An external plugin to set up /etc can use HOOK_BASEFS_UP, but the Finit
conf_monitor() must then be called *after* not before that hook, or the
inotify watchers will listen on the wrong inode.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
The file plugin.h is installed to $prefix/include and used by external
plugins. Therefore it must not include, or depend on, features set by
the configure script in config.h
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
This is the first patch in a series to add support for
log:/path/to/logfile,priority:facility.level,tag:ident
In this patch support for 'priority:' and 'tag:', log file support
coming later.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
The SVC_RECONF, SVC_LOST, SVC_START, and RUNLEVEL_CHANGE hooks are not
one-shot, they are also not regular conditions since there exist no
mechanism to reset them from flux.
One idea was to turn them into actions, but the lost + start hooks need
to be called multiple times per trigger, e.g. `initctl reload`, which
turned out to be non-trivial to implement right now.
Therefore, for (at least) the Finit v3.1 release these conditions have
been disabled ("nop") and ignored by Finit. Only actual C-style plugins
will be called.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
This patch cleans up the implmentation, adds documentation and also
makes sure to adjust the prefix path of a pid:/prefix/path argument.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
The pidfile plugin just listens to changes in files in /run or /var/run,
so when trying to locate an svc_t from a file change we need to compare
using the proper prefix path.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
A service may now have a custom PID file, possibly because it doesn't
really create one itself. This needs to be taken into account also in
the pidfile plugin.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
This patch adds support for creating, and removing, a PID file for
programs (services) that for some reason does not create one. With
Finit this is very useful since it allows us to use our condition
support even for "broken" daemons.
To enable this for a service, use the following syntax:
service pid[:[/path/to/]file[.pid]] /path/to/daemon
I.e., `pid` is a required argument if you want to enable the feature,
but the rest is optional:
pid --> /run/`basename $CMD`.pid
pid:file --> /run/file.pid
pid:file.dip --> /run/file.dip
pid:/mnt/k --> /mnt/k.pid
pid:/mnt/k.foo --> /mnt/k.foo
There is an additional feature with this patch. For daemons that run
with a different identiy one usually wants to create the PID file with
that identity, e.g.:
service inadyn -I dyn -- Dyn.com DDNS client
service inadyn -I freedns -- FreeDNS DDNS client
Here inadyn will look for /etc/dyn.conf and create /run/dyn.pid, and
/etc/freedns.conf and create /run/freedns.pid, respetively. While Finit
will in both cases assume the PID file is /run/inadyn.pid ... err fail!
To get out of this mess the `pid` argument can be used to tell Finit
what the real PID file the daemon creates is called:
service pid:!dyn inadyn -I dyn -- Dyn.com DDNS client
service pid:!freedns inadyn -I freedns -- FreeDNS DDNS client
The same composition rules listed above applies also to the `!` variant.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
The service_step() function may call svc_del() on an svc from the
service_monitor(), so we need a way to tell the latter that it no
longer can dereference the pointer.
This regression is limited to v3.1, due to the introduction of the
new non-shared-memory store of svc_t.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
Per service rlimit is new in v3.1. When an inetd service connection is
created we do a selective copy from the parent inetd. We also need to
copy the parent's rlimits, otherwise the child will be killed as soon
as the kernel starts it.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
This patch increases the default wait, 2 --> 5 sec, while pending for
Finit + watchdogd controlled reboot/shutdown. We don't want to hand
back the prompt when we're not really ready yet.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
This patch builds on top of the recently introduced "silent by default"
and support for kernel cmdline `splash` to disable silent operation.
At shutdown/reboot `log_exit()` is called, the intention is to re-enable
progress output so the user can see what's going on. So, we need to re-
set `silent` to the same setting as log_init() started with, hence the
new `muffler()`.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
Automatic redirection of output was introduced in v3.0 and for the
upcoming v3.1 a configure option was added to disable it.
