Bump version for v4.11 GA release

Signed-off-by: Joachim Wiberg <troglobit@gmail.com>

.github/CONTRIBUTING.md

@@ -24,7 +24,8 @@ the maintainer(s) and make it easier for them to include your code.
 Coding Style
 ------------
 
-> **Tip:** Always submit code that follows the style of surrounding code!
+> [!TIP]
+> Always submit code that follows the style of surrounding code!
 
 First of all, lines are allowed to be longer than 72 characters these
 days.  In fact, there exist no enforced maximum, but keeping it around

.github/workflows/coverity.yml

@@ -4,6 +4,7 @@ on:
   push:
     branches:
       - 'dev'
+  workflow_dispatch:
 
 env:
   PROJECT_NAME: finit

ChangeLog.md

@@ -3,8 +3,52 @@ Change Log
 
 All relevant changes are documented in this file.
 
+[4.11][] - 2025-03-27
+---------------------
 
-[4.9][UNRELEASED]
+### Changes
+ - Update documentation regarding transition from runlevel S to 2, how
+   run/tasks must "complete" before Finit can move to the next runlevel
+ - Improve argument handling in bundled getty when launching login
+
+### Fixes
+ - Fix #432: allow services in `setup` state to be stopped.  This fixes
+   a serious blocker introduced in 4.10 that prevented systems from
+   being rebooted when any service was in this state
+ - Fix fallback `SIGKILL` when stopping a service with `SIGTERM` does
+   not work.  This fixes a serious issue with services not responding to
+   the default `SIGTERM`
+ - Fix pre/post/ready/cleanup script execution if service has an
+   optional `env:file` (leading `-`)
+ - Fix blocked signals for pre/post/ready/cleanup scripts preventing
+   normal shutdown when a service/sysv is stopped in its setup state
+ - Fix issue where a pre:script creates a PID file in `/run` cause the
+   service to move from setup state to running.  Any `svc_t` in setup,
+   teardown, or cleanup state should be ignored in the pidfile plugin
+
+
+[4.10][] - 2025-03-03
+---------------------
+
+### Changes
+ - Add support for `cleanup:[0-3600,]script`, runs before service is
+   removed, unlike `post:script` that runs when service is stopped
+ - Improved entropy gathering in tertiary fallback of urandom plugin
+ - Reclassify `sysv` from run/task class to `service`
+
+### Fixes
+ - Fix #427: using an external getty with custom login program cause
+   Finit to use login program as the getty command
+ - Fix #428: allow `sysctl` path to be set at build time
+ - Fix #429: regression in v4.9 causing Finit to segfault at shutdown
+ - Fix possible memory leak when sourcing a service `env:file`
+ - Finit now waits until all `post:scripts` have run before returning
+   OK to `initctl <reload|runlevel>`.  This also applies to the new
+   `cleanup:scripts` introduced in this release
+ - RFC3154 conformance formatting fix for `/dev/kmsg` fallback logging
+
+
+[4.9][] - 2025-02-15
 --------------------
 
 > [!CAUTION]
@@ -29,7 +73,7 @@ All relevant changes are documented in this file.
    does not support SIGHUP.  This helps avoid unnecessary restarts of
    services that depend on a service that supports SIGHUP
  - Fix #417: leaking inotify watchers in the cgroup subsystem
- - Fix #422: honor `notfail` flag in `/etc/fstab`
+ - Fix #421: add support for `nofail` mount option in `/etc/fstab`
  - Fix #425: ordering bug in detecting .conf file changes to services
    and `initctl reload`, or runlevel changes
  - Fix `initctl touch` of template services, previously marking a
@@ -68,7 +112,7 @@ All relevant changes are documented in this file.
 
 ### Fixes
  - Fix #397: system shutdown/reboot can block on console input if action
-   is started remotely (ssh).  Caused by legact TTY screen size probing,
+   is started remotely (ssh).  Caused by legacy TTY screen size probing,
    removed from both bootstrap and shutdown/reboot
  - Fix #400: both `HOOK_MOUNT_ERROR` and `sulogin()` fail to trigger on
    either mount or `fsck` errors.  Problem caused by unresolved status
@@ -326,7 +370,7 @@ All relevant changes are documented in this file.
   `/etc/finit.conf`.
 
   For the complete experience a new top-level configuration file directive
-  `rcsd PATH` has aslo been added.  It in turn can be used by `factory.conf`
+  `rcsd PATH` has also been added.  It in turn can be used by `factory.conf`
   as follows to override `/etc/finit.d`:
 
         rcsd /etc/factory.d
@@ -909,7 +953,7 @@ make sure to read the whole change log when upgrading.
   previously hard-coded in Finit, e.g., starting the bundled watchdogd,
   various distro packed udevd and other hotplugging tools
 * A proper rescue mode has been added.  It is started extremely early
-  and is protected with a bundled `suslogin`.  Exiting rescue mode now
+  and is protected with a bundled `sulogin`.  Exiting rescue mode now
   brings up the system as a normal boot, as one expects
 * Support for `sysv` start/stop scripts as well as monitoring forking
   services, stared using `sysv` or `service` stanza
@@ -1740,44 +1784,46 @@ Major bug fix release.
 
 * Initial release
 
-[UNRELEASED]: https://github.com/troglobit/finit/compare/4.8...HEAD
-[4.9]: https://github.com/troglobit/finit/compare/4.8...4.9
-[4.8]: https://github.com/troglobit/finit/compare/4.7...4.8
-[4.7]: https://github.com/troglobit/finit/compare/4.6...4.7
-[4.6]: https://github.com/troglobit/finit/compare/4.5...4.6
-[4.5]: https://github.com/troglobit/finit/compare/4.4...4.5
-[4.4]: https://github.com/troglobit/finit/compare/4.3...4.4
-[4.3]: https://github.com/troglobit/finit/compare/4.2...4.3
-[4.2]: https://github.com/troglobit/finit/compare/4.1...4.2
-[4.1]: https://github.com/troglobit/finit/compare/4.0...4.1
-[4.0]: https://github.com/troglobit/finit/compare/3.1...4.0
-[3.1]: https://github.com/troglobit/finit/compare/3.0...3.1
-[3.0]: https://github.com/troglobit/finit/compare/2.4...3.0
-[2.4]: https://github.com/troglobit/finit/compare/2.3...2.4
-[2.3]: https://github.com/troglobit/finit/compare/2.2...2.3
-[2.2]: https://github.com/troglobit/finit/compare/2.1...2.2
-[2.1]: https://github.com/troglobit/finit/compare/2.0...2.1
-[2.0]: https://github.com/troglobit/finit/compare/1.12...2.0
+[UNRELEASED]: https://github.com/troglobit/finit/compare/4.11...HEAD
+[4.11]: https://github.com/troglobit/finit/compare/4.10...4.11
+[4.10]: https://github.com/troglobit/finit/compare/4.9...4.10
+[4.9]:  https://github.com/troglobit/finit/compare/4.8...4.9
+[4.8]:  https://github.com/troglobit/finit/compare/4.7...4.8
+[4.7]:  https://github.com/troglobit/finit/compare/4.6...4.7
+[4.6]:  https://github.com/troglobit/finit/compare/4.5...4.6
+[4.5]:  https://github.com/troglobit/finit/compare/4.4...4.5
+[4.4]:  https://github.com/troglobit/finit/compare/4.3...4.4
+[4.3]:  https://github.com/troglobit/finit/compare/4.2...4.3
+[4.2]:  https://github.com/troglobit/finit/compare/4.1...4.2
+[4.1]:  https://github.com/troglobit/finit/compare/4.0...4.1
+[4.0]:  https://github.com/troglobit/finit/compare/3.1...4.0
+[3.1]:  https://github.com/troglobit/finit/compare/3.0...3.1
+[3.0]:  https://github.com/troglobit/finit/compare/2.4...3.0
+[2.4]:  https://github.com/troglobit/finit/compare/2.3...2.4
+[2.3]:  https://github.com/troglobit/finit/compare/2.2...2.3
+[2.2]:  https://github.com/troglobit/finit/compare/2.1...2.2
+[2.1]:  https://github.com/troglobit/finit/compare/2.0...2.1
+[2.0]:  https://github.com/troglobit/finit/compare/1.12...2.0
 [1.12]: https://github.com/troglobit/finit/compare/1.11...1.12
 [1.11]: https://github.com/troglobit/finit/compare/1.10...1.11
 [1.10]: https://github.com/troglobit/finit/compare/1.9...1.10
-[1.9]: https://github.com/troglobit/finit/compare/1.8...1.9
-[1.8]: https://github.com/troglobit/finit/compare/1.7...1.8
-[1.7]: https://github.com/troglobit/finit/compare/1.6...1.7
-[1.6]: https://github.com/troglobit/finit/compare/1.5...1.6
-[1.5]: https://github.com/troglobit/finit/compare/1.4...1.5
-[1.4]: https://github.com/troglobit/finit/compare/1.3...1.4
-[1.3]: https://github.com/troglobit/finit/compare/1.2...1.3
-[1.2]: https://github.com/troglobit/finit/compare/1.1...1.2
-[1.1]: https://github.com/troglobit/finit/compare/1.0...1.1
-[1.0]: https://github.com/troglobit/finit/compare/0.9...1.0
-[0.9]: https://github.com/troglobit/finit/compare/0.8...0.9
-[0.8]: https://github.com/troglobit/finit/compare/0.7...0.8
-[0.7]: https://github.com/troglobit/finit/compare/0.6...0.7
-[0.6]: https://github.com/troglobit/finit/compare/0.5...0.6
-[0.5]: https://github.com/troglobit/finit/compare/0.4...0.5
-[0.4]: https://github.com/troglobit/finit/compare/0.3...0.4
-[0.3]: https://github.com/troglobit/finit/compare/0.2...0.3
-[0.2]: https://github.com/troglobit/finit/compare/0.1...0.2
+[1.9]:  https://github.com/troglobit/finit/compare/1.8...1.9
+[1.8]:  https://github.com/troglobit/finit/compare/1.7...1.8
+[1.7]:  https://github.com/troglobit/finit/compare/1.6...1.7
+[1.6]:  https://github.com/troglobit/finit/compare/1.5...1.6
+[1.5]:  https://github.com/troglobit/finit/compare/1.4...1.5
+[1.4]:  https://github.com/troglobit/finit/compare/1.3...1.4
+[1.3]:  https://github.com/troglobit/finit/compare/1.2...1.3
+[1.2]:  https://github.com/troglobit/finit/compare/1.1...1.2
+[1.1]:  https://github.com/troglobit/finit/compare/1.0...1.1
+[1.0]:  https://github.com/troglobit/finit/compare/0.9...1.0
+[0.9]:  https://github.com/troglobit/finit/compare/0.8...0.9
+[0.8]:  https://github.com/troglobit/finit/compare/0.7...0.8
+[0.7]:  https://github.com/troglobit/finit/compare/0.6...0.7
+[0.6]:  https://github.com/troglobit/finit/compare/0.5...0.6
+[0.5]:  https://github.com/troglobit/finit/compare/0.4...0.5
+[0.4]:  https://github.com/troglobit/finit/compare/0.3...0.4
+[0.3]:  https://github.com/troglobit/finit/compare/0.2...0.3
+[0.2]:  https://github.com/troglobit/finit/compare/0.1...0.2
 [libuEv]: https://github.com/troglobit/libuev
 [dea3ae8]: https://github.com/troglobit/finit/commit/dea3ae8

README.md

@@ -61,12 +61,13 @@ distributions:
   * [Alpine Linux](contrib/alpine/), and
   * [Debian GNU/Linux](contrib/debian/), also works on Ubuntu/Linux Mint
 
-> **Note:** support for various Linux distributions does not mean Finit
-> installs easily on all architectures.  The bundled install scripts are
-> examples for standard installations, tested on amd64 (x86_64) systems.
-> Custom setups, e.g., for embedded systems, can be found in any of the
-> following [Buildroot][] based examples: [myLinux][], [Infix][], or the
-> plain [br2-finit-demo](https://github.com/troglobit/br2-finit-demo).
+> [!NOTE]
+> Support for various Linux distributions does not mean Finit installs
+> easily on all architectures.  The bundled install scripts are examples
+> for standard installations, tested on amd64 (x86_64) systems.  Custom
+> setups, e.g., for embedded systems, can be found in the following
+> [Buildroot][] based examples: [myLinux][], [Infix][], or the plain
+> [br2-finit-demo](https://github.com/troglobit/br2-finit-demo).
 
