</a>. So, I wrote a <code>systemd-notify</code> compatible <a href="https://lukeshu.com/git/dotfiles.git/tree/.config/wmii-hg/workarounds.sh">function</a> that ultimately calls <code>socat</code>:</p> +<pre><code>## +# Just like systemd-notify(1), but slower, which is a shitty +# workaround for a race condition in systemd. +## +systemd-notify() { + local args + args="$(getopt -n systemd-notify -o h -l help,version,ready,pid::,status:,booted -- "$@")" + ret=$?; [[ $ret == 0 ]] || return $ret + eval set -- "$args" + + local arg_ready=false + local arg_pid=0 + local arg_status= + while [[ $# -gt 0 ]]; do + case "$1" in + -h|--help) command systemd-notify --help; return $?;; + --version) command systemd-notify --version; return $?;; + --ready) arg_ready=true; shift 1;; + --pid) arg_pid=${2:-$$}; shift 2;; + --status) arg_status=$2; shift 2;; + --booted) command systemd-notify --booted; return $?;; + --) shift 1; break;; + esac + done + + local our_env=() + if $arg_ready; then + our_env+=("READY=1") + fi + if [[ -n "$arg_status" ]]; then + our_env+=("STATUS=$arg_status") + fi + if [[ "$arg_pid" -gt 0 ]]; then + our_env+=("MAINPID=$arg_pid") + fi + our_env+=("$@") + local n + printf -v n '%s\n' "${our_env[@]}" + socat STDIO UNIX-SENDTO:"$NOTIFY_SOCKET" <<<"$n" +}</code></pre> +<p>So, one day when the systemd bugs have been fixed (and presumably the Linux kernel supports passing the cgroup of a process as part of its credentials), I can remove that from <code>workarounds.sh</code>, and not have to touch anything else in my WMII configuration (I do use <code>systemd-notify</code> in a couple of other, non-essential, places too; this wasn't to avoid having to change just 1 line).</p> +<p>So, now that <code>wmii@.service</code> properly has <code>Type=notify</code>, I can just stick <code>After=wmii@.service</code> into my <code>lxpanel@.service</code>, right? Wrong! Well, I <em>could</em>, but my <code>lxpanel</code> service has nothing to do with WMII; why should I couple them? Instead, I create <a href="https://lukeshu.com/git/dotfiles/tree/.config/systemd/user/wm-running@.target"><code>wm-running@.target</code></a> that can be used as a synchronization point:</p> +<pre><code># wmii@.service +Before=wm-running@%i.target + +# lxpanel@.service +After=X11@%i.target wm-running@%i.target +Requires=wm-running@%i.target</code></pre> +<p>Finally, I have my desktop started and running. Now, I'd like for programs that aren't part of the window manager to not dump their stdout and stderr into WMII's part of the journal, like to have a record of which graphical programs crashed, and like to have a prettier cgroup/process graph. So, I use <code>systemd-run</code> to run external programs from the window manager:</p> +<pre><code>runcmd() ( + ... + unset NOTIFY_SOCKET # systemd + ... + exec 8>&- # xinit/systemd handshake + exec systemd-run --user --scope -- sh -c "$*" +)</code></pre> +<p>I run them as a scope instead of a service so that they inherit environment variables, and don't have to mess with getting <code>DISPLAY</code> or <code>XAUTHORITY</code> into their units (as I <em>don't</em> want to make them global variables in my systemd user session).</p> +<p>I'd like to get <code>lxpanel</code> to also use <code>systemd-run</code> when launching programs, but it's a low priority because I don't really actually use <code>lxpanel</code> to launch programs, I just have the menu there to make sure that I didn't break the icons for programs that I package (I did that once back when I was Parabola's packager for Iceweasel and IceCat).</p> +<p>And that's how I use systemd with X11.</p> + +</article> +<footer> +<p>The content of this page is Copyright © 2016 <a href="mailto:lukeshu@sbcglobal.net">Luke Shumaker</a>.</p> +<p>This page is licensed under the <a href="https://creativecommons.org/licenses/by-sa/3.0/">CC BY-SA-3.0</a> license.</p> +</footer> +</body> +</html> |