By popular demand this patch flips the default. One now has to
explicitly enable this feature to get redirection.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
This patch adds support for restarting the API socket on SIGHUP, similar
to how SysV init and systemd use SIGUSR1 to restart their FIFO or D-Bus.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
An interface may come up with IFF_RUNNING already set, but we would
only detect IFF_UP. This patch simplfies the code a lot by pivoting
the qualifier from using ifi_change to ifi_flags, which is the actual
current status of an interface.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
This is a quite intrusive change. In addition to the already reserved
runlevels 0 and 6, halt and reboot, respectively, this patch reserves
runlevel 1 as *no networking*, like the traditional SysV init did. One
of the reasons for this change is the new 'single' user mode, introduced
earlier. Most users when entering 'S' or 'single' on the kernel cmdline
expect the single user mode (runlevel 1) to be without networking.
This leaves 2, 3, 4, 5, and 7, 8, 9 as fully user-configurable runlevels
with networking.
Of course, networking can still be enabled in runlevel 1 using ifconfig
or ifupdown, or change runlevel.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
If hostname is missing from /lib/finit/rescue.conf the system will
default to the configured default hostname (noname).
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
This patch adds a rescue (revovery) mode to Finit. Simply add `rescue`
to the kernel cmdline at boot and the following steps at bootstrap will
be skipped:
- Load of services/run/task/etc in /etc/finit.conf + /etc/finit.d/*.conf
- fsck of filesystems in /etc/fstab
- Remount of / to read-write
- Mount of all filesystems in /etc/fstab
- swapon
- A few plugin hooks:
- HOOK_ROOTFS_UP
- HOOK_MOUNT_ERROR
- HOOK_BASEFS_UP
- `runparts DIR`
- /etc/rc.local
Instead of loading /etc/finit.conf and /etc/finit.d/*.conf Finit loads
the file /lib/finit/rescue.conf, which can be overridden by the operator
if needed. If this file is removed (or lost), Finit falls back to start
a simple root shell, without any login.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
The /lib/finit/plugins path is as of this patch not longer possible to
change. Use --prefix or --libdir instead to change the leading path.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
Finit v3.1 adds support for reloading /etc/finit.conf, document what
actually is read at runtime, compared to at bootstrap, and clarify
which stanzas are for what part of the boot process.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
The difference between `runparts DIR` and `/etc/rc.local` is that one
should run just before the first runlevel change (to the configured
runlevel) at the end of bootstrap, whereas the other should run just
before launcing the TTYs, i.e. after all tasks in the configured
runlevel have been started.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
This patch is an optimization. Now that we monitor for any .conf file
changes (as soon as the event loop starts) we can skip automatic reload
of all .conf files and services if no .conf file has changed. Unless
the user issues an `initctl reload`, `init q`, or sends us SIGHUP.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
As reported in #92, systems with an udevd very early register a service
using global_rlimit[]. The change was introduced in b68c5c1 but did
not take the ordering problem into consideration, global_rlimit[] was
left uninitialized, which likely led to severely limited udevd that
was continously crashing.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
- Validate function argument
- Skip¹ shutdown() when closing a TCP socket, not needed
___
¹ also, shutdown() segfaults a late shutdown/reboot
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
On shutdown/reboot, send SIGPR as a heads-up to watchdogd that the
system is going down. All file systems are about to be unmounted.
It's just about the last process left running in the system, so we
give it two seconds to complete its work before we continue.
When everything has been unmounted and we're ready to call reboot(),
we send SIGTERM to the registered watchdog, to give ti the chance to
let the WDT reset the system.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
When Finit enters shutdown/reboot we need to advise any watchdogd that
we are going down *before* we unmount any file systems. This to let
the user save any state, e.g. reboot counters, to disk. Right before
we call reboot() in Finit we send SIGTERM to the watchdogd to give it
the chance to use the WDT to reboot the system.
The WDT cricuitry on embedded systems is often connected to the RESETn
logic of all relevant board compoents, whereas reboot() only reboots the
SoC with the CPU. Hence, a WDT reboot is more efficient and more like a
regular power cycle, which users of embedded systems often want.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
This patch makes it possible to control the default redirection of
(misbehaving) services output to /dev/null. With this disabled,
a service may write to its stdout/stderr and mess up the console.
By default all service/run/tasks stdout/stderr is redirected.
Signed-off-by: Joachim Nilsson <troglobit@gmail.com>
Joachim Nilsson