 
 Example
@@ -78,9 +79,10 @@ be placed in `/etc/finit.d/available` and enabled by an operator using
 the [initctl](#commands--status) tool.  See the above mentioned Linux
 distributions, or [myLinux][].
 
-> **Note:** as of Finit v4.4, .conf lines can be broken up using the
-> standard UNIX continuation character (`\`), also trailing comments are
-> now supported.  The latter means you need to escape any hashes used in
+> [!TIP]
+> As of Finit v4.4, `.conf` lines can be broken up using the standard
+> UNIX continuation character (`\`), trailing comments are now also
+> supported.  The latter means you need to escape any hashes used in
 > directives and descriptions (`\#`).  For more on this and examples,
 > see the [finit.conf(5)][] manual or [doc/config.md](doc/config.md).
 
@@ -285,6 +287,7 @@ run  [S]                /path/to/ident    --
 task [2] <hw/model/foo> /path/to/foo-init -- Initializing Foo board
 ```
 
+> [!TIP]
 > Notice the trick with an empty description to hide the call to `ident`
 > in the Finit progress output.
 
@@ -367,9 +370,10 @@ The `initctl` tool has three commands to help debug and optimize the
 setup and monitoring of cgroups.  See the `ps`, `top`, and `cgroup`
 commands for details.
 
-> **Note:** systems that do not support cgroups, specifically version 2,
-> are automatically detected.  On such systems the above functionality
-> is disabled early at boot.
+> [!NOTE]
+> Systems that do not support cgroups, specifically version 2, are
+> automatically detected.  On such systems the above functionality is
+> disabled early at boot.
 
 
 Runparts & /etc/rc.local
@@ -573,9 +577,10 @@ file must be used to tell Finit to stop and start it on `reload` and
 `runlevel` changes.  If `<>` holds more [conditions](doc/conditions.md),
 these will also affect how a service is maintained.
 
-> **Note:** even though it is possible to start services not belonging in
-> the current runlevel these services will not be respawned automatically
-> by Finit if they exit (crash).  Hence, if the runlevel is 2, the below
+> [!NOTE]
+> Even though it is possible to start services not belonging in the
+> current runlevel these services will not be respawned automatically by
+> Finit if they exit (crash).  Hence, if the runlevel is 2, the below
 > Dropbear SSH service will not be restarted if it is killed or exits.
 
 The `status` command is the default, it displays a quick overview of all

configure.ac

@@ -1,4 +1,4 @@
-AC_INIT([Finit], [4.9-rc1], [https://github.com/troglobit/finit/issues],
+AC_INIT([Finit], [4.11], [https://github.com/troglobit/finit/issues],
 	[finit], [https://troglobit.com/projects/finit/])
 AC_CONFIG_AUX_DIR(aux)
 AM_INIT_AUTOMAKE([1.11 foreign dist-xz subdir-objects])

contrib/alpine/README.md

@@ -1,7 +1,8 @@
 HowTo: Finit on Alpine Linux 3.4...3.19
 =======================================
 
-> **Blog** https://troglobit.com/post/2021-02-12-alpine-linux-with-finit/
+> [!TIP]
+> https://troglobit.com/post/2021-02-12-alpine-linux-with-finit/
 
 HowTo use Finit to boot an [Alpine Linux][] system.  It is assumed that
 the user has already installed make, a compiler, C library header files,

contrib/debian/README.md

@@ -20,8 +20,9 @@ tools pre-systemd work as intended.
 
     root@debian:~# apt install initscripts console-setup
 
-> **Note:** as of Debian 11 (Bullseye), libuev and libite are part of
-> the main section of Debian.  So just install the -dev packages :)
+> [!TIP]
+> As of Debian 11 (Bullseye), libuev and libite are part of the main
+> section of Debian.  So just install the -dev packages :)
 
 The following script can then be used to configure, build, install and
 set up your system to run Finit:
@@ -47,7 +48,8 @@ comands to `initctl`, followed by `reload` to activate your changes.
 You can also use a standard [/etc/rc.local](rc.local) for one-shot tasks
 and initialization like keyboard language etc.
 
-> **NOTE:** X Window system, you may need to `sudo apt install elogind`
+> [!NOTE]
+> For the X Window system, you may need to `sudo apt install elogind`
 > (Bullseye and later), followed by `initctl reload` to activate it (it
 > is enabled by default), and logout/login again.  The elogind daemon
 > ensures a regular non-root user can start and interact with an X

doc/build.md

@@ -1,5 +1,5 @@
-Building
-========
+Building Finit
+==============
 
 * [Introduction](#introduction)
 * [Configure](#configure)
@@ -17,14 +17,15 @@ optional plugins to enable.  It depends on two external libraries:
 - [libuEv][], the event loop
 - [libite][] (-lite), much needed frog DNA
 
-**NOTE:** Most free/open source software that uses `configure` default
-  to install to `/usr/local`.  However, some Linux distributions do no
-  longer search that path for installed software, e.g. Fedora and Alpine
-  Linux.  To get finit's configure script to find its dependencies you
-  have to help the `pkg-config` tool a bit if you do not change the
-  default prefix path:
-
-    PKG_CONFIG_LIBDIR=/usr/local/lib/pkgconfig ./configure
+> [!IMPORTANT]
+> Most free/open source software packages that use `configure` default
+> to install to `/usr/local`.  However, some Linux distributions do no
+> longer search that path for installed software, e.g. Fedora and Alpine
+> Linux.  To get finit's configure script to find its dependencies you
+> have to help the `pkg-config` tool a bit if you do not change the
+> default prefix path:
+>
+>     PKG_CONFIG_LIBDIR=/usr/local/lib/pkgconfig ./configure
 
 The configure script checks for all dependencies, including the correct
 version of the above mentioned libraries.  Currently required versions:
@@ -71,12 +72,13 @@ Below are a few of the main switches to configure:
 
 For more configure flags, see <kbd>./configure --help</kbd>
 
-> **Note:** the configure script is not available in the GIT sources.  It is
->           however included in (officially supported) released tarballs.  The
->           idea is that you should not need GNU autotools to build, only the
->           above mentioned dependencies, a POSIX shell, a C compiler and make.
->           Any contributing to Finit can generate it from `configure.ac` using
->           the `autogen.sh` script.
+> [!NOTE]
+> The configure script is not available in the GIT sources.  It is
+> however included in (officially supported) released tarballs.  The
+> idea is that you should not need GNU autotools to build, only the
+> above mentioned dependencies, a POSIX shell, a C compiler and make.
+> Any contributing to Finit can generate it from `configure.ac` using
+> the `autogen.sh` script.
 
 
 Example
@@ -115,10 +117,11 @@ Linux config to:
 
     CONFIG_UEVENT_HELPER_PATH="/sbin/mdev"
 
-**Note:** If you run into problems starting Finit, take a look at
-  `finit.c`.  One of the most common problems is a custom Linux kernel
-  build that lack `CONFIG_DEVTMPFS`.  Another is too much cruft in the
-  system `/etc/fstab`.
+> [!TIP]
+> If you run into problems starting Finit, take a look at `finit.c`.
+> One of the most common problems is a custom Linux kernel build that
+> lack `CONFIG_DEVTMPFS`.  Another is too much cruft in the system
+> `/etc/fstab`.
 
 
 Running
@@ -157,8 +160,10 @@ enabled.  The default Finit rescue mode configuration is installed into
 
 By default the a root shell, without login, is started.
 
-> **Note:** in this mode `initctl` will not work.  Use the `-f` flag to
-> force `reboot`, `shutdown`, or `poweroff`.
+> [!IMPORTANT]
+> In rescue mode `initctl` will not work, the same goes for `reboot`,
+> `shutdown`, and `poweroff` commands, provided they are the Finit
+> versions of these commands.  Use the `-f` flag to force the action.
 
 
 Debugging
@@ -187,8 +192,9 @@ kernel usually reboots: `configure --enable-emergency-shell`.  However,
 the behavior of Finit is severely limited when this is enabled, so use
 it only for debugging start up issues when Finit crashes.
 
-**NOTE:** Neither of these options should be enabled on production
-         systems since they can potentially give a user root access.
+> [!CAUTION]
+> None of these options should be enabled on production systems since
+> they can potentially give a user root access.
 
 
 [1]:       ftp://troglobit.com/finit/finit-4.3.tar.gz

doc/cmdline.md

@@ -4,10 +4,11 @@ Tips & Tricks with the kernel cmdline
 This document summarizes the different boot parameters that can be
 passed on the Linux kernel command line.  Not limited to Finit.
 
-The `bool` setting is one of `on, off, true false, 1, 0`.
+> [!IMPORTANT]
+> Remember to use `--` to separate kernel parameters from parameters to
+> init.  E.g., `init=/sbin/finit -- finit.debug rescue`
 
-> **NOTE:** remember to use `--` to separate kernel parameters from
-> parameters to init.  E.g., `init=/sbin/finit -- finit.debug rescue`
+The `bool` setting is one of `on, off, true false, 1, 0`.
 
 * `debug`: Enable kernel debug.  Debug messages are printed to the
    console until Finit starts up, unless `loglevel=7` (below) is used.
@@ -30,7 +31,8 @@ The `bool` setting is one of `on, off, true false, 1, 0`.
   Very useful for selecting different boot modes, e.g. manufacturing test,
   firmware upgrade, or rescue mode.
 
-  > Note: `<boot/...>` conditions cannot be cleared with `initctl`!
+> [!NOTE]
+> `<boot/...>` conditions cannot be cleared with `initctl`!
 
 * `finit.config=/path/to/alternative/finit.conf`: override the
   compile-time bootstrap config file, default:

doc/conditions.md

@@ -47,9 +47,10 @@ both the `pid/setupd` *and* `pid/zebra` conditions are satisfied.  A
 `pid/` condition is satisfied by the corresponding service's PID file
 being created, i.e., the service's default readiness notification.
 
-**NOTE:** Conditions also stop services when a condition is no longer
-  asserted.  I.e., if the Zebra process above stops or restarts, netd
-  will also stop or restart.
+> [!IMPORTANT]
+> Conditions also stop services when a condition is no longer asserted.
+> I.e., if the `zebra` process above stops or restarts, `netd` will also
+> stop or restart.
 
 Another example is `dropbear`, it does not support `SIGHUP`, but we can
 also see optional sourcing of arguments from an environment file:
@@ -106,11 +107,16 @@ on this, see [Internals](#internals).)  Thus, after a reconfiguration it
 is up to the "owner" of the condition to convey the new (or possibly
 unchanged) state of it.
 
-> **Note:** For `pid/` conditions it is expected that services "touch"
->           or recreate their PID file on `SIGHUP`.
-
 Static (one-shot) conditions, like `usr/`, never enter the `flux` state.
 
+> [!IMPORTANT]
+> For `pid/` conditions it is expected that the service reassert, i.e.,
+> "touch" or recreate, their PID file on `SIGHUP`.  This can be done by
+> calling `utimensat()` on the PID file.  Provided, of course, that the
+> service supports reloading on `SIGHUP`, otherwise it will be restarted
+> by Finit when they instead exit on the signal.  For such services, use
+> `<!>` to tell Finit the service does not support `SIGHUP`.
+
 
 Built-in Conditions
 -------------------
@@ -171,9 +177,10 @@ Built-in conditions:
 - `boot/arg`
 - `dev/node` and `dev/dir/node`
 
-**Note:** `up` means administratively up, the interface flag `IFF_UP`.
-  `running` is the `IFF_RUNNING` flag, meaning operatively up.  The
-  difference is that `running` tells if the NIC has link.
+> [!NOTE]
+> Here, `up` means administratively up, the interface flag `IFF_UP`.
+> `running` is the `IFF_RUNNING` flag, meaning operatively up.  The
+> difference is that `running` tells if the NIC has link.
 
 
 Composition
@@ -205,8 +212,9 @@ the service stanza for the daemon.  It is far from optimal since any
 synchronization of depending services may fail due to the daemon not
 having reinitialized/created their IPC sockets, or similar.
 
-> **Note:** in versions of Finit prior to v4, the PID conditions were
->           called 'svc' conditions, and they were far more complex.
+> [!NOTE]
+> In versions of Finit prior to v4, the PID conditions were called 'svc'
+> conditions, and they were far more complex.
 
 
 Debugging

doc/config.md

@@ -18,8 +18,8 @@ Finit Configuration
   * [Alternate rcS.d/](#alternate-rcs-d-)
   * [Resource Limits](#resource-limits)
   * [Runlevels](#runlevels)
-  * [One-shot Commands (sequence)](#one-shot-commands--sequence-)
-  * [One-shot Commands (parallel)](#one-shot-commands--parallel-)
+  * [One-shot Commands (sequence)](#one-shot-commands-sequence)
+  * [One-shot Commands (parallel)](#one-shot-commands-parallel)
   * [SysV Init Scripts](#sysv-init-scripts)
   * [Services](#services)
   * [Run-parts Scripts](#run-parts-scripts)
@@ -46,8 +46,9 @@ Plugins start at [hook points](plugins.md#hooks) and can run various set
 up, or install event handlers that later provide runtime services, e.g.,
 PID file monitoring, or [conditions](conditions.md).
 
-> **Tip:** see [SysV Init Compatibility](#sysv-init-compatibility) for
-> help to quickly get going with an existing SysV or BusyBox init setup.
+> [!TIP]
+> See [SysV Init Compatibility](#sysv-init-compatibility) for help to
+> quickly get going with an existing SysV or BusyBox init setup.
 
 
 ### Configuration Files
@@ -127,8 +128,9 @@ properties of the `run` statement.  For an example on the relation of
 `service` and `run` statements, and dependency handling between them,
 see [Conditional Loading](#conditional-loading), below.
 
-> **Note:** the `finit.conf` and `finit.d/` names are only defaults.
-> They can be changed at compile-time with two `configure` options:
+> [!NOTE]
+> The names `finit.conf` and `finit.d/` are only defaults.  They can be
+> changed at compile-time with two `configure` options:
 > `--with-config=/etc/foo.conf` and `--with-rcsd=/var/foo.d`.
 >
 > They can also be overridden from the [kernel command line](cmdline.md)
@@ -187,8 +189,9 @@ above, provided their respective mount point exists.
 
 With all filesystems mounted, Finit calls `swapon`.
 
-> **Tip:** to see what happens when all filesystems are mounted, have a
-> look at the [`bootmisc.so` plugin](plugins.md).
+> [!TIP]
+> To see what happens when all filesystems are mounted, have a look at
+> the [`bootmisc.so` plugin](plugins.md).
 
 At shutdown, and after having stopped all services and other lingering
 processes have been killed, filesystems are unmounted in the reverse
@@ -198,6 +201,18 @@ order, and `swapoff` is called.
 
 ### Runlevels
 
+Finit supports runlevels, but unlike other init systems runlevels are
+declared per service/run/task/sysv command.  When booting up a system
+Finit pass through three phases:
+
+ 1. Setting up the console, parsing any command line options, and other
+    housekeeping tasks like mounting all filesystems, and calling `fsck`
+ 2. Starting all run/task/services in runlevel S, then waiting for all
+    services to have started, and all run/tasks to have completed
+ 3. Go to runlevel 2, or whatever the user has set in the configuration
+
+Available runlevels:
+
  - `  S`: bootStrap
  - `  1`: Single user mode
  - `2-5`: traditional multi-user mode
@@ -205,37 +220,40 @@ order, and `swapoff` is called.
  - `7-9`: multi-user mode (extra)
  - `  0`: shutdown
 
-Runlevels are declared per service/run/task/sysv command.  Starting in
-runlevel S (bootStrap), usually only for tasks supposed to run once at
-boot, and services like `syslogd`, which you need to start and run
-throughout the whole time your system is up.
+Runlevel S (bootStrap), is for tasks supposed to run once at boot, and
+services like `syslogd`, which need to start early and run throughout
+the lifetime of your system.
 
-Before `S` is started, however, Finit performs a lot of housekeeping
-tasks like mounting all filesystems, calling `fsck` if needed, and
-making sure the everything is OK.
+Example:
 
     task [S] /lib/console-setup/console-setup.sh
     service [S12345] env:-/etc/default/rsyslog rsyslogd -n $RSYSLOGD_ARGS
 
-After runlevel S, Finit proceeds to runlevel 2.  This can be changed in
-`/etc/finit.conf` using the `runlevel N` directive, or by a script
-running in runlevel S that calls, e.g., `initctl runlevel 9`.  The
-latter is useful if startup scripts detect problems outside of Finit's
-control, e.g., critical services/devices missing or hardware problems.
+When bootstrap has completed, Finit moves to runlevel 2.  This can be
+changed in `/etc/finit.conf` using the `runlevel N` directive, or by a
+script running in runlevel S that calls, e.g., `initctl runlevel 9`.
+The latter is useful if startup scripts detect problems outside of
+Finit's control, e.g., critical services/devices missing or hardware
+problems.
 
-Before starting the services in runlevel 2, Finit first stops everything
-that is not allowed to run in 2, and then brings up networking.
-Networking is expected to be available in all runlevels except: S, 1
-(single user level), 6, and 0.  Networking is enabled either by the
-`network script` directive, or if you have an `/etc/network/interfaces`
-file, Finit calls `ifup -a` -- at the very least the loopback interface
-is brought up.
+Each runlevel must be allowed to "complete".  Meaning, all services in
+runlevel S must have started and all run/tasks have been started and
+collected (exited).  Finit waits 120 seconds for all run/tasks in S to
+complete before proceeding to 2.
 
-> **Note:** when moving from runlevel S to 2, all run/task/services that
-> were constrained to runlevel S only are dropped from bookkeeping.  So
-> when reaching the prompt, `initctl` will not show these run/tasks.
-> This is a safety mechanism to prevent bootstrap-only tasks from
-> accidentally being run again.  E.g., `console-setup.sh` above.
+Finit first stops everything that is not allowed to run in 2, and then
+brings up networking.  Networking is expected to be available in all
+runlevels except: S, 1 (single user level), 6, and 0.  Networking is
+enabled either by the `network script` directive, or if you have an
+`/etc/network/interfaces` file, Finit calls `ifup -a` -- at the very
+least the loopback interface is brought up.
+
+> [!NOTE]
+> When moving from runlevel S to 2, all run/task/services that were
+> constrained to runlevel S only are dropped from bookkeeping.  So when
+> reaching the prompt, `initctl` will not show these run/tasks.  This is
+> a safety mechanism to prevent bootstrap-only tasks from accidentally
+> being run again.  E.g., `console-setup.sh` above.
 
 
 ### Managing Services
@@ -322,7 +340,8 @@ the `ps` command we can see that the process is started with:
 
     foo -n --extra-arg=bar -s -x
 
-> **Note:** the leading `-` determines if Finit should treat a missing
+> [!NOTE]
+> The leading `-` in `env:` determines if Finit should treat a missing
 > environment file as blocking the start of the service or not.  When
 > `-` is used, a missing environment file does *not* block the start.
 
@@ -384,12 +403,13 @@ To synchronize two services the following condition can be used:
 
 For details on the syntax and options, see below.
 
-> **Note:** on `initctl reload` conditions are normally set in "flux",
-> while figuring out which to stop, start or restart.  Services that
-> need to be restarted have their `ready` condition removed prior to
-> Finit sending them SIGHUP (if they support that), or stop-starting
-> them.  A daemon is expected to reassert its readiness, e.g. systemd
-> style daemons to write `READY=1\n`.
+> [!NOTE]
+> On `initctl reload` conditions are set in "flux", while figuring out
+> which to stop, start or restart.  Services that need to be restarted
+> have their `ready` condition removed prior to Finit sending them
+> SIGHUP (if they support that), or stop-starting them.  A daemon is
+> expected to reassert its readiness, e.g. systemd style daemons to
+> write `READY=1\n`.
 >
 > However, the s6 notify mode does not support this because in s6 you
 > are expected to close your notify descriptor after having written
@@ -429,9 +449,10 @@ example employs a wrapper script in `/etc/start.d`.
         # Execute the program
         exec /usr/bin/program $OPTIONS
 
-> **Note:** the example sets `<!>` to denote that it doesn't support
->           `SIGHUP`.  That way Finit will stop/start the service
->           instead of sending SIGHUP at restart/reload events.
+> [!NOTE]
+> The example sets `<!>` to denote that it doesn't support `SIGHUP`.
+> That way Finit will stop/start the service instead of sending SIGHUP
+> at restart/reload events.
 
 
 Templating
@@ -514,6 +535,7 @@ A daemon using `SCHED_RR` currently need to run outside the default cgroups.
 
     service [...] <...> cgroup.root /path/to/daemon arg -- Real-Time process
 
+> [!NOTE]
 > Linux cgroups and details surrounding values are not explained in the
 > Finit documentation.  The Linux admin-guide cover this well:
 > <https://www.kernel.org/doc/html/latest/admin-guide/cgroup-v2.html>
@@ -539,12 +561,9 @@ service name:sysklogd [S123456789]   \
 ```
 
 The .conf files `/etc/finit.conf` and `/etc/finit.d/*` support the
-following directives:
-
-
-> **Note:** not all directives are available in `/etc/finit.d/*.conf`
-> and some directives are only available at bootstrap, runlevel `S`,
-> see [Limitations](#limitations) below for details.
+following directives.  Some directives have restrictions, e.g., only
+available at bootstrap, runlevel `S`, see [Limitations](#limitations)
+below for details.
 
 
 ### Hostname
@@ -556,7 +575,8 @@ the contents of that file is used.
 
 Deprecated.  We recommend using `/etc/hostname` instead.
 
-> **Note:** only read and executed in runlevel S (bootstrap).
+> [!NOTE]
+> Only read and executed in runlevel S (bootstrap).
 
 
 ### Kernel Modules
@@ -574,7 +594,8 @@ BusyBox mdev tool, add to `/etc/mdev.conf`:
 
     $MODALIAS=.*  root:root       0660    @modprobe -b "$MODALIAS"
 
-> **Note:** only read and executed in runlevel S (bootstrap).
+> [!NOTE]
+> Only read and executed in runlevel S (bootstrap).
 
 
 ### Networking
@@ -587,7 +608,8 @@ Deprecated.  We recommend using dedicated task/run stanzas per runlevel,
 or `/etc/network/interfaces` if you have a system with `ifupdown`, like
 Debian, Ubuntu, Linux Mint, or an embedded BusyBox system.
 
-> **Note:** only read and executed in runlevel S (bootstrap).
+> [!NOTE]
+> Only read and executed in runlevel S (bootstrap).
 
 
 ### Alternate finit.d/
@@ -608,8 +630,9 @@ another set of .conf files, e.g.:
 
     rcsd /etc/factory.d
 
-> **Note:** this directive is only available from the top-level
-> bootstrap .conf file.
+> [!NOTE]
+> This directive is only available from the top-level bootstrap .conf
+> file, usually `/etc/finit.conf`.
 
 ### Resource Limits
 
@@ -642,16 +665,18 @@ stanzas can share the same rlimits if they are in the same .conf.
 
 **Syntax:** `runlevel <N>`
 
-The system runlevel to go to after bootstrap (S) has completed.  `N`
-is the runlevel number 0-9, where 6 is reserved for reboot and 0 for
-halt.
+The system runlevel to go to after bootstrap (S) has completed.  `N` is
+the runlevel number 0-9, where 6 is reserved for reboot and 0 for halt.
+Completed in this context means all services have been started and all
+run/tasks have been started and collected.
 
 It is recommended to keep runlevel 1 as single-user mode, because
 Finit disables networking in this mode.
 
 *Default:* 2
 
-> **Note:** only read and executed in runlevel S (bootstrap).
+> [!NOTE]
+> Only read and executed in runlevel S (bootstrap).
 
 
 ### One-shot Commands (sequence)
@@ -665,10 +690,11 @@ optional arguments and description.  `run` commands are guaranteed to be
 completed before running the next command.  Useful when serialization is
 required.
 
-> **Warning:** try to avoid the `run` command.  It blocks much of the
-> functionality in Finit, like (re)starting other (perhaps crashing)
-> services while a `run` task is executing.  Use other synchronization
-> mechanisms instead, like conditions.
+> [!WARNING]
+> Try to avoid the `run` command.  It blocks much of the functionality
+> in Finit, like (re)starting other (perhaps crashing) services while a
+> `run` task is executing.  Use other synchronization mechanisms
+> instead, like conditions.
 
 Incomplete list of unsupported `initctl` commands in `run` tasks:
 
@@ -708,7 +734,7 @@ non-trivial constructs are better placed in a separate shell script.
 
 Similar to `task` is the `sysv` stanza, which can be used to call SysV
 style scripts.  The primary intention for this command is to be able to
-re-use much of existing setup and init scripts in Linux distributions.
+reuse much of existing setup and init scripts in Linux distributions.
   
 When entering an allowed runlevel, Finit calls `init-script start`, when
 entering a disallowed runlevel, Finit calls `init-script stop`, and if
@@ -723,7 +749,8 @@ file, but rather watch that file for the resulting forked-off PID.  This
 syntax also works for forking daemons that do not have a command line
 option to run it in the foreground, more on this below in `service`.
 
-> **Tip:** see also [SysV Init Compatibility](#sysv-init-compatibility).
+> [!TIP]
+> See also [SysV Init Compatibility](#sysv-init-compatibility).
 
 
 ### Services
@@ -731,13 +758,13 @@ option to run it in the foreground, more on this below in `service`.
 **Syntax:** `service [LVLS] <COND> /path/to/daemon ARGS -- Optional description`
 
 Service, or daemon, to be monitored and automatically restarted if it
-exits prematurely.  Finit tries to restart services that die, by
-default 10 times before giving up and marking them as *crashed*.
-After which they have to be manually restarted with `initctl restart
-NAME`.  The limits controling this are configurable, see the options
-below.
+exits prematurely.  Finit tries to restart services that die, by default
+10 times before giving up and marking them as *crashed*.  After which
+they have to be manually restarted with `initctl restart NAME`.  The
+limits controlling this are configurable, see the options below.
 
-> **Tip:** to allow endless restarts, see below option `respawn`
+> [!TIP]
+> To allow endless restarts, see below option `respawn`
   
 For daemons that support it, we recommend appending `--foreground`,
 `--no-background`, `-n`, `-F`, or similar command line argument to
@@ -899,32 +926,39 @@ e.g., `halt:SIGPWR`.  To change the delay between the stop signal and
 KILL, use the option `kill:<1-60>`, e.g., `kill:10` to wait 10 seconds
 before sending `SIGKILL`.
 
-Services, including the `sysv` variant, support pre/post/ready scripts:
+Services, including the `sysv` variant, support pre/post/ready and
+cleanup scripts:
 
-  * `pre:[0-3600,]script`
-  * `post:[0-3600,]script`
-  * `ready:[0-3600,]script`
+  * `pre:[0-3600,]script` -- called before the sysv/service is stated
+  * `post:[0-3600,]script` -- called after the sysv/service has stopped
+  * `ready:[0-3600,]script` -- called when the sysv/service is ready
+  * `cleanup:[0-3600,]script` -- called when run/task/sysv/service is removed
 
 The optional number (0-3600) is the timeout before Finit kills the
 script, it defaults to the kill delay value and can be disabled by
 setting it to zero.  These scripts run as the same `@USER:GROUP` as the
-service itself, with any `env:file` sourced.  The scripts must use an
-absolute path, but are executed from the `$HOME` of the given user.  The
-scripts are not called with any argument, but get a set of environment
-variables:
+service itself, with any `env:file` sourced.  The scripts are executed
+from the `$HOME` of the given user.  The scripts are not called with any
+argument, but get a set of environment variables:
 
   * `SERVICE_IDENT=foo:1`
   * `SERVICE_NAME=foo`
   * `SERVICE_ID=1`
 
 The `post:script` is called with an additional set of environment
-variables:
+variables.  Yes, the text is correct, the naming was an accident:
 
- - `EXIT_CODE=[exited,signal]`: set to one of `exited` or `signal`
+ - `EXIT_CODE=[exited,signal,crashed]`: normal exit, signaled, or
+   crashed
  - `EXIT_STATUS=[num,SIGNAME]`: set to one of exit status code from
    the program, if it exited normally, or the signal name (`HUP`,
    `TERM`, etc.) if it exited due to signal
 
+When a run/task/sys/service is removed (disable + reload) it is first
+stopped and then removed from the runlevel.  The `post:script` always
+runs when the process has stopped, and the `cleanup:script` runs when
+the the stanza has been removed from the runlevel.
+
 > [!IMPORTANT]
 > These script actions are intended for setup, cleanup, and readiness
 > notification.  It is up to the user to ensure the scripts terminate.
@@ -1027,8 +1061,9 @@ before continuing.  None of them can issue commands to start, stop, or
 restart other services.  Also, ensure all your services and programs
 either terminate or start in the background or you will block Finit.
 
-> **Note:** `runparts` scrips are only read and executed in runlevel S.
-> See [hook scripts](plugins.md#hooks) for other ways to run scripts at
+> [!NOTE]
+> `runparts` scripts are only read and executed in runlevel S.  See
+> [hook scripts](plugins.md#hooks) for other ways to run scripts at
 > certain points during the complete lifetime of the system.
 
 **Recommendations:**
@@ -1301,8 +1336,9 @@ This rescue mode can be disabled at configure time using:
 
 The rescue mode comes in two flavors; *traditional* and *fallback*.
 
-> **Note:** in this mode `initctl` will not work.  Use the `-f` flag to
-> force `reboot`, `shutdown`, or `poweroff`.
+> [!NOTE]
+> In this mode `initctl` will not work.  Use the `-f` flag to force
+> `reboot`, `shutdown`, or `poweroff`.
 
 
 ### Traditional
@@ -1314,11 +1350,12 @@ program is used instead.  If a successful login is made, or if the user
 exits (Ctrl-D), the rescue mode is ended and the system boots up
 normally.
 
-> **Note:** the bundled sulogin in Finit can at configure time be given
-> another user than the default (root).  If the sulogin user does not
-> have a password, or __the account is locked__, the user is presented
-> with a password-less `"Press enter to enter maintenance mode."`,
-> prompt which opens up a root shell.
+> [!WARNING]
+> The bundled sulogin in Finit can at configure time be given another
+> user than the default (root).  If the sulogin user does not have a
+> password, or __the account is locked__, the user is presented with a
+> prompt: `"Press enter to enter maintenance mode."`, which will open
+> up a root shell *without prompting for password*!
 
 
 ### Fallback
@@ -1366,10 +1403,11 @@ These can now also be set in any `.conf` file in `/etc/finit.d`.
 There is, however, nothing preventing you from having all configuration
 settings in `/etc/finit.conf`.
 
-> **Note:** The `/etc/finit.d` directory was previously the default
->          Finit [runparts](#run-parts-scripts) directory.  Finit >=v4.0
->          no longer has a default `runparts` directory, make sure to
->          update your setup, or the finit configuration, accordingly.
+> [!IMPORTANT]
+> The default `rcsd`, i.e., `/etc/finit.d`, was previously the Finit
+> [runparts](#run-parts-scripts) directory.  Finit >=v4.0 no longer has
+> a default `runparts` directory, make sure to update your setup, or the
+> finit configuration, accordingly.
 
 
 Watchdog

doc/distro.md

@@ -34,8 +34,9 @@ build time:
 ./configure --with-rcsd=/etc/init.d --with-config=/etc/init.d/init.conf
 ```
 
-> **Note:** remember `--prefix` et al as well, the default is likely
->           *not* what you want.  See the [build docs][1] for details.
+> [!IMPORTANT]
+> Remember `--prefix` et al as well, the default is likely *not* what
+> you want.  See the [build docs][1] for details.
 
 The resulting directory structure is depicted below.  Please notice how
 `/etc/finit.conf` now resides in the same sub-directory as a non-symlink

doc/plugins.md

@@ -35,9 +35,10 @@ For your convenience a set of *optional* plugins are available:
   files distributed with Finit.  It is read first but can be overridden
   by any of the standard tmpfiles.d directories, e.g. `/etc/tmpfiles.d`.
 
-  > **Note:** On an embedded system both `/var` and `/run` can be `tmpfs`
-  > RAM disks and `/dev` is usually a `devtmpfs`.  This must be defined
-  > in the `/etc/fstab` file and in the Linux kernel config.
+> [!NOTE]
+> On an embedded system both `/var` and `/run` can be `tmpfs` RAM
+> disks and `/dev` is usually a `devtmpfs`.  This must be defined in
+> the `/etc/fstab` file and in the Linux kernel config.
 
 * *dbus.so*: Setup and start system message bus, D-Bus, at boot.
   _Optional plugin._
@@ -63,8 +64,9 @@ For your convenience a set of *optional* plugins are available:
 
   Enabled by default.
   
-  > See the [Services](config.md#services) section in the configuration
-  > guide for an example how to run `mdevd`, alternative to plain mdev.
+> [!TIP]
+> See the [Services](config.md#services) section in the configuration
+> guide for an example how to run `mdevd`, alternative to plain mdev.
 
 * *rtc.so*: Restore and save system clock from/to RTC on boot/halt.
   Enabled by default.
@@ -89,16 +91,17 @@ For your convenience a set of *optional* plugins are available:
 
         set index 1234
 
-  **Note:** unlike the traditional .conf `module` directive, which load
-  any listed module immediately, this plugin creates a background `task`
-  which load the module(s) in the background.  The program is modprobe,
-  `/sbin/modprobe`, which you can override per .conf file:
+  Since these tasks run in the background, they return `[ OK ]` at boot,
+  unless the modprobe tool does not exist.  Check syslog for warnings
+  and the actual status of the operation using `initctl`.
 
-        set modprobe /path/to/maybe-a-modprobe-wrapper
-
-  Since these tasks run in the background, they usually return `[ OK ]`
-  at boot, unless the modprobe tool does not exist.  Check syslog for
-  warnings and the actual status of the operation using `initctl`.
+> [!IMPORTANT]
+> Unlike the traditional .conf `module` directive, which load any listed
+> module immediately, this plugin creates a background `task` which load
+> the module(s) in the background.  The program is modprobe,
+> `/sbin/modprobe`, which you can override per .conf file:
+> 
+>     set modprobe /path/to/maybe-a-modprobe-wrapper
 
 * *netlink.so*: Listens to Linux kernel Netlink events for gateway and
   interfaces.  These events are then sent to the Finit service monitor
@@ -158,8 +161,10 @@ hook points:
     EOF
     $ chmod +x /libexec/finit/hook/sys/down/foo.sh
 
-> **Note:** to use hook scripts, even for pre-bootstrap and pre-shutdown
-> tasks, you must build with `configure --enable-hook-scripts-plugin`.
+> [!IMPORTANT]
+> To use hook scripts, even for pre-bootstrap and pre-shutdown tasks,
+> you must build with `configure --enable-hook-scripts-plugin`.
+
 
 ### Bootstrap Hooks
 

doc/service.md

@@ -22,8 +22,9 @@ Finit can *not* start and monitor a daemon that:
 | ✔ | No      | No               | Yes, optionally        |
 | ✘ | Yes     | No               | No                     |
 
-> **Note:** PID files is one mechanism used to assert conditions to
-> synchronize the start and stop of other, dependent, services.
+> [!NOTE]
+> PID files is one mechanism used to assert conditions to synchronize
+> the start and stop of other, dependent, services.
 
 ### Forks to Background w/ PID File
 

man/finit.conf.5

@@ -175,13 +175,13 @@ task [s] echo "foo" | cat >/tmp/bar
 .Ed
 .It Cm sysv Oo LVLS Oc Ao COND Ac Ar /path/to/script ARGS Op -- Optional description
 Similar to
-.Cm task
+.Cm service
 is the
 .Cm sysv
 stanza, which can be used to call SysV style start/stop scripts.  The
-primary intention for this command is to be able to re-use much of
-existing setup and init scripts in Linux distributions.
-.Pp  
+primary intention for this command is to be able to re-use traditional
+SysV init scripts in Linux distributions.
+.Pp
 When entering an allowed runlevel, Finit calls
 .Cm init-script start ,
 when entering a disallowed runlevel, Finit calls
@@ -434,30 +434,42 @@ e.g.,
 .Cm kill:10
 to wait 10 seconds before sending SIGKILL.
 .Pp
-Services support the
-.Cm pre:script
-and
-.Cm post:script
-command actions as well.  These run as the same
+Services, including the
+.Cm sysv
+variant, support pre/post/ready and cleanup scripts:
+.Bl -tag -width 1n
+.It Cm pre:[0-3600,]script
+called before the sysv/service is stated
+.It Cm post:[0-3600,]script
+called after the sysv/service has stopped
+.It Cm ready:[0-3600,]script
+called when the sysv/service is ready
+.It Cm cleanup:[0-3600,]script
+called when run/task/sysv/service is removed
+.El
+.Pp
+The optional number (0-3600) is the timeout before Finit kills the
+script, it defaults to the kill delay value and can be disabled by
+setting it to zero.  These scripts run as the same
 .Cm @USER:GROUP
 as the service itself, with any
 .Cm env:file
 sourced.  The scripts must use an absolute path, but are executed from
 the
 .Cm $HOME
-of the given user.  The scripts are not called with any argument
-(currently), but both get the
-.Cm SERVICE_IDENT=foo
-environment variable set.  Here
-.Cm foo
-denotes the identity of the service, which if there are multiple
-services named
-.Cm foo ,
-may be
-.Cm foo:1 ,
-or any unique identifier specified in the .conf file.  The
+of the given user.  The scripts are not called with any argument, but
+both get a set of environment variables:
+.Pp
+.Bl -tag -width 1n -compact
+.It Cm SERVICE_IDENT=foo:1
+.It Cm SERVICE_NAME=foo
+.It Cm SERVICE_ID=1
+.El
+.Pp
+The
 .Cm post:script
-is called with an additional set of environment variables:
+is called with an additional set of environment variables, and yes the
+naming looks like it should be swapped:
 .Bl -tag -offset indent -width 1n
 .It Cm EXIT_CODE=[exited,signal,crashed]
 set to one of
@@ -471,10 +483,12 @@ set to one of exit status code from the program, if it exited normally,
 or the signal name (HUP, TERM, etc.) if it exited due to signal
 .El
 .Pp
-The scripts have a default execution time of 3 seconds before they are
-SIGKILLed, this can be adjusted using the
-.Cm kill:SEC
-modifier syntax.
+When a run/task/sys/service is removed (disable + reload) it is first
+stopped and then removed from the runlevel.  The
+.Cm post:script
+always runs when the process has stopped, and the
+.Cm cleanup:script
+runs when the the stanza has been removed from the runlevel.
 .It Cm runparts Aq DIR
 Call
 .Xr run-parts 8

plugins/modules-load.c

@@ -41,7 +41,7 @@
  *
  * Please note, the :ID is there for your benefit, it ensures that tasks
  * in Finit are unique.  If you have two tasks with the same name and ID
- * (or no ID), the last one read replaces any preceeding one!
+ * (or no ID), the last one read replaces any preceding one!
  */
 
 #include <unistd.h>

plugins/pidfile.c

@@ -80,6 +80,23 @@ static void pidfile_update_conds(char *dir, char *name, uint32_t mask)
 		return;
 	}
 
+	switch (svc->state) {
+	case SVC_TEARDOWN_STATE:
+	case SVC_CLEANUP_STATE:
+	case SVC_SETUP_STATE:
+		/*
+		 * E.g., a setup script for a container may run in lieu
+		 * of crun, trying to fetch a container image from the
+		 * network.  This script may want to advertise its PID
+		 * to the system for clean and swift reboots.
+		 */
+		dbg("%s[%d]: ignoring any PID file changes in CLEANUP/TEARDOWN/SETUP",
+		    svc_ident(svc, NULL, 0), svc->pid);
+		return;
+	default:
+		break;
+	}
+
 	/* Service condistion to set/clear */
 	mkcond(svc, cond, sizeof(cond));
 

plugins/procps.c

@@ -32,6 +32,10 @@
 #include "helpers.h"
 #include "plugin.h"
 
+#ifndef SYSCTL_PATH
+#define SYSCTL_PATH "/sbin/sysctl"
+#endif
+
 static void setup(void *arg)
 {
 	glob_t gl;
@@ -54,7 +58,7 @@ static void setup(void *arg)
 		for (i = 0; i < gl.gl_pathc; i++) {
 			char cmd[160];
 
-			snprintf(cmd, sizeof(cmd), "/sbin/sysctl -e -p %s >/dev/null", gl.gl_pathv[i]);
+			snprintf(cmd, sizeof(cmd), "%s -e -p %s >/dev/null", SYSCTL_PATH, gl.gl_pathv[i]);
 			run(cmd, "sysctl");
 		}
 		globfree(&gl);

plugins/rtc.c

@@ -338,6 +338,7 @@ PLUGIN_INIT(plugin_init)
 
 PLUGIN_EXIT(plugin_exit)
 {
+	uev_timer_stop(&rtc_timer);
 	plugin_unregister(&plugin);
 }
 

plugins/urandom.c

@@ -26,6 +26,7 @@
 #include <sys/stat.h>
 #include <sys/time.h>		/* gettimeofday() */
 #include <sys/types.h>
+#include <sys/resource.h>	/* getrusage() */
 #ifdef _LIBITE_LITE
 # include <libite/lite.h>
 #else
@@ -43,13 +44,27 @@
 #endif
 
 #ifdef RANDOMSEED
+/*
+ * This is the fallback seed function, please make sure you have drivers
+ * enabling /dev/hwrng instead.
+ */
 static void fallback(FILE *fp)
 {
+	unsigned long seed;
 	struct timeval tv;
+	struct rusage ru;
 	int iter = 128;
 
 	gettimeofday(&tv, NULL);
-	srandom(tv.tv_sec % 3600);
+	getrusage(RUSAGE_SELF, &ru);
+
+	/* Mix multiple sources of "randomness" */
+	seed = tv.tv_sec ^ (tv.tv_usec << 16);
+	seed ^= ru.ru_utime.tv_sec ^ (ru.ru_utime.tv_usec << 16);
+	seed ^= ru.ru_stime.tv_sec ^ (ru.ru_stime.tv_usec << 16);
+	seed ^= getpid() << 8;
+
+	srandom(seed);
 	while (iter--) {
 		uint32_t i, prng = random();
 

src/api.c

@@ -332,6 +332,11 @@ static void api_cb(uev_t *w, void *arg, int events)
 	int sd, lvl;
 	svc_t *svc;
 
+	if (UEV_ERROR == events) {
+		dbg("%s(): api socket %d invalid.", __func__, w->fd);
+		goto error;
+	}
+
 	sd = accept(w->fd, NULL, NULL);
 	if (sd < 0) {
 		err(1, "Failed serving API request");
@@ -557,7 +562,7 @@ static void api_cb(uev_t *w, void *arg, int events)
 			}
 
 			if (uev_io_init(ctx, &svc->notify_watcher, service_notify_cb, svc, sd, UEV_READ)) {
-				err(1, "Falied initializing %s readiness notifier", svc_ident(svc, NULL, 0));
+				err(1, "Failed initializing %s readiness notifier", svc_ident(svc, NULL, 0));
 				break;
 			}
 			return;	/* Don't close sd, used for notify */
@@ -578,8 +583,6 @@ static void api_cb(uev_t *w, void *arg, int events)
 
 leave:
 	close(sd);
-	if (UEV_ERROR == events)
-		goto error;
 	return;
 error:
 	api_exit();

src/cgroup.c

@@ -259,6 +259,11 @@ static void cgroup_events_cb(uev_t *w, void *arg, int events)
 	ssize_t sz;
 	size_t off;
 
+	if (UEV_ERROR == events) {
+		dbg("%s(): inotify socket %d invalid.", __func__, w->fd);
+		return;
+	}
+
 	sz = read(w->fd, ev_buf, sizeof(ev_buf) - 1);
 	if (sz <= 0) {
 		err(1, "invalid inotify event");
@@ -471,7 +476,7 @@ void cgroup_init(uev_ctx_t *ctx)
 	if (mount("none", FINIT_CGPATH, "cgroup2", opts, NULL)) {
 		if (errno == ENOENT)
 			logit(LOG_INFO, "Kernel does not support cgroups v2, disabling.");
-		else if (errno == EPERM) /* Probably inside an unpriviliged container */
+		else if (errno == EPERM) /* Probably inside an unprivileged container */
 			logit(LOG_INFO, "Not allowed to mount cgroups v2, disabling.");
 		else
 			err(1, "Failed mounting cgroup v2");

src/cond-w.c

@@ -113,7 +113,7 @@ static void cond_boot_strap(void)
  *
  *    => 'pid/' + 'teamd' + '' = condition <pid/teamd>
  *
- * When you add a second aggrate you need to tell Finit it's a different
+ * When you add a second aggregate you need to tell Finit it's a different
  * instance usugin the :ID syntax:
  *
  *    service :lag2 teamd -f /etc/teamd/lag2.conf -- Aggregate lag2

src/conf.c

@@ -1542,6 +1542,11 @@ static int conf_iwatch_read(int fd)
 
 static void conf_cb(uev_t *w, void *arg, int events)
 {
+	if (UEV_ERROR == events) {
+		dbg("%s(): iwatch socket %d invalid.", __func__, w->fd);
+		return;
+	}
+
 	if (conf_iwatch_read(w->fd)) {
 		err(1, "invalid inotify event");
 		return;
@@ -1574,25 +1579,26 @@ int conf_monitor(void)
 	 * have or not have symlinks in place.  We need to monitor for
 	 * changes to either symlink or target.
 	 */
-	rc += iwatch_add(&iw_conf, finit_rcsd, IN_ONLYDIR);
+	rc |= iwatch_add(&iw_conf, finit_rcsd, IN_ONLYDIR);
 	snprintf(path, sizeof(path), "%s/available/", finit_rcsd);
-	rc += iwatch_add(&iw_conf, path, IN_ONLYDIR | IN_DONT_FOLLOW);
+	rc |= iwatch_add(&iw_conf, path, IN_ONLYDIR | IN_DONT_FOLLOW);
 	snprintf(path, sizeof(path), "%s/enabled/", finit_rcsd);
-	rc += iwatch_add(&iw_conf, path, IN_ONLYDIR | IN_DONT_FOLLOW);
-	rc += iwatch_add(&iw_conf, finit_conf, 0);
+	rc |= iwatch_add(&iw_conf, path, IN_ONLYDIR | IN_DONT_FOLLOW);
+	rc |= iwatch_add(&iw_conf, finit_conf, 0);
 
 	/*
 	 * Systems with /etc/default, /etc/conf.d, or similar, can also
 	 * monitor changes in env files sourced by .conf files (above)
 	 * define your own with --with-sysconfig=/path/to/envfiles
 	 */
-	rc += iwatch_add(&iw_conf, "/etc/default/", IN_ONLYDIR);
-	rc += iwatch_add(&iw_conf, "/etc/conf.d/", IN_ONLYDIR);
+	rc |= iwatch_add(&iw_conf, "/etc/default/", IN_ONLYDIR);
+	rc |= iwatch_add(&iw_conf, "/etc/conf.d/", IN_ONLYDIR);
 #ifdef FINIT_SYSCONFIG
-	rc += iwatch_add(&iw_conf, FINIT_SYSCONFIG, IN_ONLYDIR);
+	rc |= iwatch_add(&iw_conf, FINIT_SYSCONFIG, IN_ONLYDIR);
 #endif
+	rc |= conf_reload();
 
-	return rc + conf_reload();
+	return rc;
 }
 
 /*

src/exec.c

@@ -199,7 +199,6 @@ int run(char *cmd, char *log)
 int run_interactive(char *cmd, char *fmt, ...)
 {
 	int status, oldout = 1, olderr = 2;
-	char line[LINE_SIZE];
 	FILE *fp;
 
 	if (!cmd) {
@@ -244,6 +243,7 @@ int run_interactive(char *cmd, char *fmt, ...)
 
 	/* Dump any results of cmd on stderr after we've printed [ OK ] or [FAIL]  */
 	if (fp && !debug) {
+		char line[LINE_SIZE];
 		size_t len, written;
 
 		rewind(fp);
@@ -281,7 +281,7 @@ int exec_runtask(char *cmd, char *args[])
 	return execvp(_PATH_BSHELL, argv);
 }
 
-static void prepare_tty(char *tty, speed_t speed, char *procname, struct rlimit rlimit[])
+static void prepare_tty(char *tty, speed_t speed, const char *procname, struct rlimit rlimit[])
 {
 	char name[80];
 	int fd, dummy;
@@ -351,7 +351,7 @@ static int activate_console(int noclear, int nowait)
 	dprint(STDERR_FILENO, clr, strlen(clr));
 	dprint(STDERR_FILENO, msg, strlen(msg));
 	while ((rc = read(STDIN_FILENO, &ch, 1)) > 0 && ch != '\r' && ch != '\n')
-		continue;
+		;
 
 	/* On any error (likely EINTR), we avoid starting getty */
 	if (rc == -1)

src/finit.c

@@ -66,7 +66,7 @@ int   prevlevel = -1;
 int   debug     = 0;		/* debug mode from kernel cmdline */
 int   rescue    = 0;		/* rescue mode from kernel cmdline */
 int   single    = 0;		/* single user mode from kernel cmdline */
-int   bootstrap = 1;		/* set while bootrapping (for TTYs) */
+int   bootstrap = 1;		/* set while bootstrapping (for TTYs) */
 int   kerndebug = 0;		/* set if /proc/sys/kernel/printk > 7 */
 int   syncsec   = 0;		/* reboot delay */
 int   readiness = SVC_NOTIFY_PID;
@@ -280,7 +280,7 @@ static int fsck(int pass)
 				sulogin(1);
 			}
 		}
-		rc += fsck_rc;
+		rc |= fsck_rc;
 	}
 
 	endmntent(fp);

src/getty.c

@@ -312,9 +312,9 @@ static int exec_login(char *name)
 	struct stat st;
 
 	if (passenv)
-		execl(_PATH_LOGIN, _PATH_LOGIN, "-p", name, NULL);
+		execl(_PATH_LOGIN, _PATH_LOGIN, "-p", "--", name, NULL);
 	else
-		execl(_PATH_LOGIN, _PATH_LOGIN, name, NULL);
+		execl(_PATH_LOGIN, _PATH_LOGIN, "--", name, NULL);
 
 	/*
 	 * Failed to exec login, should not happen on normal systems.

src/helpers.c

@@ -87,7 +87,7 @@ void console_init(void)
 	/* Enable line wrap, if disabled previously, e.g., qemu */
 	dprint(STDOUT_FILENO, "\033[?7h", 5);
 
-	/* Reset atttributes, background and foreground color  */
+	/* Reset attributes, background and foreground color  */
 	dprint(STDOUT_FILENO, "\033[49m\033[39m\e[2J", 14);
 
 	log_init();

src/helpers.h

@@ -141,6 +141,7 @@ static inline int ln(const char *target, const char *linkpath)
 
 	return 0;
 }
+
 static inline int dprint(int fd, const char *s, size_t len)
 {
 	size_t loop = 3;
@@ -159,7 +160,7 @@ static inline int dprint(int fd, const char *s, size_t len)
 	return rc;
 }
 
-static inline char *fgetval(char *line, const char *key, char *sep)
+static inline char *fgetval(const char *line, const char *key, char *sep)
 {
 	char *ptr, *str, *copy;
 	size_t len;

src/initctl.c

@@ -872,7 +872,7 @@ static char *svc_environ(svc_t *svc, char *buf, size_t len, int ansi)
 {
 	int bold = missing(svc);
 
-	if (!ansi || svc_checkenv(svc) || svc->env[0] == '-')
+	if (!ansi || svc_checkenv(svc))
 		bold = 0;
 
 	strlcpy(buf, bold ? "\e[1m" : "", len);

src/log.c

@@ -139,7 +139,7 @@ void logit(int prio, const char *fmt, ...)
 		goto done;
 	}
 
-	fprintf(fp, "<%d>finit[1]:", LOG_DAEMON | prio);
+	fprintf(fp, "<%d>finit[1]: ", LOG_DAEMON | prio);
 	vfprintf(fp, fmt, ap);
 	fclose(fp);
 

src/pid.c

@@ -192,7 +192,7 @@ int pid_file_set(svc_t *svc, char *file, int not)
  * pid:foo.tla  --> /run/foo.tla.pid
  * pid:/tmp/foo --> /tmp/foo.pid        (Not handled by pidfile plugin!)
  *
- * An exclamation mark, or logial NOT, in the first character position
+ * An exclamation mark, or logical NOT, in the first character position
  * indicates that Finit should *not* manage the PID file, and that the
  * process uses a "non-standard" location.  Eg. if service is 'bar'
  * Finit would assume the PID file is /run/bar.pid and the below simply

src/schedule.c

@@ -23,6 +23,7 @@
 
 #include "config.h"
 #include "finit.h"
+#include "log.h"
 #include "schedule.h"
 
 #define SC_INIT 0x494E4954	/* "INIT", see ascii(7) */
@@ -35,6 +36,7 @@ static void cb(uev_t *w, void *arg, int events)
 	struct wq *work = (struct wq *)arg;
 
 	if (UEV_ERROR == events) {
+		dbg("%s(): spurious problem with schedule work timer, restarting.", __func__);
 		uev_timer_start(w);
 		return;
 	}

src/service.c

@@ -81,7 +81,12 @@ static void service_timeout_cb(uev_t *w, void *arg, int events)
 {
 	svc_t *svc = arg;
 
-	/* Ignore any UEV_ERROR, we're a one-shot cb so just run it. */
+	if (UEV_ERROR == events) {
+		dbg("%s(): spurious problem, svc %s.", __func__, svc_ident(svc, NULL, 0));
+		uev_timer_start(w);
+		return;
+	}
+
 	if (svc->timer_cb)
 		svc->timer_cb(svc);
 }
@@ -488,28 +493,21 @@ static void source_env(svc_t *svc)
 		while (*end && !isspace(*end))
 			end++;
 		if (*end != 0) {
-			warnx("'%s=%s': not a valid identifier", key, val);
+			warnx("'%s=%s': not a valid identifier", key, value);
 			continue;	/* invalid key */
 		}
 
-		switch (wordexp(value, &we, 0)) {
-		case 0:
+		if (wordexp(value, &we, 0)) {
+			setenv(key, value, 1);
+		} else {
 			for (i = 0, *val = 0; i < we.we_wordc; i++) {
 				if (i > 0)
 					strlcat(val, " ", LINE_SIZE);
 				strlcat(val, we.we_wordv[i], LINE_SIZE);
 			}
 			setenv(key, val, 1);
-			wordfree(&we);
-			break;
-
-		case WRDE_NOSPACE:
-			wordfree(&we);
-			/* fallthrough */
-		default:
-			setenv(key, value, 1);
-			break;
 		}
+		wordfree(&we);
 	}
 
 	fclose(fp);
@@ -917,12 +915,12 @@ static void service_kill(svc_t *svc)
 		return;
 	}
 
-	dbg("%s: Sending SIGKILL to process group %d", nm, svc->pid);
+	dbg("%s: Sending SIGKILL to process %d", nm, svc->pid);
 	logit(LOG_CONSOLE | LOG_NOTICE, "Stopping %s[%d], sending SIGKILL ...", id, svc->pid);
 	if (runlevel != 1)
 		print_desc("Killing ", svc->desc);
 
-	kill(-svc->pid, SIGKILL);
+	kill(svc->pid, SIGKILL);
 
 	/* Let SIGKILLs stand out, show result as [WARN] */
 	if (runlevel != 1)
@@ -1524,6 +1522,7 @@ int service_register(int type, char *cfg, struct rlimit rlimit[], char *file)
 	char *id = NULL, *env = NULL, *cgroup = NULL;
 	char *pre_script = NULL, *post_script = NULL;
 	char *ready_script = NULL, *conflict = NULL;
+	char *cleanup_script = NULL;
 	char ident[MAX_IDENT_LEN];
 	char *ifstmt = NULL;
 	char *notify = NULL;
@@ -1630,6 +1629,8 @@ int service_register(int type, char *cfg, struct rlimit rlimit[], char *file)
 			post_script = arg;
 		else if (MATCH_CMD(cmd, "ready:", arg))
 			ready_script = arg;
+		else if (MATCH_CMD(cmd, "cleanup:", arg))
+			cleanup_script = arg;
 		else if (MATCH_CMD(cmd, "env:", arg))
 			env = arg;
 		/* catch both cgroup: and cgroup. handled in parse_cgroup() */
@@ -1674,6 +1675,7 @@ int service_register(int type, char *cfg, struct rlimit rlimit[], char *file)
 		if (tty_parse_args(&tty, cmd, &args))
 			return errno;
 
+		/* NOTE: this may result in dev == NULL! */
 		if (tty_isatcon(tty.dev))
 			dev = tty_atcon();
 		else
@@ -1706,17 +1708,19 @@ int service_register(int type, char *cfg, struct rlimit rlimit[], char *file)
 		respawn = 1;
 
 		/* Create name:id tuple for identity, e.g., tty:S0 */
-		ptr = strrchr(dev, '/');
-		if (ptr)
-			ptr++;
-		else
-			ptr = dev;
-		if (!strncmp(ptr, "tty", 3))
-			ptr += 3;
+		if (dev) {
+			ptr = strrchr(dev, '/');
+			if (ptr)
+				ptr++;
+			else
+				ptr = dev;
+			if (!strncmp(ptr, "tty", 3))
+				ptr += 3;
 
-		name = "tty";
-		if (!id || id[0] == 0)
-			id = ptr;
+			name = "tty";
+			if (!id || id[0] == 0)
+				id = ptr;
+		}
 
 		svc = svc_find_by_tty(dev);
 	} else
@@ -1817,6 +1821,11 @@ int service_register(int type, char *cfg, struct rlimit rlimit[], char *file)
 		parse_script(svc, "ready", ready_script, &svc->ready_tmo, svc->ready_script, sizeof(svc->ready_script));
 	else
 		memset(svc->ready_script, 0, sizeof(svc->ready_script));
+	if (cleanup_script)
+		parse_script(svc, "cleanup", cleanup_script, &svc->cleanup_tmo, svc->cleanup_script, sizeof(svc->cleanup_script));
+	else
+		memset(svc->cleanup_script, 0, sizeof(svc->cleanup_script));
+
 	if (!svc_is_tty(svc)) {
 		if (log)
 			parse_log(svc, log);
@@ -1916,8 +1925,8 @@ int service_register(int type, char *cfg, struct rlimit rlimit[], char *file)
 }
 
 /*
- * This function is called when cleaning up lingering (stopped) services
- * after a .conf reload.
+ * This function is called at the end of a runlevel change or reload
+ * command to delete all remnants of removed services.
  *
  * We need to ensure we properly stop the service before removing it,
  * including stopping any pending restart or SIGKILL timers before we
@@ -1962,10 +1971,14 @@ void service_monitor(pid_t lost, int status)
 		/* If the setup phase fails, drive svc to crashed. */
 		svc->status = status;
 		/* fallthrough */
+	case SVC_TEARDOWN_STATE:
 	case SVC_CLEANUP_STATE:
 		dbg("collected script %s(%d), normal exit: %d, signaled: %d, exit code: %d",
-		   svc->state == SVC_CLEANUP_STATE ? svc->post_script : svc->pre_script,
-		   lost, ok, sig, rc);
+		    (svc->state == SVC_TEARDOWN_STATE
+		     ? svc->post_script
+		     : (svc->state == SVC_CLEANUP_STATE
+			? svc->cleanup_script
+			: svc->pre_script)), lost, ok, sig, rc);
 		/* Kill all children in the same proess group, e.g. logit */
 		dbg("Killing lingering children in same process group ...");
 		kill(-svc->pid, SIGKILL);
@@ -1994,7 +2007,7 @@ void service_monitor(pid_t lost, int status)
 	kill(-svc->pid, SIGKILL);
 
 	/* Try removing PID file (in case service does not clean up after itself) */
-	if (svc_is_daemon(svc) || svc_is_tty(svc)) {
+	if (svc_is_daemon(svc) || svc_is_sysv(svc) || svc_is_tty(svc)) {
 		service_cleanup(svc);
 	} else if (svc_is_runtask(svc)) {
 		/* run/task should run at least once per runlevel */
@@ -2134,13 +2147,16 @@ static void service_pre_script(svc_t *svc)
 
 		redirect(svc);
 
+		/* Warning in service_start() after svc_checkenv() */
 		env_file = svc_getenv(svc);
-		if (env_file)
+		if (env_file && fexist(env_file))
 			snprintf(buf, sizeof(buf), ". %s; exec %s", env_file, svc->pre_script);
 		else
 			strlcpy(buf, svc->pre_script, sizeof(buf));
 
 		set_pre_post_envs(svc, "pre");
+		sig_unblock();
+
 		execvp(_PATH_BSHELL, argv);
 		_exit(EX_OSERR);
 	}
@@ -2171,12 +2187,14 @@ static void service_post_script(svc_t *svc)
 		rc = WEXITSTATUS(svc->status);
 		sig = WTERMSIG(svc->status);
 
+		/* Warning in service_start() after svc_checkenv() */
 		env_file = svc_getenv(svc);
-		if (env_file)
+		if (env_file && fexist(env_file))
 			snprintf(buf, sizeof(buf), ". %s; exec %s", env_file, svc->post_script);
 		else
 			strlcpy(buf, svc->post_script, sizeof(buf));
 		set_pre_post_envs(svc, "post");
+		sig_unblock();
 
 		if (WIFEXITED(svc->status)) {
 			char val[4];
@@ -2201,7 +2219,7 @@ static void service_post_script(svc_t *svc)
 }
 
 /*
- * Unlike the pre: and post: scripts, the ready: script cannot re-use
+ * Unlike the pre: and post: scripts, the ready: script cannot reuse
  * svc->pid, so instead we use an assoc. list for matching any scripts
  * related to the svc, handled by service_script_add() below.  When a
  * ready: script is reaped the service_monitor() checks first for the
@@ -2230,13 +2248,16 @@ void service_ready_script(svc_t *svc)
 		};
 		char *env_file;
 
+		/* Warning in service_start() after svc_checkenv() */
 		env_file = svc_getenv(svc);
-		if (env_file)
+		if (env_file && fexist(env_file))
 			snprintf(buf, sizeof(buf), ". %s; exec %s", env_file, svc->ready_script);
 		else
 			strlcpy(buf, svc->ready_script, sizeof(buf));
 
 		set_pre_post_envs(svc, "ready");
+		sig_unblock();
+
 		execvp(_PATH_BSHELL, argv);
 		_exit(EX_OSERR);
 	}
@@ -2245,6 +2266,44 @@ void service_ready_script(svc_t *svc)
 	service_script_add(svc, pid, svc->ready_tmo);
 }
 
+static void service_cleanup_script(svc_t *svc)
+{
+	svc->pid = service_fork(svc);
+	if (svc->pid < 0) {
+		err(1, "Failed forking off %s cleanup:script %s", svc_ident(svc, NULL, 0), svc->cleanup_script);
+		return;
+	}
+
+	if (svc->pid == 0) {
+		char buf[CMD_SIZE];
+		char *argv[4] = {
+			"sh",
+			"-ac",
+			buf,
+			NULL
+		};
+		char *env_file;
+
+		redirect(svc);
+
+		/* Warning in service_start() after svc_checkenv() */
+		env_file = svc_getenv(svc);
+		if (env_file && fexist(env_file))
+			snprintf(buf, sizeof(buf), ". %s; exec %s", env_file, svc->cleanup_script);
+		else
+			strlcpy(buf, svc->cleanup_script, sizeof(buf));
+
+		set_pre_post_envs(svc, "cleanup");
+		sig_unblock();
+
+		execvp(_PATH_BSHELL, argv);
+		_exit(EX_OSERR);
+	}
+
+	dbg("%s: cleanup:script %s started as PID %d", svc_ident(svc, NULL, 0), svc->cleanup_script, svc->pid);
+	service_timeout_after(svc, svc->cleanup_tmo, service_kill_script);
+}
+
 static void service_retry(svc_t *svc)
 {
 	char *restart_cnt = (char *)&svc->restart_cnt;
@@ -2349,7 +2408,7 @@ static void svc_set_state(svc_t *svc, svc_state_t new_state)
 		}
 	}
 
-	if (svc_is_daemon(svc)) {
+	if (svc_is_daemon(svc) || svc_is_sysv(svc)) {
 		char cond[MAX_COND_LEN];
 
 		snprintf(cond, sizeof(cond), "service/%s/", svc_ident(svc, NULL, 0));
@@ -2432,10 +2491,12 @@ restart:
 
 	switch (svc->state) {
 	case SVC_HALTED_STATE:
-		if (enabled)
+		if (enabled) {
 			svc_set_state(svc, SVC_WAITING_STATE);
-		else {
-			if (svc_is_conflict(svc)) {
+		} else {
+			if (svc_is_removed(svc)) {
+				svc_set_state(svc, SVC_DEAD_STATE);
+			} else if (svc_is_conflict(svc)) {
 #if 0
 				logit(svc->nowarn ? LOG_DEBUG : LOG_INFO,
 				      "%s in conflict with %s, checking again ...",
@@ -2447,6 +2508,30 @@ restart:
 		}
 		break;
 
+	case SVC_TEARDOWN_STATE:
+		if (!svc->pid) {
+			dbg("%s: post script done.", svc_ident(svc, NULL, 0));
+			service_timeout_cancel(svc);
+
+			if (svc_is_removed(svc) && svc_has_cleanup(svc)) {
+				svc_set_state(svc, SVC_CLEANUP_STATE);
+				service_cleanup_script(svc);
+			} else
+				svc_set_state(svc, SVC_HALTED_STATE);
+		}
+		break;
+
+	case SVC_CLEANUP_STATE:
+		if (!svc->pid) {
+			dbg("%s: cleanup script done.", svc_ident(svc, NULL, 0));
+			svc_set_state(svc, SVC_HALTED_STATE);
+		}
+		break;
+
+	case SVC_DEAD_STATE:
+		/* End of the line ☠ */
+		break;
+
 	case SVC_DONE_STATE:
 		if (svc_is_changed(svc))
 			svc_set_state(svc, SVC_HALTED_STATE);
@@ -2469,8 +2554,11 @@ restart:
 			case SVC_TYPE_SYSV:
 			case SVC_TYPE_TTY:
 				if (svc_has_post(svc)) {
-					svc_set_state(svc, SVC_CLEANUP_STATE);
+					svc_set_state(svc, SVC_TEARDOWN_STATE);
 					service_post_script(svc);
+				} else if (svc_is_removed(svc) && svc_has_cleanup(svc)) {
+					svc_set_state(svc, SVC_CLEANUP_STATE);
+					service_cleanup_script(svc);
 				} else
 					svc_set_state(svc, SVC_HALTED_STATE);
 				break;
@@ -2489,12 +2577,12 @@ restart:
 		}
 		break;
 
-	case SVC_CLEANUP_STATE:
-		if (!svc->pid)
-			svc_set_state(svc, SVC_HALTED_STATE);
-		break;
-
 	case SVC_SETUP_STATE:
+		if (!enabled) {
+			service_stop(svc);
+			break;
+		}
+
 		if (!svc->pid) {
 			int rc = WEXITSTATUS(svc->status);
 			int ok = WIFEXITED(svc->status);
@@ -2596,7 +2684,7 @@ restart:
 				break;
 			}
 
-			if (svc_is_runtask(svc)) { /* only run/task, not sysv here */
+			if (svc_is_runtask(svc)) {
 				svc_set_state(svc, SVC_STOPPING_STATE);
 				svc->restart_tot++;
 				svc->once++;
@@ -2818,7 +2906,7 @@ void service_notify_cb(uev_t *w, void *arg, int events)
 	ssize_t len;
 
 	if (UEV_ERROR == events) {
-		warn("Spurious problem with %s notify callback, restarting.", svc_ident(svc, NULL, 0));
+		dbg("Spurious problem with %s notify callback, restarting.", svc_ident(svc, NULL, 0));
 		uev_io_start(w);
 		return;
 	}
@@ -2868,7 +2956,7 @@ void service_notify_cb(uev_t *w, void *arg, int events)
  *
  * This does not affect the restart_tot counter, which you can see in
  * the output from 'initctl status foo', along with this instability
- * "index" in parethesis: total (cnt/max)
+ * "index" in parenthesis: total (cnt/max)
  */
 static void service_interval_cb(uev_t *w, void *arg, int events)
 {
@@ -2876,6 +2964,7 @@ static void service_interval_cb(uev_t *w, void *arg, int events)
 
 	(void)arg;
 	if (UEV_ERROR == events) {
+		dbg("%s(): spurious problem, restarting.", __func__);
 		uev_timer_start(w);
 		return;
 	}

src/sig.c

@@ -232,6 +232,8 @@ void do_iterate_proc(int (*cb)(int, void *), void *data)
 					dbg("Skipping %s ...", file);
 				else if (file[0] == '@')
 					dbg("Skipping %s ...", &file[1]);
+				else if (strstr(file, "initctl"))
+					dbg("Skipping %s ...", file);
 				else
 					if (cb(pid, data)) {
 						dbg("PID %d is still alive (%s)", pid, file);

src/sm.c

@@ -107,6 +107,9 @@ static char *sm_status(sm_state_t state)
 	case SM_BOOTSTRAP_STATE:
 		return "bootstrap";
 
+	case SM_BOOTSTRAP_WAIT_STATE:
+		return "bootstrap/wait";
+
 	case SM_RUNNING_STATE:
 		return "running";
 
@@ -116,15 +119,20 @@ static char *sm_status(sm_state_t state)
 	case SM_RUNLEVEL_WAIT_STATE:
 		return "runlevel/wait";
 
+	case SM_RUNLEVEL_CLEAN_STATE:
+		return "runlevel/clean";
+
 	case SM_RELOAD_CHANGE_STATE:
 		return "reload/change";
 
 	case SM_RELOAD_WAIT_STATE:
 		return "reload/wait";
 
-	default:
-		return "unknown";
+	case SM_RELOAD_CLEAN_STATE:
+		return "reload/clean";
 	}
+
+	return "unknown";
 }
 
 static char sm_runlevel(int lvl)
@@ -201,7 +209,7 @@ restart:
 	switch (sm->state) {
 	case SM_BOOTSTRAP_STATE:
 		dbg("Bootstrapping all services in runlevel S from %s", finit_conf);
-		service_step_all(SVC_TYPE_RUNTASK | SVC_TYPE_SERVICE);
+		service_step_all(SVC_TYPE_RUNTASK | SVC_TYPE_SERVICE | SVC_TYPE_SYSV);
 
 		sm->state = SM_BOOTSTRAP_WAIT_STATE;
 		break;
@@ -235,7 +243,7 @@ restart:
 		/* Disable progress output at normal runtime */
 		enable_progress(0);
 
-		/* System bootrapped, launch TTYs et al */
+		/* System bootstrapped, launch TTYs et al */
 		bootstrap = 0;
 		service_step_all(SVC_TYPE_RESPAWN);
 		sm->state = SM_RUNNING_STATE;
@@ -319,6 +327,22 @@ restart:
 		sm->in_teardown = 0;
 		service_step_all(SVC_TYPE_ANY);
 
+		sm->state = SM_RUNLEVEL_CLEAN_STATE;
+		break;
+
+	case SM_RUNLEVEL_CLEAN_STATE:
+		/*
+		 * Wait for post:script or cleanup:script to be collected,
+		 * which moves the svc to HALTED or DEAD state.  We will
+		 * be called by the service_monitor() on collect.
+		 */
+		svc = svc_clean_completed();
+		if (svc) {
+			dbg("Waiting to collect post/cleanup script for %s, cmd %s(%d) ...",
+			    svc_ident(svc, NULL, 0), svc->cmd, svc->pid);
+			break;
+		}
+
 		/* Cleanup stale services */
 		svc_clean_dynamic(service_unregister);
 
@@ -372,6 +396,22 @@ restart:
 		dbg("Starting services after reconf ...");
 		service_step_all(SVC_TYPE_ANY);
 
+		sm->state = SM_RELOAD_CLEAN_STATE;
+		break;
+
+	case SM_RELOAD_CLEAN_STATE:
+		/*
+		 * Wait for post:script or cleanup:script to be collected,
+		 * which moves the svc to HALTED or DEAD state.  We will
+		 * be called by the service_monitor() on collect.
+		 */
+		svc = svc_clean_completed();
+		if (svc) {
+			dbg("Waiting to collect post/cleanup script for %s, cmd %s(%d) ...",
+			    svc_ident(svc, NULL, 0), svc->cmd, svc->pid);
+			break;
+		}
+
 		/* Cleanup stale services */
 		svc_clean_dynamic(service_unregister);
 

src/sm.h

@@ -29,9 +29,11 @@ typedef enum {
 	SM_BOOTSTRAP_WAIT_STATE,  /* Waiting for bootstrap to complete */
 	SM_RUNNING_STATE,         /* Normal state, services running */
 	SM_RUNLEVEL_CHANGE_STATE, /* A runlevel change has occurred */
-	SM_RUNLEVEL_WAIT_STATE,   /* Waiting for all stopped runlevel processes to be halted */
+	SM_RUNLEVEL_WAIT_STATE,   /* Waiting for all stopped processes to halt */
+	SM_RUNLEVEL_CLEAN_STATE,  /* Wait for post:scripts and cleanup:scripts */
 	SM_RELOAD_CHANGE_STATE,   /* A reload event has occurred */
-	SM_RELOAD_WAIT_STATE,     /* Waiting for all stopped reload processes to be halted */
+	SM_RELOAD_WAIT_STATE,     /* Waiting for all stopped processes to halt */
+	SM_RELOAD_CLEAN_STATE,    /* Wait for post:scripts and cleanup:scripts */
 } sm_state_t;
 
 typedef struct sm {

src/svc.c

@@ -348,6 +348,43 @@ svc_t *svc_stop_completed(void)
 	return NULL;
 }
 
+
+/**
+ * svc_clean_completed - Have post: and cleanup: scripts finished?
+ *
+ * Called late in runlevel and reload transitions to check if all post: and
+ * cleanup:scripts have completed.  A removed service should end up in the
+ * DEAD state, while a service with a post:script ends up in HALTED.
+ *
+ * Returns:
+ * %NULL if post: and cleanup: scripts have run, otherwise a pointer to
+ * the first found svc_t waiting for post: or cleanup:script.
+ */
+svc_t *svc_clean_completed(void)
+{
+	svc_t *svc, *iter = NULL;
+
+	for (svc = svc_iterator(&iter, 1); svc; svc = svc_iterator(&iter, 0)) {
+		if (svc_enabled(svc))
+			continue;
+
+		if (svc_is_runtask(svc) && svc->state == SVC_DONE_STATE)
+			continue;
+
+		if (svc_is_removed(svc)) {
+			if (svc->state == SVC_DEAD_STATE)
+				continue; /* any cleanup:script done */
+		} else {
+			if (svc->state == SVC_HALTED_STATE)
+				continue; /* any post:script done */
+		}
+
+		return svc;
+	}
+
+	return NULL;
+}
+
 /**
  * svc_find - Find a service object by its full path name
  * @name: Full path name, e.g., /sbin/syslogd
@@ -580,7 +617,7 @@ void svc_clean_dynamic(void (*cb)(svc_t *))
 	svc_t *svc, *iter = NULL;
 
 	for (svc = svc_iterator(&iter, 1); svc; svc = svc_iterator(&iter, 0)) {
-		if (svc->removed && cb)
+		if (svc_is_removed(svc) && cb)
 			cb(svc);
 	}
 }

src/svc.h

@@ -52,14 +52,16 @@ typedef enum {
 } svc_type_t;
 
 #define SVC_TYPE_ANY          (-1)
-#define SVC_TYPE_RESPAWN      (SVC_TYPE_SERVICE | SVC_TYPE_TTY)
-#define SVC_TYPE_RUNTASK      (SVC_TYPE_RUN | SVC_TYPE_TASK | SVC_TYPE_SYSV)
+#define SVC_TYPE_RESPAWN      (SVC_TYPE_SERVICE | SVC_TYPE_SYSV | SVC_TYPE_TTY)
+#define SVC_TYPE_RUNTASK      (SVC_TYPE_RUN | SVC_TYPE_TASK)
 
 typedef enum {
 	SVC_HALTED_STATE = 0,	/* Not allowed in runlevel, or not enabled. */
 	SVC_DONE_STATE,		/* Task/Run job has been run */
+	SVC_DEAD_STATE,		/* Process is dead and scheduled for removal */
+	SVC_CLEANUP_STATE,	/* Running cleanup: script */
+	SVC_TEARDOWN_STATE,	/* Running post: script */
 	SVC_STOPPING_STATE,	/* Waiting to collect the child process */
-	SVC_CLEANUP_STATE,	/* Running post: script */
 	SVC_SETUP_STATE,	/* Running pre: script */
 	SVC_PAUSED_STATE,	/* Condition is in flux, process SIGSTOPed */
 	SVC_WAITING_STATE,	/* Enabled but condition not satisfied */
@@ -146,7 +148,7 @@ typedef struct svc {
 	char           cond[MAX_COND_LEN];
 
 	/* Instance specifics */
-	int            job;	       /* For intenal use only, canonical ref is NAME:ID */
+	int            job;	       /* For internal use only, canonical ref is NAME:ID */
 	char           name[MAX_ARG_LEN];
 	char           id[MAX_ID_LEN]; /* :ID */
 	char           ifstmt[MAX_IDENT_LEN];
@@ -196,13 +198,19 @@ typedef struct svc {
 	char           conflict[MAX_ARG_LEN];
 	char	       desc[MAX_STR_LEN];
 	char	       env[MAX_CMD_LEN];
+
 	char	       pre_script[MAX_CMD_LEN];
 	int	       pre_tmo;
+
 	char	       post_script[MAX_CMD_LEN];
 	int	       post_tmo;
+
 	char	       ready_script[MAX_CMD_LEN];
 	int	       ready_tmo;
 
+	char	       cleanup_script[MAX_CMD_LEN];
+	int	       cleanup_tmo;
+
 	/*
 	 * Used to forcefully kill services that won't shutdown on
 	 * termination and to delay restarts of crashing services.
@@ -240,6 +248,7 @@ void	    svc_foreach	           (int (*cb)(svc_t *));
 void        svc_foreach_type       (int types, int (*cb)(svc_t *));
 
 svc_t	   *svc_stop_completed	   (void);
+svc_t      *svc_clean_completed    (void);
 
 void	    svc_mark		   (svc_t *svc);
 void	    svc_mark_dynamic       (void);
@@ -270,12 +279,14 @@ static inline int svc_has_pidfile  (svc_t *svc) { return svc_is_daemon(svc) && s
 static inline int svc_has_pre      (svc_t *svc) { return svc->pre_script[0];  }
 static inline int svc_has_post     (svc_t *svc) { return svc->post_script[0]; }
 static inline int svc_has_ready    (svc_t *svc) { return svc->ready_script[0];}
+static inline int svc_has_cleanup  (svc_t *svc) { return svc->cleanup_script[0]; }
 
 static inline void svc_starting    (svc_t *svc) { if (svc) svc->starting = 1;       }
 static inline void svc_started     (svc_t *svc) { if (svc) svc->starting = 0;       }
 static inline int  svc_is_starting (svc_t *svc) { return svc && 0 != svc->starting; }
 static inline int  svc_is_running  (svc_t *svc) { return svc && svc->state == SVC_RUNNING_STATE; }
 static inline int  svc_is_stopping (svc_t *svc) { return svc && svc->state == SVC_STOPPING_STATE; }
+static inline int  svc_is_stopped  (svc_t *svc) { return svc && svc->block == SVC_BLOCK_USER; }
 
 static inline int  svc_is_removed  (svc_t *svc) { return svc && svc->removed; }
 static inline int  svc_is_changed  (svc_t *svc) { return svc &&  0 != svc->dirty; }
@@ -390,12 +401,15 @@ static inline char *svc_status(svc_t *svc)
 			return "stopping";
 		}
 
-	case SVC_CLEANUP_STATE:
-		return "cleanup";
+	case SVC_TEARDOWN_STATE:
+		return "teardown";
 
 	case SVC_SETUP_STATE:
 		return "setup";
 
+	case SVC_CLEANUP_STATE:
+		return "cleanup";
+
 	case SVC_PAUSED_STATE:
 		return "paused";
 
@@ -408,6 +422,9 @@ static inline char *svc_status(svc_t *svc)
 	case SVC_RUNNING_STATE:
 		return "running";
 
+	case SVC_DEAD_STATE:
+		return "dead";
+
 	default:
 		return "UNKNOWN";
 	}

src/tty.c

@@ -172,7 +172,7 @@ int tty_parse_args(struct tty *tty, char *cmd, char **args)
 			passenv = 1;
 		else if (!strcmp(cmd, "rescue"))
 			tty->rescue = 1;	/* for rescue shells */
-		else if (whichp(cmd))		/* in $PATH? */
+		else if (!tty->cmd && whichp(cmd))
 			tty->cmd = cmd;
 		else
 			tty->args[tty->num++] = cmd;
@@ -229,8 +229,9 @@ int tty_parse_args(struct tty *tty, char *cmd, char **args)
 		return errno = EINVAL;
 	}
 
-	dbg("Registering %s getty on TTY %s at %s baud with term %s", tty->cmd ? "external" : "built-in",
-	   tty->dev, tty->baud ?: "0", tty->term ?: "N/A");
+	dbg("Registering %s getty on TTY %s at %s baud with term %s",
+	    tty->cmd ? "external" : "built-in",
+	    tty->dev, tty->baud ?: "0", tty->term ?: "N/A");
 
 	/* Built-in getty now comes as standalone program */
 	if (!tty->cmd) {

test/depserv.sh

@@ -1,5 +1,5 @@
 #!/bin/sh
-# Verify handling of dependant services, bug #314
+# Verify handling of dependent services, bug #314
 # Two services: foo and bar, where bar depends on foo:
 #
 #   - bar must not start until foo is ready

test/notify.sh

@@ -71,7 +71,7 @@ test_one()
     assert_status "serv" "running"
     assert_cond "service/serv/ready"
 
-    # Issue #343; pidfile.so inadvertedly marked systemd services as
+    # Issue #343; pidfile.so inadvertently marked systemd services as
     # 'started' when they created their PID file.  The serv test daemon
     # waits 3 sec between pidfile and notify to trigger this bug.
     if [ "$type" != "pid" ]; then

test/run-task-tricks.sh

@@ -1,7 +1,7 @@
 #!/bin/sh
 #
 # Verifies that a run task can call initctl at bootstrap and runlevel 2,
-# without blocking, performing serveral tasks:
+# without blocking, performing several tasks:
 #
 #  1. Query running services
 #  2. Stop/Start a running service